Characteristics of Rubberised Bitumen

Binder Properties 505
Characteristics of Rubberised Bitumen

Blends
I. Widyatmoko* — R. Elliott* — J. Grenfell** — G. Airey** — A.

Collop** — S. Waite***
*Scott Wilson Limited

12 Regan Way, Chetwyn Business Park, Nottingham, NG9 6RZ, UK
Daru.widyatmoko@scottwilson.com
Richard.elliott@scottwilson.com
**Nottingham Transportation Engineering Centre

School of Civil Engineering, University Park, Nottingham, NG7 2RD, UK
James.grenfell@nottingham.ac.uk
Gordon.airey@nottingham.ac.uk
Andrew.collop@nottingham.ac.uk
***Waste & Resources Action Programme

The Old Academy, 21 Horse Fair, Banbury, Oxon, OX16 0AH, UK
Steve.waite@wrap.org.uk
ABSTRACT. This paper presents some findings from the recently completed research project
on “Rubberised Asphalt Testing to UK Standards”, funded by WRAP (Waste & Resources
Action Programme). One of the aims from this study was to provide information to potential
users of rubberised binder on the properties that can be expected from various sources/types
of crumb rubber available in the UK, when blended with typical UK bitumens, specifically
without using any additive. It was found that the selected three best blends satisfied target
requirements for ageing index, softening point and viscosity. In addition, the blends exhibited
improvements in short and long term ageing, cohesion characteristics, Fraass breaking point,
service temperature range and flash point values compared to typical penetration grade
bitumen, and satisfactory residual properties after storage stability testing.
KEYWORDS: ageing, storage stability, rheology, cohesion, resilience, rubberised bitumen.

506 Asphalt-Rubber 2009
1. Introduction
This paper presents some findings from the recently completed research project on
“Rubberised Asphalt Testing to UK Standards”, funded by WRAP (Waste & Resources Action
Programme). One of the aims from this study was to provide information to potential users of
rubberised binder on the properties that can be expected from various sources/types of crumb
rubber available in the UK, when blended with typical UK bitumens, particularly without
using any additive or oil extender. This paper describes the assessment of the rubberised
binders; a companion paper describes the assessment of the rubberised asphalt mixtures (WIDY
2009).
2. Materials for Testing
The base bitumen used to manufacture the rubberised bitumen blends comprised four
different binder types (two hard and two soft binders) that were produced from Venezuelan
and Middle Eastern crudes. Results of initial tests on these binders are presented in Table 1.
Table 1. Properties of Base Bitumen
Penetration Softening Viscosity Bitumen Composition

Bitumen at 25℃ Point at 177℃
Saturates Aromatics Resins Asphaltenes
(dmm) (℃) (cP)
(%) (%) (%) (%)
Venezuelan
54 52.2 109 7.0 55.9 20.4 16.7
40/60
Venezuelan
135 42.0 72 8.5 55.1 18.9 17.6
100/150
Middle
Eastern 50 52.5 80 3.4 60.9 21.2 14.4
40/60
Middle
Eastern 124 41.0 75 4.4 64.8 19.9 10.9
100/150
The rubber reclaimed from either car or truck tyres, from either ambient or cryogenic
production methods, were supplied in sizes available as “stock products”. Hereafter, rubber
reclaimed from car and truck tyres and produced at ambient temperature are called ambient
car (AC) and ambient truck (AT), respectively; rubber reclaimed from car tyre and produced at
cryogenic temperature is called cryogenic car (CC). Their particle size distributions (gradings)
were subsequently determined. The results are summarised in Table 2 (note: a question mark
“?” denotes the possibility of the grading falling within the stated category).
Table 2. Grading of Reclaimed Tyre Rubber (UK ‘Stock Products’)
Ambient Car
Type Cryogenic Car (CC) Ambient Truck (AT)
Target (AC)
Rubber Grading 0.4-0.63 0.63-1.4 1.4-2.0 Truck Truck Truck
Fine Coarse
Size mm mm mm No.30 No.12 4-6mm
Sieve
Size Cumulative % Passing
(mm)
5 - 100 100 100 100 100 100 100 100
3.35 - 100 100 100 100 100 100 100 92
2 100 100 100 100 100 80 100 100 61
1.18 65 – 100 100 85 12 100 9 100 77 12
0.600 20 – 100 100 23 6 88 0 87 41 2
0.300 0 – 45 71 5 2 29 0 7 34 2
0.063 0–5 2 0 0 1 0 0 0 2
PAS 107
- Fine - Granulate Fine? Granulate Fine? - Granulate?
Category
According to Publicly Available Specification (PAS) 107, rubber particles having sizes
of 1 – 10mm, 0 – 1mm, and 0 – 0.5mm are classified as granulate, powder and fine powder
respectively. These categories are also shown in the above table, for cross-reference purposes.
Some of the UK “stock products” presented in the above table do not comply with the PAS
107 categories. This is an issue that needs to be addressed by the recycled rubber industry,
and it is hoped that they can agree with standardised classes for rubber particle size (such as
that proposed in PAS 107) in the future. Nevertheless, it is expected that a blend of two rubber
sizes may be required in order to meet the target grading for use in rubberised bitumen.
3. Preliminary Screening Test
According to the ASTM D8 definition, rubberised bitumen is a blend of bitumen, reclaimed

tyre rubber and certain additives in which the rubber component is at least 15 percent by
weight of the total blend and has reacted in the hot bitumen sufficiently to cause swelling of
the rubber particles. Consequently, therefore, rubberised bitumen is not a homogenous blend;
traces of swelled rubber particles are still visible in the finished blend. Experience in the US,
Australia, South Africa and parts of Europe indicate that rubber is typically added at a rate
between 18 and 22% by weight of the total binder. Therefore, rubber additions of 15, 18.5
and 22% by weight of the total binder were adopted for this study, to cover the above range
of minimum and maximum rates of rubber addition. Four different sources of rubber (one
of which is a blend of two sources) and four different base bitumens were selected for this
project. Combinations of these variables resulted in a total of 48 rubberised bitumen blends.
As previously stated, the use of oil extender or additive was excluded from this study.
Each of the 48 rubberised bitumen blends was manufactured using a harmonised blending
procedure. This involved blending/mixing in a Silverson L4RT high shear apparatus, at 177℃
and 6000 rpm for 15 minutes and 1500 rpm for 165 minutes, resulting in 180 minutes (3
hours) of blending. The specification for rubberised binder adopted in this study required the
viscosity test results (at 177℃ and 20 rpm) of these blends to be between 1500cP – 5000cP.
Subsequently, viscosity measurement on the 48 rubberised bitumen blends and 4 unmodified
bitumen control (i.e. base bitumen) samples was carried out. The viscosity testing was
initially carried out to AASHTO; however, during the course of the testing, it was found that
another method, ASTM D6114 is more universally adopted for testing rubberised bitumen.
Consequently, the latter was adopted for the remainder of the tests.
A Brookfield Viscometer with a No. 27 spindle was used in the main viscosity test
programme at a speed of 20 rpm. In addition, a parallel test was also carried out using a
Bohlin viscometer using different spindle sizes and rotational speeds; however the results
are not presented in this paper. Variations in test results due to different test protocols and
arrangements have been observed. Subsequently, the main analysis was based on the results
obtained from the Brookfield Viscometer testing under the above test conditions. Testing
was performed on samples taken at 30, 60, 120 and 180 minutes from the commencement of
blending/mixing. In addition, penetration (BS EN 1426) and softening point (BS EN 1427)
tests were carried out on each blend at the end of the mixing period (180 minutes).
A target grading envelope for the rubber particles similar to that successfully used in the
US was adopted in this study. As mentioned previously, there were four different sources of
rubber i.e. 100% AC, 100% AT, 100% CC and a 50/50 blend of AC and AT; their particle size
gradations are summarised in Figure 1.
100
90 100% Ambient Car Tyre [50/50 blend fine/coarse]

100% Ambient Truck Tyre [No.12]
80
50%Ambient Car + 50%Ambient Truck Tyres
70 100% Cryogenic Car Tyre [0.63-1.4mm]
Upper Specification Limit
60
%Passing
Lower Specification Limit

50
40
30
20
10
0
0.01 0.1 1 10
Sieve Size (mm)
Figure 1. Target “Standard Size” Gradation for Rubber Particles
With the exception of rubber particles produced at ambient temperature from car tyre (AC),
appropriate sizes of the stock products available in the UK can be readily used to meet the
target “standard size” grading. For the AC tyre rubber, a 50/50 blend of fine and coarse car
tyre rubber was required in order to meet the grading requirements.
Results for the 48 blends are summarised according to the binder type and rubber
percentage; for example VE22 stands for Venezuelan bitumen mixed with 22% rubber and
ME22 for Middle Eastern bitumen mixed with 22% rubber. Some examples are represented in
Figure 2.
100000 100000
Viscosity (cP)
Viscosity (cP)
10000 ME22 10000 VE22
ME18.5 VE18.5
1000 ME15 1000 VE15
100 100
0 50 100 150 200 0 50 100 150 200
Blending Tim e (m ins) Blending Tim e (m ins)
40/60pen, Ambient Car 40/60pen, Ambient Car
Figure 2. Changes in viscosity with blending time and crumb rubber type
The viscosity test results were compared against target viscosity values between 1500cP
– 5000cP. In addition, one would expect the viscosity to rise as the rubber dissolved in the
bitumen and stabilise when solution was essentially complete. Where this did not occur in the
required time, or the viscosity fell with time, or was outside the permitted range, this bitumen/
rubber combination was not considered favourably. It should be noted that the sample size of
the test specimen and any lack of complete homogeneity in mixing could also lead to outlying
results.
Results from the penetration and softening point tests, carried out on samples taken at the
end of 180 minutes blending, are summarised Table 3 (note: 22% addition was also tested but
is not reported here). “Best” rubberised bitumen blends with “standard size” rubber particles
were subsequently selected based upon the viscosity (main criterion), together with the
penetration and softening (additional criteria) test results. The following “best” blends were
subjected to further assessment:
• 81.5% ME 40/60 + 18.5% AC
• 81.5% ME 40/60 + 18.5% CC
• 81.5% VE 40/60 + 18.5% AC
• 81.5% VE 40/60 + 18.5% CC
• 85% VE 40/60 + 15% AT
• 81.5% VE 100/150 + 18.5% CC

Table 3. Penetration and softening point tests rubberised bitumen
15% Rubber/85%Bitumen 18.5% Rubber/81.5%Bitumen

Binder Sample
AC/AT
AC/AT
50:50
50:50
AC
AC
CC
CC
AT
AT
Penetration at 25℃ (dmm)
Venezuelan 40/60 36 32 54 39 30 34 31 30
Venezuelan 100/150 69 54 67 61 - 41 52 66
Middle East 40/60 44 41 44 36 35 70 47 46
Middle East 100/150 63 77 91 71 74 124 80 91
Softening Point (℃)
Venezuelan 40/60 67.5 67.3 52.2 64.5 72 76.5 72 70.2
Venezuelan 100/150 59 59 59.8 57 - 71 61 63
Middle East 40/60 60.4 68.4 64 63 67.6 62.8 64 63.8
Middle East 100/150 54.8 58.2 51 55 55.6 52.4 55.8 51.4
4. Effect of rubber size
Since the adopted “standard” rubber particle size was already on the coarser side of the
grading envelope, the effect of using a finer rubber particle size (still within the grading
envelope) was evaluated using a similar suite of testing. The viscosity measurement of each
rubberised bitumen blend at 177℃ and 20 rpm was carried out by using the Brookfield
Viscometer on samples taken at 30, 60, 120 and 180 minutes after the commencement of
blending/mixing.
In this exercise, both maximum (peak-ASTM D6114) and “stabilised” (AASHTO

T316) viscosity values were recorded as “fine-peak” and “fine-stab” respectively, as well as
penetration test (BS EN 1426) and softening point test (BS EN 1427) results. Some examples
of the test results are represented in Figure 3. The viscosity results for blends using “standard”
size rubber are also reproduced in this figure to ease the comparison.
100000 100000
Viscosity (cP)
Viscosity (cP)
10000 10000
Brookfield
Brookfield
Viscometer
1000 1000 Viscometer
100 100
0 50 100 150 200 0 50 100 150 200
VE18.5-standard VE18.5-standard
Blending Time (mins) VE18.5-fine-peak Blending Time (mins) VE18.5-fine-peak
VE18.5-fine-stab VE18.5-fine-stab
40/60pen, Ambient Car 40/60pen, Cryogenic Car
Figure 3. Changes in viscosity with blending time and crumb rubber size
Finer rubber particle size, having larger surface area than the ‘standard’ size, might be
expected to yield greater reaction and interaction with base bitumen and ultimately lead to the
increased viscosity observed in the above test results. All blends incorporating a finer rubber
particle size exceeded the maximum viscosity values of 5000 cP; consequently none of them
was considered suitable for adoption in the next stage of assessment. Therefore, the options
were limited to those incorporating the “standard” size rubber as used in the preliminary tests.
The following three best blends (all incorporating the “standard” size rubber), were selected:
• 84% ME 40/60 + 16% AC
• 81.5% VE 40/60 + 18.5% AC
• 81.5% VE 40/60 + 18.5% CC
The above three blends were subsequently subjected to the following comprehensive tests:
(i) determination of Ageing Index (BBA HAPAS SG4 protocol); (ii) penetration testing at 4℃
(BS EN 1426); (iii) softening point testing (BS EN 1427); (iv) rotational viscosity at 177℃
(ASTM D6114); and, (v) resilience at 25℃ (BS EN 13880-3). The results obtained from this
assessment are detailed in Table 4.
Table 4. Comprehensive Test Results
Binder Sample Target

values for
Tests 100% ME 100% VE 16% AC 18.5% AC 18.5% CC rubberised
40/60 (Base 40/60 (Base + 84% + 81.5% + 81.5% bitumen with
Bitumen) Bitumen) ME40/60 VE40/60 VE40/60 50/70 base
bitumen
Ageing Index 5.28 7.31 3.16 2.16 2.18 < 15
G* at 25℃ (Unaged),
6.59E+05 3.98E+05 6.64E+05 8.33E+05 7.67E+05 -
Pa
G* at 25℃ (After
3.48E+06 2.91E+05 2.10E+06 1.80E+06 1.67E+06 -
RTFOT and HiPAT), Pa
Penetration (4℃, 200g,
- - 29 20 19 > 15
60s), dmm
Softening Point, ℃ 51 50 63 72 70.2 > 54
Rotational Viscosity
80 109 2550 4630 4510 1500-5000
(177℃), cP
Resilience (25℃), % 0 5 13 29 26 >25
Table 4 shows that ageing the bitumen progressively increased the G* value (complex
shear (or stiffness) modulus at a given temperature and loading rate), hence indicating a
stiffened or hardened binder. A road paving bitumen typically used in the UK is expected to
have an Ageing Index (the ratio between G*HiPAT and G*Unaged at 0.4 Hz) of less than
15; generally the lower the value, the less susceptible the binder would be expected to be
to ageing. The level of increase in softening point, viscosity and resilience, and the reduced
penetration values, can be linked to the increase in rubber content in these blends. With respect
to the typical target reference values for rubberised bitumen incorporating 50/70pen base
bitumen, the test results indicate that the selected blends satisfied the requirements for ageing
index, softening point and viscosity. Furthermore, both rubberised blends with Venezuelan
bitumen met the criterion for resilience; however, the Middle Eastern blend fell short of the
recommended resilience value although remained higher than that of the base bitumen.
5. BS EN 14023 Tests on the Three Best Performing Blends
The three preferred rubberised bitumen blends selected above were further assessed in
accordance with BS EN 14023 (i.e. the framework specification of polymer modified binder)
but excluding the elastic recovery tests. Test results are presented in Table 5.
Table 5. Binder Test Results
Binder Sample
Tests Methods 16%AC + 18.5% AC + 18.5% CC +
84%ME40/60 81.5%VE40/60 81.5%VE40/60
Penetration (25℃, 100g, 5s), dmm EN 1426 35 30 30
Softening Point, ℃ EN 1427 63 72 70.2
SHW
Cohesion - Vialit pendulum, J/cm2 1.9 1.4 1.5
Clause 939
Temperature at Peak Cohesion, ℃ 50 50 53
Temperature range at cohesion
>55 50 50
above 0.5 J/cm2, ℃
RTFOT, Change of mass, % EN 12607-1 -0.16 -0.23 -0.21
RTFOT, Penetration (25℃, 100g,
EN 1426 24 24 27
5s), dmm
RTFOT, retained penetration, % - 69% 80% 90%
RTFOT, softening point, ℃ EN 1427 65 79.5 73
RTFOT, increase in softening point, ℃ - 2 7.5 2.8
EN ISO
Flash Point, ℃ >300 298 >300
2592
Fraass breaking point, ℃ EN 12593 -12 -8 -11
EN 14023
Plasticity range, ℃ Sub-clause 75 80 81.2
5.1.9
Storage stability, penetration at the EN 13399,
39 35 30
top, dmm EN 1426
Storage stability, penetration at the EN 13399,
41 33 32
bottom, dmm EN 1426
Storage stability, difference in
- 2 2 2
penetration, dmm
Binder Sample
Tests Methods 16%AC + 18.5% AC + 18.5% CC +
84%ME40/60 81.5%VE40/60 81.5%VE40/60
Storage stability, softening point at EN 13399,
56.5 71.5 63
the top, ℃ EN 1427
Storage stability, softening point at EN 13399,
67 76 77
the bottom, ℃ EN 1427
Storage stability, difference in
- 10.5 4.5 14
softening point, ℃
HiPAT, Penetration (25℃, 100g,
EN 1426 22 18 24
5s), dmm
HiPAT, retained penetration, % - 63% 60% 80%
HiPAT, softening point, ℃ EN 1427 71 90 83.5
HiPAT, increase in softening point, ℃ - 8 18 13.3
Table 5 indicates that the increased softening point and the reduced penetration values
after the short-term ageing test (RTFOT) can be linked to the increased rubber content of
these blends. A further reduction in penetration and increase in softening point of the residual
binders after the long-term ageing test (HiPAT) is also found, consistent with the trend
observed on the RTFOT samples.
Cohesion is commonly accepted as an indicator of a bitumen’s ability to resist tensile

stress under rapid impact loading, such as due to skidding or vehicle braking, and hence the
ability to retain aggregate intact within an asphalt material (resist fretting). The higher the
cohesion value, the better the aggregate retention. Another related parameter, the temperature
range over which the cohesion is at or above 0.5J/cm2 (TR) is commonly considered to be an
indicator of service temperature range; the higher the value, the wider the service temperature
range of a bituminous binder. The peak cohesion values and the temperature range over which
cohesion values are at or above 0.5 J/cm2 for the rubberised bitumen blends are much higher
than (at least double) those of typical 40/60pen bitumen, indicating significantly improved
cohesion characteristics for these blends, implying an improved ability to retain aggregate.
Fraass breaking point (FBP) is the temperature at which thin bitumen films start to show
cracks under a small deflection. The lower the FBP value, the better the low temperature
performance of the binder. The test results show that the rubberised bitumen blends had
FBP values not greater than -8℃. This is similar to that measured for a typical UK 40/60
penetration binder (around -7℃), suggesting that the addition of rubber does not have a
detrimental effect on the low temperature properties of the binder.
The plasticity range is the numerical difference between the softening point and Fraass
breaking point, which is sometimes considered as the range of service temperatures that a
binder may be expected to perform satisfactorily over. The plasticity ranges of the rubberised
bitumen blends (i.e. 75℃ or higher) will meet the requirements for BS EN 14023 polymer
modified binder Classes 2, 3 and 4, and are considered wider than those found typically for
a 40/60pen bitumen (i.e. around 60℃). This suggests the rubberised bitumen blends can be
expected to have a relatively wide service temperature range.
As with any other bituminous material, rubberised bitumen will release combustible fumes
when heated to sufficiently high temperatures. The flash point provides an indication of the
temperature at which a heated bituminous sample will instantaneously flash in the presence
of an open flame. The flash point values of the rubberised bitumen blends comply with those
generally specified for polymer modified binders, i.e. higher than 250℃. Thus the risk that
these blends could cause a fire during material production can be expected to be no greater
than that for polymer modified bitumens. However, the blends were observed to release
visibly thick white fume with strong odour during laboratory testing, specifically when the
test temperature was increased above 200℃. It is recommended that the significance of the
presence of this fume is assessed in more detail.
The results on the stability of the blends during hot storage suggest only a marginal
difference in the penetration value of the top and bottom part of the storage container, whilst
there is up to a 25% difference in the softening point values. Whilst the latter difference is
significantly greater than that expected from 40/60pen bitumen or some polymer modified
bitumens, the results are considered not unusual and consistent with those reported in the
USA for similar materials. Rubberised bitumen blends are known to be susceptible to phase
separation (being a two phase, non-homogeneous, blend) and/or devulcanization (thinning due
to rubber degradation) after prolonged hot storage.
6. Concluding Remarks
The preliminary assessment highlighted the sensitivity of rubberised bitumen to blending

arrangements and test conditions (specifically viscosity testing). Factors that may affect the
test results include: blending protocol (duration, shear rate), viscometer (type/make, shear
rate, spindle size) and viscosity data interpretation (peak value/ASTM D6114, stabilised
value/AASHTO T316). For this research project, harmonised blending and test conditions
were adopted, i.e. high shear blending at 6000 rpm for 15 minutes, with rubber being fed in
during the first 10 minutes, followed by reducing the shear rate to 1500 rpm for the rest of the
remaining blending time of 180 minutes. A Brookfield Viscometer with spindle size No. 27
was used in the main test program.
A total of 48 rubberised bitumen blends of different compositions were manufactured and

subjected to viscosity (ASTM D6114), penetration (BS EN 1426) and softening point (BS
EN 1427) testing. None of these blends contained any additive or oil extender. The viscosity
test results were compared against target viscosity values between 1500cP – 5000cP and,
based upon the viscosity criterion and consideration from the penetration and softening
point test results, the ‘best’ six blends were selected for further assessment, of the effect
of rubber particle size on the blend properties. During this exercise, both maximum (peak-
ASTM D6114) and “stabilised” (AASHTO T316) viscosity values were recorded as well as
penetration and softening point test data. It was found that finer particle sizes, which have
a larger surface area, led to greater reaction and interaction with base bitumen and hence a
higher viscosity of the rubberised bitumen. Three best performing rubberised bitumen blends
were selected, all containing UK “standard” size rubber:
• 84% ME 40/60 + 16% ambient car tyre rubber

• 81.5% VE 40/60 + 18.5% ambient car tyre rubber
• 81.5% VE 40/60 + 18.5% cryogenic car tyre rubber
The selected blends satisfied the project requirements for ageing index, softening point
and viscosity and both rubberised blends with Venezuelan bitumen met the criterion for
resilience. The Middle Eastern blend fell short of the recommended resilience value, although
its performance was better than that of unmodified binder.
Short-term and long-term ageing test results generally suggest improved properties of
the rubberised bitumens over those expected for unmodified 40/60pen bitumen. The peak
cohesion values and the temperature range over which cohesion values are at or above 0.5
J/cm2 for the rubberised bitumen blends are much higher than those for typical 40/60pen
bitumen, indicating significantly improved cohesion characteristics for these blends and hence
improved ability to retain aggregate. Fraass breaking point values for the rubberised bitumen
were at least similar to or better than those for 40/60pen bitumen.
The plasticity ranges of the rubberised bitumen blends met the requirements for BS EN
14023 polymer modified binder Classes 2, 3 and 4, and were wider than those found typically
for 40/60pen bitumen. This suggests that the rubberised bitumen blends can be expected to
have a relatively wide service temperature range. The flash point values of the rubberised
bitumen blends comply with those specified for polymer modified binders, although the
blends were observed to release visibly thick white fume with strong odour during testing,
specifically when the test temperature was increased above 200℃. It is recommended that the
significance of the presence of this fume is assessed in more detail.
The results on the stability of the blends during hot storage suggested only a marginal
difference in the penetration value of the top and bottom part of the storage container and
a difference of up to 25% in the softening point values. However, this was not considered
unusual and is consistent with results reported in the USA for similar materials.
7. Acknowledgement
The authors would like to compliment and thank Ronald Kibuuka and the laboratory teams
in Scott Wilson and NTEC for the efficiency and effectiveness in producing the test data
in this paper. The work was won from WRAP by Scott Wilson in partnership with NTEC,
by competitive tender. WRAP helps individuals, businesses and local authorities to reduce
waste and recycle more, making better use of resources and helping to tackle climate change;
more information on WRAP’s work can be found on www.wrap.org.uk. Contributions from
Aggregate Industries, Tarmac, Nynas, UK Bitumen, Moncton Rubber Technology, Tyregenics,
Allcock and Sons, and Singleton Birch, in providing the materials and the streams of advice
from the Technical Advisor to WRAP, Mr Ian Walsh, are greatly appreciated.
8. Bibliography
APRG (1999). “The Use of Recycled Crumb Rubber”, Australian Pavement Research Group
(APRG) Technical Note 10.
Austroads (2006). “Specification framework for polymer modified binders and multigrade
bitumens”, Austroads Technical Report AP-T41/06.
BSI, “Specification for the manufacture and storage of size reduced tyre materials”, Publicly
Available Specification (PAS) 107, British Standard Institution.
CalTrans, (2003). Asphalt Rubber Usage Guide, State of California Department of

Transportation, Materials Engineering and Testing Services-MS #5, January.
COLTO (1998). Standard Specifications for Roads and Bridge Works for State Road
Authorities. Committee of Land Transport Officials-South Africa.
Hanson, D.M., and Duncan, G.M. (1995). “Characterisation of Crumb Rubber-Modified

Binder Using Strategic Highway Research Program Technology”, Transportation Research
Record 1488.
Heitzman, M. (1992). “Design and Construction of Asphalt Paving Materials with Crumb
Rubber Modifier”, Transportation Research Record 1339.
Hicks, R.G., and Epps, J.A. (2000). “Quality control for asphalt rubber binders and mixes”,
Report prepared for Rubber Pavement Association, Arizona, 9 June.
Neto, S.A.D., Farias, M.M, Pais, J.C., and Pereira, P.A.A. (2003a), “ Optimization of asphalt
rubber hot mixes based on performance laboratory tests”., Asphalt Rubber 2003.
Neto, S.A.D., Farias, M.M., Pais, J.C., Pereira, P.A.A., and Santos, L.P. (2003b), “Properties
of asphalt-rubber binders related to characteristics of the incorporated crumb rubber”,
Asphalt Rubber 2003.
Oliver, J.W.H. (1981). “Modification of Paving Asphalts by Digestion with Scrap Rubber”,
Transportation Research Record 821, Transportation Research Board, 1981.
Oliver, J.W.H. (1982). “Optimising the improvements obtained by the digestion of

comminuted scrap rubbers in paving asphalts”, Proceedings of AAPT, vol. 51, pp. 169-188.
Oliver, J.W.H. (1997). ”Crumb rubber asphalt fatigue study phase 1: Binder testing”. Contract
report CR IC6496C. ARRB Transport Research Ltd., Australia, November.
Potgieter, C.J (2004). “Bitumen Rubber Asphalt in South Africa Conventional Techniques”,
3rd Eurasphalt & Eurobitume Congress, Vienna, Paper 274.
SABITA (1997). Technical guidelines for the specification and design of bitumen-rubber
asphalt wearing courses. Manual 19. ISBN: 1-874968-13-6.
Widyatmoko, I., Elliott, R., Grenfell, J., Airey, G., Collop, A., and Waite, S. (2009), “Laboratory
Assessment of Workability, Performance and Durability of Rubberised Asphalt Mixtures”,
Asphalt Rubber 2009, Nanjing, November 2-4. (In submission)
Takallou, H.B., and Sainton, A. (1991). “Advances in Technology of Asphalt Paving Materials
Containing Used Tire Rubber, Transportation Research Record 1339, pp. 23-29.
Visser A.T. and B. Verhaege (2000). “Bitumen-rubber: Lessons learned in South Africa”.
Asphalt Rubber 2000, Portugal, November. ISBN: 972-95240-9-2.
Mechanical properties of HMA mixes using

crumb-rubber modified binders
Fernanda Pilati - Adalberto Leandro Faxina - Ana Paula Furlan

-Glauco Tulio Pessa Fabbri - Antonio Carlos Gigante
*Department of Transportation, Sao Carlos Engineering School

Sao Paulo University, Sao Carlos, Brazil
fepilati@sc.usp.br, alfaxina@sc.usp.br, afurlan@sc.usp.br, glauco@sc.usp.br, gigante@
sc.usp.br
ABSTRACT: This research aims at evaluating the effects of crumb rubber and shale-oil
residue on some mechanical properties of HMA mixes. Eight HMA mixes were produced,
using asphalt binders compounded with different proportions of crumb rubber and shale-
oil residue, in order to evaluate the effect of these two modifiers on resilient modulus and
tensile strength values. The overall results pointed out that both crumb rubber and shale-oil
residue reduce resilient modulus, but oil shows a more expressive effect. Both crumb rubber
and shale-oil residue reduce tensile strength with similar intensity. Response-surface analysis
indicated that crumb-rubber modified binders formulated with smaller proportions of crumb
rubber and shale-oil residue lead to HMA mixtures of higher tensile strength values.
KEY WORDS: crumb-rubber modified binder, shale-oil residue, resilient modulus, tensile
strength, statistical analysis
1. Introduction
Understanding how modifiers affect physical properties of asphalt binders and HMA mixes
is the key-point to pick out adequate modifier proportions. The technique of experiments with
mixtures (Cornell, 2002) has been used with success to formulate crumb-rubber modified
binders (Faxina et al., 2006a; Faxina et al., 2006b; Faxina et al., 2008a; Faxina et al., 2008b),
likewise the applications in industries of medicines, food and others, when the focus is to
model the effects that component proportions have on some properties. Modeling the effects
that modifier proportions have on properties of interest for paving industry allows one to
predict new values for non-tested conditions, avoiding extra labour to prepare samples in
specific settings and facilitating the development of researches on new materials.
Incorporating crumb rubber to asphalt binder makes the asphalt binder stiffer and
more elastic, what can make HMA mixes more resistant to rutting and fatigue and thermal
crackings. Although an asphalt binder of higher performance is obtained, workability is
usually reduced. Enhancement of asphalt binder performance is directly related to rubber
concentration, but only until a certain percentage, once viscosity at mixing and compaction
temperatures can reach impracticable levels.
Adding aromatic oils to crumb-rubber modified binders is a practical alternative to reduce

viscosity during mixing and compaction operations of HMA mixes, improving workability. On
the other hand, aromatic oils bring the asphalt binder consistence down at room temperatures,
what may damage its quality and, as a consequence, drop the performance of HMA mixes off.
Although aromatic oils are volatile, only a small amount of oil volatilizes during construction
and asphalt layer lifespan. Selecting appropriate proportions of extender oil is crucial in
obtaining crumb-rubber modified binders of good quality.
As an aromatic oil, shale-oil residue can play the role of an extender oil. But appropriate
proportions of this modifier must be chosen, depending on rubber content and asphalt-binder
grade, in order to obtain crumb-rubber modified binders which combine workability and good
mechanical response. Adding extender oils to asphalt binder reduces to a typical formulation
problem.
This research aims at evaluating the effects that the variation of the proportions of crumb
rubber and shale-oil residue, used as an extender oil, have on resilient modulus and tensile
strength of HMA mixes. Eight asphalt binders were used to produce specimens that were
submitted to resilient modulus and tensile strength tests. Regression models were fitted and,
based on these models, response-trace plots and response-surface plots were generated, in
order to evaluate both principal and interaction effects among asphalt binder, crumb rubber
and shale-oil residue.
2. Effects of crumb rubber and extender oil on resilient modulus and tensile strength
Epps (1994) pointed out that HMA mixtures containing rubber from discarded tires
can show resilient modulus higher or smaller than that obtained from asphalt mixtures
compounded with conventional asphalt binder, but typically smaller values are obtained
when crumb-rubber modified binders are used. Ayres and Witczak (1995) verified that asphalt
mixtures containing crumb-rubber modified binder show resilient modulus values from 10 to
20% smaller than that obtained from the reference mixture (AC 20). These authors pointed out
that resilient modulus values reduce as rubber proportion increases.
Leite et al. (2000) concluded that HMA mixtures containing crumb-rubber modified binder
have shown a reduction of 50% in resilient modulus compared to the reference mixture (AC
20). Faxina and Soria (2003) observed reduction of resilient modulus of HMA mixtures
containing crumb-rubber modified binder and shale-oil residue, showing that addition
of these modifiers makes asphalt mixtures more flexible. Cordeiro (2006) has come to
conclusions similar to those obtained by the last two authors. Lemes (2004) concluded that the
resilient modulus of asphaltic mixtures increases when mixtures are submitted to short-term
accelerated aging.
Faxina and Soria (2003) showed that tensile strength of HMA mixes with crumb-rubber
modified binder and shale-oil, at 25℃, reduces expressively in comparison to the reference
mixture (approximately 50%). Pinheiro et al. (2003) also obtained tensile strength values 40%
smaller when crumb-rubber modified binder was used. Other researches performed in Brazil
(Momm and Salini, 2000; Specht et al., 2003) also observed reduction of tensile strength of
HMA mixtures containing crumb-rubber modified binder. Hanson et al. (1994) got similar
tensile strength values for asphalt mixtures compacted in laboratory and cores extracted from
the road.
Bertollo et al. (2003) also concluded that tensile strength reduces as rubber proportion
increases and pointed out that rubber gradation, as used in their research, does not affect
tensile strength values. On the other hand, Leite et al. (2000) obtained tensile strength values
for an asphalt mixture using crumb-rubber modified binder 26% higher than the tensile
strength of the reference mixture (AC 20).
3. Design of the experiment
The HMA mixtures were produced with asphalt binders formulated according to an
experiment designed with the statistical tool of experiments with mixtures (Cornell 2002),
taking into account restrictions in component proportions. Denoting x1 as asphalt proportion,
x2 as rubber proportion and x3 as oil proportion, restrictions imposed in percentage are: 68≤x1
≤100, 0≤x2≤18 and 0≤x3≤14. The following restrictions also must be met: x1 + x2 + x3 = 100
and x1, x2, x3 > 0. Figure 1 shows the simplex and the sub-region where the selected mixtures
are indicated. Design mixtures are: 100-0-0, 93-0-7, 86-0-14, 91-9-0, 77-9-14, 82-18-0,
75-18-7 and 68-18-14, where the numbers represent, from left to right, the proportions of
asphalt binder, crumb rubber and shale-oil residue.
Mineral aggregate is not considered a fourth component, what would make the experiment
much more complex. As the proportion of mineral aggregate in a specimen, for a certain
binder content, does not change and each model is defined to binder contents that are
controlled, it is supposed that the proportions of mineral aggregate and asphalt binder are
constant for the eight asphalt mixtures, for each binder content. Due to this assumption, it is
possible to assess the effect of varying the proportions of asphalt, rubber and oil on resilient
modulus and tensile strength, supposing that the effects on these properties are due exclusively
to the variations of the proportions of the three components.
(a) (b)
Figure 1. (a) Simplex, restrictions and selected mixtures and (b) detailed view of the
constrained region with selected mixtures
4. Materials
The crumb rubber used is a commercial product obtained from tires of heavy vehicles
(tractors and trucks). Its particle-size distribution is shown in Table 1. Thermogravimetry
analysis indicated the following composition: 64.3% elastomers and plasticizers, 31.4%
carbon black and 4.3% inorganic materials. Base asphalt binder (AC-40) is graded as a PG
70-10. A shale-oil residue type AR-5 according to ASTM D4552 was used. Mineral aggregate
was obtained from crushed basaltic rock. Aggregate gradation is one that coincides with the
middle of the specification “C” for dense-graded mixes of the paving handbook of the Sao
Paulo State Highway Department. Aggregate composition was obtained by separating mineral
aggregate in fractions correspondent to each specification sieve with posterior mixing. Los
Angeles abrasion is 11% and apparent density of the mineral composition is 2,873 kg/m³.
Table 1. Crumb rubber particle-size distribution
sieve opening, mm 1.19 0.59 0.42 0.297 0.175 0.150 0.074

% passed 100.00 99.24 59.74 44.99 15.84 11.31 2.46
5. Experimental method
Laboratory tasks were divided in three steps: preparation of asphalt binders, specimen
compaction and acquisition of resilient modulus and tensile strength data. Eight asphalt
binders were prepared, according to the design of the experiment, and, subsequently, Marshall
specimens were produced in five binder contents: 4.5, 5.0, 5.5, 6.0 and 6.5%. Effective specific
gravity of HMA mixtures was measured according to Rice method (ASTM D 2041-00) and
the apparent specific gravity of aggregate composition was obtained by weighing the apparent
specific gravity of fine and coarse aggregates, measured according to protocols ASTM C
128-93 and ASTM C 127-91. Coarse aggregate is that retained in the 4.8mm sieve and fine
aggregate is the one that passes the 4.8 mm sieve. As the mineral aggregate showed small
asphalt absorption, this variable was not taken into account.
5.1. Preparation of crumb-rubber modified binders
Mixtures containing only asphalt binder and crumb rubber and those mixtures containing
asphalt binder, crumb rubber and shale-oil residue were prepared in a high-shear mixer
(Silverson model L4RT) at 4,000 rpm, at 170℃, during 90 minutes. Mixtures containing
only asphalt binder and shale-oil residue were prepared in a low shear mixer (Fisatom model
722D), at 400 rpm and 135℃ during 25 minutes. Seven formulations were prepared, to which
the reference asphalt binder (AC 40) is added, totalizing eight samples.
5.2. Compaction of Marshall specimens
Mineral aggregate and asphalt binder were mixed in an industrial mixer adapted to this
aim. The amount of 1,200 grams was placed in the oven during 2 hours at compaction
temperature. Seven specimens were compacted for each binder content: 4.5, 5.0, 5.5, 6.0 and
6.5%. Table 2 shows mixing and compaction temperatures.
Table 2. Mixing and compaction temperatures (°C)
asphalt binder 1 4 6 2 3 7 5 8
mixing 158 165 180 150 143 175 155 165
compaction 148 155 170 140 132 165 145 155
5.3.Acquisition of resilient modulus and tensile strength data
Resilient modulus values were obtained according to the AASHTO TP 31 protocol

and tensile strength values were obtained according to ASTM D4123 protocol, at 25℃.
For resilient modulus tests, one hundred loading cycles of 1s were applied by a pneumatic
equipment. Each pulse is divided in a loading period of 0.1s and a resting period of 0.9s.
Resilient modulus is obtained using the total displacement recovered during the resting period.
Total displacement is calculated as the difference between the maximum displacement of the
loading cycle and the displacement registered at the end of the cycle. Seven replicates were
tested for eight asphalt mixtures with five binder contents, totalizing 280 measurements.
Tensile strength values were determined through the indirect tensile test using a pneumatic
equipment. Specimens were subjected to a monotonic loading until failure, when the load
is measured. Tensile strength values were calculated based on failure loading and sample
geometry. Two replicates were tested for eight asphalt mixtures with five binder contents,
totalizing 80 measurements.
6. Analysis procedure
The statistical package Minitab 15.1 has been used to run the statistical regression.
Residual analysis has been performed through visual inspection, in order to check trends in
residuals, and through the Anderson-Darling test, to check normal distribution of residual.
Multicollinearity effects, common in experimental regions highly constrained, has not been
considered in this research, as modeling is focused exclusively on prediction. The model used
is called special cubic model, with the general format shown in equation 1.
(1)
with i, j, k = 1, 2, ... q, where q is the number of components, β i, βij and βijk are the
coefficient estimates for the terms xi, xij and xijk, respectively, with xi, xj and xk corresponding
to the proportions of asphalt, rubber and oil, respectively.
Statistical analysis is based on the evaluation of component effect plots, called response
trace plots, and of response surface plots. A response trace plot depicts the effect that each
component has on the response, in accordance with the fitted model. This kind of plot
indicates the effect of varying the proportion of one of the components, keeping the ratio
between the other two constant. This means that when analyzing the effect of varying rubber
proportion, for instance, the proportions of asphalt binder and shale-oil residue also vary
but the ratio between these two components is kept constant. Each line in the plot depicts
the effect of changing the proportion of each component along an imaginary line (direction)
linking the reference blend (usually the centroid of the simplex) to the vertex and the reference
blend to the side opposite to the vertex.
It is important to understand that the component effect analysis depicted by this kind of
plot is only one among several possibilities of analysis (several directions). The direction
chosen in this analysis procedure corresponds to the line linking the vertex to the opposite
side of the simplex, passing through the reference blend, that in this case is the centroid of the
simplex shown in Figure 1. This kind of plot can be used to identify which component is more
influential on response. The horizontal amplitude indicates in which ranges the component
proportions vary, taking account of the limits indicated in the experimental design, starting
from the reference blend to +50% of the range of variation and to -50% of this range. The
vertical amplitude indicates the intensity of the effect of changing the component proportion
on the monitored property. An easier way of evaluating the effects is to look at the percentage
of each component as increasing from left to right, in its respective range of variation.
While analyzing a response trace plot, one should keep in mind that these results are
obtained along a certain direction within the simplex. Additional analyses are possible in other
directions. This kind of graph represents a statistical artifact in order to try to understand how
each component affects response individually, as it is not possible to vary the proportion of
one of the components keeping the proportions of the others constant. In practical terms, the
three components act together and their effects on response depend on their interaction effects.
These interaction effects depend on the component proportions and are depicted in a more
effective and realistic manner by response surfaces.
When interpreting a response trace plot, it is important to keep in mind that: (a) all
components are interpreted relatively to the reference blend; (b) components with the greatest
effect on response will show the steepest response traces; (c) components with larger ranges
(upper bound – lower bound) will have longer response traces, components with smaller
ranges will have shorter response traces; (d) the total effect of a component depends on both
the range of the component and the steepness of its response trace, the total effect is defined
as the difference in the response between the effect direction point at which the component is
at its upper bound and the effect direction point at which the component is at its lower bound;
(e) components with approximately horizontal response traces, with respect to the reference
blend, have virtually no effect on response; (f) components with similar response traces will
have similar effects on response; (g) components do not have effects completely isolated: to
change the proportion of one of the components implies to change the proportion of at least
one of the complementary components.
Response surface plots depict how a response is related to mixture components, based
on the fitted model. In this kind of plot, the surface is seen as a bidimensional plan in which
all points that represents the same response value are connected to produce contour lines.
Response surface plots are useful to select ideal values for a property and the correspondent
ideal mixtures. Therefore, response surface plots represent an effective tool to delimitate
regions where shale-oil residue could be used with rubber, working as an extender oil.
It is important to remember that the models fitted in this research are restricted to the
type of asphalt binder, crumb rubber and shale-oil residue used, to the levels of the process
variables and to the experimental conditions under which mixtures have been prepared. The
fitted models are not intended to be applied to either any kind of asphalt binder, crumb rubber
and shale-oil residue or any level of process variables different from the levels used here.
7. Findings
Table 3 shows the fitted models relative to resilient modulus and tensile strength based
on which response trace plots and contour plots were generated. Figures from 2 to 6 depict
the response trace plots and Figures from 7 to 11 show the response trace plots for resilient
modulus. Figures from 12 to 16 present the response trace plots and Figures from 17 to 21
depict the contour plots for tensile strength.
Table 3. Fitted models for resilient modulus (RM) and tensile strength (TS)
property regression model R² (%) n

RM 4.5% = 15,100x1 +7,893x2 –33,538x3 64.5 55
= 16,709x1 –228,251x2 –54,181x3 +256,188x1x2 –829,002x2x3 +
RM 5.0% 96.7 54
+1,837,430x1x2x3
RM 5.5% = 14,988x1 –12,9848x2 –14,5248x3 +161,236x1x2 –217,952x1x3 90.7 55
RM 6.0% = 14,758x1 –96,283x2 –63,478x3 +112,993x1x2 + 541,389x1x2x3 95.2 53
RM 6.5% = 12,304x1 +150,432x2 +217,702x3 +180,505x1x2 –302,847x1x3 97.5 51
TS 4.5% = 2.448x1 +0.543x2 –2.756x3 58.5 16
property regression model R² (%) n

TS 5.0% = 2.716x1 +0.013x2 –2.984x3 66.0 16
TS 5.5% = 2.865x 1 +0.311x2 –4.153x3 82.2 16
TS 6.0% = 2.818x 1 +3.129x2 –7.120x3 88.1 16
TS 6.5% = 2.620x 1 –23.280x2 –6.910x3 +27.620x 1x2 +71.280x 1x2 x 3 97.1 16
x1: asphalt proportion; x2: rubber proportion; x3: oil proportion; n: number of specimens
The response trace plot for the 4.5% asphalt content (Figure 2) shows that both rubber and
oil reduce resilient modulus, but oil has a more intense effect. Rubber brings resilient modulus
down at a rate of 35MPa to each 1% rubber added and oil reduces it at a rate of 477MPa/1%.
Figure 3, relative to 5.0% asphalt content, depicts that rubber has a parabolic effect: resilient
modulus increases for rubber from 0 to 8% and decreases from 8 to 18%.
For the asphalt contents of 5.5, 6.0, and 6.5%, rubber effect is also parabolic with
maximum point around the reference mixture (84-9-7). For the asphalt contents of 5.5% (Figure
4) and 6.5% (Figure 6), resilient modulus values increase for rubber from 0 to 8%. For the 6.0%
asphalt content (Figure 5), the maximum point occurs near 11% rubber. The overall results
indicate that rubber tends to reduce resilient modulus values when intermediate to high rubber
proportions are used, for instance, superior to 8%. When rubber proportions from small to
intermediate are used, rubber tends to increase resilient modulus. This influence seems to be
unexpected, but it could be interpreted as a result of interaction effects among components.
18000 Component
resilient modulus (MPa) - 4.5%
asphalt
16000
rubber
14000 oil
12000
10000
8000
6000
-0.2 -0.1 0.0 0.1 0.2
deviation from reference blend in proportion
Figure 2. Response trace plot of resilient modulus for binder content of 4.5%
20000
Component
17500 asphalt
rubber
15000 oil
12500
10000
7500
5000
-0.2 -0.1 0.0 0.1 0.2
Figure 3 shows that oil acts reducing resilient modulus in an approximately linear way and
that its effect is less expressive for low concentrations. Oil reduces resilient modulus at a rate
of 390 MPa per 1% oil added. This linear tendency is kept for the other asphalt contents, but
with some particularities. For the 5.5 and 6.5% asphalt contents (Figures 4 and 6), oil effect
on resilient modulus is unexpressive for high proportions. For the 6.0% asphalt content (Figure
5), oil effect is linear for all concentrations. For the 5.5, 6.0 and 6.5% asphalt contents, rubber
reduces resilient modulus at the following rates: 512, 507 and 433 MPa/1% oil added.
18000
Component
16000 asphalt
rubber
14000 oil
12000
10000
8000
6000
-0.2 -0.1 0.0 0.1 0.2
16000
Component
asphalt
14000 rubber
oil
12000
10000
8000
6000
-0.2 -0.1 0.0 0.1 0.2
13000 Component
asphalt
12000
rubber
11000 oil
10000
9000
8000
7000
6000
5000
-0.2 -0.1 0.0 0.1 0.2
asphalt
1.00
10000
0.00 0.00
6000
12000
8000 4000 2000
0.32 0.68 0.32

rubber oil
Figure 7. Contour plot of resilient modulus for binder content of 4.5%
The linear model obtained for resilient modulus at 4.5% binder content generates linear
level curves (Figure 7). This contour plot indicates a small rubber effect. Higher resilient
modulus values are obtained when small proportions of rubber and oil are used (mixtures at
the top of the experimental region) and smaller values are obtained when high proportions of
rubber and oil are used (mixtures at the bottom of the experimental region).
asphalt
1.00
14000.0
0.00 0.00
10000.0
12000.0
6000.0
8000.0
2000.0 2000.0
0.32 0.68 0.32
rubber oil

asphalt
1.00
12000.0
10000.0
0.00 0.00
7000.0
8000.0
7000.0
6000.0 8000.0
0.32 0.68 0.32
rubber oil
Figure 8 (contour plot for the 5.0% binder content) points out that higher resilient
modulus values are obtained with low to intermediate rubber concentrations and with low oil
concentrations (mixtures at the top of the experimental region). This pattern repeats for the
5.5, 6.0 and 6.5% asphalt contents (Figures 9 to 11). Asphalt mixtures with lower resilient
modulus values are obtained with high concentrations of rubber and oil for the 5.0, 5.5 and
6.5% asphalt contents (mixtures at the bottom of the experimental region). Lower resilient
modulus values are also obtained for low rubber proportions and high oil proportions for the
6.5% asphalt content (mixtures at the right of the experimental region). For the 6.0% asphalt
content, smaller resilient modulus values are obtained only with small rubber proportions and
high oil proportions.
asphalt
1.00
10000.0
12000.0
0.00 0.00
6000.0
8000.0 2000.0
0.32 0.68 0.32

rubber oil
asphalt
1.00
8000.0
0.00 0.00
6000.0
4000.0 6000.0
8000.0
2000.0
0.32 0.68 0.32
rubber oil
It seems natural to associate the different patterns depicted by component effects plots
to the influence of binder content. On the other hand, it is important to keep in mind that
these plots are generated from models whose coefficients are fitted using available samples.
Variability inherent to data determinates if a certain coefficient will be statistically significant
or not, at a 95% confident level, leading to models in which not all terms are present. The
different patterns depicted by these plots are influenced not only by asphalt content but also
by model terms. Hence, asphalt content can not be pointed out as the unique factor influencing
resilient modulus behavior.
As an example, the maximum resilient modulus value of 8,000MPa was chosen in order to
find which compositions meet this requirement. Table 4 indicates rubber and oil proportions,
for the five binder contents, in order to guarantee that HMA mixtures show resilient modulus
higher than 8,000MPa. A visual inspection of response surfaces reveals that it is not easy to
decide which these ranges are. Then, ranges reported here are an approximation. The overall
results indicate that, in order to meet a resilient modulus value of 8,000 MPa, only asphalt
binders containing high rubber and oil proportions are acceptable, for binder contents of 4.5, 5.0
and 5.5%, and only asphalt binders with any rubber proportion and intermediate to high oil
proportions are acceptable, for binder contents of 6.0 and 6.5%. It is obvious that these ranges
are tied to the arbitrated value for resilient modulus. Other limits for resilient modulus would
imply in other ranges.
Table 4. Proportion ranges for resilient modulus superior to 8,000MPa
binder content proportions

4.5 rubber from 10 to 18% and oil from 12 to 14%
5.0 rubber from 16 to 18% and oil from 12 to 14%
rubber from 0 to 3% and oil from 12 to 14% and rubber
5.5
from 13 to 18% and oil from 11 to 14%
6.0 rubber between 0 e 18% and oil from 12 to 14%
6.5 rubber between 0 e 18% and oil from 6 to 14%
Figure 12 depicts the component effects on tensile strength for the 4.5% binder content.
Both crumb rubber and shale-oil residue contribute to reduce tensile strength, what can be
seen as a disadvantage. The increase of rubber proportion implies in the reduction of asphalt
available to bind the mineral particles. With less binding material available, tensile strength
reduces naturally. Crumb rubber does not play the role of binding mineral particles, but it can
affect consistency of available asphalt, once rubber particles swells during processing using
light fractions from asphalt binder.
Another aspect to be taken into account is the susceptibility of tensile strength to binder
content: generally, the highest tensile strength values lie around optimum binder content and
smaller values are obtained for binder contents above and below the optimum binder content.
In fact, tensile strength is related to film thickness formed at aggregate surface and so thicker
films would imply in higher tensile strength values until a certain binder content, as asphalt
binder in excess would damage tensile strength.
2.50
Component
asphalt
Tensile Strength (MPa) - 4.5%
2.25 rubber
oil
2.00
1.75
1.50
-0.2 -0.1 0.0 0.1 0.2

Figure 12. Response trace plot of tensile strength for binder content of 4.5%
2.75 Component
asphalt
2.50 rubber
oil
2.25
2.00
1.75
1.50
-0.2 -0.1 0.0 0.1 0.2

As 4.5% is a small binder content, taking into account the particularities of the materials
used here, the quantity of binding material available is small, what may lead to smaller tensile
strength values. By its time, the increase of oil proportion also reduces the asphalt available,
reducing the quantity of binding material. But oil plays another undesirable role: it makes the
available asphalt binder less consistent, reducing tensile strength as a consequence.
3.00
Component
asphalt
Tensile Strenght (MPa) - 5.5%
2.75
rubber
2.50 oil
2.25
2.00
1.75
1.50
-0.2 -0.1 0.0 0.1 0.2

3.00
Component
asphalt
2.75
rubber
2.50 oil
2.25
2.00
1.75
1.50
-0.2 -0.1 0.0 0.1 0.2

Figure 12 also shows that rubber and oil reduce tensile strength with different intensities.
Oil has a more intense effect than rubber: oil reduces tensile strength at a rate of 0.05MPa per
1% oil added and rubber reduces it at a rate of 0.02MPa/1%. The same pattern is observed in
Figures 13 and 14. Both rubber and oil reduce tensile strength and oil preserves its tendency
of reducing this property with more intensity than rubber. For the 5.0% binder content (Figure
13), rubber reduces tensile strength at a rate of 0.02MPa/1% and oil at a rate of 0.05MPa/1%.
For the 5.5% binder content (Figure 14), rubber reduces tensile strength at a rate of
0.02MPa/1% and oil at a rate of 0.07MPa/1%.
Rates obtained for binder contents from 4.5 to 5.5% seem to indicate that asphalt content
does not interfere either in the component effect or the component intensity. But it is worthy
to mention that the quality of the fitted models can interfere in these trends in a certain extent.
Models for 4.5 and 5.0% binder content showed a coefficient of determination relatively low,
denoting a relatively poor fitting. Looking at the coefficients of determination of the five
models, one can observe that there is a high correlation between quality of regression and
binder content, although it would be dangerous to conclude that this correlation has a true
cause-effect relationship.
A slightly different pattern is observed for the 6.0% binder content, as depicted by Figure
15. The linearity of effects is kept, as seen in the previous graphs, but two main changes are
observed: crumb rubber increases tensile strength slightly, instead of reducing it as shown in
other binder contents, and oil effect is more intense. Rubber increases tensile strength at a rate
of 0.01MPa/1% and oil reduces it at a rate of 0.09MPa/1%. This could be an explanation to
the controversy of literature about the effect rubber has on tensile strength: in fact rubber may
increase tensile strength but it depends on binder content.
2.8
Component
2.6 asphalt
2.4 rubber
oil
2.2
2.0
1.8
1.6
1.4
1.2
-0.2 -0.1 0.0 0.1 0.2
A completely different pattern is observed in Figure 16 for tensile strength. In this case,
the regression is parabolic, instead of linear as in the previous binder contents. Rubber has
a parabolic effect, with maximum point near the reference blend (84-9-7), what means that
rubber increases tensile strength from 0 to 9% and decreases it from 9 to 18%. This helps to
understand that rubber can increase tensile strength depending not only on binder content but
also on rubber proportion. For small proportions, rubber increases tensile strength, but for
intermediate to high proportions, rubber decreases it. But this conclusion is valid only for the
6.5% binder content. For the 6.5% binder content, oil maintains its linear effect on tensile
strength, reducing it at a rate of 0.07MPa/1%.
Although response trace plots clearly describe component effects, they are generated
from a statistical model that tries to interpretate nature. But one question still remains: is
crumb rubber really capable of increasing tensile strength taking into account that it can not
contribute in terms of binding potential? One would have to assume that probably there is
another factor influencing tensile strength increase rather than rubber proportion.
Optimum binder content may be this factor. As tensile strength generally decreases when
binder content is above the optimum point due to the excessive quantity of asphalt binder,
however much rubber is added to the composition the smaller the quantity of asphalt binder
available. Thus, it is not a question of rubber increasing tensile strength for a binder content
above the optimum point. But, rather, it is a question of a smaller quantity of asphalt binder
presented in a formulation with a high concentration of rubber. Higher rubber proportion, and
consequently smaller asphalt proportions, imply in higher tensile strength values.
Table 5 shows the intensity of the effects of rubber and oil of reducing tensile strength
at a rate of MPa per 1% modifier added. Negative indicate reduction and positive indicate
increase. Results indicate that binder content does not affect the intensity of rubber effect on
tensile strength expressively and that binder content has a slight effect on oil effect. Oil effect
increases with binder content for binder contents from 4.5 to 6.0%.
Table 5. Intensity of component effects on tensile strength (MPa/1% modifier added)
binder content (%)

component 4.5 5.0 5.5 6.0 6.5
rubber - 0.02 - 0.02 - 0.02 + 0.01 -
oil - 0.05 - 0.05 - 0.07 - 0.09 - 0.07
Figures from 17 to 20 are generated from linear models, so level curves are also linear.
Figures from 17 to 19 show a similar pattern: HMA mixtures using asphalt binders with
high proportions of rubber and oil present smaller tensile strength values and HMA mixtures
using asphalt binders with small proportions of rubber and oil show higher tensile strength
values. Figure 20 also shows linear level curves, but its pattern is slightly different from the
previous figures. Level curves are parallel to the side asphalt-rubber, indicating that oil effect
is predominant and rubber effect is unexpressive. Figure 21 shows nonlinear level curves, but
the same pattern is kept: HMA mixtures using asphalt binders with high proportions of rubber
and oil present smaller tensile strength values and HMA mixtures using asphalt binders with
small proportions of rubber and oil show higher tensile strength values.
asphalt
1.00
2.25
0.00 1.75 0.00

2.00
1.25
1.50 1.00
0.32 0.68 0.32
rubbe r oil
Figure 17. Contour plot of tensile strength for binder content of 4.5%
asphalt
1.00
2.25
0.00 0.00
1.75
2.00
1.25
1.50
0.32 0.68 0.32
rubbe r oil
asphalt
1.00
2.50
0.00 2.00 0.00

2.25
1.50
1.00
1.75 1.25
0.32 0.68 0.32

rubbe r oil
asphalt
1.00
2.00
0.00 0.00
1.00
0.50
1.50 0.65
2.50
0.32 0.68 0.32
rubbe r oil
Brazilian specifications require a minimum tensile strength value of 0.65 MPa for dense-
graded asphalt mixes. As response surfaces point out, all formulations within the dashed
region lead to HMA mixtures with tensile strength values superior to that required by
specification. These results drive to the conclusion that asphalt binders formulated with any
rubber proportions from 0 to 18% and any oil proportion from 0 to 14% are acceptable in
terms of tensile strength if the 0.65 MPa limit is chosen.
asphalt
1.00
2.25
1.75
0.00 0.00
2.00 1.25
0.65
1.50
1.00 1.00 0.25
0.32 0.68 0.32

rubbe r oil
8. Conclusions
This research aimed at evaluating the effect of crumb rubber and shale-oil residue on
resilient modulus and tensile strength values obtained at 25℃. Eight HMA mixtures were
produced using asphalt binders compounded with different proportions of crumb rubber and
shale-oil residue. Properties were monitored at five binder contents (4.5, 5.0, 5.5, 6.0 and 6.5%)
and data were used to fit mixture models, based on which response trace plots and contour
plots were generated.
From the response trace plot analysis, the following conclusions can be drawn:
• the overall effect of rubber on resilient modulus is parabolic, with maximum point
around the reference mixture (84-9-7);
• when binder content is 5.0, 5.5 or 6.5%, rubber increased resilient modulus for
proportions from 0 to 8% and reduced it for proportions above 8%; when binder
content is 6.0%, rubber increased resilient modulus for proportions from 0 to 11%
and reduced it for higher proportions; for 4.5% binder content rubber effect is linear,
acting to reduce resilient modulus at an approximate rate of 35MPa per each 1% rubber
added;
• rubber tends to reduce resilient modulus when intermediate to high proportions are
used.
• rubber tends to increase resilient modulus when small to intermediate proportions are
used;
• oil reduces resilient modulus linearly, at an approximate rate of 400 to 500 MPa per
each 1% oil added, depending on binder content;
• oil effect is inexpressive on resilient modulus when small concentrations (for 5.0%
asphalt binder) or high concentrations (for 5.5 and 6.5% binder content) are used;
• both rubber and oil reduce tensile strength for binder contents of 4.5, 5.0 and 5.5%; for
6.0% binder content, rubber increases tensile strength slightly and oil reduces it; for
6.5% binder content, oil reduces tensile strength and rubber shows a parabolic effect,
increasing it for proportions from 0 to 9% and reducing it for proportions from 9 to
18%;
• for small binder contents (from 4.5 to 5.5%), rubber reduces tensile strength at an
average rate of 0,02MPa/1%; oil reduces tensile strength at an average rate of 0.07
MPa/1%;
• binder content does not affect the intensity of rubber effect on tensile strength
expressively and has a slight effect on oil effect: oil effect increases for binder contents
from 4.5 to 6.0%;
• although literature reports that, in general, rubber reduces tensile strength, this research
indicates that rubber can increase tensile strength, depending on binder content and
rubber proportion: small to intermediate proportions of rubber combined with high
binder contents lead to higher tensile strength values.
From the response surface analysis, the following conclusion can be drawn:
• in order to meet a resilient modulus value of 8,000 MPa, only asphalt binders
containing high rubber and oil proportions are acceptable, for binder contents of 4.5,
5.0 and 5.5%, and only asphalt binders with any rubber proportion and intermediate to
high oil proportions are acceptable, for binder contents of 6.0 and 6.5%;
• HMA mixtures using asphalt binders with high proportions of rubber and oil present
smaller tensile strength values and HMA mixtures using asphalt binders with small
proportions of rubber and oil show higher tensile strength values;
• all formulations within the experimental region (dashed region) lead to HMA mixtures
with tensile strength values superior to that required by Brazilian specification (0.65
MPa minimum), then asphalt binders formulated with any rubber proportions from 0 to
18% and any oil proportion from 0 to 14% are acceptable in terms of tensile strength if
this limit is taken into account.
Aknowlegments
The authors would like to thank SIX-Petrobras for supplying the shale-oil residue, Artgoma
S. A. for supplying the crumb rubber, Capes for supplying a master-degree scholarship for the
first author and Fapesp for supplying a doctoral scholarship to the second author.
References
Ayres M., Witczak M. W., “Resilient modulus properties of asphalt rubber mixes from field
demonstration projects in Maryland”, Transportation Research Record, n. 1492, 1995, p.
96-107.
Bertollo S.A.M., Fernandes Jr. J.L., Bernucci L.L.B., “Mechanical properties of rubber
modified asphalt mixtures produced in Brazil”, Journal of International Pavements, v.2, n.2,
May, 2003, p.62-72.
Cordeiro W. R., « Mechanical behavior of HMA mixes using crumb-rubber modified binders
», Master Degree Thesis, Military Institute of Engineering, Rio de Janeiro, 2006, 253p. (in
Portuguese)
Cornell J. A., Experiments with mixtures: design, models, and the analysis of mixture data,
3rd. ed., New York, John Wiley & Sons, 2002.
Epps J. A., « Uses of recycled rubber tires in highways », NCHRP – Synthesis of Highway
Practice, n. 198, Washington, DC, 1994, 162p.
Faxina A. L., Soria M. H. A., “Mechanical properties of asphalt-rubber mixes using shale
extender oil”, Proceedings of the Asphalt Rubber 2003 Conference, Brasilia, Brazil, 2003, p.
693-706.
Hanson D.I., Foo K.Y., Brown E.R., Denson R., “Evaluation and characterization of a rubber-
modified hot mix asphalt pavement”, Transportation Research Record, n.1436, 1994,
p.98-107.
Leite L. F. M., Motta L. M. G., Bernucci L. B., Soares J.B., “Mechanical behavior of asphalt
rubber mixes prepared in laboratory”, Proceedings of the Asphalt Rubber 2000 Conference,
Vilamoura, Portugal, 2000, p. 309-318.
Lemes P. C. L., « Study on short-term aging of HMA mixes using modified asphalt binders »,
Master Degree Thesis, Sao Carlos Engineering School, Sao Paulo University, Sao Carlos,
2004, 131p. (in Portuguese)
Momm L., Salini, R., “Study of recycled tyre rubber in asphalt concrete mixtures”.
Proceedings of the Asphalt Rubber 2000 Conference, Vilamoura, Portugal, 2000, p.
341-357.
Specht L.P., « Evaluation of HMA mixes with recycled tire rubber », Doctoral Thesis, Rio
Grande do Sul Federal University, Porto Alegre, 2004, 279p. (in Portuguese)
Fatigue-Related Properties of Crumb

Rubber Modified Asphalt Binders
Felice Giuliani* — Filippo Merusi* — Antonio Montepara*
* Department of Civil and Environmental Engineering

University of Parma
Parma
Italy
felice.giuliani@unipr.it
filippo.merusi@nemo.unipr.it
antonio.montepara@unipr.it
ABSTRACT. The experimental investigation presented in this paper focuses on the

characterization of the fatigue behavior in crumb rubber modified (CRM) asphalt binders.
The main purpose was to identify and to describe the fatigue processes in CRM binders rather
than quantify the expected improvement in fatigue life due to crumb rubber modification. The
development of the complex modulus (G*) in cyclic shear tests (time sweeps) was recorded
under different loading conditions and the obtained trends were evaluated according to the
available failure criteria. Using both stress-controlled and strain-controlled testing mode, the
relations between the stress/strain amplitude and the number of cycle to failure (S-N curves)
were established. The results allowed to define some dynamics related to the contribute of the
crumb rubber modification in binder damage resistance with regard to the considered loading
conditions. The role of the rheometrical parameters was also investigated and the influence of
the loading mode on the damage characteristics of CRM binders was established.
KEYWORDS: Fatigue, Damage Behaviour, Performance-Related Properties.

1. Introduction
Background
With low and intermediate service temperatures, fracture propagation in asphalt pavements
frequently derives from the cyclic accumulation of fatigue damage related to stresses lower
than the amount required for the immediate failure. Different methods to evaluate fatigue in
bitumen have been thus investigated since rheological analysis was introduced in pavement
engineering. In the early ’90s, the development of the SHRP research program led to the
definition of fatigue resistance in bitumen by measuring G*·sinδ at 10 rad/s (Bahia and
Anderson, 1995). The derived criteria were also consequently used to qualify the role of
crumb rubber modifiers (CRM) in asphalt binders mechanical performances (Bahia and
Davies, 1994). However, although the SHRP parameter is still used for technical specification,
it was found inadequate for describing the real fatigue performance of modified bitumen in
the field. At this regard, the study of Stuart and Mogawer (2002), dealing with the relation
between the fatigue resistance of asphalt pavements and the SHRP fatigue parameter of
asphalt binders, is representative.
At present, different approaches have been proposed to improve fatigue testing of bitumen
and the concept of damage behavior was introduced by Bahia et al. (2001-a) to define specific
properties and test procedures related to the conditions of incipient failure in bitumen.
Moreover, the use of dynamic mechanical analysis (DMA) to evaluate fatigue properties of
bitumen was also proposed as a part of the NCHRP project 9-10 (Bahia et al., 2001-b) for the
characterization of modified asphalts in the Superpave mix design and a similar approach, still
based on the DMA, was also successfully developed by Kim et al. (2002) to study both fatigue
and healing potential in sand-asphalt mixtures and mastics. On the basis of these results it was
well known and accepted that repeated shear loading of asphalt binders in DSR causes a loss
in complex modulus. However, two different modes of failure have been observed: internal
fatigue micro-cracking and plastic edge fracture (Anderson et al., 2001). This topic was then
investigated by Bonnetti et al. (2002) and by Planche et al. (2004) and it was found that “real”
or “true” fatigue actually occurs in time sweep tests only when the initial material stiffness
exceeds a certain value, otherwise edge fracture phenomena is preponderant. Furthermore,
the application of time sweep testing to investigate the effects of different fillers and external
factors on fatigue behavior of bitumen and mastics was also recently done, yielding to an
improvement of the reliability of the method (Soenen et al., 2000, Airey et al., 2006, Bocci
et al., 2006, Santagata et al., 2008). As a consequence, the analysis of the contribution of
crumb rubber and other modifiers on bitumen fatigue behavior can be done by using time
sweep testing (Frantzis, 2003, Ajideh et al., 2006). On the other hand, the influence of the
rheometrical parameters on the final result must be carefully outlined in order to attain to a
reliable assessment of binder fatigue life in presence of non-conventional modifiers. In this
context the present experience is developed, focusing on the application of the different testing
procedures and the proposed failure criteria to the case of CRM asphalt binders in order to
attain to the assessment of their behavior with regard to the considered loading conditions.
2. Research approach
2.1 Materials
The binders used in the experiment consist of traditional bitumen B50/70 (PG 64-22, pen
= 76 dmm, TR&B = 49.5℃) subsequently modified in the laboratory at 16% and 20% CRM
content referred to the weight of the bitumen. The two CRM binders were prepared using
crumb rubber obtained by cryogenic process with maximun particles’ diameter of 0.6 mm.
Digestion time and mixing methodology are referred to as ASTM D-6114. All samples were
not subjected to artificial ageing.
In table 1 the conventional properties of binders are reported. The ring and ball softening
point procedure was done according to EN 1427 while the penetration at 25℃ was measured
according to EN 1426.
Table 1. Materials
CRM content Penetration Softening point – TR&B

Binder Type
[%] [dmm] [°C]
CRM-0 Unmodified 0.0 77 49.0
CRM-16 CRM modified 16.0 26 79.9
CRM-20 CRM modified 20.0 24 88.9
The samples preparation for rheological measurements was referred to EN 12594. CRM
binders were treated as modified binders, thus reheating and homogenization were carefully
carried out at a controlled temperature (160℃ ± 5℃) in order to obtain reproducible results
(Anderson et al., 2000). Special attention was then paid to the specimen thermal history
and storage conditions (1 hour at 25℃ ± 0.5℃) because of their influence on rheological
measurements (Soenen et al., 2005).
2.2 Experimental methods
The tests were carried out by means of a Dynamic Shear Rheometer (DSR). The
temperature during the tests was controlled by means of a Peltier conditioning cell. An air-
operating suspension system guaranteed a significant reduction in the friction between the
moving parts of the rheometer. The selected measurement system is represented by the double
plate configuration with 2.0 mm gap and 8.0 mm diameter according to expected materials’
stiffness. The samples were placed on the bottom plate, squeezed out between the two plates
and trimmed off from the edge of the plates using an hot spatula. Once these operations were
carried out, the gap was set as required in order to guarantee the correct geometry of each
samples. The test temperature was set with a maximum admitted deviation of ± 0.01℃ from
the selected temperature during the whole experiment. Before each test the samples were
subjected to a 30-minute thermal conditioning period. In order to avoid errors due to the
instrument’s sensitivity, the torque applied was higher than the minimum suggested by the
instrument producer (min. torque = 0.5 µNm).
As mentioned above, the initial binders stiffness strongly influences the failure mode in
time sweep tests. In order to ensure that true fatigue occurs during the tests, the initial binders
stiffness should be in the region of 15 MPa and in general it must be greater than 5.0 MPa
(Planche et al., 2004). Moreover it was also demonstrated that viscoelastic behavior of CRM
binders is highly dependent on strain (Kim et al., 2001). A preliminary investigation was thus
carried out in order to evaluate the binders stiffness and to check the extension of the linear
viscoelastic region. According to the SHRP procedure and to the results of Marasteanu et al.
(2000) and Airey et al. (2003) the extension of the linear viscoelastic domain was assessed
by performing strain sweeps tests and the limit of the LVE response was identify as the strain
amplitude where G* is decreased to 95% of its initial value.
Rheological testing for fatigue are performed with regard to different conditions. Strain-
controlled tests were conducted for nominal mean strain amplitude from 1.0% to 4.0%. The
stress-controlled tests were conducted for nominal mean stress amplitude from 250 kPa to 475
kPa. The loading frequency was equal to 10 Hz and the test temperature was kept constantly
equal to 15.00℃ ± 0.01℃. Defining fatigue failure during laboratory testing is quite difficult
and different conventional limit and criteria were proposed to define the number of cycle to
failure (Ghuzlan and Carpenter, 2000). During this study, fatigue life is measured with regard
to two different proposed criteria:
- fatigue life is the number of cycle (Nf) to reach 50% reduction in complex modulus (G*).
- fatigue life is the number of cycles (Np20) to achieve 20% deviation from the initial
linear trend of DER (Dissipated Energy Ratio) and calculated as follows:
(Eq. 1)
where Wi is the dissipated energy per cycle:
(Eq. 2)
With regard to the DER criteria, the Relative Crack Propagation Amplitude (RCPA)
parameter, recently proposed by Santagata et al. (2008) and defined as reported in equation 3,
was also evaluated in this study to highlight the peculiar effects caused by the CRM.
(Eq. 3)
In equation 3, NMAX is the number of cycle corresponding to the maximum value of

DER (evaluated in stress-controlled mode) and Np20 is the conventional binder fatigue life
measured as referred above.
3. Results and analysis
3.1 Linear viscoelastic properties
Results of the viscoelastic characterization are summarized in table 2, where the parameters
G*·sinδ calculated at 10 rad/s and 15℃ are presented in order to identify the SHRP fatigue
qualification of the binders. The equi-stiffness temperature relative to the value of 5.0 MPa for
G* are also presented in order to verify the stiffness condition for ‘true fatigue’ measurement
according to Anderson et al. (2001). Further data in table 2 regard the definition of the linear
viscoelastic limits expressed in terms of stress and strain amplitude.
Table 2. Linear viscoelastic limits, SHRP fatigue parameters at 15℃ and equi-stiffness
temperatures (G* = 5.0 MPa).
γLVE (10 Hz) τLVE (10 Hz) G*·sinδ @ 15°C T(G*=5.0 MPa) @
Binder
[%] [kPa] (10 rad/s) [MPa] 10 Hz [°C]
CRM-0 1.76 423 3.69 21.7
CRM-16 1.24 225 6.28 24.6
CRM-20 0.98 133 5.04 23.0
The equi-stiffness temperatures are higher than 15℃ and very similar for all binders tested.
Therefore, according to the literature findings reported above, it can be considered that the
fatigue test performed at 15℃ are internal to the ‘true fatigue’ region. However, for the stress/
strain amplitudes greater than the material relative τLVE and γLVE, specific effects related to
non-linearity must be taken into consideration.
3.2 Fatigue analysis
3.2.1. Stress-controlled approach
Results of stress-controlled fatigue tests are reported in table 3 as the number of cycles to
failure evaluated from both the 50% G* criterion and the DER criterion. Table 3 also includes
the initial complex modulus G* and the initial dissipated energy per cycle Wi evaluated after
100 loading cycles from the start of the test. All results are means of almost two independent
determinations.
Table 3. Results of stress-controlled fatigue analysis (T = 15℃)
Stress Initial complex Initial dissipated Nf Np20

Frequency
Binder amplitude modulus energy (50%G*) (DER)
[Hz]
[kPa] [MPa] [Pa] [-] [-]
10 250 25.63 5837 1.22·105 1.06·105
10 325 26.16 9837 6.59·104 5.82·104
CRM-0
10 400 24.53 15702 2.53·104 2.23·104
10 475 24.60 22050 1.72·104 1.52·104
10 250 19.32 5526 7.06·105 5.05·105
10 325 19.07 9582 3.14·104 2.42·104
CRM-16
10 400 16.70 17179 6.32·103 4.27·103
10 475 15.83 29153 2.10·103 1.43·103
10 250 13.23 8071 1.41·105 9.23·104
10 325 13.08 15116 3.13·104 2.19·104
CRM-20
10 400 12.40 21961 8.76·103 6.22·103
10 475 11.92 37856 2.98·103 2.08·103
According to the results of previous experiences (Frantzis, 2003, Ajideh et al., 2006),
binder fatigue resistance as measured by stress-controlled rheological testing depends on
the presence of CRM. In particular a reduced fatigue life is generally recorded for CRM-16
and CRM-20, especially for the higher stress amplitude used. This evidence can be firstly
explained by considering that the presence of CRM softens the binder response in the
temperature/frequency range investigated. As a consequence, for stress-controlled loading
mode, CRM binders are subjected to more extensive deformations than the unmodified binder
(Wi at 400 kPa change from 15702 Pa recorded for the unmodified binder to 21961 recorded
for the CRM-20). Moreover, a higher sensitivity to the strain level is recorded, and generally
observed (Kim et al., 2001), for the CRM binders. The consequent higher deformation levels,
together with the higher strain dependency (non-linearity in shear deformation and consequent
modulus reduction), can be firstly explain the reduced fatigue life measured on CRM binders
at τ0 = 475 kPa. However, a different assessment of the problem can be outlined by analysing
data in a graphical form. At this regard, figure 1 depicts the DER trends obtained for binder
CRM-0 and CRM-20 for a shear stress amplitude of 250 kPa. For the sake of brevity, data of
CRM-16 are not presented, however very similar tends were obtained for binders CRM-16
and CRM-20.
1 ,E + 0 5
T = 1 5 ℃ - f = 1 0 H z -τ0 = 2 5 0 kP a
9 ,E + 0 4
F a ilu re
8 ,E + 0 4
D a m a g e p ro p a g a tio n
7 ,E + 0 4
C R M -0 C R M -2 0
6 ,E + 0 4
DER - [ - ]
N o n -d a m a g e d sta g e
5 ,E + 0 4
4 ,E + 0 4
3 ,E + 0 4
2 ,E + 0 4
1 ,E + 0 4
N m a x (C R M -0 ) N m a x (C R M -2 0 )
0 ,E + 0 0
0 ,0 E + 0 0 2 ,0 E + 0 4 4 ,0 E + 0 4 6 ,0 E + 0 4 8 ,0 E + 0 4 1 ,0 E +0 5 1 ,2 E +0 5 1 ,4 E +0 5 1 ,6 E +0 5 1 ,8 E +0 5 2 ,0 E +0 5
N um ber of loading cycles - N [-]
Figure 1. Fatigue data from stress-controlled time sweep, 50%G* failure criterion. Binders
CRM-0 and CRM-20 (σ0= 250 kPa).
It is now immediately clear as the fatigue resistance of CRM binder can not be summarized
by the parameter Np20 only. On the contrary, a more correct approach can be pointed out
by involving the analysis of the whole fatigue behaviour. On the basis of the recorded trend,
we can consider that for CRM binders, the damage propagation starts as the test begins, but
its more ‘ductile’ behaviour leads to locate the complete failure threshold in correspondence
of a very higher number of cycle. Differently, for the CRM-0, the region of the damage
propagation cover a reduced number of cycle (less ductile behavior) and the failure occurs
after a few cycles after the damage starts to propagate.
Substantially, for CRM-20 the detachment of the DER line from the DER=N line occurs
after a few loading cycle. As a consequence, the 20% of DER variation from the linear trend,
which define the cycle number Np20, is located in correspondence of a number of cycle very
far from the complete failure (Nmax). For the unmodified binder CRM-0 the situation is
different and Nmax is similar to Np20. Hence, if Nmax is used in stead of Np20, fatigue life
of binders CRM-16 and CRM-20 increases up to fatigue life higher than those of CRM-0.
A quantitative evaluation of the observation reported above can be offered by the parameter
RCPA, which approaches zero for the unmodified binder only (table 4).
Table 4. RCPA (T = 15℃)
Frequency Stress amplitude RCPA

Binder
[Hz] [kPa] [-]
10 250 0.01
10 325 -0.03
CRM-0
10 400 -0.02
10 475 -0.03
10 250 0.21
10 325 0.18
CRM-16
10 400 0.41
10 475 0.41
10 250 0.42
10 325 0.49
CRM-20
10 400 0.47
10 475 0.49
3.2.2. Strain-controlled approach
Results of strain-controlled fatigue tests are reported in table 5. In addition to the number
of cycles to failure evaluated with regard to both the considered failure criteria, the initial
complex modulus and the initial dissipated energy are also reported.
Table 5. Results of strain-controlled fatigue analysis (T = 15℃)
Strain Initial complex Initial dissipated Nf Np20

Frequency
Binder amplitude modulus energy (50%G*) (DER)
[Hz]
[%] [MPa] [Pa] [-] [-]
10 1.0 23.27 5688 194350 127152
10 2.0 19.91 19980 28940 19756
CRM-0
10 3.0 21.95 51028 9570 6980
10 4.0 19.32 82579 5980 4194
10 1.0 17.73 3096 - -
10 2.0 17.01 12497 - -
CRM-16
10 3.0 13.02 23341 22760 39745
10 4.0 15.46 50486 13300 8630
10 1.0 - - - -
10 2.0 12.13 8935 - -
CRM-20
10 3.0 10.45 18405 223300 558685
10 4.0 10.93 35721 43820 53121
The first main observation is that crumb rubber modification strongly influences the fatigue
life of the base bitumen. For the strain-controlled loading mode, CRM binders shown a great
enhancement in fatigue resistance. The improvement of the fatigue life was recorded in the
entire range of strain amplitude investigated but it was mainly evident at the lower strain
amplitudes. In particular, when the tests were conducted at 1.0% and 2.0% strain amplitude,
the CRM binders did not reach the reference conventional values (50%G* or 20% DER)
of the failure criteria even after 1·106 cycles. In these cases we can hypothesize that the
conditions are approaching a sort of endurance limit, as commonly defined in general fatigue
analysis (Dowling, 1999).
25
N o n -d a m a g e d sta g e D a m a g e p ro p a g a tio n F a ilu re
C R M -0
20
S h e a r com p le x m o d u lu s - G * [M P a ]
15
C R M -20
10
T = 1 5 ℃ - f = 1 0 H z - γ0 = 2 .0 %
0
1 ,E + 0 2 1 ,E + 0 3 1 ,E +0 4 1 ,E +0 5 1 ,E +0 6
N um ber of loading cycles - N [-]
Figure 2. Fatigue data from strain-controlled time sweep – G* analysis. Binders CRM-0,
CRM-20 (γ0 = 2.0%).
As figure 2 depicts, the G*(N) curve of binder CRM-0 is characterized by an initial

horizontal plateau and by a subsequent increase in G* trend, identifying a non-damaged stage
and a steric hardening respectively (Anderson et al., 2001). Differently, the absence of such
initial characteristics in the trend of CRM-20 indicates that CRM binders have a specific
fatigue behavior. As previously discussed for the stress-controlled analysis, in this case the
region of damage progression begins soon at the start of the test and an effective non-damaged
stage can not be well identified. However, as figure 3 depicts, the rate of the decrease in G*,
extremely lower than those recorded for CRM-0, still leads to a very higher fatigue resistance.
The longer fatigue life measured for CRM-20 and CRM-16 can be explained as a
consequence of the specific loading conditions. In a strain-controlled loading mode, the
amount of stress and the dissipated energy per cycle, Wi decreases according to the increasing
damage (increasing number of cycle, N). In presence of binders with low stiffness, i.e. CRM
binders, this fact leads to reduce the values of Wi up to a threshold where damage can not
further propagate up to the complete failure. This observation can be better outlined with
regard to the DER analysis (figure 3).
1,E+08
Non-damaged stage Damage propagation Failure
T = 15℃ - f = 10 Hz - γ0 = 2.0%
1,E+07
Dissipated Energy Ratio - DER [-]
1,E+06
CRM-0
1,E+05
DER (20%)
1,E+04 CRM-20
1,E+03
d(DER)/dN = 1 d(DER)/dN > 1
1,E+02
1,E+02 1,E+03 1,E+04 Np20 1,E+05 1,E+06
Number of loading cycles - N [-]
Figure 3. Fatigue data from strain-controlled time sweep, DER failure criterion. Binders
CRM-0 and CRM-20 (γ0 = 2.0%).
3.2.3. Fatigue curves
Fatigue curves realized using the Np20 derived from the DER failure criteria are presented
in figure 5 and 6. For the binders CRM-16 and CRM-20 very similar fatigue curves were
found. As a consequence, only the data obtained for CRM-20 are reported. The fatigue curves
are realized using the initial dissipated energy in order to plot both strain-controlled data (CR
in the figures) and stress-controlled data (CS in the figures) in a same diagram. In figure 5,
only two points compose the strain-controlled data series since for the lower strain amplitude
no failure was recorded before 106 loading cycles (table 5).
In this representation, the previous observations on CRM binder fatigue behaviour are
more evident and important differences arise between the modified binders and the unmodified
one. In fact, a unique fatigue curve, that can be fitted with a unique power law in the W-N plot
can be obtained for the unmodified binder only (figure 4). Contrary, for the binder CRM-20,
a univocal fatigue curve can not be identified and data derived from strain-controlled analysis
(white points in figure 5) are located on a different line. In this case the poor value of the
coefficient R2, equal to 0.16, quantifies this observation.
1,E+05
T = 15℃ - f = 10 Hz
Initial dissipated energy - W [Pa]
1,E+04
y = 4E+07x-0,7727
R2 = 0,9809
CRM-0 (CR)
CRM-0 (CS)
1,E+03
1,E+03 1,E+04 1,E+05 1,E+06 1,E+07
Number of cycle to failure - Np20 [ - ]
Figure 4. Fatigue curve of CRM-0 from DSR time sweep, DER failure criterion (CR =
strain-controlled data, CS = stress-controlled data).
1,E+05
T = 15℃ - f = 10 Hz
Initial dissipated energy - W [Pa]
1,E+04 y = 61111x-0,1113
R2 = 0,1641
CRM-20 (CR)
CRM-20 (CS)
1,E+03
1,E+03 1,E+04 1,E+05 1,E+06 1,E+07
Number of cycle to failure - Np20 [ - ]
Figure 5. Fatigue curve of CRM-20 from DSR time sweep, DER failure criterion (CR =
strain-controlled data, CS = stress-controlled data).
4. Conclusions
The fatigue behavior of CRM binders was studied using DSR time sweeps. The binders
relative fatigue life was consequently evaluated according to different failure criteria. Based
on the experimental results, it was shown as CRM asphalt binders have a specific fatigue
behavior, strongly dependent on the loading mode, and characterized by its own fundamental
characteristics.
By analyzing the DER trends obtained from stress-controlled time sweeps it was shown as
in the case of CRM binders, the damage propagation starts as the test begins, but the presence
of a more ‘ductile’ behavior leads to locate the complete failure threshold in correspondence
of a very higher number of cycle. Differently, for the unmodified base bitumen, the region
of the damage propagation cover a reduced number of cycle (less ductile behavior) and the
failure occurs soon as the damage starts to propagate. However, in case of stress-controlled
analysis, the differences in fatigue life due to the modification become less important when
the stress amplitude is external to the linear viscoelastic region.
If the loading mode is strain-controlled, a very improved fatigue resistance was recorded
for the CRM modified binders. In particular, when the tests were conducted for low strain
amplitude, the recorded data did not meet the failure criteria even after 106 loading cycles.
Finally, it was shown as, in presence of CRM, stress-controlled data and strain-controlled data
must be described by different curves in the W-N plot.
5. Bibliography
Airey G.D., “Linear viscoelastic performance of asphaltic materials”, Road Materials and
Pavement Design, 4 (3), pp.269-292, 2003.
Airey G.D., Liao M-C., Thom N.H., “Fatigue behaviour of bitumen-filler mastics”,
Proceedings of the 10th International Conference on Asphalt Pavements, International
Society for Asphalt Pavements (ISAP), Quebec City, Canada, 2006.
Ajideh H., Burris B., Bahia H., “Optimizing the use of crumb rubber by binary modification”,
Proceedings of the Asphalt Rubber Conference 2006, pp.813-839, Palm Springs, USA,
25-27 October, 2006.
Anderson D.A., Marasteanu M.O., Mahoney J.M., Stephens J.E., “Factors affecting the
variability in the SHRP binder tests”, 79th Annual Meeting of the Transportation Research
Board (TRB), Washington D.C., USA, January, 2000.
Anderson D.A, Le Hir Y.M., Marasteanu M.O., Planche J.P., Martin D,. Gauthier G.,
“Evaluation of fatigue criteria for asphalt binders”, Transportation Research Record 1766
– Paper No. 01-3298, pp. 48-56, 2001.
Bahia, H.U., Davies, R., “Effect of Crumb Rubber Modifiers (CRM) on performance-related
properties of asphalt binders”, Journal of the Asphalt Paving Technologists, 63, pp.
414-449, 1994.
Bahia H. U., Anderson D. A., “The new proposed rheological properties of asphalt binders:
why are they required and how do they compare to conventional properties”, Physical
Properties of Asphalt Cements Binders, ASTM STP 1241 (American Society for Testing and
Materials), Philadelphia, 1995.
Bahia H.U., Zhai H., Zeng M., Hu Y., Turner P. “Development of binder specification
parameters based on characterization of damage behavior”, Journal of the Association of the
Asphalt Paving Technologists, 70, USA, 2001-a.
Bahia H.U., Hanson D.I., Zeng, M., Zhai, H., Khatri, M.A., Anderson, R.M.
“Characterization of modified asphalt binders in Superpave mix design”, National Cooperative
Highway Research Program (NCHRP), report 459, Transportation Research Board – National
Research Council, Washington, D.C., USA, 2001-b.
Bocci M., Cardone F., Cerni G., Santagata E., “Rheological characterization of fatigue
resistance of asphalt binders”, 10th International Conference on Asphalt Pavements (ISAP),
Quebec City, Canada, 2006.
Bonnetti K.S., Nam K., Bahia H.U., “Measuring and defining fatigue behavior of asphalt
binders”, Transportation Research Record 1810, 2002.
Dowling N.E., “Mechanical behavior of materials”, Prentice-Hall, Inc., Upper Saddle

River, New Jersey, USA, 1999.
Frantzis, P., “Development of crumb rubber reinforced bituminous binder under laboratory
conditions”, Journal of Materials Science, 38, pp. 1397-1401, 2003.
Ghuzlan K.A., Carpenter S.H., “An energy-derived/damage-based failure criteria for

fatigue testing”, Transportation Research Record 1723, pp.141-149, 2000.
Kim S., Loh S.W., Zhai H., Bahia H.U., “Advanced characterization of crumb rubber-
modified asphalts, using protocols developed for complex binders”, Transportation Research
Record 1767 – Paper No. 01-2986, pp. 15-24, 2001.
Kim Y.R., Little D.N., Lytton R.L., “Use of dynamic mechanical analysis (DMA) to
evaluate the fatigue and healing potential of asphalt binders in sand asphalt mixtures”, Journal
of the Association of Asphalt Paving Technologists, 71, 2002, pp. 176-206.
Loh S. W., Kim S., Bahia H.U., “Characterization of simple and complex crumb rubber
modified binders”, Report No. WI/SPR-07-01, Wisconsin Department of Transportation, July,
2000.
Planche J.P, Anderson D.A., Gauthier G., Le Hir Y.M., Martin D., “Evaluation of fatigue
properties of bituminous binders”, Materials and Structures, 37, pp. 356-359, 2004.
Marasteanu M.O., Anderson D.A., “Establishing linear viscoelastic conditions for asphalt
binders”, Transportation Research Record 1766, Paper No. 00-1374, pp. 48-56, 2000.
Santagata E., Baglieri O., Dalmazzo D., “Experimental investigation on the fatigue damage
behaviour of modified bituminous binders and mastics”, Journal of the Association of Asphalt
Paving Technologists, 77, 2008.
Soenen H., Eckmann B., “Fatigue testing of bituminous binders with a dynamic shear
rheometer”, Proceedings of the 2 nd Euroasphalt and Eurobitume Congress, Vol. I, pp.
827-834, Barcelona, Spain, 2000.
Soenen H., De Visscher J., Vanelstraete A., Redelius P., “The influence of thermal history
on binder rutting indicators”, Road Materials and Pavement Design, 6 (2), 2005.
Stuart K.D., Mogawer W.S., “Validation of the Superpave asphalt binder fatigue cracking
parameter using the FHWA’s accelerated loading facility”, Journal of the Association of
Asphalt Paving Technologists, 71, pp. 116-146, 2002.
Influence of the Bitumen Properties on the

Functional and Rheological Behaviour of
Asphalt Rubber Binders
Joana Peralta* ― Hugo M.R.D. Silva* ― Ana V.A. Machado** ―

Jorge Pais*
* Department of Civil Engineering

University of Minho, Campus de Azurém
4800-058 Guimarães, Portugal
joana@civil.uminho.pt
hugo@civil.uminho.pt
jpais@civil.uminho.pt
** IPC-Institute of Polymers and Composites

Department of Polymer Engineering
avm@dep.uminho.pt
ABSTRACT. It is estimated that about ten kilograms of tires are discarded per inhabitant
annually. The negative impact of this residue can be reduced, since rubber can be reused
as a constituent of asphalt rubber (AR) binder in road pavements. However, the materials
which constitute the AR binders and their interaction are not sufficiently characterized.
In this work several base bitumens interacted with crumb rubber, in order to produce AR
binders, which were subsequently separated, by using a modified “Basket drainage method
” to recover the residual bitumen and rubber. Additionally, a new “Sphere AR production
simulator” method was developed to reproduce the bitumen aging without the contribution of
the rubber. EN 12591 and EN 14023 standard tests and the Dynamic Shear Rheometer (DSR)
were used to evaluate the changes in the properties of the binders during AR production. It
was observed that the characteristics of the base bitumen significantly affect the AR binder
properties (mainly for interactions with softer bitumens).
KEYWORDS: “Sphere AR production simulator”, DSR, Rheology, Swelling, “Basket drainage

method”.
1. Introduction
Approximately 9 to 10 kg of rubber tires per inhabitant and year are currently discarded
in the industrialized societies. Although tires are not regarded as a dangerous residue, their
hollow shape usually brings sanitary problems and difficulties in their final deposition. When
ignited, the resulting fire is impossible to extinguish, releasing hazardous smoke. Finally,
the vulcanized rubber of tires cannot be recycled or used for the same proposes as the
unvulcanized polymer.
Furthermore, the society is wasting materials from the tire with a high intrinsic value,
namely its main constituent: vulcanized natural and synthetic rubber. Thus, the introduction of
crumb rubber in the production of Asphalt Rubber (AR) mixtures for road pavements should
be considered as a sustainable technology which transforms an unwanted residue into a new
mixture with a high resistance to fracture. However, the addition of rubber in a bituminous
mixture increases its complexity, hence being essential to carry out a study to understand the
interaction between its constituents in order to optimize the performance of AR mixtures.
The main objective of this work is to evaluate the influence of AR components, i.e.
bitumen and rubber, on its final properties. Another aim of this research is to assess the
functional and rheological changes that occur in the bitumen in the process of the AR binder
production.
Different types of bitumen from the same source and obtained in the same distillation
column were used in this study. Initially, the bitumen properties were evaluated using
conventional EN standard tests and rheological tests carried out in the Dynamic Shear
Rheometer (DSR). After that, the bitumen was left to interact with 17.5% of crumb rubber
by mass of AR. Then, a modified “Basket Drainage Method” was used to separate again the
constituents of the AR binder (residual bitumen and depolymerised rubber). The AR binder
and the residual bitumen properties were then evaluated by replicating the same tests used
for bitumen characterization, so as to assess the changes in the functional and rheological
behaviour of the binder.
A new “sphere AR production simulator” test was developed during this work to simulate
the changes that occur in the base bitumen during the production of AR that are independent
from the interaction between bitumen and rubber (such as aging). This way, it would be
possible to measure the changes in the bitumen caused exclusively by the interaction with
crumb rubber.
The main results of this work are: (i) the establishment of a relationship between the
functional and rheological properties of different bitumen and the resultant AR binders; (ii)
the assessment of the rheological changes in the properties of the AR binder and of the base
bitumen as a consequence of bitumen-rubber interaction.
2. Literature review on bitumen-rubber interaction
Presently, the pavement technology and the evaluation of bituminous mixtures are
essentially based on empirical-mechanistic studies. Frequently, the materials that compose
asphalt pavements are not sufficiently characterized, more specifically concerning their
physicochemical constitution, their rheological properties and their mutual interaction at a
microscopic scale.
Therefore, it is essential to develop knowledge about the interaction between the

constituents of bituminous materials used in flexible pavements by carrying out their
characterization at a microscopic level, so as to understand their macroscopic structural and
functional behaviour. This study becomes more significant for modified or unconventional
mixtures due to their complexity (e.g. Asphalt Rubber). Actually, beyond the traditional
constituents of the bituminous layers of the pavements, the use of crumb rubber (recycled
from used tires) should be studied as a form of environmental protection and as a pavement
performance enhancer.
Absorption of bitumen components by the rubber inevitably depletes the bitumen of the
absorbed components and, consequently, modifies its properties, by making it stiffer and
brittler (Singleton et al., 2000; Artamendi et al., 2002; Airey et al., 2003). Furthermore, the
rubber particles may also suffer some form of degradation (mainly devulcanization and
depolymerisation) when they are mixed with bitumen at high temperatures for prolonged
periods of time (Billiter et al., 1997; Zanzotto et al., 1996). The extent of swelling and
degradation depends on the nature of the rubbers, the chemical composition of bitumen and
the mixing conditions of time, temperature and degree of agitation. In addition, these processes
will determine the mechanical properties of the crumb rubber modified binders (Abdelrahman
et al., 1999). Asphalt rubber binders are also very dependent on asphalt characteristics: lower
saturates and asphaltenes contents improve the asphalt capacity to dissolve rubber (Billiter et
al., 1996).
Blending crumb rubber into asphalt is believed to improve its elastic and energy absorption
properties, which are directly related to the resistance of the binder to cracking and rutting
failures. The addition of crumb rubber also improves the resistance of asphalt binders to low
temperature cracking (Gopal et al., 2002). Bahia et al. (1995) concluded that the impact of
crumb rubber content on the reduction of low temperature cracking is a linear function of the
rubber content and that the effect of the rubber is less significant for low viscosity asphalts.
The use of low penetration grade bitumens in asphalt-rubber mixtures reduces the rate and
the amount of swelling of the crumb rubber particles. However, any changes in the rheological
properties of the binder following rubber‑bitumen interaction could result in the binder
becoming embrittled (losing flexibility and capacity of resisting cracking and fretting). The
use of high penetration grade bitumen will increase the rate and the amount of rubber swelling
and therefore the shape and rigidity of the rubber, while the binder should still have sufficient
flexibility, following the rubber-bitumen interaction, to resist cracking and fretting (Airey et
al., 2003).
The basket drainage bitumen absorption method (Rahman, 2004) was found to be a simple
and effective way to monitor rubber-bitumen interaction, when the crumb rubber particles
absorb bitumen and swell when added together at mixing temperatures. The rate and amount
of absorption is mainly dependent on the mixing temperature and on the complex chemical
nature of bitumen, but only marginally dependent on the bitumen type and grade. In addition
to normal oxidation, the residual bitumen experienced further changes in mechanical and
rheological properties in terms of increase in stiffness, elasticity, viscosity and reduction in

penetration.
Rheology is the science concerned with the aspects of deformation of real bodies under
the influence of external stresses (Ferguson et al., 1991). The concepts of rheology should be
used to define the characteristics of bituminous binders, due to the importance of the time-
dependent deformation or creep response that occurs when binders are subjected to loads (Tia
et al., 1987).
Asphalt rubber (produced by the wet method) is a very special binder if compared to
others. Extremely low phase angles at high temperatures/low frequencies and relatively
high phase angles and low stiffness at very low temperatures/high frequencies make it a
very interesting binder (van de Ven et al., 2003). The addition of crumb rubber to bitumen
decreases the elastic and viscous moduli at low temperatures and, therefore, it causes an
increase in binder flexibility. On the contrary, at high temperatures a significant increase
in both moduli and a notable drop in the loss tangent values result in a more elastic binder.
Furthermore, it can be deduced that the thermal susceptibility of the binder is clearly reduced
as a consequence of rubber addition. Consequently, enhanced resistance to permanent
deformation, low-temperature and fatigue cracking should be expected in the resulting asphalt
rubber mixtures (Navarro et al., 2005).
3. Asphalt rubber production, separation and characterisation
3.1. Bitumen selection, sample preparation and crumb rubber selection
3.1.1. Bitumen samples
In order to control the variables to be studied, all the bitumen samples were obtained from
the same batch of bitumen production on the refinery (the same distillation column), since
bitumen characteristics are different depending on the petroleum sources and their processing.
Asphalt rubber properties are very dependent on the used bitumen. Thus, this work
was begun by choosing several grade bitumens in order to evaluate their influence on the
AR characteristics. Due to the difficulty of the refinery to supply other bitumens than the
commercial line ones (10/20, 40/50, 60/70 and 150/200), these commercial bitumens were
combined in different proportions in order to obtain sixteen base bitumen samples (Table 1).
Table 1. Proportions of commercial bitumen used to obtain the sixteen base bitumen samples
Sample labelling A B C D E F G H I J K L M N O P
10/20 1.00 0.75 0.50 0.25 0.25 0.50 0.75

Commercial
40/50 0.25 0.50 0.75 1.00 0.75 0.50 0.25

bitumen
60/70 1.00 0.75 0.50 0.25 0.25 0.50 0.75

150/200 0.25 0.50 0.75 1.00 0.75 0.50 0.25
The laboratorial characterisation of these sixteen base bitumen samples was carried out in
order to evaluate the following properties:
- Functional properties, namely the penetration at 25 ºC (EN 1426), ring and ball
softening point (EN 1427), the dynamic viscosity (EN 13302) and the elastic recovery
(ASTM 5329);
- Rheological properties, using the Dynamic Shear Rheometer (only for eight of the
studied bitumens) to determine the viscosity, the modulus and the phase angle.
In addition, the four commercial bitumens were also subjected to a “sphere AR production
simulation” test developed in this work to simulate the changes that occur in the base bitumen
during the production of AR that are independent from the interaction between bitumen and
rubber. After these procedures, the changes in the functional and rheological properties of the
residual bitumens were assessed, in comparison with the AR binder residual bitumens. This
way, it was possible to measure the changes in the bitumen caused exclusively by the rubber.
3.1.2. Crumb rubber sample
The crumb rubber used in this work was produced by the cryogenic process and, according
to the supplier, was obtained by the cryogenic grinding of 30% of truck tires and 70% of car
tires. The supplied crumb rubber was sieved in order to select and use only the fraction passed
in sieve ASTM #20 (0.850 mm) and retained in sieve ASTM #40 (0.425 mm).
3.2. Production and separation of the asphalt-rubber binder
The method used to produce AR and collect the desired samples of AR, of recovered
swelled rubber and of residual bitumen was the “basket drainage method” (Rahman, 2004).
The asphalt rubber production facility is constituted by several equipments, presented in
Figure 1, assembled in a laboratory ventilation chamber.
Figure 1. Scheme and photo of the asphalt rubber production facility
The wire basket used in the AR production facility was manufactured with a wire mesh,
which was previously observed in an optical microscope (Figure 2) to measure its average
opening dimension (# 0.470 mm). The microscope was able to draw an idealized picture of
the mesh in order to assess, automatically, the referred opening dimension.
The process used to produce and collect the several samples of AR, of recovered swelled
rubber and of residual bitumen was the following one:
- Heating of 1 kg of each base bitumen (Bb) at 180 ºC and collecting of a sample;
- Introduction of 17.5% of crumb rubber by mass of asphalt rubber;
Figure 2. Microscopic photo and drawing of the wire mesh used to manufacture the basket
- Continuous heating of the asphalt rubber binder at 180 ºC for 60 minutes (digestion
time) while stirring the binder at a velocity of 230 rot/min (Figure 1);
- Collection of a sample of the produced asphalt rubber (AR);
- Suspension of the wire basket with the AR binder in an oven at 180 ºC for 15 min, so
as to separate its constituents (Figure 3) and collect a sample of the residual bitumen
(Res) (Figure 4).
Figure 3. Separation between the crumb rubber and the residual bitumen of the AR
Figure 4. Appearance of the residual bitumen and of the recovered rubber after separation
3.3. “Sphere AR production simulator” method
One of the objectives of this research is to determine the changes in the bitumen
characteristics caused by the interaction with crumb rubber.
When the bitumen is heated, smaller molecules are volatilized (aromatics), while other
molecules can associate themselves with bigger molecules (asphaltenes). Furthermore, when
bitumen interacts with rubber, the rubber particles absorb the light parts of bitumen and swell.
Since those phenomena simultaneously occur during the AR production, it seems unfeasible to
evaluate the changes in the bitumen structure caused by every previous phenomenon. Hence, a
new method was developed during this research in order to determine the ratio of the changes
exclusively induced by heating and stirring the bitumen during the AR production, entitled
“sphere AR production simulator”. The aim of this method is to simulate the AR production,
by replacing the rubber particles with metallic spheres (roughly with the same dimensions
of the rubber particles) in order to reproduce the evaporation and association processes
occurring in the bitumen during the heating period for AR production, independently from
the interaction with rubber. Before the experimental procedure, the metallic spheres were
measured (øeq ≈ 0.825 mm) using an optical microscope and their volumetric weight was
evaluated (7.75 kg/dm3).
The “Sphere AR production simulator” was carried out in a reduced scale AR production
facility (similar to the one used for AR production), and consisted in the following procedures:
- Heating of approximately 100 g of each commercial bitumen at 180 ºC;
- Introduction of approximately 240 g of metallic spheres (these spheres have a volume

equivalent to 17.5% of crumb rubber by mass of asphalt rubber);
- Continuous heating of the binder (bitumen and metallic spheres) at 180 ºC for 60
minutes (digestion time) while stirring the binder at a velocity of 230 rot/min;
- Suspension of the wire basket with the binder (bitumen and metallic spheres) in an
oven at 180 ºC for 15 min, so as to separate its constituents;
- Collection of a sample of the resulting bitumen (Sb) for evaluation of its functional
properties and future comparison with the base (Bb) and residual (Res) bitumen
properties.
3.4. Analytical determinations
During this work a series of analytical determinations were carried out in several samples
of binders obtained during the AR production, namely by evaluating the penetration at 25
ºC (Pen), the ring and ball softening point (R&B), the dynamic viscosity (DV), the elastic
recovery (ER) and the rheological properties using the Dynamic Shear Rheometer (DSR).
Table 2 systematizes the laboratorial research analysis performed in this work with the
base bitumens (Bb), asphalt rubbers (AR), residual bitumens (Res) and “Sphere simulator”
recovered bitumen (Sb).
3.5. Determination of the linear viscoelasticity range of stresses to be used in the rheological
characterization of the binders
From the sixteen initial samples of base bitumen, only eight were selected (based on their
functional properties) to evaluate their rheology. A DSR was used to assess the rheological
properties of the base bitumen, AR and residual bitumen of those eight bitumens.
Table 2. Analytical determinations carried out in this work
Sample
A B C D E F G H I J K L M N O P
labelling
Pen Bb Bb Bb Bb
Analytical determinations
Bb Bb Bb Bb Bb Bb Bb Bb Bb Bb Bb Bb
AR AR AR AR
R&B AR AR AR AR AR AR AR AR AR AR AR AR
Res Res Res Res
Res Res Res Res Res Res Res Res Res Res Res Res
DV Sb Sb Sb Sb
ER AR AR AR AR AR AR AR AR AR AR AR AR AR AR AR AR
Bb Bb Bb Bb Bb Bb Bb Bb
DSR AR AR AR AR AR AR AR AR
Res Res Res Res Res Res Res Res
The rheological tests were carried out at different temperatures (25, 35, 50, 80, 110, 140,
160 and 180 ºC) in order to determine the rheological variation of the binder properties at
all in service, laying down and mixing stages. For each temperature, frequency sweep tests
between 0.1 and 10 Hz were performed in the linear viscoelasticity range, using parallel plate
geometry with 40 mm of diameter and 1.0 mm of gap between plates.
Initially, it was necessary to establish the input stress at each DSR test temperature.
Therefore, previous stress sweep tests were performed at 1.0 Hz to determine the linear
viscoelasticity region (at each test temperature). The linear viscoelasticity range can be
identified in Figure 5 as being the range of stresses where no change is observed in the values
of the rheological characteristics measured at each test temperature.
1,00E+05
1,00E+04
1,00E+03
Measured values
1,00E+02
25 e 35
1,00E+01 50
80
110
1,00E+00 140
160
180
G'' (Pa)
1,00E-01
η* (Pa.s)
IG*I (Pa)
G*/sen(δ) (Pa)
1,00E-02
1,00E+00 1,00E+01 1,00E+02 1,00E+03
Stress (Pa)
Figure 5. Summary of the results of the rheological stress sweep tests at 1 Hz
The selection of the stress values to apply on the rheological tests at each temperature
was made to guarantee that the frequency sweep tests will be carried out in the linear
viscoelasticity region for all tested binders, thus being possible to compare their rheological
behaviour.
Rheological time sweep tests were also performed, at a constant temperature, stress and
frequency, to confirm that no structural modifications occurred in the binders throughout the
tests (60 minutes were used to evaluate the changes during AR production). The rheological
characteristics assessed in the DSR were the complex modulus (G*), storage or elastic
modulus (G’), loss or viscous modulus (G’’), complex viscosity (η*) and phase angle (δ).
4. Results and discussion
4.1. Functional characterization of the studied binders
The functional characteristics of different samples of binder collected throughout the AR

production, namely the base bitumen (base) and the asphalt rubber binder (AR), were assessed
in this part of the work in order to evaluate the changes in the binder properties caused by
the addition of rubber to the bitumen. Thus, Figure 6 presents the functional characteristics
(penetration, R&B and dynamic viscosity) of the sixteen different base bitumens (A to P) and
of the resultant AR binders produced in this study, as well as the differentials (Δ) between the
functional properties of each matching pair of base bitumen and AR binder.
Based on the values of the functional properties, it was possible to classify and sort out
the sixteen base bitumens. The results of the different tests were very consistent with the
proportions of the used commercial bitumens. Actually, the changes in the constitution of the
samples (from harder to softer bitumens) originated an ordered sequence for the evaluated
functional properties, represented by the potential trending lines presented on the left side of
Figure 6.
DVAR DVbase Pen AR Pen base R&B AR R&B base ERAR 100
1,0E+04
DV (cP); Pen (1/10 mm)
80
80
ER(%) ; R&B (ºC)

1,0E+03 60
Δ (%)
40
60
1,0E+02
20
ΔViscodity ΔPen ΔRing and Ball
1,0E+01 40 0
0 20 40 60 80 100 120 140 0 20 40 60 80 100 120 140 160
Pen at 25 ºC of base bitumen (1/10 mm) Pen at 25 ºC of base bitumen (1/10 mm)
Figure 6. Functional properties of the ARs compared to the ones of the base bitumens
For every functional property evaluated, the AR binder results presented greater
dispersions, mainly because this material is more heterogeneous than the bitumen. As it was
expected, the AR binders presented much lower values of penetration and higher values of
ring and ball temperatures and dynamic viscosity than the corresponding base bitumens. The
dynamic viscosity test is the most effective to find variations in the behaviour of the different
materials, since it distinguishes the several base bitumens (A to P) and the corresponding AR
binders very well.
The influence of the type of base bitumen in the functional characteristics is clearer in
the evaluation of the base bitumens. In opposition, the functional characteristics of the AR
binders produced with the different bitumens are very similar, mainly the dynamic viscosity
and the elastic recovery which are almost constant. The comparison between the several
types of bitumen used to produce AR also showed that the highest variations in the functional
characteristics of the binder (AR comparison with base bitumen) are obtained with the softer
bitumens. Thus, even the 150/200 bitumen (M), which is not generally used to produce asphalt
rubber and hot mix asphalts, revealed to be a good alternative to produce AR binders with a
final penetration similar to the one of the 60/70 base bitumen (I).
One of the aims of this research is to determine the changes in the bitumen characteristics
caused exclusively by the interaction with crumb rubber. The “sphere AR production
simulator” method was applied to obtain recovered bitumens (Sb) that were influenced by the
processes occurring in the bitumen during the AR production, independently of the interaction
with rubber.
The AR is a heterogeneous material, which can be separated in two different phases, the
swelled rubber and the residual bitumen. The swelled rubber, by itself, does not promote the
adhesion between the aggregates in the production of the AR mixtures. Thus, it is reasonable
to consider the swelled rubber as a special category of aggregates, being the residual bitumen
the final binder of the AR. Therefore, it is essential to evaluate the behaviour of this residual
bitumen of the AR binder in order to optimize the performance of the AR mixtures.
The measured values of the functional properties of the “sphere” and residual bitumens
are presented in Figure 7, in comparison with the corresponding base bitumens. The graph on
the left side of Figure 7 represents the evolution of the functional properties of the residual,
“sphere” and base bitumen. The right side of Figure 7 shows the proportional changes in the
functional properties of the base bitumen (differential percentage in comparison with the base
bitumen) due to the “sphere AR production simulator” method and to the interaction with
rubber.
A LB C KD EJ FI GP HO N M 100 DV Pen R&B

1,0E+03 80
ΔSphere ΔRes
Pen at 25 ºC (1/10 mm);
75
80
70
Viscosity (cP)
Ring and Ball (ºC)

1,0E+02
65 60
Δ (%)
60
40
55
1,0E+01
50 20
Res Sphere Base 45
1,0E+00
Pen Viscosity Ring and Ball 0
40
0 20 40 60 80 100 120 140
0 20 40 60 80 100 120 140
Figure 7. Functional properties of the “sphere” and residual bitumens in comparison with the
corresponding base bitumens
The changes on the bitumen constitution, after the “sphere” method and the interaction
with rubber, can be evaluated by measuring the variation in the functional properties of
binders. These changes are more perceptible in the DV test, while the R&B test is the less
sensitive one.
The results obtained for the “sphere” recovered bitumen are caused by the changes
occurred during the production of the AR that are independent from the bitumen-rubber
interaction, being mainly caused by the volatilization of light parts of the bitumen. The
variation on the functional properties of the recovered bitumen after the “spheres” method is
superior for softer bitumens and has an average value of 10% in the R&B test, 20% in the pen
test and 35% in DV test. The results obtained for the residual bitumen show that the changes
in the bitumen caused by rubber are superior to the ones observed for the “sphere” bitumen,
which indicates a significant absorption of light bitumen parts by the rubber particles.
For the conditions adopted in this work, it was noticed that the variation in the functional
properties caused by the bitumen-rubber interaction remains constant for binders obtained
from base bitumens with penetration values superior to 30 dmm (as presented in the shaded
band for each property on the right side of Figure 7). For these binders produced with base
bitumen with pen values superior to 30 dmm, the changes in the functional properties for the
AR residual bitumen doubles or triples the variation obtained for the “sphere” bitumen. Thus,
when the bitumen has some concentration of light molecules able to be absorbed by the rubber
and/or volatilized due to the temperature and agitation, these two phenomena coexist almost
equally.
However, the use of harder bitumens (pen < 30 dmm) in the AR originates binders with
ratios absorption/volatilization greatly superior to 1. For these binders the changes caused
by the evaporation of the volatile molecules of bitumen are considerably inferior to the
ones caused by the rubber-bitumen interaction. Thus, the process of diffusion of the smaller
molecules of the bitumen into the rubber is more efficient than the bitumen volatilization,
occurring initially during the AR production. The quantification of the bitumen changes
during the production of AR is difficult to evaluate, since several inter‑dependent changes
simultaneously occur in the rubber and in the bitumen, such as diffusion and volatilization.
However, the presented results allowed a better understanding of the extension of these two
phenomena.
4.2. Rheological characterization of the studied binders
The viscoelasticity of bitumen can be described by plotting the shear stress – τ (Pa)
vs. shear rate – γ̇ (s-1), as presented in Figure 8 for the base bitumen 150/200 at different
temperatures.
The bitumen M (150/200) behaves as a non-Newtonian linear viscoelastic material, or a

Bingham solid, for temperatures below 150 ºC. Above the referred temperature it behaves as a
Newtonian fluid. The slope of the several lines presented in Figure 8 increases as temperature
reduces, as a consequence of the rise in the stiffness and viscosity of the bitumen.
Only eight bitumens were selected to carry out rheological tests. In order to exemplify the
typical behaviour of the bitumen in the DSR tests, some results are only presented for the base
bitumen I (60/70) since it is the commercial bitumen typically used to produce AR binders.
1,6E+03 25 ºC
y = 17578x + 2,3877
35 ºC
y = 3766,4x + 2,3447 50 ºC
1,2E+03 y = 256,9x + 2,1545 80 ºC
110 ºC
y = 7,9568x + 1,3764 140 ºC
8,0E+02 160 ºC
τ
180 ºC
4,0E+02
y = 0,9637x + 0,0499
y = 0,2441x + 0,0638 y = 0,1396x y = 0,0949x
0,0E+00
0,0E+00 2,0E+02 4,0E+02 6,0E+02 8,0E+02
γ
.
Figure 8. Shear stress vs. shear rate curves for bitumen 150/200, at different temperatures
One of the most effective ways of representing the data obtained in the DSR tests is by
drawing master curves. These curves can be plotted because it is generally accepted that
the bitumen has a simple rheological behaviour, being applicable the time-temperature
superposition principle (TTS), by using the Arrhenius equation [1] or the Williams-Landel-
Ferry (WLF) equation [2]. The result of the superposition of curves at different times
(frequencies) and temperatures are the master curves, that cover a large range of time or
frequency domains.
η* = k.exp[Ea/(R.T)] for T < Tg [1]
Log aT = -C1.(T-Tg)/(C2+T-Tg) for T > Tg [2]
With the initial data it is not possible to determine the glass transition temperature (Tg)
value, and consequently it is not possible to identify the equation that better approaches
the data used to draw the master curves. Thus, the easiest way to choose the best equation
to fit the data is to draw a graphic of the logarithm of the stiffness parameters versus the
inverted temperature. The Arrhenius equation should be used for a linear relation between the
plotted data, or else the best approach is obtained with the WLF equation. The first graph of
Figure 9 (plot of η* vs. 1/T) shows a linear relation between the plotted parameters only for
temperatures above 100 ºC (equation [1] must be used to draw the master curve). Below 100
ºC the master curve must be drawn with equation [2]. The application of equation [2] has three
output values: C1, C2 and Tg. This last parameter is a characteristic value of each material. Its
determination depends on the geometry used to assess the rheological parameters and on the
material behaviour. In this case the Tg parameter could only be achieved by using a bending
beam geometry. However, the master curve shows a point of interception (nearly at 30 ºC) of G’
and G’’ lines (tg δ = 1) for temperatures between 25 ºC and 100 ºC, indicating some change in
the rheology of the bitumen.
Above 100 ºC the graphic of η* vs 1/T can be approached to a straight line, and thus the
Arrhenius equation was used to plot the master curve. The Arrhenius equation also provides
a characteristic value of each material, the activation energy (Ea). In this case, the Ea value
should not be considered, since it would be necessary to use other geometry (rotating
spindle) to assess more significant values. However, this master curve also shows a point of
interception of G’ and G’’ lines (nearly at 170 ºC) that corresponds to a new alteration in the
rheology of the bitumen.
1,0E+05
1,0E+03
η* (Pa.s)
1,0E+01
1,0E-01
5,0E-03 1,5E-02 2,5E-02 3,5E-02
1/T (ºC)
Figure 9. Master curves of the base bitumen I (60/70) at two different ranges of temperatures
With the rheological results obtained in this work it is not possible to draw a unique master
curve (Figure 9) covering the entire range of tested temperatures. It will be necessary to use
a variety of diverse geometries in different rheometers (for each range of temperatures), and
more temperatures should be tested to guarantee a good superposition of the obtained data.
In order to evaluate the differences in the rheological behaviour of the base bitumen, AR
binder and residual bitumen at different temperatures, the Figure 10 summarizes the main
rheological results obtained at 1.0 Hz for bitumen I (60/70). For example, the separation of the
G’’ and G*/sen(δ) lines for temperatures below 80 ºC only occurs in the AR binder, because of
the increase in the values of G’ (the rubber particles are essentially elastic) in relation to G’’,
thus reducing the values of the phase angle (δ) for similar values of G*.
G' (Pa) G'' (Pa)

tg δ η* (Pa.s)
1,0E+06 G*/ sen(δ) (Pa)
1,0E+04
Measured values
1,0E+02
1,0E+00
Base Res AR
1,0E-02
20 40 60 80 100 120 140 160 180
T (ºC)
Figure 10. Rheological characteristics of the base bitumen 60/70 and the corresponding AR
binder and residual bitumen
The inclusion of the crumb rubber in the bitumen not only rises the stiffness of the new
binder (AR), but also changes the development of the storage (G’) and loss (G’’) modulus
with the temperature. The G’ value greatly rises due to the presence of an elastic solid, the
rubber, that increases the elasticity of the AR. Actually, for temperatures below 50 ºC, the
AR presents G’ values superior to those of G’’, which is consistent with the better related
resistance of the AR to rutting, fatigue and reflective cracking distresses of the pavement at in
service temperatures.
Between 80 and 160 ºC, the rheological behaviour of the AR binder, at the chosen
frequency (1 Hz), is similar to the one of the base bitumen, even if some differences are
evident in the values of G’, G’’ and G*/sen(δ) that are higher but proportional to the ones of
the base bitumen. Thus, the AR binder is stiffer than the base bitumen, but its rheological
behaviour at production and applying temperatures is somewhat similar.
On one hand, for temperatures inferior to 100 ºC, the differences in the rheological
properties between the residual and the base bitumen are evident, presenting the maximum
discrepancy between 50 and 80 ºC. In this range of temperatures, the residual bitumen is
clearly stiffer than the base bitumen, being even stiffer than the AR binder for temperatures
inferior to 35 ºC. On the other hand, for temperatures superior to 110 ºC, the differences
between the rheological properties of the residual and the base bitumen are minor, reducing as
the temperature increases and trending to the same value at 180 ºC. However, at these higher
temperatures the AR binder still presents rheological properties clearly superior to the base
and residual bitumens.
Figure 11 shows the isochronal plots of the rheological parameters of the base bitumen and
of the resulting AR binder at a frequency of 1.0 Hz and temperatures between 25 and 180 ºC
for eight bitumens selected to carry out the DSR tests (A, B, E, I, J, M, O and P).
1,0E+06 A B 1,0E+06 A B
E I E I
1,0E+05 J O 1,0E+05 J O
P M P M
1,0E+04
1,0E+04
1,0E+03
G' (Pa)
G'' (Pa)
1,0E+03
1,0E+02
1,0E+02
1,0E+01
1,0E+01
1,0E+00
1,0E-01 1,0E+00
........ Base _____ AR
1,0E-02 1,0E-01
20 40 60 80 100 120 140 160 180 20 40 60 80 100 120 140 160 180
T(ºC) T(ºC)
1,0E+06 A B 1,0E+07 A B
E I E I
J O 1,0E+06 J O
P M P M
1,0E+04 1,0E+05
G*/sen(δ) (Pa)
1,0E+04
η* (Pa.s)
1,0E+02 1,0E+03
1,0E+02
1,0E+00 1,0E+01
1,0E+00
1,0E-02 1,0E-01
20 40 60 80 100 120 140 160 180 20 40 60 80 100 120 140 160 180
T(ºC) T(ºC)
Figure 11. Isochronal plots of the rheological properties of the base bitumens and ARs at 1.0 Hz
The AR binder is the stiffest material, being the less susceptible to the change of the type
of base bitumen (plots are very similar) and to the variation of temperature. In fact, the high
increase in the elasticity of the AR binder for all studied temperatures, caused by the presence
of rubber particles, totally changes the rheology of the binder, by reducing the influence of the
temperature in the behaviour of the binder (AR does not become a viscous liquid above 100 ºC).
Asphalt binders must have a set of characteristics in order to allow their adequate use
in road pavement construction at different temperatures. Binders should not be very rigid
at low and medium operating temperatures (to avoid cracking distresses), but they should
be sufficiently stiff at high operating temperatures to improve their resistance to permanent
deformation. The AR binders fulfil these demands for every used base bitumen, even for the
softer bitumen 150/200 (M), since the values of the rheological parameters at 25-35 ºC are
similar to the base bitumen, being much higher at 60-80 ºC. However, the several AR binders
present elevated viscosities at high temperatures, making the mixing and laying down of AR
mixes more difficult.
The changes on the viscoelastic nature of the bitumen after the interaction with crumb
rubber are also presented in Figure 12, which shows the evident differences between the tgδ
plots of the base bitumen and the AR binder for the tested range of temperatures at a frequency
of 1.0 Hz.
Base AR Freq. = 1,0 Hz

A B
1,0E+03 E I
J P
O M
1,0E+02
tgδ
1,0E+01
1,0E+00
1,0E-01
20 40 60 80 100 120 140 160 180
T (ºC)
Figure 12. Isochronal plots of tg δ for the studied binders (base bitumens and ARs) at 1.0 Hz
The maximum value of tgδ observed in Figure 12 for the several base bitumens occurred
for temperatures within 80 and 110 ºC. This means that a reversible phenomenon of relaxation
takes place in this range of temperatures for all the bitumens, because the same results were
obtained after repeating the DSR test in the same samples of bitumen. In comparison with the
base bitumen, the maximum value of tgδ and the relaxation phenomenon observed in Figure
12 for all AR binders occurred for higher temperatures within 120 and 150 ºC. Moreover, at
temperatures below 110 ºC the values of tgδ of the AR binders are inferior to the ones of the
base bitumen (due to the increased elasticity of AR), but above 110 ºC the AR and the base
bitumen values of tgδ are similar. Finally, it was observed that the AR values of tgδ are almost
independent from the type of bitumen used in the AR for temperatures between 35 and 80 ºC.
The viscoelastic rheology of the several AR binders, base and residual bitumens, can be
easily observed for temperatures between 25-50 ºC in the Cole-Cole chart shown in Figure 13.
Base Res AR A B E I
2,8E+05
J P O M
2,1E+05
G'' (Pa)
1,4E+05
7,0E+04
1,0E+00
1,0E-02 2,0E+05 4,0E+05 6,0E+05 8,0E+05
G' (Pa)
Figure 13. Comparative Cole-Cole chart for the AR binder, base and residual bitumen
After analyzing the three sets of curves (AR binder, residual and base bitumen), a
generalized hardening was observed for the residual bitumen, that is more obvious for the AR
binder, due to the increase of elasticity (G’) and decrease of viscosity (G’’). The Cole-Cole
curves of the several AR binders and residual bitumens have slightly the same form of the
corresponding base bitumens and are organized by the same order. Thus, the softer bitumen (M)
is always the lower and shorter curve and the harder bitumen (A) is the upper and longer curve
of each set, being the other six bitumens organized by the same logic between the previous
limits. Besides the stiffness increase, there is no sign of other specific changes in the residual
bitumen caused by the interaction with the rubber for temperatures between 25 and 50 ºC.
The set of Cole-Cole curves of the AR binders are wider than the ones of the base and
residual bitumens, denoting a major influence of the crumb rubber in the modification of
the softer bitumens. In fact, during the production of the AR binder, the softer base bitumen
presented the best interaction with the crumb rubber. In summary, the AR binders produced
with softer bitumens presented good rheological characteristics at in service temperatures,
thus showing the benefits of the strong interaction between that bitumen and the crumb rubber.
The changes induced in the rheology of the base bitumen during the AR production can be
evaluated by comparison with the rheological properties of the residual bitumen, as presented
in Figure 14 for all studied bitumens at the frequency of 1.0 Hz.
The analysis of the charts on the left side of Figure 14 showed that the isochronal curves
of the residual bitumen are less scattered than those of the base bitumen, particularly at higher
test temperatures, probably due to the several processes occurred during the AR production.
The rheological property that better reflects the described behaviour is the storage modulus.
The G’ isochrones defining the evolution of the elastic character of the different bitumens
became completely independent from the constitution of the different residual and base
bitumen and also from temperature, since those isochrones nearly overlap defining a single
line parallel to the horizontal axis above 140 ºC. The weak interactions that maintained the
“networks” of asphaltenes/resins no longer exist at 140 ºC and the bitumen behaves as a
Newtonian fluid, where the asphaltenes are totally peptized by the resins and dispersed in the
maltenes.
1,0E+06 Base Res 100 ΔA ΔB ΔE ΔI

A B
ΔJ ΔO ΔP ΔM
E I
J O 80
1,0E+04
P M
ΔG' (%)
60
G' (Pa)
1,0E+02
40
1,0E+00
20
1,0E-02 0
20 40 60 80 100 120 140 160 180 20 40 60 80 100 120 140 160 180
T(ºC) T(ºC)
1,0E+07
100
1,0E+05 80
G'' (Pa)
ΔG'' (%)
60
1,0E+03
40
1,0E+01
20
1,0E-01 0
20 40 60 80 100 120 140 160 180 20 40 60 80 100 120 140 160 180
T(ºC) T(ºC)
1,0E+06
100
1,0E+04 80
60
Δη*(% )
η* (Pa.s)
1,0E+02
40
1,0E+00
20
1,0E-02 0
20 40 60 80 100 120 140 160 180 20 40 60 80 100 120 140 160 180
T(ºC) T(ºC)
1,0E+03 100
80
1,0E+02
Δtg δ (%)
60
tgδ
1,0E+01
40
1,0E+00
20
1,0E-01 0
20 40 60 80 100 120 140 160 180 20 40 60 80 100 120 140 160 180
T(ºC) T(ºC)
Figure 14. Isochronal plots of the rheological characteristics of base and residual bitumens
and the corresponding differentials at 1.0 Hz
As it was expected, the graphics on the right column of Figure 14 confirmed the results
previously obtained with the comparative analysis of the functional properties of the base
and residual bitumens. Therefore, it was observed that the constitution of the base bitumen
has a great influence in the rheological and functional properties of the residual bitumen.
The softer bitumen presented the greatest changes in its rheological properties after the AR
production, mainly for temperatures between 25-35 ºC. However, the maximum variations in
the rheological properties of the base and residual bitumen were usually observed between 35
and 110 ºC.
In conclusion, the softer a base bitumen is, the more affected it will be during the AR
production process, altering more drastically its initial characteristics. Thus, the soft base
bitumens used in AR binders become significantly stiffer and eventually more adequate to be
applied in pavements conferring an increased resistance to weather and traffic actions.
4.3. Comparison between the rheological and functional properties of binders
During the pavement life cycle, the asphalt binders used in bituminous materials and
pavements should always present an adequate behaviour, initially during the mixing and lay
down phases at elevated temperatures, then during the operating phase, in order to resist to the
weather and traffic actions at low and medium temperatures and, finally, in the removal and
recycling of the distressed mixtures (this last phase was not studied in this work). The results
of the two different approaches (functional and rheological, for a reference frequency of 1 Hz)
used to assess the behaviour of the different studied binders (A, B, E, I, J, M, O and P) will be
analysed and compared in this part of the work.
At medium operating temperatures causing fatigue cracking distresses, the functional

parameter measured in this work was the penetration at 25 ºC. Thus, Figure 15 relates the
penetration value of the several binders to the rheological properties measured at 25 ºC. The
range of values of the penetration results is visibly inferior to the rheological classification
for the studied binders, and thus the vertical axis (rheology) is presented in a logarithmic
scale. Actually, the penetration values of AR binders and residual bitumens only show little
variations, being always inferior to 50 dmm.
1,0E+06 1,0E+06
G'' (Pa)
G' (Pa)
1,0E+05 1,0E+05
1,0E+04 1,0E+04
0 50 100 150 0 50 100 150
Pen. at 25 ºC (1/10mm) Pen. at 25 ºC (1/10mm)
2 1,0E+06
1,0E+05
η* (Pa.s)
tgδ
1,0E+04
0 1,0E+03
0 50 100 150 0 20 40 60 80 100 120 140
)
Pen. at 25 ºC (1/10mm Pen. at 25 ºC (1/10mm)
Figure 15. Rheological vs. functional properties of the studied binders at operating
temperatures causing fatigue cracking
The best fit to the results that relate the functional and rheological properties of the binders
at 25 ºC was obtained with potential trending lines. The exceptions were tgδ and G’’, which
presented polynomial best fit lines with significant dispersion (probably the relation between
these properties and the penetration value is not very good at 25 ºC). Comparing the rheology
of the different binders (base and residual bitumen and AR) for the same penetration value,
the AR binder is the softest material. This result is consistent with the best fatigue cracking
resistance of this material. The most considerable changes in the characteristics of the binder,
before and after the interaction with crumb rubber, were observed for bitumen 150/200 (M).
The penetration of the base bitumen M (145.3 dmm) decreased considerably for the resulting
AR binder (42.1 dmm) and residual bitumen (45.4 dmm), and the penetration becomes similar
to the one of the base bitumen 40/50 (E). However, the changes in the rheology of bitumen
M were not so severe. By opposition, the functional and rheological characteristics of base
bitumen A are nearly the same of the corresponding AR binder and residual bitumen. Thus, the
presence of rubber in the binder A does not affect the AR binder behaviour at this temperature
(25 ºC).
After that, the evolution of the rheological properties of the different AR binders was
compared with their elastic recovery in Figure 16. These results are important since the
changes occurring in harder binders (bitumen A) after the interaction with crumb rubber could
be only evaluated through the analysis of the elastic recovery tests of AR binders.
G*/ sen(δ) (Pa) G' (Pa) G'' (Pa)

1,E+07 η* (Pa.s) tg δ 0,8
Measured values
0,6
1,E+06
tgδ
0,4
1,E+05
0,2
25 ºC
1,E+04 0
40 45 50 55 60
Elastic recovery(%)
Figure 16. Relation between the elastic recovery at 25 ºC and the rheological parameters of
the different AR binders
The elastic recovery of AR binders increases as the base bitumen used for their production
is softer. The reduction of the stiffness of the residual bitumen surrounding the rubber
particles in the AR allows an easier and faster recovery of their form after deformation, while
the increase of the swelling of the rubber particles also confer them an improved volume and
elasticity. The trending lines shown in Figure 16 (except for tg δ) show a reduction in the
rheological properties of the AR with the increase in the elastic recovery. Thus, the elastic
recovery of the AR binders could be eventually increased by reducing the stiffness of the AR
binders.
At high operating temperatures causing rutting distresses, the functional parameter

measured in this work was the ring and ball softening temperature. Thus, Figure 17 relates
the ring and ball softening temperature of the several binders to the rheological properties
measured at 50 ºC.
1,0E+06 1,0E+06
1,0E+05
1,0E+05
G'' (Pa)
G' (Pa)
1,0E+04
1,0E+04
1,0E+03
1,0E+02 1,0E+03
40 50 60 70 80 90 40 50 60 70 80 90
Ring and Ball (ºC) Ring and Ball (ºC)
8 1,0E+05
6
1,0E+04
tg δ...
η* (Pa.s)
1,0E+03
2
0 1,0E+02
40 50 60 70 80 90 40 50 60 70 80 90
Ring and Ball (ºC) Ring and Ball (ºC)
Figure 17. Rheological vs. functional properties of the studied binders at high operating
temperatures causing rutting
The range of values of the ring and ball softening temperature is visibly inferior to the
rheological classification for the studied binders, and thus the vertical axis (rheology) is
presented in a logarithmic scale. The best fit to the results that relate the functional and
rheological properties of the binders at 50 ºC was obtained with exponential trending lines.
The ring and ball temperature presented a very good relation with the rheological properties.
The relations between the functional and rheological properties for the several types of binders
(base, AR and residual) are clearly different, since the trending lines obtained for every binder
are visibly detached. This occurs because the softening temperatures of the several types of
binders are dissimilar, especially the AR binder that melts above 70 ºC. In fact, the softer AR
binder is stiffer than all the base bitumens at these higher temperatures. The ring and ball
values and the rheological properties of the residual bitumen and AR binder show a significant
stiffening of the base bitumen at high operating temperatures after AR production. Unlike the
penetration results, the AR binders and residual bitumens presented a considerable variation of
their ring and ball and rheological properties in relation to the base bitumen, and the materials
produced with softer bitumens were more affected.
At mixing and laying down temperatures, the functional parameter measured in this work
was the dynamic viscosity at 180 ºC. Thus, Figure 18 relates the dynamic viscosity at 180 ºC
to the rheological properties also measured at 180 ºC.
The dynamic viscosity at 180 ºC appears to be more accurate to describe the bitumen
behaviour at this temperature than the rheological parameters, since the range of the DV
results is visibly superior to the rheological classification. Thus, both horizontal (functional)
and vertical (rheology) axes are presented in a logarithmic scale.
1,0E+00 1,0E+01
G'' (Pa)
G' (Pa)
1,0E+00
1,0E-01 1,0E-01
1,0E+01 1,0E+02 1,0E+03 1,0E+01 1,0E+02 1,0E+03
Dynanic Viscosity (cP) Dynanic Viscosity (cP)
8 1,0E+00
6
η* (Pa.s)
tgδ
4 1,0E-01
0 1,0E-02
1,0E+01 1,0E+02 1,0E+03 1,E+01 1,E+02 1,E+03
Dynanic Viscosity (cP) Dynanic Viscosity (cP)
Figure 18. Rheological vs. functional properties of the base bitumen at production and laying
down temperatures
The residual bitumens are clearly stiffer than the base bitumens at 180 ºC (with a great
increase in the dynamic viscosity and rheological properties). Finally, it was observed that the
relation between the dynamic viscosity (functional test) and the complex viscosity (DSR test) is
very good, nearly with the same results obtained in both test methods (since 1 Pa.s = 1000 cP).
5. Conclusions
The main conclusions that can be drawn from the results of this work are the following:
- during the AR binder production, the interaction of the crumb rubber particles with
softer bitumens is clearly superior than with harder bitumens;
- the functional characterization of binders is more effective for bitumens than for AR
modified binders, because the AR functional properties are poorly influenced by the
type of base bitumen used (especially at higher temperatures);
- the AR binder results present greater dispersions in some tests, mainly because of the
heterogeneity of this material;
- the 150/200 bitumen is rarely used to produce AR binders, but it greatly interacts
with the crumb rubber, being a good alternative to produce AR with very good
characteristics (similar to those obtained with 60/70 bitumens) during the life cycle of
the pavement;
- the relation between the functional and rheological properties of the studied binders
depends on the test temperature, being this relation very good for all the studied
binders (base, AR and Res) at higher temperatures (dynamic viscosity and ring and ball
temperature), but not so good at medium operating temperatures (penetration).
Acknowledgements
The authors would like to thank CEPSA and RECIPAV for their support and for the supply
of bitumen and crumb rubber, respectively.
6. Bibliography
Abdelrahman M.A., Carpenter S.H., “Mechanism of interaction of asphalt cement with crumb
rubber modifier”, Transportation Research Record, No. 1661, 1999, p. 106-113.
Airey G.D., Rahman M.M., Collop A.C., “Absorption of bitumen into crumb rubber using the
basket drainage method”, International Journal of Pavement Engineering, Vol. 4, No. 2,
2003, p. 105-119.
Artamendi I., Eastmond G.C., Khalid H., “Influence of crumb rubber modifier (CRM) from
tire waste on the rheological properties of bituminous binders”, Proc. 3rd International
Conference on Bituminous Mixtures and Pavements, Vol. 1, Thessaloniki, 21-22 November
2002.
Bahia H., Davies R., “Role of Crumb Rubber Content and Type in Changing Critical
Properties of Asphalt Binders”, Journal of the Association of Asphalt Paving Technologists,
Vol. 64, 1995, p. 130-162.
Billiter T.C., Chun J.S., Davison R.R., Glover C.J., Bullin, J.A., “Investigation of the
curing variables of asphalt-rubber binder”, American Chemical Society, Division of Fuel
Chemistry, Vol. 41, No. 4, 1996.
Billiter T.C., Davison R.R., Glover C.J., Bullin J.A., “Production of asphalt-rubber binders by
high-cure conditions”. Transportation Research Record, No. 1586, 1997, p. 50-56.
Ferguson J., Kemblowski Z., Applied Fluid Rheology, Elsevier Applied Science, Elsevier
science publishers Ltd, Essex, New York, 1991.
Gopal V.T., Sebaaly P.E., Epps J., “Effect of Crumb Rubber Particle Size and Content on the
Low Temperature Rheological Properties of Binders”, Transportation Research Board
Annual Meeting, Washington D.C., 13-17 January 2002.
Navarro F.J., Partal P., Martínez-Boza F., Gallegos C., “Influence of crumb rubber
concentration on the rheological behaviour of a crumb rubber modified bitumen”, Energy
& Fuels, No. 19, 2005, p. 1984-90.
Rahman M.M., Characterisation of Dry Process Crumb Rubber Modified Asphalt Mixtures.
Thesis submitted to the University of Nottingham for the degree of Doctor of Philosophy,
Nottingham, 2004.
Singleton T.M., Airey G.D., Collop A.C., “Effect of rubber-bitumen interaction on the
mechanical durability of impact absorbing asphalt”, Proc. 2nd Eurasphalt & Eurobitume
Congress, Vol. 1, Barcelona, 20-22 September 2000.
Tia M., Ruth B.E., “Basic rheology and rheological concepts established by H.E. Schweyer.
Asphalt rheology: Relationship to mixture”, ASTM STP 941, O.E. Briscoe Ed., American
Society for Testing and Materials, Philadelphia, 1987, p.118-145.
van de Ven M., Jenkins K., “Rheological characterization of some (polymer modified)
bitumen and bitumen-filler system at compaction and in service temperatures”, 6th RILEM
Symposium on Performance Testing and Evaluation of Bituminous Materials, Zurich,
14-16 April 2003.
Zanzotto L., Kennepohl G.J., “Development of rubber and asphalt binders by

depolymerization and devulcanization of scrap tires in asphalt”, Transportation Research
Record, No. 1530, 1996, p. 51-58.
Rheological Changes in Bitumen Caused by

Aging and by the Interaction with Rubber
Joana Peralta* ― Hugo M.R.D. Silva* ― Ana V.A. Machado** ―

Jorge Pais*
* Department of Civil Engineering

joana@civil.uminho.pt
hugo@civil.uminho.pt
jpais@civil.uminho.pt
** IPC-Institute of Polymers and Composites

Department of Polymer Engineering
avm@dep.uminho.pt
ABSTRACT. The use of bitumen modified with crumb rubber contributes for a sustainable
development of road infrastructures. However, the increasing demands on quality and
durability of pavements lead to the need of a profound knowledge in the physicochemical
changes of the materials which constitute these AR binders. In this work several base bitumens
interacted with crumb rubber, in order to produce AR binders, which were subsequently
separated, by using a modified “Basket drainage method” to recover the residual bitumen.
The effect of aging during the laying procedures was also studied by using the RTFOT test.
Additionally, a new “Sphere AR production simulator” method was developed to simulate
the bitumen aging without the contribution of the rubber. The DSR was used to evaluate the
chemical and rheological changes in the properties of AR binders. In this work, reduced
changes were observed in the binder due to the interaction with crumb rubber.
KEYWORDS: Rheological changes, DSR, Molecular weight, Molecular weight distribution,

“AR sphere simulator”, “Basket drainage method”.
1. Introduction
The use of bitumen modified with crumb rubber contributes for a sustainable development
of road infrastructures through: (i) the definition of an efficient final destination for the used
tires of cars and trucks; (ii) the improved performance of the resulting material used in the
pavement.
However, the materials that constitute these asphalt rubber (AR) binders and the
physicochemical changes in the binder during the production of the AR are not sufficiently
characterized. Thus, the main objective of this study is to characterize the chemical changes
in the constitution of the AR binder components (bitumen and rubber) and to evaluate their
interaction at a molecular scale, which will be related to the functional properties of the
binders.
Initially, the functional properties of a variety of bitumen types from the same source and
obtained in the same distillation column were assessed using conventional EN standard tests,
and their rheological behaviour was evaluated with a Dynamic Shear Rheometer (DSR).
After being characterized, the bitumen was left to interact with 17.5% of crumb rubber by
mass of AR binder. Then, a modified “Basket Drainage Method” was used to separate the
residual bitumen and the rubber constituting the AR binder. The aging of AR mixtures during
the mixing and laying down procedures were also studied by using the RTFOT method and
a new “AR sphere simulator”, which was developed to reproduce the aging of the bitumen
during AR production without the contribution of the crumb rubber interaction. Finally, the
changes in the chemical composition of the base bitumen, the AR binder and the residual
bitumen, before and after aging, were evaluated by replicating the same test used for bitumen
characterization.
The main results of this work are: (i) the identification of the chemical changes in the
bitumen as a consequence of the bitumen-rubber interaction, and (ii) a better understanding of
the aging influence in the AR binder and mixture production phases.
2. Literature review on bitumen-rubber interaction and aging – chemistry and rheology
The construction and maintenance of asphalt-surfaced roads has a long term and significant
impact on the economic vitality of a nation. The primary reasons for the deteriorating
conditions of roads include increases in overall traffic, poor asphalt binder quality coming
from high-tech refining processes, and changes in climate. Better construction processes are
needed to face these challenges. It is necessary to understand the fundamental behaviour and
properties of the roads before beginning to develop advanced construction processes (Glover,
2007).
Presently, the pavement technology and the evaluation of bituminous mixtures are essentially
based on empirical-mechanistic studies. Frequently, the materials that compose asphalt pavements
are not sufficiently characterized, more specifically concerning their physicochemical constitution,
their rheological properties and their mutual interaction at a microscopic scale. Despite of bitumen
use since biblical times, knowledge of the structure that controls its properties and its viscoelastic
nature in particular, remains incomplete (Masson et al., 2006).
Bitumen is a single-phase homogeneous mixture of many different molecules, which

may be differentiated into two broad classes: polar and non-polar. The non-polar molecules
serve as a matrix or solvent for the polar molecules, which form weak “networks” of polar-
polar associations that give bitumen its elastic properties. The polar materials are uniformly
distributed throughout the bitumen, and upon heating the weak interactions are broken to yield
a Newtonian fluid. When perturbed, in response to temperature changes and physical stresses,
these interactions break and reform to produce a new combination of interactions (Figure 1).
Figure 1. The Bitumen Model (Jones et al., 1992)
The polar molecules interact and primarily give bitumen its elastic characteristics. The
non‑polar molecules primarily contribute to the viscous behaviour of the bitumen and control
the low temperature properties of the bitumen (Youtcheff et al., 1994). Pfeiffer et al. (1940)
observed that asphaltenes, considered as the core constituents of dispersed phases, have a
marked tendency to absorb aromatic hydrocarbon solvents, approaching bitumen to a colloidal
system.
During its life cycle, bitumen hardens (aging) due to chemical reactions, including
oxidation, polymerization and condensation and/or due to physical processes, including loss
of volatiles and structural morphological changes. However, oxidative hardening has been
shown to be the principal factor responsible for a more brittle structure and an increase in
susceptibility to cracking. Oxidative hardening happens at a relatively slow rate and varies
seasonally, occurring faster at higher temperatures as diffusion of oxygen increases (Glover,
2007).
The main ageing mechanism is an irreversible one, characterised by chemical changes

of the binder, which in turn has an impact on the rheological properties. The processes
contributing to this type of ageing include oxidation (Branthaver et al., 1993), loss of volatile
components (Traxler, 1961) and exudation or migration of oily components from the bitumen
into the aggregate or modifier (Curtis et al., 1993). The second mechanism is a reversible
process called physical hardening (Bahia et al., 1993) and may be attributed to molecular
structuring, i.e. the reorganisation of bitumen molecules or microstructures to approach
an optimum thermodynamic state under a specific set of conditions (Petersen, 1984). The
particles shown in the ESEM images of Figure 2 (Rozeveld et al., 1997) may consist of
clusters of ten or more individual networks, and illustrate the relative increase in size of the
asphaltenes which occurs during the aging mechanism in the bitumen.
Figure 2. Environmental scanning electron microscopy (ESEM) images of asphaltene fraction

for (a) an AC-5 bitumen with 0.2-0.3 μm and (b) the aged AC-5 bitumen with 0.8-1.2 μm
There are several available methods to investigate the physical properties of modified
asphalts under aging conditions. One of the most used methods is the assessment of the
binder rheology with the DSR (Lima et al., 2006). The rheology of bitumen at a particular
temperature is determined by the constitution (chemical composition) and structure (physical
arrangement) of the molecules. Thus, to understand changes in the bitumen rheology, it is
essential to realize how the structure and constitution of bitumen interact to influence its
rheology (Read et al., 2003).
When crumb rubber is blended at high temperatures with bitumen to produce a modified
binder (i.e. wet process), the two materials interact once bitumen components migrate into the
rubber causing it to swell. Initially, the bitumen-rubber interaction is a non‑chemical reaction,
where the rubber particles are swollen by the absorption of the aromatic oils of bitumen (Bahia
et al., 1995). The impact of the crumb rubber modification improved the aging susceptibility,
decreasing the binder aging ratio (Martinez et al., 2006). However, the base bitumen should
be selected carefully to ensure that the content of light fractions of bitumen is large enough
at the end of the curing process (Ould-Henia et al., 2006). Generally, a higher rate of mixing
and curing temperature worsened the aging index, while higher rubber contents and asphalt
cements with lower molecular weight improved the aging characteristics of the binder (Leite
et al., 1999).
3. Asphalt rubber production, separation and characterisation
3.1. Bitumen selection, sample preparation and crumb rubber selection
3.1.1. Bitumen samples
Special attention was taken during the selection of the bitumens to be used in this work in
order to control the variables to be studied. Therefore, all the bitumen samples were obtained
from the same batch of bitumen production on the refinery (the same distillation column),
since bitumen characteristics are different depending on the petroleum sources and their
processing.
Asphalt rubber properties are very dependent on the used bitumen. Thus, this work was
begun by choosing several grade bitumens in order to evaluate their influence on the AR
characteristics. The commercial line bitumens selected to carry out this study were pen grades
10/20, 40/50, 60/70 and 150/200.
The laboratorial characterisation of sixteen base bitumen samples was carried out in order
to evaluate the following properties:
- Functional properties, namely the penetration at 25 ºC (EN 1426), ring and ball
softening point (EN 1427), the dynamic viscosity (EN 13302) and the elastic recovery
(ASTM 5329);
- Rheological properties, using the Dynamic Shear Rheometer (only for eight of the
studied bitumens) to determine the viscosity, the modulus and the phase angle;
The commercial bitumens were also subjected to other procedures in order to assess their
changes during the AR binder production, and in the mixing and laying down processes,
namely:
- The standard aging test rotating thin film oven test or RTFOT (EN 12607-1);
- The “Sphere AR production simulator” method developed in this work.
After these procedures, the functional and rheological properties of the obtained bitumen
were assessed, and the changes in the binder characteristics were measured.
3.1.2. Crumb rubber sample
The crumb rubber used in this work was produced by the cryogenic process and, according
to the supplier, was obtained by the cryogenic grinding of 30% of truck tires and 70% of car
tires. The supplied crumb rubber was sieved in order to select and use only the fraction passed
in sieve ASTM #20 (0.850 mm) and retained in sieve ASTM #40 (0.425 mm).
3.2. Production and separation of the asphalt-rubber binder
The method used to produce AR and collect the samples of AR, of recovered swelled
rubber and of residual bitumen was the “basket drainage method” (Rahman, 2004). The
AR production facility is constituted by several equipments (Figure 3) assembled in a lab
ventilation chamber.
Figure 3. Scheme and photo of the asphalt rubber production facility

The wire basket used in the AR production facility was manufactured with a wire mesh,
which was previously observed in an optical microscope to measure its average opening
dimension (# 0,470 mm). The microscope was able to draw an idealized picture of the mesh in
order to assess, automatically, the referred opening dimension.
The process used to produce and collect the several samples of AR, of recovered swelled
rubber and of residual bitumen was the following one:
- Heating of 1 kg of each base bitumen (Bb) at 180 ºC and collecting of a sample;
- Introduction of 17.5% of crumb rubber by mass of asphalt rubber;
- Continuous heating of the asphalt rubber binder at 180 ºC for 60 minutes (digestion
time) while stirring the binder at a velocity of 230 rot/min (Figure 4);
- Collection of a sample of the produced asphalt rubber (AR);
- Suspension of the wire basket with the AR binder in an oven at 180 ºC for 15 min, so
as to separate its constituents and collection of a sample of the residual (Res) bitumen
(Figure 4);
Figure 4. Separation between the crumb rubber and the residual bitumen of the AR
3.3. Reproduction of the aging during the AR production using the “Sphere AR production
simulator” method
When the bitumen is heated, smaller molecules are volatilized (aromatics), while other
molecules can associate themselves with bigger molecules (asphaltenes). Furthermore, when
bitumen interacts with rubber, the rubber particles absorb the light parts of bitumen and swell.
Since those phenomena simultaneously occur during the AR production, it seems unfeasible to
evaluate the changes in the bitumen structure caused by every previous phenomenon. Hence, a
new method was developed during this research in order to determine the ratio of the changes
exclusively induced by heating and stirring the bitumen during the AR production, entitled
“Sphere AR production simulator”.
The aim of this method is to simulate the AR production, by replacing the rubber particles
with metallic spheres (roughly with the same dimensions of the crumb rubber particles) in
order to reproduce the evaporation and association processes occurring in the bitumen during
the heating period for AR production, independently from the interaction with rubber. Before
the experimental procedure, the metallic spheres were observed and measured (øeq ≈ 0.825
mm) using an optical microscope (Figure 5) and their volumetric weight was evaluated (7.75
kg/dm3).
Figure 5. Microscopic photo of the metallic spheres used to simulate the crumb rubber
particles
The “Sphere AR production simulator” was carried out in a reduced scale AR production
facility (similar to the one used for AR production), and consisted in the following procedures:
- Heating of approximately 100 g of each commercial bitumen at 180 ºC;
- Introduction of approximately 240 g of metallic spheres (these spheres have a volume

equivalent to 17,5% of crumb rubber by mass of asphalt rubber);
- Continuous heating of the binder (bitumen and metallic spheres) at 180 ºC for 60
minutes (digestion time) while stirring the binder at a velocity of 230 rot/min;
- Suspension of the wire basket with the binder (bitumen and metallic spheres) in an
oven at 180 ºC for 15 min, so as to separate its constituents;
- Collection of a sample of the resulting bitumen (Sb) for evaluation of its functional
properties and future comparison with the base (Bb) and residual (Res) bitumen
properties.
3.4. Analytical determinations
During this work a series of analytical determinations were carried out in several samples
of binders obtained during the AR production, namely by evaluating the penetration at 25
ºC (Pen), the ring and ball softening point (R&B), the dynamic viscosity (DV), the elastic
recovery (ER) and the rheological properties using the Dynamic Shear Rheometer (DSR).
The laboratorial research carried out in this work was performed in base bitumens (Bb),
asphalt rubbers (AR) and residual bitumens (Res), before and after being aged using the
RTFOT (Bb, AR, Res) procedure, and with the “Sphere simulator” recovered bitumen (Sb).
3.5. Determination of the linear viscoelasticity range of stresses to be used in the rheological
characterization of the binders
A DSR was used to assess the rheological properties of the base bitumens, ARs and
residual bitumens. The rheological tests were carried out at different temperatures (25, 35,
50, 80, 110, 140, 160 and 180 ºC) in order to determine the rheological variation of the binder
properties at all in service, laying down and mixing stages. For each temperature, frequency
sweep tests between 0.1 and 10 Hz were performed in the linear viscoelasticity range, using
parallel plate geometry with 40 mm of diameter and 1.0 mm of gap between plates.
Initially, it was necessary to establish the input stress at each DSR test temperature.
Therefore, previous stress sweep tests were performed at 1.0 Hz to determine the linear
viscoelasticity region (at each test temperature). The linear viscoelasticity range is the range of
stresses where no change is observed in the values of the rheological characteristics measured
at each test temperature. The selection of the stress values to apply on the rheological tests at
each temperature (1000 Pa at 25 and 35 ºC, 200 Pa at 50 ºC, 40 Pa at 80 ºC, 30 Pa at 110 ºC,
20 Pa at 140 ºC, 10 Pa at 160 ºC and 3 Pa at 180 ºC) was made to guarantee that the frequency
sweep tests will be carried out in the linear viscoelasticity region for all tested binders, thus
being possible to compare their rheological behaviour.
Rheological time sweep tests were also performed, at a constant temperature, stress and
frequency, to confirm that no structural modifications occurred in the binders throughout the
tests (60 minutes were used to evaluate the changes during AR production). The rheological
characteristics assessed in the DSR were the complex modulus (G*), storage or elastic
modulus (G’), loss or viscous modulus (G’’), complex viscosity (η*) and phase angle (δ).
4.1. Functional characterization of the studied bituminous materials
To understand the changes induced by the aging process in the base bitumens and in
the AR binders, these materials were tested with the RTFOT method. The variations of the
different parameters measured before and after the RTFOT are presented in Figure 6.
DV Pen R&B ER ΔBase bitumen ΔAR

A B C KD EJ FI GP HO N M
100
80
60
Δ (%)
40
20
0
0 20 40 60 80 100 120 140
Pen at 25 ºC of base bitumen (1/10 mm)
Figure 6. Comparative plot of the functional properties of the base bitumen and AR binder
before and after the RTFOT procedure
The viscosity was the main property altered during the RTFOT aging test of the base
bitumens and AR binders. Furthermore, all the functional properties showed superior
variations for the softer base bitumen 150/220 (M), suggesting that the aging process produces
more alterations on the softer base bitumens than in the harder ones. These results confirm that
during the aging of bitumens, the smaller molecules that confer them softness are volatilized
or associate in bigger molecules that also are oxidized, and as result bitumens become stiffer.
The characterization of the AR binder with the dynamic viscosity test (EN 13302)
presented some dispersion of the results, mainly due to the heterogeneity of the AR binder
and the different density of its components. The AR binder property that was less affected by
the RTFOT aging was the elastic recovery, being possible to observe an improvement in this
characteristic after the aging process. Moreover, it was noticed that the elastic recovery of the
AR produced with the harder and the softer bitumens are almost constant after the RTFOT
aging.
The changes of the functional properties after the RTFOT aging are smaller in the AR
binder than in the base bitumen, except for the dynamic viscosity. These observations suggest
that the AR binder is less sensitive to aging, during the mixing and compaction phases, than
the base bitumen. However, during the AR production the base bitumen is already “aged”
and the RTFOT favours the continuation of the rubber swelling. Since the residual bitumen
surrounding the rubber particles has characteristics different from the base bitumen, with the
present data it is difficult to conclude if the rubber protects the AR binder from aging.
The penetration and ring and ball tests are carried out at temperatures below 100 ºC, at
which occurs a relaxation process in the bitumen. Thus, after RTFOT aging the referred
properties are mainly affected by the bitumen behaviour. In contrast, the dynamic viscosity
values are obtained for high temperatures, at which the bitumen behaves as liquid oil. In
these conditions, the changes in the behaviour of AR binders after RTFOT aging are mainly
controlled by the rubber.
The functional properties of several bitumen samples were assessed in different stages of
the AR binder production process, as well as after the application of the developed “Sphere
AR production simulator” method. The differences between the functional properties of the
processed and base bitumens are plotted in Figure 7.
The graphic on the left side of Figure 7 presents the percentage differences in the
functional properties caused in the base bitumens by the “Sphere AR production simulator”
method, the AR production process (residual bitumen) and by the RTFOT aging procedure.
The process that caused fewer changes in the functional properties of the base bitumens was
the “AR Sphere simulator” method, being the most influential the AR production. In order to
evaluate the changes caused exclusively by the rubber in the base bitumen, it was necessary to
deduct the changes strictly caused by the solubilisation, association and oxidation phenomena
that occur during the AR production, which are simulated by the “Sphere AR production
simulator” method. Thus, the shadow areas of Figure 7 show the percentage alterations caused
by the rubber, which have a considerable impact in the final residual bitumen.
Furthermore, to evaluate the aging process in the AR mixtures during the mixing and
laying down phases, in comparison to the conventional mixtures, the changes in the base
bitumen should be compared with the changes in the residual bitumen (the shadow areas in
the graphic on the left side of Figure 7), since this is the effective binder of the AR mixtures. It
was observed that the aging influence on the AR residual bitumen is visibly minor than on the
base bitumen.
100 DV Pen R&B DV Pen R&B

ΔSphere ΔRes ΔBitumen_RTFOT ΔRes_RTFOT ΔRes Δ Bitumen_RTFOT
100
A LB C KD EJ FI GP HO N M A LB C KD EJ FI GP HO N M
80
80
60
60
Δ (%)
Δ (%)
40
40
20 20
0 0
0 20 40 60 80 100 120 140 0 20 40 60 80 100 120 140
A B C KD EJ F I GP M
1,0
0,8
Weight Loss (%)
0,6
0,4
0,2 Base bitumen

AR
Res
0,0
0 20 40 60 80 100 120 140
Pen at 25 ºC of base bitumen (1/10 mm)
Figure 7. Comparative plot of the functional properties and weight loss for the “sphere”,
base and residual bitumens before and after RTFOT aging
One of the results obtained in the RTFOT aging method is the percentage of bitumen
weight loss (graph below in Figure 7). The variation of this parameter with the penetration
of the base bitumen is a linear trending line with positive slope. The mass loss increased
for softer bitumens, being the base bitumens the most influenced by aging, followed by the
residual bitumens and finally by the AR binders. It was concluded that the rubber protects the
AR binder from aging changes, namely because the AR is less affected by RTFOT aging than
the residual bitumen.
4.2. Rheological characterization of the studied binders
Initially, based on the rheological characteristics of the base bitumen I at 1.0 Hz, Figure
8 presents the changes caused in the bitumen during the AR production and by the RTFOT
aging.
Below 100 ºC, and except for tgδ, all rheological parameters have higher values after
the RTFOT aging due to the increase of the binder stiffness. For this range of temperatures
the elastic modulus of the bitumen increases more rapidly after aging, showing that the
asphaltenes are responsible for the elastic characteristics of the bitumen. Above 100 ºC, after
the relaxation of the asphaltenes, the rheological properties of the base and aged bitumens are
similar. The aged AR binder presented an unusual evolution of G’’ and G*/sen(δ), mainly for
temperatures above 80 ºC. The referred behaviour is caused by the great increase of the elastic
modulus of this material, which is even higher than the viscous modulus. This phenomenon
is not usual for the other studied materials at temperatures above 80 ºC, indicating that the
aged AR binder is clearly the stiffer and most elastic material evaluated in this work. At these
temperatures the rheology of the AR binder is largely controlled by the behaviour of the
crumb rubber particles.
G' (Pa) G'' (Pa) 1,0E+06 Base RTFOT

tg δ η* (Pa.s) G' (Pa) G'' (Pa)
1,0E+06 G*/ sen(δ) (Pa) 1,0E+05 tg δ η* (Pa.s)
Measured values
1,0E+04
Measured values
1,0E+04
1,0E+03
1,0E+02 1,0E+02
1,0E+01
1,0E+00
1,0E+00
Base Res AR
1,0E-02 1,0E-01
20 40 60 80 100 120 140 160 180 20 40 60 80 100 120 140 160 180
T(ºC) T(ºC)
1,0E+06 AR AR_RTFOT G' (Pa) G'' (Pa)

G' (Pa) G'' (Pa) Viscosity* (Pa.s) G*/ sen(α) (Pa)
1,0E+05 η* (Pa.s) G*/ sen(δ) (Pa) 1,0E+06 tg δ
tg δ
Measured values
Measured values
1,0E+04
1,0E+04
1,0E+03
1,0E+02 1,0E+02
1,0E+01
1,0E+00
1,0E+00
...... Res _____ Res_RTFOT
1,0E-01 1,0E-02
20 40 60 80 100 120 140 160 180 20 40 60 80 100 120 140 160 180
T(ºC) T(ºC)
Figure 8. Comparative plot of the rheological properties at 1.0 Hz for the base bitumen I and
the corresponding AR binder, residual bitumen and RTFOT resulting materials
However, for temperatures inferior to 80 ºC the rheological properties of the aged AR

binder are only slightly higher than for the initial AR binder. At these temperatures the
rheology of the AR binder is dominated by its bitumen component.
For temperatures above 110 ºC the G’ values of the aged AR binder increase fairly with the
temperature, while the G’’ values present a moderate decrease, originating very small values
of tgδ and the increase of the complex viscosity values. Since the bitumen is totally liquefied
at these temperatures, the referred only occurs because the rubber particles create an attrition
force over the plates opposing their movement.
In order to assess the effect of RTFOT aging in the four commercial bitumens, rheological
tests were performed in these bitumens after RTFOT aging and the results were compared
with the corresponding base bitumens. Figure 9 presents the isochronal curves obtained for
the different rheological parameters and the differentials between each base and aged bitumen.
The most significant rheological variations between the base and aged bitumen occurred
for the softer bitumen M (150/220). In opposition, the harder bitumen A showed the inverse
tendency, especially at lower temperatures. The rheological parameter that was more
influenced by the RTFOT aging was the storage modulus, especially below 100 ºC. At these
temperatures the variations of G’ were superior to 80 % for all studied bitumens. This change
is also reflected on the change of tgδ, which also presents significant variations (around 60 %)
after aging.
The shape of the curves representing the rheology of the studied bitumens is nearly the
same after aging, but is moved up and to the right (due to the higher viscosity after aging).
1,0E+06 Ab Ab_RTFOT 100 ΔA ΔE ΔI ΔM

Eb Eb_RTFOT
Ib Ib_RTFOT
Mb Mb_RTFOT 80
1,0E+04
Δ G' (%) 60
G' (Pa)
1,0E+02
40
1,0E+00
20
1,0E-02 0
20 40 60 80 100 120 140 160 180 20 40 60 80 100 120 140 160 180
T(ºC) T(ºC)
100
1,0E+05
80
Δ G' (%)
60
G'' (Pa)
1,0E+03
40
1,0E+01
20
1,0E-01 0
20 40 60 80 100 120 140 160 180 20 40 60 80 100 120 140 160 180
T(ºC) T(ºC)
1,0E+06 100
80
1,0E+04
η *(Pa.s)
60
Δη *(%)
1,0E+02
40
1,0E+00
20
1,0E-02 0
20 40 60 80 100 120 140 160 180 20 40 60 80 100 120 140 160 180
T(ºC) T(ºC)
1,0E+03
100
1,0E+02 80
Δ (tgδ) (%)
60
tg δ
1,0E+01
40
1,0E+00
20
1,0E-01 0
20 40 60 80 100 120 140 160 180 20 40 60 80 100 120 140 160 180
T(ºC) T(ºC)
Figure 9. Comparison between the rheological properties of the base and aged bitumens
The temperatures at which the maximum values of tgδ occurred after aging have also
moved to the right (higher temperatures), for higher values of tgδ, reflecting the presence of
bigger molecules in a superior concentration after the aging of the base bitumens.
In order to assess the effect of the RTFOT aging in the initial AR binder produced with the
four commercial bitumens, rheological tests were performed in these AR binders after RTFOT
aging and the results where compared with the corresponding initial AR binders. Figure
10 presents the isochronal curves obtained for the different rheological parameters and the
differentials between each initial and aged AR binder.
1,0E+06 AR AR_RTFOT
A E
100
I M
1,0E+04 80
60
G' (Pa)
ΔG' (%)
1,0E+02
40
1,0E+00
20
ΔA ΔE ΔI ΔM
1,0E-02 0
20 40 60 80 100 120 140 160 180 20 40 60 80 100 120 140 160 180
T(ºC) T(ºC)
1,0E+06 100
80
1,0E+04
60
G'' (Pa)
ΔG'' (%)
40
1,0E+02
20
1,0E+00 0
20 40 60 80 100 120 140 160 180 20 40 60 80 100 120 140 160 180
T(ºC) T(ºC)
1,0E+06
100
80
1,0E+04
60
Δη*(% )
η*(Pa.s)
40
1,0E+02
20
1,0E+00 0
20 40 60 80 100 120 140 160 180 20 40 60 80 100 120 140 160 180
T(ºC) T(ºC)
1,0E+03
100
1,0E+02 80
Δtg δ (%)
1,0E+01
tg(δ)
60
1,0E+00 40
1,0E-01 20
0
1,0E-02
20 40 60 80 100 120 140 160 180
20 40 60 80 100 120 140 160 180
T(ºC)
T(ºC)
Figure 10. Comparison between the rheological properties of the initial and aged AR binders
The evaluation of the changes in the rheology of the aged AR binder should be divided in
two phases, above and below 100 ºC. From 25 to 80 ºC there are few changes in the behaviour
of the AR binder due to aging, since the shape of the curves describing the rheological
behaviour of the initial AR and the aged AR binders is similar (even if some stiffening was
noticed). At these temperatures the behaviour of the AR is controlled by the hardened bitumen
that encloses the rubber particles in its semi-rigid structure, restraining their movements.
Moreover, the behaviour of the swelled rubber particles is influenced by the bitumen within its
reticulated structure.
Between 80 and 110 ºC, an alteration is noticed in the aged AR binder behaviour, which
coincides with the beginning of the relaxation process of the bitumen structure. After this
temperature, it is possible to observe a clearly different behaviour affected by the rubber.
During the aging of the AR binder using the RTFOT several processes took place, such as:
- The crumb rubber particles continue to swell since during the AR binder production
they were not totally saturated and thus, the volume fraction of the rubber particles
after aging of the AR binder is clearly superior to its value in the initial AR binder;
- The residual bitumen surrounding the rubber particles is subjected to a more intense
aging than the base bitumen, since it is exposed to the same test conditions in a thinner
layer;
- In the presence of sulphur, vulcanization and polymerisation processes may take place
in and within the rubber particles during the aging test, as well as during the AR binder
production;
- The rubber particles that have swollen into the saturation point start to depolymerise
and break by the action of temperature and pressure of the bitumen molecules held in
its reticules;
- At the end of the RTFOT aging, it is almost impossible to drain the AR binder from the
containers, since bitumen is entrapped between the rubber swelled particles.
The referred processes mainly occur due to the chemical characteristics of the rubber.
In fact, during the AR binder production and later during the AR mixing and laying down
phases, the rubber particles are exposed approximately to the same conditions of tire rubber
production. In the presence of vulcanization and polymerization inducers, which were not
totally consumed in the primary reaction, the broke and un‑vulcanized rubber chains can be
polymerized/vulcanized.
The results for temperatures superior to 100 ºC confirm the previous analysis. The presence
of the bitumen in the aged AR binder is imperceptible, because at these temperatures the
bitumen is softer than the swelled rubber. Furthermore, the interaction between the rubber
and the bitumen and the reactions among the swelled rubber particles, initiated during the AR
binder production and continued in the RTFOT aging, resulted in a material controlled by the
rheological behaviour of the rubber, with values of G’ visibly superior to the ones of G’’.
Evaluating the graphs of G’ and η* between 140 and 180 ºC, it is perceptible an increase
of the rheological parameters with the temperature, mainly due to the reduction of viscosity of
the residual bitumen (allowing the rubber particles to control the rheology or the AR binder)
and to the dilatation of the links in the rubber structure and increased amplitude of vibration of
the atoms induced by the temperature.
Some differences in the rheological behaviour of the aged AR binder were noticed,
considering the type of base bitumen used to produce the AR binder, namely:
- The AR produced with the softer bitumen (M) showed the highest variation between
the rheology of the initial and aged AR binder;
- The smallest variation between the aged and initial AR binder was verified for the
AR binder produced with the harder bitumen (A), except for the elastic modulus (G’)
property.
At temperatures below 80 ºC, the elastic modulus was the AR rheological property with
greater variation after the RTFOT aging. The rheological isochrones of the four aged AR
binders present the same shape and little divergence. The highest values for the rheological
parameters (except for tgδ) were observed for the aged AR produced with bitumen E, while
the lowest values were obtained for the aged AR with bitumen M. The greatest increase in the
variation of the rheological parameters due to the RTFOT aging of AR binders occurs between
50 and 110 ºC. Above 110 ºC, the variation of the rheological properties upholds a value near
100 %.
During the AR mixing and lay down processes two phenomena occur simultaneously in the
bitumen, the aging of the residual binder and the extension of the diffusion of parts of bitumen
into the rubber. Therefore, to assess exclusively the effect of aging, RTFOT tests were carried
out in the residual bitumen obtained from the initial AR binder by the basket drainage method.
Figure 11 presents the rheological properties of the residual bitumen before and after RTFOT
aging, as well as the percentage differences between them.
1,0E+06 A E ΔA ΔE
100
I M
Res Res_RTFOT ΔI ΔM
80
1,0E+04
ΔG' (%)
60
G' (Pa)
1,0E+02
40
1,0E+00
20
1,0E-02 0
20 40 60 80 100 120 140 160 180 20 40 60 80 100 120 140 160 180
Freq. = 1,0 Hz T(ºC) T(ºC)
1,0E+06 100
80
1,0E+04
ΔG'' (%)
60
G'' (Pa)
40
1,0E+02
20
1,0E+00 0
20 40 60 80 100 120 140 160 180 20 40 60 80 100 120 140 160 180
T(ºC) T(ºC)
1,0E+07 100
1,0E+05 80
60
Δη*(% )
η* (Pa.s)
1,0E+03
40
1,0E+01
20
ΔA
1,0E-01 0 Δ
20 40 60 80 100 120 140 160 180 20 40 60 80 100 120 140 160 180
T(ºC) T(ºC)
1,0E+03
100
1,0E+02 80
80
1,0E+04
ΔG'' (%)
60
G'' (Pa)
40
1,0E+02
20
5941,0E+00Asphalt-Rubber 2009 0
20 40 60 80 100 120 140 160 180
20 40 60 80 100 120 140 160 180
T(ºC) T(ºC)
1,0E+07 100
1,0E+05 80
60
Δη*(% )
η* (Pa.s)
1,0E+03
40
1,0E+01
20
ΔA
1,0E-01 0 Δ
20 40 60 80 100 120 140 160 180 20 40 60 80 100 120 140 160 180
T(ºC) T(ºC)
1,0E+03
100
1,0E+02 80
Δtg δ (%)
60
tg δ
1,0E+01
40
1,0E+00 A E 20
I M
Res Res_RTFOT ΔA
1,0E-01 0 Δ
20 40 60 80 100 120 140 160 180 20 40 60 80 100 120 140 160 180
T(ºC) T(ºC)
Figure 11. Comparison between the rheological properties of the residual bitumens before
and after RTFOT aging
According to Figure 11, the rheology of the residual bitumens before and after RTFOT
aging is similar to the one of the base and aged bitumens (presented previously), even if the
rheological properties of the residual bitumen have presented slightly higher values, except for
tgδ.
Finally, it was observed that the rheological properties of the aged residual bitumens are
higher than the aged AR binders for temperatures inferior to 80 ºC (when the behaviour is
controlled by the bitumen), thus being the residual bitumens stiffer than the corresponding
AR binders before and after aging. In addition, at the referred temperatures (< 80 ºC) the
variations of the rheological parameters of the residual bitumens before and after RTFOT
aging are greater than the observed for the AR binders, confirming the previous observations
that the presence of crumb rubber particles in the AR binder reduces the effects of aging in the
bitumen.
4.3. Changes of the bitumen during the AR production and caused by aging
The rheological study of the materials is a useful tool to assess their structure. During
the AR binder production, AR mixing and lay down phases the bitumen suffer chemical and
structural changes that can be monitored using the DSR. In fact, the rheological properties will
be plotted differently in this section to better illustrate the chemical changes that occur in the
binders. Initially, Figure 12 presents the isochrones and isotherms of tgδ for the base and aged
bitumen I.
1,00E-01 1,34E-01 1,78E-01 2,38E-01 3,17E-01 1,0E+03 25 ºC 35 ºC

4,22E-01 5,63E-01 7,50E-01 1,00E+00 1,33E+00 50 ºC 80 ºC
1,78E+00 2,37E+00 3,16E+00 4,22E+00 5,62E+00 110 ºC 140 ºC
1,0E+03 7,50E+00 1,00E+01 …….... Base ------------ RTFOT 1,0E+02 160 ºC 180 ºC
1,0E+02 1,0E+01
tg δ
1,0E+00
tg δ
1,0E+01
1,0E+00 1,0E-01
1,0E-01 1,0E-02
20 40 60 80 100 120 140 160 180 1,00E-01 1,00E+00 1,00E+01
T(ºC) Frequency (Hz)
Figure 12. Isochrones and isotherms of tgδ of the base and aged bitumen I
In the left graphic of Figure 12 (tgδ vs. temperature) the maximum value of tgδ is obtained
at a higher temperature after aging. This occurs because the proportions between the viscous
(G’’) and the elastic (G’) stiffness of the bitumen, at different temperatures, change due
to the RTFOT aging. Comparing the tgδ isotherms of the base and aged bitumen I (right
side of Figure 13), it is perceptible that the aged bitumen presents lower values of tgδ for
temperatures below 100 ºC, due to a more significant increase of G’ in comparison with G’’
after aging. For temperatures above 100 ºC the values of tgδ are quite similar between the
base and aged bitumen. According to the bitumen model (Jones et al., 1992), the relaxation
of the bitumen shall be inherent to the asphaltenes and resins molecules, since they grow and
come together during the RTFOT aging, thus being responsible for the increase in the elastic
stiffness of the aged bitumen. This can also be observed in the Cole-Cole diagrams of the base
and aged bitumen I (Figure 13).
25 ≤ T (ºC) ≤ 50 Base
Base_RTFOT
2,8E+05
2,1E+05
G'' (Pa)
1,4E+05
7,0E+04
1,0E+00
1,0E-02 2,0E+05 4,0E+05 6,0E+05 8,0E+05
G' (Pa)
Figure 13. Cole-Cole diagram for the base and aged bitumen I
As mentioned before, the Cole-Cole diagrams are a useful tool to identify changes in the
bitumen at medium in service temperatures. The more obvious change in the rheological
characteristics of the aged bitumen occurred at 25-35ºC, with a great increase in G’ values
combined with a decrease in G’’ values. Thus, at this range of temperatures the behaviour of
the aged bitumen is essentially elastic and is controlled by the asphaltenes and resins in the
bitumen.
The evolution of the values of tgδ for the aged AR binder in comparison with the initial
AR binder is presented in Figure 14 in order to assess the structural changes in this binder.
1,00E-01 1,34E-01 1,78E-01 2,38E-01 3,17E-01 1,0E+03 25 ºC 35 ºC 50 ºC 80 ºC

4,22E-01 5,63E-01 7,50E-01 1,00E+00 1,33E+00 110 ºC 140 ºC 160 ºC 180 ºC
1,0E+03 1,78E+00 2,37E+00 3,16E+00 4,22E+00 5,62E+00
7,50E+00 1,00E+01 1,0E+02
1,0E+02
1,0E+01
1,0E+01
tgδ
tgδ
1,0E+00
1,0E+00
1,0E-01
1,0E-01
......... AR ______ ARRTFOT ....... AR _____ ARRTFOT
1,0E-02 1,0E-02
20 40 60 80 100 120 140 160 180 1,00E-01 1,00E+00 1,00E+01
T (ºC) Frequency (Hz)
Figure 14. Isochrones and isotherms of tgδ of the initial and aged AR binder produced with
base bitumen I
In the previous studied materials the isochrones of tgδ present a obvious maximum and the
isotherms present a sudden change in slope, for temperatures at which the bitumen changes its
behaviour from mainly an elastic material to a viscous material (Tc). However, the isochrones
of the aged AR binder merely present a smooth maximum and only a small variation is visible
in the slope of its isotherms after RTFOT aging. The values of tgδ for the aged AR binder for
temperatures inferior to 80 ºC are similar to the initial AR binder and coherent with behaviour
of the base and aged bitumen (because the behaviour is controlled by the bitumen). However,
for temperatures above 80 ºC the values of tgδ for the aged AR binder initially stabilize and
then decrease due to the increase of G’ (because the behaviour is controlled by the rubber).
The relation between the storage and the loss modulus of the aged and initial AR binder, at
in service temperatures, is represented by the Cole-Cole diagram shown in Figure 15.
25 ≤ T (ºC) < 50 AR
2,1E+05 AR RTFOT
1,4E+05
G'' (Pa)
7,0E+04
1,0E+00
1,0E-02 2,0E+05 4,0E+05 6,0E+05
G' (Pa)
Figure 15. Cole-Cole diagram for the initial and aged AR produced with base bitumen I
The initial and aged AR binder have a similar behaviour, even if the line representing the
aged AR binder shows that this is a harder material (higher values of G*). Moreover, the ratio
between the viscous and the elastic stiffness is inferior in the aged AR binder, pointing to a
special increase in the elasticity of the material during the RTFOT aging with impact in this
range of temperatures, and consequently in the performance of the pavement in service.
The isochrones and isotherms of tgδ for the residual bitumen I before and after aging are
presented in Figure 16.
1,00E-01 1,34E-01 1,78E-01 2,38E-01 3,17E-01 1,0E+03 25 ºC 35 ºC

4,22E-01 5,63E-01 7,50E-01 1,00E+00 1,33E+00 50 ºC 80 ºC
1,0E+04 1,78E+00 2,37E+00 3,16E+00 4,22E+00 5,62E+00 110 ºC 140 ºC
7,50E+00 1,00E+01 160 ºC 180 ºC
1,0E+02
1,0E+03
1,0E+02 1,0E+01
tgδ
tgδ
1,0E+01
1,0E+00
1,0E+00
....... Res _____Res_RTFOT
1,0E-01 1,0E-01
20 40 60 80 100 120 140 160 180 1,00E-01 1,00E+00 1,00E+01
T (ºC) Frequency (Hz)
Figure 16. Isochrones and isotherms of tgδ of the residual bitumen (extracted from the AR
binder produced with base bitumen I) before and after aging
The behaviour of the residual bitumens before and after aging is similar to the one of the
base and aged bitumens (presented previously). The upper limit formed by the isochronal
lines dislocate to the right, at higher critical temperatures. This phenomenon is reflected in
the isotherm lines by the temperature at which the change of slope of the lines occur, between
50 and 80 ºC for the residual bitumen before aging and between 80 and 110 ºC after aging.
This change of behaviour is due to the variation of the molecular structure of the residual
bitumen. When the critical temperature is achieved, a relaxation of the molecular structure
of the bitumen occurs, and the weak interactions that maintain the “networks” formed by the
polar molecules of bitumen (asphaltenes and resins) are broken. Thus, the colloidal behaviour
becomes a Newtonian fluid behaviour as the asphaltenes are peptized by the resins being
totally dispersed in the maltenes phase (non-polar).
Figure 17 presents the changes that occur in the viscoelastic character of the residual
bitumen before and after aging at in service temperatures (25 ºC ≤ T < 50 ºC).
3,0E+05 25 ≤ T (ºC) ≤ 50 Res

Res_RTFOT
2,0E+05
G'' (Pa)
1,0E+05
1,0E+00
1,0E-02 2,0E+05 4,0E+05 6,0E+05 8,0E+05
G' (Pa)
Figure 17. Cole-Cole diagram for the residual bitumen (extracted from the AR produced with
base bitumen I) before and after aging
The changes in the viscoelastic characteristics of the residual bitumen induced by the
RTFOT aging indicate a stiffer aged material due to a reduction in G’’ and a slight increase
in G’, caused by the volatilization of some bitumen molecules and the increase of the particle
size of other.
The performance of the studied binders in the pavement can be obtained by analysing the
Cole-Cole diagrams of the binders before and after aging (Figure 18), where the changes of
behaviour can be clearly assessed at the lowest tested temperatures (25 and 35 ºC).
All aged base bitumens showed an increase of G’ and a reduction of G’’ at each frequency
(upper left side graphic of Figure 18). The lowest variation after aging occurs for bitumen A
and the inverse for bitumen M, as expected. The similarity between the form and the evolution
of the different curves of the aged bitumens indicates that the RTFOT is a process that
approaches the distillation process by which the different grade bitumens are produced.
Aging has a different impact over the diverse initial AR binders (upper right side graphic
of Figure 18). The effect of the RTFOT aging over the initial AR binders, between 25 and
50 ºC, caused some changes on the relation between the elastic and viscous stiffness of the
tested materials. The Cole-Cole curves of the aged AR binders produced with bitumen E and
I show a similar form to the Cole-Cole curve of the corresponding initial AR binders, even if
a small translation of the curve towards down and right means that the material gets stiffer.
In fact, the interactions between the rubber and bitumen, initiated in the production of the AR
binder, will continue during the RTFOT aging, since the rubber particles did not achieved the
saturation.
Ab_RTFOT Eb_RTFOT Ib_RTFOT Mb_RTFOT Ab Eb Ib Mb

2,5E+05 ....... AR _____ AR_RTFOT
2,8E+05
2,0E+05
2,1E+05
G'' (Pa)
1,5E+05
G'' (Pa)
1,4E+05
1,0E+05
7,0E+04 5,0E+04
A E I M
1,0E+00 0,0E+00
1,0E-02 2,0E+05 4,0E+05 6,0E+05 8,0E+05 0,0E+00 2,0E+05 4,0E+05 6,0E+05 8,0E+05
G' (Pa) G' (Pa)
2,8E+05 ....... Res _____ Res_RTFOT A E I M
2,1E+05
G'' (Pa)
1,4E+05
7,0E+04
1,0E+00
1,0E-02 2,0E+05 4,0E+05 6,0E+05 8,0E+05
G' (Pa)
Figure 18. Comparative Cole-Cole diagrams for the commercial base bitumens, AR binderss
and residual bitumens before and after aging
The Cole-Cole curves of the AR binder produced with bitumen A, before and after the
RTFOT aging, cross themselves up and no hardening is detected in the aged material at
medium in service temperatures (25 to 50 ºC). This can be explained because during the
production of the AR binder almost all the small molecules contained in the base bitumen
A (10/20) have interacted with the rubber particles. Therefore, in the RTFOT aging test the
bitumen-rubber interaction was very low.
As it would be expected, the properties of the AR binder produced with bitumen M

changed significantly during the aging process. Since the residual bitumen of the AR binder
produced with bitumen M still have a considerable amount of small and volatile molecules,
they will continue to diffuse into the rubber particles during the RTFOT aging until their
saturation (gel rubber). Then, the particles of gel rubber fracture, releasing little pieces that
increase the volume fraction of the gel rubber in the AR binder, but also reduce the stiffness
of the binder at medium in service temperatures (at these temperatures the behaviour is
controlled by the bitumen).
Finally, some dispersion can be observed in the Cole-Cole curves of the residual bitumens
before aging, due to the different effect of the rubber in the studied bitumens. However, the
influence of RTFOT aging in the behaviour of the residual bitumens is consistent with the
previous analyses presented for the base bitumens (the aged residual bitumens are stiffer).
Despite the average values of G’’ for the residual bitumens are superior to the ones of the base
bitumens and inferior to the ones of the AR binders, they always present the higher average
values of G’ (before and after RTFOT aging). Thus, the residual bitumens are the stiffest
materials studied in this work, which shows the great changes in the chemical structure and
composition of the bitumen during the production of the AR binder.
5. Conclusions
The main conclusions that can be drawn from the results of this work are the following:
- The “Sphere AR production simulator” method was found to be a simple and effective
method to determine the changes induced by the heating and stirring of the bitumen
during the AR binder production without considering the interaction with the rubber
particles;
- The assessment of the functional properties indicates that the aging process causes
more alterations on the softer base bitumens than in the harder ones;
- The elastic recovery is the functional property of the AR binder less affected by aging,
and this property improves after the RTFOT aging process;
- The functional properties of the AR binder are less affected by aging than to the ones
of the base bitumen. These observations suggest that the AR binder is less sensitive to
aging than the base bitumen during the mixing and lay down phases;
- The rubber particles continue the swelling process during the AR binder aging, because
they are not totally saturated after the AR binder production;
- The interaction between the rubber and the bitumen and the reactions among the
swelled rubber particles, initiated during the AR binder production and continued in the
RTFOT aging test, result in a material controlled by the rubber rheological behaviour
for temperatures superior to 110 ºC (with G’ values visibly superior to the values of G’’)
and by the bitumen rheological behaviour for temperatures below 110 ºC;
- The rheological properties obtained at temperatures below 80 ºC showed that the

highest variation of the AR binder characteristics after aging was obtained for the AR
binder produced with the softest bitumen M (150/220) showed, while the smallest
variation was verified for the AR binder produced with the hardest bitumen A (10/20);
- At temperatures inferior to 80 ºC, the changes in the rheological properties of the

residual bitumens after RTFOT aging are greater than the ones observed for the AR
binders, indicating that the presence of the rubber particles in the binder reduces the
aging processes;
- There is a relaxation phenomenon in the bitumen structure that occurs at a

characteristic temperature (between 80 and 140 ºC) and is related to the change in the
material behaviour, representing the change between the viscoelastic behaviour of the
bitumen (below that temperature) and the behaviour of a Newtonian fluid (above that
temperature);
Acknowledgements
The authors would like to thank to CEPSA and to RECIPAV for their support and for the
supply of bitumen and crumb rubber, respectively.
6. Bibliography
Bahia H., Davies R., “Role of Crumb Rubber Content and Type in Changing Critical
Properties of Asphalt Binders”, Journal of the Association of Asphalt Paving Technologists,
Vol. 64, 1995, p. 130-162.
Bahia H.U., Anderson D.A., “Glass transition behaviour and physical hardening of asphalt
binders”, Journal of the Association of Asphalt Paving Technologists, Vol. 62, 1993, p.
93-129.
Branthaver J.F., Petersen J.C., Robertson R.E., Duvall J.J., Kim S.S., Harsberger P.M., Mill
T., Ensley E.K., Barbour F.A., Schabron J.F., Binder Characterization and Evaluation –
Volume 2 Chemistry, SHRP A-368, SHRP, National Research Council, Washington DC,
1993.
Curtis C.W., Ensley K., Epps J., Fundamental Properties of Asphalt–Aggregate Interactions
Including Adhesion and Absorption, SHRP-A-341, National Research Council, Washington
DC, 1993.
Glover I.C., Wet and Dry Aging of Polymer-Asphalt Blends: Chemistry and Performance,
PhD dissertation submitted to the graduate faculty of the Louisiana State University and
Agricultural and Mechanical College, Department of Chemistry, 2007.
Jones D.R., Kennedy T.W., The Asphalt Model: Results of the SHRP Asphalt Research
Program, SHRP A‑001, Centre for Transportation Research University of Texas, Austin,
Texas, 1992.
Leite L.F.M., Soares B.G., “Interaction of Asphalt with Ground Tire Rubber”, Petroleum
Science and Technology, 17:9, 1999, p. 1071-1088.
Lima C.S., Tomé L.G.A., Filho C.M.S.O., Soares S.A., Soares J.B., “Aging Studies of
Asphalt-Rubber Binder”, Asphalt Rubber 2006 Conference, Palm Springs, USA, 2006, p.
797-812.
Martinez G., Caicedo B., Celis L., González D., “Rheological Behaviour of Asphalt with
Crumbed Rubber and other Modifiers”, Asphalt Rubber 2006 Conference, Palm Springs,
USA, 2006, pp. 863-884.
Masson J.-F., Leblond V., Margeson J., “Bitumen morphologies by phase-detection atomic
force microscopy”, Journal of Microscopy, Vol. 221, Issue 1, 2006, p. 17-29.
Ould-Henia M., Dumont A.G., “Assessment of the Rheological Properties of Asphalt Rubber
Binder and its Residual Phases”, Asphalt Rubber 2006 Conference, Palm Springs, USA,
2006, p. 931-948.
Petersen J.C., “Chemical Composition of Asphalt as Related to Asphalt Durability: State of

the Art”, Transportation Research Board, No.999, 1984, p.13-30.
Pfeiffer J., Saal R.N.J., “Asphaltic Bitumen as Colloid System”, The Journal of Physical
Chemistry, Vol. 44, No. 2, 1940, p. 139-149.
Rahman M.M., Characterisation of Dry Process Crumb Rubber Modified Asphalt Mixtures.
PhD Thesis submitted to the University of Nottingham for the degree of Doctor of
Philosophy, Nottingham, 2004.
Read J., Whiteoak C.D., The Shell Bitumen Handbook, fifth edition, Thomas Telford
Publishing, London, 2003.
Rozeveld S., Shin E., Bhurke A., France L., Drzal L., “Network morphology of straight and
polymer modified asphalt cements”, Microscopy Research and Technique, No. 38, 1997, p.
529-543.
Traxler R.N., “Relation between asphalt composition and hardening by volatilization and
oxidation”, Proc. Assoc. Asphalt Paving Technol., Vol. 30, 1961, p. 359-377.
Youtcheff J.S., Jones D.R., Guideline for Asphalt Refiners and Suppliers, Strategic Highway
Research Program, National Research Council, Washington DC, 1994.
Study on the Reduction of the Viscosity of

Asphalt Rubber
Pengyun Cao *— Shifeng Wang **— Jinliang Li**
* Panjin Petrochina Liaohe Asphalt Co. Ltd., Panjin Liaoning 124010, China
caopy@petrochinalhlq.sina.net
**Research Institute of Polymer Materials, Shanghai Jiaotong University

Dongchuan Rd 800, Shanghai Jiao Tong University .Shanghai 200240 China
shfwang@sjtu.edu.cn
Jinliang.Lee@sjtu.edu.cn
ABSTRACT: It is profound to reduce the viscosity of asphalt rubber (AR) at high temperature
for easy processing, transportation and application. The effects of several additives at two
typical blending temperatures (175 ℃ , 190 ℃ ) on the viscosity and the softening point of AR
were investigated. The results showed that, the viscosity and the softening point of AR were
obviously decreased after being mixed at 190 ℃ for 45minutes, compared to those mixed at
175 ℃ . Wax with high melting temperature (Tm) decreased the viscosity of AR, but increased
the softening point. Reactive agent could efficiently decrease the viscosity of AR, therefore, it
can be used to decrease the blending temperature.
KEYWORDS: Reduction, Viscosity, Reactive agent, asphalt rubber (AR)

1. Introduction
According to the definition of ASTM D8, asphalt rubber (AR) is one kind of blend
processed at high temperature (≥180 ℃) (Sun, 2007). AR has a series advantages such as low
temperature susceptibility, good elastic recovery and age resistance. The application of AR
could reduce the thickness of the road and extend the life of the road building. In the other
hand, the use of crumb rubber in the road can reduce the cost of asphalt modification and
solve the solid pollution of the used tire, and take effect in environmental protection.
There are a lot of researches and applications about AR, the properties of AR were
dependent highly on the size and the content of crumb rubber, the processing technology,
and asphalt composition (Huang, 2006; Navarro, 2007). All these studies indicated that the
viscosity at high temperature is the key factor influencing the quality of AR (Huang, 2006).
The processing temperature of AR is usually above 180℃, and the viscosity of AR is

quite high, about 10 Pa·S at 135℃. For the mixing and laying of the asphalt mixer, the
viscosity of modified asphalt should be below 3 Pa·S. The high processing temperature and
high viscosity cause a series defects of AR, such as the higher requirement of equipment, the
better age-resistance of the base asphalt, the more exposure of pollution gas and higher energy
consumption. In the meantime, the high viscosity causes serious problems of transporting and
laying. Therefore, the reduction of the viscosity and the processing temperature of AR became
an important issue.
In this paper, several additives were used to reduce the viscosity of AR, and the effect of
these additives on the viscosity at two temperatures was investigated.
2. Experimental
2.1. Materials
Crumb rubber, tread rubber from waste tires, 30 mesh, was provided by Shanghai Xiaoyou
Reclaim Rubber Factory. Asphalt, Liaohe AH-90, was produced by Panjin Liaohe Petrochina
Asphalt Co. Ltd. Additive, trans-octane rubber (TOR) (Vestnamer), was produced by Degussa
Co. Ltd. PE wax, (melting temperature of 105 ℃) was produced by Shanghai Huayi Agent
Company. Reactive agent, activator 420, was provided by Shanghai Xiaoyou Reclaim Rubber
Factory.
2.2. Processing Method
Composition: 20 wt % crumb rubber, 80 wt % asphalt. Additive content is weight content

of the AR. Processing temperature: the base asphalt was heated to the set point of temperature,
and then 20 wt % of crumb rubber was then added in.
Method 1: mixed at 175 ℃ for 45 minutes.
Method 2: mixed at 190 ℃ for 45 minutes.

2.3. Measurements
Conventional properties of softening point and rotational viscosity was tested according
to Chinese Road Engineering Asphalt and Mixture Test Procedures (JTJ052-2000), and
respectively, guideline of T0606-2000 for softening point and T0625-2000 for rotational
viscosity.
3. Results and Discussion
The size of crumb rubber particles increased when it absorbed the maltene from the
asphalt. The volume of solid particles increased and the viscosity of the left asphalt also
increased, which led to the increase of viscosity of AR (Huang, 2006). Therefore, the viscosity
of the AR could be increased by decreasing the viscosity of continuous phase and the friction
between the crumb rubber particles. The solid portion can be reduced by the degradation
of crumb rubber. The degradation part should be controlled properly, or else, the extreme
degradation will cause poor properties of the AR.
PE wax is different with the wax in the asphalt, which has high molecular weight and high
melting temperature. It has been used in the asphalt industry (Edwards, 2007).TOR is trans-
Octane rubber (Sun, 2007), which has a low crystallinity and low melt temperature (60 ℃), it
was used in the rubber industry to reduce the processing viscosity and increase the hardness
of the compound. The reactive agent is a kind of desulphurization agent, it can accelerate
the degradation speed of the crumb rubber and enhance the interaction between asphalt and
crumb rubber. In the following study, the effect of TOR, PE wax and the reactive agent on the
properties of AR were investigated.
3.1. Effect of different additives on the viscosity of AR mixed at 175 ℃
As seen from Table 1, AR has an obvious shear thinning behavior (non-Newton behavior).
Table 1. Effect of different additives on the softening point and viscosity of AR mixed at 175 ℃
additive Content Viscosity/ Pa•S (177 ℃)

Softening point/ ℃
(wt %) 1rpm 5rpm 10rpm 20rpm
0 70.0 55.0 28.0 21.0 11.0
1%TOR 70.5 78.0 29.0 16.0 10.0
1%Wax 71.0 43.8 16.0 11.0 8.0
2%Wax 73.5 44.3 16.0 12.0 9.0
0.5%Reactive agent 65.5 24.0 14.0 9.0 5.0
TOR and PE wax indeed decreased the viscosity of AR. Compared with TOR, PE wax had
a greater effect. With the addition of TOR, the viscosity decreased obviously at high speed. In
comparison with TOR, 1% wt PE wax decreased the viscosity of AR to a greater extent. The
reduction effect changed a little with the increase of PE wax. It can be also seen from Table 1,
0.5 % reactive agent sharply decreased the viscosity of AR, the viscosity reduced to the half of
the original viscosity.
Table 2. Effect of blending time on the softening point and viscosity of AR mixed at175 ℃
Blending Softening Viscosity/ Pa•S (177 ℃)

time/min point/ ℃ 1 rpm 5 rpm 10 rpm 20 rpm
45 70.0 55.0 28.0 21.0 11.0
60 68.5 44.0 18.3 14.0 9.0
120 68.0 42.0 16.8 13.0 8.6
180 66.5 40.0 15.8 12.2 8.3
Table 2 showed the effect of blending time on the softening point and viscosity of AR
mixed at 175 ℃. From Table 2, the viscosity and the softening point decreased slowly,
however, the viscosity was still very high (over 8 Pa•S) at 177 ℃, which did not meet the
application requirements of modified asphalt (Navarro.F.J, 2007). It indicated that extending
blending time at 177 ℃ could not effectively decrease the viscosity of AR.
As seen from Table 3, the effect of reactive additive on the reduction of viscosity was
obvious, and the decreasing trend became obvious with blending time, and then became
smooth after 2 h. The blend satisfied the standard specification after 1～2 h reaction time. The
effect of additive on the softening point of AR indicated that the softening point decreased in
the same trend with the viscosity of AR.
Table 3. Effect of blending time on the softening point and viscosity of AR mixed at 175 ℃ in
the presence of 0.5 wt % reactive agent
Viscosity/ Pa•S (177 ℃)

Time/min Softening time/ ℃
1 rpm 5 rpm 10 rpm 20 rpm
45 65.5 24.0 14.0 9.0 5.0
60 62.5 17.3 7.4 5.4 4.5
120 61.5 11.0 5.7 3.8 2.7
180 60.5 11.0 5.1 3.8 2.4
3.2. Effect of the additive on the reduction of viscosity of AR mixed at 190 ℃
Table 4. Effect of the kind of additive on the softening point and viscosity of AR mixed at190 ℃
Additive Content Softening point/ Viscosity/Pa•S (190 ℃)

(wt %) ℃ 1 rpm 5 rpm 20 rpm 60 rpm 90 rpm
0 60.0 15.0 6.5 3.0 2.6 2.3
1%TOR 60.5 13.5 6.0 3.1 2.7 2.3
1%Wax 60.5 9.9 4.5 2.8 2.0 1.9
2%Wax 61.0 9.8 4.5. 2.7 2.1 1.9.
Table 4 showed the effect of the different kinds of additives on the softening point and
viscosity of AR mixed at 190 ℃. It can be seen that the AR still remained shear thinning
behavior after 45 min mixing at 190 ℃, however, whose viscosity was only one fourth of AR
mixed at 175 ℃. It indicated that the original state of the crumb rubber changed and part of
the rubber was desulfurized or degraded; the percent of solid part became less, which resulted
in the decrease of the viscosity. The softening point decreased about 10 ℃, which indicated
that, the higher temperature processing not only decrease the viscosity but also decreased the
softening point of AR.
TOR and PE wax surely both decreased the viscosity of AR. PE wax had more obvious
effect. The reduction of the viscosity changed a little with the increase of the wax content.
Both TOR and PE wax had the same effect on the softening point, which lightly increased the
softening point of AR.
Compared to the viscosity of AR mixed at 175 ℃ for the same time, it was easy to find
that the AR mixed at 190 ℃ for 45minutes could meet the requirement of mixing and laying.
Compared the line 3 of Table 3 and line 4 of Table 4, it can be found that AR mixed at 175 ℃
for 120 minutes in the presence of reactive agent could be equal to the AR mixed at 190 ℃ for
45 minutes. The desulphurization agent obviously shortened this process, which made it meet
the acceptable viscosity requirement at 175 ℃ in less than an hour.
4. Conclusions
1. Mixing temperature had a great effect on the properties of AR, mixing at higher
temperature (190 ℃) is quite different with that at 175 ℃ , which greatly reduced the
viscosity and the softening point of the AR.
2. The addition of TOR could lightly reduce the viscosity and increased the softening
point of AR.
3. The addition of PE wax decreased the viscosity of AR and kept the softening point in
the same time.
4. Desulfurizaiton agent obviously decreased the viscosity in short time, however, the
softening point decreased. It could decrease the processing temperature and reduce the
blending time of AR.
References:
Edwards Y., Tasdemir Y., Isacsson U. Rheological effects of commercial waxes and poly-
phosphoric acid in bitumen 160/220 - high and medium temperature performance[J],
Construction and Building Materials, 2007, 21（10）: 1899-1908
Huang W., Zhang Y. Mechanical mechanism of asphalt rubber and standard profile [J],
petroleum asphalt, 2006, 20（4）: 61-66
Navarro F. J., Partal P., Martı′nez-Boza F. J, Gallegos C. Influence of processing conditions

on the rheological behavior of crumb tire rubber modified bitumen [J], Journal of Applied
Polymer Science, 2007, 104(3):1683-1691
Sun Z,Chen B. Asphalt rubber Application Manual[M], People’s Transportation Publication

House, 2007, First edition.
Mechanism Research of Crumb Rubber

Modified Asphalt
Zhang Xiaoying *- Xu Chuanjie* - Zhang Yuzhen*
* State Key Laboratory of Heavy Oil, China University of Petroleum

271, Bei’er Road
Dongying, Shangdong Province
P.R. China
zzxxyy1725@126.com
xucj1725@126.com
zhangyuzhen2218@163.com
ABSTRACT. This study is to search the mechanism of crumb rubber modified asphalt and to
supply some theoretical support for utilizing crumb rubber modified asphalt successfully and
reasonably. In the present study, the rule of oil component absorbed by crumb rubber from
catalyzed slurry oil, aromatic hydrocarbon extract oil and vacuum residue with temperature
and time was investigated. It was found that during the action of crumb rubber and medium,
absorbed oil component ratio first increased and then dropped as temperature and time
increased in catalyzed slurry oil and vacuum residue. However, it increased in aromatic
hydrocarbon extract oil under the same conditions. The ratio of absorbed oil component
in vacuum residuum is lower than those in catalyzed slurry oil and aromatic hydrocarbon
extract oil. And crumb rubber released some large molecule component and black carbon into
medium. At the first time the study find the proof of component exchange between crumb and
asphalt but not by presumption.
KEYWORDS:Mechanism Crumber rubber modified asphalt

1. Introduction
With the rapid development of traffic, the amount of wasted tires is increasing more
and more quickly (Yuan 2001). The importance of environmental protection is paid more
attention by people today. How to utilize wasted tires and turn them into a useful material
is a new subject. The wasted tires are used to modify asphalt to improve its performance is
an important means to consume wasted tires. The utilization of the crumb rubber to modify
asphalt in China is prevailing. This is to improve the performance of asphalt as a kind of
pavement materials and to consume some tire to decrease pollution from traffic. Some
experiment roads are being constructed by traffic and research department. Some researchers
began to modify asphalt with crumb rubber in China in 1980s (Billiter 1996; Abdelrahman
1996; Zanzotto et al. 1996; Zhang et al. 2004). The result from experiment roads is
satisfactory. And the results tell us that crumb rubber can improve the performance of asphalt
as a kind of pavement materials. However, the mechanism of modifying asphalt with crumb
rubber is not clear. Several researches have some done work in this field, but direct evidence
was scarcely obtained. They found that crumb rubber swelled at low temperature because of
absorbing oil content from asphalt. However, at higher temperature crumb rubber desulfurized
and depolymerized. Some crossing network was broken and some small molecules entered
into asphalt, which induced asphalt to resolve crumb rubber. These findings are from some
properties of crumb rubber modified asphalt and few data are found to verify the conclusion.
This paper is to find the mechanism of crumb rubber modified asphalt by studying the rule
of the ratio of oil content absorbed by rubber particle from wasted tire in catalyzed slurry oil,
aromatic hydrocarbon extract oil and residue.
2. Experimental
2.1. Materials
The rubber particles used in the study were from wasted tires in ShenYang which were
processed in environment temperature. They have coarse surface and unshaped shape. The
size of the rubber particles is about 10mm. The petroleum contents are catalyzed slurry oil,
aromatic hydrocarbon extract oil and residue. The reason to select the three medium is that
they have different four group and they are the compenents of petroleum which are the same
as asphalt .Their basic properties are listed in Table 1.
Table 1. Basic properties of catalyzed slurry oil, aromatic hydrocarbon extract oil and residue
aromatic hydrocarbon
Item catalysized slurry oil, % Residue, %
extract oil, %
saturates 46.3 12.22 20.85
aromatics 26.04 66.61 35.11
Resins 24.78 20.86 40.94
asphaltenes 2.88 0.31 3.1
Softening point，℃ - - 41.0
penetration，0.1mm - -- 164
From table 1, we find that comparing with catalyzed slurry oil, aromatic hydrocarbon
extract oil and residue, the saturates is highest and aromatics is the lowest, and oil content is
72.34% in catalyzed slurry oil. Comparing with catalyzed slurry oil and residue, the saturates
and asphaltenes are the lowest and the highest aromatics is 66%, and oil content is 78.83%
in aromatic hydrocarbon extract oil, the asphaltenes is the highest and resins is the highest
comparing with catalyzed slurry oil and aromatic hydrocarbon extract oil, and the lowest oil
content is 40.91%t in aromatic hydrocarbon extract oil.
2.2. Experimental procedure and the main analytical item
Weigh the mount rubber particle and put it into a tube, in which there is some petroleum
content, then cover the tube with aluminium foil. Put the covered tuber into a constant
temperature oven for certain period time. Pour the content in the tube and filter the content
and wash the rubber particles to remove the oil content on the surface of rubber particle,
weigh dried rubber particle to calculate the ratio of oil content absorbed by rubber particles, a
oil content left which isn’t absorbed by rubber particles is to evaluate the four group.
In the paper blending exist in the procedure of rubber particle absorbing petroleum content
in residue.
The four components are determined by thin layer chromatography, the chromatogram
column is made of silica gel，the solvents are n-heptane, toluene, tetrahydrofuran. The
method divides petroleum content into four component by their polarity and their diffusion
ability in different solvent. We gain saturates by n-heptane, aromatics by toluene, resins by
tetrahydrofuran, the left is Asphaltenes.
The temperatures were 130℃, 160℃, 190℃, 210℃, 230℃.The constant times were 1h, 3 h,
5 h, 7h in catalyzed slurry oil and aromatic hydrocarbon extract oil. The constant times were
1h, 2 h, 3 h, and 4h in residue. The mass of rubber particle and petroleum content is constant.
The data from the experiment are analyzed as following.

160
the ratio of oil content
140
120
absorbed,%
100 130℃
160℃
80
190℃
60 210℃
40 230℃
20
0
0 1 2 3 4 5 6 7 8
time，hour
Fig.1 Relationship between the ratio of oil content absorbed by rubber particle in catalyzed
slurry oil with temperature and time
160
140
120
absorbed,%
100
1h
80 3h
60 5h
7h
40
20
0
120 130 140 150 160 170 180 190 200 210 220 230 240
temperature，℃
Fig. 2 Relationship between the ratio of oil content absorbed by rubber particle in catalyzed
slurry oil with temperature and time
From figure 1, we find that the ratio of oil content absorbed by rubber particle in catalyzed
slurry oil was increasing with the increase of time at the same temperature, and the time
was shorter than five hours., the ratio of oil content absorbed by rubber particle in catalyzed
slurry oil was almost the highest when the time is 5h and the ratio of oil content absorbed by
rubber particle in catalyzed slurry oil decreased when the time was longer than 5hours. From
figure 2, we find that the ratio of oil content absorbed by rubber particle in catalyzed slurry oil
increased first when the temperature was lower than 210℃ and then decreased with increase
of temperature when the temperature was higher than 210℃ as the time keeps the same in
catalyzed slurry oil.
240
220
200
180
160 130℃
140
absorbed,%
160℃
120
100 190℃
80 210℃
60
40 230℃
20
0
0 1 2 3 4 5 6 7 8
time，h
Fig.3 Relationship between the ratio of oil content absorbed by rubber particle in aromatic
hydrocarbon extract oil between temperature and time
From figure 3 and 4, we find that the ratio of oil content absorbed by rubber particle in
aromatic hydrocarbon extract oil increase with the increasing of time and temperature. The
ratio of oil content absorbed by rubber particle in aromatic hydrocarbon extract oil at all the
temperatures and all the time does not appear. The ratio of oil content absorbed by rubber
particle in aromatic hydrocarbon extract oil is higher than that in catalyzed slurry oil when the
condition is the same.
240
220
200
180
160
140 1h
absorbed,%
120 3h
100
80 5h
60 7h
40
20
0
120 140 160 180 200 220 240
temperature，℃
Fig.4 Relationship between the ratio of oil content absorbed by rubber particle in aromatic
hydrocarbon extract oil between temperature and time
90
80
70
60
absorbed,%
50 160℃
3 40 190℃
30 210℃
20 230℃
10
0
0 1 2 3 4 5
time，h
Fig.5 Relationship between the ratio of oil content absorbed by rubber particle in residue with
temperature and time
90
80
70
60
50 1 h
absorbed,%
40 2 h
3 h
30 4 h
20
10
0
150 160 170 180 190 200 210 220 230 240
temperature,℃
Fig.6 Relationship between the ratio of oil content absorbed by rubber particle in residue with
temperature and time
From figure 5 and 6, we find，in residue the ratio of oil content absorbed by rubber
particle in increase with the increase of time and temperature when time was lower 4h and
the temperature is higher than 210℃. The ratio of oil content absorbed by rubber particle
in residue is far lower than that of in aromatic hydrocarbon extract oil and in catalyzed
slurry oil when the condition is the same. But comparing with ratio of oil content absorbed
by rubber particle in catalyzed slurry oil, the phenomenon of decreasing of the ratio of oil
content absorbed by rubber particle in residue appear in shorter time and lower temperature.
At all temperatures, the ratio of oil content absorbed by rubber particle in residue when time
is longer than 4h is far lower than that of in other temperature. This phenomenon does not
appear in aromatic hydrocarbon extract oil and catalyzed slurry oil. Stirring accelerates the
process of absorbing oil and transferring mass in residue. This makes some components in
rubber particle enter residue and make the per se mass of rubber particle reduce. So the mass
of rubber particle after rubber particle absorbed oil content in residue is lower than that of
aromatic hydrocarbon extract oil and catalyzed slurry oil. But that the mass of rubber particle
after rubber particle absorbed oil content in residue is far lower cannot be explained by
stirring. From the size of rubber particle after rubber particle absorbed oil in residue in 230℃
become small, we can infer that rubber particle take place the reaction of desulfurizing and
polymerizing and the reaction make the per se mass of rubber particle reduce rapidly. From
the four groups of three petroleums we can infer the higher content of resins and asphaltenes
have the some degree function of inducing the reaction of desulfurizing and polymerizing.
This is the reason that the ratio of oil content absorbed by rubber particle in residue and
catalyzed slurry oil appear to decrease and that of aromatic hydrocarbon extract oil appear to
increase all the while.
4. Conclusions
- The ratio of oil content absorbed by rubber particle increase with time and temperature
in aromatic hydrocarbon extract oil.
- The ratio of oil content absorbed by rubber particle increase with time and temperature
in catalysized slurry oil when the temperature is lower than 210℃. The ratio of oil
content absorbed by rubber particle decreased with time and temperature in catalyzed
slurry oil, when the temperature was higher than 210℃.
- The time and temperature of the ratio of oil content absorbed by rubber particle
decreasing with time and temperature is far shorter and far lower than that of in
catalyzed slurry oil.
- From the study, we infer that the higher content of resins and asphaltenes have the
some degree function of inducing the reaction of desulfurizing and polymerizing.
References
Abdelrahman, M. A. (1996). “Engineering characterization of the interaction of asphalt with

crumb rubber modifier (CRM)”, University of Illinois Urbana-Champaign, USA.
Billiter, T. C. (1996). “The characterization of asphalt-rubber binder”, Texas A&M University,

USA.
Yuan, C. (2001). “Prospect for application of waster rubber powder.” Synthetic Rubber
Industry, 24(2), 65-66 (in Chinese).
Zanzotto, L., Kennepohl, G. J. (1996). “Development of rubber/asphalt binders by

depolymerisation/devulcanization of scrap tires in asphalt.” Transportation Research
Record, 1530, 51-58.
Zhang, X. Y., Xu C. J., Zhang Y. Z. (2004). “The research summarizing of crumber rubber
modified asphalt (one).” Petroleum Asphalt, 18(4), 1-5 (in Chinese).
Chapter 4
Case Studies
Case Studies 619
The application of Semi-flexible Pavement

with Asphalt -Rubber on Heavy Traffic
Road
Tianqing Ling*- Wu Dan*

Wei Xia*- Zhijie Zhao**
Chu Huaxiong* - Li Changzhu*
* Institute of civil and architecture

Chongqing Jiao-tong University
No.66 Xue-fu Road, Na’an District
Chongqing City 400074, P.R.C
Lingtq@163.com
** Engineering
Highway Research Institute of M.O.C.
Beijing 100088, China
ABSTRACT: This paper focuses on the performance of semi-flexible pavement used in

heavy-traffic road according to the technical characteristics of complex pavement. The high
temperature stability, low-temperature cracking resistance and water stability of semi-flexible
pavement with asphalt-rubber were tested in the laboratory and compared to asphalt concrete
AC-16 with asphalt-rubber (AR-AC-16). In addition, an experimental section was built. The
dynamic stability of semi-flexible mixture AC-20 with asphalt-rubber (SFAC-20) was much
larger than that of the asphalt concrete AC-16 with asphalt-rubber (RAC-16). The residual
stability and freeze-thaw splitting strength ratio of SFAC-20 exceeded 21.2% and 11.7%,
respectively. Moreover, SFAC-20 mixture had larger failure tensile strain and stiffness modulus
under -10ºC temperature condition. The results showed that the semi-flexible pavement with
asphalt-rubber had better properties and was especially suitable for the surface course of
heavy-traffic road. Both laboratory tests and engineering practice showed that the semi-
flexible pavement with asphalt-rubber had excellent performance and application prospect.
KEY WORDS: Highway engineering, semi-flexible pavement, crumb rubber modified asphalt,
brittle point, strain energy density, heavy traffic
1. Introduction
With the development of industrial production of ground waste tires and its pavement
structure research, crumb rubber has become more and more popular in highway engineering
as a new road material to improve significantly the performance of asphalt concrete.
Modified asphalt preparation has been one of the efficient approaches for using crumb rubber.
Furthermore, modified agent of crumb rubber provides a good way to treat waste tires [4][5][6].
Semi-flexible pavement is usually defined a compound one prepared by porous mother

asphalt mixture filled with special cement slurry. Therefore, semi-flexible pavement combines
the properties of asphalt pavement and cement concrete pavement. It has been widely used to
improve the pavement lapse resistance in Japan and other countries. In this paper, according
to the research achievements of pouring semi-flexible pavement, the performance and the
application of crumb rubber modified semi-flexible asphalt pavement was studied considering
its energy saving and environment protection [2][7]. It helps to fill a domestic gap in the semi-
flexible field, improve the performance of semi-flexible and its application range.
2. Study on the mix design of semi-flexible pavement mixture with asphalt-rubber
2.1. Properties of raw materials
2.1.1. Crumb rubber
Crumb rubber used was of natural gradation which was ground at room temperature, and
its gradation is listed in Table 1.
Table 1. Gradation of crumb rubber
Size of sieve (mm) 2.36 1.18 0.6 0.3 0.15 0.075

Passing rate (%) 100.0 97.3 61.4 20.6 6.5 2.7
2.1.2. Asphalt-rubber
Asphalt-rubber was prepared by virgin asphalt of AH-70, and its properties are listed in
Table 2.
Case Studies 621
Table 2. Properties of asphalt-rubber
Test items Design requirements Asphalt-rubber

Content - 19%
Penetration (25ºC,0.1mm) 35~55 40
Soft point (ºC) >60 69.4
Elastic recovery at 25ºC (ºC) >70 85
Rotary viscosity(Pa·s，180ºC，20r/min) 1.5~5.0 3.9
2.1.3. Cement, fly ash and aggregate
Cement used in the study was Ordinary Portland Cement P.O42.5R from Chongqing. The
quality of fly ash was classified as the secondary class. The coarse aggregate used was crushed
by a typical limestone. A natural superfine sand from Yangtze River and limestone chip were
used as the fine aggregates. The limestone mineral powder was chosen as filler. Properties of
all materials met the requirements of standard [1].
2.2. Mix design of cement mortar
The cement mortar investigated in this paper included Ordinary Portland Cement, fly ash,
mineral powder, river sand, water, and additive (also toner in type of color mortar). According
to the related specification in Japan [2], cement mortar mix proportion was determined and is
listed in Table 3.
Table 3. Mix proportion of cement mortar
Strength at seventh day

Water/ River Mineral (MPa)
Fly- ash toner fluidity
Type cement sand powder
(%) (%) (s) Flexural Compressive
ratio (%) (%)
strength strength
Ordinary
0.65 14 10 6 - 11.4 4.4 17.2
type
Color type 0.68 14 10 6 4 11.6 4.9 18.4
2.3. Mix design of mother mixture with asphalt-rubber
The semi-flexible mixture with asphalt-rubber had a skeleton-dense structure. However,

mother asphalt mixture with asphalt-rubber had a skeleton-gap structure. Therefore, it could
supply sufficient voids for the complete filling of cement mortar based on skeleton formation.
This kind of pavement structure possessed good cementing power and its voids were filled
with high fluidity cement mortar. The volume characters of mother asphalt mixture played an
important role in the mechanical properties of semi-flexible pavement. Provided the volume
characters were mainly accounted for in the mix design process, mother asphalt mixtures
could perform very well. Designed void ratio of semi-flexible mother mixture with asphalt-
rubber was 20%-25%, and its asphalt content was 3.0%-5.0%. All indexes in details are listed
in Table 4 [2].
Table 4. Design requirement s of mother asphalt mixture
Compaction Asphalt Flow

Density Void ratio Stability
Item number content value
(g/cm3) (%) (kN)
(times) (%) (0.1mm)
Desired Each 50 at double
≥1.9 20~25 3.0~5.0 ≥3.0 20~50
value sides
2.3.1. Gradation of mother asphalt mixture
Because the asphalt film of asphalt-rubber was thicker and the void ratio of mixture with
asphalt-rubber was less than that of the virgin asphalt mixture, mineral aggregates gradation
was inclined to the lower limit in the mix design process so as to ensuring the complete
incorporation of cement mortar. Gradations of mineral aggregates are listed in Table 5.
Table 5. Composite gradation of mineral aggregates
Mix proportion Size of sieve（mm）

Size of aggregates
(%) 26.5 19 13.2 4.75 2.36 0.6 0.3 0.15 0.075
10~20mm 58.0 100 89.5 12.7 5.8 1.4 0.0 0.0 0.0 0.0
5~10mm 28.0 100 100 82.7 10.7 4.5 4.1 1.8 1.5 2.1
chip 10.0 100 100 100 78.7 47.0 16.9 12.4 4.5 2.31
Mineral powder 4.0 100 100 100 100 100 99.8 92.4 81.2 74.4
Composite gradation 100 93.9 44.2 16.9 9.9 6.5 5.2 4.1 3.8
Higher limit of gradation 100 100 60.0 24.0 18.0 14.0 10.0 8.0 6.0
Lower limit of gradation 100 93 40.0 12.0 8.0 6.0 5.0 4.0 2.0
2.3.2. Optimum asphalt- aggregate ratio
Five different asphalt contents were used based on the composite gradation and test results
are listed in Table 6. The relationship curve between asphalt-aggregate ratio and scattering loss
is shown in Figure 2. As shown in Figure 2, the minimum asphalt-aggregate ratio at inflection
point was 3.5%. From the relationship curve between asphalt-aggregate ratio and leak loss
(Figure 3), the maximum one at inflection point was 3.7%. Based on the range of 3.5%-3.7%,
the optimum asphalt-aggregate ratio of 3.6% was determined according to the designed the
void ratio.
Case Studies 623
Table 6. Indexes of mixtures at different asphalt contents
Asphalt-mineral Scattering loss Leak loss Void ratio Stability Flow value
aggregate ratio (%) (%) (%) (%) (kN) (0.1mm)
3.0 41.7 0.15 23.1 7.35 41.3
3.3 30.3 0.27 22.8 7.23 45.6
3.6 23.8 0.58 22.1 6.80 47.2
3.9 20.4 1.04 21.8 5.79 48.4
4.2 19.1 1.53 21.7 5.38 44.7
100
90 lower limit of gradation
80 higher limit of gradation
70 composite gradation
60
P (%)
50
40
30
20
10
0
26.5 19 13.2 4.75 2.36 0.6 0.3 0.15 0.075
Size （mm）
Figure 1. Composite gradation curve of mineral aggregates
45
Scatting loss ratio (%)
40
35
30
25
20
15 %
2.7 3 3.3 3.6 3.9 4.2 4.5
Asphalt-aggregate ratio
Figure 2. Relationship curve between asphalt-aggregate ratio and scatting loss ratio
1.8
1.6
Leak loss ratio (%)
1.4
1.2
1.0
0.8
0.6
0.4
0.2
0.0 %
2.7 3 3.3 3.6 3.9 4.2 4.5
Asphalt-aggregate ratio
Figure 3. Relationship curve between asphalt-aggregate and leak loss
3. Study on the properties of semi-flexible pavement mixture with asphalt-rubber
3.1. High temperature stability
The rutting test results of semi-flexible pavement mixture with asphalt-rubber were listed
in Table 7. As can be seen from Table 7, the dynamic stability of AR-AC-16 was about 2000
times/mm whereas that of semi-flexible mixture with asphalt-rubber was up to 15000 times/
mm. It was obvious that designed semi-flexible pavement mixture in the study had excellent
high temperature performance. It may due to the thinner asphalt film of semi-flexible and the
filled void by cement mortar. Therefore, the decreased degree of high temperature performance
of semi-flexible was smaller because of the asphalt softening at 60ºC temperature.
Table 7. Results of the rutting test
Type 1min 45min 60min Dynamic stability(times/mm)

Semi-flexible mixture 0.262 0.802 0.842 15750
AR-AC-16 0.635 1.972 2.243 2325
3.2. Water stability
As shown in Figure 4, compared to the AR-AC-16 asphalt mixture, the residual stability
and freeze-thaw splitting strength of semi-flexible mixtures with asphalt-rubber exceeded
21.2% and 11.7%, respectively, which showed that water stability of semi-flexible mixture
with asphalt-rubber was better than the AR-AR-16 asphalt mixture.
Case Studies 625
120
the residual stability
after soaking
100 freeze-thaw splitting
strength ratio
80
Strength ratio (%)
60
40
20
0
Semi-flexible mixture AR-AC-16
Type of mixtures
Figure 4. Comparison of water stability test results between different mixtures
According to the test results, the stability and residual stability of semi-flexible mixtures
with asphalt-rubber were 21.87kN and 21.72kN, respectively. Both of them showed a high
strength level. It is thought that the cement used in the semi-flexible mixture obviously
improved the rigidity of asphalt mixture and prevented the water into the mixture water. Also,
the stability of asphalt mixtures after 48h soaking was better than that of mixtures without
soaking. It’s because that the strength of cement slurry was not stable but increasing gradually
with time. Once ambient temperature and humidity increased, hydration of cement was greatly
enhanced and correspondingly the strength increased greatly, which resulted in the better
residual stability of mixtures after 48h soaking than that of mixtures without soaking. It was
explained from another point of view that this asphalt mixture has good water stability.
3.3. Low temperature crack resistance
Indirect tensile test was conducted to investigate the low temperature crack resistance of
semi-flexible mixture with asphalt-rubber. Test results are listed in Table 8. Compared to AR-
AC-16 asphalt mixture with asphalt-rubber, semi-flexible pavement mixture with asphalt-
rubber had lower splitting tensile strength, larger failure tensile strain, and lower stiffness
modulus, which showed that its low temperature crack resistance was better than the AR-
AC-16 asphalt. It may be the result of the thinner asphalt film and less temperature influence
of semi-flexible.
Table 8. Results of low temperature crack test
Splitting tensile Failure tensile Failure stiffness

Test condition Type of mixtures
strength (MPa) strain ( ) modulus (MPa)
Temperature is -10ºC, Semi-flexible
2.18 5136 873.52
and the loading speed mixture
is 1mm/min ARAC-16 3.83 3460 1318.74
4. A survey of the practice project construction
4.1. ProjectⅠ
A cement concrete pavement reconstruction project from Xiao nan-hai cement plant in
Chongqing was conducted in the study to investigate the construction process. This project
was a plant pavement, which had heavy traffic (rear axle load 500kN), large longitudinal
slope (the maximum up to 14.3%), and poor horizontal alignment (the minimum of horizontal
radius up to 10m). The pavement was damaged quickly during its service period with the
soaking of rainwater. Though this pavement was repaired a number of times, the whole service
quality of pavement decreased gradually, and some partial sections even became unsafe. To
restore its service quality, a design scheme was proposed. First, a 1cm-thick stress absorbing
maintenance layer with asphalt-rubber (AR-SAMI) was constructed after the treatment of
cement pavement. Then, a 4cm-thick semi-flexible pavement layer with asphalt-rubber was
paved on the top of it. The total area of this project was 7200m2 with a length of 800m and
a width of 9m. The used raw materials, the mix design and the cement mortar in the project
were similar to the laboratory.
After two service life periods with the loading of rear axle load 500kN, semi-flexible
pavement with asphalt-rubber showed an excellent lapse and rutting resistance and kept
complete compared with the ordinary concrete pavement with asphalt-rubber.
4.2. Project Ⅱ
The second project was a reconstruction one of Bei-jin road in Bei-bei district in
Chongqing. The length of this pavement was 7000m, and the width was 6.5m. The design
scheme was “treated aged cement pavement + 1cm AR-SAMI + 5cm ARSF-20”. The surface
of pavement was colored or paved with AC-13 of 3.5cm in some partial road sections to meet
the heavy traffic and landscape requirement.
Basalt crushed stone was used as aggregate in the project. Other materials were the same
as that of project one. The mix proportion of mother mineral aggregates was 85% 10-20mm
basalt crushed aggregate, 6% 5-10mm basalt crushed aggregate, 5% limestone artificial sand
and 4% limestone mineral powder. The optimum asphalt-aggregate ratio 3.3% was obtained
by the Marshall test. The performance of semi-flexible mixture with asphalt-rubber was
tested and the results are listed in Table 9. The project was constructed very well and praised
completely by the owners.
Case Studies 627
Table 9. Test results of semi-flexible mixture with asphalt-rubber
Before pouring the After pouring the cement

Index cement mortar mortar
Desired value Test result Desired value Test result
Marshall stability(kN) >3.0 4.2 >9.0 13.1
Flow value (0.1mm） 20~40 33.9 20~40 32.0
Void ratio (%) 20~28 21.28 <3.5 2.3
Bulk volume relative denstiy >1.9 2.056 -- 2.381
Dynamics stability (times/mm) 17800
4.3. Construction technology
4.3.1. Construction of AR-SAMI
Stress absorbing maintenance layer (SMI), defined as a layer located between the semi-
rigid base or cement pavement and asphalt pavement, was prepared by modified asphalt,
which has good deformability. It could result in stress relief of crack and avoid the formation
of asphalt pavement reflection crack. The construction process included: (1) obtained a
complete clean surface of base; (2) determined rubber powder percentage and prepare asphalt-
rubber; (3) sprayed asphalt-rubber with a content of 2.0-2.6 kg/m2 and aggregate with a
content of 16±2kg/m2; (4) compacted pavement using over 25t rubber-wheel roller on a clean
of loose aggregate.
4.3.2. Paving of mother mixture with asphalt-rubber
Semi-flexible pavement with asphalt-rubber was a pavement structure prepared by

including cement mortar in mother asphalt mixture. Therefore, it was necessary to control
accurately porosities of mother mixtures in the paving of mother asphalt mixture. The paving
parameters are listed in Table 10.
4.3.3. Preparation of cement mortar
Cement mortar was usually prepared by mixer or manual on site. All raw materials were
added in the following order: mineral powder, cement, fly ash, fine sand, early strength agent,
and water. Other materials should be completely mixed to be uniform before adding water,
then continued to be mixed with water for about two or three minutes to be uniform. The
manual mix was used because of the smaller amounts in Project I. The mixer was used in
Project II to improve the efficiency and the uniformity of strength.
Table 10. Construction parameters of mother mixture with asphalt-rubber
Virgin asphalt 170℃

Preparation of Temperature
Asphalt-rubber 180~190℃
asphalt-rubber
Reaction time Mixing 300r/min 45~60min
Asphalt-rubber 170~190℃
temperature
mixing Mixture 180℃
time Total mixing time 40s
Discharging temperature 190~200℃
transportation temperature
Mixture temperature on site 175~200℃
Paving speed 1~3m/min
Loose paving
1.16~1.18
paving coefficient
Paving
No less than 160℃, and discarded once lower than 140℃
temperature
double compaction three times by heavy double-
Combination of drum roller in the way of both high frequency and low
compaction
compaction frequency, final compaction one time by double-drum
roller in the way of static pressure
4.3.4. Filling of cement mortar
Firstly, the void ratio of mother mixture was determined by conducting core sample to
calculate required cement mortar amounts and design some indexes. Then, the fluidity of
cement mortar was adjusted to ensure a complete filling. Rubber harrow was first used to pave
the cement mortar repeatedly to ensure the natural filling of cement mortar. At the same time,
small vibratory roller or flat vibrator was used to enhance the filling effect. When there was
a longitudinal slope in the pavement, cement mortar need to be paved in the order of from
bottom to top so as to avoid a poor filling due to its quick flow.
Figure 5. Pouring construction of cement mortar using flat vibrator

Case Studies 629
4.3.5. Appearance treatment of pavement
It was essential that the residual cement mortar on the road surface was cleaned using
harrow to bare the irregular surface of mother mixture. Retarder was sprayed on the road
surface to obtain a good pavement structure which could be affected due to incorporation
of cement mortar in mother mixture (Figure 6). Then, when the final setting time of cement
mortar within the mother mixture ended, cement mortar on the road surface was washed
before its final setting time to ensure the strength of internal cement mortar, a good pavement
structure and the uniform pavement color (Figure 7).
Figure 6. Spraying of retarder
Figure 7. Wash of residual cement mortar
4.3.6. Curing
Curing time changed with the properties of cement mortar. The traffic was usually open
after two or three days curing, even a day provided the incorporation of early strength cement
or early strength agent in the cement mortar. The appearance of pavement after curing is
shown in Figure 8.
Figure 8. Appearance of pavement after curing
5. Conclusions
The results showed that the semi-flexible pavement with asphalt-rubber SFAC-20
had excellent high temperature performance. In addition, its residual stability and freeze-
thaw splitting strength ratio exceeded 21.2% and 11.7% respectively compared to RAC-16
mixture. Moreover, its low temperature (-10℃)crack resistance (the failure tensile strain and
stiffness modulus) was better than that of RAC-16 mixture. It was indicated that semi-flexible
pavement with asphalt-rubber could be used as an economical, feasible in technology and well
performed pavement material. It is feasible to construct the semi-flexible pavement using the
technology proposed by this study. Also, it is proved that the semi-flexible pavement is very
suitable for the pavement structure on heavy traffic road.
6. References
[1]《Technical Specifications for Construction of Highway Asphalt Pavement》(JTG

F40-2004), Beijing: China communications press, 2004.
[2] Japan Association of Roads: 《アスファルト鋪裝工事共通仕樣書解說》, Tokyo,

Maruzen Co., Ltd., December, Pingcheng 4.
[3] DONG Yingying. Study on the design parameters and construction technology of high
performance semi-flexible pavement, Master thesis of Chongqing Jiao-tong University,
2008
[4] YANG Zhifeng, LI Meijiang, WANG Xu-dong. The History and Status Quo of Rubber
Powder Used In Roadbuilding, Journal of Highway and Transportation Research and
Development, Vol.22,No.7,Jul.2005,19~22
[5] Rubber Pavement Association, How to Sell Asphalt-Rubber, Marketing Workshop,

June7,2000, Paradise Valley, Arizona.
Case Studies 631
[6] HUANG Peng, LU Weimin, ZHANG Fuqing, WEI Xiaojing. Research on performance
and technology of the rubber powder modified asphalt mixture, China Journal of
Highway and Transport, Vol.14 Sup.Dec.2001,4~7
[7] HAO Pei-wen, CHENG Lei, LIN Li. Pavement performance of semi-flexible
pavement in laboratory, Journal of Chang′an University (N atural Science Edition), Vo
l.23,No.2,Mar.2003,1~6
Case Studies 633
Asphalt Rubber - a New Concept for

Asphalt Pavements in Sweden.
Thorsten Nordgren* - Lars Preinfalk*
* Swedish Road Administration 405 33 Gothenburg Sweden

thorsten.nordgren@vv.se
lars.preinfalk@vv.se
ABSTRACT: ‘Asphalt Rubber’ is a bitumen, mixed with 15 to 20 % of rubber granules from

recycled tyres to improve the functional properties of asphalt pavements. The ‘Asphalt Rubber
’-concept for asphalt pavements has been successfully implemented in a number of US states.
The Swedish Road Administration (SRA) is currently involved in investigating the potential
implementation of this concept in asphalt pavements through a three-year development project
(2007-09). This paper presents promising test results from trials with asphalt rubber pavement
materials from the ongoing projects and evaluation studies.
KEYWORDS: ‘Asphalt Rubber’, performance, environmental effects, Sweden, Swedish Road

Administration
1. Introduction
During the autumn 2006 the Swedish Road Administration (SRA) decided to start a
development project for the period of 2007 - 2009 to investigate the potential implementation
in Sweden of asphalt pavements, produced according to the ‘Asphalt Rubber’-concept. This
concept, based on the ‘wet technique’, was developed in the United States (US) and has been
established in a number of states. ‘Asphalt Rubber’ has been produced in a large scale in US
since the end of the 1980’s. The international interest for ‘Asphalt Rubber’-pavements has
grown in recent years. At present the ‘Asphalt Rubber’-concept is being established in several
countries.
‘Asphalt Rubber’ can be characterized as a rubber-modified bitumen, where rubber

granules have been mixed with a standard bitumen (‘wet technique’). The rubber granules
originate from used tyres. The ‘Asphalt Rubber’ is manufactured in a specially developed
mixing equipment. The equipment is mobile and placed in connection to the asphalt plant,
producing the asphalt mix. The mixing of the components is carefully controlled and
followed by a procees of maturity. The content of rubber granules is 15 to 20 % of the mass of
the binder or 1,5 to 2,0 % of the mass of the asphalt mixture.
The expected long-term effects and potential benefits of the project are reduced annual and
life cycle costs, environmental advantages with respect to noise and emission of particles, and
also improved traffic safety in terms of increased friction.
Within the frame-work of the present project the following issues are evaluated:
• Technical conditions and possibilities for production
• Environmental effects and restrictions
• Technical performance of asphalt mix/pavement.
The evaluation is based on field trials/demonstration projects, where the performance

(functional properties) and possible influence on work environment as well as environment of
surroundings will be monitored.
Special focus is on the environmental consequences (both work environment and

environment of surroundings). The Swedish Chemical Agency has recommended a
restrictive use of old rubber tyres, in reference to their high contents of polycyclic aromatic
hydrocarbonates (PAH).
Asphalt pavements with rubber granules have been previously produced in Sweden, but
not according to the above ‘wet technique’. Instead a ‘dry technique’ was used, where the
rubber granules were added to the aggregate before the mixing with bitumen at the asphalt
plant. The main problem with this type of rubber asphalt was the varying quality. The ‘wet
technique’ according to the ‘Asphalt Rubber’-concept makes it possible to produce rubber
asphalt pavements of high and uniform quality.
Case Studies 635
This paper presents a progress report as of February 2009 regarding the Swedish
experience in 2007-2008 from asphalt pavements, produced according to the ‘Asphalt Rubber’
-concept..
2. Motives for the project
The Swedish Road Administration (SRA) considers that the ‘Asphalt Rubber’-development
project is a project of great economic importance and expects that the introduction of ‘Asphalt
Rubber’-pavements on the Swedish road net will result in
1. Reduced annual and lifecycle costs;
2. Environmental advantages, concerning reduction of noise and particle emissions;
3. Increased traffic safety;
Reduced annual and life cycle costs
Based on international experience the service life of ‘Asphalt Rubber’-pavements can

be doubled in comparison with corresponding standard pavements. Due to the uncertain
influence of studded tyres in Sweden on the service life of the ‘Asphalt Rubber’-pavements, it
is assumed that only half of the expected increase in service life can be utilized. The estimated
increase in costs for construction of the ‘Asphalt Rubber’-pavement is about 25 %. This
means that the annual cost for ‘Asphalt Rubber’-pavements will be about 80% of the annual
cost for the standard pavements. The annual maintenance cost for SRA’s asphalt pavements in
Sweden amounts to about 250 million Euro. Given that the ‘Asphalt Rubber’-pavements are a
marketable alternative, representing 25 % of that volume, the annual maintenance cost could
be reduced around 10 million Euro.
Environmental advantages
The tyre noise and the emission of particles, due to abrasion from studded tyres, are
expected to decrease and improve the environment.
Improved traffic safety
The traffic safety is expected to be improved on rain wet road surfaces by higher friction
between road surface and tyre and reduced formation of ‘water curtains’.
3. Objectives of the project
The project shall demonstrate the prospects for development and implementation of the
‘Asphalt Rubber’ concept for pavements on the Swedish road net. The main objectives of the
project are
1. Verification of the possible production of high quality ‘Asphalt Rubber’-pavements in

Sweden;
2. Verification of increased service life and reduced annual and life cycle costs;
3. Verification of environmental advantages;
4. Verification of management of possible health and environmental risks;
5. Verification of advantages considering traffic safety.
4. Preparatory actions and studies
The supply of technical expertise has been ascertained by engaging leading equipment
suppliers for manufacturing of rubber-modified bitumen. Contracts, covering both machine
and personal resources, have been established. In SRA’s own organisation the necessary
technical competence will be utilized during the accomplishment of the project.
A comprehensive preliminary study, including technical issues such as production

engineering and product quality, has been carried out during 2006.
Environmental aspects have also been studied by literature review and laboratory testing
and the preliminary results indicate that the environmental risks should be very limited.
Further studies of effects on health and environments form an integral part of the project and
shall be carried out in conjunction with full-scale trials.
5. Scope and time schedule of the project
The project activities during 2007-2010 are as follows.
2007 - Construction of experimental ‘Asphalt Rubber’-pavements on roads in the southern

and central parts of Sweden. Mixing equipment was supplied by SRA. Evaluation of the
hitherto achieved results. Decision was taken to continue the project.
2008 - Extended construction of experimental ‘Asphalt Rubber’-pavements, involving

different regional departments of SRA. Mixing equipment was supplied by SRA. Evaluation
of the hitherto achieved results. Decision was taken to continue the project.
2009 - Continued construction of experimental ‘Asphalt Rubber’-pavements, involving

different regional departments of SRA. Mixing equipment is supplied by SRA. Evaluation of
the hitherto achieved results. Preparation of the final report. Draft technical specification for
‘Asphalt Rubber’-pavements will be presented.
Case Studies 637
2010 - The procurement of the ‘Asphalt Rubber’-pavements will start. Project information
will be disseminated to suppliers and contractors to instil a genuine interest in the ‘Asphalt
Rubber’-concept and to create a potential market for ‘Asphalt Rubber’-pavements.
6. Manufacture and evaluation of the ‘Asphalt Rubber’-binder
The rubber granules and the manufactured ‘Asphalt Rubber’ are specified according
to the Arizona Department of Transportation “Standard specification for Road and Bridge
Construction, 2000”.
The rubber modified bitumen is manufactured in a specially designed mixing

equipment. It is mobile and placed in connection to the asphalt plant, producing the
asphalt mix. See Figure 1.
The rubber granules (see Figure 2) are produced by fragmentation and sieving of recycled
tyres, including separation of steel and fibre reinforcements. The rubber granules are mixed
with normal standard bitumen in a closed process. When the Asphalt Rubber mix has the
designed viscosity (normally 2, 5 – 3 Pas) it is ready for use as a normal bitumen in the
asphalt production. The content of rubber granules is 15 to 20 % of the mass of the binder or 1,5
to 2,0 % of the mass of the performed asphalt pavement.
The mixing equipment, used by SRA, is rented from the equipment supplier Phoenix
Industries LLC (PI), Arizona. Technical experts from PI have also been responsible for
training in production technique and technical support.
Figure 1. Mixing equipment connected to the asphalt plant

Figure 2. Rubber granules, used in SRA´s trials
Before the road trials of 2007 and 2008, the ‘Asphalt Rubber’-binder (AR) as well as
the trial mixes, produced with this binder, has been thoroughly evaluated in the laboratory
to ensure the right quality of the manufactured products. The rubber content of the rubber/
bitumen-mix has been 16-19 %. The target viscosity at 175 º C after a reaction (mature) time
of 45-60 min has been 2 to 3 Pas.
On the construction site the AR-viscosity is regularly controlled by a manual field

viscometer. If the viscosity is deviating from the target value, the rubber content is
adjusted to get the correct viscosity. The AR production temperature has been in the range
of 170 to 175 º C.
Samples of the rubber granules and the manufactured AR are taken continuously and
analysed at a certified laboratory. The basic bitumen consists of pen grade 70/100. The rubber
granules have been purchased mainly from two suppliers, Genan in 2007 and Eximlink in
2008. AR-gradings, supplied by Genan, are shown in Figure 3.
100
80
Passing by mass - %
60
40
20
0
0,063 0,125 0,25 0,5 1 2
Sieve - mm
Figure 3. Three typical gradings of rubber granules from the production in 2007.
Case Studies 639
7. Specifications and production of the ‘Asphalt Rubber’-mix
The specifications of the ‘Asphalt Rubber’-mixes for the road trials are based on
the Arizona Department of Transportation “Standard specification for Road and Bridge
Construction, 2000”. Certain modifications of the specifications for gradings and void contents
have been made to fit the Swedish practice for surface course pavements.
During 2007 and 2008 nearly 30 000 ton of ‘Asphalt Rubber’-mixes have been produced.
The major part of the produced mixes was Stone Mastic Asphalt (SMA) - mixes, designated
SMA 11 and 16. However, the filler content and the content of fine aggregate 0/2 mm have
been decreased to provide room for the rubber granules (see Figure 4). The target added
content of ‘Asphalt Rubber’ has been in the range of 8,3 to 9,0 % by mass. For the mix design
the target void content, provided by Marshall impact compaction at a compaction temperature
of about 170 ºC, has been 2,0 to 2,5 %. The temperature at the production of the ‘Asphalt
Rubber’-mix has been 165 to 170 ºC.
100
GAP 16
ABS 16
80
Passing by mass - %
60
40
20
0
0,06 0,13 0,25 0,5 1 2 4 8 16 32
Sieve - mm
Figure 4. Typical aggregate gradings for the ‘Asphalt Rubber’-mix, GAP 16, and the
reference mix, SMA 16
Two interesting minor trials with an open-graded (void content > 15 %), low noise
‘Asphalt Rubber’-pavement, based on the concept used in Arizona, have also been carried out.
However, it was observed that this concept needs some adjustment and development, mainly
because of the studded tyre use during the winter in Sweden.
8. Evaluation of performance and functional properties of the ‘Asphalt Rubber’-mix/

pavements
Asphalt mixes and constructed pavements, involved in the project, are studied by a
comprehensive monitoring programme, including field observations and measurements and
extensive laboratory studies of primarily the achieved functional properties.
ASU (Arizona State University) in Phoenix and VTI (Swedish Transport and Road
Institute) in Linkoping have been engaged in cooperative work on laboratory studies. Also
different contractor laboratories have been involved in the factory production control and
quality control and evaluation of the constructed ‘Asphalt Rubber’-pavements.
ASU has studied two asphalt mixes, laid on road E6 in the vicinity of Malmoe 2007 and
consisting of a ‘Asphalt Rubber’-mix, GAP 16 (a gap-graded asphalt mix with a nominal
upper aggregate size of 16 mm) and a reference asphalt mix, SMA 16 (stone-mastic asphalt
mix with a nominal upper aggregate size of 16 mm and a penetration bitumen with grade
70/100). The results of this study are reported separately in these AR2009 conference
proceedings.
In this paper some results from the VTI study, based on the test methods used in Sweden
are presented. At VTI the main part of the studies has been carried out on cores, extracted
from the pavement on the outer ring-road in Malmoe, constructed in 2007. The principal
purpose of the ‘Asphalt Rubber’-mix, laid on this test section, is to counteract the crack
propagation from the underlying courses. The existing pavement, constructed during
2000-2001, is a conventional asphalt pavement of 10 cm thickness (4 cm surface course and 6
cm binder course), laid on a cement bound granular base. Extensive reflection cracks from the
underlying cement bound base course appeared shortly after the traffic opening in 2001.
9. Testing conducted by VTI in Sweden
Functional properties such as flexibility (stiffness at different temperatures) resistance to

fatigue, stability, resistance to abrasion from studded tyres and durability are essential for the
assessment of the quality of an asphalt pavement. At the VTI study most functional properties
have been determined by test methods, which are in line with current European test standards.
Special test procedures, developed at VTI, have been used for the winter conditioning and
crack propagation tests. See Table 1.
Case Studies 641
Table 1. Test methods used at the VTI study
Property Type of test Method used by VTI Corresponding

European Standard
Stiffness Indirect tensile test FAS Metod 454a EN 12697-26
Resistance to fatigue Indirect tensile test VTI own method EN 12697-24
Stability (Resistant Uniaxial cyclic FAS Metod 468 a
EN 12697-25
to permanent compression test
deformations) with confinement
(Dynamic creep)
Resistance to abrasion ‘Prall’-test FAS Metod 471a EN 12697-16
from studded tyres
Water sensitivity Indirect tensile test FAS Metod 446a EN 12697-12
Durability – Freeze-thaw cycles VTI’s own method Stiffness modulus
sensitivity to winter with saltwater; according to EN
conditioning reduction of stiffness 12697-26
modulus
Crack propagation Wheel-tracking test VTI’s own method Wheel-tracking
sensitivity machine according to
EN 12697-22
Test method prepared by FAS, the former Swedish Asphalt Pavement Association.
a
9. Results of the VTI study
The results are based on cores, extracted from an ‘Asphalt Rubber’-pavement, laid on the
outer ring-road E 6 between the Sallerup and Fredriksberg interchanges. Cores with 100 mm
and 150 mm diameters, sampled September 2007. have been dispatched to VTI. Some results
are presented below. The complete results will be presented in a future VTI-publication.
9.1. Stability according to the dynamic creep test
The results of the stability/dynamic creep test are shown in Figure 5. The mean
accumulated strain after 3600 load cycles amounts to 14 500 microstrains and fulfils the most
strict requirement (<15000 microstrains) for surface courses according to SRA´s technical
specifications, 2008.
20000
Cumulative axial strain - microstrains
15000
10000
5000
0
0 1000 2000 3000 4000
Number of load applications
Figure 5. Accumulated strain as a function of the number of load cycles at 40 ºC (single test
specimens: black curves; blue average curve).
9.2. Resistance to fatigue
The loading of the fatigue test was applied in the same way as the loading of the stiffness
test and continued, until failure occurred in the test specimen. The ‘Asphalt Rubber’-test
specimens sustained about 1 million load applications at 100 microstrains and satisfied the
most strict fatigue requirement of SRA´s technical specifications, 2008.
9.3. Crack propagation sensitivity
VTI has developed a special test procedure for determination of the crack propagation
sensitivity of the rubber asphalt, based on VTI´s wheel tracking machine (WTT), specified
under clause 6.2 ‘Extra large devices’ in EN 12697-22. Test slabs are placed in the wheel-
tracking device on a divided soft rubber base in order to generate cracks through the pavement
slab. Duplicate strain gauges are mounted on the lower and upper surface of the slab, enabling
recording of the crack propagation course from the bottom to the top of the slab. The test was
conducted on 3 slabs, prepared using the ‘Asphalt Rubber’-mix, and 3 slabs, prepared using
the reference asphalt mix, SMA 16 (bitumen 70/100).
Test conditions
Slab size: 50 x 70 x 4 cm (slab mass ≈ 30 kg).
Temperature: +5 °C.
Wheel load: Adjusted to the desired strain level
Number of wheel passes: Adjusted to the crack propagation
Case Studies 643
The number of wheel passes to initiate cracking at the bottom and at the top of the slab are
recorded at a given level of initial strain (= recorded strain after 100 wheel passes). The initial
strain reflects the load applied to the test specimen. The difference between the two numbers
of wheel passes is a measure of the crack propagation rate of the test specimen. The results
for the six test slabs are shown in Figure 6. The rubber asphalt slabs are less sensitive to crack
propagation at the same initial strain level than the reference slabs.
Figure 6. Effect of initial strain on the required number of wheel passes to cause complete
crack propagation of the rubber asphalt and reference test specimens.
9.4. Sensitivity to winter conditioning
The winter conditioning method has been developed at VTI by Peet Höbeda. The
conditioning phases are intended to simulate those stresses that roads are exposed to
with alternating temperatures around zero in combination with salting. The effect of the
conditioning is assessed by measuring the stiffness modulus. The ratio of the stiffness modulus
before and after the conditioning is designated Q-winter and expressed in percent (if Q-winter
is 100 %, the winter conditioning has no influence). The advantage of the stiffness modulus
test, besides its sensitivity to variations of the properties of the test specimen, is its non-
destructive mode, i.e. the same test specimen can be tested before and after conditioning. In
this case the winter conditioning did not reduce the stiffness modulus of the ‘Asphalt Rubber’
-test specimens; the actual Q-winter value was equal to or greater than 100 %.
10. Environmental studies
Comprehensive studies have been carried out to explore the impacts of the ‘Asphalt
Rubber’-concept on the work environment and the environment of surroundings.
10.1. Work environment
Initially a laboratory study was conducted, where bitumen fumes were generated at 160 0
C and cooled down to a condensate. The condensate was analysed chemically to get a picture
of what substances could be found in the different bitumen products. The programme for the
field measurements was established based on the results from this study. The measurements
on the paving site included the content of polycyclic aromatic hydrocarbons (PAH), released
from bitumen and asphalts with rubber additives at six different pavement sites during 2008.
Measurements were conducted in the inhalation zone of the paver operator and the screed
operators, above the augers of the paver and also on the workers at the production unite for
‘Asphalt Rubber’.
The results of the measurements were as follows.
1. The highest PAH-contents arose directly above the augers. The maximum
benso(a)pyrene and naphthalene contents of the inhalated air was far below the limiting
values
- 0,03 µg/m3of benso(a)pyrene << threshold limit value = 2 µg/m3;
- 2,4 µg/m3of naphthalene << threshold limit value = 50 000 µg/m3;
2. More PAH was released, when the mixture was strongly heated, as by using a heater
especially in combination with remixing.
3. The PAH-exposure was lower at the plant, where rubber was added, than around
the asphalt paver at the site. The wind direction and velocity in conjunction with the
temperature of the asphalt mix appeared to be the factors controlling the degree of
exposure at the paver.
10.2. Leaching tests
The Swedish Geotechnical Institute (SGI) in Linkoping has investigated the leaching of
organic substances and compounds from road normal asphalts, produced with and without
additives of rubber (from recycled tyres), granulated to particles < 20 mm.
The leaching contents of most of the analysed substances and compounds were low for
both materials. The acute toxic effects of the leachate were also low.
The rubber additive caused somewhat increased leaching of other PAHs and of cresoles.
However, the contents and the accumulated amounts of these compounds were generally low.
There was no carcinogen PAH above the detection limit in any leachate.
The rubber additive generated leaching of bensotiazol, detected by GCMS-screening. The

estimated contribution of semi-quantative and accumulated leached contents of bensotiazol
to the nearby surroundings of the road were considerably higher for the asphalt with added
rubber. However, the preliminary assessment, based on the US limiting value for a closely
related compound in surface and ground water, is that leached bensotiazol should not generate
any serious contents in leachates, caused by rain on the monolithic road surface of the
Case Studies 645
material. There was also no relation between leached bensotiazol and acute toxic effects.
10.3. Particle emission caused by abrasion from studded tyres
VTI has studied the emission of inhalable abraded particles, using the road test machine,
shown in Figure 7. Two runs were conducted; one run with an ‘Asphalt Rubber’-pavement,
GAP 16, and the other one with a reference pavement, SMA 16. The particle size distribution
of the emitted particles and their chemical and morphological properties were analysed,
especially PM10 (= the mass concentration of inhalable particles finer than 10 µm, expressed
in µg/m 3), which is regulated by an EU directive and implemented in Sweden as an
environmental quality standard.
Altogether, the results showed that the ‘Asphalt Rubber’-pavement contributed to lower
emissions of both PM10 and ultra fine particles than the reference pavement.
Figure 7. VTI road test machine for abrasion and particle studies
11. Concluding remarks
The overall experiences from the studies so far are positive, and there is no reason to
discontinue the project. The results from the comprehensive tests and investigations, covering
technical, performance-related and environmental aspects, are in agreement with the results,
obtained in US.
The remaining parts of the project include:
- Development of a cost effective management of the equipment for involving rubber in

the asphalt production process;
- Reduction of the production temperature, which is advantageous for the working

environment as well as the emissions to the surroundings.
- Further development of the concept of ‘Asphalt Rubber’- pavements for “low noise
pavements” to improve the resistance to abrasion from studded tyres.
Bibliography
Arizona Department of Transportation (ADOT), ‘Standard Specifications for Road and Bridge
construction 2000’.
Gustafsson, M. Blomqvist, G. Jonson, P. Gudmundsson, A., ’Slitagepartiklar från vägbelägg-

ning med gummiinblandad bitumen’, VTI notat 10-2009.
Höbeda, P., ‘Testing the durability of asphalt mixes for severe winter conditions’, 2nd
Euroasphalt & Eurobitume Congress, Barcelona 2000.
Kriech, A. (2007) Heritage Research Group. Chemical comparison of laboratory generated

fumes from straight run vacuum distilled and two different crumb rubber modified
bitumens. Prepared for Swedish Road Administration.
Larsson, L. , ’Kolonnlakning av vägbeläggningsmaterial’, Statens Geotekniska institut (SGI),

2008.
Marcus, Hans-Olof, ’Litteraturstudie av miljö- och arbetsmiljöeffekter av inblandning av däck

i vägbeläggningar’, IVL Svenska miljöinstitutet, 2006.
Viman, L. Said, S. Hakim, H. Waldemarson, A., ‘Gummiasfalt laboratorieförsök’, VTI (report

will be published 2009).
Case Studies 647
The Use of Asphalt Rubber in a Motorway

Section Pavement Rehabilitation in Portugal
Eduardo Fung* and Dora Baptista**
*COBA, Consultores de Engenharia e Ambiente, S.A.

Av. 5 de Outubro, 323
1649-011 Lisboa
Portugal
ef@coba.pt
**BRISA, Auto-Estradas de Portugal, S.A.

Quinta da Torre de Aguilha
Edifício Brisa
2785-599 São Domingos de Rana
Portugal
maria.baptista@brisa.pt
ABSTRACT: Taking advantages of its several benefits, there has been a growing use of asphalt
rubber in the bituminous mixes for the pavement wearing course. Asphalt rubber in the
surfacing course, among other advantages, can strongly contribute to cope with the reflective
cracking and to mitigate the road noise, in maintenance and rehabilitation works.
An open graded asphalt rubber was used in the rehabilitation works of a highly trafficked
motorway section (M5), with over 120 000 vehicles per day, for the surfacing course. The
motorway section, built in Portugal in the early 40’s, was the first one with portland cement
concrete slab pavement, with a 2x2 cross section. In the early 90’s, the existing pavement
was widened to 2x3 and to 2x4, again with portland cement concrete slab pavement, overlaid
in the entire cross section with bituminous mixes. During the last years, extensive reflective
cracking was observed on the pavement surface. The paper describes the existing previous
pavement distresses, the technical solutions and the results, emphasizing the contribution of
asphalt rubber to overcome the reflective cracking and to mitigate the road noise.
KEYWORDS: asphalt rubber, pavement rehabilitation, reflective cracking, noise.

1. Introduction
The motorway section of M5 in Portugal, between Lisbon and the National Stadium
Interchange, with a traffic of over 120 000 vehicles per day, has a length of 8 km and the
pavement is composed by concrete slabs covered with bituminous layers. The original
pavement was in concrete and built in the decade of 1940, being the first motorway in
Portugal, with two carriageways of two lanes each.
In 1988 the Portuguese main motorway concessionary (BRISA) was committed to carry
out the pavement widening to three lanes in each carriageway. In early 90’s the widening and
rehabilitation works took place, with concrete slabs both inside and outside of the existing
cross-section and then the entire platform was covered with bituminous layers.
In the last years, extensive longitudinal and transverse cracking, among other localized
distresses, were observed on the pavement. BRISA decided to carry out pavement
rehabilitation works and in the year 2006 the engineering consultant company COBA was
committed to undertake the study of the sub-section between km 5+500 and km 8+000.
The rehabilitation works were carried out during the month of August of 2007, which is
the traditional period of holidays with less traffic in the motorway and were restricted to night
time and weekends.
One year later, during the month of August of 2008, the rehabilitation works of sub-section
between km 2+700 and km 5+500 took place in similar conditions and in both cases an open
graded asphalt rubber layer was laid as a finishing course with very promising results. Sub-
section between km 0+000 and km 2+700 is scheduled to be rehabilitated in August 2009.
2. Description of the Existing Pavement Condition
The distresses observed on the bituminous surface before rehabilitation were mostly
longitudinal and transverse cracks, mainly originated from the underlying concrete slabs
joints. A frequent occurring situation was the existence of a longitudinal crack approximately
along the middle of the centre lane, with a path apparently associated with the longitudinal
joint between the old concrete adjacent underlying slabs. The thickness of the bituminous
layers acknowledged from cores extracted from the pavement, varied between a minimum of
9,5 cm to a maximum of 22,5 cm, with frequent values of 14 to 18 cm.
The sub-section from km 5+500 to km 8+000, was the one which presented more defects
on the pavement. On the left carriageway, between the underpass located at km 7+850 and
about km 7+600, the pavement was comparatively in worst condition, with several types
of distresses occurring. Close to the underpass joint there was a substantial settlement on
the pavement. Deformations mainly of longitudinal surface depression type, were observed
mainly on the right lane, with substantial ruts between km 7+700 and km 7+600. On the
acceleration lane, approximately between km 7+450 and km 7+270, deformations and
cracking were observed. An illustrating schematic diagram and photo are presented in Figures
1 and 2.
Case Studies 649
7 +700 7 + 600 7 + 500
Longitudinal crack
Rut
Rut
7 + 500 7 + 400 7 +300 7 + 200
Less wider crack Wider crack
Figure 1. Schematic diagram with the location and the type of previous existing distresses on
the left carriageway pavement
Figure 2. Deformations and ruts on the left carriageway right lane pavement, approximately
at km 7+600
Deflection tests were carried out to assess the pavement structural condition, using
a Carlbro van-integrated PRI 2100 model Falling Weight Deflectometer. The results
obtained normally didn’t present values larger than 100 µm, indicating that the pavement
was generally in good structural condition. However, in some zones the values were
significantly higher. For example, close to the longitudinal crack of the central lane, on
the left carriageway, between km 7+700 and km 7+600, deflection values on the plate
center (Df1) were about 450 µm (fig 3).
FWD Tests: M5, Left Carriageway, Central Lane, From Km 5+400 to

Km 8+000
450
Deflections with 65 kN (µm)
D1 D2 D3 D4 D5 D6
400 D7 D8 D9
350
300
250
200
150
100
50
0
5+400
5+500
5+600
5+700
5+800
5+900
6+000
6+100
6+200
6+300
6+400
6+500
6+600
6+700
6+800
6+900
7+000
7+100
7+200
7+300
7+400
7+500
7+600
7+700
7+800
7+900
8+000
P.K.
Figure 3. Deflections with the Falling Weight Deflectometer on the central lane of the left
carriageway.
In order to assess the pavement structure (thickness and layers), crack depth and layers
bond, some cores samples have been taken. Figure 4 shows two of those cores samples, and it
can be observed the concrete slabs underlying the bituminous layers (the older one of 40’s and
the latest from the widening of 90’s) as well as the crack propagated through the bituminous
layers.
Figure 4. Cores samples extracted from the pavement and the crack propagation through the
bituminous layers can be observed
Case Studies 651
The sub-section between km 2+700 and km 5+500 and the sub-section between km
0+000 and km 2+700, were comparatively in better condition than the above described. The
distresses were again mostly longitudinal and transverse cracks, mainly originated from the
underlying concrete slabs joints.
3. Rehabilitation Works
3.1. Rehabilitation Solutions
Several solutions were developed to solve the pavement distresses of the sub-section
between km 5+500 to km 8+000, the first of the three sub-sections where BRISA decided
to carry out the rehabilitation works. The objective was to eliminate or retard with good
efficiency the crack propagation to the bituminous surface and reinforce the structural
capacity of the weak zones, which were detected in the deflection tests. In addition, the final
surface after rehabilitation, should also meet the BRISA strict specifications for motorways,
concerning riding quality, skid resistance and texture.
The longitudinal regularity evaluated by the International Roughness Index (IRI), in 100 m
intervals, after rehabilitation, should comply with the following:
- Max. IRI of 1,5 m/km to be achieved in each batch, in more than 50% of measurements
of the analyzed batch;
- Max. IRI of 1,8 m/km to be achieved in each batch, in more than 80% of measurements
of the analyzed batch;
- Max. IRI of 2,0 m/km to be achieved in each batch, in 100% of measurements of the
analyzed batch.
The existing values (year 2006) before rehabilitation works are summarized in Table 1:
Table 1. Values of I.R.I measured in year 2006
Eastbound
IRI (m/km) <3 <2,5 <2 <1,8 <1,5

% 92.6 77.8 22.2 11.1 0.0
Westbound
IRI (m/km) <3 <2,5 <2 <1,8 <1,5

% 100.0 82.4 58.8 41.2 17.6
The solutions were named as corrective actions and labelled from A1 to A8. A1, A2, A6
and A8 were mainly related to the milling of the existing bituminous surface layers, cracks
sealing and the placing of glass grids and bituminous mixes. A3, A4 and A5, were related to
the treatment of the damaged slabs, using high initial strength concrete and the filling of the
gap between slabs with grout.
After all the corrective actions have finished, an open graded asphalt rubber mix was laid
as surface course, with 3 cm of thickness, on the top of a underlying 5 cm asphalt concrete
layer.
The IRI final results after the rehabilitation works were finalized, at the beginning of
September 2007, are presented in Table 2:
Table 2. Values of I.R.I (m/km) measured after rehabilitation works
Eastbound
IRI <3 <2,5 <2 <1,8 <1,5

% 100,0 100,0 100,0 91,7 54,2
Westbound
IRI <3 <2,5 <2 <1,8 <1,5

% 100,0 100,0 100,0 95,8 83,3
Deflections measurements were carried out in order to assess the effectiveness of the
rehabilitations solutions, along parallel alignments in the vicinity of the former cracks.
Deflections measurements before and after rehabilitation works regarding the areas where
corrective actions A3 and A5 were carried out, are presented in Tables 3 and 4. The results
showed significant differences between the two sets of measurements, with a considerable
decrease of the deflections.
Case Studies 653
Table 3. Deflections measurements before and after rehabilitation works regarding the area
covered by corrective action A3
Before After Rehabilitation – Deflections in µm

P.K. Rehabilitation
Deflections in µm Alignment 1 Alignment 2 Alignment 3 Alignment 4
D1 D9 D1 D9 D1 D9 D1 D9 D1 D9
7+700 387 57 93 28 90 27 122 43 143 46
7+680 271 115 79 31 82 31 156 48 116 38
7+660 492 104 81 29 78 31 107 35 112 37
7+640 400 96 73 26 78 27 105 37 113 38
7+620 379 174 83 32 81 31 117 40 107 39
7+610 457 183 76 28 96 29 119 41 134 44
Table 4. Deflections measurements before and after rehabilitation works regarding the area
covered by corrective action A5
Before After Rehabilitation – Deflections in µm

P.K. Rehabilitation
Deflections in µm Alignment 1 Alignment 2 Alignment 3 Alignment 4
D1 D9 D1 D9 D1 D9 D1 D9 D1 D9
7+760 135 43 61 20 57 19 109 25 80 26
7+750 94 30 97 32 140 44 114 38
7+725 98 31 96 30 102 30 106 31
7+700 236 44 164 48 191 44 129 40 168 48
7+675 150 41 126 43 115 36 115 37
7+650 108 30 115 37 109 33 111 37
7+625 139 47 115 38 105 37 98 32 129 38
7+600 439 262 140 52 123 42 127 47 141 46
7+575 256 157 112 41 115 47 103 36 131 43
7+550 120 41 115 42 103 36 116 38
As mentioned previously, sub-section between km 2+700 and km 5+500 was generally

in better condition. In the rehabilitation works of this sub-section, the existing bituminous
surface was subjected to a micro grinding, previous to the placing of the asphalt rubber
wearing course, in order to eliminate the irregularities.
After all the corrective actions have finished, which included sealing and laying of a glass
grid on persisting cracks, an open graded asphalt rubber mix was laid as surface course, with
4 cm of thickness. This thickness was recommended as the minimum above the glass grid by
the supplier of this product.
3.2. Asphalt Rubber Mix Layer
The open graded asphalt rubber mix was composed by fine and large aggregates, crumb
tire rubber modified bitumen and a commercial filler.
A 35/50 pen grade asphalt cement base was modified adding crumb tire rubber, in a
proportion of 21% in weight (over total weight of binder), complying with specifications
(%varying between 20 and 23%), with a specified grading and in appropriate blending and
reaction conditions.
This modification is made with appropriate equipment, which is installed next to the
conventional bituminous plant, with a “by-pass” to the circuit of the original binder, between
the bitumen storage tank and the bituminous plant mixer. This equipment heats and mixes the
original binder with the rubber, allowing a chemical/physical reaction (“digestion”) between
these two materials.
The crumb rubber specification for the grading was according to Table 5:
Table 5. Grading of the crumb rubber
Grading
Sieve opening ASTM % passed
2 mm (n.º 10) 100
1,18 mm (n.º 16) 65 – 100
0,6 mm (n.º 30) 20 – 100
0,3 mm (n.º 50) 0 – 45
0,075 mm (n.º 200) 0–5
The requirements (ASTM D6114 Standard) for the rubber modified bitumen were
according to Table 6:
Case Studies 655
Table 6. Requirements for the rubber modified bitumen
Tests Requirements
Apparent Viscosity Brookfield at 175ºC, cP (AASHTO TP48) 2000-4000
20 minimum
Cone Penetration at 25º C 100 g, 5s, 1/10 mm (ASTM D5)
75 maximum
Resilience at 25ºC (ASTM D 3407) 15% minimum
Ring and Ball Temperature (ASTM D 36) 54º C minimum
The mix aggregate grading specification was according to the following (Table 7):
Table 7. Grading of mix aggregate
Sieve opening ASTM % passed

12,5 mm (1/2") 100
9,5 mm (3/8") 90 – 100
4,75 mm (nº 4) 35 – 50
2,00 mm (nº10) 6 – 10
0,425 mm (nº40) 3–7
75 µm (nº200) 2–3
The amount of commercial filler in the bituminous mix is, at least, 2% by weight of the
mineral aggregate and should be Portland cement or hydraulic lime appropriately hydrated.
The bituminous mixing formulation is composed by two distinct stages. At the first stage
the physical characteristics of the mixture are determined by the Marshall method and two
distinct mixtures, at most, are defined, to be tested at the second stage.
The following values were specified for the test results of the bituminous mixture with the
Marshall method:
- Number of blows in each specimen edge face ……………………………………… 50
- VMA value (percentage of Voids in the Aggregade mix), minimum ……………… 25%
- Porosity ……………………………………………………………………… 9 to 15%
- Retained strength minimum ……………………………………………………… 80%
- Percentage of rubber modified binder ………………………………… 9,5% to 10,5%
- Cantabro Loss, maximum ………………………………………………………… 15%

At the second stage, an experimental section is carried out whereby, after appropriate
conclusions are drawn out, the final mixing composition to be used on job site is defined and
approved by the site Engineer.
The job composition used was composed by 49% of basaltic 4/6 fraction, 39,9% of
granodioritic 6/10 fraction, 1,8% of hydraulic lime and 9,3% of rubber modified bitumen. The
asphalt rubber was produced by the “wet process”.
A thermo-adherent polymer modified emulsion tack coat was placed with a 0,35 kg/m2
rate previous to the laying of the asphalt rubber wearing course.
The rubber asphalt layer was placed and finished with wide pavers, working in parallel in
order to cover the entire carriageway, avoiding thus longitudinal construction joints.
Figure 5 shows the placing of the rubber asphalt wearing course.
Figure 5. Placing the rubber asphalt wearing course
The reduction of tire-road noise was evident, with significant difference in comparison
with the remaining sections of traditional concrete asphalt wearing course. The reductions
achieved are approximately of 5 dB, according to measurements already carried out in another
section of road in Portugal.
4. Conclusions
The use of asphalt rubber is experiencing an interesting growth due to the advantages of
its properties regarding resistance to crack propagation and noise reduction, as well as those
related to the environmental issues.
Asphalt rubber has been used since 1999 in Portugal and started with the rehabilitation of
single carriageway roads. Considering the encouraging experience and the results achieved,
the main motorways concessionary BRISA has recently decided to use this technology in the
pavement rehabilitation of some sub-sections of two motorways (M5 and M9) in the Lisbon
district.
Case Studies 657
The paper described a pavement rehabilitation case of a highly trafficked motorway

section, in which the use of an open asphalt rubber was considered a good option for the
wearing course, in combination with other solutions for the underlying layers.
5. Bibliography
Antunes, Mª de Lurdes, Baptista Fátima and Fonseca, Paulo – “Asphalt Rubber Mixtures in
Portugal: Pratical Application and Perfomance”. Asphalt Rubber 2006
Fung, Eduardo e Baptista, Maria Dora – “A5 – Auto-estrada da Costa do Estoril. Lanço
Lisboa/Estádio Nacional. Sublanço Linda-a-Velha/Estádio Nacional. Beneficiação/Reforço do
Pavimento”. Revista Engenharia e Vida, 2008.
Fung, Eduardo e Pinto, Nélia “A5 – Auto-estrada da Costa do Estoril. Lanço Lisboa/
Estádio Nacional. Sublanço Linda-a-Velha/Estádio Nacional. Beneficiação/Reforço do
Pavimento - Projecto de Execução”- COBA, 2006.
Gonzalez, Isabel – “Beneficiação e Reforço de Pavimentos na Rede de Auto-estradas da

Brisa aplicando Misturas Betuminosas com Betume Modificado com Borracha”. V Congresso
Rodoviário Português - Estrada, 2008.
Case Studies 659
Crumb Rubber From Scrap Tyres

For Use in Asphalt Pavements in the UK
Robert J Hewson RE, MICE, MInstRE, MSc, BEng.
C/o HQRE
1 UK Armd Div
Wentworth Bks
HERFORD
BFPO 15
hewson_robert@hotmail.com
ABSTRACT: An examination of the potential advantages and disadvantages of scrap tyre

crumb rubber in road construction (Crumb Rubber Modified Asphalt, CRMA), in comparison
to traditional asphalt. The paper highlights the economic influence of UK government
initiatives such as the Landfill Tax, Aggregate Tax and proposed Incineration Tax before
considering the financial implications of the performance of CRMA in comparison to
traditional asphalt and the potential reduction in pavement thickness and increase in service
life. Examining the friction properties and cold weather performance of the pavement as
well as environmental and noise pollution issues and the subject of scrap tyre recycling: The
paper brings together research, experience and best practise from around the globe to draw
conclusions for the suitability of CRMA on the UK road network.
KEYWORDS: Crumb, Rubber, Asphalt, UK, Noise, Environment.

1. Introduction
In Britain over 100,000 tyres are removed from cars, vans and trucks every day, resulting
in an annual disposal requirement for around 440,000 tonnes. Until recently the UK recovered
value from around 70% of these with the balance largely being disposed of to landfill, with
one cubic metre of landfill space required for every 17 whole tyres.
In the UK, in July 1999, this traditional, yet environmentally harmful, means of disposal
was addressed by the Landfill Directive. This prohibited the disposal of whole tyres in
new landfill sites by July 2003 and shredded tyres by July 2006. With increasing pressure
to determine an alternative means of scrap tyre disposal, one potential use crumb rubber
modified asphalt (CRMA).
Research to improve and enhance the performance of asphalt through the addition of
natural and synthetic rubber has been undertaken worldwide for some considerable time.
Historically, the objective has been to improve the asphalt’s physical characteristics through
improvements in its elasticity. Results of several demonstration projects have also indicated
increased fatigue resistance, a reduction in reflection cracking, improved skid resistance and
increased durability.
Despite apparent advantages, CRMA has received mixed reception across the world. The
United States, is currently the largest user of CRMA; by as early as 1991, Arizona had paved
in excess of 10% of its highway network with CRMA, whilst Texas had paved nearly 4,000
lane miles and recycled 3.5M tyres.
2. Aim.
This paper aims to examine the potential advantages and disadvantages of the use of
scrap tyre crumb rubber in UK road construction, concentrating upon its use in the friction or
wearing course layer.
3. Background.
A one-inch CRMA surface course recycles the equivalent of 375-2,200 scrap tyres per
lane mile. Assuming that all 203,000 miles of the UKs roads were just two lanes wide and
that just 5% were paved annually with one inch of CRMA this would require 13.8M tyres
(138,000 tonnes), almost twice the number of tyres disposed to landfill in 1999. (Calculated
conservatively at 600 tyres per lane mile.).
Used Tyre Working Group.

Manhole Adjusting Contractors Inc.
For existing landfill sites there was potential flexibility for continuing to landfill tyres beyond 2006.
Takallou & and Sainton.
Institute of Civil Engineers, Local Transport and Public Realm Survey.
6th annual report of the Used Tyre Working Group
Case Studies 661
There are many factors to consider when specifying the use of rubber in pavements
including cost, specifications, type of equipment to be used, expertise of the contractor and
subsequent recycling potential. The reported advantages and potential issues associated with
CRMA include:
Table 1. Advantages and potential issues surrounding CRMA.
Advantages Potential Issues

• Reduced pavement thickness. • Potentially high initial costs.
• Increased pavement life. • Examination of lifecycle economics.
• A reduction in reflection cracking. • An apparent lack of a specifications.
• Reduced traffic noise. • The requirement scrap tyre uniformity.
• Reduced maintenance costs. • Environmental concerns including the ease of
• Reduced pollution and increased recycling CRMA.
environmental quality. • The potential requirement for modifications to
asphalt plant or equipment.
4. Engineering Analysis
The methods of modifying asphalt, through the use of crumb rubber, are simplified into
two processes:
- The Wet Process. The Wet Process mixes the crumb rubber with the binder, allowing
time for their reaction prior to mixing with the aggregate. This results in a pavement
mixture known as Asphalt-Rubber.
- The Dry Process. The Dry Process simultaneous mixes the crumb rubber, the binder
and the aggregate. The range of materials produced using this method is collectively
known as Rubberised Asphalt.
Throughout, the paper will utilise the term crumb rubber modified asphalt (CRMA) to
include both Asphalt-Rubber and Rubberised Asphalt. In instances where information is
specific, to either Asphalt-Rubber or Rubberised Asphalt, then the material will be identified
as appropriate.
The crumb rubber utilised in CRMA is primarily derived from scrap tyres. Once collected
from the point of recovery, the tyres, with appropriate processing, will yield 4 to 5 Kg of
crumb rubber per tyre.
4.1.Collection
The detailed examination of the comprehensive collection and processing system, required
to ensure efficient and economical compliance with the requirement to recover 100% of
Emery.
discarded scrap tyres, is beyond the scope of this paper. However, initial examination reveals
that a quantum change is necessary to meet the additional demand associated with complete
recovery, primarily focussing upon greater recovery capacity and efficiency. The current cost
of transporting used tyres to the reprocessor comprises a significant element (as much as
half)10 of the overall tyre recovery charge. There is therefore potential for an increase in illegal
disposal as the Landfill Directive takes full effect.
One means of reducing illegal disposal is through increased accountability and the ability
to identify and track individual tyres from production to disposal. A proposal which would
undoubtedly incur significant administrative costs and resistance from garages, motorists and
producers alike. Conversely such a system may lead to a reduction in theft if the police were
also able to access information linking, cars, tyres and their owners.
Recycling may also be encouraged through the generation of an economically viable

situation; financial incentives offer greater potential for success than a system based upon
legislation and enforcement through accountability. Such a system is also more efficient as
compliance is increased as a result of the incentive and thus requires only limited policing,
rather than enforcement.
4.2 Tyre Shredding & CRMA Production
Shredding is an essential preliminary process, reducing bulk and minimising transportation

costs. Tyres, by design, are hardwearing and robust requiring a large amount of energy to
create crumb rubber. Cutting and shredding equipment is relatively expensive to maintain
due to the abrasiveness of the rubber and the presence of the steel cord and bead wires which
blunt the cutting blades. As a result, the main market for shredded tyres is in CRMA, or other
construction materials, where the presence of some metal residue is acceptable, albeit large
quantities may damage asphalt plant and equipment.11
The time required to disperse, blend or react the crumb rubber is dependant upon its
physical and chemical properties. Reaction time is also inversely proportional to temperature,
doubling with every 10℃ (18oF) decrease in temperature. The susceptibility of the finished
product to the physical and chemical composition of the crumb rubber and thus the method
of processing is significant. The finer the material the quicker it will react. For example,
cryogenic grinding results in a clean flat surface, leading to reduced reaction rates and lower
elastic recoveries, in comparison to the ambient grinding processes.
Variations in the rate and duration of the reaction affect the materiel characteristics.
Rubberised Asphalt applications have exhibited widely varying performances ranging from
acceptable to disastrous, the primary reason for which appears to be poor control in combining
the gradations of the aggregate and the crumb rubber. Product and thus performance
inconsistency is compounded by a lack of understanding and thus control over the volume
changes which occur as the crumb rubber swells during processing and handling.
10 Hallet.
11 Rubber Pavements Association, Crumb Rubber Modifier in Asphalt Pavement.
Case Studies 663
The Colas test section, using Colsoft material, laid in Surrey, UK, in 1999, had to be re-
laid prior to the commencement of the trial. Colas meanwhile, now recommend a maximum
transportation time of 30 minutes to prevent aggregate segregation and reaction with the
binder. Similarly the Florida Department of Transportation and Public Facilities (DoT) opted
not to develop State Specifications for this process.
Quality control is critical to the success of CRMA. Regardless of methodology, processes

must be operated and controlled as to afford the contractor, designer and client, confidence in
the all stages of the production process. This requires care and attention not only in the supply
and production of the crumb rubber but also during its subsequent incorporation with the
asphalt.
4.3 Placement of CRMA
4.3.1 Transportation
The transportation of CRMA may be undertaken using any truck typically used for
conventional asphalt. A critical factor however, is the adhesion of the material to the truck bed,
plant or equipment. In all cases the most effective means of preventing adhesion is through the
maintenance of the mix temperature.
The criticality of the compaction temperature increases with the rubber content. Thus it
is suggested that there may be a requirement for additional insulation during transportation.
Similarly, the traditional image of a fleet of sheeted asphalt trucks, waiting to discharge
into the paving machine, is one that is not compatible with CRMA, due to the increased
temperature susceptibility. The use of CRMA requires strict coordination, planning and
supervision at all stages of production and placement.
4.3.2 Placement
Information regarding the placement of CRMA is contradictory, depending upon the

origin of the source and further complicated by the absence of a nationally accepted standard
for CRMA, within the UK. The most recent guidance issued in the UK, is the document
Civil Engineering Applications of Tyres12. This states that Asphalt-Rubber should be heated
to between 149-190℃ before compaction and subsequently, laid at a temperature of at least
121℃ using standard paving equipment, with compaction completed as soon as possible.
Conversely, in the US, where Asphalt-Rubber has been an accepted pavement material for
some time, Arizona (DOT & PF) requires that the temperature of Asphalt-Rubber be at least
135℃ just prior to compaction, and be completed prior to the mix temperature reaching
104℃.
For the Dry mix it is commonly accepted that, during laying, the temperature of the
mixture should be at least 121℃ and a finishing roller must continue to compact the mixture
until it cools below 60℃ to prevent swelling due to the continued reaction between the
12 Hylands & Shulman.

bitumen and the rubber. Meanwhile, Road Research Laboratory (RRL) reports13 state that in
order to maintain good workability of the mix and to ensure adequate dispersion of the rubber
it is necessary to keep the mixing temperature of the modified materials at approximately 15℃
above those for traditional mixes.
4.3.3 Compaction
Compaction of CRMA requires additional control measures and careful supervision

if a quality surface is to be obtained. Due to the “sticky” nature of the modified asphalt,
compaction must be undertaken with either vibratory or steel-wheeled rollers. Pneumatic
compaction increases rutting, resulting in less effective compaction, as well as increased roller
“pick-up”. As such, the use of pneumatic rollers should be ignored at all times. The problem
of mix “pick-up’ maybe reduced through the addition of simple detergent to the roller water
spray systems.
The higher temperatures required for the inclusion of crumb rubber means that the rollers
are unable to begin rolling immediately behind the spreader, due to the potential for roller
“pick-up” and for pushing and shoving in the mat, at higher temperatures. Higher laying
temperatures are compounded by the cooling characteristics of the rubberised mat which, in
general, cools relatively quickly, to approximately 71-77℃, and then retains heat for longer
than traditional mixes14. As a result traffic must be kept off the surface for extended periods of
time due to the longer cooling time.
Density testing, during rolling, is usually performed using nuclear density meters.
However, on site values have been found to be lower than those subsequently obtained
from roadway core samples, a factor attributed to the presence of the rubber. Whilst such
an anomaly may be resolved through the use of a correction factor, the failure to develop a
national standard for the UK has hindered acceptance of this material.
4.4 Reluctance to Accept New ideas
Two separate agencies, or clients, control the road network within the UK. The Highways
Authority is responsible for the UK motorway and trunk road network, which comprises
approximately 5% of the road network, whilst individual local authorities are responsible for
the remaining 95%15.
The Specification for Highway Works and the Design Manual for Roads and Bridges
are the controlling documents for the use of materials in the UK motorway and trunk road
network. These documents include specifications for engineering, relating to the construction
and subsequent maintenance of roads, including the pavement surface. The use of tyres is
not included and as such any scheme utilising CRMA must seek authorisation on a scheme-
by-scheme basis with the Highways Agency requiring comprehensive information regarding
potential risk and benefits16.
13 Szatkowsk.
14 Amirkhanian.
15 ICE, Local Transport and Public Realm Survey.
16 Hylands & Shulman.
Case Studies 665
Meanwhile local authorities utilise their own specifications, which may permit the use
of alternative materials. Contractors and suppliers wishing to utilise alternative materials
are again forced to individually tailor their applications for each local authority. Such
an arrangement removes any potential savings or benefits of economy of scale and the
requirement for scheme-by-scheme approval only serves to deter many contractors from
pursuing alternative materials through increased administrative costs and uncertainty.
The UK’s performance on tyre disposal compares favourably with that of Europe.
However, there is someway to go towards matching the performance of the US. In 1998, the
UK recycled just 10.5% of its tyres whilst California recycled 26%, a figure which in 2001
had increased to 42%. It is suggested that the development of standards for the use of tyres in
civil engineering applications has contributed to this performance. In addition to regulating
the use of tyres, primarily to ensure fitness for purpose, guidance is provided on the testing of
physical properties as well as potential pollution aspects.
Thus, disjointed evolution, coupled with an over-riding lack of commitment or direction

within the UK has not assisted in the acceptance or development of alternative materials,
including CRMA as approved paving materials.
A major barrier to the widespread use of CRMA has been the increase in cost when
compared to conventional asphalt; traditionally attributed to the material being in its
developmental stage. Early Asphalt-Rubber binder had to be used within a few hours of
production, rendering potential users vulnerable to site delays, equipment breakdowns or the
vagaries of the weather. Subsequent development of a stable binder removed such financial
risk and allowed centralised production and storage. This increased product uniformity
and thus consistency of the binder qualities, which in turn aided development of consumer
confidence in the quality of the product.
Yet despite these developments in Asphalt-Rubber, the UK still appears to be concentrating

its efforts on Rubberised Asphalt. A material which, by the very absence of a reaction during
production has a very limited shelf life (less than an hour) between production and placement.
4.5 Performance of CRMA
Historically, CRMA research has aimed to utilise the beneficial properties of rubber to
improve the engineering characteristics of the asphalt. It follows that asphalt, with increased
elasticity, should show greater resistance to permanent deformation under the traffic loads.
Reflection cracking often occurs in normal asphalt wearing courses laid over existing
roads. The onset of reflection cracking can be delayed by laying a greater pavement thickness,
effectively strengthening the upper layers. Conversely, increasing the binder content of the
material reduces cracking, through increased pavement flexibility. However, this cannot be
beyond defined limits for a given geographical location, without the risk of deformation in hot
weather.
The increased viscosity of Asphalt-Rubber reduces the tendency to deform at high

temperatures, thus allowing the proportion of binder to be increased, to show an advantage

in crack prevention. Theoretically, the greater elasticity of Asphalt-Rubber, over traditional
asphalt, should increase this advantage further. Whilst Asphalt-Rubber pavements have shown
advantages as regards resistance to reflection cracking there is no discernable evidence as to
whether this is a result of the increase in binder content (achievable through the presence of
the rubber) or the contribution of the rubber to the overall material elasticity.
4.5.1 Deformation and Durability
Whilst several experimental road sections have been constructed in the UK over recent
decades, the resultant data was often inconclusive due to the incompleteness or inadequacy
of the tests. One such example was the examination of rubberised open-textured bitumen
macadam as a non-splash wearing course, conducted on the M40 at High Wycombe in 196717.
Expected to improve resistance to compaction, by traffic, of the open-textured surface, its
relative performance could not be assessed as no equivalent un-rubberised material was laid.
More recently, in 1999, Colas laid a Rubberised Asphalt test section in Surrey, however, there
is currently no indication as to the long-term performance of these sections, or the product,
Colsoft18.
The apparent lack of available information regarding CRMA experiments, in the UK,
places increased reliance upon experimentation and research conducted elsewhere. These
findings, primarily from the US, are applicable to construction in the UK, provided allowances
are made for any differences in traffic conditions and climate.
Results of several demonstration projects have indicated increased fatigue resistance,

retardation of reflective cracking, improved skid resistance and increased durability19. In 1983
the California DoT (CalTrans) compared equal thickness of asphalt-rubber and asphalt in an
experiment conducted on Route 395. In 1987, after four years of observations and testing, it
was decreed that all future projects involving Asphalt-Rubber would utilise thinner pavement
thickness. A two-to-one reduction of traditional asphalt v Asphalt-Rubber is now utilised
in most CalTrans projects, with the same reduction strategy being adopted by most local
agencies in California.
Later, in 1988, the Florida DoT commenced a comprehensive 10 year-long performance

evaluation of various CRMA mixes. Results from this period are shown below in Figure 1
to 620. It can be seen from the cracking and patching data in Figure 2 that all sections with
Asphalt-Rubber showed approximately 1 to 6% cracked areas, depending upon the amount
of rubber, whist traditional asphalt or Rubberised Asphalt returned 30% cracked areas.
Additionally, the cracking is relatively insignificant in sections containing 10 and 15% rubber,
therefore suggesting that the optimum rubber content lies within this range.
17 Thompson & Szatkowsk.

18 Harding.
19 Takallou & Sainton.
20 Choubane, Sholar et Al
Case Studies 667
Figure 1. Cracking and Patching versus Time on SR-120 Demonstration Project.
Figure 2. Cracking and Patching Versus Time on SR-16 Demonstration Project.
Figure 3. Rut depth Versus Time on I95-Southbound Demonstration Project.
Figure 4. Rut depth Versus Time on I95-Northbound Demonstration Project.

Figure 5. Rut depth Versus Time on SR-16 Demonstration Project.
Figure 6. Rut Depth Versus Time on SR-120 Demonstration Project.
Whilst crack data, as a performance indicator, shows clear advantages from the use of
Asphalt-Rubber, the benefits, when measuring rut depth, are less apparent for low rubber
contents. Figures 5 and 6 indicate that the benefits of Asphalt-Rubber are difficult to determine
when only small percentages of rubber content are utilised. However, as the rubber content
increases an improvement in performance is measurable, as shown at Figure 6, which
indicates the lowest rut depth at 17% Asphalt-Rubber whilst Rubberised Asphalt returned the
worst results.
In conclusion it is evident from the data collected during the Florida DoT evaluation that
all Asphalt-Rubber sections performed significantly better than the traditional and Rubberised
Asphalt test sections. It should be noted that research, by Takallou and Sainton, claims
Rubberised Asphalt mixes with laboratory fatigue lives ranging from 2 to 7 times longer than
conventional mixes, however, these have yet to be confirmed in the field.
As a result of the findings of the 10 year performance evaluation the Florida DoT initiated
the implementation of specifications requiring the use of crumb rubber (specified as ground
tyre rubber) in all asphalt surface mixes. Similar conclusions were drawn in May 1963, in
the UK, when inspection of the A15 at Yaxkley identified a 50% increase in pavement life
compared with normal materials. The Road Research Laboratory (RRL)21 concluded that
there was sufficient evidence to indicate that, the resistance to reflection cracking was almost
21 Thompson & Szatkowsk
Case Studies 669
entirely due to the properties of the modified asphalt, which almost entirely eliminated
cracking for the five years of observations. Despite these findings, after 45 years, CRMA is
still not a widely accepted materiel in the UK construction industry.
This early UK report also stated that Asphalt-Rubber appeared to show no advantage
on lightly trafficked roads22. On cursory examination this is a disappointing comment for
advocates of CRMA technology, however not unexpected. CRMA shows potential for
improving the performance of road pavements under vehicle loadings through utilisation
of the characteristics of the added rubber. It follows that if the stresses induced in the
pavement are negligible, then there will be limited advantage to be gained from improving
the pavements elasticity through the addition of rubber to the mix. In such scenarios the
pavements deterioration will be influenced more by the cumulative action of weathering and
ageing, than traffic loading.
4.6 Maintenance, Repair & Utility Work
It has already been identified that CRMA requires stricter control measures than for
traditional asphalt. It is therefore proposed that CRMA is less suitable for the small,
piecemeal, quantities often associated with the reinstatement of utility works. Thus, when
assessing the suitability of CRMA as a pavement material consideration must be given to the
potential impact of subsequent utilities work.
The use of dissimilar adjacent materials can lead to accentuated aging and subsequent
failure, at the material interface; where a relatively rigid pavement material (traditional
asphalt) is laid adjacent to a relatively compressible one (CRMA). CRMA may therefore be
more suited to motorways and main truck roads where minimal disturbance by future utility
work is anticipated. Conversely, CRMA is not unsuited to locations where frequent utility
work will be conducted, however additional consideration should be given to reinstatement.
It should also be noted that in the event of a road traffic incident (RTI) the characteristics
of Asphalt-Rubber differ from those of traditional asphalt. The increased viscosity, and
associated higher mixing and laying temperatures, increases the pavements resilience, to
heat damage in the event of a fire. This benefit is only marginal however as the temperature
difference required during production, 15℃, is negligible in comparison to the temperatures
experienced during a fire. The use of CRMA also increases the pavements susceptibility to
damage from fuel or chemical spillage, often associated with major RTIs. Solvents degrade
both the traditional binder and the rubber component of the pavement, increasing the volume
of material area requiring replacement in comparison to a traditional pavement.
4.7 Safety
The use of CRMA raises several potential safety issues which must be considered in its
evaluation. The potential increase in pavement service life offered by CRMA reduces the
requirement for roadworks to repair or refurbish the pavement. It has been shown that the
frequency of RTIs is up to 70% higher in roadworks than in normal flowing traffic. Thus, if
22 Szatkowsk
CRMA pavements offer an increased service life, the frequency of roadworks will be reduced,
in turn lowering the number of RTIs. The primary concern however is the pavements ability to
provide safe driving conditions.
4.7.1 Road User Safety
In general, a dry pavement surface provides good tyre-pavement friction whereas a

wet pavement reduces skid resistance. For speeds in excess of 40 mph, pavement macro-
texture assists in reducing hydroplaning, by allowing the water escape from the tyre-
pavement interface. Macro-texture testing indicates that CRMA pavements have better texture
characteristics than conventional mixes, primarily due to their improved waterproofing
characteristics23.
Figure 7. San Antonio Interstate Highway IH 35 before and after paving with Asphalt Rubber
(RPA News, Vol 7, No.4, Spring 2004).
Macro-texture may be described as a combination of stone porosity and the void ratio
of the pavement as a whole. Whilst these characteristics remain unchanged, the rubber
component within CRMA coats the stone, effectively waterproofing it, and thus increasing the
speed of dispersal of the water from the road. Thus, the use of CRMA generates better overall
surface conditions and reduces vehicle-stopping distances by allowing the surface to drain
more effectively.
Increased visibility, during adverse weather conditions, will theoretically reduce the
frequency of RTIs. However, poor visibility is a key factor in reducing vehicle speed
and increasing vehicle separation, as drivers compensate for the road conditions. Thus,
a disproportionate increase in visibility, in comparison to stopping characteristics, could
potentially be disadvantageous and therefore requires closer examination.
Examination of the accident statistics and climatic data for IH 35, San Antonia, Texas,24
23 Kuennen.
24 Rubber Pavements Association News, San Antonia Asphalt Rubber, 2004.
Case Studies 671
(shown in figure 7), shows that despite an increase of 34.6% in the number of wet days in the
year following re-surfacing, major accidents were reduced by over 43% under all conditions
and by more than 51% on wet days. Figures 8 and Figure 9, show the skid resistance data
obtained as part of the Florida DoT 10 year-long performance evaluation25 indicates that both
Asphalt-Rubber and traditional asphalt pavements provided similar friction performance.
Thus, on the basis of this information, there is no advantage or disadvantage from the use of
Asphalt-Rubber when considering the long-term skid resistance as a performance indicator.
Figure 8. Friction number versus time on I-95, Northbound, demonstration project.
Figure 9. Friction number versus time on I-95, Southbound, demonstration project.
4.7.2 Cold Weather Maintenance and Road Safety
Ice reduces friction, preventing vehicle tyres interacting with the road surface. Winter
route maintenance traditionally utilises a sand or salt mixture applied prior to and during the
formation of snow and ice on the road. The rheological properties of Rubberised Asphalt,
which increase its rutting resistance, by allowing temporary deflection under load, also serves
to reduce susceptibility to inclement weather. The increased elastic response of the material
reportedly causes ice formed on the pavement to break under transient vehicle loadings.
Takallou26, in detailed laboratory investigations and Canadian field trials considered

the effect of temperature upon Rubberised Asphalt, concluding that at –6 ℃ the rubber
25 Choubane, Sholar et Al.

26 Takallou,Hicks & Esch.
particles lose their elasticity and act as a weak aggregate. At higher temperatures however,
they act more as an elastic aggregate. Therefore, at temperatures required for ice formation,
the flexibility afforded by Rubberised Asphalt offers the potential to reduce ice formation,
increasing friction and thus potentially reduce the incidence of accidents due to inclement
road conditions.
CRMA contains Carbon Black, a component of the scrap tyre, the presence of which
results in the pavement retaining its darker appearance for a longer period of time. In addition
to generating the image of a freshly laid surface for longer and thus increasing public
perception of the performance of the local authority or highway agencies, the darker colour
has additional benefits. Both Minnesota and Alaska DOT & PF have recorded additional
de-icing benefits27, attributed to the darker colouring resulting in the absorption of more
solar energy than traditional asphalt. This serves to not only delay the onset of ice but also
causes the snow to melt faster. This benefit appears to be based upon anecdotal evidence
and whilst logical reasoning and the application of basics physics supports such a claim, the
author was unable to identify any scientific or numerical quantification of this characteristic.
Further research is therefore required to identify the impact of carbon black upon the thermal
properties of CRMA and whether this holds any potential road safety benefits.
The use of salts and chemicals during winter route maintenance programme creates
an aggressive environment, reducing the service life of road infrastructure such as bridge
bearings or other exposed metal elements. Additionally, sand, salt or chemicals will normally
only reduce stopping distances for a short period of time as they quickly become dispersed,
due to the action of passing vehicles. Frequent reapplication is therefore required, increasing
the aggressiveness of the immediate environment and the logistical burden of transporting and
distributing the de-icing material.
The continually varying road surface conditions, which result as the material is dispersed
and ice reforms, renders driving conditions more hazardous. This is compounded by the
findings of the Alaska DOT & PF who installed 12 experimental sections28 to analyse the
performance of Rubberised Asphalt29. Conducted under icy conditions with some roadway
sand occasionally present, test results indicated an average 25% reduction in stopping
distances. However, it should be noted that when tested during the summer, the Rubberised
Asphalt pavements, demonstrated lower friction numbers in comparison with the conventional
surfaces. CRMA can afford similar, if not improved, friction values when compared to its
standard counterparts, as shown earlier in Figure 8 and Figure 9. It is concluded that, had the
Rubberised Asphalt, exhibited the same summer friction values as conventional mixtures, then
the cold weather results would have been further improved, thereby strengthening the case for
the use of Rubberised Asphalt in regions susceptible to snow and ice conditions.
4.7.3 Constructor and Maintainer Safety
Whilst CRMA will not reduce or remove the dangers associated with construction, a
reduction in the requirement for road maintenance will reduce the exposure frequency of
27 Takallou, Hicks & Esch.

28 Totalling in excess of 34 lane miles constructed between 1979 and 1987.
29 Takallou, Use Of Rubber Modified Asphalt Pavements In Cold Regions.
Case Studies 673
construction workers to such potential hazards. It can also be seen that if the use of CRMA
allows for a reduction in pavement thickness, there will be a corresponding reduction in
construction activity. Thus, CRMA offers the potential to reduce the degree and frequency, of
exposure to the hazards of road construction and maintenance.
When considering any material, it is necessary to evaluate the properties and characteristics
of the material. This should include the information on the safety on the material and include
all aspects from its manufacture to disposal. One area to note is that the presence of crumb
rubber, can result in fumes and smoke at typical compaction temperatures. However, the
author was unable to find tests or statistics specifically relating to UK health and safety
legislation as regards exposure levels for emissions during asphalt production and placement.
Internationally, data is also limited in its availability and examination of Emery’s 30 findings
indicates that air emissions for both conventional and CRMA processes overlap and exhibit
wide variability. It is concluded that there is no discernable difference between the emissions,
rather that the variance is influenced by characteristics of the operation and maintenance
characteristics of each plant and rather than differences in the material.
5. Environmental Analysis
Traditional research has focused on the addition of rubber to improve the engineering
characteristics of asphalt pavements but, with increased environmental awareness backed
by European and UK directives and legislation, emphasis is now upon scrap tyre disposal.
This is recognised by the ICE which publicly31 supports the need to protect and enhance the
environment and to use resources in a way that does not disadvantage future generations; the
achievement of these objectives requires solutions which strike an informed balance in terms
of cost, benefits, sustainability and acceptability within the broader legislative framework, and
involves an evaluation of whole life costs.
The implementation of CRMA as the default material, for highways, offers the potential
for environmental savings: Thinner pavements, utilising the CalTrans findings, reduce the
materials required by 50%. Conversely the use of crumb rubber does require the rubber to
be processed, which has its own energy requirements. However, as a result of the landfill
directive the majority of tyres require processing irrespective of the disposal or recycling
regime implemented. Thus, a reduced pavement thickness continues to offer significant
environmental benefits.
CRMA must be evaluated from an environmental as well as financial and engineering

material performance perspective to determine the true magnitude of potential benefits. Due
consideration must also be given to the recovery and ease of recycling of the CRMA once its
useful life has expired. The possible long term contamination of asphalt through the addition
of rubber, without consideration of the environmental impact may only serve to exaggerate the
pollution problem.
CRMA, as a means of tyre disposal, must not be considered in isolation and whilst outside
30 Emery.
31 Institute of Civil Engineers, ICE Extra.
the scope of this report, a comparison of environmental and financial factors must be made
against other disposal methods.
5.1 Noise Reduction
Classed as environmental pollution, excessive noise in urban or residential areas is an

increasing nuisance as traffic volumes and vehicle speeds increase. It is thought that more than
half the homes in England and Wales are exposed to noise levels exceeding the World Health
Organisations recommended daytime level of 55dB32.
For automobiles, the vast majority of the noise is generated by the interaction of the tyres
and pavement whereas for heavier vehicles the engine and exhaust are the main sources.
CRMA obtains its noise-reducing properties from a combination of surface porosity and
ductility. As a result, whilst tyre noise is reduced, engine and exhaust noise are not appreciably
affected. Therefore, a pavement carrying primarily automobile traffic will exhibit a greater
decrease in traffic noise following paving with CRMA, than one which has a high percentage
of heavy vehicles.
Research has identified significant noise reduction through the use of open graded
pavement design and the inclusion of crumb rubber33. In Europe, Asphalt-Rubber surface
courses have been used to reduce noise levels and increasing skid resistance34, with results
indicating up to a 10dB reduction, equivalent to a 90% reduction in noise. Whilst the results
varied depending upon design criteria, in all instances they returned a noise reduction of no
less than 50%35. Table 2 details the findings of noise research in Sacramento36, the results
of which have been normalized for speed and traffic volume to isolate the noise-reducing
properties of the paving materials:
Table 2. Rubber Modified and Conventional Asphalt Noise Test Results, Sacramento County
Roadways.
Time elapsed after Change in noise

Ser Location Pavement type
paving level
(a) (b) (c)
(d) (e)
1. Alta Arden Expressway CRMA 1 Month -6 dB
2. 16 Months -5 dB
3. 6 Years -5 dB
4. Antelope Road CRMA 6 Months -4 dB
5. 5 Years -3 dB
6. Bond Road Conventional Asphalt 1 Month -2 dB
7. 4 Years 0 dB
32 Environment Agency, Noise from using tyres.

33 The phenomenon was first identified in Brussels in 1981 in an Asphalt-Rubber called “Drainasphalt”.
34 Manhole Adjusting Contractors Inc.
35 The Rubber Pavements Association, Noise Reduction with Asphalt-Rubber.
36 Sacramento County Public Works Agency.
Case Studies 675
Immediately after paving, traffic noise decreased along all three roadways, albeit with
some subsequent loss in noise reduction properties over time. Compaction of the asphalt
overlay reduces the porosity of the road surface, which accounts for some of the pavements
noise reduction properties. This compaction is deemed to be complete within one year of
paving. After this period, pavements paved with CRMA still exhibited good noise reduction,
whereas the noise reduction properties of the conventional asphalt have been lost.
Evaluation of noise reduction properties is complicated by several factors: pavement works

often alter roadway geometry, impacting upon both vehicle speed and noise emissions. Traffic
densities may change over time as a result of the changes in road geometry and network
configuration, as well as being subject to the vagaries of external factors not necessarily
factored into consideration. In normalisation of the figures, for comparison purposes, there is
potential for figures to be adjusted to suit the desired outcome and as such one must treat noise
reduction data with caution, unless the full facts surrounding the collection of data are known.
The use of CRMA can assist in the reduction of traffic noise, with obvious benefits for
applications in noise sensitive areas such as near schools, hospitals or residential areas.
Mitigating noise emissions can also reduce or even eliminate the requirement for expensive
and aesthetically instructive sound barriers.
5.2 A Potential Pollutant?
When assessing the suitability of a material, for inclusion in construction, consideration

must be given to its physical and chemical stability during service conditions and other
scenarios. Thus when a tyre, in whatever form (i.e. whole or crumbed) is used in a soil
environment, its total contents and the leachability of regulated elements, as well as
engineering properties must be taken into consideration.
Tyres reused on land or landfilled will leach, and research indicates that this is increased
through exposure to ultraviolet light and the effects of acidic rainwater 37. Westerberg
meanwhile concluded that tyre shreds, in unsaturated applications, have a negligible effect on
the surrounding surface or ground water quality. However, it should be noted that Westerberg
recommended that the free iron cord length be minimised, in order to reduce rust formation
and subsequent release. Additionally, that the tyre particles be washed and free from
contaminants. These recommendations, whilst predominately identified for the use of scrap
tyres as engineering fill, are equally applicable to their application in road construction if
potential pollution concerns are to be addressed.
Whilst no supporting documentary evidence was found, the author concludes that the
incorporation of rubber particles into asphalt would serve to reduce potential contamination to
levels below those indicated by Westerberg in his assessment of tyres in their raw state. The
interaction of the rubber with the asphalt binder will lead to increased stability, minimising
material or potential pollutant migration and as such the leaching of the component materials
of tyres is considered negligible. Further research or dissemination of research data in this
area is required to allay environmental fears.
37 Collins.
However, despite this apparent chemical stability, it should be noted that the presence
of organic solvents could result in the leaching of large quantities of hydrocarbons from the
rubber. Such a scenario could occur in the event of a RTI resulting in a fuel spillage. Again
this is based upon findings using tyre shreds, in their raw state. As such it is recommended that
additional research is required to examine CRMA’s susceptibility to solvents.
5.3 Recycling CRMA
Limited evidence is available regarding the environmental, economic and engineering

aspects of recycling CRMA. However, Takallou and Sainton, identify several demonstration
projects conducted in Ontario, Canada, which addressed the ability to recycle CRMA.
Their paper gives details of emission testing during the recycling process and indicates that
the CRMA proved to be recyclable. Similarly, Emery concluded that there was no reliable
evidence that asphalt pavements containing rubber could not be recycled to the same degree
as conventional pavements.
However, whilst Takallou and Sainton addressed the environmental concerns of the
recycling process, no indication was given as to the subsequent performance of the new
pavement once trafficked. This latter area is of particular interest and must be studied in detail
before specific conclusions may be drawn as to the overall ability to recycle the material.
The rubber content will undergo a degree of degradation during the recycling process and
studies are therefore required to determine the number of re-use cycles to which CRMA
may be exposed before a deterioration in pavement performance is identified. Thereafter,
consideration must be given to the subsequent disposal of CRMA once its useful life has
expired.
6. Economic Analysis
There are several areas for consideration when assessing the short and long term economic
viability of CRMA, including:
- Equipment. Specialist grinders or shredders required to produce crumb rubber, as well

as any potential effect upon traditional plant and equipment utilised.
- Serviceability and performance. Serviceability and performance must be balanced

against the whole life cost, including maintenance regimes and final materiel disposal
costs.
- Political. Any political involvement will significantly affect success, this may include
incentives to encourage reuse or recycling or legislation regarding the disposal of scrap
tyres.
- Material costs. In the UK, material costs, traditionally simple to calculate, have become
increasingly complex with the impact of the Aggregate Tax, Landfill Tax and even the
proposed Incineration Tax, as well as fluctuations in the price of crude oil must also
Case Studies 677
be evaluated. Additional factors include the pavements salvage value, the quantity of
recycled material incorporated and the ease of recycling CRMA itself.
This paper will now briefly examine some of the key financial issues, which will ultimately
determine the potential success, or financial viability of the use of CRMA, both as a disposal
medium and future construction material.
6.1 Governments Financial Influence for Change
Howlett38 identifies several tools available to government to bring about behavioural

change in an attempt to better manage waste. Financial instruments are commonly used
to penalise poor behaviour and reward desirable behaviour, amongst the more commonly
implemented ones are resource and disposal levies as well as disposal bans39. The UK
government has already gone someway towards driving behavioural change, through the use
of the Aggregate Tax, the Landfill Tax and the Landfill Directive.
The UK Landfill Directive came into force in July 1999, prohibiting the land filling of
whole tyres40 in new landfill sites by July 2003 and shredded tyres by July 2006. For existing
landfill sites there was potential flexibility for continuing to landfill tyres beyond 2006. An
earlier example of government legislative support was the Intermodal Surface Transportation
Efficiency Act of 1991, passed by the US Federal Government41. Section 1038 of the Act
prescribed that, starting in 1994, 5% of the total tonnage of federally funded roads were to be
CRMA42, with the required percentage doubling every year until reaching a maximum of 20%
in 1997.
However, this pertinent example of government intervention cannot be raised without

noting that, despite making the roads quieter, tests in the early 1990s found that the roads
were less durable than expected and as such the Federal Government did not pursue the
policy. However, prior to abandoning the Act, considerable investment and technological
development was made in the field of CRMA. As a result, the states of California, Arizona and
Florida now utilise a large proportion of their scrap tyres in asphalt roads and the pavements
have been found to be more durable than their traditional counter-parts, as well as being
quieter43.
Waste disposal levies, or landfill taxes, often seen as a financial tool, by which the
government can address environmental matters, present conflicting issues:
38 Howlett, Landfill Levy - Tax or Tactic?

39 A resource levy is a levy placed on raw materials to internalise the full costs associated with the extraction
of non-renewable resources; provides an incentive to minimise waste in production, increase\material reuse or seek
alternative materials.
40 For the purposes of the directive, the definition of tyres is taken to mean car and commercial vehicle tyres, which
form the greatest element of total used tyre arisings. The Landfill Directive includes all tyres save bicycle and tyres above
1.4 metres outside diameter, thus motorcycle and other (small plant and equipment) tyres are also included.
41 House of Commons, Select Committee on Environment, Transport and Regional Affairs, Minutes of Evidence,
12 December 2000.
42 A mix was characterised as being CRMA if it contained at least 1% of crumb rubber by weight.
43 House of Commons, Select Committee on Environment, Transport and Regional Affairs, Minutes of Evidence,
12 December 2000.
- Assessing the environmental cost of waste is difficult.
- The free market does not appear to adequately charge waste generators for the full
environmental impact of waste.
- The collection of a levy at the point of disposal significantly reduces the administrative
burden. However, whilst relatively easy to collect at the point of disposal it may serve
as a deterrent to legitimate disposal. Subsequent levy increases, to assist with the cost
of removing illegally dumped waste, would only serve to exacerbate the situation.
- A disposal tax, at the point of purchase, encourages consideration of alternative

products during the procurement process, although it increases the administrative
burden. An administrative and financial link must be established between the producer
and disposer, in order to ensure the correct distribution of monies. In comparison
to the “collection at point of disposal” system there is an increased requirement for
administrative and financial auditing due to the nature of the system and associated
transfer of responsibility/monies attributable to the waste.
6.1.1 Aggregate Tax
On the 1st April 2002, the British Government implemented the Aggregate Tax, applicable
to almost all stone leaving the quarry. In 2002 the headline rate was £1.60 per tonne (£2/T in
2009), the actual increase to the customer is far greater. Quarry by-products, and less popular
sizes, such as scalpings, fine sand and dust, are sold cheaply and it is not economically viable
to apply the full rate of tax to these products, as they often compete with exempt materials.
Therefore, the more popular sizes have to bear a higher proportion of the tax. Thus, in 2002,
the aggregate tax resulted in a mainstream aggregate price increase of between £2.00 & £2.75
per tonne44.
The impact of the aggregate tax is reflected in the price of both traditional asphalt and
CRMA. As component materials costs rises, the economic benefits afforded by a CRMA
thickness reduction strategy become increasingly significant.
6.1.2 Landfill Tax and the Landfill Tax Credit Scheme
The Landfill Tax is often described as the UK’s first environmental tax. Since its
introduction, in 1996, the rate has steadily risen. In 2008, the tax was £32 per tonne for active
wastes and £2.50 per tonne for inactive waste. There is clear indication that the Landfill Tax
“escalator” is set to continue at £8 per tonne per year albeit the rate for inactive waste will
remain extant.
The Landfill Tax Credit Scheme provides a convoluted and, it may be argued, ineffective
attempt at funding sustainable waste management. In March 2001, The Environment,
Transport and Regional Affairs Committee (ETRAC) in its report, Delivering Sustainable
Waste Management recommended that rather than attempt reform of the existing system
44 Durward.
Case Studies 679
that it be replaced in its entirety. The implementation of a new system consisting of a fund,
taking a percentage of the revenue from the Landfill Tax (and the proposed Incineration Tax)
would allow monies to be directed towards minimising, waste generation and increasing the
re-use and recycling of waste. The provision of such financial incentives for recycling and
reuse of materials, including scrap tyres in CRMA, would greatly enhance their financial
competitiveness in comparison with traditional materials.
6.1.3 Incineration Tax
Recent discussion45 on the incineration of tyres has focussed upon their high calorific
value, prompting debate as to whether tyre incineration is a means of waste disposal or merely
an alternative fuel source.
Under the Koyoto agreement, the UK government has a commitment to reduce emissions
of the greenhouse gas, carbon dioxide, by 25.6% of 1990 levels by 2010. The UTWG46
identified that this could be achieved through the increased use of carbon neutral fuels such as
tyres. In addition to burning whole tyres, a practice currently employed in areas such cement
kilns, recyclers have identified a means to reform and refine tyres and oil, from scrap cars, into
an oil which may then be burnt in existing power stations. However, the UK already burns
too much old oil and fails to comply with the EU’s Waste Oil Directive of 197547. As such,
increasing the number of tyres burnt for fuel would only compound this non-compliance.
6.2 Production and Placement Costs
6.2.1 Tyre Collection and Processing
UK companies operating in this sector effectively fall into two main groups. A handful of
companies, operating on a national basis, handle over half of those tyres currently recovered.
Tyre arisings are strategically brought together at a regional level, sorted as necessary and
subsequently distributed through reprocessors. Providing a management service between tyre
disposer and reprocessor these companies offer the standard of service required to facilitate
production of a quality sensitive material such as CRMA.
The second group is more fragmented, operating at a local level. The task of economically
pulling together widely dispersed, isolated, low level arisings is a significant one especially if
these arisings are to be matched to a particular recovery facility.
The cost of collection alone, for a truck tyre, has been estimated as being £3.25 each.
For reuse, rather than recycling, there are also additional costs associated with inspections,
documentation and retreading etc. In July 2003, the cost associated with the disposal and
recovery of used tyres was in the region of £100 - £140 per tonne of tyres48. This covers the
retailer’s costs, including the provision of storage, transport costs in getting used tyres to
disposal and recovery facilities and the gate fee raised by such facilities for accepting used
45 NCE, A burning Issue.
46 UTWG 6 report
47 M Glaskin, Crunch Time for the Scrap Industry.
48 UTWG 6 report
tyres. Gate fees charged by disposal and recovery facilities vary considerably by process and
location, and may be subject to market variations. Because of this fluidity, the paper does not
seek to give an indicative spread of costs here. However, as a rule of thumb, landfill costs, for
disposal or engineering, are significantly cheaper than the other options available.
6.2.2 Plant & Equipment Costs
From an equipment perspective it can be determined that there is generally no significant

environmental or production modification involved and the method of crumb rubber addition
will not add to the costs. Opinion on the impact upon plant production is varied but it is
generally accepted that CRMA production would result in some loss. This is attributable
to increased operating temperatures required for CRMA production and the requirement to
prevent contamination of subsequent material. The temperature sensitivity of CRMA during
compaction will also impact upon production, and thus cost, as the rate of production must be
better tailored to the actual laying rate.
6.2.3 Life Cycle Cost Analysis
To evaluate alternative pavement materials, it is necessary to consider the total cost over
its service life. The most effective method for measuring the cost-effectiveness of alternative
designs is Life Cycle Cost Analysis49 (LCCA), as shown schematically at Figure 10. The
primary purpose of LCCA is to quantify the long-term economic implications of initial
pavement decisions.
Figure 10. The Principle of Life Cycle Cost Analysis for multiple pavements.
A LCCA should include, where possible, all costs including initial construction,
rehabilitation, and maintenance costs, incurred directly by the agency responsible for the
pavement throughout the life of the pavement. Meanwhile consideration must also be given
to the indirect costs user including travel time delay costs, vehicle operating costs, accident
costs, as well as an assessment of discomfort costs both for road users and those living or
working nearby.
Development of a comprehensive LCCA should include consideration of the following

factors:
49 Emery
Case Studies 681
- Pavement recycling. Consideration must be given to the costs and practicality of

pavement recycling and its salvage value. It has been identified that further research
is required into the number of times that a CRMA pavement may be recycled whilst
retaining the beneficial properties within the rubber component.
- Winter route maintenance. Any potential benefits in winter route maintenance relating
to the thermal properties of CRMA or the reduction in ice formation due to pavement
movement under transient vehicle loads should be examined. Secondary effects,
resulting from a reduction in de-icing chemicals, may also include prolonging the
service life of elements such as bridge bearings and road furniture.
- Road user costs. Road user costs are related to the efficiency of the road network.
This must be examined at a variety of levels, from the pavements effect upon vehicle
efficiency through to the impact of roadworks for pavement repair or refurbishment:
- An increase in service life will reduce the frequency of roadworks and thus reduce
disturbance to vehicular traffic and impact upon the economy.
- Travel time costs are closely related to vehicle speed. Jong-Suk Jung 50 shows that
prolonged CRMA life results in a uniform vehicle speed throughout the pavements
life whilst the pavement speed for a traditional asphalt pavement slowly declines
as pavement roughness increases. Overall Jong-Suk Jung identifies that CRMA
pavements show substantial costs savings over a period of 25 years. Whilst there
was no discernable difference in costs in the first 5 to 10 years, thereafter the higher
maintenance costs for the conventional pavement began to make significant impact
upon the overall analysis.
6.3. Noise Reduction
American research indicates that the potential cost savings from the use of open graded
design incorporating rubberised asphalt, in the field of noise reduction at least, are in excess of
10 times the cost of constructing sound reduction structures51. Noise wall cost estimates, from
Plastral, for February 2004 indicate that, depending upon material employed, the average cost
of noise wall construction is as follows:
Assume 2.5m high noise wall, located on both sides of the carriageway.
Average noise wall cost per m2 €200/m2.
Average noise wall costs €1000 /Lm of carriageway
Noise Wall Construction £666 / Lm or £1,072,896 / mile.
It is readily apparent that significant cost savings are available through the utilisation
of CRMA’s noise reduction properties. It would be idealistic to believe that CRMA could
prevent the construction of a noise wall structure altogether, especially when considering the
50 Jong-Suk Jung, Kaloush & Way, Life Cycle Cost Analysis: Conventional versus Asphalt-Rubber Pavements.
51 Rubber Pavements Association – Noise Reduction with Asphalt Rubber, 1999.
prolonged analysis period utilised. More realistically it may be assumed that the use of CRMA
will reduce the existing noise levels. However, as traffic volumes increase, there will still be
a requirement for such noise abatement measures. Utilising the figures above for noise wall
construction costs, the potential economic benefits of delaying noise wall construction are
shown graphically in Figure 11: assuming an interest rate of 4%, a delay of just 3 years in
noise wall construction will result in present day savings of £33K. Higher interest rates or a
greater postponement in construction serves to further increase the available savings.
Figure11. Cost savings through postponement of noise wall construction per £1M.
Thus, irrespective of its other properties, considerable financial savings are available from
the use of CRMA in circumstances that would otherwise require noise wall construction. Thus
CRMA should be considered in sensitive areas such as outside schools, hospitals etc where the
aesthetic or practical considerations of noise abatement structures renders them unsuitable.
6.4 Tyre Subsidies
The additional costs associated with the incorporation of scrap tyres into asphalt production
may be simplified as follows:
Weight of passenger car tyre = 9 Kg

Disposal, recovery & processing = £ 100 / T
Number of passenger car tyres / T = 111 Tyres/T
Additional cost per tyre = Approx £1 /tyre
The author has been unable to establish exact figures for the cost of CRMA in the UK,
however Colas 52 indicate that the cost of a CRMA mix is typically 25% more than that of
conventional asphalt. Thus:
Assume cost of traditional asphalt = £ 40/T

Cost escalation for utilisation of CRMA = 25%
Cost of CRMA =£ 50/T
52 Harding.
Case Studies 683
For the purposes of this paper it is assumed that CRMA utilises 3.6 scrap tyres per Tonne.
Thus the £ 10/T cost escalation in the price of CRMA above traditional asphalt equates to £2.77
per tyre of which £1 is attributable to collection and processing.
If such a cost escalation were to be offset through the use of a disposal levy then it may be
assumed that the administration of such a charge would itself incur additional costs, associated
with the collection and distribution of such monies. For the purpose of this analysis it is
assumed that this would equate to an increase of the equivalent price of CRMA from £2.77 to
£3.00 per tyre.
The typical purchase price for an average car tyre (excluding high performance tyres etc)
is £40.00. It is proposed that a disposal charge of £ 3.00 (7.5%), against such a purchase
price, is not unreasonable for a product which is, for many, a relatively infrequent purchase.
Such a levy, applied at the point of purchase off-sets the additional costs associated with the
collection and processing of scrap tyres and encourages the use of retreaded tyres. Whilst
the figure of £3.00 has been calculated in the context of CRMA, the monies raised, if such
a scheme were to be introduced, would not be exclusive to the subsidisation of CRMA but
utilised to subsidise any disposal route.
6.5 Economic Comparison
In the simplified assessment of costs given above there has been no consideration of the
potential benefits from utilising CRMA. These benefits have significant impact upon any
LCCA. In essence however there are three possible options:
6.5.1 Ignore the potential benefits of CRMA. In this scenario the CRMA is being utilised
primarily as a scrap tyre disposal medium. It has been shown earlier that only a relatively
small disposal levy(£3) is required per tyre in order to off-set the additional costs of CRMA
in comparison with traditional asphalt. Such a levy negates any additional cost, assuming the
subsequent performance of the two materials is unaltered.
6.5.2 Implement the CalTrans reduction strategy. The CalTrans reduction strategy allows a
reduction in pavement thickness whilst utilising the same pavement service life as traditional
asphalt. Whilst the unit cost (per CuM) of conventional asphalt is less than that for CRMA,
the total cost is more due to the difference in the pavement thickness. Thus, in line with the
Caltrans reduction principle, even in circumstances where CRMA is twice as expensive as
conventional asphalt the construction costs would be the same. Whilst CRMA offers reduced
cracking and increased durability.
6.5.3 Utilise the same pavement thickness as traditional asphalt. The CalTrans strategy
utilises a reduction strategy to reduce the pavement thickness whilst returning a similar
life expectancy as for a traditional pavement. It may be assumed that rather than utilise the
beneficial properties of CRMA to reduce the pavement thickness, application of CRMA to the
same depth as traditional asphalt would serve to prolong the pavements service life and reduce
maintenance costs during this period.
7. Conclusions and Recommendations
The existing UK scrap tyre collection and re-processing framework requires consolidation
and improvement to effectively capture 100% of the UK’s tyres. Such a target will only be
achieved through an economically viable commercial framework utilising financial incentives.
The proposed solution is a disposal levy of approximately £3.00 per tyre. The alternative is
to generate a demand for scrap tyres, through their inclusion in an alternative material such
as CRMA. Without one or both of these it is suggested that there will be an increase in illegal
disposal.
With strict quality control measures, CRMA offers significant improvements over
traditional asphalt. However, there is a clear performance between Asphalt-Rubber and
Rubberised Asphalt. The performance of Rubberised Asphalt has been found to extremely
variable, it was this that led the various State transport departments within the US to
concentrate their research and development upon Asphalt-Rubber.
Asphalt-Rubber returns similar friction values to traditional asphalt whilst showing

an increased service life and a reduced maintenance requirement for the same pavement
thickness asphalt. An alternative to increasing the pavements service life is to reduce the
pavement thickness. The optimum rubber content was found to be in the region of 15-17%
when utilising cracking data or rutting depth as a measure of performance in comparison to
traditional asphalt.
The flexural properties of Asphalt-Rubber have been shown to increase safety through a
reduction ice formation. Further research is required to support the, as yet anecdotal, claims
regarding the influence of Carbon Black. Meanwhile Asphalt-Rubber pavements have been
shown to lower accident frequency by reducing the water present upon the road surface.
However the author would suggest that as a result of the higher vehicle speeds in the UK
compared to the US where the information was obtained there is a requirement to ascertain
the relationship between improved visibility, vehicle speed and the friction characteristics of
CRMA.
Serious consideration must also be given to the environmental impact of CRMA. Research
is required to support the author’s assumptions that inclusion of crumb rubber within asphalt
will serve to improve the material’s stability. Additionally, detailed research is required to
ascertain the degradation of the rubber in CRMA during the pavements service life and
subsequent recycling process. In the event that degradation is identified, a means of extracting
or recycling the rubber components must be identified to avoid the long-term contamination
of the UK road network. As a scrap tyre disposal medium, CRMA must be compared to the
spectrum of disposal methods as well as traditional asphalt. When examined against traditional
asphalt, the increase in engineering performance offsets the increased unit cost of the material.
The utilisation of CRMA to reduce current noise levels and effectively delay noise wall
construction offers the potential for significant financial savings. CRMA should also be
considered over traditional asphalt for use in sensitive locations such as near hospitals, schools
and residential areas.
Case Studies 685
The quality of Britain’s roads is in decline, demand on the road network continues to
rise and the issue of maintaining Britain’s roads becomes increasingly important. In 2000,
the Government published its 10 year transport plan containing commitments to halt the
deterioration of local roads by 2004 and eliminate the maintenance backlog by 2010. In
reality, every year since its inception in 1996, the ICE Annual Survey has shown an increase
in the road maintenance backlog.
The use of CRMA offers the opportunity to increase pavement service life and reduce
pavement maintenance costs: providing a disposal route for scrap tyres whilst potentially
assisting to improve the current state of the nations road network. As early as May 1963,
inspection of the A15, Yaxkley, UK, identified a 50% increase in pavement life compared with
normal materials. The Road Research Laboratory (RRL)53 concluded that there was sufficient
evidence to indicate that, the resistance to reflection cracking was almost entirely due to the
properties of the modified asphalt, which almost entirely eliminated cracking for the five years
of observations. Despite these findings, after 45 years, CRMA is still not a widely accepted
materiel in the UK construction industry.
It is recommended that support is required at a national level to develop guidelines,

direction and technical support for the use of CRMA and other alternative materials within
the construction industry. As such this requirement reiterate the findings of The Environment,
Transport and Regional Affairs Committee in its 2001 report “Delivering Sustainable Waste
Management” which stated that cited those in authority are:
“guilty of thinking without imagination and planning without ambition, of finding problems
instead of solutions and aiming for short-term goals without a vision of the system of resource
use and waste management which we should be striving for”.
REFERENCES
Amirkhanian S.N., Utilization of Crumb Rubber In Asphaltic Concrete Mixtures – South

Carolina’s Experience, Clemson University, USA.
Choubane B., Sholar G.A., Musselman J.A., and Page G.C., Florida Department of
Transportation, State Materials Office. Ten-year Performance Evaluation of Asphalt
Rubber Surface Mixes. Transport Research Record 1681, Paper No 99-0177
Choubane B., Sholar G.A., Musselman J.A., and Page G.C., State of Florida, State Materials
Office. Long Term Performance Evaluation of Asphalt-Rubber Surface Mixes. Research
Report FL/DOT/SMO/98-431. November 1998
Collins, K.J. et al “Scrap tyres for Marine Construction: Environmental impact”. Proceedings
of the International Symposium on Recycling and Reuse of Used Tyres, University of
Dundee, 19-21 March 2001.
53 Thompson & Szatkowsk

Cooper. Proceedings of the International Symposium on Recycling and Reuse of Used Tyres,
University of Dundee, 19-21 March 2001.
Durward R, Director British Aggregates Association, Aggregate Tax - an assessment by the

British Aggregates Association, www.axmark.co.uk/baa/aggtax10.htm
Emery J. “Evaluation of Crumb Rubber-Modified Asphalt Demonstration Projects”,

Proceedings of the 39th Annual Conference of the Canadian Asphalt Association, Nov
1994.
Environment Agency. Environment Agency Report on Tyres; Noise from using tyres. www.
environment.agency.gov.uk/commondata/105385/ea_tyres_report.pdf
Glaskin, M. “Crunch time for the scrap industry”, The Engineer, 19 March 2004.
Hallet Proceedings of the International Symposium on Recycling and Reuse of Used Tyres,
University of Dundee, 19-21 March 2001.
Harding. Telephone Conversation K Harding, Colas and R Hewson, 1 June 2004.
House of Commons, Environment, Transport and Regional Affairs Committee, Fifth Report,
Delivering Sustainable Waste Management, March 2001.
House of Commons Select Committee on Environment, Transport and Regional Affairs,

Minutes of Evidence, Examination of Witnesses, 12 December 2000.
Howlett P.F, “Landfill Levy - Tax or Tactic?” Enviro 2002 Convention & Exhibition,
Melbourne, 7-12 April 2002.
Hylands K.N. and Dr. Shulman V, Civil Engineering Applications of tyres, Viridis Report
VR5.
Institute of Civil Engineers, Local Transport and Public Realm Survey, 28 August 2003.
Institute of Civil Engineers, ICE Extra Website, Sustainability. www.ice.org.uk/about_ice/

aboutice_sustainability.asp.
Jong-Suk Jung, Kaloush K. & Way G., Arizona Department of Transportation. Life Cycle
Cost Analysis: Conventional versus Asphalt-Rubber Pavements a paper submitted to the
Rubber Pavements Association, August 2002.
Kuennen T, “Creating Friction Where Rubber Meets the Road”. Better Roads, October 2003.
Case Studies 687
Manhole Adjusting Contractors Inc Crumbling Roadways, “UFOs”, and Asphalt-Rubber, a

paper detailing the “Unidentified Favourable Observations (UFOs) of the Asphalt-Rubber
Program, 1991.
Nebraska Department of Environmental Quality, Environmental Update, “Innovative Project

recycles Tires, Improves Road Surface”, Winter 2002/2004.
New Civil Engineer, “A Burning Issue”, 24 June 2004.
Noise Management, “Tyres cut traffic impact”, March 2000, Pg 2.
Rubber Pavements Association. Publication: Noise Reduction with Asphalt-Rubber 1999.
Rubber Pavements Association. Publication: Crumb Rubber Modifier in Asphalt Pavement,

Chp 4.
Rubber Pavements Association News, San Antonia Asphalt Rubber, Permeable Friction
Course Mission Accomplished, by Doug Carlson.Vol 7, No.4, Spring 2004.
Smith J, Asphalt-Rubber – “Open Graded Friction course”. Proceedings of the International

Symposium on Recycling and Reuse of Used Tyres, University of Dundee, 19-21 March
2001.
Takallou H.B., Hicks R.G. & Esch D.C. Use Of Rubber Modified Asphalt Pavements In Cold
Regions, April 30 1987.
Takallou H.B. and Sainton A, Advances in Technology of Asphalt Paving Materials Containing
Used Tyre Rubber. Transport Research Record 1339, 1992 PP23-9.
Thompson P.D. and Szatkowsk W.S., Full-scale road experiments using rubberised surfacing
materials by Road Research Laboratory Ministry of Transport, Report LR 370, 1971.
Sacramento County Department of Environmental Review and Assessment. Effectiveness

of Rubberized Asphalt in Reducing Traffic Noise. www.rubberpavements.org/library/
sacramento_noise_study/sacnoise.html
Szatkowsk W.S., Resistance to Cracking of Rubberized Asphalt: Full-scale experiment on

Trunk Road A6 in Leicestershire, by Road Research Laboratory, Ministry of Transport,
Report LR 308, 1970.
Used Tyre Working Group, 6 Annual Report, Used Tyre Working Group Website, www.
tyredisposal.co.uk
Westerberg. Proceedings of the International Symposium on Recycling and Reuse of Used

Tyres, University of Dundee, 19-21 March 2001.
BIBLIOGRAPHY
Asphalt Rubber Technology Service, Clemson University, South Carolina, Benefits of

Rubberised Asphalt. www.ces.clemson.edu/arts/benefitsofRA.html
Asphalt Rubber Technology Service, Clemson University, South Carolina, Asphalt Rubber
Myths. www.ces.clemson.edu/arts/myths.html
Barlaz, Eleazer & Whittle, Potential to Use Waste Tires as Supplemental Fuel in Pulp and
Paper Mill Boilers, Cement Kilns and in Road Pavement, North Carolina State University
Department of Civil Engineering, January 1993. Published in Waste Management &
Research Journal Vol 11, 1993 PP 463-480.
Budiman, A.F.S “Prospects of the world rubber industry in the new decade”. International
Rubber Conference, Bogor, 12-14 September 2000.
Environment Agency, Tyre Watch Bulletin, June 2003.
Environment Agency, Tyre Watch Programme, Tyres Landfill Ban and Associated Questions.
www.environment-agency.gov.uk
Environment Agency, Landfill Directive Regulatory Guidance Note 17
Environment Council, Statutory Incentives www.the-environment-council.org.uk
Enviros, Potential Market for Recovered Tyres June 2002. Research carried out by ENVIROS
and Funded by the Department of Trade & Industry. www.londonremade.com
European Tyre Recycling Association, The Used tyre Mountain. www.etra.eu.com
Glover, Davidson, Bullin et al, A comprehensive Laboratory and Field Study of High-Cure
Crumb-Rubber Modified Asphalt Materials, Texas Department of Transportation, Research
and Technology Implementation Office, Report No 1460-1 Resubmitted Nov 2000.
Hicks R.G & Epps J.A. Dept of Civil Engineering Oregon State University Professor Of Civil
Engineering University Of Nevada, Reno. Life Cycle Cost Analysis of Asphalt-Ruber (Sic)
Paving Materials.
Jorgenson L., Asphalt-Rubber Pavement Construction, Public Works, January 2003 pp30-31
Lougheed, T.J. & Papagiannakis, A.T. “Viscosity Characteristics of Rubber-Modified

Asphalts.” Journal of Materials in Civil Engineering, August 1996
Case Studies 689
Lynch D.F. & Horthwood R.P., The Waste Tire Paving Demonstration Project at Thamseville,
Ontario Transport Association Conference, Cananda, September 1991.
Mulder E.A., & Whiteoak D., “An objective assessment of the in-service performance of TR-
Modified bitumen for road applications”. Proceedings of the 16th Australian Road Research
Board Conference, Perth, W Australia, November 1992.
Mullins, L “natural rubber in asphalt pavements NR Technology”, 1971, Part 3. Proceedings

of the International Symposium on Recycling and Reuse of Used Tyres, University of
Dundee, 19-21 March 2001.
Noise Management, March 2000, Tyres cut traffic Impact.
Satanovsky A., Jaeckel J.R., & Kuemmel D.A., “Tire/Pavement Noise Analysis”. Proceedings
of the 1998 National Conference on Noise Control Engineering, Michigan, April 1998
Sharma J.R., “Reclamation of Rubber Crumb from Army’s Discarded Tyres”. Paper presented
at ANTEC 2000, Plastics Recycling, May 7-11 2000, Orlando.
Summers, C.J. “Rubberised Hot Rolled Asphalt Wearing Course.” The Idiots Guide to
Highways Maintenance, www.highwaysmaintenance.com/rubtext
Way G.B., Flagstaff I-40 Asphalt Rubber Overlay Project, Nine Years of Success. Arizona
Department of Transportation, Transportation Research Record 1723, Paper No 00-1055.
Chapter 5
Function, Environemental and

other Aspects
Functional, Environmental and Other Aspects 693
Noise Characteristics and Field

Performance of Five Different Wearing
Courses in Arizona
Douglas D. Carlson* – George B. Way* – Ali Zareh** –

Kamil E. Kaloush*** – Krishna Prapoorna Biligiri***
* Rubber Pavements Association, Arizona, USA

dougc@rubberpavements.org
wayouta@cox.net
** Arizona Department of Transportation, Arizona, USA

azareh@azdot.gov
*** Department of Civil, Environmental and Sustainable Engineering, Arizona State

University, Tempe, AZ 85287-5306, USA
Kamil.Kaloush@asu.edu
Krishna.Biligiri@asu.edu
ABSTRACT: In 1999, the Arizona Department of Transportation placed five different asphalt
concrete pavement wearing courses as test sections on Interstate I-10, a highly trafficked
Arizona highway. The wearing courses consisted of: Asphalt Rubber Open Graded Friction
Course (AR-ACFC), Standard Open Graded Friction Course (ACFC), Polymer Modified
Open Graded Friction Course (P-ACFC), Permeable European Mixture (PEM), and Stone
Matrix Asphalt (SMA). A continuous review of these sections over the years has shown that
the pavement sections experienced different amount of cracking and wear after eight years of
service. Most importantly, AR-ACFC pavement sections experienced the least cracking and
wear after eight years of service, while the other test sections showed considerable cracking
and wear. On-Board Sound Intensity (OBSI) noise measurements were conducted in 2002 and
again in 2008 using sound Intensity equipment. The results indicated that the least noise was
observed for the AR-ACFC mixtures. In addition, 30 core samples from these test sections
were obtained to evaluate their dampening (impedance) properties in the laboratory using
an Ultrasonic Pulse Velocity (UPV) testing method. The results of these noise measurements
and laboratory testing are discussed along with the degree of surface deterioration of each
pavement test section in the field.
KEYWORDS: Asphalt rubber Open Graded Friction Course (AR-ACFC), Wearing Course,
Ultrasonic Pulse Velocity Test, Ultrasonic Pulse Time, Effective Flow Resistivity, Air Voids,
Tire / Pavement Noise, Pavement Surface Deterioration.
1. Introduction
Globally, road traffic noise has become a major problem in several metropolitan areas.
Automotive technology advancement has decreased the average noise per automobile by
about six decibels (dB) in the last 25 years. One way to reduce highway noise affecting the
adjacent residential areas is by constructing sound barriers (walls) along the city freeways.
Typical noise barrier walls cost around 1.25 million US Dollars per kilometer (Gibbs et al,
2005; Road Noise, 2004). The public has brought up so much attention to the noise problem
that transportation agencies, academic institutions and private entities are looking for
innovative ways to address the problem. In Europe and the United States, several successful
highway projects have been reported with the use of open graded or porous friction course
pavements. The success has been mainly contributed to the high air voids content of these
asphalt mixtures (Lerch et al, 2003). However, problems like clogging of pavement air voids
were reported to possibly degrade the dampening acoustic effects of these porous pavements
(Losa et al, 2003).
In Arizona, the Asphalt Rubber Asphalt Concrete Friction Course (AR-ACFC or sometimes
referred to as ARFC for brevity) mixtures have become the number one public demand item.
Residents have been asking their transportation officials to place asphalt rubber (AR) overlays
on highways nearby their neighborhoods because of the great noise reduction they provide.
In fact, legislation has been passed to overlay 185 kilometers of the freeways and highways
in the Phoenix metropolitan area with AR mixtures because of the significant noise reduction
produced by these mixes (MAG, 2005). In addition, research and field experience with these
mixes have shown that AR mixes have great resistance to rutting, fatigue and thermal cracking
(Way, 2003).
In this direction, Arizona Department of Transportation (ADOT) placed five asphalt

concrete pavement wearing courses as test sections on a highly trafficked Arizona highway
in 1999. A continuous review of these sections over the years has shown that the ARFC
pavement sections experienced the least cracking and wear after eight years of service, while
the other test sections showed considerable cracking and wear.
2. Objectives
The objective of this study was to evaluate the tire / pavement noise characteristics of five
different pavement wearing courses placed as test sections in the State of Arizona, United
States, by means of field noise measurements. In addition to the field noise measurements, 30
core samples from these test sections were obtained to evaluate their dampening (impedance)
properties using an Ultrasonic Pulse Velocity (UPV) testing method in the laboratory.
3. Wearing Course Experiment
The Arizona Department of Transportation (ADOT) conducted a preventative maintenance

pavement preservation experiment on Interstate – 10 (I-10) in Arizona during the year 1999
(Scofield, 2000). As part of this experiment, 32 test sections, with replicate cells, were
constructed constituting five asphalt concrete pavement wearing courses types. The location
of these test sections on the I-10 were randomly sequenced as shown in Figure 1. Figure
2 presents a layout of the I-10 test sections (starting at section number 99-0 and ending at
section number 99-31). The sections are marked East bound from Milepost 186.2 to 195.3.
The Annual Daily Traffic (ADT) is about 60,000 with 25% trucks. The total Equivalent Single
Axle Load (ESALs) is about 26 Million. Table 1 shows the present pavement condition for
each mile post (Zareh, 2008). The table includes recently measured pavement ride quality,
friction numbers, rutting and percent cracking at each Mile Post. ADOT utilizes a MU meter
or Side Force testing device to measure pavement surface friction. Measurements are reported
as a skid number, that is, the measured value of friction times 100. In Arizona, the intervention
level for friction reported for interstate, primary and secondary roadways is 34 (Mu Meter).
The five different pavement types under the study were as follows.
• Permeable European Mixture (PEM) as named per the project specifications
• Stone Matrix Asphalt (SMA)
• Asphalt Rubber Open Graded Friction Course (AR-ACFC)
• Polymer Modified Open Graded Friction Course (P-ACFC)
• ADOT’s Standard Open Graded Friction Course (ACFC)
Table 2 provides the aggregate size gradation, asphalt cement content and type of
Performance Grade used for the types of wearing courses used in the study. The highest
asphalt cement content used was for the AR-ACFC mixture (~9.2%). The type of polymer
used in the P-ACFC mix was either Styrene-Butadiene (SB) or Styrene-Butadiene-Styrene
(SBS). The thickness of the wearing courses were 19 mm (¾”) except for the Permeable
European Mix whose surface thickness was 32 mm (1 ¼”). Each wearing course was
supported by three structural layers beneath it. The pavement layers were designed for a 12 to
15 year service life.
Figure 1. Location of I-10 Test Sections, Arizona,

United States
Figure 2. Schematic of the I-10 Test Sections, Arizona, United States
Table 1. Present Pavement Condition of I-10 Wearing Course Experiment Test Sections by
Mile Post (Zareh, 2008)
January June 2007 January March 2007

East Bound Mile
2008 Ride, Friction 2008 Rutting Cracking
Mile Post No. Section
IRI (m/km) (Number) (mm) (%)
MP 186 186-187 1.09 50 3.8 3
MP 187 187-188 1.15 47 2.3 3
MP 188 188-189 1.95 51 5.6 6
MP 189 189-190 1.28 49 3.8 7
MP 190 190-191 0.96 51 3.3 2
MP 191 191-192 1.47 51 4.6 5
MP 192 192-193 1.39 54 4.3 5
MP 193 193-194 0.99 49 4.1 5
MP 194 194-195 1.01 46 5.6 4
Table 2. Binder and Mixture Properties of I-10 Wearing Course Experiment Test Sections
Size AR-ACFC ACFC P-ACFC PEM SMA

¾ 100 100 100 100 100
½ 100 100 100 80-90 100
3/8 100 100 100 35-60 70-90
No. 4 30-45 35-55 35-55 10-25 30-50
No. 8 4-8 9-14 9-14 5-10 20-30
No. 200 0-2.5 0-2.5 0-2.5 0-2.5 8-13
Asphalt Content (%) 9.2 6.0 6.0 6.0 6.5
Performance Grade
PG 76-22+ PG 64-16 PG 76-22+ PG 76-22+ PG 76-22+
Binder
4. Tire / Pavement Field Noise Measurements
In December of 2007, ADOT in conjunction with the Rubber Pavement Association (RPA)
and Arizona State University (ASU) performed spot check highway noise measurements on
the 32 test sections (Carlson et al, 2007; Biligiri, 2008). This 2007 noise measurements were
not performed using On-Board Sound Intensity (OBSI) technique. Rather, a hand held noise
meter was attached to the running board of a van, in such a way that the noise meter was in
close proximity to the tire / pavement interface. Freeway speeds are typically above the cross-
over speeds for all vehicle types. Above the cross-over speeds, tire pavement noise is the
dominant traffic noise source, being much higher than the aerodynamic and the engine sources
(Biligiri, 2008). This procedure was used to measure implicitly the tire pavement interaction
noise. This was a similar technique that was used in Arizona in the early 1990’s to measure
the tire / pavement noise. The sound meter was calibrated to measure the sound intensity in
the range of 80 to 130 decibels (dB), which is an appropriate range of noise measurements
in the field. A computer was connected to the sound meter to store the data. Stored data was
transferred to a PC via an RS-232 interface and analyzed using the system software. The total
time taken for one full run on the 32 test sections depended on the speed of the test vehicle.
Four runs were performed at three different speeds, namely, 100, 120 and 135 km/h (60, 72
and 75 mph) and the readings were recorded simultaneously.
4.1. Data Analysis
The sound levels were analyzed separately for each test section corresponding to each
range of Mileposts (for a mix type). A typical plot of noise measurement at 100 km/h (60 mph)
is shown in Figure 3. As seen from the figure, the least noise levels were observed for AR-
ACFC test sections. This was true for all speeds, and it was observed that the tire / pavement
noise levels were higher at higher test speeds.
For each mix type, the noise levels measured on each test section were grouped and
averaged. Figure 4 shows the average values of tire / pavement noise levels for each mix type
at the 100 km/h (60 mph) test speed. The least noise level was for the AR-ACFC mix and
the highest noise was on the P-ACFC pavement type, with a difference of around 2.5 dB. In
general, the noise level of each test section appeared to be related to the degree of surface
deterioration. The AR-ACFC experienced the least cracking and wear after eight years of
service with the other test sections showing considerable cracking and wear as illustrated in
Figure 5.
107.00
100 KMPH
106.00
TIRE / PAVEMENT NOISE (dB)
105.00
104.00
103.00
102.00
101.00
100.00
1-AR-ACFC-1/2
1-AR-ACFC-3/4
1-P-ACFC-3/4
1-ACFC-3/4
2-AR-ACFC-3/4
2-P-ACFC-3/4
2-ACFC-3/4
3-P-ACFC-3/4
3-AR-ACFC-3/4
4-AR-ACFC-3/4
6-AR-ACFC-3/4
2-AR-ACFC-1/2
3-ACFC-3/4
4-ACFC-3/4
4-P-ACFC-3/4
5-AR-ACFC-3/4
5-ACFC-3/4
5-P-ACFC-3/4
6-ACFC-3/4
6-P-ACFC-3/4
1-SMA-3/4
1-PEM-1 1/4
2-PEM-1 1/4
2-SMA-3/4
3-PEM-1 1/4
3-SMA-3/4
4-SMA-3/4
4-PEM-1 1/4
5-SMA-3/4
5-PEM-1 1/4
6-PEM-1 1/4
6-SMA-3/4
Figure 3. Tire Pavement Noise Levels (dB) for All the Test Sections at 100 km/h (60 mph)
105.5
104.77 104.98
105.0
104.48
Tire / Pavement Noise (dB)
104.5
104.0
103.5
103.0
102.69
102.40
102.5
102.0
101.5
101.0
AR-ACFC ACFC P-ACFC PEM SMA
Figure 4. Average Tire / Pavement Noise Levels for Each Mix Type at Test Speed of 100 km/h
(60 mph)
Figure 5. Illustration of 2007 Pavement Surface Deterioration after Eight Years of Service
4.2. OBSI Noise Measurements Comparison: 2002 and 2008
OBSI noise measurements were taken during the fall of 2002 by ADOT as part of the
Arizona’s Quiet Pavement Program. In addition, Dynatest Inc. measured noise levels on
the I-10 sections using an OBSI technique during March 2008 at 100 km/h (60 mph). This
independent set of Dynatest measurements were done as part of a larger California – Arizona
highway noise study (Scofield, 2008; Kohler et al, 2007; CALTRANS, 2006; Scofield and
Donovan, 2003). Figure 6 shows a comparison of noise readings between the years: fall 2002
and March 2008. In addition, the previously spot check measurements taken in December
2007 are shown on the same Figure 6 for comparative purposes. As can be observed, the data
of the average tire / pavement noise values are very similar in trend for the 2007 and 2008
measurements. The least noise observed for all years and measurements is for AR-ACFC
mixtures. This difference agrees with visual distress observations made in the fall of 2007.
Several sections that exhibited higher noise have greater amount of raveling and cracking.
106
104.77 104.98 104.48
105
104
104.68 102.69
102.84
103 102.40 102.17
102 101.56
101 100.6
99.8 101.0
100
99.94
99 99.8
98 98.9
97
96
AR-ACFC ACFC P-ACFC PEM SMA
2007 Measured at 60 MPH Dynatest 2008 at 60 MPH

Scofield-Donovan 2002 at 60 MPH
Figure 6. Comparison of Average Tire / Pavement Noise (dB) for Arizona I-10 Test Sections,
Years of 2002, 2007 and 2008
5. Laboratory Evaluation of I-10 Field Cores – Dampening Properties
In addition to the field noise measurements, 30 core samples from I-10 test sections
were obtained to evaluate their dampening (impedance) properties using an Ultrasonic
Pulse Velocity (UPV) testing method in the laboratory. This section documents the UPV test
methodology, experimental test results conducted on the field cores as well as developed
predictive model to validate laboratory test results.
5.2. Non-Destructive Ultrasonic Pulse Velocity (UPV) Test
Non destructive evaluation techniques have been widely used in industry to measure the
elastic or shear modulus of composite materials (Pellinen, 2001). The nondestructive UPV
technique is based on the measurement of wave velocities through material as described in
ASTM C597-02: Standard Test Method for Measuring Pulse Velocity through Concrete (ASTM
E494-05; ASTM C597-02). The pulse velocity, v is related to the density and elastic properties
of a solid by (Krautkämer and Krautkämer, 1990):
[1]
where:
v = velocity (in/s or m/s),
ρ = density of material (lb/ft3 or kg/m3)
E = modulus of elasticity (psi or kPa)
µ= Poisson’s ratio
In the laboratory, the pulse velocity is calculated as a ratio of the measured path length to
the measured pulse time. That is,
[2]
where:
L = Distance between centers of transducer faces (in or m)
T = Ultrasonic Pulse Time for transit (micro-sec or sec)
The Elastic Modulus is calculated as follows:
E = K ρ v2 [3]
Where:
E = Modulus of Elasticity (psi or kPa),
K = a constant dependent on the Poisson’s ratio and confinement,
ρ = density of material (lb/ft3 or kg/m3)
v = velocity (in/s or m/s)
5.3. Ultrasonic Pulse Time Transit Model
The ultrasonic pulse time transit model developed in this study was based on mathematical
calculations programmed in an Excel™ spreadsheet (Biligiri, 2008). The input parameters
included pavement volumetric properties such as asphalt cement, air voids; aggregate
gradation (and type), and any admixtures such as crumb rubber inclusions. Portland cement
concrete mixtures can be also analyzed in this spreadsheet by manipulation of input properties.
Mixture volumetrics such as Volume of air voids in the total mix, Va (%), Volume of
rubber, Vr (%), Effective Volume of asphalt or bitumen, Vb (%) and Volume of aggregate
solids, Vs (%) are calculated using input parameters. The surface areas of mix components are
based of surface factors for each sieve obtained from NCAT Handbook (Roberts et al, 1996).
The calculated mixture volumetrics as well as surface area of mix components are used to
calculate Ultrasonic Pulse Time (UPT) of total mix for each pavement material components as
follows:
Total UPT (micro-sec or sec) = Ti = tair + trubber + tbitumen + taggregate [4]
The UPT addition of various pavement materials components is justified based on the four
types of mixture properties present in the pavement material. The theory is based on the basic
assumption that the mixture is a unit volume and the wave traverses through each component.
The difficulty of measuring time through each component can be easily manipulated if one
adds transit times of each pavement material component.
For a known pavement material thickness, L, Ultrasonic Pulse Velocity (UPV) can be
calculated as follows.
UPV (m/sec or in/sec) = [5]
With the calculated value of the density from mixture volumetrics, one can calculate the
value of Impedance as well as Effective Flow Resistivity, EFR in cgs Rayls as detailed next.
The value of EFR can be directly input in the Federal Highway Administration Traffic Noise
Model (FHWA TNM).
Ultrasonic pulse velocities for standard materials have been reported in the literature (Turner
and Pretlove, 1991; Howe, 1998; Cha and Cho, 2007; Demirboga et al, 2004; Aracne-Ruddle
et al, 1999; Mochinaga et al, 2006). For example, the ultrasonic velocity in rubber is about
2,115 m/s and for bitumen, it is about 1,800 m/s. As an initial calculation, the total transit time
of two pavement materials: a conventional dense graded asphalt mixture and an asphalt rubber
mixture that is known to dampen noise were analyzed using the model. The total transit time
was calculated as a summation of transit times of different asphalt mixture’s components. It
was observed that the total transit time for the asphalt rubber friction course mixture was 1.35
times greater than the conventional mixture. This represents the dampening characteristic of
the asphalt rubber mixture over the dense graded asphalt mixture. With the available transit
velocity, one can calculate the impedance of the material, which is given by
Z = r * v [6]
Where:
Z = Impedance, N-sec/m3,
ρ = density of the material (kg/m3),
v= ultrasonic wave velocity through the material, m/sec
As explained previously, higher the transit time, lower the velocity and vice versa.
FHWA TNM used by the United States to model traffic noise uses another form of sound
absorption, the Effective Flow Resistivity, EFR which is basically the impedance of the
material; also defined as the property of the material offering resistance to acoustic wave
through an open pore (Menge et al, 1998; Sandberg and Ejsmont, 2002; Rochat et al, 2007).
The TNM Technical Manual provides values of EFR in cgs Rayls for various ground types.
Mathematically, EFR, σ is directly proportional to Impedance, Z. Both are expressed in the
same units, N-sec/m (MKS units) or cgs Rayls (cgs units). It is noteworthy that
EFR in cgs Rayls = 0.1 * EFR in N-sec/m3 [7]
Typical values of EFR for different pavement materials have been measured and reported
by Rochat et al, which is shown in Table 3 (Rochat and Hastings, 2008).
Table 3. Typical Effective Flow Resistivity Values for Different Types of Pavements (Rochat
and Hastings, 2008)
Pavement Type EFR [cgs Rayls]

Portland Cement Concrete (PCC) 20,000
Old Dense Graded Asphalt Concrete (DGAC) 14,500
New Bonded Wearing Course (BWC), 30 mm 12,300
Asphalt Rubber Friction Course (ARFC or RAC-O) 6,000 - 6,100
Open Graded Asphalt Concrete (OGAC) 2,000 - 4,200
5.4. Experimental Program and Test Results
Cores were sampled along the right travel lane from five different pavement types on the
I-10 test sections. Samples were cored with a diameter of 4 inches (100 mm) and to a depth
of anywhere between 7.5 and 10.5 inches (~187-262 mm) that also included an inlay of
dense graded asphalt concrete. Cores were brought to ASU laboratories and surface of each
of the cores were sawed to obtain regular five different wearing course pavement materials.
Thus, thickness of five different course materials varied anywhere between 0.5 and 1.5 inches
(~12.5-37.5 mm) depending on the pavement type.
Air voids of samples were estimated using a CoreLok device. After the air voids of
samples were determined, acoustic dampening properties were determined using ultrasonic
pulse velocity test as described in the previous sections.
Ultrasonic pulse velocity test methodology was conducted using 20 kHz transducers on
30 field cores. Honey was used as couplant between sample and transducer interface. UPT
was noted for the samples under investigation. The measured length of the sample was used
to calculate UPV and hence, Impedance and EFR for all pavement material samples. Table 4
summarizes the ultrasonic pulse velocity test results for the 30 field cores samples.
5.5. Model Prediction versus Laboratory Measurements
Both model predictions and actual test results on I-10 field cores were compared and the
results are shown in Figure 7. As observed in the figure, there is a good correlation between
the model predictions and laboratory measured Ultrasonic Pulse Time. It is noted that the
model under-predicts high UPT values; one reason for the under-prediction is attributed to
the input parameters used in the model, such as the use of literature reported ultrasonic wave
velocity for rubber particles and aggregates. Nevertheless, there is a rational correlation
between both the predicted and measured values.
5.6. EFR Sensitivity to Air Voids and Asphalt Content – I-10 Field Cores
The developed model was also used to understand the impact of material’s volumetric
change on EFR. An example showing changes in mixtures’ air voids is shown in Figure 8.
With an increase in air voids, there is a decrease in EFR which means that there is a positive
effect of air voids increase on noise dampening. This is rational and agrees with experts’
opinion on the role of the air voids in the mix’s noise characteristics.
Figure 9 shows a change in EFR for the different mixtures with respect to asphalt content.
The figure shows both predicted and measured EFR values for all the mixtures. It can be seen
that the predicted EFR values are higher for mixtures with lower asphalt content while the
least EFR was observed for the mixture with highest asphalt content (~9.2%). Note that the
lowest calculated EFR values were for asphalt rubber mixtures that also possessed highest air
voids levels.
Table 4. Ultrasonic Pulse Velocity Test Results for I-10 Field Cores
Calculated
Calculated
Z, Calculated Average
Average Measured Ultrasonic
Mixture Height Density Impedance Z, Measured Z,
Air Voids UPT Pulse
Type (mm) (kg/m3) (N-sec/m3 Impedance Impedance
(%) (sec) Velocity
or MKS (cgs Rayls) (cgs Rayls)
(m/sec)
Rayls)
26.4 8.20E-06 3,165 2.0785 64,460 6,446
33.1 1.02E-05 3,191 2.0924 65,437 6,544
33.2 1.00E-05 3,267 2.0556 65,812 6,581
PEM 17.14 6,738
31.1 8.60E-06 3,563 2.1445 74,878 7,488
22.1 7.30E-06 2,982 2.0386 59,578 5,958
28.8 7.60E-06 3,728 2.0287 74,111 7,411
37.5 1.17E-05 3,154 2.0712 64,015 6,401
34.7 1.05E-05 3,250 2.0789 66,224 6,622
17.4 6.50E-06 2,628 1.9631 50,551 5,055
P-ACFC 20.96 5,891
28.7 9.47E-06 2,984 1.9803 57,914 5,791
20.8 6.40E-06 3,194 1.9951 62,444 6,244
28.4 1.03E-05 2,716 1.9666 52,336 5,234
13.5 5.70E-06 2,326 2.0959 47,783 4,778
19.2 6.80E-06 2,774 2.0784 56,491 5,649
AR- 17.2 5.90E-06 2,863 2.1076 59,129 5,913
17.17 5,754
ACFC 24.7 8.60E-06 2,824 2.0846 57,700 5,770
20.4 7.30E-06 2,752 2.1422 57,775 5,778
18.3 5.60E-06 3,214 2.1064 66,352 6,635
19.2 6.00E-06 3,145 2.2768 70,172 7,017
23.5 6.00E-06 3,855 2.2100 83,492 8,349
23.9 5.60E-06 4,209 2.3652 97,557 9,756
ACFC 13.80 7,339
25.8 7.00E-06 3,624 2.2971 81,587 8,159
22.7 8.17E-06 2,739 2.0293 54,474 5,447
20.0 7.10E-06 2,773 1.9529 53,077 5,308
26.6 6.60E-06 3,974 2.0796 80,996 8,100
23.6 6.70E-06 3,472 2.1581 73,424 7,342
24.6 6.07E-06 3,991 2.1905 85,667 8,567
SMA 9.65 7,452
18.3 6.97E-06 2,581 2.1738 54,980 5,498
31.1 8.10E-06 3,784 2.0997 77,863 7,786
31.0 8.50E-06 3,593 2.1063 74,163 7,416
15
Predicted Ultrasonic Pulse Time (micro-sec)
12
9
2
R = 0.68
n = 30
0
0 3 6 9 12 15
Measured Ultrasonic Pulse Time (micro-sec)
Figure 7. Comparisons between Predicted and Measured Ultrasonic Pulse Time (micro-sec)
for I-10 Field Cores
10000 24
Calculated Effective Flow Resistivity, EFR (cgs Rayls)
9000 20 21
18
8000 18
Air Voids (%)
7339 7452
16
7000 15
6738
6000 12
11
5891
5754
10
5000 9
PEM P-ACFC AR-ACFC ACFC SMA
Figure 8. Sensitivity of EFR (cgs Rayls) to Varying Approximate Air Voids Levels – I-10 test
sections
10000 10
9000 8
6 6 6
Asphalt Content (%)

7
8000 6
7452
7339
7000 4
6738
6000 2
5891 5754
5000 0
PEM P-ACFC AR-ACFC ACFC SMA
Figure 9. Sensitivity of EFR (cgs Rayls) to Varying Asphalt Content – I-10 test sections
6. Field Measured Tire / Pavement Noise versus Laboratory Calculated EFR
Figure 10 shows a relationship between field measured tire / pavement noise and calculated
EFR (estimated from laboratory UPT measurements) for I-10 test sections. Field noise
measurements shown in the figure are as measured by Dynatest Inc. on the I-10 sections using
an OBSI technique during March 2008 at 100 km/h (60 mph). As mentioned previously, this
independent set of Dynatest measurements were done as part of a larger California – Arizona
highway noise study.
As observed, field noise was lower for mixtures with lower EFR except for P-ACFC
mixture. It must be noted that the field cores were randomly sampled from just five sections
out of the total 32 sections. Again, the sampling might not have represented exactly the five
different pavement types. Additionally, P-ACFC field cores mixtures had higher air voids
compared to the other materials and hence, calculated EFR were low.
106 9,000

105 7452 8,000
7339
6738 104.7
104 7,000
103 6,000
102.9 5891
5754
102 5,000
102.2
101.6
101 4,000
100 3,000
99.9
99 2,000
98 1,000
AR-ACFC PEM SMA ACFC P-ACFC
Figure 10. Relationship between Tire / Pavement Noise and Calculated Effective Flow
Resistivity for All Mixtures – I10 Test Sections
This apart, as mentioned previously, several sections that exhibited higher noise had
greater amount of raveling and cracking. This was particularly observed for P-ACFC
mixtures. Hence, the relationship between field measured noise and estimated EFR were not
direct. Nevertheless, four other sections showed excellent correlation between field noise and
estimated EFR. Overall, AR-ACFC mixes had the lowest recorded field noise in addition to
the lowest laboratory calculated EFR. This evidently proves that AR-ACFC mixtures had the
highest potential for noise dampening; also, the surface deterioration on these sections was the
least amongst all the wearing courses.
7. Summary and Conclusions
The purpose of this study was to conduct field noise measurements on five different
pavement wearing courses that were placed as test sections on a highly trafficked Interstate
in Arizona, United States. The pavement sections were placed in the year 1999 and included:
Permeable European Mixture (PEM), Stone Matrix Asphalt (SMA), Asphalt Rubber Open
Graded Friction Course (AR-ACFC), Polymer Modified Open Graded Friction Course
(P-ACFC), ADOT’s Standard Open Graded Friction Course (ACFC). Noise measurements
were performed using a vehicle similar to On-Board Sound Intensity (OBSI) technique.
Noise data were collected in 2002, 2007 and 2008. The least noise observed for all years
and measurements was for the AR-ACFC mixtures. In general, the noise level of each test
section appeared related to the degree of surface deterioration. The AR-ACFC experienced
the least cracking and wear after eight years of service with the other test sections showing
considerable cracking and wear.
In addition, 30 core samples from these test sections were obtained to evaluate their
dampening properties using an Ultrasonic Pulse Velocity (UPV) testing method in the
laboratory. The results of these noise measurements and laboratory testing were discussed
along with the degree of surface deterioration of each pavement test section in the field.
Correlations between predicted and laboratory measured UPT for all the pavement materials
under investigation were performed. The correlation between the model prediction and
laboratory measured UPT was good. The developed model was also used to understand
changes in materials volumetric on UPT and EFR, such as changes in air voids and asphalt
content. Relationships between field measured noise and calculated EFR (estimated from
laboratory measured UPT) for each mixture type were established. Overall, AR-ACFC mixes
had the lowest recorded field noise in addition to the lowest laboratory calculated EFR.
Bibliography
ASTM C597-02, Standard Test Method for Pulse Velocity through Concrete, 2002.
ASTM E494-05 Standard Practice for Measuring Ultrasonic Velocity in Materials, 2005.
Biligiri, K. P. (2008) Asphalt Mixtures’ Properties Indicative of Tire / Pavement Noise, Ph.D.
Dissertation, Department of Civil and Environmental Engineering, Arizona State University,
December 2008.
Carlson, D. D., Way, G. B., Zareh, A., Biligiri, K. P. and Kaloush, K. E., Tire / Pavement
Noise Characteristics of Wearing Course Asphalt Mixtures, SURF 2008, 6th Symposium on
Pavement Surface Characteristics, Portoroz, Slovenia, Europe, 20-22 October 2008.
Cha, M and Cho, G. (2007) Compression Wave Velocity of Cylindrical Rock Specimens:
Engineering Modulus Interpretation, Japanese Journal of Applied Physics, Vol.46, No. 7B,
2007, pp. 4497-4499.
Demirboga, R., Turkmen, I and Karakoc, M. (2004) Relationship between Ultrasonic Velocity
and Compressive Strength for High-Volume Mineral-Admixtured Concrete, Cement and
Concrete Research 34, 2004, pp. 2329-2336.
Gibbs, D., Iwasaki, R., Bernhard, R., Bledsoe, J., Carlson, D., Corbisier, C., Fults, K., Hearne,
T., McMullen, K., Newcomb, D., Roberts, J., Rochat, J., Scofield, L and Swanlund, M. (2005).
Quiet Pavement Systems in Europe, Report No. FHWA-PL-05-011, Submitted by American
Trade Initiatives, Virginia, Sponsored by Federal Highway Administration, American
Association of State Highway and Transportation Officials, USA, May 2005.
Governor Hull unveils plan to Add Rubberized Asphalt to Valley Freeways (2005). News,
Maricopa Association of Governments. http://www.mag.maricopa.gov/detail.cms?item=1849
Howe, M. S. (1998) Acoustics of Fluid-Structure Interactions, Cambridge University Press,

1998.
Kohler, E., Ongel, A., Harvey, J., and Rymer, B. (2007) OBSI Measurements at the UC
Pavement Research Center, Presentation made at the Transportation Research Board Summer
Meeting, San Luis Obispo, California, USA, July 22-25, 2007.
Lerch, T., Wellner, F., Hublet, J., Koeltzsch, P. and Sarradj, E. (2003). Initial Parameters for
Modeling Tire / Road Noise Reduction of Porous Asphalt Surfacing, Proceedings of Third
International Symposium on Maintenance and Rehabilitation of Pavements and Technological
Control. University of Minho, Guimaraes, Portugal, July 2003. pp. 349-358.
Losa, M., Bonomo, G., Licitra, G., and Cerchiai, M. (2003). Performance Degradation of
Porous Asphalt Pavements, Proceedings of Third International Symposium on Maintenance
and Rehabilitation of Pavements and Technological Control, University of Minho, Guimaraes,
Portugal, July 2003. pp. 475-484.
Menge, C, W., Rossano, C. F., Anderson, G. S and Bajdek, C. J. (1998) FHWA Traffic Noise
Model Technical Manual, FHWA-PD-96-010, U. S. Department of Transportation, Federal
Highway Administration, Washington, D. C., February 1998.
Mochinaga, H., Onozuka, S., Kono, F., Ogawa, T., Takahashi, A and Torigoe, T (2006)
Properties of Oil sands and Bitumen in Athabasca, 2006 CSPG – CWLS Convention, 2006.
Pellinen, T. K. (2001) Investigation of the Use of Dynamic Modulus as an Indicator of Hot-

Mix Asphalt Performance. Ph.D. Dissertation, Arizona State University, Tempe, AZ. May
2001.
Road Noise (2004) The Osgood File, American Communications Foundation website,
Presented in CBS Radio Network, 17th May 2004. http://www.acfnewsource.org/science/
road_noise.html
Roberts, F.L., P.S. Kandhal, E.R. Brown, D.Y. Lee and T.W. Kennedy, (1996) Hot Mix Asphalt
Materials, Mixture Design and Construction, Second Edition, NAPA Education Foundation,
Lantham, MD, 1996.
Rochat, J. Hastings, A. and Ferroni, M (2007) Investigating the Implementation of Pavement

Effects via OBSI Data in the FHWA Traffic Noise Model (FHWA TNM), NOISE-CON 2007,
Reno, Nevada, USA, 22-24 October 2007.
Rochat, J. and Hastings, A (2008) Volpe Center Research and Activities Update, Presentation
at the Transportation Research Board ADC40 Committee Summer Meeting 2008, Key West,
Florida, USA, July 2008.
Sandberg, U and Ejsmont, J. A (2002) Tyre / Road Noise Reference Book, INFORMEX, 2002.
Scofield, L (2000) Arizona’s Pavement Preservation Experiment, Arizona Department of

Transportation, 2000.
Scofield, L. Email correspondence, April 28, 2008.
Scofield, L. and P. Donovan (2003). Development of Arizona’s Quiet Pavement Research

Program, Proceedings of Asphalt Rubber 2003, Brasilia, Brazil, December 2003.
The California Department of Transportation (2006), Division of Environmental Analysis,

Further Development of the Sound Intensity Method of Measuring Tire Noise Performance of
In-Situ Pavements, page 9, January 2006.
Turner, J. D., and A. J. Pretlove. Acoustics for Engineers, Published by Macmillan Education
Ltd, U. K. 1991.
Way, G. B. (2003). OGFC meets CRM - Where the rubber meets the rubber - 15 years of
durable success, Conference Proceedings, Asphalt Rubber 2003, Brasilia, Brazil, December
2003.
Zareh, Ali, Pavement Management 2007-2008 Pavement Condition Data Base, Arizona
Department of Transportation, United States of America, April 2008.
Asphalt-Rubber Open Graded Friction

Course Reduces Noise-The Quiet Pavement
Program
Ali Zareh — Douglas Carlson — George B. Way
Arizona Department of Transportation

Rubber Pavements Association
Recycle Tire Engineering and Research Foundation
azareh@azdot.gov
doug.carlson@rubberpavements.org
wayouta@cox.net
ABSTRACT: The Arizona Department of Transportation recognized a need to provide the

citizens of Maricopa County living next to its freeways with a less noisy quieter freeway. The
Maricopa County freeway system is referred to as the MAG freeway system and it is composed
of over 1200 lane miles (2000 lane kilometers) of multi-lane concrete pavement. Noise is
defined as a loud sound of any sort that is disagreeable or unwanted and the people living
next to the MAG freeway highways complained that the freeways were too noisy. Throughout
the ages of civilized history, noise has been an annoying irritant to mankind. In recent years
the noise level of freeways in urban and suburban areas in the United States and the MAG
freeway system in particular has increased perceptibly as freeway traffic has increased and
more people are living closer to freeways. To reduce such a noisy irritant generated by traffic
on concrete or asphalt pavements Arizona has employed a 25 mm open graded friction
course asphalt-rubber hot mix surface course. This paper reviews the experience in Arizona
and research in California as well, with using asphalt-rubber open graded friction course
pavements to reduce the noise by 3-12 decibels. These findings have led to Arizona developing
a large scale program of covering over 1200 lane miles (2000 lane kilometers) of relatively
new and new concrete with a 25 mm (1 in.) asphalt-rubber open graded friction hot mix
surface course to substantially reduce noise. This program is referred to in Arizona as the
Quiet Pavements Program and has been very successful.
KEYWORDS: asphalt rubber, open graded friction course, noise, measurement, performance
Asphalt Rubber 2009
1. Foreword
This paper is a continuation of a paper presented at the AR2006 Conference (Ali,

2006). Since 2006 the Arizona Department of Transportation (ADOT) in cooperation with
the Maricopa Association of Governments (MAG) has continued to build and expand the
MAG freeway system which is composed of a concrete pavement covered with a 25 mm
(1 in.) asphalt-rubber (AR) open graded friction course (ARFC) to reduce noise. Typically
the placement of ARFC reduces the noise by 4 dBA. This study reports on the ten year
Quiet Pavement Program Pilot Research Study being conducted in cooperation with Federal
Highway Administration (FHWA) for purposes of determining how long the reduction of
noise associated with the ARFC placement will continue.
2. Background on Noise and Noise Measurement
Noise is defined as a loud sound of any sort that is disagreeable or unwanted. Through
the ages of civilized history noise has been an annoying irritant to mankind. The Roman
Emperor Caesar decreed in 20 BC that carts could only move through the city of Rome during
the night, since their noisy din during the day interfered with the daily business of Rome.
Carts moving through Rome at night were acceptable since Caesar and his court lived in the
mountains far from the noise of the carts in the city. As the centuries passed by mankind has
tried to live in a peaceful world. Horse drawn wagons and coaches with wooden wheels and
iron rims generated plenty of noise in the 1800’s. In the late 1800’s in England road builders
used wood blocks and even rubber blocks to deaden the wagon wheel noise. In the 1900’s
with the advent of automobiles and rubber pneumatic tires it appeared the tire/pavement noise
had finally ended for all time. However, as more automobiles and trucks took to the highways,
freeway noise gradually crept back into the city.
To measure the tire pavement noise several measurement instruments were developed. The
noise measurement instrument, which is most commonly used, is an electro-acoustical device
with a microphone that converts sound pressure (a scale value) into an electronic or voltage
signal, or vice versa. The instrument has a variety of names such as noise meter, sound meter,
or sound level meter. When placed near a sound source, it will display or readout a single
number of the corresponding decibel level in dB or dBA. The dB level refers to the sound or
noise level in general, whereas the dBA value is a sound or noise measurement adjusted to be
level to the sensitivity of the human level of hearing. The value of sound or noise is a function
of frequency typically it is obtained by averaging the sound pressure over a pre-set frequency
band, typically, from a frequency of about 100 Hz to as high as 20,000 Hz.
There are several noise measurement methods for measuring the traffic related level of
noise. The most common method is the wayside method which generally is conducted by
measuring the noise with a typical noise meter placed approximately 15.2 m (50 feet) from the
center of the roadway to be tested and a typical height of 1.5 m (5 feet), Figure 1.
Figure 1. Wayside method of measuring noise
Recently another mobile trailer form of measuring tire/pavement interface noise has been
developed. This method is called the Close Proximity method or CPX method, (a.k.a. “The
trailer method”). Figure 2 presents a picture of the trailer. The trailer method (CPX) was used
to accomplish the much of the work described in this report. The test tire is mounted in a
trailer, which is towed by a towing vehicle. Close to the test tire, generally within 0.1-0.5 m
(4 to 18 in), one or more microphones are located. The noise level is measured as an average
over a certain time interval, usually 4-60 seconds. Most trailers have an enclosure around
the microphone and test tire in order to provide screening from wind and traffic noise. Such
enclosures are lined on the inside with sound absorbing material. Some trailers may utilize
more than one test tire. The method may also utilize especially designed self-powered vehicles
that are not of the trailer type. This method is less sensitive to noise generated by other traffic.
This equipment is essentially designed for comparing road surfaces.
Figure 2. CPX trailer method of measuring noise

The newest form of measuring the tire pavement interface noise is the noise intensity
method, now referred to as the On Board Noise Intensity (OBSI) method of tire pavement
noise measurement. The American Association of State Highway and Transportation Officials
(AASHTO) and the American Society of Testing Materials (ASTM) are both developing
standard OBSI noise measurement methods. The sound intensity measuring hardware consists
of a probe (microphone pair) held next to the tire/pavement contact patch by a fixture attached
to the wheel studs of the test tire/wheel. The microphone is cabled to the interior of the
vehicle where the signals are simultaneously captured on a recorder and processed by a real
time-analyzer. The specially tuned microphone only picks up the noise of the tire pavement
interface and no other noises from wind or other vehicles or any other sound. Figure 3 shows
a typical installation.
Figure 3. Noise intensity method of measuring noise
3. Asphalt-rubber Open Graded Surface Course
To reduce such a noisy irritant generated by traffic on concrete or asphalt pavements both
California and Arizona have employed a 25 mm (1 in.) open graded AR hot mix surface
course. Asphalt-rubber is a mixture of 80 percent hot paving grade asphalt with 20 percent
ground tire rubber produced from waste tires Way (2001). The resultant asphalt-rubber binder
mixture is added hot to a hot open graded mineral aggregate to produce an asphalt rubber
friction course as the final wearing course of the pavement structure. This paper reviews the
experience in both California and Arizona with using asphalt-rubber hot mix pavements to
reduce the noise by 3-12 decibels. Research has been conducted in both states to determine
the nature of the noise and how best to measure it and to compare the results of such tests
to various pavement surfaces to reduce the noise to an acceptable level. Findings from this
research include the following observations, namely that roadside measurements have shown
that open-graded asphalt rubber friction courses (ARFC) can achieve 3 to 5 dB noise level
reduction when compared to traditional asphalt dense graded surfaces and 6 to 12 dB noise
level reduction when compared to concrete surfaces. In addition sound intensity measurements
(measurements taken close to the tire/pavement interface) have shown that open-graded
asphalt rubber friction courses (ARFC) are effective in reducing noise by 4 to 6 dB compared
with traditional dense graded asphalt concrete and by 6 to 12 dB when compared to concrete
surfaces. These findings have led to Arizona developing a large scale program of covering
over 1500 lane kilometers of relatively new concrete with ARFC to substantially reduce noise.
This program is referred to as the Quiet Pavements program and has been very successful.
4. Arizona Quiet Pavements Program
Phoenix, Arizona and its surrounding suburban sister cities have experienced a tremendous
growth in population in the last 50 years. Phoenix is one of fastest growing cities in all of
the country and now is the fifth largest city in population in the United States. With growth
in population has come the growth in automobile traffic and the need for more freeways in
the Phoenix Metropolitan area. In 1985 the citizens of Maricopa County, which includes
Phoenix and numerous sister cities, voted in favor of a 0.5 cents sales tax to fund the design
and construction of over 115 miles of new freeways. Since the funding only addressed design
and construction of the freeways and not maintenance or rehabilitation it was decided that the
freeways would be built with concrete pavements.
Starting in 1986 construction of the freeways began and with time more miles were
completed and more people bought homes built near the freeway, and they began to notice
the annoying noise of the freeways. Even though sound walls were constructed to mitigate
the noise, complaints about noise continued. In the year 2000 the Arizona Department
of Transportation (ADOT) began construction of the widening of 17 km (10 miles) of
Superstition Freeway which is in the Phoenix Metropolitan area. The freeway construction
included widening the concrete pavement from three lanes to six lanes in each direction to
accommodate the over 150,000 vehicles a day that use the freeway, Figure 4. As part of this
major construction ADOT decided to overlay all the lanes full width with an asphalt rubber
open graded friction course (ARFC). The ARFC surface was selected to provide a new surface
with a smooth ride, good skid resistance, a new surface that could be plainly re-striped and to
reduce the tire/pavement noise.
Figure 4. Superstition Freeway in 2003 with ARFC surface (Courtesy of Arizona DOT)
As construction drew to a close and the placement of the ARFC progressed to its ultimate
completion drivers, passengers and people living next to the freeway began to notice the
new ARFC riding surface was very quiet. Unexpectedly, people began to write the local
newspapers and call local radio talk shows in praise of the new quiet riding surface. Although
the quiet riding surface was an ARFC it soon began to be called by the local media and people
in general simply rubberized asphalt. From this grassroots reaction to peace in the valley,
that is the Phoenix Metropolitan area is called the Valley of the Sun, action groups sprang
up to petition government to cover all the freeway miles with rubberized asphalt. The local
governments as well as state government heard the voice of the people and developed a Quiet
Pavements Program in December 2002, (Carlson, 2003) and (Scofield, 2003). Then Governor
Jane D. Hull and ADOT Director Victor Mendez informed the public that over the next three
to four years ADOT in cooperation the Maricopa Association of Governments (MAG) would
overlay the concrete freeway system with rubberized asphalt. The cost of this Program was
estimated to be 34 million dollars ($3.50/Sq. meter / $2.90 Sq. yd.) and it was considered a
Quality of Life issue for the people of the Phoenix Metropolitan area.
In addition ADOT would work with the Federal Highway Administration (FHWA) to
collect before and after the ARFC overlay noise measurements. This research effort would
be used to determine whether the pavement surface noise reducing properties of ARFC would
remain at a beneficial level over time. If results of the study are satisfactory, they could be
mathematically modeled and become part of the FHWA National Noise Model. Presently,
the FHWA noise model does not allow the surface noise characteristics to be used as an input
into the model. Practically speaking, the national noise model only allows walls or berms to
be used as a means of reducing noise. However, asphalt-rubber open graded mixes have been
shown to reduce noise to a noticeable degree, Figure 5.
Figure 5. Superstition Freeway 2002 noise reduction due to asphalt-rubber open graded mix,
courtesy Rubber Pavements Association
5. Noise Synthesis Study
The significant reduction of the tire/pavement noise due to the use of an ARFC and the
positive way in which people responded to this triggered the Rubber Pavements Association
(RPA) to sponsor a Noise Synthesis Study (Sousa et al., 2004). The purpose of the study is to
acquaint people with what tire/pavement noise is and how an ARFC reduces the noise and to
what degree it reduces the noise. Tire/pavement noise has been known about for many years.
This report contains a summary of 47 studies based on research and data from the US, France,
Portugal, Spain, Sweden, Australia, New Zealand and Japan.
Early studies of noise focused mostly on concrete pavements and some asphalt pavements
were used as the basis for developing the FHWA noise policy in 1976. Although much
research has continued into tire/pavement research since 1976 the FHWA Policy has remained
virtually the same over all these years. Starting in the 1990’s both California and Arizona
began to investigate tire/pavement noise and to begin documenting their findings from actual
field measurements. These early projects clearly showed that an ARFC surface measurably
reduces noise. What these analytical studies did not show is to what degree people would
positively react to reducing freeway tire/pavement noise and how passionate they are about
this being a quality of life issue.
Noise is measured in decibels and the larger the decibel values the more irritating the
noise. Decibels are commonly abbreviated as dB or dBA. The dB abbreviation refers the level
of measured noise, whereas the dBA abbreviation refers to not only the measured value but
how people respond to the noise. Normal conversation in an office setting is typically about
40-50 dBA. The noise of a lawnmower is generally about 70-100 dBA depending upon how
close you are to it and of a diesel truck can be as much as 90 or more dBA. With regard to
highways the FHWA National Noise Policy sets an acceptable level of noise at 67 dBA at a
distance of 50 feet away from the centerline of the highway. By contrast the Arizona DOT
sets its acceptable level of highway noise at 64 dBA also 50 feet from the centerline of the
highway. Figure 6 shows common source of noise and the associated noise level.
Figure 6. Noise levels from quiet to very loud (Courtesy of Arizona DOT)
Highway noise is primarily generated from three point sources, the exhaust or tailpipe
noise, the engine noise and the tire/pavement noise. In addition highway noise is also a
function of traffic, more cars and trucks, more noise. Automotive engineers have very
successfully addressed both tailpipe noise and engine noise by muffling or dampening at
the point source. The tire/pavement noise has remained as a somewhat uncontrolled noise
source, primarily because the greatest emphasis has been placed on tire/pavement wet weather
friction, rather than noise.
Each of the three sources of noise contributes to the total noise. Typically when cars and
trucks reach a speed greater than about 60 km/hour (35 miles/hour) the tire/pavement noise
begins to become the dominant source of noise. At normal freeway speeds of 90 km/hour (55
miles/hour) or greater the tire/pavement interface noise can represent at least 70 percent of
the total noise as shown in Figure 7. Significantly reducing the tire/pavement point source of
noise obviously can reduce the overall noise significantly as well.
Figure 7. Noise The source of most highway noise is generated at the tire/pavement interface
(Courtesy of ACB Engineering)
What creates the tire/pavement noise is a subject of much conjecture and theory. The most
commonly held theory is that as the tire passes over the pavement a change in pressure occurs
at the tire/pavement interface and this pressure change generates the noise. Another theory
holds that the friction or rubbing of the tire against the pavement creates a noise from the
rubbing action. Also it has been observed that rolling a wheel across a concrete surface creates
more noise than rolling the same wheel across a carpet or rubber mat and thus the softness or
stiffness of the pavement itself may amplify or muffle the noise.
Reviewing the three theories about what creates the tire/pavement noise it is possible to
postulate why an ARFC would provide a very quiet ride. Empirical measurements demonstrate
that as the concrete grooves or air pockets are cut or positioned differently in relation to the
tire more or less noise develops. Likewise an asphalt open grade friction course surface has
a tremendous amount of air pockets which can dampen the pressure change gradient, thus
reducing the noise. An ARFC also has a huge amount of air pockets or air voids, which can
contribute to less of a pressure change. The ARFC surface is a much smoother riding surface
then the concrete. The ride is smoother because the ARFC mix is placed in a continuous
manner with minimal joints and the aggregate top size is 9.5 mm (3/8 in.). Such a smooth
riding surface and small top size aggregate could contribute to less tire deformation with
travel and less squeezing of air between the tire and pavement (less pressure change) and thus
less noise.
In addition the ARFC is a rubber like soft surface much like a carpet or rubber mat and this
too could reduce noise. The softness of the ARFC comes from a much higher asphalt binder
content and crumb rubber content (20 percent of the binder). Typical open graded friction
course mixes have about six percent asphalt, whereas an ARFC has 9 to 10 percent asphalt
rubber binder. Rubber is commonly used to reduce noise, thus there may be reason to believe
that rubber particles in the ARFC may very well contribute to less noise. All of these material
related aspects of an ARFC surface would tend to reduce noise and most likely collectively
do contribute to less noise. Thus there is good reason to believe that the empirical evidence of
less noise with an ARFC surface is related to a yet not fully understood scientific mechanism.
6. Tire/Pavement Interface Noise Compared to Wayside Measurements
Although the scientific mechanism of tire/pavement noise phenomenon may not be

altogether clear, clearly some surfaces are less noisy than others. Concrete freeway surfaces
can be textured or ground to give different levels of noise. All the studies appear to agree that
transverse tined concrete is the noisiest surface. Longitudinally textured concrete is somewhat
less noisy and a very specially diamond ground surface referred to by the industry as whisper
concrete, is the least noisy of any of the concrete surface textures. An ARFC surface is
even quieter that any of the concrete textures as seen in the Table 1. In this Table developed
by the Arizona DOT on concrete surfaces and ARFC pavements tested on the Phoenix
Metropolitan Freeway system it is clear that the ARFC tire/pavement noise point source
level is quietest as shown in the CPX column. The CPX column show actual close proximity
noise measurements taken very close to the tire/pavement interface. The 15 m (50 feet) from
centerline noise values are estimated from research conducted by the California Department
of Transportation that relates a tire/pavement point source measurement to the standard 15 m
(50 feet) from pavement centerline measurement which is used in the FHWA National Noise
Policy. The ARFC is the least noisy surface or quietest pavement surface in this study. For
each dBA reduction in noise, whether at the tire or farther away a noise wall height can be
reduced approximately 0.60 m (2 feet). Clearly as shown in Table 1 the ARFC is less noisy in
absolute terms but beyond this, there is the second part of the story. When an ARFC is placed
over a concrete freeway section the noise level reduction to the human ear seems even more
dramatic. From noise testing done by the Arizona DOT and California DOT it was possible to
record the noise spectrum over a wide range of frequencies.
Table 1. CPX noise measurements at the tire/pavement interface and estimated noise 15 m (50
feet) away from tire
Surface Type CPX Noise Noise 15 m

Random Trans 104.9 80.9
Transverse 102.5 78.5
Longitudinal 99.1 75.1
Whisper Grind 95.5 71.5
ARFC 91.8 67.8
The human ear can hear sounds from about 500 Hz to about 20,000 Hz. Sounds in the 1,000
to 2,000 Hz range tend to be of an annoying type to the human ear. By recording the frequency
spectrum of the concrete pavements it was observed that transverse textured concrete has a
tonal spike in the 1,000 to 2,000 Hz range (Scofield, 2003) that is particularly annoying to
human hearing as shown in Figure 8. The longitudinal and whisper texture concrete has less of
a spike, however the ARFC actually has a dip in the 1,000 to 2,000 Hz range which means to
human hearing there is very much less annoying noise being heard. In effect the ARFC is both
reducing the overall noise over all frequencies and the irritating tonal spike noise in the 1,000
to 2,000 Hz range.
Figure 8. Noise reduction in the 1,000 to 2,000 Hz range due to an ARFC surface, Scofield
(2003)
As further documentation of an Open Graded Asphalt Rubber Friction Course noise tests
from many different surfaces were compared and combined into a single illustration (Donovan,
2003), shown in Figure 9. As can be seen the least noisy surface measured to date is an Open
Graded Asphalt Rubber Friction Course (ARFC). The Open Graded Asphalt Rubber Friction
Course (ARFC) produces less noise as documented in numerous reports and by literally
millions of people driving over the freeway pavement surfaces in the Phoenix Metropolitan
area or living next to a freeway overlaid with an ARFC. The ARFC is less noisy because
it contains air pockets, air voids that reduce the pressure change. Also, because of its very
smooth riding due to the small aggregate size and because it is a layer of low modulus elastic
soft material due to the high percentage of asphalt binder and crumb rubber. The Arizona DOT
expects at least a 4-decibel reduction with the use of an ARFC and this reduction has led to
a significant improvement in the quality of life of home owners living near the freeway and
people driving on the freeway.
Figure 9. Comparison of the CPX (near the tire) tire/pavement noise for different surfaces,
(Donovan, 2003)
7. Quiet Pavement Program Pilot Research Study
The Quiet Pavement Program Pilot (QPPP) research study is a ten year study being
conducted in cooperation with Federal Highway Administration (FHWA) for the purpose
of determining how long the reduction of noise associated with the ARFC placement will
continue. The study as of 2010 is six years old. The study involves taking three types of noise
measurements. A set of OBSI measurements are being taken at selected sites within the MAG
freeway system. Wayside measurements are also being taken at selected at grade sites, and
noise measurements are also being taken within selected neighborhoods. Although the study
is not complete Table 2 shows the average level of reductions from measurements taken at the
various sites before the ARFC was placed and then after six years of service. As can be seen
the ARFC is still providing the minimum 4 dBA of noise reduction.
Table2. ADOT MAG Freeway QPPP noise measurements reduction after six years of service
Site 1- OBSI Noise Measurments

Range of Noise Reductions -4.1 dBA to -13.2 dBA
Average Noise Reduction 8.3 dBA
Site 2- Neighborhood Wayside Noise Measurements

Range of Noise Reductions +1.3 dBA to –12.3 dBA
Site 3- Roadway Wayside Noise Measurements

Range of Noise Reductions -4.4 dBA to -12.4 dBA
8. Conclusions
From the compilation of the data and analysis of the data obtained in the study the
following conclusions can be derived Sousa (Sousa, 2005):
• Roadside measurements have shown that asphalt rubber friction courses can achieve
3 to 5 dB noise level reduction when compared to traditional asphalt dense graded
surfaces and 6 to 12 dB noise level reduction when compared to concrete surfaces.
• Sound intensity measurements (CPX) have shown that asphalt rubber surfaces are
effective in reducing noise by 4 to 6 dB compared with traditional dense graded asphalt
concrete and by 6 to 12 dB when compared to concrete surfaces.
Undoubtedly more research will continue to find even better and quieter pavements, but
for now an ARFC surface has set the standard for quiet pavements. Figure 10 is a general
overview of the successful ADOT MAG Freeway that continues to reduce freeway noise to
this date.
Figure 9. ADOT MAG Freeway Quiet Pavement Program
7. References
Carlson, D. & Zhu, H., “Analysis of Traffic Noise before and After Paving with Asphalt-
Rubber”, AR 2003 Brasilia, Brazil, Dec. 2003.
Donovan, P. & Rymer, B., “Measurement of Tire Pavement Noise Sound Intensity
Methodology”, AR 2003 Brasilia, Brazil, Dec. 2003.
Scofield, L., “Development Of Arizona’s Quiet Pavement Research Program”, AR 2003

Brasilia, Brazil, Dec. 2003.
Sousa, J., Way, G. & Zhu, H, “Asphalt Rubber Noise Data Compilation (Synthesis of Current
Practices)”, Rubber Pavements Association, 2004.
Way, George, “$18 Million Benefits of Using Ground Tire Rubber in Asphalt, 11 Years of
Success”, Beneficial Use of Recycled Materials in Transportation Applications, Pages
109-119, The Recycled Materials Resource Center, November 13-15, 2001, Arlington,
Virginia.
Zareh, A., Way, G. & Kaloush, K, “Asphalt-Rubber open graded mix reduces tire pavement
noise”, AR2006 Conference, Palm Springs, California, 2006.
Development of Innovative Pavement Types

to Reduce Traffic Noise
Jürgen Haberl
Heller Ingenieurgesellschaft mbH, Austria
Karlsgasse 5
A - 1040 Wien
Austria
juergen.haberl@heller-ig.at
ABSTRACT. In recent years different innovative noise reducing road surfaces have been
developed that either use asphalt or concrete, e.g. single or double layer porous asphalt,
noise-reducing stone mastic asphalt, noise-reducing thin layers, surface treated concrete or
exposed aggregate cement concrete. The noise reduction potential of such pavements could
vary between 1 and 6dB(A) in comparison with conventional asphalt concrete surfaces
depending on the road surface type and on its condition. In the context of a cooperation of
BMVIT (Federal Ministry of Transport, Innovation and Technology of Austria), ASFINAG
(Austrian Motorway Company), ISTU (Institute for Road Construction and Road Maintenance
of the Vienna University of Technology) together with Heller Ingenieurgesellschaft mbH
(Vienna) and representatives of the road construction industry, different innovative surface
layer types have been examined for the first time in Austria under the same ambient conditions
regarding to their noise characteristics, their durability and their influence on road safety
compared with conventionally used surface layer types (stone mastic asphalt and asphalt
concrete). The test section consists of altogether 8 different test fields with the surface layer
types noise-reducing stone mastic asphalt, single and double layer porous asphalt and two
reference surfaces. Analysing the first noise measurement results the clearly noise related
advantage of porous asphalt surface layers both related to the overall traffic (SPB- index) and
the individual vehicle categories could be shown. In comparison with the examined asphalt
concrete surface layer they perform about 6-8dB(A) lower noise levels. The examined low
noise stone mastic asphalt layers exhibit about 2–4dB(A) higher passing-by noise levels than
the porous surface layer types.
KEYWORDS: traffic noise, noise reducing pavements, porous asphalt, noise measurements.
1. Introduction
Due to rising traffic volume it is more and more recommended to reduce traffic noise
directly at the source. The noise of a vehicle passing-by results on the one hand on the tyre
rolling on the road surface, and on the other hand on sound emitted by the vehicle drive
train and exhaust system. A substantial contribution to the reduction of traffic noise could be
reached by minimising the tyre/road noise which depends mainly on the driven vehicle speed,
on the type of tyre, on the pavement type and on the surface condition of the road.
2. Road traffic noise
2.1 Noise emission sources
The noise emitted by a passing vehicle on the road is a very complex situation. In general,
one can divide the different factors influencing the overall noise into the main parts
• tyre/road noise,
• power unit noise and
• wind turbulence noise.
Tyre/road noise is related to the rolling of the tyres on the road surface which depends
mainly on the driven speed, on the vehicle and tyre type and on the kind of the road surface
and their characteristics. The influence of the tyre/road noise on the overall noise emission of
vehicles passing by increases with an increased driving speed. In urban regions with speed
limits around 30–50km/h the power unit of a vehicle is an important noise factor, whereas at
highway speeds it is mostly negligible. Regarding driving speeds exceeding 120km/h wind
turbulence noise plays an increasingly dominating role on the noise emissions.
2.2 Noise-reducing pavements
In recent years, noise-reducing road surfaces have been introduced that can lead to
substantial reductions in traffic noise. However, the overall benefits of these noise-reducing
surfaces are also influenced by other factors including the tyres of the vehicles, special vehicle
parameters or the use of passive noise reduction measures such as barriers and traffic control
measures for noise reduction. In this paper Austrian experiences concerning the influence of
innovative noise-reducing pavement types are explained.
On the Austrian road network noise-reducing pavements are used for many years. An
estimation of the standard pavement types on the Austrian high level road system is shown in
Figure 1.
Figure 1. Estimated road statistics on the Austrian high level road system (Haberl, 2008)
3. Measurement methods
3.1. General statements
At present several measurement methods are used to assess traffic noise emissions, on
the one hand for the classification of road surfaces and vehicles and on the other hand for
the collection of the entire noise situation at close range of traffic routes. These methods
differ basically in the measuring technique, especially in the different positions of the used
microphones. Two principles are usually used, either rolling noise measurements with
microphones near the vehicle tyres (near- field) or pass-by measurements with microphones
on the side of the road (SILVIA, 2004).
3.1. Near-field measurements
Noise emissions in the near-field of the noise source can be measured by so called rolling
noise measurements. The measurement itself in general takes place with a trailer furnished
with a testing tyre within a sound-absorbing shroud. Because of the shroud construction of the
trailer the used microphones are to a large extent shielded in relation to the background noise.
The near-field noise emission is measured with microphones placed in direct proximity of the
testing tyre regarding a constant vehicle speed over the whole measuring distance.
3.1.1. Austrian rolling noise measurement method
In Austria a procedure for measuring rolling noise emissions is regulated in the “guidelines
and regulations for road construction” – RVS 11.06.64 (RVS, 1997) - and used either for
acceptance testing of new pavements concerning the rolling noise or for a technical evaluation
of different road surface layers concerning their noise behaviour. The tyre used is the PIARC-
tyre with flow longitudinal grooves commonly used for skid-resistance measurements.
The results of rolling noise measurements are analysed by calculating the energy
equivalent noise level over a defined road length, which is 500m according to the Austrian
standard, expressed in the value LMA. LMA is calculated as the mean value regarding the
measured frequency spectrum of two microphones over the measured road segment.
[1]
with:
LMAe ……………… energy equivalent noise level of one test run
LMAe,t,i ……………… energy equivalent noise level of one microphone
within the test run for each frequency spectrum
e ……………… number of test run
t ……………… number of microphone
i ……………… frequency spectrum
Figure 2 shows an example of the rolling noise measurement device used in Austria.
Figure 2. Schematic of the rolling noise measurement device used in Austria with detail of the
Austrian measurement tyre, photo by IFS Ziviltechniker Ges.m.b.H. (Haberl and Litzka, 2008)
3.1.2 International rolling noise measurement method
An international measuring and analysis procedure for near-field noise measurements,

the so called CPX- method regulated in the draft ISO/CD 11819-2 (ISO, 2000), is also based
on the collection of the sound pressure level near the tyre/road contact area and exhibits thus
many parallels to the Austrian method. The main differences will be found in the construction
of the measuring trailer, whereby after ISO/CD 11819-2 more liberties are given in this field,
in the use of up to 4 different test tyres and in the method of the evaluation of the measurement
results. The 4 test tyres should represent the different tyres available on the market (2 types of
summer tyres, 1 winter tyre and 1 heavy vehicle tyre).
The result of a CPX- measurement is expressed as an index value, CPXI – Close Proximity
Index, and calculated as a weighted summary of the energy equivalent noise levels of the used
test tyres (2 tyres for the survey purpose and 4 tyres for the investigatory purpose).
survey purpose:
[2]
investigatory purpose:
[3]
with:
CPXI ……………… Close Proximity Index

LA ……………… energy equivalent noise level of tyre A
(representative for a summer tyre)
LB ……………… energy equivalent noise level of tyre B
(representative for a summer tyre)
LC ……………… energy equivalent noise level of tyre C
(representative for a winter tyre)
LD ……………… energy equivalent noise level of tyre D
(representative for a heavy vehicle tyre)
An example of a rolling noise measurement trailer and the four measurement tyres
typically used in Austria is shown in Figure 3.
Figure 3. Example of a rolling noise measurement trailer used in Austria and a detail of the
4 different measurement tyres A, D, B and C used in Austria according to ISO/CD 11819 - 2
(photos by IFS Ziviltechniker Ges.m.b.H.)
3.1. Pass-by measurements
Pass-by measurements are regulated in the international standard ISO 11819 – 1, SPB,
statistical pass-by (ISO, 1997). The measurement principle is very easy; the maximum A-
weighted sound pressure level of a statistically significant number of individual vehicles
passing by is measured at a specified road-side location together with the vehicle speed.
An important fact for the comparison of measurement results is a standardised microphone
position. Therefore the horizontal distance from the microphone position to the centre of the
lane where the vehicles are passing by shall be in the range of 7,5 m ± 0,1 m, the associated
microphone height above the plane of the road surface shall be 1,2 m ± 0,1 m (see Figure 4).
The result of a SPB- measurement is expressed as an index value, SPBI – Statistical Pass-
By Index, and calculated as a weighted summary of the maximum sound pressure levels
of three different vehicle categories, cars (category 1), dual-axle heavy vehicles (category
2a) and multi-axle heavy vehicles (category 2b), at a certain reference speed. Category 2a
and category 2b can be summarised as category 2 – heavy vehicles. To ensure that random
errors do not become unacceptable large, the number of measured vehicles is minimised for
each vehicle category (cat.1 – min. 100veh., cat. 2a – min. 30veh., cat. 2b – min. 30veh. and
additionally cat. 2 – min. 80veh.)
[4]
with:
SPBI …………… Statistical Pass-By Index

L1, L2a and L2b …………… sound levels for vehicle categories 1,
2a and 2b at the reference speeds v1, v2a and v2b
W1, W2a and W2b …………… weighting factors which are equivalent
to the assumed proportions of vehicle categories in the traffic
v1, v2a and v2b …………… reference speeds of the individual
vehicle categories 1, 2a and 2b
Within pass-by measurements the situation under real traffic (different vehicles, different
driving speeds, …) can be analysed. Otherwise pass-by measurements are taking place on a
certain point only, while rolling noise measurements are recording the situation over a certain
length.
Figure 4. Schematic and reality for the test set-up for pass-by measurements according to
ISO 11819 – 1 (adopted from (Haberl and Litzka, 2008))
4. Evaluation of the noise reducing potential of different pavement types
4.1 Test section with different noise reducing asphalt pavements
4.1.1 General statements
In the context of a cooperation of the Austrian Federal Ministry of Transport, Innovation

and Technology (BMVIT), the Austrian Motorway Company (ASFINAG), the Austrian Alpine
Motorway Company (ASG), the Institute for Road Construction and Road Maintenance of
the Vienna University of Technology (ISTU) together with Heller Ingenieurgesellschaft mbH
(Vienna) and representatives of the road construction industry, a concept of a test section
“Noise-reducing road surfaces” was developed with the aim to investigate the noise-reducing
effect of three innovative surface layer types, single and double layer porous asphalt and low
noise stone mastic asphalt.
4.1.2 Location and investigated pavement types
The test section is situated in Tyrol at the motorway A12 – Inntal Autobahn – on the
carriageway direction Innsbruck (km5,80 – km10,40, see Figure 5).
Figure 5. Test section situated in Tyrol, Austria (adopted from (ASFINAG, 2009))
It consists of altogether 8 different test fields with the following surface layer types:
• LSMA 8, rubber modified bitumen, noise-reducing stone mastic asphalt

• LSMA 11, rubber modified bitumen, noise-reducing stone mastic asphalt
• LSMA 11, polymer modified bitumen, noise-reducing stone mastic asphalt
• ZDA, polymer modified bitumen, double layer porous asphalt
• ZDA, rubber modified bitumen (CTS bitumen), double layer porous asphalt
• DA 8, rubber modified bitumen (CTS bitumen), single layer porous asphalt
Additionally 2 reference surface layers, asphalt concrete (pmAB11) and stone mastic
asphalt (SMA11), were laid on the test section.
CTS bitumen is a special product of “Cts Bitumen GmbH” which contains a certain percentage of rubber
ingredients
Detailed information like information about the used binder, binder content, binder
additives, air void content and layer thickness of the investigated surface layer types is given
in Figure 6.
Figure 6. Information about investigated surface courses
4.1.3 Pass-by measurement results
Measurements conducted after a laying period of the surface layers of about 2 years show a
clearly noise-related advantage of the three porous asphalt surface layers (ZDA with polymer
modified bitumen, ZDA with rubber modified bitumen, CTS and DA8 with rubber modified
bitumen, CTS) both related to the overall traffic (SPB-Index) and the individual vehicle
categories. Analyses of the SPB-Index are shown in Figure 7.
Figure 8 shows the noise reduction potential of the investigated road surface courses
compared to stone mastic asphalt (SMA11). The difference in the SPB-Index between the
reference surface types SMA11 and pmAB11, which behave as the loudest, and the two
double layer porous asphalt layer types, which show the quietest noise performance, is about
8dB(A). A noise reduction in this range is really perceivable and audible for human beings. If
one refers to mathematical calculation, a reduction of 6dB(A) may be reached by a reduction
of the vehicle fleet to a quarter (Hoffmann et al., 2003).
Figure 7. SPB-Index of the investigated road surface courses according to the international
standard ISO 11819-1 (adopted from (Haberl and Litzka, 2008))
Figure 8. Noise reduction potential of the investigated road surface courses compared to
stone mastic asphalt SMA 11 (adopted from (Haberl and Litzka, 2008))
4.1.4 Rolling noise measurement results
Rolling noise measurements were accomplished with the two different driving speeds
v=70km/h and v=100km/h 2 years after the laying of the investigated road surface types. The
results for the 8 examined surface layer types are presented in Figure 9.
In general the results of rolling noise measurements show a similar tendency like those
of the pass-by measurements. The examined double layer porous asphalt surface layer types
show the lowest noise volumes; however the LSMA8 with rubber modified bitumen lies
more or less on the same level. This circumstance can be explained with the fact that by the
proximity of the used microphones to the examined tyre and the examined surface layer the
absorption effect of the porous asphalt road surfaces may be underestimated. Within pass-by
measurements the absorption capacity of porous asphalt layers is much more regarded.
The analyses of the influence of different vehicle speeds on the rolling noise emissions
show the same range in the LMA-values regarding the different road surface types.
Figure 9. Results of rolling noise measurements according to the Austrian standard

RVS 11.06.64, LMA-value – the first bar refers to 70km/h and the second to 100km/h (adopted
from (Haberl and Litzka, 2008))
Rolling noise measurements were also conducted according to the international standard
ISO/CD 11819-2. The analyses of the CPX-Indexes show the same tendencies than
demonstrated before regarding the Austrian standard. The results of the CPX- measurements
with driving speeds v=70km/h and v=100km/h are shown in Figure 10.
Figure 10. Results of rolling noise measurements according to the international standard
ISO/CD 11819-2, CPX- Index – the first bar refers to 70km/h and the second to 100km/h
(adopted from (Haberl and Litzka, 2008))
4.2 Long term noise behaviour of road surface courses
An analysis of the noise reduction potential of different road surface courses was
performed by the institute of road construction and maintenance of the Vienna University of
Technology. Within this research project the standard road surface types, eg. porous asphalt
and exposed aggregate cement concrete, situated on the high level road system were examined
(see Figure 11).
Looking on the very first years of the lifetime of the investigated porous asphalt layers
(DA 11) their big noise reduction potential is demonstrated. But after a period of around
5 years their noise emission performance is about the same level as the other investigated
surface courses. This could be an effect of the ongoing clogging processes or of occurring
aggregate losses, especially after hard winter times. In contrast to this the two noise-reducing
surface courses LDDH8 (noise-reducing thin layers) and LSMA8 (noise-reducing stone
mastic asphalt) could be found in the quietest range in all age classes. The remarkable thing is
that the examined exposed aggregate cement concrete layers (WB GK8, blue rectangles) are
situated in the loudest range at the beginning of their lifetime. But regarding the lifetime of
the road surface courses they are advancing to have the best (lowest) noise performance of all
examined road surfaces.
Figure 11. Long term noise behaviour of different road surface courses (Haberl and Litzka,
2005)
5. Summary and Conclusions
In the present paper a test section of different innovative noise reducing asphalt pavements
was explained in detail. The test section is situated in the western part of Austria, in Tyrol, on
the motorway A12 (Inntal Autobahn) and consists of 8 different surface layer types (3 types of
low noise stone mastic asphalt, one single layer porous asphalt, 2 types of double layer porous
asphalt and 2 reference surfaces - asphalt concrete and stone mastic asphalt).
Conducted noise measurements show the clearly noise related advantage of the three
porous asphalt surface layers in comparison with the other investigated road surface courses.
The range between the reference surface types SMA11 and pmAB11, which behave as the
loudest, and the two double layer porous asphalt layer types, which show the quietest noise
performance, is about 8dB(A).
However, a detailed statement both about the noise reducing behaviour and the lifetime
of the examined surface layer types will be possible after investigations during a minimum
investigation period of about 3-5 years only.
6. Bibliography
Haberl J., Einsatzmöglichkeiten von lärmmindernden Fahrbahndeckschichttypen zur

Reduzierung von Straßenverkehrslärm (Use of noise-reducing pavements to reduce traffic
noise, OÖ. Umweltanwaltschaft, Linz, Austria, 2008.
EU- project SILVIA, Guidance manual for the implementation of low-noise road surfaces,
final project report, GROWTH project GRD2-2000-31801 (SILVIA - Sustainable Road
Surfaces for Traffic Noise Control), Brussels, 2004.
RVS 11.06.64, Prüfverfahren – Feldprüfungen - Rollgeräuschmessung (rolling noise

measurements), Forschungsgesellschaft Straße – Schiene - Verkehr, Vienna, April 1997.
Haberl J., Litzka J., Tests of different noise reducing asphalt pavements on an Austrian
motorway, International ISAP Symposium on Asphalt Pavements and Environment,
18.08.2008 - 20.08.2008, Zürich; Schweiz, 2008.
ISO/CD 11819-2, Acoustics - Method for measuring the influence of road surfaces on traffic
noise – Part 2: The Close Proximity Method, International Organisation for Standardization
(ISO), Genf, 2000.
ISO 11819-1, Acoustics - Method for measuring the influence of road surfaces on traffic noise
– Part 1: The Statistical Pass-By Method, International Organisation for Standardization
(ISO), Genf, 1997.
Haberl, J. and Litzka, J., Versuchsstrecke lärmmindernde Straßendecken, A12 (test section for
noise reducing road surfaces, A12), report of the project 3.316, Ministry for Transportation,
Innovation and Technology (BMVIT), Straßenforschung, Vienna , 2008 (unpublished).
http://www.asfinag.at/
Hoffmann, H., von Lüpke, A., 0 Dezibel + 0 Dezibel = 3 Dezibel, Einführung in die
Grundbegriffe und die quantitative Erfassung des Lärms (fundamental terms and
quantitative survey of noise), Institut für Lärmbekämpfung des Hauptverbandes
der gewerblichen Berufsgenossenschaft, Erich Schmid Verlag, 8. Auflage, ISBN
3-503-07470-8, Berlin, Deutschland, 2003.
Haberl J., Litzka J., Bewertung der Nahfeld- Geräuschemission österreichischer

Fahrbahndeckschichten (evaluation of the near- field noise emission on Austrian road
surface courses), Ministry for Transportation, Innovation and Technology (BMVIT)
Straßenforschung Heft 554, Vienna, Austria, 2005.
Field Investigation of Tire/Pavement Noise

and Durability for Asphalt Pavements with
and without Asphalt Rubber
Qing Lu* — Erwin Kohler** —Aybike Öngel***— John T.

Harvey****
* University of California Pavement Research Center

1353 S. 46th St., Bldg. 452T
Richmond, California, 94804, USA
qlu@ucdavis.edu
** Dynatest Consulting Inc

13953 US Highway 301 South
Starke, FL 32091, USA
EKohler@dynatest.com
***Istanbul Kultur University

Department of Civil Engineering
Atakoy Campus, D-100 Yanyol, 34156
Bakirkoy, Istanbul
a.ongel@iku.edu.tr
**** University of California, Davis
Engineering III, Room 3153

One Shields Avenue
Davis, CA 95616, USA
jtharvey@ucdavis.edu
ABSTRACT. This paper presents the results of a three-year field study that compares the
performance trends of various asphalt surface mixes, with and without asphalt rubber.
The performances being investigated include tire/pavement noise, durability, friction, and
permeability. About 80 field sections on California highway system were investigated,
which covered a variety of asphalt surface mixes: rubberized open-graded mixes (RAC-O),
rubberized gap-graded asphalt concrete (RAC-G), open-graded mixes with conventional
binders (OGAC), and dense-graded asphalt concrete (DGAC). Roughness, noise, and surface
distress data were collected on each section for three consecutive years, and permeability
and skid resistance (friction) were collected under traffic closures for the first two years.
Cores were also taken in the first two years from each section to determine air-void contents
and aggregate gradations in the laboratory. The tire/pavement noise was measured by the
On-Board Sound Intensity (OBSI) method. Analysis of the data indicates that RAC-G mixes
provide some noise benefit compared to DGAC. However, the noise benefits are not as great
as those from open-graded mixes. The noise levels from RAC-G mixes appear to approach
those of DGAC after a few years in service. The data also indicate that RAC-O mixes appear
to provide longer noise reduction than OGAC mixes, while both mixes provide noise and
permeability benefits.
KEYWORDS: tire/pavement noise, asphalt pavement, rubber, durability

1. Introduction
The smoothness and quietness of pavements have received increasing attention from
both the public and transportation agencies as issues of quality of life for highway users
and neighboring residents. The concept of using quieter pavements to reduce noise is being
evaluated in California and nationwide over the past several years.
In the last few decades, open graded asphalt concrete (OGAC) surface mixes have been
placed in California and other states to reduce the dangers of hydroplaning and poor visibility
caused by splash and spray during wet weather (Caltrans, 2006). Due to its high air-void
contents and use as a sacrificial wearing course, OGAC mix can also reduce tire/pavement
noise and improve road smoothness. The durability of the mixes, particularly raveling distress,
and the long-term effectiveness of noise reduction, however, have been a concern and results
of different reports vary.
In the last two decades, hundreds of rubberized asphalt concrete (RAC) projects have
been constructed throughout California, the majority of which used asphalt rubber mixes
in overlays for maintenance and/or rehabilitation of existing asphalt concrete and Portland
cement concrete pavements. In California, asphalt rubber is specified to include 18 to 22
percent crumb rubber modifier (CRM) by total mass of the asphalt rubber blend (Caltrans,
2003). The most commonly used asphalt rubber product in California is gap-graded asphalt
rubber mixes (called rubberized asphalt concrete, gap-graded, RAC-G), which can provide
better resistance to reflective cracking and fatigue cracking than standard dense-graded asphalt
concrete (DGAC). The structural and reflection crack retardation equivalencies for RAC-G
allows its substitution for DGAC at about one-half the thickness (Caltrans, 2003). There
was some evidence from field investigation that the tire/pavement noise tends to be lower on
RAC-G pavements than that on the conventional DGAC pavements (Ongel et al. 2008; Lu et
al. 2009).
Another use of asphalt rubber product in California is asphalt rubber open graded
mixes (RAC-O). RAC-O mixes are primarily used as maintenance blankets, overlays
for rehabilitation (including restoration of surface friction), or surface courses for new
construction. They are not considered to be structural elements and are often placed about
24 to 30 mm thick (Caltrans, 2003). Compared to conventional OGAC mixes, RAC-O mixes
contain higher binder content, and can provide better resistance to pavement distresses, such
as reflective cracking, in addition to its noise-reducing property.
Although a variety of mixes are available for pavement maintenance and rehabilitation,
their noise reducing properties and the longevity of these properties compared to other mix
types are not well quantified. There is a need to identify the life of current strategies, as well
as new materials and designs, capable of maintaining smoothness and quietness for the longest
time.
The purpose of this study is to determine the noise levels, friction, smoothness, and
performance trends of different asphalt surface mixes in the field, which would allow for
identification of more durable, smoother, and quieter pavement types.
2. Methodology
2.1. Site Selection
This study presents the analysis of data collected over 3 year from 72 field pavement
sections (each with a length of 150 m) in California. The selection of the pavement sections
is a full factorial experimental design including four different asphalt pavement surface types,
three different age categories, two traffic types, and two rainfall regions. The four mix types
include open graded asphalt concrete with conventional or polymer-modified binders (OGAC),
open graded asphalt concrete with rubberized binder (RAC-O), rubberized gap graded asphalt
concrete (RAC-G), and dense graded asphalt concrete with conventional or polymer-modified
binders (DGAC). Age categories include less than 1 year old, 1–4 years old, and 4–8 years
old. Traffic type is categorized as “high” if the 2004 annual average daily traffic (AADT)
data is greater than 32,000 vehicles/day and is categorized as “low” otherwise. Rainfall is
categorized as “high” if average annual rainfall is greater than 620 mm and is categorized
as “low” otherwise. Such an experimental design provides a balanced selection of pavement
types under various traffic and climate conditions across the state.
2.2. Data Collection
Those selected pavement sections have been tested for three years. The first two years of
data included coring, condition survey, permeability, and friction tests within traffic closures,
profile and tire/pavement noise measuring within normal traffic stream, and mix property
testing on cores in the laboratory. Data collection was continued in the third year, but on a
smaller scale, in which coring, permeability, and friction tests were not conducted.
The tire/pavement noise was measured by the onboard sound intensity (OBSI) method. In
this method two locations of the sound intensity probe are used: one is at the leading edge and
the other at the trailing edge of the tire/pavement contact patch. OBSI measurements are taken
at 97 km/h. An alternative speed of 58 km/h is used if the higher speed cannot be reached
due to constraints of highway geometry or traffic conditions. Three replicate measurements
are collected in three consecutive passes on the sections. Air and pavement temperatures are
also recorded during OBSI measurements. Measurements were conducted using a Goodyear
Aquatread III tire and Dodge Stratus car in the first two years and using a standard reference
test tire (SRTT) and Dodge Stratus car on some sections in the second year and on all sections
in the third year. The first two years OBSI results measured from the Aquatread III tire were
converted into the equivalent measurements from the SRTT using a series of conversion
functions developed in this study, and are all expressed in terms of A-weighted sound intensity
levels, dB(A).
Microtexture was measured using the British pendulum tester according to ASTM E303.
Microtexture measurements were conducted at 25 m intervals and the results were expressed
in terms of British pendulum numbers (BPN). Permeability was measured using a falling-
head permeameter, a device developed at the National Center for Asphalt Technology
(NCAT) (Cooley et al. 2001). Macrotexture was measured according to ASTM E1845, using
a high sampling frequency laser profilometer on the instrumented vehicle used for the sound
intensity measurements. Macrotexture results are reported in terms of mean profile depth
(MPD). Roughness was measured with the inertial laser profiler and reported as international
roughness index (IRI). Pavement condition surveys were conducted following the Caltrans
Condition Survey Manual (version year 2000). A total of 12 cores were also collected, six in
the wheelpath and six between the wheelpath, at 25 m intervals from the selected pavement
sections to determine the air void content and aggregate gradation. Air-void contents were
calculated using the bulk specific gravity value obtained from CoreLokTM measurements and
the theoretical maximum specific gravity value obtained according to ASTM D2041. After the
asphalt from the core samples was burned off in an ignition oven, the aggregate gradation was
obtained by sieve analysis according to ASTM C136 and ASTM C117. Thicknesses of the
cores were also measured and recorded in the laboratory (Ongel et al. 2008).
3. Results
3.1. Roughness
The IRI measurements were collected every meter in both the left and right wheel paths.
The average of the two wheel path measurements along the whole length of each pavement
section was used in the analysis.
Figure 1 shows the boxplots of IRIs in three survey years for different mix types for
three age categories. As the figure shows, IRI generally increases with time. This is expected
because pavement conditions deteriorate with age due to traffic and environmental effect.
However, there are some sections, particularly those OGAC sections, showed little change in
IRI in the three-year survey period. For new mixes (Age Category “<1 year”), IRI was slightly
reduced for OGAC and RAC-O mixes in the second years. RAC-G, on the other hand, showed
significant increase in IRI in the first three years after construction. This is due to an outlier in
the data set.
In general, all pavement sections showed acceptable IRI values based on FHWA criteria
(Ongel et al. 2008), except one old DGAC pavement that has an age of 14 years at the
beginning of the survey.
Multiple regression analysis on all the mixes showed that rubberized open-graded mixes
have lower initial IRI values than nonrubberized open-graded mixes; while rubberized gap-
graded mixes have lower initial IRI values than non-rubberized dense-graded mixes. Among
the three pavement types, OGAC, RAC-G, and RAC-O, they all have lower initial IRI than
DGAC, but only RAC-O is statistically significantly different from DGAC.
Multiple regression analysis on each individual mix showed that open-graded pavements
(both OGAC and RAC-O) are smoother in high temperature regions that in low temperature
regions. For both OGAC and RAC-G pavements, IRI is affected by MPD. IRI increases with
MPD on OGAC pavements, but decreases with MPD on RAC-G pavements. Traffic volume
significantly affects IRI only on RAC-G pavements. Higher traffic volume leads to higher IRI
values.
DGAC OGAC RAC-G RAC-O

4
3
IRI (m/km)
IRI (m/km)
x
x x
x
2
x x x
x x
x x
x
x x x
x
x
x x x x
x
x x
x x x x
x x
xxx x x x
1
Survey 1231231 2 312312312 3123123123 12 31 2 3 123

211
412
713
212
213
221
222
223
231
232
233
411
413
421
422
423
433
431
432
611
612
613
621
622
623
722
731
631
632
633
711
712
721
723
732
733
Year
Phase ID, Age Category, Mix type
Age <1 1-4 >4 <1 1-4 >4 <1 1-4 >4 <1 1-4 >4
Category
Figure 1. Comparison of IRI values for different mix types at different ages for first, second,
and third survey years
3.2. Macrotexture
In this study, macrotexture is characterized by mean profile depth (MPD). Figure 2 shows
the boxplots of MPDs in three survey years (“Phase” as shown in the figure) for different
mix types for three age categories. As the figure shows, MPD generally increases with
pavement age for the same pavement section. Except for a few outliers, this increase trend is
also obvious among different pavement sections of the same mix type. Without the outlier in
RAC-G mixes, RAC-G mixes have higher MPD values than the dense-graded mixes, while
the open-graded mixes have higher MPD values than the RAC-G mixes. Among the two open-
graded mixes, OGAC mixes have higher MPD values than RAC-G mixes.
Multiple regression analysis on all the mixes showed that at the 95 percent confidence
level, age, mix type, and number of high-temperature days significantly affect macrotexture.
MPD increases with age, but decreases with the number of high-temperature days. P-values
for the interaction terms between Age and Mix Type showed that the growth rate (with age) of
MPD of OGAC pavements is significantly higher than that of DGAC pavements. The growth
rates of MPD of RAC-G and RAC-O pavements are not statistically different from that of
DGAC pavements.
Multiple regression analysis on each individual mix showed that within each mix type, air-
void content has no significant effect on the value of MPD. Fineness modulus is significant
in affecting the macrotexture of open-graded pavements, including both OGAC and RAC-O,
marginally significant in affecting the macrotexture of DGAC pavements, and insignificant
for RAC-G pavements. Generally macrotexture increases with fineness modulus. The effect
of pavement age on macrotexture is much more prominent (in terms of both statistical
significance and practical significance) on non-rubberized pavements (DGAC and OGAC)
than on rubberized pavements (RAC-G, and RAC-O).
3.3. Tire/Pavement Noise
The on-board sound intensity (OBSI) averaged from three consecutive passes is used in
the analysis. The OBSI results are given in terms of spectral content in one-third octave bands.
Summation of the one-third octave band noise levels gives the overall A-weighted sound
intensity levels. Because sound intensity measurements are highly affected by test car speed,
air density, and type of test tire, the original OBSI results were all converted to equivalent
values at the same test conditions before analysis.
Figure 3 shows the box plots of overall OBSI in three years for different mix types for
three age categories. As the figure shows, the overall tire/pavement noise generally increases
with pavement age. For newly paved pavements, the overall sound intensities measured on
OGAC, RAC-G, and RAC-O pavements are lower than the values measured on the DGAC
pavements. After the pavements were exposed to traffic, the overall sound intensity measured
on RAC-G pavements quickly approached the representative value measured on DGAC
pavements. The overall sound intensity measured on the OGAC pavements kept stable for
about five years and then increased quickly with pavement age. With a few exceptions, the
overall sound intensity measured on the RAC-O pavements kept stable for about seven years
and then increased quickly with pavement age. Based on these observations, the rank of the
four mix types in terms of noise reduction is RAC-O, OGAC, RAC-G, and DGAC (from best
to worst).

2000
x x
x
1500
(micron)
x
MPD (micron)
x x x
x x
x x
MPD
x x x
x x
x
1000
x x x x x x
x x x
x x
x
x x
x x
x x
x
500
Survey
Phase ID 1122 33 11 223311 22 33 1
1 22 33 1
1 2233 1122 331122 33 1
1 22 33 112 23 3 112 2 33 11 22 33 11 22 33
212
221
222
223
231
232
233
413
421
423
433
613
622
623
631
632
711
713
723
732
211
213
411
412
422
431
432
611
612
621
633
712
721
722
731
733
Year
Age Category <1 1-4 >4 <1 1-4 >4 <1 1-4 >4 <1 1-4 >4
Age <1 1-4 >4 <1 1-4 >4 <1 1-4 >4 <1 1-4 >4
Category
Figure 2. Comparison of MPD values for different mix types at different ages for first, second,
and third survey years
Regression analysis was conducted to determine the effects of mix properties, distresses,
traffic, and weather conditions on the sound intensity levels. A single variable regression
analysis was first conducted to check the correlation between the dependent variable and each
independent variable, and then a multiple regression model was estimated to consider the
effects of various variables simultaneously.
In the third year survey, air-void content, permeability, and friction (BPN) were not
measured in the field. To use these variables in the regression analysis, the third-year data
were extrapolated from the first two-year data. Boxplots of these three variables are shown
in Figure 4 through Figure 6. It can be observed from the boxplots that the air-void content
generally reduces with time for all mixes, and the in-situ permeability reduces with time for
OGAC, RAC-G, and RAC-O mixes. For RAC-G pavements, the in-situ permeability is as
high as at a comparable level to that of open-graded pavements in the first three years after
construction, but rapidly reduces to a near-zero level after 4 or 5 years old. Surface friction
(BPN) tends to increase slightly with pavement age. Based on these observations, linear
extrapolation was applied to estimate the third-year values of air-void content, permeability,
and BPN from the first two years’ data.
106
x
104
x
x
x
x x
OBSI(dBA)
(dBA)
x x x
x
102
x
x x
OBSI
x x x
x x x x x
Overall
xx
x x x
x x
Overall
x
100
x x x x
x
x
x
98
96
Survey 1231231 2 312312312 3123123123 12 31 2 3 123

Year
Age <1 1-4 >4 <1 1-4 >4 <1 1-4 >4 <1 1-4 >4
Category
Figure 3. Comparison of overall OBSI values for different mix types at different ages for first,
second, and third survey years

20 x
x x
x x
x
x
Content(%)
(%)
15
x x
x
Content
x
Air-void
x
x
Air-void
10
x x
x
x
x
x x x
x
x
5
x
411
632
422
621
711
731
211
212
221
222
231
232
412
421
431
432
611
612
622
631
712
721
722
732
Survey 1 2 1 2 1 2 1 2 1 2 1 2 1 2 12 1 2 1 2 1 2 1 2
Year
Age <1 1-4 >4 <1 1-4 >4 <1 1-4 >4 <1 1-4 >4
Category
Figure 4. Boxplots of air-void content for different mix types at different ages for first and
second survey years

80
x
70
x x x x x
BPN
BPN
x x x
x x x x x
x
60
x x
x x
x x
x x
x
50
40
411
431
711
731
Survey 11 22 11 2
231
1
1 2
232
2 11 2
2 11 22 1 1 22 1 1 2 2 11 2
611
621
622
11 22 11 22 1
632
712
22
722
11 22
211
212
221
222
412
421
422
432
612
631
721
732
Phase ID 2 2 1
Year
Age Category <1 1-4 >4 <1 1-4 >4 <1 1-4 >4 <1 1-4 >4
Age <1 1-4 >4 <1 1-4 >4 <1 1-4 >4 <1 1-4 >4
Category
Figure 5. Boxplots of BPN for different mix types at different ages for first and second survey
years
0.30

0.25
(cm/sec)
0.20
Permeability (cm/s)
Permeability
0.15
x
x
x
0.10
x
x x
0.05
x x x
x
x x x x
x
x x x x x x x x x
0.0
Survey
Phase ID 11 22 11 22 11 22 11 22 11 22 11 2
2 11 22 11 22 11 2
2 11 22 1
1 22 1
1 22
211
212
222
231
411
412
422
431
611
621
631
632
712
722
731
732
221
232
421
432
612
622
711
721
Year
Age Category <1 1-4 >4 <1 1-4 >4 <1 1-4 >4 <1 1-4 >4
Age <1 1-4 >4 <1 1-4 >4 <1 1-4 >4 <1 1-4 >4
Category
Figure 6. Boxplots of permeability for different mix types at different ages for first and second
survey years
After excluding highly correlated independent variables, the following multiple regression
model was estimated:
Overall Sound Intensity =103.4440+0.1488 × Age − 2.2821× ind ( MixTypeOGAC ) − 1.6271× ind ( MixTypeRAC − G )
−2.8749 × ind ( MixTypeRAC − O) − 0.0233 × NMAS − 0.0252 × Thickness − 0.0018 × NumberOfDays > 30C
−1.0635 ×10−6 × Age × AADTinCoringLane + 0.9139 × Raveling + 0.5677 × Rutting + 0.0348 × Age × ind ( MixTypeOGAC )
+0.0849 × Age × ind ( MixTypeRAC − G ) − 0.0384 × Age × ind ( MixTypeRAC − O)
(1)
where is an indicator function, 1 if the variable in the parentheses is true and 0 if

false. The estimated values and P-values of the parameters are shown below:
Table 1. Multiple regression analysis results for overall OBSI
Value Std. Error t value P-value

(Intercept) 103.4440 0.9063 114.1421 0.0000
Age 0.1488 0.0758 1.9621 0.0500
PvmntTypeOGAC -2.2821 0.5856 -3.8968 0.0001
PvmntTypeRAC-G -1.6271 0.5268 -3.0886 0.0024
PvmntTypeRAC-O -2.8749 0.5251 -5.4745 0.0000
NMAS -0.0233 0.0480 -0.4849 0.6284
Thickness -0.0252 0.0068 -3.6850 0.0003
NoDaysTempGT30 -0.0018 0.0022 -0.8497 0.3968
AgeAADTperLane -1.0635 0.0000 -0.3473 0.7289
Raveling 0.9139 0.2594 3.5229 0.0006
Rutting 0.5677 0.3921 1.4477 0.1497
AgePvmntTypeOGAC 0.0348 0.0988 0.3518 0.7255
AgePvmntTypeRAC-G 0.0849 0.1041 0.8155 0.4160
AgePvmntTypeRAC-O -0.0384 0.0936 -0.4100 0.6823
Residual standard error: 1.286 on 154 degrees of freedom; Multiple R-Squared: 0.5518
It can be seen that at the 95 percent confidence level, age, mix type, surface layer
thickness, existence of raveling significantly affect the overall sound intensity. The overall
sound intensity increases with pavement age and the existence of raveling distress, but
decreases with the surface layer thickness. Among the three pavement types, OGAC, RAC-G,
and RAC-O, they all have lower initial overall sound intensity than DGAC. The average noise
reductions (compared to DGAC pavements) for newly placed OGAC, RAC-G, and RAC-O
pavements are about 2.3, 1.6, and 2.9 dB(A), respectively.
The interaction terms between age and mix type are not statistically significant, which
indicates that the growth rate of overall sound intensity is not statistically different among
the four pavement types. This conclusion is different from the direct observations from
Figure 3. This is mostly due to the constraints applied by the multiple regression analysis.
The regression analysis assumes a linear increase of noise with age for all mixes, but Figure
3 indicates that the noise development on open-graded mixes is more like piecewise linear.
Use of different growth function forms for different mixes in the same regression model
significantly increases the complexity of parameter estimation and result interpretation,
which is not attempted in this study. Considering the total noise increase during the pavement
life covered by the data set in this study (about 10 years), the estimated parameters of the
interaction terms indicate that the noise increase is higher on OGAC and RAC-G pavements
than on DGAC pavements, and the lowest on RAC-O pavements.
3.4. Pavement Distresses
In the third-year survey, pavement conditions were evaluated in a way different from
the way used in the previous two years. In the first two years’ survey, the truck lane was
temporarily closed and pavement conditions were measured on site during the traffic closure.
During the third-year survey, however, lanes were not closed. Instead, high-resolution digital
photos were taken from the shoulder along the whole length of each section, and pavement
conditions were assessed afterwards in the office, based on the pavement surface images. It
has to be noted that some distresses, such as rutting, could not be evaluated accurately solely
based on surface images. Because of the different ways of distress assessment in the first two
years and the third year, some distresses were recorded as less severe in the third year than the
previous years. A basic assumption was made in post-processing the distress data, that is, the
third-year distress is no less than the second year.
In this study, four major distress types, including bleeding, transverse/reflective cracking,
raveling, and wheel path cracking, were analyzed.
Figure 7 shows the percentage of bleeding area measured in three consecutive years for
individual pavement sections of four mix types. In this figure, bleeding includes all three
severity levels (low, medium, and high). The figure shows that bleeding may appear two to
four years after construction on all pavement types, and it tends to appear earlier on rubberized
pavements than on nonrubberized pavements. Among the four mix types, RAC-G pavements
seem to be most susceptible to bleeding distress in terms of both the time of occurrence and
the extent of distress. Multiple regression analysis reveals that bleeding increases with age,
number of wet days, number of high-temperature days, and cumulative truck traffic, but
decreases with the number of freeze-thaw cycles. Among the four pavement types, OGAC
and RAC-O pavements have no significant difference from DGAC pavement, but RAC-G
pavement is significantly (statistically) more prone to bleeding.
Because all the sections investigated in this study are overlays of AC or PCC and it is
difficult to distinguish the thermal and reflective cracking mechanisms based only on surface
condition observations, the analysis in this study combines the thermal cracking and reflective
cracking as one distress type. Figure 8 shows the average length of transverse/reflective
cracking (at all severity levels) per unit length of pavement observed on each pavement
section in three years.
60
60
Bleeding (%)
Bleeding (%)
40
40
20
20
DGAC OGAC
0
0
0 5 10 15 20 0 5 10 15 20
Age (year) Age (year)
60
60
Bleeding (%)
Bleeding (%)
40
40
20
20
RAC-G RAC-O
0
0 5 10 15 20 0 5 10 15 20
Figure 7. Bleeding development trend in three years for various pavements
It can be seen that transverse/reflective cracking generally propagates with pavement age.
The transverse/reflective cracks seem to initiate earlier and propagates faster on the rubberized
asphalt pavements (RAC-G and RAC-O) than on the nonrubberized pavements (DGAC and
OGAC). This observation seems to suggest that use of rubber does not help prevent cracking.
However, evidence from another study has indicated that the increased cracking in the rubber
mixes may be biased by the condition of the underlying pavements, because RAC-G and
RAC-O mixes tend to be placed more on pavements with greater extent of cracking (Lee et al.
2007).
Figure 9 shows the percentage of area with raveling (including all severity levels) in the
three survey years. It can be seen that raveling may occur on all types of pavements, and
in general, raveling starts earlier on DGAC and RAC-G pavements than on open-graded
pavements. Pavements overlaid with DGAC mixes seem to experience more raveling than
pavements overlaid with other mixes (OGAC, RAC-G, and RAC-O). Raveling in the two
open-graded mixes tend to initiate and develop quickly after five years in service.
Trans./Ref. Cracking (m/m)

0.8
0.8
0.4
0.4
0.0
0.0
DGAC OGAC
0 5 10 15 20 0 5 10 15 20

0.8
0.8
0.4
0.4
0.0
0.0
RAC-G RAC-O
0 5 10 15 20 0 5 10 15 20
Figure 8. Transverse/reflective cracking development trend in three years for various

pavements
In the condition survey, all the cracks in the wheelpath were recorded as fatigue cracks,
because no data is available to determine whether they were caused by reflection. Fatigue
cracking was evaluated as the areas of cracking at three severity levels (low, moderate and
high).
Figure 10 shows the percentage of area with fatigue cracking (including all severity levels)
in the three survey years for four pavement types. It can be seen from the plots that fatigue
cracking may occur on all types of pavements, and in general, it increases with pavement age.
Limited data indicate that fatigue cracking seems to initiate earlier on DGAC and RAC-G
pavements than on open-graded pavements.
Multiple regression analysis shows that at the 95 percent confidence level, pavement age,
existence of underlying PCC slabs and cumulative truck traffic are significant in affecting
fatigue cracking. The estimated parameters indicate that fatigue cracking increases with
pavement age and cumulative truck traffic. The existence of underlying PCC slabs increases
the potential of fatigue cracking in the surface layer. Pavement type is an insignificant factor,
indicating there is no significant difference in the fatigue performance of the four mix types.
120
120
Raveling (%)
Raveling (%)
0 20 40 60 80
0 20 40 60 80
DGAC OGAC
0 5 10 15 20 0 5 10 15 20
120
120
Raveling (%)
Raveling (%)
0 20 40 60 80
0 20 40 60 80
RAC-G RAC-O
0 5 10 15 20 0 5 10 15 20
Figure 9. Raveling development trend in three years for various pavements
4. Conclusions
In this study, field data regarding tire/pavement noise, surface condition, ride quality, and
macrotexture have been collected for three consecutive years from asphalt pavements placed
with rubberized and nonrubberized mixes used in California. Analysis of the three-year data
revealed the following findings:
• All four mixes investigated provide acceptable smoothness for riders. However, there
is some evidence showing that inclusion of rubber in the mix can provide a pavement
with smoother surface.
• The macrotexture of rubberized asphalt pavements is generally smaller, and less

affected by age, than that of nonrubberized asphalt pavements.
• Compared to the average noise level of a DGAC mix, the newly placed asphalt mixes
reduce the noise by about 2.3 dB(A) for OGAC, by about 2.9 dB(A) for RAC-O,
and by about 1.6 dB(A) for RAC-G. After the pavements are exposed to traffic, this
noise reduction benefit generally changes slightly for about five to seven years and
then begins to lose quickly with pavement age. Inclusion of rubber in the open graded
mixes tends to extend the noise-reducing property for two years longer. In other words,
inclusion of asphalt rubber in the open-graded mixes can extend the noise-reducing
duration.
• RAC-G mixes are initially quieter than DGAC mixes due to its higher initial
permeability. After being opened to traffic, the permeability of RAC-G mixes reduce
significantly to near zero values, and the noise level increases quickly to near the
values measured on DGAC mixes.
• Bleeding seems to be a major distress type on asphalt rubber mixes, particularly on

RAC-G mixes.
• Inclusion of rubber does not seem to change the ravelling potential of open-graded
mixes.
• If all other conditions are the same, the study suggests that RAC-O mixes can provide
longer noise reduction benefit than conventional OGAC mixes. However, the binder
content of RAC-O needs to be selected carefully to prevent excess bleeding distress.
10 20 30 40 50 60
10 20 30 40 50 60
Fatigue Cracking (%)
DGAC OGAC
0
0 5 10 15 20 0 5 10 15 20
10 20 30 40 50 60
10 20 30 40 50 60
RAC-G RAC-O
0
0 5 10 15 20 0 5 10 15 20
Figure 10. Development trend of fatigue cracking in three years for each pavement section
Acknowledgements
The authors would like to thank California Department of Transportation (Caltrans) for
providing funding for this research. The opinions and conclusions expressed in this paper,
however, are those of the authors and do not necessarily represent those of Caltrans.
5. Bibliography
California Department of Transportation (Caltrans). Open Graded Friction Course Usage

Guide. Sacramento, California, 2006.
California Department of Transportation (Caltrans). Asphalt Rubber Usage Guide.

Sacramento, California, 2003.
California Department of Transportation (Caltrans). Quieter Pavements Road Map and Work
Plan. 2005.
A. Ongel, J. Harvey, E. Kohler, Q. Lu, and B. Steven. Investigation of Noise, Durability,

Permeability, and Friction Performance Trends for Asphaltic Pavement Surface Types:
First- and Second-Year Results. UCPRC-RR-2007-03, University of California Pavement
Research Center, California, 2008.
Cooley, Jr., L. A., E. R. Brown, and S. Maghsoodloo. Development of Critical Field

Permeability and Pavement Density Values for Coarse-Graded Superpave Pavements.
NCAT Report No. 01-03. National Center for Asphalt Technology, Sept. 2001.
Lee, C., W. Nokes, and J. Harvey. Performance of Pavement Preservation Treatments.

University of California Pavement Research Center, Davis and Berkeley, 2007.
Lu, Q., E. Kohler, J. Harvey, and A. Ongel. Investigation of Noise and Durability Performance
Trends for Asphaltic Pavement Surface Types: Three-Year Results. UCPRC-RR-2009-01,
University of California Pavement Research Center, California, 2009.
Environmental, Energy Consumption and

CO2 Aspects of Recycled Waste Tires Used
in Asphalt-Rubber
Jorge Sousa — George B. Way — Douglas Carlson
Consulpav
Recycle Tire Engineering and Research Foundation.
Rubber Pavements Association
jmbsousa@aol.com
wayouta@cox.net
doug.carlson@rubberpavements.org
ABSTRACT: Waste tires are recognized as one of the most difficult waste products to
manage in a modern society. They are not difficult individually, but are difficult collectively;
particularly when world wide almost a billion such tires are disposed of annually. Waste tires
are generated in industrialized societies at an annual rate equal to the human population
which discards them, one scrap tire per person per year. The lack of adequate disposal
methods and management systems in years past had lead to wide spread, cumbersome
collection of waste tires in unmanaged or poorly managed waste tire piles. Problems
associated with waste tire piles typically are: threat of fire and related environmental damage
from a tire pile fire and the potential increase in vectors and pests. Secondary problems are
that tire piles require substantial volume or space prior to any type of processing and are an
eyesore. One of the uses of waste tires is to recycle them into asphalt-rubber. In this paper, the
environmental, energy consumption and CO2 aspects of using ground tire rubber from waste
tires in asphalt-rubber is considered and evaluated. Various environmental studies of asphalt-
rubber are reviewed that demonstrate that asphalt-rubber is an environmentally acceptable
paving material. Also, the end uses of ground tire rubber in terms of energy consumption
from waste tires are analyzed in terms of: shredding for use in landfills as Alternate Daily
Cover, shredding for use as tire derived fuel, and crumb rubber production with an end
use in asphalt-rubber concrete pavements. This paper also touches upon a new aspect of
environmental concern recognized in the Kyoto treaty and recently ruled upon by the United
States Supreme Court namely CO2 emissions. The major goal of this paper is to investigate
the overall benefit to Society for each aspect of the use of asphalt-rubber.
KEYWORDS: scrap tires, fuel, energy, asphalt rubber, Kyoto treaty

1. Foreword
This paper is a continuation and updating of a paper presented at the AR2006 Conference
(Sousa, 2006). Since 2006 there has become a greater environmental concern about green
house gases namely CO2. This was first recognized in the Kyoto treaty and more recently
ruled upon by the United States Supreme Court in a decision where CO2 was judged to be a
pollutant.
2. Introduction
The disposal of scrap tires continues to be a major waste management issue. Approximately
300 million tires are disposed of annually each year in the USA. Tire piles can be a nuisance
as well as health hazard because they are a breeding ground for mosquitoes and harbor various
vermin. From time to time tire piles catch on fire and cause considerable environmental
damage, Figure 1.
Figure 1. The Wrong Type of Tire EnergyConsumption
Because of the huge number of tires that accumulate annually they must be managed and
processed in some way to prevent the build up of scrap tire piles, Figure 2. Many methods of
disposal or end uses of scrap tires have evolved over the years, Figure 3. There are many broad
categories of utilization. The burning of tires or derived fuel or tire derived fuel represents
the single greatest usage of about 40 percent of all the waste tires. Landfill (Alternate Daily
Cover) and civil engineering combined represent the next greatest amount representing about
24 percent of the waste tires. Ground rubber is about 10 percent of the use and is commonly
used in AR.
Figure 2. Scrap tires before processing

Figure 3. Utilization of scrap tires in the USA
3. Study Objectives
The objectives of this paper are to compare the energy consumption or saving of three
common end uses of scrap tires and to estimate the degree of CO 2 reduction associated
with use of asphalt rubber (AR). The three common end uses include shredding for use as
Alternate Daily Cover (ADC) in landfills, shredding for use as a Tire Derived Fuel (TDF)
in a combustion process and crumb rubber production with an end use in AR pavements.
The scope of the energy consumption examination is to discuss the potential energy use or
recovery benefits of each method. It should be noted that all three methods are currently
in use and serve the intended purpose of removing scrap tires from the waste stream and
represent about 74 percent of the consumption of waste tires. There are many methods of
scrap tire disposal that can be used; these three were chosen to represent the range of options.
Which method or a mix of methods used by a governmental entity to dispose of scrap tires
is a function of many factors not necessarily just the potential energy recovery benefits.
Nevertheless, using energy recovery benefits is a first start in judging the overall value of each
method to society in general.
4. Background of Analysis
The technical approach taken in this paper is consistent with a study conducted by the
Argonne National Laboratory in 1979 for the United States Department of Energy entitled
“Discarded Tires: Energy Conservation Through Alternative Uses,” (Gaines, 1979). At that
time there was an energy crisis and the usefulness of tires as a fuel source was carefully
examined. Also at that time waste disposal of tires was not an issue and air pollution
regulations were not as strict as they are today. In light of these changes, and others such as
potential global warming and green house gases namely CO2 and the future of the Kyoto
Treaty, which occurred over the past 21 years it seemed appropriate to again review this
somewhat controversial topic in some detail.
5. Analysis
For each of the three disposal methods a combustion heat fuel scorecard was created. Many
of the values were derived from the Argonne Laboratory study. Other values were obtained
from various industry sources for aggregate, steel, hauling (trucking) and tire shredding and
grinding. Table 1 is a list of typical heat combustion values for common fuels.
Table 1. Combustion Heat of Fuel
Fuel kJ/kg heat value BTU/lb heat value

Coal 26000 11000
Tire 35000 15000
Asphalt 35000 15000
Natural Gas 172000 74000
Propane 214000 92000
Gasoline 233000 100000
Diesel 319000 137000
In this study scrap tires and asphalt have the same heat of combustion value which is
slightly greater than the heat of combustion value of coal. As can be seen coal, scrap rubber
and asphalt are all at the low end of heat value. Presently, modern power generating plants
typically use natural gas as much as possible to generate electricity and meet very demanding
air pollution requirements.
6. Alternate Daily Cover
The first disposal method for scrap tires that is analyzed is called Alternate Daily Cover
(ADC). ADC involves the placement of rubber tire shreds generally about 150 mm (6 in)
square or larger being placed in a landfill to cover the daily refuse pile or layer to a depth of
150 mm (6 in), or as a light weight civil engineering fill, Figure 4 and 5.
Figure 4. Tire shreds for alternate daily cover

Figure 5. Tire shreds being placed as cover or fill
This process requires the least amount of energy of the three options. Table 2 is the heat of
combustion values for ADC.
Table 2. Utilization for Alternate Daily Cover
Process kJ/kg heat value BTU/lb heat value

Tire Shredding -93 -40
Shred Transportation -1744 -750
Gain/Loss -1837 -790
ADC is composed of scrap tires that have been shredded into approximately 100-150 mm
(4-6 in.) square tiles that are spread to a depth of 150 mm (6 in.) atop a sanitation landfill pile
at the end of each day. Regional specifications can vary on the shred size and layer depth.
This lightweight cover keeps loose material from blowing away. There is no net energy
benefit since it takes energy to shred the tires and transport the shreds and place them. The net
negative use of energy is small and under the right circumstances may be an appropriate use
of shredded tires.
7. Tire Derived Fuel
TDF is composed of whole scrap tires or shredded tires that are introduced into a coal
fired furnace to add extra heat, Figure 6.
Figure 6. Tire derived fuel whole tires combusted in coal fired cement kiln
Table 3 is an example of the heat of combustion values for TDF. Tire chips can combust
with fewer emissions than coal. In many locations, tire chips are used to help reduce the total
emissions output. There is a net positive gain in energy for the rubber used. This is a good
use of scrap tires and presently consumes about 120 million tires in the United States (Scrap,
2000).
Table 3. Utilization for Tire Derived Fuel

Combustion Energy +34900 +15000
Gain/Loss +31420 +13500
8. Asphalt Rubber
AR is composed of crumb rubber derived from the commutation of scrap tires. Table 4
represents the heat of combustion values for crumb rubber modifier (CRM) used in A-R,
Figure 7.
Figure7. Typical asphalt rubber mix paving
Table 4. kJ/kg Utilization for Asphalt rubber

Granulation -3600 -1540
Crumb Rubber Transport -1740 -750
Steel Recovery +1900 +820
Asphalt Saved +209000 to +465000 +90000 to +200000
Aggregate Saved +108000 +46000
Gain/Loss +310080 to +566080 +133030 to +243030
The crumb rubber is the size of coffee grounds and is derived from either commutation
by mechanical grinding, commonly called ambient grinding or from cryogenic commutation
using liquid nitrogen, commonly called cryogenic grinding, Figure 7. Modern plants often
employ a combination of both cryogenic and ambient technologies to obtain the most
economical product. The crumb rubber is combined with liquid asphalt and then combined
with aggregate materials and placed using conventional paving equipment. As this table shows
there is a net positive gain in energy for the rubber used.
Figure 7. Crumb Rubber after grinding to a 10 mesh or finer
The 310080 kJ/kg (133030 BTU/lb) value is consistent with the previous Argonne
Laboratory finding in 1979. In 1979 the Argonne Laboratory derived their value by examining
the use of an AR chip seal and assigning its energy savings in terms of less asphalt concrete
overlays would be needed over the life of the pavement. Since 1979, AR is now commonly
used as a binder in hot mixes in the states of Arizona, California, Texas and Florida. Energy
savings are now the result of using less than one half the thickness of routine paving material
as reported by Arizona (Way, 2000) and California (Van, 2000). The 310080 kJ/kg (133030
BTU/lb) energy savings refers to a two inch A-R overlay being used in place of a normal four
inch asphalt pavement overlay. The 566080 kJ/kg (243030 BTU/lb) energy savings refers to
an one inch open graded AR mix being placed on top of a concrete pavement in place of a
normal five inch asphalt pavement overlay.
Other energy savings that have occurred since 1979 include aggregate savings. In many
parts of the United States and Europe good quality road building aggregate is in short supply
and harder to obtain. The 108000 kJ/kg (46000 BTU/lb) energy savings refers to the mining
energy and transport energy associated with using thicker pavements compared to the thinner
AR pavements. The reclaiming of steel from tires also has considerable value not recognized
in 1979. In all the energy savings by using AR is very impressive. These energy savings
coupled with other AR benefits including less cracking, less maintenance and less noise
(Bollard, 1999) make this a very attractive and beneficial end use of scrap tires in a highway
environment. Life cycle cost analysis encouraged by the Federal Highway Administration
has shown that a substantial dollar savings can be obtained over the expected life of a project
when AR paving strategies are employed (Hicks, 2000).
9. Green House Gas CO2
Since 2006 there has become a greater environmental concern about green house gases
namely CO2. This was first recognized in the Kyoto treaty and more recently ruled upon by
the United States Supreme Court in a decision where CO2 was judged to be a pollutant, Figure
8 (US, 2007).
Figure 8. US Supreme Court ruling on CO2 as a pollutant
In examining the previous savings in energy it is possible to estimate how much CO2 can
be saved by using a thinner AR pavement. Very large amounts of CO2 can be saved by using
AR pavements which could be useful in future years as the importance of reducing CO2
becomes more significant.
Table5. CO2 Savings with AR 2 to 1 Reduced Thickness of Pavement
Use of 50 mm (2 in.) overlay SI Units US Customary

compared to 100 mm (4 in.) overlay Lane km Lane mile
Crumb Rubber kJ/kg or BTU/lb 310080 kJ/kg 133030 BTU/lb
Crumb Rubber/lane km or mile 5455 kg 20000 lbs
AR Energy/lane km or mile 1691362368 kJ 2660600000 BTU
Energy/ Liter or Gallon of Gasoline 0.0000076 kJ/Liter 0.000008 BTU/Gallon
Equivalent Liters or Gallons Saved 12800 Liters 21285 Gallons
CO2 Saved/ kg or lbs 10 kg 22 lbs
Kg or Lbs of CO2 Saved 128000 kg 468270 lbs
Tons of CO2 Saved per lane km or mile 128 tonnes 234 tons
10. Possible long term energy and CO2 savings over time
In addition to the energy and CO2 savings associated with the construction of thinner
AR pavements previous studies have shown that AR pavements can be smoother over time
(Way, 2009) as shown in Figure 9. All other things being equal smoother pavements mean
less energy consumption and therefore less CO2 emissions. As can be seen in Figure 9 after
the sixth year of service AR pavements are consistently smoother than conventional HMA
pavements by about 200 mm per km. The World Bank has developed models to relate
energy savings versus the smoothness of the pavement (World, 2001). Using their empirical
relationship the average savings in fuel for a passenger car for the 6 to 15 year period would
be approximately 0.08409 liters per 1000 km (0.03575 gallons per 1000 miles). Just as a
theoretical example of what this means 100000 cars traveling 10000 miles a year over the nine
year period of smoothness performance (year six to fifteen) the AR pavements would save a
total of about 300000 gallons of gasoline. In addition the total CO2 savings over that nine year
period would be 3300 tons of CO2.
Figure 9. Pavement smoothness in mm/Km versus years of service for HMA and asphalt-
rubber pavements
11. Discussion
The above academic exercise demonstrates the wide range of energy usefulness that scrap
tires have to offer society in general but what about the practical side of the issue? With this in
mind, experiences in the State of Arizona may be of value in discussing the pros and cons of
uses of the methods previously discussed. In the mid 1970’s tires were combusted in copper
smelters in Arizona. As air pollution laws changed Copper smelters found it more difficult to
operate in Arizona and thus by the mid 1980’s all copper smelters were shut down in Arizona.
Copper from Arizona mines is now smelted in Mexico. In the mid 1980’s cement plants
burned tires. Due to environmental concerns the cement plants decided to end the burning of
tires even though such burning can be legally permitted. Thus by 1990 no tires were being
burned in Arizona. Coincidentally the Arizona Department of Transportation (ADOT) along
with cities and counties in the state began to routinely use AR as an engineered binder in
pavements in 1988. Since that time ADOT alone has used over 10 million tires in pavements.
There was and is no special program to reuse tires in pavements in Arizona, however the
State law does encourage recycling of tires as the highest priority. Approximately 70 percent
of the scrap tires in Arizona now go into pavements, with the remainder going into various
commercial products.
11. Conclusion
Scrap tires can be used in many ways. It is important to Society to encourage their use
and removal from the waste stream. The three processes discussed meet society’s need of
preventing tire piles from accumulating and exposing the ecosystem to unnecessary risks of
increased pollution and pests. However, the potential energy used or saved in tire processing
should also be examined. Table 6 summarizes the energy values of the three tire processes
discussed in the paper.
Table 6. Comparison for the three scrap tire disposal methods
Tire Derived Material kJ Gain/loss/kg Rubber BTU Gain/loss/lb Rubber

Alternate Daily Cover -1837 -790
Tire Derived Fuel +31420 +13500
Crumb Rubber Modifier
+310080 to 566080 +133030 to +243030
In Asphalt Rubber
Besides the potential energy savings gained by using granulated tire rubber as a modifier to
asphalt pavement, it should be noted that this process can substantially improve the highway
assets maintained by our communities.
12. Acknowledgment
The authors express their appreciation to the following for their input, contributions and
research assistance:
Don Stout of FNF Construction, Inc.

Eric J. Stuart of the Steel Manufacturers Association
Michael Blumenthal of the Scrap Tire Management Council
Jim Anderson of Recovery Technologies Group
Muarry Quance of BAS Recycling
13. References
Bollard, P. “Report on the Status of Rubberized Asphalt: Traffic Noise Reduction In

Sacramento County”, Bollard & Brennan, Inc. Consultants in Acoustics and Noise Control
Engineering prepared for the Sacramento County Public Works Agency - Transportation
Division 1999.
Hicks, R.G.; Epps, J. A., “Life Cycle Costs For Asphalt Rubber Paving Materials”,
Proceedings of the Asphalt Rubber 2000 Conference, Vilamoura, Portugal, 2000.
Gaines, L.L.; Wolsky, A.M., “Discarded Tires: Energy Conservation Through Alternative
Uses”, Argonne National Laboratory for the United States Department of Energy 1979.
Scrap Tire Management Council, “1990-2000 Ten Years of Success”, Rubber Manufacturers
Association Publication: Washington, D.C., 2000.
Sousa, J, G. Way, and D. Carlson, “Energy Consumption of Alternative Scrap Tire Uses,”
Proceedings AR2006 International Conference on Asphalt Rubber, Palm Springs,
California, IBSN: 962-405-091-0, 2006.
Van Kirk, J.; Holleran, G., “Reduced Thickness Asphalt Rubber Concrete Leads to Cost
Effective Pavement Rehabilitation”’ Proceedings of the Asphalt Rubber 2000 Conference,
Vilamoura, Portugal, 2000.
Way, G.B., “Flagstaff I-40 Asphalt Rubber Overlay Project – Ten Years of Success”,
Proceedings of the Asphalt Rubber 2000 Conference, Vilamoura, Portugal, 2000.
World Bank, Christopher R. Bennett Ian D. Greenwood “HDM-4 Highway Development &
Management, volume seven, MODELLING ROAD USER AND ENVIRONMENTAL
EFFECTS IN HDM-4”, Pages 107-112, The University of Birmingham, Birmingham,
United Kingdom, 2001.
Promoting Asphalt Rubber Application

through Education
DingXin Cheng* R. Gary Hicks*
Albert M. Johnson** Shakir Shatnawi***
*California Pavement Preservation Center

California State University, Chico
202 Main Street, Chico, CA 95929
dxcheng@csuchico.edu
rhicks@csuchico.edu
**California Integrated Waste Management Board

ajohnson@CIWMB.ca.gov
***California Department of Transportation

State Pavement Engineer
shakir_shatnawi@dot.ca.gov
ABSTRACT: There are about 300 millions waste tires generated each year in the United
States. In the state of California alone, more than 40 million waste tires are generated per
year. Waste tires take up valuable landfill space and cause significant negative environmental
impact. Waste tires can be recycled as asphalt rubber to be used in rubberized hot mix asphalt
concrete. Rubberized hot mix asphalt concrete has been used in California for more than 20
years and it has been proven to be a very cost effective paving material when applied properly.
However, when using recycled materials in real applications, one faces many challenges,
especially, if the knowledge of how to use the recycled materials such as waste tires has
not been well disseminated. To overcome these challenges and barriers, teaching materials
have been developed to educate university students and professionals about asphalt rubber
usages. The teaching modules developed in this study have been taught in real university
environments. Several professor training workshops on using asphalt rubber have been
given in California of the United States and the training materials were stored on a website.
Researchers can use the materials as a source of information while practitioners can use them
as guidelines.
Keywords: Asphalt Rubber, Teaching Materials, Waste Tire

1. Background
Waste tires can occupy valuable landfill space. Waste tires in stockpiles, legally or illegally
dumped, may produce tire fires that are very hard to put out, and cause significant public
health and environmental concerns. There are approximately 300 million waste tires generated
each year in the United States. Today, most of the tires are reutilized as tire derived fuels, in
civil engineering applications, or in rubberized asphalt concrete in roadway pavements. There
are many benefits of using asphalt rubber in pavements. Asphalt rubber can reduce reflective
cracking and improve fatigue resistance; increase resistance to bleeding, flushing, drain-
down problems; allow higher binder content and increased film thickness. The open-graded
rubberized asphalt can also reduce noise and splash and spray due to rain. Even though asphalt
rubber is a good product for asphalt pavements, it doesn’t mean that people can apply asphalt
rubber freely. If engineers or contractors don’t have the knowledge on how to utilize the
material, it can fail rapidly. For example, a segment of freeway on Interstate highway 5 (I-5) in
northern California used open graded rubberized hot mix asphalt inappropriately in 2007. The
pavement failed within half year. Another example is a failed asphalt rubber overlay project in
the city of Rohnert Park, California. Many segments of the project were replaced within one
year. Therefore, it is extremely important to have engineers and contractors educated on using
asphalt rubber. This is one of the roles of the California Integrated Waste Management Board
(CIWMB) and this project in particular.
California has many different geologic features, including mountains, desert, valleys, and
coastline, and the climates of these regions are quite different. These diverse geographic and
climate conditions make California an ideal place to develop and test asphalt rubber paving
materials. California has spent significant efforts to study the usage of asphalt rubber in
pavements from laboratory performance tests, Heavy Vehicle Simulator studies, to numerous
field investigations. California has developed useful guidelines and standards for asphalt
rubber, such as Asphalt Rubber Design and Construction Guidelines (Hicks 2002), and
Caltrans Asphalt Rubber Usage Guide (Caltrans 2003).
This study is part of a bigger project about develop teaching materials for university
students on how to utilize waste tires in all areas of civil engineering applications, including
transportation, structural, geotechnical and environmental engineering. In this study, training
and teaching materials on utilizing asphalt rubber based on California experience have been
developed. The developed teaching materials contained the history and long term performance
of asphalt rubber, included case studies, and presented the ASTM standards, the latest
manuals, and specifications used in California.
2. Objectives
The objectives of this research are to:
• synthesize the knowledge of utilizing waste tires in asphalt rubber applications
• develop effective teaching materials to educate university students about utilizing

waste tire products in pavement
• promote sustainability by using waste tires in asphalt rubber applications through

university education
The goal of this paper is to summarize the curricula of asphalt rubber civil engineering
applications to make people aware of the existence of the course materials and resources
developed on this project.
3. Approach and Proposed Curricula
Utilizing waste tires in civil engineering applications is a multi-disciplinary and complex

subject. No single class currently available in civil engineering can cover all the aspects of it.
To reach university students more effectively, we developed teaching modules for different
grade levels for use in related civil engineering classes. As illustrated in Figure 1, it was
proposed to develop waste tire application teaching modules for a variety of civil engineering
courses from freshman level to senior level. Asphalt rubber applications are covered under
introduction to civil engineering design, transportation engineering and asphalt paving
materials classes, and contracts and specifications.
Introduction to Civil
Engineering Design
Contracts and
Transportation Structures Geotechnical Environmental
Specifications
Transportation Strength of Soil Mechanics Environmental

Engineering Materials Engineering
Asphalt Paving Structural Testing Foundation Waste

Materials Lab Engineering Management
Concrete
Materials
Figure 1. Roadmap for Waste Tire Applications in University Curricula
By offering teaching modules for waste tire applications at different class levels starting
from freshmen level, more students can be reached than would be by a single elective class
on utilizing asphalt rubber , for example, like a senior asphalt pavement class. By the end of
university education, a student may be exposed to asphalt rubber applications multiple times.
It was deemed as a more effective way of teaching students about unconventional materials,
such as waste tire derived materials.
4. Curricula Development
A series of course modules about asphalt rubber have been developed for a variety of
undergraduate Civil Engineering courses, including Introduction to civil engineering design,
transportation engineering, and asphalt pavement materials classes.
4.1 Introduction to Civil Engineering Design
The goal of this lecture is to introduce university students to waste tire materials and give
them an overview of utilizing waste tire products in a variety of types of civil engineering
applications. It is important for students to understand the significance of utilizing recycled
materials to preserve valuable natural resources. Students should also understand the
significance of protecting the environment and they should also learn to promote healthy and
sustainable development of our society. Figure 2 shows a picture of the Westley tire fire in the
State of California.
The lecture module also covers the benefits and challenges of using waste tire derived
products in civil and transportation engineering applications. In civil engineering, it discusses
physical properties of waste tire derived aggregate (TDA) such as TDA as backfill materials
for retaining walls and bridge abutments, lightweight fill for embankments, insulation layer
for roadway base, and vibration damping materials for rail lines.
In transportation, it introduces the rubberized hot mix asphalt for pavement. It describes
the benefits of using asphalt rubber; including prolonging paving life, reducing maintenance
costs and noise, improve safety, as well as environmental benefit of getting rid of waste tires.
It also presents a roadmap of civil engineering classes that cover waste tire applications.
FIGURE 2. Tire Fire in Westley California in September 1999 (Earth Link 2008)
4.2 Transportation Engineering
The goal of the transportation engineering lectures is to inform students of the history,
benefits, limitations and practice of using asphalt rubber (AR) as a paving material. The major
references of this class are the Caltrans Asphalt Rubber Usage Guide (Caltrans 2003), Caltrans
Highway Design Manual (Caltrans 2008), and Caltrans Maintenance Technical Advisory
Guide (Caltrans 2008). These lectures are divided into four modules, each dealing with a
different aspect of asphalt rubber applications.
The students are first introduced to the module of the history of using asphalt rubber as a
paving material. Case studies of full scale AR overlay projects in California are presented.
These studies outline the strategy of using AR as an overlay to repair existing distressed
pavements, as well as discussing the design and results of the AR overlays. The benefits of
using AR pavements as a replacement for conventional asphalt are also discussed.
The second module introduces the structural design of AR pavements. A 2005 Caltrans
study is referenced in this module to review the revised practices of using AR in new
pavements as well as an overlay (Caltrans 2005). Students are informed on the recommended
design strategies for new pavements and overlays using AR. An overview of the revised
practices for using AR in overlays and new pavement is also presented. This module also
presents cost analysis comparing AR and conventional asphalt. RHMA can be a cost effective
option when a reduce thickness design is warranted for preventing reflective cracking of the
existing pavement. Table 1 illustrates the Caltrans practice on reduce thickness on overlay
design based on reflective cracking.
Table 1. Caltrans Practice on Reduce Overlay Thickness Conversion Table for crack
retardation (Caltrans 2008)
HMA RHMA-G RHMA-G over SAMI-R

(mm) (mm) (mm)
45 30 N/A
60 30 N/A
75 45 N/A
90 45 N/A
45 if crack width < 3 mm N/A for crack width < 3 mm
60 if crack width ≥ 3 mm or 30 if crack width ≥ 3 mm and
underlying material CTB, LCB, underlying material untreated
105
or rigid pavement 45 if crack width ≥ 3 mm and
underlying material CTB, LCB, or
rigid pavement
45 over
135 60
45 HMA
Students are next introduced to the manufacturing and construction process of AR in the
third module. The module discusses the general paving process with an emphasis on the
different practices between AR and conventional asphalt. An overview of the manufacturing
process informs students how AR is produced and also highlights the operational differences
when dealing with AR such as the laydown and compaction temperatures for successful
placement of AR.
The last module of the lecture goes into detail about AR binder production, AR mix
production, inspection of paving and troubleshooting. Some or all of these modules could be
included in transportation and pavement engineering classes.
4.3 Asphalt Paving Materials
This module consists of asphalt rubber (AR) binder design, the different types of AR mixes
and cautions for using AR. The lecture defines the different types of asphalt rubber binders and
discusses how each type is produced. Crumb Rubber Modifiers (CRM) are the form of waste
tires added to the binder. The interaction between the CRM, the asphalt and the affecting
factors are explained. When designing an AR blend, it is necessary to develop a binder profile
which evaluates the compatibility, interaction, and stability between materials over a period of
time.
The students are introduced to the most commonly used types of rubberized hot mix asphalt
concrete, including Rubberized Hot Mix Asphalt – Gap graded (RHMA-G), Rubberized Hot
Mix Asphalt – Open graded (RHMA-O), and Rubberized Hot Mix Asphalt – Open graded –
High Binder content (RHMA-O-HB). The mix design, advantages, and standard specifications
are described for each rubberized asphalt mixture type. Figure 3 displays that advantage of
RHMA-O to reduce water splashing effect over traditional Dense Graded Asphalt Concrete
(DGAC). Both field blend asphalt rubber and terminal blended asphalt rubber are covered in
this module.
Figure 3. Free-Draining RHMA-O Next To DGAC (Caltrans 2003)
4.4 Contracts and Specifications
This lecture has two modules. One is on ASTM international standards; the other is for
specifications on rubberized hot mix asphalt. A series of ASTM standards related to waste tire
applications are covered. The major ones are: ASTM D8 (ASTM 2002), which defines the
terminology of asphalt rubber; and ASTM D6114 (ASTM 2002), which gives the standard
specifications of the asphalt rubber binder, and ASTM D6270 (ASTM 2004)., which has
detailed definition on tire rubber, material characterization, usage, construction practices,
guideline for fills, and leachate. The lecture also provides students the necessary background
on the ASTM International.
The specification lecture starts with various types and aspects of specifications. As
examples, standard specifications were illustrated using Caltrans standard specification on
RHMA – O (open graded rubberized hot mix asphalt) and RHMA – G (gap graded rubberized
hot mix asphalt) (Caltrans 2007).
5. Outcomes
Teaching materials for utilizing waste tire products in civil engineering applications
have been developed for ten different civil engineering courses. The teaching materials are
available on a website hosted by CSU, Chico. All these lectures have been taught at the
undergraduate level at California State University, Chico in a variety of civil engineering
courses. Students have greatly improved their knowledge on utilizing waste tire products in
civil engineering applications. They were able to demonstrate their knowledge of, and interest
in waste tire applications through their term projects, lab reports, presentations, and homework
assignments.
A website was created to store the teaching materials, which includes PowerPoint
presentations, lecture notes, sample assignments and sample solutions and student work. The
link for the website is:
http://www.ecst.csuchico.edu/cp2c/dxcheng/Curricula/CIWMBEducation.php
Professors and instructors can easily access the teaching materials by logging onto the
website. If you log in as a professor, you will be able to use all of the teaching materials.
The public can also access part of the teaching materials by logging in as a guest. Generally,
they can only access the PDF versions of the presentations. A snapshot of the webpage for
professors is shown as Figure 4.
These course materials are available to be integrated into various courses in the
undergraduate Civil Engineering curriculum and serve to introduce students to using asphalt
rubber in pavement engineering.
Three professor training workshops have been conducted in California. Over 40 professors
from 15 different universities attended the training. The general feedback from the professors
was very positive. They found the teaching materials to be very interesting and useful.
Utilizing waste tires as asphalt rubber in pavement fits into the sustainability and green
construction practices. It is a much needed part of education for the 21st century engineers.
Figure 4. Sample Webpage to Access Teaching Materials for Waste Tire Applications
6. Conclusions
Each year, more than 60 millions waste tires are consumed by civil engineering
applications and ground tire rubber usages, including asphalt rubber (RMA 2006). In order to
promote the beneficial usage of waste tires in civil and transportation engineering, educational
curricula, including a series of lectures for undergraduate courses were developed. The
following conclusions can be drawn from the curricula development project:
• Waste tire applications, especially asphalt rubber, has beneficial properties that
engineers can use. The asphalt rubber can improve the safety and performance
roadway. It can be economical too when it is used at right conditions.
• Teaching modules or lecturing materials were developed to cover freshman level

to senior level classes for asphalt rubber. The freshman class gives students an
introduction and overview of asphalt rubber and their applications. Junior classes cover
the material properties, testing, and standard specifications. Senior classes can cover
asphalt rubber applications in pavement engineering.
• Outcomes show that introduction of curricula in a variety of courses is an effective

way to teach asphalt rubber and can reach more students. Students have demonstrated a
knowledge of and interest in the sustainable use of asphalt rubber through their school
work.
Students in civil engineering are the future engineers and their knowledge of asphalt
rubber will affect the sustainable usage of recycled materials such as waste tires. It was a good
experience promoting the education of the sustainable usage of recycled waste tires in civil
and transportation engineering by developing teaching materials. The education on the use of
other recycled materials can follow a similar approach.
7. References
ASTM D 8, Standard Terminology Relating to Materials for Roads and Pavements, ASTM
International, PA, 2002.
ASTM D6114, Standard Specification for Asphalt-Rubber Binder, ASTM International,

developed in 1997, reapproved 2002.
ASTM D 6270, Standard Practice for Use of Scrap Tires in Civil Engineering Applications,
August 1998, reapproved 2004.
Caltrans, Asphalt Rubber Usage Guide, State of California Department of Transportation,

Division of Engineering Services, Materials Engineering and Testing Services-MS #5,
Office of Flexible Pavement Materials, Sacramento, CA, 2003.
Caltrans, California Department of Transportation. Recommendations for Using RAC in New

Pavements and for Structural Overlays. Sacramento, CA. December, 2005.
Caltrans Standard Specifications, Section 39, State of California Department of Transportation,

Sacramento, CA, 2007.
Caltrans Highway Design Manual, State of California Department of Transportation, Division

of Design for Project Delivery, Sacramento, CA, July 2008.
Caltrans, Maintenance Technical Advisory Guide, Volume I - Flexible Pavement Preservation,

Second Edition, State of California Department of Transportation, Office of Pavement
Preservation, Sacramento, CA, March 2008.
Earth-Link Website. http://www.earth-link.com.hk/tire3.htm. Accessed July 28, 2008.
Hicks, R.G., Asphalt Rubber Design and Construction Guidelines, Volume I – Design
Guidelines, Sacramento, CA, Jan 2002.
RMA, Scrap Tire Markets in the United States, Rubber Manufacturers Association, November
2006.
Analysis of Environmental Sustainability in

The Rehabilitation of Existing Pavements
Using Asphalt Rubber Hot Mixes
Ines Antunes* – Adriano Murachelli**
* PHD, Technical Director Asphalt Rubber Italia Srl, Italy

ines.antunes@asphaltrubberitalia.it
** Environmental engineer, consultant and vice president of AIAT - Italian Association of
Environmental and Land Planning Engineers
amurachelli@ingegneriambientali.it
ABSTRACT. Asphalt Rubber is by definition a green technology, not only because it uses a
recycled material but also because its mechanical performance allows pavement thickness
reduction, saving natural resources and money and reducing CO2 emissions.
This paper quantifies the main parameters connected to the study of the environmental
sustainability related to the production and laying of AR hot mixes.
After a brief summary of the state of AR technology in Italy, results of a quantitative analysis
of environmental benefits arising from the use of this technology are shown in terms of energy
balance, highlighting the significant savings achieved under construction, in addition to the
benefits of longer duration and less maintenance required for AR pavements.
Finally, it is reported, as a study case, the set of improvements solutions adopted for the
rehabilitation of the Florence - Pisa – Livorno highway, where the choice of using AR
technology as a rehabilitative solution was made after a detailed analysis of the environmental
benefits.
KEYWORDS: Asphalt Rubber, Wet process, Energy saving, GHG saving, CO 2 emissions
reduction.
1. Introduction
Noise reduction, safety, durability, shorter construction time, lower maintenance costs,
lower CO2 emissions, higher cost benefit ratio and natural resources saving: the list of benefits
connected to the use of Asphalt Rubber technology exceeds the most innovative material.
The present paper focuses the environmental sustainability of this technology, since
nowadays environmental factors have an increasingly decisive role in the choice of road
materials.
By re-using common waste products such as rubber from old tires it is undeniable that
there are numerous environmental advantages in saving energy and natural resources. Also,
experience has shown that by properly combining the waste product of ground tire rubber
with asphalt at high temperatures the resultant binder will make a hot mix with superior
engineering properties, including reduced fatigue and reflection cracking, greater resistance
to rutting, improved aging and oxidation resistance and better chip retention due to thick
films [1 - 3] . Plus, Asphalt Rubber pavements have demonstrated to have lower maintenance
costs, higher noise absorption, reduced splash and spray and better night-time visibility due to
contrast between pavement and striping [2, 5].
From above it can be deduced that environmental benefits of AR technology are connected
to several aspects during the entire pavement life cycle:
- On one side, the use of AR allows significant reduce in thickness and increases
pavement life (with lower maintenance), allowing a significant saving of raw materials,
saving energy during hot mix production, transport and laying, and reducing emissions
of pollutants and greenhouse gases;
- On the other side, there is the recycling of old tires, the fact that AR pavements present
more regular surface, which reduces consumption of vehicles and related emissions.
AR friction course and open graded mixes can actually reduce noise, which can avoid
the current environmental impact of sound-absorbing panels. Moreover, AR is a
recyclable hot mix, whose production does not produce more fumes than traditional
solutions [3].
2. The state of Asphalt Rubber technology in Italy
While in other European countries crumb rubber modified hot mixes are now commonly
used, in Italy, AR technology is still considered to all affects a new technology and most of
old tires are destined for burning, especially in cement industry.
The fact that the European Community since 2006 prohibits the accumulation in landfill of
used tires, has revived the interest for their possible recycling in road pavements and several
agencies have started to recommend the use of AR as a sustainable engineering solution.
With regard to the regulatory developments that are following the technological evolution
of AR in Italy, Asphalt Rubber binder (wet method) is part of the environmental directive:
“Information concerning the recycled materials and goods and articles made with recycled
materials from rubber”- as the DM 203/2003 - G.U. No 173, 27 July 2005, which encourage
the reuse of recycled materials in new constructions.
Since the first applications in the Italian network, in the end of 2006, AR pavements have
been studied and monitored and environmental expectations by increasing the application of
AR in Italy confirm the aspects introduced in the preceding paragraph.
Regarding noise, tests have demonstrated that an AR Open graded surface leads to a
reduction, depending on traffic speed, of: -3 dB for speeds <50 km/h, 4 dB for speeds between
50 and 90 km/h, 6 dB over the 90 km/h [11]. Since a noise barrier is able to obtain a reduction
of about 1 dB for every 60 cm in height, it’s possible to reduce or even eliminate them, with a
substantial reduction of environmental impact. These considerations, already tested and proven
in the U.S. and Europe, have been measured in Italy: the CIRS (Centro Interuniversitario di
Ricerca Sperimentale Stradale) measured a reduction of 4.5 dB for an average speed of 40
km/h. This result was confirmed by analysis conducted by local Environmental Agency (ARPA
of Tuscany), which has led to define a sound absorption coefficient equal to 0.75 for AR Open
graded hot mix [11].
3. Energy quantification of environmental benefits using Asphalt Rubber hot mixes for
the rehabilitation of existing pavements
To actually quantify some of the environmental benefits listed in the introduction, it is

presented the environmental improvements in terms of energy savings achievable through the
application of AR hot mixes, compared to traditional hot mixes.
The AR design solutions allow an immediate benefit in terms of energy saving, due to the
reduction of raw materials and related processes necessary during the construction phase,
besides the advantage due to the longer life of AR pavements. Since this second type of
benefit has already been examined in several studies, the present study only quantifies the
savings during construction phase, in itself enough to convince of the environmental validity
of AR solutions.
In order to illustrate energy and CO2 emission savings, the study quantifies the energy
balance of an entire rehabilitation process of a road pavement using AR hot mixes.
During the rehabilitation process it’s assumed that AR hot mixes (gap graded or open
graded) are placed at approximate half the thickness of conventional hot mixes. This
assumption is, anyway, conservative [1].
3.1 Overall energy balance during the rehabilitation process
In this paragraph it’s shown the analytical energy balance of the yard. The whole
construction process can be divided into the following phases:
- AR hot mixes production;
- Hot mixes transport;
- Milling;
- Stabilization and recycling;
- Transport for disposal;
- New pavement construction.
For each of the steps listed, consumptions and energy resources and materials are detailed,
taking as reference a pavement rehabilitation with 3 cm AR Open Graded (with nominal
maximum aggregate size of 12.5 mm), 4 cm AR Gap Graded (with nominal maximum
aggregate size of 14 mm), SAMI in Asphalt Rubber and 30 cm recycling with foamed
bitumen. This type of pavement structure as been applied several times in Italy.
The following charts illustrate each aspect of the process, highlighting their energy
consumption per m³ of AR hot mix.
- Hot mixes production: the known characteristics of the AR hot mixes plant, equipped
with a two stage gas burner (580 kW), allows to determine the energy consumption
associated with production;
- Hot mixes transport: it depends on the distance between production plant and yard;
assuming an average distance equal to 50 km it’s possible to define the energy
consumption associated with transportation;
- Cold milling: knowing the consumption of a typical cold milling machine as a

function of milling thickness (as shown in Figure 1), it’s possible to determine the
total consumption of fuel and, therefore, the equivalent energy consumption and CO2
emissions;
120
Fuel consumption [l/h]
100
80
60
40
20
0
0 10 20 30 40
Milling thickness [cm]
Figure 1. Typical fuel consumption function of a cold milling machine

- Stabilization and recycling: as for the milling machine, knowing the fuel consumption
as a function of operating conditions, depending on the thickness to be recycled
and stabilized it’s possible to find the total energy consumption and associated CO2
emissions;
- Disposal of resulting materials: materials milled and recycled are collected and
transported away from the yard. Consumptions and emissions are calculated
considering an average route equal to about 10 km, and taking into account the 15%
increase in average volume that occurs between the stage of milling and the subsequent
filling and stabilization;
- New pavement laying: in this phase, the energy balance has taken into account the
power of every machine and the total use of each of them.
The results obtained are summarized in Table 1 and Figure 2.
Table 1. Energy balance of the yard per unit of AR hot mixes
Energy consumption CO2 emissions

Process
[kWh/m³] [kg CO2/m³]
AR hot mixes production 6.206 3.600
Hot mixes transport 13.185 7.647
Milling 17.651 10.237
Stabilization and recycling 19.214 11.144
Transport for disposal 1.953 1.133
Hot mixes laying 2.960 1.717
TOTAL 61.169 35.478
25.00 Energy consumption [kWh/m³]

CO2 emissions [kg CO2/m³]
20.00 19.214
17.651
15.00
13.185
11.144
10.237
10.00
7.647
6.206
5.00 3.600
2.960
1.953 1.717
1.133
0.00
AR hot mixes Hot mixes Milling Stabilization Transport for Hot mixes
production transport and recycling disposal laying
Figure 2. Energy balance per unit of AR hot mixes

For the entire rehabilitation process, therefore, energy consumption amount to 61.169 kWh
for each m³ of AR hot mixes finally obtained, and the corresponding value of CO2 emissions
is equal to 35.478 kg CO2/m³.
3.2 Energy saving associated to AR solutions for pavement rehabilitation
The main items that, during pavement rehabilitation, conduce to a substantial energy
saving of AR applications compared with traditional solutions are:
- Reduced hot mixes production: this affects both in terms of production and
transportation. Production was estimated assuming a plant with a 580 kW gas burner.
Transport is estimated assuming, again, a 50 km average distance between plant and
construction site;
- Reduction of milling thickness: the reduction of thickness presents the double

advantage of reducing, at the same time, milling process and the amount of resulting
materials. In order to calculate energy saving during milling, it was quantified the fuel
saving analyzing the relative curve of consumption (as shown in figure 1). To taking
into account the energy saving associated to disposal needing, it was considered a 10
km average distance from construction site and disposal landfill.
Results obtained are summarized in Table 2 and Figure 3.
Table 2. Energy savings achieved for each m³ of AR hot mixes
CO2 Emission saving

Process Energy saving [kWh/m³]
[kg CO2/m³]
Hot mixes production 7.378 4.279
Hot mixes transport 15.674 9.091
Hot mixes laying 2.960 1.717
Milling 1.873 1.086
Transport for disposal 0.766 0.444
TOTAL 28.651 16.617
18.00 Energy saving [kWh/m³]
15.674 CO2 Emission Saving [kg CO2/m³]

16.00
14.00
12.00
10.00 9.091
8.00 7.378
6.00
4.279
4.00 2.960
1.717 1.873
2.00 1.086 0.766 0.444
0.00
Hot mixes Hot mixes Hot mixes laying Milling Transport for
production transport disposal
Figure 3. Energy saving and CO2 emission saving per unit of AR hot mixes
Overall, using AR hot mixes instead of traditional design solutions, energy savings amount
to 28.651 kWh for each m³ of AR hot mixes finally obtained, and the corresponding reduction
in CO2 emissions is equal to 16.617 kg CO2/m³.
That savings, compared to overall energy consumption associated to the rehabilitation

process, allow energy savings equal to 47%.
4. An Italian case study: the Florence-Pisa-Livorno Highway
The present case study regards the pavement rehabilitation along about 3+3 km of Florence
–Pisa–Livorno highway in Tuscany.
Designed and built over many years, this highway doesn’t present uniform road section
characteristics along its path, but road surface width is sufficiently constant and equal to about
18.60 m.
The measures provided are aimed at improving road safety raising the performance of
infrastructure. Interventions originally planned with traditional solutions consisted in:
- Removal of the existing pavement for a thickness equal about to 25-35 cm;
- Lime and/or cement stabilization;

- 30 cm of cold recycling (with foamed bitumen);
- 9 cm of traditional dense hot mix with modified bitumen (SBS medium)
- 5 cm of traditional binder hot mix with modified bitumen (SBS medium)
- 4 cm of traditional open graded hot mix with modified bitumen (SBS hard)
An improving AR solution was defined to increase performance and to implement, at the

same time, a significant thickness reduction of pavement rehabilitation, with correspondent
money saving.
Several researches [6 - 8] have demonstrated that, at intermediate and high temperatures,

rubber stiffens the binder and increases elasticity (proportion of recoverable deformation)
with the consequent reduction of temperature susceptibility and improvement in resistance to
permanent deformation and fatigue.
The properties conferred to bitumen by the use of rubber as modifying agent clearly reflect
on bituminous mixes manufactured with Asphalt Rubber in terms of rutting and cracking
resistance compared with conventional bituminous mixtures, as showed by several studies [9,
10].
A structural design method was applied to calculate and verify AR solution, determining
the state of tension and deformation (extension) caused by applied traffic loads, verifying the
failure criteria for the pavement and taking into account the following issues:
- AR hot mixes contribution by decreasing the extent of vertical compression at the top
of pavement foundation;
- AR hot mixes contribution by reducing traction stress at the base layers of bituminous
layers.
The structural analysis was performed using BISAR® software by Shell Research. To find
the optimal design, several possible solutions have been evaluated, choosing for the surface
AR open graded hot mixes, which presents high noise-reduction characteristics, to be applied
instead of the open graded hot mix originally planned. The chosen pavement solution also
includes an AR SAMI that provides a durable waterproof membrane that has the necessary
flexibility to withstand heavy traffic, foundation settling and climate changes. The best
advantage is the increased resistance to the propagation of cracking: it allows the sealing of
cracks on the existing pavement and prevents the spread to new surface layers, since it reduces
the stress transmitted to the upper layers.
This SAMI membrane is formed by the application of the hot asphalt rubber mix (185º),
at the rate of 2.5 Kg/m², and also by the application of chips at a rate of 10 to 12 Kg/m². The
thickness of this membrane is about 1.2 to 1.5 cm, and can be considered effective for the
pavement’s design thickness calculation.
Considering experience with Italian road materials and characteristics of aggregates that
are used in the present geographic zone and climatic database (max temperature oh 45°C and
minimum of – 15°C), it was assumed for respective AR Gap-graded and Open-graded hot
mixes an Elastic modulus of 4,000 and 1,700 MPa. SAMI was considered as a 15 mm layer
with 3500 MPa of E modulus.
Granulometric specifications for AR materials are shown in Table 3 and Table 4.
Table 3. Granulometric specifications for AR Gap Graded.
Asphalt Rubber Gap Graded MD 12.5 B AR

Mix (UNI EN 13108-1)
Dimension Limit Limit Regular
Granulometric Composition (EN 12697-2)
(mm) (min) (max) Value
16 100 100 100

0.063
0.5
10
12.5
16
2
100
14 92 100 100
90
80 12.5 83 97 90
70
10 68 82 75
60
50 8 54 68 61
40
4 25 37 31
30
20 2 12 24 18
10
0.5 7 15 11
0.063
0
12.5
0.5
10
16
2
0.063 0 3 1
UNI EN
AR BINDER CONTENT § 5.3.1.3 % MIN 7.5-8.5 8
12697-1/39
Table 4. Granulometric specifications for AR Open Graded.
Asphalt Rubber Open Graded MD 12.5 B AR

Mix (UNI EN 13108-1)
Dimension Limit Limit Regular
Granulometric Composition (EN 12697-2)
(mm) (min) (max) Value
16 100 100 100
0.063
0.5
10
12.5
16
2
8
100 12.5 93 100 100
90
80 10 88 100 95
70
60 8 68 82 75
50
40
4 23 37 30
30
2 3 15 9
20
10
0.5 2 10 6
0.063
0
12.5
0.5
10
16
0.063 0 3 1
2
UNI EN
AR BINDER CONTENT § 5.3.1.3 % MIN 8.5-9.5 9
12697-1/39
Design solution that delivers the best performance in terms of expected life, cost-benefit
analysis and noise reduction is detailed below:
- Removal of the existing pavement for a thickness equal about to 30 cm;
- Lime and/or cement stabilization;
- 30 cm foamed bitumen cold recycling;
- AR SAMI;
- 3 cm of calcareous AR Gap Graded hot mix;
- 3 cm of basaltic AR Open Graded hot mix.
Performance comparison of original and alternative solutions considered is shown in

Table 5.
Table 5. Performance comparison of original and alternative pavement solutions performed using BISAR® software by Shell Research.
Permanent
Fatigue
Total Deformation Expected life
Solutions Fail Na (Shell)
Thickness (m) et Nf ev Nv [years]
*10e-6 Shell *10e-6 Shell
Original solution: 4 cm Open mix + 5 cm
binder SBS + 9 cm dense SBS + 30 cm 0.430 52.7 1.16E+08 111.3 6.84E+08 Fatigue 1.16E+08 30
Recycling + 150 Mpa
Alternative 1: 3 cm AR open + 3 cm AR gap + Perm.
0.375 43.8 6.07E+09 149.0 2.13E+08 2.13E+08 35
AR SAMI + 30 cm Recycling + 150 MPa def.
Alternative 2: 3 cm AR open + 4 cm AR gap + Perm.
0.335 50.2 3.09E+09 169.8 1.26E+08 1.26E+08 35
Alternative 3: 3 cm AR open +5 cm AR gap + Perm.
0.345 55.5 1.87E+09 162.8 1.49E+08 1.49E+08 35
As one can see from the table, applying the traffic forecasts provided, pavement expected life obtained using the same software, parameters
and laws of fatigue to all solutions was:
• 30 years for the original solution;
• 35 years for all the three alternative solutions in AR.

Functional, Environmental and Other Aspects
787
Considering the first AR alternative solution shown above, overall energy balance of the
yard was calculated assuming the following conditions.
- Hot mixes production plant has a capacity equivalent to 220,000 kg/hours and it’s
powered by a two stage burner with a thermal power up to 581 kW.
- Production plant is about 65 km far from rehabilitation yard. Considering the average
consumption of each truck and the ratio of CO2 emissions of fuel (diesel), it is possible
to calculate energy savings and correlated CO2 emissions.
Results are summarized in the Table 6 and Figure 4.
Table 6. Overall energy balance for rehabilitation yard
Energy consumption CO2 Emissions [kg

Process
[kWh] CO2]
AR hot mixes production 7,894 4,579
Hot mixes transport 16,771 9,727
Milling 28,402 16,473
Stabilization and recycling 31,663 18,365
Transport for disposal of milled material
2,580 1,496
not recycled
Hot mixes laying 5,006 2,903
TOTAL 92,316 53,543
35,000
31,663 Energy consumption [kWh]
28,402 CO2 emissions [kg CO2]
30,000
25,000
20,000 18,365
16,771 16,473
15,000
9,727
10,000
7,894
4,579 5,006
5,000
2,580 2,903
1,496
0
AR hot mixes Hot mixes Milling Stabilization and Transport for Hot mixes laying
production transport recycling disposal
Figure 4. Overall Energy balance for the yard

Major savings in terms of materials quantities to be produced, transported, lay and/or

disposed, are related to thickness reduction of the pavement, which allows a significant
cutback in the conglomerate to be applied and in milling depth.
The respective reductions, compared with traditional solutions, of such materials quantities
are:
- reduction of hot mixes produced, equal to 2,435 m³;
- reduction of milled material to be disposed of, equal to 1,793 m³.
Such production and disposal reductions lead to energy savings quantified as reported
below. For example, considering the typical fuel consumption function of a cold milling
machine, shown in Figure 1, it is possible to quantify energy savings resulting from the 3 cm
reduction in milling thickness, as shown in Table 7.
Table 7. Energy savings achievable by reducing the milling thickness.
Thickness reduction (cm) -3.0

Fuel savings per hour (l/hours) 8.0
Milling machine work capacity (m²/hours) 500
Total milling surface (m²) 21,200
Overal work time (hours) 42
Overall fuel savings (l) 340
Energy savings (kWh) 1,551
CO2 reduction (kg CO2) 900
Moreover, the relative reduction in the amount of milled material allows further energy
savings on the transport for disposal, estimated in Table 8.
Table 8. Energy savings achievable by reducing transport for disposal of milled material.
Milled material reduction (m³) 636

Distance from yard to landfill [km] 10
Fuel consumption [km/l] 2.5
Overall fuel savings [l] 283
Energy savings (kWh) 1,290
CO2 reduction (kg CO2) 748
Overall energy savings, obtained using AR hot mixes as a more efficient alternative to
traditional hot mixes, has led to the following results shown in Table 9 and Figure 5.
Table 9. Overall energy savings achieved using AR hot mixes
Energy consumption CO2 Emissions [kg

Process
[kWh] CO2]
Hot mixes production 10,526 6,105
Hot mixes transport 22,361 12,969
Hot mixes laying 5,006 2,903
Milling 1,551 900
Transport for disposal 1,290 748
TOTAL 40,734 23,626
25,000 Energy saving [kWh]

22,361 CO2 Emission Saving [kg CO2]
20,000
15,000
12,969
10,526
10,000
6,105
5,006
5,000
2,903
1,551 900 1,290 748
0
Hot mixes Hot mixes Hot mixes laying Milling Transport for
production transport disposal
Figure 5. Overall energy savings achieved using AR hot mixes
Therefore, AR optimal design solutions allow energy savings equal to about half of total
energy consumption resulting from the rehabilitation yard energy balance.
5. Conclusions
AR hot mixes in rehabilitation processes, as well as presenting better structural and

functional performance, allow a significant reduction in environmental impact. Results of the
analysis described in this article show energy savings that amount to 28.651 kWh for each m
³ of AR hot mixes finally obtained, and a corresponding reduction in CO2 emissions equal to
16.617 kg CO2/m³.
Regarding the case-study presented, that is 3+3 km of an Italian highway, the overall
energy balance for the rehabilitation yard shows a total comsumption of 92,316 kWh and a
corresponding value of CO2 emissions equal to 53,543 kg CO2. The corresponding energy
savings are equal to 40,734 kWh and the reduction in CO2 emissions is equal to 23,626 kg
CO2, which is the equivalent of about 40,000 heavy (10%) and light (90%) vehicles emissions
that travels in that highway
Therefore, the savings achievable using AR solutions instead of pavement rehabilitations

with traditional hot mixes are equal to about half (44%) the overall energy consumption and
carbon dioxide emissions resulting.
Moreover, the AR solutions present better structural and functional performance than the
original solution. They may, indeed, support more traffic than expected and have a longer
expected life.
The alternative solution also allows a significant noise reduction due to the AR open graded
sound absorption coefficient equal to 0.75, in addition to improved adhesion characteristics.
Pavement maintenance expected in the medium to long term of this package will consist in
the restoration of the surface layer (3cm) where necessary, and not earlier than 8 years.
All this features lead the pubblic autority that manages the Firenze-Pisa-Livorno highway
to consider the AR solution the best technical alternative for the pavement rehabilitation.
6. Bibliography
[1] Antunes I., Giuliani F., Sousa J. B., Way G., “Asphalt Rubber: Il Bitume Modificato
con Polverino di Gomma di Pneumatico Riciclata”, Varirei - V International Congress
of Valorization and Recycling of Industrial Waste, L’Aquila 2005.
[2] G. B. Way, “OGFC Meets CRM. Where the rubber meets the rubber 15 years durable
success”, 2003.
[3] L. Momm, R. Salini, “Study of Recycled Tire Rubber in Asphalt Concrete Mixtures”,
Proceedings of Asphalt Rubber 2000 Conference, Portugal.
[4] ANTUNES, I., WAY, G. B., SOUSA, J., KALOUSH, K. (2006) – “The Successful
World Wide Use of Asphalt Rubber” – Proceedings of the 16th Convegno Nazionale
SIIV, Cosenza, Italy.
[5] BERNHARD, R., WAYSON, R. L. (2005) – “An Introduction to Tire/Pavement Noise

of Asphalt Pavement” – Final Research Report of the Institute of Safe, Quiet and
Durable Highways, Purdue University
[6] BERTOLLO, S. M., BERNUCCI, L. B., FERNANDES, J. L., LEITE, L. M. (2004)

– “Mechanical Properties of Asphalt Mixtures Using Recycled Tire Rubber Produced
in Brazil – A Laboratory Evaluation” – Proceedings of the 83rd TRB Annual Meeting,
Washington, D.C., USA.
[7] GIULIANI, F., MERUSI, F. (2006) - “Il Ruolo dei Bitumi Modificati con Polverino di
Gomma Riciclato nel Ripristino Funzionale delle Pavimentazioni Soggette a Fenomeni
di Ormaiamento” – Proceedings of the 16th Convegno Nazionale SIIV, Cosenza, Italy.
[8] JUNG, J., KALOUSH, K.E., WAY, G.B. (2002) – “Life Cycle Cost Analysis:
Conventional versus Asphalt Rubber Pavements” – Rubber Pavement Association,
Arizona, USA.
[9] KALOUSH, K. E., WITCZAK, M. W. (2002) – “Tertiary Flow Characteristic of

Asphalt Mixtures” – Journal of the Association of Asphalt Paving Technologists, vol.
71, pp 248-280.
[10] KALOUSH, K. E., WITCZAK, M. W., SOTIL, A. C., WAY, G. B. (2003) – “Laboratory
Evaluation of Asphalt Rubber Mixtures Using the Dynamic Modulus (E*) Test” –
Proceedings of the 82nd TRB Annual Meeting, Washington, D.C., USA.
[11] Internet site: www.asphaltrubberitalia.com

Chapter 6
Evaluation and Design of

Chip Seals
Evalutation and Design of Chip Seals 795
Asphalt Rubber Chip Seal Construction

Evaluation
J. Shawn Rizzutto, P.E.
California Department of Transportation

722 E. Barioni BLVD
Imperial, California 92251
USA
Shawn_J_Rizzutto@dot.ca.gov
Abstract: The California Department of Transportation (CALTRANS) has used Asphalt

Rubber Chip Seals (ARCS) as an effective preventative maintenance strategy for over 20
years, not only on low volume roads, but also on major roadways where high truck traffic,
high ambient and pavement temperatures have been prevalent. Due to flushing/bleeding in the
wheel paths, on one particular project in Imperial County, California, six test sections were
placed to evaluate ARCS specification refinements. Key concepts developed are the minimum
CRM content of twenty percent, use of PG 70–10 base stock asphalt, and recognition that two
descending viscosity readings are necessary for proper asphalt rubber binder reaction. This
paper will evaluate the effects of the ARCS specification refinements through the performance
of six test sections, subjected to heavy traffic loads and high temperatures.
Key Words: reacted asphalt rubber binder, pavement preservation, preventive maintenance,
asphalt rubber chip seals, descending viscosity readings.
1. Introduction
Six asphalt rubber chip seal (ARCS) test sections were placed on Route 86 from post mile
28.45 to 30.68 in Imperial County on May 12, 2005. The purpose of the test sections was
to evaluate the recommendations set forth in the Review and Evaluation of Asphalt Rubber
Chip Seals dated October 17, 2005 (ARCS 2005), which investigated several projects that
experienced flushing/bleeding. One of the projects that was investigated was on Route 86
From Metric Station 609+70.40 to 708+19.00 (PM 37.3/43.3). It was decided in March of
2005 during the final revisions of the (ARCS 2005) that the recommendations needed to be
evaluated. The best way to evaluate the recommendations and determine their soundness was
to construct test sections implementing the recommendations and subject all the test sections
to the same traffic and climatic loading experienced by one of the projects (Route 86, PM
37.3/43.3) that experienced flushing/bleeding. The test sections were laid out in the following
manner, as illustrated in Table 1:
Table 1. Test section layout
MARTIN ROAD PM 28.45, SOUTH OF BUCK ROAD PM 30.68 NORTHBOUND

HWY. 86 Imperial County, California
Description
Section Application Rate
Binder Type / Aggregate Type
1 2.264 l/m2 Hwy. 86 Binder / Hwy. 86 aggregate
2 2.264 l/m 2
Hwy. 86 Binder / New gradation
3 2.264 l/m2 * PG Binder / New gradation
5 2.490 l/m 2
* PG Binder (no asphalt modifier) / New gradation
*All PG binders utilized were made from PG 70-10 non-polymer modified
Hwy. 86 Binder utilized an AR 4000 paving asphalt
The asphalt rubber binders utilized for the test sections were the standard California
Department of Transportation Type 2 wet process asphalt rubber binder. All components
were tested and found to be in conformance with the appropriate specifications. Table 2 below
illustrates the percentage of components for each test section.
Table 2. Asphalt rubber binder composition for test sections
Test Sections 1 and 2 Test Sections 3, 4 & 6

76% AR -4000 Paving grade asphalt 76% PG 70-10 Paving grade asphalt
17% Scrap Tire crumb rubber 17% Scrap Tire crumb rubber
5% High Natural crumb rubber 5% High Natural crumb rubber
2% Raffex 120 ACB- Asphalt modifier 2% Raffex 120 ACB – Asphalt modifier
Test Section 5
76% PG 70-10 Paving grade asphalt
17% Scrap Tire crumb rubber
5% High Natural crumb rubber
No Asphalt Modifier
The aggregates shown in Figure 1 are the actual aggregates utilized for hot pre-coated
screenings, the specified gradations are shown in Table 3 and the actual aggregates for each
test sections are shown in Table 4.
Grading used on Test New 12.5mm Medium
Section 1 (current Section Maximum Grading
37 Coarse 12.5mm Grading)
Figure 1. Aggregate gradation comparison for 12.5 mm screenings

Table 3. Aggregate gradation specification comparison for 12.5 mm screenings
Table 4. Actual aggregate gradation for test sections
Sieve Size Test Section #1 Test Sections 2-6

% Passing % Passing (AVG)
19mm 100 100
12.5mm 96 81
9.5mm 54 16
4.75mm 0 1
2.36mm 0 0
75µm 0 0
(AVG) = Average of three test samples.
2. Background
Prior to the publishing of “The Review and Evaluation of Asphalt-Rubber Chip Seals”
(ARCS 2005), dated October 17, 2005, it was decided between Caltrans and Industry that due
to the binder flushing problems encountered on the three projects addressed in the report (ARCS
2005) and based on the recommendations listed in the report (ARCS 2005), six test sections
would be placed south of the original project location in the #2 Lane of Imp-86. These test
sections would vary in the asphalt binder type, binder application rate and cover aggregate size
as shown in Table 1 in the introduction. The six test sections were placed in the northbound
#2 Lane between Metric Stations 467+00 and 503+40 (PM 28.4 and 30.7) – North of Martin
Road to North of Hoskins Road on May 12, 2005. On April 27, 2006 and again on April 26,
2007, the test sections were reviewed and visually evaluated. The review and evaluation of
the performance of the ARCS test sections, at the approximate two year time period was based
on the following criteria:
2.1 Binder Type
The existing specified type 2 asphalt rubber binder type of AR–4000 was used in test
sections #1 and #2. Test sections #3, 4 and 6 used PG 70-10 base stock for the type 2 asphalt
rubber binder and test section # 5 used the PG 70-10 base stock for the type 2 asphalt rubber
binder without asphalt modifier. In the production process of asphalt rubber binder, the paving
asphalt is modified with a resinous, aromatic hydrocarbon compound (asphalt modifier/
extender oil) which has a high flash point and which with the additional aromatic contents of
the asphalt binder and high natural crumb rubber, reacts with the rubber portion of the crumb
rubber modifier (CRM), which is more appropriately defined as a “polymer swell” rather than
a chemical reaction. The crumb rubber absorbs the aromatic oils and releases some of the
similar oils used in rubber production into the asphalt binder, as illustrated in Figure 2 (HEITZ
1992). The reaction of the asphalt binder and the crumb rubber blend is affected by:
- The type and amount of the mechanical mixing energy in the blending process to react
the asphalt binder with the CRM.
- The size and texture of the CRM and the amount of the aromatic hydrocarbons in the
asphalt binder and the extender oil.
- The blending time and temperature of the blended Asphalt-Rubber binder.
- The temperature at which the reaction of the blend occurs.
- The length of time the temperature remains elevated.
In order to promote the interaction between the paving grade asphalt binder and the
asphalt rubber components, the mixture has to be reacted at high temperatures and with
aggressive agitated mixing. Since the asphalt rubber mixture possesses a resistance to flow
(high viscosity), heating the mixture to higher temperatures is necessary to keep the asphalt
rubber particles suspended in the blend. Although the reaction will be gradual at fairly low
temperatures, temperatures from 190.6ºC to 218.3ºC are necessary for the reaction to occur
and to produce timely production of the material.
“The specified reaction time should be the minimum time (at a preset temperature)
required to stabilize the binder viscosity”. (HEITZ 1992)
From SSP 37-030 (11-16-07) and NSSP 37-030

“The temperature of the blended paving asphalt and asphalt modified mixture shall not be less
than 190.6° C nor more than 226.7° C when the CRM is added. The combined material shall
be reacted for a minimum of 45 minutes after incorporation of the CRM at a temperature of
not less than 190.6° F nor more than 218.3° C. The temperature shall not be higher than 10°
below the actual flash point of the Asphalt-Rubber Binder”.
After the incorporation of the CRM into the asphalt binder, viscosity readings shall be
taken at least every hour with not less than one reading for each batch of the asphalt rubber
binder at a temperature of 218.3℃ and recorded when the reaction of the CRM with the
asphalt binder occurs, in three stages (Huff 1980).
As illustrated in Figure 2, in Stage 1, where the CRM has been initially mixed with the
asphalt binder, the CRM appears as an un-dissolved material. In Stage 2, a conversion occurs
of the CRM and asphalt binder to a more soluble “gel” state due to the additional blending
time and heating temperature. Stage 2 represents the four hour maximum time, after the 45
minute minimum reaction time period and the second descending viscosity reading has been
achieved, that the asphalt rubber binder is required to be applied to the pavement surface.
If any of the material is not used within the four hour time period after the reaction period,
heating shall be discontinued. In Stage 3, where the CRM is somewhat overreacted, the
asphalt rubber binder is returned to a material similar to the original paving grade asphalt
and does not meet the physical property specifications in SSP 37-030. Once the material
temperature cools below 218.3℃, it can be reheated no more than two times. If the material is
reheated between 218.3°C and 226.7°C additional scrap tire CRM, not exceeding 10 percent
of the total binder weight, can be added and reacted for a minimum of 45 minutes.
Figure 2. Reaction stages of asphalt and rubber (HUFF 1980)

Since the Time-Temperature-Viscosity curve, when plotted, generally depicts a bell shaped
curve, it is imperative that when the reaction of the CRM and asphalt binder has reached the
top of the curve, that the viscosity of the blend is descending on the curve. To verify this state
of the reaction, two descending viscosity readings are necessary to account for any variation in
the asphalt rubber binder during production and to verify that the viscosity meets the required
specifications. Currently, SSP 37-030 adopted 11-16-07, specifies a viscosity of 1,500 to 3,000
cP at 218.3°C. The viscosity limits need to be revised to between 1,500 and 2,500 cP as it
relates to ARCS only.
It is important to understand that the design parameters of the binder and testing procedures
are necessary to produce a satisfactory formulation. Just mixing together proportions of
arbitrarily selected asphalt binder, CRM and extender oil, even within the specified ranges,
will not necessarily produce a binder that possesses the physical properties required in the
specifications (CTARUG 2003).
2.2 Aggregate Gradation
Test section #1 used the Standard Specification – Section 37-1.02 -12.5 mm coarse
gradation with the percent passing the 9.5 mm sieve, revised by a Contract Change Order
(CCO), from 70-85 % to 50-80 %. This was the original grading used on the asphalt rubber
chip seal portion of project 11-241104 which resulted in bleeding/rutting in the wheel paths
and subsequent preparation of the October 17, 2005 report (ARCS 2005). Test sections # 2
through #6 used a revised 12.5 mm coarse gradation as recommended by the ARCS (2005)
report. This gradation results in a more one - size aggregate with no intermediate aggregate
particles occupying voids in the more uniform mat. The actual aggregates, gradations used are
illustrated in Figure 1, Table 3 and Table 3.
3. Asphalt Rubber Chip Seal Test Section Construction
3.1 Binder Application
The current asphalt rubber seal coat specifications (Standard Special Provision 37-030 -
last modified on 11-16-2007) require a binder spread rate of 2.490 liters to 2.943 liters per
square meter. A lower planned spread rate of 2.264 liters per square meters was placed on
test sections # 1 through # 3. A planned spread rate of 2.490 liters per square meter was to be
placed on test sections # 4 and # 5 while a planned spread rate of 2.716 liters per square meter
was to be placed on test section # 6. The actual spread rates were verified by certified weigh
back tickets and found to be either right on target or slightly lower than target by 0.091 l/m2.
3.2 Aggregate Spread Rate
The current asphalt rubber chip seal specifications (Standard Special Provision 37-030 –
last modified on 11-16-2007) require an aggregate spread rate of 14.65 kilograms to 21.70
kilograms per square meter. The old metric designation used on the 11-241101 project was 15
to 22 kilograms per square meter. The actual aggregate spread rate ranged from 20.61 kg/m2to
21.65 kg/m2.
3.3 Flush Coat Application
The flush coat spread rate was 0.18 l/m2 of CQS1 emulsion and 3.58 kg/m2 of sand.
4. Asphalt Rubber Chip Seal Test Section Evaluation
Reference to the control section represents the material components of the ARCS placed on
the northbound #2 lane of project 11-241104 (portion) on Imp–86 PM 37.3/43.3 on September
17, 2003 (Figure 3). The ARCS test sections were reviewed on April 27, 2006 and April 26,
2007, which represent the one year and two year construction anniversaries. The test sections
were reviewed for aesthetics, crack mitigation and were also tested for coefficient of friction
by California test method 342 and skid testing by ASTM E274-06. All test sections with the
exception of test section # 6 (0.27) were greater than the minimum coefficient of friction of 0.30
(CTM 342) and all test sections achieved skid numbers above 30 at SN 40 (ASTM E274-06).
Figure 3. Control section IMP 86 northbound number 2 lane
Test Section #1
As illustrated in Figures 4 and Figure 5 lowering the application rate of the asphalt rubber
binder from 2.58 l/m2 to 2.264 l/ m2 improved the aesthetics as compared to the control
section.
Figure 4. Test Section # 1 - 7-13-05
Figure 5. Test Section # 1 - 4-26-07 – 11:55 PM - Station 467+67

Test Section #2
As illustrated in Figures 6 and Figure 7, the wheel path flushing was less than in test
section #1 due to the larger 12.5mm coarse aggregate gradation.
Figure 6. Test Section #2 – 7-13-05
Figure 7. Test Section #2 – 4-26-07 – 12:18 PM – Station 473+80

Test Section #3
Figures 8 and Figure 9 illustrate that test section #3, was rated the best for aesthetics due to
the larger sized aggregate (New 12.5 mm coarse), the use of a harder based binder (PG70-10)
and a reduced binder spread rate (2.264 l/m2).

Test Section #4
Figure 10 and Figure 11 illustrate that test section #4, had slightly more flushing in the
wheel paths as compared to test section # 3. However, test section # 4 was rated the best when
considering crack mitigation and aesthetics.

Test Section #5
Figure 12 and Figure 13 illustrate that test section#5, had significantly more flushing in
the wheel paths as compared to test section # 4 even though all the application components
were the same except for elimination of the asphalt modifier. Without the asphalt modifier,
the binder formulation had to be heated to 221.1℃, which is above the maximum specified
temperature of 218.3℃, in order to achieve a reaction within a reasonable time period for
application.
Figure 13. Test Section #5 – 4-26-07 – 12:53PM – Station 494+40

Test Section #6
Figure 14 and Figure 15 illustrate that test section #6, as compared to test section #5
exhibited a significant improvement in crack mitigation and aesthetics. Although the binder
application rate in test section #6 was slightly higher (0.14 l/m2) than the binder application
rate in test section #5, without the asphalt modifier, the flushing was still slightly less.

4.1 General
- The six test sections validated the upper limits of the Caltrans maintenance selection
criteria (MTAG 2003).
- The changes implemented for all the test sections, excluding test section #5, resulted in
greater performance criteria, as compared to the control section based on the two year
review of the test sections.
- Coefficient of friction numbers (California TM 342) were increased in each test section
as compared to the Route 86 control section.
Satisfactory performance can de defined as the ability of the ARCS to perform under
high desert temperatures, to sustain heavy truck traffic loading without flushing/bleeding in
the wheel paths and to have a coefficient of friction value of at least 0.30 as determined by
California Test Method 342.
Changes that were implemented for all the test sections, except test section #5, resulted
in higher quality seal coats (less bleeding) as compared to the “control section”. The binder
for test section #5 eliminated the asphalt modifier, which resulted in a less than acceptable
aesthetic condition. The coefficient of friction value determined by California TM 342 was
greater than the 0.30 specified for all the test sections, except test section #6, which was 0.27.
5. Conclusions
5.1 Asphalt-Rubber Binder
- The minimum reaction time currently specified of 45 minutes is not adequate to

provide a stable binder viscosity and will not produce the desired binder properties.
- Reduction of the asphalt rubber binder spread rate from the control section spread rate
reduces flushing but can affect the mitigation of reflective cracks.
- PG 70-10 asphalt rubber binder does not mitigate reflective cracking as well as
AR-4000 but yields better results to mitigate flushing.
- The elimination of the asphalt modifier results in an increase in production time due to
the slower reaction time with the rubberized binder.
- With the elimination of the asphalt modifier from the asphalt rubber binder in test
section #5, moderate to heavy flushing was observed in the wheel paths.
- Tests validated that asphalt rubber binder can be spread less than 2.490 l/m2, under
suitable field and climatic conditions,
- Test validated that PG 70-10 non-polymer modified binder can be reacted to produce
asphalt rubber binder, but requires a longer reaction time.
5.2 Aggregate Screenings
- New 12.5 mm coarse aggregate gradation reduced flushing.
- New 12.5 mm coarse aggregate gradation, with flush coat, did not increase liability for
windshield damage.
- Reduction of the aggregate spread rate does not decrease flushing.
Use of the new 12.5mm coarse aggregate gradation and the PG 70-10 base stock resulted
in a more aesthetically acceptable seal coat for test sections 2, 3, 4 and 6 as compared to the
control section seal coat.
Note:
It should be noted that the untreated pavement areas between each of the test sections
have significantly deteriorated in the last two years with more cracking and pavement rutting
compared to the treated test sections. This confirms a well-established fact that a timely placed
preventative maintenance treatment will greatly extend the service life of a pavement.
6. Recommendations
- Implementation of the Time-Temperature-Viscosity curve with two descending

viscosity readings to account for any variation in the asphalt-rubber binder during
production.
- Increase the crumb rubber content from the current specified 20±2 % to 21±1 %. This
revision would result in a minimum crumb rubber content of 20%. Based on previous
projects, the lower specified percentage of 18% has been supplied. This change will
extend the performance and assist in reflective crack mitigation.
- Specify that the new 12.5 mm gradation, which is a more uniform, one-sized aggregate,
be used for the aggregate screenings for surface treatments.
- Specify a harder higher performance grade paving asphalt (PG 70-10 non-polymer
modified for the desert area/truck routes) for use in asphalt rubber binder formulations.
- Implementation of Non-Standard Special Provision 37-030, as presented in ARCS

2008 in Appendix I, will result in increased performance and aesthetics for asphalt
rubber chip seals throughout the State of California.
7. References
(ARCS 2005) Labar, John, Rizzutto, J. Shawn, Jones, Don, Brown, Lance, VanKirk, Jack,
Corcoran, John, Webster, Chuck, “Asphalt –Rubber Chip Seal (ARCS) Task Force
Report.” California Department of Transportation, October 2005.
(ARCS 2008) LaBar, John, Rizzutto, J. Shawn, Jones, Don, Brown, Lance, Vankirk, Jack,
Corcoran, John Webster, Chuck, “Asphalt –Rubber Chip Seal (ARCS) Task Force Report
II.” California Department of Transportation, September 2008.
(HEITZ 1992) Heitzman, Michael A; Design and Construction of Asphalt Paving Materials
with Crumb Rubber Modifier, FHWA Publication SA-92-022, May 1992.
(HUFF 1980) Huff, B.H., ARCO Concept of Asphalt-Rubber Binders, Asphalt-Rubber User-
Producer Workshop, Scottsdale, Arizona. May 7-8, 1980.
(MTAG 2003) Maintenance Technical Advisory Guide (MTAG), Caltrans Office of Pavement
Preservation, October 2003.
(CTARUG 2003) Asphalt Rubber Usage Guide, Caltrans, Division of Engineering Services,
Materials Engineering and Testing Services (METS), January 2003.
The Development of a Design Procedure

and Usage Criteria for Hot Applied Chip
Seal Applications
Jeffrey R. Smith
Assistant Marketing/Technical Director

International Surfacing Systems
P.O. Box 4770
Modesto, CA 95352
6751 West Galveston Street
Chandler, AZ 85226
jeff.smith@asphaltrubber.com
ABSTRACT. This paper will discuss the various hot applied and highly modified field blended
binder materials that are currently being manufactured in the United States and utilized in
chip and cape seal pavement preservation applications. These pavement preservation systems
serve as alternatives to expensive construction or re-construction techniques.
Further discussion will focus on a specific chip seal design method which includes binder
design and aggregate retention testing to determine materials compatibility prior to a project
being constructed. Binder and aggregate application rates, proper aggregate quality and
gradation, along with other construction issues and procedures; including traffic volume and
climatic region are variables that will also be considered as part of this “design” process.
KEYWORDS: AR/PMAR/RAB, Aggregate Retention, Pavement Preservation, PCI.

1. Introduction
Asphalt-rubber binder was formulated and developed by Mr. Charles H. McDonald in the
early to mid 1960’s. Initially the usage criterion for asphalt-rubber binder was limited to chip
seal construction also known as Stress Absorbing Membrane (SAM) and Stress Absorbing
Membrane Interlayer (SAMI) applications. This asphalt-rubber material has since gone
through major evolvement concerning binder design procedures, required testing parameters,
laboratory testing equipment, and variations of formulation. Concerning actual chip seal
design procedures, what we have learned through the research and development of asphalt-
rubber binder, has translated into a focus on the quality of the cover aggregate, further testing
and strict quality requirements of said aggregate, along with a proposed new testing protocol
to account for many of the asphalt-rubber chip seal design variables, resulting in an overall
comfort level of individual project quality and long term performance. Further evolvement
of blending and reaction equipment, construction equipment and improved construction
procedures has also contributed to the overall predictability and success of SAM and SAMI
chip seal projects.
Today there are many variations concerning the use of hot applied binder materials,
their formulations and intended use. Some of these binder materials are considered second
generation A-R (Asphalt-Rubber) and others are considered RAB (Rubberized Asphalt
Binder). The basic difference is the percentage of total modifier(s) that is (are) field blended
with the specified PG asphalt cement and the presence of particulate CRM in the binder which
relates directly to the viscosity that is tested in the field at the end of recommended interaction
time(s). This resulting viscosity that is achieved in the field, regardless of binder type, directly
relates to the improved physical properties associated with each individual binder material.
These physical properties are determined through the initial binder design profile and proper
blending procedures in the field, including vast improvements concerning field QC/QA.
2. Binder Design Effort
When considering binder design, Tables 1 and 2 demonstrate the consistency of viscosity
range and this relationship to physical properties of an A-R and PMAR (Polymer Modified
Asphalt-Rubber) binder types. A variety of interaction periods are normally conducted to
evaluate stability and consistency of the physical properties over a specific timeframe. Table
3 demonstrates the consistency of the physical properties concerning a RAB or PG 76-22TR
Binder.
Table 1. (A-R) Asphalt-Rubber Binder
Minutes of Reaction
Test Performed 45 90 240 360 1440 Specified Limits

Viscosity, Haake @ 190ºC, Pa’s,
2,200 2,500 2,500 2,400 2,100 1,500 – 4,000
(10³), or cP
Resilience @ 25ºC, %
48 ---- 48 ---- 49 18 Minimum
Rebound (ASTM D3407)
Ring and Ball Softening
72 72 72 73 73 52 – 74
Point, ºC (ASTM D36)
Cone Penetration @ 25ºC, 150g,
21 ---- 22 ---- 21 20 – 70
5 sec., 1/10 mm (ASTM D217)
Table 2. (PMAR) - Polymer Modified Asphalt-Rubber
Minutes of Reaction
Test Performed 45 90 240 360 1,440 Specified Limits

Viscosity, Haake @ 190ºC
1,900 2,100 2,200 2,000 1,800 1,500 – 4,000
Pa’s, (10³), or cP
Resilience @ 25ºC, %
52 ---- 53 ---- 55 25 Minimum
Rebound (ASTM D3407)
Ring and Ball Softening
74 75 74 76 75 55 – 80
Point, ºC (ASTM D36)
Cone Penetration @ 25ºC,
150g, 5 sec., 1/10 mm 24 ---- 25 ---- 25 20 – 65
(ASTM D217)
Table 3. (RAB) - Rubberized Asphalt Binder or PG 76-22TR (California “modified binder”)
Tests on Original Asphalt Test Method Spec. Limits Results Results

Apparent Viscosity @ 135ºC, Pa’s AASHTO T316 3.00 max. 2.795
Dynamic Shear, G*/sinδ, kPa AASHTO T315 1.00 min. ----
76ºC 2.28
82ºC 1.25
Predicted Pass/Fail Temp., ºC 83.9
Solubility in Trichloroethylene, % AASHTO T44 93 min. Report
Tests on Residue from RTFO AASHTO T240
Dynamic Shear, G*/sinδ, kPa AASHTO T315 2.20 min.
Tests on Original Asphalt Test Method Spec. Limits Results Results

76ºC 4.76
82ºC 2.64
Predicted Pass/Fail Temp., ºC 83.9
Delta (δ) at 2.20kPa, ºC 80 max. 75.54
Mass Loss, % AASHTO T240 1.00 max. 0.343
Elastic recovery, 25ºC, % 65 min. 72.0
Tests on Residue from PAV AASHTO K28 100ºC 110ºC
Dynamic Shear, G*/sinδ, kPa AASHTO T315 5000 max.
34ºC (Spec. Temp. -16 Grade) ---- 1,112
31ºC (Spec. Temp. -22 Grade) 1,533 1,589
28ºC (Spec. Temp. -28 Grade) ---- 2,162
Creep Stiffness, S, @ 60s, MPa AASHTO T313 300 max.

-12ºC 130 140
-18ºC 263 273
Slope, m-value AASHTO T313 0.300 min.

-12ºC 0.315 0.307
-18ºC 0.272 0.260
PG PG
Performance Grade AASHTO M320
76-22TR 76-22TR
The design effort, regardless of the type of hot applied binder that is specified, can be
very similar, but at the same time can differ based on the total percentage of modifier that is
required and the overall testing regiment that is specified. The design profile concerning A-R
and PMAR binder materials has remained fairly consistent over the years (see Tables 1 and
2) and is typically performed for each individual project to evaluate the compatibility and
interaction between component materials and to verify stability of the binder over time.
Apparent Viscosity – monitors the fluid consistency of the specified binder to ensure
pumpability and to identify binder changes which might affect chip seal or hot mix placement.
Measurement is achieved by the use of a rotational viscometer and the results are presented
in centipoise (cP) or Pascal Seconds (Pa-s). The required testing can be performed by either
using the Brookfield Viscometer or a handheld Haake type Viscometer. If the Brookfield
Viscometer is the required method for acceptance the Haake Viscometer should be calibrated
and correlated to the Brookfield measurement prior to use.
Penetration (ASTM D5 or AASHTO T49) – modified binder consistency can be measured

via the use of a penetrometer at low, moderate and high temperatures. Needle penetration
is usually the standard, performed at 39.2º F and 77.0º F. Cone penetration is typically used
with A-R and PMAR binder materials and is performed at the same temperatures as needle
penetration but is only utilized if the gradation of the crumb rubber modifier (CRM) is 14
mesh or larger. Results are presented in tenths of millimeter units.
Softening Point (ASTM D36 or AASHTO T53) – is an indicator of materials stiffness and
will show a tendency of the binder to flow at elevated temperatures. Measurement is achieved
by the “ring and ball” method and results are presented in Degrees Fahrenheit or Degrees
Centigrade.
Resilience (ASTM D5329) – measures the elastic properties of the modified binder and the
results are presented as percentage of rebound for the binder. Resilience is one of the most
important properties in the specifications and is considered a reliable and consistent measure
of elasticity.
3. Aggregate Requirements
Current specifications recommend the cover aggregate meet specific gradation

requirements, usually in accordance with ASTM C136 test method. The crushed coarse
aggregate shall have a percentage of wear not more than 7 percent at 100 revolutions and not
more that 30 percent at 500 revolutions, as determined by ASTM C131 test method. Ninety
percent of the crushed stone or crushed gravel should have a minimum of two fractured faces
on three sieves (#4 (4.75 mm), ⅜ inch (9 mm) and ½ inch (12 mm) and should not contain
more than 8 percent by weight of flat or elongated particles. Further aggregate testing can be
specified such as the sodium sulfate accelerated soundness test which does not allow a total
loss greater than 12 percent when the aggregate is subjected to 5 cycles in accordance with
ASTM C88 test method. Aggregate gradation should be determined by the application rate of
the specified binder which in turn should be based on the existing condition of the pavement
receiving the recommended pavement preservation treatment. Table 4 demonstrates common
gradation requirements for ⅜ inch (9 mm) and ½ inch (12 mm) aggregate materials utilized
in A-R or PMAR SAM and SAMI applications and Table 5 demonstrates common gradation
requirements for ¼ inch (4.75 mm) and ⅜ inch (9mm) aggregate materials utilized in hot
applied RAB field blend binder materials.
Table 4. Aggregate Gradation for A-R and PMAR (SAM and SAMI)
3/8 inch SAM & SAMI Gradation 1/2 inch SAM & SAMI Gradation
Sieve Size Percent Passing Sieve Size Percent Passing

½ inch (12 mm) 100 ¾ inch (15 mm) 100
⅜ inch (9mm) 70 – 85 ½ inch (12 mm) 95 – 100
¼ inch (4.75 mm) 0 – 15 ⅜ inch (9mm) 70 – 85
#8 (2.36 mm) 0-5 ¼ inch (4.75 mm) 0 – 15
#200 (75 µm) 0–1 #8 (2.36 mm) 0–5
---- ---- #200 (75 µm) 0-1
Table 5. Aggregate Gradation for RAB (PG 76-22TR – “Modified Binder”)
1/4 inch RAB Gradation 3/8 inch RAB Gradation
Sieve Size Percent Passing Sieve Size Percent Passing

½ inch (12 mm) 100 ¾ inch (15 mm) 100
⅜ inch (9 mm) 97 – 100 ½ inch (12 mm) 97 – 100
¼ inch (4.75 mm) 65 – 100 ⅜ inch (9 mm) 70 – 100
#8 (2.36 mm) 0 – 10 ¼ inch (4.75 mm) 0 – 15
#200 (75 µm) 0–1 #8 (2.36 mm) 0–5
---- ---- #200 (75 µm) 0-1
Regardless of the aggregate gradation this material is usually specified to be pre-coated or

hot pre-coated with 0.5 ± 0.25 percent emulsified binder or PG asphalt cement. When used
hot the temperature of the aggregate should be a minimum of 250ºF (121ºC) and should not
exceed 325ºF (162ºC).
4. Hot Applied Chip Seal Design
Once the specified binder has been determined and a binder design profile has been
completed, along with the aggregate being graded and confirmed to be of desirable quality,
the actual chip seal design can begin. The testing has been specified per Caltrans Materials
Laboratory “Vialit Test for Aggregate Retention in Chip Seals” (French Chip) as modified
from EN 12272-3 and further modified to accommodate hot applied binder materials, A-R,
PMAR and RAB (See Tables 1, 2 and 3). Table 6 demonstrates the procedure and test results
for this proposed test method.
Table 6. Modified Vialit Aggregate Retention
Mineral Aggregate Preparation

Aggregate Size, mm 6.3 – 9.5
#200 Washed? Yes/No Yes
Conditioning Prior to Application Dried @ 122ºF (50ºC) for 24 hours
Vialit Specimen Preparation

Binder Application Rate, g 79.0
Number of Aggregate Chips Applied 100
Binder Temperature @ Application 325ºF (163ºC)

Aggregate Temperature @ Application Ambient
Curing Condition of Specimen Plates, Phase 1 Cured @ 325ºF (163ºC) for 15 minutes
Curing Condition of Specimen Plates, Phase 2 Cured @ Ambient, 4 hours minimum
Condition of Specimen Plates @ Testing Cured @ Test Temperature for 30 minutes
Vialit Test Results

Spec. Limit
Aggregate Retention, % 90 min.
Test Temperature @ 41ºF (5ºC)
Test Temperature @ 23ºF (-5ºC)
Test Temperature @ 5ºF (-15ºC)
Test Temperature @ -8ºF (-22ºC) (Optional)
Once the specimen plates have cured at the various test temperatures then individual plates
are placed into the testing apparatus (see Figure 1) and each plate is tested according to the
following criteria. A 500 gram steel ball is dropped from a specific height, three times and the
number of stones that are retained on each plate is reported as percent aggregate retention at
the various test temperatures (see Figure 2). Ninety percent retention is the lowest value that
is acceptable at any one of the four test temperatures.
It is important to note that with each binder type that is tested for aggregate retention value,
the weight of the binder (application rate) is currently specified at 79 grams (approximately
.43 gallons per square yard). However this rate is associated for only a 3/8 inch (9 mm)
maximum size aggregate. There is now consideration to increase the weight of the binder
(application rate) to more accurately represent the actual field application rate(s) of the
A-R and PMAR binder materials and the potential use of a ½ inch (12 mm) maximum size
aggregate. Recommended minimum field application rates for RAB binder type is .40 gallons
per square yard (1.32 liters per square meter), for A-R binder type is .60 gallons per square
yard (2.86 liters per square meter) and for PMAR binder type is .65 gallons per square yard
(3.26 liters per square meter).
Figure 1. Modified Vialit Testing

Apparatus
Figure 2. Modified Vialit Test

If the aggregate retention has been verified as acceptable the chip seal design data is
documented and submitted to the specifying agency prior to the actual project beginning.
This should be required as part of the project specifications, which ensures the agency that, the
binder and aggregate materials that have been specified are compatible. This type of design
work should also be considered a base line concerning the testing of samples taken from the
project during construction, which in turn can verify the initial design work that was done.
5. Project Selection Criteria
There is much interest and a definite need concerning criteria by which the proper
binder materials and aggregate gradation can be selected as part of a pavement preservation
project. As a result of this interest there are many different concepts as to how this might be
accomplished, how the resulting data can be confirmed “repeatable” and applied to pavement
preservation projects that will potentially experience a wide range of variables, such as
differences in climate, topography, traffic volume, traffic type, etc.
Since many specifying agencies have functioning Pavement Management Systems (PMS)
this seems to be the most logical place to start the effort. A procedure needed to be identified
to accomplish this and establish a method which would potentially match existing pavement
condition with best case rehabilitation. This in turn would potentially give the agency a good
feel for what to expect in terms of constructability, cost effectiveness (initial and long term),
performance (initial and long term) and the potential for reduced maintenance costs over the
predicted life of the pavement preservation strategy being considered.
With this in mind it is then logical to look at Pavement Condition Index (PCI) to initially
determine if pavement preservation is the correct strategy. As each specifying agency is
different and conditions vary widely, these decisions should be made according to what the
agencies history is and what the expectations are to meet the end result. It is certain that
PCI is not going to give comprehensive data regarding underlying base condition, overall
pavement stability or account for other issues that may come into play during evaluation.
However, a range of PCI can give the agency an indication as to what type of pre-maintenance
should be required and potential selection criteria regarding a pavement preservation strategy
that will have the highest potential for success. As this evaluation criterion is adopted the
various agencies need to utilize their past history of maintenance activity and performance
of their chosen pavement preservation treatments to aid in the decision making process prior
to establishing “new” specifications and updated patterns of use for the various “improved”
pavement preservation strategies that may be selected.
6. Pavement Condition and Materials/System Choice
When an agency begins to review and categorize the condition of their pavements and
establish an overall PCI, the process to decide what type of rehabilitation that will be done
on these pavements can be confusing at best. There are a myriad of materials and systems
to utilize and most pavement engineers, public works directors and decision makers rely
on materials experts and sales people to guide them in the use of the specific materials and
systems that are available in the marketplace. Based on experience and performance history
the timing of any materials application is critical for that material to perform as expected.
When dealing with conventional emulsified binder, modified emulsified binder, hot applied
binder, modified hot applied binder, asphalt-rubber binder and polymer modified asphalt-
rubber binder, each material can be part of or included in a specific pavement preservation
strategy. The information in Table 7 suggests criterion of usage regarding specific pavement
preservation treatments, each having different life expectancies based on the many variables
previously discussed.
Table 7. Suggested Usage Criteria
PCI Range Recommended Pavement Preservation Method

75 - 90 Fog Seal, Scrub Seal, Slurry Seal, Conventional Emulsion Chip
Seal, Polymer Modified Emulsion (PME) Chip Seal, or Rejuvenating
Emulsion Chip Seal
50 - 75 Micro-Surface, PG 76-22PM and PG 76-22TR (field or terminal blend)
Chip Seal, Scrub Seal in Conjunction with PG 76-22 PM / TR Chip
Seal or PG 76-22 PM or TR Cape Seal
35 - 50 A-R (Asphalt-Rubber) Chip Seal, PMAR (Polymer Modified Asphalt-
Rubber) Chip Seal, RASS (Rubberized Asphalt Scrub Seal) in
Conjunction with A-R / PMAR Chip Seal, A-R / PMAR Cape Seal, or
A-R / PMAR Three Layer Cape Seal System
20 - 35 A-R / PMAR Three Layer Cape Seal System, A-R / PMAR Two Layer
Overlay System or A-R / PMAR Three Layer Overlay System
It is very important to recognize that the various binder materials mentioned in Table 7,
which may be utilized in a chip seal, cape seal or multi-layered system(s) have limitations
concerning application rate. This will in turn effect the recommendation of proper aggregate
gradation and application rate of the aggregate, which will in turn affect initial and long term
performance of the pavement preservation system being utilized. When specifying the type
of slurry seal to use as the wearing course in a cape seal pavement preservation system it is
very important to evaluate the street type (residential, collector or major arterial) and choose
accordingly. When dealing with residential street type a Type II Slurry Seal is normally
utilized. When dealing with collector or major arterial streets with traffic counts higher than
3000 vehicles per day then the Type II Slurry Seal should be upgraded to a Micro-Surface.
The following photographs (figures 3 to 7) could be considered representative of certain
pavement conditions indicated by PCI.
Figure 3. PCI of 90 (Indicates Fog Sealing as Proper Pavement Preservation Technique
Figure 4. PCI – 75 (Indicates Slurry Seal, Microsurface, Conventional Emulsion/PME Chip

Seal or Rejuvenating Emulsion Chip Seal as Proper Pavement Preservation Techniques)
Figure 5. PCI – 50 (PG 76-22TR/PM Chip Seal or Cape Seal as Proper Pavement
Preservation Techniques)
Figure 6. PCI – 35 (Indicates A-R/PMAR Chip Seal or Cape Seal or Three Layer Cape Seal
System as Proper Pavement Preservation Technique)
Figure 7. PCI – 20 (Indicates A-R/PMAR Three Layer Cape Seal, A-R/PMAR Two Layer
Overlay System, A-R/PMAR Three Layer Overlay System or Reconstruction)
7. Individual Project Variables
Each pavement preservation project has its own specific set of variables to consider.
Not only from the standpoint of existing surface condition, pre-maintenance requirements,
materials recommendations and appropriate system to specify but other variables regarding
construction issues, traffic conditions (current and future), ambient and surface temperatures,
climate and topography. Taking into consideration and accounting for as many of these and
other variables as possible will result in a pavement preservation project that will not only be
cost effective, but will also give the specifying agency the expected long term performance.
Construction variables can be overwhelming as so many different issues need to be

considered and accounted for through drafting and enforcement of proper specifications.
Construction variables range from proper binder application rate (.08 to .75 gallons per
square yard depending on the specified binder and pavement preservation technique), proper
aggregate gradation, aggregate quality and application rate (if cover aggregate is required,
based on the selection of the pavement preservation system), including the requirement of pre-
coated or hot pre-coated aggregate (pre-coating is normally associated with the use of “hot
applied binder materials”). Concerning a chip seal application proper and timely rolling with
the appropriate number of rollers is critical to the overall embedment of the aggregate into the
specified binder and end result aggregate retention value. Other variables such as ambient
temperature, surface temperature and project topography will affect the speed of the chip seal
“train”, so understanding of these variables is also critical as to how construction procedure
adjustments need to be made in the field.
8. Conclusion
Utilizing proper pavement preservation techniques is not an exact science. Accounting for
as many variables as possible through proper design, specifications, specification enforcement,
construction procedures and inspection will consistently result in a cost effective, high quality,
long lasting projects. As the specifying agency looks at maintaining pavements with any level
of PCI there is a chance for continued surface deviations due simply to the existing surface
condition, pre-maintenance that was accomplished (if any), and the pavement preservation
system that was chosen. This is certainly the case if the existing pavement condition is
below 40 on the PCI. Choosing the proper pavement preservation strategy that is initially
cost effective while giving the specifying agency the longest life cycle with minimal or
maintenance free performance should be the ultimate goal, this can only be done if focus is
placed on the variables discussed in this paper and by understanding the limitations of the
various component materials utilized as part of the many pavement preservation strategies
available today.
9. Bibliography
Caltrans, “Maintenance Technical Advisory Guide”, State of California Department of

Transportation, Sacramento California, 2003.
Cano, J.O. and Charania, E., “The Phoenix Experience Using Asphalt-Rubber”, Proceedings
of the National Seminar on Asphalt-Rubber, Kansas City Missouri, October 1989.
Griffith, A. and Hunt, E., “Asphalt Cement Chip Seals in Oregon”, State Research Project
#546, Oregon Department of Transportation, June 2000.
Huddleston, S., Kramer, T., and Stonex, A., Miscellaneous Laboratory Binder Design, Chip
Seal Design, Hot Mix Design and Various Research Projects, Mactec Engineering and
Consulting, Inc., 2002 – 2008.
Morris, G.R. and McDonald, C.H., “Asphalt-Rubber Stress Absorbing Membrane, Field
Performance and State of the Art”, Transportation Research Record 595, 1976 pp. 52-58.
Rosner, J.C. and Chehovits J.G., “Chemical and Physical Properties of Asphalt-Rubber
Mixtures – Phase III”, Volumes 1 – 5, FHWA/AZ: 82-159, Arizona D.O.T., June 1982.
Schnormeier, R.H., “Fifteen Year Pavement Condition History of Asphalt-Rubber Membranes

in Phoenix, Arizona”, 1986 Meeting of the Transportation Research Board, Washington
D.C.
Scofield, L.A., “The History, Development and Performance of Asphalt-Rubber at ADOT”,

Report Number AZ-SP-8902, ADOT, December 1989.
Senadheera, S., Tock, R.W., Hossain, S., Yazgan, B. and Das, S., “A Testing and Evaluation
Protocol to Assess Seal Coat Binder – Aggregate Compatibility”, Texas Department of
Transportation / Texas Tech University – Department of Civil Engineering, August 2003
Van Kirk, J. and Holleran, G., “Asphalt-Rubber In Chip Seals, Slurry, Microsurfacing and Hot
Mix”, 4th Annual Bitumen Asia Conference, Singapore, 1998.
Way, G. B., “Prevention of Reflective Cracking Minnetonka – East”, Report Number

1979GWI, Arizona Department of Transportation, August 1979.
Zareh, A. and Way, G. B., “35 Years of Asphalt-Rubber Use in Arizona”, Proceedings Asphalt-
Rubber 2006 Conference, Palm Springs California, October 2006.
Chapter 7
Terminal Blend Binders

(Papers on this section contain contributions with terminal blend binders that
do not satisfy ASTM ??? specifications. However they have been added to
these Proceedings not only to bring to Nanjing the authors so that later they
can be the champions of change in their respective countries but also they
provide in-site on the relative benefits of terminal blends to other binders.)
Terminal Blends (rubberized asphalts) 831
Properties of HMA Mixtures Produced

with Polymer-Modified and Tire Rubber-
Modified Asphalt Binders
P. E. Sebaaly* — H. K. Sebaaly* — E. Y. Hajj*
*Pavements/Materials Program, Dept of Civ. & Env. Engineering MS257, University of

Nevada, Reno, NV 89557, United States
psebaaly@unr.edu
haissams@unr.edu
elieh@unr.edu
ABSTRACT. This paper summarizes the impact of rubber modified binders on the laboratory
performance of hot mix asphalt mixtures (HMA). The purpose of this study is to compare
the laboratory performance of the mixtures manufactured with the PG64-28TR and PG76-
22TR rubber modified binders to the performance of mixtures manufactured with the Nevada
Department of Transportation (NDOT) currently used PG64-28NV and PG76-22NV polymer
modified binders, respectively. The PG64-28TR and the PG76-22TR are terminal blend
binders modified with 10% ground tire rubber. The mixtures were evaluated in terms of their
resistance to rutting, fatigue cracking, thermal cracking, dynamic modulus, and moisture
sensitivity. The PG64-28TR mix showed lower rutting resistance than the PG64-28NV mix
and lower laboratory fatigue resistance than the PG64-28NV mix at a strain level higher
than 500 microns. On the other hand, the PG76-22TR mix showed higher rutting resistance
than the PG76-22NV mix and lower laboratory fatigue resistance than the PG76-22NV mix
at all strain levels. The PG64-28TR mix and the PG64-28NV mix showed good resistance to
thermal cracking, and approximately similar dynamic modulus at 21.1℃. On the other hand,
the PG76-22TR mix showed higher dynamic modulus than the PG76-22NV mix at 21.1℃. All
mixtures exhibited good resistance to moisture damage.
KEYWORDS: Terminal blend, rutting, fatigue, thermal cracking, dynamic modulus, moisture
sensitivity.
1. Introduction
In the United States and Europe, approximately 300 million used tires are generated each
year (Amirkhanian et al., 2004). A typical tire contains 60% rubber, 20% steel, and 20% fiber
(Putman et al., 2004). Therefore, in many countries this has been a serious problem due to
the lack of landfill space, and environmental pollution caused by burning the rubber. There
are many applications in civil engineering to solve this problem and one them could be the
utilization of the tires rubber in pavement by including them in the asphalt mixture. The
concept of using tires rubber in asphalt mixtures follows three processes: 1) the dry process
where the crumb tire rubber is used as an aggregate in the mix, 2) the wet process where the
crumb rubber is blended into the asphalt binder on the job site and 3) the terminal blend,
introduced in the 1980s, where the tire rubber is blended into the asphalt binder at the terminal
or refinery and shipped to the hot plant.
Over the past few decades, the wet process has been more commonly used than the dry
process. The traditional method of producing the wet process is to blend the crumb rubber
into the asphalt binder on the job site at elevated temperatures and allow it to react prior to
mixing with the aggregate. The rubber modified asphalt binder produced using this technique
is referred to as “crumb rubber modified (CRM) binder”. The terminal blend technology has
been used since the mid 1980s in many states. The tire rubber is blended into the asphalt
binder at the terminal or refinery and shipped to the hot plant as a finished product (Sebaaly
et al., 1998). This technology eliminates the need for the blending unit at the job site. In
addition, this technology allows the product to be engineered specifically to meet the climatic
conditions defined by the Superpave performance graded system “PG”. The rubber modified
asphalt binder produced using this technique is referred to in this paper as “tire rubber-
modified asphalt binder”. Since the CRM binder technology has been around for quiet a long
time, several research studies have evaluated the performance of CRM mixtures. The terminal
blend technology is relatively new and therefore, has not been extensively evaluated.
The objective of this research effort is to compare the laboratory performance of Nevada’
s HMA mixtures made with rubber-modified asphalt binders to the performance of HMA
mixtures made with currently used polymer modified binders (Sebaaly et al., 2007). Two
Nevada HMA mixtures were used in the study: one mix from the south and one mix from the
north. Each mix was designed with the two types of binders which resulted in a total of four
HMA mixtures that were evaluated in this study. The laboratory performances of the mixtures
were evaluated in terms of the following properties:
- Moisture sensitivity
- Dynamic modulus
- Resistance to rutting
- Resistance to fatigue cracking
- Resistance to thermal cracking
2. Materials and mixtures characteristics
Two aggregate sources were identified, one source in northern Nevada (Lockwood) and
one source in southern Nevada (Sloan). Currently the Nevada Department of Transportation
(NDOT) specifies a PG64-28NV and a PG76-22NV polymer-modified binders for the

northern and southern parts of the state of Nevada, respectively. The NV extension indicates
that the binder is graded with the PG-special system which includes the Superpave binder
system plus the following properties: toughness, tenacity, and ductility on original and RTFO
binder at 4.4℃.
The PG64-28TR and the PG76-22TR binders are modified with 10% ground tire rubber,
and were graded following the Superpave PG system plus the elastic recovery test at 25℃ for
the original binder. All asphalt binders were supplied by the Paramount Petroleum Company,
Nevada, except for the PG76-22NV asphalt binder that was supplied by Ergon Asphalt,
Nevada. A total of four HMA mixtures were evaluated in the laboratory.
- A northern Nevada mixture designed with a polymer-modified PG64-28NV binder and

aggregate from the Lockwood source.
- A northern Nevada mixture designed with a tire rubber-modified PG64-28TR binder
and aggregates from the Lockwood source.
- A southern Nevada mixture designed with a polymer-modified PG76-22NV binder and
aggregate from the Sloan source.
- A southern Nevada mixture designed with a tire rubber-modified PG76-22TR binder
and aggregates from the Sloan source.
Figure 1 shows the gradation of the Lockwood and Sloan aggregates, respectively.
Both gradations meet the specifications for NDOT type 2C gradation. All four mixtures
were treated with 1.5% of hydrated lime by weight of aggregates according to the NDOT
specifications.
100
90
80
Type 2C upper Limit
70 Max. Density line
Lockwood
60
Percent Passing
Gradation
50 Sloan
Gradation
40
NDOT Type 2C Limits
Type 2C Lower Limit Max Density Line
30
Lockwood gradation
20 Sloan gradation
10
0
#200 (0.075mm)
#50 (0.3mm)
#30 (0.6mm)
#100 (0.15mm)
#4 (4.75mm)
#10 (2.0mm)
#40 (0.425mm)
#16 (1.18mm)
#8 (2.36mm)
3/8" (9.5mm)
1/2" (12.5mm)
1" (25.0mm)
3/4" (19.0mm)
0.45
(Sieve Opening)
Figure 1. Gradation of the Lockwood and Sloan aggregates blend

2.1. Mix designs
The NDOT Hveem Mix Design Method as outlined in the NDOT Testing Manual was
used as a guide to design the mixtures. The heated aggregate samples were mixed with
various amounts of asphalt binder so that at least two were above and at least two were
below the expected optimum asphalt content. The optimum binder content was determined
by identifying the asphalt contents which provide the specified minimum Hveem stability,
then the highest binder content that has air voids between 4 and 7%, and VMA between
12 and 22%. The mixture and volumetric properties along with the corresponding NDOT
specifications at the design binder contents are summarized in Table 1.
Table 1. Mix Design Summary and Moisture Sensitivity Properties
PG64- PG64- PG76- PG76-

Property SSRBC1
28NV 28TR 22NV 22TR
Rotational viscosity at 135℃, Pa.s 0.61 0.60 1.98 1.50 3.00 Max.
Optimum Binder Content (%DWA2) 4.8 5.0 3.8 3.9 -
Air Voids in Total Mix (%) 5.9 6.0 4.7 5.7 4-7
Void in Mineral Aggregate (%VMA) 14.1 14.8 13.2 14.3 12 - 22
Hveem Stability 40 39 42 39 37 Min.
AASHTO T-283 Moisture Sensitivity Test
Air Voids, % Average 6.8 6.6 7.2 7.1 -
Tensile Average 807 786 910 1124
448 kPa4
Strength at 25
690 kPa5
℃, kPa (Dry) CV (%) 2 10 10 10
3
Tensile Average 765 752 772 958

Strength at 25 -
℃, kPa (Wet) CV (%) 6 4 8 10
3
Tensile Strength Ratio, % 95 95 85 85 70%

1. Denotes “Standard Specification for Road and Bridge Construction”.
2. Denotes “dry weight of aggregate”.
3. Denotes “coefficient of variation” defined as the ratio of the standard deviation over the
average times 100.
4. Minimum tensile strength for the PG64-28NV mix and the PG64-28TR mix.
5. Minimum tensile strength for the PG76-22NV mix and the PG76-22TR mix.
3. Laboratory evaluation
The laboratory produced mixtures were evaluated in terms of the following properties:
moisture sensitivity, dynamic modulus, resistance to rutting, resistance to fatigue cracking,
and resistance to thermal cracking. The laboratory performances of mixtures manufactured
with tire rubber modified binders were compared to the corresponding laboratory performance
of the mixtures manufactured with polymer modified binders.
3.1. Moisture sensitivity
According to the NDOT Hveem mix design method, the dry tensile strength (TS)
should meet a minimum of 448 kPa for the PG64-28NV mix and the PG64-28TR mix,
and a minimum of 690 kPa for the PG76-22NV mix and the PG76-22TR mix. The tensile
strength ratio (TSR) should be at least 70%. The TSR is measured as the ratio of the
moisture conditioned tensile strength over the unconditioned TS of the HMA. The moisture
conditioning process followed the procedure described in AASHTO T-283 test with the
exception of using five samples at both the conditioned and unconditioned stages.
Table 1 summarizes the results of the mixtures TS and TSR at 25℃. The coefficients of
variations represent the variability of the data among the replicates for each test. The data
in Table 1 shows that all mixtures met the NDOT Hveem mix design criteria for moisture
damage. In the case of the northern mixtures, the PG64-28NV mix and the PG64-28TR mix
exhibited similar TS and TSR ratio. In the case of the southern mixtures, the PG76-22NV mix
exhibited lower TS values than the PG76-22TR mix but similar TSR value.
3.2. Dynamic modulus
Stiffness is an important property to consider for the design of any structure. In the case of
HMA mixtures the stiffness changes with temperature and frequency. Thus, when designing
a pavement structure, the moduli values must be accurately represented at all desired
temperature-frequency combinations. The dynamic modulus represents the stiffness property
of HMA mixtures. The dynamic modulus test as standardized in AASHTO TP62 was used
to develop the dynamic modulus master curve of the HMA mixtures. For linear viscoelastic
materials such as HMA mixtures, the stress-strain relationship under a continuous sinusoidal
loading is defined by its complex dynamic modulus (E*). Mathematically, the “dynamic
modulus” is defined as the absolute value of the complex modulus.
The recommended test series for the development of the E* master curve for use in
pavement response and performance analysis consists of testing at temperatures of -10,
4.4, 21.1, 37.8 and 54℃ and loading frequencies of 0.1, 0.5, 1.0, 5, 10, and 25 Hz at each
temperature. The development of a master curve allows the modulus to be determined at any
reasonable temperature-frequency combination.
The dynamic modulus test was performed on all four mixtures. Master curves were
constructed using the principle of time-temperature superposition. First, a standard reference
temperature was selected (21.1℃), and then data at various temperatures were shifted with
respect to loading time until the curves merge into a single smooth function. The 150 mm
diameter cylindrical samples were compacted in the Superpave Gyratory Compactor (SGC) to
an air voids content of 7±1%. A total of three replicates were tested for each mixture.
The loading frequency simulates the speed of traffic loads on the pavement. A higher
loading frequency represents a fast moving load and a lower loading frequency represents a
slow moving load. Figure 2 shows the dynamic modulus master curves of the various mixtures
at 21.1℃. Figure 2 shows that the PG64-28TR mix and the PG64-28NV mix have similar
dynamic modulus property while the PG76-22TR mix has a higher dynamic modulus property
than the PG76-22NV mix at all frequencies. This indicates that the PG76-22TR mix is stiffer
than the PG76-22NV mix.
a b
Figure 2. (a) Dynamic modulus of the PG64-28NV mix and the PG64- 28TR mix at 21.1℃
(b) Dynamic modulus of the PG76-22NV mix and the PG76-22TR mix at 21.1℃
3.3. Resistance to rutting
The rutting resistance of the various mixtures was evaluated in the laboratory using the
repeated load triaxial (RLT) test and the asphalt pavement analyzer (APA) test.
3.3.1 Rutting behavior of mixtures in the repeated load triaxial test
The repeated load triaxial (RLT) test measures the axial permanent deformation in the
HMA mixture as it is subjected to triaxial stress conditions. The test specimen was cored
from the center of a 150 mm by 180 mm gyratory compacted sample. The tested samples had
air voids of 7±1%. The triaxial condition was achieved by applying a static radial confining
pressure of 207 kPa using compressed air and a repeated deviatoric vertical stress. The
repeated deviatoric stress was applied for 0.1 seconds followed by a 0.6 seconds rest period.
The test was conducted at three different temperatures and on three replicates from each
mixture at each temperature for a total of 12,000 cycles with continuous measurements of the
vertical deformation along the middle 100 mm of the sample. This enabled the evaluation of
the permanent vertical strain in the sample as a function of load cycles.
The RLT test results were used to characterize the rutting behavior of the various mixtures
in the form of the performance model provided by the new mechanistic-empirical pavement
design guide (MEPDG). Equation 1 shows the performance model suggested in the MEPDG
to assess rutting in the HMA layer relates the ratio of axial permanent strain (ε p) and the
resilient axial strain (εr) to the number of loading cycles (N) and temperature (T).
[1]
In this equation, a, b, and c are experimentally determined coefficients for the permanent
deformation behavior of HMA mixtures in the secondary stage shown in Figure 3. The
primary and tertiary stages are not taken into account for developing the rutting model. The
secondary stage corresponds to small rate of rutting exhibiting a constant rate of change of
rutting that is also associated with volumetric changes; however, shear deformations also
increase.
Figure 3. Typical repeated load permanent deformation behavior of pavement materials
The data from the RLT were plotted in the form of permanent strains versus loading cycles
and the ratio of permanent strain over the resilient strain versus loading cycles as shown in
Figure 4. Each curve represents the average response of three replicate samples tested at each
of the specified testing temperatures listed on the individual plots in Figure 4. Equations 2 to 5
describe the rutting models developed by fitting a multiple linear regression for the test data at
the various temperatures for each mixture. The SAS macro called REGDIAG was used to fit
the multiple linear regressions (MLR).
εp 0.222908 1.432324
PG64-28NV mix: = − 2.10154 N T (R2 = 98.3%) [2]
εr
εp 0.257669 1.071886
PG64-28TR mix: = − 1.46011N T (R2 = 96.0%) [3]
εr
εp 0.197635 1.437127
PG76-22NV mix: = − 2.10406 N T (R2 = 99.8%) [4]
εr
εp 0.238186 1.165057
PG76-22TR mix: = − 1.74528 N T (R2 = 99.3%) [5]
εr
(a) (b)
Figure 4. (a) Permanent axial strain versus number of loading cycles

(b) Ratio of permanent axial strain to resilient axial strain versus number of loading
cycles
The rutting behavior of the tire rubber-modified asphalt mixtures were statistically
compared to the polymer modified asphalt mixtures using the least square mean statistical
technique. The ratio of the permanent axial strain to the resilient axial strain (ε p/εr) after
12,000 loading cycles were determined for each mixture using Equations 2 – 5 and compared
for any statistically significant difference. Table 2 summarizes the εp/εr values along with the
significance difference after 12,000 loading cycles. The following nomenclature is used in
Table 2 and throughout the paper:
- H: The εp/εr of the tire rubber-modified mixture is statistically significantly higher than
the corresponding polymer-modified mixture.
- L: The εp/εr of the tire rubber-modified mixture is statistically significantly lower than
the corresponding polymer-modified mixture.
- NS: No statistical significant different exist between the tire rubber-modified mixture
and the corresponding polymer-modified mixture.
Table 2. RLT Test Results at 12,000 cycles
Temperature, ℃ Mix εp/εr after 12,000 cycles Significance Level*

PG64-28NV 9.7
30 H+
PG64-28TR 16.1
PG64-28NV 17.5
50 H+
PG64-28TR 29.2
PG64-28NV 23.4
64 H+
PG64-28TR 38.8
PG76-22NV 6.9
30 L#
PG76-22TR 6.2
PG76-22NV 17.6
50 L#
PG76-22TR 15.5
PG76-22NV 27.5
70 L#
PG76-22TR 24.3
*
Comparing the NV mixtures to the TR mixtures.
+
εp/εr of the TR mixture is statistically significantly higher than the εp/εr of the NV
mixture.
#
εp/εr of the TR mixture is statistically significantly lower than the εp/εr of the NV mixture.
Table 2 shows that the PG64-28TR mix has significantly higher εp/εr than the PG64-28NV
mix at all three test temperatures of 30℃, 50℃, and 64℃. On the other hand, the PG76-22TR
mix has significantly lower εp/εr than the PG76-22NV mix at 30℃, 64℃, and 70℃. For the
same resilient strain the RLT test results shows that the PG64-28TR mix will exhibit a higher
permanent strain than the PG64-28NV mix resulting in a lower resistance to rutting for the
PG64-28TR mix compared to the PG64-28NV mix. On the other hand, for the same resilient
strain the PG76-22TR mix will exhibit a lower permanent strain than the PG76-22NV mix
resulting in a higher resistance to rutting for the PG76-22TR mix compared to the PG76-
22NV mix. However, a better laboratory rutting resistance in the RLT will not necessarily
translate to a better rutting performance in the field as the rutting performance of an asphalt
pavement is highly dependent on pavement structure, the HMA stiffness, and the rutting
characteristics of the HMA mixture. In a mechanistic pavement analysis, an HMA layer with
a higher stiffness will produce a lower vertical resilient strain in the HMA layer and hence
a better rutting life. Therefore, a mechanistic analysis needs to be conducted to correctly
evaluate rutting performance of the various mixtures.
3.3.2. Rutting behavior of mixtures in the asphalt pavement analyzer
The asphalt pavement analyzer (APA) test was used to empirically evaluate the resistance
of the mixtures to rutting. The APA test is standardized under AASHTO TP63-03, where
a loaded concave wheel travels along a pressurized rubber hose that rests upon the HMA
sample. Four 150 mm diameter cylindrical samples were compacted from each mix using the
SGC to a height of 76 mm and air voids of 7±1%. Samples are secured within form-fitting
acrylic blocks during testing. The APA wheel load is 45 kg and the hose pressure is 690 kPa.
The samples were conditioned for six hours before being tested in the dry condition at 60℃ to
8,000 cycles. A data acquisition program records rut depths at 2 points within each sample and
their average is reported.
Table 3 summarizes the rut depth data from the APA test at 60℃. The coefficients of
variations represent the variability of the data among the four replicates for each mix. A
maximum of 8 mm rut depth after 8,000 cycles at 60℃ has been used by NDOT as a general
failure criterion. The APA data in Table 3 indicate that all mixtures meet and exceed the
NDOT APA criterion. No significant difference was observed between the TR and the NV
mixtures at 60℃ indicating similar resistance to rutting. These results could be an indication
that the APA test is not sensitive enough to test for any significant difference in the rutting
resistance of the mixtures when tested at a temperature lower than the high performance
temperature of the PG binder grades specifically as is the case for the PG76-22NV and PG76-
22TR mixtures.
Following the above observations, the PG64-28NV mix and PG64-28TR mix were tested
under the APA at 64℃ and the PG76-22NV mix and PG76-22TR mix were tested under the
APA at 76℃. Table 3 shows that there is no significant difference in the rut depths between
the PG64-28NV mix and the PG64-28TR mix at 64℃, whereas the PG76-22TR mix exhibited
a slightly higher rut depth than the PG76-22NV mix at 76℃.
It should be noted that the contradiction between the APA and RLT results is due to the
fact that the APA is an empirical tests while the RLT is a mechanistic test. Therefore, the final
impact of the RLT results on the rutting performance of the mixtures must be coupled with the
dynamic modulus properties through an extensive mechanistic analysis. While the results of
the APA test are directly and empirically correlated to the rutting performance of the mixtures.
Due to length limitations, this information is not included in this paper and has been published
in other reports (Sebaaly, 2006).
Table 3. APA Rut Depths for All Mixtures
Temperature, ℃ Mix Mean Rut Depth (mm) CVfour samples rut depth, %
PG64-28NV 1.7 4.6
PG64-28TR 1.2 9.5
60
PG76-22NV 1.5 8.6
PG76-22TR 1.2 5.9
PG64-28NV 2.1 4.0
64
PG64-28TR 2.0 6.7
PG76-22NV 2.7 7.2
76
PG76-22TR 3.4 3.5
4. Fatigue characteristics of mixtures
The resistance of the HMA mixtures to fatigue cracking was evaluated using the flexural
beam fatigue test in accordance with the AASHTO T321-03. Fatigue life or failure was
defined as the number of cycles corresponding to a 50% reduction in the initial stiffness. The
fatigue behaviors of the four HMA mixtures were characterized by the following relationship.
[6]
where Nf is the fatigue life (number of load repetitions to fatigue damage), εt is the applied
tensile strain, and k1 and k2 are experimentally determined coefficients. The beam fatigue
samples were compacted using the kneading compactor. The compacted beams were 405 mm
long by 76 mm thick and 76 mm. Since fatigue cracking is more critical during the in-service
life of the pavement, the compacted beams were subjected to long-term oven aging following
the Superpave procedure of 120±3 hours at a temperature of 85℃, then cut using a diamond
blade saw to 381 mm long by 51 mm thick by 64 mm wide. The tested specimen air voids
were 7±1%. It is clearly shown from the beam fatigue relationship that the higher the k1 and
k2 values, the more fatigue resistant the mix will be. Table 4 shows the beam fatigue results of
each mix at 22.2 ℃, and the fatigue regression coefficients for all mixtures. Figure 5 shows
the fatigue characteristics of the various mixtures.
Table 4. Summary of the Fatigue Test Results and Regression Coefficients for All Mixtures
Binder Content, Average Air Initial Flexural

Mixture k1+ k2+ R2, %
% DWA* Voids, % Stiffness, MPa
PG64-28NV 4.8 6.5 1,848 1.047E-04 2.769 98.6
PG64-28TR 5.0 7.0 1,862 3.171E-17 6.541 98.7
PG76-22NV 3.8 6.8 1,924 5.656E-08 3.722 98.8
PG76-22TR 3.9 7.0 3,254 6.455E-05 3.262 97.6
*
Denotes “Dry Weight of Aggregates.”
+
k1 and k2 are experimentally determined coefficients using the following relationship:
R2 = 98.6% R2 = 98.8%
R2 = 98.7% R2 = 97.6%
Figure 5. Relationships between strain levels and number of cycles to failure of the various
mixtures at 22.2 ℃
Table 5 summarizes the fatigue data of the various mixtures in terms of the number
of cycles to failures under tensile strains of 300, 500, and 800 microns, representing low,
medium, and high levels of tensile strains at the bottom of the HMA layer of a flexible
pavement, respectively. The effect of tire rubber-modified binder on the fatigue behavior
was evaluated in terms of the ratio of fatigue life of the tire rubber-modified mixtures over
the corresponding polymer-modified mixture. A fatigue life ratio greater than one indicates a
better life for the tire rubber-modified mixtures at the same strain level. Table 5 shows a better
laboratory fatigue behavior for the polymer-modified mixtures compared to the tire rubber-
modified mixtures at all three strain levels.
Table 5. Laboratory Fatigue Life for All Mixtures at 22.2℃
Air Voids,% Number of cycles to failure, Nf

Mix average Strain Level
CV- (%)
(%) 300 microns 500 microns 800 microns
PG64-28NV 6.5 5.9 NA 144,925 39,435
PG64-28TR 7.0 1.6 NA 124,220 5,740
PG76-22NV 6.8 10 733,435 109,550 19,050
PG76-22TR 7.0 5.6 200,040 37,800 8,160
Fatigue life ratio of PG64-28TR mix to
NA 0.86 0.15
PG64-28NV mix
Fatigue life ratio of PG76-22TR mix to
0.28 0.35 0.43
PG76-22NV mix
PG64-28TR mix vs PG64-28NV mix NA+ NS# L*
PG76-22TR mix vs PG76-22NV mix L* L* L*
-
Total of 6 specimens
+
Results not reported because the test was not conducted at this strain level.
#
No significantly difference between the NV and TR mixtures.
*
Nf of the TR mixtures is significantly lower than the Nf of the NV mixtures.
Additionally, Table 5 summarizes the results of the statistical analyses for the mixtures
comparison at all three strain levels. The PG64-28TR mix has significantly less fatigue
resistance than the PG64-28NV mix at 800 microns but there is no significantly difference
between these two mixtures at 500 microns. The two mixtures are not compared at the 300
microns since they were not actually tested at this strain level. Additionally, the PG76-22TR
mix has significantly less fatigue resistance than the PG76-22NV mix at 300, 500, and 800
microns.
However, a better laboratory fatigue resistance will not necessarily translate to a better
fatigue performance in the field as the fatigue life of an asphalt pavement is highly dependent
on both the modulus and the fatigue characteristics of the HMA mixture and their interaction.
In a mechanistic pavement analysis, an HMA layer with a higher stiffness will show a lower
laboratory fatigue life but on the other hand it will produce a lower tensile strain under field
loading. Therefore, depending on the magnitude of strain reduction, the HMA layer with the
higher stiffness may result in a longer fatigue life in the field or vice versa. Therefore, a full
mechanistic analysis is needed to effectively evaluate the impact of tire rubber on the fatigue
performance of an HMA pavement.
5. Thermal cracking resistance in the thermal stress restrained specimen
The thermal stress restrained specimen test (TSRST) was used to determine the low
temperature cracking resistance of the various mixtures. The test cools down a 51 by 51 by
254 mm beam specimen at a rate of 10 ℃/hour while restraining it from contracting. While
the beam is being cooled down, tensile stresses are generated due to the ends being restrained.
The HMA mixture fractures as the internally generated stresses exceed its tensile strength. The
temperature at which fracture occurs is referred to as “fracture temperature” and represents
the field temperature under which the pavement will experience thermal cracking. The TSRST
samples were compacted using the kneading compactor to air voids of 7±1%. Since thermal
cracking is more critical during the in-service life of the pavement, the compacted beams were
subjected to long-term oven aging for 120 ± 3 hours at a temperature of 85 ℃.
It should be note that thermal cracking distress is only a concern in the northern part of the
state of Nevada. Hence, only the northern mixtures (the PG64-28NV mix and the PG64-28TR
mix) were tested in the laboratory for thermal cracking resistance.
Table 6 summarizes the fracture temperatures for the PG64-28NV mix and the PG64-
28TR mix. The coefficients of variations represent the variability of the data among the three
replicates for each test. The fracture temperatures of both mixtures were colder than the
low performance temperature of the binders of -28 ℃. Both mixtures had similar fracture
temperatures within 2 ℃ which resulted in a non significant difference at the 5% significance
level.
Table 6. Thermal Cracking Fracture Temperatures of the PG64-28NV and the PG64-28TR
mixtures
Mixture Air Voids, % CV, % Fracture Temperature ℃

PG64-28NV 7.0 3.1 -33
PG64-28TR 7.0 3.9 -31
6. Conclusion and recommendations
The objective of this research effort was to compare the performance of Nevada’s HMA
mixtures made with terminal blend (TR) rubber modified asphalt binders with Nevada’s
HMA mixtures made with polymer modified binders (Sebaaly et al., 2007). This objective
was achieved through a laboratory-based experiment that evaluated the resistance of the two
types of mixtures to rutting, fatigue, thermal cracking, and moisture damage. In addition,
the dynamic moduli of the mixtures were evaluated. Based on the data generated from the
laboratory experiment, the following conclusions can be made.
- The moisture sensitivity of the terminal blend (TR) and the polymer modified (NV)
mixtures are well above the minimum required tensile strength ratio of 70%. This
indicates that both mixtures exhibit good resistance to moisture damage.
- When the laboratory rutting resistances of the mixtures were compared in the RLT test,
the PG64-28TR mix showed lower rutting resistance than the PG64-28NV mix. On the
other hand, the PG76-22TR mix showed higher rutting resistance than the PG76-22NV
mix. Therefore, it may be advantageous to use the PG76-22TR mix in southern Nevada
where rutting is a more prevalent failure mode.
- All mixtures met the NDOT APA criterion of 8 mm at 60℃ after 8,000 loading cycles.
- When the laboratory fatigue resistances of the mixtures were compared, the PG64-
28TR mix showed lower fatigue resistance than the PG64-28NV mix at strain levels
higher than 500 microns, but the same at 500 microns, and the PG64-28TR mix
showed higher fatigue resistance than the PG64-28NV mix at strain levels lower than
500 microns. On the other hand, the laboratory fatigue resistance of the PG76-22TR
mix is lower than the laboratory fatigue resistance of the PG76-22NV mix at all strain
levels.
- The dynamic modulus of the PG64-28TR mix is approximately equal to the dynamic
modulus of the PG64-28NV mix at 21.1℃, and the dynamic modulus of the PG76-
22TR mix is higher than the dynamic modulus of the PG76-22NV mix at 21.1℃.
- The PG64-28TR mix and the PG64-28NV mix showed good resistance to thermal
cracking since the fracture temperatures of both mixtures were colder than the low
performance temperature of the binders (-28℃). Both mixtures had similar fracture
temperatures (within 2℃) which resulted in a non-significant difference.
- It is recommended that a full mechanistic analysis be performed to evaluate the

combined impacts of mixtures properties and their resistance to the various failure
modes along with pavement structure on the responses of HMA pavements to traffic
loads.
7. Bibliography
Amirkhanian S., Corley M., « Utilization of Rubberized Asphalt in the United States-an
Overview », International Symposium of Advanced Technologies in Asphalt Pavements,
South Korea. 2004, pp. 3-13.
Putman B.J., Amirkhanian S., « Utilization of Waste Fibers in stone Matrix Asphalt mixtures »,
Journal of Resources Conservation and Recycling, Vol. 42, No. 3, 2004, pp. 265-274.
Sebaaly, H., « Evaluation of Nevada’s Hot Mix Asphalt Mixtures Using Tire Rubber Modified
Binder », Master Thesis, Department of Civil Engineering, University of Nevada, Reno,
2006.
Sebaaly, P., Hand A., Epps J., « Laboratory Evaluation of Rubber-Modified Hot Mix Asphalt
Mixtures », Department of Civil Engineering, University of Nevada, Reno, 1998.
Sebaaly, P., Sebaaly, H., Hajj, E. « Evaluation of Nevada’s Hot Mix Asphalt Mixtures
Manufactured with Terminal Blend Rubber Modified Binders », Research Report, Nevada
Department of Transportation, Carson City, Nevada, 2007.
Pavement Monitoring Results After Seven

Years of Using Crumb Rubber Modified
Asphalt in Brazil
M.Sc. Rodrigo M. Vasconcellos Barros
COPAVEL Consultoria de Engenharia Ltda

Rua Catulo da Paixão Cearense 249 – São Paulo – Brazil
rodrigo@copavel.com.br
ABSTRACT: After the year 2000, when the first AR conference in Vilamoura Portugal took
place, the first trial sections began to be built using crumb rubber modified binders in major
highways in Brazil. Since then, the use of the technology has increased very much and there
are already over 2.000 km of roads using some kind of crumb rubber modified asphalt in
Brazil. Almost all of the Brazilian applications use Terminal Blend binders.
This paper is the result of the continuous monitoring of these pavements since then. Some of
these applications were done using around 15% of crumb rubber in dense graded mixes and
some of them using at least 20% in gap graded mixes. Our focus in this paper is on these gap
graded applications which have been used on some important Brazilian highways now.
The paper is pointing out the problems and difficulties presented during the asphalt mix
production and the results achieved, in terms of pavement performance. The paper shows the
properties of the binders used, such as viscosity, elastic recovery, penetration, flow etc. It also
presents the mix characteristics and all of the pavement performance related data ever since
its application, such as Traffic, FWD deflections, skid resistance and ride quality.
The majority of the applications were done within the private sector in toll concession
companies. As they are obliged by contract to survey the pavement every year, the paper
shows these monitoring surveys qualifying the pavement performance since its application.
The paper also presents some comparisons between the performance of the sections
rehabilitated with crumb rubber modified binder using thickness reduction factors to the
performance assessed in sections rehabilitated with conventional binders.
With the arrival of the asphalt rubber “in situ” mixer to Brazil, it will be possible to produce
higher viscosity binders and have even better pavement performances.
KEYWORDS: Pavement Performance, Crumb Rubber Modified Applications, Brazil

1. Introduction
In 2001 the first experiences with asphalt modification with crumb rubber took place
in Brazil, mainly in the center-southern States. At that time all the applications in the field
were made by the private sector in toll road networks. The asphalt used was, in most cases,
processed in terminal blends using at least 15% of crumb rubber. Still today 90% of the
applications are done using dense graded mixes. In some cases, where gap graded mixes were
applied, the amount of crumb rubber in the asphalt reaches up to 20%.
This paper will focus on the work performed by two big Concession Companies in the
state of São Paulo who believed in the technology and applied it on most of their network
pavement extension. They are SPVias – Rodovias Integradas do Oeste S.A. and Ecovias.
Spvias is a toll road company located the southern west region of the state of São Paulo
and has approximately 516 km of highway with almost 1.500 km in traffic lane in its network.
Their roads have a traffic varying from 5.000 to 25.000 vehicles a day in single lanes and
double lanes per direction.
Ecovias belongs to the toll Company Group Ecorodovias, which operates three more
concession companies in Brazil. It is one of the most important highway management
company in the State of São Paulo. The major roads, Via Anchieta and Imigrantes, link the
city of São Paulo to the coast, where are the most beautiful beaches of São Paulo state. These
roads also connect a large portion of the southern road network to the biggest port in South
America, the port of Santos. They hold over 30.000.000 vehicles per year.
Copavel is both responsible for the pavement design and for the yearly pavement
monitoring campaigns. This has enabled us to present this paper on behalf of the collected
data and assessed results which confirmed the performances of the roads rehabilitated using
crumb rubber binders.
2. Historical Data
The first application happened in 2001 after the first AR 2000 in Villamoura – Portugal. It
began with trial sections in the south of Brazil in two Concession Companies: Univias in Rio
Grande do Sul state and Rodonorte in Paraná State. In 2002 many trial sections were built
on many other Brazilian States and it was after 2003 that the technology was applied in larger
scale for road rehabilitation throughout the country.
In the beginning of 2002 the first gap graded mix application took place in São Paulo in
Via Anhanguera, operated by Intervias. This road links São Paulo to the northern states of
Brazil. With this gap graded mix a modified binder with 20% of crumb rubber was used. After
2003 the technology left the trial sections applications to reach large scale road rehabilitation
projects and, since then, 15.000 tons of crumb rubber modified binder have been applied on
almost 4.000 km of roads in Brazil.
The first large scale gap graded mix application was done in 2003 on Intervias
concessionaire’s roads, with almost 100 km of roads rehabilitated using a 20% crumb rubber
asphalt mix. Although they have achieved very good results and best pavement performance
ever, the Concession was sold to a different group which has changed the maintenance plan
stopping the use of crumb rubber binders.
In 2003 Spvias started the use of crumb rubber binder and has not stopped since then.
Over 23.000 tons of crumb rubber binders have been applied on almost 600 km of single lane
roads and they have plans to overlay their whole network with it. Nowadays this is the biggest
private user of crumb rubber binders in Brazil.
Ecovias started their trial sections in 2004 and today are the biggest user of gap graded
mix in Brazil. The major application have taken place after 2006 on the main roads of their
network, Via Anchieta and Rodovia dos Imigrantes. Both roads link São Paulo city to the
coastal city of Santos, where the port is. The table below presents an estimate of crumb rubber
binder consumed since the year of 2001 and also an estimate of the number of tires removed
from the environment.
Table 1. Information provided by Greca Asfaltos do Brazil crumb rubber terminal blend
producer
Quantity of Crumb Approximately Tires removed from

Year
Rubber Asphalt (ton) Road Lengh in km environment
2001 152.42 3.92 3,918.25
2002 732.25 18.82 18,823.78
2003 15,304.68 393.44 393,436.50
2004 8,024.70 206.29 206,290.49
2005 16,550.02 425.45 425,450.39
2006 28,214.09 725.30 725,297.94
2007 17,657.15 453.91 453,911.31
2008 (Jan. to Nov.) 21,246.67 546.19 546,186.89
TOTAL 107,881.98 2,773.32 2,773,315.55
Asphalt Pavement mat: 4 cm thick, 7m wide, Asphalt content 5,9%
Considerations:
(density 2,7 ton/m3)
3. Problems and Difficulties During Crumb Rubber Mixes Applications
Some of the problems encountered on the crumb rubber applications help explain the
direction that the technology has followed in Brazil in these years.
First of all and the most important, there is the fact that the Brazilian road building
industry was not well prepared to deal with such a high viscosity binder. With the addition
of 20% of crumb rubber to the binder its viscosity reached 2.000 cP. at 175 °C or more. The
asphalt plants found it difficult to pump the binder continuously and homogeneously. The
asphalt mix production was always stopping due to blocking in the binder pumping system.
This fact led the crumb rubber asphalt producers to lowering the viscosity of the binder in
order to facilitate the production of the asphalt mix. Even though the plants were later adapted
with more powerful pumps and stronger tubes, the lower viscosity binders remained as the
preferred option to reduce the production problems in the field. For that reason the use of 15%
of crumb rubber in the binder modification became almost a standard in Brazil.
The 20% crumb rubber content in binders is used only in the cases when the wearing
courses has to be built with gap graded mixes to improve macro-texture and friction
coefficients. In such cases the problem turns to be the available aggregates. In São Paulo State,
where environmental issues are of major concern nowadays, the opening of new quarries was
not always possible and the existing ones usually sell aggregates for very different purposes
such as Portland Cement Concrete, for instance.
A specific gradation and a good shape aggregate was not always that easy to find and
usually was more expensive than the regularly available aggregate. There are some regions
where, even though the rock has good mineralogical nature, the produced aggregate is not
very good. In the interior of the State of São Paulo there are good basalt rock quarries but the
shape of the produced aggregates is not so good. They are usually flat and not adequate for the
use in discontinuous grading asphalt mixes. That is also why dense graded mixes are the most
popular in our Country.
Due to such particularities the crumb rubber asphalt mixes have had lower asphalt
content when compared to traditional asphalt rubber mixes. These facts changed the
original concept of the Asphalt Rubber in Brazil, which is based on very high viscosity
binders with high asphalt content in the mixes. With the arrival of the in situ crumb
rubber mixes very high viscosity binders will be produced in even larger scales. This will
force the industry to re-adapt themselves to work with binders with viscosities around
4.000 CP. to 5.000 CP at 175°C.
4. Pavement Performance In Spvias Network
After more than 600 km of equivalent single lane roads have been rehabilitated, this paper
will focus on the results obtained from two particular applications; the first and older one is
on SP-255 highway, where a dense graded mix was applied over a cracked cement treated
base. The second application was on SP-270 highway where a gap graded mix was applied
following the California Department of Transportation (Caltrans) gradation.
4.1 Asphalt Crumb Rubber application in SP-255 Highway.
In 2002 the pavement design was conceived and it defined that the extension from Km
279,3 to Km 288,2 should have an 8 cm thick overlay for the 2,63 x 10 6 Aashto ESAL’s after
milling out the existing asphalt surface layer.
In March 2003 the section was rehabilitated using a Dense graded mix with 15% of crumb
rubber binder. The mix characteristics are shown below.
SPVIAS - SP-255 Highway - Crumb Rubber Asphalt Mix

Asphalt Institute “ Mix 4B “
Sieves Crushed
Sieves Min Max MIX Sand Filler 3/4” 3/8” P%Talbot
mm Dust
3/4
4B 4B 15 0 18 25 42
# 200 0.075 4 10 8.4 0.4 97 0 1 14.7 8.4
# 80 0.18 10 18 12.6 1.5 99.1 0 1 22.8 12.4
# 40 0.42 15 25 22.1 17.7 100 0 1.4 32.3 18.1
# 10 2.4 33 48 43.7 89.9 100 0 1.7 61.5 39.6
#4 4.8 50 70 61.2 99.4 100 0 4.8 93.6 53.9
1/2” 12.7 80 100 97.9 100 100 36.7 100 100 83.3
3/4” 19.1 100 100 100.0 100 100 100 100 100 100.0
1” 25.4 100 100 100.0 100 100 100 100 100 100.0
1 1/2” 38.1 100 100 100.0 100 100 100 100 100 100.0
2” 50.8 100 100 100.0 100 100 100 100 100 100.0
100.0
SPVIAS - SP-255 - Crumb Rubber Dense Mix
MIX 4B - Asphalt Institute
90.0
80.0
70.0
60.0
% Passing
50.0
40.0
30.0
20.0
10.0
0.0
0.01 0.1 Sieves 1 Mix 10
Talbot Mix 4B
Figure 1. Dense Graded Mix used in SPVias SP-255 Overlay Project in 2003
Table 2. Dense Graded Marshall Characteristics
Marshall Characteristics of Crumb Rubber Modified Binder

Dense Graded Mix used in SP-255 Rehabilitation Project
Description Unit Values
Density ton/m 3
2,495
Asphalt Content % 5.90%
Stability Kgf 1050 kgf/cm2
Flow mm 14.8
% Air Voids % 3.80%
% Asphalt Voids Ratio - 78
Compaction Temperature °C °C 160 °C to 165°C
Crumb Rubber Content in Asphalt % 15%
Table 3. Crumb Rubber Binder Characteristics
Crumb Rubber Modified Binder Characteristics in

DENSE GRADED MIX SP-255 Rehabilitation Project
Rotational Viscosity - Haake Viscometer - Sabita (BR 5 T) Pa.s - 175°C 1.20
Ball Penetration and Resilience Sabita (BR2 T) (0,1) mm 19.0
% - 5 min 91
Compression Recovery - Sabita (BR3 T) % - 1 hour 88
% - 4 days 54
Flow Test - Sabita (BR4 T) mm 58.3
Softening Point - (Astm - D 36) °C 54
Below are some pictures of the Rehabilitation Project.
Figure 2. Pavement surface after milling Figure 3. Asphalt mix being placed
Figure 4. Severe cracking in the CTB Figure 5. Good results after 5 years
4.1.1 Pavement Monitoring Results of SP-255 Project
The pavement has been frequently evaluated since its rehabilitation in 2003. Even though
block cracking of the cement stabilized base was very intense, it has not migrated to the
surface so far. The pavement structural capacity and ride quality have improved a lot after the
rehabilitation, as can be seen in the graphs below.
SPVIAS - SP-255 CRUMB RUBBER MODIFIED BINDER PROJECT

FWD Maximum Deflection
120
100
FWD DEFLECTION mm/100
80
60
40
20
0
278,00 279,00 280,00 281,00 282,00 283,00 284,00 285,00 286,00 287,00 288,00 289,00
2002 2006 2008
km
Figure 6. FWD Deflections

SPVIAS - SP-255 CRUMB RUBBER MODIFIED BINDER PROJECT

RIDE QUALITY - IRI - INTERNATIONAL ROUGHNESS INDEX
7,00
6,00
5,00
4,00
IRI
3,00
2,00
1,00
0,00
278,500 279,500 280,500 281,500 282,500 283,500 284,500 285,500 286,500 287,500 288,500
km 2002 2003 2005 2006 2007 2008
Figure 7. IRI Measurements
4.2. Asphalt Crumb Rubber application in SP-270 Highway.
It was possible to find good aggregates for this project and the mix was designed as a gap
graded one according to the California Department of Transportation - Caltrans guidelines.
The pavement design was done in 2004 for an estimated traffic of for 8,86 x 10 6
AASHTO ESAL’s, an 8 year service life. The pavement condition was bad with alligator
cracks and permanent deformation in localized areas. These areas were repaired with a mill
and fill technique reaching a depth of 50 mm using conventional dense graded mix. After that,
the overlay was applied with a varying thickness from 30 mm to 40 mm in the critical sectors.
The mix was applied in the end of 2004 and its properties are presented in Figure 8 below.
SPVIAS - SP-270 Highway - Crumb Rubber Modified Binder Mix

CALTRANS GRADING
Sieves Sieves Min Max MIX Filler 5/8" 3/8" Crushed Dust
mm CALTRANS C AL T R A N S 2 15 61 22
# 200 0,075 2 7 5,1 97 0,7 1,4 9,8
# 100 0,149 5 10 6,3 99,1 0,8 1,6 14,8
# 50 0,297 7 15 9,1 100 0,9 1,9 26,4
# 30 0,595 10 20 12,2 100 0,9 2,1 39,8
#8 2,38 15 25 21,8 100 1,1 3 80,9
#4 4,76 28 42 34,6 100 1,6 17,9 97,4
3/8" 9,51 78 92 90,0 100 35,5 99,5 100
1/2" 12,7 90 100 94,3 100 61,8 100 100
5/8" 15,9 100 100 100,0 100 100 100 100
3/4" 19 100 100 100,0 100 100 100 100
SPVIAS - SP-270 - Crumb Rubber Modified Binder

CALTRANS - Gap Graded
100,0
90,0
80,0
70,0
60,0
% Passing
50,0
40,0
30,0
20,0
10,0
0,0
0,01 0,1 1 10
Sieves Mix CALTRANS
Figure 8. Caltrans Grading used in SP-270 Rehabilitation Project
Table 4. Gap Graded Marshall Characteristics
Marshall Characteristics of Crumb Rubber Modified Binder Mix used in

GAP GRADED SP-270 Rehabilitation Project
Density ton/m3 2.293
Stability Kgf 780
Flow mm 13.6
% Air Voids % 5.0%
Asphalt Voids Ratio - 75

Table 5. Crumb Rubber Binder Characteristics on Gap Graded Mix
Crumb Rubber Modified Binder Characteristics in GAP GRADED

SP-270 Rehabilitation Project
Rotational Viscosity - Haake Viscometer - (Sabita BR 5 T) Pa.s - 175°C 2.56
Ball Penetration and Resilience (Sabita BR2 T) (0,1) mm 27.0
% - 5 min 93
Compression Recovery - (Sabita BR3 T) % - 1 hour 84
% - 4 days 48
Flow Test - (Sabita BR4 T) mm 30
Figure 9. SP-270 Highway with Gap Figure 10. Macro-texture detail of the
Graded Mix in Place mix applied
4.2.1 Pavement Monitoring Results of SP-270 Project
The pavement has been evaluated since its rehabilitation in 2004. After the rehabilitation,
the structural capacity, the riding quality and the skid resistance have all improved on the road.
There are some friction measurements using the Grip Tester which are presented in the graphs
below.
SPVIAS - SP-270 CRUMB RUBBER BINDER PROJECT

GAP GRADED - FWD Maximum Deflection
120
100
FWD DEFLECTION mm/100
80
60
40
20
0
117 118 119 120 121 122 123 124 125 126 127 128 129 130
2003 2004 2006
km
Figure 11. FWD Maximum Deflection
SPVIAS - SP-270 CRUMB RUBBER BINDER PROJECT

RIDE QUALITY - IRI - INTERNATIONAL ROUGHNESS INDEX
5,00
4,50
4,00
3,50
3,00
2,50
IRI
2,00
1,50
1,00
0,50
0,00
115,000 117,000 119,000 121,000 123,000 125,000 127,000 129,000
km 2002 2003 2005 2006 2007
Figure 12. IRI – International Roughness Index
There was a good structural gain in the first four kilometers of the section but around the
km 122 the deflections remained high due to drainage localized problem. This was the reason
why the pavement presented some premature failure in 2006, over this precise extension.
After solving the drainage problems the pavement was patched and it has been performing
well since then.
It is important to point out that the first three IRI surveys from 2002 to 2005 were done
with a Roughness Response Type Meter and the new ones after 2006 were done using a Laser
Profiler which, in this case, presented lower IRI values.
It was also possible to verify the levels of skid resistance provided by a gap graded mix;
the improvement is remarkable if compared with a dense graded mix. The latest results are
shown below.
SP-270 - CRUMB RUBBER BINDER GAP GRADED MIX PROJECT

SKID RESISTANCE - FRICTION SURVEY WITH GRIP TESTER
0,80
0,75
0,70
0,65
Grip Number - GN
0,60
0,55
0,50
0,45
0,40
0,35
0,30
0,25
km West lane East Lane "Minimum Required"
Figure 13. Friction Testing with Grip Tester device in 2005 after 6 months.
5. Pavement Performance In Ecovias Network
The other big project that is taking place in São Paulo is in Ecovias concessionnaire
network. The major roads of their network, Via Anchieta and Rodovia dos Imigrantes are
the link between the city of São Paulo which has an altitude of 800 m above sea level, and
the coast and both of them are being rehabilitated using crumb rubber modified binder mixes.
After doing the first trial section in 2004, they started implementing crumb rubber modified
asphalt in all new flexible pavement rehabilitation projects in their network.
5.1. Crumb Rubber Modified Binder Projects in Ecovias.
In Anchieta highway, where the first big crumb rubber asphalt mix application took place,
the road geometry is very sinuous and the traffic is one of the heaviest in São Paulo State. In
2005 the pavement design, considering a period of 4 years determined that the section from
the km 40 to 55 should have an 3 cm overlay for 2,65 x 10 7 Aashto ESAL’s, after some
patching in the existing asphalt surface layer.
The pavement structure is composed of a rigid pavement built in the sixties overlayed by
multiple dense graded asphalt mixes. Due to the stiffness of the pavement, the deflections are
usually very low and the major concerns were the crack reflection from the slab joints and
also a design a very rigid mix designed to be rut resistant and cope with all the heavy traffic.
The pavement deflections were not evaluated in this case. Table 14 below presents the mix
Marshall Characteristics.
Table 14. Gap Graded Marshall Characteristics of Ecovias’s projects
Marshall Characteristics of Crumb Modified Rubber Gap Graded

Mix used in SP-150 Rehabilitation Project - Via Anhanguera
Description Unit Values
Density ton/m 3
2,326
Stability Kgf 1065 kgf/cm2
Flow mm 10
% Air Voids % 5.0%
% Asphalt Voids Ratio - 72.5
During the mix production they had difficulties pumping the high viscosity binder in the
asphalt plant and decided to limit the viscosity to 1,8 Pa.s – 175 °C. The crumb rubber binder
was then modified to meet the contractor specifications.
In this case the amount of crumb rubber in the asphalt rubber binder was reduced to 18%.
Normally we use a minimum of 20% of crumb rubber in Brazil when applying in a gap graded
mix. The following table presents the crumb rubber binder characteristics.
Table 15. Crumb rubber binder Average Characteristics used in Ecovias’s Mixes
Crumb Rubber Binder Characteristics in GAP GRADED

Via Anhanguera - SP-150 Rehabilitation Project
Rotational Viscosity - Haake Viscometer - (Sabita BR 5 T) Pa.s - 175°C 1.80
Ball Penetration and Resilience (Sabita BR2 T) (0,1) mm 24.0
% - 5 min 93
Compression Recovery - (Sabita BR3 T) % - 1 hour 85
% - 4 days 51
Flow Test - (Sabita BR4 T) mm 60
The Anchieta highway section is located in a very humid forest area, and the mix grading
was adjusted to reduce the permeability of the surface layer. Due to the very heavy traffic, the
officials from concessionaire did not want a high content of asphalt in the mix to reduce the
risk of bleeding and rutting.
During the mix design, the mixes were tested in The São Paulo State University in an
accelerated rutting device to evaluate the rutting versus asphalt content. The results are
presented below.
POLITECHNIQUE SCHOOL OF SÃO PAULO STATE UNIVERSITY

TRANSPORT ENGINEERING DEPARTMENT
PAVEMENT TECHNOLOGY LABORATORY
CALTRANS GRADING - ECOVIAS DOS IMIGRANTES PROJECT
Permanent Deformation x Number of Cycles - GRECA ASFALTOS
100
% GROUP
R
u 10 5,25
3,85 4,53
t 2,82 3,32
2,43 3,57
2,43 2,97
1,66 2,03 3,27
1,38 2,58
D 1 1,99
1,57
e 1,21
0,96
p
t
h 0,1
100 1000 Number of Cycles 10000 100000
Ecoflex A - 6,5% Ecoflex B - 6,3% Ecoflex B - 6%
Figure 14 – Rutting accelerated testing Results
The best results were achieved for 6% of binder content with Ecoflex B (18% of crumb
rubber) which was selected as the optimum value for the mix design. This asphalt content
generated 5% air voids in the mix. In 2006 the section was rehabilitated using a Caltrans Gap
grading mix with characteristics as shown below.
ECOVIAS - SP-150 and SP-160 Highways

Via Anchieta and Rodovia dos Imigrantes
Crumb Rubber Asphalt Mix -Caltrans Gap Graded
Crushed
Sieves Sieves Min Max MIX Sand Filler 5/8" 3/8"
Dust
mm
CALTRANSCALTRANS 0 2 15 61 22
# 200 0,075 2 7 5,5 0,4 97 0,7 1,4 9,8
# 100 0,149 5 10 8,4 1,5 99,1 0,8 1,6 14,8
# 50 0,297 7 15 11,2 17,7 100 0,9 1,9 26,4
# 30 0,595 10 20 14,3 89,9 100 0,9 2,1 39,8
#8 2,38 15 25 20,5 99,4 100 1,1 3 80,9
#4 4,76 28 42 35,8 100 100 1,6 17,9 97,4
3/8" 9,51 78 92 86,3 100 100 35,5 99,5 100
1/2" 12,7 90 100 98,1 100 100 61,8 100 100
5/8" 15,9 100 100 100,0 100 100 100 100 100
3/4" 19 100 100 100,0 100 100 100 100 100
Via Anchieta SP-150 Crumb Rubber Binder Aplication

Caltrans - Gap Graded Mix
100,0
90,0
80,0
70,0
60,0
% Passing
50,0
40,0
30,0
20,0
10,0
0,0
0,01 0,1 1 10
Sieves Mix CALTRANS
Figure 15. Caltrans Grading used in SP-150- Via Anchieta Rehabilitation Project
After this first application in 2006 there has been a lot more applications in Ecovias
network in their main highways, other sections of Anchieta and Imigrantes. We are presenting
some pictures of the sections that have been rehabilitated since the year of 2006.
Figures. 16 -17 – SP-150- Via Anchieta

Gap Graded 2006 Rehabilitation Project
Figures 18 -19. SP-160- Imigrantes Gap Graded 2007 Rehabilitation Project
Figures 20-21. SP-150- Via Anchieta Gap Graded 2008 Rehabilitation Project
6. Conclusions
After these seven years of experience we can conclude that crumb rubber modified binders
represent a great contribution to the pavement rehabilitation technology. Even though we
cannot obtain the same traditional “asphalt rubber” properties using the terminal blend
binders, all the applications have presented outstanding performances so far, when compared
with conventional asphalt mix applications.
From the beginning, the crumb rubber binder use was restricted to the private sector in
toll concession companies, but now they are reaching several Highway State Departments as
well as the Federal Highway Department in Brazil. In this period Brazilian specifications were
developed to control the crumb rubber binder characteristics and its applications.
The so called “new” technology barrier has been broken and the road community is more
comfortable using these binders. Pavement designers are recommending its use more often,
and the solutions are being easily accepted by the decision makers based on good results
presented.
A lot of research has been done since the first application in 2001 with good results using
both 15% or 20% crumb rubber modified binders. The results have proven that the use of
crumb rubber in asphalt mixes is a good pavement rehabilitation alternative to make the
pavements last longer and reduce the maintenance costs. The examples presented in this paper
are from Spvias and Ecovias, road concession companies who decided to have their entire
flexible pavement network rehabilitated with crumb rubber binder mixes. It shows that the
benefit cost ratio is very positive for them no matter the amount of traffic they have on their
roads.
The crumb rubber modified binder is a powerful tool to reduce reflective cracking but, due
to the Brazilian tropical weather with a lot of rain, prior milling is imperative to guarantee
a good pavement performance. Recent research (Ceratti, 2006) has indicated that reflective
cracking on 15% crumb rubber modified dense graded mixes overlay can be 5 to 6 times
slower than the reflective cracking when using conventional dense mixes.
From now on, after the terminal blend has been accepted and the road industry has adapted
itself to deal with high viscosity binders, it is a good time to start doing the crumb rubber
mixing on the job site to have the traditional asphalt rubber, which will give us even better
results. The in situ mixers have just arrived in Brazil and we hope to present in the near
future, new results of “asphalt rubber” applications.
7. Acknowledgements
The author would like to thank Mr. Helvecio Soares and Mr. Paulo Pimenta, Spvias
President and Engineering Director for supporting the crumb rubber binder applications and
the preparation of the paper. Thanks to Mr. Humberto Gomes, Ecovias President, for allowing
the presentation of their results, and also thanks to Marcos Greca who supported and supplied
important information to the preparation of the paper. A special thanks to Savio Vasconcellos
who has helped on special advices and English revision.
8. Bibliography
Ceratti J. A. P., Nuñez W. P., Cruz L. L., Wickbolt V. S., “estudo comparativo do Desempenho
de um recapeamento utilizando asfalto Borracha”. Porto Alegre, 2006
Copavel Consultoria de Engenharia Ltda – SP-255 Highway Pavement Monitoring Reports

for Spvias – 2005, 2006, 2007,2008.
Copavel Consultoria de Engenharia Ltda – SP-270 Highway Pavement Monitoring Reports

for Spvias – 2005, 2006, 2007,2008.
Copavel Consultoria de Engenharia Ltda – SP-150 Highway Pavement Design Report for
Ecovias – RT_22_150_040_0_P13_001, 2004.
Copavel Consultoria de Engenharia Ltda – SP-150 Highway Pavement Design Report for
Ecovias – RT-22.160-011-5-P13-001, 2004.
Vasconcellos Barros, R. M, Vasconcellos F.S., “Pioneer Experience with Application of Gap

Graded Asphalt Rubber Mixes in São Paulo – Brazil”, AR-2003, Brasilia, 2003
Studies on Adhesive Performance of Waste

Crumb Rubber Modified Asphalt Mixtures
Wu Shaopeng 1* —Han Jun 2**
* Key Laboratory of Silicate Materials Science and Engineering of Ministry of Education,

Wuhan University of Technology, Wuhan 430070, CHINA
wusp@whut.edu.cn
** Key Laboratory of Silicate Materials Science and Engineering of Ministry of Education,

Wuhan University of Technology, Wuhan 430070, CHINA
hanjun@whut.edu.cn
ABSTRACT. At present, material recycling has become one of the most attractive science
alternatives, and different recycling methods are now available to specific pavement capability
and structural needs. The method that incorporates waste crumb rubber into asphalt concrete
was studied. The objective of this study was to investigate and evaluate the fatigue properties
in two of the most important factors, that is crumb rubber modifier size and temperature which
influence the performance of asphalt concrete greatly, were discussed by Direct Tensile Test
in different temperature. Using UTM25 electro-hydraulic servo-universal testing machine,
Uniaxial Direct Tensile Test was conducted on cylindrical specimens with different linearly
strain loading rate at different temperature. Through Direct Tensile Test, the result showed the
adhesive strengths were significantly enhanced. The recycling of existing waste crumb rubber
produces new pavements with considerable savings in material, money, and energy. They
have been used for many years, proving to be both economical and effective in protecting the
environment. Furthermore, waste crumb rubber modified asphalt concrete have been found,
for the most part, to perform as well as other modified asphalt.
KEYWORDS: Waste crumb rubber; modified asphalt; Direct Tensile Test; adhesive strength
1. Introduction
The use of various CRM in asphalt mixtures has increased significantly in recent years
around the world.[1-2] Research has shown that the addition of crumb rubber to base asphalt
produces binders with improved resistance to rutting, fatigue cracking, and thermal cracking
as well as reducing the thickness of asphalt overlays and reflective cracking potential.
Fatigue behavior of rubberized mixtures was found to be significantly improved compared to
conventional mixtures. [3-6]The previous research indicated that the rubber particle size is
important factor in affecting the rheological and engineering properties of mixtures because of
the interaction of CRM with the base binders. This interaction changes the physical properties,
viscosity and rheological properties of the CRM modified binder, leading to a high resistance
to the rutting of pavements [7-8]. The interaction process of CRM with binder is, essentially,
rather complicated, depending on the variables of CRM and binder (i.e., type, percentage,
size, grade, etc.) in addition to the mixing condition (temperature, mixing time, etc.). There
have been a few of researches identifying the influence of these individual variables on the
interaction. From these literatures, it was concluded that swelling of the rubber particles due to
the absorption of light fractions into the rubber particles, and stiffening of the residual binder
phase are the main mechanism of the interaction [9].
The objective of this study is to evaluate the anti-drawing properties of CRM asphalt
mixtures as a function of different size and temperatures using different methods in the
laboratory and suggest amount, size and temperature to satisfy the properties required in the
mix designs.
2. Experimental
2.1 Materials
Fine crumb rubber is made from waste wheel through crushing, which main component
is natural rubber (NR) and styrene-butadiene rubber (SBR) and rubber content is almost
55%. And classified by grain size, pulverizing crumb rubber (12-30 mesh), comminution
pulverizing crumb rubber h(30-47 mesh), fine crumb rubber(47-200 mesh)and super fine
crumb rubber(more than 200 mesh). The asphalt binder tested were base asphalt and CRM
asphalt prepared by ourselves with fine crumb rubber(CRM Asphalt-1) and super fine crumb
rubber(CRM Asphalt-2) at 142℃, whose specifications was showed in table 1.
Table 1. Specifications of asphalt binder
Item Results Index

Penetration/0.1mm 64 60-80(25℃)
Base
Softening point/℃ 48.5 44-54
asphalt
Ductility/cm ＞100 ≥100(15℃)
CRM
Softening point/℃ 56.8 ≥55
Asphalt-1
Ductility/cm 27.4 ≥20(5℃)
CRM
Softening point/℃ 60.4 ≥55
Asphalt-2
Ductility/cm 32.8 ≥20(5℃)
Gneiss coarse aggregates and fine aggregates (Bulk specific gravity of 2.68 and 2.72,
respectively) were used for the asphalt mixture. Limestone was selected as mineral filler.
The gradation type of AC 25 was selected for asphalt mixture. Mineral filler content is 2%
by weight, and preliminary asphalt content is 4.0%. Figure 1 shows the selected mixture
gradation curve.
100
upper limit of AC25
90 lower limit of AC25
middle of AC25
80
selected mixture gradation
70
percent passing(%)
60
50
40
30
20
10
0
0.01 0.1 1 10 100
Sieves (mm)
Figure1. Selected mixture gradation curve

2.2 Direct Tensile Test
The test used to evaluate the adhesion properties of the asphalt mixtures was a simple
Direct Tensile Test. The type of loading is like the one used to evaluate the asphalt-aggregate
interaction. In Figure 2 it can be seen a specimen being tested in simple tension.
Figure 2. Clamp used for DTT test
In Figure 3 is the typical result of a tensile test. The specimens were tested in simple tension
on base asphalt mixture sample at a constant deformation rate of 0.5mm/min at 15℃, quickly
reaching rupture and the maximum value of resistance. The overall results of the tensile tests
regarding the influence of temperature and strain loading rate are presented in Figure 5-7.
Based on the Strategic Highway Research Program (SHRP), Direct Tensile Test was
conducted at constant strain loading rate in this research [10,11]. The height/diameter ratio
and the maximum aggregate particle size are related and affect the final results, so the height
and diameter were both chosen 100mm at this research. The strain loading rate is 0.1, 0.5,1
and 18 mm/min. the experimental temperature is controlled at 15 and 25 ℃.
0.25
0.2
tensile strength /kN
0.15
0.1
0.05
0
0 0.5 1 1.5 2
displacement /mm
Figure 3. A typical result of the DTT test (15 ºC, 0.5 mm/min)
The test system used was a UTM25 electro-hydraulic servo-universal testing machine
made by Instron LTD, England. The clamp used for DTT test was constructed by self, as noted
in Figure 2.
2.3 Sample preparation
In this experiment, AC25 asphalt mixes was compacted into cylindrical samples. Optimum
asphalt content was 4 percent by weight of mix, and Air voids ranged between 3 and 4
percent.
Asphalt concrete specimens were prepared by gyratory compaction using 130 mm

diameter by approximately 120 mm high cylindrical specimens. After compaction, specimens
were cored and then the ends were sawed, taking special care in the cutting of the faces in
order to obtain fiat parallel faces. After this, we got specimens with 100 mm diameter and 100
mm high. The ends were cleaned before gluing them to the end caps with the epoxy resin.
Figure 4. Process of sample preparation
The sample preparation technique is importance in order to ensure that the failure plane
occurs far enough from the end caps and perpendicular to the sample axis. During the final
gluing step, we have pay especial attention to ensure proper alignment between the caps of
clamp and the specimen axis. Otherwise, shear stress will occur in the test and influence the
results.
Temperature control and experiment were all carried out in a controlled temperature
chamber. Specimens shall be placed in this chamber before test more than 3 hours and
monitored by a precise thermometer until its inner part has attained the test temperature.
3.1 Direct Tensile Strength
Direct Tensile Strength (DTS) is considered to be a very important performance

characteristic of asphalt pavement. The DTT test is carried out to define the tensile
characteristics of the asphalt concrete which can be further related to the cracking and
adhesion properties of the pavement [12].
According to the maximum load carried by a specimen at failure, the following equation is
used to compute the DTS:
P (1)
DTS = max2
π ×r
Where:
is the maximum applied load of each stress-elongation curve, as Figure 2 shows;
is the diameter of specimen.
1.8
1.6
15℃ 25℃
1.4
1.2
DTS/MPa
0.8
0.6
0.4
0.2
0
0.1 mm/min 0.5 mm/min 1 mm/min 18 mm/min
Strain loading rate/mm/min
Figure 5. DTS of base asphalt mixture
3.5
15℃ 25℃
3
2.5
DTS/MPa
1.5
0.5
0
Figure 6. DTS of CRM asphalt-1

4.5
4
15℃ 25℃
3.5
3
DTS/MPa
2.5
1.5
0.5
0
Figure 7. DTS of CRM asphalt-2
Figure 5-7 show the comparison of DTS of each group of asphalt mixtures. The results
indicated that DTS decreases sharply with the temperature increase from 15℃ to 25℃ under
the same strain loading rate. Both the base asphalt mixture and the CRM asphalt mixture have
the same relationship with the temperature. Higher strain rate seems to increase the possibility
of adhesive failure. At the same temperature, the percentage of adhesive failure at higher
strain loading rate is much more than the percentage at lower strain loading rate. And the finer
the crumb rubber size, the higher the adhesive strength of CRM asphalt.
4. Conclusions
Utilization of crumb rubber modification asphalt is effective and economical. Testing

results were analyzed through Direct Tensile Strength by UTM25 to evaluate the adhesion
properties of asphalt mixture. Conclusion can been drawn as follows:
1. The DTS decreases with the temperature increase from 15℃ to 25℃ at the same strain
loading rate.
2. At the same temperature, the percentage of adhesive failure at higher strain loading rate
is much more than the percentage at lower strain loading rate.
3. The finer the crumb rubber size, the higher the adhesive strength of CRM asphalt.
References
[1] Hicks R., Lundy J., Leahy R., Hanson D, Epps J., «Crumb rubber modifiers (CRM) in
asphalt pavements: summary of practices in Arizona, California and Florida»,Report
No. FHWA-SA-95-056, September, 1995.
[2] Ruth B., «Evaluation of ground tire rubber in asphalt concrete», Final report,
Department of Civil Engineering, University of Florida, January, 1992.
[3] Carl T., Khaldoun S., Serji A., “Effect of crumb rubber characteristics on crumb rubber
modified (CRM) binder viscosity”, Construction and Building Materials, xxx (2008)
xxx–xxx.
[4] Dantas S., Farias M., Pais J., Pereira P., Picado L., «Behavior of asphalt-rubber hot
mixes obtained with high crumb rubber contents», In Proceedings, asphalt rubber 2003
conference, Brasilia, Brazil, December 2003.
[5] Way G.B., The Rubber Pavements Association, «Technical Advisory Board leading the
way in Asphalt Rubber Research», In Proceedings, asphalt rubber 2003 conference,
Brasilia, Brazil, December, 2003.
[6] Amirkhanian S.N., «Establishment of an Asphalt-Rubber Technology Service» (ARTS),

In Proceedings, asphalt rubber 2003 conference, Brasilia, Brazil, December 2003.
[7] Airey G., Singleton T., Collop A.C., “Properties of polymer modified bitumen after
rubber–bitumen interaction”, Journal of Material Civil Engineering, 2002, 14(4), p.344
–54.
[8] Lougheed T.J., Papagiannakis A.T., “ Viscosity characteristics of rubber modified

asphalts”, Journal of Material Civil Engineering, 1996,8(3), p.153–6.
[9] Shen J., Amirkhanian S., “The Influence of crumb rubber modifier (CRM)
microstructure on the high temperature properties of CRM binder”, International
Journal Pavement Engineering, 2005, 6(4), p.265–71.
[10]Huang B., Li G., Louay N.M., “Analytical modeling and experimental study of tensile
strength of asphalt concrete composite at low temperatures”, Composites: Part B, 2003,
34, p. 705–714.
[11]Mo L., Wu S., Shui Z., “An Improvement in Electrical Properties of Asphalt Concrete”,
Journal of Wuhan University of Technology. 2002, 17(4), p. 69-72.
[12]Audrey R.C., Influence of moisture on bond strength of asphalt-aggregate systems,

Nashville, Tennessee, Vanderbilt University, 2007.
Rheological and Engineering Properties

of Rubberized Asphalt Concrete Mixtures
Containing Warm Mix Asphalt Additive
Feipeng Xiao, Serji N. Amirkhanian, Chandra Akisetty, and

Wenbin Zhao
Clemson University
Department of Civil Engineering
110 Lowry Hall
Clemson, SC 29634
feipenx@clemson.edu
abstract: In recent years, warm mix asphalt (WMA) is widely used for reducing energy
requirements and emissions in hot mix asphalt (HMA) industry. In addition, the use of
rubberized asphalt in the past has proven to be economical, environmentally sound and
effective in improving the performance of pavements across the U.S. and the world. WMA
additives reduce the mixing and compaction temperatures and achieve ideal workability of
HMA without significantly affecting the engineering properties of the mixtures. The objective
of this research was to investigate the performance characteristics (e.g. rheological and
engineering properties) of rubberized warm asphalt mixtures through a series of laboratory
tests on binders (viscometer, dynamic shear rheometer (DSR), and bending beam rheometer
(BBR) on the binders), and indirect tensile strength, rutting resistance, resilient modulus,
and fatigue behavior of mixtures. The experimental materials included one virgin binder (PG
64-22), one crumb rubber modified (CRM) binder (PG 64-22 + 10% -40 mesh rubber), two
aggregate sources, and two WMA additives (Aspha-min® and Sasobit®), and a total of twelve
mixtures were used in this study. The results of the experiments indicated that the uses of WMA
additives in rubberized HMA do not significantly reduce the rheological and engineering
properties of these mixes at lower mixing and compacting temperatures than the conventional
rubberized asphalt mixture.
keywords: Crumb rubber, Warm mix asphalt, Viscosity, G*sinδ, ITS, Rutting, Stiffness, Fatigue
life
1. Introduction
In recent years, the “warm mix asphalt” (WMA) is widely being promoted and used in
the hot mix asphalt (HMA) industry as a mean of reducing energy requirements and lowering
emissions. WMA can significantly reduce the mixing and compacting temperatures of asphalt
mixtures, by either lowering the viscosity of asphalt binders, or causing foaming in the
binders. Reduced mixing and paving temperatures would decrease the energy required to
produce HMA, reduce emissions and odors from plants, and make better working conditions
at the plant and the paving site (Kristjansdottir et al. 2007; Gandhi and Amirkhanian 2007;
Wasiuddin et al. 2007; Prowell et al. 2007; Xiao et al 2009a).
At the same time, recycling of scrap tires has also been of interest in the asphalt industry
throughout the world for over 40 years. Currently, approximately 82% of scrap tires are
utilized for such applications as tire-derived fuel, molded products, crumb rubber, and other
applications in the United States (RMA 2006). The mixing of crumb rubber with conventional
binders results in an improvement in the resistance to rutting, fatigue cracking and thermal
cracking (Button 1987; Way 2003; Xiao 2006). Many researchers have found that utilizing
crumb rubber in pavement construction is effective and economical (Airey et al. 2003; Huang
et al. 2004; Xiao et al. 2007; Xiao and Amirkhanian 2008; Xiao et al. 2009b).
However, the influence of crumb rubber and WMA additives mixed with virgin mixtures
together has not yet been identified clearly. The interaction of modified mixtures with
WMA additives is not well understood from the standpoint of binder properties and field
performance. It has been shown that the WMA additives reduce the mixing and compaction
temperatures and achieve ideal workability of HMA without significantly affecting the
engineering properties of the mixtures (Hurley and Prowell 2006; Prowell et al. 2007; Gandhi
and Amirkhanian 2007). While the addition of crumb rubber increases the demand of asphalt
binder and significantly increases the mixing and compacting temperatures, the modified
binder is helpful in resisting the high temperature deformation and extending the long-term
performance of the HMA. Because of the complicated relationships of the crumb rubber
and WMA in the modified mixtures, detailed information will be beneficial to help obtain an
optimum balance in the use of these materials.
The objective of this study was to investigate the performance characteristics of the
rubberized asphalt binder containing WMA additives and the rubberized asphalt concrete
mixtures containing the WMA additives. Experiments were carried out to evaluate rheological
properties of modified binder (i.e., viscosity, G*/sin δ, G*sin δ, and stiffness) as well as
mechanical properties of the mixtures (i.e., indirect tensile strength, rutting susceptibility, and
fatigue behavior).
2. Experimental design and procedures
2.1 Materials
One virgin binder (PG 64-22) and one crumb rubber modified (CRM) binder (PG 64-22 +
10% -40 mesh rubber) were used in this study. The rheological properties of CRM binder was
discussed in results analysis paragraph. The PG 64-22 binder was a mixture of several sources
that could not be identified by the supplier. One type of rubber, -40 mesh ambient rubber, was
produced by mechanical shredding at ambient temperature. To ensure that the consistency of
the rubber was maintained throughout the study, only one batch of crumb rubber was used
in this study (Table 1). Previous research and field projects conducted in South Carolina
indicated that the -40 mesh ambient rubber is effective in improving the engineering properties
of rubberized mixtures. Therefore, the -40 mesh rubber was employed in this study. To prepare
the modified binders, a reaction time of 30 minutes was considered suitable based on previous
studies indicating that the mixing time did not significantly influence the binder properties
(Putman 2005; Xiao et al. 2006).
Aspha-min® and Sasobit® were used in this study as the two WMA additives. Aspha-min
® is a Sodium–Aluminum–Silicate zeolite, which is hydro thermally crystallized as a very fine
powder. It contains approximately 21% crystalline water by weight. By adding it to an asphalt
mix, the fine water spray is created as all the crystalline water is released, which results in
volume expansion in the binder, therefore increasing the workability and compactability of the
mix at lower temperatures (Eurovia Services). Sasobit® is a long chain aliphatic hydrocarbon
obtained from coal gasification using the Fischer-Tropsch process. After crystallization, it
forms a lattice structure in the binder which is the basis of the structural stability of the binder
containing Sasobit® (Sasol Wax). More detail information regarding the two additives can be
found in other studies (Gandhi, 2008).
Table 1. Crumb rubber gradation
Ambient rubber (-40 mesh)

Sieve No. (mm)
% Retained % Cumulative Retained
30 (600) 0 0
40 (425) 9 9
50 (300) 31.9 40.9
80 (180) 32.9 73.8
100 (150) 7.6 81.4
200 (75) 18.6 100
Two aggregate sources (A and B) were used for preparing the samples (Table 2). Aggregate
A, granite source, is composed predominantly of quartz and potassium feldspar while
aggregate B (schist) is a metamorphic rock. Hydrated lime, used as an anti-strip additive, was
added at a rate of 1% by dry mass of aggregate. A total of 12 mixtures were evaluated in this
research.
Table 2. Aggregate Properties
Aggregate LA Absorption Sand

Abrasion Specific Gravity Soundness % Loss at 5 Cycles Hardness
Source Loss (%) (%) Equivalent
Dry SSD 11/2 to
Apparent 3/4 to 3/8 3/8 to #4
(BLK) (BLK) 3/4
A 51 0.80 2.740 2.770 2.800 0.2 0.1 0.1 - 5
B 34 0.60 2.780 2.800 2.830 0.4 0.6 0.9 35 5
2.2 Experimental Procedure
The combined aggregate gradations for a 12.5 mm mixture were selected in accordance
with the specification set by the South Carolina Department of Transportation (SCDOT).
The combined gradations for each aggregate source (A and B) are shown in Figure 1, which
shows that the design aggregate gradations for each aggregate source are the same when
using different WMA additives (Asphmin and Sasobit) at the same percentages of rubber (0%
or 10% rubber). The mixing and compaction temperatures of CRM mixture were 15-20 ℃
higher than warm mix asphalt mixtures that used 145-155 ℃ and 125-135 ℃ as mixing and
compaction temperatures, respectively.
100
80
Percent passing (%)
60
40
Agg. A
20 Agg. B
Low Range
Up Range
0
0.075 0.15 0.60 2.36 4.75 9.5 12.5 19.0
Sieve size (mm)
Figure 1. Gradation of aggregate sources
For this study, the optimum binder content, during the mix design process, was defined as
the amount required to achieve 4.0% air voids at a given number of design gyrations (Ndesign=
75). Six indirect tensile strength (ITS) specimens, compacted to 7 ± 1% air voids, were
used to evaluate the moisture susceptibility of various mixtures as modified AASHTO T283
procedures (no freeze/thaw cycle) were followed.
Six cylindrical Asphalt Pavement Analyzer (APA) specimens, for each mix type, were
compacted to 4.0 ± 0.5% air voids using a Superpave gyratory compactor. All testing with the
APA samples were carried out to 8,050 cycles to measure the rut depth of the HMA at 64℃.
The testing temperature was based on the virgin binder’s “performance grade” used in this
study.
Fatigue beams were made in the laboratory and four beams of each mixture were tested
for this study. All tests were performed in a temperature-controlled chamber at 20 ± 0.5℃.
In this study, a repeated sinusoidal loading at a frequency of 10 Hz was used; in addition, the
controlled strain mode was employed. The control and data acquisition software measured the
deflection of the beam specimen, computed the strain in the specimen and adjusted the load
applied by the loading device (AASHTO T321).
The test apparatus also recorded load cycles, applied load, and beam deflections. Failure is
assumed to occur when the stiffness reaches half of its initial value, which is determined from
the load at approximately 50 repetitions; the test is terminated automatically when this load
has diminished by 50 percent. The initial flexural stiffness of fatigue beam is determined as
follow (AASHTO T321):
aP (3l 2 − 4a 2 ) (1)
S =σ /ε =
4bδh 3
Where, S = initial stiffness; = tensile stress, in Pa; = maximum tensile stain, in m/m;
P = applied peak-to-peak load, in Newton; a = space between inside clamps, in meters; b =
average beam width, in meters; h = average beam height, in meters; δ = beam deflection at
neutral axis, in meters; and l = length of beam between outside clamps, in meters.
Statistically analyzed with 5% level of significance (0.05 probability of a Type I error) with
respect to the effects of aggregate sources and WMA additive types. For these comparisons, it
should be noted that all specimens were produced at optimum binder content.
3. Test results and discussions
3.1 Binder analysis
Table 3 indicates that the viscosity of rubberized asphalt binder decreases while the high
temperature performance (G*/sinδ) of overall binders increases with the addition of WMA
additive. The unaged binder test result shows that the Asphamin and Sasobit can improve
the workability (viscosity) and rutting resistance (G*/sinδ) of mixtures. The aged rubberized
binders show that the G*sinδ values decrease with the addition of rubber but these values
increase slightly as the WMA additives are added. It can be noted that the stiffness values
of binders have similar trends with G*sinδ values due to the addition of these materials.
Aged binder properties show that the WMA additives do not noticeably affect the long-term
performance of asphalt binder. For two WMA types, Sasobit shows a slightly greater influence
on the rheological properties regardless of the test conditions.
Table 3. Rheological properties of binders
Aging states No aging RTFO P AV

Viscosity G*/sin(δ) G*/sin(δ) G*sin(δ) Stiffness m-value
o
Test properties @ 135oC @ 64 oC @ 64 oC @ 25 oC @-12 oC @-12 C
(Pa-s) (kPa) (kPa) (kPa) (MPa)
Control (PG 64-22) 0.405 1.243 3.295 2970 183 0.311
10% CRM 1.600 0.934# 2.450# 1705 129 0.320
10% CRM+Asphamin 1.477 1.196 #
3.289# 2042 148 0.330
10% CRM+Sasobit 1.438 1.402# 3.325# 2160 151 0.290
Note: #~ testing at 76℃

3.2 ITS analysis
ITS values of mixtures made from aggregate sources A and B are shown in Figures 2 and 3,
respectively, it can be noted that the dry ITS values generally are greater than the wet ones, as
expected, but the mixtures containing WMA additive without rubber exhibit greater wet ITS
values regardless of the aggregate source. For the mixtures from aggregate A, it can be found
that the rubberized mixtures have slightly lower ITS values, conversely, the ITS values of
rubberized mixture are slightly greater as using the aggregate B.
In addition, the ITS values of all mixtures presented in Figures 2 and 3 are greater than the
value of 448 kPa (65 psi), a minimum wet ITS value set forth by SCDOT. Statistical analysis
(t-statistics) shows that there are no significant differences in the ITS values of the rubberized
mixtures containing two WMA additives. Similar results can be obtained from the control
mixtures. In addition, in comparison with two aggregate sources, it can be noted that the
mixtures from aggregate source A show lower ITS values regardless of the WMA additive or
rubber content.
3.3 TSR analysis
The TSR results are presented in Figure 4. It can be found that the ITS values of the
rubberized mixtures are lower regardless of aggregate sources and WMA types. As expected,
the rubber exhibits a negative effect on moisture resistance of asphalt mixtures. However,
these ITS values are greater than 85%, a minimum value of SCDOT’s specification. In
addition, for mixtures without rubber, it can be noted that the TSR values of those mixtures
containing WMA additive are greater than 100%, it seems that WMA additive benefits
the moisture resistance of these mixture regardless of the aggregate sources in this study.
Moreover, statistical analysis shows that there are no significant differences in TSR values of
the mixtures containing two WMA additives.
1000
Tensile Strength (kPa)
Wet ITS Dry ITS

800
600
400
200
0
0% Asphamin Sasobit 10% Asphamin Sasobit
Control Rubber Modified
Figure 2. ITS values of mixtures made with aggregate A

1000
Tensile Strength (kPa)
Wet ITS Dry ITS

800
600
400
200
0
Figure 3. ITS values of mixtures made with aggregate B
3.4 Rut depth analysis
The rut depth results shown in Figure 5 indicate that, as expected, the rut depths of the
rubberized mixtures have lower values than other mixtures. Previous research by Xiao et
al. 2007 presented a similar result. In general, the mixtures from aggregate source A show
greater rut depth, this indicates that aggregate properties play a key role in determining the rut
resistance of the mixtures. The mixtures containing two WMA additives show no significantly
different rut depth values at α = 0.05 level. As a result, the effect of WMA additives on the
rut depth of the mixture can be ignored in this study although the addition of these materials
increases the G*/sin δ value of the binder.
120
Agg. A Agg. B
100
TSR (%)
80
60
40
20
0
Figure 4. TSR values of mixtures

7.0
Agg. A Agg. B
6.0
Rut Depth (mm)
5.0
4.0
3.0
2.0
1.0
0.0
Control 10% Rubber Modified
Figure 5. Rut depth values of mixtures
3.5 Stiffness analysis
As shown in Figure 6, it can be noted that the initial stiffness values (Equation 1) of the
mixtures from aggregate source A are lower. The additions of rubber and WMA additive do
not significantly affect the initial stiffness values of the mixtures. There is also no trend for
these stiffness values based on the test results in this study.
20.0
Agg. A Agg. B
16.0
Stiffness (MPa)
12.0
8.0
4.0
0.0
0% 10% Rubber Asphamin Sasobit
0% WMA 10% rubber
Figure 6. Initial stiffness values of mixtures
3.6 Fatigue life analysis
Fatigue life values of the mixtures are shown in Figure 7. Obviously, the standard
deviations of the fatigue test results for each mixture are large since the variability of fatigue
life is generally based upon the micro-structure of beams (e.g. the aggregate-binder interface,
the void size distribution, the interconnectivity of voids, distribution of aggregate particles,
film thickness and the aged status of binder). Previous research has indicated that increasing
the number of the repeated specimens reduced the variability (Xiao et al 2007). Moreover,
in comparison with the control fatigue beam, fatigue life of mixture shows no obvious trend
when the rubber or WMA additive were added.
1,000
Agg. A Agg. B
Fatigue life (x10 3 cycle)
800
600
400
200
0
0% 10% Rubber Asphamin Sasobit
0% WMA 10% rubber
Figure 7. Fatigue life values of mixtures
4. Conclusions
Based on the experimental data shown in this limited study, the following conclusions are
reached:
• The viscosity of rubberized asphalt binder slightly decreases while the high temperature
performance (G*/sinδ) of overall binders slightly increases with the addition of WMA
additive. The unaged binder test result shows that the Asphamin and Sasobit can
improve the workability (viscosity) and rutting resistance (G*/sinδ) of mixtures.
• Statistical analysis shows that there are no significant differences in the ITS values of
rubberized mixtures containing two WMA additives. Similar results can be obtained
from the control mixtures. In addition, in comparison with two aggregate sources,
the mixtures from aggregate source A show lower ITS values regardless of the WMA
additive or rubber content. Moreover, statistical analysis shows that there are no
significant differences in TSR values of the mixtures containing two WMA additives
• Rut depths of the rubberized mixtures have lower values than other mixtures. In
general, the mixtures made with aggregate source A show greater rut depth. The
mixtures containing two WMA additives show no significantly different rut depth
values.
• The additions of rubber and WMA additive do not significantly affect the initial
stiffness values of the mixtures. There is also no trend for these stiffness values based
on the test results obtained in this study. The standard deviations of the fatigue test
results for each mixture are high since the variability of fatigue life is generally based
upon the micro-structure of beams. Fatigue life of mixture shows no obvious trend
when the rubber or WMA additive were added.
5. Acknowledgement
The authors wish to thank the financial support of South Carolina Department of Health
and Environmental Control (SC DHEC) to conduct this research work.
6. Bibliography
Airey G.D, Rahman M. M., and Collop A. C., “Absorption of bitumen into crumb rubber
using the basket drainage method”, The International Journal of Pavement Engineering,
Vol. 4(2), 2003, pp. 105-119.
Button, J. W., Little, D. N., Kim, Y., and Ahmed, J., “Mechanistic evaluation of selected asphalt
additives”, AAPT, Vol. 56, 1987, pp. 62-90
Eurovia Services: http://www.eurovia.com/en/produit/135.aspx
Gandhi, T., “Effects of Warm Asphalt Additives on Asphalt Binder and Mixture Properties”,
Ph.D. dissertation, Clemson University, Clemson, South Carolina 2008
Gandhi, T. and Amirkhanian, S., “Laboratory Investigation of Warm Asphalt Binder

Properties – A Preliminary Investigation”, MAIREPAV5 Proceedings, Vol. 5, pp. 475-480,
Park City, Utah, 2007
Huang, B., Li, G., Pang, S.S., and Eggers, J.H., “Investigation into waste tire rubber-filled
concrete”, Journal of Materials in Civil Engineering, Vol. 16, 2004, pp. 187-194
Kristjansdottir, O., Muench, S, Michael, L, and Burke, G., “Assessing Potential for Warm-Mix
Asphalt Technology Adoption”, Journal of the Transportation Research Board, No. 2040,
2007, pp. 91-99
Prowell, B., Hurley, G., and Crews, E., “Field Performance of Warm-Mix Asphalt at National
Center for Asphalt Center for Asphalt Technology Test Track”, Journal of the
Transportation Research Board, No. 1998, 2007, pp. 96-102
Putman B.J, “Quantification of the effects of crumb rubber in CRM binder”, Ph.D.
dissertation, Clemson University, South Carolina 2005
Rubber Manufacturers Association (RMA) 2006, Scrap tire markets in the United States,
2005 Edition, RMA, Washington, D.C.
Sasol Wax: http://www.sasolwax.com/Sasobit®_Technology.html
Wasiuddin, N., Selvamohan, S., Zaman, M., and Guegan, M., “Comparative Laboratory
Study of Sasobit® and Asphamin® Additives in Warm-Mix Asphalt”, Journal of the
Transportation Research Board, No. 1998, 2007, pp. 82-88
Xiao F., “Development of fatigue predictive models of rubberized asphalt concrete (RAC)
containing reclaimed asphalt pavement (RAP) mixtures” Ph.D. dissertation, Clemson
University, South Carolina 2006
Xiao F., Putman B.J., and Amirkhanian S.N., “Laboratory investigation of dimensional changes
of crumb rubber reacting with asphalt binder,” Proceedings of Asphalt Rubber 2006
Conference, Palms Spring, California, 2006, pp. 693-715
Xiao F., Amirkhanian S.N., and Juang C.H., “Rutting resistance of rubberized asphalt
concrete pavements containing reclaimed asphalt pavement mixtures”, Journal of Materials
in Civil Engineering, Vol. 19(6), 2007, pp. 475-483
Xiao F., and Amirkhanian S.N., “Resilient Modulus Behavior of Rubberized Asphalt
Concrete Mixtures Containing Reclaimed Asphalt Pavement,” Journal of Road Materials
and Pavement Design, Vol. 9 (4), 2008, pp. 633-649
Xiao F., Jordan J., and Amirkhanian S.N., “Laboratory Investigation of Moisture Damage in
Warm Mix Asphalt Containing Moist Aggregate,” Transportation Research Board (CD-
ROM), Washington, D.C., 2009a
Xiao F., Amirkhanian S.N., Shen J.N., and Putman B.J. “Influences of Crumb Rubber
Size and Type on Reclaimed Asphalt Pavement (RAP) Mixtures” Construction and
Building Materials, Vol. 23(2). 2009b, pp.1028-1034
Stability Assesment Through Solubility and

Rheological Measurements of Gtr-modified
Bitumen
A. Pérez-Lepe — A. Páez
Repsol Technology Centre
Autovía de Extremadura (N-V), km 18

Móstoles. 28931-MADRID (SPAIN)
Tlf: +34913489490
aperezlepe@repsol.com
apaezd@repsol.com
ABSTRACT. GTR(ground-tyre rubber)-modified binders or CRMB (crumb rubber-modified

binders), by the so-called wet process, represent a particular case of PMB (polymer-modified
binder), due to the specific features of this recycled-to-be rubber: a mixture of natural and
synthetic vulcanized rubbers, carbon black and plasticizers among others. When the GTR is
incorporated to the bitumen, the strong structure created through vulcanization and the strong
bond between carbon black and rubber (carbon gel or bound rubber) hinder the swelling of
the rubber and the extensive dispersion within the bitumen. Crumb rubber particles tend to
settle towards the bottom of the storage tanks.
The use in many formulations of co-additives may hide or cover up to some extent the results
of PMB-criterion stability, showing as stable binders with high concentration of decanted
fine solid particles. Therefore, a more severe assessment is proposed, based on solubilisation,
in order to quantify the concentration of “non-assimilable” rubber particles, allowing thus
for a sedimentation extent and profile kinetics evaluation, as well as an assessment of the
extent of digestion of the rubber. This method was applied to several binders manufactured
with different concentration of GTR, different additives and different processes, to support
the solubility approach. When these portions of the stored samples were analyzed through
rheological measurements, a good agreement was in most cases found with the solubility tests.
The prediction of stability was also evaluated at the industrial production site as well as the
asphalt plants storage tanks.
KEYWORDS: GTR, wet process, stability, solubility, rheology.

1. INTRODUCTION
The modification of bitumen with a range of polymers yields polymer-modified binders

(PMB) showing enhanced service properties. Depending on both polymer and bitumen, as
well as on the mixing process, the polymer is partially swollen by the light fractions of the
bitumen by the so-called physical distillation process (Lesueur, 1998). In most cases, both
systems, polymer and bitumen, are incompatible, and a compatibilization method must be
developed in order to avoid a possible destabilization and gross-phase separation phenomenon.
For many polymers (plastomers and elastomers), the mechanism by which the polymer-rich
phase ends up separated from the bitumen-rich medium, begins with a coalescence stage (Hesp
et al., 2002; Pérez-Lepe et al., 2006).
Ground-tyre-rubber (GTR) modified binders obtained by the so-called wet process

represent, due to the specific features of ground-tyre rubber, a different type of binder than
binders modified with elastomers or plastomers. GTR is generally composed by a mixture
of natural and synthetic vulcanized rubbers, carbon black and plasticizers, among others.
When the GTR is incorporated to the bitumen, the strong structure in the tyre created through
vulcanization and the strong bond between carbon black and rubber (carbon gel or bound
rubber) hinder the swelling of the rubber and the extensive dispersion within the bitumen
(Billiter et al., 1997; Panagiotis et al., 2004). Due to the greater density of the non-digested
GTR particles at common storage temperatures, a clear tendency to particle settling is
observed in most cases.
The wet process involves a mixing stage of rubber and bitumen to form a modified binder,
previous to the mixing with the aggregates. The dry process implies that the rubber is mixed
in-situ with the aggregates, with or without a short pre-mixing stage. The dry-process solution
may eventually cause some cohesion shortcomings (Páez et al., 2004) as well as quality
control and heterogeneity problems. Many efforts have been carried out in order to obtain a
GTR-modified binder such that the maximum beneficial properties of the rubber at the ground
tyre particles, is transferred to the final binder and consequently, the recycling of tyres is of
greater value. Additionally, the manufacture of modified bitumen at the binder production
site is a logistics/cost-beneficial issue, provided the binder proves to be hot-storage stable
during the transportation and storage previous-to-mix stages. Although a continuous stirring
or recirculation in storage tanks reduces to a great extent the subsequent problems of settling,
these best-practice facilities are not always readily available. The present work is focused on
the production of stable GTR-binders with enhanced properties by the wet process.
Based on the experience of GTR on the Spanish roads, the use of crumb rubber is regulated
by the standard OC 21-2007 for road construction (published by the Roads and Transportation
Spanish Government Section). Binders produced by the wet process are permitted in the top
traffic level (T00), while the dry process is restricted to a medium-level traffic demand (T2
level, in a range from T4 to T0, plus T00 as top extra level). Two groups of GTR binders are
standardized for binder-production-site manufacture: those with low content of crumb rubber
or low level of modification and those with high viscosity and high rubber content.
Various methods of stabilization, as the use of additives (synthetic polymers, reactive

polymers, waxes, acids, aromatic oils,...) and/or high-digestion processes, as well as activation
of the surface through additives or through grinding, have been carried out by different
researchers in the field in order to obtain stable binders (Soto et al., 2005; Billiter et al., 1997;
Peña et al., 2005; Bahia et al., 1994). The stability of the binder is an issue of importance,
even when the plant storage tank is sufficiently agitated and the material is periodically
recirculated (not always available).
The commonly accepted or practiced high-temperature storage stability criterion consists

of assessing the difference of penetration and softening point between the upper and bottom
binder portions of a binder tank, stored at high temperature, either a toothpaste tube in a
laboratory oven or a large scale tank . This is the same criterion used for all PMB’s (72h of
still storage at 180℃, as pointed out by EN13399 standard). In the present work, a need for
other types of tests for stability assessment is remarked.
The needle penetration test is not a reliable technique for asphalt-rubber. The cone-
penetration test, commonly used for waxes or particulate-filled systems, is envisaged as an
alternative. Even so, the use of additives in formulations may “hide” to some extent the results
of PMB-criterion stability, showing as stable binders with high concentration of decanted fine
solid particles. Therefore, a more severe assessment is proposed. The dissolution of the upper
and lower portions, as well as the original blend, with a specific solvent (THF or toluene)
renders the Concentration of “non-assimilable” rubber particles. By a simple solubility test,
an assessment of the settling extent and a kinetics profile evaluation may be inferred, as well
as an assessment of the extent of digestion of the rubber (Billiter et al., 1997; Youtcheff et
al., 2004). The digestion of rubber, and more specifically the quantity of rubber (synthetic
and natural) effectively transferred to the bitumen, can be further extracted from thermo-
gravimetric analysis.
The thermal degradation of GTR exhibits four main stages as registered from weight loss
under increasing temperature ramps. At the beginning, the plasticizers are lost within a range
of 100℃ to 300℃, approximately. From 270℃ to 440℃, the natural rubber in the particles
is decomposed. Synthetic binders appear on the thermogram between 410℃ and 520℃. A
weight loss peak related to the degradation with oxygen of carbon black is found at higher
temperatures (600℃-700℃).
By evaluating the rubber particles before and after the mixing with bitumen and right
before the application on the field, the degree of incorporation or digestion of rubber could be
attained.
Natural rubber vulcanised with common crosslinking promoters exhibits a transition due
to the break and looseness of bonds from just 170℃ upwards (as registered by DSC). This
phenomenon would eventually lead to a more viscous and sticky material that could promote
particle size growth and/or irreversibly adhere to hot surfaces. This issue could induce a heat
transfer efficiency loss within the storage tanks, with the consequent storage problems, even in
the case of agitated containers. When the majority of the loose natural rubber (not as carbon
gel) has already been transferred to the bitumen through a high-digestion process or anti-stick
additives are used, the problem is minimized.
2. RESULTS
Ground tyre rubber analysis
Crumb rubber composition was studied by a thermo-gravimetric analysis in order to

decompose at a glance the GTR particles into the different components at the reception of the
tyre powder from different sources, type of grinding and size distribution.
In all cases, the same pattern was observed, with 4 main stages of material decomposition
and loss under increasing temperature: plasticizers, followed by natural rubber, synthetic
rubbers and carbon black (CB).
Depending on the origin of tyres (mainly, the relation of heavy-load-vehicle tyres to

automobile tyres), the grinding process (cryogenic or ambient temperature) and the cleaning
process (removal of steel and textile fibres), two important features may be extracted from the
thermograms: the total amount of rubber and the final residue over the CB. As a reference,
minimum 60% weight loss before carbon black (rubber + plasticizers) and maximum 10%
of final residue is an acceptable raw material, for the storage stable binder + high digestion
approach. Volatiles content are expected to be very low at common HMA temperatures. For
example, at a maintained temperature of 185℃, a ground tyre rubber mass would lose a 0,7%
weight of volatiles (from a TGA isothermal analysis). An identification of compounds is being
carried out by TGA-MS.
In figure 1, a typical thermogram of GTR particles is presented. With the derivative of

weight (%/℃), the extension and form of the loss peaks can be evaluated.
Figure 1. Thermogram of a typical crumb tyre rubber. Weight percentage and derivative of
weight percentage with temperature are represented.
Shape and surface area of the particles are important for the digestion of the crumb rubber.
Depending basically on the grinding process, the surface area obtained can be completely
different. Cryogenic grinding is expected to yield more spherical particles with less surface
area than the ambient temperature grinding. A tearing of rubber while grinding is an option
yielding particles with high accessible area. Particle analysis with digital image processing
was found a useful technique to measure size distribution and shape of the particles, as well
as an estimate for the surface area. In figure 2 and 3, the results of the image analysis for two
different kinds of particles (ground at ambient temperature with two different processes) are
represented. In the first case, figure 2, a bimodal distribution is obtained as representing the
amount of particles (Q3 in y-axis) sorted according to size (x-axis). The surface area in the first
case is 180 /g, while the particles represented in figure 3 have an average of 430 cm2/g.
Figure 2. Image analysis size distribution of crumb rubber with low surface area. Q3(number
of particles(%)) vs size of particles (xMa_min) and q3 (number of particles (%)/size).
Figure 3. Image analysis size distribution of crumb rubber with high surface area. Q3(number
of particles(%)) vs size of particles (xMa_min) and q3 (number of particles (%)/size).
Instability and settling of GTR particles
Due to the higher density of a single crumb particle, a settling is expected to occur when
GTR is mixed (even extensively) with bitumen, during a high-temperature still storage. Crumb
rubber is therefore gradually accumulated on the bottom part of tanks with no or non-effective
agitation. A higher viscosity and tackiness is formed on the bottom part of the containers,
causing tank problems of pumping, sticking to the heating coils and heterogeneity. This
settling phenomenon can be reproduced with toothpaste tubes stored in an oven. The analysis
by needle penetration is not able to reveal the cumulative quantity of digested particles on the
bottom part of the stored binder. In order to evaluate the extent and profile of sedimentation,
a basic dissolution method was applied. After the dissolution in THF (although toluene is
also a preferred solvent) of a selected portion from the stability tube or height of the tank,
the remaining digested material is vacuum recovered through a 2,7 µm fibre glass filter. A
quantification of undigested rubber is done and the percentage weight of rubber on the bottom
part and the upper part represents a valuable stability criterion, under the more severe storage
conditions in cases where no agitation is available or required. Depending on the amount
and features of the rubber added, viscosity and chemical composition of bitumen and the
processing conditions, the settling would be ruled by a different profile. A further analysis by
TGA of the recovered rubber particles can reveal the quantity of rubber dissolved or transferred
to the bitumen phase. Following this set of experiments, an average of 23% of the total rubber
content is found to be lost through the mixing and temperature digestion process. The rest of
the rubber remains strongly bound to carbon black. The particles after the digestion show a
higher carbon black concentration and higher residue after 1000℃ (Figure 4).
Figure 4. Thermogram of crumb tyre rubber(— · —)and the recovered particles after a
stability test form the upper section(—) and bottom section ().
In a paper by the authors of the present work (2007), several binders produced without
or with the addition of anchoring, bitumen structure modification or surface treatment
compounds, revealed that the stability of a GTR-binder could not be assessed by the
penetration and softening points of the binder. In cases where the differences of these two
parameters were under 5-10 units, a high concentration of crumb particles was found on
the bottom parts of the stability tubes. On the contrary, compositions showing values for
penetration and softening point in the range 5 to 10 units, showed no accumulation of particles
on the bottom.
From a rheological point of view, the behaviour of a GTR-modified binder material and the
presence of rubber particles should be noticed through dynamic shear tests. In some cases, the
system proposed for stabilisation or co-modification is a polymer (synthetic rubber in many
cases). Some instability indices based on a comparison of modulus or dynamic viscosity at a
specific frequency, after a hot storage test, have appeared on the literature. For this purpose,
the storage modulus, G’, is a preferred selective parameter, registered in the low frequency
region and preferably at temperatures higher than 50℃, where the elastic rubber particles are
best differentiated from the bitumen matrix. If the conditions of the rheological tests are not
carefully selected, the presence of particles could not be reliably noted.
Even so, with the presence of polymers or waxes in the composition a basic rheological
study may hide the settling of particles. In Figure 5, the complex modulus, G*, and the
dynamic viscosity, η*, were registered within a range of frequencies of oscillation. The upper
and bottom portions of a stored binder were tested at 60℃. The behaviour of both samples
was very similar in the low frequency region, with only a small increase in G* for the bottom
portion. The differences of penetration and R&B softening point were 9.6 1/10mm and
4.8℃. The bottom sample had a content of filtration recovered material of 15.3% (percent
weight over the total amount), whereas the upper portion only 0.8% (the original binder had
a 10% content of crumb rubber). Even with severe settling, the presence of particles was not
quantitatively noticed.
When an effective stabilisation process is applied, the rheological curves before and after
the stabilisation period show identical behaviour (figure 6). After dissolution and filtration
of the samples, the content of undissolved matter is 7.4% , 7.7% and 7.6%, for the upper,
medium and bottom parts of the stability tube.
Industrial production of stable binders
The manufacture of stable tyre rubber binders is carried out at a central site mixing facility.
From the binder production site, the product is delivered in a tanker to the aggregate plant,
where it is stored. From the manufacture point till the moment of mixing with the aggregates,
the binder is preferably kept at a temperature close to the temperature of mixing and agitated.
The product is not rarely subjected to gradients of temperature, periods with no agitation, or
extended storage due to works delay. The product must then be completely storage stable in
order to count on a robust binder for this cost-effective production-site application.
Figure 5. Evolution of complex modulus and dynamic viscosity with angular velocity at 60℃
of the upper and lower portions after stability test (120h-163℃) of a 10% GTR binder with
polymeric additive.
The blend and reaction stage of rubber and bitumen can be carried out with a traditional
batch mixing facility. An in-line mixing of the components is preferred for production time
saving. As higher shear is achieved by adjusting mainly temperature and flow, a higher
homogeneity and digestion extent may be obtained, in the following swelling stage.
10 5
Upper
10 4
G'
G', G'' (Pa), η* (Pa穝 )
G''
η*(c)
Lower
10 3 G'
G''
η*(c)
10 2
10 1
1E-3 0,01 0,1 1 10 100
freq. (s -1)
Figure 6. Evolution of storage (●) and loss (▪) moduli and dynamic viscosity (▲) with
frequency at 60℃ of the upper (empty) and lower(full) portions after stability test (120h-163
℃) of a 10% GTR binder with polymeric additive.
Evaluation of ground tyre rubber in certain hot-mix asphalt applications. Product range and
applications
Repsol-YPF is an integrated international oil and gas company, operating in more than
30 countries and is the leader in Spain and Argentina. It is one of the ten major private oil
companies in the world and the largest private energy company in Latin America in terms
of assets. Repsol has a vast experience with the recycling of tyre rubber in roads through
modified binders produced by the wet process at the binder production site. Since year 2004,
binders for more than 400000Tm of HMA have been produced with an own developed
technology to ensure binder homogeneity and stability.
Three levels of modification have been standardized, but tailor-made products are
manufactured based on a precise application.
a) Low modification level

b) PMB equivalence
c) High-viscosity binders
a) Low modification level
These low modified rubber binders (production-site manufactured) present a response

which falls between those of base bitumen and polymer-modified binders, with no special
elastic recovery requirements. A specification for these binders was introduced in Spain for
two penetration grades: 35-50 and 50-70 1/10mm. Typical binders properties produced at
Repsol are showed in table 1.
Table 1. Low modification binder standard properties.
BINDER 35-50 grade 50-70 grade

Penetration 25 ℃, 1/10 mm 48 61
R&B SP, ℃ 59.4 53.9
PI +0,8 +0.2
Brookfield viscosity, cP
At 135º C 1170 830
At 150º C 595 425
At 180º C 216 145
Storage Stability
Penetration difference, 1/10mm 5 4
R&B difference, ℃ 2.8 3.7
Difference THF-residue (lower-
0.9 1.5
upper), %.
RTFOT
Penetration retained (%) 67 66
Variation R&B., ℃ 7.6 6.5
Variation mass (%) -0.100 0.045
As an example, the behaviour of this binder (50/70) is studied for a dense-graded asphalt
mix (20 mm nominal maximum size) and compared with base bitumen of the same penetration
grade. Specimens were analyzed following the Marshall method as a reliable HMA design
method allowing for optimum asphalt binder content determination. In table 2 and 3 Marshall
results are presented for unmodified and modified binders, respectively. The aggregates size
distribution is shown on figure 7.
SIZE DISTRIBUTION OF AGGREGATES
100
90
80
70
60
% PASSING
50
40
30
20
10
0
40 25 12,5 8 5 2,5 1,25 0,63 0,32 0,16 0,08
Sie ve UNE, mm
sieving gradation allowance range
Figure 7. Size distribution of aggregates according to a UNE sieve analysis.
As can be seen from the Marshall results, a higher density is obtained as well as higher
stability for the modified binder called BC50/70 (low-modified rubber binder). In the case of
modified binder, the optimum content was selected on order to have a recommended 4,5% air
voids. The optimum content with this criterion is higher for the modified binder than for the
unmodified bitumen.
kN Stability Density % % air voids

2,410
8,0
16,00 7,5
2,400
7,0
2,390 6,5
15,00 6,0
2,380 5,5
5,0
14,00 2,370 4,5
4,0
2,360
13,00 3,5
2,350 3,0
2,5
12,00 2,340 2,0
3,00 4,00 5,00 6,00 3,00 4,00 5,00 6,00 3,50 4,00 4,50 5,00 5,50 6,00
% binder % binder
%
m flow % % voids in MA
16,0
7,
15,5
6,
15,0
5,
14,5
4,
3, 14,0
2, 13,5
1, 13,0
3,00 3,50 4,00 4,50 5,00 5,50 6,00
3,0 4,0 5,0 6,0
% %
Figure 8. Marshall method results on a 60/70 penetration grade bitumen: stability, density, %
air voids, flow and % voids in mineral aggregates.
kN Stability density %
% air voids
2,400
9,0
19,00 2,390
8,0
17,00 2,380
2,370 7,0
15,00
2,360 6,0
13,00
2,350 5,0
11,00
2,340
9,00 4,0
2,330
7,00 2,320 3,0
5,00 2,310 2,0
3,00 4,00 5,00 6,00 3,00 4,00 5,00 6,00 3,50 4,00 4,50 5,00 5,50 6,00
% Binder % Binder % Binder
m flow % % VMA
17,0
7,
6, 16,0
5, 15,0
4, 14,0
3,
13,0
2,
1, 12,0
3,00 4,00 5,00 6,00
0,0 2,0 4,0 6,0 8,0
% Binder % Binder
Figure 9. Marshall method results on a low-modified rubber binder BC50/70: stability,

density, % air voids, flow and % voids in mineral aggregates.
Dynamic modulus experiments (NLT standard) and fatigue analysis (NLT standard) were
carried out at a content of 4,5% b/m, for unmodified and CRMB (table 2). Dynamic modulus
testing was carried out at 20℃ over tall cylindrical specimens compacted by vibrocompression
and submitted to repetitive axial loading at 10 Hz.
Table 2. Dynamic modulus and Fatigue results for bitumen and tyre-rubber binder BC50/70.
Bitumen 60/70 BC 50/70

Modulus @ 20℃ - 10 Hz ( MPa) 6116 4781
Phase angle (°) 20.01 21.92
a: 3.3672 a: 3.5869
Fatigue Law
b: -0.1966 b: -0.2412
log ε = a + b * log N
R2: 0.79 R2: 0.76
ε6 149.83 137.94
Mixes prepared with CRMB and base bitumen show a similar response towards Fatigue. A
higher resistance to permanent deformations is obtained by the rubber binder at 60℃. In table 3,
the wheel-tracking test results on both samples are presented.
Table 3. Wheel-tracking test results for bitumen and tyre-rubber binder BC50/70.
BINDER 60/70 BC 50/70

Total Deformation 2.043 mm 1.103 mm
Rate of Deformation interval 30-45 min. 14.10-3mm/min 6.10-3mm/min
Rate of Deformation interval 75-90 min. 10. 10-3mm/min 3. 10-3mm/min
Rate of Deformation interval 105-120 min. 9.10-3mm/min 2. 10-3mm/min
Laying and compaction can be done with conventional machinery, but a special attention
may be paid due to a slight more difficulty to compact the CRMB mixes due to a certain
elastic response of the rubber, even at compaction temperatures . A mixing temperature of
185℃ and compaction at 175℃ was used for this particular work.
b) PMB
Traditional PMB’s, usually made with elastomers SBR, SBS or plastomers EVA, can find
a substitute in rubber-containing binders. A crumb rubber particle has about 50%-60% rubber
in its composition. Not all the rubber can be efficiently swollen by the light fractions in the
bitumen. In order to get similar physical properties than a PMB, more crumb rubber weight
content is therefore needed. As a reference, around 10% rubber content is commonly used
for this grade. AR binders have a good performance in gap-graded mixes as well as in open-
graded mixes. The use of GTR binders in open-graded layers enables a considerable noise
reduction due to tyre/road friction interaction. A greater binder film thickness is acquired due
to the presence of CR. This is reported to help the noise reduction offered by a high number of
voids.
As an example, a HMAR with 25% VMA was extended using a medium-modification

level crumb rubber binder. The Spanish binder grade used corresponds to BM-2. Binder
properties are shown on table 3, compared to those of the corresponding PMB (with synthetic
elastomers). In figure 10, the sieving size distribution of the aggregates is presented.
The storage stability of the binder was evaluated by applying the dissolution approach. The
content of non-digested crumb was practically identical in the upper and lower portions of the
binder after hot temperature storage.
The problem of settling constitutes a heterogeneity problem and bulk viscosity evolution
during storage. It can be treated as settling of concentrated suspensions of elastic solids in a
viscoelastic matrix. A dynamic shear rheometer may be used to evaluate the evolution and
profile of settling during hot temperature storage. A tall cup filled with binder was kept heated
a several temperatures (163℃ and 180℃) and the binder was then studied with a vane rotor
coupled to the rheometer head. Although dynamic experiments can be carried out in order to
register the changes in elasticity due to rubber migration, continuous flow at low shear rate
with the upper sensor on the bottom part of the container (to register clarification) or on the
lower part (to register sedimentation), is also a good method to track the possible settling
phenomenon. In figure 11, the viscosity evolution was tracked on a DSR at 163℃ during 48
hours of storage. No increase in viscosity was reported.
Table 4. Properties of rubber-modified binder and polymer-modified binder.
BINDER BM-2(SBR) BMC-2(GTR)

PENETRATION P25℃ (1/10 mm) 43 46
R&B (℃) NLT-125 72.4 69
PI +2.85 +2.50
FRAASS T (℃) -16 -17
DUCTILITY 5℃(cm) 12 10
STORAGE STABILITY
R&B difference 1 3
Penetration difference (25℃) 3 4
ELASTIC RECOVERY Torsion 25℃ 70% 61%
VISCOSITY 135℃ (cP) 2724 3500
VISCOSITY 150℃ (cP) 1300 1900
VISCOSITY 175℃ (cP) 375 401
RHEOMETRY
AASHTO TP5 8900 9550
Complex Modulus, G* (Pa) 60℃
Phase angle, δ, (º) 60℃ 55 57

% passing
100
90
80
70
60
50
40
30
20
10
0
40 25 12, 8 5 2, 1,2 0,6 0,3 0,1 0,0
Sieve UNE, mm
Sieving gradation allowance

viscosity_settling
viscosity (Pa s)
1
0.8
0.6
0.4 η_settling
0.2
0
-0.2 0 1000 2000 3000
time (min)
Figure 11. Viscosity evolution during 48 hours of BMC-2 crumb rubber binder.
At the production site, the product was previous to serve evaluated in the storage tanks to
ensure stability. Samples from the upper part of the tank and through the bottom discharge
sampler were tested to check the complete homogeneity of the product. No problem associated
with the instability of the product was found during the aggregate mix preparation.
Due to a bad quality of the aggregates, a lack of adhesion was noted, even for a PMB of
the same grade used on purpose. An adhesion promoter additive (surfactant type) was added
during the formulation of the binder with satisfactory results. Mix design was established with
the Cantabrian (resistance to disintegration) and drainage tests for open-graded mixes. An
optimum of 4,75% b/m was selected.
Table 5. Mix-design for an open-graded asphalt-mix with PMB and GTR-modified binder.
Data for abrasion resistance tests.
Mixing Temperature 160℃

Compaction Temperature 150℃
Blows per side 50
Test Temperature 25℃
binder BMC-2 (GTR) % b/m 4.0 4.5 5.0
Binder BM-2 (PMB) % b/m 4.5
binder BMC-2 + adhesión promoter 4.0 4.5 5.0
Density g/cm 3
1.936 1.950 1.963 1.988 1.939 1.946 1.958
Air voids % 21.6 21.6 20.3 19.0 22.0 21.2 20.2
Dry Abrasion Loss % 38 30 29 14 23 22 17
Wet Abrasion Loss (24h / 60º C) % 68 37 33 25 36 32 18
c) High-viscosity binders
For non-continuous mix gradation, high contents of binders with a high viscosity and high
stiffness modulus are a preferred solution to obtain mixes with improved durability, rutting
resistance, fatigue resistance and anti-cracking properties. High-content CR binders present
higher viscosities than a PMB, enabling for higher binder contents in the mix.
As an example, a discontinuous-graded asphalt-mix was produced with a high viscosity

binder. The binder was manufactured with a high content of crumb rubber, which was high-
shear mixed with a low penetration bitumen. Data for aggregate size distribution and binder
properties are provided below.

% passing
100
90
80
70
60
50
40
30
20
10
0
40 25 12,5 8 5 2,5 1,25 0,63 0,32 0,16 0,08
Sieve, UNE mm
Sieving gradation allowance range
The mixing design was selected according to the Marshall method with a compaction
of 50bps. 7% was selected as the optimum binder content (weight percent of binder by
aggregate).
Table 6. Properties of high-viscosity asphalt rubber binder.
BINDER Standard BMAVC-1

PENETRATION P25℃ (1/10 mm) 15-30 28
R&B (℃) NLT-125 75 81
PI +3.0
FRAASS T (℃) -4 -16
STORAGE STABILITY
R&B difference 5 3
Penetration difference (25℃) 20 5
ELASTIC RECOVERY TORSION 25℃ 10% 43%
VISCOSITY 135℃ (cP) 10500
VISCOSITY 150℃ (cP) 4800
VISCOSITY 175℃ (cP) 2000 2100
RHEOMETRY
AASHTO TP5
Complex Modulus, G* (Pa) 60℃ 23000
Phase angle, δ,
(º) 60℃ 69
Table 7. High-viscosity asphalt rubber mix Marshall results.
Mixing Temperature 185ºC

Compaction temperature 175ºC
Binder content BMAVC-1 b/a % 7.0 8.0
Density g/cm 3
2.264 2.261
%VMA % 21.5 22.3
% Air voids % 7.1 6.1
Stability (EP) kN 10.12 8.80
Flow (DP) mm. 2.7 1.9
Stability/Flow kN/mm 3.8 4.5
4. CONCLUDING REMARKS
- The production of ground tyre rubber binders through a wet process by using a high
digestion process and co-additives is a preferred alternative of recycling tyres into the
pavements. The benefits of manufacturing at the production site a stable product along
the factory storage, transportation and plant storage before mixing represent a cost-
effective alternative.
- By incorporating the crumb rubber with a high digestion process, part of the rubber can
be effectively transferred to the continuous phase and can be partially swollen by the
lights fractions of bitumen. The cohesion attained in the binder by the wet process is
much higher than through dry process or low energy blending.
- The stability of the binder against settling and heterogeneity must be evaluated through
a different test than that prescribed for PMB’s. A dissolution and filtration of portions
at different heights of a binder stored at high temperatures is proposed as a best
practice for detection and quantification of possible settling phenomenon. Rheological
measurements can in most cases be a reliable technique for this purpose, as far as
the test conditions are thoughtfully selected. With the presence of certain polymeric
additives this technique may lead to misinterpretation.
- Whether the crumb rubber is incorporated to the bitumen through a high digestion
process or a semi-dry (right before the aggregate mixing) process, the material must
be carefully treated at high temperatures, due to the tendency of natural rubber to stick
together and to hot surfaces when a limit temperature is surpassed.
- A wide range of binders, from low to high content, can be manufactured from GTR,
provided the stability of the binder is assured, for a particular application: from dense-
graded to porous asphalt, as well as special applications as high-content binder anti-
cracking layers.
5. BIBLIOGRAPHY
Accorsi JV. “El impacto de la morfología y la dispersión del negro de carbono sobre la
resistencia a la intemperie del polietileno”. Simposium Internacional Wire and Cable,
Atlantic City (1999)
Anderson J, Pampulim V, Saim R, Sousa JB. “ Asphalt rubber laboratory properties related
to type and process technology of crumb rubber”. Proceedings of Asphalt Rubber 2000.
361-370Portugal. 2000
Bahia HU, Davies R. “Effect of crumb rubber modifiers (CRM) on performance-related

properties of asphalt binders”. J Assoc Asphalt Paving Technol, 63, 414–420 (1994)
Bahia HU, Zhai H. “Laboratory asphalt stability test (LAST): a new method for the storage
stability test of modified asphalts”. Reprints of the division of fuel chemistry of the ACS,
v.43, 4, 1041-1044 (1998)
Billiter TC, Chung JS, Davison RR, Glover CJ, Bullin In.”Investigation of the curing variables
of asphalt-rubber binder”. Pet Sci Technol, 15, p445 (1997)
Chaala A, C. Roy, A. Ait-Kadi. « Rheological properties of bitumen modified with pyrolytic

carbon black”. Fuel, 75, 13, 1575-1583 (1996)
Gallego Medina Juan. “Mezclas bituminosas modificadas por adición de polvo de neumáticos.
Estudio de la influencia de las características del caucho y del proceso de fabricación y
puesta en obra de la mezcla”. PhD Thesis. Madrid, 1999.
Hesp SA and Woodhams RT. “Stabilization mechanisms in polyolefin-asphalt emulsions.

Polymer modified asphalt binders”, ASTM STP 1108, Wardlaw KR and Shuler S Eds.
ASTM, Philadelphia (1992)
Lesueur, D.; Gérard, J. F.; Claudy, P.; Létoffé, J. M.; Martin, D.; Planche, J. P. “Polymer
Modified Asphalts as Viscoelastic Emulsions”. Journal of Rheology, 42, 1059-1074 (1998)
“Manual de empleo de caucho de NFU en Mezclas Bituminosas”. CEDEX. Mayo de 2007.
Navarro FJ, P. Partal, F. Martínez-Boza, C. Gallegos. “Thermo-rheological behaviour and

storage stability of ground tire rubber-modified bitumens”. Fuel 83, 2041–2049 (2004)
OC 21-2007 de la Dirección General de Carreteras del Ministerio de Fomento Español
Páez Dueñas A, P. Bernaldo de Quirós Miravent, R. Miró Recasens. “Betún modificado con
polvo de neumáticos”. VI Congreso nacional de firmes, León (2004)
Panagiotis Frantzis. “Crumb rubber-bitumen interactions: diffusion of bitumen into rubber”. J

Mat Civil Eng, July/August, p387 (2004)
Peña JL. “Desarrollo de una gama de productos a base de betún modificado con polvo de
neumáticos. Casos practices”. Panorámica actual de las mezclas bituminosas. Asefma, un
nuevo enfoque. Madrid (2005)
Pérez-Lepe A, F. J. Martínez-Boza, P. Attané, C. Gallegos. “Destabilization mechanism of

polyethylene-modified bitumen”. Journal of Applied Polymer Science, 100, 1, 260-267
(2006)
Pérez Lepe, Antonio; Páez Dueñas, Antonio. “Medida de la estabilidad del betún polvo de
neumáticos por solubilidad en disolventes orgánicos”. Carreteras. Nº. 151, 2007 , pags.
45-52
Soto JA, M. M. Colás. “Betunes modificados estables al almacenamiento mediante empleo

de polvo de caucho reactivado procedente de NFU’s, para uso en carreteras”. Panorámica
actual de las mezclas bituminosas. Asefma, un nuevo enfoque. Madrid (2005)
See H, Jiang P, Phan-Thien N. “Concentration dependence of the linear viscoelastic properties

of particle suspensions”. Rheol Acta, 38:131–137 (2000)
Ssu-Wei Loh, Sohee Kim, H.U. Bahia. “Characterization of simple and complex crumb rubber
modified binders”. Report WI/SPR-07-01. University of Wisconsin (2000)
Youtcheff J, A. Shenoy. “Evaluation of the Particulate Additive Test”. Publication FHWA-

HRT-04-110 (2004)
Comparison Between Various Bituminous

Binders Modified With Crumb Tyre Rubber
D. Lo Presti* - N. Memon** - G.D. Airey** - J.R.A. Grenfell**
* Dipartimento di Ingegneria delle Infrastrutture Viarie, Università degli studi di Palermo,

Palermo, IT
davide.lopresti@unipa.it
** Nottingham Transportation Engineering Centre, University of Nottingham, Nottingham,

UK
naeem.memon@nottingham.ac.uk
gordon.airey@nottingham.ac.uk
ABSTRACT. The aim of this paper is to show how the rheology, the performance and the
storage stability of different bituminous binders, obtained from two bitumens with different
asphaltenes content, change their properties when they are modified with recycled tyre crumb
rubber. Physical, chemical and performance characterisation followed by rheological and
storage stability analyses have been undertaken as the basis for the comparison. The results
of the investigation indicate that bitumen with lower asphaltenes content is more influenced
by the modification with the rubber. However, the modification of the bitumen with the higher
asphaltenes content gives a binder with better performance and storage stability. Moreover,
results shows that the addition of a certain amount of oil extender to the bitumens reduces the
effect of the modification and has no positive effect on storage stability.
KEYWORDS: crumb rubber, modified bitumen, performance grade, DMA, PDA

1. Introduction
Historically, recycled rubber from tyre waste has been used in asphalt mixtures using two
approaches, namely the wet process and the dry process. In the dry process, the recycled
rubber is mixed with the aggregates before introducing the binder to the mixture, the wet
process, instead, involves blending the bitumen with rubber particles (known as crumb rubber
modifier) at an elevated temperature (170 to 200°C) eventually in the presence of an oil
extender. The obtained product is a rubberised bitumen, which is then mixed with aggregate
to form a mixture. The interaction of bitumen with rubber in the wet process is known to
be affected by the blending temperature, the duration of blending, the type and amount of
mechanical blending energy, the size and texture of the rubber particles, and the aromatic
component of the bitumen. The absorption of aromatic oils from the bitumen into the rubber’s
polymer chains causes the rubber to swell and soften. The type and amount of aromatic oil in
the bitumen also plays a major role in determining the compatibility of bitumen-rubber blends
(Chehovits et al, 1982; Oliver, 1982).
This paper compares the physical, rheological and performance properties of a series of
tyre rubber-modified asphalt binders produced using a combination of different base bitumens,
an oil extender and crumb rubber. The bitumen testing varies from standard empirical tests,
such as the penetration and softening point tests, to more detailed rheological tests using a
dynamic shear rheometer. In addition to the rheological properties, the storage stability of the
tyre rubber-modified asphalt binders has been assessed.
2. Background
2.1 Bitumen composition and structure
Bitumen has a complex chemical composition, consisting mainly of hydrocarbons,

with a high carbon content of between 80% and 90% along with small percentages of
analogous hetrocyclic species and functional groups of nitrogen, sulphur and oxygen atoms.
Traditionally, bitumen is regarded as a colloidal system consisting of a suspension of high
molecular weight asphaltene micelles dispersed in a lower molecular weight oily medium
identified as maltenes (Lu et al, 1999). By means of absorption chromatography, bitumen
can be separated into four functional groups with related properties with regard to chemical
reactivity and rheological properties despite its complexity. These generic components are
generally referred to as: Asphaltenes, Resinous components (polar aromatics), Aromatics
(non-polar naphtene aromatics), and Saturates (Read and Whiteoak, 2003). In the presence
of sufficient quantities of resins and aromatics of adequate solvating power, the asphaltenes
are fully peptized, at every temperature, and the resulting micelles have good mobility within
bitumen. Peptization is provided by a shell of resins that surrounds the asphaltene core with
the thickness of this resin layer being temperature dependent (Figure 1a). At low temperature
and in the case of high asphaltene content the bitumen can show a compact structure as
shown in Figures 1b and 1c. On the other hand, insufficient quantities of resins and aromatics
(insufficient solvating power) may lead to an association of several micelles, which may be
simply expressed using the following thermal dependent equilibrium:
molecules <=> micelles <=> clusters
The system is therefore formed by a liquid part (resins dissolved in a maltene matrix), in
equilibrium with a solid part (resins peptizing the asphaltenes) and it is, of course, temperature
dependent (Yen, 1990).
Figure 1.Colloidal structures of asphalt
2.2 Rheology and Rheometry
Rheology is literally “flow science”. Rheological experiments do not merely reveal

information about the flow behaviour of liquids, but also the deformation behaviour of solids.
The connection here is that large deformations produced by shear forces cause many materials
to flow. All forms of shear behaviour, which can be described rheologically in a scientific
way, can be viewed as lying in between two extremes: the flow of idealviscous liquids on
one hand and the deformation of idealelastic solids on the other (Mezger, 2002). Bitumen is
a thermoplastic material that, under most pavement operating conditions, shows a behaviour
which is between these extremes and it is based on the combination of both the viscous and
the elastic portion and therefore is called viscoelastic
Rheometry is the measuring technology used to determine rheological properties of a

material. It includes measuring systems, instruments, tests and analysis methods. At present
the most common rheometry of bitumens includes Dynamic Shear Rheometer (DSR), which
performs a means of dynamic/cyclic analysis applying pure shear stresses in continuous/
oscillatory regimes, and Bending Beam Rheometer (BBR) which operates only at really
low temperatures at a constant rate, inducing normal stresses in a bitumen beam. The most
recent DSRs can even work at below zero degrees; this makes it probably the most suitable
equipment for a complete rheological, viscoelastic characterisation of bituminous binders.
2.3 Definition of terms and analysis methods
The principal viscoelastic parameters that are obtained from rheometrical measurements,
mainly with the DSR, are the complex shear modulus, G*, and the phase angle, δ. G* is
defined as the ratio of maximum (shear) stress to maximum strain and provides a measure
of the total resistance to deformation when the bitumen is subjected to shear loading. It
contains elastic and viscous components which are designated as the storage modulus, G’,
and loss modulus, G’’, respectively. These two components are related to the complex (shear)
modulus and to each other through the phase (or loss) angle δ which is the phase, or time, lag
between the applied shear stress and shear strain responses during a test. In order to perform
a complete rheological characterisation it is therefore fundamental to firstly collect the
rheological parameters with oscillatory tests with variable frequency and constant small strain
level within the Linear Visco-Elastic response (LVE) over a range of temperatures and then
elaborate the results through analytical techniques. One of the primary analytical techniques
used in analysing dynamic/cyclic mechanical data involves the construction of master curves,
which represents the viscoelastic behaviour of a bituminous binder at a given temperature for
a large range of frequencies. The principle that is used to relate the equivalency between time
and temperature and thereby produce the master curve is the Time Temperature Superposition
Principle (TTSP). Once TTSP is applied another interesting analysis consists in correlating
the viscoelastic parameters with temperatures at constant frequency. This kind of elaboration
provides the Isochronal plots. Another fundamental curve obtainable from the frequency
sweep data is the Black diagram. A Black diagram is a graph of the magnitude (norm) of the
complex modulus, G*, versus the phase angle, δ, obtained from a dynamic/cyclic test. The
frequency and the temperature are therefore eliminated from the plot, which allows all the
dynamic data to be presented in one plot without the need to perform TTSP manipulations of
the raw data. For this reason the Black diagram is well known even as the finger print of the
binder (Airey 2002).
3. Experimental Programme
3.1 Materials
Two different base bitumens, a PG 64-16 (bitumen A) and a PG 64-22 (bitumen B), were
used with and without an oil extender (flux) to produce four different tyre rubber-modified
asphalt binders (TR-MAB). The TR-MABs were blended using a Silverson high shear
laboratory mill to mix the base binders with fine crumb rubber obtained from used tyres.
Physical, chemical and performance properties of the base bitumens (with and without flux)
are reported in Section 4. Table 1 reports the characteristics of the crumb rubber as received
by the provider. Table 2 and Figure 2 show the results of the sieving procedure and the portion
of rubber used in this study denominated as “fine”.
Table 1. Properties of the crumb rubber
Origin : car tyre rubber reduced to crumbs by mechanical treatment

Particle shape : irregular
Fibre content : 0.5 %
Steel content : 0.1 %
100%
90%
80%
discarded
70%
total passing
60%
50%
40%
30%
20% used
10%
0%
0.01 0.1 0.5 1 10
sieve size (mm)
Figure 2. Gradation curve of the crumb rubber
Table 2. Sieving of the crumb rubber
Mass Retained retained on sieve Total passing

sieve size
cumulative individual
mm g g % %
1.0 0 0 0 100
discarded
0.6 44.1 44.1 37.1 62.9
0.5 70.9 26.8 22.5 40.4
0.3 109.5 38.6 32.4 8.0
0.25 111.6 2.1 1.8 6.2
0.212 113.5 1.9 1.6 4.6
0.15 115.4 1.9 1.6 3.0 used
0.125 116.1 0.7 0.6 2.4
0.075 118.3 2.2 1.8 0.6
0.063 118.8 0.5 0.4 0.2
119 0.2 0.2
Total mass 119 119
The following binders have therefore been incorporated into this study:
- bitumen A: PG 64-16 straight run bitumen

- bitumen B: PG 64-22 straight run bitumen
- binder Af: blend of 92.5% of PG 64-16 straight run bitumen and 7.5% of flux
- binder Bf: blend of 92.5% of PG 64-22 straight run bitumen and 7.5% of flux
- TR-MAB A: tyre rubber-modified asphalt binder produced as 85% of bitumen A and
15% of crumb rubber
- TR-MAB B: tyre rubber-modified asphalt binder produced as 85% of bitumen B and
15% of crumb rubber
- TR-MAB Af: tyre rubber-modified asphalt binder produced as 85% of base binder Af
and 15% of crumb rubber
- TR-MAB Bf: tyre rubber-modified asphalt binder produced as 85% of base binder Bf
and 15% of crumb rubber
All the binders were tested in their unaged (virgin) condition and after being subjected to
two laboratories ageing procedures:
- Short-term laboratory ageing using the Rolling Thin Film Oven test (RTFOT)
according to AASHTO T 240.
- Long-term laboratory ageing using a Pressurised Ageing Vessel (PAV) according to

AASHTO R 28.
Due to practical concerns (more details in Section 4), the short term ageing of the modified
bitumens was performed by using the Thin Film Oven test(TFOT)according to AASHTO T 240.
3.2 TR-MAB blending protocol
The blending of rubber and bituminous binders was carried out using the following
protocol:
1. The required amount of bitumen was heated at 180°C in the oven and then transferred
to a hot plate at the same temperature.
2. High shear mixing up to 2000 rpm was applied for the first 10 minutes while the flux
(only for binders Af and Bf and TR-MAB Af and Bf) and then only the fine portion of
the rubber was feed into the bitumen.
3. Time was allowed for the temperature to stabilise at 180°C.
4. Once the temperature reached 180°C, blending time was noted and mixing undertaken
at 1000 rpm for one hour.
Tables 3 and 4 summarise the mixing parameters used in the blending procedure.
Table 3. Blending protocol for TR-MAB A and TR-MAB B
mass of bitumen mass of flux mass of rubber total mixing mixing mixing
(85%) (0%) rubber (15%) size weight time speed temp.
g g g mm g min rpm °C
1700 0 300 0-0.5 2000 60 1000 180
Table 4. Blending protocol for TR-MAB Af and TR-MAB Bf
mass of bitumen mass of flux mass of rubber total mixing mixing mixing
(78.65%) (6.35%) rubber (15%) size weight time speed temp.
g g g mm g min rpm °C
1572.5 127.5 300 0-0.5 2000 60 1000 180
3.3 Physical and chemical characterization
The first part of the programme consisted of characterising the bitumens and base binders
in terms of their physical and chemical properties and to quantify the effect of the oil extender.
The physical classification consisted of the following tests: penetration (ASTM D5), softening
point (ASTM D36-95), Fraass breaking point (IP 80), ductility (10ºC, ASTM D113) and
rotational viscosity at 100 and 160ºC (ASTM T316-04). In order to have an idea of the
chemical composition of the binders, the asphaltenes content according to ASTM D6560 was
measured.
The second part of the programme consisted of comparing the different binders in terms of
their rheological characteristics using a dynamic shear rheometer (DSR). Finally, in the third
part of the programme, the storage stability of the binders was evaluated using softening point
(ASTM D36-95) and polymer dispersion of the elastomer in the binders.
3.4 Performance grading procedure
In order to classify the binders by their performance properties, all of them were subjected
to the Superpave grading procedure according to AASHTO R 29-02. The performance
grading is undertaken by testing the binder at three different ageing conditions: unaged, short-
term aged (RTFO aged) and long-term aged (PAV aged). As shown in Tables 4 and 5, the
high performance grade (high PG) is determined based on tests performed on the unaged and
RTFO aged binder.
Table 4. Overview of the procedure to find the high PG
High PG Value
Unaged binder:
- the binder is tested to determine the flash point in °C (AASHTO T 48),
- viscosity at 135°C (AASHTO T316),
- the shear modulus (G*) and phase angle (AASHTO T 315)
RTFO aged binder:

- the binder is aged in the Rolling Thin-Film Oven (AASHTO T 240)
- and the residue is tested to determine the mass loss (AASHTO T 240),
- the shear modulus (G*) and phase angle (AASHTO T 315)
Completing the sequence, the low performance grade (low PG) is determined by tests
on the PAV aged binder. The low PG of all the binders has been established using Table 1 of
AASHTO M 320-05.
Table 5. Overview of the procedure to find the low PG
Low PG Value
PAV aged binder:
- The residue from the rolling thin-film oven is aged in the Pressurised Aging
Vessel (AASHTO R 28) and this residue is tested to determine:
- the shear modulus (G*) and phase angle (AASHTO T 315),
- the creep stiffness (S) and slope, m of the log creep stiffness versus log time
relationship at different instants of load (AASHTO T 313)*,
- eventually the failure stress in Direct Tension (AASHTO T 314),
- and in post-processing the Critical Cracking Temperature (AASHTO PP 42)**
* last step for assigning the low grade using Table 1 of the specification
** last step for assigning the low grade using Table 2 of the specification
3.5 Rheological analysis
In order to obtain a rheological characterisation of all the binders, a dynamic mechanical

analysis (DMA) over a wide range of temperatures and frequency sweeps was performed
using a Bohlin GEMINI DSR. The tests were performed under the following conditions:
- Mode of loading: controlled-strain

- Temperatures: 0 to 80ºC at 5ºC intervals
- Frequencies: 0.10, 0.16, 0.25, 0.40, 0.63, 1, 1.6, 2.5, 4, 6.3 and 10 Hz
- Spindle geometries: 8 mm φ and 2 mm gap (0-45ºC),
25 mm φ and 1 mm gap (35-80ºC)
- Strain amplitude: 0.5% (within LVE response dependent on G*)
For each test, samples were prepared by means of a hot pour method, based on Alternative
1 of the AASHTO TP5 Standard. The gap between the upper and lower plates of the DSR
was set to a height of 50 µm plus the required testing gap at the mid-point of the testing
temperature range. Once the gap has been set, a sufficient amount of hot bitumen (160ºC) was
poured on to the lower plate of the DSR to ensure a slight excess of material appropriate to
the chosen testing geometry. The upper plate of the DSR was then gradually lowered to the
required nominal testing gap. The bitumen that was squeezed out between the plates was then
trimmed flush to the edge using a hot blade. Finally, the gap was closed until there was a slight
bulge around the circumference of the testing geometry.
The rheological properties of the binders were measured in terms of their complex
(shear) modulus, G*; and phase angle (viscoelastic balance of rheological behavior), δ. Once
measured, the data was used together with the Time Temperature Superposition Principle and
shift factors to produce master curves at 25°C and isochronal plots. Master curves, isochronal
plots and black diagrams have been used as the basis of all the rheological analysis in this
paper.
4. Results and discussions
4.1 Base bitumens
Two base bitumens were available and both have been characterized in terms of their
physical, chemical and performance properties. The two bitumens have similar physical
properties, but differ chemically in terms of their asphaltenes content (Table 6). This
difference in chemical composition potentially has an effect on the modification process when
they are mixed with polymers. In general a smaller percentage of asphaltenes means a lower
Colloidal Index value that leads to a system with a higher degree of solvency for the polymers
(Airey, 2003).
Table 6. Properties of base bitumen A and base bitumen B
bitumen A bitumen B
Penetration (ASTM D5) 42 dmm 54 dmm
Softening Point (ASTM D36-95) 51ºC 52.2 ºC
Fraass breaking point (IP 80) 0 ºC -2 ºC
Ductility (ASTM D113) 1000 mm 1000 mm
Rotational Viscosity @ 100ºC (ASTM T316-04) 3.86 Pa.s 5.13 Pa.s
Asphaltenes content 3.4% 16.7%
The rheological properties of the two base bitumens in terms of master curves, black
diagrams and isochronal plots are shown in Figure 3.
100
1.E+09
90
1.E+08
bitumen A 80
1.E+07
binder B bitumen A
Complex Modulus (Pa)
70
bitumen B
Phase Angle
1.E+06
60
1.E+05
50
1.E+04
40
1.E+03
30
1.E+02
20
1.E+01 10
1.E-06 1.E-05 1.E-04 1.E-03 1.E-02 1.E-01 1.E+00 1.E+01 1.E+02 1.E+03 1.E+04 1.E+05 1.E-07 1.E-06 1.E-05 1.E-04 1.E-03 1.E-02 1.E-01 1.E+00 1.E+01 1.E+02 1.E+03 1.E+04 1.E+05
Frequency (Hz) Frequency (Hz)
Figure 3a. M aster curves of complex modulus and phase angle at 25 ºC for the
base bitumens
1.E+09 1.00E+08 100
90
1.E+08 1.00E+07
80
1.E+07 1.00E+06
bitumen A 70
Complex modulus
1.E+06 1.00E+05
Phase angle
60
bitumen B
1.E+05 1.00E+04 50
40
1.E+04 1.00E+03
30
1.E+03 1.00E+02
20
1.E+02 1.00E+01 bitumen A bitumen B

10
1.E+01 1.00E+00 0
10 20 30 40 50 60 70 80 90 100 0 10 20 30 40 50 60 70 80
Phase angle (º) Temperature (ºC)
Figure 3b. Black diagram for the Figure 3c. Isochronal plots of
base bitumens complex modulus and phase angle at
0.4 Hz for the base bitumens
Although the empirical rheological properties of penetration and softening point for the
two base bitumens are similar, their detailed rheological properties shown in Figures 3a-c
indicate differences in their more detailed rheological properties. Bitumen “A” has a higher
complex modulus at high frequencies although bitumen “B” tends to be less temperature
and frequency susceptible. Both bitumens are thermorheologically simple as shown by their
smooth curves in both the master curves and Black diagram (Mezger, 2002). The higher
asphaltenes content of bitumen “B” results in a shifting of the viscoelastic balance with a
reduction in phase angles (decreasing viscous response/increasing elastic response) compared
to the low asphaltenes content bitumen “A”.
4.2 Base binders
As reported by other researchers, it is common practice to add an oil extender into the
blend to improve the solvating power of the base bitumen and increase the digestion of the
crumb rubber within the bitumen (Potgieter & Coetsee, 2003). A blend of 92.5% base bitumen
and 7.5% of oil extender (flux) was made for each bitumen obtaining “base binder Af” and
“base binder Bf”.
Table 7. Physical properties of base binder Af and base binder Bf
binder Af binder Bf
Penetration (ASTM D5) 136 dmm 157 dmm
Softening Point (ASTM D36-95) 39.8ºC 40.3ºC
Fraass breaking point (IP 80) -14ºC -17ºC
Ductility (ASTM D113) 1000 mm 1000 mm
Asphaltenes content 3.6% 16.7%
As it is noticeable from Tables 6 and 7 that the addition of the flux does not affect the
asphaltenes content. Analysing the results, it is evident that the oil extender affects the base
binder by softening it and thereby enhancing its low temperatures properties. Tables 6 and
7 show how physical properties are influenced by the flux with an increase in the cracking
resistance of both binders at low temperatures, as shown by the lower Fraass breaking point
temperature, and a softening of the binders at high temperatures, as shown by a decrease
in softening point and reduced viscosity. The modification in the performance properties
have been investigated and the results in Tables 8 and 9 show that in both cases the base
bitumen changes its performance grade moving towards lower values but maintains the
same performance temperature range. It should be noted that this shifting towards lower
performance grade values, due to the oil extender, has a greater effect on the bitumen with the
lower asphaltenes content (bitumen A).
The softening of the base bitumens throught the addition of the flux is also seen in the
detailed rheological analysis for both base bitumens and binders in Figures 4a-c and 5a-c. The
results show a uniform shifting of the complex modulus towards lower values and an increase
in the viscous behaviour at high frequencies and low temperatures (higher values of phase
angle). The Black diagrams (Figures 4c and 5c) show that even if the base binder contains
flux, it still has the same black curve as that of the base bitumen, with both binders being
considered to be thermo-rheologically simple (Mezger, 2002). Considering that elastomers
mainly modify the bitumen by increasing its performance at low frequencies and high
temperatures (Airey, 2002), the results demonstrate how the addition of the flux is important
to improve the performance of the base binder at high frequencies and low temperatures.
Table 8. Performance grading of bitumen A and binder Af
Aging states AASHTO specifications bitumen A binder Af

Rotational Viscosity (Pa.s) @135°C/ 0.40 @135°C/ 0.20
Unaged binder
G*/sin(delta) (kPa) > 1.00 @64°C/ 1.41 @52°C / 1.28
RTFO
G*/sin(delta) (kPa) > 2.20 @64°C/ 2.90 @52°C / 3.51
aged residue
G* x sin(delta) (kPa) < 5000 @28°C/ 4790 @16°C / 4970

RTFO+PAV
Stiffness@ (MPa) < 300 @-6°C / 218 @-18°C/ 275
aged residue
m-value has to be > 0.3 @-6°C/ 0.36 @-18°C/ 0.33
Flash point (°C) has to be > 230 > 300 > 300
Mass loss after RTFO ageing has to be< 1.0% 0.15% 0.09%
Performance grade PG 64 - 16 PG 52 - 28
Performance temperature range 80°C 80°C
1.E+09 100
bitumen A 90
1.E+08
binder Af 80
1.E+07
bitumen A
70
binder Af
Phase Angle
1.E+06
60
1.E+05
50
1.E+04
40
1.E+03 30
1.E+02 20
10
1.E+01 1.E-07 1.E-06 1.E-05 1.E-04 1.E-03 1.E-02 1.E-01 1.E+00 1.E+01 1.E+02 1.E+03 1.E+04 1.E+05
1.E-06 1.E-05 1.E-04 1.E-03 1.E-02 1.E-01 1.E+00 1.E+01 1.E+02 1.E+03 1.E+04 1.E+05
Figure 4a. Master curves of complex modulus and phase angle at 25 ºC for bitumen A and
binder Af
1.E+09 1.00E+08 100
90
1.E+08 1.00E+07
80
1.E+07 1.00E+06
Complex modulus
bitumen A 70
Phase angle
1.E+06
1.00E+05
binder Af 60
1.E+05
1.00E+04 50
1.E+04 40
1.00E+03
30
1.E+03
1.00E+02
bitumen A binder Af
20
1.E+02
1.00E+01
10
1.E+01
1.00E+00 0
10 20 30 40 50 60 70 80 90 100
0 10 20 30 40 50 60 70 80
Figure 4b. Black diagram for bitumen A Figure 4c. Isochronal plots of complex
and binder Af modulus and phase angle at 0.4 Hz for
bitumen A and binder Af
Table 9. Performance grading of “bitumen B” and “binder Bf”
Aging states AASHTO specifications bitumen B binder Bf

Rotational Viscosity (Pa.s) @135°C/ 0.45 @135°C/ 0.23
Unaged binder
G*/sin(delta) (kPa) > 1.00 @70°C/ 1.05 @58°C / 1.47
RTFO
G*/sin(delta) (kPa) > 2.20 @64°C/ 4.56 @58°C / 3.05
aged residue
G* x sin(delta) (kPa) < 5000 @22°C/ 5317 @16°C / 4180

RTFO+PAV
aged residue Stiffness@ (MPa) < 300 @-12°C / 225 @-18°C/ 130
m-value has to be > 0.3 @-12°C/ 0.34 @-18°C/ 0.38
Mass loss after RTFO ageing has to be< 1.0% 0.07% 0.01%
1.E+09 100
90
1.E+08
bitumen B 80
1.E+07
binder Bf 70
bitumen B
Phase Angle
1.E+06
60
1.E+05 binder Bf
50
1.E+04
40
1.E+03 30
1.E+02 20
10
1.E+01
1.E-07 1.E-06 1.E-05 1.E-04 1.E-03 1.E-02 1.E-01 1.E+00 1.E+01 1.E+02 1.E+03 1.E+04 1.E+05
1.E-06 1.E-05 1.E-04 1.E-03 1.E-02 1.E-01 1.E+00 1.E+01 1.E+02 1.E+03 1.E+04 1.E+05
Figure 5a. Master curves of complex modulus and phase angle at 25 ºC for bitumen B and
binder Bf
1.E+09 1.00E+08 100
90
1.E+08 1.00E+07
80
1.E+07 1.00E+06
bitumen B
70
Complex modulus
1.00E+05
Phase angle
1.E+06 60
binder Bf 1.00E+04 50
1.E+05
40
1.00E+03
1.E+04
30
1.00E+02
1.E+03
bitumen B binder Bf
20
1.E+02
1.00E+01
S i 3 S i 2 10
1.00E+00 0
1.E+01
0 10 20 30 40 50 60 70 80
10 20 30 40 50 60 70 80 90 100
Figure 5b. Black diagram for bitumen B Figure 5c. Isochronal plots of complex
and binder Bf modulus and phase angle at 0.4 Hz for
bitumen B and binder Bf
4.3 Modified bitumens
Four laboratory blended tyre rubber-modified asphalt binders (TR-MAB), produced using
the same kind and amount of rubber but different base binders, were tested and compared. The
comparison was made in terms of rotational viscosity (Table 10), performance properties and
rheological characteristics. Furthermore, to better understand the test results and properties, all
the modified binders were subjected to a storage stability test and then compared in terms of
softening point, polymer dispersion and rheological properties of their top and bottom sections
after a period of hot storage.
Table 10. Rotational viscosity @ 100 and 160ºC
Rotational Viscosity @ 100 ºC @ 160 ºC

(ASTM T316-04) Pa.s
TR-MAB A 45.87 2.02
TR-MAB Af 19.85 1.46
TR-MAB B 49.44 2.40
TR-MAB Bf 24.18 2.01
4.3.1 Superpave Performance grade
The grading procedure for bitumens (AASHTO R 29-02) has been followed to find the
performance grade (PG) of all the modified bitumens. Short-term ageing using the RTFO
equipment was impossible to perform, firstly, due to the rubber particles not allowing the
binder to form a uniform film (coat) on the internal surface of the glass containers and,
secondly, due to “roll out” problems. The Thin Film Oven test (AASHTO T 240) was
therefore used to simulate the short-term ageing of the TR-MAB. The performance grade test
results in Tables 11 and 12 show three significant results. Firstly, all the TR-MABs have got
a better performance temperature range in comparison with the base bitumen. Secondly, TR-
MABs obtained without adding flux perform better at high temperature and show a wider
performance temperature range, but have a higher viscosity at operating temperatures (135°C).
Finally, comparing the results between the TR-MABs originated from the two different base
bitumens, it is possible to notice that after modification with crumb rubber, bitumen B is the
one which demonstrates the biggest improvements (Figure 6).
Table 11. Performance grading of “TR-MAB A” and “TR-MAB Af”
Aging states AASHTO specifications TR-MAB A TR-MAB Af

Unaged binder Rotational Viscosity (Pa.s) @135°C/ 4.41 @135°C/ 2.75
G*/sin(delta) (kPa) > 1.00 @76°C / 1.12 @76°C / 0.99
RTFO/TFO1 G*/sin(delta) (kPa) > 2.20 @76°C / 2.38 @64°C / 2.60

aged residue
RTFO/TFO1 G* x sin(delta) (kPa) < 5000 @22°C / 4670 @13°C / 4000

+PAV
aged residue Stiffness@ (MPa) < 300 @-12°C / 181 @-18°C / 143
m-value has to be > 0.3 @-12°C / 0.35 @-18°C / 0.38
1: Thin Film Oven test (AASHTO T 240) was used to simulate the short-term ageing of the TR-MAB
Mass loss after RTFO ageing has to be < 1.0% not possible not possible
Table 12. Performance grading of “TR-MAB B” and “TR-MAB Bf”
Aging states AASHTO specifications TR-MAB B TR-MAB Bf

Unaged binder Rotational Viscosity (Pa.s) @135°C/ 6.45 @135°C/ 3.83
G*/sin(delta) (kPa) > 1.00 @88°C / 1.24 @76°C / 1.83
RTFO/TFO1 G*/sin(delta) (kPa) > 2.20 @88°C / 2.60 @76°C / 2.96

aged residue
RTFO/TFO1 G* x sin(delta) (kPa) < 5000 @19°C / 4610 @10°C / 3720

+PAV Stiffness@ (MPa) < 300 @-18°C / 183 @-24°C/ 150
aged residue
m-value has to be > 0.3 @-18°C / 0.33 @-24°C/ 0.34
1: Thin Film Oven test (AASHTO T 240) was used to simulate the short-term ageing of the TR-MAB
Mass loss after RTFO ageing has to be < 1.0% not possible not possible
130
Performance temperature range (°C)
120 Higher PG >0

110
100 Lower PG < 0
90 116
80 98 110
70 80 86 92
60 80 86
50
40 88
76 76
30 64 64 58 64
52
20
10
0
-10 -16 -22 -22
-28 -28 -28 -28 -34
-20
-30
-40
B
A
Af
Bf
Bf
f
B
A
en
en
AB
er
er
AB
AB
AB
m
m
nd
nd
M
tu
tu
M
Bi
Bi
M
-
Bi
Bi
-
TR
-
TR
-
TR
TR
Figure 6. Overview of the performance properties of all the binders
4.3.2 Rheological analysis
The frequency dependence of complex modulus and phase angle for the modified bitumens
has been assessed in Figures 7a-b and 8a-b by producing rheological master curves at a
reference temperature of 25°C and black diagrams. Isochronal plots at low frequency (0.4Hz)
were obtained to show the temperature dependency (Figures 7c and 8c). A comparison
regarding the TR-MABs and the respective original base bitumens, shows how the modified
binders have a lower frequency and temperature susceptibility. The results indicate that this
is true for both kinds of modification, with or without oil extender, but with the softer base
binder (with flux) showing, on average, lower stiffness values for the TR-MAB binders. The
polymer network effect is clearly noticeable in Figures 7a,c and 8a,c which show unique phase
angle curves due to the typical effect of the elastomers that significantly increase the elastic
response of both modified binders especially at low frequencies and high temperatures (Airey,
2003). Moreover, the presence of a slight plateau on the complex modulus curves (Figures
7a,c and 8a,c) shows how the dominance of the polymer networks improves the stiffness of
the modified binders. An overall analysis of the graphs shown in Figures 7 and 8 demonstrates
that laboratory blended TR-MABs perform better rheologically, than their original base
bitumens, showing a lower temperature and frequency susceptibility. The graphs show the
effect of the oil extender on the TR-MABs as it does with the base binder as explained in
Section 4.2. Similar to the effect seen between the base bitumens and base binders, the
addition of the oil extender for the TR-MABs does not significantly change the shape of the
black diagrams as shown in Figure 7b and 8b.
100
bitumen A
1.E+09
90
1.E+08 CR-MB A
bitumen A 80 CR-MB Af
1.E+07
CR-MB A
CR-MB Af
70
Phase Angle
1.E+06
60
1.E+05
50
1.E+04
40
1.E+03
30
1.E+02
20
1.E+01
10
1.E-06 1.E-05 1.E-04 1.E-03 1.E-02 1.E-01 1.E+00 1.E+01 1.E+02 1.E+03 1.E+04 1.E+05
1.E-07 1.E-06 1.E-05 1.E-04 1.E-03 1.E-02 1.E-01 1.E+00 1.E+01 1.E+02 1.E+03 1.E+04 1.E+05
Frequency (Hz)
Frequency (Hz)
Figure 7a. Master curves of complex modulus and phase angle at 25 ºC for “TR-MAB A”,
“TR-MAB Af” and their base “bitumen A”
1.E+09 1.00E+08 100.00
90.00
1.E+08 1.00E+07
80.00
1.E+07
1.00E+06
Complex modulus
70.00
1.E+06 1.00E+05
Phase angle
bitumen A 60.00
1.E+05 CR-MB A 1.00E+04 50.00
CR-MB Af 40.00
1.E+04 1.00E+03
30.00
1.E+03 1.00E+02
bitumen B (G*) CR-MB A (G*) 20.00
1.E+02 1.00E+01 CR-MB Af (G*) CR-MB A (δ) 10.00
1.E+01 1.00E+00
CR-MB Af (δ) bitumen B (δ) 0.00
10 20 30 40 50 60 70 80 90 100
0 10 20 30 40 50 60 70 80
Figure 7b. Black diagram for “TR-MAB Figure 7c. Isochronal plots of complex
A”, “TR-MAB Af” and their base “bitumen modulus and phase angle at 0.4 Hz for
A” “TR-MAB A” and “TR-MAB Af”
100
1.E+09
bitumen B 90
CR-MB B
1.E+08
80
1.E+07 CR-MB Bf
70
Phase Angle
1.E+06
60
1.E+05
50
1.E+04
40
1.E+03
30
bitumen B
1.E+02
20 CR-MB Bf
1.E+01 10
CR-MB B
1.E-06 1.E-05 1.E-04 1.E-03 1.E-02 1.E-01 1.E+00 1.E+01 1.E+02 1.E+03 1.E+04 1.E+05 1.E-07 1.E-06 1.E-05 1.E-04 1.E-03 1.E-02 1.E-01 1.E+00 1.E+01 1.E+02 1.E+03 1.E+04 1.E+05
Frequency (Hz)
Frequency (Hz)
Figure 8a. Master curves of complex modulus and phase angle at 25 ºC for “TR-MAB B”,
“TR-MAB Bf” and their base “bitumen B”
1.E+09 1.00E+08 100
90
1.E+08 1.00E+07
80
1.E+07 1.00E+06
Complex modulus
70
1.E+06 bitumen B 1.00E+05

Phase angle
60
CR-MB B
1.E+05
CR-MB Bf 1.00E+04 50
40
1.E+04 1.00E+03
30
CR-MB Bf (G*) bitumen B (G*)

1.E+03 1.00E+02
20
1.E+02 1.00E+01 CR-MB B (G*) bitumen B (δ) 10

CR-MB B (δ) CR-MB Bf (δ)
1.E+01 1.00E+00 0
10 20 30 40 50 60 70 80 90 100 0 10 20 30 40 50 60 70 80
Figure 8b. Black diagram for “TR-MAB Figure 8c. Isochronal plots of G* and δ at
B”,“TR-MAB Bf” and “bitumen B” 0.4 Hz: “TR-MAB B”, “TR-MAB Bf”
1.E+09
TR-MAB A TR-MAB Af
1.E+08
TR-MAB B TR-MAB Bf
1.E+07
1.E+06
1.E+05
1.E+04
1.E+03
1.E+02
1.E+01
1.E-06 1.E-05 1.E-04 1.E-03 1.E-02 1.E-01 1.E+00 1.E+01 1.E+02 1.E+03 1.E+04 1.E+05
Frequency (Hz)
Figure 9. Master curves of complex modulus at 25 ºC for all the TR-MABs
90
TR-MAB A TR-MAB Af
80
TR-MAB B TR-MAB Bf
70
60
Phase Angle
50
40
30
20
10
1.E-07 1.E-06 1.E-05 1.E-04 1.E-03 1.E-02 1.E-01 1.E+00 1.E+01 1.E+02 1.E+03 1.E+04 1.E+05
Frequency (Hz)
Figure 10. Master curves of phase angle at 25 ºC for all the TR-MABs
Further information comes from the comparison between the rheological properties of
all the TR-MABs (Figures 9 to 12). Comparing the TR-MABs originated from bitumen B
with the correspondent TR-MABs created from bitumen A. It is possible to notice that the
modified binders obtained using the higher asphaltenes content base bitumen (bitumen B) are
on average stiffer (higher G* values) and more elastic (lower phase angles) over the range
of considered frequencies and temperatures (Figures 9 and 10). However, the TR-MABs
originated from bitumen A demonstrate a greater degree of modification with the rubber. At
low frequencies and high temperatures, they show slight plateaus on the complex modulus
curves and isochronal plots (Figures 9 and 12) and have much lower values of phase angle
(Figure 10). These are typical effects seen for elastomers which tend to increase the elastic
response of the modified binders (Airey, 2003).
1.E+09
1.E+08
1.E+07
1.E+06
TR-MAB A TR-MAB Af
1.E+05
TR-MAB B TR-MAB Bf
1.E+04
1.E+03
1.E+02
1.E+01
10 20 30 40 50 60 70 80 90
Phase angle (º)
Figure 11. Black diagrams for all the TR-MABs
1.00E+08 90.00
1.00E+07 80.00
70.00
1.00E+06
l modulus
60.00
d l
Phase angle
1.00E+05
50.00
1.00E+04
Complex
40.00
1.00E+03
30.00
C
1.00E+02
20.00
TR-MAB A (G**) TR-MAB Af (G*) T
TR-MAB B (G*)
1.00E+01 TR-MAB Bf (G
G*) TR-MAB A (δ) T
TR-MAB Af (δ) 10.00
TR-MAB B (δ)) TR-MAB Bf (δ)

1.00E+00 0.00
0.0 10.0 20.0 30.0 40.0 50.0 60.0 70.0 80
0.0
Temperatu
ure (ºC)
Figure 12. Isochronal plots of complex modulus and phase angle at 0.4 Hz for all the TR-
MABs
4.3.3 Storage stability analysis
Another fundamental requirement that has been evaluated for the laboratory blended
TR-MABs was their storage stability. In this analysis all the modified bitumens have been
subjected to a hot storage test based on BS EN 13399, but modified due to the fast phase
separation of the rubber. The procedure consisted of filling toothpaste tubes (three for each
binder) with pre-heated modified binder (180°C) and then, once the tubes had been closed,
placing them in a vertical orientation into an oven at 180°C. After two different periods (3
and 5 hours), one tube for each binder was taken out from the oven and was cooled down to
ambient (room) temperature before placing into a freezer at -20°C. After that, the samples
were cut into thirds and the top and bottom sections were saved to be further analysed.
In order to have an easily determinable parameter which could give a value of the storage
stability, the softening points of both the modified binders for the top and bottom sections at
three different periods of storage were measured (Tables 13 and 14). Softening point results
show that the phase separation of the rubber starts to be relevant after three hours.
Table 13. Softening points of TR-MAB A and TR-MAB Af before and after hot storage
TR-MAB A TR-MAB Af
Softening point
top bottom top bottom
(ASTM D36-95)
°C
before hot storage 60.5 49.8
after 1h of hot storage 59.0 61.4 46.0 51.6
Table 14. Softening points of TR-MAB B and TR-MAB Bf before and after hot storage
TR-MAB B TR-MAB Bf
Softening point
top bottom top bottom
(ASTM D36-95)
°C
before hot storage 71.0 60.5
As found in previous research (Lu & Isacsson, 1997) the use of only the softening point
to evaluate the storage stability of modified binders may be inadequate. Therefore, to have
more information and more accuracy, a polymer dispersion analysis (PDA) and a rheological
characterisation of the top and bottom sections of the modified binders, after three and
five hours of hot storage, have been performed. PDA was performed using fluorescence
microscopy which provided the images shown in Figures 13 and 16. These show, with a
magnification of forty times, the binder dispersion of the crumb rubber before and after
periods of storage. Lastly, the DMA was performed using the same conditions as used in
Section 3.5 and presented in Figures 14, 15, 17 and 18.
TR-MAB A TR-MAB Af
before hot storage: coarse dispersion before hot storage: medium coarse
after 5h of hot after 5h of hot

after 3h of hot storage after 3h of hot storage
storage storage
top: medium coarse top: fine medium top: medium coarse top: medium coarse
bottom: medium coarse bottom: medium bottom: medium coarse bottom: coarse
Figure 13. Polymeric dispersion analysis of the top and bottom sections of “TR-MAB A” and
“TR-MAB Af”
1.E+09
1.E+08
1.E+07
TR-MAB A
TR-MAB A top 3h
1.E+06
TR-MAB A top 5h
TR-MAB A bottom 3h
1.E+05
TR-MAB A bottom 5h
bitumen A
1.E+04
1.E+03
1.E+02
1.E+01
10 20 30 40 50 60 70 80 90 100
Phase angle (º)
Figure 14. Black diagrams of base bitumen A and TR-MAB A before and after hot storage; a)
3h hot storage, b) 5h hot storage
1.E+09
1.E+08
1.E+07
TR-MAB Af
TR-MAB Af top 3h
1.E+06
TR-MAB Af top 5h
TR-MAB Af bottom 3h
1.E+05
TR-MAB Af bottom 5h
binder Af
1.E+04
1.E+03
1.E+02
1.E+01
10 20 30 40 50 60 70 80 90 100
Phase angle (º)
Figure 15. Black diagrams of base binder Af and TR-MAB Af before and after hot storage;
a) 3h hot storage, b) 5h hot storage
The results of these analyses show that both TR-MABs originating from bitumen A (low
asphaltenes content) begin to have a relevant phase separation after three hours of hot storage.
As shown in Figure 13, the top section of TR-MAB A has a coarse dispersion of rubber
which tends to settle toward the bottom. Comparing the images in Figure 13, TR-MAB A
shows a lower variability between the top and bottom sections than the TR-MAB Af obtained
after adding the flux to the same base bitumen. What it is suggested by the variability of the
softening points has been confirmed by the rheological characterisation. The black diagrams
in Figure 15 show how the finger print of the top section of TR-MAB Af differs considerable
from that of the bottom. Moreover, it is interesting to note how, with increasing storage time,
the top section tends to have a rheology closer to the base binder than the modified binder,
while the bottom section becomes stiffer and more elastic. The same behaviour, but less
emphasised, is shown by TR-MAB A in Figure 14.
TR-MAB B TR-MAB Bf
before hot storage: coarse dispersion before hot storage: medium dispersion
after 3h after 5h
after 3h of hot storage after 5h of hot storage
of hot storage of hot storage
top: coarse top: coarse top: medium top: fine medium
bottom: coarse bottom: coarse bottom: fine medium bottom: fine
Figure 16. Polymeric dispersion analysis of the top and bottom sections of “TR-MAB B” and
“TR-MAB Bf”
1.E+09
1.E+08
1.E+07
1.E+06
TR-MAB
BB
TR-MAB
B B top 3h
1.E+05
TR-MAB
B B top 5h
TR-MAB
B B bottom 3h
1.E+04
TR-MAB
B B bottom 5h
bitumen B
1.E+03
1.E+02
1.E+01
10 0
20 30 40 50 60 70
0 80 90 100
Phase angle (º)
Figure 17. Black diagrams of base bitumen B and TR-MAB B before and after hot storage; a)
1.E+09
1.E+08
1.E+07
1.E+06
TR-MAB Bf
TR-MAB Bf top 3h
1.E+05
TR-MAB Bf top 5h
TR-MAB Bf bottom 3h
1.E+04
TR-MAB Bf bottom 5h
binder Bf
1.E+03
1.E+02
1.E+01
10 20 30 40 50 60 70 80 90 100
Phase angle (º)
Figure 18. Black diagrams of base binder Bf and TR-MAB Bf before and after hot storage; a)
PDA and rheological analyses for TR-MABs created from bitumen B (high asphaltenes
content) confirm the results of the softening point tests (Table 14). These modified binders tend
be much more stable than the ones obtained by modifying bitumen A. TR-MAB B has good
stability showing no significant differences in terms of polymer dispersion between the top
and the bottom sections after both periods of storage. It even shows a stable rheology (Figure
17) as well as having a constant softening point (Table 14). Rheological characterisation in
Figure 18 shows how TR-MAB Bf is less stable than the base bitumen without flux. Contrary
to what was expected, the laboratory blended TR-MABs produced with base bitumens without
the addition of an oil extender tended to be more stable than the modified binders produced
with the base binder (base bitumen plus flux) especially for the base bitumen with the lower
asphaltenes content (bitumen A).
5. Conclusions
Four different Crumb-Rubber Modified Bitumens (TR-MABs), originating from two

base bitumens with different asphaltenes content, have been blended and tested. Superpave
performance grade, detailed rheological properties and storage stability analysis have been
undertaken and used as the basis for the comparison. The results have indicated that all the
modified binders have a considerable improvement in terms of their rheological, physical
and performance properties compared to the original base bitumens (40/60 pen). However,
fundamental differences have been found between the different blends:
- The TR-MABs obtained from the base bitumen with the lower asphaltenes content
(bitumen A) show a greater degree of modification when mixed with crumb rubber.
Their rheological properties demonstrate those typically found for elastomeric polymer
modified bitumens.
- Despite this fact, the TR-MABs obtained from base bitumen B (higher asphaltenes
content), show high viscosity, a better performance grade, lower temperature and
frequency susceptibility and better storage stability.
- The addition of the oil extender in the TR-MABs softens the modified binder and also
reduces performance range. In addition, the flux also appears to have a negative effect
on storage stability.
- The overall effect of adding the oil extender is the same whether it is added solely to
the base bitumen to form the base binder or used as a component in the production of
TR-MABs. Furthermore, the addition of an oil extender to a low asphaltenes content
bitumen (bitumen A) tends to have a detrimental effect on the modification in terms of
the materials stability.
- It is possible to produce a high performance TR-MAB by mixing crumb rubber with a

high asphaltenes content bitumen (bitumen B) without the need to add an oil extender,
or at least not at the high content used in this study.
6. Bibliography
Airey, G.D., “Use of black diagrams to identify inconsistencies in rheological data.” Road
Materials and Pavement Design, Vol. 3, No. 4, 2002, pp. 403-424.
Airey, G.D., “Rheological properties of styrene butadiene styrene polymer modified road
bitumens.” Fuel, Vol. 82, No. 14, 2003, pp. 1709-1719.
Chehovits, J.G., Dunning, R.L. and Morris, G.R., 1982. “Characteristics of asphalt-rubber by
the slide plate microviscometer.” Association of Asphalt Paving Technologists, Vol. 51,
pp. 240-261.
Ferry, J.D., Viscoelastic Properties of Polymers. 3rd Edition, John Wiley & Sons, Inc., New
York, 1980.
Heitzman, M. “Design and Construction of Asphalt Paving Materials with Crumb Rubber
Modifier”, Transportation Research Record 1339. 1992
Lu, X., Isacsson, U. and Ekblad, J., “Phase separation of SBS polymer modified bitumens.”
Journal of Materials in Civil Engineering, Vol. 11, 1999, pp. 51-57.
Mezger, T., The Rheology Handbook. edited by Ulrich Zorll, Vincentz. 2002.
Oliver, J.W.H., “Optimizing the improvements obtained by the digestion of comminuted scrap
rubbers in paving asphalts.” Association of Asphalt Paving Technologists, Vol. 51,
1982, pp. 169-188.
Potgieter, C.J. & Coetsee, J.S., “Bitumen Rubber Asphalt: year 2003 design and costruction
procedures in South Africa”. Proceedings Asphalt Rubber 2003 conference, Brazil,
December 2003. ISBN: 85-903997-1-0
Read, J., Whiteoak D., The Shell Bitumen handbook. Shell Bitumen, Fifth edition, 2003
Yen, T.F. “A macrostructure of petroleum asphalt.” Preprints of the Symp.: Chemistry and
characterization of Asphalt, Vol. 35, Div. Of Petroleum Vhemistry, American Chemistry
Society, Washington D.C., 1990, 314-319
Laboratory Performance Evaluation

Of Gtr-modified Sma Mixtures With
Fractionated Rap
William R. Vavrik* – Samuel H. Carpenter** –

Steven Gillen*** – Fred Garrott****
* Applied Research Associates, Inc.

100 Trade Centre Dr. Ste. 200
Champaign IL 61820
wvavrik@ara.com
** University of Illinois at Urbana-Champaign

1206 Newmark Civil Eng. Lab - MC 250
205 N. Matthews
Urbana, IL 61801
scarpent@illinois.edu
*** Illinois State Toll Highway Authority

2700 Ogden Ave
Downers Grove, IL 60515
sgillen@getipass.com
**** S.T.A.T.E. Testing

570 Rock Road, Unit K
East Dundee, IL 60118
fred@garrott.net
ABSTRACT. In 2006, the Illinois Tollway initiated a trial of hot mix asphalt (HMA) surface
mixes modified with ground tire rubber (GTR). These mixes originally were intended to
improve pavement noise, and they proved to have additional benefits for stone matrix asphalt
(SMA) and open graded asphalt mixtures.
In 2007, the Tollway embarked on a field trial and laboratory pavement performance
validation of an SMA asphalt surface mixture that uses fractionated recycled asphalt
pavement (FRAP) and GTR. This highly recycled material was evaluated during mix design,
field production, and laboratory performance testing .The Tollway selected a demonstration
project site on a rural interstate highway with high truck traffic, then proceeded with mix
designs, field production, and strict quality testing.
The mix design and field production documentation shows that GTR-modified SMA mixes with
FRAP can be designed and produced with success. The use of GTR eliminates the need for the
fibers that often are required using other modifiers.
The laboratory evaluation of performance included such tests as dynamic modulus, beam
fatigue, stability, rut testing, and tensile strength testing. The results of these tests indicate
that the performance of the GTR-modified mixes is similar to other SBS modified HMA mixes
commonly used in Illinois. These performance test data are the first validation of the use of
GTR and FRAP in SMA mixtures under heavy interstate traffic.
KEYWORDS: Fractionated Recycled Asphalt Pavement (FRAP), Stone Matrix Asphalt (SMA),
Ground Tire Rubber (GTR), Asphalt Fatigue, Asphalt Dynamic Modulus
1. Project Goals
To test fractionated recycled asphalt pavement (FRAP) materials, the Illinois Tollway,
working through its contractors and consultants, developed and conducted a project on the
applicability and feasibility of using increased amounts of recycled materials through FRAP.
The goal of the program was to answer two main questions:
− Can the Tollway design, produce, and construct high-quality hot mix asphalt (HMA)
pavements with high FRAP content mixes and specialty mixes with ground tire rubber (GTR)?
− Will these materials provide the same or better pavement performance as the standard
mixes used by the Tollway and Illinois Department of Transportation (IDOT), and with
performance that is consistent with pavement design procedures?
In the summer of 2007, a construction contract was awarded for advance pavement work
on the Jane Addams Memorial Tollway (I-90) in the Rockford area. The timing, scope, and
circumstances of the I-90 project provided a rare opportunity to evaluate several different
HMA concepts directly via plant mixing and field trials.
This paper presents the results of the mixture testing. Performance testing included
modulus, fatigue, stripping, and asphalt pavement analyzer analysis.
2. Introduction
In the summer of 2007, a construction contract was awarded for advance pavement
work on the Jane Addams Memorial Tollway (I-90) near Rockford, Illinois. To test the
new fractionated recycled asphalt pavement (FRAP) materials, Rock Road Companies and
Rockford Blacktop, the Illinois Tollway, S.T.A.T.E. Testing, and Applied Research Associates,
Inc. (ARA), with testing support provided by the Illinois Center for Transportation (ICT),
developed and conducted a research project on the applicability and feasibility of using
increased recycled asphalt pavement (RAP) contents through FRAP.
Hot mix asphalt (HMA) research testing usually is performed with laboratory-prepared
mixes. For some circumstances, this is acceptable. For others, it may be all that is available.
However, lab-prepared mixes usually are batched with oven-dried aggregates and mixed at
relatively low oven temperatures. Lab procedures do an inadequate job of imitating the high
temperatures and material handling that occurs in a plant. A plant-produced mix is better
for answering questions like “how does the mixing of RAP with virgin aggregate and liquid
asphalt in a plant affect the performance of the final mixture?”
The timing, scope, and circumstances of the I-90 project provided a rare opportunity
to evaluate several different HMA concepts directly via plant mixing and field trials. This
paper documents the activities involved in developing stone matrix asphalt (SMA) ground
tire rubber (GTR) mix designs and producing/placing those mixes. The goal of the research
program was to determine whether:
- The Tollway can design, produce, and construct high-quality HMA pavements with
GTR mixes.
- These materials will provide the same or better pavement performance as the standard
mixes used by the Tollway and Illinois Department of Transportation (IDOT), and with
performance that is consistent with pavement design procedures.
2.1 Events surrounding this research project
The Tollway faces an unprecedented rehabilitation/expansion program for its highway

network. The financial demand, market conditions, and desire to improve as much of the
network as possible require the evaluation of options for minimizing construction costs.
In the HMA industry, minimizing costs includes optimizing the selection of materials used
in the mixes. The Tollway had been looking at small pieces of this puzzle—new liquid
asphalt, different aggregates, and better performing mix designs. Nationally, fractionating
RAP is becoming recognized as an efficient way to cut the cost of a new mix and reduce the
inconsistencies of the high RAP mix properties without sacrificing quality. In May 2007,
FRAP became a serious topic of interest at the Tollway when Don Brock, President of Astec
Industries, made a convincing presentation to the Illinois HMA industry and various Illinois
government agencies.
In the summer of 2007, a large contract for preliminary work on I-90 was awarded to a
joint venture of Rock Road Companies and Rockford Blacktop. They readily agreed to work
with the Tollway to test the FRAP concept by processing the mainline overlay grindings
containing only high-quality manufactured aggregates into two fractions, category 1 fine
portion FRAP (minus #4 sieve) and category 1 coarse portion FRAP (minus ½” sieve to #4
sieve); and by processing any other reclaimed asphalt pavement containing both lower quality
natural and manufactured aggregates into two fractions, category 2 fine portion FRAP (minus
#4 sieve) and category 2 coarse portion FRAP (minus ½” sieve to #4 sieve). This concept
required additional processing and plant equipment and working with several additional
material suppliers, all without an increase in their contract price. At the same time, the
Tollway authorized S.T.A.T.E. Testing to proceed with designs for mixes which would include
a significant percentage of FRAP which would include GTR. These mixes typically contained
RAP contents 15% higher than the maximum percentages that the current IDOT or Tollway
standard specifications allow.
During late summer and fall 2007, the joint venture produced these mixes. These mixes
replaced standard mixes on the I-90 project. In most cases, each mix was placed over 2 or
more days, allowing time for mix adjustments between production runs. The joint venture
performed normal quality control (QC) testing. Quality assurance (QA) testing was performed
by S.T.A.T.E. Testing. Ultimately, the mixes were sampled for HMA materials testing at IDOT
and for more complex performance testing by the Illinois Center for Transportation at the
University of Illinois.
2.2. New concepts
As with any research, new specifications, equipment, and procedures were required to
conduct this project. These new concepts are noted below:
- A special provision was developed to implement FRAP with Illinois Tollway HMA
mixtures. It included specifications for the source material and the production of the
different FRAP products. It also included instructions on how to determine the specific
gravity of the FRAP and how to accommodate design changes.
- Contractors acquired screening equipment and made modifications to their plants to

accommodate the FRAP.
- Mix designs were created for the FRAP mixes that included non-traditional ingredient
materials, notably terminally blended GTR, modified liquid asphalt, and coarse
aggregates for the SMA mixes.
3. Materials tested
Nine HMA materials were plant produced and tested as part of this research effort, the
details of which can be found in the project report (Vavrik 2008). These mixtures, their type,
asphalt grade, and FRAP percentage are listed in table 1. The materials were selected because
they are regularly used mixes in Tollway mainline and HMA shoulder expressway pavement
sections.
The SMA mixtures included two surface mixes, one with trap rock coarse aggregate and
another with steel slag coarse aggregate, and one SMA binder that used crushed gravel as the
coarse aggregate source. The purpose of these mixtures was to evaluate the use of fine portion
FRAP and to determine if the material properties and predicted performance of the resulting
mixture were consistent with other SMA mixes previously produced in Illinois using virgin
aggregate sources only.
Table 1. FRAP SMA research mix matrix.
Research Mix Number PG grade FRAP % Fine/Coarse Coarse Agg. type

#1 –
PG 76-22 15 / 0 Crushed Gravel
SMA Binder
#2 –
PG 76-22 15 / 0 Trap Rock
SMA Surface
#3 –
PG 76-22 15 / 0 Steel Slag
SMA Surface
4. Summary of field notes
The following notes summarize the characteristics of the trial mixes. These notes were
compiled by S.T.A.T.E. Testing through their experiences from aggregates in mix designs and
through their field observations of trial mix production and compaction. The complete mix
design, mix design notes, and QA summary are provided in the project report (Vavrik 2008).
4.1 Mix 1 – SMA binder
- Quite likely the first Illinois SMA binder to include crushed gravel as the coarse
aggregate. Dolomite is the customary coarse aggregate in most SMA binder mixes.
Gravel is locally available and may better resist breakdown during compaction. The
contractor used a blend of CM13 and CM14 from two different pits.
- The liquid asphalt was GTR-modified and complied with the PG 76-22 grade, with
the exception of the tests on Residue From Rolling Thin Film Oven Test (AASHTO T
240).
- The FRAP proportion was 14% of the total mixture (fine portion category 1 FRAP).
SMA mix designs typically require the use of a manufactured fine aggregate and
require more liquid asphalt than conventional dense-graded mixes. The fine aggregate
FRAP portion source in the SMA test mix was processed from mainline overlay
grindings that consisted of high-quality crushed coarse aggregate and manufactured
fine aggregate in the original overlay materials.
- Using GTR-modified liquid asphalt eliminated the use of fibers normally used to
prevent liquid asphalt draindown during storage and transportation of SMA mixes
produced with the SBS polymer modified liquid asphalt.
The contractor used two coarse aggregate products and feeders for control. Production
took place over two nights. After the first night, minor proportioning adjustments were made
to increase voids. The resulting volumetric properties (air voids, voids in mineral aggregate
[VMA], and voids filled with asphalt [VFA]) were within specification tolerances. Limited
production quantities restricted further adjustments and testing.
4.2. Mix 2 – SMA surface (trap rock)
- This is the first Illinois SMA surface mix that includes Diabase (trap rock) as the
coarse aggregate. Diabase is a hard, non-absorptive aggregate with excellent friction
properties. Most high-volume SMA surface mixes in Illinois have included steel slag as
the coarse aggregate. The Diabase consisted of two different gradations—35% CM13
and 46% CM14.
- The liquid asphalt was GTR-modified PG 76-22.
- The FRAP proportion was 14% of the total mixture (fine portion category 1).
- The GTR liquid eliminated the customary use of fibers to prevent liquid draindown
during storage and transportation.
The contractor used two coarse aggregate products and feeders for control. Production
took place over two nights. Voids were slightly above target. VMA and VFA were within
specification limits. Limited production quantities restricted further adjustments and testing.
4.3. Mix 3 – SMA surface (steel slag)
- This mix included steel slag as the coarse aggregate. The Heritage slag made up 80%
of the aggregate blend—55% CM13 gradation and 25% CM11 gradation.
- The liquid asphalt was GTR-modified PG 76-22.
- The FRAP proportion was 15% of the total mixture (fine portion category 1 FRAP).
- The GTR liquid eliminated the customary use of fibers to prevent draindown during
storage and transportation.
The contractor used two coarse aggregate products and feeders for control. Production took
place during one day shift. The volumetric properties were within or close to specification
tolerances. Limited production quantities restricted further adjustments and testing.
5. Performance evaluation methodology
The design of full-depth HMA pavements relies on two primary HMA material properties,
the fatigue performance curve and the HMA modulus (stiffness). The dynamic modulus
determines how much the pavement section flexes under the load of a heavy truck, resulting in
strain in the asphalt layer. This strain is then evaluated against the fatigue performance curve,
which relates the strain to the allowable loads that the pavement section can carry.
Other HMA material properties that are of concern include resistance to rutting and
resistance to weathering. Relative rutting resistance is indicated by the dynamic modulus test,
which is included in this testing program, and is controlled for most projects through the mix
design process and by the selection of raw materials.
6. Asphalt mixture performance tests
Laboratory testing was performed to measure laboratory material properties and determine
the expected performance of these mixes in field applications. The laboratory program was
conducted on plant-produced mix on the second night of production. The materials were
properly sampled, and the HMA tests were performed without incident. All of the identified
mixtures were sampled at the plant and taken to the Advanced Transportation Research and
Engineering Laboratory (ATREL) at the University of Illinois. ICT research staff reheated and
split the mix, prepared test samples, and conducted all materials tests. The tests performed
included flexural beam fatigue testing and dynamic modulus testing (AASHTO TP-321 and
TP62-08, respectively).
6.1. Asphalt fatigue testing
Asphalt fatigue is a distress that develops with bottom-up cracking in a full-depth asphalt
section. Fatigue test results also can be considered an indication of the fracture toughness of
a mixture and its resistance to cracking that may develop from the top down. Asphalt fatigue
testing requires the compaction of asphalt beams that are placed in a testing apparatus and
subject to repeated bending. The equipment monitors the loads and deflections during this
bending and identifies the failure point as when the sample loses half of its initial stiffness.
Asphalt fatigue tests were performed at between four and six strain levels, corresponding
to strains that could be encountered in field conditions. These tests are used to develop a
fatigue curve and fatigue slope that can then be compared with the fatigue curve assumed in
pavement design. Fatigue curves with a higher fatigue slope than the design slope have the
potential to provide a longer fatigue life. The asphalt fatigue slopes for the SMA GTR mixes
are presented in figure 1.
6.00
5.50
5.00
Fatigue Slope
4.50
Proposed IDOT Design Slope

4.00
3.50
3.00
Current IDOT Design Slope
2.50
#1 - SMA Binder #2 - SMA Surface (trap rock) #3 - SMA Surface (steel slag)
Figure 1. Fatigue slopes for the study mixes plotted with the IDOT current and proposed
design values.
The test results show excellent fatigue performance for the SMA GTR mixes. This
is consistent with the testing of normal SMA mixtures without RAP, as shown in testing
conducted for IDOT on polymer and fiber modified liquid binder (Carpenter, 2007). The
fatigue performance of the SMA mixtures can be attributed primarily to the GTR modified
liquid that is used in these mixes, combined with the general characteristics of SMA materials.
6.2. Dynamic modulus (E*)
The modulus of HMA paving materials is one of the most important factors in establishing
the pavement cross section and thickness. How to determine that modulus has changed over
the years, but recent research has shown that the dynamic modulus is an appropriate method
to develop comparative values.
The dynamic modulus test involves applying a compressive cyclical load to the HMA test
sample. The deflection of the sample is measured by linear variable differential transformers
(LVDT’s) mounted on the HMA sample. The dynamic modulus is the maximum stress divided
by the maximum strain.
The complete characterization of HMA modulus requires testing at 20℃, 4℃, and -10℃.
The frequencies used varied from 0.01 Hz to 25 Hz. This combination of test parameters
allows a characterization covering all but the extreme high temperature regime. All design
temperatures and traffic speeds are covered by this testing program. Upon completion of the
testing, a master curve is developed using time-temperature superposition techniques.
Dynamic modulus tests were performed on samples with 4% and 7% air voids at the
temperatures and frequencies previously noted. These tests were used to develop master
curves so that modulus values could be evaluated at a variety of temperatures and frequencies
that are expected for in-service pavements. The HMA dynamic modulus at 20℃ for each of
the mixes is shown in figure 2.
The dynamic modulus test data show that, for all mixes and all speeds, the modulus of the
SMA GTR mixtures is higher than the values currently assumed for the IDOT mechanistic
design procedure. This result demonstrates that SMA high FRAP GTR mixes will not require
changes to the pavement thickness and that the performance of the roadway can be expected
to be as good as current materials and designs. These values are very similar to values
determined in testing of typical IDOT mixtures for an extended life pavement research project
(Carpenter, 2007).
6.3. Additional testing
The IDOT Bureau of Materials and Physical Research was provided samples of the three
SMA mixtures for evaluation of tensile strength ratio (TSR), stability, and testing in the
asphalt pavement analyzer (APA) and in the Indenter. As detailed in the following sections,
the SMA mixtures with FRAP and GTR have similar laboratory test results to other IDOT
tested mixes and are therefore expected to perform similarly.
6.3.1. Tensile strength ratio
TSR testing (Illinois Modified AASHTO T-283) was performed on the SMA binder with
gravel, SMA surface with steel slag, and SMA surface with trap rock. All tested samples
passed the TSR test, with the gravel mix having the lowest TSR and the highest strength.
Even after five freeze/thaw cycles, which generally will show a dramatic reduction of strength
and TSR on susceptible mixes, the gravel SMA had good strength (103 psi). It had about the
same rate of strength loss between the 140°F water bath strength and the freeze/thaw cycle
strength as it had between no conditioning strength and the 140°F water bath strength. The
TSR after five freeze/thaw cycles was 0.810, which nearly passed the Tollway criterion of 0.85
for 6-inch specimens after conditioning in the 140°F water bath only. The other two mixes had
appropriate strengths and good TSR’s. Figure 3 shows a chart of the strength and TSR data.
2,500,000
2,250,000 SMA Binder

SMA Surface (trap rock)
SMA Surface (steel slag)
2,000,000
1,750,000
Dynamic Modulus, E* (psi)
1,500,000
1,250,000
1,000,000
750,000
500,000
250,000
0
Urban Street State Route Interstate
15 mph 45 mph 60 mph
Figure 2. Dynamic modulus test data at 20℃ and various highway speeds with the IDOT
design range for northern Illinois.
6.3.2. Stability
Both 4-inch Marshall and 6-inch Superpave gyratory stability tests were performed on the
second sample of SMA binder with gravel and the trap rock samples. There was no significant
difference between the two mixes. On the 4-inch samples, the stability of the gravel was
slightly greater than that of the trap rock. For the 6-inch samples, the stability of the trap rock
was somewhat greater than that of the gravel binder. Figure 4 presents the stability data.
6.3.3. Asphalt pavement analyzer
For the APA, IDOT evaluated the number of cycles per mm of rut depth, as well as the rut
depth only, following the guidance in AASHTO TP 63-03. IDOT primarily uses the “cycles
per mm” for the steel wheels when both the rut depth and the number of cycles until the end
of the test vary. With the hoses and the steel wheels, these mixes easily lasted the full 8,000
cycles (for the hoses) and 20,000 cycles (for the submerged steel wheels test), so only the rut
depth varied.
- Hose: The trap rock had slightly more cycles per mm of rut and a slightly lower rut
depth. The steel slag had the least cycles per mm of rut and the greatest rut depth. All
the results were close and were good.
- Steel Wheel: The gravel had slightly more cycles per mm of rut and, by a slight
amount, had the lowest rut depth. The steel slag had the least cycles per mm of rut and
the greatest rut depth. Again, all the results were close and were good.
Tollway I-90, GTR & FRAP Mixes - Strength & TSR

140.0 0.98
127.9 0.969
120.0
0.96
0.957
115.9
101.8
100.0 91.4
97.4 0.94
88.6
Strength (psi)
80.0
0.92
TSR
60.0
0.906 0.90
40.0
0.88
20.0
0.0 0.86
#1 - SMA Binder (Gravel) #3 - SMA Surface (Slag) #2 - SMA Surface (Trap Rock)
Average Uncond Strength Average Cond Strength TSR
Figure 3. Strength and TSR data for the SMA mixes.
6000
5083
5000
4400
4000
Stability (lbs)
3000
2775
2517
2000
1000
4" Avg Stability
6" Avg Stability
0
SMA Binder, Gravel, GTR, FRAP #2 - SMA Surface (Trap Rock)
HMA Mix Type
Figure 4. SMA stability for 4- and 6-inch samples.

12000
SMA Binder, Gravel, GTR, FRAP
SMA, Steel Slag, GTR, & FRAP - P7-69 10076
10000
#2 - SMA Surface (Trap Rock) 9062
8639
8776
8000
7474
Cycles per mm of Rut
6944
6000
5115
4290 4313
4020
4000
3823 3788
2000
0
Hose Man Cycles / Hose Auto Cycles / SW Man Cycles / mm SW Auto Cycles / mm
mm mm
APA Mode Description
Figure 5. APA performance data for the SMA mixtures.
6.3.4. Indenter
The SMA gravel binder mix and the trap rock mix were tested in the Indenter, which
measures deformation of an HMA gyratory specimen at an in-service temperature. All Indenter
specimens start off at approximately 152 mm height (115-mm specimen and 36.7 mm for the
Indenter). These specimens all ended up around 135 to 136 mm after 300 gyrations. Compared
with all the other Indenter tests IDOT has run for the last 2 or 3 years, this is somewhere in
the middle. The deformation/number of gyrations slope is fairly straight throughout the 300
gyrations. Some mixes deform quickly then level out. Since the deformation is fairly gradual,
it is better than if it deformed quickly, even if it ended up at the same point.
7. Conclusions
Given the fatigue and modulus performance test data, two conclusions can be drawn at this
time. First, the fatigue performance of all high FRAP GTR mixes is above the current design
criterion. Second, the SMA materials with GTR and fine FRAP have material properties
similar to those of other SMA mixes. It is expected that these high RAP GTR modified SMA
mixes will perform similar to other SMA materials produced with virgin aggregate sources
and the SBS polymer modified asphalt.
8. References
S. H. Carpenter, Dynamic Modulus Performance of IDOT Mixtures, Research

Report FHWA-ICT-07-008, Illinois Center for Transportation, Urbana, December 2007.
W. Vavrik, S. Carpenter, S. Gillen, F. Garrott, J. Behnke, Evaluation of Field-

Produced Hot Mix Asphalt (HMA) Mixtures with Fractionated Recycled Asphalt Pavement
(RAP): 2007 Illinois Tollway Field Mix Trials, Research Report ICT-08-030, Illinois
Center for Transportation, Urbana, October 2008.
Steady Shear Properties of Crumb Rubber

Modified Bitumen
Biswanath Saha 1 , Sonal Maheswari 1 , P. Senthivel 1 , N. V.

Choudury1 and J. Murali Krishnan2
1
BPCL-Corporate R&D Centre, Plot No. 2A, Udyog Kendra, P.O. Surajpur, Greater Noida
(UP) 201 306, India
2
Department of Civil Engineering, Indian Institute of Technology Madras, Chennai 600036,
India,
E-mail: jmk@iitm.ac.in (Corresponding Author)
ABSTRACT. The rheological properties of crumb rubber modified bitumen poses considerable
challenges in terms of quantification. This investigation concerns the measurement of shear
stress and normal stress difference when samples of crumb rubber modified bitumen were
tested in a dynamic shear rheometer in a steady shear mode for a range of temperature and
shear rates. Two commercially available crumb rubbers (I and II) were blended with an air
blown bitumen. Three different dosage rates were used and the testing was carried out. It is
seen that the material exhibit considerable normal stresses and normal stress difference when
subjected to shear. A coarser graded crumb rubber (crumb rubber-I) results in considerable
reduction in shear stress and normal stress difference for any dosage rate, shear rate and
temperature. A finer graded crumb rubber (crumb rubber-II) results in considerable higher
stress-overshoot in shear stress and normal stress difference for any dosage rate, shear
rate, and temperature. For a given crumb rubber type and temperature, as the shear rate is
increased, the normal stress differences increases at a much higher rate than the shear stress.
These measurements clearly depict the difference in terms of crumb rubber type and dosage
rates. These kinds of fundamental rheological experiments, when appropriately interpreted,
shed considerable light on the interaction of crumb rubber with bitumen.
KEYWORDS: Steady Shear, Shear Stress, First Normal Stress Difference, Crumb Rubber, Air
Blown Bitumen, Stress Overshoot
1. Introduction
India is currently proceeding with one of the ambitious bituminous pavement building
exercise. More than US$ 50 Billion is being spent on connecting the length and breadth of the
country with four to six lane roads. The first phase of the road construction is completed and
efforts are on to quantify the performance of the constructed facilities. It is interesting to note
that considerable amount of crumb rubber modified bitumen is being used in the pavement
construction.
Bitumen is processed in India through various methods. Depending on the crude source
and the viscosity/penetration grade required, bitumen is air-blown or blended (IS73, 2006).
In the air-blowing process, a conventional method of blowing compressed air by simple
distribution rings is used. Air-blowing is also carried out in some refineries through a patented
BITUROX unit. Considerable variability exists related to the crude source and processing
methods across the country and efforts are underway to take into account the variability
related to the physical, chemical and rheological characteristics.
Research on modification of bituminous binders with crumb rubber has gained significant
interest world-wide due to the increased environmental concerns related to the disposal of
scrap tire. It is claimed that addition of crumb rubber can increase the stiffness of the binder
at high temperature and reduce the stiffening of the binder at low temperature thus effectively
expanding the temperature range to which the binder can be used (Morrison and Hesp 1995).
Interesting history related to the use of crumb rubber in bitumen and bituminous mixtures and
the details related to dry and wet process can be seen from the reference of Morrison and Hesp
(1995). Rheological investigations related to the use of devulcanized rubber tyre particles
in bitumen and their fracture resistance are also discussed in Morrison and Hesp (1995).
Considerable investigation and understanding related to crumb rubber – asphalt interaction is
mainly through the studies of Amirkhanian and co-workers (Amirkhanian 2003; Lee at al.
2006; Putman and Amirkhanian 2006; Shen et al. 2009; Thodesen et al. 2009a).
It is well known that crumb rubber produced by the cryogenic process differ significantly
when compared with the crumb rubber produced in the ambient conditions (Blumenthal
1994). The surface area for a given particle size is much higher for crumb rubber produced
in ambient condition due to the mechanical tearing action. Complex reactions take place
when crumb rubber is ingested with bitumen. Depending on the type of bitumen (high/low
viscosity grade) and depending on the type of crumb rubber (content, particle size distribution
and surface area characteristics), one can postulate a wide variety of reactions. Absorption of
aromatic oils, swelling of rubber particles, possible dissolution of rubber components and de-
volatilization and cross-linking in rubber can be some of the reactions (Shen et al. 2009). It
is also postulated that the increase of apparent viscosity of crumb rubber modified asphalt is
due to two main mechanisms, the first one is due to the swelling of crumb rubber particles
resulting in reduced space for the particles to move around and the second is due to the
stiffening of the binder (Abdelrahman and Carpenter 1999; Shen and Amirkhanian 2005).
Most of the work related to crumb rubber modified bitumen have been mainly related to
the measurement of viscoelastic properties assuming that the material is linear viscoelastic.
Hence, it is not uncommon to see investigations trying to elicit the performance grade
properties of crumb rubber modified in terms of crumb rubber properties (see for instance
Thodesen et al. 2009b). One need not overemphasise the necessity to characterize the
behaviour of the material in terms of fundamental rheological tests and then build appropriate
constitutive models towards the same. The experimental investigations reported in this study
are aimed towards that step.
In the following we first discuss the rheology of steady shear flow. Two types of crumb
rubber particles were blended with an air blown bitumen at different dosage rates and the
blend was tested at four different temperatures for four shear rates. These details are given
in section 3. In section 4, we summarize the results and discuss the possible reasons for such
wide range of behaviour when two different types of crumb rubber are used.
2. Rheology of Steady Shear Flow
One of the most fundamental viscometric flows is the simple shear flow. The velocity fields
for such a flow is given as follows,
[1]
where is the velocity gradient and , , and are the velocity fields. For steady shear
flow, the shear rate (the absolute value of velocity gradient) is independent of time. This is one
of the classic flows which are normally used for characterizing the rheological properties of
viscoelastic fluids.
Constitutive equations relate the strains developed in the material to the stresses and
bulk of the experimental work is conducted to establish these relationships. The constitution
of the material and its mechanical response plays an important role in developing a proper
constitutive model. For instance, if the fluid is Newtonian (viscosity independent of shear
rate), the only component for the stress tensor which is non-zero is the shear stress for the
velocity field given in equation [1]. It is interesting to note here that when a viscoelastic fluid
is subjected to shearing, one sees considerable complexity in terms of the components of the
stress tensor. Here it is possible that all the six independent components of the stress tensor
may be non-zero. The most general form of the stress tensor for the case of simple shearing
flow for such fluids is given as follows (Bird et al. 1987),
[2]
where is the stress tensor, is the thermodynamic pressure and is the total stress. Now
the stresses that are customarily used in characterizing the shear flow are essentially shear
stresses given by and the normal stress differences. As can be seen from equation [2], there
are two normal stress differences, the first normal stress difference given by and the
second normal stress difference given by . Using the above basic definition, in non-
Newtonian fluid mechanics literature, it is customary to define a non-Newtonian viscosity or
shear-rate dependent viscosity and two normal stress coefficients and all these
three functions are collectively called as viscometric functions (Bird et al. 1987).
In a dynamic shear rheometer normally used for characterizing the viscoelastic properties
of bitumen, the top plate is rotated and the bottom plate is kept fixed. The torque and the
normal force engendered during this motion are measured. Typically, in bitumen literature,
the torque (and hence the shear stress) and the angular velocity (and hence the shear rate)
are of main interest and different types of experiments are performed to elicit the mechanical
characteristics of bitumen. However, over and above the torque and angular velocity
measurements, the normal forces hold considerable information related to the rheological
characteristics of the material. Typically, in a steady shear flow where the upper plate is
subjected to constant angular velocity, one can measure in a dynamic shear rheometer the
torque required to achieve this rotation as well as the total normal force required to maintain
the disks at a separation of any given height. When a material like bitumen is sheared,
considerable normal forces are developed in the material (equation [2]) and these forces
tend to push the upper plate apart. Since the shearing is conducted at a constant height, the
dynamic shear rheometer applies an equal and opposite force during shear to maintain this
shear and this normal force along with the torque are the raw data which can be collected
during measurement. An equipment specific material constant can be used to convert torque
to shear stress, angular velocity to shear rate and the normal force to first and second normal
stress differences.
The measurement of normal forces during shear is as old as experimental mechanics.

In his seminal paper on granular materials, Reynolds (1885) discussed a quintessential
feature concerning the response of granular solids, that of “dilatancy”. Granular materials
when sheared undergo a change in volume and in order to keep the volume constant, one
has to apply a force, i.e., shearing engenders forces perpendicular to the plane of shear.
This phenomenon is a characteristic shared in common by many non-linear materials and
the underlying physics goes by different names in different fields. It is called the “Poynting
effect” in non-linear elasticity (Poynting 1909), “Weissenberg effect” in non-Newtonian fluid
mechanics (Weissenberg 1947) and “dilatancy” in the theory of granular solids (Reynolds
1885). Review of literature related to bitumen and bituminous mixtures show that scant
literature exists related to measurement of this aspect (Stastna et al. 2003, Kasula et al. 2005,
Wekumbra et al. 2007) let alone developing sophisticated constitutive models to predict such
response.
It is noted here that the shear rate is not constant in the parallel plate rheometer. The
velocity field for the parallel plate is given as and the shear rate
is Here the angular velocity is given as w and it is seen that the shear rate varies
across the plate. The shear rate at the periphery is maximum and this is the shear rate which
is used in all the material parametric calculations. There are also considerable disadvantages
due to inertia as well as edge fracture in the parallel plate assembly. It is also interesting to
point out here that edge fracture due to the extremely high shear rate at the periphery is an
artifact of the equipment and it is not related in any manner to the rheological properties of the
binder. Interesting discussions related to whether one can quantify the edge fracture seen in
oscillatory motion of a dynamic shear rheometer as depicting the fatigue and fracture response
of asphalt films can be seen in the paper by Anderson et al. 2001.
Another important parameter during steady shear flow is the behavior of viscoelastic
materials during transient shear (start-up) flow (Bird et al. 1987). Transient studies have
revealed striking nonlinear phenomena, e.g., a shear stress overshoot after sudden imposition
of a fixed finite shear rate. In fact, these nonlinear overshoot effects have proved to be
extremely useful in studies concerned with developing models that depict the real material
response. While this aspect has been studied extensively in polymer rheology, this is not
studied at all in detail in bitumen rheology and very few experiments have been conducted to
characterize the startup of the steady shear flow. In fact one of the investigations that reports
few experimental results related to the stress growth coefficient in polymer modified asphalt
is due to Wekumbura et al. 2007. In these experiments, essentially for very small shear rates
only, the shear stress approaches its steady state monotonically (Bird et al. 1987). For large
shear rates, the shear stress reaches a maximum and then approaches the steady state value.
In this investigation, two fundamental aspects of steady shear flow are investigated. The
first one is the development of shear stress and normal stress differences when the material
is subjected to steady shear. The second aspect is the overshoot of shear stress during steady
shear conditions and the influence of shear rate on this overshoot.
From the discussion related to the mechanism of bitumen – crumb rubber interaction, it
is seen that considerable swelling takes place when crumb rubber is ingested with bitumen.
When this swelled fluid mixture is subjected to steady shear, the resistance of the particle
during shear engenders normal forces perpendicular to the plane of shear. Hence, one
should ideally notice considerable increase of normal forces as the dosage of crumb rubber
is increased. As the normal forces are dependent on shear rate, one should also notice the
incidence of higher normal forces as the shear rate is increased for a given crumb rubber
percentage. In the following we show through experimental investigation normal stress
differences and stress overshoot.
3. Experimental Investigations
3.1. Materials
Two types of commercially available crumb rubber were used. Table 1 show their gradation
and table 2 gives their relevant properties. Both these crumb rubber were manufactured under
ambient conditions.
Table 1. Sieve Analysis of Crumb Rubber
Sieve number Crumb rubber – I Crumb rubber – II

(mesh size in micron) (percentage retained) (percentage retained)
30 (600) 0.48 0.20
40 (425) 3.46 1.77
60 (250) 50.02 34.76
100 (150) 40.95 54.82
200 (75) 4.17 6.26
Pan 0.92 2.19
Table 2. Crumb Rubber Properties
Property Crumb rubber – I Crumb rubber – II

Moisture + volatile content
0.27% 0.28%
(up to 300 ℃)
Ash content 4.22% 10.78%
Toluene insoluble 84.7% 80.9%
The bitumen used was provided by Mumbai Refinery, India and confirmed to the VG-30
grade of IS: 73-2006. In terms of penetration grade, this material confirms to the 60/70 grade.
Bitumen and the crumb rubber were prepared by mixing together at 180 ℃ and the blend was
kept in the agitation with high speed homogenizer for 1 hour at 4000 rpm with temperature
around 175-180 ℃. The samples were then stored in separate cups for further testing.
During the start of each and every testing, the sample was slightly warmed up in a water bath
maintained at 60 ℃ and a steel spatula was used for completely mixing and homogenizing the
blend. Considerable variations are normally noticed when the testing was conducted without
this step.
3.2. Steady Shear Experiments
Anton-Paar dynamic shear rheometer (MCR301) was used for measurement in all the
investigations. In addition, these experiments were repeated at the R&D facility of Bharat
Petroleum Corporation Limited in the Anton-Paar dynamic shear rheometer (MCR501) for
reproducibility. Parallel plate measurement system of 25 mm diameter was used with a gap
of 0.8 mm. The plate diameter and the gap were chosen taking into account the torque and
normal force sensor capabilities of both the equipments. The material was sheared for 300
second with data points collected every 0.01 second. The test matrix used is given below in
table 3.
Table 3. Test Matrix
Variables Details
Crumb rubber type I and II
Dosage rate 8, 10, 12 %
Shear rate (temperature) 1, 2, 5 and 8 s-1 (60 ℃)
5, 10, 15 and 20 s-1 (70 ℃)
30, 40, 50 and 60 s-1 (80 ℃)
40, 50, 60 and 70 s-1 (90 ℃)
4. Results and Discussions
Figures 1(a) – (d) show the results of the steady shear experiments conducted at 60 ℃
for the two types of crumb rubber used in this investigation. For the sake of illustration, only
shear rates 1 and 8 s -1 are shown here. The influence of the size fractions of the two types of
crumb rubber are clearly seen here. As the material is sheared, bitumen blend with finer crumb
rubber fractions exhibit significant shear and normal stresses.
Figure 1-a. Crumb Rubber – I at 1 s-1 for Figure 1-b. Crumb Rubber – II at 1 s-1 for
various dosages various dosages
Figure 1-c. Crumb Rubber – I at 8 s-1 for Figure 1-d. Crumb Rubber – II at 8 s-1 for
various dosages various dosages
It is also interesting to note that at higher dosage rates and for the same shear rates, crumb
rubber particles with finer particle size exhibit first normal stress difference higher than shear
stress. As the numbers of particles are more for the given percentage, the particle surface area
is increased resulting in increased capacity for absorption of aromatics. When this blend is
sheared, the normal stresses also increases.
Figure 2-a. Crumb Rubber – I at 10 s-1 Figure 2-b. Crumb Rubber – II at 10 s-1
for various dosages for various dosages
Figure 2-c. Crumb Rubber – I at 20 s-1 Figure 2-d. Crumb Rubber – II at 20 s-1
for various dosages for various dosages
Figures 2(a)-(d) show the results for 70 ℃ for various dosage rates for the two crumb
rubbers tested. The manner in which the shear stresses and first normal stress differences are
evolving with time as the material is being sheared is clearly seen. It is also interesting to note
the considerable amount of stress overshoot and decay in case of first normal stress difference
for the finer crumb rubber particles. Due to the increased packing associated with finer
particles, the resistance increases initially and as the material is being systematically sheared,
this resistance systematically decays with time. The similar behaviour is also seen at 80 ℃ as
shown in figures 3(a)-(d).
Figure 3-a. Crumb Rubber – I at 30 s-1 for Figure 3-b. Crumb Rubber – II at 30 s-1
various dosages for various dosages
Figure 3-c. Crumb Rubber – I at 60 s-1 for Figure 3-d. Crumb Rubber – II at 60 s-1
It is noted from figures 2(a)-(d) and 3(a)-(d) that as the shear rate is increased for the same
temperature, the first normal stress difference is much higher than the shear stress for a higher
dosage. However, the increased magnitudes of first normal stress difference with shear stress
occurs at a lower dosage rate for the crumb rubber-II grade material as the particles in this
case are much finer. Figures 4(a)-(d) show shear stress and normal stresses change for two
shear rates for all the dosage rates at 90 ℃.
Figure 4-c. Crumb Rubber – I at 60 s-1 for Figure 4-d. Crumb Rubber – II at 60 s-1
It is seen that as the temperature increases, the mobility of the particles to slide past each
other increases and hence this results in the first normal stress difference always less than the
shear stress. However, the blend with crumb rubber-II at a higher dosage rate of 12% still
exhibits higher normal stress difference initially but decays to have value equal to that of the
shear stress (figure 4-d).
One important manifestation of the non-linear behaviour of viscoelastic fluid is the

tendency to exhibit stress-overshoot during the inception of steady shear flow. To highlight the
stress-overshoot exhibited, two representative graphs are shown below in figures 5-a and 5-b.
The influence of particle size distribution is more pronounced in terms of first normal stress
difference. This is clearly evident from figure 5-b where one can see that the crumb rubber-I
blend reaches its plateau much earlier than blend having crumb rubber-I for all dosage rates.
In case of shear stress (figure 5-a), the magnitude of the shear stress shoot depends clearly on
the dosage rate and on the type of crumb rubber used.
5. Conclusions
Interaction of crumb rubber particles with bitumen is quite complex. A large number of
factors related to the crumb rubber particle shape, size, surface area, reaction time, processing
temperature and the type of bitumen influence the rheological behaviour of the blend.
Selective absorption of oil by rubber particles and their related swelling, increased stiffening
of the parent binder and the reduction in mobility of rubber particles as the dosage rate is
increased are possibly some of the cause for this complex behaviour. The present investigation
reported on capturing the mechanical characteristics of this material when subjected to steady
shear. Due to the inherent non-linearity, one sees considerable normal stresses when subjected
to steady shear. Some of the important observations related to this study can be summarized
below:
• A coarser graded crumb rubber (crumb rubber-I) results in considerable reduction

in shear stress and normal stress difference for any dosage rate, shear rate and
temperature.
• For a given crumb rubber type and shear rate, as the dosage is increased, the difference
between the shear stress and normal stress differences keeps decreasing. However, as
the temperature is increased, the onset of normal stress difference is considerably lower
signifying the reduction of resistance to mobility among rubber particles.
• A Finer graded crumb rubber (crumb rubber-II) results in considerable higher stress-
overshoot in shear stress and normal stress difference for any dosage rate, shear rate,
and temperature.
• For a given crumb rubber type and temperature, as the shear rate is increased, the
normal stress differences increases at a much higher rate than the shear stress.
The measurement of normal stresses and their difference along with shear stress hold
the key to development of robust constitutive models for crumb rubber modified bitumen.
The influence of the particle size distribution as well as the dosage rate was clearly captured
during the steady shear motion reported here. Efforts are currently underway to quantify the
influence of size and shape characteristics of crumb rubber particles and their relationship to
the development of normal stresses and shear stresses.
Acknowledgements: J. Murali Krishnan acknowledges the support of Polymer Engineering

Laboratory, IIT Madras on the conduct of experiments.
References
Abdelrahman, MA and Carpenter, SH., “Mechanism of interaction of asphalt cement with

crumb rubber modifie”r. TRR No. 1661, 1999, 106–113.
Amirkhanian SN, “Establishment of an Asphalt-Rubber Technology Service (ARTS”), In:

Proceedings, asphalt rubber 2003 conference, 2003, Brasilia, Brazil.
Anderson, DA., Hir, YML., Marasteanu, MO., Planche, J–P, Martin, D. and Gauthier, G.,.
« Evaluation of fatigue criteria for asphalt binders ». Transportation Research Record
1766, 2001, 48–56.
Bahia, H.U., and Davies, R,. « Effect of crumb rubber modifiers (CRM) on performance-
related properties of asphalt binders » AAPT 63, 1994, 414–438.
Bird R.B., Armstrong R.C. and Hassager O., “Dynamics of polymeric liquids”, Volume I,
John Wiley, New York, 1987.
Blumenthal MH. « Producing ground scrap tire rubber: a comparison between ambient and
cryogenic technologies » Scrap Tire Management Council, Washington, DC, 1994.
IS 73 -2006, Specification for paving grade bitumen, Bureau of Indian Standards, India.
Kasula, L.R. Krishnan, JM. Rajagopal, KR. and Little, DN., 2005. « Normal stress and
stress relaxation data for sand asphalt undergoing torsional flow » Mechanics Research
Communication 32 (1), 43–52.
Lee SJ, Amirkhanian SN, Thodesen C, Shatanawi K,. “Effect of compaction temperature on
rubberized asphalt mixes” In: Proceedings, asphalt rubber, 2006, Palm Springs, CA.
Morrison, GR, Hesp, SAM,. « A new look at rubber-modified asphalt bindes », Journal of
Materials Science, 30, 1995, 2584-2590.
Poynting, JH.,. « On pressure perpendicular to the shear planes in finite pure shears, and
on the lengthening of loaded wires when twisted » Proceedings of the Royal Society of
London 82 557, 1909, 546–559.
Putman BJ, Amirkhanian SN. « Crumb rubber modification of binders: interaction and
particle effects » Proceedings, Asphalt Rubber, Asphalt Rubber 2006, Palm Springs, CA,
November 2006.
Reynolds, O.,. « On the dilatancy of media composed of rigid particles in contact. With
experimental illustrations » The London, Edinburgh, and Dublin Philosophical Magazine
and Journal of Science 20 127, 1885, 469–481 (Series 5)
Shen, J, and Amirkhanian, SN.,. « The Influence of crumb rubber modifier (CRM)
microstructure on the high temperature properties of CRM binder » Int J Pavement Eng 6
(4), 2005, 265–271.
Shen, J, Amirkhanian, SN., Feipeng Xiao and Boming Tang,. « Influence of surface area and
size of crumb rubber on high temperature properties of crumb rubber modified binders »
Construction and Building Materials, Volume 23, Issue 1, 2009, 304-310.
Stastna J., Zanzotto L., and Vacin O. J.,.“Viscosity function in polymer modified asphalts”
Journal of Colloid and Interface Science, 259, 2003, 200–207.
Thodesen,C, Shatanawi, K, Amirkhanian, SN,. « Effect of crumb rubber characteristics on

crumb rubber modified (CRM) binder viscosity » Construction and Building Materials,
Volume 23, Issue 1, 2009a ,295-303.
Thodesen, C, Shatanawia, K, Amirkhanian SN, and Bridges, W,. « Development of an

empirical model for determining G*/sin δ in crumb rubbernext term modified binders »
Construction and Building Materials, 2009b, http://dx.doi.org/10.1016/j.conbuildmat.2008
.09.002 (Article in Press)
Weissenberg, K.,. « A continuum theory of rheological phenomena » Nature 159 March 1947,
310–311
Wekumbura C., Stastna J. and Zanzotto L.,. “Destruction and recovery of internal structure
in polymer-modified asphalts” Journal of Materials in Civil Engineering, Vol.19, 2007,
227-232.
Xiao, F, Amirkhanian, SN, Shen, J, and Putman, B, 2009. Influences of crumb rubber size and
type on reclaimed asphalt pavement (RAP) mixtures, Construction and Building Materials,
Volume 23, Issue 2, 1028-1034.
Chapter 8
Invited papers on related

aspects
(As invited papers the presentations in this Chapter have not been subjected to
the Technical Committee blind peer review)
Invited Papers on Related Aspects 955
Experimental Study on Strength Developing

Law of Epoxy Asphalt Mixture during its
Curing Reaction
Huang Wei — Chen Lei-lei — Qian Zhen-dong
Intelligent Transportation System Research Center

Southeast University
Nanjing, 210096
China
Email: hhhwei2005@126.com
Email: leilei1985117@163.com
Email: qianzd@seu.edu.cn
ABSTRACT. Epoxy asphalt mixture (EAM) is widely used in the pavement engineering
recently. As a thermosetting material, its strength developing law is different from
thermoplastic paving material such as HMA, which lead to a more complex construction
control. This paper presents an experimental program on the strength developing law of
EAM. Firstly, the viscosities of epoxy asphalt binder (EAB) in different temperatures were
tested and the viscosity-time curve was constructed to investigate its developing law with
time and temperature. Then the EAM strength developing law with reserved time and curing
time in different temperatures were studied respectively using Marshall test from the view of
phenomenology. Finally, some important time and temperature points of EAM construction
control were concluded according to the researches and case studies were conducted to verify
the effectiveness of the conclusion.
KEYWORDS: epoxy asphalt mixture, strength developing law, viscosity-time curve,

construction control
1. Introduction
Epoxy Asphalt Mixture (EAM), mixed by epoxy asphalt binder (EAB) and high quality
aggregate, has been widely used in the pavement engineering for its excellent performance.
However, as a thermosetting material, EAM has been practically observed with different
performances compared to the common thermoplastic materials as HMA, especially on the
strength developing law, which affects the practical construction directly. Therefore, in order
to instruct the construction of EAM, researches on the strength developing law in the curing
process is needed.
As an irreversibly cure material, thermosetting material can be completed through heating

and chemical reactions, by which the curing process transforms the resin into a plastic or
rubber by a cross-linking process. Due to the added energy or catalysts, the molecular chains
react at chemically active sites and are linked into a rigid, 3-D structure. Theoretically, the
material will obtain a higher melting point when compared to the surrounding ambient
temperature because of the larger molecular weight caused by the cross-linking process.
However, uncontrolled reheating usually causes the material reaching the decomposition
temperature before the melting point is obtained. Therefore, a thermosetting material cannot
be melted and re-shaped after it is cured (Milton Keynes, 2000).
There have been many researches and studies being conducted to study the strength
mechanism of paving materials. The Transportation Materials Research Center (TMRC) of
the Michigan Technological University has researched the strength mechanism of the asphalt
mixtures using the finite element and discrete element methods (Dai Q, 2007; You Z, 2007;
Dai Q, 2006), Ranja Bandyopadhyay (Ranja Bandyopadhyay, 2007) and Martin H. Sadd
(M.H. Sadd, 2005) have also investigated the strength formation of asphalt mixtures from
microscopic view. However, these researches are all focusing on the thermoplastic material
such as HMA; few studies are conducted on the thermosetting material. The Intelligent
Transportation System Research Center of the Southeast University has taken a number of
researches on EAM in recent years (Huang, 2003; Luo, 2006; Chen, 2006), nevertheless, most
of the researches are aiming at the performances of EAM after curing, and further studies on
the strength developing law of EAM is still needed.
The former studies show that the performances of EAM are influenced significantly by
its reserved time and curing time. The reserved time is defined as the time from the mix of
asphalt and aggregate to the compaction of the mixture, an optimum reserved time will lead
to the best performance of EAM. Also, the strength developing law of EAM in different
temperatures will help to determine the curing time after practical EAM construction. So, in
this paper, an experimental program is put on to investigate the strength developing law of
EAM. Firstly, a viscosity test is conducted to investigate the viscosity developing law of EAB;
secondly, the performance of EAM under different reserved time and curing time in different
temperatures is studied to investigate the strength developing law of EAM; finally, the test
results are analyzed and case studies are conducted.
2. Materials and methods
2.1. Materials Preparation
Two main materials are involved in the test. The details of preparing for both materials and
specimen are described as follows.
2.1.1. Epoxy Asphalt
The binder used in the test is 2910-type local epoxy asphalt, which is composed of two
components marked as A and B. Component A is the epoxy resin while component B consists
of petroleum asphalt and curing agent. The basic information of the material is given in Table 1.
Table 1. Technical Index of 2910-Type Local Epoxy Asphalt
Technical Indexes Measured Value Criteria Test Method

Mass ratio（A∶B） 100:290 100:290
Tensile strength（MPa ，23℃） 3.26 ≥2.0 ASTM D 638
Fracture elongation（%，23℃） 242 ≥200 ASTM D 638
Duration of viscosity up to 1Pa·s（min） 110 ≥50 JTJ052-2000
2.1.2 Aggregate
Aggregates take about 93 to 94 percentage of the weight for the total mixture. Therefore,
careful considerations should be given to the selection of type and quality of the aggregates.
In this study, the basalt aggregate and the limestone powder special for steel bridge pavement
are selected based on the practical engineering, and the max aggregate diameter is 13.2mm.
The basic information of the aggregate is shown in Table2 and Table 3, the gradation curve for
the aggregate is shown in Figure 1.
Table 2. Technical Index of the Basalt Aggregate

Compressive strength (MPa) 140 ≥120 JTG E41-2005(T0221-2005)
Los Angeles abrasion value (%) 11.5 ≤22.0 JTG E42-2005(T0317-2005)
Crushing value (%) 8.9 ≤12 JTG E42-2005(T0316-2005)
1# 2.924
2# 2.983
Apparent density 3# 2.912
≥2.65 JTG E42-2005(T0304-2005)
(g/cm3) 4# 2.900
5# 2.967
6# 2.803
Water sbsorption 1# 0.93
≤1.5 JTG E42-2005(T0304-2005)
(%) 2# 1.22
Table 3. Technical Index of the Limestone Powder

Density (g/cm3) 2.703 ≥2.500 JTG E42-2005(T0352-2005)
Hydrophilic coefficient (%) 0.63 ≤1 JTG E42-2005(T0353-2005)
Plasticity index (%) 3.2 4.0 JTG E42-2005(T0354-2005)
Figure 1. Designed gradation in the test
2.2. Specimen Preparation
Considering the convenience of the specimen forming, the Marshall test is adopted to
investigate the performance of EAM, Marshall stabilities and air voids are also two main
indexes in the practical EAM construction. Based on the Marshall mixture design test, a 6.5%
binder content is determined as the optimum asphalt content, the Marshall mixtures are shaped
using impact compaction method, five replicates are prepared for each testing.
2.3. Experimental program
Viscosity of binders has closed relationship with performance of the mixtures, however,
there are many other clues that affect the performance of the mixture, so the strength
developing laws of EAM with reserved time and curing time are also studied by investigating
the Marshall stability of the curing specimens.
2.3.1 Viscosity test
110°C, 115°C, 120°C, 125°C and 130°C are selected as test temperatures according to the
practical construction temperatures. The viscosity of EAB is measured using a Brookfield
rotational dial viscometer with a rotating speed of 100 rpm (29#), then, the viscosity-time
curves are constructed to investigate the viscosity developing law in different temperatures.
2.3.2 Study on reserved time
The epoxy asphalt and the aggregates are mixed at 110°C, 115°C, 120°C, 125°C and 130
°C respectively, and the reserved times are selected as 30min, 50min, 70min, 90min and
110min. The stability and air void of each specimen is measured and studied.
2.3.3 Study on curing time
The Marshall mixtures are formed in the same condition at temperature 120°C and
reserved time 50min, then curing at the temperatures of 120°C, 125°C and 130°C respectively,
the Marshall stability and air void at different curing time are tested. Meanwhile, other
Marshall mixtures shaped in the same conditions are curing in the room temperature at about
23°C to simulate the actual condition in EAM construction. The Marshall tests are conducted
as well at different curing time to investigate the strength developing law of EAM with the
curing time.
3. Results and discussions
3.1. Viscosity test
The viscosity-time curves of EAB in different temperatures are shown in Figure 2.As
shown in Figure 2, the viscosity of EAB increased as the reaction time increasing, and the
viscosity increase slowly at the beginning of the test, while it increases faster and faster
with the time passing. It is also observed that the higher the test temperature is, the faster
the viscosity grows. This can be described using the dual-Arrhenius model (Min, 2006;
Min2007) as
(1)
where
= the viscosity,
= the Arrhenius Pre-factor,
= the flow activation energy,
Ea = the reaction activation energy,
R = the gas constant,
T = the absolute temperature,
t = the time, and A is the factor.
The Equation (1) presents that the viscosity of EAB is a function of time and temperature,
so the time and temperature construction process of EAM should be strictly controlled.
Figure 2. Viscosity-time curve of EAB
It can be observed in detail that before the viscosity of EAB reaches about 1000mPa·s,
the viscosity grows slowly, this goes against the initial strength formation of EAM; after the
viscosity of EAB reaches about 3000mPa·s, the viscosity grows rapidly, and this will affect
the rolling effect of EAM. So, it can be drown that the viscosity range that suitable to the
EAM construction is about 1000mPa·s to 3000mPa·s.
3.2. Study on reserved time
The tests results of EAM stability and air void are shown in Figure 3.The air void and
Marshall stability controlled indexes in the practical EMA construction are under 3.0 and over
40kN respectively. It can be observed from Figure 3(a) that the air void of EAM grows with
the reserved time increasing, when the test temperature is under 120°C, the air void increasing
slowly as the times passing while the air void increasing rapidly when the test temperature
is over 120°C, which make the air void difficult to control in rolling process. From Figure
3(b), it can be seen that there is an optimum reserved time corresponding to each temperature,
the stability at the temperatures of 120°C and 130°C are higher than others. Taking a
comprehensive consideration on the air void and stability, the optimum temperature of EAM
construction is 120°C, and the reserved time is no longer than 70min.
(a) Air void – reserved time curve

(b) Stability – reserved time curve
Figure 3. Experimental results of study on the reserved time
A further study on reserved time of EAM is conducted and the reserved times
corresponding to every temperature from 110°C to 130°C are concluded as shown in Table
4. The data in Table 4 indicates that there is a decrescent trend of reserved time as the
temperatures increasing and the shortest and the longest reserved time are both decreasing. It’s
also can be seen that while the temperatures rising, the reserved time range of EAM becomes
narrow, there is 48min for EAM construction at 110°C while it decreases to 25min when the
temperature is 130°C. According to the comprehensive considering of the laboratorial results
and the need of practical construction, it is better to construct EAM at the temperature from
115°C to 120°C.
Table 4. The reserved time range of different temperatures
Temperature Shortest Reserved Longest Reserved Reserved Time

(℃) Time(min) Time(min) Range (min)
110 42 90 48
111 42 87 45
112 41 84 43
113 40 82 42
114 39 79 40
115 38 76 38
116 37 75 38
117 35 74 39
118 33 72 39
119 32 71 39
120 30 70 40
121 30 69 39
Temperature Shortest Reserved Longest Reserved Reserved Time

(℃) Time(min) Time(min) Range (min)
122 30 68 38
123 30 67 37
124 30 66 36
125 30 65 35
126 30 63 33
127 30 61 31
128 30 59 29
129 30 57 27
130 30 55 25
3.3. Study on curing time
The tests results of EAM stability and air void in different curing times are shown in
Figure 4. Figure 4(a) presents the strength developing law of EAM in laboratorial curing
condition, it can be seen that the stability grows as the curing time increasing until it achieves
the maximum value, which means the curing reaction is finish. It also can be seen that curing
time is shorter when the curing temperature is higher, 5 hours at 130°C is the optimum
laboratorial curing condition.
Figure 4(b) shows the strength developing law of EAM at room temperature. It can be
observed that curing process in this condition is much slower than in laboratorial condition;
under this circumstance, the EAM need about 40 days to reach the stability of 40kN.
(a) EAM stability in the laboratorial curing conditions

(b) EAM stability at the room temperature
Figure 4. Experimental results of study on the curing time
4. Case study
4.1. Tianxingzhou Highway-railway Yangtze River Bridge
The Tianxingzhou Yangtze River Bridge is taken as a case to verify the key points
of EAM construction control which are concluded through the experimental study. The
Tianxingzhou Bridge is the first cable-stayed highway-railway bridge that contains 4 railway
lines in the world, the total length of the bridge is 766m with a main span of 504m. For the
excellent performance, EAM is selected as paving material of Tianxingzhou steel deck paving
engineering which is started in November 2008.
According to the results of laboratorial study, the construction temperature of EMA is

determined as 115°C to 120°C, and the reserved time range in Table 4 is abided. The core-
drilling test was conducted after curing for three days, the test results show that the average air
void of the core-drilling specimen is 2.0, less than the control standard 3.0, it proves that 115°
C to 120°C is a suitable condition for EAM construction.
For the construction season is later autumn, the average temperature during the
construction process is about 10°C, lower than the experimental temperature. So Marshall
specimens were formed and cured in the natural environment, the Marshall test is conducted
to monitor the strength develop of the EAM, the test result is presented in Figure5. It can be
observed from Figure 5 that the strength of EAM grows more slowly than the experimental
study for the low temperature, it takes over 100 days that when the stability reaches 40kN, and
the stability reaches maximum value after curing for about 140 days.
Figure 5. Strength developing law of EAM in Tianxingzhou Bridge
4.2. Shanghai Yangtze River Bridge
The Shanghai Yangtze River Bridge is taken as another case. The Shanghai Yangtze River
Bridge is a 1430m cable-stayed bridge with a main span of 730m, which is the main bridge
of crossing-river channels in Chongming of Shanghai. EAM is selected to be used in the steel
deck paving project, which is the first time of EMA use in paving project of Shanghai. The
project is started in June 2009, a hot summer in China.
As in the steel deck paving project of Tianxingzhou Bridge, the construction temperature
of EMA is determined as 115°C to 120°C, and the reserved time range in Table 4 is abided
also. The core-drilling test results show that the average air void of the core-drilling specimen
is 1.8, less than the control standard 3.0, it can be concluded again that 115°C to 120°C is
suitable for EAM construction.
During the construction of the EAM, Marshall specimens were formed and cured in the
natural environment. The average temperature during the construction term is about 25°
C, about 15°C higher than that of Tianxingzhou Bridge. The EAM Marshall test results of
Shanghai Bridge are shown in Figure6. It can be observed from Figure 5 that the strength of
EAM grows more quickly than that of Tianxingzhou Bridge, the stability of EAM reaches
40kN after only 18 days curing, about 90 days earlier than that in Tianxingzhou Bridge. It
indicates that the curing time is affected a lot by temperature and it is better to construct EAM
in hot summer.
Figure 6. Strength developing law of EAM in Shanghai Yangtze River Bridge

5. Conclusion
This paper presents an experimental program to study the strength developing law of
EAM, and conclusions can be drawn as follows:
1. The viscosity of EAB is a function of time and temperature, it increases with the
reaction time passing, and it grows quicker when the reaction temperature is higher. It
would be better to roll EAM when the viscosity of EAB is in the range of 1000mPa·s
to 3000mPa·s.
2. The air void grows with the revered time increasing, and the optimum condition for
EAM construction is 120°C, 70min according to the laboratorial study, which is also
verified in the steel deck paving engineering project of Tianxingzhou Highway-railway
Yangtze river Bridge and Shanghai Yangtze river Bridge.
3. The strength of EAM grows with the curing time going, and it reaches the maximum
value more rapidly in a higher temperature.130°C, 5h is suitable for the completely
curing of EAM in the laboratory.
4. The Phenomenological method is adopted to determine the curing time needed in the
natural curing condition, the results show that it takes only about 18 days curing when
the stability of EAM reaches 40kN in summer while the curing time is over 100days in
later autumn. So it is better to construct EAM in hot summer.
Acknowledgment
The authors would like to thank the financial support to this research from the Chinese
National Natural Science Foundation (No. 50878053) and Chinese New Century Excellent
Talents Program of Education Ministry.
Bibliography
Bandyopadhyaya R. et al. “Numerical simulation of mechanical behavior of asphalt mix”.

Constr Build Mater, Vol. 23, 2007, pp.1-8.
Chen T. J. “Research on Crack Behavior of Long-span Steel Bridges Deck Epoxy Asphalt
Pavement”, Southeast University(CHN), 2006.
Dai Q., Z. You, and M.H. Sadd. “A Micromechanical Visco-elastoplastic Model for Asphalt
Mixture”, Geotechnical Special Publication (GSP): Asphalt Concrete: Simulation,
Modeling, and Experimental Characterization, American Society of Civil Engineers
(ASCE), 2006, pp. 12-20.
Dai Q. and You Z, “Prediction of Creep Stiffness of Asphalt Mixture with Micromechanical
Finite Element and Discrete Element Models”, Journal of Engineering Mechanics,
American Society of Civil Engineers (ASCE), Vol. 133 No.2, 2007, pp. 163-173.
Huang W., Z. D. Qian and G. Cheng, et al. “Epoxy asphalt concrete paving on the deck of
long-span steel bridges”. Chinese Science Bulletin, Vol. 48 No. 21, 2003, pp. 2391- 2394.
Luo S, J.W. Wang, and Z.D. Qian. “Research on the performance of locally developed epoxy
asphalt mixture”, Proceedings of the 26TH Annual Southern African Transport Conference,
July 2007.
M.H. Sadd and Q.Dai. “A comparison of micro-mechanical modeling of asphalt materials

using finite elements and doublet mechanics”. Mechanics of Materials. Vol. 37, 2005, pp.
641–662.
Milton Keynes. T838 “Design and Manufacture with Polymers: Introduction to Polymers”,
The Open University (UK), 2000, pp. 9.
Min Z. H. and W. Huang. “Study on viscosity and construction property of epoxy asphalt”.
Journal of Highway and Transportation Research and Development, Vol. 23 No. 8, 2006,
pp 45-48. (in Chinese)
Min Z. H., Z. J. Zhang, Z.D. Qian and W. Huang. “Dependency of time-temperature of epoxy
asphalt mixture strength”. China Journal of Highway and Transport, Vol. 20 No.3, 2007,
pp. 1-4. (in Chinese)
You Z. and Dai Q. “A Review of Advances in Micromechanical Modeling of Aggregate-

Aggregate Interaction in Asphalt Mixture”. Canadian Journal of Civil Engineering /Rev.
can. génie civ. Vol. 34 No. 2, 2007, pp. 239-252.
Asphalt Rubber: policy disclosure in Italy
F. Canestrari* - E. Pasquini* - F. A. Santagata* - I. Antunes**
*Università Politecnica delle Marche

Dipartimento ISAC – Sezione Strade
Via Brecce Bianche
60131 Ancona, Italy
f.canestrari@univpm.it
e.pasquini@univpm.it
f.a.santagata@univpm.it
** Asphalt Rubber Italia Srl

Via Giusti 67
51017 Pescia, Italy
ines.antunes@asphaltrubberitalia.it
ABSTRACT. In the present paper, it is presented a proposal for an evaluation systemic

approach aimed at the characterization of Asphalt Rubber Hot Mixes (ARHM). This
constituted the input to develop a careful scientific-technical assessment process to assure
the quality control of ARHM. In particular, this kind of approach allows i) to take immediate
action to technical production and laying of ARHM that should also guarantee “traditional
” Italian road materials specifications ii) to develop AR expertise in terms of mechanical
performance (stiffness, fatigue resistance and permanent deformation resistance) and
functional properties (noise and skid resistance), trying to bridge the gap of knowledge with
the international experience of forty years of successful use iii) to develop innovative scientific
research to provide original contributions to engineering road materials investigation.
KEYWORDS:Evaluation systemic approach, Asphalt Rubber Hot Mixes, Wet process

1. Introduction
Asphalt Rubber (AR) is a mixture of hot asphalt and crumb rubber derived from waste
scrap tires (ASTM D-6114). This type of blending process is called “wet”. It was invented in
the United States in the State of Arizona in the late 1960’s. AR is commonly used as binder
for wearing course hot mix asphalts in order to improve smoothness and skid resistance and
to reduce cracking and traffic noise. While in many countries all over the world it is used
extensively and successfully in the highway paving industry, in Italy AR technology is quite
new and first applications were made in 2006 (Santagata et al., 2008).
The study presented in this paper came about as a result of the ever-growing demand for
high performances in the field of road constructions, which lead investigation into asphalt
binder properties, surpassing the current traditional empiricist approach in favor of a more
specific characterization. Particularly, the interest in developing this scientific investigation is
aimed at driving the worth of industrial utilization of asphalt binders over traditional asphalts,
and systemizing its classification, based on the most innovative specifications. Adding to all
this, there is also an environmental concern linked to the possibilities of re-using a significant
amount of reclaimed tires given by Asphalt Rubber Hot Mixes (ARHM).
This paper focuses on an evaluation systemic approach aimed at the characterization of

ARHM, describing the first steps adopted to have Asphalt Rubber “made in” Italy: a global
strategy which strongly involved the, at the moment, unique Italian producer of Asphalt
Rubber binder (Asphalt Rubber Italia), CIRS (Experimental Interuniversity Road and Airport
Research Centre) representing the most important scientific entities in Italy and some local
agencies.
2. Systemic approach
2.1. Formulation of Asphalt Rubber binder
Experience has shown that by properly combining the waste product of ground tire rubber
(Crumb Rubber Modifier - CRM) with asphalt at high temperatures the resultant Asphalt
Rubber binder will have many improved superior engineering properties. Such improved
engineering properties include reduced fatigue and reflection cracking, greater resistance to
rutting, improved aging and oxidation resistance and better chip retention due to thicker binder
films (Partl et al., 2009; Santagata et al., 2007; Souza et al., 2005; Zborowski et al., 2004;
Potgieter et al., 2002; Kaloush et al., 2003; Cook et al., 2006; Kumar et al., 2005; Bertollo
et al., 2004). Also Asphalt Rubber pavements have demonstrated to have lower maintenance
costs (Way, 2000; Jung et al., 2002), lower noise generation (Pasquini, 2009; Antunes et
al., 2006a; Leung et al., 2006), higher skid resistance and better night-time visibility due to
contrast in the pavement and stripping (Antunes et al., 2005).
Asphalt Rubber binders have specific and unique characteristics. Several studies have
underlined that the enhanced performance depends on the higher percentage of asphalt rubber
within the hot mix and on the percentage of crumb rubber used in the asphalt rubber binder
prepared by the wet process (Way, 2003).
Conventional binder tests have shown how the CRM modified binders are extremely
dependent on the processing conditions, particularly to what concerns the temperature and
time of reaction. The time required to disperse, blend and react the crumb rubber is dependent
on a number of factors including the chemistry of the asphalt cement and crumb rubber as
well as the temperature of the blended material. The same tests put in evidence as the content
of CRM is determinant in binder’s consistency and elastic behavior (Antunes et al., 2004;
Antunes et al., 2006b).
The main effort to manage the production of AR in Italy consisted in adapting the
consolidated AR technology knowledge from abroad to the Italian reality (raw materials,
weather, environment, specifications, etc), starting by mixing several bitumen-base (virgin
asphalt) with several type and size of crumb rubber, before finding the “optimal” blend to be
tested in an industrial scale.
In laboratory, several virgin asphalt (penetration classes: 40/50, 50/70 and 70/100) and
high percentages of crumb rubber (higher than 15%) were blended into a homogeneous
asphalt-rubber system at a temperature of 190°C, which then reacted at 180°C for a minimum
of 45 minutes. Then, the blend characteristics were compared with the specification required
by ASTM D-6114 which defines Asphalt Rubber binder (Table 1).
Table 1. Requirements for Asphalt Rubber binder design (ASTM D 6114)
Tests Requirements
Viscosity Brookfield @175ºC: cP (ASTM D2196) Min. 1500 Max 5000
Penetration @25ºC: 1/10 mm (ASTM D5) Min. 25 Max 75
Resilience @25ºC: % (ASTM D 5329) Min. 20
Softening Point, Ring&Ball: ºC (ASTM D36) Min. 54.5
These limits along with the results attained for each percentage of crumb rubber blended
into the bitumen allow the determination of the “optimal” content of crumb rubber, which is
usually the percentage that improves the softening point temperature, resiliency and ductility.
The results show that increasing the CRM percentage corresponds to increasing the
softening point, resilience and viscosity values as depicted in Figures 1 and 2.
Variation of Viscosity with Softening Point Temperature for AR binders

85
M in. ASTM D 6114

80 40/50 + 18%
Softening Point Tem perature [°C]
CRM
75 70/100 A
50/70 B
70
M ax. AST M D 6114

65 70/100 A + 15% CRM
70/100 A + 16% CRM
60 70/100 A + 18% CRM
Min. ASTM D6114 - 60°C 70/100 A + 20% CRM
70/100 A + 22% CRM
55
50/70 B + 15% CRM
50/70 B + 16% CRM
50 50/70 B + 18% CRM
500 1000 1500 2000 2500 3000 3500 4000 4500 B + 20% 5000
50/70 CRM
Viscosity @175°C [cP] 40/50 + 18% CRM
Figure 1. Variation of viscosity with the softening point for Asphalt Rubber with different
percentage of CRM
Variation of Resilience with Softening Point Temperature for AR binders

80
M in. A STM D 6114 -
40/50 + 18%
75 CRM 70/100 A
Softening Point T emperature [°C ]
15%
20%
70 50/70 B
70/100 A + 15% CRM

65 70/100 A + 16% CRM
70/100 A + 18% CRM
Min. ASTM D6114 - 60°C
70/100 A + 20% CRM
60 70/100 A + 22% CRM
50/70 B + 15% CRM
50/70 B + 16% CRM
55 50/70 B + 18% CRM
50/70 B + 20% CRM
40/50 + 18% CRM
50
0 10 20 30 40 50 60 70
Resilience @25°C [%]
Figure 2. Variation of resilience with softening point for Asphalt Rubber with different
percentage of CRM
It can be seen that some asphalts are more compatible with CRM than others and thus
binder design pre-testing is necessary to determine that the proper asphalt grade is being used
to formulate the asphalt rubber.
Taking in consideration the results above and the Italian weather and ambient contest, three
blends are now commonly produced at the AR industrial producing plant: 40/50 with 18%
CRM, 50/70 with 20% CRM and 70/100 with 20% CRM, depending on the use.
2.2. Characterization of Asphalt Rubber mixes
Following the Arizona Department Of Transportation (ADOT) specification, in Italy, the

AR producer began experimenting with two asphalt rubber mixes: an open graded mixture
(ARFC) and a gap graded mixture (ARAC) (Scofield, 1989). To fully utilize AR properties,
two aggregate gradations, that would provide a high voids in the mineral aggregate (VMA),
have been adopted as shown in Table 2.
Table 2. Italian ARFC (Open Graded) and ARAC (Gap Graded) Gradations
AR MIX Open Graded ARFC Gap Graded ARAC

Sieves (mm) Min Max Min Max
16 100 100 100 100
12.5 93 100 83 97
10 88 100 68 82
8 68 82 54 68
4 23 37 25 37
2 3 15 12 24
0.5 2 10 7 15
0.063 0 3 0 3
AR Binder 8.5% 9.5% 7.5% 8.5%
To what concerns the study of ARHM, the adopted approach consisted in developing
theoretical and experimental parallel activities, with the main purpose of:
- to disclose experimental findings to encourage the use of ARHM in the Italian road
network in the awareness of potential technical, environmental and economic benefits;
- to develop innovative and relevant topics to provide original contributions to scientific

research.
In particular, it was provided technical support during the production of asphalt rubber
mixes to assure that ARHM respect Italian “traditional” standard specifications. This support
was divided in time by monitoring the characteristics of several materials produced and
implemented for many applications on the road in recent years. The intent is to gather as much
information as possible to set up a large database, which allows the establishment of real
performance-type specifications.
At the same time, advanced research on AR materials through innovative test methods
have been developed.
3. Operative support for road applications
In this chapter the technical activities conducted within the routine characterization of
asphalt rubber mixes are illustrated. This activity is geared to provide immediate answers to
ARHM producers and potential users of these innovative materials in the Italian market. In
that sense several productions of ARHM were analyzed in terms of composition (aggregate
gradation and AR binder content), volumetric properties (void content) and mechanical
performance (Marshall Stability, indirect tensile strength, loss of strength after immersion in
water).
As an example, some results obtained during the characterization of AR gap graded –

ARAC (Figures 3, 4, 5, and 6) and AR open graded – ARFC (Figures 7, 8, 9, 10) are shown.
Aspahlt Rubber mix GAP GRADED
9.00
8.00
7.00
6.00
Air voids (%)
5.00
4.00
3.00
2.00
1.00
0.00
_1
_2
_3
_4
_5
_6
_7
_8
_9
8
_1
_1
_1
_1
_1
_1
_1
_1
_1
50
50
50
50
50
50
50
50
50
50
50
50
50
50
50
50
50
50
M
Sample
Figure 3. Volumetric properties of a Gap Graded AR mix
Asphalt Rubber mix GAP GRADED

15.00
Marshall Stability (kN)
12.00
9.00
6.00
3.00
0.00
M50_1 M50_2 M50_3 M50_4 M50_5 M50_6
Sample
Figure 4. Marshall Stability of a Gap Graded AR mix

4.00
Marshall quotient (kN/mm)

3.50
3.00
2.50
2.00
1.50
1.00
0.50
0.00
M50_1 M50_2 M50_3 M50_4 M50_5 M50_6
Sample
Figure 5. Marshall Quotient of a Gap Graded AR mix

Loss of Marshall Stability after
30
15 days of immersion (%)
25
20
15
10
5
0
M50_7 M50_8 M50_9 M50_10 M50_11 M50_12
Sample
Figure 6. Loss of Marshall Stability of a Gap Graded AR mix
Obviously, the results presented refer only to two specific materials, gap and open graded,
respectively, packaged with a particular compaction methodology and compaction energy. In
that sense, this results represent only a model of the type of data obtained during this phase of
the study. In particular, it is possible to notice the good repeatability of results for each tested
material. These studies were repeated systematically over the years on the production in order
to verify also the quality of the process.
Asphalt Rubber mix OPEN GRADED
20.00
18.00
16.00
14.00
Air voids (%)
12.00
10.00
8.00
6.00
4.00
2.00
0.00
AR_1 AR_2 AR_3 AR_4 AR_5 AR_6 AR_7 AR_8 AR_9 AR_10
Sample
Figure 7. Volumetric properties of an Open Graded AR mix

15.00
Marshall Stability (kN)
12.00
9.00
6.00
3.00
0.00
Md_2 Md_3 Md_5 Md_6 Md_7 Md_9 Md_10
Sample
Figure 8. Marshall Stability of an Open Graded AR mix
4.00
Marshall quotient (kN/mm)
3.50
3.00
2.50
2.00
1.50
1.00
0.50
0.00
Md_2 Md_3 Md_5 Md_6 Md_7 Md_9 Md_10
Sample
Figure 9. Marshall Quotient of an Open Graded AR mix
30
Loos of Marshall Stability after
15 days of immersion (%)
25
20
15
10
0
Mw_2 Mw_3 Mw_5 Mw_6 Mw_8 Mw_9
Sample
Figure 10. Loss of Marshall Stability of an Open Graded AR mix
The detailed and rigorous monitoring of these requirements allowed making a continuous
optimization of ARHM and also AR binders. In addition, this extensive data enable to make
a preliminary draft of a specification for ARHM following both Italian standard specification
for “traditional” road materials and international AR requirements. Table 3 is an example of
adaptation of different standards of specifications to develop AR gap graded (ARAC) mixture
optimized taking into account the specificities of the Italian reality.
Italian specifications, even those more advanced, follow the current traditional empiricist
approach and only provide a marginal concept of performance of a material (such as stiffness,
fatigue and permanent deformation). This fact is mainly due to the lack of a database that has
a significance level and statistical feedback in terms of real performance so it is not always
possible to prescribe certain standards. Producers are not familiar with non empirical tests and
it is often the same thing to for Agencies. The introduction of ARHM in the Italian market
opened a new scientific approach to road materials and, with the technical and scientific
support of CIRS, the first Standard Specifications for Italian ARAC and ARFC were made in
2007. These specifications are constantly changing since the monitoring of AR materials in
laboratory and under real traffic flows is still in progress.
Table 3. Italian Traditional requirements for ARAC (Gap-Graded)
Requirements
Test condition Test method Unit
Binder course Wearing course
Compaction EN 12697-34 n° of blows 50 50
Requirements
Marshall Stability EN 12697-34 kN >9 >9
Marshall quotient EN 12697-34 kN/mm 1,5÷3,0 1,5÷3,0
Air voids EN 12697-8 % 5–8 5–8
Loss of Marshall Stability
CNR n. 149/92 % ≤25 ≤25
after 15 days of immersion
Nowadays, it is considered that for an objective evaluation of the performance of road

materials, specifications should refer to performance-related test methods that can effectively
evaluate the real potential of the material. In the near future, therefore, the aim is that
these rules begin to be understood as a requirement and replace those that are traditionally
and empirically used. In this sense, it is extremely important to introduce the concept of
“performance” of a material, which includes its mechanical and functional characterization.
In this light, the proposed specifications require to evaluate certain performance parameters
of the ARHM taken from the site and compacted in the laboratory until reaching the voids
produced in situ. These parameters must be representative of the visco-elastic properties and
the cracking resistance of the mixes, according to EN 12697 series, in order to provide useful
numerical elements for the design and reference values for non-destructive in situ testing.
4. Advanced performance characterization
As already introduced, there are many studies in the international literature focused on the
mechanical performance (stiffness, fatigue and rutting) and functional properties (noise and
friction) of ARHM. In order to bridge the gap of knowledge with the international experience,
the same kind of approach was applied in Italy, and the performance achieved by these kind of
materials was assessed experimentally through standardized testing protocols. This choice is
considered appropriate in consideration of local particularities mainly related to the different
characteristics of raw materials (bitumen and aggregates).
The laboratory and in situ collected data for AR mixes and the comparison with more
traditional materials will allow the drawing up performance-related specifications with
obvious impact to the quality of road infrastructure.
The advanced characterization concerns two main aspects: mechanical and functional
properties.
4.1. Mechanical properties
The mechanical behavior of AR materials have been analyzed by means of the evaluation
of three main properties: stiffness modulus, permanent deformation and fatigue behavior.
Several productions of ARHM have been analyzed from this point of view and the study is
still ongoing. As an example, the Authors show the results obtained during the performance
characterization of some ARHM.
Materials under study were first characterized in terms of stiffness performance. In Figure
11 it is possible to observe the results for some of the investigated materials. These results
were determined by means of indirect tensile configuration tests, according to European EN
12697-24, at 20 ℃.
3500
3000 3297
Mean Stiffness Modulus (MPa)
2500
2371
2000
1963
1500
1000
500
0
GAP OPEN_1 OPEN_2
Figure 11. Indirect Tensile Stiffness Modulus of several AR mixtures @ 20°C

It should be noted that the ARAC (Gap Graded) mix is characterized by an air void
content of about 6.5% while ARFC (Open Graded) mix was optimized in order to reach
values between 15% and 20%. Bearing in mind also the great amount of binder that is used
in ARHM, these mixes are characterized by a significantly high VMA (Voids in Mineral
Aggregate) that further penalize the stiffness properties of ARHM. In this sense, the
properties of the AR binder shall ensure that the entire ARHM can still have good rheological
performance.
In particular, ARFC (open) shows obviously lower stiffness due to the higher percentage
of voids. However, stiffness modulus values of Open Graded ARHM are always greater than
those observed for conventional porous asphalts (Canestrari et al., 2007), despite the higher
binder content and VMA. This aspect confirms the relevant rheological properties of asphalt
rubber binder.
Fatigue @ 20°C
1000
AR OPEN
initial microstrainε
POROUS
100
1000 10000 100000 1000000
cycles to failure N
Figure 12. Fatigue behavior of an ARFC Open Graded AR mixture
For each tested material, the fatigue behavior was also analyzed on the basis of dynamic
tests in indirect tensile configuration according to EN 12697-24. As an example, Figure 12
represents the fatigue curve of the ARFC (OPEN_1) previously introduced. The results are
compared with those obtained in the case of a conventional porous hot mix (void content =
20%) prepared with SBS modified bituminous binder.
Even though porous layers provide a limited contribution to the fatigue resistance of the
whole pavement, the results presented above clear indicate that the mix made with Asphalt
Rubber binder offers significantly better performance than a material not prepared with AR
binder and having similar volumetric properties. This fact seems to confirm the remarkable
properties of asphalt rubber binder.
The advanced systemic approach described in this chapter finally provides the evaluation
of rutting resistance. This property has been assessed by rutting tests performed according to
BS 598:110 at 60 ℃. This evaluation is essential because asphalt rubber mixtures, having high
binder contents, are potentially more prone to permanent deformation-related distresses. The
assessments carried out on different materials have had a huge success: even in comparison
with referenced materials with strong rutting resistance (Canestrari et al., 2007, Santagata
et al., 2007), ARHMs have showed very high rutting resistance. Indeed, the deformation
obtained after several hours test duration was found to be negligible (less than 1 mm), putting
in evidence, once again, the valuable rheological properties that AR binder confers to the
ARHM (Figure 13).
WTT @ 60°C
AR GAP
1
AR OPEN
0.9
SMA
0.8
POROUS
0.7
rut depth (mm)
0.6
0.5
0.4
0.3
0.2
0.1
0
0 60 120 180 240 300 360 420 480 540
time (min)
Figure 13. Rutting behavior of AR mixtures
4.2. Functional properties
In the evaluation of performance of AR mixtures, it is essential to monitor the also

functional properties (friction and noise) by paying special attention to the evaluation of
rolling noise reduction. This problem is nowadays an increasingly strong issue that often
affects the choice of materials for road pavement wearing courses.
The evaluation of functional properties have been done both in situ, through the
measurement of surface characteristics and the evaluation of noise levels along the side of
the roadway, and in laboratory, through the determination of sound-absorption properties of
materials studied.
As an example, it is interesting to show the results obtained during the acoustic

characterization of ARHM, both gap and open graded, in two different trial sections.
4.2.1. Trial sections
A Gap Graded and an Open Graded AR mixes were laid down on an urban trial section
about 700 m long in Florence. The thickness of both wearing courses was 30 mm. The section
covered with ARAC (gap graded) is composed of two lanes (one for ordinary traffic and one
reserved for buses) while ARFC (open graded) trial section has only one lane.
Another trial section realized with a Gap Graded AR wearing course was constructed on
an urban way about 1000 m long in Imola. Asphalt Rubber material was staggered with some
short sections of a traditional HMA wearing course selected as reference surface for traffic
noise survey. The thickness of the new ARAC wearing course was 30 mm.
4.2.2. In situ acoustic analysis
Noise levels recorded along the side of sections constructed with the studied materials
were related with noise levels recorded at the same time on near different sections covered
with traditional asphalt concretes and interested by quite the same traffic load.
The acoustic characterization of materials under study was carried out performing noise
measurements surveys, each lasting one week, according to the Italian technical specifications.
Measurement boxes, according to Class 1 of EN 60651 and EN 60804, were fixed at about 4
m high to lampposts situated along the side of the experimental roadway (Figure 14).
Measurement boxes were able to determine the A-weighted equivalent sound level Leq for
each measuring hour and the results were summarized in one mean A-weighted sound pressure
level for night (Leq,wn) and day (Leq,wd) periods, as shown in the example of Figure 15.
In order to make the comparative study of the “in situ” acoustic measurement objective,
it was supported by traffic investigation. Each investigation was carried out for 24 hours
once a week, during the traffic noise finding weeks, recording the number of vehicles and the
corresponding speed and length.
Figure 14. Measurements boxes

90
Leq,day
80
Leq,night
Leq,wn =56,3 dBA Leq,wd =63,2 dBA
70
60
Leq (dBA)
50
40
30
20
10
0
08/10/2007 09/10/2007 10/10/2007 11/10/2007 12/10/2007 13/10/2007 14/10/2007 15/10/2007
Measuring days
In situ noise measurements (one week)
Figure 15. In situ noise measurements (one week)
Observing the experimental results obtained for the trial section of Florence (Table 4), it
could seem that noise level recorded for the reference material was not comparable with that
of asphalt rubber mixtures because of the not negligible difference of traffic flow and speed
between the different sections. But actually, the measurement box corresponding to reference
material was placed at a greater distance from traffic stream than those mounted along the
asphalt rubber experimental sections in such a way that the distance counterbalanced the
different traffic conditions. In fact, it is possible to estimate through, for example, the Italian
CNR prediction model (Canale et al., 1986) that the greater noise level recordable for the
reference material due to the higher traffic flow and speed is roughly counterbalanced by the
greater distance between measurement box and traffic stream.
Table 4. “In situ” acoustic results – Florence trial section
ARAC (gap) ARFC (open) Reference

Leq,wd (dBA) 65.1 63.2 67.9
Leq,wn (dBA) 57.9 56.3 61.5
Traffic (vehicles/day) 6694 5656 8967
Mean speed, day (km/h) 37.7 40.8 48.3
Mean speed, night (km/h) 42.0 44.9 58.3
Heavy vehicles (%) 10.16 2.05 3.51
Thus, ARAC mix proved to be about 3 dB(A) quieter than a traditional dense graded
asphalt concrete principally thanks to the asphalt rubber binder employed. It has to keep in
mind that 3 dB(A) noise reduction corresponds to halving traffic flow or doubling the distance
between the source and the receiver.
A further 2 dB(A) noise reduction was demonstrated to be achieved through the

employment of open graded AR mixture that is able to combine acoustic benefits arising from
AR binder with those obtainable from high air void content in terms of sound absorption
capabilities. Moreover, this mix is characterized also by a reduced maximum chipping size
that further enhanced rolling noise reduction properties.
Results obtained from in situ acoustic and traffic investigations for the trial section of
Imola are presented in Table 5.
Table 5. “In situ” acoustic results – Imola trial section
ARAC (gap) Reference

Leq,wd (dBA) 66.5 72.7
Leq,wn (dBA) 60.4 66.4
Traffic (vehicles/day) 9305 12396
Mean speed, day (km/h) 74.3 74.2
Mean speed, night (km/h) 77.5 77.3
Heavy vehicles (%) 5.52 5.53
Differently to what happened for the previous case, the noise pressure levels recorded were
not directly comparable. In fact, the two sections were characterized by the same mean vehicle
speed and heavy vehicle content but different total traffic flow, 25% lower for ARAC material
owing to an intermediate intersection.
Moreover, the sound measurement box corresponding to ARAC section had to be placed
at a distance from traffic stream (Figure 16 on the right) sensibly greater than that of reference
material road section (Figure 16 on the left).
Thus, in the same way as the previous situation, it is possible to estimate that the difference
between noise level in ARAC section and in reference material section has to be reduced of
about 3 dB(A) considering these two aspects. Taking into account the previous considerations,
ARAC proved again to be about 3 dB(A) quieter than a traditional dense graded asphalt
concrete having quite the same void content and maximum chipping size. As a consequence,
this noise reduction, corresponding to halving the traffic flow, has to be imputed to the use of
the asphalt rubber binder that confers more elastic properties to the bituminous mixture.
Figure 16. Measurement boxes position – Imola trial section
4.2.3. Laboratory acoustic results
The acoustic laboratory experimental program consisted in the determination of the sound
absorption coefficient α of the investigated materials by means of the impedance tube (Figure
17) according to the EN ISO 10534-1.
Samples subjected to sound absorption investigation were 95 mm diameter cylindrical

specimens cored from slabs prepared in laboratory with the Roller Compactor (EN
12607-33). Slabs were prepared with materials taken during the construction of the respective
experimental road sections. The thickness and air void content of the slabs were preferably
chosen equal to those obtained in situ.
Each sample was subjected to 7 different test frequencies: 400, 500, 630, 800, 1000, 1250
and 1600 Hz.
Figure 17. Sound absorption test setup
The Open Graded mixture showed a not pronounced sound absorption coefficient
notwithstanding the 14% air void content (Figure 18). This fact seems to confirm that a void
content lower than 15% is not able to guarantee good absorption properties because the
pores are probably not totally interconnected between them. Moreover, the selected reduced
maximum chipping size probably enhanced the air flow resistance of the material limiting
sound absorption characteristics.
SOUND ABSORPTION COEFFICIENT α

Frequency (Hz)
Specimen
400 500 630 800 1000 1250 1600
OG-ARsitu_1 0.17 0.29 0.38 0.38 0.23 0.10 0.07
OG-ARsitu_2 0.10 0.18 0.34 0.31 0.25 0.05 0.05
OG-ARsitu_3 0.19 0.20 0.35 0.32 0.14 0.03 0.04
OG-ARsitu_4 0.11 0.23 0.45 0.31 0.15 0.03 0.03
OG-ARsitu_5 0.16 0.25 0.34 0.24 0.09 0.05 0.06
mean values 0.15 0.23 0.37 0.31 0.17 0.05 0.05
OG-ARsitu Absorption Coefficients

1.0
0.9 OG-ARsitu_1
Absorption Coefficient α
0.8 OG-ARsitu_2
Voids=14% OG-ARsitu_3
0.7
0.6 OG-ARsitu_4
0.5 OG-ARsitu_5
0.4 mean values
0.3
0.2
0.1
0.0
0 200 400 600 800 1000 1200 1400 1600 1800
Frequency (Hz)
Figure 18. Absorption coefficients of Open Graded AR mix
Furthermore, it is very interesting to note that ARFC – Open Graded bituminous mixtures
demonstrated a low peak frequency of absorption corresponding to 630 Hz (Figure 18).
According to (Sandberg et al., 2002), this may be due to the higher tortuosity, i.e. a pores-
shape parameter, that arise from the reduced maximum chipping size coupled with to the high
binder content that created narrow channels linking up pores.
As a matter of fact, the not elevated sound absorption coefficients recorded for the ARFC
material indirectly proved that the very good anti-noise performance demonstrated in situ by
this mixture is principally due to the acoustic properties, in terms of reduction of rolling noise
generation, arising from the adding of asphalt rubber binder.
Finally, low absorption coefficients were showed by the Gap Graded AR mixture (Figure
19). This fact proved once again that the asphalt rubber binder is the main author of the
acoustic benefits demonstrated by this kind of mixture .
SOUND ABSORPTION COEFFICIENT α

Frequency (Hz)
Specimen
400 500 630 800 1000 1250 1600
GG-AR_1 0.11 0.14 0.14 0.25 0.10 0.04 0.03
GG-AR_2 0.13 0.18 0.12 0.29 0.02 0.07 0.09
GG-AR_3 0.10 0.15 0.18 0.20 0.11 0.01 0.04
GG-AR_4 0.08 0.12 0.18 0.29 0.08 0.05 0.04
GG-AR_5 0.12 0.18 0.20 0.20 0.06 0.05 0.02
mean values 0.11 0.15 0.16 0.25 0.07 0.04 0.04
GG-AR Absorption Coefficients

1.0
0.9 GG-AR_1
Absorption Coefficient α
0.8 GG-AR_2
0.7
Voids=6% GG-AR_3
0.6 GG-AR_4
0.5 GG-AR_5
0.4 mean values
0.3
0.2
0.1
0.0
0 200 400 600 800 1000 1200 1400 1600 1800
Frequency (Hz)
Figure 19. Absorption coefficients of Gap Graded AR mix
5. Innovative researches
The characterization of a road material should be completed with the development of

advanced studies that will provide original contributions to scientific research. In this regard,
several research items have been developed in the recent years, aimed at the evaluation of
specific performance-related properties through innovative test methods.
As an example, it is possible to mention studies on the durability of Asphalt Rubber

mixes with particular interest to water damage. This aspect is being investigated through two
innovative testing methods (Pasquini, 2009, Partl et al., 2009, Santagata et al., 2009): the
CAST, implemented in the EMPA laboratories; and the PATTI test, specifically modified for
the determination of the adhesion and cohesion properties of bitumen-aggregate system.
In particular, CAST equipment allows testing of laterally deformation constrained donut

shaped specimens under simultaneous action of cyclic mechanical loading, temperature
changes and water exposure. Tests are performed in a conventional, temperature controlled,
servo-hydraulic tension-compression machine. (Figure 20).
As an example, Figure 21 shows a typical output of a strain-controlled fatigue test

performed at 10 Hz and with temperature cycles. The complex modulus E* and phase angle of
the material are calculated based on a finite element model of the test.
Figure 20. CAST setup for dry test (left) and wet test (right)
1 Modulus Phase Temperature 60
55
Temperature (°C) - Phase angle (°)

0.8
50
Normalized Modulus
0.6
45
40
0.4
35
0.2
30
0 25
0 180000 360000 540000 720000 900000 1080000 1260000 1440000
Number of cycles
Figure 21. CAST fatigue test experimental output
The CAST results demonstrated that open graded mixtures with asphalt rubber binder
had not only superior fatigue resistance but also significantly reduced moisture sensitivity as
compared to traditional porous or semi-porous asphalt mixtures.
6. Conclusions
Progress in the use of recycled materials has been appreciable in the highway community
over the last 20 years. However, further development is dependent on more cooperation
among various disciplines: industry and government, highway engineers and environmental
specialists.
In this paper, the Authors presented a complete proposal for an evaluation systemic
approach aimed at the characterization of Asphalt Rubber mixtures (ARHM) and useful for
areas where the Asphalt Rubber wet process technology has recently been introduced.
The selected approach provides parallel activities that give immediate answers for ARHM
producers and applicators but that are also able to develop relevant scientific research.
The results so far obtained, and in part presented in this paper, are constantly in progress
as there is also a continuous research for mixture production optimization. In any case, they
clearly showed the enhanced performance of Asphalt Rubber mixtures in terms of mechanical
and functional properties.
Despite the huge literature focusing on Asphalt Rubber study, this work shows how it is
still possible and essential to enhance research addressing specific issues in order to implement
and update the knowledge to date consolidated.
Acknowledgements
Part of this work is based on research activities funded by Asphalt Rubber Italia S.r.l. (Italy)
7. Bibliography
Antunes I., Giuliani F., Sousa J.B., “Caracterização Reológica de Betumes Modificados com
CRM Sujeitos a Envelhecimento Artificial”, 1st European Road Congress, Lisbona, 2004.
Antunes I., Giuliani F., Sousa J.B., Way G., “Asphalt Rubber: Il Bitume Modificato con
Polverino di Gomma di Pneumatico Riciclata”, Varirei - V International Congress of
Valorization and Recycling of Industrial Waste, L’Aquila, 2005.
Antunes I., Way G.B., Sousa J., Kaloush K., “The Successful World Wide Use of Asphalt
Rubber”, 16th Convegno Nazionale SIIV, Cosenza, 2006a.
Antunes I., Giuliani F., Sousa J.B., “Chemical Modification of Asphalt Rubber with
Polyphosphoric Acid” Asphalt Rubber 2006 Conference, San Diego, 2006b.
Bertollo S.M., Bernucci L.B., Fernandes J.L., Leite L.M., “Mechanical Properties of Asphalt
Mixtures Using Recycled Tire Rubber Produced in Brazil – A Laboratory Evaluation”, 83rd
TRB Annual Meeting, Washington, 2004.
Canale S., Ventura F., “Rumore da traffico e progettazione stradale – metodologie di

previsione e possibilità di attenuazione”, Autostrade, Vol. 2, 1986.
Canestrari F., Bocci M., Ferrotti G., Pasquini E., “Mechanical characterization
of environmentally friendly mixtures” International Conference on Advanced
Characterization of Pavement and Soil Engineering Materials, Athens, 2007.
Cook M.C., Bressette T., Holikatti S., Zhou H., Hicks R.G., “A Full Scale Experiment on
Asphalt Rubber Modified Pavements in California”, 10th International Conference on
Asphalt Pavements, Quebec City, 2006.
Jung J., Kaloush K.E., Way G.B. «Life Cycle Cost Analysis: Conventional versus Asphalt
Rubber Pavements», 2002, Rubber Pavement Association.
Kaloush K.E., Witczak M W., Sotil A.C., Way G.B., “Laboratory Evaluation of Asphalt
Rubber Mixtures Using the Dynamic Modulus (E*) Test”, 82nd TRB Annual Meeting,
Washington, 2003.
Kumar P., Chandra S., Bose S., “Comparative Study on Stone Matrix Asphalt Mixes with
Different Stabilizers”, 84th TRB Annual Meeting, Washington, 2005.
Leung F., Tighe S., Macdonald G., Pento, S., “Noise Reducing Asphalt Pavements: a Canadian
Case Study”, 10th International Conference on Asphalt Pavements, Quebec City, 2006.
Partl M.N., Pasquini E., Canestrari F., Virgili A., “Analysis of water and thermal sensitivity of
open graded AR mixtures”, submitted to Construction and Building Materials, 2009.
Pasquini E., «Advanced characterization of innovative environmentally friendly bituminous

mixtures», Ph.D. Thesis, 2009, Università Politecnica delle Marche.
Potgieter C.J., Sadler D.E., De Villiers E.M., “Bitumen Rubber Asphalt: Report on the Long
Term Performance in South Africa”, 9th International Conference on Asphalt Pavements,
Copenhagen, 2002.
Sandberg U., Easement J.A., Tire/Road Noise Reference Book, Kisa, Informex, 2002.
Santagata F.A., Canestrari F., Pasquini E., “Mechanical characterization of asphalt rubber -
wet process” 4th International SIIV Congress, Palermo, 2007.
Santagata F.A., Antunes I., Canestrari F., Pasquini E., “Asphalt Rubber: Primeiros Resultados
em Itália” Estrada 2008, V Congresso Rodoviário Português, Estoril, 2008.
Santagata F.A., Cardone F., Canestrari F., Bahia H.U., “Modified PATTI test for the
characterization of adhesion and cohesion properties of asphalt binders” The 6 th
International Conference on Maintenance and Rehabilitation of Pavements and
Technological Control MAIREPAV6, Torino, 2009.
Scofield L.A., «The History, Development, and Performance of Asphalt Rubber at ADOT»,
Report Number AZ-SP-8902, December 1989, ADOT.
Souza R., Himeno K., Kobayashi A., “Evaluation of Japanese Asphalt Binders and Asphalt
Mixtures after Mixing Ground Rubber”, 3rd International SIIV Congress, Bari, 2005.
Way G.B., “OGFC Meets CRM. Where the Rubber meets the Rubber. 12 Years of Durable
Success”, Asphalt Rubber 2000 Conference, Vilamoura, 2000.
Way G.B., “OGFC Meets CRM. Where the rubber meets the rubber 15 years durable success”,
Asphalt Rubber 2003 Conference, Brasilia, 2003.
Zborowski A., Sotil A., Kaloush K.E., Way G.B., “Material Characteristics of Asphalt Rubber
Mixtures”, 83rd TRB Annual Meeting, Washington, 2004.
De-icing Characteristics of Rubber Concrete

Pavements
Fred Milani—H.B.Takallou
CRM Company, USA

26691 Plaza Drive
Mission Viejo, CA 92691
femilani@aol.com
hbtak@aol.com
ABSTRACT. The performance of rubberized asphalt has been investigated extensively and it
is proven that rubberized asphalt is superior to conventional asphalt. The de-icing property
of rubberized asphalt has attracted the attention of many Transportation Departments in the
United States since early 1980. In few field test, significant reduction on stopping distance
were observed during icy conditions as compared to control sections in Alaska, New Jersey,
and Minnesota.
Simulated traffic (wheel passage) tests were used to study ice dis-bonding from the surface of
rubberized asphalt, the results were somewhat encouraging. Ice dis-bonding were observed
where the rubber particles where present. This results lead to the idea of “Chunk Rubber
Asphalt Concrete” (CRAC).
The purpose of this paper is to demonstrate that de-icing is an inherent property of rubberized
asphalt by correlating the de-icing property of rubberized asphalt to the interfacial surface
tension between water and rubber modified bitumen and interpret the above contradictory
results in terms of interfacial surface tension and rubber particle size.
KEYWORDS. De-Icing, Rubberized Asphalt, Interfacial Surface Tension

1. Introduction
Current techniques to remove ice and compacted snow from pavement surfaces are costly
because they needlessly crush or melt most of the ice layer. Also, in many cases they do
not accomplish dis-bonding or clear the ice from the surface of the pavement. In addition,
deicing chemicals currently in use (salts) are environmentally objectionable. More efficient
alternatives to current physical and chemical methods are needed which will selectively apply
environmentally acceptable energy to the ice - pavement interface to accomplish dis-bonding.
In 1980’s and 1990’s most of the Transportation Departments in the cold states installed
rubberized asphalt in test sections in order to compare the performance of rubberized asphalt
with conventional asphalt. It was believed that iced formed on the surface of rubberized
asphalt will dis-bond from the surface of asphalt presumably because rubber acts as aggregates
that flexes on the pavement surface under traffic. The flex helps to break the ice and the wind
generated by passing vehicles moves the ice to the side of the road. In these experimental
sections, from 1 to 3 percent granulated crumb rubber by weight of the total mix was added to
the paving mix. The granulated rubber consisted of rubber particles ranging in size from 4.2
mm (1/4 in) to 2.0 mm (Number 10 sieve).
Alaska Department of Transportation installed experimental pavement sections using

rubber with granulated rubber particle size ranging from 1.6 to 6.4 mm. On measuring vehicle
stopping distances, significant reduction in distance were observed during icy conditions as
compared to control sections (Takallou, et al., 1987). Some raveling was detected. Test section
in route NJ41 in Cherry Hill, New Jersey showed improved skid resistance, however, some
raveling were reported.
From test sections in other cold states that had used rubber with the above specifications
it had been concluded that there is no difference between the performances of rubberized
asphalt and conventional asphalt in cold climate (Stuart et al.,1991; Stuart et al.,1991; Stuart
et al.,1991; Mogawer et al.,1989). Raveling was one of the major problems in these test
sections.
Recently, the New Jersey Department of Transportation paved a section of Route 95 with
asphalt rubber. In this project unlike previous projects fine crumb rubber (100 % passing 20
mesh screen), were used. Skid test showed that the rubber improved the skid resistance of the
pavement and no raveling is reported so far (Asphalt 2008).
On laboratory level, Albert F. Wuori (Wuori,1993) studied the dis-bonding of ice from
the surface of rubberized asphalt by simulated traffic method. Wheel passage tests of ice
grown on the rubber-aggregate asphalt samples did not show significant ice dis-bonding
from the surface, but the result where somewhat encouraging in a sense that ice-dis-bonding
was observed on those areas of the surface that was covered by large rubber particles. The
conclusion was that by increasing the percentage of larger rubber particles the potential for
ice dis-bonding under wheel loading will increase. The technology referred to “Chunk Rubber
Asphalt Concrete (CRAC)” used crumb rubber particles larger than 4.75 mm with rubber
particle as large as 10 mm on the top of asphalt with the crumb rubber concentrations of 3,
6 and 12 percent. Simulated traffic (wheel passing) on the CRAC indicated that as rubber
content increases, ice breaks faster from the surface. Ice dis-bonding was observed where ever
the rubber particles were present, but the raveling was the major problem.
Raveling of rubber particles could be attributed to the available surface area of rubber
particles for surface-surface interaction with bitumen and flexibility of rubber particles. For a
constant weight of rubber, as the particle size of rubber decreases, its surface area increases.
The surface area of 1 gram of 0.375 mm rubber particles is about 45 times larger than 1
gram of 10 mm rubber particles. Therefore the surface-surface bonding between rubber and
asphalt of the latter is about 45 times stronger and particle flexibility in the latter practically is
diminished.
The dis-bonding of ice from the surface of rubberized asphalt could be explained in terms
of interfacial surface energy between water and the surface and flexibility of chunk rubber in
CRAC technology. In order to demonstrate that interfacial surface tension and de-icing are
correlated we will examine the interfacial surface tension of water-glass, water-Teflon and
water-rubber.
The interfacial surface tension between water and glass is 144 ergs/cm2. Water wets the
surface of glass and ice formed on the surface of glass sticks to the glass and will not break
down easily.
The interfacial surface tension between water and Teflon is 28 ergs/cm2. Water does not
wet the surface of Teflon and the ice formed on the surface of Teflon easily breaks down from
the surface.
The interfacial surface tension between water and bitumen is 84 ergs/cm2 and the
interfacial surface tension of water-rubber is 48 ergs/cm2. The ice formed on the surface of
chunk rubber portion of CRAC is not bonded to the surface as strongly as to the bitumen
portion of the asphalt; therefore, the ice formed over the rubber portion of CRAC will break
under the weight of moving vehicles. The flexibility of rubber chunk helps the separation
of ice from the surface the same way that flexibility of plastic in ice plastic trays helps the
removal of ice cubes with a minor pressure.
Recently the surface free energy (SFE) characteristics of two Oklahoma aggregates
with and without Styrene – Butadiene- Rubber (SBR) were evaluated for moisture induced
damaged potential using a Universal Sorption Device (USD). SBR coating altered the
aggregate surface from hydrophilic to hydrophobic .
The purpose of this paper is to demonstrate that de-icing is an inherent property of

rubberized asphalt by correlating the de-icing property of rubberized asphalt to the interfacial
surface tension between water and rubber modified bitumen and interpret the above
contradictory results in terms of interfacial surface tension and rubber particle size.
2. Laboratory Experimental Design
The drop shape method is been used to measure the contact angle of a water drop formed
on the surface of the sample. The interfacial surface energy is calculated by Young equation:
W S/L = γ L/A (1+cos Θ )
Measuring method is based on taking images of a steady drop on the surface of the sample.
The images then were analyzed for contact angle measurements.
Two different rubber gradations were used in these investigations. Table 1 represents a
typical gradation analysis for these two samples. The surface area of second sample is about
4 times more than the surface area of the first sample. Rubber was added to hot GP80-60
bitumen at 200 °C and mixed for 30 minutes. A drop of colored water was placed at room
temperature on the surface of bitumen and bitumen modified with crumb rubber samples on
the range of 2 to 16%. The images of the water drop on right angle to drop were recorded by a
digital camera,
Table 1: Gradation analysis for two types of crumb rubber
Sieve size Sample 1 Sample 2

% Passing % Passing
2.36-mm (#8) 100 100
2.00-mm (#10) 100 100
1.18-mm (#16) 50 100
600-µm (#30) 9 70
300-µm (#50) 3 25
150-µm (#100) 1 6
75-µm (#200) 0 1
The images were analyzed for contact angles measurements. The interfacial surface free
energies were calculated for these samples by Young equation.
3. Laboratory Test Results
The interfacial surface tensions for between water drop and rubber modified bitumen
are tabulated in table 2. The errors in interfacial surface measurements are attributed to two
factors. The error related to the angle Θ measurements for the lack of a sharp boundary
between the drop and the surface and the error related to the surface of the sample covered by
rubber particles. The second sample covers more area and the coverage is more uniform. As
the result the error in interfacial surface measurements in the second sample is less than the
first one.
Table 2. Interfacial surface tensions for bitumen modified rubber
I.S.T ergs/cm2 I.S.T ergs/cm2

%Rubber
Sample 1 Sample 2
0 84±1 84±1
2 83±1 83±1
4 81±2 81±2
6 77±3 76±3
8 74±4 73±3
10 70±4 69±3
12 69±3 68±2
14 68±3 67±2
16 68±3 67±2
In both samples the interfacial surface tension decreases by adding rubber to bitumen
and levels up at about 15% rubber. The interfacial surface tension of about 67-68 ergs/cm2 is
considerably less than that of bitumen (84 ergs/cm2), but larger than that of rubber (48 ergs/
cm2), Therefore, it is expected that ice dis-bounding from the surface of asphalt modified with
fine crumb rubber not to be as pronounced as dis-bounding from the surface of chunk rubber
portion of CRAC, but to be more pronounced than none rubber portion of CRAC.
4. Conclusion
The interfacial surface tension of about 67 ergs/cm2 of asphalt modified with 15-20%
fine crumb rubber indicates that de-icing is an inherent property of rubberized asphalt. The
disbanding of ice from such a surface will be uniform, the surface- surface interaction of
rubber – bitumen will be much stronger and raveling practically will be diminished as been
illustrated by many successful projects in states with warm climates and recent projects in
New Jersey.
5. Bibliography
Asphalt: The Magazine of the Asphalt Institute: Posted March 6th, 2008
Mogawer W. S., Stuart K. D., Lee K. W., Evaluation of the Effects of Deicing Additives on
Properties of Asphalt Mixtures,” Transportation Research Record 1228, Asphalt Mixtures
and Asphalt Chemistry, Transportation Research Board, Washington, D.C., 1989, p. 41-53.
Stuart K. D., Mogawer W. S., “Laboratory Evaluation of Verglimit and PlusRide,” Public
Roads, Volume 55, Number 3, Federal Highway Administration, McLean, VA, December
1991” p. 79-86.
Stuart K. D., Mogawer W. S., “Laboratory Evaluation of Verglimit and PlusRide,” FHWA-
RD- 90-013, Federal Highway Administration, McLean, VA, March 1991, p. 119.
Takallou H.B., Hicks R. G., Esh D. C., “Use of rubber-modified Asphalt Pavement in Cold
Regions”, Paving in Cold Areas. Proceedings of 3rd Canada/Japan Workshop (Vol. 2). July
1987 p. 541-572
Wasiuddin, N. M; Zaman, M; O’Rear, E. “Effect of Polymeric Aggregate Treatment Using

Styrene-Butadiene Rubber (SBR) for Moisture-Induced Damage Potential. Transportation
Research Board Annual Meeting 2007 Paper #07-3248.
Wuori, A.F., Keweenaw Research Center Michigan Technological University. Strategic

Highway Research Program National Research Council Washington, DC 1993,
Author’s Index
A N.Choudury 941
A.Collop 317,505
G.Airey 317,505,903
C.Akisetty 873 D
S.N.Amirkhanian 437,873
I.Antunes 777,967 O.Dašek 415
C.Dimitroplos 401
B Q.Ding 209
Y.S.Doh 285
Bai Q. 109
D.Baptista 647 E
R.Barros 847
K.P.Biligiri 255,387, 693 R.Elliott 317,505
D.Biondi 493
F
S.Biro 461,475
P.Burch 401 G.Fabbri 519
A.Faxina 519
C
M.Farias 341,355
B.Caicedo 171 A.S.Figueroa 329
Cao B.223 S.Filippi 493
Cao P. 603 L.Fontes 269,369
R.Cao 97,109 E.Fung 647
F.Canestrari 967 A.Furlan 519
D.Carlson 693,711,755
G
S.Carpenter 929
L.Celis 171 F.Garrott 929
Chen L. 955 A.Geiger 461
D.Cheng 767 A.Gigante 519
Chu H.619 S.Gillen 929
F.Giuliani 493,541 L
D. Gonzáles 171
J.Grenfell 317,505,903 Li C. 619
Li J.449,603
H Li Z. 317
T. Ling 619
J.Haberl 725 F.R.Lizcano 329
E.Hajj 831 Q.Lu 737
Han J. 865
J.Harvey 189,737 M
R.Hewson 659
G.Hicks 119,767 A.Machado 555,579
S.Hu 209 S.Maheswari 941
C.Huang 209 G.Martinez 171
S.Huang 209 N.Memon 903
Huang W. 955 F.Merusi 493,541

F.Milani 989
J M.Minhoto 369
L.N.Mohammad 233
A.Johnson 767 A.Montepara 541
D.Jones 189 A.Murachelli 777
S.B.Cooper, Jr 233
N
K
A.Nadkarni 401
K.E.Kaloush 119,137,153,255,387,4 S.Neto 341,355
01,693 T.Nordgren 387,633
J.Kay 461,475
H.H.Kim 285 O
J.C.Kim 285
K.W.Kim 285 A. Öngel 737
J.V.Kirk 291
P
E.Kohler 737
J.Krishnan 941 A. Páez 885
J.Kudrna 415 J.C.Pais 47,269,341,355,369,555,579
E.Pasquini 967 J.B.Sousa 47,73,755
J.Peralta 555, 579 M.Stroup-Gardner 119
P.Pereira 269,369
T
A. Pérez-Lepe 885
F.Pilati 519 H.B.Takallou 989
G.Polacco 493 Tan Y. 303
D.Lo Presti 903 C.Thodesen 475
L.Preinfalk 633 G.Triches 269,369
Q U
Qian Z.109,955 P.Ullidtz 189
R V
J X.Reed 255,291,387, W.Vavik 929

J.Rizautto 795
M.Ryu 285 W
M.C.Rodezno 137,387
S.Waite 317,505
H.Rondon 329
Wang H.449
Wang S. 449,603
S
G.B.Way
B.Saha 941 25,47,73,119,693,711,755
F.A.Santagata 967 I.Widyatmoko 317,505
P.Sebaaly 831 R.Wu 189
P.Senthivel 941 Wu D. 619
S.Shatnawi 119,767 Wu S.865
J.Shen 437 Wei X. 619

H.Silva 555,579
X
J.Smith 813
Song X. 303 F.Xiao 873
M.I.Souliman 387 Xu C.609
Y
Yang J. 223
Y.Yang 209
Yu B. 223
A.Zareh 25,693,711
A.Zborowski 153
W.A.Zeiada 387,401
Zhou W. 303
Zhang X. 609
Zhang Y. 449
Zhang Y. 449
Zhang Y. 609
Z. Zhao 619
W.Zhao 873
Zhu Y.449

Characteristics of Rubberised Bitumen

Uploaded by

Copyright:

Available Formats

Characteristics of Rubberised Bitumen

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Characteristics of Rubberised Bitumen

Uploaded by

Copyright:

Available Formats

Binder Properties 505

Characteristics of Rubberised Bitumen

I. Widyatmoko* — R. Elliott* — J. Grenfell** — G. Airey** — A.

*Scott Wilson Limited

**Nottingham Transportation Engineering Centre

***Waste & Resources Action Programme

KEYWORDS: ageing, storage stability, rheology, cohesion, resilience, rubberised bitumen.

2. Materials for Testing

Table 1. Properties of Base Bitumen

Penetration Softening Viscosity Bitumen Composition

Table 2. Grading of Reclaimed Tyre Rubber (UK ‘Stock Products’)

3. Preliminary Screening Test

According to the ASTM D8 definition, rubberised bitumen is a blend of bitumen, reclaimed

90 100% Ambient Car Tyre [50/50 blend fine/coarse]

Lower Specification Limit

Figure 1. Target “Standard Size” Gradation for Rubber Particles

1000 ME15 1000 VE15

40/60pen, Ambient Car 40/60pen, Ambient Car

• 81.5% ME 40/60 + 18.5% AC

• 81.5% ME 40/60 + 18.5% CC

• 81.5% VE 40/60 + 18.5% AC

• 81.5% VE 40/60 + 18.5% CC

• 85% VE 40/60 + 15% AT

• 81.5% VE 100/150 + 18.5% CC

Table 3. Penetration and softening point tests rubberised bitumen

15% Rubber/85%Bitumen 18.5% Rubber/81.5%Bitumen

4. Effect of rubber size

In this exercise, both maximum (peak-ASTM D6114) and “stabilised” (AASHTO

40/60pen, Ambient Car 40/60pen, Cryogenic Car

• 84% ME 40/60 + 16% AC

• 81.5% VE 40/60 + 18.5% AC

• 81.5% VE 40/60 + 18.5% CC

Table 4. Comprehensive Test Results

Binder Sample Target

5. BS EN 14023 Tests on the Three Best Performing Blends

Table 5. Binder Test Results

Cohesion is commonly accepted as an indicator of a bitumen’s ability to resist tensile

The preliminary assessment highlighted the sensitivity of rubberised bitumen to blending

A total of 48 rubberised bitumen blends of different compositions were manufactured and

• 84% ME 40/60 + 16% ambient car tyre rubber

CalTrans, (2003). Asphalt Rubber Usage Guide, State of California Department of

Hanson, D.M., and Duncan, G.M. (1995). “Characterisation of Crumb Rubber-Modified

Oliver, J.W.H. (1982). “Optimising the improvements obtained by the digestion of

Mechanical properties of HMA mixes using

Fernanda Pilati - Adalberto Leandro Faxina - Ana Paula Furlan

*Department of Transportation, Sao Carlos Engineering School

Adding aromatic oils to crumb-rubber modified binders is a practical alternative to reduce

3. Design of the experiment

Table 1. Crumb rubber particle-size distribution

sieve opening, mm 1.19 0.59 0.42 0.297 0.175 0.150 0.074

5.1. Preparation of crumb-rubber modified binders

5.2. Compaction of Marshall specimens

Table 2. Mixing and compaction temperatures (°C)

5.3.Acquisition of resilient modulus and tensile strength data

Resilient modulus values were obtained according to the AASHTO TP 31 protocol

property regression model R² (%) n

property regression model R² (%) n

8000 4000 2000

0.32 0.68 0.32

I. Widyatmoko* — R. Elliott* — J. Grenfell — G. Airey — A.