Compression and Recovery Behaviour of Polyamide-6 Based Foams

Download as pdf or txt
Download as pdf or txt
You are on page 1of 14

Compression and Recovery Behaviour of Polyamide-6 Based Foams

Compression and Recovery Behaviour of Polyamide-6 Based Foams


V.G. Izzard*, H. Hadavinia*1, V.J. Morris*, P.J.S. Foot*, L. Wilson** and K. Hewson**
*
Materials Research Centre, Kingston University, London, UK
**
Zotefoams plc, Croydon, UK
Received: 30 August 2011, Accepted: 21 December 2011

SUMMARY
This paper presents experimentally determined compression sets and compression behaviour of two closed-cell,
crosslinked foam materials; ZOTEK N B50 (Nylon 6) and ZOTEK N A30 (nylon/polyolefin alloy). This work
forms the basis of future investigations into the post-impact behaviour of these foams.
Compression set performance was measured at 25.0, 37.5 and 50.0% strain at nine temperatures ranging from
-5 to 90 °C. Compression tests at constant strain rates were conducted at four temperatures between 23 to 90 °C.
Finally compression tests at 23 °C were repeated at four strain rates between 0.3 and 550 hr-1 to determine strain
rate dependency. The Nagy and Williams-Landel-Ferry scaling factor for strain rate and temperature were applied
to the experimental results and equations were derived which allowed the performance of the two polyamide
based foams to be interpolated over the strain rate and temperature range of study. The required material
properties of interest of the base polymers have been assessed and are presented and discussed in relation to the
performance of the foam materials.

Keywords: Cellular solid, Polyamide, Nylon-6, Polyolefin, Compression set, Stress-strain, Scanning
electron microscope, Glass transition temperature, Strain rate, Foam

1. INTRODUCTION foam requires detailed characterisation As given by Mills et al. 6 and


of the foam material’s response. Many Rodriguez-Perez et al.7, it is well
Polymer foams are multi-phase authors, including Mills1, Gibson and established that the driving forces
materials made of a matrix material Ashby2, have reported on work that for recovery of closed cell foams
(polymer) and cells (gas). Depending characterises cellular solids at the are related to (i) the compressed gas
on the matrix material as well as the micro level using structural mechanics inside the cells and (ii) the viscoelastic
cell microstructure, foams exhibit modelling based on assumed lattice recovery of the deformed cell walls.
dramatically different properties. structures. As stated by Kraynik and Additionally, it is also widely known
In the majority of polymer foam Warren3, the mechanical behaviour that the mechanical properties of
applications such as automotive, sport of cellular solids is dependent on the base polymers are sensitive to
and packaging, foams are subjected to the mechanical behaviour of its both temperature and strain rate,
stresses. Recovery after relaxation of an components. That is to say, in order especially at or near to their glass
applied stress is an important property to predict foam performance, the transition temperatures (Tg)8. It is
in the subsequent behaviour of the models developed must be based therefore crucial that any cellular
foams. In addition, one of the primary on an accurate description of the solid performance predictions are
uses for polymer based cellular solids base polymer behaviour. However, based, not only on the understanding
is to act as energy-absorbing materials many of these models assume that of the micro structure, but also on
during impact. In reality, impacts occur variations in base polymer properties a detailed knowledge of the base
over a wide range of temperatures and are either simply characterised or polymer behaviour. For any model
at velocities that may vary over many linearly dependent with regard to to be truly useful it must consistently
orders of magnitude and at angles environmental conditions 4,5 , and and realistically predict the foam
of incidence that are unpredictable. therefore cannot be applied to a range performance over a wide range of
Understanding the behaviour of the of foams5 or situations. conditions. Of equal importance is
the foam behaviour during post-
compression recovery. A robust
1
Corresponding Author: e-mail: h.hadavinia@kingston.ac.uk (H. Hadavinia), model should be consistent over a
Tel.:+44 20 8417 4864, Fax:+44 20 8417 4992
range of temperatures with regard
©
Smithers Rapra Technology, 2012 to these parameters. This study

Polymers & Polymer Composites, Vol. 20, No. 5, 2012 411


V.G. Izzard, H. Hadavinia, V.J. Morris, P.J.S. Foot, L. Wilson and K. Hewson

provides a range of experimental 2. BASE POLYMER (DSC). A small amount of material


measurements fitted by scaling PROPERTIES (circa 5 mg) was sealed in an aluminium
and empirical equations to allow DSC crucible using a pierced lid.
predictive interpolation of static The fundamental behaviour of any The samples were heated in turn
compressive behaviour as a precursor cellular solid is influenced by the from -100 to 180 °C, and then cooled
for future studies into the dynamic properties of the base polymer. To back to -100 °C before reheating the
performance of these materials. predict the mechanical performance sample back to 180 °C. Heating and
and possible responses of the foam, it cooling rates of 10 °C/min were used
Zotefoams plc is a world leader in is important to gain an understanding throughout.
the manufacture and distribution of the polymer properties matrix. These
of speciality foam materials, properties are essential for further
numerical studies for modelling of
2.2 Results of Base Polymer
manufactured using a unique, Investigation
p r o p r i e t a r y t e c h n o l o g y 9. T h e the foams under general multiaxial
technology uses nitrogen gas as a loading. The mechanical properties of both base
Physical Blowing Agent (PBA) which polymers are shown in Table 1. These
is dissolved into the base polymer 2.1 Experimental Procedure are based on the mean of three separate
in a high pressure/temperature measurements for each material. The
Tensile tests were undertaken on tensile stress-strain curves of N A30
process. Once equilibrium is
base polymer samples of both and N B50 are shown in Figure 1b. The
achieved, thermodynamic instability
foam materials to determine the elongation to break and elastic modulus
is introduced to initiate nucleation
ultimate tensile stress, elongation at for the N A30 base polymer show it is
of the foam structure. This method
break, elastic modulus in tension and significantly more ductile than N B50
of production minimises impurities
Poisson’s ratio. but has lower stiffness. It would be
in the product whilst also optimising
properties and performance. It is also expected that the foam materials would
considered environmentally friendly For each base polymer, three tensile also demonstrate similar differences
since nitrogen is the only major tests were carried out with reference at equivalent density. The properties
by-product of the process. A further to BS 2782-3 standard 10. Strain presented in Table 1 are consistent with
advantage of this technology is the measurements were made using two previously published data11,12.
wide range of materials to which it VISHAY general purpose strain gauges
can be applied. CEA-06-240UZ-120. A third strain 3. FOAM PERFORMANCE
gauge was used to determine Poisson’s
Examples of such materials are the
ratio of the material, Figure 1a. These 3.1 Compression Set
gauges have a stated accuracy of 0.1% Compression set is a standard test
newly developed closed cell foams
which far exceeds the accuracy of 2% procedure applied to foams. It provides
based on Polyamide-6 (PA6) under
in measurements, as required by BS data on one of the basic mechanical
study here. For materials such as
2782-3. Load and strain data were responses of foams: recovery after
PA6, foaming has traditionally
been technically challenging with recorded using VISHAY data logging fixed compressive strain loading. In
limitations on any possible reduction system 5000 operating at 10Hz. literature, there are many instances
in density. ZOTEK N B50 and where compression set tests have been
ZOTEK N A30 grade foams have All dimensional measurements were used to monitor the effects on foams
nominal densities of 50 kg/m 3 carried out using a MID-F125/150 following chemical or mechanical
and 30 kg/m 3, respectively. The Mitutoyo Digimatic digital micrometer ageing or when process/formulation
characteristics of these foams with an accuracy of ±0.006 mm variations were being investigated.
reflect the general properties of fitted with an 8 mm flat anvil. For For example, the effects of blending
this class of polymer with good mass determination, a Mettler AJ150 two or more polymers13,14, the effect of
high temperature performance, balance with an accuracy of ±0.001 g the addition of fillers15,16, the effect of
resistance to hydrocarbon fuels was used. Prior to testing, all samples the level of crosslinking17 and aspects
and oils combined with impact were conditioned at 23 °C and 50% of foam ageing18,19. Little however
performance, buoyancy and low RH for a minimum of 24 hours, but has been reported on the data from
thermal conductivity which derive no control of humidity was applied the compression set test itself, or on
from the cellular structure. These during the testing. compression set data outside the scope
foams are considered ideal for shock defined by the most commonly used
and impact absorption, acoustic The glass transition temperature (Tg) ISO20 and ASTM21 standards. Often
dampers and cushioning applications was measured on the base polymers compression set data are recorded at
in extreme environmental conditions. using differential scanning calorimetry only a few points of recovery time,

412 Polymers & Polymer Composites, Vol. 20, No. 5, 2012


Compression and Recovery Behaviour of Polyamide-6 Based Foams

and no predictive interpretation of the Figure 1. (a) Tensile test specimen of base materials, (b) Stress-strain behaviour of
foam performance outside of the range N B50 and N A30 base polymer
of these measurements is available.
(a)
Literature does contain studies on
recovery rates after loading for
crosslinked, closed-cell foams.
Strain gauges 1 & 2
Mills and Gilchrist22 investigated 3rd strain gauge fitted at
fitted to each side of
the recovery rates of low density 90° to strain gauges 1
polymer sample
polyethylene (LDPE) and ethylene & 2 for Poisson’s ratio
vinyl acetate copolymer (EVA) foams measurement
on completion of 7 day creep tests.
Their work shows that at relatively high
strains, the majority of the gas in the
cells escapes. For low density foams,
viscoelastic recovery is hindered by
the pressure drop in the cells, making
recovery a slow process. This study
does not investigate the permeability
of cell walls to air but it is a likely
contributing factor to the recovery of
the foams studied in this work. (b)

3.1.1 Compression Set


Experimental Procedure
In order to reduce equipment-
dependent uncertainties, all
dimensional measurements were
carried out in accordance with
ISO 192323 using a MID-F125/150
Mitutoyo Digimatic digital micrometer
with an accuracy of ±0.006 mm fitted
with an 8 mm flat anvil. For mass
determination, a Mettler AJ150
balance with an accuracy of ±0.001 g
was used. The density of the foam
sheet was calculated according to
ISO 84524. To minimise the effects
of density variation upon the results,
all samples were taken from two
foam sheets manufactured in the
same batch. These samples were Table 1. Base polymer properties at 23oC
conditioned for a minimum of 6 days at Property* Nylon-6 Nylon /Polyolefin Alloy
23 °C (±2 °C) and 50% (±5%) relative N B50 N A30
humidity prior to testing. Standard Standard
Deviation Deviation
Compression set tests were conducted Density (kg/m3) 1143 3.8 1099 0.4
in accordance with ISO 185620. Three Elastic Modulus Es (GPa) 3.19 0.02 2.11 0.01
samples were tested for each test Max Stress ss (MPa) 73.9 0.9 47.2 0.4
condition and the mean value reported. Poisson’s Ratio 0.398 0.001 0.319 0.012
Nominal test sample size was 50 mm
Elongation % 3.4 N/A 66 N/A
× 50 mm × 25 mm. The dimensions of
Glass Transition Temperature (oC) 45-51
the ZOTEK N B50 and ZOTEK N A30
samples were recorded and compressed *Based on the mean of 3 samples

Polymers & Polymer Composites, Vol. 20, No. 5, 2012 413


V.G. Izzard, H. Hadavinia, V.J. Morris, P.J.S. Foot, L. Wilson and K. Hewson

to a fixed strain (25.0, 37.5 and 50.0% Figure 2. Compression set results of N B50 and N A30 at temperatures between
of original thickness) and acclimatised -5 °C and 90 °C at c.s. 50.0%
at experimental temperature (-5, 10,
23, 30, 40, 50, 60, 70 and 90 °C) for a
duration of 22 hours.

On release of the strain, the samples


were maintained at the relevant
experimental temperature during
recovery. The dimensions of the
samples were re-measured at various
time intervals from 30 minutes through
to 24 hours post release. Recovery at
temperature is beyond the scope of
ISO 185620.

All compression set (c.s.) data are


reported according to ISO 185620,
using Eq. (1)

do − dr
c.s. = x100
do
(1)

where do is the original thickness of


the sample and dr is the thickness after
recovery at times of interest.

3.2 Results of Compression Set


Testing
Table 2 provides a comparison of the
compression set results for both N
B50 and N A30 at the temperature and
recovery times specified in ISO 185620.

The N B50 results from Table 2 show


good agreement with published data9.
This confirms that the methods adopted
are transferable to additional test
conditions and to other polyamide-6
based foams.

Table 2 also shows that the N B50 foam


recovers at approximately the same rate Table 2. Comparison of compression set tests results
for the first 30 minutes of the recovery. Parameter N B50 N A30
This rate is not affected significantly by Density 47.96 kg/m3 31.62 kg/m3
the level of initial fixed strain. c.s. (25.0%, 30 min, 23 oC) 14.7% 7.0%
c.s. (25.0%, 1440 min, 23 C)
o
10.9% 3.4%
Figure 2 shows the experimental c.s. (37.5%, 30 min, 23 oC)* 22.7% 10.9%
compression sets as a function of c.s. (37.5%, 1440 min, 23 oC)* 19.3% 7.6%
recovery time at various temperatures
c.s. (50.0%, 30 min, 23 oC) 28.6% 10.9%
after initial fixed 50% compression
c.s. (50.0%, 1440 min, 23 oC) 23.4% 6.9%
strain for N B50 and N A30. The
curves represent the best fit for each * Initial fixed strain characteristic not explicitly included in ISO 1856

414 Polymers & Polymer Composites, Vol. 20, No. 5, 2012


Compression and Recovery Behaviour of Polyamide-6 Based Foams

temperature studied. The results show Figure 3. Comparison of experimental compression set data at 30 and 1440
that N B50 has a lower recovery rate minutes with empirical curve fit- N A30
than the N A30 across all measured
temperatures. A wider spread of data
is also noticeable with respect to
temperature for N A30. This would
suggest that the N A30 recovery is
more sensitive to temperature than
that of N B50. In addition, the results
indicate that neither material returns to
a zero compression set state (i.e. full
recovery) after 24 h. This is expected,
given the short timescale and the closed
cell nature of these materials.

As previously reported by Izzard et


al.25; Eq. (2) below is an empirical,
predictive equation for recovery after
compression set as a function of time
for N B50. The present study has
extended the temperature range and
it also includes 37.5% initial strain.

c.s. (%) = a ln(t) + b Table 3. Coefficients of logarithmic empirical fitting to compression set data for
(30 < t < 1440) (2) ZOTEK N B50
Test Compressive Strain Compressive Strain Compressive Strain
where t is the recovery time in minutes Temperature = 25.0% = 37.5% = 50.0%
and ‘a’ and ‘b’ are temperature (°C) a b a b a b
dependent coefficients as presented
-5 -0.934 16.511 -0.812 21.246 -1.208 32.092
in Table 3.
10 -1.340 18.744 -0.619 19.596 -1.497 35.509
The same conditions are applied to the 23 -1.004 18.339 -0.870 25.699 -1.323 32.884
lower density, N A30 foam material, 30 -0.936 20.644 -1.242 25.410 -1.798 36.023
and results showed a similar pattern 40 -1.224 19.359 -1.827 31.347 -1.855 32.000
to those as previously reported 25. 50 -1.275 22.352 -1.662 31.238 -3.314 44.367
Eq. (2) can therefore apply to N A30 60 -1.266 22.138 -1.431 29.447 -1.745 33.623
foam with the temperature-dependent
70 -1.050 19.644 -1.200 27.012 -2.334 37.426
coefficients given in Table 4. In both
90 -1.360 21.431 -1.591 29.680 -1.265 32.904
cases the data, when plotted, exhibits
a high logarithmic dependency as a
Table 4. Coefficients of logarithmic empirical fitting to compression set data for
function of recovery time.
ZOTEK N A30
Test Compressive Strain Compressive Strain Compressive Strain
Eq. (3) predicts the compression set as Temperature = 25.0% = 37.5% = 50.0%
a function of recovery time at a specific (°C) a b a b a b
temperature. Although Figure 2 would
-5 -0.373 4.997 -0.578 8.857 -0.900 11.986
indicate a very limited temperature
effect on the compression set behaviour 10 -0.632 7.583 -0.660 8.439 -0.460 8.268
for N B50, it shows a definite upward 23 -0.961 10.221 -0.827 13.612 -1.093 14.743
trend for N A30, indicating that a 30 -1.448 16.435 -1.063 13.043 -1.629 20.955
temperature dependent function is 40 -1.101 11.524 -1.279 15.559 -1.137 16.649
required. A surface modelling package, 50 -1.215 14.100 -1.417 18.000 -1.548 22.333
3D table curve26, was used to derive a 60 -1.082 11.427 -1.098 16.148 -1.447 20.754
surface fit. Equation (3) uses a 3rd order
70 -1.282 14.067 -1.392 19.370 -1.434 22.969
polynomial for temperature which
90 -1.326 16.911 -1.598 24.060 -1.731 30.159
includes a logarithmic time component.

Polymers & Polymer Composites, Vol. 20, No. 5, 2012 415


V.G. Izzard, H. Hadavinia, V.J. Morris, P.J.S. Foot, L. Wilson and K. Hewson

This function is an empirical surface tube of length 300 mm was positioned dampened the temperature fluctuations
fit which predicts the compression over the sample and located on a of the heating jacket. Temperature was
set performance of these foams as a preinstalled base cap. A second 50 mm monitored and controlled by the use
function of both temperature and time. aluminium disc and rod arrangement of thermocouples; one on the foam
was placed in the tube and held above sample and one in the brass tube. Once
c.s. (%) = a + b T + c T2 + d T3 + the foam sample by the upper jaw of the thermal equilibrium had been achieved,
e ln(t) (-5 < T < 90, 30 < t < 1440) universal testing machine. The brass the sample was left to condition for a
(3) tube was capped using a second brass further 2 hours.
plate. A heating jacket was wrapped
where t is the recovery time in minutes,
around the assembly and set to the Compressive loading was applied
T is the recovery temperature in oC
required temperature (see Figure 4). to the sample at a constant strain
and ‘a’ to ‘e’ are the foam dependent
constant coefficients presented in The brass tube served as a heat sink and rate of 0.01 s-1 until a strain of 80%
Tables 5 and 6 for N B50 and N A30,
respectively.
Table 5. Coefficients of surface fitting to compression set data at various
temperatures N B50
Compression set at any given time
Compressive Coefficient
for N B50 is almost independent
Strain a b c d e
of temperature but for N A30 the
compression set increases as the 25.0% 17.7925 8.2464 x 10 -2
-1.0413 x 10 -4
-5.7929 x 10-6 -1.1572
temperature rise as shown in Figure 3. 37.5% 23.5808 1.1710 x 10-1 -1.5032 x 10-5 -8.9340 x 10-6 -1.2686
50.0% 35.7303 4.2558 x 10-2 -2.0852 x 10-3 -1.7781 x 10 -5 -1.7773
3.3 Compression Testing
The compressive stress-strain Table 6. Coefficients of surface fitting to compression set data at various
temperatures N A30
performance of the foam is a
characteristic of its energy-absorbing Compressive Coefficient
Strain a b c d e
capabilities, and can be determined by
the methods described in BS3386-127. 25.0% 9.1989 1.8785 x 10-1 -4.0553 x10-3 3.0341 x 10-5 -1.0598
It should be noted that the level of 37.5% 11.4089 8.2848 x 10-2 -4.2795 x 10-4 8.1484 x 10-6 -1.0745
accuracy for force and displacement 50.0% 13.5568 1.0628 x 10-1
-2.0541 x10 -4
8.0655 x 10-6 -1.2642
used in this study far exceeds those
specified in the standard. Further
to this, BS3386-1 does not include Figure 4. High temperature compression test set up
investigation at varying strain rates or
environmental temperature. This study
used BS3386-1 as a guide and develops
the method to study the effect of these
parameters.

3.3.1 Experimental Procedure -


Effect of Temperature at Constant
Strain Rate
All samples were 50 mm diameter
25 mm thick foam cylinders which
were conditioned and measured as per
Section 3.1.1. Cylinder diameter was
measured using Engineer’s stainless
steel Vernier Calliper to an accuracy
of ±0.02 mm.

A sample was placed on an aluminium


disc of a slightly larger diameter held by
an aluminium rod in the lower jaw of
the universal testing machine. A brass

416 Polymers & Polymer Composites, Vol. 20, No. 5, 2012


Compression and Recovery Behaviour of Polyamide-6 Based Foams

was reached. Compression in the Figure 5. Experimental and fitted stress-strain curves over the temperature range
sample was measured using a LVDT 50 °C to 90 °C. (a) N B50 (b) N A 30
branched from the loading arm (see
Figure 4). Both load and compressive
displacement were logged using a
Vishay Measurements Group System
5100 (data logger) set at 10Hz sample
rate. The foams were tested at four
different temperatures of 23, 50, 70
and 90 °C.

3.3.2 Results of Compression Tests


- Effect of Temperature at Constant
Strain Rate
The compression tests results at
different temperatures are shown in
Figure 5. The elastic modulus of the
foam shows a temperature dependency;
the higher the temperature, the lower
the elastic modulus. The elastic
modulus will vary both above and
below Tg. Proportional change is lower
below Tg than it is above.

The observed temperature dependent


compressive stress-strain behaviour
of the foams can be described using
the Williams-Landel-Ferry (WLF)28
shift factor:

 −C ( T − Tr ) 
L(T) = EXP  1 
 C2 + ( T − Tr ) 
(4)

where C 1 and C 2 are material


dependent constants, Tr is the reference
temperature (often taken as the polymer
glass transition temperature Tg) and T
is the temperature of interest.

The compressive behaviour of N A30


does not exhibit the behaviour as
predicted by using the shift factor in
Eq. (4). It is thought that the change in
cell gas pressure due to temperature is expected. For higher density foams, and more ductile foams, the gas
the dominating effect, rather than any where there is more polymer in the cell pressure effect must be extracted from
characteristic of the polymer behaviour. walls or struts, and foams made with the data prior to the application of the
At increasing strain increments, stiffer polymers, the gas pressure has a WLF shifting factor.
the volume of the cells decreases. less dominating effect on their buckling
Raising the temperature at those strain capability. Therefore the convergence From the ideal gas law, assuming
increments also results in the cell gas at higher strains seen in Figure 5b is the number of moles is constant and
pressure being higher. This effect not observed in higher density/stiffer Poisson’s ratio is zero, the gas pressure
counteracts the temperature ‘softening’ foams. In order to understand the curves for given temperatures can be
of the polymer that would normally be polymer performance in low density determined from Eq. (5):

Polymers & Polymer Composites, Vol. 20, No. 5, 2012 417


V.G. Izzard, H. Hadavinia, V.J. Morris, P.J.S. Foot, L. Wilson and K. Hewson

Cell Gas Pressure (MPa gauge)= A single sample, whose nominal n(ε )


 
dimensions were 25 mm x 140 mm ε
 T   ε   σ(ε) = σ 0   (7)
  x   x0.1013 x 140 mm, was tested at each test  ε 
 To  1 − ε − (ρFoam / ρPolymer )   o
condition. The load and linear variable where:
(5) differential transducers (LVDTs)
s0 = Nominal stress at quasi-static
outputs were digitally logged using
strain rate ε 0
where T is the temperature of interest, VISHAY data logging system 5000.
To is the temperature at which the ε = Strain rate
foam cell pressure is in equilibrium The foam sample remained unloaded ε 0 = Strain rate at quasi-static test
with atmospheric pressure, ε is strain, during this process through the use of n(ε) = Rate dependency coefficient
rFoam and ρPolymer are the foam and base 26 mm spacers (see Figure 6). The rig where n(e) = a + be.
polymer densities, respectively. was installed into a Mayes, 2 column
universal testing machine with a 50 kN The material rate dependency
In this case, To is taken to be 23 °C load cell. Four LVDTs were fitted to coefficient n(ε) was derived by plotting
as the foam is conditioned at this each corner of the test rig and zeroed. the stress ratio against strain rate ratio
temperature. This is to achieve gas The spacers were removed and the on a log-log scale at specific strain
equilibrium within the foam. sample compressed to approximately levels. As shown in Figure 8a for N
80% of its original thickness. An A30, a family of straight lines with
If the effect of gas pressure within the average displacement from the LVDTs tangent n was formed for both foams.
cell is added to the WLF shifting factor, was calculated and used to determine A regression was then used to fit the
strain. Samples were tested at strain tangents of these lines, n, against strain.
the stress strain performance of N A30
rates between 0.300 to 550 h-1 at 23 °C From this the functional form of n is
can be accurately predicted as shown in
for both N B50 and N A30. derived as shown in Figure 8b for N
Figure 5 using the following equation:
A30. The results of n(ε) for N A30 and
σ (ε) = 3.3.4 Results of Compression Tests N B50 are:
T
- Effect of Strain Rate at Constant
σ 0 (ε)L(T) + Cell Gas Pressure (6) Temperature n(ε) =
where sΤ (ε) is the stress at temperatures The foam elastic moduli E* at 23 °C 
of interest between 50 and 90 °C, sο (ε) for N B50 and N A30 at different strain  0.0052 − 0.07424 log(ε) N A30
is the stress at the lowest temperature rates are given in Table 8 and shown  0.0785 − 0.0674ε N B50
studied above Tg, in this case 50 °C. graphically in Figure 7. The lines
(8)
indicate that there was a near linear
variation of elastic moduli with log Although the test results scaled
Eq. (6) was fitted to N B50 and N A30
of strain rate over the ranges studied. accurately for N B50, they did not scale
experimental data. The relative density
is based on foam densities as given in so well for N A30. It was found that
Table 2. Constants C1 and C2 given Nagy et al.29 stated that based on quasi- N A30 foam softened with increasing
in Table 7 obtained using iterative static test data, it is possible to model strain. This behaviour was not observed
stress as a function of strain and strain rate for N B50 and therefore no softening
methods.
using the following constitutive equation: at higher strains occurred.
3.3.3 Experimental Procedure -
Effect of Strain Rate at Constant
Figure 6. Compression test set up at different strain rates
Temperature
All samples were measured and
conditioned as per Section 3.1.1. Foam
breadth and width were measured using
an engineer’s stainless steel Vernier
Calliper to an accuracy of ±0.02 mm.

Table 7. WLF coefficients for


temperature shifting factor in Eq. (4)
Foam Material C1 C2
N B50 11.0 880
N A30 3.4 180

418 Polymers & Polymer Composites, Vol. 20, No. 5, 2012


Compression and Recovery Behaviour of Polyamide-6 Based Foams

In order to improve the scaling for low Table 8. Variance in elastic Young’s modulus at different strain rate at 23 °C
stiffness foams, the Nagy equation N B50 N A30
was modified to include an additional
Strain Rate Elastic Modulus Strain Rate Elastic Modulus E*
strain-dependent term as follows: (hr-1) E* (MPa) (hr-1) (MPa)
0.3044 6.52 0.308 0.97
n(ε )
  3.020 7.79 3.050 1.29
ε
σ(ε) = A(ε)σ 0   54.64 10.08 30.5 2.04
 ε 
o
(9) 540.7 10.81 305.0 2.78

where A(ε) is a linear strain dependent


Figure 7. Foam Young’s modulus at various strain rates at 23 °C
function for low stiffness materials.
By using iterative methods it was
found that:


A(e) = 1.3 − 0.42ε N A30
1 N B50
(10)

4. SCANNING ELECTRON
MICROSCOPY STUDY OF
FOAM
The reliability of Eq. (3) can be
supported to some extent through
the analysis of Scanning Electron
Microscope (SEM) images of the
foam structure before and after
testing to investigate any potential
damage. Since SEM operates at low
pressure and high energy levels,
initial tests were conducted at
prolonged exposure to determine
any detrimental effects; none were
observed. Nonetheless, the SEM
images shown represent a small part
of the total bulk foam structure and 4.1 SEM Images Experimental 4.2 Interpretation and
the extent of any damage observed
Procedure Discussion of SEM Images
may not be representative of the
whole sample. Samples were cut Compression set samples for 50% SEM images of N A30 and N
using a surgical knife. initial compressive strains were B50 at 50% strain are presented
selected from the extremes of the in Figure 9a,c,d and f for each
As reported by Mills and Gilchrist22 temperature range. For comparative of the compression set samples at
and discussed in Section 3.1, recovery purposes, an uncompressed sample both high (90 °C) and low (-5 °C)
would be slowed by loss of gas (zero compression) was also analysed temperature. Figures 9b and 9e
pressure within the foam cells and as a control. show control samples which have
plastic damage in the cell walls. experienced no compression testing.
While diffusion of the gas through Three 10 mm X 10 mm X 5 mm (5 mm Figure 9a-f reveals no obvious
the cell window membranes is one in direction of observation) test pieces damage in the form of burst cells.
mechanism, it may also be expected were cut, from each of the compression These images represent the extremes
that, depending on strain levels and set samples. The magnification settings of test temperature and compressive
polymer matrix, evidence of ruptured are approximately 200x and 100x strain for the compression set tests.
cell walls could also be observed in for the N B50 and N A30 images, Assuming these to be representative
the SEM images. respectively. of the foam structure, it is reasonable

Polymers & Polymer Composites, Vol. 20, No. 5, 2012 419


V.G. Izzard, H. Hadavinia, V.J. Morris, P.J.S. Foot, L. Wilson and K. Hewson

Figure 8. Experimental and fitted stress over the strain range 0.05 to 0.8 at 23 °C for the strain rates studied (a, b and c) N
A30, (d) N B50

to assume no observable rupture of the Figure 10a-d shows SEM images the base polymer mechanical properties
cell windows would have been caused of the samples which have been presented in Figure 1b.
by the application of the lower strain compressed to 80% of the initial
levels and intermediate temperatures thickness during the compression
5. DISCUSSION
of the other compression set tests. tests for both foams. N B50 sample
images show significant plastic damage The compression set data show that
Figure 9a shows slightly more resulting, in some cases, in rupture over a 24 hour period neither of the
buckling and wrinkling within the of the cell windows, Figure 10a foam materials returned to its original
polymer film than is observed in the and 10b. This characteristic is also thickness. This is expected behaviour
control, Figure 9b. It is therefore evident in the N A30 sample images of closed cell polymer foams, and
possible that the N B50 foam will not with increased plastic damage but no suggests that mechanisms of plastic
recover its original thickness as some observed bursting cell wall. These damage due to cell wall buckling
plastic damage may have occurred due images confirm that the base polymer and gas loss through diffusion have
to buckling in which the material has of N A30 foam has greater ductility contributed to irrecoverable set in
significantly exceeded the elastic limit than that of the more brittle N B50 the bulk cellular structure. SEM
of the polymer. polymer base, a result also indicated by images have identified that there is no

420 Polymers & Polymer Composites, Vol. 20, No. 5, 2012


Compression and Recovery Behaviour of Polyamide-6 Based Foams

Figure 9. SEM results- 50% compression set tests (a) N B50 at 90 °C, (b) N B50 Figure 10. SEM results- 80%
no compression at 23 °C, (c) N B50 at -5 °C, (d) N A30 at 90 °C, (e) N A30 no compression set tests (a) N B50 at
compression at 23 °C, (f) N A30 at -5 °C 90 °C, (b) N B50 at -5 °C, (c) N A30 at
90 °C, (d) N A30 at -5 °C

consistent cell window rupture evident cell wall rupture which suggests that at
within the structure of the materials at strains greater than 50% compression
strain up to 50%, but there are cell wall cell walls did fail by this mechanism.
buckling effects visible in a number This contributed to a reduction in the
of the high strain samples. This is mechanical performance of the bulk
indicative of plastic damage caused foam. There is no evidence in this
by the initial compressive stage of study to suggest that the foam material
the compression set test contributing will not continue to recover beyond
to irrecoverable strain. Furthermore, 24 hours since the base polymers are
the same behaviour was observed viscoelastic. It is reasonable to assume
when additional N A30 samples that, the foam would recover to, and
were compressed to 80%. However, then plateau, at some lower level of
in N B50 samples under the same compression set, but that time frame
compression there was evidence of far exceeds the limits of this study.

Polymers & Polymer Composites, Vol. 20, No. 5, 2012 421


V.G. Izzard, H. Hadavinia, V.J. Morris, P.J.S. Foot, L. Wilson and K. Hewson

Theoretically on application of the for N B50 over the strain rate range Although, this theory has not been
empirical Eq. (3), and assuming no studied, as shown in Table 8. investigated in this work, it may be
plastic damage occurs, it could be applicable using the developed Eq. (6)
predicted that the time required for The manufacturing process of these and (8) for each material. It should be
the material to achieve a near zero foams incorporates a process by noted that this must be supported with
compression set is measured in years. which the polymer base material is additional experimentation.
crosslinked9. It is reported8 that the
Literature reports that the Tg of level of crosslinking vastly reduces
6. CONCLUSIONS
Nylon is approximately 50 °C11,12. It the change in elastic modulus that can
is noteworthy that on or around the occur over the Tg region. It is postulated In this paper, time and temperature
Tg region, the mechanical properties that the crosslinking characteristic dependent predictive functions have
of a polymer exhibited discontinuity may have limited the change observed been developed to approximate
in performance particularly in elastic in elastic modulus to approximately the compression set and recovery
modulus. Compression set results show 200% over the ranges studied. This performance of two types of closed cell
anomalies at or around a temperature is beneficial in creating a consistent crosslinked polyamide-6 based foams.
of 50 °C; compressive stress-strain performance of such materials when These equations are limited to the
data on the same foam indicate a local operating at or above the Tg region, ranges studied, and any extrapolation
maximum around this point. where otherwise performance may beyond these limits would need to be
be highly variable. The sensitivity to supported by further experimentation.
At temperatures near Tg the modulus compressive performance based on
of elasticity is sensitive to strain rate. the level of crosslinking may vary The Nagy constitutive equation and
In order to investigate this further, it depending on the polymer type. WLF scaling factor have been applied
would be necessary to measure the to derive accurate strain rate and
change in the elastic modulus as a In addition to crosslinking, this temperature dependent functions for
function of strain rate at a constant investigation has found that cell both foams. The coefficients for the
temperature. This experimentation structure, polymer volume and the scaling equations are dependent on
was carried out at four different strain effect of gas pressure on these the temperature and strain rates and
rates, as discussed in Section 3. Over characteristics were also governing they are determined in this study for
the strain rate range studied, the elastic factors when predicting stress-strain two foams.
modulus increases by approximately performance of low density foams.
100% and 200% for the N B50 and N This investigation found that the
A30 foams, respectively. From SEM studies, it was observed that
performance of low density foams
at all temperatures samples compressed
was sensitive to temperature. WLF
up to 50% strain do not show cell walls
Stress-strain data were fitted as a temperature shifting factors do not
rupture. This confirms that all results
function of strain rate using Nagy take into account the effect of the gas
for the 50% or less compression set
constitutive equation30 for both N A30 pressure-temperature dependency
tests are unaffected by temperature
and N B50. The fit for N B50 was found within the cell. In low density foam,
and therefore supports the application
to be good. However, the fit for N A30 the cell gas pressure can account for
approximately 75% of the material’s of empirical equation (Eq. 3).
was not so good. Nagy’s constitutive
equation has been modified by performance. For this reason any
including a strain dependent coefficient describing equation must include an SEM images of the N A30 foam
and the fit was much improved. additional term, based on the ideal gas samples show no significant damage
law, which accounts for the effect of in the form of cell wall rupture; even
The modulus of elasticity and stress temperature on the gas within the cell. after 80% compression strain is
at particular strain values are sensitive This then allows for the prediction of achieved. However, buckling of the
to temperatures close to the glass the combined behaviour of the polymer cell walls is evident. SEM images of
transition30. Since polyamides are and gas within the cells, thus giving an the N B50 foam samples show evidence
thermoplastics, literature indicates accurate determination of bulk material of cell wall rupture (approximately
that the modulus of elasticity of the stress-strain properties without the 10%) when 80% compression strain
base polymer, and foams made from need for additional experimentation. is achieved. This is consistent with
this polymer, may vary by 3 orders differences in the ductility of the base
of magnitude dependent upon either As stated by Zhang et al.28, the Nagy polymers.
change in temperature and or strain and WLF equations can be linked
rate2. This study found that the elastic to derive a combined strain rate and Both experimental and direct
modulus tripled for N A30 and doubled temperature equation for each material. measurement has determined that for

422 Polymers & Polymer Composites, Vol. 20, No. 5, 2012


Compression and Recovery Behaviour of Polyamide-6 Based Foams

the base polymers, the Tg is 48 ± 3 °C. 8. Ashby M.F. and Jones D.R.H., Isocyanate in Humid Aging and
This value is in agreement within the Engineering Materials 2 – An Durability of Molded seating Foam,
ranges given in literature for Nylon 6. Introduction to Microsctructures, Journal of Cellular Plastics 35
Processing & Design, 2nd edition, (1999) 492.
Butterworth and Heinemann, 1998. 20. BS EN ISO 1856, 2001
Acknowledgements 9. www.zotefoams.com, Date of Incorporating Amendments No
Access February 2011 1, Flexible Cellular Polymeric
The authors would like to acknowledge Materials – Determination of
10. BS 2782-3 METHODS 320A
the financial support from EPSRC for Compression Set.
TO 320F, 1976 Incorporating
project CASE/CNA/06/29 and from Amendment No.1 Plastics Part 3, 21. ASTM D395-03,2008 Standard
Zotefoams plc. Many thanks go to the Mechanical Properties - Methods Test Methods for Rubber Property
technical staff at Zotefoams plc for the 320A to 320F, Tensile Strength, Compression Set
supply of the samples and to Richard Elongation and Elastic Modulus. 22. N.J. Mills and A. Gilchrist, Creep
Giddens and Colin Bradsell for their 11. Greco R., Nicodemo L. and and Recovery of Polyolefin Foams
technical expertise. In addition our Nicolais L., The Glass Transition – Deformation Mechanisms,
thanks also go to Bryan Ruby and of Nylon 6, Macromolecules, 9(4) Journal of Cellular Plastics, 33
Trevor Nell for the manufacture of the (1976) 686-687. (1997) 264-292.
compression test fixtures. 12. Mills N.J., Plastics Microstructure, 23. BS EN ISO 1923,1995
Properties and Applications, Cellular Plastics and Rubbers
Edward Arnold, 1986. – Determination of Linear
REFERENCES Dimensions
13. Park K.W., Kim G.H. and
1. Mills N.J., Polymer Foams Chowdhury S.R., Improvement 24. BS EN ISO 845,1995 Incorporating
Handbook, Engineering and of Compression set Property of Amendments No.1 Cellular Plastics
Biomechanics Applications and Ethylene Acetate Copolymer/ and Rubbers- Determination of
Design Guide, 1st Edition, Chapter Ethylene-1-Butene Copolymer/ Apparent (Bulk) Density.
7, Butterworth-Heinemann, 2007. Organoclay Nanocomposite Foams, 25. Izzard V.G., Bradsell C.H.,
2. Gibson L.J. and Ashby F., Cellular Polymer Engineering and Science, Hadavinia H., Morris V.J., Foot
Solids, Structures and Properties, (June 2008) 1183. P.J.S. and Witten N., Investigation
2nd edition, Pergammon Press, 14. Sreeja T.D. and Kutty S.K.N., of the Compression Recovery
Oxford, 1999. Studies on Acrylonitrile Butadiene Properties of Polyamide-6 Cellular
3. Kraynik A.M. and Warren W.E., Rubber/Reclaimed Rubber Blends, Solid Over the Temperature Range
Low Density Cellular Plastics, Journal of Elastomers and Plastics, of -5oC to 90oC, Key Engineering
Physical Basis of Behaviour, 1st 34 (2002) 45. Materials 417-418 (2010) 933-936.
Edition, Chapter 7 - The Elastic 15. Jhlham I.S. and Maita I.J., Testing 26. SYSTAT User Manual, TableCurve
Behaviour of Low-Density Cellular and Evaluation of rubber-base 3D, Automated Surface Fitting and
Plastics, Chapman and Hall, 1994. Composites Reinforced with Equation Discovery Version 4.0 for
4. Mills N.J., Stampfli R., Marone Silica Sand, Journal of Composite Windows, 2002.
F. and Bruhwiler P.A., Finite Materials, 40 (2006) 2099. 27. BS EN 3386-1, 1998 Polymeric
Element Micromechanics Model 16. Nayak N.C. and Tripathy D.K., Materials, Cellular Flexible
of Impact Compression of Closed- Studies on Composites Based on Determination of stress-strain
Cell Polymer Foams, International Acrylonitrile Butadiene Rubber Characteristics in Compression,
Journal of Solids and Structures, 46 and High styrene Resin, Journal of Low Density Material.
(2009) 667-697. Elastomers and Plastics, 33 (2001) 28. Zhang J., Kikuchi N., Li V., Yees
5. Mills N.J. and Zhu H.X., The High 179. A. and Nusholtz G., Constitutive
Strain Compression of Closed-Cell 17. Bacci D., Marchini R. and Scrivani Modeling of Polymeric Foam
Polymer Foams. Journal of the M.T., Peroxide Crosslinking Material Subjected to Dynamic
Mechanics and Physics of Solids, of Ziegler Natta Thermoplastic Crash Loading, Int. J. Impact
47 (1999) 669-695. Polyolefins, Polymer Engineering Engineering 21, (1998) 369-386.
6. Mills N.J. and Hwang A.M.H., The and Science, 44 (2004) 131. 29. Nagy A., Ko W.L. and Lindholm
Multiple - Impact Performance of 18. Coons J.E., McKay M.D. and U.S., Mechanical Behaviour of
High - Density Polyethylene Foam, Hamada M.S., A Bayesian Analysis Foamed Materials under Dynamic
Cellular Polymers, 9 (1989) 259-276. of the Compression Set and Stress- Compression. J. Cellular Plastics,
7. Rodriguez-Perez M.A., Diez- Strain Behaviour in a Thermally 10, (1964) 127-134
Gutierrez S., De Saja J.A., The Aged Silicone Foam, Polymer 30. Mano J.F. and Viana J.C., Effects
recovery behaviour of crosslinked Degradation and Stability, (2006) at the Strain Rate and Temperature
closed cell polyolefin foams, 1824. in Stress Strain Tests, Study of the
Polymer Engineering and Science, 19. Neff R.A. and Marsalko T.M., Glass Transition of a Polymide – 6,
38 (1998), 831-837. Roles of Convention Polyol and Polymer Testing, 20 (2001) 937-943.

Polymers & Polymer Composites, Vol. 20, No. 5, 2012 423


V.G. Izzard, H. Hadavinia, V.J. Morris, P.J.S. Foot, L. Wilson and K. Hewson

424 Polymers & Polymer Composites, Vol. 20, No. 5, 2012

View publication stats

You might also like