Progress in Organic Coatings 50 (2004) 6267

The influence of pigment volume concentration (PVC)

on the properties of an epoxy coating
Part I. Thermal and mechanical properties
M.T. Rodrguez a, , J.J. Gracenea b , A.H. Kudama a , J.J. Suay a
a

Technology Department, Materials Science and Metallurgical Engineering, Jaume I University,

Campus Riu Sec. Avda. Sos Baynat, s/n, 12071 Castelln, Spain
b Naber Pinturas y Barnices, S.L., N-332, km 279, 46469 Beniparrell-Valencia, Spain

Received 15 March 2003; received in revised form 26 September 2003; accepted 23 October 2003

Abstract
The effect of the ratio between pigment volume concentration (PVC) and critical pigment volume concentration (CPVC), denoted
hereafter by , on the thermal and mechanical properties of an epoxy primer was investigated by thermogravimetric analysis (TGA),
differential scanning calorimetry (DSC), dynamic-mechanical analysis (DMA), stressstrain tests and measurement of Persoz hardness.
Paints based on epoxy resin were prepared at different values (0.73, 0.80, 0.88, 0.93, 0.98, 1.08, 1.15 and 1.24). It was concluded that
thermal and mechanical properties can be greatly affected by the PVC/CPVC ratio and in a complex manner.
2004 Elsevier B.V. All rights reserved.
Keywords: CPVC; Primer; DSC; DMA; Mechanical analysis

1. Introduction
The critical pigment volume concentration (CPVC) was
first clearly defined and experimentally measured in the seminal work of Asbeck and Van Loo from 1949 [1]. Since that
time, the CPVC concept has been the basis of much work,
including that of Asbeck [2], which examines the performance of coating according to how it is affected by variations in the level of pigmentation. The practical significance
of determining this property is that the CPVC is a transition
point and many properties of the paint undergo a marked
change above the CPVC [3]. Bierwagen and Hay, in a 1975
paper [4], divide the affected properties into three groups:
(1) Mechanical and thermal properties. Flexibility, tensile
strength, block resistance, glass transition temperature
(Tg ), scrub resistance, and cold crack resistance. These
properties depend significantly on binder performance,
which can be modified by pigment addition.
(2) Transport properties. Permeability, scrubbability, stain
removal, corrosion resistance, blistering, and enamel
Corresponding author. Present address: INIA, Dpt. Mejora Genetica
y Biotecnologia, Ctra. La Coruna km 7, Madrid 28040, Spain.
Tel.: +34-91-347-39-13; fax: +34-91-357-22-93.
E-mail address: rblasco@tec.uji.es (M.T. Rodrguez).

0300-9440/$ see front matter 2004 Elsevier B.V. All rights reserved.
doi:10.1016/j.porgcoat.2003.10.013

holdout, penetration over porous substrates, and electrical resistance. These are dependent on the flow of materials or electrical current through the coating.
(3) Optical properties. Gloss, opacity, and tint acceptance.
These are properties of the dry films appearance.
At PVC below the CPVC, a dry coating film is a continuous coating, a composite consisting of pigment particles
randomly embedded in a continuously connected matrix of
polymer. Above the CPVC, there are void structures in the
film due to insufficient polymer, but the pigment particles
can still be thought of as being continuously connected. The
polymer is still continuously connected globally, but voids
will cause the polymer to lose local connectivity, and thus the
sharp drop-off in mechanical properties of the coating just
above the CPVC. A new phase, air, is now present in the film
and its properties drastically affect those of the film, especially with regard to density, mechanical, thermal, transport
properties and optical properties [5]. Above CPVC there is
not enough resin to cover the pigment surface and an important amount of the polymer is absorbed in the pigment.
The main aim of the present work was to study the effect of
varying the PVC/CPVC ratio of an epoxy coating system by
means of different characterization (mechanical and thermal)
techniques.

M.T. Rodrguez et al. / Progress in Organic Coatings 50 (2004) 6267

2. Experimental

63

free polymer films by just leaving the material to dry and

cure over a non-adhesive polymer substrate.

2.1. Paints
2.3. Thermogravimetric analysis (TGA)
A primer was formulated using a type I epoxy resin based
on Bisphenol A; an aliphatic polyamine was added in the
component B of the formulation and a tertiary amine was
used as a catalyst. The values used in this work were
obtained by the procedures developed by Bierwagen [6].
Using this methodology, CPVC values were calculated from
pigment properties according to the following relationship:
CPVC =
P

(1)

a
i=1 Xi (1 + Ui )

where is the densest random packing factor for dry paint;

i in the subscript describes the increases in particle size 1, 2,
. . . , m; P the number of pigments in the paint; Xi represents
the volume fraction of pigment containing particles of size
i and Uia the volume of oil absorbed in pigment of particle
size i.
The pigment used in the study was talc and calcium carbonate mixture including silica-based ion exchange compound as corrosion inhibitor.
A primer without pigments was prepared in order to establish better properties in comparison with pigmented paints.
Obviously, in this case the primer contains resin, crosslinking agent and catalyst only. Characteristics of the different
mixtures that were formulated are shown in Table 1.
2.2. Substrates
Cold-rolled low-carbon steel test panels were used in order to ensure uniform substrate surfaces for testing purposes.
All test panels were degreased with acetone, and paints were
deposited by means of a graduated application bar. The systems were allowed to cure under ambient air conditions for
1 week prior to testing. Paint dry film thickness was determined by using an Elcometer and the thicknesses were
maintained within the 50 5 m range.
Samples used to perform TGA, DSC, DMA and
stressstrain tests were not obtained by material deposition
on any metallic substrates. They were usually obtained as
Table 1
Specifications of studied samples
Solid in
weight (%)

Solid in
volume (%)

PVC (%)

PVC/CPVC

70.8
70.3
69.8
69.5
69.2
68.7
68.4
68.0

55.4
53.7
52.2
51.1
50.0
48.3
47.0
45.5

36.00
39.66
43.28
45.82
48.62
53.19
57.00
61.38

0.73
0.80
0.88
0.93
0.98
1.08
1.15
1.24

In order to determine whether any residual organic solvent had been trapped because of pigment addition, thermogravimetric tests were carried out for each of the primers
formulated, which had been left to cure for 1 week. A Mettler thermogravimetric analyzer connected to a Balzers mass
spectrometer was used. Samples of approximately 1520 mg
were scanned up to 900 C at 10 K min1 . All scans were
performed with an argon flow of 50 cm3 min1 .
2.4. Differential scanning calorimetry (DSC)
A Perkin Elmer DSC7 differential scanning calorimeter
was employed for isothermal and dynamic scans in order to
determine the reaction heat of curing of samples with different and to measure the Tg of samples cured at ambient temperature during 1 week. The DSC was calibrated
with a high purity indium sample. The test samples weighed
from 8 to 12 mg. In the isothermal curing experiments, a
steady isothermal baseline was established at the curing temperature by means of two empty sample pans. A flow of
20 cm3 min1 of nitrogen was used as purge gas. Isothermal
results did not show any good results probably as a consequence of the presence of solvent in the formulated but not
cured samples. DSC scans at 10 K min1 were performed
on cured samples (at ambient temperature during 1 week)
to detect any residual exothermic reactions and to determine
the midpoint Tg .
2.5. Mechanical tests (stressstrain curves)
Stressstrain curves of cured primers with different
were obtained with an Instron universal machine. Cured
samples were standardly cut and a 100 N cell was used to
perform traction tests according to ASTM D1708-96 [7].
2.6. Dynamic-mechanical thermal analysis (DMA)
A dynamic-mechanical Analyzer (Perkin Elmer DMA)
was employed to perform mechanical measurements from
100 to 150 C at 1.5 K min1 on samples with different
PVC/CPVC. The tests were carried out in the dual cantilever
mode and at a frequency of 1 Hz. All samples tested were
cured previously for 1 week at ambient temperature.
2.7. Persoz hardness
The procedure for the measurement of hardness of organic
coatings by pendulum damping test (according to ASTM
D 4366-95 [7]) is based on the principle that the amplitude of oscillation of a pendulum touching a surface de-

M.T. Rodrguez et al. / Progress in Organic Coatings 50 (2004) 6267

creases more rapidly the softer the surface is. A pendulum resting on a coating surface is made to oscillate and
the oscillation number decreases by a specified, measured
amount (in this study the oscillation decrease considered
was from 12 to 4 ). Three tests were performed to each
coated sample and the final result was the average of the
three.
3. Results and discussion

110
100

Weight Loss TG (%)

64

90
80
70

=1.24
=1.08

60

=0.88

50

=0.73

40

3.1. Thermogravimetric analysis

Fig. 1 shows thermogravimetric analysis results for cured

primer formulated at = 0.93. As can be seen, thermogravimetric mass loss signal (TG) and the derived mass loss
versus time signal (DTG) have quite complicated spectra.
This can be attributed to the number of different materials
used in the primer.
There is an initial weight loss until 100 C that can be correlated (using the mass spectrometer signal) to water evaporation. Up to 200 C there is a second weight loss that is
attributed to plastifying agent degradation and finally until
550 C there is a significant peak in the DTG signal which
corresponds to the epoxy resin (used as a binder) degradation. Up to this temperature there are two peaks in the DTG
signal, which can be correlated to some degradations concerning inorganic pigments.
Fig. 2 shows TG signals of cured primers formulated
with = 0.73, 0.88, 1.08 and 1.24. It can be clearly seen
that the spectra are quite similar and how, as increases,
the total weight loss at 550 C decreases because there is
less epoxy resin present in the formulation. These thermogravimetric tests did not manage to detect any presence
of solvent retained in the dry, cured samples used in this
study.

200

400

600

800

1000

T(C)

Fig. 2. Mass loss signal (TG, %) vs. time signal for dry, cured primer
formulated with = 0.73, 0.88, 1.08 and 1.24.

3.2. Thermal analysis (DSC)

Isothermal and dynamic scans were carried out on uncured samples in order to determine curing reaction kinetics. Results did not offer good results, probably due to the
presence of solvent in all these samples. Good results were
only obtained when DSC scans were performed on cured
samples. In all these tests some residual enthalpies (Hresid )
were detected when temperature scans were applied to cured
primers formulated with a different parameter.
It can be seen in Fig. 3 that the Tg value varied significantly with pigment volume content. Tg behavior depends
on whether the primer was formulated with a value lower
than 1 (in this case as increases, Tg decreases, for example, in 25 C for = 0.98 with respect to primer formulated
with 0.73), or whether the primer was formulated with a
value higher than 1 (in this case as increases, Tg increases
too, for example, at 28 C when = 1.08 in comparison
with primer formulated with = 0.98).
0,2

100

0,1
90

-0,1

80

-0,2
70
-0,3

DTG (% min -1) 2

Weight Loss TG (%)

0,0

-0,4

60

-0,5
50
0

200

400

600

800

1000

T (C)
Fig. 1. Mass loss signal (TG, %) () and derived mass loss vs. time signal DTG (% min1 )2 (- - -) for dry, cured primer formulated with = 0.93.

M.T. Rodrguez et al. / Progress in Organic Coatings 50 (2004) 6267

65

120

10

10

10

10

10
150

90

-10

30

Tan

-20

T (C)

60

-15

-1

10

-25
0

Modulus (Pa)

10

-5

Tg (C)

10

-30
-2

-30
0.0

0.2

0.4

0.6

0.8

1.0

10

-150

1.2

-100

-50

When > 1, there can be a greater adsorbance of the

plastifying agent by pigment particles increasing the
crosslinking density.
When > 1, pigment particles reduce their distance and
the epoxy system can be rigidized because of a confinement phenomena.
When > 1, the epoxy resin curing reactions only take
place inside the pigment particle or on its surface (there
is only enough resin to wet the pigment because of the
high level of PVC), and no reactive groups are hindered
by the inorganic material, so there is no decrease in the
encounter probability between reactive groups.

100

Fig. 4. Dynamic-mechanical loss tangent (tan ) and storage modulus (E )

for primers formulated with different ( < 1): 0 ( ), 0.73 (), 0.80
(), 0.88 ( ).

It is probable that, as a consequence of these results, when

PVC increases (above CPVC) the epoxy curing mechanism
can change and there is an increase in the system crosslinking density which allows a decrease in the free volume in it
and a reduction of its permeability to ions, water and gases
(not in the global primer but only in the binder).
3.3. Dynamic-mechanical results
Figs. 4 and 5 show the curves for loss tangent (tan ) and
the real part of the complex modulus, the storage modulus
(E ) of the dynamic-mechanical test performed for cured
primers formulated with values lower than 1 (Fig. 4) and
values higher than 1 (Fig. 5).
The main dynamic-mechanical (or ) relaxation of amorphous polymers is triggered by the glass transition of the
system. With the increase of PVC (and while PVC < CPVC)
the maximum of the loss tangent shifts to much lower temperatures (Figs. 3 and 4). The temperature shifts of this maximum of the relaxation spectra (T ) are of the same order

-1

10

Tan

It is known that the Tg of an epoxy system increases with

crosslinking density as the curing reaction progresses [8,9],
so Tg is a good indicator of the curing degree of the system.
This is specially true in the last stages of the reaction, when,
in other thermal tests (like isothermal or dynamic scans), the
heat flow signal is below the sensitivity of the apparatus [10]
or when the presence of other materials (such as solvent) do
not allow us to obtain good representative DSC results in
isothermal or dynamic scans.
When primers are formulated with increasing values,
but which are below 1, a trapping effect by the pigments
can reduce the mobility of the molecules of the reactive
system, which decreases the encounter probability between
the epoxy resin and the curing agent species, giving rise to a
proportional increase in viscosity of the medium. Regions in
the bulk material can thus be formed where epoxy and amide
groups remain non-reactive and this has a plasticizing effect
and reduces Tg [1113]. As a consequence of these results
it can be stated that when PVC increases (until CPVC) there
is a crosslinking density reduction in the polymeric system
which will allow an increase in free volume in the epoxy
resin and an increased permeability to water, ions and gases.
When primers with a PVC above the CPVC are tested, a
significant increase in the Tg value is observed. Possible explanations can lie in a combination of physical and chemical
phenomena, such as:

50

T (C)

PVC/CPVC

Fig. 3. Tg () and T () (temperature of the maximum of the principal

mechanical relaxation peak) of cured samples formulated at different
values.

-2

10

-100

-50

50

10

11

10

10

10

10

10

10

Modulus (Pa)

-35

100

T (C)

Fig. 5. Dynamic-mechanical loss tangent (tan ) and storage modulus (E )

for primers formulated with different : PVC/CPVC ( 1): 0.98 (),
1.08 (), 1.15 (+), 1.24 ( ).

66

M.T. Rodrguez et al. / Progress in Organic Coatings 50 (2004) 6267

10

10

10

1.24
1.15
9

E25C

10

10

10

Stress (MPa)

10

E-100C

10

1.08

0.88

0.93
0.80

0.98

0.73

1
6

10
0.0

0.2

0.4

0.6

0.8

1.0

1.2

PVC/CPVC

100 C

25 C

Fig. 6. Vitreous at
() and rubbery at
for primers formulated with different values.

() storage modulus

of magnitude as the Tg differences found by DSC. A lower

crosslinking density because of the plastifying effect due to
unreacted epoxy and amine groups as the PVC increases
may be responsible for this phenomenon [7].
With further increases in PVC (and when PVC > CPVC),
the maximum of the loss tangent shifts to much higher temperatures (Figs. 3 and 5). A higher crosslinking density as
is increased may be responsible for this phenomenon.
Fig. 6 shows the logarithm of the storage modulus in

the glassy state (log E100
C ) and in the rubbery state

(log E25 C ) of the primers formulated with different
values.
There are two processes that can affect these results: a
mechanical effect due to reinforcement of pigment particles, and epoxy binder changes in crosslinking density and
rigidity.
When increases there are no significant changes in

log E25
C until = 1. If PVC increases above CPVC,

log E25
C increases because of the reinforcement effect of
the pigment and because of an increase in the crosslinking
density and in the rigidity of the binder.

However, an important decrease in log E100
C occurs
when 0.73 < < 1. This is probably due to big changes
in the matrix nature (a decrease in crosslinking density) and
mainly to changes in the adhesion between pigment parti
cles. When > 1, there is a notable increase in log E100
C
because of the reinforcement effect of the pigment and
an increase in the crosslinking density and rigidity of the
binder.
There is also an increase in the heterogeneity of the types
of matrix chains that constitute the epoxy binder as the
value increases, which is responsible for the broadened of
the relaxation that can be clearly seen in the tangent
signal (Figs. 4 and 5). The reason for this phenomena may
be that there is more and more epoxy resin that cures on
the surface of the pigments or inside them, and less polymer
curing as bulk material as increases. This process may be
responsible for a different curing mechanism in the epoxy

0,1

0,2

0,3

0,4

0,5

0,6

Strain (%)

Fig. 7. Stressstrain curves for cured primers formulated with different

values.

resin that will have a less homogeneous chemical structure

as PVC increases.
3.4. Mechanical tests (stressstrain curves)
Fig. 7 shows stressstrain curves for cured primers formulated with different values. Table 2 shows the mechanical
parameters for these samples: Youngs modulus (E), elastic limit stress ( elastic limit ), elastic limit strain (elastic limit ),
rupture stress ( rupture ), and rupture strain (rupture ).
It can be clearly observed that the pigment volume present
in the primer has a big effect on the mechanical properties.
As PVC increases the coating becomes more and more rigid
and fragile, and less ductile (E, elastic limit , rupture parameter values increasing and elastic limit and rupture parameter
values decreasing). This effect is especially significant in
coatings formulated with PVC > CPVC.
As PVC increases until CPVC there are not too many
changes in Youngs modulus (the same effect as the storage modulus). However, the coating is much less ductile
(as can be observed in the stressstrain curves and in the
rupture parameter).
Table 2
Mechanical parameters for cured primers formulated with different
values

E (MPa)

elastic limit
(MPa)

elastic limit
(%)

rupture
(MPa)

rupture
(%)

0.73
0.80
0.88
0.93
0.98
1.08
1.15
1.24

9.532
16.784
26.790
20.767
46.141
76.691
248.067
228.587

0.163
0.292
0.488
0.448
0.610
1.279
1.798
2.158

0.017
0.017
0.019
0.022
0.017
0.018
0.010
0.012

2.240
2.373
2.869
2.810
2.452
3.024
3.568
3.452

0.552
0.348
0.299
0.311
0.168
0.111
0.063
0.037

M.T. Rodrguez et al. / Progress in Organic Coatings 50 (2004) 6267

Changes can be due to two effects that act in opposite

directions: on the one hand, a reinforcement process due to
pigment particle addition and, on the other hand, a decrease
in the crosslinking density of the binder combined with a
loss of adhesion between particles.
When 1 (samples with = 0.98, 1.08, 1.15 and
1.24), mechanical changes are far more significant. In this
case there are two effects that act in the same direction: a
reinforcement effect because the pigment content increases,
and higher crosslinking density and rigidity in the polymer
matrix.
Primer formulators are especially interested in getting
the maximum adhesion between the organic coating and the
metal substrate in order to ensure better anticorrosion primer
properties. Since high mechanical parameter values such as
E, elastic limit or rupture correspond to high rigidity characteristics, when these types of epoxy coatings are formulated
at high PVC values a good adhesion to the substrate cannot
be expected because of the fragility of the coatings. Epoxy
binder in the coating is normally under stress conditions
(internal stress generated during the curing process, water
uptake, thermal cycles, and so on) and if it is fragile a delamination process can take place at the coating/metal interface.

67

faster with time and reaches higher values. Higher pigmentation volume content allows faster solvent evaporation, thus
reducing physical drying time. On the other hand, when
PVC > CPVC much of the resin can cure inside pigment
particle or on its surface which can produce a catalytic effect
in the resin epoxy resin process. An increase in the curing
kinetics of this type of anticorrosive primers is especially
important to enhance recoatability with polyurethane top
coats.

4. Conclusions
The thermal and mechanical characteristics of different
paints based on epoxy resin and prepared at different
values (0.73, 0.80, 0.88, 0.93, 0.98, 1.08, 1.15 and 1.24)
were studied.
It has been determined that as values increase, there
is a significant change in Tg , T , Persoz hardness and
stressstrain curves. The trend of some parameters (Tg and
T ) was completely different when > 1 or < 1.
Acknowledgements

3.5. Persoz hardness

One possible reason for formulating at high PVC is to try
to enhance the curing process. Fig. 8 shows Persoz hardness
versus time for each composition with a different value.
As the parameter increases, Persoz hardness increases

References

160
=1.24

140
120
=1.15

Oscillations

The authors are grateful to Ms. Eva Romero, Mr. David

Roures and Mr. Jos Ortega for their assistance during experiments and to project MAT2000-0123-P4-03 for the financial support.

100
=1.08

80

=0.98
=0.93
=0.73

60

=0.88

40
20
0

12

24

36

48

60

Time (hours)

Fig. 8. Persoz hardness evolution (in number of oscillations) vs. time for
primers formulated with different values = 0.73 (), 0.88 ( ), 0.93
(), 0.98 (), 1.08 (), 1.15 (+), 1.24 ( ) and applied over carbon steel
substrates.

[1] W.K. Asbeck, M. Van Loo, Critical pigment volume relationships,

Ind. Eng. Chem. 41 (1949) 1470.
[2] W.K. Asbeck, J. Coat. Technol. 64 (806) (1992) 4758.
[3] R.W. Braunshausen Jr., R.A. Baltrus, L. de Bolt, J. Coat. Technol.
64 (810) (1992) 5154.
[4] G.P. Bierwagen, T.K. Hay, Prog. Org. Coat. 3 (1975) 281.
[5] G.P. Bierwagen, J. Coat. Technol. 64 (806) (1992) 7175.
[6] G.P. Bierwagen, J. Paint Technol. 44 (1972) 47.
[7] Annual Book of ASTM Standards, American Society for Testing
and Materials, Philadelphia, PA, 1995.
[8] J. Mijovic, Appl. Polym. Sci. 3 (1986) 1177.
[9] D.J. Plazek, J. Appl. Polym. Sci. 24 (1986) 1303.
[10] G. Wisanrakkit, J.K. Gillham, J.B. Enns, J. Appl. Polym. Sci. 4
(1990) 1895.
[11] M.F. Grenier-Loustatot, P. Grenier, Polymer 33 (1992) 6.
[12] J.J. Suay, M. Monlen, J.L. Gmez, Polym. Int. 48 (1999) 1269
1276.
[13] J.J. Suay, M. Monlen, J.L. Gmez, Polym. Eng. Sci. 40 (8) (2000)
17251735.