Polymer 41 (2000) 579–594

Molecular transport of aromatic solvents in isotactic polypropylene/

acrylonitrile-co-butadiene rubber blends
S. George a, K.T. Varughese b, S. Thomas a,*
a
School of Chemical Sciences, Mahatma Gandhi University, Priyadarshini Hills P.O., Kottayam, Kerala 686 560, India
b
Polymer Laboratory, Central Power Research Institute, Bangalore 560 094, India
Received 28 October 1998; received in revised form 15 February 1999; accepted 12 March 1999

Abstract
The sorption and diffusion of a series of aromatic hydrocarbons through blends of isotactic polypropylene/nitrile rubber have been
investigated. The effects of blend ratio, vulcanizing systems and different fillers on the transport behavior were studied. In the blends, the
uptake and diffusion coefficient increase with an increase in rubber concentration, and decrease with the molar volume of solvents. The
diffusion data were analyzed and it has been observed that most of the systems follow a Fickian mode of transport. The effect of temperature
on the transport behavior was investigated at four temperatures. The activation energy for the diffusion, and the thermodynamic parameters
like entropy and enthalpy were calculated. The experimental diffusion data were correlated with various theoretical predictions. q 1999
Elsevier Science Ltd. All rights reserved.
Keywords: Blends; Polypropylene; Acrylonitrile-co-butadiene rubber

1. Introduction in many applications like oil seals, gaskets, etc. For the last
few decades, the improvement of the performance of the
During the last few decades the importance of polymer nitrile rubber has been tried by blending with various poly-
blends have increased, since it is possible to achieve desir- mers [13–16]. The blending of NBR with polypropylene
able properties by simple blending of polymers. Among the was found to improve the physical and mechanical proper-
different types of polymer blends thermoplastic elastomers ties of NBR [17,18]. Hence it is necessary to analyze the
have their own advantages as they combine the processa- transport behavior and the mechanism of transport of
bility of plastics with the performance of vulcanized rubbers various organic solvents through PP/NBR blends in detail.
[1,2]. The possibility of developing new thermoplastic elas- The transport of small molecules through polymers has
tomers like polypropylene (PP)/natural rubber (NR), PP/ been widely studied by various research groups [19–27]. It
poly(ethylene-co-vinyl acetate) (EVA), PP/ethylene propy- was found that the transport of the solvents through poly-
lene diene rubber (EPDM), high density polyethylene mers is influenced by the physical and chemical structure of
(HDPE)/NR, HDPE/NBR, nylon/EPDM, etc. has been polymers, the crosslink density, the shape and size of
investigated by various researchers [3–7]. However, many solvent molecules and temperature. The effect of various
of these polymer blends are incompatible or immiscible and fillers on the transport behavior of rubbery polymers was
are characterized by narrow interface and weak interfacial investigated [28–30], and it was observed that the presence
interaction. They often exhibit poor mechanical properties of fillers makes a tortuous path to the transport of solvents
[8]. These problems associated with immiscible polymer through polymer samples and thereby reduces the solvent
blends could be alleviated by using different techniques uptake. However, the transport behavior is also affected by
like compatibilization and dynamic vulcanization [9–11]. the interaction of filler between the polymer and the solvent
The transport behavior of various organic solvents and [30].
gases through polymers is of great technological impor- With the increasing importance of polymer blends in the
tance, since nowadays the polymer membranes are increas- transport behavior of organic solvents and gases, studies
ingly used in various barrier applications [12]. Nitrile have been directed at polymer blends as well [31–34]. In
rubber, which is an oil resistant elastomer, is widely used polymer blends, the transport behavior depends on the
miscibility of the component polymers as well as the
* Corresponding author. Fax: 191-481-561190. morphology of the system [31]. Hence the transport
0032-3861/00/$ - see front matter q 1999 Elsevier Science Ltd. All rights reserved.
PII: S0032-386 1(99)00208-6
580 S. George et al. / Polymer 41 (2000) 579–594

Table 1 of error in these measurements is that arising during weigh-

Formulation of dynamically vulcanized blends ing since the sample has to be taken out from the solvent for
Sample P70S P50S P50C P50M P30S P0S weighing. However, since the weighing is completed within
20–30 s, the error can be neglected [30]. The results of
PP 70 50 50 50 30 – diffusion experiments were expressed as moles of solvent
NBR 30 50 50 50 70 100
uptake by 100 g of polymer sample, Qt mol%.
ZnO 5 5 – 5 5 5
Stearic acid 2 2 – 2 2 2
 
Mass of solvent sorbed
Sulfur 0.2 0.2 – 0.1 0.2 0.2
Molar mass of solvent
TMTD a 2.5 2.5 – 2.5 2.5 2.5 Qt mol% ˆ × 100 …1†
CBS b 2 2 – 2 2 2 Mass of polymer
DCP c – – 2 1 – –

a
Tetramethyl thiuram disulfide.
b
N-cyclohexyl benzothiazyl sulfenamide. 3. Results and discussion
c
Dicumyl peroxide.
3.1. Effect of blend ratio
phenomena in polymer blends can be used as a character-
ization technique, i.e. in order to understand the miscibility The transport behaviors of unvulcanized and dynamic
as well as morphology of the system. In this paper we have vulcanized thermoplastic elastomers from PP and NBR
investigated the transport of a series of aromatic solvents were analyzed. The results of the analysis of diffusion
through PP/NBR blends. The effects of blend ratio, type of experiments of the unvulcanized and vulcanized blends in
crosslinking and different fillers on the transport phenomena toluene are shown in Fig. 1(a) and (b), respectively, as
have been studied. The experimental results have been mol% uptake (Qt) vs. square root of time. Polypropylene
compared with theoretical predictions. has the minimum uptake of toluene while NBR has the
maximum level. The blends show an intermediate behavior,
i.e. between that of the pure components. The lowest value
2. Experimental
of mol% uptake for polypropylene in spite of the match in
solubility parameter between PP and toluene, is due to the
2.1. Materials
crystallinity of PP (Table 2). In a semicrystalline polymer,
Isotactic polypropylene having an MFI of 3 g/10 min was some amorphous part is also present along with crystalline
supplied by IPCL, Baroda. Acrylonitrile-co-butadiene regions. Only this amorphous region will contribute to the
rubber (NBR) with an acrylonitrile content of 34% was uptake of solvent and hence PP has the lowest uptake value
purchased from Synthetics and Chemicals, Bareli, U.P. among all the proportions of the blend. In the case of blends,
The solvents benzene, toluene and xylene were of analytical as the concentration of NBR increases, the crystalline
grade. The fillers used were HAF-black cork and silica trea- content decreases. Hence the hindrance for the transport
ted with silane coupling agent. The blends were prepared in of toluene decreases and uptake increases. In the blends
a Brabender plasticorder by melt mixing at 1808 C. The the crystalline PP phase makes a tortuous path to the trans-
binary blends are denoted as P100, P70, P50, P30, P0, where port of solvent through the amorphous regions in blends.
the subscripts denote the wt% of NBR in the blend. The Fig. 2 shows the variation of Q∞ of PP/NBR blends with
formulation of dynamic vulcanized blends is shown in blend composition. As the concentration of NBR increases,
Table 1. In filled blends the filler loading of 30 phr is used the Q∞ value increases linearly up to 50 wt% NBR after
and the blends are designated as P50Tsi30, P50C30 and P50K30 which a change in slope of the curve is observed. This
respectively for silane treated silica, carbon black and cork difference in Q∞ values with blend composition can be
filled P50 blends. correlated with the morphology of the system. The scanning
Samples for transport studies were prepared by compres- electron micrographs of P70, P50 and P30 blends are shown in
sion molding the sample in a hydraulic press into 2 mm Fig. 3. It is seen that in P70 and in P50 NBR is dispersed as
thick sheets at 1808C. The circular shaped samples were spheres in the continuous PP matrix. In the P30 blend both
punched out from the sheet using a sharp edged die. NBR and PP form continuous phase leading to a co-contin-
uous morphology. Because of the dispersed nature of NBR
2.2. Sorption experiments phase in P70, the continuous PP phase acts as a tortuous path
for the diffusion process of solvent and hence the uptake is
The samples were soaked in 20 ml solvent in diffusion less. This is schematically shown in Fig. 4. However, as the
bottles and kept at constant temperature by keeping in a concentration of NBR increases from 30 to 50 wt%, the size
thermostatically controlled heating oven. The weight of of NBR domains increases which increases the contact
the swollen samples was measured at frequent intervals between NBR particles and crystallinity of the blends
until equilibrium swelling is reached. The experiments decreases (Table 2). This leads to the high uptake in P50
were conducted at 30, 40, 50 and 658C. A possible source compared to P70. In P30, where NBR forms the continuous
 S. George et al. / Polymer 41 (2000) 579–594 581

Fig. 1. (a) Variation of mol% toluene uptake (Qt) with square root of time of unvulcanized PP/NBR blends; (b) variation of mol% toluene uptake (Qt) with
square root of time of dynamically vulcanized PP/NBR blends.

phase the diffusion process is continuous through NBR where Qt and Q∞ are the mol% sorption at time t and at
phase and hence a sharp increase in uptake is observed. equilibrium, respectively, and k is a constant that depends
The mechanism of transport of PP/NBR blends was on the structural characteristics of the polymer and gives
analyzed using the empirical relation [5]. information about the interaction between the polymer
and solvents. The value of n ˆ 0.5 indicates a Fickian
log…Qt =Q∞ † ˆ log k 1 nlog t …2† mode of transport, while n ˆ 1 indicates case II (relaxation
582 S. George et al. / Polymer 41 (2000) 579–594

Table 2 i.e. the penetrant molecules are first sorbed by the polymer
Solubility parameter difference and crystallinity of PP/NBR blends followed by diffusion through the polymer. The net diffu-
Sample Solubility % Crystallinity (from DSC data) sion through polymer depends on the difference in the
parameter amount of penetrant molecules between the two surfaces.
difference Hence, the permeability [19]
between
polymer PˆD×S …3†
and solvent
(cal/cm 3) 1/2 where D is the diffusivity and S, the solubility, and the
values of S are taken as grams of liquid sorbed per gram
P100 0.20 55.30 of rubber.
P70 0.12 33.90
P50 0.34 20.90
The kinetic parameter, the diffusion coefficient D can be
P30 0.56 13.70 calculated using the equation [19,26],
P0 0.88 –
X∞  
Qt 8
e2…2n11† p …Dt=h †
2 2 2
ˆ12 …4†
Q∞ nˆ0 …2n 1 1† p
2 2

controlled) transport. The value of n between 0.5 and 1

indicates an anomalous transport behavior. The values of where t is the time and h is the initial thickness of the
n and k for PP/NBR blends are obtained by regression analy- polymer sheet. Although this equation can be solved readily,
sis of log…Qt =Q∞ † vs. log t plot. The results of the analysis it is instructive to examine the short-time limiting expres-
are given in Table 3. Since the values range between 0.46 sion as well.
and 0.6, i.e. in PP/NBR blends, the mode of transport is Qt
close to Fickian. From the table it is seen that as the concen- ˆ ‰4=p1=2 Š‰Dt=h2 Š1=2 …5†
Q∞
tration of NBR increases the values of n slightly increase
and approach the Fickian mode. For the Fickian mode of From a plot of Qt vs. t 1/2, a single master curve is obtained
transport, the rate of diffusion of permeant molecules is which is initially linear. Thus D can be calculated from a
much less than the relaxation rate of the polymer chains. rearrangement of Eq. (5) as
Usually rubbers and semicrystalline polymers exhibit Fick-  
hu 2
ian mode of diffusion [31]. Dˆp …6†
4Q∞
The transport of small molecules through polymers
generally occurs through a solution diffusion mechanism, where h is the sample thickness, and u is the slope of the

Fig. 2. Variation of equilibrium uptake (Q∞) of PP/NBR blends with weight percentage of NBR.
 S. George et al. / Polymer 41 (2000) 579–594 583

with increase in rubber content is due to the increase in

concentration of NBR with a high diffusion coefficient
and also due to the fact that the tortuosity exhibited by PP
to the penetration of solvent molecules decreases with an
increase in NBR content. The change in P and D with rubber
content is small up to 50 wt% NBR, after which a sharp
increase is observed in these values. This sharp change in
the values is due to the phase inversion of NBR from
dispersed to continuous phase on passing from 50 to
70 wt% NBR, which leads to an increased diffusion process.
In the case of heterogeneous polymer blends, the perme-
ability can be interpreted in terms of various theoretical
models. Robeson’s two limiting models, namely series
and parallel models, are generally used in the case of poly-
mer blends.
According to the parallel model
Pc ˆ P 1 f 1 1 P 2 f 2 …7†
and by the series model
Pc ˆ P1 P2 =…f1 P2 1 f2 P1 † …8†
where Pc, P1 and P2 are the permeation coefficients of the
blend, component I and component II, respectively, and f 1
and f 2 are the volume fractions of components I and II,
respectively.
Further, for a conducting spherical filler, the overall
composite permeation coefficient is given by Maxwell’s
equations as [31,35]
Pd 1 2Pm 2 2fd …Pm 2 Pd †
Pc ˆ Pm …9†
Pd 1 2Pm 1 fd …Pm 2 Pd †
where the subscripts d and m correspond to the dispersed
phase and the matrix, respectively.
Robeson [35] extended Maxwell’s analysis to include the
continuous and discontinuous characteristics of both phases
at intermediate compositions and expressed the equations as:
" #
P2 1 2P1 2 2f2 …P1 2 P2 †
Pc ˆ Xa P1
P2 1 2P1 1 f2 …P1 2 P2 †
" #
P1 1 2P2 2 2f1 …P2 2 P1 †
1 X b P2 …10†
P1 1 2P21 1 f1 …P2 2 P1 †

where Xa and Xb are the fractional contributions to the contin-

uous phase so that Xa 1 Xb ˆ 1.
Fig. 3. SEM micrographs of PP/NBR blends: (a) 70/30; (b) 50/50 and; (c) Fig. 6 shows the various theoretical and experimental
30/70 PP/NBR blends.
curves showing the variation of P with NBR volume frac-
tion. The P values of PP/NBR blends show an intermediate
initial linear portion of sorption curves, i.e. before the attain- behavior in between the two limiting models, parallel and
ment of 50% of equilibrium uptake. The variation of diffu- series. The experimental data are close to the Maxwell
sion and permeation coefficients with blend composition for model with PP phase continuous at 70 wt% PP, and at
PP/NBR blends is shown in Fig. 5. It is seen from the figure 30 wt% PP the data is close to the Maxwell model with
that PP shows the lowest value for diffusion and permeation the NBR phase continuous. It is seen from the figure that
coefficient and NBR shows the highest. The blends take the experimental curve coincides with Robeson’s model for
intermediate values. As the concentration of NBR increases, a co-continuous morphology at Xa ˆ 0.6 and this suggests
the D and P values increase. The increase in D and P values that a phase inversion occurs for NBR at this volume
584 S. George et al. / Polymer 41 (2000) 579–594

Fig. 4. Schematic representation of the tortuous path exhibited by the crystalline PP phase to the transport of solvent.

fraction. This result is consistent with our early observation leads to the formation of S–S linkages while DCP vulcani-
from SEM micrographs that phase inversion occurs between zation gives rise to rigid C–C linkages. In mixed systems,
50 and 70 wt% of NBR in the blend. both types of crosslinks are present (Fig. 8). Since in the
DCP system the diffusion of solvents is more difficult due to
3.2. Effect of type of crosslinking the rigid nature of C–C linkages, it shows the lowest uptake
and low diffusion coefficient values. The presence of more
The physical and mechanical properties of polymers are
flexible S–S linkages permits the solvents to permeate more
found to depend on the nature of crosslinking viz., sulfur or
easily through the sulfur vulcanized samples and hence they
peroxide curing system. In this paper we have investigated
show the highest uptake. In mixed vulcanized system, an
the effect of three different crosslinking systems, sulfur,
intermediate behavior is expected, since it contains both
peroxide and a mixed system comprising sulfur and perox-
C–C and S–S linkages. But the results indicate that the
ide, on diffusion. Fig. 7 shows the Qt vs. t 1/2 curves of
mixed system shows almost the same behavior as that
dynamically vulcanized P50 blends. Among the different
of the sulfur system. This can be explained using cross-
vulcanizing systems used, the DCP crosslinked system
link density and molar mass between the crosslinks for
shows the lowest uptake. Sulfur cured system shows the
the different vulcanizing systems.
highest solvent uptake and the mixed system takes an inter-
The molar mass between crosslinks (Mc) of the network
mediate position. The values of D, P, n and k for these
polymer chain was calculated using the relation developed
samples are given in Table 4. It is seen from the table that
by Flory–Rehner [36].
the values of n vary from 0.46 to 0.052 for the three systems,
i.e. sulfur system shows the lowest n value while the DCP 2rp Vs …f1=3 2 f=2†
system shows the highest. All the three systems follow an Mc ˆ …11†
ln…1 2 f† 1 f 1 xf2
almost Fickian mode of transport. When we examine the k
values it is seen that the DCP crosslinked systems show the where r p is the density of the polymer, Vs, the molar volume
lowest k value and sulfur and mixed crosslinked systems of solvent and f , the volume fraction of swollen rubber in
show almost similar values which is higher than that of the fully swollen sample. The volume fraction of the swol-
the DCP system. Hence the interaction between solvent len rubber is estimated by considering PP as a filler in NBR
and polymer blend is low for DCP crosslinks compared to since the uptake of PP is low, which is given by
sulfur and mixed systems. The diffusion and permeation
…d 2 fw†r21
p
coefficients of toluene are low in the case of DCP compared fˆ …12†
to sulfur and mixed systems. The difference in the transport …d 2 fw†r21 21
p 1 A s rs
parameters of these three different vulcanizing systems where d is the deswollen weight of the polymer, f the
arises from the type of crosslinks formed and also due to volume fraction of filler; w the initial weight of sample; r p
the extent of crosslinking, i.e. the crosslink density. Among the density of the polymer, r s the density of the solvent and
the three vulcanizing systems used, sulfur vulcanization As, the weight of the solvent in the swollen sample.
Table 3 Vs …ds 2 dp †2
n and k values for diffusion of toluene through PP/NBR blends xˆb1 …13†
RT
n
Sample n k (g/g min )
where b is the lattice constant ( ˆ 0.34), d the solubility
P100 0.46 0.01 parameter, and Vs the molar volume of the solvent.
P70S 0.47 0.04 The values of Mc and the crosslink density are given in
P50S 0.47 0.05 Table 5. It is seen that the Mc values are in the order DCP ,
P30S 0.51 0.05
mixed ù sulfur system, i.e. the molar mass between the
P0S 0.60 0.03
crosslinks is more in the case of mixed vulcanized system
 S. George et al. / Polymer 41 (2000) 579–594 585

Fig. 5. Variation of diffusion and permeation coefficient with weight percentage of NBR.

and lowest for the DCP system. Since Mc is inversely permeates less solvent molecules compared to sulfur and
proportional to crosslink density, the crosslink density mixed vulcanized systems. Since blends vulcanized with
follows the order DCP . mixed ù sulfur, i.e. the DCP sulfur and mixed system have the same crosslink density,
vulcanized system with the highest crosslink density the uptake may be comparable. Hence the combined effect

Fig. 6. Experimental and theoretical version of permeation coefficient with volume fraction of NBR is the blends.
586 S. George et al. / Polymer 41 (2000) 579–594

Fig. 7. Variation of mol% uptake of toluene (Qt) with square root of time for various dynamically crosslinked P50 blends.

of the crosslink density and rigidity of bonds accounts for

the transport behavior in sulfur and mixed systems.

3.3. Effect of penetrant size

The sorption behavior of organic solvents through poly-

mer samples is affected by the size, shape and polarity of
penetrant molecules. Fig. 9 shows the sorption curves of P50
blend crosslinked with sulfur for benzene, toluene and
xylene. As expected, as the size of penetrant molecules
increases the solvent uptake decreases, i.e. the low molecu-
lar weight solvent benzene shows the highest uptake and
xylene, the high molecular weight solvent shows the lowest
uptake. The variation of Q∞ with molar volume of solvent
for different blend compositions is shown in Fig. 10. It is
seen that the Q∞ values decrease linearly with the molar
volume of penetrant for all the blend compositions. The
rate of decrease in Q∞ with molar volume increases with
increase in concentration of NBR.
The influence of penetrant size on the mechanism of
transport, diffusion and permeation coefficients of various

Table 4
Transport parameters for diffusion of toluene through dynamically vulca-
nized P50 blends

Sample n k (g/g min n) D ( × 10 5 cm 2/s) P ( × 10 5 cm 2/s)

P50S 0.46 0.05 2.03 1.64

P50C 0.52 0.03 1.81 1.43
P50M 0.47 0.05 2.22 1.78 Fig. 8. Schematic representation of the types of crosslinks formed in (a)
sulfur; (b) peroxide and; (c) mixed cured P50 blends.
 S. George et al. / Polymer 41 (2000) 579–594 587

Table 5 unfilled blend the n values increase for silica and C-black
Mc and crosslink density values of dynamically vulcanized P50 blends loaded blends, while in cork filled sample n decreases. The k
Sample Mc Crosslink density values increase upon the incorporation of cork while in the
( × 10 24 g mol/cm 3) other two cases, the k value shows a decrease. While con-
sidering the diffusion and permeation coefficients, it is seen
P50S 1673.02 5.97
that the values are higher than the unfilled blend in the case
P50C 1252.08 7.99
P50M 1671.78 5.98 of C-black and silica-filled samples and lower for the cork-
filled blend. The difference in the behavior of the filled
blends on the diffusivity and permeability and in the equi-
polymer blends was analyzed and the results are given in librium values may be related to the interaction between the
Table 6. The n values decrease with increase in molecular filler and polymer. It has been reported that in thermoplastic
weight of the penetrants. In the case of k values, there is no elastomers, the filler particles generally concentrate in the
regular trend. However, in general, toluene shows the high- rubber phase. Since silica filler surface is polar, it shows
est k value and hence more interaction with the polymer. better interaction with the polar NBR particles which may
Both D and P values, decrease with the penetrant size for the lead to more phase separation, i.e. interaction between PP
pure NBR and blends. and NBR decreases, which leads to weak interphases. This
weak interphase may enhance the diffusion and permeation
3.4. Effect of fillers of solvents through the membrane which increases the D
and P values. However, the final equilibrium values depend
The incorporation of filler into polymer networks is found mainly on the extent of crosslink density and reinforcement
to decrease the sorption and diffusion of solvents. Fig. 11 exerted by the filler.
depicts the sorption curves of unfilled and 30 phr loaded The extent of reinforcement is assessed by using the
silica, carbon black and cork filled P50 blends. It is seen Kraus equation [37], according to which
from the figure that the initial uptake of toluene is not
much affected by the presence of fillers. However, the equi- Vr0 =Vrf ˆ 1 2 m‰f =1 2 f Š …14†
librium uptake values are lower for the filled blends
compared with unfilled ones. Among the three fillers used, where Vrf is the volume fraction of swollen rubber in the
the C-black filled blends showed the lowest equilibrium fully swollen filled sample, f is the volume fraction of filler
uptake. The uptake behavior of different blends follows and m is a measure of the extent of reinforcement.
the order P50 . P50K30 . P50Tsi30 . P50C30. Table 7 shows The crosslink density and extent of reinforcement (m) of
the n, k, D and P values of these blends. Compared with the different fillers obtained from the Kraus equation for various

Fig. 9. Variation of mol% uptake (Qt) with square root of time for the diffusion of benzene, toluene and xylene through P50 blends.
588 S. George et al. / Polymer 41 (2000) 579–594

Fig. 10. Variation of equilibrium uptake (Q∞) of toluene in P50 blends with molar volume of solvents.

filled and unfilled blends are shown in Table 8. It is seen that with the highest crosslink density and intermediate rein-
the crosslink density decreases in filled blends as compared forcement shows the lowest uptake.
with unfilled P50 blend. Hence, the reduction in equilibrium
uptake in filled blends is due to the reinforcement exerted by 3.5. Effect of temperature
the fillers. Among the filled blends, the crosslink density
varies in the order P50C30 . P50Tsi30 . P50K30 and the extent The effect of temperature on mol% uptake of P50 blends is
of reinforcement follows the order: P50Tsi30 . shown in Fig. 12. As the temperature increases, the uptake
P50C30 . P50K30. Hence cork-filled P50 blend with the lowest of solvent increases as expected. The rates of diffusion and
crosslink density and reinforcement shows the highest permeation also increase with increase in temperature. The
uptake among the filled samples. The C-black filled sample temperature dependence of transport properties can be used
to evaluate the activation energy for the diffusion and
permeation processes using the Arrhenius relation
Table 6
X ˆ X0 exp 2 …Ea =RT† …15†
Transport parameters for diffusion of various solvents (B ˆ benzene, T ˆ
toluene, X ˆ xylene) through PP/NBR blends where X is P or D; Ea, the activation energy, R, the universal
n k (g/g min )n
D ( × 10 cm /s)
5 2
P( × 10 cm /s)
5 2 gas constant; and T, the absolute temperature. The Arrhe-
nius plots of log D and log P vs. 1/T for different blend
P100 B 0.05 0.01 0.07 0.01 compositions are shown in Figs. 13 and 14, respectively.
T 0.46 0.01 0.22 0.02
From the Arrhenius plots, the values of activation energy
X 0.53 0.01 0.76 0.01
P70S B 0.47 0.04 0.01 0.01 for permeation and diffusion were calculated using regres-
T 0.47 0.04 0.99 0.45 sion analysis. The results are shown in Tables 9 and 10. In
X 0.42 0.04 0.58 0.23 different blend compositions, the activation energy for diffu-
P50S B 0.50 0.04 2.26 2.11 sion decreases with increase in rubber concentration, while
T 0.47 0.05 2.03 1.65
activation energy for permeation did not show a regular
X 0.48 0.05 1.44 0.98
P30S B 0.51 0.05 3.54 5.53 trend. Among the different vulcanizing system used, the
T 0.51 0.05 3.50 4.79 DCP vulcanized system shows the highest activation energy
X 0.48 0.05 2.50 2.66 and mixed system shows the lowest value. The highest
P0S B 0.63 0.03 5.53 11.65 activation energy observed in DCP crosslinked system
T 0.60 0.03 4.45 11.28
may arise from the rigidity of C–C linkages compared
X 0.58 0.03 3.73 7.08
with S–S linkages present in sulfur and mixed systems. It
 S. George et al. / Polymer 41 (2000) 579–594 589

Fig. 11. Variation of mol% uptake of toluene with square root of time for various filled P50 blends.

is also interesting to note that the activation energy the sorption proceeds through creation of new sites or pores
decreases upon the incorporation of filler. The toluene in the polymer. In the case of blends with different NBR
shows the highest activation energy among the solvents compositions, the DH value decreases with increase in NBR
used. content.

3.6. Thermodynamic parameters 3.7. Kinetics of diffusion

The thermodynamics parameters for diffusion, DH and The first-order kinetic model has been used to follow the
DS can be calculated using van’t Hoff’s relation: kinetics of sorption and diffusion of solvents through PP/
NBR blends. In order to apply this kinetic model it is
DS8 DH8s assumed that during sorption of solvents, structural changes
logKs ˆ 2 …16†
2:303R 2:303RT may occur in polymer chains which require a rearrangement
where Ks is equilibrium sorption constant which is given by of the polymer segments that can dominate the kinetic beha-
vior. According to the first-order kinetic equation,
No: of mols of solvent sorbed at equilibrium
Ks ˆ : …17† dc=dt ˆ k 0 …C∞ 2 Ct † …18†
Mass of the polymer
The values of DS and DH are obtained by regression where k 0 is the first-order rate constant, and Ct and C∞
analysis of the plots of log Ks vs. 1/T. The values of DH8 represent the concentrations at time t and at equilibrium,
and DS8 are given in Table 11. It is seen from the table that respectively. Integration of the above equation gives
the DH values are positive and vary from 2.2 to 4.6 kJ/mol. k 0 t ˆ 2:303log‰C∞ =…C∞ 2 Ct †Š …19†
The positive values of DH indicate that the sorption is an
endothermic process and is dominated by Henry’s mode, i.e. Figs. 15 and 16 show plots of log…C∞ 2 Ct † vs. t for various
blend compositions in toluene. From these plots, the first-
Table 7 order rate constants were calculated by regression analysis
Transport parameters for diffusion of toluene through filled P50 blends and the results are given in Table 12. As expected, the rate
n k (g/g min n) D ( × 10 5 cm 2/s) P ( × 10 5 cm 2/s) constant values increase with an increase in concentration of
rubber in the blends. Among the vulcanized systems, the
T-silica 0.56 0.03 2.36 1.78 DCP vulcanized system shows the lowest k 0 value, while
C-black 0.51 0.04 2.36 1.64
the mixed system shows the highest value. The sulfur vulca-
Cork 0.44 0.06 1.93 1.50
nized blend shows an intermediate behavior. In filled
590 S. George et al. / Polymer 41 (2000) 579–594

Table 8 Table 13
Crosslink density and extent of reinforcement of filled P50 blends Theoretical and experimental Mc values for dynamically vulcanized PP/
NBR blends
Crosslink density m
( × 10 24 g mol/cm 3) Mc

Unfilled 2.28 Experimental Affine Phantom

T-silica 1.87 2 1.71
C-black 2.11 2 0.53 P70S 1196.32 520.40 173.46
Cork 1.35 0.82 P50S 1673.02 1283.94 427.98
P50P 1252.02 978.94 326.31
P50M 1671.78 1282.99 427.66
Table 9 P30S 2992.19 2261.16 753.64
Activation energy for diffusion and permeation of toluene in PP/NBR
blends
systems, the rate constant values are higher as compared
EP (kJ/mol) ED (kJ/mol)
with unfilled systems.
P70S 21.54 26.15
P50S 20.31 23.66
P30S 18.03 29.86 3.8. Comparison with theory
P50P 23.48 26.31
P50M 12.27 17.04
P50Tsi30 19.15 21.29 The chemical crosslink density calculated was correlated
with those of affine and phantom network models [21]. In
the affine network model, it is assumed that the junction
Table 10
Activation energy for diffusion and permeation of various solvents (B ˆ
points are embedded in the network without fluctuations,
benzene, T ˆ toluene, X ˆ xylene) in PP/NBR blends so that the components of each chain vector transform line-
arly with macroscopic deformation.
ED (kJ/mol) EP (kJ/mol) According to the affine network model, the molecular
P50S B 17.95 21.48 weight between crosslinks is given by
T 20.31 23.66  
X 18.40 21.36 m 1=3
rVs v2=3 v 1=3
1 2 v
P50Tsi30 B 16.90 20.57 2c 2m
n 2m
T 19.15 21.29 Mc …aff† ˆ …20†
2 …ln…1 2 v2m † 1 v2m 1 xv22m †
X 16.04 18.20

where Vs is the molar volume of solvent, m , the number of

Table 11 effective chains, n , the number of junctions, v2m, the poly-
Thermodynamic parameters for diffusion of toluene through PP/NBR mer volume fraction at equilibrium swelling, v2c, the poly-
blends
mer volume fraction during crosslinking, and r , the polymer
Sample DH8 (kJ/mol) DS8 (J/mol/deg) 2 DG density.
In the case of the phantom network model, the chains can
P70S 4.60 28.78 4.02
move freely through one another, i.e. the junction points
P50S 3.16 28.86 5.49
P30S 2.77 25.90 4.99 fluctuate over time around their mean position without
P50M 4.70 24.00 2.49 any hindrance from the neighboring molecules. In the phan-
P50P 3.71 27.32 4.48 tom network model, Mc is calculated using the relation
P50Tsi30 2.13 32.94 4.48

2
rVs v2=3
12 1=3
2c v2m
Table 12 f
Mc …ph† ˆ …21†
Rate constant values for the transport of toluene in PP/NBR blends 2 …ln…1 2 v2m † 1 v2m 1 xv22m †
Sample Rate constant k ( × 10 3 min 21)
where f is the junction functionality.
P100 0.22 The results obtained from these calculations are shown in
P70S 1.04 Table 13. From the table it is seen that the experimental
P50S 1.95
values are close to the affine network model. Hence in PP/
P30S 2.80
P0S 2.66 NBR blends, the crosslinked junctions are embedded in the
P50P 1.55 network, so that they cannot fluctuate freely and the chain
P50M 2.03 vector transforms linearly with macroscopic deformation.
P50Tsi30 2.25 Attempts have been made to compare the experimental
P50K30 2.19
diffusion curves with the theoretical diffusion profile. The
P50C30 2.14
theoretical curves are constructed using the equation which
 S. George et al. / Polymer 41 (2000) 579–594 591

Fig. 12. Variation of mol% uptake (Qt) of toluene with square root of time at various temperatures.

Fig. 13. Arrhenius plots of log D vs. 1/T of dynamically vulcanized P50 blends.
592 S. George et al. / Polymer 41 (2000) 579–594

Fig. 14. Arrhenius plots of log P vs. 1/T of dynamically vulcanized P50 blends.

describes the Fickian diffusion model: In order to generate the diffusion curves, the experimen-
Qt 8 X∞ tally determined D values are substituted in Eq. (22). Fig. 17
ˆ12 2 1=…2n 1 1†2 exp‰2D…2n 1 1†2 p2 t=h2 Š: shows the theoretical and experimental sorption curves of
Q∞ p nˆ0
P70, P50 and P30 blends in toluene. The overall agreement
(22) between the experimental and theoretical curves is fairly

Fig. 15. Variation of log…C∞ 2 Ct † vs. time of PP/NBR blends.

 S. George et al. / Polymer 41 (2000) 579–594 593

Fig. 16. Variation of log…C∞ 2 Ct † vs. time of filled P50 blends.

Fig. 17. Experimental and theoretical diffusion curves of P70, P50 and P30 blends.
594 S. George et al. / Polymer 41 (2000) 579–594

good. A similar behavior was observed for other blend [2] Walker BM, editor. Handbook of thermoplastic elastomers New
systems and solvents as well. York, NY: Van Nonstrand Reinhold, 1979.
[3] Yoon LK, Choi CH, Kim BY. J Appl Polym Sci 1995;56:239.
[4] Campbell DS, Elliot DJ, Wheelans MA. NR Technol 1978;9:21.
4. Conclusions [5] Kuriakose B, De SK. Mater Chem Phys 1985;12:157.
[6] Coran AY, Patel R, Williams D. Rubber Chem Technol 1982;55:116.
[7] Koros WJ, editor. Barrier polymers and structures ACS Symposium
The transport behavior of various aromatic solvents Series, 423. Washington, DC: American Chemical Society, 1990.
through PP/NBR blends was investigated. The transport [8] Paul DR, editor. Polymer blends, II. New York, NY: Academic Press,
phenomena follows a Fickian trend in PP/NBR blends. As 1976 Chapter 12.
the concentration of NBR increases in the blend, the equili- [9] Thomas S, Prud’homme RE. Polymer 1992;33:4260.
brium uptake values and diffusion and permeation coeffi- [10] Miettinen RMH, Seppala JV, Ikkala OT, Reima IT. Polym Engng Sci
1994;5:395.
cients are increased. The variation in transport parameters [11] Lia FS, Su AC, Hsu T-C. Polymer 1994;35:2579.
was correlated with the morphology of the system. The [12] Varughese KT, Nando GB, De PP, De SK. J Mater Sci 1988;23:3894.
equilibrium uptake values show a sharp increase after [13] Liu NC, Xie HQ, Baker WE. Polymer 1993;34:4680.
50 wt% NBR which is due to the phase inversion of NBR [14] George J, Joseph R, Varughese KT, Thomas S. J Appl Polym Sci
from dispersed to continuous phase. The experimental 1995;57:449.
[15] Coran AY, Patel R. Rubber Chem Technol 1983;56:1045.
permeation coefficients were compared with various theore- [16] George S, Joseph R, Varughese KT, Thomas S. Polymer
tical predictions. Maxwell and Robeson’s models have been 1995;36(23):4405.
used to predict the experimental permeation coefficients. [17] George S, Neelakantan NR, Varughese KT, Thomas S. J Polym Sci B:
The effect of type of crosslinking on transport behavior Polym Phys 1997;35:2309.
was studied. Among the three vulcanizing systems used [18] George S, Ramamurthy K, Anand JS, Varughese KT, Thomas S.
Submitted to Polymer for publication.
(sulfur, DCP and mixed (S 1 DCP) systems), the sulfur
[19] Shivaputrappa B, Harogappad A, Aminabhavi TM. Macromolecules
system showed the highest uptake and DCP the lowest 1991;24:2598.
and mixed system showed intermediate behavior. The [20] Schneider NS, Illinger JL, Cleaves MA. Polym Engng Sci
difference in the behavior was correlated with type of cross- 1986;26(22):1547.
links formed, i.e. S–S, C–C and C–C 1 S–S linkages and [21] Liao DC, Chern YC, Han JL, Hseih KH. J Polym Sci B: Polym Phys
crosslinking density. The diffusion and permeation coeffi- 1997;35:1747.
[22] Unnikrishnan G, Thomas S. J Polym Sci B: Polym Phys 1997;35:725.
cients and Q∞ values decreased with increase in molar [23] Mathew AP, Pakirisamy S, Kumaran MG, Thomas S. Polymer
volume of solvents. The effect of different types of filler 1995;36(26):4935.
on transport properties was also investigated. In the filled [24] Johnson T, Thomas S. J Macromol Sci Phys 1997;B36(3):401.
blends, the uptake values follows the order cork . silica . [25] Mathai AE, Thomas S. J Macromol Sci Phys 1996;B35(2):229.
carbon black. The activation energy for diffusion and [26] George SC, Thomas S, Ninan KN. Polymer 1996;37(26):5839.
[27] Aminabhavi TM, Aithan US, Shukla SS. JMS Rev Macromol Chem
permeation was calculated. The activation energy for diffu- Phys 1989;C29(2/3):319.
sion decreases with an increase in rubber concentration. [28] Lawandy SN, Botros SH. J Appl Polym Sci 1991;42:137.
Comparison of the crosslink density values with phantom [29] Horkay F, Zrinyi M, Geissler E, Hecht AM, Pruvost P. Polymer
and affine network models indicates that in PP/NBR blends, 1991;32:835.
the network structure could be modeled by the affine theory. [30] Unnikrishnan G, Thomas S, Varghese S. Polymer 1996;37(13):2687.
[31] Hopfenberg HB, Paul DR. In: Paul DR, editor. Polymer blends, I.
The experimental and theoretical diffusion curves are in
New York, NY: Academic Press, 1976.
good agreement for the blend systems. [32] Kolarik J, Gueskens G. Polym Networks Blends 1997;7(1):13.
[33] Aminabhavi TM, Phayde HTS. J Appl Polym Sci 1995;55:1335.
[34] Asaletha R, Kumaran MG, Thomas S. Submitted to Polymer for
Acknowledgements publication.
[35] Robeson LM, Noshay A, Matzner M, Merriam CN. Die Angew
One of the authors (S.G.) is thankful to the Council of Makromol Chem 1973;29/30:47.
Scientific and Industrial Research, New Delhi, for the finan- [36] Flory PJ. Principles of polymer chemistry, Ithaca, NY: Cornell
cial assistance for this work. University Press, 1953.
[37] Mathew G, Singh RP, Lakshminarayan R, Thomas S. J Appl Polym
Sci 1996;61:2035.
References

[1] De SK, Bhowmick AK, editors. Thermoplastic elastomers from

rubber–plastic blends New York, NY: Ellis Horwood, 1990.