Full Paper

Diffusion Crosslinking Kinetics of Thermoset

Polymers
Stanislav A. Patlazhan,* Alexander A. Berlin, Michel Bouquey,
Christophe A. Serra, Rene Muller

Relative crosslink concentration

The problem of the delayed crosslinking arising due to the diffusion of a crosslinker molecule
from a vector particle within a thermoset polymer matrix is discussed. The crosslinker
molecules, initially embedded in the vector particles, diffuse from the particle to the polymer
matrix and react with the functional groups grafted to polymer chains. The two-step diffusion
crosslink kinetics is studied theoretically on the example of polyethylene functionalized with
0.6
anhydride groups reacting with 1,4-butanediol as a crosslinker molecule. This provides the independent of time
t =1000
limiting distribution functions of the concentration of
0.4
crosslinks, unreacted and intermediate functional
40
groups, which are shown to be rather sensitive to system
10
parameters as rates of chemical reactions, ratio of diffu0.2
sion coefficients of matrix to vector particles, stoichio4
metric composition of the reactive groups, etc. The
2
nonmonotonic distribution of the crosslink concen0.5
0.0
tration around the vector particles at certain range of
0.0
1.0
2.0
3.0
4.0
5.0
the parameters is revealed.
Relative distance from vector particle

Chemical crosslinking is one of the most important

methods that allow extending the range of applications
of commercial polymers. However, the mechanical processing of thermoset polymers is usually hampered by the fast
growing viscosity of the material due to the chain

S. A. Patlazhan, A. A. Berlin
Semenov Institute of Chemical Physics of Russian Academy of
Sciences, 4, Kosygin Street, Moscow, 119991 Russia
Fax: (7)-499-137-67-11; E-mail: [email protected]
M. Bouquey, C. A. Serra, R. Muller
Groupe dIntensification et dIntrapolation des Procedes
Polyme`res (G2IP), Laboratoire dIngenierie des Polyme`res pour les
Hautes Technologies, Ecole de Chimie Polyme`res et Materiaux,
Universite de Strasbourg, 25 rue Becquerel, Strasbourg Cedex 2,
F-67087, France
Macromol. Theory Simul. 2011, 20, 000000
2011 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim

crosslinking. Various methods have been proposed to

overcome this difficulty. For instance, prepreg technology
is always used in the preparation of layered engineering
materials preliminary impregnated by a thermosetting
binder.[1] In this case the molecules of crosslinker diffuse
into the conjugated layers resulting in hardening of the
product. However, this method is unsuitable for the
synthesis of homogeneous and isotropic materials.
In a previous paper new polymeric materials that are able
to crosslink after the melt processing stage were
described.[2] To prevent an early reaction in the melt, vector
particles were used, whose function was to temporarily
trap the crosslinker molecules and to release them mainly
after processing. With this method there were achieved: (i) a
limited crosslinking reaction at the end of the melt
processing stage, (ii) the continuation of the crosslinking
reaction at moderate temperature after melt processing,
(iii) the possibility to control the uniformity of crosslinking

wileyonlinelibrary.com

DOI: 10.1002/mats.201000102

Early View Publication; these are NOT the final page numbers, use DOI for citation !!

Introduction

S. A. Patlazhan, A. A. Berlin, M. Bouquey, C. A. Serra, R. Muller

www.mts-journal.de

density, even in thick parts, due to the presence of vector

particles distributed in the material, and (iv) the possibility
to work with high viscous thermoset polymers. Nevertheless, it appears that the final properties of crosslinked
polymers according to this process depend simultaneously
on chemical parameters (nature of the crosslinker, reaction
catalyst, etc.) and process parameters (curing temperature,
nature, and weight concentration of the vector particles,
quality of their dispersion in the polymer matrix, etc.).
In order to optimize the final composition of the reactive
system and eventually controlling the microstructure of the
material, we present in this paper a numerical simulation of
this post-crosslinking process by taking into account the
effect of the competitive diffusion and reaction kinetics of
the crosslinker on the uniformity of the crosslinking density
around vector particles.
Description of the Chemical System
In ref.[2] the particular case of anhydride-functionalized
polyolefins where the formation of crosslinks between
polymer chains follows a two-step mechanism involving
butanediol as the crosslinker molecule was focused on.
Initially the butanediol molecules are trapped in Orgasol
polyamide-12 particles playing the role of vector particles.
Then they diffuse within the polymer matrix where they
react with the polymer functional groups.
The general scheme of a crosslink formation resulting in
the two chemical reaction of 1,4-butanediol with two
succinic anhydride (P) bearing by two different polymer
chains is presented in Figure 1a. In the first reaction, a
butanediol (C) is grafted on the polymer backbone through
the reaction of one hydroxyl function with the anhydride
group (P) which are carried by the polymer chain. This
reaction results in opening of the anhydride ring by an
alcohol function leading to formation of two intermediate
species. The first one is a carboxylic acid function (I) is
weakly reactive with other alcohol functions.[35] The
second one is an ester function bearing the second hydroxyl
group (see Figure 1b). In the second step, the intermediate
species I react with another anhydride function, which are
carried by another polymer backbone (Figure 1a), and form
a crosslinking point (B).
The model developed in the present paper takes into
account the kinetics of the two above reactions, as well as
the diffusion of butanediol both in the Orgasol vector
particles and in the polymer matrix. It appears that the final
properties of crosslinked polymers according to this process
will depend on the spatial distribution of crosslink density
around the particles. The problem of predicting this
distribution is considered here by taking into account
simultaneous chemical reaction and diffusion processes. As
will be shown, the final distribution around a vector particle
is not necessarily uniform, which may lead to microstruc-

Figure 1. Crosslinking of the functionalized polyethylene: (a) general scheme and (b) reaction of succinic anhydride (P) with
1,4-butanediol followed by formation of intermediate hydroxyl
groups (I).

turing of the material and potentially to new original

properties.
According to this process, the parameters that can be
adjusted are the nature and the concentration of the
reactants (C and P) as well as the volume fraction and
therefore the average distance between the neighboring
vector particles. The temperature of the process will affect
both the chemical reaction and diffusion kinetics, although
the two processes may be characterized by different
activation energies.
Theoretical Model
We consider a matrix of functionalized polymer filled with a
small amount of vector particles of volume fraction f. Each
particle contains the crosslinker molecules of an initial
concentration C0. Butanediol diffuses from the particles
into the polymer matrix and reacts with the functional
anhydride groups of the polymer chains presenting in the
polymer matrix at an initial concentration P0 . A spherical
composite cell of radius R containing a particle of radius a in
its center can be considered as a representative volume
element of a system (see Figure 2). In this case the volume
fraction of vector particles is equal to f a=R3
This approach implicitly assumes either isotropic or
regular periodic distribution of vector particles in the
polymer matrix. The continuity condition between two
adjacent cells then requires that the flux of crosslinker
molecules be equal to zero at the outer radius of the
representative volume element. The model could of course

Macromol. Theory Simul. 2011, 20, 000000

2011 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim

www.MaterialsViews.com

Early View Publication; these are NOT the final page numbers, use DOI for citation !!

Diffusion Crosslinking Kinetics . . .

www.mts-journal.de

of the chemical reagents C, P and their products I, B.

Assuming spherical symmetry, the concentration of the
crosslinker agent inside the vector particle is described by
the following diffusion equation:

 2

@C
@ C
2 @C
Din

@t
@r2
r @r

Figure 2. Representative volume element of the system.

be improved by taking into account a random distribution

of vector particles at the cost of a more time-consuming 3D
simulation on a larger representative cell which should
contain several vector particles. This approach is outside the
scope of the present paper and it will be shown that the
simplified 1D model already leads to a satisfactory
description of the reaction/diffusion process.
The initial ratio of hydroxyl to anhydride groups
is characterized by the stoichiometric coefficient
j fC0 f=P0 1f where f 2 is the butanediol functionality. The numbers of hydroxyl and anhydride reactive
groups initially present in the system coincide when j 1,
which corresponds to the stoichiometric composition.
Butanediol molecules diffuse through the vector particles
and polymer matrix with the diffusion coefficients Din and
Dout, respectively. Their ratio d Din =Dout dependents on
the nature of the particles and matrix. On the other hand,
due to the very slow diffusion of polymer chains compared
to low molecular weight species, we assume that the
diffusion of P, I, and B functions are negligible.
Once in the polymer matrix, the crosslinker molecules C
react with anhydride groups P forming intermediate
species I linked to the polymeric chains. Owing to the
bifunctionality of butanediol molecules, the intermediate
groups keep chemical activity and could react with
anhydride groups belonging to other macromolecules. As
a result of such reactions, covalent crosslinks B are created.
Intramolecular crosslinking can be neglected at small
fractions of anhydride groups on the polymer chains. The
kCP
rate constant kCP of the first chemical reaction C P ! I
exceeds the rate constant kIP of the second reaction
kIP
I P ! B at least twice because in contrast to butanediol
the intermediate species I has only one reactive hydroxyl
group. Moreover the mobility of free butanediol molecules
is much higher than that of intermediate groups grafted to
polymer chains. Therefore the ratio n kIP =kCP falls in the
range between 0 and 0.5.[6,7]
This allows us to write out the kinetic equations
describing the time evolution and the spatial distribution

The rate of change of crosslinker concentration in the

polymer matrix is governed both by a diffusion process and
by the chemical reactions. The first chemical reaction of
scheme presented in Figure 1 reduces the concentration of
crosslinker due to chemical reaction with the polymer
functional groups. As a result and still assuming spherical
symmetry, the concentration of crosslinker obeys the
following kinetic equation:
 2

@C
@ C 2 @C
Dout
kCP CP

@t
@r2 r @r

(2a)

Evolution of concentration of the intermediate species I is

defined by two opposite processes: (i) its birth due to the
kCP
first reaction C P ! I of the scheme and (ii) its
kIP
disappearance caused by the second reaction I P ! B.
This is governed by the following equation assuming that
the crosslinking does not affect significantly the matrixs
volume and 1st order reactions:
@I
kCP CPkIP IP
@t

(2b)

On the other hand, concentration of the polymer

functional groups P steadily decreases according to
equation
@P
kCP CPkIP IP;
@t

(2c)

while that of crosslinks B steadily increases as

@B
kIP IP:
@t

(2d)

The following initial conditions

C t 0 C0 in the vector particle;

(3a)

Ct 0 0; I t 0 0; Pt 0
P0 ; Bt 0
0 in polymer matrix;

(3b)

reflect a lack of crosslinker, intermediate groups, and

crosslinks in polymer matrix as well as homogeneous

Macromol. Theory Simul. 2011, 20, 000000

2011 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim

Early View Publication; these are NOT the final page numbers, use DOI for citation !!

www.MaterialsViews.com

(1)

S. A. Patlazhan, A. A. Berlin, M. Bouquey, C. A. Serra, R. Muller

www.mts-journal.de

distribution of crosslinker and anhydride groups in the

vector particle and matrix, respectively, in the initial time.
The boundary conditions take into consideration disappearance of crosslinker flux at the center of the vector
particle and at the outer boundary of the representative
volume element as well as the dynamic balance at the
interface between the particle and the matrix:




@C
@C
@C
@C

0; Din 
Dout 
@r r 0
@r r R
@r r a0
@r r a0

(4)

The above formulated boundary value problem (1)(4)

can be expressed in dimensionless form. For this purpose,
we define the average particle radius a as a length unit,
t a2 =Dout as a time unit, and the initial concentration of
the polymer functional groups P0 as a concentration unit.
As a consequence, the dimensionless radial coordinate,
time and concentrations are respectively equal to r~
r=a; t~ t=t; C~ C=P0 ; I~ I=P0 ; P~ P=P0 and B~ B=P0 ,.
Then the diffusion Equation (1) inside the particle can be
rewritten as
 2~

@C~
@ C
2 @C~
;

d
r~ @r~
@t~
@r~2

(5)

while the kinetic Equations (2) in the matrix become

~
@C~
@2 C~
2 @C~
~ @I K C~PKn
~
~
 K C~P;
2
I~P;
~
~
~
r @r
@t
@t~
@r~
@P~
@B~
~
~
~
K C~PKn
I~P;
KnI~P:
~
@t
@t~

(6)

An effective rate constant K kCP P0 a2 =Dout is introduced

in Equation (6) as a combination of the reaction rate
constant kCP , the diffusion coefficient Dout of the crosslinker
in the polymer matrix, the initial concentration of polymer
functional groups P0 , and the particle average radius a. The
initial conditions (3) are transformed to


1f
C~ t~ 0 j
in the particle;
ff






C~ t~ 0 0; I~ t~ 0 0; P~ t~ 0


1; B~ t~ 0 0 in the polymer matrix;

(7a)

(7b)

while the boundary conditions (4) are represented as





@C~
@C~
@C~
@C~

0;
d

:
@r~r~ 0
@r~r~ f1=3
@r~r~ 10
@r~r~ 1 0

(8)

Thus the spatial distribution of all chemical reagents and

products are determined by the five dimensionless parameters: K, f, j, d, and n.

Results and Discussion

The solution of the boundary value problem (5)(8) was
obtained numerically for the following data: the reaction
rate constant of butanediol with anhydride was taken to be
equal to kCP 106 m3 mol1 s1 ; the diffusion coefficient of
butanediol in the polyolefin matrix was taken as
Dout 1016 m2 s1 ; the initial concentration of anhydride
groups in the polymer matrix was taken as
P0 15 mol  m3 ; the average radius of vector particles
was taken as a 2 mm and their volume fraction as
f 0:0078. This corresponds to the dimensionless reaction
rate constant K 0.6.
To discuss an influence of system parameters controlling
final distribution of crosslink density, a reference system
was determined by the following characteristics: j 1,
d 1, K 0:6, and n 0:5. Distributions of dimensionless
~ I,
~ and B~ in the representative volume
~ P,
concentrations C,
element calculated for this case at different times are
presented in Figure 3. It is seen that concentration C~ of the
free crosslinker molecules quickly decreases: it does not
exceed 3  103 already at the dimensionless time t~ 40
(Figure 3a). This process is accompanied by the consumption of anhydride groups on the polymeric chains
(Figure 3b). In a first step, the decrease of concentration P~
is mainly due to the formation of the chemically active
intermediate species following the reaction of butanediol
with anhydride: the concentration of I~ rapidly increases as
soon as the crosslinker starts to diffuse into polymer matrix


(Figure 3c). Up to time t~ 0:5, the curves I~ r~; t~ are


practically symmetric to the curves of P~ r~; t~ with respect
to a horizontal axis at 0.5, indicating that the sum P~ I~ is
nearly constant and that almost no crosslinks B formed at
this stage. On the other hand Figure 3d shows that the
concentration B~ slowly increases with distance from the
particle. The small amount of crosslinks close to the vector
particle is due to similar kinetics of chemical reaction and
diffusion of the crosslinker. As a result, in the vicinity of the
particlematrix interface the majority of anhydride groups
are transformed into intermediate groups I which cannot
enter into reaction with each other. Only the small fraction
of anhydride which did not react with the crosslinker will
react with the intermediate species forming a small
amount of crosslinks B. The concentration of the intermediate species I~ decreases with the distance from the
particle while the number of unreacted functional groups
on the polymeric chains increases. This is favorable to an
increase of the crosslink concentration B~ away from the
particle interface. However, at higher distances from the
particle, the concentration of intermediate species I~
decreases to such an extent that the crosslink concentration
formed in this area also decreases thus leading to a
nonmonotonic variation of the crosslink distribution
function.

Macromol. Theory Simul. 2011, 20, 000000

2011 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim

www.MaterialsViews.com

Early View Publication; these are NOT the final page numbers, use DOI for citation !!

Diffusion Crosslinking Kinetics . . .

www.mts-journal.de

()

%
1.0
C

(b) %
1.0
P

%
C60
1

40

0.8

20

0.6

0
0.0

0.8

1.0

2.0

3.0

4.0

%
r5.0

0.6

4
0.4

0.4

3
6

0.2

0.2
5
6

0.0
0.0

1.0

2.0

3.0

%
r5.0

4.0

0.0
0.0

1.0

2.0

3.0

%
r5.0

4.0

(d) B
%
0.6

(c) 1.0
I%
0.8

7
0.4

0.6

5
0.4

5
6

0.2

0.2

1
0.0
0.0

1.0

1
2.0

3.0

4.0

5.0
%

0.0
0.0

1.0

2.0

3.0

4.0

5.0
r%

~ ~r (a), unreacted groups on polymer chains P~~r (b), intermediate

Figure 3. Dimensionless concentrations of the crosslinker molecules C
~~r (d) versus dimensionless distance ~r from the vector particle center at different
groups attached to a single chain ~I~r (c), and crosslinks B
time points: ~t 0:1 (1), 0.5 (2), 2 (3), 4 (4), 10 (5), 40 (6), 1 000 (7).

In the course of time, the spatial distribution functions of

concentrations of the intermediate species, crosslinks, and
polymer functional groups approach to the asymptotic
(limiting) functions I~r~, B~r~, and P~r~. Depending on the
system parameters such a condition was found to be
achieved at the dimensionless time between 103 and 104
with an accuracy of 103. The radial position at which the
concentration of the intermediate species becomes negligibly small (r~ 3:7 in Figure 3) coincides with the point
beyond which remaining anhydride functions exist
(P~r~ > 0). This position also corresponds to the maximum
value of crosslink concentration B~r~ and corresponds to the
layer where the crosslinking is the most efficient (if all
initially present P functions are involved in crosslinks at the
end of the reaction, a maximum final value of B~ 0:5 is
obtained). It should be noted that at the end of the
crosslinking process a certain amount of unreacted

Macromol. Theory Simul. 2011, 20, 000000

2011 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim

Early View Publication; these are NOT the final page numbers, use DOI for citation !!

www.MaterialsViews.com

intermediate species and polymer functional groups I

and P remains in the system. These species are spatially
separated and therefore unable to react with each other,
since they are unable to diffuse.
The final distribution of crosslink concentration around
the particles will control the mechanical properties of the
material. Therefore it is important to estimate the influence
of the different system parameters namely: the dimensionless reaction rate constant K, the stoichiometric coefficient
j, the ratio d of the diffusion coefficients, and that n of the
reactivity of the intermediate species over that of the free
crosslinker, on the limiting functions B~r~, I~r~, and P~r~.
Figure 4 shows the influence of parameter K on the limiting
concentration distributions of crosslinks, intermediate
species, and unreacted anhydride groups at the stoichiometric composition and at d 1 and n 0:5. A reduction of
K results in a more uniform distribution function for the

S. A. Patlazhan, A. A. Berlin, M. Bouquey, C. A. Serra, R. Muller

www.mts-journal.de

%
0.6
B

()

5
0.4

4
3
2

0.2

0.0
0.0

1.0

2.0

(b) %
I
1.0

3.0

4.0

%
r5.0

(c) %
0.40
P

0.8

0.30

1
0.6

2
3

0.20

0.4

0.10

0.2

3
5

0.0
0.0

1.0

2.0

3.0

4.0

5.0
r%

0.00
0.0

1.0

2.0

3.0

4.0

5.0
r%

~ ~r (a), intermediate species ~I~r (b), and unreacted polymer functional

Figure 4. Limiting concentration distribution functions of crosslinks B
groups P~~r (c) at the stoichiometric composition, d 1 and n 0:5 for different values of dimensionless reaction rate constant: K 1:8 (1), 0.6
(2), 0.2 (3), 0.06 (4), 0.01 (5).

crosslinks B~r~ which also tends toward the maximum

possible value of 0.5, while the final concentrations I~r~ and
P~r~ decrease. At K  0:01 the concentration distribution of
crosslinks B~r~ and intermediate species I~r~ turn almost
homogeneous while the concentration of anhydride groups
becomes negligible. This is not surprising since a reduction
of the dimensionless reaction rate constant leads to
predominance of the diffusion rate over the rate of chemical
reactions and to more uniform distribution of the free
crosslinker C at any time. As a consequence, a more uniform
distribution of reaction products B~r~ and I~r~ is expected.
One should notice that a reduction of K can be reached
either by an increase of the diffusion coefficient Dout or a
decrease of the reaction constant kCP , or the vector particle
size a, or of the initial concentration of anhydride groups P0.
On the contrary, an increase of K results in the formation of
a sharp maximum for the limiting crosslink distribution

function B~r~ at a dimensionless distance r~ of about 3.6 from

the center of the vector particle.
In practice it might be easier to achieve controlled
variations of the stoichiometric coefficient j than matching
the dimensionless reaction rate constant K. The limiting
concentration distribution functions B~r~, I~r~, and P~r~
calculated for different values of j are presented in Figure 5
at K 0:6, d 1, and n 0:5 for a reaction time t  1000. It
can be seen that reduction of the stoichiometric coefficient
leads to narrowing the limiting distribution of crosslink
concentration B~r~ and shifting its maximum (always
corresponding to B~ 0:5) toward the particlematrix
interface (Figure 5a). In doing so, the crosslink concentration in the immediate vicinity of the particle quickly
increases with decreasing j. However, below j 0:0625, the
distribution of crosslink concentration B~r~ becomes a
monotonously decaying function of r~. The observed

Macromol. Theory Simul. 2011, 20, 000000

2011 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim

www.MaterialsViews.com

Early View Publication; these are NOT the final page numbers, use DOI for citation !!

Diffusion Crosslinking Kinetics . . .

www.mts-journal.de

() 0.6
B%
2

3
6

0.4

0.2

0.0
0.0

1.0

2.0

3.0

4.0

5.0
r%

(c) P%

(b) I%
0.8

0.80

7
6
5

1
0.4

0.40
2

0.0
0.0

1.0

2.0

3.0

4.0

5.0
r%

0.00
0.0

1.0

2.0

3.0

4.0

5.0
r%

~~r (a), intermediate groups ~I~r (b), and unreacted polymer functional
Figure 5. Limiting concentration distribution functions of crosslinks B
groups P~~r (c) for different stoichiometric coefficients: j 2 (1), 1.2 (2), 1 (3), 0.5 (4), 0.25 (5), 0.125 (6), 0.0625 (7) at K 0:6, d 1, n 0:5.

behavior for B~r~ can be attributed to the decrease of the

limiting concentration of the intermediate groups I~r~
(Figure 5b) and the corresponding increase of the limiting
concentration P~r~ of anhydride groups (Figure 5c). In the
vicinity of the particlematrix interface, this leads to an
Kn
increased probability of the reaction I P ! B and of the
formation of crosslinks. Below j 0:0625, all intermediated
species which are formed will react with anhydride
functions and lead to the formation of crosslinks. No
residual intermediate species remains at long reaction
times and a further decrease of j will lead to a decrease of
crosslink concentration at any position. On the other hand,
the limiting crosslink distribution function B~r~ is widened
at j > 1. Its maximum is shifted toward the outer zone of the
representative volume element while the domain corresponding to small concentrations of crosslinks tends to
extend. Note that, as already observed in Figure 3 and 4, the
positions of the maxima in the limiting functions B~r~

Macromol. Theory Simul. 2011, 20, 000000

2011 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim

Early View Publication; these are NOT the final page numbers, use DOI for citation !!

www.MaterialsViews.com

coincide with those where the limiting I~r~ and P~r~

functions become zero.
Using vector particles of different nature may modify the
diffusion coefficient Din of the crosslinker inside the particle
and therefore the ratio d Din =Dout . Figure 6 shows limiting
crosslink concentration distributions B~r~ calculated at
different values of d at the stoichiometric composition j 1
and at K 0:6 and n 0:5. It can be seen that decreasing the
diffusion coefficient of the crosslinker inside the particle
leads to a more uniform distribution function B~r~ and
consequently to a reduction of the concentration I~r~ of the
unreacted intermediate species in the vicinity of the
particle. Such a behavior is similar to the result found for
small values of the dimensionless reaction rate constant K.
This is not surprising since a reduction of the diffusion
coefficient inside the particle with respect to that in the
polymer matrix (corresponding to a reduction in d) will lead
to a more uniform distribution of the free crosslinker inside

S. A. Patlazhan, A. A. Berlin, M. Bouquey, C. A. Serra, R. Muller

www.mts-journal.de

%
0.6
B

Also, a decrease in n results in the saturation of the

considered distribution functions.

0.4

Conclusion

0.2

2
1
0.0
0.0

1.0

2.0

3.0

4.0

%
r5.0

Figure 6. Limiting distribution function of crosslink concentration

~ ~r for different ratios of diffusion coefficients in the vector
B
particle and polymer matrix: d 1 (1), 0.1 (2), 0.01 (3), 0.001 (4) at
j 1, K 0:6, n 0:5.

the matrix at any reaction time. A reduction of K has exactly

the same effect. At values of d larger than 1 (Din > Dout ) the
limiting crosslink distribution function B~r~ is very close to
that obtained when d 1 (curve 1 in Figure 6). In this case
the molar flux of crosslinker from the particle toward the
matrix is only limited by the diffusion coefficient Dout in the
polymer matrix, and the diffusion kinetics inside the
particle has negligible influence.
Finally, it makes sense to estimate an influence of the
ratio n kIP =kCP of rate of the reaction of intermediate
species I with the anhydride groups P to that of the reaction
of the crosslinker C with P representing relative difference
in mobility of the crosslinker molecules C and intermediate
groups I attached to polymer chains. Theoretically, n is
ranging from 0 to 0.5. Figure 7a shows that a decrease in n
results in some reduction of the limiting distribution
function B~r~ in the vicinity of the vector particles as well as
in the middle area between the neighbor inclusions
whereas the maximum of this function is independent
of n. These features of B~r~ are accompanied by different
limiting residual concentrations of the intermediate
product I~r~ and anhydride groups P~r~ corresponding to
different n (see Figure 7b and c). It is seen that a decrease in n
leads to an increase of I~r~ near the vector particle while P~r~
grows between the particles. This is originated by the
slowing down reaction between I and P groups which
inevitably leads to the growing concentration of the
unreacted anhydride groups P~r~. As seen from Figure 7,
the variation in the ratio n of rates of chemical reactions
does not practically influences to positions of zero values of
I~r~ and P~r~ as well as of that of maximum of B~r~ function.

The limiting distribution function B~r~ of crosslink concentration for an initially stoichiometric composition of an
anhydride functionalized polymer and vector particles
containing initially the crosslinker (butanediol) molecules
is predicted in most cases to be nonmonotonic: its
maximum is located at some distance away from the
particlematrix interface while the crosslink concentration
remains small in the close vicinity of the particles.
Variations of the stoichiometric coefficient j representing
the initial ratio of reactive groups carried by the free
crosslinker and by the polymer chains leads to significant
modifications of the final crosslink concentration distribution: as j is reduced, the distribution function B~r~ is
narrowed and its maximum is shifted toward the particle
surface whereas its maximum value remains equal to the
theoretical value of 0.5.
In a certain range of the stoichiometric coefficient j, the
crosslink concentration turns out to be quite small both
near and in a distance from the particle. This means that
eventually each vector particle will be surrounded by a
none or slightly crosslinked domain, which will itself be
encapsulated in a more densely crosslinked polymer shell.
Outside these capsules, the material again corresponds to a
none or weakly crosslinked domain, depending on the
critical conversion inherent to the polymer system.[8] The
schematic sketch of such a structure is presented in
Figure 8a.
A further decrease of the stoichiometric coefficient j
(below the critical value of 0.0625 for the above reference
system) tends to increase the crosslink density in the
neighborhood of the particle. As a result, each particle will
be surrounded by a densely crosslinked shell while the rest
of the matrix remains in the state of a none or weekly
crosslinked domain (see Figure 8b). If j  1 the domain with
the highest crosslink concentration is shifted toward the
periphery of the representative volume element where the
polymer matrix is highly crosslinked. At the same time,
each vector particle will be surrounded by a none or weakly
crosslinked domain (Figure 8c).
Vector particles with a relatively small diffusion
coefficient Din of the crosslinker may be used to produce
materials with a homogeneous crosslink concentration.
The same result can be reached with a significant
decrease of the dimensionless reaction rate constant
K kCP P0 a2 =Dout which may for instance be obtained by
reducing the mean radius of the vector particles or the
initial concentration P0 of polymer functional groups.

Macromol. Theory Simul. 2011, 20, 000000

2011 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim

www.MaterialsViews.com

Early View Publication; these are NOT the final page numbers, use DOI for citation !!

Diffusion Crosslinking Kinetics . . .

www.mts-journal.de

%
() B
0.6

0.4

0.2

3
2
0.0
0.0

1.0

2.0

3.0

%
r5.0

4.0

(c) 0.30
P%

(b) 1.0
I%
3

3
2

0.8

0.20
0.6

1
0.4

0.10
0.2

0.0
0.0

1.0

2.0

3.0

4.0

5.0
r%

0.00
0.0

1.0

2.0

3.0

4.0

5.0
r%

~~r (a), intermediate groups ~I~r (b), and unreacted polymer functional
Figure 7. Limiting concentration distribution functions of crosslinks B
groups P~~r (c) for different ratios n kIP =kCP of rates of chemical reactions: n 0:5 (1), 0.1 (2), 0.01 (3) at j 1, K 0:6, d 1.

()

(b)

(c)

Figure 8. Possible material morphologies after polymer crosslinking. The dark domains correspond to a densely crosslinked network, the
light gray and gray domains represent a none or weakly crosslinked polymer networks, respectively.

Macromol. Theory Simul. 2011, 20, 000000

2011 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim

Early View Publication; these are NOT the final page numbers, use DOI for citation !!

www.MaterialsViews.com

S. A. Patlazhan, A. A. Berlin, M. Bouquey, C. A. Serra, R. Muller

www.mts-journal.de

In the end of the crosslinking process a certain amount of

intermediate and polymer functional groups I and P remain
unreacted: being spatially separated from each other and
having no possibility to be put into contact by diffusion.
Received: December 26, 2010; Revised: March 19, 2011; Published
online: DOI: 10.1002/mats.201000102
Keywords: crosslinking; diffusion; gelation; kinetics (polym.); matrix
[1] B. T. Astro
m, Manufacturing of Polymer Composites, Chapman
& Hall, London 1997.

[2] L. Abbas, M. Bouquey, J. J. Flat, R. Muller, Eur. Polym. J. 2008, 44,

1238.
[3] T. Salmi, E. Paatero, P. Nyholm, M. Still, K. Na
rhi, Chem. Eng. Sci.
1994, 49, 5053.
[4] S.-A. Chen, K.-C. Wu, J. Polym. Sci., Polym. Chem. Ed. 1982, 20,
1819.
[5] R. Jedlovcnik, A. Sebenik, J. Golob, J. Korbar, Pol. Eng. Sci. 1995,
35, 1413.
[6] A. A. Berlin, V. G. Oshmyan, Polym. Sci. USSR 1976, 18A, 2282.
[7] L. K. Pakhomova, O. B. Salamatina, S. A. Artemenko, A. A. Berlin,
Polym. Sci. USSR 1978, 20B, 554. 2.
[8] P.-G. De Gennes, Scaling Concepts in Polymer Physics, Cornell
University Press, Ithaca and London 1979.

Macromol. Theory Simul. 2011, 20, 000000

2011 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim

10

www.MaterialsViews.com

Early View Publication; these are NOT the final page numbers, use DOI for citation !!