Linear Temperature Programming Rheometers for the Cure of Rubbers.

Effect of the Various Parameters

Linear Temperature Programming Rheometers for the Cure of

Rubbers. Effect of the Various Parameters

I.D. Roscaa, R. Grangerb and J.M. Vergnaudb,1

a
Department of Chemical Engineering, University Politehnica, Bucarest, Romania
b
Faculty of Sciences, University of St-Etienne, 42023, Saint-Etienne, France

Received: 25 April 2003 Accepted: 14 August 2003

SUMMARY
In spite of their drawbacks, rheometers are usually run under isothermal conditions for
the cure of rubbers: At least three experiments are needed at different temperatures
selected within a narrow temperature window to determine the kinetic parameters of the
cure of rubbers, and it takes some time for thermal equilibrium between the sample,
initially at room temperature, and the dies to be established. With the rheometer in
scanning mode, the dies and the sample are heated from room temperature up to the
selected final temperature at a constant rate. The heating rate is crucial in this new
method. The rubber used in this study is an ethylene propylene diene monomer
vulcanised with 20% sulphur and with a rather high cure enthalpy. The state-of-cure
versus temperature curves obtained in scanning mode have been calculated using a
numerical model that takes into account the heat transfer by conduction through the
rubber, the coefficient of heat transfer at the die-rubber interface, and the kinetics of the
heat generated by the cure reaction. The kinetic parameters were determined
mathematically from these state-of-cure versus temperature curves, which are similar to
torque-temperature curves, by obtaining straight lines from them or using a least squares
method. The kinetic parameters calculated in these ways were similar to those introduced
by obtaining the state-of-cure versus temperature curves using the numerical model.
Whatever the heating rate, ranging from 2 to 10 K/min, and whatever the coefficient of
heat transfer at the die-rubber interface which expresses the quality of the contact, the
moving die rheometer can be used in scanning mode to determine the kinetics of the cure
of rubbers, even when the enthalpy of cure is high.

1. INTRODUCTION involving heat transfer by conduction through the

sample thickness, the other resulting from the slightly
Many rubber compounds are routinely prepared and exothermic cure reaction. Moreover, the cure reaction
analysed1,2, in order to be used in various applications, itself is highly complex, kinetically speaking3, even
so any studies of their cure processes are of great if it is often expressed in terms of temperature and of
importance3,4. In the same way as for thermosetting the activating agent concentration by an Arrhenius’
resins, when curing rubbers the samples should be equation3,4 over a 30 K range. From the point of view
heated up to a temperature at which the irreversible of economics, the curing rubbers is of great importance
reaction starts, whereupon the fresh plastic uncured if we are to optimise the cure time and minimise the
material is changed into an elastic three-dimensional production of defective products.
molecular network. The curing of rubbers is complex,
since the heating process consists of two stages, one
Because of their low enthalpy of cure, rubbers have
been successfully studied by various methods such
as the monitoring the viscosity or torque change of the
1 Corresponding author. Prof. J.M. Vergnaud, Chemin privé de
sample during the cure, while this sample is kept at
Grange Bruyas, route de Chavanne, 42400 Saint-Chamond, France,
a constant temperature. Thus, the Mooney viscometer,
tel.: 33.4.77.31 86 73; fax: 33-4-77-31 86 73 the Wallace-Shawbury curometer, the oscillating disc
E-mail address: Vergnaud.Jean-Maurice@wanadoo.fr rheometer (ODR) and the moving die rheometer

Polymers & Polymer Composites, Vol. 12, No. 1, 2004 87

 I.D.Rosca, R. Granger and J.M.Vergnaua

(MDR) have been developed and marketed, the MDR On the other hand, the process of cure has often been
being introduced most recently in 19855. In the MDR, analysed by taking into account heat transfer by
a thin sheet of rubber (around 2 mm) previously at conduction through the rubber and also the kinetics
room temperature is introduced between two dies of the heat generated by the reaction. In one study, the
kept at a pre-determined temperature. Then the lower state of cure was evaluated from the swelling in
die oscillates with a reaction torque transducer toluene of thin rubber samples cured at various times
positioned above the upper die. Because of the faster and temperatures, but as stated by the authors12, ‘the
heat transfer resulting from thinner samples and a method was tedious and highly time-consuming’.
better die design, the MDR was found to give shorter Other studies involved measuring and calculating
times of cure than the ODR5. The MDR run under the temperature profiles developed through the
isothermal conditions should be used successively at thickness of the sample at various times: one13, by
three different temperatures selected from within a introducing a temperature reference point at 149°C,
temperature window of around 30°K, to evaluate the leading to a rate coefficient of cure increasing by a
kinetics of the cure reaction. The activation energy factor of 1.85 per 10°C; another14, assuming that the
and the pre-exponential factor have been calculated rate of cure doubles for each 10°C increase in
from the fractional modulus-time values obtained at temperature around the reference temperature of
these three temperatures for EPDM compounds6,7-10, 150°C. A constant time conversion of 1.3 per 10°C
using the MDR. Very often the modulus changes with was also found for the sulphur vulcanisation of
time in accordance with first order kinetics, and the EPDM blends6 around the reference temperature of
rate constant varies with temperature according to an 150°C . Numerical models have been built and tested
Arrhenius’ equation 7-10. The heat generated by the for the cure of NR with sulphur4,15,16 and of EPDM
cure reaction is not considered, which is acceptable with peroxide17. They took into account all the known
when the cure enthalpy is low. But various drawbacks facts, especially heat conduction through the sample
appear with the MDR run under isothermal conditions: and the kinetics of the heat generated by the cure
reaction. These models were able to evaluate not only
(i) There is only a narrow temperature range over the temperature profiles developed through the
which meaningful data can be obtained3,4,11, sample, but also the profiles of state-of-cure, provided
because a low temperature leads to very long that the kinetic parameters of the cure reaction were
cure times, while too high a temperature is known.
ruled out by the existence of an induction
period before the cure initiates. The main difference between rubbers and
(ii) The applied temperature is the set apparatus thermosetting resins is the enthalpy of cure, which is
temperature and not the true sample far larger for thermosetting resins, this fact being the
temperature. At the beginning, the sample is at main consideration when selecting methods for
room temperature when it isintroduced into the determining the parameters of the kinetics of cure.
dies of the MDR, which is kept at the pre- Calorimetry, either under isothermal conditions or in
determined temperature. It takes some time for scanning mode, has been widely used to study the
thermal equilibrium to be established8. When cure of thermosetting resins. In this technique, the
the cure enthalpy is not negligible, temperature heat flux emitted by the exothermic reaction is
gradients develop through the sample, resulting recorded as a function of time when the apparatus is
from the internal heat generated by the reaction4. kept at constant temperature (DC) or alternatively of
temperature in scanning mode (DSC). In the seventies
Thus, in order to develop the MDR, which up to now and the early eighties, most workers preferred using
has been used only under isothermal conditions, the the constant temperature method11,18,19. But in-depth
theory of the process has been developed for the MDR studies20,21 showed that in this case, the process is far
to be run in scanning mode, at a constant heating from simple. Fresh uncured resin, initially at room
rate9, or even increasing it in proportion to the square temperature, when dropped into the calorimeter kept
root of time10. The advantages of this new method9 at the selected temperature, leads to a very large
are evident, as only one experiment is necessary to endothermic heat flux associated with the conduction
evaluate the kinetic parameters of the cure. Both the heating stage. This hides the first part of the reaction,
temperature profiles and those for the state of cure are and it is only when the temperature of the sample has
rather flat in the sample, proving the good homogeneity reached that of the oven, that the exothermic heat flux
of the state of the cure. Of course, the main variable in resulting from the exothermic cure reaction can be
this new way of using the MDR technique is the rate observed. Moreover, there is only a narrow
of heating,9,10 which should be studied in-depth. temperature window over which meaningful data

88 Polymers & Polymer Composites, Vol. 12, No. 1, 2004

 Linear Temperature Programming Rheometers for the Cure of Rubbers. Effect of the Various Parameters

can be obtained3,4. As a result DSC proved to be the heated in the MDR in scanning mode with the
best method4,20. A few attempts have been made to linear temperature programming.
use the various calorimetry techniques to study the
cure of rubbers. DSC was used to determine the (ii) The treatment of the state-of-cure versus
temperature at which the reaction starts11,19, and to temperature curves, which are similar to the
measure the cure enthalpy3,4,11,19-21. The same torque-temperature curves obtained when the
technique was also used to evaluate the kinetics MDR is run in scanning mode with linear
parameters of the cure of natural rubber with temperature programming.
sulphur18, and of ethylene/propylene/diene monomer
cured with peroxide17, assuming that the rate of cure 2.1 Calculation of the State of Cure-
varies with temperature according to an Arrhenius’ Temperature Curves with the MDR in
equation. Scanning Mode
2.1.1 Assumptions
The first objective of this present project was to
In order to define the process precisely, the following
evaluate the torque-temperature histories obtained
assumptions were made:
with the MDR run in scanning mode at various
heating rates. The simple way to calculate these
curves, using the torque-time histories and expressing (i) the rubber sheet is heated by unidirectional
them as a function of temperature has not been conduction through its thickness.
considered here, because a similar method is used in
(ii) the contact at the rubber-die interface, defined
the opposite sense to estimate the kinetic parameters
by a coefficient of heat transfer, is either perfect
from the torque-temperature curves. As a
or not.
proportionality was found between the state of cure
expressed in terms of heat of cure and the torque (iii) the temperature of the dies is increased linearly
measured in the MDR8, the state-of-cure versus with time at a constant heating rate.
temperature histories in the sample have been
calculated by using a numerical method, taking into (iv) the internal heat generated by the overall reaction
account the heat transfer by conduction through the of cure follows an nth order, the rate of which is
sample and at the die-rubber interface, as well as the expressed by the Arrhenius’ relation.
heat generated by the cure reaction. (v) the state of cure at time t is expressed as the heat
evolved up to time t as a fraction of the total heat
The other objective of this project was to define evolved.
precisely the role played by various parameters, e.g. (vi) the state of cure is proportional to the torque as
the heating rate. A high rate is responsible not only for a fraction of its maximal value, as shown in a
a decrease in the cure time, but also for the steeper previous paper8.
temperature profiles through the thickness of the
sample9,10,20. Another parameter, the coefficient of 2.1.2 Mathematical Treatment
heat transfer at the rubber-die interface, measures the
quality of the contact, and should be examined. The basic equations of the heat transfer are given4,
Finally, the effect of a rather high enthalpy for the with the unidirectional heat conduction through the
cure of rubbers was to be considered. Thus it was thickness of the sheet, where the contribution of the
considered appropriate to compare the values of the heat conduction and the heat generated by the cure
kinetic parameters obtained by the mathematical reaction appears in the right-hand side:
treatment of these curves with those introduced in
the numerical method for calculating the state-of-
∂T ∂2 T ∂Q
cure versus temperature curves. ρC = λ 2 +ρ (1)
∂t ∂x ∂t

2. THEORETICAL and with the coefficient of transfer h at the die-rubber

interface:
Two aspects were considered:

(i) The calculation of the state of cure developed ∂T

−λ = (Tt ,die − T1 ,S ) (2)
through the thickness of the sample when it is ∂x

Polymers & Polymer Composites, Vol. 12, No. 1, 2004 89

 I.D.Rosca, R. Granger and J.M.Vergnaua

Upon putting SOC=100.Z (8)

Torque1 SOC1 while Z is obtained by integrating the equation (5):

Z= = (3)
Torque∞ SOC∞
Z = 1 - exp(-S) (9)

where the torque and state of cure SOC are at time t,

and infinite time (which in fact corresponds with the where the function S is defined by:
time necessary for the cure to be achieved).

 E 
1

The rate of heating b is defined by:

0
∫
S1 = k 0 exp − dt
 RT1 
(10)

T1 = T0 + b.t (4)
with the recurrent relationship:

The differential equation under isothermal condition

expressing the change with time of the partial heat  E 
evolved from the cure reaction as a fraction of the total S1+ ∆t = S1 + k 0 ∆T.exp −  (11)
 RT1 
heat evolved (which is also the change with time of
the torque value8) , is written as follows:
and:

dZ  E 
= k 0 (1 − Z)n exp −  (5) S0 = 0 (12)
dt  RT 

Thus, the state-of-cure versus temperature, or the

2.1.3 Numerical Treatment torque-temperature histories can be calculated at any
As no analytical solution exists for the problem, it is temperature and time, whatever the rate of heating.
resolved by a numerical treatment4. The thickness of
the sample in the dies of the MDR was divided into 2.2 Mathematical Treatment of the Torque-
slices of constant thickness ∆x, each slice being Temperature Curves
defined by an integer n, while the time is divided into
increments of time ∆t. The change in the state of cure with time given by
equation 5 should be expressed as a function of
temperature, in scanning mode with the linear
The principle is as follows: the new temperature TNn temperature programming, leading to:
at position n after elapse of time ∆t is expressed in
terms of the previous temperature at the same and
adjacent places: dZ k 0  E 
= (1 − Z)n .exp −  (13)
dT b  RT 
∆Q
[Tn −1 + ( M − 2)Tn + Tn +1 ] +
1
TN n = (6)
M C which becomes in the logarithm form:

with the dimensionless number M

 dZ   k0  E
Ln  = Ln  + n.Ln(1 − Z) − (14)
 dT   b RT
( ∆x )2 ρC
M= (7)
∆t λ
 dZ 
By plotting Ln  versus Ln(1 - Z) in equation 14,
 dT 
The state of cure expressed in percent is given by:
the order n of the overall reaction of cure can be

90 Polymers & Polymer Composites, Vol. 12, No. 1, 2004

 Linear Temperature Programming Rheometers for the Cure of Rubbers. Effect of the Various Parameters

obtained, being equal to the slope, by neglecting the 3.3 Characteristics of the Cure and of Heat
E Transfer
last term which is very low.
RT The values of the kinetic parameters of the cure and
the thermal parameters are shown in Table 2.
 dZ  n
In the same way, by plotting Ln   (1 − T)  versus
 dT   4. RESULTS
1 4.1 Questions to be Answered
the reciprocal temperature , the slope gives the
T Two questions were considered in succession:
value of the energy of activation.

(i) One was concerned with the calculation of the

3. EXPERIMENTAL torque-temperature curves in the MDR apparatus
run in scanning mode with linear temperature
3.1 Materials used for the Cure
programming. It would have been easy to
EPDM compounds containing 20% sulphur were calculate the torque-temperature curves by using
used (Hutchinson). Their main composition is given equation 13, expressing the change in the torque
as follows (w/w(%) in Table 1. with temperature, and the kinetic parameters
obtained from the MDR measurements run under
isothermal conditions. But, as the same
Table 1 Composition of the sample (w/w)
equation 13 is the basis for the mathematical
EPDM elastomer: 100 C black: 30 plasticizer: 4 treatment necessary to evaluate the kinetic
sulphur as other chemical: 2 parameters from these torque-temperature
vulcanising agent:19 curves, the same method would be used in both
senses.
Thus, another method was preferred, based on
3.2 Apparatus the calculation of the state-of-cure versus
temperature curves, the state of cure being
A Monsanto 2000 E Moving Die Rheometer (MDR proportional to the torque9, using the numerical
Alpha Technologie) was used under isothermal model. This method of calculation allows one
conditions within the 140-160°C range, to determine to evaluate the effect of parameters of interest
the kinetic parameters of the cure, except for the such as the heating rate, the enthalpy of cure
enthalpy of cure which was obtained by calorimetry. and the quality of the contact at the rubber-die
interface.
A calorimeter DSC 111 (SETARAM) was run in (ii) The other related to two methods of calculation
scanning mode at a heating rate of 2 K/min to determine of the kinetic parameters from the state-of-cure
the enthalpy of cure. The heat flux was recorded as a versus temperature curves. These methods were
function of temperature (or time). Samples around described and evaluated by comparing them to
0.15 g were placed into a cylindrical holder with an the original data introduced when calculating
internal diameter of 6 mm. The thermal parameters, these curves using the numerical treatment.
such as the thermal conductivity, the density and the
thermal capacity were obtained from Hutchinson.

Table 2 Kinetic and thermal parameters

Lnk0 = 21.9 k0(s-1) E/R = 11,550 K ρ = 1,170 kg.m-3

C = 2.34 kJ.kg-1.K-1 λ = 0.25 W.m-1.K-1 n =1.18

h infinite hL Enthalpy of cure = 104 kJ kg-1

h finite with =1
λ

Polymers & Polymer Composites, Vol. 12, No. 1, 2004 91

 I.D.Rosca, R. Granger and J.M.Vergnaua

4.2 Calculation of the Change in Torque (or linearly with time. The heat transfer developed
State of Cure) with Linear Temperature through the thickness of the sample when the dies of
Programming the MDR apparatus were heated at a constant rate,
and the state of cure developed through the rubber
4.2.1 Parameters of Cure Introduced for sample are both calculated using the numerical model.
Calculating the Curves
The parameters of cure were determined from the Two values of the coefficient of heat transfer at the
results obtained with the MDR run under isothermal die-rubber interface were selected, one being infinite
conditions at various temperatures within the 140- (meaning that the temperature of the rubber surface is
160°C range, using an EPDM rubber cured with 20% always equal to that of the die with a perfect contact)
sulphur. These data are typical, in the sense that the and the other being very low4. Uniform profiles of the
order of the overall reaction is far from being equal to state of cure through the thickness of the sample are
1, as often found in the literature7-9,15,17 with EPDM found in Figure 1, as already shown for 2% peroxide
rubbers vulcanised with peroxide. In fact, it would in a previous paper9. This is because of the scanning
not have been necessary to use experimental values, system, where the rubber and the dies initially at
as the purpose of the present study was to test a room temperature are heated at a rather low heating
method of evaluation of these parameters, provided rate, and also because the rubber sample is thin
that they are taken in a reasonable range. The enthalpy enough for the heat transfer to be fast in spite of the
of cure has been measured by calorimetry in scanning low heat conductivity of the rubber. This is not the
mode. The values of the kinetic parameters of the case in the MDR run under isothermal conditions,
overall reaction, as well as the cure enthalpy, and the where steep gradients of temperature and state of
conduction heat through the rubber, are shown in cure through the thickness of the sample have been
Table 2. found, especially at the beginning of the process
when the sample at room temperature is introduced
in the heated dies8,9. Thus with the torque being
4.2.2 Calculation of the Torque- proportional to the state of cure9,10,17, the state-of-
Temperature (State of Cure-Temperature) cure versus temperature curves calculated in the
Curves scanning mode of the MDR are able to represent the
In scanning mode, the temperature of the dies and of torque-temperature curves used to evaluate the
the rubber, both starting at room temperature, increases parameters of the cure, with the help of equation 14.

Figure 1 Profiles of temperature (left) and state of cure (right) developed through the thickness of the 2 mm thick
sample during the scanning at 10 K/min of the MDR, with 20% sulphur. Times are shown in seconds

92 Polymers & Polymer Composites, Vol. 12, No. 1, 2004

 Linear Temperature Programming Rheometers for the Cure of Rubbers. Effect of the Various Parameters

The state-of-cure versus temperature curves are shown (i) The change in state-of-cure versus temperature
in Figure 2 for various heating rates, and the (torque-temperature), as shown in Figure 2,
corresponding state-of-cure versus time curves are follows S-shaped curves, starting at about the
shown in Figure 3. The curves calculated by using the same temperature of the die, whatever the value
original kinetic parameters and shown in Table 2 are of the heating rate selected through the rather
denote by ‘orig’ in these two figures. The following large range of 2-10 K/min.
conclusions are worth noting from Figures 2-5:

Figure 2 State of cure - temperature (torque-temperature) histories as a function of temperature calculated with the
MDR apparatus run in scanning mode with linear temperature programming, for various heating rates (2; 5; 10; in
K/min). The curves are calculated with various kinetic parameters: original values (orig); estimated values from the
slopes in Figures 6 and 7 (estim); fitted with the least-squares method (fit), with 20% sulphur

Figure 3 State of cure - temperature (torque-temperature) histories as a function of time obtained with the MDR
apparatus run in scanning mode with linear temperature programming, for various heating rates (2; 5; 10; in K/min).
The curves are calculated with various kinetic parameters: original values (orig); estimated values from the slopes
in Figures 6 and 7 (estim); fitted with the least-squares method (fit), with 20% sulphur

Polymers & Polymer Composites, Vol. 12, No. 1, 2004 93

 I.D.Rosca, R. Granger and J.M.Vergnaua

Figure 4 MDR in scanning mode, with linear temperature programming of 10 K/min. Evaluation of the effect of the
quality of the contact at the rubber-die interface, with h infinite (full line) or finite (dotted line)

Figure 5 Temperature (left) and state of cure (right) at the mid-plane of the sample as a function of time, during the
scanning of the MDR at 10 K/min.. Effect of the quality of the contact at the dies-rubber interfaces. Perfect contact:
full line. Poor quality of contact: dotted line

(ii) The effect of the heating rate was of great (iii) The temperature of the dies for which the state
importance, the curves being shifted to higher of cure (torque) is a maximum depends on the
temperature when the heating rate was heating rate. The higher the heating rate, the
increased, as shown in Figure 2. The curves higher the temperature necessary for the torque
were shifted through a translation in the same to reach its maximum value.
way as has been observed for the linear
temperature programming of EPDM cured with (iv) The change in the state of cure (torque) expressed
peroxide9. as a function of time, in Figure 3, is of interest

94 Polymers & Polymer Composites, Vol. 12, No. 1, 2004

 Linear Temperature Programming Rheometers for the Cure of Rubbers. Effect of the Various Parameters

from the point of view of economics. Thus, (vii) The effect of the quality of the thermal contact
other S-curves were also obtained, but the effect at the rubber-die interface, as well as the enthalpy
of the heating rate appears to be of great of cure also appears in Figure 5. Even at a high
importance, as much longer times are necessary heating rate (10 K/min), the curves describing
to perform experiments at a lower heating rate, the increase in temperature at the mid-plane of
the times of the experiment being a little less the rubber with time, and the corresponding
than inversely proportional to the heating rate. increase in the state of cure were much the
same. Of course, the temperature-time curve
(v) The effect of the quality of the contact at the die-
was shifted to longer times for a lower coefficient
rubber interface, measured by the coefficient of
of heat transfer, with exceptions at around 800 s
heat transfer at this interface, clearly appears in
and 150°C where these curves intersect, as well
Figure 4 at the high heating rate of 10°K/min,
as at around 950 s and 175°C. This fact, surprising
where the state-of-cure versus temperature
at first sight, results from the internal heat
curves obtained (either with a low coefficient of
generated by the cure reaction, which dissipates
heat transfer at the interface or with a perfect
into the dies more slowly when the coefficient
contact), were nearly the same. As shown in this
of heat transfer at the interface is low4 .
figure, the quality of the thermal contact between
the dies and the sample, measured by the
4.3 Determination of the Kinetic Parameters
coefficient of heat transfer at this interface, does
of the Cure with the MDR in Scanning Mode
not play an important role in the process. In fact,
the two values of this coefficient of heat transfer The kinetic parameters of the cure were determined
were quite different, the one being infinite and by a mathematical treatment based on equation 14,
the other rather low4 , but no air layer was using the state-of-cure versus temperature (torque-
assumed to stay at this interface. temperature) curves obtained with the MDR run in
scanning mode when the temperature was increased
(vi) As shown in previous studies8,9, the effect of the
linearly with time.
heat generated by the cure reaction is negligible,
because of the low enthalpy of cure. Moreover,
the temperature gradients leading to other By plotting ln(dZ/dT) versus ln(1-Z), as shown in
gradients of state of cure were nearly flat when Figure 6, an almost straight line was obtained, the
the MDR was used in scanning mode9. slope of which equalled the order of the overall

Figure 6 Estimation of the order of the overall reaction, for a heating rate of 10 K/min

Polymers & Polymer Composites, Vol. 12, No. 1, 2004 95

 I.D.Rosca, R. Granger and J.M.Vergnaua

Figure 7 Estimation of the values of the energy of activation of the overall reaction, with the heating rate of 10 K/min

reaction of cure, as given in equation 14, when the expressing the rate of change of the state of cure
terms in k0 and E are negligible. (or modulus) as a function of temperature, is far
easier to use.
By plotting ln[(dZ/dT)(1-Z)-n] versus the reciprocal (iii) A straight line was obtained over a wide range,
temperature expressed in K, a straight line was either for determining the order of the overall
obtained, which slope equal to E/R (Figure 7). The reaction (Figure 6) or for the two parameters
pre-exponential factor k0 could be obtained in two expressing the temperature dependence of the
ways, by extrapolating the curve in Figure 7, and by rate of the reaction (Figure 7) with the energy of
using Equation 13. activation.
(iv) Finally, more precise values of these three
The following conclusions can be drawn from these parameters (n, E, k0) can be obtained by using a
results: method of least-squares, as shown in Table 3,
where the values of these parameters are
(i) The Freeman-Carroll’s method22 is not a collected for the three values of the heating rate,
convenient one, being highly complex, either as well as the original parameters introduced
for obtaining or for interpreting the data, as for evaluating the state-of-cure versus
already shown temperature curves. This method of least-squares
takes into account not only the values estimated
(ii) The method based on using Equation 14,

Table 3 Values of the kinetic parameters, introduced and calculated

Parameters Introduced Estimated Obtained by Least-squares Method
Heating Rate (K/min) 2 5 10 2 5 10
Order n 1.184 0.914 0.915 0.916 1.18 1.18 1.18
Ln(k0) 21.9 21.9 21.8 21.75 21.8 21.8 21.75
E/R (K) 11550 11550 11490 11480 11485 11480 11470

96 Polymers & Polymer Composites, Vol. 12, No. 1, 2004

 Linear Temperature Programming Rheometers for the Cure of Rubbers. Effect of the Various Parameters

for each kinetic parameter, but also the others successfully tested. It consists of a two step method,
intervening in equation 14. It is thus necessary the first being necessary to estimate the values by
in starting this method to introduce the values obtaining linear curves through making
estimated from the slopes of the curves drawn approximations, followed by the second using a
in Figures 6 and 7. This conclusion also appears least-square method and taking into account all the
in Figures 2 and 3, where the curves calculated parameters.
by this two-stage method were nearly the same
as the ones obtained with the original values of
Interesting results are obtained. The one shows that
the parameters.
the calculated kinetic parameters of the overall
(v) Of course, the enthalpy of cure was not obtained reaction are broadly similar, whatever the heating
from the MDR technique, in the same way as in rate; the other shows that the quality of the contact at
the case of the MDR run under isothermal the die-rubber interface is of minor importance; and
conditions. that the method can be used for a high value of the
enthalpy of cure.
(vi) An interesting result also appeared with the
heating rate, since the kinetic parameters
obtained by this two-stage calculation were REFERENCES
similar, whatever the value given to it.
1. Wake W.C., The analysis of rubber & rubber-like
polymers. McLaren, London (1968).
5. CONCLUSIONS
2. Brown R.P., Physical Testing of Rubbers.
From this analysis of the two methods of running the Applied Science, London (1979).
MDR apparatus, either under isothermal conditions
3. Brazier D.W.; Grassie N. editor. ‘Developments
or in scanning mode, it appears that there is potential
in Polymer Degradation’; Applied Science,
for using the MDR with temperature programming.
London, vol.2 (1987) 27.
From the principle of each method, the advantages
are clear. Thus the MDR requires at least three 4. Vergnaud J.M. and Bouzon J., Cure of
experiments when used under isothermal conditions. Thermosetting Resins, Springer, London (1992).
They have to be carried out at various temperatures
which need to be selected carefully within the narrow 5. Di Mauro P., Proceedings of Internat. Rubber
temperature window over which meaningful data Conference, (1985) 50, Kyoto, Japan. Japan
can be obtained. On the other hand, the MDR driven Society of the Rubber Industry.
in scanning mode needs only one experiment. 6. Sengupta A. and Konar B.B., J Applied Polymer
Sci., 66 (1997) 1231.
The most important factor in this new way of using
the MDR is the heating rate. A value ranging from 2 to 7. Mishra S., Baweja B. and Chandra R.J., J. Applied
10°C/min can be used, a low value being responsible Polymer Sci., 74 (1999) 2756.
for longer experiment times (The time of the 8. Rosca I.D. and Vergnaud J.M., Plast. Rubber
experiment is approximately inversely proportional Composites, 30 (2001) 275.
to the heating rate).
9. Rosca I.D. and Vergnaud J.M., Polymer, 43
(2002) 195.
The torque measured in the MDR was proportional to
the state of cure evaluated from the partial heat 10. Rosca I.D. and Vergnaud J.M., Polymer, 44
evolved from the overall reaction, expressed as a (2003) 4067.
fraction of the total heat evolved. It was possible to
calculate the curves expressing the torque as a function 11. Brazier D.W., Rubber Chem. Technol., 53 (1980)
of the temperature in the MDR run in scanning mode. 437.
This way of working makes it possible to determine 12. Hands D and Horsfall F., Rapra Report. A new
the effect of the quality of the contact between the dies method for simulating industrial cure process,
and the rubber. Vol. 44/3, (1980).

A method for calculating the kinetic parameters from 13. Prentice G.A and Williams M.C., Rubber Chem
the state-of-cure versus temperature curves, similar Technol., 53 (1980) 1023.
to the torque-temperature curves, has been

Polymers & Polymer Composites, Vol. 12, No. 1, 2004 97

 I.D.Rosca, R. Granger and J.M.Vergnaua

14. Kong D., White J.L., Weissert F.C. and Nakajima NOMENCLATURE
N.J., Rubber Chem.Technol., 60 (1987) 140.
b heating rate K/min
15. Accetta A., Le Parlouer P. and Vergnaud J.M., C heat capacity of the rubber in Equation 1
Thermochim. Acta, 59 (1982) 149.
∆t increment of time for calculation of the state of
16. Khouider A., Rochette J.B. and Vergnaud J.M., cure
Thermochim. Acta, 89 (1985) 81. ∆x increment of space for calculation of the state
17. Azaar K., Lamine B.,.Granger R., Rosca I.D. and of cure
Vergnaud J.M., Plast. Rubber Composites, 29 E activation energy of the overall cure reaction
(2000) 253. n order of the overall cure reaction
18. Rochette J.B., Sadr A., Abdul M. andVergnaud λ thermal conductivity of the rubber
J.M., Thermochim. Acta, 79 (1984) 187. k0 pre-exponential factor for the rate of the cure
reaction in Equation 5
19. Brazier D.W. and Nickel G.H., Rubber Chem.
Technol., 48 (1975) 26. h coefficient of heat transfer at the die-rubber
interface
20. Vergnaud J.M., Thermochim. Acta, 114 (1987) L thickness of the rubber sample (2 mm)
15.
M dimensionless number expressed by
21. Brazier D.W., Nickel G.H. and Szentgyorgyi Z., Equation 7
Rubber Chem Technol., 53 (1980) 160. Q heat generated by the overall cure reaction in
Equation 1
22. Freeman E.S. and Carroll B.J, J. Phys. Chem., 62
(1958) 394. ρ density of the rubber
S function defined by Equation 10 used for
calculating the state of cure of the rubber
T temperature (K)
TNn new temperature after elapse of time ∆t at
position n
x longitudinal abscissa through the thickness of
the rubber
Z value of the torque as a fraction of its maximum
value in Equation 3

98 Polymers & Polymer Composites, Vol. 12, No. 1, 2004