Polymer Testing 102 (2021) 107310

Contents lists available at ScienceDirect

Polymer Testing
journal homepage: www.elsevier.com/locate/polytest

Experimental investigation for modeling the hardening of thermosetting

polymers during curing
Bilen Emek Abali a ,∗, Mert Yücel Yardımcı c,d , Michele Zecchini b , Gilda Daissè b ,
Flávio H. Marchesini e , Geert De Schutter c , Roman Wan-Wendner b,c
a
Uppsala University, Division of Applied Mechanics, Department of Materials Science and Engineering, Box 534, 751 21 Uppsala, Sweden
b University of Natural Resources and Life Sciences, Vienna, Christian Doppler Laboratory LiCRoFast, Peter-Jordan-Straße 82, 1190 Vienna, Austria
c Ghent University, Department of Structural Engineering and Building Materials, Magnel–Vandepitte Laboratory, Technologiepark-Zwijnaarde

60, Zwijnaarde 9052 Ghent, Belgium

d
Istanbul Okan University, Department of Civil Engineering, Tuzla, 34959, Istanbul, Turkey
e
Ghent University, Department of Materials, Textiles and Chemical Engineering, Technologiepark-Zwijnaarde 130, Zwijnaarde 9052 Ghent, Belgium

ARTICLE INFO ABSTRACT

Keywords: During curing of thermosetting polymers, crosslinking results in hardening or stiffening of the material. In
Thermosetting polymers electronics, for example in encapsulating integrated circuits (die bonding), thermosets are fully cured in a
Curing controlled environment (under UV-light or within a thermal oven) such that the highest stiffness possible
Rheometer
has been achieved. In building materials, specifically in thermosets used in fastening systems (adhesive
Inverse analysis
anchoring), hardening occurs at environmental temperature. Daily temperature variations alter the curing
process and possibly lead to a lower stiffness. We demonstrate a modeling approach for the mechanical
response dependency on the degree of cure by means of rheometer measurements under a specific temperature
profile. Precisely, we perform oscillatory rheometric tests and convert the storage and loss moduli to material
parameters depending on the degree of cure. Moreover, the temperature dependency as well as chemical
shrinkage have been determined by the same experimental protocol. The presented approach has been applied
to a commercially available (epoxy) thermoset used as an adhesive. We have observed a hardening after a
gelation point of 0.7 and an adequate fit for mechanical response by polynomial functions of degree four.

1. Introduction Curing is modeled by a scalar quantity called degree of cure, which

is normalized to the maximum possible curing state. Mechanical re-
Description of a rheological behavior in polymer systems is chal- sponse depends on this degree of cure. One particular application in
lenging because of its intrinsic non-linear nature [1–4]. One class of structural engineering relies on the epoxy based adhesives used in
polymers, called thermosetting polymers, starts an irreversible curing fastening systems. A post-installed rod is anchored to the hardened
process under an activation energy supplied from the environment. concrete by drilling a hole and attaching with the adhesive [5–7].
This energy is in form of UV-light or temperature. Especially for two The curing process starts immediately after filling the hole by the two
components—one is the base material and the other is the hardener— component mixture and it continues at the ambient temperature. Cure
the chemical reaction starts after mixing the components, where in- kinetics and mechanical response depending on the degree of cure is of
dividual chains form covalent bonds by using subgroups leading to interest in thermosetting polymers also in processing such as additive
crosslinking. This process is exothermal and often autocatalytic, so once manufacturing [8,9]. We stress that the curing process is irreversible
the reaction starts, it continues effected by the energy produced by and the degree of cure varies between 0 and 1. The cure kinetics is
its own exothermal reaction. Generally speaking, the curing process well studied for different systems with differential scanning calorimetry
increases the stiffness of the material, thus, the mixture as a fluid (DSC), for example we refer to [10,11] for phenomenological modeling
material starts behaving as a solid body. The transition from fluid to and [12–14] for formal analysis. In this work, we highlight the relation
solid is called the gelation point. We aim at studying the stiffening of the mechanical response with the degree of cure. Mechanical dy-
process from the fluid to the solid behavior. namical experiments exist in the literature [15–17], they often report

∗ Corresponding author.
E-mail address: [email protected] (B.E. Abali).

https://doi.org/10.1016/j.polymertesting.2021.107310
Received 6 March 2021; Received in revised form 22 July 2021; Accepted 1 August 2021
Available online 28 August 2021
0142-9418/© 2021 The Authors. Published by Elsevier Ltd. This is an open access article under the CC BY license (http://creativecommons.org/licenses/by/4.0/).
B.E. Abali et al. Polymer Testing 102 (2021) 107310

the results without obtaining a material equation. We briefly explain 3. Experiments in rheometer and calorimeter
the model used in this work and simulate the experimental process in
order to obtain the degree of cure during experimental characterization. All measurements were performed at the Magnel–Vandepitte Lab
The main purpose is a methodology for combining mechanical response at Ghent University (Belgium) on an Anton Paar Modular Compact
dependence on the degree of cure. Rheometer MCR 102 and its software running on a Windows 10 com-
A rotational viscometer (rheometer) deforms the specimen in tor- puter. Disposable aluminum upper plates were used on a flat bottom
sion, which is simple shear, monotonously or under a cyclic motion. surface that is covered by an aluminum tape. A photo depicting the
The cyclic or so-called oscillatory measurement is harmonic that is configuration is seen in Fig. 1. The bottom plate was fixed and the
optimized for linear rheology with outputs of so-called storage and upper plate was driven by a given sinusoidal function steering the mo-
loss moduli, responsible for reversible (elastic) and irreversible (dissi- tion. All measurements were displacement controlled in order to make
pative, viscous) processes, respectively. Rheological analyses are well- sure that the deformation was within the small displacement regime,
established for a variety of materials, from cellulose [18] to cement thus, elastic. At a constant frequency of 1 Hz and 0.1% amplitude,
paste [19–21]. Extension to nonlinear models is possible as well, we shear strain was given by an internal feedback controller regulating
refer to [22,23] in the case of a rheometer, [24,25] for postprocess- the torque for the torsional motion leading to simple shear. By using
ing, [26,27] for modeling approaches. D-PP25 upper plate complied with the ISO6721-10, we measured the
In order to use the experimental results in simulations, we need an storage and loss moduli by using the conversion factors, which utilize
adequate material model [28]. Therefore, we propose a methodology a mean shear rate and stress as described in DIN 53018. The diameter
by presenting experimental results and explain a possible connection by of the upper plates was 25.000 mm regarding the datasheet, i.e. with
introducing a scalar function for the dependency between mechanical 0.1 μm dimensional accuracy, and the specimens were trimmed in all
response and the degree of cure. We model the experimentally observed experiments by the same person in order to increase the repeatability.
behavior accurately by determining all parameters by means of a non- The initial height was set to 1 mm as also suggested in standards. The
linear regression. This simple yet effective methodology can be used for volume of the specimen was depending on the height, which did change
all chemically hardening substances like thermosets. We demonstrate during crosslinking known as the chemical shrinkage. Normal force
the approach by applying it to a commercially available epoxy based and current thickness were recorded as well. Precise active temperature
thermosetting polymer used as an adhesive in anchors. steering was acquired with installed Peltier plates (air ventilation). At
the bottom plate the temperature was measured and assumed to be con-
2. Mechanical response stant through the 1 mm thickness. Prior to starting each rheological test,
the temperature of the bottom plate was reduced to 10 ◦ C and then the
In linear rheology, the stress-strain relation is modeled by using material to be tested was put on the bottom plate. After arranging the
simple rheological models composed of springs and dashpots in series 1 mm material thickness (corresponding to the gap distance between
and parallel. In the case of rheometric experiments, shear stress and the parallel plates) the rheological protocol was initiated. During the
shear strain are modeled by this mechanical model. Especially for solid rheological test:
bodies showing a creep behavior, which is the case for the examined
thermosetting polymer, the standard linear solid model, also called • step I, temperature increase from 10 ◦ C to 85 ◦ C in 10 min;
Zener model, is used between the shear stress component, 𝜎, and shear • step II, constant temperature at 85 ◦ C for 50 min;
strain component, 𝜀, as follows: • and step III, temperature decrease back to 10 ◦ C in 10 min.

𝜎 + 𝐾1 𝜎 • = 𝐾2 𝜀 + 𝐾3 𝜀• , (1) The outcome of the measurement, which is used in the inverse analysis,
is visualized in Fig. 2. The protocol was selected for the following
under the assumption that the same relation holds for all components reason. First, in step I, a sufficiently quick temperature increase was
of stress and strain (co-linearity). The coefficients for a spring under realized in order to reach 85 ◦ C being slightly higher than the glass
shear, 𝐺1 , and dashpot, 𝜂1 , in parallel connected to a spring, 𝐺0 , in transition temperature at the fully cured state. This choice ensured the
series read complete curing without vitrification. Moreover, the increase was quick
𝜂1 𝐺0 𝐺1 𝐺0 𝜂1 enough that the gelation point was not reached during this heating
𝐾1 = , 𝐾2 = , 𝐾3 = . (2)
𝐺0 + 𝐺1 𝐺0 + 𝐺1 𝐺0 + 𝐺1 phase. We stress that the 1 mm thickness of the specimen is key for
To be determined material constants are linear springs denoting mod- enabling this relatively fast heating process. Second, in step II, there
uli, 𝐺𝑖 , and a linear dashpot with the viscosity, 𝜂1 . Under the constant was a constant (in time) temperature state designed for isothermal
mechanical loading with constant (in time) stress, 𝜎0 , the solution of curing. We emphasize that the chemical reaction had been exothermal
this ordinary differential equation in time reads such that the active temperature control was of importance. The active
( temperature control is a standard functionality in all Peltier augmented
( 𝐺 ))
1 1 bottom plate rheometers. The step II was 50 min long, which was indeed
𝜀(𝑡) = 𝐽 (𝑡)𝜎0 , 𝐽 (𝑡) = + 1 − exp − 1 𝑡 . (3)
𝐺0 𝐺1 𝜂1 adequate for the specific choice of the specimen material. Third, step III
The creep response, 𝐽 (𝑡), is to be determined by experiments (not was a cooling down process for measuring the temperature dependency
studied herein). The modulus, 𝐺1 , in Pa and viscosity, 𝜂1 in Pa s are also to be analyzed in the next section.
converted to a creep time, 𝜏1 = 𝜂1 ∕𝐺1 , which indicates the response We observed that the suggested temperature variation was reason-
time of the viscous relation. Depending on the underlying material, able with the air ventilation Peltier plates. Only at the change from
for example different chain lengths of the thermosets, we expect that the linear increase to the steady state, a slight deviation was recorded
different response times affect the system response. In this case, the right after 10 min corrected by the feedback controller to the set value.
generalized Zener model reads The disposable upper plate was necessary since the adhesive attained a
( ( )) high stiffness making it otherwise challenging to peal off the adhering
1 ∑ 1 𝑡 𝜂
𝐽 (𝑡) = + 1 − exp − , 𝜏𝑖 = 𝑖 . (4) surface. Thus, the upper plate itself was often deformed by getting off
𝐺0 𝑖
𝐺𝑖 𝜏 𝑖 𝐺 𝑖
the specimen. For each experiment, a new upper plate was used. As
All moduli terms, 𝐺𝑖 , incorporate a reversible (elastic) response and 𝜂𝑖 the material, a commercial product was used—epoxy based adhesive
(correspondingly 𝜏𝑖 ) an irreversible (dissipative, viscous) response. The thermosetting polymer for fastening applications. This product comes
only assumption is that the material behaves linearly, for a nonlinear in a cartridge with a static mixer ensuring the same mixing ratio of
response and its modeling, we refer to [29, Sect. 9] for a detailed the base and hardener. For each specimen, a new cartridge was used.
investigation regarding oscillatory measurements. The time between mixing at room temperature until the start of the

2
B.E. Abali et al. Polymer Testing 102 (2021) 107310

Fig. 3. DSC isothermal measurement results at 30, 40, 60, and 80 ◦ C after normalizing,
time, 𝑡, in hour as 𝑥-axis and degree of cure, 𝜔, as 𝑦-axis.

curing reaction. Specifically, a Netzsch DSC F3 Maia (Netzsch Geraete-

bau GmbH, Germany) was utilized for all experiments with its software.
The temperature was controlled, so that isothermal measurements for
different temperature values of 30, 40, 60, and 80 ◦ C were applied. In
each experiment, the heat flux was recorded that was necessary to hold
the temperature constant during the autocatalytic chemical process.
Around 10 mg of the mixed, fresh material was used in a cell that
contained both the sample and a reference. Mass of specimen was
measured prior to measurement and assumed to remain constant—
we ignored possible gases leaving the specimen during the chemical
process. The highest temperature of 80 ◦ C was adequate for obtaining
a fully cured state. This statement was examined by an additional
dynamic test from 20 ◦ C to 100 ◦ C. In this way, we observed by the
recorded heat flux that the isothermal experiment at 80 ◦ C did result
in a complete curing. The enthalpy was calculated based on this ex-
periment with the complete curing. By using the enthalpy of the fully
cured measurement, we normalized all experiments as shown in Fig. 3.
By using additional experiments, the glass transition temperature for
different degrees of cure was obtained and used for modeling the curing
phenomenon in the following section. Effected by the glass transition
Fig. 1. Experimental setup, top: equipment, bottom left: parallel plate-plate with 1 mm temperature dependency on the degree of cure, lower temperature
thick specimen, bottom right: after the experiment, hardened specimen of 25 mm isothermal experiments failed to achieve a fully cured state, known as
diameter taken out of the upper plate, put behind the specimen. vitrification.

4. Analysis of results
experiment was less than a minute, so we presumed that the curing is
insignificant in this short period (regarding the time until the gelation In order to examine the repeatability of the procedure, we demon-
point). In order to minimize the air trapping within the specimen, the strate three subsequent experiments where the storage and loss moduli
mixture was put into the middle of the bottom plate and squeezed by are obtained directly as results shown in Fig. 4(a) and (b). Obviously,
at first sight, the repeatability was missing. Especially the final stiffness
the upper plate to the set 1 mm thickness. No preforming was used.
values represented by the storage values after 60 min were different.
Differential Scanning Calorimetry (DSC) measurements were exe- Approximately after 17 minutes, we observed the so-called gelation
cuted in Tulln (Austria) for measuring the heat produced during the point, 𝜔g , where the curing surpasses to a solid state with an increased

Fig. 2. Measurables 𝐺′ (left) and 𝐺′′ (right) are shown over time in min. On the right axes, the controlled temperature is plotted indicating from 10 ◦ C to 85 ◦ C in 10 min, holding
for 50 min, cooling down to 20 ◦ C in 10 min.

3
B.E. Abali et al. Polymer Testing 102 (2021) 107310

crosslinking density rate such that the hardening process was acceler- 5. Modeling the curing process
ated. As this gelation point was visible around the same time for all
repetitions, we examined the hardening behavior by normalizing the An exothermal chemical process is of autocatalytic character for
values by their end values. As seen in Fig. 4(c) and (d), the normalized the curing of thermosets [30–33]; therefore, once the process starts,
results show an adequate consistency. Thus, we conclude that the it evolves until the fully cured state has been reached. However, the
hardening procedure is the same; however, the overall stiffness differs. reaction speed may be so slow that the process is practically observed
Under taking off the specimen by bending and peeling off without as stopped—vitrification. As the energy threshold for this phenomenon,
a so-called glass transition temperature, 𝑇g , is introduced. The glassy
damage, we compared their porosity values because of trapped air
regime has its name because viscous fluid (here resin) shows a power
(bubbles) in Fig. 5 by a visual inspection. Although the preparation was
law increase in viscous characteristics [34] as temperature decreases
done by the same person, the porosity was different and one of three
below the glass transition temperature. Thus, below and above 𝑇g ,
specimens was ‘‘bubble-free.’’ The air trapping into the specimen is
the differing curing rates are taken into account, we refer to phe-
expected in real life conditions as well and we report that the stiffness is
nomenological observations [35] and models based on the free volume
depending on this significantly, at least for thin structures as herein. A
reduction [36]. We characterize glassy and rubbery states as occurring
volumetric ratio like porosity seems to be inadequate for this cylindrical
simultaneously with their individual autocatalytic reactions involv-
geometry, where the stress distribution fails to be constant along the ing a diffusion controlled mechanism, 𝐾diff , as well as a chemically
radial direction. We circumvent ourselves from a relation between steered mechanism, 𝐾chem . This model has been introduced in [37] and
porosity and stiffness in general owing to the lack of the statistical amended in [38]. Effected by the epoxy based mixture used herein, we
relevance. We emphasize that each test is more than one hour using model primary and secondary amines via autocatalysis and impurity
disposable upper plate and a new cartridge. catalysis (denoted by 𝑐 in index) by the following evolution equation:
The storage modulus, 𝐺′ , is responsible for the reversible deforma- ( )
𝐾 1 1 1 1 1 1
tion. Analogously, the loss modulus, 𝐺′′ , is steered by the irreversible 𝜔 = 𝐾1 (1 − 𝜔)2 𝜔 + 1𝑐 ,
•
= + , = + ,
′ and loss modulus
𝐾1 𝐾1 𝐾1,chem 𝐾diff 𝐾1𝑐 𝐾1𝑐,chem 𝐾diff
response. By normalizing the storage modulus by 𝐺m ( ) ( )
′′ , it is obvious that the underlying experiment may be used to
𝐸3 𝐸4
by 𝐺m 𝐾1,chem = 𝐴3 exp − , 𝐾1𝑐,chem = 𝐴4 exp − ,
𝑅𝑇 𝑅𝑇
establish the hardening relation with respect to the degree of cure. ( )
𝐶1 (𝑇 − 𝑇g )
Therefore, we propose to decompose the mechanical parameters, 𝐸𝑖 𝐾diff = 𝑘𝑇g exp .
𝐶2 + |𝑇 − 𝑇g |
and 𝜂𝑖 that depend on the temperature, 𝑇 , and degree of cure, 𝜔, in
a multiplicative way (7)
Modeled by the Williams–Landel–Ferry (WLF) equation in the rubbery
𝐸𝑖 (𝑇 , 𝜔) = 𝐸̄ 𝑖 𝜃𝐸 (𝑇 )𝜙𝐸 (𝜔) , 𝜂𝑖 (𝑇 , 𝜔) = 𝜂̄𝑖 𝜃𝜂 (𝑇 )𝜙𝜂 (𝜔) , (5)
regime, the diffusion rate, 𝐾diff , is several orders of magnitude greater
with the amplitudes, 𝐸̄ 𝑖 , 𝜂̄𝑖 , to be determined by additional (creep) than the chemical rate, 𝐾chem . As aforementioned, in the glassy state,
tests. The proposed decomposition has the formal benefit of combining diffusion rate is nearly zero compared to the chemical rate. Therefore,
different measurements: although the curing process is evolving, we measure no change in the
degree of curing known as vitrification. Under these conditions, instead
• one mechanical (creep) test for determining the amplitudes, 𝐸̄ 𝑖 , of hours, the reaction needs years to be completed. Modeling 𝑇g needs
𝜂̄𝑖 , at the reference temperature, 𝜃𝐸 = 𝜃𝜂 = 1, and the fully cured to be established by additional calorimetric measurements by applying
state, 𝜙𝐸 = 𝜙𝜂 = 1 (not demonstrated herein), a fit function [38–41], as follows:
(
• the presented rheometer test during the curing process in order (1 − 𝜔) ln(𝑇g,0 ) + 𝛥𝐶𝜔 ln(𝑇g,∞ ) )
𝑇g = exp . (8)
to acquire, 𝜙𝐸 , 𝜙𝜂 , at an isothermal state (1 − 𝜔) + 𝛥𝐶𝜔
• and another thermal test for detecting the temperature depen- For determining all the parameters 𝑇g,0 , 𝑇g,∞ , 𝛥𝐶 as well as 𝐶1 , 𝐶2 ,
dency by fitting 𝜃𝐸 (𝑇 ), 𝜃𝜂 (𝑇 ) at the fully cured state, 𝜙𝐸 = 𝜙𝜂 = 𝑘𝑇g , 𝐴3 , 𝐴4 , 𝐸3 , 𝐸4 from calorimetry experiments, there exists differ-
1. ent proposals in the literature [42,43] based on simplifications that
parameters may be obtained separately. Instead of following these
As we need a normalized 𝜙𝐸 , 𝜙𝜂 in order to obtain 𝜙𝐸 (𝜔 = 1) = 𝜙𝜂 (𝜔 =
protocols, we solve the inverse analysis directly by a method developed
1) = 1, it is obvious for the linear viscoelastic material response to
in Python language. In this way, the proposed method is also applicable
propose
for different evolution equations than given by Eq. (7). Technically
𝐺′ 𝐺′′ speaking, the degree of cure is used for computational models [44–
𝜙𝐸 = ′
, 𝜙𝜂 = ′′ . (6)
𝐺m 𝐺m 47] in a coupled manner. However, in the experimental setting, as
By using the proposed multiplicative decomposition leading to a sepa- the temperature distribution is held constant within the specimen, we
decouple the problem and reduce it to an ordinary differential equation.
ration of concerns, the amount of experiments are reduced drastically.
This equation for the evolution of degree of cure is solved numerically
The sole assumption for this decomposition relies on the same depen-
by SciPy packages in Python, specifically, the fourth-order Runge–Kutta
dency for all curing states. From a physical chemistry point of view, this
is used with a tolerance value of 10−6 in each iteration. By using an
assumption is often taken for granted. As we are interested above the
optimization problem based on minimization of squared error by means
gelation point, 𝜔g , it is admissible to claim that the crosslinking above
of trust region reflective algorithm augmented by a Cauchy loss, we
the gelation point hardens the material without significantly affecting
acquire parameters as compiled in Table 1 and the agreement with
its material response. In other words, a viscoelastic response above the the experimental results are depicted in Fig. 7. For details of the exact
gelation point fails to switch to a brittle material response or vice versa. procedure and DSC curves, we refer to [48].
We aim for determining the hardening relations, 𝜙𝐸 (𝜔), 𝜙𝜂 (𝜔), We obtain the degree of cure, 𝜔, over time as visualized in Fig. 8.
uniquely. In this regard the temperature rate is assumed to be insignif- By using the degree of cure simulation for different protocols with two
icant. This claim was examined by additional experiments as seen in distinct temperature rates, all experimental results are compared by
Fig. 6. For different temperature rates, the response was qualitatively using Eq. (6). This comparison is of interest in terms of repeatability
the same. After converting the results by means of the degree of and to be depicted in Fig. 9. Around 𝜔g = 0.7 we observe the gelation
cure, in the forthcoming section, we will observe that the results were point and around 𝜔 = 0.85 we observe that half of the stiffness has
quantitatively the same as well. been reached. Different cartridges and tests under different temperature

4
B.E. Abali et al. Polymer Testing 102 (2021) 107310

Fig. 4. Three subsequent measurements by the same protocol with storage modulus, 𝐺′ in pascal, (a) and loss modulus, 𝐺′′ in pascal, (b) over time, 𝑡 in minute, during hardening.
By maximum moduli, 𝐺m ′ , 𝐺′′ , normalized results, (c) and (d), showing the consistency of the hardening over time and pointing out the deviation of the final stiffness but not the
m
hardening procedure.

Fig. 5. Specimens of 1 mm thicknesses after the experiment, left: no visible bubbles, middle: two visible bubbles, right: three large and three small bubbles are visible.

Fig. 6. Measurements with the aforementioned temperature rate (blue) and half of this value (brown) leading to the same value (both cases are ‘‘bubble-free’’), time 𝑡 in minute
on 𝑥-axis and storage, loss moduli, 𝐺′ , 𝐺′′ on 𝑦-axes.

rates are reporting an adequate repeatability that is demonstrating the specimen, 𝑡. Curing induced volumetric change is called chemical
the feasibility of the assumptions used herein. Hence, we conclude shrinkage. By using the protocol as above, we isolate the chemi-
that 𝜙𝐸 and 𝜙𝜂 functions are admissible in order to scale elastic and cal shrinkage from the volumetric expansion due to the temperature
viscous parameters by means of the degree of cure. By using the cooling change, since curing is obtained by a constant temperature. The gap
after reaching the fully cured state, as seen in Fig. 2, we obtain the between the upper and lower plates is initially 1 mm and is mea-
temperature dependence, 𝜃𝐸 and 𝜃𝜂 , as shown in Fig. 10. sured with 1 μm accuracy throughout the hardening. Hence, the strain
Furthermore, we obtain the volumetric shrinkage during curing, is determined by measuring the thickness, 𝑡, leading to a shrinkage
roughly after the gelation point, by measuring the current thickness of parameter for modeling the behavior depicted in Fig. 11.

5
B.E. Abali et al. Polymer Testing 102 (2021) 107310

Table 1 Table 2
Determined material parameters for the curing model, Determined material parameters for cure and temperature dependencies in Eqs. (9),
for details, see material B in [48]. (11), (12).
Variable Value Unit 𝜙𝐸 (𝜔) 𝜙𝜂 (𝜔) 𝜃𝐸 (𝑇 ) 𝜃𝜂 (𝑇 )
𝑇g,0 = 9 ◦
C 𝐴0 = −59.70 𝐵0 = 73.77 𝐴̄ 0 = 646.00 𝐵̄ 0 = 1265.98
𝑇g,∞ = 83 ◦C 𝐴1 = 123.16 𝐵1 = −179.69 𝐴̄ 1 = −2647.25 𝐵̄ 1 = −5313.72
𝛥𝐶 = 0.332 – 𝐴2 = 2.99 𝐵2 = 24.62 𝐴̄ 2 = 4054.93 𝐵̄ 2 = 8363.15
𝐴3 = −136.03 𝐵3 = 200.35 𝐴̄ 3 = −2735.37 𝐵̄ 3 = −5848.28
𝐴3 = 96 748 1/s
𝐴4 = 70.58 𝐵4 = −118.05 𝐴̄ 4 = 682.76 𝐵̄ 4 = 1533.86
𝐴4 = 11 012 1/s
𝜔g = 0.80
𝐸3 = 46 440 G
𝐸4 = 49 462 G 𝜔ref = 0.64 𝛽 = −5.25 × 10−2
𝑘𝑇 𝑔 = 10.3 G
𝐶1 = 17.3 –
𝐶2 = 12.3 –
Unknowns, 𝒖 = { 𝜔g , 𝐴0 , 𝐴1 , 𝐴2 , 𝐴3 , 𝐴4 , 𝐵0 , 𝐵1 , 𝐵2 , 𝐵3 , 𝐵4 }, are
determined by an inverse analysis, again by using Scipy packages and
implementing an error minimization algorithm based on [49, Sect. 4.4].
This least squared error, which is the norm of the difference between
fit functions, 𝜙̂ × and experimental results 𝜙× , is minimized
( )
𝒖 = arg min ‖𝜙𝐸 − 𝜙̂ 𝐸 ‖2 + ‖𝜙𝜂 − 𝜙̂ 𝜂 ‖2 . (10)
w.r.t. 𝒖

Since the chosen fit functions result in a nonlinear minimization prob-

lem with many solutions (local minima), we use the Levenberg–
Marquardt approach with a stochastic initial guess within [−10, 10] for
𝐴× , 𝐵× and [0, 1] for 𝜔g running several times. Additionally, we utilize a
penalty term for fulfilling results between [0, 1] as well as zero at 𝜔g and
one at 𝜔 = 1. Analogously, the thermal dependency is approximated by
a polynomial fit as follows:
𝜃̂𝐸 (𝑇 ) = 𝐴̄ 0 + 𝐴̄ 1 𝑇 + 𝐴̄ 2 𝑇 2 + 𝐴̄ 3 𝑇 3 + 𝐴̄ 4 𝑇 4 ,
(11)
𝜃̂𝜂 (𝑇 ) = 𝐵̄ 0 + 𝐵̄ 1 𝑇 + 𝐵̄ 2 𝑇 2 + 𝐵̄ 3 𝑇 3 + 𝐵̄ 4 𝑇 4 .
Fig. 7. Validation of inverse analysis results in modeling the curing phenomenon
(continuous line) by means of the isothermal DSC experiments (dots), time, 𝑡 in minute, Polynomial degree of four is often used in standards. For the chemical
on 𝑥-axis and degree of cure, 𝜔, on 𝑦-axis.. shrinkage, we use the simple linear relation for an isotropic material
with the strain depending only on the degree of cure, as follows:

𝜺sh = (𝜔 − 𝜔ref )𝛽𝑰 , (12)

with the Kronecker delta (identity), 𝑰. By referring to Fig. 11, the linear
relationship is justified. Obviously, the strain is small such that we
relate the thickness, 𝑡, and the initial thickness, 𝑡0 = 1 mm, for obtaining
the linear strain measure
𝑡 − 𝑡0
= (𝜔 − 𝜔ref )𝛽 . (13)
𝑡0
By using the aforementioned optimization algorithm for this regression
problem, we obtain unknowns 𝛽 and 𝜔ref .

7. Results

The proposed method uses the decomposition of hardening related

terms and thermal dependency as in Eq. (5). The inverse analysis results
Fig. 8. Experimental temperature regulation (top) with temperature, 𝑇 in degree
Celsius, on 𝑦-axis given in time, 𝑡 in minute, on 𝑥-axis. Expected curing evolution in the coefficients as compiled in Table 2. The inverse analysis is used
(bottom) given by degree of cure, 𝜔, over time, 𝑡. by means of the curing model such that Table 1 is compulsory for the
accuracy of utilizing Eqs. (9), (11), (12). The good agreement with the
experimental results are depicted in Figs. 12, 13 showing that the error
6. Inverse analysis for experiments minimization has been working well. Both results in Figs. 12 and 13
are obtained independently, we basically assume that 𝜔g and 𝜔ref may
be different. Shrinkage during curing is dominated by the crosslinking
Using the experimental results 𝐺′ and 𝐺′′ as converted into 𝜙𝐸 and
phenomena that is accelerated beyond the gelation point. Therefore,
𝜙𝜂 as shown in Fig. 9, there are five experimental data sets with each
we expect 𝜔g and 𝜔ref have similar values, as observed in Table 2.
3600 data points corresponding to one hour experiments with a 1 Hz
The fit in Fig. 12 is adequate owing to the polynomial function of
data acquisition rate. All of the data has been used to fit a polynomial
degree four. For the fit in degree of cure dependency, using such a
curve of degree four above the gelation point, 𝜔g , as follows:
high order polynomial is essential in order to obtain the asymptotic
{
0 , if 𝜔 < 𝜔g approach to unity, 𝜙× → 1, as degree of cure converges to the fully
𝜙̂ 𝐸 (𝜔) = , cured state, 𝜔 → 1.
𝐴0 + 𝐴1 𝜔 + 𝐴2 𝜔2 + 𝐴3 𝜔3 + 𝐴4 𝜔4 , else
{ (9) Rheometer measurements by using the proposed protocol have been
0 , if 𝜔 < 𝜔g fruitful for obtaining the stiffness dependence on temperature and de-
𝜙̂ 𝜂 (𝜔) = .
𝐵0 + 𝐵1 𝜔 + 𝐵2 𝜔2 + 𝐵3 𝜔3 + 𝐵4 𝜔4 , else gree of cure; as well as the chemical shrinkage. The thermal expansion

6
B.E. Abali et al. Polymer Testing 102 (2021) 107310

Fig. 9. Experimental mechanical properties to degree of cure relation, 𝜙𝐸 and 𝜙𝜂 , obtained from 𝐺′ and 𝐺′′ values in time, on 𝑦-axis versus degree of cure, 𝜔, on 𝑥-axis.

Fig. 10. Experimental 𝜃𝐸 and 𝜃𝜂 results, obtained from the cooling part in 𝐺′ and 𝐺′′ values in time, on 𝑦-axis versus degree of cure, 𝜔, on 𝑥-axis.

model known in rheology and investigate moduli and viscosities de-

pendency on the degree of cure. Curing kinetics for such systems are
well-known and we have used an evolution equation incorporating
the glass transition temperature depending on the degree of cure.
Parameters for this model are obtained by following standards with
isothermal and dynamic calorimetry measurements. The work proposes
a methodology based on a specifically chosen temperature profile
during rheometer measurements. Degree of cure is simulated by the
evolution equation and the methodology allows to obtain dependency
of moduli and viscosities on the degree of cure. More specifically, we
have used a temperature protocol with steps I, II, and III that leads to
an isothermal curing process right before gelation point up to the fully
cured state, which is then followed by a temperature decrease. This
simple experimental setting allows obtaining enough information for
setting up an inverse analysis and determine fit functions describing the
Fig. 11. Experimental thickness change, 𝑡 in millimeter, on 𝑦-axis with respect to the complete dependency of moduli and viscosities on the degree of cure.
degree of cure, 𝜔, on 𝑥-axis. Thickness change is related to the chemical shrinkage Moreover, the final step allows to obtain the temperature dependency
effected by the curing, starting from a reference degree of cure around the gelation for an interval as well. Since the rheometer records the volume change,
point until the fully cured state (measured under constant temperature).
additionally, we have modeled the chemical shrinkage from the same
set of experiments. The benefit of the proposed strategy is to use a
standard rheometer with an active temperature control. No tweaks or
is rather small such that for this material the accuracy of the current special feedback controller tuning is necessary. The raw data is directly
gap measurement is not sufficient, although the protocol allows this used in the inverse analysis. The same procedure may be used for other
in step III by a temperature reduction after the fully cured state has thermosets, especially in applications, where a complete curing is not
been attained such that thermal expansion is isolated from the chemical expected, the mechanical response dependence on the degree of cure is
shrinkage. of paramount importance.

8. Conclusion
Acknowledgment
Thermosetting polymers harden and shrink during curing. We have
investigated this phenomenon and modeled by using material equations This research was funded by the Christian Doppler Forschungsge-
with parameters acquired by an inverse analysis with the aid of open sellschaft, Austria, Grant LICROFAST to Roman Wan-Wendner. The
source packages from SciPy in the Python language. The presented financial support by the Austrian Federal Ministry for Digital and Eco-
approach may be adequate for several thermosetting polymers, herein, nomic Affairs and the National Foundation for Research, Technology,
we have analyzed an epoxy based commercial product with supe- and Development is acknowledged, as is the partial financial support by
rior mechanical properties used as an adhesive. As the mechanical the Austrian Science Fund (FWF) under Grant P 31203-N32. The finan-
properties depend on the degree of cure, we give a linear standard cial support by the Special Research Fund (BOF) of Ghent University,

7
B.E. Abali et al. Polymer Testing 102 (2021) 107310

Fig. 12. Experimental 𝜙𝐸 and 𝜙𝜂 results as data points and the determined fit curves (dashed line) in Eqs. (9), (11).

[7] J. Fischer, P.R. Bradler, D. Schmidtbauer, R.W. Lang, R. Wan-Wendner, Long-

term creep behavior of resin-based polymers in the construction industry, Mater.
Today Commun. 18 (2019) 60–65.
[8] A. Redmann, P. Oehlmann, T. Scheffler, L. Kagermeier, T.A. Osswald, Thermal
curing kinetics optimization of epoxy resin in digital light synthesis, Addit.
Manuf. 32 (2020) 101018.
[9] A. Redmann, T.A. Osswald, A model for modulus development of dual-cure resin
systems, Polym. Eng. Sci. 61 (3) (2021) 830–835.
[10] M. Kiasat, Curing Shrinkage and Residual Stresses in Viscoelastic Thermosetting
Resins and Composites (Ph.D. thesis), Technical University of Delft, 2000.
[11] A. Lion, P. Höfer, On the phenomenological representation of curing phenomena
in continuum mechanics, Arch. Mech. 59 (1) (2007) 59–89.
[12] R. Mahnken, Thermodynamic consistent modeling of polymer curing coupled to
visco–elasticity at large strains, Int. J. Solids Struct. 50 (13) (2013) 2003–2021.
[13] M. Hossain, Modelling the curing process in particle-filled electro-active polymers
with a dispersion anisotropy, Contin. Mech. Thermodyn. 32 (2020) 351–367.
[14] Z. Liao, M. Hossain, X. Yao, M. Mehnert, P. Steinmann, On thermo-viscoelastic
experimental characterization and numerical modelling of VHB polymer, Int. J.
Fig. 13. Experimental results for the chemical shrinkage, 𝜀sh in percent, on 𝑦-axis
Non-Linear Mech. 118 (2020) 103263.
versus degree of cure, 𝜔, on 𝑥-axis, and determined fit curve (dashed line) by using
[15] J.A. Kim, D.G. Seong, T.J. Kang, J.R. Youn, Effects of surface modification on
the model in Eq. (12).
rheological and mechanical properties of CNT/epoxy composites, Carbon 44 (10)
(2006) 1898–1905.
[16] H. Rastin, M.R. Saeb, M. Nonahal, M. Shabanian, H. Vahabi, K. Formela, X.
Belgium, Concerted Research Action (GOA), project BOF18-GOA-009, Gabrion, F. Seidi, P. Zarrintaj, M.G. Sari, Transparent nanocomposite coatings
based on epoxy and layered double hydroxide: Nonisothermal cure kinetics and
is acknowledged.
viscoelastic behavior assessments, Prog. Org. Coat. 113 (2017) 126–135.
[17] A. Zotti, S. Zuppolini, A. Borriello, M. Zarrelli, Thermal and mechanical
References characterization of an aeronautical graded epoxy resin loaded with hybrid
nanoparticles, Nanomaterials 10 (7) (2020) 1388.
[18] B.E. Abali, Inverse analysis of cellulose by using the energy-based method and
[1] G. Marrucci, G. Ianniruberto, et al., Modelling nonlinear polymer rheology is a rotational rheometer, Appl. Sci. 8 (8) (2018) 1354.
still challenging, Korea-Aust. Rheol. J. 17 (3) (2005) 111–116. [19] Y. Qian, S. Ma, S. Kawashima, G. De Schutter, Rheological characterization of the
[2] R. Long, K. Mayumi, C. Creton, T. Narita, C.-Y. Hui, Rheology of a dual viscoelastic solid-like properties of fresh cement pastes with nanoclay addition,
crosslink self-healing gel: Theory and measurement using parallel-plate torsional Theor. Appl. Fract. Mech. 103 (2019) 102262.
rheometry, J. Rheol. 59 (3) (2015) 643–665. [20] X. Dai, S. Aydın, M.Y. Yardımcı, K. Lesage, G. De Schutter, Effects of activator
[3] B.E. Abali, C.-C. Wu, W.H. Müller, An energy-based method to determine mate- properties and GGBFS/FA ratio on the structural build-up and rheology of AAC,
rial constants in nonlinear rheology with applications, Contin. Mech. Thermodyn. Cem. Concr. Res. 138 (2020) 106253.
28 (5) (2016) 1221–1246. [21] X. Dai, S. Aydin, M.Y. Yardımcı, K. Lesage, G. De Schutter, Influence of water to
[4] I. Argatov, A. Iantchenko, V. Kocherbitov, How to define the storage and loss binder ratio on the rheology and structural build-up of alkali-activated slag/fly
moduli for a rheologically nonlinear material? Contin. Mech. Thermodyn. 29 ash mixtures, Constr. Build. Mater. 264 (2020) 120253.
(2017) 1375–1387. [22] R.H. Ewoldt, A. Hosoi, G.H. McKinley, New measures for characterizing nonlinear
[5] G. Singer, G. Sinn, H.C. Lichtenegger, S. Veigel, M. Zecchini, R. Wan-Wendner, viscoelasticity in large amplitude oscillatory shear, J. Rheol. (1978-Present) 52
Evaluation of in-situ shrinkage and expansion properties of polymer composite (6) (2008) 1427–1458.
materials for adhesive anchor systems by a novel approach based on digital [23] R. Ewoldt, P. Winter, J. Maxey, G. McKinley, Large amplitude oscillatory shear
image correlation, Polym. Test. 79 (2019) 106035. of pseudoplastic and elastoviscoplastic materials, Rheol. Acta 49 (2) (2010)
[6] P.R. Bradler, J. Fischer, R. Wan-Wendner, R.W. Lang, Shear test equipment 191–212.
for testing various polymeric materials by using standardized multipurpose [24] M. Wilhelm, Fourier-transform rheology, Macromol. Mater. Eng. 287 (2) (2002)
specimens with minor adaptions, Polym. Test. 75 (2019) 93–98. 83–105.

8
B.E. Abali et al. Polymer Testing 102 (2021) 107310

[25] C. Klein, H. Spiess, A. Calin, C. Balan, M. Wilhelm, Separation of the nonlinear [37] E. Rabinowitch, Collision, co-ordination, diffusion and reaction velocity in
oscillatory response into a superposition of linear, strain hardening, strain condensed systems, Trans. Faraday Soc. 33 (1937) 1225–1233.
softening, and wall slip response, Macromolecules 40 (12) (2007) 4250–4259. [38] C.W. Wise, W.D. Cook, A.A. Goodwin, Chemico-diffusion kinetics of model
[26] B.E. Abali, H. Yang, Magnetorheological elastomer’s material modeling and epoxy-amine resins, Polymer 38 (13) (1997) 3251–3261.
parameter determination by using the energy-based method, in: B.E. Abali, H. [39] E. DiBenedetto, Continuity of weak solutions to a general porous medium
Altenbach, F. dell’Isola, V.A. Eremeyev, A. Öchsner (Eds.), New Achievements in equation, Indiana Univ. Math. J. 32 (1) (1983) 83–118.
Continuum Mechanics and Thermodynamics, Springer Nature, Singapore, 2019, [40] D. Hesekamp, H.C. Broecker, M.H. Pahl, Chemo-rheology of cross-linking poly-
pp. 1–15. mers, Chem. Eng. Technol.: Ind. Chem.-Plant Equip.-Process Eng.-Biotechnol. 21
[27] F.H. Marchesini, R.M. Oliveira, H. Althoff, P.R. de Souza Mendes, Irreversible (2) (1998) 149–153.
time-dependent rheological behavior of cement slurries: Constitutive model and [41] R. Venditti, J. Gillham, A relationship between the glass transition temperature
experiments, J. Rheol. 63 (2) (2019) 247–262. (Tg) and fractional conversion for thermosetting systems, J. Appl. Polym. Sci.
[28] B.E. Abali, J. Vorel, R. Wan-Wendner, Thermo-mechano-chemical modeling and 64 (1) (1997) 3–14.
computation of thermosetting polymers used in post-installed fastening systems [42] M. Jouyandeh, S.M.R. Paran, A. Jannesari, D. Puglia, M.R. Saeb, Protocol for
in concrete structures, Contin. Mech. Thermodyn. (2020) 1–20. nonisothermal cure analysis of thermoset composites, Prog. Org. Coat. 131
[29] B.E. Abali, Thermodynamically Compatible Modeling, Determination of Material (2019) 333–339.
Parameters, and Numerical Analysis of Nonlinear Rheological Materials (Ph.D. [43] M. Nonahal, H. Rastin, M.R. Saeb, M.G. Sari, M.H. Moghadam, P. Zarrintaj, B.
thesis), Technische Universität Berlin, Institute of Mechanics, 2014. Ramezanzadeh, Epoxy/PAMAM dendrimer-modified graphene oxide nanocom-
[30] L. Shechter, J. Wynstra, R.P. Kurkjy, Glycidyl ether reactions with amines, Ind. posite coatings: Nonisothermal cure kinetics study, Prog. Org. Coat. 114 (2018)
Eng. Chem. 48 (1) (1956) 94–97. 233–243.
[31] M. Kamal, S. Sourour, Kinetics and thermal characterization of thermoset cure, [44] M. Hossain, Modelling and Computation of Polymer Curing (Ph.D. thesis),
Polym. Eng. Sci. 13 (1) (1973) 59–64. Friedrich-Alexander-Universität Erlangen-Nürnberg, Germany, 2010.
[32] M.R. Kamal, Thermoset characterization for moldability analysis, Polym. Eng. [45] J. Mergheim, G. Possart, P. Steinmann, Modelling and computation of curing
Sci. 14 (3) (1974) 231–239. and damage of thermosets, Comput. Mater. Sci. 53 (1) (2012) 359–367.
[33] C. Heinrich, M. Aldridge, A.S. Wineman, J. Kieffer, A.M. Waas, K.W. Shahwan, [46] R.A. Sauer, A survey of computational models for adhesion, J. Adhes. 92 (2)
Generation of heat and stress during the cure of polymers used in fiber (2016) 81–120.
composites, Internat. J. Engrg. Sci. (53) (2012) 85–111. [47] R. Landgraf, Modellierung und Simulation der Aushärtung polymerer Werk-
[34] C.A. Angell, Perspective on the glass transition, J. Phys. Chem. Solids 49 (8) stoffe (Ph.D. thesis), TU Dresden, Germany, 2016.
(1988) 863–871. [48] B.E. Abali, M. Zecchini, G. Daissè, I. Czabany, W. Gindl-Altmutter, R. Wan-
[35] C.-S. Chern, G.W. Poehlein, A kinetic model for curing reactions of epoxides with Wendner, Cure kinetics and inverse analysis of epoxy-amine based adhesive used
amines, Polym. Eng. Sci. 27 (11) (1987) 788–795. for fastening systems, Materials 14 (14) (2021) 3853.
[36] H.-J. Flammersheim, J.R. Opfermann, Investigation of epoxide curing reactions [49] B.E. Abali, C. Cakiroglu, Numerische Merhoden Für Ingenieure, in: Lehrbuch,
by differential scanning calorimetry–formal kinetic evaluation, Macromol. Mater. Technik : Technomathematik, Springer Vieweg, 2020.
Eng. 286 (3) (2001) 143–150.