Study of Polyester-Based Coil Coatings

By Using Thermal Analysis Methods

M. Korhonen, P. Starck, and B. Löfgren—Helsinki University of Technology*

P. Mikkilä** and O. Hormi—University of Oulu†

INTRODUCTION

I
n commercial coating applications, thermosets are
rarely used without the incorporation of other mate-
rials, since a successful coating requires properties
which cannot provide any one component alone.
The curing of polyester-based clearcoats has
Usually the coatings consist of catalysts, hardeners, been studied by using thermal analysis meth-
fillers, and additives. Catalysts and accelerators are ods. Polyester resins are used as raw materials
employed to reduce the processing time. Hardeners will
affect not only the reaction rate, but also the ultimate for coil coatings. The aim of this work was to
properties of the thermoset, such as its network struc- study the viscoelastic properties of coil coat-
ture, Tg∝ , and dynamic mechanical properties. Fillers
are used both to provide or enhance physical properties
ings. Differential scanning calorimetry (DSC)
such as modulus, thermal expansion and dimensional was found to be useful in curing studies of lac-
stability, thermal and electrical conductivity, and to
quers, for example, to predict conversion as a
reduce costs. Additives are commonly employed to
enhance the thermal stability or burning characteristics function of cure time or cure temperature.
of materials. Water and solvents that are part of the Dynamic mechanical analysis (DMA) was
process can become involved in the chemistry of curing
and, moreover, plasticize the cured thermoset.1 used to determine the gel times of some
Polyester resins, in combination with amine resins or clearcoats. The DSC measurements were done
as polyurethane base resins, dominate the scene of coil with different apparatus and software pro-
coating paints with a share of about 60%. Polyester coat-
ings with amino resin crosslinkers show a particularly grams. This study concentrates on the use of
well-balanced profile. It is not a star performer in any these modern techniques to predict the curing
particular test, but has no weak points either. In the coil
coating line, the coated strip is heated in an oven by hot and correlation of the results.
air with temperatures in the range of 300 to 500°C. The
dwell time of coil coatings is from 15 to 60 sec and as an
upper limit for metal temperature is 250°C (PMT = peak resistance to building-site damage is obviously
metal temperature), above which organic materials may required. 3
oxidize.2 This kind of manufacturing process requires a Curing of coil coatings, as with coatings in general, is
fast curing system which presents quite a challenge in usually characterized by gelation and vitrification. In
obtaining the right or actually optimum cure profile for the thermosetting process gelation and vitrification take
different applications. When the coating has cured, it place during cure as a consequence of chemical reac-
must be capable of forming around sharp bends, must tions which convert a liquid resin-hardener mixture to a
have excellent exterior durability, and must withstand solid. Gelation corresponds to the incipient formation of
corrosion in marine or industrial environments. In addi- an infinite tridimensional molecule. Vitrification, which
tion, specific chemical resistance may be needed, and usually follows gelation, entails a physical transforma-
tion from the liquid or rubbery state to the glassy state
* Polymer Science Centre, Innopoli B2, P.O. Box 356, FIN-02157 Espoo, Finland as a result of an increase in the average molecular
† Department of Chemistry, P.O. Box 3000, FIN-90014 University of Oulu, Finland
** Author to whom correspondence should be addressed, email: [email protected]. weight and in the crosslink density of the material. At

Vol. 75, No. 937, February 2003 67

M. Korhonen et al.

Table 1—Samples Studied in This Work. Polyesters Used Were S1, S2, and S3, with a The Mettler DSC 20(30) using nitrogen
Polyester/Curing Agent Ratio of 82.5/17.5 and the Catalyst Was p-TSA DSC cell purge was calibrated for tem-
Sample S1 82.5/17.5 pTSA S2 82.5/17.5 pTSA S3 82.5/17.5 pTSA perature using indium and zinc, and for
enthalpy using indium. Measurements
AV solid/mg KOH/g . . . . 6.7 7.1 4.3 were made by using medium pressure
OHV solid/mg KOH/g . . . 46 32 41 steel pans in order to avoid the evapora-
Visc.diluted/P (23°C) . . . . . 18.7 26.9 16.8
Mn g/mol . . . . . . . . . . . . . 3210 3760 3280
tion of solvents and to separate simulta-
Mw g/mol . . . . . . . . . . . . . 9030 9250 7840 neous curing and evaporation reactions.
MEL/OH-ratio . . . . . . . . . . 3.32 4.77 3.72 The curing experiments were performed
Catalyst % on binder . . . 0.32 0.32 0.32 at three different temperatures: 150°C,
Catalyst/HMMM . . . . . . . 0.018 0.018 0.018 160°C and 170°C for 60 min. Samples in
Solid content/ % . . . . . . . 64.2 63.6 63.3
the range 6-20 mg were accurately
weighed into the DSC sample pans
vitrification, the chemical reaction slows down signifi- which were then hermetically sealed. Some of the sam-
cantly. 4 The end-use performance of coil coatings is ples were also measured on Mettler DSC 821e and the
affected both by the coating formulation and the coating results obtained with these two apparatus were com-
process; thus, information on the curing process is pared.
required to ensure that the properties of the material are The degree of curing and the reaction rates were cal-
optimized for a specific application. A number of arti- culated from isothermal DSC curves when using
cles have been published which describe curing more Mettler DSC 20. From the exothermic DSC curves, the
thoroughly and present several thermal analysis tech- degree of curing was calculated by dividing the
niques, differential scanning calorimetry (DSC), enthalpy at time, t, by the total enthalpy of the curing
dynamic mechanical analysis (DMA), torsional braid exotherm. No residual curing exotherm was obtained.
analysis, dielectric analysis, and thermal stimulated The reaction rate was calculated by dividing the peak
current, that have been used in curing studies. 5-17 height at a certain time by the total enthalpy of the cur-
This paper concentrates on utilizing the modern ther- ing exotherm and sample weight.
mal analysis techniques to study the viscoelastic prop- In the DSC measurements done by Mettler 821e the
erties of liquid polyester-based coil coating systems. Model Free Kinetics program proposed by Vyazov-
Modeling and property prediction overall in coatings kin18-20 was used. Vyazovkin based his method in the
formulation is not simple. Today there are several com- Arrhenius equation in which the temperature depend -
mercially available kinetic software applications for cur- ence is assumed to reside in the rate constant. The used
ing studies. In the present work we investigated the model-free kinetics software collects data from multiple
correlation between two different apparatus and soft- experiments, usually three. Instead of a single point of
ware programs. Properties of coil coatings were studied conversion, the approach follows all points of conver-
experimentally by DSC and DMA. sion from the multiple experiments. Based on the
results, predictions can then be made on how the reac-
tion would proceed at various temperatures. Model-free
kinetics program does not presume any information
EXPERIMENTAL about the order of the reaction and thereby permits the
solving of practical problems in the case of complex
Materials
processes.
The samples and some of their physical characteristics Another set of DSC measurements was performed
are presented in Table 1. The table shows the amount of using Perkin-Elmer DSC-7 unit operating a Windows™
polyester and curing agent in the samples. The catalyst
95/NT-based Pyris software. The DSC was calibrated
was p-toluenesulfonic acid (pTSA) and the curing agent for temperature and enthalpic responses using indium
was hexamethoxymethylmelamine (HMMM). The sam- (156.5°C and 28.34 J/g). Six to 15 mg of clearcoat were
ples were analyzed as received from the coating indus-
sealed in a stainless steel high-pressure pan (part num-
try and the exact composition of samples was unknown. ber 0319-0218). These pans eliminated the interfering
The polyesters, S1, S2, and S3, differ in crosslinking den- effects of the heat of vaporization by suppressing the
sity and molar mass. Also, the acid values and hydroxyl
vaporization of solvents or by containing volatile reac-
numbers of the samples differ. The catalyst on binder
and the catalyst on curing agent is the same for each
sample and the solid content is very similar. Table 2—Results of Sample S3 82.5/17.5 pTSA Analyzed by
Scanning Kinetics Program. Reaction Profile (20, 30, 50, 75,
Methods and 90%) as a Function of Time at Isothermal Temperatures
150° and 170°C
DIFFERENTIAL SCANNING C ALORIMETRY (DSC): Curing
of thermosetting polyester clear lacquers has been stud- Isothermal Isothermal
Conversion (%) Temperature 150°C Temperature 170°C
ied by using differential scanning calorimeter methods.
The polyester-amino resin ratio is generally in the range 20 ................ 3.6 min 0.6 min
90:10 to 70:30. In this study we used a ratio of 82.5:17.5. 30 ................ 7.2 min 1.2 min
50 ................ 13.3 min 1.8 min
DSC measurements were done both with Mettler’s and 75 ................ 24.8 min 4.2 min
Perkin-Elmer’s techniques. 90 ................ 43.0 min 6.6 min

68 Journal of Coatings Technology

 Study of Polyester-Based Coil Coatings

tion products. Samples were analyzed in a nitrogen RESULTS AND DISCUSSION

atmosphere and scanned from 30° to 250°C at 10°, 15°,
and 20°C/min. In addition, some clearcoats were run Curing Studies by Using Different Kinetics Programs
isothermally at 160°, 170° and 180°C for 30 min. For
Sample S3 82.5/17.5 pTSA was measured by a different
these isothermal runs, the cell was allowed to stabilize
at chosen isothermal conditions before setting the sam- DSC apparatus and the results were processed by differ-
ent kinetics programs: Mettler DSC 821e Model-Free
ple. The results were calculated by using the kinetic
Kinetics program and Perkin-Elmers’ Scanning Kinetics
software program. To perform the isothermal kinetics
calculations, the software fits three to six data curves program. The results are compared in the following sec-
tion.
that have been taken at a constant temperature, to the
The results obtained by the DSC methods are sum-
Arrhenius relationship and thereby determines the pre-
exponential factor, activation energy, and order of reac- marized in Table 2 and Figure 1. Table 2 shows the
obtained results of sample S3 82.5/17.5 pTSA calculated
tion. With this program, the behavior of the material can
by Perkin-Elmers’ Scanning Kinetics program. Figure 1
be predicted under isothermal conditions.21
presents the results of the same sample analyzed using
Mettlers’ Model-Free Kinetics program. The conversion
DYNAMIC MECHANICAL A NALYSIS (DMA): DMA times can also be seen in Table 5. Scanning Kinetics pro-
experiments were carried out on Perkin-Elmer DMA-7 gram uses a multilinear regression to fit a single data
unit operating a Windows-based Pyris software using curve, which has been taken at constant heating rate, in
liquid nitrogen cooling. The DMA-7 was calibrated for this study 10°C/min, to the Arrhenius relationship and
temperature using indium. Frequency scans were used thereby determine reaction parameters. Model-Free
in order to evaluate gelation. The samples were run Kinetics program collects data from multiple experi-
over a frequency range of 50 to 1 Hz at three different ments, usually three, as was the case in this study: heat-
temperatures: 150°, 160°, and 170°C. The cup-and-plate ing rates were 10°, 15°, and 20°C/min.
geometry was used, since it enables characterization of Good correlation was found at isothermal tempera-
materials from the under-cured state through to the ture 150°C for conversions lower than 75%. At 75 and
fully cured without changing the sample geometry. 90% conversion with Perkin-Elmers’ Scanning Kinetics

Figure 1—Dynamic scans and conversion predictions as a function of time by Mettler Model-Free Kinetics
program for sample 3 82.5/17.5 pTSA.

Vol. 75, No. 937, February 2003 69

M. Korhonen et al.

program gave longer times than Mettlers’. Deviations

observed at isothermal 170°C could be attributed to dif-
ferent heating rates. The Scanning Kinetics program
uses only one data 10°C/min and Model-Free Kinetics
program uses three different heating rates: 10°, 15°, and
20°C/min. Although the sample was the same, the
experiments were done independently at different peri-
ods of time; the sample could be older and not so reac-
tive which could also affect the measurements.
Figures 1 and 2a-c show the curves by kinetics pro-
grams: extent of reaction as a function of cure time
(Figure 2a), extent of reaction as a function of an isother-
mal cure temperature (Figure 2b), cure time as a function
of an isothermal cure temperature (Figure 2c), and con-
version as a function of time (Figure 1).
Based on the dynamic measurements, the conversion
predictions of sample S3 82.5/17.5 pTSA are shown in
(2a)
Figure 1. The figure shows that the higher the tempera-
ture the smaller the curing time. At temperature 230°C
the conversion of 75% is reached in 0.61 min.
Considering that the oven temperature in coil coating
process is more or less 300°C, a curing time less than
one minute should be enough to achieve a well-cured
product.
As can be seen in Figure 2a, predicted conversion
increases with temperature and the reaction is very fast
at the higher temperatures 190°, 210°, and 230°C. It
takes only 10 sec to reach 90% conversion. In the coil
coating process the dwell time is only 15-60 sec and a
fast reaction is essential to reach a desired optimum
degree of cure.
Figure 2b shows the predicted extent of reaction as a
function of an isothermal cure temperature and it can be
seen that cure time curves 0.6, 0.8, and 1 min differ
(2b)
clearly from the other cure time plots. These kinds of
data afford us an opportunity to evaluate the cure pro-
file of a coating and to further develop desired proper-
ties of a final product. Roughly, in this case, 0.6-1 min
curing seems to develop desired structure-property
relations, whereas a shorter cure time seems to give too
low a conversion and degree of cure. Using cure times
higher than 5 min, the temperature used could be much
lower than 200-230°C.
Figure 2c shows predicted cure time as a function of
an isothermal cure temperature. These kinds of data
allows for the evaluation of the optimal baking temper-
ature.

Mettler DSC 20 vs. Mettler DSC 821e

The curing measurements and curing time predic-
tions made with Mettlers’ apparatus were mostly based
(2c)
on the DSC 20, but some measurements were done with
the new Mettler DSC 821e and the results are compared
here. The samples under study were S1 82.5/17.5 pTSA,
Figure 2—Curves obtained by using Perkin- S2 82.5/17.5 pTSA, and S3 82.5/17.5 pTSA. DSC 20
Elmers´ Scanning Kinetics program for sample measurements were isothermal for 60 min at tempera-
S3 82.5/17.5 pTSA: (a) Conversion as a func- tures 150°, 160°, and 170°C. The degree of curing was
tion of cure time; (b) Conversion as a func- calculated from the measurements by dividing the
tion of isothermal cure temperature; and (c) enthalpy at time, t, by the total enthalpy of the curing
Cure time as a function of isothermal cure exotherm. The Mettler DSC 821e measurements were
temperature. done in the scanning mode with three different heating

70 Journal of Coatings Technology

 Study of Polyester-Based Coil Coatings

Table 3—Conversion Times at Different Temperatures for Sample S1 82.5/17.5 pTSA

150°C 160°C 170°C

Conversion (%) DSC 20 DSC 821e DSC 20 DSC 821e DSC 20 DSC 821e

20 .............. 3.45 6.44 3.12 3.68 1.72 2.16

30 ............. 5.92 8.74 5.28 5.05 2.91 2.99
50 ............. 10.45 13.33 7.00 7.84 5.49 4.73
75 ............. 17.34 20.11 15.83 12.12 9.91 7.47
90 ............. 27.36 26.35 24.56 16.11 17.34 10.06

Table 4—Conversion Times at Different Temperatures for Sample S2 82.5/17.5 pTSA

150°C 160°C 170°C

Conversion (%) DSC 20 DSC 821e DSC 20 DSC 821e DSC 20 DSC 821e

20 ............. 6.03 5.43 3.02 3.13 2.80 1.85

30 ............. 8.40 8.11 5.06 4.70 3.88 2.79
50 ............. 15.83 13.94 8.51 8.30 7.76 5.07
75 ............. 27.47 23.71 16.16 14.53 14 9.10
90 ............. 38.24 32.23 23.91 20.26 21.01 13.00

rates: 5°, 10°, and 20°C/min in the temperature range DMA Studies
25-230°C. After these dynamical measurements the In the DMA measurements, frequency scans were
Model-Free Kinetics program simulated the curing used to characterize gelation of a clearcoat, S1 poly-
times at isothermal conditions. Tables 3-5 show the time ester/curing agent ratio 82.5/17.5 with catalyst pTSA,
to reach different conversions at different temperatures at temperatures 150°, 160°, and 170°C. The properties of
by two DSCs for samples S1 82.5/17.5 pTSA, S2 liquid materials can be measured by a cup-and-plate
82.5/17.5 pTSA, and S3 82.5/17.5 pTSA. measuring system where a 13-mm diameter cup was
The values obtained by DSC 20 were based on real mounted at the base and a parallel 10-mm diameter
measurements and the curing times were calculated plate at the top. This measuring system enables charac-
manually. The DSC 821e used dynamic measurements terization of clearcoats from the uncured low-modulus
to simulate and predict the curing times at different state to the fully cured high-modulus state. Earlier we
isothermal temperatures, and the program calculated had succesfully investigated gelation of paint samples
the time to reach a certain conversion. The results by the cup-and-plate system using thermal scans,22
obtained by these two different DSCs are in agreement, whereas in the present work, we used frequency scans,
especially when the conversion is 75% or smaller, and and in this case clearcoat sample was scanned over the
the temperature is above 160°C. Very good similarity is frequency range from 50 to 1 Hz.
seen on S3 in Table 5 at 75% conversion. At 150°C the The principle of DMA is based on a study of the time
curing time by DSC 20 is 25.10 min and by DSC 821e it dependent mechanical behavior of materials.
is 24.87 min. At 160°C the values are 15.30 and 15.66 min
Mechanical properties of materials are determined in
and at 170°C 11.63 and 10.08 min. Some differences may dynamic modes by applying an oscillating force and
be due to aging of sample, incidental failure in DSC- measuring the material’s response. The results are inter-
run, or uncertainty of manual calculation by DSC 20.
preted using fundamental relationships such as
The effect of polyester resin on curing times can also Hooke’s Law (for elastic behavior) and Newton’s Law
be seen in Tables 3-5. In samples S1 and S2 the polyesters (for damping or viscous behavior). These results can be
are slightly branched and S3 is linear. If we concentrate normalized for sample dimensions to obtain standard
on conversions of 50 and 75%, the most reactive poly- reporting values such as storage modulus E′, and loss
ester is the S3 at almost every temperature and with modulus E′′, and tan δ as a function of temperature or
both DSCs. Only when the temperature is 160°C and frequency. This method is highly sensitive in detecting
when using DSC 20 does the S2 have a slightly faster changes in internal molecular mobility. 23-25
curing time than S3. As can be seen from Table 1, the
No particular sample preparation was needed, since
polyester S3 has the lowest molecular weight, which
could be the reason for faster curing times. the clearcoat was used as delivered and the measure-

Table 5—Conversion Times at Different Temperatures for Sample S3 82.5/17.5 pTSA

150°C 160°C 170°C

Conversion (%) DSC 20 DSC 821e DSC 20 DSC 821e DSC 20 DSC 821e

20 . . . . . . . . . . . . . 5.06 12.40 3.02 6.06 1.94 3.05

30 . . . . . . . . . . . . . 7.00 14.58 4.96 7.82 3.34 4.32
50 . . . . . . . . . . . . . 12.06 18.70 7.97 11.03 5.71 6.66
75 . . . . . . . . . . . . . 25.10 24.87 15.30 15.66 11.63 10.08
90 . . . . . . . . . . . . . 32.53 31.99 24.56 20.60 17.02 13.52

Vol. 75, No. 937, February 2003 71

M. Korhonen et al.

Figure 3—Storage modulus E′ and loss modu- Figure 4—Variation of storage and loss modu-
lus E ′′ of clearcoat S1 82.5/17.5 pTSA at lus with time for three different temperatures:
isothermal temperature 170°C. 150°, 160°, and 170°C.

ments were done as described. The sample cup was be seen, the gel times decreased with an increase in tem-
filled about half way to 2 mm. Then it was placed into perature: 150°C at 2.4 min, 160°C at 1.6 min, and 170°C
the instrument at ambient temperature and brought to at 1.1 min. This was expected since an increase of tem-
the cure temperature as fast as possible. The top plate perature leads to a faster cure reaction and gelation,
was supported until the desired temperature was indicating the incipient formation of an infinite tridi-
obtained and then lowered so that it touched the sample mensional structure occurred earlier.
but did not penetrate it. This was done by applying As can be concluded from the obtained DSC data,
“floating forces,” that is, static forces with such forces DSC measurements provide valuable information on
causing the probe to float at sample height. The runs the extent of reaction and the rate of it, but do not pro-
were started using a dynamic force which gave an vide information on the gel point of the system or the
amplitude of about 5 µm throughout the run. resulting network structure. DMA measurements, on
DMA plots, storage modulus (E′) and loss modulus the other hand, do give information on the gel point and
(E′′), as a function of time for sample S1 82.5/17.5 pTSA also on the network structure, crosslinking efficiency. In
at 170°C, are shown in Figure 3. Gelation is one of the addition the viscosity of the reactive mixture can be
most important kinetics characteristics of a curing measured during the crosslinking reaction. In general it
process. The gel time value depends on the criterion seems that combined use of two or more thermal analy-
used for its determination and therefore it is important sis techniques is the most advantageous in coating
to specify criterion. Tung and Dynes 26 found a correla- investigations. An interesting technique is an isothermal
tion between the gel point and the intersection point of time-temperature-transformation (TTT) cure diagram,
the curves of storage modulus E′ and loss modulus E′′. introduced by Gillham, 28 which provides a useful tool
Gillham 27 considered the gel point to the point at which for understanding the curing and analyzing the curing
the storage modulus E′ reaches its maximum in isother- processes of thermosets. With the aid of the TTT dia-
mal curing studies. Lange et al.28 investigated gelation gram it is possible to design an optimum cure cycle for
using a torsional dynamic mechanical analyzer and desired material.4,29
found that gelation occurred when G′=G′′ or tan δ=1. In
this work gel times were determined as the crossover of
the elastic (E′) and viscous (E′′) modulus, i.e. E′=E′′.
CONCLUSIONS
Figure 3 shows a typical result we obtained in a fre-
quency scan experiment. At the beginning, when the DSC and DMA analyses were used to characterize the
clearcoat was in a liquid state, the viscous modulus pre- cure of a polyester resin system used in coil coating
dominate and E′′>E′. At the gel point E′′=E′ and then E′ applications. The coil coating process requires a fast cur-
modulus increased with time. Finally, E′ modulus ing system and it is essential to find an optimum cure
increases still more and reaches a plateau, indicating profile for different applications.
that the sample has been cured. At isothermal tempera- DSC can be a very useful tool in predicting curing
ture 170°C the E′ curve begins to change to plateau after times for different samples. The curing times can be eas-
10 min which correlates well with measurements of con- ily and quite reliably obtained by any DSC apparatus
versions by DSC 821e. Table 3 shows predicted values of despite their age and lack of modern computer pro-
conversion for this same sample: 50% at 4.73 min, 75% grams. Nevertheless, the modern software programs
at 7.47 min, and 90% at 10.06 min. proved to be very practical and time saving aids in cur-
Figure 4 shows modulus curves for S1 82.5/17.5 ing studies. The same kind of results were obtained by
pTSA at three different isothermal temperatures. As can both Mettler and Perkin-Elmer DSCs and the differ-

72 Journal of Coatings Technology

 Study of Polyester-Based Coil Coatings

ences in the results were probably due to the different (10) Gillham, J.K., “Structure/Property Relationship in Epoxy
heating rates and aging of the sample. However, the Resins,” J. Appl. Polym. Sci., 19, 676 (1979).
(11) Gualpa, M.C., Riccardi, C.C., and Vázquez, A., “Study of the
results were in good agreement in the predicted conver- Kinetic and Crosslinking Reaction of Novolak with Epoxy
sion range 20-75%. Different polyester types were also Resin,” Polymer, 39, 2247 (1998).
compared and the S3 was found to be the most reactive (12) Martin, J.S., Laza, J.M., Morrás, M.L., Rodríguez, M., and León,
one. It gave the fastest curing time at every temperature L.M., “Study of the Curing Process of a Vinyl Ester Resin by
Means of TSR and DMTA,” Polymer, 41, 4203 (2000).
at conversions 50 and 75%. (13) Vilas, J.L., Laza, J.M., Garay, M.T., Rodríguez, M., and León, L.
By using the DMA technique, liquid samples like M., “Unsaturated Polyester Resins Cure: Kinetic, Rheologic, and
clearcoats can be analyzed with cup-and-plate geome- Mechanical-Dynamical Analysis. I. Cure Kinetics by DSC and
try. Gelation cannot be observed directly by techniques TSR,” J. Appl. Polym. Sci., 79, 447 (2001).
(14) Simitzis, J., Stamboulis, A., Tsoros, D., and Martakis, N.,
sensitive only to the chemical reaction, such as DSC, “Kinetics of Curing of Unsaturated Polyesters in the Presence of
whereas DMA is capable of measuring and estimating Organic and Inorganic Fillers,” Polym. Int., 43, 380 (1997).
the gel point. (15) Wang, C.S. and Lin, C.H., “Properties and Curing Kinetic of
Diglycidyl Ether of Bisphenol A Cured with a Phosphorus-
Containing Diamine,” J. Appl. Polym. Sci., 74, 1635 (1999).
(16) Li, Z.F., Callahan, R.F., Hanson, C., and Howard III, J.M., “In-situ
ACKNOWLEDGMENTS Cure Characterization of Coatings by Thermally Stimulated
Current (TSC) Spectroscopy and Relaxation Map Analysis
The authors would like to thank Neste Chemicals, (RMA),” Proc. for Relaxation Phenomenon in Polymers Symposium,
Rautaruukki Steel, and Teknos Steel for samples and Dearborn, MI, pp.1-4, Nov. 8, 1994.
financial support. In addition, the financial support (17) Neag, M.C., “Coatings Characterization by Thermal Analysis,”
in Paint and Coating Testing Manual, Koleske, J.V. (Ed.) 14th Ed.,
from the National Technology Agency (TEKES) is grate- ASTM, Philadelphia, pp. 855-858, 1995.
fully acknowledged. (18) Vyazovkin, S. and Wight, C.A., “Model-Free and Model-Fitting
Approaches to Kinetic Analysis of Isothermal and Isothermal
Data,” Thermochimica Acta, 340-341, 53 (1999).
References (19) Vyazovkin, S. and Goryachko, V., “Potentialities of Software for
(1) Prime, R.B., Thermosets in Thermal Characterization of Polymeric Kinetic Processing of Thermoanalytical Data by the
Materials, Turi, E.A. (Ed.), 2nd edition, Vol. 2, Academic Press, Isoconversion Method,” Thermochimica Acta, 194, 221 (1992).
New York, 1659-1663, 1997. (20) Kelsey, M.S., Applications of Model-Free Kinetics, American
(2) Schmithenner, M., “Basic Resins for Coil Coatings,” Euro. Coat. J., Laboratory, pp. 13-16, 1996.
p. 9, Presented at the 63rd Annual Conference, October 7-9, 1997. (21) Pyris Kinetics Software Guide, Perkin-Elmer Instruments, USA,
(3) Turner, G.P.A., General Industrial Paints in Paint and Surface pp. 7-59, 1999.
Coatings Theory and Practice, Lambourne, R. and Strivens, T.A. (22) Korhonen, M., Starck, P., and Löfgren, B., “Characterization of
(Eds.), 2nd edition, William Andrew Publishing, USA, 502-528, Coil Coatings by Thermal Analysis,” J. Appl. Polym. Sci., in press.
1999. (23) Nielsen, L.E. and Landel, R.F., Mechanical Properties of Polymers
(4) Barral, L., Cano, J., López, J., Nogueira, P., Ramirez, C., and and Composites, 2nd ed., rev. and expanded, Marcel Dekker Inc.,
Abad, M.J., “Isothermal Cure of an Epoxy System Containing New York, pp. 131-180, 1994.
Tetraglycidyl-4,4´-diaminodiphenylmethane (TGDDM), a (24) Hill, L.W., “Mechanical Properties of Coatings,” Federation Series
Multifunctional Novolac Glycidyl Ether and 4,4´- on Coatings Technology, Federation of Societies for Coatings
Diaminophenylsulphone (DDS): Vitrification and Gelation,” Technology, Philadelphia, p. 25, 1987.
Polym. Int., 42, 301 (1997). (25) Campbell, D. and White, J.R., Polymer Characterization Physical
(5) Cook, W.D., Lau, M., Mehrabi, M., Dean, K., and Zipper, M., Techniques, Chapman and Hall Ltd., The University Press,
“Control of Gel Time and Exotherm Behaviour During Cure of Cambridge, pp. 304-313, 318-326, 1989.
Unsaturated Polyester Resins,” Polym. Int., 50, 129 (2001). (26) Tung, C. and Dynes, P.J., “Relationship Between Viscoelastic
(6) Auad, M.L., Aranguren, M.I., Elicabe, G., and Borrajo, J. “Curing Properties and Gelation in Thermosetting,” J. Appl. Polym. Sci.,
Kinetics of Divinyl Ester Resins with Styrene,” J. Appl. Polym. 27, 569 (1982).
Sci., 74, 1044 (1999). (27) Lange, J., Johansson, M., Kelly, C.T., and Halley, J.P., “Gelation
(7) Provder, T., “Cure Characterization in Product Research and Behaviour During Chainwise Crosslinking Polymerisation of
Development,” J OURNAL OF COATINGS TECHNOLOGY, 61, No. 770, Methacrylate Resins,” Polymer, 40, 5699 (1999).
33 (1989). (28) Enns, J.B. and Gillham, J.K., “Time-Temperature-Transformation
(8) Salla, J.M. and Martin, J.L., “Dynamic, Isothermal and Residual (TTT) Cure Diagram: Modeling the Cure Behaviour of
Heats of Curing of an Unsaturated Polyester Resin,” Thermosets,” J. Appl. Polym. Sci., 28, 2567 (1983).
Thermochimica Acta, 126, 339 (1988). (29) Gillham, J.K., “Molecular Design of Thermosetting Systems:
(9) De Nograro, F.F., Guerrero, P., Corcuera, M.A., and Mondragon, Evolution of Properties During Cure in Terms of Transitions,”
I., “Effects of Chemical Structure of Hardener on Curing Proc. ANTEC 2000 Volume II Materials, pp. 2086-2090, Orlando,
Evolution and on the Dynamic Mechanical Behavior of Epoxy FL, May 7-11, 2000.
Resins,” J. Appl. Polym. Sci., 56, 177 (1995).

Vol. 75, No. 937, February 2003 73