1/2004

Information for users of METTLERTOLEDO thermal analysis systems

19TA Tip

Dear Customer,This year METTLER TOLEDO celebrates 40 years of outstanding success in thermal analy-sis. In 1964 we introduced the TA1, an advanced, research grade TGA/DTA instrument.Since then, we have continuously expanded our TA product line with new ideas and solu-tions. And now we are proud to introduce the TMA/SDTA841e.Thermomechanical analysis is more sensitive than DSC for a number of applications. Wedemonstrated this in UserComs 17 and 18 for the determination of glass transition tem-peratures. TMA is in fact the ideal complement to a DSC system.

DSC measurements at high heating rates –advantages and limitationsDr. Matthias Wagner, Dr. Rod Bottom, Philippe Larbanois, Dr. Jürgen Schawe

IntroductionDSC measurements are usually performed at heating rates of 10 to 20 K/min. This is agood compromise between accuracy, resolution, sensitivity and actual measurementtime.

The use of high heating rates represents an attempt to measure the sample in its origi-nal state, i.e. so rapidly that the sample has no time to change during the measurement(e.g. cold crystallization).

The recent increased discussion on this topic can be summarized as follows: Measure-ments at high heating rates lead to• much shorter experiment times,• increased sample throughput,• better sensitivity with certain effects (e.g. the determination of weak glass transitions)

because of the greater heat flow signals at higher heating rates and• less reorganization in the sample during the DSC measurement.

Contents

TA Tip- DSC measurements at high

heating rates – advantagesand limitations 1

New in our sales program- TMA/SDTA841e 6- UV-DSC 7

Applications- Kinetic studies of complex

reactions. Part 2: Descriptionof diffusion control 8

- Curing kinetics of phenol-formaldehyde resins 10

- Curing of powder coatingsusing UV light 13

- Quality assurance of plasticmolded parts by DSC.Part 1: Incoming materials 16

Dates- Exhibitions 19- Courses and Seminars 19

UserCom 1/20042

In principle, the software allows high heat-ing rates (up to 400 K/min and more). Theaccuracy of the results so obtained must bechecked with respect to the physical andchemical facts. The laws governing physi-cal and chemical events such as reactionsand transitions must be taken into ac-count.

Such events include:• Heating rate-dependent effects:

- kinetic events- crystallization- chemical reactions- glass transitions

• Heating rate-independent effects:- melting- physical properties such as

specific heat capacity

Besides this, sample measurement is influ-enced by the following sample propertiesand instrument parameters:Sample• Situation within the sample:

- sample size- sample homogeneity- thermal gradients- thermal conductivity- decomposition- thermal contact

Instrument• Instrumental effects:

- crucible- signal-time constant- baseline stability- stabilization time- calibration

To gain a better understanding of the effectof these different factors, a number ofmeasurements will now be presented anddiscussed. All the measurements were per-formed with a DSC822e equipped with anautomatic sample robot. The instrumentwas calibrated with indium and zinc.

Application examplesMelting of metalsIn general, the melting point of pure sub-stances is determined as the onset tem-perature of the peak. For physical reasons,the side of the melting peak before meltingis a straight line. The enthalpy of fusion is

determined from the peak area. As a resultof the given thermal conductivity, meltingtakes place over a certain period of time.The peaks are therefore much broader athigher heating rates.

With the DSC822e, the melting point ofindium can be determined over a widerange of heating rates (Fig. 1). Due to theintegrated FlexCal® calibration system,onset temperatures and the enthalpies offusion are independent of the heating rate.This ensures that the DSC measures reli-ably at high heating rates and that accu-rate temperature and heat flow values areobtained.

To improve heat transfer, it is advisable toperform fast measurements in light cru-cibles. The measurements used for theevaluation in Figure 1 were done in 20-µlaluminum crucibles.

to higher final temperatures. At high heat-ing rates, the reaction peak is possiblyshifted to such high temperatures that thesample begins to decompose. The occur-rence of a decomposition reaction can beverified through measurement of the glasstransition temperature after the reaction.

The example in Figure 2 shows the resultsof a curing reaction using a reactive coat-ing powder (KU600) as an example. Heat-ing rates of up to 200 K/min were used.As expected, the glass transition tempera-ture is shifted to higher temperatures athigher heating rates. It can also be seenthat the chemical reaction is shifted tohigher temperature. The temperature ofthe peak maximum of the reaction showsa linear relationship with the logarithm ofthe heating rate. The increase is deter-mined by the activation energy. As shownin Figure 3, the measurement result corre-

In summary, one can say that- onset temperatures can be determined

accurately- the same calibration procedures as with

conventional DSC (heating rate to20 K/min) can be used.

Curing of resinsReaction temperatures should also increasewith increasing heating rates. To followthe complete reaction, one must measure

sponds to the predicted behavior. The reac-tion enthalpy is practically independent ofthe heating rate.The sample has a built-in «temperaturecalibration system» – the accelerator,which melts at 208 °C. As expected, thismelting process is independent of the heat-ing rate and proves that the shifts of theother characteristic temperatures are notartifacts. The results are summarized inTable 1.

Fig. 1. Onset temperatures and enthalpies of the melting peak of indium at different heating rates.

3UserCom 1/2004

Melting of PETAmorphous polyethylene terephthalate(PET) exhibits cold crystallization and theheating rate determines how much crystal-line material is formed and hence the de-gree of melting that takes place. Figure 4displays several curves of PET normalizedin heat capacity units measured at heatingrates between 10 and 300 K/min. After theglass transition at about 80 °C, cold crys-tallization occurs. The PET used for thesemeasurements melts at 240 °C. With in-creasing heating rates, cold crystallizationshifts to higher temperatures due to the ki-netic processes involved in nucleation. Thesmaller peak areas indicate that with in-creasing heating rates the degree of crys-tallinity decreases. The curves measured atlower heating rates do not characterize theoriginal material because the main effects(crystallization and melting) originateduring the actual measurement itself. Thecurve measured at 200 K/min shows onlyweakly induced processes and a very smallmelting peak. The absence of the meltingpeak at 240 °C on the curve measured at300 K/min shows that in this case no coldcrystallization at all has occurred. Thiscurve characterizes the original material.At first sight only the glass transition isvisible. A closer examination of the curverevealed a very small endothermic peak atabout 150 °C (marked with an arrow inFig. 4).

This peak, which occurs between 140 and160 °C, is shown on an expanded ordinatescale in Figure 5. At lower heating rates thepeak is masked by cold crystallization. Theeffect responsible for this peak is the melt-ing of a small amount of crystals that wereformed during cooling in manufacture.The peak characterizes the original crys-tallinity of the material (about 1%). Thelow melting temperature indicates the im-perfect structure of the crystallites.The experimental results show that at ahigh heating rate of 300 K/min, the origi-nal state of the material can be character-ized because no cold crystallization occursduring the experiment.Table 1. Characteristic thermoanalytical data of the curing reaction of KU600.

Heating rate Tg onset Tpeak reaction Tf accelerator ∆Hr[K/min] [°C] [°C] [°C] [J/g]

20 71.2 215.5 207.3 56.56

100 75.4 257.7 208.3 57.97

200 75.7 279.4 208.7 57.54

Fig. 2. Curing of KU600 powder at high heating rates.

Fig. 3. Peak temperature of the curing reaction of KU600 as function of the logarithmic heating rate.

UserCom 1/20044

transfer in the sample. The small liquidcrystal transition can still however clearlybe seen as a shoulder.The inserted diagram shows that the twoeffects can however be clearly separatedusing a deconvolution routine.

Polymorphic transitionsThe phase behavior of organic substancescan also be investigated at high heatingrates. This is shown using phenylbutazoneas an example. In the crystalline form thissubstance melts at 106 °C. If the melt isshock-cooled, amorphous material is pro-duced. On heating, the glass transitionoccurs at -5 °C followed by cold crystalli-zation and then melting of two polymor-phic phases. This behavior is shown inFigure 7 using heating rates of 50, 100 and150 K/min.It can be seen that, at higher heating rates,the crystallization peak is shifted to highertemperatures due to kinetic reasons. Thedecrease of the enthalpy of crystallizationwith increasing heating rate shows thatless material is able to crystallize. Due tothe high heating rate, the time for crystalgrowth between the beginning of cold crys-tallization and melting is very short.The polymorphic phases formed melt at 98and 106 °C. The transition temperatures ofboth polymorphic phases are independentof the heating rate. The areas of the meltingpeak and crystallization peak correspondto one another.

SummaryThe following table summarizes the ad-vantages of using high heating rates:

The diagram inserted in Figure 4 displaysthe step height of the glass transition atdifferent heating rates.

As expected, ∆cp is independent of the heat-ing rate. This result shows that it is possi-ble to achieve correct quantitative resultseven at high heating rates.

Liquid crystal transitionsAzoxydianisole melts at 118 °C and thenexhibits a liquid crystal transition at135 °C. Both effects are independent of theheating rate. This substance can also bemeasured at very high heating rates asshown in Figure 6. At 200 K/min the melt-ing peak is broadened due to the finite heat

Fig. 4. DSC curves of amorphous PET, normalized to initial sample weight and the heating rate (specific heatcapacity). Sample weight was less than 1 mg. Data sampling interval was 0.1 seconds.

Fig. 5. Melting peak of small crystals formed in amorphous PET during cooling in production.

Advantages

Suppression of kinetically controlled ef-fects (e.g. cold crystallization � meas-urement of the original crystallinity).

Simulation of production conditions.

Signal amplification due to the morerapid heating rate for- small samples- for samples that exhibit only weak

transitions.

High heating rates appreciably shortenthe measurement time.

5UserCom 1/2004

As with many other techniques, the advan-tages are offset by certain limitations inother respects.Possible limitations that should be consid-ered are:

ConclusionsThe DSC822e can measure samples at veryhigh heating rates. FlexCal® technologyensures that the measured temperaturesremain independent of the heating rate.The short response time of the DSC sensorresults in rapid stabilization of the signalso that starting temperatures do not haveto be very low.The influence of the heating rate on theeffect to be investigated should always becarefully considered beforehand. In gen-eral, higher heating rates lead to largerthermal gradients in the sample. This caninfluence the kinetics of processes, e.g. nu-cleation in connection with crystallizationis significantly affected.Furthermore, peaks are broadened due tosmearing. To reduce these effects, it is bestto use light crucibles and optimize the ther-mal contact between the sample and thecrucible. Small, thin samples reduce theformation of temperature gradients withinthe sample. The use of commercially avail-

able crucibles allows an automatic samplechanger to be used.

Higher heating rates permit kinetic proc-esses to be studied over a large dynamic

range. Undesired processes that occur dur-ing the measurement (e.g. cold crystalliza-tion) can be suppressed so that the meas-urement curve characterizes the originalmaterial.

Fig. 6. Melting peaks of azoxydianisole measured at different heating rates.

Fig. 7. Phenylbutazone measured at different heating rates.

Limitations

High heating rates degrade temperatureresolution (e.g. separation of peaks be-comes poorer or even impossible).

There are no standards (up until now)for the application of high heating rates.

Small samples are not representative ofthe sample as a whole.Sample preparation demands special care.

Higher heating rates necessitate higherend temperatures in methods because thepeaks become broader. At higher end tem-peratures the risk of sample decomposi-tion is greater.

Higher heating rates necessitate a highermeasurement signal data sampling rate.

UserCom 1/20046

New in our sales program

TMA/SDTA841e

We are once again pleased to introduce anew instrument.The TMA/SDTA841e complements theTMA/SDTA840 and extends the previoustemperature range available in thermo-mechanical analysis (TMA). The newmeasuring module allows you to analyzepolymeric and other materials in the range-150 °C to 600 °C.

Thermomechanical analysis measures thedimensional changes of a sample as a func-tion of temperature. The characteristic val-ues and properties most frequently studiedare:• expansion coefficients• glass transition temperatures• softening temperatures• solid-solid transitions• melting behavior• swelling behavior and• elastic behavior.

SDTA signalThe sample temperature is measured veryclose to the sample by means of a sheathedK-type thermocouple.A thin quartz glass coating protects it

against direct contact and contamination.This means that temperature adjustmentcan be performed using the melting pointsof pure metals and/or the length change ofsuitable standards.

Fig. 2. TMA is an excellent method for determining glass transitions and expansion coefficients. The figureshows the expansion behavior of two rubber samples (E-SBR: emulsion SBR; L-SBR: solution SBR). Glasstransitions are often difficult to see on TMA curves. They are however immediately apparent as a step changeon the expansion coefficient curve.

Fig. 1. The new TMA/SDTA841e module.

Typical areas of application

Industry Applications

Plastics (elastomers, thermosets, Coefficient of thermal expansion, glassthermoplastics) transition, creep behavior

Electronics industry Glass transition, coefficient of thermalexpansion, delamination

Paint/lacquers/adhesives/coatings Softening point

Textile fibers Expansion and shrinkage behavior

Packaging (films) Expansion and shrinkage behavior

Chemical (organic and inorganic Coefficient of thermal expansion andmaterials, metals, pharmaceutical expansion behavior, solid-solid transitionsproducts)

Universities Dimensional changes under the most variedconditions

7UserCom 1/2004

The SDTA signal is the difference betweenthe reference temperature calculated witha model and the measured sample tem-perature (US Patent 6146013).This means that, besides the length change,the simultaneously measured DTA signal(SDTA) is also available as a measurementquantity. This can often be of great helpwhen interpreting a measurement curve.

Gastight measuring cellThe measuring cell can be purged with gasin order to measure samples under definedatmospheric conditions just as in theTMA/SDTA840.

AutomationThe furnace opens and closes at the strokeof a key, guaranteeing easy and reliableoperation.

CoolingThe liquid nitrogen cooling system hasbeen completely automated; manual refill-ing is not necessary. This simplifies opera-tion, improves reproducibility and allowsmeasurements to be performed over a longperiod of time.

ThermostatingThe mechanical part of the measuring cellis incorporated in a thermostated housing.This minimizes drift and guarantees excel-lent accuracy for the determination of ex-pansion coefficients.

Deformation modesDifferent accessories enable you to performmeasurements on solid samples, foams,films and fibers.

The TMA/SDTA841e supports the followingmeasurement modes:• Expansion• Penetration• Tension• Bending• Swelling

The DLTMA mode (dynamic load TMA) al-lows you to study the elastic behavior ofmaterials.In DLTMA the force exerted on the probealternates automatically between twovalues.

UV-DSCThe DSC module is easily converted to aUV-DSC system by means of a special ac-cessory.

The UV-DSC system allows you to study theinfluence of temperature, UV light inten-sity and exposure time on curing processes.The sample is held isothermally at a par-ticular temperature and exposed to UVlight for a defined period of time. The re-action of the sample to the light is meas-ured with the DSC.

Overview of the TMA/SDTA841e specifications

Wide measurement range +/- 5.0 mm

High resolution 0.6 nm (over the whole measurement range)

Long sample lengths up to 20 mm

Large force range -0.1 to 1 N

High force resolution 0.13 mN

Wide temperature range -150 °C to 600 °C

Cooling system automatic liquid nitrogen cooling

High temperature accuracy +/- 0.35 K

SDTA resolution 0.005 K

Automation automatic opening of the furnace with keystroke

DLTMA mode ≤ 1 Hz

Options gas controllerswitched line socket

Light curing systems are used• to shorten processing times and/or• to avoid exposing sensitive materials

to high temperatures.

Nowadays, light curing systems are usedfor all kinds of coating applications (e.g.in the automobile industry) and in den-tistry (curing fillings).

The UV-DSC accessory can be easily in-stalled or removed by the user.Fig.1. DSC822e with the UV-DSC accessory.

Fig. 3. Quartz glass sample support with the K-typethermocouple for sample temperature measurement.

UserCom 1/20048

Applications

Kinetic studies of complex reactionsPart 2: Description of diffusion controlDr. Jürgen Schawe

IntroductionCross-linking reactions of «hot curing sys-tems» are complex reactions. Dependingon the reaction conditions, an initiallychemically controlled reaction can changeand become diffusion controlled. The reac-tion rate thereby decreases rapidly. The re-action almost stops. The reason for this ischemically induced vitrification as a resultof which the material changes from a liq-uid to a glassy state.In the first part of this study [1], the cur-ing reaction of the a two-component epoxyresin consisting of the diglycidylether of bis-phenol A (DGEBA) and diaminodiphenyl-methane (DDM) as hardener or curingagent was investigated by DSC and evalu-ated using the Model Free Kinetics (MFK)software. Here it was shown that curing atheating rates greater than 1 K/min ischemically controlled over the entire reac-tion range.

By means of MFK, it is possible to correctlypredict the course of an isothermal reac-tion right up to vitrification. If the glass

transition temperature is known as a func-tion of conversion, MFK can be used toestimate the time at which the materialvitrifies in an isothermal reaction.MFK must be extended in order to describethe complete course of a curing reactionthat includes the transition from chemi-cally to diffusion-controlled kinetics [2-4].In this article, MFK is enhanced sufficientlyfor it to take the effect of diffusion controlon the reaction kinetics into account.

Prediction of chemical behaviorThe conversion curveIf MFK is applied to DSC curves that weremeasured at sufficiently high heatingrates, one obtains activation energy curvesas a function of conversion that describe apurely chemically controlled reaction(curve a in Fig. 2 of Part 1 [1]). The acti-vation energy curve can then be used tocalculate the isothermal course of the re-action. A comparison of the conversioncurve calculated using MFK and the meas-ured conversion curve for Treact = 100 °C isdisplayed in Figure 1.

Up until about 40 min, the measured curveagrees well with the MFK prediction. Inthis range the cross-linking reaction ischemically controlled. Afterward, the influ-ence of diffusion control increases and themeasured reaction proceeds more slowly.A comparison of the two curves allows theinfluence of diffusion control on the reac-tion kinetics to be estimated.The reaction rate is described by the equa-tion

where dα/dt is the reaction rate (derivativewith respect to time of the measured con-version curve), [dα/dt]chem is the reactionrate of the chemically controlled reaction(obtained using MFK) and ƒd(α) is thediffusion control function that describesthe influence of diffusion control on thereaction kinetics. ƒd can be determined ac-cording to eq (1) by division of the meas-ured reaction rate and the reaction ratecalculated by MFK (see Fig. 1).

Fig. 1. Conversion as a function of time at a reaction temperature of 100 °C. αmis the measured curve and αchem is the prediction with MFK using the activationenergy curve (a) in Figure 2 from Part 1 [1].The inserted diagram displays the diffusion control function, ƒd. The continuouscurve is calculated from the conversion curves according to eq (1). The dashedline corresponds to the model function from eq (2).

Fig. 2. Conversion as a function of reaction time for isothermal reactions at differ-ent temperatures. Dashed line: result from MFK (αchem); continuous line: resultof the calculation according to eq (4), circles: measured curves. The line with thesquares marks the vitrification. These values were taken from Figure 4.

9UserCom 1/2004

Since the diffusion control is connectedwith the vitrification, a model function isintroduced for ƒd that takes into accountthe conversion-dependence of the glasstransition temperature [4]:

The only parameter in this equation is ∆T.This parameter describes the width of theglass transition. ∆T is set to 20 K, the widthof the glass transition of the fully curedmaterial.A modified DiBenedetto equation [5] isused to describe the glass transition tem-perature as a function of conversion:

where Tg0 and Tg1 are the glass transitiontemperatures of the unreacted and fullyreacted sample. λ is a fit parameter. Theparameters λ = 0.31 and Tg1 = 169.3 °Cwere determined [1] by means of curve fit-ting to the measurement results from post-curing experiments [1]. Tg0 = -18.3 °Cwas measured by DSC at 10 K/min [1].As can be seen in Figure 1, the model func-tion calculated according to eq (2) agreesvery well with the experimentally deter-mined diffusion control function. Usingeqs (1) and (2), a complete prediction ofthe kinetics of such complex reactions canbe made based on MFK.

The isothermal conversion curves were cal-culated using a spreadsheet. The change ofconversion at a reaction step ∆αi is givenby

where αi is the sum of all the previous re-action steps:

αchem,i are the individual values of theMFK prediction for the isothermal reactionat Treact.At the beginning: ∆α0 = α(treact = 0) = 0and ∆α1 = α1 = αchem,1.The indices i and j (i,j = 0,1,2,3...) describethe number of the base points.

The conversion calculated according to eq(4), the results of the MFK prediction andthe measurements are displayed in Figure2 and show agreement of the measure-ments with the predictions of MFK beforevitrification begins. The results of the cal-culation taking into account the diffusioncontrol function describe the entire reac-tion.

Maximum conversionAfter vitrification, the reaction becomesmore and more diffusion controlled andthen practically stops at a maximum con-version, αmax, which is smaller at lower re-

Fig. 3. The maximum conversion for an isothermal reaction as a function of thereaction temperature.

Fig. 4. Vitrification time as a function of reaction temperature for an isothermal re-action. The measured data (filled circles) were determined from postcuring meas-urements (Fig. 3 in Part 1 [1]). The empty circles show the MFK results and thesquares have been calculated taking the diffusion control function into account.

action temperatures. αmax is an importanttechnological quantity because it indicatesthe conversion up to which the materialcan react. A greater conversion can only bereached by raising the reaction tempera-ture.The calculated values of αmax originatefrom the end points of the calculations ac-cording to eq (4) (Fig. 2). Figure 3 showsthat they agree with the measured valueswithin the limits of measurement uncer-tainty.

The vitrification timeAnother important factor for describingcuring reactions is the vitrification time,tv. This is important from the technologi-cal point of view because it indicates thetime after which the resin is no longer liq-uid. At the vitrification time, Tg = Treact.The conversion-dependence of Tg is usedto calculate α at tv (eq (4) and Fig. 4 in[1]). The time for the conversion curve toreach this value is called the vitrificationtime. tv was determined from the two con-version curves, one value calculated byMFK [1] and the other according to eq (4)taking the diffusion control function intoaccount. The values are displayed togetherwith the experimental results from postcur-ing measurements (Fig. 3 in Part 1 [1]) inFigure 4. It can be seen that the vitrifica-tion time can be predicted with very goodaccuracy on the basis of both calculations.Only at high reaction temperatures (greaterthan 130 °C) is the time calculated byMFK somewhat too short. If the diffusion

UserCom 1/200410

control function is taken into account, tv iscorrectly predicted over the entire tempera-ture range. MFK alone without taking dif-fusion control into account is adequate topredict the storage life.

ConclusionsMFK adequately describes the reaction ki-netics if the DSC measurements are madeat sufficiently high heating rates so thatthe reaction remains chemically control-led. If the material vitrifies during the re-action, the kinetics changes from beingchemically controlled to diffusion control-led. In this case, a better description of thekinetics can be achieved through the intro-duction of a diffusion-control function.This function enhances MFK in such a way

that it also describes the transition to dif-fusion control. It is then possible to calcu-late the maximum conversion that can bereached.

To predict the kinetics of the curing reac-tion with MFK requires at least three dy-namic DSC measurements performed atdifferent heating rates. Furthermore, atleast four (better six) postcuring measure-ments to determine the conversion depend-ence of the glass transition temperatureare needed. The glass transition tempera-ture of a fully cured sample should be de-termined in an additional experiment. Incomparison to the time required for iso-thermal measurements, the use of MFKpresents a considerable saving of time.

Literature[1] J. Schawe, UserCom 18 (2003) 13.[2] S. Vyazovkin, New J. Chem., 24 (2000)

913.[3] S. Vyazovkin N. and Sbirrazzuoli,

Macromol. Rapid Commun, 21 (2000)85.

[4] J.E.K. Schawe, J. Thermal. Anal. Cal., 64(2001) 599.

[5] J.P. Pascault and R.J.J. Williams,J. Polym. Sci.: Part B: Polym. Phys. 28(1990) 85.

Previous UserCom articles can be down-loaded viahttp://www.mt.com/home/products/en/lab/thermal/analysis/141401.asp

Curing kinetics of phenol-formaldehyde resinsFelisa Chan, Guangbo He, Yves Bedard and Bernard Riedl, Université Laval, Dept of Wood Science and CERSIM, Ste-Foy, Qc, Canada, G1K 7P4

IntroductionDSC and TGA are well-known techniquescommonly used to perform kinetic studiesof materials in thermal analysis (TA).Knowledge of the influence of temperatureon the reaction rate of phenol-formalde-hyde resins (PF resols) allows the polym-erization rate to be predicted. This is ofgreat practical importance for example inthe manufacture of wood composite mate-rials or for the storage of the resins.

Most of the resol cure kinetics reported inthe literature have been obtained using theBorchardt-Daniels method. Here only onerun at one heating rate (dynamic meas-urement) is necessary to obtain the desiredkinetic parameters. The kinetic parametersare, however, usually heating rate depend-ent and care must be taken when evaluat-ing and interpreting the results. Likewise,the activation energy is often assumed tobe constant throughout the reaction. Thisis, however, not true for complex curing re-actions such as occur in phenolic resins. A

phenol-formaldehyde (PF) resin may be oflow viscosity at the start of the run and fitthe assumptions for dilute solutions. Butas the curing reaction proceeds, the mate-rial undergoes gelation, i.e. changes froma liquid to a rubbery state, and possibly vit-rifies (transition from a rubbery to a glassystate). The cross-linking reduces molecu-lar mobility and results in the processchanging from being kinetically controlledto diffusion-controlled [1].

The model-free kinetics (MFK) method,developed by Vyazovkin [3-5] is based onthe assumption that the activation energy,Ea, is dependent on the conversion (α). Ata particular conversion, the activation en-ergy, Ea, is independent of the heating rate.Evaluation by MFK requires at least threedynamic measurements performed at dif-ferent heating rates.Since the reaction of phenol and formalde-hyde under alkaline conditions to a PF resinis complex, MFK evaluation seems a suit-able method to describe the curing behavior.

Experimental detailsA liquid PF resol consisting of low and highmolecular weight fractions (1:1 by weight)was synthesized in the laboratory.

The high molecular weight fractions wereprepared with a formaldehyde-to-phenolratio of 2.2 and a ratio of NaOH to phenolof 0.35; the low molecular weight fractionshad ratios of 2.2 and 0.17. The resultingresin had a solid content of 46% («panmethod»).

A METTLER TOLEDO DSC 20 with STARe

software was used for all the thermoana-lytical measurements. Since such adhesiveformulations contain water, re-usable 30-µlhigh-pressure steel crucibles capable ofwithstanding vapor pressures up to 10 Mpawere used. The dynamic DSC measure-ments were performed at heating rates of5, 10 and 20 K/min in the temperaturerange 25 to 250 °C. The DSC curves wereevaluated with different kinetic methods(STARe software options).

11UserCom 1/2004

Measurements and resultsEvaluation according to Borchardt-DanielsFigure 1 shows the DSC curves of the PFresol measured at heating rates of 5, 10and 20 K/min. The exothermic reactionpeaks were shifted to higher temperaturesat increasing heating rates due to kineticeffects. The beginning of an endothermicdecomposition reaction can be seen from200 °C onward especially in the curvemeasured at 20 K/min. This overlaps theend of the curing reaction. A «spline»baseline type was therefore used to inte-grate the DSC peaks to determine the con-version and the reaction rate. Care wastaken to make sure that the total reactionenthalpy was about the same for eachevaluation because it can be assumed thatit is independent of the heating rate. Thekinetic parameters for each measurementcurve were determined using the nth orderkinetics method according to Borchardt-Daniels (see Fig. 1 and Table 1). As thetable shows, with the Borchardt-Danielsmethod the results change with the heat-ing rate. Ea at 20 K/min has about thesame magnitude as the value published byAmen-Chen et al. [6] for liquid laboratory-synthesized PF resin (Ea = 129 kJ/mol at10 K/min).To demonstrate how well the three valuesdescribe the curing process, the kinetic pa-rameters were used to simulate the DSCcurves. As Figure 1 shows, this was donewith the parameter set for 5 K/min. The

measured curve is relatively well simu-lated; at higher heating rates, however, thesimulated curve is displaced by more than5 K from the corresponding measurementcurve. This shows that the kinetic descrip-tion is only valid for the measured condi-tions, and that other situations cannot beaccurately predicted.Table 1 also lists the evaluation accordingto ASTM E698. From the three measure-ments, only the peak temperatures areused. The reaction order, n, is set to 1. Eaand ln(k0) differ significantly from theBorchardt-Daniels values. Differences inthe kinetic parameters from multiple dy-namic measurements and from isothermalmeasurements have also been described in[2]. The activation energy from individualdynamic measurements was more than25% greater than that determined frommultiple measurements.

Evaluation with model free kineticsFigure 2 shows the evaluation with modelfree kinetics. The first step is to determinethe conversion curve as a function of tem-perature for each of the three heatingrates. The same baselines as in Figure 1were used. The diagram clearly shows thatthe reaction curves are shifted to highertemperatures at higher heating rates. Forheating rates of 5, 10 and 20 K/min, 50%conversion was for example reached at131.9, 141.8 and 153.3 °C. This informa-tion can already be valuable for perform-ing the reaction process on a technicalscale.In a second step, the activation energy (Ea)curve is calculated as a function of conver-sion from the conversion curves. At the be-ginning of the reaction, Ea increases byabout 5%, and then attains a practicallyconstant value of about 86 kJ/mol between

Fig. 1. DSC curves of the PF resol measured in high-pressure crucibles at 5, 10and 20 K/min. The parameters of the nth order kinetics are entered for each curve.The simulated curves (red dashed lines) were calculated with the kinetic parame-ters from the measurement at 5 K/min.

Fig. 2. Model free kinetics: Above left: the conversion curves calculated from theDSC curves (see Fig. 1) for heating rates of 5, 10 and 20 K/min. Above right: theactivation energy, Ea, as a function of reaction conversion. The result block showsEa and ln(k0) evaluated according to ASTM E698. Bottom half: the simulatedDSC curves plotted together with the measured DSC curves. The dashed curvesshow the curves predicted from the ASTM parameters; the continuous red curvesshow the curves from MFK.

Table 1. Summary of the numerical results of the PF resol curing reaction.

Kinetic evaluation Heating rate Ea n ln(k0) ∆Hr

K/min kJ/mol - - J/g

Borchardt-Daniels 5 117 1.17 29.4 428.3

Borchardt-Daniels 10 116 1.12 28.9 426.8

Borchardt-Daniels 20 124 1.33 31.2 423.3

ASTM E698 88.6 1 21.0

MFK (α = 50%) 86.4

MFK (α: 30 to 80%) 86.4 ± 0.3

UserCom 1/200412

40 and 80%. This value is very similar tothe value obtained from the ASTM E698method. The rapid increase of Ea above 90%conversion indicates a significant changein the reaction mechanism. The results ofAmen-Chen et al. [6] also show a practi-cally constant activation energy from 20%conversion onward.The results from MFK can then be used tosimulate DSC curves (Fig. 2). Comparisonwith the measured curves shows excellentagreement. There are small deviations to-ward the end of the reaction where the DSCcurves are deformed through overlappingeffects (decomposition).

For comparison Figure 2 also shows theDSC curves calculated using the kinetic pa-rameters obtained from ASTM E698. Com-parison with the measured curves showsthat in spite of good agreement of the Eavalues from ASTM and MFK evaluation,the curve simulated with the ASTM param-eters is significantly poorer than the curvesimulated with the results obtained fromMFK.He et al. [1] report a value ofEa = 60 kJ/mol (MFK method) for the PFresol described in their publication.Borchardt-Daniels evaluations tend to givehigher activation energies compared withMFK and ASTM698 values. A similar obser-vation was made by Amen-Chen et al. [6]in their pyrolysis reactions of oil-PF resins.

Prediction of reaction behaviorThe results of the kinetic analysis allowone to predict the course of the curing re-action under other reaction conditions.Besides the simulation of DSC curves al-ready shown (Figs. 1 and 2), predictions ofa more technical interest are summarizedin Figure 3. The diagram on the left thatshows conversion as a function of timeanswers the question, «How long does thereaction take at a particular isothermal re-action temperature?» The other diagramon the right shows how the reaction timechanges with temperature for a given con-version. According to the MFK prediction,

it takes for example 40 minutes to obtain a95% degree of cure at a constant tempera-ture of 120 °C. In contrast the predictionmade on the basis of the ASTM results gives23 min, which in view of the poor simula-tion of the DSC curves (Fig. 2) seems im-probable.These types of predictions are important forpractical processes using PF resins, e.g. inhot press molding cycles in the manufac-ture of wood composite materials. Further-more, by properly choosing the conditions,predictions can also be made with respectto the storage life of the resins.

ConclusionsThe reaction behavior of phenol-formalde-hyde resins can be studied with differentkinetic evaluation methods. From the re-sults, the conversion-temperature-timebehavior can be predicted and answersgiven to technological questions. The acti-vation energy is often the key quantity. Acomparison of the methods has shown thatonly model free kinetics (MFK) is able toadequately describe the complex curingreaction in most respects. The Borchardt-Daniels method can hardly determine theactivation energy properly and the ASTM

E698 method yields the activation energyonly at the highest reaction rate (DSC peaktemperature). MFK however shows the acti-vation energy curve as a function of conver-sion. This can provide useful informationfor the interpretation of the curing mecha-nisms besides yielding trustworthy simula-tion data. It has for example been shownfor PF resins in wood composites (He et. al.[1]) that the degree of cure of the resin inthe wood is not so high as when the resinis used in the pure form. Diffusion of theresin in the wood, which has a porousstructure, is hindered or even preventedthrough gelation of the adhesive.

Literature[1] He, G., Riedl B., Aït-Kadi A., J. Applied

Polymer Science, 87 (2003), p. 433.[2] Prime, R.B., in: Thermal Characteriza-

tion of Polymeric Materials. AcademicPress, San Diego, (1997) p. 1636-1646.

[3] Vyazovkin, S., Lesnikovich, A. 165Thermochimca Acta (1990), p. 273.

[4] Vyazovkin, S. Thermochimca Acta 194(1992), p. 221.

[5] Vyazovkin, S., N. Sbirrazzuoli. Macro-molecules (1996) 29, p. 1867.

[6] Amen-Chen, C., B. Riedl, C. Roy.Holzforschung. 56(3), (2002) p. 273

Fig. 3. Applied kinetics: prediction of conversion behavior. Left diagram: conver-sion as a function of time for three different reaction temperatures. Right diagram:reaction time as a function of the isothermal curing time for three different degreesof conversion. The red continuous lines show the predictions based on model freekinetics, the black dashed curves show the behavior based on the results of theASTM E698 kinetics evaluation.

13UserCom 1/2004

Curing of powder coatings using UV lightDr. Markus Schubnell

IntroductionToday, powder coating technology is ap-plied to a wide range of different materials(wood, plastics, metals, etc.).

Besides the excellent properties of suchcoatings, their use also offers importantecological advantages. For example, unlikeliquid paints, no solvents are used so that

only negligible amounts of volatile organiccompounds (VOCs) are released into theatmosphere.The powder coating is usually sprayed ontothe parts to be coated and then cured. Thecuring or cross-linking process is then per-formed either thermally in an oven (typi-cally at about 180 °C) or by means of UVlight. Curing with UV light has the great

Experimental detailsDesign of the UV-DSC systemThe measurements were performed with aMETTLER TOLEDO DSC822e equipped withan accessory that allowed the sample to beexposed to UV light.The UV-DSC system is pictured in Figure 1and a schematic diagram is shown in Fig-ure 2.The light source used was a Hamamatsu«Lightningcure 200» system. The spec-trum of the built-in mercury-xenon lampis shown in Figure 3.

advantage that materials sensitive to tem-perature (such as wood or plastic products)can be coated.In practice, a combined IR/UV processingline is used for the UV curing of powdercoatings. In the IR zone, the powder«melts» under the influence of the infra-red (IR) light and forms a homogeneousfilm on the substrate to be coated. The liq-uid film so formed is then cured in the UVzone within seconds or minutes.This article describes how DSC can be usedto investigate the UV curing behavior ofpowder coatings.

Fig. 3. Spectrum of the UV lamp used in the Hamamatsu«Lightningcure 200». The lamp emits UV light mainlyin the so-called UV-A region (315-400 nm). The UVlight emitted in the 290-315 nm wavelength regionis referred to as UV-B light. UV light of still shorterwavelengths (down to 100 nm) is called UV-C light.

Sample preparation and evaluation ofthe measurementsThe sample investigated was a commer-cially available powder coating materialused as a clear primer for wood fiberboardand chipboard surfaces. Typically 8.5 mgof the material was spread evenly over thebottom of the crucible forming a layerabout 0.8 mm thick. In each case, thesamples were held for 25 minutes at thetemperature at which curing was to be per-formed. During this time interval, thesample was exposed to UV light for 15 min-utes. The degree of cure was afterward

Fig. 1. Left: the UV-DSC accessory. Right: support for the quartz glass fiber bundle.

Fig. 2. Optical setup for projecting UV light into theDSC822e.

UserCom 1/200414

checked by measuring the glass transitiontemperature of the sample in a separateheating experiment at 10 K/min.The baseline during the curing process isof course required to determine the en-thalpy of cure. To a good approximation,this unknown baseline can be determinedby repeating the measurement with thecured sample under exactly the same con-ditions (including UV exposure). Anotherpossibility is simply to integrate the exo-thermic curing peak during the UV expo-sure interval using a horizontal baseline.Figure 4 shows that the two methods infact yield similar results. For our experi-ments, only the second method was used.

Measurements and resultsWith UV curing, three basic questionsarise:1. What influence do the temperature and

UV light intensity have on the curingprocess?

2. How long does the sample have to beexposed to UV light to achieve an ad-equate degree of cure or cross-linking?

3. What are the optimum parameters forthe cross-linking reaction?

The first two questions are investigated inmore detail in the following sections. The

ture (Fig. 6). The peak width here is definedas the full width of the peak at half themaximum peak height (Full Width at HalfMaximum, FWHM). The maximum heatflow (Fig. 7) changes by about 0.05 mWfor a temperature change of 10 K. At alight intensity of 1.12 mW/cm2, this corre-sponds to about 5% of the maximum heatflow. These results show that the curingtemperature has an insignificant (but nev-ertheless clearly measurable) influence onthe actual cross-linking process. In con-trast, temperature has a large effect on theviscosity (and hence on flow behavior) andin turn on the uniformity of the thicknessof the coating.

latter point is discussed in the final con-clusions.

Influence of temperature and UV lightintensityTo investigate the influence of temperatureand UV light intensity on the curing reac-tion, isothermal DSC curves were measuredat four different temperatures using threedifferent UV light intensities.An example is shown in Figure 5. The fig-ure displays curves of samples that weremeasured at 100, 110, 120 and 130 °Cwhile being exposed to a UV light intensityof 1.12 mW/cm2 (falling on the sample).The results show that temperature does nothave any appreciable influence on theenthalpy of cure at this level of light inten-sity. This also means that glass transitiontemperatures measured after curing are in-dependent of temperature. At higher tem-peratures, however, the maximum heatflow increases and the peak width de-creases, i.e. the curing or cross-linking re-action takes place more rapidly.

Figures 6 and 7 summarize these relation-ships for the different UV light intensitiesused. With a light intensity of 1.12 mW/cm2,the width of the curing peak changes byabout 20 s for a 10 K change in tempera-

Fig. 6. Influence of temperature on the curing time fordifferent UV light intensities. The FWHM increases byabout 0.3 minutes when the temperature is reducedby 10 K.

Fig. 5. Curing of a powder coating system at different temperatures using a UVlight intensity of 1.12 mW/cm2 (on the sample). The sample was held isother-mally for 25 minutes at the given temperatures and exposed to UV light for 15 mi-nutes. The left side of the figure displays the isothermally measured curves andthe evaluation of the enthalpy of cure. The right side of the diagram shows theheating curves recorded (at 10 K/min) immediately after curing. The conclusionis that temperature has no significant influence on the degree of cure at this levelof light intensity. In contrast, the widths (and therefore the heights) of the curingpeaks vary with temperature (Figs. 6 and 7).

Fig. 4. Determination of the enthalpy of curing in UV-DSC experiments. The sam-ple was held isothermally for 25 minutes at 110 °C. After 6 minutes, the shutterof the lamp was opened and the sample exposed to UV light for 15 minutes dur-ing which time the curing reaction took place. The reaction enthalpy is obtainedfrom the measured curve (black, continuous) by integration of the exothermiccuring peak. If baseline correction is not applied, a horizontal baseline is drawnat equilibrium heat flow after curing (21 minutes). Correction of the original meas-ured curve with a baseline yields the red curve. The baseline is the black dottedDSC curve that results when the cured sample is again exposed to UV light of thesame intensity for the same interval. Calculation of the reaction enthalpy, i.e. theenthalpy of cure, yields the same value with both methods within the limits ofmeasurement accuracy.

15UserCom 1/2004

The two figures also show that the UV lightintensity strongly influences the curingtime. This relationship is shown in Figure8. If the data is approximated to a simple

exponential function, the FWHM (andtherefore the curing time) doubles whenthe light intensity is reduced by about0.4 mW/cm2.

Thermal curing (without exposure toUV light)The powder coating system studied herecan also undergo thermal curing or cross-linking. This is illustrated in the DSCmeasurements shown in Figure 9. The sec-ond heating run was used as baseline forthe determination of the reaction enthalpyor enthalpy of cure. The measurementsshow that thermal cross-linking does notbegin until about 180 °C. This proves thatno significant degree of thermal cross-link-ing takes place in the temperature range inwhich the UV curing experiments were per-formed.

Influence of the UV exposure time onthe degree of cureFor a given UV light intensity, the degree ofcure and hence the glass transition tem-perature depend mainly on the UV exposuretime. Figure 10 shows the heat flow curvesfor the isothermal curing of samples ex-posed to UV light for different periods oftime. The sample temperature was 110 °Cand the light intensity 1.12 mW/cm2. Theinserted diagram in the upper right part ofthe figure shows heating curves of thesamples measured immediately after cur-ing (heating rate 10 K/min). From thesecurves, the relationship between the glasstransition temperature and the degree of

cure, i.e. conversion or degree of cross-linking can be plotted (Fig. 11). The de-gree of cure corresponds to the ratio of thereaction enthalpies of a partially cured anda fully cured sample. For practical applica-tions, the coating should have a glass tran-sition temperature of about 70 °C. This isachieved with a degree of cure or conver-sion of 80%. Figure 12 displays the conver-sion as a function of time for a samplethat was exposed to a UV light intensity of1.12 mW/cm2 at 110 °C until it was fullycured. The diagram shows that a conver-sion or degree of cure of 80% is reached af-ter an exposure time of about 5 minutes.

ConclusionsPowder coating technology with UV lightcuring is used to apply high quality coat-ings to the surfaces of temperature sensi-tive materials. The influence of tempera-ture, UV light intensity and exposure timeon the curing process can easily be deter-mined using a DSC822e equipped with aUV accessory.The experiments clearly showed that tem-perature has a negligible influence on theactual curing or cross-linking process forthe powder coating system studied here.Temperature does however have an impor-tant influence on the flow properties of theuncured coating material and hence onthe homogeneity of the coating producedin the cross-linking reaction. A number ofother studies have established that the flowbehavior of this powder coating at 110 °Cis optimal for the UV cross-linking process.

Fig. 7. Influence of temperature on the maximum heatflow during curing for different UV light intensities. Themaximum heat flow increases by about 0.05 mWwhen the temperature is increased by 10 K.

Fig. 8. Relationship between the FWHM and the UVlight intensity at different temperatures. The diagramshows that the influence of temperature on the curingor cross-linking process can be neglected in the tem-perature and light intensity ranges used here. If anexponential relationship is assumed for the FWHMand light intensity, it shows that the FWHM approxi-mately doubles when the light intensity is reduced by0.4 mW/cm2.

Fig. 9. Thermal curing or cross-linking reaction of a powder-coating system. Thecuring process begins at about 180 °C at the heating rate used here (10 K/min).Evaluation of the glass transition of the cured material yielded a midpoint valueof 71.9 °C.

Fig. 10. Influence of UV exposure time on the degree of cure. Sample temperature110 °C; light intensity 1.12 mW/cm2. Upper right: the glass transitions of the par-tially cured samples (heating rate 10 K/min).

UserCom 1/200416

Quality assurance of plastic molded parts by DSCPart 1: Incoming materialsDr. Achim Frick, Claudia Stern, Fachhochschule Aalen, Kunststofftechnik, Aalen, Germany

The article describes several practical ap-plications that illustrate the use of DSC forthe quality control of plastic molded parts.It explains how DSC can be used to identifymaterials and detect differences betweenbatches. The influence of additives such ascolorants and stabilizers is also discussed.

IntroductionThe quality of incoming plastic moldingmaterials is often only assessed by charac-terizing their flow properties in the melt orin solution, in particular by measuring themelt flow index (MFI) or the viscosity in-dex (VI) in solution. These rheologicalmethods do not however identify the mate-rial or allow information to be obtainedabout melting and solidification behavior,

although the latter is directly relevant tothe properties of the molded parts produced.This article discusses different applicationsof DSC for the quality control of incomingmaterials using practical examples.Part 2 of this series will deal with the useof DSC for process and production control.

Material identificationSemicrystalline polymeric materials meltover a temperature range characteristic ofthe particular type of polymer concernedwhereas amorphous molding materialsusually exhibit a pronounced glass transi-tion. For this reason, melting temperaturesare used to identify semicrystalline thermo-plastic molding materials, and glass tran-sition temperatures amorphous materials.

Figure 1 shows the DSC second heatingcurves of several different polymers. Priorto each measurement, the sample was firstheated to eliminate the possible effects ofthermal and mechanical history and thencooled reproducibly under control. Themelting temperature is taken to be themaximum of the melting peak.

DSC can also distinguish between differentvariations within a particular class of poly-mers. This is illustrated in Figure 2, whichshows the melting behavior of a POM(polyoxymethylene) homopolymer and aPOM copolymer. While the molecules ofthe homopolymer consist of just one typeof monomer, the molecules of the copoly-mer contain different monomers. These

UV-DSC experiments showed that about 5minutes were needed to attain an adequatedegree of cross-linking at this temperatureand with a light intensity of 1.12 mW/cm2.If the UV light intensity is doubled, the ex-

posure time is reduced to about 3 minutes.There is however a limit to the possible re-duction in UV exposure time: light inten-sities that are too high lead to incompletecuring of the powder coating. On the other

hand, UV exposure times that are too longlead to an increase in temperature on thesurface of the substrate. This inevitablyresults in the coating having poorer me-chanical properties.

Fig. 12. The UV exposure time necessary to achieve a given conversion or degreeof cure, e.g. of 80%, can be determined from the curing curve of a sample that isexposed to UV light until it is fully cured.

Fig. 11. Glass transition temperature as a function of conversion or degree ofcure.

17UserCom 1/2004

the width and temperature maximum ofthe crystallization peaks. Under definedcooling conditions, different morphologi-cal structures are formed when the poly-mer melt solidifies. This influences theproperties of the individual molded parts.Applied to the field of plastics processing,this means that the functional propertiesof the molded parts can differ due to differ-ences in the materials from batch to batcheven though the same processing condi-tions were used in production. The proper-ties of the parts with respect to quality maytherefore also be different.

StabilizersRaw polymers are stabilized to preventthermal and oxidative aging. In the DSC

method, the effect of the stabilizer is as-sessed by measuring the sample in an oxi-dative atmosphere (oxygen or air as purgegas). In particular, the beginning of thethermal oxidative degradation of the poly-mer is measured. Figure 4 shows thebehavior of four polypropylene (PP) sam-ples containing different concentrations ofthermostabilizers. The decomposition ofthe PP sample without stabilizer beginsimmediately after melting is completed.With larger stabilizer contents, the onsetof decomposition is shifted to higher tem-peratures. A stabilizer content of 2% ormore ensures that the molding material isadequately stabilized with regard to aging.A closer examination of the curves showsthe presence of a small peak at 116 °C.

differences in molecular structure influ-ence the melting behavior of the materials.In this particular case, the melting tem-perature of the copolymer is about 10 Klower than that of the homopolymer.

Differences between batchesPlastic molding materials are usually pro-duced in batches. Individual batches maydiffer because of variations in the polymersynthesis and in compounding. Such dif-ference can be measured by DSC andevaluated. For example, Figure 3 shows theDSC heating and cooling curves of threedifferent batches of the same standard typeof POM (POM black). Significant differ-ences in crystallization behavior can beseen in the cooling curves, particularly in

Fig. 1. DSC heating curves (second measurement) of several different polymers:granules; heating rate 20 K/min; purge gas N2.

Fig. 3. Cooling curves followed by (second) heating curves of different batches ofPOM (Hostaform C13021 black). Cooling and heating rates 20 K/min; purgegas N2; sample weight 5.0 ± 0.1 mg).

Fig. 2. Melting curves of samples of original POM homopolymer (Delrin 500 Clnatur) and original POM copolymer (Hostaform C13021 natur). Heating rate20 K/min; purge gas N2; sample weight 5.0 ± 0.1 mg.

Fig. 4. Oxidation Onset Temperature (OOT) of PP containing different amountsof stabilizer. The inserted diagram displays the melting region of the stabilizeron an expanded scale. Heating rate 20 K/min; purge gas O2; sample weight5.7 ± 0.1 mg.

UserCom 1/200418

This can be assigned to the melting of thethermostabilizer. The peak is displayed onan expanded scale in the inserted diagramin Figure 4.

ColorColored molded parts are manufacturedeither from bulk-colored molding materialor by adding a standard color concentrate,a so-called color masterbatch, to the orig-inal uncolored molding material. Thebase material of such color masterbatchesis usually a resin formulation or a polyo-lefine. Processing and technical problemscan arise if the melting range of the mas-terbatch and that of the molding materialto be colored are too far apart. To obtainmolded parts with optimum properties,e.g. excellent visual homogeneity andprocessibility, masterbatches based on apolymer matrix should be used for coloringthe molding material. The polymer matrixmust be compatible with the molding ma-terial and melt in the same temperaturerange otherwise problems arise when themolding material is processed.

DSC measurements can be used to char-acterize the masterbatch and to estimatethe temperature range for processing themolding material. Figure 5 shows themelting curves of a reinforced PP moldingmaterial and the color masterbatch. Themasterbatch has already completely meltedby 118 °C, while the PP molding materialonly just begins to melt at this tempera-ture. The melting temperatures in fact dif-fer by 48 K, which can lead to processingproblems. The melting peak temperatureof 117 °C for the granulated coloring ma-

terial is typical for a color masterbatch.Carbon black is often used to render mate-rials black. Figure 5 shows the DSC curvesof unreinforced polybutylene terephthalate(PBT) molding material with and withoutcarbon black. On cooling, the blackmolding material crystallizes with a peakmaximum at 189 °C, whereas the originaluncolored molding material has a maxi-mum at 167 °C. The reason for this is theincreased crystallization rate due to the ef-fect of the carbon black particles acting ascenters for nucleation. This leads to a finercrystalline structure and results in plasticmolded parts whose properties are differentto those formed from the uncolored poly-mer melt under the same cooling condi-tions. The difference in structure is alsothe reason for the difference in meltingbehavior. The second heating run of theblack sample exhibits two peaks with the

first maximum at about 212 °C. This rela-tively low melting temperature indicatesthat small crystallites are formed due tothe large number of nucleation centers.

ConclusionsDSC is an excellent method for the qualitycontrol and characterization of incomingmaterials. The analysis of glass transitions,melting and crystallization processes, andchemical reactions provides important in-formation on the identity and stability ofthe molding materials as well as the influ-ence of different types of additives.

Comprehensive information on the samplepreparation and application possibilities ofDSC in plastics technology is given in theDSC technical manual. This book can beobtained from Prof. Dr.-Ing. Achim Frickat the Fachhochschule Aalen.

Fig. 5: DSC curves (second heating runs) of differently colored molding materials:a) melting curve of the PP-GF30 polymer and the color masterbatch; b) coolingand second heating curves of differently colored PBT, Ultradur B4520. Heating/cooling rate 20 K/min; purge gas N2; sample weight 3.5 ± 0.1 mg).

19UserCom 1/2004

Dates

Exhibitions, Conferences and Seminars - Veranstaltungen, Konferenzen und SeminareAIMAT June 29-July 2, 2004 Ancona (Italy)Calorimetry & Thermal Effects in Catalysis July 6-9, 2004 Lyon (France)5th International Conf. on Polymer-solvent Complexes & Intercalates July 11-13, 2004 Lorient (France)SLAP 2004 July 11-16, 2004 Valencia (Spain)ACS August 23-25, 2004 Philadelphia, Pennsylvania, (USA)ICTAC September 12-19, 2004 Chia Laguna, Sardinia (Italy)NATAS October 4-6, 2004 Williamsburg, Virginia, (USA)Het Instrument 2004 November 1-5, 2004 Utrecht (Netherlands)

TA Customer Courses and Seminars in Switzerland - Information and Course Registration:TA-Kundenkurse und Seminare in der Schweiz - Auskunft und Anmeldung bei:Frau Esther Andreato, Mettler-Toledo, Analytical, Schwerzenbach, Tel: ++41 1 806 73 57, Fax: ++41 1 806 72 40, e-mail: esther.andreato@mt.comCourses / KurseTMA/DMA Basic/SW Basic (Deutsch) 20. September, 2004 TMA/DMA Basic/SW Basic (English) September 27, 2004TGA/DMA Advanced (Deutsch) 21. September, 2004 TGA/DMA Advanced (English) September 28, 2004DSC Basic (Deutsch) 22. September, 2004 DSC Basic (English) September 29, 2004DSC Advanced (Deutsch) 23. September, 2004 DSC Advanced (English) September 30, 2004SW Advanced (Deutsch) 24. September, 2004 SW Advanced (English) October 1, 2004

TA-Kundenkurse und Seminare in DeutschlandFür nähere Informationen wenden Sie sich bitte an: Frau Inga Kuhn, Mettler-Toledo GmbH, Giessen, Tel: ++49 641 507 404,e-mail: inga.kuhn@mt.com

Fachseminar TA-Methoden/Kunststoffe 22.06.2004 Frankfurt/M.Fachseminar TA-Methoden/Pharma 24.06.2004 MünchenFachseminar „Dynamisch Mechanische Analyse – die Zukunft fürinnovative Materialentwicklung» 01.07.2004 Frankfurt/M.Branchenworkshop „TA-Methoden in der Pharmazeutik» 14/15.10. 2004 GiessenBranchenworkshop «TA&IR an Kunststoffen» 20/21.10.2004 GiessenWeitere Informationen zu diesen Veranstaltungen finden Sei unter: www.labtalk.de

TA-Kundenkurse und Informationstage in ÖsterreichAnmeldung und nähere Informationen: Frau Geraldine Braun, Tel. +43-1-604 19 80 - 48, e-mail: geraldine.braun@mt.com

TA Semináre v Ceské RepublicePro blizsí informace prosím kontaktujte: Helena Beránková, Mettler-Toledo Praha, Tel: +420 272 123 152, e-mail: helena.berankova@mt.com

TA szemináriumok MagyarországonRészletekkel kapcsolatban szívesen adunk tájékoztatást a Mettler-Toledo magyarországi irodájában.Tel: +36 1 288-4040, Fax: +36 1 288-4050, e-mail: ilona.matus@mt.com

Cours et séminaires d’Analyse Thermique en FranceRenseignements et inscriptions parChristine Fauvarque, Mettler-Toledo S.A., 18-20 Av. de la pépinière, 78222 Viroflay Cedex, Tél: ++33 1 3097 1439, Fax: ++33 1 3097 1660Cours clients (2ème session)DMA/TMA et le logiciel STARe 4 octobre 2004 Viroflay (France) TGA et logiciel STARe 7 octobre 2004 Viroflay (France)DSC les bases et logiciel STARe 5 octobre 2004 Viroflay (France) Logiciel STARe 8 octobre 2004 Viroflay (France)DSC avancé et logiciel STARe 6 octobre 2004 Viroflay (France)

Seminars for Thermal Analysis in BelgiumFor more details, please contact Philippe Larbanois, N.V. Mettler-Toledo S.A., Zaventem, Tél: ++32 2 334 02 11, Fax: ++32 2 334 03 34.e-mail: philippe.larbanois@mt.com

TA Customer Courses and Seminars in the NetherlandsVoor verdere informatie kunt U kontakt opnemen met: Hay Berden of Ko Schaap, Mettler-Toledo B.V., Tiel, Tel: ++31 344 63 83 63

UserCom 1/200420

Corsi e Seminari di Analisi Termica in ItaliaPer ulteriori informazioni Vi preghiamo di contattare:Marina Varallo, Mettler-Toledo S.p.A., Novate Milanese, Tel: ++39 02 333 321, Fax: ++39 02 356 2973, e-mail: marina.varallo@mt.com

DSC base 28 Settembre 2004 Novate Milanese TGA 30 Settembre 2004 Novate MilaneseDSC avanzato 29 Settembre 2004 Novate Milanese TMA 1 Ottobre 2004 Novate Milanese

Cursos y Seminarios de TA en EspañaPara detalles acerca de los cursos y seminarios, por favor, contacte con:Francesc Catala, Mettler-Toledo S.A.E., Tel: ++34 93 223 76 00, e-mail: francesc.catala@mt.com

Seminario de aplicaciones TA: Septiembre 21, 2004 Madrid Seminario de aplicaciones TA: Septiembre 28, 2004 BarcelonaSeminario para usuarios STARe: Septiembre 22, 2004 Madrid Seminario para usuarios STARe: Septiembre 29, 2004 Barcelona

TA Customer Courses and Seminars for Sweden and the Nordic countriesFor details of training courses and seminars, please contact:Fredrik Einarsson, Mettler-Toledo AB, Tel: ++46 455 30 0080, Fax: ++46 8 642 45 62, e-mail: fredrik.einarsson@mt.com

TA Customer Courses and Seminars in the UKFor details of training courses and seminars, please contact:Rod Bottom, Mettler-Toledo Ltd, Leicester, Tel: ++44 116 234 5025, Fax: ++44 116 236 5500

DSC Basic 26 October, 2004 Leicester DSC Basic 27 October, 2004 Leicester

TA Customer Courses and Seminars in the USA and CanadaBasic Thermal Analysis Training based on the STARe System is being offered at various locations.For information, please contact: Tom Basalik at +1 614 438 4687, Fax: +1 614 438 4693 or by e-mail to: tom.basalik@mt.com

Curve Interpretation September 2, 2004 Blue Ash, OH Curve Interpretation September 29, 2004 Durham, NCSample Preparation September 2, 2004 Blue Ash, OH Sample Preparation September 29, 2004 Durham, NCSTARe User Training September 21-22, 2004 Columbus, OH Automotive Analysis by DMA October 20, 2004 Southfield, MI

Sample Preparation October 20, 2004 Southfield, MI

TA Customer Courses in the South East Asia Regional Office, Kuala LumpurFor information on dates, please contact:Malaysia: Ricky Tan at ++603 78455773, Fax: 603 78458773 Thailand: W. Techakasembundit at ++662 7230336, Fax: 662 7196479Singapore: Joslyn Yeo at ++65 8900011, Fax: 65 8900013 or SEA regional office: Soosay P. at ++603 78455373, Fax: 603 78453478

TA Seminars in TaiwanFor details of seminars please contact: Kimmy Wu at Mettler-Toledo Taiwan, Tel: ++886 2 6578898, Fax: ++886 2 26570776,e-mail to kimmy.wu@mt.com and visit our home page at http://www.tw.mt.com

TA Customer Courses and Seminars in JapanFor details of training courses and seminars please contact:Yasushi Ikeda at Mettler-Toledo Japan, Tel: ++81 3 5762 0746, Fax: ++81 3 5762 0758, or by e-mail to yasushi.ikeda@mt.com

TA/FP Workshop August 3, 2004 Tokyo Service Center TA/FP Workshop August 5, 2004 Osaka Service CenterInfoday Seminar October 19, 2004 Tokyo Technical Center Infoday Seminar October 21, 2004 Osaka Branch

Cursos y Seminarios de TA en Latin AmericaPara detalles acerca de los cursos y seminarios, por favor, contacte con:Francesc Català, Mettler-Toledo, LA, Tel: ++34-932 237 615 (Spain), e-mail: francesc.catala@mt.com

Editorial team

Dr. J. Schawe, Dr. R. Riesen, J. Widmann, Dr. M. Schubnell, C. Darribère, Dr. M. Wagner, Dr. D. P. May, Ni Jing, Urs Jörimann,Physicist Chem. Engineer Chem. Engineer Physicist Chem. Engineer Chemist Chemist Chemist Electr. Engineer

METTLER-TOLEDO GmbH, Analytical, Sonnenbergstrasse 74, CH-8603 Schwerzenbach, SwitzerlandKontakt: urs.joerimann@mt.com, Tel: ++41 1 806 73 87, Fax: ++41 1 806 72 60Internet: http:/www.mt.com/ta

© 05/2004 Mettler-Toledo GmbH, ME-51724379, Printed in Switzerland