UserCom6

USER COM December 97 1

USER COM6

Information for users ofMETTLER TOLEDO thermal analysis systems December 1997

TA TIP1.1 General observationsA calibration shows you whether your module supplies you with correctmeasured values. Depending on the number of sensors in your instrument,you have to perform different calibrations (heat flux, temperature, mass,force and length).

Adjustment: Change of instrument parametersCalibration: Determination of the deviationReference substance: Substance with known literature values

needed for the calibrationLimits of permissible Permissible deviation from desirederror: value

All STARe calibration methods can be used for the calibration and the adjust-ment. The decision whether the determined deviation should be corrected ornot always rests with you. If you press CANCEL in the determination of thedeviation, you leave things as they are. On the other hand, with OK the oldadjustment data are replaced by the values you have just determined.

1.2 Prior considerationsBefore the first calibration, thought should be given to the following threepoints:1. Limits of permissible error2. Calibration interval3. Possible measurement combinations (module, crucible and gas)

1.2.1 Limits of permissible errorIn the determination of the limits of permissible error and the calibration in-terval, one should abide by the principle "as accurately and as frequently asnecessary".Starting from your (validated) methods, you can define the limits of permis-sible error appropriate to your situation. The adjustment and calibrationmethods available in the database of METTLER TOLEDO are based on the

Dear customer

Yet another successful year forMettler Toledo is drawing to aclose. In the field of thermalanalysis once again we succeededin launching many new productsin 1997.We would like to take this oppor-tunity to thank you for the numer-ous recommendations whichcontinue to help us optimize ourproducts.The Mettler-Toledo TA team wishyou a successful 1998.

Calibration

Contents

TA TiP:– Calibration

New in the sales program:– TMA/SDTA840– STARe SW V5.1– Food application handbook– Pharmaceutical application

handbook

Applications– Modern building materials (FTIR)– Selection of MaxRes parameters– Investigation of the memory effect

of polyethylenes– Investigation of copolymers with

DSC30– Denaturation of proteins– Thermoanalytical investigation of

hydrate– Phase correction in ADSC

measurements in glass transition– ADSC during the glass transition

2 USER COM December 97

STARe instrument specifications andthus have very narrow limits of per-missible error.

1.2.2 Calibration intervalWith regard to the calibration inter-val, we recommend a monthly testbased on Fig. 1: Flow chart of meas-urements. If the test results are cor-rect 2 months in succession, the cali-bration interval can be doubled, oth-erwise it must be halved.If you have very exacting demands,you can perform each measurementbetween two calibrations

1.2.3 Measurement combinationFor each module, crucible and purgegas combination (= measurement

combination), a set of adjustmentdata can be stored in the database.The correct adjustment data appro-priate to the experimental conditionsat hand are transferred to the modulein the measurement so that youmeasure correctly even if the gas isswitched from helium to oxygen.

1.2.4 Adapting the adjustmentand calibration methods

As there are innumerable measure-ment combinations, adjustment andcalibration methods are stored in thedatabase as examples only for thestandard case (STARe module, stand-ard aluminum crucible 40 µl (DSC),aluminum oxide crucible 70 µl(TGA) and air).

You must adapt these methods toyour measurement combination(s)1. Ensure that your module can at-

tain the defined heating andcooling rates, otherwise youmust adapt the rates

2. In the case of crucibles with arelatively large mass and/or rela-tively poor heat conduction, thetemperature range must be in-creased as the stabilization timeis longer

3. The crucible type must bechanged if you do not measurewith the standard crucibles

4. The reference substances shouldbe selected so that they lie in thetemperature range you are inter-ested in (extrapolation has an ad-verse effect on the accuracy ofthe results)

5. Select the correct purge gas.To achieve the greatest accuracy, allmeasurement combinations usedlater should be adjusted beforehand.Without an adjustment, the adjust-ment data for a nonadjusted meas-urement combination are extrapo-lated from the standard adjustmentdata and determined factors, but thesame high measurement accuracycan not be achieved by this means.

1.3 Adjustment procedureIn the adjustment we recommendthe following sequence:1. Tau lag adjustment2. Temperature adjustment (neces-

sary only for your standard mea-surement combination)The furnace sensor (temperatureadjustment) is not influenced bythe purge gas and the crucibleand need not therefore be read-justed for every combination

3. Sensor adjustment (heat flux,sample temperature, force orlength)

The "total adjustment" makes allthree adjustments (tau lag, tempera-ture and sensor) using a single meas-urement thus saving you a great dealof time.

Fig. 1: Flow chart of measuremnets

Adapting the adjustmentand calibration methods

Determination of thelimits of permissible error

Determination of themeasurement combination(s)

Determination of thecalibration interval

Adjustment

Calibration

Measurements

Calibration

Not OK

OK


been previously adjusted, the selec-tion of the heating rate is no longerimportant here (the only restriction isthat this should not be less or greaterthan the heating rates selected for thetau lag adjustment).Again we recommend an adjustmentwith at least two substances with dif-ferent onset temperatures.

1.3.3 Sensor adjustmentThe sensor adjustment corrects thesignal value (DSC: heat flux, TMA:length, SDTA: sample temperature).This is another case where it is ad-visable to perform the adjustmentwith at least two substances with dif-ferent literature values to attain ahigh accuracy in the temperaturerange of interest.

1.3.4 Total adjustmentWith the total adjustment we offeryou the possibility to make all ad-

justments mentioned above with asingle measurement.Attention must be paid to the follow-ing:1. The reference substances must

not exhibit any mutual interfer-ence (local separation).

2. For the TMA40 there is no totaladjustment as we know of no

1.3.1 Tau lag adjustmentWith the tau lag adjustment the dy-namic behavior of the measuring cellis corrected. Two simple controlmeasurements suffice to determinewhether a tau lag adjustment is nec-essary. Determine the onset tempera-ture of indium with two differentheating rates of 5 and 20 K/min. Ifthe onset temperatures are different,which is physically not possible, atau lag adjustment is needed.We recommend performing this ad-justment with at least two differentsubstances with different onset tem-peratures which lie in the tempera-ture range of interest. A module withcorrect tau lag values shows no de-pendence of the temperature on theheating rate

1.3.2 Temperature adjustmentThe temperature adjustment ensuresthat the onset temperatures corre-

spond exactly to the literature values.Depending on the literature data, thevalues vary slightly but the values inthe database can be adjusted veryeasily as long as the substance hasnever been used. If this is not thecase, the substance can be stored un-der a new name with a new value.As the dynamics of the module have

material that generates reversiblelength changes and is recognizedas a reference substance.

3. For the DSC modules the heatflux measurement is performedonly on the basis of the sub-stance with the lowest meltingtemperature (entry of the samplemass possible only for one sub-stance)

4. You should use the total adjust-ment only for your standardmeasurement combination

1.4 Calibration(dynamic case)

Select a heating rate in the region ofthe tau lag adjustment and measure aknown substance. After the auto-matic or manual evaluation, youknow whether or not you meet therequirements you have defined re-garding temperature or signal size.

1.5 Calibration(isothermal case)

With the isothermal step method youcan determine the isothermal tem-perature very accurately. If the iso-thermal temperature is not correct, itcan be used for the temperature ad-justment. Without a correct tau lagadjustment, however, it is never pos-sible to measure accurate tempera-tures in either dynamic or isothermalmeasurements.Care is also required when the mod-ule is used for both isothermal anddynamic methods. In this case, it isessential to make a tau lag adjust-ment beforehand.Selection of the step (in °C) in theisothermal step gives the temperatureaccuracy.Example of a test at 200 °C:1. Isostep method for indium (with

steps of 0.1 °C)2. Isostep method for tin (with

steps of 0.1 °C)3. Draw diagram as follows: Iso-

thermal temperature deviationsas a function of temperature

Fig. 2: Shift in the onset temperature owing to wrong tau lag values


Fig. 3: Onset temperatures after the tau lag adjust-ment independent of the heating rate, but still withwrong absolute temperature and wrong heat flux

Fig. 4: Onset temperatures after the tau lag adjust-ment, but still with wrong heat flux

Fig. 5: Fully adjusted DSC module Fig. 6: Isostep measurement for indium and tin for anadjusted module

Fig. 7: Isostep measurement for indium and tin for a mod-ule without tau lag adjustment


4. The temperature deviation at200°C can now be taken fromthe diagram; in the measured ex-amples (with tau lag adjustment:0.0 °C, without tau lag adjust-ment 1.6 °C).

As an example of a module withoutor with a wrong tau lag adjustment

1.6 Reference substancesThe reference substances needed for the calibration and adjustment are marketed by METTLER TOLEDO. The fol-lowing reference substances are available and can be traced to the producer:Indium In 156.6 °C 28.5 J/g ME 119442Tin Sn 231.9 °C 60.1 J/g ME 51140621Lead Pb 327.5 °C 23.0 J/g ME 650013Zinc Zn 419.6 °C 107.5 J/g ME 119441Aluminum Al 660.3 °C 397.0 J/g ME 51119701

The following are also available for the TMA40 and TG50 measuring cell, but are without a producer's certificate:

Aluminum AI 660.3 °C ME 29593Silver Ag 961.8 °C ME 29594Isatherm(1) NiMn3AI 144.5 °C ME 29800Nickel Ni 357.0 °C ME 29799Trafoperm(2) Fe 748.0 °C ME 29798

(1) Isatherm minus (Ni (94 %), Mn (1-3 %), AI (1-2%), Si (1-2 %), Fe (0.3 %), Mg (0.15 %), Cu (0.1 %),C (0.05 %))

(2) Trafoperm C5 (Fe (95 %), Si ( 3.5-4.5 %, Mn (<0.02 %), Cu(<0.2 %), C (<0.07 %), Cr (<0.06 %),P (0.03 %), S (0.02 %))

If you require certified reference substances which can be traced to a national or international standard, you can ob-tain these from LGC, for example:

LGCThe Office of Reference MaterialsLaboratory of the Government ChemistQueens RoadTeddingtonMiddlesex, TW11 OLY, UK

Tel: +44 (0) 181 943 75 65Fax: +44 (0) 181 943 75 54E-mail: [email protected]

and subsequent, dynamic tempera-ture adjustment the isostep measure-ments were repeated.It is clear that large isothermal devia-tions appear although the tempera-ture adjustment for the module wascorrect.

Based on the error diagram, the iso-thermal temperature can now becorrected for a desired temperaturefor this case. A tau lag adjustment ispreferable to such a correction to en-sure the module measures correctlyin the dynamic and in the isothermalcase.

Fig. 8: Isothermal temperature deviation as a function of temperature

Temperature


TMA/SDTA840

We are pleased to be able to intro-duce you to our new TMA/SDTA840module less than a year after thelaunch of the TGA/SDTA851e. Thismodule replaces the TMA40 measur-ing cell, which has been and willcontinue to be used for years in hun-dreds of labs.

SDTA signalAnalogous to the TGA/SDTA851e

module, we now measure the sampletemperature right at the sample. TheR type thermocouple is protectedagainst contamination by a thinquartz glass coating. This now al-lows an adjustment with lengthchange and/or with melting points ofpure metals (your own choice).In addition, the DTA signal can becalculated. Analogous to TGA, si-multaneously with the length changein the DTA signal you have anothermeasured quantity available whichcan facilitate your interpretation.

Gas-tight measuring cellThe measuring cell is gas tight sothat measurements can be performedunder a defined atmosphere.

ThermostatingThe mechanical part of the measur-ing cell is accommodated in athermostated housing thus guaran-teeing the highest accuracy in the de-termination of the coefficients of ex-pansion.

Length measurementThe entire electronics have been up-graded to state of the art. The rangeswitching in the length measurementis a thing of the past

AutomationThe furnace can be opened andclosed at the press of a button mak-ing the operation simpler and moredependable.

Features of the TMA/SDTA840 module at a glance

Large measurement range ± 5.0 mmHigh resolution 10 nmLong sample lengths up to 20 mmLarge force range -0.1 ... 1 NHigh force resolution 1.3 mNLarge temperature range RT ... 1100 °C (*)High temperature accuracy ± 0.25 °CSDTA resolution 0.005 °CAutomation Automatic opening and closing of the

furnace by pressing a buttonDLTMA mode < 1 Hz (user definable up to 1 Hz)Options Gas controller, MS/FTIR coupling,

switched line socketCooling Cryostat cooling

(*) Low temperature option in preparation

Fig 9: The new TMA/SDTA840


Software V5.1

The STARe SW V5.1 replaces the STARe SW V5.0. The new software version incorporates control of the new TMA/SDTA840 module.In addition, a few of your wishes have again been fulfilled with this software release:• MaxRes for TMA curves (SW Option MaxRes)• ADSC phase correction (SW Option ADSC)• Configurable "segment separation ..." evaluation• E module evaluation (SW Option TMA)• X-Y signal value transferrable to diagram• Multiplication and division with curves (SW Option Mathematics)• Logarithmic representation• Superimposable coordinate network• Configurable data export (selectable number of data points)• Selectable line type• Several databases (one must be selected on startup)

Food application handbookThe new food application handbookhas been produced in cooperationwith Dr Behlau from the FraunhoferInstitute for Food Technology andPackaging in Freising. Using 20 in-teresting examples, it shows howthermal analysis can be used in the

food industry and supplements thecollection of application handbooksfrom Mettler-Toledo (German51725003, English 51725004).

Pharmaceutical applicationhandbookTogether with Dr Pffeffer, Dr Giron

and Ms Schwarz from Novartis (exSandoz), we have produced an appli-cation handbook for the pharmaceu-tical industry with over 40 examples.Examples from actual practice areshown in which thermal analysis cansupply valuable information (Ger-man 51725005, English 51725006).

Fig. 10: E module calculation

Minor improvements have also been implemented such as the insertion of arrows, copy/paste of texts, marker shift-ing using the keypad and many more.


qualitative nature of the substanceinvolved can seldom be answered.Additional information and compari-son measured curves of known sys-tems would be necessary for this.However, the nature of the gasesevolved in the decomposition proc-esses allows deductions regardingthe type of substance which has re-leased them. This also permits quali-tative compositions to be deter-mined.

The following example shows thethermogravimetric analysis of amodified tile adhesive with coupledinfrared spectrometry of the evolvedgases.The measurement module TGA/SDTA851e is connected to a FTIRspectrometer via an electricallyheated copper transfer line with aninner glass coating.The sample vapors are transferred bythe carrier gas (here nitrogen with

approx. 100 ml/min) to a heatedglass cell with KBr windows in theray path of the spectrometer.The timing of the measurements ofthe two instruments is electricallysynchronized so that time-dependentfunctions such as chemigrams andGram-Schmidt curves can be com-bined in real time with the TGcurves. The data of the FTIRspectrometer are stored in theJCAMP format using its own soft-

ware and can then be read into theSTARe software.Figure 11 first shows the experimen-tal curve in the temperature range25 °C to 1000 °C at a heating rate of20 K/min. Several step-like masslosses can be seen here which stemin part from water losses from hy-drate salts and in part from the de-composition processes of the poly-mers and carbon dioxide loss fromcarbonates. The time-differentiated

Modern building materials

Today's building materials such astile adhesives not only comprise theclassical ingredients such as cementand additives - sand, lime, etc. - butalso a complex mixture of polymersin the form of dispersions. Thesepolymer additives provide the build-ing materials with the required prop-erties in regard to elasticity, adhesivepower, frost resistance, etc.Such complex formulas also requirea corresponding number of qualitycontrols during the production. Inaddition, the increasing competitionin this sector necessitates pressingahead with further development andcomparing rival products with one'sown.In addition to the classical mechani-cal and chemical test methods, ther-mal analysis including primarilythermogravimetry has gained greatimportance and significant applica-tion possibilities. The advantage ofthis method is that a type of "finger-print" of the substance mixture underinvestigation is obtained in very littletime and without the use of chemi-cals. This is in the form of an experi-mental curve which describes themass change as a function of tem-perature. Quantitative percentages ofthe individual constituents can thusbe evaluated. In the following exam-ple, the mass decrease, e.g. between300 °C and 500 °C corresponds tothe decomposition of the polymerfraction and hence its content. Fromthe mass loss between 550 °C and750 °C, the loss of carbon dioxidefrom calcium carbonate, the contentof the latter can be calculated.While thermogravimetry by itselfprovides quantitative content infor-mation, this is generally group infor-mation; the individual polymer typesin a mixture can usually not be sepa-rated. The question regarding the

Fig. 11: TGA and FTIR curves of a tile adhesive


curve (DTG curve) provides addi-tional clarification of these proc-esses, it is a measure of the masschange rate. In addition, several timefunctions from the IR spectra are re-corded. The Gram-Schmidt diagramshows the integral change of thespectra over the entire wavelengthregion. As a result, the temperatureand time regions in which processeswith drastic changes in gas composi-tion occur can be identified. It can beeasily seen how this curve reflectsthe mass changes (DTG curve).The chemigrams provide analogousinformation, but in a greatly re-stricted wavelength region in which,for example, typical absorptionbands of assumed decompositiongases appear. In this manner, the ap-pearance of particular substances canbe shown during the measurement.The example shows a chemigram be-tween 2765 and 2646 cm-1 as well asone between 1817 and 1753 cm-1.The first represents the appearanceof a typical absorption band of ana,b-unsaturated aldehyde, the secondthe carbonyl bands of ketones or es-ters.The IR spectra stored during themeasurement time can be comparedwith those in a database of IR spec-tra of substances in the gas phase.Using the example of the spectrumafter 31 minutes measurement time(first significant mass change step),Fig. 12: FTIR spectrum after 31 minutes

2-butenal (crotonaldehyde) wasfound. Its spectrum together with themeasured spectrum is shown in Fig-ure 12. The other spectra were as-signed to the following compoundsor compound classes in the samemanner: 42 minutes: acetic acid; 50minutes: acetone; 59 minutes hydro-carbon mixture (including methane);90 minutes: carbon dioxide; 115minutes: sulfur dioxide . The twolast mentioned gases probably arisefrom additives containing carbonateand sulfate. The assignments men-tioned can naturally be assessed onlyin a limited sense as the "hit rate" de-

pends on the one hand on the size ofthe spectra database and on the otherhand on the homogeneity of thetransferred gases. While accompany-ing gases such as water vapor or CO

2

among others may be removed fromthe spectra by software, they mayhave an adverse effect on the de-pendability that the most suitablecompounds have been traced.From the fragments found in thismanner, the substances in the samplematrix can now be reconstructed.Their amount can then be calculatedfrom the TG mass losses.


MaxRes: Event-controlled adaption ofthe heating rate

SummaryToday, most thermoanalytical meas-urements are performed withpredefined constant heating rates.Despite the great success and wideapplication of this method, it hasbeen known for forty years thatconcepts such as thermal effects of asample exist which must be enlistedin the determination of the experi-mental conditions [1 - 4], i.e. par-ticularly how the heating rate can beadapted during the analysis.This article describes the basic dif-ferences between these two methodswith constant or effect-dependentheating rate, their specific advan-tages and disadvantages and offerstips on the selection of valid experi-mental parameters using the exampleof the MaxRes method inthermogravimetry [5].

IntroductionIn the determination of the condi-tions for a conventional thermalanalysis, the user is forced to decidebetween two different, frequentlycontradictory goals (experiment timeor resolution). The selection of theheating rate is a determining factor:while a high heating rate leads to anamplification of an effect and toshort experiment times, it can alsoresult in an unsatisfactory separationof thermal events; a low heating rateallows the separation of thermal ef-

fects in close proximity, but leads tolong experiment times.Although modern instruments with ahigh sensitivity and a quick-responsefurnace have reduced the timeneeded for an experiment, high reso-lution measurements always need agreat deal of time.A dynamic switching of the heatingrate as available with MaxRes allowsmeasurements with a clearly higherresolution during the same experi-ment time (see also example 2).

Dr Benedikt Schenker1, Dr Rudolf Riesen21 Technisch-Chemisches Laboratorium, ETH Zurich, CH-8092 Zurich, Switzerland, Tel: +41 (1) 632 30 59,FAX: +41 (1) 632 12 22, E-mail: [email protected] Mettler Toledo, CH-8603 Schwerzenbach.

Selection of the methodIn the decision regarding whether ameasurement should be performedwith a constant or effect-dependentheating rate, the following pointsmust be considered:• Time available• Simultaneous DTA measurement

necessary or not?• Is the mass change and/or the tem-

perature of a particular event of in-terest?

• Repeat measurements necessary?

Table 1 shows an overview of the advantages and disadvantages of bothmethods.

Constant heating rate: Effect-dependent heating rate(MaxRes):

Adv.: Simple parameter selection Increased resolutionDTG allows step separation Time savingSimultaneous DTA meas. possible Quasi-equilibrium conditionsSuitable for curve comparisons during the effectsEvaluations with kinetic modelsSubtraction of baseline possible

Disad.: Insufficient step separation Simultaneous DTA measure-with high heating rates ment not interpretableTime consuming with low Change in heating rate canheating rates simulate a step

No subtraction of baseline

Table 1: Advantages and disadvantages of constant and effect-dependentheating rate


MaxRes - algorithmA MaxRes experiment starts in ex-actly the same manner as a conven-tional experiment. The furnace isheated at a constant heating rate.However, when a signal change ap-pears and its derivative exceeds acertain amount (high thresholdvalue), the heating rate is reduced.As soon as the derivative drops be-low a different, defined amount (lowthreshold value), the heating rate israised again. This algorithm allows ahigh heating rate to be used in re-gions in which nothing changes anda low heating rate to be employed inregions in which a transformationtakes place.An initial measurement with a large,constant heating rate should never be

Example 1Figure 13 shows the decompositionof calcium oxalate monohydrate(CaC

2O

4 * H

2O). The solid lines

show the set of curves which weremeasured with the parameters in thevalue range of Table 2. It is clearlyapparent that all parameter sets usedseparate the various steps (decompo-sitions) very well. The different stepheights and the different transforma-tion temperatures can be determinedwith high accuracy.The curve which was measured witha high heating rate is shown as adashed line.The constant heating rate was se-lected so that this experiment lastedtwice as long as the fastest of theMaxRes experiments shown. Despite

omitted. This measurement allowsthe selection of optimum MaxResparameters. In addition, from the de-rivative of the signal the number ofeffects can be found and whether ornot these overlap. It should be notedthat the MaxRes measurements cannever be corrected by the baseline.In the case of low steps whichstretch over a large temperaturerange, the error can already lead toinaccurate results owing to the den-sity change during heating of the fur-nace atmosphere (Archimedes).

double the experiment time, thesteps are clearly less well separated(poorer resolution).

Example 2The thermal decomposition of car-bonates takes place over a widetemperature range even if a rela-tively low heating rate is used. AsFigure 14 shows, the weight losssteps with a mixture of strontiumand barium carbonates overlap ex-tensively even at the selected heat-ing rate of 12 K/min. With a

MaxRes measurement the two stepsare clearly separated and within thesame amount of time available. Inboth measurements, the total weightloss corresponds to the theoreticalvalue for the sample weights used. Inthe MaxRes experiment, 21.032 mgstrontium carbonate and 20.136 mgbarium carbonate were mixed andheated from 500 to 1600 °C, result-ing in a weight loss of 10.75 mgCO

2. The CO

2 partial pressure of the

incipient BaCO3

decomposition de-layed the SrCO

3 decomposition

somewhat (appreciably less in theMaxRes measurement). As a result,the measurement shows the firstweight step is too small comparedwith the theoretical value (28.14 %instead of 29.8 %) and that of the se–

second step correspondingly toolarge (24.0 % instead of 22.3 %).Parameters used : Standard settingsat a maximum and minimum heatingrate of 40 and 1.5 K/min.

ConclusionsMaxRes allows measurements ofsamples with overlapping thermaleffects with a higher resolution andwithin a considerably shorter time.Thanks to better separation, theevaluation and interpretation iseasier and more dependable.

Recommended parametersFor MaxRes experiments the parameters listed in Table 2 are recommended, but for special analyses primarily achange in the heating rate and secondly the threshold values should be effected.

Parameter Recommended value Minimum recommended Maximum recommend (standard setting) value value

Min. heating rate 1 K/min 0.1 K/min 5 K/minMax. heating rate 20 K/min 5 K/min 50 K/minUpper threshold value 3 µg/s 1 µg/s 20 µg/sLower threshold value 1 µg/s 0.33 µg/s 7 µg/sFactor 2 2 2Time delay 0.3 min 0.1 min 0.5 minFilter 10 s 5 s 30 s

Table 2: Recommended parameters: The ratio of the upper to the lower threshold value should be at least threeto one.


References[1] L. Edrey, F. Paulik, and J.

Paulik, Hungarian Patent No.152 197 (1962)

[2] J. Rouquerol, Bull. Soc. Chim.Fr., (1964), pp 31-32

[3] F. Paulik, Special Trends inThermal Analysis, John Wiley& Sons, Chichester (1995)

[4] H.G. Wiedemann, D. Nehring,Zeitschrift für anorganischeund allgemeine Chemie, 304(1960) 137

[5] USER COM 4, Information forusers of Mettler Toledo thermalanalysis systems December 1996, page 4

Fig. 13: Decomposition of CaC2O

4 • H

2O.

Fig. 14: TGA curves as well as the associated temperature profiles of thedecomposition of a carbonate mixture, once with event-controlled adaptionof the heating rate (MaxRes) and at constant heating rate (12 K/min) withthe same experiment time.


Investigation of the memory effect ofpolyethylenes

SummaryWhen a PE sample is exposed totemperature in the melting range ofthe crystalline fractions for a certaintime, the crystal structure is changed.This phenomenon is known as thememory effect ("thermal memory ca-pacity") and can be measured usingDSC. The memory effect is mani-fested as a deviation in comparisonwith a DSC curve of a sample with-out a thermal prehistory.The aim of this article is to show thatthese deviations have a reproduciblemagnitude which depends on thetemperature and the treatment time.As a result, it is possible to deter-mine how long a sample has beensubjected to a particular temperature.

IntroductionPartially crystalline PE, which is fre-quently used as an insulation mate-rial for cables, comprises:• Amorphous zones, also known as

the matrix, in which polymerchains are not arranged geometri-cally. The structure of these zonesis equivalent to those of a liquid.

• Crystalline zones in which thepolymer chains have a geometricarrangement (zigzag).

In the melting of these crystals in therange of approx. 50 to approx.130 °C, DSC shows an endothermicpeak. The larger the crystalline frac-tion, the larger this melting peak.A completely crystalline PE has anenthalpy of fusion of 290 J/g, a com-pletely amorphous PE one of 0 J/g.The crystalline zones are alwaysformed when the melted sample iscooled.The crystal formation is promoted byconditioning, e.g. for 20 min. at70 °C . Here, additional crystals witha melting range somewhat above

70 °C are formed. These crystalsproduce an additional peak on theDSC curve on reheating of previ-ously cooled sample. Only condi-tioning within the crystal melting

Fig. 15: PE with and without thermal prehistory

Practical procedureThe principle is relatively simple: The samples are conditioned at differenttemperatures for different lengths of time. They are then measured usingDSC and the deviation evaluated. Relatively small samples of approx. 3 to5 mg mass are used to avoid a smearing of the effects owing to low ther-mal conductivity of the sample.

Measurement program: 30 ... 130 °C at 10 K/minMeasuring cell: DSC30Evaluation: The additional peak is integrated in comparison

with the PE without thermal prehistory, whichserves as a baseline (integral over the subtractedcurve)

Sample: PE low density (PE LE4244) before and aftercrosslinking

It is important to perform the measurements using the same heating rate.

ResultsThe area of the additional peak due to the memory effect increases withincreasing conditioning time and finally reaches a limit value.

range causes this effect. If the treat-ment temperature exceeds the melt-ing range, PE changes to the amor-phous state and its thermal prehis-tory is lost.

L. Pheulpin, Cable Alcatel Suisse SA


ConclusionsIt follows from the representation ofthe heat of fusion of the conditioningpeak against conditioning time that:• the heat of fusion increases with

increasing conditioning time• the heat of fusion approaches a

maximum value asymptotically(equilibrium condition)

• the thermal prehistory is moremarked with crosslinked PE (ex-cept at 90 °C). The exception canperhaps be explained by the factthat after crosslinking the polymerchains are already somewhat or-dered and are better suited forcrystallization.

If the function of the heat of fusionof the conditioning peak against theconditioning time is known, DSC al-lows determination of how long aPE-insulated cable has been exposedto a particular temperature. This ena-bles quality assurance of the thermalaftertreatment of the insulating mate-rial.

Fig. 16: Memory effect of crosslinked PE

Fig. 17: Memory effect of uncrosslinked PE


Investigation of copolymers with DSC30

IntroductionThis article investigates the influenceof additives as well as the oxidativedecomposition on the crystallizationof copolymers. Copolymers are usedfrequently in industry as they arefavorably priced and their propertiescan be very easily changed.In the shaping of thermoplastic, theyare crystallized from the melt bycooling. This process inducesnonisothermal crystallization whichis of great interest for the processing,e.g. by injection molding.The investigated material is a PP-rich polypropylene-polyethylene(PP-PE) copolymer (95 % weightpercent). The two samples called Aand B contain REPSOL PPR 1042(melting index 66.5, melting pointtemperature 162 °C, density 0.903 g/cm3). Various additives were used:• Antioxidants (Tinuvin 327,

Tinuvin 770, Tinuvin 770-DF,Kronos CL 220 and Irganox BZ15)

• UV stabilizers (Quimasorb 144and Chimasorb 944)

• Dyestuffs (Iagacolor 10401, ElfTex 415, Cromoftal A3R,Cromoftal DPP-BO and CinquasiaRRT 891D)

DSC is used to measure the heats offusion and crystallization. The influ-ence of additives on the kinetics ofthe crystallization and on the meltingtemperatures of copolymers was al-ready investigated in earlier work[1].

ResultsAll measurements were performedwith a DSC30 measuring cell underdry air. The samples were broughtfrom the solid to the liquid state at5 K/min and then the thermal prehis-

tory destroyed with an isotherm of5 minutes. Finally, the samples werecooled to room temperature at con-stant cooling rates of 2.5, 5, 10, 20and 40 K/min. All experiments wererepeated at least 3 times.If the sample is heated to just abovethe melting range, the DSC curveshows only the endothermic meltingpeak. After cooling, the sample crys-tallizes exothermically. If heating isappreciably higher, e.g. up to 300 °Ca certain oxidative degradation takesplace which results in the samplesubsequently crystallizing less well.In the extreme case, the material re-mains in the amorphous condition.The lower the maximum tempera-ture, the greater the enthalpy of crys-tallization.

Fig. 18 shows DSC curves with twodifferent maximum temperatures.Kinetic investigations under isother-mal conditions can be performedwith the aid of the Avrami equation.The Avrami equation describes thedependence of the degree of crystal-lization on the crystallization time.In the case of nonisothermal crystal-lization, a modified Avrami equationcan be used [2, 3].The left side of Figure 19 shows thecrystallization peaks of samples Aand B. The conversion curve (crys-talline fraction) represents the basicinformation for the crystallization ki-netics. On the right side of the dia-gram the conversion curves togetherwith the associated tabulated valuesare represented.

Fig.18: Cooling curves of PP copolymer at -10 K/min in dry air.Top: heated to maximum 185°C. The enthalpy of crystallization is 81.5 J/g.Bottom: heated to maximum 240°C, where clearly oxidation takes place.This sample has has a heat of crystallization of only 58 J/g.

J.J. Suñol, J. Saurina, R. Berlanga, J. Farjas, Materials Research Group,Universitat de Girona Av. Lluís Santaló s/n, E-17071, Girona. E-mail: [email protected].


Fig. 19: Crystallization of PP copolymer that has been heated to 185 °C.The cooling rate is -10 K/min. Sample A is the basic polymer, sample Bcontains nucleating agent. Left: DSC curves, right: resulting conversioncurves with the corresponding tabulated values for the crystallized frac-tion.

Sample B (base polymer plus nucle-ating agent) already crystallizes at ahigher temperature than A (basepolymer) thus proving the effective-ness of the nucleating agent.

ConclusionDSC has been used to measure PP-PE copolymers with various addi-tives. Small amounts of additiveshave a considerable influence on thecrystallization temperature. Thecrystallization progress can easily bemeasured with DSC. It is also appar-ent that with increasing oxidativedegradation during conditioning inair, the tendency to crystallize onsubsequent cooling decreases. Ow-ing to the heating, the enthalpy ofcrystallization is increased.

Referenzen[1] J.J. Suñol, J. Saurina,

D. Herreros, P. Pagès,F. Carrasco, „Análisiscalorimétricø de copolímeros debase polipropileno-polietileno“Afinidad 1996, submitted forpuplication.

[2] J.J. Suñol. Thesis, 1996.Universitat Autònoma de Barcelona.

[3] N. Clavaguera, J. Saurina,J. Lheritier, J. Masse, A. Chauvet,M.T. Clavaguera-Mora, „Eutecticmixtures for Pharmaceutical Ap-plications: A Thermodynamic andkinetic study“ Termochimica Acta1997, accepted.


Denaturation of proteins

Fig. 20: DSC measurements of different proteins

In addition to fats and carbohydrates,proteins, long chain molecules com-prising amino acids, are one the mostimportant food components.Depending on the process param-

eters used, their processing may in-volve denaturation. As the extent ofthis denaturation influences the sub-sequent properties such as waterbinding power, emulsifiability, etc.,

Samples commercial raw materials (wheat, spelt, lupin, soya) from a health food store.

Meas. conditions Measuring cell: DSC20Crucible: 40 µl aluminum, hermetically sealed, reference side filled with same

amount of waterSample preparation: grind seeds, aqueous extraction at pH 4.5, centrifuge, discard

supernatant, extract residue again at pH 8.5, supernatant evaporated todesired dry substance content under vacuum.

Sample weight: between 4 – 6 mgMeasurement program: 30 – 110 °C at 5 K/min

Atmosphere: stationary air

it is of great importance for the proc-ess control. As the following exam-ple shows, the denaturation can eas-ily be measured using DSC.

Interpretation4 different, untreated plant proteinswere investigated. All show a char-acteristic, endothermic peak, that ofthe protein denaturation. It lies in the

usual range of 72 to 100°C. Theproteins differ in their thermal stabil-ity and in the enthalpy of denaturation.The enthalpies of denaturation forgrain and oil seed proteins lie in the

range 3 to 10 J/g protein. Theenthalpies are dependent on the cur-rent phase of the proteins; with seedproteins in the original state theenthalpy is less than in the dissolved


Thermoanalytical investigation ofhydrateMany substances used in the pharmaceutical industry have the ability to form so-called hydrates or solvates. Inthese, water is not only present on the surface as moisture, but also bound permanently in the crystal. This propertyis known as pseudopolymorphism and usually leads to a really complex melting behavior being obeserved withsuch hydrates. A combination of DSC and TGA can be used for a complete characterization.

Samples Glucose, monohydrate and anhydrate

Meas. conditions Measuring cells: DSC820 and TGA850Crucible: DSC 40 µl aluminum, pierced lid

TGA 100 µl aluminum, pierced lidSample preparation: as deliveredSample weight: between 5 and 15 mgMeasurement program: DSC 30 – 250 °C at 20 K/min

TGA 30 – 300 °C at 20 K/minAtmosphere: Nitrogen, DSC 50 ml/min and TGA 80 ml/min

state in aqueous solutions. On the other hand, processed protein usually exhibits lower reaction enthalpies as possi-bly parts of the protein are denatured during the treatment.Spelt has temperature sensitive proteins with the highest enthalpy of denaturation Whereas wheat proteins are al-ready extensively denatured at 86 °C, the main fraction of the soya proteins is still native. Here, denaturation doesnot start until 88 °C.

EvaluationSample DS content Weight Reaction enthalpy Tpeak

(DS, mg) (J/gDS) (°C)1) Soya 13.0 % 4.78 5.2 94.92) Lupin 23.0 % 6.80 4.2 90.93) Wheat 12.5 % 3.44 3.5 80.24) Spelt 20.0 % 5.40 17.2 80.7

DS = dry substance

ConclusionThe above measurements show that in principle each product containing protein can be investigated and character-ized using the DSC method. The DSC curve is a type of "fingerprint" of the measured sample. It shows characteris-tically the condition of the protein. On the other hand, identification of plant proteins alone using a single DSCcurve is difficult as the enthalpies can fluctuate as a function of the variety and growth condition. A native samplemust therefore always be measured as a reference to assess the degree of denaturation.

InterpretationGlucose can occur as the monohy-drate or anhydrate, depending on theprevailing conditions. As the TGAand the DSC curves show, there is a

clear difference between the twosamples. The anhydrate shows nonoticeable moisture loss, on the otherhand with the monohydrate a step ofaround 7 % can be observed which

can be ascribed to the loss of hydratewater and does not fully agree withthe stoichiometry.In DSC, the anhydrate shows only itsmelting peak at around 161 °C. In


Bild Gluose, Pseudopolymorismph

Fig. 21: TGA and DSC curve of glucose, present as anhydrate and monohydrate.

the case of the monohydrate, severaleffects, namely water loss, meltingand recrystallization overlap. Thisresults in a broad peak in the rangebetween 60 and 125 °C. Theanhydrate formed in the recrystalli–zation then melts again at around154 °C. This shift is due to the some-what lower purity of the substance.The evaporated amount of water canalso be roughly estimated from thefirst DSC peak if a heat of evapora-tion of around 2300 J/g is assumedfor water. This would give here avalue of about 0.45 g water(around 9%), somewhat higher thanthe value determnined with TGA.However, this method should beused only if the ongoing reaction isknown exactly. The crucible shouldalso be back weighed after the meas-urement.

EvaluationsDSC is used to determine the onsets and TGA the weight loss.SampleTGA: Step %Monohydrate 7.09 (hydrate water loss)Anhydrate < 0.1 (residual moisture)

DSC: Range °C Onset °CMonohydrate 60 – 125 (water loss) 153.8 (melting anhydrate)Anhydrate 160.9 (melting anhydrate)

ConclusionThe combination of DSC and TGA curves allows indentification of pseudopolymorphism as well as a clear distinc-tion between anhydrate and monohydrate.


Theoretically the phase angle in theglass transition region of a polymeris predicted to be zero before and af-ter the transition and to deviate fromzero only during the transition [2,3].The experimentally measured phaseangle, however, behaves in a quitedifferent fashion. The phase angle isnot zero before the transition, devi-ates from this value at approximatelythe same temperature as the stepchange in Cp* and then returns to adifferent non-zero value. This devia-tion from zero before and after thephase transition is easy to understandand is caused by the problem of heattransfer between the instrument andthe sample and within the sample it-self, an experimental difficulty notincluded in the current theoreticalanalysis [4]. This problem of heattransfer will exist irrespective of themake of differential scanning calo-rimeter used in the experiments. It isimportant for the user to be aware,therefore, that if the phase angle ap-pears to be zero before and after thetransition this indicates that the rawdata must have undergone a correc-tion procedure.The simplest method of correctingthe phase angle would be to manipu-late the data in such a way as to arbi-trarily make the phase angle zero be-fore and after the transition, but suchan approach is quite clearly unsatis-factory! To develop a scientificallyacceptable correction procedure wemust consider how the phase angledepends on the experimental vari-ables. It has been shown that, provid-ing the phase angle is small, which itinvariably is, then,

φht ≈ m Cp

*ω/K (4)where the subscript ht indicates that

this is a phase angle due to heattransfer, m andCp are the mass andspecific heat capacity of the sample,respectively, K is Newton’s law con-stant and ω is the frequency ofmodulation [5]. It is important tonote that φht does not include anycontribution from relaxation proc-esses in the sample. From this equa-tion we can see that the phase angelφht increases linearly as the specificheat capacity increases. Figure 22shows how the heat capacity ofpolycarbonate varies with tempera-ture, and it increases by ca 15 % inthe glassy region. If we now look atthe dependence of the negative phaseangle on temperature in the glassyregion φht then that too increases bythe same relative amount, as pre-dicted by equation (4). Weyer et al [6]suggested that this relationship couldbe used in correcting the phase anglefor heat transfer effects. Specifically,the baseline used to correct the phaseangle should be a sigmoidal curvewhich mirrors the change in C

p*

with temperature. Thus, if we takethe phase angle curve and subtractfrom it the heat capacity curve,scaled to make incremental differ-ence before and after the transitionequal to that seen in the phase angle,we obtain a phase angle which has avalue of zero either side of the tran-sition as predicted by theory. Thisprocedure removes the effects ofheat transfer and we obtain the cor-rected phase angle which arises fromrelaxation processes within the sam-ple, see figure 22. Using this phaseangle the in-phase and out-of-phaseheat capacities may now be calcu-lated.

Phase correction in ADSC measure-ments in glass transition

In alternating differential scanningcalorimetry (ADSC), the temperatureis varied sinusoidally as a function oftime and is superimposed on a con-stant underlying heating rate.Schawe [1] has proposed that thedata obtained using the techniquemay be best interpreted in terms of acomplex heat capacity (C

p*) an in-

phase heat capacity (Cp') and an out-

of-phase heat capacity (Cp") which

are defined as:Cp

* = AHF / Aq (1)In-phase heat capacity:Cp

' = Cp* cosφ (2)

Out-of-phase heat capacity:Cp

" = Cp* sinφ (3)

where AHF

is the amplitude of theheat flux modulations, Aq is the am-plitude of the heating rate modula-tions and φ is the phase angle. Thephase angle is defined as a negativequantity within the Mettler-ToledoSTARe software, which is intuitivelypleasing because it indicates that theheat flux modulations lag behind theheating rate modulations.Considerable excitement surroundsthis technique and the reason for thislies to a large extent with our ability,for the first time, to measure theseheat capacities. It is true that theirphysical significance has yet to beestablished but here we have previ-ously unattainable informationwhich should enhance our under-standing of how materials respond tochanges in temperature. Clearly theevaluation of the in-phase and out-of-phase heat capacities relies on theevaluation of the phase angle and sowe must understand the factors thatinfluence the measurement of φ anddevelop methods to correct for spuri-ous instrumental effects.

Corrie T. Imriea, Zhong Jianga and John M. Hutchinsonb

Departments of Chemistrya and Engineeringb, University of Aberdeen, Scotland, UK


[1] J. E. K. Schawe, Thermochim.Acta, 260 (1995) 1

[2] J. M. Hutchinson and S.Montserrat, J. Thermal Anal,47, 103 (1996)

[3] J. M. Hutchinson and S.Monserrat, Thermochim. Acta,286, 263 (1996)

[4] B. Schenker, J. Thermal Anal.,48 (1997) 1097

[5] B. Wunderlich, Y. Jin and A.Boller, Thermochim, Acta, 238,277 (1994)

[6] S. Weyer, A. Hensel and C.Schick, Thermochim, Acta, inthe press (1997)

Fig.22: Complex heat capcity of polycarbonate


ADSC in the glass transition region:the appearance of ripples

heat flux and phase angle traces.They are also visible in the heat ca-pacity traces at an enlarged scale.The origin of these ripples lies in theFourier transformation of the modu-lated heat flux signals in the ADSCprocedure and arises from the“windowing” procedure that is used.This can be seen rather clearly in thetheoretical model predictions [3] forthe phase angle shown in Figure 24.Here the main effect is the departureof phase angle from zero as the glasstransition interval is traversed onheating. Superimposed on this areripples, which appear very similar tothose observed in practice (e.g. Fig-ure 23). The window used in thetheoretical model was a single cycleof the heating rate, which, as thecontrolled variable, was a perfect si-nusoidal modulation.

In Figure 24, the heating rate modu-lations had a period of 60 s and anamplitude of ± 3.1 K/min around andunderlying heating rate of 0.5 K/min.The heat flux modulations, on theother hand, are not perfectly sinusoi-dal in the transition region, as a re-sult of the relaxation processes thatoccur there (in practice, there arealso additional effects due to prob-lems of heat transfer, which are notpresent in this idealised theoreticalmodel). What happens is that, as theunderlying endothermic process be-gins, the heat flux begins to lag theheating rate, and as a consequencethe period of the heat flux modula-tions increases slightly relative tothat of the imposed heating ratemodulations. The Fourier transfor-mation interprets this as a phase lagof the first harmonic component, but

The application of ADSC to the glasstransition is now reasonably well un-derstood in respect of the complexspecific heat capacity C

p*, its in-

phase and out-of-phase components,C

p’ and C

p’’ respectively, and the

phase angle f between heating rateand heat flux [1-3]. For example, the“step-change” in C

p* and C

p’ oc-

curs in a temperature range wherethe timescale for molecular motion isapproximately equal to the period ofthe ADSC modulations. At lowertemperatures, the moleculartimescale is much longer than the pe-riod, so that the response is glassy,with the value of C

p corresponding

to that of a glass; on the other hand,at higher temperatures, the moleculartimescale is much shorter than theperiod, so that the response is liquid-like and gives a value of C

p corre-

sponding to that of a liquid. Clearly,the temperature at which this “step-change” occurs will depend on thechosen period of the modulations;the shorter is the period, the higherthe temperature at which this transi-tion occurs.This behaviour, as well as othercommon features of ADSC traces inthe glass transition region (for exam-ple a peak in the negative phase an-gle) can be explained by a theoreticalmodel [1-3] and are universally ac-cepted as resulting from the relaxa-tion processes. Another feature, how-ever, that is often observed in ADSCtraces in the glass transition regionhas yet to obtain such universal rec-ognition. This is the appearance of“ripples” superimposed on thecurves which result from the ADSCanalysis. An example is shown inFigure 23, which was obtained for asample of polyvinyl acetate, wherethese ripples are easily seen in the

John M. Hutchinsonb, Corrie T. Imriea and Zhong Jianga

Departments of Chemistrya and Engineeringb, University of Aberdeen, Scotland, UK

Fig. 23: ADSC evaluation of PVAc in the glass transition from -10 to 70 °C.The selected experimental parameters were: Period = 60 s, temperature am-plitude = 1 °C, mean heating rate = 1 K/min and sample mass = 12.38 mg.


the use of sliding window equal toone cycle of heating rate modula-tions, which is of a slightly differentfrequency from that of the heat fluxmodulations, results in the ripplesobserved in the phase angle.An exact consideration of the rippleon the phase curve in Figure 24shows that the period of these ripplescorresponds exactly to half the im-posed period of the heating rate.On the mean value curve of the heatflux signal, the period is alwaysequal to the period of the heatingrate and this can be explained by theFourier algorithm.In conclusion, the ripples seen on theADSC traces of Fourier transformedquantities are a natural consequenceof the data analysis procedureadopted. Ripples will always bepresent in the glass transition region,where the phase angle departs fromzero, unless a data smoothing routineis used. They should not, however,be confused with noise on the signal;on the contrary, their appearance rep-resents an ability to measure phaseangle changes with great sensitivity.

Tips:ADSCThe following points must be considered in ADSC measurements:1. Use thin samples with a large contact area with the crucible to assure optimum heat conduction between the

crucible and the sample2. Use small samples to minimize errors due to the thermal conductivity of the sample itself3. With helium as the purge gas, the heat conductance to the sample can be optimized4. Press sample to the crucible bottom with the crucible lid to ensure the contact area does not change during the

measurement5. At least 6 periods should be measured during the effect

–> determines the maximum mean heating rate6. The amplitude of the temperature modulation must be selected so that the maximum heating and cooling rate of

the instrument is not exceeded or dropped belowThe follows holds as a rule of thumb: "1-1-1" (mean heating rate: 1 K/min, amplitude: 1 °C, period 1 min)

Storage life of tapes and DATsAccording to the information from tape and DAT manufacturers, your backups can be reliably read for 10 years.After this time, it is advisable to transfer the backup to a new tape or DAT if the data have to be available for alonger period.A tape or DAT should not be used for more than 100 backups.

ReproducibilityIf you do not have your TA module in permanent operation and rely on the highest reproducibility of the measure-ment (e.g. in determination of the residue or determination of the cp function), you should discard the first meas-urement after putting the instrument back into operation.

[1] J.M. Hutchinson and S. Montserrat, J. Thermal Anal, 47, 103 (1996)[2] J.M. Hutchinson and S. Monserrat, Thermochim. Acta, 286, 263 (1996)[3] J.M. Hutchinson and S. Monserrat, Thermochim. Acta, in press (1996)

Fig. 24: Theoretical phase during the glass transition. The selected experi-mental parameters were: Period = 60 s, temperature amplitude = 0.3 °C,mean heating rate = 0.5 K/min. The temperature range was T

R - 10 °C to T

R

+ 10 °C, where TR is chosen at random.

-10 -8 -6 -4 -2 0 2 4 6 8 10

0.02

0

-0.02

-0.04

-0.10

-0.06

-0.08

-0.12

-0.14

temperature, T – Tr / K

phas

e an

gle

/ rad


Exibitions, conferences and coursesPITTCON March 1 – 6, 1998 New OrleansForum Laboratoire March 31 – April 3, 1998 ParisAnalytica April 21 – 24, 1998 Munich5th Lähnwitzseminar on Calorimetry June 7 – 12, 1998 Kühlungsborn

(near Rostock)ESTAC 7 August 30 – September 4, 1998 Balatonfüred, HungaryNATAS September 13 – 15, 1998 ClevelandK October 22 – 29, 1998 DüsseldorfICTAC 2000 August 14 – 18, 2000 Copenhagen

TA customer courses and seminars (CH)Information and registration: Helga JudexTel: ++41 1 806 7265, Fax ++41 1 806 7240, E-Mail: [email protected]

TA course 16 (French) April 15 – 17, 1998 Greifensee, SwitzerlandTA course 17 (German) April 21 – 23, 1998 Greifensee, SwitzerlandTA course 18 (English) April 28 – 30, 1998 Greifensee, SwitzerlandTA course 47 (German) November 17 – 19, 1998 Greifensee, SwitzerlandTA course 48 (English) November 24 – 26, 1998 Greifensee, Switzerland

Thermal Analysis in PolymerCharacterizationby Prof. Edith A. Turi May 25 – 29, 1998 Zurich, Switzerland

TA customer courses and seminars (Germany)For further information please contact Mettler Toledo GiessenTel. 641 507 431 (Suzanne Pouchie)

Workshop DSC February 10 – 11, 1998 GiessenGEFTA meeting June 4 – 5, 1998 Stuttgart

TA information days and training courses (USA)Please contact your local instrument specialist or Jon ForemanTel. 1 800 METTLER (63 88 537) 88 21 oder Fax 1 609 448 47 77

Regional TA Seminars (UK)Customised on-site training couresContact Rod Bottom for more information Tel. 0116 234 50 25, Fax 0116 236 55 00

If you have any questions concerning other meetings, the products or applications, please contact your localMETTLER TOLEDO dealer.Internet: http://www.mt.com

Editorial officeMettler-Toledo GmbH, AnalyticalSonnenbergstrasse 74CH-8603 Schwerzenbach, SwitzerlandTel. ++41 1 806 73 87; Fax ++41 1 806 72 60E-Mail: [email protected]

Dr. J. de Buhr, Dr. B. Benzler, Dr. R. Riesen, Th. Nitschke, J. Widmann, U. Jörimann,Layout und Production: Promotion & Documentation Schwerzenbach, G. UnterwegnerME–51709782

UserCom6

Documents