Copyright� 2001, The Clay Minerals Society 433

Clays and Clay Minerals, Vol. 49, No. 5, 433–443, 2001.

BASELINE STUDIES OF THE CLAY MINERALS SOCIETY SOURCECLAYS: THERMAL ANALYSIS

STEPHEN GUGGENHEIM AND A.F. KOSTER VAN GROOS‡Department of Earth and Environmental Sciences, University of Illinois at Chicago, Chicago, Illinois 60607, USA

Figure 1. Parameters defined for a simple DTA or DTGpeak (a) and for a TG curve (b).

Figure 2. Parameters defined for complex DTA or DTGpeaks.

INTRODUCTION

Thermal analysis involves a dynamic phenomeno-logical approach to the study of materials by observingthe response of these materials to a change in temper-ature. This approach differs fundamentally from staticmethods of analysis, such as structural or chemicalanalyses, which rely on direct observations of a basicproperty of material (e.g. crystal structure or chemicalcomposition) at a well-defined set of conditions (e.g.temperature, pressure, humidity). Clay minerals arehighly susceptible to significant compositional changesin response to subtle changes in conditions. For ex-ample, changes in the fugacity of water affect the sta-bility of interlayer H2O in a clay mineral (see below).Therefore, care must be taken that all experimentalconditions are known with accuracy and precision.

Differential thermal analysis (DTA), thermal gravi-metric analysis (TG or TGA), and derivative thermalgravimetric (DTG) analysis are reported for each ofthe eight Source Clay minerals using commonly avail-able commercial instruments. The DTA curves showthe effect of energy changes (endothermic or exother-mic reactions) in a sample. For clays, endothermic re-actions involve desorption of surface H2O (e.g. H2Oon exterior surfaces) and dehydration (e.g. interlayerH2O) at low temperatures (�100�C), dehydration anddehydroxylation at more elevated temperatures, and,eventually, melting. Exothermic reactions are relatedto recrystallization at high temperatures that may benearly concurrent with or after dehydroxylation andmelting. Discriminating between desorption and de-hydration or dehydration and dehydroxylation may beproblematic. The TG curves ideally show only weightchanges during heating. The derivative of the TGcurve, the DTG curve, shows changes in the TG slopethat may not be obvious from the TG curve. Thus, theDTG curve and the DTA curve may show strong sim-ilarities for those reactions that involve weight and en-thalpy changes, such as desorption, dehydration anddehydroxylation reactions.

In thermal analytical studies of clay minerals, re-sults from different laboratories often show significantvariations in the desorption and dehydration propertiesof these minerals. Koster van Groos, Guggenheim andco-workers (for a summary, see Guggenheim and Kos-ter van Groos, 1992a,b), found that the temperature ofan event involving H2O is greatly affected by the fu-

43

4C

laysand

Clay

Minerals

Gu

gg

en

he

ima

nd

Ko

ster

Va

nG

roo

s

Table 1. Themogravimetric analysis.

Sample

Adsorbed water/Dehydration/Other

To Tc � wt.% To Tc � wt.% To Tc � wt. %

Other

To Tc � wt.%

Dehydroxylation

To Tc � wt.% To Tc � wt.%

KGa-1b*KGa-1b�

KGa-2*KGa-2�

�20�20�20�20

————

————

————

————

————

————

————

————

————

469474457462

0.401.680.552.07

468474457462

567566541561

12.66º12.26º12.08º12.19º

————

————

————

SWy-2�

SWy-2�SAz-1*SAz-1�

SAz-1�

�20�20�20�20�20

697490

10092

6.228.267.39

13.6515.22

697490

10092

151158155165162

1.321.462.803.904.00

—————

—————

—————

151158155165162

631619541575567

1.701.391.052.012.32

631619541575567

719713683674666

3.203.472.932.442.82

—————

—————

—————

STx-1*STx-1�

STx-1�Syn-1�

Syn-1�

�20�20�20�20�20

10793767064

5.0410.618.973.815.01

1079376——

173165148——

7.402.662.89——

—————

—————

—————

1731651487064

598617545409408

1.001.451.822.692.75

598617545409408

709706703531523

2.732.632.894.814.81

———531523

———643639

———

2.722.66

SHCa-1�

SHCa-1�PFl-1*PFl-1�

PFl-1�

�20�20�20�20�20

7873716486

7.856.702.123.926.33

——716486

——11597

170

——

4.734.311.51

——187191170

——245247220

——

2.242.202.37

7873

245247220

562564361382354

1.621.700.491.010.94

562564361382354

658717506497510

0.641.204.303.774.53

858717———

711743———

0.190.60———

See Figure 1b and text for definition of column headings and other abbreviations.* UIC, nitrogen purge;� UT, no nitrogen purge;� UT, nitrogen purge.º � wt.% based on sample weight just prior to dehydroxylation.SHCa-1� Four additional dehydroxylation events are at: To� 711�C, Tc � 741�C; � wt.% � 0.63; To� 741�C, Tc � 806�C, � wt.% � 0.73; To� 806�C, Tc � 850�C, �

wt.% � 0.92; To� 850�C, Tc � not available,� wt.% � 1.78.SHCa1� Three additional dehydroxylation events are at: To� 743�C, Tc � 809�C, � wt.% � 0.90; To� 809�C, Tc � 848�C, � wt.% � 0.80; To� 848�C, Tc � 894�C,

� wt.% � 0.60.PFl-1* Additional dehydration occurred at: To� 115�C, Tc � 187�C, � wt.% � 1.16.PFl-1� Additional dehydration occurred at: To� 97�C, Tc � 191�C, � wt.% � 1.59.PFl-1� Additional dehydration occurred at: To� 115�C, Tc � 187�C, � wt.% � 1.16.

Vo

l.4

9,

No

.5

,2

00

14

35

Th

erm

ala

na

lysis

Table 2. Derivative thermal gravimetry.

Sample

Adsorbed water/Dehydration

Peak (1)

Po Sm Sr Pm So1 Sm1 So2 Sm2 Pr

Peak (2)

Po Sm Sr Pm So Sm Pr

Dehydroxylation

Po Sm Sr Pm So Sm Pr

KGa-1b*KGa-1b�

KGa-2*KGa-2�

�20�20�20�20

————

————

————

————

————

————

————

————

————

————

————

————

————

————

————

448462437462

————

————

519517506509

————

————

581575553549

SWy-2�

SWy-2�SAz-1*SAz-�

SAz-1**

�20�20�20�20�20

—————

—————

5159786564

1061091127270

1151191368076

123127—121117

139148—144146

165164175172174

—————

—————

—————

—————

—————

—————

—————

592598512546542

622612———

640628———

689685633639633

—————

—————

734726680691689

STx-1*STx-1�

STx-1�

Syn-1�

Syn-1�

�20�20�20�20�20

——

——

——

——

8269513733

————52

————61

—————

—————

122113898485

126122104——

—————

—————

150146130——

—————

—————

173169153——

457595451378366

503—485441—

553—547473—

669679663494483

———554544

———600602

722719712669666

SHCa-1�

SHCa-1�PFl-1*PFl-1�

PFl-1�

�20�20�20�20�20

———4343

———5752

6454938167

———113114

———130120

—————

—————

8578

128146128

——163169150

———199—

———207—

——215220196

———237—

———250—

——260270225

617528311344305

—————

—————

636617441442444

—————

—————

663711530527538

See Figures 1a and 2 and the text for definition of column headings and other abbreviations.* UIC, nitrogen purge;� UT, no nitrogen purge;� UT, nitrogen purge.STx-1� One additional dehydroxylation peak at: Po� 474�C, Pm� 505�C, Pr � 529�C.SHCa-1� Three additional dehydroxylation peaks with: Po� 708�C, Pm� 729�C, Pr � 752�C; Po � 798�C, Pm� 827�C, Pr � 867�C; Po � 867�C, Pm� 890�C, Pr �

945�C.SHCa-1� Three additional dehydroxylation peaks with: Po� 711�C, Pm� 735�C, Pr � 756�C; Po� 784�C, Pm� 829�C, Pr � 868�C; Po� 868�C, Pm� 908�C, Pr, not

available.

43

6C

laysand

Clay

Minerals

Gu

gg

en

he

ima

nd

Ko

ster

Va

nG

roo

s

Table 3. Differential thermal analysis.

Dehydration Dehydroxylation Melting Recrystallization

Sample

Peak (1)

Po Sm Sr Pm So1 Sm1 So2 Sm2 Pr

Peak (2)

Po Pm Pr Po Sm Sr Pm So Sm Pr Po Pm Pr Po Pm So Sm Pr

KGa-1b*KGa-1b�

KGa-2*KGa-2�

————

————

————

————

————

————

————

————

————

————

————

————

483452474439

————

————

575518568513

————

————

618577604557

————

————

————

974980961965

990993977984

————

————

994997981994

SWy-2*SAz-1*STx-1*Syn-1*SHCa-1*PFL-1*

90�70�70

90�70

97

——————

——————

135178167139156152

—204198———

—228217———

170269249—200—

185327327—250—

209368351188265196

234——251265209

332——319331272

374——353360307

601548593392727377

———458741—

———502758—

684641681532827455

—695—563——

—704—578——

737721730673888526

—815829—

1122—

—854894—

1175—

—883931—

1196—

889950982939—850

923979

1008990—872

948992

1023———

96010131080———

976110611061009

—915

See Figures 1a and 2 and the text for definition of column headings and other abbreviations.* TK; � UT.Syn-1* Two additional dehydroxylation peaks are at 489�C and 626�C

SHCa-1* One small additional dehydroxylation peak with: Po� 645�C, Pm� 695�C and Pr� 717�C was observed.

Vol. 49, No. 5, 2001 437Thermal analysis

Table 4. Comments on color changes after thermal analysisfor samples analyzed at UIC.

Sample Comments

KGa-1bKGa-2SWy-2SAz-1STx-1Syn-1SHCa-1PFl-1

no changechange from tan to orangechange from tan to red-brownchange from brown-white to orangeno changeno changeno changetan to dark red-brown

Figure 3. Experimental curves for KGa-1b: (a) TGA/DTG (UIC), N2 purge; (b) TGA/DTG (UT); (c) DTA (TK); (d) DTA(UT).

gacity of water,f , at the hydrated site. Hence, theH O2

response of materials during thermal analysis will bestrongly influenced by a variety of factors, includingthe humidity surrounding the sample at the time of theexperiment. Evidence for this is shown by the effectof purging the atmosphere surrounding the samplewith nitrogen. In addition, the grain size of clay ag-gregates, as well as the size of the individual crystals,will affect the diffusivity of H2O and, consequently,f at the hydrated site. More uniform experimentalH O2

conditions may be obtained using high-pressure ther-mal analysis (HP-DTA), because the sample capsulemay be sealed after water is added, thus controllingf .H O2

Other experimental variables which influence theapparent temperatures of water-loss reactions includesample size, packing, sample holder configuration,heating rate, particle distribution, contaminants,etc.Most of these affect the ability of water to equilibratearound the clay sample during a dynamic experiment.Although the storage of samples over a saturated so-lution of Mg(NO3)2·6H2O prior to thermal analysismaintains constancy of relative humidity at 55% andthus helps in maintaining a fixed water pressure earlyin the thermal analysis, this technique is only effectiveduring the procedure at the very lowest temperaturerange near room temperature. For a more completediscussion on the importance of each variable, seeMackenzie (1972). Shipment of samples to the partic-ipating laboratories precluded maintaining these sam-ples at 55% relative humidity. Thus, the samples werestored under ambient humidity conditions at each lab-oratory.

438 Clays and Clay MineralsGuggenheim and Koster Van Groos

Figure 4. Experimental curves for KGa-2: (a) TGA/DTG (UIC), N2 purge; (b) TGA/DTG (UT); (c) DTA (TK); (d) DTA(UT).

EXPERIMENTAL METHOD

Samples were prepared as described by Costanzo(2001) and distributed for thermal analysis to severaldifferent laboratories with different instruments. Threelaboratories were involved: University of Utrecht(UT), University of Illinois at Chicago (UIC), andThiele Kaolin Company (TK). The UT laboratory useda Du Pont Thermal Analysis System 1090 for the DTAexperiments and a TA Instruments 2000 system con-troller with a 2950 TGA module. Du Pont and TAInstruments were previously affiliated and are run us-ing the same software. The UIC laboratory used a TAInstruments 1600 for the DTA experiments and a mod-ule 51 (TA Instruments) for TGA; the software is iden-tical to that used by UT. The TK instrumentation wasa Perkin-Elmer 7 Series Thermal Analysis System.

The UT and UIC laboratories used platinum cruci-bles with Pt/Pt13%Rh thermalcouples adjacent to thebottom of the pans. The TK laboratory used aluminacrucibles with Pt/Pt10%Rh thermalcouples. Samples

were pressed by hand into the crucibles at UT, com-pressed using a Puritan applicator at TK, and tappedinto the crucibles at UIC. At UIC, however, samplesSWy-2, SHCa-1 and Syn-1 were first wetted with wa-ter and then dried before placing into the crucible ow-ing to difficulty in handling the freeze-dried (‘fluffy’)material. The N2 flow rate was 50 cc/min for DTAexperiments and either the same flow rate or no N2

flow for TG experiments. At TK, the purge gas wasair with a flow rate of�26 psi. The DTA experimentsused an�-alumina powder as the reference. For allexperiments, a heating rate of 10�C/min was main-tained.

RESULTS

The DTA experiments from the TK laboratory andthe TG and DTG curves from the UT and UIC labo-ratories are reported here. Temperatures for ‘simple’DTA and DTG peaks are defined (Figure 1a) with anextrapolated onset (Po), peak maximum (Pm), and ex-

Vol. 49, No. 5, 2001 439Thermal analysis

Figure 5. Experimental curves for SWy-2: (a) TGA/DTG(UT); (b) TGA/DTG (UT); N2 purge, (c) DTA (TK).

trapolated return (Pr� Peak return) to baseline. TheTG curves have an extrapolated onset (To) and ex-trapolated completion (Tc), as shown in Figure 1b.Composite peaks, where peak shape is difficult to de-fine owing to the complexity of two or more nearlysuperimposing peaks are illustrated in Figure 2. Thus,

in addition to the main peak (Pm), simple shoulderpeaks are shown on the low-temperature limb, withthe maximum (Sm) and the associated return temper-ature (Sr). On the high-temperature limb, the param-eters So1, So2,etc. and Sm1, Sm2,etc. define theonset and maximum temperatures of the shoulderpeaks. Experimental curves are given in Figures 3–10and significant temperatures for the curves are tabu-lated in Tables 1–3.

Table 4 provides color-change information for thesamples after thermal analysis at UIC.

DISCUSSION

The use of several laboratories with different equip-ment and different operators shows the variability ofresults for thermal analysis. As is common for samplesnot stored at constant-humidity conditions prior tothermal analysis, variations will occur at temperaturesbelow �100�C owing to surface-adsorbed H2O. Dif-ferent techniques in purging (e.g. with N2, air, or nopurge) will also produce variations at these tempera-tures. Above�100�C, differences in packing of thesample may produce significant differences in the re-sults. In these cases, evolving water vapor affectsf around the sample, with results depending on theH O2

ability of this vapor to disperse away from the sample.Therefore, caution must be exercised in comparing re-sults to those presented here.

Mackenzie and Callie`re (1979) noted that weightloss is affected by the presence of ammonia in thesynthetic mica-montmorillonite (Syn-1). Figure 8shows a slow and continuous loss in weight from 200to 400�C, a significant loss in weight from 400 to600�C, and a DTA event near 320�C. Although a de-termination of ammonium loss at any of these tem-peratures would require additional study, ammonialoss is certainly a possible explanation. It is notewor-thy that the DTA curve for SWy-2 also shows an eventat �330�C.

Mineralogical analysis of the samples as determinedby X-ray diffraction (Chipera and Bish, 2001) indi-cates the presence of impurities in the studied frac-tions. We see no evidence of impurities affecting thethermal analysis, except possibly in SHCa-1, in whicha 740�C peak in the TGA could represent dolomitedecarbonation, and in Syn-1, in which the 430�C eventin TGA and the 570�C event in the DTA could maska response from boehmite. However, in SHCa-1, nocorresponding evidence for dolomite was observed inDTA. In addition, visual examination of SAz-1 priorto thermal analysis showed white and brown flakes,clearly indicating inhomogeneity of the sample.

ACKNOWLEDGMENTS

We thank S. van der Laan and Paul Anten, both at EarthSciences, Budapestlaan, Utrecht, The Netherlands, J. Elzea

440 Clays and Clay MineralsGuggenheim and Koster Van Groos

Figure 6. Experimental curves for SAz-1: (a) TGA/DTG (UIC), N2 purge; (b) TGA/DTG (UT); (c) TGA/DTG (UT); N2purge; (d) DTA (TK).

Kogel, Thiele Kaolin Company, and P. Messersmith and K.Thorne, both at the Department of Restorative Dentistry, Uni-versity of Illinois at Chicago, for helping in the facilitationof the experiments. We thank Ross Giese, S.U.N.Y. at Buf-falo, and Fred Wicks, Royal Ontario Museum, for reviews.Portions of this work were funded by the Donors of the Pe-troleum Research Fund, administered by the American Chem-ical Society, under grant 32858-AC5 and by the National Sci-ence Foundation under grant EAR-0001122.

REFERENCES

Chipera, S.J. and Bish, D. (2001) Baseline studies of the ClayMinerals Society Source Clays: powder diffraction analy-ses.Clays and Clay Minerals, 49, 398–409.

Costanzo, P.M. (2001) Baseline studies of the Clay MineralsSociety Source Clays: introduction.Clays and Clay Min-erals, 49, 372–373.

Guggenheim S. and Koster van Groos, A.F. (1992a) High-Pressure Differential Thermal Analysis (HP-DTA) I. De-hydration reactions at elevated pressures in phyllosilicates.Journal of Thermal Analysis 38, 1701–1728.

Guggenheim S. and Koster van Groos, A.F. (1992b) High-Pressure Differential Thermal Analysis (HP-DTA) II. De-hydroxylation reactions at elevated pressures in phyllosili-cates.Journal of Thermal Analysis 38, 2529–2548.

Mackenzie, R.C. (editor) (1972)Differential Thermal Analy-sis, vol. 2. Applications. Academic Press, London and NewYork.

Mackenzie, R.C. and Callie`re, S. (1979) Thermal analysis,DTA, TG, DTG. Pp. 243–284 in:Data Handbook for ClayMaterials and other Non-metallic Minerals (H. van Olphenand J.J. Fripiat, editors). Pergamon, Oxford, UK.

E-mail of corresponding author: xtal@uic.edu

Vol. 49, No. 5, 2001 441Thermal analysis

Figure 7. Experimental curves for STx-1: (a) TGA/DTG (UIC), N2 purge; (b) TGA/DTG (UT); (c) TGA/DTG (UT), N2purge; (d) DTA (TK).

442 Clays and Clay MineralsGuggenheim and Koster Van Groos

Figure 8. Experimental curves for Syn-1: (a) TGA/DTG(UT); (b) TGA/DTG (UT), N2 purge; (c) DTA (TK). Figure 9. Experimental curves for SHCa-1: (a) TGA/DTG

(UT); (b) TGA/DTG (UT), N2 purge; (c) DTA (TK).

Vol. 49, No. 5, 2001 443Thermal analysis

Figure 10. Experimental curves for PFl-1: (a) TGA/DTG (UIC), N2 purge; (b) TGA/DTG (UT); (c) TGA/DTG (UT), N2purge; (d) DTA (TK).