QUANTITATIVE DIFFERENTIAL THERMAL ANALY-SES OF CLAY AND OTHER MINERALS

H. W. vAN DER N{anol,

Agricultural, E*periment Station and, Institute Jor

S oil Re s ear ch, T . N .O., Gr o nin gen, I{ etheil and's.

Ansrnecr

The shape and the intensity of the thermal curve of minerals, when analyzed by the

il.t.a. method, are strongly influenced by amorphous coatings and disordered structures on

the surface of the particles (Beilby layer), and furthermore by difierences in particle and/or

crystallite size, the degree of crystallinity of the crystallites and ion substitutions in the

crystal structure. Examples are given.

Especially in the group of natural clay minerals, variations in these phenomena are

numerous. Consequently an accurate quantitative determination of the amount of a clay

mineral in a sample by means of. t]ne tl.t.a. method, is impossible. Thus it was found that

even kaolinite from pure weII known deposits has a heat of transformation which varies

from 100 to 176 and from 23 to 43 cal./g. for its endothermal (ca 600" c.) and exothermal

reaction (ca 980" C.) respectively.For other minerals reliable results can be obtained only if the conditions under which the

specimens are formed in nature are practically always the same and the minerals are more-

over of pure chemical composition, i.e., circumstances, which are mostly seldom found.

These conclusions are based on experiments where errors due to the differential thermal

analysis il.t.a. apparatrts are eliminated by calibration of the instrument with a standard

chemical, use of the same sample holder and thermocouple, dilution of the mineral with

the inert reference material and packing the mixtures always in the same way in the sample

holder.

INrnonucrtox

Quantitative d.t.a.has been applied by several authors as a rapid, in-

expensive and accurate method for the determination of the amount of a

mineral-see e.g. Norton (1940) for kaolinite, hydrargillite and diaspore

in bauxites from Dutch Guiana, Kiyoura and Sata (1950)' and Murray

et al. (1951) for calcite in limestones. This method of analysis was also

used by Vold (19a9) to determine the amount of stearic or benzoic acid

in a sample from its heat of fusion. The result should be accurate to

within a few per cents according to the authors. Grimshaw and Roberts(1953) suggest that the accuracy of the d.t.o. method can be increased by

diluting the samples 1:3 with inert (calcined) alumina before the test.

For in this case the conductivity of the samples investigated should be

nearer to that of the inert a-AIzOa reference material which is used in a

d,.t.a.Ilowever, Speil et al. (19a5) found for undiluted samples of kaolinite

from various origins, deviations of about 30/6 in their heats of trans-formation. De Bruijn and Van der Nlarel (1954) who investigated aside

from kaolinite also other minerals from various localities, arrived at

222

QUANTITATIVE DIFFERENTIAL THERMAL ANALYSES OF CLAY 223

deviations of 100 to 200/6. These large deviations cannot be ascribed

It is the purpose of this article to obtain data about the variation inthermal efiects of minerals from various origins, and also about the mostimportant factors by which heat of transformation of a certain mineral isaffected. All the d.t.a. tests described in the experimental part of thisarticle have been performed with the apparatus constructed and de-scribed by De Bruijn (1954) and which has been in operation in the Lab-oratory of Soil Mechanics at Delft (Netherlands) since 1952.1

The nickel-block oven contains 6 holes in which 5 samples can be an-alyzed simultaneously, the remaining hole being reserved for the inertreference material, usually a-AlzOa previously heated for several hoursat ca. 1300" c. It is provided with pt/pt, Rh thermocouples, a Boersmaamplifier and a Brown recorder.

ExppnruBNrs

Factors by which the thermal transformations of a mineral are af-fected when it is heated and the effect registered by the d..t.a. method,are:

(I) Parti.cle size oJ the mineral, investigated.

Many authors have found that the f.ner the particle size of the sample,the lower is the actual temperature at its transformation and the smaller

Milne (1953), De Bruijn and van der Marel (1954) for muscovite. DeBruijn and Van der Marel (1954) for pyrophyllite, Faust (1949), Graf

I The author is indebted to Prof. rr. E. c. w. A. Geuze, Director of the Laboratory forsoil Mechanics, Delft, for use of this apparatus and to H. Labrie for assistence in theanalvses.

224 H. W. VAN DER MAREL

(1es2).Apparently, the smaller the particle or the crystallite size, the smaller

ur. th" forces needed for the transformations inside the crystals. Even

and calcite.

when inhalated continuouslY.I twas found tha t thene i l by laye rcou ldbe removedby t rea t i ng the

sample with a borate bufier-see Clelland et al' (1952)-or by etching

with HF-see clelland and Ritchie (Ig52), Nagelschmidt et al. (1952).

C ho l c cdon

TM uscov i tc I

IL

Ouort, I

tL

t

QT]ANTITATIVE DIFFERENTIAL THERMAL ANALYSES OF CLAY 225

A t

400 60 -2 ro f

< 2 P

powdc fad

powdc r cd

. 2 P

Bio t i t c

powdercd

. 2 P

powdcrcd

. 2 P

2rO-soof

. aop

p rcc i p i t o t cd( co ro ! )

Pyro-phy t l i ta

Co lc i t c

Notc : Tha

Dc

roo 200 300 400 soo 600 700 800

tc m Dcr oturc

dc r i ved f r omcurv?s o f quor tz ,muscov i tc ond pyrophy l l i ta o r

Brui jn ond Von dcr Morer (9Sa) - Port t r . p9. 4O7.

Fro. l. D.t.a. of various minerals in difierent grades of fineness.

Weight of samples analyzed in mg.

226 II. W. VAN DER ,IIAREL

Grimshaw and Roberts (1953)-and that the intensity of the originala/B transformation could be restored.2

(2) Degree of crystallinity oJ the mineral inaestigated,

According to Grimshaw et al. (1945), Grim (1947) and Murray (1954)the endo- and exothermal reaction of the poorly crystallized kaolinites(fire clay, ball clay) are broader, of less intensity and. they begin at alower decomposition temperature than the corresponding reactions ofthe better crystallized ones. The first is caused by loss of OH from thecrystal structure and the latter by the formation of T Al:Oa.

Cailldre and H6nin (1948b) ascribe the difierence in peak temperatureof the endothermal reaction of dickite, kaolinite and metahalloysite,s be-ing ca. 580o C., 5500 C. and 5250 C., respectively, to differences in the de-gree of crystallinity. The first is the best crystallized and the last thepoorest. Apparently a mineral with a high degree of crystallinity needsmore energy for its transformation to a new structure than a mineral ofpoor crystallinity. According to Speil et al. (1945), Kerr and Kulp (1948),Bramao et al. (1950, 1952) and Glass (1954), metahalloysite (and hal-loysite) may also be distinguished from kaolinite by the shape of the en-dothermal reaction, i.e. the peak for metahalloysite (halloysite) returnsto the base line at a faster rate than it departs, whereas well crystallizedkaolinites show about equal rates. The shape ratio of the thermogram,i.e. tan aftan B, in which a:the angle between the perpendicular tothe peak and the descending side and 0: the corresponding angle on theascending side, has been proposed by Bramao et al. (1952) as a conven-ient means to distinguish kaolinite from metahalloysite (halloysite). Thereason of this difierent behaviour of the above minerals is, according toKerr and Kulp (19a8), that the sheets in the metahalloysite (haltoysite)structure are superimposed in a less orderly manner than in kaolinite.The occurrence of halloysite (metahalloysite) in nature as small lath-shaped and not as plate-like particles such as kaolinite, is also supposed

2 The thickness of the Beilby layer on quartz particles was computed by Clelland (1951),Dempster (1951), Clelland et al. (1952), Clelland and Ritchie (1952) and Dempster andRitchie (1952, 1953) from the decrease in density, the decrease in d..t.o. effect and the in-crease in SiOz dissolved by the borate bufier between the original and the ground kaolinitesample. They found 0.03-0.05 p, 0.11-O.15 p and 0.02-O.03 p respectively. Nagelschmidtet al. (1952), Gibb et al. (1953) found from r-ray and electron diffraction analyses 0.03 pand 0.0H.06 p, respectively. Meldau and Robertson (1952) arrived at 0.03 to 0.20 pfor augite, hornblende, fluorite and aragonite by electron difiraction analysis.

3 In this paper the terms halloysite and metahalloysite, the latter resulting from theformer by heating at ca. 45o C., have been used to designate the phases Alroi.2SiO2.4II2Oand AlzOs' 2SiOr' 2HzO, respectively, also called endellite and halloysite, respectively byother investigators.

QUANTITATIVE DIFFERENTIAL THERMAL ANALYSES OF CLAI. 221

to be caused by stresses in the crystal resulting from this particularstructure.

By x-ray analysis the poorer degree of crystallinity of halloysite andmetahalloysite as compared with kaolinite can easily be observed on thephoto-see e.g. van der Marel (1950). Therefore, due to a lack of per-fection of the individual crystallites, rather than to their smallet size,the d..t.a. curve of halloysite (metahalloysite) is not symmetrical as inkaoiinite, but asymmetrical. Ilowever, de Bruijn and van der Marel(1954) and Robertson et al. (1954) observed many well crystallized kaolin-ites, verified by r-ray analysis, to have an asymmetric endothermalreaction. In Plate r are some examples. Thus the size oI the crystallitesand amorphous coatings-see under (l)-on the particles have also af-fected the shape and the intensity of the endothermal efiect of kaolinite.

(3) Ion substitutions ,i.n the crystal, structure of the mineral inaesti.gated,

According to Orcel (1935) and afterwards Kelley and Page (1943),Cail lbre and H6nin (1947, 1948a,b,l94g), Kulp et al. (1951), Graf (1952)the introduction of iron into the structure of a mineral, substitutingaluminum, magnesium or silicon, shifts the endothermal peak tempera-ture to lower grades and changes its intensity. According to Cailldre andH6nin (1947,I948a,b, t949) the kind of binding of OH in the crystalstructure of a mineral is related to its decomposition temperature. fnthis way OH of talc, antigorite and brucite is bound with decreasingstrength in the same order as here mentioned as they have endothermalefiects at 950o C., 650o C. and 400" C., respectively.

Page (1943) and Kelley and Page (1943) found that when Al is the pre-dominant constituent of the octahedral sheet of the 2:1 minerals, thewater is given off at a lower temperature than when Mg is the predomi-nant constituent. Grim and Rowland (1942) found that iron-free Texasmontmorillonite does not show its exothermal peak (caused by the for-mation of a spinel) until at ca. 10500 C. As a contrast magnesium-richOtay montmorillonite does not show a distinct exothermal reaction be-cause enstatite is formed at ca. 950o C. instead of spinel-see Earley etal. (1953). In Fig. 2 is demonstrated the evident effect of ion substitutionson the shape and the intensity oI the d.t.a. curve for some montmorillo-nites, kaolinites and carbonates-see also the examples given by otherinvestigators.

(4) Diferent e*changeable cations in the mineral, iwestigated

Hendricks et al. (1940), Caitlbre and H6nin (1944), Barshad (1948,1950) and Arens (1951) found that the kind of exchangeable cation af-fects the shape and the intensity of the low temperature endothermal

228

Meto-ho l loys i te

2A3

Koo l i n i t ?95a

K o o l i n i t c292

Koo l i n i t et25

r-r2 r33 | 7l

E. W. VAN DER MAREL

t o 3 t o 32.5r 5.O2f | 502 2.Sl l: l l 133 | 12

f o t 2 3 4 5

Ac m

2 8 3

9 5 4

2 9 2

! 2 5

M ort ins bcrg

Du tch Gu iono

F roncc

Zcltlitz

Mor t i nsbe rg

Du tch Gu iono

F r o n c ?

Zc t t l i t z

Note: all the samples were dried t hour at 105' C. before the d.t.a.

Prarn I. X-ray and il,.t.a. analyses of metahalloysite and kaolinite (2p with an asym-metric shape of their endothermal reaction at 600o C. Furthermore of kaolinite (2p from

Zettlitz (Czechoslov.) with a symmetric shape as usual.

tcmPcroture in oc

OU A N T I T ATI V E DI II I? EREN T I A L

P y f o p h y l l i t cT q l c

oB c i d e l l i t ?

S o p o n i t c

N o n t r o n i t c

K o o l i n i t ?

A n t i g o r i t c

C r o n s t e d t i t c

c o l c i t c

M o g n c s i t c

S i d c r i t c

S t r o n t i o n i t c

W i t h c r i t c

229T'IIERMAI, ANALY.SES OF CLAY

o-

Co

o

oo

A I

M 9

F Q

A I

M 9

F c

CI,il- __.i

M 9

oco

o

too 200 300 400 500 600 700 800 900 looo

l cmPc ro tu r c i n oc

I The author is indebted for these valuable samples to Dr. S. B. Hendricks,

U. S. Department of Agriculture, Beltsville, Maryland, (U.S.A.).

Frc.2. D.t.a. of pyrophyllite and talc and some montmorillonites,

kaolinites and carbonates. Weight of samples analyzed in mg.

H. W. VAN DER MAREL

reaction of montmorillonite. Faust (1951) arrived at the same result forsauconite, e.g. a mineral representing the Zn-end member of the mont-morillonite series.a The low endothermal reaction of montmorillonite iscaused by loss of planar water, broken bound water and water bound toadsorbed cations. The latter is the most firmly held and thus has an en-dothermal reaction at higher temperature than the foregoing.

A very large efiect on the d.t.a. diagram of saturating a mineral withdifferent cations, has been observed for vermiculite by Barshad (1948),Walker and Milne (1950) and Arens (1951). This mineral contracts itsplates when saturated with K+, NHr+, Rb+ and Cs+ ions thereby ex-pelling the water molecules between the lavers, when saturated withCa+, Mg++, or Sr+ ions-see Gruner (1939), Barshad (1948, 1950),Walker (1950). Mackenzie (1950) has attempted to correlate the peaktemperatures as found by Hendricks et al. (1940) for montmorillonitesaturated with various cations, with the hydration energies of these ionsaccording to Bernal and Fowler (1933). A correlation was found, exceptfor lithium. However Greene-Kelley (1953) pointed out that the activa-tion energy detected by the d.t.a. method is not directly related to thehydration energy of the cation concerned. It will also depend on thecharge density between the layers. In this way cations which give doublepeaks for vermiculite, may give single peaks for montmorillonite. Ofcourse minerals with a small cation exchange capacity such as kaolinite-see Arens (1951)-will not manifest significantly the phenomenon ofits d.t.o. curve being influenced by the kind of adsorbed cation. In Fig.3 are some examples demonstrating the efiect of the kind of adsorbedcations on the low temperature thermal reaction of vermiculite and openillite (ammersooite). The latter mineral also contracts its plates whentreated with K+, NHa+, Rb+ and Cs+ ions thereby expelling its cationbound HrO molecules from between the layers-see Van der Marel(1954). In order to eliminate the disturbing effect of the weaker boundHrO molecules, the samples were dried (in this case t hour at 1050 C.)before they were analyzed by the d.t.a. method. Moreover the sampleswere analyzed under conditions where errors due to the apparatus couldnot disturb the results. As pointed out by De Josselin de Jong (1956)these errors are: dilution of the sample with the inert reference material(o-ALOr), packing the samples always in the same way in the sampleholder, covering the thermocouples with sufficient material so that theheat flow in the neighbourhood of the thermocouples always follows thesame geometrical (cylindrical) pattern. Finally the sample holder*its

a By montmorillonite is meant here, as in common usage, the high aluminum memberof the montmorillonite series with some slight replacement of Al3+ by MglF and no replace-ment of Sia+ by Al*.

QUANTITATIVE DIFFEREN:TIAL THERMAL ANALYSES OF CLAY 231

thermocouple was calibrated with a standard chemical (in this caseCuSOn'sHrO) before and after each mineral/ot-AlzOs mixture was ana-lyzed,.

As may be observed from the results, the effect of the kind of exchange-able cation is smaller for the open illite as for vermiculite. However,the total cation exchange capacity of the open illite is only ca. 65 m.e./100

3m /'oo '".'..",,i?l ," :T 3oo

Note: All samples dried t hour before the il.t.o. and diluted 3: 1 with a-AlzOa.

Frc. 3. D.t.a. of vermiculite and open illite (ammersooite)

saturated with difierent cations.

g. as against ca. 120 m.e./100 g. for vermiculite. Therefore, also the totalamount of HzO molecules which influence the thermal effect of openillite is smaller than that of vermiculite.

(5) Chernical reactions behteen minerals in the sample inaestigated,

Decomposition of pure dolomite, CaMg(COe)r takes place in twosteps. First the magnesite component loses its COz at ca. 8000 C. andthereafter the calcite component loses its COz at ca. 9000 C.-see Flood(1950). However, Stutterheim et al. (1951), Heady (1952) and Webb

232 E. W. VAN DER MAREL

(1953) found that in an intimate mixture of dolomite and Ca(OH)2, thenormal large endothermal reaction due to the decomposition of MgCO3in the dolomite, was hardly noticeable. This efiect was caused by theinstantaneous, exothermal reaction between the COz evolved from theMgCOr fraction of the dolomite with the CaO resulting from the priordecomposition of the Ca(OH)r. In consequence the first endothermaleffect of dolomite was almost completely cancelled, but the second endo-thermal efiect of dolomite was enlarged as its effect was increased by theendothermal reaction of the formed CaCOa having the same peak tem-perature as the CaCO3 compon€nt of dolomite. On the other hand Webb(1953) found for molar mixtures of MgCOa and CaO or Ca(OH)2, thata MgCO3 peak was obtained which was followed by only a small CaCOapeak. Thus the same reaction took place as in the foregoing examplebut only less readily. In mixtures of quartz and calcite, the endothermalreaction of calcite is hindered by an exothermal reaction of calcium sili-cate, following immediately upon the former-see Lippmann (1952).

(6) Other d,isturbing components in the sample inaest'igoted.

According to Grimshaw et al. (1945), Heady (1952) and De Bruijnand Van der Marel (1954) exothermal efiects caused by small amounts oforganic matter (humus) are often so pronounced as to reduce, or evento eliminate any endothermal efiect which may take place simultaneous-Iy. By treating the sample before the d.t.a. with HzOr, slightly decom-posable organic material may be oxidized. Many soils and shales, how-ever, contain organic matter which cannot be destroyed with HrOz,e.g. cellulose, lignin and elemental carbon. According to Mackenzie andLakin (1953) graphitic (crystallized) carbon is less easily oxidized in airthan amorphous carbon and thus shows an oxidation peak at a muchhigher temperature.

It should be mentioned here that Allaway (9a\ could magnify theendothermal effect of montmorillonite, beidellite and nontronite byinvestigating samples saturated with piperidine. The latter is first re-duced to carbon and then yields hydrogen with the water vapour escap-ing from the crystal at its decomposition temperature and which has agreat heat of combustion when reacting with the oxygen of the air. How-ever, for kaolinite having only a small base exchange capacity, or forsamples mixed with finely divided carbon, this method has proved tobe valueless. Evidently, a mechanical mixture of inert clay with piperi-dine or carbon burns in a way much different from a clay which containsadsorbed piperidine cations.

If organic matter is present, the only way to avoid errors is to analyzethe sample in vacuum or argon atmosphere-see Rowland and Lewis

QT,IANTITATIVE DIFFERENTIAL TEERMAL ANALYSES OF CLAY 233

(1951). A number of workers-see Budnikov and Bobrovink (1938)'

Berg (1943, 1945), Schwob (1950) and Graf (1952)-have described the

evident, decreasing effect of only small amounts of soluble alkali andalkaline earth chlorides and carbonates, upon the theimal decompositionof dolomite. Caillbre and H6nin (19480) observed for mixtures of NaCI

to.3 to.3t ' t r.g. ' .7' .,t o

5 4 g 2 | o | 2 3 4 s c f r

297

423

2 9 7

4 2 3

mon tmor i l l on i t c+ 9 l yca fo l

k o o l i n i t e

4OO mg

3OO mg

t cmpe ro tu r c i n oC

Note: the samples were dried I hour at 105' C. before the d'.t.a.

Prlre II. X-ray and il.t.a. analysis of montmorillonite (2p and

kaolinite (2p which contain Na2COB.

and KCI that the endothermal peak temperature-being for each com-ponent in its pure state ca. 800o C.-is even shifted to 660" C. in the 5O7omixture. Moreover, they found for NaCI, KCI, CaCOa, CaSOa mixtures

that the endothermal reaction of calcite at 900o C. may disappear com-pletely and that new ones are produced. This is due to various reactionsoccurring during the heating of the sample. Gruver et al. (1949) found a

strong suppression of the thermal reactions of kaolinite when 5/6 NaCl

234 H, W. VAN DER MAREL

or 5To NarCOs was added to the sample and that the suppression of theexothermal reaction was decreased more than that of the endothermalreaction.

Plate II shows the results of x-ray and d..t.a. analyses of kaolinite andmontmorillonite which have been purified by sedimentation in Atterbergcylinders, 0.01 N NaOH being used as peptisator. As with these fineclays the NaOH could not be washed ofi entirely, the dried sampleswhich were analyzed,, still contained (ca 47d NazCO (through absorp-tion of COz from the air). It may be concluded that the thermal curvesof these samples, which by *-ray analysis are pure well crystallized kaolin-ite and montmorillonite, have evidently been changed by the impurity.

(7) Volume changes in the sample inaestigated

Shrinking, sintering or melting of the sample during heating, afiect itsthermal properties and therefore also the shape and the intensity of thed..t.a. curve-see Norton (1939), Berg (1945) and De Josselin de Jong(1956). In Fig. 4 are some striking examples-see also the examplesgiven by other investigators. Thus magnesite, calcite, brucite, hydrargil-lite lose ca. 52/e, 4470, 3070 and 30/s respectively of their weight at thedecomposition temperature. As a result of the decreased conductivity,after the decomposition of these minerals, the base line comes at a higherlevel than before. This effect may be decreased by mixing the sampleswith a-AIzO3. Zinnwaldite and lepidolite melt at their decompositiontemperature:ca. 900o C. Thereby the thermocouple is covered with atight glassy substance and the contact between the sample and the sam-ple holder is disturbed. As a result of the decreased conductivity of thesample, the d,.l.o. curve moves downwards, thus making any quantita-tive analysis impossible. Another inconvenience is that after the testthe thermocouple has become useless for further analysis. Incompleteoxidation of carbo lignin during heating caused by lack of sufficient oxy-gen, is manifested by a gradually downward movement of the d,.t.a.curve. In this case only reliable results for the heat of combustion ofthis organic-see Fig. 4-can be obtained if the sample is diluted 1:29with a-Al:Os.

From the above it may be concluded that with a d.t.a. great difier-ences may be expected in the intensity, the area and the shape of thethermal peaks of minerals from difierent origins. This holds especiallyfor clay minerals where variations in particle size of the crystallites andchemical composition (ion substitutions) in the tetrahedral and octa-hedral layers, may be numerous in nature-see f.e. the analyses of Rossand Hendricks (1954) and Early et al. (1953)-and the minerals beingmoreover, mostly coated with various kinds of amorphous substanceswhich also influence the thermal reactions.

QUANTITATIVE DIFFERENTIAL THERMAL ANALYSES OF CLAY 235

Figure 5 shows the d..t.a. results ol 12pt samples of montmorillonitesfrom various, pure deposits. They were obtained by sedimentation of theoriginal samples in Atterberg cylinders with 0.05 N NH4OH as peptisa-

H y d r o r g i l l i t c

m q . < - A , l O- 2 a

Z i n n w o l d i t e

L c p i d o l i t c

C orbo- t ign in + 2 e o m g A - A l 2 O 3

iL

Q u o r t z

A I b i t c

A n o t o s e

mg + 2oo mg ,{ -A l o

4OO m9

4OO m9

4OO mg

loo 200 300 4m 500 600 7@ 800 900 looo

t cmpe ro tu l e i n oC

Frc.4. D.t.a. of minerals which lose large amounts of water (hydrargillite), which melt(zinnwaldite, lepidolite) or which burn ofi (carbo-lignin) when heated. Furthermore ofquartz and of some inert minerals (albite, anatase).

tor. The excess of electrolytes was washed out by filtration and the purityof the dried samples was checked by *-ray diffraction after the glycerolmethod of MacEwan (1946). The curves show evidently that the peak

5 See for this oualification note 4.

236 H. W. VAN DER MAREL

temperature, the intensity and even the relative intensities between thehigh temperature endothermal and exothermal reaction of this claymineral vary considerably when it is registered by the d.t.a. method.

Figure 6 shows the d.t.a. results oI the 12p, separate of soils of various

C loy Spur-Wyoming

Li t i l€ Rbck -Aikonsos

Po l kv i l l c - 'M i ssou r i

Bu rns -M ies i ss i pp i

B. Fourchc -S. Dokoto

C hombers -A r i zono

Yovopo i -A r i zono

G o u g e c l o y - N c v o d o

Otoy c l oy -Co t i f o rn i o

M o r o c c o - N A f r i c o

S u r rey -E ng l ond

Gc i scnheim -Gcrmony

Vouc lusc -F foncc

Bon tom - l ndoncs io

too 200 300 400 500 600 7c,0 goo too tooo

tcmpero ture in og

Fre. 5. D.t.a, of dried (1 hour at 105' C.) samples(2 p, each 400 mg. of pure (verified byr-ray analysis) montmorillonite from difierent localities.

A t

oo

7 5

6 7

a 2

E ng l ond

JOVO

Born?o

S urnot r o

D u t c h

G u iono

C ?y l on

G o l d C o a s t

P r a t o r i o

B r o z i r

roo 200 300 400 500 600 700 800 900 tooorcmpc ro tu re i n oQ

Note: Endothermal effect at ca 325" C. and exothermal efiect at ca 350' C. of some

samples caused by small amounts of hydrargillite and iron oxides respectively.

Frc. 6. D.t.a. of dried (1 hour at 105' C.) samples 12p, each300 mg. of pure white

kaolinite from Cornwall (England). Furthermore of high graded kaolinites (2p from soils

f rom difierent localities,

238 H. W. VAN DER MAREL

origin which alter x-ray analysis contain practically only kaolinite. Theseparate was obtained by first treating the original samples with HzOr(to destroy easily decomposable organic matter). They were afterwardstreated in the same way in Atterberg cylinders, etc., as is described formontmorillonite. The d.t.a. curves show that also in this case, the peaktemperature, the intensity and the relative peak areas of the endothermal

Tasrn 1. Pne Arue ol Punn, Wnrrr, Wor-r- Cnysrnrlrzeo (Vrnrlreo sv r-RA.vANar.vsrs) Dnrno (1 Houn er 105' C.) Keor,rNrrr (2p ewo ol Dnreo (1 Houn er 105' C,)Powonnrr CuSOr'5HzO (Srnxoml) Drurno wrtr a-AhOs 1.50:2.50 Equer.rv Pacxront rnn Saupr,n Hor"onn elro ANer.vzno wrrn Tr{E Sew 1s"*lro"ovptx,. D.t.a, ol Kao-

r,rNrrES Ar,tpnNerrrc wrrrr rrrosE ol Sraxoano CuSOr.5HzO Sauplrs

KAOLINITE STANDARD

Endothermal(ca 600' C.)

in cm.2

Exothermalca( 950' C )

in cm.s

Endothermal

Exothermal

Endothermal(ca 150" C.)

in cm.2

rol

o ( 4 1

122016332540

292460

1 46531282023263 2 73293301 1 5328

1 9 . 814.62 2 . O26.61 9 . 51 1 '

1 , 7 . 21 8 82 2 . 32 r . 820.9t 7 . 2, L 7

2 6 . 52 2 12 4 . O2 2 . 7

4 . 6

4 8J O

4 . 84 . 0+ . 45 . 34 . 43 . 9

6 0

6 . 0J J

4 . 3 04 . 2 94 . 5 84 . 7 54 . 1 54 . 6 24 . 3 04 . 2 74 2 14 . 9 55 . 1 34 . 9 14 . 1 24 . 2 13 6 84 . 3 64 8 34 . 6 7

3 1 . 83 3 33 2 . 83 2 . 23 2 . 63 2 . O3r .23 2 . 63 3 +3 1 83 2 . 63 1 . 53 3 . 23r .73 0 . 330.73 3 . 53 1 . 13 0 . 5t 1 7

Dutch Guiana

Gabon, CongoCornwall, EnglandZettlitz, TsechoslovProvencel -R"i-, f"

tuot"

JavaBangkafem. J lan tar l ^

. iJumauaranol. 5atel./Dhong Thuan, ThailandMurfreesboro, Ark, l

, ls a r n ' L i l o l '

. l - , ^ .Macon, Ueorgra lU 54.Dry Branch, GeorgialNew ]ersey )Mesa Alta, Nw. Mexico

Arithmetical mean (r) 3 2 . 0 6

Standard error of the mean (s4) :o 22 cm.2 :O.69Ta.Standard error of the single determination (s):0.97 cm.z:3.02To.

and the exothermal reaction vary considerably.To determine the magnitude of the variations in the heat of trans-

formation of kaolinite, being the least variable clay minera"l, d,.t.a. wereperformed on the 12 p separates of pure, white, well crystallized (verifiedby *-ray analysis) kaolinite and under conditions where errors due thed.t.a. method are eliminated.

The results are summarized in Table 1. It may be concluded that underthese optimum conditions, the measured heat of transformation of the

Note: By calculation was lound for the CuSOo' SHzO-CuSOr' HrO transformation at ca 150o C.

QUANTITATIVE DIFFERENTIAL THERMAL ANALYSES OF CLAY 239

standard chemical used, is practically constant. rt has a standard devia-tion of the single determination (s) of only 3.02/p for its first (CuSOn.5HrO-+CuSOn.HzO) endothermal effect (peak temperature: ca. 150o C.).However, the endothermal and the exothermal reaction of kaolinitesfrom difierent origin has a peak area which varies from 14.6 cmJ to 25.7cm.2 and from 3.4 cm.2 to 6.3 cm.2 respectively.o The relative intensities ofthe two thermal reactions varies from 3.7 to 5.1. Another diffrculty isthat many minerals have reactions which overlap each other for a largepart e.g. kaolinite and quartz, kaolinite and illite, glauconite and illite-see for further samples also Table 1 of De Bruijn and Van der Marel(1954), Part II, page 412. Only exceptionally the overlapping can beeliminated by use of a slow-heating rate, or by heating the sample invacuum-, in nitrogen- or in carbon dioxide atmosphere-see for the latterRowland and Lewis (1951) and Rowland and Beck (1952) in case ofdolomite.

Then there is still the difficulty that varieties of a mineral may occurin naLure, which are of the same structure and composition, but differinglargely in their thermal efiect-see ior a quartz McDowall and Dunn(1947), McDowall and Vose (1947 ,1952), Fieldes (1952), Keith and Tut-tle (1952) and Lewcook and Wylde (1953).

DrscussroN

Particle size, degree of crystallinity of the crystallites, kind andamount of ion substitutions and amorphous coatings (Beilby layer) mayvary considerably for a certain mineral of different origin. As they greatlyinfluence the heat of transformation of a mineral as registered by a d,.t.a.the application of this method for quantitative purposes will thereforebe very restricted. This conclusion holds especially for the group of clayminerals. The same factors have so far resisted also a quantitativedetermination of this group of minerals by the r-ray method.

The quantitative d.t.a. method can only be applied to well crystallizedminerals of well defined chemical composition if they give sharp diagramsof great intensity and if they are formed in nature under practically thesame conditions. However, attention should also be paid in this case to thesample investigated, as very fine crystallites and small amounts of chlo-

6 For the decomposition of 150 mg. CuSOr. SHrO to CuSOr.HzO at ca 150o C. 33 cal-ories are needed (see tables of constants). This thermal effect is registered by the it.t.a.apparatus here used as a peak area of. 32.06 cm.2 (standard deviation of the mean:0.22cm.2). Thus thevariation in the thermal transformations of pure kaolinitewhen heated ina d,,t.a. oven, is as follows:

endothermal reaction 100 to 176 cal. per gramexothermal reaction 23 to 43 cal. per gram

240 E. W. VAN DER MAREL

rides or carbonates of alkali and alkaline earths may largely disturb the

thermal effects. However, there lies an open and at present only acci-

dentally explored field for the d.t.a. method in the study of the heat of

trans{ormation of HzO and other polar molecules bound to rest valences

at the surface of minerals, or to the exchangeable cations of these miner-

als. Another possibility for the d.t.a. method is the study of ion substitu-

tions in a certain mineral and their place in the crystal structure in com-

bination with chemical and r-ray analysis. In these cases small difierencesin the physical behaviour of the HrO molecules (e.g. when stirred to an

ice-like state) or in the strength by which OH ions are bound in a crystal,

may cause large difierences in the peak area or the peak temperature of

a d.t.a. Their registration can be made more accurate and therefore the

application of. the d.t.a. technique as a very sensitive calorimetric method

for quantitative purposes enlarged, (1) bV carrying out the analyses at a

high heating rate in vacuum-, argon- or carbon dioxide atmosphere; (2)

by use of the more sensitive Au, Pd/Pt, Rh (Pallaplat) thermocouplewhich is also resistant to oxidation at the higher oven temperatureslT(3) bV placing the thermocouple not inside the sample holder as usual,

but as recently suggested by Boersma (1955), outside it. In that case thepeak area is no more dependent on the heat conductivity, the heat ca-pacity or volume changes of the mineral investigated, but merely on theproduced reaction heat of the mineral sample and a certain calibration

constant of the apparatus.s

AcxNowr,nocMENT

The author is indebted to Ir. G. De Josselin De Jong (Soil Mechanics

Laboratory, Research Department, Delft, Netherlands) for collabora-

tion.

Rrlrnewcrs

Ar-rewlv, W. H. (1949), Difierential thermal analysis of clays treated with organic cationsas an aid in the study of soil colloids: Soil Sci,. Soc. Amer. Proc.,13,183-188 (1948).

Anr,r.rs, P. L. (1951), A study on the difierential thermal analysis of clays and clay min-

erals: Thesis Agricult. Univ. of Wageningen (Netherlands) 131 pp.

Bensueo, I. (1948), Vermiculite and its relation to biotite as revealed by base exchange

reactions, r-ray analysis, difierential thermal curves and water content: Am. Mineral.,

33, 655-678.

? Manufactured by W. C. Iferaeus, G.M.B.H., Hanau (Germany)' It has a E.M.F. at

100' C., 300" C., 600' C. and 900o C. ol 2.86, 10.60, 24.72 and 40.35 mV, respectively.

The corresponding thermal efiects of aPt/Pt, Rh therrnocouple are 0.64, 2.31,5.23 and

8.45, respectively.s See for details on the factors afiecting the application of quantitative mineral analysis

with the commonly rsed il.l.o. equipments and the construction and the perspectives of

the Boersma sample holder, the articles by De Josselin de Jong (1956) and Boermsa (1955).

QUANTITATIVE DIFFBRENTIAL TEEKMAL ANALYSES OF CLAY 241

Blnosnm, I. (1950), The efiect of the interlayer cations on the expansion of the mica typeof crystal lattice: Am. MineroL.,35, 225-238.

Brrmv, G. (1921), Aggregation and flow of solids: Macmillan, London.Bnnc, L. G. (1943), Influence of salt admixtures upon dissociation of dolomite: Compt.

Rmd'. (Dokl'od'y) d'e l.'Aca.il. il. Sci. de /I/.R.S.S., 38, 2+-27 .Rrno, L. G. (1945), On area measurements in thermograms for quantitative estimations

and the determination of heats of reaction: Compt. Rend,. (Dokla'd.y) ilel'Acad. d.. Sci.I/.R.S.S., 49,64845r.

Brnrcurauen, L. H. (1944), Difierential thermal analysis of quartz: U. S. Bw. Mi'nes,Rept. 3763.

Bnnrvel, J. D., ano FowLER, R. H. (1933), A theory of water and ionic solution withparticularreferencetohydrogenandhydroxylions:,I. Chmt.Phys., l'515-548.

Bornsu.l, S. L. (1955), Theory of difierential thermal analysis and new methods of meas-uring and interpretation: J. Am. Ceram. Soc.,38, 281-284.

Bnl-or,ev, W. T., Bunst, J. F., auo Gnen, D. L. (1953), Crystal chemistry and difierentialthermal effects of dolomite: Am. Minerol,., 38, 207-217 .

Bnmrao, L., C,lov, J. G., ar.ro Hrrnntcrs, S. B. (1950), The determination of halloysiteand its use in the study of red soils : 4llz I nternatl. C ong. S oi.l, S ci,., P ort. l, pp. 313

Bnaueo, L., Clov, J. G., Hnwonrcr<s, J. 8., eNo Swrnor-ow, M. (1952), Criteria for thecharacterization of kaolinite, halloysite and a related mineral in clays and soils: SoilSci., 73, 273-287 .

Bnuryx, C. M. A. DE, AND M.arul, H. W. vaN nnn, (1954), Mineralogical analysis of soilclays. Part I: Introduction and difierential thermal analysis; Part II: Examples ofmineral analysis by r-ray difiraction and difierential thermal analysis: Geologie enMijnbouw, 16 (nw. ser.), 69-83, and 407-428, respectively.

BuoxtKov, P. P., arvo BonnovrNr, D. P. (1938), The influence of admixtures on thedecarbonation of dolomite: J. AppI. Chem. (U.S.S.R.), f f' 1151-1154.

Cerr,r,inu, S., awo HfNrN, S. (1944), Sur I'Origine de quelques anomalies, pr6sent6es parIes courbes thermiques de certains montmorillonites: C. R. Acail. Sci.,2l9' 685-686.

Cerlr,inr, S., alrm HfNnv, S. (1947) Application de I'analyse thermique differentiellee l'6tude des argiles des sols : Ann. A gron., 17, 23-7 2.

CerlrDru, S., ewo Hr6wrN, S. (1948d), Essais de synthbse des mineraux argileux: Verre etSilic. Ind.,13,63-64.

C.lrr,rinr, S., exo Hfrnv, S. (1947) Sur une synth6se de I'antigorite i basse temp6rature:C. R. ilel'Aco|l. et 5ci.,224, 1439-lll4JJ.

Carrr"tnr, S., nmo H6wrn, S. (1948b),1'Analyse therrnique et son interpretalion: ActesC ongr. I nt ernatl. C tr am. H ol,lan de, S chnenin gen, 13 7-1 5 1.

Clrr,r.ine, S., and Hfr.rrN, S. (1949), Sur la signification des resultats de I'analyse ther-mique difierentielle: Veme et Sil,. Ind,.,13, 114-118.

Cr,nrr,,tlro, D. W. (1951), Thesis University of Glasgow, England.Clorlervo, D. W., CuwNrNc, W. M., Ar.ro Rrtcrro, P. D. (1952), Physico-chemical studies

ondusts:IAhigh-solubi l i tylayeronsi l iciousdustssurfaces:J. APpl,.Chettt ' . ,2,3141.Clrr.r-aNo, D. W., arr'o Rtrcnm, P. D. (1952), Physico-chemical studies on dusts: II

Nature a.nd regeneration of the high-solubility layer on silicious dusts: "I. Appl'. Chem.2 ,4248.

Drrpsmn. P. B. (1951), Thesis University of Glasgow, England.Dnrresrrn, P. 8., ewo Rrrcrrm, P. D. (1952), Surface of finely ground silica: Nature, 169,

538.Dnlmsrrn, P. 8., .lxo Rrrcnre, P. D. (1953), Examination of finely ground quartz by

difierential thermal analysis and other physical methods: J . ApPl. Chem., 3, 182-t92-

a A a H, W. VAN DER MAREL

DnAcsnonr, R. D., KrssrNcnn, H. E., aNl Pumnls, A. T. (1951), X-ray study of the de-composition of kaolinite : S o il, S ci., 7 l, 439-448.

Eanr,v, J. W., Mrr,Nt, J. H., erll Mc. Vnlon, W. J. (1953), Thermal dehydration andr-ray studies on montmorillonite: Am. MineraL.,38, 770-783.

Eenr,v, J. W., Osrrraus, B. 8., ,rlro Mrr,re, J. H. (1953), Purification and properties ofmontmorillonite Am. MineraJ., 38, 7 07-724.

Fausr, G. T. (1949), Dedolomitization and its relation to a possible derivation of a mag-nesium-rich hydrothermal solution: Am. Mineral., 34, 7 89-823.

Feusr, G. T. (1951), Thermal analysis and *-ray studies of sauconite and some zinc min-erals of the same paragenetic association: Am. Mineral.,36,795-822.

Frrlons, M. (1952), Abnormal thermal behaviour of a-quartz from some New Zealandsoils: N atur e, 17 O, 366-367 .

Fr,oon, H. (1950), Note on the thermal decomposition of dolomite: Det. Kongelige NorskeVi.denskabers S elskab F ord,hand.l.inger, 22, 188-192.

Grrn, J. G., Rrrcnrn, P. D., alro Sreme, J. W. (1953), Electron optical examination offinely ground silica: J . Appl. Chem.,31 213-218.

Gress, A. D. (1954), High temperature phases from kaolinite and halloysite; Am. Minerol.,39, t93-207.

Gner, D. L' (1952), Preliminary report on the variations in the difierential thermal curvesof ironlow dolomites: Am. Mineral., J7, l-27.

Gmrrvn-Korr,nv, R. (1953), Interpretation ol d..t.a. diagrams; the low temperature endo-thermic peak: Clay Min. Bull.,2, 79_€,3.

Gnru, R. E., AND Rowr,aNo, R. A. (1942), Difierential thermal analysis of minerals andother hydrous materials: Am. Minero.l.,27,746-761, and 801-818.

Gnru, R. E. (1947), Differential thermal curves of prepared mixtures of clay minerals:Am. M iner aJ., 32, 493-50I.

Gnnrsruw, R. W., Hrerow, E., AND Ronenrs, A. L. (1945), II Thermalanalysis methods:Trans. Brit. Ceram. Soc.,44r 7642.

Gnrusu.l,w, R. W., awo Ronenrs, A. L. (1953), The quantitative determination of someminerals in ceramic materials by thermal means: Trans. Bri,t. Ceram. Soc., 52, 50-61.

Gnuron, J. W. (1939), Ammonium mica synthesized from vermiculite: Am. MineroJ.,19,JJ /- . ) / . )-

Gnuvon, R. M., Hnlrny, E. C., .a,r.ro Hevsmr, I. 0g4g), Suppression of thermal reactionsin kaolinite: Am. Minaal., J4, 869-873.

Ilrenv, H. H. (1952), Difierential thermal study of Colorado oil shale: Am. Minera.l.,37,804-811.

IfoNnnrcrs, S. B., Nnrsox, R. A., allo Ar,nxa-wonn, L. T. (1940), Hydration mechanismof the clay mineral montmorillonite saturated with various cations: J. Amer. chem.Soc.,62, 1457-146/'.

Jossrr.rx on JoNG, G. DE, (1956), calibration of the difierential thermal analysis apparatusfor quantitative purposes: To appear in J. Am. Ceram. Soc.

Knr-lev, w. P., arvo Pacn, J. B. (1943), criteria for the identification of the constituentsof soil colloids: Soi.l Sci. Soc. Amer. Proc.,7,175-lgl (1942).

Knrerr, M. L., allo Turrr,n, O. F. (1952), Significance in variation in the high-low inversionof quattz: Am. J . Sci..,l, 203.

Kann, R. F., aNo Kulr, J. L. (1948), Multiple differential thermal analysis: Am. Minera.l,.,33, 387-419.

KtyounA, R., ,LNo Sar,l, T. (1950), The quantitative analysis of the CaCOrCa(OH)rMg(OH)e system by differential thermal analysis: f. Ceram. Assoc. fapon, 58,3-6.

Kun, J. L., el[o Turns, A. (1951), Differential thermal analysis of natural hydrous ferricoxrdes : A m. M iner atr., 36, 23-4.

QUANTITATIVE DIFFERENTIAL THERMAL ANALYSES OF CLAY 243

Kur,e, J. L., KrNr, P., ,r.rn Knn_n, p. F. (1951), Thermal study of the Ca_Mg_Fe carbonateminerals: Am. Minerd., J6, &H70.

Lews, W. D., am Pecn, J. B. (1946), Changes produced in kaolinite by dry grinding:S oil. S c,i., 62, 3 l9-336.

Lrwcoor, W., aNo Wvr.ou, J. H. (1953), The behaviour of silicaTr ans act. Brit. Cer am. S oc., 52, 2ZS-258.

LnnltANN, F. (1952),-see Dreiziger Jahrestagung der Deutschensellschaft von 1. bis 10. September 1952 in Regensburg referred315.

MacEweN, D. M. c. (1946), The identification and estimation of the montmorillonitegroup of minerals, with special reference to soil clays: soe . chem. Ind., 6s, 2gg-307 .

Mac<nNzrn, R. c. (1950), Some notes on the hydration of montmorinonite: cray. Min.Bu l l . ,1 ,115-119.

MacrrNzrr, R. C., e*l Mrr-r.rr, A. A. (1953), The effect of grinding of micas: I Muscovite:Mi,neral. M ag., 30, 178-185.

MacxrNzrr, J., aNo Lerrn, J. R. (1953),-see the Discussion at the end of the article ofGrimshaw, R. W., and Roberts, A. L. (1953) in Trans. Brit. Ceram.Soc., 52, 50_61.

Mennl, H. w. ven nnn (1950), rdentification of minerals in soil clay by x-ray diffractionpatterns : Sail Sci., 70, 109-136.

Marut, H. W. vmq onn (1954), Potassium fixation in Dutch soils: Mineralopical anal"tsis:S o il S ci.., 7 8, 163-17 9.

Mantnt, R. T., axo Russer,l, M, B. (1952), Clay minerals of four Southern New yorkSoils: Soil Sci., 74, 267-279.

McDowerr,, I., luo DuNN, L. R. L. (1947), New Zeal. pottery and. Ceram. Res. Assoc.Techn. Pap. no. l.

McDowerr-, r., eNl Voso, w. (1947), New Zeal. pottery anir, ceram. Res. assoc. Techn,Pap. no. l.

McDowar,r, I., exn Vosn, W. (1952), Anomalous properties in quartz: Nature, 17Or 366.Mer,oA', R., ero RouunrsoN, H. s. (1952), Die Bruchfliichen feingemahlener Mineral-

pulver : T onind,. Ztg., 76, 365-368.Munnav, J. A., Frscnnn, H. C., arvo Srreon, R. W. (1951), Thermal analysis of limestones:

Proc. 49th AnnuaL. Conl. Natl,. Lime Assoc., Woshi.npton.Munuv, H. H. (1954), Structural variations of some kaolinites in relation

halloysite: Am. Mineral., 39, 97-108.to dehydrated

bricks on reheating:

Mineralogischen Ge-to in Tonind. Ztg.,76,

Nacnrscnuror, G., GonnoN, R. L., aNn Gnu.lrw, O. G. (1952), Surface of finely groundsili,ca : N ature, 169, 538.

NreuwrNsenc,C.J.vnN,eNnPrnrnns,H.A.J.(929),Thedehydrat ionvapourpressureof kaolin : Rec. Traa. C him. P oys B as., 49, 40H16.

NonroN, F. H. (1939), critical study of the difierential thermal method for the identifica-tion of the clay minerals: -I. Amer. Ceram. 50c.,22,5L43.

Nonro*, F. H. (1940), Analysis of high alumina clays by the thermal method: J, Amer.Cer am. S oc., 23, 281-282.

onctr., J. (1935), I'Emploi de I'analyse thermique difierentielle dans la determination desconstituents des argiles des laterites et des bauxites:7th Congr. Internatl. Mi.nes. Met.Geol,. Appl. Paris, Secti,on Geol. appl,.,l, 359-373.

Pecr, J. B. (1943), Difierential thermal analysis of montmorillonite: Soil sci,., s6,273-2g3.Panxnnr, C. W., PenxrNs, A. T., ,llfl) DRAosoonr,, R. D. (1950), Decomposition of min_

erals by grindingl. Kans. Acad. Sci. Trans., 53,3g6-397.PemrNs, A. T. (1950), Reactions of bentonite as influenced by adsorbed cations and

grinding: Soi,l. Sci. Soc. Amer. proc., 14, 95-96 (lg4g).

H. W. VAN DER MAREL

RosrnrsoN, R. H. S., BrlNor.nv, G. W., .lr,rn MAcrcNztl, R. C. (1954)' Mineralogy of

kaolin clays from Pugu Tanganyika: Am. Mineral',3g' 118-138'Ross, Cr.nnnNcr, S., aNn llnNontcrs, S. B. (1945), Minerals of the montmorillonite group:

tJ. S. Deport. Int. Profess. Pap.2O5-8,23-79.Rowr-.trto, R. A., exo Lawrs, D. R. (1951), Furnace atmosphere control in difierential

thermal analysis: Am. Miner.,36' 80-91.Rowr-m{D, R. A., exo Bncr, C. W. (1952), Determination of small quantities of dolomite

by difierential thermal analys is : A m. M i,ner aJ., 37, 7 6-82.Seuerrnn, G. (1950), Sur I'influence de la dimension des cristaux de chlorites sur leurs

courbes d'analyse thermique difierentielle: Bull. Soc. Franc. Mineral'.,73r 43-48'

ScnuylnNsoncn, J. VaN, .a.Nl AnnNs, P. L. (1950), The electrokinetic behaviour of freshly

prepared a and 'y FeOOH: Rec.d.. Trao. Chi.m. Pays Bas,69, 1559-1565'Scnwon, Y. (1950), Les carbonates rhombo6driques simples et complexes de calcium,

magn6sium et Ier: Publ.icotion lechni,qte no. 22, centre il.'Etud,es et ile Recherches ile l,''

Inilustri,e iles Liants H yd'rauJi'ques, Pais, 105 pp.Srnrr, S., BerxnlneMrn, L. H., Pnsr, J. A., eNo DAvtns, B. (1945), Difierential thermal

analysis, its application to clays and other aluminous materials: U. S. Bur' oJ Mi'nes,

Techn. PaPer n0.664.SrurtnmEru, N., Wnnr, T. L., aro UneNovsrY, B. (1951), Developments in research on

the burning and hydration of lime and on its use in building: Proc. Brit. Bl.tlg. Res.

C ongr. P ar l'. 2, 12W126.Vor.o, Ml.nloarr , J . 0949), Difierential thermal analysis: ,4 nol. Chem., 21, 683_{,88'

wnss, T. L. (1953), Some applications of differential thermal analysis and findings using

this technique: S. AJrican InL. Chem',7r22+-232.w4r,xrr, G. F. (1950), Trioctahedral minerals in the soil clays of northeast scotland:

M inerol. M ag., 29, 7 2-U'wa1gn, G. F., lNo Mnrvn, A. (1950), Hydration of vermiculites saturated with various

cations: Trotcs.4th Internatl'. Congr. Soil Sci. Amsterilamr2 6247.

Manuscript receiaed. Apri.I 22, 1955.