THERMAL ANALYSIS STUDY OF THE NATROLITE GROUP

C. J. PnNc*

AssrRlcr

The ttrree members of tle natrolite group, natrolite, scolecite, and mesolite, from tendifierent Iocalities have been examined by both static and difierential thermal analysismettrods. The thermal changes taking place in the specimens at reaction peak temperaturesgiven by the DTA curves have been investigated by coordinated optical and r-ray diffrac-tion methods. Despite their close similarity in composition and crystal structure, the ttrreezeolites show different thermal behavior. Natrolite gives off its water rapidly in a singletemperature range, whereas scolecite and mesolite lose their water in two and three stagesrespectively. It is thus inferred that there is only one type of water molecule in natrolite,while t}le water molecules in scolecite and mesolite consist, respectively, of two and threetypes witl different bond strengtls. This is supported on structural grounds. The range ofstability also is different for the three zeolites. Natrolite will not break down completelyuntil about 940o C., but scolecite collapses structurally at 560' C. and mesolite at a stilllower temperature of 490o C. This is explained as a consequence of their different degreesof hydration.

INtnooucrroN

Much study has been made of the dehydration phenomenon of zeo-lites and their changes in physical properties that accompany the efiectof decomposition. The method generally employed is to determine theIoss of water at various elevated temperatures under equilibrium condi-tions with or without control of the water vapor pressure. Owing to cer-tain technical difierences and possible variations in the composition ofthe specimens studied, the results are not consistent and hence the inter-pretations of their thermal history vary.

The purpose of this report is to give the results of an investigation ofthe thermal behavior of the natrolite group by both static dehydrationand difierential thermal analysis methods. The natrolite group wasselected because its members, natrolite, scolecite, and mesolite, have arelatively simple and definite composition and have been nearly complete-ly studied structurally. An attempt has also been made to correlate thethermal efiects with the structures of the zeolites.

Although the three zeolites are isostructural and show markedly simi-lar physical properties, they do not undergo the same thermal changes.Further, the intensity of those thermal reactions due to dehydration andthe temperatures at which they take place are different for the three be-cause each contains different amounts of water and the water moleculesmay occupy different lattice positions. Thus, the thermal reaction data

* Department of Geology, Columbia University, New York, N. Y. Present address:University of Wisconsin, Madizun, Wisconsin.

834

THERMAL ANALYilS STUDY OF THE NATROLITE GROUP 835

should be useful not only for the purpose of identification but also in the

interpretation of the structure.

AcrNowr,BlGMENT

The research was undertaken at Columbia University at the suggestion

of Prof . P. F. Kerr. The writer wishes to express his indebtedness to Prof.

Kerr for his interest and advice throughout the study and critical reading

of the paper. He is also grateful to Prof. J. L. Kulp and Prof' R. J'Holmes for their helpful suggestions and criticism. The encouragementand assistance given by Prof. Chas. H. Behre, Jr. during the course of

the work are highly appreciated.

MarBnrar-s

The specimens used for the study were obtained from the Egleston

Mineralogical Collection of Columbia University and included the

following:

Natrolite, Giant's Causeway, Ireland. White, radiating needle-Iike crystals.

Natrolite, West Paterson, N. J. White, radiating hairy needles.

Natrolite, Brevig, Norway. White accicular crystals, partly admixed with chlorite.

Natrolite, Auvergne, France. Groups of white, stout, translucent crystals with

good prisms and pyramids.

Scolecite, Poonah, India. White, slender, glassy needles in radiating groups'

Scolecite, Bombay, India. Thick clusters of long white glassy crystals.

Scolecite, Moore, Mercer County, N. J. White accicular crystals.

Mesolite, Peter's Point, Nova Scotia. Divergent groups of white slender needles.

Mesolite, Cape d'Or, Nova Scotia. A mass of radiating white, fibrous crystals

densely matted togettrer.(10) Mesolite, Scotland. Groups of white fibrous crystals.

The three zeolites, natrolite (NarAlzSiaOro'2HzO), scolecite (CaAh-

SirOro' 3HzO), and mesolite (CarNarAlusisOgo' SHzO), have practically a

constant composition as first pointed out by Winchell (1925) and show

only limited ionic substitutions of the types Na2--+Qa and Na-+K (Hey

& Bannister ,1932,1933, 1936). Chemical analyses of previously describedspecimens probably from the same localities as a number of the specimenshere studied are given in the following table (Table 1). It is believed that

they are representative of the chemical composition of the three zeolitessince all the tested specimens have been thoroughly examined opticallyand by means of. r-ray diffraction to ensure purity and identity.

ExpnmrunNrAr EeurpMENT AND PnocBpunB

Specimens were first examined with a binocular microscope and materi-al was selected for further study. The standard immersion method wasused to determine the optical constants. X-ray powder photographs of

(1)(2)(3)(4)

i/ q\

(6)(7)(8)(e)

836 C. J. PENG

all specimens were taken by means of a Philips r-ray unit to confirm theidentity of the minerals. X-ray spectrometer measurements were madeon some of the specimens. Copper-K radiation with nickel frlter wasused throughout.

Both static dehydration and differential thermal analysis methodswere applied to the study of the thermal behavior of the zeolites. Theprocedure to obtain dehydration data is similar to that generally followed.Material was heated in a Freas electric oven for successive periods of

T.lsrr 1. Cnnurclr ANlr.vsus or Narnorrrr. Scor,rcrre aNn Mnsor.rrs

Natrolite Scolecite Mesolite

t , , nsio:AlzOsNuroKzOCaOHrO

2 7 . 2 115 .860.06

9 . 7 0

47 .882 6 . 1 215 .63

0.459.80

47 .6027 .4015 .360 .230. 139 .47

45 .1625.900 . 1 60.06

14.8613 .66

46.O126.664.660 . 2 09.88

12.69

46.2626.484 . 9 8

9 . 2 413.04

Total 99.88 100.19 99.80 100.36 100. 10 100.00

1. Kinbane (White Ilead), County Antrim (Giant's Causeway), Ireland. AnalystF. N. Ashcroft , Mineral. Mag., L7, 3O7 (1916).

2. Tour de Gevillat, Auvergne, France. Analyst F. Gonnard, Boill. Soc. Franc. Mi.nerol,.,14, 170 (1891).

3. Puy de Marmant, Puy-de-Dome, France. Analyst M. H. Hey, Minerd. Mag.,23, 246 (t932).

4. B. M. 33887, Syhadree Mts., Bombay, fndia. Analyst M. H. Hey, Mineral. Mag.,24,228 (t936).

5. Poonah, Bombay, India. Analyst G. Tschermak, Sitzungsber. Akad. Wiss. W'ian,nahnwi,ss. Kl., Abt. I , 126, 541 (1917).

6. Cape d'Or, Nova Scotia. Analyst E. W. Todd, Unia. Toronto Stud.i,es, Geol,. Sa, no.14,57 (t922).

7. Isle of Skye, Scotland. Analyst M. F. Heddle, Mi.neral,. Mag.,5, 118 (1883).

20 hours at about 20o intervals up to 3500 C. The temperature wasmeasured with a thermometer inserted into the furnace through a holeon the top of the chamber. For higher temperatures, heating was carriedout in a rheostat-controlled electric mufle for 3 hours at about 50o C.intervalsl the temperature was measured with a Brown portable electricpyrometer with an accuracy of about 5o. After a desired temperaturewas reached, the sample was cooled to room temperature in a desiccatorand then weighed. This was followed by reheating it at the same tempera-

TEERMAL ANALYSIS STUDY OF THE NATROLITE GROUP 837

ture for another 2 hours in order to check the weight previously obtained.

Weight losses were practically the same before and after reheating and

only in a few cases was further heating called for to reach a constant

weight, indicating that heating over a period of 20 hours up to 3500 C.

and 3 hours at higher temperatures is sufficient to bring about equilibrium

at a given temperature. The process was then repeated at higher tempera-

tures. Each sample to be heated was ground to pass a 200-mesh sieve.

No attempt was made to control the water vapor pressure in the furnace.

The difierential thermal analysis method not only provides dehydra-tion data, but also reveals those thermal reactions that are not accom-panied by changes in weight, such as structural breakdown, recrystalliza-tion, phase transition, etc. The multiple DTA :uui,it developed in theMineralogical Laboratory of Columbia University (Kerr & Kulp, 1948)was used in the present study. The pulverized samples were packed

uniformly in the specimen recesses which, ! in. in radius and I in. deepfor each, are spaced symmetrically with the alundum holes with respectto the vertical axis of the furnace. Heating was carried out without cover-

ing from room temperature up to about 10500 C. at the rate of 12|o per

minute. Two or three runs were made on each sample. The high sensitiv-ity scale 2000 for the d.c. amplifier was used throughout, for which theamplification is about two. For identifying each thermal reaction, the

specimen was first heated up to the peak reaction temperature and then

examined by optical and r-ray powder methods.The endothermic peaks on the DTA cuwes of the zeolites due to loss

of water in general correlate well with the shoulders on their dehydrationcurves. But as a rule, they occur at higher temperatures than correspond-ing shoulders, although initial decomposition usually takes place at ap-proximately the same temperature in both cases. The difierence in tem-perature maximum between these two kinds of thermal curves observedin the present study is about 1000 C. The reason for the reaction-tempera-ture lag in the DTA curves lies in the dynamic character of the methodby which the sample is heated at a constant rate, thus extending thereaction over a longer temperature range, while each water loss shownon a dehydration curve is determined at a constant temperature underequilibrium conditions.

With the zeolites some experimental difficulties were encountered and

the most serious one concerned the DTA method. The three zeolites allfuse at about 10000 C. Since they adhere firmly to the sample holder afterfusing, the thermocouple head was often found to be either dislocatedor broken when a heated sample was taken out of the specimen hole forfurther study. This caused considerable trouble and discouraged theapplication of the technique to other zeolites.

838 C. I. PENG

TuBnlrar- ANer,vsBs

(A) Dehydration Curaes

Two specimens of each member of the natrolite group were selectedfor the static thermal analysis in order to verify their dehydration course.Two groups of specimens were tested, one comprising (1) natrolite,Giant's Causeway, Ireland, (2) scolecite, Bombay, India, and (3) meso-lite, Peter's Point, Nova Scotia; and the other group including (1)natrolite, Auvergne, France, (2) scolecite, Moore, N. J., and (3) mesolite,Cape d'Or, Nova Scotia. The specimens of each group were heated simul-taneously, but the study of the first group was made in January whilethe second group in June of the same year. The dehydration curve ofeach of the zeolites is essentially constant, although the specimens of thefirst group generally showed a water-loss at a lower temperature than thecorresponding ones of the second group (FiS.1 and Table 2). The dis-crepancies are explained as being chiefly due to the usual lower humidityof the air in winter.

The dehydration curves of the three zeolites are markedly distinctive.Scolecite loses its water in two stages and mesolite in three as shown bythe shoulders on their curves, while natrolite completes its dehydrationin a single stage, represented by a smooth straight line. Further, eachstage represents the loss of a definite number of water molecules in pro-portion to the total water content of the mineral. The last shoulder onthe curves of both scolecite and mesolite does not level off sharply butslopes upward gently over a large temperature range. Such an extendedand slow dehydration may result from an increasing firmness of bindingof water molecules in the structure at higher temperatures so that theirprobability of escape from the tightest lattice positions may be decreasedexponentially. But the considerably pronounced secondary break of thelast shoulder on the scolecite curve at about 4500 C., although not sharpenough, indicates a possibility that the two water molecules removedduring the shoulder may differ slightly in structural capacity.

The general shape and water-temperature relationship of the curvesof natrolite and scolecite are in good agreement with those of previousworkers (Rinne, 1890; Walker & Parsons,1922; Cavinato, 1927;H.ey &.Bannister, 1932, 1935). Milligan and Weiser (1937) reported that natroliteloses its water in two stages; but an examination of their curve revealsthat it is very similar to their scolecite curve and hence suggests thattheir specimen might be actually scolecite.

Some of the published curves of mesolite show a nearly continuousdehydration (Zambonini, 1908; Pelacani, 1908; Hey & Bannister, 1933),but others bear a striking similarity to the writer's curve (Fig.2). It isbelieved that the discrepancies are not a difference of matter but of de-

THERMAL ANALYSIS STUDY OF THE NATROLITE GROUP

TETPERATUSE II{ 'C

roo 4oo 5@ 600 79O

)t

IE

p s= 7

B GO to

3 r

t l

t 0

9'a

3cU -

3 s

o 5

$ toJ 5

I

o

t 4

t 5

l2

t l)!

. 8

E 'o

6 3o! s

a

2

I

o

ta

t!

tz

t l

toEUF

@U

UJo2

of

o

E

@=z

NATROLITE

I Giont 's CousewoY,

l r e l ond

2.Auvergne, Fronce

S C O L E C I T E

l .Bomboy , I nd io

2 'Moo rg N ' J '

UJ

)o=/o

GUo=?

@

-UJo-o

o

E

o-tz

l l MESoLTTE

l l t B2 'Pe le r ' s Po in i

I novo sco l io

a 3 'Copc d 'o r ,

Frc. 1. Dehydration curves of natrolite, scolecite' and mesolite.

840 C. J. PENG

Teern 2, Dnnvpr-a,rroN De.re or N,mnor,rrr, Scornctm, axo MBsor.rrn

(Percentage of water lost up to temperature stated)

Natrolite Scolecite MesoliteTemperature

in C.

67"82"

101"1270147'162"184"200"205'222"262"3000310'325"350"400'450'500"570'610'655'670'700'765"770"8000810"865"

0.080.040 . 1 10 . 1 40.300.300.30

0 . 3 60.482 . 7 08.40

9.459 . 5 09 . 5 59 . 3 19 . O 79 . 3 69.459.62

9 . 8 19 . 8 8

9.92

9 . 9 2

0.020.050 . 1 30 . 1 60 . 1 90 . 2 40 . 3 20.42

0 . 5 01 . 0 0

8.479 . 2 89 . 0 1

9 . 4 2

9 . 2 19 . 2 +

9.349 . 4 1

9.43

9.639 . 7 2

o.o20.050.050.050 .330.942 . M

4.064 .074 .564.89

5 . 147 .40

t0.4710.9011 .59t2 .6 r12 .84t3.23

13 .58I J . / J

13 .81

13 .83

0 .010 .010.060.090 . 1 00 . r 2o.943 .80

3 .994 . 1 5

4 .855.Ol5 . 8 29 .79

10.8810.99r1.9412.62

1 3 . 1 613.41

13 .53

13 .5613 .59

0 .01o .020 .220 .220 .621 . 0 12 . 5 7

3 . 2 33 .526.007 .o2

8 .309 .81

1 1 . 1 01 1 . 4 81 1 . 8 312.4812.51t2 .74

12.821 2 . u

12 .85

12 .85

0.030 .070 . 1 5o .230 . 2 50 .311.032 .89

3 .585 . 4 1

5 .949.46

10.8211 .30lt.6912.1112.5412.71

t2.8612.86

t2 .90

12.9212.96

0 . 130 .27o.410 .590 .620 .93r . 7 93 .61

3 .965 .92

7 . 2 rt t .2711 .3611.92t 2 . Mt2.49t2.9012.99

13.0813.04

1 3 . 1 3

1 3 . 1 913.23

Natrolite: 1. Giant's Causeway, Ireland; 2. Auvergne, France.Scolecite: 1. Bombay, India; 2. Moore, N. J.Mesolite:1. Peter's Point, Nova Scotia;2. Same as 1;3. Cape d'Or, Nova Scotia.

gree and may be explained as being due to the relatively small differencesin volatility between the water molecules given ofi in the three stages sothat the shoulders are not always well defined and sometimes may evenfail to show up.

(B) Diferential Thermal, Analysis

The only published DTA data for zeolites are the study of natroliteby Sveshnikovr and Kuznetsov (1946), but no details of their work areavailable. In the present investigation, four specimens of natrolite and

I

THERMAL ANALVSIS STIIDY OF THE NATROLITE GROUP 841

three each of scolecite and mesolite from different localities have been

examined by the DTA method. The specimens were fractioned into

three particle-size groups: (1) minus 80 mesh, (2) minus 80 and plus 200

mesh, and (3) minus 200 mesh. The DTA curves of the three fractions

of each specimen are essentially identical (Fig. 3 and Fig. 4) ' Some endo-

thermic peaks showed a small shift in peak temperature, but the varia-

tions are so irregular that it would be impossible to state definitely

l . m P . r o l u r . I n ' 0

Frc. 2. Dehydration curve of mesolite.

whether they are due to the factor of particle size or not. However, the

efiect of grain size on the sharpness of an endothermic peak is more dis-

tinct, since almost all endothermic peaks shown by the minus 200 mesh

fractions are sharper than the coarser ones. Furthermore, the endothermic

d.oublets were also only shown by the finer fractions' Since particles of

smaller grain-size should react thermally more uniformly, the sharpness

of the endothermic peaks seems to be expected.The DTA curves of both natrolite and mesolite show a high-tempera-

ture exothermic peak just above 10000 C. but experimental work showed

that this peak on both curves varied considerably in shape and peak

temperature. The exothermic reaction on the natrolite curve has a very

small intensity and was shown only by the minus 80 mesh fractions. It

results from recrystallization, but the small intensity clearly indicates

A C a v l . a l .

! I a l l . r a P a t . . n .

G l e l i c r

u2 C. J, PENG

Couscwoy, il r c l o n d , I

N A T R O L I T E

OEGREES CENTIGRAOEloo 200 500 a@ loo 600 7@ 8oo too tooo iloo

l l l l l l l l l r rIOO 2OO 5OO ,tOO 5OO COO 7OO tOO 9Oo |OOO ilOO

DEGREES CENTIORAM

- 80m!sh

2 . S o m a ,8O-20Omcah

4 tV Po lc r lon ,N J- 80 mc!h

5 S o m r ,8O-2OOmrth

6 S onc ,- zoom.rh

7 Auvcrgna,F r o n c a ,8O-2OOn.rh

8 Som. ,- 2OOmrrh

9 B r c v i g ,Norroy'- EOmarh

l O S o m g8O-2OOmtrh

| | Som! ,- 2OOmcth

l l l l l l

l l l l l l

Frc. 3. Difierential thermal analvsis curve of natrolite.

THERMAL ANALYSIS STT]DY OF THE NATROLITE GROUP 843

oEcRtlS cEirleiu

'oo rr 5r 1r 1@ To

71o .T r?o "ls nto

A. SCOLECITEI Moorq f l J

- EOn.h I

2soha. EO-zOOntth :

3 som., - 2OOn..h

4 Eohboyr Indlo- 8 O h . . h

5 somq EO-zOOmr.h I

6 Som., -2OOmath

7 Poonoh, Indio8O-2OO n..h

8 S o n r , - 2 O O m . r h

B. MEso|-rrEI S c o l l o n d , - 8 0 h a . h

l O S o m ! , 8 0 - 2 o o m c l h

l l S o m . , - 2 O O m a r h

l 2 P . l c r ' ! P o i n l ,N o v o S c o t i o- 8 0 m . B h

1 5 S o h . , g O - 2 O O n . t h

1 4 S o m c , - 2 O O m c ! h

15 cop. d'o.,l {ovd Scol io

80- 2OO n..h

16 Scn., -2OOn..h

C. t7 Mittur. of 33.gst L Ino l ro l i la (W Po lc r .o \

- \

N J ) o n d 6 6 . 6 6 4r c o l a c l l a ( 8 o n b c t , | |I n d l o ) . - 2 O O h . . i | |

r l l l l l l l l l lr @ 2 @ w @ 5 @ @ ? @ a @ m r @ [ o o

oEeiEE3 CEiTteiAE

Fro. 4. Differential thermal analysis curves of scolecite and mesolite.

C. J. PENG

that the thermal efiect must not be strong. fn other words, the reactionmay be a simple and easy one so that not much heat is evolved. Sincefiner particles should facilitate recrystallization, then this weak thermalefiect may have been so subdued in the fractions of smaller grain sizesas to be unable to show up on the D?,4 curves.

The exothermic peak on the mesolite curve is a sharp and strong one.While it appeared at a constant peak temperature but showing increas-ing sharpness with decreasing particle size on the curves of the Scotlandspecimen, it occurred at a higher temperature on the curve of the coarsefraction of the Peter's Point specimen than on the curves of its finerfractions. Since the results are not consistent, the evidence for the effectof grain size on the exothermic peak does not seem conclusive.

Other factors may have also contributed to the variations of the reac-tion peaks. A reaction may be delayed if the sensitivity and position of thethermocouple head in the sample hole cannot be maintained reasonablyconstant in comparative runs, as often occurred in the present work aspreviously pointed out. Any change in the sample packing conditions inthe sample holes may also impose some irregularities on the peak shapeand temperature of reactions. Since limited ionic substitutions are knownfor the zeolites as discussed in an early section, shift of reaction peaksdue to the presence of certain "impurities" is thus possible.

In spite of the aforesaid variations, the DTA curves of the three zeo-lites are as consistent and distinguishing as their dehydration curves.Each of the curves consists of a series of thermal reaction peaks due toloss of water and/or structural changes, but the intensity and tempera-tures at which they take place are markedly different for the three sili-cates. Further, those endothermic effects due to loss of water can be wellcorrelated with the shoulders of corresponding dehydration curves, al-though they are found at higher temperatures for the reasons previouslydiscussed. Also, the sharpness of these reaction peaks and their relativehigh temperatures indicate that the water is relatively firmly bound inthe structures of the zeolites.

The DTA data and interpretations of the thermal reactions are givenbelow:

(l) Natrolite

The natrolite curve is characterizedby an extremely sharp and strongendothermic peak at 4550 C. Sveshnikovr and Kuznetsov (1946) reportedthe peak as at 3500 C., but there is no way to determine the cause oftheir lower peak temperature because of lack of detailed information.The reaction starts at about 3000 C. and ends abruptly at 490" C. It isobviously consistent with the sharp break on the dehydration curve of

TtrERMAL ANALYSIS STUDY OF TH.E NATROLITE GROUP 84s

the zeolite and both are due to the release oI the two water molecules.The dehydrated' material showed little changes in appearance and opticalproperties, but there was some lattice expansion (Table 3).

Tler-t 3. Ixmnpr,nNea SplcrNc lon Nernor,rrr nlo hs Monrlrcerrows,

Natrolite

Anhydrous natrolite,at the first

endothermic peaktemperature (455" C.)

d(A)

Metanatrolite,at the second

endothermic peaktemperature (565" C.)

d(A)d(A)

12JA

56789

101 1t213l41 516l 71819202 l2223242526272829303 132J J

34353637383940t 1

1 n

434445

6 . 4 4s . 8 14 . 5 74 . 3 24 . 1 03 . 4 83 . 1 8

2 . 9 32 . 8 32 . 6 62 . 5 72.432 . 4 12 . 3 22 . 2 52 . 1 72 . 0 61 . 9 61 . 8 71 . 8 0I . / J

1 . 7 31 . 7 0 I n1 .68 In1 .65 In1 . 6 2 I n1 .60 In1 .57 In1 . 5 3 I nr .401 . 4 2r . 3 9l . J /

1 . 3 51 . 3 21 . 3 11 . 2 91 . 2 71 . 2 41 . 2 2r . l 9t . r ot . t 4

6 .46s .854 . 6 24 . 3 54 . 1 43 . 4 8J - 1 6

s . 1 43 . 0 92 . 9 42 . 8 5

2 . 5 72 . M2 . 4 12 . 3 22 . 2 5 I n2 . 1 82 . 0 6r . 9 61 .931 . 8 0

1 . 7 21.70 In1 . 6 8I .65r . 6 21 . 6 01 . 5 71 .531 . 4 6I A n

1 .391 . 3 7 I n1 .34 In1 .33 In1 . 3 1

1 . 2 21 . 1 9 I n

1 . 1 4 I n

6 . 3 85 . 7 94.424 . 3 04.09

3 . 1 3

2 . 9 3 D2 . 8 3

2 . 5 6, 4 '

2 . 3 22 . 2 4 D2 . 1 82.05 In1 .96 In1 .881 . 8 0

1 . 6 9

1 .60 In1 . 5 7 I n1 .53 In1 .45 In1 . 4 2r . 3 9

I . 3s1 .32 In1 .30 In

7 .22 rn1 .19 In

1010

J

106I

10

A

1013

2z61124

6

8J

85

. J

75

+10

J

11

. 51

. 5

. 5I

n

. J

. 5

. 5

I

J

. J

1. 5. 5. 5I

8102

104

10

J

10

5a

I. J

z

1

3

5

. J11\I

. 5

. 5

. 5

1. 5

2. 5

XCuKdr:1.5405 A

In-Indistinct; D-Diffuse.

846 C. J. PENG

At 565o C. there is a very small but persistent endothermic peak.

Optical and r-ray diffraction data of the Giant's Causeway specimen

heated up to that temperature indicate the presence of a new phase

which may correspond to Rinne's metanatrolite (1890) and hence, the

endothermic reaction may represent the transformation.

After the 565" C. endothermic effect, the curve is almost level until

between 910o and 940o C. where there is a small rounded endothermic

trough with an immediately preceded small hump. Since the f-ray pow-

der photograph of the materials heated to that temperature showed

only a few diffuse bands, the thermal reaction may be properly identified

with structural destruction.The zeolite shows a smali exothermic reaction peak at about 10100

C. as mentioned previously. Samples heated up to the peak temperature

were essentially isotropic but with scattered weakly birefringent granules.

Nevertheless, they all gave a distinct r-ray powder pattern. Most lines

are those of nepheiine, although some could not be accounted for (Table

4). It is thus clear that the exothermic reaction must be caused by re-

crystallization of fused natrolite into nepheline and probably another

phase, although glass is still dominant at that temperature. The forma-

tion of nepheline and glass from natrolite was reported by C. Doelter

long ago (18S0) and the exothermic reaction was also noted by Sveshni-

kovr and Kuznetsov (1946), but the latter did not identify the products

and the peak temperature given by them is much lower, namely, be-

tween 8950 and 960o C.

(2) Scolecite

The DTA curve of scolecite shows three sharp endothermic peaks. The

first reaction starts at about 1700 C. and produces a broad V-shaped

peak at 310' C. But the large breadth of the peak and its strange change

in slope ron the high-temperature side suggest that it may be a doublet

with a small and poorly defined. subsidiary peak at about 3200 C. (Fig. a).

The major peak may be well correlated with the first shoulder on the de-

hydration curve of the zeolite ending around 200o C', if about 100o is

allowed for the temperature difference between these two kinds of ther-

mal curves. Thus it would represent a loss of one of the three water mole-

cules in the mineral. The possible small subpeak may be due to the trans-

formation of the mineral to metascolecite, which has long been recognized

as a high-temperature form resulting from the removal of one water

molecule from the zeolite (Rinne, 1890; Cavinato,1927; Hey & Bannister,

1936). It seems probable that since these two reactions, partial dehydra-

tion and phase transition, are so closely related, their thermal effects

may overlap and hence cannot be well separated on the DTA qtrve'

THERMAL ANALYSIS STUDY OF THE NATROLITE GROUP 847

Teer.n 4. Pownnn Drr.lnecrrow Dlre rnolt Nnrnor,nr Hrlrno ro1010" C. eno Nnpnu,rNE, ICuKdl:1.5405 A

Natrolite heated to1010" c.

Nepheline,Magnet Cove, Ark.

1z.t4q

6789

101 112I J

l41 5l6l 718t9202 12223242526272829303 132J J

343536J I

3839404 l42+5M4546474849505 15253< t

.).'t

56

drA) d.6)

4.210

3 .833 . 5 9J . Z l

J . l /

3 . 0 12 . 8 72 . 5 82 . 4 82 . 3 82 . 3 +2 . 2 92 . t l

2 . 0 82 . 0 31 . 9 8I .93I . 8 81 . 8 41 . 8 1

r . 7 61 . 7 21 . 6 91 . 6 41 . 6 21 . 6 0

1 . 5 6

\ - + lr . 4 61 . 4 31 . 3 91 . 3 8r ..1.)t . J . )

r . 3 21 . 3 11 . 2 81 . 2 61 . 2 61 . 2 51 . 2 41 2 2l 2 l1 . 1 91 . 1 8t . t 6t . t 4

8 .6944.989

4.2564. r4 t3 .830

3.260

2.9992 . 8 7 62 . s 6 72.486

2 . 3 3 6

2 . 0 7 72.0261 . 9 7 41 .9231 .881

1 . 7 9 11 . 7 8 6r - / . ) Jt . 7 0 4

1 .6631 A)O

t .612I . .)9.)

l . ) J /

1 . 5 2 3t - + o l1 .4551.4261 .3801 . 3 7 01 .335

1 .312

r .278t .265t .263

1 .233

1 .2187.2011 .1821 . 1 5 11 .1371 .028

8

8A

7z

10764

87z

o

A

o4n

6

+6A

262

o657o844L

427452

678z46

77

101010

10

10o64

10

811+I

111A

114182J

1o811

6

711

1

167t

- 1

d(A)

4 . 9 5 84.634

4.1873 . 8 3 93.6443 . 2 6 7

3 . 0 1 92 .8882.5822 . 5 0 22.4052 . 3 5 62.3162 . 1 7 52. t322 . 1 0 5

1 . 9 8 01.931r .8961 . 8 5 01 . 8 1 0t - l l J

r .704

1.643| . 6 2 71 .610

1 . 5 7 3I .5351 . 4 7 8r .4681.4401 . 3 9 31 . 3 8 41 . 3 5 5

r .325

1 . 2 8 8r .2761 . 2 6 6

1 . 2 4 41 . 2 2 7| 2131 . 1 9 5r . lo.t1 . 1 4 8r . l 4 l

II

810I9

1086436

1I

IJ

21z

1

z

1

1

61

I351I

z

3I2

1I

J

111

848 C. J. PENG

The second endothermic efiect starts at about 4400 C. and forms a

doublet with a small auxiliary peak at about 470o C. and a sharp majorpeak at 5000 C. It may be correlated with the final shoulder on the de-hydration curve due to the liberation of the other two water molecules.

Its doublet form also seems to be in accord with the appearance of the

shoulder which has a secondary break and both suggest difiering volatili-

ties of the two water molecules.The third endothermic peak merges with the second doublet at about

525o C. and reaches its maximum at about 560o C. The powder r-ray

difiraction photograph of the Bombay specimen heated up to that tem-perature gave only a few diffuse lines, thus indicating the reaction as a

result of structural disintegration (Table 5).Then the curve levels ofi until the final firing temperature at 10500 C.

used in the present study. Heated samples were a porous porcellaneous

mass, isotropic between crossed nicols, and yielded no diffraction lines.

Hence, melting of the zeolite resulting in glass and anorthite as reported

by Doelter (1890) cannot be proved.

(3) Mesolite

The DTA curve of mesolite features two major and one minor endo-

thermic reactions and two exothermic peaks. As in the case of scolecite,thermal decomposition of this zeolite starts at a relatively low tempera-

ture, probably in the neighborhood of 1500 C. and then proceeds slowly

until it reaches the first endothermic peak. The latter is a doublet with a

minor subpeak at about 310o C. and a major one at 3250 C., but the

minor subpeak is not always developed. The second endothermic reac-

tion takes place sharply at 4000 C. and reaches its peak at about 440oC.

It is a strong symmetrical peak, but the base line shifts to higher level

on the high temperature side, probably due to a change in the thermal

conductivity and specific heat of the material.As demonstrated by the dehydration curve presented above, mesolite

gives ofi its water successively in three stages with 2, 2, and 4 molecules

respectively. If about 100o is assumed as the temperature difference be-

tween the two kinds of thermal data, the first endothermic doublet would

be correlative with the first two shoulders on the dehydration curve and

the second sharp endothermic peak with the last large shoulder. Then,

it would follow that the first and second endothermic reactions may be

attributed to the loss of (2+2) and 4 water molecules respectively.

Samples heated up to the peak temperatures of the two thermal effects

showed an increase in the refractive indices, but little changes in appear-

ance. The heated material just after the first endothermic peak, like

natrolite, had an expanded lattice, but after the second endothermicpeak it showed some contraction (Table 6).

THERMAL ANALYSIS STUDY OF THE NATROLITE GROUP

T.csln 5. Iwrcplaxan Spacrwc ron Scor.rcrtr mm Monrlr.clrroNs ATVemous Po,q' TrupBneruREs, trCuKar:1.5405 A

Firstendothermic

peak(320'C.)

Secondendothermic

peak(s00" c.)

Thridendothermic

peak(550'-570' C.)

I

4

6789

101 1t213l415t6l718t9202 l22231L

2526272829303132JJ

J+

J . )

36J /

383940A 1

4243M4546A J

484950

JI

J

J

I121

10. 2

48161

J

110

. 5I

. J

2

t

1I

I

. 5

. 5

. 5

. 5

. 5

. 2

. 5

O . J J

5 . 8 14.694 . 3 74 . t 93 . 6 33 . 1 53.092 . 8 62 . 6 82 .582 . 4 72 . 4 2' , 1)

2 .262 .202 . r 72 .O72.03r .991 .951 .901 .861 .80 Bt . 7 5r . 7 2

1 .66 B1 .64 Br . 6 11 .601 . 5 2I .50t . 471 .43

1 .38

1 .33

1 . 3 r

r . 2 7r . 2 4r . 2 2t . 2 t1 . 1 9t . r 71 . 1 51 . 1 5

6 . 5 25 . 8 14 . 6 t4 . 3 54 . 1 63 . 6 33 . 1 73 . 0 72 . 8 7

2 . 5 7 82 . M

2 . 2 52 . 1 8 D2 . 1 5 D2 . 0 4

L . 9 6r . 8 7

1 . 8 01 . 7 5 8r . 7 3 B1 . 6 81 . 6 61 .631.62 ln1 . 6 01 . 5 1 I n

1 . 4 8I . + J

1 . 4 01 . 3 9r . 3 71 .331 . 3 1 B1 . 2 9I .28 In

L . Z J

1 . 2 1 L n1 . 1 9 I n1 . 1 7 I nt . 1 4

di.p,)

6 . 4 1 B5 . 7 5 B4.644 . 3 24 . 1 63 . 6 13 . 1 63 . 0 62 . 8 6

2 . 5 6 82 . 4 2 8

2 . 2 72 . 1 9

2 . O 4 8

1 . 9 51 .86

1 . 8 0 Dl . / J

1 . 6 8

r .63

1 . 5 8 D1 . 5 2 D

t . 4 7 D1 . 4 3 D

1 . 3 8 D

1 .331 . 3 1r . 2 71 . 2 4

t . 2 l

1 . 1 8

d(L)

Difiuse

Diffuse

Difiuse

3. 2I

22

1

2

. 5I1

. J

. 2

I

1

1

1

. 5

. 5

. 5

. 5

. 2

J

8(8I1

210

2

. 5A

. 5I

n

1 . 5. 515

J

. J

I. 5

II

1

12

. J

. 5

. 5

. 5

. 5

. 5

d.6) | r.

B-Broad; In-Indistinct ; D-Diffuse.

C. J. PENG

Tllrn 6. IwrnnlleNan SpecrNc lon Mrsor,rrn AND Irs Monrrrcerroxs-erVenrous Tnrnu,tr, Rrectow Prlr Trlponarunrs, trCuKar:1.5405 A

Line

I)3

6789

101 1t 2I J

t4l 5

16L I

1819202 l2223242526272829303132JJ

34J J

3637383940414243M4546i ,

d(A)

6 . M B5 . 7 95.464 . 6 64 . 3 5 B4 . t o3 .893 . 1 8 D3 . 0 8 D2 . 8 62 . 5 7 DN A '

2 . 4 12 . 3 42 . 2 72 . 1 9 D

2 .05

1 .951 .861 . 8 1I . / . ' )

1 . 7 21 .68

1 . 6 41 . 5 91 . 5 4r . 5 21 . 4 71 . 4 31 . 4 01 . 3 91 . 3 5

1 . 2 4 8t . 2 l1 . 1 9

1 . 1 51 . 1 51 . 1 5

Firstendothermic

peaktemperaturer

325" C.

46

46z1

4z

10z

2- l- 1

- 11

- lI2LI

- 1_ I

- l

_ I

1

1

211

I

1

I1

- 1- 1

I- 1

1

- 1- l- 1

d(A)

6. s45 . 8 1

4 . 6 24 . 3 7 D4 . 1 73 . 6 63 . 1 8 D3 . 0 92 . 8 82 . 5 82 . 4 8 82 . M2 . 3 32 . 2 72 . 2 r2 . 0 92 . 0 52 . U )r . 9 61 . 8 61 . 8 1 D1 . 7 6 Dr . 7 3 D1 . 6 9r . 6 71 . 6 41 . 6 01 . 5 5 DI . J J

1 .48t . Mr . 4 11 .391 . 3 7 D1 . 3 5 DI . J J

t . 3 21 .301 . 2 81 a A

1 . 2 21 . 1 91 . 1 8 D1 . 1 61 . 1 5

1

1- l

J

z

101

I1

1

J

1

1

11

1

I- 1- l- 1

31I1

- l

I- l

- 1

- l- l- l

Secondendothermic

peaktemperature,

440" C.

d(A)

2 . 1 8 D2 . 0 6 D

6 . 45 . 8 1

4 . 6 64 . 3 3 D4.O9

3 . O 72 . 8 5 D2 . 5 5 B2 . 4 1 D

1 . 8 1 D

1 . 6 4 D1 . 6 1 D

r . 4 7

1 . 3 8 D

1 . 3 1 D

1 . 1 3 D1 . 1 0

Thirdendothermic

peaktemperature,

490 'C .

d(A)

46

1A

2Difiuse

Difiuse

Weak

Weak

B-Broad ; In-Indistinct ; D-Difiuse.

THERMAL ANALYilS STUDY OF THE NATROLITE GROUP 851

Another endothermic peak occurs at 490o C. and is of small intensity'

Since the Peter's Point specimen heated to that temperatule gave no dis-

tinct diffraction lines, the reaction is interpreted as a structural break-

down. Just prior to this peak there is a very small swell at 470o C. indica-

tive of a weak exothermic reaction, but its explanation is not clear.

diffraction pattern (Table 7).It has been well established that the three zeolites ale independent

species but isostructural and mesolite is not an intermediate mixture

(Gorgey, 1909; Bowman, 1909; Winchell ' 1925; Cavinato, 1926; Hey &

Bannister, 1932-33-36; Taylor et al ',1933; Wyart, 1933; Berman, 1937)'

This is fully confirmed by the present study. Both their dehydration and

DL4 curves are distinctive and the DTA curve of mesolite does not ap-

pear to be a combination of the natrolite and scolecite curves. Additional

evidence is provided by the DTA atwe of an artificial mixture of one

part of natrolite and two parts of scolecite proportional to the composi-

tion of mesolite (Fig.4, No. 17).It produced a typical composite curve

of the two components showing no resemblance whatever to the mesolite

curve. However, although the overall shape of the mesolite curve is

characteristic, its first endothermic peak well matches the first endo-

thermic peak of scolecite and its second endothermic peak falls almost

within the same temperature range as the intense endothermic peak of

natrolite. Since all of the peaks mainly result from dehydration, it ap-

pears that the water molecules in the three zeolites released during cor-

responding endothermic reactions may be structurally of the same type

so that upon heating they behave in the same manner.

Cnvsrer, SrnucrunB aul TnBnlrar, BBnavron

The fundamental cause of thermal decomposition of a mineral is the

disruption of certain bonds within its crystal structure. Since the bonds

which hold atoms to their lattice positions have specific enelgy values,

852 C. J. 7ENG

Tnsrn 7. Poworn Dmr.nacrroN Dam lnou Mnsolrre Hnarno ro 1040. C.

Labradorite(A.S.T.M.)

d.(A)

4 . O 7

J - l I

3.64

3 . 2 03 . 0 02 . 9 22 . 8 42 . 6 42 . 5 32.402 . 2 92 . 2 12 . 1 22 . 0 t

1 . 9 2

I .83

t . 7 7

1 . 7 1r . 6 2I .561 . 5 31 . 4 8r . 3 71 .351 . 3 21 . 2 91 . 2 71 . 2 5t . 2 l1 . 1 61 . 1 3r .o7

Mesolite heatedto 1040" C.

Labradorite,Labrador

d.6) d(A)

1

34

6789

101 1T21374

16r718t9202 l2223242526272829303132JJ

34J.)

36J /

38394041

4

1 . 31 . 1

10t . l1 . 10 . 50 . 52 . 7I510 . 7u . 5

o . 4

0 . 7

0 . 5

0 . 30 . 30 . 10 . 10 . 40 . 10 . 40 . 10 . 10 . 20 . 10 .40 . 30 . 30 . 1

3 . 9 7 5

3 . 6 7 8

3 . 3 7 8

3 . 1 6 3

2 .8882 . 7 8 72.6492.4982.401

2.130

1.999t .912

1.829

1 . 7 6 5

1 .7081 .6051 .5601 .5251.480

1 . 3 5 01.319

r . 2 7 11 . 2 5 31.219

6

1

10

- 1- l

- l

n

1

n

- 1- l

1

- l,,

- l

1

- l

I

6.4884.6894.0383 .885J . / O J

3 .6833 .4883 .3653.2153.0202.9402.8452.6602 .500

2 .300

2.L282.025r.9961.9251 .8901 .8351 .800l . / o 5

1 . 7 5 01 . 7 2 51.6081.545

1.4631 3881 . 3 5 51.328

t . 2 7 2

1.224

1 . 1 3 5

1

I.)I5i

II

10I3

a

J

I

n

1o

t

I

1

1

eNo LasRADonrrr, trCuKa1:1.5405 A

THERMAL ANALYSIS STUDY OF THE NATROLITE GROUP 853

natrolite, probably with the same cavities fiiled. But since they have

different water contents and cations, they must differ in certain details

of their arrangements. These structural differences are believed to be

chiefly responsible for the difiering thermal behavior of the zeolites.The diagnostic feature of the thermal behavior of natrolite is the in-

tense endothermic effect at 4550 C. with its two water molecules removedat the same time. Evidently, the two must occupy equivalent lattice posi-

tions so that they have the same volatility. This is in perfect agreementwith the structure suggested by Taylor et al. (Fig.5a). After dehydrationthe structure is essentially preserved and will not break down until about9400 C.

( G ) ( b ) ( c )

Frc. 5. The arrangement of cations and water molecules in natrolite by W. H. Taylor

et al. (o), in scolecite by L. Pauling (b), and in scolecite by Taylor et al. (c). The large

circles represent oxygen atoms in the SiOr and AlOr tetrahedra. The figures denote heights

in Angstrom units above the plane of projection. (After W. H. Taylor et ol. 1933)

The dehydration of scolecite'proceeds discontinuously with one watermolecule coming ofi during the 3100 C. endothermic reaction and the

other two in the 500o C. endothermic doublet, indicating that the formermust be more loosely bound in the structure than the latter. This tends to

bear out Pauling's arrangement of the water in the zeolite (Fig. 5D) in

which one molecule is placed at 3.6 A from the single cation of Ca whilethe other two are 2.5 A away as in the natrolite structure. It is, however,at variance with the structure proposed by Taylor et al. (Fig. 5c) inwhich one water molecule is a little closer to the cation than the other twoat distances of 2.3 A, and 2.5 A respectively, and hence the single watermolecule would come ofi later than the other two. But Hey (1936) foundit satisfactory to explain his dehydration data of the zeolite. Anotherpoint brought up by the present study is that both rhe DTA and staticcurves suggest that even the other two water molecules may not occupyexactly the same lattice positions as they do in the structure suggested

I

854 C. J. PENG

either by Pauling or by Taylor et al., since they are removed in a doubletreaction. Further structural study thus seems desirable.

While the natrolite structure persists up to about 9400 C. as alreadydiscussed, scolecite begins to show signs of structural collapse after finaldehydration at 5000 C. and breaks down completely at 5600 C. Their dif-ferent structural stability cannot be explained by the nature of the cat-ions, since Ca and Na ions are almost of the same size and the strengthof the Ca-HzO bond (2/7) is even stronger than that of the Na-HzO bond(1/6) according to the structures of the two zeolites worked out byTaylor et al.It seems certain that the much lower stability of scolecitemay result from the introduction of the extra water molecule into thenatrolite structure, which, according to Pauling's suggestion (Fig. 5b),takes up one of the two Na positions in natrolite. Since it must be muchmore loosely bound to other ions than the corresponding Na ion in natro-Iite, the entire structure would be weakened accordingly. Furthermore,since it is more readily driven out upon heating than the other two watermolecules, its removal would further damage the structure. But the othertwo water molecules are given ofi at slightly higher temperatures (about500' C.) than the two in natrolite (at 455o C.) although they probablyoccupy similar lattice positions. This is explained as being due to thestrength difierence between the Ca-HqO and Na-HzO bonds.

No detailed r-ray analysis has yet been made of mesolite. Since itsr-ray diffraction pattern is almost identical to that of both natrolite andscolecite and it has a composition corresponding to one natrolite and twoscolecite molecules, Taylor et al. (1933) have suggested that the mesolitestructure must be similar in all its essentials to that of the other two zeo-lites except that its cell size is three times as large. Further, 6 of the 8water molecules occupy the general positions as in natrolite and the othertwo are more strongly tied to the Ca ions as in scolecite. Such an arrange-ment of the water molecules fails, however, to explain the mode of de-hydration of the zeolites as revealed by their thermal curves, which showthat 2 water molecules each are removed earlier in two closely spacedstages and the other 4 in a single higher-temperature reaction. It is thusindicated that the water in mesolite may comprise two volatile groups oftwo molecules each and one less volatile group of four. Further, as hasbeen noted above, the two volatile groups come off in about the sametemperature range as the first water molecule of scolecite and the lessvolatile one at about the peak temperature of the intense endothermicefiect of natrolite. Then it may be reasonably assumed that only 4 of the8 water molecules may occupy the general positions as in natrolite and theother 4 may be placed in similar positions as the extra water molecule inscolecite. fn the absence of detailed structural data at present, however, itwould be merely speculative to assign any specific positions to the water.

TEERMAL ANALYSIS STUDY OF THE NATROLITE GROUP 855

After final dehydration at 4400 C. mesolite has already shown consider-able optical and structural changes and breaks down completely at 4900C. The still lower stability of the zeolite may be similarily explained asfor scolecite as being due to addition of extra water molecules to thenatrolite structure. Thus, it appears that the higher the water contentof a zeolite, the smaller is its stability range, which seems to be in linewith the view of J. W. Gruner (1950) as to the reaction energies ofsilicates.

CoNcr,usrorqs

The distinctive shoulders on the dehydration curves and their corre-sponding endothermic peaks on the DTA curves of the three zeolitesclearly indicate that the water must hold definite Iattice positions. Al-though the three are isostructural, they give ofi their water at difierenttemperatures and in difierent amounts, The conclusion is irresistiblethat the water occupies different positions in their structures. On thebasis of their thermal data, it is suggested that natrolite contains a singlegroup of two water molecules of equal volatility, scolecite has a volatilegroup of one molecule and another less volatile group of two, and mesolitecomprises two groups of two molecules each oI high and slightly differentvolatilities and a group of four molecules of equal and lower volatility.This confirms the validity of the methods to arrange water molecules innatrolite and scolecite proposed by Taylor et al. and by Pauling. But it isassumed by Taylor et al. that the water in mesolite consists of a groupof six molecuies and another group of two.

The physical properties of the three zeolites show marked changesduring the course of dehydration and upon heating up to about 10500 C.They do not, however, undergo the same kind of changes at the sametemperature. While natrolite shows only slight changes in the opticalproperties and crystal structure after complete dehydration at 455o C.and retains its structure until about 940" C, the structures of scoleciteand mesolite are completely destroyed respectively at about 5600 C.and 490o C. immediately after their final dehydration. Their stabilityranges thus seem to depend upon the degree of hydration-apparentlythe higher the water content of the mineral, the lower is its stability.

The thermal curves of the three zeolites are so characteristic that theyshould prove to be a useful means of identification. Further, they mayserve as an additional proof that the three do not constitute an isomor-phous series but are independent species.

RprnnrNcns

l. Am. Soc. Testing Materiols U Nat'l Res. Counci) (f950), X-ray Diffraction Data Cards(in three sets).

8s6 C. J. PENG

2. Asrcnolr, F. N. (1916), The natrolite occurring near Kinbane (White llead), CountyAntrim: Minzral. Mag., 17,307.

3. Bnnuex, H. (1937), Constitution and classification of the natural silicates: ,42r.Mineral.,22, 374.

4. BowueN, H. L. (1909), On the identity of poonahlite with mesolite: Minaal. Mag.,t5,216.

5. Cevrwero, A, (1926), Sulla mesolite: Atti Coc. Itai. Sci. Nat.,65, L04-114.6. - (1927), Nuove osservazioni sulle zeoliti del gruppo della natrolite: Mem. Rend..

Accad. Lincei. cl.. sci. fis. mat. nat. Roma, ser. 6,2, 320-350.7. Dorrren, C (1890), Uber die ki.instliche Darstellung und die chemiche Constitution

einiger Zeolitle: Neues Jahrb. Mimeratr., l, 13+-137.8. Gowrvnr-o, F. (1891), Sur le groupe mdsotype dans le Puy-de-Dome: BuJl. Soc. Franc.

Minaal., 14, 170.9. Goncev, R. (1909), iiber Mesolith: lbid.,28,77-106.

10. Gnurtrn, J. W. (1950), An attempt to arrange silicates in the order of reaction energiesat relatively low temperatures: Am. Mi.naal.,35, 143.

11. Heoor.o, M, F. (1883), On a new mineral locality: Mi,neraL. Mag.,5,718.12. Hnv, M. H., ,l.r.'n Bervxrsrnn, F. A. (1932), Studies on the zeolites, Pt. 3, Natrolite

and metanatroLte : Ibiil., 23, 243-289.13. - (1933), Studies on the zeolites, Pt. 5, Mesolite: Ibid,.,23r 42l-47.14. - (1935), Studies on the zeolites, Pt. 8, A theory of the vapour pressure of the

zeolites and of the diffusion of water or gases in a zeolite crystal: Ibid.,24r 110-130.15. - (1936), Studies on the zeolites, Pt. 9, Scolecite and metascolecite: ILid.,24,

227-253.16. Krn:n, P. F., aNn Kwr, J. L. (1948), Multiple difierential thermal analysis: Am.

Mineral.,33, 387-419.17. Mrr-r.rclN, W. O., arlo Wrrsnn, H. B. (1937), The mechanism of the dehydration of

zeolites: Jow. Phys. Chem.,4l, 1029-104o.18. Peurwc, L. (1930), Tbe structure of some sodium and calcium aluminosilicates: Proc.

Nat. Acad. Sci,., U.S.A., 16, 453-459.19. Pnr.ecaNr, L. (1908), Studio chemico delle zeoliti de Montresta (Sordegna): Atti,.

Accad. Limei. Renil. cl. fis. mat. nat., ser. 5, 17, sem. 2, 66{8.20. Rnvrvo, F. (1890), ijber Umiinderungen, welche die Zeolithe durch Erwiirmen bei und

nach dem Triibewerden erfahren: Sitzungsber. d. k. preuss, Akadem. d.. Wissensch. z.Berli,n,46, 1163.

21. SvnspxIxow, V. N., em Kuzrvnrsov, V. G. (1946), Structural relations betweenzeolites and natural kaolin and their transformations on heating: Izaest. Akail. Nauk,U.S.,S.R. Otilel,, Kbim. Nauk, 1946,25-36; Chen. Abst.,42,6200 (1948).

22. Tl^ton, W. H., Mrnr, C. A., eno Jacrsow, W. W. (1933), The structures of fibrouszeolites: Zei.ts. Krist., 84, 37 3-398.

23. Tscrnnuar, G, (1917-1918), Der chemische Bestand und das Verhalten der Zeolithe,I. Teil.: Si,tzungsber. Akad. Wiss. Wien, Math.-nah.crwi.ss. Kl., Abt. I, 126, 541406;II Teil.: Ibid.,127, 177-289.

24. Wx,xrn, T. L.,.lNo Pensows, A.L. (1922), The zeolites of Nova Scotia: Uni.u. TorontoStudies, Geol. Ser.,14, l+-22.

25. Wnrcrer.r,, A. N. (1925), A new theory of the composition of the zeolites: Am. Mi.neral.,lO,712-116; l7l-174.

26. Wvenr, J. (1933), Recherches sur les zeolites: Bull. Soc. Franc. Mineral., 56, 81-187.27. ZeusoNrrvr, F. (1908), Contributo allo studio dei silcati i&atui Atti. il. Reale Accail. il.

Sci. fis. e. mot. il. Napoli, 16, l.

Manuscript receiaed. fuJy 26, 1954.