American Mineralogist, Volume 67, pages 1229-1241, 1982

The thermal behavior of scapolites

Groncto GnezreNr eNo Sr,ncIo LuccsBsI

Istituto di Mineralogia e Petrografia, Universitd degli StudiPiazzale Aldo Moro 5,00186 Rome, Italy

Abstract

Eight scapolite samples from different geologic settings were selected so as to fullyrepresent the marialite-meionite series.

Thermal analyses (TG and DTA), as well as mass spectrometer analysis with a heatingdevice, were used to define the release of the volatile components during heating. In thetemperature range up to 800'C, the emission of HzO, SOz, COz, NaCl, and KCI wasobserved to occur continuously, but with five evident steps. The presence of SO2 in thegaseous phase may be due to decomposition reactrons.

X-ray powder diffractograms reccirded every 100'C during heating up to 800'Cindicated an increase in a and V, which occurred linearly in the whole marialite-meioniteseries, with a trend inversely proportional to the percentage of meionite present in eachsample.

The relationships between increases in lattice parameters, with rising temperature, andpercentage of meionite may be expressed by the equations:

The trends in a and V are reversible since samples treated to 800' C revert to initial cellparameters upon cooling. Temperature increases did not clearly affect the value of c, whichremained constant. Beyond 800' C the unit cell parameter a increased anomalously, in anirreversible way, while the breakdown of the mineral started producing various com-pounds: plagioclase, which invariably has a lower calcium content than the scapolite fromwhich it derives, halite and calcium aluminum silicates.

Refractive index measurements carried out during heating and on the cooled sample,confirmed such behavior, at least to a first approximation, showing for both e and c,r a linearincrease during heating and a reversal of this trend to the initial values upon cooling.

daldT = -1.59 x 10-6 Me% + 1.85 x l0-4dv ldT : -0 .3x l \ - 3MeVo +0 .034

(r2 : 0.996)(r, : 0.999)

ed as SO?-, as demonstrated by X-ray spectrosco-py (Chappel and White, 1968) and electron micro-probe determinations (Lovering and Widdowson,1968). The sulphur content can reach the highestpercentage among the volatile components (Knor-ring and Kennedy, 1958), being favored by highpressure (Lovering and white, 1969).

Infrared absorption spectroscopic studies havemade it possible to determine the relation betweenthe percentage content of CO2 and SOr (Schwarczand Speelman, 1965):

SOr (wt.%) : 0.704 COz (wt.Vo) - 0.800

The scapolite framework is characterized by twotypes of four-membered silicate tetrahedral rings,both lying parallel to the (001) plane (Papike, 1964;

t229

Introduction

The composition of natural scapolite is related tosolid solution between two theoretical end-mem-bers: marialite, Naa(Al3SieOz+)Cl and meionite,Cae(AleSioOz)COz (Strunz, 1978, p. 484). Suchminerals show a rather complicated substitutionscheme with three coupled replacements in whichNa*, Si4* and Cl- are replaced by Ca2*, Al3* andCO|-, respectively (Papike, 1964). This anionicsubstitution occurs only for percentages of meionitebelow 75, since at higher values the anion site isfilled with CO3- @apike,1964; Evans et al., 1969).Scapolites are characterized by a substantialamount of volatiles, which, besides Cl-, CO!- andF-, include H2O, whose structural role is not welldefined (Evans et al.,1969), and sulphur, coordinat-

0003-004)v82l | | l2-1229s02.w

GRAZIANI AND LUCCHESI: THERMAL BEHAVIOR OF SCAPOLITES

Papike and Zoltai, 1965; Papike and Stephenson,1966; Lin and Burley, 1973a,1973b, and 1975). Asthe meionite content increases, rings of the firsttype rotate in a counter-clockwise direction, andthose of the second in a clockwise direction. Thearray of rings delimits cavities of smaller and largersize, the former occupied by cations and the latterby anions. An increase in meionite content alsoaffects the values of the lattice parameters, resultingin a linear increase in a and a change in c (Ulbrich,1973).

The correlation between crystallographic dataand meionite content was also investigated by Bur-ley et al. (1961), who recognizeda sharp decrease inthe angular distance (A) between the reflections ofthe (400) and (112) planes as scapolite calciumcontent increased.

Levien and Papike (1976) investigated the prob-lem of the effects of temperature increase on a 33.5percent meionite sample. They identified a rotationof the four-membered rings of tetrahedra in the(001) plane, during heating, in an opposite directionto that due to increasing meionite content. Suchbehavior resulted in a "more open structure",leading to an increase in a and V values while cremained constant.

The correlation between chemical compositionand physical properties was also investigated byShaw (1960), who found linear relationships be-tween meionite content, volatile components, den-sity and the mean refractive indices.

The presence of scapolite as a primary constitu-ent of granulite inclusions found within eclogite andultrabasic rocks (Lovering and White, 1964), and itscharacteristic high temperature and pressure stabil-ity (Newton and Goldsmith, 1976), suggested thatthis mineral could be a storage site for volatiles,especially CO2, in the lower crust (Goldsmith,1976). The petrologic and mineralogic interest inscapolites suggested that it would be useful toinvestigate the thermal behavior of the marialite-meionite series in more detail defining the effects oftemperature increases and the order in which thevolatile components are released.

Material and methods

Various scapolite samples from different geologicsettings and representing most of the marialite-meionite series were selected from numerous exam-ined specimens of crystals free from macroscopicinclusions. With the intent of preliminary checkingof the meionite content of each idiomorphous sam-

ple, the mean refractive indices were evaluatedusing oriented laminae parallel to the {100} prism(Shaw, 1960). The immersion method was adopted.A water cell connected with the central plate of themicroscope controlled the temperature of the circu-lating liquid, within a range of about 50" C. AnAbb6 refractometer, connected in parallel to thewater system, allowed the refractive index of themounting liquid (Cargille Laboratory) to be deter-mined at any time, for the sodium doublet. Themeionite content of the massive samples was evalu-ated by means of X-ray powder diffraction, adopt-ing the angular shift (A) method, as suggested byBurley et al. (1961).

Samples I and 2, from the Umba deposit, NETanzania, consist of idiomorphous, gem-qualitycrystals of considerable size (cc. 5 mm x l0 mm x20 mm). They are both without inclusions, andrespectively violet and yellow in color. Their settingis characterized by pegmatitic intrusions related todunites and other ultrabasic rocks, as well as bymetamorphic rocks of various type, mainly pre-Cambrian gneisses and marbles (Solesbury, 1967).Scapolite crystals have been found both in contactzones and in secondary deposits (Zwaan, l97l).

Samples 3 and 5 are scapolite crystals originallymined at Ankazob6 and Gabenja respectively, inMadagascar. The crystal from the first locality,which belonged to the Eppler Collection, Munich,West Germany, is idiomorphous, large in size (ca.30 mm x 30 mm x 60 mm), and of a lighter greycolor; it is crossed by numerous growth channelsand two-phase fluid inclusions, most of them anhe-dral. The samples from Gabenja, is an idiomor-phous, yellowish crystal (ca.2 mm x 2.5 mm x 10mm), without inclusions, kindly supplied by theNational Museum of Natural History, SmithsonianInstitution, sample 141364. In the former localityscapolite is usually found in heterogeneous beryl-bearing pegmatite enriched in euxenite and morta-zite (Lacroix, 1919).

Two massive Canadian scapolites, from Gooder-ham, near Haliburton, Ontario, sample 4, and Gren-ville, Quebec, sample 7, were kindly made availableby the National Museum of Natural History, Smith-sonian Institution, samples 126018 and R 13120.The former was found among altered gabbros,marbles, skarns, and nepheline and corundum syen-ites (Adams and Barlow, 1911). The setting of theGrenville sample, on the other hand, comprises avariety of pre-Cambrian rocks including metagab-bros, amphibolites, gneisses and marbles in which

GRAZIANI AND LUCCHESI: THERMAL BEHAVIOR OF SCAPOLITES t23l

scapolite presumably crystallized through the intro-duction of chlorine from granite magma (Budding-ton. 1939).

Sample 6 came from Manchester, New Hamp-shire, U.S.A.; it is a transparent crystalline frag-ment (ca. 5 mm x 7 mm x 8 mm) with an idiomor-phous habit. Such scapolite occurs in irregularcolumnar masses, in a biotite schist with narrowcalcareous bands interbedded with granite and peg-matite lenses (Stewart, l94l).

Sample 8, which was kindly made available bythe Museo di Mineralogia, Istituto di Mineralogia ePetrografia of the University of Rome, Italy, sample7536120, was found at Pianura, in the MountSomma-Vesuvius area, Italy. The idiomorphous,colorless, transparent crystal of fairly small size(ca. 4 mm x 5 mm x l0 mm) includes greenpyroxenes; it was found in a geode from volcanicejecta mainly composed of limestone associatedwith pyroxene and leucite (Scherillo, 1935).

Electron probe microanalyses were then per-formed on a JEoL rsu-5Oe using natural and synthet-ic standards. Matrix corrections were made accord-ing to the EMIADR VII program described by Ruck-lidge and Gasparrini (1969). All iron was calculatedas FeO. The COz content of the total gaseous phaseemitted by the substance after heating at 1000'Cwas determined by means of the high sensitivityanalyzer of Scarano and Calcagno (1975), using anair current with constant COz content as carrier.The H2O content was estimated by assessing thedifference between the thermogravimetric resultsand the chemical data for the other volatile compo-nents. The density of all the samples was measuredwith a Berman balance.

Differential thermal analyses (DTA) were carriedout on a BDL micro-differential thermal analyzerequipped with a semi-microprobe (capacity:25 p.l).The thermocouples were made with Platinel II andthe experiments were performed in flowing air (5ml/min), with a thermal gradient of 10' C/min in the25-1400" C temperature range. Under the sameexperimental conditions used for DTA, thermalgravimetric analyses were carried out with a TGS-2Perkin Elmer thermogravimetric system usingabout l0 mg of material for each analysis.

Serious obstacles were faced in performing thethermal analyses, since striking variations in thequantitative ratios between each weight loss be-came apparent when the experiments were carriedout under differing experimental conditions. Evenwhen the chemical homogeneity of a sample had

been checked by electron microprobe analyses, itwas found that the thermal studies may be stronglyinfluenced by experimental conditions. A current ofair gave better resolution than nitrogen or even astatic atmosphere. The way the samples were pre-pared turned out to be very important, especiallyregarding grinding and loading techniques. Thus,such factors may cause great difficulties in obtain-ing and assessing data on the amounts of H2O, CO2and SO3 present in a sample.

Mass spectrometric experiments were performedin an attempt to identify the gaseous componentsreleased in the same temperature range, 25-1400' C, under high vacuum conditions. The equip-ment used was a Nuclide magnetic sector massspectrometer, equipped with a Knudsen cellsource: an alumina-lined tantalum crucible wasplaced in the cell and the temperatures were mea-sured with a chromel-alumel thermocouple insertedon the bottom ofthe cell. The background pressurewas maintained at about 10-6-10-8 torr throughoutthe evaporation experiments. Full identification ofcarbon dioxide and water vapor, among the vola-tiles released during heating, was hindered by thefact that they are non-condensable gases, in theexperimental conditions chosen, and by the un-avoidable presence of small amounts of them in theapparatus.

The samples were also analyzed in a dry nitrogencurrent of 5 ml/min in the temperature range of 25-1000'C every 100'-15" C, with an X-ray powderdiffractometer equipped with an AF 3000 heatingcamera. Ni-filtered CuKo radiation, scanning froml0o to 70" at ll2'2O per minute, was used to detectany structural variations within the sample beingheated. Scapolite was kept at the chosen tempera-ture for 30 minutes before each diffractogran wasrecorded, to ensure the stabilization ofevery speci-men at the experimental temperature, since thisisothermal stage proved to be the most suitable forthis purpose. After heating, the same experimentswere performed on each sample at room tempera-ture, using the same procedures.

Unit cell parameters were estimated from X-raypowder diffraction data indexed by Gibbs and Bloss(1961) using a least-squares refinement (Farinatoand Loreto, 1975) with semiconductor-grade siliconmetal (Jarrel Ash J. M., spectroscopy impurity lessthan 300 ppm) as an internal standard.

To investigate any changes in the refractive indi-ces occurring during heating, the minimum devi-ation method was adopted; a polarizing Stoe goni-

1232

meter was used, and measurements were carriedout with a sodium light at 589.0 nm. The heatingdevice consisted of a single crystal heater similar tothat described by Brown et al. (1973). Two prismswere prepared for each sample. They were orientedparallel to the optic axis of the sample, and theirorientation was checked by means of a universalstage. The maximum angle between the reflecting

GRAZIANI AND LUCCHESI: THERMAL BEHAVIOR OF SCAPOLITES

faces was 68'and the minimum was 60o. This valuewas calculated by assuming the hypothetical varia-tion in the refractive indices that might have oc-curred during heating, and its effect on the variationof the minimum deviation angle (Franziani, 1965).

The measurements taken during heating wereperformed with a heating rate of 10'C/min, every50oi5o C, in the temperature range of 25-450" C,

Table l. Electron microprobe analyses and physical data of scapolites

sio2 58.67

T io2 o .o l

Arzo3 20.74

F.O a O.O9

MnO O.Ol

H8O O,1 l

CaO 4 .11

Na2O 10.59

K.O t .22ot^t lA-2-_

I o,s4Bzo Ico2! l .o8sol o.o3F t r .

c r 3 . 0 1

55.77

22.15

0 . 1 7

o.09

6 . 9 3

9 . 3 1

o . 7 9

0 . 3 5

0 . 9 1

o.o5

2 . 9 6

52.78

o . 0 2

24.L6

o . 4 8

o.01

8 . 9 7

7 . 8 5

1 . 0 5

o , 4 2

I . 5 3

o_ro

2 . 4 2

52.01

24, t3

o . 3 0

o , 2 6

1 o . 1 1

o.86

o . 5 8

1 . 6 1

o . 8 l

2 . O 7

50.98

o.01

o . 2 5

o,04

L2.94

0 . 8 5

o . 3 8

2 . a f

o . 1 1

4 9 . l O

o . 2 4

0 . 0 1

L4.20

4 . 9 0

0 . 8 9

o . 2 l

L . 2 9

0 . o 4

4 7 . 2 2

o.04

25,76

0.oE

o.09

1 5 . 1 5

4 . O 5

o . 5 2

I . 5 3

2 . 7 9

L . 2 6

o . 4 5

4L.28

o . 2 a

20,77

1 . 4 0

o . 4 5

o , 2 9

4 . 3 0

o . 3 4

0 . 2 0

loo .27

o . 6 8

roo.090,66

100.27 99.99

o . 4 7

100. 17

o , 3 1

too .10

o.30

99.95

0 . t o

too.90

o.05o = c l , r

Tot. l 9 9 . 5 9 99.42 99.72 99.52 9 9 , E 5 9 9 . 8 0

Nubers of ionc on the ba! i ! of 12 (si , Al)

5 1

T i

F e

Mn

Mg

C .

Nr

K

H

F

u c l

r.ort I ,r.oo3.52e I

o ,520

8'roa I rz.oo

3 . 8 9 6 '

:.;i1 ,,,o.r r , I

f.2.oo 7 '758

4.242

0.037

3 . 9 9 0 . 0 5 8

1 . 6 1 6

2.O97

o , 1 6 4

o . 5 7 7

o.rza Io.e6 o:ott

Io.rr, I

12.oo 7 '59o

4 . 4 1 0

o.oo l

o .031

3,97 0 .OO9

2,064

I . ) I J

o . 1 5 3

o , 3 7 7

0 . 5 8 3

0 . 9 4 o . o 1 2

o . 3 4 3

1 2 . o o 7 ' 4 7 L

4.529

o.03 l .

3 .78 0 .OO2

2 . 3 1 5

1 , 4 4 6

o . 2 1 3

o . 5 5 3

o.g4 o 'L47

0 . o 1 9

o . ! 2 5

59.2

12.oo 7 '3o4

4.696

o.oo5

o.o l0

3 . 9 7 0 . O 2 0

2 . 6 7 9

L . 2 L 4

0 . 1 0 2

L . t l 6

o . 5 8 9

0 , 1 4 5

o , 1 1 8

6 7 . 3

1 2 . o o e ' I s I r z . o o5, G85 J

o . @ 5

4.03 0 .064

o.295

o.898

0 .85 o ' o39 l o . ss

7 . 7 9 5

4.205

o.oo2

o,059

o.oo l

o .059

1 . 4 1 9

2,244

o.200

o . 4 1 4

o.rzs Io.o22 \

". raa I3 8 . 6

o . 4 1 5

o.o90

0 . 2 1 3

0.0o3

o . 7 3 6

o . 9 5

t 7 . 6

o.o52

8 7 , 3

t r .538ro(8)c l 1 . 549 r (1 )p (g/co3) 2.586(4)g ( f ) 1?,or4(3)c (A) 7 .597 (3)

l .54oo1(8) r.s4762(6)

1 . 5 5 8 6 7 ( 6 ) 1 . 5 5 1 8 ( 1 )

2 .594(4) 2 .607(s )

r2 .o45(4) 12 .083(3)

7 . s 8 7 ( 3 ) 7 . 5 8 3 ( 3 )

1 , 5 4 9 3 3 ( 9 ) r . 5 5 s 4 3 ( 7 )

1 , 5 7 4 3 ( 1 ) 1 . 5 7 6 8 7 ( 6 )

2 .688(J ) 2 .583(5)

L2 ,L34(4) 12 . r47(3)

7 . 5 7 L ( 4 ) 7 . 5 5 6 ( 3 )

- r .555O2 (5)

- r . 5 9 2 4 ( r )

2 , 1 0 6 ( 3 ) 2 . 7 4 L ( 4 )

L 2 . L 4 7 ( 3 ) 1 2 . r 8 5 ( 5 )

7 .562(3) 7 .s78<4)

2.62r(4)

r2 .087 (5 )

7 ,577 (3 \

a Total i ron rr EeO.

I oetemined ui th a therugraviretr ic enalyccr.

Meslured ui th a CO2 amlyrer.

GRAZIANI AND LUCCHESI: THERMAL BEHAVIOR OF SCAPOLITES t233

while those taken at room temperature after heatingwere carried out on samples heated to a maximumof 800' C.

Results

Chemical composition

The presence of minute inclusions in scapolitenecessitated the use of an electron microprobe forchemical analyses (Ingamells and Gittings, 1967).Each analysis shown in Table I gives the mean ofthe data collected from at least five different posi-tions.

The percentage of meionite in each sample wasevaluated by the relationship between the compo-nents: MeVo : [(Ca + Mg + Fe * Mn * Ti)/(Na +K * Ca + Mg * Fe * Mn + Ti)l x 100, suggestedby Shaw (1960). These data were in good agreementwith those obtained during the preliminary exami-nations, using optical and structural methods(Shaw, 1960; Burley et a|.,1961; Ulbrich,1973).

Thermal study

The thermogravimetric curves (TG) for the eightsamples show similar trends, with steady lossesintemrpted by five steps of more conspicuousweight loss. The continuity of the steady weightloss makes it hard to exactly identify the initial andfinal emission temperatures of each of the five steps(Fig. 1). The first of these stages is not well defined.It begins at about 80" C and ends at about 130" C.The second starts at 230-240" C and continues to350" C, varying considerably in intensity from sam-ple to sample and showing no clear correlation withmeionite content. The next weight-loss step is simi-lar beginning at about 420" C and ending at about520 'C.

In all the eight samples the fourth and fifthweight-loss steps appear to be the most significant.The fourth starts at about 660" C and ends at about800'C, with weight loss increasing with the per-centage meionite in the sample. Conversely, theweight-loss corresponding to the fifth step is in-versely correlated with meionite content; this stepbegins at 940" C and ends between 1250" and1300'C. This behavior is made conspicuous by thederivative thermogravimetric curve (DTG), which,in all samples, has an almost identical pattern.

The differential thermal curve (DTA) shows threeclear endothermic peaks, with peak temperatures ofabout 100o, 710" and 1030'C, corresponding to thefirst, fourth and fifth weight losses. A wide-ranging

fat ---:r - r , / #t

\'/ zoo Pvl

T " C

Fig. l. Thermogravimetric curves of scapolites showing acommon trend, with steady overall weight loss interrupted byfive steps of sudden weight loss. Such behavior is confirmed bythe DTG and DTA experiments; a representative curve is shownfor each.

broad exothermic effect is recognizable between150' and 650'C. The peak temperature of thesecond endothermic effect corresponds to that re-ported by Kauffman and Dilling (1950), althoughthey did not note the other two.

To understand the order in which the volatilecomponents were released the samples were ana-lyzed by a mass spectrometer designed for vapor-ization studies on solids at high temperatures, andthe gaseous phase evolved was checked for: H2O,HF, Na, HCl, F2, K, COz, CaO, NaCl, KCl, Cl2,SOz, SOr, NazO, K2O, Na2Cl2, &ndK2Cl2.

First, the whole spectrum was scanned; then thesingle compounds were examined repeatedly atconstant temperature, so as to define any variationsin their partial pressure and clarify the releasemodalities (Fie.2).

These experiments provided evidence that theH2O loss, which starts at 70-100' C, because of thehigh vacuum operating conditions, reaches a maxi-mum between 150o and 250'C, gradually exhaust-ing and stopping at about 400" C. The emission ofCO2 begins at about 100'C, increases slowly,

1234

reaches its maximum partial pressure value at about600'C, then decreases gradually and ends at850'C. The release of sulphur as SO2 begins at atemperature between 300" and 350o C, reaches amaximum at about 450'C end then gradually de-creases, persisting at trace levels until almost1200" c.

The emission of NaCl begins at a temperaturebetween 750" and 850'C, becomes quite conspicu-ous between 900' and 1000" C, then diminishesfairly rapidly, and stops at about 1250'C. Thepresence of KCI was recognized in the same tem-perature range. At the same time, small amounts ofNa, Cl and K were detected; their presence wasattributed to the dissociation of such halogen com-pounds.

In addition to these volatiles, particularly in themass spectra recorded beyond 600-800'C, smallamounts of other compounds have been identified.Their intensities were roughly 0.1-0.01 percent ofCOz and NaCl, the most abundant species in suchexperimental conditions. Also intermediate reac-tion products and perhaps hydrocarbon moleculeswere noted.

In the gaseous phase, no sulphur, SO3 or HCI wasnoted, even though the latter has been reported inscapolites (Donnay et al., 1978).

X-ray dffiaction

X-ray powder diffraction patterns were run foreach sample, at different temperatures and at room

H z o

Coz

soz

N a C l

K C I

0 200 600 1000 1400T o c

Fig. 2. The release range of volatile components duringheating. Continuous weight loss is shown by the length of eachsilhouette, while the main emission temperature is marked by themaximum.

GRAZIANI AND LUCCHESI: THERMAL BEHAVIOR OF SCAPOLITES

temperature after heating, to assess the relation-ships between temperature, percentage meionitecontent and any changes in unit cell parameters.

It was noted that each lattice parameter shows apeculiar trend during temperature increase, inde-pendent of the meionite content of the varioussamples (Table 2). During heating the a valuesshowed a steady, almost linear increase that wasinversely proportional to the percentage meionitecontent of the sample (Fig. 3).

The slope values of these straight lines were thenplotted vs. each chemical composition, by means ofa linear regression, yielding the following relation(f : 0.996):

daldT: -1.59 x l0-6MeVo + 1.85 x 10-4

where daldT represents the slope of the straight linerelative to the increase in a during heating andMeVois the meionite percentage of the heated sample(Fig. a). The value of c remained almost constant asnoted by Levien and Papike (1976).

The unit cell volume increases linearly in inverseproportion to the meionite content of the scapolitesample (Fig. a). The correlation between its in-crease and the chemical composition may be ex-pressed by the relation Q'2 : 0.998):

dvldT: -0.3 x l \ -3MeVo + 0.034

where dVldT represents the slope of the straight linerelative to the increase in the volume of the unit cellversus temperature and MeVo is the meionite con-tent.

When the experiments were carried out at roomtemperature, after heating the a, c, and V valueswere identical to the initial values, within the limitsof experimental error, independent of the tempera-ture reached by the sample. At the same time, thevalues of the angular shift (A) between the (400) and(112) reflections, whose variations were correlatedwith meionite content by Burley et al. (1961), weremeasured for both experimental cycles. A decreaseduring heating and an unchanged value after eachcycle, were evidenced.

In investigating the relationships between the avalue of each sample, measured at preselectedtemperatures and its chemical composition, it wasfound that the differences between the values of thea parameter recorded for different samples de-creased steadily with rising temperatures, up to700" C. Thus the difference between the a values ofthe two end-members, marialite and meionite, fellby 44 percent, from 0.171 to 0.095, when the

temperature was raised from 100" to 700'C. Athigher temperatures this uniformity broke down,because of incipient breakdown phenomena, dis-played by the release of NaCl and KCl, especially inthe samples with the lowest percentage of meionite.

X-ray powder diffractograms recorded at tem-peratures over 700o C, at 100'C intervals, revealedthe presence of various compounds, among whichplagioclase and halite have been recognized. Theformer displayed a sodium content higher than thatof the scapolite from which it was derived (Fie. 5).The albite content of this plagioclase has beenevaluated from the X-ray powder patterns using the

t235

A(131), L(241) methods proposed by Bambauer eral. (1967).

Optical results

The refractive indices of the six transparent sam-ples were first measured at high temperatures, andthen at room temperature, after cooling. The formermeasurements continued up to a maximum of450o C, since beyond this temperature the radiatingenergy of the heated scapolite prisms made it im-possible to clearly distinguish the emission doubletof sodium. The measurements at room temperaturewere performed after the sample had been heated to

GRAZIANI AND LUCCHESI: THERMAL BEHAVIOR OF SCAPOLITES

Table 2. Crystal data for scapolites at different temperatures

srDple Unit cel l

nsber p6r6@ter! (A) zs' too' 2oo' 3oo'r ( 'c)

4oo' 5oo' 600' 7oo' 8oo'

. a; c

vv

atc

vv

tzccvv

ttc

vv

aa

cvv

I

t

ccvv

1 2 . 0 1 4 ( 3 ) r 1 2 . 0 1 9 ( 5 )1 2 . 0 r 4 ( 3 ) r 2 . 0 1 8 ( 3 )7 .597(3 ' 7 ,s88(3)7 ,597(3) 7 .598(2 '

1096.7(8) 1096(1)r o 9 5 . 7 ( 8 ) 1 0 9 7 . 4 ( 7 )

12 .o45(4) 12 .063(5)r2 .045(4) r2 .04r (3 )7 . 5 8 7 ( 3 ) 7 . 5 8 2 ( 3 )7 ,s81 (3 ) 7 .584 (3 )l loo(1) r lo3( r )r loo(1) 1099. 7 (8 )

12 .O83(3) 12 .o88(4)12 .083(3) 12 .081(3)7 , s 8 3 ( 3 ) 7 . 5 7 8 ( 4 )7 . s 8 3 ( 3 ) 7 . s 8 r ( 3 )

r r o 7 . 2 ( 7 ) 1 1 0 7 . 4 ( 9 )t to? .2(7) 1106.5(8)

12 .087(5) 12 .109(4)12 .087(5) r2 ,o93(3)7 ,5?7 (3> 7 ,57 t (4 t7 .571(3 ' 7 .568(4)

1ro7( r ) r l ro0)l loT(1) 1105,8(8)

1 2 . 1 3 4 ( 4 ) 1 2 . 1 4 9 ( 3 )1 2 . 1 3 4 ( 4 ) r 2 . l 3 0 ( 3 )7 . 5 7 1 ( 4 ) 7 . 5 7 8 ( 4 )7 ,s7L(1) 7 .568(3)

I U 4 . 9 ( 9 ) 1 1 1 8 . 6 ( 8 )1 1 1 4 . 9 ( 9 ) 1 r 1 3 . 7 ( 8 )

12 ,147(3) 12 .153(3)t2 . r47(3 ' 12 .141(4)7 ,566(3 ' 7 .568(4)r ,565(3) 7 .s66(3)

u 1 5 . 5 ( 7 ) r r 1 7 . 8 ( 8 )1 1 1 6 . 5 ( 7 ) r 1 1 5 . 3 ( 9 )

L2 ,L47(3 ' 12 . r53(3)L2 .L47 (3 ' L2 ,149(3)7 .552(3) 7 .556(3)7 .562(3 ' 7 .561(3)

r r 1 5 . 9 ( 7 ) 1 1 1 6 . 1 ( 7 )r 1 1 5 . 9 ( 7 ) r 1 1 6 . 2 ( 8 )

1 2 . 1 8 5 ( s ) r 2 , 1 9 6 ( 3 )12 ,185(5) 12 .16 l (3 )r .578(4) 7 .579(4 'r .578(4) 7 .s73(2 '

l r 2 s ( 1 ) 1 r 2 7 , s ( 9 )1 1 2 5 ( 1 ) r 1 2 3 . 8 ( 8 )

12 .041(3) 12 .o57 (4 )1 2 . o 1 7 ( 4 ) 1 2 , 0 1 r ( 3 )7 , s 9 6 ( 3 ) 7 . 6 0 1 ( 3 )7 . 5 9 4 ( 3 ) 7 . 5 9 0 ( 4 )

r l o l . 3 ( 7 ) r 1 0 5 . 1 ( 9 )1096.7(9) ro95.o(9)

L2 .O77 14) 12 ,089(4)12.049(4) 12.046(3)7 . 5 8 8 ( 2 ) 7 . 5 7 8 ( 3 )7 . 5 8 9 ( 3 ) 7 , s 7 9 ( 2 )

1106.8(9) r loT(1)1 r 0 r . 9 ( 9 ) 1 0 9 9 . 9 ( 7 )

L2 .Lo2(4) 12 .117(3)12 .0E8(2) r2 .083(3)7 . 5 8 2 ( 3 ) 7 . 5 8 9 ( 4 )7 . 5 8 3 ( 3 ) 7 . s 8 8 ( 4 )

r r r o . 5 ( 7 ) 1 r 1 4 , 3 ( 9 )1 1 0 8 , 2 ( 6 ) r r o 8 . l ( 8 )

12 ,113(2) L2 , r27(3)r2 ,o9o(4) r2 .087(3)7 .576( t ' 7 ,579(4)7 ,574( l ) 7 ,576(3)

r n 1 . 7 ( 7 ) 1 r 1 4 , 8 ( 9 )11o7.3(9) 1106,9(8)

1 2 . r s s ( 3 ) 1 2 . 1 6 3 ( 3 )1 2 . 1 3 5 ( 3 ) r 2 . r 3 7 ( 3 )7 .570(3) 7 .569(2)7 .5 f4 (3) 7 .510(4)

r r 1 8 , 6 ( 8 ) 1 r 1 9 . 7 ( 7 )r r 1 5 . 5 ( 7 ) 1 r 1 5 . 1 ( 9 )

L2 , t64(4 ' 12 . r75(3)12 .145(3) t2 .142(2 '7 .56r (4 ) 7 ,557 (3 't .55r (4 ) 1 ,s67 (3 'r r 1 8 ( r ) r r 2 1 . 8 ( 7 )

r 1 1 5 . 4 { 8 ) r r 1 5 . 7 ( 7 )

1 2 . r 5 8 ( 4 ) r 2 . r 5 7 ( 3 )1 2 . r 4 3 ( 4 ) r 2 , r 4 5 ( 3 )7 . 5 6 1 ( 4 ) 7 , 5 6 3 ( 3 )7 . 5 7 6 ( 3 ) 7 . 5 6 0 ( 4 )

1 1 r 7 . 8 ( 9 ) r 1 r 9 . 7 ( 8 )r r r 5 . 9 ( 9 ) r r 1 5 . 3 ( 8 )

r2 .196(3) l2 ,2OO(4)r 2 . 1 8 7 ( 3 ) r 2 . 1 8 8 ( 3 )

7 .584(3) 7 ,577(3)7 .517 (4 ) 7 .578(3)

1 r 2 8 . 2 ( 7 ) r 1 2 7 . 8 ( 9 )1 r 2 s . 4 ( 9 ) r 1 2 5 . e ( 8 )

1 2 . o 5 8 ( 4 ) r 2 . 0 8 7 ( 4 ) r 2 . 0 9 8 ( 3 )1 2 . 0 o 9 ( 3 ) 1 2 . 0 1 1 ( 3 ) r 2 . o r 8 ( 4 )7 . 5 9 8 ( 4 ) 7 . 5 8 9 ( 3 ) 7 . s 9 1 ( 3 )7 .597 (3 ) 7 ,599(3) 7 .592(3)

1 1 0 6 . s ( 9 ) 1 1 0 8 ( r ) 1 r r r . 2 ( 9 )1 0 9 5 . 7 ( 7 ) r o 9 5 . 4 ( 7 ) 1 0 9 6 . 7 ( 9 )

L 2 , O 9 7 ( 3 ) 1 2 . 1 1 6 ( 3 ) 1 2 . 1 3 0 ( 4 )L2 ,O44( t ) r2 .039(2) 12 .Os2(3)7 ,592(3 ' 7 .585(4) 7 .s82(4)7 .585(3) 7 ,586(2) 7 .s89(4)

r r r 1 . 1 ( 8 ) r 1 1 3 , 8 ( 9 ) r r r 5 . 7 ( 9 )1 1 o o , 5 ( 8 ) 1 0 9 9 , 7 ( 5 ) r r o 2 . 3 ( 8 )

1 2 . r 2 8 ( 4 ) 1 2 , 1 4 0 ( 3 ) 1 2 . 1 5 0 ( 4 )L2 .079(4) 12 .084(3) 12 ,080(3)7 . 5 8 r ( 3 ) 7 . 5 E 3 ( 3 ) 7 . 5 8 4 ( 3 )7 ,578(4) 7 ,586(2) 7 ,588(3)

1 1 1 s . 2 ( 9 ) 1 1 1 7 . 7 ( 8 ) 1 1 r 9 . 7 ( 8 )11o5.7(9) 1108.o(8) 11o7,3(8)

12 . r r9 (3) L2 . r29(4)12 .oL5(2 ' 12 .o21(3)

7 .598(4) 7 .s97 (2 '7 . 6 0 3 ( 4 ) 7 . 5 9 9 ( 3 )

1 r 1 6 . 1 ( 8 ) 1 1 1 7 . 7 ( 9 )1 0 9 7 . 7 ( 8 ) 1 0 9 8 . 2 ( 7 )

L2 ,L42<3 ' 12 . 16r (4 )r2 ,046(4) 12 .041(3)7 . 5 8 5 ( 3 ) 7 . 5 9 o ( 3 )7 .588 (4 ) 7 .587 (3 )

r l r 8 . 3 ( 9 ) 1 1 2 2 ( 1 )1 1 0 1 . r ( 9 ) r r o o . r ( E )

12.166(4) 12 .178(2)1 2 . 0 8 4 ( 3 ) r 2 . o 8 5 ( 3 )

7 . 5 8 5 ( 3 ) 7 . 5 8 8 ( 3 )7 .583(3) 7 ,582(a)

r 1 2 2 ( r ) r r 2 s . s ( 6 )1 r o 7 , 4 ( 7 ) 1 1 0 7 , 5 ( 9 )

12 . r73(2) 12 , r85(2)r2 .o8s(3) r2 .o91(2)7 ,577(4) 7 .572(3 '7 . 5 7 r ( 3 ) 7 . 5 7 5 ( 4 '

1123.0(8) r r24 ,4(7)1 1 o 5 . 8 ( 7 ) 1 1 o 7 . 5 ( 8 )

12 ,2O3(2) r2 .2 rO(4)1 2 . r 3 3 ( 4 ) 1 2 . 1 4 0 ( 3 )7 ,576<4) 7 .569(4)7 , 5 7 5 ( 2 ) 7 . 5 7 r ( 3 )

r r 2 8 . 3 ( 8 ) r l 2 8 ( 1 )r r 1 5 . 3 ( 8 ) 1 1 1 5 . 9 ( 8 )

t2 .2O9(3) 12 .218(4)r 2 . 1 4 5 ( 5 ) 1 2 . r 4 r ( 4 )

7 .s62(4) 7 .565(3)7 ,s68(2 ' 7 .558(3)

1L27.3(A) r129,4(9)1 r 1 6 ( r ) 1 1 1 4 , 1 ( 9 )

12 .196(3) L2 .203(2)r 2 . 1 4 8 ( 3 ) r 2 , 1 5 5 ( 4 )7 .560(4) 7 ,562(3)7 .s59(3) 1 .564(4)

1 1 2 4 . 6 ( 9 ) r r 2 6 . 2 ( 7 )1 l l5 . t (E) r I r7 ( r )

12,22O(4) L2.224(2)r 2 . 1 2 6 ( 3 ) 1 2 . 1 8 2 ( 2 )7 .575(3) 7 ,582(4 '7 .579(3) 7 .574(3)

1 1 3 r . 2 ( 9 ) r r 3 3 , 2 ( 7 )1 1 2 5 . 6 ( 8 ) 1 1 2 4 . 1 ( 7 )

L2 . r72( t ) r2 . r87(3) 12 . r98(3)L2 . r29(2) r2 . r32(3) 12 . r38(3)

7 ,514( t ' 7 ,566(1 ' 1 ,572(3 '7 .567(3) 7 .569(2 ' 7 ,572(3)

1 1 2 2 . 3 ( 8 ) 1 1 2 3 . 8 ( 8 ) 1 1 2 6 . 7 ( 8 )1 r 1 3 . 2 ( 7 ) 1 1 1 4 , 1 ( 7 ) r 1 1 5 . 7 ( 7 )

1 2 . 1 5 3 ( s ) 1 2 . 1 6 0 ( 3 )12 .089(4) t2 .O97 (2 )

7 . s 8 1 ( s ) 7 . 5 7 8 ( 3 )7 ,s75(4) 7 .579(2)r 1 l 9 ( r ) 1 1 2 0 . 6 ( 7 )

r r o 7 . o ( 9 ) r 1 0 9 . 2 ( 6 )

L2,L92(2) t2,20t(2'L2 ,L48(4 ' 12 .144(3)7 . 5 7 1 ( 3 ) 7 , 5 6 7 ( 3 '7 ,512(3) 7 ,567 (4)

r 1 2 4 . o ( 7 ) n 2 6 . 6 ( 6 )1 1 1 7 . 5 ( 9 ) 1 1 1 5 . 0 ( 9 )

1 2 . 1 4 o ( 3 )r2 .o89(3)7 .559(4)7 ,577 (3 )

1 1 1 5 , 6 ( 9 )r r o T , 5 ( 7 )

t2,L84(2)r 2 , 1 5 2 ( 3 )

7 .564(2)7,565(3)

1 r 2 3 . 0 ( s )r r 1 7 . 3 ( 8 )

a

a

vv

t2 ,L74(4 ' 12 ,L82(2) r2 . r92(4)12 .147(3) 12 .148(3) L2 .L52(2)7 ,559(2) 7 .563(3) 7 ,s67(3)7 . s s 8 ( 3 ) 7 . s 6 3 ( 3 ) 7 . 5 6 0 ( 3 )

r t 2 o , 3 ( 7 ) r r 2 2 , s ( 6 ) 1 1 2 4 ( 1 )1 r r 5 , 3 ( 8 ) r r r 6 . 2 ( 8 ) r l r 5 . 6 ( 6 )

12,209(3) t2,209(2' t2,2L6(2)1 2 , 1 8 5 ( 4 ) 1 2 . 1 8 3 ( 4 ) 1 2 , r 8 7 ( 3 )7 ,s8r (3 ) 7 .576(4 ' 7 ,579(3)7 .s8r (4 ) 7 .575(3) 7 ,576(4)

1 1 3 0 . r ( 7 ) r 1 2 9 . 5 ( 8 ) 1 r 3 r . 1 ( 7 )r 1 2 5 ( 1 ) r r 2 4 . 5 ( 9 ) 1 r 2 s . 3 ( 8 )

I

!gcvF

aAI

Drta col lcccrd durid8 hc.t ln8.

D.t . col lect.d r f tcr hert in8.

Strod.rd d.vi l t ionr fs lodic.rcd in prrrnthcr. ! in lenr of rhc l l r t l i8ni f ic.nt f iguE.i .

S A M P L E1 7 6 % M e

GRAZIANI AND LUCCHESI: THERMAL BEHAVIOR OF SCAPOLITES

12 .16

1 2 , 0 8

12 .OO

S A M P L E 2 I2 8 8 % M e ] /

1-c,/,/-

o/tv/-v/

9

12.20

1 2 . 1 2

12.O4

t

,'rrr", r' ,l

S A M P L E 33 8 6 % M e

S A M P L E 55 5 7 % M e

1 2 . 2 4

1 2 . 1 6

400 800 0 400 800

T o cFig. 3. Variations in a during heating. All the samples show a linear trend with a slope inversely proportional to the meionite

content of the specimen.

a maximum of 800'c, a limit slightly below thetemperature at which plagioclase and sodium andpotassium chloride form.

The heating experiments indicated that risingtemperatures lead to a linear increase in the valuesof the refractive indices, up to 350-375" C, at aslightly higher rate in the case of e than in that of ar.As a first approximation, thii increase was propor-tional to the meionite content of the sample.

The slope values relative to linear increase ofboth e and (e + o)12 during heating, have beenplotted vs. chemical composition, by means of alinear regression yielding the following relation:

d e l d T : 1 . 3 0 x l \ - 7 M e % o + 1 5 . 2 3 x 1 0 - 7

d[(e + a) l2] ldT: l . l3 x l \ -7MeVo + 5.22 x 10-7

where deldT and d[(e + a)lTUdT refer to the in-crease of e and the mean refractive index duringheating, while MeVo is the meionite percentage ofthe heated sample. However, a wide data disper-sion was noted, since the correlation coemcients

are respectively of I : 0.907 and F : 0.850, whichdoes not allow a clear linear correlation. At highertemperatures, a sudden increase was recognized forboth the refractive indices, inversely proportionalto the meionite content (Fig. 6).

The refractive indices measured after cooling toroom temperature were almost unchanged for eachsample over the whole temperature range investi-gated. Slight variations from the linear trend werenoted, however, when measurements were per-formed after heating to temperatures above 600' C,especially on sulphur rich samples and on thosecontaining a low percentage of meionite. The varia-tions were at the estimated error limits of about-'-0.001. A decrease in birefringence values was alsofound in all the samples, following a linear trendinversely proportional to meionite content.

Color, too, was afected by heating. There was aprogressive fall in color intensity, until the samplesbecame completely colorless at temperature be-tween 350'and 400'C.

GRAZIANI AND LUCCHESI: THERMAL BEHAVIOR OF SCAPOLITES t237

100

M e %

Fig. 5. Variation in the anorthite content of the plagioclaseformed vs. meionite content of the scapolite.

tained by investigating the relationship betweenmeionite content and the weight loss occurringbetween 390" and 850'C; in this way only theamounts of COz and SO2 released were involved(F ig .7 ) .

The release of NaCl and KCl, which takes placebetween 800" C and 1250' C, displays a linear trendwhen plotted against the meionite content of thespecimen; the highest values being found in thesodium-rich samples (Fig. 7), similar to the resultsof Evans et al. (1969).

The thermogravimetric data, supported by stoi-chiometric considerations, suggest that all the chlo-rine present is released as halogen compounds inthis temperature range. The appreciable degree ofscattering seen in the upper graph in Figure 7 maybe due to the presence of potassium chloride, whichis always found, even if in small amounts, inscapolites. The presence of potassium, which isconsidered to be structurally equivalent to sodiumin such minerals, increases the total weight loss tobe expected from a sample with a given meionitecontent.

A number of chemical data reported by Shaw(1960), Ingamells and Gittings (1967), and Evans e/al. (1969) were plotted using the least squaresmethod against meionite content, yielding a linearrelation. This was done by calculating the totalweight of NaCl and KCI which could be released by

Discussion

The thermal analyses showed a similar trend forall the examined samples, independent of theamounts of the various gaseous compounds re-leased.

Up to the temperature of 850" C, the release ofH2O, CO2, and SO2 can be roughly correlated withthe meionite content of the sample examined. Itshould be stressed, however, that CO2 and SO3content can be related to the percentage of meioniteon a stoichiometric basis (Shaw, 1960), while this isnot true for H2O, whose structural role has not yetbeen clearly defined (Evans et al., 1969). An exces-sive uncertainity, nearly 30-40 percent, may, thus,occur in assessing the meionite content of thescapolite samples. Thus better results were ob-

r l I I r --r---T----r---l

21C i4

4 . 1 6 "

.\

\ .\..

\

0 50 100

M e %Fig. 4. Increasing rate of lattice parameter variation during

heating, vs. the meionite content of the sample.

o-Q

50

r00

F

o l

F

> l

GRAZIANI AND LUCCHESI: THERMAL BEHAVIOR OF SCAPOLITES

.//,4

u - 6 '- - 4 1 '

# v c

1 . 5 5 2

1 . 5 5 0

1 . 5 5 3

ence of SO?- using X-ray fluorescence, while infra-red studies by Schwarcz and Speelman (1965) re-vealed low amounts of SO3-, while SiOX- did notallow the recognition of SO?-. The emission of SO2,revealed by mass spectrometry, could suggest thatit derives from SO3-, according to the equilibrium:2SO3 C 2SO2 + 02. It should, however, be notedthat it has been reported that the light alkaline metalsulphates vaporize by decomposition to the metalelement, plus SO2 and Oz (Ficalora et al., 1968).This reaction would account for the absence of SO3in the released gaseous phase. In this last casesulphur should be coordinated as SO?- in scapo-lites.

Papike and Stevenson (1966) investigated theeffect of the chemical composition of scapolite on

---Ya- 4

-./94

olq,/

,./l

srt

1 . 5 4 9

1 . 5 8 3

1 . 5 7 9

1 . 5 7 0

S A M P L E 6 S A M P L E 8

t . 5 6 6

500 0 Toc 500Fig. 6. Increase in e (filled circles) and mean refractive index (open circles) during heating. All the samples show a linear trend up to

350-375'C proportional to tbe meionite content, followed by a sudden increase.

' | P L E 3 S A M P L E 5 1 / .

, / /

o'vy' y'ta

-a"- ..i-'

.rt

the sample, assuming a complete loss of chlorine,and a Na/K ratio the same in the solid and gaseousphases. The fact that the straight line in the uppergraph in Figure 7 cuts the abscissa at about 80percent Me can be explained by the fact that athigher meionite content the anion sites contain onlycol-.

The H2O release occurs until ca.400" C with twomain weight losses respectively starting at 70-100'C and 150-250'C, thus suggesting the exis-tence of two types of H2O in the scapolites. Howev-er, the absence of a wide emission gap between thetwo weight losses, and the continuous H2O release,do not support such a hypothesis. Continuous emis-sion is also characteristic of CO2 and SO2.

Chappel and White (1968) recognized the pres-

L I

0 50 1ooM e %

Fig. 7. Relationship between the release of volatiles andmeionite content. The total CO2 + SO2 weight loss rises as thepercentage of meionite falls, while the toral NaCl + KCI weiehtloss fails, reaching zero at about 80Vo Me.

its structure, suggesting that with rising meionitecontent, Al3+ cations enter Sia+ tetrahedra, makingthem rotate and thus increasing the value of c.

The thermal expansion of scapolites which leadsto an increase in a and constant c during heating,has already been investigated in a 33.5 percentmeionite scapolite by Levien and Papike (1976) andcould be attributed to the rotation of Si and Altetrahedra in a direction opposite to that caused bychemical composition; this would lead to an open-ing of the cation channels and thus to an increase inthe values of a.

The c parameter, on the other hand, remainsconstant with both compositional and thermal vari-ation because it is difficult to rotate tetrahedra thatlink in that direction.

It is, however, striking that the present study,carried out on specimens representative of thewhole marialite-meionite series, indicates that dur-ing heating the increases in a and V are inverselyproportional to the meionite content. The relation-

1239

ship could be attributed to the effect of the reverserotation of a ring of tetrahedra due to simultaneouschanges in chemical composition and temperature.This hypothesis would account for the reversibilityof the thermal expansion of this mineral, since insamples cooled from temperatures no higher than800o C, a returns to its initial value at room tem-perature (Table 2). It should be noted that slightdiscrepancies from this behavior may be caused bycooling rate.

When heating exceeds 1000" C, an irreversibleincrease in a to anomalously high values occurs inall the examined samples. This may be due to asmall degree of disordering of Al3+ and Sia* in thetetrahedral sites (Levien and Papike, 1976), and tothe breakdown of scapolite. Such breakdown pro-duces various materials: halite, plagioclase andcalcium compounds. The former easily sublimeseven during short isothermal stages above 1000'C.

The plagioclase produced displays an anorthitepercentage which is lower than that of the scapolitefrom which it derives. In addition, the emission ofNaCl and KCI raises the calcium content of scapo-lite during the formation of plagioclase. It is remark-able that, under equilibrium conditions, the anor-thite content of plagioclase, when crystallized withscapolite, is normally 20-25 percent lower than thecoexisting meionite content of scapolite (Heitanen,1967), though such behavior does not always occur,especially in calcium-rich scapolites (Haughton,1971). Calcium compounds have not been clearlyrecognized in the X-ray diffraction patterns buttheir existence has been inferred by stoichiometricconsiderations bearing in mind that the unmixing ofmore sodic plagioclase from scapolite implies acalcium remnant which should give way to theformation of calcium aluminum silicates with lowersilicon content than the contemporaneously formedplagioclase.

As a first approximation, the refractive indices,too, follow the general trend increasing their valuesduring heating, that is, consistent with more meioni-tic values. However, the considerable disagreementnoted in the linear regression between such increaseand meionite percentage may be ascribed to theclose dependence of the refractive indices of scapo-lite on the content of components such as S, K, Hand excess carbon (Shaw, 1960). Thus the suddenand conspicuous increase in the refractive indicesof all the samples examined at temperatures above375' C could be attributed to the second release ofH2O. Conversely, the decrease in birefringence

GRAZIANI AND LUCCHESI: THERMAL BEHAVIOR OF SCAPOLITES

Ae

3

Y+

oz

0

6

AR

3

oa+

N

P0

1240

values followed a linear trend during heating in allthe samples, and did not seem to be influenced bythe sudden increase in the corresponding refractiveindices.

- Conclusions

(1) For all the scapolites examined, the thermalanalyses indicate a common trend, in which therelease of volatiles occurred uninterruptedly overthe whole temperature range, but with five evidentspecific steps corresponding to the loss of H2O,SO2, CO2, NaCl, and KCl. The weight loss of thelast four was correlated with meionite content.

(2) Heating resulted in linear increases in a and Vthat were inversely proportional to the meioniticpercentage of the examined sample. The relation-ships between the increases in the lattice parame-ters, with rising temperature, and the percentage ofmeionite, can be expressed by the equations:

daldT: -1.59 x l0-6MeVo + 1.85 x 10-aG : o'ee6)

dv ldT: -0 .3 x lO-3MeVo + 0 .034 e :0 .998)

The lattice parameters in samples cooled to roomtemperature after heating to a maximum of 800" Creturned to their initial values. In both cases cremained constant.

(3) An irreversible increase in a beyond ca.1000" C was followed by the formation halite, pla-gioclase and possibly calcium aluminum silicates.The unmixed plagioclase was less calcic than thescapolite from which it was derived.

(4) The optical data confirmed these findings,showing an increase of the refractive indices duringheating and a sharp increase corresponding to thesecond release of H2O.

AcknowledgmentsWe wish to thank Prof. W. F. Eppler, Munich, West Germa-

ny, and Dr. P. E. Desautels, Dept. of Mineral Sciences, Smith-sonian Institution, Washington, U.S.A., for their collaborationin providing some scapolite samples. We are grateful to Proff. F.Baroncelli, CNEN Laboratory of Analytical Chemistry, Casac-cia, and A. D. Magri, Institute of Analytical Chemistry, Univer-sity of Rome, Italy, for their kind assistance in carrying outthermal analyses. Thanks are also extended to Prof. G. Balducciand G. Gigli, Institute of Physical Chemistry, University ofRome, Italy, for performing and interpreting the mass spectrom-eter experiments on heating. The constructive suggestions ofProf. M. Fornaseri, Institute of Geochemistry, University ofRome, Italy, are gratefully acknowledged.

This work was supported both by the grant from the Universi-ty of Rome for inclusions studies and from the ltalian NationalResearch Council (CNR), CT 80.02608.05.

GRAZIANI AND LTJCCHESI: THERMAL BEHAVIOR OF SCAPOLITES

References

Adams, F. D. and Barlow, A. E. (l9ll) G6ologie des r6gionsd'Haliburton et Bancroft, Provincie d'Ontario. GeologicalSurvey of Canada Mdmoire, 6.

Bambauer. H. U., Corlett, M., and Eberhard, E. (1967) Dia-grams for the Determination of Plagioclase using X-ray Pow-der Methods (Part III of Laboratory Investigations on Plagio-clase). Schweizerische Mineralogische und PetrographischeMitteilungen, 47, 33f-349.

Brown, G. E., Sueno, S., and Prewitt,C. T. (1973) A new singlecrystal heater for the precession camera and four-circle dif-fractometer. American Mineralogist, 58, 698-704.

Buddington, A. F. (1939) Adirondack igneous rocks and theirmetamorphism. Geological Society of America Memoirs, 7.

Burley, B. J., Freeman, E. 8., and Shaw, D. M. (1961) Studieson scapolite. Canadian Mineralogist, 6, 670-679.

Chappel, B. W. and White, A. J. R. (1968) The X-ray spectro-graphic determination of sulphur coordination in scapolite.American Mineralogist, 53, 1735-1738.

Donnay, G., Shaw III , C. F., Butler, I . S. and O'Neil , J. R.(1978) The presence of HCI in scapolites. Canadian Mineral-ogist, 16,341-345.

Evans, B. W., Shaw, D. M., and Haughton, D. R. (1969) Scapo-lite stoichiometry. Contributions to Mineralogy and Petrology,24,293-305.

Farinato, R. and Loreto, L. (1975) A least-squares refinement ofthe crystal lattice constants and evaluation of their errors,using the direct unit-cell. Rendiconti della Societi ltaliana diMineralogia e Petrologia, 31, 487-500.

Ficalora, P. J., Uy, O. M., Muenow, D. W., and Margrave,J. L. (1968) Mass spectrometric studies at high temperatures.XXIX, Thermal decomposition and sublimation of alkali metalsulphates. Journal of the American Ceramic Society, 51, 574-577.

Franzini, M. (1965) L'errore nella misura dell'indice di rifra-zione, il metodo della deviazione minima nei prismi. Atti dellaSocitir Toscana di Scienze Naturali, Serie A, 72, 589-591.

Gibbs, G. V. and Bloss, F. D. (1961) Indexed powder difractiondata for scapolite. American Mineralogist, 46, 1493-149'7.

Goldsmith, J. R. (1976) Scapolites, granulites and volatiles in thelower crust. Geological Society of America Bulletin, 87, l6l-168.

Haughton, D. R. (1971) Plagioclase-scapolite equilibrium. Cana-dian Mineralogist, 10, 854-870.

Hietanen, A. (1967) Scapolite in the Belt Series in the St. Joe-Clearwater Region, Idaho. Geological Society of America,Special Paper 86.

Kauffman, A. J. and Dilling, E. D. (1950) Differential thermalcurves of certain hydrous and anhydrous minerals, with adescription of the apparatus used. Economic Geology, 45,222-2M.

Knorring, O von and Kennedy, W. Q. (1958) The mineralparagenesis and rnetamorphic status of garnet-horneblende-pyroxene-scapolite gneiss from Ghana (Gold Coast), Mineral-ogical Magazine, 31, 846-859.

Ingamells, C. O. and Gittins, J. (1967) The stoichiometry ofscapolite. Canadian Mineralogist, 9, 214-236.

Lacroix, A. (1919) Sur une scapolite des pegmatites de Madagas-car constituant une gemme. Comptes Rendus de I'Accademiedes Sciences, Paris, 169, 261-264.

Levien, L. and Papike, J. J. (1976) Scapolite crystal chemistry:aluminum-silicon distributions, carbonate group disorder, andthermal expansion. American Mineralogist, 61, 8U-877.

Lin, S. B. and Burley, B.J. (1973a) Crystal structure of asodium and chlorine-rich scapolite. Acta Crystallographica,829,1272-t278.

Lin, S. B. and Burley, B. J. (1973b) The crystal structure ofmeionite. Acta Crystallographica, 829, 2024-2026.

Lin, S. B. and Burley, B. J. (1975) The crystal strucrure of anintermediate scapolite-wernerite. Acta Crystallographica,83 l , 1806-1814.

Lovering, J. F. and White, A. J. R. (1964) The significance ofprimary scapolite in granulitic inclusions from deep-seatedpipes. Journal of Petrology, 5, 195-218.

Lovering, J. F. and White, A. J. R. (1969) Granulitic and eclogi-tic inclusions from basic pipes at Elegate, Australia. Contribu-tions to Mineralogy and Petrology, 21,9-52.

Lovering, J. F. and Widdowson, J. R. (1968) Electron micro-probe determination of sulphur coordination in minerals. Lith-os, 1,264-267.

Newton, R. C. and Goldsmith, J. R. (1976) Stability of the end-member scapolites: 3NaAlSi3O6 + NaCl, 3CaAl2SirO. +CaCO3, 3CaAl2Si2Os + CaSoa. Zeitschrift ftir Kristallogra-phie, 143, 333-353.

Papike, J. J. (1964) The crystal structure and crystal chemistry ofscapolite. Ph.D. Thesis, University of Minnesota.

Papike, J. J. and Zoltai, T. (1965) The crystal structure of amarialite scapolite. American Mineralogist, 50, 641-655.

Papike, J. J. and Stephenson, N. C. (1966) The crystal structureof mizzonite, calcium- and carbonate-rich scapolite. AmericanMineralogist, 51, 1014-1027 .

l24l

Rucklidge, J. and Gasperrini, E. L. (1969) Electron microprobeanalytical data reduction. EMeADR VII. Manual reissue, De-partment of Geology, University of Toronto, Toronto, Cana-da.

Scarano, E. and Calcagno, C. (1975) High sensitivity carbondioxide analyser. Analytical Chemistry, 47, 1055-1065.

Scherillo, A. (1935) La meionite del Somma-Vesuvio. Periodicodi Mineralogia, 6, 227-239.

Schwarcz, H. P. and Speelman, E. L. (1965) Determination ofsulphur and carbon coordination in scapolite by infraredabsorption sp€ctrophotometry. American Mineralogist, 50,656-ffi6.

Shaw, D. M. (1960) The geochemistry of scapolite. Part I.Previous work and general mineralogy. Journal of Petrology,1,218-2ff i .

Solesbury, F. W. (1967) Gem corundum pegmatites in NETanganyka. Economic Geology, 62, 981-991.

Stewart, G. W. (1941) ldocrase and scapolite from Manchester,New Hampshire. American Mineralogist, 26, 509-51l.

Strunz, H. (1978) Mineralogische Tabellen. Geest and Porting,Leipzig.

Ulbrich, H. H. (1973) Cristallographic data and refractive indi-ces of scapolites. American Mineralogist, 58, 8l-92.

Zwaan,P. C. (1971) Yellow scapolite, another gem-mineral fromUmba, Tanzania. Journal of Gemmology, 12,304-3W.

Manuscript received, August 5, l98l;acceptedfor publication, April 28, 1962.

GRAZIANI AND LUCCHESI: THERMAL BEHAVIOR OF SCAPOLITES