a calorimetric study of synthetic amphiboles along the ... · a calorimetric study of synthetic...

American Mineralogist, Volume 7 9, pages I I I 0- I I 2 2, I 994

A calorimetric study of synthetic amphiboles along the tremolite-tschermakite joinand the heats of formation of magnesiohornblende and tschermakite

EucnNr A. Snnrr.rxDepartment of Geological and Geophysical Sciences and Princeton Materials Institute, Princeton University, Princeton,

New Jersey 08544, U.S.A.

Dlvro M. JpxrrnsDepartment of Geological and Environmental Sciences, State University of New York at Binghamton, Binghamton,

New York 13901. U.S.A.

Ar,nx,lNon c, NlvnorsxYDepartment of Geological and Geophysical Sciences and Princeton Materials Institute, Princeton University, Princeton,

New Jersev 08544. U.S.A.

Arstnlcr

A suite of synthetic calcic amphiboles in the tremolite-tschermakite (Tr-Ts) series hasbeen studied by high-temperature drop-solution calorimetry in molten 2PbO.BrOr. Thepseudobinary join, which is shifted toward magnesiocummingtonite (Mc) by l0 mol0/0,extends from tremolite (Tr), Ca,rMgrrSirOrr(OH)r, to magnesiohornblende (MgHb),Ca,,MgrAlrsirorr(OH)r, and represents the operation of one full Mg-Tschermak substi-tution, I6rMg,l4rsi = IoJdl,t+l[], into the tremolite fOrmula. Transmission electron micro-scope (TEM) examination of the amphiboles shows that all the samples contain a lowconcentration of (0 l0) chain multiplicity faults (CMFs), suggesting that the increase of Alin the amphibole does not significantly increase or decrease the tendency ofthese defectsto form. The calorimetric measurements indicate that the energy change associated withone full Mg-Tschermak substitution is quite small, on the order of -5 to -10 kJlmol.Enthalpies of formation from the elements at 298.15 K and I bar have been calculatedusing the new calorimetric data for the magnesiohornblende (MgHb) and tschermakite(Ts)end-members.ThevaluesofAlf forMgHbandTsare -12401.2 + 10.6 and-12527.7+ 16.4 kJlmol, respectively. The entropy of MgHb, at 298.15 K and I bar, has also beencalculated using four different site-mixing activity models involving the octahedral (M2)and tetrahedral (Tl) sites and the experimental phase-equilibrium data of Jenkins (1994).Two of the models, the two-site and four-site coupled models, are essentially indistin-guishable and gave the most consistent values for ^q"Hb, 575.1 + 3.9 and 575.6 + 3.9J/(mol. K), respectively.

Irvrnooucrrox

The general form of the Tschermak cation exchangecan be expressed as r61M,l4lsi = I6lAl,t41Al, where M nor-mally refers to a divalent cation such as Fe, Mg, or Ca.This exchange is of fundamental importance in many sil-icate solid solutions, including the pyroxenes, amphi-boles, micas, chlorites, and melilites. In amphiboles, thiscoupled substitution is of primary importance in two sol-id-solution series, the orthorhombic ferromagnesian se-ries and the monoclinic calcic series. Recent transmissionanalytical electron microscope (TEM-AEM) studies haveshown that this substitution dominates during exsolutionreactions between anthophyllite and gedrite in the ferro-magnesian series (Smelik and Veblen, 1993) and betweenactinolite and hornblende in the calcic series (Smelik eta l . , l99 l ) .

Any attempt to relate amphibole composition to a spe-cific P and ?" or to an exchange reaction in P-Z spacerequires accurate thermodynamic data for certain end-member amphibole compositions. For calcic amphiboles,these data have generally come from phase-equilibriumstudies (e.g., Welch and Pawley, l99l1, L6ger and Ferry,l99l; Jenkins, 1994). Thermodynamic data for a largenumber of rock-forming minerals, including amphiboles,are tabulated in various thermodynamic data bases (Ro-bie et al., 1979; Helgeson et al., 1978; Berman, 1988;Holland and Powell, 1990).

In this paper we present results from a calorimetric studyof synthetic amphiboles in the tremolite-tschermakite (Tr-Ts) series [CarMgrSirOrr(OH)r-CarMgrAloSiuOrr(OH)r].The primary goals of the study are to (l) investigate theenergy change associated with the Mg-Tschermak exchangein calcic amphibole and (2) determine independently the

0003-004x/94 / | | r2-t I I 0$02.00 1 1 1 0

SMELIK ET AL.: CALORIMETRIC STUDY OF SYNTHETIC AMPHIBOLES 1 l 1 t

Ca2Mg3Al2[Al2Si6] q2(OH)2

Ts

MgHb

AMPH I44

AMPH 28-2

AMPH 19-5

AMPH 25-1

AMPH 29-1

TREM 23-IO, TREM 23-5

Tr tM4lCa = tM4lMg MC Tr IVICCa2Mg5[Sfu]Qz(oH)2. Mg7[Sis]Q2(oH)2

Fig. l. (a) Plot of tremolite-tschermakite-magnesiocummingtonite (Tr-Ts-Mc) showing starting compositions for amphibolesynthesis, made at intervals of 10 molo/o along a pseudobinary join shifted toward the Mc apex by l0 mol9o (circles). The solubilitylimit for Al in tremolite is near the MgHb end-member. (b) Representative electron microprobe analyses of the synthetic amphiboles.The dashed line at 0.5 Ts represents the operation of one full Mg-Tschermak substitution, i.e., the Al content of magnesiohornblende.

a.

f

heats of formation for the amphibole end-members mag-nesiohornblende (MgHb), CarMgoAlrSi'Orr(OH)r, andtschermakite, and compare these results with those fromphase-equilibrium studies. By incorporating thermo-chemical data from Berman (1988) and Holland andPowell (1990) in our analysis, the results presented hereshould be compatible with those internally consistent datasets.

Enpnnnrrcxur. METHoDS

Amphibole synthesis

All amphiboles were synthesized from mixtures of re-agent grade CaCOr, MgO, and AlrOr; SiO, was preparedfrom desiccated silicic acid. Previous experimental workby Jenkins (1988) and Cho and Ernst (1991) has shownthat the solubility limit of Al in tremolite approaches themagnesium hornblende end-member composition,Ca,MgAlrSi?Orr(OH)r, which is equivalent to one fullMg-Tschermak substitution into the tremolite formula.For the synthesis of amphiboles for calorimetry, we pur-posely chose a join parallel to the true join of tremoliteand magnesium-hornblende, offset by l0 molo/o magne-siocummingtonite (Mc), to obtain the purest amphiboleyields possible. The magnesium-cummingtonite compo-nent is reflected by a constant substitution of 0.2 Mg forCa in the M4 site across the entire join (Fig. la). We choseto offset thejoin on the basis offindings ofprevious work-ers, who reported a 5-l0o/o enrichment of Mg in the M4site in their attempts to synthesize end-member tremolite

(Jenkins, 1987; Graham et al., 1989; Pawley et al., 1993).Therefore, the amphiboles were synthesized at l0 molo/oincrements along the join Ca, , Mg, , Si, Or, (OH)r -Ca, rMgorAlrsirorr(OH)r. Electron microprobe (EMP)analysis of the amphiboles (discussed below) and previ-ous work on similar compositions by Jenkins (199a) in-dicate that the compositional offset is preserved along thejoin. Approximately I wto/o qvarlz was mixed into eachexperiment to compensate for the loss of silica to theambient fluid during hydrothermal treatment (Jenkins,1987). Syntheses were performed by loading a stoichio-metric oxide-carbonate mixture into a Pt capsule sealedat one end, heating it with a butane torch to drive of CO,from CaCOr, and then sealing the other end after adding25-30 wto/o HrO.

Syntheses below l0 kbar were performed in internallyheated vessels using Ar as the pressure medium. Pres-sures were measured with both a PPI bourdon-tube andHarwood manganin resistance-coil gauges, the latter be-ing factory calibrated against a dead-weight piston gauge.Pressure measurements are believed to be accurate to + 50bars. Temperatures were measured with two chromel-alumel thermocouples calibrated against the freezing pointof NaCl and are believed accurate to + I K. Synthesesabove I 0 kbar were carried out in a piston-cylinder presst/z in. rn diameter using NaCl as the pressure medium.The syntheses were performed in a hot piston-out fashionby pressurizing the assemblage to about 2 kbar below thedesired pressure and then using the thermal expansion ofthe assemblage to reach the desired pressure. Pressure

. 1 4 4o 28-26 19-5o 25-lo 29-ltr 23-10,23-5

ttt2

TABLE 1. Details of amphibole synthesis

SMELIK ET AL.: CALORIMETRIC STUDY OF SYNTHETIC AMPHIBOLES

Sample Composition r(K) P (kbar) r (h) Synthesis products'

TREM 23.10

TREM 23-5AMPH 29.1AMPH 25-1AMPH 19-5

AMPH 28-2AMPH 14.4

Ca, 6Mg5rSisOr2(OH),

Ca, sMgs,SisOr2(OH),Ca, sM95oAlo4SiTsOz(OH),Cal sMg4sAlosSiT"Oz(OH),Cal sM946A11 2Si7oO"r(OH),

Ca, sMgliAll 6SiTrOz(OH),Ca, sMg4 rAl2oSiT oOz(OH),

113q10 )

1 1 1 1 (17 )1 1 37(5)1 1 48(1 0)1 134(5)

1 1 15(s)1 096(s)

s.20(70)

6.10(10)6.35{5)6.0q5)

13.0(3)

12.9(3)12.0(5)

239

188237109120

amph, trace cpx, andtrace opx

amph and trace cpxampnamphamph, trace opx, and

trace qtzampnamph, cor (1.6 wtTo)*,

and trace opx

14770

/Vote: the error ranges for P-f conditions (shown in parentheses) take into account small P and ffluctuations during the course of the experiment.'Amph : amphibole; cor : corundum; opx : orthopyroxene; cpx : clinopyroxene; qtz : quartz; trace : below detection limit on powder XRD

(about 1.0 wto/").** Amount determined by Rietveld modal analysis.

transmission to the sample was calibrated by the reactionalbite : jadeite + qvartz at 873 K to a precision of +250bars, and it was found that virtually no correction wasrequired to obtain the accepted value of 16.25 kbar (Jo-hannes et al., l97l;, Holland, 1980). Temperatures weremeasured with chromel-alumel thermocouples newlymade for each experiment and have an estimated uncer-tainty of t 5 K. The synthesis conditions for the amphi-boles are summarized in Table l.

Amphibole characterization

Optical examination. All experimental products wereexamined with the petrographic microscope using an oilof refractive index 1.60, which matches n, of the amphi-bole, to identify the presence ofother phases. Only traceamounts (<lolo) of impurities were detected for somesamples (Table l).

Electron microprobe. The synthetic amphiboles wereanalyzed using the Cameca SX-50 fully automated elec-tron microprobe (EMP) at the Princeton Materials Insti-tute, in wavelength-dispersive mode, operated at l5 keVand 20 nA. The beam diameter was typically I pm. Min-eral standards for Mg, Ca, Si, and Al were MgO, CaSiOr,and AlrSiOr. Analytical corrections were made using thePAP version of the fubA method (Pouchou and Pichoir,1985). Performing quantitative EMP analysis on ex-tremely fine-grained synthetic amphibole, typically oc-curring as 5 x 20 pm laths, can be problematic. We haveprepared our samples using conventional epoxy-mount-ing and polishing techniques (e.g., Graham et al., 1989;Raudsepp et al., l99l). The technique satisfies the re-quirement for a flat, polished surface but introduces thepossibility that if the grain is thin, the analysis may becontaminated by other grains below it. For further dis-cussion of this technique and grain-dispersal techniques,see Jenkins (1994).

Only the largest grains were analyzed, and even thenwe were never absolutely certain that the excitation vol-ume of the beam was totally within the amphibole crys-tal. When it was not, the result usually was low analyticaltotals. An analysis was considered acceptable when theanalytical sum was between 95 and 99o/o, and reasonable

amphibole stoichiometry was achieved when the analysiswas normalized to 23 O atoms. Representative EMPanalyses are plotted on the Tr-Ts-Mc triangle in Figurelb for comparison with the end-member compositions(Fig. la). Despite the scatter in the probe data, most ofthe synthetic compositions cluster well along the offsetTr-Ts join. The most Al-rich one, AMPH l4-4, does notmatch the composition of the MgHb end-member. Op-tical examination of this sample indicated the presenceof unreacted corundum from the original oxide mix inthe experiment products (Table l). The amount of co-rundum was quantifiable by Rietveld modal analysis, asdiscussed below.

X-ray diffraction. All synthetic amphiboles were sub-ject to Rietveld structure refinement. Step-scans weremade of each sample on a Scintag XDS-2000 powderdiffractometer operated in the vertical d-d configurationusing Cu radiation and a Si(Li) solid-state detector. Thestep increment for all amphiboles except sample TREM23-10 was 0.1'20 for counting times of I s (500-1000counts on the maximum peaks), using a 2-mm primaryslit and a 0.3-mm receiving slit. For TREM 23-10 theconditions were changed to a step size of 0.04 20, 5-scounting time, primary slit of 0.5 mm, and a receivingslit of 0.02 mm. Rietveld refinements were performedusing the program DBWS-9006 (Young, 1993) yieldingR"..* values of about 0.08.

For AMPH l4-4, containing a small amount of corun-dum, a Rietveld modal analysis was performed followingthe method of Bish (1988) and Bish and Post (1993). Themodal analysis indicates about 1.6 wto/o corundum. Sincethe initial oxide mix for this composition contained 12.6wto/o AlrOr, it appears that only I1.0 wto/o went into theamphibole, which equates to 1.74 total Al cations or 43.5molo/o tschermakite (0.87 Xr.). This result agrees very wellwith the compositions measured for AMPH l4-4 by EMP(10.38-11.68 wto/o AlrO,, Fig. lb), which are shiftedslightly toward the Tr apex.

The refined unit-cell parameters for the synthetic am-phiboles are summarized in Table 2 and plotted againstthe composition in Figure 2. We used two batches ofTrroMc,o tremolite (TREM 23-5 and TREM 23-10) in

Trsle 2. Crystallographic data for synthetic tremolite-tscher-makite amphiboles

sampre a(A) b(A) c(A) B0 v(A')

l l l 3

9.85

9.80

23-10,,o3-s 29_I O

25-lL4-4

,99 ,"9 o

18.04

18.02

18.00

r7.98

5.275

ro4.9104.8

104.7

104.6

i04.5

f,g

oo

ooo

O O

o

o o

ooo

oo

oo o oo

0.0 1.0 t.74 2.0A{olcations Per 23 O

Fig. 2. Unit-cell parameters for synthetic tremolite-tscher-makite amphiboles (with 10 molo/o Mc) vs. total Al content. Allcell parameters show nearly linear trends with composition' sug-gesting a complete solid solution at the synthesis conditions.

Maresch et al. (1990), and Ahn et al. (1991) has shownvarying degrees of defect concentration in synthetic am-phiboles. The dominant defect type is the chain-multi-plicity fault (CMF), which occurs parallel to (010) of theamphibole structure. These planar faults can be either thesingle-chain variety, which are essentially unit-cell slabsof pyroxene structure, or the multiple-chain variety (>2),which increases the proportion of the mica-like slab ofthe amphibole structure. Characteization of these faultsis important for two reasons. First, if the defect concen-tration is high enough, it can significantly shift the am-phibole stoichiometry from the expected value. Second,a highly defective structure may not be energetically sim-


o{

BrREM 23-5 9.807(2) 18.054(3)TREM 23-10 9.815(1) 18.0580)AMPH 29-1 9.801(1) 18.042(21AMPH 2s-1 9.784(1) 18.027(2)AMPH 19-5 9.771(1) 18.013(2)AMPH 28-2 9.767(1) 17.994(2)AMPH 14-4 9.769(1) 17.995(2)

s.276(1) 104.56(1) 904.2(21s.279(1) 104.58(1) 905.49(9)s.279(1) 104.68(1) 902.9(2)5.280(1) 104.74(6) 900.6(2)s.284(1) 104.80(1) 899.2(2)s.288{1) 1M.90(1) 898.0(2)5.287(11 104.91(1) 898.2(2)

Note: the numbers in parentheses indicate the enor in the last decimalplace.

this study and we were surprised to find that their refinedunit-cell volumes differ by about 1.3 A'. We initiallythought that this might be due to different instrumentalconditions for the data collections, but we rejected thishypothesis on the basis of data for corundum refinements(multiple scans at different conditions), which indicate nosystematic variation in cell parameters as a function ofscanning condition. However, there does appear to be asmuch as 0.lolo variation in the unit-cell volumes frommultiple scans, which could account for the 1.3-A' dif-ference between TREM 23-5 and TREM 23-10. On thebasis of that, the EMP results, and the TEM results, wedo not feel that the higher cell volume for TREM 23-10indicates something fundamentally different about thesample; it is simply on the high side of statistical varia-tions for volume determinations.

All the unit-cell parameters vary linearly as a functionof composition (Fig. 2). These trends suggest a completesolid solution across the join at the temperatures of for-mation (1093-1148 K). The linear decrease in cell vol-ume with increasing Al content is consistent with otherstudies in the Tr-Ts system (Jenkins, 1988, 1994), andagrees well with the trends observed for the Mg-Tscher-mak substitution in the pyroxene system along the joinof diopside and Ca-Tschermak (Newton et al., 1977) andin the mica system along the phlogopite-eastonite join(Circone et al., l99l). Extrapolation of a least-squaresregression line through the volume data (R'z: 0.96) yieldsa cell volume for the hypothetical Ts end-member (withl0 molo/o Mc) of 888.3 A', which agrees well with thestudy of Jenkins (1994, 888 + 2 L). This cell volumeshould be regarded as a minimum volume for tscher-makite, since there is no correction for the approximatelyl0 molo/o Mg in the M4 site.

Transmission electron microscopy. The synthetic am-phiboles were also examined with a Philips CM2O-STtransmission electron microscope (TEM) operated at 200keV. The samples were gently ground and dispersed inanhydrous alcohol. A drop ofeach suspension was thenapplied to a 3.0-mm Ni grid with a holey carbon supportfilm.

The primary purpose of the TEM examination was todetermine the nature and concentration of defects in theamphiboles. Previous high-resolution TEM (HRTEM)work by Maresch and Czank (1988), Graham et al. (1989),

o{

I

o

o{

q)

l l 1 4

Fig. 3. High-resolution TEM (HRTEM) image of Al-freesample TREM 23-10 taken down [1001. The (020) fringes areclearly visible, running from top to bottom in the image. Planardefects are indicated by arrows. These defects are single-chainfaults and represent narrow slabs ofdiopsideJike material. Tri-ple-chain faults were also observed.

ilar to one poor in defects, even if the composition isnearly identical.

Representative TEM images of the synthetic amphi-boles are presented in Figures 3-5. Most of the imageswere taken down a [u}wlzone, usually [100] or Il0l], sothat the CMFs were parallel to the electron beam andeasily visible. All the synthetic amphiboles were found tocontain CMFs. Figures 3, 4, and 5 are HRTEM imagesof Al-free TREM 23-10, AMPH 25-l (0.8 total Al), andAMPH l4-4 (1.74 total Al), respectively. The most com-monly observed defect types for the amphiboles were sin-gle-chain errors (pyroxene material) and triple-chainsfaults. The most common observation was several planardefects scattered randomly throughout a given crystal. Noexamples of ordered polysomes of the biopyribole series(e.g., clinojimthompsonite or chesterite) were observed.For the vast majority of crystals, the estimated volumepercent of defective material was less than 5o/o. The TEMobservations did indicate, however, that the tremolitecompositions (TREM 23-5, TREM 23-10) had a slightlyhigher frequency and abundance of defects than the Al-


Fig. 4. HRTEM image of sample AMPH 25-l with approx-imately 0.8 total Al cations pfu. This image, also taken down a,shows four individual triple-chain faults embedded in an oth-erwise perfect amphibole structure. This microstructure, con-sisting of several randomly scattered CMFs in a given crystal,was the most commonly observed one in the synthetic tremolite-tschermakite amphiboles.

bearing samples, but they did not approach the defectconcentration in synthetic end-member tremolite sam-ples examined by Ahn et al. (1991). Pawley et al. (1993)also reported a relatively low CMF concentration in theMg-enriched synthetic tremolite (TrrrMcr) used in theircalorimetric study. These authors further indicated thatthe increase of richterite component virtually eliminatedthe CMFs from the amphibole structure. The present re-sults for tremolite-tschermakite amphiboles indicate thatthe Mg-Tschermak substitution does not appear to inhib-it the formation of planar defects, as the richterite sub-stitution does.

Maresch et al. (1990), in a HRTEM study of synthetictremolite, found the defect character to vary with synthe-sis conditions. For synthesis conditions comparable withours, they describe armored diopside relics, up to l0 nm

wide, on which the amphibole appears to nucleate. Theyalso noted that CMFs are minor in both the diopside andtremolite. We have found no such diopside relics in ourTEM examinations. Maresch et al. (1990) also studiedsynthetic anthophyllite-gedrite samples and found thatthere was a significant decrease of CMFs with only a slightincrease of gedrite component into the anthophyllite. Theywere unable to determine whether the cause of this wasdue to the edenite substitution, the Tschermak substitu-tion, or a combination of both. Even though our samplesbelong to the calcic, rather than the ferromagnesian se-ries, the present TEM observations suggest that the eden-ite substitution may be largely responsible for the elimi-nation of CMFs in synthetic orthorhombic amphiboles,rather than the Tschermak substitution.

As mentioned above, one major concern is whether theamphibole stoichiometry is significantly shifted as a re-sult of the presence of defects. Simple calculations follow-ing the method of Veblen and Buseck (1979) show thateven the most defective crystals in our study, with per-haps 100/o defect concentration, do not cause a significantcompositional shift in the amphibole. Furthermore, sincein most crystals there was a combination of single-chainand triple-chain defects, any net compositional shift wouldbe even smaller, since these defect types affect the stoi-chiometry in opposite directions.

Although the EMP results presented above (Fig. lb)show good agreement with the chosen compositional join,the microprobe provides no information about compo-sitional heterogeneity on the submicron scale. Therefore,we have examined several of the samples using analyticalelectron microscopy (AEM) to check the compositions ofnondefective and defective amphibole areas. For theseanalyses, we used a probe size of about 10 nm. The anal-yses were collected with an EDAX Si(Li) detector with afixed take-offangle of 25'. We were unable to detect anysignificant compositional differences between defectiveand nondefective areas using AEM. This is due in part tothe poor precision of AEM, compared with EMP tech-nique, and also to the fact that the small compositionaldifferences that must exist in the defective areas are with-in a region <10 nm, which is the optimal resolution ofthe AEM system.

Drop-solution calorimetry

The calorimetric technique used in this study is basedon the method described by Navrotsky (1977), employ-ing a Calvet-type twin microcalorimeter. Two Pt cruci-bles, each containing 30 g of 2PbO.BrO. solvent, are heldat a constant temperature of 97 5 K in the calorimeter.Relative changes in heat flow within the calorimeter, fromone sample chamber to the other, are measured by twothermopiles, connected in opposition. The samples areintroduced into the solvent by dropping them from roomtemperature (-298 K) through a long, silica-glass tube,hence the term drop solution. The heat effect is a com-bination of the heat of solution of the amphibole in thelead borate and the heat effect from raising the temper-

1 1 1 5

Fig. 5. Bright-field TEM image of sample AMPH l4-4 withabout 1.74 Al pfu. The central part of rhe crystal (indicated byarrows) contains a considerable number of CMFs. This examplerepresents the most extreme defect concentration observed, whichstill represents less than 5-100/o of the total volume of the crystal.Similar defect concentrations were observed, though rarely, inall the synthetic amphiboles, regardless ofthe Al content.

ature of the sample from room temperature to the calori-meter temperature. This experiment differs from a reg-ular-solution experiment, in which the sample isequilibrated at calorimeter temperature before dissolu-tion. The calibration constant, relating microvolts tojoules, was determined by the Pt-drop method (Navrots-ky, 1977). Overall, the methodology is essentially iden-tical to that used by Pawley et al. (1993) for tremolite-richterite amphiboles.

Recent work by Navrotsky et al. (1994) has addressedthe problem of the evolution of volatiles during calori-metric experiments on hydrous phases and carbonates.They found that performing calorimetric experiments un-der a flowing gas atmosphere resulted in a negligible en-thalpy of interaction between the evolved HrO or CO,and the solvent, suggesting that very little, ifany, ofthevolatile remains dissolved in the solvent. Under flowinggas conditions, they were able to obtain reproducible re-sults, with excellent base-line behavior, for several hy-drous silicates and carbonates. The present drop-solutionexperiments were performed in the same manner, underan atmosphere of flowing Ar at 30-50 mllmin.

For TREM 23-5,tlrre sample was prepared for calorim-etry in two ways. The first series of drops employed theuse of small silica-glass capsules, in which 6-12 mg ofamphibole powder was placed. This technique, describedby Burnley et al. (1992), was developed to overcome dis-solution problems that commonly occur with Mg-richcompounds. Of course, using the capsules introduces theneed for an additional calibration constant, and ulti-mately increases the likelihood of additional errors in themeasurements. The results of these first experimentsshowed significant scatter (a little over 2o/o of the total


I 1 1 6 SMELIK ET AL.: CALORIMETRIC STUDY OF SYNTHETIC AMPHIBOLES

TABLE 3. Enthalpies of drop solution for synthetic tremolite-tschermakite amphiboles with 10 mol% Mc

SAMPIE TREM 23-5 TREM 23-10 AMPH 29.1 AMPH 25-1 AMPH 19-5 AMPH 28-2 AMPH 14.4

Al-AH AS'd (kJ/moD

MeanStd Err"

0.0931 .7937.2-973.5'975.5977.O986.4'

963.518.8

0.0946.2948.6952.3952.6956.8959.2966.9

954.65.3

0.4936.4945.6948.2953.4953.7955.9967.2

951.57.2

0.8937.9939.5967.4967.6968.5978.8

959.913.9

1 . 2928.5935.1936.6949.7949.8959.2970.4977.9950.912.4

1 . 6938.4949.2950.2953.8957.1961.3967.1969.5955.8

7.2

1.7493it.9941.1946.7949.0950.2951.4952.8954.3947.4

4.8. These values for TREM 23-5 were obtained using SiO, glass capsules; the remaining TREM 23-5 values and all other values were obtained using

oellets." The standard error equals 2ol\aN.

heat effect) and showed some evidence of slow dissolu-tion; therefore the capsule technique was abandoned.

A second batch of experiments for TREM 23-5, andall the remaining calorimetric experiments, were per-formed using pressed pellets of the amphiboles, whichhave been shown to dissolve well in molten 2PbO.BrO3(Pawley et al., 1993). The pellets were about 3 mm indiameter and weighed from l0 to 20 mg. Each samplewas carefully weighed before it was dropped into the cal-orimeter. The calorimetric peak normally returned to baseline in 30-50 minutes, indicating complete, and rapid,dissolution.

Rrsur,rsThe enthalpies of drop-solution for synthetic tremolite-

tschermakite amphiboles (in kJlmol) are given in Table3. Six to eight enthalpy measurements were made foreach composition. For sample TREM 23-5, the meanvalue for the enthalpy of drop solution, Ma.op *r, includesexperiments performed with capsules and with pellets.The uncertainty for this sample, given as 2 sd of the mean,is rather high, about 2o/o of the total heat effect, and maycontain errors due to problems with the silica-glass cap-sules. We unfortunately exhausted the material before wecould complete a full series of pellet drops for this sam-ple.

For the other tremolite sample and the remaining am-phibole compositions, the uncertainties are in the 0.5-l.5olo range, expressed as 5-13 kJ/mol, which agree wellwith the magrritude of uncertainties encountered in thePt-calibration experiments. The magnitude of the heateffect is in the 950 kJlmol range for these amphiboles,and it becomes immediately obvious by examining themean values of A-Flo.oo *, as a function of Al content thatthere appears to be only a small energetic difference ofabout -10 kJlmol or less between the Al-free and Al-rich compositions (Table 3). This small reduction in theenthalpy ofsolution is consistent with the effect that theTschermak substitution has on the enthalpy of solutionof akermanite (Charlu et al., l98l) and diopside (Newtonet al., 1977). It also agrees with the initial effect of theTschermak substitution in phlogopite (Circone and Na-

vrotsky, 1992), but the nonlinear trend of the mica dataextrapolate to a slightly higher Af/o.., *, for end-membereastonite.

DrscussroxMfu,oo-, for MgHb and Ts

One of the primary goals of this study is to obtain, fromcalorimetric data, values for the enthalpies of formationfrom the elements for the amphibole end-members mag-nesiohornblende and tschermakite at 298.15 K. In orderto extract this information we first need values for Afl*.o *,for these compositions. The mean values for AFld.oo *, fromTable 3 are plotted against the Al content of the amphi-bole in Figure 6. Unlike Graham and Navrotsky (1986)and Pawley et al. (1993), we had only compositions forthe Tr-rich half of the full Tr-Ts join, from Tr to MgHb,because more Al-rich compositions could not be synthe-sized. The lack of calorimetric data for Xr" > 0.5 and thealmost constant Ma,op"or for 0 < Xr" < 0.5 preclude adefinite statement about enthalpies of mixing along theTr-Ts join. The present data do not justify anything morethan a least-squares regression through the data (Fig. 6).

Because ofthe large and variable uncertainties, the lin-ear regression was perfbrrned by weighting each A110."o.",by the inverse ofits uncertainty squared (e.g., Bevingtonand Robinson, 1992, p. 106). The equation for the lineis Af/o,oo,o, : 955.0 (+4.4) - 3.0 (+3.6;41,",. The straightline fit is rather poor, with a low R value (0.563). Becausethe energy change associated with the Mg-Tschermaksubstitution is quite small, we feel the least-squares re-gression line is the best approach for extracting A/1d.op*rvalues for the end-member MgHb and for a tentativeextrapolation to Ts. The question still remains, however,whether the measured enthalpy values for the composi-tions with l0 molo/o Mc can be used to represent true Tr-Ts amphiboles.

To check the possible effect of the l0 molo/o Mc com-ponent, a series of drop-solution experiments was per-formed on a sample of very pure natural tremolite, kindlyprovided by Colin Graham. This tremolite was used byGraham et al. (1984) in a study of H isotopic exchange

between amphibole and HrO. It was also used in a phase-equilibrium study of the tremolite breakdown reactiontremolite:2diopside + 1.5 enstatite + Pquartz + HrOby Welch and Pawley (1991). These authors reported anaverage composition for this tremolite of Ca,nrFeoor-MgrouAloo,SirerO22(OH)r, on the basis of 3l EMP anal-yses. This tremolite was also used by Pawley et al. (1993)in a calorimetric study of synthetic amphiboles along thetremolite-richterite join. The mean value of Aflo.op,or forthis sample was 959.6 + ll.2 kJlmol (8 drops), whichcompares well with the results of Pawley et al. (1993,958.0 kJ/mol). These enthalpy values are in generalagreement with our measured values for the Mc-substi-tuted tremolite shown in Table 3. On the basis of thisresult, we conclude that the small amount of Mg in theM4 site has a minimal effect on the enthalpies of solution,and therefore we may calculate values of A11o,oo

"", for

MgHb and Ts using the equation above. The results are949.0 t 8.5 and 943.0 + l5.l kJlmol, respectively.

A.EIP for MgHb and Ts

The enthalpies of formation for MgHb and Ts may becalculated using a combination of the present calorimet-ric data and data from the literature. The thermodynamiccycle is based on the following equations for MgHb andTs, respectively:

2CaMgSirOu + MgO + AlrO3 + 3SiO, + Mg(OH),: CarMgAlAlSi?Orr(OH), (l)

and

2CaMgSirOu + 2Al2O3 + 2SiO, + Mg(OH),: CarMg.AlrAlrSi6Orr(OH)r. (2)

The data required for the calculations include enthal-pies of formation from the elements and enthalpies ofdrop solution in lead borate at -97 3 K for diopside, peri-clase, corundum, quartz, and brucite. For this paper wehave chosen heat of formation values from two of thepublished internally consistent thermodynamic data sets,those of Holland and Powell (1990) and Berman (1988).

The heat of solution data for the reactant mineralsavailable from the literature are summarized in Table 4.For quartz and diopside, both heats ofdrop solution andheats of solution in 2PbO.BrO, have been reported,whereas for periclase and corundum only heats of solu-tion are available, and for brucite only drop-solution ex-periments have been done. To calculate a heat of dropsolution from solution data, it is only necessary to addthe value of heat content, Hr*, - Hrnr,r, for the mineralof interest. The value for Hr*, - Hr'",, (Z*, : calorimetertemperature) varies depending upon which heat-capacityexpression is chosen. To assess the extent of variability,we have calculated heats ofdrop solution for all five re-actant minerals in Equations I and 2, using different pub-lished values for heats of solution and several heat-ca-pacity expressions (Table 4), and have compared thesewith available measured values for AFld.oo *r.

For all the phases except quartz, the calculated values

l l tT

o

AHa,op"or = 955.0 - 3.0(Ald R=0.563

0 0.4 0.8 r.2 1.6 2

Al.o,

Fig. 6. Plot of AFld,op*r vs. A1,.. for synthetic tremolite-tschermakite amphiboles. A weighted least-squares regressionline is fitted through the drop-solution data. The energy changeassociated with the Mg-Tschermak exchange, reflected by theheat ofdrop-solution data, is very small and negative, suggestedby the slight negative slope of the line. The MgHb end-memberis indicated by the solid square. Two points for synthetic trem-olite (A1,., : 0) are shown, TREM 23-5 and TREM 23-10, thelatter hawing the smaller error.

for A-F/o,oo.o, show close agreement, independent of theheat-capacity expression chosen. The calculated valuesfor quartz are complicated by the enthalpy of the a-0transition at 844 K. A summary of published values forAII*', ranging from 0.335 to l.2l kJ/mol, has been givenby Ghiorso et al. (1979). Recently, Hemingway (1987)provided new heat-capacity data for quartz and reevalu-ated the literature in an attempt to resolve the publisheddisparities in heat-capacity data in the 425-1025 K range.He concluded that all recent tabulations of thermody-namic data for quartz are low by about l-2o/o. In light ofthese results, we have included in Table 4 calculationsusing the heat-capacity expression for quartz given byHemingway (1987) and have incorporated his value forAH..p(0.625 kJ/mol) in all the Hr-, - -Flrn, ', calculations.

Because the variations in AlIfL; *' values are small, theerrors resulting from differences in values from Table 4are small, compared with the elTor in the heats of dropsolution measured for the synthetic amphiboles. We havechosen the following values as the preferred values ofA/Id.oo

"or for the calculation of amphibole heats of for-

mation: diopside and quartz, the values measured by Chaiand Navrotsky (1993), 40.02 + 0.22 and231.4 ! l. l kJ/mol, respectively; periclase and corundum, the calculatedvalues for Ma.op.or using the solution data of Navrotskyet al. (1994, 1986) combined with the heat-capacity ex-pressions of Holland and Powell (1990), 36.51 and 107.63kJ/mol, respectively; brucite, the measured value of Na-vrotsky et al. (1994), 144.1 + 3.6 kJlmol, which wasobtained using the methodology employed for the syn-thetic amphiboles in this study.

Table 5 shows the calculated values of A.F{ for MgHband Ts, using the preferred thermodynamic data for thereactant minerals from Table 4 and the AH! data for the

SMELIKET AL.:CALORIMETRIC STUDY OF SYNTHETIC AMPHIBOLES

980

! szo€1 sooEF sso

\{R noo

930

1 l 1 8 SMELIK ET AL.: CALORIMETRIC STUDY OF SYNTHETIC AMPHIBOLES

TABLE 4, Calorimetric data for silicates near 973 K from the literature

Hrut - Hn"' aF,5.3-aF,s*

Diopside

Periclase

Corundum

149.1 8

31.71

74.84

43.48

69.62

149.97

31 .68

74.91

42.33

70 31

1 50 .14

31.71

74.87

42.36

70.33

236.83235.08235.04234,59

36.6536.5236.5836.51

107.1 I107.39107.6038.3340.3039.2539.97

45.1 3

87.65 + 1.0585.90 + 1.0585.86 + 1.5585.4'l + 1.47

4.94 + 0.334.81 + 0.584.87 + 0.624.80 + 1.00

32.34 + 0.3332.55 + 0.4632.76 + 0.33-5.15 + 0.29-3 .18 + 0 .42-4.23 + 0.21-3 .51 + 0 .18

237.62 237.79235.87 236.04236.00 235 83235.38 235.55

36.62 36.6s36.49 36.5236.55 36.5836.48 36.51

107 .25 107.21107.46 107 .42107.67 107.6337.18 37.2139.15 39.1838.10 38.1338.82 38.85

Brucite

/Votei boldfaced values represent high and low values of AFfiSd for each mineral. ltalicized values are the preferred values used in our calculations.Units of measure are kJ/mol.

-Hea t - capac i t yexp ress ionsused :R :Rob iee ta l . ( 1979 ) ;B :Be rman (1988 ) ;H -P :Ho l l andandPowe l l ( 1990 ) ;H :Hemingway (1987 , f o rqua r t z

only).'* Calorimetric exoeriments done at 986 K.t Drop solution under static air conditions near 973 K.+ Drop solution under static air conditions neat 977 K.$ Drop solution under flowing Ar gas near 975 K.

reactant minerals from Holland and Powell (1990) andBerman (1988).

For magnesiohornblende, the calculated enthalpy offormation at 298.15 K is - 12401.2 kJlmol. For tscher-makite, the calculated enthalpy of formation is - 12527 .7kJlmol. If the enthalpies for these amphiboles are calcu-lated using the minimum and maximum thermodynamicvalues shown in Table 4, a range of about 25 kJ/mol(-12390.9 to -12419.5 and - 12519.3 to -12543.5 kJ/mol, for MgHb and Ts, respectively) is obtained. Thisrange can be largely attributed to variations in the quartzand diopside calorimetric data. The differences that resultusing the Berman (1988) AH! data are insignificant, shift-ing the values by a maximum of about I kJ/mol. It isdifficult to determine accurately the error range associ-ated with these calculations because some studies giveuncertainties as I or 2 sd, whereas others give uncertain-ties as the standard deviation of the mean, and somestudies give no errors at all. By taking the square root ofthe sum of the squares of all the reported errors for allthe data used in the calculations (using the Holland andPowell AH? data), we obtain uncertainties of + 10.6 and+ 16.4 kJlmol for MgHb and Ts, respectively.

Comparison with values obtained fromphase equilibrium

Our preferred values for the enthalpies of formationfor MgHb and Ts are compared with values from theliterature in Table 6. As far as we know. there have been

no previous studies, either phase-equilibrium or calori-metric, aimed at determining thermodynamic propertiesfor the magnesiohornblende composition. Holland andPowell (1990), however, have derived a value of - 12420.3+ 12.7 kJlmol for AfI! of MgHb, for inclusion in theirinternally consistent thermodynamic data base (their Ta-ble 7, reliability level 2). Their value agrees closely withthe value we calculate using the older quartz heat-capac-ity data but is somewhat more negative than our pre-ferred Af19. This diference probably reflects Holland andPowell's (1990) use ofthe older heat-capacity expressionsfor quartz in their least-squares fitting process.

The enthalpy of formation for the end-member tscher-makite composition has been derived in several phase-equilibrium studies, and the published values are com-pared with our preferred value in Table 6. Our resultsagree well with the results ofthe recent phase-equilibriumstudy ofJenkins (1994) and the study ofl,6ger and Ferry(1991), which combines experimental data with naturalmineral parageneses. The At{ reported by Chermak andRimstidt (1989), obtained by summing the estimatedpolyhedral contributions for each oxide and hydroxidecomponent, is significantly more negative than our pre-ferred value, which corroborates that these authors ad-mitted difficulties at predicting thermodynamic proper-ties for calcium aluminum amphiboles and pyroxenesusing the polyhedral approach. The value of Liebermanand Kamber (1991), derived by least-squares fitting to acombination of natural parageneses and experimentaldata, is considerably more positive than our value.

SMELIK ET AL.: CALORIMETRIC STUDY OF SYNTHETIC AMPHIBOLES I 1 1 9

TABLE 5. Calculated enthalpies of formation for MgHb and Tsat 1 bar and 298.15 K using calorimetric data

AmphibolePreferred value'

(kJ/mol)

-'t240't.2-12402-3-12527.7-'t2528.7

Nofei H-P : Holland and Powell (1990); B: Berman (1988)."Calculated using preferred values of AHEBd and AHm'd from

Table 4.

reaction at To and Po (298.15 K, I bar), AS. is the totalentropy change of the reaction, and AV, and K" are thevolume change and equilibrium constant for the reaction,respectively.

Because Reactions 3 and 4 involve only solids and con-serve HrO, and since no Co data are available for MgHb,we will assume as a first approximation that Aq : 0 forboth reactions and that AZ, (solids) is constant over theP-T range (l-8 kbar and 873-1073 K) of Jenkins's re-versal experiments. These simplifications allow Equation5 to be rewritten as

AG, : 0 : AGg - A^SKr - T) + LV,(P - Po)+ RZln K". (6)

Noting that AGp : AH? - Z.ASP and substituting intoExpression 6, we arrive at

AG,:0 : LH? - rA,S9 + LV,(P - &) + Rrh K". (7)

For Reactions 3 and 4, the entropy term in Equation 6,AS!, can be expanded as shown in Equation 8:

ASf : AS!,n.* - Sfl,r", (8)

where AS!*.." is Sg, + ,Sl" - SBt - ,S&,, for Reaction 3and Sg. + ,S8" - S$. for Reaction 4. Substituting this intoEquation 7 and rearranging gives the following expressionfor the entropy of MgHb at 298.15 K and I bar:

Sfln ru

_ _ At19 - IAS8*"." + A4(P - %) + Rin ln K, /o\I '

\ a t

Using Equation 9, we can test various site-mixing modelsfor MgHb to see which, if any, give constant values ofSflorro. The equilibrium constant, K,, for Reaction 3 canbe expanded as

K"- -

and for Reaction 4 as

(afli'n)(aol;")(10)

(a.i&tiiXariXag%,)

Taate 4.-Continued

H ReferenceData base

Navrotsky and Coons (1976)".Newton et al. (1977)Kiseleva et al (1979)Navrotsky et al. (1980)..Chai and Navrotsky (1993)Charlu et al. (1975)Navrotsky and Coons (1976)-'Davies and Navrotsky (1981)--Navrotsky et al. (1994)Charlu et al. (1975)Newton et al (1980)Navrotsky et al. (1986)Charlu et al. (1975)Navrotsky and Coons (1976)-.Kiseleva et al. (1979)Akaogi and Navrotsky (1984)Chai and Navrotsky (1993)Kiseleva and Ogorodova (1984)tClemens et al. (1987)tCircone and Navrotsky (1992)+Navrotsky et al. (1994)$

MgHb

Ts

H.PB

H-PB237.40 ! 1.10

39.9841 .9540.9041.62

40.02 + 0.221 1 1 . 5 0 + 1 . 5 9120.04 + 2.22117.76 + 1.54144.10 + 3.60

Site-rnixing models and entropy calculations

Jenkins (1994) examined experimentally two divariantreactions involving Al substitution in tremolitic amphi-bole to derive activity-composition relationships for am-phiboles in the Tr-Ts pseudobinary system, as well asstandard-state thermodynamic properties for end-mem-ber tschermakite. To complement Jenkins's study, wehave used our AIf data, determined independently fromcalorimetry, combined with the experimental reversal dataof Jenkins (1994), to derive the standard entropy forMgHb and test various site-mixing models for aluminumtremolites.

We have reformulated the divariant Reactions 2 and 3studied by Jenkins (1994) in terms of the MgHb com-ponent, rather than the Ts component, to obtain the tworeactions

Ca,MgAlrSi,Orr(OH), + CaMgSirO6 + SiO,amphibole (MgHb) diopside quartz

: CarMgrSirOrr(OH), + CaAlrSirO, (3)mphibol€ (Tr) plagidare

CarMgoAlrSirOrr(OH), * MgrSiOnamphibole (MgHb) fonterite

: CarMgrSirOrr(OH), + MgAlrO4. @)amphibole (Tr) spinel

Following the methodology of Jenkins (199a), at equilib-rium we may write the following expression for the Gibbsfree energy ofreaction at P and T:

AG,: o : ^G? - l" ^t. o, * J'" ̂ vdP

+ R?" ln K. (5)

where AG! is the change in the Gibbs free energy of the

( l l )

Electron microprobe analyses of plagioclase, diopside,quartz, forsterite, and spinel by Jenkins (1994) indicatedthat these phases are essentially pure. By accepting unit

l 120 SMELIKET AL.: CALORIMETRIC STUDY OF SYNTHETIC AMPHIBOLES

TABLE 6. Comparison with AHP and I values from the literature for MgHb and Ts

References (methods)'s0

J/(mol.K)AHto

kJ/mol

-12401.2 + ',t0.6-12420.29 + 12.70

-12527.7 + '16.4-12534.4 + 13.0

- 1 2578.949-12449.O94

-12606.0 + 36.0-12534.67

MgHb575.3 + 3.9

551.0Ts

542.5 + 12.7514.687542.72

538.6

This study'- (C)Holland and Powell 0990) (LS)

This study'- (C)Jenkins (1 994)'- (PE)Mader and Berman (1992) (PE)Lieberman and Kamber (1991)t (LS, PE)Chermak and Rimstidt (1989)+ [rH)L6ger and Ferry (1991) (PE)

- Methods: C : calorimetry; LS : least squares; PE : phase equilibrium; TH : theoretical.'- Using Holland and Powell (1990) data base; entropy is average of two-site and four-site activity-composition models.f Using Berman (1988) data base.+ Calculated based on summation of polyhedral contributions wilh all OH as Mg(OH),.

activity for the pure solids at P and Z of interest as thestandard state, Equations l0 and I I can be expressed interms of amphibole activities only:

The strategy here is to solve Equation 9 using the ex-perimental reversal data for Reactions 3 and 4 of Jenkins(1994) to see which of the four site-mixing activity-com-position models gives the most constant value for S$*0.We have used the value for the AHP for MgHb from Table6 and enthalpy of formation values for the other phasesfrom Holland and Powell (1990), except for tremolite,which came from Jenkins et al. (1991), to calculate Alf.We used the entropies for diopside, tremolite, anorthite,qlrarIz, forsterite, and spinel from Holland and Powell(1990) to calculate AS3**. Using the cell volume forAMPH 28-2 (Table 2) and the planar extrapolation meth-od described by Jenkins (1994), we obtain a volume of270.8 + 0.1 cm3/mol (899.16 + 0.33 A,; for MgHb. Us-ing the molar volumes from the Holland and Powell(1990) data set for all other phases, we calculate for LV,a value of + 13.8 cm3 (+ 1.38 J/bar) for Reaction 3 and- 1.98 cm3 (-0. 198 J/bar) for Reaction 4.

The results for the four site-mixing models are sum-marized in Table 7. The uncertainties reflect the propa-gation of all errors entering into the calculation of Sfln*6by Equation 9. The mean values represent the weightedaverage Sflrro t I sd of the mean. Although all activity-composition models examined provide a satisfactory rangeofconstant entropy values, the two-site coupled and four-site coupled mixing models produce the least amount ofscatter (smallest standard deviations). These two mixingmodels also yield nearly identical mean values for Sfloo*.The average value of Sflrrro from these two mixing modelsis listed in Table 6.

Coucr,usroxsThese results represent the first data on heat of for-

mation from calorimetric measurements for calcic am-phiboles approximately along the join between tremoliteand tschermakite and allow the estimation of thermo-dynamic properties for the important amphibole end-member compositions, magnesiohornblende and tscher-makite. The calorimetric data indicate that the energeticchange associated with the Mg-Tschermak substitution(l61Mg,t4rsi + tolfil,tnlfil) is small. A weighted least-squares

(r2)

Four mixing models will be considered here: (l) coupledsubstitution of octahedral and tetrahedral Al onto onlyone M2 site (one-site coupled model), (2) random mixingof Al on both M2 sites but not on the tetrahedral sites(two-site coupled model), (3) random mixing of Al andSi on the four Tl sites but not the M2 sites (four-sitecoupled model), and (4) complete random mixing of Alon the two M2 sites and Al and Si on the four Tl sites(random mixing model). These models represent ideal-ized site-mixing configurations that serve as a startingpoint for the following calculations and are not meant toimply long-range ordering arrangements that would beinconsistent with amphibole crystallography or crystalchemistry. Ideal activity-composition models for thesemixing schemes have been derived according to the pro-cedure of Price (1985). The resultant expressions for rK"for the four mixing models are given below, with X de-fined as X{,'z (mole fraction of Al on the M2 site):

(l) one-site coupled model:

,, (t - 2x)n^ : 2x

(2) two-site coupled model:

(t - Y1zn " :

4 x . t _ x \

(3) four-site coupled model:

( t - %),n": ,t"tr-t<t -tS

(4) random mixing model:

,. aTo"I\" : ---------

s MsHb


TABLE 7. Calculated values of S$n"o values using different activity-composition models

IL2L

PTXdata point' One-site coupled model Two-site coupled model Four-site coupled model Random mixing model

'|2345o7

571.44 + 14.91570.55 + 13.65570.97 + 12.34572.56 + 13.85572.02 + 12-56573.38 + 14.12572.02 + 13.13

575.37 + 12.18570.62 + 12.61571.53 + 12.33570.25 + 12.78

571.89 + 3.94

Reaction 3576.68 + 14.90575.68 + 13.62575.91 + 12.30577.21 + 13.82576.52 + 12.53577.44 + 14.11575.97 + 13.00

Reaction 4574.48 + 12.05572.47 + 12.60S72.Ag + 12.90572.32 + 12.75

Mean values575.09 + 3.92

577.U + 14.90576.79 + 13.60576.95 + 12.29578.13 + 13.81577.37 + 12.52578 .11 + 14 .11576.61 + 12.96

573.89 + 12.04572.46 + 12.59572.68 + 12.29572.37 + 12.74

575.57 + 3.92

565.92 + 15.19566.11 + 14.16568.48 + 12.73572.13 + 14.07572.39 + 12.82575.59 + 14.17574.02 + 13.58

580.35 + 12.13575.81 + 12.63576.85 + 12.35575.29 + 12.82

573.38 + 3.99

I9

1 01 1

Nofe.' units of measure for models are J/(mol. K).. PTX data points are reversal experiments of Jenkins (1 994).

linear regression through the calorimetric data suggests aslight decrease in A.F155;*, with increasing Al content. Thistrend is also observed in other silicate systems in whichthe same substitution is operating, such as initially alongthe phlogopite-eastonite join (Circone and Navrotsky,1992) and the join of diopside and Ca-Tschermak pyrox-ene (Newton et al., 1977). Because of the limited rangeof compositions available for this study (only the Al-poorhalf of the join), the data for the Al-rich end-membertschermakite were determined by linear extrapolation.

The thermodynamic data calculated from the mea-surements of the heat of drop solution and other calori-metric data from the literature agree well with previousvalues derived from phase-equilibrium studies. Our finalpreferred values for A/{ for MgHb and Ts are -12401.2+ 10.6 and -12527.7 + 16.4 kJ/mol, respectively. Fourideal activity-composition relationships were tested usingthe calorimetric data. Although all four models gave sat-isfactory results, the two-site and four-site coupled mod-els gave the least scattered values for Shs'b, 575.1 + 3.9and 575.6 + 3.9, respectively. These values are signifi-cantly higher than the one published value in Hollandand Powell's (1990) data base.

Although these results are insufficient for the precisecalculation of phase equilibria involving aluminous calcicamphiboles, they represent another step in the persistenteffort to understand the activity-composition relation-ships in complex amphibole solid solutions. It is hopedthat the results from this paper can be incorporated intothe ever-evolving internally consistent thermodynamicdata sets and provide some critical data needed for ther-modynamic modeling of mineral reactions involvingtschermakitic calcic amphiboles.

AcxNowr,nncMENTs

E.A.S. would like to thank Ed Vicenzi, for help with the microprobeanalyses, and Vincent lamberti and Peter Heaney, for helpful discussions.Constructive reviews by D. Perkins and H.T. Haselton, Jr. improved themanuscript. The arnphibole synthesis work was carried out at the exper-

imental petrology laboratory in the Department of Geological Sciences at

the State University ofNew York at Binghamton, supported by NSF grant

EAR-9104266 to D.M.J. The TEM and EMP characterization was done

at the electron microbeam facility of the Princeton Materials Institute,

and the calorimetry was performed in the calorimetry laboratory in the

Department of Geological and Geophysical Sciences, both at Princeton

University. This research was supported by DOE grant DE-FG02-

85ER13437 to A.N.

Rmnnrxcns crrEDAhn, J.H., Cho, M., Jenkins, D.M., and Buseck, P.R. (1991) Structural

defects in synthetic tremolitic amphiboles. American Mineralogist, 76,l 8 l l - 1823 .

Akaogi, M., and Navrotsky, A. (l 984) The quartz-coesite-stishovite trans-formations: New calorimetric measurements and calculation of phase

diagrams. Physics ofthe Earth and Planetary Interiors, 36,124-134.Berman, R.G. (1988) Internally-consistent thermodynamic data for rnin-

erals in the system Na'O-KrO-CaO-MgO-FeO-Fe'O.-AlrOr-SiOr-TiO'-H,O-CO'. Journal of Petrology, 29, 445-522.

Bevington, P.R., and Robinson, D.K. (1992) Data reduction and erroranalysis for the physical sciences (2nd edition), 328 p. McGraw-Hill,New York.

Bish, D.L. (1988) Quantitative phase analysis using the Rietveld method.

Journal ofApplied Crystallography, 21, 86-91.Bish, D.L., and Post, J.E. (1993) Quantitative mineralogical analysis using

the Rietveld full-pattern fitting method. American Mineralogist, 78,932-940.

Burnley, P., Navrotsky, A., and Bose, K. (1992) Heat of formation ofsynthetic and natural talc by drop solution calorimetry: A test ofa newtechnique (abs.). Eos, 73, 522.

Chai, L., and Navrotsky, A. (1993) Thermochemistry ofcarbonate-py-roxene equilibria. Contributions to Mineralogy and Petrology, 1 14, 139-147.

Charlu, T.V., Newton, R.C., and Kleppa, O.J. (1975) Enthalpies of for-mation at 970 K of compounds in the system MgO-AlrOr-SiO' fromhigh temperature solution calorimetry. Geochimica et CosmochimicaActa.39. 1487-1497.

-(1981) Thermochemistry of synthetic CarAlrSiO, (gehlenite)-

Ca,MgSi,O, (ikermanite) melilites. Geochimica et Cosmochimica Acta,45. 1609-1617.

Chermak, J.A., and Rimstidt, J.D. (1989) Estimating the thermodynamicproperties (AG? and Alf) ofsilicate minerals at 298 K from the sumof polyhedral contributions. American Mineralogist, 7 4, lO23- lO3 l.

Cho, M., and Ernst, W.G. (1991) An experimental determination of calcicamphibole solid solution along the join rremolite-tschermakite. Amer-ican Mineralogist, 76, 985-1001.

tr22

Circone, S., and Navrotsky, A. (1992) Substitution oft6lAl in phlogopite:High-temperature solution calorimetry, heat capacities, and thermo-dynamic properties of the phlogopite-eastonite join. American Miner-aloglst, 7 7, I I 9 l-l 205.

Circone, S., Navrotsky, A., Kirkpatrick, R.J., and Graham, C.M. (1991)Substitution of t6alAl in phlogopite: Mica characterization, unit-cellvariation, ,?Al and -Si MAS-NMR spectroscopy, and Al-Si distribu-tion in the tetrahedral sheet. American Mineralogist, 76, 1485-1501.

Clernens, J.D., Circone, S., Navrotsky, A., and McMillan, P.F. (1987)Phlogopite: High temperature solution calorimetry, thermodynamicproperties, Al-Si and stacking disorder, and phase equilibria. Geochim-ica et Cosmochimica Acta, 51,2569-2578.

Davies, P.K., and Navrotsky, A. (1981) Thermodynamics of solid solu-tion formation in NiO-MgO and NiO-ZnO. Joumal of Solid StateChemistry, 38,264-276.

Ghiorso, M.S., Carmichael, I.S.E., and Moret, L.K. (1979) Inverted high-temperature quartz: Unit cell parameters and properties ofthe d-p in-version. Contributions to Mineralogy and Petrology, 68,307-323.

Graham, C.M., and Navrotsky, A. (1986) Thermochemistry of the trem-olite-edenite using fluorine analogues, and applications to amphibole-plagioclase-quartz equilibria. Contributions to Mineralogy and Petrol-ogy, 93, l8-32.

Graham, C.M., Harmon, R.S., and Sheppard, S.M.F. (1984) Experimen-tal hydrogen isotope studies: Hydrogen isotop€ exchange between am-phibole and water. American Mineralogist, 69, 128-138

Graham, C.M., Maresch, W.V., Welch, M.D., and Pawley, A.R. (1989)Experimental studies on amphiboles: A review with thermodynamicperspectives. European Journal of Mineralogy, l, 5 3 5-5 5 5.

Helgeson, H.C , Delany, J.M., Nesbitt, H.W., and Bird, D K. (1978) Sum-mary and critique of the thermodynamic properties of rock-formingminerals. American Journal of Science. 278A. l-229.

Hemingway, B.S. (1987) Quartz: }{eat capacities from 340 to 1000 K andrevised values for the thermodynamic properties. American Mineral-ogist,72, 273-279-

Holland, T.J.B. (l 980) The reaction albite : jadeite + quartz dererminedexperimentally in the range 600-1200 qC. American Mineralogist, 65,t29-r34.

Holland, T.J.B., and Powell, R. (1990) An enlarged and updated inter-nally consistent thermodynamic dataset with uncertainties and corre-lations: The system KrO-Na,O-CaO-MgO-MnO-FeO-Fe,O,-AlrOr-TiOr-SiO,-H,-Or. Journal of Metamorphic Geology, 8, 89-124.

Jenkins, D.M. (1987) Synthesis and characterization oftrernolite in thesystem HrO-CaO-MgO-SiOr. American Mineralogist, 7 2, 7O7 -7 | 5.

-(1988) Experimental study of the join tremolite-tschermakite: Areinvestigation. Contributions to Mineralogy and Petrology, 99, 392-400.

-(1994) Experimental reversal of the aluminum content in tremo-litic amphiboles in the system HrO-CaO-MgO-AlrOr-SiO,. AmericanJournaf of Sciencn. 294. 593-620.

Jenkins, D.M, Holland, T.J.B., and Clare, A.K. (1991) Experimental de-termination of the pressure-temperature stability field and therrno-chemical properties of synthetic tremolite. American Mineralogist, 76,4s8-469

Johannes, W., Bell, P.M., Mao, H.K., Boettcher, A.L., Chipman, D.W.,Hays, J.F., Newton, R.C., and Seifert, F. (1971) An interlaborarorycomparison ofpiston-cylinder pressure and calibration using the albitebreakdown reaction Contributions to Mineralogy and Petrology, 32,24-38.

Kiseleva, I.A., and Ogorodova, LP. (1984) High-temperature solutioncalorimetry for determining the enthalpies of formation for hydroxyl-containing minerals such as talc and tremolite. Geochemistry Inter-national. 21. 36-46

Kiseleva, I.A., Ogorodova, L.P., Topor, N.D., and Chigareva, O.G. (1979)A thermochemical study of the CaO-MgO-SiO, system. GeochemistryInternational. 16. 122-134.

L6ger, A., and Ferry, J.M. (1991) Highly atuminous hornblende from low-pressure metacarbonates and a preliminary thermodynamic model forthe Al content of calcic amphibole. American Mineralogist, 76, lOO2-l 0 l 7 .


Lieberman, J.E., and Kamber, B. (1991) Consistent thermodynamic prop-erties for tschermakite from experimental and empirical constraints(abs.). Eos, 72, 564.

Miider, U.K., and Berman, R.G. (1992) Arnphibole thermobarometry: Athermodynamic approach. Current research: Part E. Geological Surveyof Canada, Paper 92-lE, 393-400.

Maresch, W.V, and Czank, M. (1988) Crystal chemistry, growth kineticsand phase relationships of structurally disordered (Mn'+, Mg) amphi-boles. Fonschritte der Mineralogie, 66, 69-121.

Maresch, W.V., Czank, M., and Seifert, F. (1990) Realbau (defect struc-ture) in synthetic tremolite and anthophyllite-gedrite solid solutions.Terra Abstracts. 2. 24.

Navrotsky, A,. (1977) Progress and new directions in high temperaturecalorimetry Physics and Chemistry of Minerals, 2, 89-104.

Navrotsky, A., and Coons, W.E. (1976) Thermochemistry of some py-roxenes and related compounds. Geochimica et Cosmochimica Acta,40, l28l-1288.

Navrotsky, A., Hon, R., Weill, D.F., and Henry, D.J. (1980) Tbermo-chernistry of glasses and liquids in the systems CaMgSi,06-CaAlrsirOs-NaAlSi.O., SiO,-CaAlrSi,O.-NaAlSiOr, and SiO,-AlrO,-CaO-Na,O.Geochimica et Cosmochimica Act^. 4 4. | 409 - | 423.

Nawotsky, A., Wechsler, B.A,. Geisinger, K., and Seifert, F. (1986) Ther-mochemistry of MgAl,Oo-AlqOo defect spinels. Journal of the Ameri-can Ceramic Society, 69, 418-422.

Nawotsky, A., Rapp, R.P., Smelik, E.A., Burnley, P., Circone, S., Chai,L., Bose, K., and Westrich, H.R. (1994) The behavior of HrO and COtin high-temperature lead borate solution calorimetry of volatile-bearingphases. American Mineralogist, 79, 1099-1 109.

Nemon, R.C, Charlu, T.V., and Kleppa, O.J. (1977) Thermochemistryof high pressure garnets and clinopyroxenes in the system CaO-MgO-Al,O3-SiOr. Geochimica et Cosmochimica Acl^, 41, 369-377.

-(1980) Thermochemistry of the high structural state plagioclases.Geochimica et Cosmochimica Acta, 44, 933-941.

Pawley, A.R., Graham, C.M., and Navrotsky, A. (1993) Tremolite-rich-terite arnphiboles: Synthesis, compositional and structural character-ization, and thermochemistry. American Mineralogist, 7 8, 23-35.

Pouchou, J.L., and Pichoir, F. (1 98 S)'PAP' g(pZ) procedure for improvedquantitative microanalysis. Microbeam Analysis, 1985, 104-106.

Price, J.G. (1985) Ideal site mixing in solid solutions, with an applicationto two-feldspar geothemometry American Mineralogist, 70, 69 6-7 0 l.

Raudsepp, M., Turncock, A.C., and Hamhome, F.C. (1991) Amphibolesynthesis at low pressure: What grows and what doesn't. EuropeanJournal of Mineralogy, 3, 983-1004.

Robie, R.A., Hemingway, B.S., and Fischer, J.R. (1979) Thermodynamicproperties ofminerals and related substances af 298.15 K and I bar(105 Pascals) pressure and at higher temperatures. U.S. Geological Sur-vey, Bulletin 1452,456 p.

Smelik, E.A., and Veblen, D.R. (1993) A transmission and analltrcalelectron microscope study of exsolution microstructures and mecha-nisms in the orthoamphiboles anthophyllite and gedrite. AmericanMineralogist, 7 8, 5l l-532.

Smelik, E.A., Nyman, M.W., and Veblen, D.R. (1991) Pervasive exso-lution within the calcic amphibole series: TEM evidence for a misci-bility gap between actinolite and hornblende in natural samples. Amer-ican Mineralogist, 7 6, | 184-1204.

Veblen, D.R., and Buseck, P.R. (1979) Chain-width order and disorderin biopyriboles. American Mineralogist, 64, 687 -7 O0.

Welch, MD., and Pawley, A.R. (1991) Tremolite: New enthalpy andentropy data from a phase equilibrium study of the reaction tremolite: 2 diopside + 1.5 orthoenstatite + B-quartz + HrO. American Min-eralogist, 7 6, 193l-1939.

Young, R.A. (1993) Introduction to the Rietveld method. In R.A. Young,Ed., The Rietveld method, p.39-42. Oxford University Press, Oxford,U.K.

M,rNuscnrP.r RECEIVED APrIr 4, 1994Mem;scnrrr AccEp,rED Jvrv 25, 1994

a calorimetric study of synthetic amphiboles along the ... · a calorimetric study of synthetic...

Documents