American Mineralogist, Volume 76, pages 1589-1596, 1991

Thermodynamic properties of anthophyllite and talc:Corrections and discussion of calorimetric data

Bnucn S. HnrvrrNcwl,vU.S. Geological Survey, Reston, Virginia 22092,U.5.4.

Ansrnlcr

Arithmetic errors in calculating heat capacity values (Krupka, 1984; Krupka et al.,1985a) for anthophyllite and several errors in the Hess cycles utilized to derive enthalpiesof formation of anthophyllite and talc are identified, and revised values are reported. Thecorrected values are: for anthophyllite, ^Sns - ,$ : 554.2 + 3 J/mol'K and A1Hln, :-12163 + 8kJ/mol.KforthecompositionMgurFeo,SirOrr(OH)ri andSlnr:534.5 + 3.5J/mol'K, ArHorrr: -12070 a 8 kJ/mol and ArGln": -11343 + 8 kJlmol for the com-position MgrSirOrr(OH)r; and for talc [MgrSioO,o(OH)r], \H9n": -5900 + 2 kJ/mol andArG2nr: -5520 + 2 kJ/mol. Corrected heat capacity equations for anthophyllite are: forMgurFeorSirOrr(OH),

cg :1285 .774 + 0 .0116487 - 8602 .0 tT 05 - 111780002 - '

and for MgrSirOrr(OH),

cg: 1259.938 + 0.0188847 - 8127.407-05 - 11642700T 'z

in J/mol.K, with the valid range 298.15 to 1200 K. The revised calorimetric values arein good agreement with evaluations by linear and mathematical programming (Day et al.,1985; Berman et al., 1986; Berman, 1988), and they support the values determined bysolubility studies (Hemley eI al., 1977 a, 1977b; Bricker et al., 1973).

Irvrnooucrrou

Knowledge of the thermodynamic properties and henceof the stability of hydrous minerals is important for mod-els of the origin and evolution of metamorphic HrO aswell as for processes that may involve metamorphic HrO(e.g., the formation of ore bodies and the elastic or non-elastic response of rocks under stress). Enthalpies andGibbs free energies of formation for anthophyllite andtalc are in question because results derived from aqueoussolubility studies (Hemley et al., 1977a, 1977b; Brickeret al., 19731' Hostetler et al., l97l) differ from those ob-tained from aqueous HF (Weeks, 1956; Benninglon, 1956;Barany, 1963) and molten salt (Kiseleva and Ogorodova,1984) calorimetry by about 16-23 kJlmol. Day et al.(1985), Berman et al. (1986), and Berman (1988) haveevaluated thermodynamic properties for these phases us-ing the techniques of linear and mathematical program-ming and combining the calorimetric and solubility datawith available phase equilibrium data. The optimized re-sults selected in these studies agree with the values ob-tained from solubility studies (e.g., Hemley et al., 1977a,1977b). However, Day et al. (1985) have shown that theexperimental data can be fitted in such a way that goodagreement with the calorimetric data results, albeit withsomewhat larger residuals. Because solubility and phaseequilibrium experiments measure properties through sur-face reactions and calorimetric experiments measure

properties of the bulk sample, disparate data must beexamined to determine whether some measurements arein error or whether a process other than that thought tobe measured has occurred. Therefore, a reevaluation ofthe calorimetric Hess cycles is indicated.

The entropy and molar heat capacities for anthophyl-lite of the composition MgrFe',Si8Orr(OH), (Krupka,1984; Krupka et al., 1985a, 1985b) are in error. This factwas discovered simultaneously and independently byBernard Evans of the University of Washington and theauthor. These data are recalculated in this studv.

ExpnnrlrnNTAr. HEAT cApAcrrrEs AND ENTRopTESFOR ANTHOPHYLLITE

Molar values for the heat capacities and entropies foranthophyllite of the composition MgurFeorSirOrr(OH),(Krupka, 1984; Krupka et al., 1985a, 1985b) are in errorbecause of arithmetic errors in the input data set pro-cessed with the standard computer software used in ourlaboratory. The specific heats for anthophyllite reportedin these studies are correct. It is important to emphasizethat the errors were associated only with the data set andnot with the computer software. Thus no other data setssuffer the same errors. Nor do the errors result from prob-lems in the measurement of heat capacities, instrumentcalibration, or other experimental detail. Rather, the er-rors involve the use of incorrect molar masses and an

0003-004x/9 l/09 l0-l 589$02.00 I 589

I 590

incorrect value for the mass of FerOr. Preliminary heatcapacities were calculated by Krupka during initial datacollection using a molar mass of 780.82 g for the naturalanthophyllite sample. When corrections were made fordeviations of the sample composition from Mgu.FeorSir-Orr(OH)2, the specific heats were recalculated using a val-ue for the molar mass of 801 .638 g instead of 780.82 g.The specific heats calculated in this manner were toosmall, and this calculation represents the primary error.Other errors of lesser consequence were: (l) summationof the correction terms yielded a value of 801.506 g ratherthan 801.638 g, (2) the molar mass was reported to be802.900 g when the calculations were based upon a valueof 801.506 g, and (3) the correction for FerO, should havebeen 5.0 g rather than the 5.7 g used. The heat capaatyand entropy of the natural anthophyllite sample at298.15K are 644.2 and 540.7 J/mol'K, respectively, based upona molar mass of 780.82 g and integration of the datagiven in Table 3 of Krupka et al. (1985a). The specificheat and entropy are 0.8250 and 0.6925 J/g'K.

The heat capacity and entropy of MgurFeorSisOrr(OH),at 298.15 K may be calculated using 881.858 g for themass of I mol of the natural sample (anthophyllite plusimpurities) and subtracting the contribution of the fol-lowing components (correction terms in moles of sub-stance): 0.2325 SiOr, 0.0477 AlrO3, 0.0318 FerO., 0.0564NaAlSi.Or, 0.0124 KAlSi,O8, 0.0980 CaSiO., 0.1461Mg,SiOo, 0.0075 MnrSiOo, 0.0070 NirSiO4, 0.0018 PrO5,0.0018 CO,, 0.0009 TiOr, 0.0062 CrrOr, and 0.0293 HrO.For Mgu.Feo?Si8Orr(OH), of mass 802.906 g, the heatcapacity and entropy are 665.3 + 1.5 and 554.2 + 3.0J/mol.K, respectively. The corrected values are signifi-cantly larger Ihan 647.2 + 1.5 and 538.9 + 3.0 J/mol.Kreported by Krupka (1984) and Krupka et al. (1985a) forthe heat capacity and entropy, respectively. Heat capacityand entropy data for the component phases are somewhatdifferent from those used by Krupka et al. (1985a) andwere taken from Robie et al. (1979, 1982a, 1984) andKrupka et al. ( I 98 5a). Data for albite and microcline wereused in these calculations rather than the data for theglasses of these compositions that were used by Krupka(1984) and Krupka et al. (1985a).

The heat capacity and entropy of magnesio-antho-phyllite at 298.15 K were calculated in a similar mannerusing926.269 g as the mass of I mol of the natural sam-ple and subtracting the contributions from the followingcomponents (in moles): 0.1409 amosite (Bennington etal., 1978), 0.0447 AlrO3, 0.0253 Fe,Or, 0.0600 NaAl-Si3O8, 0.01 32 KAlSi3Oo, 0. 1040 CaSiOr, 0.0021 MnrSiOo,0.0075 Nirsio4, 0.0019 p'os, 0.0019 cor, 0.0009 Tior,and 0.0066 CrrO.. The amosite sample measured by Ben-nington et al. (1978) was a mixed Mg-Fe crystal. Thecorrection listed above is based upon removal of all Fer*and some Fe3*, Al, and Mn using the amosite data. Theuse of the amosite data also results in removal of Mg, Si,and OH in excess. Therefore, 0.0454 Mg(OH)r, 0.0844MgrSiOo, and 0.2972 SiO, must be added. The use ofthermodynamic data for amosite rather than that of a

HEMINGWAY: THERMODYNAMIC DATA FOR ANTHOPHYLLITE AND TALC

phase like fayalite is justified on the basis of the similarmagnetic contribution of Fe at low temperatures in an-thophyllite and amosite. For Mg,SirOrr(OH), of mass780.82 g, the heat capacity and entropy at 298.15 K are662.4 + 2.0 and 534.3 + 3.5 J/mol.K, respectively. Thecorrection was also calculated using the data for grunerite(Anovitz et al., 1988) rather than that for amosite. Thiscalculation yielded 534.7 + 3.5 J/mol.K for the entropyof magnesio-anthophyllite, in good agreement with thefirst calculation. The average of these values is recom-mended.

The results of the latter calculations are of greatest im-portance because the entropy and heat capacity of mag-nesio-anthophyllite are the primary thermodynamic val-ues that have been utilized. The change from the heatcapacity given by Krupka et al. (1985a) to the heat ca-pacity at 298.15 K as corrected in this study is negligible.The entropy at 298.15 K is 0.50/o lower than that calcu-lated by Krupka (1984) and Krupka et al. (1985a) but iswithin the reported uncertainty. The difference in thesevalues largely reflects the difference in the correction pro-cedure for Fe used in each study.

The entropy reported here for magnesio-anthophylliteis in excellent agreement with the value obtained byBerman et al. (1986) in their derivation of an internallyconsistent thermodynamic data set for the system MgO-SiO,-H,O. However, the reactions limiting anthophyllitestability show significant curvature in P-Z space, and theexperimental phase reversal data exhibit significant scat-ter and spread. Thus, entropies (which are slope terms inthe calculations) calculated in this manner are subject tolarge uncertainties. Therefore, the agreement may onlybe viewed as encouraging.

Specific heats for anthophyllite measured by differen-tial scanning calorimetry (Krupka, 1984; Krupka et al.,1985b) have been corrected using the procedures andcomponents discussed above. Additional values have beenestimated to 1200 K. The heat capacities of the two cor-rected compositions were separately fitted by least squaresto an equation ofthe form suggested by Haas and Fisher(1976). For each fit, the average deviation between thedata and the equation was 0.4010. For the temperatureinterval 298.15 to 1200 K, the heat capacities (in J/mol.K) may be calculated from

cop: 1285.774 + 0.0116487 - 8602.01r-05- l l 1780007-,

for MgrFeorSisOrr(OH), and from

Cg: 1259.938 + 0.0188842 - 8127.407-05- 116427007-2

for MgrSirOrdOH)r. As discussed earlier, the heat capac-ities for the composition Mg rFeo rSisOrr(OH), differ sig-nificantly from those given in Krupka (1984) and Krupkaet al. (1985b), being 2.8o/o larger at 298.15 K and 2.4o/olarger at 700 K. Heat capacities calculated for Mgr-Si8Orr(OH), differ by less than 0.2010.

HEMINGWAY: THERMODYNAMIC DATA FOR ANTHOPHYLLITE AND TALC

TABLE 1. Thermodynamic properties for anthophyllite and talc from selected sources

1 5 9 1

ReterenceL,H%ekJ/mol

A,Gg". Sgr"kJ/mol J/mol.K

Weeks (1956)Naumov et al. (1974)Helgeson et al. (1978)Hemley et al (1977b)Day er ar. (1985)

Berman et al. (1986)Berman (1988)This study

Bennington (1 956)Barany (1963)Hostetler et al. (1971)Bricker et al. (1973)

Naumov et al. (1974)Hemfey et al. (1977a1Helgeson et al. (1978)Robie et al. (1979)Kiseleva and Ogorodova (1984)Day et al. (1985)

Berman et al. (1986)Berman (1988)This study

Anthophyllite-11358.3.-1 1364.6-11361 .4-11344.2-11345.0-11365.7-11345.9-11342.6-113|i}.3

Talc-5501.8'-55/t3.0-5518.7-5521.6-5528.7-5543.4-5524.6-5523.7-5536.0-5542.4'-5518.2-5529.2- 5519.6-5517.7-5520.0

- 12086.8-12090.9- 1 2086.5-12072.7"-12072.8-12093.5-12072.4- 1 2069.0-12070.0

-5881.8-5922.5-5898.7--5895.3-5902.4-5922.9-5904.6'-5903.3- 5915.9-5922.4-5898.2-5909.2-5899.s-5897.4-5900.0

HF calorimetryEvaluationEvaluationSolubilityEvaluationEvaluationEvaluationEvaluationHF calorimetry

HF calorimetryHF calorimetrySolubilityEvaluationSolubilityEvaluationSolubilityEvaluationEvaluationMolten saltEvaluationEvaluationEvaluationEvaluationHF calorimetry

534.7538.1

537.0537.0535.0535.2534.5

260.&!

280.58280.58260.79

260.83260.&t

260.8i!260.84t260.76261.19260.8t!

. Calculated trom the measured value and the entropy used in this study.

ENrrur-py oF FoRMATToN oFANTHOPHYLLITE AND TALC

The stability of magnesio-anthophyllite is in questionbecause results for the enthalpies and consequently theGibbs free energy of formation from two experimentaltechniques differ significantly. Early estimates of the en-thalpies and Gibbs free energies of formation of magne-sio-anthophyllite were based upon the calorimetric re-sults given in Weeks (1956) (e.g., Naumov et al., 1974;Helgeson et al., 1978). More recently, the solubility dataof Hemley eI al. (1977d have been utilized (e.g., Day etal., 1985; Berman et al., 1986; Berman, 1988). These dis-parate results (see Table I ) appear to be the critical valuesupon which the enthalpies and Gibbs free energies of for-mation of magnesio-anthophyllite are based.

Phase equilibrium reactions (e.g., Reactions l-5) pro-vide constraints on the thermodynamic properties ofmagnesio-anthophyllite:

magnesio-anthophyllite: 7 enstatite + q\artz + H2O (1)

3 magnesio-anthophyllite + 4 quartz + 4 HrO:7 talc (2)

magnesio-anthophyllite + forsterite:9 enstatite + HrO (3)

5 megnesio-anthophyllite + 4 HrO:9ta lc+4forster i te (4)

magnesio-anthophyllite: talc * 4 enstatite. (5)

Because of the large stoichiometric coefficients, phaseequilibrium reactions cannot be utilized to resolve the

disparity discussed above (see also Day et al., 1985). Forexample, a change of -2kJ/mol in the enthalpy of for-mation of enstatite (a change well within the uncertaintyof the value) would require a change of - 14 kJlmol (forReaction l) or -18 kJ/mol (for Reaction 3) in the en-thalpy of formation of magnesio-anthophyllite.

A similar disparity exists between aqueous HF andmolten salt calorimetric data (Barany, 1963; Bennington,I 956; Kiseleva and Ogorodova, 1984) and solubility data(Hemley et al., 1977a; Bricker et al., 1973; Hostetler etal., l97l) for talc. Because talc enters into several ofthephase equilibrium reactions, a review ofthe data for talcmay help resolve the disparate data for anthophyllite.

Three determinations of the Gibbs energy of formationof talc by solubility methods (Hemley et al., 1977a;Bnck-er et al., 1973; Hostetler et al., l97l) show good agree-ment but estimate a Gibbs free energy of formation fortalc that is 20 kJ less stable than that derived by HF andmolten salt calorimetry. The consistency of the resultsargues against a systematic error in the studies. If thecalorimetric data were considered correct. then the solu-bility data could be reinterepreted as representing equil-ibration with a phase less stable than talc, for example,vermiculite. However, before resorting to such an inter-pretation, a review of the calorimetric thermodynamiccycles is in order.

The cycle used by Barany (1963) appears to contain anerror in the enthalpy of solution of MgO and thus will beconsidered first. The thermodynamic cycles followed inthe studies ofBennington (1956) and Barany (1963) areslightly different, with Bennington (1956) using Mg(OH),rather than MgO and with dissolution in a different sol-vent.

t592

Barany (1963) measured the enthalpies of solution oftalc and of the constituent oxides and reported -187.8+ 1.4 kJ for the enthalpy ofthe reaction

4SiO, + 3MgO + HrO: Mg.SioO,o(OH), (6)

in a mixture of 10.050/o HCI and 10.050/o HF. Barany(1963) reported the following enthalpies of solution: fortalc, -849.94 t 0.80; SiO,, -143.70 + 0.27; MgO,-155.02 + 0.l l; and HrO, 2.09 + 0.21, all in kJ/mol.King et al. (1967) have also reported values for the en-thalpies of solution of two of the oxides in a solvent ofthe same composition, at the same temperature (disso-lution at 73.7 "C), and in the same equipment. A differentsample of MgO was used in the latter study. The resultsreported by King et al. (1967) are: for MgO, - 148.31 +0.21 and for HrO, 2.82 + 0.08, both in kJ/mol. Thesevalues differ significantly from the earlier results reportedby Barany (1963); however, the authors provide no dis-cussion ofthis fact.

The results provided for the enthalpies of solution ofMgO and HrO by King et al. (1967) are preferred for thefollowing reasons. The enthalpy of formation of forsteritehas been determined in several studies, and good agree-ment between these results and those of King etal. (1967)is found (e.g., King etal., 1967, -2171.7; Torgeson andSahama, 1948, -2177.1and -2174.9; Charlu etal.,1975,-2175.1; Kiseleva et al., 1979, -2171.9; Robie et al.,1982b, -2170.4; Brousse et al.,1984, -2172.0; all in kJ/mol with an uncertainty of about 12.5 kl/mol or less).If the reactions reported by Barany (1963) were substi-tuted for those used by King et al. (1967), the resultingenthalpy of formation for forsterite would be -2185.1

kJ/mol, a value in poor agreement with other studies.The enthalpies of solution of MgO reported in the two

studies are significantly different. The result reported byBarany (1963) is considered to be in error (too large) be-cause the sample was very fine grained (reagent gradematerial). Taylor and Wells (1938) have shown that fine-grained MgO has an excess enthalpy of solution. King etal. (1967) specifically stated that macrocrystalline MgOwas used in their study, and thus no excess enthalpy ofsolution is expected.

The enthalpy of formation of talc may now be calcu-lated from a Hess cycle that includes the enthalpy of so-lution oftalc from Barany (1963) and the enthalpies ofsolution of the component oxides from King et al. (1967).The enthalpy of formation of talc from the oxides is-166.97 + 1.5 U/mol and from the elements is -5900.07+ 2.1 kJlmol.

Bennington (1956) determined the enthalpy of forma-tion of talc from the oxides (-148.66 + l.5l kJlmol).Bennington (1956) measured the enthalpy of solution oftalc (-845.1| + 1.27 kJ/mol) in a solution of l0o/o HCIand l0o/o HF at 8l 'C. The ancillary enthalpies of solutionrequired for his cycle represented dissolutions in acid ofanother composition.

Problems in the Hess cycle used by Bennington (1956)are obvious from a comparison ofthe accepted enthalpies

HEMINGWAY: THERMODYNAMIC DATA FOR ANTHOPHYLLITE AND TALC

of solution of SiO, (Barany, 1963) and MgO (King et al.,1967) it a solution of l0o/o HCI and l0o/o HF and thevalues of -137.70 + 0.13 kJ/mol for SiO, (20.10/o HF)and -146.ll + 0.61 kJ/mol for MgO (2jo/oHCl and 50/oHF) used by Bennington (1956). The values used by Ben-nington are for dissolution at 8l "C, whereas those ofKing et al. (1967) represent dissolution at73.7'C. Basedupon the studies of Kilday and Prosen (1973) and Hem-ingway et al. (1988), we can estimate a correction of - l.lkJ to Bennington's (1956) value for quartz dissolution,g1ving73.7 "C. A similar correction for MgO would berequired.

We may substitute Bennington's (1956) enthalpy of so-lution of talc for that of Barany (1963) in the Hess cycledescribed above. In addition, we must assume no signif-icant change in the enthalpy of solution of talc between73.7 and 8l "C, and we must correct for the heat content(Hg, - Hl.,) of talc. The corrected enthalpy of talc dis-solution is -847.82 + 4 kJ/mol. The enthalpies of for-mation of talc derived from this value are -169.09 + 4and -5902.19 + 4 kJlmol from the oxides and elements,respectively.

The calorimetric values derived from the corrected Hesscycles described above are in good agreement amongthemselves and with estimates made from evaluation ofexperimental data or by using other experimental tech-niques (see Table l). Thermodynamic properties for talcat I bar and based upon the CODATA key values for theelements (Cox et al., 1989) are listed in Table 2.

Recently, molten salt solution calorimetry has been ap-plied to hydrous phases, and the enthalpy of formationof talc may be derived from data contained in three stud-ies (Kiseleva and Ogorodova, 1984; Akaogi et al., 1984;Clemens et al., 1987). Kiseleva and Ogorodova (1984)determined the enthalpy of formation of talc at 298.15K (-5922.4 + 5.6 kJ/mol) from Reaction 7 with disso-lution

2MgO + 4SiO, + Mg(OH),: Mg,SioO,o(OH), (7)

at 973 K. Clemens et al. (1987) determined the enthalpyof formation of phlogopite from measurements at 987 K.Values were reported for the enthalpies of solution forMgO, SiOr, and Mg(OH), that may be substituted intothe Hess cycle listed above, yielding -5913.1 + 5 kJ/mol for the enthalpy of formation of talc at 298.15 K.The Hess cycle may be modified slightly by substitutingthe enthalpy ofsolution offorsterite (Akaogi et al., 1984)for 2MgO + SiOr. This version of the Hess cycle yields-5911.9 + 5 kJ/mol. In the latter calculations, the en-thalpy of solution of talc aI 987 K was calculated fromthe value given by Kiseleva and Ogorodova (1984)through the addition of the enthalpy difference Hnr, -

Hn , and from the enthalpy of formation of forsterite fromKing et al. (1967). The enthalpy of solution of talc wasassumed to be constant over the temperature interval of973 to 987 K. If the enthalpy of solution of talc changedin accordance with changes observed for the enthalpiesof solution of MgO and SiO, (Shearer and Kleppa, 1973),

the calculated enthalpies of formation of talc would in-crease by -2.5 kJlmol.

Enthalpies of formation of talc calculated from the re-sults of molten salt calorimetry are sigrrificantly morenegative than are those derived from solubility data andthe corrected Hess cycle based upon aqueous HF calo-rimetry. However, the values are not internally consis-tent, although all are based upon the same value for theenthalpy of solution for talc. The results represent earlyefforts to study hydrous minerals or minerals containingvolatile components, and they suggest that systematic er-rors may be present. The enthalpy of formation of for-sterite may be calculated from the enthalpy of solutionofforsterite (Akaogi et al., 1984) and from the enthalpiesof solution of MgO and SiO, taken from Clemens et al.(1987, -2172.9 kl/mol) and from Kiseleva and Ogoro-dova (1984, -2177.2 kJ/mol). For the anhydrous phase,forsterite, the results are in reasonable agreement withvalues determined by other methods (cited above).

There are several experimental problems in the appli-cation of the molten salt solution calorimeter to the studyof hydrous or volatile component bearing phases that maylimit the usefirlness of the technique. For example, thereference phase, Mg(OH)r, decomposes at a temperaturelower than that of the molten salt solvent, but talc de-hydrates at a temperature above that used for dissolution.Thus the processes involved in the dissolution of eachphase may be different and result in a different final statefor each process (essentially a failure to maintain stoichi-ometry through a failure to match the concentration ofthe volatile component dissolved in the solvent). In ad-dition, accumulation of HrO vapor within the calorime-ter will result in a change in the thermal conductivity ofthe cell and, therefore, a change in the calibration oftheenergy equivalent of the calorimeter. It is not clear thata valid Hess cycle can be established for hydrous phasesusing the molten salt solution technique. It is clear thatthe enthalpy of formation of talc derived by the moltensalt technique deviates significantly from results ofothertechniques and is probably in error as a consequence ofundetected systematic errors.

The calorimetric data for magnesio-anthophyllite mustbe in error if the calorimetric values for forsterite and talc(discussed above) are correct and the phase equilibriumdata for Reactions 2 and 4 are correct. Two proceduresfollowed by Weeks (1956) may be faulted. Like Benning-ton (1956), Weeks (1956) used a mixed acid (20o/o HCIand 5olo HF) for sample dissolution. However, Weeks(1956) used the dissolution of quartz in a solution of 20.1o/oHF as part of his Hess cycle. The anthophyllite sampleused by Weeks (1956) contained Fe. Weeks (1956) useda questionable graphical technique to correct the enthalpyofsolution ofanthophyllite for the effect ofFe.

The enthalpy of solution of quartz used by Weeks (1956)appears to be fairly good despite the difference in acidsused. Weeks (1956) reported and used a value of -137.70kJ/mol for the enthalpy of solution of quartz in 20.1o/oHF at 8l oC. Neuvonen (1952) measured the enthalpy of

I 593

Trer-e 2. Thermodynamic properties for talc, Mg"SioO,o(OH),

Formation fromelements

HEMINGWAY: THERMODYNAMIC DATA FOR ANTHOPHYLLITE AND TALC

Tem-pera- Heat

GibbsEnthalpy energy

ture capacity Entropy function function Enthalpy energyK -----------J/mol.K ------ kJ/mot '.' '---.--

298.15 321.7'l+0.63

300 323jt7350 359.20400 386.604s0 406.22500 420.63550 432.51600 444.26650 457.93700 475.20750 497.49800 525.95

260.830 0.000 260.830+0.63262.825 1.988 260.836315.448 50.538 264.909365.298 90.925 274.373412.030 124.929 287.102455.610 153.811 301.799496.268 178.617 317.651534.398 200.258 334.141570.479 219.535 350.944605.020 237.154 367.866638.533 253.735 384.798671.513 269.824 401.689

-5900.0 -5s20.0+2.0 +2.O

-5900.1 -5517.8-5901.2 -5454.0-5901.2 -5390.1-5900.3 -5326.2-5899.0 -5262.5-5897.5 -5198.9-5895.7 -5135.5-5893.5 -5072.2-5890.9 -5009.2-5887.5 -4946.3-5883.2 -4883.7

solution of "extremely fine quartz powder" and reported- 139.7 5 kJ/mol for the enthalpy of dissolution in 20o/oHCI and 5olo HF at 75 "C. The difference of -2 kJ/molis well within the range of excess enthalpy of solution thatcan be expected for fine quartz powder (Hemingway andRobie, 1977; Hemingway et al., 1988). For further veri-fication, the enthalpy of solution of quarlz in 20o/o HCIand 5o/o HF may be calculated from the enthalpy of for-mation from the oxides of chrysotile (King et al., 1967)and the enthalpies of solution of chrysotile and ancillarycomponents in 20o/o HCI and 50/o HF as reported by Ben-nington ( I 956). The result of this calculation is - I 38.38kJ/mol for the enthalpy of solution of quartz.

Thus it would appear that the enthalpy of solution ofquartz in 20.1o/o HF and in 20o/o HCI and 5olo HF arenearly equivalent at 70-80 oC, but the value measured inl0o/o HCI and l0o/o HF is larger. This may explain whyBennington's (1956) Hess cycle for chrysotile (reactionsin 20o/o HCI and 50/o HF except quartz in 20.10/o HF)yields a value for the enthalpy of formation of chrysotilein good agreement with other studies, but his Hess cyclefor talc (talc dissolution in l0o/o HCI and l0o/o HF andother reactions as above) yields an erroneous result.

The graphical solution used by Weeks (1956) to correcthis measured enthalpy of solution of anthophyllite for theeffect of Fe is questionable because Hemingway and Ro-bie (1973) have shown that the enthalpy of solution ofCaO varies with the amount of MgO dissolved in an HCIsolvent. Using a plot of enthalpy of solution vs. sampleFe content, Weeks (1956) plotted the enthalpies of solu-tion of natural anthophyllite, tremolite, and ferrotremo-lite. He connected the values for the two tremolite com-positions and then constructed a line parallel to this linethrough the value for the natural anthophyllite. Weeks(1956) used the constructed line to estimate the enthalpyof solution of magnesio-anthophyllite. Because tremolitecontains both Ca and Mg and because the Mg content ofthe model sample increases as Fe is replaced by Mg, theenthalpic effect of the interaction between Ca and Mg canbe expected to increase. This effect would not be present

1594 HEMINGWAY: THERMODYNAMIC DATA FOR ANTHOPHYLLITE AND TALC

Tlele 3. Thermodynamic properties for magnesio-anthophyllite, Mg'SirOrr(OHL

Formation trom elements

Temperature Heat capacity Entropy Enthalpy functionK ---. J/mol.K -------'

Enthalpy Gibbs energy-- kJ/mol

Gibbs energyfunction

298.15

3003504004505005506006s0700750800850900

95010001 1001 200

663.90+2.O667.00737.08788.36827.81859.34885.28907.13925.87942.21956.63969.51981.1 1991.65

1001 .291 010.171 026.041039.90

534.500+3.5

538.616647.005748.937844.1 63933.071

1016.2291094.22'l1 1 67.5881236.8141302.3211364.4791423.6091479.991

1533.8681585.4571 682.4981772.382

0.000

4.104104.093186.572255.717314.558365.298409.565448.578483.266514.353542.405567.875591 .130

612.467632.1 33667.239697.728

534.500

534.513542.911562.3il588.446618 .513650.931684.656719.01 1753.548787.968822.075855.734888.861

921.401953.323

1015.2601 074.654

-12070.O+8.0

-12070.0- 1 2069.6-12067.O-12063.1- 1 2058.1-12052.5- 1 2046.3-12039.8- 12031 .0-12025.9- 12018.6-12011.3-12003.9

- 11343.3r8.0

-11338.8- 1 1216.8- 11094.8-10973.1-10851.7-10730.7-10610.0- 10489.8- 10368.0- 10250.5- 10131.4- 10012.8-9894.s

-12056.3- 1 2049.1-12033.4-12016.0

-9774.7-9653.7-9412.4-9173.0

in anthophyllite and thus invalidates Weeks'(1956) mod-el.

A comparison may be made of the enthalpies of solu-tion of Fe-Mg analogue pairs with no Ca as a systemcomponent. The enthalpies of solution of MgTiO' andFeTiO. in 10.050/o HCI and 10.050/o HF (Kelley et al.,1954) are -2023.4 and - l3l1.6 J/9, respectively. Theenthalpy of solution of MgrSiOo in the same acid (Kinget al., 1967) is -2717.5 J/g, and that for FerSiOo in 20o/oHCI and 50/o HF (King, 1952) is -1654.4 J/g. These re-sults suggest that the enthalpy of solution per gram ofthe Mg-bearing phase should be larger than that of theFe-bearing analogue. Weeks (1956) used -2162.3 and-2169.6 J/g, respectively, for the enthalpies of solutionof tremolite and ferrotremolite.

The enthalpy of solution of anthophyllite reported byWeeks (1956) may be corrected for the efect of Fe andminor Ca by a Hess cycle utilizing the enthalpies of so-lution of MgTiOr, FeTiOr, and CaTiO, (Kelley et al.,1954). The corrections are: 205.35 J for FeTiOr, 36.90 Jfor CaTiOr, and -283.80 J for MgTiOr. A correction of0.35 J is also necessary for Al (see Weeks, 1956). Sum-mation of these values yields -2368.13 J/g for the en-thalpy of solution of magrresio-anthophyllite (or - 1849.55kJ/mol).

A corrected enthalpy of formation for magnesio-antho-phyllite may be calculated from the corrected enthalpy ofsolution of anthophyllite and the Hess cycle used by Weeks(1956) and, for comparison, from a Hess cycle based uponthe enthalpy of solution of chrysotile (Bennington, 1956).The revised enthalpy of formation of magnesio-antho-phyllite from the oxides is -288.59 kJ/mol and -1207O

+ 8 kJ/mol from the elements. The enthalpy of forma-tion of magnesio-anthophyllite also may be calculatedfrom the reaction

2Me'Si'o'(o"t I HfS l fiL?ri,o,,(oH),. (8)

The Hess cycle based upon Reaction 8 yields a value of42.25 kI for the enthalpy of the reaction and - 12073 kJ/mol for the enthalpy of formation of magnesio-antho-phyllite (from the elements and using -4364.3 kJ/molfor the enthalpy of formation of chrysotile, e.g., Berman,l 988) .

The corrected values for the enthalpy of formation ofmagnesio-anthophyllite are in good agreement with thevalues predicted by Berman et al. (1986, -12072 kJlmol)and Berman (1988, - 12069 kI/mol). The results supportthe value derived by Hemley et al. (1977b, -12069 kJ/mol). Thermodynamic properties for magnesio-antho-phyllite at I bar and based upon the CODATA key valuesfor the elements (Cox et al., 1989) are listed in Table 3.

The enthalpy of formation of Mg.FeorSi'Orr(OH), maybe calculated using a similar procedure to that describedabove for magnesio-anthophyllite. Corrections to the ob-served enthalpy of solution of anthophyllite are: 34.22 Jfor FeTiO., 36.90 J for CaTiO., and -80.67 J for Mg-TiO.. The enthalpy of solution of 802.906 g of antho-phyllite is - 1876.21 kJlmol. The enthalpy of formationfrom the elements is derived from the reaction

6.3MgO + 0.35Fe,SiO4 + 7.65SiO, + H,O: MgrFeorSisorr(OH)r. (9)

The enthalpy of Reaction 9 is -603.555 kJ based uponthe enthalpies of solution for the component oxides givenby Weeks (1956) and that for fayalite from King (1952).The enthalpy of formation from the elements ofMgu.Feo,SirOrr(OH),is -12163 + 8 kJlmol based uponthe enthalpies of formation of the oxides components fromRobie et al. (1979) and of fayalite from Robie et al.(l 982b).

Survrprl,nv AND coNcLUSroNS

Errors in the derivation ofthe heat capacities and en-tropy of anthophyllite of the composition Mgu..Feo 7Si8-

Orr(OH), (Krupka, 1984; Krupka et al., 1985a, 1985b)have been documented, and revised values have beencalculated. Revised equations for the heat capacity ofthetwo anthophyllite compositions are given for the tem-perature interval 298.15 to 1200 K.

Separate problems in the calorimetric Hess cycles usedby Barany (1963), Bennington (1956), and Weeks (1956)caused the enthalpies of formation of talc and magnesio-anthophyllite reported by these authors to be in error.These problems have been discussed, and revised calori-metric values are reported. The revised values are inagreement with solubility data (Hemley et al., 1977a,1977b; Bricker et al., 1973; Hostetler et al., l97l) andwith optimized solutions derived by the techniques oflinear and mathematical programming (Day et al., 1985;Berman et al., 1986; Berman, 1988).

Analysis of the results of molten salt solution calorim-etry applied to hydrous phases suggests that systematicerrors may exist in the procedures. These errors may leadto incorrect values for enthalpies of solution for phasesnot stable at the temperature of the solvent. Further workwill be required to validate the molten salt solution tech-nique in this application.

The recommended thermodynamic properties for an-thophyllite and talc derived from the calorimetric studiesare: S!r, - ^S8: 554.2 + 3 J/mol'Kand \Hgnr: -12163+ 8 kJ/mol for Mgo rFeo,SirOrr(OH)r; S!n, : Jl{.J + 3.5J/mol.K, LrH|n": -12070 + 8 kJ/mol, and ArG!,, :-11343 + 8 kJ/mol for Mg,SirO,r(OHL; and A,Hlnr:-5900 + 2N/mol and ArGlr, : -5520 + 2 kJ/mol fortalc.

RnrnnrNcps crrEDAkaogi, M., Ross, N.L., McMillan, P., and Nawotsky, A. (1984) The

Mg,SiO" polyrnorphs (olivine, modified spinel and spinel)-Thermo-dynamic properties from oxide rnelt solution calorimetry, phase rela-tions, and models of lattice vibrations. American Mineralogist, 69,499-512.

Anovitz, L.M., Essene, E.J., Hemingway, 8.S., Komada, N.L., and Wes-trum, 8.F., Jr. (1988) The heat-capacities ofgrunerite and deerite: Phaseequilibria in the system Fe-Si-C-O-H and implications for the meta-morphism of banded iron formations. Eos. 69. 5 I 5.

Barany, R- (1963) Heats offormation ofgehlenite and talc. U.S. Bureauof Mines Report of Investigations 625 l.

Bennington, K.O. (1956) Role of shearing stress and pressure in differ-entiation as illustrated by some mineral reactions in the system MgO-SiO,-HrO. Journal of Geology, 64, 558-577 .

Bennin4on, K.O., Ferr-ante, M.J., and Stuve, J.M. (1978) Thermodynam-ic data on the amphibole asbestos minerals amosite and crocidolite.U.S. Bureau of Mines Report of Investigations 8265.

Berman, R.G. (1988) Intemally-consistent thermodynamic data for rnin-erals in the system NarO-K,O-CaO-MgO-FeO-FerOr-AlrO3-SiO,-TiOr-HrO-CO,. Joumal of Petrology, 29, 445-522.

Berman, R.G., Engi, M., Greenwood, H.J., and Brown, T.H. (1986) Der-ivation of intemally-consistent thermodynamic data by the techniqueof mathematical programming, a review with application to ilte systemMgO-SiO,-H,O. Journal of Petrology, 27 , 133l-1364.

Bricker, O.P., Nesbitt, H.W., and Gunter, W.D. (1973) The stability oftalc. American Mineralogist, 58, 64-72.

Brousse, C., Newton, R.C., and Kleppa, O.J. (1984) Enthalpy of forma-tion of fo$terite, enstatite, akermanite, monticellite. and merwinite at1073 K determined by alkali borate solution calorimetry. Geochimicaet cosmochimica Acta. 48. l08l-1088.

1 595

Charlu, T.V., Newton, R.C., and Kleppa, O.J. (1975) Enthalpies of for-mation at 970 K of compounds in the system MgO-AlrO.-SiO, fromhigh temperature solution calorimetry. Geochimica et CosmochimicaActa.39. 1487-1497.

Clemens, J.D., Circone, S., Navrotsky, A., McMillan, P.F., Smith, B.K.,and Wall, V.J. (1987) Phlogopite: High temperature solution calorim-etry, themodynamic properties, Al-Si and stacking disorder, and phaseequilibria. Geochirnica et Cosmochimica Acta, 5l, 2569-2578.

Cox, J.D., Wagman, D.D., and Medvedev, V.A. (1989) CODATA keyvalues for thermodynamics. Hemisphere Publishing Corporation, NewYork.

Day, H.W., Chernosky, J.V., and Kumin, H.J. (1985) Equilibria in thesystem MgO-SiO,-H,O: A thermodynamic analysis. American Miner-alogist, 70, 237-248.

Haas, J.L., Jr., and Fisher, J.R. (1976) Simultaneous evaluation and cor-relation of thermodynamic data. American Joumal of Science, 276,525-545.

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

Hemingway, B.S., and Robie, R.A. (1973) A calorimetric determinationof the standard enthalpies of formation of huntite, CaMgr(CO,)., andartinite, Mg,(OH),COr.3HrO, and their standard Gibbs free energiesof formation. Journal of Research of the U.S. Geological Suwey, l,535-54 l .

Hemingway,8.S., and Robie, R.A. (1977) Enthalpies of formation of lowalbite (NaAlSi,O,), gibbsite (Al(OH)j) and NaAlO,: Revised values forA}ljr* and AGlr* of some aluminosilicate minerals. Journal of Re-search ofthe U.S. Geological Survey, 5, 413-429.

Hemingway, 8.S., Nitkiewicz, A., and Donahoe, R.J. (1988) Key valuesfor therrnodynamic properties of geologic materials: I. Enthalpy of so-lution of quartz in hydrofluoric acid, a preliminary evaluation. U.S.Geological Survey Open-file Report 88-252.

Hemley, J.J., Montoya, J.W., Christ, C.L., and Hostetler, P.B. (1977a)Mineral equilibria in the MgO-SiOlH,O system: I. Talc-chrysotile-forsterite-brucite stability relations. American Joumal of Science, 277,322-35r.

Hemley, J.J., Shaw, D.R., and Luce, R.W. (1977b) Mineral equilibria inthe MgO-SiOr-HrO system: II. Talc-antigorite-forsterite-anthophyllite-enstatite stability relations and some geologic implications in the sys-tem. American Joumal of Science. 277 . 353-383.

Hostetler, P.B., Hemley, J.J., Christ, C.L., and Montoya, J.W. (1971)Talc-chrysotile equilibrium in aqueous solutions. Geological Society ofAmerica Abstracts with Programs, 3, 605-606.

Kelley, K.K., Todd, S.S., and King, E.G. (1954) Heat and free energy darafor titanates ofiron and the alkaline-earth rnetals. U.S. Bureau ofMinesReport of Investigations 5059.

Kilday, M.V., and Prosen, E.J. (1973) The enthalpy of solution of lowquartz (c-quartz) in aqueous hydrofluoric acid. U.S. National Bureauof Standards Journal of Research . 7 7 A. 20 5-21 5.

King, E.G. (1952) Heats of formation of manganous metasilicate (rho-donite) and ferrous orthosilicate (fayalite). American Chemical SocietyJournal. 74. 4446-4448.

King, E.G., Barany, R., Weller, W.W., and Pankmtz, L.B. (1967) Ther-modynamic properties of forsterite and serpentine. U.S. Bureau of MinesReport of Investigations 6962.

Kiseleva, I.A., and Ogorodova, L.P. (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 (translated from Geokhirniya, 1745-1755, 1983).

Kiseleva, I.A., Ogorodova, L.P., Topor, N.D., and Chigareva, O.G. (1979)A thermochemical study of the CaO-MgO-SiO, system. Geokhimiya,l 8 l l - 1825 .

Krupka, K.M. (1984) Thermodynamic analysis of some equilibria in thesystem MgO-SiO,-H,O. Ph.D. thesis, The Pennsylvania State Univer-sity, University Park, Pennsylvania.

Ikupka, K.M., Robie, R.A., Hemingway, B.S., and Kerrick, D.M. (1985a)Inw-temperature heat capacities and derived thermodynamic proper-ties of anthophyllite, diopside, enstatite, bronzite, and wollastonite.American Mineralogist, 7 O, 249-260.

Krupka, K.M., Hemingway, B.S., Robie, R.A., and Kerrick, D.M. (1985b)

HEMINGWAY: THERMODYNAMIC DATA FOR ANTHOPHYLLITE AND TALC

r 596

High-temperature heat capacities and derived thermodynamic prop-

erties of anthophyllite, diopside, dolomite, enstatite, bronzite, talc,tremolite, and wollastonite. American Mineralogist, 70,261-27 L.

Naumov, G.B., Ryzhenko, 8.N., and Khodakovsky, I.L. (1974) Hand-book ofthermodynamic data. National Technical Information ServicePB-226,722.

Neuvonen, K.J. (1952) Thermochemical investigation of the Akereman-ite-gehlenite series. Bulletin de la Commission G6ologique de Finlande,1 5 8 , 5 1 p .

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

Robie, R.A., Hemingway, B.S, and Takei, H. (1982a) Heat capacitiesand entropies of MgrSiOo, Mn,SiOo, and CorSiO" between 5 and 380K. American Mineralogist, 67, 470-482.

Robie, R.A., Finch, C.B., and Hemingway, B.S. (1982b) Heat capacityand entropy offayalite (FerSiOn) between 5.1 and 383 K: Comparisonof calorimetric and equilibrium values for the QFM buffer reaction.American Mineralogist, 67 , 463-469.

HEMINGWAY: THERMODYNAMIC DATA FOR ANTHOPHYLLITE AND TALC

Robie, R.A., Hemingway, B.S., Ito, J., and Krupka, K.M. (1984) Heat

capacity and entropy of NirSiOn-olivine from 5 to 1000 K and heat

capacity of Co,SiO" from 360 to 1000 K. American Mineralogist, 69,

1096 -1 l 0 l .Shearer, J.A., and Kleppa, O.J. (1973) The enthalpies of formation of

MgAl,O4, MgSiO., MgrSiOo, and Al,SiO, by oxide melt solution calo-

rimetry. Journal of Inorganic and Nuclear Chemistry, 35, 1073-1078.

Taylor, K., and Wells, L.S. (1938) Studies ofheat ofsolution ofcalcium

and magnesium oxides and hydroxides. U.S' National Bureau ofStan-

dards Journal of Research, 21, 133-149.Torgeson, D.R., and Sahama, Th.G. (1948) A hydrofluoric acid solution

calorimeter and determination of the heats of formation of M&SiOo,

MgSiOr, and CaSiO,. American Chemical Society Journal, 70,2156-

2160 .Weeks, W.F. (1956) Heats of formation of metamorphic minerals in the

system CaO-MgO-SiO1H'O and their petrological significance. Joumalof Geology, 64, 456-472.

MeNuscnrpr REcETvED Novrrvrsen 21, l99OMeNuscnrrr ACcEPTED Mav 20, l99l