Condensation in Dust-enriched Systems, by D.S. Ebel and L. Grossman,
Geochimica et Cosmochimica Acta, 1999

Technique: Thermodynamic Data and Models for Solid Phases

The internally consistent thermodynamic database of Berman (1988), or a combination of the internally consistent databases of Berman and Brown (1985) and Berman (1983) were used wherever possible for all potential condensates in Table 3 and for most end-member components of the solid solution series in Table 4, except for the metal alloy. This means that Berman (1988) was the source of end-member data for the melilite and feldspar solid solutions, not the references cited for the solution models for these phases. The JANAF data (Chase et al., 1985) for pyrrhotite, Fe_0.877S, are based on estimation of heat capacities from 600 to 1475K. Recent work below 1000K by Grønvold and Stølen (1992) indicates that these data cause over-stabilization of pyrrhotite by ~5 kJ at 1000K. Therefore, Gibbs energies of formation of pyrrhotite from the JANAF tables were revised upward by this amount in the calculation. This revision lowers the appearance temperature of pyrrhotite by ~50K, compared to the JANAF data.

The solid solution models implemented in the MELTS program (circa 1993; Table 4) were used in all calculations, except that Ca-pyroxenes (Sack and Ghiorso, 1994a,b,c) were constrained to have 1 total atom of Ca + Na per 6 oxygen atoms. These represent the most comprehensive treatments of the anhydrous igneous rock-forming minerals presently available, and are the solid solution models against which the MELTS silicate liquid model is calibrated. In addition, solid Fe-Ni-Si-Cr-Co alloy was modeled using JANAF data (Chase et al., 1985) for pure metal end-members, and an asymmetric binary solution model calibrated against activity data for the binary systems of Chuang et al. (1986b) for Fe-Ni, Sakao and Elliott (1975) for Fe-Si, and Normanton et al. (1976) for Fe-Cr, with Fe-Co treated as ideal. Such a calibration is justifiable for the dilute alloys found at high temperature in this work.

Some cations of great interest in condensation are not contained in some of the liquid or solid solution models used here. These are the first condensation calculations in which the TiO₂ content of spinel is modeled, and extraordinarily high TiO₂ contents are predicted at very high temperatures. In all such cases, however, spinel coexists with a CMAS liquid into which TiO₂ is artificially prevented from dissolving. Partitioning experiments (Connolly and Burnett, 1999) suggest that these high TiO₂ contents may be spurious. Insufficient experimental work exists to justify inclusion of Ti₃+ or Cr₃+ in the pyroxene model. No solution model is used for Mn, S, P or C in the metal alloy, and this could artificially enhance the stabilities of troilite, pyrrhotite and whitlockite. Similarly, our inability to account for Ni or Co in troilite or pyrrhotite, nor for Cr, Ti or Al in olivine, may artificially destabilize these phases slightly. Although Hirschmann (1991) has modeled Ni, Co and Mn in olivine, these elements are not addressed by the pyroxene model, nor are Ni and Co included in the spinel model employed here. Because inclusion of Ni, Co or Mn in only one of these phases would artificially stabilize that phase and cause it to contain excessive amounts of these cations, these cations were not included in the olivine model. This omission, however, artificially stabilizes MnTiO₃-rich rhombohedral oxide solid solutions and crystalline MnO.

Miscellaneous solid phases^a
Table 3. Pure solid phases considered in the calculation, and sources of thermodynamic data.
Aenigmatite	Na₂Fe₅TiSi₆O₂₀	M
Andalusite	Al₂SiO₅	B8	Al	MgS
Anhydrite	CaSO₄	R	Al₄C₃	Mg₂Si
Anthophyllite	Mg₇Si₉O₂₂(OH)₂	B8	AlN	MgSO₄
Apatite	Ca₅(PO₄)₃OH	M	Al₂S₃	MgTi₂O₅
Brucite	Mg(OH)₂	B8	Al₆Si₂O₁₃	Na
Ca-aluminate	CaAl₂O₄	B5	alpha Ca	alpha Na₃AlF₆
Calcite	CaCO₃	B8	beta Ca	beta Na₃AlF₆
Cohenite	Fe₃C	R	CaCl₂	NaAlO₂
Cordierite	Mg₂Al₄Si₅O₁₈	B8	CaF₂	NaCl
Corundum	Al₂O₃	B8	Ca(OH)₂	NaCN
Cristobalite	SiO₂	M	CaS	Na₂CO₃
Dolomite	CaMg(CO₃)₂	B8	CoO	NaF
Grossite	CaAl₄O₇	B5	Cr₃C₂	NaH
Hatrurite	Ca₃SiO₅	B5	CrN	NaO₂
Hibonite	CaAl₁₂O₁₉	B3	Cr₂N	Na₂O
Kalsilite	KAlSiO₄	M	Cr₂O₃	Na₂O₂
Leucite	KAlSi₂O₆	M	FeCl₂	NaOH
Lime	CaO	B8	FeF₂	Na₂S
Magnesite	MgCO₃	B8	Fe_0.947O	Na₂S₂
Manganosite	MnO	R	FeO	Na₂SiO₃
Merwinite	Ca₃MgSi₂O₉	B8	Fe(OH)₂	Na₂Si₂O₅
Nepheline	NaAlSiO₄	M	Fe(OH)₃	Na₂SO₄(I-V)
Periclase	MgO	B8	FeS2 (Pyrite)	NH₄Cl
Perovskite	CaTiO₃	R	FeSO₄	P
Pyrrhotite	Fe_0.877S	J	Fe₂(SO₄)₃	monocl S
Quartz	SiO₂	M	Graphite	ortho S
Rankinite	Ca₃Si₂O₇	B5	K	alpha SiC
Rutile	TiO₂	B8	KCl	beta SiC
Sapphirine	Mg₄Al₁₀Si₂O₂₃	B3	KF	Si₃N₄
Sillimanite	Al₂SiO₅	B8	KF₂H	SiS₂
Sinoite	Si₂N₂O	F	KH	alpha Ti
Sphene	CaTiSiO₅	B8	K₂O	beta Ti
Talc	Mg₃Si₄O₁₀(OH)₂	B8	KOH	TiC
Tialite	Al₂TiO₅	R	K₂S	TiH₂
Tri-Ca aluminate	Ca₃Al₂O₆	B5	K₂SO₄	TiN
Tridymite	SiO₂	M	K₂SiO₃	alpha TiO
Troilite	FeS	C	Mg	beta TiO
Whitlockite	Ca₃(PO₄)₂	M	MgC₂	Ti₂O₃
Wollastonite	CaSiO₃	B8	Mg₂C₃	Ti₄O₇
			MgCl₂	alpha Ti₃O₅
			MgH₂	beta Ti₃O₅
			MgF₂
			Mg₃N₂
note: a) Symbols for data are: B5=Cp from Berman and Brown (1985), 298 K data from Berman (1983); B3=Berman (1983); B8=Berman (1988); C=Hsieh et al. (1987); R=Robie et al. (1978); F=Fegley (1981); M='MELTS' software database (Ghiorso & Sack, 1995); J= Chase et al. (1985) modified for consistency with Grønvold & Stølen (1992).

Table 4. Solid solutions considered in the calculation, and sources of solution models.
Metal alloy (this work)			Feldspar (Elkins and Grove, 1990)
Iron	Fe	Albite		NaAlSi₃O₈
Nickel	Ni	Anorthite		CaAl₂Si₂O₈
Silicon	Si	Sanidine		KAlSi₃O₈
Chromium	Cr
Cobalt	Co	Spinel (Sack and Ghiorso, 1991a,b)
		Chromite		FeCr₂O₄
Olivine (Sack and Ghiorso, 1989, 1994b)			Hercynite		FeAl₂O₄
Fayalite	Fe₂SiO₄	Magnetite		Fe₃O₄
Forsterite	Mg₂SiO₄	Spinel		MgAl₂O₄
Monticellite	CaMgSiO₄	Ulvospinel		Fe₂TiO₄

Melilite (Charlu et al., 1981)			Rhombohedral oxide (Ghiorso, 1990)
Åkermanite	Ca₂MgSi₂O₇	Geikielite		MgTiO₃
Gehlenite	Ca₂Al₂SiO₇	Hematite		Fe₂O₃
		Ilmenite		FeTiO₃
Orthopyroxene (Sack & Ghiorso, 1989, 1994b)			Pyrophanite		MnTiO₃
Enstatite	Mg₂Si₂O₆
Ferrosilite	Fe₂Si₂O₆

Ca-pyroxene (Sack & Ghiorso, 1994a,b,c)
Diopside	CaMgSi₂O₆
Hedenbergite	CaFeSi₂O₆
Alumino-buffonite	CaTi_0.5Mg_0.5AlSiO₆
Buffonite	CaTi_0.5Mg_0.5FeSiO₆
Essenite	CaFeAlSiO₆
Jadeite	NaAlSi₂O₆

CONDENSATION
in
DUST-ENRICHED SYSTEMS

Denton S. Ebel (1)

Lawrence Grossman(1,2)

(1) Department of The Geophysical Sciences
The University of Chicago
5734 South Ellis Ave.
Chicago, IL 60637

(2) Enrico Fermi Institute
The University of Chicago
5640 South Ellis Ave.
Chicago, IL 60637

Submitted December 22, 1998 to

Geochimica et Cosmochimica Acta

Revised version submitted June 30, 1999

Abstract

Introduction
Technique

Bulk Composition

Method of Calculation

Data for Elements and Gas Species

Data and Models for Solids

Data and Models for Silicate Liquids

Test of MELTS: Peridotite KLB-1

Transition Between Liquid Models

Results

Vapor of Solar Composition

General Effects of Dust Enrichment and Total Pressure