Fe,Cr)'* (Ti,Fe)Rsn, a New Mineralin the Busteeenstatite Achondrite

American Minerulogist, Volume 59, pages 465470, 1974

Heideite,(Fe,Cr)'* (Ti,Fe)rSn, A New Mineralin the BusteeEnstatite Achondrite

Kr-lus Knu- Departmentof Geologyand Institute of Meteoritics, Uniuersityof New Mexico, Albuquerque,New Mexico 87131

Rosrx Bnnrt NASA, JohnsonSpace Center, Houston.Texas 77058

Absfi?ct

Heideite, (Fe, Cr)r*,(Ti, Fe)sSroccurs as rare anhedral grains up to 100 pm in diameterin the Bustee enstatite achondrite. Other sulfides in Bustee are titanian troilite, ferroan alabandite, and daubreelite. Heideite (named after Professor Fritz Heide) has a distorted NiAs structure. It is probably pono- clinic (12/m). Indexed X-ray powder data are given (14 lines). The strongest lines are: 2.06 100 Iu, z.g 75 202, 1.72 75 310, 2.97 so |rc. It is cream white in color in reflected light and exhibits moderately strong reflection birefringence. Electron microprobe analysesand crystallogtaphic data indicate that heideite has the structural formula (Fe2+oesCr2+orr)(Ti3+r.zoFe2+o.ao)Sr. One grain of a chromian variety of heideite was found to contain 7.9 wt percent Cr substituting for Fe. The phasewas synthesizedat 650oC.Troilite in Busteeranges in Ti content from 0.2 to 16.3 wt percent, suggestingthat the sulfide assemblageformed and cooled under disequilibrium conditions.

Introduction other silicate minerals. Small quantities of metallic Grains of a phasethat is cream-coloredin reflected nickel-iron and osbornite (TiN) are also present. light caught our attention in the Bustee enstatite Ninety sulfidegrains in onethin sectionof Busteewere achondrite. Electron microprobe analysesindicate a analyzed;grains in order of abundanceare: titanian formula of the type (Ti, Fe, Cr)r*,Sn in which Ti is troilite (78 percent),ferroan alabandite(11 percent), more abundantthan Fe, thus establishingit as a new daubreelite(10 percent),chromian heideite (l percent). mineral. Titanium-rich troilite had previously been Compositional ranges of the above minerals are reported in Bustee and other enstatite achondrites listed in Table l; Ca was not determinedexcept for by Keil (1969) who recorded 16.3 wt percent Ti in alabandite. Selectedanalyses of troilite, alabandite, one troilite grain from Bustee.He pointed out the and daubreeliteare listed in Table 2. Figure I shows strongly chalcophile behavior of Ti in enstatite the extent of Ti solid solution in titanian troilite in chondrites and achondritesand also mentioned the Bustee.It is noteworthy that a large proportion of discoveryof a Ti, Fe sulfide in which Ti ) Fe. This compositionslie off the FeS-TiSjoin and that some latter phaseis the subject of the presentpaper. The compositionslie closer to the FeS-TisS"join than to mineral is namedin honor of the late ProfessorFritz the FeS-TiSjoin. Most grains in Busteeare homo' Heide of Jena, Germany, in recognition of his many geneous; however, widespread heterogeneityfrom contributionsto the field of meteoriticsat a time when grain to grain indicates that chemical equilibrium very little work was being done in this important was not achievedbetween distant grains. field. The name and the mineral have been approved Heideite occurs as rare, discrete,anhedral grains by the Commission on New Minerals and Mineral at the grain boundaries of silicate minerals in all Namesof the InternationalMineralogical Association. three polishedthin sectionsof Busteethat we examined. Grains range in diameter up to 100 pm, but Occurrence and Co-existing Sulfide Minerals most are of the order of 40 prmin diameter.One to The Busteeenstatite achondrite consistspredomi- four grains of heideite were found in each section nantly of enstatiteand diopsidewith minor amountsof (approximately I cm') that we examined.Heideite 465 466 K. KEIL AND R. BRETT

TesLB l Compositional Ranges(wt percent)of Troilite, Alabandite, and Daubreelite in Busteeas Measuredby Electron Microprobe Techniques

Mineral- IA I"lc

Troilite L6.3 - 0,2 51.8 - \l+.? 1.4 - O.2 0.5 - 0,01 0.5 - 0.05 * l+0.1+- 36.0 Alabanttite 0.09 - 0.Ol- 16.2 - ff.3 0.18- 0.11 h\.9 - l.r 7.3 - r.6 o.72- O.2o \o.7 - 37.6 Daubreelite 1.4 - 0.05 15.8 - ?.3 35.1+- 3l+.1 r.? - o.B o.\5- o,25 * 45.0- \3.3

* = not d.eternined..

is nowherein contact with any other opaquemineral Ti, Cr, and S, the phase also contains about 200 in the sectionsexamined. The mineral has a creamy- ppm Mn and lessthan 0.5 wt percentMg. One grain white color in reflected light and has moderately should be termed chromian heideite; it contains strong reflection birefringenceranging from purple 7.9 wt percentCr and is correspondinglydeficient in gay to cream gray. Based on the appearanceof Fe with respectto other heideitegrains. This suggests scratches and the relative polishing hardness of that heideite may be part of a solid solution series troilite and heideitein the samepolished section, the towards brezinaite(CrrSn). hardness of heideite is similar to that of troilite. It is clear from Table 3 that in order to maintain Surprisingly, the Bustee mineral has no discernible the chargebalance, Ti must be trivalent and Cr must cleavage.The small grain sizeand rarity of the phase' be divalent.Such valence states are consistentwith the precluded a determination of cleavage, magnetic highly reduced conditions under which enstatite properties,specific gravity, and hardness. achondriteshave formed (e.g., Keil, 1969). We consideredthat the formula of heideitemay be Composition of the type (Ti, Fe, Cr)r-,S, and be an extremely Heideitewas analyzed with an Anr-nux-snaelectron Ti-rich pyrrhotite solid solution, but rejected this microprobeX-ray analyznrat an excitation potential possibility for the following reasons: of 15 kV. Minerals and synthetic compounds were (l) x in the aboveformula is 0.21 for heideite and used as standards,and corrections for instrumental is thus higher than for natural or syntheticpyrrhotite. and matrix effectswere made following procedures Such cation deficiency would be highly unlikely, describedby Keil (1967). An analysisof heideite is especially for a phase apparently formed in the listed in Table 3, with a calculatedformula and charge presenceof metallic iron. balance.Heideite varies little in composition. In the (2) Experimental work by Viaene, Kullerud, and six grains analyzed,Ti ranges from 27.5 to 29.6 wt Taylor (1971)indicates a compositionalgap between percent and S ranges from 43.7 to 46.1 wt percent. (Fe, Ti),-"S solid solution and FeTirSnsolid solution. The range in Cr content with the exception of one A similar gap is seen between Ti-rich troilite and grain is less than 0.1 wt percent. In addition to Fe, heideitein Bustee(Fig. 1).

TasLB 2. SelectedCompositions of Daubreelite (1,2), Alabandite (3), and Troilite (4-12)in Busteeas Obtained by Electron Microprobe Techniques

Ti 0.07 0.36 o.ot 14.r 13.9 7.8 5.93 \.83 3.57 2.B'( 1 .l+8 O.20 Fe l+\ r5.3 T.)+ l.6.2 ,7 )+3. 7 ,2.o 55.9 )o.z )o.'z 57,8 ,9.\ 63.0 Cr 35.4 35.1 0.18 o.T8 o.9r 0.82 0.l+5 o.55 o.l+t 0.31 r.4o o.\2 I{n 2.O l_0.3 l+r.5 o.\9 o.ol_ o.ot o.ol+ 0.02 0.0\ o.or_ o.o5 o,o2 Mg 0.27 0.1+5 r.75 0.32 o.3B o.3B o.35 o.3L 0.26 o.16 o.3B o.zg s 45.r \5.2 38.2 \o.2 l+0.3 38.7 38.7 36.8 37.7 35.6 36.5 36.,

TotaL 99.r 98.8 97.9 r-00.6 99.2 99.7 100,4 98.9 98.7 98. o 99.2 loo ,l+ HEIDEITE, A NEW MINERAL 467

ToS IoS

FeS ris lo 20 30 40 50 60 70 80

Frc. 1. Plot of TiS contents of Bustee troilites (mole percent).H is composition of natural heideite' if Cr is recalculatedas Fe.

Phase F,quilibria includes FeTirSn and presumably heideite, extends join. We havesynthesized heideite from elementalTi, Fe, towardsthe FeS and S in evacuatedsilica glasstubes held at 650"C Strucfure for 2 weeks(Table 3). The synthesizedmaterial has On the basis of X-ray powder diffraction studies, optical properties to natural heideite and identical Hahn, Harder, and Brockmtiller (1956) state that gives a similar X-ray powder pattern (Table 4). synthetic FeTirSn prepared at 600oC has a NiAs Syntheticheideite is not attracted to a hand magnet. structure(Table 4). Plovnick,Vlasse, and Wold (1968) Etch tests of 4 minutes duration on the polished found that FeTirSnprepared over a range of tem- heideite using standard etch surface of synthetic peraturesas high as 3(X)"Chas the CrrSn structure gave the following results: HNO' (1 : l), reagents which is a monoclinic defect NiAs structure (space effervescence,iridescent blue etch;HCI (l : 1),negative; group I2/m). This structure has cation layers with KOIJ (40To,by wt.), negative;KCN (20%, by wt.), ordered vacanciesalternating with completely filled negative;HgCl, (s%by wt.), iridescentblue-to-brown cation layers. Plovnick, Vlasse,and Wold show that tarnish; FeCl,"(207o by wt.), iridescentbrown-to-blue there is a limited compositionalrange for the ordered tarnish. structure. Appreciable deviation from the FeTirS. We performed differential thermal analysis on composition results in formation of a disordered syntheticheideite in evacuatedsilica glasstubes using structure with trigonal symmetry (3m'1. Plovnick, the method of Kullerud (1971).Heating and cooling Vlasse, and Wold state that no disorder occurs in rates were 5'C/minute. The temperaturerange of the FeTirSnup to about 1000o,unless decomposition and experimentswas from room temperatureto 900'C. lack of stoichiometry occur. Nakajima, Horinchi, No exothermic or endothermic reactions were and Morimoto (1972) confirm the structure of observed.N. Morimoto and S. Nakajima (written Plovnick, Vlasse,and Wold (1968).The filled cation communication,1973) find that high and low tempera- layers contain Ti atoms; the layers with vacancies ture polymorphs exist for FeTirSn.Either heideite is contain Fe. With increase of Fe/Ti, Fe occupies sufficiently removed compositionally from FeTirSn that thesepolymorphs do not occur,or any transitions Trslr 3. Composition (wt percent) and Formula for Heideite that occur are too sluggishor have AH too small to be and Composition of Synthetic Heideite* recordedat the heatingand cooling ratesused.

Hahn, Harder, and Brockmiiller (1956)made runs L Z Fomula and charges for Heldelte in the system FeS-TiS-TiS' at 600"C and found a solid solution field that extendedfrom FeS to about ri 28,s zs.s (ref+r,c'l+rr) (rrfro r.f+ro) so Fe 25.1 25.I 45 mole percentTiS and at least to l0 mole percent TiSr. Their runs were exploratory, however,and no cr 2.9 1.98 + 0.32 + 5.10 + 0.60 = 8.00 s 44.9 45.2 phase boundaries can be drawn. Viaene, Kullerud, and Taylor (1971) fgund an extensivemonosulfide Total 101,4 99.8 solid solution betweenFe'-"S and Ti'-,S at 800'C. 7: Hei.deite; avewrge of 5 gtains utaLyzed, They found that a ternary solid solution, which 2: Suntheti,c hei,deite; -- = not detected. K. KEIL AND R. BRETT

Tralp 4. d-values in A of Natural and synthetic Heideite and synthetic FeTizSr*

Natu:'al- heideite qrrhthati^ h6iaai+6 FaTi s. ** ry -1-

dobs d"aJ-c r AAT c- c _ -t oos caac oDs caac

002 5.7r '.703 20 5.7)8 5.7\3 2 5.72 5.7\6 '-l::o"u 101 5.28 5.280 25 5.3o2 5.2TT 2 ^-:- ^--^-. 1t_0 2.97 2.970 50 2.9't' 2.975 z.YI l.ylo veek

t!1 2.5t+ 2.66 2.6t+\ 2.6\Z oo 2.63 2.6\r *:::u 1O)+ 2.57 2.r70 5 -L I4 z.uo t.u>) 100 2.068 z.o6t roo 2.o5-) 2.066 v. strong 1.9\ 1.91+o 015 :: :: tlgo rlirt v. weak 020 r.7z 1.7r-1 7' r.7r9 r.7L9 50 I.71 1.?16 n. strong

1.64 r.5\3 v. wea.]< tfo, tao, zuo rlZor.Zos 5 r.609 r.509 2 1.60 r.612 n. strong 222 1.l+)+t.)+36 10 I.l+1+5t.l+\0 10 I .l+2 I .l+l{0 m. strong

IL7, l+o\, [o]+, 224, 1.32 r.32 5 1.3r- f.318 mediun 2rT, 0r8, 22\

veak 118, 413, 2oB, l-t8

4]5 f.17 ].rT7 strong-ned. J-4 pOSS]-Ol.e refl-ections ':lo '-10 q f.r0 l_.105 v. srrong 1.05- 1.055 eediuu.

r nl.n 417 '-lt ':l!' t.v)L L.v+> 'r tlo,* tl6u mediu:n 20 r.0l-0 r,010

* calculated d-Dalues and indering uere dertoed fron the eomputer pwg"on of D. Applenot .. aaowning a monoclinic celL (T2ld). fntensittee utere esti,nated oisually. ** Valuee obtcirpd by Ealtn, Ean'den, qtd Broe\otilLLer (L956).

sufficientsites to fiIl the Ti layer accordingto Naka- probe analysis upon completion of the diffraction jima, Horiuchi, and Morimoto (1972).Morimoto and studies to confirm that the diffraction pattern was, Nakajima (written communication, 1973)have done in fact, that of heideite.Due to the small amount of further single crystal X-ray diffraction studieson the material available, the X-ray patterns are so poor structure of FeTirSn and find that there is a high- that d-valuesare givon only to the second decimal temperaturepolymorph with NiAs structure stable place (Table 4). Similarly, sincewe could not obtain above 540oC in addition to the low temperature diffractometer data, the intensity data for heideite polymorph whose structure was determined by (Table 4) are semi-quantitativeonly. Plovnick, Vlasse, and Wold (1968). The low form Debye-Scherrerdata werealso obtainedon heideite decomposesbelow about 300oCaccording to Naka- synthesizedat 650'C (Table 4) using Mn-filtered Fe jima and Morimoto. They state that decomposition radiation. Note the good agreement in d-values is extremelysluggish. between natural and synthetic heideite. Synthetic Becauseall grains of heideitethat we examinedin FeTirSn(data from Hahn, Harder, and Brockmiiller, Busteeare anhedral,there are no data on the external 1956)has similar d-values(Table 4). The d-valuesof morphology of the phase.The small grain size and heideiteare sufficientlysimilar to those of troilite to rarity of the phase preventedany attempt to study suggestthat heideite probably has a distorted NiAs heideiteby single-crystalX-ray techniquesbut X-ray structure. diffraction films were obtained using the Debye- A computerprogram written by D. Appleman that Scherrer technique using Ni-filtered Cu radiation. indexes and calculatestheoretical d-valuesand cell The heideitegrains were mounted for electronmipro- dimensionsfor an assumedsymmetry was used to HEIDEITE, A NEW MINERAL calculatethe best struotureand cell dimensionsfrom Tlsr-n 5. Unit Cell Dimensions, Cell Volume, and Calculated for Heideite the X-ray data (Tables 4, 5). The two polymorphs Density Natural and Synthetic

of FeTirSndiscussed above were chosenas possible Heideite HeideitesFthetic FeTi2s!* FeTi2S!*d structuresin addition to a number of other distorted 5.97 5,936 ,.94 5.921 3,\2 3.1+38 3.44 3.\26 NiAs structures.The best fit of the X-ray data for u.! il,!8 11.5 rr.163 heideite, synthetic heideite, and for Hahn's d-values 9f .0 90,0 To.8 70.10 for FeTirSnwas obtainedby assumingthe low temper- !. ol 3.9 3.993 polymorph FeTirSn (space group I2/m) ature of * FeTl^s,., after Hahn et d. (1956). *rlov Feti^sr after Morihoto dd sakaJi@, kitten coM. (19?3). I n"no- (Table 4). ' , obtained by Morimoto and Nakajima clinic " cell (I2/n) dd z=2 sere asshedl, after the vork of Morlnoto The fit is by no means perfect, perhaps becauseof ed Na}sJiDa (uitten cotu., 19?3) on los FeTi2SL. the poor X-ray data. One can positively state that heideitehas a distortedNiAs structureand is probably monoclinic (12/m), in which casethe cell dimensions a minimum temperatureof formation. Busteeferroan and cell volume are those listed in Table 5. This alabandite contains from about 0.4 to 1.3 mole structure is the same as that of brezinaite, Cr'Sn percent Ca; this indicatesa minimum temperatureof (Bunch and Fuchs, 1969),which is consistentwith the formation ranging from about 24fr"C to nearly possible solution of heideite toward brezinaite 500"C, which is approximatelythe range obtainedby reportedabove. Assuming that this structureis correct, Skinner and Luce (1971) for alabandite coexisting the calculated density is 4.1 g/cc; Morimoto and with oldhamitein enstatitechondrites. Nakajima (written communication, 1973) obtain a (5) Viaene, Kullerud, and Taylor (1971) have calculateddensity of 3.993g/cc for the low tempera- shown that at 800'C there is an extensive solid ture form of FeTirSn. solution which includes FeTirSn and compositions By analogy with the structure obtained for low toward the Fe-S join. This solid solution includes FeTi,Snby Nakajima, Horiuchi, and Morimoto (1972) heideite,and is not in equilibrium with metallic Fe. in which Ti layers are filled, we have written the Equilibrium betweenmetal and heideite appears to structural formula as shown in Table 3. The fact that be possibleat low temperatures.Either heideite was FeTi"Snprepared by Hahn, Harder, and Brockmilller formed under disequilibrium conditions at high (1956) at 600oC appearsto be the low temperature temperaturein a parent body that cooled rapidly, or form suggeststhat the high-low transition is rapid. heideite formed at relatively low temperatures.It is puzzling that no heideite grains are associatedwith Conclusions other sulfide or metal grains, and that its range of The following conclusionscan be made concerning composition is small. This suggeststhat heideite was the conditionsof formation of heideitein Bustee: formed as a discretephase at relatively low tempera- (l) The high Ti content of some troilite in Bustee ture that restrictedthe solid solution field for heideite. suggeststhat the meteorite cooled rapidly from Heideiteis certainlya mineral of limited o@urrenoe, elevatedtemperatures-16.1 wt percent Ti is surely but we speculatethat other examplesof it may be not stablein troilite at room temperature. found in other enstatiteachondrites. (2) The presenceof titanian troilites that do not lie on the FeS-TirS' join, but rather in the area Acknowledgments boundedby FeS,TiS, and Ti,S' (Fig. l) suggeststhat We thank Carleton B. Moore, Arizona State University, for both divalentand trivalent Ti are present,thus placing providing samples of the Bustee meteorite from the Nininger constraintson the redox conditionspertaining during meteorite collection; Daniel Appleman and Judy Konnert, powder pattern; J. the cooling of the Busteeparent body. U.S.G.S., for help in indedng the heideite Kocmaneck,then of the GeophysicalLaboratory, for performing (3) The broad range of Ti content of troilite in the DTA runs; Maxwell Blanchard, NASA Ames Research Bustee indicates that, at most, only conditions of Center, for assistancein X-ray diffraction studies; and George local equilibrium prevailedduring cooling. H. Conrad, University of New Mexico, for help in electron (4) Skinner and Luce (1971) have determined a microprobe work. We also thank N. Morimoto and S. Nakajima N. Morimoto for solvus for CaS in. alabandite in equilibrium with for allowing us to use unpublished data, and reviewing the manuscript. This work is supported in part by the oldhamite and troilite. Ferroan alabanditein Bustee National Aeronautics and SpaceAdministration, Grants NGL doesnot co-existwith oldhamiteand troilite, and any 32-C0+063 and NGL 32-ffi+O@ (Klaus Keil, Principal In- temperatureobtained from the abovesolvus represents vestigator). 470 K, KEIL AND R. BRETT

References niques lor High Pressureand High Temperature, Springer- Verlag, New York, Inc., p. 289-315. BuNcH, T. E., eNt L. H. FucHs (1969) A new mineral: NeKerrM4 S., H. HonrucHr, AND N. MoRrMoro (1972) brezinaite, CroSn,and the Tucson meteorite. Am. Mineral. 54, l503-1518. Crystal structures of (Fe,Ti)"Sn (abstr.). Acta Crystallogr. A2E. 61. H,lrrN, H., B. Henonn, eNo W. Bnocxvijr_rBn (1956) Provlrcx, R. H., M. VLAssE,eNo A. Woro (1968) Prepara- Untersuchungen der terniiren Chalcogenide. X. Versuche tion and structural properties of some ternary chalco. zur Umsetzung von Titansulfiden mit Sulfiden zwei- genices of titanium. Inorg. Chem. 7, 127-129. wretiger Ubergangsmetalle. Z. anorg. aIIg. Chem, 2EE, 260-268. SrrNNrn, B. J., eNo F. D. LucB (1971) Solid solutions of the type (Ca, Mg, Mn, Fe)S and their use as geo- KtIr, K. (1967) The electron microprobe x-ray analyzer thermometers for the enstatite chondrites. Am. Mineral. and its application in mineralogy. Fortschr. Mineral. 44, 4--66. 56, 1269-1296. VnrNr, W., G. Kur,r"rnuq eNo L. A. TAyLoR (1971) (1969) Titanium distribution in enstatite chon- Phase equilibria studies in the system Ti-S and Fe-Ti-S drites and achondrites and its bearing on their origin, (abstr.). Geol. Soc. Am. Abstr. Programs, Earth Planet. Sci. Lett. 7, 243-248. 3, 739. KuttERUr, G. (1971) Experimental techniques in dry Manuscript receited, Ianuary 14, 1974; accepted sulfide research. In, G. C. Ulmer, Ed., Researth Tech- lor publication, February 1, 1974.