<<

Canadian Mineralogist Vol.23, pp. 6l-76 (1985)

THE RELATIONSHIPBETWEEN , BONDING AND CELLDIMENSIONS IN THE GOPPERSULFIDES

RONALD J. GOBLE Deportment of Geology, University of Nebrasko-Lincoln, 433 Morrill Holl, Lincoln, Nebrosko 68588-0340,U.S.A,

ABSTRACT de cui'ire de structureconnue indiquent que la structure de la yarrowiteet cellede la spionkopiteressembleraient and copper- sulfidescan be classifiedinro surplusieurs points i cellede la covelline,mais on doit con- three generalgroups: (1) anilite, , geerite,, siddrercomme probable une occupation statistique des sites. , haycockite, , and bor- Lestentatives de ddtermination de la structurede la spion- nite with structuresbased upon approximatecubic close- kopiteont 6t€entravdes par I'imperfectiondes cristaux et packing of the atoms; (2) and le grand nombrede sitesde cuivrevacants. La stru4ure with structures based upon approximate hexagonal close- de la geeriteest rhomboddrique (fi32?), a 15.77A, a packing oftle sulfur atoms; (3) covelline, yarrowite, spion- 13o56', Z = l, et ressembleraiti celle de la digdnite. kopite and idaite with a combination of hexagonalclose- packing and covalent bonding of the sulfur atoms. The (Traduit par la R6daction) average spacing D between layers in all groups can be expressedas D = 2.M3 + 0.654(Cu:S) + l. 183@e:S). The Mots-clds:yarrowite, spionkopite, geerite, sulfurc de cui- ionic radius R of sulfur for group-l mineralsis R, = 2t7 vre, sulfuresde cuivre et fer, covellineblaubleibender, Q'l2A), where D is from the previous expression;for structurecristalline, liaisons. group-2 , Rz = 1.856+ 0.060 (Cu:S)+ 0.023 @e:S);for group-3minerals, R: = 1.857+ 0.039(Cu:S) - 0.020 @e:S). Consideration of bond lengths in co- INTRODUCTIoN ordination polyhedra of known copper structures indicatesthat major portions ofthe yarrowite and spionko- Eight copper-sulfideminerals have beenidentified pite structures will resemblethe covelline structure. vdth to date: covellineCul.6gs, yarrowite Cu1.12S,spion- probable statistical site-occupancy.Attempts at the deter- kopite Cu,.*S, geeriteCu1.6eS, anilite Cut.75S,dige- mination of the spionkopitestructure were hamperedby the imperfection of the crystals and the partial occupancy nite Cu,.*S, djurleite Cu1.eSand chalcociteCuz.mS. of most structural positionsof copper. The geeritestruc- Partial or completestructural determinationshave ture is rhombohedral(R3n?) with a 15.77A, o 13.56,. been carried out for covelline (Oftedal 1932, Berry Z = l, and will probably resemblethe digenitestructure. 1954,Kalbskopf et al. 1975,Evans & Konnert 1976), anilite (Koto & Morimoto 1970),digenite (Donnay Keywords: yarrowite, spionkopite, geerite,copper , et al. 1958,Morimoto & Kullerud 1963),djurleite copper-iron sulfides, blaubleibender covelline, crystal (Takedaet al. 1967a,b,Evans 1979),and chalcocite structure, bonding, (Sadanagaet al. 1965,Evans l97l). The structures of yarrowite, spionkopiteand geeritehave not been SoMMAIRE determined.Yarrowite and spionkopite,two of the blaubleibenderor "blue-remaining" covellines,were On peut regrouperles sulfuresde cuivre et de cuivre r for many yearsdescribed in terms of the hexagonal fer en trois grandesfamilles: l) anilite, dig6nite,geerite, cubanite, chalcopyrite,haycockite, talnakhite, mooihoe- unit-cell of covelline,based upon similaritiesof X- kite et , dont la structure contient un empilement ray powder patterns(e.9., Frenzel1959, Moh 1971, approximativement cubique compact des atomes de sou- Rickard 1972,Putnis et ql. 1977),However, Goble fre; 2) djurl6ite et chalcocite,dont la structuremontre un (1980)has shown that the unit cellsof yarrowiteand empilementhexagonal compact (6rossomodo) des atomes spionkopite are not the sameas that of covelline. de soufre, et 3) covelline,yarrowite, spionkopiteet idaite, Geeriteis a pseudocubiccopper sulfide that has only qui montrent une combinaison de I'agencementhexagonal recently been reported (Goble & Robinson 1980). compacfdes atomes de soufre avecliaisons covalentes. Pour The known structurescan be divided into three lestrois groupes,la s6parationmoyenneD descouches est generalgroups based upon the natlue of packingof 6galed I'expression2.063 + 0.654(Cu:S) + 1.183(Fe:S). the sulfur atoms: (l) anilite and digenite,with struc- Le rayon ionique.R du soufre Cansles min6raux du pre- mier groupe est 6gal d D/2 l2A. Pour les mindraux ou turesbased upon approximatecubic close-packing, deuxidmegroupe, on a .rR2: 1.856+ 0.060(Cu:S) + 0.023 (2) djurleite and chalcocite,with structuresbased (Fe:S), et pour ceux du troisiime groupe, upon approximate hexagonalclose-packing, and (3) R: = l.8SZ + 0.039 (Cu:S) - 0.020 @e:S).Les longueurs covelline, with a combination of hexagonalclose- de liaisons dans les polybdres de coordination des sulfures packing and covalentbonding of the sulfur atoms. 61 62 THE CANADIAN MINERALOGIST

The resemblancebetween the well-developedyar- geeritefrom the studiesof Goble (1980)and Goble rowite and spionkopite subcellsand the covelline & Robinson (1980)were used for all single-crystal unit-cell (Goble 1980)suggests that theseminerals patterns. Precession,Weissenberg and integrated belong to group 3; the pseudocubicnature and Weissenbergphotographs were preparedfor selec- resemblanceof the geeriteunit-cell to a structurepro- ted orientations of the yarrowite and spionkopite ducedby the leachingof anilite (Goble 1981)sug- reciprocallattices. Precession photographs of covel- gest that geeritebelongs to group 1. Single-crystal line in the equivalentorientations were also prepa- X-ray studiesprovide a method of examining the red in order to checkthe supposedstructural simi- structuresof yarrowite, spionkopiteand geeriteand larity of the blaubleibendercovellines to this . of determiningthe true relationshipbetween these For geerite,only precessionphotographs were pre- structure$and those already known for other cop- pared. Filtered Cu-radiationwas usedfor all Weis- per sulfides. senbergfilms and Mo radiation for all precession films. The fragmentsof yarrowite, spionkopiteand PRoCEDURES geeriteused measure approximately 0.14 x 0.08 x 0.03,0.25 x 0.08 x 0.08and 0.07 x 0.06 x 0.02 Cleavagefragments of yarrowite, spionkopite and mm, respectively. For intensity determinations,0-, l- and 2level integratedWeissenberg photographs (rotation axis c), were taken of the yarrowite and spionkopite frag- ments.Exposure times for all levelswere on the order of 200 hours using a Philips fine-focus Cu X-ray tube (point-focus port) with a Philips PW 1008/85self- rectified high-voltagepower supply operatingat 40 Ft.c kV and 12 mA. Three-film packswere used, the films f being separated by one sheet of black paper. ]t Individual films of eachpack werescaled using the "film factor" of Morimoto & Uyeda(1963). Differ- ent levelswere scaled by a comparisonof reflections present on two or more of the levels, with the A. COVELUNE B. COVELUNE 0120)ptone assumptionthat the unit cellsare approximately hex- agonal. Intensitieswere estimatedby visual comparison with the {110} reflection of covellineon standard- scalefilms preparedfrom multiple exposures.Unob- servedreflections were assignedmaximum intensi- ties of one-half the minimum observableintensity at that point on the film. Raw intensity-datawere correctedfor Lorentz and polarization effectsand, in the caseof spionkopite,for the presenceof minor (= l59o) intergrownoriented yarrowite; absorption I*I correctionswere not applied. Wilson plots (Wilsgn 1942)show^overall temperature-factors of 0.45 A2 T l_Jl and 0.75 A2 for yarrowite and spionkopite, re- IT tY I spectively. -t_ l..-/...... ---.....n)-) A C.'ANIUTE" D. E. METASTABLE (100)phne DI6EMTE STRUCTURALDATA Frc. 1. The crystal structure of covelline(after Wuensch group-3 1974),"anilite", sphaleriteand metastabledigenite. The The crystal strucure of covelline, a "anilite" structureis an idealizedversion (see text). In mineral,is shownin Figure la (after Wuensch1974); B t,Cu and tvCu indicate copper atoms in triangular equivalentsites for all atomslie on the (l l0) plane, and tetrahedralco-ordination polyhedra, respectively; and the structure can be representedby a section S. indicates covalently bonded sulfur; lC and lI through the unit cell on this plane as shown in Figure representthe spacingsof one covalentlybonded and one lb. The crystal structure of anilite, a group-l (ionically bonded) layer, respectively.Filled tetrahedral mineral, can be representedin a similar fashion if in A, B and D indicate fully occupied triangular circles & Morimoto (1970)are co-ordination polyhedra; filled and open triangles in C the structural data of Koto and E indicate fully and partly occupiedtetrahedral co- transformed into an idealized cubic close-packed ordination polyhedra,respectively, with atomic displace- unit-cell. Copperatoms in triangularly co-ordinated ment toward the four surroundingtriangular faces. sitesare assignedto the tetrahedra of which these CRYSTAL STRUCTURE OF THE COPPER SULFIDES 63

a11.tlt3

Oooe

r663 Oli3 a88e

a-LZ1

.057

.T7.nn 4776 )zza

.n.n.i o635 )zzt

olo.fit.m a$r

415.15.15 Ods-0 oiiS

.t3.f3.tz a66s-

an.ji.i3 a7.t7 i" FIc. 2. The reciprocallattice of geeritewith rhombohedralindices. Both rhombo- hedral and cubic crystallographicaxes are indicated. sites form a face, and displacementof the copper cil of Canada,Ottawa, Ontario KIA 0S2.For yar- atoms from thesetetrahedrally co-ordinatedposi- rowite, I 17nonequivalent reflections were observed, '(idealized" tions areignored. A sectionthrough this 623 were unobserved(only lF1,e1lvalues are listed structure of anilite perpendicularto the close-packed in Table l, lF*:'l = lFnrrl). For spionkopite,147 layersof sulfur is shown in Figure lc; reduction of nonequivalentreflections were observed,406 were the cell dimensionsnecessitates averaging full and unobserved;all supercellreflections of the tpe hkl empty tetrahedraand resultsin partial (%) occupancy wfih h or k + 6n were ignored (i.e., the by copper for some tetrahedra. The structure of a' : a/6 = 3.8?i/A subcellwas used). Examination metastabledigenite (Morimoto & Kullerud 1963)in of the structure factors indicatesthat the spacegxoup the equivalent orientation is shown in Figure le. The for both yarrowite and spionkopiteis one of .P5ra1, crystal structuresof djurleite and chalcocite,group-2 P3ml, P321,Hlm, Hlm or Bl2. minerals,involve extensivetriangular co-ordination Geeritewas describedby Goble & Robinson (1980) of Cu and cannotreadily be representedon sections asbeing pseudocubic (possibly orthorhombic), space suchas lb, lc and le. Becauseofthe strongresem- group F43m. Careful re-examinationof precession blance betweenthe X-ray patterns of geerite and photographsshows that geeritecan be assignedcon- sphalerite,a sectionthrough the sphaleriteunit-cell sistentindices based upon a rhombohedralunit-cell in this orientation is representedin Figure ld. similar to that proposedfor digeniteby Donnay e/ Goble (1980,Figs. l, 2) showedthat the recipro- cal latticesof yarrowite and spionkopiteare similar to but distinct from that of covelline.Yarrowite X- TABTE3. ANDCUBIC II{DICES FOR I1IE GEERITEPOIIDER IATA ray data were lndexed on a hexagonal cell with a 3.800,c 67.26A; spiopkopitewas show4 to be hex- r/r1 Rhmbohedral Ind'l @s Cubic lndices qsonal, with a 22.962A (i.e., 6 x 3.8n X), c 4t.429 3. t28 100 221, 555 A. Well-developedsubcells with the approximate 'I.9182.712 l0 334 50 776 {imensionsof the covellineunit-c€ll (a 3.796,c 16.36 1.870* l0 A) were noted. Tables I and 2, which list the l0 '1.576 30 ll3,889 3ll observedstructure-factors for yarrowite '10l0 442,10.10.10 222 and spion- 1.247 997,15.15.t5 kopite, have been submitted to the Depository of 1.109 20 il2, 13.t3.14 422 UnpublishedData, CISTI, National ResearchCoun- rrcflectlon not obseryed on sJngle-crystal patterns 64 THE CANADIAN MINERALOGIST o a

o o O

a Oor o o w36ut72n

O o a 1u11u32u1. orarOsvtovlev, a a hlflrrr o a hl01c o a ' o ouraraoz!'or ln:rnifr.OO, oT cilTtVtTn o o o ota o aoautsznln o

a o a

a o

FIc. 3. The reciprocal lattice of anilite after leachingin ferric sulfate solution for twelve hours. Rhombohedralindices consistent with the geeriteunit-cell are shown. Open circleslabeled T representreflections attributed to the secondindividual of a twin. a/. (1958).The geeritereciprocal lattice, indexedon Relationship between structure and composition this basis,is reproducedin Figure2; the X-ray pow- der pattern with both pseudocubicand rhombo- The five known structures have hedral indexingis givenin Table 3. Geeriteis rhom- either approximately cubic or hexagonal close- bohedral Cu6S5(diffraction aspect R*1) with a packedlayers of sulfur or, in the caseof covelline, 15.77A, a 13"56', Z = l, spacegroup R3m, R3m, a combination of hexagonalclose-packed with cova- or R32. These data are consistentwith the weak lently bonded layers of sulfur. In an ideal close- bireflectanceand moderateanisotropism observed packed netwqlk the distance D between layers is by Goble & Robinson (1980). equalto 2R l2A, whereR is the radius of the sulfur The choice of the rhombohedral cell for geerite atoms. In copper sulfides such as covelline, with is substantiated by the leaching experiments con- ductedby Goble (1981),which produceda structure similar to that of geerite.The reciprocal lattice of TABLE4. AVERAGEDISTAI{CE BnIIEEII SULruR TAYERS (O) AI{ORADIUS OF this material is shown in Figure 3; indicesrefer to SULFURAIO.IS UTIFIN SULFIJRTAYERS (R) IN C'OPPERSULFIDES the geeriterhombohedral cell. Fractional indicesindi- cate that the leachedphase has approximately tJree nlreral D (8) hkr R (i) hkr data source times the a dimension of geerite (a,1,47.L2 A, a @ell'lm 2.735 006 1.8971 ll0 Poltet & Evans(1976) 4"40'). Careful re-examinationof reflection inten- vafMlte 2.7 0.0.24 1.899 ll0 Goble(1980) ;plonkopite 2.964 0.0. 14 1.910 660 Coble (19B0) sitieson precessionfiLns of leachedanilite showsthat geerite 3.128 555 1.918 776 Table 3 anlllte 3.20B [email protected] 1.%3 040,400, Potter & Evans (1976) reflections indicated by open circles in Figure 3 are 224 dlgenlte 3.216 555 1.969 10.10.0 ilorlrcto & Kullefud (1963) producedby twinning about [00]*]" or [110*]". 3.200 1.959 oonnayet al. (1958) dJur'lelte 3.62 BO0 1.9556 046 potter & Evans (1976) Converting the rhombohedral cell tp a hexagonalcell chalcoclte 3.373 h4 1.974n (30 Potter & Evans (1976)

producesa length cpu of l4\.2 A, approximately pl'nes For anilite D is taken on the averageof the four nonDarallel-il1!i twice that in yarrowite (67.26A), thus explaining the in the pseudocublc cell, that-ls, the orthoah@bic (0221 arld 12021 olaFs rith sDacinqsof 3.198 X and 3.2i8 A, respectively; R ls resemblanceof the two phaseson X-ray powder pat- i.aken as the rciqhted averaqe of the six mn-parallel t440' pseu@- cubic planes, th;t ls,-the odiorh@blc (40!), {040}, an! t224} plares terns, a$ noted by Goble (1981). xith spacingsof 1.977X' 1.953I, and 1.962A. respectlvelv. CRYSTAL STRUCTURE OF THE COPPER SULFIDES 65

layersof covalently bonded sulfur, D will be an aver- age of the covalent, C and close-packedor tetra- hedral, I interlayer spacings,weighted as to the rela- tive number of each (see,for saamplelFig. l). As noted by Goble (1981), planes with spacingscor- D (A) respondingto R and D are readily identifiable in the copper sulfides; theseare listed in Table 4 and shown as functions of compgqition in Figure 4. The ideal relationship,D = 2R.l2A,is also shown in Figure 4. For spionkopite and geerite, the compositions Cu1.a6Sand Cu,.6oSbetter fit the data of Figure 4 than the compositions Cu1.r2Sand Cu1.53Sdeter- mined from the microprobe data of Goble & Smith (1973),Goble (1980)and Goble& Robinson(1980). DigeniteD and djurleite R valuesare anomalouswith respectto the data of Figure 4. R6= 1.860.0.058(Cu:S),' Figure 4 confirms that geerite has approximate --2-;lg-z-? close-packingof sulfur atoms, whereasyarrowite and ;oiJ_ spionkopite have structures with a combination of t close-packing and covalent bonding of the sulfur ,, t atoms. In Table 5 the spacingbetween close-packed R (A) = .12A, vfui-\, layers,D' 2R has beencalculated from the '--'l' R values of Table 4 and combined with the spacing R,= 1.850.0.036(Cu:s) betweenlayqrs of covalently bonded sulfur in covel- I line, 2.071 A @vans& Konnert 1976), in order to determine the number of each type of layer parallel to the c axesof yarrowite and spionkopite. Compara- ble data are presentedfor covelline, geeriteand anil- ite. The geerite21d anilifs data are consistentwith pseudocubicunit-cells containing three and six layers 1.0 1.2 1.8 2.0 of close-packedsulfur, respectively; the yarrowite ti,, '."1r" and spionkopite data are consistent with a number Frc. 4. The relationship(A) betweenaverage distance D of different combinations of covalently bonded and betweensulfur layers and composition and @) between close-packedlayers corresponding to different unit- radiusR of sulfur atomswithin sulfur layersand com- cell contents. positionfor thecopper sullides (after Goble l98l). Open Becauseof the small sample-size,the presenceof circlesfor spionkopiteand geeriteindicate analyzed (Cur.:zS,Cu1.53S); for impurities and the scarcity of yarrowite compositions open circles and digenite(Cu1.6S) indicate two possibledistances; open spionkopite, no attempt was made to measuretheir possible yersar circlesfor djurleiteindicate two compositions density. Instead, the composition density rela- (Cu1.e3S,Cur.c7). Correlations were determinedby tionships for known copper sulfides shown in Table omittingall datarepresented by opencirclc exceptthose 5 and Figure 5 wereused to determinethe most prob- for digenite,which were averaged. The dashedline in able densityof yarrowite and spionkopite.Compo- B representsthe equationderived in A if ideal close- sitions of Cu1.,2Sand Cu132Swere used, as deter- packingof sulfur atomsis present. mined by Goble & Smith (1973)and confirmed by Goble (1980). A spionkopite composition of Cu1.asS,consistent with the data of Figure 4, was with lessthan 1.75 copper atoms per sulfur atom, also used.For spionkopite,the data of Table 5 and there is an increasingdegree of developmentof cova- Figure 5 are consistentwith a unit-cell content of 14 lent sulfur-sulfur bonding with loss of copper, as Cu,.rr_,.*S;for yarrowite, cell contentsof either 24 shownby the data of Table 5. Thesedata werecon- Cus.12Sor 25 Cu1.12Sare acceptable(although 24 vertedto aratio of thenumber of interlayersof cova- Cu1.12Sseems more likely), depending upon the lently bondedsulfur to the number of close-packed choice of the density-compositionrelationship in or tetrahedrally co-ordinated interlayers, C/7, and Figure 5. plotted as a function of composition in Figure 6. For The difficulty in the choice of a unit-cell content spionkopitea C/Tvalue of 0.167,corresponding to and density for yarrowite was resolved by examin- 2 covalent and 12 tetrahedral interlayers in the cell, ing the relationship betweenratios of the differently is consistentwith the data; for yanowite, of the three bonded types of layers (close-packedor covalently possibleC/T values,only 0.412, correspondingto boqded) and composition. In those copper sulfides 7 covalent and 17 tetrahedral interlayers, is consis- 66 THE CANADIAN MINERALOGISI

TABLT 5. MSSIBLE NUI.IBERSOF COVALEIIY BONDEDSULRJR LAYERS AI{D CLOSE-PACKEDSULFUR LAYERS IlI ONEHEIAGOML UNII CELLOF STIECTEDCOPPER SI'LFIDES 0. olss oss

uoor{i) D' (A) nC + nI o*r.(i) ' i;i*jj pt 0./. covell lle 16.341 3.098 zc+ 47 4.68 9le lamrlte 67.26 3.101 lC { 217 67.L9 22 4.48 -t>oto 4Cr 197 67.20 23 4.69 7CI I7f 67.21 24 4.89 0.3 10cr l5T 67.23 25 5.09 :l: splonkoplte 4t.429 3.119 2C 1 l2'l 41.57 14 5C+ 107 41.55 15 5.49/5.74 atr 8c+87 4t.52 16 5.8/6.12 o23S geer'lte 9.37 3.132 0c+ 3T 9.40 3 3.42/5.61 EI;0.2 anll lte 19.24 3.189 0c+67 19.14 6 dlgenlt€ 5.63/5.71 014S dJurlelte 5.76/5.82a EE chalcrclte 5.79 5+0.1 clo 9ob;: obseryeda cell dlrensloni geerlte and anlllte cells ,ere converted olq b,thelr hexagonalequtyalents. R: observedrad.lus Els of sulfur atons wlthln >to o22S sulfur layerslU': colculated dlstsnce betreen lryers of close-Dacked ol sulfur (" 2Rr'2l3i"R.frm Table 4). C: layers of covalenily bond;d sulfuri ot spacing ls 2.071 A (Evanst Konnert 1976). T: layers of cl;se-Dacked 0.0 'lengthiSulfuri spoclng ls Dr A. nC + eT: n@ber of C and T lavers ln'one e 1.2 1.tl 1.6 other values of n lead to nonlntegral values oia. o.,t.: calcu- loted-d cell dimnslol. Z: nmber of suliur atom in lexagon'ii"unii-ceit; I Cu:S rotio - the a' a/6. 3.827 A subcell of splonkoplte ls used. Deislty is calcu- lated frcn unit-cell parareteE. Densltles of splonkopjte apfor Cu1.32S/Cu1.asS.Densitles of geerlte are for Cur_crSTCurrnS. Data Ftc. 6. Ratio between the number of layers of covalently sources are as in Table 4 ercept for dJurlelte (at;-iaken-lr6m rakeda at a/. (1967b1. bondedsulfur to the number of close-packedor tetra- hedral sulfur layers (ionically bonded) as a function of composition. Different points shown for yarrowite, spi onkopite and geerite representdifferent possiblestruc- tures and compositions. 5.8 GPtss TABLE6. AVERAGEDISTNCE BfllEEN sU.FURLAYERS, D, ANDMDIUS OF SIJI.FURATOMS HITHIN SOFUR LAYERS,R, IN COPPER-IRONSULFIDES 56 {/ oineral Fe:s D (i) hkr R data cu:s d) $urce E -u sllfur 0.0 0.0 2.048 calc 1.856 calc 1,2 cubanlte 0.33 0.667 3.12 002 1.867 123,330 3 ;'|5'4 chalcopyr'lte 0.5 0.5 3.03 lI2 1.865 220 3 6Pts haycocklte 0.5 0.625 3.07 226 1.889 440 4 talnakhlte 0.55 0.55 3.06 222 1.874 4N 5 @lh@klte 0.5625 0.5625 3.07 221 1.87',t 4& 4 ldalte 0.833 0.167 2,82 006 1.890 uZ 6 s.z 'idalte ? 0.845 0.155 2,792 0U 1.887 l',lo 7 fi bonlte I.25 0.25 3.18 224 (uc 1.937 440,408 3 T' The value of D for sulfur ls tlE average covalent S-S dlstane deter- nlred tun the strocture of o.thorhotrblc sulfur. The value of R for 5.0 sulfur ls half the mlnlmm lntemolecular dlstance of Ab.ah@ (1955) scaled to allow for the revlsed cell-parileteB ol cooper et al.(196x). olta sources: (1) lbrahms (1955)i (2) Cooperet al. (1961)i (3) Berry & Tlmpson (1962); (a) Cabri & Hol1 (1972); (5) Hiller & Prcbsthaln (1956); (6) Frenzel (1959)i (7) Yund (1963).

oY%rs The crystal-structure - composition relationships shown in Figure 4 can be extendedto include sulfur 1.0 1.2 1.t, 1.6 1.8 2.0 and copper-iron sulfides,as shown in Table 6 and Cu:Srotio Figure 7. In Figure 7a the interlayer distance, for is plotted as a function of the Ftc. 5. Calculated density of the copper sulfides func- copper-iron sulfides, as a (Cu plotted tion of composition,Filled circlesrepresent values cal- ratio of metal + Fe) to sulfur. The lines culatedfrom known structures;open circlesrepresent are for different ratios of copper to total metal and possibledensities for yarrowite and spionkopite. Differ- satisfy the equation: D = 2.O63+ 0.654 ent points shown for a single mineral represent differ- (Cu:S)+ 1.183@e:S), R2 = 0.993.In Figure7b the ent possiblecompositions for this mineral. The two ionic radius R of sulfur for the samecopper-iron curves represent the range fo densitiesto be expected. sulfides is shown as a function of metal-to-sulfur ratio. Each copper-to-metalratio is representedby three straight-line segments.The central high-slope tent with the data. Once again, compositions of segment o!_each line represents the equation Cu1.a6Sand Cu1.66Sfor spionkopite and geerite bet- R: D/Q V%) plotted for the valuesof D in Figure ter fit the data than do the analyzed compositions 7a, and was used to determine whether a given sul- of Culj2s and Cu1.sS. fide belongs to the "low-metal" (eft) or "high- CRYSTAL STRUCTURE OF THE COPPER SULFIDES 67

3./,

32

30

:sJ8 o2.6

2.1

22

02 0.1 0.6 08 1.0 12 1A 16 t8 29 (Cu.Fe):S

1.98 1.0(r) Rl = 1.857.0.039(Cu:5) - 0.020(Fe:S) Rh= 0a3(4 1.96 1.856.0060(Cu:S) " 0.023(Fe:5) R =Dr(2lm)

1.91 0.5(.)

0.33(+) _1.92 :g Et.9o 0.0

1.88

1.86 {------

1.81t 0.0 q2 0.1 06 0.8 1.0 12 1! 16 18 2p (Cu'Fe): S Ftc. 7. The relationship (A) betweenaverage distance D betweensulfur layers and compositionand (B) betweenradius R of sulfur atoms within sulfur layersand compositionfor the copperand copper-iron sulfides.Values of the Cu,z(Cu+ Fe) ratiorepresented are 1.0 (-.-),0.S3 (-4. .),0.5 (-+),0.33 (..+..) and 0 (-). The open diamond in B representsa Cu,/(Cu+Fe) value of 0.44,buta separateset of curvesfor this value is not included; this point is obscuredby a closeddiamond in A. In B, each group of minerals with a singlevalue of the Cu,/(Cu + Fe) ratio is representedby tlree straight-line segmpqts;the middle, high- slopesegment represents the ideal relationshiF,Rc: D/2,JrA, with D taken from A; the high-Cu (right-hand) segment represenrsthe calculated relationship Ra: 1.856+ 0.060 (Cu:S)+ 0.023 (Fe:S); rhe low-Cu (efr-hand) segment representsthe calculatedrelationship Rr : 1.857+ 0.039 (Cu:S) - 0,020 (Fe:S). metal" (right) eroup of sulfides. Approximate equa- The positions of the lines shown are more tenuous tions usedfor the low-metaland high-metalcopper- than in Figure 7a becauseof the possibility of a sul- iron sulfidesare: R1= 1.857+ 0.039(Cu:S) - 0.020 fide having a behavior similar to that of high-Cu sui. (Fe:S), R2:0.992, and Ra: 1.856+ 0.060 fides (anilite to chalcocite)or low-Cu sulfides (covei- (Cu:S)+ 0.0234 (Fe:S), R2:0.999, respectively. line to geerite). 68 THE CANADIAN MINERALOGIST

A,COVELUNE "--.-....----.- L-'F \ O /' -\ ,J flc aq \ /a €> $"/ e F^$ =.4 -ua /fi( /cca \*6

FIc. 8. /20/Patterson projections for covelline,yarrowite and spionkopite.Peak posi- tions are indicated by filled circles.

In Figure 7, an increasein iron content resultsin an increasein D but a decreasein R. This is due td the presenceof iron in the structuresinhibiting the formation of covalentS-S layers(1.e., covelline and chalcopyriteboth have a metal-to-sulfur ratio of one, but only covelline has layers of covalent sulfur). BecauseD is an averagespacing and doesnot include any of the small covalenti^nterlayer distances in the copper-ironsulfides (- 2 A for a covalentinterlayer as opposedto = 3 A for a tetrahedralinterlayer), -u an increasein iron content increasesthe averagespa- Y cing of interlayers.Also, the lack of covalentlayers U o- resultsin fewer metal atoms, on average,per tetra- U hedral interlayerin the copper-iron sulfidesthan in F the correspondingcopper sulfides, and a decrease u in R. As an example,covelline and chalcopyriteboth E have a metal-to-sulfur ratio of one. D for chal- /lL copyrite is the averageof six tetraJnedralinterlayers ..ANtLIIE" of approximately3 A each,or 3 A; D for covelline is the average oj four tetrahedral interlayers of approximately 3 A eqchand two covalent interlayers of approximately2 A each,or 2.7 A. Similarly,in chalcopyrite there is one metal atom for eachsulfur or one per tetrahedralinterlayer, whereasin covel- 0 2 1 6 8,(A)10 12 14 16 l8 20 line, there are six metal atoms spreadthrough four tetrahedral interlayers or one and one-half per tetra- Ftc. 9. Pattersonsections along x: 0,./ = 0 for coveline, hedral interlayer. This increasesthe S-S distanceand, yarrowite, spionkopite,sphalerite and "anilite". Lines therefore,the radipsof sulfur R from 1.80A in chal- indicate the interatomic distancesof Table 7. copyrite to 1.90 A covelline. CRYSTAL STRUCTURE OF THE COPPER SULFIDES 69

It is interestingto note that the minerals shown in Figure 7b tend to clusteron or near points at which the high-slopeR*. Iines or extensionsof the R"d" lines cut the observedR1 and R7,lines for the vari- ous metal-to-sulfur ratios, evenif the various lines havedifferent valuesof the ratio Cu:(Cu + Fe) [i.e., idaite lies at the point where the R.4. line for a B.YARROWITE Cul(Cu + Fe) value of 0.33 cuts the R, line for a Cu/(Cu + Fe) value of 0.331.These presumably representparticularly stable structural arrangements in the copper-iron sulfides.It is also interestingto note that the chalcopyrite,mooihoekite, haycock- UJ C.SPIONKOPITE ite, talnakhite group of mineralsrepresents a series I extending along a single Rcal" line at Y u Cul(Cu + Fe) :0.5 (haycockite is actually on a L u slightly different line). A secondseries extends from : anilite and digenite to geerite.L. Whiteside (pers. comm. 1983)has observedfive intermediatestruc- trl E tures in a leachingstudy of membersof this series. In Figures4 andT, the spacingsR and D represent the radius of sulfur atoms and the averagespacing betweensulfur layersin the copperand copper-iron sulfides,respectively. For the coppersulfides in the compositionranges Cul.e_,.6Sand Cur.rr_2.6S,R would appear to vary continuously. However, although not representedin this way in the figures, D must vary in a stepwisefashion sinceit depends partly upon the number of times a constant quan- 0 2 4 6 I -10 12 11 16 l8 20 tity, the spaqingof layersof covalentlybonded sul- z (A) fur (= 2.07 A) is averagedwith a variablequantity, FIc. 10. Pattersonsections along x = a/3, y = 2a/3 for 2R l2A. A similar relationship would apply to the covelline,yarrowite, spionkopite, sphalerite and "anil- various compositionranges shown for copper-iron ite' ' . Linesindicate the interatomic distances of Table7 . sulfidesin Figure 7. Theiefore, for eachmineral in Figures4 and7, D will be nearly constant,particu- larly becauseof the limited solid-solutionshown by of sphalerite,which are also presentedin Figures9 most copper and copper-iron sulfides. and 10. Representativeinteratomic distancesfrom Iron sulfide data are not shown in Table 6 and the covelline,sphalerite and "anilite" structuresof Figure 7. Both and pyrrhotite have octahedral Figure I are listed in Table 7 and indicated in Figures Fe co-ordination and do not fit the data presented. 9 and 10. The sphaleriteand "anilitd" distanceshave TheoreticalFeS in the sphaleritestructure is consis- beenscaled to the covellinecell and are therefore dis- tent with the data of Figure 7a but not 7b. placedslightly from the positions of the Patterson peaks. Pqtterson synthesesand space-group determination Severalfeatures are immediatelyapparent on the yarrowite, spionkopite, covelline, sphalerite and Pattersonprojections on (010)for yarrowite and "anilite" one-dimensionalsyntheses: spionkopite are shown in Figures 8b and c (space a) The Pattersonsyntheses of yarrowite and spi- groupBml); filled circlesrepresent Patterson peaks. onkopite are, in general,intermediate in form with A similar projection for covellineprepared from the respectto the covellineand "anilite" syntheses.This data of Berry (1954)is shown in Figure 8a. One- is to be expectedbecause these minerals are inter- dimensionalPatterson syntheses prepared for lines mediate in the alteration sequenceanilite - spion- throughx: 0, a/3 and?-a/3for theyarrowite,spion- kopite - yarrowite* covelline and presumably kopite, covelline ldata from Berry (1954)]and "ideal- representstructures intermediate to those of anilite ized" anilite unit-cells are shown in Figures9 (x = 0), and covelline.However, the Pattersonsyntheses for l0 (x : a/3) and I I (x = 2a/3), Becauseof a lack of yarrowite and spionkopite more closely resemble reliable data, Pattersonsyntheses could not be pre- those of covellinethan those of "anilite". Patter- pared for geerite.However, the strong resemblance sonslntheses of yarrowite closelyresemble those of of the geeriteand sphaleriteX-ray patternssuggests covelline,whereas spionkopite synthesesmore closely that the Pattersonsyntheses would resemblethose resemblethose of sphalerite. 70 THE CANADIAN MINERALOGIST

b) Sets of pseudomirror^planes ile present at A.COVELUNE approxir4ately16.8 and 8.4 A for yarrowiteand 17.6 and 8.8 A for spionkppite.These reflect the presence of the I 6.8 and I 7.6 A subcellsnoted by Goble(1980) B.YARROWTE and puggestpossible additignal subcellsat 8.4 and 8.8 A. The 16.8and 17.6A subcellsshow a closp correspondencein peak positions with the 16.36A unit-cell of covellineand explain the application of C.SPIONKOPITE F the covelline unit-cell to the bloubleibender covel- (, lines by earlier workers. A similar set of pseudo- U mirror planesexists on the Pattersoq synthesesof Y "anilite", representingthe basic 5.5 A cubic close- tl D.SPHALERIIE (:seite) packed subcell. u c) The strongpeak at a z of 8.210 8.8 A Q = c/2 for covelline)is shifted somewhatfrom x = a/3 in u the covellinesynthesis to x: 0 in the yarrowite and E spionkopitesyntheses. The shift is more pronounced in the spionkopite than in the yarrowite syntheses in the syntheses. E:ANIUTE" and is further developed "anilite" d) The Patterson synthesesof yarrowite and spion- kopite show small peaks at x:a/3, z:0, cor- respondingto triangularly bondedcopper atoms. The small sizeof thesepeaks and the lack of suchpeaks on P6/mmm Patterson maps indicate that space 0 2 1 6 8 .10 12 11 16 t8 20 groups having equivalentpositions with x differing (A) z by a/3 ard zequal(Hlm, P312,P3lm) arenot pos- siblefor theseminerals. Therefore, the spacegroup = y = FIo. 11. Patterson sections along r 2a/3, a/3 for for yarrowiteand spionkopitemust be oneof Hml, covelline, yarrowite, spionkopite, sphalerite and Hml or P321. "anilite". e) The peak at x = 0; z = 4.6 A on the Patterson synthesescorresponds to the distancebetween two copper atoms occupying apex-sharingcopper-sulfur tetrahedra(i.e.,iucu-iucu in covelline).It is moder- ately strong in covelline,very strong in "anilite", TABLE7. OISTAI{CESBFmEEN ATOl'lS IN THEcOvELtINE, SPHALER]TE A{D'NILI'TE' S1RUCTURES but weak in both yarrowite and spionkopite. This suggeststhat in yarrowite and spionkopite the fac- distance befreen atons along z-*t. (E) occupiedcopper-sulfur tetrahedra,rather than - nJneral atons Ar,Ay, " 0,0 atms Lx'Af ol3'2a/3 ing alternatelyup and down along c as in covelline and "anilite" (seeFig. l), face either up or down !c-sc 2.07(: lC) s_tinc, 0.00 s--atcu 0.76 along c in a given region of the cell. Ai'ss.:.'lsr.. Lrcu-i.i.nc! 2.29 LzrCi-. Cu i:ii5.88 Sc-z,Cu 2.83 f) There is a generalbroadening of peaksin the Pat- 5c-5c 6.10 s--s 3.05(= lT) yarrowite !c:sc 8. t7 irci-iucu 3.60 tersonsyntheses of both and spionkopite, Cu-S 8.17 S-:S 5.12 s:-1ucu 5.34 suggestingthat there may be considerabledisplace- !i ",Cu 5,8a from the ideal tetrahedral and s.-zocu 7.41 ment of copper atoms aocu_arc! g.13 triangular co-ordination polyhedra found in covel- 5-s 8.17 line and in the idealized "anilite" structure of Figure sphale|ite s-d'cu 2.2g s-ducu 0.76 s-< l-u5 l. This will probably be reflected in statistical fuiu-ducu 3.05 "cu-s 5.34 occupancyof the four triangularly co-ordinated sites "anilite" 5.d'c,t 2.2g ircu-s 0.76 surrounding a given tetrahedrallyco-ordinated site ""Cu-LuCu 4.58 5-S 3.05 Lttu-Luu 3.05 in the structure. S-",cq s.34 Lrc!-Lrcu 7.62 Relationship between structure qnd bonding iiicu ls @pper ln trlangular coordinat'lon. ircu ls coppe|in tetrahedral coordlnation. Sc ls covalently bondedsulfur. Structureshave been proposed for covelline,ani- S is lonlcally bondedsulfur. itcu-s The ordJr of ltms l4-.aton palrs -1garblt.ary (i.e., may refer lite, digenite,djurleite and chalcocite.In the covel- to both distances ""Cu-S and S-""Cu). TtE dlstances listed for sphalerlte and "allllte" are based upon the line structure (Fig. l), Evans& Konnert (197Q deter- S-s and ,tcu-s dlstances of covelline (1.e., are mt to the sue scale as the unlt cells for these nlnerals). Thls result! in mined copper to be in tetrahedral (4 sites) and negatlve dlsplacments of these dJstances frm the Patterson peaks 'ln Flgures 5 and 6. triangular (2 sites) co-ordination, with Cu-S dis- CRYSTAL STRUCTURE OF THE COPPER SULFIDES 7l

tancesof 2.31 and2,l9 A, respectively.In the ani- IABLE 8. CALCULATDAND OBSERVED BOND.LA{6IHS IN COPPERSULFIDES lite structure (approximated in Fig. l), Koto & Morimoto (1970)found copper in distorted tetra- iil7,,_e iv",,_. ^ averaqe Cu-s data nlneral (olcul-aic-a, 8) (catiirlaiea, X) (olseivea, X) source hedral (8 sites)and triangular (20 sites)co-ordinatiory avelllne 2.'191 2.323 2.19,2.31 with Cu-S distancesvaryinC from 2.28 to 2.52 $ yarMlte 2.193 2,326 (weighted splonkopite 2.205 2.339 average 2.37^L) and 2.2.4 to 2.35 A ger'lte 2,215 z.Ug (weightedaverage 2.30 A), respectively.Of the 112 mll.lte 2.267* Z,cBA, 2.{,2,31 2 dlgenlte 2.262- 2.274* 2.399- 2.412* 2.29 3,4 possiblebonds in the tetrahedralco-ordination poly- djurlelte 2.255- 2.3.9,* 2.462 2.2' 2.31 5 - 2,2,2.31 5 hedrq 84 clusterclosely around a Cu-S distanceof chalc@lte 2.280 2.352* 2.477 2.30A, with therem*inder having distances of 2.52 Bond lengths are calculated for ideal closFpacked structures. *For anlllte the 'ld@llzed" cublc close-packed cell is used; A (8 bonds), 2.94 A (8 bonds), arLd3.22 L 62 varlatlons in the dlgenlte data reflect dlfferent data sources. warlations ln the djurleite and chalcoclte data reflect differences bonds). In the structure proposed for metastable ln locatlon and shapeof tie oordlnation polyhedra (parallel or digenite (Frg. l) by Morimoto & Kullerud (1963), obllque to the close-packed layers. copper atoms are in distorted tetrahedral co- Data sources: (l) Evans& Konnert (1976); (2) Koto & t{orlrcto (1970); (3) l'lorlmto & Kullerud (1963)i (a) Horlmto (1964); (5) Evans (1981). ordination, with copper atoms statistically displaced toward the triangular facesof the tetrahedra.Cu-S distancesare not given by the authors but, by anal- ogy with the relatedstructure of metastablebornire superscriptson the copper ions refer to the co- (Morimoto 1964),the averagedistance from the cop- ordination number: per atom to the thrye closestatoms of sulfur can be a) The formula of covelline can be written as estimated t'cu-s as 2.29 A, whereasthe averagedirlance (tcu+)4(',Cu2*)r(Sr),r(S,-)r; -2.32 A and from the fourth atom of sulfur would be 2.74 A,.In 'tu-S - 2.19 A, Overall charge-balanceis main- the djurleite and chalcocitestructures, Evans (1981) tained; chargebalance between the close-packedsul- determinedthat the copper atoms weremainly in dis- fur layersis also maintained, as shown in Figure l2a. torted triangularco-ordinatjon, with Cu-S {istances b) The formula of chalcocite.panbe written as varyingfrom 2.18to 2.90A (averageZ.3l A). Two (tiCu+)re2(S2-)n;'iiiCu-S - 231 L. overall charge- copper atoms in chalcociteand one in djurleite have balance and charge balance betweenthe layers is or approachlinear two-fold co-ordigation,with Cu- maintainedas shown in Figure l2b (modified after S distancesof approximately2.2 A. Evansl98l). Calculatedand observedCu-S bond lengthsfor c) The formula of djurleite can be written as (it,Cu+)601iigrz+)r(s2-)32for cul.eoS or as the copper sulfide minerals are presentedin Table * * 8. Observe{ bondJengthsclustef at approximately 1,tu lur1ttcu2)z(52) n for Cu1.e7S.Evans (l 979) 2.2 and2.3 A. Comparisonwith the calculatedbond- reported only 62 copper atoms in the structure, of lengthsin Table 8 showsthat yarrowite, spionkopite which one, Cu{62),is in linear two-fold co-ordination - and geerite will all readily accommodatecopper (tu-S 2.2 A); a second,Cu(13), would appearto atoms in both triangular and tetrahedral co- approachlinear co-ordination. The remainder of the ordination without sienificantly distorting thesesites. copper ions is approximately in trianeular three-fold - Assuming that the differencein size of the copper co-ordination (ttcu-S 2.3 A). As shown in Figure atoms in thesesites reflects a difference in charge l2c (modified after Evans l98l), whether or not (l'.e., Cu+ being the larger and Cu2+ the smaller overall charge-balanceand a balance of charge ion), it may be possibleto predict the relativedistri- betweenthe layersare maintainedwould dependon bution of copper atoms betweenthe two types of the location of a secondcupric ion or on the loca- sites.Whereas studies such as those of Tossell(1978) tion of the "missing" cuprous ion in Cu6S32. and Vaughan & Tossell (1980) show that such a d) Distortions in tetrahedrally co-ordinated sitesin bonding model is grosslyoversimplified, it may be anilite and displacements from tetrahedrally co- used as a first approximation in predicting co- ordinated sitesin metastabledigenite reflect the fact ordination. that neither tetrahedral nor triangular co-ordination polyhedrain theseminerals fit our model. Instead, If sulfur is regardedas having a chargeof -2, the both Cu+ and Cu2* tend to be in intermediateposi- covalently bonded 52 pair in covelline a chargeof tions. -2, the tetrahedrally bonded Cu in covelline a charge e) Figure 4 showsthat there is a basic difference of + l, and the triangularly bondedCu in covelline betweencopper sulfides with Cu:S > 1.75and those a chargeof +2, one can make the following empir- with Cu:S < 1.60. This is reflected in the co- ical observationsusing an ionic model (whetheror ordination number of cuprous and cupric ions. not the reducedCu-S distahcein triangularly bonded Structurally, yarrowite, spionkopite, and geeritewill Cu is due to Cu* bondedto S- rather than to Cu2+ approximate the behavior of covelline and contain bonded to S2-,as suggestedby Folmer & Jellinek cuprousions in tetrahedrallyco-ordinated sites and (1980), is immaterial; the result is the same).The cupric ions in triangularly co-ordinatedsites. Also, 72 THE CANADIAN MINERALOGIST

tu-2 r 91 - rrl i"-r Cu-l l^ cr1 si ) i"-z Cs2 5 c( Cu{ &{ c!-1 92 s Cu-3 &-2 Cu? Cu-2 5 tu-l 3Yi"64$:5 - -6-.;:;u*,, A.COVELLINE C.DJURLEIIE

C) ovr l-llPil1n .uL 3it-) l-rt Nr4ti o&- .3L t^ .% o3# llt lni4|( .78 l^ .7tg _1 Fl'' 1 t0 !1il.] 11 .1 l0 .36 -11 ffi) om- q2 AJ

D.YARROWITE E.SPIONKOFNE F.OEERITEF.OEERITE

Frc. 12. Ionic bonding in covelline, low chalcocite(Evans 1981),djurleite @vans l98l), yarrowite, spionkopiteand geerite.Numbers within the covellinestructure in A representionic charges;in other structuresthey representsite occupancies other than one. Numbersto the right of eachstructure represent charges; anows indicatethe amount of chargeabove (f) and below (l) the ion.

TABLE 9. 'IRTJCTUMLDATA FOR COVEIIIIIS. YARROIITTE, SPIONrcPII! AilD CEERITE atoms in the structuresof yarrowite and spionkop- ite are, like thosein covelline,constrained to atomic @velllne ylrelts spJonkopite g4rlte positionswith x: 0, o/3,2o/3 by peakpositions on Pattersonsyntheses. analyzedconposltion: tot.ooS Cul.l2S cu1.32s cur.s3s structuralcmposltlon'-- f) The formula of yarrowite can be writtpn as (fm FJgs. 6'9): Crt-ooS Cut.]lS c'l.36-r.4oscur.60-r.6rs 'vcu-s - 2.33A and crystdl syst4r heFgonsl iExagonal huagonal rhmbohedral (,cu*)r(,iCu'*)r(sttr(stl10; spacesroupr P6y'@c PSnt, P32l RSr? rttcu-S=2.19 A. To maintain charge balance Tl,rHi or P301 - the sulfur layers, a distribution of copper @tldimnslons: a.3.7$88 a.3.eO0E a 22.s628 a " 15.77I between c"16.34lX c"67,264 cE4l.429I a:13056' atomssuch as in Figure l2d is required. Within this c'rg.s\q c!8sg fomla: Cus hnSzq generalmodel, sulfur atomscan be shifted to many unlicell @ntent: 2"6 7'1 2"36 z-r g/q3 possiblepositions as long as the symmetryrequire- catdlated d€nsliy: +.oe E/o3 4.91 q/cti3 5.33 g/m3 5.61 nunber of sulfut '14 mentsof the spacegroup are maintainedand seven 5 layeB in unft celt: 6 24 layersof covalentlybonded sulfur are retained.Cu nwber of covalently-'layers bondedsulfur co-ordinatedsites can be shifted ln unlt cell: 2 7 20 atomsin triangularly avefage n@bet of perpendicularto c, and Cu atoms in tetrahedrally @pper a&ns pea close-Dacked co-ordinatedsites can be shiftedwithin the two sites sllfur laFr: 1.50 1.58 1.59-1.63 1.53-1.60 in any layer. Data are fw Pottet & Evans (1976), coble a smith (1973)' 0ob1e (1980)' and e) The formula of spionkopite can be written Goble& Roblnson(1980). The denslty of splonkoplteis ronglv repoded iucu-s - as 5.13 g/@3 by coble (1980); thls ls th value for the analyzed@mosl- as 36^('ucu*)rr(t'icu2*)c.s(FJ2-r(S2-)'oi tlon, Cll.32S. 234 A and 'tu-S - 2,20 A. To maintain charge CRYSTAL STRUCTURE OF THE COPPER SULFIDES 73 balancebetween the sulfur layers,a distribution of copperatoms for the a' subcellsuch as in Figure 12e is required. Shifts in the position ofsulfur and cop- per atoms are possibleas long as the requirements of symmetry are maintained and two layers of cova- lently bonded sulfur are retained in the structure. h) The formula of geerite can be^ written as * * (tcu )6(,:cu2 )r(sr-),;,-,cu-s = 2 3 s x and ttr.- S - 2.22A. To maintain chargebalance between the sulfur layers,a distribution of copperatoms such as is shown in Figure l2f is suggested.This structure is basedupon that presentedfor metastabledigenite by Morimoto & Kullerud (1963),with the restrictions that copperatoms be locatedonly in undistoftedtri- angular and tetrahedral co-ordination polyhedra and that rather than one out of everyfive Cu-S-Cu sand- wicheshaving one empty Cu position, two of five must be empty. It is to be expectedthat in eachof thesestructures (yarrowite, spionkopite,geerite), there will be some distortion of the ideal Cu-S triangular and tetra- hedral co-ordination polyhedra, as has beennoted in previously determinedcopper sulfide structures. The basicstructural data for covelline,yarrowite, spionkopiteand geeriteare summarizedin Table 9. For each structure, the averagenumber of copper atoms per tetrahedral layer is calculatedbased on the Frc. 13. A, B, C, D: Possiblearrangement of layers of bonded and covalentlybonded sulfur in spi- ideal compositionas determinedfrom Figures4 ionically and onkopite. E. Attempted crystal-structuredetermination 6 and structural data. For yarrowile, spionkopite and of spionkopite;fractions indicate site occupancy; scat- geerite,this numberapproximates 1.60, whereas for tering factors for copper were used for all positions. covellineit is 1.50. This suggeststhat in the tetra- F. Possibleinterpretation of the partial crystal-structue hedral layersbetween the layersof covalentlybonded of E. sulfur, the structuresofyarrowite, spionkopiteand geeriteprobably have a similar co-ordinationof cop- per atoms, and that this is somewhatdifferent than in covelline. of the 41.4A cellwith atomicpositions indicated for PARTIAL DETERMINATION that structure giving the lowest R-value. R for this OF THE CRYSTAL STNUCTUNE OF SPIONKOPITE structureis 0.280for observeddata but increasesto A.404 if unobserved data are included. This was The numerousmethods of arranging 2 covalent found to be the generalcase for spionkopite; R could and 12tetrahedral layers parallel to the c axis of the readily be reducedto below 0.20 for observeddata, spionkopite cell may be reducedto four types as but inclusion of unobservedreflections causeda shown in Figure l3a,b,c,d (assumedspace gxoup sharpincrease in R. For this reason,an attemptwas IF,ml). All other possiblearrangements may be der- made to find the relationship governingthe unob- ived by shifting the sulfur atoms parallel to the c axis servedreflections. It should be noted that for most (i.e., by changingthe origin). All possiblepositions of the attemptsto determinethe structure, scatter- for copper in tetrahedral co-ordination are indicated; ing factors for copper were used; sulfur would be triangularly co-ordinatedsites have been omitted. expectedto show up as Cu with approximately Vz Numegousattempts were made^at solving the 8.8 and occupancy.Attempts to changeselected positions of 1?.6A subcellsand the 41,4A cell of spionkopite, copper to sulfur did not result in an improvement starting with the basicstrucfural arrangementsshown in the R factor. A possibleinterpretation of the struc- in Figure 13 and assumingthat large areasof the sub- ture in Figure l3e is shownin Figure l3f. The inter- cells would resembleeither the covellineor 'Jdeal- pretedstructure is inconsistentwith a centricspace- ized anilite" structures(thea' = q/6 = 3.827 Asub- group. It is also missing severalcopper atoms. cell was used in all attempts at determining the An examinationof the structurefactors (Table2) structure). In every case,splitting of atomic positions showsthat for the spionkopite cell (a' = a/6), the becamea problem.Figure l3e showsthe bottom half / Miller index for reflectionsof the type ft minus fr 74 "l;"-l"l--.:lTHE CANADIAN MINERALOGIST 045 o o20 8i: o45 9ro [_] 1__j o20 o AB

hrin oxis f

FIc. 14.A. Atomic positionsin the 8.9 A subcellof spionkopitenecessary to account for the systematicabsences; fractional occuppnciesare indicatedas percentages. B. Positjon of Pattersonpeaks for the 8.9 A subcell. C. Structure derived for the 8.9 A subcell:positions onx = a/3 and:c = 2a/3havethe sameoccupancy. D. A twiqned-sphaleritesubcell for comparisonwith C. E, Structurederived for the 35.5A subcell;positions on x: a/3 and x: 2a/3 havethe sameoccupancy. F. A possibleinterpretation of the structurein E. G. A tqinned-sphaleritesubcell for comparisonwith E. H. A twinned-sphalerite41.4 A cell for spionkopite.

equals3n is 4,7, 11,14, 18, 21,25,29,32^36. If In Figure 14, lessthan full occupancyof the sites the cell is re-indexedon the basisof a 35.5 A pseu- is indicated as a percentage. o docell, this index approximates3, 6, 9, 12, 15, 18, Various attemptswere made at solving the 8.9 A 21, 24, 27, 30. The data setre-indexed on this basis re-indexedsubcell of spionkopite using least-squares has two setsof extinctions:if I minus k is equal to refinement and the minimum residual method o.f Ito 3n, then / is not equalto 3n, and if Z minus k is not (1973).The final structurederived for the 8.9 A re- equal to 3r, then / is equal to 3n (but / is not equal indexed subcell is shown in Figure 14c. R for all to 0). Theseextinctions require that atomsbe placed reflectionsis 0.292 (againscattering factors wereused in the cell in the positions0, 0, zi V3,2A,z + tA; for copper only). The correspondencewith the 2/t,rA, z, * 2A;r/t,2A, z + 2/z;2/t,rA, z + rA(frml atomic positions derived from consideration of spacegroup). The atom at 0, 0, z must havedouble extinction data is obvious, as is the similarity to a the weight of the remaining atoms. Theseatomic 'jtwinned-sphalerite"structure (Fig. lad). This 8.9 positions (with z = 0) are indicated on Figure l4a, A re-indexedsubcell *u, e"p*d"Jinto ihe 35.5 A togetherwith peak positions for a Pattersonsynthesis re-indexedcell as shown in Figure l4e. A possible on an 8.9A subcellbased upon the abovere-indexed interpretation is s.lrownin Figure l4f. A "twinned- data (Figure l4b). The peak positions on this Pat- sphalerite" 35.5 A cell is again shown for compari- terson map are identical with those of a sphalerite- son (Fig. l4g). The interpretedcell shown in Figure type $tructureusing the body diagonal as a c axis. l4f is againinconsistent with a centricunit-cell and, CRYSTAL STRUCTURE OF THE COPPER SULFIDES 75 once again, severalatoms of copper are missing. The ing the specimenof geeriteused in the single-crystal prediction that there would be fewer apex-sharing studies.Discussions with Dr. J.D. Scott were par- tetrahedrathan in the covellinestructure is shown ticularly helpful during the attemptsto solvethe crys- by the distributign of copperatoms. All attemptsat tal structureof spionkopite.Finally, I thank Mr. L. solving the 41.4 A cell of spionkopiteusing thesedata Whitesidefor helpful discussionsduring the writing failed. No atomic arrangementcould be derived that of this paper and Helen Betenbaughfor tlping the would account for the observedsystematic extinc- manuscript. tions, afthough expansio! would appearto take place by the insertion of a 6^A segmentbetween the two halves of the 35.5 A re-indexed subcell, such REFERENCES as is slown in Figure l4h for a "twinned-sphalerite" 41.4 A structure. ASRAHAMS,S.C. (1955): The crystal and molecular structure of orthorhombic sulfur. Acta Cryst. E, DISCUSSIoN 66t-671.

In attempting to solve the structure of spionko- Bsnnv, L.G. (1954): The crystal structure of , pite, we are trying to derive a minimum of seven- CuS, and klockmannite, CuSe.Amer. Mineral.39, teen independentatomic positions using just 147 s04-509. observedreflections. If, as is suggestedby the par- tial structures shown in Figures 13f and l4f, the & THor"rpsoN,R.M. (1962): X-ray powder data PeacockAtlas, Geol. Soc. structure is acentric rather than centric (Azl rather for minerals:the Amer. Mem. E5. than frml), this number would increaseto thirty- four. The full cell is in fact thirty-six times this large. Casm, L.J. & Har.r-,S.R. (1972):Mooihoekite and In addition, featuressuch as the diffuse nature of haycockite,two newcopper-iron sulfides, and their the X-ray patterns, broadeningof peakson Patter- relationshipto chalcopyriteand talnakhite.Amer. son synthesesand splitting of atomic positionsdur- Mineral. 57, 689-708. ing structure determinationsall suggestthat Cu is statistically distributed between the two possible Coopsn,A.S., Bor.ro,W.L. & ASRAHAMS,S.C. (1961): tetrahedrally co-ordinatedsites of copper between The latticeand molecularconstants in ofihorhom- sulfur layersand among the four possibletriangu- bic sulfur.Acta Cryst.14, 1008. larly co-ordinatedsites of copper situated around eachtetrahedrally co-ordinated site. Any attemptat DoNNev,G., DoNNav,J.D.H. & Kur.lsnuo,G. (1958): removing constraintson the x and y atomic positions Crystal and twin structure of digenite, CueSr. during determinations invariably lead to splitting of Amer. Mineral. 43, 228-242. peak positions in these directions as well as along EvaNs,H.T., Jn. (1971):Crystal low z. Given theseobservations, it is extremelydoubt- structureof chal' cocite.Noture 232. 69-70. ful that the structure of spionkopite can be fully solved with the currently available data. (1979):Djurleite (Cur.e4s)and low chalcocite Most of the problemsthat exist in attemptingto (Cu2S):new crystalstructure studies. Science 203, solvethe spionkopite structure can be applied equally 356-358. well to the yarrowite structure.The only factor that might make it more amenableto structuredetermi- (1981):Copper coordination in low chalcocite nation would be the previousobservation that large and djurleite and other copper-rich sllfides. Amer. areasof the cell should strongly resemblethe covel- Mineral. 66. 807-818. line structure(e.9., Fig. l2d). However, attemptsat & Kovmnr, J.A. (1976):Crystal structure solvingthe spionkopitestructure using the structure refinement of covellite. Amer. Mineral. 61. of covelline as a model were unsuccessful. 996-1000. The structure of geeritemay be solvable. Peaks observedon X-ray patternsare relativelysharp. The Fottr,mn,J.C.W. & JBrrrurr, F. (1980): The valence of major hindrancein determiningthe structure may copper in sulphides and selenides:an X-ray photo- lie in the scarcityand small sizeof natural material. electron spectroscopystudy. J. Ler;sCommon Metals Experiments such as those conducted by Globe 76. 153-162. (1981)may provide a methodfor synthesizinglarger (1959): fragments of geeritefor single-crystalstudies. FnENzBr,G. Idait und "blaubleibender Covel- lin". Neues Jahrb. Mineral. Abh.93,87-132.

ACKNOWLEDGEMENTS Gogre, R.J. (1980): Copper sulfides from Alberta: yar- rowite CueSr and spionkopite Cu3eS2s, Can. The author thanks Dr. G.W. Robinson for provid- Minerol. lE. 5ll-518. 76 THE CANADIAN MINERALOGIST

- (1981):The leaching of copper from anilite and PurNrs, A., Gnacr, J. & CavenoN, W.E. (1977): the production of a metastablecopper sulfide struc- Blaubleibender covellite and its relationship to nor- tnre. Can. Mineral. 19. 583-591. mal covellite. Contr. Mineral. PetrologJt 60,2@-217.

Rosrxsox, G. (1980): Geerite, Cu1.56S,a new & Rrcrenp, D.T. (1972): Covellite formation in low tem- rer sulfide from deKalb Township,Township. New York. coppersulfide perature aqueous solutions. Minerql. Deposita 7, Can. Mineral. lE, 519-523. 180-188. & Svrrr, D.G.W. (1973):Electron microprobe investigationof coppersulphides in the Precambrian SeoeNeca,R., OHuasa, M. & Monluoro, N. (1965): Lewisseries of S.W. Alberta. Canada.Can. Mineral. On the statistical distribution of copper ions in the 12,95-103, structure of 6-chalcocite, Mineral. J. 4' 275-290.

Hnrsn, J.-E. & PnossruerN,K. (1956):Thermische Tarepa, H., DowNav, J.D.H. & Apprnvar, D.E. und rdntgenographischeUntersuchungen an Kup- (1967a):Djurleite twinning. Z. Krist. 125, 414-422. ferkies.Z. Krist.l0E, 108-129. Iro, T. (1973): On the application of a minimum Rosssool\a,E.H. & ApprsueN, D.E. residualmethod to the structuredetermination of (1967b): The crystallography of djurleite, Cu1.e7S. superstructures,Z, Krist. 137, 399-411. Z. Krist. 125, 404-413.

Kalssropr, R., PEnrrm,F. & ZSBMAN,J. (1975):Ver- Tossrlr, J.A. (1978): Theoretical studies of the elec- feinerungder Kristallstrukturdes Covellins, CuS, tronic structure of copper in tetrahedral and trian- mit Einkristall daten. Tsche r ma ks Mineral. Petro g. gular coordination with sllfur. Phys. Chem. Mitt. 22. U2-249. Minerals 2,225-236.

Koro, K. & Monnaoro,N. (1970):The crystalstructure VaucneN, D.J. & Tossplr, J.A. (1980):The chemical of anilite. Acta Cryst. 826,915-924. bond and the properties of sulfide minerals. l. Zn, Cu in tetrahedral and triangular co- (1971): Fe and Mon, G.H. Blue-remainingcovellite and its ordinations with sulfur. Can. Mineral. lE, 157-163. relationsto phasesin the sulfur rich portion of the copper-sulfursystem at low temperatures.Mineral. Soc.Japan Spec. Pap.1,226-232. Wrrsor, A.J.C. (1942): Determination of absolute from relative X-ray intensity data. Nature 150, Monnaoro,H. & Uvene,R. (1963):A secondcom- 15t-t52. parison of various commerciallyavailable X-ray films,Acta Cryst.16, 1107-1119. WurNscu, B.J. (1974): Sulfide crystal . /n Sulfide Mineralocy (P.H. Ribbe, ed.). Mineral. Soc. Monurroro,N. (1964):Structures of two polymorphic Amer., Rev. Mineral. 7, W2l-M. forms of Cu5FeSa.Acta Cryst. 17, 351-360. & Kur-renup,G. (1963): Polymorphismin YuNo, R.A. (1963): Crystal data for synthetic digenite.Amer. Mineral. 4E, 1.10-123. Cu5.5ferS6.5,(idaite). Amer. Mineral. 48, 672-676. OrrspAr, I. (1932):Die Kristallstrukturdes Covellins (CuS).Z. Krist.83,9-25. Porrnn,R.W. II & EvaNs,H.T., Jn.(1976): Definitive X-ray powderdata for covellite,anilite, djurleite, Received February 7, 1984, revised manuscript accepted andchalcocite. J. Res.U.S. Geol.Sun. 4.205-212. July 31, 1984.