t11s

The Canadian M in eralo gist Vol. 36, pp. ll 15-l 137 (1998)

CHEMISTRY MINERAL CRYSTALLOGRAPHY, OF GOLDFIELDITE, NOMENCLATURE ANDCHEMICAL SERIES SOLID.SOLUTION THETELLURIAN MEMBER OF THETETRAHEDRITE
ALFONSOG. TRTJDU'
Department of Eanh Sciences, Monosh University, Claytott, Victoria 3168, Aastrali.a

TJLRICHIC\IffTEL,
Insrttut fi)r Minzralogie und lagerstiinenlehre, Rhzinischc-Westfiilische Technische Hochschule, Wiillncrstrasse 2, D-52056 Aachzn, Germany ABsrRAcr and tellurian phases of the tetrahedrite solid-solution series This comprehensive review on goldfieldite, Cu19TeaS13, (teEahedritJ, contains a summary of crystallographic properties and compositional data. A new classification scheme for these minerals is proposed. Furthermore, the association of these minerals to specific ore environments is discussed. The bond angles and the optical properties of goldfieldite and tetrahedrite are very similar. The cell dimension of goldfieldite is intermediate between that of tennantite and tetrahedrite; thus it is strongly controlled by the ionic radius of Te. As expected, Cu vacancies in goldfieldite reduce the size of the unit cell. The best electron-microprobe analyses are obtained when S is stendardized on tetrahedrite, bismuthinite or stibnite. The prefened procedure for calculating structural formulae of tellurian elamples of the tetrahedritess from electron-microprobe results is based on total numbers of atoms, as this distributes possible analytical errors over all elements. Owing to the occurrence of vacancies in goldfieldite, the calculation should be based on 29 atoms per formula unt (apfu) for minerals with 2 or less Te apfu Fe/(Te + As + Sb + Bi + Te) < 0.51. For more Te-rich compositions, formula calculation on the basis of 29 - 4f[e{Ie + As + Sb + Bi) - 0.51 apfu is r*,ommended. A review of available electron-microprobe data on goldfieldite and tellurian members of the tetrahedrite,, supports previous observations on the substitutions. For up to 2 Te aplu, substitution of Te4 for (As, Sb, Bi)3+ is coupled with Cu+ for (Fe,Zn)z* subsdrution, whereas higher levels of Te result in vacancies (reduced Cu contents). A nomenclature grid for goldfieldite and the other Cu-dominant examples of the tetrahedrite"" is proposed on the basis of their semimetal content. The name "goldfieldite" is reserved for those compositions of the tetrahedrite"" that contein more than3 apfu of.Te. For those minerals with 2 to 3 apfuof Te, modifying adjectives such as arsenoan,stiboan or minerals of the tetrahedrite"" bismuthoan should be used, according to the second most abundant element. The remaining @11urian should be named according to the most abundant semimetal (tennantite for the As-rich member, tetrahedrite for the Sb-rich one) preceded by "tellurian", provided that at least I apfa of Te is present. Most of these tellurian minerals of the tetrahedrite"" are found in high-sulfidation epithermal Au deposits, regardless whether they are As- or Sb-rich. A few cases have been reported from porphyry Cu and volcanogenic massive sulfide deposits. Keywords: goldfieldite, tetrahedrite solid-solution series, electron-microprobe data, substitutiotrs, strucnual formula, nomenclanue, ore deposits. SoMrr{arne Cette dvaluation compr6hensive de la goldfieldite, CuroTeaSrr, et des phases riches en tellure de la solution solide dite tetraedritess, contient un sommaire des propri6t6s cristallographiques et des intervalles de composition. Nous ptoposons un ds sss min6raux avec des contextes nouveau sch6made classification pour ces min6raux. De plus, nous discutons de I'asso"136i6a de min6ralisation sp6cifiques. Les angles des liaisons et les propri6t6s optiques de Ia goldfieldite et de la ra6drite sont trCs semblables. Le parambre r6ticulaire de la goldfieldite est interm6diaire entre celui de la tennantite et celui de la t6traddrite; iI est donc fortement d6pendent du rayon ionique du Te. Tel que pr6dit, les lacunes dans la position Cu r6duisent le paramCtrer6ticulaire. Les meilleures donndes obtenues par nicrosonde 6lectronique sont attendues lorsque la srandardisation pour le soufre s'effectue sur la t6tra6drite, la bismuthinite ou la stibnite. Le protocole recommand6 pour le calcul d'une formule structurale de cristau( de la s6rie t6na6drite,, riche en Te i partir des r6sultats de microsonde est fond6 sur le nombre total d'atomes, ce qui distribue les

21069, alfonso.trudu@ Ausralia E-matl a.ddress: Presentaddress: CSIRO - Minerals, P.O. Box 83, Kenmore,Queensland minerals.csifo.au E-mail aidress: knittel@rwth-aachen.de shouldbe addressed. Corresponding author,to whom reprint requests

I116
erreurs amlytiques Possibles sur tous les 6l6ments. Vu la pr6sence de lacunes dans la goldfieldite, le calcul devrait supposer 29 alomes par t'nit6 formulaire (apzj pour les cas oil le Te 6quivaut A2 apuf ou morns [Te(Ie + As + Sb + Bi + Te) < 0.51.Dans le cas de compositions plus enrichies en Te, le calcul de la formule se fait sur une base de29 -4(lel(Te + As + Sb + Bi) - 0.51 apuf. Un examen des donndes disponibles de microsonde dlectronique portant sur la goldfieldite and et les membres enrichis en tellure de la sdrie t6tra6drite"" 6taye les observations ant6rieures d propos des sch6mas de substitution. Oil le Te 6quivarlitd2 aptt ou moins, la substitution de Te# d I'ensemble (As, Sb, Bi)3* est coupl6e au remplacement de Cu+ par I'ensemble (Feln)2+; en revanche, les niveaux plus 61ev6sde Te mbnent d des lacunes (teneur r6duite en Cu). Nous pr6sentons une grille de nomenclature pour la goldfieldile et les aufes min6raux d dominance de Cu de la s6rie t6tra6drite"" selon leur teneur en semim6tal. Le nom "goldfieldite" est r6serv6 pour les compositions de la s6rie t6tra6drite""contenant plus de 3 apuf de Te. Pour les compositions contenant entre 2 et3 apufdeTe,nous proposonsles qualificatifs arsenical,stibi6, ou bismuthifdre, selon lequel des trois 6l6ments As, Sb ou Bi est le plus enrichi. Les autres cas porteurs de Te de la s6rie t6tra6drite"" devraient porter le nom du min6ral selon l'ldment senfm6tellique le plus abondant (tennantite pour le membre arsenical, tdtraedrite pour le membre stibi6) pr6c6d6 par "tellureux", moyennant au moins I apuf deTe.Laplupart des exemples de la s6rie t6tra6drite"" enrichis en tellwium proviennent des gisernentsd'or 6pithermaux h sulfidation 6lev6e, qu'ils soient riches en As ou en Sb. Quelques cas ont 6t6 signal6s dans les gisements de cuivre de type porphyrique et les gisements volcalogdniques de sulfures massifs. (Iraduit par la Rddaction) Mots-cl6s: goldfieldite, solution solide de la t6tra6drite, donn6es de microsonde 6lectronique, subsfitutions, formule structurale, nomenclature, gftes min6raux.

IlrRopucuorl Goldfieldite, ideally Cu1sTe4S13, as shown by Kalbskopf (1974) and Johnson & Jeanloz (1983), is the tellurian end-member of the tetrahedrite-tennantite solid-solution series (henceforth tetrahedriteJ. The pure end-member has not been found as natural mineral; a maximum of 3.4 Te (out of a possible 4) atoms per formula unit (apfu) has been reported (Trudu & Kninel, in press). Goldfieldite, tellurian tetrahedrite and tennantite have been reported to occur in various types of ore deposit, inclurling epithermal gold deposis, porphyry-Cu deposits, and volcanogenic massive sulfide deposits. Goldfieldite was fint describedby Sharwood (1907) and later identified by Ransome et al. (1909) n ores from the Mohawk mine at Goldfiel4 Nevada. Thompson (1946) summarizedthe controversy regarding the interpretation of this mineral, which was thought to be possibly a mixture of different known minerals (e.g., tetrahedrile + calaverite, or famatinite + bismuthinite + calaverite or sylvanite), even if it appeared to be homogeneous. For a goldfieldite specimen from the Claremont mine at Goldfield, he obtained an X-ray pattem analogous to that of tetrahedrite and hence, proved it to be a member of the tetrahedrile solid-solution series. ln his classifications of sulfosalts, Nowacki (1969) regarded goldfieldite as a variefy of tetrahedrite and believed that Te substituted for S. Kato & Sakurai (1970a) were the fust to infer that Te substitutes for As and Sb. Kalbskopf (1974) ndependently came to rhe s4me conclusion and suggested the end-member formula Cu lqTeaS 1 3. Charlar & lt'vy ( I 974) speculatedthat in goldfieldite, Te can be both a cation, substituting for As and Sb, and an anion, replacing S. Kase (1986) proposed that substitution of Tea+ for the trivalenr semimetals (As, Sb and Bi) is compensatedby an in-

(mainlyCu with minorAg) crease of monovalent atoms from 10 to 12 apfu at the expense of a decrease of the (Fe andZn) from2 to Oapfu leading divalent elements to the formulaCu12(As,Sb,Bi)zTezSrs. For higherconcentrations of Te, the chargebalancein tetrahedrite"" is maintainedthroughvacancies in one of the sitesoccupied by monovalent metal atoms(Kalbskopf1974, Dmitriyevaet al. 1987). These relationships havebeen confirmed for natural goldfieldite (Trudu & Knittel l99l,in presso Shimizu& Stanley1991). (l97la) investigated De Medicis& Giasson thesystem Cu-S-Te at 340"C, but did not report the occurrence of goldfieldite. In contrast, Karup-Mgller (in prep.)found goldfielditeat 350' and450'C in his experimental investigation of the same system. We havecollectedresultsof overfour hundredelectron-microprobeanalysesof natural tellurian tetrahedrite,"plus a few from syntheticmaterial.Thesedata form thebasis for thisreview,in whichwe alsosummarjze anddiscuss the crystnllography, opticalproperties andmineralchemistryof naturalandsynthetictellurian tetrahedrite"". We proposea new schemeof classification basedon mineral chemistryand discussthe occlurenceof thesemineralsin different types of ore deposits. Cnysranocnaprilc PARATmTERS Goldfieldite, as all mineralsof the tetrahedrite"", is cubic, a derivativeof the sphaleritestructure,space group l43m(tt/lensch 1964) with two formulaunits per cell. The formula of a half unit cell of tetrahedrite"" is generallygiven as: M*rcfuP*zF*+S13, whereM+ = Cu, Ag; lE+ = Fe, Zn, M4 Cd, and Hg, andX = Sb, As, Bi, Te. Half of the twelve M sitesare tetrahedrallycoordinated(Ml), andthe remainder arein a triangularplanar

GOLDFIELDITE, TTIE TELLURIAN TETRAHEDRITESJ

lllT
Jl I I I

M 1 6M 2 4( X Y r ) o Z

'uM1 =cu @

I I

(D "'M2=cu

'ux=Te 'uY=s

@ u' @ z=s

Frc. l.Hafunit-cellofpuregoldfieldite,modiJiedafterpzuling&Newman(1934)andCharnocketal. depicted only 213 full to account for two vacancies.

(1989a).TheM2sitesare

coordination (M2).T\e X atoms occupy four 3-coordrnated sites, forming a trigonal pyramid. This arrangement creates a void in the structure, the site of a lone pair of electrons originating from the X atoms (Moore 1981, Johnson & Burnham 1985). In end-member goldfieldite (Cu1sTeaS13), trvo of the M2 sitesare vacanl and the others are occupied by monovalent atoms (predominaaly Cu) to balancethe chargeresulting from thereplacement of trivalent As, Sb and Bi by tetravalent Te (Kalbskopf 1974, Dmitriyeva et al. 1987). Amongst the anions, twelve (usually labeled Y) are in 4- fold coordination, and the 13th atom (normally referred to as Z; is octahedrally coordinated. A halfunitcell of goldfieldite is drawn in Figure 1; as they are supposedto be statistically occupied, the M2 sites are all represented as 2A full. Radii for five elements relevant to this studv (rvCu*, ryFe2*,rv7n2+'5.2^d S2-) were given by Shannon (1981). This author also showed that the radius of SL in sulfides is independent of its coordination number. However, no radii are available for the other elements that may be incorporated in the structure. Owing to this lack of data, we must rely on radii measuredin oxides and halides by Shannon (1976). Furthermore, for each element, ionic radii are available only for some of the valence and coordination numbers of interest to this study. Estimates of ionic radii required for the interpretation of variations in bond leneths and cell dimensions

in tellurian members of the tetrahedrite"" ate based on the following observations(e.9., Shannon 1976): (l) for each atomic species, the ionic radius decreaseslinearly with oxidation state and increases with coordination number, nd (2) for C-u,Fe, and Zn,radti measuredin sulfides are about 0.1 A smaller than those in oxides. In

o.8 o.7 r (A) 0.6 o.5 o.4 1.0 r (A) 0.8 0.6

-/-)

Cu+ (oldde)

./"1cu+

(sulfide)

lf*y1j----tsf*

As3+

o.4 23456 coordination number

Frc. 2 . PIot of coorditration number versas ionic radius, r, in Angstrtim (A), for Cu and As, Sb and Bi. Measwed radii from Shannon (1976, l98l) are shown by filled symbols, and estimated values by open symbols.

1 ll 8
TABLE I. IOMC RADtr FOR MA'OR AI{D MINOR ELEMENTS RtrRRE) TOINTEISI WORK TABLE 2. MEAIIURED AI{D CAI.CUI.ATED BONIIITNGIIilI IN SYNTIIETIC GOLDFIBDTT4 ARSENOA}I GOLDFIEIDII4

ciddimde B@d Gr Cu Ch F9 7^ AB Ag A8 A8 Ar To Ts IR@ uid

80d9

CN /,.100 IRe

EIR'i

neaercd

A(elc.@.)

tr u
+2 +3 +3 +5 +5 +4 +4 +3 +3 +3 +5
+J

IV Iv IV Itr
VI

0.@ o.674 0.74

0.n 0.74
0.495 0.57 o.T2 o.475 0.60 0.65 0.80 0.E6r 0.90 0.94 o.74

057r odul 0.66 0.6

Ml-Y

At:O.63s I

M.61

ry ry
III Iv

S:1.70 2.335 2.37(2)ffi 2.313(9) acfl 2.33(E) 16 2344qrsi.

-O.O35(2) 0.02<9, -{.0048) -o.00(o

tuD-f

sb sb sb sb sb
Bi Bi Bi

VI Itr ry

o.ort 0.t0

6t:O.57

IU

8.61 gtr.7o 2.2'1 2.lqt8Gd

0.11(t 2.24q10)scfl 0.6q10) 2.2sE9) T@ 0.0149) 2.2.rXDTcr 4.@2fi) 0.14(3) 0.0t8(7) 0.06(8) o.6qt

tufr-Z Qt:0.57

Itr

t.6l

S:1.70 227

ry
vt ry
IV

VI Iu

o.a

2.13(3)scdl 2.'l2{Ascdl z2aa\\Ta 2.a4(5lTf!

+3 +3

i.to
l.l7 1.56 1.W

0^91. 1.Cl t.15 t-70

X-f

Ss

r"st

To:0.66 4 (3To+lAs) As:0.495 3 Sb:0.86 3

m Itr m m

3.92 S:1.70 2.36 2.41(3) 8cfl -0.05(3) -0.04{9) S:1.70(2.319)2.364(9)aGfl 6.06 S:1.70 2.195 2.245(s)To -O.051(5) E.61S:l.7O 2.56 2.44qrTea 0.114(3)

t Vabs elorlced bllwirg tho procedm dqsibed h ths tem od shm ir Egrs 2. Tb nmbq ddigin iegiw o in the oigitrsl !*renco. CN: cordincim nmb6, IR tonicndii rspdsd by Sh@on(19O, SR ndiiin rulfdc rcported by Sh@@ (l9El). The dat' appfieblototfcaheddt& appsin bold fre

AIl vatue m in Angffon uitr. The qiddio dst (Gid. stde) of S is iMisbb 2-, but its mdincim Dnb@ (cN) m W ad VI fu ths I ad Z ritcl respeottudy. f fiactimd ioric ah!ruts (s t(E). IR imia ndiu. scfil: d@ o goldfisldits rychreic goldfieldm frm Kalbskopf (1974). aGfr: dda o mm ftm ffiiyffi aI (19E7). Tu .l"ta m t@ffits to@ Cru@cl (1964). Tet: "t ftm Wu@schstaf (1966). dce q trfrahedrile

contrast, semimetals have nearly the same ionic radius in both types ofcrystals Cfable l). Ionic radii for oxides (Shannon 1976) and sulfides (Shannonl98l) for Cu, Fe,7n, As,Te, Sb, Bi, S and Se in the oxidation states and coordinations relevant to this investigation are listed in Table l. Figure 2 shows the graphic exffapolation of the radii of Cu, As, Sb and Bi in 3-fold coordination. The radius of IIrCu+ in sulfides was calculated as-the equivalent value measured in oxides minus 0.1 A (Fig. 2). Because the radii for As, Te, Sb and Bi are very close (t0.03 A or better) in oxides and sulfides (cl tuAss+, ryTe4o vgi:* and vrBi3* in Table l), data obtained from oxides (e.g., for Sb) and sulfides (e.g., Bi) may be used interchangeably.The extrapolation method usedto estimate the radius of As is similar to that employed for Cu. In0erestingly, although the radius of 52- is different in oxides and sulfides, no such changeis observedfor Se. Bon"d lengths and angles Crystal-structure data has been reported so far only by Kalbskopf (1974), who gathered powder-diffraction data from synthetic material, and by Dmitriyeva et al. (1987), who studied a single crystal from the Yelshishche deposit in Bulgaria. Bond lengths reported in these studies, as well as data for tetrahedrire (Wuensch 1964) and tennantite (Wuensch et aI. 1966), are summarized in Table 2. The effect of Zn and Fe in the M2 site on the bond lengths of tennantite and tetrahedrite was neglecled, because the radii of Fe and

Zn differ from that of Cu by only 0.02 A. The analytical uncertainties in the bond lengths measuredby Kalbskopf (1974) are relatively large (t0.02 to 10.05 A), because they were calculated from powder-diffraction data, whereas the single-crystal data of Dmitiyeva et al. (1987), Wuensch (1964) and Wuensch et al. (1966) are associated with substantially lower uncertainties (+0.003 to 0.010 A). There are four distinct bonds in the tetrahedrite"" structures. Atoms in the Ml and X sites are bonded to 4-coordinated Z anions, whereas the atoms n the"M2 sites are bonded to those in both the Y aad Z sites. As the bond length in theory is equal to the sum of the radii of the two atoms linked together, Table 2 also compares theoretical bond-lengths based on the data of Table 1 with the direct measurements of bond lengths in tetrahedrite"s.Increasing covalency (Gamble 1974) and delocalization of electrons (Shannon I976) lead to measured bond-lengths tlat are smaller than the sum of ionic radii. T\e Ml-Ybond-length should be about the same in all Cu-dominated minerals of the tetrahedritess,because substitution in the Xposition does not affect thesesites. The data in Table 2 support this interference, considering the uncertainties of the meutsurements. There is also good agreement between bond lengths calculated from ionic radii (2.34 A) and measuredbond-lengths (2.312.37 A). The MZ-Y bond lengths for tennantite, tetrahedrite, and the arsenoangoldfieldite of Dminiyevaet al. (1987) are very similar, though a decreasefrom tetrahedrite to

GOLDFIELDME, TIIE TELLURIAN TETRAHEDRITES

1 t1 9

0.16
synth.gfd.C!
tt

r Gamble (1974)

o.12
synth.gfd.tr
I

tr M2-Y o M2-Z

vl

o
I

C'

E
0.08
(J .E

ars.gfd.O tenn. o

0.04

o.00 0 0.10 0.20 0.30

fi
Frc. 3. Plot of fractional ionic character (J) versu.sthe difference beflreen calculated and measured bond-lengths (in A). The line represents the linear regression through the data of Gamble (1974), which is representedby filled squares.See text for further explanation. synth. gfd.: synthetic goldfieldite (Kalbskopf 1974), ars. gfd.: arsenoan goldfieldite @mitriyeva et aI. 7987), tenn.: tennantite (Wuensch et aI. 1966), ter.: Etrahedrite (Wlensch 1964).

goldfieldite is indicated.In contennantiteto arsenoan goldfielditehasa significantlyshorter trasl thesynthetic (2.16 t 0.05 A versus M2-Y bond-length 2.240t 0.0I 0 that a relaA); however,Kalbskopf(1974)proposed to that value owing tively largeerror might be attached to an increased temperafure-factor. All measured M2-Z bond lengths are shorter than the calculatedvalues.Measuredbond-lengthsshorter concerned thanthesumof theionic radii of theelements wereobserved by Gamble(1974)to be inverselypro(/) in sulfides, portionalto thefractionalionic character (Frg.3), with fi definedas: selenides andtellurides fi (A - B) 1 - exp [-{.25 (xa - xs)2] (l)

T\e X-Y bond lengths show some contradictory results. The measured Te-S and As-S distances in synthetic goldfieldite, arsenoan goldfieldite and tennantite are longer than those calculated, a fact in contrast with Gamble's (1974) observation.These discrepancies may be the result of analytical errors related to the chemical composition the mineral, errors in the determination of the bond lengths, possible changesin coordination of the semimetals as a result of vacancies in the structure, or inapplicable ionic radii. The possibility of an analytical error was discussed by Wuensch et al. (1966) in their study of the structure of tennantite; the mineral they used could have contained minor amounts of Sb, thus increasing the bond length. A similar error could not have affected synthetic goldfieldite. Substantial errors in measurements of the bond lengths are unlikely, though Kalbskopt (1974) reported a relatively large uncertainry of 10.03 A for the measuredTo-S bond length in synthetic goldfieldite. This leaves the possibility that the estimated ionic radius for mTe is slightly too low at 0.66 A. Vacancies are reported only in goldfieldite, but they should not affect the coordination of Te (Kalbskopf I97 4). If they did, these would result in an increase in the radius of Te, which would make the difference befween calculated and measured bond-lengths positive (Table 2), in agreementwith Gamble's (1974) general prediction for all minerals. Thus, with the data currently available, no viable hypothesis can be proposed to explain why the measured bond-lengths for synthetic goldfieldite, natural arsenoan goldfieldite and tennantite exceed the calculated values. More accurate measurements can possibly help to solve this problem. The X-Y bond lengths in tetrahedrite are in agreement with Gamble' s (1974) prediction, though we used the radii applicable to sulfides. There are very limited published data regarding the bond angles in goldfieldite. Kalbskopf (1974) teported that in pure goldfieldite, the S-Te-S angle is 95.7 + 0.7", which is very close to the corresponding value (95.1 t 0.2") for tetrahedrite measured by Wuensch (1964), as are reported values for the two S(I0{u(M1)-S(D angles in arsenoangoldfieldite @mitriyevaet al. 1987). Cell dimensions Currently available data for the cell dimensions of goldfieldite are summarized in Table 3, which also lists reference values for synthetic tetrahedrite and natural tennantite. The prediction of cell dimensions as a function of the different types of substitutions in tetrahedritess was flust attempted by Charlat & I-Evy (1975) and subsequently refined by Mozgova et al. (1979) arrdJohnson et aI. (1986,1987). However, only Johnson et al. (1986,1987) provided details on the type of statistical analysis used. Te-for-(As,Sb) substitution is only taken into account by the equation given by Mozgova et al. (1979):

@auling1948),whereA andB are the two atomsinvolved in the bond, and 1, their electronegativities, as calculated by Pauling(1948).However,use of the smallerradii observed in sulfides(Shannon1981; between Table I of this paper)led to good agreement bond-lengths calculated ard observed for MI-Y ard M2-Y bonds in tetrahedrite,tennantite and arsenian goldfieldite. Onlytheshorterbond-length in goldfieldite with the relationship observed by would be consistent in Gamble (L974). Hence,the shorterbond-lengths goldfielditemight be dueto thevacancies in fheM2 site, ionic character of the bond. ratherthan to decreasing

t120

TIIE CANADIAN MINERALOGIST TABLB 3. UNIT@I SIU OF GOLDFIELDITE TETRAHEDRTTEA},{D TENNANTTTE

40 (A)

Sghecic cytals Goldfddite frroTeoss Tstrahe&fte C\luSb4SE Tcrahsddte c\oGe,zntsb{fu Ndral dtrtals Aneoom goL{ffslditg GquAq.rSb@TqrSF Arseom goldfddbE GruAq:T%gSsr Stibom goldfeldite Cu,"nArorSbrrTerS* Stibm goHfleldto OqEAtuSbuTqrSEa Sdmlo goldfleldite CqDAsuSboaT%rSurS Tarnorreg Teffin*te CqFenln,Aqr\** TeanhedritoCurSoorZn"rSbnrSo*t 10.263(5) 10.3190) 10.383(l) raftnkcpf(9+) Hall 0972) HanQYt2)

10.304 10 7(2)

PDF (29s31) Dltrihiyswatal (1987)

10.35 10.32q5) 10.3t 10.186 10.235(t 10.391(1)

Sptidmoveral (t9t4) TqinetaLQgr' Spirftlmry& Olrtgin (t9S5) PDF (l l-102) Wu@cn f al- (l 966') Wumsch (1964)

reported.L.G. Berry (as cited in PDF#29-531) reported a value of 10.186 A for natural tennantitecontaining up to 4.77 wt.VoAg; as the Ag-for-Cu substitution in the tetrahedrite," causes an increase in the cell edge (Hall 1972), it can be predicted that a pure tennantite endmember (Cu16[Fe,Zn]2As+Sr:) should have a cell size slightly smaller than 10.186 A. On the basis of the available data (Table 3), there is a considerablecontraction (about 0. I A) in the cell edge as the proportion of vacanciesincreases^ in goldfieldite. The largest cell-edge reportd 10.38 A, measuredby Spiridonov & Okrugin (1985), pertains to the selenian variety. It is evident that the increase in the cell dimension with respect to that of pure goldfieldite is due to the Se-for-S substitution (c/ Table t). Orrrcar, Pnopmrns ANDSpEcrRoscoprc REsULTS

Quantitative measurements of the reflectance of goldfieldite were reported by Spiridonov et al. (1984)l values in the range of 29.5 to 3l7o were measuredfor the As-bearing variety. High Te contentsin goldfieldite SChmicat conpodtionis Dotavaih[o * Chmiel 64ssido[ is wliable md c ruimce with spafion'Eicdo m ths me mdedal cmpocition protOty cfore marginally increase its reflectance, but large amounts tD ChrZDuFfuAs$Su (Wi:mch aal 156). *. IGflilts of old estnis SPIitrgE ofSb have the opposite effect. These authorsconcluded (1959) rcportd the blwirg coryosfim ftr ndsial A'od thi! l@lit4 CurZn*FqrSbnSo. that this mingral is indistinguishable from tetrahedrite, as both are grey with a pinkish to brownish hue. In contrast, unpublished observations suggest that the Te content of goldfieldite may affect its color. One of the aa= 10.379 + 0.031Fe + 0.0287n + 0.096Hg + present authors (UK) notes that Te-rich tetrahedrilessis 0.0072MelS - 0.040As + 0.093Bi - 0.003Te + distinctly pinkish where in contact with Te-free Ag/(21.9-r.or As) (2) tetrahedrite"", in particular in oil immersion. At the La Mejicana base- and precious-metal epitherrnal deposit This equation is valid only for tetrahedrite"" with less (Argentina), two types of goldfieldite were recognized than3.7apfu Agandfor mineralswith a total of 29 apfu (A. Losada-Calder6n, pers. commun.): a low-Te (Mozgovaet al. 1979).The frst conditionis easilymeq (a apfu) variety, which occurs at deeper levels and is as the highestAg contentrecordedin goldfieldi[eis texturally associa[edwith gold, shows a greenish tint 2.2 apfu (Kovalenker et aI. 1980), bur the second resembling tennantite, and a high-Te (>2 apfu) variety, excludescrysta.ls with more than 2 Te apfu ow:mgto which is presentat shallower levels, is grey, like normal vacancies inthe M2 site.The validify of Equation 2 has goldfieldite. The presenceof Se, as shown by material beentestedagainstthe measured sizeof the unit cell of from the Ozemoye Au deposit, in Kamchatka, Russia, goldfieldite reported in Table 3 (PDF #29-531; does not seem to affect the color of the mineral. Spiridonov et al. 1984)with the following results: cell Spiridonov et al. (1984) measuredalso the Vickers dimensionscalculatedfrom mineral-chemistry data microhardness (load = 30 g) of goldfieldite and obtained are smallerthan thosemeasured [10.280A versus values decreasing from 334 for tetrahedrite with a very 10.304 A for thearsenoan goldfieldite ofPDF #29-531, low Te content (0.54 wt.Vo) to 205 for goldfieldite and10.32 Aversus 10.35A for thestiboan goldfieldite (Te = 14.98 wt.Vo). of Spiridonovet al. (1984)1. In contrast,Basu et al. At present, there are only two studies concerned with (1981)claimedthat Equation2 canpredicrcell sizes the spectroscopic examinationofgoldfieldite. In a recent within the analyticaluncertaintyof the measured value. laser Raman investigation of some geologically imporAs expectedo thecell sizeof synthetic goldfielditeis lant sulfides,Mernagh & 1ru6u (1993) observedRaman intermediate in sizebetweenthat of tennantiteandthat bands at 354 and 324 crtrr , which represent an intermeof tetrahedrite in spite of the two vacancies present, diate position between those of tennantite and of However,thereis still somedebate aboutthe exactcell- tetrahedrite. They interpreted these results as typical dimension of tennantite. Wuensch et al. (1.966) ques- one-mode behavioro with frequencies decreasing tionedthe purity of the naturalcrystalthey^analyze.d,continuously from lighter (tennantite) to heavier + 0.005A, whichis (tetrahedrite and goldfieldite) minerals of the same because it yieldeda valueof 10.235 toolargeifcompared to thevalueof 10.19 A previously group.

GOLDFIELDITE, TIIE TELLURIAN TE'TRAHEDRTES

1l2l

bismuthinite and tetrahedrite produced markedly better results (Fig. 4). A similar behavior in electron-microprobe data was reported by Harris (1990) for Ni-Hg sulfides: HgS analyzed with S standardizedon NiS at 12 -20 kV yielded values up to 7 wt . Vohtgfier than expected, o -9td9*" whereas the opposite effect took place when NiS was M* * * analyzedwith S standardizationon HgS. Harris (1990) 11 ascribed these poor results to the correction programs, o\o-_ which inadequately dealt with the absorption edge of the standards. 10 Analytical problems with Cu in tetrahedrite that is 4.0 3.0 2.5 3.5 2.O not fully substituted have been discussed by Lind & Makovicky (1982). However, the available data suggest Te that tellurian tenahedrite"" is in general fi.rlly substituted. It is not known whether Se, the other anion possibly present in goldfieldite, should be calibrated under conditions similar to those of S. Se-bearinggoldfieldite is very rare. The only published data pertain to materials from the Ozernoye Au deposit in Kamchatka, Russia. Spiridonov & Okrugin (1985) did not mention which Se standard they used. For Te and most of the other elements, standardization on pure metal pfoduced good results (Trudu & Knittel, in press). This shows that matrix effects have a negligible effect on the measured concentrations of the cations, as already pointed out by Hanis (1990). Altogether, 19 elements have been reported to occur in tellurian tehahedrite"", but only the following are routinely detected: S, Cu, Fe,Zn, As, Se, Ag, Sb, Te and Bi. Of these, Bi and Se tend to be neglected in routine TOTALStSeATOMS analysis of goldfieldite. Bi substitutes for Te, and its theoretical end-member will be used in the new classiFrc. 4. Plots showing the effect of the choice of the reference fication scheme for goldfieldite (see below). Neglectmaterial for S calibration on the analytical data- All the data ing to analyze for Se, even if presentin amounts as low in this diagram are ftom the Tirad porphyry Cu-Au prosas I wt.%o,can lead to the erroneous assumption that pect, northwestem Luzon, Philippines (Trudu and Knittel, low S is caused by a poor calibration of the electron in press). Legend: analyses with S calibration carried out microprobe. It must be kept in mind that low concentraon ttrahedrite, stibnite and lis6uthinite (circles in upper tions of Se may not appear on an elemental spectrum of diagram and stippling in lower diagram), and on anenogoldfieldite, becausecritical X-ray absorption lines of pyrite (crosses in upper diagram and harching in lower diagram). All concentrations are expressed in aplu. Se partially overlap thoseof As, which is usually present in larger amounts than Se in this mineral. Elements of secondary importance, Au and Pb, and In the second study, a l2lsb Mtissbauer investiga- minor or trace amounts of Hg, Mn, Cd, Sn, Co, V and tion ofseveral Sb-bearingminerals, Hedges & Stephens Ge have been reported only occasionally. (1977) determined that the Sb in goldfieldite is 3-coordinated, confirming the result of Kalbskopf (1974). Fonuure Cer,curanow oF THESTRUcTuRAL

13

-d"ail"

..s16?&L**

CarmnnrroN oF ELEcTRoN-MrcRopRoBE DATA The quality of electron-microprobe data on goldfi eldite, generally obtainedin the wavelength-dispersion mode, surprisingly seems to depend on the choice of the reference materials used for calibration, in particular for analysisof the samplesfor S. Figure 4 shows that calibration of S on an arsenopyrite standard gave low values for S in goldfieldite, resulting in too high a number of atoms per formula unit, when formulae were calculated on the basis of 13 S. Calibration on stibnite.

Structural formulae for tetrahedrite"" are usually calculated on the basis of 13 anions (S + Se) or 29 total atoms. Seal et al. (1990) suggestedthat a calculation procedure basedon 4 semimetal atoms (As, Sb, Bi, Te) be adopted, although they admit that small analytical errors for the semimetals can cause larger errors in recasting the number of sulfur and metal atoms. This problem is also inherent in the calculation of the structural formula on the basis of 13 anions, in particular, if Se is not sought, but is present in substantial2mounts. In addition, analytical data for S may be problematical,

r122
as outlined above. It is even less suitable to calculate the struchral formula for goldfieldite based on 12metal (Cu, Ag, Zn, Fe,etc.) atoms(e.g., Charlat&Itvy 1975) becausevacancies occur for more than 2 atoms of Te (apfu). T\e same reason should make the calculation procedure based upon a fixed number of total atoms suspect, as this varies from 29 for Te S 2 apfu to 27 apfu for end- member goldfieldite. Keeping in mind all these considerations, we prefer to calculate the structural formula of goldfieldite and other tellurian minerals of the tetrahedrite"" using the method basedon the total number of atoms.This procedure has the advantage of spreading the analytical error over all the elements sought. In order to take into account the occurrenceofvacancies forTe>2 apfu" we suggest that l) for Te/(Te + As + Sb + Bi) < 0.5, the calcularion be basedon 29 apfu, as no vacanciesoccur in the crystal structure; 2) for Tel(Te + As + Sb + Bi) > 0.5, rhe formulae are to be calculated on the basis of 29 - 4lTel(Te + As + Sb + Bi) - O.5l apfu. Mwenar CrnpusrnY and Mtissbauer spectroscopy, and Makovicky et al. (1990), who studiedsynthetic letrahedriteby Mdssbauer spectroscopy. These studies in part arrived at different conclusions relevant to tellurian tetrahedriters. Copper, the most abundant metal in goldfieldite, occupies both the Ml and.M2 site. Its oxidation state has been the source of speculation; however, X-ray photoelectronspectroscopyofabout twenty sulfides and sulfosalts,including brahedrite, by Nakai et aI. (1976), as quoted by Kase (1986), indicates that Cu tends to be monovalent in these minerals. Reported concenffations of Cu in the tellurian members of the tetrahedriters vary from 5.88 to 12.74 apfu, but values of less than cc. 9.5 Cu apfuwere obtained only for Au-rich goldfieldite (Figs. 5, 6; Kovalenker & Troneva 1980), which probably is not really goldfieldite but a mixture of different minerals (see discussion below), Problems of electronmicroprobe analysis of tetrahedrite not fully substituted with Cu were discussedby Lind & Makovicky (1982), but the trend displayed in Figure 6 suggests that Cu Thc M sites invariably is in excessof lO apfu, where this is required In tetrahedrite"", there are two sites occupied by by coupled substitution as a result of Te incorporation. metals, Ml and, M2, whrchare 4- and 3-fold coordinated, The Cu contents are a function of two types of respectively. Cu is the main occupant of the M sites; it substitutions,Ag + Au for Cu, and Te for (As, Sb, BD. may bereplacedby variable amountsof Fe, Zn, Ag, Mn, The plot shown as Figure 5 indicates that the fust type Cd and Hg. In the tellurian members of the tetrahedrite"", of substitution is of minor importance, as the concenX-ray-diffraction studies (Kalbskopf 1974, Dmiriyeva tration of the precious metals rarely exceeds 0.5 apfu. et al. 1987) have shown that up to two vacanciesoccur The data of Charnock er aI. (1989a, b) indicate that in the M2 site where the Te contents exceed 2 apfu, as Ag-for-Cu substitution occurs at the M2 site in both previously discussed. natural and synthetic tehahedrite. Figure 6 shows the 11s srystalIographic positions of the metals (Fe, Zn, relationships between concentrations of monovalent Ag, Mn, Cd and Hg) substituting for Cu in the tetrametals (mainly Cu) and Te; Cu increases from l0 to hedrite*, and thus in goldfieldite, must be re-assessed 12 opfu as Te concentrations reach up to 2 apfu. For in light of the findings of Charn ock et al. ( t 989a, b), who higher amounts of Te, the Cu content decreases to used extended X-ray absorption fine structure @XAFS) lO apfu, as vacanciesn the M2 site appear(Kase 1986, We have assembled a databasecontaining all results of electron-microprobe analyses of tellurian members ofthe tenahedriters group available to us, using an arbitrary cut-off of 0.3VoTe (ir all, 354 compositions, as shown in Table 4). No attempt was made to evaluate the qualiry ofthese data,becausevarious types oferrors can affect them, e.g., counting statistics, matrix correction, use ofinappropriate standards,etc.,blt we have excluded all compositions with totals outside the range 97-1O2Vo.In the following.discussiono we will, however, point out inconsistenciesin some data sets. In the following sectionsowe summarize what is known about the distribution of the elements on the different structural positions. Where no information is available, we will discuss the possibilities based on Goldschmidt's rules of substitution, as summarized in Faure (1991), and with reference to other studies of tetrahedrite-group minerals.
TABLE 4. LIST OF SOURGS OF ELECTRON.MIG,OPROBEDATA ASSEMBLEDFOR TITIS STTJDE AND NTJMBER OF DATASETS

Borisowsra! (1986) (1987) Dditiyffistal Frwzd,et aI (197, Igum(1986) (1990) JCPDS Kmp-MsXq (1992)(sydhstio Kre(1986) Kdo & Sakumi(1970a) Kato & Sshmi (1970b) Ktritol(tEpubl. dda) Knitel (1989) Kovaldksetal (1989) Kovsloks&Rusitrqv(l9EO Kovalmko&Trom(1980) Koralskqeta, (1980) Kovalskseral (1989)* Totd

la l0 I 1 2a II 27 9l 8 18 20

Koval@kset4l (1990) r.&y(r 7) Ipgircvaal (1983) (1983) Mozow & Ts'pitr N@gorcdmerql (1978) Nwiello (19E9) Sakhrcwsral (1984) Shimia & Stslsy (1991) Spiridmov (1984) Spiridonov(I987) Spiidowstal (1984) Spiridow&Okrugh(1985) Spdrgs(l%g) Tru&&Kdttel(mpm) Twpt\etql (l9n) Wilga[iseral (1990)

35 I 4 8 19 3 6 II I 3 6 3 2 32 I 2

' ccluding wa omporitim o.lredy nputed in Kmloku & Ruinov (198Q. Tho omplete dcahs mybo obtaiaedu m E(CEL file fron the snd ailor.

GOLDFIELDITE, TTIE TELLIJRIAN TETRAHEDRITE$

rl23

l3 12 11 10 cu9
a Eb tr_ U trA tr-

_i.ji"::-:.
4
aA .tr A

I--------.I 7 6 5 0.0 0.5 1.0

oa

crl.isl;,:A

1.5

z.o

2.5

Ag+Au
Ftc. 5. Plot ofthe level ofAg + Au yerszs Cu in tellurian tetrahedrite",. All the values ale in apfu. T\e dashed lines indicate the range of possible substitutions. Compositions of alleged auriferous goldfieldite (Kovalenker & Troneva 1980) are shown as filled diatoo4.. ftiangles representsamplesfrom the Kochbulashk deposit (Novgorodovaet al. 1978, Kovalenker et al. 1980, Kovalenker & Rusinov 1986); squaresdepict all other samples (see Table 4).

13
trtr

12 11 +10
CD
? f

tr

-#"6ffi-"LFffi. -Ea A.o -R:

E-S --_!Hufrg.tr12tr-

{:^:-'
-a

".O-ad-

tr

--__

*o

;-;i>.-\<

,-9
aa

(J8

7
6

0.0

0.5

1.0

1.5

Te

2.5

3.0

3.s

4.O

Ftc. 6. Plot of Te content verszs that of monovalent metals M" (Cu, Ag, Au) in tellurian tetrahedrite," minerals. All the values are in apfu. The solid line shows the trend of the ideal Te versus M+ relattonship, whereas the dashed lines enclose the data that deviate by less than 0.5 apfufromthe ideal relationships. The circles refer to synthetic compositions. Other symbols as in Figure 5.

Trudu & Knittel 1991) to balance the increase in charges causedby the substitution of Te for trivalent semimetals. Figwe 7 shows the relationship between monovalent (mainly Cu) and divalent (6ainly Fe and Zn) cations in tellurian tetrahedrite"s. The plot confirms the substi-

tution of divalent by monovalentatomsto balancethe excesschargesresulting from incorporation of tetravalent Te (up to 2 Te apfu). only if the Te Iron occursin tellurian tetrahedrite"" contentis less than 2 apfu, owing to charge-balance

lt24

T}IE CANADIAN MINERALOGIST

1?

=
+ l1 U'
T

tr

tr E REt6 Fffift -ru tr

M + +M 2 + = 1 2 fonaa

A A -A A

= (J

10

tr tr

aa
I 0.0 0.5 1.0 1.5 2.O 2.5 3.0 3.5 4.O

Fe+Zn
Hc. 7. Plot of divaleatlt/P* (Fe,Zn)versus M+ (Cu, Ag, Au) metals in tellurian tetrahedrit"" minerals. All the values are n apfu. Symbols as in Figure 6.

4 Fe+Zn
o a

A -trn A

.^
A tr trtr trtr

ooo A

trtr *-E

atr tr n-a us !tu rt--\A tu-

tr fl

P4^

0.0

0.s

1.0

l.s

2.o Te

2.5

3.0

3.5

4.O

Frc. 8. Plot of proportion of divalent ions lVa + (Fe, Zn) versw Te tn tellurian tetrahedritess. All the values ue in apfu. The solid line in the diagram shows the fend of the ideal Te versus Ar12* relationship. Symbols as in Figure 5.

* versusM + substitution. (Frg.8). Chamocket constraints investigated ArP This is not the casefor al. (1989b) the oxidation state and coordination of Fe in the tellurian tetrahedritesr, do plot as most compositions tetrahedrite"s and concludedthat divalent and trivalent closeto theline ofideal substitution. In contrastto these Fe occurinthe Ml andM2 sites,respectively, but the findings, on the basis of their study of synthetic lower oxidationstatepredominates in naturalminerals. tetrahedrite, found that at low Makovicky et al. (199O) Insynthetictetrahedrite"",thepresenceoftrivalentFeis iron contents,iron is predominantlyferric, whereas proportional (Cu + between to theexcess of monovalent iron is precations CurrFerSb+Sr: andCuroFezSbaSl3, Ag> lO apfu) resultingin a change in trendfor the ideal dominantlyferrous.Whetherthese findings arerelevant

GOLDFIELDITE, THB TELLURIAN TETRAHEDRITE$

lt25

to tellurian tetrahedritess is uncertain, because Cu in excessof 10 apfu is compensatedby Te4*(Fig. 5). Furthermore, Makovicky et al. (1990) concluded thatprobably all iron resides atthe M2 site, regardlessof its oxidation state. Ttre zinc-for-tron substitution in tetrahedrite"" was studied experimentally by Cambi et aI. (1965). As Zn only occurs in the divalent state, it must substitute for Fe in the Ml site with 4-fold coordination. A histogram of Fe/@e +Zn) n goldfieldite (Frg. 9) shows that rhere is complete sotd-solution befween pure Fe andZnendmembers; however, most compositions analyzed are either Zn- or Fe-rich. This could be a function of the composition of the hydrothermal fluids from which these crystals precipitated. The silver content of tellurian tetrahedriter" rarely exceedsI apfu; evenin thosefew casesin which it does (e.9., Kovalenker et al. 1980), the values are erratic. Harris (1990) cautioned that Ag enrichment can t2ke place during electron-microprobe analysis of Ag-rich tetrahedrite owing to migration of Ag toward the analyzed spot. Using EXAFS spectroscopy,Charnock er a/. (1988) have ascertained that in tetrahedriters, Ag occurs in the 3-coordinated M2 site, where it substitutes for Cu. Figure 5 supports this substitution, as previously discussed. However, the lack of data and the low concentrations of Ag in the tellurian tetrahedrite"" prevent us from evaluating the effect of Ag-for-Cu substitution on the cell dimension. It must be pointed out that samples of Ag-rich natural tehahedrite are characterized by an anomalously small unit-cell Glall 1972, Riley 1974). Among the other metals reported in goldfieldite, gold is the most common @orjsova et al. 1986, Knittel & Johnson,unpubl. data, Kovalenker & Troneva 1980, Kovalenker et al. 1980, Sakharova et al. 1984, Trudu & Knittel, in press). In general, it occurs in concentrations below lVo, the only possible exception being goldfieldite from epithermal gold depositsin the C.I.S. (Kovalenker & Troneva 1980, Kovalenker et al. 1980), where amounts of up to 10.68 wt.Vo Att have been reported. Kovalenker & Troneva (1980) concluded that Au substitutesfor Cu in tetrahedrite"". Spiridonov (1984) disputed the existence of Au-rich goldfieldite on the basis that the examples of auriferous goldfieldite with Au concenffations in excessof I wt.Vocontain quantities of Fe and Zn (tp to 4.16 apfu in one sample; Fig. 8) that are incompatible with the maximum content of 2 apfu of divalent elements in tetrahedrite,". Furthermore, X-ray-diffraction patterns of auriferous goldfieldite by Kovalenker & Troneva (1980) indicate that aralyzed,grains are inhomogeneous(Spiridonov 1984) and may be considered as intergrowths of goldfieldite, Fe-sulfi de and calaverite. In Figures 5 and 6, the data ofKovalenker & Troneva (1980) for Au-rich goldfieldite display a rrend differenr from the l-to-l substitution between Cu and the

precious metals. Furthermore, the structural formulae calculated for these samples are not charge-balanced. TheseobservationssupportSpiridonovos(1984) conclusion that the examples of auriferous (>l wt.Vo Au) goldfieldite of Kovalenker & Troneva (1980) probably are mixtures of several minerals. Therefore, it appears that only a limited degree of Au incorporation is possible in goldfieldite; we infer that Au, which falls in the same column of the periodic table as Ag, probably occupies M2 sites and is monovalent. Cadmium concentrationsof less than O.lO wt. Voin tellurian tefrahedrite"" phases were reported by Spiridonov et al. (1984),Boisovaet al. (1986) and Spiridonov (1987). [n these minerals, the Cd-for-(Fe,Zn) substihrtion may be similal to that proposedfor tetrahedriteby Johnson et al. (1986) and confirmed by Charnock et al. (1989a), in spite of the substantial difference in ionic radii between 4-coordinated divalent Fe and Za (0.66 and 0.64 A, respectively,from Table l) and Cd (0.84 A: Shannon1981). Manganesehas been reporled to occur only in tellurian tetrahedrite""in concentrations up to 0.24 wt. Vo (Spiridonov et al. 1984, Shimizu & Stanley 1991) and is also uncommon in tetrahedrite""in general. Basu et aI. (1984) reported Mn concentrations as high as 5 .74 wt.Vo from the Rajpura - Dariba deposit in Raj astan (India), characterized by low contents of Fe and Zn, indicating that Mn in goldfieldite occupies a tetrahedral position, substituting for divalent Fe and Zn as in tetrahedrite. Mercury, not surprisingly, has been found in goldfieldite, as there is a pure mercurian end-member, schwazite of the tetrahedrile"" series. However, Hg in tellurian tetrahedrite," has been reported only from the

(J

z--

co

=.
l!

O r4l 40 &

10

20

30

4A

50

60

70

80

90

100

1 0 0F e Fe+Zn Flc. 9. Histogram of theratio 100fFe/(Fe + Zn)] in tellurian

tetrahedrite"".

tt26

TT{E CANADTAN MINERALOGIST

Akturpak deposit in Russia (Kovalenker & Rusinov 1986). Kalbskopf (1971) showed that divalent Hg is 4-coordinated in mercurian tetrahedrite, thus substituting for other divalent metals such as Fe and Zn. ln 4-coordination, Hg has a radius of 0.84 A lshannon 1981),the sameas Cd. The X site The X site is of considerable interest, becauseit is where Te-for-(As,Sb,Bi) substitution takes place. Figure 10 shows the occupancy of this site; there is considerable scatter, if all data are considered (from 3.4 to 4.6 apfu), but the values cluster within the 3.9 to 4.2 apfu ntewal. The relatively large spread of the data, which does not occur in the data obtained by Spriager (1969), Kase (1986) and Trudu & Knittel (in press), may, in some instances,be due to the standardization of S on unsuitable reference materials, as previously discussed. Tellurium. In nature, tellurian tetrahedrite"" crystals contain up to 3.4 apfu of Te; there are no gaps in this sotd solution (Figs. 6, 8, I 1). Figure 11 presentsa plot of the concentrationof Te yersas the sum of As, Sb and Bi concentrations. These data points show a very high correlation coefficient (0.97), supportingthe assumption that Te substitutes for any of the three semimetals. Arsenic, antimony and, to a lesser extent, bismuth, are the most common elements substituting for Te in goldfieldite. Johnson et al. (1986) inferred that Sb-for-Te substitution may be prevalent, but our data, plotted as Figure 12, indicates that there is no preferred substitution between any of As, Sb and Bi. Whereas virtually unlimited substitutions is observed among As, Sb and Te (up to a maximum of 3.5 Te ap[u), the c,,rrently known maximum Bi contentsis I apfu (Fig. l2E). This is not surprising, because a pure Bi-bearing tetrahedrite""end-member has never been doc'mented to occur in nafure. TheYan"dZsites In goldfieldite, the twelve Yand one Z sites are occupied by S and Se, and not by Te, as believed earlier (e.9., Nowacki 1969). Figure 13 presentshistograms of the total number of anions per formula unit; the upper diagram ofFigure 13, basedon all available data"shows a rather broad spread from 11.3 to 14.0 apfu S + Se, with most data clustering around 13 apfu.ln the lower histogram, the data of Kase (1986) and Trudu & Knittel (in press) are displayed; these cover the entire range of the Te versus (As + Sb + Bi) substitution and were obtained under optimal analytical conditions. These data cluster in the very limited range of l3.O X 0.3 apfu S + Se. This suggeststhat the large spread in fhe complete dataset (including data of Knittel 1989) results from the use of an inappropriate S standard, as previouslv discussed.

S is replaced to a maximum of 2.8 apfu by Se (Fig. l4). Tellurian tetrahedrite with an elevated Se content was found only in samplesfrom the Ozernoye epithermal Au deposit (Kamchatk4 Russia: Kovalenker et al. 1989, Spiridonov & Okrugin 1985), and ftom the Wild Dog epithermal Au systemon Bougainville Island @apua New Guinea: Noviello 1989). As the experiments at 340"C by de Medicis & Giasson (1971b) on the system Cu-Te-Se failed to produce the Se-bearing homologues of goldfielditeo we cannot assesswhether complete Se-for-S substitution exists in goldfieldite, as it does in sulfides @etlke & Barton 196l). Other elements In addition to the aforementioned elements, V, Co, Ge, Sn and Pb have in cases been reported in fface amounts in goldfieldite. Vanadium so far has only been reported by Spiridonov et al. (1984), who found 0.02wt.Va V in goldfieldite from Uzbekistan, a value so
160

(J lll f

z cy
rl! E,
L

3.2 30

4.2

4.4

o
3.8 4.0
4.4

TOTAL SEMFMETAL

ATOMS

PER FORMULA UNIT

Flc. 10. Histograms of the occupancy of the X site by semimetal (As, Sb, Te and Bi) ions. The upper diagram represents all data on tellurian tetrahedrite""; the lower histogram shows compositions reported by Kase (1986; hatched pattern) and Trudu and Knittel (in press.; dotted pattem) for comparison.

GOLDFIELDITE. TTIE TELLTJRIAN TETRAHEDRITE."

tln

E3 +
.cl a4
T

,2

tt,

-?\l{l tr oo oG

-I?

trr -tr _U

As+Sb+Bi+Te=4

o
0.0 0.5 1.0 1.5 2.0 Te 2.5 3.0 3.5 4.O

Flc. 1I . Plot of Te contgntversw the sum of As, Sb, ard Bi. The solid line in the diagram showsthe trend of the ideal Te-for-(As + Sb + Bi) substitution.All valuesarc h apfu, symbols asin Figure5.

close to the detection limits that the reality of V incorporation in goldfieldite may be questioned. Besides, in an exploratory study of minor-element incorporation in synthetic tetrahedrite, Makovicky & Karup-Mg1ler (1994) found that V does not enter this mineral. Reported contents of cobalt likewise are very low [Spiridonov et al. (1984) reported 0.03 wt.% Co in the same sample in which V occursl, but Makovicky & Karup-Mgller (1.994) were able to synthesize the Co end-member of tetrahedrite. Concentrationsof up to O.l7 *.Vo germaniumwere reported by Spiridonov et al. (1984). Considering that Ge is invariably tetravalent (like Te), the most likely location for Ge is the X position, where it may substitute for As, Sb, Bi or Te. The only problem is that Ge tends to be 4- or 6-coordinated in sulfosalts, whereas the X site is a trigonal pyramid. This fact probably accounts for the very low Ge concentrations in goldfieldite. Tin may be a relatively common trace constituent in the tellurian tetrahedriter" from the Kochbulashk epithermal gold deposit in Uzbekistan: Kovalenker et al. (1980) reported up to 0.22 wt.7o Sn, and Novgorodova et al. (1978), up to 2.95 wt.% Sn for tetrahedritess from this deposit. However, the data by Novgorodova et al. (1978) referred ro tetrahedrite samplesthat cont2in only trace emounts (< | wL.Vo) of Te. Considering that Sn falls in the 54ms sohrmn of the periodic table as Ge and tends to occur in nature in its tetravalent oxidation state (e.9., in srannite, colusite, etc.), it is very likely that this element can best be accommodated within the X sites of the tellurian mem-

bers of the tetrahedrite"", substiruting for the semimetals like Ge. Izad, is another element that rarely occurs in goldfieldite (Kovalenker et al. 1980,Kovalenker & Rusinov 1986); its maximum concentration reaches0.41 wt.Vo. In tetrahedrite"", however, values up to 12.3 wt.%ohave been recorded by Vavelidis & Melfos ( 1997) and Bishop et al. (1977). These studies show that plumbian teDnantite is low in Fe and Zn; therefore, these authors infer that Pb substitutes for these divalent elements. In contrast, based upon data from letrahedritess minerals with Pb concentrationsof less than 2 w.Vo,Basu et al. (1981) raised the possibility that Pb substitutesfor the semimetals in the X site, becausein some sulfides and sulfosalts, it can occur in 3-fold pyramidal coordination, which is the same as that of the X site. They still suggestedthat Pb should be divalent. Makovicky & KarupMpller (1994) synthesized tetrahedrite with up to O.4Pb apfu, whereas attempts to synthesize Pb-rich tennantite failed (<0.1 Pb apfu). All tetrahedrile"" containing significant 4mounts of Pb (both synthetic and natural material) contain ca. 4 (As,Sb) apfu,hence we believe that Pb substitr.rtesfor Fe and Zn. The preceding discussion leads us to propose the following chemical formulae: 1) for tellurian tetrahedrite and tennantite (Frg. 17 shows that tellurian Bi-bearine tetrahedrite does not occur in nature): N ]Cn' a*SF e,Zn Mn,Co,Cd,Hg,Pb)z*z-,1 m11Cu,Ag,Au)*6lN[(As,Sb,Bi)3*q-,(Te,Ge,Sn)4"]

rv11s,se;i12u11s,se;1

lr28

THE CANADIAN MINERALOGIST

a
4

As

Sb

i-f-ifu
+ .rrffi. i f*--+

2.O
1.5
+

(c)

11 . 5

12.5

13.0

T O T A LS + S e A T O M S
Frc. 13. Histograms of the occupancy of the I and Z sites by (S t Se) in tellurian tetraledrite"". All values ue n apfu. The upper diagram represents all available data on tellurian tetrahedrite""; the lower histogram shows analytical results of Trudu & Ifuiael (in press; dotted pattrn) and Kase (1986; hatched pattem) for comparison.

Bi 1.0 0.5 0.0

+ + ++ I f*+++

i** +.+

+ -l # . .Sf+t+ + * .* - 'tl f +

r+ -c*i{.-J +

*+

Te
yer.rus (a)As, (b) Sband(c) Bi. All Ftc. 12.Plotsof Te content the valuesaren apfu.

Notr,mNcraruns No official system of nomenclature exists to define tetrahedrite"r minerals according to their Te content. Two definitions of goldfieldite exist: Kato & Sakurai (l97oa) and Spiridonov (1984) suggested16 name the tetrahedrite"" goldfieldite where Te is the most abundant whereasDmitriyeva et al. (1987) suggested semimetalo that the mineral be named goldfieldite only where Te is greater than the sum of the other three semimetals. The former definition conforms with the rules on the nomenclature of the tetrahedrite"" minerals set by the IMA (Spiridonov 1984). Kovalenker et al. (1990) proposed an equilateral triangular projection with Te (goldfieldite), Sb + Bi

w h e r e0 ! x < 2 , a n d 2) for goldfieldite, which according to Figure 17 can be pure, arsenoanor stiboan, the formula is: rv1cu*61r11Cu,Ag,Au)+6,lIv[(As,Sb,Bi)3+2-, (Te,Ge,Sn)421rv;1S,Se;t21u11S,Se;1 w h e r e0 < x 1 2 .

GOLDFIELDITE, TTIE TELLTJRIAN TETRAHEDRITESS

It29

Se

Wild Dog (Noviello, 1989) Kamchatla (Kovalenker et al. 1989)

a Ozernoye(Spiridonov & Okrugin, 1985)

I

s\ "q

"rj

10

1i

12

t5

FIG. 14. Plot of S content yers us Se (apfu) illustrating the substitution between these elements. The line in the diagram corresponds to a total of 13 (S + Se) aplz. Symbols as in Figure 5, except for ttrose for Se-rich compositions, which are identifred in the figure.

(tetrahedrite) and As (tennantite) at the vertices. They divided it into six fields: Sb-goldfieldite, As-goldfieldite, Te-tetrahedrite, Te-tennantite, As-tehahedrite and Sb-tennantite.We believe that this nomenclature for the tetrahedrite"" is inadequate for the following reasons: 1) it implies that the Sb substitutes for Bi only, an interpretation that is not always correct (see below). 2) Some misnaming is likely 1s take place near the vertices, where nearly pure end-memberswill have the prefix of the second most abundant elemento which is presentin insignificant concentrations.3) The terms used by Kovalenker er aI. (1,990) to name the different fields do not conform the IMA recommendations (Nickel & Mandarino 1987): the use of a hyphenated chemical symbol preceding the name of a mineral is incorrect and is to be avoided. Thereforeowe suggest a new chemical nomenclature gdd for tetrahedrite"", in which the M sites are predominantly occupied by Cu, and S is the main anion. Ideally, asthere is probably completesolid-solution amongstAs, Sb, Te and Bi in the tetrahedrite"", a tetrahedron with each end-member occupying a vertex constitutes the most appropriate representation(see center of Fig. l5). We neglect ofher elements, such as Ge and Sn, which may be accommodatedin theXsiteo becausethey occur in insignificant amountsonly in goldfieldite. The values used for this plot must be in apfu and normalized to 1007o.However, as three-dimensional representations are very difficult to portray on a sheet of paper, it is more convenient to use the four faces of the tetrahedron individually. If only three of the four elements are present, then the selection of the appropriate triangle is obvious. If all four elements occur in the mineral, then an approximation has to be made: it is proposedthat the

triangle containing the three most abundant elements be used. Geometrically, this form of approximation is performed by extending the line connecting the vertex of Ieast abundant element to the point inside the tetrahedron representing the composition of the mineral until this line intersects the face of the tetrahedron. Numerically, this procedlue is followed by recasting the quantities of the three most abundant semimetals to lOOVo. This geomeric approximation has the advantage of retaining the proportions amongst the three most abundant semimetals as they occur ia the original structural formula. Considering that in tetrahedrite""otle least abundant semimetal, where present,occurs in very low concentrations, the error introduced by this procedure is negligible. The four triangles of the "exploded" tetrahedron representing the tetrahedrite"" are shown in Figure 15. Regarding the nomenclature of the different fields in the triangles, we have followed these criteria: 1) the tetrahedrite"" mineral is named after the most abundant semimetal present, as is usually done. 2) The same principle is applied to distinguish between a pure end-member (i.e., it must have a ratio greater than 3:1 between the two most abundant semimetals) and one that has substantial amounts of another element. In the latter case, the appropriate adjectival modifier has been used, as outlined by Nickel & Mandarino (1987). 3) The lack of an approved name for the Bi-bearing end-member of the tetrahedriters creates some problems in nomenclature. In the Soviet literature, the term "anniviteo', eponymous after the Annivier Valley in Switzerland, is commonly used for this mineral. However, this name is not officially recognizedby the IMA

I 130

THE CANADIAN MINERALOGIST

As
Te

sb

Ae

group in which the Ftc. 15. New nomenclature grid for minel4ls of the tetrahedritess M sites are predominsndy 6ssupied by Cu. The triangles correspondto the faces ofthe tetrahedron shown in the center of the figure. Abbreviations: Gfd: goldfieldite, Ten: tennantite, Tet tetrahedrite, BlTet: Bi-bearing tetrahedrite, Ars: arsenoan, Bis: bismuthoan, Stb: stiboan, and Tell: tellurian.

@.H. Nickel, pers. commun. 1991), becausea natural mineral close to Culq(Fe,Zn)2BiaS13 hasnot beenfound; one of the highest Bi contents in a tetrahedrite"" mineral is 1.60 apfu reported by Lur'ye and coworkers (Jambor & V4nko 1990). Therefore, the name of "Bibearing tetrahedrite" has been chosenobearing in mind that the name 'letrahedrite" does not imply that Bi substitutes exclusively for Sb. The proposed chenical nomenclature aims at emphasizing the chemistry of tetrahedrite"" minerals. It conforms to the recommendation of the IMA for temary solid-solutions (Nickel 1992) n the sense that in each triangle the three mrin areas subdivided by the 5OVorule keep the name of tle end-member. The further subdivisions are informal. The structural formulae of all the samples in the database have been calculated following the criteria previously outlined; the results,which exclude synthetic goldfieldite, are plotted in Figure 16 according to the

of Figure 15.Most of the compositions nomenclature fall within theTe-Sb-As triangle;only a few plot in the Te-As-Bi triangle;in only four is Te the leastabundant fall Only about llVo of the compositions semimetal. within the goldfieldite field, the majority being either The remainor stiboan, but not bismuthoan. zrsenoan evendistribushowanapproximately ing compositions tion in all the other fields (tellurian tetrahedriteand andstiboantennantite); tennantite, arsenoan tetrahedrite only a few clusterin the fields of the pure brahedrite and tennantite. DrscussroN pointof view, two aspects Froma crystallochemical involvdeserve furtherclarification:(a) thesubstitutions ing atoms with relatively large differencesin radii (e.g., Cd for Fe and Zn), and (b) the absence of

GOLDFIELDITE, THE TELLURIAN TETRAHEDRITE$

I 131

goldfieldite with a Te content greater than 3.5 apfu in nature. Although goldfieldi0e and tellurian tetrahedrite"" are ore minerals, they have neverbeen reportedto occur in quantities of economic significance within orebodies, unlike tehahedrite, tennantite and freibergite. Sakharova et al. (1984) regarded goldfieldite as "typomorphic of gold-telluride deposits localized within volcanic regions". Knittel (1989) suggested thar As-rich goldfieldite is more common in "polymetallic Cu mineralizationo',whereas the Sb-rich variety seems to be prevalent in'base-metal-poor precious metal deposits". We will reassess these hypotheseson the basis of our database. Subsfintions of relatively large atoms in tellurian tetrahe drit e* minerals One of Goldschmidt's rules (Faure l99l) on substitutions in minerals statesthat the radii of the two atoms involved in the substitution should not differ by more than 157o.In tellurian members of tehahedrite"",there are a few casesthat do not follow this rule: they involve Cd and Hg for the Ml site, Ag for M2, and As for X. Considering that tetrahedrite"" minerals have a sfructure that is sufficiently "flexible" to allow distortions, the replacement of smaller atoms (e.9., Cu, Fe and Zn) by larger ones (Cd and Hg) can be accommodated through rotation of tetrahedra around their 4-fold axes CN.E. Johnson,pers. commun., 1992). This is very likely the case for goldfieldite, in which the Cd and Hg concentrations are extremely low. The substitution of Ag for Cu in tetrahedrite""has puzzled scientists for a long time, as with more than 4 Ag apfu the cell size decreases, rather than increases, as suggested by the larger size of Ag. Chanock et al. (1988) anributed this to Ag - Sb interaction as deduced ftom EXAFS spectroscopic results. In goldfieldite, however, the maximrrm observedAg content barely exceeds I apfu;thtsitoutside the high-Ag region, and it is likely that the Ag-for-Cu substitution is accommodated through distortion. The Te-for-As substitution is reflected by a change in the cell dimension, as previously shown. Absence of end-membergoffieldite in nature There appear to be no crystallographic reasons to explain why end-member goldfieldite does not form in nature. Hydrothermal synthesisof goldfieldite was carried out by Kalbskopf (1974): his runs lasted 25 hours at 230oC under vapor-saturated conditicins and produced goldfi eldite (88 wt.%o), digenite (8 wt.Vo)and native Te (4wt.Vo). Kalbskopf (1974) infened that 180'C should be considered as the lower limit of stability of goldfieldite, and favorable pH conditions are neutral to mildly alkaline. Kovalenker & Rusinov (1986) suggested that the formation of goldfieldite requires oxidizing conditions and high activity of sulfur. Dry

Bi-Tet I o"

Sb

ArsTet

StbTen

AS

Frc. 16. Classification of 361 samples of natural goldfieldite (electron-microprobe data). In the Te-Sb-As plot, the crossesindicate compositions for which Bi was not sought, whereas the circles refer to compositions with values for Bi. The abbreviations of the fields are as the same as in Figure 15.

tt32

THE CANADIAN MINERALOGIST

synthesis within the system Cu-Te-S was undertaken by S. Karup-Moller (unpubl.) to determine phase relationships at 350o, 450o, 550o, 675', 800' and 900"C. Only runs at 350o and 450oC produced goldfieldite: at the former temperature, it coexists with digenite, covelliteonative Te, phaseA (CusoTez:S+z) and a liquid of composition Te13Ss7, whereas at 450"C, goldfieldite is in eqnilibrium with digenite, covellite and a liquid of composition Cu y9.sTes6.5Ss.5. These experiments indicate that pure goldfieldite is stable at certain conditions and could precipitate under natural conditions given the optimum composition of hydrothermal fluid. However, natural hydrothermal fluids from which tellurian tetrahedrites"precipitate probably always contain at least minor amounts of As, Sb and Bi, hence these elements are incorporated to some extent in the solid solution. Ore deposits hosting goldfieldite Our review shows that tellurian tetrahedrite"" is most commonly found in epithennal gold deposits.Although we have only scant information on the mineralogy and alteration of the majority of these deposits, it appears that high-sulfidation epithermal gold deposits are the most common host of this sulfosalt. This is not sumrising, as high base-metal contents and the pr"."n"" of sulfosalts (e.9., enargite - luzonite) are two of the characteristics of high-sulfidation systems (White & Hedenquist 1990). Five occurrences of goldfieldite in porphyry environments are reported. At Tirad, in the Philippines, goldfieldite is associated with deep-seatedadvanced argillic alteration (Trudu & Knittel, in press), which cross-cuts the main porphyry-style mineralization and shows strong similarities to the high-sulfidation l,epanto Cu-Au-As deposit, where Gonzalez (1956) reported the occurrence of tellurian phases. At the Marian gold deposit, also in the Philippines, goldfieldite belongs to an assemblageof minerals typical of low-sulfidation epithermal systems that precipitated in polymetallic veins at the periphery of porphyry-Cu minqall2alisl (Ituittel1989). At Butte, Montanao Springer (1969) reported the occurrence ofgoldfieldite in association with colusile, but he did not give precise indications regarding the timing of precipitation of these minerals. There is only very limited referenceto the timing of the deposition of colusite at Butte: Thompson (1973) reported it in the "horsetail zone", which is characterized by predominantly sericitic and minor advanced argillic alterations (Meyer et al. 1968), indicating that colusite and the associatedgoldfieldite precipitated toward the end of the geochemical evolution of the porphyry-style minslalization at Butte. In the Goldfield area Nevada.orebodies form irregular sheets and pipes within systems of silicified veins hosted by extensively acid-sulfate (alunite - quaftz) altered rhyodacite, and it has been speculated that a porphyry-type system may lie at depth

nearby, although probably not directly beneath Goldfield (Wallace 1979). Finally, at the Murgul copper deposit in Turkey (Willgallis et al. l99O), bismuthoan tennantite with up to 0.87 wlVo Te has been analyzed, but no data on its paragenetic position were provided. Schneider et al. (1988) showedthat "fahlore" is related to the latest stage of alteration, which is siliceous in character; although they infer a volcanogenic origin for the Murgul deposit, most of their observations appear to be indicative of a porphyry-Cu style mineralization. The occurence of goldfieldite in the Besshi-type depositsin Japan(Kase 1986) suggeststhat this mineral also is associated with volcanogenic massive sulfide (VMS) mineralization. Kase (1986) reportedthat loweramphibolite-grade metamorphism affected the Sazare ard Ikadazu deposits, and that tennantite and tellurian tennantite, together with chalcopyrite, bornite and sphalerite, occur in the interstices of granular pyrite. The description does not provide enough detail to infer whether the tellurian minerals of the tetrahedrite,, forrned prior or during metamorphism. The Bittibulashk deposit, in Russia, where tellurian tennantite occurs (Loginov et al. 1983), likewise appears to be a VMS deposit. We conclude that goldfieldite and other tellurian members of tetrahedrite"" occur in both high- and lowsulfidation epithermal Au mineralization, but also can form during the last stages of porphyry-style mineralization and in VMS deposits.Our review also shows that As and Sb enrichment in goldfieldite is not related to specific styles of mineralization, as Knittel (1989) suspected. Triangular Te - Sb - As and tetrahedral Te - Sb - As - Bi plots @g. 17) show that goldfieldite from the Tirad deposit in the Philippines (Trudu & Knittel, in press) and the Megradzor deposit in Armenia are enriched in As. In contrast, at the Goldfield deposit, Nevada (Knittel & Johnsonoin prep.), the Au deposits of the Kamchatka Peninsula, Russia (Kovalenker et al. 1989) and the Au deposits from the Kuraminski Range in Uzbekistan (Novgorodova el aL 1978), the tellurian members of the tetrahedritess are enriched in Sb. All these deposits display characteristics typical of epithermal-style mineralization. Another interesting aspect of the data displayed in Figure 17 is that in each deposit, the ratio of the two most abundant trivalent semimetals is approximately constant, whereas the Te content is variable. This may reflect constant As/Sb and variable Te conlents in the hydrothermal fluids from which the tellurian tefrahedrite precipitated, but may also reflect a variable abiliry of tetrahedrite"s to accommodate Te, depending on the physicochemical conditions. Thermodynamic data regarding the As-Sb-Te exchange are not available. In their study of the thermodynamic properties of tetrahedrite and tennantite, Sack & Loucks (1985) stated that "tetrahedrites-lennantites with intermediate Sb/(As + Sb) ratios have greater As/Sb ratio than the

GOLDFIELDITE, TIIE TEIITURIAN TETRAHEDRITEss

I 133

Tlrad deposlt, Phlllpplnes (Trudu and Knittd, in prep.)

Megradzor deposit, Armenia, C.l.S. (Kovalenker et al., 1990)

Sazare and lkadazu deposits, Japan (Krq 1886)

Goldfreld deposit, Nevad4 U.s.A. (Knifiel and Johnsorx in prep)

Karnchatka Peninsula depodts, Russla, C.l.S. (Kovalenker et al., 1989)

te

Te

Kuraminski Range dPostts' uzbekistan, Cl-S(Novgorodova et al., 1978)

As

As

Frc. 17. Triangular plots of tellurian tetrahedrite,, minerals from selected ore deposits. The dashed lines represent the best fit to the data of each olot.

n34

THE CANADIAN MINERALOGIST

hydrothermal solutions from which they precipitated". Therefore, the ratios amongst semimetals in tetrahedrite""minerals are not directly indicative of the composition of the fluid from which they precipitated. In several ore deposits in which tetrahedrite-tennantiteis the only As- and Sb-bearing phase, its Sb content increases with the evolution of the hvdrothermal fluid (Sack & Loucks 1985). Cor.tcl-ustows In the present study, we show that goldfieldite and other tellurian phases of the tehahedrite"" seem to be relatively rare in comparisonto other speciesof the seme group, such as tehahedrite, tennantite and freibergite. This may be due to the fact that Te is rarely enriched in nature. Furthermore, Te has not been routinely sought during electron-microprobe studies of sulfosalts. Incorporation of Te into tetrahedritessleads to coupled substitutions to maintain an overall charge-balance. These substitutions, initially proposed by Kase (1986) and confirmed by data obtained by Trudu & Knittel (1991) and Shimizu & Stanley (1991), can be summarized as follows: For replacement of up to two atoms of trivalent semimetal by letravalent Te, divalent Fe and Zn are replaced by monovalent Cu. Higher Te contentsare compensated by^a reduced number of Cu atoms. Interestingly, Cu Il2 * substitution in goldfieldite takes place in the Ml site, whereas the vacanciesoccur in the M2 position. Where substitutions iavolving elements of different valence take place, then the charge balance in the mineral can be maintained in two ways, either through a coupled substitution or through a vacancy in the structure. In renierite, which according to Bernstein et aL (1989) has the formula Curo(Znr-,Cu")(Gez*,As)FeaS 16 where 0 (-r ( 1, on-lythe coupled substitution occurs: as trivalent As replaces tetravalent Ge, monovalent Cu takes the place of divalentZn. An example of the second mechanism is shown by colusite (S. Merlino, pers. commun., 1992): the charge balance in the solid solution between the Sn-rich (Cu26V2Sn2AsaS32) and Sn-poor varieties (Cu2aV2As6S3) is maintainedthrough vacancies in one of the Cu sites (note that As is pentavalent in this mineral) with a reduction of Cu atoms from 26 to 24. As shown throughout this paper, goldfieldite seems to be unusual in that both mechanisms of maintaining charge balance occur, as Te progressively substitutesfor As, Sb and Bi. We hope that this work will make geoscientistsmore aware that Te (and also Bi and Se) may be significant components in minerals of the tetrahedrite$ series. With more analytical data, we will be able to re-assess whether a specific signature in mineral composition is fypical of a certaia style of mineralization, Petersen& McMillan (1992) have shown an application of the compositional data of the letrahedrite"" minerals to min-

eral exploration and mine development; however, they have not extended their work to the tellurian members of this series. We have experimental work in progress on the tetrahedrite"" with variable Te content in order to measure the coordination of Te through EXAFS spectroscopy and the variation in cell size as a function of crystal chemistry. We hope that one day the work on the thermodynamic properties of tetrahedrite and tennantite will be extended to goldfieldite as wello so that we can gain a better understandingof the conditions under which it precipitates in ore environments. AcIc.{owLEDGEMENTS Special fi1anksgo to Neil Johnson for thoroughly reviewing the manuscript. Gus Mumme, Nick Oliver and Mike Solomon commentedon some drafu. Discussions with Aldo Cundari, Alex Losada-Calder6n and Marc Marshall were very helpful. Ernst Burke, Stefano Merlino and Ernie Nickel made important suggestions. Sven Karup-Mgller kindly provided a preprint of his manuscript on the Cu-Te-S system. Tom Hall's assistance with the geometrical approximation for the classification of goldfieldite is appreciated.Finally, we thank E. Makovicky for a very insightful review, an anonymous reviewer for his positive review, and the editor, Robert F. Martin, for editorial help. RErenslrcEs BAsu, K., BoRTNtr<ov, N.S., Moorcrynrsr,N.N., Mozcove, A.V., Srvrsov,A.V., TsEprN, A.I. & Vmmr-evsxa;e, Z.V. (1984):Raremineralsfrom Rajpura- Dariba,Rajasran, India. V. The fust recordedoccurrence of manganoan fahlore. Neues. Jahrb.Mineral.Abh. I49.105-112. & Tsepnr,A.I. (1981):Raremineralsfrom Rajpura- Dariba,Rajastan, India. III. Plumbian letrahedrite.Neues.Jahrb. Mineral. Abh. t41,280-289. BERNsrEnr, L.R., Rercuel, D.G. & MERLTNo, S. (1989): Renieritecrystal structurerefined from Rieweld analysis of powder neutron-diffractiondata.Am. Mineral. 74, tI77 -1,18t. BEnrKE, P.M. & BenroN,P.B.,Jn. (1961):Unit-cell dimensionsversus composition in the systems PbS{dS, PbSPbSe,ZnS-ZnSe,and CuFeSl.e6{uFeSe1.e6. U.S.GeoI. Surv.,Prof. Pap. 4249,266- 270. Btssop,A.C.,Cnronlr, A.J.& CLARI<, A.M. (1977): Plumbian tennantite from Sark"Channel lslands. Mineral Mag.4t, 59-63. Bonrsova, E.A.,BoRoDAEv, Yu.S.&BrcHARova,G.I. (1986): Raretypesof terahedrite- tennantite oresof thea gold ore deposit.7np. Vses. Mineral. Obshthest.lLs,63-72.

GOLDFIELDITE, THE TELLURIAN TETRAHEDRITES

I 135

Caprnr,L., Er-r-r, M. & TANcERTNT, I. (1965): Prossesi Icu\,trrovA, series: tennantite, N.S. (1986):New isomorphous idrotermali.IV. Tetraedritisostituite.La Chimica e goldfieldite, Il5, annlite. Zap. Vses. Minzral. Obshchzst. L'lrdustria 47, 703-715. 449454 (in Russ.). Cnenrer, M. & LEw, C. (1974):Substitutions multiplesdans la s6rietennantite- t6ffaedrite.BuII. Soc. fr. Mindral. Cristallogr. m, 241-25O. D.A. (1990): New mineralnarnes. Am. Jnwon, J.L. & VANKo, Mincral. 75,706-713.

(1990):CD-ROMMineralpowderdiffraction: JCPDS JCPDS - IntemationalCenfe, Swarthmore, Pennsylvania. (1975):Influencedes principales & _ substitutions sur le parambtre cristallin dansla s6rie C.W. (1985):Crystalstrucnue M.L. & BunNsaM, tennantite- t6tra6drite. Bull. soc.fr. Min'ral. Crtsmllogr. JornqsoN argentian tetrahedrite. refinementof an arsenic-bearing 98,152-158. O. Am. M ineral.70, | 65-I'1 CHenNocx, J.M., Gannsn,C.D., PArrRrcr, R.A.D. & modelfor & Jnaxroz,R. (1983):A Brillouin-zone Veucuax, D.J. (1988):lnvestigationinto the nafire of variationi-ntetrahedrite. Am. Mineral. 68, compositional copper and silver sites in argentiantetrahedriteusing 220-226. EXAFS spectroscopy. Phys.Chem.Minerals 15, 296-299. &(1989a): Coordination sites of metals in tetrahedrite minerals determined by EXAFS. J. Solid State Chem.82,279-289.

JonNsoN, N.E., Cnnrc, J.R. & RrusuDr, J.D. (1986): Can. Mineral. 24, Compositional trendsin tetrahedrite. 385-397. &25,237-244. desQuecksilbers im KAr-nsKopF, R. (1971):Die Koordination r. Mix. 18, 173-17 5. Schwazit.Tschermaks Minzral Petrog (1974): Synthese und Kristallstrukturvon Curz-*Te+Sr:, dem Tellur-Endgliedder Fahlerze. Mineral.Petrogr.Mitt.2I, l-l0. Tschzrmaks S. (1992): The phasesystemCu-Te-S at KARnI-MoLLER, 350',450', 675', 800o, and900'C.Unpubl.Report. from the Besshi-rype Kese, K. (1986):Tellurian tennantite metarnorphic belg lapan.Can. depositsin the Sambagawa Mineral.?4,39940r'. ofgoldfieldite K. (19'10a): Re-definition Karo, A. & SAKURAT, Cu2(Te,As,Sb)aS;p'. J. Mineral. Soc.Japan lO, 122-123 (in Japanese). & _(1970b): Goldfieldite from theKawazu
mine, Shizuoka prefecture, Japan. J. Mineral. Soc. Japan 10,122 (in Japanese). KwrrreU U. (1989): Composition and association of arsenian goldfieldite from the Marian gold deposit northern Luzon, Philippines. Mineral. Petrol. 40, 145-154. KovAr-ENrcR, V.A., NEKRAsov, I.J., SeNporm.srua, S.M., Nernasov,c, A.N., Malov, V.S., DANcHENKo, V.J. & DurrntvBve, M.T. (1989): Sulfur, selenium, tellurium mineralization in the Kuriles, Kamchatka volcanic belt. Mincral. 21L ll,3-18 (in Russ.).

g1,at65 &-(1989b); and Mdssbauer spectroscopic study of Fe-bearing tetrahedrites. Mineral. Mag. 53, 193-199. DEMEDrcrs, R. & Gnssor.r,G. (l97la): l,e systdme Cu-S-Te. C.R. Acad.Sci. Paris,SCr. C,?i72,1848-1851.
& _ (l97lb): Le systdme Cu-Se-Te. C.R.

(1987):Effect of substio-

tions on the cell dimension of tetrahedrite. Can. Mineral.

Acad. Sci.Paris,Sdr.D, n2, 513-515. DMrrRryBvA, M.T., YEFREMov, V.A. & Kovernrrcn, V.A. (1987):Crystalstrucnreof As- goldfieldite. Dokl. Acol. Sci. U,SSR, Earth Sci.Sect.2W, 141-1,44. Fnune, G. (1991): Principles and Applications of lrnrganic Geochzmistry.MacMillan Publishing Co.,New York"N.Y. FneNzpr J., Ar--TaneqcHarr, M. & NusER, B. , G., OfisMAr..T{, (1975):The Calabona copperore depositof Alghero, SardiniaNeues Jahrb.Mfucral.,Abh. I25.107-155. Gel,mr4 F.R. (1974):Ionicity, atomicradii, andstructue in rhe layereddichalcogenides of Group IVb, Vb, and VIb transition metals. "L SoltdStateChem.9,358-367. GoNzAr-Es, A. (1956):Geologyof the Lepantocoppermine, Mankayan, Mountainkovince. In Copper Deposits of the (A.J.Ilnkel,Jr. et al.,eds.). Philippines Philippines Bureau of Mines,Spec. PubL 16, 17-50. HAr, A.J. (1972):Substitution of Cu by Zn, Fe, and Ag in synthetictetrahedrite,Cu12SbaS13. Bull. Soc,fr. Minlral. -594. Cri stallogr. 95, 583 Henss, D.C. (1990):Electron-microprobe analysis. /z AdvancedMicroscopicStudiesof Ore Minerals (J.L. Jambor & D.J.Vaughan, els.).Mincral. Assoc. Can-, Shon-Course Vol.L7,319-339. Heocss,S.W. & SrEvENs, J.G. (1977):r2lsb Mdssbauer - antimony - sulfur spectroscopy. I. Studyof somecopper minerals. Rev.Chim rninlrale 14.331-338.

& Rusnrov, V.L. (1986):Goldfieldite:chemical

composition, paragenesisand formation conditions. Mizeral Zh. 8, 57-7O (in Russ.).

& TnoNeva,N.V. (1980):On the gold-bearing

fahlore. Proc. Twelfih Gen. Meeting, Int. Mfurcral. Assoc., 75-83.

I 136

TIIE CANADIAN MINERALOGIST

& DosnoN'rqm.rKo, V.V. (1980): Unusual composition of the main ore- forming minerals of the pipe-shaped ore deposit of Kochbulashk. /r Methods for the Investigation of Ore-Forming Sulfides and their Paragenesis. Moscow, USSR (140-164; in Russ.). ZALBEKJAN, M.A., LApurr{A, I.P., Mercv, V.S.,

in mineralnomenclanre. Nrcru, E.H. (1992):Solidsolutions Can. M ineral. 30. 23| -234. J.A. (1987):Procedures involving & Manpannqo, on New Minerals and Mineral the IMA Commission on mineralnomenclature. Can.MinNames, andguidelines eral. 25,353-377 .

Snwoovrnsrare, S.M.,Genasro,M.I. & Msnan:AN, D.V. M.I., TspDl, A.I. & Dwrnrrrva, M.T. (1978): (1990): mineralization Sulfide- telluride of theMegradzor Novconooova, New isomorphousseries in a group of tetrahedriteorefield, Armenia. Int. Geol,Rev.32,705-720. 107, goldfieldite ores.7np Vses.Mineral. Obshchest. l0 (in Russ.). 100-1 LEvY,C. (1967):Contribution i la min$mlegig des sulfures de cuiwe du type Cu3XS4. Bur. Rech.G6ol. Mini/res, Fluid InclusionSmdy, Novmlrc, S.P.(1989):Paragenzsis, Mdm.54. Microprobe Analysis and Geochemistryof the GoldTelluride Mineralizing Flails at thc Wild Dog Prospect, LND, I.L. & MAKovrcKy, E. (1982): Phase relations in thesysMonashUniv., Clayton,VictoPNG.B.Sc.(Hon.)thesis, tem Cu-Sb-S a120O"C,108 Paby hydrothermal analysis. ria. AustraliaMicroprobe analysesof tetrahedrite L warrrir'5.Neues Jahrb.Mineral.,Abh. 145"13+156. und Kristallchemie NowAcKI,W. (1969):Zur Klassification Mineral. Petrogr. Mitt. 49, der Sulfosalze. Schweiz. Loctrrov,V.P., MAGRBT, A.A., RusrNov, V.L., BoRrsovsKu, 109-156. S.E.& Nosx, L.P. (1983): Firstoccurrence of goldfieldite in a pyrite deposit Dokl. Al<ad.Nauk SSSR 273, 437440 S.,Duc}rt,G. & VrzzeuvI, G. (1981): Onr-aml,P.,MsRLn{o, (in Russ.). Colusite:a new occurenceand crystalchemistry.Caz. Mineral. 19,423427. Merovlcxx, E., Foncnrn, K., LorrenuosER, W. & AlvrmAUER, G. (1990):The role of Fe2"andFe3*in syn- PAULNG, L. (1948):ThzNatureof thz ChcmicalBond.Comell theticFe-substitutrd tetahedite. M ineral.Penol. 43,73-87. Universityhess, Ithac4 N.Y. & Kenw-Mor-Lsn,S. (1994):Exploratory snrdies & Nswvml, E.W. (1934): The crystal structureof on the substitutionof minor elementsin synthetic and binnite,(Cu, Fe)AsS,and the chemicalcomposition tetrahedrite. I. Substitution by Fe, Zn, Co, Ni, Mn, Cr, V structure of minerals of the tetrahedritegroup. Z. andPb.Unit cell parameter changes on substitutionandthe Kristallogr. 88, 54-62. structural role of "Cnzp'.Neues Jahrb.Mineral.,Abh.167, 89-123. E.U. & McMu-aN, D. (1992):Electronmicroprobe PerERsEN, in explorationand mine mineral analysis:applications development. Mining Eng.4, L39-143. MrrutncH,T.P.& Trupu, A.G. (1993):A laserRaman microprobe study of some geologicallyimportant sulfide Geology andoredeposits of Goldfield, RaNsot"m, F.L. (1909): minerals. ChzmGeol.I03. Il3-I27. Nevada.U.S.GeoI.Sarv.,Prof. Pap. ffi. MEvER,C., Suna, E.P., GoDDARD, C.C., ZuurN, L.G., GursERr,J.M.,MllB" R.N.,McAr-enn, Bno& G.B., Rn-ev,J.F. (1974): T\e tetrahedrite freibergite series,with J.F., orebody.Mincral. to the Mount Isa Pb-Zn-Ag reference INcnnsou, R.G., Bunr.s,G.J. & Wlcer, T. (1968):Ore Deposia9, lLl-124. depositsof Butte, Montana-Iz Ore Depositsof the United States, 1933-1967: Vol. tr (J.D.Ridge, the Graton-Sales propered.). Am. Insr Mining Metall. PetroleumEngineers, New Secr, R.O. & [.oucrs, R.R. (1985):Thermodynamic constraintson the interdeties of tetrahedrite-tennantites: York, N.Y. (137 3 -1416). pendence of Ag * Cu, Fe * Zn, C't * Fe, and As - Sb 0-1289. exchange reacttons. Am. Mineral.70, 127 (1981): Moonr, P.B. Lone-pairelectrons, anion packing, sulfosalts:towarda unffied picnre. Geol.Soc.Am-,Abstr. Sarnenovn, M.S.. Lsnsnsva, N.V. & Cnunnnov,V.M. Programs l3r 5l4. (1984):Fint find in Kamchatka of the rare Te minerals goldfieldite, rucklidgeite and native Te. DoH. Acol. Sci. Mozcova, N.N. & TsBpw, A.I. (1983)Fablores. NaukaPress, ?.il8,165-167. USSR, Earth Sci.Sect. (in Russ.). Muscor,,/, Russia C.M. (1988):RelaScr.rmen, H.-J.,Ozcun,N. & Par-qcros, Oznove, N.N., BoRTNtr<ov, N.S. & rareearthelement distribution, tionshipbetween alteration, Tnoweva,N.V. (1979):Mercuriferous grey copperores. and mineralizationof the Murgul copper deposit, northhp. Vses. Mineral. Obshch.est. lO8, 437453 (in Russ.). westernTurkey. Econ. Geol. 83, 1238-1246. NAKAT, I., SucrrANr, Y., Nrwa,Y. & Nacnsrmvre, W.C. (1990): K. (1976): Sear-, E.J. & KELLEv, R.R., II, EsseNe, Application of X-ray photoelectron spectra for copperore Tetrahedriteand tennantite:evaluationof thermodynamic minerals. Mincral. Soc.Japan, Abstr.,B4L (in Japanese). equilibria.Can.Minzral. ?4,725-738. dataandphase

GOLDFIELDITE, TIIE TELLI'RIAN

TETRA}IEDRITE5S

ll37

SnauroN, R.D. (1976): Revised effective ionic radii and systematic studies of interatomic distances in halides a:rd chalcogenides.A cta Cry stallogr. A32, 75 | -7 67 . _ (1981): Bond distancesin sulfides and apreliminary table of sulfide crystal radii. /z Structure and Bonding in Crystals 2 (M. O'Keefe & A. Nawotsky, eds.). Academic hess, New York, N.Y. (53-70).

TRUDU,A.G. & KNnruL U. (1991): Substitutions in hightellwium goldfieldite, the Te-bearing end-member of the tetrahedrite"" grotp. Geol. Assoc. Can. - Mineral. Assoc. Can. - Soc. Econ Geol., Progratn Abstr.16" I25. & -(1998): Ore textures and minglal shs6-

Srnnwooo, W.J. (1907): Gold tellurides. Mining and Scienrific Press 94,731-732. Sm,uu, M. & SreNmv, C.J. (1991): Coupled substitutions in goldfieldite - tetrahedrite minerals from the Ikiri mine, J apan.M ineral. Mag. 55, 5 15-519. SennoNov, E.M. (1984): Species and variefies of fahlore (tetrahedrite - tennantite) minerals and their rational nomenclatue. DokI. Acal..Sci. USSR,Eanh Sci. Sect. 279, 166-172. (l 987): Stoichiometry of fahlore compositions.

istry of goldfieldite, Cu-Te-Fe sulfides and Cu-Bi sulfosalts from the Tirad porphyry Cu-Au prospect, Guinaoang,northwesternLuzon, Philippines. Can. Mineral.36 (in press). M.I. & Dmrnvwa,M.T. (1977): TsepN, A.I., Novconouove, First find of goldfielditein the USSR.Dokl. Acad. Sci. USSR, Earth Sci. Sect.2M,156-158. VavsLrus, M. & MBr-ros,V. (1997): Two plumbian from Maronia area tetrahedrite- tennantiteocculTences (Ihrace) and Milos Island (AegeanSez).Eur. J. Minzral. . 9,653-65'7 of buried porWALLAcE, A.B. (1979): Possiblesignatures phyry-copper deposits in middleto late Tertiaryvolcanic Nev.Bur. MinesGeol.,Rep.33, rocksof westem Nevada. 69-76. environWnrre,N.C.& HsnENernsr, J.W.(1990): Epithermal variations and causes, mentand stylesof mineralization: for exploration. J. Geochem. Expl.36, and guidelines 445474. SeleniumN. & Sncvntw, E. (1990): Wu-rcnr-r-rs, A., Ozcr,n, of sulfrdes in thecopperoredeposits andtellurium-bearing Murgul,NE Turkey.Eur. J. Mirrcral.2, 143-148. of tetrahedrite Wuwscn, B.J. (1964):The crystalstructure 453. Z. Kristallogr.ll9, 43'7 Cu12SbaS13. Y. & Nowncn, W. (1966):The crystal TAIGucril, Z. Kristallogr,I?3, l-20. structure of binniteCurzAsaSp. May 20, Received July 7, 1992, revisedmanuscriptaccepted 1998.

Dokl. Acad.Sci. USS&Eanh Sci.Sect. 2Vl, 159-163. Cnvnrva, T.N., Soxor-ova, N.F.,Om,ucnt,V.M. & Bnnnr,ov,A.S. (1984):Effect of telluriumon optical propenies of fahlore (teEahedrite-tennantite)minerals. Dokl. Acad. Sci. USSR, Earth Sci.Sect.279. 193-196. & Olc,uctrr,V.M. (1985):Seleniungoldfieldite, a
new fahlore variety. Dokl. Acad. Sct. USSR,Earth Sci. Sect.

x0,474478.
SpRn{cER, G. (1969): Electronprobe analyses of tetrahedrite. Neues Jahrb. Mineral., Monatsh.,24- 32. TuowsoN, B.J. (1973): Disribution of primary mineralization and hydrothermal alteration wirhin the Berkeley Pit, Bune District, Monlana. In Guidebook for the Butte Field Meering (R.N. Miller, ed.)., Ll-L7. THoMpsoN,R.M. (1946): Goldfieldite = tellurian tetrahedrite: Univ. of Toronto Snd., Geol. Ser.50,77-78.