THE CANADIAN MINERALOGIST
Journal of the MineralogicalAssociation of Canada
Volume 23
Part.4
NOVEMBER 1985
Canadian Mineralogist
Vol. 23, pp. 517-534(1985)
A RE-EXAMINATIONOF THE ARSENOPVRITEGEOTHERMOMETER:
PRESSURECONSIDERATIONSAND APPLICATIONSTO NATURAL ASSEMBLAGES*
ZACHARY D. SHARP, ERIC J. ESSENE ANDWILLIAM C. KELLY
Deportmentof GeologicalSciences,The Universityof Michigon,Ann Arbor, Michigan48109,U.S.A.
Sotvtuetnr
ABsrRAcr
Experiments on the stability of arsenopyrite in the system Fe-As-S haveled to the useof arsenopyritecompositions as a geothermometer(Clark 1960a,b,Kretschmar&
fton 1976),but the effect of pressurehas not beenrcolved.
Volume calculationsshow that pressurewill shift sulfidation buffer-curvesin the systemFe-As-S to higher sulfur
fugacity/(S). Increasingpressurewill decreasethe arsenic
content of arsenopyritecoexistingwith loellingite, and will
increasethe upper stability of arsenopyritecoexisting with
pyrite by approximately t4"C/kbar. Partial molar volumes
for arsenopyritesolid-solutionsindicatethat pressurehas
a negligibleeffect on the compositionof arsenopyritecoexisting with pyrite, but sulfidation buffer-curves will be
shifted to higher /(S) by up to one log unit. If
arsenopyriteFeAsl -, 51*, Iies on the FeS2-FeAs2
binary
join, its composition should be fixed when buffered by
either pyrite or loellingite alone at fixed P and T. Experimentson the ternary systemFe-As-S havebeeninterpreted
to show that two phasesare required to fix the composi
tion of arsenopyrite at any P and T (Kretschmar & Scott
1976), suggestingmajor deviations from ideality for
arsenopyriteor possibleerrors in the experimentaldata.
Application of the arsenopyrite geothermometerappears
valid for depositsmetamorphosedto greenschistand lower
amphibolite facies(Oriental, Homestake: - 5@'C infened
and calculated), but yields low temperatures for deposits
metamorphosedto upper amphibolite and granulite facies
(Batmat, Geco: -650oC inferred, 5@oC calculated),and
inconsistent temperaturesfor low-temperature hydrothermal deposits(Panasqueira:-300'C inferred, 500'C calculated).
Keywords: arsenopyrite,geothermometry, geobarometry,
sulfide stability, sulfur fugacity, sulfide thermodynamics.
Les donn6esdisponiblessur la stabilii6de I'arsenopyrite
dans le systbrneFe-As-S ont permis l'utilisation des compositions de ce mindral comme g6othermomdtre (Clark
1960a,b, Kretschmar& Scott 1976),mais I'influence de
la pressionrestaitmeconnue.Le calcul desvolumesmontre qu'une augmentation de la pression ddplace les courbes des assemblages-tampondans les rfuctions de sulfidation du syst0meFe-As-S vers les fugacit€s plus 6lev6esdu
soufre. Aussi, cetteaugmentationdiminue la proportion
de I'arsenic dans I'arsenopyrite qui coexiste avec loellingite, et augmentele champs de stabilite de I'arsenopyrite
qui coexisteavecla pyrite d'environ |4oClkbar. Lesvolumesmolaires partiels pour les solutions sotdes d'arsenopyrite indiquent que la pression exerceun effet ndeligeable
sur la composition de cette phase en coexistenceavec la
pyrite, mais la courbe de la r6action-tampon est ddplac€e
n/(S) plus €lev€epar un ordre de grandeur. Si I'arsenopyrite FeAsl-rS1*, fait partie du systeme binaire
FeS2-FeAs2,sa composition serait fix6e par la seulepr6sencede pyrite ou loellingite d P et T fixes. On d€duit, d
partir des r€sultats d'expdriencesdans le systBmeternaire
Fe-As-S, que deux phasessont ndcessairespour fixer la
composition de I'arsenopyriteen termesde P et T (Kretschinar & Scott 1976), ce qui laisse prdvoir des d6viations
importantesde l'iddalitd de l'arsenopyriteet deserreurspossibles dans les donn6esexp€rimentales.L'application du
g6othermombtreparalt valide pour lesgitesmin6raux m6tamorphisdsaux faciesschistesverts et amphibolite inf6rieur
(Oriental, Homestake:5@oC, temp€ratured6duiteet calcul6e). Le gdothermombtre donne des temperatures trop
bassespour les gftesm6tamorphis€sdansles faciesamphibolite supdrieur et granulite @almat, Geco: temp6rature
ddduite 650"C, calcul6e500oC) et des r€sultats erratiques
pour les gisementshydrothermaux de bassetemperature
(Panasqueira:temp€ratured6duite300oC,calcul€e5@'C).
(Traduit par la Redaction)
*Contribution No..403 from the Mineralogical Laboratory,
Depafrmentof GeologicalSciences,University of Michigan,
Mots-cl*s: arsenopyrite,g6othermomdtrie,gdobarom€trie,
517
518
THE CANADIAN MINERALOGIST
stabilitedessulfures,fugacitddu soufre,analysethermodynamiquedessulfures.
temperature when equilibrated with either pyrite or
loellingite. Experimentson the binary join were carried out by Clark (19@a,b),who found that buffered
INTRODUCTIoN
arsenopyriteis enrichedin arsenicwith increasing
Many metamdrphosedore-depositshave signifi- temperature
and enriched in sulfur with increasing
cant economicimportance(e.g., Broken Hill, New pressure(Frg.
1). He concludedthat the maximum
South Wales; Homestake,South Dakota; Balmat, stability
of coexistingarsenopyriteand pyrite is apNew York), and an understandingof their meta- proximately 490oC
at I bar, but is extendedto 520oC
morphic history is warranted. Efforts to find sulfides
at 2 kbar. Kretschmar& Scott (1970 concludedthat
suitablefor geothermometryand geobarometryhave two Fe-As-S phases
are needed to buffer arbeen limited by the generally nonrefractory nature
senopyritecomposition at a given pressureand temof this group of minerals.Arsenopyriteis one of the perature,
and inferred that slight deviationsfrom the
few sulfides, besidessphalerite, that is relatively ideal
cation,/anionratio causearsenopyriteto plot
refractory and exhibits appreciablesolid-solution
off the simple FeS2-FeAs2.binaryjoin. Their exsensitiveto pressureor temperature(Clark 1960a,b, perimentally
determined eraph of temperature yerBarton 1970,Scott 1974,1976).Experimentson the
sus at.tlo As in arsenopyriteis shown in Figure 2.
stability of arsenopyrite in the system Fe-As-S
Kretschmar& Scott (1976)wereunableto duplicate
(Clark 1960a,b,Barton 1969,Kretschmar& Scott
the pressuredependence
observedby Clark and pro1976)were limited by extremelyslow rates of reac- posed
that the pressureeffect on compositionis less
tion, preciselybecauseof the refractory nature of
than Clark had reported. They suggestedthat the
arsenopyrite.We have evaluatedthe arsenopyrite gold
in the capsulesthat Clark had usedfor his exgeothermometerby analyzingnatural assemblages
perimentsmay have diffused into the arsenopyrite
whoseP-T historieshavepreviouslybeenindepencharge,and circumventedthe problem by usingglassdently established.Theseinclude fluid-inclusion fillwalled capsules(seeKretschmar 1973for details).
ing temperaturescorrectedfor pressure,isotopefractionation and silicate phase-equilibria.
Pnocnpunr
Ideal arsenopyrite(FeAsr_1S1ay,
rs 10.13l)Iies
Arsenopyrite-bearingsamplesfrom localities with
on the binary join pyrite (FeS, - loellingite (FeAsr,
estfunates
were rmaalthough minor deviationsfrom the stoichiometric accuratepressure-temperature
ratio Fe,/(As+ S) = y have beenreported (Klemm lyzed quantitatively for Fe, As, S, Ni, Bi, Sb and
1965,Kretschmar& Scott 1976).The composition Co usingan ARL-EMX electronmiiroprobe at the
of ideal arsenopyriteis fixed at a givenpressureand University of Michigan. Operatingconditionswere
as follows; acceleratingvoltagel5 kY; beamcurrent
150 pA; samplecurrent on arsenopyrite0.015 pA.
A grain ofanalyzed arsenopyrite(asp200)from Dr.
S.D. Scott was used as a standard for.Fe, As and
S. SyntheticNiS, CorSo,Sb2S,and Bi metal were
usedas standardsfor the other elements.Each analT (oc)
ysisconsistedof a minimum of ten point-countsper
grain at 20 secondsper point. Up to ten grainswere
analyzedwithin a singlesample.Samplesreanalyzed
600
using arsenicmetal and pyrite as standardsfo1 Fe,
As and S gavecomparableresults.Volatilization and
migration of elementswereevaluatedby successive
measurementsat a single point.
50
Unit-cell refinements of several samples of
arsenopyritewereobtainedto confirm the volumecompositiondata of Morimoto & Clark (1961).Xray measurementswere madewith a powder diffractometer using CuKa radiation, a graphite
34
monochrometer and reagent-grade CaF, checked
36
38
against quartz as an internal standard. A least,
Mole7" As
program after Burnham (l 96l)
squares-refinement
Flc. l. Stability of arsenopyriteat I bar (open circles) and
wasusedto calculatecell parametersof arsenopyrite,
at 5 kbar (solid circles).Limbs of the stability field are
assuminga monoclinic cell.
for compositions
stable with end-members pyrite
(As-poor limb) and loellingite (As-rich limb). Compositions were determinedfrom the d131data of Clark
(1960b) by using the composition equation from
Kretschmar& Scott (1976).
RESULTS
For eachlocality, pressure-temperatureestimates,
buffering assemblages,and microprobe data are
A RE.EXAMINATION OF THE ARSENOPYRITE GEOTHERMOMETER
+u
T)2
700
519
osp+lii+L
Thir rfudy:
O O o l p + g o + L , o r p + 9 y + A sb c l o r
56O'C ond 32 ot.% A!
+ orp+lit+L
orp+L
*
-Q- osp+ 15+9o
tr^I orp+py+L
0 o r g+ 1 6 + A r
OO orp+py+p
A
Y
Clsrh (1960o)r
orp+lii+Ar
o t p + p o + L 1o t p + p y + L
bolor 491'C
(J
c
osp+ po+ L
o
5
Io, soo
CI
E
!,
l-
o s p+ p y + L
"I
lf
/
//
oo6
/l
'iifd
{,
I
I
rsr.r /
osp+lti+po
e't
t
./
/
"/
/
^l
FeAs1.sS9.9
FcAss.15S9.65
3233i343536
Alomic % Arsenicin Arsenopyrite
Ftc.2. T-X diagram with arsenopyrite-bufferedcurves(reprinted from Kretschmar& Scott 1976).
listed in Table l. Microprobe data are reported to in this study is given in the section Analysis of
two significant figuresto avoid rounding errors. A Natural Assemblages.Compositionsof arsenopyrite
more completedescriptionof eachlocality analyzed all lie closeto the line representingthe stoichiometric
THE CANADIAN MINERALOGIST
520
IIALE l.
Saopl€
Or FORl.lATIoll
ARSEIIOPIRITE
CO!.IPOSITIOil,
CoD(ISlItlC MIIIEBALSIr| THE Fe-As-S SYSTEU,AllD P-T CONDITIONS
LosaLtiy
onp. (at.t)
S
As
Fe
conp. (ri.i)
feSAs
Hoeestake, SD t6
a
n7
OrlentaL Ulns, CA
9-13-252
9-'tr-1't2
35.18 20.63 45.t12 1 0 1 . 2 3
3 { . 1 1 1 9 . 7 3 t 1 6 . 6 0 l00.lllr
Ceco, oniarlo
Balnat, l{Y
Araca, Boll,v1a.
ARA-7!'
35.49 2l.116 44.85 1 0 r . 8 0
33.52 2't.'10 \5.58 100.20
ARA-Io?2
ARA-58
Kalahuyo, Eollvla
cHA-23
n
Llaltagua, Bollvlar
Sayaquira, BoLlYla
sAr-la
sAY-27
Tasm, BollYla
rsN-36
35.30 22.21
35.76 22.20
3!.?9
\1.65
\2.91
99.r6
100.87
20.lr5 116.16 1 0 l . r l o
31.80 t8.l?
17.1
33.q
48.56 r00.93
r{8.5
99.0
33.90 r9.76
34.11 20.2
rt6.72 100.38
46.6
10,t.2
gg.g2
PCL-24rr
T
(o c)
(1)
(r)
2-3
50q50
3 3 . 6 t 3 6 . 8 3 29,56
3 3 . 6 0 3 6 . 3 4 30.06
30.9
3 3 . 3 9 3 5 . 1 6 3r.45
32.',t5 35.26 32.59
El,t
tsl
Xn
Eil
ril
po-py
py
'
(1 )
(I )
( 1)
(1)
(1)
0.2-1
ll90!rl0
33.19 33.98 32.83
32.2
32.8
3 1 . 3 1 3 1 . 2 0 35.63
33.6
33.0
36. 1r
35.8
Ell
Xn
xR
EM
B4
xn
po
i
32.87 33.37 33.76
33.0
33.7
33.3
33.5
32.9
32.62 33.82 3 3 . 5 6
Eu
E!4
xn
xn
fc
po-py(?)
'
n
i
(1)
(7)
(1)
(6)
(10)
33.3
33.3
33.1
Eu
xn
EM
po
"
po-py
(7)
(6)
(?)
33.5
33.8
El.'l
xR
po
i
(7)
(6)
&,l
Erl
EH
xR
Er.r
XR
trt
po-py(?)
'
'
<1.0
"i
'
.
(l)
(l)
(1 )
(1 )
(l)
(1)
(l)
t!:5
20.0
100.9
33.2
33.5
3\.2
19.7
\6.7
10 0 . 6
33.1
33.2
34.0
19.5
\5.9
99.q
33.3
33.2
17.58 49.rr3 r00.72
P
( kbar )
po-py(?)
po
,t6.,r
33.71
ref.
EU
Er.r
'r5.56
Panasqueira, PortuSal
PcL-5
34.26 18.q! t18.99 101.69
pcl_40
33.59 17.95 r18.91 100.45
pcL-2't?
3q.0! 18.r0 rr8.95 101.09
otneral.
ss@b.
3 3 . 5 r 3 4 . 2 3 32.25
33.30 33.65 33.30
19.65
33.01
analyt.
Eothod
33.29 31,22 35.q9
33.16 30.86 35.98
3 3 . 3 6 3 0 . 8 9 35.75
35.1
33.32 30.26 36.42
36.q
33.56 3r. o? 35.37
pcL-303
3 r r . 16 1 0 .r 6
oni.
P-23
35.02 22.7\
Clevel.andMlne,
Tas@nla
Moreton'aHrbr, NfId.
3rt.1 19.9
{e.:o
roo.de
\1.75
99.51
115.3
99.3
33.3
33.8
B r o k e nH 1 1 1 , t { S I '
r?.r
31.8
Madelelne Mlne, Caspe, P.Q.
17.6
Conteoyto L, il.!1.T."" 3 3 . 3
'.
1
9.0
33.9
5O.O
98.9
32.2
30.1
t19.5
100.t1
33.0
47.2
100.1
py
"
Lo-po
po
.
Tlmtns,
33.1!
33.2
(1)
(1)
(6)
(1 )
(7 )
(6)
6.010.1 650130
6.5!0.5 625!25
0.19-1.0 q9eq0
280!50
ref.
Q,16)
(2,16)
(3)
(rr)
(5)
(6)
(e)
El.
py
(1)
1.0-1.5 320130
(15)
3 3 . 6 1 1 . 0X B
po
(9)
1.510.5 rl9011o
(9)
B4
32.9
32.5!2.5 *'i
El,t
37.7
po-lo
i
po-1o
(?)
(12)
(7''
1.210.3 350150
(12)
5.5F0.5 75@50
(11)
30.11 36.6
Er4
po-py
(7)
1.210.3 350150
(13)
32.4
El.'t
po-1o
(7)
37.rt6 2 9 , \ 3
3rl.1t
57tt25
(ltl)
CooposltlonE fron pubLtahed d... values uere obtalned uslng the equtlon of KretschEar & Scoft (1976). (B'l-eleciron
dlcroprob€, XR-X-ray, ltc-wet 6ieotcal, ?-posslble equlllbrlun phase). References for TabLe lr (1) Present study; (2) chtnn
( 1 9 6 9 ) a n d p r e s e n ! s l u d y ; ( 3 ) C o v e n e y( 1 9 7 2 , l 9 8 l ) t ( 1 t ) P e t e r e e n ( t 9 8 3 ) ; ( 5 ) B r o w ne l a l . ( 1 9 7 8 ) r ( 6 ) ( e L l y & T u r n e a u r e
( 1 9 7 0 ) ; ( 7 ) K r e t a c h r a r( t 9 7 3 ) ; ( 8 ) K e l l y i R y € ( 1 9 7 9 ) ; ( 9 ) c o u t n e ( 1 9 8 3 ) t ( 1 0 ) M o r t r c t o & c l a r k ( 1 9 6 1 ) r ( 1 t ) P h l l L l p s &
xall (1981): (12) l(ay & Strons (1983); (13) shelton (1983); (14) Bostock (1968); (15) lralsh et, al. (1985); (16) Petersen,
perg. com.
!'
l'
2'
5' o.1t co.
o.3 rt.t co.
z.o, co, 0.3! Ni, o.rt sb. 3' 0.61 co, o.gtr Et, 0.2, Nl.
c o n t a l n ec o , t l l , s b
composition FeAsl_rS1a' except for one s€rmple
wherethe apparentcation-deficiencyis satisfiedby
cobalt (Fig. 3). No other sample analyzedin this
study has appreciableconcentrationsof elements
other than Fe, As or S. Most arsenopyrilegrains
showsmall, apparentlyrandom, microscopicvariations in composition. This same feature has been
observed.byother investigators(Berglund& Ekstrdm
1980,Lowell & Gasparrinil982,Kay & Strong 1983,
J.K. Bohlke,pers.comm. 1984).No obviouscorelo-rim variations were observedin any grains. In
addition, no migration or volatilization of elements
was detectedduring microprobe analysis.
Unit-cell data are listed in Table2 and plotted as
composition yersasvolume (Fig. 4) along with the
data of Morimoto & Clark (1961). Our S-rich
arsenob,'ritedata correspondfairly well with those
of Morimoto & Clark. For more As-rich compositions, our data indicatea sigaificantly smallervolume
(- l9o) than the datum of Morimoto & Clark for
a syntheticarsenopyriteal 0.5@ X(FeAs) on the
FeS2-FeAs2binary join. They noted significant
differencesin the powder patternsbetweenthe synthetic and natural samplesand suggestedthat the synthetic material could probably be indexedusing a
pseudo-orthorhombic
cell (Morimoto & Clark 1961).
This higher symmetry is probably due to enhanced
disorder and twinning associatedwih rapid synthesis and may affect both unit-cell dimensionsand
d,r, values. Ignoring this datum, two best-fit curves
,
A RE-EXAMINATION
521
OF THE ARSENOPYRITE CEOTHERMOMETER
(po)
Poriugot
" PANASQUEIRA,
OBOLIVIAho)
tl
(po,lo)
o HOMESTAKE,
SD(py,po)
V
oORIENTAL,Co(py,po)
v BALMAT,NY(py)
o
aGECO,
Ont
^TIMMINS,
ont (py)
I FeAsS
&
V
vv
V
Fe
v
As
35)30
Fe/(As+$=Y'1/
Ftc. 3. Arsenopyritecompositionsin at.t7oin the ternary systemFe-As-S. Coexistingphasesshown in parentheses.
have been drawn through the remaining data, the appearsto be a valid compositional indicator for
first (solid) basedonly on our new values,and the natural samples.
second(dashed)including the data of Morimoto &
Clark (1961)for natural samples.
Comparisonsof arsenopyritecompositions deterEXPERIMENTAL AND
mined by electron microprobe and those determined
TurnvooyNaurc CoNsnnnerroNs
using the d13, equation of Kretschmar & Scott
(1976) are in excellent agreement(Table 2). For
The stability field of arsenopyriteas a function of
arsenopyritecompositions without appreciablesub- /(SJ and temperature (Frg. 5) (Barton 1969) is
stitution by other elements, the value of d 6, bounded by the sulfidation reactions
TIALE 2. CELL PAR.AMERS,ffOLM VOLI'I{ESAM CALCULATMd131 YALUESFORARSMOPIRITE
S@ple l8a11ty
at.t !s tn 6p
(Proh)
(x?aY)
(AnA-74)
bllvla
Boltvra (CM-23)
orl6ntaL (9-15-112)
(PCL-?I7)
Pesquelra
Pan6quelra (P&-244)
32.5(3)
33.6(tl)
29.9(3)
36,0(5)
36.5(4)
Coll
In
odges
and al-valu€s
32.2
33.5
30.9
35,7
36.4
Angstro@t
'298
5.748(4)
5.752(3)
5.?1r9(3)
5.761(3'
5.773(\)
g 1n degoest
5.6't5(2)
5.683(l)
5.666(l)
5.699(1)
5.702(1)
V in
on3/mle.
5.?64(3)
5.779(3)
5,717(3)
5.78Ot2'
5.783(3)
112,04(5)
112.20(3)
11r.87(3)
112.46(3)
112,48(4)
26.2\(3'
26,34(2,
26.20(3'
26.4312)
26.rr8(4)
ot3t
1.6308
1.6323
1,6293
1 .63118
1,6356
522
THE CANADIAN MINERALOGIST
llll
I
I
molor
volume
(cm3)
26.4
a
t)
;j/
ar'?
'l
26.O
r
.50
46
l
.g
tt"o?
Frc. 4. Volume - compositionplot for arsenopyrite.Trianglesare from Morimoto
& Clark (1961);circleswith error bars are from'this study. Errors in the data of
Morimoto & Clark are not shown. Solid best-fit line is for our data exclusively;
dashed line includes data of Morimoto & Clark. The datum for synthetic
arsenopyriteat X FeAs2= 0.56 (square)is ignored for the fitting of the line.
/zS2+ FeAs2+ FeS: 2FeAsS
S, + loellingite + pyrrhotite = arsenopyrite
(l)
'/zS2+ FeAsS: FeSz+ As(liq)
52 +.arsenopyrite- pyrite + arsenicliquid
(2)
(3)
1/zS,+ FeAsS: FeS+ (As,SXliq)
52+ arsenopyrits: pyrrhotite * arsenic-sulfur liquid
%s, + FeAsS: FeSz+ As(s)
52 + arsenopyrilg= pyrite + arsenicmetal
(4)
Within the stability field of arsenopyrite,two additional sulfidation reactions
t/zS2+FeS:FeSz
(5)
Sr + pyrrhotits : p,"rite
and
(6)
7uS2+FeAs2= FeAsS+ As(s)
52+ loellingite = arsenopyrite+ arsenicmetal
serveto buffer /(S) at a given temperature,uniquely defining the composition of a given
arsenopyrite in equilibrium with the buffering
phases. Whereas other reactions involving Asr,
H2S, etc,, might also be considered,in no way do
they invalidatethesereactions.We have continued
to consider reactions with S, following Barton
(1969,ln0i and Kretschmar& Scott(1976)for direct
comparison with their figures. The shift in these
curveswith pressurecan be estimatedusingexisiing
thermodynamic data. For the sulfidation curves,the
expression
AGPr-AGor: J,"o- AV.dP + RTtn ff-,
*
(7)
can be simplified to
J,tr- av.dP = RTlnB!t-
(8)
since at equilibrium AGPr: 0 and /(S)1 6- is
chosenso that AGor - 0, and the activities of the
solid phasesare independentof pressure.Unit-cell
data for solid solutionsof arsenopyrite@g. 4) and
pyrrhotite (Toulmin & Barton 1964)show a nearly
linear dependenceof their partial molar volumeswith
composition, consistentwith the assumptionof constant activity at all pressures. The AVs for each
reaction was calculatedat P and T using the computer program EQUILI. Thermodynamicdata for
all calculationsard given in Table 3.
The pressureeffect on the melting curvesin Figure
5 could not be calculatedthermodynamicallybecause
of inadequate volume and entropy data for sulfide
liquids. A plot of experimentallydeterminedsulfidemelting curves @ig. 6) shows that the solidus is
extendedto higher temperaturesat higher pressures
by an averageof + l4'C,/kbar. This shift is in excellent agreementwith the experimentally determined
value of Clark (1960b), but later questionedby
Kretschmar& Scott (1976).We have acceptedthe
l4"C/kbar shift in arsenopyrite-pyritemelting curves
as a preliminary estimatefor the purposesof this
A RE-EXAMINATION OF THE ARSENOPYRITE GEOTHERMOMETER
523
tosl(sz)
-8
-t2
400
T("C)
500
600
700
Ftc. 5. Sulfidation curvesin the systemFe-As-s in the stability range of arsenopyrite
(shaded).After Barron (1969).
paper. Using the calculated5 kbar invariant point
for reactions(l) and (6) as a starting point, the stability field of arsenopyritehas been constructedat
elevatedpressure(Frg. 7). All invariant points are
shifted to higher /(SJ and temperature,although
the amount of the shift varies for each reaction.
Arsenopyrite may coexistwith pyrite up to 550oC
at 5 kbar (Fig. 7), significantlyhigherthan the commonly acceptedupper limit of 490.C for the stability of arsenopyrite + pyrite at all pressures.
Kretschmar& Scott superimposedtheir experimentally determinedcompositional data for arsenopyrite
onto Barton's /(S)-T diagram, connectingequal
compositionsbetweenbuffering curveswith straightline isopleths (Fig. 8). A fixed composition of
arsenopyritein the pyrrhotite field (not buffered by
either pyrite or loellingite) definesa univariant curve
in/(St-T spaceat constant pressure.Arsenopyrite
buffered only by pynhotite may lie anpvhere along
the isopleth for that specific composition.
Arsenopyrite isoplethswere extendedby Kretschmar
& Scott (1976)with a shallow slopeinto the pyrite
and loellingite fields. Thesefields representassemblages that include the equilibrium pairs
arsenopyrite-pyrite and arsenopyrite-loellingite,
respectively. If arsenopyrite lies on the true
FeSr-FeAsr binary join, its composition will be
fixed when buffered by either pyrite or loellingite at
constanttemperatureand pressure.The arsenopyrite
compositionalisoplethswould then haveto be vertical in/(Sr-T spacein thesefields, permitting variations in sulfur fugacity, but not temperature,for
a constant composition (Fie. 9). The reaction
FeS, + t/zAst: FeAsS+ %S,
(9)
PYrite+ Asr: alt.nopYrite + S,
illustratesthat log/(Sj will vary inverselywith log
524
THE CANADIAN MINERALOGIST
IIBLE
€ru
bs€n1o
be enopy!'1te
Lo6UiaSllE
Pyfl t€
Pyfrhottto
3. THEMOD$IAI.IIC DATA UAED FOR CALCULATIONS
sou@
12.963
26.r13
26.61
23.943
18.20
(1)
(2',
(3)
(rr)
( rl)
r.zoxroll
1.0EX10_;
1 . 1 0 X 1 0:
tI.9fi10_; -2.5\X10-;
8.rl2xl0-ir
-1.lC!9-; 9 . r 1 X r ^0
4.33X1
o-;
-r.3rxro-l
4.',12]It0: 2,62X10 :
q.58X10
1 .8911o-<
a.szxroli
;:;,,il;:
2.86XlO-:
- 2 . 3 1 X lo a
-i:ili;-?
l.39X1o-o
E
Soaco
3.16
0.99
0.82
0.69
0.12
( 5 , 6)
l7,6')
(8,9)
(tl )
(q)
* wzrootl
a 81 + cT2 { DT3)li codpfosolblltty
ra or3zor..
For oxpandtvtiy,
v!-tw
-',
faotor E-1,000(vo-v).P-'.v
P tn kb. so8oosr (1) nobio et al. (19?8)r (2) ror!@to & Clak
(1961); (3) Redohr (1969)t ({) &oti (1973)i (5) Krlsh!6 et a1. (1979) (6) Blrch (1966)i
(?) Data for coasr. KJskshw & n*ko (19?7), (Iaostruotwal
rlth f€Ass)r (8) r,Jeksh@ & Rakke (1977)
(9) Data for E3 6slte.
Blroh (1966), (IsoslruolFal
etgh FeAar).
v!,
/(As) in equilibrium with arsenopyrite and pyrite.
For a given a(FeAsS)in arsenopyrite, the relationship betweensulfur fugacity and arsenic fugacity at
constanttemperatrueand pressureis fixed within the
stability field of binary arsenopyrite and pyrite. A
similar €ugumentappliesto binary arsenopyrite and
loellingite. The isoplethsin the pyrite and loellingite
fields in Figure 8 were constructed to fit the
experimentalresults of Kretschmar& Scott (1976)
(Fig. 2). The isoplethsin the pyrite and loellingite
fields would only be vertical ifreactions (2), (4) and
(5) coincide, and reactions(l) and (6) coincide in
Figure 2. If arsenopyritelies on the pyrite-loellingite
binary join, the location of thesereactionsmust coincide. Kretschmar& Scott (1976)attributed the sloping isoplethsin the pyrite and loellingite fields to
small deviationsfrom the ideal binary join. Slight
deviations from the stoichiometric ratio
Fel(As + S): y2 would place arsenopyriteoff the
binary join and in the three-phasefield Fe-As-S (Fig.
l0). Two additional phaseswould then be required
to buffer arsenopyrite,allowing the isoplethsin the
pyrite and loellingite fields to shift from a vertical
slope. However, the very shallowly sloping isopleths
in the pyrite- and loellingite-bearing fields (Fig. 8)
suggestan evengreater dependenceon temperature
than in the pyrrhotite field. The minor nonstoichiometry observed in arsenopyrite with respect to
Fel(As + S) requireseither highly nonideal behavior
in arsenopyriteto accountfor the shallowlysloping
isoplethsin the pyrite and loellingitefields, or some
other explanation. In any case, the assemblage
arsenopyrite+ pyrrhotite is stableon both sidesof
reaction (6) in Figure 8. In the absenceof loellingite, arsenopyriteisoplethsshould not changetheir
slope acrossthis boundary. Natural arsenopyritemay provide tests
loellingite - pyrrhotite assemblages
of the alternative topologies for arsenopyrite
isoplethsin the loellingite field, becausedifferent
temperaturesare predicted for this assemblagefrom
each model.
Kretschmar& Scott (1976)determinedthe location of reaction (6) from their two unreversed
experimentaldata-points at 480oand 550'C (Fig. 2).
However, Clark's (1960a)five unreversedarsenopyrite-loellingite-arsenic experimentscoincide with
6
I
,
I
-.'f
'f
oo
p y / tl
P
(kbor)
pv po
i
I
tl
c y= d g * l
2
200
600
T ("C)
tooo
Ftc.6. P-T plot ofsulfide melting curves.Data from Clark (1960b),Clark (1966),
Kullerud (1964),Ryzhenko& Kennedy(1973).Symbols:cv covellite,dg digenite.
A RE-EXAMINATION
525
OF THE ARSENOPYRITE GEOTHERMOMETER
lbori
400
T("C)
500
5 i
600
700
Flc. 7. Stability field of arsenopyriteat I bar and at 5 kbar. Also shown is the maximum thermal stability of coexistingpyrite and arsenopyriteat I bar and at 5 kbar.
reaction (l) in Figure 2, suggestingthat the two
curves(l) and (6) may actually coincide.The separation of the two loellingite-bearing curves and the
three pyrite-bearing curves (Fig. 2) by Kretschmar
& Scott may not be valid, and isopleths may actually be vertical in the pyrite- and loellingite-bearing
fields @ig. 9). Tight compositional reversals are
neededto test whetherreactions(t) and (6) and reactions (2), (4) and (5) coincide, or whether arsenopyrite showssignificant ternary solid-solution, permitting deviation of (l) from (O and (2) from (4)
from (5).
If the isopleths in the pyrite and loellingite fields
arevertical @ig. 9), then their positions will be determined by the intersection of the isoplethsin the pyrrhotite field with the curvesdefining the limits of the
loellingite and pyrite fields. The shift in the
arsenopyrite isopleths in the pyrrhotite field with
pressurewas calculatedfrom the shift in the reaction
FeS+ %S2= FeSz(asP)
Po + Sz: FeSz(asP).
(10)
The partial molar volume of FeS2in arsenopyrite
was determined by extending the unit-cell data in
Figure 4 back to end-member FeS, (Fig. ll).
Whereas the partial molar volume of FeS2(asp)is
still imprecisely known, the available data yield a
higher partial molar volume than the molar volume
of pyrite. Using 24.85 cm3,/molefor the partial
molar volume of FeS2in arsenopyrite,the shift of
in reaction (10) at 5 kbar and constanttemper"f(Sr)
ature was then salculated (Frg. l2). The difference
between the/(St shift in reaction (5) and reaction
(10) is very small (Fig. 12). The cornposition of
arsenopyrite coexisting with pyrrhotite and pyrite is
practically unaffected by pressure,but is shifted by
approximately I log unit/(S) at 5 kbar (Fig. l2).
The compositionof arsenopyritecoexistingwith loel-
526
THE CANANIAN MINERALOGIST
loq J(S2)
€orfc) 5oo
600
700
Ftc. 8. Isoplethsof arsenopyritecompositionsuperimposed
on Barton's (1969)/(StT plot (Fig. 5). After Kretschmar& Scott (1976).
lingite is shifted to lessAs-rich compositionsat higher
pressuresas a result of reaction (O beingshifled more
than reaction (5) with increasingpressure(Fig. 12).
The shift in the isopleths in Figure 12 is constructed
assumingthat they retain the sameslope at all pressures,but the isoplethscould changetheir slopeso
as to exactly offset the shift in the arsenopyritearsenic-loellingitecurve (reagtion6) for all pressures.
This cannot be calculated,but there is no reasonto
expect that the two shifts should exactly coincide.
If the changein the slopeof the isoplethsis minimal,
arsenopyrite equilibrated with loellingite at a high
pressurewill record a low temperaturewhen plotted
on the I bar isopletl graph of Kretschmar& Scott
(Fig. 8). The pressureshift we predict is not very large
for any buffered assemblageand should only be consideredfor high-pressureassemblages.
For the assemblage arsenic + arsenopyrite + loellingite
equilibrated at 5 kilobars, the temperature error
associatedwith usingthe I bar calibration of Kretschmar & Scott(1970 will be on the order of 20oC.This
small shift is probably lessthan the precision of the
arsenopyrite geothermometer and can be ignored.
The shift in sulfur fugacity is not trivial, however,
and we suggest that buffered assemblages
equilibrated at high pressrueshave a higher sulfur
fugacity (-0.1 loe /(S)/kbar, cf. Fie 12) than
predictedby Figure 8.
OsssnvarIoNs ON NATURALASSEMBLAGES
Arsenopyrite was analyzed from sevenlocalities
characterizedby varying physicochemicalconditions
of formation (Table 1). Each locality is describedin
greater detail below.
A RE.EXAMINATION OF THE ARSENOPYRITE GEOTHERMOMETER
527
(K)
rooo/T
{.
loq j(S2)
400 T/on\
500
600
7.'/-l'
Ftc. 9. Isoplethsin Figure 8 redrawn vertically in the pyrite and loellingite fields.
Panasqueira, Portugal
mation. Filling temperaturesat this stagerangefrom
230"to 360oC,withaclusterof dataat280'C. This
The tin-tungsten deposits at Panasqueira are correspondsto a pressureof lessthan I kbar (Kelly
hostedby a sequenceof tightly folded pelitic schists & Rye 1979).Boiling ceasedafter the early oxide
(Kelly & Rye 1979).The mineralizationoccurredin stage.Filling temperaturesdo not decrease,however,
four distinct stages.Arsenopyrite deposition was and Kelly & Rye attributed the lack of boiling in these
associatedonly with the first two stages.In the later inclusionsto a reduction of dissolvedCO, iu
earliest oxide-silicatestageof mineralization, two the fluids rather than to an increasein hydrostitic
eventsof arsenopyritedepositionare recognized.The pressure.Depositional temperaturesand pressures
main sulfide stageproduced arsenopyriteintergrown are taken as 280 + 50"C and I kbar for the
with pynhotite. The third stage of mineralization Panasqueiraarsenopyrite.
involved hypogenealteration ofpyrrhotite to pyrite
and marcasite. The closing stage of mineralization Homestake mine, South Dokota
accounts for only very small amounts of sulfide
precipitation.
The Homestakedepositin the Black Hills, South
Fluid inclusions from the oxide-silicate stage Dakota, consistsof intenselydeformed Precambrian
formed during boiling conditionsand needno pres- garnet-biotite schistsand retrograde chlorite schists
sure correction to reconstructtemperaturesof for- (Slaughter1968).Compositionsof coexistinggarnet
528
THE CANADIAN MINERALOGIST
Frc. l0a. Phaserelationsin the systemFe-As-S between363 and 491'C (after Kretschmar& Scott 1976).The stability
field of arsenopfrite (shaded)is exaggeratedto show the effect of slight deviations from the FeS2-FeAs2binary join.
b. Explodedview of lOa in the region of arsenopyrite.The compositionof arsenopyritein equilibrium with pyrite
at a constantpressureand temperaturecan be anywherealong the bold line betweenA and B, dependingon/(S).
The composition of arsenopyritein equilibrium with pyrite and pyrrhotite must lie.at A. c,d, Phaserelations in
the systemFe-As-S for arsenopyritelying on the binary join FeS2-FeAsr.In this case,arsenopyritecoexistingwith
pyrite must lie at point C,
and biotite (Chinn 1969) yield temperatures of
500t 50oC using the calibration of Ferry & Spear
(1978).Temperaturesbasedon coexistingCO2-and
HrO-rich fluid inclusionsin quartz are 415 I l5oc,
and correspondingpressuresare 2.5 t 0.5 kbar @.
Petersen, pers. comm. 1984). Deposition of
arsenopyrite and pyrite occurred at the end of the
final eventoffolding (Noble 1950,Slaughter1968).
Chinn's (1969) textural evidencefor equilibrium
betweengarnet, quaftz, biotite, cummingtonite and
arsenopyrite sugge$ts that temperatures of
arsenopyrite formation are close to metamorphic
temperatures.Textural evidencesuggeststhe simultaneous crystallization of arsenopyrite, pyrite and
pyrrhotite in some samples.
Oriental mine, California
The host rock of the Oriental mine, California,,
metasediments
consistsof upper-greenschist-facies
with a few small adjacentgranitic intrusive bodies
(Coveney1972, l98l), Arsenopyritewas deposited
contemporaneouslywith pyrite; temperaturesof
fluid-inclusion homogenizationrangefrom 200" to
280'C. Pressureestimatesof 0.7 to 2.5 kbar suggest
temperaturesof formation in excessof 340o, and
possiblyashigh as5@oC(Coveney1972,l98l). The
magnitude of the oxygen-isotope fractionation
betweenmica and quartz is consistent with temperatures of 340! 40"C (J.K. Bohlke, pers. comm.
r985).
Bolivia
Mineralization in the tin-tungsten depositsin the
Cordillera Real, Bolivia, is associatedwith a group
of small granitic stocks (Kelly & Turneaure 1970).
Arsenopyrite formed in the early stagesof minerali-
A RE.EXAMINATION
OF THE ARSENOPYRITE GEOTHERMOMETER
529
V'(RAs2)
molor
volume
(cnF)
lo
?6
Vt s.)
24
pv
.2
.4
.6
.8
to
FeAs,
Ftc. 1l, Extension of unit-cell data on arsenopyrite@ig, 4) back to end-memberFeSr
lld FeAs2;dashedand solid lines havethe sameslopeas in Figure 4. The pointi
labeledpy and lo representthe molar volumesof pyrite and loellingite, respectively. Partial molar volume of Fes2(asp)usedin ialculations is thJ averageof
the intersectionsof the dashedand the solid line with the FeS, axis.
zation, and is associatedwith pyrrhotite in all samples and loellingite in sample ARA-107. Fluidinclusion filling temperaturescorrectedfor pressures
of 1.0 kbar correspondto 450-530"C for the early
vein-stageeventof mineralization (Kelly & Turneaure
1970).
Geco, Ontario
auriferouSpyrite disseminatedin quartz-carbonate
veins, with quartz-pyrite-arsenopyritedeposition
occurring in the early stagesof vein mineralization
(Smith et al. 1984),The arsenopyrite-pyritesample
is from the Pamour mine in the easternpart of the
Timmins gold region. Temperatureand pressureestimates for this deposit are bracketed by the
homogenization temperature of the two-phase
(CO2-CHr-H2O fluid inclusionsand by the intersecrion of rhe (co2-cHo) and (co2-cHr-H2o
fluid-inclusion isochores(Walsh 1985).Theseesti
matesare 320t 30"C and l-1.5 kilobars,broadly
consistent with oxygen-isotopetemperatures of
400 J 80'C for the region (Kerrich& Hodder 1982).
The arsenopyritehas no detectableminor elemenrs
and a low As/(As * S) value, as expected for
arsenopyriteequilibrated at a low temperAture.
The Geco massive-sulfidedeposit, part of the
Manitouwadgemining district, northern Ontario, is
located in the westernAbitibi-Wawa belt and has
beenstudiedby Pelersen(1983).Petersen'sestimates
of peak metaurorphicconditions are 650 t 30"C and
6 + I kbar using well-calibratedmetamorphicreactions and other geothermometersand geobarometers.Thesetemperaturesare slightly higherthan the
less preciseestimatesby Jameset al. (1978). The
Manitouwadge massive-sulfide deposits were Balmot, New York
emplacedprior to metamorphism(Friesenet ol. 1982,
Petersen1983).The single sampleof arsenopyrite
Metamorphic temperaturesand pressuresin the
analyzedconsistsof four small grains enclosedin an Balmat-Edwardsmining district, New York, were
unidentified silver-antimony sulfide.
determinedby Brown et al, (1978) using calcitedolomite thermometry(Goldsmith& Newton 1969)
Timmins, Ontario
and sphaleritebarometry(Scott& Barnes1971,Scott
1973),and by Bohlen et al, (1985) using garnetThe gold deposits near Timmins, Ontario, are rutile-ilrtrenite-sillimanite-quartz barometry. Their
hosted in metavolcanic rocks and metagreywackes estimatesshow that the district reached peak temin the southwesternpart of the Abitibi greenstone peraturesand pressuresof.625x.25"C arrd6.5 t 0.5
bett (Smith et al, 1984).Mineralization consistsof kbar.
530
THE CANADIAN MINERALOGIST
400
T EC)
soo
%
5kbor
l bor
FeS2(osp)
-6
tos J(sr)
5 kbor
I bor
-8
-to
5 kbor
l bor
t.5
t.4
1.3
tooo/T(K)
t.?
Frc. 12. Illustration of the shift in arsenopyriteisoplethsat pressure.For simplicity, only the py-po, FeSz(asp)-po
and lo-asp-As curvesare shown(seeFig. 2). The 33%oisoplethsis shifted to higher/(S) when buffered bY PY-Fo,
and shiftedto lower at.goAsin aspwhen buffered by As and lo, Reactions(5) and (10)are shiftedby slightly different amounts at 5 kbar, sincethe partial molar volume of FeS2in arsenopyriteis slightly larger than the molar volume
of pyrite.
Application of the arsenopyritegeothermometerto
them is invalid as an indicator of peak metamorphic
temperatures.Ore depositsmetamorphosedto conArsenopyrite compositionsfrom this study and ditions of the greenschistand lower amphibolite
others (Morimoto & Clark 1961,Kretschmar1973, faciesplot within the proper compositionalfields,
Collins 1983) are plotted on the composition- but severalsamplesof low-temperaturehydrothertemperaturediagram of Kretschmar& Scott (1976) mal arsenopyriteare unusuallyAs-rich. In particular, arsenopyritefrom Panasqueiralies well outside
using temperatureestimatesfrom Table I (Fig. l3).
Uncertainty in the temperature of formation and the stability field of arsenopyritefor the temperacompositionalrangeof eacharsenopyritesampleis ture of its formation. Applytng the phasediagram
illustrated by the size of ,the sample boxes. of Kretschmar& Scott @g. 2) to thesecompositions
yields higher temperaturesthan actual temperatures
Arsenopyrite from all high-gradedepositsplots outsidethe stability field of arsenoplrite. The temper- of formation. Kretschmar (1973) reported that
arsenopyritecoexistingwith pyrite and pyrrhotite
aturespredicted by the arsenopyritecompositionsin
thesesamplesare severalhundred degreeslower tharr from the Madeleinemine contains 36.6 at.9o As,
actualtemperaturesof formation. Thesearsenopyrite contrary to the claim by Kretschmar& Scott (1976)
sampleshave either resetor are retrogradephases. that in all arsenopyrite-pyritepairs, arsenopyrite
APPLICATION OF THE ARSENOPYRITE
GSOTHTRMOIVIETER TO NATURAL ASSEMBLACES
A RE.EXAMINATION OF THE ARSENOPYRITE GEOTHERMOMETER
531
800
T ("C)
400
PI, PO
30
34
3B
oL70 As in osp
Ftc. 13.Assemblage,temperatureof formation, and buffering assemblage
for var!
ous natural samplesof arsenopyritesuperimposedon Figure 2. l) Broken Hill,
New SouthWales,2) Geco,Ontario, 3) Balmat, New york,4) Contwoyto Lake,
Northwest Territories, 5) Bolivia, 6) Homestake, South Dakota, ?) Oriental,
California, 8) Moretons Harbour, Newfoundland,9) Panasqueira,Portugal, lOj
Madeleine,Gasp6,euebec, ll) Clevelandmine, Tasmania, 12)Timmins, Ontario.
The apparently anomalous arsenopyritecompositionsfrom Panasqueiraand
Madeleinemay possiblybe explainedby changein arsenopyriteisoplethsbelow
the hexagonal-monoclinictransition in pynhotite (seetext).
contains less than 33.3 at.rlo As. Criteria for
equilibriuminhydrothermaldeposits
aredifficultto
assess' and apparent equilibrium
mineralassemblages
precipitated
have
metastaPlyactually
blyunderwidelyfluctuatingconditionsofsupersaturation.
Drscussro^r
can be ignored for low-presspre hydrothermal
deposits. Fressur" will affict the composition of
arsenopyrite in metamorphosed deposits, and should
not be ignored. The most dramaiic effect of pressure is to extend the stability field of coexiiting
arsenopyrite and pyrite to higher temperatures by
illi:iifri':?.'r1""?|:nrffi;'"THgJ,ffi
bly incorrect for regional metamorphic assemblages
'
. Tqe effect of pressureon arsenopyritecomposi- of moderateto high gxade.
tion buffered by two other phasesis not large and
Arsenopyritegeothermometryhas beenusedfor
532
THE CANADIAN MINERALOGIST
temperatureestimationwithout additional data from
other conventionalgeothermometers(e.g., Berglund
& Ekstrdm 1980,Lowell & Gasparrini 1982,Kay &
Strong 1983,Sundbladet al. 1984,Kalogeropoulos
1984). Berglund & Ekstrom (1980) reported on
arsenopyritewith X(A0 ranging from 3l to 33 mole
9o coexisting with sphalerite, pyrite and pyrrhotite
at five sulfide ore deposits in northern Sweden.
Sphaleritebarometry (Scott 1983)yields anomalously
high pressures(6-8 kbar), and arsenopyrite thsrmometry give anomalously low temperatures
(360-480'C) for rocks associatedwith middleamphibolite-faciesschists.
Lowell & Gasparrini (1982)applied arsenopyrite
geothermometryto samples in equilibrium with
pyrite from granite-hosted sulfide veins in
southeasternMissouri. The arsenopyritecompositions X(As) clusterat 31.U329 and32.5-32.9at.Vo
for two separateveins, correspondingto temperatures of 430 t 60o and 470 !20oC, respectively.
These estimates are reasonable but have not been
checked against other geothermometers.
The compositionsof arsenopyritein hydrothermal
veinsin the Moretons Harbour Area, Newfoundland
(Kay & Strong 1983),are widely scattered[X(As) =
29-37 at.6/ol. Temperature estimates from fluidinclusion data are 350t 50'C. The broad rangeof
arsenopyritecompositionsis inconsistentwith the
geothermometer of Kretschmar & Ssott since
arsenopyrite with X(As) > 34 at,alois not stable at
350"C.
Sundblad et al. (1984) applied arsenopyrite thermometry to several ore deposits in the Swedish
Caledonidesfor samplesbuffered by pyrite and pyrrhotite. After excluding arsenopyrite with greater
than 0.2 wt.Vo Co, Ni or Sb, they obtained somewhat scatteredtemperaturesof 370a 80oC [X(As)
29.5-33.3 at.9ol for chlorile-zone deposits and
430 t 50'C [X(As) 30.8-33.0] for biotite-zone
deposits.Whereasthesetemperaturesare in good
agreement with thermometer$ applied to other
chlorite- and biotite-zonerocks (e.9., Essene1982)'
other well-calibrated thermometers (e.g., calcitedolomite) should also be applied to these Swedish
deposits.
Kalogeropoulos(1984)extrapolatedthe arsenopyrite field to 250oCfor application of arsenopyrite
thermometry to a Pb-Zn massive-sulfide deposit.
However, a straight-line extrapolation of the reactions involving arsenopyrite and pyrite in Figure 2
is not valid (Kretschmar& Scott 1976).Arsenopyrite
compositionaldata suggestinglow temperaturesof
formation are inconclusivewitlout independelrtthermometric evidence.Sphaleriteisoplethschangeradically below the hexagonal-monoclinic transition in
pyrrhotite (Scott & Kissin 1973).The pynhotite transition may cause the arsenopyrite reaction-curves
(Fig. 2) to shift to more As-rich compositions at low
temperatures, thereby fitting the arsenopyrite data
from the anomaloushydrothermal deposits@ig. 13).
Arsenopyrite geothermometryappearsto be valid
for ore deposilsmetamorphosedat low or intermediate grades,but the accuracyis lessthan desirable.
The range of compositions analyzedwithin a single
samplemay be too broad to reasonablyconstrain
temperatures of formation. The compositions
reported by Kay & Strong (1983)for arsenopyrite
from Moretons Harbour, Newfoundland rangefrom
29 to 36 at.9o As, spanningthe entire known compositional rangeof arsenopyrite.Arsenopyriteapparently buffered by pyrite and pyrrhotite at the
Homestakemine rangesfrom 32.5 to 34.0 atVoAs,
equivalentto a temperaturerangeof 465oto 535oC.
The composition of arsenopyritecoexisting wilh
pyrite and pyrrhotite from the Oriental mine,
California, only limits temperaturesto less than
400'C. The arsenopyrite+ pyrrhotite + loellingite
assemblagefrom Bolivia limits temperaturesto
between490' and 610oC. Coexistingarsenopyrite
and pyrrhotite restrict temperatureto a lesserdegree.
The composition of arsenopyrite from Bolivia
buffered by pyrrhotite constrainstemperaturesto less
than 525oC.The coexistenceof arsenopyritefrom
Panasqueirawith pyrrhotite indicatestemperatures
between410' and 635oC,whereasactual temperaby fluid-inclusion data
tures of formation as assessed
are less than 350oC. The arsenopyrite + pyrite
assemblagefrom Timmins limits temperatureto less
than 330oC.
Application of the arsenopyrite geothermometer
to samples formed under conditions other than
greenschistand amphibolite facies is not recommended.Arsenopyrite from high-grademetamorphic
rocks appearsto havereset,and arsenopyritefrom
hydrothermal deposits may not have attained
equilibrium. Wide compositional ranges of
arsenopyritehave beeninterpreted as effects of temperature variation or fluctuation in sulfur fugacity
within the hydrothermalsystem(Lowell & Gasparrini 1982, Kalogeropoulos 1984). Berglund &
Ekstrdm (1980)proposedthat local/(SJ variations
exist in medium-grademetamorphic rocks, basedon
varying compositionsof arsenopyriteon a fine scale.
We suggestthat the arsenopyritegeothermometerbe
usedwith discretion, preferably in conjunction with
other, better-calibratedthermometers(e.g., Essene
1982)Tightly reversedexperimentsare sti[ needed
before the arsenopyrite geothermometer can be
regardedas fully quantified. Unfortunately, there are
few calibrated geothermometers other than the
arsenopyritegeothermometerthat can be applied to
massive-sulfidedeposits.An understandingof the
low-temperature behavior of arsenopyrite and the
effect'of minor-elementsubstitutionwould be valu' able. Thesetopics, combinedwith new high-pressure
experiments involving compositional reversals,
A RE.EXAMINATION
OF THE ARSENOPYRITE GEOTHERMOMETER
should be the focus of further investigationsin the
systemFe-As-S.
tionsandapplications.
Econ.Geol,55,1345-1381,
t63r-1652.
(1960b):The Fe-As-S system.Variationsof
arsenopyritecompositionsas a function of T and
P. CornegieInst, LTash.YeqrBook 59,127-130.
AcKNowLEDcEMENTs
The authorsare grateful to Dr. StephenD. Scott,
of the University of Toronto, for his helpful commentsand for providing an arsenopyritemicroprobe
standard.Mr. John K. Bohlke went beyondthe call
of duty with his adviceand information. Dr. Erich
U. Petersen'sfluid-inclusion data for samplesfrom
Homestakeweregreatly appreciated.Dr. Donald R.
Peacor'sinterest and adviceon mineralogicalconsiderations were beneficial. We also thank Mr.
LawrenceM. Anovitz and Dr. Wilbur C. Bigelow
for their help with the electronmicroprobe.Thanks
are also due to Mr. JonathanA. Gorman, Mr. James
F. Walsh and Dr. PeteJ. Dunn for arsenopyritesarnples from various localities. Finally, we acknowledge
the helpful and informative reviewsof Dr. Ulrich
Kretschmar, Dr. Robert Martin, and an anonymous
referee. Funding from WCK's NSF grant no.
EAR-80-26263 helped in the compilation of this
project.
RSFERENCES
Banrolq,P.B., Jn.(1969):Thermochemical
studyof the
systemFe-As-5. Geochim.Cosmochim.Acta 33,
841-857.
533
Clanr, S.P., Jn. (1966): High-pressurephase
equilibria.,tz Handbookof PhysicalConstants(S.P.
ClarkJr., ed.).Geol.Soc.Amer.Mem.97,345-370.
Cou-rNs,
P.L.F. (1983):GeologyandMineralbationot
the ClevelandMine, WestemTasmania.Ph.D. dissertation,Univ. Tasmania,Hobart, Australia.
Covpnry, R.M. (1972): Gold Mineralizationat the
OrientalMine, Alleghany,Californio. Ph.D. dissertation, Univ. Michigan,Ann Arbor, Mich.
-
(1981):Goldquartzveinsandauriferousgranite
at tlte Orientalmine, AlleghanyDistrict, California.
Econ. Geol. 76, 2176-2199.
EsssNe,E,J. (1982): Geologic thermometryand
barometry. In Characterizationof Metamorphism
Through Mineral Equilibria (J.M. Ferry, ed.).
Mineral. Soc.Amer., Rev.Mineral. t0, 153-206.
Fenny, J.M. & Sppan,F.S. (1978): Experimental
calibrationof the partitioningof Fe and Mg between
biotite and garnet. Contr. Mineral, Petrology 66,
I 1 3 -1
l 7.
(1970):Sulfidepetrology.Mineral Soc,Amer.
Spec.Pap.3, 187-198.
FnrEsBN,
R.G., Prcncn,G.A. & Wrsrs, R.M. (1982):
Geologyof the Gecobasemetaldeposit./z Precambrian SulphideDeposits(R.W. Hutchinson,C.D.
Bsncr.uND,
S. & Ersrn0rr,r,
T.K. (1980):Arsenopyrite
Spence& J.M. Franklin, eds.).Geol.Assoc.Can.
and sphaleriteasT-P indicatorsin sulfideoresfrom
Spec.Pap.25,343-363.
northernSweden.Mineral. DepositalS, l7S-187.
Gotosurru,J.R. & Nrwror, R.C. (1969):P-T-XrelaBrncs,F. (196Q:Compressibility;
elasticconstants.
.In
tions in the systemCaCO3-MgCO3at high temperHandbookof PhysicalConstants(S.P. Clark Jr.,
aturesand pressures.
A mer. J. Sci.267-A, I 60-I 90.
ed.).Geol. Soc.Amer. Mem. 97,97-173.
Jnvrs,R.S., Gnrcvn,R.A.F. & Paur, L. (1978):The
BoHrrN,S.R., Valr.nv,J.W. & EssrNs,E.J. (1985):
petrologyof cordierite-anthophyllite
gneisses
and
Metamorphism
in theAdfuondacks.
L Pressure
and
associatedmafic and pelitic gneissesat
temperature.J. Petrology26, (in press).
Manitouwadge,Ontario.Amer. J, Sci.27t, 4l -63.
- loellingite
Bosrocr,H.S. (1968):Gold - arsenopyrite
- pyrrhotitedepositsin amphibolite,ItchenLake- KaLocrnopouros,S.I. (1984): Composition of
arsenopyritefrom the OlympiasPb-Zn massivesulContwoytoLake area,District of Mackenzie,Geol.
fide deposit,Chalkidiki Peninsula,N. Greece.-A/eaes
Sam. Con. Pap. 68-18,72-76.
Jahrb. Mineral. Monatsh., 296-300.
BnorvN,P.E., Esseue,E.J. & Kelr-v,W.C. (1928): Kav, A. & SrnoNc,D.F. (1983):Geologicand fluid
Sphaleritegeobarometryin the Balmat-Edwardsdiscontrols on As-Sb-Au mineralizationin the Moretrict, New York. Amer. Mineral. 63, 2SO-257.
tonsHarbourarea,Newfoundlard,.
Econ.Geol.78,
1590-1604.
Bur.NHau,C.W. (1961):Lattice constantrefinemenr.
CornegieInst..Ilash. YearBook 61, 132-135,
KeLLv,W.C. & Ryr, R.O. (1979):Geologic,fluid
CuruN, W. (1969): Structural and Mineralogical
inclusion, and stableisotopestudiesof the tinStudies ot the Homestoke Mine, Leod, South
tungstendepositsof Panasqueira,P ofiugal. Econ,
Dokota.Ph.D. dissertation,Univ. California,BerGeol.74, l72t-1822,
keley,California.
& TunNBaunr,
F.S. (1970):Mineralogy,paraCI,{RK,L.A, (19@a):The Fe-As-Ssystem:phaserelagenesisand geothermometry of the tin and tungsten
534
THE CANADIAN MINERALOGIST
depositsof the easternAndes, Bolivia. Econ. Geol.
65, 609-680.
KsnnrcH, R. & HoooBn, R.W. (1982): Archean lode
gold and basemetal deposits;chemical evidencefor
metal separation into independent hydrothermal systems. Can. Inst. Mining Metall. Spec. Vol. A'
t44-160.
Roarc,R.A., HevttNcwev,B.S. & Frsspn,J.R. (1978):
Thermodynamic properties of minerals and related
substancesat 298.15K and I bar (1S Pascals)pressure and at higher temperatures. U.S. Geol, Sum.
Bull. 1452 (revised edition).
Scorr, S.D. (1973): Experimental calibration of the
sphalerite geobarometer. Econ. Geol. 6E, 466'474.
Krnrsnus, A. & Raxrn, T. (1977): High temperature
studies of marcasite and arsenopyrite type compounds. Acta Chem. Scand. A3l,517-529.
(19?4): Experimental methods in sulfide synthesis. /n Sulfide Mineralogy (P.H. Ribbe, ed.).
Mineral. Soc. Amer., Rev. Mineral. 1' Sl-38.
Kr-Elrv, D.D. (1965): Synthesenund Analysen in den
Dreiecksdiagrammen FeAsS-CoAsS-NiAsS und
FeS2-CoS2-NiS2. Neues Jahrb. Mineral. Abh. 103,
205-255.
(1976): Application of the sphalerite geobarometer to regionally metamorphosed tetruns. Amer,
KnerscHMAR,U. (1973): Phase Relations Involving
Arsenopyrite in the System Fe-As-S and Their
Application. Ph.D. thesis,Univ. Toronto, Toronto,
Ont.
& Scorr, S.D. (1976):Phaserelationsinvolving arsenopyrite in the system Fe-As-S and their
application. Can. Mineral. 14, 364-386,
Knrsrner, R.S., Snmvasal, R. & DsveNanavalaN, S.
(1979): Thermal Expansion of Crystals. Pergamon
Press, London.
Mineral. 61, 661-670.
(1983): Chemical behaviour of sphalerite and
arsenopyrite in hydrothermal and metamorphic
environments. Mineral. Mag. 41, 427-435.
& BanNes, H.L. (1971): Sphalerite geothermometry and geobarometry. Econ. Geol, 66,
653-669.
& KrssrN,S.A. (1973):Sphaleritecomposition
in the Zn-Fe-S system below 300oC. Econ. Geol.
6t, 475-479.
Kurrsruo, G. (19@): Review and evaluation of recent
research on geologically significant sulfide-type systems. Fortschr. Mineral. 41, 221-270.
SnBrroN,K.L. (1983): Composition and origin of oreforming fluids in a carbonate-hosted porphyry copper and skarn deposit: a fluid inclusion and stable
isotope study of Mines Gasp6,Qudbec. Econ. Geol.
78.387-421.
Lowell, G.R. & Gaspannm, C. (1982): Composition
of arsenopyrite from topaz greisen veins in
southeastern Missouri. Mineral. Deposita 17,
229-238.
SreucHrEn,A.L. (1968): The Homestake mine. /r Ore
Deposits of the United States (J.D. Ridge, ed.).
Amer. Inst. Mining Metoll. Petroleum Engin.'
1436-1459.
Monrnaoro, N. & Cum, L.A. (1961): Arsenopyrite
crystal-chemical relations. Amer, Mineral, 46,
1448-t469.
Surrn, T.J., Clorr, P.L. & Krslen, S.E. (1984):
Geochemistry of fluid inclusions from the
gold deposit, Timmins,
Mclntyre-Hollinger
Ontario, Canada. Econ. Geol. 19, 1265-1285.
NoBrB, J.A. (1950): Ore mineralization in the
Homestake gold mine, Lead, South Dakota. Geol.
Soc. Amer. Bull. 61.221-252.
PprenseN, E.U. (1983): Metamorphism ond
Geochemistry of the Geco Massive Suffide Deposit
and its Enclosing Wall-rocks. Ph.D. thesis, Univ.
Michigan, Ann Arbor, Mich.
Pnrr-rrps,G.N. & Warr, V.J. (1981): Evaluation of
prograde regional metamorphic conditions: their
implications for the heat source and water activity
during metamorphism in the Willyama Complex,
Broken Hill, Australia. Bull. Min€raL 104, 801-810.
SuNpsr-AD,K,, Zacnntssoll, E., Srraros,S.-A.' BBncLUND, S. & ArIuoBn, C. (1984): Sphalerite
geobarometry and arsenopyrite geothermometry
applied to metamorphosed sulfide ores in the
Swedish Caledonides. Econ. Geol. 79' 1660-1668.
Touuran, P., III & BanroN, P.B., Jn. (1964):A thermodynamic study of pyrite and pyrrhotite. Geochim.
Cosmochim. Acta 2E, 641-671.
Warsn, J.F. (1985): Physical and Chemical Controls
on Gold Mineralization at the Pamour #1 Mine'
Timmins, Ontario, Canada, M.S. thesis, Univ.
Michigan, Ann Arbor, Mich.
Revrnonr., P. (1969): The Ore Minerals and their Intergrowths. Pergamon Press, London.
RyznBNro, B. & KeNNsov,G.C. (1973): The effect of
pressure on the eutectic in the system Fe-FeS. :4mer.
J. Sci. 273.803-810.
Received December 17, 1984, revised manuscript
acceptedMay 3, 1985,