American Mineralogist, Volume 58, pages 1-10, 1973

Mertieite, A New Palladium Mineral from Goodnews Bay, Alaska'

Gsonce A. Dnssonoucn, U. S. Geological Suraey, Denuer, Colorado 80225J. J. FrNNnv, Colorado School of Mines, Golden, Colorado 80401, nNo

B. F. LeoNeno, U. S. Geological Suruey, Denuer, Colorado 80225

Abstract

Precious-metal placer concentrates provided by John B. Mertie, Jr., and obtained from the

Goodnews Bay Mining Company carry a new palladium mineral, mertieite, that has the ap-proximate composition Pd"(Sb, As)r and contains 0-2.5 weight percent copper. Mertieite in

polished section is brassy yellow, bireflectant, and distinctly anisotropic. At 546 nm, in air'

Rp'ranges from 50.8 to 56.4 percent and Rg'from 53.6 to 59.0 percent; R:55.5 percent =

Rm. HV* - 561-593, mean 578. Precession photographs show that mertieite is pseudohexa-gonal, possibly monoclinic, with pseudohexagonal a - 15.O4, c - 22.41 A. Strongest lines in

X-ray powder photographs are 2.28 us, 2.23 mw-s, 2.17 m-us, 2.Ol-2-02 m. A total of 43 re-

flections are indexable according to the pseudohexagonal cell. Electron microprobe analysis of

eight mertieite grains show Pd 70.8-74.0, Cu ( 0.1-2.5, Sb 15.2-25.5, as 2.8-9.4 wt percent'

sum 97.8-100.8. The name mertieite is in honor of John B. Mertie, Jr., of the U.S. Geological

Survey, who studied the Goodnews Bay, Alaska, platinum placers. The name and the mineral

have been approved by the Commission on New Minerals and Mineral Names of the Inter-

national Mineralogical Association.An inhomogeneous but stoichiometric compound whose bulk composition is Pd"(Sb, As),

has been synthesized. The X-ray powder pattern of this compound is essentially the same as

that of mertieite.Quantitative microprobe analysis of X-ray stibiopalladinite from the locality of its original

description shows that the chemical formula of this mineral should be PduSb, rather than PdaSb.

A stable phase corresponding to Pd"Sb could not be synthesized. Arsenic is not detected in

stibiopalladinite but is present in all grains of mertieite. Precession photographs show that

stibiopalladinite is different from mertieite and is either hexagonal or perhaps orthorhombic.

Powder data for stibiopalladinite can be indexed provisionally according to a cell having a =

12.80, b - 15.04, c = 11.36 A. If stibiopalladinite is truly orthorhombic, its space group is

P22,2,bgt the space group assignment must be regarded as tentative. Optically, stibiopalladinite

resembles mertieite but differs in magnitude and sign of bireflectance. For stibiopalladinite, at

546 nm in air, Rp ranges from 51.8 to 54.9 percent and Rg'from 52.3 to 55.4 percent; R =

53.6 and Rm' - 54.4 percent. HVm = 589-644, mean 607.

fnftoduction

Sweral brassy-colored grains, which range from0.10 to 0.50 mm in maximum dimensions but aremostly less than 0.25 mm, occur in precious-metalplacer concentrates at Goodnews Bay, Alaska. Grainsin the concentrates are chiefly alloys of platinum,iridium, and osmium. The palladium mineral, whenpolished, appears as a brassy yellow phase, exhibitsslight bireflectance, and is distinctly anisotropic withcrossed nicols; in oil its appearance is unchanged. Itis the only major palladium mineral observed in theplacer concentrates. Its recognition is based on quali-tative electron microprobe and X-ray fluorescence

1 Publication authorized by the Director, U.S. GeologicalSurvey.

milliprobe analysis of about 250 alloy grains of theprecious-metal concentrates of which palladium isonly a minor constituent. Palladium constitutes lessthan 0.5 wt percent of the platinum metals of the dis-trict (Mertie,1969, Table 37). For analyses of con-centrates, see Mertie (1.940, p.76-85; 1969, p. 86-8 7 ) .

Two of the palladium mineral grains contain ir-regularly distributed intergrown native gold' Silver is

the only element alloyed with the gold, and it con-

stitutes about 35 wt percent. Minerals other thangold were not observed in contact with the palladiummineral. The precious metal placer concentrates ap-pear to be derived frorn ultramafic bedrock, because

chromite is a common constituent in the concentratesand also occurs intergrown with platinum-group

DESBOROUGH, FINNEY, AND LEONARD

TASLE I. COMPOSITION AND PHYSICAI PROPERTIES* OF MERTIEITE AND STIBIOPAI.L"A,DINITE

CoEpoolt lon (wetght percent)

Grouo I

Four gralns

Rege Me!!n

liertleite

cfoup II

Four graLns (flve areae)

Range uean tLlL I

Stlbiopalladlnite

Seven Brains

It"* "VPd 70.8-74 .0 72 .9

C u < 0 . 1 - 2 . 5 - ! , 2

sb 15 .2 -15 . 3 15 . 3

A s 9 . 0 - 9 , 4 9 . 2

r . 4- 1 . 1

0 . 0 6

o . 2

7 r . 5 - 7 3 . L 7 2 . 3 0 . 8< 0 . 1 - 0 . 2 < 0 . 1 - 0 . 1

23 .4 -25 .5 24 .7 0 .8

2 . 8 - 4 . 2 3 . 3 0 . 5

67 .8r{ . 5

1 .9 { { . 1

31.2{{ .5

97 .6 -99 .9 -98 .6 1 . 0 9 9 . 4 - 1 0 0 . 8 1 0 0 . 4 0 . 6 100 .9Ponula (reo)

lbeEt approxlmtlon for @rtLelte taken as (pd,Cu)ra*(Sb,As)r_*, where r - 0.1-0.2]

lGrtlelte - Group I

(Pdt .oocuo. to )

($o .9zko. go)

Mertielte - Group II

(Pd5 . r3c t .o . 02 ) (sbr .

s l&0. r l )

StlbiopalLadlnl te

(Pdr,. a3cto. zg) sbt. g4

Re f lect ance

gAVe-

length, nnFour gral.n€

470 546Four grains

470 546 470Eight gralns

546 589

&:#/E!.'REr+Rp '

2

4 5 . 7 - 5 4 . 8 5 4 . 6 - 5 8 . 5

50. 3 54 .9

45.0-50 .8 53 .3-54 .9

5 0 . 2 5 5 . 5

5 7 . 7 - 6 L . 4 4 8 . 9 - 5 5 . 0 5 3 . 6 - 5 9 . 0

5 7 . 2

56.2-57 .L 48 .9-52 .2 50 .8-56 ,4

5 8 . 0 5 L . 2 5 5 . 6

Vickers hardness, 50-grm load

s5 . 3 -61 .4

52,7-58 .8

1 1 t

47.0-5L .7 52 .3-55 .4 55 .1-58 .2

4 8 . 4 5 4 . 4 5 7 . 0

45.5-50 ,3 51 .8-54 .9 55 .L-57 .2

4 8 . 6 5 3 . 6 5 6 . 6

Means

Range

Mean

I\{o gralns

570,s93

5 7 8*Rg' = mxiauo ref lectace for obseped or lentat ion of graln ; RE = lnteredlate pr incipal ref lectece; Rp ' = ulninu ref lect&ce

tor observed or lentat ion of grain; n.d. = not detemlned.A Stadard deviation of the saryle.

U RD of rertlelte detemined from section of nul1 bireflectance observed ln one grain only, Rr' of sctblopalladinlte deterntneds ta t la t i ca1 ly .

1\ro gralns

582,584

Flve grains--ungrouped data

Seven gralns

590-644

607

metal alloys as well as with minor laurite (RuSz)in a nugget (Leonard et al.,1969).

Qnantitative Physical Properties of Mertieite

For seven grains of mertieite, the reflectance val-ues measured at 470,546, and 589 nm with a Reich-ert photoelectric microphotometer reveal no constantvalue for any particular wavelength. Thus mertieitecannot be tetragonal or hexagonal. The United Statesprimary germanium standard calibrated by the Na-tional Physical Laboratory, Teddington, England,was used as the standard for comparison. Values forthe germanium standard are 47.0,51.3, and 52.0percent at 470, 546, and 589 nm (nanometers),respectively.

The indentation or Vickers hardness of mertieite,as measured with a Leitz hardness indenter using a5O-gram load (Table 1), is moderately high andcompaxable to that of magnetite.

Quantitative Chemical Analysis andX-ray Powder Data

Quantitative electron microprobe analysis usingthe Applied Research Laboratories instrument of theU.S. Geological Survey was carried out at 20 kVand a specimen current of 1.5 X 10-8 amp and con-stant beam current. The following synthetic materialswere used as standards (numbers express wt percent):(1) Pd wire (99.99 percent); (2) Pd85.sFere.zt (3)Au73.52Pd26.ae; (4) Ptee.orPdrr.ssi (5) Sbroo.o; Ptre.u,ASrs.+zi (7) Agge.zzSb+r.ar$r.aei (8) A&s.zoSb2e.rsSrz.ze i (9) Agea.roSbu.zoSzz sri ( 10) Feso osCurn.ueSsr.s+; (11) Pd6e.65Sn3e.3s. Selection of these stan-dards was determined by similarity in compositionand average atomic number (21 to the unknown.The standard of nearest concentration to the un-known was used for each element determined. Lin-earity of element concentration versus X-ray line

MERTIETTE

intensity for the standards precluded use of othercorrections. Deviations of X-ray line intensity versuselement concentration are included in the "uncer-tainty" of analysis. Wavelength scans were made onthe analyzed grains by means of LiF and ADP crys-tals for detection of major and minor elements. Sim-ultaneous quantitative analyses for each of threeelements were performed with the following analyz-ing crystals for the indicated spectral lines: ( 1) ADPfor PdLa, LiF for CuKa, and ADP for SbLa; (2)ADP for SbL", LiF for AsKa, and ADP for PdLa.Results of analysis of eight mertieite grains are givenin Table 1. The maximum uncertainty value is basedon the reproducibility of analysis at three separatetimes and on the variations in line intensity attrib-uted to instrument drift or instability, as judged bymeasurements on the standards both before and afteranalysis of the Pd mineral. The values listed are theaverage of determinations of five to ten areas on eachof eight grains during two separate periods of anal-ysis. Six of the eight grains are chemically homoge-neous within the limits of analysis. Mineragraphicstudies show that one of the slightly inhomogeneousgrains is composed of several different optical unitswhich correspond to the compositional variationsobserved.

Data for the eight analyzed grains (Table 1) havebeen combined into groups, at the editor's sugges-tion. The grains of group I differ significantly fromthose of group II in Sb and As content, even whenallowance is made for the maximum uncertainty inthe microprobe analysis. However, grains of the twogroups do not differ significantly in reflectance, mi-crohardness, or X-ray powder data.

Results of microprobe analysis indicate mutualsubstitution of antimony and arsenic but a fixed ratioof palladium to antimony plus arsenic (Table 1). Ageneral atomic formula of Pd5(Sb,As)z is indicatedby the chemical data; however, a more satisfactoryexpression is (Pd,Cu)5*,(Sb,As)2-,, where x = 0. 1-0.2. Copper may be nonessential in the formula in-asmuch as five of the eight grains analyzed containless than 0.1 wt percent Cu (Table 1).

The formula of mertieite, Pdr(Sb,As)2, is similarto the formulas ascribed to stibiopalladinite, PdaSb,to arsenopalladinite, Pd3As, and to the compound,Pd5As2, synthesized by Raub and Webb (1963).The crystal structure has not been determined, butX-ray powder data show the five strongest lines to be2.20 ( lO) , 2 .33 (8 ) , 2 .365 ( 8 ) , 2 .268 (4 ) , 2 .116 (4 ) .Chemical data indicate that mertieite differs from

stibiopalladinite in that 2.8 to 9.4 wt percent arsenic

is present in all grains of mertieite. The syntheticcompound PdgAs is tetragonal and has the Fe3Pstructure (Schubert et al., 196O). X-ray powder datafor stibiopalladinite have been published (Berry andThompson, 1962; Ramdohr, 1960; Genkin, 1968)'but, to our knowledge, the structure has not beendetermined. The difierences in listed d-values andintensities for the X-ray reflections of Pds(Sb,As)zshown for each mertieite grain sampled (Table 2)are partly the result of spottiness of powder patterns

and partly of actual differences in the smallest d-val-ues. Compositional difterences in the two differentgrains are probably responsible for some of the vari-ation in spacings (Tables I and 2)' Actual differ-ences in the position of the arcs representing highestvalues of 20 are evident by overlaying the powderpatterns. The X-ray data of mertieite are dealt withmore fully in the following comparison of mertieiteand stibiopalladinite.

Comparison of Mertieite and Stibiopalladinite

For direct comparison, stibiopalladinite was ob-tained from the U.S. National Museum (USNM

specimen No. R6483). This specimen is from FarmTweefontein, 16 miles from Potgietersrust, Transvaal,South Africa, and is the same as that used for X-raydiffraction by Berry and Thompson (1962, p. 32)'

In the Tweefontein specimen subhedral grains of

stibiopalladinite less than 0.5 mm in diameter aredisseminated along one side of a mafic silicate rockwhich appears to be hortonolite dunite. Several grainswere removed for X-ray diffraction studies and elec-tron microprobe analysis.

Two separate powder spindles of stibiopalladinitewere prepared from fragments of several differentgrains. The resulting X-ray diffraction data (Table

2) differ somewhat from those reported for stibiopal-ladinite by Ramdohr (1960), Berry and Thompson(1962), and Genkin (1968, Table 17)' The X-rayreflections listed by Ramdohr ( 1960) as the strong-est for stibiopalladinite are in poor agreement withthose given by Berry and Thompson (1962). Possi-bly these discrepancies result from compositional dif-ferences.

The distinctive but rather inconspicuous differ-ences in the powder data of mertieite and stibiopal-ladinite (Table 2) were confirmed by precessionphotographs, for which even minute single crystalswere difficult to obtain. Mertieite is pseudohexagonal,monoclinic(?), with pseudohexagonal a = 15.04,

DESBOROUGH, FINNEY. AND LEONARD

TAILE 2. X-MY POIIDER DATA* FOR HERTIEITE, ITS SYMHETIC ANALOGUB, AND STIBIOPAI.I.ADINI?E

Syochetl.c enalogue**of oert l .el te Uert iei te-Pd5 (Sb,As ) 2

Sttbiopal ladini te, Pd5Sb2

Group I Group I I usNM R5483

Cu/Nl Cu/Nld ( o b s . ) l r d ( o b s . ) A I

CulNid ( o b s . ) l

Splndle #8Cu/Nl

d ( o b s . ) l r

Sptndle #9cu/Ni

d ( o b e . ) A I

Splndle #5Fe/!'tn

d ( o b s , ) l r d ( o b s . ) l

3 . 3 4

3 . 8 8

3 . 5 0

3 . 3 5

3 . 1 3

2 . 9 5

2 . 8 4

2 . 7 8

2 . 5 5

2 . 4 8

2 , 4 0

3 . 3 6

m

3 . 5 3 w

2 . 7 8 6 2 . 7 8 3 2 . 7 7 9 2 . 7 7 0

3 . 1 8 w

2 . 9 6 w

2 .83 \ry

2 . 5 5 w

2 . 4 t M

2 . 4 0 w

2 . 5 9 7

2 . 5 1 6

2 . 4 5 6

2 . 3 4 2

2.592

2 . 5 2 2

2.589

2 . 5 1 8

2 . 4 5 2

2 . 3 9 2

2 . 5 4 6

2 .505

2 . 4 5 6

2 . 4 0 02 . 3 9 5

2 . 3 5 9

2 . 2 7 8

2 . 2 3 2

2 . 1 7 r

2 .135

2 . 0 7 8

2.0r7

1 . 9 1 8

r . 8 6 1

1 . 8 1 6

r . 7 9 8

2.3L5

2 . 2 7 9

2 . 2 3 0

2. r59

2. t32

2 . 2 7 9

2 . 2 3 3

2 . r 7 2

2 . 3 3

2 . 2 7 0

2 . 3 4

m --------------- ------------ 2.126

2 . 1 9 4 v s 2 . 1 9 4

2 . O 7 8

2.Ot9

r . 9 2 3

1 . 8 6 0

1 . 8 2 0

2 . 2 7 4

2 . 2 3 !

2 . r 8 0

fi n

2.28r

2 . 7 7 r

Gs

b

G

2 . 0 5 8

2.0 t6

L .920

1 . 8 7 9

1, 825

t . 7 3 2

1 . 6 0 8

I . 5 7 3

1 . 5 4 4

2 . 0 7 6

2.O74

r . 9 2 0

1 . 8 6 1

t .920

1 . 8 6 4

D

M

D

E

m

n

D

D

2 . 0 1 5 w

L . 9 5 4 v

1 . 8 8 0 0

1 . 9 6 6 w

L . 6 t 2 W

w 1 . 8 2 3

) - . 6 7 5 7 , 7 9 1

W

r . 5 1 3 r . 572

1 . 4 9 I

1 . 4 7 7

r . 4 4 9

t . 4 3 4

1 . 3 4 0

D I . ) / J

w L . 4 9 1

v 1-.429

w 1 . 4 1 0

1 . 5 0 7 w

1 . 4 E 0 w

r . 3 6 5

r , 359

r . 2 8 2

7 . 5 7 9 n

f . ) r l w

1 . 4 8 5 w

o

M

1 . 4 8 1

W

M

1 4 3 1

I 4 2 7 w 1 4 0 6

1 . 4 1 3

1 . 3 4 1

W

D

t . 4 1 6

1 . 3 4 3

r . 388

1 . 3 3 5

1 . 3 1 8

t . 2 8 4

L . 2 7 2

L . 2 5 4

t . 2 7 2

1 . 2 5 2

1 . 3 1 7

t . 2 7 2

I .254at

c = 22.41 A. Stibiopalladinite seems to be hexa-gonal or perhaps orthorhombic. Pseudohexagonal aand c of mertieite correspond to hexagonal d and cof stibiopalladinite, but c of stibiopalladinite is somemultiple of c of mertieite. The hol reflections ofstibiopalladinite are streaked where h * 3n; this sug-gests disordering along c of stibiopalladinite. Quali-tatively at least, the precession photographs of mer-

tieite and stibiopalladinite support the findings fromprobe analyses and X-ray powder photography thatthe two minerals are different. The pseudohexagonal,rather than truly hexagonal, symmetry shown by pre-cession photographs of mertieite confirms the deduc-tion from optical study that the symmetry of mer-tieite must be lower than tetragonal or hexagonal.

The interplanar spacings of mertieite, averaged

MERTIEITE

Mert iete-Pd5 (Sb ,As) 2Stlblopal ladtntte, Pd5Sb2

Group I Group I l usNu R6483

sptndle #8 splndle /19 sptndle /15cu/Nt cu/Ni- Fe/Mn - cu/Nl^ cu/Ni^

d( ; ; : : t ; . r d ( ; ; ; . t i r d ( ; ; ; . J i r d (obs . ) i r d (obs . ) i r d (obs . ) i r

r . 222

r . 196

1 . 1 8 1

t . 2 2 5

1 . 2 0 0 1 . 2 4 7

1 . 2 2 7

1.194

r .182

t . 0 9 7

n L . 2 2 2

n 1 . 1 9 7

w 1 . 2 2 2 a I D

i l 1 . 1 9 7 d 1 n

w 1 . 1 8 0 o ,

n

L , 2 6 5 m L , 2 6 8 f f i

t . 1 8 0

t . 1 7 0

L . ! 4 2

w L 2 4 5 w

7 . 2 2 7

1 . 1 9 4

1 . 1 8 1

1 . 0 9 7

wm

1 . 1 0 3

t .089dt

1 . 0 8 r ?

n

n

m

l 06Io,

t . 0 5 2 0 1

1 . 0 4 0 0 d l

1 . 0 2 8 9 0 1

I . 0 0 0 1 o I

1 , 0 5 5 q I

1 . 0 4 3 3 d

w I . 0 4 9

D

n

Dt . 0 3 9 2 d l

d

n

1.0017a1

0.9920o,

0 .9826oI

1 .004

0.9874

0.9629

0 . 9 4 8 8

0 . 9 3 8 3

0 . 9 1 7 6

0 , 9 1 1 0

0 . 8 9 9 3

0. 8819c ,

0 . 8 7 4 9

0.862L

0 . 9 6 1 8

0 . 9 4 8 3

0 . 9 2 3 5 a 1

0 . 9 0 8 9 a ,

0 . 3 8 2 6 a 1 w

o.9270

0 . 9 0 9 7

0 . 9 5 5 8

0 . 9 2 7 7

0 . 9 0 9 8

0 . 8 9 8 1 a I

0 . 8 9 1 3 d 1

0 . 8 8 2 7 d 1

0 . 8 7 5 3 c ,

0 . 8 5 7 0 c ,

0 . 8 5 3 1 d l

0 , 8 3 9 8 d 1

0 . 8 2 0 8 o 1

0 . 8 1 8 7 a 1

0 . 8 1 5 0 d I

0 . 8 1 2 2 o 1

0 . 7 9 9 7 a . '

O , 7 9 4 2 d 7

0 . 7 8 7 8 o ,

0 . 7 8 0 4 c a

0 . 8 7 5 6 0 1 n

0 . 8 6 6 8 0 1 n

0 . 8 5 2 8 c , o

0 . 8 2 1 4 0 1

0 . 8 1 4 6 o 1

0 . 8 1 2 3 a 1

0 . 7 9 9 4 a 1

0 . 7 9 4 2 d 1

0 . 7 4 7 9 d 1

0 . 7 8 0 5 a I

0 . 8 2 9 5 o ,

0- 8185o1

0 . 8 2 8 6 a 1 w

M

n

@

M

E

G

0 . 8 0 7 7 d 1 M 0 . 8 0 7 8 c 1 m

114,6-@dianeEer pdaler c@ra. Fi lb shr lnkage correct ims nade where necessary. Intenslt les ate viaual est i@!e6. w - weak' D - EedluD'

s = sErong, v - very, ----- - not obseneal, Cu/Ni = nickel- f l l teted copper radiat lotr , Fe/Mn = mganese-f i l teted i ron radiet ion

k g " r P d 7 0 . 3 _ 7 3 . g , s b t g . g _ Z S . g , e A . O _ 2 . 0 , l n h o @ g e n e o u s b u t o n e p r o d u c t

from the powder data reported in Table 2, can beindexed according to the pseudohexagonal cell havinga = 15.04, c = 22.41. A (Table 3).

Reflectiuity Data

Nineteen fragments from seven grains of stibiopal-ladinite were mounted and polished. Determination

of Vickers hardness of stibiopalladinite with a 50-grarn load for seven grains gave an average value of607 with a range of 589-644, which is similar to

but slightly higher than that of mertieite (Table 1)'

Qualitative optical properties of the Tweefonteinstibiopalladinite are also similar to those of mertieite,according to examination under a comparison micro-scope.

DESBOROUGH, FINNEY, AND LEONARD

o{ 55

oi

450 500 550 600)r , nm

Frc. 1. Dispersion of the reflectance of mertieite (cor-responding to analyses of group I) and of stibiopalladinite.Tho least value of Rp' and the greatest value of Rg, areassumed to approximate Rp and Rg. For mertieite, Rmwas determined on sections of null bireflectance. For stibio-palladinite, Rm'was determined statistically by the methodof minimum overlap,

Stibiopalladinite is yellowish white in air; in oil,it is very slightly darker and slightly richer yellow.Reflection pleochroism in air is weak, pale yellow toyellowish white. In some orientations, the yellowshows a very faint greenish, pinkish, or lavender tint.In oil, bireflectance is slightly enhanced, but reflec-tion pleochroism is diminished and the pinkish andlavender tints are suppressed. Anisotropism is dis-tinct in air and in oil. Color effects are generallylacking, though some grains are purplish gray togreenish gray in the position of maximum illumina-tion. Internal reflection is absent. Conoscopically,stibiopalladinite shows some dispersion of the ellip-ticity, the effect approximating that of several nativemetals. Stibiopalladinite has three microscopic part-ings, two whose traces are mutually perpendicularand a third oblique thereto. Replacement by a grayalteration product favors one of the cleavage traces.

A faint wedge- or flameJike structure, presumablytwinning, is barely detectable in one grain. Sectoredtwinning was sought but not found. Its presence inRietfontein stibiopalladinite led Ramdohr (1969) tonote that orthorhornbic or lower symmetry might beinferred for the mineral. The ubiquitous twinningshown in single-crystal X-ray photographs of ourstibiopalladinite is evidently a submicroscopic fea-ture, though one of several grains etched with HCIconc. + KCIOB showed a few twin lamellae 1 to 10microns wide.

The reflectance of 8 grains of stibiopalladinite isshown in Table 1, together with the microhardnessof 7 grains. Stibiopalladinite is only a little harderthan mertieite, but the reflectance curves (Fig. 1)are significantly different. The mean reflectance ofthe two minerals is close, but mertieite has strongerbireflectance. For both minerals, the bireflectance issignless at 47O nm, within the limits of measurement.However, at 546 and 589 nm the sign of the bire-flectance is (+) for mertieite but (-) for stibiopal-ladinite. Like mertieite, stibiopalladinite has no con-stant value of R, at each of the measured wavelengths,for one extinction position of each grain and there-fore cannot be tetragonal or hexagonal. The red-lightreflectance of the minerals could not be reliably de-termined, owing to the insensitivity of the apparatusfor measuring very small areas at 650 nm.

Composition

Analysis of 19 polished grain fragments of theTweefontein stibiopalladinite for arsenic, using thesame standards and conditions listed previously, dem-onstrates that, unlike mertieite, stibiopalladinite con-tains no arsenic. However, all the grains contain 1.9:t 0.1 wt percent copper, indicating that copper maybe an essential constituent of stibiopalladinite. TheTweefontein stibiopalladinite contains only Pd, Sb,and perhaps Cu as essential elements. Other elements(Bi, S, Sn, Fe, T, Rh, Ni, Au, Pt, Ru, Os, Ir, Co)were sought but not detected. Quantitative micro-probe analysis of 10 grain fragments for Pd, Sb, andCu gives the following results in weight percent withindicated standard deviations: Pd = 67.8 f 0.5, Sb= 31.2 * 0.5, and Cu : 1.9 :L 0.1. This does notcorrespond to PdsSb where Pd : 72.42 and Sb =27.58, but rather to the formula Pda.65Cus 15Sb2where Pd = 67.09, Sb = 31.66, and Cu = 1..23.Thus the formula for Tweefontein stibiopalladiniteis similar to that of mertieite (two grains of mertieitewere reanalyzed for direct comparison of composi-

MERTIEITE

tion). In this connection, it is noteworthy that in theoriginal description of stibiopalladinite, Adam (1927)

reported values of 70.4 and 70.35 weight percentPd for two determinations on a concentrate fromPotgietersrust.

Comparison with Other SamPles

Genkin ( 1968) investigated stibiopalladinite frornthe Noril'sk deposits and compared this with mate-rial from other localities by X-ray diffraction andreflectivity studies. One X-ray microanalyzer anal-ysis of material frorn chalcopyrite veins indicatedPd?3sb2b (Borovskii et aI., 1959), but no statementis given regarding use of standards, detection limits,or precision of analysis. Genkin ( 1968) reported thatmicrospectrographic analyses of crystalline grains ofthis stibiopalladinite show only Pd and Sb. X-raydiffraction data were presented by Genkin ( 1968,Table 17) for material identified as stibiopalladinite,as well as for two synthetic Pd-Sb phases. He recog-nized essentially two structural types and diffeten-tiated these on the basis of a strong line at 2.28-2.30 A. The Noril'sk samples Pt-I4, N-51, andRamdohr's Potgietersrust, South Africa, sample ex-hibit this line, but the specimens of platinum con-centrate (N-56) and Berry and Thompson's Pot-gietersrust sample do not. The sample of Berry andThompson (1962, p.32) is from the same specimen(USNM R6483) as that from which we obtainedmaterial for the electron microprobe analyses pre-sented above and for the X-ray diffraction data pre-sented in Table 2. We conclude that Genkin's ( 1968,Table 17) two groups of stibiopalladinite, recognizedby him on the basis of X-ray diffraction data. are infact identical to our chemical and X-ray diffractiondistinction between mertieite - Pdb(Sb,As)s - andstibiopalladinite. In addition, it appears that the for-mula for stibiopalladinite corresponds more closelyto PdrSbg than to Pdssb. Evidence of this is previded by our chemical data and by Genkin's diffrac-tion data for the synthetic compound PdoSbz (Gen-kin, 1968, Table 17, 5-4); the spacings of Genkin'ssynthetic Pdbsb2 correspond closely to those of hisstibiopalladinite N-56 and to those of Berry andThompson's Tweefontein stibiopalladinite (1962, p.32).

Similqrities and Difierences

There is no question that the structures of mer-tieite - Pdr(Sb,As): - and stibiopalladinite - Pdb

IABLE 3. OBSERVED AXD CALCI'IATED INTERPINIAR SPACINGS FOR MERTIEI1E(IN ANGSTRO}TS)

d .- b a . {a rc .

h k l

3 . 6 5

3, 34

3 . 0 1

2 , 7 8 r

2 ,590

2,5 t5

2 . 4 5 4

2 . 3 9 3

2.362

2.3L6

2 . 2 7 9

2.232

2 , 7 7 1

2.L34

2 . 0 7 4

2.Or7

1 . 9 2 0

L,452

I , 8 I4

1 . 7 9 8

L . 5 7 4

1 . 4 9 r

r . 4 7 8

t . q ) J

1 ,433r . 413

r ,3421 . 317| . 27 rr . 253L . 2 2 3r , 198I . I8O

I . I 7 0

1 . 1 4 5

1 . 1 0 3

1 . 0 8 9

1 . 0 8 1

1 . 0 6 1

1 . 0 5 4

1 . 0 4 0 7

1 , 0 2 8 9

r .0002

3.360

2.984

2 , 7 7 5

2.635

2.587

2,506

2 . 4 5 3

2,400

2,362

2.327

2.248

2 , 2 3 2

2. t71

2,L32

2.O77

2.016

L.920

1 . 8 6 r

r .816

1.800

r . 4 9 1

! . 4 7 8

r . 4 5 3

t . 4 J )

L . 4 L 4

1,342

1 . 3 1 6

L . 2 7 2

1 , 2 5 3

t .223

L. I97

1 . 1 7 9

1 , 1 7 0

1 . 1 0 3

1.080

1.050

1 . 0 5 5

1.0413

1.0291

1.0018

3 1 . 0

32,O

32.3

5 0 , r

3 3 , 0

5O.3 , 42 .L , 3 j .2

5 0 . 4

5 1 , 1

33.4

) I . J

50,0

4 3 . 1 , 6 0 . 2

5 2 . 1

5 O . 4 , 5 2 . 1

6 r . 3

7 0 . 0 , 5 3 . 0

5 1 . 5

62.L

) ) , 2 , D 4 . r

55 .3 , 64 .2

7 3 . 2

90.2

7 3 . 4 , 8 2 . 2

6 s . 3 , 9 L . 3 , 7 4 , 2

8 3 , 1 , 8 3 . 2 ?

9 2 . 2 , l O . O . 4

5 6 , 0

8 t .2 , 66 ,4

93.2

1 0 . 2 . 0

1 1 . 0 . 3

8 5 , 1 , 1 1 , 0 . 5

1 0 . 3 . 1

72,O.2

9 5 . 0

9 5 . 2 , 1 2 . 0 . 5

95.4 , 86 .5

8 6 . 5

1 3 . 0 . 0

* dob". 1" averaged froD Table 2.

** calculatedofor pseudohexagodal cel l for uhtch a ' 15.04'

c - 2 2 , 4 7 A .

Sb, - have some atomic arrangement in common'This can be deduced from the presence of a 1.56-

1.59 A line of medium intensity as shown inTable 2

and also by Genkin (1968, Table 17), as well as

from the presence of the2.17-2.20 Aand2.25-2.28A lines, all of which are present in mertieite and

stibiopalladinite. The 2.23 A line of mertieite readilydistineuishes the two minerals (Table 2). It is also

i *:obs . , { * *t a l c .

TABLE 4. OBSERVED AND CALCIII.A,TED IMERPLAMR SPACINGS OF

STIBIOPAII.ADINITE (IN ANGSTROMS)

DESBOROUGH, FINNEY. AND LEONARD

h k l

Stibiopalladinite

a : 12.80 A,b -- rs.o4 Lc : 1 1 . 3 6 Au : 2,186.9 L"

Mertieite

4 cos 30o : 13.02 Ab : 15.04 Ac : 22.41 A,u : 4,388.3 A'

No attempt was made to determine a hexagonal callfor stibiopalladinite because the optical data excludehexagonal symmetry and because an orthorhombiccell seemed more likely from comparison of the pow-der data of stibiopalladinite and mertieite. Nor wasan attempt made to vary b for stibiopalladinite inorder to achieve a closer fit between the cells of thetwo minerals. If the provisional primitive orthor-hombic cell is indeed the true cell, the space groupof stibiopalladinite is uniquely determined as P22r2.(Al1 orders of reflections are represented except for0ft0, which has k = 2n.) An unequivo'cal determina-tion of the space group and unit cell cannot be madefrom our single-crystal fragments.

The provisional cell edges of stibiopalladiniteUSNM R6483 differ substantially from aoboco = 2.8,5.2,4.6 A reported for stibiopalladinite by Ramdohr(1969, p . a13) .

Synthesis of Palladium Compounds

High purity metals, certified to contain less than50 parts per million (ppm) of metals as impurities,were utilized in attempts to synthesize compoundscomposed of Pd and Sb; Pd, Sb, and As; Pd, Sb,and Cu; Pd, Sb, Cu, and As; and Pd3Sn2. Each com-pound was weighed to the desired proportion of met-als, sealed in a silica tube evacuated to less than 2Amicrons pressure, melted in a hydrogen-oxygen flame,and either quenched frorn a melt or annealed in afurnace after melting. A major objective of this ef-fort was directed at synthesis of compounds corre-sponding to Pd3Sb, Pde(Sb,As), Pd3(Sb,Cu), or Pdg(Sb,As,Cu), because PdBSb was the formula givenfor stibiopalladinite by Adam (1927). We were un-able to obtain a stable phase corresponding to PdgSbor Pda(Sb,As,Cu) despite many attempts. Previousinvestigators of the Pd-Sb system have not recog-rized a Pdssb phase as one which crystallizes withfixed composition (Hansen and Anderko, 1958; Prattet al., 1968). Furthermore, configuration of thePd-Sb diagram (Hansen and Anderko, 1958, Fig.610) indicates that crystallization of PfuSb from aliquid would require quenching at a temperature ofl20O"C, which is geologically unreasonable.

3 . 88

3 . 5 0

3. 3s1 1 1

t o (

2 . 8 4

2 . 7 8

2 , 5 5

2 . 4 8

2 . 4 0

2 . 3 3

2 .L94

2 . 0 t 5

r . 9 5 4

1 . 8 8 0

1 . 860

1 . 820

r . 7 9 1

L . > t I

1 . 5 0 7

1 . 4 8 0

L . 4 2 7

1 . 385

L . 5 ) t

1 . 3 3 5

t . 315

r . 2 8 2

3 . 8 6 0

3 .526

3 . 3 5 4

3 . r 31

2 . 9 4 5

2 . 8 4 0

) a Q 1

2 . 5 4 8

2 . 4 7 L

2 . 4 0 3

t 2 2 <

2 . 2 7 r

2 . L 9 2

2,0 t3

r . 9 5 3

1. 880

1. 850

r . d l o

1 . 7 9 1

L . J I b

1 . 5 0 7

I . 480

L . 4 2 6

1 . 386

r , 361

r . J J 4

I . 3 1 6

r .282

311

321

o23,3r2

410

420

004

014 ,104

50o,2r4,342

034

520,441

260,3r4

005

450

072,62L

5 3 3

080

180

206,7t0,372

470 , L55 , 27 3

811

0' Lo. o ,822

49O,750,555

057 ,705

208

4 . 1 0 . 0

086

706,690

0 5 8 , 1 0 . 0 . 0

[Remining 15 Llnes not lndexed]

* dob" . taken f ron pa t te rn B , Tab le 2 .

* * Ca lcu la ted fo r o r thorhonb ic (?) ce t l fo r wh icha = 1 2 . 8 0 ; b = 1 5 . 0 4 ; c = l - 1 . 3 6 A .

of interest that the value of the other strong line(2.27 A) for the stibiopalladinite in Table 2 is notin good agreement with the value of 2.25 A givenfor the same specimen by Berry and Thompson(1962, p. 32). The gross similarity but distinctivedifference in the structure of mertieite and stibiopal-ladinite is confirmed by precession photographs.

The powder data for stibiopalladinite USNMR6483 can be indexed (Table 4) on a primitiveorthorhombic cell which compares closely to thepseudohexagonal mertieite cell:

MERTIEITE

The present synthetic studies suggest that if As or

Cu is available. either or both will become incorpo-rated in the Pd-Sb compound. In general, copperand arsenic apparently substitute for antimony, ac-cording to simultaneous analysis of Cu and Sb or Asand Sb on the electron microprobe' However, anal-ysis of one synthetic phase - Pd;g.zSbz7 5Cu1s 5As3 7in wt percent - indicates Cu substitution for Pd,with the resulting formula (Pd+.oCur.o) (Sb1.7As6 3)'

All the single synthetic compounds were either toofinely crystalline or too inhomogeneous for single-crystal X-ray diffraction studies. The two-productsynthetic compounds were very intimately inter-grown, precluding X-ray diffraction study of eachseparately.

One charge with a bulk composition of Pdt".uoSbsg.zzAso 04 wt percent yielded an inhomogeneousproduct with a composition of Pdzo e-za aSbrs.e-rn eAs+ o-o.z wt percent. This material exhibits the majorX-ray powder lines shown by natural mertieite(Table 1). Simultaneous microprobe analysis of Pd,Sb, and As on this synthetic product shows substitu-tion between As and Sb; results of more than 40analyses demonstrate that although the product issomewhat inhomogeneous, each point analysis corre-sponds to the formula Pds(Sb,As)2. Furthermore,the range in Pd, Sb, and As content of this inhomo-geneous synthetic product corresponds remarkablywell to that shown for the various grains of naturalmertieite presented in Table 1.

Of four additional synthetic compounds of Pd andSb combined with Cu and/or As, all have a 2.2O-2.23 A strong line, 1.56-1.57 A and 1.27 A lines,and other lines in common. These data seem to cor-roborate the conclusion that there is some basicatomic arrangement common to this group of com-pounds. The arrangement is apparently controlled byappropriate substitution of As for Sb, Cu for Sb, andin one case, Cu for Pd. However, thermal history isobviously a very important aspect, and thus, exceptfor the synthetic product showing the X-ray powderpattern which corresponds to mertieite (Table 1),direct comparisons with natural minerals should beavoided.

Note added in proof: El-Boragy, Bhan, and Schu-bert ( 1970, Iour. Less-C,o'rnmon Metals ?2' 445-458) determined by single-crystal work that thesynthetic compound Pd 71.5 Sb 28.5, or Pd5*62eSbz-n o, at 760"C has what they term the PdsSbz

type of structure, P63cm, a = 7.60a, c : 13.86s A.

The compositionally similar compound Pd 72.0 Sb

28.0, or Pdri*oz:Sbz ozz, at 50O'C gives a different

powder pattern-ope indicating a superlattice whose

symmetry is slightly lower than P63cm. The cell edges

of the second compound are a = 7 '6e, c : 42.2n A.

They consider the structure of the second compound(their "Pd;*Sbz-") to be a variant of the structure

of the first-a part of the same phase bundle.

Though one is inclined to regard the structural dif-

ference as the result of ordering at lower tempera-

ture, the authors state that the difference is due to

the slight change in composition. The recalculated

powder data of the two compounds are not identical

with those of mertieite, of the synthetic analogue of

mertieite, or of stibiopalladinite, though the three

strongest reflections of synthetic Pd 72.O Sb 28.0(d 2.265,2.199, 1.5 '19 A) do match those of our

st ib iopal ladin i te (d 2.272,2.194,1.578 A). We can

only echo the interpretation of El-Boragy et al.: fot

the minerals, as well as for the related synthetic

compounds near PdrSbz, small changes in composi-

tion may be attended by substantial variations or

changes in structure.

Acknowledgments

We are grateful to John B. Mertie, Jr., and the Good-news Bay Mining Company for providing the materialfor study. We thank lohn White, U.S. National Museum,for providing us with stibiopalladinite, and MichaelFleischer, U.S. Geological Survey, for translating part ofGenkin's Noril'sk monograph.

References

Aneu, H. R. (1927) Note on a new palladium mineral fromPotgietersrust platinum field. ./. Chem. Met. Mining Soc'S. Alrica,27,249-250.

BEnnv, L. G., lNo R. M. Tnorvrpsow (1962) X-ray powderdata for ore minerals-The Peacock atlas, Geol' Soc'Amer. Mem.85' 281 pP.

Bonovsxu, I. 8., A. N. Dsrv, eNo I' D. Mlncuuxovl(1959) Primenenie metoda lokal'nogo rentgenospektral'-nogo analiza k issledovaniyu mineralov platinovoi gruppy

lApplication of the method of local X-ray spectrum

analysis to investigation of the platinum group mineralsl'Geol. Rud. Mestorozh. 6, 68-73. [In Russian]'

GENrrN. A. D. (1968) Mineraly platinovykh metallov i

ikh assotsiatsii v medno-nikelevykh rudakh Noril'skogomestorozhdeniya [Minerals of the platinum metals and

their associations in copper-nickel ores of the Noril'sk

depositl. Akad. Nauk,SSSR, lnst' Geol. Rud' Mestorozh',

Petrogr. Mineral. Geokhim., 106 pp' [In Russian]'

10 DESBOROUGH, FINNEY. AND LEONARD

HANSEN, Mrx, lNo K. P. ANonmo (1958) Constitution olBinary Alloys, 2d Ed. McGraw-Hill. New york. 1305pp.

LroNeno, B. F., G. A. DnssonoucH, AND N. J. pece(1969) Ore microscopy and chemical composition ofsome laurites. Amer. Mineral. 54, lj3}-1346.

Mnnrrr, J. B., Jn. (1940) The Goodnews platinum deposits,Alaska. U.S. Geol. Suru. Bull. 91t,97 pp.

(1969) Economic geology of the platinum metals.U.S. Geol. Sura. Prof . Pap. 630, 120 pp.

Pne,rr, J. N., K. M. Mvrns, J. B. Dennv, Jn., ewo M. H.Murrrnn (1968) X-ray studies of palladium-cadmiumand palladium-antimony alloy. t. Less-Common Metals14. 427433.

Rervroonn, Peul (1960) Die Erzmineralien und ihre Ver-wachsungen, 3d Ed. Akad,.-Verlag, Berlin. 1089 pp.

(1969) The ore minerals and their intergrowths, 3d.Ed. Pergamon Press, Oxford. 1174 pp. [English transla-tionl

Rlua, C. J., eNo G. W. Wrss (1963) An investigation ofthe phase-diagram palladium-arsenic in connection withsuperconductivity. l. Less-Common Metals S, 27l-277.

Scnurrnr, K., S. Bn.lN, W. BunxrHnor, R. Gonre, H. G.MrrssNen, M. PcirzscmB, enp E. Sror_z (1960) Einigestrukturelle Ergebnisse an metallischen Phasen (5).Naturwissenschalten 47, 303. lin German]

Manuscript receiued December 13, 1968; accepted forpublication, September 6, 1972.