THE AMERICAN
MINERALOGIST,
VOL. 50, JULY'AUGUST,
1965
THE CRYSTAL STRUCTURES OF ORDERED
AND DISORDERED COBALTITE
R. F. Groso, Jn.1.LNnP. F. Krnn, DepartmentoJ Geology,
Colwmbia Lrni,tersit"t.New York.
ABSTRACT
Single crystal o-ray studies indicate that the mineral cobaltite (CoAsS) is orthorhombic
with the arsenic and sulfur partially disordered Heating in vacuum to 850' C. completely
disorders the arsenic and sulfur The space group of this form is Pa3 and the structure is
that of pyrite. Annealing in vacuum from 450oC. orders the arsenic and sulphur completely'
The space group is Pca21 (pseudoisometric). Examination of the related mineral gersdorffite
(NiAsS) shows that it belongs to spacegroup P213 and has an ullmanite (NiSbS) structure.
INrnopucrrox
The mineral cobaltite is a sulfarsenideof cobalt. CoAsS. in which
minor amounts of iron and nickel substitute for cobalt, and antimony
substitutes for arsenic (Beutell, 1911). It is found in high temperature
veins or as a constituent of disseminateddeposits in metamorphosed
rocks. Crystals are common and the most frequent forms are the cube
and pyritohedron, and, less frequentlv, the octahedron (Palache et al.
1941).
On the basis of crystal morphology, cobaltite has generaily been assumed to have the symmetry of the pyrite group. This symmetry is isometric-diploidal-2/m 3 (seefor example,Dana, 1892).With the determination of the pvrite structure (Bragg, 1914)it becameclearthat cobaltite must have a lower symmetr;r, since the substitution of an As-S unit
for the S-Sunit of pyrite would destroy the symmetry centersof the S-S
units. This lower symmetry was thought to be tetartoidal -23, and
minerals similar to cobaltite (ullmanite and gersdorffite) were placed in
t h e " c o b a l t i t eg r o u p " ( P a l a c h ee t a l . 1 9 4 4 ) .
The first ir-ray study of cobaltite was done by Mechling (1921). From
the symmetry of Laue photographs,he proposed a structure based on
(1922) examined a
the spacegroup P213. Shortly after, Schneiderh<thn
and found it
microscope
single specimenof cobaltite with the reflecting
had unthat
cobaltite
to be anisotropicin polarizedlight. He suggested
orthorhombic
dergone an inversion after formation and was actually
(the symmetry by analogv with pyrite and marcasite). This optical
anisotropy has been confi.rmed by several others (Sampson, 1923;
Fliirke, 1923,1926;Ramdohr, 1950;Onorato,1957).When heated above
850" C. and quenched,cobaltite was reported to be optically isotropic
(Fl<;rke,1926).
1 Present Address: Research and Development Division, The Carborundum Company,
Niagara Falls, New York.
1002
ORDEREDAND DISORDEREDCOBALTITE
1OO3
Ransdell consideredthe questionin 1925and confirmed the apparent
isometriccharacterof powder photographs.He suggestedthat the optical
anisotropy might be either (1) the result of deformation twinning caused
by grinding and polishing the specimen,or (2) evidence that cobaltite is
pseudoisometric.Peacock and Henry (1947) on the basis of intensity
measurementsof powder photographs decided that cobaltite had a disordered structure in which the arsenic and sulfur were substituting for
each other. This is identical with the structure of pyrite. Onorato (1957)
using intensity data obtained from Weissenbergphotographs,proposed
the symmetry P21/c. The agreementbetween observedand calculated
intensities was poor. It has recently come to the authors' attention,
through the kindnessof Dr. Winterberger, that an r-ray study of cobaltite
was done by Le Damany in t962. The study concernedonly natural
cobaltite and indicated a partial disorder of arsenic and sulfur.
PnBtrlrrwanY INVEsrrGATroN
Several of the previous investigations (Peacock and Henry, 1947;
Onorato, 1957) used copper radiation. This resulted in a very high
fluorescent background from the cobalt. As an initial check on the accuracy of the space group determination, several samples of cobaltite
were examined by powder diffraction techniques with iron radiation and
a very small sample. The results (Table 1) clearly showed several reflections forbidden by the spacegroup assignedby Peacockand Henry. A
small sampleand long exposurewere necessaryto show theseweak lines.
AII such lines were indexable using the cobaltite unit cell, indicating that
they were probably not caused by impurities. To resolve the apparent
contradictionsin cobaltite, it was decidedto do a structure determination using single cr1'stal techniques.
Spacp Gnoup DBrBnurNarroN
The cobaltite samplesused in the single crystal studies were from Cobalt, Ontario (Columbia Collection No. 200-83)and were in the form of
small cubes (1-2 mm edge),some being modified by octahedra.Using a
small cleavagefragment, the precessioncamera was used to take orthogonal zero layer and first upper layer photographs with filtered molybdenum radiation. The photographs indicated a primitive cell and
orthogonal glide planes as in spacegroup Po3. However, several weak
odd order prism reflections were present which should have been extinguished by the glide planes. It was later learned that these "forbidden
reflections"werealso observedby Tak6uchi (1956).Thesereflectionssuggested the possibility of some sort of disorder. One natural form of disorder would be a partial substitution of sulfur and arsenic. Complete
R. F. GIESD, IR. AND P. F. KERR
1004
T.tsln 1. Var-t'ns lon SolrB Colalrrre
SprcrlrnNs
hkl
1001
1101
111
200
210
2rl
220
300,221
3101
311
222
320
32r
400
4t0,322
411,330
331
420
421
332
422
500,430
510,431
511,333
520,432
52t
440
d obs.
d obs.
d calc.
d obs.
5.56
3.94
3.30
2.77
2.49
2.27
1.973
1.856
1.762
1.680
1.608
1.545
1.490
1.393
1.350
1.310
1.277
r.246
r.216
1.189
1.139
1.115
1.093
1.073
1.033
1.018
.987
5.58
2
3.95
1
2
3.22
2.79
8
2.50
10
2.28
9
1.97+ 6
1.861 1
5.58
3.94
2
1
J.JT
Z
I. /OJ
3.28
2
283
6
2.54
10
8
2.29
1.988 z
1 . 6 8 2 10
1 610
2
4
1 5+7
6
r 490
1.681 10
1608
+
1.548 2
r491
8
1.690
I
;
1.683 10
1.611 1
1.548 4
1.492 I
1.400
1.3s4
1.316
r.28r
+
+
+
+
|.248
l2t8
1.190
1.139
3
5
3
3
|.116 +
1.09s +
r.074 8
1.037
1. 0 1 9
.987
8
2.79
10
2.50
2.28
9
1.972 3
3
3.35
2.78
7
10
2.49
2.28
9
1.970 2
5
3
5
I
1
1.276 t
124e +
111
1 074
1.036
1. 0 1 9
.987
6
6
1
6
1.559
1.496
1.403
r.r41 +
1.253
t.222
1.198
1 .1 4 8
r.076
1.038
|.02r
.988
1080
r.042
1.022
.989
1.2t8
1.189
1.217
1.189
1.139
1
|
6
5
3
7
9
1.60e +
3
1
+
4
2
1
3
1. Cobalt, Ont., 20 hrs., 0.3 mm. glass capillary.
2. Tunaberg, Sr,veden,18 hrs., very small fiber.
3. Cobalt, Ont., 8 hrs., 0.3 mm. glass cap. (USNM 95740).
4. Cobalt, Ont., ?, ?(Peacock and Henry, 1947).
I Reflections forbidden in space group proposed by Peacock and Henry (1947).
disorder would yield the structure suggestedb1' Peacock and Henry
(1947)and have the spacegroup Po3.
Drsonlnnno Srnucrune DBtenutN.q.rroN
Crystals of cobaltite were sealedunder vacuum in fused quartz tubes.
These were piaced in a muffie furnace and held at an elevatedtemperature for one or two days and then quenchedin cold water. After quenching, a small fragment was examinedby precessionmethods.It was hoped
ORDERED AND DISORDERED COBALTITE
1005
that an increasein temperature would permit the arsenicand sulfur to
randomly substitute for eachother to a greaterdegreethan in the natural
material. As the temperatureprior to quenchingincreased,the forbidden
reflectionsbecameweaker.
At the highest temperatures () 700' C.), decomposition became a
problem. After quenching,the crystal surfaceswere found to be coated
with a mixture of decompositionproducts. However, in the center of the
crystal, there appearedto be little alteration. Precessionphotographsindicated the central material was still cobaltite. This method was used by
Florke (1926)in his high temperatureoptical studies.
A sampleheld at 800-850' C. for two days showedonly the systematic
extinctionsof spacegroup Po3. This crystal was used to collect (hko) intensity data. It was a roughiy orthogonal cleavagefragment with the
d i m e n s i o n s. 2 0 X . 1 3 X . 0 4 m m . I t w a s m o u n t e d ( o n t h e p r e c e s s i o n
camera) with the smallestdimensionparallel to the r-ray beam. No correction was made for absorption errors. Diffraction intensitieswere estimated by comparison with a standard scale. The usual Lorentz and
p o i a r i z a t i o nc o r r e c t i o n sw e r e m a d e .
Assuming the structure to be of the pyrite type with arsenicand sulfur substituting for each other, the structure is determinedby one positional parameter, that of the (arsenic*sulfur) atom. Structure factors
were calculated using various values for this parameter and the normal
R-factor was plotted against the parameter. The atomic scattering factors were taken from the Internationalle Tabellen, V. 2. The results are
s h o w n i n F i g . l a n d T a b l e 2 . T h e b e s t f i t i s o b t a i n e dw h e n x : . 3 8 0 . N o
temperaturefactor rvasused in the calculations.
Onnnnno SrnucrunB Dotonltrw.+ttoN
A singlecrystal from the samelocalitv was sealedin a quartz tube and
heatedto 450' C. It was cooledslowly to room temperatureover a period
of ten hours in an attempt to order the arsenicand sulfur. Weissenberg
and precessionphotographs indicated that the space group was either
Pca21or Pbcm. The cell dimensionsremainedessentiallythe same (5.582
+.002A) and the cell was isometric within measurementerrors.A cleavage fragment (.25X.10X.07 mm.) was used to collect(hko) intensity
data with a precessioncameraand (hoi) data were collectedwith a Weissenberg camera. Intensities were estimated as before and reduced to
structure factors in the usual wa1..
Regardingthis as an order-disordermechanism,the spacegroup of the
orderedform must be a subgroupof the disorderedform (Buerger,1947).
This rules out the centrosymmetric space grotp Pbcm.
A trial structure was found in space group Pca2r based upon an order-
1006
R. F. GIESE, TR. AND P. F. KERR
.r6
\
.14
I
/
\
.12
/
R
\
.lo
\
.o8
/
\
/
\
.o6
.o4
.376
.380
.384
x
Frc. 1. Plot of the R factor wrsus the variable parameter, x, in disordered cobaltite.
ing of the arsenic and sulfur atoms. The parameterswere refined using
Ieastsquaresmethods and the scatteringtablesin Volume 3 of the International Tables for X-ray Crystallography. The final R factors for (hko)
Tl.srn
hkr
200
400
600
210
220
230
240
250
260
270
410
X:not
2. Drsonorrro
F Obs.
91.0
-48 4
3 0 .s
9 7. 7
72.r
88.6
4 9. 9
-s4.4
42.7
-62.0
8.8
observed.
85.0
52.5
2 7. 2
95.2
70.0
93.5
s 3. 8
s9.3
39 5
6 1. 2
136
CoAsS-Sp,lcB
Gnoup Pa3
hkl
F Calc.
F Obs
420
430
440
450
460
610
620
630
640
650
49.9
88
134.3
- 5.9
52.4
-56.4
42.7
-60.6
52.4
40.6
5 1. 8
8.9
129.O
X
56.9
52.4
40.5
62.2
52.1
40.8
ORDEREDAND DISORDEREDCOBALTITD
1OO7
and (hol) are .10 and .12. The observedand calculated intensities are
listed in Table 3.
Porrsnen SrcrroN Sruly
To verify the results obtained from the r-ray studies, tr,vosampleswere
prepared for microscopic study. Both came from Sweden (locality unknown) and appearedto be single crystalsl one a cube and the other a
pyritohedron.
Tlrr,n 3. OrsrnvnnaNnCar,cur,amtSrnucrunnFecronsron OnonnooCoAsS
hkl
F Obs.
o20
0.30
040
050
060
070
120
130
140
1.;0
160
170
200
220
230
240
250
260
310
320
330
93.4
- s 6 1.
449
308
49.7
992
31.4
1 5. 6
30.4
6 7. 2
l.t .t
9 9 .1
73.6
12.3
529
218
480
32.1
90.4
376
li Calc
-97 8
307
-55.0
439
-27 3
-46.i
-102.6
27.5
1 10
-26.5
6 3 .I
-11 3
ro0 2
-78r
-8 5
504
180
246
-99 2
hkl
F Obs.
340
350
360
400
410
420
430
440
.r50
19.6
35.2
6q5
52.5
402
. 5 85.
173
143 5
1t 7
599
352
640
88
78
68
31 0
10.3
480
620
460
510
520
-\l0
540
550
600
610
620
640
7"0
002
849
F ('alc
12.5
-31.6
673
-53.5
387
-52 7
-44 5
147 8
-86
-56 1
-27 9
560
-46
-4.6
67
25.4
to.2
-44 3
602
-33.8
hk1
F Obs.
004
006
200
201
202
203
201
205
206
400
401
102
403
404
405
600
601
602
603
504
- s 7I
45.1
99.1
r2r 2
92.9
los 8
52.1
478
38 2
52 5
29 2
91 6
20i
t43 4
169
310
780
59 1
828
783
F Calc.
54.6
59.7
100.2
113.2
78 8
102.3
52 3
650
46.7
.5.35
20.4
75.0
r23
148 3
9.9
2.54
658
466
i77
612
1 , 1 34
The anisotropy of both was weak and difficult to photograph using a
normal lighting arrangement.The photographs(Figs. 2, a-c) were taken
using a very high intensity xenon light source.Figure 2a shows the cube
cobaltite and Fig. 2b the pyritohedral cobaltite. These clearly show the
previously reported anisotropy. Somesmall fragments of the latter were
sealedin an evacuatedsilica tube and heated to 825oC. for one hour and
quenchedin water. Figure 2c shows a small fragment of this material
which is almost completely isotropic. X-ray diffraction confirmed that
the material was still cobaltite.
DrscussroN
Electron density projections of the ordered and disorderedforms of
cobaltite are shown in Fig. 3. The contour interval is on an arbitrary
scale. The final atomic parameters, atomic radii and interatomic dis-
Frc. 2. Polished section photographs of cobaltite (150X). Figs.2a, b-natural
material. Fig. 2c-material quenched from 850'C.
!)
Frc. 3. Electron density projections of disordered (3a) and ordered cobaltite (3b, 3c).
The contours are drawn with an arbitrary interval. The first solid contour is at height zero.
Figure 3a is projected onto (001). Figures 3b and 3c are projected onto (001) and (010) respectively.
1010
Ir. F. GIESIj, IR. AND P. F'. KERR
tancesare listed in Table 4. To more easily comparethe orderedand disordered forms, the parameters of the ordered form have been transformed to coincide with the disorderedcell. It is seen that the shifts,
causedby ordering,are small.
T,rlr,n 4. Arourc PARAMETERS,Boxo LeNcrns nNr Arolrrc Relrr rn Omrnro
axn Drsonornro Cogatrrrn
Parameters
Disordered
Ordered
XYZXYZ
Cobatt
Arsenic
Sulfur
-.010
.380
-.380
.006
.381
.380
.011
.383
-.382
0
.380
-.380
0
.380
-.380
0
.380
-.380
Bond Lengths
Ordered
Cobalt-Arsenic
Cobalt-Arsenic
Cobalt-Arsenic
Cobalt-Sulfur
Cobalt-Sulfur
Cobalt-Sulfur
Arsenic-Sulfur
(Arsenicf Sulfur) -(Arsenicf
Cobalt-(Arsenic* Sulfur)
Disordered
2 . 3 1A
2.39
2.36
2.36
2.26
2 -29
2.30
Sulfur)
z-Jl
2.31
Atomic Radii
Disordered
Cobattite
Iron
Cobalt
Arsenic
Sulfur
(Arsenic and Sulfur)
Ordered
Cobaltite
Arsenopyrite
(Morimoto and
Clark, 1961)
1 . 1 4A
1 . 1 6A
1.16
t.l7 A
1.18
t.t2
1 . 1 5( A v . )
121
1.11
1 . 1 6( A v . )
The disorderedstructure, as mentioned before is the pyrite structure
(Fig. aa). The cobalt atoms (smaller spheres)have a face-centeredarrangement and the eight (arsenic*sulfur) atoms (larger spheres)are in
the eight subcubesof the unit cell. In eachsubcube,the (arsenicf sulfur)
atom lies on a line joining oppositecornersof the subcube(heavy dotted
Iine). Thesebody diagonalsare axesof 3-fold symmetry.This arrangement
ORDERED AND DISORDI|RIID COBALTITD,
1011
Frc. 4. Structure of cobaltite. 4a. Disordered cobaltite. The smaller spheres are cobalt
and the larger spheres are (arsenicfsulfur).4b.
Ordered cobaltite. The smaller spheres are
cobalt and the larger spheres are arsenic (plain) and sulfur (double circles).
placesa cobalt atom at the center of an octahedron whose cornersare
(arsenicf sulfur) atoms. Each (arsenicf sulfur) atom is coordinatedby
three cobalt atoms and one (arsenic*sulfur) atom in a distorted rerrahedral arrangement.
t012
R. F GIESE, ]]1. AND P. F. KERR
The ordered structure (Fig. ab) has the same spatial arrangementas
shown in Fig. 4a. The eight (arsenicf sulfur) atoms becomefour sulfur
and four arsenic atoms. The larger spheresin the lower four subcubes
becomesulfur and the larger spheresin the upper four subcubesbecome
All atoms move slightly away from their positions
arsenic(or a'ice-aersa).
in the disorderedstructure. A cobalt atom in the ordered form is coordinated by three arsenic and three sulfur atoms in an octagonal
arrangement.An arsenicatom is coordinatedb,v three cobalt atoms and
one sulfur atom and similarly a sulfur atorn is coordinatedby three cobalt atoms and one arsenicatom in a distorted tetrahedral arrangement.
The abilitl'of arsenicand sulfur to substitute for each other suggests
the possibility of similar behavior in arsenopvrite(FeAsS) and gersdorffite (NiAsS).
In a restudl' of the cr1'stalstructure of arsenopyrite,Morimoto and
Clark (1961) conclude that the space group is P1 approa"chingP2v/c
as the arseniccontent increases.This is equivalentto saying that replacement of sulfur br. arsenicincreasesthe symmetry. In fact, the authors
concludedthat from the evidenceof electrondensity maps,30 per cent of
It is unfortunate that
the sulfur substituted for arsenic and r.tice-verso.
their quenchingexperimentswere not done at temperaturesabove 600o
C. since this is approximately the lor,verlimit for observablesvmmetrY
changesin cobaltite.
Crystals of gersdorffiteare rather rare and no report could be found in
the literature regarding its space group and structure based on single
crystals. Por,vderphotographs indicated space group Pa3 and a pyrite
structure based on a disordering of arsenic and sulfur (Peacock and
IIenry, 1947). Previous reports indicated space group P2$ (Palache
et al.1944).
Three smali octahedral crystals (origin unknown) were found in the
Columbia Collection.They were identified as gersdorffiteb1' a comparison of the d values with those of Peacockand Henry and an emission
spectrographicanalysiswhich showed: trace-iron, copper' silver, antimony; moderate-gold, bismuth, cobalt; major-nickel, arsenic'Sulfur
was not included in the analysis. Precessionand Weissenbergphotographs of a fragment of one of the crystals clearly-showed spacegroup
P2fi. Yery long exposuresfailed to show an,v reflectionsviolating this
spacegroup. A comparisonof the (hko) intensitieswith those of ullmanite (Tak6uchi, 1956) showed them to be isostructural. Hence, there
seemsto be no evidenceof substitution between arsenic and sulfur in
gersdorffite.The sample available rvasinsufficient to allow any quenching
experiments.
ORDL,RED
AND DISORDERED
COBALTITF:,
1113
Cowcr,usrons
We are now in a position to explain the apparentlv contradictory
p r o p e r t i e so f c o b a tl i t e .
1. The optical anisotropy resultsfrom the non-isometrics1'mmetryof
n a i u r a lc o b a l t i t e .
2. The completetransformation to isometricsymmetrl.at 800-850oC.
confirms Fkirke's observation of the optical isotrop.,'of cobaltite
above 830" C.
3. The isometric appearanceof the powder photographs of cobaltite
results from the near identity of the two structures in terms of
atomic positions and the close similarity in interatomic distances
in both forms.
4. The forbidden r-ra1. reflectionsobservedin natural cobaltite are
those which occur in the ordered form and result from the partial
disorderingof the natural material.
AcrNowreocMENTS
The authors expresstheir appreciation to ProfessorRaiph Holmes,
Columbia University, Dr. William Croft, Sperry Rand ResearchCenter
and Dr. S. R. Kamhi, MassachusettsGeneralHospital for their interest
and helpful suggestions.The research was supported in part by the
Atomic Energy Commission.One of us (RFG) was aided also b1' a grant
from the Department of Geologv, Columbia Universit_v.The initial
computational work was done through the kindness of llr. Kenneth
King at the Watson ResearchLaboratorv, Columbia Universitr'.
Rnlrneltcls
Brurrr.r., A. (1911) Chemisch-Mineralogische Lintersuchungen am Glanzkobalt. Centralb.
Minerol. 663-673.
Bnacc, W. L. (1914) The analysis of crystals by the:r-ray spectrometerProc. Royal Soc.oJ
L o n d o n ,A , 8 9 , 4 6 8 4 8 9 .
Bunncon, M. J. (1947) Derivative crystal structures. Jour. Chem.Physics,15, 1-16.
DeNe, E. S., 1892,A SystemoJ Mineralogy John Wiley and Sons,6 Ed.
Flcinra, W. (1923) Mikrographische Beobachtungen An Ni-Und Co-Erzen. Metall u,nd.
Erz,B.2O, 198-206.
(1926) Zur Polymorphie des CoAsS. Centralb Mineratr. Abt A , 337-8.
LDeunNv,
I 0962) Contribution a 1'etude cristallographique de la copaltine Mem
Preseq.t€,Fac. Sci.. Unit. Payis.
Mrcnr.rxc, M. (1921) Abh Sachs. Ges.,8.38, in Strukturbericht, 1913-1926: Zeit Krist ,
1931, 283-284.
Monrrtoro, N. eNo L. A. Cr.mr (1961) Arsenopyrite crystal-chemical relations. -4r2.
Mineral.46, lM8-1469.
Oxonero, E. (1957) Riintgenographische Untersuchungen iiber den Kobaltglanz: Neues
Jahrb Mineral. Abh.,B.91,41-53.
1014
R. F. GIESE. TR, AND
P. F. KERR
Per-acnn, C., H. Benuax nno C. FnoNotr- (1944) Dana's System of Mineral'ogy,I,7 F,d.,
John Wiley and Sons, Inc., 834 p.
Pracoct<, M. A. aNo W. G. HnNnv (19a7) The crystal structuresof cobaltite,gersdorffite
and ullmanite. U nit. T or onto S tuil i.es,Geol. S er. 52, 7 l-80.
Rnrr,tnorrn, P. (1950), Die Erzmineralien und lhre Verwachsungen. Academie Verlag
(Berlin).
Rausnor.r, L. S. (1925) The crystal structure of some metallic sulphides. Am. Iulinerol. lO,
281104.
SaunsoN, E. (1923) Econ. Geol.18, 604t-611.Review of Schneiderhdhn'sBook.
ScnNnroetrrrirrn, H. (1922) Anlei.tung rul Mi.crographischen Bestimmung und Unterstuhxtngen ron Erzen im Aufall,enden Licht. Gesellschaft Deutscher Metallhiitter und
Bergleute (Berlin).
T,rr6ucnr, Y. (1957) The absolute structure of ullmanite. Mi'neral. f our' (fapan),2,90'
lo2.
M anuscript receitd., N owmber 4, 1964; aceepted,
Jor publication, Ianuary 21, 1965.