JOURNAL
OF SOLID
STATE
CHEMISTRY
19,75-80
(1976)
The Preparation and Characterization
CuFe,Ge,-,S, (0.5 < x < 1.0)
J. ACKERMANN,
S. SOLED,
of the Solid Solution Series
A. WOLD,*
AND E. KOSTINERf
* Department qf Chemistry, Brown University, Providence, Rhode Island 02912,
and the TDepartment of Chemistry and Institute of Materials Science, University of
Connecticut, Storrs, Connecticut 06268
Received March 26, 1976
Single crystals of the solid solution series CuFe,Gel-,Sz (0.5 < x i 1.0) have been prepared by
the chemical vapor transport technique. X-ray diffraction analysis and density measurements
have indicated that all members of this system crystallize with the chalcopyrite structure. Mijssbauer
spectra show that these crystals contain both iron(H) and iron(II1) on tetrahedral sites and that
the iron concentration agrees with that determined by chemical analysis. Magnetic susceptibilities
for x = 0.53 display antiferromagnetic behavior. The Nkel temperature of 12°K and an effective
moment of 5.0 BM is observed, which approaches the calculated spin-only moment of 4.92 BM.
As x increases, deviations from spin-only behavior occur, indicating complex magnetic interactions.
prepare members of the intermediate
solid
solution series CuFe,Ge,&
(0.5 < x < 1.0)
either as polycrystalline powders or, preferably
as single crystals. The members of this solution
series would contain both Fe”’ and Fe”
according to the stoichiometry CuFe,Ge,-,S,
= CuFe~‘_,Fe~:‘_,Ge,-,S, (0.5 < x < 1.O). The
growth, crystallographic, and magnetic properties of crystals which contain d5 and d6 iron
on tetrahedral sites are reported.
Introduction
Chalcopyrite, CuFeS,, crystallizes with an
ordered zincblende structure, space group
132d, in which all the atoms occupy slightly
distorted tetrahedral sites (I). The metal atoms
are located at the centers of sulfur tetrahedra,
each corner of which is shared by two neighboring tetrahedra (Fig. 1). Both MGssbauer
(2, 3) and neutron diffraction (4) studies indicate that chalcopyrite is antiferromagnetic
with a Ntel temperature of 853°K. These
studies also confirm that iron is present in the
high spin d5 electron configuration (Fe”‘).’
Briartite, CuFeo.sGeo.&, has been reported
to crystallize with a stannite-related structure
in the space group Id2m (Fig. 2) (5). Wintenberger et a/. (6, 7) have shown that briartite
orders antiferromagnetically
at 12°K and
contains iron in the high spin d6 state (Fe”).
It would be of interest to see if it is possible to
Experimental
Single crystals of the solid solution series
CuFe,Ge,-,S,
(0.5 < x < 1.O) were grown by
the chemical vapor transport technique
(8).
The elements were purified before use. High
purity copper (Matthey-Johnson
99.999 %)
was reduced in a 15 y0 HZ/85 ‘A Ar atmosphere
at 600°C for 4 hr. High purity iron (Leico
99.9999%) was reduced in a similar atmosphere at 950°C for 20 hr. Sulfur (Atomergic
99.9999 %) was vacuum sublimed at 120°C and
iodine (Baker 99.95 7;) at 50” to remove trace
impurities. Electronic grade germanium (Gal-
* To whom all correspondence should be addressed.
1 In this paper, Fen and Fe”’ will be used to
denote the d6 and d5 electron configurations of iron,
respectively.
Copyright
Q 1976 by Academic
Press, Inc.
All rights of reproduction
in any form reserved.
Printed in Great Britain
75
76
ACKERMAN
ET AL.
0 cu
0 Fe
OS
@ Sn
-Cl-
Chalcopyr
Stannite
FIG. 2. The stannite structure, space group Ii12m.
Note the ordering of the iron and tin atoms.
ite
FIG. 1. The structure of chalcopyrite, space group
md.
lard-Schlesinger 99.9999%) was used without
further treatment.
Nominal stoichiometries of Cu, Fe, Ge, and
S were introduced into an II-mm i.d. silica
tube and evacuated to 1 FmHg. After iodine
(5 mg/cm”) was sublimed inside (9), the silica
tube was sealed. The sealed tube was wrapped
with a tightly wound Kanthal wire (0.057-in.
diam.) coil to even out extraneous temperature
gradients within the transport furnace. The
tube was heated for 12 hr, during which the
growth zone was held 100°C higher than the
charge zone. This process back-transported
all nucleation sites from the growth zone.
After thermal equilibration of the growth and
charge zones, the temperature of the growth
zone was then reduced over a period of 5 days
to the values given in Table I. After the
growth period, the tube was cooled slowly to
room temperature. The crystals were removed
and washed with absolute ethanol in order
to extract the iodine. Using this method,
crystals were produced with dimensions
10 x 4 x 1 mm and weights up to 60 mg. The
crystals grew as platelets, with colors varying
from black, for those with low iron content,
to gold, for those with high iron content.
TABLE
I
CONDITIONS OF PREPARATION OF CuFe,Ge,-,S,
Charge stoichiometry
CuFeO&eO.St
CuFeo.&ed32
CuF%.7Geo.3S2
CuRdkb.&
CuJkd3eo.A
Growth stoichiometry
Transport
agent
CuFeo.53Ge0.47SZ
12
CuFeO.&eO.&
12
Cul%.&%&
12
CuFeo.s4Geo.l.&
CuFeo.dhdS
12
12
Temperature gradient
(charge zone-growth zone) Duration
(weeks)
(“C)
850-790
840-780
825-750
830-770
850-790
3
3
3t
3
3
SOLID SOLUTION SERIES
Chemical Analysis
In chemical vapor transport reactions of
mixed solution series there is the possibility
that the crystal growth stoichiometry differs
from the charge (nominal) stoichiometry. To
determine the true stoichiometry of the single
crystals, iron analysis was performed using the
photometric thiocyanate method (10, II).
Known weights of the unknowns were
dissolved in 50 ml of aqua regia. Oxides of
nitrogen and sulfur were dispelled from the
solution by boiling. Precipitated sulfur was
removed by filtration. After 400 ml of water
and 25 ml of 0.5 M potassium thiocyanate
were added, the solution was diluted volumetrically to 500 ml with water. The absorbance at 480 pm of the red (Fe(H,O),SCN)‘+
complex was measured at once on a Cary 14
spectrophotometer. Weighed amounts of
freshly reducedhigh-purity iron were dissolved
and used as references.A 0.025 M potassium
thiocyanate solution was used as the blank.
The charge and growth stoichiometries are
listed in Table I.
77
CuFe,Ge,-,S,
Chicago Corp.). Isomer shifts are reported
with respect to the zero position of a crystal of
sodium iron(U) nitropentacyanide dihydratc,
Na,(Fe(CN),NO)*2H,O (National Bureau of
Standards, Standard Reference Material No.
725). The quadrupole splitting of the standard
was taken as 1.7048mm/set (12).
Results, X-Ray Measurements, and Densities
X-ray powder diffraction patterns were
obtained for ground single crystals of the
members of the solution series CuFe,Ge,-,S,
(0.5 < x < 1.0) with a Norelco diffractometer
using monochromatic radiation (AMR Focusing Monochromator) and a high-intensity
copper source (CuKor, = 1.5405A). Scanning
rates of 1a 20/min and *” 28/min were used.
High purity silicon (a, = 5.4301 A) was added
as an internal standard in order to obtain
precise lattice dimensions. Both stepcounting
on ground crystals and long-exposure precession techniques on single crystals were used
to search for reflections causedby the ordering
of the iron and germanium atoms. Observation of the (101) reflection and the absence of
the (110) and (002) reflections for all members
MSssbauer Measurements
of this seriesshows that the members adopt the
The iron-57 Mossbauer spectra of ground I32a’ chalcopyrite space group in which the
singlecrystals of CuFe,Ge,_,S, (0.5 < x < 1.O) Fe”, Fe”‘, and Ge atoms are distributed
were measuredwith a model NS-1 Mossbauer randomly upon the Fe sites of chalcopyrite.
spectrometer (Nuclear Science and Engineer- The precision lattice constants for the meming Corp.) operating in the constant accelera- bers of this series are given in Table II. The
tion mode. The 14.4-keV radiation emitted introduction of iron into briartite, CuFe,.,from 20 mCi of 57Co diffused into Pd was Ge0.5SZ,replaces germanium with the larger
detected with a gas proportional counter and FelI1 atoms. In order to maintain electroneucollected with a 400-channel analyzer (Nuclear trality, some of the Fe” is converted into
TABLE
II
CRYSTALLOGRAPHIC PARAMETERSAND DENSITIESFOR CuFe,Ge,-,S,
x
0.53
0.65
0.75
0.84
0.95
1.0
c (4
v (A”)
P&S
PEdC
10.531(l)
10.536(l)
10.531(l)
10.517(l)
10.488(l)
10.409(l)
299.4
299.7
298.8
295.8
293.2
290.2
4.25(l)
4.20(l)
4.18(l)
4.19(l)
4.17(l)
4.18(l)
4.25
4.20
4.17
4.18
4.17
4.18
a (4
-__
~__
5.332(l)
5.334(l)
5.327(l)
5.304(l)
5.287(l)
5.280(l)
:8
ACKERMAN
Il.5
E.hE
6.K
FIG. 3. The variation
composition.
x 8.72
R.ES
ET AL.
changes in the a,- and c,,-axeswith stoichiometry are shown in Figs. 3 and 4 and disp!ay a
nonlinear behavior.
Densities of single crystals of the solid
solution series CuFe,Ge,-,S, (0.5 < x < 1.0)
were obtained by the hydrostatic technique
(13) using pertluoro-(1-methyldecalin) as the
fluid medium. Prior to each measurement,the
fluid was calibrated against a single crystal of
silicon with a density of 2.328 g/cm”. All
measurements were made at 24 f 0.2”C and
31 + 3 % relative humidity. The observed
densities are given in Table II with the values
calculated from the analyzed stoichiometries
and observed lattice constants.
8.9s
of the a0 parameter with
Miissbauer Spectra
-
Il.%+
Mossbauer spectra of the members of the
solid solution series CuFe,Ge,-,S, where
x = 0.53, 0.65, and 0.84 have been obtained.
The spectra display two quadrupole sp!it
absorptions. From the quadrupole splittings
it was possible to assign one of the doublets to
Fe” (QS 2.52-2.57 mm/set) and the other
doublet to Fen’ (QS 0.46-0.56 mm/set). The
ratio of the area under the Fen’ doublet to the
area under the Fen’ + Fe” doublets (Fe”‘/
Fe,,,,‘) is presented in Table III, together with
the ratios calculated from chemical analysis.
Observation of the two distinct doublets
displays the presenceof both Fen and Fe”’ in
the chalcopyrite structure. Furthermore, the
agreement between observed and calculated
Fel’VkotaI ratios shows the concentrations of
Fe” and Fe”’ to be consistent with chemical
D
D
ll.SZ..
D
D
18.99..
.z
.z
””
00
18.96..
18.9..
ID
D
18.W.
Il.92
B.SS
8.62 x 1.72
FIG. 4. The variation
composition.
8.82
I.%
of the cg parameter with
smaller Fe”‘. In such double substitutions the
change in lattice dimension with stoichiometry may be expected to be complex. The
TABLE III
THE ISOMER SHIFTS (IS) AND QUADRUPOLE
SPLITTINGS
(QS)
x = 0.53,0.65,
FOR MEMBERS OF THE SERIES
AND 0.84”
CuFe,Ge,-,Sz
WHERE
Ratio of Fe”‘/Fe tota,
Inner peaks (Fe”‘)
Outer peaks (Fe”)
x
IS
Qs
IS
QS
Mijssbauer
Analysis
0.53
0.65
0.84
0.568
0.546
0.567
0.55
0.47
0.49
0.884
0.857
0.858
2.52
2.57
2.48
0.13
0.42
0.70
0.11
0.46
0.81
a Ail spectra were recorded at 22 f 2°C.
CuFe,Ge,-,S,
SOLID SOLUTIONSERIES
analysis. For members of the solution series
with x 2 0.84, a six-line magnetic hyperfine
spectrum is seen; however, the lack of resolution of the peaks in this spectrum prevents an
accurate determination of the area ratios and
the Fe” and Fe”’ assignments.
Magnetic Measurements
Magnetic measurements were made using
the Faraday balance described by Morris and
Wold (14). For every sample the bulk magnetic susceptibility was measured as a function
of field strength at 300 and 77°K (HondaOwens method; 15, 16). The results are shown
in Figs. 5 and 6. The lack of any field dependence indicates the absence of ferromagnetic
impurities or bulk ferromagnetic order in these
79
Z-,’ :
giY 21..
_I,,’.*.*:
,l
>
,, -- .*
E Iq.’
,‘ ..*
,’
,’
z
,1.’ *.- Do
7..
,,I’ **..
__-‘*....
.a.-/
e...
‘_ _:..:.__
I
~Lb,,nl~bdLn~~noDI~oD~~nn~
.,AfAA~~Yh
X18
~ @a0
4b04”
; 68..
$
-no 0”
FIGS. 7 and 8. x,’
versus temperature for members
series Cu(Fe,Ge,-,)S,.
0.0,
Cu(Feo.&ecdSZ;
---, Cu (Feo.&eo.&h;
0 0 0,
Cu(Feo.75Ge~.&2; 000, Cu(Feo.&eo.ldS2;
A A A,
Cu(Feo.95Ge0.0s)S2.
of
FIGS. 5 and 6. The variation of the molar susceptibility with field strength at 300‘K (Fig. 5) and at 77’K
(Fig.6)forCu(FexGe,-,)S,.(A)x
=0.53;(B)x = 0.65;
(C) .Y= 0.75; (D) x = 0.84; (E) s = 0.95.
the
solution
samples. The susceptibility was also measured
as a function of temperature from 3 to 575°K.
The plot of inverse susceptibility vs temperature for CuFe0.53 Ge0.47S2 is shown in Fig. 7.
This material shows antiferromagnetic
order
with a N&l temperature of 12°K and a
paramagnetic moment of 5.0 BM, which is
consistent with a spin-only moment due to iron
in the high spin d6 configuration. For materials
with a greater Fe concentration,
complex
magnetic behavior involving Fe” and Fe”’
situated randomly in a chalcopyrite structure
vs
appears. In the inverse susceptibility
temperature curves (Figs. 7 and S), this complex magnetic behavior is observed by a
broadening of the Neel points as well as a
reduction of the paramagnetic moment from
the spin-only values. The magnetic
parameters for these materials are summarized
in
Table IV.
80
ACKERMAN
TABLE
ET AL.
IV
MAGNETIC PARAMETERS
OF CuFe,Gei-,S2
x
e (“K)
TN (“W
Perr(obs)(BM)
Perr<eaic)
(BM)
Temperature of
paramagnetic behavior (“K)
0.53
0.65
0.75
0.84
0.95
-12
33
80
112
12
16
16
35
4.99(2)
4.30(2)
3.56(2)
2.81(2)
5.00
5.36
5.57
5.72
40-575
100-575
185-575
300-575
1.0
NCel temperature and paramagnetic region
above decomposition temperature 853 (3, 18)
Bull. Sot.
Acknowledgment
6. J. ALLEMANLI
The assistance of the Army Research Office,
Durham, North Carolina, the Materials Research
Laboratory, Brown University, and the University of
Connecticut Research Foundation is acknowledged.
The authors would like to thank Mr. Don Schleich
for communicating the technique of using wire coils
to improve crystal growth by chemical vapor transport.
7. P. IMBERT,F. VARRET, AND M. WINTENBERGER,
References
I. N. W. BUERGER AND M. J. BUERGER, Amer.
Mineral. 19, 289 (1934).
2. E. FRANK, II, Nuovo Cim. B 58,407 (1968).
3. M. DIGUISEPPE, J. STEGER, A. WOLD, AND
E. KOSTINER, Znorg. Chem. 13,182s (1974).
4.
CORLISS, J.
ELLIOT, AND J. HASTINGS, P&x
G. DONNAY,
N.
L. M.
D. H.
DONNAY,
Rev. 112,1917
(1958).
5.
W. SCHAEFER AND R. NIETSCHE,
-645 (1974).
Mat. Res. Bull. 9,
AND M.
Fr. Mineral.
Cristallogr.
WINTENBERGER,
93, 14 (1970).
J. Phys. Chem. Solids 34,1675 (1973).
8. H. SCHAEFER,“Chemical Transport Reactions,”
Academic Press, New York (1964).
9. A. FINLEY, D. SCHLEICH, J. ACKERMANN, S. SOLED,
AND A. WOLD, Mat. Res. Bull. 9, 1665 (1974).
10. G. LUNDELL, H. BRIGHT, AND J. HOFFMAN,
“Applied Inorganic Analysis,” Wiley, New York
(1953).
II. “Standard Methods of Chemical Analysis,” Vol.
1, Van Nostrand Princeton, N.J. (1962).
12. R. W. GRANT, R. M. HOUSELEY,AND U. GONSER,
Phys. Rev. 178, 523 (1969).
13. R. L. ADAMS, Ph.D. thesis, Brown University
(1973).
14.
B. MORRIS
AND A. WOLD,
Rev. Sci. Znstrum. 39,
1937 (1968).
IS. K. HONDA, Ann. Phys. (N. Y.) 32,1048 (1910).
16. M. OWENS, Ann. Phys. (N. Y.) 37, 657 (1912).
17. T. TERANISHI, J. Phys. Sot. Japan 16,188l (1961).