Simultaneous diffraction and resistance measurements of metal sulphides

Article abstract of DOI:10.1016/j.jpcs.2006.12.018

We describe the construction of a cell that enables the electrical resistance of a material to be measured simultaneously with the collection of powder neutron diffraction data at elevated temperatures. The cell has been used to investigate the origin of the anomalous electron-transport properties of the monosulphide, CrS. The data reveal that a decrease in electrical resistance above 573 K is associated with the disappearance of semiconducting monoclinic CrS and the concomitant growth of a disordered CdI₂-type metallic phase.

Full text of DOI:10.1016/j.jpcs.2006.12.018

ARTICLE IN PRESS
Journal of Physics and Chemistry of Solids 68 (2007) 1052–1056
www.elsevier.com/locate/jpcs
Simultaneous diffraction and resistance measurements
of metal sulphides
a,
Ã
a
a
Anthony V. Powell , T. Erinc Engin , Paz Vaqueiro , Steve Hull
b
a
Department of Chemistry, Heriot-Watt University, Edinburgh EH14 4AS, UK
b
ISIS Facility, Rutherford Appleton Laboratory, Didcot, Oxfordshire OX11 0QX, UK
Abstract
We describe the construction of a cell that enables the electrical resistance of a material to be measured simultaneously with the
collection of powder neutron diffraction data at elevated temperatures. The cell has been used to investigate the origin of the anomalous
electron-transport properties of the monosulphide, CrS. The data reveal that a decrease in electrical resistance above 573 K is associated
2
with the disappearance of semiconducting monoclinic CrS and the concomitant growth of a disordered CdI -type metallic phase.
r 2007 Elsevier Ltd. All rights reserved.
Keywords: A. Inorganic compounds; A. Chalcogenides; C. Neutron scattering; D. Transport properties
1
. Introduction
We have sought to construct sample cells that enable the
collection of powder neutron diffraction data to be coupled
with the measurement of the key physical property of
electrical conduction. Initial development and testing of
this approach was carried out on the Polaris diffractometer
at the ISIS facility. These proof-of-principle studies have
focused on a range of metal sulphides that exhibit
anomalies in the temperature dependence of their electrical
conductivity. Here, we present an investigation of struc-
ture–property relationships in the binary sulphide, CrS,
which owing to the degenerate ground state associated with
The high count-rates available on modern powder
neutron diffractometers afford the opportunity to collect
diffraction data at small increments over a wide range of
some external variable within a realistic timescale. This
allows, for example, the evolution of both nuclear and
magnetic structure with temperature to be followed in real
time. Such measurements have provided considerable
insights into complex physical phenomena such as charge
and orbital ordering [1], superconductivity [2], ionic
conductivity [3] and magnetoresistance [4]. However, the
value of studies of structure as a function of temperature
may be greatly enhanced if relevant bulk properties can be
simultaneously investigated, since a more direct correlation
of subtle structural changes with physical properties is
achieved and variations between measurements obtained
on samples from different preparations eliminated. Neu-
trons offer considerable advantages over X-rays in this
regard as they also probe magnetism, which is intimately
linked to the above phenomena. However, such ‘diffrac-
tion-plus’ measurements require the development of
dedicated sample cells.
4
the d electron configuration of Cr(II), is susceptible to a
Jahn–Teller distortion [5]. It was proposed [6] that the
suppression of this distortion is responsible for an apparent
insulator to metal transition on heating above 673 K. The
first results of our programme, reported here, demonstrate
the feasibility of simultaneous diffraction–resistance mea-
surements and also reveal new insights into the origin of
the electronic anomalies in CrS.
2. Experimental
A sample of CrS was synthesised using the flux method
described by Popma and van Bruggen [7]. Appropriate
quantities of chromium (Aldrich, 99%), sulphur (Aldrich,
99.9%) and iodine (Fisher 99.9%) corresponding to a
Ã
Corresponding author. Fax: +44 131 451 3180.
E-mail address: a.v.powell@hw.ac.uk (A.V. Powell).
0022-3697/$ - see front matter r 2007 Elsevier Ltd. All rights reserved.
doi:10.1016/j.jpcs.2006.12.018

ARTICLE IN PRESS
A.V. Powell et al. / Journal of Physics and Chemistry of Solids 68 (2007) 1052–1056
1053
1
:1 wt% ratio of CrS and CrI were mixed and sealed into
2
were collected, whilst simultaneously recording the elec-
trical resistance, for ca. 2 h at each temperature point in the
range 298pT/Kp873. Additional data sets were collected
at 523 and 973 K on a powdered sample contained in an
evacuated silica ampoule. Data collected for an empty
silica tube were used to provide a background that was
subtracted from the diffraction data before structural
refinement. Data were processed using the Genie package
[10] and Rietveld analysis carried out using GSAS [11].
ꢀ
4
an evacuated (o10 Torr) silica tube. The reaction
mixture was heated at 1373 K for 24 h, cooled to 973 K
ꢀ
1
at 3 K min , after which it was removed from the furnace
and allowed to cool to room temperature. The CrI2
flux was removed by washing with water and the solid
product dried at room temperature. The sample was
characterised by powder X-ray diffraction using a Philips
PA2000 diffractometer operating with nickel-filtered
Cu-K radiation.
a
The in-situ cell for diffraction-resistance measurements
at elevated temperatures, consists of 10 mm diameter high-
s
purity (Spectrosil ) fused silica tubing terminated by two
3. Results
Powder X-ray diffraction data for the flux-grown
product reveal that in addition to reflections arising from
the monoclinic form of CrS [5], there are additional peaks
which can be indexed on an hexagonal sub-cell. This is
consistent with the presence of a second phase, also with a
NiAs-related structure. All attempts to prepare CrS under
a variety of synthetic conditions produced essentially the
same powder X-ray diffraction pattern. A reflection from
quartz-to-metal seals (Fig. 1(a)). The glass selected has a
low (o1%) boron content to minimise neutron absorption.
A cylindrical pellet (5 mm diameter ꢁ 8 mm length) of CrS
was fabricated from the powdered sample by pressing in a
5
mm die and sintering at 973 K in an evacuated
ꢀ
4
(
o10 Torr) silica tube. Electrical contact to the sample
was made by two stainless-steel electrodes, welded to the
glass–metal seal after encapsulating the pelletised material
in the cell. Stainless-steel springs were incorporated into
each electrode in order to compensate for differential
thermal expansion of cell components under the conditions
of the experiment. After assembly, the cell was evacuated
˚
this second phase at ca. 5.77 A violates the reflection
conditions for the P6 /mmc space group of hexagonal CrS
3
[12]. However, it can be indexed as the (0 0 1) reflection of a
CdI -type structure in which octahedral sites between
2
successive pairs of close-packed sulphide layers are
alternately fully and partially occupied. The absence of
any additional low-angle peaks suggests that there is no
two-dimensional ordering of defects within the partially
occupied layer. Therefore this second phase was accounted
for using a Cr S -type structural model [5] described in the
ꢀ
4
to o10 Torr via a glass side arm which was then sealed
with a blowtorch. The sealed enclosure prevented loss of
sulphur that has previously been observed at high
temperatures under the dynamic vacuum within the
furnace [8]. In an effort to minimise the background
scattering from the cell, the stainless-steel electrodes were
shielded by wrapping the top and bottom of the cell with
gadolinium foil. A thermocouple was attached to the
exterior of the cell and shielded by the gadolinium foil. The
whole assembly was mounted on a standard ISIS furnace
centre stick (Fig. 1(b)), located in a furnace with vanadium
7
8
space group P-3 m1, in which the octahedral sites between
pairs of octahedral CrS layers are partially occupied in a
2
disordered fashion. A refined site-occupancy factor of
0.70(3), indicates a slightly cation-deficient Cr S -type
phase with a stoichiometry of Cr6.8S8.
7
8
A section of the powder neutron diffraction data from
the highest resolution back-scattering bank of detectors,
collected in the resistance cell over the temperature range
298pT/Kp873 is shown in Fig. 2. Data from room
temperature to 573 K are well fitted by a combination of a
monoclinic CrS phase, together with the trigonal cation-
deficient Cr S -type structure (Fig. 3(a)). In addition to
ꢀ
4
windows that was then evacuated to o10 Torr. Electrical
connections for four-probe DC resistance measurements
were spot welded to the ends of the stainless-steel
electrodes and connections made to a Keithley 2000
multimeter. Time-of-flight powder neutron diffraction data
were collected on the Polaris diffractometer [9] at the ISIS
facility, Rutherford Appleton Laboratory. Neutron data
7
8
scattering from the cell (predominantly austenite), most
Fig. 1. Construction of the diffraction–resistance cell for high-temperature measurements: (a) location of the pelletised chromium sulphide sample
between the two electrodes sealed into the evacuated cell (Gd shielding removed) and (b) the cell assembly mounted on a furnace centre stick.

ARTICLE IN PRESS
A.V. Powell et al. / Journal of Physics and Chemistry of Solids 68 (2007) 1052–1056
1054
3
2
2
1
1
0
5
0
5
0
5
0
5
8
73
5
73
73
4
-
-
10
1.0 1.2 1.4 1.6 1.8 2.0 2.2 2.4 2.6 2.8 3.0
d/Å
298
b
5
4
3
0
0
0
1
.90
1.95
2.00
2.05
d/Å
2.10
2.15
2.20
Fig. 2. A region of the powder neutron diffraction data from the back-
scattering detector bank, obtained over the temperature range 298pT/K
p873. The onset of the structural change at 573 K is evidenced by the
˚
sudden divergence of the (1 0 2) reflection (ca. 2.14 A at 573 K) of the
˚
-type phase, from the strong reflection at 2.12 A that arises from the
20
7
Cr S
8
cell.
10
˚
evident at ca. 2.1 A, reflections arising from a small
quantity of Cr are present in diffraction patterns at all
temperatures: the amount present varies slightly (6–8 wt%)
between different preparations. Data from the low-angle
0
-
10
1
.0 1.2 1.4 1.6 1.8 2.0 2.2 2.4 2.6 2.8 3.0
detector bank also exhibit an additional intense peak at ca.
˚
d/Å
4
.2 A, which is magnetic in origin. The disappearance of
this reflection at temperatures between 433 and 453 K,
compares favourably with the previously reported Neel
Fig. 3. Observed (points), calculated (full line) and difference (lower full
line) profiles for CrS at (a) 523 K (Rwp ¼ 4.7%); reflection markers are for
´
temperature for CrS of T ¼ 460 K [13]. Detailed analysis
7 8
the trigonal Cr S -type phase (lower), monoclinic CrS (middle) and a-Cr
N
(
upper); (b) 973 K (Rwp ¼ 2.5%); reflection markers correspond to the
of the magnetic properties of CrS will be presented
elsewhere in due course. At temperatures below 453 K,
the magnetic contribution to the scattering in the highest-
trigonal Cr -type phase (lower) and a-Cr (upper). Scattering from the
7 8
S
˚
cell at ca. 2.1 A is excluded from the refinement.
˚
resolution bank is most evident at ca. 2.9 A. This region
was therefore excluded from structural refinements. The
small amount of remaining magnetic scattering below TN
slightly modifies the derived phase fractions, and is
responsible for their apparent minor variation between
room temperature and ca. 430 K. No anomalies in the
lattice parameters of either the monoclinic CrS or Cr S -
type phase are observed in this region, suggesting that the
resistance change at ca. 400 K is an artefact associated with
take up of thermal stresses in the cell. Data obtained
remains effectively constant up to 573 K, above which, the
weight fraction of the latter increases dramatically, with a
corresponding decrease in the amount of the monoclinic
phase. There is a strong correlation between the increase in
weight fraction of the trigonal phase above 573 K and the
onset of an almost linear decrease in the logarithm of
resistance of the sample (Fig. 4(b)). Both the a and c lattice
parameters of the trigonal phase exhibit anomalies in their
temperature dependence at 573 K, which is reflected in a
marked change in dV/dT at this temperature (Fig. 4(c)).
7
8
at T4573 K are well described (Fig. 3(b)) by
a
single chromium sulphide phase of the Cr7S8 type,
described in the space group P-3m1, together with
elemental Cr (ca. 8%).
4. Discussion
Rietveld analysis using data collected in the range
˚
.5–3 A was used to determine the weight fraction of each
of the phases present. Refinement reveals (Fig. 4(a)) that
the relative amounts of the monoclinic and trigonal phases
Jellinek [5] first characterised the structure of monoclinic
CrS from a crystal fragment that was found to consist
of two phases: trigonal Cr S and monoclinic CrS. Powder
0
7
8
X-ray diffraction data also indicated the presence of a-Cr

ARTICLE IN PRESS
A.V. Powell et al. / Journal of Physics and Chemistry of Solids 68 (2007) 1052–1056
1055
1
00
with lattice parameters related to those of the primitive
pffiffi
a
b
c
CrS
Cr
9
8
7
6
5
4
3
2
1
0
0
0
0
0
0
0
0
0
hexagonal unit cell by a ꢂ b ꢂ 2 3ap, and c ꢂ 2cp. By
contrast, others have argued that the apparent superlattice
lines are the result of impurities [16] or of l/2 contamina-
tion [7]. Similarly, the powder neutron diffraction data
presented here provide no evidence for such an enlarged
unit cell.
Cr S
7
8
Popma and van Bruggen [7] determined that a quenched
sample of Cr0.96S, which is close to the composition
prepared here, consisted of a mixture of monoclinic and
NiAs-type CrS at room temperature. However, on heating,
separation of the NiAs phase occurred at 350 K to yield the
same mixture of phases as obtained here. At 550 K, the
resulting Cr S was converted into a NiAs phase while at
100
7
8
870 K the monoclinic CrS disappeared. The distinction
between hexagonal NiAs and trigonal Cr S (CdI type) is
based upon the presence or otherwise of the (0 0 1)
10
7
8
2
1
˚
diffraction peak at ca. 5.8 A. This reflection is clearly
present in our X-ray diffraction data at room temperature.
However, in the neutron diffraction data, this weak
reflection occurs in the low-angle detector bank where the
statistics are relatively poor, making unambiguous identi-
fication difficult. For cation-deficient materials, the NiAs
and CdI2 structure types correspond to interlayer and
intralayer disorder of cation vacancies, respectively. The
temperature variation of lattice parameters show no signs
of a discontinuity below the phase coalescence temperature
6
6
6
3
2
1
6
0
(
573pT/Kp623) that would indicate a change in the
300
400
500
600
700
800
900
nature of the vacancy disorder. Therefore, we believe that
the intralayer disorder persists through the electronic
anomaly. The Cr7S8 to NiAs transition temperature
depends on the difference between inter and intralayer
vacancy–vacancy interaction energies [17]. When these are
T/K
Fig. 4. (a) Temperature variation of the refined weight fractions of the
three phases present, (b) the corresponding change in the electrical
resistance of the sample over the temperature range 298pT/Kp873, and
(
c) the temperature variation of the unit-cell volume of the trigonal
approximately equal, the CdI form of a cation-deficient
2
7 8
Cr S -type phase.
phase may never be stabilised [8]. Conversely if the
interlayer interaction is much higher than the intralayer
energy, only the CdI phase will be obtained, even up to
relatively high temperatures [18], as appears to be the case
in the material investigated here.
and it was proposed that the monoclinic phase was formed
from a solid-state reaction involving diffusion of Cr into
Cr S . An investigation of the transport properties of CrS
x
2
7
8
materials with 1.0pxp1.2 identified a large decrease in
resistivity in the temperature range 573–773 K for materials
with xp1.12 [6]. It was proposed [14] that the large
temperature range over which the decrease in resistivity
takes place, is due to the occurrence of two phase
transitions, requiring exsolution and transformation of
CrSx of unstated composition from the trigonal phase
around 600 K followed by the monoclinic-hexagonal
transformation at 870 K of CrS originally present. The
refined weight fractions determined in the present work do
not appear to support this model as they reveal that the
loss of the monoclinic phase that begins at ca. 600 K is
continuous to high temperatures.
The d-electron states in transition-metal sulphides are
located between the filled anion 3p-bands and the empty
cation 4s-levels. The d-levels may be localised, resulting in
semiconducting behaviour, or may overlap with either the
valence or conduction bands, resulting in metallic beha-
viour. Furthermore, mixing of the d-levels with anion
orbitals leads to the formation of d-bands, which if
sufficiently broad, can result in itinerant electron beha-
viour. Band structure calculations [19] performed for non-
magnetic hexagonal CrS, indicate that t2g and e levels are
g
significantly broadened by metal–metal and metal–sulphur
covalency. Intraatomic exchange lifts the spin degeneracy
4
of these levels. For a high-spin d configuration, this would
The change in electrical properties has been associated
with a transition from a mixture of monoclinic- and NiAs-
type CrS to a single phase of the NiAs structure [6].
However, these authors indexed the monoclinic phase on
the basis of a superlattice, first proposed by Haraldsen [15],
give rise to a partially occupied s-sub-band of mainly
cation eg character and metallic behaviour would be
expected. However, the cooperative Jahn–Teller distortion
around Cr(II):d in the monoclinic phase removes the
orbital degeneracy of this s band, leading to electron
4

ARTICLE IN PRESS
A.V. Powell et al. / Journal of Physics and Chemistry of Solids 68 (2007) 1052–1056
1056
localisation in the room temperature phase. By contrast,
the trigonal phase of Cr S shows no such distortion and is
metallic at all temperatures [14].
The decrease in electrical resistance of CrS at tempera-
tures above 573 K could arise [6] from a transformation
from a semiconducting monoclinic phase, in which the
Appendix A. Supplementary materials
7
8
A diagrammatic representation illustrating the compo-
nents of the cell is available (Supplementary information).
Supplementary data associated with this article can
be found in the online version at doi:10.1016/j.jpcs.
2006.12.018.
Jahn–Teller distortion localises the e electron, to a metallic
g
hexagonal NiAs-type form, in which the Jahn–Teller
distortion is suppressed and the eg electron transports
charge in a narrow s band. However, the data presented
here suggest that the fall in the resistance of CrS above
References
[
1] P.G. Radaelli, Y. Horibe, M.J. Gutmann, H. Ishibashi, C.H. Chen,
R.M. Ibberson, Y. Koyama, Y.S. Hor, V. Kiryukhin, S.W. Cheong,
Nature 416 (2002) 155.
5
disordered CdI -type metallic phase (Cr S ) with concomi-
73 K is associated with the progressive growth of a
2
7 8
tant loss of the semiconducting monoclinic CrS compo-
nent. The results demonstrate that this process occurs over
a range of temperatures, suggesting that it is driven by a
change in the stability range of the cation-deficient Cr7S8
phase with increasing temperature. This is reflected in an
increase in the site-occupancy factor associated with the
site in the partially occupied layer, from 0.72(1) at 523 K to
[
2] J.M. Tranquada, B.J. Sternlieb, J.D. Axe, Y. Nakamura, S. Uchida,
Nature 375 (1995) 561.
[3] S. Hull, Rep. Prog. Phys. 67 (2004) 1233.
[
[
[
[
[
4] M.B. Salamon, M. Jaime, Rev. Mod. Phys. 73 (2001) 583.
5] F. Jellinek, Acta Crystallogr. 10 (1957) 620.
6] T. Kamigaichi, J. Phys. Soc. Jpn. 15 (1960) 1355.
7] T.J.A. Popma, C.F. van Bruggen, J. Inorg. Nucl. Chem. 31 (1969) 73.
8] P. Vaqueiro, S. Hull, B. Lebech, A.V. Powell, J. Mater. Chem. 9
(1999) 2859.
0
.87(1) at 973 K. The changing composition of the trigonal
[
9] S. Hull, R.I. Smith, W.I.F. David, A.C. Hannon, J. Mayers, R.
Cywinski, Physica B 180 (1992) 1000.
phase is responsible for the observed discontinuity in the
unit-cell volume at 573 K. The decrease in defect concen-
tration in the trigonal phase requires diffusion of chro-
mium into the defective structure, through a solid-state
reaction, as proposed by Jellinek [5], although the data
presented here suggest that the stoichiometric CrS compo-
sition is not reached even at high temperatures.
[
10] W.I.F. David, M.W. Johnson, K.J. Knowles, C.M. Moreton-Smith;
G.D. Crosbie, E.P. Campbell, S.P. Graham, J.S. Lyall, Rutherford
Appleton Laboratory Report, RAL-86-102, 1986.
[11] A.C. Larson, R.B. von Dreele, General Structure Analysis System,
Los Alamos Laboratory, 1994 (Report LAUR 85-748).
[
[
[
[
[
[
12] N. Schoenberg, Acta Metall. 2 (1954) 427.
13] J.T. Sparks, T. Komoto, J. Phys. (Paris) 25 (1964) 567.
14] V.V. Sokolovich, Sov. Phys. Solid State 34 (1992) 371.
15] H. Haraldsen, Z. Anorg. Allg. Chem. 234 (1937) 372.
16] F. Jellinek, Ark. Kemi. 20 (1963) 447.
Acknowledgements
17] K. Kosuge, Chemistry of Non-stoichiometric Compounds, Oxford
University Press, Oxford, 1993.
We wish to thank the UK EPSRC and the CCLRC
Centre for Materials Physics and Chemistry for supporting
this work.
[
[
18] Y. Oka, K. Kosuge, S. Kachi, J. Solid State Chem. 23 (1978) 11.
19] J. Dijkstra, C.F. Van Bruggen. C. Haas, R.A. de Groot, J. Phys.:
Condens. Matter 1 (1989) 9163.