Anda di halaman 1dari 6

ARTICLE IN PRESS

Journal of Physics and Chemistry of Solids 71 (2010) 993–998

Contents lists available at ScienceDirect

Journal of Physics and Chemistry of Solids


journal homepage: www.elsevier.com/locate/jpcs

Magnetic properties for the Cu2MnSnSe4 and Cu2FeSnSe4 compounds


E. Quintero a, M. Quintero a,n, E. Moreno a, L. Lara a, M. Morocoima a, F. Pineda a, P. Grima a,
R. Tovar a, P. Bocaranda a, J.A. Henao b, M.A. Macı́as b
a
Centro de Estudios de Semiconductores, Departamento de Fı́sica, Facultad de Ciencias, Universidad de Los Andes, Mérida 5101, Venezuela
b
Grupo de Investigación en Quı́mica Estructural (GIQUE), Facultad de Ciencias, Escuela de Quı́mica, Universidad Industrial de Santander,
Apartado aéreo 678, Bucaramanga, Colombia

a r t i c l e in fo abstract

Article history: Magnetic susceptibility w measurements in the range from 2 to 300 K were carried out on samples of
Received 21 October 2009 the Cu2FeSnSe4 and Cu2MnSnSe4 compounds. It was found that Cu2FeSnSe4 was antiferromagnetic
Received in revised form showing ideal Curie–Weiss behavior with a Néel temperature TN of about 19 K and Curie–Weiss
1 April 2010
temperature y ¼  200 K, while for Cu2MnSnSe4 the behavior was spin-glass with a freezing
Accepted 11 April 2010
temperature Tf of about 22 K and Curie–Weiss temperature y ¼  25 K. The spin-glass order parameter
q(T), determined from the susceptibility data, was found to be in agreement with the prediction of
Keywords: conventional spin-glass theory.
A. Inorganic compounds & 2010 Elsevier Ltd. All rights reserved.
A. Semiconductors
D. Magnetic properties

1. Introduction compounds by replacing the II cations with Mn, Fe, Co and/or


Ni ions. These compounds can be found in the section
Magnetic semiconducting materials (MSM) are of interest (I2–IV)1  xII3xVI3 at x¼0.25 with I¼ Cu, Ag, II ¼Mn or Fe,
because of the manner in which the magnetic behavior associated IV¼Ge, Sn, Pb and VI¼S, Se, Te. The crystal structure of various
with the concerned magnetic ion can modify and complement the I2–II–IV–VI4 compounds has been investigated by several authors
semiconductor properties [1,2]. These MSM have received atten- [3,4,5], and it has been found that in almost all cases, the
tion because of their potential application in optoelectronic and compounds showed either the tetrahedral tetragonal stannite
magnetic devices. It has been found that the particular magnetic (I42m) structure based on zinc-blende, or an orthorhombic
behavior occurring in any given case depends to a large extent on superstructure derived from wurtzite (known as wurtz-stannite,
the distribution of the magnetic atoms in the lattice. Thus for Pmn21).
MSMs in which these atoms are at random in the cation lattice the With regard to the work done on the I2–II–IV–VI4 materials,
material mainly shows spin-glass form [1]. For MSMs, where the the crystallographic and magnetic properties of the Cu2FeGeSe4
arrangement of magnetic atoms on the cation lattice is regular, a and Cu2FeGeTe4 compounds have been given in earlier work [6],
variety of magnetic states can occur, viz antiferromagnetism, and it was found that Cu2FeGeSe4 was antiferromagnetic with
ferromagnetism, canted ferromagnetism, magnetic competition bound magnetic polarons (BMPs) observed at low temperature,
between magnetic neighbors leading to magnetic frustration, etc. while that the obtained results for Cu2FeGeTe4 could be analyzed
[2,3]. The materials that have been most studied are the using the Néel theory of ferrimagnetism [6]. Guen and Glaun-
semimagnetic semiconductor alloys obtained from the tetrahed- singer [7] measured the magnetic susceptibility, using the
rally coordinated II–VI semiconductor compounds by replacing a Faraday method, on crystal and powder samples of Cu2MnSiS4,
fraction of the group II cations with a paramagnetic ion as Mn + 2, Cu2MnGeS4, Cu2FeGeS4, Cu2MnSnS4, Cu2MnGeSe4 and
Fe + 2, etc., giving alloys such as Cd1  zMnzTe, which show spin- Cu2MnSnSe4. Their results showed that the first four compounds
glass behavior, very large magneto-optical effects, bound mag- were antiferromagnetic, while that the last two ones, i.e.
netic polarons (BMPs), etc. [1,2]. It was recently suggested [3,4] Cu2MnGeSe4 and Cu2MnSnSe4, were ferromagnetic. However,
that another set of magnetic compounds and alloys, which could according to the magnetic measurements of Chen et al. [8], which
show larger magneto-optical effect than the II–VI derived alloys, were made using a SQUID magnetometer on 19 polycrystalline
can be obtained from the tetrahedral bonded I2–II–IV–VI4 samples of I2–Mn–IV–VI4 with I ¼Cu, Ag, IV¼Si, Ge; Sn, Pb and
VI¼Se,Te, these two compounds were found to be antiferromag-
netic. More recently, in the work made on the magnetic properties
n
Corresponding author. Tel.: + 58 274 2716979; fax: + 58 274 2401286. of the Cu2Cd1  zMnzGeSe4 system [9], it was found that the
E-mail address: mquinter10@intercable.net.ve (M. Quintero). Cu2MnGeSe4 compound is antiferromagnetic. This result is in

0022-3697/$ - see front matter & 2010 Elsevier Ltd. All rights reserved.
doi:10.1016/j.jpcs.2010.04.010
ARTICLE IN PRESS
994 E. Quintero et al. / Journal of Physics and Chemistry of Solids 71 (2010) 993–998

agreement with the data of Chen et al. [8], but it is in conflict with 100 G, using a quantum design SQUID magnetometer. The
the result of Guen and Glaunsinger [7] for this compound. The magnetic measurements showed that, in each case, the amount
above results suggest that the type of magnetic interaction in of secondary phase for the compressed samples was considerably
some of these materials is still not clear. Hence, in the present smaller than for the as-grown sample. For the magnetic
work the properties of the Cu2MnSnSe4 and Cu2FeSnSe4 com- parameter values could be estimated for the compressed samples,
pounds are investigated. The X-ray powder diffraction studies these magnetic results will be further discussed below. In the
made on these compounds have been reported in previous works case of Cu2FeSnSe4, measurements of magnetization M as a
[10,11], and it was found that, at room temperature, these function of applied field B up to 5 T were made at various fixed
compounds have the tetragonal stannite structure (I42m), with temperatures.
lattice parameter values a ¼5.7362(5) Å, c¼11.401(5) Å for
Cu2MnSnSe4 and a ¼5.7086 (2) Å, c¼11.2786 (2) Å for Cu2-
FeSnSe4. The phase transition temperatures for the Cu2MnSnSe4 3. Results, analysis and discussion
and Cu2FeSnSe4 compounds have been obtained in earlier work
[11], using differential thermal analysis (DTA) measurements. It was indicated above that the traces of secondary phase were
From these DTA experiments, it was concluded that both observed for each of the as grown compound, and that the
compounds melt incongruently, with a value of the liquid magnetic measurements showed that the amounts of any extra
temperature TL of about 800 1C for Cu2MnSnSe4 and 675 1C for phase for the compressed samples were considerably smaller than
Cu2FeSnSe4. Regarding the available magnetic data for these for the as grown samples. Thus, for the as grown samples, while it
compounds, as was indicated above, in the case of Cu2MnSnSe4 was possible to determine separate values of the magnetic
there are some discrepancies on the type of magnetic behavior of transition temperature for each phase, no values for the magnetic
this compound. Information on the magnetic parameter values for parameter y or C, to be discussed below, could be estimated in
the Cu2FeSnSe4 compound has not yet been given. Here, results of these cases. This is illustrated in Figs. 1 and 2, where the
magnetic susceptibility w measurements as a function of tem- variations of the inverse of the magnetic susceptibility 1/w with
perature T are presented for the Cu2MnSnSe4 and Cu2FeSnSe4 temperature T, for as grown and compressed samples, are,
compounds. In addition, magnetization measurements made on respectively, shown for Cu2MnSnSe4 and Cu2FeSnSe4. It can be
Cu2FeSnSe4 at various temperatures are given. seen from Fig. 1 that, for temperatures above 50 K, the 1/w vs T
curve obtained for the as grown sample presents the typical form
shown by the secondary MnSe phase [14], while that the amount
2. Sample preparation and experimental techniques of this extra phase is significantly reduced for the compressed
sample. Furthermore, the 1/w vs T curve obtained for the
The samples were produced by the melt and anneal technique. In compressed sample is linear at higher temperatures and the
each case, the components of 1 g sample were sealed under vacuum extrapolation of this line to 1/w ¼0 gives a negative value of the
(E10  5 Torr) in a small quartz ampoule, and then the components Curie–Weiss y   25 K indicating that this compound is
were heated up to 200 1C and kept for about 1–2 h, then the antiferromagnetic. It is also seen in the inset of Fig. 1 that the
temperature was raised to 500 1C using a rate of 40 K/h, and held at susceptibility w measurements carried out under zero-field
this temperature for 14 h. After the samples were heated from 500 to cooling ZFC (heating run) and field cooling FC (cooling run) gave
800 1C at a rate of 30 K/h and kept at this temperature for another different results, i.e. temperature hysteresis is observed below
14 h, the temperature was raised to 1150 1C at 60 K/h, and the Tf  22 K. The difference in the two experiments is a common
components were melted together at this temperature, initially, for characteristic of magnetic disorder and frustration and could be a
about 1 h. The furnace temperature was brought slowly (4 K/h) down hallmark of a spin-glass state, in which the magnetic disorder is
to 600 1C, and the samples were annealed at this temperature for 1 quenched and the interaction between spins are in conflict with
month. Then, the samples were slowly cooled to room temperature each other leading to frustration, this state will be discussed
using a rate of about 2 K/h. Hence, the resulting as-grown samples below. For antiferromagnetic behavior, the variation of 1/w with T
were investigated by X-ray powder diffraction, with a Rigaku D/MAX above the Néel temperature TN is given by the relation [15]
IIIB diffractometer equipped with an X-ray tube (CuKa radiation:
1=w ¼ ðTyÞ=C ð1Þ
l ¼1.5406 Å; 40 kV, 30 mA) using a diffracted beam graphite
monochromator. The X-ray diffraction patterns, obtained for each where y is the Curie–Weiss temperature and C the Curie constant.
sample, were indexed with the computer program DICVOL04 [12] The theoretical value of C is given by [15]
using an absolute error of 0.031 (2y) in the calculations, lattice
C ¼ NA g 2 mB 2 JðJ þ 1Þ=3KB W ð2Þ
parameter values were estimated. The space group was established
using CHECKCELL program [13]. It was found that, for each as-grown where NA is the Avogadro number, mB the Bohr magneton and W
compound, few extra lines were observed in the diffraction pattern the molecular weight of the compound. Thus, the experimental
and these were found to be due to the presence of small amounts of a data for T4100 K in Fig. 1, for Cu2MnSnSe4, were fitted to the
secondary phase [11]. Each sample, showing to be two phase, was above equation and the results were C ¼6.6  10  3 emu K/g
carefully crushed and the final powder was compressed in a new (which gives an experimental effective magnetic moment
quartz ampoule. Then, the compressed sample was re-melted at mexp  5.7 mB) and y ¼  25 K. The experimental value of C is,
1150 1C for about 3 or more hours, and the annealing procedure within the limits of the experimental errors, close to the
indicated above was repeated. It was found that, for the compressed theoretical value C ¼7.0  10  3 emu K/g (meff ¼5.9 mB), obtained
samples, the intensities of the extra lines were considerably reduced, from Eq. (2) using J¼ S¼5/2, L¼0 and g ¼2 for the Mn2 + ion, this
and no sensible variation of lattice parameter values of the main result indicates that the effect from MnSe is practically negligible
phase was observed for the as grown and compressed samples, both for the compressed sample. It is to be mentioned that the
giving similar parameter values [11]. Also, in each case, the stannite magnetic results obtained here for the Cu2MnSnSe4, as well as
structure I42m was observed. those obtained earlier for Cu2MnGeSe4 [9], are in agreement with
Measurements of magnetic susceptibility w as a function of the ones given by Chen et al. [8], but are in conflict with those
temperature T in the range 2–300 K were made, on the as grown reported by Guen and Glaunsinger [7]. From their magnetic
and compressed samples, with a fixed value of magnetic field B of susceptibility results on Cu2MnSnSe4, they concluded that this
ARTICLE IN PRESS
E. Quintero et al. / Journal of Physics and Chemistry of Solids 71 (2010) 993–998 995

60000
0.030 FC Cu2MnSnSe4
Compressed sample
0.025
50000
0.020

χ (emu/g)
0.015

40000 0.010

ZFC
1/χ (g/emu) 0.005

30000 0.000
10 15 20 25 30 35
T (K)

20000

As grown sample
10000
Compressed sample

0
-50 0 50 100 150 200 250 300
T (K)

Fig. 1. Variation of the inverse of the susceptibility 1/w with temperature T for Cu2MnSnSe4. Squares: as grown sample. Circles: compressed and melted for 3 h. Inset:
variation of magnetic susceptibility w with T. The direction of the ZFC (heating run), or FC (cooling run), is indicated by the corresponding arrow.

50000 4.9 mB. This is consistent with the fact that Fe2+ often exhibits an
As-grown
orbital contribution to the magnetic moment. The main factor that
Compressed and melted for 3 h can affect the value of the magnetic parameter C, so as to give a higher
Melted for more than 3 h value than the theoretical C obtained with J¼ 4, L¼S¼2 and g¼1.5, is
40000
the mass m of the material measured, since the value determined for
Cu2FeSnSe4 w depends upon this mass m. Thus, if an amount of a secondary phase
1/χ (g/emu)

or an impurity is present, the weighed value of the sample would not


30000
be accurate for m. In this case, the mass of the main phase contained
in m will be lower than the real one resulting in a high value of C, and
hence in an apparent high value of mexp. This fact is the main reason
20000 for the greater experimental mexp values obtained here, compared to
the full theoretical meff ¼6.7 mB obtained using J¼4 and g¼ 1.5 for
Fe2 + . In contrast to the Cu2MnSnSe4 case, it was observed that for the
10000 Cu2FeSnSe4 compound the ZFC and FC susceptibility measurements
gave similar results showing a peak at about 19 K, i.e. no temperature
hysteresis was detected, so that spin-glass form at low temperature is
0 ruled out here. This result together with the negative value of y
-200 -100 0 100 200 300 (¼  200 K) suggest that the Cu2FeSnSe4 compound is antiferromag-
T (K) netic with a Néel temperature TN of about 19 K.
Coming back to the w vs T results shown in the inset of Fig. 1
Fig. 2. Variation of the inverse of the susceptibility 1/w with temperature T for for Cu2MnSnSe4, it has been indicated above that the difference in
Cu2FeSnSe4. Circles: as grown sample. Squares: compressed and melted for 3 h.
the ZFC and FC experiments is a common characteristic of
Triangles: melted for more than 3 h.
magnetic disorder and frustration typical of a spin-glass state, in
which the magnetic disorder is quenched and the interaction
compound was ferromagnetic with an experimental effective between spins are in conflict with each other leading to
magnetic moment mexp 4.67 mB and a positive value of y of frustration. To describe the spin-glass state theoretically, Edwards
19.8 K. The large reduction of m from the value of spin-only and Anderson [16] introduced a spin-glass order parameter q as a
meff ¼5.9 mB for Mn + 2 was explained as possibly due to the strong function of temperature T. Based on the appropriate form of the
covalence of the Mn–Se bond. mean-field theory MFT, Sherrington and Kirkpatrick [17,18]
It can be seen from Fig. 2 that a similar behavior occurs for as predicted that the spin-glass order parameter q(T) vanishes at
grown and compressed samples of the Cu2FeSnSe4 compound. the freezing temperature Tf and it is proportional to (1  T/Tf)
However, in this case, it was found that the compressed sample below Tf. The value of q(T) can be extracted from the magnetic
melted for 3 h gives an experimental value of C of about 1.9  10  2 susceptibility, but treating C and y as effective spin-glass
emu K/g (mexp  9.6 mB), while that the sample melted for more than parameters rather than the true constants, i.e. the effective spin-
3 h gave a value of 1.0  10  2 emu K/g (mexp  7 mB), which is closer glass Curie constant CSG, and the effective spin-glass Curie–Weiss
to the theoretical value C¼0.91  10  2 emu K/g (meff ¼6.7 mB) temperature ySG, as follows [19]:
determined from Eq. (2) using L¼S¼2, J¼4 and a
g-factor of 1.5 for the Fe + 2 ion. The measured magnetic moment of T wðTÞ
qðTÞ ¼ 1 ð3Þ
Fe + 2 is found to be greater than the theoretical spin-only value of CSG þ ySG wðTÞ
ARTICLE IN PRESS
996 E. Quintero et al. / Journal of Physics and Chemistry of Solids 71 (2010) 993–998

1.2
1.0

0.8
1.0
0.6

q (T)
0.8 0.4

0.2
q (T)

0.6 0.0
0.0 0.2 0.4 0.6 0.8 1.0
(Tf-T)/Tf
0.4
Tf = 22.9 K

0.2

0.0
0 5 10 15 20 25
T (K)

Fig. 3. Variation of the order parameter q as a function of temperature T, the solid line indicates the mean field theory prediction. Inset: variation of the normalized order
parameter q(T/Tf) vs (Tf  T)/Tf , the line represents the fit to Eq. (4).

Thus, in the present work, it was of interest to estimate the 3


order parameter q(T) and Tf. Fig. 3 shows the q(T) curve obtained
2K Cu2FeSnSe4
using the w vs T data shown in the inset of Fig. 1, with
4K
CSG ¼2.45 emu K/mol Oe and ySG ¼22.3 K. It is seen that the
10 K
value of q(T) drops to zero at about Tf E22.9 K, which, as
20 K
expected, is higher than the temperature given by the maximum
2 40 K
in the cusp of the ZFC curve. In the inset of Fig. 3, the normalized
M (Am2/Kg)

order parameter q(T/Tf) is plotted vs the normalized temperature 50 K


(Tf  T)/Tf together with the resulting data fitted by the equation:
 
TTf b
qðT=Tf Þ ¼ q0 ðT=Tf Þ þa ð4Þ
Tf 1
From the fit we obtained q0 ¼0.006 70.002, a ¼1.01270.002
and b¼1.050 70.008, close to the values predicted by the mean-
field theory (MFT) [17].
Measurements of the magnetization M as a function of the
applied magnetic field B, at various fixed temperatures, were 0
made on the Cu2FeSnSe4 compound. The resulting M vs B curves
for values of B up to 5 T are shown in Fig. 4. It is seen from this 0 1 2 3 4 5
figure, that, for each temperature, the form of the M vs B curve is B (T)
linear and magnetic hysteresis was not observed. The curves
Fig. 4. Variation of magnetization M as a function of applied field B for a range of
obtained below 20 K are in the antiferromagnetic region, and temperatures.
neither spin-flop behavior and/or the saturation field could be
observed in the range of the present measurements. It is to be
crystal structure and values of y and TN are known for a given
noted that the M(B) curves do not have the typical form given by a
antiferromagnetic material, it is possible to determine values of
material showing bound magnetic polarons BMPs, and that the
the exchange interaction parameters between the various
values of M are smaller than those obtained when BMPs are
magnetic ions. The mean field theory gives the following
present in the material. This is illustrated in Fig. 5, where results
equations for y and TN [15]:
obtained at 4 K for the present Cu2FeSnSe4 together with those
reported previously for Cu2FeGeSe4 [20], Cu2Cd0.25Fe0.75GeSe4 2SðS þ 1Þ X
y¼ mi Ji ð5Þ
[21] and Cu2FeGeTe4 [21] are shown, no BMPs were observed for 3k
the Te compound. Moreover, when the data in Fig. 4 was analyzed
using the Langevin relation [20,21], it was found that good fits 2SðS þ 1Þ X
TN ¼ Zi mi Ji ð6Þ
could be obtained. However, in contrast to Cu2FeGeSe4 a 3k
temperature range where meff ¼ms could not be observed for the where the summation is over sets of equidistant magnetic neighbors
compound. Again, these results indicated that BMPs are not from a chosen magnetic atom, mi and Ji being the number and
present in the Cu2FeSnSe4 compound. exchange interaction for the ith set, S being the total spin of each
Another point of interest here is to discuss the exchange magnetic ion and Zi being +1 for antiparallel and 1 for parallel spin.
interaction parameters Ji in the present compounds. When the Values of Ji can now be obtained from Eqs. (5) and (6), but only two
ARTICLE IN PRESS
E. Quintero et al. / Journal of Physics and Chemistry of Solids 71 (2010) 993–998 997

6 significant value can be given to Z. However, in analogy with the


Cu2FeGeSe4 T = 5 K MnIn2Te4 compound, if some fractions of Cu and/or Mn atoms are
Cu2Cd0.25Fe0.75GeSe4 T = 5 K distributed at random on the 2a and 4d sites and with the Sn
Cu2FeGeTe4 T = 5 K atoms fixed on the 2b position, then there will be 8 nearest
magnetic neighbors. Also, considering that the Cu2MnSnSe4 is
Cu2FeSnSe4 T = 4 K
equivalent to a sample with z¼0.25 of the II1  zMnzVI alloys and
4
taking y ¼ 25 K and S ¼5/2 for Mn + 2, Eq. (5) gives J1/k¼  2.14 K
M (Am2/Kg)

for this compound. It is seen that in this case the value of J1/k is
five times smaller than the typical one for the II1  zMnzVI alloys.
Then a possible reason for this smaller value would be that y for a
random alloy is about seven times higher than that for the present
2 Mn compound.
With regard to the disorder of the Mn atoms, it is worth to
point out that two well known independent techniques are used
to investigate the local structure of a crystal: the X-ray absorption
fine structure (EXAFS) analysis and the fourier transform infra red
(FTIR) spectroscopy. Particularly, from experimental EXAFS
0
observations it is possible to obtain data on configuration
0 1 2 3 4 5 quantified probability deviation with respect to random conse-
B (T) quent to site occupation preferences, i.e. information on the
random distribution of the corresponding magnetic ion on the
Fig. 5. Variation of magnetization M as a function of applied field B at 4–5 K. Full
squares: Cu2FeSnSe4. Open circles: Cu2FeGeSe4 (from Ref. [20]). Open triangles: cation sub-lattice. This knowledge is important because, as was
Cu2Cd0.25Fe0.75GeSe4 (from Ref. [21]). Open squares: Cu2FeGeTe4 (from Ref. [21]). indicated above, site occupation preferences may affect crystal
behavior, characteristics and properties of the material.

independent parameter values can be obtained. Thus, if various values


of Ji are treated as independent, as is frequently done, only two values 4. Conclusions
of Ji can be found. As it was indicated above, the present compounds
have the tetragonal stannite structure I42m. Generally, in these I2–II– The magnetic results suggest that Cu2MnSnSe4 is spin-glass
IV–VI4 materials the cation positions are I—4d, II—2a, IV—2b, so that while Cu2FeSnSe4 shows ideal antiferromagnetic Curie–Weiss
the Mn2 + and/or Fe2 + are on 2a. Thus, assuming that the total spin behavior. In the case of Cu2MnSnSe4, the variation of the order
system can be treated as two interpenetrating sub-lattices, which parameter q(T) was found to be typical of a spin-glass material,
have no resultant interaction between them, then the magnetic unit giving parameter values in good agreement with the prediction of
cell with this spin arrangement has dimensions 2a, 2a, 2c. Values of Ji the MFT. In the case of Cu2FeSnSe4, the saturation field and spin-
can now be determined from Eqs. (5) and (6), but only two flop transitions as well as presence of BMPs were not observed in
independent exchange parameter values can be obtained. As it was the present M vs B measurements. The exchange integral values
discussed previously [9], for a fully ordered stannite structure, such as estimated for the present compounds are found to be smaller than
for the present Cu2FeSnSe4, with only J2 and J3, Eqs. (5) and (6) reduce those obtained for the II1  zMnzVI alloys. It was found that the
to y ¼A(4J2 +8J3) and TN ¼  A(4J2), where J2 and J3 are, respectively, amounts of secondary phase present in the as grown samples
the nearest magnetic neighbor and next nearest magnetic neighbor were considerably reduced for the compressed compounds.
exchange interaction terms, and A¼2S(S+1)/3k. These are clearly too
approximate to give any realistic value of Ji. If another term is added,
i.e. y ¼A (4J2 +8J3 +4J4) and TN ¼ A (4J2  4J4), then, in this case, it can Acknowledgements
be assumed that J3  J4 since r3 and r4 are close. With the values of
TN ¼20 K and y ¼ 200 K given above for Cu2FeSnSe4, and taking This work was partially supported by the Consejo de
S¼2 for Fe2+ , this gives J2/k¼ 4.06 K and J3/k¼ 2.81 K. The Desarrollo Cientı́fico, Humanistico y Tecnológico CDCHT-ULA
resulting value of J2/k is about two orders of magnitude smaller than (Projects C-1436-06-05-AA, C-1437-06-05-Ed, C-1547-08-05-B,
J1/k for the II1 zMnzVI random alloys, typically J1/k 10 K [22]. This C-1532-07-05-B and C-1630-09-05-B). Thanks are due to the Plan
result is possible since the value of y is of the same order of de Desarrollo de Talentos Humanos de Alto Nivel of Fondo
magnitude than for II1 zMnzVI, but the number of nearest magnetic Nacional de Ciencia, Tecnologı́a e Investigación FONACIT, through
neighbors connected by J2 for the ordered stannite is m2 ¼4 Grant ULA-proyecto no. 2008002062 (Contract no. 200900476),
(separated by a distance d¼a) compared to m1 ¼12 (separated by Venezuela.
d¼a/O2) for random alloys.
Turning to the Cu2MnSnSe4, from the magnetic results it can References
be concluded that this compound shows spin-glass behavior
similar to the II1  zMnzVI alloys [1,22]. The main property giving [1] J.K. Furdyna., J. Kossut., Diluted Magnetic Semiconductors, in: R.K. Willardson,
spin-glass form is the random distribution of the corresponding A.C. Beer (Eds.), Semiconductors and Semimetals, vol. 25, Academic Press,
New York, 1988.
magnetic ion on the cation sub-lattice. Thus, the difference in [2] Y. Shapira, E.J. McNiff. Jr., N.F. Oliveira. Jr., E.D. Honig., K. Dwight., A. Wold.,
magnetic behavior between Cu2FeSnSe4 and Cu2MnSnSe4 can be Phys. Rev. B 37 (1988) 411.
attributed to the fact that in the former the Fe atoms are ordered [3] G.H. McCabe., T. Fries., M.T. Liu., Y. Shapira., L.R. Ram-Mohan., R. Kershaw., A.
Wold., C. Fau., M. Averous., E.J. McNiff. Jr, Phys. Rev. B 56 (1997) 6673.
on the cation sub-lattice while in the latter the Mn is disordered.
[4] W. Shafer, R. Nitsche, Mater. Res. Bull. 9 (1974) 645.
A similar behavior was reported for the MnIn2Te4 compound, [5] M. Quintero, A. Barreto, P. Grima, R. Tovar, E. Quintero E., G. Sanchez Porras, J.
which also is I42m with the Mn and In atoms arranged at random Ruiz, J.C. Woolley, G. Lamarche, A.-M. Lamarche, Mater. Res. Bull. 34 (1999) 2263.
on the cation sites [23,24]. With the spin-glass behavior of this [6] M. Quintero, R. Tovar, A. Barreto, E. Quintero, J. González, G. Sánchez Porras, J.
Ruiz, P. Bocaranda, J.M. Broto, H. Rakoto, R. Barbaste, Phys. Status. solidi (b)
Cu2MnSnSe4 compound, the spin orientation of the Mn ions below 209 (1998) 135.
Tf will be random so that Eq. (6) cannot be used, since no [7] L. Guen, W.S. Glaunsinger, J. Solid State Chem. 35 (1980) 10.
ARTICLE IN PRESS
998 E. Quintero et al. / Journal of Physics and Chemistry of Solids 71 (2010) 993–998

[8] X.L. Chen, A.-M. Lamarche, G. Lamarche, J.C. Woolley, J. Magn. Magn. Mater. [15] J.S. Smart, in: Effective Field Theory of Magnetism, Saunders Company,
118 (1993) 119–128. Philadelphia and London, 1966.
[9] E. Quintero, M. Quintero, E. Moreno, M. Morocoima, P. Grima, J. Henao, J. [16] S.F. Edwards, P.W. Anderson, J. Phys. F: Met. Phys. 5 (1975) 965.
Pinilla, J. Alloys Compd. 471 (1–2) (2009) 16–20. [17] D. Sherrington, S. Kirkpatrick, Phys. Rev. Lett. 35 (1975) 1792.
[10] V.P. Sachanyuk, I.D. Olekseyuk, O.V. Parasyuk, Phys. Status Solidi (a) 203 [18] S. Kirkpatrick, D. Sherrington, Phys. Rev. B 17 (1978) 4384.
(2006) 459. [19] T. Mizoguchi, T.R. McGuire, S. Kirkpatrick, J.R. Gambino, Phys. Rev. Lett. 38
[11] E. Moreno, M. Quintero, M. Morocoima, E. Quintero, P. Grimaa, R. Tovar, P. (1977) 89.
Bocaranda, G.E. Delgado, J.E. Contreras, A.E. Mora, J.M. Briceño, R. Avila [20] E. Quintero, R. Tovar, M. Quintero, J. González, J. Ruiz, P. Bocaranda, J.M. Broto,
Godoy, J.L. Fernandez, J.A. Henao, M.A. Macı́as, J. Alloys Compd. , doi:10.1016/ H. Rakoto, R. Barbaste, J. Magn. Magn. Mater. 210 (2000) 208–214.
j.jallcom.2009.07.066. [21] E. Quintero, M. Quintero, M. Morocoima, P. Bocaranda, J. Appl. Phys. 102 (8)
[12] D. Loüer, A. Boultif, J. Appl. Crystallogr. 24 (1991) 987. (2007) 083905.
[13] Jean Laugier and B. Bochu, CHEKCELL: LMGP-Suite Suite of Programs for the [22] J. Spalek, A. Lewicki, Z. Tarnawski, J.K. Furdyna, R.R. Galazka, Z. Obuszko, Phys.
interpretation of X-ray. Experiments, ENSP/Laboratoire des Matériaux et du Rev. B 33 (1986) 3407.
Génie Physique, BP 46. 38042 Saint Martin d’He res, France /http://www. [23] K.-J. Range, H.-J. Hubner, Z. Naturf. 30b (1975) 145.
inpg.fr/LMGPS and /http://www.ccp14.ac.uk/tutorial/lmgp/S, 2002. [24] M. Morocoima, F. Pineda, M. Quintero, E. Quintero, E. Moreno, P. Grima, R.
[14] J.B.C. Efrem D’Sa, et al., Pramana- J. Phys. 63 (2) (2004). Tovar, P. Bocaranda, J.A. Henao, Physica B 404 (2009) 1819–1825.

Anda mungkin juga menyukai