Hexagonal CoSe formation in mechanical alloyed Co75Se 25 mixture

Article abstract of DOI:10.1016/j.ssc.2004.03.044

A hexagonal CoSe alloy with NiAs-type structure was obtained by mechanical alloying starting from a mixture of pure crystalline powders with nominal composition Co₇₅Se₂₅. X-ray diffraction (XRD), differential scanning calorimetry (DSC), M?ssbauer spectroscopy (MS) and Raman scattering (RS) techniques were used to follow the structural, thermal, magnetic and optical properties of the binary mixture as a function of milling time. XRD results show the formation of a nanometric hexagonal CoSe phase between 3 and 70 h of milling coexisting with non-reacted Co phases, also in nanometric scale. DSC and RS results showed some changes in the thermal and optical properties of the crystalline phases when the milling time increases. The Raman active modes of the CoSe and Co oxide phases were observed. MS results showed practically no iron in the samples milled up to 15 h, while for extended milling times (70 h), they showed the presence of some α-Fe and the formation of other iron alloys due to the contamination by the milling media.

Full text of DOI:10.1016/j.ssc.2004.03.044

Solid State Communications 131 (2004) 265–270
www.elsevier.com/locate/ssc
Hexagonal CoSe formation in mechanical alloyed Co Se mixture
25
75
a,
C.E.M. Campos , J.C. de Lima , T.A. Grandi , K.D. Machado , V. Drago , P.S. Pizani
a
a
a
a
b
*
a
Departamento de F ´ı sica, Universidade Federal de Santa Catarina, C.P. 476, 88 040-900 Florian o´ polis, SC, Brazil
b
Departamento de F ´ı sica, Universidade Federal de S a˜ o Carlos, C.P. 676, 13565-905 S a˜ o Carlos, SP, Brazil
Received 15 October 2003; received in revised form 23 February 2004; accepted 23 March 2004 by B. Jusserand
Available online 12 April 2004
Abstract
A hexagonal CoSe alloy with NiAs-type structure was obtained by mechanical alloying starting from a mixture of pure
crystalline powders with nominal composition Co75Se25. X-ray diffraction (XRD), differential scanning calorimetry (DSC),
M o¨ ssbauer spectroscopy (MS) and Raman scattering (RS) techniques were used to follow the structural, thermal, magnetic and
optical properties of the binary mixture as a function of milling time. XRD results show the formation of a nanometric
hexagonal CoSe phase between 3 and 70 h of milling coexisting with non-reacted Co phases, also in nanometric scale. DSC and
RS results showed some changes in the thermal and optical properties of the crystalline phases when the milling time increases.
The Raman active modes of the CoSe and Co oxide phases were observed. MS results showed practically no iron in the samples
milled up to 15 h, while for extended milling times (70 h), they showed the presence of some a-Fe and the formation of other
iron alloys due to the contamination by the milling media.
q 2004 Elsevier Ltd. All rights reserved.
PACS: 61.46. þ w; 61.82. 2 d; 61.10. 2 i; 65.50. þ m; 63.20. 2 e
Keywords: A. Mechanical alloying; A. Nanocrystalline materials; C. X-ray diffraction; C. Differential scanning calorimetry; E. Raman
spectroscopy
1
. Introduction
Starting from a mixture with Fe25Se75 nominal composition,
depending on the milling time, orthorhombic FeSe , hexagonal
NiAs-type Fe Se and FeSe structures were obtained [11,12].
Thermal treatments at 200 and 400 8C transform Fe Se , which
is a ferrimagnetic superstructure, to other Fe–Se phases,
including FeSe with NiAs- and PbO-structures [12]. Starting
from a mixture with nominal composition Fe75Se25, only the
hexagonal FeSe phase was obtained [13].
2
Transition-metal–chalcogenide alloys show significant
7
8
changes in their physical properties depending on the specific
stoichiometry. Moreover, in many cases, structural or magnetic
phase transitions occur upon varying the temperature, pressure
or composition (non-stoichiometric samples). Very interesting
7
8
examples are the Fe12xSe
x
alloys. Together with several closely
Usually, transition-metal–selenides are produced by
normal solid-state reaction [6–8], selenization of aqueous
solutions containing transition-metal cations [9], or mol-
ecular precursors [10]. In the special case of cobalt
selenides, according to the Co–Se phase diagram [14],
there are two homogeneous and stable phases at room
related transition-metal–chalcogenide systems (V12xSe
x
,
Co12xSex and Ni12xSe ), Fe12xSex has been extensively
x
investigated by many groups [1–10]. Recently, we produced
some of these alloys via mechanical alloying (MA) and their
structural, thermal, magnetic and optical properties were
investigated by several experimental techniques [11–13].
temperature (CoSe
compositions (Co Se
preparation has not been proved yet. Recently, Co
containing CoSe phase in a minority quantity were
produced by Kazuhiro et al. [15] and Song et al. [16].
Most of the samples of CoSe have a pyrite structure with
2
and CoSe) and two other possible
and Co Se ), whose homogeneity of
Se films
3
4
2
3
*
Corresponding author. Address: Laboratoire de Physique des
Milieux Condens e´ s, Universit e´ Pierre et Marie Curie, Campus
3
4
Boucicaut, 75015 Paris, France. Tel.: þ55-48-331-6832; fax: þ55-
2
48-331-9068.
E-mail address: pcemc@fisica.ufsc.br (C.E.M. Campos).
2
0038-1098/$ - see front matter q 2004 Elsevier Ltd. All rights reserved.
doi:10.1016/j.ssc.2004.03.044

266
C.E.M. Campos et al. / Solid State Communications 131 (2004) 265–270
cubic symmetry. Ramdohr and Schmidt [17] discovered the
marcasite structure of CoSe with orthorhombic symmetry
in a mineral in 1955. Since then, few methods have been
particle size ,10 mm) and Se (99.999% purity, particle size
,150 mm) in a Co75Se25 composition was milled at room
temperature in a Spex 8000 shaker mill. Several steel balls
with mean diameter of 10 mm were put together with the
powdermixture ina steel containertoperformthe milling. The
ball to powder weight ratio used was 10:1. The container was
always closed inside a glove box withargon gasbeforestarting
the milling. After selected times (3, 10, 15, 30, 50 and 70 h),
the milling was stopped to collect small amounts of the
mixture. In order to determine possible contamination by
milling media in extended milling times, the samples milled
for 50 and 70 h were analyzed by energy dispersive
spectrometry (EDS). The EDS results for the sample milled
for 50 h were: 56.2 at.% Co, 20.5 at.% Se, 15.0 at.% O and
8.3 at.% Fe, while for the sample milled for 70 h, they were:
48.9 at.% Co, 17.0 at.% Se, 18.5 at.% O and 15.6 at.% Fe. It is
interesting to note that, although the contamination increased,
the Co/Se ratio remained practically unchanged. Due to the
high Fe content, M o¨ ssbauer measurements were performed at
room temperature using an electromechanical transducer
driven in the constant-acceleration mode, in conjunction
witha 512-multichanelanalyzer. The gamma-raysourcewasa
2
reported for the preparation of CoSe with this structure, as,
2
for instance, the hydrothermal technique [18]. Recently, we
2
showed that the orthorhombic CoSe , together with the
cubic one (minority) can be produced by MA after a few
ꢀ
hours of milling [19]. The cubic structure ða ¼ 5:857 AÞ of
2 2
CoSe compound is isotypic with the FeS (C2) structure.
The lattice parameters of the orthorhombic structure of
ꢀ
ꢀ
ꢀ
CoSe2 are a ¼ 3:600 A; b ¼ 4:840 A; c ¼ 5:720 A: The
ꢀ
ꢀ
hexagonal CoSe structure (a ¼ 3:621 A; c ¼ 5:289 A) is
isotypic with the NiAs (B8) and undergoes a polymorphic
transformation at about 892 8C. At small concentrations of
Se (1.6–2.8 wt%), the phase diagram also displays a
reversible 1 $ a transformation of Co. The 1 ! a trans-
formation is raised from 467 to 520 8C by 2.8 wt% Se and
the a ! 1 transformation lowered from 426 to 400 8C by
1
.6 wt% Se, indicating that these amounts are soluble at
about 400 and 520 8C, respectively. The (cubic) a-Co phase
is a high temperature one and during the Co purification
process, small amounts can be retained at room temperature.
MA is a dry milling process [20] in which a mixture of
powdersisactivelydeformedinacontrolledatmosphere,under
a highly energetic ball charge, to produce nanostructured
alloys, amorphous materials and unstable and metastable
phases[11,19,21–26]. The difficulty in preparing binary alloys
of constituents with a large difference in their melting points
and vapor pressures by usual melting techniques, as it happens
to alloys of cobalt and selenium, can be overcome using MA.
However, due to the low temperatures usually reached in MA,
the final products show poor crystallization. For poorly
crystallized materials, several difficulties arise in the interpret-
ation of the X-ray diffraction (XRD) patterns due to the peak
broadening caused by two main facts: (i) the final product
containslatticedefects and strains;(ii)the meancrystallitesize
5
7
Web Research 30 mCi Co in a Rh matrix. The XRD
measurements were performed at room temperature, in a
Miniflex Rigaku diffractometer using Cu K
radiation. DSC
a
measurements were done with N gas flux in a TA2010 DSC
2
cell and using a heating rate of 10 8C/min. RS measurements
were performed at room temperature using a T64000 Jobin-
Yvon triple-monochromator coupled to an optical micro-
scopy, a CCD detector and a conventional photon counting
˚
system. The5145 A line (,5 mW) of argonion laser was used
as excitation light always in the backscattering geometry.
3
. Results and discussion
(
some tens of nanometers). MA has several intrinsic
3
.1. XRD measurements
advantages, such as relatively inexpensive equipment, and
the possibility of scaling up. A disadvantage of MA is that the
physical mechanisms involved are not yet fully understood,
and a better understanding of these mechanisms is desirable.
Inthispaper, results onthe hexagonalCoSe alloy produced
by MA starting from a Co-rich mixture are reported. The
choice of the Co Se composition was made to verify the
Fig. 1 shows the XRD patterns of the Co Se mixture
7
5
25
as a function of milling time. The pattern of the starting
mixture (0 h) displays the peaks of trigonal Se phase
(
(
marked with triangles) as well as those of both hexagonal
diamonds) and cubic (squares) Co phases. It is interesting
75
25
to note that this XRD pattern is very noisy, and the Co peaks
intensities are smaller than those of the Se that is in minority
amount. This fact can be explained with a possible oxidation
of the very small Co particles. This suggestion will be
further discussed in Section 3.4. After 3 h of milling, the
XRD pattern shows no Se peaks, and the Co ones were
broadened and their intensities were drastically reduced.
This pattern also shows the presence of a new phase, whose
peaks are marked with opened circles. This new phase was
indexed to the CoSe phase, JCPDS card number 15-464 [27],
which has a hexagonal structure, with lattice parameters a ¼
possibility to form Co-rich alloys by MA, once that this
technique allows to extend the solubility range in several
alloys system [20]. XRD, differential scanning calorimetry
(
(
DSC), M o¨ ssbauer spectroscopy (MS) and Raman scattering
RS) techniques were used to follow the evolution of the
physical and chemical properties as a function of milling time.
2
. Experimental procedure
ꢀ
ꢀ
The starting mixture of elemental Co (99.7% purity,
3:6315 A and c ¼ 5:3011 A and space group P63=mmc: For

C.E.M. Campos et al. / Solid State Communications 131 (2004) 265–270
267
The initial CoSe phase content was 56% (3 h) of the
crystalline phases, after 15 h of milling it was 37% and for
longer milling times it increased to about 60%. In the initial
mixture (0 h) the cubic Co phase content represents 6% of
the crystalline phases, after 15 h of milling it reaches a
maximum value of 40%, while the hexagonal one, that
represents 60% of the crystalline phases at 0 h, goes to 23%.
For milling times longer than 15 h, an increase in the
hexagonal Co content up to 60% and a drastic reduction of
the cubic one is observed, since only small amounts of it
(,2%) are found at 50 and 70 h of milling. Huang et al. [31]
reported results about reversible 1 $ a transformations of
Co induced by ball milling. They observed that for low- and
intermediate-milling intensity conditions (using a planetary
ball mill) the hexagonal is the majority phase. This result
was also observed in the present paper. The structural
change in the non-reacted Co after 15 h of milling, which is
predominantly cubic, can be associated with the reduction of
hexagonal CoSe phase at this milling time.
The lattice parameters ratio c=a and density values for the
CoSe phase were calculated as a function of milling time.
The c=a values obtained were: 1.469 (3 h), 1.475 (15 h),
1
.448 (50 h), and 1.470 (70 h). It is interesting to observe
that the c=a ratio suffer a drastic reduction after 50 h of
milling. This fact can be associated with excessive stress
over the CoSe phase at this milling time, probably due to the
increasing of contaminants contents, as suggested by the
EDS analysis. The density showed a reduction approxi-
mately linear with increasing milling time, starting from
Fig. 1. XRD patterns of the Co75Se25 mixture as a function of
milling time.
longer milling times (10– 70 h), all XRD patterns are similar
in shape, indicating that no new phases were nucleated.
According to EDS analysis results, there is sufficient iron in
those samples (at least, that milled for 50 and 70 h) to be
detectedbyXRDanalysis,butsincetheFepatternhasthemost
intensepeaks inthe same2u region asthe CoSeand Cophases,
3
3
7
.8 g/cm (3 h) and reaching 7.6 g/cm (70 h). The last value
3
is very close to that given in the JCPDS card (7.56 g/cm ).
(
see crosses in Fig. 1) its identification is a difficult task. This
suggestion will be further discussed in Section 3.2. The XRD
patterns alsoshowthatthemillingprocess causesabroadening
and reduction inthe intensities ofthepeaks, suggesting thatthe
mean crystallite size of the crystallites is reduced with
increasing milling time.
3.2. MS measurements
Fig. 2 shows the M o¨ ssbauer spectra obtained for the
Co75Se25 samples milled for 3, 15 and 70 h. The first two
spectra (3 and 15 h) showed basically residual or even no
traces of iron, while the third one showed one magnetically
ordered sextet and at least two paramagnetic quadrupole
doublets. The sextet parameters are typical of a-Fe phase
and it is responsible for 30% of the total absorption area of
the spectrum. The hyperfine parameters of the doublets
clearly indicate that they correspond to high-spin iron (II)-
and iron (III)-containing complexes resulting from a
reaction with the non-reacted Co, O, and maybe Se. It is
difficult to be more specific about the exact nature of these
paramagnetic components, but it is clear that the hyperfine
parameters neither correspond exactly to Fe–Co oxide films
In order to confirm the reduction in the crystallite size, all
the XRD patterns were simulated using the Rietveld structure
refinement procedure [28]. The structural data of the crystal-
line phases used in the simulations were obtained from TAPP
software [29]. From the simulations lattice parameters,
densities and full widths at half maximum (FWHM) of the
peaksofthephasesweredetermined. Althoughthepatternsare
very noisy, the mostintense peaks located atsmall angles were
sufficiently defined to confirm numerically the intensity
reduction and broadening of all peaks as well as to show
some changes in the relative amount of crystalline phases with
increasing milling time. The mean crystallite size was
obtained using the FWHM and u position of the peaks and
Scherrer formula [30]. The obtained values after 3 h of milling
were 7 nm for CoSe, 8 nm for cubic Co and 5 nm for
hexagonal Co. For longer milling times there are only small
fluctuations around these values.
[
2 4 2 2
32], CoFe O [33,34], Co Fe and CoSe [35] compounds,
nor to amorphous Fe–Se–O [36]. This result agrees with the
EDS analysis of this sample (milled for 70 h) and also
showed that the contamination is relevant just for extended
milling times.

268
C.E.M. Campos et al. / Solid State Communications 131 (2004) 265–270
Fig. 2. M o¨ ssbauer spectra for Co75Se25 samples as a function of
milling time.
Fig. 3. DSCcurvesoftheCo75Se25 mixture as a function of milling time.
3
.3. DSC measurements
and its enthalpy variation increased approximately 3 J/g. It is
interesting to note that the CoSe content showed a minimum
value at this milling time, suggesting an increase in defect
centers. The maximum intensity of the second exothermic
band was shifted to lower temperatures (334 8C).
Fig. 3 shows the DSC curves of the Co Se mixture as
7
5
25
a function of milling time. The initial mixture curve shows
the endothermic peak at 217 8C of the fusion of trigonal
Se and several exothermic features that take place for
temperatures above 400 8C. According to the Co–Se phase
diagram [14], all these high temperature features are located
in the temperature range corresponding to the reversible
For longer milling times (50 and 70 h), important changes
occurred in the DSC curves. However, it is possible to see at
least two features, which can be associated with the effects
already described. The first exothermic peak was now shifted
to low temperatures and its enthalpy variation changed,
reaching the highest value at 50 h of milling (18 J/g). After
70 h ofmilling, the enthalpy variation of thispeakwas reduced
totheminimumvalue(8 J/g, respectively). Thesefactssuggest
the reduction of the defect centers concentration from 50 to
70 h of milling. It is interesting to note that for the sample
milled for 70 h, according to the M o¨ ssbauer and EDS
measurements, an high a-Fe content and other two iron alloys
were observed. The second exothermic band, located between
250 and 600 8C, shows features similar to those observed for
15 h of milling. The small changes observed are associated
with the slight changes in the chemicalenvironments of the Co
phases due to the milling process.
1
$ a transformation of Co (300–600 8C). Based on this
fact and considering the large amount of elemental Co used
in the initial mixture, these exothermic features can be
attributed to1 ! a transformation and/or structuralrelaxation
of Co. The curve corresponding to 3 h of milling displays
two broad exothermic bands, whose maximum intensities
are located around 170 and 390 8C, and one endothermic
peak at 525 8C. The first asymmetric exothermic band is
associated with partial annealing of the defect centers and/or
structural relaxation in the crystalline components. The
second exothermic band coexists with two features: one
exothermic, located between 340 and 425 8C, and another
endothermic, located at 525 8C. Since the second band is in
the same temperature range of the exothermic reactions
observed for the elemental Co in the initial mixture, this
band was attributed to the same effects already described.
After 15 h of milling the DSC curve shows features
similar to those already described, except the absence of the
endothermic peak. Until now, we have any reasonable
explanation neither for the origin nor for the disappearance
of the endothermic peak. The maximum intensity of the first
exothermic band was shifted to higher temperature (210 8C)
3.4. RS measurements
Fig. 4 shows the Raman spectra of pure elemental powders
(c-Se and c-Co) and Co Se mixture as a function of milling
7
5
25
time. Table 1 shows all the Raman parameters obtained from
the fitting of these spectra by using Lorentzian functions. The
c-Se spectrum shows three lines located at 140, 234 and
2
1
237 cm , which are associated with the characteristic

C.E.M. Campos et al. / Solid State Communications 131 (2004) 265–270
269
vibrations of the Se chains [37,38]. The c-Co spectrum shows
n
21
a very small line at 190 cm , a broad band that is composed
2
1
by at least two components (centered at 485 and 525 cm
)
and an intense Raman line at the high frequency range (with a
21
maximum at 690 cm ). The broad band of this spectrum is a
strong evidence of the contamination of the c-Co sample, and
probably is associated with O–O bond vibrations in a very
disordered system [39]. However, there are huge experimental
limitations to store Raman lines of metals like Co, Fe, Ni,
which sometimes (depending of their structural configur-
ations) even do not show Raman active modes. Then, the
Raman lines observed for the c-Co sample were identified as
3 4
Co–ObondvibrationsofamixedCooxide(CoO·Co O )[39],
suggesting a Co-oxide coating of the Co particles in the
starting mixture, in good agreement with the weak XRD
pattern collected for this sample (see Fig. 1 in Section 3.1).
Considering this fact, EDS analysis of the elemental Co
powder used in the initial mixture Co75Se25 was performed.
Theresultsshowedapproximately7.0 at.%O.Thepeaksofthe
XRD patterns of both CoO and Co O are shown in Fig. 1
3
4
(
starsandbars, respectively), and onlythe mostintensepeakof
the CoO can be observed in the pattern of the samples milled
for 50 and 70 h, which made difficult their inclusion in the
XRD simulations. Although ultra pure elements were used in
this study, the oxidation of very fine particles is well reported
in the literature [40,41].
Fig. 4. Raman spectra for pure crystalline elements (c-Co and c-Se)
and for Co75Se25 mixture as a function of milling time.
After the milling process, the Raman spectra become
Table 1
2
Raman frequency v (cm ), linewidth G (cm ) and relative intensity ratio obtained by using a set of Lorentzian functions
1
21
Sample
Se
n
3 4
CoO·Co O
0
00
E
E
A1
1
2
3
4
5
c-Se
c-Co
v
140.6
9.1
233.8
7.4
237.5
5.6
–
–
–
–
–
–
–
–
–
–
–
–
–
–
–
G
I=IA1
0.26
1.30
1.00
v
–
–
–
–
–
–
–
–
–
190.7
9.5
483.7
9.5
526.9
16.3
678.0
23.7
691.5
12.2
G
I=I5
0.03
0.10
0.05
0.23
1.00
Co75Se25
CoSe
3
1
5
7
h
v
173.5
12.1
194.6
6.1
476.0
12.9
517.6
15.8
612.2
6.1
681.1
14.5
G
I=I4
0.07
0.17
0.21
0.19
0.03
1.00
5 h
0 h
0 h
v
168.6
6.5
190.4
11.7
0.12
474.3
13.0
516.4
15.8
616.2
8.4
680.0
16.3
G
I=I5
0.05
0.18
0.16
0.05
1.00
v
173.5
12.5
193.6
11.2
0.15
475.6
14.2
518.5
23.8
611.3
6.0
681.0
17.0
G
I=I5
0.15
0.18
0.22
0.05
1.00
v
174.9
15.5
194.9
6.0
475.3
13.1
517.0
14.1
611.6
6.7
680.2
15.6
G
I=I4
0.08
0.18
0.19
0.17
0.04
1.00

270
C.E.M. Campos et al. / Solid State Communications 131 (2004) 265–270
very noisy and the c-Se features completely disappeared,
indicating the CoSe nucleation, which is in agreement with
the XRD and DSC results. The Raman lines observed in the
with milling time increasing, indicate drastic changes in the
chemical environment of the nanometric particles of CoSe.
3
approximately 10 cm , but the broad band is absent and a
h spectrum seems to be the Co-oxide ones, downshifted by
2
1
Acknowledgements
2
1
new Raman line was observed at about 173 cm
.
We thank the Brazilian agencies CAPES, FINEP,
FAPESP and CNPq for the financial support. We are also
indebted with J.C. Chervin and B. Canny from the
Laboratoire de Physique des Milieux Condens e´ s due to
technical support in some Raman measurements.
According to our previous study [19], there is a strong
Raman line associated with the CoSe phase at about
2
66 cm . Then, although there is a frequency up shift,
1
1
the new Raman line can be attributed to CoSe phase bond
vibrations. These Raman shifts indicate that the Co oxide
and CoSe particles are under high-stressed conditions.
At 15 h of milling, the same features observed at 3 h
were detected. However, now the linewidths of Co-oxide
lines increased (see Table 1), making certain interference in
References
[1] S. Stolen, et al. Phys. Solids, 48 (1087) 1213.
[2] K. Aggarwal, et al., Physica 90B (1977) 269.
[3] M. Rea, et al., Solid State Commun. 20 (1976) 325.
2
1
the fitting of the line located at 170 cm , which seems to
enlarge and to suffer a frequency downshift. Then, at least
one conclusion can be made; the chemical environment of
the CoSe phase is modifing as the milling time proceeds.
The Raman spectrum of the samples milled for 50 h
showed basically the same features found for the smaller
milling times and no other new Raman lines, such as the Fe–
Se bond vibration modes [11,13,19], were observed. The
frequency shifts observed for this milling time is attributed to
changes in the stress conditions imposed to the Co-oxide and
CoSe particles in the powder mixture. Moreover, the changes
in the stress conditions imposed to the CoSe phase are in good
agreement with the c=a ratio reduction, showed by the XRD
pattern simulation. All these structural changes in the powder
mixture are attributed to the increase in the contamination
contents, as showed by the EDS analyses.
[
[
[
[
[
[
4] Y. Nishihara, et al., J. Chem. Phys. 71 (1979) 3796.
5] M. Knecht, et al., J. Alloys Compd. 246 (1997) 166.
6] P. Terzieff, Physica 103B (1981) 158.
7] K.V. Reddy, et al., Phys. Stat. Sol. (a) (1975) 32.
8] A. Goswami, et al., J. Cryst. Growth 8 (1971) 247.
9] N. Hamdadou, et al., J. Cryst. Growth 241 (2002) 313.
[
[
10] S. Stuczynski, et al., J. Organomet. Chem. 449 (1993) 301.
11] C.E.M. Campos, et al., Solid State Commun. 123 (2002) 179.
[12] C.E.M. Campos, et al., J. Magn. Magn. Mater. 269 (2004) 6.
[13] C.E.M. Campos, et al., J. Magn. Magn. Mater. 270 (2004) 89.
[14] M. Hansen, Constitution of Binary Alloys, Geminuim Publ.
Co, New York, 1985, pp. 502–503.
[
[
[
[
[
[
15] T. Kazuhiro, et al., Bull. Chem. Soc. Jpn 65 (1992) 329.
16] X. Song, et al., J. Chem. Soc., Dalton Trans. (1997) 2689.
17] P. Ranmdohr, et al., Neues Jahrb Mineral Monatsh 6 (1955) 133.
18] W. Zhang, et al., Mater. Res. Bull. 35 (2000) 2403.
19] C.E.M. Campos, et al., Physica B 324 (2002) 409.
20] C. Suryanarayana, Prog. Mater. Sci. 46 (2001) 1.
The CoSe Raman line of the sample milled for 70 h is
worst resolved than in the previous spectra, probably due to
the high a-Fe content, as well as, due to the formation of at
least other two iron alloys in this sample, as showed by the
M o¨ ssbauer results (see Fig. 2). This features, confirming the
drastic changes in chemical environment of the CoSe phase
with milling time increasing.
[21] M. Abbate, et al., J. Phys.: Condens. Matter 13 (2001) 5723.
[22] J.C. de Lima, et al., J. Alloys Compd. 234 (1996) 43.
[23] A.W. Weeber, et al., Physica B 153 (1988) 93.
[
[
[
[
24] D.K. Mukhopadhyay, et al., Scr. Metall. Mater. 30 (1994) 133.
25] A.R. Yavari, et al., Phys. Rev. Lett. 68 (1992) 2235.
26] J.C. de Lima, et al., J. Non-Cryst. Solids 318 (2003) 121.
27] JCPDS—Powder Diffraction File Search Manual, International
Centre for Diffraction Data, Pennsylvania, USA, 1981.
28] H.M. Rietveld, J. Appl. Crystallogr. 2 (1969) 65.
29] TAPP, version 2.2, E. S. Microware, Inc., 2234 Wade Court,
Hamilton, OH 45013.
4
. Conclusions
[
[
The nanocrystalline CoSe phase was successfully
obtained between 3 and 70 h of milling;
[
30] Z. Strnad, Glass–Ceramic Materials, Elsevier, Amsterdam,
The nanometric non-reacted Co suffered structural
changes (from hcp to fcc), as milling time reached 15 h,
and after 50 h, the remaining non-reacted Co was reverted to
hcp phase;
1
986, p. 161.
[31] J.Y. Hunag, et al., Appl. Phys. Lett. 66 (1995) 308.
[32] S. Nasu, et al., IEEE Trans. Magn. 23 (1987) 2257.
[
[
[
[
[
33] Y. Ahn, et al., Mater. Lett. A50 (2001) 47.
34] V. Blaskov, et al., J. Magn. Magn. Mater. 162 (1996) 331.
35] V.H. McCann, et al., J. Phys. Chem. Solids 38 (1977) 991.
36] P.S. Thind, et al., Polyhedron 5 (1986) 1629.
37] J.B. Renucci, PhD Thesis, Universite Paul Sabatier, Toulouse,
France, 1974.
The M o¨ ssbauer results showed only iron traces in the
samples milled up to 15 h, while for extended milling times
(70 h) they showed the presence of some a-Fe and the
formation of other iron alloys;
The Raman results showed very intense Raman lines of the
CoO·Co O oxide for all milled samples and even for the
[
38] F.Q. Guo, et al., Phys. Rev. B 57 (1998) 10414.
3
4
[39] L.C. Shcumacher, et al., Electrochem. Acta 35 (1990) 975.
[40] O. Bockman, et al., Hidrometallurgy 55 (2000) 93.
elemental Co used in the start mixture. The Raman line of
CoSe phase was detected, and the changes in its parameters
[41] J. Xu, et al., J. Phys. B.: At. Mol. Opt. Phys. 29 (1996) 6227.