The X-ray diffraction patterns of bimetallic FexM100-x (M = Mo and W) nanoclusters

Article abstract of DOI:10.1021/jp003128b

We produced the bimetallic Fe_xMO_100-x and Fe_xW_100-x nanoclusters with cluster size of 1.5-4.5 nm by thermally decomposing a mixture of two metal carbonyl vapor species and recorded the X-ray diffraction (XRD) patterns of them for the mole percent (x)range of 0 a?? x a?? 100. The XRD patterns changed from that of pure M (M = Mo and W) nanoclusters to that of pure Fe nanoclusters as x increased. The solubilities of bimetallic nanoclusters were 2-3 times higher than those of the corresponding bulk alloys, respectively. We observed that the cluster size decreased whereas the atomic level strain increased as the degree of alloying increased. We also observed that the cluster size effect on the broadening of the full width at half-maximum (fwhm) of the XRD peaks was stronger than the strain effect. Both cluster size decrease and atomic level strain increase of Fe_xW_100-x nanoclusters were larger than those of Fe_xMo_100-x nanoclusters.

Full text of DOI:10.1021/jp003128b

5856
J. Phys. Chem. B 2001, 105, 5856-5861
The X-ray Diffraction Patterns of Bimetallic FexM100-x (M ) Mo and W) Nanoclusters
G. H. Lee,* S. H. Huh, J. W. Park, and H. K. Kim
Department of Chemistry, College of Natural Sciences, Kyungpook National UniVersity,
Taegu 702-701, South Korea
ReceiVed: August 30, 2000; In Final Form: December 24, 2000
x x
We produced the bimetallic Fe Mo100-x and Fe W100-x nanoclusters with cluster size of 1.5-4.5 nm by thermally
decomposing a mixture of two metal carbonyl vapor species and recorded the X-ray diffraction (XRD) patterns
of them for the mole percent (x) range of 0 e x e 100. The XRD patterns changed from that of pure M (M
)
Mo and W) nanoclusters to that of pure Fe nanoclusters as x increased. The solubilities of bimetallic
nanoclusters were 2-3 times higher than those of the corresponding bulk alloys, respectively. We observed
that the cluster size decreased whereas the atomic level strain increased as the degree of alloying increased.
We also observed that the cluster size effect on the broadening of the full width at half-maximum (fwhm) of
the XRD peaks was stronger than the strain effect. Both cluster size decrease and atomic level strain increase
x x
of Fe W100-x nanoclusters were larger than those of Fe Mo100-x nanoclusters.
Introduction
bimetallic nanoclusters and alloys but higher than those of the
corresponding equilibrium phase bulk alloys, respectively.
Alloy nanoclusters have a variety of physical and chemical
properties that are different from those of bulk alloys. Their
properties also depend on composition and cluster size. Thus,
alloy nanoclusters are new materials that can be used in a variety
of areas such as catalysts and electric and magnetic materials,
etc.
Experimental Section
(
i) Production. The thermal decomposition method has been
2
5
described previously, and only details pertinent to the present
study are described here. The thermal decomposition method
combines thermal decomposition of a gas mixture of two metal
carbonyl vapor species Fe(CO)5 and M(CO)6 (M ) Mo and
W) by using a hot filament and collision-induced clustering
between the decomposed neutral Fe and M metal atoms.
Nichrome wire with 1 mm thickness was used as a filament.
But any kind of resistive heater can be used as a filament. The
black and bimetallic nanoclusters were produced as the final
products. We noticed that the COs produced during the metal
carbonyl decomposition process did not degrade the purity of
the bimetallic nanoclusters and were evacuated after the
experiment. Typical running conditions were as follows: fila-
ment voltages were 15-25 V, which corresponded to the
Bimetallic nanoclusters have been produced by mechanical
1-3
4
5,6
attrition, sputtering, reduction of cations in solution, ultra-
7
8
violet (UV) photolysis, inert gas condensation, and chemical
vapor deposition, etc. Nonequilibrium phase bimetallic nano-
9
clusters were often produced by using the first two methods,
whereas equilibrium phase bimetallic nanoclusters were pro-
duced by using the latter three methods. The first is a well-
known method by which a variety of nonequilibrium phase alloy
nanoclusters can be produced in any composition, and thus has
served as a general method of producing a variety of bimetallic
nanoclusters. Studies on bimetallic nanoclusters produced by
mechanical attrition include characterizations of the structure
26
by using X-ray diffraction (XRD) spectrometery,1
properties by using differential scanning calorimeter (DSC),
magnetic properties by using superconducting quantum interfer-
0-17
filament temperature of 350-450 °C; the metal carbonyl vapor
pressures were 10-100 Torr, the diluent Ar pressure was 100-
thermal
1
4-18
3
00 Torr, and the reaction chamber temperature was ∼150 °C
1
9-23
in order to vaporize the solid sample M(CO)6. At the above
filament temperatures, bimetallic nanoclusters with purity of
greater than 95% were obtained. After the experiment, the
reaction chamber was exposed to air in order to collect the
nanoclusters. Bimetallic nanoclusters were, however, immedi-
ately kept in a vacuum in order to minimize oxidation and
moisture absorption. In fact, the X-ray photoelectron spectros-
copy (XPS) showed some surface oxidation in Fe, Mo, and W
nanoclusters, especially for Fe nanoclusters when exposed to
air for approximately an hour. To control the value of the Fe
mole percent x (0 e x e 100) in FexM100-x nanoclusters, various
different ratios in the vapor pressure of Fe(CO)5 to that of
M(CO)6 were used and the final exact value of the mole percent
x was determined by using the elemental analysis with an
inductively coupled plasma atomic emission spectrometer.
ence device (SQUID) and M o¨ ssbauer spectrometery,
and
size distribution by using transmission electron microscopy
TEM).16, 24
(
Although there exist many studies on nonequilibrium phase
bimetallic nanoclusters produced by mechanical attrition, there
are only a few studies on equilibrium phase bimetallic nano-
clusters. In this work, we produced the equilibrium phase
FexMo100-x and FexW100-x nanoclusters (0 e x e 100) by
thermally decomposing two metal carbonyl vapor species and
investigated structures by recording the XRD patterns. We
discussed the solubility, the cluster size decrease, and the atomic
level strain increase due to alloying, and the differences between
FexMo100-x and FexW100-x nanoclusters. We found that the
solubilities of FexMo100-x and FexW100-x nanoclusters were
lower than those of the corresponding nonequilibrium phase
(ii) X-ray Diffraction. The Philips X-ray diffraction spec-
trometer (X-PERT model) was used. The X-ray diffraction
*
To whom correspondence should be addressed. E-mail: ghlee@
bh.knu.ac.kr. Fax: 82-53-950-6330.
(XRD) patterns were recorded between 2θ ) 20° and 80°. The
1
0.1021/jp003128b CCC: $20.00 © 2001 American Chemical Society
Published on Web 06/05/2001

XRD Patterns of Bimetallic FexM100-x Nanoclusters
J. Phys. Chem. B, Vol. 105, No. 25, 2001 5857
Figure 2. XRD patterns of bimetallic Fe
indicates the mole percent of Fe atoms in bimetallic nanoclusters. The
hkl) assigned on the top of each peak indicates the Miller indices. O
and 4 indicate the peaks that originate from the bimetallic fcc and bcc
Fe 100-x nanoclusters, respectively. b indicates the amorphous peaks
x
W100-x nanoclusters where x
x
Figure 1. XRD patterns of bimetallic Fe Mo100-x nanoclusters where
x indicates the mole percent of Fe atoms in bimetallic nanoclusters.
The (hkl) assigned on the top of each peak indicates the Miller indices.
O and 4 indicate the peaks that originate from the bimetallic fcc and
(
x
W
x
bcc Fe Mo100-x nanoclusters, respectively.
that originate from W metal in bimetallic fcc Fe
clusters.
x
W100-x nano-
step size and the sitting time at each point were 0.04° and 0.5
s, respectively. Both 1.54056 Å and 1.54439 Å of the Cu KR
radiation with intensity ratio of 2 to 1 were used without filtering
in order to increase the signal-to-noise ratio. To improve the
signal-to-noise ratio, both the sample holder and the detector
were also rotated simultaneously. The nanoclusters were
squeezed onto the glass substrate with a spot diameter of ∼1
cm. The weak and broad peak around 2θ ) 25° in some XRD
patterns in Figures1 and 2 was due to the glass substrate. In
general, the peak intensity of the bimetallic nanoclusters was
low because only a small amount was produced after each run
with the present small reaction chamber and because the
bimetallic nanoclusters could not be added together after several
runs because the same composition could not be obtained after
each run, whereas the pure nanoclusters could be added together
after several runs to improve the signal-to-noise ratio. Of course,
the signal-to-noise ratio could be improved by using a longer
sitting time for scan than the present 0.5 s.
function. Since the peak intensity was generally low, except
for that of the pure Fe and M (M ) Mo and W) nanoclusters,
we used the peak center of the most intense peak, i.e., the (111)
peak for the fcc structure and the (110) peak for the bcc structure
when we calculated the lattice constant. Due to the overlap of
the bcc (110) peak with the fcc (111) peak, a somewhat large
error in determining the peak position of the bcc (110) peak
and, thus, a somewhat larger error in the lattice constant of the
bcc structure than that of the fcc structure occurred accordingly,
especially for the peaks with x near to 50.
Following are the special features observed in the XRD
patterns, which will be discussed in this paper. First, the full
width at half-maximum (fwhm) of the fcc peaks (i.e., the peaks
from Fe atoms dissolved M nanoclusters) increased noticeably
as x increased, whereas those of the bcc peaks (i.e., the peaks
from M atoms dissolved Fe nanoclusters) increased only slightly
as x decreased. Also, note that the fwhm of the fcc FexW100-x
nanoclusters were broader than those of the fcc FexMo100-x
nanoclusters. In the fcc FexW100-x nanoclusters, the XRD
patterns became amorphous at x g 12.8. The broadening of the
fwhm in the FexMo100-x and FexW100-x nanoclusters and the
amorphous structure in the fcc FexW100-x nanoclusters were due
to a small cluster size and a large strain, as will be discussed.30
(ii) Solubility. Figure 3 (a and b) represents the lattice
constants of the fcc FexMo100-x and the fcc FexW100-x nano-
clusters plotted as a function of the Fe mole percent x,
Results and Discussion
(i) XRD Pattern. Figures 1 and 2 represent the XRD patterns
of bimetallic FexMo100-x and FexW100-x nanoclusters, respec-
tively, as the Fe mole percent x increases. In general, the XRD
pattern of bimetallic nanoclusters changed from that of pure M
(M ) Mo and W) nanoclusters to that of pure Fe nanoclusters
as x increased. The XRD patterns of pure nanoclusters had been
discussed in the previous works. Briefly, Fe nanoclusters
consisted of the body centered-cubic (bcc) structure, whereas
respectively. As Figure 3 (a and b) shows, each plot deviates
from the corresponding Vegard’s law, respectively,1
7,31
which
both Mo and W nanoclusters consisted of the face centered-
cubic (fcc) structure.2
7-29
The assignment of the peaks of the
indicates that solubility of Fe atoms in both the fcc Mo and the
fcc W nanoclusters is very low. Figure 4 (a and b) represents
the lattice constants of the bcc FexMo100-x and the bcc FexW100-x
nanoclusters plotted as a function of the mole percent x,
respectively. As Figure 4 (a and b) shows, each plot also deviates
from the corresponding Vegard’s law, respectively, which
indicates that both solubilities of Mo and W atoms in the bcc
Fe nanoclusters are also very low.
pure nanoclusters is provided in the corresponding XRD patterns
as shown in Figures 1 and 2 and the assignment of the peaks of
the bimetallic nanoclusters follows that of the pure nanoclusters.
Here, since the fcc (200) peak occurs near to the bcc (110) peak,
an overlap of these two peaks occurred.
The peak center and the full width at half-maximum (fwhm)
of the peak were determined by fitting the peak to the Lorentzian

5858 J. Phys. Chem. B, Vol. 105, No. 25, 2001
Lee et al.
The solubility of the minor atoms at each composition is
obtained by reading the mole percent of that metal in the
abscissa which corresponds to the same lattice constant in the
Vegard’s law as that of the bimetallic nanoclusters. We took
the average solubilities for the observed cluster size range
because the solubilities were not much different from one
another. As provided in Table 1, solubilities of both FexMo100-x
and FexW100-x nanoclusters were generally low but were 2-3
times higher than those of the corresponding bulk alloys,3
respectively. These low solubilities in fact can be noticed from
the almost zero enthalpies of mixing of the corresponding bulk
4-36
3
7-39
alloys.
These low solubilities reconfirm that the produced
FexMo100-x and FexW100-x nanoclusters were the equilibrium
phase bimetallic nanoclusters.
We compared the solubilities of the present FexMo100-x and
FexW100-x nanoclusters with those of the nonequilibrium phase
bimetallic nanoclusters and alloys produced by other methods.
The solubilities of Fe-W nanoclusters produced by mechanical
10
attrition were 10-19 mole percent, which is much higher than
those of the present equilibrium phase FexW100-x nanoclusters.
In the case of FexMo100-x nanoclusters, the nonequilibrium phase
amorphous alloys produced by ion-induced amorphization of
the Fe-Mo multilayered films showed an enhanced solubility
Figure 3. Plot of the lattice constant of (a) bimetallic fcc Fe
x
Mo100-x
nanoclusters and (b) bimetallic fcc Fe 100-x nanoclusters as a function
x
W
of the Fe mole percent x. The fcc (111) peaks were used in calculating
the lattice constants. Large deviations from the corresponding Vegard’s
laws indicate low solubilities of Fe atoms in both Mo and W
nanoclusters, respectively. The errors in the lattice constants originate
from the errors in the peak centers.
40,41
of >10 mole percent.
Thus, the equilibrium phase FexMo100-x
and FexW100-x nanoclusters produced in this work have much
lower solubilities than the corresponding nonequilibrium phase
bimetallic nanoclusters and alloys, respectively, as expected.
Since the solubilities are very low, it is expected that both
FexMo100-x and FexW100-x nanoclusters are separated in phase
into Fe and M (M ) Mo and W) nanocluster regions as in bulk
alloys. Such a phase separation had been suggested in Au-Pd
nanoclusters produced by reduction method through extended
5
X-ray absorption fine structure (EXAFS) analysis and Fe-Cu
and Fe-Ag nanoclusters produced by inert gas condensation
8
method through M o¨ ssbauer spectroscopy. The latter work had
also suggested that the alloying occurred at the intermetallic
cluster boundary regions. Thus, we expect that a similar alloying
also occurred in the FexMo100-x and FexW100-x nanoclusters.
This explains why we observed both XRD peaks originating
from Fe and M nanoclusters for some x range near to 50 (see
Figures 1 and 2).
(iii) Cluster Size and Strain. The broad fwhm of the
bimetallic nanoclusters is originated from both a small cluster
size and a large strain. To extract each contribution from the
observed fwhm () (δs)o), we used the well-known Cauchy size
broadening and Gaussian strain broadening functions,1
1,14,30
Figure 4. The plot of the lattice constant of (a) bimetallic bcc
Fe
x
Mo100-x nanoclusters and (b) bimetallic bcc Fe
x
W
100-x nanoclusters
2
2
1
/(δs) ) D - 6.25D<e >[s/(δs) ]
o
o
as a function of the Fe mole percent x. The bcc (110) peaks were used
in calculating the lattice constants. Large deviations from the corre-
sponding Vegard’s laws indicate low solubilities of both Mo and W
atoms in Fe nanoclusters, respectively. The errors in the lattice constants
originate from the errors in the peak centers.
2
in which D is the cluster diameter, <e > is the atomic level
mean square strain () 0.64 (∆d/d) ) in which d and ∆d are the
lattice distance and the uncertainty in d, respectively, and s is
the reciprocal lattice distance, i.e., 2sin θ/λ in which θ and λ
are the diffraction angle and the X-ray wavelength () 0.1542
nm), respectively. After calibrating the observed (δs)o against
the fwhm of Fe nanoclusters in order to remove both instru-
2
A close look at the lattice constants showed that the lattice
constants slightly decreased as x increased in Figure 3 (a and
b) and slightly increased as x decreased in Figure 4 (a and b),
which indicates that the solubilities slightly increased as the
degree of alloying increased in all bimetallic nanoclusters. This
is due to the capillarity effects originating from a large surface
tension of the small nanoclusters.3 According to the capillarity
effects, the smaller the cluster size, the higher the solubility.
Note that we observed that the cluster size decreased as the
degree of alloying increased as will be discussed later, in
agreement with the observed slight solubility increase as the
degree of alloying increased.
42
mental and X-ray KR intensity broadenings, we plotted 1/(δs)o
2
2
as a function of [s/(δs)o] to extract D and <e > from the
intercept and the slope, respectively. Note that the above relation
requires several XRD peaks in each XRD pattern in order for
it to be used. Thus, only the fcc peaks were analyzed by using
the above relation and the fwhm values obtained from the
Lorentzian function fit of the peaks are provided in Table 2.
(iii-a) Cluster Size. Figure 5 shows the transmission electron
microscope (TEM) and high resolution transmission electron
2,33

XRD Patterns of Bimetallic FexM100-x Nanoclusters
J. Phys. Chem. B, Vol. 105, No. 25, 2001 5859
TABLE 1: Bulk Enthalpy of Mixing in kJ/mol, Bulk Solubility in mol %, and Solubility of Nanocluster in mol %
bulk solubility (600-700 °C)b nanocluster solubility (100-150 °C)c
solubility of
solubility of
M atoms in bcc
Fe nanocluster
solubility of
Fe atoms in
fcc M nanocluster
bimetallic
alloy
bulk enthalpy
of mixing
M atoms in R-Fe
M ) Mo, W
solubility of
Fe atoms in R-M
a
Fe-Mo
Fe-W
-1 ∼ -3
2.27
1.68
0.5
∼2.0
8.7 ( 3.5
6.8 ( 3.5
6.0 ( 3.5
4.7 ( 3.5
0
a
Ref 37. b Ref 34 and ref 35. c This work and the temperature corresponds to the reaction chamber temperature.
TABLE 2: Full Width at Half Maximums (fwhm) of the
Face Centered Cubic (fcc) Peaks Used in Calculating the
Cluster Diameters and the Atomic Level Strains
fwhm of the fcc peaks in 2θ degree
bimetallic
nanoclusters
x
(111)
(200)
(220)
(311)
Fe
x
Mo100-x
0
0.81 ( 0.02 0.85 ( 0.02 0.94 ( 0.03 1.02 ( 0.05
1
2
4
6.8 1.59 ( 0.02 1.78 ( 0.05 2.26 ( 0.10 2.51 ( 0.10
5.4 1.93 ( 0.10 2.08 ( 0.15 2.70 ( 0.20 3.15 ( 0.30
3.6 2.34 ( 0.20 2.58 ( 0.30
Fe
x
W
100-x
0
6
0.74 ( 0.02 0.78 ( 0.02 0.86 ( 0.03 0.94 ( 0.05
.9 2.5 ( 0.2 2.7 ( 0.3 3.5 ( 0.5 4.0 ( 1.0
Figure 6. Plot of the cluster diameter D of the bimetallic fcc Fe
and fcc Fe 100-x nanoclusters as a function of the Fe mole percent x.
This plot shows that D decreased as x increased.
x
Mo100-x
x
W
agreement with the amorphous XRD pattern. Figure 6 represents
a plot of the cluster diameter D of the fcc FexMo100-x and the
fcc FexW100-x nanoclusters as a function of the Fe mole percent
x. As Figure 6 shows, D decreased as x increased in both fcc
FexMo100-x and fcc FexW100-x nanoclusters, in agreement with
the HRTEM measurements represented in Figure 5. A similar
behavior was also observed in the nonequilibrium phase
1
1,14
15
16
bimetallic Fe-Cu,
Al-Ru, and Zr-Al nanoclusters
produced by mechanical attrition. The cluster size decrease in
the nonequilibrium phase bimetallic nanoclusters produced by
mechanical attrition was explained by using the mechanism of
plastic deformation and solution hardening balanced with
11
recovery behavior. However, for the present equilibrium phase
FexMo100-x and FexW100-x nanoclusters, there was no mechan-
ical force on the bimetallic nanoclusters during formation
process. Remember that both FexMo100-x and FexW100-x nano-
clusters in this work were produced through steady-state
collisions between atoms in gas phase. Although not clear for
the moment, the cluster size decrease with x might be related
to the surface tension change of the bimetallic nanoclusters due
to the minor atoms, because the surface tension is generally
reduced as foreign atoms are accumulated on the cluster surface.
This is confirmed from a larger cluster size change in the fcc
FexW100-x nanoclusters than in the fcc FexMo100-x nanoclusters
as shown in Figure 6, because the W metal has a larger bulk
4
3
surface tension than Mo metal and thus, a larger surface
tension change in the fcc FexW100-x nanoclusters is expected
than in the fcc FexMo100-x nanoclusters.
Figure 5. Transmission electron microscope (TEM) micrographs of
(
(
a) pure Fe nanoclusters, (b) bimetallic Fe25.4Mo74.6 nanoclusters, and
c) pure Mo nanoclusters. Also shown are high-resolution transmission
(
iii-b) Strain. Figure 7 represents a plot of the atomic level
2
1/2
root-mean-square strain <e > as a function of the Fe mole
electron microscope (HRTEM) micrographs of (d) bimetallic Fe7.2Mo92.8
nanoclusters, (e) bimetallic Fe22.0 78.0 nanoclusters, and (f) bimetallic
27.7 nanoclusters. In d-f, the arrow indicates one cluster.
2
1/2
percent x. As can be seen in Figure 7, the <e > of both
bimetallic nanoclusters was large and increased as x increased.
This strain increase with x was somewhat due to the solubility
increased with x, even though the solubility increase with x was
tiny, as discussed before, because a more lattice distortion (or
a more strain) is induced as the solubility increases. Note that
the mismatch in physical dimensions such as the lattice constant
and the lattice structure between Fe and M (M ) Mo and W)
nanoclusters induces the lattice distortions. Thus, the larger the
W
Fe72.3W
microscope (HRTEM) micrographs. As can be seen in Figure
, the cluster diameter generally decreased as the degree of
5
alloying increased. Interestingly, the HRTEM micrograph of
Figure 5f corresponds to the amorphous pattern. As can be seen
in Figure 5f, the lattice structures of individual clusters with
cluster diameter of less than 1 nm were almost distorted, in

5860 J. Phys. Chem. B, Vol. 105, No. 25, 2001
Lee et al.
2
1/2
Figure 7. Plot of the atomic level root-mean-square strain <e > of
the bimetallic fcc Fe Mo100-x and fcc Fe 100-x nanoclusters as a
Figure 9. Plot of (δs)o,D/(δs)o,e as a function of the Fe mole percent x
in which the (δs)o,D and the (δs)o,e are the fwhm due to the cluster
x
x
W
2
1/2
2
1/2
function of the Fe mole percent x. This plot shows that <e >
diameter D and the atomic level root-mean-square strain <e > ,
increased as x increased.
respectively. This plot shows that the contribution of the cluster size
to the observed fwhm is more important than that of the atomic level
strain.
of the fcc FexMo100-x nanoclusters, in agreement with the plot
shown in Figure 7.
It is interesting to know which effect among the strain effect
and the cluster size effect contributed more to the observed
fwhm () (δs)o). To estimate which effect contributed more to
the observed (δs)o than the other, we estimated the (δs)o,D/(δs)o,e
2
1/2
(
) 1/2.5<e > sD), the ratio in the fwhm due to the cluster
2
1/2
diameter D to that due to the atomic level strain <e > . Figure
represents a plot of (δs)o,D/(δs)o,e as a function of the Fe mole
9
percent x for the fcc FexMo100-x nanoclusters. Here, the fcc
FexMo100-x nanoclusters were used instead of the fcc FexW100-x
nanoclusters simply because the fcc FexMo100-x nanoclusters
have more data points than the fcc FexW100-x nanoclusters, as
can be seen in both Figure 6 and Figure 7. Figure 9 clearly
shows that the (δs)o,D/(δs)o,e is much greater than 1 for all x,
which indicates that the (δs)o,D is the major contribution to the
observed (δs)o. Note that the (δs)o,D will be large for the small
clusters because of the (δs)o,D ) 1/D relationship and that the
produced cluster size in this work was small (see Figure 6).
Also, note that the (δs)o,e will be small because the lattice
distortion was minimal due to a low solubility of Fe atoms in
the fcc Mo nanoclusters. Thus, these two results will give the
result why the (δs)o,D was larger than the (δs)o,e.
Figure 8. Plot of ∆r as a function of the Fe mole percent x.
mismatch, the more the strain. Note that in the previous works
we found that Fe, Mo, and W nanoclusters had bcc, fcc, and
fcc structures with the lattice constants of 0.286, 0.419, and
7-29
0
.416 nm, respectively.2
Thus, both structural and lattice
constant differences between Fe and M nanoclusters are large,
which explains the large atomic level strain and the large atomic
level strain increase with x shown in Figure 7.
It is interesting to note that the atomic level strain of the fcc
FexW100-x nanoclusters was somewhat larger than that of the
12,13
Fe-W nanoclusters produced by mechanical attrition.
From
this work and other works,1 it was observed that the atomic
level strain increased as the cluster size decreased. Considering
that the cluster sizes of the bimetallic nanoclusters produced in
this work are much smaller than those produced by mechanical
attrition, it is likely that the atomic level strain of the fcc
FexW100-x nanoclusters produced in this work is somewhat larger
than that of the Fe-W nanoclusters produced by mechanical
attrition. This also explains why the atomic level strain of the
fcc FexW100-x nanoclusters was larger than that of the fcc
FexMo100-x nanoclusters because we observed that the cluster
size of the fcc FexW100-x nanoclusters was smaller than that of
the fcc FexMo100-x nanoclusters, as can be seen in Figure 6.
The increase of the atomic level strain with x can be directly
noticed from the increase of the bond length change with x.
Figure 8 represents the bond length change percent ∆r defined
by [(ro - r)/ro] × 100, in which ro and r are the bond lengths
of the pure fcc M (M ) Mo and W) nanoclusters and the fcc
FexM100-x nanoclusters, respectively, plotted as a function of
the Fe mole percent x. Figure 8 clearly shows that the ∆r
increased as x increased, which confirms the atomic level strain
increase with x shown in Figure 7. Also, Figure 8 shows that
the ∆r of the fcc FexW100-x nanoclusters is larger than that of
the fcc FexMo100-x nanoclusters, which indicates that the atomic
level strain of the fcc FexW100-x nanoclusters is larger than that
4,15
(iv) Why the fwhm of fcc Peak is Greater Than the fwhm
of the bcc Peak. As mentioned previously, the fwhm of the
bcc peaks in FexM100-x (M ) Mo and W) nanoclusters increased
only slightly as x decreased, whereas those of the fcc peaks
increased noticeably as x increased (see Figures 1 and 2). This
was principally due to the much larger cluster sizes of the bcc
FexM100-x nanoclusters than those of the fcc FexM100-x nano-
4
2
clusters. From the present and previous works, the cluster
diameters of pure nanoclusters are D(Fe) > D(Mo) ≈ D(W).
Thus, although the cluster size of the bcc FexMo100-x and bcc
FexW100-x nanoclusters will decrease as the degree of alloying
increases, the fwhm due to the cluster size () (δs)o,D) will not
increase much for large clusters because of the (δs)o,D ) 1/D
relationship. For this reason, the fwhm of the bcc peaks in
FexM100-x (M ) Mo and W) nanoclusters increased only slightly
as x decreased, whereas those of the fcc peaks increased
noticeably as x increased.
(v) Amorphous Structure. The amorphous structure ob-
served in the FexW100-x nanoclusters at x g 12.8 is peculiar
because it was not observed in the FexMo100-x nanoclusters. In
the previous work, we observed the amorphous structure in both
Mo and W nanoclusters when the cluster diameter was below
27-29
∼1 nm.
At this cluster diameter, there exist many structural

XRD Patterns of Bimetallic FexM100-x Nanoclusters
J. Phys. Chem. B, Vol. 105, No. 25, 2001 5861
isomers such as molecular clusters, icosahedron, decahedron,
cuboctahedron, and their deformed structures.44 The molecular
clusters and the deformed structures of icosahedron, decahedron,
and cuboctahedron provide the real amorphous pattern. In the
present case, the regular structures of icosahedron, decahedron,
and cuboctahedron also exist, although these structures were
not distinguishable in the observed XRD patterns, probably
because the amount of the nanoclusters with the regular
structures was not enough to show their characteristic XRD
patterns, respectively. That is, the present amorphous pattern
in FexW100-x nanoclusters contains all these contributions. Note
that the amorphous structure in the FexW100-x nanoclusters began
to appear at x ) 12.8 whose cluster diameter is expected to be
somewhat below 1.6 nm as judged from Figure 6. This implies
that the amorphous structure in the FexW100-x nanoclusters was
also primarily due to a tiny cluster size as in the pure
nanoclusters. On the other hand, note that the amorphous
structure was also observed in both Fe-Mo and Fe-W
(4) Chien, C. L.; Liou, S. H.; Kofalt, D.; Yu, W.; Egami, T.; McGuire,
T. R. Phys. ReV. B 1986, 33, 3247.
(
5) Toshima, N.; Harada, M.; Yamazaki, Y.; Asakura, K. J. Phys.
Chem. 1992, 96, 9927.
(6) Torigoe, K.; Nakajima, Y.; Esumi, K. J. Phys. Chem. 1993, 97,
8
5
304.
(7) Esumi, K.; Wakabayashi, M.; Torigoe, K. Colloids Surf. 1996, 109,
5.
(8) Herr, U.; Jing, J.; Gonser, U.; Gleiter, H. Solid State Commun. 1990,
76, 197.
(9) Chang, L. L.; Esaki, L. Surf. Sci. 1980, 98, 70.
(
(
10) Huang, X.; Mashimo, T. J. Alloys Comp. 2000, 296, 183.
11) Eckert, J.; Holzer, J. C.; Krill, C. E., III; Johnson, W. L. J. Mater.
Res. 1992, 7, 1980.
12) Boldrick, M. S.; Yang, E.; Wagner, C. N. J. J. Non-Cryst. Solids
992, 150, 478.
13) Shen, T. D.; Wang, K. Y.; Quan, M. X.; Wang, J. T. J. Appl. Phys.
1992, 71, 1967.
14) Eckert, J.; Holzer, J. C.; Krill, C. E., III; Johnson, W. L. J. Appl.
Phys. 1993, 73, 2794.
15) Hellstern, E.; Fecht, H. J.; Fu, Z.; Johnson, W. L. J. Appl. Phys.
(
1
(
(
(
1
6
1
989, 65, 305.
(16) Fecht, H. J.; Han, G.; Fu, Z.; Johnson, W. L. J. Appl. Phys. 1990,
7, 1744.
nanoclusters with the cluster diameter of 2-15 nm produced
(
17) Eckert, J.; Holzer, J. C.; Johnson, W. L. J. Appl. Phys. 1993, 73,
31.
(18) Ma, E.; Atzmon, M. Phys. ReV. Lett. 1991, 67, 1126.
19) Hernando, A.; Crespo, P.; Escorial, A. G.; Barandiaran, J. M. Phys.
ReV. Lett. 1993, 70, 3521.
20) Yavari, A. R.; Desr e´ , P. J.; Benameur, T. Phys. ReV. Lett. 1992,
68, 2235.
by mechanical attrition.1
0,13,45
The amorphous structure observed
in mechanical attrition was, however, not due to a small cluster
size but due to a nonequilibrium phase composition because
the amorphous structure can be produced at any cluster size in
case of a nonequilibrium phase composition.
(
(
(
21) Aning, A.; Oliver, F. W.; Courtney, T. H.; May, L. J. Appl. Phys.
1
993, 73, 5749.
(
Conclusion
22) Crespo, P.; Hernando, A.; Yavari, R.; Drbohlav, O.; Escorial, A.
G.; Barandiaran, J. M.; Or u´ e, I. Phys. ReV. B 1993, 48, 7134.
23) Ca e¨ r, G. L.; Delcroix, P.; Shen, T. D.; Malaman, B. Phys. ReV. B
996, 54, 12775.
24) Lemoine, C.; Fnidiki, A.; Lemarchand, D.; Teillet, J. J. Magn.
Magn. Mater. 1999, 203, 184.
25) Huh, S. H.; Oh, S. J.; Kim, Y. N.; Lee, G. H. ReV. Sci. Instrum.
999, 70, 4366.
26) The filament temperature was obtained from the technical data
In this work, we produced the equilibrium phase FexMo100-x
and FexW100-x nanoclusters with cluster size of 1.5-4.5 nm by
thermally decomposing a mixture of two metal carbonyl vapor
species. The fwhm of the XRD peaks became broader as the
degree of alloying increased. We noticed that the cluster size
effect on the fwhm of the XRD peaks was stronger than the
strain effect because the cluster size of the bimetallic nanoclus-
ters produced in this work was so small. The solubilities of
equilibrium phase FexMo100-x and FexW100-x nanoclusters
produced in this work were much lower than those of the
nonequilibrium phase bimetallic nanoclusters produced by
mechanical attrition and those of the nonequilibrium phase alloys
produced by ion mixing, but 2-3 times higher than those of
the bulk alloys. It is important to note that the higher solubilities
of the bimetallic nanoclusters than those of the corresponding
bulk alloys allow us to produce bimetallic nanoclusters having
electric, magnetic, and mechanical properties that are somewhat
different from those of the corresponding bulk alloys, which
are important for industrial applications. This allows us to design
nanodevices with special functions. Also, the equilibrium phase
bimetallic nanoclusters produced by the present method are
thermally more stable in composition than the nonequilibrium
phase bimetallic nanoclusters produced by mechanical attrition
and thus will be useful for high-temperature applications.
(
1
(
(
1
(
provided by the Pelican Wire Company, Inc., 6266 Taylor Road, Naples,
FL 34109-1896, USA.
(
(
27) Lee, G. H.; Huh, S. H.; Jung, H. I. J. Mol. Struct. 1998, 440, 141.
28) Huh, S. H.; Kim, H. K.; Park, J. W.; Lee, G. H. Phys. ReV. B 2000,
62, 2937.
(29) Oh, S. J.; Huh, S. H.; Kim, H. K.; Park, J. W.; Lee, G. H. J. Chem.
Phys. 1999, 111, 7402.
30) Klug, H. P.; Alexander, L. E. X-ray Diffraction Procedures, 2nd
(
ed.; Wiley: New York, 1974; Chapter 9.
(31) Cullity, B. D. Elements of X-ray Diffraction, 2nd ed.; Addison-
Wesley: Reading, 1978; Chapter 12.
(32) Jesser, W. A.; Shiflet, G. J.; Allen, G. L.; Crawford, J. L. Mater.
Res. InnoVat. 1999, 2, 211.
(33) Couchman, P. R.; Jesser, W. A. Nature 1977, 269, 481.
(
34) Abrahamson, E. P., II; Lopata, S. L. Trans. Met. Soc. AIME 1966,
36, 76.
35) Binary Alloy Phase Diagrams; Massalski, T. B., Ed.; American
2
(
Society of Metals: Metals Park, 1986.
(36) Hume-Rothery, W. The Structures of Alloys of Iron; Pergamon:
Oxford, 1966; Chapter 5.
(37) de Boer, F. R.; Boom, R.; Mattens, W. C. M.; Miedema, A. R.;
Niessen, A. K. Cohesion in Metals; North-Holland: Amsterdam, 1988;
Chapter 3.
(
(
(
(
38) Miedema, A. R. Philips Tech. ReV. 1976, 36, 217.
39) Watson, R. E.; Bennett, L. H. Phys. ReV. Lett. 1979, 43, 1130.
40) Zhang, Z. J. J. Appl. Phys. 1997, 81, 2027.
Acknowledgment. This work was supported by the KNU
Research Fund (2000) and KRF (2000-041-D00137). We thank
the Korea Basic Science Institute for allowing us to use the
X-ray diffraction spectrometer and inductively coupled plasma
atomic emission spectrometer at a membership rate.
41) Lin, C.; Chen, Y. G.; Liu, B. X. Nucl. Instrum. Methods B 1999,
148, 946.
(42) Jung, H. I.; Huh, S. H.; Oh, S. J.; Kim, Y. N.; Kim, H. K.; Park,
J. W.; Chung, J. J.; Lee, G. H. J. Korean Phys. Soc. 1999, 35, 265. We
multiplied the observed fwhm provided in Table 2 by 1.85.
(43) CRC, Handbook of Chemistry and Physics, 65th ed.; Weast, R. C.,
References and Notes
Astle, M. J., Beyer, W. H., Eds.; CRC Press: Boca Raton, Florida, 1985;
p. F-23 and p. F-26.
(44) Vogel, W. Cryst. Res. Technol. 1998, 33, 1141.
(45) Schaaf, P.; Rixecker, G.; Yang, E.; Wagner, C. N. J.; Gonser, U.
Hyperfine Interact. 1994, 94, 2239.
(1) Benjamin, J. S. Metall. Trans. 1970, 1, 2943.
(2) Weeber, A. W.; Bakker, H. Physica B 1988, 153, 93.
(3) Koch, C. C. Nanostruct. Mater. 1993, 2, 109.