Synthesis of skeletal silver from rapidly solidified Al-Ag precursor

Article abstract of DOI:10.1016/S0925-8388(02)00798-3

Skeletal silver was synthesized by applying an alkaline leaching process to rapidly solidified Al-Ag granular precursors. The structure of the leached materials was examined by scanning electron microscopy, EPMA, X-ray diffraction and differential scanning calorimetry. Most of the aluminum was leached out and only a few percent aluminum remained in the as-leached state. The crystallographic structure of the leached material was face centered cubic and it was formed by rearrangement of silver atoms dissolved in the α-Al solid solution during leaching. The Ag₂Al phase was too stable for the alkaline solution so that it did not contribute to the formation of skeletal silver. The formation ratio of skeletal silver to the total silver in the precursor depended on the amount of dissolved silver in the α phase. Rapid solidification of Al-Ag alloys was quite effective for obtaining the super-saturated α phase. More than 80% of the silver in the precursor turned into skeletal silver using the rapidly solidified precursor. The specific surface area of the skeletal silver was about 9 m²/g. The mean diameter of the skeletal silver was about 20 nm as derived using the X-ray line broadening method. This value was 1.2-1.5 times smaller than that evaluated from the specific surface area assuming spherical silver particles. The difference may result from shape degeneration. Elsevier Science B.V. All rights reserved.

Full text of DOI:10.1016/S0925-8388(02)00798-3

Journal of Alloys and Compounds 348 (2003) 270–277
L
www.elsevier.com/locate/jallcom
Synthesis of skeletal silver from rapidly solidified Al–Ag precursor
*
I. Yamauchi , T. Mase, T. Kajiwara, M. Saraoka
Department of Material Science and Engineering, School of Engineering of Osaka University, 2-1, Yamadaoka, Suita, Osaka 565-0871, Japan
Received 2 May 2002; received in revised form 17 May 2002; accepted 17 May 2002
Abstract
Skeletal silver was synthesized by applying an alkaline leaching process to rapidly solidified Al–Ag granular precursors. The structure
of the leached materials was examined by scanning electron microscopy, EPMA, X-ray diffraction and differential scanning calorimetry.
Most of the aluminum was leached out and only a few percent aluminum remained in the as-leached state. The crystallographic structure
of the leached material was face centered cubic and it was formed by rearrangement of silver atoms dissolved in the a-Al solid solution
during leaching. The Ag2Al phase was too stable for the alkaline solution so that it did not contribute to the formation of skeletal silver.
The formation ratio of skeletal silver to the total silver in the precursor depended on the amount of dissolved silver in the a phase. Rapid
solidification of Al–Ag alloys was quite effective for obtaining the super-saturated a phase. More than 80% of the silver in the precursor
turned into skeletal silver using the rapidly solidified precursor. The specific surface area of the skeletal silver was about 9 m2 /g. The
mean diameter of the skeletal silver was about 20 nm as derived using the X-ray line broadening method. This value was 1.2–1.5 times
smaller than that evaluated from the specific surface area assuming spherical silver particles. The difference may result from shape
degeneration.
2002 Elsevier Science B.V. All rights reserved.
Keywords: Alloys; Rapid solidification; Precipitation; Metallography
1. Introduction
with decreasing temperature and becomes negligible at
room temperature.
Skeletal silver is a candidate cathode oxygen reduction
catalyst [1–4]. It is prepared by chemical leaching of an
aluminum–silver alloy. In the leaching process, skeletal
silver is formed by chemical removal of aluminum atoms
from lattice sites of the particular phase and rearrangement
of residual silver atoms into the Ag phase. If the particular
phase has a low solubility for silver, a small amount of
skeletal silver may be formed.
In the case of skeletal copper, the maximum solubility of
copper in the a-Al phase is less than 3 at%. However,
skeletal copper is usually formed from the Al2Cu phase
[5–7]. Therefore, the low solubility of copper in the a
phase is not important.
If skeletal silver is formed from silver dissolved in the a
phase, its formation ratio to the initial silver content in the
precursor alloy may depend on the solubility of silver in
the a phase. Therefore, it is necessary to increase the
solubility to increase the ratio. However, there are only a
few reports on this subject for skeletal silver.
In a previous paper, we reported that rapid solidification
of Al–Ag alloys was quite effective in forming a highly
saturated solid solution and it can suppress the formation
of the Ag2Al phase [9].
In this paper, we examine the effect of rapid solidifica-
tion or heat treatment of Al–Ag alloys on the formation
ratio of skeletal silver.
The equilibrium phase diagram of the Al–Ag system
(Fig. 1) [8] shows that both the a and Ag2Al phases can
exist at room temperature for Al-rich alloys. The maxi-
mum solubility of silver in the a phase is 23.5 at% at the
eutectic temperature. The solubility drastically decreases
2. Experimental procedure
2.1. Chemical leaching of precursors
Various Al–Ag alloys containing silver from 15 to 39
at% were used, and rapidly solidified granular precursors
were directly prepared from the melt by the Rotating-
*
Corresponding author. Tel.: 181-6-6879-7524; fax: 181-6-6879-
7536.
E-mail address: yamauchi@mat.eng.osaka-u.ac.jp (I. Yamauchi).
0925-8388/02/$ – see front matter
PII: S0925-8388(02)00798-3
2002 Elsevier Science B.V. All rights reserved.

I. Yamauchi et al. / Journal of Alloys and Compounds 348 (2003) 270–277
271
Fig. 1. Equilibrium phase diagram of the Al–Ag system [8].
Water-Atomization process [7,10] as described previously
[9]. To clarify the effect of rapid solidification on the
formation ratio of skeletal silver, some precursors were
annealed at 573 K for 1.8 ks for precipitation treatment or
at 823 K for 3.6 ks for solution treatment.
About 3 g of each granular precursor with a mean
diameter of 200 mm was used for alkaline leaching. The
leaching was carried out in a basic solution of 20 mass%
NaOH at 333 K for 1.8 ks to remove Al atoms from lattice
sites of the precursor. This is the most popular procedure
for the synthesis of Raney (skeletal) catalysts [5]. The
leaching was carried out under bubbling N2 gas to avoid
oxidization of the reactant.
There was no significant difference, except that the surface
smoothness decreased slightly on leaching. Similar results
were obtained for other precursors [7].
Fig. 3 shows the X-ray diffraction patterns of rapidly
solidified powders. With the exception of the highest
silver-containing alloy (Al61Ag39), the formation of the
Ag2Al phase was quite low. Most of the silver dissolved in
the a phase by rapid solidification. Quantitative evaluation
of the solubility by Vegard’s law was difficult because of
the similarity of the atomic radii of aluminum and silver.
The amount of exothermic heat produced during precipi-
tation of the Ag2Al phase may be used instead of Vegard’s
law. This value is proportional to the solubility. Differen-
tial scanning calorimetry (DSC) was applied for various
rapidly solidified alloys. All DSC curves showed a clear
exothermic reaction corresponding to precipitation of the
Ag2Al phase [9]. The result showed a maximum for the
Al70Ag30 alloy [9]. The solubility of silver in the a phase
was maximum for the Al70Ag30 alloy and was .23.5 at%,
as expected from the equilibrium phase diagram (Fig. 1).
The Rotating-Water-Atomization (RWA) process [7,10]
developed by the authors for the production of rapidly
solidified powder has two major advantages. The first is
that it is much more convenient to prepare relatively small
amounts of super-saturated alloy powders than with other
atomization processes. The second is the direct formation
of the granular precursor from the melt. In the RWA
process, no additional pulverization process is required.
Fig. 4 shows the XRD patterns of leached specimens.
Except for the Al61Ag39 specimen, the major phase was of
f.c.c. structure and a small amount of the Ag2Al phase.
The diffraction lines of the f.c.c. structure were similar to
2.2. Characterization of the structure of the leached
specimens
The structures of the leached specimens were examined
by monochromated X-ray diffraction (XRD), scanning
electron microscopy (SEM) and differential scanning
calorimetry (DSC). The specific surface area was mea-
sured by the BET method. The chemical composition of
the leached specimens was determined by an electron
probe micro analyzer (EPMA).
3. Results and discussion
3.1. Structural change on chemical leaching
Fig. 2 shows the change of the external appearance of
rapidly solidified Al85Ag15 alloy powder on leaching.

272
I. Yamauchi et al. / Journal of Alloys and Compounds 348 (2003) 270–277
Fig. 3. X-ray diffraction patterns of rapidly solidified precursors.
silver content and this value included a certain amount of
the residual Ag2Al phase. By subtracting the contribution
of the Ag2Al phase, the actual aluminum content of the
newly formed silver phase may be 2.5% at most.
Fig. 2. Change of shape by leaching (Al85Ag15 alloy powder).
those of the a-Al phase. In the Al85Ag15 alloy, the lattice
parameter of the a-Al f.c.c. phase was 0.40526 nm, similar
to that of pure Al (4.0497 nm). On the other hand, the
lattice parameter of the f.c.c. phase after leaching was
0.40812 nm, similar to that of pure silver (0.40863 nm).
The change of the lattice constant on leaching suggests the
formation of a different f.c.c. phase from the a-Al f.c.c.
phase.
Table 1 shows the chemical compositions of the leached
specimens evaluated by EPMA. The leached structure is
finely distributed and composed of the f.c.c. phase and
Ag2Al phase, so that the evaluated composition is the
average of the specimen. However, the amount of residual
Ag2Al phase in the leached specimens is negligible except
for the Al61Ag39 specimen because the strongest diffrac-
tion peak (0002) of the Ag2Al phase was very weak for
the alloys of lower silver content. Therefore, the evaluated
compositions were similar to those of the newly formed
silver phase. The aluminum content of the leached speci-
men increased with increasing silver in the precursor. The
aluminum content was .10% for the precursor of highest
Fig. 4. X-ray diffraction patterns of leached specimens.

I. Yamauchi et al. / Journal of Alloys and Compounds 348 (2003) 270–277
273
Table 1
Results of chemical analysis of leached specimens (at%)
From Eqs. (1) and (4):
Vpore
rprecursor
]
Specimen
Al
2.5
5.4
7.4
10.2
24.5
Ag
]
5 1 2
mAg
(5)
Vprecursor
rAg
Al85Ag15
Al80Ag20
Al75Ag25
Al70Ag30
Al61Ag39
97.5
94.6
92.6
89.8
75.5
Substituting the values rprecursor 5 3.89 g/cm3, rAg 5 10.5
g/cm3 and mAg 5 0.416 for the Al85Ag15 alloy into Eq.
(5), we obtain
Vpore
]
¯ 0.85
(6)
Figs. 3 and 4 clearly show that the a-Al phase complete-
ly disappears on leaching and turns into the silver phase.
Thus, the phase transformation from the a-Al phase to the
silver phase occurred below 373 K by changing the
composition. The conventional phase transformation oc-
curs by changing the temperature. But in this case it
occurred by changing the composition instead of the
temperature. Although the composition and crystallo-
graphic structure varied on leaching, the external appear-
ance of the particles was retained during leaching, as
shown in Fig. 2. This means that the apparent volume of
the particles was maintained after leaching. On the other
hand, the density of the silver phase is much higher than
that of the a-Al phase. To keep the volume constant during
leaching, pores must be introduced into the particle to
compensate for the density difference. For evaluation of
the total volume of introduced pores, we assume that only
the a-Al single phase existed before leaching and all silver
atoms in a particle turned into the silver phase after
leaching.
Vprecursor
About 85% porosity can be introduced into a particle by
leaching. The aluminum atoms were removed from the
lattice sites of the a-Al f.c.c. phase by leaching and the
residual silver atoms rearranged and formed the new f.c.c.
silver phase. The mechanism of the rearrangement is still
not clear.
If one nucleus of the silver phase is formed, then some
of the silver atoms around the nucleus will migrate to it
and growth may occur. However, the distance moved may
be limited so that new nuclei will form. The aluminum-
and silver-free region will form a pore. Thus a skeletal
silver structure composed of fine silver granules and pores
will be formed.
3.2. Stability of the Ag2 Al phase on leaching
The diffraction lines of the Ag2Al phase still remained
after leaching, as shown in Fig. 4. The relative peak ratio
(0002)Ag2Al/(111)Ag increased with increasing silver.
Especially for the Al61Ag39 alloy, the (111) peak height of
Ag was much lower than that of the (0002) peak of Ag2Al,
as shown in Fig. 4. This indicates a decrease of the silver
phase because the amount of Ag2Al phase did not vary
after leaching.
An alloy (Al35Ag65 alloy) of Ag2Al composition was
cast into a metallic mold of 10 mm inner diameter and
leached to confirm whether the silver phase was formed
from the Ag2Al phase.
Fig. 5 shows the XRD patterns of the as-solidified and
as-leached specimens. There is no significant difference
between them. The Ag2Al phase was stable with respect to
leaching and formation of the silver phase from the Ag2Al
phase was negligible.
The mass amount of silver (WAg) in a particle can be
evaluated from
*
W
Ag 5 (V precursor rprecursor)*mAg
(1)
where Vprecursor is the volume of the particle, mAg is the
mass composition of the particle (precursor) and rprecursor
is the density of the particle (precursor).
The mass composition can be derived from the atomic
weight of Al and Ag and the atomic composition of the
particle (precursor):
CAgMAg
]
mAg 5
(2)
CAgMAg 1 (1 2 CAg)MAl
where CAg is the atomic percent of precursor, MAg is the
atomic weight of silver, and MAl is the atomic weight of
aluminum.
3.3. Effect of rapid solidification on the formation of
skeletal silver
From this assumption, the mass amount of silver in a
particle is conserved during leaching and can be evaluated
from
The amount of silver phase was evaluated from Fig. 4
using a calibration curve based on the X-ray peak ratio
(0002)Ag2Al/(111)Ag and the volume fraction ratio
Ag2Al/Ag. The formation ratio of skeletal silver to initial
silver is shown in Fig. 6. About 80% of the initial silver
was transformed into skeletal silver in the alloys of lower
silver content. The formation ratio drastically decreased
beyond 30 at% silver. This curve is qualitatively in
W
Ag 5 rAgV
(3)
Ag
where rAg is the density of silver and V is the volume of
the silver phase in a leached particle.
Ag
The pore volume Vpore can be expressed by
Vpore 5Vprecursor 2V
(4)
Ag

274
I. Yamauchi et al. / Journal of Alloys and Compounds 348 (2003) 270–277
Fig. 5. Change of X-ray diffraction pattern on leaching for Al35Ag65 alloy.
agreement with the solubility curve for the as-solidified
specimens, as shown in a previous paper [9]. The solubility
of silver in the precursor was an important factor for
increasing the ratio. To confirm this effect, the rapidly
solidified powders were heat treated at 573 K for 3.63103
s to precipitate the Ag2Al phase. After heat treatment, the
Ag2Al phase was precipitated from the a-Al phase and the
solubility of silver in the a phase became quite low.
In this case, the formation ratio for most precursors was
near zero, as shown in Fig. 6. This clearly shows the effect
of the solubility of silver on the formation ratio. Usually,
the shape of the precursor is preferably granular for easy
leaching and for use as a granular catalyst. In the case of
granular precursors with low solubility, solution treatment
can be applied to increase the solubility from the phase
diagram (Fig. 1). To confirm this possibility, granular
precursors encapsulated in an evacuated quartz tube of 10
mm outer diameter were heated at 823 K for 1.83103 s
and the tube was then quenched in water. At 823 K, the
equilibrium solubility of silver is about 21 at%. Theoret-
ically, the a single phase can be obtained at room
temperature for an Al85Ag15 specimen if precipitation is
suppressed during water quenching. However, Fig. 7
¨
shows a very fine Widmanstatten structure inside a rather
Fig. 6. Effect of rapid solidification on formation ratio of skeletal silver
to initial silver.
coarse network of Ag2Al precipitates for an as-quenched

I. Yamauchi et al. / Journal of Alloys and Compounds 348 (2003) 270–277
275
¨
Fig. 7. Fine Widmanstatten structure in the coarse network of the Ag2Al
phase (Al85Ag15 alloy annealed at 823 K for 1.83103 s and then
water-quenched).
specimen. This indicates that the precipitation of the
Ag2Al phase was not suppressed by water quenching.
The quenching rate of the powder may be lower than
that of the bulk due to the apparent low thermal con-
ductivity of porous material filled into a tube without any
pressure. As described in a previous paper [9], the precipi-
tation rate of the Ag2Al phase from the a phase was quite
high, so that the quenching rate was not high enough to
suppress precipitation. X-ray diffraction patterns of these
quenched specimens are shown in Fig. 8.
Fig. 8. X-ray diffraction patterns of solution-treated specimens (823 K,
1.83103 s).
The major phase of the specimens was the Ag2Al phase
and this is consistent with the microstructure shown in Fig.
7. The X-ray diffraction patterns (Fig. 9) of the as-leached
state of these specimens showed Ag2Al as the major phase
and a small amount of skeletal silver. The formation ratio
of the solution-treated specimens is also plotted in Fig. 6.
Even though solution treatment was carried out, it was not
so effective in obtaining skeletal silver with high per-
formance. High solubility specimens are necessary to
increase the quenching rate. If a bulk specimen is
quenched into water, a higher quenching rate can be
attained due to the thermal conductivity being higher than
that of the powder specimen. However, an additional
pulverization process is required in the case of a bulk
specimen.
phases. The specific surface area of the Al61Ag39 alloy is
still high, although the fraction of skeletal silver is low. It
is suggested that the specific surface of the Ag2Al phase is
not negligible. If the surface area ratio of each phase can
be converted from the volume ratio as described by Eq.
(7), it can be obtained from the relation of the volume ratio
and the X-ray diffraction intensity ratio:
2 / 3
SAg /SAg Al 5 (V /V
)
(7)
Ag Ag Al
2
2
where SAg and SAg are the surface areas of, respectively,
Al
2
Ag and Ag2Al in 1 g of leached specimen, and V /V
Ag Ag Al
is the volume ratio of Ag to Ag2Al obtained from X-r2ay
diffraction.
]
The specific area Sleached of the leached specimen is
defined as
Thus, the rapidly solidified powder produced by the
Rotating-Water-Atomization process has good performance
and is an advantage for skeletal silver production.
]
(SAg 1 SAg Al)/1 5Sleach
(8)
2
and is given in column 1 of Table 2. Combining Eqs. (7)
and (8), SAg can be obtained.
The mass of the leached specimen is given by
3.4. Specific surface area and size of skeletal silver
granules
Table 2 shows the specific surface area per unit mass of
as-leached specimens obtained by the BET method. As
described above, the leached specimens were composed of
both skeletal silver and the Ag2Al phase. The values in
column 1 of Table 2 include the specific areas of both
rAgVAg 1 rAg AlV Al 5 1
(9)
Ag
2
2
where rAg and rAg are the densities of the Ag and Ag2Al
Al
2
phases, respectively.
The value of V can be determined From Eqs. (7) and
Ag

276
I. Yamauchi et al. / Journal of Alloys and Compounds 348 (2003) 270–277
with high fractal dimension. After leaching, only the a
phase was leached out and a small amount of skeletal
silver was formed from the a phase, but the Ag2Al phase
was stable and no structural change occurred on leaching.
However, the dendrite skeleton of the Ag2Al phase
remained after leaching, resulting in a high specific surface
area.
The mean diameter of the skeletal silver was evaluated
to be about 20 nm from X-ray line broadening (Scherrer’s
method). This value is 1.2–1.5 times smaller than that
evaluated from the specific surface area assuming the
shape of the skeletal granules to be spherical. The differ-
ence is due to the above assumption; the shape is not
spherical, but elongated, as observed by Smith et al. [11]
for skeletal copper.
4. Conclusions
Skeletal silver was formed by chemical leaching via
arrangement of silver atoms dissolved in the a phase. The
Ag2Al phase was stable with respect to alkaline solution.
More than 80% of the silver in the precursor re-arranged
into skeletal silver using rapidly solidified powder with
high solubility. Solution treatment of the powder to
increase the solubility of silver was not effective due to the
fast precipitation of the Ag2Al phase during water quench-
ing. The specific surface area of the skeletal silver was
about 9 m2 /g, independent of the initial composition of the
precursor. The mean diameter of the skeletal silver was
about 20 nm, as derived from X-ray line broadening. This
value is 1.2–1.5 times smaller than that evaluated from the
specific surface area assuming spherical silver granules.
This suggests that the shape of the skeletal silver granules
is not spherical.
Fig. 9. X-ray diffraction patterns of leached specimens after solution
treatment (823 K for 3.63103 s).
]
(9). Finally, the specific area SAg of skeletal silver can be
evaluated from
]
S
Ag 5 SAg /(rAg ?V )
(10)
Ag
Column 2 of Table 2 shows the specific surface area of
skeletal silver calculated using the above method.
The specific surface area of skeletal silver is about 9
m2 /g independent of the composition of the precursor.
This suggests that the size of skeletal silver granules is
almost the same in all alloys. The above results also
indicate that the specific surface area of the stable Ag2Al
phase is high. The reason why it is high in the Al61Ag39
alloy can be interpreted as follows. The primary Ag2Al
phase was formed in the rapidly solidified Al61Ag39 alloy.
Its structural morphology consisted of quite fine dendrites
Acknowledgements
The authors would like to express their thanks to the
Hosokawa Powder Technology Foundation for financial
support.
References
Table 2
[1] H. Binder, A. Kothling, G. Sandstade, in: G. Sandstade (Ed.), From
Electrocatalysis to Fuel Cells, University of Washington Press,
1972.
[2] Y. Matsuno, K. Suzawa, A. Tsutsumi, K. Yoshida, Int. J. Hydrogen
Energy 21 (1996) 195.
Surface area of a leached specimen and skeletal silver
Specimen
Surface area per
leached specimen
(m2 /g)
Surface area per
skeletal silver
(m2 /g)
[3] Y. Matsuno, K. Suzawa, A. Tsutsumi, K. Yoshida, Int. J. Hydrogen
Energy 22 (1997) 615.
[4] S. Rahman, M.A. Al-Saleh, J. Chem. Eng. 72 (1999) 229.
[5] D.J. Young, M.S. Wainwright, R.B. Anderson, J. Catal. 64 (1980)
116.
Al85Ag15
Al80Ag20
Al75Ag25
Al705Ag305
Al615Ag39
9.8
11.5
10.6
12.7
9.7
8.6
9.1
8.1
9.5
8.6
[6] M.S. Wainwright, Chem. Ind. 68 (1996) 213.

I. Yamauchi et al. / Journal of Alloys and Compounds 348 (2003) 270–277
277
[7] I. Ohnaka, I. Yamauchi, Mater. Sci. Eng. A 182 (1994) 1190.
[8] T.B. Massalsky (Ed.), Binary Phase Diagrams, ASM, 1986, p. 8.
[9] I. Yamauchi, T. Kajiwara, T. Mase, M. Saraoka, J. Alloys Comp. (in
press).
[10] I. Ohnaka, I. Yamauchi, S. Kawamoto, T. Fukusako, J. Mater. Sci.
20 (1985) 2148.
[11] A.J. Smith, P.R. Munroe, T. Tran, M.S. Wainwright, J. Mater. Sci.
36 (2001) 3519.