Ultrasound-assisted preparation of a highly active and selective Co-B amorphous alloy catalyst in uniform spherical nanoparticles

Article abstract of DOI:10.1016/j.jcat.2006.12.014

Uniform spherical Co-B amorphous alloy nanoparticles were prepared by ultrasound-assisted reduction of Co(NH₃)²⁺₆ with BH^-₄ in aqueous solution, and the particle size was adjusted by changing either the ultrasound power or the ultrasonication time. During liquid-phase cinnamaldehyde (CMA) hydrogenation, the as-prepared Co-B catalyst exhibited much higher activity and better selectivity to cinnamyl alcohol (CMO) than the regular Co-B obtained by direct reduction of Co²⁺ with BH^-₄. The higher activity can be attributed to both the higher dispersion of Co active sites (S_Co) and the higher intrinsic activity (R^S). The higher selectivity can be attributed to both the uniform Co-B amorphous alloy particles and the strong electronic interaction between Co and B, which enhances the competitive adsorption of C{double bond, long}O group against C{double bond, long}C group in the CMA molecule. Meanwhile, the stronger adsorption for hydrogen on Co active sites was more favorable for C{double bond, long}O hydrogenation in comparison with the C{double bond, long}C hydrogenation.

Full text of DOI:10.1016/j.jcat.2006.12.014

Journal of Catalysis 246 (2007) 301–307
www.elsevier.com/locate/jcat
Ultrasound-assisted preparation of a highly active and selective Co-B
amorphous alloy catalyst in uniform spherical nanoparticles
Hexing Li a,∗, Hui Li a, Jing Zhang a, Weilin Dai b, Minghua Qiao b
a
Department of Chemistry, Shanghai Normal University, Shanghai 200234, PR China
b
Department of Chemistry, Fudan University, Shanghai 200433, PR China
Received 6 October 2006; revised 13 December 2006; accepted 17 December 2006
Available online 22 January 2007
Abstract
Uniform spherical Co-B amorphous alloy nanoparticles were prepared by ultrasound-assisted reduction of Co(NH ) with BH in aqueous
2+
−
4
3
6
solution, and the particle size was adjusted by changing either the ultrasound power or the ultrasonication time. During liquid-phase cinnamalde-
hyde (CMA) hydrogenation, the as-prepared Co-B catalyst exhibited much higher activity and better selectivity to cinnamyl alcohol (CMO) than
−
2+
the regular Co-B obtained by direct reduction of Co with BH . The higher activity can be attributed to both the higher dispersion of Co active
4
S
sites (S ) and the higher intrinsic activity (R ). The higher selectivity can be attributed to both the uniform Co-B amorphous alloy particles and
Co
the strong electronic interaction between Co and B, which enhances the competitive adsorption of C=O group against C=C group in the CMA
molecule. Meanwhile, the stronger adsorption for hydrogen on Co active sites was more favorable for C=O hydrogenation in comparison with
the C=C hydrogenation.
© 2006 Elsevier Inc. All rights reserved.
Keywords: Sonochemical preparation; Co-B amorphous alloy catalyst; Uniform particles; Cinnamaldehyde; Hydrogenation; Cinnamyl alcohol
1. Introduction
the coverage of the Co active sites. One promising approach
to designing uniform Co-B amorphous alloy particles is to re-
duce cobalt complexes by BH−4 , as is often used in preparing
Ni-P amorphous alloys [12]. However, the poor reduction de-
gree seems problematic. Ultrasonication has proven a useful
technique for inducing chemical reaction and inhibiting particle
agglomeration due to the chemical effects from acoustic cavita-
tion, which produces unusual chemical environments [13–16].
In this paper, we report a Co-B amorphous alloy catalyst
in uniform nanoparticles with controllable size prepared by
ultrasound-assisted reduction of Co(NH3)26+ with BH4−. Such
Co-B catalyst exhibits higher activity and even better selec-
tivity than the regular Co-B obtained via traditional methods
during liquid-phase cinnamaldehyde (CMA) hydrogenation to
cinnamyl alcohol (CMO). This is of industrial importance be-
cause CMO is widely used in organic synthesis and in the pro-
duction of perfumes, flavorings, pharmaceuticals, and other fine
chemicals [17–22]. Correlation of the catalytic performance to
the structural properties is discussed briefly based on various
characterizations.
Amorphous alloy catalysts have received much attention ow-
ing to their higher activity, better selectivity, and stronger sulfur
resistance in many hydrogenation reactions. Co-B is one of
the most widely studied of these catalysts [1–3]. Chemical re-
duction is most frequently used in preparing Co-B amorphous
alloys [4–7]; however, the regular Co-B samples obtained via
direct reduction of Co2+ by BH−4 in aqueous solution usually
display low surface area, broadly distributed particle size, and
poor thermal stability against crystallization due to aggregation,
because the reduction reaction is vigorous and exothermic [8,9].
At very low temperatures or in ethanol solution, the reaction
between Co2+ and BH−4 can proceed smoothly, providing the
Co-B amorphous alloy catalysts with high surface area [10,11].
But very high amounts of boron oxides coexisting with the Co-
B amorphous alloy are harmful to catalytic performance due to
*
Corresponding author. Fax: +86 21 64322272.
E-mail address: hexing-li@shnu.edu.cn (H. Li).
0021-9517/$ – see front matter © 2006 Elsevier Inc. All rights reserved.
doi:10.1016/j.jcat.2006.12.014

302
H. Li et al. / Journal of Catalysis 246 (2007) 301–307
using AlKα radiation), during which all of the Co-B samples
were dried and pretreated in situ in pure Ar atmosphere to avoid
oxidation. All of the binding energy values were calibrated us-
ing C 1s = 284.6 eV as a reference.
2.3. Activity test
Liquid-phase hydrogenation of CMA was carried out in a
500-ml stainless autoclave containing 5.0 g of catalyst, 15 ml of
CMA, 105 ml of EtOH, and 1.0 MPa of H2 at 383 K. The reac-
tion system was stirred vigorously (1000 rpm) to eliminate the
diffusion effect [25]. According to the PH drop with the reac-
2
tion time, both the specific activity (i.e., H2 uptake rate per gram
of cobalt; Rm = mmol/hg Co) and the areal activity (i.e., H2
uptake rate per m2 of the active surface area; RS = mmol/hm2
Co) were calculated by using the ideal gas equation. The RS
could be considered the intrinsic activity, because the effect of
metal dispersion was excluded. During the reaction, the H2 was
refilled to maintain 1.0 MPa, and samples were withdrawn from
the reaction mixture every 30 min for product analysis on a gas
chromatograph (GC 102) with a flame ionization detector (FID)
by means of 15% Apiezon (L)/Gas Chrom (red) at 523 K with
N2 as carrier gas, from which both the CMA conversion and the
CMO yield could be determined. The reproducibility of the re-
sults was checked by repeating the runs at least three times and
was found to be within acceptable limits ( 5%).
Fig. 1. A self-designed reactor used for sonochemical preparation of the Co-B
catalyst.
2. Experimental
2.1. Catalyst preparation
Fig. 1 schematically illustrates the self-designed reactor used
for sonochemical preparation of the Co-B sample. In the typ-
ical synthesis, 80 ml of 25% NH3·H2O solution was mixed
with 20 ml of CoCl2 solution containing 1.0 g of Co to form a
Co(NH3)2+ complex. Then 32 ml of 2.0 M KBH4 aqueous so-
lution wa6s added at 283 K in N2 atmosphere. No significant re-
action occurred in the absence of ultrasound-irradiation. Ultra-
sonication induced a smooth reduction of Co(NH3)26+ by BH4−,
resulting in a black solid that was then washed thoroughly with
H2O and absolute alcohol (EtOH) and finally stored in EtOH
until the time of use. The as-prepared Co-B samples were des-
ignated Co-B-x-y, where x and y represent ultrasound power
(50, 75, and 100 W) and ultrasonication time (30 and 90 min),
respectively. Ultrasound power <50 W or ultrasonication time
<30 min was not used due to the incomplete reduction of the
Co(NH3)26+ in aqueous solution. For comparison, the regular
Co-B sample was also prepared via direct reduction of CoCl2
by KBH4 [5] and was designated Co-B-0.
3. Results and discussion
3.1. Structural and electronic characteristics
As shown in Fig. 2, the SAED pictures revealed that, similar
to the regular Co-B (Co-B-0), the fresh Co-B samples prepared
by the present method also displayed diffractional cycles in-
dicative of the amorphous structure [26]. Meanwhile, the XRD
patterns (Fig. 3) also demonstrated that both the fresh Co-B-
50-30 and the fresh Co-B-0 samples were present in typical
amorphous alloy structure, because only one broad peak around
2θ = 45◦ was observed [27], which could be further confirmed
by subsequent XPS and EXAFS characterizations. Treatment
of the Co-B-50-30 at 773 K in N2 flow for 2 h resulted in
various diffractional peaks corresponding to both the metal-
lic Co and the Co2B alloy, implying crystallization, together
with partial decomposition of the Co-B amorphous alloy. Fur-
ther evidence for this conclusion was obtained from EXAFS.
2.2. Characterization
The bulk composition was analyzed by inductively coupled
plasma optical emission spectrometry (ICP; Varian VISTA-
MPX). The amorphous structure was determined by X-ray pow-
der diffraction (XRD; Rigaku D/Max-RB with CuKα radia-
tion), selective area electronic diffraction (SAED; JEM-2010),
and extended X-ray absorption fine structure (EXAFS; BL-
10B). Surface morphology and particle size were evaluated by
transmission electron microscopy (TEM; JEOL JEM2010). H2
temperature-programmed desorption (TPD) and H2 chemisorp-
tion were conducted on a Quantachrome CHEMBET-3000 sys-
tem, from which the metallic surface area (SCo) was calculated.
The surface electronic states were investigated by X-ray photo-
electron spectroscopy (XPS; Perkin–Elmer PHI 5000C ESCA
Fig. 2. SAED pictures of the fresh Co-B-0 and Co-B-50-30 samples.

H. Li et al. / Journal of Catalysis 246 (2007) 301–307
303
Fig. 3. XRD patterns of (a) the fresh Co-B-0, (b) the fresh Co-B-50-30, and
(c) the Co-B-50-30 after being treated at 773 K in N flow for 2 h.
2
Fig. 5. TEM morphologies of the fresh Co-B samples. (a) Co-B-0, (b) Co-B-50-
30, (c) Co-B-75-30, (d) Co-B-100-30, and (e) Co-B-50-90.
Co-B-0, Co-B-50-30 was present in uniform spherical nanopar-
ticles of relatively larger sizes. The role of the ultrasonication
can be interpreted in term of several factors. First, the intense
shock waves resulting from the cavitation effect [30] could in-
duce the reaction between Co(NH3)26+ complex and BH4− to
produce the Co-B amorphous alloy. The agglomeration of Co-B
particles could be avoided, because of the smooth reaction and
the rapid movement of Co-B particles caused by these shock
waves. Meanwhile, the formation, growth, and implosive col-
lapse of bubbles in liquid induced by ultrasonication might cre-
ate an extremely high energy level of localized supersaturation
due to evaporation of the solvent in the bubbles at high tem-
perature, which triggers nucleation [31]. Constant ultrasound
power and irradiation time resulted in a similar microenviron-
ment of localized supersaturation, possibly accounting for the
uniform shape and particle size of the Co-B sample. As shown
in Figs. 5c and 5d, the Co-B particle size increased with increas-
ing ultrasound power due to the increased energy level of the lo-
calized supersaturation. Extremely extensive ultrasound power
(100 W) or very long ultrasonication time (90 min) resulted in
an abrupt increase in particle size and even the irregular particle
shape (Figs. 5d and 5e), obviously due to the melting agglom-
eration during particle collisions [32].
Fig. 4. RDF curves of (a) the fresh Co-B-0, (b) the fresh Co-B-50-30, and (c) the
Co-B-50-30 after being treated at 773 K in N flow for 2 h.
2
As shown in Fig. 4, the radial distribution functions (RDFs) ob-
tained from the Coχ(k)k3 edge by fast Fourier transformation
exhibited only one broad peak at around R = 1.5–2.5 Å for both
the fresh Co-B-50-30 and fresh Co-B-0 samples, indicating that
there was no long-range ordering structure but only the short-
range ordering structure confined within the first-near-neighbor
atom layer [28]. Meanwhile, a shoulder peak around 1.7 Å also
demonstrated Co-B alloy formation [29]. Furthermore, calcu-
lations based on EXAFS revealed a much shorter Co-B bond
length than the sum of the Co radius and the B radius, obvi-
ously due to formation of the Co-B alloy. After being treated
at 773 K in N2 flow for 2 h, the intensity of the original peak
increased greatly, and two small additional peaks appeared at
longer distances, indicating the transformation from the amor-
phous structure to the crystalline structure.
The TEM image (Fig. 5) revealed that the Co-B-0 sample
was present in almost shapeless particles with extremely broad
size distribution, from <10 nm to >300 nm. The presence of
the large particles could be attributed to particle agglomeration
[6], whereas the small particles could be attributed to the for-
mation of boron oxides (as confirmed by the subsequent XPS
spectra), which hampered the growing of Co particles. Unlike
The XPS spectra (Fig. 6) revealed that the cobalt species
in all of the Co-B samples were present in the metallic state,
whereas the boron species were present in two states, the B al-
loying with metallic Co(Balloy) and the B oxide (B2O3). The
boron oxide can be attributed to the hydrolysis of BH−4 rather
than the surface oxidation of the alloying B, because the prepa-

304
H. Li et al. / Journal of Catalysis 246 (2007) 301–307
Fig. 6. XPS spectra of the fresh Co-B-0 and Co-B-50-30 samples.
Table 1
a
Structural and catalytic properties of different Co-B amorphous alloy catalysts
m
S
b
Catalyst
Bulk comp.
(atom%)
S
R
R
Yield
(%)
Co
2
2
(m /g)
(mmol/h g Co)
(mmol/h m Co)
Co-B-0
Co-B-50-30
Co-B-75-30
Co-B-100-30
Co-B-50-90
Co
Co
Co
Co
Co
Co
B
15.8
25.5
17.3
8.1
5.3
9.4
17.4
40.8
26.0
13.0
8.5
1.1
1.6
1.5
1.6
1.6
0.82
87.6
91.2
90.7
89.0
89.3
76.9
75.4 24.6
B
22.3
77.7
B
78.3 21.7
B
78.2 21.8
B
77.9 22.1
c
Crystallized Co-B-50-30
B
22.3
7.7
77.7
a
Reaction conditions: catalyst, 5.0 g; CMA, 15 ml; EtOH, 105 ml; T , 383 K; P 2 , 1.0 MPa; stirring rate, 1000 rpm.
H
b
c
The maximum yield obtained under the present reaction conditions.
Obtained by treating the Co-B-50-30 amorphous alloy at 773 K in N flow for 2 h.
2
ration and treatment of all of the Co-B samples were carried
out in an absolutely anoxic reactor. Based on the areas of the
XPS peaks indicative of the Balloy and the B oxide, the surface
molar ratios of the alloying B to the metallic Co(Balloy/Co) and
the alloying B to the total B(Balloy/Btotal, Btotal = Balloy + B
oxide) were determined using 0.13 and 2.50 as the PHI sensi-
tivity factors to B 1s and Co 2p3/2, respectively [23,24]. The
surface Balloy/Co ratio in the Co-B-50-30 was determined to be
32.3/67.7, much higher than that in the Co-B-0 (28.0/72.0), in-
dicating that the Co-B-50-30 was more surface-enriched with
the boron species than the Co-B-0. Meanwhile, the surface
Co 2p3/2 + 2BE Co 3s). Thus, the “final state” effect indicating
dispersion degree might be insignificant. The BE shift at boron
can be attributed mainly to the “initial state” effect resulting
from the stronger electronic interaction between the metallic
Co and the alloying B, making the B more electron-deficient
and the Co more electron-enriched. The failure to observe the
BE shift of the metallic Co can be understood by considering its
relatively greater atomic weight compared with the B atom [23].
The hydrogen TPD curves (Fig. 7) for Co-B-50-30 exhibit
only one hydrogen desorption peak at around 600 K, imply-
ing that it contains unique Co active sites. In contrast, Co-B-0
displays four hydrogen desorption peaks at around 493, 593,
640, and 822 K. Thus, we can conclude that the Co-B-0 sample
might contain at least four kinds of Co active sites, possibly cor-
responding to the metallic Co and the Co2B, Co3B, and Co4B2
alloys [29].
Other structural parameters are summarized in Table 1. ICP
analysis revealed that all of the Co-B samples prepared by
ultrasound-assisted reduction of Co(NH3)26+ by BH4− exhibited
similar bulk compositions regardless of the ultrasound power
and the ultrasonication time. Co-B-50-30 exhibited higher Co
content in the composition than Co-B-0, possibly due to a lower
amount of B oxides. Co-B-50-30 had a significantly greater
metallic area (SCo) than Co-B-0, obviously due to good dis-
persion of the Co-B particles. Further increases in ultrasound
power resulted in decreased SCo due to the increasing Co-B
particle size. Extremely extensive ultrasound power (100 W)
Balloy/Btotal ratio was much higher in the Co-B-50-30 (68/100)
compared with the Co-B-0 (51/100), indicating that the Co-B-
50-30 contained less B oxides, possibly attributed to the inhibi-
tion of BH−4 hydrolysis during the ultrasound-assisted reduction
of Co(NH3)26+ and the removal of the oxidized B species by ul-
trasound due to its cleaning effect. Thus, the Co-B sample pre-
pared through the present method contained fewer impurities.
Compared with the standard binding energy (BE) of the pure
B (187.1 eV), a positive BE shift was seen in the alloying B
in both Co-B-0 (188.2 eV) and Co-B-50-30 (188.7 eV), im-
plying partial electron transference from alloying B to metallic
Co [5,33]. The BE of the alloying B in Co-B-50-30 was even
0.5 eV higher than that in Co-B-0. Estimation on the Auger pa-
rameters demonstrated that both samples displayed the similar
“final state” values [α = BE (Co 2p3/2) + KE Co LMM] but
quite different “initial state” values (β = KE Co LMM + BE

H. Li et al. / Journal of Catalysis 246 (2007) 301–307
305
Fig. 7. Hydrogen TPD curves of the fresh Co-B-0 and Co-B-50-30 samples.
Scheme 1. A plausible reaction mechanism of the CMA hydrogenation.
Fig. 8. Reaction profiles of the CMA hydrogenation over the Co-B-0 and the
Co-B-50-30 amorphous alloy catalysts. Reaction conditions: catalyst, 5.0 g;
or very long ultrasonication time caused an abrupt drop in SCo
obviously due to the melting agglomeration of the Co-B alloy
particles, as shown in Fig. 5.
,
CMA, 15 ml; EtOH, 105 ml; T , 383 K; P 2 , 1.0 MPa; stirring rate, 1000 rpm.
H
no significant influence on RS, but did result in a decrease in
3.2. Catalytic performance
overall hydrogenation activity (Rm) due to the decreased SCo
.
The higher selectivity to CMO on Co-B-50-30 than on Co-
B-0 can be attributed mainly to the absence of HCMA byprod-
uct resulting from hydrogenation of the C=C group in the CMA
molecule. These results can be explained by considering the fol-
lowing factors.
CMA hydrogenation has been widely studied; a plausible
reaction route is described in Scheme 1 [34,35]. Besides the
main product CMO, the competitive hydrogenation of the C=C
against the C=O in the CMA molecule resulted in dihydrocin-
namaldehyde (HCMA) byproduct, and the further hydrogena-
tion of either the CMO or the HCMA resulted in another
byproduct, 3-phenylpropanol (HCMO). The CMA hydrogena-
tion curves over the Co-B-0 and Co-B-50-30 amorphous alloy
catalysts given in Fig. 8 clearly show that Co-B-50-30 ex-
hibits much higher activity and better selectivity to CMO than
Co-B-0. Other catalytic properties are summarized in Table 1.
The maximum CMO selectivity (99.0% at CMA conversion
of 92.1%) was obtained on Co-B-50-30. This value is much
higher than those values obtained on the Raney Co catalyst
(70.0 and 23.0% at CMA conversions of 70.8 and 98.7%, re-
spectively) [36] under similar reaction conditions. The superior
catalytic properties of the as-prepared catalysts compared with
the Raney Co catalyst can be attributed to their favorable amor-
phous structure, supported by the fact that fresh Co-B-50-30
exhibited much higher Rm and RS, as well as selectivity to
CMO, than the corresponding crystallized catalyst obtained by
treating the fresh catalyst at 773 K in N2 flow for 2 h.
1. The TEM morphologies (Fig. 5) demonstrated the presence
of Co-B-50-30 in uniform spherical particles of relatively
larger size, which could inhibit the adsorption of the C=C
due to steric hindrance from the neighboring benzene ring
[34] and, in turn, inhibit the hydrogenation of C=C, result-
ing in increased selectivity to CMO.
2. According to the hydrogen TPD curves (Fig. 7), Co-B-50-
30 contained only one kind of Co active site. Meanwhile,
comparing the positions of the principal hydrogen desorp-
tion peaks clearly shows stronger hydrogen adsorption for
Co-B-50-30 than for Co-B-0. The hydrogen atoms strongly
adsorbed on the Co active sites were favorable for the com-
petitive hydrogenation of the C=O group against the C=C
group [37] coexisting in the CMA molecule, resulting in
the higher selectivity to CMO. This also can account for
the higher RS of Co-B-50-30 compared with Co-B-0.
3. Blyholder and Shibabi confirmed that a metal–oxygen bond
rather than a metal–carbon bond was responsible for the
bonding between the carbonyl group and Co metal [38].
Meanwhile, Noller and Lin also confirmed that the C=O
The greater activity of Co-B-50-30 compared with Co-B-0
can be attributed to the enhanced dispersion of the Co active
sites (see the SCo values) and intrinsic activity (see the RS val-
ues). Increases in ultrasound power or ultrasonication time had

306
H. Li et al. / Journal of Catalysis 246 (2007) 301–307
molecule; and (4) the strong adsorption of the hydrogen, which
is favorable for competitive hydrogenation of the C=O against
the C=C coexisting in the CMA molecule. The optimum ul-
trasound power and ultrasonication time were determined to be
50 W and 30 min (i.e., Co-B-50-30). Extremely extensive ul-
trasound power or very long ultrasonication time is harmful for
both the activity and selectivity to CMO due to melting agglom-
eration of the Co-B particles.
Scheme 2. A plausible model of CMA adsorption and hydrogenation on the
Co-B.
hydrogenation started from the attack of H− on the C atom
[39]. Thus, the adsorption of the C=O group and its hy-
drogenation over the Co-B amorphous alloy catalyst can be
described in the model shown in Scheme 2 [25].
Acknowledgments
This work was supported by the National Natural Science
Foundation of China (20377031, 20603023), the Preliminary
973 Project (2005CCA01100), the Shanghai Science and Tech-
nology Committee (05QMX1442, 0552nm036), and the Shang-
hai Eduction Committee (T0402, 05DZ20).
The XPS spectra (Fig. 6) show that Co-B-50-30 exhibited
a stronger electronic interaction between the metallic Co and
the alloying B compared with Co-B-0, making the Co more
electron-enriched and the alloying B more electron-deficient.
On one hand, adsorption of hydrogen on the Co active sites
with higher electron density might facilitate the formation of
H− species [39] and thus, enhance hydrogenation activity to-
ward the C=O group. On the other hand, the electronic inter-
action between the C=O group and the Co active sites was the
forward donation from the highest occupied molecular orbital
(HOMO) of the C=O bonding (i.e., from 5σ of the O atom to
the surface dz2 and s orbitals of the Co atom) and the back-
donatio∗n from the dx2−y2 orbital of the Co atom to the LUMO
(i.e., πC=O) [40]. Because πC∗=O is an antibonding orbital,
the increased back-donation due to the higher electron den-
sity on the Co atom could weaken the C=O bond, favoring the
dissociation of the C=O bond. In addition, the more electron-
deficient B alloying with the metallic Co could further enhance
the adsorption for the C=O group via a side-bond interaction
(Scheme 2), thereby activating the C=O bond. Meanwhile, ul-
trasound treatment increased the dispersion and thus increased
hydrogen and/or C=O adsorption as well as the back-donation
of electrons to the antibonding orbital, because electron transfer
is facilitated on the kink, corner sites. These results also can ac-
count for the higher intrinsic activity (RS) and better selectivity
to CMO on Co-B-50-30 compared with Co-B-0.
References
[1] A. Molnar, G.V. Smith, M. Bartok, Adv. Catal. 36 (1989) 329.
[2] Y. Chen, Catal. Today 44 (1998) 3.
[3] J.F. Deng, H.X. Li, W.J. Wang, Catal. Today 51 (1999) 113.
[4] A.H. Uken, C.H. Bartholomew, J. Catal. 65 (1980) 402.
[5] H.X. Li, Y. Wu, H. Luo, M. Wang, Y. Xu, J. Catal. 214 (2003) 15.
[6] H.X. Li, H. Li, W.L. Dai, M.H. Qiao, Appl. Catal. A Gen. 238 (2003) 119.
[7] H.X. Li, H. Li, M. Wang, Appl. Catal. A Gen. 207 (2001) 129.
[8] J.Y. Shen, Z.Y. Li, Q.J. Yan, Y. Chen, J. Phys. Chem. 97 (1993) 8504.
[9] H.X. Li, H. Luo, W.L. Dai, M.H. Qiao, J. Mol. Catal. A Chem. 203 (2003)
267.
[10] Y.Z. Chen, J.S. Wu, J. Chin. Inst. Chem. Eng. 23 (1992) 119.
[11] Z.B. Yu, M.H. Qiao, H.X. Li, J.F. Deng, Appl. Catal. A Gen. 163
(1997) 1.
[12] H.X. Li, W.J. Wang, H. Li, J.F. Deng, J. Catal. 194 (2000) 211.
[13] K.S. Suslick, S.B. Choe, A. Cichowlas, M.W. Grinstaff, Nature 353 (1991)
414.
[14] P. Mulvaney, M. Cooper, F. Grieser, D. Meisel, J. Phys. Chem. 94 (1990)
8339.
[15] N.A. Dhas, A. Gedanken, Chem. Mater. 9 (1997) 3144.
[16] L.H. Thompson, L.K. Doraiswamy, Ind. Eng. Chem. Res. 38 (1999) 1215.
[17] T.B.L.M. Marinelli, S. Nabuurs, V. Ponec, J. Catal. 151 (1995) 431.
[18] A. Ghosh, R. Kumar, Micropor. Mesopor. Mater. 87 (2005) 33.
[19] X.Q. Wang, R.Y. Saleh, U.S. Ozkan, J. Catal. 231 (2005) 20.
[20] Á. Zsigmond, I. Balatoni, K. Bogár, F. Notheisz, F. Joó, J. Catal. 227
(2004) 428.
[21] S.I. Fujita, Y. Sano, B.M. Bhanage, M. Arai, J. Catal. 225 (2004) 95.
[22] F. Delbecq, P. Sautet, J. Catal. 220 (2003) 115.
[23] H. Li, H.X. Li, W.L. Dai, Z. Fang, J.F. Deng, Appl. Surf. Sci. 152 (1999)
25.
[24] Physical Electronic Division, Perkin–Elmer, Operator’s Reference Manual
for PHI PC Windows Software Version 1.2b, p. 274.
[25] H.X. Li, X. Chen, Y. Xu, Appl. Catal. A Gen. 225 (2002) 117.
[26] S. Klein, J.A. Martens, R. Parton, K. Vercruysse, P.A. Jacobs, W.F. Maier,
Catal. Lett. 38 (1996) 209.
4. Conclusion
The ultrasound-assisted chemical reduction of Co(NH3)62+
by BH−4 in aqueous solution has proven to be a promising
approach to designing uniform spherical Co-B amorphous al-
loy particles of controllable size. Such a Co-B catalyst exhib-
ited much higher activity and better selectivity to CMO during
liquid-phase CMA hydrogenation compared with regular Co-B
prepared by direct reduction of Co2+ with BH4−. This find-
ing can be attributed to (1) the single type of Co active site;
(2) the uniform Co-B alloy particles with relatively large size,
which might inhibit adsorption for the C=C group in the CMA
molecule and thus inhibit C=C hydrogenation to form HCMA;
(3) the strong electronic interaction between the metallic Co
and the alloying B, resulting in more electron-enriched Co and
more electron-deficient B, which enhance the adsorption and
activation of both hydrogen and the C=O group in the CMA
[27] A. Yokoyama, H. Komiyama, H. Inoue, T. Masumoto, H.M. Kimura,
J. Catal. 68 (1981) 355.
[28] J.A. Schwarz, C. Contescu, A. Contescu, Chem. Rev. 95 (1995) 477.
[29] B. Shen, S. Wei, K. Fan, J.F. Deng, Appl. Phys. A 65 (1997) 295.
[30] K.S. Suslick, S.J. Doctycz, Science 247 (1990) 1067.
[31] S. Devarakonda, J.M.B. Evans, A.S. Myerson, Cryst. Growth Des. 3
(2003) 741.
[32] T. Prozorov, R. Prozorov, K.S. Suslick, J. Am. Chem. Soc. 126 (2004)
13,890.
[33] Z.G. Fang, B.R. Shen, K.N. Fan, Chin. J. Chem. Phys. 15 (2002) 17.
[34] T.B.L.M. Marinelli, V. Ponec, J. Catal. 156 (1995) 51.

H. Li et al. / Journal of Catalysis 246 (2007) 301–307
307
[35] Y. Nitta, Y. Hiramatsu, T. Imanaka, Chem. Express 4 (1989) 281.
[36] B.J. Liu, L.H. Lu, T.X. Cai, K. Iwatani, Appl. Catal. A 180 (1999) 105.
[37] H. Hu, M. Qiao, S. Wang, K. Fan, H.X. Li, B. Zong, X. Zhang, J. Catal.
221 (2004) 612.
[38] G. Blyholder, D. Shihabi, J. Catal. 46 (1977) 91.
[39] H. Noller, W.M. Lin, J. Catal. 85 (1984) 25.
[40] E. Boellaard, R.J. Vreeburg, O.L.J. Gijzeman, J.W. Geus, J. Mol. Catal. 92
(1994) 299.