Characterization and hydrodesulfurization properties of catalysts derived from amorphous metal-boron materials

Article abstract of DOI:10.1016/j.jcat.2006.12.009

Unsupported and silica-supported amorphous metal-boron materials (Ni-B, Mo-O-B, and Ni-Mo-O-B) were prepared by NaBH₄ reduction of aqueous or impregnated metal salts. The resulting materials were characterized by a range of techniques, including conventional and time-resolved X-ray diffraction. The latter technique was used to determine the onset of crystallization of the amorphous materials during annealing in He flow and to identify the phases formed. Annealing of unsupported Ni-B resulted in the crystallization of predominantly Ni₃B, followed by Ni metal, whereas Ni-B/SiO₂ formed Ni and then NiO. There was no evidence for crystallization of B-containing phases for Mo-O-B or Mo-O-B/SiO₂ on annealing; instead, the predominant phase formed was MoO₂. In general, the phases formed for Ni-Mo-O-B and Ni-Mo-O-B/SiO₂ were consistent with those formed in the monometallic materials, but at higher annealing temperatures. Catalysts prepared by sulfiding Ni-B/SiO₂ and Ni-Mo-O-B/SiO₂ materials had significantly higher thiophene HDS activities than conventionally prepared sulfided Ni/SiO₂ and Ni-Mo/SiO₂ catalysts, whereas a sulfided Mo-O-B/SiO₂ catalyst had a dramatically lower HDS activity than a sulfided Mo/SiO₂ catalyst.

Full text of DOI:10.1016/j.jcat.2006.12.009

Journal of Catalysis 246 (2007) 277–292
www.elsevier.com/locate/jcat
Characterization and hydrodesulfurization properties of catalysts derived
from amorphous metal-boron materials
Greg L. Parks a, Melissa L. Pease a, Autumn W. Burns a, Kathryn A. Layman a, Mark E. Bussell a,∗,
Xianqin Wang b, Jonathon Hanson b, José A. Rodriguez b
a
Department of Chemistry, MS-9150, Western Washington University, Bellingham, WA 98225, USA
b
Department of Chemistry, Brookhaven National Laboratory, Upton, NY 11973, USA
Received 20 July 2006; revised 5 December 2006; accepted 8 December 2006
Available online 22 January 2007
Abstract
Unsupported and silica-supported amorphous metal-boron materials (Ni-B, Mo-O-B, and Ni-Mo-O-B) were prepared by NaBH reduction of
4
aqueous or impregnated metal salts. The resulting materials were characterized by a range of techniques, including conventional and time-resolved
X-ray diffraction. The latter technique was used to determine the onset of crystallization of the amorphous materials during annealing in He flow
and to identify the phases formed. Annealing of unsupported Ni-B resulted in the crystallization of predominantly Ni B, followed by Ni metal,
3
whereas Ni-B/SiO formed Ni and then NiO. There was no evidence for crystallization of B-containing phases for Mo-O-B or Mo-O-B/SiO on
2
2
annealing; instead, the predominant phase formed was MoO . In general, the phases formed for Ni-Mo-O-B and Ni-Mo-O-B/SiO were consistent
2
2
with those formed in the monometallic materials, but at higher annealing temperatures. Catalysts prepared by sulfiding Ni-B/SiO and Ni-Mo-
2
O-B/SiO materials had significantly higher thiophene HDS activities than conventionally prepared sulfided Ni/SiO and Ni-Mo/SiO catalysts,
2
2
2
whereas a sulfided Mo-O-B/SiO catalyst had a dramatically lower HDS activity than a sulfided Mo/SiO catalyst.
2
2
© 2007 Elsevier Inc. All rights reserved.
1. Introduction
traced to strong interactions between adsorbed S and surface
B atoms, preventing the poisoning of active Ni hydrogenation
sites.
Amorphous metal-boron materials have attracted interest as
heterogeneous catalysts since the early work of Brown and
co-workers, who found that unsupported Ni-B catalyzed the
production of hydrogen from aqueous sodium borohydride
(NaBH4) [1,2]. Since then, investigation of the catalytic proper-
ties of amorphous metal-boron materials has focused primarily
on hydrogenation reactions, although they have been shown to
be active for other processes as well, including polymeriza-
tion and desulfurization [2]. Of particular interest to the current
study, Ni-B catalysts have demonstrated greater sulfur tolerance
than B-free Ni catalysts. For example, a Ni-B/SiO2 catalyst was
observed to be not only more active and selective for cyclopen-
tadiene hydrogenation than a Ni/SiO2 catalyst, but also more
sulfur tolerant [3]. Based on theoretical calculations, Luo et al.
[4] concluded that the S resistance of Ni-B catalysts can be
To the best of our knowledge, there has been only one report
in the literature describing the use of amorphous metal-boron
materials as catalysts for hydrodesulfurization (HDS). Cheng
et al. [5] measured thiophene HDS activities for alumina- and
titania-supported Ni-B catalysts. The Ni-B/TiO2 catalyst was
significantly more active than the Ni-B/Al2O3 catalyst, which
the authors attributed to differences in the thiophene HDS ac-
tivation energies over the two catalysts [5]. Suslick and co-
workers [6] recently reported the results of thiophene HDS
measurements for crystalline Co2B and Ni3B, both in unsup-
ported form, and found the catalysts to be unstable under HDS
conditions and to have relatively low activities. The use of
boron as an additive in sulfide-based catalysts has received far
greater attention in the literature than the use of amorphous
or crystalline boron-containing materials as catalysts [7]. As
described by others [8–10], the addition of small amounts of
B (<1 wt%) to sulfided Co-Mo/Al2O3 and Ni-Mo/Al2O3 cat-
*
Corresponding author. Fax: +1 360 650 2826.
E-mail address: mark.bussell@wwu.edu (M.E. Bussell).
0021-9517/$ – see front matter © 2007 Elsevier Inc. All rights reserved.
doi:10.1016/j.jcat.2006.12.009

278
G.L. Parks et al. / Journal of Catalysis 246 (2007) 277–292
alysts has been observed to increase HDS activity, whereas
greater amounts resulted in decreased activity. In each of these
cases, the B was added directly to the γ -Al2O3 support in the
form of boric acid (H3BO3) and then calcined before impreg-
nation with metal salts. However, if B was added (as H3BO3)
after the metals to Co-Mo/Al2O3 catalysts only a decrease in
HDS activity was observed [11]. The enhanced HDS activity
of sulfided Co-Mo/Al2O3 and Ni-Mo/Al2O3 catalysts to which
small amounts of B were added to the support before the metals
has been traced to a number of factors, including increased sup-
port acidity, improved Co and Mo dispersion, and altered MoS2
particle morphologies [7,12].
The focus on catalysts derived from amorphous metal-boron
materials in the current study stems from our interest in the hy-
drotreating properties of nonsulfide materials, which has led us
to investigate metal carbides, nitrides, and phosphides [13–17].
Oxide-supported molybdenum carbide (β-Mo2C), nitride (γ -
Mo2N), and phosphide (MoP), for example, have been ob-
served to have higher thiophene HDS activities than oxide-
supported sulfided Mo catalysts [13–15,18,19]. Most dramat-
ically, a silica-supported nickel phosphide catalyst (Ni2P/SiO2)
was observed to be more than 20 times more active than sul-
fided nickel on the same support (sulf. Ni/SiO2) [17].
A silica-supported Ni-B (Ni-B/SiO2) catalyst with a nominal
30 wt% Ni loading was synthesized using a modified version of
the procedure described above for the unsupported Ni-B cata-
lysts. A 5-g sample of silica (SiO2, Cabot, M-7D, 200 m2/g),
previously calcined at 773 K for 3 h, was impregnated with
an aqueous solution of 6.45 g Ni(C2H3O2)2·4H2O until in-
cipient wetness. Multiple impregnations were needed, and the
Ni(C2H3O2)2/SiO2 precursor was dried at 393 K after each im-
pregnation. A 2-g sample of the Ni(C2H3O2)2/SiO2 precursor
was placed in a 500-ml, three-necked, round-bottomed flask
and purged with N2 followed by H2 as described above. An
aqueous solution of 0.80 g of NaBH4, degassed with N2, was
added dropwise to the Ni(C2H3O2)2/SiO2 precursor under con-
stant stirring. The Ni(C2H3O2)2/SiO2 precursor turned black on
addition of the NaBH4 solution, accompanied by effervescence
and the liberation of heat. The mixture was stirred for 30 min,
after which the Ni-B/SiO2 product was collected by vacuum fil-
tration and washed three times each with nanopure water and
then ethanol. The Ni-B/SiO2 product, kept moist during the
washings and filtration, was transferred to a vacuum dessica-
tor, dried, and passivated as described above.
2.1.2. Mo-O-B and Mo-O-B/SiO2
Although we successfully prepared bulk and silica-supported
Ni-B materials in the current work, our attempts to prepare
bulk and supported Mo-B and Ni-Mo-B materials resulted in
only partial reduction of Mo species by aqueous NaBH4. How-
ever, the resulting materials were amorphous and, in the case
of the Ni-Mo-O-B materials, are novel precursors to highly
active thiophene HDS catalysts. In addition to describing the
HDS properties of catalysts derived from the amorphous metal-
boron materials, we report the detailed characterization of the
amorphous-to-crystallization transition of these materials dur-
ing annealing in flowing He using time-resolved X-ray diffrac-
tion (XRD).
An amorphous molybdenum catalyst, called Mo-O-B here-
inafter, was-prepared by reduction of ammonium heptamolyb-
date ((NH4)6Mo7O24·4H2O, Fisher Scientific, A.C.S. Grade)
with NaBH4 by a procedure similar to that described for Ni-
B. An aqueous solution of 1.04 g NaBH4 was added dropwise
to an aqueous solution of 3.90 g (NH4)6Mo7O24·4H2O under
constant stirring. After continued stirring for 30 min, the Mo-
B product was filtered under vacuum and washed three times
each with water and then ethanol. The Mo-O-B product was
dried and passivated as described above.
Synthesis of a silica-supported Mo-O-B (Mo-O-B/SiO2) cat-
alyst with a nominal Mo loading of 30 wt% was carried out
following a procedure similar to that used for the preparation
of Ni-B/SiO2 catalysts, as described above. A 5.00 g sample of
SiO2, calcined at 773 K for 3 h, was impregnated with an aque-
ous solution of 3.94 g (NH4)6Mo7O24·4H2O. Multiple impreg-
nations were necessary, and the (NH4)6Mo7O24/SiO2 catalyst
precursor was dried at 393 K after each impregnation.
An aqueous solution of 0.20 g NaBH4 was added dropwise
to 2.00 g of the (NH4)6Mo7O24/SiO2 catalyst precursor with
constant stirring. After NaBH4 addition, the mixture was stirred
for 30 min to ensure that the reaction reached completion. The
Mo-B/SiO2 product was vacuum filtered, washed, dried, and
passivated as described above.
2. Experimental
2.1. Catalyst synthesis
2.1.1. Ni-B and Ni-B/SiO2
An amorphous nickel-boron (Ni-B) catalyst was-prepared
by the reduction of nickel acetate (Ni(C2H3O2)2·4H2O, Alfa
Aesar, >99% purity) with sodium borohydride (NaBH4, EM
Science, 98% purity) using a procedure adapted from that of
Glavee et al. [20]. An aqueous solution of 3.01 g of Ni(C2H3-
O2)2·4H2O was placed in a 500-ml three-necked flask and then
flushed with N2 (Airgas, 99.999%) and then ∼10 L of hydrogen
(H2, Airgas, 99.999%). An aqueous solution of 1.97 g NaBH4
degassed with N2 was added dropwise with constant mixing.
After stirring for 30 min, the solid product was collected by
vacuum filtration, then washed three times each with water and
then ethanol (C2H5OH, Pharmco, 99.5%). While still moist, the
product was transferred to a vacuum dessicator and dried under
vacuum for 3 h. The dessicator was then back-filled to 1 atm
with a 1 mol% O2/He mixture to passivate the Ni-B product
before exposure to the ambient atmosphere.
2.1.3. Ni-Mo-B and Ni-Mo-B/SiO2
Amorphous nickel-molybdenum-boron (Ni-Mo-O-B) cat-
alysts with a range of Ni/Mo molar ratios were prepared
by reduction of Ni(NO3)2·6H2O and (NH4)6Mo7O24·4H2O
with NaBH4 following a procedure similar to that described
above for Ni-B. The synthesis procedure for a Ni-Mo-O-B
catalyst with an expected molar ratio of Ni/Mo = 1.0 is de-
scribed here. An aqueous solution of 2.00 g NaBH4 was added
dropwise to an aqueous solution containing 1.50 g of nickel

G.L. Parks et al. / Journal of Catalysis 246 (2007) 277–292
279
nitrate [Ni(NO3)2·6H2O, Alfa Aesar, 99.999%] and 1.51 g
(NH4)6Mo7O24·4H2O under stirring. After continued stirring
for 30 min, the Ni-Mo-O-B product was filtered under vacuum,
washed three times each with water and then ethanol, then dried
and passivated as described above.
Synthesis of a silica-supported Ni-Mo-O-B (Ni-Mo-O-B/
SiO2) catalyst with a theoretical Ni/Mo molar ratio of 0.75
was carried out as follows. A 5.00-g sample of calcined
SiO2 was impregnated with an aqueous solution of 2.44 g
(NH4)6Mo7O24·4H2O, followed by drying at 393 K. The
(NH4)6Mo7O24/SiO2 precursor was then impregnated with
an aqueous solution of 2.44 g Ni(NO3)2·6H2O with sub-
sequent drying at 393 K. An aqueous solution of 2.01 g
NaBH4 was added dropwise to 2.00 g of the Ni(NO3)2-
(NH4)6Mo7O24/SiO2 precursor under stirring. After continued
stirring for 30 min, the product was filtered, washed, dried and
passivated as described above.
flow of H2 at 623 K for 1 h, and then degassed in 45 ml/min
He at 650 K for 1 h. The chemisorption capacity measurements
were carried out at 196 K as described previously [18].
2.2.3. Conventional and time-resolved X-ray diffraction
measurements
Conventional X-ray diffraction (XRD) patterns were ac-
quired using a Rigaku GeigerFlex X-ray diffractometer equip-
ped with a copper X-ray source (λ = 1.54178 Å). XRD patterns
were collected for as-prepared catalysts, as well as for cata-
lyst samples annealed to different temperatures in flowing He.
For the annealing experiments, approximately 0.1 g of a cat-
alyst was supported on a quartz wool plug placed in a quartz
U-tube, where it was purged for 30 min in a 50-ml/min He (Air-
gas, 99.999%) flow. The sample was then heated to the desired
temperature, ranging from 573 to 973 K, in 30 min while con-
tinuing the He flow. The annealing temperature was maintained
for 30 min, followed by sample cooling to room temperature in
the He flow. The sample was passivated with a 50-ml/min flow
of a 1 mol% O2/He flow for 1 h before sample preparation for
the XRD measurement.
2.1.4. NiO/SiO2, MoO3/SiO2 and NiO-MoO3/SiO2
Oxidic precursors of silica-supported, sulfided Ni, Mo, and
Ni-Mo catalysts were prepared as described previously [14,16].
The time-resolved XRD data were collected at the beamline
X7B (λ = 0.922 Å) of the National Synchrotron Light Source
(NSLS) in Brookhaven National Laboratory (BNL) using a
MAR345 area detector. The sample was loaded into a sapphire
capillary cell attached to a flow system [21–23]. A small resis-
tance heater was wrapped around the capillary, and the tempera-
ture was monitored with a 0.1-mm chromel-alumel thermocou-
ple placed in the capillary near the sample. Under a helium flow
of ∼20 ml/min, the sample was heated from 298 to 1073 K
within 2 h, then cooled to 298 K over 1 h. Diffraction pat-
terns were collected during the heating and cooling processes.
The data analysis consisted of several steps. The original pow-
der rings were first integrated with the FIT2D code [24]; the
FIT2D parameters for the integration of the data were obtained
from LaB6 powder patterns. The time-resolved XRD pattern
files were finally plotted using IDL software [25].
2.2. Catalyst characterization
2.2.1. Elemental composition and X-ray photoelectron
spectroscopy
The metal (Ni, Mo) and B contents of the unsupported and
silica-supported catalysts were determined using inductively
coupled plasma–atomic emission spectroscopy (ICP-AES) by
Huffman Laboratories, Inc. X-ray photoelectron spectroscopy
(XPS) measurements were carried out using a Physical Elec-
tronics Quantum 2000 Scanning ESCA Microprobe system
with a focused monochromatic AlKα X-ray (1486.7 eV) source
and a spherical section analyzer. The XPS spectra were col-
lected at a pass energy of 23.5 or 46.95 eV. The spectra were ref-
erenced to an energy scale with binding energies for Cu(2p3/2
)
at 932.67 0.05 eV and Au(4f) 84.0 0.05 eV. Binding ener-
gies were corrected for sample charging using the C(1s) peak at
284.6 eV for adventitious carbon as a reference. Low-energy
electrons and argon ions were used for specimen neutraliza-
tion.
2.2.4. Infrared spectroscopy of adsorbed CO
The IR spectroscopic experiments were conducted in an ion-
pumped (110 L/s) ultra-high-vacuum (UHV) chamber with a
base pressure of ∼5×10−9 Torr, equipped with a Mattson RS-1
FTIR spectrometer outfitted with a narrow-band MCT detector
interfaced to a personal computer for data acquisition and treat-
ment. The chamber also contains a high-pressure cell that can
be isolated from the UHV chamber. This system has been de-
scribed in detail previously [26].
Approximately 7 mg of the desired catalyst was pressed
at 10,000 psi into a nickel metal mesh (50 × 50 mesh size,
0.002 in. wire diameter); the area of the pressed samples was
0.80 cm2. A chromel-alumel thermocouple was spot-welded to
the nickel mesh to monitor the temperature of the sample. This
assembly was then mounted onto a sample holder equipped
with resistive heating and liquid nitrogen cooling. After mount-
ing in the UHV system, the catalyst samples were evacuated to
10−3 Torr over a period of ∼30 min. Unless otherwise stated,
the catalysts were then either reduced or sulfided in situ. Specif-
2.2.2. BET surface area and oxygen chemisorption
Single-point BET surface area and O2 pulsed chemisorp-
tion measurements were obtained using a Micromeritics Pulse-
Chemisorb 2700 apparatus. Catalyst samples (∼0.10 g) were
placed in a quartz U-tube, degassed in a 60 ml/min flow of He
for 30 min at room temperature, followed by a 2 h degassing
in a 45 ml/min flow of He at 623 K, then cooled to room tem-
perature under flowing He. The BET measurements were then
carried out as described previously [18].
Chemisorption capacity measurements were conducted us-
ing a 10.3 mol% O2/He mixture (Airco) as the probe gas. After
treatment in flowing He as described above, samples were sul-
fided by heating from room temperature to 650 K (5.9 K/min)
in a 60-ml/min flow of 3 mol% H2S/H2 and holding at 650 K
for 2 h. The sulfided samples were then reduced in a 60-ml/min

280
G.L. Parks et al. / Journal of Catalysis 246 (2007) 277–292
ically, the catalysts were reduced in flowing H2 (60 sccm) at
475 K for 1 h or sulfided in 100 Torr of a 3 mol% H2S/H2 mix-
ture at 650 K for 15 min. To remove weakly adsorbed species
from the surface of the reduced and sulfided catalysts, the high-
pressure cell was then evacuated to ∼1 × 10−7 Torr before the
sample was annealed at 475 K (reduction) or 650 K (sulfidation)
for 1 min. After pretreatment, the reduced/sulfided sample was
cooled to room temperature at a pressure of ∼1 × 10−8 Torr,
and a background IR spectrum was acquired. An IR spectrum
of adsorbed CO was then collected at 298 K while the catalyst
sample was in the presence of 5.0 Torr CO.
the flow was switched to a 3.2 mol% thiophene/H2 reactor feed
(50 ml/min). The reaction was carried out for 24 h, with auto-
mated sampling of the gas effluent done at 1-h intervals. Thio-
phene HDS activities (nmol Th/g cat s) were calculated from
the total product peak areas taken from the chromatogram af-
ter 24 h.
3. Results
3.1. Catalyst characterization
Transmission FTIR spectra were acquired in the 4000–
1000 cm−1 range by collecting 128 scans at 4-cm−1 resolution.
The sample spectrum was ratioed against a background spec-
trum acquired using a blank nickel mesh mounted in the sample
holder and a pressure of 5.0 Torr CO. All IR spectra were pre-
pared by subtracting the IR spectrum obtained before dosing
from the IR spectrum acquired after dosing. The IR spectra pre-
sented in this study have been reproduced without any smooth-
ing treatment.
The metal (Ni, Mo) and B contents of the unsupported and
silica-supported amorphous materials prepared in this study
(as well as surface compositions in some cases), along with
the BET surface areas and O2 chemisorption capacities of the
supported materials, are listed in Table 1. The monometallic
Ni materials have significant B content and hereinafter are re-
ferred to as Ni-B for the bulk phase and Ni-B/SiO2 for the
supported phase. The composition determined for unsupported
Ni-B (Ni1.95B1.00) is similar to that reported by others, indi-
cating that Ni:B molar ratios near 2:1 are typical [2,27]. Re-
cently, however, compositions ranging from Ni1.33B1.00 [28] to
Ni3.10B1.00 [29] have been reported. The Ni-B/SiO2 catalyst
is more B-rich than the unsupported Ni-B, but some of the B
likely is associated with the silica support. Others have reported
compositions for silica-supported Ni-B materials ranging from
Ni1.57B1.00 [30] to Ni3.35B1.00 [31].
The monometallic Mo materials have low B content and, as
described later, significant O content; consequently, hereinafter
these materials are referred to as Mo-O-B for the bulk phase and
Mo-O-B/SiO2 for the supported phase. Similarly, the bimetallic
Ni-Mo materials contain significant amounts of O (in addition
to B) and are referred to as Ni-Mo-O-B for the bulk phase
and Ni-Mo-O-B/SiO2 for the supported material. A search of
the literature revealed two reports describing the synthesis of
amorphous Ni-Mo materials by borohydride reduction, with an
aqueous mixture of nickel chloride and sodium molybdate used
in both cases [32,33]. The only characterization information re-
ported was that the majority of the Mo in the amorphous Ni-Mo
material from the most recent study was in the +3 oxidation
state as determined by XPS [33].
2.3. Thiophene HDS activity measurements
Thiophene HDS activity measurements were carried out us-
ing an atmospheric pressure flow reactor outfitted with a gas
chromatograph (HP 5890 Series II). The gas chromatograph
was equipped with a flame ionization detector for on-line analy-
sis of thiophene and hydrocarbon products. The flow reactor
system and the specifics of the HDS activity measurements
have been described in detail previously [18].
Before measurement of thiophene HDS activities, the silica-
supported Ni-B, Mo-O-B, and Ni-Mo-O-B catalysts were sub-
jected to one of two different pretreatments: (1) reduction in H2
by heating the catalyst sample from room temperature to 650 K
(5.9 K/min) in a 60-ml/min flow of H2 and holding at 650 K
for 2 h, or (2) sulfidation by heating from room temperature to
650 K (5.9 K/min) in a 60 ml/min flow of 3 mol% H2S/H2 and
holding at 650 K for 2 h. The oxidic precursors of the silica-
supported Ni, Mo, and Ni-Mo catalysts were subjected only to
the sulfidation procedure. After catalyst pretreatment, the tem-
perature was adjusted to the reaction temperature of 643 K, and
Table 1
Properties of unsupported and silica-supported Ni-B, Mo-O-B, and Ni-Mo-O-B catalysts
Catalyst
Bulk
composition
Surface
composition
BET surface
area (m /g)
O chemisorption
2
2
(µmol O /g)
2
Ni-B
Mo-O-B
Ni-Mo-O-B (Ni/Mo = 1.0)
Mo-O-B/SiO
Ni
B
Ni
Mo
B
–
–
–
–
–
–
1.95 1.00
1.6 1.0
Mo
Ni
B
B
16.6 1.00
12.6 1.0
Mo
B
Ni Mo B
1.1 1.9 1.0
1.10
1.11 1.00
Mo
Ni
B
Mo
–
–
–
B
8.9 1.0
72
75
84
88
82
92
93
120
109
22
10
18
13
17
26
93
195
251
2
9.81 1.00
Mo
Ni-Mo-O-B/SiO (Ni/Mo = 0.12)
B
B
B
B
B
B
B
2
0.43
3.59 1.00
Ni-Mo-O-B/SiO (Ni/Mo = 0.26)
Ni
Ni
Ni
Ni
Ni
Ni
Ni
Mo
Mo
Mo
Mo
Mo
Mo
2
0.78
1.13
1.44
1.93
1.32
1.71
3.03 1.00
Ni-Mo-O-B/SiO (Ni/Mo = 0.33)
2
3.38 1.00
Ni-Mo-O-B/SiO (Ni/Mo = 0.45)
–
2
3.22 1.00
Ni-Mo-O-B/SiO (Ni/Mo = 0.99)
Ni Mo B
2
1.95 1.00
2.9 3.4 1.00
Ni-Mo-O-B/SiO (Ni/Mo = 2.22)
–
–
2
0.59 1.00
Ni-Mo-O-B/SiO (Ni/Mo = 5.52)
2
0.31 1.00
Ni-B/SiO
B
Ni B
2.8 1.0
2
1.65 1.00

G.L. Parks et al. / Journal of Catalysis 246 (2007) 277–292
281
(a)
(b)
Fig. 1. TEM images of (a) unsupported and (b) silica-supported Ni-B catalysts in as-prepared form.
A TEM image of unsupported Ni-B (Fig. 1a) reveals that
the sample comprises aggregates of spherical-shaped, amor-
phous particles of ∼5 nm in diameter, as confirmed by higher-
resolution images (not shown). These observations are consis-
tent with the results recently reported by Li et al. [27] and
Fang et al. [29], who observed spherical amorphous Ni-B parti-
cles but with larger diameters (∼10 and ∼35 nm, respectively).
TEM images of unsupported Ni-B annealed to 773 K in flow-
ing He indicate that the small Ni-B particles have sintered to
give much larger, crystalline particles, which is consistent with
the XRD pattern of a Ni-B sample annealed to 773 K (see
Fig. 5a). Interestingly, the TEM image of the Ni-B/SiO2 cat-
alyst (21.6 wt% Ni) shown in Fig. 1b shows relatively large
Ni-B particles with diameters of 20–40 nm. This range of Ni-B
particle sizes for the Ni-B/SiO2 catalyst is similar to the ob-
servations of Chen et al. [28] for a Ni-B/MCM-41 catalyst (no
loading given), in which Ni-B particle sizes of 20–60 nm were
observed, but different from those for a Ni-B/SBA-15 cata-
lyst (9.84 wt% Ni) with more uniformly sized Ni-B particles
∼6.5 nm in diameter. TEM images of the Ni-B/SiO2 catalyst
annealed to 773 K in He revealed no substantial changes in the
supported Ni-B particles other than crystallization.
TEM images of unsupported Mo-O-B and Mo-O-B/SiO2
(19.2 wt% Mo) are shown in Fig. 2. The amorphous Mo-O-B
particles are quite large in both materials, with particle diam-
eters of 80–100 nm for Mo-O-B and 60–80 nm for Mo-O-
B/SiO2. Fig. 3 shows TEM images for unsupported and silica-
supported Ni-Mo-O-B. The unsupported Ni-Mo-B material has
particle sizes of 90–100 nm, similar to the unsupported Mo-
O-B sample. For the Ni-Mo-O-B/SiO2 catalyst (7.5 wt%, Ni,
12.3 wt% Mo, Ni/Mo = 0.99), the Ni-Mo-O-B particle size is
∼80 nm. Using TEM, Fang et al. [29] observed unsupported Ni-
Cr-B with a composition Ni3.57Cr0.17B1.00 to have particle sizes
of ∼7 nm, substantially smaller than the ∼35 nm particles of
unsupported Ni-B. Not surprisingly, the Ni-Cr-B material had a
surface area four times higher than that of the unsupported Ni-
B. The Cr was determined to have a +3 oxidation state, which
the authors assigned to the partial reduction of aqueous CrO42−
to Cr2O3 on addition of KBH4 [29].
XPS spectra for the silica-supported Ni-B, Mo-O-B, and Ni-
Mo-O-B (Ni/Mo = 0.99) catalysts are shown in Fig. 4. Because
the catalysts were passivated in a 1 mol% O2/He mixture af-
ter synthesis, peaks for oxidized metal and B species dominate
the XPS spectra. For the Ni-B/SiO2 catalyst, two peaks are
apparent in the Ni(2p3/2) region at 851.5 and 855.2 eV. The
high-binding energy species is assigned to Ni2+ species in the
passivation layer surrounding the Ni-B particles and likely is
in the form of Ni(OH)2, for which Ni(2p3/2) binding energies
of 855.6–856.6 eV have been reported [34]. The low-binding
energy Ni species is assigned to Ni bonded directly to B; its
binding energy of 851.5 eV is below that of nickel metal (Ni0,
852.5–852.9 eV [35]) as well as the Ni(2p3/2) binding energies
of 852.2–853.0 eV reported by others for Ni-B in unsupported
and supported forms [31,36–38]. It has been suggested that B
donates electron density to Ni in Ni-B [3,36]. The only peak ob-
served in the B(1s) region is located at 191.1 eV, which lies be-
tween the binding energies of elemental B (187.3 eV [39]) and
B in B2O3 (192.9 eV [39]). The B(1s) binding energy measured
for the Ni-B/SiO2 catalyst lies between the values reported by
others for unsupported Ni-B catalysts (188.2–188.4 eV) [28]
and Ni-B/Al2O3 catalysts (192.1–192.2 eV) [38]; two peaks
were observed in each case. The peak at 188.2–188.4 was as-
signed to B bonded to Ni; based on the binding energy, the
authors concluded that B donates electron density to Ni, in
agreement with earlier work by Okamoto et al. [36,40]. The
peak at 192.1–192.2 eV was assigned to oxidized B species,
presumably at the surface of the Ni-B particles.

282
G.L. Parks et al. / Journal of Catalysis 246 (2007) 277–292
(a)
(b)
Fig. 2. TEM images of (a) unsupported and (b) silica-supported Mo-O-B catalysts in as-prepared form.
(a)
(b)
Fig. 3. TEM images of (a) unsupported and (b) silica-supported Ni-Mo-O-B catalysts in as-prepared form.
The XPS spectrum of the Mo-O-B/SiO2 catalyst indicates
Mo in Mo2B5 (227.9 eV [39]) and MoB2 (227.3 eV [39]). The
B(1s) binding energy of 191.2 eV is substantially higher than
that of B in Mo2B5 (187.7 eV [39]) and MoB2 (188.4 eV [39])
and is indicative of oxidized B species, as was observed for the
Ni-B/SiO2 catalyst.
The XPS spectrum of the Ni-Mo-O-B/SiO2 catalyst is es-
sentially a composite of those of the silica-supported Ni-B and
Mo-O-B catalysts, except that the peaks associated with the
most reduced Ni and Mo species are of lower intensity. Thus,
although the same Ni, Mo, and B species are present in the Ni-
that only partial reduction of the Mo6+ ions occurred dur-
ing treatment of the dried (NH4)6Mo7O24/SiO2 precursor with
aqueous NaBH4. Two peaks, at 229.8 and 231.3 eV, are ap-
parent in the Mo(3d5/2) region for the Mo-O-B/SiO2 catalyst.
The former peak has a binding energy consistent with Mo4+
in MoO2 (229.1–230.9 eV [39]); the latter peak’s binding en-
ergy is similar to that of Mo5+ in Mo2O5, (231.7 eV [41]).
The binding energies of the Mo species in Mo-O-B/SiO2 are
at higher values than for Mo metal (227.6–228.1 eV [39]), and

G.L. Parks et al. / Journal of Catalysis 246 (2007) 277–292
283
regimes in which the various phases formed are stable in a flow-
ing He atmosphere. Fig. 5a shows XRD patterns for as-prepared
Ni-B and a Ni-B sample annealed at 773 K, as well as a refer-
ence pattern for Ni3B (card no. 73-1792 [42]). The XRD pattern
for the as-prepared Ni-B exhibits a single broad peak at ∼45.0◦
that is characteristic of amorphous Ni-B alloys [27–29,43]. The
XRD pattern of a Ni-B sample annealed to 773 K in flowing
He is in good agreement with the reference pattern of Ni3B.
Depending on the synthesis and annealing conditions, single
phases or combinations of Ni3B, Ni2B, or Ni metal are typically
observed [2,27,28,33,43]. The overlap of the XRD patterns for
Ni3B and Ni2B makes it difficult to determine whether Ni2B
is present in our Ni-B sample annealed to 773 K, but we can
conclude that if it is present at all, its concentration is quite
low. Examining the sequence of time-resolved XRD patterns in
Fig. 5b allows us to determine the temperature of crystallization
of amorphous Ni-B to give Ni3B, as well as the subsequent de-
composition of Ni3B to give crystalline Ni metal. Noting that
the X-ray wavelength is 0.9220 Å for the time-resolved ex-
periments, the broad peak characteristic of amorphous Ni-B is
located at ∼28◦ in the first XRD pattern, which corresponds
to the as-prepared material at 298 K. During heating in flow-
ing He, the XRD patterns remain unchanged up to 633 K, at
which point peaks associated with crystalline Ni3B first appear.
With continued heating, the Ni3B XRD peaks strengthen in in-
tensity, reaching a maximum at ∼723 K, and then disappear
by ∼933 K. Although difficult to distinguish from Ni3B peaks,
Fig. 4. XPS spectra of silica-supported Ni-B, Mo-O-B, and Ni-Mo-O-B cata-
lysts in as-prepared form.
Mo-O-B/SiO2 catalyst, the oxidized surface layer covering the
particles is apparently thicker.
A number of techniques, including XRD and differential
scanning calorimetry (DSC), have been used by others to char-
acterize amorphous-to-crystalline transitions for unsupported
and silica-supported Ni-B catalysts. XRD has been used to
identify the crystalline phases formed on annealing. In the cur-
rent study, we combined conventional and time-resolved XRD
to carefully pinpoint the amorphous-to-crystalline transitions
for the different materials and also to identify the temperature
(a)
(b)
Fig. 5. (a) Conventional XRD patterns (λ = 1.54178 Å) of an as-prepared unsupported Ni-B catalyst, a Ni-B catalyst annealed to 773 K in flowing He, and a reference
pattern for Ni B. (b) Time-resolved XRD patterns (λ = 0.9220 Å) acquired while heating (6.5 K/min) as-prepared unsupported Ni-B catalyst in flowing He.
3

284
G.L. Parks et al. / Journal of Catalysis 246 (2007) 277–292
(a)
(b)
Fig. 6. (a) Conventional XRD patterns (λ = 1.54178 Å) of an as-prepared Ni-B/SiO catalyst, a Ni-B/SiO catalyst annealed to 773 K in flowing He, and a reference
2
2
pattern for Ni metal. (b) Time-resolved XRD patterns (λ = 0.9220 Å) acquired while heating (6.5 K/min) an as-prepared Ni-B/SiO catalyst in flowing He.
2
new peaks (e.g., 30.3◦) begin to appear at ∼713 K that are as-
signed to the XRD pattern of Ni metal. The Ni metal peaks
grow steadily in intensity before temporarily reaching a plateau
at ∼873 K, then significantly increasing in intensity starting at
∼933 K. The growth of the Ni metal XRD peaks above ∼933 K
corresponds with the decomposition of Ni3B, as determined
by the loss of intensity of Ni3B XRD peaks. Once the Ni3B
has completely decomposed at ∼933 K, the intensity of the Ni
metal XRD peaks stabilizes.
For silica-supported Ni-B, the XRD pattern for the as-
prepared catalyst (Fig. 6a), shows an increasing background at
decreasing Bragg angles associated with the amorphous silica
support, as well as a broad low-intensity peak at ∼43◦ assigned
to the supported amorphous Ni-B particles. The XRD pattern
of a sample of the Ni-B/SiO2 catalyst annealed to 773 K in
flowing He shows relatively low-intensity peaks that can be
readily assigned to Ni metal, with no indication of a crystalline
B-containing phase. This observation is consistent with those
of others for Ni-B/SiO2, Ni-B/C, and Ni-B/Al2O3 catalysts an-
nealed to >600 K [33,38,44]. The first XRD pattern in Fig. 6b,
which corresponds to as-prepared Ni-B/SiO2 at 298 K, shows
two broad peaks at 20◦ and 35◦ that do not appear at the cor-
responding 2θ values in the conventional XRD pattern of the
same catalyst (see Fig. 6a) and cannot be attributed to any
expected phase(s). The sequence of time-resolved XRD pat-
terns for the Ni-B/SiO2 catalyst shows the crystallization of
silica-supported Ni-B to give Ni metal beginning at ∼533 K,
as indicated by the growth of an XRD peak at 26.0◦. This and
other Ni metal XRD peaks reach a maximum in intensity at
∼813 K, above which the intensity begins to decrease. Above
∼813 K, the Ni metal undergoes a reaction with surface O to
form NiO (card no. 06-595 [42]), as indicated by the growth of
XRD peaks associated with NiO (e.g., 21.8◦) at the expense of
Ni peaks.
As described above, the amorphous materials produced by
NaBH4 reduction of aqueous and silica-supported (NH4)6Mo7-
O24 have low B content. XRD patterns for the annealed mate-
rials are consistent with these observations. Shown in Fig. 7a
are XRD patterns for as-prepared unsupported Mo-O-B and
a sample of this material annealed to 773 K in flowing He,
along with some reference patterns. The XRD pattern for the
as-prepared sample shows a single broad peak at ∼27.5◦, indi-
cating that the material is amorphous, whereas the pattern for
the sample annealed to 773 K in flowing He shows peaks char-
acteristic of one or more crystalline phases. The most intense
peaks in the pattern for the annealed Mo-O-B sample corre-
spond to MoO2 (card no. 32-0671 [42]), whereas less-intense
peaks can be assigned to minor phases of MoO3 (card no. 35-
0609 [42]) and possibly MoB (card no. 06-0644 [42]). If all of
the B in the Mo-O-B sample were present in the form of MoB,
then this boride phase would constitute approximately 8 mol%
of the sample. As a result, it is not surprising that the peaks
at 30.3◦ and 42.4◦ are of very weak intensity. The sequence
of time-resolved XRD patterns shown in Fig. 7b indicates that

G.L. Parks et al. / Journal of Catalysis 246 (2007) 277–292
285
(a)
(b)
Fig. 7. (a) Conventional XRD patterns (λ = 1.54178 Å) of an as-prepared unsupported Mo-O-B catalyst, a Mo-O-B catalyst annealed to 773 K in flowing He,
and reference patterns for MoO , MoO , and MoB. (b) Time-resolved XRD patterns (λ = 0.9220 Å) acquired while heating (6.5 K/min) as-prepared unsupported
2
3
Mo-O-B catalyst in flowing He.
the unsupported Mo-O-B remains amorphous until ∼613 K, at
which point the sample rapidly crystallizes to give primarily
MoO2 and some MoO3, as indicated by the growth of strong
and weak peaks at 15.5◦ and 16.4◦, respectively. There is no
evidence in the sequence of XRD patterns indicating forma-
tion of a crystalline B-containing phase. Due to experimental
problems, it was not possible to continue the time-resolved
XRD measurements above ∼733 K for the unsupported Mo-
O-B sample.
The XRD results for the Mo-O-B/SiO2 catalyst are con-
sistent with those of the unsupported Mo-O-B sample. XRD
patterns for as-prepared and annealed (773 K) samples of a
Mo-O-B/SiO2 catalyst, along with reference patterns for MoO2
and MoO3, are shown in Fig. 8. The XRD pattern for the
as-prepared sample shows only an increasing background, in-
dicating that the material is amorphous, whereas the pattern
for the sample annealed to 773 K shows peaks characteristic
of MoO2 and MoO3. As for the unsupported Mo-O-B, crys-
tallization of the amorphous silica-supported Mo-O-B occurs
rapidly, starting at a temperature of ∼623 K, to give almost
exclusively MoO2 with a minor amount of MoO3. After their
rapid increase in size, the peaks associated with MoO2 increase
gradually with increased temperature, whereas the peaks as-
sociated with MoO3 decrease in intensity and disappear by
873 K.
and 10, respectively. The XRD patterns for the as-prepared
samples are indicative of amorphous materials, whereas pat-
terns for annealed samples exhibit XRD peaks associated with
Ni metal and MoO2, as well as possibly MoB in the silica-
supported material. XRD peaks assigned to Ni and MoO2 are
narrow and intense for the unsupported Ni-Mo-O-B sample
annealed to 873 K, as is the MoO2 reflection at 26◦ for the
annealed Ni-Mo-O-B/SiO2 catalyst, but the most prominent Ni
peak at 44.6◦ is barely discernible above the background for the
silica-supported material.
The time-resolved XRD measurements for the Ni-Mo-O-B
and Ni-Mo-O-B/SiO2 materials show distinct differences from
those of the unsupported and silica-supported Ni-B and Mo-O-
B materials. For the unsupported Ni-Mo-B, the onset of crys-
tallization of Ni and MoO2 is observed at ∼550 and ∼795 K,
respectively; no formation of Ni3B or MoO3 is observed. The
formation of crystalline Ni in Ni-Mo-O-B occurs at a substan-
tially lower temperature than for Ni-B, for which crystalline Ni
formation starts at ∼713 K when crystalline Ni3B (formed at
633 K) begins to decompose. On the other hand, the onset of
crystallization of MoO2 is shifted from ∼613 K for Mo-O-B to
∼795 K for Ni-Mo-O-B. For the silica-supported materials, the
onset of crystallization is seen at ∼675 K for Ni and ∼790 K for
MoO2. These crystallization temperatures are shifted substan-
tially higher than for Ni-B/SiO2 and Mo-O-B/SiO2, for which
Ni and MoO2 crystallization start at 523 and 633 K, respec-
tively.
XRD patterns for as-prepared and annealed (873 K) Ni-
Mo-O-B and Ni-Mo-O-B/SiO2 materials are shown in Figs. 9

286
G.L. Parks et al. / Journal of Catalysis 246 (2007) 277–292
(a)
(b)
Fig. 8. (a) Conventional XRD patterns (λ = 1.54178 Å) of an as-prepared Mo-O-B/SiO catalyst, a Mo-O-B/SiO catalyst annealed to 773 K in flowing He,
2
2
and a reference pattern for MoO . (b) Time-resolved XRD patterns (λ = 0.9220 Å) acquired while heating (6.5 K/min) an as-prepared Mo-O-B/SiO catalyst in
2
2
flowing He.
Infrared spectroscopy of adsorbed CO on reduced and sul-
fided Ni-Mo-O-B/SiO2 and Ni-Mo/SiO2 catalysts at room tem-
perature are shown in Fig. 11. The IR spectra for the reduced
catalysts (Fig. 11a), which were treated in flowing H2 at 475 K,
reveal significant differences. The IR spectrum for the reduced
Ni-Mo/SiO2 catalyst shows νCO absorbances at 2048, 2141,
and 2189 cm−1 that can be assigned to CO adsorbed on Ni0,
Ni+, and Mo4+ sites, respectively, with the CO bonded linearly
in each case [45–50]. The peak centered at 2048 cm−1 is quite
broad, most likely because it comprises overlapping absorbance
features corresponding to CO adsorbed on various Ni0 sites, as
well as possibly on Ni-promoted Moδ+ sites. The observation
of CO adsorbed on oxidized Ni and Mo sites is not surpris-
ing given the mild reducing conditions used. The IR spectrum
of adsorbed CO on a reduced Ni-Mo-O-B/SiO2 catalyst shows
dramatic differences, with νCO absorbance features at ∼1903,
1953, 2006, 2040, 2059, 2127, and 2187 cm−1. Consistent with
the reduced Ni-Mo/SiO2 catalyst, the peaks at 2127 and 2187
cm−1 are assigned to CO adsorbed to Ni+ and Mo4+ sites,
respectively. The envelope of absorbance features in the 1975–
2075 cm−1 region is much more intense than the broad peak
centered at 2048 cm−1 in the IR spectrum of the reduced Ni-
Mo/SiO2 catalyst, indicating that there are substantially more
sites at the surface of the reduced Ni-Mo-O-B/SiO2 catalyst.
The νCO absorbances at ∼1903 and 1953 cm−1 are assigned to
bridge-bonded CO on Ni0 sites, and the peak at 2040 cm−1 is
assigned to linearly bonded CO on Ni0 sites [45,46]. The peak
at 2059 cm−1 is tentatively assigned to multiply coordinated
CO molecules adsorbed on highly defected Ni0 sites [46], al-
though absorbance due to adsorbed nickel tetracarbonyl species
or CO adsorbed to Ni-promoted Moδ+ sites also would be
expected in this region. Assigning the peak at 2006 cm−1 is
more difficult, because this peak is located at a wavenum-
ber somewhere between the ranges assigned to bridge-bonded
(1800–1980 cm−1) and linearly bonded (2040–2070 cm−1) CO
on Ni0 sites and at a different wavenumber than the νCO ab-
sorbances observed for adsorbed CO on a reduced Ni-Mo/SiO2
catalyst (see above) or on a mildly reduced Ni-Mo/Al2O3 cat-
alyst (1950, 2065, and 2170 cm−1) [50]. Instead, based on the
assignments of others for peaks in the 1997–2025 cm−1 range
for adsorbed CO on reduced Mo/Al2O3 catalysts [47,51], the
peak at 2006 cm−1 is tentatively assigned to CO adsorbed to
Mo0 sites.
In contrast to the IR spectral results for the reduced catalysts,
the IR spectra of adsorbed CO on the sulfided Ni-Mo/SiO2
and Ni-Mo-O-B/SiO2 catalysts shown in Fig. 11b indicate that
the adsorption sites exposed on these catalysts are quite sim-
ilar. For the sulfided Ni-Mo/SiO2 catalyst, νCO absorbances
are observed at 2095 and ∼2112 cm−1; these are assigned to
Ni-promoted Moδ+ and unpromoted Mo2+ sites, respectively
[52–55]. The IR spectrum of adsorbed CO on the Ni-Mo-O-
B/SiO2 catalyst exhibits νCO absorbance features at ∼1970,
2040, 2088, and 2144 cm−1, respectively. The peaks at ∼1970
and 2040 cm−1 are assigned to bridge-bonded and linearly

G.L. Parks et al. / Journal of Catalysis 246 (2007) 277–292
287
(a)
(b)
Fig. 9. (a) Conventional XRD patterns (λ = 1.54178 Å) of an as-prepared unsupported Ni-Mo-O-B catalyst, a Ni-Mo-O-B catalyst annealed to 873 K in flowing He,
and reference patterns for MoO and Ni. (b) Time-resolved XRD patterns (λ = 0.9220 Å) acquired while heating (6.5 K/min) as-prepared unsupported Ni-Mo-O-B
2
catalyst in flowing He.
(a)
(b)
Fig. 10. (a) Conventional XRD patterns (λ = 1.54178 Å) of an as-prepared Ni-Mo-O-B/SiO catalyst, a Ni-Mo-O-B/SiO catalyst annealed to 873 K in flowing
2
2
He, and reference patterns for MoO and Ni. (b) Time-resolved XRD patterns (λ = 0.9220 Å) acquired while heating (6.5 K/min) an as-prepared Ni-Mo-O-B/SiO
2
2
catalyst in flowing He.

288
G.L. Parks et al. / Journal of Catalysis 246 (2007) 277–292
(a)
(b)
Fig. 11. IR spectra of adsorbed CO on (a) reduced and (b) sulfided Ni-Mo-O-B/SiO and Ni-Mo/SiO catalysts.
2
2
bonded CO on Ni0 sites [45,46], indicating the presence of
some unsulfided Ni species on the catalyst surface. Finally, the
νCO absorbances at 2088 and 2144 cm−1 are assigned to CO
adsorbed on Ni-promoted Moδ+ and Ni+ sites, respectively, as
described above for Ni-Mo/SiO2.
3.2. HDS activity
The thiophene HDS activity versus time for reduced and sul-
fided Ni-Mo-O-B/SiO2 catalysts and of a sulfided Ni-Mo/SiO2
catalyst are shown in Fig. 12. The reduced Ni-Mo-O-B/SiO2
catalyst exhibited a steady HDS activity over the entire 24 h on
stream, but its activity after 24 h was lower than that of either
the sulfided Ni-Mo/SiO2 or Ni-Mo-O-B/SiO2 catalysts. These
latter two catalysts both exhibited decreasing HDS activity over
time, but the activity of the sulfided Ni-Mo/SiO2 catalyst de-
creased more rapidly such that its activity was just 70% of that
of the sulfided Ni-Mo-O-B/SiO2 catalyst after 24 h on stream.
The thiophene HDS activities of the Ni-B/SiO2, Mo-O-
B/SiO2, and Ni-Mo-O-B/SiO2 catalysts are plotted in Fig. 13 as
a function of the Ni mole fraction (excluding B content) of the
catalysts. The HDS activities are for the amorphous metal boron
catalysts subjected to a sulfidation pretreatment (see Section 2).
XRD patterns of sulfided Ni-B/SiO2, Mo-O-B/SiO2, and Ni-
Mo-O-B/SiO2 (Ni/Mo = 0.99) catalysts are shown in Fig. 14.
The XRD patterns exhibit a rapidly dropping background simi-
lar to those of the as-prepared catalysts, but the patterns for the
sulfided Ni-B/SiO2 and Ni-Mo-O-B/SiO2 catalysts show ad-
ditional features. The XRD pattern of the sulfided Ni-B/SiO2
catalyst shows peaks at 21.5◦ and 31.3◦ that can be assigned
Fig. 12. Thiophene HDS activities of reduced and sulfided Ni-Mo-O-B/SiO
2
catalysts (Ni/Mo = 0.99) and a sulfided Ni-Mo/SiO catalyst (Ni/Mo = 0.50)
2
as a function of the time on-stream.
to Ni9S8 (card no. 22-1193), and the pattern for the sulfided
Ni-Mo-O-B/SiO2 catalyst also exhibits a peak at 21.5◦.
HDS activities of sulfided Ni/SiO2, Mo/SiO2, and Ni-
Mo/SiO2 catalysts prepared by conventional preparation meth-
ods are also plotted in Fig. 13. The trend of HDS activities
for the conventionally prepared Ni-Mo/SiO2 catalysts is con-
sistent with those reported previously [7]. A broad maximum

G.L. Parks et al. / Journal of Catalysis 246 (2007) 277–292
289
NiO, 72 µmol O2/g), while having an O2 chemisorption ca-
pacity 3.5 times higher than that of the sulfided Ni/SiO2 cata-
lyst. Comparing the Ni-Mo catalysts, the most active sulfided
Ni-Mo-O-B/SiO2 catalyst (Ni/Mo = 0.99, 9.5 wt% NiO and
18.4 wt% MoO3 equivalents, 26 µmol O2/g) was 1.4 times
more active than the sulfided Ni-Mo/SiO2 catalyst with the op-
timal Ni loading (Ni/Mo = 0.5, 7.9 wt% NiO, 30.4 wt% MoO3,
23 µmol O2/g).
4. Discussion
As part of a larger research effort to explore the HDS prop-
erties of nonsulfide materials, the focus of this study was cata-
lysts derived from amorphous metal-boron materials. Unsup-
ported and silica-supported Ni-B and Ni-Mo-O-B materials
were observed to have substantial B content (B/Me = 0.2–0.5,
Me = Ni+Mo), whereas unsupported and silica-supported Mo-
O-B had minimal B contents (B/Mo = 0.08–0.1). As summa-
rized in Fig. 13, the Ni-B/SiO2 and Ni-Mo-O-B/SiO2 catalysts
were more active than their sulfide counterparts (Ni/SiO2 and
Ni-Mo/SiO2), whereas the Mo-O-B/SiO2 catalyst was substan-
tially less active than a sulfided Mo/SiO2 catalyst. Interestingly,
the HDS activity versus time data plotted in Fig. 12 indicates
that Ni-Mo-O-B/SiO2 catalysts have distinct HDS properties
relative to conventional sulfided Ni-Mo/SiO2 catalysts. Typi-
cal of the activity trends observed for sulfided Mo and Ni-Mo
catalysts tested in our laboratory [16,18], the HDS activity of a
sulfided Ni-Mo/SiO2 catalyst (Ni/Mo = 0.5) decreases steadily
over the 24-h test period. On the other hand, the reduced Ni-
Mo-O-B/SiO2 catalyst shows a steady HDS activity, whereas
the sulfided Ni-Mo-O-B/SiO2 catalyst exhibits a slowly de-
creasing activity that appears to be nearing steady-state HDS
activity after 24 h. As discussed later, insight into the HDS
properties of the catalysts derived from the amorphous Ni-Mo-
O-B/SiO2 materials can be gained from the characterization
studies carried out in this work.
Fig. 13. Thiophene HDS activities of sulfided Ni-Mo-O-B/SiO , and Ni-Mo/
2
SiO catalysts as a function of the Ni mole fraction.
2
Unsupported and supported Ni-B materials have received
the greatest attention of amorphous metal-boron materials as
catalysts and, as a result, more is understood about the proper-
ties of these materials. DSC has been used to indirectly probe
the crystallization of unsupported Ni-B and Ni-B on a range
of supports. For unsupported Ni-B, DSC reveals a very strong
exothermic peak centered at ∼615 K [27]. Li et al. also ac-
quired XRD patterns (independent of the DSC measurements)
after annealing Ni-B to increasing temperatures. No changes in
the XRD patterns were observed below 573 K, whereas above
this temperature crystallization occurred to give predominantly
Ni3B (and a small amount of Ni2B) at 673–773 K and then
metallic Ni above 773 K. The intense DSC peak at ∼615 K,
which extends over the temperature range of 500–800 K, corre-
sponds to the crystallization of Ni-B to give Ni3B and its subse-
quent decomposition to give Ni metal. This is generally consis-
tent with our conventional XRD results (Fig. 5a), in which we
observed a pattern nearly identical to that of a reference pattern
for Ni3B after annealing to 773 K, as well as our time-resolved
XRD results (Fig. 5b) that show rapid crystallization of Ni-B
starting at ∼633 K. Crystallization of the Ni-B yields Ni3B, for
Fig. 14. X-ray diffraction patterns of sulfided Ni-B/SiO , Mo-O-B/SiO , and
2
2
Ni-Mo-O-B/SiO catalysts.
2
in activity is observed for catalysts with molar ratios of Ni/Mo
∼0.5, with an enhancement of activity of approximately four
measured relative to an unpromoted sulfided Mo/SiO2 cata-
lyst. For the sulfided Ni-Mo-O-B/SiO2 catalysts, the maximum
in HDS activity is shifted to a catalyst with a molar ratio of
Ni/Mo = 0.99 and corresponds to an activity enhancement of
a factor of ∼33 relative to a sulfided Mo-O-B/SiO2 catalyst,
which had a quite low HDS activity. The sulfided Mo-O-B/SiO2
catalyst (28.7 wt% MoO3 equivalent, 22 µmol O2/g) was six
times less active than the sulfided Mo/SiO2 catalyst (30.4 wt%
MoO3, 18 µmol O2/g), despite having a slightly higher O2
chemisorption capacity. On the other hand, the sulfided Ni-
B/SiO2 catalyst (27.4 wt% NiO equivalent, 251 µmol O2/g)
was twice as active as the sulfided Ni/SiO2 catalyst (30.1 wt%

290
G.L. Parks et al. / Journal of Catalysis 246 (2007) 277–292
which XRD peaks are at a maximum at ∼723 K, followed by
formation of Ni metal starting at ∼713 K. The time-resolved
XRD measurements provide greater detail concerning the ex-
tent of crystallization and the identity of the phases formed than
is possible with DSC and conventional XRD.
active Ni9S8 phase. Due to the high reaction temperature of
723 K, which is above typical HDS processing temperatures
(573–673 K) [7], directly comparing the results of Suslick et al.
with those of the current study is difficult. At 643 K, we observe
that the sulfided Ni-B/SiO2 catalyst has twice the steady-state
HDS activity as a B-free sulfided Ni/SiO2 catalyst, despite ev-
idence of some bulk Ni9S8 formation during presulfidation of
the Ni-B/SiO2 catalyst.
Despite having a higher B/Ni molar ratio than the unsup-
ported Ni-B, there is no evidence in the conventional and time-
resolved XRD patterns for the Ni-B/SiO2 catalyst for the forma-
tion of crystalline B-containing phases. Crystalline Ni is formed
starting at ∼533 K, which subsequently converts to NiO start-
ing at ∼813 K, presumably forming via reaction of Ni with
OH groups on the support surface and/or surface O in the pas-
sivation layer formed on the catalyst surface after synthesis.
Interestingly, XRD patterns acquired for a Ni-B/C catalyst in-
dicate a higher crystallization temperature (>600 K) for Ni-B
particles on the carbon support, than on either SiO2 or γ -Al2O3
[33,38,44]. In all cases, the only crystalline phase observed to
form on the different supports is metallic Ni. Li et al. [38] ob-
served a shift of the exothermic peak in their DSC traces for
Ni-B/Al2O3 catalysts to lower temperatures with increased Ni
loading (11–15 wt% NiO equivalent). Given the high loading
of our Ni-B/SiO2 catalyst (27.4 wt% NiO equivalent), it is not
surprising that we observe the first indication of crystallization
at the lower temperature of ∼533 K using time-resolved XRD.
The XPS peak assigned to reduced Ni species (i.e., not in the
In contrast to the Ni-B/SiO2 catalyst, the Mo-O-B/SiO2
catalyst had an HDS activity six times lower than a sulfided
Mo/SiO2 catalyst with a similar Mo loading. This observation
is surprising, given that the Mo-O-B/SiO2 catalyst has a some-
what higher O2 chemisorption capacity and a very low B con-
tent (B/Mo = 0.10) that makes up just 0.22 wt% of the catalyst.
The conventional and time-resolved XRD patterns indicate that
both unsupported and silica-supported Mo-O-B crystallize to
give predominantly MoO2, with some MoO3 also present. For
the Mo-O-B/SiO2 catalyst, time-resolved XRD revealed rapid
onset of MoO2 crystallization at ∼623 K, below the sulfida-
tion temperature of 650 K. Thus, it is likely that crystalline
MoO2 on the silica support is the Mo phase that undergoes
sulfidation for the Mo-O-B/SiO2 catalyst. Thomazeau et al.
[56] compared sulfided Mo/SiO2 catalysts prepared from silica-
supported MoO2 and MoO3 precursors and measured similar
thiophene HDS activities. The addition of small amounts of B
(<1 wt%) in the form of H3BO3 to γ -Al2O3 before the prepara-
tion of sulfided Co-Mo/Al2O3 and Ni-Mo/Al2O3 catalysts has
been observed to increase HDS activity [8–10]. On the other
hand, adding B (as H3BO3) after the metals to Co-Mo/Al2O3
catalysts produced only a decrease in HDS activity [11]. For
the Mo-O-B/SiO2 catalyst, TEM revealed a large particle size
(60–80 nm), but the O2 chemisorption capacity of the catalyst
in sulfided form is somewhat higher than that for the sulfided
Mo/SiO2 catalyst, indicating that the dramatically lower HDS
activity is not due to dispersion effects. The XRD pattern of
the sulfided Mo-O-B/SiO2 catalyst shows no evidence of crys-
talline phases, indicating either that the Mo-containing phase
remains amorphous or that the crystallite size is below the de-
tection limit of XRD. Based on the foregoing discussion, no
explanation is apparent for the dramatically lower HDS ac-
tivity of the sulfided Mo-O-B/SiO2 catalyst, which has a low
B content incorporated during NaBH4 (partial) reduction of
a (NH4)6Mo7O24/SiO2 precursor, relative to a B-free sulfided
Mo/SiO2 catalyst.
passive oxide layer) in the Ni-B/SiO2 catalyst has a Ni(2p3/2
)
binding energy (851.5 eV) below that of Ni metal (853.0 eV).
As has been suggested before, B associated with Ni likely do-
nates electron density to Ni atoms in the silica-supported Ni-B.
Thus, although no crystalline B-containing phases are observed
on annealing our Ni-B/SiO2 catalyst, the substantial B con-
tent in the amorphous Ni-B particles significantly modifies the
properties of the Ni. The Ni-B/SiO2 catalyst, when presulfided,
had a thiophene HDS activity nearly twice that of a sulfided
Ni/SiO2 catalyst after 24 h on stream. This observation is not
surprising given that the sulfided Ni-B/SiO2 catalyst had an O2
chemisorption capacity 3.5 times higher than that of the sulfided
Ni/SiO2 catalyst. TEM images of an as-prepared Ni-B/SiO2
catalyst (e.g., Fig. 1b) reveal relatively large amorphous Ni-B
particles on the support (20–40 nm), so the high chemisorption
capacity of this catalyst appears unrelated to a high dispersion
of the Ni-B phase. Apparently, the high B content of this cata-
lyst (B/Ni = 0.61) facilitates a high site density on the sulfided
Ni-B/SiO2 catalyst (relative to the sulfided Ni/SiO2 catalyst)
despite the relatively large particle size. An XRD pattern of a
sulfided Ni-B/SiO2 catalyst (Fig. 14) indicates the formation
of some Ni9S8 in the catalyst as a result of a sulfidation pre-
treatment at 650 K. Either this Ni sulfide phase formed on the
Ni-B particles is highly dispersed or a significant number of ad-
sorption sites exist on unsulfided or partially sulfided portions
of the supported Ni-B particles. Suslick et al. [6] observed un-
supported Ni3B to become partially converted to Ni9S8 after
thiophene HDS for 20 h at the high temperature of 723 K and
fully converted to Ni9S8 after sulfidation in a 10 mol% H2S/H2
mixture at 723 K for 12 h. The HDS activity of the unsupported
Ni3B steadily increased over 15 h on stream, which the authors
concluded was due to its conversion to the more catalytically
The most important observation of this study is that Ni-Mo-
O-B/SiO2 catalysts with Ni/Mo > ∼0.5, prepared by NaBH4
reduction of metal salt precursors, have higher thiophene HDS
activities than conventionally prepared Ni-Mo/SiO2 catalysts.
For example, a Ni-Mo-O-B/SiO2 catalyst with Ni/Mo = 0.99,
when presulfided, had an HDS activity 1.4 times higher than
a sulfided Ni-Mo/SiO2 catalyst with the optimal molar ratio
of Ni/Mo = 0.5. These observations are somewhat surprising
given that the unpromoted Mo-O-B/SiO2 catalyst is six times
less active than a conventionally prepared sulfided Mo/SiO2
catalyst, and both amorphous materials crystallize to give the
predominant Mo-containing phase of MoO2. However, the on-
set of MoO2 crystallization is shifted >150 degrees higher (to

G.L. Parks et al. / Journal of Catalysis 246 (2007) 277–292
291
∼790 K) for the Ni-Mo-O-B/SiO2 catalyst. Similarly, the crys-
tallization of Ni shifts approximately 150 degrees to higher
temperature (∼675 K) for the Ni-Mo-O-B/SiO2 catalyst com-
pared with a Ni-B/SiO2 catalyst. These results (as well as those
for the unsupported materials) indicate that the simultaneous
reduction of Ni and Mo salts to give a Ni-Mo-O-B phase on
the silica support yields an amorphous material that is sub-
stantially more resistant to crystallization than when the metal
salts are reduced alone. It is important to note, however, that
whereas the unsupported and silica-supported Ni-Mo-O-B ma-
terials are apparently homogeneous in the as-prepared form,
they do still phase-segregate on heating in flowing He to give
principally Ni metal and MoO2, the phases formed from heat-
ing the monometallic amorphous metal-boron phases. High-
resolution TEM images are consistent with the as-prepared un-
supported and silica-supported Ni-Mo-O-B materials being ho-
mogeneous as they reveal particles of uniform morphology. The
Ni in the Ni-Mo-O-B/SiO2 catalyst (Ni/Mo = 0.99) is appar-
ently quite well dispersed as the conventional and time-resolved
XRD patterns exhibit only very weak peaks associated with Ni
metal (and no evidence for Ni-B crystalline phases). The higher
dispersion of Ni in the Ni-Mo-O-B/SiO2 catalysts than in the
Ni-Mo/SiO2 catalysts is further supported by the IR spectral re-
sults for adsorbed CO on the catalysts. For the sulfided catalysts
(Fig. 11b), the absorbance feature assigned to Ni-promoted
Moδ+ sites is substantially larger for the Ni-Mo-O-B/SiO2 cat-
alyst than for the Ni-Mo/SiO2 catalyst, and the spectrum of
the latter also shows an absorbance feature associated with un-
promoted Mo2+ sites that is not apparent in the spectrum of
the sulfided amorphous metal-boron catalyst. Although directly
comparing the IR spectra for the reduced catalysts (Fig. 11a)
is more difficult, because some of the Ni in the Ni-Mo-O-
B/SiO2 catalyst is already reduced in the as-prepared catalysts
before the H2 treatment (as indicated by XPS), the dramatically
larger absorbance features in the νCO region associated with
CO linearly bonded to Ni0 for the Ni-Mo-O-B/SiO2 catalyst
also point to a higher Ni dispersion for this catalyst relative to
the Ni-Mo/SiO2 catalyst. Therefore, we conclude that the use of
NaBH4 to reduce the Ni and Mo salts yields Ni-Mo-O-B/SiO2
catalysts with very highly dispersed Ni species. This likely ex-
plains why the optimal Ni/Mo molar ratio of 0.99 for the Ni-
Mo-O-B/SiO2 catalysts is twice that of the conventionally pre-
pared Ni-Mo/SiO2 catalyst, for which the optimal Ni/Mo molar
ratio is 0.50, because more Ni that effectively promotes Mo
can be accommodated in catalysts prepared via NaBH4-based
synthesis. It follows that the Ni-Mo-O-B/SiO2 catalyst with
the optimal Ni/Mo molar ratios is significantly more active for
thiophene HDS compared with the conventionally prepared Ni-
Mo/SiO2 catalyst with its optimal Ni/Mo molar ratio. An XRD
pattern of a sulfided Ni-Mo-O-B/SiO2 catalyst (Ni/Mo = 0.99)
reveals a small peak at 21.5◦ suggesting that a small amount
of Ni has been converted to Ni9S8, a phase with low HDS
activity. Previous studies of alumina-supported Co3Mo3N and
Ni3Mo3N materials sulfided under similar conditions showed
substantially stronger XRD peaks for Co9S8 in the former and
MoS2 in both catalysts [13,57]. By comparison, the amorphous
Ni-Mo-O-B/SiO2 catalyst is less susceptible to the formation of
crystalline sulfide phases as a result of sulfidation pretreatment.
This resistance, along with the high Ni dispersion, contributes
to the high HDS activity of the Ni-Mo-O-B/SiO2 catalysts.
What remains unclear is the chemical environment of B in
the Ni-Mo-O-B/SiO2 catalysts. The Ni-B and Ni-Mo-O-B ma-
terials have similar boron-to-metal molar ratios (B/Me ∼0.5)
and similar XPS spectra in the Ni(2p) and B(1s) regions. How-
ever, whereas unsupported Ni-B crystallizes to give Ni3B start-
ing at 633 K, unsupported Ni-Mo-O-B begins to crystallize
at 550 K to give Ni metal, and no crystalline B-containing
phases are formed at higher temperatures. Apparently, B is less
strongly associated with Ni in Ni-Mo-O-B than in Ni-B, result-
ing in crystallization of the Ni at a lower temperature. For the
silica-supported phases, the greatest difference between Ni-B
and Ni-Mo-O-B is that very little crystalline Ni is formed in the
Ni-Mo-O-B/SiO2 on annealing, whereas a substantial Ni peak
forms for the Ni-B/SiO2, again reinforcing the conclusion that
Ni is highly dispersed in the former catalyst. Finally, although
no definitive conclusions can be drawn about the chemical en-
vironment of B in the catalysts, given the significant O content
of the unsupported and silica-supported Ni-Mo-O-B (relative
to the Ni-B materials) and the foregoing discussion, it seems
likely that at least some of the B is bonded primarily to O in
a form that does not yield a crystalline phase on annealing in
flowing He.
5. Conclusion
Sulfided Ni-B/SiO2 and Ni-Mo-O-B/SiO2 catalysts, in which
the precursors were prepared by NaBH4 reduction of silica-
supported metal salts, had significantly higher thiophene HDS
activities than conventionally prepared sulfided Ni/SiO2 and
Ni-Mo/SiO2 catalysts. On the other hand, a sulfided Mo-O-
B/SiO2 catalyst had dramatically lower HDS activity than
a sulfided Mo/SiO2 catalyst. Based on the results of vari-
ous physicochemical characterization measurements, including
time-resolved XRD, the high HDS activities of sulfided Ni-
B/SiO2 and Ni-Mo-O-B/SiO2 catalysts are traced to a high
density of active sites on these materials compared with con-
ventionally prepared catalysts.
Acknowledgments
This research was supported by the National Science Foun-
dation under grant CHE-0101690 and the Camille and Henry
Dreyfus Scholar/Fellow Program for Undergraduate Institu-
tions. Some of the research described in this paper (TEM, XPS)
was performed in the Environmental Molecular Sciences Lab-
oratory (EMSL), a national scientific user facility sponsored
by the Department of Energy’s Office of Biological and Envi-
ronmental Research and located at Pacific Northwest National
Laboratory. The research carried out at the NSLS beam line
X7b was supported under contract DE-AC02-98CH10886 with
the U.S. Department of Energy, Office of Basic Energy Sci-
ences, Chemical Science Division. The NSLS is supported by
the Divisions of Materials and Chemical Sciences of the U.S.
Department of Energy.

292
G.L. Parks et al. / Journal of Catalysis 246 (2007) 277–292
References
[28] Y.-Z. Chen, B.-J. Liaw, S.-J. Chiang, Appl. Catal. A 284 (2005) 97.
[29] J. Fang, X. Chen, S. Yan, M. Qiao, H. Li, H. He, K. Fan, J. Catal. 229
(2005) 97.
[30] S.-T. Wong, J.-F. Lee, J.-M. Chen, C.-Y. Mou, J. Mol. Catal. A 165 (2001)
159.
[31] H. Li, H. Li, W.-L. Dai, J.-F. Deng, Appl. Catal. A 207 (2001) 151.
[32] R. Paul, P. Buisson, N. Joseph, Ind. Eng. Chem. 44 (1952) 1006.
[33] H. Li, H. Li, J.-F. Deng, Catal. Today 74 (2002) 53.
[34] P. Dufresne, E. Payen, J. Grimblot, J.P. Bonnelle, J. Phys. Chem. 85 (1981)
2344.
[1] H.I. Schlessinger, H.C. Brown, A.E. Finholt, J.R. Gilbreath, J. Am. Chem.
Soc. 75 (1953) 215.
[2] R.C. Wade, Catal. Rev. Sci. Eng. 14 (1976) 211.
[3] J.-F. Deng, H. Li, W. Wang, Catal. Today 51 (1999) 113.
[4] C. Luo, W.-N. Wang, M.-H. Qiao, K.-N. Fan, J. Mol. Catal. A 184 (2002)
379.
[5] Q.-Y. Cheng, W. Li, J. Wu, N.-J. Guan, K.-Y. Tao, Ran Liao Hua Xue Xue
Bao 28 (2000) 249.
[35] D. Briggs, M.P. Seah (Eds.), Practical Surface Analysis by Auger and X-
ray Photoelectron Spectroscopy, Wiley, New York, 1983.
[36] Y. Okamoto, Y. Nitta, T. Imanaka, S. Teranishi, J. Chem. Soc. Faraday
Trans. 75 (1980) 2027.
[6] S. Skrabalak, K. Suslick, Chem. Mater. 18 (2006) 3103.
[7] H. Topsøe, B. Clausen, F.E. Massoth, in: J.R. Anderson, M. Boudart
(Eds.), Catalysis: Science and Technology, vol. 11, Springer-Verlag, Ber-
lin, 1996, p. 1.
[37] J.A. Schreifels, P.C. Maybury, W.E. Swartz, J. Catal. 65 (1980) 195.
[38] J. Li, M. Qiao, J.-F. Deng, J. Mol. Catal. A 169 (2001) 295.
[39] C.D. Wagner, A.V. Naumkin, A. Kraut-Vass, J.W. Allison, C.J. Powell,
J.R. Rumble Jr., NIST X-ray Photoelectron Spectroscopy Database, ver-
sion 3.4, vol. 2005, National Institute of Standards and Technology, 2003.
[40] Y. Okamoto, K. Fukino, T. Imanaka, S. Teranishi, J. Catal. 74 (1982) 173.
[41] H. Belatel, H. Al-Kandari, F. Al-Khorafi, A. Katrib, F. Garin, Appl. Catal.
A 275 (2004) 141.
[42] JCPDS Powder Diffraction File, International Centre for Diffraction Data,
Swarthmore, PA, 2000.
[43] Z. Jiang, H. Yang, Z. Wei, Z. Xie, W. Zhong, S. Wei, Appl. Catal. A 279
(2005) 165.
[44] Y. He, M. Qiao, H. Hu, J.-F. Deng, K. Fan, Appl. Catal. A 228 (2002) 29.
[45] C.H. Rochester, R.J. Terrel, J. Chem. Soc. Faraday Trans. 73 (1977) 609.
[46] D.G. Blackmond, E.I. Ko, J. Catal. 96 (1985) 210.
[47] J.B. Peri, J. Phys. Chem. 86 (1982) 1615.
[48] J.L.G. Fierro, A. Lopez Agudo, L.G. Tejuca, J. Chem. Soc. Faraday Trans.
81 (1985) 1203.
[49] A.A. Davydov, Infrared Spectroscopy of Adsorbed Species on the Surface
of Transition Metal Oxides, Wiley, New York, 1990.
[50] E.C. DeCanio, J.C. Edwards, T.R. Scalzo, D.A. Storm, J.W. Bruno,
J. Catal. 132 (1991) 498.
[51] E.C. DeCanio, D.A. Storm, J. Catal. 132 (1991) 375.
[52] M. Angulo, F. Mauge, J.C. Duchet, J.C. LaValley, Bull. Soc. Chim. Belg.
96 (1987) 925.
[8] J. Ramirez, P. Castillo, L. Cedeno, R. Cuevas, M. Castillo, J.M. Palacios,
A. Lopez-Agudo, Appl. Catal. A 132 (1995) 317.
[9] D. Li, T. Sato, M. Imamura, H. Shimada, A. Nishijima, J. Catal. 170 (1997)
357.
[10] Usman, T. Kubota, Y. Araki, K. Ishida, Y. Okamoto, J. Catal. 227 (2004)
523.
[11] G. Muralidhar, F.E. Massoth, J. Shabtai, J. Catal. 85 (1984) 44.
[12] T. Kabe, A. Ishihara, W. Qian, Hydrodesulfurization and Hydrodenitro-
genation: Chemistry and Engineering, Wiley–VCH, Weinheim, 1999.
[13] B. Diaz, S.J. Sawhill, D.H. Bale, R. Main, D.C. Phillips, S. Korlann,
R. Self, M.E. Bussell, Catal. Today 86 (2003) 191.
[14] D.C. Phillips, S.J. Sawhill, R. Self, M.E. Bussell, J. Catal. 207 (2002) 266.
[15] J.A. Rodriguez, J.-Y. Kim, J.C. Hanson, S.J. Sawhill, M.E. Bussell,
J. Phys. Chem. B 107 (2003) 6276.
[16] S.J. Sawhill, D.C. Phillips, M.E. Bussell, J. Catal. 215 (2003) 208.
[17] S.J. Sawhill, K.A. Layman, D.R. Van Wyk, M.H. Engelhard, C. Wang,
M.E. Bussell, J. Catal. 231 (2005) 300.
[18] P.A. Aegerter, W.W.C. Quigley, G.J. Simpson, D.D. Ziegler, J.W. Logan,
K.R. McCrea, S. Glazier, M.E. Bussell, J. Catal. 164 (1996) 109.
[19] K.R. McCrea, J.W. Logan, T.L. Tarbuck, J.L. Heiser, M.E. Bussell,
J. Catal. 171 (1997) 255.
[20] G.N. Glavee, K.J. Klabunde, C.M. Sorenson, Langmuir 10 (1994) 4726.
[21] P. Norby, J.C. Hanson, Catal. Today 39 (1998) 301.
[22] P.J. Chupas, M.F. Ciraolo, J.C. Hanson, C.P. Grey, J. Am. Chem. Soc. 123
(2001) 1694.
[53] F. Maugé, J.C. Lavalley, J. Catal. 137 (1992) 69.
[54] B. Müller, A.D. van Langeveld, J.A. Moulijn, H. Knözinger, J. Phys.
Chem. 97 (1993) 9028.
[55] B.M. Vogelaar, P. Steiner, A.D. van Langeveld, S. Eijsbouts, J.A. Moulijn,
Appl. Catal. A 251 (2003) 85.
[23] X. Wang, J.C. Hanson, G. Liu, J.A. Rodriguez, A. Iglesias-Juez, M. Fer-
nandez-Garcia, J. Chem. Phys. 121 (2004) 5434.
[24] A.P. Hammersley, S.O. Svensson, A. Thompson, Nucl. Instrum. Methods
Phys. Res. A 346 (1994) 312.
[25] Research Systems, Inc., Boulder, CO, 2002.
[56] C. Thomazeau, V. Martin, P. Afanasiev, Appl. Catal. A 199 (2000) 61.
[57] J.W. Logan, J.L. Heiser, K.R. McCrea, B.D. Gates, M.E. Bussell, Catal.
Lett. 56 (1998) 165.
[26] A.L. Diaz, M.E. Bussell, J. Phys. Chem. 97 (1993) 470.
[27] H. Li, Y. Wu, Y. Wan, J. Zhang, W. Dai, M. Qiao, Catal. Today 93–95
(2004) 493.