Carbazole hydrodenitrogenation over nickel phosphide and Ni-rich bimetallic phosphide catalysts

Article abstract of DOI:10.1016/j.apcata.2014.05.026

Silica-supported nickel phosphide (Ni₂P) and Ni-rich bimetallic phosphide catalysts were investigated for the hydrodenitrogenation (HDN) of carbazole with and without a benzothiophene co-feed. The Ni₂P/ SiO₂, Co_0.1Ni_1.9P/SiO₂, and Fe _0.03Ni_1.97P/SiO₂ catalysts exhibited high carbazole HDN activities and out-performed a commercial sulfided Ni-Mo/Al ₂O₃ catalyst under the testing conditions employed. Co-feeding of benzothiophene inhibited carbazole HDN over the metal phosphides, but the Ni₂P/SiO₂ and Ni-rich bimetallic phosphide catalysts maintained higher activities than the sulfided Ni-Mo/Al ₂O₃ catalyst. The product selectivities of the metal phosphide and sulfide catalysts were similar, except that the sulfided Ni-Mo/Al₂O₃ catalyst yielded more hydrogenated carbazoles while the metal phosphide catalysts produced more of the partially hydrogenated hydrocarbon cyclohexylbenzene. The metal phosphide catalysts favored ring-opened (hexylcyclohexane) and ring-contracted ((2-methylcyclopentyl)cyclohexane) products, indicating the presence of metal and Br?nsted acid sites on the catalyst surfaces. The Ni₂P/SiO₂ and bimetallic phosphide catalysts exhibited excellent stabilities in the HDN conditions employed (carbazole-only and mixed feeds), as indicated by phase purity, average crystallite size and resistance to sulfur incorporation of the supported metal phosphide particles.

Full text of DOI:10.1016/j.apcata.2014.05.026

Accepted Manuscript
Title: Carbazole Hydrodenitrogenation over Nickel Phosphide
and Ni-rich Bimetallic Phosphide Catalysts
Author: Richard H. Bowker Boris Ilic Bo.A. Carrillo Michael
A. Reynolds Brendan D. Murray Mark E. Bussell
PII:
S0926-860X(14)00360-3
DOI:
Reference:
http://dx.doi.org/doi:10.1016/j.apcata.2014.05.026
APCATA 14846
To appear in:
Applied Catalysis A: General
Received date:
Revised date:
Accepted date:
10-3-2014
20-5-2014
23-5-2014
Please cite this article as: R.H. Bowker, B. Ilic, Bo.A. Carrillo, M.A. Reynolds,
B.D. Murray, M.E. Bussell, Carbazole Hydrodenitrogenation over Nickel Phosphide
and Ni-rich Bimetallic Phosphide Catalysts, Applied Catalysis A, General (2014),
http://dx.doi.org/10.1016/j.apcata.2014.05.026
This is a PDF file of an unedited manuscript that has been accepted for publication.
As a service to our customers we are providing this early version of the manuscript.
The manuscript will undergo copyediting, typesetting, and review of the resulting proof
before it is published in its final form. Please note that during the production process
errors may be discovered which could affect the content, and all legal disclaimers that
apply to the journal pertain.

Submitted to Applied Catalysis A, March 10, 2014; Revised May 19, 2014.
Carbazole Hydrodenitrogenation over
Nickel Phosphide and Ni-rich Bimetallic Phosphide Catalysts
Richard H. Bowker,1,2 Boris Ilic, 1,2 Bo A. Carrillo,1,2
Michael A. Reynolds,3 Brendan D. Murray3 and Mark E. Bussell1,2*
1Department of Chemistry, MS-9150
Western Washington University
Bellingham, WA 98225
2Advanced Materials Science and Engineering Center
Western Washington University
Bellingham, WA 98225
3Shell Oil Company
Westhollow Technology Center
3333 Highway 6 South
Houston, TX 77082
*Author to whom correspondence should be addressed.
Correspondence Address
Prof. Mark E. Bussell
Department of Chemistry, MS-9150
Western Washington University
516 High Street
Bellingham, WA 98225
Tel: 360-650-3145
Fax: 360-650-2826
E-mail: Mark.Bussell@wwu.edu
Page 1 of 38

Highlights
• Nickel-based phosphide catalysts out-performed a commercial Ni-Mo catalyst for
carbazole HDN.
• Nickel phosphide catalysts exhibited excellent stability in a carbazole /
benzothiophene feed.
• Nickel-based phosphide catalysts favored ring-opened and ring-contracted
products.
• Ni2P produced more of the partially hydrogenated hydrocarbon
cyclohexylbenzene than a Ni-Mo catalyst.
ABSTRACT
Silica-supported nickel phosphide (Ni2P) and Ni-rich bimetallic phosphide catalysts were
investigated for the hydrodenitrogenation (HDN) of carbazole with and without a
benzothiophene co-feed. The Ni2P/SiO2, Co0.1Ni1.9P/SiO2, and Fe0.03Ni1.97P/SiO2 catalysts
exhibited high carbazole HDN activities and out-performed a commercial sulfided Ni-Mo/Al2O3
catalyst under the testing conditions employed. Co-feeding of benzothiophene inhibited
carbazole HDN over the metal phosphides, but the Ni2P/SiO2 and Ni-rich bimetallic phosphide
catalysts maintained higher activities than the sulfided Ni-Mo/Al2O3 catalyst. The product
selectivities of the metal phosphide and sulfide catalysts were similar, except that the sulfided
1
Page 2 of 38

Ni-Mo/Al2O3 catalyst yielded more hydrogenated carbazoles while the metal phosphide catalysts
produced more of the partially hydrogenated hydrocarbon cyclohexylbenzene. The metal
phosphide catalysts favored ring-opened (hexylcyclohexane) and ring-contracted ((2-
methylcyclopentyl)cyclohexane) products, indicating the presence of metal and Brönsted acid
sites on the catalyst surfaces. The Ni2P/SiO2 and bimetallic phosphide catalysts exhibited
excellent stabilities in the HDN conditions employed (carbazole-only and mixed feeds), as
indicated by phase purity, average crystallite size and resistance to sulfur incorporation of the
supported metal phosphide particles.
KEYWORDS
Hydrodenitrogenation, HDN, carbazole, HDS, benzothiophene, nickel phosphide, bimetallic
phosphide, cobalt nickel phosphide, iron nickel phosphide
2
Page 3 of 38

1. INTRODUCTION
The removal of sulfur and nitrogen impurities from fossil fuel feedstocks is a critical
processing step in producing ultralow sulfur transportation fuels [1,2]. The removal of
organonitrogen compounds is necessary to achieve the ultralow sulfur levels since these
compounds inhibit sulfur removal [3,4]. In addition, organonitrogen compounds poison catalysts
used in down-stream refinery processes such as hydrocracking, catalytic cracking and reforming
[5]. While sulfided Co-Mo/Al2O3 and Ni-Mo/Al2O3 catalysts are the workhorses of commercial
hydrotreating processes, a considerable research effort is ongoing to assess the potential of
alternate catalytic materials such as metal phosphides for the hydrodenitrogenation (HDN) and
hydrodesulfurization (HDS) reactions [6,7]. A number of laboratories have investigated the
HDS properties of the monometallic phosphides of the first-row transition metals (e.g. Ni2P) and
of MoP and WP, typically supported on silica, as well as of bimetallic phosphide materials.
Among the monometallic phosphides, Ni2P has generally been identified as being the most active
for HDS of thiophene, dibenzothiophene (DBT) and 4,6-dimethyldibenzothiophene (4,6-
DMDBT) [6,8-12]. Some bimetallic phosphide catalysts, Ni2P containing a small amount of a
second first-row metal (e.g. Fe0.03Ni1.97P/SiO2 [13]), for example, have exhibited HDS activities
that are higher than those of the binary phosphides of the same metals.
The HDN properties of metal phosphides have been less studied than have the HDS
properties of these materials, but for feedstocks rich in nitrogen such as the Canadian oil sands
[14] it is at least as important to assess the nitrogen removal capabilities of phosphide-based
hydrotreating catalysts. Depending upon whether the metal phosphide was in the bulk phase or
supported, and the organonitrogen compound used, MoP, Co2P and Ni2P have exhibited high
HDN activity [8,15-17]. Consistent with what was observed for HDS, some bimetallic
3
Page 4 of 38

phosphides (NixMoP, Co0.4Ni2P) have been observed to have HDN activities higher than those of
monometallic phosphide catalysts [18,19].
While metal phosphides have exhibited promising HDS and HDN properties, evaluating
their potential necessitates that they be directly compared to commercially available catalysts
employed in current industrial hydrotreating applications. This study addresses this issue as
metal phosphide phases supported on silica are compared to a commercial Ni-Mo/Al2O3 catalyst
for the HDN of carbazole with and without a benzothiophene co-feed. Ni2P and nickel-rich
bimetallic phosphides (e.g. Fe0.03Ni1.97P, Co0.1Ni1.9P) show excellent promise for development
into a new generation of commercial hydrotreating catalysts for hydrocarbon feeds containing
high levels of nitrogen impurities.
2. EXPERIMENTAL METHODS
2.1 Catalyst Synthesis
25 wt% Ni2P/SiO2, Fe0.03Ni1.97P/SiO2 and Co0.1Ni1.9P/SiO2 catalysts were prepared from
precursors having phosphorus-to-metal molar ratios (P/M) of 0.9 or 1.0 as described below.
Catalyst precursors were prepared either by sequential or co-impregnation, using ammonium
dihydrogen phosphate as the phosphorus salt. Following impregnation, the catalyst precursors
were calcined and subjected to temperature programmed reduction (TPR) to give the metal
phosphide catalysts.
Ni2P/SiO2. The precursor of a 25 wt% Ni2P/SiO2 catalyst (19.8 wt% Ni) having P/Ni = 0.9 was
prepared by incipient wetness impregnation of a solution consisting of 1.9599 g nickel nitrate
hexahydrate (Ni(NO3)2·6H2O, Alfa Aesar, 99.9985%) and 0.6978 g of ammonium dihydrogen
phosphate (NH4H2PO4, Alfa Aesar, 98%) onto 1.50 g silica support (Cab-O-Sil, M7D). Prior to
impregnation, the pH of the solution was adjusted to ~2.1 to inhibit precipitate formation. The
catalyst precursor was dried at 393 K for 1 h between impregnations and 24 h after the final
4
Page 5 of 38

solution addition. The dried catalyst precursor was calcined in air for 3 h at 773 K, heating from
room temperature at 40 K/min. Following calcination and cooling to room temperature,
approximately 0.5 g of the precursor was transferred to a quartz U-tube and degassed in 60
mL/min He (Airgas, 99.999%) for 30 min. The precursor was then subjected to TPR in H2
(Airgas, 99.999%) flowed at 150 mL/min by heating from room temperature to 923 K at a
heating rate of 1 K/min. Following TPR, the Ni2P/SiO2 catalyst was cooled to room temperature
in continued H2 flow, purged for 30 min in 60 mL/min He, and then passivated in 60 mL/min 1
mol% O2/He (Airgas, 99%) for 2 h.
Fe0.03Ni1.97P/SiO2. The precursor of a 25 wt% Fe0.03Ni1.97P/SiO2 catalyst (0.24 wt% Fe, 19.6
wt% Ni) was prepared by impregnation of a solution containing 1.9366 g nickel nitrate and
0.0341 g iron nitrate nonahydrate (Fe(NO3)3·9H2O, Fischer, ACS grade) dissolved in 5 mL
ultrapure water (18 MΩ resistance) onto 1.50 g of silica. The supported metal precursor was
dried at 393 K between impregnations and 24 h after the final addition. A solution containing
0.7758 g ammonium dihydrogen phosphate dissolved in 4 mL ultrapure water was then
impregnated onto the silica-supported iron-nickel nitrate precursor as described above to give a
precursor having a P/M = 1.0. Prior to impregnation, the pH of the solutions was adjusted to
~2.1. Following drying in a 393 K oven for 24 h, the catalyst precursor was calcined, reduced
via TPR, and passivated as described above.
Co0.1Ni1.9P/SiO2. The precursor of a 25 wt% Co0.1Ni1.9P/SiO2 catalyst (0.96 wt% Co, 18.9 wt%
Ni) was prepared in a similar manner as the Ni2P/SiO2 catalyst, except that the impregnating
solution contained 1.8628 g nickel nitrate hexahydrate, 0.0596 g cobalt nitrate hexahydrate
(Co(NO3)2·6H2O, Alfa Aesar, ACS grade) and 0.6978 g of ammonium dihydrogen phosphate.
Prior to impregnation, the pH of the solution was adjusted to ~2.1 to inhibit precipitate
formation. The solution was impregnated onto 1.50 g of silica support to give a precursor having
5
Page 6 of 38

P/M = 0.9. Following drying in a 393 K oven for 24 h, the catalyst precursor was calcined,
reduced via TPR, and passivated as described above.
Ni-Mo/Al2O3. The Ni-Mo/Al2O3 catalyst used in this study was provided by Shell Oil Company
(Shell 424) and had a composition of 4 wt% NiO, 19.5 wt% MoO3 and 8 wt% P2O5 in its oxidic
form. The Ni-Mo/Al2O3 catalyst was sulfided prior to use as described below.
2.2 Catalyst Characterization
X-ray diffraction (XRD) patterns of the as-prepared and HDN-tested Ni2P/SiO2,
Fe0.03Ni1.97P/SiO2 and Co0.1Ni1.9P/SiO2 catalysts were obtained using a PANalytical X’Pert Pro
diffractometer equipped with a monochromatic Cu-Kα source (λ = 1.54050 Å). BET surface area
measurements were acquired using a Micromeritics PulseChemisorb 2700 instrument.
Approximately 0.1000 g of catalyst was placed in a quartz sample tube and degassed at room
temperature in a 60 mL/min He flow for 30 min. The sample was treated in a flow of He (45
mL/min) for 2 h at 623 K and then cooled to room temperature in a continued He flow. The
BET measurements were carried out as described elsewhere [20].
Carbon monoxide (CO) and oxygen (O2) pulsed chemisorption measurements were also
obtained using the Micromeritics PulseChemisorb 2700 instrument. Approximately 0.1000 g of
catalyst was degassed in 60 mL/min He at room temperature for 30 min. Prior to the
measurements, the samples were either reduced (metal phosphides) or sulfided (Ni-Mo/Al2O3)
in-situ. For reduction, samples were heated from room temperature to 650 K in a 60 mL/min
flow of H2 and held at this temperature for 2 h. For sulfidation, samples were heated from room
temperature to 650 K in a 60 mL/min flow of a 3.0 mol% H2S/H2 mixture, held at this
temperature for 2 h, and then reduced in a 60 mL/min flow of H2 at 623 K for 1 h. All of the
samples were then degassed in 45 mL/min He at 673 K for 1 h prior to the chemisorption
measurement. CO chemisorption capacities were measured by injecting a calibrated sample
6
Page 7 of 38

volume of CO gas (Messer, 99.99%) at 1 min intervals into an He flow (45 sccm/min) passing
over the catalyst sample until CO uptake ceased. Prior to injection, the CO was passed through a
coil of 1/8” stainless steel tubing submerged in a pentane slush (142 K) to remove metal carbonyl
impurities. Catalyst samples were maintained at a temperature of 273 K during CO
chemisorption measurements. The O2 chemisorption capacity measurements were carried out at
196 K using a procedure described elsewhere [20]. A 10.3 mol% O2/He mixture (Airco) was
used to obtain the O2 chemisorption capacity measurements.
X-ray photoelectron spectroscopy (XPS) spectra were acquired using a SAGE 100 XPS
System (SPECS GmbH) outfitted with a Mg Kα X-ray source (1253.6 eV) and a 95 mm
hemispherical energy analyzer. The XPS measurements were carried out on the as-prepared
Ni2P/SiO2, Fe0.03Ni1.97P/SiO2, Co0.1Ni1.9P/SiO2 catalysts following passivation and transfer
through air to the spectrometer. The spectra were collected with a pass energy of 15 eV.
Binding energies were corrected for sample charging using the C(1s) peak at 284.6 eV for
adventitious carbon as a reference. Surface compositions were determined for the metal
phosphide catalysts using SpecsLab software and the peak areas in the the Ni(3p), Co(3p),
Fe(3p) and P(2p) regions.
Samples of the HDN-tested catalysts were collected for X-ray diffraction and bulk carbon
and sulfur analyses. Following completion of an HDN activity measurement, the flow of liquid
feed was stopped while maintaining the H2 gas flow for 1 h at the reaction temperature. The
reactor was then cooled to room temperature, depressurized, and then purged with 60 mL/min He
for 30 min. The system was then opened and allowed to sit for a minimum of 3 h to permit the
slow exposure of the catalyst to ambient air. Bulk carbon and sulfur analyses of HDN-tested
catalyst samples were carried out using a LECO SC-144DR Sulfur and Carbon Analyzer.
Approximately 0.10 g of catalyst was transferred into a ceramic boat, which was then loaded into
7
Page 8 of 38

a furnace where the sample was combusted in an oxygen-rich environment at ~1625 K for 3 min.
Combusted carbon (CO2) and sulfur (SO2) that evolved from the catalyst sample was quantified
via IR detection and reported as wt% C and S. An additional ~0.05 g sample of the HDN-tested
catalysts was used to acquire an X-ray diffraction pattern as described above.
2.3 Carbazole HDN activity measurements
Carbazole HDN activity measurements were carried out using a fixed-bed, continuous
flow reactor operating at a total pressure of 3.0 MPa and temperatures in the range of 548-673 K.
The reactor feed consisted of either 1000 ppm carbazole and 500 ppm dodecane in a 39.85 wt%
p-xylene/decane solution or 3000 ppm benzothiophene, 1000 ppm carbazole, and 500 ppm
dodecane in a 39.55 wt% p-xylene/decane solution. The 500 ppm dodecane served as an internal
standard for gas chromatographic analysis of the reactor effluent. The liquid feed (5.4 mL/h)
was injected into a 50 mL/min flow of hydrogen and vaporized prior to entry into the reactor.
0.25 g of catalyst (16-20 mesh size) was diluted with quartz flakes to a total volume of 5 mL and
loaded into a reactor tube having a diameter of 1.1 cm and length of 40 cm. The internal reactor
temperature was monitored with an axially mounted thermocouple in direct contact with the
catalyst bed. Metal phosphide and bimetallic phosphide catalysts were pretreated by heating
from room temperature to 650 K in 1 h in a 60 mL/min flow of H2 and held at this temperature
for 2 h, then cooled to room temperature in continued H2 flow. A commercial Ni-Mo/Al2O3
catalyst (Shell 424) was subjected to a sulfidation pretreatment by heating from room
temperature to 650 K in 1 h in a 60 mL/min flow of 3.0 mol% H2S/H2 mixture, held at this
temperature for 2 h, and then cooled to room temperature in continued H2S/H2 flow.
After pretreatment, with the catalyst samples at room temperature, the reactor was
pressurized to 3.0 MPa with H2. The catalyst bed was heated to 548 K over 15 min in a 50
mL/min flow of H2 and liquid feed injection was begun once the reactor reached the operating
8
Page 9 of 38

temperature. The reactor was allowed to stabilize under operating conditions for 3 h prior to
sampling reactor effluent at 30 min intervals over 2.5 h. The catalyst temperature was raised 25
K, the reactor stabilized for 3 h, followed by sampling of the reactor effluent at 30 min intervals.
This procedure was repeated until sampling at the maximum catalyst temperature (623-673 K)
was completed. The first effluent sample collected at each temperature was discarded; the four
subsequent effluent samples were analyzed off-line using a gas chromatograph (Agilent 6890N)
equipped with an HP-5 column and a flame ionization detector. HDN product identification was
conducted by GC/MS and the GC was calibrated using serial diluted product standard solutions
to allow product quantification of reactor effluent samples. Based on this method, carbon
balances of 90-95% were achieved, indicating that the analyses of the reactor effluent accounted
for all of the major reaction products. The majority of the HDN activity measurements were
done in duplicate to ensure reproducibility.
3. RESULTS
X-ray diffraction patterns for the 25 wt% Ni2P/SiO2, Fe0.03Ni1.97P/SiO2 and
Co0.1Ni1.9P/SiO2 catalysts prepared in this research are shown in Figure 1 along with a reference
pattern for Ni2P from the powder diffraction file (card no. 089-2742 [21]). For the bimetallic
phosphide compositions investigated, solid solutions are formed [13,22-24], which is reflected in
the XRD patterns for the Fe0.03Ni1.97P/SiO2 and Co0.1Ni1.9P/SiO2 catalysts closely resembling that
of the Ni2P/SiO2 catalyst. There is no evidence in the XRD patterns for the formation of
impurity phases such as the more metal-rich Ni12P5 (card no. 22-1190 [21]). Using the Scherrer
equation and the full-width at half-maximum (FWHM) of the (111) reflection observed at
~40.8°, average crystallite sizes of 12-15 nm were calculated for the supported metal phosphide
particles. Additional characterization data for the catalysts, including BET surface areas and
chemisorption capacities, are listed in Table 1. The metal phosphide catalysts had similar
9
Page 10 of 38

surface areas and chemisorption (CO, O2) capacities; the O2 chemisorption capacities of the
metal phosphide catalysts were approximately four times greater than that of the sulfided Ni-
Mo/Al2O3 catalyst. The surface compositions of the as-prepared metal phosphide catalysts,
determined using XPS, are listed in Table 1. The as-prepared Ni2P/SiO2, Fe0.03Ni1.97P/SiO2 and
Co0.1Ni1.9P/SiO2 catalysts had surface compositions that were enriched in P relative to the
expected stoichiometry of the bulk phase materials (P/M = 0.50); this was particularly true for
the bimetallic phosphide catalysts. The amounts of Fe and Co at the surfaces of the
Fe0.03Ni1.97P/SiO2 and Co0.1Ni1.9P/SiO2 catalysts were below the detection limit of the XPS
instrument used. It should be noted that the surface compositions of the metal phosphide
catalysts under reaction conditions may differ from those of the passivated catalysts measured ex
situ. Previous studies in our laboratory of FexNi2-xP/SiO2 and CoxNi2-xP/SiO2 catalysts before
and after HDS testing revealed the tested catalysts to have more metal-rich compositions
(relative to the as-prepared catalysts) and to have low, but measureable S concentrations [13,24].
The silica-supported Ni2P and bimetallic phosphide catalysts were tested for carbazole
HDN conversion with and without a benzothiophene co-feed. The reaction network for
carbazole HDN is complex and includes a number of hydrogenated carbazole compounds
[3,19]). The reaction network shown in Scheme 1 is consistent with the hydrogenated carbazoles
and hydrocarbon products identified via GC-MS of the liquid reactor effluent at the different
reaction temperatures. In this study, the carbazole HDN conversion is defined as the conversion
of carbazole and hydrogenated carbazoles to hydrocarbon products that do not contain nitrogen.
For some of the catalysts investigated, the carbazole and carbazole HDN conversions differed
significantly due to the presence of hydrogenated carbazoles in the reactor effluent. Shown in
Figure 2 are the carbazole HDN conversions as a function of reaction temperature (498-623 K)
for the Ni2P/SiO2, Fe0.03Ni1.97P/SiO2 and Co0.1Ni1.9P/SiO2 catalysts as well as for the sulfided Ni-
10
Page 11 of 38

Mo/Al2O3 catalyst. All of the catalysts exhibited increasing HDN conversions with increasing
temperature, with the metal phosphide catalysts having conversions ±4% of each other at each
temperature. The Ni2P/SiO2, Fe0.03Ni1.97P/SiO2 and Co0.1Ni1.9P/SiO2 catalysts were more active
than the sulfided Ni-Mo/Al2O3 catalyst in the temperature range 573-623 K. At 623 K, the metal
phosphide catalysts had carbazole HDN conversions of 95% or higher while the sulfided Ni-
Mo/Al2O3 catalyst converted 75% of the carbazole to hydrocarbon products at this same
temperature. The HDN product selectivities for the catalysts will be discussed in the following
section. The average crystallite sizes for the metal phosphides in the HDN-tested catalysts were
unchanged, nor was there evidence for impurity phases in the XRD patterns (Figure 3 and
supporting information). Carbon and sulfur analysis of the HDN-tested catalyst samples showed
the metal phosphide catalysts to have carbon contents of 1-1.4 wt%. The HDN-tested sulfided
Ni-Mo/Al2O3 catalyst had C and S contents of 1.67 and 7.10 wt%, respectively; the high sulfur
content indicated that the catalyst remained sulfided despite the absence of sulfur compounds in
the reactor feed.
A second set of carbazole HDN experiments, with a 3000 ppm co-feed of
benzothiophene, were conducted for the Ni2P/SiO2, Fe0.03Ni1.97P/SiO2, Co0.1Ni1.9P/SiO2 and
sulfided Ni-Mo/Al2O3 catalysts and the HDN conversion vs. temperature data for these
experiments are plotted in Figure 4. Co-feeding of benzothiophene inhibited HDN of carbazole
for the three metal phosphide and sulfided Ni-Mo/Al2O3 catalysts, particularly at low reaction
temperatures. At 598 K, the Ni2P/SiO2, Fe0.03Ni1.97P/SiO2 and Co0.1Ni1.9P/SiO2 catalysts had
carbazole HDN conversions of 62, 48 and 66%, respectively, with the benzothiophene co-feed,
while these same catalysts had carbazole HDN conversions of 89, 96, and 89%, respectively, for
the carbazole-only feed. The sulfided Ni-Mo/Al2O3 catalyst had carbazole HDN conversions at
598 K of 58% for the mixed feed and 73% for the carbazole-only feed. At 623 K, however, the
11
Page 12 of 38

sulfided Ni-Mo/Al2O3 catalyst had a higher carbazole HDN conversion of 77% vs. 75% for the
carbazole-only feed, while mild inhibition of HDN activity was still observed for the metal
phosphide catalysts at this temperature (Tables 2 and 3). Consistent with the carbazole HDN
results for the carbazole-only feed, however, the metal phosphide catalysts were more active than
the sulfided Ni-Mo/Al2O3 catalyst at higher temperatures (623-673 K). At 648 K, for example,
the Ni2P/SiO2, Fe0.03Ni1.97P/SiO2, Co0.1Ni1.9P/SiO2 catalysts converted 97, 91 and 96% of the
carbazole to hydrocarbon products, while the sulfided Ni-Mo/Al2O3 catalyst had a carbazole
HDN conversion of 84%.
The benzothiophene HDS conversions for the mixed feed experiments over metal
phosphide and sulfided Ni-Mo/Al2O3 catalysts are shown in Figure 5. The metal phosphide and
sulfided Ni-Mo/Al2O3 catalysts exhibited high benzothiophene HDS conversions at temperatures
of 598 K and above. A detailed analysis of the HDS product selectivity was not carried out, but
the major products were identified as ethylbenzene and ethylcyclohexane.
In contrast to what was observed for the carbazole-only feed, the average crystallite sizes
for the metal phosphides increased slightly from 12-15 nm in the as-prepared catalysts to 14-16
nm in the HDN-tested samples when the mixed carbazole / benzothiophene feed was used.
There was no evidence for impurity phases in the XRD patterns of the HDN-tested catalysts
(Figure 3, supporting information). Carbon and sulfur analysis of the HDN-tested metal
phosphide catalysts revealed somewhat lower carbon contents of 0.67-1.27 wt% relative to the
carbazole-only feed and low, but measurable S contents of 0.16-0.35 wt%. The sulfided Ni-
Mo/Al2O3 catalyst also had a somewhat lower C content of 1.41 wt%, but a higher S content
(8.56 wt% S) than was observed for the carbazole-only feed.
The carbazole conversions and product selectivities (mixed carbazole / benzothiophene
feed) for the metal phosphide and sulfided Ni-Mo/Al2O3 catalysts are plotted as a function of
12
Page 13 of 38

temperature in Figures 6-9 and the selectivities for the carbzole-only and mixed feeds at 623 K
are compared in Table 4. As indicated by the product selectivity data, the reaction of carbazole
over the metal phosphide and sulfided Ni-Mo/Al2O3 catalysts yields four different hydrocarbon
products – cyclohexylbenzene (CHB), bicyclohexane (BCH), (2-methylcyclopentyl)cyclohexane
(MCPCH), and hexylcyclohexane (HCH) – as well as hydrogenated carbazoles (4H-CARB, 6H-
CARB, 12H-CARB). Biphenyl (BP) was not observed in significant amounts (≤ 0.2 mol%).
The reaction pathways leading to the different product species are summarized in the reaction
network shown in Scheme 1. Focusing initially on the sulfided Ni-Mo/Al2O3 catalyst,
hydrogenated carbazoles comprise 43 mol% of the products at 573 K but this amount decreases
to 5-7 mol% at 623-673 K. The major hydrocarbon product at low temperatures (573-598 K) is
bicyclohexane (30-35 mol%); however, the amount of BCH in the products decreases
substantially with increasing temperature.
At 623-673 K, the major product is (2-
methylcyclopentyl)cyclohexane (50-55 mol%), which is formed by ring contraction of one of the
six-member rings in BCH. The product selectivities for the metal phosphide catalysts exhibit
qualitatively similar trends as observed for the sulfided Ni-Mo/Al2O3 catalyst, but with some
important differences. The metal phosphides produced less hydrogenated carbazoles (< 3 mol%)
at high reaction temperatures (623-673 K), resulting in higher carbazole HDN conversions, and
yielded higher amounts of the partially hydrogenated product cyclohexylbenzene than did the
sulfided Ni-Mo catalyst. The Ni2P/SiO2, Fe0.03Ni1.97P/SiO2, Co0.1Ni1.9P/SiO2 catalysts had CHB
selectivities of 26, 30 and 32 mol% at 673 K, while the sulfided Ni-Mo/Al2O3 catalyst produced
just 13 mol% CHB at this temperature.
The HDN product selectivities are listed in Table 4 for the carbazole-only and mixed
feeds at a reaction temperature of 623 K, showing how benzothiophene affects product
selectivity. Significant differences in the HDN product selectivities are apparent when
13
Page 14 of 38

comparing the two different feeds, with similar changes observed for the sulfided Ni-Mo/Al2O3
and metal phosphide catalysts. When benzothiophene is present in the reactor feed, the
selectivity towards BCH decreased from 33-40 mol% to 22-29 mol%, while the selectivities
towards CHB, MCPCH and HCH increased by approximately the same amount in total. As
indicated in Scheme 1, MCPCH and HCH are products formed via ring contraction and ring
opening, respectively, of one of the six-member rings of BCH.
4. DISCUSSION
The catalytic removal of nitrogen in crude oil fractions via HDN reactions is critical in
the production of ultra-low sulfur fuels since organonitrogen compounds inhibit HDS and other
oil conversion reactions [3,5]. For hydrotreating reactions, strong adsorption of organonitrogen
molecules on catalyst surfaces blocks active sites for HDS and can limit the supply of available
hydrogen species [4]. This is particularly problematic for HDS of refractory sulfur compounds
such as 4,6-DMDBT. Nitrogen-containing compounds in crude oil can be divided into basic and
non-basic compounds, with the basic compounds consisting primarily of molecules containing
six-membered N-containing rings (e.g. quinoline) and the non-basic compounds generally
consisting of molecules having five-membered N-containing rings (e.g. carbazole). Studies have
shown that crude oil fractions from different sources contain different amounts and distributions
of organonitrogen compounds. Heavier feeds typically contain higher levels of nitrogen (and
sulfur) compounds [3]. Middle distillate cuts from shale oil and petroleum were observed to
contain predominantly pyridine compounds (six-membered N rings) while the same cut from a
Canadian oil sands-derived feed showed higher levels of indole and carbazole compounds (five-
membered N rings) than pyridines [3,25]. Carbazole and alkyl-substituted carbazoles are among
14
Page 15 of 38

the most refractory organonitrogen compounds and, for this reason, carbazole was selected as the
model compound for this HDN study.
Conventional Ni-Mo sulfide catalysts used in hydrotreating frequently contain
phosphorus as an additive, most frequently for HDN processing due to the stronger promotion of
HDN activity than of HDS. Indeed, the commercial Ni-Mo/Al2O3 catalyst used in this study
contained 8 wt% P2O5, which corresponds to a phosphorus-to-metal mole ratio (P/(Ni+Mo)) of
0.6. The properties of phosphorus as a promoter of sulfide-based hydrotreating catalysts have
been discussed in reviews by others [3,26,27]. Phosphorus is typically added to the
impregnation solution in the form of phosphate (e.g. H3PO4, NH4H2PO4), resulting in increased
solubility of molybdate salts via formation of phosphomolybdate species.
A number of
explanations have been advanced to explain the promotional effect of P on HDN activity
including morphological changes of the MoS2 crystallites (increased stacking of MoS2 layers),
leading to increased hydrogenation activity, and the formation of highly dispersed Ni and/or Mo
phosphides (or phosphosulfides) at the edges of the MoS2 under reaction conditions [3,26,27].
Oxide-supported metal phosphides represent a possible step-out catalyst development
technology for the removal of S and N impurities from crude oil feedstocks. The goal of the
current study was to compare the catalytic properties of silica-supported nickel phosphide and
Ni-rich bimetallic phosphides with those of a commercial Ni-Mo sulfide catalyst for the HDN of
carbazole. The key finding of this study is that Ni2P/SiO2 and Ni-rich bimetallic phosphide
(Fe0.03Ni1.97P/SiO2, Co0.1Ni1.9P/SiO2) are highly active and stable HDN catalysts, with properties
superior to those of a commercial sulfided Ni-Mo/Al2O3 catalyst under the reaction conditions
employed. Studies of carbazole HDN over unsupported and supported metal phosphides have
been reported [18,19]. For a series of unsupported NixMoP catalysts (0 ≤ x ≤ 1.11), Abu and
Smith measured significantly higher carbazole conversions (>85%) for bimetallic NixMoP
15
Page 16 of 38

phases than for MoP (54.5%) at 583 K on a mass of catalyst basis [19]. The most active
phosphide, Ni0.07MoP, produced mostly BCH (85.6%) with smaller amounts of 4H-CARB
(7.9%) and isomerized or cracked hydrocarbon products. The higher HDN activities of the
NixMoP catalysts relative to MoP were attributed to enhanced hydrogenation properties for the
Ni-containing phosphides. For alumina-supported phases, Co0.4Ni2P/Al2O3 and Ni0.3MoP/Al2O3
exhibited high HDN activity for carbazole, but the Mo-based phosphide was the more active of
the two [18]. This carried over for a mixed light gas oil feed for which the Ni0.3MoP/Al2O3
catalyst was more active and the Co0.4Ni2P/Al2O3 was less active than a sulfided Ni-Mo/Al2O3
catalyst.
In the present study, Ni2P/SiO2, Fe0.03Ni1.97P/SiO2, Co0.1Ni1.9P/SiO2 catalysts exhibited
higher carbazole HDN conversions than a sulfided Ni-Mo/Al2O3 catalyst (for the same mass of
catalyst) using both carbazole-only and mixed carbazole / benzothiophene feeds. Of particular
importance, the maximum carbazole HDN conversions achieved for the metal phosphide
catalysts (92-98%) were significantly higher than those of the sulfided Ni-Mo/Al2O3 catalyst (75-
84%) for both feeds. In addition, the nickel phosphide and Ni-rich bimetallic phosphide catalysts
yielded substantially higher amounts of partially-hydrogenated hydrocarbon products than did
the sulfide catalyst, therefore consuming less hydrogen while achieving the higher HDN
conversions. The metal phosphide catalysts also exhibited excellent stability, as indicated by no
loss of phase purity and minimal increase in average crystallite size, as well as minimal S
incorporation in the catalysts for the mixed carbazole / benzothiophene feed.
As indicated by the HDN conversion plots in Figures 2 and 4, the Ni2P/SiO2,
Fe0.03Ni1.97P/SiO2, Co0.1Ni1.9P/SiO2 catalysts generally exhibit higher carbazole HDN
conversions than the sulfided Ni-Mo/Al2O3 catalyst over the temperature range 548-673 K, but
the difference is greatest at high conversions as noted above. While not a focus of this study, a
16
Page 17 of 38

possible explanation of the lower HDN conversion for the sulfided Ni-Mo/Al2O3 catalyst at high
temperatures relative to the metal phosphide catalysts is a greater sensitivity to poisoning by
NH3. The catalyst activities are best compared at lower conversions and this is done using the
data at 573 K for the carbazole-only feed and at 598 K for the mixed carbazole / benzothiophene
feed. The catalyst activities are compared at these temperatures on the basis of mass of catalyst
loaded into the reactor (0.25 g), as well as by using turnover frequencies (TOFs) as listed in
Table 5. For the TOF calculations, chemisorption capacities (273 K) were used for the metal
phosphide catalysts and O2 chemisorption capacity (196 K) was used for the sulfided Ni-
Mo/Al2O3 catalyst. On a mass basis, the metal phosphide catalysts have higher carbazole HDN
activities than the sulfided Ni-Mo/Al2O3 catalyst (except for the Fe0.03Ni1.97P/SiO2 catalyst with
the mixed feed), but the sulfide catalyst has a higher TOF at the comparison temperatures. These
results point to the inherent limitation of Mo sulfide-based catalysts of having relatively small
numbers of active sites, with the low site densities traced to the anisotropic structure of MoS2
slabs that expose active sites only edge planes. The nickel phosphide based catalysts, on the
other hand, have substantially higher active site densities as measured by their chemisorption
capacities, which translate into higher carbazole HDN activities despite having lower TOFs.
The HDN activity differences among the three metal phosphide catalysts were generally
small, with the Ni2P/SiO2 and Co0.1Ni1.9P/SiO2 catalysts somewhat more active than the
Fe0.03Ni1.97P/SiO2 catalyst. In earlier work, the bimetallic phosphide compositions were
observed to have higher thiophene and dibenzothiophene HDS activities than Ni2P/SiO2, which
was hypothesized to be due to P-enrichment of the Ni-rich bimetallic phosphide surfaces and
transfer of electron density from the Co or Fe to Ni in the phosphides [7,13,24]. As indicated by
the surface compositions in Table 2, the Co0.1Ni1.9P/SiO2 and Fe0.03Ni1.97P/SiO2 catalysts are
enriched in P at their surfaces (P/M = 1.2-1.5) relative to the Ni2P/SiO2 catalyst (P/Ni = 0.68),
17
Page 18 of 38

but no significant differences in catalytic properties are observed. Other than the small HDN
activity differences, the Co0.1Ni1.9P/SiO2, Fe0.03Ni1.97P/SiO2 and Ni2P/SiO2 catalysts show
important similarities – excellent resistance to S incorporation for catalytic measurements using
the carbazole / benzothiophene feed and no loss of phase purity as indicated by XRD.
The addition of benzothiophene to the reactor feed inhibited carbazole HDN over the
metal phosphide and Ni-Mo/Al2O3 catalysts as indicated by the conversion data presented in
Figures 2 and 4. Similar inhibition of carbazole HDN by organosulfur compounds was reported
for sulfided Mo/Al2O3 and Ni-Mo/Al2O3 catalysts previously [28,29]. For sulfided Mo/Al2O3,
carbazole HDN activity decreased with increasing concentrations of S compounds in the feed;
the strongest inhibition was observed for dibenzothiophene and carbon disulfide with lesser
inhibition for ethanethiol, thiophene and dimethyl sulfide. 2-Methyl-2-propanethiol, when co-
fed at levels in the 500-3000 ppm range, reduced the carbazole conversion over a Ni-Mo/Al2O3
catalyst by approximately 20 mol% at 603 K, which is similar to the decrease observed in the
current study at 598 K (carbazole conversion: 73 mol% - carbazole-only feed, 58 mol% - mixed
feed). Nagai et al. concluded that the inhibition of HDN over Ni-Mo/Al2O3 was due to the
organosulfur compounds competing with carbazole for hydrogenation sites at low temperatures
(< 593 K) and hydrogenated carbazoles competing with the sulfur compounds (including H2S)
for C-N hydrogenolysis sites at higher temperatures (≥ 593 K) [29]. For the Ni2P/SiO2,
Co0.1Ni1.9P/SiO2 and Fe0.03Ni1.97P/SiO2 catalysts, the decreases in carbazole conversion at 598 K
associated with co-feeding benzothiophene were larger (e.g. Ni2P/SiO2, carbazole conversion: 89
mol% - carbazole-only feed, 62 mol% - mixed feed) than observed for the Ni-Mo/Al2O3
catalyst, indicating a greater sensitivity to the organosulfur compound for the metal phosphide
catalysts at this temperature. Only the Fe0.03Ni1.97P/SiO2 catalyst exhibited significant HDN
18
Page 19 of 38

inhibition by benzothiophene at 623 K, at which temperature over 95 mol% of the
benzothiophene was converted to hydrocarbon products.
Based on studies of carbazole HDN over Ni-Mo/Al2O3 catalysts [3,18,29-31] as well as
the present work, the reaction network in Scheme 1 is proposed. For the metal phosphide
catalysts (and Ni-Mo/Al2O3), insignificant amounts of biphenyl were detected in the reactor
effluent indicating that direct denitrogenation (DDN) is an inconsequential contributor to
carbazole HDN over the catalysts in this study. On the other hand, hydrogenated carbazoles
(4H-CARB, 6H-CARB, 12H-CARB) were observed, particularly at low reaction temperatures,
and these are intermediates in the hydrogenation (HYD) pathway for carbazole HDN. Four
hydrocarbon products of the HYD pathway were detected – CHB, BCH, HCH and MCPCH –
but with BCH and MCPCH composing over 70 mol% of the products. HCH and MCPCH are
isomerization products formed via ring opening and ring contraction, respectively, of one of the
six-member rings of BCH. A similar product selectivity was observed for unsupported W2C and
Pt-promoted W2C for carbazole HDN using a mixed carbazole / 4,6-DMDBT feed [32].
Reactions of BCH to give HCH and MCPCH in carbazole HDN over the tungsten carbide
catalysts were proposed to involve both metal and Brönsted acid sites [32]. The existence of
metal and Brönsted acid sites on the surface of Ni2P and CoxNi2P catalysts was invoked in a
study of 4,6-DMDBT HDS [33], for which increased activity via the direct desulfurization
pathway was associated with isomerization of the organosulfur compound (methyl migration) to
give methyl-substituted DBTs that more readily undergo C-S bond cleavage reactions. The high
HDS activity via the DDS pathway was most pronounced for unsupported metal phosphides (e.g.
Co0.08Ni2P) having P-enriched surfaces; the higher surface P content resulted in a higher
concentration of Brönsted acid sites at the catalyst surface in the form of incompletely reduced P
species (HxPO4x-3) [33]. The Ni2P/SiO2, Co0.1Ni1.9P/SiO2 and Fe0.03Ni1.97P/SiO2 catalysts in the
19
Page 20 of 38

current study exhibit varying levels of surface P enrichment and the high MCPCH product
selectivities are attributed to the presence of both metal and acid sites at the surfaces of these
catalysts. Interestingly, the high selectivities for isomerization products (HCH and MCPCH) are
more pronounced for the mixed feed than for the carbazole-only feed, suggesting surface
modification due to inclusion of benzothiophene in the feed (e.g. formation of a phosphosulfide
surface layer [6,7]) that favors further reaction of BCH on the catalyst surface.
As noted in the results section, the trends in product selectivity are similar in most details
for the metal phosphide and Ni-Mo catalysts. The most striking differences are the persistence
of hydrogenated carbazoles (~6 mol%) in the products at higher temperatures (623-673 K) over
the sulfided Ni-Mo/Al2O3 catalyst compared to the metal phosphide catalysts (< 3 mol%) and the
substantially higher CHB selectivities of the metal phosphide catalysts. At 623-673 K, the CHB
selectivity over the sulfided Ni-Mo/Al2O3 catalyst was only 5-13 mol% while over the metal
phosphide catalysts it was 10-32 mol%, with the CHB selectivities increasing with temperature
in each case. Taken together, these selectivity results suggest that C-N hydrogenolysis of 6H-
CARB to give CHB is a more active reaction channel for the metal phosphide catalysts than for
sulfided Ni-Mo/Al2O3, resulting in higher CHB selectivities and higher HDN conversions. The
higher CHB selectivity of the metal phosphide catalysts relative to the Ni-Mo/Al2O3 catalyst also
results in lower hydrogen consumption, an important consideration for catalyst selection for
hydrotreating applications.
5. CONCLUSIONS
The carbazole HDN properties of Ni2P/SiO2, Co0.1Ni1.9P/SiO2 and Fe0.03Ni1.97P/SiO2
catalysts were investigated for carbazole-only and mixed carbazole – benzothiophene feeds.
The metal phosphide catalysts exhibited high carbazole HDN activities and out-performed a
commercial sulfided Ni-Mo/Al2O3 catalyst on a mass of catalyst basis, with the largest
20
Page 21 of 38

differences observed at temperatures at which the HDN conversions were high. Co-feeding of
benzothiophene inhibited carbazole HDN over the metal phosphides, but the Ni2P/SiO2 and Ni-
rich bimetallic phosphide catalysts maintained higher activities than the sulfided Ni-Mo/Al2O3
catalyst. The product selectivities of the metal phosphide and sulfide catalysts were similar,
except that the sulfided Ni-Mo/Al2O3 catalyst yielded more hydrogenated carbazoles while the
metal phosphide catalysts produced more of the partially hydrogenated hydrocarbon
cyclohexylbenzene. The metal phosphide catalysts favored the isomerization products HCH and
MCPCH, formed by ring opening and ring contraction of one of the six-member rings of BCH,
respectively, that indicate the presence of metal and Brönsted acid sites on the catalyst surfaces.
The Ni2P/SiO2 and bimetallic phosphide catalysts exhibited excellent stabilities in the HDN
conditions employed (carbazole-only and mixed feeds), as indicated by phase purity, average
crystallite size and resistance to sulfur incorporation of the supported metal phosphide particles.
6. ACKNOWLEDGEMENTS
This research was supported by a grant from Shell Global Solutions (U.S.) Inc. The
authors acknowledge Dr. Jerimiah M. Purcell of Shell Oil Company for carrying out the GC-MS
analyses on samples of the reactor effluent.
21
Page 22 of 38

7. REFERENCES
1.
H. Topsøe, B. Clausen, F.E. Massoth, Hydrotreating Catalysis, in: J.R. Anderson, M. Boudart (Eds.)
Catalysis: Science and Technology, Springer-Verlag, Berlin, 1996, pp. 1-310.
2.
3.
4.
R. Prins, Hydrotreating, in: G. Ertl (Ed.) Handbook of Heterogeneous Catalysis, 2008, pp. 2695-2718.
E. Furimsky, F.E. Massoth, Hydrodenitrogenation of Petroleum, Catal. Rev. Sci. Eng., 47 (2005) 297-489.
T.C. Ho, L. Qiao, Competitive adsorption of nitrogen species in HDS: Kinetic characterization of
hydrogenation and hydrogenolysis sites, J. Catal., 269 (2010) 291-301.
5.
P. Wiwel, B. Hinnemann, A. Hidalgo-Vivas, P. Zeuthen, B.O. Petersen, J.Ø. Duus, Characterization and
Identification of the most Refractory Nitrogen Compounds in Hydroprocessed Vacuum Gas Oil, Ind. Eng.
Chem. Res., 49 (2010) 3184-3193.
6.
S.T. Oyama, T. Gott, H. Zhao, Y.-K. Lee, Transition Metal Phosphide Hydroprocessing Catalysts, Catal.
Today, 143 (2009) 94-107.
7.
R. Prins, M.E. Bussell, Metal Phosphides: Preparation, Characterization and Catalytic Reactivity, Catal.
Lett., 142 (2012) 1413-1436.
8.
X. Wang, P. Clark, S.T. Oyama, Synthesis, Characterization and Hydrotreating Activity of Several Iron
Group Transition Metal Phosphides, J. Catal., 208 (2002) 321-331.
9.
S.T. Oyama, Novel Catalysts for Advanced Hydroprocessing: Transition Metal Phosphides, J. Catal., 216
(2003) 343-352.
10.
11.
T. Korányi, Phosphorus Promotion of Ni (Co)-Containing Mo-free Catalysts in Thiophene
Hydrodesulfurization, Appl. Catal. A, 239 (2003) 253-267.
F. Sun, W. Wu, Z. Wu, J. Guo, Z. Wei, Y. Yang, Z. Jiang, F. Tian, C. Li, Dibenzothiophene
Hydrodesulfurization Activity and Surface Sites of Silica-Supported MoP, Ni2P and Ni-Mo-P Catalysts, J.
Catal., 228 (2004) 298-310.
12.
13.
14.
15.
S.J. Sawhill, K.A. Layman, D.R. Van Wyk, M.H. Engelhard, C. Wang, M.E. Bussell, Thiophene
Hydrodesulfurization over Nickel Phosphide Catalysts: Effect of the Precursor Composition and Support,
J. Catal., 231 (2005) 300-313.
A.F. Gaudette, A.W. Burns, J.R. Hayes, M.C. Smith, T. Seda, M.E. Bussell, Mössbauer Spectroscopy
Investigation and Hydrodesulfurization Properties of Iron-Nickel Phosphide Catalysts, J. Catal., 272 (2010)
18-27.
W. Kanda, I. Siu, J. Adjaye, A.E. Nelson, M. Gray, Inhibition and Deactivation of
Hydrodenitrogenation(HDN) Catalysts by Narrow-Boiling Fractions of Athabasca Coker Gas Oil, Energy
Fuels, 18 (2004) 539-546.
W.R.A.M. Robinson, J.N.M. van Gestel, T.I. Korányi, S. Eijsbouts, A.M. van der Kraan, J.A.R. van Veen,
V.H.J. de Beer, Phosphorous Promotion of Ni(Co)-Containing Mo-Free Catalysts in Quinoline
Hydrodenitrogenation, J. Catal., 161 (1996) 539-550.
16.
17.
V. Zuzaniuk, C. Stinner, R. Prins, T. Weber, Transition Metal Phosphides: Novel Hydrodenitrogenation
Catalysts, Stud. Surf. Sci. Catal., 143 (2002) 247-255.
V. Zuzaniuk, R. Prins, Synthesis and Characterization of Silica-Supported Transition-Metal Phosphides as
HDN Catalysts, J. Catal., 219 (2003) 85-96.
22
Page 23 of 38

18.
19.
20.
I.I. Abu, K.J. Smith, HDN and HDS of model compounds and light gas oil derived from Athabasca
bitumen using supported metal phosphide catalysts, Appl. Catal. A, 328 (2007) 58-67.
I.I. Abu, K.J. Smith, Hydrodenitrogenation of Carbazole over a Series of Bulk NixMoP Catalysts, Catal.
Today, 125 (2007) 248-255.
S.J. Sawhill, D.C. Phillips, M.E. Bussell, Thiophene Hydrodesulfurization over Supported Nickel
Phosphide Catalysts, J. Catal., 215 (2003) 208-219.
21.
22.
JCPDS Powder Diffraction File, International Centre for Diffraction Data, Swarthmore, PA, 2000.
J.B. Goodenough, Interpretation of Structure and Magnetism in Transition-Metal Pnictides M2X and (M1-
xM'x)2X, J. Solid State Chem., 7 (1973) 428-447.
23.
24.
25.
26.
R. Fruchart, A. Roger, J.P. Senateur, Crystallographic and Magnetic Properties of Solid Solutions of the
Phosphides M2P, M = Cr, Mn, Fe, Co and Ni, J. Appl. Phys., 40 (1969) 1250-1257.
A.W. Burns, A.F. Gaudette, M.E. Bussell, Hydrodesulfurization Properties of Cobalt-Nickel Phosphide
Catalysts: Ni-rich Materials are Highly Active, J. Catal., 260 (2008) 262-269.
S. Kasztelan, T. des Courières, M. Breysse, Hydrodenitrogenation of petroleum distillates : industrial
aspects, Catal. Today, 10 (1991) 433-445.
R. Iwamoto, J. Grimblot, Influence of Phosphorus on the Properties of Alumina-Based Hydrotreating
Catalysts, Adv. Catal., 44 (1999) 417-503.
27.
28.
R. Prins, Catalytic Hydrodenitrogenation, Adv. Catal., 46 (2001) 399-464.
M. Nagai, T. Masunaga, N. Hana-oka, Hydrodenitrogenation of Carbazole on a Mo/Al2O3 Catalyst. Effcts
of Sulfiding and Sulfur Compounds, Energy Fuels, 2 (1988) 645-651.
29.
30.
M. Nagai, Inhibiting Effect of Sulfur and Oxygen Containing Compounds on Carbazole
Hydrodenitrogenation on a NiMo/Al2O3 Catalyst, Sekiyu Gakkaishi, 36 (1993) 502-506.
C. Moreau, C. Aubert, R. Durand, N. Zmimita, P. Geneste, Structure-Activity Relationships in
Hydroprocessing of Aromatic And Heteroaromatic Model Compounds over Sulphided NiO-MoO3/Al2O3
and NiO-WO3/Al2O3 Catalysts: Chemical Evidence for the Existence of Two Types of Catalytic Sites,
Catal. Today, 4 (1988) 117-131.
31.
32.
F. Sánchez-Minero, J. Ramirez, Kinetic Study of the Reaction System 4,6-DMDBT + Naphthalene +
Carbazole, Using NiMo/Al2O3-SiO2(x) Catalysts, Ind. Eng. Chem. Res., 50 (2011) 2671-2677.
M. Lewandowski, P. Da Costa, D. Benichou, C. Sayag, Catalytic Performance of Platinum Doped
Tungsten Carbide in Simultaneous Hydrodenitrogenation and Hydrodesulphurization, Appl. Catal. B., 93
(2010) 241-249.
33.
I.I. Abu, K.J. Smith, The Effect of Cobalt Addition to Bulk MoP and Ni2P Catalysts for the
Hydrodesulfurization of 4,6-dimethyldibenzothiophene, J. Catal., 241 (2006) 356-366.
23
Page 24 of 38

TABLES
Table 1: Catalyst Characterization Data for as-prepared Catalysts
Precursor Average
Chemisorption
Capacity
Nominal
XPS Surface
Composition
BET Surface
Area(m2/g)
P/Metal Crystallite
Composition
CO
O2
Ratio
Size(nm)
(μmol/g)
(μmol/g)
Ni2P/SiO2
0.9
1.0
0.9
----
12
15
Ni1.46P1.00
Fe≤0.05Ni0.80P1.00
Co≤0.05Ni0.66P1.00
----
132
124
122
110
113
103
107
----
271
258
248
65
Fe0.03Ni1.97P/SiO2
Co0.1Ni1.9P/SiO2
Sulf. Ni-Mo/Al2O3
13
----
Table 2: Hydrotreating Activity and Characterization Data for Tested Catalysts –
Carbazole-only Feed
Post-HDN
Carbon
Content
(wt% C)
Sulfur
Content
(wt% S)
Average
Crystallite
Size(nm)
Carbazole HDN
Conversion (%) at 623 K
Catalyst
Ni2P/SiO2
12
15
1.06
1.10
1.39
1.67
----
----
96
95
98
75
Fe0.03Ni1.97P/SiO2
Co0.1Ni1.9P/SiO2
Sulf. Ni-Mo/Al2O3
13
----
----
7.10
24
Page 25 of 38

Table 3: Hydrotreating Activity and Characterization Data for Tested Catalysts – Mixed
Carbazole / Benzothiophene Feed
Post-HDN
Average
Post-HDN Catalyst
C and S Contents
Conversion (%) at 623 K
Catalyst
Crystallite
Size (nm)
Carbazole
HDN
Benzothiophene
wt% C
wt% S
HDS
Ni2P/SiO2
15
16
0.67
1.27
0.88
1.41
0.31
0.16
0.35
8.56
96
80
94
77
98
Fe0.03Ni1.97P/SiO2
Co0.1Ni1.9P/SiO2
Sulf. Ni-Mo/Al2O3
96
14
98
----
95
Table 4: Carbazole HDN Product Selectivity at 623 K
Catalyst
Feed
Sulf. Ni-Mo/Al2O3
Ni2P/SiO2
Fe0.03Ni1.97P/SiO2
Co0.1Ni1.9P/SiO2
CARB
CARB
CARB
CARB
CARB
CARB
CARB
CARB
+ BT
+ BT
+ BT
+ BT
0.1
12.9
33.4
48.3
5.2
0.2
5.0
0.2
6.0
0.2
0.1
13.2
33.8
47.3
5.6
0.2
14.9
22.4
52.7
9.8
0.1
8.9
0.2
BP
11.4
26.4
52.5
9.5
11.8
22.7
56.9
8.4
CHB
38.3
47.8
8.7
29.1
53.7
11.0
40.3
46.0
4.7
BCH
MCPCH
HCH
Table 5: Carbazole HDN Activities and Turnover Frequencies
Carbazole-only Feed (573 K)
Carbazole / Benzothiophene Feed (598 K)
Catalyst
HDN Activity
(nmol/g⋅s)
HDN Activity
(nmol/g⋅s)
HDN TOF
(s-1) x 104
HDN TOF
(s-1) x 104
Ni2P/SiO2
15.9
13.2
14.9
10.8
1.41
1.28
1.40
1.67
17.4
13.3
18.4
16.3
1.54
1.29
1.72
2.50
Fe0.03Ni1.97P/SiO2
Co0.1Ni1.9P/SiO2
Sulf. Ni-Mo/Al2O3
25
Page 26 of 38

SCHEME CAPTION
Scheme 1: Carbazole HDN reaction network (adapted from [3,19]).
FIGURE CAPTIONS
Figure 1: XRD patterns for as-prepared Ni2P/SiO2, Fe0.03Ni1.97P/SiO2 and Co0.1Ni1.9P/SiO2
catalysts.
Figure 2: Carbazole HDN conversion vs. temperature for Ni2P/SiO2, Fe0.03Ni1.97P/SiO2,
Co0.1Ni1.9P/SiO2 and sulfided Ni-Mo/Al2O3 catalysts (carbazole only feed).
Figure 3: XRD patterns for Ni2P/SiO2 catalysts (a) in as-prepared form, (b) after carbazole
HDN testing, and (c) after carbazole HDN testing with benzothiophene co-feed.
Figure 4: Carbazole HDN conversion vs. temperature for Ni2P/SiO2, Fe0.03Ni1.97P/SiO2,
Co0.1Ni1.9P/SiO2 and sulfided Ni-Mo/Al2O3 catalysts (mixed carbazole /
benzothiophene feed).
Figure 5: Benzothiophene HDS conversion vs. temperature for Ni2P/SiO2, Fe0.03Ni1.97P/SiO2,
Co0.1Ni1.9P/SiO2 and sulfided Ni-Mo/Al2O3 catalysts (mixed carbazole /
benzothiophene feed).
Figure 6: Carbazole HDN product selectivity vs. temperature for a sulfided Ni-Mo/Al2O3
catalyst (mixed carbazole / benzothiophene feed).
Figure 7: Carbazole HDN product selectivity vs. temperature for a Ni2P/SiO2 catalyst (mixed
carbazole / benzothiophene feed).
Figure 8: Carbazole HDN product selectivity vs. temperature for a Fe0.03Ni1.97P/SiO2 catalyst
(mixed carbazole / benzothiophene feed).
Figure 9: Carbazole HDN product selectivity vs. temperature for a Co0.1Ni1.9P/SiO2 catalyst
(mixed carbazole / benzothiophene feed).
26
Page 27 of 38

*Graphical Abstract (for review)
hydrocarbons
Ni2P
SiO2
Page 28 of 38

Scheme 1
Page 29 of 38