Journal of Catalysis 383 (2020) 331–342

Contents lists available at ScienceDirect

Journal of Catalysis
journal homepage: www.elsevier.com/locate/jcat

Desulfurization of 2-phenylcyclohexanethiol over transition-metal

phosphides
Xuerong Zhou a, Xiang Li b,⇑, Roel Prins c, Anjie Wang a,d, Lin Wang a, Shengnan Liu a, Qiang Sheng b
a
State Key Laboratory of Fine Chemicals, School of Chemical Engineering, Dalian University of Technology, Dalian 116024, People’s Republic of China
b
Tianjin Key Laboratory of Brine Chemical Engineering and Resource Eco-utilization, College of Chemical Engineering and Materials Science, Tianjin University of Science
and Technology, Tianjin 300457, People’s Republic of China
c
Institute of Chemical and Bioengineering, ETH Zurich, 8093 Zurich, Switzerland
d
Penn State and Dalian University of Technology Joint Center for Energy Research (JCER), Dalian University of Technology, Dalian 116024, People’s Republic of China

a r t i c l e i n f o a b s t r a c t

Article history: The desulfurization of 2-phenylcyclohexanethiol (2-PCHT) was studied over Ni2P, MoP, and WP under
Received 27 October 2019 4.0 MPa H2 or Ar at 240 °C. The hydrodesulfurization of 2-PCHT proceeded through three parallel path-
Revised 19 January 2020 ways: b-elimination, hydrogenolysis, and dehydrogenation. Under Ar, the parallel pathways were b-
Accepted 20 January 2020
elimination, hydrogenolysis or homolytic C–S bond cleavage, and dehydrogenation. MoP and WP were
more active than Ni2P. b-Elimination dominated the hydrodesulfurization of 2-PCHT over Ni2P, while
hydrogenolysis was as fast as b-elimination over MoP and WP. Under Ar, b-elimination and dehydrogena-
Keywords:
tion pathways were about equal over Ni2P, whereas b-elimination was the major pathway over MoP and
2-Phenylcyclohexanethiol
Cycloalkyl C–S bond
WP. The inhibiting effects of piperidine depended on the reaction and catalyst. Phosphosulfide phases
Desulfurization were formed under both H2 and Ar, but the sulfidation behavior of Ni2P was different from that of
Ni2P MoP and WP. Ni2P was more difficult to sulfide than MoP and WP under H2.
MoP Ó 2020 Elsevier Inc. All rights reserved.
WP
Transition-metal phosphosulfide
Piperidine

1. Introduction bond. Hydrogenolysis is catalyzed by metal atoms [9,10], whereas

both basic sites (to abstract the b-hydrogen atom) and acid or
Hydrodesulfurization (HDS) of petroleum fractions is one of the vacancy sites (to bind the sulfur of the reactant) are involved in
major catalytic operations in the petroleum industry [1]. Driven by the b-elimination mechanism [11]. In addition, hydrogen is a reac-
the growing demand for ultraclean fuels and the increasing use of tant in hydrogenolysis, but is not required for b-elimination.
low-quality crude oil feed stocks, deep HDS has received great DBT and DBTs undergo HDS by two parallel pathways: direct
attention. To achieve deep HDS, the most refractory sulfur- desulfurization (DDS) and hydrogenation (HYD) [12,13]. DDS leads
containing compounds, dibenzothiophene (DBT) and its alkylated to the formation of biphenyls, while HYD yields not only the final
derivatives (DBTs), must be removed from transportation fuels desulfurized products, cyclohexylbenzenes and bicyclohexyls, but
[2–4]. Hence, an in-depth understanding of the HDS mechanisms also the tetrahydro, hexahydro, and dodecahydro hydrogenated
of these molecules is prerequisite for the development of efficient sulfur-containing intermediates (Scheme S1 in the Supplementary
HDS catalysts. The two most commonly employed mechanisms for Material). There are three types of C–S bonds (the aryl C–S bond,
the cleavage of C–S bonds are hydrogenolysis and b-elimination. In the vinyl C–S bond, and the cycloalkyl C–S bond) in the HDS net-
hydrogenolysis, a C–S bond is broken by reaction with H atoms, works of DBT and DBTs. We demonstrated that over transition-
while C–H and S–H bonds form simultaneously [5–7]. For instance, metal phosphides [14–16], the cleavage of the aryl and vinyl C–S
hydrogenolysis of the C–S bond of thiophene yields 1,3-butadiene bonds in DBT and 1,2,3,4-tetrahydro-dibenzothiophene (TH-DBT)
[6], while biphenyl (BP) is the product of the hydrogenolysis of DBT occurs mainly through hydrogenolysis, while in the initial
[8]. b-Elimination involves the breaking of the a-C–S bond and a cycloalkyl C–S bond-breaking in 1,2,3,4,4a,9b-hexahydro-dibenzo
b-C–H bond, leading to the formation of H2S and a C=C double thiophene (HH-DBT), both b-elimination and hydrogenolysis can
be involved.
Transition-metal phosphides (e.g., Ni2P, MoP, WP, CoP, and FeP)
⇑ Corresponding author. are a novel class of HDS catalysts. They have high activity and
E-mail address: [email protected] (X. Li).

https://doi.org/10.1016/j.jcat.2020.01.023
0021-9517/Ó 2020 Elsevier Inc. All rights reserved.
332 X. Zhou et al. / Journal of Catalysis 383 (2020) 331–342

stability under hydrotreating conditions [17,18]. Among the phos- pelleted, crushed, and sieved to 20–40 mesh. A total of 0.10 g pre-
phides investigated, Ni2P, MoP, and WP are the three most active cursor was used for each run. Prior to the reaction, the precursor
catalysts for HDS and HDN of petroleum feedstocks [17,18]. Their was transformed into the phosphide phase by in situ H2
initial activity increases gradually with time on stream during temperature-programmed reduction. It was reduced in a 100 mL/
HDS reactions [19]. It is suggested that sulfur is incorporated into min H2 flow at 1.0 MPa by heating it quickly from room tempera-
the surface of these phosphide catalysts, leading to the formation ture to 120 °C and maintaining this temperature for 1 h, then
of a more active surface phosphosulfide phase [20,21]. The forma- heated to 400 °C at 5 °C/min, and finally heated at 1 °C/min to
tion, the preparation, and the nature and the catalytic perfor- 500 °C (for the preparation of Ni2P) or 650 °C (for the preparation
mances of these novel sulfur-containing phases still need to be of MoP and WP) and holding at that temperature for 2 h.
studied. Our previous results suggested that the formation of the Because the reaction of 2-PCHT is quite fast, it was studied at a
molybdenum phosphosulfide phase was due to the direct reaction temperature as low as 240 °C. After the precursor had been con-
of the phosphide catalyst with sulfur-containing molecules during verted into the phosphide phase, the reactor was cooled to the
HDS, rather than to the sulfidation of the MoP surface by H2S/H2 reaction temperature, and the gas-phase feed was introduced. It
mixtures with low H2S partial pressures (5 or 35 kPa H2S in consisted of 1.5 kPa 2-PCHT, 164 kPa decalin (as solvent), and
4.0 MPa H2) [22]. For an in-depth understanding of the cleavage about 3.8 MPa H2 or Ar both in the absence or presence of
of the cycloalkyl C–S bonds in the hydrogenated intermediates of 0.25 kPa piperidine. The space–time (or the contact time) calcu-
DBT and DBTs, the catalytic properties of transition-metal phos- lated based on the catalyst weight was defined as s = wcat/nfeed,
phides, and the formation of the phosphosulfide phase, we studied where wcat denotes the catalyst weight (g) and nfeed the total molar
the desulfurization of 2-phenylcyclohexanethiol (2-PCHT) over flow to the reactor (mol/min) [25]. The space–time was changed by
bulk Ni2P, MoP, and WP catalysts, under both a hydrogen atmo- varying the flow rates of the liquid (from 0.1 to 0.3 mL/min) and
sphere and an inert gas atmosphere (Ar). The molecular structure the gas, while keeping their ratio constant (750 v/v). The reaction
of 2-PCHT is similar to that of HH-DBT, but only possesses a products were analyzed off line by an Agilent-6890N gas chro-
cycloalkyl C–S bond. It is a chemically reasonable intermediate in matograph equipped with an HP-5 column and a flame ionization
the HDS of DBT, although it has never been detected so far. As pro- detector. Because 3-phenyl-1-cyclohexene (3-PCHE) and phenylcy-
posed by Wang and Prins [23], it can be formed by the hydrogenol- clohexane (or cyclohexylbenzene, CHB) cannot be separated by the
ysis of the aryl C–S bond in HH-DBT. nonpolar HP-5 column, an HP-INNOWax polyethylene glycol col-
umn was used to quantify the 1-PCHE, 3-PCHE, and 4-phenyl-1-
cyclohexene (4-PCHE) present in the desulfurization products
2. Experimental methods
[14]. The structures and the acronyms of the main compounds
involved in the present work are listed in Table 1.
2.1. The preparation of catalyst precursors
Table 1
The phosphate precursors of Ni2P, MoP, and WP were prepared Structures and acronyms of the main compounds involved in the present work.
by adding an aqueous solution of 3.0 g (NH4)2HPO4 in 10 mL deion-
Name Structure Acronym
ized water dropwise to a solution of 6.6 g Ni(NO3)2
6H2O, 4.0 g
(NH4)6Mo7O24
4H2O, or 5.6 g (NH4)6W12O40
xH2O in 15 mL deion- 2-Phenylcyclohexanethiol 2-PCHT
ized water while stirring. The resulting mixtures were stirred while
the water was evaporated to obtain the solid products, which were
dried at 120 °C for 12 h and calcined at 500 °C for 3 h. The molar
compositions of the precursors were NiO∙0.5P2O5, MoO3∙0.5P2O5, 1,2,3,4-Tetrahydro-dibenzothiophene TH-DBT
and WO3∙0.5P2O5.

2.2. The synthesis of 2-PCHT

1,2,3,4,4a,9b-Hexahydro- HH-DBT
dibenzothiophene
2-PCHT was synthesized by free radical addition of 1-phenyl-1-
cyclohexene (1-PCHE) and thioacetic acid, followed by hydrolysis
with 10% potassium hydroxide in an alcohol–water mixture (1:1,
v/v) and extraction with ethyl acetate [24]. The extracted oily layer 1-Phenyl-1-cyclohexene 1-PCHE
was washed successively with saturated NaHCO3 and NaCl solu-
tions and then dried by anhydrous MgSO4. The obtained yellow 3-Phenyl-1-cyclohexene 3-PCHE
oil was further purified by column chromatography on silica gel
with a mixture of petroleum ether (b.p. 60–90 °C) and ethyl acetate 4-Phenyl-1-cyclohexene 4-PCHE
(20:1, v/v). The synthesized 2-PCHT possessed the cis configura-
tion. 1H NMR (400 MHz, CDCl3): 7.28 (t, J = 7.5 Hz, 2H), 7.18 (t, Cyclohexylbenzene CHB
J = 8.4 Hz, 3H), 3.57 (s, 1H), 2.95 (d, J = 12.4 Hz, 1H), 2.13 (m,
6.4 Hz, 1H), 2.00–1.86 (m, 3H), 1.74–1.62 (m, 2H), 1.52 (d, Biphenyl BP
J = 13.6 Hz, 1H), 1.41–1.30 (m, 1H), 0.94 (d, J = 5.3 Hz, 1H). 13C
NMR (100 MHz, CDCl3): 144.36, 128.15, 127.69, 126.36, 47.21,
(Cyclopentenylmethyl)benzene CPEMB
44.17, 34.28, 26.50, 24.07, 19.99. EI-MS m/z (%): 192 (90, M + ),
158 (18), 143 (13), 129 (26), 117 (48), 101 (99), 91 (1 0 0), 77
(17), 65 (10), 51 (9), 39 (9). Cyclopentylmethylbenzene CPMB

2.3. The desulfurization reactions 2-Pheny-l-cyclohexenethiol 2-PCHET

The desulfurization reactions were carried out in a stainless

steel tubular reactor (8.0 mm i.d.). The catalyst precursors were
 X. Zhou et al. / Journal of Catalysis 383 (2020) 331–342 333

We found that independent of the catalyst (Ni2P, MoP, or WP) 3. Results

and reaction conditions (under H2 or Ar, in the presence or absence
of piperidine), all reactions showed substantial deactivation (25– 3.1. Characterization
30%) at the beginning of the reaction. For instance, the conversion
of 2-PCHT over MoP under 4.0 MPa Ar at 240 °C, 0 kPa piperidine, Since the structures of Ni2P, MoP, and WP are stable under HDS
and s = 0.74 g min/mol decreased from ca. 75% to 50% in the first conditions (in other words, they will not transfer to other phases),
6 h and then became constant (Fig. S1 in the Supplementary Mate- we only measured the XRD patterns of the spent catalysts after the
rial). To stabilize the catalysts, each reaction was started at high desulfurization of 2-PCHT under Ar, which were denoted as Ni2P-
space–time (s = 0.74 g min/mol) and then the space–time was Ar, MoP-Ar, and WP-Ar, respectively. Only diffraction peaks due
decreased in three steps to s = 0.25 g min/mol, after which it to the phosphide phases were observed (Fig. S2 in the Supplemen-
was increased again to s = 0.74 g min/mol. After this procedure, tary Material), indicating that the structures of these phosphide
the performance of all phosphide catalysts became stable under catalysts were well preserved. This is confirmed by the XPS results
all conditions. The product distribution over Ni2P hardly changed (Fig. S3 in the Supplementary Material). The major peaks or dou-
during this run-in period, while over MoP and WP the selectivity blets (due to spin–orbit interaction) observed in the Ni2p, Mo3d,
for CHB decreased, whereas the selectivities for phenylcyclohexe- and W4f spectra of the spent catalysts after the reactions under
nes (PCHEs) increased. In the following, we will show only the H2 (denoted as Ni2P-H, MoP-H, and WP-H) and Ar can be assigned
results obtained with run-in catalysts. to the reduced metal atoms in the phosphide phases. These peaks
are at 853.0 eV for Nid+ (0 < d < 2) in Ni2P [31], 231.6 and 228.4 eV
2.4. Characterization for Mod+ (0 < d < 4) in MoP [32], and 33.6 and 31.4 eV for Wd+ (0 -
< d < 4) in WP [33]. In the P2p region, the predominant doublets at
Before characterization, the fresh catalyst samples were pre- ca. 129.5 and 130.3 eV (Fig. S3 in the Supplementary Material) are
pared under the same reduction conditions as used in the in situ related to the P atoms bonded to Ni, Mo, and W phosphides [31–
reduction, followed by passivation with 0.5% (vol.) O2 in Ar. The 33]. The minor peaks or doublets at higher binding energies are
XRD patterns of the catalysts were measured on a Rigaku D/Max due to the oxidized metal atoms and PO3 4 species generated by
2400 diffractometer. Nitrogen physisorption was performed on a air exposure of the phosphide catalysts before XPS measurement
Quantachrome Autosorb-1-MP analyzer. The specific surface areas [32]. Considering the measurement errors, the binding energies
of the bulk Ni2P (4.3 m2/g), MoP (3.8 m2/g), and WP (4.1 m2/g) cat- (and, thus, the electronic states of these elements) in the spent cat-
alysts were about equally low. The compositions of the three fresh alysts under H2 and Ar were equal. The surface compositions of the
catalysts were determined by inductively coupled plasma atomic spent catalysts determined by XPS are given in Table S1 in the Sup-
emission spectrometry (ICP-AES, Optima 2000DV). The P/metal plementary Material, and the surface elemental ratios of phospho-
ratios of Ni2P, MoP, and WP were found to be 0.84, 0.97, and rus/metal (P/M) and sulfur/metal (S/M) are summarized in Table 2.
0.91, respectively, close to those of the corresponding precursors. The XPS P/Ni ratios of the spent Ni2P catalysts (ca. 2.2) were higher
CO adsorption was carried out using pulsed chemisorption as than the P/Ni ratios of the nickel phosphate precursor (1.0) and the
described elsewhere [26]. About 0.2 g of passivated, air-exposed fresh catalyst (0.84), whereas the XPS P/M ratios of the spent MoP
catalyst sample was re-reduced in a H2 flow to remove the passiva- and WP catalysts were close to those of their precursors (1.0) and
tion layer (80 mL/min flow rate, heating from 25 to 500 °C at a rate the fresh catalysts (0.97 for MoP and 0.91 for WP). It seems that
of 10 °C/min, and then holding at the final temperature for 1 h). phosphorus is enriched in the surface of Ni2P.
The reactor was then cooled to 30 °C in a flow of H2. An Ar flow Sulfur was detected in the surface of the spent catalysts. The
at 80 mL/min was used to flush the catalyst for 30 min to achieve S2p spectra of all the catalysts were similar (Fig. 1). They were
an adsorbate-free reduced catalyst surface. After pretreatment, composed of a major peak at ca. 161.3 eV and a shoulder at higher
1.25 mL pulses of 1% CO in Ar were injected into a flow of Ar binding energies (ca. 162.5 eV), which are in agreement with those
(80 mL/min), and the CO uptake was measured using a TCD detec- of the sulfur species in the transition-metal phosphosulfide phases
tor. CO pulses were repeatedly injected until the response from the [31–33]. The S2p binding energies of the spent Ni2P catalysts were
detector showed no further CO uptake. The CO uptakes of the Ni2P, about 0.5 eV lower than those of the spent MoP and WP catalysts.
MoP, and WP catalysts were measured to be 4.7, 2.2, and 2.4 lmol/ The S/M ratios of WP-Ar and MoP-Ar were lower than those of WP-
g, respectively. H and MoP-H, but the S/Ni ratio of Ni2P after the reaction in Ar was
X-ray photoelectron spectroscopy (XPS) spectra of the spent more than five times as high as that after the reaction in H2. The S/
catalysts were obtained with a Thermo ESCALAB 250 X-ray photo- M ratio of Ni2P-Ar was the highest among the catalysts studied,
electron spectrometer equipped with a high-resolution double- whereas that of Ni2P-H was the second lowest. WP-Ar possessed
focusing hemispherical analyzer, using an AlKa (1486.6 eV) source. the lowest S/M ratio, only half of the value of WP-H. The situation
The base pressure before the analysis was better than 5  106 Pa. is different for MoP. The S/M ratios of Mo-H and MoP-Ar were close
All the spectra were obtained in a constant analyzer energy mode. to each other.
A constant pass energy of 20 eV was used for the narrowscan; the
widescan spectra were acquired with a pass energy of 100 eV. The
instrument was operated in Large Area XL mode and the aperture 3.2. Desulfurization of 2-PCHT over Ni2P
diameter was 0.5 mm. The accuracy of the binding energy values
was ±0.05 eV. All binding energies were referenced to the C1s peak Because the isomerization of cis-2-PCHT to trans-2-PCHT was
at 284.6 eV. The atomic concentrations of different chemical ele- fast under our conditions, trans-2-PCHT and cis-2-PCHT were con-
ments were extracted from XPS intensity ratios after proper nor- sidered a single compound. The conversion of 2-PCHT over Ni2P
malization using atomic sensitivity factors [27]. The spent reached 44% at s = 0.74 g min/mol under H2 (Fig. 2A). The main
catalysts were used directly for XPS measurements after being four products were the desulfurized products 1-PCHE, 3-PCHE,
flushed with n-heptane and dried in an Ar flow. The samples were CHB, and BP (Fig. 2A). Their yields at the highest space–time
transferred to the XPS spectrometer through air. (0.74 g min/mol) are summarized in Table 3. 1-PCHE and 3-
The other basic characterizations of the transition-metal phos- PCHE, the two isomers of PCHE, were the two most abundant prod-
phide catalysts used in the present work can be found in our pre- ucts (Fig. 2A and Table 3). While the selectivity of 1-PCHE
vious publications [15,16,28-30]. decreased and that of 3-PCHE increased with space–time
334 X. Zhou et al. / Journal of Catalysis 383 (2020) 331–342

Table 2
XPS phosphorus/metal (P/M) and sulfur/metal (S/M) ratios of the spent phosphide catalysts.

Catalyst Ni2P-H Ni2P-Ar MoP-H MoP-Ar WP-H WP-Ar

P/M 2.27 2.11 1.10 1.18 1.27 1.32
S/M 0.19 0.96 0.31 0.28 0.39 0.17

ing to 5.2 at s = 0.74 g min/mol), while under H2 it was almost

lower than one. The total selectivity of PCHEs was almost constant
Ni2P-H MoP-H WP-H (Fig. S4 in the Supplementary Material). TH-DBT was detected as
the fifth product (Fig. 3A and Table 3), but 4-PCHE and CPEMB
could hardly be detected. Piperidine showed a negative influence
Intensity (a.u.)

on the conversion of 2-PCHT over Ni2P under Ar, and only 1-

PCHE, 3-PCHE, CHB, and BP were observed (Fig. 3C). Their selectiv-
Ni2P-Ar MoP-Ar WP-Ar ities were similar to those in the absence of piperidine (cf. Fig. 3B
and D); only the total selectivity of PCHEs was a little bit higher
in the presence of piperidine (Fig. S4 in the Supplementary
Material).

166 164 162 160 166 164 162 160 166 164 162 160 3.3. Desulfurization of 2-PCHT over MoP
Binding energy (eV)
MoP was more active than Ni2P in the desulfurization of 2-PCHT
Fig. 1. XPS spectra of the spent Ni2P, MoP, and WP samples in the S2p region.
under 4.0 MPa H2 at 0 kPa piperidine (Fig. 4A). The conversion of 2-
PCHT was ca. 60% at s = 0.74 g min/mol (Fig. 4A). CHB was the most
abundant product, and its yield and selectivity increased with
(Fig. 2B), the total selectivity of the three PCHE isomers (1-PCHE, space-time (Fig. 4A and B). 1-PCHE and 3-PCHE were the second
3-PCHE, and 4-PCHE) was nearly constant (ca. 78%; Fig. S4 in the and third most abundant products, and they had similar yields
Supplementary Material). CHB and BP were the third and fourth (Fig. 4A). The yield of 3-PCHE increased with space–time, while
major products. Their BP/CHB ratio varied from 1.2 at low space– that of 1-PCHE passed through a maximum between 0.50 and
time (0.25 g min/mol) to 0.5 at high space–time (0.74 g min/mol). 0.74 g min/mol (Fig. 4A), and as a result the selectivity of 1-PCHE
BP behaved like a primary product, as its selectivity decreased with decreased with space–time, whereas the selectivity of 3-PCHE
space–time (Fig. 2B). Four other compounds were observed at low hardly changed (Fig. 4B). The yield of 4-PCHE was very low (less
concentrations: 4-PCHE > TH-DBT ~ cyclopentenylmethylbenzene than 3%). The total selectivity of PCHEs was much lower than that
(CPEMB) > cyclopentylmethylbenzene (CPMB). CPMB and CPEMB over Ni2P under identical conditions and decreased slightly with
are isomers of CHB and PCHEs, respectively. space–time (Fig. S4 in Supplementary Material). BP, CPMB, and
Piperidine strongly inhibited the HDS of 2-PCHT. The 2-PCHT CPEMB were detected in trace amounts and TH-DBT was not
conversion at s = 0.74 g min/mol decreased significantly from detected.
44% at 0 kPa piperidine (Fig. 2A) to 14% at 0.25 kPa piperidine The HDS of 2-PCHT over MoP was inhibited by piperidine. At
(Fig. 2C). 1-PCHE and BP were the two most abundant compounds s = 0.74 g min/mol and 0.25 kPa piperidine, the conversion of 2-
at s < 0.5 g min/mol, with about equal yields (Fig. 2C) and selectiv- PCHT was 37% (Fig. 4C). Only four products, CHB, 3-PCHE, 1-
ities (Fig. 2D). At s > 0.5 g min/mol, the BP yield leveled off and CHB PCHE, and a trace amount of BP, were detected, with CHB as the
became the second most abundant product, just behind 1-PCHE. most abundant product (Fig. 4C). The yield of 3-PCHE was higher
The yield of BP in the presence of piperidine was even higher than than that of 1-PCHE. The selectivity of CHB increased slightly from
in the absence of piperidine (Fig. 2A and C and Table 3). This led to 47% to 53% with space–time, whereas that of 1-PCHE decreased
a higher BP/CHB ratio (changing from 2.65 to 0.94 with increasing from 22% to 16% (Fig. 4D). The total selectivity of PCHEs was more
space–time) in the presence of piperidine. The selectivities of 1- than 46% (Fig. S4 in the Supplementary Material). Because of the
PCHE and BP decreased with space–time (Fig. 2D), suggesting that low yields of BP (Fig. 4A and C and Table 3), the BP/CHB ratio
1-PCHE and BP are primary products of the HDS of 2-PCHT. The was very low (less than 0.03) in the HDS of 2-PCHT over MoP in
total selectivity of PCHEs was ca. 45%, lower than in the absence the absence as well as in the presence of piperidine.
of piperidine (ca. 78%), and was nearly constant with space–time The reactivity of 2-PCHT over MoP under Ar in the absence of
(Fig. S4 in the Supplementary Material). Only trace amounts of piperidine (Fig. 5A) was lower than that under H2 (Fig. 4A), but
TH-DBT and 4-PCHE were detected, and the formation of CPMB higher than over Ni2P under Ar (Fig. 3A). 1-PCHE was the major
and CPEMB was completely inhibited by piperidine. product, with a selectivity around 43% (Fig. 5B). The yields of the
Desulfurization of 2-PCHT also occurred under Ar (Fig. 3A), but other products decreased in the order BP ~ 3-PCHE > TH-DBT ~
the reactivity of 2-PCHT was lower than under H2 (cf. Figs. 2A and CHB > 4-PCHE ~ CPEMB (Fig. 5A). The selectivities of all products
3A). 1-PCHE and BP were the two most abundant compounds, with were nearly constant (Fig. 5B).
about equal yields (Fig. 3A) and selectivities (Fig. 3B). The selectiv- The desulfurization of 2-PCHT under Ar over MoP was inhibited
ity of 1-PCHE was independent of space–time, whereas that of BP by piperidine as well, and the yields of the main products
decreased with space–time (Fig. 3B). The yields of 3-PCHE and decreased in the order 1-PCHE > BP > CHB > 3-PCHE (Fig. 5C). With
CHB were about equal (Fig. 3A). Whereas under H2 the main prod- increasing space–time, the selectivities of 1-PCHE and BP
ucts were 1-PCHE and 3-PCHE, under Ar their yields had decreased decreased, whereas that of 3-PCHE remained unchanged
considerably (Fig. 3A and Table 3). Under Ar the yield of CHB plus (Fig. 5D). The total selectivity of PCHEs was about 10% lower than
BP was about equal to that under H2, but the ratio BP/CHB was in the absence of piperidine (Fig. S4 in the Supplementary Mate-
greater than one (as high as 10 at s = 0.25 g min/mol and decreas- rial). CHB behaved like a final product, its yield and selectivity
 X. Zhou et al. / Journal of Catalysis 383 (2020) 331–342 335

100 50 increased with space-time (Fig. 5C and D). 4-PCHE and TH-DBT
A were detected in trace amounts.
2-PCHT concentration (%)

Product concentration (%)

80 40
3.4. Desulfurization of 2-PCHT over WP

60 1-PCHE CPEMB 30
3-PCHE CPMB WP and MoP behaved similarly in the desulfurization of 2-PCHT
4-PCHE TH-DBT (Figs. S5 and S6 in the Supplementary Material). Minor differences
40 CHB 2-PCHT 20 are as follows. WP was as active as MoP in the HDS of 2-PCHT, but
BP
was more active than MoP in the desulfurization of 2-PCHT under
Ar. Under H2 and 0 kPa piperidine, the selectivity of CHB over WP
20 10
was slightly lower than that over MoP, while the selectivities of 1-
PCHE and 3-PCHE were slightly higher (Fig. 4B and Fig. S5B in the
0 0 Supplementary Material). In the presence of piperidine, however,
0.0 0.2 0.4 0.6 0.8
Space-time (g·min/mol )
the product selectivities over the two catalysts were almost equal
under H2 (Fig. 4D and Fig. S5D in the Supplementary Material).
100
Under 4.0 MPa Ar and 0 kPa piperidine, WP showed a higher selec-
B 1-PCHE BP tivity to 1-PCHE but a lower selectivity to 3-PCHE than MoP
80 3-PCHE CPEMB (Fig. 5B and Fig. S6B in the Supplementary Material). The formation
4-PCHE CPMB
of BP under Ar was more inhibited by piperidine over WP than over
CHB TH-DBT
Selectivity (%)

MoP (Table 3).

60

4. Discussion
40
4.1. The cleavage of the cycloalkyl C–S bond in 2-PCHT
20
It is well known that thiols undergo hydrogenolysis to alkanes
and H2S elimination to alkenes [34–37]. Therefore, in our case,
0
0.2 0.3 0.4 0.5 0.6 0.7 0.8 1-PCHE and 3-PCHE are the products of the elimination of the SH
Space-time (g·min/mol) group from 2-PCHT, and the hydrogenolysis of 2-PCHT yields
CHB under hydrogen. CHB can also be formed by the hydrogena-
100 10
tion of PCHEs. Nevertheless, the variations of the selectivities for
CHB as well as those of the total selectivities for PCHEs with
Product concentration (%)
2-PCHT concentration (%)

80 8 space–time were minor. In particular, the selectivities of CHB, 1-

1-PCHE C
3-PCHE PCHE, and 3-PCHE were almost constant in the HDS of 2-PCHT over
4-PCHE
60 CHB 6 MoP and WP at 0.25 kPa piperidine (Fig. 4D and Fig. S5D in the
BP Supplementary Material). These observations suggest that the
TH-DBT hydrogenation of PCHEs to CHB over the phosphide catalysts is
40 2-PCHT 4 slow at 240 °C in the presence of sulfur-containing compounds.
In other words, the HDS of 2-PCHT to PCHEs and to CHB are two
20 2 parallel reactions. This is confirmed by the low reactivity of 1-
PCHE over Ni2P in the presence of 0.5 kPa benzothiophene at
240 °C, 4.0 MPa H2, and s = 0.74 g min/mol. The conversion of 1-
0 0
0.0 0.2 0.4 0.6 0.8 PCHE and the yield of CHB were less than 5% and 4.3%, respectively.
Space-time (g·min/mol ) Hence, hydrogenation reactions are negligible over the phosphide
catalysts at 240 °C.
100
According to our previous work [14], 4-PCHE is mainly formed
1-PCHE by the isomerization of 3-PCHE. This is probably because of the
D
80 3-PCHE conjugation between the double bond and the aromatic ring in
4-PCHE
1-PCHE, which makes it more stable than 3-PCHE. Over Ni2P under
CHB
Selectivity (%)

60 BP H2 (4.0 MPa), 1-PCHE did not undergo isomerization to 3-PCHE and

TH-DBT 4-PCHE at a higher temperature (340 °C) [14]. In the desulfuriza-
tion of 2-PCHT over the phosphide catalysts, the yields of 4-PCHE
40 were always very low. Additionally, the selectivities of 3-PCHE
were nearly constant over all three phosphide catalysts under Ar,
20 indicating that its further reaction is negligible. These results sug-
gest that the isomerization of 1-PCHE and 3-PCHE is slow under
the present conditions, which may be due to the low acidity of
0
0.2 0.3 0.4 0.5 0.6 0.7 0.8 these bulk phosphide catalysts [16,29,38]. The low acidity of the
Space-time (g·min/mol) phosphides is further proved by the very low yields of CPEMB
and CPMB, the skeleton isomers of PCHE and CHB or the products
Fig. 2. Concentrations of the reactant and products (A and C) as well as the product
selectivities (B and D) in the desulfurization of 2-PCHT at 240 °C over Ni2P under
of the heterolytic cleavage of the cycloalkyl C–S bond in 2-PCHT
4.0 MPa H2 in the absence (A and B) or presence (C and D) of 0.25 kPa piperidine as a [16]. The two compounds disappeared after the addition of a base
function of space–time. (piperidine). It should be noted that not only acid sites but also
336 X. Zhou et al. / Journal of Catalysis 383 (2020) 331–342

Table 3
Yields of the main products in the reaction of 2-PCHT at 240 °C and s = 0.74 g min/mol over Ni2P, MoP, and WP under H2 or Ar in the absence or presence of 0.25 kPa piperidine.

Yield (%)
In the absence of piperidine In the presence of piperidine
Catalyst Atmosphere 1-PCHE 3-PCHE CHB BP TH-DBT 1-PCHE 3-PCHE CHB BP
Ni2P H2 20.9 12.9 4.8 2.5 0.7 4.3 1.5 4.0 3.7
Ar 7.5 1.3 1.4 7.2 0.8 3.3 0.8 0.7 3.2
MoP H2 9.4 12.2 33.5 0.1 –a 5.9 11.3 19.7 0.2
Ar 15.5 5.8 3.1 6.1 3.4 4.0 1.9 2.8 3.0
WP H2 11.0 12.0 29.3 0.1 –a 4.0 7.5 15.9 0.5
Ar 24.8 6.4 3.6 9.0 3.6 4.4 2.4 3.3 2.8
a
Not detected.

metal surfaces are able to catalyze skeletal rearrangements of of oxidative insertion that occurs in the HDS of thiophenic com-
hydrocarbons [39,40]. Thus, because of the low acidity of these pounds catalyzed by organometallic compounds in homogeneous
bulk transition-metal phosphide catalysts and the low yields of systems [43]. Anyway, these results suggest that the breaking of
the isomerization products, we cannot exclude the possibility that the aryl C–S bond is also a reversible reaction.
the formation of CPEMB and CPMB is due to the metallic character Under Ar or in the absence of H2, CHB could be a product of the
of the transition-metal phosphides. disproportionation of PCHE. However, CHB did not behave like a
BP was detected as one of the major products in the desulfuriza- secondary product during the desulfurization of 2-PCHT under
tion of 2-PCHT over all three phosphide catalysts under Ar at 0 kPa Ar. As shown in Figs. 3B and 5B as well as Fig. S6B in the Supple-
piperidine. It was the most abundant product over Ni2P (Fig. 3A) mentary Material, the selectivities of CHB were nearly constant.
and the second most abundant product over MoP (Fig. 5A) and Actually, not only the selectivities of CHB but also those of the
WP (Fig. S6A in the Supplementary Material). BP can be formed other products were almost independent of space–time over MoP
by the dehydrogenation and/or the disproportionation of PCHEs. (Fig. 5B) and WP (Fig. S6B in the Supplementary Material). This is
The disproportionation of PCHEs yields BP and CHB with a BP/ further confirmed by the low reactivity of 1-PCHE and the very
CHB ratio of 0.5; that is, one PCHE molecule donates four hydrogen low yield of CHB (less than 0.3%) at 240 °C and s = 0.74 g min/mol
atoms to two PCHE molecules, forming one BP and two CHB mole- over the fresh and spent Ni2P catalysts under 4.0 MPa Ar. Hence,
cules [41]. However, the BP/CHB ratios were higher than 0.5 in mechanisms other than disproportionation of PCHE must be taken
most cases. More importantly, the selectivities of BP over the phos- into account for the formation of CHB under inert atmosphere. One
phide catalysts under Ar decreased with space–time, suggesting possible mechanism is homolytic C–S bond cleavage via free radi-
that it is unlikely to be a secondary product of the further reaction cal intermediates, which has been demonstrated to be the desulfu-
of PCHEs. To check this, we carried out the reaction of 1-PCHE at rization mechanism for the HDS of cyclopropylmethylthiol both in
240 °C and s = 0.74 g min/mol over Ni2P-H, Ni2P-Ar, and a fresh a homogeneous system [44] and in a heterogeneous system using
Ni2P catalyst under 4.0 MPa Ar. The conversion of 1-PCHE was less an actual CoMo sulfide HDS catalyst [45]. The other possibility is
than 3.5%. Although BP was the most abundant product over the the hydrogenolysis of 2-PCHT, because the dehydrogenation of
fresh Ni2P catalyst, its yield was less than 1.7% at s = 0.74 g min/- 2-PCHT under Ar generates H2. The present data are not sufficient
mol. Since the formation of BP from 2-PCHT involves the cleavage to verify or rule out this possibility, which will be studied in our
of the C–S bond and the dehydrogenation of the saturated hydro- subsequent work.
carbon ring, these results suggest that the dehydrogenation of Based on analysis of the product distributions, we propose the
the cyclohexane ring of 2-PCHT must occur prior to or simultane- following reaction networks for the desulfurization of 2-PCHT
ously with the cleavage of the C–S bond (the dehydrogenation–d under H2 and Ar over the transition-metal phosphides under the
esulfurization pathway). This explains the low yield or absence of reaction conditions of the present work (Scheme 1). 2-PCHT under-
BP in the HDS of 2-PCHT under H2, because the dehydrogenation goes HDS by three parallel pathways: b-elimination, hydrogenoly-
reaction is not favored under high hydrogen pressure. sis, and dehydrogenation. The dehydrogenation pathway also
One explanation for the formation of a minor amount of TH-DBT includes two parallel reactions: the dehydrogenation-ring closure
in the desulfurization of 2-PCHT over Ni2P (Fig. 3A), MoP (Fig. 5A), reaction and the dehydrogenation-desulfurization reaction with
and WP (Fig. S6A in the Supplementary Material) under Ar as well TH-DBT and BP as the products, respectively. The rate of the dehy-
as in the HDS of 2-PCHT over Ni2P (Fig. 2A) could be that ring clo- drogenation pathway is low (over Ni2P) or almost zero (over MoP
sure of 2-PCHT yields HH-DBT, which subsequently dehydro- and WP) under 4.0 MPa H2. Under inert atmosphere, the three par-
genates to TH-DBT. However, we never observed HH-DBT. We allel pathways of 2-PCHT desulfurization are b-elimination,
confirmed that the conversion of HH-DBT over Ni2P was almost hydrogenolysis or homolytic cleavage of the C–S bond, and
zero below 280 °C under 4.0 MPa Ar or H2. Therefore, TH-DBT is dehydrogenation.
likely formed by a dehydrogenation-ring closure pathway; that
is, 2-PCHT first dehydrogenates to 2-phenyl-1-cyclohexenethiol 4.2. Comparison of the catalytic performances of Ni2P, MoP, and WP
(2-PCHET), which then undergoes a ring closure reaction to form
TH-DBT. The ring closure reaction must be a quick reaction, The numbers of surface active sites of Ni2P, MoP, and WP were
because TH-DBT behaved like a primary product with nonzero ini- determined by CO chemisorption. The turnover frequencies (TOF)
tial selectivity. The mechanism of the ring closure of 2-PCHET of 2-PCHT under H2 and Ar over these catalysts at 0 and 0.25 kPa
remains a subject for future study. It could be a reverse reaction piperidine were calculated based on the 2-PCHT conversions at
of the hydrogenolysis of the aryl C–S bond or might follow a dehy- the lowest space–time (0.25 g min/mol) and their CO uptakes,
drogenative coupling mechanism. The latter is usually used to
describe the intermolecular and intramolecular coupling of silane TOF = [F 0
X 2PCHT ]/[wcat
(CO uptake)], ð1Þ
Si–H bonds with aromatic and aliphatic C–H bonds catalyzed by
where F0 is the initial molar flow rate of 2-PCHT, lmol/s; X2-PCHT is
metals [42]. This reaction can be regarded as a reverse reaction
the conversion of 2-PCHT; and wcat represents the weight of the cat-
 X. Zhou et al. / Journal of Catalysis 383 (2020) 331–342 337

100 18 alyst charged in g. The results are shown in Fig. 6. MoP and WP were
A much more active than Ni2P in the desulfurization of 2-PCHT both
15 under H2 and Ar. This seems to contradict previous reports that

Product concentration (%)

2-PCHT concentration (%)

80
1-PCHE BP
Ni2P is more active than MoP and WP in the HDS of DBT [17]. There
3-PCHE CPEMB 12 are two possibilities to explain this contradiction. (1) Compared
60 4-PCHE TH-DBT with MoP and WP, Ni2P is less active in the desulfurization of 2-
CHB 2-PCHT
9 PCHT but more active in the HDS of DBT. (2) The activity of Ni2P
40 is strongly inhibited by excess phosphorus present in the catalyst.
6 The P/Ni ratios of the Ni2P catalyst determined by ICP-AES and
XPS were 0.84 and ca. 2.2 (Table 2), respectively. They were both
20
3 higher than the stoichiometric value (0.5). Excess phosphorus has
to be used for the preparation of Ni2P by temperature-
0 0 programmed reduction of nickel phosphate precursors; otherwise
0.0 0.2 0.4 0.6 0.8 metal-rich nickel phosphides like Ni12P5 with low HDS activities
Space-time (g ·min/mol)
will be formed [46]. The large excess of phosphorus is known to
100 have a strong negative influence on the dispersion of Ni2P and its
B 1-PCHE 3-PCHE HDS performance [47]. On the other hand, the surface P/metal ratios
80 4-PCHE CHB of WP and MoP determined by XPS (Table 2) were close to their sto-
BP CPEMB ichiometric values. To clarify this question, it is necessary to prepare
TH-DBT
Ni2P with stoichiometric P/Ni ratio by conventional TPR, which is
Selectivity (%)

60
challenging.
Under H2, the activity of the phosphides decreased in the order
40 MoP > WP > Ni2P, while under Ar the order was WP > MoP > Ni2P.
At 0 kPa piperidine and s = 0.25 g min/mol, the turnover frequen-
20 cies of 2-PCHT over Ni2P and MoP under H2 were higher than under
Ar, whereas that over WP under Ar was higher than under H2.
Piperidine strongly inhibited the desulfurization of 2-PCHT over
0 all three phosphide catalysts, and WP seemed to be more sensitive
0.2 0.3 0.4 0.5 0.6 0.7 0.8
Space-time (g·min/mol) to piperidine inhibition than MoP. The TOF of WP under Ar was
higher than that of MoP at 0 kPa piperidine, but the turnover fre-
100 10 quencies of the two catalysts were about equal in the presence of
0.25 kPa piperidine. This is consistent with our previous finding
2-PCHT concentration (%)

Product concentration (%)

80 C 8 that WP was more sensitive to piperidine inhibition than MoP in

1-PCHE the HDS of DBT [16].
3-PCHE The desulfurization of 2-PCHT over Ni2P under H2 mainly took
60 CHB 6
BP
place by b-elimination to 1-PCHE and 3-PCHE (Fig. 2A). The total
2-PCHT selectivity of PCHEs was around 80% (Fig. S4 in the Supplementary
40 4 Material). Piperidine strongly inhibited the HDS of 2-PCHT over
Ni2P mainly by inhibiting the b-elimination pathway. At
20 2 s = 0.74 g min/mol, the total yield of 1-PCHE and 3-PCHE decreased
significantly from 33.8% in the absence of piperidine to 5.8% in the
presence of piperidine (Table 3). In contrast, the addition of piper-
0 0 idine hardly inhibited the formation of CHB and BP or the
0.0 0.2 0.4 0.6 0.8
Space-time (g ·min/mol ) hydrogenolysis and dehydrogenation pathways. The yield of BP
in the presence of piperidine was even slightly higher than in the
100 absence of piperidine (Fig. 2A and C, as well as Table 3). Conse-
D 1-PCHE quently, the total selectivity of PCHEs decreased from ca. 80% at
80 3-PCHE 0 kPa piperidine to ca. 45% at 0.25 kPa piperidine (Fig. S4 in the
CHB
Supplementary Material). Over Ni2P under 4.0 MPa Ar, the selectiv-
Selectivity (%)

BP
ity for BP was very close to that of 1-PCHE (Fig. 3B), indicating that
60
the dehydrogenation pathway is as fast as the b-elimination path-
way. These two pathways were equally inhibited by piperidine,
40 because no significant change in product selectivities was observed
before and after the addition of piperidine (Fig. 3D), except for the
20 disappearance of TH-DBT (the dehydrogenation-ring closure pro-
duct) as well as 4-PCHE and CPEMB. These compounds were
detected in trace amounts in the desulfurization of 2-PCHT over
0
0.2 0.3 0.4 0.5 0.6 0.7 0.8 Ni2P under Ar in the absence of piperidine (Fig. 3A).
Space-time (g·min/mol) The product distributions of 2-PCHT desulfurization over MoP
and WP (the two group VIB transition-metal phosphides) were
Fig. 3. Concentrations of the reactant and products (A and C) as well as the product similar, and very different from that over Ni2P. MoP and WP exhib-
selectivities (B and D) in the desulfurization of 2-PCHT at 240 °C over Ni2P under
ited much higher hydrogenolysis activities than Ni2P. The selectiv-
4.0 MPa Ar in the absence (A and B) or presence (C and D) of 0.25 kPa piperidine as a
function of space–time. ities of CHB and the total selectivities of PCHEs were both around
338 X. Zhou et al. / Journal of Catalysis 383 (2020) 331–342

100 100
A B 1-PCHE 3-PCHE
80 80 4-PCHE CHB
BP CPEMB
Concentration (%)
CPMB

Selectivity (%)
1-PCHE
60 3-PCHE 60
4-PCHE
CHB
40 BP 40
CPEMB
CPMB
20 2-PCHT 20

0 0
0.0 0.2 0.4 0.6 0.8 0.2 0.3 0.4 0.5 0.6 0.7 0.8
Space-time (g·min/mol) Space-time (g·min/mol)
100 100
C D 1-PCHE
80 80 3-PCHE
CHB
BP
Concentration (%)

Selectivity (%)
1-PCHE
60 60
3-PCHE
CHB
BP
40 2-PCHT 40

20 20

0 0
0.0 0.2 0.4 0.6 0.8 0.2 0.3 0.4 0.5 0.6 0.7 0.8
Space-time (g·min/mol) Space-time (g·min/mol)

Fig. 4. Concentrations of the reactant and products (A and C) as well as the product selectivities (B and D) in the desulfurization of 2-PCHT at 240 °C over MoP under 4.0 MPa
H2 in the absence (A and B) or presence (C and D) of 0.25 kPa piperidine as a function of space–time.

100 100
A B 1-PCHE 3-PCHE
80 80 4-PCHE CHB
BP CPEMB
Selectivity (%)
Concentration (%)

TH-DBT
60 1-PCHE BP 60
3-PCHE CPEMB
4-PCHE 2-PCHT
40 CHB TH-DBT 40

20 20

0 0
0.0 0.2 0.4 0.6 0.8 0.2 0.3 0.4 0.5 0.6 0.7 0.8
Space-time (g·min/mol) Space-time (g·min/mol)

100 10 100

D
2-PCHT concentration (%)

1-PCHE
Product concentration (%)

3-PCHE
80
1-PCHE
C 8 80
4-PCHE
3-PCHE CHB
Selectivity (%)

4-PCHE BP
60 6 60 TH-DBT
CHB
BP
TH-DBT
40 2-PCHT 4 40

20 2 20

0 0 0
0.0 0.2 0.4 0.6 0.8 0.2 0.3 0.4 0.5 0.6 0.7 0.8
Space-time (g·min/mol) Space-time (g·min/mol)

Fig. 5. Concentrations of the reactant and products (A and C) as well as the product selectivities (B and D) in the desulfurization of 2-PCHT at 240 °C over MoP under 4.0 MPa
Ar in the absence (A and B) or presence (C and D) of 0.25 kPa piperidine as a function of space–time.
 X. Zhou et al. / Journal of Catalysis 383 (2020) 331–342 339

Scheme 1. Reaction networks of 2-PCHT desulfurization at 240 °C over the transition-metal phosphide catalysts under 4.0 MPa H2 and Ar. The notations of the rate constants
of some steps in the network under H2 are given.

under Ar (Table 3). As a consequence, MoP performed best in the

0.5 0 kPa piperidine HDS of 2-PCHT (Fig. 6). The addition of piperidine hardly affected
0.25 kPa piperidine the selectivity of the three reaction pathways over MoP under H2
0.4 (Fig. 4), except for the increase in the 3-PCHE/1-PCHE ratio and
the disappearance of 4-PCHE, CPMB, and CPEMB. As a base, piper-
idine is expected to inhibit isomerization reactions. For WP under
TOF (S-1)

0.3 H2, the b-elimination pathway was more inhibited by piperidine

than the hydrogenolysis pathway (Fig. S5 in the Supplementary
0.2 Material). The total selectivity for PCHEs at 0.25 kPa piperidine
was about 4–8% lower than that at 0 kPa piperidine (Fig. S4 in
the Supplementary Material).
0.1
It is worth noting that the ratios of the PCHE isomers over the
phosphide catalysts under Ar differed from those under H2.
0.0 Because of the conjugation between the double bond and the aro-
H2 Ar H2 Ar H2 Ar
matic ring in 1-PCHE, it is more energetically stable than 3-PCHE.
Ni2P MoP WP
However, the b-elimination of the H atom attached to carbon atom
C(2) of 2-PCHT to form 1-PCHE (the b1 pathway) is more sterically
Fig. 6. Turnover frequencies of 2-PCHT desulfurization at 240 °C and s = 0.25 g
min/mol over Ni2P, MoP, and WP under 4.0 MPa H2 or Ar in the absence or presence hindered than the b-elimination of the H atom attached to carbon
of 0.25 kPa piperidine. atom C(6) to yield 3-PCHE (the b2 pathway). For example, trans
elimination of H2S from trans-2-PCHT can only occur by b-H elim-
ination from C(6), giving 3-PCHE. Therefore, the ratio of the PCHE
50% (Fig. 4B, as well as Figs. S4 and S5B in the Supplementary isomers is affected by more than one factor. Over Ni2P, the b-
Material), demonstrating that the hydrogenolysis and b- elimination reaction preferred to occur through the b1 pathway,
elimination pathways are equally fast. b-Elimination dominates both under H2 and Ar. The 3-PCHE/1-PCHE ratio under H2 (ca.
the desulfurization of 2-PCHT under Ar over MoP and WP, and 0.5) was much higher than that under Ar (ca. 0.15), indicating that
the dehydrogenation pathway is faster than the hydrogenolysis the b2 pathway was favored under H2 (Figs. 2A and 3A). The situ-
or homolytic cleavage pathway (Fig. 5 and Fig. S6 in the Supple- ation is quite different over MoP and WP. The 3-PCHE/1-PCHE
mentary Material). The b-elimination pathway was more inhibited ratios were close to 1 under H2 (Fig. 4A and Fig. S5A in the Supple-
by piperidine than the latter two pathways. Under Ar, the total mentary Material). At 0.25 kPa piperidine the ratios were even
selectivity of PCHEs decreased from more than 61% over MoP or higher than one (Fig. 4C and Fig. S5C in the Supplementary Mate-
64% over WP at 0 kPa piperidine to about 52% or 54% at 0.25 kPa rial). Under Ar, however, the 3-PCHE/1-PCHE ratios were much
piperidine, respectively (Fig. S4 in the Supplementary Material). lower than one (Fig. 5 and Fig. S6 in the Supplementary Material).
The CO uptake of MoP (2.2 lmol/g) and WP (2.4 lmol/g) was This suggests that the surface active sites or reactions under H2 and
about equal, which allowed direct comparison of their catalytic Ar are different. To fully understand these important questions,
performance based on product distributions. Under Ar, WP showed further experimental and theoretical work is needed.
higher b-elimination activity than MoP. The total yield of 1-PCHE We studied the kinetics of the desulfurization of 2-PCHT by a
and 3-PCHE over WP was about 10% higher than that over MoP pseudo-first-order model, which is usually applied to describe
(Fig. 5A and Fig. S6A in the Supplementary Material, as well as the HDS of DBT and its derivatives [12,48]. The rate equation of
Table 3). The dehydrogenation activity of WP was also slightly the pseudo-first-order model for a plug-flow reactor is
higher than that of MoP (Fig. 5A and Fig. S6A in the Supplementary
Material). Hence, WP exhibited the highest desulfurization activity -ln(1 - X 2PCHT ) = k2PCHT
s, ð2Þ
under Ar (Fig. 6). Under 4.0 MPa H2, MoP was the most active of the
three phosphide catalysts in the hydrogenolysis of 2-PCHT to CHB where k2-PCHT is the rate constant, mol g1 min1. ln(1  X2-PCHT) is
(Fig. 6). Additionally, the b-elimination of 2-PCHT to PCHEs over then plotted against s, and the rate constant k2-PCHT can be deter-
MoP under H2 was about as fast as that under Ar, whereas the b- mined from the slope of each fitted line. The regression analysis
elimination of 2-PCHT over WP under H2 was slower than that (Figs. S7 and S8 in the Supplementary Material) demonstrated that
340 X. Zhou et al. / Journal of Catalysis 383 (2020) 331–342

the HDS of 2-PCHT over the phosphide catalysts can be fitted by a compounds) under H2, or by the reduction of a sulfur-containing
first-order rate model, but the desulfurization of 2-PCHT under Ar precursor (Ni2P2S6) [21,22,31,32]. All these methods require hydro-
cannot be described by the power-law rate expression. The calcu- gen. As revealed by the S2p XPS spectra (Fig. 1), sulfur species
lated concentration-s curves of 2-PCHT desulfurization under H2 assigned to those in phosphosulfide phases were detected in the
and Ar from the obtained pseudo-first-order rate constants are surfaces of the spent phosphide catalysts after the desulfurization
shown in Figs. S9 and S10 in the Supplementary Material, respec- of 2-PCHT both under H2 and under Ar. This not only confirms the
tively. Consistent with regression analysis, the experimental data formation of transition-metal phosphosulfide phases, but also indi-
of 2-PCHT concentrations under H2 agreed with the calculated cates that the phosphosulfide phases can be formed by direct reac-
concentration-s curves, whereas those under Ar deviated largely tion of the phosphide catalysts with the sulfur-containing
from the curves predicted by the first-order model. This is reason- compounds, even in the absence of hydrogen, which provides a
able. As shown in Scheme 1, the reaction network of 2-PCHT desul- more practical and safer method for their preparation. According
furization involves several kinds of reactions that proceed on to the covalent nature of transition-metal phosphides and the
different active sites. In the present work, we only calculated the XPS analysis, which demonstrated that the metal, phosphorus,
rate constants of the HDS of 2-PCHT by assuming pseudo-first- and sulfur species in the phosphosulfide phases were not in their
order kinetics. The constants of the three primary parallel steps ionic and elemental states (Fig. 1 and Fig. S3 in the Supplementary
and some of the following steps were obtained from these rate con- Material), the phosphosulfide phases might be covalent in nature.
stants and the initial product selectivities estimated by extrapolat- A typical S2p XPS spectrum is composed of a doublet structure
ing the selectivity-s curves to s = 0. A detailed study of the kinetics due to the presence of S2p1/2 and 2p3/2 levels with spin–orbit split-
of 2-PCHT desulfurization will be an important subject of our future ting of 1.18 eV and area ratio of 1:2. The S2p XPS spectra of all the
research. These rate constants are helpful for a semiquantitative spent catalysts featured a broad asymmetric peak with a shoulder
comparison of the HDS performances of the phosphide catalysts. at higher binding energy. These peaks cannot be fitted well by con-
As shown in Table 4, the hydrogenolysis activity of Ni2P was very sidering single sulfur species (Fig. S11 in the Supplementary Mate-
low. The rate of the hydrogenolysis pathway (khyg) at 0 kPa piperi- rial), suggesting that several sulfur species may be present in the
dine was about one order of magnitude lower than that of the b- phosphosulfide phases. A DFT calculation revealed that two kinds
elimination pathway (kb). It was even lower than the rate constant of sulfur species may be present in the phosphosulfide phase, the
of the dehydrogenation pathway (kd). MoP and WP all exhibited sulfur species that replaces the phosphorus species in the phos-
high hydrogenolysis activity. Their hydrogenolysis pathway rate phide phase and atomic sulfur deposited on the threefold hollow
constants were 8.0 and 6.3 times as high as that of Ni2P, respec- sites [49]. Therefore, we fitted the S2p spectra by assuming the
tively. The rate constants of the b-elimination pathway of all the coexistence of two sulfur species (the S-I and S-II species) in the
three phosphide catalysts were comparable to each other. Ni2P phosphosulfide phases. The following assumptions were made in
was more active than MoP and WP in the dehydrogenation of 2- the curve fitting. (1) A Shirley background was used and the peaks
PCHT. As far as the b-elimination pathway is concerned, the ratio were all fitted with Gaussian functions. (2) The energy difference
of the rate constant of the b1 pathway (kb1) to that of the b2 path- between the S2p1/2 and 2p3/2 states, as well as their area ratio, were
way (kb2) in the absence of piperidine decreased in the order Ni2P fixed to the theoretical values, 1.18 eV and 1:2, respectively. (3)
(2.8) > WP (1.6) > MoP (1.3). The b1 pathway is favored over Ni2P. The full width at half-maximum (FWHM) values of the S2p1/2
Both the hydrogenolysis pathway and the b-elimination pathway and 2p3/2 doublet peaks of each sulfur species were equal and were
were inhibited by piperidine, whereas the dehydrogenation path- fixed for all the catalysts (Table 5). That is, the FWHM of the dou-
way was hardly affected (Ni2P) or even improved (MoP and WP) blet peaks of the S-I species was fixed to 1.5, while that of the S-II
after the addition of piperidine. In summary, piperidine showed dif- species was fixed to 1.0. According to the fitting results (Fig. 1 and
ferent effects on different reaction pathways in the desulfurization Table 5), the S-II species has a lower oxidation state than the S-I
of 2-PCHT over different phosphide catalysts, which can be an indi- species. The binding energies of the doublets of the S-I and S-II spe-
cation of the complexity of the cleavage of the cycloalkyl C–S bond cies were close to those of the sulfur species in a thiolate-type
in 2-PCHT and the active sites in the surfaces of the transition-metal environment and a sulfide species (S2), respectively [32]. No sul-
phosphide catalysts. fur oxide was detected. In the fields of optoelectronic devices and
photovoltaics, sulfur passivation by (NH4)2S or H2S is used to
4.3. The phosphosulfide phases of the transition-metal catalysts improve electronic properties of group III–V semiconductors, such
as InP and GaP [50]. We found in our previous work [30] that a
It has been established that the real active phases for the work- H2S-passivated Ni2P/MCM-41 catalyst showed higher HDS activity
ing Ni2P, MoP, and WP catalysts during HDS reactions are phospho- than its O2-passivated counterpart. The H2S-passivated Ni2P/MCM-
sulfide phases. These phases can be obtained during the course of 41 catalyst was stable up to 150 days (kept in air). Moreover, it was
HDS reactions, or by the treatment of the transition-metal phos- not necessary to re-reduce the H2S-passivated Ni2P/MCM-41 cata-
phide with sulfidation agents (H2S or organic sulfur-containing lyst prior to the HDS reaction. This suggests that transition-metal

Table 4
Pseudo-first-order rate constants of some reactions in the network of the desulfurization of 2-PCHT at 240 °C over Ni2P, MoP, and WP under 4.0 MPa H2 in the absence or presence
of piperidine.

Catalyst Piperidine partial pressure (kPa) Pseudo-first-order rate constant (mol g1 min1)
kba kb1 kb2 khyg kdb kdd kdrc
Ni2P 0 0.57 0.42 0.15 0.08 0.11 0.10 0.01
0.25 0.097 0.091 0.006 0.022 0.10 0.10 0.004
MoP 0 0.63 0.35 0.28 0.64 0.007 0.007 0
0.25 0.39 0.19 0.20 0.27 0.01 0.01 0
WP 0 0.66 0.41 0.25 0.51 0.01 0.01 0
0.25 0.24 0.11 0.13 0.23 0.02 0.02 0
a
kb is the sum of kb1 and kb2.
b
kd is the sum of kdd and kdrc.
 X. Zhou et al. / Journal of Catalysis 383 (2020) 331–342 341

Table 5
Deconvolution results of the S2p XPS spectra of the spent Ni2P, MoP, and WP catalysts.

Catalyst S-I S-II S-II/S-I

2p3/2 2p1/2 2p3/2 2p1/2
BE (eV) FWHM (eV) BE (eV) FWHM (eV) BE (eV) FWHM (eV) BE (eV) FWHM (eV)
Ni2P-H 162.2 1.5 163.4 1.5 161.4 1.0 162.6 1.0 1.72
Ni2P-Ar 162.2 1.5 163.4 1.5 161.1 1.0 162.3 1.0 3.24
MoP-H 162.7 1.5 163.9 1.5 161.8 0.9 163.0 0.9 2.50
MoP-Ar 162.6 1.5 163.8 1.5 161.6 1.0 162.8 1.0 1.74
WP-H 162.5 1.5 163.7 1.5 161.6 1.0 162.8 1.0 2.91
WP-Ar 162.0 1.5 163.2 1.5 161.2 0.9 162.4 0.9 2.03

phosphosulfide phases could be more stable under air or less 2-PCHT underwent hydrodesulfurization over Ni2P mainly via the
affected by air than their corresponding sulfur-free phases. b-elimination pathway. Under Ar, the b-elimination and dehydro-
It must be noted that since the nature of the phosphosulfide genation pathways were about equal. Over MoP and WP, the
phase is still not well understood. Our fitting results cannot be hydrogenolysis pathway was as fast as the b-elimination pathway
used to interpret the real structures of the phosphosulfide phases. under H2, while b-elimination dominated the desulfurization of 2-
They are only helpful for a quantitative comparison of the phos- PCHT under Ar. Piperidine inhibited different reaction pathways to
phosulfide phases formed under H2 and Ar. The different binding different extents, which may indicate the complexity of the cleav-
energies and the S-II/S-I ratios of the sulfur species suggest differ- age of the cycloalkyl C–S bond in 2-PCHT and the active sites in the
ent interactions between sulfur and the phosphide catalysts surfaces of the transition-metal phosphides. Phosphosulfide
(Table 5). The S-II/S-I ratios of the spent MoP and WP catalysts phases were formed under both H2 and Ar, demonstrating that
under H2 were higher than those under Ar. Moreover, the S/M their formation may be due to the reaction of the phosphide cata-
ratios of MoP-H and WP-H were also higher than those of MoP- lysts with the sulfur-containing compounds and that H2 is not nec-
Ar and WP-Ar, respectively (Table 2), which suggests that the sul- essary for the reaction. The sulfidation degree of spent MoP and
fidation degree of the spent MoP and WP is higher under H2 than WP under H2 was higher than under Ar, whereas the formation
under Ar. Although Ni2P-Ar possessed the highest S/M ratio among of the nickel phosphosulfide phase was difficult under H2.
all the spent catalysts, the S/M ratio of Ni2P-H was low (Table 2). It
was only 1/5 of the value of Ni2P-Ar, and was very close to the S/M Declaration of Competing Interest
ratio of WP-Ar, the lowest one (Table 2). The S-II/S-I ratio of Ni2P-H
was the lowest among all the spent catalysts (Table 5). Thus, the The authors declare that they have no known competing finan-
formation of nickel phosphosulfide, particularly the formation of cial interests or personal relationships that could have appeared
the sulfur-containing species with sulfur in a lower oxidation state to influence the work reported in this paper.
(the S-II species), is difficult under H2. This suggests that the nickel
phosphosulfide is more easily reduced under H2 than the molybde- Acknowledgments
num and tungsten phosphosulfides. This may explain why the for-
mation of the nickel phosphosulfide is more difficult than that of This research was financially supported by the Natural Science
the molybdenum phosphosulfide [51]. By means of CO-IR, Sun Foundation of China (21673029) and the Natural Science Founda-
et al. reported that a MoP/SiO2 catalyst can be sulfided by a mix- tion of Tianjin (19JCZDJC31700).
ture of thiophene/H2 (10/100 Torr), but more severe sulfiding con-
ditions were required to transform a Ni2P/SiO2 catalyst to its
Appendix A. Supplementary material
phosphosulfide counterpart. They obtained the nickel phosphosul-
fide phase using a high concentration H2S/H2 mixture (10 vol%) as
Supplementary data to this article can be found online at
the sulfidation agent [51]. The different sulfidation behavior of
https://doi.org/10.1016/j.jcat.2020.01.023.
Ni2P, MoP, and WP might be due to their different nature. The sul-
fidation behavior of the two group VIB transition-metal phos-
References
phides (MoP and WP) was similar, and different from that of
Ni2P, a group VIII transition-metal phosphide. At present we can- [1] R.V. Mom, J.N. Louwen, J.W.M. Frenken, I.M.N. Groot, In situ observations of an
not give a definite answer to this important question, which will active MoS2 model hydrodesulfurization catalyst, Nat. Commun. 10 (2019) 1–
be studied in our subsequent work. 8.
[2] F. van Looij, P. van der Laan, W.H.J. Stork, D.J. DiCamillo, J. Swain, Key
parameters in deep hydrodesulfurization of diesel fuel, Appl. Catal. A 170
(1998) 1–12.
5. Conclusions [3] E.M. Morales-Valencia, C.O. Castillo-Araiza, S.A. Giraldo, V.G. Baldovino-
Medrano, Kinetic assessment of the simultaneous hydrodesulfurization of
dibenzothiophene and the hydrogenation of diverse polyaromatic structures,
The desulfurization of 2-PCHT over the transition metal phos- ACS Catal. 8 (2018) 3926–3942.
phides at 240 °C under 4.0 MPa H2 proceeded mainly through three [4] L. Zhang, W. Fu, Q. Yu, T. Tang, Y. Zhao, H. Zhao, Y. Li, Ni2P clusters on zeolite
nanosheet assemblies with high activity and good stability in the
parallel pathways: b-elimination, hydrogenolysis, and dehydro- hydrodesulfurization of 4,6-dimethyldibenzothiophene, J. Catal. 338 (2016)
genation. The dehydrogenation pathway consisted of two parallel 210–221.
reactions: the dehydrogenation-ring closure reaction with TH- [5] J.H. Sinfelt, Catalytic hydrogenolysis on metals, Catal. Lett. 9 (1991) 159–171.
[6] J.M.J.G. Lipsch, G.C.A. Schuit, The CoO-MoO3-Al2O3 catalyst: III. Catalytic
DBT as the product and the dehydrogenation-desulfurization reac-
properties, J. Catal. 15 (1969) 179–189.
tion with BP as the product. Under 4.0 MPa Ar, the three parallel [7] E.J.M. Hensen, M.J. Vissenberg, V.H.J. de Beer, J.A.R. van Veen, R.A. van Santen,
pathways are b-elimination, homolytic C–S bond cleavage or Kinetics and mechanism of thiophene hydrodesulfurization over carbon-
hydrogenolysis, and dehydrogenation. MoP and WP exhibited supported transition metal sulfides, J. Catal. 163 (1996) 429–435.
[8] Th. Weber, J.A.R. van Veen, A density functional theory study of the
much higher activity than Ni2P in the desulfurization of 2-PCHT. hydrodesulfurization reaction of dibenzothiophene to biphenyl on a single-
They were the most active catalysts under H2 and Ar, respectively. layer NiMoS cluster, Catal. Today 130 (2008) 170–177.
342 X. Zhou et al. / Journal of Catalysis 383 (2020) 331–342

[9] D.C. Grenoble, The chemistry and catalysis of the toluene hydrodealkylation [31] S. Tian, X. Li, A. Wang, R. Prins, Y. Chen, Y. Hu, Facile preparation of Ni2P with a
reaction: I. The specific activities and selectivities of group VIIB and group VIII sulfur-containing surface layer by low-temperature reduction of Ni2P2S6,
metals supported on alumina, J. Catal. 56 (1979) 32–39. Angew. Chem. Int. Ed. 55 (2016) 4030–4034.
[10] D.C. Grenoble, The chemistry and catalysis of the toluene hydrodealkylation [32] J. Kibsgaard, T.F. Jaramillo, Molybdenum phosphosulfide: An active, acid-
reaction: II. Kinetic analysis, J. Catal. 56 (1979) 40–46. stable, earth abundant catalyst for the hydrogen evolution reaction, Angew.
[11] F. Bataille, J.L. Lemberton, P. Michaud, G. Pérot, M. Vrinat, M. Lemaire, E. Chem. 126 (2014) 14661–14665.
Schulz, M. Breysse, S. Kasztelan, Alkyldibenzothiophenes [33] X. Li, S. Tian, A. Wang, R. Prins, C. Li, Y. Chen, XPS study of a bulk WP
hydrodesulfurization-promoter effect, reactivity, and reaction mechanism, J. hydrodesulfurization catalyst, J. Catal. 352 (2017) 557–561.
Catal. 191 (2000) 409–422. [34] R.L. Wilson, C. Kemball, Catalytic reactions of methyl mercaptan on disulfides
[12] Y. Sun, R. Prins, Mechanistic studies and kinetics of the hydrodesulfurization of of molybdenum and tungsten, J. Catal. 3 (1964) 426–437.
dibenzothiophene on Co–MoS2/c–Al2O3, J. Catal. 267 (2009) 193–201. [35] P. Kieran, C. Kemball, Catalytic reactions of ethyl mercaptan on disulfides of
[13] J.G. Jang, Y.-K. Lee, Promotional effect of Ga for Ni2P catalyst on molybdenum and tungsten, J. Catal. 4 (1965) 380–393.
hydrodesulfurization of 4,6-DMDBT, Appl. Catal. B 250 (2019) 181–188. [36] F. Rota, R. Prins, Role of hydrogenolysis and nucleophilic substitution in
[14] X. Li, J. Bai, A. Wang, R. Prins, Y. Wang, Hydrodesulfurization of hydrodenitrogenation over sulfided NiMo/c-Al2O3, J. Catal. 202 (2001) 195–
dibenzothiophene and its hydrogenated intermediates over bulk Ni2P, Top. 199.
Catal. 54 (2011) 290–298. [37] T. Todorova, R. Prins, Th. Weber, A density functional theory study of the
[15] J. Bai, X. Li, A. Wang, R. Prins, Y. Wang, Hydrodesulfurization of hydrogenolysis and elimination reactions of C2H5SH on the catalytically active
dibenzothiophene and its hydrogenated intermediates over bulk MoP, J. (100) edge of 2H-MoS2, J. Catal. 246 (2007) 109–117.
Catal. 287 (2012) 161–169. [38] Y.-K. Lee, S.T. Oyama, Bifunctional nature of a SiO2-supported Ni2P catalyst for
[16] L. Yang, X. Li, A. Wang, R. Prins, Y. Chen, X. Duan, Hydrodesulfurization of hydrotreating: EXAFS and FTIR studies, J. Catal. 239 (2006) 376–389.
dibenzothiophene, 4,6-dimethyldibenzothiophene, and their hydrogenated [39] F.G. Gault, Mechanisms of skeletal isomerization of hydrocarbons on metals,
intermediates over bulk tungsten phosphide, J. Catal. 330 (2015) 330–343. Adv. Catal. 30 (1981) 1–95.
[17] S.T. Oyama, Novel catalysts for advanced hydroprocessing: Transition metal [40] G. Maire, F. Garin, Mechanisms of skeletal rearrangements of hydrocarbons on
phosphides, J. Catal. 216 (2003) 343–352. metal: elementary steps, J. Mol. Catal. 48 (1988) 99–116.
[18] R. Prins, M.E. Bussell, Metal phosphides: preparation, characterization and [41] S. Carrā, P. Beltrame, V. Ragaini, Kinetics of the hydrogen transfer reaction of
catalytic reactivity, Catal. Lett. 142 (2012) 1413–1436. cyclohexene on palladium powder, J. Catal. 3 (1964) 353–362.
[19] S.J. Sawhill, D.C. Phillips, M.E. Bussell, Thiophene hydrodesulfurization over [42] N. Tsukada, J.F. Hartwig, Intermolecular and intramolecular, platinum-
supported nickel phosphide catalysts, J. Catal. 215 (2003) 208–219. catalyzed, acceptorless dehydrogenative coupling of hydrosilanes with aryl
[20] D.C. Phillips, S.J. Sawhill, R. Self, M.E. Bussell, Synthesis, characterization, and and aliphatic methyl C-H bonds, J. Am. Chem. Soc. 127 (2005) 5022–5023.
hydrodesulfurization properties of silica-supported molybdenum phosphide [43] J.J. Garcia, A. Arevalo, S. Capella, A. Chehata, M. Hernandez, V. Montiel, G.
catalysts, J. Catal. 207 (2002) 266–273. Picazo, F. Del Rio, R.A. Toscano, H. Adams, P.M. Maitlis, Analysis of a
[21] K.K. Bando, T. Wada, T. Miyamoto, K. Miyazaki, S. Takakusagi, Y. Koike, Y. hydrodesulfurization process – 2 [1]. The reactions of 2- and 3-
Inada, M. Nomura, A. Yamaguchi, T. Gott, S.T. Oyama, K. Asakura, Combined methylthiophenes with tris(triethylphosphine)platinum(0), Polyhedron 16
in situ QXAFS and FTIR analysis of a Ni phosphide catalyst under (1997) 3185–3195.
hydrodesulfurization conditions, J. Catal. 286 (2012) 165–171. [44] S.H. Druker, M.D. Curtis, Activation of C-S bond homolysis by coordination to a
[22] J. Bai, X. Li, A. Wang, R. Prins, Y. Wang, Different role of H2S and Mo/Co/S-containing cluster, J. Am. Chem. Soc. 117 (1995) 6366–6367.
dibenzothiophene in the incorporation of sulfur in the surface of bulk MoP [45] K.E. Dungey, M.D. Curtis, Homolytic C-S bond cleavage on a heterogeneous Co/
during hydrodesulfurization, J. Catal. 300 (2013) 197–200. Mo/S hydrodesulfurization catalyst, J. Am. Chem. Soc. 119 (1997) 842–843.
[23] H. Wang, R. Prins, Hydrodesulfurization of dibenzothiophene and its [46] A. Wang, L. Ruan, Y. Teng, X. Li, M. Lu, J. Ren, Y. Wang, Y. Hu,
hydrogenated intermediates over sulfided Mo/c-Al2O3, J. Catal. 258 (2008) Hydrodesulfurization of dibenzothiophene over siliceous MCM-41-supported
153–164. nickel phosphide catalysts, J. Catal. 229 (2005) 314–321.
[24] H.L. Goering, D.I. Relyea, D.W. Larsen, The stereochemistry of radical additions. [47] S.J. Sawhill, K.A. Layman, D.R. Van Wyk, M.H. Engelhard, C. Wang, M.E. Bussell,
III. The radical addition of hydrogen sulfide, thiophenol and thioacetic acid to Thiophene hydrodesulfurization over nickel phosphide catalysts: effect of the
1-chlorocyclohexene, J. Am. Chem. Soc. 78 (1956) 348–353. precursor composition and support, J. Catal. 231 (2005) 300–313.
[25] O. Levenspiel, Chemical Reaction Engineering, third ed., Wiley, 1998, p. 394. [48] Y. Wang, Z. Sun, A. Wang, L. Ruan, M. Lu, J. Ren, X. Li, C. Li, Y. Hu, P. Yao, Kinetics
[26] I.I. Abu, K.J. Smith, The effect of cobalt addition to bulk MoP and Ni2P catalysts of hydrodesulfurization of dibenzothiophene catalyzed by sulfided Co-Mo/
for the hydrodesulfurization of 4,6-dimethyldibenzothiophene, J. Catal. 241 MCM-41, Ind. Eng. Chem. Res. 43 (2004) 2324–2329.
(2006) 356–366. [49] A.E. Nelson, M. Sun, A.S.M. Junaid, On the structure and composition of the
[27] J.H. Scofield, Hartree-Slater subshell photoionization cross-sections at 1254 phosphosulfide overlayer on Ni2P at hydrotreating conditions, J. Catal. 241
and 1487 eV, J. Electron Spectrosc. Relat. Phenom. 8 (1976) 129–137. (2006) 180–188.
[28] X. Li, Z. Sun, A. Wang, X. Yang, Y. Wang, Effect of TiO2 on the [50] S. Sloboshanin, R.K. Gebhardt, J.A. Schaefer, T. Chasse, Photoemission study of
hydrodesulfurization performance of bulk Ni2P, Appl. Catal. A 417–418 the interface reaction between Ag and H2S treated InP (001), Surf. Sci. 431
(2012) 19–25. (1999) 252–259.
[29] L. Yang, X. Li, A. Wang, R. Prins, Y. Wang, Y. Chen, X. Duan, [51] F. Sun, W. Wu, Z. Wu, J. Guo, Z. Wei, Y. Yang, Z. Jiang, F. Tian, C. Li,
Hydrodesulfurization of 4,6-dimethyldibenzothiophene and its hydrogenated Dibenzothiophene hydrodesulfurization activity and surface sites of silica-
intermediates over bulk Ni2P, J. Catal. 317 (2014) 144–152. supported MoP, Ni2P, and Ni-Mo-P catalysts, J. Catal. 228 (2004) 298–310.
[30] X. Duan, Y. Teng, A. Wang, V.M. Kogan, X. Li, Y. Wang, Role of sulfur in
hydrotreating catalysis over nickel phosphide, J. Catal. 261 (2009) 232–240.