On the multifaceted roles of NiSx in hydrodearomatization reactions catalyzed by unsupported Ni-promoted MoS2

Article abstract of DOI:10.1016/j.jcat.2020.08.026

A series of unsupported Ni-Mo sulfide catalysts with varying Ni contents (0.13–0.72 mol_Ni mol_Ni+Mo^-1) was post-synthetically treated with concentrated HCl to remove large crystallites of accessible NiS_x. These sulfide particles inevitably form and grow at Ni concentrations required for the synthesis, but generally have very low activities. In all cases, Ni concentrations were greatly reduced by the HCl treatment. While this ‘leaching’ strategy successfully improved the reaction rates of high Ni-content (>0.4 mol_Ni mol_metal^-1) catalysts for the hydrogenation of phenanthrene, modest to drastic decreases in catalytic rates (×0.1–0.8) were registered for catalysts with lower Ni concentrations. For the lowest Ni-loaded catalyst (0.05 mol_Ni mol_metal^-1), HCl treatment caused a dramatic loss of specific surface area and catalytic activity by more than a factor of 6 and shifted the selectivity pattern to that of pure MoS₂. These observations allow us to conclude that Ni atoms incorporated at the slab edges are inherently susceptible to HCl attack. NiS_x, however, are the preferential sites at which HCl induces dissolution. Experiments with inter-particle mixtures and segregated beds of NiS_x and MoS₂ demonstrate that NiS_x not only activates H₂, but also acts as a reservoir to dynamically incorporate Ni in the MoS₂ slabs at reaction conditions. These beneficial effects are reduced, as nickel sulfide particles become excessively abundant as typical for high Ni-content catalysts, for which edge substitution by Ni is near or at its maximum. The areal activity and concentration of chemisorbed nitric oxide (NO) are well correlated for the leached catalysts, with the exception of the lowest Ni-containing catalyst that has a low degree of Ni edge substitution (<20% of total edge atoms) and predominantly unpromoted sites. This linear correlation shows that the Ni-promoted sites are more than five-fold as active as the unpromoted sites.

Full text of DOI:10.1016/j.jcat.2020.08.026

Journal of Catalysis 391 (2020) 212–223
Contents lists available at ScienceDirect
Journal of Catalysis
journal homepage: www.elsevier.com/locate/jcat
On the multifaceted roles of NiSx in hydrodearomatization reactions
catalyzed by unsupported Ni-promoted MoS2
b,
Ferdinand Vogelgsang b, Yinjie Ji b, Hui Shi a,b, , Johannes A. Lercher
⇑
⇑
a School of Chemistry and Chemical Engineering, Yangzhou University, Yangzhou, Jiangsu 225009, China
b Technical University of Munich, Department of Chemistry, Catalysis Research Center, 85747 Garching, Germany
a r t i c l e i n f o
a b s t r a c t
-1
Article history:
A series of unsupported Ni-Mo sulfide catalysts with varying Ni contents (0.13–0.72 molNi mol
) was
Ni+Mo
Received 14 April 2020
Revised 20 August 2020
Accepted 22 August 2020
Available online 1 September 2020
post-synthetically treated with concentrated HCl to remove large crystallites of accessible NiSx. These
sulfide particles inevitably form and grow at Ni concentrations required for the synthesis, but generally
have very low activities. In all cases, Ni concentrations were greatly reduced by the HCl treatment. While
-1
this ‘leaching’ strategy successfully improved the reaction rates of high Ni-content (>0.4 molNi mol
)
metal
catalysts for the hydrogenation of phenanthrene, modest to drastic decreases in catalytic rates (ꢀ0.1–0.8)
Keywords:
were registered for catalysts with lower Ni concentrations. For the lowest Ni-loaded catalyst (0.05 molNi
Hydrodearomatization
Polynuclear aromatic compounds
Phenanthrene
NO chemisorption
Nickel sulfides
-1
mol
), HCl treatment caused a dramatic loss of specific surface area and catalytic activity by more than
metal
a factor of 6 and shifted the selectivity pattern to that of pure MoS2. These observations allow us to con-
clude that Ni atoms incorporated at the slab edges are inherently susceptible to HCl attack. NiSx, however,
are the preferential sites at which HCl induces dissolution. Experiments with inter-particle mixtures and
segregated beds of NiSx and MoS2 demonstrate that NiSx not only activates H2, but also acts as a reservoir
to dynamically incorporate Ni in the MoS2 slabs at reaction conditions. These beneficial effects are
reduced, as nickel sulfide particles become excessively abundant as typical for high Ni-content catalysts,
for which edge substitution by Ni is near or at its maximum. The areal activity and concentration of che-
misorbed nitric oxide (NO) are well correlated for the leached catalysts, with the exception of the lowest
Ni-containing catalyst that has a low degree of Ni edge substitution (<20% of total edge atoms) and pre-
dominantly unpromoted sites. This linear correlation shows that the Ni-promoted sites are more than
five-fold as active as the unpromoted sites.
Unsupported sulfide catalysts
Ó 2020 Elsevier Inc. All rights reserved.
1. Introduction
to accommodate all Ni (or Co). Such segregated promoter phases
alone typically have low to moderate catalytic activities in
Binary and ternary transition metal sulfides (TMS) are promis-
ing catalysts for the hydrogenation of heavier crudes with high pol-
yaromatics and heteroatom contents [1–11]. These supported and
bulk mixed-phase sulfide catalysts contain a primary active struc-
ture (i.e., Ni(Co)-Mo(W)-S) with varying degrees of substitution at
the edge of the slabs, often together with segregated sulfide phases
of the promoter element (e.g., Ni or Co) [2,12–18]. The formation of
segregated promoter sulfides (NiSx or CoSx) is thermodynamically
favored and inevitable in TMS catalysts with high promoter con-
centrations [16,19,20], because the edge surface area is insufficient
hydrotreating reactions [20–23], and are generally considered to
be detrimental because they reduce the mass-specific and metal-
based reaction rates. Yet their ubiquitous presence requires a more
comprehensive understanding of their roles in hydrogenation and
hydrogenolysis reactions beyond their own limited catalytic
propensity.
Various models have been advanced to elucidate the chemical
structure and function of the promoter [13–15,23–32]. The major-
ity of the community nowadays describes the hydrotreating TMS
catalysts with the ‘‘Co-Mo-S” model and its analogues
[13–15,32]. However, there is sufficient evidence supporting alter-
native theories, such as the ‘‘remote control/contact synergy”
model, in which the catalytic synergy is attributed to an intimate
contact and cooperative action between the Mo(W)S2 and NiSx or
CoSx phases without necessarily forming a mixed ‘‘Ni(Co)-Mo
(W)-S” phase [25,26,31].
⇑
Corresponding authors at: School of Chemistry and Chemical Engineering,
Yangzhou University, Yangzhou, Jiangsu 225009, China (H. Shi); Technical Univer-
sity of Munich, Department of Chemistry, Catalysis Research Center, 85747
Garching, Germany (J.A. Lercher).
E-mail addresses: shihui@yzu.edu.cn (H. Shi), johannes.lercher@ch.tum.de
(J.A. Lercher).
https://doi.org/10.1016/j.jcat.2020.08.026
0021-9517/Ó 2020 Elsevier Inc. All rights reserved.

F. Vogelgsang et al. / Journal of Catalysis 391 (2020) 212–223
213
Eijsbouts consolidated the various models into a unified depic-
tion of structural evolution that reflects the dynamic nature of the
multi-component TMS catalyst in its working state [16]. At low Ni
(Co)-to-Mo(W) ratios, promoter atoms are atomically dispersed at
the edges of Mo(W)S2, while at very high ratios, particles of segre-
gated promoter sulfides are in intimate contact with and ‘‘deco-
rate” the Mo(W)S2-slabs, corresponding structurally to the
contact synergy model. While the presence of CoSx and NiSx crys-
tals in the close vicinity of the primary active structure partially
limits the accessibility of edge sites in the slab, this spatial proxim-
ity could also enable more efficient spillover of hydrogen (contact
synergy model), necessitated by the short diffusion path, from CoSx
or NiSx crystals to the adjacent promoted Mo(W)S2 phase. More-
over, the dispersion and segregation of the promoter likely occurs
in a dynamic fashion under reaction conditions [16]. It is therefore
conceivable that the manifestations of the promoter effect in a
given reaction, e.g., the variations of rates and selectivities to the
changes in the promoter concentration, should depend on the
quantity, size, location, and nature of the segregated phases, the
degree of Ni substitution in the Ni(Co)-Mo(W)-S phase, as well as
the interactions among these phases and with the support.
solution (Alfa Aesar, 40–44% w/w aq. solution) were added in
excess (S/Mo = 10). The mixed solution was stirred at 60 °C for
1 h. After that, the dark red solution was put on ice for 3 h to allow
complete precipitation of dark red ATM crystals, which were then
filtered off and washed with cold isopropanol. Next, the tetraalky-
lammonium thiomolybdate precursor was synthesized via a co-
precipitation method described by Alonso et al. [39] The ATM crys-
tals previously obtained were dissolved in deionized water and a
solution containing stoichiometric amounts (1:1 with respect to
Mo) of hexamethonium bromide, (CH3)3N(Br)(CH2)6N(Br)(CH3)3
(Alfa Aesar, 98%) was added. Once the two solutions came in con-
tact, an orange water-insoluble precipitate (hexamethonium
tetrathiomolybdate) formed immediately, which upon full sedi-
mentation was filtered off and washed with copious amounts of
water to remove Br-. The solid was dried in dynamic vacuum in a
desiccator and kept there before use. The choice of hexametho-
nium tetrathiomolybdate was originally intended for achieving
higher surface areas of the sulfides (upon decomposition) than in
a previous work by Alonso et al., where (R4N)2MoVIS4 (R = H,
CH3, n-C4H9) was used [40]. As shown later, this precursor eventu-
ally led to unsupported MoS2 with a specific surface area of
140 m2 gꢂ1, exceeding the conventional range of surface areas
(up to 100 m2 gꢂ1) reported for (NH4)MoS4-derived MoS2 while
being less than that (250 m2 gꢂ1) of (C4H9)4N)2MoS4-derived MoS2.
MoS2 and Ni-containing MoS2 catalysts were obtained by ther-
mal decomposition of the thiomolybdate in organic solvent in a
high pressure batch reactor (300 mL, Parr instruments) using a
method adapted from that described elsewhere [41]. First, 3 g of
the precursor, hexamethonium tetrathiomolybdate, were loaded
into the batch reactor and 100 mL of decalin were added. For Ni-
containing catalysts, additionally a solution of nickel naphthenate,
Ni(C11H7O2)2 (Alfa Aesar, 8.09 wt% Ni according to the certificate of
this specific batch of chemical) in decalin was added in a quantity
corresponding to the target ratio of Ni/(Ni + Mo) (varying from 0 to
0.7). The batch reactor was purged three times with 45 bar H2 and
pressurized to that pressure. Subsequently, it was heated to 350 °C
and maintained at that temperature and H2 pressure for 3 h. After
cooling down to room temperature overnight, the liquid was dec-
anted and the separated solid was put into a 50 mL centrifuge tube.
Then, it was washed with 40 mL of n-hexane, centrifuged and dec-
anted again. This procedure was repeated 5 times. Finally, the cat-
alyst was dried in dynamic vacuum. Before pelletizing and packing
into the reactor, the NiaMoS2 was re-sulfided with a 10 mL minꢂ1
flow of 10 vol% H2S in H2 (Westfalen, certified mixture of 98% pur-
ity H2S and 99.999% purity H2) at 400 °C for 8 h. Those catalysts are
denoted as NiaMoS2-P (‘P’ stands for ‘parent’, while ‘a’ represents
the molar fraction of Ni among total metals, i.e., Ni/(Ni + Mo)).
In addition, a series of catalysts was prepared by treating the
parent samples with aqueous concentrated HCl (~37 wt%, 12 M).
Specifically, 400 mg of the parent NiaMoS2 were put into a 5 mL
glass tube and 3 mL of concentrated HCl solution were added. After
1 h of vigorous reaction, the solution was centrifuged and the green
aqueous phase was removed. This procedure was repeated twice to
the remaining solid; no further sign of dissolution (coloring of the
aqueous phase, gas evolution) was observable after the third HCl
treatment. Then, the solid sample was washed with copious
amounts of deionized water to remove remaining chloride ions,
dried in dynamic vacuum overnight and stored in a desiccator.
Before pelletizing, the samples were re-sulfided with a 10 mL minꢂ1
flow of 10 vol% H2S in H2 at 400 °C for 8 h. These ‘leached’ samples
are denoted with an ‘L’ as the suffix.
Inspired by the work of Eijsbouts et al. [19,20] on the roles of
nickel sulfides (NiSx) and interested in exploiting the higher hydro-
genation activity of Ni-promoted MoS2/WS2 compared to the Co-
promoted counterparts, we recently undertook a major effort to
reduce the concentrations of NiSx in high-Ni-content mixed Ni-
Mo and Ni–W sulfides supported on
c-Al2O3 and in self-
supported forms [33–35]. This was enabled by a ‘leaching’ proce-
dure, which relied on a selective reaction between concentrated
HCl and accessible NiSx phases, but not Mo(W)S2. While these
studies proved the efficacy of this strategy in enhancing catalytic
rates of highly Ni-loaded catalysts, the detailed impact of the HCl
treatment on the catalysis with unsupported sulfide catalysts of
lower nickel contents was not explored.
Thus, we address here the consequences of removing NiSx by
HCl treatment for the physicochemical properties and catalytic
rates of unsupported Ni-Mo sulfide catalysts containing low to
high Ni contents (Ni/(Ni + Mo) = 0.13–0.72 mol/mol). We started
with two hypotheses, i.e., (i) NiSx and Ni incorporated at the slab
edge are in dynamic equilibrium and (ii) removing NiSx will
improve the structure-activity correlation using probe molecules
that adsorb on both the primary phase and NiSx. To address the
first hypothesis, hydrogenation of phenanthrene was chosen as
the model hydrodearomatization reaction. The regioselectivity of
this reaction exhibits
a marked sensitivity to Ni promotion
[36,37], allowing us to gain insights into the changes induced by
HCl to the edge substitution degree in the primary slabs. To test
the second hypothesis, we employed chemisorption of nitric oxide
(NO) as a means to quantify the concentration of coordinatively
unsaturated sites (CUS) that mediate H2 activation and hydrogen
addition.
2. Experimental
2.1. Synthesis of catalysts
The parent NiaMoS2 (‘a’ represents the nominal molar fraction
of Ni among total metals, i.e., Ni/(Ni + Mo) in the synthesis mix-
ture) catalysts were synthesized via a thiosalt method to allow a
good control over the Ni content. First, ammonium tetrathiomolyb-
date ((NH4)2MoS4, abbreviated as ATM hereafter) was synthesized
following a method reported by de Brimont et al. [38] Specifically,
A nickel sulfide sample (denoted as NimSn hereafter) was syn-
thesized by a method described by Bezverkhyy et al., according
to which mixtures of Ni3S2 and NiS would be synthesized [42].
Nickel nitrate hexahydrate, Ni(NO3)2ꢁ6H2O (Acros organics, 99%),
was dissolved in water and Na2Sꢁ9H2O (VWR, 98%) was added.
8 g of ammonium heptamolybdate tetrahydrate, (NH4)6Mo7O24
-
ꢁ4H2O (Merck, 99%), were dissolved in 20 g of ammonium hydrox-
ide solution (Merck, 32 wt% in water), and 72 g of a (NH4)2S

214
F. Vogelgsang et al. / Journal of Catalysis 391 (2020) 212–223
The solution was stirred at room temperature for 1 h and the pre-
cipitate was filtered off. The precipitate was washed several times
with water and once with acetone before leaving it over night to
dry. Finally, the dried sample was re-sulfided in a 10 mL minꢂ1
flow of 10 vol% H2S/H2 at 400 °C for 8 h prior to catalytic use;
the catalytic activity was hardly measurable. This sample could
be completely dissolved by the same HCl treatment.
were diluted with SiC (sieved to mesh sizes of 63–90 mm) in a
1:20 w/w ratio. The filling pattern is shown in Fig. 1 (left panel).
After conditioning the catalyst at 360 °C for 21 h, the
hydrodearomatization reaction was carried out at 300 °C and
60 bar total pressure and steady-state catalytic rates were mea-
-1
sured by varying the space time from 10 to 30 (gcatꢁh) mol
.
Phe,0
Deactivation was < 5% of the initial activity over time on stream.
Liquid samples were taken in a time interval of 30 min and ana-
lyzed with an offline GC (Shimadzu 2010, flame ionization detec-
tor, 50 m HP-1 column with 32 mm inner diameter).
2.2. Characterization
The contents of H, C, N and S were quantified by oxidative com-
bustion using an automated element analyzer instrument (Vario EL
CHN Analyzer, ELEMENTAR). The gases CO2, H2O, N2 and SO2 were
detected by gas chromatography. The concentrations of Ni and Mo
were photometrically measured after digestion of the solids.
Physisorption of N2 at ꢂ196 °C was performed on an automated
nitrogen adsorption analyzer Sorptomatic 1990 Series (Thermo
Finnigan), where BET surface areas and total pore volumes (Gur-
vich method) of ~250 mg of sample were determined. The samples
were outgassed in vacuum at 250 °C for 2 h prior to adsorption.
Powder X-ray diffraction (XRD) patterns were recorded on an
Empyrean system from PANalytical equipped with a Cu X-ray tube
Conversion rates of phenanthrene were obtained from the
conversion-space time plots. H2-addition rates were determined
using the following equation using initial rates and selectivities
(extrapolated to zero contact time):
H2-addition rate = Phenanthrene hydrogenation rate ꢀ [Sel.DiHPhe
(%) + 2 ꢀ Sel.TetHPhe (%) + (4 ꢀ (Sel.sym-OHPhe + Sel.asym-OHPhe) (%)]
2.4. Experiments with physical mixtures and stacked beds of NimSn and
MoS2
The as-obtained NimSn sample was mixed with MoS2 (synthe-
sized from thermal decomposition of hexamethonium
tetrathiomolybdate, Section 2.1), both in powder form and the
total mass being ~ 120 mg, in a 5 mL glass tube where n-hexane
was added. The slurry was ultrasonicated to mix the powders until
the n-hexane evaporated. Typically, 15 mg of the pelletized and
sieved (250–355 mm) physical mixture was loaded into the trickle
bed reactor described above. The filling pattern in such experi-
ments is illustrated in Fig. 1 (right panel). Variations of catalyst
mix were positioned within the isothermal zone of the reactor fur-
nace: diluted MoS2, undiluted MoS2, NimSn/MoS2 mixtures with
mass ratios of 1:1, 3:1, 6:1 and 9:1, as well as stacked NimSn and
MoS2 beds in a mass ratio of 3:1 separated by an intervening bed
(Cu-K radiation, 0.154 nm), a nickel Kb-filter, and solid-state
a
detector (PIXcel1D) operated at 45 kV and 40 mA with step size
of 0.017° and scan time of 115 s per step. The peaks were identified
using the Highscore software.
Transmission electron microscopy (TEM) images were taken on
a JEOL JEM 2010 instrument equipped with a LaB6-cathode with an
accelerating voltage of 120 kV. The samples were prepared by
grinding a small amount of specimen and dispersing it ultrasoni-
cally in absolute ethanol. Subsequently, drops of this suspension
were applied on a Cu grid (200 mesh) with a lacey carbon film
and ethanol was evaporated at room temperature. For each sam-
ple, >100 of slabs were measured to allow a statistical analysis of
the average slab length and stacking degree.
of 4 mm of c-Al2O3.
Nitric oxide (NO) chemisorption was carried out at ambient
temperature and pressure on a continuous flow reactor connected
to a mass spectrometer (Pfeiffer Vacuum QME 200). The catalyst
(80 mg) was packed into a quartz tube (4 mm inner diameter,
isothermal zone 5 cm) and re-sulfided in a flow of H2S (10 vol%
in H2, 100 mL minꢂ1 gꢂ1) by heating to 400 °C (10 °C minꢂ1) and
holding at this temperature for 1 h. Subsequently, the reactor
was cooled to 250 °C at which the H2S/H2 flow was stopped, and
then was cooled to 35 °C in a constant flow of 8 mL minꢂ1 N2
(i.e., 100 mL minꢂ1 gꢂ1). For the pulsing experiments, a loop with
a defined volume (1.69 mL) was filled with 10 vol% NO in helium,
which was pulsed (6.9 mmol of NO per pulse) through an auto-
mated valve into the reactor. This was repeated 24 times in
30 min intervals and the breakthrough of NO was detected with
a mass spectrometer (monitoring the signal at m/z = 30). Total
NO uptakes were calculated by subtracting each residual peak area
(i.e., NO not adsorbed) from the averaged peak area after saturation
with NO and adding up the differences.
3. Results and discussion
3.1. Elemental and phase compositions of NiaMoS2 catalysts before and
after leaching
The contents of Ni, Mo and S in the parent NiaMoS2 (where ‘‘a”
refers to the molar fraction of Ni among all metals) and those after
HCl treatment are compiled in Table 1. The total contents of C, H
and N from organic residues were < 5 wt% for most cases
(Table S1, SI). The Ni content in NiaMoS2 was close to the nominal
value in the synthesis mixture, and the residual carbon in the par-
ent sample was lower (C/Mo = 0.89 mol molꢂ1 for MoS2 and 0.04–
0.31 mol molꢂ1 for Ni-containing MoS2 samples) with the synthe-
sis method employed here than with (R4N)2MoS4 precursors (C/
Mo = 1.2–2.0 in MoS2) [40]. The molar ratios of S to metals (Ni
and Mo) varied between 1.3 and 2.0 for the parent series. The S/
metal ratio decreased monotonically with increasing Ni content.
This is attributed to a much lower average stoichiometric ratio of
S to Ni (<2 or even < 1) in various NiSx phases (e.g., Ni3S2, Ni9S8),
present in increasing concentrations as the overall Ni content
increased.
2.3. Catalytic evaluation
Phenanthrene hydrogenation was carried out in a down-flow
trickle bed reactor system equipped with an HPLC pump (Gilson
302) and mass flow controllers (Wagner Mess- und Regeltechnik
GmbH), to control the volumetric flows of the liquid feed, H2
(Westfalen, 99.999%) and N2 (Westfalen, 99.999%). The liquid feed
contained 1 wt% phenanthrene (Alfa Aesar, 98%), 2 wt% tetrade-
cane (Alfa Aesar, 99%) as internal standard and 1000 ppmw sulfur
as dimethyl disulfide (DMDS, Aldrich, 98%) in decalin (Merck, cis-
and trans-isomeric mixture, 99.0%) as the solvent. In a typical
experiment, 15 mg of catalyst (250–355 mm pellets, resulfided)
By adding concentrated HCl to the suspension containing solid
particles, the supernatant turned green, indicating the presence
of Ni2+ in the solution. In parallel, H2S evolved. The loss of Ni is evi-
dent when comparing the parent and leached samples (Table 1).
The Ni concentration decreased in the order: 62.5% (Ni0.1
-
MoS2) > 50.0% (Ni0.2MoS2) > 44.4% (Ni0.6MoS2) > 37.5% (Ni0.7MoS2)
ꢃ 37.0% (Ni0.4MoS2) ꢃ 34.3% (Ni0.5MoS2). In contrast, the Mo con-
centration generally increased with the original Ni content: 13%

F. Vogelgsang et al. / Journal of Catalysis 391 (2020) 212–223
215
Fig. 1. Illustration of the filling patterns in conventional operation (left) and experiments with physical mixtures and stacked beds (right).
Table 1
Elemental compositions of the parent and HCl-treated NiaMoS2.
Ni[a] [mmol g
]
Mo[a] [mmol g
]
S[a] [mmol g
]
cat
S/metal [mol/mol]
Atomic Ni fraction [molNi mol
-1
-1
-1
-1
]
metal
Catalysts
cat
cat
Ni0.1MoS2
Ni0.2MoS2
Ni0.4MoS2
Ni0.5MoS2
Ni0.6MoS2
Ni0.7MoS2
MoS2
0.8 (0.3)
1.6 (0.8)
2.7 (1.7)
3.5 (2.3)
5.4 (3.0)
6.4 (4.0)
0
5.3 (6.0)
5.1 (5.5)
4.2 (5.4)
3.9 (4.8)
3.1 (4.3)
2.5 (3.8)
5.3
12.3 (11.8)
12.1 (12.9)
11.8 (11.2)
11.6 (12.5)
11.5 (12.5)
11.9 (12.1)
12.4
2.0 (1.9)
1.8 (2.0)
1.7 (1.6)
1.6 (1.8)
1.4 (1.7)
1.3 (1.6)
2.3
0.13 (0.05)
0.24 (0.13)
0.39 (0.24)
0.48 (0.33)
0.63 (0.41)
0.72 (0.52)
0
NimSn
11.3
0
10.1
0.89
1
[a]
Values in the parentheses correspond to the leached samples.
(Ni0.1MoS2), 7% (Ni0.2MoS2), 28% (Ni0.4MoS2), 23% (Ni0.5MoS2), 38%
(Ni0.6MoS2) and 52% (Ni0.7MoS2). Without applying an electric
potential or strong oxidative driving force, pure MoS2 does not
react with concentrated HCl (even at 12 M) at these conditions,
according to the thermodynamic data (for bulk phase) [43]. Thus,
we hypothesized that the (Ni)MoS2 slab structure stayed largely
intact during the treatment, while crystalline NiSx reacted with
HCl to produce H2S and dissolved. Independent analysis of the fil-
trate solution after leaching of parent MoS2 revealed a relatively
minor loss of Mo (~0.5% of total Mo), suggesting a small extent of
destruction (~2–3% of edge Mo) at the slab edge. However, this
level of Mo leaching did not seem to be substantial. As a result,
the observed increase in Mo content after leaching, for the Ni-
containing samples, must be due to the loss of mass in the form
of NiSx. Consequently, also the overall ratio of S/metal should
increase. This was observed experimentally for the majority of
samples except Ni0.1MoS2 and Ni0.4MoS2. Aside from possible mea-
surement errors (e.g., 10% for Mo) in elemental analysis, the
minor decrease in this ratio for Ni0.1MoS2 and Ni0.4MoS2 suggests
that some phases with a S/Me ratio of > 2 have also been involved
in the dissolution. It is, however, unclear as to why these two sam-
ples differ from the others. It should be mentioned at this point
that Ni incorporated in the slabs might also be attacked by HCl, a
possibility that will be examined in detail later. While the elemen-
tal analysis results reveal the variations in the bulk fractions of seg-
regated NiSx and Ni-promoted slabs, we note that a careful XPS
study combined with a well-advised peak decomposition approach
could provide a means to resolve and quantify the different Ni spe-
cies on the surface in the parent and leached samples, which would
be worthwhile to explore in future studies.
Fig. 2 (a) and (b) depict the powder X-ray diffraction patterns of
the parent and leached NiaMoS2 catalysts, respectively. Every for-
mulation shows the underlying diffraction pattern of MoS2, with
peaks at 2h = 14°, 32°, 40° and 59° corresponding to the reflections
of (002), (010), (013) and (110) planes. Conventionally, the (002)
plane is used to calculate the stacking degree of the MoS2 slabs by
determining the d-spacing of the slabs and their crystal size by the
Scherrer equation. Note, however, that the curvature and imperfect
stacking (i.e., turbostraticity and disorder) of sulfide slabs also
causes the loss of scattering coherency, thus, resulting in asymmet-

216
F. Vogelgsang et al. / Journal of Catalysis 391 (2020) 212–223
Fig. 2. (a) XRD patterns of the parent NiaMoS2 formulations with increasing Ni content ordered from bottom to top. The stick patterns of four reference substances (obtained
from Highscore database) are shown at the bottom: blue, godlevskite (Ni9S8); green, millerite (NiS); orange, heazlewoodite (Ni3S2); red, polydymite (Ni3S4). (b) XRD patterns
of the ‘leached’ formulations. The reflections of MoS2 are denoted along with indices of lattice planes.
ric broadening and precluding the observation of sharp reflections.
The sharp reflections in the patterns of high-Ni content catalysts
at inaccessible locations (e.g., entrapped or ingrown). Note that
the leached samples had been sulfided before measurements and
we cannot exclude the formation of new crystallites of NiSx during
that re-sulfidation process.
are attributed to NiSx phases (primarily Ni9S8 and Ni3S2), which
-1
were hardly detectable at Ni contents up to 0.4 molNi mol
.
metal
While the absence of sharp reflections in the XRD patterns of other
low-Ni-content NiaMoS2 (a = 0.1–0.4) samples does not necessarily
indicate the absence of segregated nickel sulfide phases [19], they
are most likely small in size and become undetectable by XRD.
The patterns of the ‘leached’ NiaMoS2 catalysts evidence a suc-
cessful removal of a majority of large nickel sulfide crystallites
originally present in the parent samples. Some sharp reflections
3.2. Textural and slab properties of NiaMoS2 catalysts before and after
leaching
BET surface areas and total pore volumes (Gurvich method) are
compiled in Table 2; the corresponding N2 sorption isotherms and
pore size distribution histograms are presented in Figs. S1 and S2.
Pure MoS2 displayed a specific surface area of 140 m2 gꢂ1 and a
pore volume of 0.538 cm3 gꢂ1, comparable to those of the parent
Ni0.1MoS2. Increasing the Ni content led to a decrease in the speci-
of NiSx remained in the XRD patterns of formulations with Ni con-
-1
tents higher than 0.5 molNi mol
, albeit showing a smaller num-
metal
ber of peaks at lower intensities. Their persistence could reflect
that the severity and duration of the present leaching protocol is
insufficient to remove all leachable NiSx or that some is present
fic surface area and pore volume of the parent samples. Up to 0.5
-1
molNi mol
, the surface area and pore volume remained no
metal

F. Vogelgsang et al. / Journal of Catalysis 391 (2020) 212–223
217
Table 2
Physicochemical characteristics of NiaMoS2, MoS2 and NimSn.
-1
-1
[a]
Catalysts
BET surface area [m2
g
]
Pore volume [cm3
g
]
cat
Average number of slabs in stacks (TEM) [–]
Slab length (TEM) [nm]
fMo [–]
cat
Ni0.1MoS2
Ni0.2MoS2
Ni0.4MoS2
Ni0.5MoS2
Ni0.6MoS2
Ni0.7MoS2
MoS2
140 (22)
129 (106)
91 (45)
90 (80)
26 (44)
34 (61)
140 (160)
11
0.394 (0.061)
0.384 (0.278)
0.256 (0.095)
0.224 (0.166)
0.049 (0.077)
0.081 (0.153)
0.538 (0.400)
0.0455
4.1 0.9 (3.8 1.1)
3.5 0.9 (3.7 1.5)
3.7 1.2 (3.5 0.9)
3.9 1.3 (3.8 1.1)
3.2 0.9 (3.8 0.8)
3.8 1.0 (3.5 1.2)
4.1 1.1 (4.3 1.0)
5.3 1.4 (4.0 1.1)
4.9 1.6 (4.8 1.7)
6.6 3.0 (4.9 1.6)
5.3 1.7 (4.8 1.4)
4.4 1.2 (4.3 1.4)
5.2 2.4 (6.2 2.7)
5.6 1.3 (4.7 1.5)
0.21 (0.27)
0.22 (0.23)
0.15 (0.22)
0.21 (0.23)
0.25 (0.25)
0.19 (0.17)
0.20 (0.23)
[b]
[b]
[b]
NimSn
–
–
–
P
m
i¼1 6ni ꢂ6
[a]
[b]
P
Calculated via fMo
Not applicable.
¼
with the number of Mo atoms along the edge ni ¼ L=6:4Å þ 0:5 [44,45].
3n2i ꢂ3ni þ1
m
i¼1
lower than 90 m2 gꢂ1 and 0.2 cm3 gꢂ1, respectively. For Ni0.6MoS2
and Ni0.7MoS2 in their parent form, the surface areas were rather
low at 26 and 34 m2 gꢂ1 and their pore volumes were also much
lower than catalysts containing less Ni. Because NiSx domains in
parent samples with lower Ni contents are less abundant and sig-
nificantly smaller in size, these smaller entities do not block the
pores to the extent that large crystallites do and, hence, the varia-
tions in pore volumes and specific surface area is less drastic than
for materials with larger NiSx domains.
The effect of HCl treatment on the specific surface area and pore
volume of the NiaMoS2 varied with the Ni content. For the two
highest-Ni-loaded samples, which, according to XRD, contained a
larger number of nickel sulfide crystallites, the specific surface
areas and pore volumes increased substantially after leaching. This
is in agreement with our most recent work also showing a signifi-
cant increase in pore volume and specific surface area upon HCl
agglomeration of the Ni-promoted MoS2 primary particles and
the intimacy of NiSx and NiMoS2 in the parent samples.
These parent and leached samples were investigated by TEM to
see if the leaching process had caused any major modifications of
the slab characteristics. Representative images are shown in
Fig. S3. For the parent series of catalysts, the average slab length
varied between 4 and 5 nm, while the average stacking degree
was between 3 and 4 slabs per particle, both being marginally
affected by the Ni content. The primary slab structure was not sig-
nificantly affected by HCl treatment. The fraction of edge atoms
among all atoms in a slab was estimated from the average slab
length by assuming an equilateral hexagon shape for the slab
(see the last column of Table 2). According to these estimates,
the total number of Ni atoms before leaching should exceed the
total number of edge atoms (decreasing from 1.0 to 0.5 mmoledge
gꢂ1 with Ni concentration increasing from 0.8 to 6.4 mmolNi
atom
treatment [34] for a series of high-Ni-content (0.67–0.77 molNi
gꢂ1) on all parent samples except Ni0.1MoS2, which means that NiSx
must be present in those samples. For the leached samples, the
total amount of Ni in both Ni0.1MoS2 and Ni0.2MoS2 would not be
sufficient to attain a full substitution at the edge. In particular,
the calculated Ni edge substitution degree would be less than
20% of the total edge atoms for the leached Ni0.1MoS2 sample, thus,
leaving a high statistical probability (>60%) of exposing unpro-
moted sites (i.e., neighboring Mo-Mo) on this sample.
-1
mol
) Mo(W) sulfide materials prepared by co-precipitation.
metal
Such increase is attributed to the removal of a large quantity of
NiSx with low surface area and pore volume (see the last entry of
Table 2). The deleterious effect of this leaching on the textural
properties at lower Ni contents is, however, not understood at this
stage and may point to a complex action of the HCl treatment that
goes beyond the simple removal of NiSx. While we speculate that
agglomeration of the (Ni)MoS2 phase occurred to a significant
extent for the low-Ni-content materials during the HCl treatment,
TEM and SEM pictures did not provide strong support, nor did they
evidence any morphological changes (e.g., NiSx obstructing the
edges, or amorphization of the slabs) that could be held responsi-
ble for the loss of specific surface area in these samples.
There appears to be a ‘zigzag’ trend in the variation of specific
surface area caused by the HCl treatment: the most drastic loss
was exhibited by Ni0.1MoS2 (a factor of 6 decrease), followed by
Ni0.4MoS2 that lost more than a half of its original pore volume
and surface area, whereas such decreases were much less pro-
nounced for Ni0.2MoS2 and Ni0.5MoS2 (Table 2). The varying extents
of losses and gains in BET surface area for the individual samples
after HCl treatment reflect a complex interplay of multiple factors
(e.g., removal of NiSx with different sizes and surface areas in the
parent sample, the different amounts of retained organics in the
parent samples and the removal of organic residues during HCl
treatment). It appears plausible that organic residues in some of
the parent samples (Ni0.1MoS2, Ni0.4MoS2 and Ni0.5MoS2) helped
in resisting agglomeration of particles and thus their removal dur-
ing HCl treatment (Table S1) caused the particles to agglomerate.
Although the exact reason is not quantitatively understood at
present, the ‘zigzag’ trend across the series could be a result of
removing organics and NiSx during the HCl treatment, which have
opposing consequences for the surface area and textural
properties, and thus the variations in surface area might depend
on the quantities of these species removed and perhaps also the
3.3. Phenanthrene hydrogenation on NiaMoS2 catalysts
3.3.1. Reaction pathways and initial selectivities
Ni-promoted MoS2 catalysts with different Ni contents, before
and after HCl treatment, were examined in the hydrogenation of
phenanthrene. In all reactions, four products were identified,
namely
9,10-dihydrophenanthrene
(DiHPhe),
1,2,3,4-
tetrahydrophenanthrene (TetHPhe), 1,2,3,4,5,6,7,8-octahydrophe
nanthrene (sym-OHPhe) and 1,2,3,4,4a,9,10,10a-octahydrophenan
threne (asym-OHPhe). Perhydrophenanthrene, the fully hydro-
genated product, was not observed at the reaction temperature
of 300 °C and no lighter products were observed so the carbon bal-
ance was always closed within 5%.
A delplot analysis of first and second rank [46] allowed us to
identify the nature of the detected products. Two primary products
were observed on all catalysts, DiHPhe (two H-added) and TetHPhe
(four H-added), while sym- and asym-OHPhe (eight-H added) were
secondary kinetic products. Therefore, a reaction network includ-
ing two routes for hydrogenation is proposed (Scheme 1). Along
path 1, phenanthrene first undergoes the hydrogenation across
the olefin-like 9,10-bond to DiHPhe followed by further hydro-
genation of a lateral ring to asym-OHPhe, while along path 2,
hydrogenation of one of the lateral rings forms TetHPhe first, fol-
lowed by the hydrogenation of the other lateral ring to form
sym-OHPhe. The DiHPhe formation is reversible (Keq = 1.6 for this

218
F. Vogelgsang et al. / Journal of Catalysis 391 (2020) 212–223
the Ni promotion of the MoS2 slab [36]; specifically, the selectivity
shifts in favor of DiHPhe with decreasing degree of Ni incorpora-
tion at the slab edge. Selectivities were quite similar among Ni-
promoted samples, irrespective of the Ni content.
For most catalysts, HCl treatment hardly changed the pathway
ratio or slightly shifted it towards path 2, whereas the ‘leached’
Ni0.1MoS2 catalyst showed a marked shift towards path 1, moving
closer to that of MoS2 (Fig. 3). Given the relation between Ni pro-
motion and product distribution [36], the observation for Ni0.1
-
MoS2 suggests that most of the promoted active sites have been
eliminated after ‘leaching’ of this catalyst. This, in turn, strongly
suggests that HCl is able to attack Ni2+ in the slab structure, in
addition to segregated NiSx. The same treatment, however, did
not seem to cause the loss of Ni-promoted edge sites in other Nia-
MoS2 catalysts. Thus, we hypothesize that HCl preferentially
attacks NiSx, unless the amount of such entities is small as in the
case of Ni0.1MoS2. In this connection, we would like to point to
the statistically less than 20% Ni edge substitution for Ni0.1MoS2,
being much lower than for the other samples.
Scheme 1. Simplified reaction pathways of phenanthrene hydrogenation.
step at 60 bar of H2 and 300 °C; calculated from thermodynamic
data in ref [47]). Other hydrogenation reactions remain essentially
irreversible at typical conditions (e.g., Keq = 50 and larger for
[TetHPhe]/[Phe] and [OHPhe]/[Phe] at 60 bar H2 and 300 °C).
The initial selectivities of the products, obtained by extrapola-
tion to zero contact time in a selectivity-conversion plot, are shown
in Fig. 3. Detailed values are compiled in Table S2, which also
includes the ratio of selectivities along the two paths (i.e.,
Sel.DiHPhe + Sel.asym-OHPhe for path 1 and Sel.TetHPhe + Sel.sym-OHPhe
for path 2). For all parent and HCl-treated formulations, the two
primary products DiHPhe and TetHPhe accounted for a selectivity
of more than 90% combined. Of the ‘deeper’ hydrogenated
products, asym-OHPhe was hardly observed, while sym-OHPhe
3.3.2. Effects of Ni content and HCl treatment on the hydrogenation
rates
Table 3 lists the conversion rates of phenanthrene and the
hydrogen-addition rates for the parent and HCl-treated NiaMoS2
catalysts. Hydrogen addition rates plotted against the molar frac-
tion of Ni among total metal are illustrated in Fig. 4 for both sets
of catalysts, while other plots are shown in Figs. S4-S7. The rates
follow a volcano-shaped curve for both parent and HCl-treated ser-
was found for catalysts with Ni contents higher than
ies of catalysts, with the maximum rate appearing at 0.32 molNi
-1
-1
0.15 molNi mol
. Here, for non-promoted bulk MoS2 the initial
metal
mol
(Ni0.5MoS2-L). The maximum rate was rather close to that
metal
selectivities were 86% for DiHPhe and 14% for TetHPhe, respec-
tively, while no deep hydrogenation was observed. On the parent
NiaMoS2, the selectivities to DiHPhe and TetHPhe became compa-
rable, a greater extent of deep hydrogenation (OHPhe formation)
was observed and the overall reaction shifted in favor of path 2,
in contrast to the unpromoted MoS2. As previously observed for
measured on Ni0.2MoS2-P. In terms of conversion rates normalized
to total Mo content, the bulk NiaMoS2 catalysts peaked at 1.9
molPhe mol hꢂ1
,
similar to the maximum rate (1.6 molPhe
-1
Mo
mol hꢂ1, also appearing at 0.33 molNi mol
) measured for
metal
-1
-1
Mo
Al2O3-supported NiMoS2 catalysts (with similar average slab
lengths) at the same conditions [36]. Similarly, other normalized
rates at the maximum were also comparable on bulk and
Al2O3-supported NiMoS2 catalysts.
Ni-promoted MoS2 supported on
c-Al2O3, the ratio of initial selec-
tivities in phenanthrene hydrogenation is a strong indication for
Fig. 3. Initial selectivities obtained from extrapolation of product selectivities (measured at phenanthrene conversion of 10–20%) to zero contact time at 300 °C and 60 bar
total pressure. The arrow indicates a significant shift in the pathway ratio on the HCl-treated Ni0.1MoS2 toward that of a pure MoS2 catalyst. Blue bar: DiHPhe; green bar:
TetHPhe; yellow bar: asym-OPhe; orange bar: sym-OPhe.

F. Vogelgsang et al. / Journal of Catalysis 391 (2020) 212–223
219
Table 3
Hydrogenation rates of phenanthrene on the parent and HCl-treated formulations. The rates are obtained at 300 °C and 60 bar H2 pressure by varying the space time between 7
and 40 (hꢁgcat)/molPhe,0. The feed contained 1 wt% Phe, 1000 ppmw S as dimethyl disulfide (DMDS) and 2 wt% tetradecane as the internal standard in decalin. The ratio of H2/
hydrocarbons was 300 (v/v).
Catalysts
Conversion rates
H2 addition ratesa
hꢂ1
]
[molPhe mol
hꢂ1
-1
metal
]
[molPhe mol hꢂ1
]
[molH2 mol hꢂ1
]
[molH2 mol
hꢂ1
]
-1
metal
-1
-1
-1
[mmolPhe
g
cat
Mo
Mo
Ni0.1MoS2-P
Ni0.1MoS2-L
Ni0.2MoS2-P
Ni0.2MoS2-L
Ni0.4MoS2-P
Ni0.4MoS2-L
Ni0.5MoS2-P
Ni0.5MoS2-L
Ni0.6MoS2-P
Ni0.6MoS2-L
Ni0.7MoS2-P
Ni0.7MoS2-L
4.6
0.7
8.6
4.6
7.5
5.8
6.0
9.1
3.6
6.4
2.7
5.2
0.7
0.1
1.3
0.7
1.1
0.8
0.8
1.3
0.4
0.9
0.3
0.7
0.9
0.1
1.7
0.8
1.8
1.1
1.6
1.9
1.1
1.5
1.1
1.4
1.4
0.1
3.0
1.5
3.0
2.0
2.8
3.5
2.0
2.7
1.8
2.4
1.2
0.1
2.3
1.3
1.8
1.5
1.5
2.3
0.8
1.6
0.5
1.2
a
Calculated from the phenanthrene consumption and the moles of hydrogen consumed along each route.
undesired inert mass, leading to an eventual drop of reaction rates
on a mass or total metal basis. A fraction of (promoted) sites of the
(Ni)MoS2 slabs may be blocked by NiSx, if they are located nearby,
leading to a decrease in reaction rates even on a Mo basis.
Our prior study showed that HCl treatment of Al2O3-supported
NiWS2 catalysts selectively removed NiSx and increased the mass-
specific catalytic activity at high Ni loadings, while at low Ni load-
ings, HCl treatment did not change the overall activity [33]. In the
present work on unsupported NiMo sulfide catalysts, the effect of
HCl treatment appears to be more intricate, depending on the nat-
ure of the parent samples, especially the original Ni concentration.
For parent catalysts with Ni concentrations higher than 0.4 molNi
-1
mol
, H2-addition rates (total metal based) increased on their
metal
HCl-treated counterparts, by 53% for Ni0.5MoS2, 100% for Ni0.6MoS2
and 140% for Ni0.7MoS2. For catalysts with original Ni contents
-1
lower than 0.4 molNi mol
, the HCl treatment had a negative
metal
impact on the reaction rate, which dropped by 17% on Ni0.4MoS2,
43% on Ni0.2MoS2 and 83% on Ni0.1MoS2 (Table 4).
As discussed above, increases in reaction rates are partially due
to the loss of inactive phases in the TMS catalysts. While a direct
measurement of the mass loss during leaching was subject to large
errors because of the difficulty in quantitative recovery of solids,
the extent of mass loss can be calculated based on the elemental
composition, by assuming the mass conservation of Mo that is
minimally leachable in concentrated HCl. Since the removed mass
is almost only composed of NiSx, this factor (Table 4) also reflects a
theoretical correction for the rate of a catalyst after removing (a
portion of) these inactive phases.
Fig. 4. Hydrogen addition rates (calculated from the phenanthrene consumption
and the moles of hydrogen consumed along each route) over the parent and HCl-
treated NiaMoS2 as a function of Ni content. Full squares represent the parent
samples, while open squares correspond to the HCl-treated formulations.
Let us now take a closer look at the individual series. Within the
parent samples, the total-metal-based H2-addition rate increased
-1
-1
from 1.2 molH2 mol
hꢂ1 (Ni0.1MoS2) to 2.3 molH2 mol
hꢂ1
metal
metal
(Ni0.2MoS2). After passing the maximum, the rate decreased with
Following this rationale, the reaction rate is expected to
-1
increasing Ni content in the order of Ni0.4MoS2 (1.8 molH2 mol
-
metal
increase for all catalysts after HCl treatment. This is experimentally
hꢂ1)>Ni0.5MoS2 (1.5molH2 molm-1etal hꢂ1)>Ni0.6MoS2 (0.8
-1
observed for Ni contents higher than 0.4 molNi mol
, yet the
metal
molH2 mol
hꢂ1) > Ni0.7MoS2 (0.5 molH2 mol
hꢂ1). This trend
metal
-1
-1
metal
expected extents of rate enhancement merely due to inert mass
removal are always smaller than the observed changes (Table 4).
Therefore, we conclude that high Ni-loaded catalysts benefit from
is explained by the presence of both active and inactive Ni-
associated phases, the relative population of which changes with
the Ni concentration. Increasing the Ni content initially increases
the edge substitution by Ni and the number of promoted sites at
the edge, leading to a better hydrogenation activity. Here, the frac-
tion of metal atoms located at the slab edges is ~ 0.2 on average
Table 4
Comparison of the observed rate enhancement due to ‘leaching’ and the expected
enhancement due to the loss of inactive mass.
(Table 2). Therefore, a full Ni substitution of the edges was reached
-1
Catalysts
Observed rate
enhancement for H2
addition (rleached/rparent
Expected factor of H2-addition rate
increase due to the removal of inert
mass (nMo,leached/nMo,parent)
on Ni0.2MoS2 (0.24 molNi mol
), which was the most active for-
metal
mulation. Upon further increasing the amount of Ni in the cata-
lysts, more and larger NiSx crystallites form and segregate from
the original slabs. Al2O3-supported nickel sulfide phase was previ-
ously shown to be inactive for the hydrogenation of phenanthrene
[36] and bulk NiSx (at least the large crystallites) were found in this
study to be inactive as well. Thus, these segregated Ni species con-
tribute negligibly to the hydrogenation activity and represent
)
Ni0.1MoS2 0.17
Ni0.2MoS2 0.57
Ni0.4MoS2 0.83
Ni0.5MoS2 1.53
Ni0.6MoS2 2.00
Ni0.7MoS2 2.40
1.13
1.08
1.29
1.23
1.39
1.52

220
F. Vogelgsang et al. / Journal of Catalysis 391 (2020) 212–223
leaching, by way of losing inactive large NiSx particles and
unblocking of active sites. In contrast, the effect of HCl treatment
was more destructive for those with lower Ni contents, likely
caused by the removal of small NiSx clusters located at the edges
of the (Ni)MoS2 slabs or even the removal of promoting Ni atoms.
Thus, we hypothesize that NiSx has positive as well as negative
roles in the hydrogenation of phenanthrene.
3.5. Structure-activity correlations for promoted and unpromoted sites
in bulk NiaMoS2
Transmission infrared (IR) spectroscopy was previously applied
to determine the surface density of (potentially) catalytically rele-
vant sites on supported sulfide samples with good IR light trans-
missibility [33,37]. However, for bulk sulfide catalysts, we failed
to obtain IR spectra. In this final section, we attempt to correlate
the hydrogenation rates of the HCl-treated NiaMoS2 catalysts with
the concentrations of coordinatively unsaturated sites (CUS) deter-
mined using the chemisorption of nitrogen monoxide (NO) [15,52].
Values of NO uptakes are compiled in Table S3. As described
earlier, HCl treatment caused varying degrees of changes to the
textural properties of the bulk catalysts, especially Ni0.1MoS2 and
Ni0.4MoS2, though the primary slab structure was not significantly
affected. For a fair comparison of activities of different parent and
HCl-treated formulations, H2-addition rates and NO uptakes were
normalized by the specific surface area (Fig. 5).
3.4. Phenanthrene hydrogenation on physical mixtures and stacked
beds of NimSn and MoS2
To better understand the roles of NiSx particles in the hydro-
genation of phenanthrene, the reaction was carried out on various
physical mixtures of a homemade NimSn sample and MoS2. These
experiments were designed in a way akin to earlier works by Del-
mon and coworkers [48–51]. Note that this NimSn sample was not
active for this reaction. Mo-based H2-addition rates (Table 5) were
substantially enhanced (by a factor of 3.0–3.1 compared to pure
MoS2) on inter-particle mixtures containing NimSn and MoS2 in
mass ratios of 1:1 and 3:1. More remarkably, all physical mixtures
showed a selectivity pattern similar to that of the Ni-promoted cat-
alysts, with comparable selectivities to DiHPhe and TetHPhe and a
lower-than-unity ratio (0.9) of path 1 and path 2 (Table 6). This
means that a promotional effect arises even by physical contact
of NimSn and MoS2 particles, reminiscent of the ‘‘remote control/-
contact synergy” model [25,26,31].
Several NiSx phases exhibit Tammann temperatures lower than
270 °C. Thus, nickel sulfide particles are hypothesized to be highly
mobile at the reaction temperature of 300 °C. We hypothesize that
nickel sulfide particles in an interparticle mixture provide a means
to dynamically incorporate Ni into MoS2 slabs under reaction con-
ditions, accounting for the observed rate enhancement and selec-
tivity shifts with physical mixtures of NimSn and MoS2. At even
higher mixing ratios of NimSn and MoS2, H2-addition rates normal-
ized to Mo decreased and approached that of pure MoS2, likely due
to potential blockage of active sites with excessive presence of
nickel sulfides in the vicinity of MoS2.
The NimSn sample had a small BET surface area (11 m2 gꢂ1) but
showed a considerable NO uptake (35 l
mol gꢂ1). This sample was
inactive for phenanthrene hydrogenation under the conditions
studied. Since parent catalysts with high Ni contents, especially
Ni0.6MoS2 and Ni0.7MoS2, contain a great number of nickel sulfide
crystallites, the correlation between catalytic activity and NO
uptake was poor among the parent samples (filled squares in
Fig. 5). HCl treatment removes a substantial portion of the large
nickel sulfide crystallites and allows gaining increased access of
reactant and site-specific probe molecules (such as NO) to the
active sites located at the edge of the (Ni)MoS2 slab. Given that NiSx
at accessible surfaces had been successfully removed by HCl, an
improved correlation between catalytic activity and NO uptake
was expected for the leached series, as was indeed observed,
except for the leached Ni0.1MoS2 (Fig. 5). Parent samples without
too many nickel sulfide particles, NiaMoS2 (a = 0.1–0.5), appear
to lie along the same trend line, albeit with pronounced scatter.
By normalizing the measured rates against the corresponding
NO uptakes, we were able to obtain quasi-TOF (‘‘TOF”), a site-
specific activity that does not take into account the dependence
of site density on the conditions (primarily, temperature, pressure
and H2S/H2 ratio). For all leached samples except Ni0.1MoS2-L, a
steeper slope is observed, corresponding to an estimated ‘‘TOF”
of 100 hꢂ1 that represents the H2-addition activity of Ni-
promoted CUS at the slab of MoS2. It is likely that the degree of
Ni edge substitution is high in those samples and the activity is
predominantly contributed by promoted CUS, which is probed by
NO and not heavily blocked by NiSx. The Ni0.1MoS2-L catalyst can
be readily identified as an outlier in the leached series, exhibiting
a significantly smaller H2-addition ‘‘TOF” of 17 hꢂ1 that is compa-
rable to that of pure MoS2. Recall that this catalyst behaved kinet-
ically similar to a non-promoted MoS2 catalyst in terms of the
Additionally, an experiment with separated beds of NimSn and
MoS2 by 4 mm of
c-Al2O3 was carried out (see Experimental).
Separating NimSn from MoS2 increased the H2-addition rate by
83% (from 0.6 to 1.1 molH2 mol hꢂ1), indicating that H2 can be
-1
Mo
activated on NimSn and spill over to separated MoS2 placed down-
stream. At the same time, the pathway ratio remained essentially
the same as for the pure MoS2, which indicates that no direct pro-
motion occurred without physical contact and that the modest rate
enhancement is only due to the increased supply of H-species via
spillover. Since nickel sulfide particles are unable to traverse the
long distance (4 mm) of the Al2O3 bed [50], this experiment serves
to substantiate the local effect of NiSx on the dynamic formation of
Ni-promoted sites as observed for intimately mixed particles.
Table 5
Reaction rates of phenanthrene hydrogenation at 300 °C and 60 bar total pressure on physical mixtures and stacked beds of NimSn and MoS2.
Catalystsa
Ni/Mo [mol molꢂ1
]
Reaction rates
[molPhe/(molmetal h)]
[molPhe/(molMo h)]
[molH2 molMo hꢂ1
]
MoS2
0
0
0.53
0.54
0.39
0.16
0.04
0.03
0.13
0.53
0.54
1.17
1.16
0.47
0.61
0.93
0.61
0.63
1.9
1.8
0.7
MoS2 (undiluted)
NimSn/MoS2(1:1)
NimSn/MoS2(3:1)
NimSn/MoS2(6:1)
NimSn/MoS2(9:1)
NimSn–Al2O3–MoS2(3:1)b
2.0
6.2
12.2
22.7
6.3
0.9
1.1
a
Values in parentheses are the mass ratio of NimSn and MoS2. The total mass of the catalyst bed was kept identical in all experiments.
NimSn and MoS2 was separated by a 4 mm bed of c-Al2O3.
b

F. Vogelgsang et al. / Journal of Catalysis 391 (2020) 212–223
221
Table 6
Initial product selectivities of phenanthrene hydrogenation at 300 °C and 60 bar total pressure obtained by extrapolating to zero contact time. Pathway ratio (path 1/path 2) is
constant for NimSn/MoS2 mixtures. For stacked beds of NimSn and MoS2 separated by 4 mm -Al2O3, the ratio was shifted to that measured for pure MoS2.
c
Catalystsa
Initial selectivities (%)
DiHPhe
Pathway ratio
TetHPhe
sym-OHPhe
asym-OHPhe
NimSn/MoS2(1:1)
NimSn/MoS2(3:1)
NimSn/MoS2(6:1)
NimSn/MoS2(9:1)
NimSn–Al2O3–MoS2(3:1)b
45
47
48
47
84
52
52
52
53
16
2
1
0
0
0
1
0
0
0
0
0.9
0.9
0.9
0.9
5.3
a
Values in parentheses are the mass ratio of NimSn and MoS2. The total mass of the catalyst bed was kept identical in all experiments.
NimSn and MoS2 was separated by a 4 mm bed of c-Al2O3.
b
350
300
250
200
150
100
50
samples with relatively high Ni loadings, i.e., the edge-
incorporated Ni remained intact upon HCl attack.
However, the effects of HCl treatment turned out to be more
intricate than we had thought. Two distinct observations have sur-
faced from this study: (1) HCl treatment caused the removal of
edge-incorporated Ni from the slab for the lowest-Ni-content sam-
ple (Ni0.1MoS2); (2) HCl treatment caused the decrease in catalytic
activity for less Ni-loaded catalysts. It appears that removing NiSx
on high-Ni-loaded samples universally increases the accessible
surfaces; this is seen from the generally increased surface areas
and pore volumes after HCl treatment of such samples in our pre-
vious and this work. This is the main driver behind the rate
Ni0.6MoS2-P
Ni0.7MoS2-P
enhancement. On the other hand, for parent catalysts with lower
-1
Ni contents (i.e., < 0.4 molNi mol
) and less (and likely smaller)
metal
NiSx, the accessibility of the NiMoS phase is most likely no longer
the limiting factor, and the major consequence of removing NiSx
turns detrimental, probably because (i) sites on NiSx surfaces can
serve to dissociate H2 and (ii) structural stability against HCl attack
is compromised, the cause of which remains unclear. In the
extreme case of Ni0.1MoS2, without enough NiSx at the periphery,
a loss of edge-incorporated Ni even occurred, leading to a drastic
rate loss (Table 3).
Summarizing, only high-Ni-loaded unsupported sulfides were
used in our previous works [34,35], while this study expanded
the range of Ni contents to encompass those compositions that cor-
responded to less-than-complete edge substitution. The added
value of this work lies in this extension and the implications of
the complex trends as discussed above, providing a more compre-
hensive view on the roles of NiSx. Admittedly, there is a lot more to
understand about this leaching procedure and a more systematic
investigation of the process variables (duration, temperature,
exposure to air, etc.) needs to be carried out in the future.
Ni0.1MoS2-L
MoS2
1.0
0
0.0
2.0
3.0
4.0
5.0
6.0
-2
NO uptake / surface area [µmol m ]
Fig. 5. Hydrogen addition rate (calculated from the phenanthrene consumption and
the moles of hydrogen consumed along each route) as a function of NO uptake, both
normalized to the specific surface area. Open squares represent the leached
versions of the parent samples (filled squares), unpromoted MoS2 is identified via
the filled circle. The slopes of the lines can be interpreted as quasi-TOFs for the
promoted (100 hꢂ1) and non-promoted sites (17 hꢂ1), respectively.
pathway ratio. Together, these results speak strongly for the fact
that a majority of Ni atoms incorporated at the slab edge have been
removed in this particular sample. The TOFs calculated based on
phenanthrene conversion are 55 and 8.7 hꢂ1 for promoted and
non-promoted CUS, respectively. Notably, the 6-time difference
in the ‘‘TOFs” between promoted and unpromoted sites is similar
in magnitude to that reported previously for Al2O3-supported Ni-
Mo sulfide catalysts at the same reaction conditions [37]; in that
work, TOFs of the Ni-promoted and non-promoted sites were 40
and 6 hꢂ1, respectively, albeit with a different method (based on
SH counts and IR spectroscopy) to enumerate active sites.
4. Conclusion
We have shown for the catalytic hydrogenation of phenan-
threne that the ubiquitous presence of separate NiSx phases in
unsupported Ni-Mo binary sulfide catalysts has multiple conse-
quences, some being detrimental to catalytic activity, some being
beneficial. These phases are inherently much less active than
unpromoted and promoted sites in (Ni)MoS2 slabs, but under reac-
tion conditions, they are potentially able to dynamically incorpo-
rate Ni into slabs originally without Ni or with low edge
substitution degrees. NiSx can also activate H2, and because of
the higher surface-to-volume ratios, small particles, which are
more prevalent in low Ni-content samples, do so more efficiently
than large particles, which are more frequently found on high Ni-
content samples. Therefore, the combined effects of nickel sulfide
phases on the hydrogenation activity depend on the Ni content,
which determines the abundance and size of these entities.
3.6. On the complexity of the HCl treatment
In a most recent study [35], we applied the HCl treatment to a
series of unsupported NiMo sulfides containing high concentra-
tions of Ni (0.48–0.60 atomic metal fraction). For those compact
(less porous) NiMo sulfides, a significant portion (but not all) of
NiSx was found to be selectively removed by the HCl extraction
whereas the decorated NiMoS phase stayed intact. Unsupported
NiMo sulfides studied in the present work exhibited very different
textural properties and morphologies (Figs. S1-S3); in general, the
samples used here were a lot more porous, and the Ni contents var-
ied in a much wider range (0.13–0.72 atomic metal fraction). For
all Ni-loaded samples except for Ni0.1MoS2, we reached the same
conclusion as in the previous work for unsupported NiMo sulfide
Post-synthetic treatment of high Ni-content (e.g., > 0.4 molNi
-1
mol
as investigated in this work) bulk sulfide catalysts with
metal

222
F. Vogelgsang et al. / Journal of Catalysis 391 (2020) 212–223
[5] M. Egorova, R. Prins, Hydrodesulfurization of dibenzothiophene and 4,6-
dimethyldibenzothiophene over sulfided NiMo/ -Al2O3, CoMo/ -Al2O3, and
Mo/ -Al2O3 catalysts, J. Catal. 225 (2004) 417–427.
[6] L. Van Haandel, M. Bremmer, P.J. Kooyman, J.A.R. Van Veen, T. Weber, E.J.M.
Hensen, Structure-activity correlations in hydrodesulfurization reactions over
Ni-promoted MoxW(1–x)S2/Al2O3 catalysts, ACS Catal. 5 (2015) 7276–7287.
[7] S. Eijsbouts, S.W. Mayo, K. Fujita, Unsupported transition metal sulfide
catalysts: From fundamentals to industrial application, Appl. Catal. A Gen.
322 (2007) 58–66.
[8] R.R. Chianelli, G. Berhault, B. Torres, Unsupported transition metal sulfide
catalysts: 100 years of science and application, Catal. Today. 147 (2009) 275–
286.
[9] J. Bocarando, R. Huirache-Acuña, W. Bensch, Z.D. Huang, V. Petranovskii, S.
Fuentes, G. Alonso-Núñez, Unsupported Ni-Mo-W sulphide HDS catalysts with
the varying nickel concentration, Appl. Catal. A Gen. 363 (2009) 45–51.
[10] A. Olivas, D.H. Galván, G. Alonso, S. Fuentes, Trimetallic NiMoW unsupported
catalysts for HDS, Appl. Catal. A Gen. 352 (2009) 10–16.
[11] S. Shan, H. Liu, Y. Yue, G. Shi, X. Bao, Trimetallic WMoNi diesel ultra-deep
hydrodesulfurization catalysts with enhanced synergism prepared from
inorganic–organic hybrid nanocrystals, J. Catal. 344 (2016) 325–333.
[12] J. Hein, O.Y. Gutiérrez, E. Schachtl, P. Xu, N.D. Browning, A. Jentys, J.A. Lercher,
Distribution of metal cations in Ni-Mo-W sulfide catalysts, ChemCatChem. 7
(2015) 3692–3704.
[13] H. Topsøe, B.S. Clausen, R. Candia, C. Wivel, S. Mørup, In situ Mössbauer
emission spectroscopy studies of unsupported and supported sulfided Co-Mo
hydrodesulfurization catalysts: evidence for and nature of a Co-Mo-S phase, J.
Catal. 68 (1981) 433–452.
[14] C. Wivel, R. Candia, B.S. Clausen, S. Mørup, H. Topsøe, On the catalytic
significance of a CoMoS phase in CoMo/Al2O3 hydrodesulfurization catalysts:
Combined in situ Mössbauer emission spectroscopy and activity studies, J.
Catal. 68 (1981) 453–463.
[15] N.-Y. Topsøe, H. Topsøe, Characterization of the structures and active sites in
sulfided Co-Mo/Al2O3 and Ni-Mo/Al2O3 catalysts by NO chemisorption, J. Catal.
84 (1983) 386–401.
concentrated HCl successfully dissolves large crystallites of NiSx at
accessible surfaces, leading to a reduction of catalytically inactive
sulfides and an increased exposure of active sites. These two pri-
mary positive consequences overcompensate the loss of some enti-
ties that are able to provide additional H2 activation sites, and,
thus, the mass-specific and total-metal-based reaction rates of
these catalysts are significantly enhanced. This leaching strategy,
however, is associated with the risk of leaching Ni-promoted active
sites within the slabs of the main active phase, particularly when
the amount of NiSx is insufficient to retard the attack of HCl on
c
c
c
the primary slab structure. For the lowest Ni-loaded catalyst stud-
-1
ied here (0.05 molNi mol
), HCl treatment caused a catastrophic
metal
loss of catalytic activity by a factor of > 6, and even shifted the
selectivity pattern to that of pure MoS2. Such selectivity shifts were
not captured on other catalysts. These observations, together with
the sensitivity of the regioselectivity to Ni promotion for this reac-
tion, lead us to conclude that Ni atoms incorporated at the slab
edges are inherently susceptible to HCl attack, but NiSx appear to
be the preferential sites where HCl attack and dissolution starts.
Disrupting the Ni-promoted sites and reducing the small NiSx are
the two major causes for the substantially decreased rates on
low Ni-content samples upon the HCl treatment.
Finally, we have demonstrated that HCl treatment leads to a
greatly improved correlation between areal activity and the areal
concentration of chemisorbed NO on the leached catalysts, with
the exception of the lowest Ni-containing one. By estimating its
Ni content and the fraction of edge atoms, we conclude that this
catalyst has a low edge substitution degree of Ni (<30% at the slab
edge) and contains a predominant fraction of edge sites as unpro-
moted sites. The slope of the linear correlation indicates that the
Ni-promoted active sites in NiMoS2 are more than five times as
active as the unpromoted sites, a similar enhancement factor com-
pared to that observed earlier for the Al2O3-supported
counterparts.
[16] S. Eijsbouts, On the flexibility of the active phase in hydrotreating catalysts,
Appl. Catal. A Gen. 158 (1997) 53–92.
[17] E. Krebs, B. Silvi, P. Raybaud, Mixed sites and promoter segregation: A DFT
study of the manifestation of Le Chatelier’s principle for the Co(Ni)MoS active
phase in reaction conditions, Catal. Today. 130 (2008) 160–169.
[18] B. Guichard, M. Roy-Auberger, E. Devers, C. Legens, P. Raybaud, Aging of
Co(Ni)MoP/Al2O3 catalysts in working state, Catal. Today. 130 (2008) 97–
108.
[19] S. Eijsbouts, L.C.A. Van Den Oetelaar, R.R. Van Puijenbroek, MoS2 morphology
and promoter segregation in commercial Type 2 Ni-Mo/Al2O3 and Co-Mo/
Al2O3 hydroprocessing catalysts, J. Catal. 229 (2005) 352–364.
[20] S. Eijsbouts, X. Li, J. Bergwerff, J. Louwen, L. Woning, J. Loos, Nickel sulfide
crystals in Ni-Mo and Ni-W catalysts: Eye-catching inactive feature or an
active phase in its own right?, Catal. Today. 292 (2017) 38–50.
[21] B.S. Clausen, H. Topsøe, F.E. Massoth, J.R. Anderson, M. Boudart (Eds.),
Hydrotreating Catalysis, Science and Technology, Vol 11, Springer, Berlin,
1996.
[22] R.R. Chianelli, Fundamental studies of transition metal sulfide
hydrodesulfurization catalysts, Catal. Rev. 26 (1984) 361–393.
[23] M. Karroua, H. Matralis, P. Grange, B. Delmon, Unsupported NiMo catalysts.
influence of the sulfiding temperature and evolution of the unsupported
NiMoS phase during reaction, Bull. Soc. Chim. Beig. 104 (1995) 11–18.
[24] M. Karroua, A. Centeno, H.K. Matralis, P. Grange, B. Delmon, Synergy in
hydrodesulphurization and hydrogenation on mechanical mixtures of
cobalt sulphide on carbon and MoS2 on alumina, Appl. Catal. 51 (1989)
L21–L26.
Declaration of Competing Interest
The authors declare that they have no known competing finan-
cial interests or personal relationships that could have appeared
to influence the work reported in this paper.
Acknowledgement
This work was financially supported by the Chevron Energy
Technology Company. The authors would like to especially thank
Dr. Alexander Kuperman, Dr. Axel Brait, and Prof. Roel Prins for
fruitful discussions. We are also grateful to the members of TU
München for their support in different ways, especially Ulrike
Ammari (Elemental Analysis), Roland Weindl (TEM) and Xaver
Hecht (technical support).
[25] M. Karroua, H. Matralis, P. Grange, B. Delmon, Synergy between ‘‘NiMoS” and
Co9S8 in the hydrogenation of cyclohexene and hydrodesulfurization of
thiophene, J. Catal. 139 (1993) 371–374.
[26] B. Delmon, Mobility of atomic species on catalyst surfaces (spillover): Effects
on creation of active sites and stability of catalysts, Appl. Catal. A Gen. 113
(1994) 121–124.
[27] J.M.J.G. Lipsch, G.C.A. Schuit, The CoO-MoO3-Al2O3 catalyst II. The structure of
the catalyst, J. Catal. 15 (1969) 174–178.
Appendix A. Supplementary material
[28] R.J.H. Voorhoeve, J.C.M. Stuiver, Kinetics of hydrogenation on supported and
bulk Nickel-Tungsten sulfide catalysts, J. Catal. 23 (1971) 228–235.
[29] R.J.H. Voorhoeve, J.C.M. Stuiver, The mechanism of the hydrogenation of
cyclohexene and benzene on Nickel-Tungsten sulfide catalysts, J. Catal. 23
(1971) 243–252.
Supplementary data to this article can be found online at
https://doi.org/10.1016/j.jcat.2020.08.026.
[30] R.J.H. Voorhoeve, Electron spin resonance study of active centers in Nickel-
Tungsten sulfide hydrogenation catalysts, J. Catal. 23 (1971) 236–242.
[31] B. Delmon, A new hypothesis explaining synergy between two phases in
heterogeneous catalysis: the case of hydrodesulfurization catalysts, Bull. Soc.
Chim. Belg. 88 (1979) 979–987.
[32] P. Ratnasamy, S. Sivasanker, Structural chemistry of Co-Mo-alumina catalysts,
Catal. Rev. Sci. Eng. 22 (1980) 401–429.
[33] W. Luo, H. Shi, M. Wagenhofer, O. Gutiérrez, J. Lercher, Maximizing active site
concentrations at Ni-substituted WS2 edges for hydrogenation of aromatic
molecules, J. Phys. Chem. Lett. 10 (2019) 5617–5622.
References
[1] H. Toulhoat, P. Raybaud (Eds.), Catalysis by Transition Metal Sulfides: From
Molecular Theory to Industrial Application, Éditions Technip, Paris, 2013.
[2] H. Topsøe, B.S. Clausen, Importance of Co-Mo-S type structures in
hydrodesulfurization, Catal. Rev. 26 (1984) 395–420.
[3] J.V. Lauritsen, S. Helveg, E. Lægsgaard, I. Stensgaard, B.S. Clausen, H. Topsøe, F.
Besenbacher, Atomic-scale structure of Co-Mo-S nanoclusters in hydrotreating
catalysts, J. Catal. 197 (2001) 1–5.
[34] S. Albersberger, H. Shi, M. Wagenhofer, J. Han, O.Y. Gutiérrez, J.A. Lercher, On
the enhanced catalytic activity of acid-treated, trimetallic Ni-Mo-W sulfides
for quinoline hydrodenitrogenation, J. Catal. 380 (2019) 332–342.
[4] M. Lacroix, N. Boutarfa, C. Guillard, M. Vrinat, M. Breysse, Hydrogenating
properties of unsupported transition metal sulphides, J. Catal. 120 (1989) 473–
477.

F. Vogelgsang et al. / Journal of Catalysis 391 (2020) 212–223
223
[35] M.F. Wagenhofer, H. Shi, O.Y. Gutiérrez, A. Jentys, J.A. Lercher, Enhancing
hydrogenation activity of Ni-Mo sulfide hydrodesulfurization catalysts, Sci.
Adv. 6 (2020) eaax5331.
[44] E.J.M. Hensen, P.J. Kooyman, Y. Van der Meer, A.M. Van der Kraan, V.H.J. De
Beer, J.A.R. Van Veen, R.A. Van Santen, P.J. Kooyman, Y. Van der Meer, A.M. Van
der Kraan, The relation between morphology and hydrotreating activity for
supported MoS2 particles, J. Catal. 199 (2001) 224–235.
[36] E. Schachtl, L. Zhong, E. Kondratieva, J. Hein, O.Y. Gutiérrez, A. Jentys, J.A.
Lercher, Understanding Ni promotion of MoS2/
the hydrogenation of phenanthrene, ChemCatChem.
4130.
c
-Al2O3 and its implications for
[45] S. Kasztelan, H. Toulhoat, J. Grimblot, J.P. Bonnelle, A geometrical model of the
active phase of hydrotreating catalysts, Appl. Catal. 13 (1984) 127–159.
[46] N.A. Bhore, M.T. Klein, K.B. Bischoff, The Delplot technique: a new method for
reaction pathway analysis the First-Rank Delplot: identification of products of
primary rank, Ind. Eng. Chem. Res. 29 (1990) 313–316.
7 (2015) 4118–
[37] W. Luo, H. Shi, E. Schachtl, O.Y. Gutiérrez, J.A. Lercher, Active sites on Nickel-
promoted transition-metal sulfides that catalyze hydrogenation of aromatic
compounds, Angew. Chemie - Int. Ed. 57 (2018) 14555–14559.
[47] C.G. Frye, Equilibria in the hydrogenation of polycylic aromatics, J. Chem. Eng.
Data. 7 (1962) 592–595.
[38] M. Ruinart De Brimont, C. Dupont, A. Daudin, C. Geantet, P. Raybaud,
Deoxygenation mechanisms on Ni-promoted MoS2 bulk catalysts:
A
[48] S.G. de León, P. Grange, B. Delmon, New evidence of the promotion of activity
and selectivity through hydrogen spill-over activation of biphasic sulfided
hydrotreating catalysts, Stud. Surf. Sci. Catal. 345–348 (1993).
[49] J. Ojeda, N. Escalona, P. Baeza, M. Escudey, F.J. Gil-Llambías, Synergy between
Mo/SiO2 and Co/SiO2 beds in HDS: A remote control effect?, Chem. Commun. 3
(2003) 1608–1609.
combined experimental and theoretical study, J. Catal. 286 (2012) 153–164.
[39] G. Alonso, M.H. Siadati, G. Berhault, A. Aguilar, S. Fuentes, R.R. Chianelli,
Synthesis of tetraalkylammonium thiometallate precursors and their
concurrent
in
situ
activation
during
hydrodesulfurization
of
dibenzothiophene, Appl. Catal. A Gen. 263 (2004) 109–117.
[40] G. Alonso, M. Del Valle, J. Cruz, V. Petranovskii, A. Licea-Claverie, S. Fuentes,
Preparation of MoS2 catalysts by in situ decomposition of
tetraalkylammonium thiomolybdates, Catal. Today. 43 (1998) 117–123.
[41] B. Yoosuk, J.H. Kim, C. Song, C. Ngamcharussrivichai, P. Prasassarakich, Highly
active MoS2, CoMoS2 and NiMoS2 unsupported catalysts prepared by
[50] P. Baeza, M.S. Ureta-Zañartu, N. Escalona, J. Ojeda, F.J. Gil-Llambías, B. Delmon,
Migration of surface species on supports:
A proof of their role on the
synergism between CoSx or NiSx and MoS2 in HDS, Appl. Catal. A Gen. 274
(2004) 303–309.
[51] N. Escalona, R. García, G. Lagos, C. Navarrete, P. Baeza, F.J. Gil-Llambías, Effect
of the hydrogen spillover on the selectivity of dibenzothiophene
hydrothermal
synthesis
for
hydrodesulfurization
of
4,6-
dimethyldibenzothiophene, Catal. Today. 130 (2008) 14–23.
hydrodesulfurization over CoSx/
c-Al2O3, NiSx/c-Al2O3 and MoS2/c-Al2O3
[42] I. Bezverkhyy, M. Danot, P. Afanasiev, New low-temperature preparations of
some simple and mixed Co and Ni dispersed sulfides and their chemical
behavior in reducing atmosphere, Inorg. Chem. 42 (2003) 1764–1768.
[43] E. Peters, Direct leaching of sulfides: chemistry and applications, Metall. Trans.
B. 4 (1976) 505.
catalysts, Catal. Commun. 7 (2006) 1053–1056.
[52] N. Koizumi, S. Jung, Y. Hamabe, H. Suzuki, M. Yamada, Investigation of Surface
Site of Ni Species on NiMo/Al2O3 Hydrodesulfurization Catalyst Sulfided at
High-Pressure by Means of DRIFTS Combined with Low-Temperature NO
Adsorption, Catal. Lett. 135 (2010) 175–181.