Elucidation of molybdenum-based catalysts using a radioisotope tracer method part 2? promotion effect of cobalt on molybdena/alumina catalyst

Article abstract of DOI:10.1039/a705224c

Hydrodesulfurization (HDS) reactions of radioactive ³⁵S-labelled dibenzothiophene were carried out over a series of cobalt-promoted molybdena/alumina catalysts at temperatures between 240 and 320°C and under 50 kg cm^-2 pressure. HDS and hydrogenation (HYD) activities of Mo/Al₂O₃ catalysts were remarkably enhanced with the addition of cobalt, the maximum promotional effect occurring with a Co/Mo molar ratio of ca. 0.5. The amount of labile sulfur and the rate constant of sulfur exchange increased significantly with the addition of cobalt, indicating that cobalt makes the sulfur more mobile and that the active phases in the promoted catalysts are different from that in the unpromoted catalyst. The promoting effect of cobalt was thus attributed to the formation of more active sites. On the other hand, the amount of labile sulfur increased linearly with the Co/Mo ratio up to ca. 0.5 whereas the rate constant of sulfur exchange remained almost constant. Moreover, the enhancement in catalytic activity did not change significantly with the cobalt content. The increase in the catalytic activity with the cobalt content was thus ascribed to an increase in the number of the same active sites.

Full text of DOI:10.1039/a705224c

View Article Online / Journal Homepage / Table of Contents for this issue
Elucidation of molybdenum-based catalysts using a radioisotope
tracer method
Part 2¤ Promotion eþ ect of cobalt on molybdena/alumina catalyst
Weihua Qian, Atsushi Ishihara, Yuzo Okoshi, Waka Nakakami, Masazumi Godo and
Toshiaki Kabe*
Department of Applied Chemistry, T okyo University of Agriculture and T echnology, 2-24-16
Nakacho, Koganei, T okyo 184, Japan
Hydrodesulfurization (HDS) reactions of radioactive 35S-labelled dibenzothiophene were carried out over a series of cobalt-
promoted molybdena/alumina catalysts at temperatures between 240 and 320 ¡C and under 50 kg cm~2 pressure. HDS and
hydrogenation (HYD) activities of Mo/Al O catalysts were remarkably enhanced with the addition of cobalt, the maximum
2
3
promotional e†ect occurring with a Co/Mo molar ratio of ca. 0.5. The amount of labile sulfur and the rate constant of sulfur
exchange increased signiÐcantly with the addition of cobalt, indicating that cobalt makes the sulfur more mobile and that the
active phases in the promoted catalysts are di†erent from that in the unpromoted catalyst. The promoting e†ect of cobalt was
thus attributed to the formation of more active sites. On the other hand, the amount of labile sulfur increased linearly with the
Co/Mo ratio up to ca. 0.5 whereas the rate constant of sulfur exchange remained almost constant. Moreover, the enhancement in
catalytic activity did not change signiÐcantly with the cobalt content. The increase in the catalytic activity with the cobalt content
was thus ascribed to an increase in the number of the same active sites.
HDS of sulfur-containing petroleum fractions is a major cata-
lytic process in the petroleum industry. Alumina-supported
cobalt or nickel-promoted molybdenum-based catalysts are
most widely used. Recently, the deep desulfurization of light
gas oil has become an urgent problem because of ever stricter
environmental regulations.1h3 It is, therefore, necessary to
develop a new catalyst with high activity and selectivity. There
have been various attempts to improve the performance of
conventional CoÈMo/Al O catalysts.4,5 One conventional
of cobalt was investigated and the correlation of HDS activity
with sulfur exchange data was also discussed.
Experimental
Materials
Decalin, used as a solvent, was commercially available GR
grade (Kishida Chemicals). Dibenzothiophene (32S-DBT) and
35S-labelled dibenzothiophene (35S-DBT) were synthesized
according to a previous method.7 Hydrogen (99.99%) was
obtained from Tohei Chemicals. Hydrogen sulÐde in hydro-
2
3
method is to increase the content of active metal species.
Therefore, the dispersion state of molybdenum on the alumina
support for the base catalyst Mo/Al O , and the promotion
2
3
gen (H S 5.0%) was obtained from Takachio Chemicals.
e†ect of cobalt must be clariÐed. Structural and catalytic
aspects are now relatively well understood, however, problems
remain concerning the behaviour of sulfur on working cata-
lysts. We have developed a direct method to estimate the
behaviour of sulfur on the working catalysts. Radioactive 35S-
labelled dibenzothiophene (DBT) was synthesized and its
HDS was carried out on sulÐded catalysts under practical
HDS conditions.6h8 By tracing the change in unreacted
2
Ammonium
heptamolybdate
tetrahydrate
[(NH ) Mo O É 4H O] and cobalt nitrate hexahydrate
4 6 24
[Co(NO ) É 6H O] were commercial GR grade (Kishida
7
2
3 2
2
Chemicals). All scintillation solvents for measuring radioac-
tivity were obtained from Packard Japan Co. Ltd. The c-
Al O used as support was supplied as 1/32 inch extrudates
and obtained from Nippon Ketjen Co. Ltd. (KF-13775). The
physical properties of the support are surface area 256 m2
g~1, pore volume 0.712 ml g~1 and mean pore diameter 8.2
nm.
2
3
35S-DBT and the H 35S formed the behaviour of sulfur on the
2
working catalyst can be monitored in situ.
In a previous paper, we investigated the behaviour of sulfur
on a series of sulÐded Mo/Al O catalysts, containing 6 to 20
2
3
Preparation of catalysts
wt.% molybdena, using a 35S radioisotope pulse tracer
method (RPTM).9 The sulfur exchange rates were approx-
imately the same for all samples and apparent activation ener-
gies were also the same. These results suggested that the
mechanism of HDS and the nature of the active sites did not
vary with Mo content. Moreover, the amount of labile sulfur
increased linearly with the molybdena content up to 16 wt.%
and then levelled o† with further loading of molybdena. This
suggests that monolayer dispersion of molybdenum sulÐde on
alumina was maintained up to 16 wt.% but some crystalline
molybdenum sulÐde was formed at higher molybdena con-
tents. On the basis of these results, the Mo/Al O catalyst
On the basis of the results reported in ref. 9, a Mo/Al O
2
3
catalyst containing 16 wt.% MoO was selected as base cata-
3
lyst. The cobalt content was varied between 0 and 7.7 wt.%
CoO. The catalysts were prepared by the usual impregnation
method: the alumina supports were initially impregnated with
an aqueous solution of ammonium heptamolybdate tetra-
hydrate and subsequently with an aqueous solution of cobalt
nitrate hexahydrate as required. After impregnation of every
metal precursor, the catalyst was dried at 120 ¡C for 3 h and
then calcined in air at 450 ¡C for 15 h. The catalysts were
denoted as Mo(16), Co(1)ÈMo(16), Co(2)ÈMo(16), Co(3)È
Mo(16), Co(5)ÈMo(16) and Co(8)ÈMo(16), respectively. The
2
3
containing 16 wt.% of MoO was selected as a base catalyst
3
for CoÈMo catalysts in this study. A series of Co-promoted
value in parentheses indicates the wt.% of CoO or MoO in
3
catalysts with 16 wt.% MoO and 1.0È7.7 wt.% Co were pre-
the catalyst e.g. Co(1)ÈMo(16) ca. 1.0 wt.% CoO and 16.0
3
pared and HDS of 35S-DBT carried out. The promoting e†ect
wt.% MoO . The physical properties of these catalysts are
3
given in Table 1. The molar ratios, Co/Mo, for the catalysts
¤ Part 1: ref. 9.
were ca. 0, 0.12, 0.24, 0.40, 0.60 and 1.00.
J. Chem. Soc., Faraday T rans., 1997, 93(24), 4395È4400
4395

View Article Online
Table 1 Physical properties of prepared catalysts
chemical composition
BET
Mo content
Co content
catalyst
Mo(16)
Co(1)ÈMo(16)
Co(2)ÈMo(16)
Co(3)ÈMo(16)
Co(5)ÈMo(16)
Co(8)ÈMo(16)
MoO (wt.%)
CoO (wt.%)
surface area/m2 g~1
/atom nm~2
/atom nm~2
Co/Mo
3
15.8
15.8
15.5
15.3
15.3
14.6
0.0
229
227
224
220
215
206
2.89
2.89
2.90
2.91
2.98
2.97
0.00
0.35
0.70
1.18
1.78
3.00
0.0
0.99
1.96
3.22
4.77
7.69
0.12
0.24
0.40
0.60
1.01
35S RPTM
wt.% solution of 35S-DBT in decalin was then substituted for
the 32S-DBT after ca. 90 min and was reacted until the forma-
35S-DBT tracer experiments were carried out with a Ðxed-bed
reactor of conventional design, the details are described else-
where.9 After calcination at 430 ¡C in air overnight, the cata-
lyst was presulÐded with a mixture of 5% H S in H Ñowing
at 1.8 l h~1, atmospheric pressure, and 400 ¡C for 3 h. After
presulÐdation, the reactor was cooled, in an H SÈH stream,
to the reaction temperature and pressurized with hydrogen. A
decalin solution of DBT was then fed into the reactor by a
high-pressure liquid pump (Kyowa Seimitsu KHD-16).
tion amount of 35S-H S became constant. After 35S-DBT was
2
introduced, the radioactivity of the unreacted 35S-DBT in the
liquid product increased and approached a steady state imme-
diately. However, it took ca. 210 min to achieve a steady state
2
2
in radioactivity of the H 35S produced. In order to determine
2
the extent of exchanged sulfur more accurately, the 35S-DBT
2
2
solution was replaced by decalin solvent after ca. 315 min. It
was observed that a portion of 35S, which is represented with
the shaded area A in Fig. 1, remained on the catalyst. When
Typical reaction conditions were H Ñow rate 25 l h~1,
decalin and H were used to purge the catalyst for ca. 2.5 h,
2
2
WHSV 28 h~1, reaction pressure 50 kg cm~2, concentration
less release of H 35S was detected. This indicates that the
2
of DBT in decalin 1.0 wt.% and reaction temperature 240È
320 ¡C. A typical operation procedure was described in the
previous paper.9 The liquid product was collected from a gasÈ
sulfur accommodated on the catalyst could not be eluted
without a supply of sulfur derived from HDS of DBT.
However, when the reactant solution was replaced with
32S-DBT after ca. 480 min, this portion of 35S can be released
liquid separator. The H S produced was absorbed by bub-
2
bling through
a
commercial basic scintillation solution
again as H 35S, as shown in Fig. 1 (shaded area B). This
2
(Carbosorb, Packard Japan Co. Ltd.). For each run, the liquid
products and absorbed H S solution samples were collected
every 15 min. The components of the liquid products were
portion of 35S was approximately equal to the shaded area A,
which represented the total amount of labile sulfur on the
catalyst under these reaction condition. According to a
method reported in ref. 9, the amount of labile sulfur can be
2
analysed by gas chromatography using an FID detector
(Shimazu 17 A, Shimazu Co. Ltd.) and a commercial capillary
column (DB-1). Radioactivities of the unreacted 35S-DBT in
calculated from the total radioactivity of the H 35S released
2
after 32S-DBT was re-introduced i.e. the shaded area B.
the liquid product and the 35S-H S formed in the absorbed
To discuss the release process of H 35S more accurately, the
2
2
solution were measured with a liquid scintillation counter
release rate constant was determined. As reported previously,9
(Beckman LS-6500, Beckman, Co. Ltd.), after adding proper
scintillator solvent (PermaÑuor or InstaÑuor, Packard Japan
Co. Ltd.) to each fraction. Liquid scintillation counting is
described in ref. 7.
the release of H 35S in the range of area B can be Ðtted with a
2
Ðrst-order reaction. Thus, the release rate constant of H 35S
2
(k ), i.e. the rate constant of sulfur exchange was determined.
ER
To investigate the e†ect of reaction temperature, the same
radioisotope tracer experiments over Co(5)ÈMo(16) catalyst
were conducted with a solution of 1.0 wt.% 35S-DBT at 240,
260 and 280 ¡C under 50 kg cm~2 pressure. Fig. 2 shows the
Measurement of catalytic activity
The 35S radioisotope tracer method allows us to evaluate the
HDS and HYD activity simultaneously by estimating conver-
sion of dibenzothiophene using gas chromatography. The
detectable products were biphenyl (BP) and cyclo-
hexylbenzene (CHB). The rate constants (k ) of the pseudo-
Ðrst-order reaction of HDS of dibenzothiophene for this series
e†ect of temperature on the release of H 35S in the HDS of
2
35S-DBT. The steady state of radioactivity of 35S-DBT in the
liquid product was achieved immediately at every reaction
temperature and after replacing 35S-DBT with decalin solvent,
HDS
of catalysts were determined using
k
\ [ln(1 [ x/100)F/A
(1)
HDS
where k
is the rate constant of HDS, [(mol DBT converted)
HDS
s~1 (nm2 catalyst)~1], x is the percentage conversion of DBT,
F is the feed rate [(mol DBT) s~1], A is the surface area of the
catalyst, (nm2). Similarly, the formation of CHB was used to
describe the HYD reaction of DBT, and the rate constants
(k ) of the pseudo-Ðrst-order reaction of hydrogenation of
HYD
DBT were similarly determined from the formation of CHB.
Results
35S RPTM
The HDS of 35S-DBT was carried out on
a sulÐded
CoÈMo/Al O catalyst [Co(1)ÈMo(16)] at 260 ¡C and 50 kg
2
3
cm~2. Fig. 1 shows a typical result using the 35S RPTM. Ini-
tially, a decalin solution of 1 wt.% 32S-DBT was pumped into
the reactor until the conversion of DBT became constant. A 1
Fig. 1 Change in the radioactivity of unreacted 35S-DBT (L) and
H 35S formed (…) with reaction time. 50 kg cm~2, 260 ¡C, Co(1)È
2
Mo(16).
4396
J. Chem. Soc., Faraday T rans., 1997, V ol. 93

View Article Online
Fig. 2 E†ect of temperature on the release of 35S-H S. 50 kg cm~2,
Fig. 3 E†ect of Co/Mo ratio on sulfur exchange. 50 kg cm~2, 280 ¡C.
(|) Mo(16), (…) Co(1)ÈMo(16), (K) Co(2)ÈMo(16), (=) Co(3)ÈMo(16),
(L) Co(5)-Mo(16).
2
Co(5)ÈMo(16).
the radioactivity of 35S-DBT also decreased immediately at
every temperature, as shown in Fig. 1. Thus, only the change
in radioactivity of released H 35S was shown in Fig. 2. The
2
radioactivity of 35S remaining on the catalyst at a di†erent
temperature (area B), which corresponds to the amount of
labile sulfur, increased with temperature. Similarly, the
amounts of labile sulfur (S ) and rate constant of sulfur
0
exchange (k ) were determined and the values are sum-
ER
marized in Table 2, both increased with increasing tem-
perature.
The e†ect of the Co/Mo ratio on the behaviour of sulfur on
the catalysts was also investigated. Fig. 3 shows a signiÐcant
increase in area B at 280 ¡C for cobalt-promoted catalysts.
The values of S , corresponding to area B in Fig. 3, are given
0
in Table 3. Similarly, the k for all promoted catalysts were
ER
much greater than that for the unpromoted catalyst, whereas
little di†erence in k was observed among the promoted cata-
lysts at the same temperature.
In order to understand the promotion e†ect of cobalt, the
ER
Fig. 4 E†ect of Co/Mo ratio on the increase in S (+) and k (>)
0
ER
relative to those of Mo(16) catalyst at 260 ¡C
increase in S and k at 260 ¡C relative to the values of the
0
ER
Mo(16) catalyst (Table 4) with the Co/Mo ratio is shown in
of CHB increased markedly at lower Co/Mo molar ratio and
then decreased slightly at higher Co/Mo molar ratio.
Fig. 4. S increased markedly on addition of cobalt, then
increased linearly with increasing Co/Mo ratio up to ca. 0.5
and decreased slightly above this ratio. In contrast k
increased markedly on addition of Co and then remained con-
0
k
and k
were determined and are listed in Tables 2È4.
HDS
/k
HDS HDS0
6, k and k
HYD
ER
k
is plotted vs. Co/Mo for reaction at 260 ¡C in Fig.
being the rate constants of HDS on CoÈMo
stant with increasing Co/Mo ratio.
HDS
catalyst and Mo(16), respectively. Similarly, k /k
HYD HYD0
being the rate constants of
HDS0
is
plotted vs. Co/Mo, k
and k
HDS and HYD activity
HYD
HYD0
HYD on CoÈMo catalyst and Mo(16), respectively. A signiÐ-
Fig. 5 shows the e†ect of Co/Mo molar ratio on the conver-
sion of DBT, and the yield of CHB, on a series of CoÈMo
catalysts at 260 ¡C. The total conversion of DBT and the yield
cant promotion e†ect of cobalt on the molybdena catalyst for
both HDS and HYD activity was observed. k
and k
HDS
HYD
increased linearly with increasing Co content at lower Co
Table 2 Amount of labile sulfur, release rate constant of H S and activity of HYD and HDS on sulÐded Co(5)ÈMo(16) catalyst
2
temperature/¡C
conversion (%)
S /atom nm~2
k
/10~4 s~1
k
/10~4 atom s~1 nm~2
k
/10~4 atom s~1 nm~2
0
ER
HDS
HYD
240
260
280
35.8
67.0
94.7
0.91
1.21
1.38
4.95
6.23
8.25
5.3
13.1
34.9
0.38
0.67
1.14
Table 3 Amount of labile sulfur, release rate constant of H S and activity of HYD and HDS at 280 ¡C on sulÐded Co-promoted Mo/Al O
2
2
3
catalysts
catalyst
conversion (%)
S /atom nm~2
k
/10~4 s~1
k
/10~4 atom s~1 nm~2
k
/10~4 atom s~1 nm~2
0
ER
HDS
HYD
Mo(16)
9.4
64.3
83.5
91.8
94.7
0.27
0.87
1.13
1.30
1.38
2.43
8.12
7.92
8.00
8.25
1.1
11.6
20.5
29.1
34.9
0.17
0.66
1.08
0.94
1.14
Co(1)ÈMo(16)
Co(2)ÈMo(16)
Co(3)ÈMo(16)
Co(5)ÈMo(16)
J. Chem. Soc., Faraday T rans., 1997, V ol. 93
4397

View Article Online
Table 4 Amount of labile sulfur, release rate constant of H 35S and activity of HYD and HDS at 260 ¡C on sulÐded Co-promoted Mo/Al O
2
2 3
catalysts
catalyst
conversion (%)
S /atom nm~2
k
/10~4 s~1
k
/10~4 atom s~1 nm~2
k
/10~4 atom s~1 nm~2
0
ER
HDS
HYD
Mo(16)
4.8
38.2
50.3
61.0
67.0
56.7
0.21
0.64
0.91
1.10
1.21
0.88
1.73
6.15
6.22
6.13
6.23
7.30
0.54
5.41
7.99
10.9
13.1
10.4
0.08
0.31
0.48
0.52
0.67
0.51
Co(1)ÈMo(16)
Co(2)ÈMo(16)
Co(3)ÈMo(16)
Co(5)ÈMo(16)
Co(8)ÈMo(16)
content. The maximum increases in the HDS and HYD activ-
The e†ect of temperature and Co/Mo ratio on k for this
series of catalysts was investigated. Within experimental error,
ER
ities occurred on the catalyst with Co/Mo B 0.5 where the
values of k /k
and 8, respectively. The activities for HDS and HYD
decreased slightly with further increase in the cobalt content.
and k /k
were approximately 25
k
at the same temperature did not vary with Co/Mo for all
HDS HDS0
HYD HYD0
ER
promoted catalysts, although it was greater than that for the
unpromoted catalyst. The activation energy of the sulfur
exchange reaction was also determined from an Arrhenius
plot and for all promoted catalysts was found to be 7 ^ 1 kcal
mol~1, less than that of the unpromoted catalyst (10 ^ 1 kcal
mol~1). This indicates that the sulfur on the Co-promoted
catalyst is more mobile than that on the unpromoted catalyst.
This implies that a new active site is formed on CoÈMo/Al O
Discussion
The apparent activation energies of HDS and HYD for all
catalysts were obtained from Arrhenius plots of the reaction
rate constants. There was no signiÐcant di†erence in the
apparent activation energies: 23 ^ 2 and 25 ^ 2 kcal mol~1
for HDS and HYD, respectively. This indicates that there is
essentially no di†erence in the mechanism of reaction on addi-
2
3
with addition of cobalt. Therefore, the promoting e†ect of
cobalt can be explained by the formation of a more active
phase, i.e. the so-called CoMoS phase. On the other hand, the
tion of cobalt. The enhancement in k
tion was 10È25-fold and 4È8-fold, respectively, but the
and k
on Co addi-
HDS rate increased almost linearly with increasing S k in
HDS
HYD
0 ER
virtually the same manner for all promoted catalysts at the
enhancement in k and k did not signiÐcantly increase
same temperature. This indicates that the new active phase
derived from the addition of cobalt is the same for all the
promoted catalysts. Therefore, the increase in catalytic activity
with Co/Mo is due to the increase in the number of new active
sites with increasing Co/Mo.
HDS
HYD
with increasing Co content. The maximum enhancement was
only a factor of ca. 2 when the Co/Mo ratio increased from
0.12 to 0.6. Hence, the promoting e†ect of cobalt is considered
to arise from the formation of a more active phase and the
increase in catalytic activity to be due to increase in the
number of the same active sites.
Since the addition of promoters has been one of the main
routes for the preparation of improved catalysts, considerable
e†orts have been devoted to understanding the origin of pro-
motion and the di†erent types of promotional behaviour. In
recent years, signiÐcant progress has been achieved as it has
become possible to obtain detailed information at the atomic
level about the active sulÐded state of promoted catalysts. In
order to explain the promotional behaviour numerous models
and proposals have been reported.7 Generally, with increasing
promoter concentration, the activity may increase signiÐcantly
and typically passes through a maximum at a Co/Mo or
Ni/Mo molar ratio of 0.3È1.0.17h19 Many studies have shown
that the activity of promoted catalysts is related to the pres-
ence of highly active CoÈMoÈS type structures. TopsÔe et al.
proposed the CoÈMoÈS model for CoÈMo/Al O cata-
For alumina-supported Mo-based catalysts, coordinately
unsaturated sites (CUS) or anion vacancy sites of sulÐded
catalysts are generally accepted to be the active sites for HDS
and HYD.10h15 The recently developed RPTM enables one to
determine the amount of labile sulfur on the working catalyst,
which is considered to represent the amount of sulfur that can
be transformed to vacancies, and the rate constant of sulfur
exchange, which is considered to represent the ease of this
transformation process.16 In order to investigate the corre-
lation of the HDS rate of DBT with S and k , the product
0
ER
of S and k at several temperatures on the series of catalysts
0
was plotted vs. the HDS reaction rate (r ). A good straight
line plot through the origin was obtained. This result can be
HDS
2
3
explained as follows. Under a given reaction condition, S is
lysts.20h22 The reducible Co-oxide species in the
0
correlated with all the sulfur which can be transformed to
CoÈMo/Al O catalyst were well distributed and associated
2
3
vacancies (active sites), and k is correlated with an average
as the CoÈMoÈO phase, which was considered to be the pre-
cursor of the highly active CoÈMoÈS phase. Recently, the exis-
tence of the CoÈMoÈS phase in the sulÐded CoÈMo/Al O
ER
frequency of the transformation of labile sulfur to an active
site. Therefore, S k will be correlated with the HDS rate.
0 ER
2
3
Fig. 5 E†ect of Co/Mo ratio on the conversion of DBT and the yield
of CHB at 260 ¡C
Fig. 6 E†ect of Co/Mo ratio on k
HDS
Mo(16) catalyst at 260 ¡C. (…) k /k
HDS HDS0
and k
relative to those of
HYD
; (=) k /k
HYD HYD0
.
4398
J. Chem. Soc., Faraday T rans., 1997, V ol. 93

View Article Online
catalyst was also veriÐed by the X-ray photoelectron spectro-
scopic study of Lee et al.23
It is commonly accepted that molybdenum interacts with a
hydroxy group on the alumina carrier surface, resulting in the
formation of a monolayer structure.22,24,25 In our previous
paper,9 the apparent activation energies were found to be
approximately the same for a series of sulÐded molybdena
alumina catalysts containing 6È20 wt.% molybdena, indicat-
ing that the mechanism of HDS and the nature of active sites
did not vary with Mo content. Moreover, the amount of labile
sulfur and the HDS activity increased monotonically with
increasing molybdena content up to 16 wt.%, indicating that
monolayer dispersion of molybdenum sulÐde on alumina was
maintained up to 16 wt.%. This agrees with our early work,26
where the monolayer dispersion of molybdenum on alumina
was maintained up to 8È12 wt.% molybdenum, i.e. 12È18
wt.% molybdena. Thus, it is an acceptable hypothesis that the
MoS phase is present as a single-slab structure Ñat on the
2
surface of the alumina, as shown in Scheme 1(a).8
Recently, the di†erence in sulfur behaviour between
Mo/Al O and CoÈMo/Al O has been investigated, and a
2
3
2 3
structure of sulÐded CoÈMo/Al O was proposed, as shown
2
3
in Scheme 1 (structure b).8 In this model, it was considered
that the structure of MoS was a deformed tetrahedral struc-
2
ture and only the locations of the other two weak MowS
Scheme 2 HDS mechanism of DBT on a sulÐded CoÈMo/Al O
2
3
bonds within the layers were changed. The Mo wS or
catalyst. S: 32S, S*: 35S, (K): anion vacancy.
1
d1
Mo wS , and Mo wS or Mo wS bonds were con-
2
d0
1
d5
2
d4
sidered to be the two weak bonds assigned to the van der
Waals type. In this model, one cobalt atom will promote four
ated. When 35S in DBT occupies the vacancy and the
carbonÈsulfur bonds were cleaved in the HDS reaction, 35S
will remain on the catalyst as an MowS species and the old
sulfur atoms in the two neighbouring MoS phases. If this is
2
the case, the promotion e†ect of cobalt on the Mo/Al O
anion vacancy disappears. As
a result, sulfur exchange
2
3
catalyst will occur at a molar ratio of Co/Mo of 0.5, which is
in good agreement with the results obtained in this work. The
rates of both HDS and HYD were promoted with increasing
Co/Mo molar ratio up to ca. 0.5, suggesting that cobalt could
be well dispersed on Mo/Al O and formed a CoÈMoÈS
between the sulfur in DBT and sulfur on the catalyst will
occur. Therefore, the promotion e†ect of cobalt on the
Mo/Al O catalyst is considered to arise when cobalt lowers
2
3
the strength of the sulfurÈmolybdenum bond leading to easier
formation of H S from labile sulfur and subsequent desorp-
2
3
2
phase with the two neighbouring MoS phases at Co/Mo
\0.5.
tion, i.e. the addition of cobalt will enhance the formation and
2
desorption of H S from the labile sulfur, which may be the
2
The promoting e†ect of cobalt was postulated to be due to
increased lability of the sulfur in the so-called CoÈSÈMo
phase where the bond strength of sulfur was considered to be
the weakest.16 This is consistent with the results shown in Fig.
4, where both S and k increased markedly with the addi-
rate-determining step for HDS reactions.
Conclusions
The HDS and HYD activities of Mo/Al O catalysts were
0
ER
2
3
tion of cobalt. Scheme 2 illustrates the mechanism of sulfur
exchange in the HDS reaction on cobalt-promoted
remarkably enhanced on addition of cobalt. The maximum
a
promotional e†ect occurred with a Co/Mo molar ratio of ca.
molybdena/alumina catalyst. It was considered that a portion
of sulfur bonded with Co and Mo atoms, i.e. the sulfur in the
CoÈSÈMo phase, was more labile and was present as SH
groups in the H atmosphere. When H S is formed and subse-
0.5, where the relative increase in k
and k
was ca. 25-
HDS
HYD
and 8-fold, respectively. The apparent activation energies of
HDS and HYD reactions were 23 ^ 2, and 25 ^ 2 kcal
mol~1, respectively, for all catalysts. For the Mo(16) catalyst,
the amount of labile sulfur and the rate constant of sulfur
exchange increased signiÐcantly on addition of cobalt, indicat-
ing that cobalt makes the sulfur more mobile and that the
active phases in the promoted catalysts are di†erent from that
in the unpromoted catalyst. It was suggested, therefore, that
the promoting e†ect of cobalt is attributed to the change in
the nature of the active sites, i.e. the formation of more active
2
2
quently desorbs from the catalyst, an anion vacancy is gener-
sites. On the other hand, S increased linearly with Co/Mo
0
ratio, up to ca. 0.5, whereas k remained almost constant.
ER
Moreover, the enhancement in k
signiÐcantly with increasing Co content. The maximum
enhancement in both k and k was only by a factor of
ca. 2 when the ratio of cobalt to molybdenum increased from
0.12 to 0.6. Therefore, it can be considered that the increase in
catalytic activity with the Co/Mo ratio is due to the increase
in the number of the same active sites.
and k
did not increase
HDS
HYD
HDS
HYD
References
Scheme 1 (a) Tetrahedral structure of MoS on a sulÐded Mo/Al O
2
2 3
catalyst; (b) deformed tetrahedral structure of MoS and structure of
1
2
A. Ishihara and T. Kabe, Ind. Eng. Chem. Res., 1993, 32, 753.
T. Kabe, A. Ishihara and Q. Zhang, Appl. Catal. A, 1993, 97, L1.
2
CowSwMo on a sulÐded CoÈMo/Al O .
2
3
J. Chem. Soc., Faraday T rans., 1997, V ol. 93
4399

View Article Online
3
4
5
6
7
Q. Zhang, W. Qian, A. Ishihara and T. Kabe, J. Jp n ., Petrol. Inst.,
1997, 40, 185.
16 W. Qian, A. Ishihara, G. Wang, T. Tsuzuki, M. Godo and T.
Kabe, J. Catal., 1997, 170, 286.
17 R. Candia, B. S. Clausen and H. TopsÔe, J. Catal., 1982, 77, 564.
18 M. Boudart, J. S. Arrieta and R. Dalla Betta, J. Am. Chem. Soc.,
1983, 105, 6501.
19 H. TopsÔe, B. S. Clausen, N-Y. TopsÔe, J. Hyldtoft and J. K.
NÔrskov, ACS, Petrol Div. Prepr., 1993, 38, 638.
20 H. TopsÔe, B. S. Clausen, R. Candia, C. Wivel and S. MÔrup, J.
Catal., 1981, 68, 433.
21 C. Wivel, R. Candia, B. S. Clausen S. MÔrup and H. TopsÔe, J.
Catal., 1981, 68, 453.
22 N. Y. TopsÔe and H. TopsÔe, J. Catal., 1993, 139, 631; 641.
23 D. K. Lee, H. T. Lee, I. C. Lee, S. K. Park, S. Y. Bse, C. H. Kim
and S. I. Woo, J. Catal., 1996, 159, 219.
24 G. C. A. Schuit and B. C. Gates, AIChE J., 1973, 19, 417.
25 F. E. Massoth, Advances in Catalysis and Related Subjects, Aca-
demic Press, New York, 1978, vol. 27 p. 265.
H. TopsÔe, B. S. Clausen, N-Y. TopsÔe and E. Pedersen, Ind.
Eng. Chem. Fund., 1986, 25, 25.
S. M. A. M. Bouwens, J. A. R. van Veen, D. C. Koningsberger,
V. H. J. de Beer and R. Prins, J. Phys. Chem., 1991, 95, 123.
W. Qian, A. Ishihara, S. Ogawa and T. Kabe, J. Phys. Chem.,
1994, 98, 907.
T. Kabe, W. Qian, S. Ogawa and A. Ishihara, J. Catal., 1993, 143,
239.
T. Kabe, W. Qian and A. Ishihara, J. Catal., 1994, 149, 171.
W. Qian, Q. Zhang, Y. Okoshi, A. Ishihara and T. Kabe, J.
Chem. Soc., Faraday T rans., 1997, 93, 1821.
8
9
10 N. K. Nag, K. S. P. Rao, K. V. R. Chary, B. R. Rao and V. S.
Subrahmanyam, Appl. Catal., 1988, 41, 165.
11 B. M. Reddy, K. V. R. Chary, V. S. Subrahmanyam and N. K.
Nag, J. Chem. Soc., Faraday T rans. 1, 1985, 81, 1655.
12 K. Ramanathan and S. W. Weller, J. Catal., 1985, 95, 249.
13 W. Zmierczak, G. Muralidhar and F. E. Massoth, J. Catal., 1982,
77, 432.
26 T. Kabe, S. Yamada, M. Oba and Y. Miki, Int. Chem. Eng., 1972,
12, 366.
14 J. Valyon and W. K. Hall, J. Catal., 1983, 84, 216.
15 J. Bachelier, J. C. Duchet and D. Cornet, J. Catal., 1984, 87, 283.
Paper 7/05224C; Received 21st July, 1997
4400
J. Chem. Soc., Faraday T rans., 1997, V ol. 93