Formation and catalytic properties of edge-bonded molybdenum sulfide catalysts on TiO2

Article abstract of DOI:10.1006/jcat.2002.3534

The effect of preparation conditions (calcination atmosphere, sulfidation atmosphere, and sulfidation temperature) on the orientation of MoS₂ clusters on TiO₂ supports was studied. Edge-bonded MoS₂ clusters formed when the catalyst was sulfided in a flow of H₂S/N₂ at 573 or 673 K. However, when sulfided in H₂S/N₂ at higher temperatures than 773 K, the edge-bonded MoS₂ clusters transformed to highly aggregated basal-bonded MoS₂ clusters. Catalytic activity tests, using hydrogenation of 1-methylnaphthalene as a model test reaction, revealed that the turnover frequency on the catalyst with edge-bonded MoS₂ clusters prepared by sulfiding at 573 K in H₂S/N₂ was higher than that on the catalyst with basal-bonded MoS₂ clusters prepared by sulfiding in H₂S/H₂.

Full text of DOI:10.1006/jcat.2002.3534

Journal of Catalysis 207, 361–370 (2002)
doi:10.1006/jcat.2002.3534, available online at http://www.idealibrary.com on
Formation and Catalytic Properties of Edge-Bonded Molybdenum
Sulfide Catalysts on TiO2
∗
∗
,1
Yasuhiro Araki, Kosaku Honna, and Hiromichi Shimada†
∗
Tsukuba Branch of Advanced Catalysts Research Laboratory, Petroleum Energy Center, 1-1 Higashi, Tsukuba, Ibaraki 305-8565, Japan;
and †National Institute of Advanced Industrial Science and Technology (AIST), 1-1-1 Higashi, Tsukuba, Ibaraki 305-8561, Japan
Received October 19, 2001; revised December 28, 2001; accepted December 28, 2001
In the late 1980s, Topsøe et al. (4) proposed that the active
The effect of preparation conditions (calcination atmosphere, catalytic sites of Co- or Ni-promoted Mo sulfide catalysts
sulfidation atmosphere, and sulfidation temperature) on the orien- are located on the so-called “Co (Ni)–Mo–S” structure, in
tation of MoS2 clusters on TiO2 supports was studied. Edge-bonded
MoS2 clusters formed when the catalyst was sulfided in a flow of
H2S/N2 at 573 or 673 K. However, when sulfided in H2S/N2 at higher
temperatures than 773 K, the edge-bonded MoS2 clusters trans-
formed to highly aggregated basal-bonded MoS2 clusters. Cata-
lytic sites in industrial catalysts. Therefore, the Co–Mo–S
lytic activity tests, using hydrogenation of 1-methylnaphthalene
as a model test reaction, revealed that the turnover frequency
on the catalyst with edge-bonded MoS2 clusters prepared by sul-
fiding at 573 K in H2S/N2 was higher than that on the cata-
lyst with basal-bonded MoS2 clusters prepared by sulfiding in Co–Mo–S structures depends on the morphology of the
which Co (Ni) atoms are bonded to the edges of MoS2 crys-
tallites. Later, Topsøe et al. (5) suggested that there are dif-
ferent types of “Co–Mo–S” structures and that only some
of these Co–Mo–S structures function as highly active cata-
structures on the support must be clarified in detail and the
catalytic activities of those structures must be determined.
Numerous studies indicate that the catalytic activity of
H2S/H2.
ꢀc 2002 Elsevier Science (USA)
MoS2 clusters on the support, because the layered MoS2
Key Words: hydrodesulfurization (HDS) catalysts; catalyst dis- structure is highly anisotropic. Candia et al. (6) claimed
persion; morphology; orientation.
that there are at least two types of Co–Mo–S structures:
one called “Co–Mo–S(I),” which has relatively strong in-
teraction with the support and is less catalytically active
than the other structure, called “Co–Mo–S(II),” which has
INTRODUCTION
weak interaction with the support (7–12). Single-layered
MoS2 clusters with Co at their edges probably are Co–Mo–
S(I), whereas multilayered MoS2 clusters with Co, except
on the bottom layer, are Co–Mo–S(II). Whitehurst et al.
Alumina-supported Mo sulfide catalysts with Co or Ni
as promoter are widely called hydrodesulfurization (HDS)
catalysts and have long been used for hydrotreatment of
petroleum fractions. Within the past decade, the catalytic
(
13) suggested that, due to steric hindrance, the bottom
performance of HDS catalysts has been significantly im-
proved to satisfy a wide range of requirements. The most
recent requirement is to drastically reduce the sulfur level
in diesel fuels, to at least below 30 ppm, to meet more strin-
gent environmental regulations that will be introduced in
the near future. Such reduction in sulfur (deep desulfuriza-
tion) requires that next-generation HDS catalysts possess
the ability to remove sulfur from hard-to-desulfurize com-
pounds such as 4,6-dimethyldibenzothiophene (1–3). This
ability requires the development of HDS catalysts that have
full dispersion of highly active catalytic sites. In addition,
high hydrogenation activity as well as high HDS activity
arerequiredforsuchdeepdesulfurizationcatalyststoelimi-
nate the steric hindrance that lowers the reactivity of the
above hard-to-desulfurize compounds.
layers of Co–Mo–S structures that have multilayers might
be less active than the other layers. Vrinat et al. (14) re-
ported that in Mo sulfide catalysts without Co promoters,
only the topmost layers of MoS2 clusters on the support are
catalytically active in the HDS reaction of thiophene. Fur-
thermore, Daage et al. (15) claimed that all the edge planes
of MoS2 clusters possess HDS activity, whereas due to steric
hindrance, only the “rim,” i.e., the top and bottom edges, of
unsupported MoS2 clusters possess hydrogenation activity.
These three studies indicate that the morphology of MoS2
clusters, in particular the aspect ratio of the layered struc-
ture (i.e., lateral dimension/thickness), significantly affects
the catalytic performance, irrespective of the presence or
absence of Co- or Ni-promoters.
Based on the above discussion, the catalytic performance
of Co–Mo–S structures depends also on the orientation
1
of the MoS clusters on the support, because the upper
To whom correspondence should be addressed. Fax: +81-298-61-2371.
E-mail: h-shimada@aist.go.jp.
2
edge sites of the “edge-bonded” MoS2 clusters that are
361
0
021-9517/02 $35.00
c 2002 Elsevier Science (USA)
All rights reserved.
ꢀ

3
62
ARAKI, HONNA, AND SHIMADA
of aromatic rings, in which the effect of steric hindrance
is larger than in the HDS reactions. Finally, we clarified
the characteristics of the edge-bonded and basal-bonded
MoS2 clusters needed in the design of highly active catalytic
sites.
EXPERIMENTAL
Catalyst Preparation
The TiO2 support used here was ultrafine anatase-type
particles (Nanophase Technologies Ltd.) with an average
particle diameter of 30 nm, a BET surface area of 50 m /g,
2
FIG. 1. Schematic of the orientation of MoS2 clusters on supports.
and a purity of 99.95%. The Mo sulfide catalysts were pre-
pared by using an equilibrium adsorption method (21, 22).
In the equilibrium adsorption, 10 g of the TiO2 pow-
perpendicular to the support surface (Fig. 1a) have weaker
electronic interaction with the support than do single-
layered MoS2 clusters that are “basal-bonded” (parallel) to
the support (Fig. 1b). In addition, the upper edge sites of the
edge-bonded MoS2 clusters have less steric hindrance than
do either the edge sites of the basal-bonded single-layered
MoS2 clusters (Fig. 1b) or the edge sites of the bottom layers
of the basal-bonded multilayered MoS2 clusters (Fig. 1c).
Until recently, no clear evidence was reported for such
edge-bonded MoS2 clusters on γ -Al2O3, although there
were some discussion about the edge-bonded MoS2 clus-
ters based on transmission electron microscopy (TEM) pic-
tures (16–18). In 1999, using γ -Al2O3 single crystals with
different indices, Sakashita and Yoneda (19) showed that
there is an epitaxial relationship between the orientation
of MoS2 clusters and the surface structures of γ -Al2O3;
namely, when the spacing of MoS2 layers is identical to
the lattice distance of a surface plane of an γ -Al2O3 crys-
tal, MoS2 clusters are edge-bonded to the surface. As a re-
sult, edge-bonded clusters form on (100) γ -Al2O3, whereas
basal-bonded MoS2 clusters form on other planes, such
as (111) and (110), or on amorphous planes of γ -Al2O3.
In a subsequent study, Sakashita et al. (20) confirmed an
3
der was mixed with 400 cm of an aqueous solution of
(
NH4)6Mo7O24 (0.007M)andkeptat323Kfor24h. ThepH
of the solution was kept constant at 2.0 by adding HNO3.
The pH value of 2.0 for equilibrium adsorption was cho-
sen to achieve high loading of molybdate on the TiO2 sup-
port but to not exceed too much that of the monolayer
coverage.
After filtration and rinsing with pure water, the solid was
dried in air at 298 K for 16 h and then at 323 K for 3 h. The
calcination was done at 723 K for 3 h in a flow of either
nitrogen or dry air to clarify the effect of the inert-gas at-
mosphere that was used in the previous study (20) on the
morphology and orientation of MoS2 clusters on the sup-
port. Sulfidation of the resulting calcined catalysts was done
for 2 h at various temperatures, ranging from 573 to 873 K,
in a flow of either 5% H2S/H2 or 5% H2S/N2. Induced cou-
pled plasma emission spectrometry analysis indicated that
the Mo loading was 4.6 wt% as metal.
Catalyst Characterization
Observation of the catalysts after sulfidation was carried
epitaxial relationship for the orientation of MoS2 clusters out using transmission electron microscopy (TEM) with a
on anatase-type TiO2 powders; namely, edge-bonded MoS2 Hitachi H-800 operated at an accelerated voltage of 200 kV.
clusters form on the spherical anatase powders that mainly Each catalyst sample was ground into powder (by using a
expose the (001) plane.
In our current study, we determined the preparation con- copper grid with holey carbon films in n-heptane.
ditions (calcination atmosphere, sulfidation atmosphere, X-ray photoelectron spectroscopy (XPS) spectra were
mortar and a pestle) and then ultrasonically dispersed on a
and sulfidation temperature) that enhance the formation of obtained by using a PHI 5500 photoelectron spectrometer
edge-bonded MoS2 clusters on an anatase-type TiO2 sup- with monochromatic Al Kα excitation (1486.6 eV, 150 W).
port. First, we clarified the effects of calcination, either in The energy scale of the spectrometer was calibrated using
nitrogen or dry air, and the sulfidation conditions, using ei- the Au 4f5/2 (84.0 eV) line of a pure Au plate and the Cu
ther H2S/N2 or H2S/H2 at temperatures ranging from 573 to 2p3/2 (932.4 eV) line of a sputtered Cu plate. All peak ener-
8
73 K, on the orientation of MoS2 clusters on the support. gies were corrected using the C 1s line of the adventitious
Then, we compared the catalytic activities of these prepared carbon at 285.0 eV. As a reference for the assignment of
MoS2-supported catalysts using a model test reaction. We S 2p lines, amorphous MoS3 was prepared by thermal de-
chose the hydrogenation of 1-methylnaphthalene (1-MN) composition of ammonium-tetrathiomolybdate at 523 K for
as a model test reaction, because more prominent effects 1 h in a flow of N2 according to a method in the literature
of edge-bonded clusters are expected in the hydrogenation (23, 24).

CATALYTIC PROPERTIES OF EDGE-BONDED MoS2 CATALYSTS
363
The dispersion of MoS2 on the support was measured by
Figure 3 shows representative TEM images of Mo/TiO2
using a NO chemisorption method. After being sulfided in catalysts sulfided in H2S/N2 at 573, 673, or 773 K after cal-
a stream of either 5% H2S/H2 or 5% H2S/N2, about 0.2 g of cination in air. The catalyst sulfided at 573 K (Fig. 3a) had
the catalyst was treated with a flow of H2 for 1 h at 603 K. many edge-bonded MoS2 clusters that were less than 2 nm
3
Then, 10% NO/He pulses, each having a volume of 2.1 cm , long (indicated by a circle) but had no basal-bonded MoS2
were introduced to the catalyst at 303 K. The amount of clusters. The number of Mo atoms estimated from the ob-
NO at the exit of the reactor was monitored using a thermal served MoS2 clusters in Fig. 3a was less than that expected
conductivity detector so that saturation of the adsorption from the Mo loading of 4.6%. This indicates high disper-
of NO on the catalyst could be detected.
sion of MoS2 clusters or insufficient sulfidation, similar to
the results for the catalyst sulfided in H2S/H2. The catalyst
sulfided at 673 K (Fig. 3b) had a relatively large number of
edge-bonded MoS2 clusters (indicated by a circle), result-
ing from the growth of small MoS2 clusters that were not
detected in the image, and a small number of basal-bonded
MoS2 clusters (indicated by a dotted circle). The catalyst
sulfided at 773 K (Fig. 3c) mostly had large (>5 nm), mul-
tilayered basal-bonded MoS2 clusters on the support (indi-
cated by a dotted circle) but had relatively few edge-bonded
MoS2 clusters.
Figure 4 shows representative TEM images of Mo/TiO2
catalysts sulfided in either H2S/H2 or H2S/N2 after calcina-
tion in N2. The catalyst sulfided at 573 K in H2S/N2 (Fig. 4a)
had edge-bonded MoS2 clusters whose lengths were longer
than those for the catalyst calcined in air (Fig. 3a). Com-
parison of the images of the catalysts calcined in N2 with
those of the corresponding catalysts calcined in air reveals
the same trends; namely, the calcination atmosphere did
not affect the orientation of the MoS2 clusters, whereas the
average length of the MoS2 clusters was longer when the
catalysts were calcined in N2.
Catalytic Activity Test
Catalytic activity was evaluated by using hydrogenation
of 1-MN as a model test reaction. Before use, 1-MN was
purified by column chromatography to remove N- and S-
containing compounds in the reagent. The reactions were
carried out at 603 K for 1 h in a microautoclave (inner vol-
3
3
ume of 35 cm ) containing 10 cm of the feed (25 wt%
-MN with the balance being tetradecane) and hydrogen
1
with an initial pressure of 6 MPa at 298 K. The feed and
product were analyzed by using gas chromatography with a
HP Ultra #1 capillary column. The reaction rate constants
were calculated using the conversion data obtained for dif-
ferent amounts of the catalyst, ranging from 0.05 to 0.5 g,
and assuming a pseudo-first-order kinetics. To apply this as-
sumption, the reaction conditions were determined so that
the total conversion would not exceed 20%. Under these
conditions, decreases in the hydrogen partial pressure dur-
ing the reactions were negligible. The products obtained
were only 1- and 5-methyltetralin.
The TiO2 support used in the present study was spherical
with a nonporous structure. The TEM image presenting a
whole particle, as shown in Figs. 2–4, gives the complete
image of catalyst particles. Thus, we conclude that these
RESULTS AND DISCUSSION
Catalyst Characterization
Figure 2 shows representative TEM images of Mo/TiO2 TEM images are representative of the MoS2 structures in
catalysts sulfided in H2S/H2 at 573, 673, or 773 K after calci- the catalysts.
nation in air. The catalysts sulfided at 573 and 673 K (Figs. 2a
Table 1 summarizes the above TEM observation results,
and 2b) had small numbers of MoS2 clusters as a monolayer showing that the orientation of MoS2 clusters on the sup-
or as two layers (indicated by arrows in Fig. 2a). All of the port depended on the sulfidation conditions. Sulfidation of
layers were parallel to the surface of TiO2, indicating basal- the catalyst in H2S/N2 enhanced the formation of edge-
bonding on the support. Because the Mo loading was 4.6% bonded MoS2 clusters; however, these clusters changed
(
see Section Catalyst Preparation above), the major part theirorientationtobasal-bondedwhensulfidedatatemper-
of the Mo species was not detected in these images. This ature higher than 673 K. Sulfidation of the catalyst in H2S/
can be attributed either to high dispersion of MoS2 clusters H2 resulted in highly dispersed basal-bonded MoS2 clusters.
that contain fewer than seven Mo atoms per layer (25) or Increasing the sulfidation temperature in H2S/H2 caused
to insufficient sulfidation, particularly in the catalysts sul- aggregation of MoS2 clusters, but less than the aggregation
fided at 573 K. The catalyst sulfided at 773 K (Fig. 2c) had caused when the catalysts were sulfided in H2S/N2.
many MoS2 clusters with multilayered structures (indicated
Figure 5 shows the effect of sulfidation conditions on
by an arrow), indicating that aggregation and stacking of the Mo 3d XPS spectra of MoS2/TiO2 catalysts calcined in
MoS2 layers was proceeded by higher-temperature sulfida- air. As reported in numerous studies(26–30), when the de-
tion. Note that Mo loading of 4.6% as metal corresponds to gree of sulfidation was increased, two features, at 232.5 and
2
6+
5
.8 Mo atom/nm that is slightly higher than that for mono- 235.5 eV, which are assigned to Mo features (Mo 3d5/2
2
layer coverage (5.0 Mo atom/nm ) but not so high that it and Mo 3d3/2) from oxide species, shifted to 228.5 and
forms crystalline species at the oxide stage, as described in 232.0 eV, which are assigned to Mo features from sulfide
a previous paper (22). species. Simultaneously, S 2s features appeared at 226.0 eV.
4
+

3
64
ARAKI, HONNA, AND SHIMADA
FIG. 2. Transmission electron microscopy (TEM) images of MoS2/TiO2 catalysts calcined in air and then sulfided in H2S/H2 at (a) 573 K (solid
arrow indicates monolayered MoS2 cluster and dotted arrow indicates two-layered MoS2 cluster), (b) 673 K, and (c) 773 K (solid arrow indicates
multilayered MoS2 cluster). (×1,000,000.)
Asymmetric features observed in the spectra of the cata- and S 2s. The curve fitting for each spectrum was done by us-
lysts sulfided at low temperatures, typically those observed ing a least-squares method (included in the software pack-
for the catalyst sulfided in H2S/H2 at 573 K (Fig. 5b), indi- age that accompanied the spectrometer, PHI 5500) with the
cate the presence of Mo5 species, likely oxisulfide, in the following restrictions:
+
transition stage from Mo6 to Mo , as described in pre-
vious papers (28–30). In the Mo 3d spectra of MoS2/TiO2
catalysts calcined in N2 (not shown here), changes similar
to those of the MoS2/TiO2 catalysts calcined in air (Fig. 5)
were observed, although the line widths of the features for
the catalysts calcined in N2 were slightly larger.
+
4+
1. Peak area ratio of Mo 3d5/2/ Mo 3d3/2 was fixed at
3/2.
2. Peak energy difference between Mo 3d5/2 and Mo
3d3/2 was assumed constant.
6+
Based on this curve fitting, the relative ratios of Mo
,
5
+
4+
To quantitatively discuss the changes in Mo species, each Mo , and Mo in the MoS /TiO catalysts calcined in air
2
2
spectrum was deconvoluted into seven peaks assigned to were calculated and are shown in Fig. 6. As expected from
6
5
+
6+
5+
5+
4+
4+
Mo 3d , Mo 3d , Mo 3d , Mo 3d , Mo 3d , Mo 3d
,
2 2
Fig. 5, sulfidation in H S/H at 573 K yielded nearly 80%
/2
3/2
5/2
3/2
5/2
3/2

CATALYTIC PROPERTIES OF EDGE-BONDED MoS2 CATALYSTS
365
FIG. 3. Transmission electron microscopy (TEM) images of MoS2/TiO2 catalysts calcined in air and then sulfided in H2S/N2 at (a) 573 K (solid
circle indicates edge-bonded MoS2 cluster), (b) 673 K (solid circle indicates edge-bonded MoS2 cluster and dotted circle indicates basal-bonded MoS2
cluster), and (c) 773 K (dotted circle indicates basal-bonded MoS2 cluster). (×1,000,000.)
Mo4 , with the remaining assigned to Mo . Sulfidation sulfided in H2S/H2 (Figs. 7a–7c), changes in the S 2p spectra
+
5+
of the catalyst in H2S/N2 at 573 or 673 K resulted in higher corresponded to changes in the Mo 3d XPS spectra; namely,
4
+
ratios of Mo than sulfidation in H2S/H2, whereas at 773 K, sulfidation at 573 K yielded mostly MoS2-like species and
the ratio remained unaffected. These results indicate that the characteristics of the spectra approached those of pure
Mo oxide on TiO2 was sulfided more rapidly in H2S/N2 than MoS2 with increasing sulfidation temperature.
in H2S/H2.
The spectrum for the catalyst sulfided in H2S/N2 at 573 K
Figure 7 shows the effect of sulfidation conditions on the (Fig. 7d) showed features different from those for the cata-
S 2p XPS spectra of MoS2/TiO2 catalysts calcined in air. The lyst sulfided in H2S/H2 (Figs. 7a–7c). Figure 7 also shows
doublets with a main peak at 161.8 eV observed in the sul- the spectra of amorphous MoS3 and elemental sulfur as
fided catalysts were consistent with that of pure MoS2. The references. Evidently, the spectrum of the catalyst sul-
major peak was assigned to S 2p3/2 of S2 in MoS2, and the fided in H2S/N2 at 573 K corresponded to the spectrum
−
minor one assigned to S 2p1/2. In the spectra of the catalysts of MoS3. According to previous studies (23, 31), the S 2p3/2

3
66
ARAKI, HONNA, AND SHIMADA
FIG. 4. Transmission electron microscopy (TEM) images of MoS2/TiO2 catalysts calcined in N2 and then sulfided in (a) H2S/N2 at 573 K (circle
indicates edge-bonded MoS2 cluster), (b) H2S/N2 at 673 K (circle indicates edge-bonded MoS2 cluster; this same image appeared in a previous paper
(20)), (c) H2S/H2 at 673 K (circle indicates basal-bonded MoS2 cluster), and (d) H2S/N2 at 873 K (circle indicates basal-bonded MoS2 cluster).
(×1,000,000.)
2
−
binding energy of bridging S in MoS3 is 162.9 ± 0.2 eV, for the Mo 3d spectra (see above) except that the peak area
2
2−
2−
those of terminal S2 and S in MoS3 are 161.6 ± 0.2 eV, ratio of S 2p3/2/S 2p1/2 was fixed at 2.
0
that of S in elemental sulfur is 164.0 eV, and that of bridging
S
Figure 8 shows the effect of sulfidation temperature on
2
−
in MoS2 is 161.8 eV. Based on these binding energies, each of these fractions. The catalyst sulfided in H2S/N2 at
the spectra in Fig. 7 were deconvoluted into the following 573 K contained 60% of the S (ME) assigned to bridging
2
2
−
three doublets (S 2p3/2 and S 2p1/2): S (HE) with S 2p3/2
S
inMoS3. Thisfractiondecreasedwithincreasingsulfida-
at 164.0 ± 0.2 eV, S(ME) with S 2p3/2 at 162.9 ± 0.2 eV, tion temperature and finally disappeared at 773 K. Figure 8
and S (LE) with S 2p3/2 at 161.6 ± 0.2 eV. Thus, S (HE) indicates that sulfidation of the catalyst in H2S/H2 also
is attributed to elemental sulfur, S (ME) to MoS3, and S yielded MoS3, although the fraction was smaller than that
(
LE) to MoS2 and MoS3. The fraction of each doublet was in the catalyst sulfided in H2S/N2. The observation of
calculated by a curve fitting using the same procedure used MoS3 in the catalyst sulfided in H2S/H2 is consistent with

CATALYTIC PROPERTIES OF EDGE-BONDED MoS2 CATALYSTS
367
TABLE 1
Formation Characteristics of MoS2/TiO2 Catalysts Calcined and Sulfided under Various Conditions (Revealed by TEM)
Calcination
atmosphere
Sulfidation
atmosphere
Sulfidation
temperature (K)
Max. cluster
length (nm)
Ave. cluster
length (nm)
Number
of layers
Ave. number
of layers
Orientation
Air
H2S/H2
H2S/N2
573
Few MoS2 clusters were observed.
Few MoS2 clusters were observed.
Basal bonding
Basal bonding
Basal bonding
Edge bonding
Edge + basal
6
7
73
73
12
5.1
1–5
2.1
573
Few MoS2 clusters were observed.
673
12 (Edge)
0 (Basal)
60 (Basal)
5.7 (Edge)
8.0 (Basal)
3.5 (Edge)
3–20 (Edge)
1–5 (Basal)
3–9 (Edge)
1–9 (Basal)
9.3 (Edge)
2.0 (Basal)
6.3 (Edge)
3.0 (Basal)
2
773
Edge + basal
24.9 (Basal)
N2
H2S/H2
H2S/N2
573
Few MoS2 clusters were observed.
Few MoS2 clusters were observed.
Basal bonding
Basal bonding
Basal bonding
Edge bonding
Edge + basal
6
8
73
73
40
7
15.5
3
1–10
1–20
3.2
4.8
573
673
10 (Edge)
0 (Basal)
6.2 (Edge)
9.7 (Basal)
33.2
5–20 (Edge)
1–5 (Basal)
1–15
10.4 (Edge)
2.1 (Basal)
3.1
2
873
100
Basal bonding
Note. The average length and average number of MoS2 layers were obtained by averaging all the MoS2 layers observed in three pictures with a
magnification of 1,000,000 (15 cm × 20 cm).
previous results (32, 33) that report the formation of MoS3 into MoS2 at higher temperature. Sulfidation of the cata-
by sulfidation of the catalyst in H2S/H2 at low tempera- lyst in H2S/N2 is more likely to maintain the amorphous
ture. No spectra contained S (HE) from elemental sulfur; MoS3 structures at higher temperature than does sulfida-
this is also consistent with previous results (27) that report tion in H2S/H2. Note that the exact fraction of MoS3 is not
the absence of the formation of elemental sulfur during the the same as the fraction in Fig. 8, because S (LE) contains
sulfiding of Mo oxide catalysts.
contributions from both MoS2 and MoS3.
Based on the above discussion, sulfidation of oxide cata-
The trends observed in the S 2p spectra of MoS2/TiO2
lysts at first yielded MoS3, which subsequently transformed catalysts calcined in N2 (not shown here) were similar to
those observed in the S 2p XPS spectra of MoS2/TiO2 cata-
lysts calcined in air (Fig. 7). The line width of each feature
FIG. 5. Mo 3d X-ray photoelectron spectra of MoS2/TiO2 cata-
lysts calcined in air (a) before sulfidation, and then after sulfidation in
(
b) H2S/H2 at 573 K, (c) H2S/H2 at 673 K, (d) H2S/H2 at 773 K, (e) H2S/N2
FIG. 6. Fraction of Mo4 , Mo , and Mo in MoS2/TiO2 catalysts
calcined in air.
+
5+
6+
at 573 K, (f) H2S/N2 at 673 K, and (g) H2S/N2 at 773 K. (h) That of pure
MoS2.

3
68
ARAKI, HONNA, AND SHIMADA
FIG. 8. Fraction of S (ME) and S (HE) in MoS2/TiO2 catalysts cal-
cined in air. Closed triangle, S (ME) from MoS3 in the catalyst sulfided by
H2S/N2; open triangle, S (LE) from MoS2 and MoS3 in the catalyst sul-
fided by H2S/N2; closed circle, S (ME) from MoS3 in the catalyst sulfided
by H2S/H2; and open circle, S (LE) from MoS2 and MoS3 in the catalyst
sulfided by H2S/H2.
dispersion of MoS2 clusters was higher in the catalyst sul-
fided in H2S/H2 than in the catalyst sulfided in H2S/N2.
In both sulfidation atmospheres, the dispersion decreased
by increasing the sulfidation temperature. The NO/Mo val-
ues obtained for the present catalysts sulfided at 673 K
(
0
0.096 mol/mol for the catalyst sulfided in H2S/N2 and
.145 mol/mol for the catalyst sulfided in H2S/H2) were al-
most the same or higher than that obtained for a laboratory-
prepared Mo/Al2O3 catalyst sulfided at 673 K in H2S/H2
(
0.095 mol/mol). Based on the TEM results, XPS results,
and NO uptake measurements, the MoS2 clusters in the
catalysts calcined in air and sulfided in H2S/H2 at low tem-
perature were highly dispersed and most were undetectable
in the TEM images in Fig. 2.
The results of catalyst characterization are summarized
as follows.
1
. Sulfidation of the catalyst in H2S/N2 at the low tem-
FIG. 7. S 2p X-ray photoelectron spectra of MoS2/TiO2 catalysts cal-
cined in air and then sulfided in (a) H2S/H2 at 573 K, (b) H2S/H2 at 673 K,
perature of 573 K yielded edge-bonded MoS2 clusters. With
(
(
c) H2S/H2 at 773 K, (d) H2S/N2 at 573 K, (e) H2S/N2 at 673 K, and
f) H2S/N2 at 773 K; that of (g) pure MoS2, (h) MoS3, and (i) elemental S.
in the spectrum for MoS2/TiO2 catalysts calcined in N2 was
larger than that of the corresponding feature of these cata-
lysts calcined in air, similar to the correspondence between
the Mo 3d spectra. Based on the XPS and TEM results, for-
mation of edge-bonded MoS2 clusters was likely related to
the formation of MoS3 during sulfidation. A short lifetime
forMoS3 duringsulfidationinH2S/H2 presumablyhindered
the formation of edge-bonded MoS2 clusters on TiO2. Fur-
thermore, H2 in H2S/H2 possibly led to surface hydroxyl
groups on TiO2, which yielded basal-bonded MoS2 clusters
when sulfidation was done in H2S/H2.
Figure 9 shows the effect of sulfidation conditions on
FIG. 9. NO uptake on the MoS2/TiO2 catalysts calcined in air and
the NO uptake by MoS2/TiO2 catalysts calcined in air. The then sulfided under different atmospheres and temperatures.

CATALYTIC PROPERTIES OF EDGE-BONDED MoS2 CATALYSTS
369
increasing sulfidation temperatures, however, the edge- The catalysts calcined in air showed higher activity than
bonded clusters transformed to highly aggregated basal- the corresponding catalysts calcined in N2. Among the cata-
bonded clusters.
lysts sulfided in H2S/N2, the catalytic activity gradually de-
2
. Sulfidation of the catalyst in H2S/H2 yielded highly creased with increasing sulfidation temperatures in each
dispersed MoS2 clusters. Although most of these clusters series of the catalysts calcined in air or N2. In contrast,
were undetectable in the TEM images, some of those among the catalysts sulfided in H2S/H2, the catalytic activ-
that were detected showed preferential formation of basal- ity increased with increasing sulfidation temperatures up to
bonded MoS2 clusters.
773 K. The results show that the catalyst calcined in air and
3
. Calcination atmosphere did not affect the orienta- sulfided in H2S/N2 at 573 K yielded the highest activity.
tion of MoS2 clusters on the support, either edge-bonded
Among the catalysts sulfided in H2S/N2, the change in
or basal-bonded, although calcination in air yielded higher the catalytic activity (Fig. 10) corresponded with that in
dispersion of MoS2 clusters after sulfidation than did calci- the dispersion (Fig. 9). This indicates that catalytic activ-
nation in N2.
. More than 80% of the Mo oxide transformed into Mo lytic sites and that the turnover frequency (TOF) on the
sulfide by sulfidation at 573 K. The effect of the remaining edge-bonded MoS2 clusters was relatively independent of
ity decreased with a decrease in the number of active cata-
4
6
+
Mo species on the catalytic activity tests (discussed in the the dispersion. In contrast, the catalytic activity of the cata-
next section) can be assumed to be relatively small. The lyst sulfided in H2S/H2 increased with increasing sulfidation
sulfided Mo species gradually aggregated with increasing temperatures, despite a decrease in the dispersion (Fig. 9).
sulfidation temperatures.
The TOF on the basal-bonded MoS2 clusters increased with
increasing sulfidation temperatures. This increase is pre-
sumably due to the decrease in the electronic interaction
between the basal-bonded MoS2 clusters and TiO2 sup-
port, corresponding to the transformation of single-layered
MoS2 clusters to multilayered MoS2 clusters shown in Fig. 1.
This is in good agreement with the increase in the cata-
lytic activity of Co–Mo sulfide catalysts from Co–Mo–S(I)
to Co–Mo–S(II), as reported in the literature (7–12). Fur-
thermore, the TOF on the edge-bonded MoS2 clusters was
higher than that on the basal-bonded MoS2 clusters, be-
cause the number of active catalytic sites estimated by us-
ing the NO uptake for the edge-bonded clusters was smaller
than that for the edge-bonded clusters at each sulfidation
In a previous paper (20), we showed that edge-bonded
MoS2 clusters form on a TiO2 support when the catalyst
is calcined in N2 and then sulfided in H2S/N2. Our current
results indicate that sulfiding the catalyst in H2S/N2 was es-
sential for the formation of edge-bonded clusters, whereas
calcining in N2 had no effect on such formation. Instead,
although calcining in N2 resulted in the formation of aggre-
gated structures, this aggregation could have increased the
possibility of detecting the edge-bonded MoS2 clusters in
the TEM observation in our previous study.
Catalytic Activity
Figure 10 compares the hydrogenation activities of the temperature (Fig. 9).
MoS2/TiO2 catalysts prepared under different conditions.
2 2
Sulfidation of the catalyst in H S/N at a temperature
higher than 673 K, however, yielded rapid aggregation of
MoS2 clusters. As a result, the activity of the catalyst sul-
fided in H2S/H2 was superior to that of the catalyst sulfided
in H2S/N2 when the sulfidation temperature was 673 K.
Edge-bonded MoS2 clusters probably aggregate more eas-
ily than do basal-bonded MoS2 clusters, because edge-
bonded clusters have weaker electronic interaction with the
support.
Numerous studies have reported that the catalytic ac-
tivities of TiO2-supported MoS2 catalysts are superior to
those of Al2O3-supported catalysts (34–39). Consequently,
numerous discussions on the possible reasons for this su-
periority of TiO2-supported catalysts have been reported,
although no definite conclusion has been reached. One pos-
sible reason for the superior activity is that the formation
of edge-bonded MoS2 clusters might contribute to the high
activity of TiO2-supported catalysts.
Our current results reveal, however, that edge-bonded
clusters are not as stable as basal-bonded clusters. In addi-
tion, populating the edge sites of a MoS2 cluster with Co
is essential for practical preparation of catalysts. Further
FIG. 10. Catalytic activity of MoS2/TiO2 catalysts calcined and sul-
fided under different calcination atmospheres, sulfidation atmospheres,
and sulfidation temperatures.

3
70
ARAKI, HONNA, AND SHIMADA
studies are needed to apply the edge-bonded clusters dis- 12. Louwers, S. P. A., Craje, M. W. J., van der Kraan, A. M., Geantet, C.,
and Prins, R., J. Catal. 144, 579 (1993).
cussed in our study to industrial catalysts.
1
1
1
3. Whitehurst, D. D., Isoda, T., and Mochida, I., Adv. Catal. 42, 345
1998).
(
CONCLUSION
4. Vrinat, M., Breysse, M., Geantet, C., Ramirez, J., and Massoth, F.,
Catal. Lett. 26, 25 (1994).
5. Daage, M., and Chianelli, R. R., J. Catal. 149, 414 (1994).
Formation and catalytic properties of edge-bonded MoS2
clusters on TiO2 supports were studied. The following con- 16. Pratt, K. C., Sanders, J. V., and Christov, V., J. Catal. 124, 416
1990).
7. Hayden, T. F., and Dumesic, J. A., J. Catal. 103, 366 (1987).
(
clusions were obtained.
1
1
. Edge-bonded MoS2 clusters formed on TiO2 sup- 18. Stockmann, R. M., Zandbergen, H. W., van Langeveld, A. D., and
Moulijn, J. A., J. Mol. Catal. A 102, 147 (1995).
ports by sulfiding the oxide precursor in H2S/N2 at low tem-
peratures, such as 573 and 673 K. The edge-bonded MoS2
clusters transformed to aggregated basal-bonded clusters
by sulfidation at high temperatures, such as 773 K.
1
2
9. Sakashita, Y., and Yoneda, T., J. Catal. 185, 487 (1999).
0. Sakashita, Y., Araki, Y., Honna, K., and Shimada, H., Appl. Catal. A
197, 247 (2000).
21. Wang, L., and Hall, K., J. Catal. 77, 232 (1982).
2
. In the hydrogenation of 1-methylnaphthalene, edge- 22. Kim, D. S., Kurusu, Y., Wachs, I. E., Hardcastle, F. D., and Segawa, K.,
J. Catal. 120, 325 (1989).
bonded MoS2 clusters on TiO2 showed a higher turnover
frequency than did basal-bonded MoS2 clusters on TiO2.
23. Weber, Th., Muijsers, J. C., and Niemantsverdriet, J. W., J. Phys.
Chem. 99, 9194 (1995).
24. Chang, C. H., and Chan, S. S., J. Catal. 72, 139 (1981).
REFERENCES
25. Eijsbouts, S., and Heinerman, J. J. L., Appl. Catal. A 105, 53 (1993).
6. Brinen, J. S., and Armstrong, W. D., J. Catal. 54, 57 (1978).
27. Li, C. P., and Hercules, D. M., J. Phys. Chem. 88, 456 (1984).
Kwart, H., Preprint of Am. Chem. Soc. Div. Petroleum Chem., 22, 28. Makovsky, L. E., Stencel, J. M., Brown, F. R., Tischer, R. E., and
2
1
2
. Houalla, M., Broderick, D., deBeer, V. J. H., Gates, B. C., and
9
41 (1977).
Pollack, S. S., J. Catal. 89, 334 (1984).
29. Jepsen, J. S., and Rase, H. F., Ind. Eng. Chem. Prod. Res. Dev. 20, 467
(1981).
30. Jong, A. M., Borg, H. J., van Ijzendoorn, L. J., Soudant, V. G. F. M.,
de Beer, V. H. J., van Veen, J. A. R., and Niemantsverdriet, J. W.,
J. Phys. Chem. 97, 6477 (1993).
. Katti, S. S., Westerman, D. W. B., Gates, B. C., Youngless, T., and
Petrakis, L., Ind. Eng. Chem. Process Des. Dev. 23, 773 (1984).
. Kabe, T., Ishihara, A., and Zang, Q., Appl. Catal. A 97, L1 (1993).
. Topsøe, H., Clausen, B. S., Candia, R., Wivel, C., and Mørup, S.,
J. Catal. 68, 433 (1981).
3
4
5
. Topsøe, H., Clausen, B. S., and Massoth, F. E., in “Hydrotreating Cata-
lysis” (J. R. Anderson and M. Boudart, Eds.), Vol. 11, p. 162. Springer-
Verlag, Berlin/Heidelberg, 1996.
31. Muijsers, J. C., Weber, Th., van Hardeveld, R. M., Zandbergen, H. W.,
and Niemantsverdriet, J. W., J. Catal. 157, 698 (1995).
32. Payen, E., Kasztelan, S., Houssenbay, S., Szymanski, R., and
Grimblot, J., J. Phys. Chem. 93, 6501 (1989).
33. de Boer, M., van Dillen, A. J., Koningsberger, D. C., and Geus, J. W.,
J. Phys. Chem. 98, 7862 (1994).
34. Matsuda, S., and Kato, A., Appl. Catal. 8, 149 (1983).
35. Ng, K. Y. S., and Gulari, E., J. Catal. 95, 33 (1985).
36. Shimada, H., Sato, T., Yoshimura, Y., Hiraishi, J., and Nishijima, A.,
J. Catal. 110, 275 (1988).
6
. Candia, R., Sørensen, O., Villadsen, J., Topsøe, N., Clausen, B. S., and
Topsøe, H., Bull. Soc. Chim. Belg. 93, 763 (1984).
7
. Bouwens, S. M. A. M., van Zon, F. B. M., van Dijk, M. P., van der
Kraan, A. M., de Beer, V. H. J., van Veen, J. A. R., and Koningsberger,
D. C., J. Catal. 146, 375 (1994).
8
. Topsøe, H., Clausen, B. S., Topsøe, N.-Y., and Zeuthen, P., Stud. Surf.
Sci. Catal. 53, 77 (1990).
9
. Ramirez, J., Fuentes, S., Diaz, G., Vrinat, M., Breysse, M., and
Lacroix, M., Appl. Catal. 52, 211 (1989).
37. Ramirez, J., Fuentes, S., Diaz, G., Vrinat, M., Breysse, M., and
Lacroix, M., Appl. Catal. 52, 211 (1989).
1
1
0. Eijsbouts, S., Appl. Catal. A 158, 53 (1997).
1. Brown, V. M., Louwers, S. P. A., and Prins, R., Catal. Today 10, 345
38. Okamoto, Y., Maezawa, A., and Imanaka, T., J. Catal. 120, 29
(1989).
(1991).
39. Luck, F., Bull. Chim. Soc. Belg. 100, 781 (1991).