Platinum-supporting hollandite-type vanadium-chromium mixed oxides as efficient heterogeneous catalysts for deoxygenation of sulfoxides under atmospheric H2 pressure

Article abstract of DOI:10.1039/c7cy00547d

Deoxygenation of sulfoxides to their corresponding sulfides is an important reaction, and the development of efficient heterogeneous catalysts that can utilize molecular hydrogen (H₂) as a reducing agent is highly desired. In this study, we successfully developed a new concept in heterogeneous catalysts, platinum-supporting hollandite-type vanadium-chromium mixed oxide catalysts (Pt/V_(1-x)Cr_x-Hol, x: Cr/(V + Cr) molar ratio of the precursor solution), for sulfoxide deoxygenation with H₂. Hollandite-type vanadium-chromium mixed oxides (V_(1-x)Cr_x-Hol) that did not support platinum could stoichiometrically deoxygenate methyl phenyl sulfoxide to methyl phenyl sulfide, and the V³⁺ species in V_(1-x)Cr_x-Hol were oxidized to V⁴⁺ after deoxygenation. However, the V⁴⁺ species could not be re-reduced with H₂. By supporting platinum nanoparticles on V_(1-x)Cr_x-Hol, the re-reduction of the oxidized vanadium species with H₂ became possible. Thus, in the presence of Pt/V_(1-x)Cr_x-Hol, catalytic sulfoxide deoxygenation with H₂ as the reducing agent can be realized; the deoxygenation occurs on V_(1-x)Cr_x-Hol, and the role of platinum is to re-reduce the oxidized vanadium species with H₂. In the presence of Pt/V_0.7Cr_0.3-Hol, various types of structurally diverse sulfoxides could be selectively converted into their corresponding sulfides under atmospheric H₂ pressure (1 atm). In addition, the present system was applicable to the deoxygenation of pyridine N-oxides to their corresponding pyridines. The observed catalysis was truly heterogeneous, and the Pt/V_0.7Cr_0.3-Hol catalyst could be reused for sulfoxide deoxygenation, although the performance was reduced.

Full text of DOI:10.1039/c7cy00547d

Catalysis
Science &
Technology
View Article Online
View Journal
PAPER
Platinum-supporting hollandite-type vanadium–
chromium mixed oxides as efficient
Cite this: DOI: 10.1039/c7cy00547d
heterogeneous catalysts for deoxygenation of
sulfoxides under atmospheric H2 pressure†
Tsubasa Uematsu, Yoshiyuki Ogasawara, Kosuke Suzuki,
*
*
Kazuya Yamaguchi
and Noritaka Mizuno
Deoxygenation of sulfoxides to their corresponding sulfides is an important reaction, and the development
of efficient heterogeneous catalysts that can utilize molecular hydrogen (H2) as a reducing agent is highly
desired. In this study, we successfully developed a new concept in heterogeneous catalysts, platinum-
supporting hollandite-type vanadium–chromium mixed oxide catalysts (Pt/V(1−x)Crx-Hol, x: Cr/(V + Cr) mo-
lar ratio of the precursor solution), for sulfoxide deoxygenation with H2. Hollandite-type vanadium–chro-
mium mixed oxides (V(1−x)Crx-Hol) that did not support platinum could stoichiometrically deoxygenate
methyl phenyl sulfoxide to methyl phenyl sulfide, and the V3+ species in V(1−x)Crx-Hol were oxidized to V4+
after deoxygenation. However, the V4+ species could not be re-reduced with H2. By supporting platinum
nanoparticles on V(1−x)Crx-Hol, the re-reduction of the oxidized vanadium species with H2 became possible.
Thus, in the presence of Pt/V(1−x)Crx-Hol, catalytic sulfoxide deoxygenation with H2 as the reducing agent
can be realized; the deoxygenation occurs on V(1−x)Crx-Hol, and the role of platinum is to re-reduce the
oxidized vanadium species with H2. In the presence of Pt/V0.7Cr0.3-Hol, various types of structurally diverse
sulfoxides could be selectively converted into their corresponding sulfides under atmospheric H2 pressure
(1 atm). In addition, the present system was applicable to the deoxygenation of pyridine N-oxides to their
corresponding pyridines. The observed catalysis was truly heterogeneous, and the Pt/V0.7Cr0.3-Hol catalyst
could be reused for sulfoxide deoxygenation, although the performance was reduced.
Received 21st March 2017,
Accepted 3rd April 2017
DOI: 10.1039/c7cy00547d
rsc.li/catalysis
Mitsudome et al. demonstrated sulfoxide deoxygenation under at-
mospheric H2 pressure (1 atm) with a heterogeneous Ru/TiO2 cat-
Introduction
Sulfoxides are now widely utilized as chirons in numerous asym-
metric transformations wherein the sulfoxide groups, after stereo-
selective introduction, are removed by deoxygenation to their cor-
responding sulfides followed by hydrodesulfurization.1a,b In other
transformations such as homologation of aromatic aldehydes, de-
oxygenation of the sulfoxide moieties plays a crucial role.1 To
date, a variety of catalytic procedures have been developed for the
deoxygenation of sulfoxides; these typically utilize reducing
agents such as zinc metal, silanes, boranes, phosphines, and al-
cohols.2 The use of molecular hydrogen (H2) as a reducing agent
in combination with heterogeneous catalysts is more desirable
from the viewpoint of green chemistry. However, to the best of
our knowledge, there are only a few reports on heterogeneously
catalyzed deoxygenation of sulfoxides with H2.3 In 2014,
alyst for the first time.3a Shimizu et al. subsequently developed
an efficient catalytic deoxygenation system under high pressure
H2 conditions (H2 7 atm) by utilizing a Pt–MoOx/TiO2 catalyst.3b
In this study, we developed a new concept in heteroge-
neous catalysts for the deoxygenation of sulfoxides to their
corresponding sulfides based on the properties of hollandite-
type vanadium–chromium mixed oxides (Fig. 1). Hollandite-
type oxides are composed of double-chains of edge-sharing
MO6 octahedra (M = transition metal), forming 2 × 2 and 1 ×
1 tunnel structures, and various types of cations or anions
can be situated in the 2 × 2 tunnel.4 Owing to the presence of
the tunnel ions, the lattice transition metals tend to possess
mixed valence states, which lead to specific properties of
hollandite-type oxides such as specific electron and ion con-
ductivities.5 Not only these properties but also the sizes, mor-
phologies, and stabilities of hollandite-type oxides can be
finely tuned by substitutional doping into the lattice and/or
changing the tunnel ions.6 Thus, we envisaged that suitably
designed hollandite-type oxides could probably act as efficient
Department of Applied Chemistry, School of Engineering, The University of Tokyo,
7-3-1 Hongo, Bunkyo-ku, Tokyo 113-8656, Japan.
E-mail: kyama@appchem.t.u-tokyo.ac.jp, tmizuno@mail.ecc.u-tokyo.ac.jp
† Electronic supplementary information (ESI) available: Fig. S1–S6. See DOI:
10.1039/c7cy00547d
This journal is © The Royal Society of Chemistry 2017
Catal. Sci. Technol.

View Article Online
Paper
Catalysis Science & Technology
Fig. 1 Design concept of heterogeneous catalysts for sulfoxide deoxygenation with H2 based on hollandite-type vanadium–chromium mixed
oxides.
heterogeneous catalysts for various types of functional group
transformations. Hollandite-type manganese oxides are reported
to be quite effective for numerous oxidation reactions.7 Also, we
recently reported that hollandite-type manganese oxides show
higher catalytic performances and stability for several oxidation
reactions such as aerobic oxygenation of sulfides and alkylarenes
than manganese oxides with other structures and that the per-
formance can be much improved by substitutional doping of
several transition metals into the hollandite framework.8
The noticeable new findings in this study are as follows
(Fig. 1): (i) the stoichiometric deoxygenation of methyl phenyl
sulfoxide to methyl phenyl sulfide proceeded efficiently in
the presence of a hollandite-type vanadium oxide composed
of V3+ and V4+ (V-Hol), while the reaction hardly proceeded
when V2O3 and VO2 were used; (ii) by substitutional doping
of Cr3+ into the framework to form hollandite-type vana-
dium–chromium mixed oxides (V(1−x)Crx-Hol, x: Cr/(V + Cr)
molar ratio of the precursor solution), the specific surface
areas significantly increased in comparison with that of V-
Hol, and V(1−x)Crx-Hol oxides were more stable than V-Hol;
(iii) after deoxygenation, the V3+ species in V(1−x)Crx-Hol were
oxidized to V4+ with retention of the hollandite structures,
but the V4+ species could not be re-reduced with H2; (iv) by
alcohol (see the Experimental section).9 The X-ray diffraction
(XRD) pattern of V-Hol matched well with that reported in
the ICDD PDF database (JCPDS 50-1797), corresponding to
V8.18O16·1.46H2O with a hollandite-type structure (Fig. 2).
According to the previous report, V-Hol contains oxide ions
in the 2 × 2 tunnel.9 The specific surface area of the V-Hol
sample prepared in this study was 18 m2 g−1, which was not
significantly high with respect to its surface area when used
as a catalyst. Substitutional doping of additional metal cat-
ions into the metal oxide framework is one of the most com-
monly utilized strategies to control the catalytic properties,
sizes, and morphologies. In our previous study, we found
that substitutional doping of Mo6+ into hollandite-type man-
ganese oxides can significantly increase their specific surface
areas.8 It is expected that six coordinated cations having crys-
tal radii similar to those of vanadium cations (V3+, V4+, and
V5+) can be readily introduced into the octahedral hollandite
framework. Thus, to increase the specific surface area, we
attempted to introduce several transition metal cations such
as Ti4+, Cr3+, Mn3+, Fe3+, Co2+, Ni2+, and Cu2+. The prepara-
tion of pure hollandite-type mixed oxides was successful only
when Cr3+ was used.
In this study, five types of hollandite-type vanadium–chro-
mium mixed oxides with various vanadium and chromium
supporting platinum nanoparticles on
(1−x)Crx-Hol), the re-reduction of the oxidized vanadium spe-
cies with H2 was made possible; (v) in the presence of Pt/
0.7Cr0.3-Hol, various types of structurally diverse sulfoxides
V(1−x)Crx-Hol (Pt/
V
V
could be converted into their corresponding sulfides under
atmospheric H2 pressure, and the deoxygenation of pyridine
N-oxides to their corresponding pyridines was also possible;
and (vi) the observed catalysis was truly heterogeneous, and
Pt/V0.7Cr0.3-Hol could be reused for sulfoxide deoxygenation.
Results and discussion
Synthesis of hollandite-type oxides and their use in sulfoxide
deoxygenation
To begin with, we prepared a hollandite-type vanadium oxide
(V-Hol) according to a previous report with minor modifica-
tions involving the liquid-phase reaction of VOCl3 with benzyl
Fig. 2 XRD patterns of V-Hol and V(1−x)Crx-Hol.
Catal. Sci. Technol.
This journal is © The Royal Society of Chemistry 2017

View Article Online
Catalysis Science & Technology
Paper
contents (V(1−x)Crx-Hol, x: Cr/(V + Cr) molar ratio of the pre-
cursor solution, x = 0.1, 0.2, 0.3, 0.4, and 0.5) were success-
fully prepared using essentially the same procedure as that
for V-Hol with CrIJacac)3 (acac = acetylacetonate) as the chro-
mium source (see the Experimental section). The specific sur-
face areas and the elemental analysis data for V(1−x)Crx-Hol
are summarized in Table 1. The specific surface areas in-
creased with the chromium content; the specific surface
areas of V0.9Cr0.1-Hol, V0.8Cr0.2-Hol, V0.7Cr0.3-Hol, V0.6Cr0.4-Hol,
and V0.5Cr0.5-Hol were 37, 61, 81, 115, and 180 m2 g−1, respec-
tively, which were significantly larger than that of the parent
undoped V-Hol (18 m2 g−1) (Table 1). The XRD patterns of
V(1−x)Crx-Hol are shown in Fig. 2. From these XRD patterns,
we concluded that all V(1−x)Crx-Hol samples possessed a
hollandite-type structure. No additional diffraction peaks at-
tributed to segregated Cr2O3 were observed, and small
changes in the lattice parameters of V(1−x)Crx-Hol with respect
to chromium were observed as a result of changing the dop-
ing level. These results indicated that Cr3+ species were effi-
ciently doped into the hollandite framework. The XRD peak
intensities became weaker as the chromium content was in-
creased, suggesting the low crystallinity and/or formation of
smaller crystals of
V(1−x)Crx-Hol. Transmission electron
microscopy (TEM) images revealed that V-Hol possessed a
rod-like morphology, and the average length and width of the
rods were 200 50 and 50 5 nm, respectively (Fig. 3a). As
shown in Fig. 3, the lengths and widths decreased as the
chromium content was increased, resulting in the formation
of grain-like aggregates of ultrafine nanocrystals; for example,
the average length and width of V0.5Cr0.5-Hol were 30 10
and 5 2.5 nm, respectively (Fig. 3f). Therefore, the larger
specific surface areas of V(1−x)Crx-Hol possibly resulted from
the formation of nanocrystals. The normalized vanadium
K-edge X-ray absorption near edge structure (XANES) spectra
of V-Hol and V(1−x)Crx-Hol are shown in Fig. 4a. The absorp-
tion edges of V-Hol and V(1−x)Crx-Hol were located between
those of V2O3 (V3+ reference) and VO2 (V4+ reference)
(Fig. 4a). Therefore, the vanadium species in V-Hol and
Fig. 3 TEM images of (a) V-Hol, (b) V0.9Cr0.1-Hol, (c) V0.8Cr0.2-Hol, (d)
V
0.7Cr0.3-Hol, (e) V0.6Cr0.4-Hol, and (f) V0.5Cr0.5-Hol.
cess amount of V-Hol (300 mg, ca. 6.3 equiv. vanadium spe-
cies with respect to 1a), the deoxygenation proceeded effi-
ciently at 100 °C in n-decane, affording 2a in 74% yield
(Table 2, entry 1). After the deoxygenation of 1a, V-Hol was re-
trieved by filtration, and the XRD pattern was obtained. As
can be seen in Fig. 5, the hollandite structure was intrinsi-
cally preserved after deoxygenation, while peak shifts toward
higher angles were observed. Therefore, we consider that the
V3+ species in V-Hol may have been partially oxidized during
deoxygenation.10,11 The oxidized vanadium species could not
be re-reduced, even when the reaction was carried out under
a H2 atmosphere. Although the hollandite structure of V-Hol
was intrinsically stable during the reaction in non-polar sol-
vents, such as n-decane, the use of polar solvents was prob-
lematic. When the reaction was performed in N,N-
dimethylformamide (DMF), V-Hol was partially converted into
vanadium oxide(s) with different structureIJs) (Fig. S1, ESI†).
To realize catalytic deoxygenation with H2, we attempted to
prepare a platinum-supporting V-Hol catalyst by a precipita-
tion method in an aqueous medium (see the Experimental
section). However, the catalyst preparation failed because
V-Hol was completely converted into V6O13 during the prepa-
ration (Fig. S2, ESI†). Owing to the aforementioned
V
(1−x)Crx-Hol possibly possess mixed valence states of V3+ and
V4+. The normalized chromium K-edge XANES spectra of
(1−x)Crx-Hol reveal that the oxidation state of the chromium
V
species in V(1−x)Crx-Hol was +3 (Fig. 4b).
The deoxygenation of methyl phenyl sulfoxide (1a) to
methyl phenyl sulfide (2a) was performed with hollandite-
type oxides under an Ar atmosphere. In the presence of an ex-
Table 1 Properties of V-Hol and V(1−x)Crx-Hol
BET
Content (wt%)
Cr/(V +
Cr)
(mol%)
surface area
Sample
(m2 g−1
)
V
Cr
V-Hol
18
37
61
81
115
180
53.7
48.4
42.1
38.2
29.6
25.8
—
5.5
8.5
12.8
18.9
20.8
—
V0.9Cr0.1-Hol
V0.8Cr0.2-Hol
V0.7Cr0.3-Hol
V0.6Cr0.4-Hol
V0.5Cr0.5-Hol
10.1
16.5
24.8
38.5
44.1
This journal is © The Royal Society of Chemistry 2017
Catal. Sci. Technol.

View Article Online
Paper
Catalysis Science & Technology
Fig. 5 XRD patterns of the as-prepared V-Hol and V(1−x)Crx-Hol (black
lines), and the patterns of the retrieved V-Hol and V(1−x)Crx-Hol after
the deoxygenation of 1a under the conditions described in Table 2 (red
lines).
When V(1−x)Crx-Hol (300 mg, ca. 3.0–5.7 equiv. of vana-
dium species with respect to 1a) were used, the deoxygen-
ation of 1a also proceeded significantly, giving 2a in 20–83%
yields (Table 2, entries 2–6). The XRD analysis of V(1−x)Crx-Hol
retrieved after the deoxygenation revealed that the hollandite
structure was intrinsically preserved after the reaction
(Fig. 5). In these cases, the XRD peaks were shifted toward
higher angles, indicating the oxidation of V(1−x)Crx-Hol after
the reaction. As mentioned above, in the vanadium K-edge
XANES spectrum of the freshly prepared V0.7Cr0.3-Hol, the ab-
sorption edge was located between those of V2O3 (V3+ refer-
ence) and VO2 (V4+ reference), indicating that V0.7Cr0.3-Hol
possesses mixed valence states of V3+ and V4+ (Fig. 6a). After
this had been used for the deoxygenation, the absorption
edge was shifted toward higher energy, and the spectrum
approached that of VO2 (Fig. 6a), indicating that the V3+ spe-
cies in V0.7Cr0.3-Hol were oxidized to V4+ during deoxygen-
ation. In contrast, the chromium K-edge XANES spectrum of
Fig. 4 Normalized (a) vanadium and (b) chromium K-edge XANES
spectra of V-Hol, V(1−x)Crx-Hol, and reference materials (V2O3, VO2,
and Cr2O3).
instability, we did not utilize V-Hol for further examinations
of sulfoxide deoxygenation. In contrast, the doping of Cr3+
species could stabilize the hollandite structure; V(1−x)Crx-Hol
oxides were highly stable even in polar solvents and strong
aqueous basic media (Fig. S1 and S2, ESI†).
V0.7Cr0.3-Hol was almost unchanged after deoxygenation
Table 2 Deoxygenation of 1a using various vanadium oxides
Entry
Sample
Vanadium content (mmol)
Yield of 2a (%)
1
2
3
4
5
6
7
8
V-Hol
3.16
2.85
2.48
2.25
1.74
1.52
3.62
4.00
74
66
83
71
50
20
<1
3
V0.9Cr0.1-Hol
V0.8Cr0.2-Hol
V0.7Cr0.3-Hol
V0.6Cr0.4-Hol
V0.5Cr0.5-Hol
VO2
V2O3
Reaction conditions: vanadium oxides (300 mg), 1a (0.5 mmol), n-decane (1 mL), Ar (1 atm), 100 °C, 24 h. Yields were determined via GC analy-
sis using naphthalene as an internal standard.
Catal. Sci. Technol.
This journal is © The Royal Society of Chemistry 2017

View Article Online
Catalysis Science & Technology
Paper
(Fig. 6b). Consequently, the V3+ species in V0.7Cr0.3-Hol acted
as the reducing agent for deoxygenation,11 whereas the Cr3+
species did not. Also, in the case of V0.7Cr0.3-Hol, the oxidized
vanadium species could not be re-reduced, even when the re-
action was performed under a H2 atmosphere.
the oxide ion is smoothly transferred from the sulfoxide to
the tunnel of the hollandite-type oxides.
Catalytic sulfoxide deoxygenation with H2 using platinum-
supporting hollandite-type oxides
No production of 2a was observed when VO2 was used for
the deoxygenation of 1a (Table 2, entry 7). Although the V3+
species in the hollandite-type oxides used in this study could
effectively deoxygenate 1a (Table 2, entries 1–6), the reaction
using V2O3, which contains V3+, gave 2a in only 3% yield
(Table 2, entry 8). Although the specific surface area of V2O3
was 5 m2 g−1, which was slightly smaller than that of V-Hol
(18 m2 g−1), the performance of V-Hol was much better than
that of V2O3 (Table 2, entry 1 vs. entry 8). Therefore, the
higher performance of the hollandite-type oxides as com-
pared with V2O3 is probably not caused by the difference in
their specific surface areas. Although we cannot conclusively
explain the reason for the higher performance of the
hollandite-type oxides, we consider that the specific proper-
ties of hollandite-type oxides, such as specific electron and
ion conductivities, play an important role.5 It has been
reported that V-Hol contains oxide ions for charge compensa-
tion in the 2 × 2 tunnel.9 Thus, after the two-electron transfer
from the vanadium species to a sulfoxide, it is possible that
As described in the previous section, the oxidized vanadium
species in V(1−x)Crx-Hol could not be re-reduced with H2.
Herein, we prepared platinum-supporting V(1−x)Crx-Hol cata-
lysts (Pt/V(1−x)Crx-Hol). Catalytic sulfoxide deoxygenation with
H2 is expected to be realized through the following catalytic
mechanism: (i) two-electron transfer from V(1−x)Crx-Hol to a
sulfoxide and successive oxide ion transfer from the sulfoxide
to V(1−x)Crx-Hol, producing the desired sulfide; (ii) H2 dissoci-
ation on platinum; (iii) two-electron transfer from two hydro-
gen atoms to the oxidized V(1−x)Crx-Hol via platinum (re-re-
duction of V(1−x)Crx-Hol); and (iv) reaction of two protons and
the oxide ion, producing water as the co-product. Four types
of effective Pt/V(1−x)Crx-Hol catalysts (x = 0.2, 0.3, 0.4, and 0.5;
Pt contents: 1.0–1.4 wt%)12 were successfully prepared by a
simple precipitation method in an aqueous medium (see the
Experimental section). As we expected, catalytic sulfoxide de-
oxygenation with H2 as the reducing agent was realized in
the presence of Pt/V(1−x)Crx-Hol, and Pt/V0.7Cr0.3-Hol was the
best catalyst (Fig. S3, ESI†);13 for example, when Pt/V0.7Cr0.3
-
Hol (30 mg, Pt: 0.43 mol%, V: 45 mol%) was used, the deoxy-
genation of 1a proceeded efficiently under mild conditions
(H2 1 atm, 100 °C, in n-decane), affording the desired sulfide
2a in 98% yield after 2 h (Fig. 7). Under the same reaction
conditions, commonly utilized supported platinum catalysts
such as Pt/Al2O3 (30 mg, Pt: 0.43 mol%) and Pt/C (9.0 mg, Pt:
0.46 mol%) were not effective for the deoxygenation of 1a
and afforded 2a in only 5% and 6% yields, respectively, thus
indicating that the vanadium species (not platinum) play an
important role in the present Pt/V(1−x)Crx-Hol-catalyzed sulfox-
ide deoxygenation.14
The XRD pattern of Pt/V0.7Cr0.3-Hol just before its use in
the deoxygenation (just after the H2 pretreatment, see the Ex-
perimental section) is shown as black lines in Fig. 8. In the
XRD pattern of Pt/V0.7Cr0.3-Hol retrieved after the deoxygen-
ation of 1a under an Ar atmosphere, the diffraction peaks
were shifted toward higher angles (Fig. 8, red lines), which
was virtually the same as the effect observed in Fig. 5, indi-
cating the oxidation of the V3+ species in Pt/V0.7Cr0.3-Hol.
When the oxidized catalyst was treated under the conditions
described in Fig. 7 without sulfoxides (H2 1 atm, 100 °C, in
n-decane), the diffraction peaks were shifted to their original
positions (Fig. 8, blue lines), thus showing that the oxidized
vanadium species were re-reduced by H2 through platinum
catalysis. All the aforementioned experimental results sup-
port the proposed mechanism. This mechanism is completely
different from the reported one using Pt–MoOx/TiO2 under
rather harsh conditions (typically, H2 7 atm, 120 °C, solvent-
free).3b
Fig. 6 Normalized (a) vanadium and (b) chromium K-edge XANES
spectra of the as-prepared V0.7Cr0.3-Hol, the retrieved V0.7Cr0.3-Hol af-
ter the deoxygenation of 1a under the conditions described in Table 2,
and the reference materials (V2O3, VO2, and Cr2O3).
To verify whether the observed catalysis was truly caused
by solid Pt/V0.7Cr0.3-Hol or leached metal species, the
This journal is © The Royal Society of Chemistry 2017
Catal. Sci. Technol.

View Article Online
Paper
Catalysis Science & Technology
Fig. 7 Scope of the Pt/V0.7Cr0.3-Hol-catalyzed-deoxygenation of (a) sulfoxides and (b) pyridine N-oxides. Reaction conditions: Pt/V0.7Cr0.3-Hol (30 mg,
Pt: 0.43 mol%, V: 45 mol%), sulfoxides or pyridine N-oxides (0.5 mmol), n-decane (1 mL), H2 (1 atm), 100 °C. Pt/V0.7Cr0.3-Hol was utilized after
pretreatment in 1 atm of H2 at 150 °C for 30 min. Yields were determined via GC analysis using naphthalene as an internal standard. aReuse experiment.
following control experiments were conducted. The deoxygen-
ation of 1a was performed under the conditions described in
Fig. 7, and the Pt/V0.7Cr0.3-Hol catalyst was removed from the
reaction mixture by hot filtration at ca. 40% yield of 2a. Then,
the filtrate was again heated at 100 °C under atmospheric H2
pressure (1 atm). In this case, no further production of 2a was
observed (Fig. S4, ESI†). In addition, we confirmed via induc-
tively coupled plasma atomic emission spectroscopy (ICP-AES)
that no vanadium, chromium, or platinum species were pres-
ent in the filtrate (below 0.1%). These results clearly indicate
that the observed catalysis for the present Pt/V0.7Cr0.3-Hol-
catalyzed sulfoxide deoxygenation is truly heterogeneous.15 Af-
ter the reaction was completed, Pt/V0.7Cr0.3-Hol could be easily
retrieved from the reaction mixture by simple filtration. We
performed a reuse experiment using the retrieved catalyst. Un-
fortunately, the catalytic performance of Pt/V0.7Cr0.3-Hol some-
what decreased; a 98% yield of 2a was obtained with the freshly
prepared catalyst, while a 75% yield was obtained with the
reused catalyst for the deoxygenation of 1a under the condi-
tions described in Fig. 7. We confirmed by TEM analysis that
the average particle size of platinum increased after its use; the
average particle size of the freshly prepared Pt/V0.7Cr0.3-Hol was
0.6 nm, and this increased to 1.8 nm after use (Fig. S5, ESI†).
Therefore, the deactivation is probably caused by the increase in
the platinum particle sizes.
Finally, we turn our attention to the examination of the
substrate scope for the current Pt/V0.7Cr0.3-Hol-catalyzed deox-
ygenation. In the presence of Pt/V0.7Cr0.3-Hol, various types of
structurally diverse sulfoxides could be converted into their
corresponding sulfides under atmospheric H2 pressure (1
atm) (Fig. 7). Aryl sulfoxides could be almost quantitatively
converted into the corresponding sulfides. Notably, the
present Pt/V0.7Cr0.3-Hol-catalyzed system was effective for sulf-
oxide substrates with other reducible substituents such as
CC double bonds, Cl, and Br; the deoxygenation of sulfox-
ides proceeded selectively to produce the corresponding sul-
fides without the reduction of these functionalities. An alkyl
sulfoxide was also converted into its corresponding alkyl sul-
fide, although a longer reaction time was required to attain a
high yield in comparison with the reaction with aryl sulfox-
ides. Interestingly, the present Pt/V0.7Cr0.3-Hol-catalyzed sys-
tem was also applicable to the deoxygenation of pyridine
N-oxides. Deoxygenation of pyridine N-oxides is an important
reaction for the synthesis of substituted pyridines, and the
development of an efficient catalytic system is highly desir-
able.16 In the presence of Pt/V0.7Cr0.3-Hol, pyridine N-oxide
and its derivatives, which possess electron-donating as well
as electron-withdrawing substituents on the pyridine rings,
were converted into their corresponding pyridines under at-
mospheric H2 pressure (1 atm) (Fig. 7).
Catal. Sci. Technol.
This journal is © The Royal Society of Chemistry 2017

View Article Online
Catalysis Science & Technology
Paper
Experimental section
Materials
VOCl3, VO2, V2O3, VIJacac)3, VOIJacac)2, and H2PtIJOH)6 were
purchased from Aldrich. CrIJacac)3 was purchased from Kanto
Chemical. Pt/C (Pt content: 5.0 wt%) was purchased from N.
E. Chemcat Corporation. An aqueous solution of polyIJN-vinyl-
2-pyrrolidone)-stabilized Pt colloidal nanoparticles (Pt-PVP, Pt:
4 wt%) was purchased from Tanaka Kikinzoku Kogyo. Sub-
strates, solvents, and naphthalene (internal standard for GC
analysis) were purchased from Tokyo Chemical Industry,
Kanto Chemical, Wako Pure Chemical Industry, or Apollo Sci-
entific. All reagents were used as received without further
purification.
Preparation of catalysts
V-Hol was prepared according to the following procedure.9
VOCl3 (2.0 mL; 3.68 g, 21.24 mmol) was slowly added to ben-
zyl alcohol (80 mL) under Ar. The mixture was stirred at room
temperature for several minutes. The solution gradually
turned from red to green–blue. The reaction solution was
transferred to a Teflon-lined vessel for solvothermal reaction
and heated at 150 °C for 24 h. The resulting black suspension
was filtered, and the precipitate was washed with ethanol
(500 mL) and dried at 60 °C in open air, affording 0.8 g of
V-Hol. For the preparation of V(1−x)Crx-Hol, CrIJacac)3 was ini-
tially added to benzyl alcohol (10–50 mol% with respect to
the total metal content), followed by the above procedure,
affording 1.0–3.0 g of V(1−x)Crx-Hol. Pt/V(1−x)Crx-Hol catalysts
were prepared as follows. H2PtIJOH)6 (1.5 wt% with respect to
the supports) was dispersed in water (30 mL). The pH of the
suspension was adjusted to 12.9 with aqueous NaOH (0.1 M).
At this pH, H2PtIJOH)6 was dissolved completely. Then,
Fig. 8 XRD patterns ((a) overall views and (b) magnified views) of Pt/
V0.7Cr0.3-Hol just after the H2 pretreatment (black lines), the retrieved
Pt/V0.7Cr0.3-Hol after the deoxygenation of 1a under an Ar atmosphere
(red lines), and Pt/V0.7Cr0.3-Hol after treatment under the conditions
described in Fig. 7 without sulfoxides (H2 1 atm, 100 °C, in n-decane)
(blue lines).
Conclusions
We have successfully developed a new concept in heteroge-
neous sulfoxide deoxygenation catalysts, Pt/V(1−x)Crx-Hol,
based on hollandite-type vanadium–chromium mixed oxides.
In the presence of Pt/V0.7Cr0.3-Hol, various types of structur-
ally diverse aryl and alkyl sulfoxides could be converted into
their corresponding sulfides under atmospheric H2 pressure
(1 atm). Other reducible functional groups were completely
tolerant to reduction, and the deoxygenation of sulfoxides
proceeded selectively. Moreover, the deoxygenation of pyri-
dine N-oxides to their corresponding pyridines proceeded ef-
ficiently in the presence of Pt/V0.7Cr0.3-Hol. The observed ca-
talysis was truly heterogeneous, and the Pt/V0.7Cr0.3-Hol
catalyst could be reused for sulfoxide deoxygenation, though
the performance was somewhat decreased. Our proposed
sulfoxide deoxygenation proceeds through the following cata-
lytic mechanism: (i) two-electron transfer from V(1−x)Crx-Hol
to a sulfoxide, and successive oxide ion transfer from the
sulfoxide to V(1−x)Crx-Hol, producing the desired sulfide; (ii)
H2 dissociation on platinum; (iii) two-electron transfer from
two hydrogen atoms to the oxidized V(1−x)Crx-Hol via plati-
num; and (iv) reaction of two protons and the oxide ion, with
water as the co-product. This mechanism is significantly dif-
ferent from previously reported ones.
V(1−x)Crx-Hol (0.5 g) was added to the platinum-containing so-
lution, and the mixture was stirred at room temperature. Af-
ter 15 min, the pH of the suspension was kept at 7.5 for 15
min using aqueous HNO3 (0.1 M), and then the solution was
stirred for a further 6 h. The precipitate was recovered by fil-
tration, washed with a large amount of water (300 mL), and
dried in vacuo at room temperature overnight, affording Pt/
V(1−x)Crx-Hol. The Pt contents in Pt/V0.8Cr0.2-Hol, Pt/V0.7Cr0.3-
Hol, Pt/V0.6Cr0.4-Hol, and Pt/V0.5Cr0.5-Hol were 1.1, 1.4, 1.1,
and 1.0 wt%, respectively. Pt/Al2O3 (Pt content: 1.4 wt%) was
prepared using the same procedure. Pt/V2O5 (Pt content: 0.66
wt%), Pt/VO2 (Pt content: 0.60 wt%), and Pt/V2O3 (Pt content:
0.69 wt%) were prepared via an impregnation method using
an aqueous solution of Pt–PVP.
Characterization of V-Hol and V(1−x)Crx-Hol
The XRD patterns were recorded using a Rigaku SmartLab in-
strument with Cu Kα radiation (λ = 1.5418 Å, 45 kV, 200 mA).
The BET surface areas were measured by N2 adsorption at
−196 °C using a Micromeritics ASAP 2010 instrument. The
contents of vanadium, chromium, and platinum in the cata-
lyst samples were determined via ICP-AES using a Shimadzu
This journal is © The Royal Society of Chemistry 2017
Catal. Sci. Technol.

View Article Online
Paper
Catalysis Science & Technology
ICPS-8100 apparatus. The TEM images were obtained using a
JEM-2000EX II apparatus at an accelerating voltage of 200 kV.
The TEM specimens were prepared by dispersing the samples
in ethanol, sonicating them for 10 min, and leaving them to
stand for 30 min before depositing on a carbon-coated cop-
per grid. The XANES spectra were measured at the beam line
BL11S2 of the Aichi Synchrotron Radiation Center. Each sam-
ple was formed into a pellet, doubly wrapped in an Al lami-
nate pack, and heat-sealed. The samples were attached to a
sample holder. Vanadium and chromium foils were used for
energy calibration.
Catal. Lett., 2011, 141, 833; (g) S. Enthaler, ChemCatChem,
2011, 3, 666; (h) N. García, P. García-García, M. A.
Fernández-Rodríguez, R. Rubio, M. R. Pedrosa, F. J. Arnáiz
and R. Sanz, Adv. Synth. Catal., 2012, 354, 321; (i) N. García,
P. García-García, M. A. Fernández-Rodríguez, D. García,
M. R. Pedrosa, F. J. Arnáiz and R. Sanz, Green Chem.,
2013, 15, 999 and references cited therein; ( j) K. Oh and
W. E. Knabe, Tetrahedron, 2009, 65, 2966.
3 (a) T. Mitsudome, Y. Takahashi, T. Mizugaki, K. Jitsukawa
and K. Kaneda, Angew. Chem., Int. Ed., 2014, 53, 8348; (b)
A. S. Touchy, S. M. A. H. Siddiki, W. Onodera, K. Kon and K.
Shimizu, Green Chem., 2016, 18, 2554.
4 T. Gao, M. Glerup, F. Krumeich, R. Nesper, H. Fjellvåg and
P. Norby, J. Phys. Chem. C, 2008, 112, 13134.
Deoxygenation reactions
GC quantitative analyses were performed on a Shimadzu GC-
2014 apparatus with an FID detector equipped with a TC-1 or
Stabilwax capillary column. GC-MS analyses for product iden-
tification were performed on a GCMS-QP2010 apparatus
equipped with an InertCap 5 capillary column at an ioniza-
tion voltage of 70 eV. Sulfoxide deoxygenation was performed
as follows. Pt/V(1−x)Crx-Hol (30 mg) or vanadium-based oxide
(300 mg), sulfoxide (0.5 mmol), n-decane (1 mL), and naph-
thalene (internal standard, 0.2 mmol) were placed in a
Schlenk-type reactor. Then, H2 or Ar (1 atm) was sealed into
the reactor, and the reaction was performed at 100 °C for 2–
24 h with stirring. For the catalytic deoxygenation using Pt/
5 (a) R. N. DeGuzman, Y.-F. Shen, E. J. Neth, S. L. Suib, C.-L.
O'Young, S. Levine and J. M. Newsam, Chem. Mater., 1994, 6,
815; (b) Y. Yuan, C. Zhan, K. He, H. Chen, W. Yao, S. Sharifi-
Asl, B. Song, Z. Yang, A. Nie, X. Luo, H. Wang, S. M. Wood,
K. Amine, M. S. Islam, J. Lu and R. Shahbazian-Yassar, Nat.
Commun., 2016, 7, 13374.
6 (a) C. K. King'ondu, N. Opembe, C.-H. Chen, K. Ngala, H.
Huang, A. Iyer, H. F. Garcés and S. L. Suib, Adv. Funct.
Mater., 2011, 21, 312; (b) P. Moetakef, A. M. Larson, B. C.
Hodges, P. Zavalij, K. J. Gaskell, P. M. Piccoli and E. E.
Rodriguez, J. Solid State Chem., 2014, 220, 45; (c) Q. Feng, H.
Kanoh, Y. Miyai and K. Ooi, Chem. Mater., 1995, 7, 148.
7 (a) Y.-C. Son, V. D. Makwana, A. R. Howell and S. L. Suib,
Angew. Chem., Int. Ed., 2001, 40, 4280; (b) N. N. Opembe,
Y.-C. Son, T. Sriskandakumar and S. L. Suib, ChemSusChem,
2008, 1, 182; (c) K. Yamaguchi, Y. Wang and N. Mizuno,
ChemCatChem, 2013, 5, 2835; (d) S. Dharmarathna, C. K.
King'ondu, L. Pahalagedara, C.-H. Kuo, Y. Zhang and S. L.
Suib, Appl. Catal., B, 2014, 147, 124; (e) K. Yamaguchi, K.
Sakagami, Y. Miyamoto, X. Jin and N. Mizuno, Org. Biomol.
Chem., 2014, 12, 9200; ( f ) K. Yamaguchi, H. Kobayashi, T.
Oishi and N. Mizuno, Angew. Chem., Int. Ed., 2012, 51, 544;
(g) K. Yamaguchi, H. Kobayashi, Y. Wang, T. Oishi, Y.
Ogasawara and N. Mizuno, Catal. Sci. Technol., 2013, 3, 318
and references cited therein; (h) Y. Wang, H. Kobayashi,
K. Yamaguchi and N. Mizuno, Chem. Commun., 2012, 48,
2642.
V(1−x)Crx-Hol, the catalyst was pretreated at 150 °C under 1
atm of H2 for 30 min just before use. It is well known and we
confirmed in a separate experiment that Pt(0) nanoparticles
are formed after such H2 treatment. After the reaction, the re-
action mixture was diluted with toluene or acetone (3 mL),
and the spent catalyst was separated by filtration and washed
with acetone. Then, the filtrate was analyzed by GC and GC-
MS. The products were confirmed by comparing their GC re-
tention times and GC-MS spectra with authentic data. The
spent catalyst was washed with acetone and water and then
dried in vacuo at room temperature before being reused.
Acknowledgements
This work was supported in part by JSPS KAKENHI Grant No.
15H05797 in Precisely Designed Catalysts with Customized
Scaffolding.
8 T. Uematsu, Y. Miyamoto, Y. Ogasawara, K. Suzuki, K.
Yamaguchi and N. Mizuno, Catal. Sci. Technol., 2016, 6, 222.
9 I. Djerdj, D. Sheptyakov, F. Gozzo, D. Arčon, R. Nesper and
M. Niederberger, J. Am. Chem. Soc., 2008, 130, 11364.
Notes and references
1 (a) M. Madesclaire, Tetrahedron, 1988, 44, 6537; (b) M. C.
Carreño, Chem. Rev., 1995, 95, 1717; (c) K. Ogura, M.
Yamashita and G. Tsuchihashi, Synthesis, 1975, 385.
10 It is known that various kinds of reduction reactions
efficiently proceed in the presence of low valence vanadium
compounds (in particular, compounds of V2+ and V3+): (a)
T.-L. Ho and G. A. Olah, Synthesis, 1976, 807; (b) G. A. Olah
and T.-L. Ho, Synthesis, 1976, 798; (c) T.-L. Ho and G. A.
Olah, Synthesis, 1977, 170; (d) D. H. R. Barton, D. O. Jang
and J. C. Jaszberenyi, Tetrahedron Lett., 1992, 33, 2311; (e)
T.-L. Ho and G. A. Olah, Synthesis, 1976, 815; ( f ) Y. Hayasi
and J. Schwartz, Inorg. Chem., 1981, 20, 3473; (g) G. A. Olah,
G. K. S. Prakash and T.-L. Ho, Synthesis, 1976, 810; (h) G. A.
Olah, M. Arvanaghi and G. K. Surya, Synthesis, 1980, 220.
2 (a) S. C. A. Sousa, J. R. Bernardo, M. Wolff, B. Machura and
A. C. Fernandes, Eur. J. Org. Chem., 2014, 1855; (b) S. Krackl,
A. Company, S. Enthaler and M. Driess, ChemCatChem,
2011, 3, 1186; (c) A. C. Fernandes and C. C. Romão,
Tetrahedron Lett., 2007, 48, 9176; (d) A. C. Fernandes, J. A.
Fernandes, C. C. Romão, L. F. Veiros and M. J. Calhorda,
Organometallics, 2010, 29, 5517; (e) S. Enthaler, Catal. Sci.
Technol., 2011, 1, 104; ( f ) S. Enthaler and M. Weidauer,
Catal. Sci. Technol.
This journal is © The Royal Society of Chemistry 2017

View Article Online
Catalysis Science & Technology
Paper
11 In the presence of VIJacac)3 (equimolar amount with respect
to 1a), the deoxygenation of 1a proceeded to some extent
under the conditions described in Table 2, giving 2a in 10%
yield for 3 h. Conversely, VOIJacac)2 was not effective for the
deoxygenation. Therefore, the V3+ species likely played an
important role in sulfoxide deoxygenation.
12 The preparation of platinum-supporting catalysts using
V-Hol and V0.9Cr0.1-Hol was unsuccessful because of the low
stabilities of V-Hol and V0.9Cr0.1-Hol in the catalyst prepara-
tion aqueous medium (see the Experimental section).
13 The specific surface areas increased with the chromium
content (Table 1). Differential thermal analysis (DTA)
revealed that the stabilities of V(1−x)Crx-Hol against oxidation
increased with the chromium content (Fig. S6†). The
increases in their specific surface areas and stabilities
should positively affect their catalytic activities. However, the
vanadium contents in V(1−x)Crx-Hol decreased (Table 1) and
the average oxidation states of vanadium species in V(1−x)Crx-
Hol gradually increased (Fig. 4a) as the chromium content
was increased. Therefore, the amounts of V3+ species (active
species) in the Pt/V(1−x)Crx-Hol catalysts decreased as the
chromium content was increased. For the above-mentioned
reasons, a volcano-shaped relationship was likely observed
between the chromium contents and the reaction rates for
the deoxygenation of 1a (Fig. S3b†).
14 Platinum-supporting vanadium oxide catalysts such as Pt/
V2O5 (30 mg, Pt: 0.20 mol%), Pt/VO2 (30 mg, Pt: 0.18 mol%),
and Pt/V2O3 (30 mg, Pt: 0.21 mol%) were less effective than
Pt/V0.7Cr0.3-Hol and gave 2a in 37%, 12%, and 10% yields
even after 4 h, respectively, under the conditions described
in Fig. 7.
15 R. A. Sheldon, M. Wallau, I. W. C. E. Arends and U.
Schuchardt, Acc. Chem. Res., 1998, 31, 485.
16 (a) L.-C. Campeau, S. Rousseaux and K. Fagnou, J. Am.
Chem. Soc., 2005, 127, 18020; (b) S. H. Cho, S. J. Hwang and
S. Chang, J. Am. Chem. Soc., 2008, 130, 9254; (c) Y. Mikami,
A. Noujima, T. Mitsudome, T. Mizugaki, K. Jitsukawa and K.
Kaneda, Chem. – Eur. J., 2011, 17, 1768; (d) S. Donck, E.
Gravel, N. Shah, D. V. Jawale, E. Doris and I. N. N.
Namboothiri, RSC Adv., 2015, 5, 50865; (e) H. P. Kokatla,
P. F. Thomson, S. Bae, V. R. Doddi and M. K. Lakshman,
J. Org. Chem., 2011, 76, 7842 and references cited therein;
( f ) R. Ma, A.-H. Liu, C.-B. Huang, X.-D. Li and L.-N. He,
Green Chem., 2013, 15, 1274.
This journal is © The Royal Society of Chemistry 2017
Catal. Sci. Technol.