Selective photo-assisted catalytic oxidation of methane and ethane to oxygenates using supported vanadium oxide catalysts

Article abstract of DOI:10.1039/a800938d

Selective photooxidation of light alkanes, mainly methane and ethane, into the corresponding aldehydes was achieved using silica-supported vanadium oxide catalysts under UV irradiation at elevated temperature. Photooxidation of methane using the V₂O₅/SiO₂-IW (incipient wetness) (0.6 mol% V) catalyst at 493 K for 2 h gave 68 μmol of methanal, which corresponds to 76 mol% selectivity and 0.48 mol% one-pass yield. Photooxidation of ethane using V₂O₅/SiO₂ (calcined at 1023 K) -IW (0.6 mol% V) catalyst for 1 h gave 85 μmol of ethanal, which corresponds to 90% selectivity and 1.1% one-pass yield. The catalysts prepared by the sol-gel method also showed activity, especially for the reaction of ethane. Both UV irradiation and a reaction temperature as high as 500 K were essential. The photo-assisted catalytic reactions were very sensitive to the reaction temperature, method of preparation of the catalyst, and addition of water vapour. While the reaction of methane was inhibited by the addition of water vapour, the photooxidation of ethane and propane was promoted in the presence of a controlled amount of water vapour. In addition, the reaction with methane required UV irradiation at a wavelength <310 nm, whereas the reaction with ethane or propane proceeded by irradiation at longer wavelength. According to Raman and UV diffuse reflectance spectroscopic studies and XRD studies, only isolated four-coordinated vanadium oxide surface species were considered to show catalytic activity toward the photooxidation of methane, while accumulated vanadium surface species, not crystalline V₂O₅, were shown to be active for the photooxidation of ethane and propane.

Full text of DOI:10.1039/a800938d

View Article Online / Journal Homepage / Table of Contents for this issue
Selective photo-assisted catalytic oxidation of methane and ethane to
oxygenates using supported vanadium oxide catalysts
Kenji Wada,*¤ Hiroshi Yamada, Yoshihisa Watanabe and Take-aki Mitsudo*
Department of Energy and Hydrocarbon Chemistry, Graduate School of Engineering, Kyoto
University, Sakyo-ku, Kyoto 606-8501, Japan
Selective photooxidation of light alkanes, mainly methane and ethane, into the corresponding aldehydes was achieved using
silica-supported vanadium oxide catalysts under UV irradiation at elevated temperature. Photooxidation of methane using the
V O /SiO -IW (incipient wetness) (0.6 mol% V) catalyst at 493 K for 2 h gave 68 lmol of methanal, which corresponds to 76
2
5
2
mol% selectivity and 0.48 mol% one-pass yield. Photooxidation of ethane using V O /SiO (calcined at 1023 K) -IW (0.6 mol%
2
5
2
V) catalyst for 1 h gave 85 lmol of ethanal, which corresponds to 90% selectivity and 1.1% one-pass yield. The catalysts prepared
by the solÈgel method also showed activity, especially for the reaction of ethane. Both UV irradiation and a reaction temperature
as high as 500 K were essential. The photo-assisted catalytic reactions were very sensitive to the reaction temperature, method of
preparation of the catalyst, and addition of water vapour. While the reaction of methane was inhibited by the addition of water
vapour, the photooxidation of ethane and propane was promoted in the presence of a controlled amount of water vapour. In
addition, the reaction with methane required UV irradiation at a wavelength \310 nm, whereas the reaction with ethane or
propane proceeded by irradiation at longer wavelength. According to Raman and UV di†use reÑectance spectroscopic studies
and XRD studies, only isolated four-coordinated vanadium oxide surface species were considered to show catalytic activity
toward the photooxidation of methane, while accumulated vanadium surface species, not crystalline V O , were shown to be
2
5
active for the photooxidation of ethane and propane.
The development of the direct transformation of light alkanes
into more valuable compounds with high conversions and
selectivities is desirable.1h3 However, there have only been a
few successful attempts. In particular, the most expected selec-
tive aerobic oxidation of methane to oxygen-containing
chemicals is still extremely difficult, in spite of recent inter-
esting organometallic approaches.2,3 As a method to over-
been reported.14 For example, O~ radical anion species on
silica-supported vanadium oxide, produced by the decomposi-
tion of adsorbed N O or O , were reported to be relatively
2
2
stable, even above 500 K, whereas O ~ species were shown to
2
have a shorter lifetime. Since desorption of reaction products
from the surface is promoted at high temperatures, we exam-
ined the photooxidation of light alkanes at elevated reaction
temperature. Successful results for the selective oxidation of
methane, ethane and propane were obtained using silica-
supported molybdenum oxide catalysts15h17 or an n-type
solid metal oxide semiconductor18,19 at ca. 500 K. In these
studies using silica-supported catalysts, we used a Ñuidized
catalyst bed, in order to keep the temperature of the whole
catalyst bed uniform, and to irradiate all particles of the cata-
lyst. The reactions selectively yielded the corresponding alde-
hydes, but the conversions of methane did not reach the
desired levels. In our preliminary account, the supported
vanadium oxide catalysts were found to show excellent activ-
ity, especially for the reaction of methane.20
In the present paper our e†orts were mainly focused on the
design of optimum catalysts for the photooxidation of
methane. The e†ects of the method of preparation was exam-
ined, in order to investigate the correlation between the activ-
ity and the surface state of the catalyst. The e†ects of reaction
conditions, catalyst pretreatment and irradiation wavelength
were also investigated.
come these barriers, the reaction of methane with N O in the
2
presence of supported molybdenum oxide catalysts has been
studied.4 Selective formation of methanol was reported. The
O~ radical anion on the surface was proposed to be the active
and selective species. The results suggest that, if certain
oxygen species can be produced on the surface of the catalyst,
then selective oxidation could be achieved, even in the pres-
ence of dioxygen. In order to produce such oxygen species, the
e†ect of UV irradiation of the surface of various metal oxide
catalysts has been examined over the last two decades. The
activities of both supported metal oxide catalysts and n-type
solid oxide semiconductors towards photo-assisted catalytic
oxidation of light alkanes have been examined.5h13 These
studies were mainly performed at ambient or lower tem-
perature. Kaliaguine et al.5 reported the reaction of methane
and ethane by the use of c-irradiation of metal complexes sup-
ported on silica; the major products were carbon oxides,
together with a small amount of methanal. Ward et al.6
reported the photooxidation of methane using silica-
supported MoO or CuMoO catalysts at 313È373 K to give
3
4
a small amount of methanol without deep oxidation, but the
formation of methanal was not conÐrmed.
Experimental
Materials
In these earlier studies, secondary deep oxidation of the
adsorbed intermediates or products was considered to be
mainly responsible for their low selectivities. In addition, irra-
diation of the surface of certain metal oxides was reported to
induce spontaneous formation of various kinds of active
oxygen species, which essentially decrease the selectivity. Dif-
ferences in the thermal stabilities of active oxygen species have
Silica-supported vanadium oxide catalysts were prepared by
several methods: incipient wetness, designated V O /SiO -
2 5
2
IW, evaporation to dryness (V O /SiO -ED), and solÈgel
2 5
V O /SiO -SG). To prepare the IW-catalysts, 5 g of silica gel
2 5
2
(
(
2
Alfa, BET surface area 260 m2 g~1, mean particle size 10 lm)
was impregnated overnight in an oxalic acid solution of
ammonium metavanadate (10 cm3), followed by vacuum
evaporation at 323 K and calcination in air at 823 K for 2 h.
¤
E-mail: wadaken=scl.kyoto-u.ac.jp
J. Chem. Soc., Faraday T rans., 1998, 94(12), 1771È1778
1771

View Article Online
The V O /SiO -ED catalysts were prepared similarly, using
emission in the range 5 \ 2h/degrees \ 70. UVÈVIS di†use
reÑectance spectra were measured with a Perkin-Elmer
Lambda-19 multipurpose spectrophotometer, in which reÑec-
ted beams were gathered by an integrating sphere (100 mm
id). A UV cell (10 mm ] 40 mm, inner thickness 1.0 mm)
equipped with a branched chamber and a stop valve was used
in order to avoid any contact with moisture. The catalyst was
heated in the branched chamber at 573 K under atmospheric
pressure for 30 min, then evacuated at 823 K for 30 min, fol-
lowed by the treatment under 150 Torr of oxygen at 823 K for
2 h. The catalyst was then transferred into the UV cell, and
spectra were measured in vacuo. Magnesium oxide (Wako)
was used as a reÑectance standard. All the spectra were modi-
Ðed in terms of the KubelkaÈMunk function. X-Ray photo-
electron spectra (XPS) of the catalysts were acquired with an
ULVAC-PHI 5500MT system equipped with a hemispherical
energy analyser. The catalyst, pretreated in air at 823 K for 2
h, was mounted on indium foil and then transferred to an
XPS analyser chamber. The residual gas pressure in the
chamber during data acquisition was less than 1 ] 10~8 Torr.
2
5
2
3
0 cm3 of the solution. The V O /SiO -SG catalysts were pre-
2
5
2
pared as follows: tetraethoxysilane (e.g. 49 mmol) in ethanol (5
cm3) was mixed with vanadium oxyacetylacetonate (e.g. 1
mmol), dimethyl formamide (50 mmol), and water (10 cm3).
After gelation, the catalysts was dried and then calcined in air
at 823 K for 2 h. The amount of vanadium in the catalyst after
calcination was measured by atomic absorption and X-ray
Ñuorescence analyses. The catalyst mounted in the reaction
tube was again subjected to in situ calcination in a stream of
air at 823 K for 2 h.
Highly puriÐed methane ([99.9%, including less than
.01% of ethane) was used without further puriÐcation. The
0
as-supplied ethane, containing a small amount of CO and
2
water vapour, was slowly passed over a column Ðlled with
pulverized fused potassium hydroxide and molecular sieve 5A,
and then fed to the reactor. Helium, propane, and oxygen
were used as received. Other organic reagents were obtained
from Wako Co., Ltd., and distilled just before use.
The spectra were measured at ca. 173 K using Mg-Ka
radi-
Apparatus and procedures
1, 2
ation (15 kV, 400 W). The electron take-o† angle was set at
45¡. Binding energies were referenced to the Si 2p level of the
The reaction was carried out using an upstream Ñow-type
reactor equipped with a preheater and a window for UV irra-
diation (10 mm ] 20 mm, inner thickness: 1.0 mm) as
described previously.16 A quartz glass tube was Ðtted over the
whole reactor in order to maintain the temperature of the
catalyst bed uniformly by passing heated air around the
reactor. The temperature of the irradiated surface of the cata-
lyst was monitored by a thermocouple inserted directly into
the catalyst bed. We conÐrmed the absence of the inÑuence of
this thermocouple on the yields of the products. There were
only small variations in the temperature between the top and
the bottom of the catalyst bed (ca. 15 K). UV irradiation was
carried out with a high-pressure mercury vapour lamp (200
W, arc length; 75 mm) with a water Ðlter. Chemical actinom-
etry using iron(III) oxalate revealed that the number of
photons irradiated to the catalyst bed was 3.3 ] 10~7 einstein
s~1 for 250È500 nm, 0.9 ] 10~7 einstein s~1 for 250È300 nm,
for both V O /SiO -IW and SG catalysts (0.025 g). Typical
catalyst support, SiO . The apparent shape and size of the
2
catalyst particles were observed with a Hitachi S-2500 CX
scanning electron microscope (SEM). The Raman spectra
were obtained with an NEC GLG3260 Ar` ion laser (514.5
nm, 50 mW) and a Jobin-Yvon T-64000 spectrometer with
CCD detector. The scattered light was collected in a back-
scattering geometry. The accumulation time was 300 s. The
sample was treated under 150 Torr of oxygen at 823 K for 2 h,
evacuated at room temperature for 30 min, and then sealed in
a tubular Raman cell made of Pyrex glass (8 mm od).
Results
E†ect of method of catalyst preparation
The activities of various silica-supported group 5 and 6 metal
oxide catalysts towards the photo-assisted oxidation of
methane and ethane was examined at elevated reaction tem-
perature, 493 K. Table 1 shows the yields of products from
methane and ethane. The runs lasted 2 h for methane and 1 h
for ethane. In the photooxidation of methane, silica-supported
vanadium and molybdenum oxide catalysts showed signiÐcant
activity. In particular, a silica-supported vanadium oxide cata-
lyst (IW, 1.0 mol% V) was most e†ective, yielding 62 lmol of
methanal together with 28 lmol of carbon oxides. The yield,
based on methane feed, and the selectivity for methanal were
2
5
2
reaction conditions were: amount of catalyst, 0.025 g; feed
rate of alkane, 7.5 mmol h~1 (180 cm3 h~1);
alkane : oxygen : helium \ 3 : 1 : 10 in molar ratio; W /F, 0.71
g h mol~1 (3.0 ] 10~5 g h cm~3). Liquid products were col-
lected by passing through an iceÈwater trap at 273 K for the
products of oxidation of methane, an acetone trap at 195 K
for those from ethane and a methanol trap at 195 K for those
from propane. Gaseous products were collected into gas-
sampling bags.
0.41 mol% and 69 mol%, respectively. For the photooxidation
of ethane, both the silica-supported vanadium and molyb-
denum oxide catalysts showed high efficiency. SigniÐcant
changes in the formation rate of the products were not
observed during prolonged runs for 9 h, with both molyb-
denum and vanadium oxide catalysts. Note that, with molyb-
denum oxide catalysts, formation of methanal, the product via
the dissociation of the carbonÈcarbon bond, was signiÐcant.
The e†ect of the loading level of vanadium species on the
photooxidation of methane was examined using V O /SiO -
IW catalysts at 493 K. As shown in Fig. 1(a), the rate of for-
mation of methanal had a maximum value of 34 lmol h~1
(0.15 lmol h~1 m~2, see also Table 2) at 0.6 mol% V, which
corresponds to 76 mol% selectivity and 0.48 mol% one-pass
yield. Turnover frequency (TOF) was 8.2 mol (mol V)~1 h~1.
Further loading drastically decreased the rate of formation.
The reaction of ethane also showed a similar dependence [Fig.
1(b)]. The reaction with the IW catalyst with 0.6 mol% V
a†orded the desired products in the highest yields for both
reactants. The rate of formation of carbon oxides also showed
similar behaviour.
Analysis
The products were analysed by gas chromatography: a
Porapak-Q column at 353 K with an FID for C ÈC hydro-
carbons, and at 433 K for methanol, ethanal and ethanol; a
Porapak-N column at 433 K with an FID for acetone and
propanal; a TSG-1 on Shimarite-F column (from Shimadzu
Co. Ltd.) at 393 K with a TCD for methanal; a molecular
sieve 5A column at 323 K with a TCD for oxygen; and an
1
4
2
5
2
active carbon column at 323 K with a TCD for H , CO and
2
CO .
2
The atomic absorption analysis was carried out using an
AA-8200 atomic absorption/Ñame emission spectrometer
(
Nippon Jarrell Ash Inc.). X-Ray Ñuorescence (XRF) analyses
were carried out using KEVEX EDX-771 apparatus in the
Environment Preservation Center of Kyoto University. The
BET surface area of the catalyst was measured with a
BELSORP 28, a microprocessor-controlled automatic system
using N at 77 K, from BEL Japan Inc. An X-ray di†raction
2
study was performed using a Shimadzu XD-D1 with Cu-Ka
1772
J. Chem. Soc., Faraday T rans., 1998, V ol. 94

View Article Online
Table 1 Photo-assisted catalytic oxidation of methane and ethane using silica-supported group 5 and 6 metal oxide catalysts at 493 Ka
yield/lmol
catalyst
(prep., mol%[metal]b)
alkane
(conversion %)
time/
h
selectivity for
aldehydes (%)
run
CH CHO
HCHO
CO
CO2
3
1
2
3
4
5
6
7
8
9
V O /SiO (ED, 2)
CH (0.48)
2
2
2
2
2
1
1
1
1
È
È
È
È
È
41
62
42
3
60
62
46
tr.
26
9
7
22
0
12
16
4
0
6
0
1
2
0
tr
12
6
tr
4
7
2
83
69
82
0
72
93
98
92
2
5
2
4
V O /SiO (IW, 1)
CH (0.60)
2
5
2
4
V O /SiO (SG, 1)
CH (0.37)
2
5
2
4
Nb O /SiO (IW, 2)
MoO /SiO (ED, 1)
CH (0)
2
5
2
4
CH (0.24)
3
2
4
V O /SiO (ED, 2)
C H (0.65)
2
5
2
2 6
V O /SiO (SG, 1)
C H (0.89)
2
5
2
2 6
MoO /SiO (ED, 1)
C H (0.77)
7
tr
3
2
2 6
WO /SiO (IW, 1)
C H (0.04)
B100
3
3
2 6
a Amount of catalyst 0.025 g, W /F 0.71 g h mol~1, alkane feed rate 7.5 mmol~1, alkane : O : He \ 3 : 1 : 10. b Preparation method and loading
2
level in preparation.
In order to investigate the e†ect of the pretreatment of the
silica support, the reactions were performed using V O /SiO -
IW (0.6 mol% V) catalysts prepared using silica calcined at
various temperatures. As shown in Fig. 2, the amounts of
methanal from methane decreased at higher calcination tem-
perature, whereas those of ethanal from ethane increased. The
V O /SiO (calcined at 1023 K)-IW (0.6 mol% V) catalyst
2
5
2
2
5
2
yielded 85 lmol of ethanal (0.43 lmol h~1 m~2). Provided
that the reaction requires one photon to produce one mol-
ecule of ethanal from ethane, this formation rate corresponds
to ca. 7% of the number of photons (ca. 250È500 nm) irradi-
ated into the catalyst layer as measured by chemical actinom-
etry (vide supra). The yields of carbon oxides from methane
also decreased with the use of calcined silica, thus the selec-
tivity for methanal was nearly unchanged. On the other hand,
the amounts of carbon monoxide and dioxide from ethane
were almost constant, irrespective of the calcination tem-
perature of the silica, implying improved selectivity for ethanal
by high-temperature pretreatment.
Since the catalysts prepared by the solÈgel method were
expected to have uniform surface vanadium species, their
Fig. 1 E†ect of loading level of V on photooxidation of (a) methane
and (b) ethane using V O /SiO -IW catalysts: Rate of formation of
2
5
2
(
j) ethanal, (g) methanal, (…) carbon monoxide and (=) carbon
dioxide. Amount of catalyst, 0.025 g; feed rate of alkane, 7.5 mmol
h~1; alkane : oxygen : helium \ 3 : 1 : 10 in molar ratio; W /F, 0.71 g h
mol~1; 493 K.
Table 2 BET surface areas of selected silica-supported vanadium
oxide catalysts
catalyst (mol%)
BET surface area/m2 g~1
IW 6.6
IW 0.6a
IW 0.6
SG 4.0
SG 2.0
SG 1.0
SG 0.2
230
200
220
480
520
580
750
Fig. 2 E†ect of calcination temperature of silica gel used for the
preparation of the V O /SiO -IW (0.6 mol% V) catalysts on photoox-
2
5
2
idation of (a) methane and (b) ethane, Rate of formation of (j)
ethanal, (g) methanal, (…) carbon monoxide and (=) carbon dioxide.
Amount of catalyst, 0.025 g; feed rate of alkane, 7.5 mmol h~1;
alkane : oxygen : helium \ 3 : 1 : 10 in molar ratio; W /F, 0.71 g h
mol~1; 493 K.
a Prepared using silica calcined in air at 1023 K for 2 h.
J. Chem. Soc., Faraday T rans., 1998, V ol. 94
1773

View Article Online
e†ects on the reaction of methane and ethane were examined.
As shown in Table 1, runs 3 and 7, the SG catalyst generally
showed high activities, especially towards the reaction of
ethane. Fig. 3 shows the e†ect of increasing the concentrations
of vanadium on the reactions of methane and ethane. For
both reactions, the catalyst containing 1.0 mol% V showed
the maximum yield of aldehydes. Further increasing the vana-
dium concentrations resulted in decreasing conversions, espe-
cially markedly in the case of methane. TOF of the catalyst
containing 0.2 mol% V exhibited a much higher value, indi-
cating that the catalysts with lower V concentration were suit-
able for both reactants. In every case examined, the selectivity
for aldehyde formation was [90%. Most of by-products were
carbon oxides. As can be seen in Fig. 3(c), the activities for the
reactions with ethane per unit surface area of the SG catalysts
were at best 0.11 lmol h~1 m~2, and were generally lower
than those of the IW catalysts [because of lack of BET surface
area, the data for SG (4.9 mol%) catalysts were not included
in Fig. 3(c)].
E†ects of reaction conditions, addition of water and irradiation
wavelength
As demonstrated previously,20 the temperature crucially
a†ected the reactions. In the photooxidation of methane using
a silica-supported vanadium oxide catalyst (IW, 1.0 mol% V),
the reaction at \400 K gave only trace amounts of the pro-
ducts. The yields of the products increased with increasing
temperature. This is probably because of promotion of the
desorption of water and the products from the surface. The
yield of methanal showed a maximum at ca. 500 K, and
decreased markedly at higher temperatures. The yields of
carbon oxides also decreased above 500 K. In addition, for-
mation of carbonaceous materials on the surface during reac-
tions at high temperatures was not observed, since air
oxidation of these used catalysts at 823 K for 1 h produced
only a trace amount of carbon dioxide. These results indicate
that the decrease in the yield of methanal above 500 K was
not due to secondary deep oxidation of methanal. On the
other hand, the ethane reactivity showed a less marked depen-
dence on the reaction temperature. The reaction at ca. 350 K
yielded 4 lmol h~1 of ethanal, and the rate gradually
increased with increasing temperature, even at [500 K.
The e†ects of water vapour on the photooxidation of
methane and ethane were examined using the V O /SiO (IW,
2 5 2
.6 mol% V) catalyst, as shown in Fig. 4. The amount of
0
methanal from methane decreased drastically on introduction
of water vapour. On the other hand, the amount of ethanal
from ethane improved slightly in the presence of 9 Torr of
water vapour. Further increase in the vapour pressure of
water decreased the yields.
The photo-assisted oxidation of propane produced propa-
nal, acetone, ethanal and carbon dioxide, the main products
being propanal and ethanal. The rate of formation of carbon
dioxide was 6 lmol h~1, thus the selectivity for oxygen-
containing chemicals was [90%. As shown in Fig. 5, the reac-
tion of propane also showed a similar dependence on water
Fig. 3 E†ect of content of vanadium species in the V O /SiO -SG
2
5
2
catalysts on photooxidation of methane and ethane: (a) Rate of for-
mation, (b) TOF and (c) rate of formation per unit surface area of (j)
methanal from methane, (…) ethanal from ethane. Amount of catalyst,
Fig. 4 E†ect of water vapour on photooxidation of methane and
ethane using the V O /SiO -IW (0.6 mol% V) catalyst: Rate of for-
2
5
2
mation of (j) methanal from methane, (…) ethanal from ethane.
Amount of the catalyst, 0.025 g; feed rate of alkane, 7.5 mmol h~1;
alkane : oxygen : helium \ 3 : 1 : 10 in molar ratio; W /F, 0.71 g h
mol~1; 493 K.
0
.025
g;
feed
rate
of
alkane,
7.5
mmol
h~1;
alkane : oxygen : helium \ 3 : 1 : 10 in molar ratio; W /F, 0.71 g h
mol~1; 493 K.
1774
J. Chem. Soc., Faraday T rans., 1998, V ol. 94

View Article Online
UV-35 Ðlter (passes j [ 350 nm) a†orded a small amount of
ethanal.
Analysis of the catalysts
Table 2 shows the BET surface areas of selected catalysts. For
V O /SiO -IW catalysts of loading levels below 6.6 mol%,
2
5
2
there were no signiÐcant changes in surface areas; they
showed slightly lower surface areas than that of the parent
silica support. The use of silica calcined at high temperature
slightly decreased the surface area. These results indicate that
surface area was not a signiÐcant factor. On the other hand,
the SG catalyst with 0.2 mol% V showed a large surface area
of 750 m2 g~1. However, increases in the concentration of
vanadium oxide greatly decreased the surface areas. These
marked changes in surface area would a†ect the surface con-
centration and dispersion of vanadium species and, eventually,
the rate and selectivity of the reaction (vide infra).
Fig. 5 E†ect of water vapour on photooxidation of propane using
The catalysts were analysed by means of Raman spectros-
copy. Selected spectra of the IW catalysts are shown in Fig. 7.
The V O /SiO -IW (0.6È1.7 mol% V) catalyst showed a
the V O /SiO -IW (0.6 mol% V) catalyst. Amount of catalyst,
2
5
2
0
.025
g;
feed
rate
of
propane,
7.5
mmol
h~1;
propane : oxygen : helium \ 3 : 1 : 10 in molar ratio; W /F, 0.71 g h
2
5
2
mol~1; 493 K; 1 h.
Raman band at 1043 cm~1, corresponding to isolated vana-
dium oxide (VO ) surface species,21,22 whereas, with the
V O /SiO -IW (6.6 mol% V) catalysts the formation of V O
microcrystals (996 cm~1, terminal VxO bond; 703 cm~1,
VwOwV vibrations) was observed, in addition to VO -type
4
vapour to that of the reaction of ethane. The reaction was 9.2
torr of water vapour increased the rate of formation of the
products, especially ethanal and carbon dioxide. Further addi-
tion of water vapour decreased the conversion of propane, but
there were changes in the selectivities. The selectivity for
ethanal increased further, whereas the formation of carbon
dioxide decreased. These changes in the selectivities suggest
variations in the active oxygen species or surface state of the
catalysts (vide infra).
2
5
2
2 5
4
The e†ects of irradiation wavelength on photooxidation of
methane and ethane using the V O /SiO -IW (0.6 mol% V)
2
5
2
catalyst were investigated. The results are shown in Fig. 6. The
amount of methanal formed by the photooxidation of
methane decreased drastically by cutting o† the light of wave-
length shorter than 310 nm using a UV-31 Ðlter. This indi-
cates that, for the reaction of methane, irradiation with light
of wavelength \300 nm was essential. On the other hand, the
reaction of ethane when using a UV-31 Ðlter produced a sig-
niÐcant amount of ethanal. Even the reaction when using a
Fig. 6 E†ect of irradiation wavelength on photooxidation of
methane and ethane using the V O /SiO -IW (0.6 mol% V) catalyst.
2
5
2
Amount of the catalyst, 0.025 g; feed rate of alkane, 7.5 mmol h~1;
alkane : oxygen : helium \ 3 : 1 : 10 in molar ratio; W /F, 0.71 g h
mol~1; 493 K.
Fig. 7 Raman spectra under dehydrated conditions of the
V O /SiO -IW and -SG catalysts. (\) prepared using silica calcined at
2
5
2
1023 K.
J. Chem. Soc., Faraday T rans., 1998, V ol. 94
1775

View Article Online
Fig. 8 XRD patterns of the V O /SiO -IW and -SG catalysts. (\)
2
5
2
prepared using silica calcined at 1023 K.
species. Raman bands around 975 and 800 cm~1 were assign-
ed to silanol groups and siloxane bridges, respectively.22 Two
weak Raman bands at around 1075 and ca. 930È910 cm~1
were observed, characteristic of SiO~ and Si(O~) functional-
2
ities.22 Hence, peaks other than at 1043, 996 and 703 cm~1
can be assigned to the silica support. The spectra of
V O /SiO -SG (1.0 mol% V) catalyst also showed a Raman
2
5
2
Fig. 10 UV di†use reÑectance spectra of V O /SiO -IW catalysts
band at 1043 cm~1 and none of bands corresponding to
2
5
2
under dehydrated conditions
V O , indicating the formation of only four-coordinated
2
5
vanadium oxide species.
used in this study.23 For all the catalysts examined, most of
Fig. 8 shows XRD patterns of the IW and SG catalysts. The
formation of crystalline V O was observed with the catalyst
their surface was composed of SiO , together with small
2
amount of carbon and vanadium. As shown in Fig. 9(b), which
2
5
with 6.6 mol% V. On the other hand, the spectra of the cata-
lysts of lower loading level did not show any di†raction pat-
terns other than the common patterns of crystalline silica. The
SG catalysts examined also did not show any patterns except
for those of crystalline silica, indicating the absence of micro-
crystals of V O .
shows the relative molar ratio of V and Si vs. the overall
atomic ratio (N /N ), the IW catalysts of 0.2 mol% V showed
V
Si
the highest vanadium observability. Combined with the
results of XRD and Raman analyses, the formation of rather
accumulated vanadium species on the surface with IW cata-
lysts of high loading level is demonstrated, whereas the IW
catalysts of low loading level and the ED catalyst seem to
have more dispersed species on their surface. Note that there
were no signiÐcant di†erences in the surface atomic ratio of V
to Si between IW catalysts of 0.6 mol% V prepared using
silica with or without calcination. They showed the same
atomic ratio of vanadium to silicon (0.84 atom%). Fig. 9(c)
and (d) show the atomic ratios of each of the elements and the
relative ratio of the molar ratio of V and Si for SG catalysts,
respectively. There were similar tendencies in the vanadium
observability to those of the IW catalysts, indicating the for-
mation of rather accumulated vanadium species with SG cata-
lysts of high vanadium concentration.
2
5
XPS analysis of the catalysts revealed the presence of Si, O,
C and V on their surface. There were no signs of other ele-
ments by wide-range survey analysis. Fig. 9(a) shows the
atomic ratio of each of the elements on the surface, deter-
mined by peak intensities of Si 2p, O 1s, C 1s and V 2p vs.
overall atomic ratio (N /N ). These atomic ratios were calcu-
V
Si
lated based on sensitivity factors reported for the spectrometer
Fig. 10 shows selected UV di†use reÑectance spectra of the
catalysts. A major absorption band around 260 nm was
detected for all the catalysts. With V O /SiO -IW (1.7 mol%
2
5
2
V) and -IW (1.0 mol% V) catalysts a shoulder peak at ca. 300
nm was also observed. On the other hand, the catalyst with
6.6 mol% V exhibited a strong absorption band at ca. 330 nm,
in addition to that at 260 nm. Note that there was weak
absorption at 300 nm in the spectrum of the 0.6 mol% V cata-
lyst using calcined silica, whereas the spectrum of the catalyst
using silica without calcination showed no such shoulder.
Discussion
One of the probable reaction mechanisms, based on that pro-
posed by Kazansky and coworkers5 on the basis of HartreeÈ
Fock calculations is shown in Scheme 1. As a Ðrst step, the
surface four-coordinated vanadium oxide having a VxO
double bond is excited by irradiation to form a charge-
transfer complex. An oxygen is trapped by a positive hole to
produce a strongly electrophilic O~ radical species, and an
Fig. 9 (a) Surface composition and (b) V/Si atomic ratio of IW (open
symbols) and ED (Ðlled symbols) catalysts, (c) surface composition
and (d) V/Si atomic ratio of SG catalysts, measured by XPS.
1776
J. Chem. Soc., Faraday T rans., 1998, V ol. 94

View Article Online
suitable surface state of the catalysts and/or the conditions
between the reaction of methane and other light alkanes. Only
very highly isolated vanadium oxide surface species, which
absorb light of wavelength \300 nm, would show catalytic
activity towards the selective photooxidation of methane,
whereas rather accumulated vanadium surface species, not
crystalline V O , would activate other light alkanes.
2
5
On the other hand, the reaction of methane was inhibited
by the addition of water, whereas the reaction of ethane was
sometimes slightly promoted. In addition, in the photooxida-
tion of propane, the rate of formation of the products
increased greatly in the presence of a small amount of water,
and changes in selectivities were observed. In particular, the
yields of carbon dioxide and ethanal were markedly pro-
moted. Further addition of water vapour, conversely,
decreased the yields. These results suggest changes in the
active oxygen species and/or surface state which governed the
reactions.
These phenomena were most likely due to OH radicals on
the surfaces. Enhancement of photoreactions by water vapour
have been previously observed e.g. on titanium oxide.26
Similar dependences of the oxidation rate on humidity have
been reported and explained as being the result of competitive
absorption of water and reactants.26 On the other hand, the
addition of water would cause surface hydration, which was
reported to increase the coordination number of the vana-
dium species and weaken the VxO double bond.27 Provided
that, for the reaction with methane, the presence of four-
coordinated species on the surface was crucial, and surface
OH radicals did not show signiÐcant activity, then it would be
reasonable for the addition of water to retard markedly the
formation of methanol from methane.
Scheme 1
electron transfers to the vanadium. Such excitation has been
generally accepted for the photooxidation of various
alkenes.24 Some of activated species return to the ground state
by recombination of the hole and the electron. An alkane is
adsorbed on the photoactivated species and its CwH bond is
activated. Molecular oxygen would adsorb onto an electron-
rich site of the intermediate. Abstraction of another hydrogen
by an adsorbed oxygen species would produce aldehydes,
together with reoxidation of the surface species. At relatively
high reaction temperature, transformation of the intermediate
to aldehyde, desorption of the products and water, and reoxi-
dation of the vanadium surface species would be promoted.
According to the catalytic runs and XRD study, formation
of crystalline V O seems not to have a positive e†ect in the
2
5
present reactions. Studies of the Raman spectra indicate that
in general four-coordinated vanadium oxide species seem to
be necessary. In addition, XPS and catalytic runs suggest that
formation of more dispersed species on the surface deÐnitely
a†ects the yields of aldehydes.
Conclusions
The results in Fig. 3 indicate that SG catalysts with a lower
concentration of vanadium species yield aldehydes in higher
yields, although the activities per unit surface area were gener-
ally lower than those of the IW catalysts. Since SG catalysts
with lower V concentration showed higher surface areas, the
density of the surface vanadium species would be decreased
with decreasing concentration. In addition, the XPS study
suggests the formation of rather accumulated vanadium
species on the surface of the catalysts with high vanadium
concentration [Fig. 9(d)]. These results suggest the e†ec-
tiveness of vanadium species in low density for aldehyde for-
mation.
However, the reactions with methane and higher alkanes
are sometimes di†erently a†ected by the reaction conditions
and catalyst preparation. In general, the reaction of methane
was apt to be very sensitive, and very delicate tuning of the
surface species seems to be required. For example, although
calcination of the silica support showed completely di†erent
e†ects between the reactions of methane and those of ethane,
no signiÐcant di†erences were found in the spectroscopic
study. The BET surface area of the catalyst using non-calcined
silica and that using calcined silica were 220 and 200 m2 g~1,
respectively. These catalysts showed similar surface atomic
ratios of V to Si as measured by XPS. There were no signiÐ-
cant di†erences in the Raman spectra. The only di†erence was
in the UVÈVIS di†use reÑectance spectra. As shown in Fig.
Selective photooxidation of light alkanes, mainly methane and
ethane, into the corresponding aldehydes was achieved using
silica-supported vanadium oxide catalysts under UV irradia-
tion at ca. 500 K. For the photooxidation of methane, the
V O /SiO -IW (0.6 mol% V) catalyst showed the highest
2
5
2
activity. For the reaction with ethane, the IW catalyst pre-
pared using calcined silica showed promoted activity. The
catalysts prepared by the solÈgel method also showed activity,
especially for the reaction of ethane, although the activity per
unit surface area was generally lower than for the IW cata-
lysts. While the reaction of methane was inhibited by the addi-
tion of water vapour, the photooxidation of ethane and
propane was promoted in the presence of a controlled amount
of water vapour. The reaction with methane required UV irra-
diation of wavelength \310 nm, whereas the reaction with
ethane or propane proceeded on irradiation at longer wave-
length. According to the surface analysis of the catalysts, only
isolated four-coordinated vanadium oxide surface species were
considered to show catalytic activity towards the photooxida-
tion of methane, while rather accumulated vanadium surface
species were shown to be active for the photooxidation of
other light alkanes.
The authors thank Professor Zempachi Ogumi, Dr. Minoru
Inaba and Mr. Hiroyuki Yoshida of Kyoto University for
instruction in the Raman spectroscopy. The authors are also
grateful to Mr. Yoshiji Honda of the Environment Preser-
vation Center of Kyoto University for XRF analyses.
10, a shoulder absorption at 300 nm was characteristic for the
catalyst using calcined silica. Previous studies of UV spectra25
suggest that a shoulder absorption at 300 nm is responsible
for rather aggregated non-octahedral vanadium species. For-
mation of such aggregated species is reasonable, since calcina-
tion at high temperature would reduce the number of surface
hydroxy groups.
References
1
2
P. Pitchai and K. Klier, Catal. Rev.-Sci. Eng., 1986, 28, 13.
R. A. Periana, D. J. Taube, E. R. Evitt, D. G. Loffler, P. R.
Wetrcek, G. Voss and T. Masuda, Science, 1993, 259, 340.
T. Nishiguchi, K. Nakata, K. Takai and Y. Fujiwara, Chem. L ett.,
1992, 1141.
Ž
Combined with the dependence on the irradiation wave-
length, we can conclude that there are di†erences in the most
3
J. Chem. Soc., Faraday T rans., 1998, V ol. 94
1777

View Article Online
4
5
6
7
H. F. Liu, M. Iwamoto and J. H. Lunsford, J. Chem. Soc., Chem.
Commun., 1982, 78.
18 K. Wada, K. Yoshida, Y. Watanabe and T. Suzuki, J. Chem. Soc.,
Chem. Commun., 1991, 726.
19 K. Wada, K. Yoshida, T. Takatani and Y. Watanabe, Appl.
Catal. A, 1993, 99, 21.
20 K. Wada and Y. Watanabe, in Methane and Alkane Conversion
Chemistry, ed. M. M. Bhasin and D. W. Slocum, Plenum, New
York, 1995, p. 179.
S. L. Kaliaguine, B. N. Shelimov and V. B. Kazansky, J. Catal.,
1
978, 55, 384.
M. D. Ward, J. F. Brazdil, S. P. Mehandru and A. B. Anderson,
J. Phys. Chem., 1987, 91, 6515.
N. Djeghri, M. Formenti, F. Juillet and S. J. Techner, Faraday
Discuss. Chem. Soc., 1974, 58, 185.
21 S. S. Chen, I. E. Wachs, L. L. Murrel, L. Wang and W. K. Hall, J.
8
9
K. R. Thampi, J. Kiwi and M. Gra
M. Gra
128.
Ž tzel, Catal. L ett., 1989, 109.
Phys. Chem., 1984, 88, 5831.
Ž
tzel, K. R. Thampi and J. Kiwi, J. Phys. Chem., 1989, 93,
22 N. Das, H. Eckert, H. Hu, I. E. Wachs, J. F. Bwalzer and F. J.
Feher, J. Phys. Chem., 1993, 97, 8240.
4
1
1
1
1
0
1
2
3
M. Daroux, D. Klvana, M. Duran and M. Bideau, Can. J. Chem.
23 F. Moulder, W. F. Stickle, P. E. Sobol and K. D. Bomben, Hand-
book of X-ray Photoelectron Spectroscopy, Perkin-Elmer Co.,
Eden Prairie, 1992.
Eng., 1985, 63, 668.
S. Takenaka, T. Kuriyama, T. Tanaka, T. Funabiki and S.
Yoshida, J. Catal., 1995, 155, 196.
24 M. Anpo and H. Yamashita, in Surface Photochemistry, Series in
Photoscience and Photoengineering, vol. 1, ed. M. Anpo, Wiley,
Chichester, 1997, p. 117.
K. Marcinkowska, S. Kaliaguine and P. C. Roberge, J. Catal.,
1
984, 90, 49.
K. Marcinkowska, S. Kaliaguine and P. C. Roberge, Proceedings
of the 9th Congress on Catalysis, ed. M. J. Phillips and M.
Ternan, 1988, vol. 4, p. 521.
25 G. Lischke, W. Hanke, H. G. Jerschkewitz and G. ^hlmann, J.
Catal., 1985, 91, 54.
26 T. N. Obee and R. T. Brown, Environ. Sci. T echnol., 1995, 29,
1223.
1
1
4
5
M. Volodin and M.BolÏshov, Kinet. Katal., 1993, 34, 127.
T. Suzuki, K. Wada, M. Shima and Y. Watanabe, J. Chem. Soc.,
Chem. Commun., 1990, 1059.
27 S. Yoshida, T. Tanaka, Y. Nishimura, H. Mizutani and T. Fun-
abiki, in ref. 13, vol. 3, p. 1473.
1
6
7
K. Wada, K. Yoshida, Y. Watanabe and T. Suzuki, Appl. Catal.,
1
991, 74, L1.
1
K. Wada, K. Yoshida and Y. Watanabe, J. Chem. Soc., Faraday
T rans., 1995, 91, 21.
Paper 8/00938D; Received 3rd February, 1998
1778
J. Chem. Soc., Faraday T rans., 1998, V ol. 94