Carbon monoxide as an intermediate product in the photocatalytic steam reforming of methane with lanthanum-doped sodium tantalate

Article abstract of DOI:10.1039/d1cy00264c

Photocatalytic steam reforming of methane (PSRM) has been studied as an attractive method to produce hydrogen by utilizing photoenergy like solar energy at around room temperature with metal-loaded photocatalysts, where methane and water are selectively converted to carbon dioxide and hydrogen. In the present study, we used a PSRM system using a flow reactor at around room temperature to yield the partially oxidized product, carbon monoxide (CO). It was found that some La-doped NaTaO3 samples can produce carbon monoxide constantly in addition to hydrogen and carbon dioxide. Among the prepared samples, a La(2 mol%)-doped NaTaO3 photocatalyst without any cocatalyst exhibited the highest photocatalytic activity and the highest CO selectivity of 24%. The CO yield depended on the photocatalysts and the reaction conditions. Suitable reaction conditions for CO yield were high light intensity, a higher flow rate, and a moderately high methane/water ratio. Some additional reaction tests revealed that water gas shift (WGS) can take place as an undesirable successive reaction, i.e., the produced carbon monoxide can successively react with water to form carbon dioxide, which would restrict the CO yield significantly.

Full text of DOI:10.1039/d1cy00264c

Catalysis
Science &
Technology
View Article Online
View Journal
PAPER
Carbon monoxide as an intermediate product in
the photocatalytic steam reforming of methane
with lanthanum-doped sodium tantalate†
Cite this: DOI: 10.1039/d1cy00264c
ab
a
a
Wirya Sarwana, Akihiko Anzai, Daichi Takami,
ac
ac
Akira Yamamoto
and Hisao Yoshida
*
Photocatalytic steam reforming of methane (PSRM) has been studied as an attractive method to produce
hydrogen by utilizing photoenergy like solar energy at around room temperature with metal-loaded
photocatalysts, where methane and water are selectively converted to carbon dioxide and hydrogen. In the
present study, we used a PSRM system using a flow reactor at around room temperature to yield the
partially oxidized product, carbon monoxide (CO). It was found that some La-doped NaTaO
produce carbon monoxide constantly in addition to hydrogen and carbon dioxide. Among the prepared
samples, a La(2 mol%)-doped NaTaO photocatalyst without any cocatalyst exhibited the highest
3
samples can
3
Received 11th February 2021,
Accepted 14th June 2021
photocatalytic activity and the highest CO selectivity of 24%. The CO yield depended on the photocatalysts
and the reaction conditions. Suitable reaction conditions for CO yield were high light intensity, a higher
flow rate, and a moderately high methane/water ratio. Some additional reaction tests revealed that water
gas shift (WGS) can take place as an undesirable successive reaction, i.e., the produced carbon monoxide
can successively react with water to form carbon dioxide, which would restrict the CO yield significantly.
DOI: 10.1039/d1cy00264c
rsc.li/catalysis
CH4 þ 2H2O → CO2 þ 4H2 ΔG°298K ¼ 113:5 kJ mol−
1
(3)
Introduction
4
Methane (CH ) is a ubiquitous substance obtained from
underground and biomass. Considering it as a carbon
resource and not as a fuel, the conversion of methane to
other industrially valuable chemical compounds such as
These reactions are highly endergonic except for WGS and
need a huge amount of energy. Conventionally, methane
combustion supplies heat for the required energy and thus
the reaction temperature increases, which necessitates
several requirements such as inhibition of carbon deposition
and an expensive heat-resistant reactor. Therefore, a new
reaction system that can work at low temperature is highly
desirable for these endergonic reactions.
Photocatalytic reactions have been shown as a promising
way to promote thermodynamically difficult reactions by
utilizing solar light as photoenergy and thus the reaction can
occur even at room temperature. Photocatalytic steam
carbon monoxide (CO) and hydrogen (H
researchers.
2
) is very attractive to
The mixture of CO and H2 known as syngas
could be catalytically obtained in steam reforming of methane
SRM, eqn (1)). For the industrial production of hydrogen,
successive water gas shift reaction of CO gives additional H2
and CO (WGS, eqn (2)) and the overall reaction gives
hydrogen and CO without CO production (eqn (3)).
1
–7
(
2
2
CH4 þ H2O → CO þ 3H2 ΔG°298K ¼ 142:1 kJ mol− 1 (1)
4 2
reforming of methane (PSRM) can convert CH and H O directly
to H and CO (eqn (3)) even under mild conditions although it
CO þ H2O → CO2 þ H2 ΔG°298K ¼ −28:6 kJ mol−
1
(2)
2
2
8,9
is endergonic. This reaction was originally developed by our
1
0,11
group using a Pt-loaded TiO
2
(Pt/TiO
2
) photocatalyst
and a
1
0,12,13
Pt-loaded La-doped NaTaO photocatalyst (Pt/NaTaO :La).
3
3
a
After that, various photocatalysts have also been reported for
Graduate School of Human and Environmental Studies, Kyoto University, Kyoto
06-8501, Japan. E-mail: yoshida.hisao.2a@kyoto-u.ac.jp
14,15
16
6
b
the PSRM such as Pt/CaTiO
3
,
Pt/CaTaO
These studies have focused
only on hydrogen production, where the observed molar ratio
has been always H /CO = 4 (eqn (3)).
3
:La,
Rh/
17,18
19–21
Department of Mechanical Engineering, Sumbawa University of Technology, Olat
K Ti O ,
and Pt/β-Ga O .
2 3
2
6
13
Maras, Sumbawa, West Nusa Tenggara, 84371, Indonesia
c
Elements Strategy Initiative for Catalysts and Batteries (ESICB), Kyoto University,
2
2
Kyoto 615-8520, Japan
So far, photocatalytic conversion of methane and water
has been reported also for the production of valuable
†
Electronic supplementary information (ESI) available. See DOI: 10.1039/
d1cy00264c
This journal is © The Royal Society of Chemistry 2021
Catal. Sci. Technol.

View Article Online
Paper
Catalysis Science & Technology
2
2–27
28
29
chemicals such as methanol,
ethanol, and aldehydes.
The NTO:La samples (2 g) were soaked and stirred in an
aqueous solution of the precursor (100 ml; 0.05, 0.1 or 1 mM
at 353 K) until the water completely evaporated, and dried
overnight in an oven at 353 K. Before use, metal loaded NTO:
La was calcined at 673 K for 2 h. This was denoted M(y)/NTO:
La where y is the wt% of the metal cocatalyst. To investigate
the role of La doping, two additional La loaded samples were
also prepared. The precursor used for the loading process
In the present study, we attempted to change the selectivity
of the PSRM to produce CO that is one of the partially
oxidized products from methane and very valuable as a
chemical intermediate. In the previous study, we found that
a metal cocatalyst could change the product selectivity in
PSRM. The Rh cocatalyst loaded either on the K Ti O
2
6 13
surface or the NaTi O surface always gave H and CO2
selectively without any formation of CO.
6
13
2
1
8,30
In contrast, a Pt
3 3 2
was La(NO ) ·6H O (Nakalai Tesque, 99.9%). The 0.3 wt% La
loaded photocatalyst gave CO as a minor by-product with a
species was loaded on the surface of the NTO and NTO:La(1)
samples by the impregnation method, and these samples
were labelled La(0.3)/NTO and La(0.3)/NTO:La(1).
low selectivity of 9–10% over Pt/K Ti O (ref. 18) and Pt/
2
6 13
2 3
Ga O (ref. 31) photocatalysts. There has been, however, no
report focusing on CO production in the PSRM.
Some photocatalysts such as Ga O (Kojundo, 99.99%),
2
3
Here, we report the PSRM to produce CO by using a La-
ZnO (Kojundo, 99.99%), and TiO (Ishihara Sangyo Kaisha,
2
2
−1
doped NaTaO
3
(NaTaO
3
:La) photocatalyst without loading any
ST-01, 300 m g ) were also employed for comparison. All of
these photocatalysts were used as received without any
pretreatment.
cocatalysts, where the CO selectivity among the oxidative
products reached 24% for the first time. It is also revealed
that the low CO selectivity is due to the successive oxidation
to CO ; in other words, CO is an intermediate product in the
2
Characterization
PSRM.
The content of La in the samples was determined by X-ray
fluorescence (XRF) analysis on an EDX-8000 (Shimadzu) with
a calibration curve obtained using the samples prepared by a
conventional impregnation method.
Experimental
Catalyst preparation
The scanning electron microscopy (SEM) images were
captured using a JEOL JSM-890. The average particle size was
evaluated from the SEM images.
X-ray diffraction (XRD) patterns of the samples were
recorded at room temperature using a Shimadzu Lab X XRD-
Sodium tantalate (NaTaO3, referred to as NTO) and
lanthanum-doped sodium tantalate samples (NaTaO :La,
3
referred to as NTO:La) were mainly prepared by a flux
method. The starting materials, Na CO , Ta O (Rare
2
3
2 5
Metallic, 99.99%), and La O (Kishida, 99.99%), were mixed
2
3
6000. The radiation used was Cu Kα (40 kV, 30 mA). The
together with a flux, NaCl (Kishida, 99.5%), in an alumina
crystallite size was determined by the Scherrer equation using
the full width at half maximum (FWHM) of the diffraction
line at 2θ = 22.8° in the XRD patterns.
mortar for 15 min. The molar ratio for NaTaO :La was Na :
3
Ta : La = 100 : 100 : x, where x is from 0 to 5, and the solute
concentration in the molten mixture was 70%, i.e., the ratio
The BET specific surface area was measured by using the
was NaTaO : NaCl = 70 : 30. The mixture was placed in a
3
adsorbed amount of N on the sample surface at 77 K using
2
platinum crucible and heated by using an electric muffle
a Quantachrome Monosorb MS-21.
The diffuse reflectance (DR) UV-visible spectra were
recorded using a JASCO V-670 equipped with an integrating
−
1
furnace, where the temperature was increased at 200 K h
from room temperature to 1273 K and kept for 5 h. It was
−
1
cooled down at 100 K h to 773 K, and naturally to ambient
temperature in the furnace. The product obtained was
washed with hot distilled water (353 K, 500 ml) 3 times to
remove the flux and then dried at 353 K overnight. These
samples were labelled NTO:La(x), where x is the amount of
La (mol%) used as a dopant.
4
sphere, where BaSO was used as a reference. The bandgap
3
2
was estimated by using a Tauc plot.
Photocatalytic activity test
Photocatalytic reaction tests for the PSRM were carried out
1
3
Another sample was prepared by a solid-state method with
the same procedure as mentioned above only without using
the NaCl flux. The La doping was 2 mol%. The starting
materials were mixed, heated, and cooled, followed by
washing under the same conditions as mentioned above.
This sample was referred to as NTO:La(2)SS.
Some precious metals such as Au, Ag, Pd, and Pt and
transition metals such as Ni, Cu, and Zn were loaded as
cocatalysts on the NTO:La surface by an impregnation
method. The precursors used for the loading process were
using a fixed bed flow reactor as shown in Fig. S1.† The
photocatalyst powder was pressed under 40 MPa for 1 min,
which gave a pellet. The pellet obtained was crushed and
sifted using 25 mesh and 50 mesh sieves. The remaining
granules on the 50 mesh sieve were used for the reaction
tests. The catalyst granules (1.2 g) were introduced inside a
3
quartz reactor (60 × 20 × 1 mm ), where the irradiated area of
2
the photocatalyst was 6.0 cm . The feed gas mixture used for
this reaction test typically consisted of CH (25%) and H O
4
2
(2.4%) with argon as a carrier gas. The flow rate of the feed
−
1
HAuCl
H PtCl
4
·6H
2
O, AgNO
3
(Kishida, 99.8%), PdCl
2
(Kishida, 99%),
gas was 15 ml min . A 300 W xenon lamp (PE300BUV) was
used as a light source for photoirradiation without any
(Wako, 99.9%), Ni(NO )·6H O (Wako, 99%),
2
6
3
2
−
2
Cu(NO ) ·3H O, and Zn(NO ) ·6H O (Nacalai Tesque, 99%).
optical filter. The light intensity was 165 mW cm that was
3
2
2
3 2
2
Catal. Sci. Technol.
This journal is © The Royal Society of Chemistry 2021

View Article Online
Catalysis Science & Technology
Paper
measured in the range of 254 ± 10 nm wavelength. The
temperature of the reaction became 323 K due to the
photoirradiation. The gas produced in the reaction was
analyzed by on-line gas chromatography equipped with a
thermal conductivity detector (Shimadzu, GC-8A, TCD). The
interval of sampling was ca. 30 min. Since the sensitivity to
0.55 μm (Fig. 1e). As for the NTO:La(2)SS sample, the
morphology of the crystal was more roundish with somewhat
irregular shapes (Fig. 1f). These results are also quite similar
1
3
to the previous report.
Fig. S2† shows the XRD patterns of the non-doped NTO
and various NTO:La samples. The diffraction lines of all the
samples indicate the presence of the NaTaO3 perovskite
phase. No impurity phases related to La O or La were
CO in the argon carrier gas was low, the experimental error
2
for CO determination was relatively large.
2
2 3
detected, although this might be due to the low
concentration of the La dopant. The average crystallite sizes
of each sample were estimated and are listed in Table 1. As
Results and discussion
Characterization of the photocatalysts
3
shown, the more La dopant introduced into NaTaO , the
Fig. 1 shows the SEM images of the non-doped NTO and
various NTO:La samples prepared by the flux method and the
solid-state reaction method. The average particle sizes of
these samples estimated from these images are listed in
Table 1. The non-doped NTO sample prepared by the flux
method consisted of roundish cubic particles sized in the
range of 0.2–2.5 μm (Fig. 1a), where the average particle size
was 0.8 μm (Table 1). The La-doped samples prepared by the
flux method showed cubic crystals sized in the range of 0.04–
smaller the crystallite size obtained. The NTO:La(2)SS sample
had a similar crystallite size to the other NTO:La samples
with low La doping such as the NTO:La(0.5) and NTO:La(1)
samples. Note that the crystallite sizes determined by XRD
were much smaller than the particle size observed in the
SEM images (Table 1). This means that the crystals observed
in the SEM images were not single crystals but
1
3,16
polycrystals.
In the NTO:La(2) and NTO:La(5) samples,
the crystallite size decreased with increasing La doping while
the particle size increased. It is suggested that the larger
amount of La ions enhanced the aggregation of the particles
of the smaller crystallites.
1
0
.0 μm (Fig. 1b–e), and the average particle sizes were 0.21–
.55 μm, which were much smaller than those of the non-
doped NTO sample. This result confirmed that La doping can
3
3–35
inhibit the crystal growth of NaTaO3.
The NTO:La(1)
The BET specific surface areas of the various samples are
also listed in Table 1. As shown, the NTO:La(1) sample has
the highest specific surface area compared to the other
samples, supporting that doped lanthanum cations inhibit
the crystal growth thus increasing the specific surface area.
The increase of the specific surface area was consistent with
the increasing La-doping amount in the range of 0–1 mol%.
However, in the case of the NTO:La(2) and NTO:La(5)
samples, the specific surface areas decreased with increasing
La doping, suggesting that the aggregation of the particles
decreased the specific surface area. Sun et al. reported that a
sample showed the smallest average particle size of 0.21 μm.
However, the NTO:La(5) sample with the higher La doping
showed a slightly different morphology with a larger size of
3
higher Sr doping amount in NaTaO samples could decrease
3
6
the surface area.
Simply stated, the surface area should be inversely
proportional to the particle size. Thus, we examined the
relationship between the specific surface area and the average
particle size based on the values of the samples with 0–1% La
(
Table 1) and we confirmed an inversely proportional
correlation (Fig. S3†). Such correlation was not obtained
between the specific surface area and the crystallite size. This
means that the BET specific surface area reflects the surface
area of the polycrystal particles observed in the SEM images,
not the crystallite size. It was found that the specific surface
areas for the NTO:La(2) and NTO:La(5) samples were higher
than expected based on the correlation, suggesting that the
aggregation provides interspaces between the particles as a
certain pore structure in the NTO:La(5) sample.
These powder samples were pelletized to form granules
for the photocatalytic reaction tests. In some cases, the
specific surface areas of the granule samples were lower than
those of the powder samples but not significantly decreased
(Table 1).
Fig. 1 SEM images of the prepared samples: (a) non-doped NTO, (b)
NTO:La(0.5), (c) NTO:La(1), (d) NTO:La(2), (e) NTO:La(5), and (f) NTO:
La(2)SS.
This journal is © The Royal Society of Chemistry 2021
Catal. Sci. Technol.

View Article Online
Paper
Catalysis Science & Technology
Table 1 Structural and optical properties of the prepared samples
a
b
c
d
e
f
La content
(mol%)
Average particle size
(μm)
Crystallite size
(nm)
S
BET powder
S
BET granule
Bandgap
(eV)
2
−1
2
−1
Entry
Sample
(m g
)
(m g )
1
2
3
4
5
6
NTO
0.0
0.3
0.7
1.7
4.8
1.5
0.79
0.29
0.21
0.25
0.55
0.41
68
50
42
35
27
43
2.4
5.3
8.3
7.5
5.6
4.1
1.9
5.4
7.7
7.4
5.5
3.9
4.0
4.1
4.1
4.1
4.1
4.1
NTO:La(0.5)
NTO:La(1)
NTO:La(2)
NTO:La(5)
NTO:La(2)SS
a
b
c
The actual La dopant content measured by XRF. The average particle size estimated from the SEM images. The average crystallite size
d
e
calculated from a line width in the XRD patterns. The specific surface area of the powder sample measured by the BET method. The
specific surface area of the granule samples measured by the BET method. The bandgap estimated by a Tauc plot from the DR UV-vis spectra.
f
The DR UV-visible spectra of the samples are shown in
Fig. S4.† The bandgaps of each sample were calculated by
using a Tauc plot and are listed in Table 1. These results
online GC-TCD. The product formation rates and the
product selectivity were constant and the reaction
continuously proceeded for a long time, for at least 20
hours. The ratio of photoexcited electrons and holes that
3
suggest that La doping can enlarge the bandgap of NaTaO ,
1
3
−
+
which is consistent with the previous report.
were consumed for the product formation, R(e /h ), was
estimated from the formation rates of H , CO, and CO
which are RH2, RCO, and RCO2, respectively, according to eqn
2
2
,
Photocatalytic activity test
−
+
(
4). As shown in Fig. 2, this ratio was almost unity, R(e /h )
= 1, meaning that the products should be almost limited to
, CO, and CO . Based on these results, the CO selectivity
Fig. 2 shows the time course of the product formation rates
in the photocatalytic reaction test for the PSRM with the
H
2
2
NTO:La(2) sample in the flow of feed gas (25% CH , 2.4%
among the oxidative products, SCO(%), can be calculated
according to eqn (5) and the CO selectivity was as high as
24% with the NTO:La(2) photocatalyst. The apparent
quantum efficiency, AQE (%), defined as the ratio of the
4
H O and 72.6% Ar) at atmospheric pressure and room
2
temperature. Although the production rates of H
2 2
and CO
decreased slightly for the initial 4 hours, the production rates
−
1
of H2 and CO2 became constant at 21 and 4 μmol h ,
2
number of photons used for H formation to the number of
respectively. It was found that CO was also produced in the
incident photons that can be absorbed by the photocatalyst,
was estimated to be 0.12%. No products were obtained
addition of H2 and CO . The CO production rate was
2
−
1
constant from the start at 1 μmol h . Other gaseous
under dark conditions or without employing
a
oxidation products such as ethane were not detected with the
photocatalyst. These facts obviously indicate that the
reaction takes place photocatalytically.
−
+
R(e /h ) = RH2/(3 × RCO + 4 × RCO2
)
(4)
(5)
S
CO(%) = 100 × RCO/(RCO + RCO2
)
Fig. 3 shows the photocatalytic activity of the various
samples. All the prepared La-doped NaTaO samples without
3
any cocatalysts showed photocatalytic activity to produce CO
in the PSRM. As shown in Fig. 3A, the NTO:La(x) samples
exhibited higher photocatalytic activities to produce H
and CO than the non-doped NTO sample (Fig. 3Aa–e). It is
confirmed that La doping could increase the photocatalytic
2
, CO,
2
1
2,33
activity of NaTaO
3
.
Among them, the NTO:La(0.5), NTO:
La(1), and NTO:La(2) samples exhibited the highest
photocatalytic activity, e.g., NTO:La(2) exhibited production
−
1
−1
rates of 23 μmol h for H
2
, 1 μmol h for CO, and 4 μmol
−
1
Fig. 2 Time course of the production rate of H
(
2
(diamonds), CO
squares), and CO (black circles) and R(e /h ) (white circles) that is the
2
h
for CO (Fig. 3Ad). The NTO:La(2)SS sample also showed
2
−
+
high production rates (Fig. 3Af) but they were lower
compared to the NTO:La(2) sample (Fig. 3Ad). This result
reveals that the sample prepared by the flux method has
better photocatalytic activity in the PSRM, which is quite
consumed electron/hole ratio calculated from the production rates.
The NTO:La(2) photocatalyst (1.2 g) was used in a flow of the gas
mixture (25% CH
4
, 72.6% Ar, and 2.4% steam) at the flow rate of 15 ml
−
1
min under photoirradiation. The light intensity used was 165 mW
−
2
13
cm when measured at 254 ± 10 nm.
similar to the previous report. The ratios of the electrons
Catal. Sci. Technol.
This journal is © The Royal Society of Chemistry 2021

View Article Online
Catalysis Science & Technology
Paper
electron and hole consumption in the NTO:La(5) sample
might be caused by the production of some undetected
oxidative products, such as coke.
To compare the CO production rate and selectivity, the
enlarged graph is shown in Fig. 3B. The NTO:La(x) and NTO:
La(2)SS samples produce CO as a minor product. The La-
doped samples showed a higher CO production rate than the
non-doped NTO sample although the NTO:La(5) sample
exhibited a low CO production rate among them. However,
the CO selectivity with each photocatalyst was in the range of
15% to 24%, meaning that the CO selectivity did not vary so
much with the amount of La dopant. Meanwhile, the non-
doped NTO sample could not produce CO.
In order to investigate the role of La doping in crystals,
another series of samples were prepared, i.e., a La oxide species
was loaded on the surface of the non-doped NTO and NTO:
La(1) samples, and they were used for the photocatalytic
reaction test (Fig. 3C). Although the photocatalytic
production rate over the NTO sample was 1.26 μmol h
H
2
−
1
−1
(Fig. 3Ca), that over the La(0.3)/NTO sample was 0.53 μmol h
(Fig. 3Cb), meaning that the La oxide species on the NTO
surface decreased the photocatalytic activity to less than half its
original value. This negative effect was also observed over the
NTO:La(1) sample (Fig. 3C, c and d). These results suggested
that the surface La oxide species would change the properties of
the photocatalytic active sites, decreasing the photocatalytic
activity. Thus, the La doping effect on the photocatalytic activity
3+
of the NTO:La photocatalysts should be generated by the La
cations incorporated in the crystal structure and not by the La
oxide species located on the surface. Onishi et al. reported that
the electron–hole recombination process was suppressed by La
38,39
doping in the NaTaO
3
host.
The gradient of electrostatic
potential due to La doping promotes more efficient charge
38
separation. However, the photocatalytic activity of the NTO:
La(5) sample was almost the same as that of the bare NTO
sample (Fig. 3A, a and e), which is also consistent with the
38
literature. The excess La dopant in NaTaO might produce the
3
La oxide species on the surface, decreasing the photocatalytic
activity.
To know the effect of metal cocatalysts, various metal
loaded NTO:La(1) samples were examined as listed in Table
S1.† These metal loaded samples exhibited lower
photocatalytic activity for CO production than the bare NTO:
La(1) sample. The addition of a Pt cocatalyst on the NTO:
La(1) sample increased the hydrogen production rate but
suppressed CO formation, which is consistent with the
previous study. Loading a Pd cocatalyst also could change the
photocatalyst selectivity. Since the NTO:La(1) sample without
any loaded cocatalysts could produce H , CO, and CO , it is
Fig. 3 (A) Production rates of H
2 2
(gray bar), CO (black bar), and CO
(white bar) and the ratio of the consumed electrons and holes (white
circle) in the various samples: (a) NTO, (b) NTO:La(0.5), (c) NTO:La(1), (d)
NTO:La(2), (e) NTO:La(5), and (f) NTO:La(2) SS. (B) Enlarged graph showing
2
the CO and CO production rates and the CO selectivity, SCO (white
triangle). (C) Photocatalytic production rates and CO selectivity with
another series of samples on which a La oxide species was loaded by the
impregnation method: (a) NTO, (b) La(0.3)/NTO, (c) NTO:La(1), (d) La(0.3)/
2
2
NTO:La(1), and (e) Pt(0.1)/NTO:La(1). Photocatalyst: 1.2 g. Photoirradiation
demonstrated that the surface sites of the bare NTO:La
photocatalyst could originally produce H , CO, and CO and
Pt and Pd cocatalysts only promote the formation of H and
2
area: 6 cm . The feed gas consists of 25% CH
4
, 72.6% Ar, and 2.4% steam
−
1
−2
2
2
(
total flow rate: 15 ml min ). Light intensity: 165 mW cm for (A) and (B)
−2
2
and 27 mW cm for (C). Sampling was carried out after 4 hours irradiation.
1
0–14,16,37
CO
2
.
It is usually considered that Pt can function as
a good cocatalyst to enhance the electron and hole separation
1
3–16,40,41
and holes consumed for all the samples were almost unity
except for the NTO:La(5) sample (Fig. 3Ae). The unbalance of
and thus enhance the photocatalytic activity.
Thus,
even if the photocatalyst produces CO, the Pt metal cocatalyst
This journal is © The Royal Society of Chemistry 2021
Catal. Sci. Technol.

View Article Online
Paper
Catalysis Science & Technology
might promote successive conversion of CO with H O to form
2
CO . On the other hand, the other metal species such as Ag,
2
Ni, Cu, and Zn show negative effects on both the
photocatalytic activity and CO selectivity (Table S1†). These
cocatalysts would interact with the surface active sites of the
bare photocatalyst and make them less reactive.
As a comparison, we also checked the photocatalytic
activity and CO selectivity of some other photocatalyst
2 3 2
samples such as Ga O , TiO , and ZnO as listed in Table S1†
entries 10–12. These samples were used directly without any
additional pretreatment. As a result, NTO:La(1) exhibited the
highest photocatalytic activity and CO selectivity among
them. The Ga O sample also showed photocatalytic activity
2
3
to produce H , CO, and CO with a high CO selectivity, 19%.
2
2
In contrast, non-doped NTO, ZnO and TiO
2
showed low
−
Fig. 4 Production rate of H
2
(gray bar) and CO
2
(white) and RWGS(e /
+
photocatalytic activities for H and no CO or CO formation.
h ) (open circle) in the photocatalytic water gas shift (PWGS) reaction
2
2
test over the samples: (a) bare NTO, (b) NTO:La(1), (c) NTO:La(2), and
2
(d) NTO:La(5). Photocatalyst: 1.2 g, photoirradiation area:
6
cm ,
Reaction scheme for CO production
reactant: 15% CO, 2.7% H
2
O and Ar as the balance, flow rate: 15 ml
min . Sampling was carried out after 2 hours irradiation.
−1
To elucidate the possibilities of the successive reactions of
produced CO to give CO, such as photocatalytic dry
2
reforming of methane (PDRM, eqn (6)) and photocatalytic
reverse water gas shift (PRWGS, eqn (7)), we performed
In the PWGS experiment, the ratio of the consumed
electrons and holes was unity with these samples, which is
consistent with eqn (2). These photocatalysts, especially the
NTO:La photocatalysts, were highly active for the
photocatalytic WGS (PWGS). These high production rates
would be related to the lower reaction Gibbs energy for the
WGS. It is known that the steam reforming of methane
photocatalytic
photocatalyst.
reaction
tests
with
the
NTO:La(1)
CO2 þ CH4 → 2CO þ 2H2 ΔG°298K ¼ 170:7 kJ mol−1
CO2 þ H2 → CO þ H2O ΔG°298K ¼ 28:6 kJ mol−
(6)
(7)
1
(
SRM) is a highly endergonic reaction while the water gas
In the PDRM experiment, the concentration of CO in the
feed gas mixture was 20%, that of CH was 20% and the rest
4
shift (WGS) is an exergonic reaction as mentioned above (eqn
(2) and (3)). This result revealed that the photocatalytic water
gas shift can occur in the photocatalytic steam reforming of
methane (PSRM) with the NTO and NTO:La photocatalysts
and the latter is more active. This is the first report that a
2
was Ar gas, while in the PRWGS experiment the composition
of the feed gas was 20% CO , 20% H , and 60% Ar gas.
2
2
2
Although this CO concentration was much higher than that
achieved in the PSRM reaction experiments in this study,
there were no gaseous products in both reactions, meaning
that these reactions did not take place successively. Although
bare semiconductor photocatalyst without
promotes water gas shift (WGS) with high selectivity.
a cocatalyst
These results suggest that CO was produced in the
photocatalytic steam reforming (PSRM, eqn (1)) and
converted to CO2 in the photocatalytic water gas shift
reaction (PWGS, eqn (2)). This provides a steady state for the
CO production, which is the reason why the CO selectivity
was low being 15–24% in the present system. Further, the
successive reaction of CO with CH (PDRM, eqn (6)) and H
2
CO reduction using water as an electron source is another
2
probable reaction, Nakanishi et al. already reported that La
doped NaTaO
producing CO through CO reduction with water.
3
without a cocatalyst was not active for
4
2
2
On the other hand, to elucidate the successive reaction of
produced CO with water, the tests for photocatalytic water-
gas shift (PWGS, eqn (2)) were carried out with various NTO
and NTO:La photocatalysts (Fig. 4), where the feed gas
2
4
(PRWGS, eqn (7)) to produce CO could scarcely take place
under the present conditions as mentioned. Thus, the
proposed scheme for CO production is summarized in
Scheme 1. If the successive PWGS can be controlled by the
development of photocatalysts or reaction conditions, higher
CO selectivity will be obtained.
mixture consisted of 15% CO, 2.7% H
2
O, and 82.3% Ar.
Under dark conditions, there were no H and CO observed
2
2
(not shown). However, upon light irradiation, the gaseous
products were clearly detected. It was found that the WGS
can occur photocatalytically under these conditions with
−
+
these photocatalysts. The ratio of the products, RWGS(e /h ),
was estimated by using eqn (8) where RH2 and RCO2 are the
production rates for H and CO , respectively.
Reaction conditions
2
2
The influence of the reaction conditions, such as the light
intensity, the flow rate of reactants, and the CH4
concentration, on the photocatalytic activity and the CO
−
+
R
WGS(e /h ) = RH2/RCO2
(8)
Catal. Sci. Technol.
This journal is © The Royal Society of Chemistry 2021

View Article Online
Catalysis Science & Technology
Paper
Scheme
1 Proposed reaction scheme for CO production in the
photocatalytic steam reforming of methane over
photocatalyst based on the experimental results.
3
a NaTaO :La
selectivity was also investigated. Fig. 5A shows the effect of
the irradiation light intensity on the photocatalytic
performance of the NTO:La(2) photocatalyst, where the light
−
2
intensities utilized were 27, 35, and 165 mW cm without
using an optical filter, which represent incident photon
1
8
18
18 −1
numbers of 1.1 × 10 , 1.4 × 10 , and 6.6 × 10 s ,
respectively. As mentioned, no reaction occurred in the dark.
Higher light intensity provides higher production rates,
which is quite reasonable for photocatalysis. Interestingly, it
was found that the CO selectivity also slightly increased with
increasing light intensity from 18% to 24%. It is suggested
that the CO selectivity should be related to the difference of
the reaction rates of the first PSRM (eqn (1)) and the
successive PWGS (eqn (2)). As shown in eqn (1), it required 6
pairs of electrons and holes to promote CO generation in the
PSRM, while the PWGS only needs 2 pairs of charge carriers.
By increasing the light intensity, more electrons and holes
will be available for these photocatalytic reactions, which
might be more critically beneficial to the first PSRM reaction
rather than the successive PWGS reaction, resulting in the
higher CO formation rate. Higher temperature originating
from the high light intensity might also be another
possibility for increasing CO selectivity since the PSRM is an
endothermic reaction while PWGS is exothermic.
Fig. 5B shows the effect of the increasing total flow rate of
−
1
the reaction mixture (15, 30, and 50 ml min ), which
decreases the contact time (2.4, 1.2, and 0.7 s, respectively),
on the methane conversion and the CO selectivity over the
NTO:La(2) photocatalyst. As expected, the methane
conversion decreased with the increasing flow rate because of
less contact time. The CO2 production rate decreased with
the increasing flow rate (Fig. S5A†), while the CO production
rate was observed to be almost the same for various flow
rates, resulting in a slightly higher CO selectivity at the
highest flow rate. In principle, a high flow rate with a short
contact time should lower both the reaction rates, the CH
conversion to CO (eqn (1)) and the successive reaction of CO
to CO (eqn (2)). In the present case, it was revealed that a
4
Fig. 5 Formation rate of CO (circle), CO
together with CO selectivity (triangle) and CH
under different reaction conditions: (A) various incident photon
2
(square), and H
2
(diamond)
4
conversion (asterisk)
2
short contact time can further limit the successive reaction of
−
1
4
numbers; flow rate, 30 ml min ; feed gas composition: CH (35%),
CO to CO because practically there is less contact time for
2
steam (2%), and Ar (balance); photocatalyst, the NTO:La(2) sample, (B)
the successive reaction. It is concluded here that contact time
limitation can slightly increase the CO selectivity.
Fig. 5C and S5B† show the effect of the composition of
the feed gas mixture on the photocatalysis with the NTO:
−1
various flow rates of the feed gas mixture: 15, 30, and 50 ml min
;
−
2
light intensity, 35 mW cm ; feed gas composition: CH
4
(35%), steam
(
2%), and Ar (balance); photocatalyst, the NTO:La(2) sample, and (C)
various CH /H O ratios in the feed gas: CH (10–40%), steam (1.9–
.8%), and Ar (balance); light intensity, 165 mW cm ; flow rate, 15 ml
4
2
4
−2
2
La(1) photocatalyst, where the CH
4
concentration was varied
−1
min ; photocatalyst, the NTO:La(1) sample. The mass of photocatalyst
used was 1.2 g and the irradiation area was 6 cm for each experiment.
with a constant H O concentration to provide various ratios
2
2
of CH
4
to H
2
O. The ratios of CH
4
/H
2
O examined were from
See also Fig. S5 in the ESI.†
This journal is © The Royal Society of Chemistry 2021
Catal. Sci. Technol.

View Article Online
Paper
Catalysis Science & Technology
3
.5 to 21, much higher than the CH /H O stoichiometric Acknowledgements
4
2
ratio, 0.5. The graph shows that by increasing the CH /H O
4
2
This work was financially supported by ISHIZUE 2020 of
Kyoto University Research Development Program and the
Program for Elements Strategy Initiative for Catalysts and
Batteries (ESICB, JPMXP0112101003), commissioned by the
MEXT of Japan. W. Sarwana is grateful to the Indonesia
Endowment Fund for Education (LPDP), Ministry of Finance
Indonesia Republic for financial support during the study.
4
ratio, the CH conversion decreased gradually and became
stable. The CO selectivity increased in the lower range of the
CH /H O ratio from 3.5 to 11, and became constant at the
4
2
higher ratio range. Water would be strongly adsorbed on the
photocatalyst surface like liquid film at this low
a
temperature, so that an excess amount of CH is required to
4
perform PSRM (eqn (1)), but a further increase does not
influence the CO selectivity as much. Moreover, one
additional experiment was conducted in the flow of a gas
Notes and references
mixture with a very high CH
4 2 4
/H O ratio (90% CH and 0.3%
−
1
H O in Ar) at the flow rate of 15 ml min (Fig. S6†). The ratio
1 J. R. Rostrup-Nielsen, Catal. Today, 1993, 18, 305–324.
2 J. M. Cormier and I. Rusu, J. Phys. D: Appl. Phys., 2001, 34,
2798–2803.
2
−
+
of the consumed electrons and holes was not unity; R(e /h )
was around 2. Since the CH concentration was very high, it
4
was suggested that methane decomposition occurs to form
carbon although the reaction time was not long enough to
change the color of the photocatalyst. However, these
conditions gave a high CO selectivity of 39%, due to the
further acceleration of the first PSRM (eqn (1)) and the
limitation of the successive PWGS reaction (eqn (2)).
3 T. V. Choudhary and V. R. Choudhary, Angew. Chem., Int. Ed.,
2008, 47, 1828–1847.
4 X. Zhu, H. Wang, Y. Wei, K. Li and X. Cheng, J. Rare Earths,
2010, 28, 907–913.
5 N. Salhi, A. Boulahouache, C. Petit, A. Kiennemann and C.
Rabia, Int. J. Hydrogen Energy, 2011, 36, 11433–11439.
6 K. Y. Koo, S. Lee, U. H. Jung, H.-S. Roh and W. L. Yoon, Fuel
Process. Technol., 2014, 119, 151–157.
These results under various reaction conditions support
the proposed scheme mentioned above (Scheme 1).
7
M. A. Nieva, M. M. Villaverde, A. Monzón, T. F. Garetto and
A. J. Marchi, Chem. Eng. J., 2014, 235, 158–166.
Conclusions
8 L. Yuliati and H. Yoshida, Chem. Soc. Rev., 2008, 37, 1592.
K. Shimura and H. Yoshida, Energy Environ. Sci., 2011, 4,
467.
9
Photocatalytic steam reforming of methane (PSRM) to
produce CO was successfully uncovered by employing a La-
doped NaTaO3 photocatalyst without
2
1
1
1
1
1
0 H. Yoshida, S. Kato, K. Hirao, J. Nishimoto and T. Hattori,
a cocatalyst. This
Chem. Lett., 2007, 36, 430–431.
reaction produces CO and CO simultaneously. The presence
2
1 H. Yoshida, K. Hirao, J. Nishimoto, K. Shimura, S. Kato, H.
Itoh and T. Hattori, J. Phys. Chem. C, 2008, 112, 5542–5551.
2 K. Shimura, S. Kato, T. Yoshida, H. Itoh, T. Hattori and H.
Yoshida, J. Phys. Chem. C, 2010, 114, 3493–3503.
3 A. Yamamoto, S. Mizuba, Y. Saeki and H. Yoshida, Appl.
Catal., A, 2016, 521, 125–132.
4 K. Shimura, H. Miyanaga and H. Yoshida, in Studies in
Surface Science and Catalysis, ed. E. M. Gaigneaux, M.
Devillers, S. Hermans, P. A. Jacobs, J. A. Martens and P.
Ruiz, Elsevier, 2010, vol. 175, pp. 85–92.
of water and CO also initiates the photocatalytic water gas
shift reaction (PWGS), which decreases the CO production
rate. The selectivity to CO in the PSRM was controlled by the
photocatalyst properties and the reaction conditions such as
the light intensity, the flow rate of the reactant, and the ratio
4 2
of CH to H O in the feed gas mixture to some extent.
However, an excellent achievement is very difficult to be
obtained by changing these parameters in the present ranges
with the current photocatalyst and the reaction system. Thus,
the development of the photocatalyst and the reactor would
be desirable for further improvement.
1
1
1
1
1
2
2
2
5 K. Shimura and H. Yoshida, Energy Environ. Sci., 2010, 3,
6
15–617.
6 A. Anzai, K. Fujiwara, A. Yamamoto and H. Yoshida, Catal.
Today, 2020, 352, 1–9.
7 K. Shimura, H. Kawai, T. Yoshida and H. Yoshida, Chem.
Commun., 2011, 47, 8958–8960.
8 K. Shimura, H. Kawai, T. Yoshida and H. Yoshida, ACS
Catal., 2012, 2, 2126–2134.
9 K. Shimura, T. Yoshida and H. Yoshida, J. Phys. Chem. C,
Author contributions
Wirya Sarwana: conceptualization, investigation, and writing
–
original draft. Akihiko Anzai and Daichi Takami:
investigation. Akira Yamamoto: methodology and funding
acquisition. Hisao Yoshida: conceptualization, funding
acquisition, project administration, supervision, and writing
2
010, 114, 11466–11474.
0 K. Shimura, K. Maeda and H. Yoshida, J. Phys. Chem. C,
011, 115, 9041–9047.
1 K. Shimura and H. Yoshida, Phys. Chem. Chem. Phys.,
012, 14, 2678–2684.
–
review & editing.
2
2
Conflicts of interest
2 C. E. Taylor and R. P. Noceti, Catal. Today, 2000, 55,
259–267.
There are no conflicts to declare.
Catal. Sci. Technol.
This journal is © The Royal Society of Chemistry 2021

View Article Online
Catalysis Science & Technology
Paper
2
2
2
2
3 M. A. Gondal, A. Hameed, Z. H. Yamani and A. Arfaj, Chem.
33 H. Kato, K. Asakura and A. Kudo, J. Am. Chem. Soc.,
2003, 125, 3082–3089.
34 X. Li and J. Zang, Catal. Commun., 2011, 12, 1380–1383.
35 H. Sudrajat, S. Babel, I. Thushari and K. Laohhasurayotin,
J. Alloys Compd., 2019, 775, 1277–1285.
36 J. Sun, G. Chen, J. Pei, R. Jin, Q. Wang and X. Guang,
J. Mater. Chem., 2012, 22, 5609–5614.
37 S. P. Singh, A. Anzai, S. Kawaharasaki, A. Yamamoto and H.
Yoshida, Catal. Today, 2021, 375, 264–272.
Phys. Lett., 2004, 392, 372–377.
4 S. Murcia-López, K. Villa, T. Andreu and J. R. Morante, ACS
Catal., 2014, 4, 3013–3019.
5 K. Villa, S. Murcia-López, J. R. Morante and T. Andreu, Appl.
Catal., B, 2016, 187, 30–36.
6 S. Murcia-López, M. C. Bacariza, K. Villa, J. M. Lopes, C.
Henriques, J. R. Morante and T. Andreu, ACS Catal., 2017, 7,
2
878–2885.
2
7 J. Xie, R. Jin, A. Li, Y. Bi, Q. Ruan, Y. Deng, Y. Zhang, S. Yao,
G. Sankar, D. Ma and J. Tang, Nat. Catal., 2018, 1, 889–896.
8 Y. Zhou, L. Zhang and W. Wang, Nat. Commun., 2019, 10, 506.
9 J. Du, W. Chen, G. Wu, Y. Song, X. Dong, G. Li, J. Fang, W.
Wei and Y. Sun, Catalysts, 2020, 10, 196.
38 A. Yamakata, T. Ishibashi, H. Kato, A. Kudo and H. Onishi,
J. Phys. Chem. B, 2003, 107, 14383–14387.
39 M. Maruyama, A. Iwase, H. Kato, A. Kudo and H. Onishi,
J. Phys. Chem. C, 2009, 113, 13918–13923.
40 M. Yoshida, A. Yamakata, K. Takanabe, J. Kubota, M.
Osawa and K. Domen, J. Am. Chem. Soc., 2009, 131,
13218–13219.
41 J. Yu, L. Qi and M. Jaroniec, J. Phys. Chem. C, 2010, 114,
13118–13125.
2
2
3
3
3
0 H. Yoshida, S. Mizuba and A. Yamamoto, Catal. Today,
2
019, 334, 30–36.
1 F. Amano, C. Akamoto, M. Ishimaru, S. Inagaki and H.
Yoshida, Chem. Commun., 2020, 56, 6348–6351.
2 J. Tauc, R. Grigorovici and A. Vancu, Phys. Status Solidi B,
42 H. Nakanishi, K. Iizuka, T. Takayama, A. Iwase and A. Kudo,
ChemSusChem, 2017, 10, 112–118.
1
966, 15, 627–637.
This journal is © The Royal Society of Chemistry 2021
Catal. Sci. Technol.