Photocatalytic homolysis of methyl formate to dry formaldehyde on PdO/TiO2: Photocatalytic reverse Tishchenko reaction of methyl formate

Article abstract of DOI:10.1039/c4ra05744a

Photocatalytic homolysis of dry methyl formate (MF) to dry formaldehyde readily takes place in high selectivity (≥80%) upon irradiation of MF vapour on PdO/TiO₂ in the 385-1050 nm region. The Pd(i)Pd(0) nanoparticles supported on Ti(iii)-bearin

Full text of DOI:10.1039/c4ra05744a

RSC Advances
View Article Online
View Journal | View Issue
COMMUNICATION
Photocatalytic homolysis of methyl formate to dry
formaldehyde on PdO/TiO : photocatalytic reverse
2
Cite this: RSC Adv., 2014, 4, 33144
Tishchenko reaction of methyl formate†
Received 1st April 2014
Accepted 14th July 2014
ab
b
ab
Agni Raj Koirala, Son Docao and Kyung Byung Yoon*
DOI: 10.1039/c4ra05744a
www.rsc.org/advances
Photocatalytic homolysis of dry methyl formate (MF) to dry formal-
In the case of unimolecular photolysis of MF with UV lights
dehyde readily takes place in high selectivity ($80%) upon irradiation (193.3 and 234 nm), the CO elimination has been shown to be
1,12
of MF vapour on PdO/TiO
nanoparticles supported on Ti(III)-bearing TiO
reaction are proposed to be the active form of the catalyst for this products were CO, CO
2
in the 385–1050 nm region. The Pd(I)Pd(0) the major reaction.
In the case of photocatalytic decompo-
2
produced during the sition of MF on M/TiO
2
(M ¼ Pt, Pd, Rh, Ir, and Ru) with UV, the
2
2
, and H , indicating that the formation of
novel reaction.
CO and CH
CH OH to HCOOH and the decomposition of HCOOH to H
CO
3
OH takes place rst and the dehydrogenation of
and
as the catalyst
3
2
15,16
2
take place subsequently.
With bare TiO
2
Methyl formate (MF) has been used as a reagent for the
production of perfumes and food avouring agents and as a
solvent. During these applications and through vegetation it is
oen released to the atmosphere. Dimethyl ether and dime-
thoxy methane have been used as diesel substituents or as
diesel additives as a means to reduce the emission of carbon
particulates from automobiles. MF is also a key intermediate
during their combustion. In this respect, thermolysis, base-
catalyzed thermolysis,
14
the photolysis of MF does not take place, unless O is present.
2
Thus, the heterolysis of MF to CO and CH OH (CO elimi-
3
1,5–13,15,16
nation) has been the exclusive or primary reaction.
In
1
,2
this respect, nding a novel reaction that leads to homolysis of
MF to two CH O molecules or the unprecedented reverse
2
Tishchenko reaction of MF would be a highly valuable addition
to the chemistry of MF.
3–5
5–9
Formaldehyde should be highly pure and anhydrous to be
used for the preparation of commercially valuable high poly-
10,11
1,12,13
photolysis with ultraviolet light,
and catalytic photolysis of MF on various metal supported
17
18
meric polyoxymethylene and various aminoplast adhesives.
Aqueous CH O solution can be dehydrated to produce anhy-
14–16
TiO
2
have received great attention.
2
The theoretical studies on the unimolecular decomposition
of MF have shown that there are four sets of decomposition
products, (1) CH OH + CO, (2) two CH O, (3) CH + CO , and (4)
2
drous CH O. However, the drying and concentrating processes
are not only highly cumbersome but during these processes
3
2
4
2
HCOH + CH O, with the activation energy increasing in the
2
9
order of (1) < (2) < (3) < (4) (Fig. 1). Consistent with this, various
experiments have demonstrated that the formation of CH
3
OH
and CO (CO elimination from MF) is the major reaction during
5
–9
MF thermolysis.
Even during the related base-catalyzed
10,11
thermolysis, the CO elimination occurs exclusively.
In this
respect, MF has also been regarded as a CO carrier or a CO
8
storage compound.
a
Program of Integrated Biotechnology, Sogang University, Seoul 121-742, Korea
b
Korea Center for Articial Photosynthesis, Department of Chemistry, Sogang
University, Seoul 121-742, Korea. E-mail: yoonkb@sogang.ac.kr; Fax: +82-2-706-
4
269; Tel: +82-2-715-2569
†
Electronic supplementary information (ESI) available: Fig. SI-3. The UV parts of
the solar simulated light and the natural solar light both under the AM 1.5
condition. See DOI: 10.1039/c4ra05744a
Fig. 1 The potential surfaces for four different types of unimolecular
decomposition of MF proposed by Francisco and the coworkers.
9
33144 | RSC Adv., 2014, 4, 33144–33148
This journal is © The Royal Society of Chemistry 2014

View Article Online
Communication
RSC Advances
formic acid, low molecular weight polymers, and other by-
19
products are also formed, which must be removed. Accord-
ingly, the highly pure anhydrous CH O obtained by this method
2
becomes very expensive. Instead of this, anhydrous CH O has
2
20
been prepared by dehydrogenation of methanol on Ag, Ag
17
21,22
supported on SiO
2
/MgO/Al
2
O
3
,
and Fe
2
O
3
/MoO
3
.
However,
the catalysts are not stable and the reactions are carried out at
ꢀ
23
high temperatures (350–700 C). In this respect, the novel
catalysts that are highly stable and can be used unlimitedly
under mild conditions should be developed.
We now report that MF vapour undergoes homolysis to two
anhydrous CH O in high selectivity (>80%) upon passing it over
2
PdO-supported TiO
and the catalyst is stable for an unlimited period of time.
PdO/TiO was prepared by impregnation of Pd(NH
onto TiO powder (Degussa P25), followed by calcination at
00 C for 4 h. The amount of Pd loading was 1%. Similarly, Pt-
loaded TiO2 (PtO/TiO ), Au O -loaded TiO , (AuO/TiO ) and
2 2
(PdO/TiO ) under a solar simulated light
2
3 4 2
) Cl
2
ꢀ
4
2
2
3
2
2
CuO-loaded TiO2 (CuO/TiO ) were prepared. A metal oxide
2
Fig. 2 Plots of power density of the solar simulated light and UV light
(
MO)-loaded TiO (MO/TiO ) powder (300 mg) was placed on a
2 2
used in this study and the diffuse reflectance spectra of PdO/TiO
blue) and PdO/TiO irradiated with the solar simulated light in the
2
2
Petri dish and the MO/TiO -containing Petri dish was placed at
(
2
the bottom of a stainless steel reactor having a quartz window at presence of dry MF vapor (pink).
ꢀ
the top. The temperature of the reactor was maintained at 30 C.
The MF vapor was introduced into the reactor in the vapor
phase by passing a carrier gas (Ar, H , or O ) through a glass
2 2
of PdO/TiO and the PdO/TiO irradiated with light A in the
2
2
presence of the MF vapour are also shown.
bubbler lled with MF. The MF used in this study was carefully
dried in a glove box charged with dry Ar by storing it in a bottle
charged with dried Na-zeolite A pellets for 1 day. The small
amount of methanol that exists in MF as the stabilizer was not
completely removed. So the MF used in this work contained
2
The irradiated area was 6 cm . For dark reactions the quartz
window of the reactor was covered with an opaque lid. The
stream of vapours from the outlet of the reactor was introduced
on-line into one or two gas chromatographs (GCs) equipped
with a ame ionization detector (FID) and a pulsed discharge
detector (PDD). Several different types of GC columns were used
to analyze the products (ESI†). The amounts of reactants and
0
.4% of methanol (ESI†).
The ow rate of the carrier gas was 6 mL min . The
ꢁ1
temperature of the MF-containing bubbler was maintained
2
ꢀ
products are reported in terms of mmol per a unit area (1 cm )
between ꢁ40 and ꢁ1 C. The rate of MF input into the reactor
ꢁ
2
ꢁ1
ꢁ1
ꢁ2
and per h (mmol cm
h ).
was between 5.5 and 680 mmol h cm . The MF conversions
varied between 33 and 0.1% depending on the rate of MF input,
the nature of catalyst, and the reaction condition. In many
cases, the rate of MF input was maintained to be very high (>450
With the dry MF vapour in high input rates (>450
ꢁ2
ꢁ1
mmol cm
2
h ), PdO/TiO as the photocatalyst, light A as the
light source, and Ar as the carrier gas, the reacted amount of MF
ꢁ
2
ꢁ1
ꢁ1
ꢁ2
was 16.0 mmol cm
h
2
(Fig. 3A). With O as the carrier gas the
mmol h cm ) to minimize the secondary reaction of the
primary product by rapidly sweeping the primary products away
from the catalyst. To see the effect of the moisture on the
reaction rate and the product selectivity, the gas stream of MF
and a carrier gas was passed through a water bubbler to
simultaneously introduce moisture and MF into the reactor.
ꢁ2
reacted amount increased substantially (to 20.0 mmol cm
ꢁ1
h
), indicating that the cheaper anhydrous air can be used as a
better carrier gas. However, with H2 as the carrier gas, the
ꢁ
2
ꢁ1
reacted amount decreased sharply (6.0 mmol cm
h ).
The nature of MO also sensitively affected the reacted
amount. Thus, with PdO, PtO, AuO, and CuO as MO, respec-
tively, the reacted amounts were 16.0 to 6.2, 2.2, and to 2.0 mmol
The MF-to-moisture (MF/H
types of lights were used for photocatalytic reactions. The AM-
2
O) ratio was 17 ꢂ 3. Four different
ꢁ
2
ꢁ1
cm
2
h , respectively (Fig. 3B). Thus, PdO/TiO is most active.
1
3
.5 solar simulated light but with the spectral width between
20 and 1050 nm (light A) was used as the primary irradiation
The nature of light also sensitively affected the reacted amount.
ꢁ
2
ꢁ1
Thus, while light A gave 16.0 mmol cm
and light D gave 8.0, 2.6, 31.0 mmol cm
Fig. 3C).
Interestingly, when moist MF (MF/H
into the reactor under the condition of light A, PdO/TiO
h
, light B, light C,
source (Fig. 2). The power of the solar simulated light used in
ꢁ
2
ꢁ1
ꢁ
2
h
, respectively
our work was 72 mW cm due to the absence of the infrared
(
lights in the 1150–3200 nm region. By using long pass cut off
2
O ratio ¼ 17) was fed
, Ar, the

lters the solar simulated lights with the spectral widths
between 385 and 1050 nm (light B) and between 515-1050 nm
light C) were also produced and used. We also used a UV light
2
ꢁ2
ꢁ1
reacted amount increased by 2.5 times (to 41.0 mmol cm
with respect to that produced under the dry condition (Fig. 3D).
Under this moist condition, the use of O or H as the carrier gas
h )
(
with the spectral width between 235 and 400 nm (light D). The
plots of the power density of the lights with respect to the
wavelength are shown in Fig. 2. The diffuse-reectance spectra
2
2
gave rise to a decrease in the reacted amount by ꢃ50%.
This journal is © The Royal Society of Chemistry 2014
RSC Adv., 2014, 4, 33144–33148 | 33145

View Article Online
RSC Advances
Communication
Other products and the overall product selectivities are
shown in Fig. 4. The minor products are CH OH, CO , CO, and
3
2
H . The CH O selectivity remained at ꢃ80% regardless of the
2
2
carrier gas under the condition of light A, PdO/TiO , and dry MF
2
(Fig. 4A). The high selectivity maintained until the conversion
reached to 33% by decreasing the input rate of dry MF to 5.5
ꢁ
2
ꢁ1
mmol cm
affected the CH
Thus, with PdO, PtO, AuO, and CuO, respectively, the CH
h
. In contrast, the nature of MO sensitively
2
O selectivity and the product distribution.
2
O
selectivities were 84, 89, 59, and 34%, respectively (Fig. 4B). The
nature of the light also affected the product distribution
(
Fig. 4C). Thus, the CH O selectivities with lights A, B, C, and D
2
were 84, 97, 86, and 65%, respectively. In the presence of
moisture (MF/H O selectivity deceased sharply
O ¼ 17), the CH
to below 55% with Ar as the carrier gas and even to below 30%
with O and H as the carrier gases. Thus, the elimination of
moisture is crucial for higher CH O selectivities.
2
2
2
2
2
The PdO/TiO2 powder is nearly white because the loaded
amount of PdO (black) is only 1%. Accordingly, the diffuse-
reectance spectrum of (blue spectrum in Fig. 2) has only a
weak absorption tail between 400 and 700 nm. However, the
Fig. 3 (A) Effect of the carrier gas on the reacted amount of dry MF
during the photocatalytic reaction on PdO/TiO with light A. (B) Effect
2
of the nature of MO (PdO, PtO, AuO, and CuO) on the reacted amount
of dry MF during the photocatalytic reaction with the solar simulated
light and Ar as the carrier gas. (C) Effect of the irradiation light (A–D) on color of PdO/TiO
2
immediately turned grey and a very broad
the reacted amount of dry MF during the photocatalytic reaction on absorption band appeared over the entire visible region (pink
PdO/TiO
reacted amount of moist MF (MF/H
reaction on PdO/TiO with light A.
2
Ar as the carrier gas. (D) Effect of the carrier gas on the
spectrum in Fig. 2) upon adsorption of MF either in the dark or
under the light. We attribute the broad new absorption band to
the absorption by the reduced forms of PdO, such as Pd(I) and
Pd(0) species.
2
O ¼ 17) during the photocatalytic
2
The EPR spectra of the related various samples taken at 25 K
in the dark and under light B are shown in Fig. 5. TiO
the dark gave a very weak signal due to Ti(III) at g ¼ 1.975
Fig. 5A). Upon irradiation with light B, two strong signals due
2
alone in
(
+
to the hole (h , g ¼ 2.01) and Ti(III) (g ¼ 1.99, 1.97, and 1.95)
appeared (Fig. 5B). This phenomenon happens because TiO
2
absorbs light in the 385–400 nm region. Indeed, the two signals
Fig. 4 (A) Effect of the carrier gas on the product distribution during
the photocatalytic reaction of dry MF on PdO/TiO with light A. (B)
2
Effect of the nature of MO (PdO, PtO, AuO, and CuO) on the product
distribution during the photocatalytic reaction of dry MF with the solar
simulated light and Ar as the carrier gas. (C) Effect of the irradiation
light (A–D) on the product distribution during the photocatalytic
reaction of dry MF on PdO/TiO
carrier gas on the product distribution during the photocatalytic
reaction of moist MF (MF/H with light A.
O ¼ 17) on PdO/TiO
2
Ar as the carrier gas. (D) Effect of the
Fig. 5 EPR spectra of various samples (as indicated) taken at 25 K in
the dark and under the light (light B) in the presence and absence of
MF.
2
2
33146 | RSC Adv., 2014, 4, 33144–33148
This journal is © The Royal Society of Chemistry 2014

View Article Online
Communication
RSC Advances
became more intense when TiO
2
was irradiated with UV (light secondary reactions while we used on-line reaction system with
D). The MF-adsorbed TiO (TiO -MF) does not show any EPR high input rates of reactants to minimize the secondary reac-
2
2
signal in the dark except the original weak Ti(III) signal (Fig. 5C). tions of the primary products by rapidly sweeping the primary
+
However, a weak h signal and a strong Ti(III) signal appeared products away from the catalyst. The production of H and CO2
2
upon irradiation of TiO
that the electron transfer from MF to h took place. PdO/TiO
2
-MF with light B (Fig. 5D), indicating without CH
4
indicates that the reaction system of Solymosi's
group contains a signicant amount of moisture. In such a case
did not give additional ESR signals due to h and Ti(III) in the MF will primarily undergo hydrolysis to CH OH and HCO H,
dark (Fig. 5E). Upon irradiation with light B, it gave a weak and which subsequently undergo photocatalytic decomposition into
+
2
+
3
2
24
broad signal at g ¼ 2.1 due to Pd(I), together with the signals CO
2
2
and H . The presence of CO also indicates that CO elimi-
+
due to h and Ti(III) (Fig. 5F). In the presence of MF, either in the nation to CH OH and CO also took place.
3
dark (Fig. 5G) or under light B (Fig. 5H), PdO/TiO2 gave a
In summary, we have demonstrated the photocatalyic
homolysis of dry MF vapour into dry CH O vapour in high
+
+
stronger Pd(I) signal and no h signal, indicating that the h
scavenging by MF took place, giving rise to the formation of selectivity (>80%). PdO/TiO
Pd(I) species and Ti(III). Coupled with the broad UV-vis spectrum CH O in high selectivity with an inert gas or more preferentially
due to Pd nanoparticles, we conclude that MF converts PdO into with dry air as the carrier gas. The UV-free solar simulated light
Pd(I)/Pd(0) nanoparticles and TiO into the electron rich Ti(III)- is best to produce dry CH O in high selectivity and in high yield.
trapping TiO . We conclude that this is the active catalyst which
2
2
is the best catalyst to produce dry
2
2
2
2
carries out the photocatalytic conversion of MF to two CH O.
2
Acknowledgements
Although a detailed study is necessary to elucidate the
precise mechanism, we tentatively propose the reaction mech-
anism as follows. Thus, MF undergoes an oxidative addition by
cleaving the C–O bond to the electron rich Pd(I)Pd(0) nano-
This work was supported by the Korea Center for Articial
Photosynthesis, funded by the Ministry of Science, ICT and
Future Planning through the National Research Foundation of
Korea, no. 2009-0093886 and no. 2012R1A2A3A01009806. We
thank J. Y. Lee for the help in preparing the manuscript and
drawing gures. We also thank KBSI (Korea Basic Science
Institute) at Ewha Womans University for ESR measurements.
2
particles which are also supported on electron rich TiO (Fig. 6,
step A). The hydrogen atom transfer from the Pd-bound
methoxy group to the Pd-bound formyl group and the reduc-
tive elimination of two CH O takes place to give two CH O
2
2
molecules (Fig. 6, step B). For this reaction we propose that the
formation of electron rich Pd nanoparticles and TiO is the key
2
to this photocatalytic reaction. The activation of the surface
plasmon of the Pd nanoparticles by photons as well as the
Notes and references
maintenance of the electron rich Pd and TiO
essential for the oxidative addition and reductive elimination
reactions.
2
seems to be
1 S. H. Lee, J. Chem. Phys., 2008, 129, 194304.
2 T. J. Wallington, M. D. Hurley, T. Maurer, I. Barnes,
K. H. Becker, G. S. Tyndall, J. J. Orlando, A. S. Pimentel
and M. Bilde, J. Phys. Chem. A, 2001, 105, 5146–5154.
3 S. Dooley, F. L. Dryer, B. Yang, J. Wang, T. A. Cool, T. Kasper
and N. Hansen, Combust. Flame, 2011, 158, 732–741.
4 E. M. Fisher, W. J. Pitz, H. J. Currane and C. K. Westbrook,
Proc. Combust. Inst., 2000, 28, 1579–1586.
In the presence of H O, signicant amounts of CH OH, CO ,
2
3
2
and H were formed while the produced amount of CO is very
2
small. In this case, we conclude that CH OH was not produced
3
from the heterolysis of MF (CO elimination process) but from
the photocatalytic hydrolysis of MF yielding CH
HCO H, which decomposes into CO and H according to the
recent report.
Solymosi and the coworkers recently reported that CO
3
OH and
2
2
2
5 S. Dooley, M. P. Burke, M. Chaos, Y. Stein, F. L. Dryer,
V. P. Zhukov, O. Finch, J. M. Simmie and H. J. Currane,
Int. J. Chem. Kinet., 2010, 42, 527–549.
25
2
, CO
and H2 are the major products during the photocatalytic
6 W. K. Metcalfe, J. M. Simmie and H. J. Curran, J. Phys. Chem.
A, 2010, 114, 5478–5484.
7 B. W. Higdon Gray, C. C. Hobbs and M. J. Onore, US Pat.,
3,812,210, 1971.
decomposition of MF on M/TiO (M ¼ Pt, Pd, Rh, Ir, and Ru) by
2
irradiating the catalyst with a UV light with the spectral width
15
between 250–440 nm with the maximum intensity at 254 nm.
As noted, this result is quite different from ours. We think the
difference arises from the fact that Solymosi's group adopted
the circulation system in which the primary products undergo
8 F. R. Vega, J. C. Cl ´e ment and H. D. Abbayes, Tetrahedron Lett.,
1993, 34, 8117–8118.
9 J. S. Francisco, J. Am. Chem. Soc., 2003, 125, 10475–10480.
1
1
0 F. Q. Ma, D. S. Lu and Z. Y. Guo, J. Catal., 1992, 134, 644–653.
1 F. Q. Ma, D. S. Lu and Z. Y. Guo, J. Mol. Catal., 1993, 78, 309–
3
25.
2 M. H. Chao, P. Y. Tsai and K. C. Lin, Phys. Chem. Chem. Phys.,
011, 13, 7154–7161.
3 G. Cui, F. Zhang and W. Fang, J. Chem. Phys., 2010, 132,
34306.
1
1
1
2
0
4 C. C. Chuang, W. C. Wu, M. C. Huang, I. C. Huang and
Fig. 6 Schematic illustration of the proposed mechanism.
J. L. Lin, J. Catal., 1999, 185, 423–434.
This journal is © The Royal Society of Chemistry 2014
RSC Adv., 2014, 4, 33144–33148 | 33147

View Article Online
RSC Advances
Communication
1
5 G. Schubert, T. B ´a ns ´a gi and F. Solymosi, J. Phys. Chem. C,
013, 117, 22797–22804.
6 A. C. Lukaski and D. S. Muggli, J. Catal., 2004, 223, 250–261.
Ullmann, W. Gerhartz, Y. S. Yamamoto, F. T. Campbell, R.
Pfefferkorn and J. F. Rounsaville, VCH, Deereld Beach,
FL, USA, 5th edn, 1985, pp. 619–651.
2
1
1
7 L. P. Ren, W. L. Dai, Y. Cao, H. Li and K. Fan, Chem. 22 H. Gerberich and G. C. Seaman, Kirk-Otmer Encyclopedia of
Commun., 2003, 3030–3031.
8 T. Salthammer, S. Mentese and R. Marutzky, Chem. Rev.,
Chemical Technology, John Wiley & Sons, New York, 5th
edn, 2004.
1
2
010, 110, 2536–2572.
23 A. C. Lausche, J. S. Hummelshøj, F. Abild-Pedersen, F. Studt
and J. K. Nørskov, J. Catal., 2012, 291, 133–137.
1
2
9 R. S. Aries, US Pat., 2,953,602, 1960.
0 C. A. Bazilio, W. J. Thomas, U. Ullah and K. E. Hayes, Proc. R. 24 M. Narayana, J. Michalik, S. Contarini and L. Kevan, J. Phys.
Soc. London, Ser. A, 1989, 399(1816), 181–194. Chem., 1985, 89, 3895–3899.
1 G. Reuss, W. Disteldorf, O. Grundler and A. Hilt, in 25 A. Gazsi, G. Schubert, P. Pusztai and F. Solymosi, Int. J.
Ullmann's Encyclopedia of Industrial Chemistry, ed. F. Hydrogen Energy, 2013, 38, 7756–7766.
2
33148 | RSC Adv., 2014, 4, 33144–33148
This journal is © The Royal Society of Chemistry 2014