Solar light induced carbon-carbon bond formation via TiO2 photocatalysis

Article abstract of DOI:10.1039/a709172i

Solar light irradiation of a TiO₂ suspension in MeCN containing maleic anhydride and 4-methoxybenzyl(trimethyl)silane gives benzylated succinic acid (or anhydride) on a gram scale.

Full text of DOI:10.1039/a709172i

View Article Online / Journal Homepage / Table of Contents for this issue
Solar light induced carbon–carbon bond formation via TiO2 photocatalysis
Laura Cermenati,a Christoph Richterb and Angelo Albinia†
a Department of Organic Chemistry, University of Pavia, via. Taramelli 10, 27100 Pavia, Italy
b
Plataforma Solar de Almeria, PO Box, 22, 04200 Tabernas, Almeria, Spain
Solar light irradiation of a TiO2 suspension in MeCN
containing maleic anhydride and 4-methoxybenzyl(trime-
thyl)silane gives benzylated succinic acid (or anhydride) on
a gram scale.
R1X+· ? R· ? Products
(4)
must contain both a donor and an acceptor with the suitable
redox potential [see eqn. (1)] since return electron transfer to the
surface [e.g. eqn. (2)] would otherwise immediately quench the
key intermediate. One of the radical ions should react rapidly in
order to prevent back electron transfer after diffusion [eqn. (3)]
from quenching the reaction. In our case, the reaction is
cleavage of the radical cation leading to a radical, which is then
trapped in the desired chemical reaction [eqn. (4)]. This implies
that the radical anion should not fragment, because otherwise
statistical radical coupling (R1· + R2·) would result, and that
efficient trapping prevents dimerisation of R1·.
In the present case, 4-methoxybenzyl(trimethyl)silane (see
Scheme 1, Eox 1.31 V vs. SCE in MeCN, cf. Evb 2.2 V for TiO2)
functions as the fragmentable donor and either maleic acid or
maleic anhydride (Ered 20.84 V, cf. Ecb 20.8 V) has the double
role of electron acceptor and radical trap.
Thus, an MeCN solution (1 l) of the silane (3.88 g, 0.02 m)
and maleic anhydride (2.16 g, 0.022 m) containing TiO2 (1.4 g)
(untreated Degussa P25 pigment) was pumped by means of a
peristaltic pump through a refrigerated tube sitting in the focus
of a parabolic mirror (surface exposed to the sun 0.2 m2), while
maintaining a slow flux of nitrogen and keeping the temperature
at 15 °C. This arrangement was sufficient for maintaining a
uniform suspension of TiO2 in the tube. After 10 h (July,
Almeria, sunny day) of exposure to solar light, the suspension
was filtered, the solvent evaporated and the residue recrystal-
lized to give 2.86 g (65%) of 2-(4-methoxybenzyl)succinic
anhydride. Minor products were 4,4A-bis(4-methoxy)bibenzyl
and 2,3-bis(4-methoxybenzyl)succinic anhydride.
Irradiation of a semiconductor (l ! DEbg = Ecb 2 Evb, see
Scheme 1) causes charge separation. If a significant part of the
solar emission is absorbed, methods for solar energy conversion
based on this principle can be devised. Most of the research in
this field makes use of titanium dioxide, an inexpensive,
chemically stable and atoxic semiconductor which absorbs all
of the UV component of the solar spectrum.1–4 Photo-
electrochemical solar cells have been devised and their
efficiencies have been considerably improved over the years.
However, solar light is diffuse, and electrical power produced in
this way remains expensive.5 The same holds for the production
of fuels, e.g. hydrogen, which is also feasible using this
principle.5 On the other hand, the production of fine chemicals
may be rewarding. We report here an example of the use of solar
light for the most typical reaction of organic synthesis, carbon–
carbon bond formation, via TiO2 photocatalysis. The reaction
considered is the radical alkylation of electron-withdrawing
substituted olefins.
When planning a synthesis based on TiO2 photocatalysis, the
problem is to translate a transient charge separation (electron-
hole recombination on the semiconductor surface takes place in
ca. 30 ps)6 into an irreversible and selective reaction. The
system we used is outlined in eqns. (1)–(4). The solution
TiO2(h+, e2) + R1X + R2Z ? TiO2 + R1X+· + R2Z2·
R1X+· + TiO2(e2) ? TiO2 + R1X
(1)
(2)
(3)
Monitoring the reaction by GC showed that it followed the
expected zero-order kinetics, and reagent consumption and
R1X+· + R2Z2· ? R1X + R2Z
E
E
CH2Ar
E
– •
E
E
E
e–
ECB
e –
H +
CH2Ar
E
•
∆Ebg
hν
E
krec
E
E
EVB
h+
+ •
•
ArCH2SiMe3
ArCH2
Ar = 4-MeOC6H4
E = CO2H or E—E = (CO)2O
ArCH2SiMe3
Scheme 1
Chem. Commun., 1998
805

View Article Online
a good yield of alkylated products is obtained on a reasonable
scale by this simple method, both solvent and semiconductor are
easily recovered and reused, and little excess reagents or by-
products remain. This may induce further exploration of
heterogeneous photocatalysis for organic synthesis beyond the
few cases that are known to date,8,10 and lead to its use, along
with homogeneous photoreactions,11 for the exploitation of
solar light.
0.020
0.016
0.012
0.008
0.004
0.000
Support of this work by the EC under the TMR program is
gratefully acknowledged.
0.0
0.4
0.8
1.2
1.6
2.0
2.4
Notes and References
Incident photons / mol
† E-mail: albini@chifis.unipv.it
Fig. 1 Formation of (-) 4-Methoxybenzylsuccinic anhydride or (<) acid vs.
‡ This was measured by means of a calibrated photometer (l < 400 nm).
§ A 30% yield of dihydrofuranyldiphenylhydrazine (also on a gram scale)
has been obtained by Kish via CdS photocatalysis from azobenzene and
dihydrofuran; in that case, however, the key step is radical coupling (see ref.
8).
¶ Based on the total flux incident on the reflecting mirror surface exposed
to the sun, 0.2 m2) and the fact that this is concentrated on a Pyrex tube (™
3 cm, length 1 m), coaxial with the mirror, in which the solution flows. No
account is taken of any loss by refraction or reflection.
incident UV photons
product formation were proportional to the integrated incident
flux‡ (Fig. 1). A similar reaction course and a similar yield of
the benzylated succinic acid was obtained when maleic acid was
used in the place of the anhydride, although this required 22 h
exposure overall. Furthermore, the alkylation could be effected
with other donors, e.g. 4-methoxyphenylacetic acid.
These reactions are initiated by hole/electron transfer to
produce a pair of radical ions (see Scheme 1). The alkene radical
anion is stable, while the silane radical cation fragments and
gives the 4-methoxybenzyl radical. Under the present condi-
tions this is mainly trapped by maleic anhydride (or acid), rather
than coupling. The adduct radical is reduced by the persistent
radical anion of the acceptor or by electron transfer at the
semiconductor surface, and protonated by water present in
MeCN to give the final products. Operation of a similar
mechanism for photoinduced electron transfer initiated alkyla-
tion of unsatured acid derivatives had been previously demon-
strated to occur when a soluble sensitizer was used (although in
that case no benzylation was observed),7 but has no precedent in
photocatalysis.§ The change to a heterogeneous sensitiser
simplifies the method, since introduction of other species into
the solution is avoided, work-up is simpler and the semi-
conductor can be recovered.
1 A. Fujishima and K. Honda, Nature, 1972, 238, 37.
2 K. C. Chang, A. Heller, B. Schwartz, S. Menezes and B. Miller, Science,
1977, 196, 1097.
3 A. Heller, Acc. Chem. Res., 1995, 28, 503.
4 M. Graetzel, J. Sol-Gel Sci. Technol., 1994, 2, 673.
5 L. Jacob, E. Oliveros, O. Legrini and A. M. Braun, Trace Met. Environ.,
1993, 3, 511.
6 D. E. Skinner, D. P. Colombo, J. A. Cavalieri and R. M. Bowman,
J. Phys. Chem., 1995, 99, 7853; N. Serpone, D. Lawless, R.
Khairutdinov and E. Pelizzetti, J. Phys. Chem., 1995, 99, 16 655.
7 M. Fagnoni, M. Mella and A. Albini, J. Am. Chem. Soc., 1995, 117,
7877; M. Mella, M. Fagnoni, M. Freccero, E. Fasani and A. Albini,
Chem. Soc. Rev., 1998, 81.
8 R. Kuenneth, C. Feldmer, F. Knoch and H. Kisch, Chem. Eur. J., 1995,
1, 441; W. Schindler, F. Knoch and H. Kisch, Chem. Ber., 1996, 129,
925.
9 N. Serpone, J. Photochem. Photobiol. A, 1997, 97, 1.
10 M. A. Fox, Top. Curr. Chem., 1987, 142, 71; M. A. Fox and M. T.
Dulay, Chem. Rev., 1993, 93, 417; H. Kish, J. Prakt. Chem., 1994, 336,
635.
Exploitation of solar light in this way is not efficient (at the
present stage ‘apparent’¶ quantum yields are ca. 3% for the
anhydride and 1% for the acid, see Fig. 1, and the literature
shows that in general this quantity does not exceed 10% with
TiO2 sensitised reactions,9 in part due to reflection). However,
11 P Esser, B. Pohlmann and H. D. Scharf, Angew. Chem., Int. Ed. Engl.,
1994, 33, 2009; Angew. Chem., 1994, 106, 2085.
Received in Cambridge, UK, 22nd December 1997; 7/09172I
806
Chem. Commun., 1998