Decatungstate as photoredox catalyst: Benzylation of electron-poor olefins

Article abstract of DOI:10.1021/ol301900p

Excited tetrabutylammonium decatungstate (TBADT), known to activate a variety of compounds via hydrogen atom transfer (HAT), has now been applied as a photoredox catalyst for the effective oxidative cleavage of benzyl silanes and radical benzylation of reducible olefins occurring in isolated yields from poor to excellent.

Full text of DOI:10.1021/ol301900p

ORGANIC
LETTERS
2012
Vol. 14, No. 16
4218–4221
Decatungstate As Photoredox Catalyst:
Benzylation of Electron-Poor Olefins
Sara Montanaro, Davide Ravelli, Daniele Merli, Maurizio Fagnoni,* and Angelo Albini
PhotoGreen Lab, Department of Chemistry, University of Pavia, Viale Taramelli 12,
27100 Pavia, Italy
fagnoni@unipv.it
Received July 10, 2012
ABSTRACT
Excited tetrabutylammonium decatungstate (TBADT), known to activate a variety of compounds via hydrogen atom transfer (HAT), has now been
applied as a photoredox catalyst for the effective oxidative cleavage of benzyl silanes and radical benzylation of reducible olefins occurring in
isolated yields from poor to excellent.
Polyoxometalates (POMs) are metalꢀoxygen clusters
showing anamazingvarietyofproperties.1aꢀe SomePOMs
show interesting photochemical activity, as demonstrated
by the pioneering work by Hill.1f The most widely studied
among them is the decatungstate anion ([W10O32]4ꢀ),2a
usually employed as a tetrabutylammonium salt (TBADT).
This salt found some applications as a photocatalyst
both under aerated conditions, as in the photooxidation
of alkanes, alkenes, and alcohols,2 and under anaerobic
conditions for the photoactivation of CꢀH bonds in
alkanes,3a,b amides,3c aldehydes,3d ethers,3e and aromatic
alcohols,3f and as an innovative photocatalyst for the
degradation of organic pollutants in water.2a Two photo-
chemical processes compete upon TBADT excitation, viz.
hydrogen atom transfer (HAT)4 and electron transfer (ET).5
The former mainly operates in the presence of substrates
containing weak CꢀH bonds (as in the case of alcohols),3f,6
while the latter takes place when the reaction involves easily
oxidizable substrates, such as aromatic amines or aromatic
hydrocarbons.7a,b Of the two paths, only HAT has been
exploited for synthetic applications.2a Thus, to make the ET
path viable, the redox potential of reagent R (E(R•þ/R))
must be lower than the redox potential of the decatungstate
excited state (E([W10O32]4ꢀ*/[W10O32]5ꢀ)). Assessing this
point is complicated by the fact that the reactive state is a
relaxed excited state (the wO species), whose identity is still
(1) (a) Long, D.-L.; Tsunashima, R.; Cronin, L. Angew. Chem., Int.
Ed. 2010, 49, 1736. (b) Yu, R.; Kuang, X.-F.; Wu, X.-Y.; Lu, C.-Z.;
Donahue, J. P. Coord. Chem. Rev. 2009, 253, 2872. (c) Nomiya, K.;
ꢀ
Sakai, Y.; Matsunaga, S. Eur. J. Inorg. Chem. 2011, 179. (d) Dupre, N.;
ꢀ
^
Remy, P.; Micoine, K.; Boglio, C.; Thorimbert, S.; Lacote, E.;
Hasenknopf, B.; Malacria, M. Chem.;Eur. J. 2010, 16, 7256. (e)
Yunxiang, Q.; Zhenshan, H. Curr. Org. Chem. 2009, 13, 1347. (f) Hill,
C. L. Synlett 1995, 127.
(4) (a) Tanielian, C.; Schweitzer, C.; Seghrouchni, R.; Esch, M.;
Mechin, R. Photochem. Photobiol. Sci. 2003, 2, 297. (b) Kothe, T.;
Martschke, R.; Fischer, H. J. Chem. Soc., Perkin Trans. 2 1998, 503.
(5) (a) Tanielian, C.; Seghrouchni, R.; Schweitzer, C. J. Phys. Chem.
A 2003, 107, 1102. (b) Nosaka, Y.; Takei, T.; Fujii, N. J. Photochem.
Photobiol. A: Chem. 1995, 92, 173.
(6) Lykakis, I. N.; Vougioukalakis, G. C.; Orfanopoulos, M. Org.
Lett. 2003, 5, 2875.
(7) (a) Texier, I.; Delaire, J. A.; Giannotti, C. Phys. Chem. Chem.
Phys. 2000, 2, 1205. (b) Duncan, D. C.; Fox, M. A. J. Phys. Chem. A
1998, 102, 4559. (c) Duncan, D. C.; Netzel, T. L.; Hill, C. L. Inorg. Chem.
1995, 34, 4640.
(2) (a) Tzirakis, M. D.; Lykakis, I. N.; Orfanopoulos, M. Chem. Soc.
Rev. 2009, 38, 2609. (b) Molinari, A.; Bratovcic, A.; Magnacca, G.;
Maldotti, A. Dalton Trans. 2010, 39, 7826.
(3) (a) Dondi, D.; Cardarelli, A. M.; Fagnoni, M; Albini, A. Tetra-
hedron 2006, 62, 5527. (b) Ryu, I.; Tani, A.; Fukuyama, T.; Ravelli, D.;
Fagnoni, M.; Albini, A. Angew. Chem., Int. Ed. 2011, 50, 1869.
(c) Angioni, S.; Ravelli, D.; Emma, D.; Dondi, D.; Fagnoni, M.; Albini,
A. Adv. Synth. Catal. 2008, 350, 2209. (d) Esposti, S.; Dondi, D.;
Fagnoni, M.; Albini, A. Angew. Chem., Int. Ed. 2007, 46, 2531. (e)
Tzirakis, M. D.; Orfanopoulos, M. Org. Lett. 2008, 10, 873. (f) Lykakis,
I. N.; Tanielian, C.; Seghrouchni, R.; Orfanopoulos, M. J. Mol. Catal.
A: Chem. 2007, 262, 176.
r
10.1021/ol301900p
Published on Web 08/02/2012
2012 American Chemical Society

applications would be using benzyltrimethylsilanes as
radical precursors9,11 and maleic anhydride and related
alkenes as electrophilic traps/oxidizing agents.12 The results
are presented below (Tables 1, 2). Preparative experiments
involved the TBADT-photocatalyzed reactions of a number
of benzylsilanes (1ꢀ8) with electrophilic alkenes (9ꢀ13)
upon irradiation for 24 h by means of phosphor-coated
lamps (λirr centered at 310 nm). Blank experiments showed
that parent benzyltrimethylsilane (1) was not appreciably
consumed when irradiated in neat acetonitrile for 24 h under
these conditions, either alone or with maleic anhydride (9).
The results are gathered in Table 2, where both the silane
conversion (initial concentration 0.1 M) and the yield of the
benzylated alkene are reported. An excess of the electrophilic
olefin (0.2 M) was routinely used to ensure the full con-
sumption of the starting silane. The excess olefin could be
limited, however, at least in some cases. As an example, the
starting concentration of olefin 9 could be decreased to
0.15 M in the benzylation with 1 (0.1 M) that gave open
chain 2-benzylsuccinic acid (14) in ca. 90% yield. The
reactivity of benzyltrimethylsilane in the presence of a
catalytic amount (2 mol %) of TBADT was examined in
neat acetonitrile (condition A), in neat MeCN/H2O 5:1
(condition B) and in MeCN/H2O 5:1 solution containing
0.5 M LiClO4 (condition C). The irradiation of parent 1 in
the presence of conjugated alkenes 9to 13 under condition A
gave variable results, from 27% conversion and no benzy-
lated product with maleimide 11c (see Table S1; in this case a
precipitate was formed on the reaction vessel walls) to 100%
conversion and 90% benzylation with maleic anhydride 9.
Scheme 1. Possible Pathways in Tetrabutylammonium Deca-
tungstate (TBADT) Photocatalyzed Reactions
not fully known.7c Previous studies estimated that the wO
potential ranges from 2.26 to 2.61 V (vs SCE),7a thus taking
the lower limit E(R•þ/R) < 2.26. An approach for facilitat-
ing the ET path is to incorporate an electroauxiliary group
(EG) in the reagent. This would lower E(R•þ/R) and open a
reactive path for the radical cation.8 The trimethylsilyl group
(TMS) has been largely used in this role in electrochemical
or photoinduced electron transfer (PET) reactions8,9 and
is split off from the radical cation generating a radical.8
However, introducing a TMS group is no guarantee that
the reaction follows an ET path. Indeed, HAT can remain
the major path, as shown in the case of N-methyl-
N-(trimethylsilylmethyl)acetamide,3c where only silicon-
containing products were formed despite the fact that the
oxidation potential of this amide was estimated to be
E1/2 < 2.29 V vs SCE and thus suitable for undergoing
a monoelectronic oxidation by wO. In addition, if the
process has to be photocatalytic, TBADT (used in a few
percent proportion) must be regenerated; otherwise the
conversion will stop when this is consumed. In the HAT
mechanism, [HW10O32]4ꢀ typically undergoes a back hydro-
gen transfer reaction (Scheme 1, path a0) and this regenerates
the photocatalyst.
Table 1. Formal Redox Potential (E°0) of the Compounds Used
in This Work
Correspondingly, in the ET mechanism, efficient reox-
idation of the reduced cluster (or of the two-electron
reduced species formed by dismutation)10 is required,
also because back electron transfer to the donor radical
cation would otherwise compete and quench any chemical
reactions. Thus, a successful ET catalyzed alkylation
requires reoxidation of [W10O32]5ꢀ by a sacrificial acceptor
(A, path b0). Oxygen could be used in this role, but would
also trap alkyl radicals, thus restricting the palette of viable
processes to oxygenation reactions. To avoid this limita-
tion, a reducible trap can be used, such as an electrophilic
olefin in the twofold role of trap for the radicals and
electron acceptor. In view of the above, a favored case
for testing TBADT as a photoredox catalyst for synthetic
Under condition B both parameters increased, in some
cases remarkably, except in the reaction with 9, where the
(8) Yoshida, J.-i.; Kataoka, K.; Horcajada, R.; Nagaki, A. Chem.
Rev. 2008, 108, 2265.
(11) (a) Zhuikov, V. V. Russ. J. Gen. Chem. 1997, 67, 887. (b) Taylor,
K.; Miura, K.; Akinfaderin, F.; Fry, A. J. J. Electrochem. Soc. 2003, 150,
D85.
(12) Some of these systems have been previously investigated under
TiO2 photocatalysis. See: (a) Cermenati, L.; Richter, C.; Albini, A.
Chem. Commun. 1998, 805.
(9) Albini, A.; Fagnoni, M. In Molecular and Supramolecular Photo-
chemistry; Griesbeck, A. G., Mattay, J., Eds.; Synthetic Organic Photo-
chemistry; Dekker: New York, 2005; Vol. 12, p 453.
(10) Yamase, T.; Takabayashi, N.; Kaji, M. J. Chem. Soc., Dalton
Trans. 1984, 793.
Org. Lett., Vol. 14, No. 16, 2012
4219

(in the range 52 to 97%). Ring-substituted benzylsilanes 2 to
7 were examined in the presence of fumaronitrile 10 and
in representative cases of maleic anhydride 9. Again, the
conversion was variable under condition A (31 to 80%),
with a benzylation yield ranging from 27 (for the 4-cyano
derivative 7) to 84% (for the 4-methyl derivative 3). Like-
wise, improved results were obtained under condition B,
while the conversion reached 100% under condition C
for the 4- and 3-methoxy (2 and 6), the 4-t-butyl (4),
and 4-chloro (5) derivatives, with a benzylation yield of
80ꢀ93%. Less successful were the reactions of the 4-methyl
(3) (78% conversion, 54% benzylation) and 4-cyano (7)
derivatives (respectively 80 and 54%). In the case of 7,
however, increasing the irradiation time to 40 h allowed
improvement of both parameters. The effect of chain sub-
stitution was also tested by using the R-methylbenzylsilane 8
which gave satisfactory results already under condition A
and excellent results under B and C. On the other hand,
acrylonitrile, cyclohexenone and methyl acrylate were poorly
alkylated by 1 under the present conditions and the silane
remained essentially unaffected (not reported in Table 2).
To gain indications of the mechanism involved, flash
photolysis experiments were carried out. In accordance
with the literature, the reactive excited state of decatung-
state (wO) was revealed at λ = 780 nm, with a lifetime τ =
54 ns in neat acetonitrile.7a In the presence of p-methox-
ybenzylsilane (2), this transient was quenched with a linear
SternꢀVolmer plot (see Figure S3), from which a rate
constant of 7.1(5) ꢁ 108 Mꢀ1 sꢀ1 was determined. No
persistent absorption was observed, contrary to the case of
H-donating quenchers, for which a long-lived signal has
been observed and attributed to [HW10O32]4ꢀ that absorbs
in the same region.7a
Table 2. TBADT Photocatalyzed Benzylation of Electron-Poor
Olefins
The above results demonstrate that benzylation reac-
tions have been obtained under decatungstate photocata-
lysis. Such reactions play an important role in fine
chemistry and in the synthesis of pharmaceuticals and
agrochemicals.13a,b These involve cationic alkylations of
aromatics and heteroaromatics13c and the more recently
developed transition-metal catalyzed benzylations.13a,b,d,e
Benzyl radicals are usually generated upon single elec-
tron transfer oxidation of benzyl silanes or phenylacetates
(see below). However, these are rather persistent inter-
mediates and are inefficiently trapped, as witnessed by the
high proportion of competitive homocouplings to give
bibenzyls that are generally observed in such reactions.13f
This contrasts with the more favorable course of various
a Reaction mixture irradiated for 24 h with ten 15 W phosphor-
coated lamps (λirr = 310 nm); Condition A: MeCN as the solvent; B:
MeCN/H2O 5:1 as the reaction medium; C: 0.5 M LiClO4 in MeCN/
H2O 5:1 solution as the reaction medium. Conversions and yields are
based on silane consumption. b 0.15 M olefin. c 0.1 M olefin. d 40 h
irradiation time. e Equimolar mixture of diastereoisomers. f 2,3-Diphe-
nylbutane 260 (as a mixture of diastereoisomers) was obtained in 3%
yield (see Supporting Information for details).
(13) (a) Weaver, J. D.; Recio, A., III; Grenning, A. J.; Tunge, J. A.
ꢀ
Chem. Rev. 2011, 111, 1846. (b) Liegault, B.; Renaud, J.-L.; Bruneau, C.
Chem. Soc. Rev. 2008, 37, 290. (c) Noji, M.; Konno, Y.; Ishii, K. J. Org.
€
Chem. 2007, 72, 5161. (d) Ackermann, L.; Barfusser, S.; Kornhaass, C.;
Kapdi, A. R. Org. Lett. 2011, 13, 3082. (e) Mertins, K.; Iovel, I.; Kischel,
J.; Zapf, A.; Beller, M. Angew. Chem., Int. Ed. 2005, 44, 238.
(f) Baciocchi, E.; Rol, C.; Rosato, G. C.; Sebastiani, G. V. J. Chem.
Soc., Chem. Commun. 1992, 59.
(14) For example, see: Freccero, M.; Fasani, E.; Albini, A. J. Org.
Chem. 1998, 58, 1740. Hatoum, F.; Gallagher, S.; Baragwanath, L.; Lex,
J.; Oelgemoller, M. Tetrahedron Lett. 2009, 50, 6335. Cermenati, L.;
Freccero, M.; Venturello, P.; Albini, A. J. Am. Chem. Soc. 1995, 117,
7869. Brivio, L.; Caronna, T.; Fontana, F.; Gambarotti, C.; Mele, A.;
Natali Sora, I.; Punta, C.; Recupero, F. J. Photochem. Photobiol. A:
Chem. 2010, 214, 112.
change was limited. Under condition C the conversion was
further increased and was consistently in the range
85ꢀ100%, except for the reactions with dimetyl fumarate
(67%) and dimethyl maleate (48%), while the benzylation
yield generally did not change with respect to condition B
€
4220
Org. Lett., Vol. 14, No. 16, 2012

benzylations via photoinduced electron transfer. In the latter
case, however, benzyl radicals coupled with another odd-
electron species, viz. the radical anion of the acceptor,
likewise formed in the ET step.9,14 In one case the benzyl
radical was generated from a benzyl silane by a thermal
SET reaction and trapped by an N-acyliminium ion.15 The
benzylation of olefins considered here has only a few
precedents, including the radical addition onto an enamine
generated in situ by organocatalysis16a or onto the photo-
generated radical anion of highly reducible olefins (such as
1,1-dicyanoethenes).16b Very recently, Naito et al. reported
the functionalization of electron-poor olefins based on the
CꢀH activation in (substituted) toluenes by using the
Et3B/O2 system as a radical initiator, albeit the use of a
high temperature (180 °C) and of the H-donor as neat
reaction medium undermined the synthetic potential of the
process.16c A milder functionalization was again obtained
photochemically by using TiO2 as the photocatalyst.12a
The reactions carried out in the present work can be safely
recognized as involving ET from the benzylic derivatives to
excited decatungstate (see Scheme 2), as supported by
various evidence. Thus, quenching of the reactive decatug-
state wO does not form persistent [HW10O32]4ꢀ and occurs
Scheme 2. Proposed Reaction Mechanism
with that of the ground state of the photoactive species
E([W10O32]4ꢀ/[W10O32]5ꢀ) that was measured to be
around ꢀ0.9 V vs SCE.18 The comparison is only quali-
tative, since the actual value of this potential depends on
the water content of the medium10 and on the dispropor-
tionation of [W10O32]5ꢀ that generates the more reducing
anion [W10O32]6ꢀ (ca. ꢀ1.4 V vs SCE).10,18 ET to the
alkene at the second branch of the photocatalytic process
causes two positive effects on the efficiency of the reaction.
First, back electron transfer to R•þ is limited; second, the
radical anion of the alkene is formed. This is a much better
trap for benzyl radicals than the neutral form and explains
the success of the benzylation. Under these conditions the
olefin is used in (almost) an equimolar amount, making
the reaction convenient for application. A contrario evi-
dence includes attempts to extend the reaction to less
reducible olefins such as acrylonitrile (E1/2 = ꢀ2.17 V vs
SCE),19a cyclohexenone (E1/2 = ꢀ2.20 V vs SCE),19b and
methyl acrylate (E1/2 = ca. ꢀ2.90 V vs SCE)19c that led to
only small amounts of benzylated derivatives, while most
of the starting silane remained unreacted.
To our knowledge, the reactions above are the first
preparative application of a TBADT-photocatalyzed
synthesis involving the ET rather than the more common
HAT mechanism. Perhaps more importantly, this smooth
procedure, which is formally analogous to a reductive
Heck,20 adds to the small group of known benzylations
of alkenes via benzyl radicals. Prediction of the success of
the reaction can be reached by taking into account the
redox potential of both the benzylsilane and the unsatu-
rated trap, with reference to Scheme 2.
at a high rate (7 ꢁ 108 Mꢀ1
s
ꢀ1), quite similar to those
reported for bona fide single electron donors, such as
naphthalene or phenanthrene.7a The benzylic derivatives
considered here are oxidized at 1.28 to 1.81 V vs SCE
(Table 1), and thus ET is exothermic by 10ꢀ30 kcal molꢀ1
,
taking into account the uncertainty of the wO potential.
Fragmentation of the silane radical cation is then thermo-
dynamically allowed (log kcl ca. 6.36 for compound 2).17a
Indeed, only Si-free products were obtained; e.g., with
methyl-substituted silane 3 no competitive deprotonation
occurs as judged from the product structure, in accordance
with the indication that desilylation is ca. 100 times faster
with respect to deprotonation.17a The cleavage is known to be
nucleophile-assisted,17b and this is well in accordance with the
increased yields in water admixed solvent (condition B) and
when the ionic strength is increased (condition C).17c
Finally, the high yield of benzylation requires some
comment. Persistent species such as benzyl radicals react
sluggishly with modest electrophiles such as those tested
in this work. The typical result is a significant or predo-
minant proportion of bibenzyl derivatives. In the present
case, however, no such products were formed, except
for a small amount in the case of the hindered R-radical
from 8. A viable rationalization is that the olefin serves
the function of sacrificial acceptor. The redox potentials
of the alkenes used here (E(A/A•ꢀ)) are in the range
ꢀ0.88 to ꢀ1.65 V vs SCE and should be compared
1
Supporting Information Available. H NMR, 13C NMR
spectra for all new compounds. This material is available
free of charge via the Internet at http://pubs.acs.org.
(18) Renneke, R. F.; Pasquali, M.; Hill, C. L. J. Am. Chem. Soc. 1990,
112, 6585.
(15) Maruyama, T.; Mizuno, Y.; Shimizu, I.; Suga, S.; Yoshida, J.-i.
J. Am. Chem. Soc. 2007, 129, 1902.
(19) (a) Iwai, K.; Takemura, F.; Furue, M.; Nozakura, S.-i. Bull.
Chem. Soc. Jpn. 1984, 57, 763. (b) Harada, J.; Sakakibara, Y.; Kunai, A.;
Sasaki, K. Bull. Chem. Soc. Jpn. 1984, 57, 611. (c) Arnold, D. R.; Borg,
R. M.; Albini, A. J. Chem. Soc., Chem. Commun. 1981, 138.
(20) Minatti, A.; Zheng, X.; Buchwald, S. L. J. Org. Chem. 2007, 72,
9253. Goksu, G.; Ocal, N.; Kaufmann, D. E. Molecules 2010, 15, 1302.
(16) (a) Shih, H.-W.; Vander Wal, M. N.; Grange, R. L.; MacMillan,
D. W. C. J. Am. Chem. Soc. 2010, 132, 13600. (b) Hayamizu, T.; Maeda,
H.; Ikeda, M.; Mizuno, K. Tetrahedron Lett. 2001, 42, 2361. (c) Ueda,
M.; Kondoh, E.; Ito, Y.; Shono, H.; Kakiuchi, M.; Ichii, Y.; Kimura, T.;
Miyoshi, T.; Naito, T.; Miyata, O. Org. Biomol. Chem. 2011, 9, 2062.
(17) (a) Mella, M.; Fagnoni, M.; Freccero, M.; Fasani, E.; Albini., A.
Chem. Soc. Rev. 1998, 27, 81. (b) Dinnocenzo, J. P.; Farid, S.; Goodman,
J. L.; Gould, I. R.; Todd, W. P.; Mattes, S. L. J. Am. Chem. Soc. 1989, 111,
8973. (c) Mattay, J.; Vondenhof, M. Top. Curr. Chem. 1991, 159, 219.
The authors declare no competing financial interest.
Org. Lett., Vol. 14, No. 16, 2012
4221