Tetrabutylammonium decatungstate-photosensitized alkylation of electrophilic alkenes: Convenient functionalization of aliphatic C-H bonds

Article abstract of DOI:10.1002/chem.200501216

Tetrabutylammonium decatungstate (TBADT, 2×10^-3M) is an effective photocatalyst for the alkylation of electrophilic alkenes (0.1 M, α,β-unsaturated nitriles, esters, ketones) by alkanes, alcohols, and ethers. The products are in most cases obtained in >70% isolated yields, through an experimentally very simple procedure. The kinetics of the radical processes following initial hydrogen abstraction by excited TBADT in deoxygenated MeCN have been studied. In the absence of a trap, back hydrogen transfer from reduced tungstate is the main pathway for alkyl radicals, while α-hydroxyalkyl radicals are oxidized to ketones by ground-state TBADT. With both radical types the reaction ceases at a few percent conversion. However, trapping by electrophilic alkenes is followed by reduction of the radical adduct and regeneration of the catalyst, which allows the alkylation to proceed up to complete alkene conversion with the mentioned good yields of products. With a nucleophilic (α-hydroxyalkyl) radical, alkylation is efficient (Φ = 0.58) and can also be carried out when degassing is omitted, the only difference being a short induction period. With a less reactive (cyclohexyl) radical, the quantum yield is lower (Φ = 0.06) and the reaction is considerably slowed in aerated solutions, but the chemical yield remains good.

Full text of DOI:10.1002/chem.200501216

FULL PAPER
DOI: 10.1002/chem.200501216
Tetrabutylammonium Decatungstate-PhotosensitizedAlkylation of
ꢀ
Electrophilic Alkenes: Convenient Functionalization of Aliphatic C H
Bonds
[a]
Daniele Dondi,[a, b] Maurizio Fagnoni,[a] andAngelo Albini*
Abstract: Tetrabutylammonium deca-
tungstate (TBADT, 2”10ꢀ3 m) is an ef-
fective photocatalyst for the alkylation
of electrophilic alkenes (0.1m, a,b-un-
saturated nitriles, esters, ketones) by al-
kanes, alcohols, and ethers.The prod-
ucts are in most cases obtained in
>70% isolated yields, through an ex-
perimentally very simple procedure.
The kinetics of the radical processes
following initial hydrogen abstraction
by excited TBADT in deoxygenated
MeCN have been studied.In the ab-
sence of a trap, back hydrogen transfer
from reduced tungstate is the main
pathway for alkyl radicals, while a-hy-
droxyalkyl radicals are oxidized to ke-
tones by ground-state TBADT.With
both radical types the reaction ceases
at a few percent conversion.However,
trapping by electrophilic alkenes is fol-
lowed by reduction of the radical
adduct and regeneration of the catalyst,
which allows the alkylation to proceed
up to complete alkene conversion with
the mentioned good yields of products.
With a nucleophilic (a-hydroxyalkyl)
radical, alkylation is efficient (F
=
0.58) and can also be carried out when
degassing is omitted, the only differ-
ence being a short induction period.
With a less reactive (cyclohexyl) radi-
cal, the quantum yield is lower (F =
0.06) and the reaction is considerably
slowed in aerated solutions, but the
chemical yield remains good.
ꢀ
Keywords:
A
alkylation
·
C H
chemistry · photo-
catalysis · radicals
Introduction
versatile conditions.[4] The excited states of alkanes and
many simple aliphatic derivatives are high in energy and un-
dergo unselective fragmentation, so photochemistry has
little preparative value in this case.However, chemical acti-
vation through reaction with a photoexcited compound can
be achieved under mild conditions and probably deserves
more attention than it has received up to now.In this
method, an activated intermediate is generated; in many in-
stances this is a carbon-centered radical arising by hydrogen
ꢀ
The mild and selective functionalization of aliphatic C H
bonds remains a major goal in synthesis,[1] in particular in
the context of current interest in sustainable methods that
adhere to the principle of atom economy by minimizing the
use of reagents and auxiliary agents and operate under mild,
energy-saving conditions.[2] Catalysis by transition metal
complexes is a promising possibility,[3] although metal com-
plexation of compounds not containing p orbitals is less
common than that of alkenes or aromatics, thus limiting the
versatility of this approach.An appealing alternative would
be a photochemical reaction.The photon is the ideal
“green” reagent, since it leaves no residue and transfers
large amounts of energy, while operating under mild and
ꢀ
abstraction from a C H bond [Eq.(1)].
These reactions include photoinduced radical chain proc-
esses (quantum yield @1, Scheme 1a), such as halogenation
and related reactions.Other reactions, such as oxygenation,
take place thermally following a chain course at high tem-
peratures, but can be carried out photochemically at room
temperature by stoichiometric generation and trapping of
the radicals.A triplet ketone, a photogenerated acyl radical,
[a] Dr.D.Dondi, Dr.M.Fagnoni, Prof.A.Albini
Department of Organic Chemistry
University of Pavia (Italy)
v.Taramelli 10, 27100 Pavia (Italy)
Fax : (+39)0382-98323
E-mail: angelo.albini@unipv.it
ꢀ
or a photoexcited semiconductor can be used for the C H
bond activation.An excellent choice for a photoinitiator is a
[b] Dr.D.Dondi
INCA Research Unit, University of Pavia (Italy)
Chem. Eur. J. 2006, 12, 4153 – 4163
ꢀ 2006 Wiley-VCH Verlag GmbH & Co.KGaA, Weinheim
4153

ꢀ
Addition to a C C multiple bond, in contrast, is the most
useful radical process in synthesis and occurs as a chain re-
action (after initiation, in several cases photochemical), pro-
ꢀ
vided that an easily fragmentable precursor (R X, with
X = Br or, better, I) and a mediator, typically a tin derivative,
are used (Scheme 1c).[10] Extension of this principle to a
chain radical process based on the cleavage of a strong bond
ꢀ
such as a C H one is out of the question.However, an ap-
pealing possibility is a non-chain, photocatalyzed process as
shown in Scheme 1d, involving H abstraction by the excited
ꢀ
initiator, radical addition to a C C p bond, and back hydro-
gen transfer to the radical adduct, leading to the regenera-
tion of the photoinitiator.In comparison to Scheme 1c, this
path has the disadvantage of being a stoichiometric (in
terms of photons; the oxometalate is a regenerated catalyst)
rather than a chain process, but this shortcoming is counter-
balanced by the excellent atom economy, since the reagents
are 100% incorporated in the products (unlike in Scheme
1c, in which the atom or group X is lost).Furthermore, no
mediator such as a tin derivative is used, conditions are
mild, and toxic reagents are avoided.A first attempt to
apply this concept—in the photocatalyzed addition of cyclo-
alkanes to a,b-unsaturated nitriles—was successful[11] and
we now report a study on the scope and mechanism of CACHTREUNGH
ꢀ
activation and C C bond formation through radical alkyla-
tion of a,b-unsaturated compounds under photocatalytic
conditions.
Scheme 1.Some of the different mechanisms for photoinduced reactions
of aliphatic derivatives.a) (Photoinitiated) radical chain process.b) Non-
chain photoreaction catalyzed by a polyoxometalate (Mn+=O).c) (Photo-
initiated) radical chain alkyation reaction.d) Non-chain photocatalyzed
alkylation reaction.
Results
Products studies: In this work, the easily prepared
tetrabutylACHRTEaUNG mmonium decatungstate (TBADT) was used as
polyoxometalate (Mn+=O in Scheme 1b),[5] provided that it
is regenerated in the course of the reaction and functions as
a photochemically activated catalyst (i.e., it is not consumed
over several cycles), although the reaction of the organic
compound is stoichiometric in terms of adsorbed photons.
Photocatalyzed oxygenation of alkanes with use of a poly-
oxometalate has been particularly well studied (sodium, po-
tassium, or tetraalkylammonium decatungstate are most
often used).Hydroperoxides or, under different conditions,
ketones have been selectively obtained, sometimes with
high levels of alkane conversion,[6] in contrast to the mix-
tures obtained under conventional (peroxide-initiated) au-
tooxidation conditions.Under these conditions, the catalyst
is reoxidized back by oxygen.Oxygenation may be taken as
an example of radical addition to a multiple bond A=B (O=
O in this case), but the extension of this concept to other re-
actions, in particular to addition to carbon–carbon multiple
the photocatalyst in acetonitrile, with alkanes, alcohols,
ethers, and acetals serving as the radical precursors and elec-
trophilic alkenes as radical traps.The reaction kinetics of
radical generation and trapping were studied in detail in
representative cases.
The photocatalyzed reactions were studied both in prepa-
rative experiments (100 mL argon-flushed solutions) for
product isolation and characterization and in small-scale ex-
periments (2 mL samples, deoxygenated by several freeze-
degas-thaw cycles) for quantum yield measurements.
TBADT was quite photostable when irradiated in oxygen-
free MeCN, and under our conditions several hours were re-
quired for observable reduction to take place.Irradiation of
this salt (0.002m) in a nitrogen-flushed acetonitrile solution
containing propan-2-ol (0.5m), however, caused the develop-
ment after a few minutes of the intense blue color character-
istic of the reduced tungsten species.The only organic prod-
uct was acetone, quantified by HPLC as its 2,4-dinitrophe-
nylhydrazone.
ꢀ
bonds, with the goal of achieving synthetically desirable C
C bond formation, has so far met with limited success.Radi-
cal addition of alkanes to alkenes and alkynes,[7a] or to
CO,[7b] through polyoxometalate photocatalysis has been
demonstrated but occurs only up to small degrees of conver-
sion of the reagent (ca.5–8%), although good yields of ni-
triles and/or a-iminoesters have been obtained by addition
to cyanoformates.[8,9]
When a similar solution containing acrylonitrile (AN,
0.1m) was irradiated, the color was again apparent and per-
sisted, but a new product besides acetone was formed and
was identified after separation by distillation (isolated yield
72% on the basis of the starting amount of AN) as 4-hy-
4154
www.chemeurj.org
ꢀ 2006 Wiley-VCH Verlag GmbH & Co.KGaA, Weinheim
Chem. Eur. J. 2006, 12, 4153 – 4163

Photosensitized Alkylation of Electrophilic Alkenes
FULL PAPER
droxypentanenitrile (1a; see Scheme 2, Table 1).Similarly,
photocatalysis of propan-2-ol in the presence of isopropyl-
ACHTREUNGidenemalonitrile (IPMN, 0.1m) under the same conditions
gave the corresponding hydroxyalkylmalononitrile 1b in a
above alkenes (see Table 1), while small amounts of bicyclo-
hexyl and cyclohexyl methyl ketone (5% each) were formed
in the absence of such additives.[11] Alkylation with cyclo-
hexane was now extended to other electrophilic alkenes.In
particular, alkylation of both methyl acrylate and cyclohex-
en-2-one was satisfactorily achieved, giving products 5 and
6, respectively (Table 1).In contrast, attempted alkylation
with amines was unsuccessful, irradiation of TBADT in the
presence of Et3N resulting in a rapid darkening of the solu-
tion, from which no products could be isolated.
satisfactory yield (50%).
As indicated, the above irradiations had been carried out
after deoxygenation of the solutions.However, this step
could be omitted.Thus, when carried out in an open tube,
the irradiation of TBADT in the presence of propan-2-ol
and AN gave product 1a in a yield only slightly lower
(64%) than had been obtained under air exclusion after the
same irradiation time.The blue color was barely detectable
in this case.The alkylation in air-equilibrated solution could
also be carried out with cyclohexane, though in that case the
required irradiation time was four times longer than under
air exclusion.
The source of the b-hydrogen atom in the nitriles formed
was tested by repeating the photocatalyzed addition of tetra-
hydrofuran to AN first in CD3CN and then in CH3CN con-
taining 0.5% D2O.No measurable deuteration resulted in
the first case, while product 2a was quantitatively obtained
as the 2-[D1] derivative under the latter set of conditions.
Exactly the same result was obtained in the alkylation of
AN with cyclohexane, which quantitatively gave the 2-[D1]
derivative of 4a in the presence of 0.5% D2O.
Scheme 2.Photoproducts arising from the irradiation of TBADT in the
presence of various aliphatic derivatives and electrophilic alkenes.
Table 1.Isolated yields of TBADT-photocatalyzed (00.02 m) alkylation of
electrophilic alkenes in acetonitrile (lir 310 nm).
Donor
[0.5m]
Alkene
[0.1m]
Alkylated product
(yield [%])
G
G
ACHTREUNG
2-PrOH
THF
CH2=CHCN
CMe2=C(CN)2
CH2=CHCN
CMe2=C(CN)2
CH2=CHCN
CMe2=C(CN)2
CH2=CHCN
1a (72)
1b (50)
2a (78)[a]
2b (75)
3a (76)
3b (50)
4a (63)[b]
5 (65)
Monitoring the photoreaction: The above photocatalyzed
reactions were conveniently carried out by irradiation at
310 nm, since TBADT absorbs strongly at this wavelength,
where the blue reduced tungsten derivative that developed
during the reaction absorbed poorly.The latter species cor-
responded, as previously determined, to a mixture of the
one- and two-electron reduced decatungstate salts
(HW10O324ꢀ/H2W10O324ꢀ),[12–14] with the latter species pre-
dominating.Thanks to the 310 nm absorption of the residual
catalyst, alkylation could be pursued up to virtually com-
plete alkene conversion, though at a gradually decreasing
rate.The course of the reaction was monitored and the time
profiles for the formation of acetone from propan-2-ol and
of product 1a from the alcohol in the presence of AN are
shown in Figure 1.These data can be compared with the
moles of electrons accepted by the catalyst to form the tung-
state reduced species, which were calculated from the
known extinction coefficients of such compounds, as shown
in Figure 2.
DIOX
C6H12
CH2=CHCO2Me
cyclohexen-2-one
6 (43)
[a] Irradiation in MeCN containing 0.5% D2O gave 2-[D1]-2a (>90%
specific deuteration).[b] MeCN containing 05. % D 2O, 2-[D1]-4a (>90%
deuteration) was obtained.
Other alcohols, such as methanol, were oxidized under
these conditions and similarly caused a fast development of
the blue color.Tetrahydrofuran (THF) and 1,3-dioxolane
(DIOX) were then examined as a representative ether and
acetal, respectively.Irradiation of the tungstate in the pres-
ence of both substrates again gave the blue color, but no
GC-detectable products.When the reactions were carried
out in the presence of a,b-unsaturated nitriles, alkylation
took place satisfactorily.Tetrahydrofuran thus gave 3-(2-tet-
rahydrofuranyl)propanenitrile (2a) from AN in 78% yield
and dinitrile 2b from IPMN in 75% yield.Similarly, the di-
oxolane gave the dioxolanylpropanenitrile 3a in 76% yield
with AN and dinitrile 3b in 50% yield with IPMN.
The reduction of the catalyst proceeded up to over 90%
conversion in a few minutes.When it was the only additive,
2-PrOH was oxidized to acetone, reaching a 1”10ꢀ3 m con-
centration in the same time interval.Further irradiation pro-
duced only slow growth.When AN was also present, the
rate of reduction of TBADT decreased to about two thirds
of the previous value and that of formation of acetone to
We had previously reported that cyclohexane gave alky-
lated nitriles 4a (63%) and 4b (66%) in the presence of the
Chem. Eur. J. 2006, 12, 4153 – 4163
ꢀ 2006 Wiley-VCH Verlag GmbH & Co.KGaA, Weinheim
www.chemeurj.org
4155

A.Albini et al.
Figure 1.Products formed ( m) upon irradiation of a degassed (except
when noted) 2”10ꢀ3 m TBADT solution in MeCN in the presence of vari-
Figure 3.Turnover number of the photocatalyst for the formation of:
~
!
1) ( ) product 1a, and 2) ( ) acetone (multiplied ”10) from a deaerated
solution of 2-PrOH (0.5m) and AN (0.1m) in MeCN versus the AN con-
*
ous reagents: ( ) acetone (multiplied ”10 for better visibility) in the
&
^
*
presence of 2-PrOH (0.5m); ( ) acetone (”10) and ( ) product 1a in the
version.( ) The same for product 1a in air-equlibrated MeCN solution.
~
&
presence of 2-PrOH (0.5m) and AN (0.1m); ( ) product 4a in the pres-
( ) The same for product 4a in deaerated MeCN in the presence of 0.5m
!
ence of cyclohexane (0.5m) and AN (0.1m); ( ) product 1a in the pres-
cyclohexane and 0.1m AN.
ence of 2-PrOH (0.5m) and AN (0.1m) in an air-equilibrated solution.
The above reactions were carried out at a 0.5m starting
concentration of the hydrogen donors.The initial quantum
yields for the above reactions were estimated by extrapola-
tion of the curves in Figure 1 and Figure 2 and are reported
in Table 2.Further experiments were carried out with lower
Table 2.Initial quantum yields for the photochemical reduction of
TBADT (Fr) and for photocatalyzed product formation (Fp) in MeCN
and in the presence of AN (lir = 313 nm).
Neat MeCN
products (Fp)
0.1m AN in MeCN
products (Fp)
ꢀ
R H
Fr
Fr
none
0.004
2-PrOH 0.52
Me2CO (0.31)
0.23
Me2CO (0.09), 1a (0.58)
C6H12
0.013
A
0.037 4a (0.06)
Figure 2.Electrons ( m) accepted by the catalyst upon irradiation of a de-
gassed 2”10ꢀ3 m TBADT solution in MeCN in the presence of various
C6H12
MeOH 0.25
THF
DIOX
0.004 5 (0.05)[a]
0.04
&
*
additives: ( ) in the presence of 2-PrOH (0.5m); ( ) in the presence of
0.47
0.36
0.05
0.04
2a (0.20)
3a (.38)
!
2-PrOH (0.5m) and AN (0.1m); ( ) in the presence of cyclohexane
~
(0.5m); ( ) in the presence of cyclohexane (0.5m) and AN (0.1m).
[a] In the presence of 0.1m methyl acrylate.
one third, while formation of product 1a proceeded steadily,
though at a gradually decreasing rate, even when 90% of
the catalyst had been reduced.The turnover number for 1a
(the ratio between the moles of product formed and the
moles of electrons transferred to the catalyst) grew from ca.
12 in the initial phase to >35 at advanced conversion (see
Figure 3).The same experiment in aerated solution showed
an induction phase of a few minutes, but the amount of
product 1a formed later increased over that observed in de-
gassed solution (as mentioned above, very little reduced
TBADT was formed in this case).
With cyclohexane, reduction of the catalyst took place at
a much slower pace, although it continued, reaching a pla-
teau at which about 30% of the starting amount had been
reduced after 8 h.In contrast with the 2-PrOH case, the rate
of TBADT reduction increased in the presence of AN, while
alkylation proceeded effectively up to turnover numbers of
50 and above.
concentrations of propan-2-ol and cyclohexane and the ini-
tial quantum yields of the TBADT reduction were likewise
evaluated and are plotted in Figure 4a and Figure 5a.
The reaction course was also monitored by flash photoly-
[15,16]
sis (FP).Previous studies by Hill et al,.
Giannotti et
al.,[17] and Tanielian et al.[13] and others had found evidence
of complex behavior after excitation of TBADT in MeCN.
Picosecond transients preceded the reactive excited state (t
= 55 ns) and this gave the reduced polytungstate (first half-
life ca.700 ns), all of these species absorbing around
760 nm.In MeCN, the yield and the lifetime of the last spe-
cies decreased in the presence of oxygen, and its lifetime
was shortened, while addition of 2-PrOH (in air-saturated
solution) increased the yield and made this species persis-
tent over several hundred ns.
On repetition of the flash experiment in oxygen-free
MeCN the above transient was observed (t = 58ꢁ4 ns)
with a small persistent component (>5 ms).The long-lived
4156
www.chemeurj.org
ꢀ 2006 Wiley-VCH Verlag GmbH & Co.KGaA, Weinheim
Chem. Eur. J. 2006, 12, 4153 – 4163

Photosensitized Alkylation of Electrophilic Alkenes
FULL PAPER
Figure 4.Initial quantum yields for photoinduced electron transfer to
TBADT (2”10ꢀ3 m) by irradiation in MeCN in the presence of 2-PrOH
as determined: a) by steady-state irradiation, and b) by flash photolysis.
Figure 6.Time profile of the absorption at 760 nm after flashing of a 6”
10ꢀ5 m solution of TBADT in MeCN in the presence: a) of propan-2-ol
(0, 0.05, 0.1, 0.2 and 0.5m, bottom-to-top), and b) of cyclohexane (0, 0.1,
0.2, 0.3, and 0.5m).The curves were identical in the presence of 01.
m
AN.
can be seen in Figure 1, the amount of this product formed
reaches a plateau slightly above 1 mm.At this stage most of
the catalyst (starting concentration 2 mm) is in the reduced
form.The observed stoichiometry is fitted well by Equa-
tion (2).
4ꢀ
2 W10O32 þ Me2CHOH
Figure 5.Initial quantum yields for photoinduced electron transfer to
TBADT (2”10ꢀ3 m) by irradiation in MeCN in the presence of cyclohex-
ane as determined: a) by steady-state irradiation, and b) by flash photoly-
sis.
ð2Þ
! 2 HW10O32 ðor H2W10O324ꢀÞ þ Me2C¼O
4ꢀ
As for cyclohexane, this gives a little bicyclohexyl and cy-
clohexyl methyl ketone, but again a plateau is reached at a
few percent conversion (not shown in the Figure 1).
transient increased greatly and become dominant in the
presence of increasing amounts of both cyclohexane and 2-
PrOH (Figure 6).The addition of 01. m AN to these solu-
tions caused no appreciable change.
The quantum yields for TBADT reduction under FP con-
ditions were measured by use of benzophenone as actinome-
ter.The results are plotted in Figure 4b and Figure 5b as a
function of the additive concentration and the key parame-
ters are reported in Table 2.
ꢀ
The tungstate thus activates C H bonds, but it behaves as
a stoichiometric oxidant and accumulates in the inactive re-
duced state.Known reoxidation mechanisms are hydrogen
evolution—demonstrated to occur, although inefficiently, in
the presence of platinum[12]—or reaction with oxygen, which
also traps the radicals to form peroxy derivatives.[6a]
In this work the reaction was carried out in the presence
of electrophilic alkenes and alkylation was found to occur
efficiently, overcoming the other photoinduced processes
either for the major part (with propan-2-ol) or completely
(with cyclohexane, THF, or DIOX).Under these conditions,
TBADT is not a stoichiometric reagent, but rather a photo-
catalyst (TON>35 at the end of the reaction).Although the
rate of alkylation may vary considerably—the irradiation
time required for converting propan-2-ol, for example, is
much shorter than that for cyclohexane (by a factor of ca.
Discussion
Course of the reaction: The irradiation of TBADT in deoxy-
genated MeCN has been found to induce CACHTREHUGN activation of
various aliphatic derivatives and formation of dehydrogenat-
ed products.Thus, propan-2-ol is oxidized to acetone.As
Chem. Eur. J. 2006, 12, 4153 – 4163
ꢀ 2006 Wiley-VCH Verlag GmbH & Co.KGaA, Weinheim
www.chemeurj.org
4157

A.Albini et al.
10)—the alkylation proceeds steadily in all cases, as shown
by the regular increase in TON with conversion (see
Figure 3).
Figure 5b versus the hydrogen donor concentration.The
lines connecting the experimental points were drawn accord-
ing to Equation (4) and the values were obtained: K = 5.3
and 2.1mꢀ1 for propan-2-ol and cyclohexane, respectively.
These correspond to the rate constants kr = 10.1 and 4.0”
107 mꢀ1 sꢀ1 when account is taken of the measured TBADT
lifetime in MeCN (58 ns; see Figure 6 and Table 3), close to
Hydrogen abstraction: As mentioned in the Introduction,
the nature and reactivity of the decatungstateꢁs excited
states and the primary events in the presence of several ali-
phatic derivatives have been clarified previously.[13,15–17] The
the previously measured values (kr
=
8 and 3.7”
4ꢀ
reactive excited state of the W10O32 ion, indicated simply
107 mꢀ1 sꢀ1).[13] Some reduction occurs in neat MeCN (F8 =
0.09), indicating that the solvent and the tetrabutylammoni-
um counter cation contribute,
as W* in Scheme 3, is characterized by a hole in a O-based
although inefficiently, to the
photoreduction
(see
Scheme 3).If such contribu-
tion were due to MeCN only,
this would give k’r
=
9”
104 mꢀ1 sꢀ1.It may be noted
that the rate of hydrogen ab-
straction by W* is about two
orders of magnitude faster
than by excited states of or-
ganic molecules similarly char-
acterized by holes localized on
Scheme 3.Competing paths in the photocatalyzed activation of alkanes.
MO and accordingly efficiently abstracts hydrogen from ali-
4ꢀ
Table 3.Parameters for the TBADT-photoinduced activation of cyclo-
hexane and propan-2-ol in MeCN.
phatic derivatives to yield reduced HW10O32 (indicated as
HW in Scheme 3) and alkyl radicals.The discussion below
on the early steps is very limited, because the current data
essentially confirm previous work, while the subsequent rad-
ical reactions and their exploitation in synthesis are the
target of this study and are discussed in some detail.
Reagent
Flim
K [mꢀ1
]
kq [mꢀ1 sꢀ1
]
[a]
C6H12
0.52
0.52
0.52
0.55
2.1
0.024
5.3
4.0”107
[b]
C6H12
2-PrOH[a]
2-PrOH[b]
1.0”108
18.4
Reduced polytungstate is easily monitored by FP through
the strong absorption in the visible region and, according to
Scheme 3, the quantum yield associated with the reduction
(Fred) is given by Equation (3), where h is the formation ef-
ficiency of the reactive excited state W* (from the literature
hꢂ0.55[13]).Reduction of the tungstate involves two compo-
nents: reaction with the hydrogen donor RH (kr; see
Scheme 3) and reaction with the solvent (kb; this path in-
cludes any contribution to the reduction by the counter
cation TBA+ or by any impurity) in competition with decay
of the excited state (kd).
FACHTRE(UNG neat)
0.09
0.004
MeCN[a,c]
MeCN[b]
9”104
[a] From flash photolysis experiments.[b] From steady-state experiments.
[c] Under the simplified assumption that the polytungstate is reduced
only by the solvent; compare reference [13].
oxygen, such as the ketone triplet state.As an example, ben-
zophenone has kr = 7.5”105, 1.8”106, and 1.3”102 mꢀ1 sꢀ1
with cyclohexane, propan-2-ol, and MeCN.[19a–c] Similarly,
TBADT has a much broader scope as a photocatalyst than
titanium dioxide, to which it is often compared, since it is a
stronger oxidant, able to activate alkanes and their simple
derivatives, on which the latter is hardly active.[19d]
kb þ kr½RHŠ
ð3Þ
Fred
¼
h
kd þ kb þ kr½RHŠ
This can be expressed by means of the Stern–Volmer type
Equation (4),[18] where K = k /
r A(kd + kb) and the value of
Fred in neat acetonitrile [F8 = hkb/A(kd + kb)], as well as the
Radical reactions—neat solvent: As is apparent from com-
parison of Figure 4a,b and Figure 5a,b, the quantum yields
for polytungstate reduction (Fred) in flash photolysis (FP)
and in the steady-state (SS) experiments are different.The
initial values of Fred (SS) with cyclohexane are much lower
(about 2%) than the corresponding FP values at the same
RH concentration, while Fred (SS) with propan-2-ol is larger
than Fred (FP), about twice as much at low alcohol concen-
tration.The quantum yield of reduction in neat MeCN—
F8 (SS)—is also much lower (4% of the FP value).
CTHREUNG
CTHREUNG
limiting value from extrapolation at growing [RH] (Flim
h) are introduced.
ꢃ
Fo þ Flim K½RHŠ
ð4Þ
Fred
¼
l þ K½RHŠ
The quantum yields of formation of (monoelectronic) re-
duced tungstate in the presence of either cyclohexane or
propan-2-ol in FP experiments are plotted in Figure 4b and
This is due to further reactions of the radicals formed in
the primary step, in particular to paths leading to reoxida-
4158
www.chemeurj.org
ꢀ 2006 Wiley-VCH Verlag GmbH & Co.KGaA, Weinheim
Chem. Eur. J. 2006, 12, 4153 – 4163

Photosensitized Alkylation of Electrophilic Alkenes
FULL PAPER
tion of the catalyst (see Scheme 3).As it appears from
Figure 6, the amount of reduced polytungstate remains un-
changed over several ms.On a longer timescale, however, a
further evolution takes place.[12] This involves disproportio-
Steady accumulation of HW causes a dramatic drop in the
steady-state concentration of cyclohexyl radicals during the
course of the process, and this actually becomes so low after
a small (ca.5%) amount of conversion that formation of
coupling products terminates.Such a limited conversion has
been observed with other donors upon photocatalysis in the
absence of oxygen.[5a]
4ꢀ
nation of the monoelectronic reduced anion HW10O32
(HW) to yield the two-electron reduced polytungstate
4ꢀ
H2W10O32 (kds = 9”104 mꢀ1 sꢀ1).[12] Furthermore, HW is in
part reoxidized by reaction with the alkyl radicals, either
through back hydrogen transfer (kbh) or through electron
transfer to give a carbanion (krd; see Scheme 3).However,
the alkyl radical R· is also consumed through reactions not
involving HW, which are typically quite fast (kc >1”
109 mꢀ1 sꢀ1).[20] The unbalance of kds and kc causes the pro-
gressive accumulation of reduced tungstate during the pro-
cess and correspondingly the decrease of the alkyl radicals,
Conversion back to MeCN is also important for the
·CH2CN radicals formed by H abstraction from the solvent.
Succinonitrile has been detected in small amounts under
these conditions, but this is clearly a minor process.[12a] Ac-
tually, the steady-state value of Fred (SS) in neat MeCN is
0.004, 22 times smaller than the flash photolysis value, indi-
cating that regeneration of TBADT and MeCN (rate con-
stant k’bh, see Scheme 3, right-hand side) efficiently com-
petes with coupling of the cyanomethyl radicals (k’c = 1.2”
109 mꢀ1 sꢀ1).[25]
C
since the [HW]/[R ] ratio is determined by the reaction be-
tween the two species (kbh + krd).This is a classical case of
the persistent radical effect, as evidenced by Fischer et al.[21]
In the case of cyclohexane, the radicals undergo both di-
merization to bicyclohexyl and disproportionation to cyclo-
hexene and cyclohexane[16,20] [the latter reaction has not
been specifically assessed here, but it obviously contributes
to the observed Fred (SS)].According to Scheme 3, the
steady-state yield of reduced tungstate is a fraction of that
initially formed, as expressed by the efficiency factor
[Eq.(5)]; see Equation (6).
Unlike in the cyclohexane case, the steady-state quantum
yield of tungstate reduction with propan-2-ol is larger than
the flash photolysis value.Thus, at [2-PrOH] = 0.5m, the in-
itial value of Fred is 1.86 times larger than Fred (FP), and
plotting of Fred (SS) versus the starting alcohol concentra-
tion (Figure 4a) gives an apparent K value 3.4 times larger
[12,21,26]
than the FP-derived value.As previously suggested,
but not quantitatively assessed, this is due to the oxidation
of the a-hydroxyalkyl radical by the decatungstate (indi-
cated as W) as shown in Scheme 4 and Equation (7).
kc½RCŠ
kc½RCŠ
ð5Þ
ð6Þ
a ¼
ffi
Me2CCOH þ W ! Me2CO þ HW
ð7Þ
ðkbh½HWŠ þ krd½HWŠ þ kc½RCŠ fðkbh þ krdÞ½HWŠg
kb þ akr½RHŠ
Fo þ aFlim K½RHŠ
l þ K½RHŠ
Fred ðSSÞ ¼ h
¼
kd þ kb þ kr½RHŠ
The Fred (SS) versus [C6H12] plot (Figure 5a) and compari-
son with Figure 5b indicate that a is slightly over 1%.The
C
rate constant for the back reaction HW + R can be deter-
mined (8”109 mꢀ1 sꢀ1; see below) through the competition
with addition to electrophilic alkenes.Hydrogen transfer
(kbh) is by far the main path in this reaction (Table 4), while
formation of the cyclohexyl anion, as deduced from the pro-
portion of cyclohexyl methyl ketone produced through trap-
ping by MeCN, corresponds to about 5% (krd ꢂ4”
108 mꢀ1 sꢀ1).[22] This relatively low value is reasonable, be-
cause electron transfer is largely endothermic as evaluated
on the basis of the reduction potential of the sec-butyl radi-
Scheme 4.Competing paths in the photocatalyzed activation of propan-2-
ol.
Indeed, at a low starting propan-2-ol concentration (=
0.1m), Fred (SS)ffi2”Fred (FD), indicating that virtually every
ketyl radical is oxidized.This is in accord with the easy oxi-
dation of such radicals—E(Me2C·OH/Me2C+OH)
=
ꢀ
[23]
C
cal [E
A
compare EAHCTRE(UNG HW/
ꢀ0.61 V versus SCE[27]—and, coupled with the rapid depro-
tonation of the cation to give acetone, supports electron
transfer according to Equation (7) as a viable pathway.[19]
The rate constant for radical reaction (kox) was again deter-
mined by competition with addition to electrophilic alkenes
(see below).It may be noted that the excess of Fred (SS)
over Fred (FP) diminishes with increasing alcohol concentra-
tion (compare Figure 4a/b), reasonably because when a rela-
W) = ꢀ0.97 V[24].]
Table 4.Bimolecular rate constants ( k, mꢀ1 sꢀ1) for the radical reactions
in the TBADT photocatalytic processes in MeCN (see Scheme 3,
Scheme 4).
C6H12
2-PrOH
kc 1”109
kad 2”106
kbh 8”109
krd 4”108
kad 1.5”108
kox 2”1010
C
tively high Me2C OH concentration is generated, oxidation
according to Equation (7) results in local depletion of the
Chem. Eur. J. 2006, 12, 4153 – 4163
ꢀ 2006 Wiley-VCH Verlag GmbH & Co.KGaA, Weinheim
www.chemeurj.org
4159

A.Albini et al.
decatungstate and thus in incomplete oxidation of the radi-
cals.[28a] As a result, the limiting value of Fred (SS) is only
slightly larger than Fred (FP).
Figure 1 and Figure 3 show that the rather efficient oxida-
tion of propan-2-ol to acetone (quantum yield 0.32 at 0.5m
alcohol starting concentration) stops when all of the starting
amount of decatungstate has been reduced.[28b]
Thermal oxidation of the initially formed a-oxygen-substi-
tuted radicals may also occur with ethers and acetals such as
tetrahydrofuran and 1,3-dioxolane, in view of the relatively
accessible oxidation potentials of such species (oxidation po-
tential of the 2-tetrahydrofuranyl radical = ꢀ0.35 V versus
SCE, of the dioxolanyl radical = ꢀ0.9 V;[29] deprotonation,
in this case from the a carbon, is a fast process).[30] In con-
trast, oxidation is excluded for alkyl radicals, for which the
oxidation potential has a positive value versus SCE.[30] In
the last case, oxidation (again coupled with deprotonation)
occurs when a more easily reduced polyoxometalate (e.g., a
phosphapolytungstate)is used.[24]
ping and back hydrogen transfer is expressed in Equa-
tion (8).
Fad
kad½ANŠ
ð8Þ
¼
Fred ðFPÞ
fðkbh þ krdÞ½HWŠ þ kad½ANŠg
This allows the rate of the latter process to be evaluated,
since Fred is measured under the same conditions as Fad.On
the basis of the moderate rate of addition to AN (kad ca.2”
106 mꢀ1 sꢀ1),[32] this gives (kbh +krd) = 8”109 mꢀ1 sꢀ1.In con-
junction with the a value determined above, this gives a
·
[HW]/[cyclo-C6H11 ] ratio of 800 under steady-state condi-
tions in neat MeCN in the early phase of the reaction (com-
·
pare with [HW]/[cyclo-C5H9 ] = 900, measured by EPR in
the reaction with cyclopentane).[21] The proportion of re-
duced anion increases during the conversion, resulting in a
decreased rate of alkylation, which nevertheless proceeds
steadily and reaches
a satisfactorily turnover number
(Figure 3).The final step in the alkylation process is reduc-
tion of the adduct radical with regeneration of TBADT.
This occurs by electron followed by proton transfer, as indi-
cated by the experiments in 0.5% D2O/CH3CN and in
CD3CN, in accord with the much easier reduction of a-
cyano-substituted alkyl radicals (of known high electron af-
finity)[33] with respect to their unsubstituted counterparts.
This is obviously a key step, both for allowing the catalytic
role of the tungstate[34a] and for the reaction selectivity.As
for the latter point, this introduces some limit to alkene con-
centration.Indeed, when a 1 m AN concentration is used,
trapping of the radical adduct by the alkene (a relatively
slow process, kꢂ1”104 mꢀ1 sꢀ1)[32] becomes competitive with
electron transfer from HW and certain amounts of the di-
meric and trimeric adducts are formed.[11]
Radical reactions—trapping by electrophilic alkenes: As dis-
cussed above, with alcohols TBADT is a consumed reagent
rather than a catalyst, whist with alkanes efficient back hy-
drogen transfer and HW accumulation severely limit the
alkyl radical concentration.Thus, although for opposite rea-
sons, the reactions stop at an early stage in both cases, de-
priving an otherwise promising (high rate constant for H ab-
ꢀ
straction; see above) method of photoinduced C H bond
activation of preparative interest.The reaction continues in
oxygen-equilibrated solutions because molecular oxygen re-
generates TBADT, thus maintaining the operation of the
photocatalytic cycle, besides trapping the organic radical to
give hydroperoxides.[6a,31] Under these conditions, Tanielian
indeed found that a plot of the SS quantum yield of oxygen
consumption versus cyclohexane concentration gave a value
of K quite close to that obtained from the tungstate reduc-
tion in flash photolysis experiments.[13] Furthermore, the
yield of hydroperoxides from cyclohexane corresponded to
the amount of oxygen adsorbed, at least up to a certain
level of conversion.Scheme 1b is therefore operative and
reduced tungstate does not accumulate.
In the case of propan-2-ol, initial H abstraction is much
more efficient, resulting in fast accumulation of the reduced
form HW.[34b] On the other hand, trapping by AN is also
more efficient (kad ꢂ1.5”108 mꢀ1 sꢀ1).[32] Since the alkylation
cycle regenerates TBADT (Scheme 4), reduced HW is accu-
mulated to a smaller extent with respect to neat solvent (see
Figure 2), in contrast with what happens in the slow alkyla-
tion by cyclohexane, where the determining factor is back H
transfer.The initial quantum yield of alkylation is somewhat
larger than Fred (FP), while the quantum yield for the oxida-
tion to acetone is 29% of the value in the absence of the
trap.These data confirm that oxidation is practically the
only process from the a-hydroxyalkyl radical.The ratio be-
tween the quantum yield of acetone in the absence and in
As the above data show, electrophilic alkenes are likewise
efficient traps.Table 1 shows that a,b-unsaturated nitriles,
dinitriles, esters and ketones (0.1m) are smoothly alkylated
under these condition up to complete conversion and, except
for the case of cyclohexenone (where competitive light ad-
sorption may have a negative effect), the isolated yields of
the alkylation products are always >50% and in most cases
>70%.
the presence of AN corresponds to the ratio kox[W]/ACHTREUNG(kox[W]
+ kad [AN]) and allows calculation of kox = 2”1010 mꢀ1 sꢀ1.
From the mechanistic point of view, the data show that
trapping by AN contributes significantly to the decay of cy-
clohexyl radicals, which brings about an increase in accumu-
lated HW (see Figure 2) due to the decreased contribution
of back hydrogen transfer.Trapping is not complete, though,
and the initial quantum yield of alkylation (0.06) corre-
sponds to 22% of the initial yield of radicals as determined
by flash photolysis Fred (FP).The competition between trap-
This high (diffusion-controlled) value is consistent with oxi-
C
dation being the main process for the Me2C OH radical
under these conditions (kox [W]@kbh [HW]), as mentioned
above.Other oxygen-substituted radicals such as those from
tetrahydrofuran and dioxolane are likewise efficiently trap-
ped.As with simple alkyl radicals, the final step is electron
transfer to the (a-cyano) radical adduct, as established by
deuteration experiments in the THF case.
4160
www.chemeurj.org
ꢀ 2006 Wiley-VCH Verlag GmbH & Co.KGaA, Weinheim
Chem. Eur. J. 2006, 12, 4153 – 4163

Photosensitized Alkylation of Electrophilic Alkenes
FULL PAPER
It is interesting to compare the effects of oxygen and of
alkenes on TBADT-photoinduced reactions.Both reagents
are effective radical traps and the tungstate is regenerated
in their presence, allowing conversion to proceed.The reac-
tion with oxygen is complicated by the fact that the peroxy
radical formed may give different reactions according to the
substrate chosen and the conditions.[6b] Alcohols or ketones
are thus formed from alkane hydroperoxides in different
amounts.Trapping by unsaturated nitriles, esters, and ke-
tones is a clean process, since it produces an electrophilic
radical that is efficiently reduced by HW, avoiding compet-
ing processes.Furthermore, the alkylation can also be car-
ried out in air-equilibrated solutions.As shown in Figure 1,
in the case of propan-2-ol the alkylation efficiency actually
becomes somewhat larger than under oxygen exclusion,
after a short induction time, and reduced tungstate accumu-
kylation[34a] are chain reactions, and the limiting quantum
yield is in both cases about 0.5, close to the expected value
Flim = h.As one might expect, alkylation is more efficient
with more nucleophilic radicals—Fad is 10 times larger with
propan-2-ol than with cyclohexane, for example.This makes
reaction with a “moderately activated” (a-hydroxy-substitut-
ed) alkyl derivative particularly convenient (e.g., it can also
be carried out in air-equilibrated solution; indeed, it gives
slightly better results under those conditions), but does not
detract from the preparative value of alkylation with an
alkane, since there are no significant competing processes
and so a good chemical yield can be obtained, although
after a longer irradiation time.
These results demonstrate that TBADT photocatalysis is
a viable and rather general method for the mild functionali-
ꢀ
zation of aliphatic C H bonds, and that, besides for oxygen-
lates to a small degree.Dissolved oxygen ([O ] = 2”10ꢀ3 m,
ations, it can be conveniently used for C C bond-forming
ꢀ
2
with kO ffi5”109 mꢀ1 sꢀ1)[35] competes with AN ([AN]
=
reactions.Thermal analogy for such process is quite limited,
variously involving the use of high temperature and pressur-
e,[37a] the preparation and cleavage of peroxides,[37b] or a cat-
alyzed process based on some sophisticate metal complex
that requires a much more elaborate setup than the photo-
chemical synthesis.[38,39] The reactions presented above are
paradigmatic examples of 100% atom efficiency and the
small amount of catalyst used, as well as the simple experi-
mental setup, both during irradiation (e.g., no anhydrifica-
tion or deoxygenation required) and during workup, make
this method an example of “green” synthesis, comparing fa-
2
0.1m, with kad ꢂ1.5”108 mꢀ1 sꢀ1) for the short-lived radicals
but on the other hand contributes to the reoxidation of HW,
a persistent species, and thus to the overall efficiency of the
process.The first effect causes the induction period, but the
oxygen concentration then decreases, since equilibration
with the atmosphere does not compensate for chemical con-
sumption, and the alkylation proceeds under favorable con-
ditions.With cyclohexane, on the other hand, alkylation is
considerably slowed down (the rate becomes a quarter of its
original value) in an aerated solution, consistently with the
fact that the lower rate of addition to AN (kad ꢂ2”
106 mꢀ1 sꢀ1) makes trapping of the radicals by oxygen more
important.Nevertheless, a good yield of the alkylated prod-
uct is also obtained in this case, though after a longer irradi-
ation time.
vorably with other approaches to CACHTREHUNG functionalization.
Experimental Section
Materials: Tetrabutylammonium decatungstate (TBADT)[12a] and unsatu-
rated dinitrile IPMN[41] were prepared by published procedures.The
other materials and the solvents were of commercial origin.
Conclusion
Preparative irradiations: These were carried out in 1 cm diameter quartz
tubes containing a solution (6 mL) of TBADT (40 mg, 0.002m), the
alkane (0.5m), and the electrophilic alkene (0.1m).These were purged
with purified argon for 10 min, septum-capped, and irradiated for 15–
20 h in a multilamp apparatus fitted with six 15-W phosphor-coated
lamps (center of emission 310 nm).The irradiated solution was flushed
with oxygen until colorless, passed through a 4-cm layer of neutral alumi-
na, and concentrated.The residue was first examined by GC/MS and
then chromatographed on silica gel with cyclohexane and cyclohexane/
ethyl acetate 98:2 mixture as eluents.With IPMN essentially the same re-
sults were also obtained when degassing was omitted.
Compounds 1a,[40,42] 2a,[43] 3a,[44] 4a,[11] and 5[45] had physical characteris-
tics identical to those of the compounds prepared by alternative proce-
dures as reported in the literature.Bicyclohexyl and cyclohexyl methyl
ketone were recognized and determined by comparison of their GC/MS
spectra with those of authentic samples.
As mentioned in the Introduction, the wealth of data from
previous research support the contention that the high reac-
tivity of photoexcited decatungstates might make these ma-
ꢀ
terials highly suited for the mild activation of aliphatic C H
bonds.However, this rarely produces high chemical conver-
sion, either because of a rapid drop in the alkyl radical con-
centration after a few percent conversion (due to reaction
with the persistent reduced tungstate), or because of the cat-
alyst consumption when an easily oxidized radical is used.
Useful reactions are obtained when the radical is trapped ef-
ficiently by a path that regenerates the catalyst,[36] as is
indeed the case for oxygenation, which can be carried out to
a rather high level of conversion (>50%).[6] In the present
case, the alkene serves the same function as oxygen: it both
traps the radicals and regenerates the catalyst (in this case
indirectly, via the radical adduct as shown in the cycle in
Scheme 3).Competition between trapping by the above al-
kenes and the other available paths allows the key kinetic
parameters for the reactions of the radicals generated by
photocatalysis to be obtained.Neither oxygenation [31] nor al-
2-Cyano-4-hydroxy-3,3,4-trimethylpentanenitrile (1b): Colorless crystals,
m.p. 138–1408C; 1H NMR (300 MHz, CDCl3, TMS): d = 1.15 (s, 6H),
1.3 (s, 6H), 4.4 (brs, each, 1H), 5.2 ppm (s, 1H); IR (KBr): n˜ = 3415,
2170 cmꢀ1
.
3-(2-Tetrahydrofuranyl)propanenitrile (2a) [D1]-2: Oil that solidifies on
standing; 1H NMR (300 MHz, CDCl3, TMS): d = 1.5 (m, 1H), 1.7–2.1
(m, 5H), 2.5 (m, 1H—2H in the non-deuterated analogue), 3.7–4.0 ppm
(m, 3H); 13C NMR (75 MHz, CDCl3): d = 13.9 (t, J = 20.6 Hz, CHD),
Chem. Eur. J. 2006, 12, 4153 – 4163
ꢀ 2006 Wiley-VCH Verlag GmbH & Co.KGaA, Weinheim
www.chemeurj.org
4161

A.Albini et al.
25.6 (CH2), 30.9 (CH2), 31.1 (CH2), 67.7 (CH2), 77.0 (CH), 119.6 ppm
(CN).
[9] Addition of a carbanion, formed by reduction of a primarily formed
radical, to MeCN to give an imine—and a methyl ketone from it—
proceeds to a high level of conversion with some polycyclic alkanes
2-Cyano-3-methyl-3-(2-tetrahydrofuranyl)butanenitrile (2b): Oil that sol-
1
(see further below) C.M. Prosser-McCartha, C.L. Hill,
J. Am.
idifies on standing; H NMR (300 MHz, CDCl3, TMS): d = 1.15 (s, 3H),
Chem. Soc. 1990, 112, 3671.R.F.Renneke, M.Kadkhodayan, M.
Pasquali, C.L.Hill, J. Am. Chem. Soc. 1991, 113, 8357.
[10] Radicals in Organic Synthesis, (Eds:. P.Renaud, M.P.Sibi), Wiley-
VCH, 2001.
1.3 (s, 3H), 1.7 (m, 1H), 1.95 (m, 3H), 3.75–3.9 (m, 3H), 4.0 ppm (s,
1H); 13C NMR (75 MHz, CDCl3): d
= 18.2 (CH3), 21.1 (CH3), 26.1
(CH2), 26.2 (CH2), 32.8 (CH), 41.7 (C), 68.8 (CH2), 81.5 (CH), 111.9
(CN), 112.3 ppm (CN); IR (KBr): n˜ = 2230, 1070 cmꢀ1
.
[11] D.Dondi, M.Fagnoni, A.Molinari, A.Maldotti, A.Albini,
Eur. J. 2004, 10, 142.
[12] a) T.Yamase, N.Takabayashi, M.Kaji, J. Chem. Soc. Dalton Trans.
1984, 793; b) A.Hiskia, E.Papaconstantinou, Inorg. Chem. 1992, 31,
163.
Chem.
2-Cyano-3-[2-(1,3-dioxolanyl)]-3-methylbutanenitrile
(3b):
1H NMR
(300 MHz, CDCl3, TMS):
d = 1.25 (s, 6H), 3.9 (s, 1H), 3.95–4.1
(AA’BB’, 4H), 4.8 ppm (s, 1H); 13C NMR (75 MHz, CDCl3): d = 20.1
(CH3), 29.8 (CH), 41.3 (C), 65.5 (CH2), 105.8 (CH2), 111.6 ppm (CN); IR
(KBr): n˜ = 2230, 1070 cmꢀ1
.
[13] C.Tanielian, K.Duffy, A.Jones, J. Phys. Chem. A J. Phys. Chem. B
Steady-state measurements: Solutions (2 mL) of TBADT in MeCN
(0.002m) containing the appropriate additives in 1-cm optical path spec-
trophotometric cuvettes were degassed by five freeze–degas–thaw cycles
(to 10ꢀ6 Torr).The samples were irradiated by means of a focalized
Osram 150 W high-pressure mercury arc fitted with an interference filter
at 313 nm or a band-pass filter (the former for quantum yield measure-
ments, the latter for comparing reactivities).Light absorbed was deter-
mined by photometer and the light flux was measured by ferrioxalate ac-
tinometry.Formation of the alkylated products was determined by GC
on the basis of calibration curves with use of benzophenone as internal
standard).Formation of the two reduced forms of TBADT was moni-
tored by spectroscopy in the visible region on the basis of the known mo-
lecular extinction coefficients of such molecules.[12,13b] Acetone was deter-
mined by HPLC by the method reported by Franco.[46]
1997, 101, 4276.C.Tanielian, R.Seghrouchni, C.Schweitzer,
J. Phys.
Chem. A J. Phys. Chem. B 2003, 107, 1102.Tanielian and co-workers
compared the results with sodium and tetrabutylammonium deca-
tungstate in neat MeCN and concluded that about 40% of the
quenching was due to the solvent.
[14] A.Chernseddine, C.Sanchez, J.Livage, J.P.Launay, M.Fournier,
Inorg. Chem. 1984, 23, 2609.
[15] The initially generated LMCT excited state of this anion decays
(30 ps) to a further state with a lifetime of about 55 ns in MeCN.
The electronic structure of the latter state is similar to that of the
first, with a predominantly W-based HOMO; see reference [16].
[16] D.C.Duncan, T.L. Netzel, C.L.Hill,
[17] L.P. Ermolenko, J.A. Delaire, C. Giannotti,
Inorg. Chem. 1995, 34, 4640.
J. Chem. Soc. Perkin
Trans. 1 1997, 25; I.Texier, J.F.Delouis, J.A.Delaire, C.Giannotti,
P.Plaza, M.M.Martin, Chem. Phys. Lett. 1999, 311, 139; A.Hiskia,
E.Papaconstantinou, Polyhedron 1988, 7, 477.
Flash photolysis measurements: Flash photolysis measurements were car-
ried out with an Applied Photophysics kinetic spectrometer and use of
the fourth harmonic of a Nd-YAG Lumonics HY 200 laser.The quantum
yield of transient formation was determined by comparison with the
[18] Equation (4) is obtained from Equation (3) by dividing numerator
and denominator by (kd + kb).This approach has been previously
used by Tanielian; see reference [13].
ketyl radical formed by excitation of benzophenone in propan-2-ol (e540
[47]
= 3220mꢀ1 cmꢀ1
)
from benzophenone as the reference.
[19] a) L.Giering, M.Berger, C.Steel, J. Am. Chem. Soc. 1974, 96, 953;
b) L.J.Johnston, D.J.Lougnot, V.Wintgens, J.C.Scaiano,
J. Am.
Chem. Soc. 1988, 110, 518; c) Y.M.A. Naguib, C. Steel, S.G.
Cohen, M.A. Young, J. Phys. Chem. 1987, 91, 3033; d) L.Cerme-
nati, D.Dondi, M.Fagnoni, A.Albini,
Tetrahedron 2003, 59, 6409.
[20] M.Simic, E.Hayon, J. Phys. Chem. 1971, 75, 1677.
Acknowledgements
[21] T.Kothe, R.Martshke, H.Fischer,
J. Chem. Soc. Perkin Trans. 2
1998, 503.
Support of this work by the MURT, Rome, through a PRIN Program is
gratefully acknowledged.
[22] The alkane may in part result from protonation of the carbanion, a
path that has previously been demonstrated to have some role with
polycyclic alkanes; see further below.
[23] D.Occhialini, S.U.Pedersen, H.Lund,
Acta Chem. Scand. 1990, 44,
[1] R.H.Crabtree, Chem. Rev. 1985, 85, 245.R.H.Crabtree,
J. Chem.
715.
Soc. Dalton Trans. 2001, 2437.A.Maldotti, A.Molinari, R.Amadel-
li, Chem. Rev. 2002, 102, 3811.
[2] B.M. Trost, Acc. Chem. Res. 2002, 35, 695–705.P.Anastas, J.
Warner, Green Chemistry: Theory and Practice, Oxford University
Press, Oxford, 1998.S.Luo, Y.Peng, B.Zhang, PG. .Wang, JP. .
Cheng, Curr. Org. Synth. 2004, 1, 405–429.
[24] R.F.Renneke, M.Pasquali, C.L.Hill,
6585.
[25] H.G.Korth, P.Lommes, W.Sicking, R.Gustmann,
J. Am. Chem. Soc. 1990, 112,
Int. J. Chem.
Kinet. 1983, 15, 267.
[26] D.Sattari, C.L.Hill,
Chem. Commun. 1990, 634.D.Sattari, C.L.
Hill, J. Am. Chem. Soc. 1993, 115, 4649.
[27] T.Lund, D.D.M.Wayner, M.Jonsson, A.G.Larsen, K.Daasbjerg,
J. Am. Chem. Soc. 2001, 123, 12590.
[3] A.E.Shilov, G.B.Shul ’pin, Chem. Rev. 1997, 97, 2879; G.A.Olah,
A.Molnar, Hydrocarbon Chemistry, Wiley, New York, 1995; R. H.
Crabtree, Chem. Rev. 1995, 95, 987; S.Niu, M.B.Hall,
Chem. Rev.
[28] a) Take into account that in steady-state irradiation with a practical-
ly useful concentration of the photocatalyst the absorbance is essen-
tially limited to a thin layer of solution; the local concentration of
the transients may thus be higher than expected.b) Oxidation was a
clean process and acetone was not significantly consumed in the
timescale required for its formation from propan-2-ol.Papaconstan-
tinou had previously found that the polytungstate-photocatalyzed
oxidation of acetone in water was much slower than that of propan-
2-ol; A.Mylonas, A.Hiskia, E.Androulaki, D.Dimotikali, E.Papa-
constantinou, Phys. Chem. Chem. Phys. 1999, 1, 437.
2000, 100, 353; M.Torrent, M.Sola, G.Frenking, Chem. Rev. 2000,
100, 439; Activation and Functionalization of Alkanes, (Ed.: C. L.
Hill), Wiley, New York, 1989; A.A. Fokin, P.R. Schreiner, Chem.
Rev. 2002, 102, 1551.
[4] A.Albini, M.Fagnoni, M.Mella, Pure Appl. Chem. 2000, 72, 1321;
A.Albini, M.Fagnoni, Green Chem. 2004, 6, 1.
[5] a) C.L. Hill, Synlett 1995, 127; b) C.Tanielian, Coord. Chem. Rev.
1998, 178–180, 1165.
[6] a) R.C. Chambers, C.H. Hill, Inorg. Chem. 1989, 28, 2509; b) A.
Maldotti, R.Amadelli, V.Carassiti, A.Molinari,
Inorg. Chim. Acta
[29] D.D.M. Wayner, J.J. Dannenberg, D. Griller,
1988, 110, 131.D.D.M.Wayner, A.Houman,
1998, 52, 377.
J. Am. Chem. Soc.
Acta Chem. Scand.
1997, 256, 309; c) A.Molinari, R.Amadelli, A.Mazzacani, G.Sar-
tori, A.Maldotti, Langmuir 2002, 18, 5400.
[7] a) B.S. Jaynes, C.L. Hill,
J. Am. Chem. Soc. 1993, 115, 12212;
b) B.S.Jaynes, C.L.Hill, J. Am. Chem. Soc. 1995, 117, 4704.
[8] Z.Zeng, C.L.Hill, Chem. Commun. 1998, 2467.
[30] E.Fasani, M.Mella, A.Albini,
449.
J. Chem. Soc. Perkin Trans. 2 1995,
4162
www.chemeurj.org
ꢀ 2006 Wiley-VCH Verlag GmbH & Co.KGaA, Weinheim
Chem. Eur. J. 2006, 12, 4153 – 4163

Photosensitized Alkylation of Electrophilic Alkenes
FULL PAPER
[31] A contribution by radical chain processes through peroxy radicals
cannot be excluded in alkane photooxygenation, but seems to have
no major effect on the kinetics, see references [13, 16b] and A.
Ioannidis, E.Papaconstantinou, Inorg. Chem. 1992, 31, 163.
efficiently with regeneration of the catalyst; see krd in Scheme 3 and
C.M.Prosser-McCartha, C.L.Hill,
3671.
J. Am. Chem. Soc. 1990, 112,
[37] a) JO. .Metzger, K.Klenke, J.Hartmanns, D.Eisermann,
Chem.
[32] H.Fischer, L.Radom,
Angew. Chem. 2001, 113, 1380; Angew.
Ber. 1986, 119, 508; b) S.Araneo, F.Fontana, F.Minisci, F.Recu-
pero, A.Serri, J. Chem. Soc. Chem. Commun. 1995, 1399.G.G.Ga-
lust’yan, V.G.Glukhovtsev, Khim. Geterotsikl. Soedin. 1984, 20, 21.
[38] K.A. Miller, J.M. Bartolin, R.M. OꢁNeill, R.D. Sweeder, T.M.
Chem. Int. Ed. 2001, 40, 1340.The addition of alkyl radicals to elec-
trophilic alkenes is largely exothermic: DH for the addition of the
ꢀ1
C
(CH)3C OH radical to acetonitrile, for example, is ꢀ27 kcalmol
(K.Hꢂberger, H.Fischer, Int. J. Chem. Kinet. 1993, 25, 913).It is
not expected that the reaction should be reversible at room temper-
ature.
Owens, J.W. Kampf, M.M. Banaszak Holl, N.J. Wells,
J. Am.
Chem. Soc. 2003, 125, 8986.
[39] In the case of alkylation by alcohols, an obvious alternative is to
prepare the ketyl radical by reduction of the corresponding ketone;
for the synthesis of 1a, for example, see reference [40].
[33] G.N.Merril, G.D.Dahlke, S.R.Kass,
4462.
J. Am. Chem. Soc. 1996, 118,
[34] a) A significant contribution from hydrogen H abstraction from the
aliphatic derivatives, which would produce a chain process, is unlike-
ly at room temperature.In fact, even in the most favorable case ( a-
cyanoalkyl radical and propan-2-ol) the process remains endother-
mic (by ca.4 kcalmol ꢀ1 in that case), and experiments with a non-re-
generated photoinitiator (e.g., benzophenone) clearly show that no
chain process is involved (unpublished work by the authors).A
chain process is certainly operative in high-temperature non-cata-
lyzed thermal alkylations; see, for example, reference [37a].b) The
ratio [HW]/[R·] must be larger than in the cyclohexane case, and has
indeed been found to be “large” by epr; see reference [21].
[40] M.Atobe, S.Fujiwara, T.Naneko,
123.
Bull. Chem. Soc. Jpn. 2002, 75,
[41] G.Buchbauer, E.Dworan, Monatsh. Chem. 1980, 111, 1165.
[42] C.A.Grob, A.Waldner, Helv. Chim. Acta 1979, 62, 1854.
[43] M.Agourrody, E.Montaudon, B.Maillard, Can. J. Chem. 1987, 65,
2694.H.Yorimitsu, K.Wakabayashi, H.Shinokubo, K.Oshima,
Bull. Chem. Soc. Jpn. 2001, 74, 1963.
[44] P.B.Shresta-Dawadi, J.Lugtenburg, Eur. J. Org. Chem. 2003, 4564.
[45] N.Matsumoto, J.Kumanotani, Tetrahedron Lett. 1975, 16, 3643.
[46] D.R. Cardoso, S.M. Bettin, E.V. Reche, B.S. Lima-Neto, S.W.
Franco, J. Food Compos. Anal. 2003, 16, 563.
[35] B.Maillard, K.C.Ingold, J.C.Scaiano,
5095.
J. Am. Chem. Soc. 1983, 105,
[47] E.J.Land, Proc. R. Soc. London Ser. A 1968, 305, 407.
[36] The same holds when another transient derived from the radical is
trapped, as is the case for the carbanions derived from some polycy-
clic alkanes that, contrary to the case of simple alkanes, are formed
Received: October 5, 2005
Published online: March 7, 2006
Chem. Eur. J. 2006, 12, 4153 – 4163
ꢀ 2006 Wiley-VCH Verlag GmbH & Co.KGaA, Weinheim
www.chemeurj.org
4163