Photoredox transformations with dimeric gold complexes

Article abstract of DOI:10.1002/anie.201306727

Let the sunshine in! Unactivated alkyl and aryl bromides underwent a light-enabled reductive radical cyclization in the presence of a dimeric phosphine-gold complex as a photocatalyst (see scheme; X=C(CO ₂Et)₂, NR, O). Sunlight can be used as the energy source for this simple and efficient radical reaction, which does not require potentially hazardous and toxic chemical reagents, such as organostannanes and chemical initiators.

Full text of DOI:10.1002/anie.201306727

.
Angewandte
Communications
DOI: 10.1002/anie.201306727
Radical Reactions
Photoredox Transformations with Dimeric Gold Complexes**
Guillaume Revol, Terry McCallum, Mathieu Morin, Fabien Gagosz, and Louis Barriault*
Dedicated to Professor Pierre Deslongchamps on the occasion of his 75th birthday
The conversion of solar energy into chemical energy in
photosynthesis has enthralled scientists for decades. Sunlight
could be used as an inexpensive, green, and sustainable source
of energy to induce chemical reactions.[1,2] However, many
organic compounds do not absorb sunlight or visible light, and
the application of photochemistry in synthesis is therefore
restricted. In response to this limitation, several light-
absorbing photocatalysts have been developed in an effort
to mimic the light-harvesting abilities of biocomplexes found
in natural processes, such as photosynthesis.[3] The inorganic
complex [Ru(bpy)3Cl2] (1) is one of the most used one-
electron photoredox catalysts for research in energy storage,
water splitting, and photovoltaic devices.[4,5] Increased atten-
tion is being placed on visible-light-mediated photoredox
processes for the development of efficient and waste-mini-
mizing processes of general applicability that should decrease
our dependence on toxic chemical products.[6]
During the last decade, several metal complexes, such as
[Ru(bpy)3Cl2] (1), [Ir(ppy)2(dtbbpy)PF6] (2), and fac-[Ir-
Scheme 1. General mechanism for metal-catalyzed photoredox reac-
tions. EWG=electron-withdrawing group.
(ppy)3] (3), have proven their potential in absorbing light
and transforming it into electronic energy that can be used in
ꢀ
ꢀ
ꢀ
organic synthesis to generate new C C, C N, and C O bonds
through a single-electron-transfer (SET) process.[6f] In gen-
eral, the photoredox transformation can operate through two
distinctive pathways, the oxidative quenching cycle (path A)
or the reductive quenching cycle (path B; Scheme 1). In the
former, the photoexcited metal complex transfers an electron
ꢀ
The reduction of carbon–halogen (C X) bonds to pro-
duce carbon-centered radical intermediates is one of the most
useful processes for accessing new reactivity and has broad
scope for application in chemical synthesis.[7] Typically,
methods to generate alkyl, vinyl, and aryl radicals from
carbon–halogen bonds require the use of potentially hazard-
ous and/or toxic chemical reagents, such as organostannane
compounds and chemical initiators (AIBN, Et3B/O2,
K2S2O8).[8] The use of photoexcited RuII (path A) and IrII/III
polypyridine complexes 1–3 is an attractive alternative for
efficient synthetic transformations involving the cleavage of
carbon–halogen bonds (path A).[9–12] However, the reduction
potential of these catalysts limits the range of possible radical
intermediates to those derived from substrates with highly
activated or weak carbon–halogen bonds, such as polyhalo-
methanes, bromomalonates, electron-deficient benzyl halides,
and alkyl/aryl iodides [Eq. (1), Scheme 1].[11,12] Although
important advances in photocatalysis have been made in the
last decade, one might concede that efficient methods for the
photocatalyzed reductive scission of unactivated carbon–
halogen bonds have yet to be described.
ꢀ
to an oxidative quencher, such as an alkyl or aryl halide (R
X). The radical intermediate is converted into a product,
whereas the oxidized metal returns to its original oxidation
state by accepting an electron from a tertiary amine base
(R3ND). In the latter pathway, the photoexcited metal complex
acts as an oxidant and accepts an electron from a sacrificial
electron donor source (R3ND) to generate a complex [M]nꢀ1
that may function as a reducing agent.
[*] Dr. G. Revol, T. McCallum, M. Morin, Prof. L. Barriault
Centre For Catalysis, Research and Innovation
Department of Chemistry, University of Ottawa
10 Marie Curie, Ottawa, On, K1N 6N5 (Canada)
E-mail: lbarriau@uottawa.ca
Dr. F. Gagosz
Dꢀpartement de Chimie, UMR 7652, CNRS/Ecole Polytechnique
91128 Palaiseau (France)
In this context, the design of a photocatalytic system of
broader applicability must fulfill specific criteria. Aside from
an appropriate redox-potential window, the catalyst must
possess a long-lived excited state to undergo productive
oxidative quenching (path A) with alkyl, vinyl, and aryl
halides. Furthermore, photoexcitation of the catalyst should
occur within a readily accessible wavelength range and thus
[**] We thank the Natural Sciences and Engineering Research Council
(for Accelerator and Discovery grants to L.B.) and the University of
Ottawa (for a University Research Chair to L.B.) for support of this
research. We also thank Prof. Derek Pratt (University of Ottawa) for
insightful discussions.
Supporting information for this article is available on the WWW
under http://dx.doi.org/10.1002/anie.201306727.
13342
ꢀ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2013, 52, 13342 –13345

Angewandte
Chemie
require no special equipment. On the basis of the rich
photochemistry of dimeric gold complexes, we report herein
a conceptually novel photocatalytic system for the efficient
generation of carbon-centered radical intermediates from
unactivated alkyl (E8reduction = ꢀ1.90 to ꢀ2.50 V versus the
saturated calomel electrode (SCE)) and aryl bromides
(E8reduction = ꢀ2.05 to ꢀ2.57 V versus SCE) under remarkably
mild conditions [Eq. (2)].[13]
no cyclization product 6 was observed when the visible-light
photocatalysts 1–3 were used under standard conditions (with
a 23 W compact fluorescent light (CFL) or white-light light-
emitting diodes (LEDs); Table 1, entries 10–12).[11,12] These
results confirm without ambiguity that 1) no background
reactions took place in the absence of sunlight (UVA) or the
gold(I) dimeric complex and 2) photoexcited [Ru(bpy)3Cl2]
(1), [Ir(ppy)2(dtbbpy)PF6] (2), and fac-[Ir(ppy)3] (3) cannot
reduce unactivated alkyl carbon–bromide bonds by metal-to-
ligand charge transfer (MLCT).
On the basis of studies by Che and co-workers, we
considered the use of cationic [Au2(m-dppm)2]2+ (dppm =
bis(diphenylphosphanyl)methane) species as potential photo-
catalysts.[14] These dimeric gold species absorb light in the UV
region (lmax = 295 nm) to produce an in situ high-energy and
long-lived photoexcited state (quantum yield: 0.23) with
The scope of the reaction with respect to the alkyl or aryl
bromide substrate was explored (Tables 2 and 3). The radical
cyclization of alkyl bromides 7–10 with iPr2NEt (2 equiv;
procedure A) in acetonitrile gave the corresponding cyclic
products 14–17 in yields ranging from 58 to 93% (Table 2,
entries 1–7). Irradiation with sunlight rather than UVA light
led to a slight improvement in the reaction yield (Table 2,
entries 2, 4, and 6). Surprisingly, complete conversion of the
tosyl allylamine 11 into the corresponding pyrrolidine 18
required a prolonged irradiation time of 36 h (Table 2,
entry 8). To decrease the reaction time, we examined various
hydrogen donors. It was found that the addition of formic acid
in the presence of iPr2NEt greatly enhanced the rate of the
reaction.[17] Indeed, the conversion of 11 was completed after
1 h to produce 18 in 66% yield (Table 2, entry 9). As
expected, the cyclization of bromides 12 and 13 gave the
desired pyrrolidines 19 and 20 in 86 and 63% yield,
respectively (Table 2, entries 10 and 11). Labeling experi-
ments with DCO2D (2 equiv) and iPr2NEt (2 equiv) indicated
that iPr2NEt is the main hydrogen-atom source [Table 2,
Eq. (3)].[18] Although superior reactivity was observed for
photoredox reactions with formic acid, insufficient data are
available at present to permit a detailed mechanistic dis-
cussion.[19]
3+
a strong reduction potential [E8(Au2 !Au22+*) = ꢀ1.6 to
-ꢀ1.7 V versus SSCE (sodium-saturated calomel elec-
trode)].[15] To demonstrate the applicability of dimeric gold
complexes in synthesis, we selected the radical cyclization of
bromoalkene 5 as the benchmark reaction. A survey of
dimeric gold catalysts in conjunction with trialkyl amine bases
as sacrificial electron and hydrogen donors revealed that the
radical cyclization of alkyl bromide 5 to give 6 proceeded in
excellent yield in the presence of diisopropylethylamine
(iPr2NEt) in acetonitrile (MeCN) with sunlight as the light
source (Table 1, entry 2). Although the counteranion had
little effect on the conversion of the radical cyclization, we
found the gold catalyst 4e[16] to be the most robust and
photostable dimeric gold complex (Table 1, entries 3–6).
UVA light (315–400 nm) proved to be a good surrogate for
sunlight (Table 1, entry 7). Rigorous control experiments
showed complete recovery of the starting material (0%
conversion) when the reaction was carried out in the absence
of a gold catalyst 4 or light (Table 1, entries 8 and 9). Similarly,
Table 1: Optimization of the photoredox reaction.[a]
Having confirmed the applicability of the dimeric gold
complexes in photoredox processes with alkyl bromides, we
examined the reductive cleavage of aryl bromides (Table 3).
The reductive radical cyclization of aryl bromide 21 afforded
the cyclized product 27 as a mixture (71:29) of isopropyl- and
isopropenyl-substituted compounds in 74% yield (Table 3,
entry 1). Biaryl compounds 28 and 29 were formed in
excellent yields (Table 3, entries 2–5), and the cyclization of
sulfonamide 24 proceeded mainly by ipso substitution as
anticipated to give the biaryl amine 30 as the major product
along with tricycle 31 in 93% overall yield (Table 3,
entry 6).[20] The reductive dehalogenation of aryl bromides
25 and 26 by procedure B led to the formation of indoline 32
and sulfonanilide 33 in 90 and 94% yield, respectively. As for
the alkyl bromides series, control experiments confirmed that
the SET reduction of aryl bromides did not operate in the
absence of the dinuclear gold complex or with catalyst 2 or
3.[21]
Entry
Catalyst
Light source
Yield [%]
1[b]
2
3
4
5
[Au2(m-dppm)2](OTf)2 (4a)
[Au2(m-dppm)2](OTf)2 (4a)
[Au2(m-dppm)2](NTf2)2 (4b)
[Au2(m-dppm)2](BF4)2 (4c)
[Au2(m-dppm)2](SbF6)2 (4d)
[Au2(m-dppm)2]Cl2 (4e)
[Au2(m-dppm)2]Cl2 (4e)
–
[Au2(m-dppm)2]Cl2 (4e)
[Ru(bpy)3Cl2] (1)
[Ir(ppy)2(dtbbpy)]PF6 (2)
fac-[Ir(ppy)3] (3)
sunlight
sunlight
sunlight
sunlight
sunlight
sunlight
UVA (315–400 nm)
sunlight/UVA
–
23 W CFL/LEDs
23 W CFL/LEDs
23 W CFL/LEDs
20
94
72
70
90
86
74
s.m.
s.m.
s.m.
s.m.
<5
6
7
8[c]
9[d]
10[e]
11[e]
12[e]
[a] All reactions were performed in degassed solvent with pyrex glass-
ware. [b] Et3N was used instead of iPr2NEt. [c] The reaction mixture was
irradiated for 18 h. [d] The reaction mixture was heated at 608C for 24 h.
[e] The reaction mixture was irradiated for 36 h. bpy=2,2’-bipyridyl,
dtbbpy=di-tert-butyl-2,2’-bipyridine, ppy=2-phenylpyridinate,
s.m.=starting material, Tf=trifluoromethanesulfonyl.
Angew. Chem. Int. Ed. 2013, 52, 13342 –13345
ꢀ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.angewandte.org 13343

.
Angewandte
Communications
Table 2: Photoredox cyclization of alkyl bromides.
Table 3: Photoredox cyclization and dehalogenation of aryl bromides.
Entry Substrate
Product
Procedure[a]
(type of light) (ratio)[b]
Yield [%]
Entry
Substrate
Product
Procedure[a]
(type of light)
Yield [%]
(ratio)
74
(71:29)
1
2
A (UVA)
A (sunlight)
81 (73:27)
92 (65:35)
1
A (sunlight)
21
27a,b
7
14a,b
15a,b
16
2
3
A (UVA)
A (sunlight)
75[b]
80
3
4
A (UVA)
A (sunlight)
58 (84:16)
81 (86:14)
8
22
23
28
29
4
5
A (UVA)
A (sunlight)
90[b]
91
5
6
A (UVA)
A (sunlight)
58
60
9
7
A (UVA)
93
10
17
18
93[b]
(63:37)
6
A (UVA)
8
9
A (UVA)
B (UVA)
63
66[c]
11
7
8
B (UVA)
B (UVA)
90[b]
10
B (UVA)
B (UVA)
86 (50:50)
25
26
32
33
12
19a,b
11
63
94[b]
13
20
[a] Procedure A: [Au2(m-dppm)2]Cl2 (2.5 mol%), iPr2NEt (2 equiv),
MeCN, 2–8 h; procedure B: [Au2(m-dppm)2]Cl2 (2.5 mol%), iPr2NEt
(5 equiv), HCO2H (2 equiv), MeCN, 1–4 h. [b] Ratio of the isopropyl- to
the isopropenyl-substituted product. [c] The reaction was completed in
1 h.
[a] Procedure A: [Au2(m-dppm)2]Cl2 (2.5 mol%), iPr2NEt (2 equiv),
MeCN, 2–8 h; procedure B: [Au2(m-dppm)2]Cl2 (2.5 mol%), iPr2NEt
(5 equiv), HCO2H (2 equiv), MeCN, 1–4 h. [b] The reaction mixture was
irradiated for 8–16 h; [Au2(m-dppm)2]Cl2 (5.0 mol%) was used.
We sought to further validate the gold-catalyzed photo-
induced carbon–carbon bond-forming reaction by including
intermolecular processes (Scheme 2). The radical addition of
1-bromopropanol (34) and N-chlorophthalidimide (37) to the
diallylated malonate 35 produced the cis-substituted cyclo-
pentanes 36 and 38 in 77 and 38% yield (d.r. 5:1), respectively
[Eqs. (4) and (5), Scheme 2]. Notably, aryl bromide 39 was
converted into dihydrophenanthrene 40 in the presence of
formic acid (procedure B) in 44% yield [Eq. (6), Scheme 2].
To gain more insight into the reaction mechanism, we exposed
aryl bromide 39 to UVA light for 16 h in the presence of
iPr2NEt and obtained the dibromodibenzyl product 41 in
33% yield. Irradiation of compound 41 according to proce-
dure B produced dihydrophenanthrene 40 in 64% yield along
with the dehalogenation product 42 in 9% yield [Eq. (7),
Scheme 2]. These results suggest that the transformation most
likely proceeds by an initial recombination of benzyl radi-
cals.[22] The ring closure leading to compound 40 could then be
explained by the subsequent generation of an aryl radical,
which would attack the other aromatic ring in a formal ipso
substitution of the bromine atom.
13344 www.angewandte.org
ꢀ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2013, 52, 13342 –13345

Angewandte
Chemie
[4] K. Kalyanasundaram, Coord. Chem. Rev. 1982, 46, 159.
[5] A. Juris, V. Balzani, F. Barigelletti, S. Campagna, P. Belser, A.
von Zelewsky, Coord. Chem. Rev. 1988, 84, 85.
[6] For recent reviews, see: a) L. Shi, W. Xia, Chem. Soc. Rev. 2012,
41, 7687; b) J. Xuan, W.-J. Xiao, Angew. Chem. 2012, 124, 6934;
Angew. Chem. Int. Ed. 2012, 51, 6828; c) J. M. R. Narayanam,
´
C. R. J. Stephenson, Chem. Soc. Rev. 2011, 40, 102; d) F. Teply,
Collect. Czech. Chem. Commun. 2011, 76, 859; e) T. P. Yoon,
M. A. Ischay, J. Du, Nat. Chem. 2010, 2, 527; f) D. W. MacMillan,
D. A. Rankic, C. A. Prier, Chem. Rev. 2013, 113, 5322.
[7] a) Radicals in Organic Synthesis (Eds.: P. Renaud, M. P. Sibi),
Wiley-VCH, Weinheim, 2001; b) D. P. Curran in Comprehensive
Organic Synthesis, Vol. 4 (Eds.: B. M. Trost, I. Fleming, M. F.
Semmelhack), Pergamon, Oxford, 1991, pp. 715 – 779; c) Topics
in Current Chemistry, Radicals in Synthesis I and II, Vol. 263 –
264 (Ed.: A. Gansꢀuer), Springer, Berlin, 2006; d) Encyclopedia
of Radicals in Chemistry, Biology and Materials, Vol. 1 and 2
(Eds.: C. Chatgilialoglu, A. Studer), Wiley, Chichester, 2012.
[8] For examples of alternatives to tin hydride mediators, see: a) B.
Quiclet-Sire, S. Z. Zard, Pure Appl. Chem. 2011, 83, 519; b) C.
Chatgilialoglu, Acc. Chem. Res. 1992, 25, 188; c) P. A. Baguley,
J. C. Walton, Angew. Chem. 1998, 110, 3272; Angew. Chem. Int.
Ed. 1998, 37, 3072; d) C. Ollivier, P. Renaud, Chem. Rev. 2001,
101, 3415; e) A. Studer, S. Amrein, Synthesis 2002, 835; f) B. C.
Gilbert, A. F. Parsons, J. Chem. Soc. Perkin Trans. 2 2002, 367;
g) M. E. Weiss, L. M. Kreis, A. Lauber, E. M. Carreira, Angew.
Chem. 2011, 123, 11321; Angew. Chem. Int. Ed. 2011, 50, 11125;
h) J. A. Murphy, T. A. Khan, S. Z. Zhou, D. W. Thomson, M.
Mahesh, Angew. Chem. 2005, 117, 1380; Angew. Chem. Int. Ed.
2005, 44, 1356; i) A. Ekomiꢁ, G. Lefꢂvre, L. Fensterbank, E.
Lacꢃte, M. Malacria, C. Ollivier, A. Jutand, Angew. Chem. 2012,
124, 7048; Angew. Chem. Int. Ed. 2012, 51, 6942.
[9] D. A. Nicewicz, D. W. C. MacMillan, Science 2008, 322, 77.
[10] H. W. Shih, M. N. Vander Wal, R. L. Grange, D. W. C. MacMil-
lan, J. Am. Chem. Soc. 2010, 132, 13600.
[11] J. D. Nguyen, E. M. DꢄAmato, J. M. R. Narayanam, C. R. J.
Stephenson, Nat. Chem. 2012, 4, 854.
[12] H. Kim, C. Lee, Angew. Chem. 2012, 124, 12469; Angew. Chem.
Int. Ed. 2012, 51, 12303.
Scheme 2. Gold-catalyzed intermolecular radical cyclization reactions.
[13] a) S. Rondinini, P. R. Mussini, P. Muttini, G. Sello, Electrochim.
Acta 2001, 46, 3245; b) A. J. Fry, R. L. Krieger, J. Org. Chem.
1976, 41, 54.
[14] H.-L. Kwong, V. Wing-Wah Yam, D. Dan Li, C.-M. Che, J.
Chem. Soc. Dalton Trans. 1992, 3325.
[15] C.-M. Che, H.-L. Kwong, V. Wing-Wah Yam, K.-C. Choc, J.
Chem. Soc. Chem. Commun. 1989, 885.
[16] C. King, J.-C. Wang, M. N. I. Khan, J. P. Fackler, Jr., Inorg.
Chem. 1989, 28, 2145.
[17] M. R. J. Narayanam, J. W. Tucker, C. R. J. Stephenson, J. Am.
Chem. Soc. 2009, 131, 8756.
[18] No deuterium incorporation was observed when the reaction
was performed in CD3CN. These results are in concordance with
the deuteration experiments described by Stephenson and co-
workers.[17]
In summary, we have described a light-enabled reductive
radical reaction of unactivated alkyl and aryl bromides in the
presence of a dimeric phosphine–gold complex as the photo-
catalyst. It was demonstrated that the applicability of this
photoredox process is not limited to intramolecular processes,
but that intermolecular transformations are also possible.
Although no clear mechanism has been established with
respect to the use of formic acid, our results clearly
demonstrate the uniqueness of this photoredox process and
its high synthetic potential. We are currently exploring the
application of this method in the context of synthesis and
working toward a full understanding of the reaction mecha-
nism.
[19] No significant differences in the isopropyl/isopropenyl ratio
were observed for entries 1–4 and 10 in Table 2 when proce-
dure B (with formic acid) was used instead of procedure A. Poor
conversion was observed when the Hantzsch ester and NADH
(the reduced form of nicotinamide adenine dinucleotide) were
used as hydrogen donors.
[20] W. B. Motherwell, A. M. K. Pennell, J. Chem. Soc. Chem.
Commun. 1991, 877.
[21] See the Supporting Information for control experiments.
[22] a) K. Hironaka, S. Fukuzumi, T. Tanaka, J. Chem. Soc. Perkin
Trans. 2 1984, 1705; b) J.-M. Kern, J.-P. Sauvage, J. Chem. Soc.
Chem. Commun. 1987, 546.
Received: August 1, 2013
Revised: September 11, 2013
Published online: October 16, 2013
Keywords: cyclization · gold · photoredox catalysis ·
.
radical reactions · sunlight
[1] G. Ciamician, Science 1912, 36, 385.
[2] A. Albini, M. Fagnoni, Green Chem. 2004, 6, 1.
[3] M. Fagnoni, D. Dondi, D. Ravelli, A. Albini, Chem. Rev. 2007,
107, 2725.
Angew. Chem. Int. Ed. 2013, 52, 13342 –13345
ꢀ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.angewandte.org 13345