Photochemical activity of a key donor-acceptor complex can drive stereoselective catalytic α-alkylation of aldehydes

Article abstract of DOI:10.1038/nchem.1727

Asymmetric catalytic variants of sunlight-driven photochemical processes hold extraordinary potential for the sustainable preparation of chiral molecules. However, the involvement of short-lived electronically excited states inherent to any photochemical reaction makes it challenging for a chiral catalyst to dictate the stereochemistry of the products. Here, we report that readily available chiral organic catalysts, with well-known utility in thermal asymmetric processes, can also confer a high level of stereocontrol in synthetically relevant intermolecular carbon-carbon bond-forming reactions driven by visible light. A unique mechanism of catalysis is proposed, wherein the catalyst is involved actively in both the photochemical activation of the substrates (by inducing the transient formation of chiral electron donor-acceptor complexes) and the stereoselectivity-defining event. We use this approach to enable transformations that are extremely difficult under thermal conditions, such as the asymmetric α-alkylation of aldehydes with alkyl halides, the formation of all-carbon quaternary stereocentres and the control of remote stereochemistry.

Full text of DOI:10.1038/nchem.1727

ARTICLES
PUBLISHED ONLINE: 11 AUGUST 2013 | DOI: 10.1038/NCHEM.1727
Photochemical activity of a key donor–acceptor
complex can drive stereoselective catalytic
a-alkylation of aldehydes
1
†
1
†
1
1,2
´
´
Elena Arceo , Igor D. Jurberg , Ana Alvarez-Fernandez and Paolo Melchiorre
*
Asymmetric catalytic variants of sunlight-driven photochemical processes hold extraordinary potential for the sustainable
preparation of chiral molecules. However, the involvement of short-lived electronically excited states inherent to any
photochemical reaction makes it challenging for a chiral catalyst to dictate the stereochemistry of the products. Here, we
report that readily available chiral organic catalysts, with well-known utility in thermal asymmetric processes, can also
confer a high level of stereocontrol in synthetically relevant intermolecular carbon–carbon bond-forming reactions driven
by visible light. A unique mechanism of catalysis is proposed, wherein the catalyst is involved actively in both the
photochemical activation of the substrates (by inducing the transient formation of chiral electron donor–acceptor
complexes) and the stereoselectivity-defining event. We use this approach to enable transformations that are extremely
difficult under thermal conditions, such as the asymmetric a-alkylation of aldehydes with alkyl halides, the formation of
all-carbon quaternary stereocentres and the control of remote stereochemistry.
ascinated by the ability of natural photosynthetic systems to reactive radical species under the action of a photoredox catalyst12,13
,
convert solar energy into chemical energy1, the scientific com- is differentiated clearly from the stereoselective ground-state process
munity long ago recognized the potential of light-driven reac- controlled by a distinct chiral catalyst.
F
tions (photochemistry) as a powerful approach to chemical
Herein we describe how some of the effective tools available for
synthesis2. From the high-energy intermediates generated by photo- catalytic asymmetric reactions in the ground state can be translated
induced excitation of organic molecules, unique reaction manifolds into the realm of photochemical activation to address the challenges
can be accessed that are generally unavailable to conventional of enantioselective photochemistry. We found that readily available
thermal pathways. This explains why photochemical reactions con- chiral amines, with an established profile as catalysts of thermal
siderably enrich the synthetic repertoire of modern organic che- asymmetric processes, can exert high stereocontrol in synthetically
mists3,4. However, there has been very limited success in relevant intermolecular carbon–carbon bond-forming reactions
developing asymmetric catalytic photoreactions in solution that driven by visible light. The catalysts do not contain any photosensi-
create chiral molecules with a well-defined three-dimensional tive unit, but rather they guide the photoactivation of the substrates
spatial arrangement5,6. This has instilled the general perception by inducing the transient formation of photon-absorbing chiral
that photochemistry is too unselective to parallel the impressive electron donor–acceptor (EDA) complexes14,15. A mechanism
levels of efficiency reached by the asymmetric catalysis of thermal (Fig. 1) that involves an in-cage radical combination as the stereode-
reactions7. Historically, two fundamental issues frustrated the devel- fining step is proposed and discussed based on diverse experimental
opment of asymmetric photochemical variants: (1) the inability of evidence. To our knowledge this study constitutes the first example
most organic molecules to absorb light in the visible spectrum, of photochemical asymmetric catalysis that involves an EDA
which calls on specialized apparatus for high-energy irradiation, complex of the reacting partners as the radiation acceptor.
and (2) the low-barrier, very rapid processes that proceed from
short-lived electronic excited states. The latter condition challenges Results and discussion
the ability of a chiral catalyst to bias and organize the molecular Our previous studies in asymmetric organocatalysis led us to
architecture of the fleetingly excited states of photoreactions, a pre- wonder if the synthetic potential of this approach16 could be
requisite for high stereoinduction.
expanded by combining it with the seemingly distinct field of
One effective catalytic strategy, among the very few reported photochemistry. It is well-established that chiral secondary amines
so far8, uses a chiral organic catalyst appropriately adorned with of type I (Fig. 1) effectively condense with aldehydes to form reactive
hydrogen-bonding motifs to bind a specific substrate selectively9,10
.
nucleophilic enamine intermediates II, and confer high stereoselec-
The catalyst, modified to act as a photosensitizer, excites the tivity to their subsequent thermal reactions17. However, the enam-
substrate via photoinduced electron transfer (ET), thus directing ine’s potential to participate actively in the photoexcitation of
the resulting intramolecular cyclization towards a stereoselective substrates remains almost unexplored18,19. Still, the ability of tertiary
pathway. The approach is limited by the intrinsic weakness of the amines to form EDA complexes with electron-accepting molecules
hydrogen-bonding interactions11, which complicates the geometri- of high electron affinity15,20 suggests that the lone pair of the pyrro-
cal organization of the catalyst–substrate assembly, and requires a lidine ring in the enamine II (Fig. 1) could engage in such a molecu-
specially tailored catalyst. Recently, a viable alternative was ident- lar aggregation. A few precedents in the literature18, and the low
ified in that the photochemical activation step, which furnishes ionization potentials of the pyrrolidine-based enamines of type II
1
2
´
Institute of Chemical Research of Catalonia, Avenguda Pa¨ısos Catalans 16-43007 Tarragona, Spain, Institucio Catalana de Recerca i Estudis Avanc¸ats,
Passeig Llu´ıs Companys 23-08010 Barcelona, Spain; †These authors contributed equally to this work. e-mail: pmelchiorre@iciq.es
*
NATURE CHEMISTRY | ADVANCE ONLINE PUBLICATION | www.nature.com/naturechemistry
1
© 2013 Macmillan Publishers Limited. All rights reserved.

NATURE CHEMISTRY DOI: 10.1038/NCHEM.1727
ARTICLES
We first examined the possibility of EDA associations between
enamines II and electron acceptors by combining an excess of
butanal (2a, 15 equiv.) and amines 1a and 1b (1 equiv.) in
methyl tert-butyl ether (MTBE) with 2,4-dinitrobenzyl bromide
(3a, 1 equiv., Table 1). The commercially available diarylprolinol
silylether catalysts 1a and 1b were selected because of their potential
to induce high enantioselectivity in thermal reactions of aldehydes
that proceed through enamine formation28. The choice of the nitro-
benzyl bromide 3a was motivated by the low reduction potentials of
nitroaromatics, which allow them to engage in EDA complexes with
tertiary amines by means of n ꢀ p* interactions20. The bromide
would be the leaving group needed to originate a reactive benzylic
radical within the chiral intermediate of type VI. Immediately
after mixing with the bromide 3a, the MTBE solution developed a
marked yellow–orange colour, and its optical absorption spectrum
showed a bathochromic displacement in the visible spectral
region, diagnostic of an EDA complex (Fig. 2a). We recognized
the formation of the coloured EDA complex IV and the tendency
of alkyl bromide radical anions to dissociate29 as key to developing
an asymmetric catalytic photochemical strategy without the need for
a photosensitizer.
Preliminary results (Table 1) revealed the feasibility of the
approach. When using a household 23 W compact fluorescent
light (CFL) bulb to irradiate an MTBE solution of 2a (3 equiv.)
and 3a in the presence of catalysts 1a and 1b, we observed the
quantitative formation of the desired a-benzylated product 4a
with a promising level of optical purity (Table 1, entries 1 and
2, 83% enantiomeric excess (e.e.) with catalyst 1b). Analogous
results were obtained using either 2,6-lutidine or sodium acetate
as base (Table 1, entries 2 and 3, respectively). Careful exclusion
of light or of the aminocatalyst completely inhibited the reaction,
even on heating at 50 8C (Table 1, entries 4–6). The use of a blue
light-emitting diode (LED) as a monochromatic light source
(lmax ¼ 460 nm) reduced the reaction rate only slightly
(Table 1, entry 7), which indicates that absorption in the visible
O
EWG
O
H
H
R
R
Established tools for
thermal reactions
N
H
+
N
I
EWG
N
Donor
H
II
R
Colourless
R
Thermal reactivity
Photochemical reactivity
Acceptor
LG
+
N
EWG
H
III
R
VI
EWG
LG–
Colourless
LG
LG
+
N
N
e
– transfer
–
H
IV
V
R
R
hν
EWG
EWG
Coloured EDA complex
Chiral radical ion pair
Figure 1 | Mechanistic proposal for asymmetric catalytic photochemical
processes. Transient generation of chiral enamine-based EDA complexes IV
exploited to access photochemical pathways. Central to this study is the
possibility for substances that do not absorb visible light to become coloured
on ground-state EDA association. EWG, electron-withdrawing group; LG,
leaving group; filled grey circles represent the chiral fragment of the
aminocatalyst scaffold.
(for example, 1-(but-1-enyl)pyrrolidine has an ionization potential
of 7.2 eV)21, qualify them as potential donors for facilitating EDA
associations in the ground state.
Table 1 | Explorative studies on the feasibility of the catalytic
asymmetric photochemical process.
O
Catalyst 1 (20 mol%)
NO2
O
NO2
Typically, EDA complexes are characterized by the appearance of
a weak absorption band, the charge-transfer band, associated with
an ET transition from donor to acceptor22. In many cases, the
energy of this transition lies within the visible-frequency range.
This raises the tantalizing possibility of using easily managed
visible light to activate substances that would not normally absorb
in the visible spectrum. Surprisingly, although the photoexcitation
of EDA complexes is studied extensively23, their use in chemical
synthesis is scarcely reported24–26. This is mainly because of an
unproductive, fast reverse ET, which restores the ground-state
EDA complex and thus renders any further reactivity improbable.
Given these intrinsic difficulties, the question arises as to whether
the enamine-based EDA complex IV could be harnessed to access
a synthetically useful photochemical pathway. A viable path can
be envisaged in that visible light irradiation of IV might induce
ET to occur, and so produce the chiral contact radical ion pair V.
As a critical factor, the presence of a suitable leaving group (LG in
V) within the radical anion partner may trigger a fragmentation
event rapid enough to compete with the reverse ET. This would pro-
ductively render the positively charged intermediate pair VI, which
brings two radicals within a geometrically restricted chiral space and
in very close proximity. This condition should facilitate a stereocon-
trolled radical combination to form a new carbon–carbon bond and
forge the stereogenic centre. This photochemical process would
result in the asymmetric intermolecular a-alkylation of aldehydes
with alkyl halides12,13, a synthetically useful catalytic transformation
that cannot be realized under thermal control27.
Visible light (23 W CFL)
H
H
Br
+
2,6-lutidine (1 equiv.)
2 MTBE, [3a]0 = 0.5 M, 25 °C
Me
NO2
3 equiv.
Me
NO
4a
2a
3a
Catalysts used in this study
H
Ar
Ar
Ar
N
Ar
OTMS
N
H
OTMS
H
H
1c
Ar = 3,5-(CF3)2-C6H3
1a Ar = C6H5
1b Ar = 3,5-(CF3)2-C6H3
1c
Entry
Catalyst
Light
Time
Yield (%)
e.e. (%) 4a
1
1a
1b
1b
1b
1b
–
1b
1b
1c
ON
ON
ON
OFF
6 h
6 h
6 h
48 h
48 h
48 h
16 h
40 h
48 h
98
98
94
0
0
0
89
78
87
75
83
82
–
–
–
82
84
92
2
3*
4
5
6
7
8
9
OFF, 50 8C
ON
ON, LED†
ON, in air
ON
*Reaction performed using
13.8 W m22. TMS, trimethylsilyl.
1
equiv. NaOAc instead of 2,6-lutidine. †460 nm LED, irradiance
2
NATURE CHEMISTRY | ADVANCE ONLINE PUBLICATION | www.nature.com/naturechemistry
© 2013 Macmillan Publishers Limited. All rights reserved.

NATURE CHEMISTRY DOI: 10.1038/NCHEM.1727
ARTICLES
a
0.7
0.6
0.5
0.4
0.3
0.2
0.1
0.0
2a:1b:3a (15:1:1)
2a:1b (15:1)
2a
2a:1b
(15:1)
3a
1b:3a
(1:1)
2a:1b:3a
(15:1:1)
3a
2a
[3a] = 0.2 M
400
500
600
700
Wavelength (nm)
b
3.0
2.5
2.0
1.5
1.0
0.5
0.0
4.0
3.5
3.0
2.5
2.0
1.5
1.0
0.5
0.0
Br
2a:1b:3g (15:1:1)
3g
2a:1b:3h (15:1:1)
3h
Br
O
O
N
N
2a:1b (15:1)
2a:1b (15:1)
Coloured EDA complex
Coloured EDA complex
[3g] = 0.2 M
[3h] = 0.2 M
365
465
565
665
365
465
565
665
3.0
2.5
2.0
1.5
1.0
0.5
0.0
4.0
Aldehyde:1b:
3a (3:0.2:1)
3a
Aldehyde:
1b (3:0.2)
Aldehyde:1b:3g
(3:0.2:1)
3.5
3.0
2.5
2.0
1.5
1.0
0.5
0.0
Br
Br
Aldehyde:1b
(3:0.2)
O
N
N
O2N
3g
NO2
Coloured EDA complex
[3g] = 0.5 M
Coloured EDA complex
[3a] = 0.5 M
390
490
590
Wavelength (nm)
690
790
370
470
570
Wavelength (nm)
670
770
c
Catalyst 1b (20 mol%)
O
O
O
O
Ar
Ar
OTMS
Light
Dark Light
R
Light
Dark Light
R
R
R
H
H
Br
+
H
H
N
+
+
Br
H
2,6-lutidine (1 equiv.)
MTBE, [3]0 = 0.25 M, 25 °C
2,6-lutidine (1 equiv.)
MTBE, [3]0 = 0.5 M, 25 °C
3 equiv.
Me
Me
Me
Me
2a
3 equiv.
3a or 3g
1b
4a or 4g
2a
3a or 3g
4a or 4g
1 equiv. 1 equiv.
100
100
80
60
40
20
0
80
60
40
20
0
3a conversion
4a e.e.
3a conversion
3g conversion
3g conversion
0
5
10
0
1
2
3
4
5
6
7
8
9
Time (h)
Time (h)
Figure 2 | Mechanistic investigations. a, Optical absorption spectra (recorded in MTBE in 1 mm path quartz cuvettes using a Shimadzu 2401PC UV-visible
spectrophotometer) and visual appearance of the separate reaction components and of the coloured EDA complex. b, Representative optical absorption
spectra of EDA complexes with enamines and extended enamines formed in situ (recorded in MTBE in 1 mm path quartz cuvettes using a Shimadzu 2401PC
UV-visible spectrophotometer). c, Successive intervals of irradiation and dark periods for reactions of bromobenzyl 3a and bromophenacyl 3g promoted by
20 mol% (left) and 1 equiv. (right) of catalyst 1b. a.u, arbitrary units.
region is sufficient for the reaction to occur. The reduced reactivity
As for the stereoselectivity of the reaction, we found that it could
observed in the presence of oxygen (Table 1, entry 8) is consonant be improved by using amine 1c as catalyst (92% e.e., Table 1, entry 9).
with a radical mechanism.
This is presumably because of the increased steric demand of the
NATURE CHEMISTRY | ADVANCE ONLINE PUBLICATION | www.nature.com/naturechemistry
3
© 2013 Macmillan Publishers Limited. All rights reserved.

NATURE CHEMISTRY DOI: 10.1038/NCHEM.1727
ARTICLES
Table 2 | Survey of the aldehydes and alkyl bromides that can participate in the catalytic asymmetric photochemical
alkylation.
1c (20 mol%)
hν (23 W CFL)
O
O
H
+
H
Br
EWG
EWG
2,6-lutidine (1 equiv.)
MTBE, [3]0 = 0.5 M, 25 °C
R
R
2
3
4
Entry
Substrate
Product
Entry
Substrate
Product
O
O
O
NO2
O
1
7
Br
H
H
2b
4b
H
( )5
( )5
Me
3h
O
4h
Me
O
NO2
Me
70% yield, 86% e.e.
81% yield, 92% e.e.
Ph
O
O
NO2
O
2
3
8
9
Br
H
H
H
2c
4c
O
4i
3i
Me
O
Me
Me
O
Me
Ph
NO2
Me
94% yield, 86% e.e.
80% yield, 93% e.e.
NO2
O
O
H
H
Br
NO2
2d
4d
NO2
H
NO2
NO2
O
3j
Me
O
4j
92% yield, 93% e.e.
86% yield, 86% e.e.
O
O
O
NO2
4
5
10
Br
H
H
H
2e
( )3
4e
( )3
4k
70% yield, 83% e.e.
O
3k
TIPSO
TIPSO
NO2
NO2
Me
73% yield, 90% e.e.
CN
O
O
CN
11
O
Br
Br
H
Br
H
Br
3l
NO2
Me
NO2
O
4l
3f
4f
Me
96% yield, 87% e.e.
95% yield, 84% e.e.
O
O
O
S
O
6*
12
Br
Br
H
H
S
O
O
3g
4g
3m
Me
92% yield, 87% e.e.
Me
4m
89% yield, 94% e.e.
Entries 1–4 were carried out in the presence of bromide 3a; entries 5–12 were carried out in the presence of aldehyde 2a. *Reaction performed under natural sunlight irradiation on the roof-top of the Institute of
Chemical Research of Catalonia, Tarragona (Spain), on a partially cloudy day (27 Feb 2013, from 13:00 untill 18:00). TIPS, triisopropylsilyl.
catalyst’s chiral fragment, which should impose greater confor- Pyrex vessel on a rooftop effected the asymmetric catalytic alky-
mational constraints on the short-lived diradical intermediate of lation of 2a with phenacyl bromide 3g (Table 2, entry 6, 5 hours
type VI (Fig. 1).
reaction time, 89% yield, 94% e.e.).
As shown in Table 2, aldehydes that bear sterically hindered
The strategy also enabled transformations that are extremely dif-
chains or heteroatom moieties were also benzylated stereoselectively ficult through thermal mechanisms. It forged all-carbon quaternary
in the presence of 3a (Table 2, entries 1–4). Moreover, we found that stereocentres with high fidelity, as demonstrated by the reaction of
other bromide-containing acceptors combined productively with 2-phenylpropanal 5 with 3a to afford the enantioenriched product
chiral enamines in photon-absorbing EDA associations. In addition 6 (86% e.e., Fig. 3a). In addition, this approach demonstrates a
to electron-deficient benzylic systems (Table 2, entry 5), a broad great potential for targeting stereocentres remote from the carbonyl
array of phenacyl bromides participated effectively in the enantiose- moiety. The absorption spectra recorded for the EDA complexes of
lective alkylation of butanal 2a (Table 2, entries 6–12, 70–96% iso- bromides 3a and 3g and the extended enamines, generated by con-
lated yield, 83–94% e.e.). The reaction protocol is operationally densation of the aminocatalyst with unsaturated aldehydes, are red
simple, conducted at ambient temperature with readily available shifted compared to those of the enamine of type II, and new bands
substrates and catalysts, and using household CFL bulbs as the in the visible region become well resolved (Fig. 2b). We found that
light source. Additionally, natural solar light effectively promoted the extended enamine intermediates efficiently initiated light-
the process: simply placing the reaction mixture in an ordinary induced asymmetric alkylation setting a new stereocentre at a
4
NATURE CHEMISTRY | ADVANCE ONLINE PUBLICATION | www.nature.com/naturechemistry
© 2013 Macmillan Publishers Limited. All rights reserved.

NATURE CHEMISTRY DOI: 10.1038/NCHEM.1727
ARTICLES
hydrogen abstraction or dimerization derivatives. In addition,
experiments with successive intervals of irradiation and dark
periods were performed; these resulted in a total interruption of
the reaction progress in the absence of light, and recuperation of
reactivity on further illumination (Fig. 2c). These results
demonstrate that light is a necessary component of the reaction.
Although they do not definitively rule out a radical-chain
mechanism, the data show that any chain-propagation process
must be short-lived.
a
1a (20 mol%)
O
O
NO2
(23 W CFL)
hν
Me
H
+
3a
H
2,6-lutidine (1 equiv.)
Me Ph
Ph
6
°
NO2
MTBE, 25 C, 40 h
rac-5
93% yield, 86% e.e.
O
b
Ph
Complete γ-selectivity
1a (20 mol%)
H
NO2
We conducted further experiments to challenge the possibility of
an out-of-cage radical diffusion and subsequent chain propagation.
An alternative mechanism, detailed in Fig. 4b, can be envisaged
O
3a
3l
8a
Me
Ph
γ
Me
NO2
(23 W CFL)
hν
H
90% yield, 50% e.e.
to proceed via
a Kornblum–Russell SRN1-type alkylation
2,6-lutidine (1 equiv.)
MTBE, 25 °C, 15–48 h
pathway19,31–33, because both electron-poor benzyl31 and phenacyl32
bromides are suitable substrates. The process would proceed
through a traditional enamine catalysis pathway, with the
photoexcitation of the EDA complex (IV in Fig. 1) serving only
to initiate a radical-chain mechanism. The radical anion within V
would transpire out of the solvent cage to render, on facile
fragmentation, a carbon-centred radical species VIII. That free
radical could then add to the transiently generated enamine II to
give an a-amino radical IX, which could be oxidized by compounds
with low reduction potentials, such as the electron-poor alkyl
bromide III. ET to III would regenerate the radical anion VII,
and thus propagate the chain. Although in our proposed
mechanism (Fig. 1 and Fig. 4a) the enamine radical cation within
V is involved actively in the C–C bond-forming event, in the
alternative mechanistic hypothesis, depicted in Fig. 4b, it would
be an unproductive species.
O
7
Ph
H
Ph-3-Br
8l
Me
O
87% yield, 40% e.e.
c
O
O
1b (20 mol%)
(23 W CFL)
NO2
α
H
H
hν
+
3k
γ
Me
Me
2,6-lutidine (1 equiv.)
MTBE, 25 °C, 45 h
γ
'
O
9
10
70% yield
26% e.e.
10:1 γ vs γ' selectivity
Me
Me
Me
Me
Figure 3 | Evaluating the scope and the strategy’s potential to address
synthetically relevant problems. a, Forging an all-carbon quaternary
stereogenic centre. b, Remote stereocontrol and complete g-site selectivity.
c, Capacity to differentiate between three potential reactive centres.
We recognized that the implementation of a suitable radical
clock-containing aldehyde substrate should allow us to better deci-
pher the role played by the enamine radical cation within the ion
distant position. Accordingly, treatment of the a-branched enal 7
with the benzyl bromide 3a or phenacyl bromide 3l enabled
stereoselective alkylation exclusively at the aldehyde’s g-carbon
(Fig. 3b, 8a and 8l formed in high yield and 50% and 40% e.e.,
respectively). The capacity for controlling the site selectivity was
corroborated further by the experiments detailed in Fig. 3c, in
which three possible reactive centres were partitioned. Citral 9
was functionalized at the more substituted g-site position, with
the enantioenriched adduct 10 being the main product detected
(70% yield, 10:1 g-regioselectivity). Our rationale for the
observed site selectivity at the more distant position of polyun-
saturated aldehydes relies on the preferential formation of the
conjugated (rather than non-conjugated) radical iminium ion inter-
mediates of type VI.
a
Br
+
In cage
combination
O
N
EWG
hν
EDA complex
H
IV
–
H
Same pathway
as in Fig.1
R
V
R
EWG
b
Leaves
the cage
Br
+
N
EWG
–
Br–
H
EWG
R
VII
Mechanistic considerations. Our proposed mechanism (Fig. 1)
involves highly organized EDA complexes30 as critical
intermediates that play an explicit role in determining the
reactivity and the stereoselectivity. Control experiments performed
for all the reactions presented here confirmed that the absence of
the aminocatalyst 1 or of visible-light illumination suppressed the
process completely. These observations suggest that the formation
of a photon-absorbing enamine-derived EDA complex IV is
necessary to initiate the photochemical process. On photoinduced
ET, the intermediate V is formed. This contact radical ion pair is
stabilized by Coulombic attractive forces that preserve the original
well-defined mutual orientation. The alkyl radical generated on
R–Br dissociative cleavage then undergoes combination with its
chiral enamine radical cation partner in the solvent cavity of their
EWG
N
e– transfer
II
VIII
R
Br
III
N
EWG
Acceptor
EWG
IX
R
origin (intermediate VI). A variety of data are consonant with Figure 4 | Two possible reaction mechanisms. a, Our proposed mechanism
radical combination that occurs prior to diffusive separation of based on the in-cage radical combination driven by EDA formation (see also
the radical pair out of the solvent cage. For example, we did not Fig. 1). b, A plausible Kornblum–Russell alkylation pathway31,32 via a radical-
detect any by-products derived from free radicals, such as chain SRN1 mechanism.
NATURE CHEMISTRY | ADVANCE ONLINE PUBLICATION | www.nature.com/naturechemistry
© 2013 Macmillan Publishers Limited. All rights reserved.
5

NATURE CHEMISTRY DOI: 10.1038/NCHEM.1727
ARTICLES
4a was not detected after prolonged exposure to light (24 h).
However, the oxyamination adducts of the aldehyde (TEMPO
addition in the a position) and of the acceptor (trapping of the
benzylic radical) were formed in a 1:1 ratio. Additionally, substoichio-
metric amounts of TEMPO (10 mol% with respect to 3a) only
retarded the commencement of the alkylation process, which then
proceeded normally with a slight decrease in yield. A similar effect
was observed with the galvinoxyl radical. Although this outcome is
interesting per se, we do not wish to base the mechanistic proposal
on this result. There is a plausible explanation, based on the behaviour
of TEMPO as an electron donor in EDA complexes, that questions its
validity as a trapping experiment (see Supplementary Scheme S8 for
the alternative mechanism).
We then conducted the alkylation reaction in dimethyl sulfoxide
(DMSO) in the presence of a substoichiometric amount of FeBr2
(ref. 35), a one-electron transfer reagent (Fig. 5b). The reaction
was performed under the rigorous exclusion of light to suppress
completely a mechanism initiated by EDA complex excitation.
Under these conditions, the ET from FeBr2 to 3a, leading to the
radical anion VII depicted in Fig. 4b, is the only source of free
radical intermediates amenable for a trapping event by the elec-
tron-rich enamine. This experiment mimics the conditions for an
SRN1-type chain mechanism. Accordingly, these conditions pro-
vided the alkylation adduct 4a as a minor product (26% yield,
81% e.e.) together with large quantities of the by-products expected
from diffusion-controlled free radical reactions, such as alcohol 14,
bibenzyl 15 and toluene derivative 16 (Fig. 5b, pathway B). This
product distribution contrasts sharply with that obtained under illu-
mination and excluding the ET reagent, in which none of these by-
products were detected (Fig. 5b, pathway A). Interestingly, in the
absence of FeBr2 adduct 4a was formed quantitatively (98% yield)
in a lower optical purity (72% e.e.), which suggests that the stereo-
selectivity of this process is governed by a structurally different tran-
sition-state assembly.
a
O
1b (20 mol%)
O
NO2
(23 W CFL)
hν
H
3a
H
+
2,6-lutidine (1 equiv.)
MTBE, [3a]0 = 0.5 M, 25 o
NO2
C
Ph
Ph
trans-12
cis-11
SET
R'
1b
3
R'
R''
+
R''
H
+
N
R''
R'
+
N
N
H
H
Ph
13
Stereorandom
rearrangement
Ph
Ph
b
O
NO2
H
98% yield
72% e.e.
Visible light
Pathway A
Me
NO2
O2N
4a
Catalyst 1b
(20 mol%)
2a
+
HO
2,6-lutidine
3a
DMSO, 16 h
4a
Pathway B
Dark
26% yield
NO2
14
81% e.e.
+
40% yield
FeBr2 (50 mol%)
Me
NO2
NO2
O2N
+
NO2
NO2
16
15
O2N
Traces
12% yield
c
Last, as depicted in Fig. 5c, aldehyde alkylation with
phenacyl bromide 3g under irradiation by a monochromatic LED
1b (20 mol%)
O
O
O
(lmax ¼ 460 nm, irradiance 13.8 W m22
) proceeded smoothly,
LED* (λ = 460 nm)
H
H
+
which confirmed that residual near-ultraviolet light from the CFL
source was not responsible for the observed reactivity. The use of
low-energy monochromatic visible light excludes the possibility
for homolytic generation of a carbon-centred radical as an initiation
step because 3g is stable under these conditions.
Taken together, our investigations support a mechanism that
occurs predominantly through an enamine radical cation VI, the
formation of which is triggered by an ET event within an EDA
assembly and subsequent in-cage radical combination (Fig. 1).
Although a radical-chain process cannot be ruled out completely,
its contribution appears negligible. Continuing studies are aimed
2,6-lutidine (1 equiv.)
Br
O
4g
Me
Me
°
MTBE, 25 C, 40 h
2a
3g
72% yield, 84% e.e.
Figure 5 | Mechanistic investigations. a, Differentiating a radical cation-
based pathway and enamine addition. b,c, Challenging the possibility of an
RN1-type alkylation pathway (b) and of homolytic generation of carbon-
centred radicals (c). LED: l ¼ 460 nm, irradiance ¼ 13.8 W m22 using an
optical filter with a cutoff wavelength of 400 nm to prevent irradiation by
ultraviolet harmonic frequencies.
S
pair V. When the cis-cyclopropyl-substituted aldehyde 11 was to further delineate the reaction mechanism and explore potential
exposed to the alkylation protocol in the presence of bromide 3a, applications to other asymmetric catalytic photoreactions.
the trans-substituted cyclopropane product 12 was formed exclusively
Conclusions
(Fig. 5a). These observations provide evidence that the corresponding
cyclopropylcarbinyl radical cation 13, which undergoes ring opening/
closing prior to productive C–C bond formation, is an intermediate
along the main reaction pathway. Figure 5a shows the mechanism
of cyclopropane stereomutation enabled by the radical enamine 13,
which led to the thermodynamically more stable trans-cyclopropyl
product 1234. In contrast, such radical intermediates are absent from
the propagation steps of the enamine addition mechanism, as depicted
in Fig. 4b. Therefore, cis-substituted cyclopropane products should be
dominant if a radical-chain pathway prevails.
We found a bridge, the chiral EDA complex, to connect two power-
ful fields of molecule activation: asymmetric organocatalysis and
photochemistry. Specifically, we demonstrate that chiral enamines,
key intermediates in ground-state organocatalytic asymmetric pro-
cesses, have the potential to participate actively in the photoexcita-
tion of substrates, without the need for an external photosensitizer.
The strategy differs from, but complements, the approach of photo-
redox catalysis36, a rapidly developing area of modern chemical
research. These findings are expected to open new avenues for reac-
tion design in the field of asymmetric photochemical processes.
Also, we performed trapping experiments in the presence of
radical scavengers (details are given in Supplementary
Scheme S7). When 2,2,6,6-tetramethylpiperidine 1-oxyl (TEMPO,
1 equiv.) was added to the reaction mixture of butanal 2a,
Methods
All reactions were performed under an argon atmosphere in oven-dried glassware
using standard Schlenk techniques, unless otherwise noted. Synthesis-grade
bromide 3a and catalyst 1b, the corresponding alkylation product solvents were used as purchased and the reaction mixtures were deoxygenated by
6
NATURE CHEMISTRY | ADVANCE ONLINE PUBLICATION | www.nature.com/naturechemistry
© 2013 Macmillan Publishers Limited. All rights reserved.

NATURE CHEMISTRY DOI: 10.1038/NCHEM.1727
ARTICLES
three cycles of freeze–pump–thaw prior to illumination. Full experimental
details and characterization data for all new compounds are included in the
Supplementary Information.
23. Hilinski, E. F., Masnovi, J. M., Amatore, C., Kochi, J. K. & Rentzepis, P. M.
Charge-transfer excitation of electron donor–acceptor complexes. Direct
observation of ion pairs by time-resolved (picosecond) spectroscopy. J. Am.
Chem. Soc. 105, 6167–6168 (1983).
Received 6 May 2013; accepted 3 July 2013;
published online 11 August 2013
24. Rathore, R. & Kochi, J. K. Donor/acceptor organizations and the electron-
transfer paradigm for organic reactivity. Adv. Phys. Org. Chem. 35, 193–318 (2000).
25. Gotoh, T., Padias, A. B. & Hall, J. H. K. An electron donor–acceptor complex
and thermal triplex as intermediates in the cycloaddition reaction of
N-vinylcarbazole with dimethyl 2,2-dicyanoethylene-1,1-dicarboxylate.
J. Am. Chem. Soc. 113, 1308–1312 (1991).
26. Berionni, G., Bertelle, P-A., Marrot, J. & Goumont, R. X-ray structure of a CT
complex relevant to Diels–Alder reactivity of anthracenes. J. Am. Chem. Soc.
131, 18224–18225 (2009).
References
1. Balzani, V., Credi, A. & Venturi, M. Photochemical conversion of solar energy.
ChemSusChem, 1, 26–58 (2008).
2. Ciamician, G. The photochemistry of the future. Science 36, 385–394 (1912).
3. Albini, A. & Fagnoni, M. Handbook of Synthetic Photochemistry
(Wiley-VCH, 2010).
27. Melchiorre, P. Light in aminocatalysis: the asymmetric intermolecular
a-alkylation of aldehydes. Angew. Chem. Int. Ed. 48, 1360–1363 (2009).
28. Jensen, K. L., Dickmeiss, G., Jiang, H., Albrecht, Ł. & Jørgensen, K. A. The
diarylprolinol silyl ether system: a general organocatalyst. Acc. Chem. Res. 45,
248–264 (2012).
4. Bach, T. & Hehn, J. P. Photochemical reactions as key steps in natural product
synthesis. Angew. Chem. Int. Ed. 50, 1000–1045 (2011).
5. Inoue, Y. & Ramamurthy, V. Chiral Photochemistry (Marcel Dekker, 2004).
6. Mu¨ller, C. & Bach, T. Chirality control in photochemical reactions:
enantioselective formation of complex photoproducts in solution. Aust. J.
Chem. 61, 557–564 (2008).
7. Ojima, I. Catalytic Asymmetric Synthesis (Wiley, 2010).
8. Guo, H., Herdtweck, E. & Bach, T. Enantioselective Lewis acid catalysis in
intramolecular [2þ2] photocycloaddition reactions of coumarins. Angew. Chem.
Int. Ed. 49, 7782–7785 (2010).
9. Bauer, A., Westkamper, F., Grimme, S. & Bach, T. Catalytic enantioselective
reactions driven by photoinduced electron transfer. Nature 436,
1139–1140 (2005).
10. Mu¨ller, C., Bauer, A. & Bach, T. Light-driven enantioselective organocatalysis.
Angew. Chem. Int. Ed. 48, 6640–6642 (2009).
11. Knowles, R. R. & Jacobsen, E. N. Attractive noncovalent interactions in
asymmetric catalysis: links between enzymes and small molecules catalysts.
Proc. Natl Acad. Sci. U.S.A. 107, 20678–20685 (2010).
12. Nicewicz, D. A. & MacMillan, D. W. C. Merging photoredox catalysis with
organocatalysis: the direct asymmetric alkylation of aldehydes. Science 322,
77–80 (2008).
13. Shih, H-W., Vander Wal, M. N., Grange, R. L. & MacMillan, D. W. C.
Enantioselective a-benzylation of aldehydes via photoredox organocatalysis.
J. Am. Chem. Soc. 132, 13600–13603 (2010).
14. Mulliken, R. S. Molecular compounds and their spectra. II. J. Am. Chem. Soc. 74,
811–824 (1952).
15. Foster, R. Electron donor–acceptor complexes. J. Phys. Chem. 84,
2135–2141 (1980).
16. MacMillan, D. W. C. The advent and development of organocatalysis. Nature
455, 304–308 (2008).
17. Mukherjee, S., Yang, J. W., Hoffmann, S. & List, B. Asymmetric enamine
catalysis. Chem. Rev. 107, 5471–5569 (2007).
18. Cantacuzene, D., Wakselman, C. & Dorme, R. Condensation of perfluoroalkyl
iodides with unsaturated nitrogen compounds. J. Chem. Soc. Perkin Trans. 1
1365–1371 (1977).
19. Russell, G. A. & Wang, K. Electron transfer processes. 53. Homolytic alkylation
of enamines by electrophilic radicals. J. Org. Chem. 56, 3475–3479 (1991).
20. Singh, J. O., Anunziata, J. D. & Silber, J. J. n–p Electron donor–acceptor
complexes. II. Aliphatic amines with dinitrobenzenes. Can. J. Chem. 63,
903–907 (1985).
21. Mu¨ller, K., Previdoli, F. & Desilvestro, H. Enamines, II. A theoretical and
photoelectron spectroscopic study of the molecular and electronic structures of
aliphatic enamines. Helv. Chim. Acta 64, 2497–2507 (1981).
22. Rosokha, S. V. & Kochi, J. K. Fresh look at electron-transfer mechanisms via the
´
29. Costentin, C., Robert, M. & Saveant, J-M. Electron transfer and bond breaking:
recent advances. Chem. Phys. 324, 40–56 (2006).
30. Mannschreck, A., Roza, P., Brockmann, H. & Kemmer, T. Demonstration of
diastereomeric electron-donor–acceptor complexes by 1H-NMR spectroscopy.
Angew. Chem. Int. Ed. Engl. 17, 940–942 (1978).
31. Kornblum, N. Substitution reactions which proceed via radical anion
intermediate. Angew. Chem. Int. Ed. Engl. 14, 734–745 (1975).
32. Russell, G. A., & Ros, F. Reactions of a-halo ketones with nucleophiles J. Am.
Chem. Soc. 107, 2506–2511 (1985).
´
33. Rossi, R. A., Pierini, A. B. & Pen˜en˜ory, A. B. Nucleophilic substitution reactions
by electron transfer. Chem. Rev. 103, 71–168 (2003).
34. Van Humbeck, J. F., Simonovich, S. P., Knowles, R. R. & MacMillan, D. W. C.
Concerning the mechanism of the FeCl3-catalyzed a-oxyamination of
aldehydes: evidence for a non-SOMO activation pathway. J. Am. Chem. Soc. 132,
10012–10014 (2010).
35. Nazareno, M. A. & Rossi, R. A. SRN1 reactions of 7-iodobicyclo[4.1.0]heptane,
1-iodoadamantane, and neopentyl iodide with carbanions induced by FeBr2 in
DMSO. J. Org. Chem. 61, 1645–1649 (1996).
36. Narayanam, J. M. R. & Stephenson, C. R. J. Visible light photoredox catalysis:
applications in organic synthesis. Chem. Soc. Rev. 40, 102–113 (2011).
Acknowledgements
This work was supported by the Institute of Chemical Research of Catalonia Foundation
and by the European Research Council under the European Community’s Seventh
Framework Program (FP7 2007–2013)/ERC Grant agreement 278541. This work is
dedicated to P. G. Cozzi on the occasion of his 50th birthday.
Author contributions
E.A. and I.D.J. were involved in the discovery and subsequent development of the light-
´
driven alkylation reactions. E.A., I.D.J. and A.A-F. performed the experiments. E.A., I.D.J.,
´
A.A-F. and P.M. designed and analysed the experiments. P.M. conceived and directed the
project and wrote the manuscript with contributions from E.A.
Additional information
Supplementary information and chemical compound information are available in the
online version of the paper. Reprints and permissions information is available online at
www.nature.com/reprints. Correspondence and requests for materials should be
addressed to P.M.
donor/acceptor bindings in the critical encounter complex. Acc. Chem. Res. 41, Competing financial interests
641–653 (2008).
The authors declare no competing financial interests.
NATURE CHEMISTRY | ADVANCE ONLINE PUBLICATION | www.nature.com/naturechemistry
7
© 2013 Macmillan Publishers Limited. All rights reserved.