Journal of the American Chemical Society
Communication
complexes as a suitable class of PCs because of two
considerations: (1) the requisite chiral information can be
introduced into the structure of the pairing anion, which would
not affect the photocatalytic activity of the PC;6 (2) the
installation of DGs with an anion-recognition ability onto the
substrate would ensure its association with the PC. Assuming a
case with an electron-rich substrate embedded with such DG
functionality, we reasoned that capturing the chiral anion
component of the PC by the DG moiety would form a chiral
supramolecular ion pair. Upon light irradiation, single-electron
transfer (SET) from the substrate to the accompanying
excited-state cationic Ir complex could generate the corre-
sponding radical-ion pair with a defined three-dimensional
arrangement. This would enable the resultant radical cation to
undergo ensuing stereoselective bond formation within the
asymmetric environment created by the chiral anion. Herein,
we report the successful operation of this system using urea as
an anion-recognizable, redox active DG, and a cationic Ir
complex-chiral borate ion pair as a PC in achieving an efficient
and highly diastereo- and enantioselective [3 + 2]-cyclo-
addition of cyclopropylamine with alkenes under visible-light
irradiation.
To substantiate our hypothesis, we chose the [3 + 2]-
cycloaddition of cyclopropylamines with alkenes as a model
reaction,7−10 which is initiated by single-electron oxidation of
the amine reactant, primarily because N-carbamoylation of the
amine readily sets the urea functionality that is expected to act
as an oxidizable DG with a distinct anion-binding capability.
Importantly, the affinity of the urea moiety toward an anion
would be increased in the corresponding radical cation
generated via single-electron oxidation by an appropriate PC,
such as an Ir-polypyridyl complex. As an anionic component of
the PC, we employed weakly coordinating chiral borate 111,12
(Table 1 scheme) in anticipation that its negligible
nucleophilicity would be beneficial for the pairing Ir complex
to exert full potential as a PC,13 while its characteristic as a
hydrogen-bond acceptor would allow interaction with the urea
moiety. A preparatory step of the presumed reaction pathway
is the assembly of the photoactive Ir-chiral borate ion pair,
[PC][1], and N-cyclopropylurea derivative 2 into the supra-
molecular ion pair [PC] [1⊂2] (Figure 2). The excitation of
the Ir complex [PC]+ with irradiation of visible light followed
by SET leads to the generation of the radical-ion pair [1⊂2]•
with concomitant release of the reduced, noncharged Ir
complex [PC]. In this assembly, the radical cation derived
from 2 is considered to exist as a formal equilibrium mixture of
N-radical cation and distonic radical cation,14 and the latter
would react with alkene 3. The intermediary alkyl radical
would undergo stereoselective five-exo cyclization under the
guidance of accompanying 1 to afford aminocyclopentane 4 as
a form of N-radical cation pairing with 1. After single-electron
reduction of the N-radical cation by the reduced Ir complex
[PC], the desired product 4 is liberated via the transfer of
chiral anion 1 to 2 to complete the catalytic cycle. An
alternative chain process could also be operative if the N-
radical cation of 4 acts as an oxidant of 2 to directly generate
the key radical-ion pair [1⊂2]• via simultaneous transfer of the
hole and 1, although the same transition-state structure would
be involved in the stereodetermining step.
Figure 2. Proposed catalytic cycle. PC = photocatalyst (chromo-
phore), B* = chiral borate ion 1.
[rac-Ir(dFCF3ppy)2(dtbbpy)][1a] (*Ered = 1.42 V vs SCE)15
(5 mol %) in dichloromethane under irradiation with blue
LEDs (470 nm) at −30 °C. After stirring for 24 h, the desired
[3 + 2]-cycloadduct 4a was isolated in 40% yield as a mixture
of diastereomers (1.6:1), and a certain enantiomeric excess
(ee) was detected (18% ee for the major diastereomer) (Table
1, entry 1). Interestingly, the parallel reaction with [rac-
Ir(dFCF3ppy)2(dtbbpy)][PF6] as the PC under otherwise
similar conditions for obtaining racemic 4a was very sluggish
(<10% yield), suggesting that the property of the anionic
component of the PC is tightly associated with the reaction
efficiency. This observation prompted us to evaluate the
relationship between the anion structure and the catalytic
activity of the iridium-based PC by examining the reaction with
a series of iridium complexes possessing various anions. This
revealed that the reactivity trend was indeed dependent on the
anion characteristics (Table S1). Considering the anion effect
reported for ruthenium-based photocatalysis,13 we assumed
that the observed reactivity difference could be ascribed to the
stabilization of the excited-state iridium complex by the
counterion and thus we measured the fluorescence spectrum
of each iridium complex. However, all the complexes exhibited
essentially identical spectra (Figure S1), unlike the case with
ruthenium complexes. Therefore, we considered the influence
of the ability of the anion to coordinate with the urea on
reactivity and determined the association constants (Ka) of the
anions with 2 by 1H NMR titration experiments (Figure S2).16
By plotting the Ka values against cycloaddition conversion, an
evident correlation was revealed (Figure 3). While the reaction
with the iridium complex bearing a noncoordinating anion
afforded very low conversion, the conversion gradually
increased as the Ka value increased, reaching a maximum
when Ka was approximately 60 M−1, and then declined with
further increases in Ka. Although elucidation of the origin of
this profile should await detailed mechanistic studies, one
possibility is that the equilibrium between the radical cation of
2 and the corresponding distonic radical cation would favor the
The validity of this mechanistic blueprint was initially
assessed by attempting the cycloaddition of 3,5-xylyl cyclo-
propylurea (2) (Eox = 1.37 V vs saturated calomel electrode
(SCE)) and α-methylstyrene (3a) with a catalytic amount of
B
J. Am. Chem. Soc. XXXX, XXX, XXX−XXX