Article

pubs.acs.org/accounts

The Prowess of Photogenerated Amine Radical Cations in Cascade

Reactions: From Carbocycles to Heterocycles
Published as part of the Accounts of Chemical Research special issue “Photoredox Catalysis in Organic
Chemistry”.
Scott A. Morris, Jiang Wang, and Nan Zheng*
Department of Chemistry and Biochemistry, University of Arkansas, Fayetteville, Arkansas 72701, United States

CONSPECTUS: Cascade reactions represent a class of ideal organic reactions because they
empower efficiency, elegance, and novelty. However, development of cascade reactions
remains a daunting task for synthetic chemists. Radicals are known to be well suited for
cascade reactions. Compared with widely used carbon-based radicals, nitrogen-based radicals,
such as neutral aminyl radicals and protonated aminyl radicals (amine radical cations), are
underutilized, although they are behind some notable synthetic methods such as the
Hofmann−Löffler−Freytag reaction. The constraint on their usage is generally attributed to
the limited number of available stable precursors. Since amine radical cations offer increased
reactivity and selectivity in chemical transformations compared with neutral aminyl radicals,
their generation is of utmost importance. Recently, a surge of reports has been revealed using
visible light photoredox catalysis. It has been demonstrated that amines can act as an electron
donor in a reductive quenching cycle while the amine itself is oxidized to the amine radical
cation. Although a number of methods exist to generate amine radical cations, the
photochemical formation of these species offers many practical advantages.
In this Account, we discuss our journey to the development of annulation reactions with various π-bonds and electrophilic
addition reactions to alkenes using photogenerated amine radical cations. Various carbocycles and heterocycles are produced by
these reactions. In our annulation work, we first show that single electron photooxidation of cyclopropylanilines to the amine
radical cations triggers ring opening of the strained carbocycle, producing distonic radical cations. These odd-electron species are
shown to react with alkenes and alkynes to yield the corresponding cyclopentanes and cyclopentenes in an overall redox neutral
process. Further development of this annulation reaction allows us to achieve the [4 + 2] annulation of cyclobutylanilines with
alkynes. In our work on electrophilic addition reactions to alkenes, we reveal that photogenerated amine radical cations are
capable of undergoing the electrophilic addition reactions to alkenes to form a variety of indoles and indolines. This chemistry
represents a rare oxidative C−N bond-forming reaction using visible light. Conclusions drawn from observational results and
proposed mechanisms are outlined in this Account. Additionally, open discussion of our successes and deficiencies in our
experiences will give readers helpful insights as to how these species tend to react.
The overall utility of photogenerated amine radical cations has yet to reach its full potential. With our current results, we
anticipate more new transformations can still be derived from the ring opening processes of cyclopropylanilines and
cyclobutylanilines under visible light photocatalysis. Additionally, since utilizing photogenerated amine radical cations in C−N
bond-forming reactions has practically been absent in literature, we are confident more new reactions have yet been exploited.

■ INTRODUCTION
Compared with carbon radicals, which are considered to be
mainstream in synthetic chemists’ toolbox, nitrogen-centered
radicals are rather underutilized. The limited number of
available precursors often constrains their wide use in organic
synthesis.1 Synthetically, amine radical cations (protonated
aminyl radicals) are typically preferred over neutral aminyl
radicals in large part due to their increased electrophilicity, their
ability to add to alkenes, and their propensity to undergo
synthetically useful transformations, such as the formation of
cyclic amines as seen in the Hofmann−Lö ffler−Freytag
reaction.2 In addition, amine radical cations show heightened
reactivity and selectivity over neutral aminyl radicals, adding
further credence to their importance in chemical synthesis.3
The chemistry of amine radical cations is usually determined by
the functionality around its periphery. For instance, although
deprotonation is a major pathway, amine radical cations have
been shown to undergo irreversible C−C bond cleavage,
electrophilic addition, and hydrogen atom abstraction, which
are all determined by the nitrogen atom’s substituents.4
Our group has recently been intrigued by the synthetic
potential of photogenerated amine radical cations, particularly
their use in cascade reactions. Chemists have generated amine
radical cations using various approaches, including the use of
UV light with a photosensitizer,5 strong stoichiometric
Received: May 29, 2016

© XXXX American Chemical Society A DOI: 10.1021/acs.accounts.6b00263

Acc. Chem. Res. XXXX, XXX, XXX−XXX
Accounts of Chemical Research
oxidants,6 enzyme oxidation,7 anodic oxidation,8,9 and visible
light photochemistry.10 We specifically target the formation of
amine radical cations by direct one-electron oxidation of the
corresponding amines using visible light photoredox catalysis.
This approach overcomes some commonly observed short-
comings in other methods, such as instability of precursors, use
of toxic reagents (e.g., Bu3SnH), and harsh conditions (e.g.,
strongly acidic conditions) and the need for special apparatus
(e.g., quartz).
Visible light photoredox catalysis has been a subject of great
importance recently, largely due to its ability to generate odd-
electron species such as carbon- and nitrogen-centered radicals
from stable organic substrates with full valence electrons.11
Amines are typically used as stoichiometric sacrificial electron
donors to generate the highly reducing species that initializes
radical reactions with organic substrates, and amine radical
cations are concomitantly formed. This process is referred to as
a reductive quenching cycle, one of the three main cycles for
quenching the photoexcited state (Scheme 1).4,12 However,
Scheme 1. Photogenerated Amine Radical Cations Using a
Reductive Quenching Cycle
chemists have only recently exploited amines’ dual role as the
sacrificial electron donor and a substrate, which allows the
amine to not only reduce the photocatalyst but also serve as a
reactive substrate for useful transformations. In our attempts to
exploit amines’ dual role, we were able to form a variety of
carbocycles and heterocycles using cascade processes initialized
by amine radical cations. Our first report, focusing on
annulation reactions of cyclopropyl- and cyclobutyl-anilines
with alkenes and alkynes, reveals a redox-neutral pathway
centered on a ring-opening process to form a reactive distonic
ion. Our second usage afforded indoles and indolines by
utilizing photogenerated amine radical cations as electrophilic
addition partners to alkenes. This work was carried out using
benign aerobic reaction conditions and stands in contrast to
similar indole syntheses.
■ C−C BOND CLEAVAGE AS A POWERFUL TOOL
FOR THE CONSTRUCTION OF CARBOCYCLES
Our group’s initial work in photoredox-catalyzed C−C and C−
N bond formation tandems began with [3 + 2] annulation
reactions of cyclopropylanilines with olefins. The discovery of
this chemistry began with a serendipitous result stemming from
our previous work in photoinduced N−N bond cleavage of
aromatic hydrazines.13
In what appeared to be a trivial transformation, photo-
cleavage of the aromatic hydrazine bearing a cyclopropyl group
B DOI: 10.1021/acs.accounts.6b00263
Article
yielded an unexpected adduct which, upon further analysis, was
determined to be the endoperoxide adduct derived from a
[3 + 2] annulation with oxygen (Scheme 2A). We speculated
Scheme 2. Initial Discovery of the [3 + 2] Annulation Using
Cyclopropylanilines
that, as anticipated, the N−N bond cleavage would occur first
to produce cyclopropylaniline, which could then undergo the
[3 + 2] annulation with oxygen. To verify this hypothesis, we
prepared cyclopropylaniline and subjected it to the identical
reactions conditions, which yielded the same endoperoxide
(Scheme 2B), giving further credence to the proposed [3 + 2]
annulation pathway.
In order for the annulation to occur, the highly strained
cyclopropane ring would need to open first, react with the π
bond of the reaction partner, and then close to complete the
transformation. The key to this annulation reaction centers on
the well-established ring opening reaction of the radical cation
of cyclopropylanilines, which would give rise to the
corresponding distonic radical cation capable of participating
in the annulation in a stepwise fashion (Scheme 3).
Scheme 3. Cyclopropylaniline Ring Opening to Generate a
Reactive Distonic Radical Cation
Since our group’s goal was to exploit the reactivity of
photogenerated amine radical cations, this system seemed
appropriate for further study and examination. Furthermore,
although distonic radical cations are well-established species in
the gas phase, their reactivity in solution phase is not well
understood. From the limited published data, they exhibit
unique reactivity distinct from conventional radical cations and
related even-electron ions.14 We rationalized that this distonic
radical cation possesses a dual functionality in which the
nucleophilic radical has the capacity to act as a nucleophile with
acceptors, such as alkenes in a Giese reaction,15 while the
electrophilic iminium ion could engage in a subsequent ring
closure as an electrophile with the resulting radical from the
Giese reaction to complete the transformation.16
■
[3 + 2] ANNULATION USING ALKENES AND
CYCLOPROPYLANILINES
We immediately decided to test this reaction system using
styrene in the presence of oxygen to generate the
corresponding carbocycle.17 Our initial results, obtained by
performing the reaction with 2 mol% Ru(bpz)3(PF6)2 in
Acc. Chem. Res. XXXX, XXX, XXX−XXX
Accounts of Chemical Research
nitromethane, gave modest amounts of annulation product 2
(21%) although 100% conversion of 1 was observed
(Table 1, entry 1). Presumably, the annulation of the distonic
Table 1. Reaction Optimization for the [3 + 2] Annulation
entry conditions catalyst time (h) yield of 2 (%)
1 open to air Ru(bpz)3(PF6)2 12 21
2 degassed Ru(bpz)3(PF6)2 3 96
3 degassed Ru(bpy)3(PF6)2 12 79
4 degassed Ir(ppy)2(dtbbpy)PF6 12 73
radical cation with oxygen was competitive against with styrene,
resulting in low yield of 2. Further optimization revealed that
the desired [3 + 2] annulation required degassed conditions,
which yielded 2 as a single regioisomer in 96% yield, albeit as a
1:1 mixture of diastereomers (entry 2). We also noted that
under the optimized conditions, other photoredox catalysts
failed to match the productivity compared with Ru(bpz)32+
(entries 3 and 4), which bolsters an impressive 1.31 V
reduction potential for its excited state (Ru(bpz)32+*/Ru-
(bpz)3+ versus Ag/AgCl).
We subsequently developed a substrate scope that took into
consideration both olefin and N-cyclopropylaniline alterations,
which is summarized in Scheme 4.
As anticipated, we were able to generate a variety of
annulation products in moderate to high yields (40−87%).
Although the highest reaction efficiency was observed using
styrene as the olefin component, we were delighted to find that
acrylonitrile and 1-phenyl-1,3-butadiene participated in the
annulation as well. We discovered that internal alkenes failed to
react under the prescribed conditions, presumably due to steric
hindrance, so it was not surprising that the conjugated diene’s
terminal double bond was more reactive in the [3 + 2]
annulation.
Our next goal was to test the viability of using bicyclic
cyclopropanes, readily synthesized via Kulinkovich−de Meijere
reaction of amide precursors,9,18 to give the corresponding 5,5-
and 6,5-fused carbocycles. Furthermore, we rationalized that
conformational bias imposed by the constraint of bicyclic
cyclopropanes could improve the poor diastereoselectivity
Scheme 4. Summarized Substrate Scope Using Monocyclic Cyclopropylanilines
C DOI: 10.1021/acs.accounts.6b00263
Article
observed in the monocyclic case. When we tested our
hypothesis using styrene and an angular-methyl-bearing bicycle,
we were pleased to isolate the annulation product 3 as a 4:1
mixture of diastereomers in 77% combined yield, which
presumably proceeds through the distonic radical cation int 1
(Scheme 5). This result stands in direct contrast to the
monocyclic case in which tertiary amines failed to react. We
reasoned that this phenomenon was likely due to the higher
ring strain built in fused bicyclic cyclopropanes.
Carbocycles 4 and 5 were acquired using the optimized
conditions, both of which display variability on the bicycle,
albeit low diastereoselectivity. However, in order to further
enhance the diastereoselectivity, an angular tert-butyl-bearing
bicyclic cyclopropylaniline was subjected to the annulation
conditions with the para-methoxy analogue of styrene, which
yielded fused 5,5-carbocycle 6 as a single isomer in 30% yield.
We attributed the low yield to a sluggish ring opening process
or slow addition to the iminium ion to close the ring, since we
recovered roughly 50% of the starting material.
To justify the observed diastereoselectivity in the [3 + 2]
annulation using bicyclic cyclopropylanilines and alkenes, we
applied the Beckwith−Houk model, which is generally used to
predict the diastereoselectivity of the cyclization of hexenyl
radicals (Figure 1).19 We determined that if the hydrogen on
the radical carbon is axial to the R1 on the iminium, then the
1,3-diaxial interaction between R1 and R2 is obviated.
■ REACTION EXPANSION: UTILIZING TERMINAL
ALKYNES, ENYNES, AND DIYNES
Following this report, our group revealed further improvements
to the substrate scope, documented in subsequent accounts.20
In these instances, we aimed to eliminate diastereoselectivity by
using alkynes as the annulation partner to yield the
corresponding cyclopentenes. We were also intrigued by the
possibility to introduce an alkene functionality into the
carbocycle, thus allowing for further elaboration into other
useful products. Since alkynes innately have lower reactivity
toward the radical addition than the alkene counterparts, a
stronger light was shown to be necessary for an efficient
annulation.21 As expected, however, the optimized conditions
were identical to those described in our first annulation report.
Initially, we found that terminal alkynes bearing radical-
stabilizing groups, such as the phenyl group, and electron-
withdrawing groups were effective partners in these trans-
formations. However, we were particularly impressed when a
variety of diynes and enynes reacted well with cyclopropylani-
Acc. Chem. Res. XXXX, XXX, XXX−XXX
Accounts of Chemical Research
Scheme 5. Summarized Substrate Scope Using Bicyclic Cyclopropylanilines
Figure 1. Diastereoselectivity model for bicyclic cyclopropylamines.
lines to give added functionality to the annulation product
(Scheme 6). The diynes, whether symmetrical or asymmetrical,
exhibited excellent regioselectivity and provided the highest
yields when bearing electron-withdrawing groups. Enyne
Scheme 6. Expansion of the [3 + 2] Annulation Using Various Alkynes
D DOI: 10.1021/acs.accounts.6b00263
Article
reactivity was more complicated, because the olefin’s steric
bulk influenced which π system (alkene or alkyne) participated
in the annulation. We found that when dealing with sterically
unhindered alkenes, the alkene moiety was the major reacting
site whereas the alkyne moiety was favored in the cases
involving sterically hindered alkenes. In certain instances when
distinguishing between the alkene and alkyne by steric bulk or
electronic factors became less clear, poor selectivity was often
observed, resulting in a complicated mixture of both predictable
and unidentified annulation products. We also noted that, as
expected from literature precedence22 and our previous
annulation report using alkenes,17 tertiary amines failed to
react in this transformation, likely due to a slower ring opening
process.
Acc. Chem. Res. XXXX, XXX, XXX−XXX
Accounts of Chemical Research
Scheme 7. Rationale for Observed Regioselectivity Using Diynes
Scheme 8. Utilizing Substituted Cyclopropanes in the [3 + 2] Annulation
Scheme 9. Oxidative Cleavage of Requisite Aryl Group
Since regioselectivity became a focused point of discussion
with diynes and enynes, we decided to probe further into this
phenomenon. Using diynes as a representative, we performed
DFT calculations on model systems such as vinyl radicals 9 and
10 that reflected the proposed intermediates 7 and 8,
respectively (Scheme 7). As expected, resonance-stabilized
vinyl radical 9, which led to the major regioisomer, was found
to be 10.5 kcal/mol more stable than vinyl radical 10, which led
to the minor regioisomer.20b
We then shifted our focus to studying substituted cyclo-
propylanilines, focusing specifically on the cyclopropane ring.
We realized that decorating the cyclopropane could result in a
larger substrate scope and further improve the utility of this
chemistry, since, in our previous reports, we had only examined
modified cyclopropanes in the bicyclic examples with alkenes.17
E DOI: 10.1021/acs.accounts.6b00263
Article
We started our study by subjecting α-methylcyclopropylaniline
to the optimized conditions with phenylacetylene. After 15 h,
we were delighted to observe the α-substituted annulation
product 11, albeit in only 33% yield (Scheme 8). We next
turned our attention to β-substituted cyclopropylanilines. Both
methyl- and phenyl-substitutions were amenable to the
annulation, yielding carbocycles 12 and 13 in 46% and 30%
yields, respectively. Additionally, we note that the observed
regioselectivity of the final product is derived from
intermediates similar to int 2, which possess a stabilized carbon
radical.
The N-aryl substituent quickly emerged to be a critical
constituent for these transformations as they were shown to
lower the redox potential of amines.23 Cyclic voltammetry
experiments on both N-cyclopropylaniline and an N-
Acc. Chem. Res. XXXX, XXX, XXX−XXX
Accounts of Chemical Research
alkylcyclopropylamine revealed roughly a 0.6 V lower oxidation
peak potential when the nitrogen atom bears a phenyl rather
than an alkyl group.17 We realized that this posed a limitation
to our chemistry; therefore, we set our final focus on the
removal of the requisite aryl group. Oxidative cleavage has
shown to be effective in the removal of electron-rich N-aryl
groups, typically being mediated by strong oxidants like ceric
ammonium nitrate (CAN), periodic acid, or trichloroisocyanu-
ric acid (TCCA).24 We synthesized a variety of cyclopropyl-
amines bearing a removable aryl group in hopes they would
participate in the annulation reaction. Among the groups
screened, the para-methoxy phenyl (PMP) group was found to
be the best, being capable of undergoing the annulation
reaction while also being removable. Indeed, we were excited to
see that the annulation reaction proceeded in 47% yield to give
the corresponding PMP-bearing cyclopentene 14, which after
much screening was cleaved using CAN/H2SO4 and then
acylated to give amide 15 in 68% yield over 2 steps (Scheme 9).
In accordance with the data presented in these reports, an
overall redox-neutral pathway was proposed (Scheme 10).
Scheme 10. Proposed Mechanism for the [3 + 2] Annulation
Upon irradiation with visible light, the photocatalyst gets
excited to its triplet state, which is capable of oxidizing the
cyclopropylaniline to the corresponding amine radical cation.
Subsequent C−C bond cleavage results in the distonic radical
cation, which adds via a radical pathway to the π bond, whether
alkene or alkyne. The resulting alkyl or vinyl radical then
cyclizes onto the iminium ion to yield another amine radical
cation that can readily be reduced by the Ru+ complex, yielding
both the annulation product and starting photocatalyst.
■
[4 + 2] ANNULATION: OPENING CYCLOBUTANE
USING VISIBLE LIGHT
Encouraged by the successful results of the [3 + 2] annulation
of cyclopropylanilines with various π bonds, we questioned
whether cyclobutylanilines could participate in the annulation
reaction in an analogous manner to produce the [4 + 2]
annulation products. Similar ring strain between cyclobutane
and cyclopropane (27.5 vs 26.7 kcal/mol)25 as well as oxidation
peak potential between cyclopropyl- and cyclobutyl-aniline lent
credence to this idea, as shown in Scheme 11A. However, ring
opening of cyclobutylcarbinyl radicals was reported to be
significantly slower than that of cyclopropylcarbinyl radicals.26
The same phenomenon was observed with respect to
cyclobutyl-n-propylaminyl and cyclopropyl-n-propylaminyl rad-
icals, whereas the rate constant of the latter was determined to
be more than 100 times larger than the former (Scheme
F DOI: 10.1021/acs.accounts.6b00263
Article
Scheme 11. Key Physical Data of Cyclobutylaniline and
Cyclopropylaniline
11B).27 On the other hand, it was reported that fragmentation
of amine radical cations was significantly faster than their
neutral aminyl radical counterpart.28 Since amine radical cations
were the proposed intermediate to undergo the ring opening,
we felt that the ring opening’s rate might be sufficient for the
annulation to occur.
With these in mind, we used 4-tert-butyl-N-cyclobutylaniline
16 and phenylacetylene as the model substrates to examine the
proposed [4 + 2] annulation (Table 2).29 The catalyst system
Table 2. Reaction Optimization for the [4 + 2] Annulation
time GC yield of 17
entry catalyst conditions (h) (%)
1 Ru(bpz)3(PF6)2 CH3NO2, 2 16 30
LED
2 Ru(bpz)3(PF6)2 MeOH, 2 LED 16 37
3 Ir(ppy)2(dtbbpy)PF6 MeOH, 2 LED 12 97
4 Ir(ppy)2(dtbbpy)PF6 MeOH, 1 LED 16 27
optimized for the [3 + 2] annulation, 2 mol% Ru(bpz)3(PF6)2
in nitromethane under nitrogen, was initially examined. The
desired [4 + 2] annulation product 17 was obtained in 30%
yield after irradiatioin for 16 h by two 18 W LED light bulbs
(entry 1). Encouraged by this success, we underwent systematic
screening of the catalyst system. Methanol was found to be a
better solvent than nitromethane (entry 2). Switching the
catalyst to Ir(ppy)2(dtbbpy)PF6 made the largest impact on the
annulation, affording product 17 in 97% yield (entry 3). The
use of two lights was critical because they provided more
photon flux and higher temperature. By comparison, irradiation
with one LED light furnished product 17 in 27% yield (entry
4). The need for stronger light in the [4 + 2] annulation was
consistent with the premise that the amine radical cation of
cyclobutylanilines undergoes the ring opening more slowly than
that of cyclopropylanilines (vide supra).
We next established the substrate scope by varying
substituents on the N-aryl group and alkyne. Substituents
with various electronic and steric characters were generally
tolerated, and heterocycles showed no ill effects toward the
annulation (Scheme 12). With respect to alkynes, we also
observed a notable difference in the substrate scope between
the [3 + 2] and [4 + 2] annulation. Unlike the [3 + 2]
annulation, the [4 + 2] annulation was not limited to the
annulation with terminal alkynes, and internal alkynes also
Acc. Chem. Res. XXXX, XXX, XXX−XXX
Accounts of Chemical Research
Scheme 12. [4 + 2] Annulation Scope of Monocyclic
Cyclobutylanilines
reacted with predictable regioselectivity based on the
substituent’s ability to stabilize the vinyl radical. We attributed
this difference to cyclobutylanilines’ increased stability toward
light and the primary carbon radical of the distonic radical
cation derived from cyclobutylanilines being more nucleo-
philic.30
We sought to further expand the substrate scope by studying
cis fused 5,4-membered and 6,4-membered bicyclic cyclo-
butylanilines, which were readily prepared in four steps. These
cyclobutylanilines generally gave higher yields of the annulation
products than the monocyclic ones, and modest to good
diastereoselectivities were achieved (Scheme 13). The opening
of bicyclic cyclobutylanilines was completely regioselective,
producing the more stable carbon radical.
■
[4 + 2] ANNULATION IN FLOW
In line with other reported photochemistries, we encountered
long reaction time and difficulty in scale up during our studies
of the [4 + 2] annulation. Flow is often considered to be the de
facto solution to these issues. We applied flow to the [4 + 2]
annulation of cyclobutylanilines with alkenes, alkynes, and
diynes.31 The reaction time was often cut by half in flow
(Scheme 14A). A gram scale annulation was also successfully
achieved in flow with a slightly lower yield (Scheme 14B).
Scheme 13. [4 + 2] Annulation Scope of Bicyclic Cyclobutylanilines
G DOI: 10.1021/acs.accounts.6b00263
Article
■ AMINE RADIAL CATIONS AS A WAY TO FORM
SUBSTITUTED HETEROCYCLES
One of the hallmark features of carbon radicals is their ability to
engage in inter- and intramolecular addition to various types of
π bonds in cascade reactions. We wondered whether photo-
generated amine radical cations could mimic carbon radicals in
these types of cascade reactions leading to indoles and
indolines, two important motifs in medicinal chemistry. Similar
transformations have been successfully realized using metals,
such as Pd. Our approach was centered on a C−N bond
forming photoredox cascade in which an Umpolung reaction of
anilines was employed. The amine radical cations of anilines
acted as an electrophile, while alkenes functioned as a
nucleophile. Thus, it was mechanistically distinct from the
Pd-catalyzed process and offered advantages, including lower
reaction temperatures. More importantly, our approach utilized
the cascade reaction initialized by the photogenerated amine
radical cations, which could furnish fused indoles and indolines.
Initially, we designed styrene 18 bearing an ortho
sulfonamide as the substrate to test the cascade reaction, as
shown in Scheme 15A. The sulfonamide group was chosen as
the nucleophile because it was not photo-oxidizable.32 The
alkene moiety would be oxidized to the electrophilic alkene
radical cation by single electron photooxidants, such as the
photoexcited Ru(bpz)32+*. The targeted C−N bond formation
would be realized via nucleophilic capture of the radical cation
by the sulfonamide group. Instead of yielding the anticipated α-
phenyl indole, dimerization adduct 19 was observed. We
immediately revised our strategy and decided to attempt an
Umpolung approach in which an electron-rich amine, such as
aniline, would be oxidized to the amine radical cation by the
photoexcited Ru(bpz)32+*. The amine radical cation would
then act as an electrophile in the addition reaction to the
alkene. The use of amine radical cations in electrophilic
addition reactions to alkenes is an established pathway and has
been applied to the synthesis of many interesting molecules.33
The sulfonamide group was replaced by an aniline, as shown in
Scheme 15B with compound 20, which resulted in no reaction.
This was a puzzling observation, since cyclic voltammetry
experiments revealed an oxidation peak potential of 0.9 V for
20, which suggested that 20 should be readily oxidized by the
photoexcited Ru(bpz)3(PF6)2 (1.31 V reduction potential for
Ru(bpz)32+*/Ru(bpz)3+ vs Ag/AgCl).34,35 Nonetheless, since
electron-rich para-alkoxyphenyl groups lower the oxidation
Acc. Chem. Res. XXXX, XXX, XXX−XXX
Accounts of Chemical Research
Scheme 14. [4 + 2] Annulation in Flow
Scheme 15. Preliminary Data for Indole Synthesis Using Visible Light
peak potential of anilines, the phenyl group was exchanged for a
para-methoxyphenyl (PMP) group (compound 21). When this
compound was subjected to the photochemistry conditions, the
desired indole 22 was formed. This result marked the first-
reported example of electrophilic addition to an olefin using
photogenerated amine radical cations. With this new reaction in
hand, we set out to find the optimal reaction conditions and a
substrate scope tolerant of this process.34
H DOI: 10.1021/acs.accounts.6b00263
Article
■ VISIBLE LIGHT PHOTOCATALYZED SYNTHESIS OF
SUBSTITUTED INDOLES
Initial screening revealed Ru(bpz)3(PF6)2 as the optimal
photocatalyst in acetonitrile (Table 3, entries 1 and 2). It was
found that silica gel significantly increased reaction efficiency,
probably due to its mild acidity and ability to adsorb oxygen
(entry 2). This reaction only proceeded under aerobic
conditions, suggesting a critical role of atmospheric oxygen
for catalyst turnover (entry 3). Ru(bpy)3(PF6)2, possessing a
Acc. Chem. Res. XXXX, XXX, XXX−XXX
Accounts of Chemical Research Article

Table 3. Reaction Screen ■ FUSED INDOLINE SYNTHESIS USING TETHERED

STYRENYL ANILINES
We later expanded the scope of this chemistry by exploiting the
presumable benzylic cation to generate fused indolines.38 We
wondered whether a tethered styrenyl aniline bearing a
nucleophile could efficiently trap the cation in an intra-
time molecular SN1 fashion to yield the C2,C3-fused indoline. To
entry additive catalyst (h) yield of 23 (%) test this hypothesis, we synthesized styrenyl aniline 26 and
1 none Ru(bpz)3(PF6)2 24 31 subjected it to the photochemistry conditions (Scheme 18).
2 silica gel Ru(bpz)3(PF6)2 5 88 Fused indoline 27 was obtained as a single diastereomer.
3 silica gel (degassed) Ru(bpz)3(PF6)2 24 0 Further reaction screening identified the optimal conditions to
4 silica gel Ru(bpy)3(PF6)2 24 19 include Ru(bpy)3(PF6)2 and acetic acid with exposure to air.
5 TFA Ru(bpz)3(PF6)2 24 21 A substrate scope was also developed for this chemistry,
which included a variety of oxygen- and nitrogen-based
nucleophiles with a range of structural modifications (Scheme
19A). As expected, the olefin’s geometry had no effect on the
lower reduction potential for its photoexcited state (0.78 V),
diastereoselectivity of the reaction, since >99:1 dr was observed
was less effective in the transformation (entry 4). In order to
in all cases. Additionally, the 3,4,5-trimethoxyphenyl group was
test the acidity parameters, trifluoroacetic acid was examined,
identified as the removable N-aryl group, which was cleaved
and a lower yield of the indole product 23 was obtained due to
using CAN in sulfuric acid, thus improving the substrate scope
immediate decomposition of the product by the strong acid
(Scheme 19B). These findings revealed a photocatalyzed
(entry 5).
indoline formation by way of diamination39 or amino-
A substrate scope was then developed to demonstrate
hydroxylation40 of alkenes to form two consecutive chiral
applicability of this new reaction. Virtually every position of the
centers, including a tetrasubstituted carbon center.

indole ring could be varied when generating monosubstituted
indoles. In fact, the only structural requirement for this
chemistry was the para-alkoxyphenyl group on the nitrogen
atom. A condensed substrate scope is shown in Scheme 16.
An additional observation was made when using internal
olefins such as 24 in Scheme 17. In this case, a 1,2-alkyl shift/
aromatization generated the 2,3-disubstituted indole product.
This phenomenon has been documented by the Driver group
when a benzylic cation intermediate is formed.36 Therefore, we
concluded that our chemistry would likely go through a similar
intermediate. In all our 1,2-shift cases, aromatic groups were
found to migrate much faster than alkyl groups, resulting in a
single disubstituted indole product of type 25.
This chemistry marked a new way to use photogenerated
amine radical cations, effectively achieving amination of vinylic
■
MECHANISM
A mechanism for these transformations is detailed in Scheme
20. Upon visible light irradiation, the photogenerated amine
radical cation, which remains in the protonated form
presumably due to the presence of the weak acid, adds
electrophilically to the styrene, generating the quaternary
ammonium salt. Although we present this as an electrophilic
addition, we cannot rule out a 5π pericyclic Nazarov-like
cyclization,41 which could give the same intermediate.
Subsequent deprotonations and redox manipulations provide
the key benzylic cation, which sets up the indole formation by
way of deprotonation or 1,2-shift/deprotonation. Alternatively,
the key benzylic cation is trapped intramolecularly by a tethered
nucleophile to afford a fused indoline.

■
C−H bonds using mild, aerobic conditions. The only other
known example to date was reported by Knowles, in which his
group generated substituted pyrrolidines and piperidines CONCLUSION
through olefin hydroamination using photogenerated amine In summary, photogenerated amine radical cations were
radical cations.37 exploited to generate a variety of carbocycles and heterocycles.

Scheme 16. Summarized Substrate Scope

I DOI: 10.1021/acs.accounts.6b00263
Acc. Chem. Res. XXXX, XXX, XXX−XXX
Accounts of Chemical Research
Scheme 17. 1,2-Shift Substrate Scope
Scheme 18. Examination of Indoline Formation
Scheme 19. Substrate Scope and Removal of para-
Alkoxyphenyl Component
Our initial reports centered on the C−C bond cleavage of
photogenerated amine radical cations derived from cyclo-
propyl- and cyclobutyl-anilines to generate distonic radical
cations. These odd-electron species displayed some unique
reactivity, permitting the annulation reaction with various
alkenes and alkynes in an overall redox neutral process. We also
developed a new application of photogenerated amine radical
cations in electrophilic addition to tethered alkenes. Substituted
indoles and indolines were synthesized from styryl anilines. The
amine radical cations set up the cascade reaction including the
electrophilic addition to the styrene, the formation of a
J DOI: 10.1021/acs.accounts.6b00263
Article
Scheme 20. Proposed Mechanism
synthetically useful benzylic cation intermediate, and a series of
redox manipulations. All of our chemistries display the power of
photogenerated amine radical cations to participate in cascade
reactions to provide structurally diverse and interesting
products.
■ AUTHOR INFORMATION
Corresponding Author
*E-mail: nzheng@uark.edu.
Funding
This work was funded by the University of Arkansas, the
Arkansas Biosciences Institute, NIGMS (Grant P30
GM103450), and NSF (Career CHE-1255539).
Notes
The authors declare no competing financial interest.
Biographies
Scott Morris received a B.S. in Biochemistry from Stephen F. Austin
State University in 2010. He is now completing his Ph.D. in organic
chemistry at the University of Arkansas under the guidance of
Acc. Chem. Res. XXXX, XXX, XXX−XXX
Accounts of Chemical Research
Professor Nan Zheng, focusing on the formation of C−N bonds using
visible light photoredox catalysis.
Jiang Wang received his B.S. in Chemistry from Lanzhou University,
China. He joined the University of Arkansas in 2011 and is currently a
Ph.D. candidate under the direction of Professor Nan Zheng, studying
photoredox catalysis.
Nan Zheng did his undergraduate studies at the University of Science
and Technology of China and obtained his Ph.D. from the University
of Michigan with Professor William R. Roush. He then performed
postdoctoral studies at MIT with Stephen L. Buchwald before starting
his independent career at the University of Arkansas in 2008. His
research interest is centered on photoredox catalysis.
■
ACKNOWLEDGMENTS
We are grateful to all our co-workers for their outstanding
intellectual and experimental contribution.
■
REFERENCES
(1) For reviews, see: (a) Stella, L. Radicals in Organic Synthesis, 1st
ed.; Wiley-VCH: Weinheim, Germany, 2001; Vol. 2, pp 407−424.
(b) Zard, S. Z. Recent progress in the generation and use of nitrogen-
centered radicals. Chem. Soc. Rev. 2008, 37, 1603−1618.
(2) Hioe, J.; Sakic, D.; Vrcek, V.; Zipse, H. The stability of nitrogen-
centered radicals. Org. Biomol. Chem. 2015, 13, 157−169.
(3) Newcomb, M.; Marquardt, D. J.; Deeb, T. M. N-Hydroxypyr-
idine-2-thione carbamates. IV. A comparison of 5-exo cyclizations of
an aminyl radical and an aminium cation radical. Tetrahedron 1990, 46,
2317−2328.
(4) Hu, J.; Wang, J.; Nguyen, T. H.; Zheng, N. The chemistry of
amine radical cations produced by visible light photoredox catalysis.
Beilstein J. Org. Chem. 2013, 9, 1977−2001.
(5) Lee, J.; U, J. S.; Blackstock, S. C.; Cha, J. K. Facile Ring Opening
of Tertiary Aminocyclopropanes by Photooxidation. J. Am. Chem. Soc.
1997, 119, 10241−10242.
(6) (a) Beeson, T. D.; Mastracchio, A.; Hong, J.-B.; Ashton, K.;
MacMillan, D. W. C. Enantioselective Organocatalysis Using SOMO
Activation. Science 2007, 316, 582−585. (b) Lee, H. B.; Sung, M. J.;
Blackstock, S. C.; Cha, J. K. Radical Cation-Mediated Annulation.
Stereoselective Construction of Bicyclo[5.3.0]decan-3-ones by Aerobic
Oxidation of Cyclopropylamines. J. Am. Chem. Soc. 2001, 123, 11322−
11324.
(7) (a) Zhong, B.; Silverman, R. B. Identification of the Active Site
Cysteine in Bovine Liver Monoamine Oxidase B. J. Am. Chem. Soc.
1997, 119, 6690−6691. (b) Wessjohann, L. A.; Brandt, W.; Thiemann,
T. Biosynthesis and Metabolism of Cyclopropane Rings in Natural
Compounds. Chem. Rev. 2003, 103, 1625−1647.
(8) Shono, T.; Matsumura, Y.; Inoue, K.; Ohmizu, H.; Kashimura, S.
Electroorganic chemistry. 62. Reaction of iminium ion with
nucleophile: a versatile synthesis of tetrahydroquinolines and
julolidines. J. Am. Chem. Soc. 1982, 104, 5753−5757.
(9) Madelaine, C.; Six, Y.; Buriez, O. Electrochemical aerobic
oxidation of aminocyclopropanes to endoperoxides. Angew. Chem., Int.
Ed. 2007, 46, 8046−8049.
(10) Condie, A. G.; Gonzalez-Gomez, J. C.; Stephenson, C. R. J.
Visible-Light Photoredox Catalysis: Aza-Henry Reactions via C-H
Functionalization. J. Am. Chem. Soc. 2010, 132, 1464−1465.
(11) For reviews, see: (a) Teply, F. Photoredox catalysis by
[Ru(bpy)3]2+ to trigger transformations of organic molecules.
Organic synthesis using visible-light photocatalysis and its 20th
century roots. Collect. Czech. Chem. Commun. 2011, 76, 859−917.
(b) Xuan, J.; Xiao, W.-J. Visible-Light Photoredox Catalysis. Angew.
Chem., Int. Ed. 2012, 51, 6828−6838. (c) Prier, C. K.; Rankic, D. A.;
MacMillan, D. W. C. Visible Light Photoredox Catalysis with
Transition Metal Complexes: Applications in Organic Synthesis.
Chem. Rev. 2013, 113, 5322−5363.
K DOI: 10.1021/acs.accounts.6b00263
Article
(12) For reviews, see: (a) Maity, S.; Zheng, N. A photo touch on
amines: new synthetic adventures of nitrogen radical cations. Synlett
2012, 23, 1851−1856. (b) Chen, J.-R.; Hu, X.-Q.; Lu, L.-Q.; Xiao, W.-
J. Visible light photoredox-controlled reactions of N-radicals and
radical ions. Chem. Soc. Rev. 2016, 45, 2044−2056.
(13) Zhu, M.; Zheng, N. Photoinduced cleavage of N-N bonds of
aromatic hydrazines and hydrazides by visible light. Synthesis 2011,
2011, 2223−2236.
(14) Stirk, K. M.; Kiminkinen, L. K. M.; Kenttamaa, H. I. Ion−
molecule reactions of distonic radical cations. Chem. Rev. 1992, 92,
1649−1665.
(15) (a) Giese, B. Formation of carbon-carbon bonds by addition of
radicals to alkenes. Angew. Chem., Int. Ed. Engl. 1983, 22, 753−754.
(b) Cekovic, Z.; Saicic, R. Free radical annulation of cyclopentane ring.
Tetrahedron Lett. 1986, 27, 5893−5896. (c) Saicic, R. N.; Cekovic, Z.
Cyclopentane ring formation in the cycloaddition reaction of 3-alkenyl
radicals to radicophilic olefins. Tetrahedron 1990, 46, 3627−3640.
(16) For reviews and examples of these processes, see: (a) Reissig,
H.-U.; Zimmer, R. Donor-Acceptor-Substituted Cyclopropane De-
rivatives and Their Application in Organic Synthesis. Chem. Rev. 2003,
103, 1151−1196. (b) Yu, M.; Pagenkopf, B. L. Recent advances in
donor-acceptor (DA) cyclopropanes. Tetrahedron 2005, 61, 321−347.
(c) Lebold, T. P.; Kerr, M. A. Intramolecular annulations of donor-
acceptor cyclopropanes. Pure Appl. Chem. 2010, 82, 1797−1812.
(d) Mel’nikov, M. Y.; Budynina, E. M.; Ivanova, O. A.; Trushkov, I. V.
Recent advances in ring-forming reactions of donor-acceptor cyclo-
propanes. Mendeleev Commun. 2011, 21, 293−301. (e) Cavitt, M. A.;
Phun, L. H.; France, S. Intramolecular donor-acceptor cyclopropane
ring-opening cyclizations. Chem. Soc. Rev. 2014, 43, 804−818.
(f) Schneider, T. F.; Kaschel, J.; Werz, D. B. A new golden age for
donor-acceptor cyclopropanes. Angew. Chem., Int. Ed. 2014, 53, 5504−
5523. (g) Grover, H. K.; Emmett, M. R.; Kerr, M. A. Carbocycles from
donor-acceptor cyclopropanes. Org. Biomol. Chem. 2015, 13, 655−671.
(h) Novikov, R. A.; Tomilov, Y. V. Dimerization of donor-acceptor
cyclopropanes. Mendeleev Commun. 2015, 25, 1−10.
(17) Maity, S.; Zhu, M.; Shinabery, R. S.; Zheng, N. Intermolecular
[3 + 2] Cycloaddition of Cyclopropylamines with Olefins by Visible-
Light Photocatalysis. Angew. Chem., Int. Ed. 2012, 51, 222−226.
(18) Madelaine, C.; Buzas, A. K.; Kowalska-Six, J. A.; Six, Y.; Crousse,
B. Diastereoselective Ti-mediated preparation of bicyclic amino-
cyclopropanes from N-alkenyl amides. Tetrahedron Lett. 2009, 50,
5367−5371.
(19) Spellmeyer, D. C.; Houk, K. N. Force-field model for
intramolecular radical additions. J. Org. Chem. 1987, 52, 959−974.
(20) (a) Nguyen, T. H.; Maity, S.; Zheng, N. Visible light mediated
intermolecular [3 + 2] annulation of cyclopropylanilines with alkynes.
Beilstein J. Org. Chem. 2014, 10, 975−980. (b) Nguyen, T. H.; Morris,
S. A.; Zheng, N. Intermolecular [3 + 2] Annulation of Cyclopropylani-
lines with Alkynes, Enynes, and Diynes via Visible Light Photo-
catalysis. Adv. Synth. Catal. 2014, 356, 2831−2837.
(21) Wille, U. Radical Cascades Initiated by Intermolecular Radical
Addition to Alkynes and Related Triple Bond Systems. Chem. Rev.
2013, 113, 813−853.
(22) Li, X.; Grimm, M. L.; Igarashi, K.; Castagnoli, N., Jr.; Tanko, J.
M. The first calibration of an aminiumyl radical ion clock: Why N-
cyclopropylanilines may be poor mechanistic probes for single electron
transfer. Chem. Commun. 2007, 25, 2648−2650.
(23) Kavarnos, G. J.; Turro, N. J. Photosensitization by reversible
electron transfer: theories, experimental evidence, and examples. Chem.
Rev. 1986, 86, 401−449.
(24) Verkade, J. M. M.; van Hemert, L. J. C.; Quaedflieg, P. J. L. M.;
Alsters, P. L.; van Delft, F. L.; Rutjes, F. P. J. T. Mild and efficient
deprotection of the amine protecting p-methoxyphenyl (PMP) group.
Tetrahedron Lett. 2006, 47, 8109−8113.
(25) Wiberg, K. B. The Chemistry of Cyclobutanes; Wiley: Chichester,
U.K., 2005. Khoury, P. R.; Goddard, J. D.; Tam, W. Ring strain
energies: substituted rings, norbornanes, norbornenes and norborna-
dienes. Tetrahedron 2004, 60, 8103−8112.
Acc. Chem. Res. XXXX, XXX, XXX−XXX
Accounts of Chemical Research Article

(26) Jin, J.; Newcomb, M. Rate Constants for Reactions of Alkyl

Radicals with Water and Methanol Complexes of Triethylborane. J.
Org. Chem. 2007, 72, 5098−5103.
(27) Maeda, Y.; Ingold, K. U. Kinetic applications of electron
paramagnetic resonance spectroscopy. 35. The search for a
dialkylaminyl rearrangement. Ring opening of N-cyclobutyl-N-n-
propylaminyl. J. Am. Chem. Soc. 1980, 102, 328−331.
(28) Horner, J. H.; Martinez, F. N.; Musa, O. M.; Newcomb, M.;
Shahin, H. E. Kinetics of Dialkylaminium Cation Radical Reactions:
Radical Clocks, Solvent Effects, Acidity Constants, and Rate Constants
for Reactions with Hydrogen Atom Donors. J. Am. Chem. Soc. 1995,
117, 11124−11133.
(29) Wang, J.; Zheng, N. The cleavage of a C-C Bond in
cyclobutylanilines by visible-light photoredox catalysis: Development
of a [4 + 2] annulation method. Angew. Chem., Int. Ed. 2015, 54,
11424−11427.
(30) Minisci, F.; Zammori, P.; Bernardi, R.; Cecere, M.; Galli, R.
Nucleophilic character of alkyl radicals generated in redox processes.
III. Reactivity of alkyl radicals towards conjugated olefins. Tetrahedron
1970, 26, 4153−4166.
(31) Wang, J.; Nguyen, T. H.; Zheng, N. Photoredox-catalyzed [4 +
2] annulation of cyclobutylanilines with alkenes, alkynes, and diynes in
continuous flow. Sci. China: Chem. 2016, 59, 180−183.
(32) Ross, S. D.; Finkelstein, M.; Rudd, E. J. Anodic oxidations. VIII.
Anodic oxidation of N,N-dimethylmethanesulfonamide in alcohols and
in acetic acid. J. Org. Chem. 1972, 37, 2387−2891.
(33) (a) Karady, S.; Corley, E. G.; Abramson, N. L.; Weinstock, L. M.
Cyclization of electrochemically generated nitrogen radicals. A novel
synthesis of 11-substituted dibenzo[a,d]cycloheptenimines. Tetrahe-
dron Lett. 1989, 30, 2191−2194. (b) Newcomb, M.; Deeb, T. M. N-
Hydroxypyridine-2-thione carbamates as aminyl and aminium radical
precursors. Cyclizations for synthesis of the pyrrolidine nucleus. J. Am.
Chem. Soc. 1987, 109, 3163−3165.
(34) Maity, S.; Zheng, N. A Visible-Light-Mediated Oxidative C-N
Bond Formation/Aromatization Cascade: Photocatalytic Preparation
of N-Arylindoles. Angew. Chem., Int. Ed. 2012, 51, 9562−9566.
(35) Crutchley, R. J.; Lever, A. B. P. Ruthenium(II) tris(bipyrazyl)
dication - a new photocatalyst. J. Am. Chem. Soc. 1980, 102, 7128−
7129.
(36) Stokes, B. J.; Liu, S.; Driver, T. G. Rh2(II)-Catalyzed Nitro-
Group Migration Reactions: Selective Synthesis of 3-Nitroindoles
from β-Nitro Styryl Azides. J. Am. Chem. Soc. 2011, 133, 4702−4705.
(37) Musacchio, A. J.; Nguyen, L. Q.; Beard, G. H.; Knowles, R. R.
Catalytic Olefin Hydroamination with Aminium Radical Cations: A
Photoredox Method for Direct C-N Bond Formation. J. Am. Chem.
Soc. 2014, 136, 12217−12220.
(38) Morris, S. A.; Nguyen, T. H.; Zheng, N. Diastereoselective
Oxidative C-N/C-O and C-N/C-N Bond Formation Tandems
Initiated by Visible Light: Synthesis of Fused N-Arylindolines. Adv.
Synth. Catal. 2015, 357, 2311−2316.
(39) De Jong, S.; Nosal, D. G.; Wardrop, D. J. Methods for direct
alkene diamination, new & old. Tetrahedron 2012, 68, 4067−4105.
(40) Nilov, D.; Reiser, O. The Sharpless asymmetric amino-
hydroxylation - scope and limitation. Adv. Synth. Catal. 2002, 344,
1169−1173.
(41) Radosevich, A. T.; Wiest, O. Quantum Mechanical Study of the
Ring-Closing Reaction of the Hexatriene Radical Cation. J. Org. Chem.
2001, 66, 5808−5813.

L DOI: 10.1021/acs.accounts.6b00263
Acc. Chem. Res. XXXX, XXX, XXX−XXX