How Radical Are "radical" Photocatalysts? A Closed-Shell Meisenheimer Complex Is Identified as a Super-Reducing Photoreagent

Article abstract of DOI:10.1021/jacs.1c06844

Super-reducing excited states have the potential to activate strong bonds, leading to unprecedented photoreactivity. Excited states of radical anions, accessed via reduction of a precatalyst followed by light absorption, have been proposed to drive photoredox transformations under super-reducing conditions. Here, we investigate the radical anion of naphthalene monoimide as a photoreductant and find that the radical doublet excited state has a lifetime of 24 ps, which is too short to facilitate photoredox activity. To account for the apparent photoreactivity of the radical anion, we identify an emissive two-electron reduced Meisenheimer complex of naphthalene monoimide, [NMI(H)]-. The singlet excited state of [NMI(H)]- is a potent reductant (-3.08 V vs Fc/Fc+), is long-lived (20 ns), and its emission can be dynamically quenched by chloroarenes to drive a radical photochemistry, establishing that it is this emissive excited state that is competent for reported C-C and C-P coupling reactivity. These results provide a mechanistic basis for photoreactivity at highly reducing potentials via singlet excited state manifolds and lays out a clear path for the development of exceptionally reducing photoreagents derived from electron-rich closed-shell anions.

Full text of DOI:10.1021/jacs.1c06844

pubs.acs.org/JACS
Article
How Radical Are “Radical” Photocatalysts? A Closed-Shell
Meisenheimer Complex Is Identified as a Super-Reducing
Photoreagent
Adam J. Rieth, Miguel I. Gonzalez, Bryan Kudisch, Matthew Nava, and Daniel G. Nocera*
Cite This: https://doi.org/10.1021/jacs.1c06844
Read Online
sı
*
Metrics & More
Article Recommendations
Supporting Information
ACCESS
ABSTRACT: Super-reducing excited states have the potential to
activate strong bonds, leading to unprecedented photoreactivity.
Excited states of radical anions, accessed via reduction of a
precatalyst followed by light absorption, have been proposed to
drive photoredox transformations under super-reducing condi-
tions. Here, we investigate the radical anion of naphthalene
monoimide as a photoreductant and find that the radical doublet
excited state has a lifetime of 24 ps, which is too short to facilitate
photoredox activity. To account for the apparent photoreactivity of
the radical anion, we identify an emissive two-electron reduced
Meisenheimer complex of naphthalene monoimide, [NMI(H)]−.
The singlet excited state of [NMI(H)]− is a potent reductant
(−3.08 V vs Fc/Fc+), is long-lived (20 ns), and its emission can be
dynamically quenched by chloroarenes to drive a radical photochemistry, establishing that it is this emissive excited state that is
competent for reported C−C and C−P coupling reactivity. These results provide a mechanistic basis for photoreactivity at highly
reducing potentials via singlet excited state manifolds and lays out a clear path for the development of exceptionally reducing
photoreagents derived from electron-rich closed-shell anions.
conversion for such a ΔS = 0 nonradiative decay to a doublet
ground state is exceptionally fast. Spectroscopic investigations
of organic radicals have demonstrated nearly universally short
excited-state lifetimes on the ps time scale,14−19 which may be
even shorter than a single vibrational period (7 fs).20 Although
electron transfer from radical anions does occur intra-
molecularly in covalently linked donor−acceptor systems,21,22
the short lifetime of doublet excited states cannot support
intermolecular processes of photoredox reactions, as the
lifetime (τo) of the active excited state should be longer than
the average time between molecular encounters, that is, τo ≥
1/kd[sub] where kd is the rate of diffusion (typically ∼109 −
1010 M−1 s−1), and [sub] is the concentration of substrate. For
a substrate concentration of 0.1 M, an approximate lower
bound for a SET process requiring diffusion is τo ≥ 1−10 ns.
Therefore, the measured excited state lifetimes of nearly all
radical doublet species are orders of magnitude too short to
result in productive photochemistry. Indeed, excited state
INTRODUCTION
■
Exploration of the redox potential limits accessible by
photoinduced single-electron transfer can unlock new reac-
tivity and expand the range of processes amenable to
photoredox chemistry and the design of attendant synthetic
methods. Using standard photocatalysts with highest occupied
molecular orbital (HOMO) energy levels near to or more
oxidizing than that of the normal hydrogen electrode (NHE =
0), photoredox processes using visible light have been
practically limited to less than ∼2 V in either direction.1,2
Recently, these limits have been circumvented by shifting the
ground state potential of the photocatalyst via redox processes
prior to light absorption; oxidation of a precatalyst followed by
light absorption generates strongly oxidizing excited states3−6
and reduction of a precatalyst followed by light absorption can
generate super-reducing excited states.7−11 These purported
two-photon or electrophotocatalytic processes dramatically
expand the range of synthetically useful bond activation
photoprocesses.7−11 Additionally, because the processes are
photoactivated, they are selective for single electron transfer
(SET) even in cases where direct electrolysis would result in
overoxidation or -reduction.7 Although the highly reducing
nature of excited organic radical anions has long been
recognized,17,12,13 the mechanism and energetics of a process
driven by a doublet excited state is quizzical as internal
Received: July 1, 2021
https://doi.org/10.1021/jacs.1c06844
J. Am. Chem. Soc. XXXX, XXX, XXX−XXX
© XXXX American Chemical Society
A

Journal of the American Chemical Society
pubs.acs.org/JACS
Article
spectral features with lifetimes longer than 1 ns measured for
several presumptive radical anions23,24 were later shown to
instead arise from longer-lived closed shell impurities. For
instance, the radical anions of both 9,10-dicyanoanthracene
and anthraquinone were later found to contain the emissive
byproducts cyanoanthrolate or bianthrolate, respec-
tively.14−19,24 Therefore, there is an imperative to better
understand the nature of super-reducing radical photore-
agents.7,9,10 A proper mechanistic understanding of the origins
of super-reducing photochemistry can facilitate their applica-
tion in the design of photoredox methods and expand the
possibilities for bond activation via reductive SET.25
Here, we show that for one recently reported photo-
electrocatalyst, naphthlalene monoimide (NMI),7 the photo-
active component is not a radical but a two-electron reduced
species. Using a combination of cyclic voltammetry,
spectroelectrochemistry, and ultrafast spectroscopy, we dem-
onstrate that the two-electron reduction of NMI produces an
emissive Meisenheimer complex,26,27 [NMI(H)]−[TBA]+
(TBA = tetrabutylammonium), which has been isolated and
crystallographically characterized. Time-resolved spectroscopy
reveals that the lifetime of the doublet excited state of the
radical anion (τ[2NMI−•]*) is 24 ps, which is too short to
participate in photoredox chemistry. By contrast, the lifetime
of the singlet excited state of the Meisenheimer complex
(τ[1NMI(H)−]*) is 20 ns. This long fluorescent lifetime
together with a−3.08 V vs Fc/Fc+ excited state energy gives
rise to super-reducing chemistry. Indeed, the emissive excited
state of the doubly reduced Meisenheimer complex is
dynamically quenched by haloarene substrates and leads to
the previously observed C−C and C−P coupling reactivity.7
These results provide a clear mechanistic underpinning for
super-reducing photo/electro-catalysis, laying out a path for
the further development of powerfully reducing closed-shell
anions as photoreagents for novel bond activation pathways.
Figure 1. (A) Cyclic voltammograms of 1 mM NMI in 0.1 M
TBAPF6 propylene carbonate solution, scanning at 50 mV s−1 from
0.2 to −2.0 V (green) and 0.2 to −2.8 V (orange), using a glassy
carbon working electrode, Pt mesh counter electrode, and leak-free
Ag/AgCl reference electrode.
RESULTS
■
Cyclic voltammetry (CV) of NMI in propylene carbonate
containing 0.1 M tetrabutylammonium hexafluorophosphate
(TBAPF6) revealed two reductive events (Figure 1). The first
reduction wave occurs at −1.7 V vs Fc/Fc+ and is fully
reversible, so long as the sweep potential is limited to avoid the
second reductive wave (Figure 1, green). The first reductive
event corresponds to a single electron transfer to reversibly
form the NMI radical anion, as previously observed.7,18 The
second reductive feature is irreversible, at approximately −2.6
V (Figure 1, orange), and also corresponds to a one-electron
process, coupled to a chemical step. The appearance of this
irreversible reductive wave is accompanied by the emergence
of an oxidative feature at −0.7 V, as well as a decrease in the
peak height for reoxidation of the previously reversible feature
at −1.7 V. Increasing the scan rate results in the formation of
two distinct oxidation features near −0.7 V, as well as partial
reversibility of the −0.7 V feature upon reduction (Figure S1,
S2). The cyclic voltammetry features are not altered after 10
cycles (Figure S3). Upon scanning past −2.6 V, the
reversibility of the radical anion feature at −1.7 V can be
recovered when the scan is limited to −2.0 V on the
subsequent cycle (Figure S4).
Figure 2. UV−visible spectroelectrochemistry of 1 mM NMI in 0.1 M
TBAPF6 dimethylacetamide solution. The Pt working electrode was
poised at the 1e− reducing potential of Eappl = −2.3 V (A) and the 2e−
reducing potential of Eappl = −3.0 V (B).
absorption features appear at 422 and 493 nm, as well as
weaker absorption signals at 750 and 850 nm (Figure 2A,
green), all of which are characteristic of a naphthalene imide
radical anion.18 Bulk electrolysis at −2.3 V vs Fc/Fc+ in a W-
cell generated a bright lime green solution in the catholyte
(Figure S5), with the same absorption features in Figure 2A.
Examination of this electrolyzed solution by ultrafast transient
absorption (λexc = 420 nm) revealed a bleach signal as well as
an excited state absorption feature, both of which decay with a
Figure 2 shows spectroelectrochemistry data when the
working electrode in a two-compartment H-cell is polarized to
−2.3 V vs Fc/Fc+, which is sufficiently beyond the first wave to
drive the one electron reduction of NMI. Prominent
B
https://doi.org/10.1021/jacs.1c06844
J. Am. Chem. Soc. XXXX, XXX, XXX−XXX

Journal of the American Chemical Society
pubs.acs.org/JACS
Article
Figure 3. (A) Transient absorption spectra for NMI electrolyzed at −2.3 V results in a transient absorption feature at λmax = 500 nm and a
bleaching signal between 380 to 425 nm; the bleach signal is interrupted by a sharp feature at 420 nm due to the excitation pump used to excite the
most prominent absorptive feature of the NMI radical anion. (B) Single exponential decay fits (red lines) to the transient absorption data in panel
■
■
A for the excited state absorption intensity averaged between 495−525 nm ( , light blue) and the bleach feature between 385−395 nm ( , sea
green), both of which exhibit 24 ps lifetimes. (C) Normalized absorption spectra of NMI (black), NMI electrolyzed at −2.3 V (one-electron
reduced species, green), NMI electrolyzed at −3 V (two-electron reduced species, burnt orange), overlaid with the normalized emission spectrum
of the −3.0 V electrolyzed sample (orange dash) and the excitation scan for the emission at 600 nm (light orange dash). (D) Normalized
absorption (solid) and emission (dashed) spectra for NMI electrolyzed at −3.0 V vs Fc/Fc+ (burnt orange), NMI treated with TBAF in DMAc
(pink), and NMI treated with NaBH4 in DME (blue). (E) fs-pumped streak camera emission image (λexc = 480 nm) for NMI treated with TBAF in
DMAc. Colors represent photon counts (purple is low and red is high). The red line shows a monoexponential fit of data averaged in a 540−560
nm spectral window. (F) Dynamic Stern−Volmer quenching plots of solutions of NMI treated with TBAF in DMAc in the presence of 4-
methylchlorobenzoate or chlorobenzene (fitted data presented in Figures S9 and S10).
lifetime of 24 ps (Figure 3A,B). This result is in agreement
with previous measurements of rylene imide radical anion
excited states.18
nm. However, no emission is observed arising from the NMI
radical anion, even when excited at its peak absorption
wavelength λmax = 420 nm and the emission detection
wavelength is scanned into the infrared spectral region (λdet
= 800−1500 nm, InGaAs detector) (Figure S8).
NMI may be reduced chemically using either fluoride-
mediated solvent oxidation with tetrabutylammonium fluoride
(TBAF) in N,N′-dimethylacetamide (DMAc) or treatment
with NaBH4 in dimethoxyethane (DME). For both reduction
reactions, the absorption and emission spectra of the reduced
product match those obtained for the electrochemically two-
electron reduced NMI (Figure 3D). Emission lifetime
measurements were performed on NMI solutions in DMAc
treated with TBAF using a fs-pumped streak camera (Figure
3E). The data were averaged in the 540 to 560 nm window and
fitted to a single exponential decay to furnish a lifetime of 20.1
ns (Figure 3E, red line).
Figure 2B shows the spectroelectrochemistry of NMI when
the applied potential in Figure 2A is shifted to −3.0 V, beyond
the second reductive wave in the CV. Strong absorption
features at 385 and 490 nm grow (Figure 2B, orange) with the
concomitant disappearance of the NMI radical anion
absorption features. Bulk electrolysis in a W-cell at −3.0 V
vs Fc/Fc+ generated a bright orange solution in the catholyte,
with visible green emission at λmax = 550−600 nm (Figure 3C
and Figure S5), just beyond the lowest energy absorption
feature of the two-electron reduced NMI. A fluorescence
excitation spectrum (Figure 3C, light orange dash) coincides
with the absorption features of the doubly reduced NMI
species, with the most intense emission arising from excitation
of the absorption bands at 385 and 490 nm. A lower-intensity
emission with λmax = 550−600 nm is also observed for samples
electrolyzed at −2.3 V, and its excitation spectrum also
matches that of the two-electron reduced NMI (Figure S7).
This result is consistent with the CV in Figure 1, which shows
the onset of the second reduction occurs at −2.3 V. In view of
the lowest energy absorption of the NMI radical anion,
emission, if present, would be expected to occur beyond 850
Addition of chloroarene substrates to NMI treated with
TBAF in DMAc resulted in dynamic quenching of the
emission lifetime (Figures S9 and S10). Figure 3F shows
well-behaved Stern−Volmer kinetics as measured by lifetime
quenching. For methyl-4-chlorobenzoate, the quenching rate
constant is kq = 3 × 109 M−1 s−1, and for the more challenging
C
https://doi.org/10.1021/jacs.1c06844
J. Am. Chem. Soc. XXXX, XXX, XXX−XXX

Journal of the American Chemical Society
pubs.acs.org/JACS
Article
Figure 4. (A) Single-electron reduction of naphthalene monoimide (NMI) furnishes the radical anion with a doublet excited state ([2NMI−• ]*)
lifetime of 24 ps. A second electron transfer followed by protonation furnishes a two-electron reduced, emissive Meisenheimer complex with a
singlet excited state ([1NMI(H)−]*) lifetime of 20 ns. Right: crystal structure of [NMI(H)]−[TBA]+: C, gray; N, blue; O, red; H, white. H atoms
except those on the sp3 carbon of the naphthalene ring are omitted and disorder is removed for clarity. Thermal ellipsoids shown at 50%
probability. (B) Simplified energy level diagram of the NMI system; all potentials given vs Fc/Fc+. Energy levels for each species is aligned with the
compounds in A. SOMO = singly occupied molecular orbital. SUMO = singly unoccupied molecular orbital. Energetics of [NMI(H)]−are assigned
in Figure S24.
reduction of chlorobenzene, the quenching rate constant is
reduced to kq = 107 M−1 s−1.
countercation. A slight twist was evident in the imide ring, with
a dihedral angle of 2.5°. Further examination of the structure
revealed that the two carbon atoms at the 2- and 7-positions of
the reduced naphthalene ring exhibited C−C bond lengths that
were longer than expected for sp2-hybridized carbon atoms. In
addition, the residual electron density around these carbon
atoms suggested the presence of both C(sp3)H2 and C(sp2)H
hydrogen atoms. Accordingly, the structure was best refined
with one sp3 hybridized carbon and one sp2 hybridized carbon
disordered over the 2- and 7-positions. Consistent with the
quinodal structure of a hydride-reduced Meisenheimer
complex, the average carbonyl C−O bond distances were
lengthened by 0.035(6) Å and the average C−C bond
distances between the carbonyl groups and the reduced
naphthalene ring were shortened by 0.048(5) Å with respect to
the independently crystallized neutral NMI (Table S1). These
relative changes in bond lengths are consistent with ab initio
calculations for NMI and ortho-[NMI(H)]− (Figures S16, S17
and Tables S3, S4). Other structures of formal hydride
addition products of NMI were also optimized, including
para-, meta-, and carbo-[NMI(H)]−, which did not match the
crystallographic bond length changes (Figures S18−S20 and
Tables S5−S7)
Isolation of the doubly reduced, emissive NMI species
proved possible via treatment of NMI with NaBH4 in
dimethoxyethane. Anion exchange of the resulting product
with tetrabutylammonium chloride resulted in a bright orange
1
solid. H and Heteronuclear Single Quantum Correlation−
Distortionless Enhancement by Polarization Transfer (HSQC-
DEPT) NMR experiments offered initial evidence of a
hydride-reduced naphthalene ring: the phase of the resonance
near 4.1 ppm in 1H NMR indicates a CH2 group and not a CH
or CH3 group (Figure S11). Experimental NMR spectra were
consistent with ab initio structure and NMR chemical shift
calculations of two NMI hydride addition isomers, ortho- and
para-[NMI(H)]− (Figures S12, S13 and Tables S9, S10 for
1H; Tables S14, S15 for 13C; and Tables S18, S19, Figure S23
for NMR peak assignments). Additional calculations revealed
that the NMR data is inconsistent with two other conceivable
isomers, meta- and carbo-[NMI(H)]− (Tables S11, S12 for 1H;
Tables S16, S17 for 13C). Treatment of NMI with NaBD4,
followed by oxidation in air, leads to a partially deuterated
naphthalene ring, evidenced by reduced integrated intensity in
1H NMR largely at the ortho-, though also at the para-position
(Figure S14).
A cyclic voltammogram of the isolated [NMI(H)]−[TBA]+
reveals a prominent irreversible oxidative wave near −0.7 V
(Figure S6), matching the features observed upon the two-
electron reduction of the neutral NMI. Additionally, using
[NMI(H)]−[TBA]+ as a stoichiometric photoreagent, the
reactivity previously reported to be due to the NMI radical
anion is reproduced.7 Specifically, treatment of [NMI-
(H)]−[TBA]+ with stoichiometric methyl-4-chlorobenzoate
Recrystallization of the NaBH4-reduced NMI via slow
diffusion of diethyl ether into a toluene solution of the
product resulted in orange blade-like crystals suitable for
single-crystal X-ray diffraction analysis (Figure S15). Figure 4A
shows the structure of the product, which corresponds to a
reduced anionic NMI species with a charge-balancing [TBA]+
D
https://doi.org/10.1021/jacs.1c06844
J. Am. Chem. Soc. XXXX, XXX, XXX−XXX

Journal of the American Chemical Society
pubs.acs.org/JACS
Article
in C6D6 resulted in no reaction in the dark (Figure S25). Upon
440 nm irradiation, the reductive homocoupling product
dimethyl-biphenyl-4,4′-dicarboxylate is observed as the major
product, along with a smaller amount of the hydrodehaloge-
nation product methyl benzoate (Figure S25). In the presence
of excess N-methyl pyrrole, the C−C coupling product
resulting from benzoate radical addition to N-methyl pyrrole
is also observed (Figure S26). In the presence of 5 eq
triethylphosphite, the C−P coupling product methyl 4-
(diethoxyphosphoryl)benzoate was obtained as the major
product (Figures S27, S28).
one-electron process. Moreover, the irreversibility of the
second reductive wave is a characteristic of an EC process28
in which NMI radical anion undergoes a facile chemical
reaction following the addition of an electron. The observation
of a remnant amount of radical anion upon the anodic scan
likely results from a limiting proton inventory needed for the
chemical step of the EC process (vide infra). The anodic wave
at Ep,a = −0.65 V (orange trace) is associated with the product
of the EC process of the NMI radical anion. At fast scan rates,
the wave exhibits quasi-reversible behavior (Figure S2),
indicative of a redox process associated with the product of
the two-electron EC process of NMI. These data indicate that
deprotonation of the neutral NMI(H) species is rapid and
occurs within the electrode double layer. Deprotonation results
in the NMI radical anion, which, at −0.65 V, is also rapidly
oxidized to form NMI.
Spectroelectrochemistry of NMI solutions under an applied
voltage of −3.0 V reveals the two-electron species may cleanly
be generated and is stable in the absence of oxygen, as
indicated by the permanence of the absorption spectrum
(Figure 2B). Bulk electrolysis of the NMI solution at −3.0 V
generates an orange solution with a pronounced emission
band, λem,max = 563 nm (Figure 3C). Excitation scans reveal
that the fluorescence tracks the 385 and 490 nm absorption
envelopes of the two-electron reduced NMI species (Figure
3C). We note that the excitation spectrum in Figure 3C is the
same as that recently assigned to an excitation spectrum of a
purported quartet excited state of NMI−•.29 As the excitation
spectrum is not that of the NMI radical anion, our result
establishes that emission originates from the excited state of
the two-electron reduced NMI species and not from either a
spin-allowed or a spin-forbidden excited state of the radical
anion.
The cyclic voltammetry and spectroelectrochemistry results
suggested that the doubly reduced NMI species generated
from the EC process could be chemically generated and
isolated. Treatment of NMI with NaBH4, followed by cation
exchange, allowed for the isolation and crystallographic
characterization of [NMI(H)]−[TBA]+ (Figures 4 and S15).
NMR characterization and comparison to DFT NMR shift
calculations indicated the likely presence of two isomers,
wherein the sp3 carbon on the naphthalene ring is either ortho-
or para- to the imide ring (Figures S11 and S23, Tables S18
and S19). Crystallographic refinement gave bond metrics
consistent with the ortho- isomer (Table S1, Figure S15).
Cyclic voltammetry of the isolated [NMI(H)]−[TBA]+
revealed the same peaks as for neutral NMI, with a prominent
initial oxidative wave at −0.7 V vs Fc/Fc+ (Figure S6). These
results confirm the assignment of the two-electron reduced
NMI observed in electrochemistry as the hydride-reduced
species, [NMI(H)]−.
NMI may also be conveniently reduced by DMAc.
Treatment of rylene imides with fluoride sources in some
polar aprotic solvents was previously reported to result in one
electron reduction, in the dark.34,35 The mechanism of this
process has been debated in the literature. Initial reports
suggested direct fluoride oxidation34,30 later revised by reports
of fluoride-mediated solvent oxidation,35 wherein a proton-
coupled electron transfer (PCET) process involving solvent
deprotonation and formation of bifluoride is concomitant with
electron transfer from solvent to rylene imide. In the case of
rylene bisimides, for which two-electron reductions are fully
reversible, this process was demonstrated to terminate at one
DISCUSSION
■
Figure 4A summarizes the results of NMI redox chemistry and
attendant excited state energetics and dynamics. NMI is
reduced at −1.73 V vs Fc/Fc+ and when the voltammetry scan
range is limited to less than −2.3 V, the one-electron wave is
reversible, indicating the formation of a stable radical anion,
NMI−• (Figure 1, green). Bulk electrolysis of NMI at −2.3 V
vs Fc/Fc+ yields the well-defined and characteristic absorption
spectrum of NMI−• shown in Figure 2A (green trace). The
absorption spectrum of solutions in the dark are indefinitely
stable in the glovebox, allowing for the NMI−• excited state
properties to be examined upon its production by bulk
electrolysis. The absence of emission from solutions of the
radical anion is noteworthy, as the quantum yield for emission
is
Φ = kr/(kr + knr)
(1)
e
where kr is the rate of radiative relaxation and knr is the rate of
nonradiative relaxation. The lack of emission is indicative of
fast nonradiative decay subject to the condition that knr ≫ kr.
This contention is verified by transient absorption spectrosco-
py (Figure 3A), which is able to capture the dynamics of the
radical anion inasmuch as there is no emission signal available
to monitor excited state decay kinetics. The bleaching signal as
a result of ground state depletion (λ = 385−395 nm) recovers
with the same lifetime (24 ps) as the decay of the excited state
absorption signal (λ = 495−525 nm) of the NMI radical anion.
A 24 ps lifetime of the radical anion excited state (Figure 3B) is
too short to support bimolecular chemistry. Assuming an
electron transfer rate constant at the diffusion limit of 1010 M−1
s−1, at the concentrations that substrate has been employed for
photoreactivity (0.1 M), the chromophore will encounter a
substrate molecule on average every 1 ns. Therefore, with a
lifetime of 24 ps, ∼42 lifetimes of the excited state will pass
between chromophore-substrate interactions, resulting in an
excited state concentration of ∼10−19 times the original
concentration of excited state. The passage of so many excited
state lifetimes between chromophore-substrate collisions
would give rise to a vanishingly small quantum yield that is
too low to support observed photoredox conversion yields,
which are achieved in a time scale of hours. In the absence of a
viable photoredox chemistry of the radical anion, we sought to
identify the species responsible for effecting the single-electron
reductive chemistry of haloarenes, as such a photoreagent must
possess potent reducing properties, which could be useful for a
wide range of chemical conversions.
Extending the cathodic window of CV scans to more
negative potential, a second wave is observed beyond the NMI
radical anion (Figure 1, orange). The wave in Figure 1 is
extremely informative. The cathodic current matches that
observed for reduction of NMI to NMI−•, thus indicating a
E
https://doi.org/10.1021/jacs.1c06844
J. Am. Chem. Soc. XXXX, XXX, XXX−XXX

Journal of the American Chemical Society
pubs.acs.org/JACS
Article
electron.35,31 Here, for naphthalene monoimide, fluoride
treatment in DMAc results in the formation of the two-
electron reduced NMI, with absorption and emission spectra
matching those of the doubly reduced species by spectroelec-
trochemistry. The same spectra are obtained for treatment of
NMI with hydride transfer reagents such as NaBH4 (Figure
3D). We envision three plausible mechanisms for the
generation of [NMI(H)]− by fluoride treatment of DMAc:
solvent deprotonation followed by direct hydride transfer from
solvent to NMI, solvent anion σ-addition to NMI followed by
hydride elimination, or fluoride-mediated solvent oxidation to
generate the NMI radical anion, followed by disproportiona-
tion. Other mechanisms involving nucleophilic fluoride
addition to NMI, similar to that proposed in a recent
publication,10 may also be operative.
is the plausible result of radical−radical coupling, which, in the
absence of a transition metal catalyst, normally requires high
radical concentrations.
The use of electrophotochemistry to generate the NMI
radical anion is not selective, as [NMI(H)]− may also be
produced. The onset potential of the EC process is at −2.3 V
vs Fc/Fc+. Consequently, the use of −2.3 V in electro-
photochemistry methods will result in a Nernstian production
of [NMI(H)]−. The issue of nonselective production of the
NMI radical anion is exacerbated by employing constant-
current electrolysis, as has routinely been utilized to date, to
generate the photoactive NMI species because the applied
potential at the working electrode is uncontrolled. In a
constant-current electrolysis, the working electrode is driven to
a potential needed to maintain a constant current and thus will
greatly exceed the onset potential of the EC process of the
radical anion. Indeed, when the applied potential to the
working electrode was controlled (not constant-current but
constant potential) and set to a value slightly beyond the first
reductive wave of NMI, product yields were substantially
attenuated.7 In line with these results, we observe a small
amount of fluorescence from a sample electrolyzed at −2.3 V.
The fluorescence excitation scan is not that of the NMI radical
anion but matches that of the UV−vis spectrum of [NMI-
(H)]− (Figure S7). Thus, under constant current electro-
photochemistry conditions, [NMI(H)]− is invariably formed
and it is this species that is responsible for the observed
photoreactivity.
Highly colored Meisenheimer complexes have a rich
chemistry.26,27 They are key intermediates in SNAr chemistry,32
formed as σ-addition complexes of nucleophiles with arenes,
and hydride Meisenheimer complexes have been implicated in
biological metabolic pathways.33 We demonstrate herein the
utility of Meisenheimer complexes as super-reducing photo-
catalysts. Beyond the hydride σ-complex of NMI reported
here, other nucleophiles could produce similar Meisenheimer
adducts. Fluoride-mediated solvent oxidation may result in
dimethylacetamidyl (from DMAc) or dimsyl (from DMSO)
anion addition to NMI.34,35 Additionally, of possible relevance
to several photochemical cycles invoking two-photon radical
anion photocatalysis, a photon-mediated formation of a
triethylamine (TEA) σ-complex may be envisioned wherein
photooxidation of TEA followed by deprotonation produces a
radical pair consisting of the NMI radical anion and the neutral
TEA radical. A σ-complex formed from this radical pair is
another possible photoactive Meisenheimer adduct. Moreover,
the reduction potential for [NMI(H)]− is 1 V positive of the
reduction potential for the formation of the NMI radical anion,
allowing for the possibility of [NMI(H)]− to function as a
useful two-electron hydride transfer reagent. Accordingly, the
development of Meisenheimer complexes may produce a
palate of useful powerfully reducing one- and two-electron
photoreagents.
Generation of [NMI(H)]− by TBAF in DMAc proved to be
a convenient method to study the emissive lifetime and
dynamic quenching with substrates. The lifetime of the
emissive excited state of [NMI(H)]−, 20 ns, is just long
enough to react in a useful fashion in solution limited by
diffusion. With this lifetime, reactivity requires high substrate
concentrations for efficient quenching. The high excited state
reduction potential of [NMI(H)]− allows it to react with
haloarenes. The two-electron reduced excited state species is
dynamically quenched by aryl chlorides previously used as
substrates for electrophotocatalytic C−C and C−P coupling
reactions. Moreover the fluorescence at 550 nm is quenched
dynamically, indicating that substrate and excited state
photocatalyst directly interact, providing an additional
relaxation pathway for the photocatalyst which reduces its
lifetime. The quenching rate for 4-methylchlorobenzoate, on
the order of 109 M−1 s−1, is near the diffusion limit (Figures 3F
and S8). However, the substantially slower rate for
chlorobenzene, on the order of 107 M−1 s−1 (Figures 3F and
S9), indicates that this substrate is approaching the limit of
redox potential for the excited state.
The reducing power of the [NMI(H)]− excited state,
[NMI(H)]−*, may be determined from a Latimer diagram
(Figure S24C). With properly assigned spectral properties, the
overlap of the absorption and emission profiles of the two-
electron reduced species (Figure S24B) furnishes an E00 value
of 533 nm, establishing an excited state energy of 2.33 eV. The
reduction potential of the two-electron reduced NMI is
approximated by the anodic wave at −0.75 V vs Fc/Fc+ in
cyclic voltammetry (Figure S24A), furnishing a reduction
potential of E([NMI(H)]0/− = −3.08 V vs Fc+/Fc (Figure
S24C). The energetics of the [NMI(H)]− anion are
summarized on the energy level diagram of Figure 4B together
with those of its NMI and NMI radical anion congeners.
As indicated by the Stern−Volmer quenching results and
excited state redox potential, [NMI(H)]−* is super-reducing,
capable of supporting SET to haloarene substrates. Treatment
of [NMI(H)]−[TBA]+ with stoichiometric methyl-4-chloro-
benzoate in C6D6, along with 440 nm light and the radical
traps N-methyl pyrrole or triethylphosphite resulted in C−C
and C−P coupled products (Figures S26−S28) that
recapitulated the photon-mediated reactivity previously
attributed to the NMI radical anion. Additionally, in the
absence of a radical trap, irradiated solutions of methyl-4-
chlorobenzoate and [NMI(H)]−[TBA]+ produced the aryl
homocoupling product dimethyl-biphenyl-4,4′-dicarboxylate,
along with a smaller amount of the hydrodehalogenation
product methyl benzoate (Figure S25). The biphenyl product
The results reported herein for the NMI Meisenheimer
complex may likely be generalized to organic transformations
reported to be driven by other radical anion excited states, as a
preponderance of results demonstrate such doublet excited
states to be too short-lived14−19 to participate in bimolecular
reactions with substrates. For instance, hydrodehalogenation at
potentials more reducing than −2.0 V vs saturated calomel
electrode (SCE), initially proposed to be driven by the
perylene diimide (PDI) radical anion excited state,36 is likely
due to a closed-shell photoactive species that is formed by
F
https://doi.org/10.1021/jacs.1c06844
J. Am. Chem. Soc. XXXX, XXX, XXX−XXX

Journal of the American Chemical Society
pubs.acs.org/JACS
Article
oxidative coupling between PDI and various substrates.37,38
Chloroarene electrophotoreductions have also been reported
to result from a photoactive species that is proposed to be the
radical anion of dicyanoanthracene.9 However, the non-
emissive dicyanoanthracene radical anion is frequently
accompanied by emissive anthrolate or cyanoanthrolate
impurities resulting from a bimolecular oxidation reaction of
the neutral species.12,17,39 Such closed-shell anthrolate anions
are known to be competent photocatalysts for very similar
photoreductions of chloroarenes.40−42 In summary, the
ultrafast relaxation resulting from internal conversion of
doublet excited states to a doublet ground states suggests
that synthetically useful photochemically active radical species
would be extraordinary, as opposed to their reaction to
produce closed shell species in which photoredox processes are
driven from the more common singlet/triplet lowest energy
excited state manifold.
Cambridge Crystallographic Data Centre, 12 Union Road,
Cambridge CB2 1EZ, UK; fax: +44 1223 336033.
AUTHOR INFORMATION
■
Corresponding Author
Daniel G. Nocera − Department of Chemistry and Chemical
Biology, Harvard University, Cambridge, Massachusetts
02138, United States; orcid.org/0000-0001-5055-320X;
Email: dnocera@fas.harvard.edu
Authors
Adam J. Rieth − Department of Chemistry and Chemical
Biology, Harvard University, Cambridge, Massachusetts
02138, United States; orcid.org/0000-0002-9890-1346
Miguel I. Gonzalez − Department of Chemistry and Chemical
Biology, Harvard University, Cambridge, Massachusetts
02138, United States; orcid.org/0000-0003-4250-9035
Bryan Kudisch − Department of Chemistry and Chemical
Biology, Harvard University, Cambridge, Massachusetts
02138, United States; orcid.org/0000-0003-3352-5383
Matthew Nava − Department of Chemistry and Chemical
Biology, Harvard University, Cambridge, Massachusetts
02138, United States
CONCLUSIONS
■
Spectro/electrochemical data combined with ultrafast transient
absorption and emission spectroscopies demonstrates that the
photoactive NMI species is a two-electron reduced singlet
excited state [NMI(H)]−* and not the excited state of the
radical anion NMI−•. The closed-shell electron rich photo-
catalyst results from two sequential one electron reductions of
NMI, followed by a chemical step, which furnish a hydride.
Whereas the short lifetime of 24 ps (Figure 3A,B) of the radical
anion doublet excited state precludes intermolecular electron
transfer, as imposed by diffusional limitations, the two-electron
closed shell singlet excited state [NMI(H)]−* possesses a
lifetime of 20 ns, which is typical of closed shell singlet excited
states. This sufficiently long-lived excited state together with an
exceptional excited state reduction potential bestows [NMI-
(H)]− with the properties to react with aryl halide substrates
and effect C−C and C−P bond coupling, initiated by single
electron transfer.
Complete contact information is available at:
https://pubs.acs.org/10.1021/jacs.1c06844
Notes
The authors declare no competing financial interest.
ACKNOWLEDGMENTS
■
This work was supported by the National Science Foundation
under grant CHE-1855531. M.I.G. acknowledges the Arnold
and Mabel Beckman Foundation for an Arnold O. Beckman
Postdoctoral Fellowship. NSF’s ChemMatCARS Sector 15 is
supported by the Divisions of Chemistry and Materials
Research, National Science Foundation, under grant number
CHE-1834750. Use of the Advanced Photon Source, an Office
of Science User Facility operated for the U.S. Department of
Energy (DOE) Office of Science by Argonne National
Laboratory, was supported by the DOE under Contract No.
DE-AC02-06CH11357.
There are few excited state species, to date, with super-
reducing properties and sufficiently long lifetimes to drive
photoredox events. To this end, [NMI(H)]− should find utility
as a reagent for a variety of photochemical transformations that
require the activation of strong bonds. Finally, the work
described herein highlights that doublet excited states are
generally too short-lived to support photoredox chemistry
driven by electronically excited radical anions. We suspect that
for systems in which radical anions have been proposed to
drive photoredox chemistry, it may be prudent to reconsider
the possibility of a more stable, closed shell, super-reducing
species. Such re-evaluation will be useful for the future
elaboration of super-reducing excited state species under-
pinning new synthetic methodologies.
REFERENCES
■
(1) Shaw, M. H.; Twilton, J.; MacMillan, D. W. C. Photoredox
Catalysis in Organic Chemistry. J. Org. Chem. 2016, 81, 6898−6926.
(2) Romero, N. A.; Nicewicz, D. A. Organic Photoredox Catalysis.
Chem. Rev. 2016, 116, 10075−10166.
(3) Huang, H.; Lambert, T. H. Electrophotocatalytic SNAr
Reactions of Unactivated Aryl Fluorides at Ambient Temperature
and Without Base. Angew. Chem., Int. Ed. 2020, 59, 658−662.
(4) Huang, H.; Strater, Z. M.; Rauch, M.; Shee, J.; Sisto, T. J.;
Nuckolls, C.; Lambert, T. H. Electrophotocatalysis with a
Trisaminocyclopropenium Radical Dication. Angew. Chem., Int. Ed.
2019, 58, 13318−13322.
(5) Shen, T.; Lambert, T. H. Electrophotocatalytic Diamination of
Vicinal C−H Bonds. Science 2021, 371, 620−626.
(6) Huang, H.; Strater, Z. M.; Lambert, T. H. Electrophotocatalytic
C−H Functionalization of Ethers with High Regioselectivity. J. Am.
Chem. Soc. 2020, 142, 1698−1703.
ASSOCIATED CONTENT
* Supporting Information
The Supporting Information is available free of charge at
https://pubs.acs.org/doi/10.1021/jacs.1c06844.
■
sı
Experimental procedures, synthetic methods, character-
ization, crystallographic data (PDF)
Accession Codes
CCDC 2092941−2092942 contain the supplementary crys-
tallographic data for this paper. These data can be obtained
free of charge via www.ccdc.cam.ac.uk/data_request/cif, or by
emailing data_request@ccdc.cam.ac.uk, or by contacting The
(7) Cowper, N. G. W.; Chernowsky, C. P.; Williams, O. P.; Wickens,
Z. K. Potent Reductants via Electron-Primed Photoredox Catalysis:
Unlocking Aryl Chlorides for Radical Coupling. J. Am. Chem. Soc.
2020, 142, 2093−2099.
G
https://doi.org/10.1021/jacs.1c06844
J. Am. Chem. Soc. XXXX, XXX, XXX−XXX

Journal of the American Chemical Society
pubs.acs.org/JACS
Article
(8) MacKenzie, I. A.; Wang, L.; Onuska, N. P. R.; Williams, O. F.;
Begam, K.; Moran, A. M.; Dunietz, B. D.; Nicewicz, D. A. Discovery
and Characterization of an Acridine Radical Photoreductant. Nature
2020, 580, 76−80.
(9) Kim, H.; Kim, H.; Lambert, T. H.; Lin, S. Reductive
Electrophotocatalysis: Merging Electricity and Light to Achieve
Extreme Reduction Potentials. J. Am. Chem. Soc. 2020, 142, 2087−
2092.
(28) Savéant, J.-M.; Costentin, C. Elements of Molecular and
Biomolecular Electrochemistry. 2nd Ed. Wiley: Hoboken, NJ, 2019.
(29) Tian, X.; Karl, T. A.; Reiter, S.; Yakubov, S.; de Vivie-Riedle, R.;
König, B.; Barham, J. P. Electro-Mediated PhotoRedox Catalysis for
Selective C(sp3)−O Cleavages of Phosphinated Alcohols to
Carbanions. Angew. Chem., Int. Ed. 2021, DOI: 10.1002/
anie.202105895.
(30) Guha, S.; Saha, S. Fluoride Ion Sensing by an Anion-π
Interaction. J. Am. Chem. Soc. 2010, 132, 17674−17677.
(31) Wentz, H. C.; Skorupskii, G.; Bonfim, A. B.; Mancuso, J. L.;
Hendon, C. H.; Oriel, E. H.; Sazama, G. T.; Campbell, M. G.
Switchable Electrical Conductivity in a Three-Dimensional Metal-
Organic Framework: Via Reversible Ligand n-Doping. Chem. Sci.
2020, 11, 1342−1346.
(10) Cole, J. P.; Chen, D.; Kudisch, M.; Pearson, R. M.; Lim, C.-H.;
Miyake, G. M. Organocatalyzed Birch Reduction Driven by Visible
Light. J. Am. Chem. Soc. 2020, 142, 13573−13581.
(11) Costentin, C.; Fortage, J.; Collomb, M.-N. Electrophotocatal-
ysis: Cyclic Voltammetry as an Analytical Tool. J. Phys. Chem. Lett.
2020, 11, 6097−6104.
(32) Neumann, C. N.; Hooker, J. M.; Ritter, T. Concerted
Nucleophilic Aromatic Substitution with 19F− and 18F−. Nature
2016, 534, 369−373.
(12) Fox, M. A. The Photoexcited States of Organic Anions. Chem.
Rev. 1979, 79, 253−273.
(13) Shukla, S. S.; Rusling, J. F. Photoelectrocatalytic Reduction of
4-Chlorobiphenyl Using Anion Radicals and Visible Light. J. Phys.
Chem. 1985, 89, 3353−3358.
(33) Lenke, H.; Knackmuss, H. J. Initial Hydrogenation during
Catabolism of Picric Acid by Rhodococcus Erythropolis HL 24−2.
Appl. Environ. Microbiol. 1992, 58, 2933−2937.
(14) Fujitsuka, M.; Kim, S. S.; Lu, C.; Tojo, S.; Majima, T.
Intermolecular and Intramolecular Electron Transfer Processes from
Excited Naphthalene Diimide Radical Anions. J. Phys. Chem. B 2015,
119, 7275−7282.
(34) Saha, S. Anion-Induced Electron Transfer. Acc. Chem. Res.
2018, 51, 2225−2236.
(35) Bélanger-Chabot, G.; Ali, A.; Gabbaï, F. P. On the Reaction of
Naphthalene Diimides with Fluoride Ions: Acid/Base versus Redox
Reactions. Angew. Chem., Int. Ed. 2017, 56, 9958−9961.
(36) Ghosh, I.; Ghosh, T.; Bardagi, J. I.; Konig, B. Reduction of Aryl
Halides by Consecutive Visible Light-Induced Electron Transfer
Processes. Science 2014, 346, 725−728.
(15) Fujitsuka, M.; Majima, T. Reaction Dynamics of Excited
Radical Ions Revealed by Femtosecond Laser Flash Photolysis. J.
Photochem. Photobiol., C 2018, 35, 25−37.
(16) Fujita, M.; Ishida, A.; Majima, T.; Takamuku, S. Lifetimes of
Radical Anions of Dicyanoanthracene, Phenazine, and Anthraquinone
in the Excited State from the Selective Electron-Transfer Quenching.
J. Phys. Chem. 1996, 100, 5382−5387.
(17) Breslin, D. T.; Fox, M. A. Excited-State Behavior of Thermally
Stable Radical Ions. J. Phys. Chem. 1994, 98, 408−411.
(18) Gosztola, D.; Niemczyk, M. P.; Svec, W.; Lukas, A. S.;
Wasielewski, M. R. Excited Doublet States of Electrochemically
Generated Aromatic Imide and Diimide Radical Anions. J. Phys.
Chem. A 2000, 104, 6545−6551.
(19) Gumy, J. C.; Vauthey, E. Investigation of the Excited-State
Dynamics of Radical Ions in the Condensed Phase Using the
Picosecond Transient Grating Technique. J. Phys. Chem. A 1997, 101,
8575−8580.
(20) Zinchenko, K. S.; Ardana-Lamas, F.; Seidu, I.; Neville, S. P.;
Van Der Veen, J.; Lanfaloni, V. U.; Schuurman, M. S.; Wörner, H. J.
Sub-7-fs Conical-Intersection Dynamics Probed at the Carbon K-
Edge. Science 2021, 371, 489−494.
(21) Martinez, J. F.; La Porte, N. T.; Wasielewski, M. R. Electron
Transfer from Photoexcited Naphthalene Diimide Radical Anion to
Electrocatalytically Active Re(bpy)(CO)3Cl in a Molecular Triad. J.
Phys. Chem. C 2018, 122, 2608−2617.
(22) Christensen, J. A.; Phelan, B. T.; Chaudhuri, S.; Acharya, A.;
Batista, V. S.; Wasielewski, M. R. Phenothiazine Radical Cation
Excited States as Super-Oxidants for Energy-Demanding Reactions. J.
Am. Chem. Soc. 2018, 140, 5290−5299.
(37) Marchini, M.; Gualandi, A.; Mengozzi, L.; Franchi, P.; Lucarini,
M.; Cozzi, P. G.; Balzani, V.; Ceroni, P. Mechanistic Insights into
Two-Photon-Driven Photocatalysis in Organic Synthesis. Phys. Chem.
Chem. Phys. 2018, 20, 8071−8076.
(38) Zeman, C. J.; Kim, S.; Zhang, F.; Schanze, K. S. Direct
Observation of the Reduction of Aryl Halides by a Photoexcited
Perylene Diimide Radical Anion. J. Am. Chem. Soc. 2020, 142, 2204−
2207.
(39) Janzen, E. G.; Rudy, B. C.; Lopp, I. G.; Happ, J. W.
Chemiluminescent Oxygenation of 9,10-Dicyanoanthracene Radical
Anion. J. Chem. Soc. D 1970, 491−492.
(40) Schmalzbauer, M.; Marcon, M.; König, B. Excited State Anions
in Organic Transformations. Angew. Chem., Int. Ed. 2021, 60, 6270−
6292.
(41) Schmalzbauer, M.; Svejstrup, T. D.; Fricke, F.; Brandt, P.;
Johansson, M. J.; Bergonzini, G.; König, B. Redox-Neutral Photo-
catalytic C−H Carboxylation of Arenes and Styrenes with CO2. Chem.
2020, 6, 2658−2672.
(42) Schmalzbauer, M.; Ghosh, I.; König, B. Utilising Excited State
Organic Anions for Photoredox Catalysis: Activation of (Hetero)Aryl
Chlorides by Visible Light-Absorbing 9-Anthrolate Anions. Faraday
Discuss. 2019, 215, 364−378.
(23) Eriksen, J.; Joergensen, K. A.; Linderberg, J.; Lund, H. Electron-
Transfer Fluorescence Quenching of Radical Ions. Experimental Work
and Theoretical Calculations. J. Am. Chem. Soc. 1984, 106, 5083−
5087.
(24) Eriksen, J.; Lund, H.; Nyvad, A. I. Electron-Transfer
Fluorescence Quenching of Radical Ions. J. Am. Chem. Soc. 1984,
37b, 459−466.
(25) Chakraborty, B.; Menezes, P. W.; Driess, M. Beyond CO2
Reduction: Vistas on Electrochemical Reduction of Heavy Non-Metal
Oxides with Very Strong E−O Bonds (E = Si, P, S). J. Am. Chem. Soc.
2020, 142, 14772−14788.
(26) Meisenheimer, J. Ueber Reactionen Aromatischer Nitrokorper.
Justus Liebig 1902, 323, 205−242.
(27) Jackson, C. J.; Gazzolo, F. H. On Certain Colored Substances
Derived from Nitro Compounds. Third Paper. Proc. Am. Acad. Arts
Sci. 1900, 35, 263−281.
H
https://doi.org/10.1021/jacs.1c06844
J. Am. Chem. Soc. XXXX, XXX, XXX−XXX