Photoredox catalysis on unactivated substrates with strongly reducing iridium photosensitizers

Article abstract of DOI:10.1039/d0sc06306a

Photoredox catalysis has emerged as a powerful strategy in synthetic organic chemistry, but substrates that are difficult to reduce either require complex reaction conditions or are not amenable at all to photoredox transformations. In this work, we show that strong bis-cyclometalated iridium photoreductants with electron-rich β-diketiminate (NacNac) ancillary ligands enable high-yielding photoredox transformations of challenging substrates with very simple reaction conditions that require only a single sacrificial reagent. Using blue or green visible-light activation we demonstrate a variety of reactions, which include hydrodehalogenation, cyclization, intramolecular radical addition, and prenylationviaradical-mediated pathways, with optimized conditions that only require the photocatalyst and a sacrificial reductant/hydrogen atom donor. Many of these reactions involve organobromide and organochloride substrates which in the past have had limited utility in photoredox catalysis. This work paves the way for the continued expansion of the substrate scope in photoredox catalysis.

Full text of DOI:10.1039/d0sc06306a

Chemical
Science
View Article Online
View Journal | View Issue
EDGE ARTICLE
Photoredox catalysis on unactivated substrates
with strongly reducing iridium photosensitizers†
Cite this: Chem. Sci., 2021, 12, 4069
a
All publication charges for this article
have been paid for by the Royal Society
of Chemistry
Jong-Hwa Shon,
Jimmie Weaver
Dooyoung Kim,a Manjula D. Rathnayake,b Steven Sittel,a
b
a
*
and Thomas S. Teets
Photoredox catalysis has emerged as a powerful strategy in synthetic organic chemistry, but substrates that
are difficult to reduce either require complex reaction conditions or are not amenable at all to photoredox
transformations. In this work, we show that strong bis-cyclometalated iridium photoreductants with
electron-rich b-diketiminate (NacNac) ancillary ligands enable high-yielding photoredox transformations
of challenging substrates with very simple reaction conditions that require only a single sacrificial
reagent. Using blue or green visible-light activation we demonstrate a variety of reactions, which include
hydrodehalogenation, cyclization, intramolecular radical addition, and prenylation via radical-mediated
pathways, with optimized conditions that only require the photocatalyst and a sacrificial reductant/
hydrogen atom donor. Many of these reactions involve organobromide and organochloride substrates
which in the past have had limited utility in photoredox catalysis. This work paves the way for the
continued expansion of the substrate scope in photoredox catalysis.
Received 17th November 2020
Accepted 28th January 2021
DOI: 10.1039/d0sc06306a
rsc.li/chemical-science
photosensitizer which has become ubiquitous in photoredox
catalysis, fac-Ir(ppy)3 (ppy ¼ 2-phenylpyridine),9 demonstrating
Introduction
hydrodeiodination of unactivated aryl, alkyl, and vinyl iodide
substrates.10 This photosensitizer has an excited-state reduction
potential almost 1 V more negative than the common ruthe-
nium polypyridyl photosensitizers, and has become a very
popular choice for photoredox methodology involving less
activated substrates.10,11 It is also possible to activate the
same substrate classes for cyclization reactions, using
[Ir(ppy)2(dtbbpy)]+ (dtbbpy ¼ 4,40-di-tert-butyl-2,20-bipyridine)
as the photosensitizer.12
Despite the recent successes expanding photoredox chem-
istry to a variety of unactivated organoiodide substrates, trans-
formations involving more economical and widely available
bromide and chloride substrates remain rare and have only
been developed very recently. There are several photoredox
transformations on unactivated organobromide and organo-
chloride substrates that require UV excitation.13–17 In addition,
recent work from Jui's group demonstrated intramolecular
hydroarylation reactions involving aryl bromide substrates,
Visible-light photoredox catalysis has emerged as a powerful,
versatile, and increasingly important methodology in organic
synthesis.1 In this strategy, a photosensitizer, usually a chro-
mophoric organic compound or transition-metal coordination
complex, absorbs visible light and initiates charge transfer to
the substrate, either directly, via a redox mediator, or with the
assistance of a co-catalyst. Such an approach had already
become mainstream in the area of solar fuels,2,3 before being
adopted by the small-molecule organic synthesis community
and even more recently becoming widespread in polymer
synthesis.4
One of the most signicant remaining challenges in photo-
redox catalysis is to extend the substrate scope to molecules that
are difficult to reduce or oxidize. In photoredox transformations
initiated by electron transfer, highly activated substrates, like a-
carbonyl halides, benzyl halides, sulfones, and sulfoniums are
commonly used.5–8 These substrate classes can be readily acti-
vated by homoleptic ruthenium photosensitizers of the general
formula [Ru(N^N)3]2+, where N^N is 2,20-bipyridine or a related
derivative, as well as certain organic photosensitizers. To
expand the substrate scope to molecules that are more difficult
to reduce, Stephenson's group turned to another
¨
using visible light, an organic photosensitizer, and Hunig's base
as the only reductant.18 However, the vast majority of visible-
light photoredox transformations on unactivated organo-
chloride or organobromide substrates require either additional
energy input in the form of an applied potential or a second
photon, or the use of additional costly additives beyond the
sacricial reductant. A selection of these recent advances in
visible-light photoredox activation of challenging organohalide
substrates is summarized in Fig. 1. MacMillan's and Ste-
phenson's groups both revealed strategies for photoredox
aDepartment of Chemistry, University of Houston, 3585 Cullen Blvd., Room 112,
Houston, TX 77204-5003, USA. E-mail: tteets@uh.edu
bDepartment of Chemistry, Oklahoma State University, 107, Physical Science,
Stillwater, OK 74078, USA
† Electronic supplementary information (ESI) available: Experimental details, GC
traces, and NMR spectra for catalysis studies. See DOI: 10.1039/d0sc06306a
© 2021 The Author(s). Published by the Royal Society of Chemistry
Chem. Sci., 2021, 12, 4069–4078 | 4069

View Article Online
Chemical Science
Edge Article
Fig. 1 Previously reported methods for photoredox activation of challenging substrates and an introduction to this work.
activation of aryl and alkyl bromide substrates which involve challenging substrates. There have been some promising efforts
a
superstoichiometric amount of tris(trimethylsilyl)silane with low-valent group 6 (Cr, Mo, W) isocyanide complexes with
(TTMSS) as an additive.19,20 Another recent approach for pho- very appealing photophysical and redox attributes,27,28 and some
toredox activation of challenging substrates is to use two- early success in photoredox transformations of difficult
21
substrates.28,29 The highly reducing organic photosensitizer 10-
¨
photon activation. Konig's work with perylene diimide (PDI)
and Nicewicz's work with acridinium photosensitizers22 have phenylphenothiazine (PTH) has also been shown to be capable of
shown that certain organic photosensitizers can be reductively promoting hydrodehalogenation of a wide range of aryl halide
quenched to a radical state, which upon second excitation substrates, albeit with UV activation.30 Our group has introduced
produces a strong reductant capable of reacting with a wide a new class of heteroleptic bis-cyclometalated iridium photo-
variety of aryl halide substrates. In another approach which sensitizers with the general formula Ir(ppy)2(NacNac), where
requires two-photon excitation, high-power laser irradiation of NacNac is a variably substituted b-diketiminate ligand, which
a water-soluble fac-Ir(ppy)3 analogue produces hydrated elec- have excited-state reduction potentials more potent than fac-
trons which can reduce aryl or benzyl halides in aqueous solu- Ir(ppy)3, by ꢀ300–500 mV.31 With some of these photosensitizers
tion.23,24 And nally, in two works published concurrently, Lin's we have shown very efficient hydrodebromination of a few aryl
and Wickens's groups developed strategies that combine elec- bromide substrates, using only blue LED irradiation and an
trochemical and photochemical activation of organic photo- amine base in concert with the photosensitizer.32 Relatedly,
sensitizers, generating reductants strong enough to react with Connell et al. have shown that the well-known photosensitizer
aryl bromide and aryl chloride compounds.25,26
[Ir(ppy)2(dtbbpy)]+ is converted under photoredox reaction
All of these aforementioned visible-light approaches, while conditions to a more reducing charge–neutral complex, via
effective, require some combination of technically challenging semireduction of dtbbpy, and that this modied complex is
reaction conditions, atom-inefficient silane additives, or high- a much more potent photoreductant and promotes photoredox
powered laser light sources. An approach that has been much transformations of a variety of aryl halide and some alkyl
less explored until recently is to tailor new, more reactive bromide substrates.33
photosensitizers capable of reacting directly with more
4070 | Chem. Sci., 2021, 12, 4069–4078
© 2021 The Author(s). Published by the Royal Society of Chemistry

View Article Online
Edge Article
Chemical Science
In this work, we show that the strongly photoreducing
Ir(ppy)2(NacNac) complexes previously developed by our group
(Fig. 1) are versatile and efficient photoredox catalysts. Using
simple reaction conditions with only a single additive needed,
we demonstrate high-yielding hydrodehalogenation trans-
formations of aryl bromide, aryl chloride, aryl uoride, and
alkyl bromide substrates. In addition to hydrodehalogenation,
we also show that these newly developed iridium photocatalysts
are effective for other synthetically useful functional group
transformations, including ether C–O cleavage, radical addition
and cyclization, arylation, prenylation, and ketone reduction.
We show with one of the catalysts that green light irradiation
results in good product yields for most of these trans-
formations, which may be useful for more advanced applica-
tions involving wavelength-control of dual-catalyst systems.34
This work shows that strongly reducing photosensitizers can
enable reactions on a wider range of substrates to include those
not easily amenable to photoredox catalysis, using simple,
generalizable reaction conditions for many different
transformations.
Results and discussion
Choice of photosensitizer and sacricial reagent
To comment briey on our choice of photosensitizers, we chose
complexes Ir1 and Ir2 (Fig. 1) for this study, noting that they
performed best of the 10 photosensitizers we previously evalu-
ated for aryl bromide hydrodebromination.32 These catalysts are
synthesized by a general and simple route involving the readily
accessible precursor [Ir(ppy)2(m-Cl)]2 and the potassium salt of
the respective b-diketiminate. Whereas stock solutions of Ir1
and Ir2 slowly degrade under ambient conditions, and reactions
described here were typically set up in a glovebox, if necessary
solids of Ir1 and Ir2 can be handled on the benchtop and
reactions can be set up outside of a glovebox, as we have
previously shown.32 Ir1 and Ir2 are soluble and stable in most
aromatic and polar organic solvents and have signicant visible
absorption beyond 400 nm (Fig. S1 of the ESI†), making them
suitable for visible-light photoredox catalysis. The catalysts and
their degradation products can easily be separated from organic
reaction products by ltering through a short plug of silica. As
shown in Fig. 2, the ground- and excited-state redox potentials
of Ir1 and Ir2 are markedly different than the iridium photo-
sensitizers fac-Ir(ppy)3 and [Ir(ppy)2(dtbbpy)]+, both commonly
used in photoredox activation of aryl halide substrates.32,35 The
biggest effect of the b-diketiminate ancillary ligand is to desta-
bilize the HOMO, making Ir1 and Ir2 both much easier to
oxidize in the ground-state. Their formal IrIV/IrIII redox couples
occur below the ferrocene couple, which in combination with
their triplet excited-state energies (ET1) leads to the prediction
that Ir1 and Ir2 are much stronger excited-state reductants than
fac-Ir(ppy)3 and [Ir(ppy)2(dtbbpy)]+, with E(IrIV/*IrIII) ¼ ꢁ2.6 V
and ꢁ2.4 V for Ir1 and Ir2, respectively. We have previously
shown that these more potent excited-state reduction potentials
correlate with faster photoinduced electron-transfer to model
substrates.31,32 On the other hand, Ir1 and Ir2 are expected to be
much weaker excited-state oxidants than the two reference
Fig. 2 Comparison of ground- and excited-state redox potentials for
widely used Ir photosensitizers with Ir1 and Ir2.
compounds. In short, we expect Ir1 and Ir2 to be thermody-
namically and kinetically superior for photoredox trans-
formations that involve excited-state reduction steps, which are
the types of reactions we studied in this work.
Our previous conditions for hydrodebromination used tet-
ramethylethylenediamine (TMEDA) as the sacricial reagent. In
hopes of further optimizing the reaction conditions and being
able to access even more challenging substrates, we evaluated
our choice of sacricial reagent. The role of the sacricial
reductant depends on the reaction mechanism, and either
involves reductive quenching of the photosensitizer excited
state, or reduction of the oxidized photosensitizer to regenerate
the photosensitizer ground state if oxidative quenching is
operative. Whichever mechanism is occurring, the reaction
could benet from a more strongly reducing sacricial reagent.
Complexes Ir1 and Ir2 are difficult to reduce both in the excited
state E(*IrIII/IrII < 0 V) and in the oxidized ground state, with
E(IrIV/IrIII) respectively ꢁ0.26 and ꢁ0.39 V (all potentials are
referenced to the ferrocenium/ferrocene couple). The oxidation
potentials of tertiary amine bases fall in the range of 0.17–
0.52 V, and Pellegrin and Odobel have suggested a list of
alternative sacricial reagents for photocatalysis, many of
which are more reducing.36 Aer some preliminary screening of
strongly reducing thiolate or dithiocarbamate salts which did
not lead to quenching or to particularly robust catalysis, we
settled on 1,3-dimethyl-2,3-dihydro-2-phenyl-benzimidazole
© 2021 The Author(s). Published by the Royal Society of Chemistry
Chem. Sci., 2021, 12, 4069–4078 | 4071

View Article Online
Chemical Science
Edge Article
(BIH, Eox ¼ ꢁ0.07 V). BIH is known to be a good electron and Ir2 with green LED irradiation (520–525 nm; see Fig. S2† for
hydrogen atom donor, and has been adopted in multiple pictures of reaction apparatuses), we observe comparable
studies including both photoredox catalysis37,38 and photo- product yields across the range of substrates tested. Control
catalytic CO2 reduction.39,40
experiments
with
the
substrates
4-tri-
uoromethylbromobenzene (S2) and 3-methoxybromobenzene
(S3), using identical conditions but without the Ir catalyst
added, did not lead to the formation of the respective arene
products P2 and P3, indicating that BIH is not a catalyst on its
own. One possible complication with BIH or even amine-based
sacricial reagent in photoredox catalysis is that the radical
cation state (BIHc+ or NR3c+) can participate in a dark radical
chain reaction41 with the substrate, following an initiation step
involving the photosensitizer and light. To check the possibility
of radical chain reactions, we have measured the photochem-
ical quantum yield (F) for the formation of product from
hydrodebromination of S2 and S7. The measured quantum
yields are 4.7% for the conversion of S2 to P2, and 1.4% for the
conversion of S7 to P7. Whereas a quantum yield >100% must
indicate dark radical chain processes and the quantum yields
we obtained do not explicitly rule out chain reactions following
a slow initiation step, the results are in line with other recently
reported quantum yields for photocatalytic transformations
and suggest that radical chain reactions are not a dominant
pathway.37,42,43 One other experiment in support of this
conclusion is summarized in Fig. S3 of the ESI,† where we
monitored the hydrodebromination of 2-bromomesitylene with
alternating light/dark cycles. As shown in Fig. S3,† hydro-
debromination progress ceases in the absence of light, indi-
cating that if there are radical chain processes, they are short-
lived and not persistent over long timescales.
Hydrodebromination of aryl bromides
Replacing TMEDA with BIH as the sacricial reagent allows
further optimization of the conditions for aryl bromide hydro-
debromination and an expansion of the substrate scope,
summarized in Scheme 1. In our previous work where TMEDA
was used as the sacricial reagent, the reactions required
2.5 mol% Ir catalyst and 48 h reaction times for optimal yields at
45 ꢂC, but with BIH we can decrease catalyst loading to 1 mol%,
decrease reaction time to 24 h, and still achieve similar or even
better product yields. Eight substrates were tested, and in all
cases hydrodebromination occurs with moderate to excellent
yields. The reaction tolerates electron-poor, electron-rich, and
sterically encumbered aryl bromide substrates. Comparing
catalyst Ir1 with blue LED irradiation (465–470 nm) and catalyst
Hydrodechlorination of aryl chlorides
Using the same conditions optimized for hydrodebromination
of aryl bromides, we moved on to more challenging aryl chloride
substrates. We did achieve successful hydrodechlorination
reactions of aryl chlorides, also summarized in Scheme 1, which
is rare in visible-light photoredox catalysis.23,25,26,33 These reac-
tions required higher catalyst loading (2.5 mol%) and a larger
excess of BIH (3 equiv.) for optimization, but otherwise condi-
tions and reaction times are the same as those in Scheme 1 for
hydrodebromination. One limitation is that the substrate for
hydrodechlorination must have one or more electron-
withdrawing groups to be activated, suggesting that the
electron-transfer step is limiting with these substrates. Never-
theless, with optimized conditions we were able to achieve high-
yielding hydrodechlorination of four aryl chloride substrates,
with quantitative yields when using catalyst Ir1 and blue LED
irradiation, and slightly lower yields with catalyst Ir2 and green
LED irradiation.
Hydrodeuorination
Recently photoredox catalysis has been shown to be a promising
strategy to break even strong and recalcitrant aryl C–F bonds
and is limited by the necessary deep reduction. There has been
a surge of recent interest in photoredox catalysis on aryl uo-
ride44–48 or triuoromethylated substrates,45,49 which provides
Scheme 1 Hydrodehalogenation of aryl halides.
4072 | Chem. Sci., 2021, 12, 4069–4078
© 2021 The Author(s). Published by the Royal Society of Chemistry

View Article Online
Edge Article
Chemical Science
convenient access to pharmaceutically relevant partially uori-
nated aromatics. Whereas all recent developments in extreme
photoredox reductions have focused on organobromide and
organochloride substrates,20,21,23,25,26,29,33 we are also interested
in the efficacy of catalysts Ir1 and Ir2 for hydrodeuorination.
Using similar conditions to those optimized by Weaver's group,
we demonstrated high-yielding para-selective hydro-
deuorination (HDF) of the substrate S13, forming product P13
in good yield using either Ir1 or Ir2 as the catalyst with blue LED
illumination (Scheme 1). The regioselectivity of the HDF
remains the same, and with only half the catalyst loading the
yield is slightly higher than that obtained with fac-Ir(ppy)3 as the
catalyst.44 As part of our optimization of the HDF reaction we
screened several other catalyst variants (Ir3–Ir8, see Table S1†),
with the HDF results summarized in Table S2.† A few other
catalysts, Ir3–Ir6, performed as well as Ir1 and Ir2, but none
were clearly superior for HDF.
Hydrodebromination of alkyl bromides
Hydrodebromination of alkyl bromides remains rare in photo-
redox catalysis, with the most systematic study from Ste-
phenson's group using reaction conditions that include excess
TTMSS as an additive, in addition to excess DIPEA,20 and Con-
nell et al. recently showed that TTMSS additive is not required
for two alkyl bromide substrates.33 As shown in Scheme 2, the
newly optimized hydrodehalogenation conditions presented
here are also operable for unactivated alkyl bromide substrates.
The reactions require 2.5 mol% of Ir catalyst with 3 equiv. of
BIH, avoiding wasteful silane additives that were previously
used with alkyl bromide substrates.19,20
Scheme 2 Hydrodebromination and intramolecular radical addition
The reactions tolerate primary (S14–S17), secondary (S18),
and tertiary alkyl bromide (S19) substrates. The primary alkyl
bromide substrates tested all have tethered aryl rings, and three
involved side reactions and thus did not exclusively produce the
straight-chain alkane product. For substrate S14 HBr elimina-
tion competed with hydrodebromination, resulting in a 1 : 1
mixture of the alkane (P14a) and substituted styrene (P14b)
products. As shown in Fig. S4,† replacing BIH with DIPEA or
TMEDA improves the selectivity for P14a, albeit at the expense
of lower conversion. For substrates S15 (alkane) and S16 (ether),
intramolecular radical addition can occur following C–Br
cleavage, resulting in cyclized side products. Similar intra-
of unactivated alkyl bromides.
the linear product P17 is observed in good yield (Fig. S8†). For
the secondary (S18) and tertiary (S19) alkyl bromide substrates,
we observe hydrodebromination exclusively. Unlike Ir1 and Ir2,
fac-Ir(ppy)3 does not promote effective hydrodebromination of
1-bromoadamantane (S19), with majority starting material
remaining aer 24 h (Fig. S9†).
Radical cyclization
molecular radical addition reactions to an aryl group have been Having demonstrated efficient hydrodebromination of alkyl
studied by Stephenson et al.50 and Zhang et al.51 For S15 nearly and aryl bromides (Schemes 1 and 2), and observing cyclization
equal amounts of the linear (P15a) and cyclized (P15b) products side products in some of the cases (Scheme 2), we next turned
are formed (Fig. S5†), with the ratio depending slightly on the our attention to substrates with tethered allyl substituents,
catalyst choice and excitation wavelength. Efforts to optimize which are poised for radical cyclization following C–Br bond
the ratio between P15a and P15b did not result in further cleavage. Scheme 3 summarizes the results with three such
improvement (Fig. S6 and S7†). Reactions are more sluggish in substrates, where the cyclization is initiated either by cleavage
MeCN compared to DMF (Fig. S7†) and replacing BIH with of an aryl C–Br bond (S20) or an alkyl C–Br bond (S21 and S22).
TMEDA does not improve conversion or yield (Fig. S6†). For S16, Cyclization of these substrates occurs under the same condi-
the linear hydrodebromination product P16a was preferred, tions as previously optimized for hydrodebromination, and
forming in a 6 : 1 to 7 : 1 ratio with the cyclic ether product P16b forms substituted heterocyclic indoline (P20) or pyrrolidine
(Fig. S8†). For the rest of the reactions, hydrodebromination is (P21 and P22) products. In all cases, moderate yields of the
the only outcome observed. Substrate S17 does not undergo cyclized products are obtained, with no major differences
radical addition to form a fused ve-membered ring, and only between catalyst Ir1(blue LED) or Ir2 (green LED). For the
© 2021 The Author(s). Published by the Royal Society of Chemistry
Chem. Sci., 2021, 12, 4069–4078 | 4073

View Article Online
Chemical Science
Edge Article
the same transformation using the alkyl bromide substrates S18
and S19, but with N-methylpyrrole present, we still only
observed hydrodebromination products.
We also investigated prenylation reactions52 of the aryl
uoride substrate S13 (Scheme 3), intercepting the radical with
6 equiv. of a secondary prenylation reagent. These reactions
were screened with eight different catalysts from our library
(Table S3†), and in all cases we observed modest conversion and
modest yield of the desired product P25. The catalyst Ir2 was
among the best for this transformation, with P25 formed in 45%
yield, but even in this case we observed 14% of the HDF product
P13 and 41% of unconverted starting material aer 22 h, with
no further conversion aer 68 h. Also, as shown in Table S3 in
the ESI,† we observe quantitative mass balance in these reac-
tions, indicating there are not deleterious side reactions
involving S13. Taken together, the results of these studies seem
to indicate catalyst decomposition under the reaction condi-
tions, as in all cases modest conversion is observed aer 22 h,
but no improvement in yield is observed aer 68 h. We suspect
that the rather acidic pentauorophenol byproduct, formed
following prenyl group transfer, reacts deleteriously with the
electron-rich, highly basic b-diketiminate ligands in Ir1 and Ir2.
Activation of C–O bonds
Stephenson and coworkers have introduced a two-step strategy
for cleavage of b-O-4 linkages in substrates relevant to lignin
depolymerization, which proceeds via oxidation of the substrate
secondary alcohol to a ketone, followed by Ca–O cleavage under
photoredox conditions using [Ir(ppy)2(dtbbpy)]+ as the cata-
lyst.53,54 Using the typical reaction conditions described here,
with 1 mol% of Ir1 and 2 equiv of BIH, we observed near-
quantitative cleavage of S26 to the respective ketone (P26a)
and phenol (P26b) products (Scheme 4). We did not observe any
cleavage of native lignin model substrates that had not been
pre-oxidized to the ketone. Finally, as also summarized in
Scheme 4, the combination of Ir1 and BIH was effective for the
hydrogenation of benzophenone (S27) to diphenylmethanol
(P27). Benzophenone is difficult to reduce and has a reduction
potential of ꢁ2.2 V vs. Fc+/Fc.55,56 Thus, photoredox activation is
uncommon although it has been previously reported,57 and
Scheme
reactions.
3 Intramolecular and intermolecular functionalization
secondary alkyl bromide substrate S22, the two diastereomeric
products are formed in a 1 : 1 ratio, further supporting the
radical nature of the reaction.
Intermolecular functionalizations
We also investigated intermolecular functionalization reactions
catalyzed by Ir1 and Ir2, where the organic radical that is
generated is intercepted by a second reagent. We initially chose
the radical acceptor N-methylpyrrole for these studies, previ-
42
¨
ously used by Konig and coworkers. We subjected aryl
bromide substrate S2 and aryl chloride substrate S10 to our
optimized conditions in the presence of 10 equiv. of N-methyl-
pyrrole, which resulted in moderate yields of the arylated
products (P23 and P24) along with 20–25% yield of the respec-
tive hydrodehalogenation product (Scheme 3). We attempted Scheme 4 Activation of substrates with C–O bonds.
4074 | Chem. Sci., 2021, 12, 4069–4078
© 2021 The Author(s). Published by the Royal Society of Chemistry

View Article Online
Edge Article
Chemical Science
chemical reduction typically requires hydridic reagents like compared to 91% and 89% with Ir1 and Ir2, respectively. Thus,
sodium borohydride. Quantitative reduction of S27 to P27 was for most of the tested hydrodebromination reactions Ir1 and Ir2
observed with 1 mol% Ir1, 2 equiv of BIH, and blue LED irra- outperform fac-Ir(ppy)3.
diation for 24 h. This outcome deviates from previously re-
ported photoredox reactions of benzophenone and other
aromatic ketones, which use amine sacricial reagents and
Mechanistic considerations
proceed via reductive C–C coupling to form substituted pinacol As summarized in Scheme 5, there are two limiting mecha-
products.16,57 In our reaction conditions, where BIH is used as nisms for photoredox activation of C–X bonds. Experiments
the sacricial reductant in combination with Ir1 or Ir2, the ketyl described above and summarized in the ESI (Section VI and
radical intermediate that is formed via SET to the ketone is Fig. S3†) speak against a radical chain mechanism, leaving
trapped before it can dimerize.
more typical oxidative and reductive quenching pathways as the
This either suggests that the ketyl radical does not build up likely mechanisms. In the oxidative quenching mechanism, the
to a substantial concentration in our reaction mixture to allow excited Ir photosensitizer ([Ir]*) donates an electron to the
bimolecular combination, or that BIH is more effective at substrate directly, which fragments to form an aryl or alkyl
trapping ketyl radicals via hydrogen atom transfer than an radical. In the presence of BIH donation of a hydrogen atom
amine is. We hypothesize that two equivalents of BIH are furnishes the hydrodehalogenation product, but if a suitable
involved in this transformation to generate the H2 equivalent intra- or intermolecular radical trap is present, functionaliza-
that adds to benzophenone.
tion can occur. For the reductive quenching mechanism, the net
outcome is the same, but the photosensitizer in its excited state
instead accepts an electron from BIH, and the reduced photo-
sensitizer ([Ir]ꢁ) reacts with the substrate to generate the
Comparison with fac-Ir(ppy)3
Also of interest in this study is how the catalytic performance of substrate radical. The formation of Rc and Xꢁ following electron
complexes Ir1 and Ir2 compares with fac-Ir(ppy)3, historically transfer to R–X is typically a stepwise process, where electron
the most widely used photosensitizer for challenging reductive transfer generates [R–X]cꢁ, which then fragments.58,59 Unpro-
transformations. We also carried out hydrodehalogenation on ductive geminate recombination is thus possible if the reduced
select substrates using fac-Ir(ppy)3 as the photocatalyst under [R–X]cꢁ species has a sufficient lifetime, and this may factor into
blue light irradiation, using otherwise identical conditions to the low product-formation quantum yields we observe with S2
those outlined in Schemes 1 and 2. The product yields are and S7 (see above). Acknowledging the importance of the
summarized in Table 1 and compared directly with the yields stepwise fragmentation of R–X, we still show it in Scheme 5 as
obtained with Ir1 (blue light irradiation) and Ir2 (green light a single step for simplicity, and once formed the fate of Rc
irradiation). For one aryl bromide substrate, S1, and one aryl depends on what other reagents are available to trap it.
chloride substrate, S9, we obtained near-quantitative yields with
Thermodynamically, the oxidative quenching pathway
fac-Ir(ppy)3, as we did with Ir1 and Ir2. However, for the should be preferred. The excited-state IrIV/*IrIII potentials of Ir1
remainder of the tested substrates we obtained lower yields with and Ir2 are ꢁ2.6 and ꢁ2.4 V vs. Fc+/Fc (see Fig. 2), which is
fac-Ir(ppy)3. This includes the previously reported hydro- enough reducing power to transfer an electron to most aryl
deuorination of S13, which proceeded to 75% yield with fac- halide substrates.59,60 In contrast, the excited-state *IrIII/IrII
Ir(ppy)3 with 0.5 mol% loading,44 compared to 89% and 88% potentials of Ir1 and Ir2 are estimated to be ꢁ0.4 V and ꢁ0.7 V,
with Ir1 and Ir2, at half the catalyst loading. In addition, for aryl more negative than the Eox of BIH (ꢁ0.07 V),61 indicating that
bromide substrates S2, S5, and S6, hydrodebromination yields reductive quenching by BIH should not be thermodynamically
ranged from 66% to 74% with fac-Ir(ppy)3, compared with 81% favorable. We do note that the *IrIII/IrII potentials are a crude
to 99% with Ir1 and Ir2. Finally, for adamantyl bromide (S19) estimate since we could not identify a clear IrIII/IrII wave in Ir1
the yield of hydrodebromination is only 25% with fac-Ir(ppy)3, and Ir2 and the triplet-state energy has an uncertainty of
ꢃ0.1 eV. Nevertheless, it seems that of the two initial photore-
actions outlined in Scheme 5, oxidative quenching by the
Table
1 Comparison of hydrodehalogenation yields for select
substrate is thermodynamically more likely.
substrates, using fac-Ir(ppy)3, Ir1, and Ir2 as photosensitizers
In an attempt to determine the kinetically preferred pathway
for Rc formation, we performed Stern–Volmer quenching
experiments of Ir1 with two substrates and BIH. Our prior
investigations, conducted in MeCN with several aryl bromide
substrates, including S1, S3, S6, and S7, revealed that oxidative
quenching by the substrate (kq ¼ 0.078–4.6 ꢄ 108 Mꢁ1 sꢁ1 in
MeCN)32 was generally preferred over reductive quenching by
the sacricial reagent TMEDA (kq ¼ 6.3 ꢄ 106 Mꢁ1 sꢁ1 in
MeCN). In this work our typical optimized reactions were
carried out in a different solvent (DMF) with a different sacri-
cial reagent (BIH), prompting us to reevaluate the kinetics of
some of the key electron-transfer steps. Despite reductive
a
Substrate
fac-Ir(ppy)3
Ir1
Ir2
S1
S2
S5
S6
S9
S13
S19
99%
66%
69%
74%
86%
89%
81%
99%
99%
89%
91%
99%
88%
97%
97%
99%
88%
89%
99%
75% (ref. 44)b
25%
a
Conditions identical to those shown in Schemes 1 and 2 unless
otherwise noted. For fac-Ir(ppy)3 blue light irradiation was used.
b
0.5 mol% fac-Ir(ppy)3.
© 2021 The Author(s). Published by the Royal Society of Chemistry
Chem. Sci., 2021, 12, 4069–4078 | 4075

View Article Online
Chemical Science
Edge Article
Scheme 5 Proposed quenching pathways involving iridium photosensitizers, [Ir], and substrates to generate organic radical intermediates.
quenching seemingly being thermodynamically disfavored, BIH using Ir1 as the photosensitizer. These quenching rate
quenches the excited state of Ir1 (Fig. S10†) with an observed constants exceed those of fac-Ir(ppy)3 with the same substrates,
rate constant of 2.3 ꢄ 108 Mꢁ1 sꢁ1 in DMF. We note a similar, by a factor of 1.4 (S7, kq ¼ 1.4 ꢄ 108 Mꢁ1 sꢁ1) and 1.5 (S10, kq ¼
albeit slightly smaller quenching rate constant of 1.5 ꢄ 5.3 ꢄ 107 Mꢁ1 cmꢁ1), continuing the trend that oxidative
108 Mꢁ1 cmꢁ1 for fac-Ir(ppy)3 with BIH, which on the basis of quenching with the more strongly photoreducing Ir1 is faster
redox potentials (Fig. 1) should be faster if reductive quenching than fac-Ir(ppy)3. Thus, for transformations where oxidative
were occurring. Noting also that Stern–Volmer alone does not quenching is an important elementary step, the more reducing
identify the quenching mechanism,62 we cannot conclusively excited state of Ir1 can lead to signicantly faster rates of
state that the quenching by BIH is in fact electron-transfer excited-state electron transfer, compared to fac-Ir(ppy)3.
quenching, and it may be caused by unproductive triplet–
The thermodynamic and kinetic data described above is only
triplet energy-transfer without electron transfer. Regardless of relevant to electron-transfer steps that occur from the photo-
the precise quenching pathway with BIH, its kq value exceeds sensitizer excited state. Another interesting mechanistic
the quenching rate constant for aryl bromide substrate S7 (1.9 consideration is the precise role of the BIH sacricial reagent,
ꢄ 108 Mꢁ1 sꢁ1) and aryl chloride substrate S10 (8.1 ꢄ 107 Mꢁ1 particularly as it pertains to the hydrogen-atom transfer steps.
sꢁ1), which almost certainly involve oxidative electron-transfer The electron and hydrogen-atom transfer chemistry of BIH is
quenching. This observation, in concert with the practical well understood.61,63 BIH itself is a reasonably strong reductant,
consideration that all reactions in this study were conducted with E(BIHc+/BIH) ¼ ꢁ0.07 V. In addition, the C–H bond
with excess BIH (1.5–3 equiv.), indicates that quenching by BIH strength in BIH is considerably weaker than that of a typical
is the kinetically preferred pathway. That said, this outcome alkene or arene. As shown in Scheme 5 it is likely that the
does not guarantee that the quenching by BIH is productive, organic radical that forms, Rc, can be trapped by BIH to form R–
and the rate constants for the two quenching pathways appear H. However, it is worth noting that the C–H bond strength in
to be similar. For some of the more easily reduced substrates BIHc+, which would form following electron-transfer from BIH,
(e.g. S1 and S9) direct quenching by the substrate is likely is over 40 kcal molꢁ1 weaker than that of BIH,61 so it is possible
competitive with or even faster than quenching by BIH. Thus, that this radical cation can serve as the hydrogen atom donor
these studies don't conclusively identify a mechanism, but they provided it is able to diffuse to the organic radical. It is also
do suggest that the quenching pathways involving BIH and the possible to consider a radical chain mechanism that propagates
substrate are kinetically competitive.
via the BIc radical, although this alternative is not favored by the
Considering the likely importance of substrate oxidative available thermodynamic and mechanistic data. BIc, which
quenching, we also investigated select quenching rate constants would form following hydrogen atom transfer from BIH, is itself
of fac-Ir(ppy)3 in comparison with Ir1. We have previously a potent reductant with E(BI+/BIc) ¼ ꢁ2.06 V.61 This reducing
described oxidative quenching rate constants of model potential is below that required to reduce a typical aryl halide or
substrates methyl viologen and benzophenone with fac-Ir(ppy)3, alkyl halide, for which Ered lies beyond ꢁ2.3 V.59,60 Moreover, the
Ir1, Ir2, and several other NacNac-supported bis-cyclometalated excited states of Ir1 and Ir2 (E(IrIV/IrIII) ¼ ꢁ2.4 to ꢁ2.6 V) and
iridium photosensitizers.31,32 We showed in this previous work their reduced state that would form via reductive quenching
that the greater reducing power of Ir1 and Ir2 leads to signi- (E(IrIII/IrII) ꢀ ꢁ2.7 V) are both substantially more potent as
cantly faster excited-state electron transfer, particularly for reductants. In addition, as discussed above the quantum yield
benzophenone which has a reduction potential similar to data and light “on/off” experiments argue against a dark radical
typical aryl halides.56 As noted above we measured quenching chain mechanism. For these reasons, we favor the oxidative/
rate constants of aryl bromide substrate S7 (1.9 ꢄ 108 Mꢁ1 sꢁ1
)
reductive quenching mechanisms shown in Scheme 5, as
and aryl chloride substrate S10 (8.1 ꢄ 107 Mꢁ1 sꢁ1) in DMF, opposed to the radical chain process involving electron transfer
4076 | Chem. Sci., 2021, 12, 4069–4078
© 2021 The Author(s). Published by the Royal Society of Chemistry

View Article Online
Edge Article
Chemical Science
¨
¨
to the substrate by BIc and hydrogen atom transfer to the
8 M. Neumann, S. Fuldner, B. Konig and K. Zeitler, Angew.
substrate radical by BIH.
Chem., Int. Ed., 2011, 50, 951–954.
9 K. Dedeian, P. I. Djurovich, F. O. Garces, G. Carlson and
R. J. Watts, Inorg. Chem., 1991, 30, 1685–1687.
10 J. D. Nguyen, E. M. D'Amato, J. M. R. Narayanam and
C. R. J. Stephenson, Nat. Chem., 2012, 4, 854–859.
Conclusions
We have demonstrated that bis-cyclometalated iridium
complexes with b-diketiminate ancillary ligands are versatile 11 E. D. Nacsa and D. W. C. MacMillan, J. Am. Chem. Soc., 2018,
photosensitizers for a variety of photoredox transformations. 140, 3322–3330.
The highly reducing nature of their excited state permits an 12 H. Kim and C. Lee, Angew. Chem., Int. Ed., 2012, 51, 12303–
expansion of the substrate scope to challenging organobromide
12306.
and organochloride substrates, which traditionally are inert to 13 S. O. Poelma, G. L. Burnett, E. H. Discekici, K. M. Mattson,
photoredox catalysis or require forcing, complex reaction
conditions to activate. These reactions generate an organic
radical following photoinduced electron transfer, and the fate
N. J. Treat, Y. Luo, Z. M. Hudson, S. L. Shankel,
P. G. Clark, J. W. Kramer, C. J. Hawker and J. Read de
Alaniz, J. Org. Chem., 2016, 81, 7155–7160.
of this radical depends on the nature of the substrate and the 14 R. Matsubara, T. Yabuta, U. Md Idros, M. Hayashi, F. Ema,
reaction conditions. In most cases under typical reaction Y. Kobori and K. Sakata, J. Org. Chem., 2018, 83, 9381–9390.
conditions, hydrodehalogenation is the outcome, which estab- 15 T.-H. Ding, J.-P. Qu and Y.-B. Kang, Org. Lett., 2020, 22, 3084–
lishes the breadth of the substrate scope enabled by these
3088.
reaction conditions. If the substrate has an appropriately posi- 16 K. Li, Q. Wan, C. Yang, X.-Y. Chang, K.-H. Low and
tioned aryl or alkenyl group, radical cyclization is possible, and C.-M. Che, Angew. Chem., Int. Ed., 2018, 57, 14129–14133.
reactions conducted in the presence of a second radical trap 17 L. Zhang, Z. Wu and L. Jiao, Angew. Chem., Int. Ed., 2020, 59,
substrate can result in bimolecular functionalization. Overall, 2095–2099.
this work motivates the continued pursuit of designer photo- 18 A. R. Flynn, K. A. McDaniel, M. E. Hughes, D. B. Vogt and
catalysts and modied photoredox reaction conditions that will N. T. Jui, J. Am. Chem. Soc., 2020, 142, 9163–9168.
allow a much wider range of cheap, readily available substrates 19 P. Zhang, C. (Chip) Le and D. W. C. MacMillan, J. Am. Chem.
to be used in synthetically valuable photoredox
transformations.
Soc., 2016, 138, 8084–8087.
20 J. J. Devery, J. D. Nguyen, C. Dai and C. R. J. Stephenson, ACS
Catal., 2016, 6, 5962–5967.
¨
21 I. Ghosh, T. Ghosh, J. I. Bardagi and B. Konig, Science, 2014,
Conflicts of interest
346, 725–728.
There are no conicts to declare.
22 I. A. MacKenzie, L. Wang, N. P. R. Onuska, O. F. Williams,
K. Begam, A. M. Moran, B. D. Dunietz and D. A. Nicewicz,
Nature, 2020, 580, 76–80.
23 C. Kerzig, X. Guo and O. S. Wenger, J. Am. Chem. Soc., 2019,
141, 2122–2127.
Acknowledgements
We acknowledge the National Science Foundation (NSF) (CHE-
1846831 to T. S. T. and CHE-1453891 to J. D. W.) for support. T. 24 C. Kerzig and O. S. Wenger, Chem. Sci., 2019, 10, 11023–
S. T. also acknowledges the Welch Foundation (Grant No. E- 11029.
1887) for funding this research. We thank Prof. Loi Do for 25 H. Kim, H. Kim, T. H. Lambert and S. Lin, J. Am. Chem. Soc.,
access to the gas chromatography.
2020, 142, 2087–2092.
26 N. G. W. Cowper, C. P. Chernowsky, O. P. Williams and
Z. K. Wickens, J. Am. Chem. Soc., 2020, 142, 2093–2099.
27 W. Sattler, M. E. Ener, J. D. Blakemore, A. A. Rachford,
P. J. LaBeaume, J. W. Thackeray, J. F. Cameron,
J. R. Winkler and H. B. Gray, J. Am. Chem. Soc., 2013, 135,
10614–10617.
Notes and references
1 C. K. Prier, D. A. Rankic and D. W. C. MacMillan, Chem. Rev.,
2013, 113, 5322–5363.
2 T. R. Cook, D. K. Dogutan, S. Y. Reece, Y. Surendranath,
¨
T. S. Teets and D. G. Nocera, Chem. Rev., 2010, 110, 6474– 28 L. A. Buldt, X. Guo, A. Prescimone and O. S. Wenger, Angew.
6502. Chem., Int. Ed., 2016, 55, 11247–11250.
3 T. S. Teets and D. G. Nocera, Chem. Commun., 2011, 47, 9268– 29 P. Herr, F. Glaser, L. A. Buldt, C. B. Larsen and O. S. Wenger,
9274. J. Am. Chem. Soc., 2019, 141, 14394–14402.
4 M. Chen, M. Zhong and J. A. Johnson, Chem. Rev., 2016, 116, 30 E. H. Discekici, N. J. Treat, S. O. Poelma, K. M. Mattson,
¨
10167–10211.
Z. M. Hudson, Y. Luo, C. J. Hawker and J. R. de Alaniz,
Chem. Commun., 2015, 51, 11705–11708.
31 J.-H. Shon and T. S. Teets, Inorg. Chem., 2017, 56, 15295–
15303.
5 D. A. Nagib and D. W. C. MacMillan, Nature, 2011, 480, 224–
228.
6 E. B. Corcoran, M. T. Pirnot, S. Lin, S. D. Dreher,
D. A. DiRocco, I. W. Davies, S. L. Buchwald and 32 J.-H. Shon, S. Sittel and T. S. Teets, ACS Catal., 2019, 8646–
D. W. C. MacMillan, Science, 2016, 353, 279–283.
8658.
7 D. M. Schultz and T. P. Yoon, Science, 2014, 343, 1239176.
© 2021 The Author(s). Published by the Royal Society of Chemistry
Chem. Sci., 2021, 12, 4069–4078 | 4077

View Article Online
Chemical Science
Edge Article
33 T. U. Connell, C. L. Fraser, M. L. Czyz, Z. M. Smith, 48 J. Xie, M. Rudolph, F. Rominger and A. S. K. Hashmi, Angew.
D. J. Hayne, E. H. Doeven, J. Agugiaro, D. J. D. Wilson, Chem., Int. Ed., 2017, 56, 7266–7270.
J. L. Adcock, A. D. Scully, D. E. Gomez, N. W. Barnett, 49 J. B. I. Sap, N. J. W. Straathof, T. Knauber, C. F. Meyer,
´
´
M. Medebielle, L. Buglioni, C. Genicot, A. A. Trabanco,
A. Polyzos and P. S. Francis, J. Am. Chem. Soc., 2019, 141,
17646–17658.
¨
T. Noel, C. W. am Ende and V. Gouverneur, J. Am. Chem.
34 V. Kottisch, Q. Michaudel and B. P. Fors, J. Am. Chem. Soc.,
2017, 139, 10665–10668.
35 J.-H. Shon and T. S. Teets, Comments Inorg. Chem., 2020, 40,
53–85.
Soc., 2020, 142, 9181–9187.
50 J. W. Tucker, J. M. R. Narayanam, S. W. Krabbe and
C. R. J. Stephenson, Org. Lett., 2010, 12, 368–371.
51 B. Hu, Y. Li, W. Dong, K. Ren, X. Xie, J. Wan and Z. Zhang,
Chem. Commun., 2016, 52, 3709–3712.
36 Y. Pellegrin and F. Odobel, C. R. Chim., 2017, 20, 283–295.
37 Y. Zhang, T. S. Lee, J. L. Petersen and C. Milsmann, J. Am. 52 S. Priya and J. D. Weaver, J. Am. Chem. Soc., 2018, 140, 16020–
Chem. Soc., 2018, 140, 5934–5947. 16025.
38 E. Hasegawa, N. Izumiya, T. Miura, T. Ikoma, H. Iwamoto, 53 J. D. Nguyen, B. S. Matsuura and C. R. J. Stephenson, J. Am.
S. Takizawa and S. Murata, J. Org. Chem., 2018, 83, 3921–
3927.
39 Y. Tamaki, K. Koike, T. Morimoto and O. Ishitani, J. Catal.,
2013, 304, 22–28.
40 H. Takeda, K. Ohashi, A. Sekine and O. Ishitani, J. Am. Chem.
Soc., 2016, 138, 4354–4357.
Chem. Soc., 2014, 136, 1218–1221.
¨
¨
54 I. Bosque, G. Magallanes, M. Rigoulet, M. D. Karkas and
C. R. J. Stephenson, ACS Cent. Sci., 2017, 3, 621–628.
55 N. G. Connelly and W. E. Geiger, Chem. Rev., 1996, 96, 877–
910.
56 N. G. Tsierkezos, J. Solution Chem., 2007, 36, 1301–1310.
41 M. A. Cismesia and T. P. Yoon, Chem. Sci., 2015, 6, 5426– 57 M. Nakajima, E. Fava, S. Loescher, Z. Jiang and M. Rueping,
5434. Angew. Chem., Int. Ed., 2015, 54, 8828–8832.
42 I. Ghosh, R. S. Shaikh and B. Konig, Angew. Chem., Int. Ed., 58 C. Costentin, M. Robert and J.-M. Saveant, J. Am. Chem. Soc.,
2017, 56, 8544–8549. 2004, 126, 16051–16057.
43 A. Rosas-Hernandez, C. Steinlechner, H. Junge and 59 L. Pause, M. Robert and J.-M. Saveant, J. Am. Chem. Soc.,
M. Beller, Green Chem., 2017, 19, 2356–2360. 1999, 121, 7158–7159.
44 S. M. Senaweera, A. Singh and J. D. Weaver, J. Am. Chem. Soc., 60 H. Roth, N. Romero and D. Nicewicz, Synlett, 2015, 27, 714–
2014, 136, 3002–3005. 723.
45 W. Xu, H. Jiang, J. Leng, H. Ong and J. Wu, Angew. Chem., Int. 61 X.-Q. Zhu, M.-T. Zhang, A. Yu, C.-H. Wang and J.-P. Cheng, J.
Ed., 2020, 59, 4009–4016. Am. Chem. Soc., 2008, 130, 2501–2516.
46 S. Senaweera and J. D. Weaver, J. Am. Chem. Soc., 2016, 138, 62 D. M. Arias-Rotondo and J. K. McCusker, Chem. Soc. Rev.,
2520–2523. 2016, 45, 5803–5820.
47 A. Singh, J. J. Kubik and J. D. Weaver, Chem. Sci., 2015, 6, 63 S. Ilic, U. Pandey Kadel, Y. Basdogan, J. A. Keith and
¨
´
´
´
7206–7212.
K. D. Glusac, J. Am. Chem. Soc., 2018, 140, 4569–4579.
4078 | Chem. Sci., 2021, 12, 4069–4078
© 2021 The Author(s). Published by the Royal Society of Chemistry