Organic Letters
Letter
The field of PhotoRedox Catalysis (PRC) evolved rapidly
over the past two decades inspired by photocatalytic reactions
using Ru(II)- or Ir(III)-bipyridyl complexes.23−28 A paradigm
shift toward the use of organic dye and first-row transition
metal photocatalysts has provided an alternative to the use of
unsustainable, toxic, and expensive Ru- and Ir-based photo-
catalysts.29−31 Another recently trending concept in PRC
completely obviates the need for an exogenous “photocatalyst”
to harness visible photon energy as chemical energy. This
approach leverages the photoactivity of a ground state
assembly of an electron donor and an electron acceptor,
referred to as an electron donor−acceptor (EDA) complex.
While the electron donor (D) and acceptor (A) may be
colorless on their own, their charge transfer interaction results
in a bathochromic shift to afford a visible light absorbing
colored species that, upon photoexcitation, affords reactive
intermediates (radical anions and cations) via SET (Figure
2).32−34 Although the substitution of photocatalysts by EDA
Table 1. Organophotocatalyst Screening and Control
Reactions
a
Reaction conditions: 0.135 M, (NH4)2S2O8 (5 equiv), 24 h.
Other oxidants were attempted but led to no reaction. Changes
in oxidant loading either severely decreased yields (decom-
position) or gave no reaction. Efforts to conduct reactions in
other aprotic solvents with a lower excess of amide (15 equiv)
did not achieve satisfactory yields, consistent with previous
reports where amides were required in excess.19,21,22
Anhydrous conditions were essential while strict degassing
was not required. A Design of Experiments (see SI) approach
varying overall concentration, temperature, and reaction time
found that the optimum conditions affording the best
compromise of monoadduct yield and monoadduct selectivity
were 0.135 M 1 in DMF, 24 h, rt, and 5 equiv of persulfate,
providing 77% (1H NMR) of 3 with 12.8:1 (monoadduct/
bisadduct) selectivity. When the optimal reaction conditions
were tested in the dark, (i) 3’s yield was halved, (ii) conversion
was decreased, and (iii) the proportion of productive
conversion was lower, indicating more decomposition.
With the optimized conditions in hand, the substrate scope
was examined first with respect to arene partners (Scheme 1).
Where some methoxyarene substrates and derivatives were not
converted quantitatively, longer reaction times were examined
to improve conversion and monoadduct yield. A number of
electron-rich arenes underwent Friedel−Crafts amidoalkylation
in moderate to high (24−84%) yields.
Formation of an EDA complex requires compatibility in the
electronics and/or sterics of its electron-rich and its electron-
poor partners.35 Lower substrate conversions (and yields of 5
and 10) were observed for 1,3-dimethoxybenzene and 1,2,3,5-
tetramethoxybenzene than 1,3,5-TMB. Substrates such as
1,2,3-TMB and 1,2,4-TMB which have the same number of
methoxy groups as 1 gave almost no reaction (see SI). The
reaction of 1,4-dimethoxybenzene (1,4-DMB) also failed to
proceed. This suggests a balance of electronics is required in
forming the EDA complex. Interestingly, a substrate primed for
intramolecular amidoalkylation en route to 7 resulted only in
intermolecular amidoalkylation. Comparable product yields
were obtained for the reactions of 1 with dimethylformamide
(DMF), dimethylacetamide (DMA), and N-methyl-2-pyrroli-
done (NMP). In products 8, 9, and 18, para-methoxybenzylic
positions were tolerated with no C(sp3)−H functionalization
on the anisole. An exocyclic five-membered amide afforded a
Figure 2. Formation of radical ion pairs via an EDA complex after
complexation and photoexcitation.
complexes represents a “greener” solution for C−H function-
alization, their prediction of occurrence and photophysical
characterization is more challenging.35 Moreover, their proper-
ties highly depend upon the individual components which
restricts their substrate scope of applications. Herein, we
disclose a photocatalyst-free visible-light-mediated aryl func-
tionalization of amides via an EDA complex.
Our initial investigation focused on finding a nontoxic and
cheap organophotocatalyst as an alternative to Ru(bpy)3Cl2
(II)/(III)
(E1/2*
= −0.81 V vs SCE) in the photochemical
Friedel−Crafts amidoalkylation reaction.19 1,3,5-Trimethoxy-
benzene (1,3,5-TMB) 1 was selected as an electron-rich arene
and ammonium persulfate as an oxidant. Finding similar yields
for several different organophotocatalysts screened prompted
us to conduct a control reaction in the absence of any
photocatalyst, which was just as efficient. This observation was
surprising, since such a control reaction in the hands of
Stephenson and co-workers gave <5% conversion. We
hypothesized an EDA complex was responsible. We speculated
that a combination of (i) our LEDs likely being higher power
than those employed in “early” PRC papers and (ii) our
reaction configuration placing reactors in closer proximity to
the LEDs may have resulted in an overall greater radiant flux to
the reaction to enable such reactivity. Although LED
specifications were not available in that paper, other
publications from the group around the same time period
(2011−2012) suggested common use of a 1 W input power
blue LED strip spread around a beaker at some (approximately
a few centimeters) distance from the reaction (∼5 mL).36−38
In contrast, each LED herein had an optical power of ∼0.85 W
and reactions (1 mL) were placed directly atop of LEDs for the
highest possible light penetration.
From entry 4 (Table 1) as a starting point, we began a
2003
Org. Lett. 2021, 23, 2002−2006
Kaur, Jaspreet
