Journal of the American Chemical Society
Article
a
detected when 2,2,6,6-tetramethylpiperidine-1-oxyl (TEMPO)
was added to the reaction mixtures (entry 11).
Table 1. Optimization of the Reaction Conditions
After robust optimization of the reaction conditions, we next
investigated the scope of various unactivated hydrocarbons to
evaluate the generality of this selective C−H pyridylation
(Table 2). First, we evaluated the broadly available feedstock
alkane family in terms of site selectivity and reactivity.
Pleasingly, the current method proceeded selectively at the
tertiary C−H bonds among multiple hydridic protons (>20:1
r.r. if not denoted otherwise) in various linear alkane chains,
such as 2,5-dimethylhexane (3a), 2,3-dimethylbutane (3b),
2,4-dimethylpentane (3c), 2,2,5-trimethylhexane (3d), iso-
pentane (3e), and 2-methylpentane (3f). Notably, this
protocol tolerated versatile functional groups such as
phthalimide (3i), benzoate (3j and 3k), ester (3l), bromide
(3g), and chloride (3h), and the intriguing synthetic value of
this transformation was further demonstrated by its excellent
site- and chemoselectivity for methine positions (bond
dissociation energy (BDE) = 96.5 kcal/mol)80 over the α-
position of heteroatoms in alkane substrates. Remarkably, for
substrates bearing more than one tertiary C−H bond with
subtle differences in terms of steric effects and BDEs, this
catalytic system has a marked preference for the abstraction of
slightly more electron-rich methine C−H bonds, yielding
targeted products 3m and 3n. Next, the pivotal importance of
small feedstock alkanes prompted us to evaluate the
applicability of this method to more challenging gaseous n-
butane. Gratifyingly, we observed that this protocol enables the
efficient functionalization of n-butane to afford synthetically
useful building blocks (3q) with excellent selectivity despite
the small difference in BDEs between secondary (BDE = 98.6
kcal/mol) and primary hydrogens (BDE = 101.1 kcal/mol).
Likewise, n-pentane and n-hexane were successfully trans-
formed, and trapping of the pyridyl group with two competing
methylene C−H bonds resulted in two isomeric mixtures (3r
and 3s). Employing a substituted cyclic alkane resulted in the
formation of the corresponding product 3o with excellent
selectivity for the tertiary C−H bond. Similarly, the
adamantane derivative was amenable to the standard
conditions with an excellent degree of selectivity at the
methine position (3p). The current method can be extended
to norbornane and oxabicycloheptane, and methylene C−H
bonds on the ethylene bridge were selectively activated,
demonstrating an excellent level of regio- and diastereose-
lectivity (3t and 3u). Expanding the scope from electronically
neutral hydrocarbons to a diverse range of heteroatom-
containing substrates, such as methanol, 1,4-dioxane, thioether,
and aliphatic amides, was possible to successfully afford the
coupling products (3v−3z) with the incorporation of pyridine
exclusively at the carbon adjacent to the heteroatom. To
establish the broad utility of the unified protocol, substrates
bearing different classes of hydrogen donors (aldehyde, amide,
phosphine oxide, and alkylsilane units) were further examined.
Under the same reaction conditions, the phosphorus group
was compatible with the current conditions to afford 3aa and
3ab. Alkyl silane can also be utilized as a building block for this
protocol to insert a pyridine moiety (3ac). In addition,
aldehydes and amides can be successfully applied to achieve
umpolung reactivity exclusively at the acyl C−H bonds over
the tertiary and alpha positions of heteroatoms, accessing
diverse acylation-type products (3ad−3ai). Similarly, cyclic
alkanes were also competent substrates, which were efficiently
converted to the desired products with excellent yield
b
c
entry
salt
PC (1.0 mol%)
yield (%)
selectivity (C2/C3)
d
1
2a
2b
2b
2b
2b
2c
2d
2a
2c
2c
2c
fac-Ir(ppy)3
eosin Y
TBADT
benzophenone
AQ
AQ
AQ
AQ
AQ
trace
39
23
44
67
78
46
15
2
>20:1
9:1
>20:1
>20:1
>20:1
18:1
e
3
e
4
5
6
7
d
8
f
>20:1
9
n.d.
n.d.
n.d.
10
11
g
AQ
a
Reaction conditions: 1a (0.5 mmol), 2 (0.1 mmol), and catalyst (1.0
mol%) in acetonitrile (1.0 mL) under irradiation using blue LEDs
(440 nm) at rt for 48 h under N2. Yields were determined by H
b
1
c
1
NMR spectroscopy. The ratio of selectivity was determined by H
d
NMR spectroscopy and GC-MS. NaHCO3 (1.2 equiv) was added.
e
Without NaHCO3, no product was detected. A 390 nm Kessil LED
f
g
was used. Reaction was carried out in the dark. TEMPO (2.0 equiv)
was added. TBADT = tetrabutylammonium decatungstate. 1,2-DCE =
1,2-dichloroethane. DCM = dichloromethane. n.d. = not detected.
accompanied by a decrease in reactivity (entry 3). Next, we
considered the use of aryl ketones and quinones that have been
reported as useful visible-light HAT photocatalysts by merging
with Ni or Cu catalysts to achieve promising potential in the
catalytic activation of inert C−H bonds.67−79 For example,
Martin and co-workers reported C−H arylation and alkylation
under dual diaryl ketone HAT/Ni catalysis.69 In addition, the
Gong group achieved stereoselective C−H functionalization of
hydrocarbons under dual pentacenetetrone HAT/Cu catal-
ysis.70 Our evaluation of various members of this class of HAT
photocatalysts revealed that AQ derivatives improved the
product yield and selectivity (see the SI for details). Notably,
we found that the employment of AQ as the HAT catalyst
resulted in increased reactivity. Under this HAT/RHAT
catalytic system, an external base was not required. Solvent
screening experiments revealed that reaction was best
conducted in acetonitrile and that the C−H bond in
acetonitrile is inert under the given HAT catalytic systems
(see the SI for details). The effect of the sulfonamide group on
the N-aminopyridinium salt was next investigated by tuning its
electronic effect (entries 5−8). Replacing the p-tolyl group
with a more electron-deficient 4-trifluoromethyl phenyl moiety
(2c) further increased the yield (entry 6, 78%), probably
because the generated alkyl radical is more favorably
intercepted by electron-deficient salt. We carried out control
experiments to confirm the critical role of light and the
photocatalyst (entries 9 and 10). As anticipated, 3a was not
3005
J. Am. Chem. Soc. 2021, 143, 3003−3012