Ni-Catalyzed Divergent Cyclization/Carboxylation of Unactivated Primary and Secondary Alkyl Halides with CO2

Article abstract of DOI:10.1021/jacs.5b03340

A user-friendly Ni-catalyzed reductive cyclization/carboxylation of unactivated alkyl halides with CO₂ is described. The protocol operates under mild conditions with an excellent chemoselectivity profile and a divergent syn/anti selectivity pat

Full text of DOI:10.1021/jacs.5b03340

Communication
pubs.acs.org/JACS
Ni-Catalyzed Divergent Cyclization/Carboxylation of Unactivated
Primary and Secondary Alkyl Halides with CO2
Xueqiang Wang,†,§ Yu Liu,†,§ and Ruben Martin*,†,‡
†Institute of Chemical Research of Catalonia (ICIQ), Av. Països Catalans 16, 43007 Tarragona, Spain
‡Catalan Institution for Research and Advanced Studies (ICREA), Passeig Lluïs Companys 23, 08010 Barcelona, Spain
S
* Supporting Information
opportunity to control parasitic β-hydride elimination path-
ways10 while resulting in carboxylated carbocyclic skeletons
ABSTRACT: A user-friendly Ni-catalyzed reductive cyc-
lization/carboxylation of unactivated alkyl halides with CO2
is described. The protocol operates under mild conditions
with an excellent chemoselectivity profile and a divergent
syn/anti selectivity pattern that can be easily modulated by
the substrate utilized.
from simple precursors via distal catalytic CO2 fixation. We
speculated that a technique capable of modulating, at will, the
anti/syn selectivity of the cyclization event would set the
standards for catalytic biomimetic cascade carboxylation
events.12 Herein, we report a mild and user-friendly reductive
cyclization/carboxylation of unactivated alkyl halides with CO2
en route to elusive tetrasubstituted olefins (Scheme 2, path
atalytic reductive coupling reactions of organic halides
C
have evolved from mere curiosities to robust tools that
rapidly build up molecular complexity from simple precursors.1
At present, this field of expertise remains essentially confined to
bond-formation events at the initial site (Scheme 1, path a).
Scheme 2. Cyclization/Functionalization of Alkyl Halides
Scheme 1. Bond Formation via Electrophile Couplings
b).13 In sharp contrast to syn-carbometalation techniques using
stoichiometric, well-defined, and in many instances air-sensitive
organometallics (Scheme 2, path a),14 our protocol is
characterized by its exquisite chemoselectivity profile while
obviating the need for sensitive species. Importantly, this
transformation is distinguished by an unconventional diver-
gence in syn/anti- selectivity that can be easily dictated by the
ligand backbone or substrate utilized.
We began our investigations by studying the catalytic cascade
cyclization/carboxylation reaction of 1aa with CO2 (1 atm)
utilizing NiCl2·glyme as the catalyst and Mn as the reductant in
N,N-dimethylformamide at room temperature (Table 1).15 As
for related catalytic reductive coupling processes,1 we
anticipated that subtle differences in the ligand backbone
would exert a profound influence on the reaction outcome. As
shown in Table 1, this turned out to be the case; while L1−L4
predominantly resulted in nonproductive β-hydride elimination
pathways (entries 1−4), the inclusion of ortho substituents in
the phenanthroline backbone cleanly produced 2a. Among
them, L5 and L6 allowed 2a to be obtained in respectable
yields of 33% and 39%, respectively, with no observable side
Intriguingly, the ability to promote cascade reactions of
unactivated alkyl electrophiles via multiple C−C bond formations
(path b) has been virtually unexplored.2,3 If successful, such
protocols would offer a unique opportunity to increase our
chemical portfolio for rapidly preparing carbocyclic skeletons
while dealing with bond-formation events at distal sites.
In recent years, we4 and others5 have designed new catalytic
techniques for reductive carboxylation of organic halides using
CO2, probably the greenest C1 synthon in nature.6 Unlike the
utilization of stoichiometric amounts of organometallic
complexes,7 many of these protocols operate under mild
conditions in the absence of sensitive reagents, thus
representing a straightforward, yet practical, alternative for
preparing carboxylic acids, privileged motifs in a myriad of
pharmaceuticals.8 At the outset of our investigations, however,
it was unclear whether CO2 could participate in cascade
reductive coupling reactions via multiple bond-forming
reactions.9 Although we anticipated that reductive cascade
processes based on the use of unactivated alkyl halides,10
probably the most challenging substrates in the cross-coupling
arena, would be rather problematic, we were attracted to the
challenge.11 Specifically, such a route would offer the unique
Received: March 31, 2015
© XXXX American Chemical Society
A
DOI: 10.1021/jacs.5b03340
J. Am. Chem. Soc. XXXX, XXX, XXX−XXX

Journal of the American Chemical Society
Communication
a
a b
,
Table 1. Optimization of the Reaction Conditions
Table 2. Scope Unactivated Primary Alkyl Bromides
a
1aa (0.30 mmol), Ni catalyst (10 mol %), L (20 mol %), Mn (2.20
b
equiv), DMF (0.15 M), and CO2 (1 atm) at rt overnight. Determined
by HPLC using naphthalene as an internal standard. NiBr2·diglyme (5
mol %). Using 1a (0.30 mmol); isolated yield. Without Mn or with
Zn as the reductant. DMA as the solvent. MeCN as the solvent.
c
a
b
d
e
Conditions: see Table 1, entry 10. Isolated yields (averages of at
c
f
g
least two independent runs) are shown. E/Z = 19:1.
via competitive trimerization pathways.17 As anticipated from a
classical syn-carbometalation reaction via in situ generated alkyl
nickel species,14,18 we obtained 2r and 2s. The structure of 2r
was unequivocally established by X-ray crystallographic
analysis.15
products (entries 5 and 6). Strikingly, the choice of precatalyst
(entries 7−9), solvent (entries 11 and 12), and reductant
(entry 10) had a non-negligible effect on the reactivity,
suggesting an intimate interplay among all of the reaction
parameters. While not anticipated, we obtained the best results
with the a priori less activated alkyl bromide 1a (entry 10 vs 8),
which gave 2a in 85% isolated yield.16 Importantly, not even a
trace of 3a was found in the crude reaction mixtures. In line
with our expectations, control experiments revealed that all of
the reaction parameters (NiBr2·diglyme, L6, Mn, and DMF)
were critical for success.15
Encouraged by these results, we turned our attention to the
preparative scope of our catalytic cyclization/carboxylation
reaction (Table 2). Particularly noteworthy was the functional
group tolerance of our protocol, as ketones (2d), ethers (2b,
2i), esters (2e, 2j), amides (2f), alkenes (2m), and heterocycles
(2o, 2p) all were perfectly accommodated. Undoubtedly, the
exquisite chemoselectivity profile of our transformation
represents a bonus compared with classical carbometalation
techniques based on the utilization of organolithium or
Grignard reagents, among others (Scheme 2, path a).14 As
shown for 2g, the inclusion of ortho substituents on the
aromatic motif did not hamper the reaction. Interestingly, we
found that the cyclization/carboxylation event could even be
conducted in the presence of electrophilic partners that are
suited for Ni-catalyzed reductive carboxylation reactions, such
as aryl chlorides (2l),5c tosylates (2k), or pivalates (2j);4c
notably, no traces of the corresponding benzoic acids derived
from C−Cl or C−O bond cleavage were detected in the crude
reaction mixtures, thus providing ample opportunities for
further functionalization. While one might argue that such a
protocol would essentially be restricted to five-membered rings
or alkyne residues possessing aromatic motifs, the preparation
of 2n, 2o, 2p, and 2q clearly indicates otherwise. Strikingly, free
alkynes posed no problems (2h); such a finding is certainly
remarkable in view of the proclivity of terminal alkynes to react
Next, we focused our attention on a more challenging
scenario dealing with unactivated secondary alkyl halides.10
These substrates are particularly problematic because of their
reluctance to undergo oxidative addition and their propensity
to undergo nonproductive β-hydride elimination, thus con-
stituting an opportunity to explore the robustness of our
cyclization/carboxylation event (Table 3). Strikingly, the
reaction of 4a using L6 under conditions otherwise similar to
those of Table 2 resulted in an unexpected selectivity switch
(5a:5a′ = 3.3:1). In a formal sense, 5a can be derived from a
rather elusive anti-carbometalation event.19 Although 5a was
fully characterized by NMR spectroscopic analysis, X-ray
crystallography unambiguously identified the abnormal anti-
selective motion.15 It is worth noting that the preparation of 5a
represents the first reductive carboxylation that can be
conducted with unactivated secondary alkyl electrophiles.
Interestingly, the anti selectivity could be modulated by the
choice of ligand. Specifically, we found that L5 uniquely
afforded 5a, with only small amounts of 5a′ being present in
the crude mixtures (5a:5a′ = 12.5:1). At present, we have no
explanation for this intriguing behavior.20
On the basis of these results, we wondered whether the
observed anti-selectivity switch for 5a could be applied to other
substrate combinations. As shown in Table 3, this was indeed
the case, and a host of differently substituted secondary alkyl
bromides could be coupled in high yields with high anti
selectivities.21 Notably, six-membered carbocyclic skeletons
could also be accommodated, albeit in lower yields (5f). A
simple comparison of 5a vs 5b and 5c clearly shows that the
anti selectivity is favored with bulkier substituents on the side
chain. A similar effect was found with ortho-substituted
aromatic motifs (5d vs 5a). Less counterintuitive was the
B
DOI: 10.1021/jacs.5b03340
J. Am. Chem. Soc. XXXX, XXX, XXX−XXX

Journal of the American Chemical Society
Communication
a b
,
Mn, the targeted cyclization/carboxylation product (2i or 4i)
was cleanly produced in the presence of the reducing agent.
Although premature, we believe that these experiments tacitly
suggest that the carboxylation event does not occur from in situ
generated Ni(II) species but rather from putative Ni(I) reaction
intermediates.22,29
Table 3. Scope Unactivated Secondary Alkyl Bromides
In conclusion, we have developed a mild, robust, and user-
friendly Ni-catalyzed cascade reductive cyclization/carboxyla-
tion using CO2 at atmospheric pressure in which the selectivity
pattern is dictated by an appropriate substrate and/or ligand
selection. Further investigations into related processes as well as
the development of an asymmetric version are currently
underway.
ASSOCIATED CONTENT
■
S
* Supporting Information
Experimental procedures and spectral data. The Supporting
Information is available free of charge on the ACS Publications
website at DOI: 10.1021/jacs.5b03340.
AUTHOR INFORMATION
a
b
■
Conditions: see Table 1, entry 10. Isolated yields (averages of at
c
Corresponding Author
*rmartinromo@iciq.es
least two independent runs) are shown. L6 was used as the ligand.
d
e
L5 was used as the ligand. L4 was used as the ligand.
Author Contributions
§X.W. and Y.L. contributed equally.
observation that the ligand backbone exerted a profound
influence on the selectivity pattern, with L5 or L4 providing the
best anti/syn selectivities, thus showing the subtleties of our
system.20 At present, we believe that the anti-selectivity switch
in secondary alkyl bromides might be attributed to the
intermediacy of vinyl radical species that undergo rapid
isomerization prior to recombination with Ni(I)BrLn spe-
cies.22−25 Taken together, the data shown in Tables 2 and 3
illustrate the prospective impact of our Ni-catalyzed reductive
cyclization/carboxylation event from simple building blocks by
promoting distal CO2 fixation while controlling the syn/anti
selectivity pattern of the cyclization event.
Notes
The authors declare no competing financial interest.
ACKNOWLEDGMENTS
■
We thank ICIQ, the European Research Council (ERC-
277883), MINECO (CTQ2012-34054 and Severo Ochoa
Excellence Accreditation 2014-2018, SEV-2013-0319) and the
Cellex Foundation for support. Johnson Matthey, Umicore, and
Nippon Chemical Industrial are acknowledged for gifts of metal
and ligand sources. Y.L. thanks COFUND for a fellowship. We
thank Dr. J. Cornella for preparing 8 and Eduardo Escudero for
all of the X-ray crystallographic data.
Although a detailed picture requires further studies, we
decided to shed light on the mechanism by studying the
REFERENCES
■
Scheme 3. Mechanistic Experiments
(1) For selected reviews, see: (a) Moragas, T.; Correa, A.; Martin, R.
Chem.Eur. J. 2014, 20, 8242. (b) Tasker, S. Z.; Standley, E. A.;
Jamison, T. F. Nature 2014, 509, 299. (c) Everson, D. A.; Weix, D. J. J.
Org. Chem. 2014, 79, 4793. (d) Knappe, C. E. I.; Grupe, S.; Gartner,
̈
D.; Corpet, M.; Gosmini, C.; Jacobi von Wangelin, A. Chem.Eur. J.
2014, 20, 6828. (e) Montgomery, J. Organonickel Chemistry. In
Organometallics in Synthesis; Lipshutz, B. H., Ed.; Wiley; Hoboken, NJ,
2013; pp 319−428.
(2) For cascade reductive coupling reactions of aryl electrophiles, see:
Durandetti, M.; Hardou, L.; Lhermet, R.; Rouen, M.; Maddaluno, J.
Chem.Eur. J. 2011, 17, 12773.
(3) For an elegant isolated example of a cyclization of alkyl halides
followed by C−C bond formation, see: Zhao, C.; Jia, X.; Wang, X.;
Gong, H. J. Am. Chem. Soc. 2014, 136, 17645.
(4) (a) Moragas, T.; Cornella, J.; Martin, R. J. Am. Chem. Soc. 2014,
136, 17702. (b) Liu, Y.; Cornella, J.; Martin, R. J. Am. Chem. Soc. 2014,
136, 11212. (c) Correa, A.; Leon, T.; Martin, R. J. Am. Chem. Soc.
2014, 136, 1062. (d) Leon, T.; Correa, A.; Martin, R. J. Am. Chem. Soc.
2013, 135, 1221. (e) Correa, A.; Martin, R. J. Am. Chem. Soc. 2009,
131, 15974.
(5) (a) Nogi, K.; Fujihara, T.; Terao, J.; Tsuji, Y. Chem. Commun.
2014, 50, 13052. (b) Tran-Vu, H.; Daugulis, O. ACS Catal. 2013, 3,
2417. (c) Fujihara, T.; Nogi, K.; Xu, T.; Terao, J.; Tsuji, Y. J. Am.
Chem. Soc. 2012, 134, 9106.
stereochemical course of 6 (Scheme 3, top). As shown, careful
1H NMR spectroscopic analysis revealed that the reaction
exclusively afforded 7,15 an observation that is consistent with a
scenario consisting of an initial oxidative addition with
inversion of configuration.26,27 Next, we turned our attention
to the reactivity of air-sensitive 8, which is easily accessible
simply by reacting Ni(COD)2 with L6 in tetrahydrofuran
(Scheme 3, bottom).28 Importantly, while no reaction took
place upon exposure of 8 to either 1i or 4b in the absence of
C
DOI: 10.1021/jacs.5b03340
J. Am. Chem. Soc. XXXX, XXX, XXX−XXX

Journal of the American Chemical Society
Communication
(6) For selected reviews, see: (a) Zhang, L.; Hou, Z. Chem. Sci. 2013,
4, 3395. (b) Tsuji, Y.; Fujihara, T. Chem. Commun. 2012, 48, 9956.
(c) Cokoja, M.; Bruckmeier, C.; Rieger, B.; Herrmann, W. A.; Kuhn, F.
E. Angew. Chem., Int. Ed. 2011, 50, 8510. (d) Huang, K.; Sun, C.-L.;
Shi, Z.-J. Chem. Soc. Rev. 2011, 40, 2435. (e) Martin, R.; Kleij, A. W.
ChemSusChem 2011, 4, 1259. (f) Sakakura, T.; Choi, J. C.; Yasuda, H.
Chem. Rev. 2007, 107, 2365.
(7) For selected examples, see: (a) Correa, A.; Martin, R. Angew.
Chem., Int. Ed. 2009, 48, 6201. (b) Ukai, K.; Aoki, M.; Takaya, J.;
Iwasawa, N. J. Am. Chem. Soc. 2006, 128, 8706. (c) Takaya, J.; Tadami,
S.; Ukai, K.; Iwasawa, N. Org. Lett. 2008, 10, 2697. (d) Ohishi, T.;
Nishiura, M.; Hou, Z. Angew. Chem., Int. Ed. 2008, 47, 5792.
(e) Yeung, C. S.; Dong, V. M. J. Am. Chem. Soc. 2008, 130, 7826.
(8) (a) Patai, S. The Chemistry of Acid Derivatives; Wiley: New York,
1992. (b) Goossen, L. J.; Rodríguez, N.; Goossen, K. Angew. Chem.,
Int. Ed. 2008, 47, 3100. (c) Maag, H. Prodrugs of Carboxylic Acids;
Springer: New York, 2007.
Hall, R. E.; Haley, A. D.; Brandon, R. J.; Konovalova, T.; Desrochers,
P. J.; Pulay, P.; Vicic, D. A. J. Am. Chem. Soc. 2006, 128, 13175.
(f) Ciszewski, J. T.; Mikhaylov, D. Y.; Holin, K. V.; Kadirov, M. K.;
Budnikova, Y. H.; Sinyashin, O.; Vicic, D. A. Inorg. Chem. 2011, 50,
8630.
(23) (a) Galli, C.; Gentili, P.; Guarnieri, A.; Zvi, R. J. Org. Chem.
1996, 61, 8878. (b) Jenkins, P. R.; Symons, M. C. R.; Booth, S. E.;
Swain, C. J. Tetrahedron Lett. 1992, 33, 3543. (c) Fukuyama, T.;
Nishitani, S.; Inouye, T.; Morimoto, K.; Ryu, I. Org. Lett. 2006, 8,
1383. (d) Martinez-Grau, A.; Curran, D. P. Tetrahedron 1997, 53,
5679. (e) Reference 10c.
(24) For a mechanistic rationale, see ref 15. The intermediacy of
radical species gains credence from the observation that the
carboxylation of iodo(2-methoxycyclopentylidene)methylbenzene
(Z/E = 2:1) produced a single carboxylic acid in 22% yield with L5
as the ligand (E/Z > 95:5). At present, we cannot rule out an E/Z
isomerization of well-defined Ni(II) species. For example, see:
(a) Huggins, J. M.; Bergman, R. G. J. Am. Chem. Soc. 1981, 103,
3002. (b) Conn, R. S. E.; Karras, M.; Snider, B. B. Isr. J. Chem. 1984,
24, 108. (c) Reference 2.
(25) We found that radical scavengers such as TEMPO, among
others, did not inhibit the carboxylation of 1b. Interestingly, an
otherwise similar experiment employing 4b as the substrate gave little
conversion, if any, to 5b/5b′. These results reinforce the notion that
single-electron transfer (SET) processes come into play when dealing
with secondary alkyl halides.
(9) For an excellent review of tandem catalysis, see: Fogg, D. E.; dos
Santos, E. N. Coord. Chem. Rev. 2004, 248, 2365.
(10) (a) Kambe, N.; Iwasaki, T.; Terao, J. Chem. Soc. Rev. 2011, 40,
4937. (b) Jana, R.; Pathak, T. P.; Sigman, M. S. Chem. Rev. 2011, 111,
1417. (c) Rudolph, A.; Lautens, M. Angew. Chem., Int. Ed. 2009, 48,
2656.
(11) For selected catalytic cyclization techniques of unactivated alkyl
halides using nucleophile/electrophile regimes, see: (a) Venning, A. R.
O.; Bohan, P. T.; Alexanian, E. J. J. Am. Chem. Soc. 2015, 137, 3731.
(b) Cong, H.; Fu, G. C. J. Am. Chem. Soc. 2014, 136, 3788.
(c) Fruchey, E. R.; Monks, B. M.; Patterson, A. M.; Cook, S. P. Org.
Lett. 2013, 15, 4362. (d) Monks, B. M.; Cook, S. P. J. Am. Chem. Soc.
2012, 134, 15297. (e) Bloome, K. S.; McMahen, R. L.; Alexanian, E. J.
J. Am. Chem. Soc. 2011, 133, 20146. (f) Kim, H.; Lee, C. Org. Lett.
2011, 13, 2050. (g) Firmansjah, L.; Fu, G. C. J. Am. Chem. Soc. 2007,
129, 11340.
(12) Yoder, R. A.; Johnston, J. N. Chem. Rev. 2005, 105, 4730.
(13) Flynn, A. B.; Ogilvie, W. W. Chem. Rev. 2007, 107, 4698.
(14) For selected cyclization of well-defined alkyl organometallics
with alkyne side chains, see: (a) Kambe, N.; Moriwaki, Y.; Fujii, Y.;
Iwasaki, T.; Terao, J. Org. Lett. 2011, 13, 4656. (b) Rao, S. A.; Knochel,
P. J. Am. Chem. Soc. 1991, 113, 5735. (c) Bailey, W. F.; Ovaska, T. V.;
Leipert, T. K. Tetrahedron Lett. 1989, 30, 3901. (d) Bailey, W. F.;
Ovaska, T. V. J. Am. Chem. Soc. 1983, 115, 3080. (e) Richey, H. G., Jr.;
Rothman, A. M. Tetrahedron Lett. 1968, 9, 1457.
(26) For related isotopic-labeling studies with Pd catalysts see ref 10c
and: (a) Netherton, M. R.; Fu, G. C. Angew. Chem., Int. Ed. 2002, 41,
3910. (b) Stokes, B. J.; Opra, S. M.; Sigman, M. S. J. Am. Chem. Soc.
2012, 134, 11408.
(27) Such a labeling pattern differs from previous work without
pendant alkynes in the side chain (ref 4b). This observation is
consistent with the ability of alkynes to act as noninnocent ancillary
ligands. See ref 11d for an example of such behavior.
(28) Powers, D. C.; Anderson, B. L.; Nocera, D. G. J. Am. Chem. Soc.
2013, 135, 18876.
(29) Vinylnickel(I) species have been proposed to be generated from
vinylnickel(II) by SET. For example, see ref 5c and: (a) Dunach, E.;
̃
Esteves, A. P.; Medeiros, M. J.; Olivero, S. New J. Chem. 2005, 29, 633.
(b) Fujihara, T.; Horimoto, Y.; Mizoe, T.; Sayyed, F. B.; Tani, Y.;
Terao, J.; Sakaki, S.; Tsuji, Y. Org. Lett. 2014, 16, 4960. (c) Nadal, M.
L.; Bosch, J.; Vila, J. M.; Klein, G.; Ricart, S.; Moreto,
Chem. Soc. 2005, 127, 10476.
́
J. M. J. Am.
(15) See the Supporting Information for details.
(16) Homodimerization accounts for the mass balance.
(17) (a) Galan, B. R.; Rovis, T. Angew. Chem., Int. Ed. 2009, 48, 2830.
(b) Gandon, V.; Aubert, C.; Malacria, M. Chem. Commun. 2006, 2209.
(c) Chopade, P. R.; Louie, J. Adv. Synth. Catal. 2006, 348, 2307.
(18) (a) Hartwig, J. F. Organotransition Metal Chemistry: From
Bonding to Catalysis; University Science Books: Mill Valley, CA, 2010;
p 379. (b) Marek, I.; Minko, Y. In Metal-Catalyzed Cross-Coupling
Reactions and More; de Meijere, A., Brase, S., Oestreich, M., Eds.;
̈
Wiley-VCH: Weinheim, Germany, 2014; Chapter 10, pp 763−874.
(19) For selected anti-carbometalation techniques using stoichio-
́
metric organometallic species, see: (a) Fressigne, C.; Girard, A.-L.;
Durandetti, M.; Maddaluno, J. Angew. Chem., Int. Ed. 2008, 47, 891.
(b) Simaan, S.; Marek, I. Org. Lett. 2007, 9, 2569. (c) Ma, S.; Negishi,
E.-I. J. Org. Chem. 1997, 62, 784. (d) Lu, Z.; Ma, S. J. Org. Chem. 2006,
71, 2655.
(20) At present, the intermediacy of “redox-noninnocent ligands”
cannot be ruled out. See: Hu, X. Chem. Sci. 2011, 2, 1867.
(21) Unfortunately, the use of tertiary alkyl halides resulted in
recovered starting material or reduced products.
(22) For the intermediacy of Ni species in odd oxidation states, see:
(a) Laskowski, C. A.; Bungum, D. J.; Baldwin, S. M.; Del Ciello, S. A.;
Iluc, V. M.; Hillhouse, G. L. J. Am. Chem. Soc. 2013, 135, 18272.
(b) Breitenfeld, J.; Ruiz, J.; Wodrich, M. D.; Hu, X. J. Am. Chem. Soc.
2013, 135, 12004. (c) Biswas, S.; Weix, D. J. J. Am. Chem. Soc. 2013,
135, 16192. (d) Schley, N. D.; Fu, G. C. J. Am. Chem. Soc. 2014, 136,
16588. (e) Jones, G. D.; Martin, J. L.; McFarland, C.; Allen, O. R.;
D
DOI: 10.1021/jacs.5b03340
J. Am. Chem. Soc. XXXX, XXX, XXX−XXX