Co-Catalyzed Direct Addition of Allylic C(sp3)-H Bonds to Ketones

Article abstract of DOI:10.1021/acs.orglett.7b02871

By using Co(acac)₂/Xantphos with AlMe₃, the C(sp³)-H bonds of allylarene derivatives were cleaved for reaction with various ketones, affording the homoallylic alcohols in moderate to good yields. The branch/linear selectivity depended on the steric and electronic factors of the ketone electrophiles. The intermediate in this reaction is thought to be a low-valent allylcobalt(I) species, which exhibits high nucleophilicity toward ketones.

Full text of DOI:10.1021/acs.orglett.7b02871

Letter
pubs.acs.org/OrgLett
Cite This: Org. Lett. XXXX, XXX, XXX-XXX
Co-Catalyzed Direct Addition of Allylic C(sp3)−H Bonds to Ketones
Tsuyoshi Mita,* Satoshi Hanagata, Kenichi Michigami, and Yoshihiro Sato*
Faculty of Pharmaceutical Sciences, Hokkaido University, Sapporo 060-0812, Japan
S
* Supporting Information
ABSTRACT: By using Co(acac)2/Xantphos with AlMe3, the C(sp3)−H bonds of allylarene derivatives were cleaved for reaction
with various ketones, affording the homoallylic alcohols in moderate to good yields. The branch/linear selectivity depended on
the steric and electronic factors of the ketone electrophiles. The intermediate in this reaction is thought to be a low-valent
allylcobalt(I) species, which exhibits high nucleophilicity toward ketones.
he C−H activation reaction has become a powerful tool,
umpolung under reductive conditions cannot be achieved in the
Tespecially for the construction of C−C bonds directly from samepot. Totacklethisproblem,adiboronreagentwasemployed
for the generation of a nucleophilic allylboron under oxidative
conditions, which allows the reaction of carbonyl electrophiles
(eq 2).9 However, direct conversion from analkene byC(sp3)−H
activation without the generation of a reactive intermediate
(boron/silane/stannane) would be a powerful strategy that is
preferable in terms of atom and step economy (eq 3).
poorly reactive C−H bonds (the pKa values are usually more than
ca. 30) without using a strong base or acid.1 Of particular interest
is the functionalization of C(sp3)−H bonds2 because of its high
potential for the synthesis of 3D complex molecules from simple
aliphatic compounds. In this regard, polar carbonyl electrophiles
have been less exploited,3 although there are many successful
examplesofC(sp2)−Hadditiontoaldehydes, aldimines, andtheir
equivalents, especially in the presence of Rh(III) and Co(III)
complexes as catalysts.4,5 We turned our attention to the allylic
C(sp3)−Hbondsadjacenttoalkenes, withtheaimofdevelopinga
new class of carbonyl insertion reactions to C(sp3)−H bonds.
Cleavage of the allylic C(sp3)−H bonds of alkenes using a
To develop a new electrophilic transformation of the allylic
C(sp3)−H bonds of alkenes, we considered generating a low-
valent η3-allylmetal complex. Alkylated transition-metal species
such as alkylcobalt(I) and alkylrhodium(I) were reported to
cleave the allylic C−H bond of alkenes to generate an η3-
allylmetal(I) complex with the release of an alkane.10 However,
the reactivities of these complexes were not fully studied, which
attracted our initial interest. After considerable experimentation,
allylcobalt(I) was found to show high nucleophilicity toward
CO2, a poorly reactive carbonyl electrophile.11 A combination of
Co(acac)2/Xantphos with AlMe3 led to smooth allylic carbox-
ylation of alkenes with CO2, affording the corresponding β,γ-
unsaturated carboxylic acids in high yields with high linear
selectivities. Encouraged by this successful carboxylation, we
focused on the development of direct addition12 to ketones,
which have not been commonly employed in C−H activation
chemistry due to the low electrophilicity and potential
reversibility.13 We herein reveal a direct nucleophilic addition of
allylic C(sp3)−H bonds to various ketone electrophiles.
Pd(II) salt was first reported by Huttel,6a,b and the reactive
̈
intermediate wasemployed forallylic alkylation byTrost.6c White
further extended this stoichiometric activation to a catalytic
variant by using a stoichiometric amount of the oxidant and a
sulfoxide ligand.7 Since the generated η3-allylpalladium(II)
complex exhibited high reactivity toward a variety of nucleophiles
(Scheme 1, eq 1), it cannot be allowed to react with electrophiles.
Umpolung of the electrophilic η3-allylpalladium(II) complex was
established by using various reductants or alkylmetal reagents
(Zn, InI, SnCl2, SmI2, ZnEt2, or BEt3).8 Yet, catalytic cleavage of
C(sp3)−H bonds by Pd(II) under oxidative conditions and
Scheme 1. Previous Works
First, the reaction parameters were investigated using methyl-
4-allylbenzoate (1a) and acetone (2a) in the presence of various
transition-metal catalysts and an alkylmetal source. According to
our previously developed allylic carboxylation with CO2,11
Co(acac)2 (10 mol %), Xantphos (20 mol %), and AlMe3 (1.5
equiv) were employed for the reaction of allylarene 1a and 2a in
DMF (Table 1). The corresponding homoallylic alcohol 3aa was
obtained in 59% yield with modest branch selectivity (b/l = 86/
Received: September 13, 2017
© XXXX American Chemical Society
A
DOI: 10.1021/acs.orglett.7b02871
Org. Lett. XXXX, XXX, XXX−XXX

Organic Letters
Letter
Table 1. Screening of Reaction Conditions
a
yield (%)
entry
metal salt
solv
DMF
MRn
3aa (b/l)
4a
5a
rec 1a
1
2
Co(acac)2
Co(acac)3
Co(acac)2
Rh(acac)3
[Rh(cod)Cl]2
Co(acac)2
Co(acac)2
Co(acac)2
Co(acac)2
Co(acac)2
Co(acac)2
Co(acac)2
AlMe3
AlMe3
AlMe3
AlMe3
AlMe3
AlEt3
59 (86/14)
59 (86/14)
29 (69/31)
−
−
−
11 (−/100)
−
−
51 (80/20)
49 (86/14)
75 (83/17)
75 (81/19)
28
27
29
−
42
63
25
71
88
37
23
15
16
12
13
8
−
−
3
DMF
DMF
DMF
DMF
DMF
DMF
DMF
DMF
THF
b
3
4
5
−
−
−
2
90
58
−
54
−
−
2
c
6
7
ZnMe2
ZnEt2
BEt3
d
8
e
−
−
10
18
10
8
9
10
11
12
AlMe3
AlMe3
AlMe3
AlMe3
1,4-dioxane
DMA
DMA
−
−
−
f
13
Co(acac)2
a
1
Yields were determined by H NMR analysis using 1,3,5-trimethoxybenzene as an internal standard. The ratio of branched to linear products is
given in parentheses. 1.0 equiv of CsF was added. 8% of methyl 4-propylbenzoate was observed. 29% of methyl 4-propylbenzoate was observed.
4% of methyl 4-propylbenzoate was observed. 1.0 equiv of AlMe3 was used.
b
c
d
e
f
14), alongwiththeolefinisomerizationproduct4a(E/Zmixture)
and methylated product 5a (entry 1). When Co(acac)3 was
employed instead of Co(acac)2, a similar product distribution was
observed (entry 2). The addition of CsF, which was effective for
carboxylation, did not work well in this case (entry 3). The use of
RhcatalystssuchasRh(acac)3 and[Rh(cod)Cl]2 didnotpromote
the desired carbonyl allylation (entries 4 and 5). Other alkylmetal
reagents such as AlEt3, ZnMe2, ZnEt2, and BEt3 were not
promising (entries 6−9), and a small amount of hydrogenated
product, methyl 4-propylbenzoate, was observed when MEt3 (M
= Al, Zn, and B) was employed. Screening of potential solvents
such as THF, 1,4-dioxane, and DMA revealed that DMA was the
best in terms of the yield of 3aa (entries 10−12). Eventually, the
amount of AlMe3 was reduced to 1.0 equiv, and the yield of 3aa
was maintained at 75% (entry 13).
With the optimal conditions (Co(acac)2 (10 mol %), Xantphos
(20 mol %), and AlMe3 (1.0 equiv) in DMA at 60 °C) in hand, we
investigated the substrate scope and limitations of ketone
electrophiles (Figure 1). The carbonyl character of the employed
ketones was elucidated by the CO stretch in the IR spectrum.
The reaction was applicable to not only acetone (2a) but also
other aliphatic ketones, including 2-butanone (2b) and 3-
pentanone (2c). However, the amount of linear products
increased as the alkyl chain was elongated. Since these three
ketones appeared to be equally reactive based on the CO
stretch (2a: 1715, 2b:1717, and 2c: 1716 cm−1), the difference in
their regioselectivities mainly depended on steric bulkiness. In
contrast, aromatic ketones such as acetophenone (2d) and
benzophenone (2e), which are much bulkier than aliphatic
ketones, resulted inselectivegenerationofthelinearproduct. The
double bond character of 2d (1686 cm−1) and 2e (1659 cm−1)
was lower than that of 2a (1715 cm−1) because of the conjugation
effect from the aromatic ring. To exclude the steric factor of
electrophiles, substituted acetophenones 2f−2h were examined:
electron-withdrawing substituents favored the formation of
Figure 1. Substrate scope and limitations. Isolated yields are shown. a5
equiv of ketones was employed. bWith 1 mmol of 1a.
branched products, whereas electron-donating substituents
selectively promoted linear addition. This trend was in
accordance with the strength of the CO bonds (2f: 1703, 2g:
1694, and 2h: 1675 cm−1). In the case of cyclic ketones, the linear
selectivity increased with an increase in the ring size. In these
cases, as the strength of the CO bond (2i: 1783; 2j: 1746; 2k:
1715; 2l: 1701; 2m: 1698 cm−1) was reduced by ring expansion,
the ratio of the linear product gradually increased. On the basis of
the structural bulkiness and the CO stretching of 2, we
concludedthatstericallycompactand/orreactiveketones(higher
doublebondcharacter) suchasacetone(2a), cyclobutanone (2i),
B
DOI: 10.1021/acs.orglett.7b02871
Org. Lett. XXXX, XXX, XXX−XXX

Organic Letters
Letter
and cyclopentanone (2j) were favored for branch selectivity,
whereas sterically crowded ketones were favored for linear
selectivity.
Next, several allylarenes were investigated using acetone (2a)
and acetophenone (2d) (Figure 2). High branch selectivity was
generation of the linear isomer. Since enantioselective C−H
activation is still limited,15 there is much potential for further
research in this direction. This result also suggested that the
allylcobalt species bearing a diphosphine ligand, and not an
allylaluminum species, was the actual intermediate for ketone
addition.
A proposed mechanism for this reaction is depicted in Figure 3.
First, a low-valent methylcobalt(I) species would be generated
Figure 2. Substrate scope and limitations. Isolated yields are shown.
observed in the reaction of arylarenes 1a, 1b, 1c, and 1d with 2a.
Conversely, the reaction with 2d led to a dramatic change in the
regioselectivity to yield linear products predominantly, although
the diastereoselectivities of 3bd−3dd were not very high.
The C(sp3)−H addition to ketones was expanded to an
intramolecular variant using ketones 6a and 6b. Under the
optimal conditions, cyclized products 7a and 7b were obtained in
63% and 72% yields, respectively, without generation of
intermolecular products (Scheme 2). The structure of the
major diastereomer was determined to be syn by NOE
spectroscopy.14
Figure 3. Plausible catalytic cycle.
from a cobalt salt and AlMe3.16 Dimethylcobalt(II) is known to
undergo disproportionation to form methylcobalt(I) and
trimethylcobalt(III),17 and the latter is further reduced to
methylcobalt(I) via reductive elimination. Next, the allylic
C(sp3)−H bond of an alkene is oxidatively added to
methylcobalt(I) to generate intermediate I, which would be
under equilibrium to generate II and III. Intermediate II, which
can be stabilized by an aryl substituent,18 would be further
reduced to form low-valent η1-allylcobalt(I) IV along with the
release of methane. Subsequently, C−C bond formation with
compact and/or reactive ketones would proceed preferentially at
the α-position to produce cobalt alkoxide V.19,20 Finally, V is
transmetalated with AlMe3 to afford aluminum alkoxide VI with
regeneration of the methylcobalt(I) species. In contrast, when
sterically demanded ketones are used, γ-addition8,11−13 of IV is
preferable due to steric repulsion between the ketone and the
Xantphos ligand, affording cobalt alkoxide VII. Olefin isomer-
ization product 4 and methylated product 5 would be generated
through the reductive elimination of III. The oxygen atom in
Xantphos plays a crucial role in coordinating to cobalt, which
would assist the tautomerization of η3-allylcobalt I to η1-
allylcobalt complex II.21,22
Scheme 2. Intramolecular Cyclization
Finally, we conducted the reaction of 1b with acetone (2a) in
the presence of a chiral ligand (Scheme 3). Screening of several
potential chiral ligands revealed that (R)-DIFLUORPHOS was
the most promising in terms of enantioselectivity. Branched 3ba
was obtained in 25% yield, with 81% ee, in DMF without the
Scheme 3. Enantioselective Variant Using Chiral Ligands
In conclusion, we have successfully developed the first cobalt-
catalyzed nucleophilic addition of the C(sp3)−H bonds of
allylarene derivatives to ketones. We are now conducting
computational studies to reveal the detailed mechanism for
explaining the observed regioselectivity. Much effort toward the
development of an asymmetric variant is also ongoing, and the
results will be reported in due course.
C
DOI: 10.1021/acs.orglett.7b02871
Org. Lett. XXXX, XXX, XXX−XXX

Organic Letters
Letter
(6) (a) Huttel, R.; Kratzer, J.; Bechter, M. Chem. Ber. 1961, 94, 766.
̈
ASSOCIATED CONTENT
* Supporting Information
■
(b) Huttel, R.; Christ, H. Chem. Ber. 1963, 96, 3101. (c) Trost, B. M.;
̈
S
Fullerton, T. J. J. Am. Chem. Soc. 1973, 95, 292.
TheSupportingInformationisavailablefreeofchargeontheACS
Publications website at DOI: 10.1021/acs.orglett.7b02871.
(7) Recent examples of Pd(II)-catalyzed allylic C(sp3)−H functional-
izations: (a) Chen, M. S.; White, M. C. J. Am. Chem. Soc. 2004, 126, 1346.
(b) Young, A. J.; White, M. C. J. Am. Chem. Soc. 2008, 130, 14090.
(c) Kondo, H.; Yu, F.; Yamaguchi, J.; Liu, G.;Itami, K. Org. Lett. 2014, 16,
4212. (d) Ammann, S. E.; Liu, W.; White, M. C. Angew. Chem., Int. Ed.
2016, 55, 9571. For Rh(III) catalysis, see: (e) Cochet, T.; Bellosta, V.;
Roche, D.; Ortholand, J.-Y.; Greiner, A.; Cossy, J. Chem. Commun. 2012,
48, 10745. (f)Shibata, Y.;Kudo, E.;Sugiyama, H.;Uekusa, H.;Tanaka, K.
Organometallics 2016, 35, 1547 and references therein.
Experimental details and characterization data (PDF)
AUTHOR INFORMATION
Corresponding Authors
■
*E-mail: tmita@pharm.hokudai.ac.jp.
*E-mail: biyo@pharm.hokudai.ac.jp.
ORCID
(8) Reviews of nucleophilic η1-allylpalladium species: (a) Tamaru, Y. J.
Organomet. Chem. 1999, 576, 215. (b) Marshall, J. A. Chem. Rev. 2000,
́ ́
100, 3163. (c) Szabo, K. J. Chem. - Eur. J. 2004, 10, 5268. (d) Szabo, K. J.
Tsuyoshi Mita: 0000-0002-6655-3439
Yoshihiro Sato: 0000-0003-2540-5525
Notes
Synlett 2006, 2006, 811. (e) Zanoni, G.; Pontiroli, A.; Marchetti, A.;
Vidari, G. Eur. J. Org. Chem. 2007, 2007, 3599. Our recent achievements
of palladium-catalyzed allylic carboxylation using ZnEt2: (f) Mita, T.;
Higuchi, Y.; Sato, Y. Chem. - Eur. J. 2015, 21, 16391. (g) Mita, T.; Tanaka,
H.; Higuchi, Y.; Sato, Y. Org. Lett. 2016, 18, 2754. (h) Higuchi, Y.; Mita,
T.; Sato, Y. Org. Lett. 2017, 19, 2710.
The authors declare no competing financial interest.
(9) (a) Tao, Z.-L.; Li, X.-H.; Han, Z.-Y.; Gong, L.-Z. J. Am. Chem. Soc.
ACKNOWLEDGMENTS
■
2015, 137, 4054. Iridium-catalyzed allylic C(sp3)−H borylation
This work was financially supported by Grant-in-Aid for Scientific
Research (C) (No. 26410108), Grant-in-Aid for Scientific
Research (B) (No. 26293001) from JSPS, and also by JST
ACT-C (No. JPMJCR12YM). K.M. thanks JSPS for a fellowship
(No. 14J08052).
followed by carbonyl addition: (b) Olsson, V. J.; Szabo,
́
K. J. Angew.
Chem., Int. Ed. 2007, 46, 6891. (c) Olsson, V. J.; Szabo,
2009, 74, 7715.
(10) (a) Klein, H.-F.; Helwig, M.; Braun, S. Chem. Ber. 1994, 127, 1563.
(b) Urtel, H.; Meier, C.; Eisentrager, F.; Rominger, F.; Joschek, J. P.;
́
K. J. J. Org. Chem.
̈
Hofmann, P. Angew. Chem., Int. Ed. 2001, 40, 781. (c) Beck, R.; Florke,
̈
U.; Klein, H.-F. Organometallics 2015, 34, 1454.
REFERENCES
■
(11) Michigami, K.; Mita, T.; Sato, Y. J. Am. Chem. Soc. 2017, 139, 6094.
(12) Base-catalyzed allylic functionalization: Bao, W.; Kossen, H.;
Schneider, U. J. Am. Chem. Soc. 2017, 139, 4362.
(13) A recent achievement for ketone addition: Wei, X.-F.; Xie, X.-W.;
Shimizu, Y.; Kanai, M. J. Am. Chem. Soc. 2017, 139, 4647.
(1) Selected reviews on C−H activation: (a) Lyons, T. W.; Sanford, M.
S. Chem. Rev. 2010, 110, 1147. (b) Wendlandt, A. E.; Suess, A. M.; Stahl,
S. S. Angew. Chem., Int. Ed. 2011, 50, 11062. (c) Guo, X.-X.; Gu, D.-W.;
Wu, Z.; Zhang, W. Chem. Rev. 2015, 115, 1622. (d) Hartwig, J. F.; Larsen,
M. A. ACS Cent. Sci. 2016, 2, 281. (e) Gensch, T.; Hopkinson, M. N.;
Glorius, F.; Wencel-Delord, J. Chem. Soc. Rev. 2016, 45, 2900.
(14) See the Supporting Information (SI).
(15) Reviews on enantioselective C−H functionalization: (a) Engle, K.
M.; Yu, J.-Q. J. Org. Chem. 2013, 78, 8927. (b) Newton, C. G.; Wang, S.-
G.; Oliveira, C. C.; Cramer, N. Chem. Rev. 2017, 117, 8908.
(16) (a) Komiya, S.; Bundo, M.; Yamamoto, T.; Yamamoto, A. J.
Organomet. Chem. 1979, 174, 343. (b) Mohtachemi, R.; Kannert, G.;
Schumann, H. J. Organomet. Chem. 1986, 310, 107. (c) Rudkin, M. E.;
Joensuu, P. M.; MacLachlan, W. S.; Lam, H. W. Org. Lett. 2008, 10, 2939.
(17) Bryndza, H. E.; Bergman, R. G. J. Am. Chem. Soc. 1979, 101, 4766.
(2) Selected reviews on C(sp3)−H activation: (a) Jazzar, R.; Hitce, J.;
Renaudat, A.; Sofack-Kreutzer, J.; Baudoin, O. Chem. - Eur. J. 2010, 16,
2654. (b) Dastbaravardeh, N.; Christakakou, M.; Haider, M.; Schnurch,
M. Synthesis 2014, 46, 1421. (c) He, J.; Wasa, M.;Chan, K. S. L.; Shao, Q.;
Yu, J.-Q. Chem. Rev. 2017, 117, 8754.
̈
(3)(a)Qian, B.;Guo, S.;Shao, J.;Zhu, Q.;Yang, L.;Xia, C.;Huang, H. J.
Am. Chem. Soc. 2010, 132, 3650. (b) Qian, B.; Guo, S.; Xia, C.; Huang, H.
Adv. Synth. Catal. 2010, 352, 3195. (c) Komai, H.; Yoshino, T.;
Matsunaga, S.; Kanai, M. Org. Lett. 2011, 13, 1706. (d) Rueping, M.;
Tolstoluzhsky, N. Org. Lett. 2011, 13, 1095. (e) Qian, B.; Xie, P.; Xie, Y.;
Huang, H. Org. Lett. 2011, 13, 2580. (f)Best, D.;Kujawa, S.;Lam, H. W. J.
Am. Chem. Soc. 2012, 134, 18193. (g) Komai, H.; Yoshino, T.;
Matsunaga, S.; Kanai, M. Synthesis 2012, 44, 2185. (h) Wu, X.; Zhao,
Y.; Ge, H. J. Am. Chem. Soc. 2015, 137, 4924.
(4) Reviews on transition-metal-catalyzed direct additions of C−H
bonds to carbon−heteroatom multiple bonds: (a) Yan, G.; Wu, X.; Yang,
M. Org. Biomol. Chem. 2013, 11, 5558. (b)Zhang, X.-S.; Chen, K.;Shi, Z.-
J. Chem. Sci. 2014, 5, 2146. (c) Hummel, J. R.; Boerth, J. A.; Ellman, J. A.
Chem. Rev. 2017, 117, 9163.
(18) (a) Galamb, V.; Pal
(b) Zucchi, C.; Palyi, G.; Galamb, V.; Sam
Li, P.; Alper, H. Organometallics 1996, 15, 3222.
́
yi, G. J. Chem. Soc., Chem. Commun. 1982, 487.
́
́
par-Szerencses, E.; Marko, L.;
́
́
́
(19) Since the allylcobalt(I)−Xantphos IV complex is thought to have
an electron-rich cobalt center, the back-donation of d-orbitals of
cobalt(I) to π* of CO of reactive ketones would be expected.
(20) A low-valent allylcobalt species generated from allylacetate and
CoBr2 in the presence of Zn smoothly reacted with ketones: Gomes, P.;
́
Gosmini, C.; Perichon, J. Synthesis 2003, 2003, 1909.
(21) Xantphos in a Rh(I) complex worked as a POP-type pincer ligand:
Ren, P.; Pike, S. D.; Pernik, I.; Weller, A. S.; Willis, M. C. Organometallics
2015, 34, 711.
(5) Selected examples on Rh(III)- and Co(III)-catalyzed nucleophilic
additions of C(sp2)−H bonds to aldehydes, imines, and isocyanates:
(a)Tsai, A. S.; Tauchert, M. E.; Bergman, R. G.; Ellman, J. A. J. Am. Chem.
Soc. 2011, 133, 1248. (b) Yang, L.; Correia, C. A.; Li, C.-J. Adv. Synth.
Catal. 2011, 353, 1269. (c) Park, J.; Park, E.; Kim, A.; Lee, Y.; Chi, K.-W.;
Kwak, J. H.; Jung, Y. H.; Kim, I. S. Org. Lett. 2011, 13, 4390. (d) Hesp, K.
D.; Bergman, R. G.; Ellman, J. A. J. Am. Chem. Soc. 2011, 133, 11430.
(e) Tauchert, M. E.; Incarvito, C. D.; Rheingold, A. L.; Bergman, R. G.;
Ellman, J. A. J. Am. Chem. Soc. 2012, 134, 1482. (f) Yoshino, T.; Ikemoto,
H.; Matsunaga, S.; Kanai, M. Angew. Chem., Int. Ed. 2013, 52, 2207.
(g) Wangweerawong, A.; Bergman, R. G.; Ellman, J. A. J. Am. Chem. Soc.
2014, 136, 8520. (h) Hummel, J. R.; Ellman, J. A. J. Am. Chem. Soc. 2015,
137, 490. Co(I)-catalyzed a nucleophilic addition of 2-phenylpyridines
to aldimines: (i) Gao, K.; Yoshikai, N. Chem. Commun. 2012, 48, 4305.
(22) Other Co/Xantphos-catalyzed transformations: (a) Jin, M.-Y.;
Yoshikai, N. J. Org. Chem. 2011, 76, 1972. (b) Ding, Z.; Yoshikai, N.
Synthesis 2011, 2561. (c) Wang, C.; Teo, W. J.; Ge, S. ACS Catal. 2017, 7,
855.
D
DOI: 10.1021/acs.orglett.7b02871
Org. Lett. XXXX, XXX, XXX−XXX