Arylation of hydrocarbons enabled by organosilicon reagents and weakly coordinating anions

Article abstract of DOI:10.1126/science.aam7975

Over the past 80 years, phenyl cation intermediates have been implicated in a variety of C-H arylation reactions. Although these examples have inspired several theoretical and mechanistic studies, aryl cation equivalents have received limited attention in organic methodology. Their high-energy, promiscuous reactivity profiles have hampered applications in selective intermolecular processes. We report a reaction design that overcomes these challenges. Specifically, we found that b-silicon-stabilized aryl cation equivalents, generated via silylium-mediated fluoride activation, undergo insertion into sp3 and sp2 C-H bonds. This reaction manifold provides a framework for the catalytic arylation of hydrocarbons, including simple alkanes such as methane. This process uses low loadings of Earth-abundant initiators (1 to 5 mole percent) and occurs under mild conditions (30° to 100°C).

Full text of DOI:10.1126/science.aam7975

RESEARCH
◥
We sought to discover a phenyl cation equiva-
lent that could be used in catalytic, intermolecular
C–H arylation reactions. We were inspired by
seminal reports of analogous reactivity over the
past 80 years. In the 1930s, Mascarelli reported
that the thermolysis of aryl diazonium salt 1 led
to the formation of fluorene 2, presumably via
C–H insertion of a phenyl cation intermediate
(Fig. 1A) (12). More recently, reports from Albini’s
group (13, 14) suggested that photogenerated
phenyl cations participate in intramolecular C–H
insertion reactions; Siegel and co-workers (15, 16)
proposed the intermediacy of phenyl cations
in the silylium-mediated intramolecular Friedel-
Crafts reaction of aryl fluorides (Fig. 1B); and
Reed and co-workers (17) presented evidence
for incipient aryl cations, where the structure
of a phenyl halonium salt of undecachlorinated
monocarba-closo-dodecaborate anion (3; Fig. 1C)
was elucidated by x-ray crystallography (17).
Despite these fundamental breakthroughs, cata-
lytic intermolecular reactions of phenyl cation
equivalents have been elusive. Considering
these seminal efforts, as well as the frontier
molecular orbital analogy that can be made
with singlet carbene species (18, 19), we pursued
the application of phenyl cation equivalents
in catalytic, intermolecular C–H functional-
ization reactions such as those depicted in
Fig. 1C.
REPORT
ORGANIC CHEMISTRY
Arylation of hydrocarbons enabled by
organosilicon reagents and weakly
coordinating anions
Brian Shao,* Alex L. Bagdasarian,* Stasik Popov, Hosea M. Nelson†
Over the past 80 years, phenyl cation intermediates have been implicated
in a variety of C–H arylation reactions. Although these examples have inspired
several theoretical and mechanistic studies, aryl cation equivalents have received
limited attention in organic methodology. Their high-energy, promiscuous reactivity
profiles have hampered applications in selective intermolecular processes. We
report a reaction design that overcomes these challenges. Specifically, we found
that b-silicon–stabilized aryl cation equivalents, generated via silylium-mediated
fluoride activation, undergo insertion into sp3 and sp2 C–H bonds. This reaction
manifold provides a framework for the catalytic arylation of hydrocarbons,
including simple alkanes such as methane. This process uses low loadings of
Earth-abundant initiators (1 to 5 mole percent) and occurs under mild conditions
(30° to 100°C).
ince 1891, when Merling unknowingly pre-
pared an aromatic tropylium ion (1), carbo-
cations have played a substantial role in
the development of organic chemistry as
a field of scientific and practical endeav-
cations that are divalent, and/or are not stabilized
by resonance donating groups or hyperconju-
gation, are not easily manipulated or used in
routine transformations (5). Although exquis-
ite fundamental studies of these more reactive
species have revealed remarkable reactivity pro-
files (6), studies aimed at methodological advances
in this area are rare (7, 8). This is particularly
true for phenyl cations. Their reactive nature has
thwarted characterization in the condensed phase,
and even their existence as reactive intermediates
remains a matter of debate (9–11).
We envisioned that b-silylated aryl fluorides
(4a; Fig. 2A) would be particularly well suited
as phenyl cation precursors for several reasons:
(i) We anticipated that b-silicon stabilization
would lower the barrier for fluoride abstrac-
tion and temper the s-electrophilicity of the
S
or (2). Conceptually, carbenium ions serve as
retrons, guiding the design of retrosynthetic
analyses and elucidating the selection of synthet-
ic equivalents (3). Practically, carbocations are
equally important, as stabilized variants are rou-
tinely generated and used in standard synthetic
transformations (4). On the other hand, carbo-
Department of Chemistry and Biochemistry, University of
California, Los Angeles, CA 90095, USA.
*These authors contributed equally to this work. †Corresponding
author. Email: hosea@chem.ucla.edu
Fig. 1. Reactions involving putative phenyl cations. (A) Mascarelli’s reaction (12). (B) Siegel’s intramolecular Friedel-Crafts reaction of aryl fluorides (15).
(C) Our dual C–F/C–H functionalization strategy. R = ethyl or triisopropyl; R1 = aryl, alkyl, halide, or silyl ether; WCA, weakly coordinating anion; TMS,
trimethylsilyl; Mes, mesityl; Me, methyl; cat., catalytic.
Shao et al., Science 355, 1403–1407 (2017) 31 March 2017
1 of 5

RESEARCH
| REPORT
resulting phenyl cation equivalent 5 (20–24);
(ii) we envisioned that silicon substitution would
enhance the nucleophilicity of the arene p-system,
perhaps improving the insertion reactivity of
phenyl cation equivalent 5 (25); and (iii) we hy-
pothesized that elimination of the b-silicon group
from a reactive intermediate such as arenium 6
could regenerate the key reactive silicon species.
Taking into account these reaction design
concepts, we envisioned the cationic chain pro-
cess depicted in Fig. 2A. Here, a substoichiometric
silylium-carborane initiator (26), generated via
Bartlett-Condon-Schneider hydride transfer (27),
could abstract a fluoride from fluoroarene 4a to
generate aryl cation equivalent 5 (Fig. 2A). Sub-
sequent insertion into the hydrocarbon C–H
bond would then yield b-silicon–stabilized Wheland
intermediate 6. Elimination of trimethylsilylium
would then afford C–H arylation product 7.
This elementary step would generate the active
trimethylsilylium-carborane salt that proceeds
through the catalytic cycle (28).
Table 1. Scope of arylation reaction. Reactions were performed at 0.1 M fluoroarene in ben-
zene solvent. *Yield determined by gas chromatography–flame ionization detector (GC-FID) using
nonane as an internal standard. †Yield determined by NMR using an internal standard. ‡Isolated
yield. TBS, tert-butyldimethylsilyl; n-Bu, butyl; Et, ethyl.
We report the successful execution of this
methodological hypothesis, wherein a broad
scope of b-silylated aryl fluorides are shown
to be competent reagents for the arylation of
unactivated sp3 and sp2 C–H bonds, including
the characteristically inert bonds in methane
(Fig. 1C). After a short examination of reaction
conditions (table S1) (29), we found that exposure
of fluoride 4a to 1 mol % of [Ph3C]+ [HCB11Cl11]–
and 2 mol % of triethylsilane in benzene solvent
resulted in the facile formation of biphenyl (7a;
Fig. 2B) in 55% yield at 30°C in 1 hour. Appli-
cation of Siegel’s intramolecular reaction con-
ditions to fluorobenzene (4b; Fig. 2B) resulted only
in trace product 7a, and use of meta- and para-
silylated aryl fluorides 4c and 4d did not result
in product formation.
With this initial finding in hand, we inves-
tigated the scope of this arylation reaction. We
were pleased to observe selective C–F function-
alization in the presence of weaker carbon-halogen
(C–X) bonds, similar to previous reports from
Ozerov and co-workers (30). Specifically, entries
1 to 4 in Table 1 highlight the fluorophilicity of
the silylium-carborane catalyst. In these cases,
weaker C–X bonds that have less steric encum-
brance than the C–F bond do not undergo ion-
ization. This selectivity stands in stark contrast
to many traditional reactions of aryl halides,
where reactivity is often inversely proportional
to bond dissociation energy (31, 32). To further
investigate this unusual selectivity and to sep-
arate b-silicon effects from fluorophilicity, we
exposed (2-bromo-6-fluorophenyl)trimethylsilane
(Table 1, entry 5) to our reaction conditions.
Remarkably, m-bromobiphenyl was formed in
good yield, supporting our claim of halide se-
lectivity. In general, halide substitution was well
tolerated.
Polycyclic aromatic fluorides (Table 1, entry 6)
also were competent under the reaction condi-
tions, as demonstrated by the formation of 1-
phenylnaphthalene in 49% yield. Additionally,
aryl and alkyl substitution (Table 1, entries 7 to
10) were tolerated under the reaction conditions,
Shao et al., Science 355, 1403–1407 (2017) 31 March 2017
2 of 5

RESEARCH
| REPORT
Fig. 2. Catalytic design. (A) Proposed catalytic cycle. (B) Initial investigation of aryl fluorides. Ph, phenyl; rt, room temperature.
providing phenylated aromatics in moderate to
Table 2. sp3 C–H functionalization. Top: Arylation of cyclic and acyclic alkanes performed at 0.05 M
fluoroarene in alkane solvent. Bottom: Arylation of methane performed at 0.1 M fluoroarene in C6F6 solvent.
*Yield determined by GC-FID using nonane as an internal standard. †Yield determined by NMR using an
internal standard. ‡Reaction was performed in the absence of o-dichlorobenzene. i-Pr, isopropyl.
excellent yields (45 to 99%). Consecutive arylation
of difluorides was also possible, as demonstrated
by the formation of o-terphenyl (Table 1, entry
11), albeit in a diminished 36% yield. Finally,
the presence of heteroatom donor substituents,
typically incompatible with silylium catalysis
(Table 1, entry 12), provided 29% yield of the
desired phenol derivative. In cases where the
yields were moderate, we observed several by-
products, including fluoroarenes resulting from
protodesilylation of the starting materials (pre-
sumably resulting from highly acidic arenium
intermediates) as well as products resulting from
a second arylation event (29, 33).
Throughout our scope studies, some general
reactivity trends were apparent. Positional se-
lectivity was preserved in all cases, including
those with lower yields. Halide substituents
required higher reaction temperatures, whereas
alkyl and aryl substituents allowed faster, lower-
temperature arylation. These observations are
consistent with the intermediacy of a cationic
aryl species.
Bolstered by these results, we began our in-
vestigation into the arylation of alkanes. After
optimization of reaction conditions (table S2)
(29), we found that cyclohexane could be phenyl-
ated by aryl fluoride 4a in 41% yield (Table 2,
entry 1). We were surprised to find that this
alkane arylation reaction proceeded at 60°C
in 2 hours. Likewise, cyclopentane underwent
smooth arylation under similar conditions in
54% yield (Table 2, entry 2). Cycloheptane
could also be arylated in 40% yield (Table 2,
entry 3). With these results in hand, we set
out to investigate reactivity with acyclic alkanes.
We were pleased to find that n-hexane un-
derwent arylation to yield all three phenyl-
hexane isomers in 40% overall yield (Table 2,
entry 4). This C–H arylation reaction displayed
terminal selectivity, with an a:b:g ratio of 5:2:1.
Shao et al., Science 355, 1403–1407 (2017) 31 March 2017
3 of 5

RESEARCH
| REPORT
In a similar fashion, n-pentane also under-
went terminal-selective arylation to yield
phenylpentane isomers in 42% yield with a
10:3:1 ratio (Table 2, entry 5). In an earlier
report, direct and terminal-selective aryla-
tion of saturated hydrocarbons required zeo-
lite catalysts at >200°C and delivered <5%
yield (34).
The terminal C–H bonds of alkanes, although
kinetically more accessible, have higher bond
dissociation energies (98 kcal/mol) than their
internal counterparts (93 to 95 kcal/mol) (35).
Given our ability to arylate the strongest of C–H
bonds (Table 2, entries 4 and 5) and the long-
standing interest in hydrocarbon gas func-
tionalization (36, 37), we became intrigued by
the prospect of arylating methane gas (C–H
bond dissociation energy ~105 kcal/mol) (32).
Initial experiments were thwarted by delete-
rious arylation of solvent. After optimization
(table S3) (29), we found that use of C6F6 solvent
allowed for the arylation of methane in 32%
yield at low temperatures (60°C) and syntheti-
cally relevant pressures (35 bar) (Table 2, lower
panel). The conversion of naphthalene 8 to 1-
methylnaphthalene 9 serves as a rare example
of methane gas functionalization using main-
group catalysis (38).
Several experiments were performed to probe
the nature of the reactive intermediate. Arynes,
as well as their transition metal complexes, have
been shown to undergo C–H insertion reac-
tions (39, 40); however, intermolecular inser-
tion into sp3 C–H bonds has not been reported
(41, 42). The different products observed in Table 1,
entries 1 versus 4 and 7 versus 8, which would
presumably go through an identical aryne inter-
mediate (10; Fig. 3A), suggest that an aryne
intermediate is not active. Instead, these en-
tries suggest that an electrophilic site is local-
ized to the fluorine-bound carbon. As a further
means of investigation, we prepared butyl de-
rivative 11 (Fig. 3A). The intermediacy of aryne
10 (R = butyl) would lead to the formation of
biaryl 14 from both fluorides 11 and 13. How-
ever, unlike the arylation reactivity observed
in Table 1, entry 10 (i.e., 13 → 14), compound
14 was not observed in the reaction of butyl
derivative 11 (Fig. 3A). Instead, we observed
rapid intramolecular C–H insertion to form 1-
methylindane (12) in 35% yield (43% after op-
timization) (29).
Fig. 3. Mechanistic studies. (A) Comparison of products obtained from exposure of isomers 11 and
13 to standard reaction conditions. (B) Isotopic labeling studies that support a concerted insertion
mechanism and intermolecular competition experiment to determine the presence of a kinetic
isotope effect, ruling out C–H insertion as rate-determining step.
These observations do not rule out an un-
symmetrical silylium-dicarbenoid species anal-
ogous to the silver complex proposed by Lee
and co-workers (40). However, it is clear that
the reactive intermediate is not an aryne, as
C–C bond formation occurs at the C–F carbon
exclusively. Furthermore, although we draw
the reactive intermediate 5 (Fig. 2A) as a
formal carbocation, it is possible that a free
phenyl cation is not operative. As one of sev-
eral mechanistic possibilities, the C–H inser-
tion event may occur through reactive aryl
cation-halide adducts that form reversibly
under the reaction conditions. One could en-
vision a species analogous to the unreactive
phenyl chloronium salt described in (17), where-
in the b-trimethylsilyl group enhances reac-
tivity through steric buttressing or electronic
perturbation. Additionally, a trialkylsilyl phenyl
fluoronium salt, resulting from the interac-
tion between the fluoroarene substrate and a
cationic silicon species, could be operative
(fig. S51).
To investigate the nature of the key C–H in-
sertion event, we carried out isotopic labeling
studies. Using cyclohexane-d12, we observed for-
mation of phenylcyclohexane-d12 with an overall
deuterium incorporation of 80% (Fig. 3B). Al-
though deuterium was incorporated primarily
at the ortho-positions, enrichment of the meta-
and para-positions was also observed. We attri-
bute the mixture of D12 isomers to rapid hydride
shifts of Wheland intermediate 6 (Fig. 2A) (43).
However, the lack of apparent deuterium cross-
over (i.e., the presence of D12 n products) supports
a concerted C–H insertion process. Furthermore,
a competition experiment using a 1:1 mixture of
C6D12 and C6H12 provided a kinetic isotope effect
of 1.08 (Fig. 3B). This intermolecular competition
experiment rules out C–H insertion as the rate-
determining step (44).
Our work constitutes a methodology for gen-
erating aryl cation equivalents that engage in
Shao et al., Science 355, 1403–1407 (2017) 31 March 2017
4 of 5

RESEARCH
| REPORT
the intermolecular C–H arylation of arenes and
alkanes through main-group catalysis. Reagents
were designed wherein b-silicon substitution
facilitated C–F bond activation and catalyst
turnover. Furthermore, we found that the hyper-
fluorophilicity of the silylium-carborane cata-
lyst mediates selective functionalization of
C–F bonds in the presence of weaker C–X
bonds. This selectivity trend is complementary
to transition metal–catalyzed cross-coupling
reactions and traditional nucleophilic substi-
tution reactions. More fundamentally, this
work represents an exciting paradigm in cat-
alysis, where strong bonds (C–F and C–H) are
directly engaged in C–C bond-forming cross-
coupling reactions.
16. O. Allemann, K. K. Baldridge, J. S. Siegel, Org. Chem. Front. 2,
1018–1021 (2015).
17. S. Duttwyler et al., Angew. Chem. Int. Ed. 210, 7681–7684
(2010).
18. P. S. Skell, A. V. Garner, J. Am. Chem. Soc. 78, 5430–5433
(1956).
35. D. F. McMillen, D. M. Golden, Annu. Rev. Phys. Chem. 33,
493–532 (1982).
36. A. K. Cook, S. D. Schimler, A. J. Matzger, M. S. Sanford, Science
351, 1421–1424 (2016).
37. K. T. Smith et al., Science 351, 1424–1427 (2016).
38. B. G. Hashiguchi et al., Science 343, 1232–1237 (2014).
39. T. Truong, O. Daugulis, J. Am. Chem. Soc. 133, 4243–4245
(2011).
40. S. Y. Yun, K. P. Wang, N. K. Lee, P. Mamidipalli, D. Lee, J. Am.
Chem. Soc. 135, 4668–4671 (2013).
41. Although arynes are not known to form C–C bonds through
intermolecular sp3 C–H activation, metal-aryne complexes
have been shown to form metal-alkyl bonds through this
process (42).
42. K. Wada et al., J. Am. Chem. Soc. 125, 7035–7048 (2003).
43. S. Fornarini, M. E. Crestoni, Acc. Chem. Res. 31, 827–834
(1998).
19. The filled p-orbitals and empty sp2 orbital of phenyl cations
(17) are analogous to the orthogonal unoccupied p-orbital and
occupied s-orbital of singlet carbenes (18).
20. J. B. Lambert, Tetrahedron 46, 2677–2689 (1990).
21. Y. Apeloig, D. Arad, J. Am. Chem. Soc. 107, 5285–5286 (1985).
22. Y. Himeshima, H. Kobayashi, T. Sonoda, J. Am. Chem. Soc.
107, 5286–5288 (1985).
23. V. Dichiarante et al., J. Am. Chem. Soc. 129, 15919–15926
(2007).
24. K. K. Laali, G. Rasul, G. K. S. Prakash, G. A. Olah, J. Org. Chem.
67, 2913–2918 (2002).
25. W. Weber, Silicon Reagents for Organic Synthesis (Springer-
Verlag, 1983).
26. C. Douvris, O. V. Ozerov, Science 321, 1188–1190 (2008).
27. P. D. Bartlett, F. E. Condon, A. Schneider, J. Am. Chem. Soc.
66, 1531–1539 (1944).
28. Mechanistically, parallels can be drawn to Brönsted acid
catalysis, wherein an incident cationic species (proton versus
silylium) is eliminated from the cycle during the activation step
and regenerated anew from the penultimate intermediate. In
this case, triethylsilane was used for initiation because it is
easier to manipulate than gaseous trimethylsilane.
29. See supplementary materials.
30. C. Douvris, C. M. Nagaraja, C.-H. Chen, B. M. Foxman,
O. V. Ozerov, J. Am. Chem. Soc. 132, 4946–4953 (2010).
31. A. F. Littke, G. C. Fu, Angew. Chem. Int. Ed. 41, 4176–4211
(2002).
44. E. M. Simmons, J. F. Hartwig, Angew. Chem. Int. Ed. 51,
3066–3072 (2012).
ACKNOWLEDGMENTS
REFERENCES AND NOTES
The authors declare no competing financial interests. Support was
generously provided by the University of California, Los Angeles
(UCLA). We thank V. Lavallo and J. R. Gordon for assistance with
the syntheses of carba-closo-monocarborane anions, K. N. Houk
and E. M. Sletten for useful discussions regarding mechanism,
N. K. Garg and A. M. Spokoyny for assistance in manuscript
preparation, the UCLA Molecular Instrumentation Center for NMR
instrumentation, and the Mass Spectrometry Facility at the
University of California, Irvine. Experimental procedures and
characterization data are provided in the supplementary materials.
1. G. Merling, Ber. Dtsch. Chem. Ges. 24, 3108–3126 (1891).
2. C. K. Ingold, Structure and Mechanism in Organic Chemistry
(Cornell Univ. Press, 1953).
3. E. J. Corey, X. M. Cheng, The Logic of Chemical Synthesis
(Wiley, 1995).
4. R. R. Naredla, D. A. Klumpp, Chem. Rev. 113, 6905–6948 (2013).
5. G. A. Olah, R. H. Schlosberg, J. Am. Chem. Soc. 90, 2726–2727
(1968).
6. G. A. Olah, J. Org. Chem. 66, 5943–5957 (2001).
7. R. Pellicciari et al., J. Am. Chem. Soc. 118, 1–12 (1996).
8. F. Zhang et al., J. Am. Chem. Soc. 136, 8851–8854 (2014).
9. H. Zollinger, Acc. Chem. Res. 6, 335–341 (1973).
10. H. B. Ambroz, T. J. Kemp, Chem. Soc. Rev. 8, 353–365 (1979).
11. Z. Wu, R. Glaser, J. Am. Chem. Soc. 126, 10632–10639 (2004).
12. L. Mascarelli, Gazz. Chim. Ital. 66, 843–850 (1936).
13. E. Fasani et al., Chem. Commun. 1997, 1329–1330 (1997).
14. M. Fagnoni, A. Albini, Acc. Chem. Res. 38, 713–721 (2005).
SUPPLEMENTARY MATERIALS
32. S. J. Blanksby, G. B. Ellison, Acc. Chem. Res. 36, 255–263
(2003).
www.sciencemag.org/content/355/6332/1403/suppl/DC1
Materials and Methods
Figs. S1 to S51
Tables S1 to S3
References (45–69)
33. For many substrates, the by-products were volatile and
difficult to detect. Several crude nuclear magnetic resonance
(NMR) spectra are presented in the supplementary materials
that show the apparent purity of the products subsequent to
rotary evaporation.
Spectral Data
15. O. Allemann, S. Duttwyler, P. Romanato, K. K. Baldridge,
J. S. Siegel, Science 332, 574–577 (2011).
34. N. Danilina, E. L. Payrer, J. A. van Bokhoven, Chem. Commun.
46, 1509–1510 (2010).
30 January 2017; accepted 9 March 2017
10.1126/science.aam7975
Shao et al., Science 355, 1403–1407 (2017) 31 March 2017
5 of 5

Arylation of hydrocarbons enabled by organosilicon reagents
and weakly coordinating anions
Brian Shao, Alex L. Bagdasarian, Stasik Popov and Hosea M.
Nelson (March 30, 2017)
Science 355 (6332), 1403-1407. [doi: 10.1126/science.aam7975]
Editor's Summary
Turning benzene into a C−H bond cleaver
Ask chemists for the best way to break a strong bond, and they will tell you to make an even
stronger one. Shao et al. applied this principle by using silicon-fluorine bonds to break
carbon-hydrogen bonds. They prepared benzene rings with adjacent fluorine and silicon substituents.
Then they used a little extra activated silicon, paired with a carborane, to prime a cycle that draws away
the fluorine to produce a cation-like aryl intermediate. This intermediate can then slice through the t
ypically inert C−H bonds in alkanes, including methane. The alkylated rings go on to release their
silicon, which keeps the process going.
Science, this issue p. 1403
This copy is for your personal, non-commercial use only.
Visit the online version of this article to access the personalization and
article tools:
Article Tools
Permissions
http://science.sciencemag.org/content/355/6332/1403
Obtain information about reproducing this article:
http://www.sciencemag.org/about/permissions.dtl
Science (print ISSN 0036-8075; online ISSN 1095-9203) is published weekly, except the last week
in December, by the American Association for the Advancement of Science, 1200 New York
Avenue NW, Washington, DC 20005. Copyright 2016 by the American Association for the
Advancement of Science; all rights reserved. The title Science is a registered trademark of AAAS.