Gold-catalyzed oxidative coupling of arylsilanes and arenes: Origin of selectivity and improved precatalyst

Article abstract of DOI:10.1021/ja408712e

The mechanism of gold-catalyzed coupling of arenes with aryltrimethylsilanes has been investigated, employing an improved precatalyst (thtAuBr₃) to facilitate kinetic analysis. In combination with linear free-energy relationships, kinetic isotope effects, and stoichiometric experiments, the data support a mechanism involving an Au(I)/Au(III) redox cycle in which sequential electrophilic aromatic substitution of the arylsilane and the arene by Au(III) precedes product-forming reductive elimination and subsequent cycle-closing reoxidation of the metal. Despite the fundamental mechanistic similarities between the two auration events, high selectivity is observed for heterocoupling (C-Si then C-H auration) over homocoupling of either the arylsilane or the arene (C-Si then C-Si, or C-H then C-H auration); this chemoselectivity originates from differences in the product-determining elementary steps of each electrophilic substitution. The turnover-limiting step of the reaction involves associative substitution en route to an arene π-complex. The ramifications of this insight for implementation of the methodology are discussed.

Full text of DOI:10.1021/ja408712e

Article
pubs.acs.org/JACS
Gold-Catalyzed Oxidative Coupling of Arylsilanes and Arenes:
Origin of Selectivity and Improved Precatalyst
Liam T. Ball,†,‡ Guy C. Lloyd-Jones,*,† and Christopher A. Russell*,‡
†School of Chemistry, University of Edinburgh, Joseph Black Building, West Mains Road, Edinburgh EH9 3JJ, U.K.
‡School of Chemistry, University of Bristol, Cantock’s Close, Bristol BS8 1TS, U.K.
S
* Supporting Information
ABSTRACT: The mechanism of gold-catalyzed coupling of
arenes with aryltrimethylsilanes has been investigated, employ-
ing an improved precatalyst (thtAuBr3) to facilitate kinetic
analysis. In combination with linear free-energy relationships,
kinetic isotope effects, and stoichiometric experiments, the
data support a mechanism involving an Au(I)/Au(III) redox
cycle in which sequential electrophilic aromatic substitution of
the arylsilane and the arene by Au(III) precedes product-
forming reductive elimination and subsequent cycle-closing
reoxidation of the metal. Despite the fundamental mechanistic similarities between the two auration events, high selectivity is
observed for heterocoupling (C−Si then C−H auration) over homocoupling of either the arylsilane or the arene (C−Si then C−
Si, or C−H then C−H auration); this chemoselectivity originates from differences in the product-determining elementary steps
of each electrophilic substitution. The turnover-limiting step of the reaction involves associative substitution en route to an arene
π-complex. The ramifications of this insight for implementation of the methodology are discussed.
(low loading of a relatively benign metal catalyst,6 room
temperature, aerobic atmosphere, reagent-grade solvents, and
near-equal stoichiometries of coupling partners/oxidant). The
oxidative nature of the reaction and the passivity of the
INTRODUCTION
■
Advances in synthetic methodology benefit from mechanistic
insight. In cross-coupling this has allowed rational design of
both catalyst1 and reagent;2 in direct arylation,3 a potential
arylsilane coupling partner tolerate a broad range of
alternative to classical cross-coupling, experimental and
functionality, including not only esters, aldehydes, and primary
computational investigations have illuminated the main
alcohols but also arylhalides and triflates, rendering the
modes of reactivity,4 in particular the C−H activation step.
methodology orthogonal to traditional cross-coupling.
Since Periana’s seminal report in 2004,7 the field of oxidative
We recently reported the gold-catalyzed arylation of arenes by
aryltrimethylsilanes (Scheme 1).5 The reaction exhibits high
gold catalysis has expanded to encompass a diverse spectrum of
C−C and C−X bond-forming reactions.8 However, for such
regio- and chemoselectivity, without reliance on ortho-directing
groups, and proceeds under mild and convenient conditions
processes, there is generally little evidence regarding the
oxidation states of the metal that are accessed,9 and only an
empirical appreciation of the factors affecting reaction rates.
Thus, it is widely appreciated8b,c that further developments will
benefit from a deeper understanding of the elementary
reactivity involved.10 For the process outlined in Scheme 1,
both the relative reactivity of the arene and the regioselectivity
of its arylation are consistent with the established rules of
electrophilic aromatic substitution (SEAr).5 Similar observations
have been made in stoichiometric auration of simple arenes by
Au(III),11 and in their Au-catalyzed oxidative functionaliza-
tion.12 In contrast, the interaction of Au(I) cations with arenes
results in either labile complexes13 or benzylic C−H
activation.14 Given this precedent, and the well-documented
ipso SEAr reactivity of aryltrimethylsilanes, we envisaged that
net C−H and C−Si auration15 by an electrophilic Au(III)
Scheme 1. Representative Scope of Au-Catalyzed Arene−
Arylsilane Oxidative Coupling5
Received: September 4, 2013
© XXXX American Chemical Society
A
dx.doi.org/10.1021/ja408712e | J. Am. Chem. Soc. XXXX, XXX, XXX−XXX

Journal of the American Chemical Society
Article
species may constitute key elementary steps in the cycle.16
Herein, we report a mechanistic investigation into Au-catalyzed
arylation, with the aim of elucidating the origin of the
chemoselectivity that results in heterocoupling of the arylsilane
with the arene, rather than homocoupling of either species.
Scheme 2. Dichotomous Au-Catalyzed Arylation Cycles
RESULTS AND DISCUSSION
■
Preliminary Studies. Our initial mechanistic hypotheses
were based on prior observations5 and were reinforced by
preliminary stoichiometric studies. Treatment of (4-
fluorophenyl)trimethylsilane (ArFSiMe3, 1) and/or mesitylene
(2), with Ph3PAuOTs resulted in no change in either the Au(I)
which afforded fluorene-derivative 5 as the sole product (eq 6).
Operation of cycle B with irreversible C−H auration would
complex or the aromatic substrates (31P{1H}, H, 19F NMR)
1
(eq 1).17 Attempts to generate a more-electrophilic, phosphine-
require reaction to occur with exceptional selectivity for C-3,
despite the fact that C-5 is similarly electronically activated and
more sterically accessible. Indeed, arylation of des-silyl analogue
6 afforded 7 as a single regioisomer (eq 7). Alternatively, C−H
auration may be reversible, with rapid protodeauration of a
monoarylgold species allowing formal migration of the metal:
once C-3 is reached, C−Si auration and reductive elimination
would afford 5. Conducting cyclization of 4 with CD3OD as
cosolvent, however, resulted in no isotopic enrichment of either
5 or residual 4 (at ∼90% conversion). In contrast, initial C−Si
auration renders a single aromatic C−H bond (C-3) accessible
in a subsequent intramolecular C−H auration, consistent with
the quantitative formation of 5. These observations point to
C−Si auration preceding C−H auration in 4, as illustrated in
cycle A.
free Au(I) species in situ by addition of AgOTf to thtAuCl (tht
= tetrahydrothiophene) resulted in deposition of metallic gold
and no reaction of either 1 or 2 (eq 2).14As the lack of
reactivity of Au(I) toward 1 and 2 (eqs 1 and 2) does not
preclude intermediacy of aryl−Au(I) species, their involvement
was tested under catalytically relevant conditions (eqs 3 and 4).
thtAuC6H4F decomposed in the presence of acid,18 as did
thtAuC6F5,19 while Ph3PAuC6H4F reacted with excess PhI-
(OH)OTs to give Ph3PAuOTs and the corresponding
diaryliodonium salt.20 Finally, stoichiometric reaction of 1
and 2-bromothiophene 3 in the presence Au(OAc)3 and
camphorsulfonic acid (CSA) afforded both hetero- and
homocoupling products (eq 5). Together, these observations
disfavor a catalytic cycle involving auration by Au(I), and/or
the involvement of discrete aryl−Au(I) intermediates, and
demonstrate the feasibility of aryl−aryl bond formation
mediated by Au(III). The operation of a Au(I)/Au(II)-dimer
redox cycle, as proposed for olefin heteroarylation,10a was
discounted following isolation of reactive intermediates (vide
infra).
After screening a wide range of fluorinated reactants19 as
candidates for in situ 19F NMR kinetic studies, we selected a
combination of fluorophenylsilane 1 and thiophene 8 with a
fluorinated oxidant (F-IBDA, 4-fluoroiodobenzene diacetate).
Although arylation proceeds cleanly to form 9 as a single
regioisomer within an appropriate time frame, an induction
period of ∼45 min (Figure 1, left) complicated analysis. The
origin of this induction period was investigated, vide infra, to
We were thus able to outline a catalytic cycle in which
oxidation of Au(I) by iodine(III) generates a reactive Au(III)
species capable of aurating both the arylsilane C−Si bond and
the arene C−H bond. Subsequent, product-forming reductive
elimination from the diaryl Au(III) intermediate closes the
catalytic cycle. However, this initial hypothesis (Scheme 2)
raises the question of whether C−Si auration precedes C−H
auration, or vice versa (cycle A vs cycle B). Discrimination
between these two scenarios was achieved by subjecting
tethered substrate 4 to the standard coupling conditions,5
Figure 1. Representative first-order log-plots with Ph3PAuOTs (left)
and thtAuBr3 (right) as precatalyst.
B
dx.doi.org/10.1021/ja408712e | J. Am. Chem. Soc. XXXX, XXX, XXX−XXX

Journal of the American Chemical Society
Article
allow its circumvention and ultimately facilitate a more
thorough treatment of reaction kinetics (Figure 1, right).
Induction Period. Origin of the Induction Period. Sulfonic
acids react with iodobenzene(dicarboxylates) to generate the
more reactive [hydroxy(sulfonyloxy)iodo]benzene oxidants.21
However, reaction of F-IBDA with CSA to generate hydroxy-
(camphorsulfonyloxy)iodo]fluorobenzene (F-HCIB), or a
similar species,19 is complete upon mixing (19F NMR).
Furthermore, an identical induction period is observed when
arylation is conducted with preformed F-HCIB, suggesting that
formation of an active catalyst from Ph3PAuOTs is instead
responsible for the induction period.22 Under catalytically
relevant conditions, but in the absence of 1 or 8, treatment of
Ph3PAuOTs with F-IBDA/CSA led to consumption of 2 equiv
of oxidant over ∼120 min (Figure 2, left).
For a series of precatalysts, (p-RC6H4)3PAuOTs, the
duration of the induction period increases as R = Me < H <
CF3. As more strongly donating ligands facilitate metal
oxidation25 and retard ligand dissociation, the observed trend
disfavors path A. In contrast, if Bi and Ci are fast, the induction
period associated with paths B and C will depend differently on
ligand basicity: while phosphine dissociation (Bii) is expected to
be slower for R = Me than for R = CF3, reductive elimination of
a phosphonium salt (Cii) is expected to be accelerated by a
more nucleophilic phosphine.26 Path C is thus most consistent
with the observed dependence of the induction period on R.
Changing from tri(p-tolyl)phosphine to tri(o-tolyl)-
phosphine resulted in lengthening of the induction period to
∼3 h, consistent with the more sterically demanding
phosphine27 disfavoring the geometry change (linear to
square-planar) that accompanies oxidation of the Au(I) center
(steps Bi and Ci).25 With this insight it is now possible to
rationalize the ligand-dependence observed during initial
reaction development:5 (p-RC6H4)3PAuOTs precatalysts gave
uniformly high yields and selectivities, suggesting that the same
active catalyst is formed, independent of ligand. With bulkier
ligands, including NHCs,19 very slow release of the active
catalyst results in greatly diminished yields.
Development of an Improved Precatalyst. Having
established that phosphine oxidation occurs during the
induction period, a range of ‘ligand-free’ Au(I) and Au(III)
species was tested as arylation precatalysts. AuBr3 was identified
as a source of highly active catalyst that initiated turnover
without induction,19 although the hygroscopicity and limited
solubility of the salt rendered its use inconvenient. Through
further precatalyst screening19 we found that thtAuBr3 also led
to high catalyst turnover without a discernible induction period
(Figure 1, right). In addition to simplifying kinetic analyses,
thtAuBr3 is conveniently soluble in organic media and stable
toward air, light, and moisture. Couplings between a diverse
range of partners exhibited chemo- and regioselectivities
identical to reactions performed with Ph3PAuOTs as
precatalyst.5
Figure 2. Reaction of Ph3PAuOTs (left), and thtAuBr3/8 (right), with
F-IBDA/CSA at 30 °C in CDCl3/CD3OD. 4-Fluoroiodobenzene was
monitored as a proxy for oxidant consumption.
Monitoring the process by 31P{1H} NMR revealed
quantitative formation of Ph3PO.23 The induction period is
thus associated with the net two-electron oxidation of both the
metal and the ligand. Pathways for liberation of a phosphine-
free catalyst from Ph3PAuOTs (Scheme 3) include: dissociation
Scheme 3. Liberation of a Phosphine-Free Catalyst
The high selectivity for heterocoupling (leading to 9) with
thtAuBr3, thtAuBr, or Ph3PAuOTs, contrasts strongly with the
reduced activity and selectivity obtained with chloride-based
precatalysts such as Ph3PAuCl or thtAuCl.19 The origins of
these differences became apparent upon detailed analysis of the
crude reaction mixtures. Aromatic bromination products were
detected in reactions employing thtAuBr3 as precatalyst,28 the
net loss and oxidation of bromide resulting in the in situ
formation of a gold sulfonate (vide infra). The ancillary sulfide
is also rapidly and quantitatively oxidized on addition of F-
IBDA/CSA, forming tetrahydrothiophene 1-oxide and a
product of ring opening.19 The same ratio of tht-derived
byproducts19 is observed under stoichiometric and catalytic
conditions and is independent of the presence of substrates.
Monitoring the reaction of thtAuBr3 with F-IBDA/CSA in the
presence of thiophene 8 indicated that, relative to [Au],
formation of 2 equiv of bromination product 10, and
consumption of 5 equiv of oxidant, accompanies catalyst
activation (Figure 2, right). In contrast, chlorination side
products were not observed with thtAuCl as precatalyst, and
the only tht-derived byproduct was the sulfoxide.29
of PPh3 from Au(I) prior to independent oxidation of each
component (path A), oxidation of phosphine-ligated Au(I) to
Au(III), followed by either dissociation and oxidation of PPh3
(B), or reductive elimination of phosphorus(V)24 and
reoxidation of Au(I) (C). Rapid initial consumption of one
equivalent of oxidant (relative to [Au]) is accompanied by
formation of only ∼0.2 equiv of Ph3PO (Figure 2, left). Slower,
subsequent consumption of the second equivalent of oxidant
occurs at a rate that parallels formation of Ph3PO. It follows
that the initial burst of oxidant consumption corresponds to
oxidation of Au(I) which, taken with the observation that free
PPh3 is oxidized very rapidly under the reaction conditions,19
suggests that path A is not a major contributor to catalyst
activation.
Reaction of 8 with [nBu4N]Br under the conditions of
Scheme 1, but in the absence of gold, led to clean
bromination.30 In contrast, [nBu4N]Cl did not chlorinate 8,
suggesting that the chloride in Ph3PAuCl and thtAuCl remains
C
dx.doi.org/10.1021/ja408712e | J. Am. Chem. Soc. XXXX, XXX, XXX−XXX

Journal of the American Chemical Society
Article
available to ligate gold. This was corroborated by addition of
catalytic [nBu4N]Cl to a thtAuBr3-catalyzed coupling of 1 and 8
([Cl]/[Au] = 1), which resulted in a substantial drop in
selectivity and activity, analogous to that observed with both of
the AuCl-based precatalysts.19
Stoichiometric Studies. With the insight that precatalyst
activation affords ‘ligand-free’ Au(III), we conducted stoichio-
metric reactions (Scheme 4) to confirm the viability of the
dibenzoaurole as anion 16 or cation 17, respectively. That
arylation of biphenyl with 1 occurs solely at the 4-position,19
while (o-biphenyl)silane 15 is aurated at the 2′-position, is
again consistent with C−Si auration preceding C−H auration.
The acceleration of C−Si auration by steric decompression
(vide infra) and the stability of the resulting dibenzoaurole
toward reductive elimination render 15 a powerful catalyst
poison.19
Reaction Kinetics. Catalytic Rate Law; Activation
Parameters. Having identified thtAuBr3 as an efficient
precatalyst, the coupling of fluorophenylsilane 1 and thiophene
8 was investigated in detail by in situ 19F NMR, allowing
formulation of an empirical rate law (Figure 3, left). The rate of
Scheme 4. Stoichiometric Experiments
Figure 3. Predicted vs experimentally observed rates (left, d[9]/dt =
kobs[thtAuBr3][8][CSA]/[MeOH]0.5; kobs = 3.09 M−1.5s−1), and Eyring
analysis (right), of the arylation of 8 with 1 (Figure 1).
arylation exhibits first-order dependence on the concentrations
of thtAuBr3, arene 8, and CSA34 and is independent of silane 1
and F-IBDA. The clean first-order dependence of rate on
[thtAuBr3] over an 8-fold concentration range strongly suggests
a monomeric, and presumably mononuclear, catalyst. In
addition to acting as an essential cosolvent and sequestering
the TMS moiety liberated upon C−Si auration, monomeric
MeOH is an inhibitor.35 Analysis over the temperature range
25−55 °C afforded apparent activation parameters of ΔH =
15.2 0.5 kcal mol−1 and ΔS = −26.5 0.5 cal K−1 mol−1 for
arylation of 8 with 1 (Figure 3, right).36 The large, negative
value of ΔS is consistent with a significant increase in order
during the turnover-limiting event, although care should be
taken in the interpretation of this value, due to the possibility of
contributions to ΔS from several elementary steps.
sequential C−Si/C−H auration of cycle A (Scheme 2). The
isolation of discrete Au−C σ-bonded species, and the absence
of products arising from benzylic activation, provided support
for ionic, rather than SET-type mechanisms.31
While the auration of simple arenes by Au(III) was
documented by Kharash as early as 1931,11a the corresponding
auration of arylsilanes has not been reported. By analogy to
Fuchita’s modification11d of the original C−H auration
procedure,11a reaction of phenyltrimethylsilane 11 with
(AuCl3)2, and addition of 2,6-lutidine, afforded trans-
AuPhCl2(lut.) as evidence of C−Si auration by a well-defined
Au(III) precursor (Scheme 4A). The same species was isolated
from a thtAuBr3-catalyzed reaction following addition of 2,6-
lutidine and its HCl salt.19 Phenylthiophene 12 was obtained
upon addition of AgOTf to trans-AuPhCl2(lut.) and 2-
bromothiophene 3 in CDCl3/CD3OD, thereby demonstrating
that further arylation of a coordinatively unsaturated
intermediate, and subsequent reductive elimination, are also
feasible (cf. eq 5).32
Although thtAuBr3 itself does not react with either coupling
partner, formation of an active Au(III) species under
catalytically relevant conditions and its reactivity toward
arylsilanes were confirmed by treating the precatalyst with
C6D5I(OH)OMs in the presence of arylsilane 13 (Scheme 4B).
Addition of excess [nBu4N]Cl allowed trichloroaurate 14 to be
characterized in situ (1H NMR and HR ESI-MS). Alternatively,
the product of sequential C−Si/C−H auration could be
generated from (o-biphenyl)trimethylsilane 15 under analogous
conditions (Scheme 4C). Addition of either [nBu4N]Cl or 4-
(dimethylamino)pyridine (DMAP) allowed isolation of the
Deuterium Kinetic Isotope Effects (KIEs). Measurement
of absolute rate, as well as a series of intra- and intermolecular
competitions, afforded KIEs for both the arylsilane and the
1
arene. There was no detectable (1H NMR) transfer of H/2H
between substrates/products and the cosolvent (CD3OD or
MeOH). Arylation of 2-bromothiophene 3 with mixtures of
either d0-1/2,6-d2-1 (→18) or d0- and 2,6-d2-(p-tolyl)-
trimethylsilane (→19) gave no significant β-SKIE (Chart 1),
suggesting that a substantial change in hybridization β to Si
does not occur on approach to the product-determining
transition state (TS) for C−Si auration. Negligible SKIEs were
also observed for arylation of 3 with mixtures of d0- and α,α,α-
d3-(p-tolyl)trimethylsilane (→19), or d0- and α,α,α-d3-(o-
tolyl)trimethylsilane (→20).
The absolute rate of arylation of thiophene 8 with silane 1
proved insensitive to isotopic substitution at either the 4- or 5-
positions of the thiophene (Scheme 5A). Similarly, intermo-
lecular competitions between isotopologues of 8 for 1 yielded
no significant KIE (Scheme 5B). Intramolecular competitions
were performed by reaction of a single isotopomer of 3-
D
dx.doi.org/10.1021/ja408712e | J. Am. Chem. Soc. XXXX, XXX, XXX−XXX

Journal of the American Chemical Society
Article
Chart 1. SKIEs for C−Si Bond Auration (Intermolecular
Competition)
Scheme 5. PKIEs for C−H Bond Auration
Figure 4. Hammett plot for silanes derived from absolute rates of
arylation of 3 and relative rates of arylation of mesitylene 2.
Figure 5. Hammett plot of rates of arylation (vs σ) and rates of
acylation (vs σ+) of 2-substituted-9,9-dimethylfluorenes.
While arylation is accelerated by electron-releasing substituents,
the magnitude of the reaction constant (ρ = −1.8) is relatively
small due to the insulating effect of the fluorene system.38 In
comparison, Friedel−Crafts acylation of the same fluorenes
gave ρ+ = −3.4,19 2.7-fold less than the ρ+ value reported for
acylation of simple arenes under the same conditions (ρ+ =
−9.1).39 Scaling by the same factor thus gives ρ ≈ −4.9 for Au-
catalyzed arylation of simple arenes, similar to the value (ρ =
−4.7) found for electrophilic, benzylic auration of para-
substituted toluenes.40 Comparable reaction constants (all
based on σ+) have been reported for electrophilic aromatic
metalation by Hg (ρ+ = −4.0 to −6.4),39 Rh (ρ+ = −4.5),39 and
Ru (ρ+ = −4.4).41 In contrast, a negligible reaction constant is
observed for stoichiometric C−H activation via σ-bond
metathesis (σBM) at Sc,42 while Ru-mediated C−H activation
via σBM is predicted to have a positive reaction constant (ρ =
2.6 to 3.2).43 The significance of the reaction constant for Au-
catalyzed arylation is discussed in later sections.
Alkylbenzenes. Wheland Intermediate and π-Complex
Stability. For SEAr reactions, the Hammond postulate suggests
that, if the relative rate of reaction of a series of arenes
correlates with the relative stability of the intermediate π-
complexes, then the rate-limiting TS resembles a π-complex;
similarly, correlation against Wheland Intermediate (WI)
stability suggests rate-limiting formation of a WI.39 In practice,
the relevant π-complexes/WIs cannot usually be observed
directly, and thus model complexes with e.g., silver cations, HCl,
or Br2 (model π-complex), or superacids (model WI) are
employed. Relative rates of Au-catalyzed arylation of a series of
methylbenzenes with silane 1 were derived from absolute rate
bromothiophene 21 with 1; in contrast to the previous
experiments, small normal PKIEs were observed (Scheme
5C). Together, these results indicate that the turnover-limiting
elementary step of C−H auration involves neither C−H/D
cleavage nor a change in hybridization/geometry and that C−
H/D cleavage is product determining only in the partitioning
that leads to regioisomers.
Linear Free Energy Relationships (LFERs). Arylsilane.
Hammett plots derived from analysis of both absolute rates of
arylation (single arene/single silane; A, Figure 4), or from
competition between two arylsilanes for a single arene (B,
Figure 4), correlated against standard σ-values. This confirms
the absence of direct resonance interaction between the
aromatic substituent and the reaction center in the product-
determining TS; a similar observation has been made for
mercuridesilylation.37 Since the silane does not appear in the
empirical rate equation, the dependence of kobs on the identity
of the arylsilane was, as expected, negligible (ρ = −0.2). In
contrast, partitioning between competing arylsilanes confirmed
that C−Si auration is accelerated by electron-releasing
substituents (ρ = −1.6).
Arene. The absolute rates of arylation of 2-substituted-9,9-
dimethylfluorenes gave a good correlation against σ (Figure 5).
E
dx.doi.org/10.1021/ja408712e | J. Am. Chem. Soc. XXXX, XXX, XXX−XXX

Journal of the American Chemical Society
Article
data and were analyzed as a function of the relative stabilities of
the π-complexes (Figure 6, left) or WIs (right) formed by the
same arenes with AgNO3 or HF/BF3,45 respectively.
44
Figure 7. Relative rates of Au-catalyzed arylation of alkylbenzenes with
4-TMS-(N-Me-N-Pv)aniline vs relative rates of mercuration (A), and
of thiophene 8 with (p-alkylphenyl) silanes vs relative rates of
mercuridesilylation (B). Pv = pivaloyl.
Figure 6. Absolute rates of arylation of polymethylbenzenes with 1 vs
relative stabilities of the corresponding π-complex (left; R2 = 0.97) and
WI (right; R2 = 0.99). Rates are plotted as krel with respect to toluene;
see Supporting Information [SI] for details.
of the arylsilanes is insensitive to minor electronic changes in
R′, (2) there is minimal charge delocalization in the product-
determining step, or (3) the opposing electronic and solvation
effects cancel.
The rates of arylation of substrates with one or no methyl
substituents ortho to the reacting C−H bond correlate against
π-complex stabilities, suggesting that the turnover-limiting
elementary step proceeds via a TS resembling a π-complex.
The converse is true of substrates with methyl substituents at
both positions ortho to the reacting site. Within the context of
SEAr, this can be rationalized by considering that evolution
from a π-complex to a WI requires the electrophile (in this case
Au) to move toward the aryl plane, which would be met by an
increasing steric penalty as the number of methyl groups ortho
to the reactive site increases. A similar phenomenon has been
noted in electrophilic mercuration, for which the barrier to
metalation of di-ortho-substituted arenes is greater than twice
that for mono-ortho-substituted arenes, an effect that is
amplified by buttressing between adjacent substituents.38a By
analogy to electrophilic mercuration and thallation,31b the steric
component of the energy barrier to π-complex formation is
largely unchanged by increasing methylation, such that the
overall π-complexation barrier would decrease with increasing
arene methylation (i.e., π-basicity). In contrast, the stability of
the WI would show significant sensitivity to steric factors, with
the energy barrier to WI formation exceeding that associated
with π-complexation in di-ortho-methylated substrates.
o-Methyl Arylsilanes. Comparison of the absolute rates of
reaction of thiophene 8 with either (4-fluorophenyl)silane 1 or
(o-tolyl)trimethylsilane revealed that arylation with the less
electron-rich arylsilane occurred some 4.6 times faster (Chart
2A); given the low sensitivity of kobs to the electronic nature of
Chart 2. Reactions of o-Methyl Arylsilanes
Baker−Nathan Reactivity. The relative rates of Au-catalyzed
the arylsilane (Figure 4), this can largely be attributed to steric
effects in the turnover-limiting auration of thiophene 8. In
contrast, competition between o- and p-tolylsilanes for 8 resulted
in substantial partitioning in favor of the more sterically
congested ortho-isomer (Chart 2B). Ortho-methylation of
arylsilane 1 resulted in a small increase in regioselectivity for
arylation of 3-bromothiophene 21 (Chart 2C); a more
appreciable increase in regioselectivity was observed for
arylation of toluene with 4-chloro-2-methylphenyltrimethylsi-
lane 13 relative to (4-chlorophenyl)trimethylsilane (Chart 2D).
Thus, introduction of a methyl substituent ortho to the silyl
moiety results in an increase in both the relative reactivity of the
arylsilane and the regioselectivity with which the arene is
arylated, while the overall rate of catalytic turnover is decreased.
Proposed Mechanism. The data presented above and in
the SI, taken with the discussion below, allow a more detailed
catalytic cycle to be proposed (Figure 8). Following precatalyst
activation, electrophilic auration of the arylsilane by ‘ligand-free’
Au(III) species I affords, via II and III, the catalyst resting state
arylation of monoalkylbenzenes (PhR) decrease as R = Me > Et
i
t
> Pr > Bu, opposing the increasing arene activation by
induction/hyperconjugation. A similar trend is observed for
stoichiometric electrophilic mercuration,46 with krel(arylation)
correlating linearly with krel(mercuration) (A, Figure 7). The
Baker−Nathan order arises from increasing hindrance by R to
the solvation of a charged Ar−R intermediate,47 suggesting that
some positive charge does develop on the arene during the
turnover-limiting event.
In contrast to the arene, the relative rates of C−Si auration in
the competition between a series of p-R′C6H4SiMe3 for 8 are
almost insensitive to the nature of R′ (B, Figure 7); similarly,
mercuridesilylation also shows smaller dependence on R′ than
does mercuration on R (in PhR).48 It has been proposed that,
due to the high SEAr reactivity of arylsilanes, the TS for
electrophilic mercuration is early and resembles the reactants,
such that there is little electronic demand on ring
substituents.48 The observed trend suggests that: (1) auration
F
dx.doi.org/10.1021/ja408712e | J. Am. Chem. Soc. XXXX, XXX, XXX−XXX

Journal of the American Chemical Society
Article
Figure 8. Overall catalytic cycle proposed for Au-catalyzed arylation of arenes with aryltrimethylsilanes. [‘X’ = undefined ligand(s), likely comprising
a mixture of MeOH/CSA and conjugate bases. The identity of ‘X’ and its protonation state is expected to change during the cycle in response to
electronic and steric requirements of the catalyst, in turn determining the stereochemistry and net charge of the complex. Inset: qualitative reaction
coordinate for C−H auration of 8 (solid line) and 4,5-d2-8 (dashed line).]
IV. Exchange of methoxide in IV for camphorsulfonate in V
provides a more electrophilic metal, which undergoes turnover-
limiting, associative substitution by the arene to afford π-
complex VI. This species evolves to WI VII; for certain
substrate classes (e.g., 3-bromothiophene 21), formation of VII
is reversible (or the Au fluxional), such that C−H cleavage is
regioselectivity determining. Collapse of VII to VIII precedes
ligand dissociation, with reductive elimination from three-
coordinate diaryl−Au(III) intermediate IX releasing the biaryl
product. Oxidation of the resulting Au(I) species X regenerates
I, thereby closing the cycle.
Catalyst Speciation. Precatalyst activation proceeds with
irreversible loss of the supporting phosphine/sulfide. A
corollary is that, while neither the selectivity nor activity of
the catalyst can be modulated by tuning the ancilliary ligand,
substrate-specific precatalysts are not required. Identification of
the active Au(III) species that is liberated from the precatalyst
is challenging due to the dynamics of ligand exchange. Indeed,
the nature of the supporting ligands/anions (e.g., MeOH, CSA,
AcOH, methoxide, camphorsulfonate, acetate, π-bound arene,
etc.), the net charge of the complex, and its geometry will likely
change in response to the different electronic and steric
requirements of the metal at various stages of the cycle.
Nonetheless, we suggest that camphorsulfonate anions, and up
to one methoxide/hydroxide, are associated with the Au(III)
center, alongside datively bound methanol/water.19
supports reactivity of arylsilanes via SEAr, including arylation of
late transition metals such as Pd(II) and Pt(II).52
Protodesilylation of arylsilanes exhibits an inverse β-SKIE,53
typical for an SEAr reaction with rate-limiting formation of a
WI. The absence of a SKIE in intermolecular competition
between d0-/2,6-d2-arylsilanes (Chart 1) suggests that there is
little rehybridization β to Si, and the absence of a normal SKIE
for arylation of 2-bromothiophene 3 with the CH3/CD3
isotopologues of tolyltrimethylsilanes implies that there is little
delocalization of charge onto the aromatic system during the
product-determining step of C−Si auration.54 Furthermore, the
low sensitivity of the reaction rate to the identity of p-alkyl
substituents (B, Figure 7) may also be indicative of minimal
charge delocalization in the rate-limiting step of C−Si auration.
That the Hammett reaction constant (ρ = −1.6) is significantly
smaller than values for typical SEAr substitutions ipso to
silicon55 and that good correlation is found against σ rather
than σ+ (Figure 4) are also consistent with small electronic
demands being made of the arylsilane.
The increase in relative reactivity upon ortho-methylation
(Chart 2B) of the arylsilane arises from steric decompression, a
phenomenon observed in other SEAr processes, including
mercuridesilylation56 and protodesilylation.57 Steric interac-
tions that destabilize the arylsilane in the ground state are
reduced upon formation of a WI, in which the TMS moiety
moves out of the aryl plane. Thus, for the Au-catalyzed process,
the product-determining event must involve distortion of the
arylsilane from planarity.
Three general scenarios can be envisaged in which the rate of
C−Si auration, and selectivity for C−Si over C−H, is
controlled by (a) π-complexation of the arylsilane (I → II),
(b) cleavage of the C−Si bond via σBM (II → III′ → IV),58 or
C−Si Auration. Precedent exists for transmetalation of aryl
moieties to Au(III) from various organometallic reagents33,49

including boroxines50although no mechanistic studies have
been reported to date. While formation of Au−CF3 species via
reaction of Au(I) and Au(III) phenoxides with TMS-CF3 has
been proposed to proceed via σBM,51 a large body of data
G
dx.doi.org/10.1021/ja408712e | J. Am. Chem. Soc. XXXX, XXX, XXX−XXX

Journal of the American Chemical Society
Article
(c) formation of a WI on an SEAr pathway (II → III). Scenario
(a) is consistent with the small, negative reaction constant (and
the correlation against σ), the insensitivity to the nature of p-
alkyl substituents and the lack of KIEs. However, it fails to
account for the accelerating effect of silane ortho-methylation,
and provides no explanation for the selectivity between the
silane and the arene (i.e., chemoselectivity; vide infra). As an
alternative, scenario (b) fails to account for both the
accelerating effects of ortho-methylation and the reaction
constant (a negligible or positive ρ value is expected for
σBM),42,43 although invoking a reaction manifold fundamen-
tally different from that operating for C−H auration offers a
clear basis for chemoselectivity. Finally, rate-limiting formation
of a WI, scenario (c), accounts not only for the accelerating
effect of ortho-methylation but also, in the limit of an early
TS,48 for the absence of KIEs, the insensitivity to the nature of
alkyl substituents, and the magnitude/direction of the reaction
constant. Thus, when taken with the high ipso-factors of
arylsilanes,39 an SEAr mechanism with rate-limiting formation
of a WI (III) (scenario (c)) is the preferred pathway for C−Si
auration.40
C−H Auration. The empirical rate law (Figure 8) is
consistent with rate-governing generation of a π-complex
between the arene substrate and the catalyst.59 The first-order
dependence on the concentration of CSA, and inhibition by
MeOH, implies a pre-equilibrium between methoxide resting
state IV and camphorsulfonate V, with the fractional depend-
ence on the concentration of MeOH resulting from the
extensive dimerization of MeOH in CHCl3.60 Following
turnover-limiting formation of the π-complex VI, evolution to
VIII can nominally occur via σBM, concerted metalation−
deprotonation, or SEAr, pathways, although the latter is
preferred (vide infra).61
That V → VI is the turnover-limiting step is corroborated by
the KIEs determined for reactions of deuterated thiophenes
(Scheme 5). The absence of any KIE on the absolute rate
suggests that the turnover-limiting step does not involve
cleavage of the C−H bond (VII → VIII) or significant
rehybridization en route to WI VII.62 The absence of KIEs in
intermolecular competition is consistent with irreversible
substrate coordination being both turnover limiting and
product determining.63 The observation of small PKIEs
(Scheme 5C) in intramolecular competition indicates that
cleavage of the C−H bond (i.e., collapse of VII to VIII, via an
early and/or nonlinear TS) controls regioselectivity, which in
turn requires that formation of the WI is formally reversible and
that π-complex VI is fluxional.59 Figure 8 includes a qualitative
reaction coordinate diagram illustrating these points; however,
this does not necessarily extend to other substrates, as the
strong resonance stabilization of a thiophene-derived WI will
significantly lower the energy of the TS by which it is formed.
Such an effect will be absent in toluene, for example, so that
formation of the WI may become irreversible and regiose-
lectivity determining. In hindered substrates, e.g., mesitylene 2,
the energy of the WI may be raised sufficiently for its formation
to become both product determining and turnover limiting
(vide supra).
demanding gold center may be more discriminating during
formation of VII from fluxional π-complex VI. In the second
case, if formation of VII is rapidly reversible, the regioselectivity
of auration will depend on the rate of collapse of VII to VIII,
which will be influenced by the steric requirements of the metal
center as it moves into the aryl plane. The Baker−Nathan
reactivity (A, Figure 7) suggests that charge develops on the
arene (e.g., formation of a WI) during the turnover-limiting TS,
contrary to the small reaction constant (σ rather than σ+, Figure
5). However, the effect can also be rationalized in terms of
steric hindrance to formation of VI via π-complexation. The
generation of VI has the characteristics of an associative
process:64 the rate is first order in arene (ρ = −1.8) with a large
apparent negative entropy of activation (Figure 3, right).
Orientation of the silane-derived aryl ring perpendicular to the
square plane of V results in one axial site of the metal being
blocked;65 as such, the approach of the arene en route to
formation of VI is hindered, and the turnover rate is reduced
with ortho-methyl arylsilanes (Chart 2A).
Origin of Chemoselectivity. The high selectivity for hetero-
over homocoupling (Scheme 1) requires a high degree of
discrimination between the arene and the silane in two
superficially similar SEAr processes. An explanation is, however,
provided by considering how the coupling partners are able to
meet the different demands made by the metal during the two
chemoselectivity-determining steps. Thus, while the high ipso
factor of the silyl moiety activates the arylsilane toward
formation of a WI (III) by stabilizing the TS leading to it,
activation of the ground state through donation of electron
density into the π-system of the arylsilane is minimal. This is
reflected in the low resonance-effect parameter and substituent
constant of the TMS moiety (R = −0.08, σp = −0.07), as
compared to e.g., MeO (R = −0.56, σp = −0.27) or Me (R =
−0.18, σp = −0.17).66 In contrast, the arene is unactivated
toward substitution ipso to H; the substrates employed,
however, all feature functionality that increases π-electron
density (i.e., π-basicity). Thus, if the first auration proceeds via
chemoselectivity-determining formation of a WI (III), the
silane will generally out-compete the arene. In this model rapid
equilibrium between arene and arylsilane π-complexes (II′ ↔
II) prior to the first auration is inconsequential. The converse is
true for the second auration, in which irreversible π-
complexation by the most π-basic arene provides the second
stage of chemoselectivity, i.e., formation of VI outcompetes
formation of VI′. The steric bulk of the TMS moiety may
augment this selectivity by destabilizing π-complex VI′,
comparable to the Baker−Nathan selectivity observed for the
arene, thereby accounting for the suppression of homocoupling
observed with the triethylsilyl analogue of 1.5 The overall
selectivity is thus controlled by two different elementary
processes in the sequential auration events. This shift in the
selectivity-determining step requires a change in the Au(III)
electrophile in the second auration, as provided by the silane-
derived σ-bonded aryl. Chemoselectivity begins to break down
in the limits of π-basic arylsilanes, or less π-basic arenes (e.g.,
benzene, fluorobenzene). Both situations result in the silane
competing effectively in the second auration event (VI′) and
thus homocoupling; the impact of chloride is discussed in the
SI.
The increase in regioselectivity observed upon arylation of 3-
bromothiophene 21 or toluene with ortho-substituted
arylsilanes (Chart 2C and D) cannot be used to distinguish
between scenarios in which formation of WI VII is either
irreversible and product determining, or reversible, although
KIEs support the latter. In the first case, a more sterically
Reductive Elimination. Investigations into C−C bond-
forming reductive elimination from a Au(III) center have
found generation of a coordinatively unsaturated intermediate
to be rate limiting.67 The conditions developed for gold-
H
dx.doi.org/10.1021/ja408712e | J. Am. Chem. Soc. XXXX, XXX, XXX−XXX

Journal of the American Chemical Society
Article
catalyzed arylation5 thus favor reductive elimination: ancillary
ligands are removed by oxidation during the induction period, a
polar, but poorly coordinating, solvent system is employed, and
weakly coordinating sulfonate anions are introduced through
both the acid additive and the precatalyst.
Oxidant Speciation. A key requirement of oxidative
arylation is the regeneration of a Au(III) species via
iodine(III)-mediated oxidation of Au(I). Despite the frequent
invocation of such a transformation in oxidative gold catalysis8
and its regular stoichiometric exploitation,68 the intimate
mechanism of the redox event has yet to be elucidated. Indeed,
the mechanism for the analogous oxidation of Pd(II) to Pd(IV)
has only very recently been investigated (DFT).69 Thorough
investigation of Au(I) oxidation is hampered by the distribution
of iodine(III) species in solution;19 we anticipate that insight
into this elementary step will prove central to further
development of the field.
sequester the catalyst via generation of a stable cycloaurate.
MeOH is found to perform three roles, acting as cosolvent,
TMS-scavenger and inhibitor. Given the inhibitory effect of the
alcohol, there is a practical benefit to performing reactions at
lower MeOH concentration, which is facilitated by use of a
more-soluble oxidant (generated with CSA rather than TsOH,
for example). Although reactions can be run under air, catalytic
turnover is suppressed by adventitious water.19 Inhibition by
Brønsted- and Lewis-bases arises from the dependence of rate
on CSA concentration, and the possibility of coordinative
saturation of the electrophilic metal center. The chemo-
selectivity results from electrophilic C−Si and C−H auration
in which formation of a Wheland Intermediate and a π-
complex, respectively, are selectivity determining, allowing
informed retrosynthetic disconnection to an electron-poor
arylsilane and an electron-rich arene; this insight may also
facilitate extension of the methodology to other arylating
agents/coupling partners.
CONCLUSIONS
Finally, we note that the evidence in support of a Au(I)/
Au(III) redox cycle suggests that this manifold may also
function in other oxidative, Au-catalyzed protocols employing
iodine(III) oxidants. This insight may facilitate informed
development of new catalysts and new catalytic cycles, leading
ultimately to alternative, more efficient oxidants.
■
The investigation reported herein has provided preliminary
insight into the mechanism of Au-catalyzed coupling of arenes
with aryltrimethylsilanes, from initial activation of the
precatalyst and the general structure of the catalytic cycle, to
the nature of some of the key elementary processes and the
origins of chemoselectivity and regioselectivity. A mechanistic
proposal that is consistent with all of our experimental
observations is illustrated in Figure 8, together with points
for consideration during practical application of the method-
ology.
ASSOCIATED CONTENT
* Supporting Information
■
S
Additional discussion, experimental procedures, kinetic data,
characterization data, and NMR spectra. This material is
available free of charge via the Internet at http://pubs.acs.org.
Kinetic analysis shows that PR3-ligated precatalysts are
associated with long induction periods, which are extended
by sterically demanding and/or less basic phosphines. This
allows not only a posteriori rationalization of previously
observed ligand-dependence, but also facilitates a priori
prediction of precatalyst suitability. In this regard, thtAuBr3
catalyzes arylation without an observable induction period, and
is both conveniently stable (air, moisture, and light) and
soluble. Rapid activation of thtAuBr3 is accompanied by
generation of bromination byproducts. The analogous in situ
anion exchange does not occur for gold chlorides, and this
appears to be the basis of the lower reactivity and selectivity
obtained with these precatalysts.19 Similar active species are
generated from Ph3PAuOTs and thtAuBr3 precatalysts via an
oxidative process, accounting for the empirical use of excess
iodine(III) reagent. The on-cycle metal is essentially ‘ligand-
free’, improving economy and making for simple precatalyst
selection, albeit at the expense of ligand-based catalyst tuning.
The extent to which ‘ligand-free’ conditions apply to related
oxidative gold-catalyzed processes8 remains to be fully explored.
Elucidation of the empirical catalytic rate law suggests that
the originally reported conditions5 (2 mol% Au and 1 equiv
arene vs 1 mol% Au and 2 equiv arene) are equivalent in terms
of reaction kinetics (but not necessarily in terms of selectivity).
The rate of arylation depends on the arene concentration, but
not on the silane concentration, validating use of the latter as
the limiting reagent. However, an excess of arylsilane may be
valuable in some circumstances, such as in the reaction of
strongly activated arenes with deactivated arylsilanes. The
effects of ortho-substitution of the arylsilane must also be
considered in synthetic strategy: although C−Si auration is
accelerated via steric decompression, and the regioselectivity of
C−H auration improved, the overall turnover frequency is
reduced. It is also important to note that o-biphenylsilanes
AUTHOR INFORMATION
■
Corresponding Authors
guy.lloyd-jones@ed.ac.uk
chris.russell@bristol.ac.uk
Notes
The authors declare no competing financial interest.
ACKNOWLEDGMENTS
■
Miss N. S. Townsend and Miss L. C. Forfar for X-ray
crystallography; Dr Z. M. Hudson for cyclic voltammetry;
Umicore AG & Co. KG for HAuCl4; the Bristol Chemical
Synthesis CDT (studentship to L.T.B.; EPSRC EP/G036764/
1).
REFERENCES
■
(1) (a) O’Brien, C. J.; Kantchev, E. A. B.; Valente, C.; Hadei, N.;
Chass, G. A.; Lough, A.; Hopkinson, A. C.; Organ, M. G. Chem.Eur.
J. 2006, 12, 4743. (b) Organ, M. G.; Avola, S.; Dubovyk, I.; Hadei, N.;
Kantchev, E. A. B.; O’Brien, C. J.; Valente, C. Chem.Eur. J. 2006, 12,
4749. (c) Martin, R.; Buchwald, S. L. Acc. Chem. Res. 2008, 41, 1461.
(d) Biscoe, M. R.; Fors, B. P.; Buchwald, S. L. J. Am. Chem. Soc. 2008,
130, 6686. (e) Lundgren, R. J.; Stradiotto, M. Chem.Eur. J. 2012, 18,
9758.
(2) (a) Darses, S.; Genet, J.-P. Chem. Rev. 2007, 108, 288.
(b) Molander, G. A.; Ellis, N. Acc. Chem. Res. 2007, 40, 275.
(c) Denmark, S. E.; Regens, C. S. Acc. Chem. Res. 2008, 41, 1486.
(d) Gillis, E. P.; Burke, M. D. Aldrichimica Acta 2009, 42, 17.
(e) Knapp, D. M.; Gillis, E. P.; Burke, M. D. J. Am. Chem. Soc. 2009,
131, 6961. (f) Lennox, A. J. J.; Lloyd-Jones, G. C. Isr. J. Chem. 2010,
50, 664.
(3) (a) Alberico, D.; Scott, M. E.; Lautens, M. Chem. Rev. 2007, 107,
174. (b) Ackermann, L.; Vicente, R.; Kapdi, A. R. Angew. Chem., Int.
Ed. 2009, 48, 9792. (c) McGlacken, G. P.; Bateman, L. M. Chem. Soc.
I
dx.doi.org/10.1021/ja408712e | J. Am. Chem. Soc. XXXX, XXX, XXX−XXX

Journal of the American Chemical Society
Article
Rev. 2009, 38, 2447. (d) Lyons, T. W.; Sanford, M. S. Chem. Rev. 2010,
110, 1147.
(b) Funaki, K.; Kawai, H.; Sato, T.; Oi, S. Chem. Lett. 2011, 40, 1050.
(c) Funaki, K.; Sato, T.; Oi, S. Org. Lett. 2012, 14, 6186.
(17) Stoichiometric arylation of Au(I) has been reported in the
presence of base using electron-deficient arenes (a) Lu, P.; Boorman,
T. C.; Slawin, A. M. Z.; Larrosa, I. J. Am. Chem. Soc. 2010, 132, 5580,
boronic acids (b) Partyka, D. V.; Zeller, M.; Hunter, A. D.; Gray, T. G.
Angew. Chem., Int. Ed. 2006, 45, 8188. (c) Partyka, D. V.; Zeller, M.;
Hunter, A. D.; Gray, T. G. Inorg. Chem. 2012, 51, 8394. See also ref
10f), and siloxanes (ref 16a), in the absence of base using tetraaryl
borates (d) Weber, S. G.; Zahner, D.; Rominger, F.; Straub, B. F.
Chem. Commun. 2012, 48, 11325, and stannanes (e) Meyer, N.;
Sivanathan, S.; Mohr, F. J. Organomet. Chem. 2011, 696, 1244.
(18) Roth, K. E.; Blum, S. A. Organometallics 2010, 29, 1712.
(19) See SI.
(4) (a) Boutadla, Y.; Davies, D. L.; Macgregor, S. A.; Poblador-
Bahamonde, A. I. Dalton Trans. 2009, 5820. (b) Balcells, D.; Clot, E.;
Eisenstein, O. Chem. Rev. 2010, 110, 749. (c) Livendahl, M.;
Echavarren, A. M. Isr. J. Chem. 2010, 50, 630. (d) Ackermann, L.
Chem. Rev. 2011, 111, 1315.
(5) Ball, L. T.; Lloyd-Jones, G. C.; Russell, C. A. Science 2012, 337,
1644.
(6) Very low catalyst loadings of have been achieved for
mechanistically simpler Au-catalyzed reactions: Blanco Jaimes, M. C.;
Bohling, C. R. N.; Serrano-Becerra, J. M.; Hashmi, A. S. K. Angew.
̈
Chem., Int. Ed. 2013, 52, 7963.
(7) Jones, C. J.; Taube, D.; Ziatdinov, V. R.; Periana, R. A.; Nielsen,
R. J.; Oxgaard, J.; Goddard, W. A., III. Angew. Chem., Int. Ed. 2004, 43,
4626.
(8) (a) Wegner, H. A. Chimia 2009, 63, 44. (b) Wegner, H. A.;
Auzias, M. Angew. Chem., Int. Ed. 2011, 50, 8236. (c) Hopkinson, M.
N.; Gee, A. D.; Gouverneur, V. Chem.Eur. J. 2011, 17, 8248. (d) de
Haro, T.; Nevado, C. Synthesis 2011, 2530.
(9) Calculations support a Au(I)/Au(II)-dimer redox-couple for
oxidative olefin heteroarylation (ref 10a); a Au(I)/Au(III) couple was
suggested for a closely related transformation: (a) Wang, W.; Jasinski,
J.; Hammond, G. B.; Xu, B. Angew. Chem., Int. Ed. 2010, 49, 7247.
Homocoupling during gold-catalyzed cyclisations has been ascribed to
stoichiometric reduction of Au(III): (b) Hashmi, A. S. K.; Blanco, M.
C.; Fischer, D.; Bats, J. W. Eur. J. Org. Chem. 2006, 1387.
(10) Mechanistic investigations have so far focused on Selectfluor-
mediated couplings of ArB(OH)2: (a) Tkatchouk, E.; Mankad, N. P.;
Benitez, D.; Goddard, W. A., III; Toste, F. D. J. Am. Chem. Soc. 2011,
(20) Control experiments detailed in ref 5 indicate diaryliodonium
salts to be unreactive. See also ref 10g.
(21) (a) Neiland, O. Y.; Karele, B. Y. Zh. Org. Khim. 1970, 6, 885.
(b) Koser, G. F.; Wettach, R. H. J. Org. Chem. 1977, 42, 1476.
(c) Stang, P. J.; Surber, B. W.; Chen, Z. C.; Roberts, K. A.; Anderson,
A. G. J. Am. Chem. Soc. 1987, 109, 228. (d) Yusubov, M. S.; Wirth, T.
Org. Lett. 2005, 7, 519.
(22) Crabtree, R. H. Chem. Rev. 2011, 112, 1536.
(23) A range of cyclic voltammetry experiments that aimed to
characterize a ‘ligand-free’ Au(III) species were uninformative due to
interference by iodine(III) reduction, see SI. Formation of Ph3PO
from Ph3PAuCl has also been noted with Selectfluor: see ref 10c.
Oxidative oxyarylation under similar conditions forms (Ph3P)2Au+:
(a) Brenzovich, W. E.; Benitez, D.; Lackner, A. D.; Shunatona, H. P.;
Tkatchouk, E.; Goddard, W. A., III; Toste, F. D. Angew. Chem., Int. Ed.
2010, 49, 5519.
́
133, 14293. (b) Faza, O. N.; Lopez, C. S. J. Org. Chem. 2013, 78, 4929.
Selected early examples of ‘F+’-mediated processes for which the
Au(I)/Au(III) manifold has been proposed: (c) Zhang, G.; Peng, Y.;
Cui, L.; Zhang, L. Angew. Chem., Int. Ed. 2009, 48, 3112. (d) Cui, L.;
Zhang, G. Z.; Zhang, L. M. Bioorg. Med. Chem. Lett. 2009, 19, 3884.
(e) Hopkinson, M. N.; Tessier, A.; Salisbury, A.; Giuffredi, G. T.;
Combettes, L. E.; Gee, A. D.; Gouverneur, V. Chem.Eur. J. 2010, 16,
4739. (f) Hashmi, A. S. K.; Ramamurthi, T. D.; Rominger, F. J.
Organomet. Chem. 2009, 694, 592. (g) Hashmi, A. S. K.; Ramamurthi,
T. D.; Todd, M. H.; Tsang, A. S.-K.; Graf, K. Aust. J. Chem. 2010, 63,
1619.
(11) Intermolecular auration in aprotic/nondonor media: (a) Khar-
asch, M. S.; Isbell, H. S. J. Am. Chem. Soc. 1931, 53, 3053. (b) Liddle,
K. S.; Parkin, C. J. Chem. Soc., Chem. Commun. 1972, 26. (c) de Graaf,
P. W. J.; Boersma, J.; Vanderkerk, G. J. M. J. Organomet. Chem. 1976,
105, 399. (d) Fuchita, Y.; Utsunomiya, Y.; Yasutake, M. J. Chem. Soc.,
Dalton Trans. 2001, 2330. In protic/donor media: (e) Davidson, J.
M.; Triggs, C. J. Chem. Soc. A 1968, 1324. (f) Nizova, G. V.; Shulpin,
G. B. React. Kinet. Catal. Lett. 1982, 20, 69.
(24) Initially formed Ph3PX2 will rapidly solvolyze to Ph3PO.
(25) For example, oxidation of R3PAuI with I2 occurs for PR3 =
PMe2Ph but not PMePh2 or PPh3, and for PMe3 but not PtBu3. See:
(a) Schneider, D.; Schier, A.; Schmidbaur, H. Dalton Trans. 2004,
1995.
(26) Reductive elimination of Ar3P+-O2CCF3 from Pd is accelerated
by basic phosphines, as the TS resembles nucleophilic attack by Ar3P
on the trifluoroacetate oxygen. The converse is true for Pd(OAc)2.
See: (a) Amatore, C.; Carre, E.; Jutand, A.; M’Barki, M. A.
Organometallics 1995, 14, 1818. (b) Amatore, C.; Jutand, A.;
Lemaitre, F.; Ricard, J. L.; Kozuch, S.; Shaik, S. J. Organomet. Chem.
2004, 689, 3728.
(27) Clavier, H.; Nolan, S. P. Chem. Commun. 2010, 46, 841.
(28) Whether the bromination side product is derived from the
arylsilane or the arene (or both) depends on their relative reactivity.
For example, in the coupling between 1 and 8, the latter is brominated
(→ 10), whereas in the coupling of 1 and toluene, bromodesilylation
of the silane generates 4-bromofluorobenzene (19F NMR, GC−MS).
(29) Two equivalents of oxidant are consumed in the reaction of
thtAuCl with F-IBDA/CSA. See SI.
(12) (a) Kar, A.; Mangu, N.; Kaiser, H. M.; Beller, M.; Tse, M. K.
Chem. Commun. 2008, 386. (b) Kar, A.; Mangu, N.; Kaiser, H. M.;
Tse, M. K. J. Organomet. Chem. 2009, 694, 524. (c) de Haro, T.;
Nevado, C. J. Am. Chem. Soc. 2010, 132, 1512. (d) Qiu, D.; Zheng, Z.;
Mo, F.; Xiao, Q.; Tian, Y.; Zhang, Y.; Wang, J. Org. Lett. 2011, 13,
4988. (e) Pradal, A.; Toullec, P. Y.; Michelet, V. Org. Lett. 2011, 13,
6086.
(30) For nonmetal-mediated oxidative bromination, see: Braddock,
D. C.; Cansell, G.; Hermitage, S. A. Synlett 2004, 461.
(31) For examples of benzylic activation during electrophilic
metalation, see: (a) Kochi, J. K.; Tang, R. T.; Bernath, T. J. Am.
Chem. Soc. 1973, 95, 7114. (b) Lau, W.; Kochi, J. K. J. Am. Chem. Soc.
1986, 108, 6720. Reduction of Au(I) in the presence of silver salts
results in benzylic activation (ref 14).
(32) Stoichiometric cross-coupling between an aryl−Au(III) species
and an arene has been reported: (a) Cambeiro, X. C.; Boorman, T. C.;
Lu, P. F.; Larrosa, I. Angew. Chem., Int. Ed. 2013, 52, 1781. (b) Hofer,
M.; Nevado, C. Tetrahedron 2013, 69, 5751.
(33) Uson, R.; Vicente, J.; Cirac, J. A.; Chicote, M. T. J. Organomet.
́
Chem. 1980, 198, 105.
(34) Reaction is inhibited when [CSA] < [F-IBDA]; see SI.
(35) For the role of MeOH in electrophilic palladation of arylsilanes,
see: Steenwinkel, P.; Gossage, R. A.; Maunula, T.; Grove, D. M.; van
Koten, G. Chem.Eur. J. 1998, 4, 763.
́ ́
(13) (a) Herrero-Gomez, E.; Nieto-Oberhuber, C.; Lopez, S.; Benet-
Buchholz, J.; Echavarren, A. M. Angew. Chem., Int. Ed. 2006, 45, 5455.
(b) Lavallo, V.; Frey, G. D.; Kousar, S.; Donnadieu, B.; Bertrand, G.
Proc. Natl. Acad. Sci. U.S.A. 2007, 104, 13569.
(14) Savjani, N.; Bew, S. P.; Hughes, D. L.; Lancaster, S. J.;
Bochmann, M. Organometallics 2011, 31, 2534.
(15) ‘C−H/C−Si auration’ refers to the net transformation from free
reactants to Au-σ-bonded arene, not solely the C−X cleavage step.
(16) Transmetalation via a siliconate is unlikely as electronegative
substituents on Si and/or basic additives are required. We are aware of
one example of Si to Au aryl transmetalation employing arylsiloxane
and Au(I): (a) Dupuy, S.; Slawin, A. M. Z.; Nolan, S. P. Chem.Eur. J.
2012, 18, 14923. For aryltrimethylsilanes in cross-coupling, see:
J
dx.doi.org/10.1021/ja408712e | J. Am. Chem. Soc. XXXX, XXX, XXX−XXX

Journal of the American Chemical Society
Article
(36) Here, ‘apparent’ activation parameters (ΔH and ΔS) refer to the
net barriers that must be overcome when >1 elementary step is
required to reach the turnover-limiting TS from the resting state.
(37) Chipperfield, J. R.; France, G. D.; Webster, D. E. J. Chem. Soc.,
Perkin Trans. 2 1972, 405.
(38) The low ρ value is also consistent with a TS in which the
partially formed Au−C bond stabilizes the developing WI (cf. ref 18),
or with attack by a largely ionic electrophile: (a) Brown, H. C.;
McGary, C. W. J. Am. Chem. Soc. 1955, 77, 2310.
(39) Taylor, R. Electrophilic Aromatic Substitution; John Wiley & Sons
Ltd: Chichester, 1990.
(63) Arylation of Au(I) under basic conditions, as determined by
intermolecular competition, exhibits a primary KIE of 5 (ref 17a).
(64) (a) Hartwig, J. F. Organotransition Metal Chemistry: From
Bonding to Catalysis; University Science Books: CA, 2010. For
associative substitution at Au(III): (b) Baddley, W. H.; Basolo, F.
Inorg. Chem. 1964, 3, 1087. Neutral Au(III): (c) Cattalini, L.;
Ricevuto, V.; Orio, A.; Tobe, M. L. Inorg. Chem. 1968, 7, 51. Anionic
Au(III): (d) Cattalini, L.; Orio, A.; Tobe, M. L. J. Am. Chem. Soc. 1967,
89, 3130. (e) Robb, W. Inorg. Chem. 1967, 6, 382.
(65) Basolo, F.; Chatt, J.; Gray, H. B.; Pearson, R. G.; Shaw, B. L. J.
Chem. Soc. 1961, 2207.
(66) Hansch, C.; Leo, A.; Taft, R. W. Chem. Rev. 1991, 91, 165.
(67) (a) Tamaki, A.; Magennis, S. A.; Kochi, J. K. J. Am. Chem. Soc.
1974, 96, 6140. (b) Komiya, S.; Albright, T. A.; Hoffmann, R.; Kochi,
J. K. J. Am. Chem. Soc. 1976, 98, 7255. (c) Komiya, S.; Kochi, J. K. J.
Am. Chem. Soc. 1976, 98, 7599. (d) Lawrence Kuch, P.; Stuart Tobias,
R. J. Organomet. Chem. 1976, 122, 429. (e) Komiya, S.; Shibue, A.
(40) Ess, D. H.; Nielsen, R. J.; Goddard, W. A., III; Periana, R. A. J.
Am. Chem. Soc. 2009, 131, 11686.
(41) Yi, C. S.; Yun, S. Y. J. Am. Chem. Soc. 2005, 127, 17000.
(42) Thompson, M. E.; Baxter, S. M.; Bulls, A. R.; Burger, B. J.;
Nolan, M. C.; Santarsiero, B. D.; Schaefer, W. P.; Bercaw, J. E. J. Am.
Chem. Soc. 1987, 109, 203.
(43) DeYonker, N. J.; Foley, N. A.; Cundari, T. R.; Gunnoe, T. B.;
Petersen, J. L. Organometallics 2007, 26, 6604.
(44) Andrews, L. J.; Keefer, R. M. Molecular Complexes in Organic
Chemistry; Holden-Day: San Francisco, 1964.
(45) Olah, G. A.; Kuhn, S. J.; Flood, S. H. J. Am. Chem. Soc. 1961, 83,
Organometallics 1985, 4, 684. (f) Vicente, J.; Dolores Bermudez, M.;
́
Escribano, J. Organometallics 1991, 10, 3380.
(68) (a) Casado, R.; Contel, M.; Laguna, M.; Romero, P.; Sanz, S. J.
Am. Chem. Soc. 2003, 125, 11925. (b) Teets, T. S.; Nocera, D. G. J.
Am. Chem. Soc. 2009, 131, 7411. (c) Gaillard, S.; Slawin, A. M. Z.;
Bonura, A. T.; Stevens, E. D.; Nolan, S. P. Organometallics 2009, 29,
394. (d) Hofer, M.; Nevado, C. Eur. J. Inorg. Chem. 2012, 2012, 1338.
(69) Lian, B.; Zhang, L.; Chass, G. A.; Fang, D.-C. J. Org. Chem.
2013, 78, 8376.
4571.
(46) Olah, G. A.; Hashimoto, I.; Lin, H. C. Proc. Natl. Acad. Sci.
U.S.A. 1977, 74, 4121.
(47) Exner, O.; Bohm, S. J. Chem. Soc., Perkin Trans. 2 1997, 1235.
(48) Benkeser, R. A.; Liston, T. V.; Stanton, G. M. Tetrahedron Lett.
1960, 1, 1.
(49) (a) Vicente, J.; Bermudez, M.-D.; Chicote, M.-T.; San
Santano, M.-J. J. Organomet. Chem. 1990, 381, 285. (b) Vicente, J.;
́
chez-
́
Bermudez, M.-D.; Chicote, M.-T.; San
́
chez-Santano, M.-J. J. Chem.
, R.; Laguna, A. Coord. Chem.
́
Soc., Chem. Commun. 1989, 141. (c) Uson
́
Rev. 1986, 70, 1.
(50) Price, G. A.; Flower, K. R.; Pritchard, R. G.; Brisdon, A. K.;
Quayle, P. Dalton Trans. 2011, 40, 11696.
(51) Usui, Y.; Noma, J.; Hirano, M.; Komiya, S. Inorg. Chim. Acta
2000, 309, 151.
(52) (a) Akhrem, I. S.; Chistova, N. M.; Mysov, E. I.; Volpin, M. E. J.
Organomet. Chem. 1974, 72, 163. (b) Akhrem, I. S.; Chistovalova, N.
M.; Safoyan, H. N.; Volpin, M. E.; Moiseev, I. J. Organomet. Chem.
1979, 178, C23. (c) Akhrem, I. S.; Chistovalova, N. M.; Malysheva, A.
V.; Volpin, M. E. B. Acad. Sci. USSR Ch. 1983, 32, 2417. (d) Mansuy,
D.; Bartoli, J. F.; Chottard, J. C. J. Organomet. Chem. 1974, 71, C32.
(e) Mansuy, D.; Pusset, J.; Chottard, J. C. J. Organomet. Chem. 1976,
105, 169.
(53) Szele, I. Helv. Chim. Acta 1981, 64, 2733.
(54) Note that KIEs from CH3/CD3 substitution are typically small:
(a) Swain, C. G.; Knee, T. E.; Kresge, A. J. J. Am. Chem. Soc. 1957, 79,
505. (b) Lauer, W.; Koons, C. J. Org. Chem. 1959, 24, 1169.
(55) For example: protodesilylation, ρ+ = −5.3; bromodesilylation, ρ+
= −6.8 (ref 39).
(56) Benkeser, R. A.; Hoke, D. I.; Hickner, R. A. J. Am. Chem. Soc.
1958, 80, 5294.
(57) Benkeser, R. A.; Krysiak, H. R. J. Am. Chem. Soc. 1954, 76, 6353.
(58) The importance of the counteranion has recently been
demonstrated for the transmetallation from Au(I) to Pd(II), which
fails both in silico and in vitro in the absence of a halide source:
Hansmann, M. M.; Pernpointner, M.; Dopp, R.; Hashmi, A. S. K.
̈
Chem.Eur. J. 2013, 19, 15290.
(59) Ghosh, M. K.; Tilset, M.; Venugopal, A.; Heyn, R. H.; Swang, O.
J. Phys. Chem. A 2010, 114, 8135.
(60) Bellamy, L. J.; Morgan, K. J.; Pace, R. J. Spectrochim. Acta 1966,
22, 535.
(61) Intramolecular exchange of phenyl/benzene ligands at Au(III)
likely proceeds via σ-complex assisted metathesis (ref 59).
(62) Inverse SKIEs have been reported for SEAr reactions in which
WI formation is kinetically significant, see ref 45 and Borodkin, G. I.;
Zaikin, P. A.; Shubin, V. G. Tetrahedron Lett. 2006, 47, 2639.
K
dx.doi.org/10.1021/ja408712e | J. Am. Chem. Soc. XXXX, XXX, XXX−XXX