nucleophiles, the olefin scope has been entirely restricted to
substrates bearing a single alkyl substituent.[16] Herein, we
report the gold-catalysed, three-component methoxyaryla-
tion of both styrenes and gem-disubstituted olefins in the
successful (entry 1), the addition of two equivalents of p-tol-
uenesulfonic acid monohydrate (p-TSA) afforded the de-
sired product 1 in a promising yield of 20% (entry 2). The
facile nature of ligand exchange at iodine
that the active oxidant was likely based on [hydroxy-
(tosyloxy)iodo]benzene (HTIB, Koserꢀs reagent)[18] rather
AHCTUNGTRENNUNG
(III)[17] suggested
ACHTUNGTRENNUNG
presence of an iodineACTHNUGTRNEUNG(III) oxidant.
than PhIACTHNUTRGENNG(U OAc)2; this supposition was supported by the very
Results and Discussion
similar yields obtained with either freshly prepared[19] or
commercially sourced HTIB (entry 3).[20]
Oxidant development and reaction optimisation: The pres-
ence of basic/nucleophilic additives (e.g., Cs2CO3, TBAF=
tetrabutylammonium fluoride) was shown to have a detri-
mental effect on the Selectfluor-mediated oxyarylation of
monosubstituted olefins.[10] In addition, the combination of
nucleophilic promoters (e.g., TBAF, TBAT=tetrabutylam-
monium triphenyldifluorosilicate) and alternative oxidants
The yield of 1 was doubled by portionwise addition of
HTIB (entry 4), but, despite significant efforts, could not be
improved further. Subsequent investigation of iodineACHTUNGTRENNUNG(III) re-
agents derived from 2-iodobenzoic acid[21] revealed that the
use of the tosyloxy analogue IBA-Ts (entry 6) was prefera-
ble to either the parent benziodoxolone IBA (1-hydroxy-
1,2-benziodoxol-3ACTHNUTRGNEUNG(1H)-one; entry 5) or its acyclic relative
(e.g., PhI
(OAc)2) afforded only traces of the desired oxyary-
HTIB. The yield of 1 was increased to 83% by employing a
combination of IBA and two equivalents of p-TSA (entry 7)
in place of preformed IBA-Ts; this result could not be im-
proved by varying the acid additive (entries 8–10)[23] or the
aryl source (entries 11–13). IBX (2-iodoxybenzoic acid)
proved an unsuitable oxidant (entries 14 and 15), presuma-
bly owing to the competing reaction between the iodine(V)
reagent and the alcohol.[24] Importantly, control reactions[25]
indicated the need for both the acid and the oxidant, as well
as a gold precatalyst; the failure of palladium, and various
other late transition metals, to catalyse the transformation
provided evidence against low-level impurities being respon-
sible for the catalytic activity ascribed to gold.[26]
(Table 1). Thus, whilst employing PhIACTHNUTRGENUG(N OAc)2 in place of Se-
lectfluor (under previously optimised conditions) proved un-
Table 1. Oxidant screen.[a]
Subsequent reaction optimisation[25] revealed that both
AuI and AuIII species—as well-defined complexes, simple
salts or in combination with AgI salts—were able to act as
precatalysts for the IBA/p-TSA-mediated oxyarylation
(Table 2, entries 1–12). As with our previously reported
methodology,[9] the reaction was sensitive to the steric and
electronic environment about the gold centre: [AuClACHTUNGTRENNUNG(PPh3)]
proved a more efficient precatalyst than complexes of bulky
and/or strong donor ligands (entry 1 vs. entries 6–10). Inter-
Entry
Oxidant
Additive
Yield [%][b]
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
PhI
PhI
U
–
<1
20
23
48[c]
0
55
83[d]
65
15
4
p-TSA
–
–
–
–
p-TSA
MsOH
TfOH
TFA
p-TSA
p-TSA
p-TSA
–
estingly, the reactivity of [AuCl
ACHTUNGTRENNUNG
HTIB
HTIB
IBA
IBA-Ts
IBA
IBA
IBA
IBA
IBA
IBA
IBA
IBX
IBX
by the digold complexes [(AuX) CTHNUGTRENNUNG
dppm=bis(diphenylphosphino)methane; X=Cl, Br).[27]
Whilst a significant halide effect was observed for the Se-
lectfluor system at approximately 508C, this was found to di-
minish at higher temperatures,[6b,8a,10] as is reflected in our
own results (compare entry 11, X=Cl with entry 12, X=
Br).
3[e]
0[f]
0[g]
44
37
Notably, upon changing the reaction solvent[25] to benzo-
trifluoride (PhCF3) and diluting from 0.1m (entry 13) to
0.05m (entry 14), the yield of 1 was significantly improved
and the formation of biaryl side products was largely sup-
pressed (<3%).[28] However, no additional benefits resulted
from further dilution. Subsequent investigations indicated
that, whilst the amounts of arylsilane, IBA and p-TSA could
be reduced, two molar equivalents of each reagent were re-
quired for consistently high yields, particularly when em-
ploying more challenging substrates.[29] In addition, the reac-
tion proceeded most efficiently at 708C; hence, optimised
p-TSA
[a] Conditions: [AuClACHTUNGTRENNUNG(PPh3)] (5 mol%), 1-octene (0.1 mmol), PhSiMe3
(0.2 mmol), oxidant (0.2 mmol), additive (0.2 mmol), MeCN/MeOH (9:1,
1 mL), 708C, 15 h. Ms=methanesulfonyl; Tf= trifluoromethanesulfonyl;
TFA=trifluoroacetic acid. [b] Yields are an average of at least 2 repeti-
tions and were determined by GC-FID (FID=flame ionisation detector)
against an internal standard. [c] HTIB was added in six equal portions
over 100 min. [d] 1 was not formed in the absence of [AuCl
[e] PhB(OH)2 was used in place of PhSiMe3. [f] PhSi(OMe)3 was used in
place of PhSiMe3. [g] PhSiMe2OH was used in place of PhSiMe3.
(PPh3)].[25]
ACHTUNGTRENNUNG
AHCTUNGTRENNUNG
2932
ꢁ 2012 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2012, 18, 2931 – 2937