Aryl trifluoroborates in Suzuki-Miyaura coupling: The roles of endogenous aryl boronic acid and fluoride

Article abstract of DOI:10.1002/anie.201001522

Undercover agents: The biaryl coupling of an aryltrifluoroborate with an aryl bromide involves in situ hydrolysis of the boron reagent. The hydrolysis products are key components in ensuring that the reaction proceeds with high efficiency and avoids the e

Full text of DOI:10.1002/anie.201001522

Communications
DOI: 10.1002/anie.201001522
Reaction Mechanisms
Aryl Trifluoroborates in Suzuki–Miyaura Coupling: The Roles of
Endogenous Aryl Boronic Acid and Fluoride**
Mike Butters, Jeremy N. Harvey, Jesus Jover, Alastair J. J. Lennox, Guy C. Lloyd-Jones,* and
Paul M. Murray
A wide range of organoboron reagents can be used as
alternative reagents to boronic acids in Suzuki–Miyaura (SM)
coupling reactions.[1] The readily prepared,[2] convenient to
handle potassium trifluoroborates, RBF3K, which have been
developed by the groups of Genet[3] and Molander,[4] are often
the reagents of choice for these transformations. Although
extensive optimization of the base, solvent, and temperature
is required for each class of substrate,[4,5] their utility in SM
coupling reactions has led to their widespread commercial
availability. Apart from a preliminary study in 2003,[5a] their
Scheme 1. SM coupling of trifluoroborate 1 with bromide 2 to generate
biaryl 3 together with the three major side products arising from
mode of action has not been investigated in detail, and the
protodeboronation (6), homocoupling (9), and oxidation (10).
origin of their efficacy[4,5a] remains to be elucidated. Herein,
we report the SM coupling of aryl trifluoroborate 1 with aryl
bromide 2 to generate biaryl[5a,j] 3 (Scheme 1). We show that
endogenous aryl boronic acid 4 and fluoride, both arising
from 1, play key roles in the coupling reaction, being involved
at all stages: from catalyst activation and catalytic turnover,
through to the inhibition of side reactions. Collectively, these
phenomena result in the exceptional performance of the
reagent in the SM coupling.[4]
The SM coupling of 1 with 2 was studied in a toluene/
water (3:1)[5b–e] biphasic solution, and in a tetrahydrofuran/
water (10:1)[5f–i] solution, both systems being commonly
employed for the SM coupling of trifluoroborates.[5] The
reactions in toluene/water, failed to go to completion:
turnover ceased after 6 hours, affording 55% of the base-
catalyzed[6] protodeboronation product 6[7] and ꢀ 32% of
coupling product 3. In aqueous tetrahydrofuran (Scheme 1)
the reaction proceeded much more efficiently (5.5 h; > 95%
yield of 3), with few side products (ꢀ 0.1–2%), even when the
reaction was performed in air.
The performance of aryl boronic acid reagents can be
improved by the addition of KF,[8] whereas trifluoroborate
reagents require aqueous solvent systems for SM coupling
with standard substrates.[9–11] This observation has led to
suggestions that mixed borates, [RBF(3ꢁn)(OH)n]ꢁ,[12,13] are
the true transmetalating species.[4b,5a,10,13b] Base titration of 1
in a solely aqueous medium (D2O) was monitored by 19F and
11B NMR spectroscopy. Trifluoroborate 1 underwent hydrol-
ysis via boronic acid 4 to give boronate 5; the transformation
required approximately three equivalents of K2CO3 or
Cs2CO3, or six equivalents of KOH to proceed to completion.
At ambient temperature, boronate 5 slowly gave rise to
fluorobenzene 6 by protodeboronation;[6] the process was
substantially faster at 558C. Rapid equilibrium between 4 and
5 gave rise to time-averaged 19F NMR chemical shifts (p-F-Ar
nuclei), from which analysis of DdF values versus [base] was
used to establish the mol% of boronate 5 (e.g. Figure 1a).
When the dibasic nature of M2CO3 was taken into
account, there was no significant difference in the curve
In contrast, reaction of the boronic acid (4) under
identical conditions, gave 3 in variable yield, and afforded
substantially more of side products 9/10 (2–40%), compared
to trifluoroborate substrate 1.
[*] Prof. Dr. J. N. Harvey, Dr. J. Jover, A. J. J. Lennox,
Prof. Dr. G. C. Lloyd-Jones
School of Chemistry, University of Bristol, Cantock’s Close
Bristol, BS8 1TS (UK)
Fax: (+44)117-929-8611
E-mail: guy.lloyd-jones@bris.ac.uk
Dr. M. Butters, Dr. P. M. Murray
AstraZeneca, Severn Road, Hallen, Bristol BS10 7ZE (UK)
[**] We thank AstraZeneca PR&D for generous funding. G.C.L.J. is a
Figure 1. a) Titration of 1 and 4 with K2CO3 in D2O (71 mm) and in
THF/D2O (8 mm). There is no significant difference in using Cs2CO3 or
KOH as the base. b) The effect of D2O concentration on the fraction of
5 (mol%) in 4/5, generated from 1+4 equiv K2CO3 in THF.
Royal Society Wolfson Research Merit Award Holder.
Supporting information for this article is available on the WWW
under http://dx.doi.org/10.1002/anie.201001522.
5156
ꢀ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2010, 49, 5156 –5160

Angewandte
Chemie
obtained with K2CO3, Cs2CO3, or KOH, and after addition of
the theoretical amount of base required to replace the three
fluoride ions in 1, the shape of the curve was identical to that
observed for the boronic acid 4. At low concentrations of
base, the effect of additional fluoride (4, 8, and 12 equiv KF)
on the titration of 1 with K2CO3 was to slightly increase the
proportion of 5, because of the slight increase in the pH value
caused by the aqueous fluoride. However, under the concen-
trations of base employed for the SM coupling, the effect was
negligible (Scheme 1).
Protodeboronation to give 6 is the major pathway for the
inefficient SM coupling of 1 in a toluene/water biphase,[7] but
is not detected under the standard aqueous tetrahydrofuran
conditions[5f–i] (see Scheme 1). The effect of tetrahydrofuran
on homogeneous solutions of 5, generated in situ from 1 with
4 equivalents K2CO3 in D2O, was studied by 19F NMR
spectroscopy. As the D2O concentration was reduced from
55m (100% D2O) to 0m (100% THF; see Figure 1b), the
equilibrium population shifted from ꢂ 98% boronate (5) to
ꢂ 98% boronic acid (4), thereby resulting in complete
suppression of the protodeboronation reaction (5!6). In
5.1m D2O (10:1 THF/water, Scheme 1),[5f–i] where the equi-
librium population of 5 is almost independent of base
concentration (Figure 1a), protodeboronation is negligible:
< 0.1% 6 was generated in 12 days, as opposed to 46% 6 in
47m D2O.
Scheme 3. Interconversion of [D4]-4/[D0]-1 with [D0]-4/[D4]-1 in 10:1
THF/D2O (5.1m D2O) with 3 equiv Cs2CO3 at 558C. Also shown is the
palladium-catalyzed SM coupling with 2; see Figure 3a.
spectroscopy, although they must be present, albeit at low
concentrations (ꢀ 60 mm), to facilitate the indirect equilibri-
um between 1 and 4. To contribute significantly to the
catalytic flux for SM coupling, these mixed species, such as 7
and 8, need to be exceptionally active towards transmetala-
tion. To assess this aspect, we studied (B3LYP/6-
31G*,lacv3p)[16] phenyl transfer from [PhBX3]ꢁ (X = F, OH)
to [Pd(Br)Ar(L)n] (L = PPh3; n = 1,2), generated by oxidative
addition of PhBr to [Pd(PPh3)2], in tetrahydrofuran. As
illustrated in Figure 2, the energy barrier towards transmeta-
19F NMR analysis of the SM coupling reaction confirmed
that 1 undergoes hydrolysis,[13–15] but the growth in biaryl
product (3) was substantially in advance of the boronic acid 4
for the first 20–30 turnovers. No intermediate mixed species
(7, 8, or the related neutral intermediates, ArBX2; X =
F,OH,OD) were observed under these conditions.[13a] How-
ever, at higher water concentrations, an intermediate was
ꢁ
detected, reaching a maximum (8% of the total Ar B
species) when 0.5 equivalents K2CO3 was added in a THF/
D2O (1:10, 45m D2O) solution (Scheme 2).
Figure 2. Energies (kcalmolꢁ1) for Ph transfer to [Pd(Br)Ar(L)n] from
[PhBX3]ꢁ (X=F, OH; L=PPh3); lower energy (n=1) complexes are
shown; (n=2) follows the same order, but are ca. 20 kcalmolꢁ1
higher: IA2/TSA2=ꢁ33.9:5.5); ꢁ31.9:13.2; ꢁ31.2:17.6; ꢁ26.9:10.9.[16]
Scheme 2. 19F EXSY contacts (a t=25–40 ms) observed between
7,8 and 5 on hydrolysis of [d0]-1 in 1:10 THF/D2O (45m D2O).
19F EXSY analysis showed that this species, tentatively
assigned by integration as [ArBF2(OD)]ꢁK+ (8),[16] was in
rapid equilibrium (t ꢀ 40 ms) with boronate 5 and KF, but not
with trifluoroborate 1. However, reversible generation of
trifluoroborate 1 was confirmed under the SM coupling
conditions from the observation of a smooth equilibration
between [D0]-1 and [D4]-4 with [D4]-1 and [D0]-4, respectively
(Scheme 3).
lation increases with increasing ligation of boron by fluoride,
consistent with the decreasing ability of the boron reagent to
complex to Pd (intermediate IA2) and the reduced nucleo-
philicity of the phenyl moiety (transition state TSA2). We
expect catalytic flux to proceed almost exclusively through
phenyl transfer from PhB(OH)2 (cf 4) to [Pd(OH)-
(PPh3)(Ph)], or from [PhB(OH)3]ꢁ (cf 5ꢁ) to [Pd(PPh3)(Ph)]+;
without significant contribution from 1, 7, or 8.
In parallel with F/OH exchange at boron, fluoride
sequestration by the base and the glassware[15] progressively
drove the equilibrium towards the boronic acid 4 (> 98%).
No intermediate species were detected by 11B or 19F NMR
This conclusion was tested by performing the SM coupling
reactions of mixtures of [D0]-1 and [D4]-4, in initial ratios
90:10, 70:30, 50:50, 30:70, and 10:90 (Figure 3a).[16] Catalytic
turnover (2!3) proceeded in parallel with the equilibration
Angew. Chem. Int. Ed. 2010, 49, 5156 –5160
ꢀ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.angewandte.org
5157

Communications
and its stoichiometry is in direct proportion to the catalyst
loading.[16] However, in reactions employing trifluoroborate 1,
which generates 4 in situ, there is no trace of 9 (< 0.1%), even
with high catalyst loadings (5 mol%). The quantitative
reduction of [PdCl2(PPh3)2] must therefore take place
before any significant build up[17] of 4 has occurred. Endog-
enous[18] fluoride can catalyze[19] the reduction[20] of the
precatalyst, [PdCl2(PPh3)2], for example, by hydrolysis of a
fluorophosphonium species generated from a phosphorane[21]
complex (11). 31P NMR spectroscopic analysis of the SM
coupling of 1 in THF/18OH2 (10:1) confirmed that hydrolytic
catalyst activation generates Ph3P 18O (DdP = 0.04 ppm), and,
=
nominally, a monophosphane Pd0 complex.[22]
Figure 3. a) Isotope distribution (% [D4]) in 3 as a function of the
conversion of 2, during coupling with [D0]-1/[D4]-4 in the initial ratios
indicated. b) SM coupling efficiency ([3]/[9+10]) as a function of [2]av.
and [4]av. in reactions conducted in air. Solid lines are solely as a guide
for the eye.[16]
between 1 and 4, and, after complete conversion of 2, the
isotope distribution in the SM coupling product 3 (%[D4]),
reflects the initial [D0]/[D4] ratio of 1:4. However, prior to full
equilibration, the isotope distribution in the evolving SM
coupling product 3 reflects the isotope distribution in the
reactive component. The product that was initially generated
was predominantly [D4]-3, even if the mixture of boron
reagents contained just 10% [D4]-4, thus confirming that
boronic acid 4 is the dominant transmetalating species. These
results reinforce the conclusion that 1 serves as a reservoir for
4, via an unobserved intermediate species, such as 7 and 8,
which themselves do not contribute directly to the SM
coupling.
The question then arises as to why the indirect reaction,
that is, starting from 1 rather than 4, is more efficient. The
significant difference lies not in the rate, but in the substantial
decrease in the amount of side products obtained with
trifluoroborate 1. Indeed, reactions starting directly from
boronic acid 4 proceeded to completion faster than reactions
starting from 1; predominantly because the slow and some-
what variable[15] rate of hydrolysis (1!4) can limit the rate of
catalytic turnover.
For SM coupling in aqueous tetrahydrofuran, the rate of
protodeboronation (5!6) is negligible, and thus only two of
the three side products (6, 9, and 10) differentiate the use of
trifluoroborate reagent 1 from boronic acid 4. Under routine
synthetic conditions, that is, reagents and solvents that have
not been rigorously purified, the SM coupling of 2 with 4 in
aqueous tetrahydrofuran generated 9 and 10 in two stages:
rapid initial production followed by a progressive accumu-
lation during the catalytic turnover. As shown below, these
side products arise from three different processes: one
generates 9, one generates 10, and one generates both. All
three processes are substantially suppressed in reactions
starting from 1, in which 4 is generated in situ.
Phenol 10 is also produced (observed by 19F NMR
spectroscopy) at the beginning of SM coupling reactions in
which 4 is employed directly, and in quantities that depend on
the purity of the tetrahydrofuran. However, phenol 10 is
absent in reactions that employ 1 to generate 4 in situ. A
tetrahydrofuran-derived oxidant[23] is thus consumed by a
non-phenolic pathway, prior to significant hydrolysis of 1.
Control experiments confirmed that at 558C, trifluoroborate
1 is able to mediate the decomposition of aqueous tetrahy-
drofuran solutions of tetrahydrofuran-2-hydroperoxide.[23] In
contrast, boronic acid 4 is efficiently oxidized[24] into phenol
10 under the same conditions. The difference in outcome
between 1 and 4 may arise from a bifurcation in the
conventional oxidation pathway (12, X = OH), in which an
[1,2]-Ar shift leads to 10. The increased Lewis acidity of the
boron center (X = F) would both reduce the migratory
ꢁ
aptitude of the Ar group, and affect the polarity of the O
O moiety, for example, to induce its regioisomeric cleavage
via a [1,2]-H/CH2 shift from the tetrahydrofuran ring,[25a] or an
a-CH elimination.[25b]
In the SM coupling of boronic acid 4, incomplete
degassing of the aqueous tetrahydrofuran solvent prior to
reaction, or an ingress of air during sampling of the reaction,
results in the generation of side products 9 and 10 in a 1:1 ratio
throughout catalytic turnover. This process is almost absent
when 1 is employed, even though the reactions proceed
through 4.[26] The mechanism of the palladium-catalyzed
aerobic oxidative homocoupling[27] of aryl boronic acids has
been shown by Amatore, Jutand, and co-workers[28] to involve
a peroxo intermediate [Pd(k2-O2)]. This species reacts with
two molecules of aryl boronic acid and water to generate a
biaryl species (cf 9) and perboric acid, B(OH)2OOH. The
latter oxidizes a third molecule of aryl boronic acid to the
phenol (10).
In reactions that employ 4 directly, the precatalyst,
[PdCl2(PPh3)2], is reductively activated by a two-stage trans-
metalation/reductive elimination sequence to generate [Pd-
(PPh3)2] and homocoupled product 9. The latter is detected by
19F NMR spectroscopy immediately after turnover begins,
Consideration of a mechanism for SM coupling in
competition with oxidative homocoupling (Scheme 4) sug-
gests that a Pd0 intermediate (13) is partitioned between
oxidative addition of the aryl bromide (2) versus reaction with
oxygen and the boronic acid to form 14.
5158
www.angewandte.org
ꢀ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2010, 49, 5156 –5160

Angewandte
Chemie
concentrations of boronic acid 4 (Figure 3b). Slow hydrol-
ysis of trifluoroborate 1 relative to catalytic turnover
ensures low concentrations of 4.[30]
The above factors account for the exceptional perfor-
mance of the RBF3K reagent 1 under routine synthetic
conditions.[4] This insight will be of use in the scale-up of biaryl
SM coupling reactions. For example, the slow addition of
boronic acid 4 circumvents prior conversion into trifluorobo-
rate 1.[2–4] This procedure avoids corrosion of the reaction
vessel by fluoride, requires just one equivalent of base instead
of three, and still benefits from a hydrolytic PdII precatalyst
activation.[19]
Scheme 4. Schematic mechanism for the SM coupling of ArBr 2 (by
oxidative addition of 2 to Pd0 complex 13 leading to biaryl 3) versus
the oxidative homocoupling of 4 with O2 via 14, which leads to biaryl 9
and phenol 10. [Pd]=cis/trans [Pd(PPh3)n], where n=0,1,2; X=Br, OH.
Conducting the SM coupling of 4 under air, so that [O2] is
approximately constant, confirmed a linear relationship
between [2] and [3]/([9] + [10]), see the open circles in
Figure 3b. An analogous, but inverse relationship was found
for the boronic acid ([3]/([9] + [10]) / [4]ꢁ1), thus suggesting a
reversible,[29] or boronic acid facilitated O2 addition (13!14,
via [Pd(k2-O2)]). The high sensitivity of the oxidative pathway
to the concentration of boronic acid is key to the efficiency of
the indirect process that starts from trifluoroborate 1: slow
hydrolysis[30] of 1 relative to turnover keeps the concentration
of 4 low, thus reducing the partitioning of 13 via the [Pd(k2-
O2)(4)] intermediate 14. This conclusion was tested by
syringe-pump addition (0.2 mm sꢁ1)[16] of a THF/H2O (10:1)
solution of 4 + Cs2CO3 to an SM coupling of 2 that was
conducted under air. The effect was substantial, affording a
[3]/([9] + [10]) ratio of > 9.5, as compared to a ratio of only 3.5
with 1 under the standard conditions (Scheme 1).
Seven key points emerging from this study are outlined
below. These may also be of importance when considering
mechanisms for turnover, for precatalyst activation, and for
oxidative side reactions in other areas of transition-metal
catalysis that use RBF3K reagents under aqueous condi-
tions.[2,3,31]
1) Reagent 1 undergoes hydrolytic equilibrium with boronic
acid 4; an accompanying sequestration of fluoride by base
and glassware[15] resulted in slow but complete conversion
into 4.
2) Low water concentrations reduce the equilibrium popu-
lation of borate 5 (Figure 2b), thus suppressing proto-
deboronation.
3) Mixed intermediates 7 and 8 are not detected (< 60 mm)
and computationally (Figure 2) are found to be less
efficient than 4 at aryl transfer to PdII in the trans-
metalation step.
The extensive optimization of base, solvent, and temper-
ature in SM coupling reactions with other trifluoroborates
and other organohalides and pseudohalides[4,5] may reflect the
requirement to balance the rate of hydrolysis with the rate of
catalytic turnover, so that side reactions, such as protode-
boronation and oxidative homocoupling, are suppressed.[5a]
Our analysis also suggests that partially fluorinated borates
(R-BF3ꢁnXn, X = OH, OR) can participate in the transmeta-
lation event, albeit less efficiently than the boronic acid;
however, most studies show that mixed intermediates
undergo rapid disproportionation to the fully fluorinated
and non-fluorinated species.[5a,12,13]
Received: March 13, 2010
Published online: June 11, 2010
Keywords: boron · fluorides · homogeneous catalysis ·
.
palladium · reaction mechanisms
[1] a) N. Miyaura, A. Suzuki, Chem. Rev. 1995, 95, 2457 – 2483; b) F.
Bellina, A. Carpita, R. Rossi, Synthesis 2004, 2419 – 2440; c) R.
Martin, S. L. Buchwald, Acc. Chem. Res. 2008, 41, 1461 – 1473;
d) H. Doucet, Eur. J. Org. Chem. 2008, 2013 – 2030.
[2] a) E. Vedejs, R. W. Chapman, S. C. Fields, S. Lin, M. R.
Schrimpf, J. Org. Chem. 1995, 60, 3020 – 3027; b) for a review
see: H. A. Stefani, R. Cella, A. S. Vieira, Tetrahedron 2007, 63,
3623 – 3658; c) V. Bagutski, A. Ros, V. K. Aggarwal, Tetrahedron
2009, 65, 9956 – 9960.
[3] The use of RBF3K reagents in catalysis was pioneered by Darses
and Genet; for examples, see: a) S. Darses, J. P. Genet,
Tetrahedron Lett. 1997, 38, 4393 – 4396; b) S. Darses, J. P.
Genet, Eur. J. Org. Chem. 2003, 4313 – 4327; c) S. Darses, J. P.
Genet, Chem. Rev. 2008, 108, 288 – 325.
[4] For an account, see: a) G. A. Molander, N. Ellis, Acc. Chem. Res.
2007, 40, 275 – 286; for a review, see: b) G. A. Molander, B.
Canturk, Angew. Chem. 2009, 121, 9404 – 9425; Angew. Chem.
Int. Ed. 2009, 48, 9240 – 9261.
[5] For the use of trifluoroborates in biaryl synthesis, see: a) G. A.
Molander, B. Biolatto, J. Org. Chem. 2003, 68, 4302 – 4314; in
toluene/water: b) G. A. Molander, F. Vargas, Org. Lett. 2007, 9,
203 – 206; c) G. A. Molander, J. Hama, D. G. Seapy, Tetrahedron
2007, 63, 768 – 775; d) G. A. Molander, L. Jean-Gerard, J. Org.
Chem. 2009, 74, 1297 – 1303; e) G. A. Molander, D. E. Petrillo,
Org. Lett. 2008, 10, 1795 – 1798; in THF/water: f) G. A.
Molander, L. A. Felix, J. Org. Chem. 2005, 70, 3950 – 3956;
g) G. A. Molander, D. E. Petrillo, N. R. Landzberg, J. C.
Rohanna, B. Biolatto, Synlett 2005, 1763 – 1766; h) G. A.
Molander, P. E. Gormisky, D. L. Sandrock, J. Org. Chem. 2008,
73, 2052 – 2057; i) G. A. Molander, A. R. Brown, J. Org. Chem.
4) The majority of catalytic turnover proceeds via 4 (Fig-
ure 3a).
5) The use of 1 results in fluoride-catalyzed[19] hydrolytic
reduction[20] of the precatalyst, possibly via 11, bypassing
the generation of 9.
6) Trifluoroborate 1 mediates the decomposition of traces of
tetrahydrofuran hydroperoxide in the aqueous tetrahy-
drofuran, possibly via a Lewis acid induced bifurcation in
12, thus avoiding generation of the phenol 10.
7) The competing Pd0-catalyzed aerobic homocoupling via
14, which consumes three molecules of boronic acid 4 per
cycle, is suppressed by high concentrations of 2 and low
Angew. Chem. Int. Ed. 2010, 49, 5156 –5160
ꢀ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.angewandte.org
5159

Communications
2006, 71, 9681 – 9686; j) G. A. Molander, B. Canturk, L. E.
Kennedy, J. Org. Chem. 2009, 74, 973 – 980.
[6] H. G. Kuivila, J. F. Reuwer, Jr., J. A. Mangravite, Can. J. Chem.
1963, 41, 3081 – 3090.
[19] If fluoride sources are absent and 4 is in low concentration
(<0.4 mm), aqueous basic tetrahydrofuran effects slow hydro-
lytic precatalyst activation, possibly catalyzed by the glass
surface, thus bypassing generation of 9.
[7] 19F NMR analysis of reactions with [1]0 = 47 mm (total volume)
showed triarylboroxine (generated in situ), 4, and 2 in the
toluene phase, and 5, 3KF, and 6 in the aqueous phase.
Metastable dispersions of 6 in the aqueous phase migrate
slowly into the toluene phase. In more dilute reactions ([1]0 =
8 mm), all species except KF were located in the toluene phase.
19F NMR analysis of a reaction conducted without toluene (just
water as solvent) yielded < 16% of SM coupled product (3) over
a 6 hour period.
[20] Verkade and co-workers have detailed the fluoride-mediated
and -catalyzed reduction of palladium halides by phosphanes:
a) M. R. Mason, J. G. Verkade, Organometallics 1990, 9, 864 –
865; b) M. R. Mason, J. G. Verkade, Organometallics 1992, 11,
2212 – 2220; c) P. A. McLaughlin, J. G. Verkade, Organometallics
1998, 17, 5937 – 5940.
[21] S. A. Macgregor, D. C. Roe, W. J. Marshall, K. M. Bloch, V. I.
Bakhmutov, V. V. Grushin, J. Am. Chem. Soc. 2005, 127, 15304 –
15321.
[8] a) S. W. Wright, D. L. Hageman, L. D. McClure, J. Org. Chem.
1994, 59, 6095 – 6097; b) G. A. Molander, D. L. Sandrock, Org.
Lett. 2009, 11, 2369 – 2372.
[9] a) G. A. Molander, D. L. Sandrock, J. Am. Chem. Soc. 2008, 130,
15792 – 15793; b) G. C. Fu, A. F. Littke, C. Dai, J. Am. Chem.
Soc. 2000, 122, 4020 – 4028.
[22] The generation of nominally different catalysts from the reaction
of [PdCl2(PPh3)2] with 1 versus 4, did not appear to be the origin
of the better performance of 1. Both systems are phosphane-
deficient (31P{1H} NMR analysis with internal standard, Tol3P =
O) and gave positive tests for heterogeneous catalysis using the
mercury droplet method (C. Pad, W. Hartman, Ber. Dtsch.
Chem. Ges. 1918, 51, 711). Traces of oxidant, such as
B(OH)2OOH, will presumably oxidize the free phosphane
ligand. A sample of Ph3P18O (0.12 mm, 32% 18O) was unchanged
after heating at 558C overnight in 10:1 THF/H2[18O0.7] and
Cs2CO3 (24 mm).
[23] THF undergoes aerobic oxidation, particularly when free of
stabilizer, for example, after distillation, and exposed to light or a
metal catalyst; a) A. Robertson, Nature 1948, 162, 153; b) G. I.
Nikishin, V. G. Glukhovtsev, M. A. Peikova, A. V. Ignatenko,
Izv. Akad. Nauk SSSR Ser. Khim. 1971, 10, 2323 – 2325; Control
experiments confirmed that exposure of aq THF (10:1) to air led
to ꢂ 8.0 mm hydroperoxide which rapidly oxidizes 4 into 10 at
558C. Neither base (e.g. Cs2CO3) nor water is required, and
neither KF nor TBAF decomposed the peroxide. Use of [D]0-1/
[D]4-4 mixtures confirmed that 4, but not 1, is oxidized into 10.
[24] a) H. G. Kuivila, J. Am. Chem. Soc. 1954, 76, 870 – 874; b) H. G.
Kuivila, A. G. Armour, J. Am. Chem. Soc. 1957, 79, 5659 – 5662.
[25] a) W. Ockels, J. Stein, H. Budzikiewicz, Z. Naturforsch. B 1984,
39, 90 – 94; b) S. J. Blanksby, G. B. Ellison, V. M. Bierbaum, S.
Kato, J. Am. Chem. Soc. 2002, 124, 3196 – 3197.
[10] Alcohols are also used as a reaction medium,[5a] but need to be
slightly wet (laboratory grade) to be effective. See: T. E. Barder
S. L. Buchwald, Org. Lett. 2004, 6, 2649 – 2652. Analogous effects
were found for the SM coupling of 1 with 2: in laboratory-grade
MeOH, the reaction went to completion in ꢀ 30 minutes;
whereas in anhydrous MeOH, the coupling proceeded to 80%
completion in 9 hours at 808C, wherein protonation/decarbox-
ylation of the carbonate base potentially provides a source of
water.
ꢁ
[11] ArN2+BF4 (J.-P. GenÞt, S. Darses, J.-L. Brayer, J.-P. Demoute,
Tetrahedron Lett. 1997, 38, 4393 – 4396) and Ar2I+BF4 (C.-Z.
ꢁ
Chen, M. Xia, Synth. Commun. 1999, 29, 2457 – 2465) salts
undergo SM coupling with ArBF3K reagents under base-free
ꢁ
conditions, probably through competition between ArBF3 and
ꢁ
BF4 as a complexed fluoride ligand for coordination to the
palladium cation (see Figure 2).
[12] Vedejs et al. noted that aqueous solutions of PhBF3K “are
somewhat acidic, suggesting the existence of an equilibrium
between the tetracoordinated ’ate’ species and products of
hydrolytic cleavage or ligand exchange.” (see Ref. 2a).
[13] a) R. Ting, C. W. Harwig, J. Lo, Y. Li, M. J. Adam, T. J. Ruth,
D. M. Perrin, J. Org. Chem. 2008, 73, 4662 – 4670; b) R. A. Batey,
T. D. Quach, Tetrahedron Lett. 2001, 42, 9099 – 9103; c) C. A.
Hutton, A. K. L. Yuen, Tetrahedron Lett. 2005, 46, 7899 – 7903.
[14] During the course of this work, the use of wet silica gel to
convert RBF3K into RB(OH)2 was reported: G. A. Molander,
L. N. Cavalcanti, B. Canturk, P.-S. Pan, L. E. Kennedy, J. Org.
Chem. 2009, 74, 7364 – 7369.
[26] Control experiments eliminated modification of the glass surface
by 1, the presence of phenol (10) as a ligand for palladium, or
lower base concentrations as being responsible for this differ-
ence.
[27] a) M. Moreno-Maꢀas, M. Pꢁrez, R. Pleixats, J. Org. Chem. 1996,
61, 2346 – 2351; b) M. A. Aramendꢂa, F. Lafont, M. Moreno-
Maꢀas, R. Pleixats, A. Roglans, J. Org. Chem. 1999, 64, 3592 –
3594.
[15] The rate of hydrolysis is unaffected by [2] or [Pd], but increases
with the glass-surface/reaction-volume ratio. Samples hydrolyze
efficiently in the absence of base in a glass vessel, but not at all if
conducted in a teflon vessel. In glass vessels, the presence of base
caused variable induction periods before the onset of hydrolysis,
after which the rates are identical to runs without base present.
Solution-phase fluoride (in aq THF) could not be detected by
19F NMR spectroscopy, owing to precipitation (as MF; M = K,
Cs) or sequestration by the glass surface.[14]
[28] C. Adamo, C. Amatore, I. Ciofini, A. Jutand, H. Lakmini, J. Am.
Chem. Soc. 2006, 128, 6829 – 6836.
[29] Reversible O2 binding might arise from phosphane defi-
ciency.[21,28]
[30] a) “MIDA” boronates are employed as slow release agents:
D. M. Knapp, E. P. Gillis, M. D. Burke, J. Am. Chem. Soc. 2009,
131, 6961 – 6963; b) for a recent comparison of the use of
“MIDA” boronates with organotrifluoroborates, see: K. Brak,
J. A. Ellman, J. Org. Chem. 2010, 75, 3147 – 3150.
[16] For full details, see the Supporting Information.
[17] The addition of up to 50% of 4 to 1 was required to regenerate 9
by transmetalation in the precatalyst activation.
[18] SM coupling of 4 + TBAF (3 equiv) completely suppressed the
generation of 9 through activation of a precatalyst. Significant
quantities of 1 were generated, together with 8 (7% Ar), BF4K
(16% B), and B(OH)F3K (27% B). Fluoride (exogenous or
endogenous) did not inhibit the aerobic oxidative catalytic
generation of 9 and 10. See also: S. Punna, D. D. Diaz, M. G.
Finn, Synlett 2004, 2351 – 2354.
[31] For examples of copper-catalyzed ether formation, see: a) R. A.
Batey, T. D. Quach, Org. Lett. 2003, 5, 1381 – 1384; rhodium-
catalyzed 1,2-additions: b) A. Ros, V. K. Aggarwal, Angew.
Chem. 2009, 121, 6407 – 6410; Angew. Chem. Int. Ed. 2009, 48,
6289 – 6292; rhodium-catalyzed 1,4-additions: c) L. Navarre, M.
Pucheault, S. Darses, J.-P. Genet, Tetrahedron Lett. 2005, 46,
4247 – 4250; d) C. G. Frost, S. D. Penrose, R. Gleave, Org.
Biomol. Chem. 2008, 6, 4340 – 4347.
5160
www.angewandte.org
ꢀ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2010, 49, 5156 –5160