Reactivity and stability of dinuclear Pd(I) complexes: Studies on the active catalytic species, insights into precatalyst activation and deactivation, and application in highly selective cross-coupling reactions

Article abstract of DOI:10.1021/ja209424z

The catalysis derived from the dinuclear Pd(I)-Pd(I) complex, {[PtBu ₃]PdBr}₂, has been studied with experimental, computational, and spectroscopic techniques. Experimental selectivity studies were performed, and the reactivity was s

Full text of DOI:10.1021/ja209424z

Article
pubs.acs.org/JACS
Reactivity and Stability of Dinuclear Pd(I) Complexes: Studies on the
Active Catalytic Species, Insights into Precatalyst Activation and
Deactivation, and Application in Highly Selective Cross-Coupling
Reactions
Fabien Proutiere, Marialuisa Aufiero, and Franziska Schoenebeck*
Laboratory for Organic Chemistry, ETH Zurich, Wolfgang-Pauli-Strasse 10, 8093 Zurich, Switzerland
̈
̈
S
* Supporting Information
ABSTRACT: The catalysis derived from the dinuclear Pd(I)−Pd(I) complex, {[PtBu3]-
PdBr}2, has been studied with experimental, computational, and spectroscopic techniques.
Experimental selectivity studies were performed, and the reactivity was subsequently
investigated with density functional theory (B3LYP-D and M06L) to deduce information on
the likely active catalytic species. The reactivity with aryl chlorides and bromides was found
to be inconsistent with direct catalytic involvement of the Pd(I) dimer but consistent with
mononuclear Pd(0) catalysis. Computational studies suggest that precatalyst transformation
to the active catalytic species does not proceed via a direct disproportionation mechanism; a
reductive pathway is the most likely scenario instead. Through 31P NMR investigations it was
identified that the combination of ArB(OH)2, KF, and water triggers the conversion of the
precatalyst to Pd(PtBu3)2 and, most likely, Pd-black as a competing side process, explaining
the incomplete conversions of aryl chlorides in Suzuki cross-coupling reactions under Pd(I) dimer conditions. New applications
in highly regio- and chemoselective transformations in short reaction times at room temperature are also demonstrated.
This report will focus on the detailed study of the catalysis
derived from the Pd(I) dimer, {[PtBu3]PdBr}2. It has been
suggested that the high reactivity of this Pd(I) dimer might be
due to the formation of the monoligated 12-electron species,
Pd(0)PtBu3, in solution (see Figure 1).3 This could occur via
INTRODUCTION
■
A number of dinuclear Pd(I)−Pd(I) complexes have been
synthesized over the past 70 years, and new reactivities at such
multiple palladium sites have been achieved.1 The di-tert-
butylphosphane-ligated palladium(I) dimer, {[PtBu3]PdBr}2,
for example, was first synthesized by Mingos, Vilar, and co-
workers and was shown to be reactive with a number of small
molecules.2 The catalytic potential was explored by Hartwig
and co-workers3 and Prashad et al.4,5 The Pd(I) dimer
precatalyst gave rise to very high reaction rates, leading to
Pd-catalyzed transformations of aryl bromides in as few as 15
min at room temperature in Suzuki and amination reactions.3,6
With unactivated aryl chlorides, Suzuki cross-coupling reactions
were less efficient, and even prolonged reaction times did not
give rise to full conversion.3 Syntheses of structural alternatives
of dinuclear Pd(I)−Pd(I) complexes have been reported,2 and
their formation as side products from mononuclear Pd(II)
intermediates has been discussed.7 Related to that, the
formation of a dinuclear Pd(I) dimer from a Pd(II)−NHC
complex under catalytically relevant conditions has recently
been reported also.8 Further remarkable activity of dinuclear Pd
species in different oxidation states, i.e., Pd(III)−Pd(III), has
been documented for C−H insertion,9 and even Pd(I)−Pd(II)
dimers have recently been isolated in C−H halogenations.10
Despite these advances, there is still relatively little under-
standing of the precise role and mode of action of such multiple
Pd sites in catalysis. Nevertheless, the development of Pd−Pd
bonds into promising larger sized clusters, i.e., nanoparticles,
has progressed significantly in recent years.11
Figure 1. Possible mechanisms of formation of the active species from
Pd(I) dimer, or direct involvement in catalysis.
disproportionation of the Pd(I) dimer into Pd(0)PtBu3 and
Pd(II)(PtBu3)Br2 species,2b,3a,5b,6 or by reduction of the Pd(I)
dimer to Pd(0)PtBu3 (see Figure 1).3a,c Even anionic
[Pd(0)BrPtBu3]¯ might be formed.3a,12 Homolytic scission
into Pd(I) radicals, potentially followed by reduction, might
also be a possibility.13 However, other reports favor a direct
catalytic involvement of Pd(I) dimers of related structure in
catalysis and suggest a catalytic cycle in which the Pd−Pd bond
stays intact.14
Received: October 6, 2011
Published: December 2, 2011
© 2011 American Chemical Society
606
dx.doi.org/10.1021/ja209424z | J. Am. Chem.Soc. 2012, 134, 606−612

Journal of the American Chemical Society
Article
We herein report our combined experimental and computa-
tional efforts to gain detailed mechanistic insights into the
catalysis derived from Pd(I) dimers, focusing on {[PtBu3]-
PdBr}2 to address the questions of (i) the likely active catalytic
species and (ii) precatalyst activation and deactivation. We also
(iii) provide an explanation why aryl chlorides react less
efficiently under Pd(I) dimer conditions in Suzuki coupling
reactions and (iv) present novel applications in highly
regioselective cross-coupling reactions in rapid reaction times
at room temperature.
RESULTS AND DISCUSSION
■
We have previously shown the usefulness of selectivity studies
to gain indirect information on the likely active catalytic species
of a given transformation. For this, we had performed a
combined computational and experimental study to elucidate
the origin of selectivity reversal of a Pd-catalyzed cross-coupling
reaction in different solvents (see Figure 2).15 Fu and co-
Figure 3. Previous computational investigations on the origin of
selectivity reversal in polar solvent: (i) oxidative addition TSs in
MeCN (top left), (ii) solvent coordination in TS (top right), and (iii)
TSs for oxidative addition by anionic PdLX¯ as active species (with X
= F or ArB(OH)O).15
We also provided experimental support of these conclusions;
i.e., we performed cross-coupling reactions in the absence of
coordinating additives in polar solvent to disfavor an anionic Pd
species (Stille reaction, Figure 2). This gave rise to
predominant C−Cl insertion, consistent with Pd(0)PtBu3 as
the active species (Figure 2).15
Experimental Selectivity Studies and Reaction Rates
under Pd(I) Dimer, {[PtBu3]PdBr}2, Conditions. The
extensive studies performed on the selectivity of substrate 1
led to a reactivity picture that allows us to use it as a
mechanistic probe for our current study.15,16 If, as assumed,
{[PtBu3]PdBr}2 indeed formed [Pd(0)PtBu3] in solution, we
would expect C−Cl insertion to be favored in reactions of the
dimer with substrate 1. If, on the other hand, anionic
[Pd(0)(X)PtBu3]¯ (with X = Br, F, ArB(OH)O) were to be
formed as active species, we would expect triflate insertion (see
Figure 2).15 Thus, we reacted 1 with 1.5 mol % {[PtBu3]PdBr}2
in the presence of KF and ArB(OH)2 in MeCN and THF and
followed the reactions by calibrated GC-MS analysis.17 In both
reactions, only a single isomer was detected. Selectivity for C−
Cl insertion was seen in THF, whereas in MeCN, exclusive
addition to C−OTf took place. As such, the identical selectivity
is observed under Pd(I) dimer conditions as with the
Pd2(dba)3/PtBu3 catalytic system.
Figure 4 gives the conversions of the reactions over time. In
both MeCN and THF, the reactions had reached high
conversion after 30 min (78% in MeCN and 76% in THF)
employing the Pd(I) dimer. Although very similar conversions
are reached after 30 min, the initial reaction rate is lower in
MeCN, which could be explained by the lower solubility of the
Pd(I) dimer in MeCN compared to THF. The analogous
reactions involving Pd2(dba)3/PtBu3 in comparison (Figure 5)
were much slower and led to only 3−5% conversion after 30
min.18
Figure 2. Previous chemoselectivity investigations involving
Pd2(dba)3/PtBu3 in Suzuki and Stille cross-coupling reactions with 1
(the formation of the predominant products is illustrated).15,16
workers had established that the reaction of Pd2(dba)3/PtBu3
with 4-chlorophenyl triflate (1) in THF gives rise to selective
C−Cl insertion, consistent with PdPtBu3 as active catalytic
species.16 We discovered that employing the identical
conditions in polar solvents, such as MeCN, gives rise to a
complete selectivity reversal with exclusive transformation of
the C−OTf bond.15 We also demonstrated this selectivity
reversal for intermolecular competition experiments and
established that it is general in arylboronic acid.15 Through
detailed computational studies, we were able to show that
neither electrostatic stabilization nor coordination by the polar
solvent accounts for the selectivity reversal (see Figure 3).
Instead, a different active species is likely the cause of selectivity
reversal in polar solvent: the reactivity of Pd2(dba)3/PtBu3 in
polar solvents was found to be consistent with anionic
[Pd(0)PtBu3X]¯ as active catalytic species if coordinating
additives or cross-coupling partners are present in the mixture,
leading to preferential C−OTf addition. In nonpolar solvent,
the reactivity is consistent with neutral [Pd(0)PtBu3], giving
C−Cl addition.
607
dx.doi.org/10.1021/ja209424z | J. Am. Chem.Soc. 2012, 134, 606−612

Journal of the American Chemical Society
Article
complex to the active catalytic species therefore needs to take
place prior to reaction.
Could a Pd(I) Radical Be Reactive? Such a species could
result from homolytic dimer dissociation (see Figure 1). The
bond dissociation energy of a related Pd(I)−Pd(I) dimer was
previously determined to be between 21 and 29 kcal/mol.21
Using dispersion-corrected unrestricted density functional
theory (DFT-D, see Figure 6),22 we calculate an energy
Figure 4. Reactivity of 1 with Pd(I) dimer over time. The reaction was
monitored with calibrated GC-MS analysis.
Figure 6. Free energy reaction profile for the oxidative addition of
[Pd(I)Br(PtBu3)] to 1 [at UB3LYP-D/6-31+G(d) and SDD (for
Pd)].29
penalty of 26.1 kcal/mol in the gas phase for the homolytic
dimer scission (see Figure 6).23,24 Calculation at the M06L
level of theory36to use an alternative methodand
application of solvation correction for THF (using a CPCM
model) gives 24.8 kcal/mol as dissociation energy.25 Thus,
homolytic dimer dissociation might potentially be a feasible
process.
However, would the resulting Pd(I) radical species also be
reactive? To investigate this, we calculated the transition states
for oxidative addition by the Pd(I) radical, [Pd(I)Br(PtBu3)],
using DFT-D.26,27 The barriers for oxidative addition by Pd(I)
are calculated to be rather high: an overall 14 kcal/mol higher
energy barrier was calculated for C−Cl insertion compared to
the analogous reaction relative to Pd(PtBu3)2, making this
mechanistic possibility therefore unlikely (see Figure 4).
Moreover, the Pd(I) radical was calculated to favorably insert
into C−Cl in the gas phase, in THF, and in MeCN,28 and
would thus be inconsistent with the opposing selectivities in
different media.
Figure 5. Reactivity of Pd2(dba)3/PtBu3 with 1. The reaction was
monitored with calibrated GC-MS analysis.
Using an alternative base, KOH, instead of KF in the Suzuki
coupling reaction of 1 under Pd(I) dimer conditions gave very
similar results: we obtained 80% yield of 2 in THF and 82% of
3 in MeCN after 30 min reaction time.
At first sight, the selectivities observed in the reactions of the
Pd(I) dimer with 1 appear to be consistent with [Pd(0)PtBu3]
as active catalytic species in nonpolar solvents and [Pd(0)(X)-
PtBu3]¯ (with X = Br or F or ArB(OH)O or OH) in polar
solvents.19 As such, the observed selectivity would be consistent
with “standard” Pd(0) catalysis. However, an alternative
scenario could be the direct involvement of the Pd(I) dimer
in catalysiswould this be consistent with the observed
reactivity also?
Is Bimetallic Catalysis a Likely Scenario? Recent reports
suggested the direct involvement of structurally related
dinuclear Pd(I) species in Suzuki cross-coupling reactions
with aryl bromides.14 We therefore set out to explore whether
{[PtBu3]PdBr}2 could possibly be the reactive species also in
our case, rather than a Pd(0) species derived from it. To test
this, we mixed the Pd(I) dimer with 4-chlorophenyl triflate (1)
in THF in a glovebox and examined the mixture by 31P NMR
spectroscopy. Over the course of 30 min, the only peak
observed was that corresponding to the dimer itself (87 ppm).6
We did not detect a species resulting from oxidative addition to
1.
To test an additional, significantly easier target for oxidative
addition,20 we repeated the study with bromobenzene in place
of 1. Using bromobenzene as the solvent at room temperature,
however, also did not give rise to reaction with the Pd(I) dimer.
This suggests that the Pd(I) dimer does not react directly with
aryl bromides or chlorides. Activation of the dinuclear Pd(I)
The reactivity derived from the dinculear Pd(I) complex,
{[PtBu3]PdBr}2, is thus inconsistent with dincuclear and
mononuclear Pd(I) catalysis. Instead, the reactivity is consistent
with “standard” mononuclear Pd(0) catalysis (giving neutral
Pd(0) in THF) or anionic Pd(0) in MeCN as active catalytic
species) in Suzuki cross-coupling reactions, completely
analogous to the catalytic system involving Pd2(dba)3/PtBu3.15
Further Applications of Pd(I) Dimer in Regioselective
Cross-Coupling Reactions of Dihalogenated Hetero-
cycles. In light of the excellent chemoselectivity achieved
with substrate 1, we were intrigued to further explore the scope
of the Pd(I) dimer in regioselective cross-coupling reactions. In
particular, functionalization of polyhalogenated heterocycles via
regioselective and iterative cross-coupling reactions is a
straightforward and attractive route to synthesize important
building blocks for pharmaceutical and materials applica-
tions.30,31 To the best of our knowledge, Pd(I)−Pd(I)
precatalysts have not been explored for their potential in
608
dx.doi.org/10.1021/ja209424z | J. Am. Chem.Soc. 2012, 134, 606−612

Journal of the American Chemical Society
Article
regioselective cross-coupling reactions. Thus, we selected N-
and S-containing five- and six-membered heterocycles (see
dichloropyrimidine, the predicted ΔΔG⧧ is smallest (1 kcal/
mol), which is in line with the observed formation of the
second, minor regioisomeric product. Thus, the computation-
ally predicted selectivity is in line with the experimental
selectivities. This further supports Pd(0)PtBu3 as the active
catalytic species under the conditions applied (i.e., nonpolar
solvent THF), consistent with our findings in Figure 4.34
Moreover, these results reinforce that high selectivities can be
achieved without having to sacrifice reaction rates.
Scheme 1. Selective Couplings of Dihalogenated
Heterocycles
Studies on the Precatalyst Activation Mechanism
Computational Studies. In standard Pd(0) catalysis, high
reactivities can be achieved through the use of bulky ligands in
combination with a Pd(0) precatalyst (e.g., Pd2(dba)3).35
Sterically demanding ligands, such as PtBu3, usually give rise to
PdL2 as catalyst resting state (L = phosphine ligand) after
displacement of dba. The displacement of dba is rather rapid in
THF (e.g., for the reaction shown in Figure 5, we already
observed the peak corresponding to PdL2 as the exclusive P
signal in 31P NMR when we examined the mixture after 10 min
reaction time).36 However, overall such catalytic systems are
presumed to favor a monoligated Pd as the active catalytic
species (in nonpolar solvent, see ref 15 and Figures 2 and 3
above). The PdL1 species in turn is believed to be the origin of
enhanced reactivity.3c Although no such monoligated Pd
species has ever been spectroscopically indentified nor detected
otherwise, experimental and computational data support ligand
dissociation to Pd(0)L1 and L prior to oxidative addition.22b
This initial ligand dissociation step (PdL2 → PdL + L) is
frequently considered to be a reactivity-limiting step and may
occur via direct ligand dissociation or substrate-assisted ligand
displacement.22c,3c Recent computational studies suggest an
energy barrier of ca. 30 kcal/mol for an associated displacement
of PtBu3 from PdL2 (i.e., substitution of one ligand by the
substrate takes place prior to oxidative addition).22b,c
Table 1. Results of Regioselective Couplings (See Scheme 1)
and Computed ΔΔG⧧ Preference (Calculated with B3LYP-
D)29 for the Illustrated Regioisomer (in kcal/mol)
As the reaction rates under Pd(I) dimer precatalyst
conditions are much higher than those under Pd(0) precatalyst
conditions (see Figures 4 and 5), it has been suggested that
precatalyst activation to form the putative active Pd(0) species
might be a very facile process.2,3,5 Thus, the precatalyst
activation should have an activation barrier significantly smaller
than 30 kcal/mol.
a
9% of the alternative regioisomeric product formed also.
Considering the possibilities for precatalyst breakdown (e.g.,
dimer dissociation or disproportionation, see Figure 1), we
calculated37 the energetic penalties for the different possibilities
and compared those to the PdL2 → PdL + L dissociation
energy. As a first approximation, we studied the direct
dissociative pathways. Table 2 summarizes the results. We
Table 1) and performed Suzuki cross-coupling reactions with
these substrates (Scheme 1) in THF. Brominated thiophenes
usually give preferential cross-coupling at the 2-position.
However, those couplings are reported to require high reaction
temperature (ca. 80−100 °C) for hours.33 We found that use of
the Pd(I) dimer gives highly selective transformation in 30 min
at room temperature (see entries 1 and 2). We achieved similar
success (82% yield) in the Suzuki coupling of 2,4-dichloropyr-
imidine applying identical conditions (see entry 3). The rapid
and completely selective couplings of 2,4-dibromothiazole
(entries 4−6) at room temperature were particularly pleasing,
since it has potential in the synthesis of natural products32 or
relevant drugs. We analyzed the reaction mixtures in all cases by
GC-MS prior to workup and did not detect the alternative
regioisomers for entries 1, 2, and 4−6 (Table 1). For
dichloropyrimidine (entry 3), we observed the second isomer
in 9% yield.
Table 2. Calculated Dissociation Energies for Potential
Dimer Dissociations25,36 (Compare with Figure 1; L =
PtBu3)
a
dissociation
ΔGdiss (kcal/mol)
Pd(0)L2 → Pd(0)L+L
{Pd(I)BrL}2 → 2 Pd(I)LBr
{Pd(I)BrL}2 → Pd(II)Br2L + Pd(0)
{Pd(I)BrL}2 → Pd(II)Br2 + Pd(0)L2
{Pd(I)BrL}2 → LPd2Br2 + L
27.2
24.8
38.1
65.4
36.2
L
a
M06L(THF)//B3LYP/6-31+G(d) with SDD (for Pd).
Calculations of the transition states for oxidative addition to
C−halogen by Pd(0)PtBu3 predict a clear preference (ΔΔG⧧ =
3.7 −4.5 kcal/mol) for the 2-positions for dibromothiophene
and -thiazole (see with Table 1). For oxidative addition to 2,4-
employed the M06L functional that has been shown to be
appropriate for the treatment of metal−ligand binding
interaction,36 and used a CPCM solvation model for THF.
609
dx.doi.org/10.1021/ja209424z | J. Am. Chem.Soc. 2012, 134, 606−612

Journal of the American Chemical Society
Article
For the “benchmark” dissociation of PdL2 → PdL + L, we
calculated an energetic penalty of 27.2 kcal/mol (entry 1),
which is in the range of the previously calculated 30 kcal/mol
for the alternative substrate-assisted displacement mechanism
(see above).
Pd(I) dimer (87 ppm)6 after 5 min. We identified this peak as
Pd(PtBu3)2, verified through independent mixing of Pd-
(PtBu3)2, 2-methylphenylboronic acid, water, KF, and internal
standard in THF. After 20 min reaction time, Pd(PtBu3)2 is the
only phosphine-containing species present in the mixture.
Pd(PtBu3)2 in turn is the resting state observed in catalysis
For the Pd(I) dimer precatalyst activation, disproportiona-
tion has frequently been assumed as the likely mechanism to
form Pd(0)L.2,3,5 However, the calculations suggest that
disproportionation of Pd(I) dimer to PdBr2PtBu3 and PdPtBu3
is too high an energy process (38.1 kcal/mol) and therefore
disfavored (see entry 3, Table 2). Thus, the generally assumed
facile liberation of highly reactive Pd(0)PtBu3 does not seem to
occur in a direct disproportionate manner from the precatalyst.
As discussed previously, homolytic dissociation into the Pd(I)
radical seems to be feasible (24.8 kcal/mol), and this pathway
for precatalyst transformation is of lowest energy, at least
among the possibilities that we calculated in Table 2. However,
the involvement of Pd(I) radicals in the catalytic cycle was
already excluded above (Figure 6). If the precatalyst activation
were to occur via homolytic cleavage, reduction of the Pd(I)
radical to Pd(0) would need to subsequently take place.
Thus, the most likely scenario for precatalyst activation is via
a reductive mechanism.38
16
derived from Pd2(dba)3/PtBu3. Under those conditions (and
when 1:1 Pd:L ratios are employed), a second, phosphine-free
Pd(0) species is presumed to be present in the mixture also.
Does the observation of Pd(PtBu3)2 in the 31P NMR
spectrum mean that the Pd(I) dimer reactivity in fact involves
the intermediacy of Pd(PtBu3)2? To examine this, we
performed further spectroscopic studies and carried out a
stepwise addition sequence of the reagents. We separately
mixed the Pd(I) dimer, water,39 arylboronic acid, and KF along
with the internal standard in THF. This led to partial
conversion of dimer to Pd(PtBu3)2 (peak at 85.5 ppm,16 see
Figure 8). Through examination of different addition
Stability of the Pd(I) DimerSpectroscopic Studies.
To get additional insights into the transformation and stability
of the dinuclear Pd(I) complex, we performed 31P NMR
investigations on the reaction involving the Pd(I) dimer
precatalyst with 1 under Suzuki cross-coupling conditions.
We mixed 2-methylphenylboronic acid, substrate 1, KF, the
Pd(I) dimer (under catalytic conditions with 2.5 mol % catalyst
loading), and THF in a glovebox (t = 0 s) and subsequently
followed the reaction by 31P NMR spectroscopy versus an
internal standard (Me3PO4). Figure 7 shows the 31P NMR
Figure 8. 31P NMR study versus internal standard (Me3PO4).
sequences,40 we found that only the combination of water,
arylboronic acid, and KF gives rise to conversion of the Pd(I)−
Pd(I) precatalyst to Pd(PtBu3)2. During this process of Pd(I)
dimer transformation to Pd(PtBu3)2, we also noted a black
precipitation from solution, which we presume to be Pd-black.
As such, the conversion to Pd(PtBu3)2 is independent of the
substrate and the catalytic cycle and constitutes either an
activating or deactivating process of the precatalyst.
We next set out to test whether the transformation of Pd(I)
dimer to Pd(PtBu3)2 is in fact crucial for catalysis. To
investigate this, we fully transformed the Pd(I) dimer in situ
using 2-methylphenylboronic acid, KF, and H2O in THF (see
Scheme 2). Once full conversion to Pd(PtBu3)2 had taken place
(as judged by 31P NMR), 1 was added. Thirty minutes later, we
examined the reaction mixture by calibrated 1H NMR
spectroscopic analysis, which indicated that product 3 had
formed in only 5% yield. Using commercially available
Pd(PtBu3)2 in a separate experiment gave rise to the same
conversion with substrate 1 (see Scheme 2). For comparison,
the reaction with Pd(I) dimer with only 1.5 mol % catalyst
loading gave 76% yield after the identical reaction time (see
Figure 4). The conversion from Pd(I) dimer to Pd(PtBu3)2
(and Pd-black) must thus be a deactivation pathway of the
Figure 7. 31P NMR study versus internal standard (Me3PO4).
spectra after (i) 5 and (ii) 20 min reaction time, which is the
point when the conversion reached a plateau at ca. 80% and did
not reach full conversion thereafter. Interestingly, we observed
a second peak in the 31P NMR at 85.5 ppm along with the
610
dx.doi.org/10.1021/ja209424z | J. Am. Chem.Soc. 2012, 134, 606−612

Journal of the American Chemical Society
Article
a
This material is available free of charge via the Internet at
http://pubs.acs.org.
Scheme 2. Reactivity of 1 with Pd(PtBu3)2
AUTHOR INFORMATION
■
Corresponding Author
Schoenebeck@org.chem.ethz.ch
ACKNOWLEDGMENTS
■
We thank the Swiss National Science Foundation and ETH
Zurich for financial support. Calculations were performed on
the ETH high performance cluster “Brutus”.
̈
REFERENCES
■
(1) (a) First preparation of a Pd(I) dimer: Gelman, A. D.; Meilakh,
E. Dokl. Akad. Nauk SSSR 1942, 36, 188. (b) For a review: Murahashi,
T.; Kurosawa, H. Coord. Chem. Rev. 2002, 231, 207. (c) Werner, H.
Adv. Organomet. Chem. 1981, 19, 155.
(2) (a) Vilar, R.; Mingos, D. M. P.; Cardin, C. J. J. Chem. Soc., Dalton
Trans. 1996, 4313. (b) Dura-Vila, V.; Mingos, D. M. P.; Vilar, R.;
White, A. J. P.; Williams, D. J. J. Organomet. Chem. 2000, 600, 198.
For further studies, see: (c) Christmann, U.; Vilar, R.; White, A. J. P.;
Williams, D. J. Chem. Commun. 2004, 1294. (d) Christmann, U.;
Pantazis, D. A.; Benet-Buchholz, J.; McGrady, J. E.; Maseras, F.; Vilar,
R. J. Am. Chem. Soc. 2006, 128, 6376.
(3) (a) Stambuli, J. P.; Kuwano, R.; Hartwig, J. F. Angew. Chem., Int.
Ed. 2002, 41, 4746. (b) Hama, T.; Liu, X.; Culkin, D. A.; Hartwig, J. F.
J. Am. Chem. Soc. 2003, 125, 11176. (c) Christmann, U.; Vilar, R.
Angew. Chem., Int. Ed. 2005, 44, 366.
a
Conditions: KF (3.0 equiv), ArB(OH)2 (1.01 equiv), H2O (3.0
equiv).
precatalyst. Whether the same process is also responsible for
generating the active Pd(0) species can be neither confirmed
nor excluded. However, this process could explain why
reactions with aryl chlorides do not reach full conversion
under Pd(I) dimer conditions, even with greater catalyst
loadings or prolonged reaction times.3 Wheras Pd(PtBu3)2 is
also a proficient catalyst for aryl bromides (e.g., under
alternative autocatalytic pathways7b), it is an inefficient species
for the transformation of aryl chlorides. The competing Pd(I)
dimer to Pd(PtBu3)2 transformation is thus the likely cause of
subquantitative conversions in Suzuki cross-coupling reactions
of {[PtBu3]PdBr}2 with aryl chlorides.
(4) Prashad, M.; Mak, X. Y.; Liu, Y.; Repic, O. J. Org. Chem. 2003, 68,
1163.
(5) For other transformations using Pd(I) dimers, see: (a) Hama, T.;
Culkin, D. A.; Hartwig, J. F. J. Am. Chem. Soc. 2006, 128, 4976.
(b) Hill, L. L.; Crowell, J. L.; Tutwiler, S. L.; Massie, N. L.; Hines, C.
C.; Griffin, S. T.; Rogers, R. D.; Shaughnessy, K. H. J. Org. Chem.
2010, 75, 6477. (c) Denmark, S. E.; Baird, J. D. Org. Lett. 2006, 8, 793.
(d) Barder, T. E. J. Am. Chem. Soc. 2006, 128, 898. (e) Han, X.; Weng,
Z.; Hor, T. S. A. J. Organomet. Chem. 2007, 692, 5690.
(6) Colacot, T. J. Platinum Metals Rev. 2009, 53, 183.
(7) In efforts to synthesize Pd(II) complexes in the presence of dba,
dimer formation was observed: (a) Strambuli, J. P.; Incarvito, C. D.;
CONCLUSIONS
■
In conclusion, through computational, experimental, and NMR
studies, we have shown that the reactivity derived from the
dinuclear Pd(I)−Pd(I) precatalyst, {[PtBu3]PdBr}2, is con-
sistent with monocuclear Pd(0) catalysis. The Suzuki cross-
coupling reactivity observed in nonpolar solvents is consistent
with Pd(0)L and with anionic Pd(0)LX¯ in polar solvents,
completely analogous to reactivity under Pd2(dba)3/PtBu3
conditions. Computational studies suggest that precatalyst
transformation to the active catalytic species does not proceed
via a direct disproportionation mechanism, and a reductive
pathway is likely instead. Through 31P NMR investigations it
was identified that the combination of ArB(OH)2, KF, and
water triggers the conversion of the precatalyst to Pd(PtBu3)2
and, most likely, Pd-black as a competing side-process,
explaining the subquantitative conversions of unactivated aryl
chlorides in Suzuki cross-coupling reactions. Highly regio- and
chemoselective transformations have also been demonstrated in
short reaction times at room temperature employing Pd(I)
dimer conditions.
Buhl, M.; Hartwig, J. F. J. Am. Chem. Soc. 2004, 126, 1184. See also:
̈
(b) Barrios-Landeros, F.; Carrow, B. P.; Hartwig, J. F. J. Am. Chem. Soc.
2008, 130, 5842.
(8) Boyd, P. D. W.; Edwards, A. J.; Gardiner, M. G.; Ho, C. C.;
Lemee-Cailleau, M.-H.; McGuinness, D. S.; Riapanitra, A.; Steed, J.
́
W.; Stringer, D. N.; Yates, B. F. Angew. Chem., Int. Ed. 2010, 49, 6315.
(9) (a) Powers, D. C.; Ritter, T. Nat. Chem. 2009, 1, 302. (b) Deprez,
N. R.; Sanford, M. S. J. Am. Chem. Soc. 2009, 131, 11234.
(10) Bedford, R. B.; Haddow, M. F.; Mitchell, C. J.; Webster, R. L.
Angew. Chem., Int. Ed. 2011, 50, 5524.
(11) Nanoparticles and Catalysis; Astruc, D., Ed.; Wiley-VCH:
Weinheim, 2007.
(12) Amatore, C.; Jutand, A. Acc. Chem. Res. 2000, 33, 314.
(13) The bond dissociation energy of a related Pd(I)−Pd(I) dimer
was previously determined to be between 21 and 29 kcal/mol. See:
Lee, C. L.; James, B. R.; Nelson, D. A.; Hallen, R. T. Organometallics
1984, 3, 1360.
(14) (a) Das, R. K.; Saha, B.; Rahaman, S. M. W.; Bera, J. K. Chem.
Eur. J. 2010, 16, 14459. (b) Murahashi, T.; Takase, K.; Oka, M.;
Ogoshi, S. J. Am. Chem. Soc. 2011, 133, 14908. For discussion, see
also: (c) Han, X.; Weng, Z.; Hor, T. S. A. J. Organomet. Chem. 2007,
692, 5690. (d) See also ref 8.
ASSOCIATED CONTENT
■
S
* Supporting Information
(15) Proutiere, F.; Schoenebeck, F. Angew. Chem., Int. Ed. 2011, 50,
8192.
(16) Littke, A. F.; Dai, C. Y.; Fu, G. C. J. Am. Chem. Soc. 2000, 122,
Details on experimental procedures, spectroscopic data,
computational information, Cartesian coordinates of all
calculated species and thermal data, and complete ref 26.
4020.
611
dx.doi.org/10.1021/ja209424z | J. Am. Chem.Soc. 2012, 134, 606−612

Journal of the American Chemical Society
Article
(17) The boronic acid used was deoxygenated prior to use. Boronic
acid, 1, KF, and water were dissolved in solvent. The mixture was
stirred for 5 min and the catalyst subsequently added (t = 0).
(18) This is in line with previous comparative kinetic studies of
various precatalysts with the Pd(I) dimer in amination reactions, see:
Hooper, M. W.; Utsunomiya, M.; Hartwig, J. F. J. Org. Chem. 2003, 68,
2861.
(19) We performed the Pd(I) dimer reactions also in DMF and
toluene (under otherwise identical conditions). In DMF, exclusive C−
OTf insertion was seen (61% of 3 was formed, and 39% of 1 was
recovered). In toluene, exclusive C−Cl insertion was detected (2 was
formed in 75% yield and 22% of 1 was recovered).
9:1 selectivity). Reaction of dibromothiazole with 4-methoxyphenyl-
boronic acid gave 91% of the product arising from coupling in the 2-
position as the exclusive product.
(35) (a) Phan, N. T. S.; Van Der Sluys, M.; Jones, C. W. Adv. Synth.
Catal. 2006, 348, 609. (b) See also ref 3c.
(36) In THF, previous studies suggest an equilibration time of 5 min
to form Pd(PtBu3)2. See ref 18.
(37) For appropriateness of M06L to calculate binding energies, see:
(a) Sieffert, N.; Buhl, M. Inorg. Chem. 2009, 48, 4622. (b) Zhao, Y.;
̈
Schultz, N. E.; Truhlar, D. G. J. Chem. Theory Comput. 2006, 2, 364.
(c) Zhao, Y.; Truhlar, D. G. J. Chem. Phys. 2006, 125, 194101.
(d) Zhao, Y.; Truhlar, D. G. J. Phys. Chem. A 2006, 110, 13126.
(38) The mechanism of this reductive process is the topic of future
work.
(20) Metal-catalyzed cross-coupling reactions; Diederich, F., Stang, P. J.,
Eds.; Wiley-VCH: New York, 1998.
(39) (a) On the importance of water in the transmetallation, see:
Carrow, B. P.; Hartwig, J. F. J. Am. Chem. Soc. 2011, 133, 2116.
Butters, M.; Harvey, J.; Jover, J.; Lennox, A.; Lloyd-Jones, G.; Murray,
P. Angew. Chem., Int. Ed. 2010, 49, 5156. (b) On the importance of
water in precatalyst activation in amination reactions, see: Fors, B. P.;
Krattiger, P.; Strieter, E.; Buchwald, S. L. Org. Lett. 2008, 10, 3505.
(c) Lou, S.; Fu, G. C. Adv. Synth. Catal. 2010, 352, 2081. (d) Miyaura,
N.; Suzuki, A. Chem. Rev. 1995, 95, 2457.
(21) Lee, C. L.; James, B. R.; Nelson, D. A.; Hallen, R. T.
Organometallics 1984, 3, 1360.
(22) (a) Grimme, S. J. Comput. Chem. 2006, 27, 1787. For
application of DFT-D for the study of organometallic catalysis, see:
(b) Kozuch, S.; Martin, J. M. L. ACS Catal. 2011, 1, 246.
(c) McMullin, C. L.; Jover, J.; Harvey, J. N.; Fey, N. Dalton Trans.
2010, 39, 10833.
(23) Dispersion correction was previously shown to be beneficial in
the treatment of dimeric metal complexes, see: Grimme, S.; Djukic, J.-
P. Inorg. Chem. 2011, 50, 2619.
(24) Regarding the appropriateness of DFT for radicals, see:
(a) Riley, K. E.; Op’t Holt, B. T.; Merz, K. M. J. Chem. Theory
Comput. 2007, 3, 407. (b) Lyaskovskyy, V.; Olivos Suarez, A. I.; Lu,
H.; Jiang, H.; Zhang, X. P.; de Bruin, B. J. Am. Chem. Soc. 2011, 133,
12264. (c) Valcarcel, A.; Clotet, A.; Ricart, J. M.; Delbecq, F.; Sautet,
P. J. Phys. Chem. B 2005, 109, 14175.
(40) We examined mixing of (i) dimer + boronic acid + water, then
KF; (ii) dimer + KF + water, then boronic acid; and (iii) dimer +
boronic acid + KF, then H2O.
(25) CPCM (THF) M06L/6-31+G(d)//B3LYP/6-31+G(d) with
SDD (for Pd) was applied. The standard state was converted to 1 M in
solution.
(26) Frisch, M. J.;et al. Gaussian09, Revision A.01 [see SI for full
reference].
(27) The ⟨S2⟩ values for the oxidative addition TSs are ⟨S2⟩ = 0.763
for C−Cl insertion by Pd(I) radical and ⟨S2⟩ = 0.772 for C−OTf
insertion, which do not deviate too much from the expected values for
a radical.
(28) The preference for C−Cl insertion is calculated to be ΔΔG⧧=
8.7 kcal/mol in THF and 8.4 kcal/mol in MeCN. Calculated with
CPCM B3LYP-D/6-31+G(d) and SDD (for Pd). See SI for details.
(29) Gas-phase energies are given. See ref 22 for application of D-
DFT in Pd catalysis. B3LYP was previously found to be an adequate
method to study such transformations: de Jong, G. T.; Bickelhaupt, F.
M. J. Chem. Theory Comput. 2006, 2, 322.
(30) (a) Pozharskii, A. F.; Soldatenkov, A. T.; Katrizky, A. R.
Heterocycles in Life and Society; Wiley: New York, 1997. (b) Reynolds,
J.; Skotheim, T. A. Handbook of Conducting Polymers, 3rd ed.; CRC
Press: Boca Raton, FL, 2007. (c) Electronic Materials: The Oligomer
Approach; Mullen, K., Wegner, G., Eds.; Wiley-VCH: Weinheim, 1998.
̈
(d) Sariciftci, N. S.; Smilowitz, L.; Heeger, A. J.; Wudl, F. Science 1992,
258, 1474.
(31) (a) Handy, S. T.; Sabatini, J. J. Org. Lett. 2006, 8, 1537.
(b) Strotman, N. A.; Chobanian, H. R.; He, J.; Guo, Y.; Dormer, P. G.;
Jones, C. M.; Steves, J. E. J. Org. Chem. 2010, 75, 1733. (c) Fairlamb, I.
J. S. Chem. Soc. Rev. 2007, 36, 1036. (d) For a review, see: Schroter, S.;
Stock, C.; Bach, T. Tetrahedron 2005, 61, 2245.
(32) (a) Gallant, M.; Belley, M.; Carriere, M.-C.; Chateauneuf, A.;
Denis, D.; Lachance, N.; Lamontagne, S.; Metters, K. M.; Sawyer, N.;
Slipetz, D.; Truchon, J. F.; Labelle, M. Bioorg. Med. Chem. Lett. 2003,
13, 3813. (b) Samantha, V.; Handy, S. T. Synthesis 2009, 1, 138.
(33) (a) Muller, H. M.; Delgado, O.; Bach, T. Angew. Chem., Int. Ed.
̈
2007, 46, 4771. (b) Delgado, O.; Muller, H. M.; Bach, T. Chem.Eur.
̈
J. 2008, 14, 2322.
(34) We also investigated the regioselectivities under Pd2(dba)3 (2.5
mol%)/PtBu3 (5 mol %) conditions: reaction of dichloropyrimidine
with 4-methoxyphenylboronic acid gave similar selectivity as entry 3,
Table 1, but lower conversion after 30 min at room temperature (31%,
612
dx.doi.org/10.1021/ja209424z | J. Am. Chem.Soc. 2012, 134, 606−612