Observation of a hidden intermediate in the stille reaction. Study of the reversal of the transmetalation step

Article abstract of DOI:10.1021/ja802994v

A study of the reaction of cis-[PdRf₂(AsPh₃)₂] (Rf = 3,5-C₆Cl₂F₃) with ISnBu₃ (that is the reversal of the natural Stille reaction of [PdRfI(AsPh₃)₂] with RfSnBu₃) allows for the observation of cis-[PdRf₂(AsPh₃)(ISnBu₃)], the expected intermediate from a cyclic transmetalation in the direct Stille reaction, thus providing experimental support to the operation of cyclic transmetalation pathways. Copyright

Full text of DOI:10.1021/ja802994v

Published on Web 07/18/2008
Observation of a Hidden Intermediate in the Stille Reaction. Study of the
Reversal of the Transmetalation Step
Mo´nica H. Pe´rez-Temprano,† Ainara Nova,‡ Juan A. Casares,*,† and Pablo Espinet*,†
IU CINQUIMA/Qu´ımica Inorga´nica, Facultad de Ciencias, UniVersidad de Valladolid, E-47071 Valladolid, Spain,
and Qu´ımica F´ısica, Edifici C.n. UniVersitat Auto`noma de Barcelona, E-08193 Bellaterra, Catalonia, Spain
Received May 5, 2008; E-mail: espinet@qi.uva.es; casares@qi.uva.es
Scheme 1
Among the palladium-catalyzed cross-coupling processes, the
Migita-Stille reaction is perhaps the most deeply studied experi-
mentally. The accepted mechanism can follow the open or cyclic
pathways shown in Scheme 1. A comprehensive account of the
mechanistic studies was published in 2004.1 Further mechanistic
studies,2–4 including detailed theoretical (DFT) calculations,5,6 have
been published recently.
The steps preceding the transmetalation have been experimentally
confirmed by observation of intermediates. Of the two intermediates
produced by transmetalation in an open mechanism (lower part of
Scheme 1), trans-[PdR1R2L2] (for which the reductive elimination
is rate-determining) was observed by NMR, whereas cis-[PdR1R2L2]
(for which the reductive elimination is faster than transmetalation)
was unambiguously deduced kinetically.7,8 However, for the cyclic
mechanism (upper part of Scheme 1), the transmetalation is usually
rate-determining9 since the cis arrangement of the R1 and R2
substituents obtained from the transmetalation makes their coupling
fast. This has precluded so far the observation of any putative
transmetalated intermediate that would further support experimen-
tally the proposed cyclic mechanism.
Our recent DFT calculations for the catalyzed coupling of PhBr
and Sn(vinyl)Me3 suggest that the formation of the intermediate
cis-[PdR1R2L2] from the starting reagents is thermodynamically
disfavored, so a successful coupling requires an easy irreversible
R1-R2 reductive elimination to pull the reaction forward.5 More-
over, a post-transmetalation intermediate with ISnBu3 acting as a
I donor ligand, [PdR1R2(AsPh3)(ISnBu3)] (I1), was predicted. The
detection of this intermediate would provide strong experimental
support to the cyclic mechanism since this intermediate does not
appear in the open mechanism.
Scheme 2
A reaction of XSnBu3 with a cis-[PdR1R2L2] complex reluctant
to reductively eliminate R1-R2 might behave (obeying the micro-
scopic reversibility principle) as the reverse of the Stille reaction,
to give trans-[PdR1XL2] and R2SnBu3. This reverse reaction should
allow for study of the hidden post-transmetalation steps in the cyclic
mechanism; since in the retro-transmetalation10 sense the intermedi-
ate I1 precedes the transmetalation step, which is the highest barrier
for the process, it might become observable. cis-[PdRf2(AsPh3)2]
(R1 ) R2 ) Rf ) 3,5-C6Cl2F3), 1,11 a stable compound that
additionally offers easy monitoring of the reactions by 19F NMR,12
was chosen to study the retro-transmetalation of the Stille reaction
(Scheme 2).
transmetalation) were irrelevant for the kinetic study of the retro-
transmetalation process, as they were observed only at times larger
than those used in the kinetic study. The other species detected by
19F NMR were the pursued intermediate [PdRf2(AsPh3)(ISnBu3)]
(I1) and the dimer [Pd2Rf2(µ-I)2(AsPh3)2] (5), already reported.13
1
In an independent study, the fast equilibrium 3 + 4 ) /2 5 + L
The reaction of 1 and ISnBu3 (2) in THF at 323 K yielded a
mixture of trans-[PdRfI(AsPh3)2] (3) and RfSnBu3 (4) (Scheme 2).
Minor side reactions (some decomposition to Pd(0) or formation
of some trans-[PdRf2(AsPh3)2], which does not undergo retro-
was fully characterized (K4(323 K) ) 3.7 × 10-5 mol L-1; ∆H0
) 19.4 ( 2.0 kcal mol-1; ∆S0 ) 39.7 ( 6.2 cal K-1 mol-1).14
These values are important for the retro-transmetalation kinetics
since the AsPh3 released modifies the rate of transmetalation.
Moreover, due to the high ∆S0 value, 5 is not observed in the NMR
spectra at room or lower temperatures. The relevance of these
† Universidad de Valladolid.
‡ Universitat Auto`noma de Barcelona.
9
10518 J. AM. CHEM. SOC. 2008, 130, 10518–10520
10.1021/ja802994v CCC: $40.75
2008 American Chemical Society

C O M M U N I C A T I O N S
Figure 1. 19F NMR spectrum (Fortho range) of the reaction in Scheme 2,
recorded at 230 K, showing the signals of intermediate I1 and a schematic
representation of the through-space 19F-19F coupling (dotted curves in the
structure). The signal / belongs to trans-[PdRf2(AsPh3)2]. 13C satellites are
observed for the very intense signals of cis-[PdRf2(AsPh3)2] and trans-
[PdRfI2(AsPh3)2].
Figure 2. Observed concentrations as a function of time of (b) the
palladium reactant 1; (9) the trans product 3 (corrected for the formation
of dimer 4). Solid lines are the best fit using Gepasi.
Scheme 3
ISnBu3. These values were used to calculate ∆Gq for the pre-
equilibrium at other temperatures.
In an independent study, ∆Gq for the transmetalation (TS2) was
determined at 323 K by monitoring the variations of concentrations
of 1 and 3 during the retro-transmetalation reaction. At this
temperature, the concentration of I1 reached a maximum very
quickly and then kept steady until the final steps of the reaction.
The data of eleven runs were fit to the multistep kinetic model
proposed in Scheme 2, by nonlinear least-squares (NLLS) regres-
sion, using the program Gepasi.21 The experimental values for the
equilibrium constants K1 (K1 ) k1/k-1; see Scheme 1) and K4 were
entered into the adjust program. Figure 2 shows the results of a
typical run (consumption of reactant and formation of trans product
as a function of time) and the curves of the Gepasi fit. Luckily, it
was found that the value of k2 was little sensitive to the value
assumed for k3 (which could not be determined experimentally).
Thus, although the Gibbs energy of TS3 could not be assigned, a
reliable value of ∆Gq(323 K) ) 26.5 kcal/mol (k2 ) 5.28 × 10-3
s-1) was calculated for TS2.
dimeric species in the Stille reaction, usually carried out at reflux,
has been noticed before.13,15–17
The predicted intermediate [PdRf2(AsPh3)(ISnBu3)] (I1) could
be unambiguously detected in the course of the kinetic experiments
below 283 K, as a deceptively simple pair of doublets of an AA′BB′
spin system, due to the through-space J coupling of ortho-fluorine
atoms of the nonequivalent perhaloaryl groups in conditions of slow
rotation of the fluoroaryl group about the C-Pd bond (Figure 1).
Irradiation of one of the signals simplifies the other to a sharp
singlet, while at 323 K the signals appear as broad triplets. This
spectroscopic and dynamic behavior has been reported in depth
for other [PdRf2L1L2] complexes.12,18,19
DFT calculations were carried out on a model where AsH3 and
ISnMe3 are used instead of AsPh3 and ISnBu3 (the computational
details are given in the Supporting Information). The calculated
∆E(gas) and ∆E(solvent, THF) values for the whole reaction are
depicted in Figure 3, along with the experimental ∆G values
determined in THF, so that calculated and experimental results can
be directly compared. Note that the experimental data only reach
to TS2 since, beyond that point, only the formation of 3 + 4 is
experimentally observed, along with the formation of dimer 5. In
agreement with the calculated ∆E0 values of 7-8 kcal/mol for the
retro-transmetalation equilibrium 1 + 2 ) 3 + 4 (including in 3
the amount corresponding to its dimerization to 5), the retro-
transmetalation reaction is completely shifted to the right, that is,
to the formation of the Stille reagents, as observed by 19F NMR.
For intermediate I1 and for TS2 (Figure 4), the total number of
particles does not change with respect to the starting system (in
each case, taking into account the molecule of ligand released, the
exchange is two particles for two particles), and a moderate entropic
contribution is expected. Consistent with this prediction, the
respective experimental values of their Gibbs energies lie between
those calculated for E in the gas phase and in solvent. This
The identity of I1 was confirmed independently. Attempts to
prepare [PdRf2(AsPh3)(THF)] (Scheme 3) as a possible precursor
for I1 afforded a complex mixture of 1 and 6-9 in THF,18,20 which
were easily identified. The addition to this mixture of a large excess
of ISnBu3 gave, as major species, cis-[PdRf2(AsPh3)2] (1), cis-
[PdRf2(ISnBu3)2] (10), and [PdRf2(AsPh3)(ISnBu3)] (I1). In addition
to the 19F NMR spectra, the 119Sn NMR spectrum confirmed three
Sn-containing species, appearing at 63.2 ppm (m) (the most intense,
ISnBu3, in excess in solution), 39.7 (I1), and 104.6 ppm (10).
The activation parameters for the formation of I1 were deter-
mined by an Eyring study using 19F NMR (at 376.50 MHz) in the
range of 225-270 K. Fortunately, at these temperatures, the slow
transmetalation step is not observed within the time of experiment.
The study afforded for TS1 the following values: ∆Hq ) 15.0 (
1.1 kcal mol-1; ∆Sq ) -17.4 ( 4.5 cal K-1 mol-1. The negative
value for ∆Sq supports an associative substitution of AsPh3 by
9
J. AM. CHEM. SOC. VOL. 130, NO. 32, 2008 10519

C O M M U N I C A T I O N S
assign associative versus dissociative processes in other cases where
only ∆Gq values were experimentally available.
In summary, an intermediate of the type cis-[PdR1R2L(ISnBu3)],
which was proposed to be the result of a cyclic transmetalation
but had never been detected in the direct reaction, has been
spectroscopically observed in a retro-transmetalation reaction. The
formation of this intermediate is a strong experimental support for
the occurrence of cyclic mechanisms in some Migita-Stille
reactions.
Acknowledgment. We thank the Ministerio de Ciencia y
Tecnolog´ıa (Project CTQ2007-67411/BQU, CTQ2005-09000-C02-
01, and a studentship to M.H.P-T.), INTECAT Consolider Ingenio
2010 (CSD2006-0003 and CSD2007-0006), and the Junta de
Castilla y Leo´n (VA-060-04) for support.
Supporting Information Available: Details of syntheses, charac-
terization of the complexes, kinetic experiments, and DFT calculations
including calculated structures. This material is available free of charge
via the Internet at http://pubs.acs.org.
References
Figure 3. Profile of the retro-transmetalation reaction 1 + 2, showing the
experimental values in THF and the calculated values ∆E(gas) and
∆E(solvent) in kcal/mol. The calculated structure of the detected intermedi-
ate I1 is also shown.
(1) Espinet, P.; Echavarren, A. M. Angew. Chem., Int. Ed. 2004, 43, 4704–
4734.
(2) Nilsson, P.; Puxty, G.; Wendt, O. F. Organometallics 2006, 27, 1285–
1292.
(3) Ariafard, A.; Lin, Z.; Fairlamb, J. S. Organometallics 2006, 27, 5788–
5794.
(4) Santos, L. S.; Rosso, G. B.; Pilli, R. A.; Eberlin, M. N. J. Org. Chem.
2007, 72, 5809–5812.
(5) Nova, A.; Ujaque, G.; Maseras, F.; Lledos, A.; Espinet, P. J. Am. Chem.
Soc. 2006, 128, 14571–14588.
´
(6) (a) Alvarez, R.; Faza, O. N.; Lopez, C. S.; de Lera, A. R. Org. Lett. 2006,
8, 35–38. (b) Alvarez, R.; Faza, O. N.; de Lera, A. R.; Cardenas, D. J.
AdV. Synth. Catal. 2007, 349, 887–906.
(7) Casado, A. L.; Espinet, P.; Gallego, A. M.; Mart´ınez-Ilarduya, J. M. Chem.
Commun. 2001, 339–340.
(8) Casado, A. L.; Espinet, P.; Gallego, A. M J. Am. Chem. Soc. 2000, 122,
11771–11782.
(9) Casado, A. L.; Espinet, P. J. Am. Chem. Soc. 1998, 120, 8978–8985.
(10) We use the word “retrotransmetalation” to indicate the exchange of an R
on Pd for an X on Sn, the reverse of what happens in the so-called
transmetalation.
(11) Espinet, P.; Mart´ınez-Ilarduya, J. M.; Pe´rez-Briso, C.; Casado, A.; Alonso,
M. A. J. Organomet. Chem. 1998, 551, 9–20.
(12) Espinet, P.; Albe´niz, A. C.; Casares, J. A.; Mart´ınez-Ilarduya, J. M. Coord.
Chem. ReV., in press.
Figure 4. Calculated geometry (DFT) for the observed intermediate I1,
and for the transmetalation transition state TS2. Other calculated geometries
are given in the Supporting Information.
(13) Casares, J. A.; Espinet, P.; Salas, G. Chem.sEur. J. 2002, 8, 4843–4853.
(14) From the integrated 19F spectra recorded between 302.6 and 329.9 K.
(15) (a) Mann, G.; Shelby, Q.; Roy, A. H.; Hartwig, J. F. Organometallics 2003,
22, 2775–2789. (b) Roy, A. H.; Hartwig, J. F. Organometallics 2004, 23,
1533–1541.
coincidence supports the idea that the calculated values give a
reasonable estimation of the system as far as entropy changes are
small.
(16) Ricci, A.; Angeluci, F.; Bassetti, M.; Lo Sterzo, C. J. Am. Chem. Soc. 2002,
124, 1060–1071.
(17) Amatore, C.; Bahsoun, A. A.; Jutand, A.; Meyer, G.; Ntpede, A. N.; Ricard,
L. J. Am. Chem. Soc. 2003, 125, 4212–4222.
(18) Casares, J. A.; Espinet, P.; Mart´ınez-Ilarduya, J. M.; Mucientes, J. J.; Salas,
G. Inorg. Chem. 2007, 46, 1027–1032.
In contrast, the E values calculated for TS1 are noticeably lower
than the experimental Gibbs energy value, but the calculated
ETS1(solvent, THF) value (12.5 kcal/mol) is not far from the
experimental ∆Hq value (15.0 ( 1.1 kcal/mol). As discussed above,
the experimental data point to an associative transition state (hence
with a considerable entropic contribution). Interestingly, a correction
for entropy of about 8-10 kcal/mol to the ETS1(solvent, THF) value
would bring the calculation very close to the ∆GqTS1 value obtained
experimentally (20.7 kcal/mol).22 Thus, in spite of some deviations,
arising from the simplification of the system using a calculation
model with smaller ligands, the estimation of E(solvent) and the
crude correction for entropy (based only in the number of particles)
behave remarkably well and should possibly be reliable enough to
(19) (a) Alonso, M. A.; Casares, J. A.; Espinet, P.; Mart´ınez-Ilarduya, J. M.;
Pe´rez-Briso, C. Eur. J. Inorg. Chem. 1998, 1745–1753. (b) de la Cruz, R.;
Espinet, P.; Gallego, A. M.; Mart´ın-Alvarez, J. M.; Mart´ınez-Ilarduya, J. M.
J. Organomet. Chem. 2002, 663, 108–117.
(20) Similar complexes cis-[PdRf2(solvent)(H2O)] and cis-[PdRf2(H2O)2], as well
as a detailed study of equilibria in acetone solution for analogous Ni
complexes
cis-[NiRf2(solvent)(H2O)],
cis-[NiRf2(H2O)2],
cis-
[NiRf2(solvent)L], and cis-[NiRf2L(H2O)], (L ) monodentate ligands,
including AsPh3), have been reported (see ref 18).
(21) (a) Mendes, P. Comput. Appl. Biosci. 1993, 9, 563–571. (b) Mendes, P.
Trends Biochem. Sci. 1997, 22, 361–363. (c) Mendes, P.; Kell, D. B.
Bioinformatics 1998, 14, 869–883. (d) Martins, A. M.; Mendes, P.;
Cordeiro, C.; Freire, A. P. Eur. J. Biochem. 2001, 268, 3930–3936.
(22) The estimated entropic contribution for only the variation in the number
of particles (considered as rigid spheres) is 8-10 kcal/mol; see: Watson,
L. A.; Eisenstein, O. J. Chem. Educ. 2002, 79, 1269–1277.
JA802994V
9
10520 J. AM. CHEM. SOC. VOL. 130, NO. 32, 2008