Angewandte
Chemie
the palladium center resonates at À13.3 ppm, and appears as
a double doublet, with coupling constants of 57.7 and
23.3 Hz.[22] Attempts to obtain high-resolution crystallo-
graphic data were unsuccessful, as complex 1 crystallized as
a twin, thus allowing us to only show a lower resolution X-ray
structure in Figure 3. However, the various collected spec-
troscopic data unambiguously confirmed the identity of the
PdII complex 1.
With complex 1 in hand, its ability to reductively eliminate
PhCF3 was subsequently explored. We studied the reductive
elimination of PdII complex 1 in the presence of an additional
equivalent of ligand in toluene at four different temperatures
(between 608C and 908C). The conversions were monitored
by 19F NMR spectroscopy. To our delight, the designed
complex 1 indeed gave rise to clean formation of PhCF3,
and quantitative reductive elimination was achieved in
100 min at 808C. Thus, the computational reactivity design
was successful. Figure 4 illustrates the kinetic profiles. Further
plexes for R2PCH2CH2PR2, with R = Ph (dppe), R = CF3
(dfmpe), and R = F (dfpe).[13,25] The n(CO) stretch is widely
used as a measure of the electronic properties of phosphine
ligands. Our designed ligand dfmpe (R = CF3) indeed gives
rise to a higher frequency n(CO) stretch (2116 cmÀ1) than
dppe (R = Ph, 2049 cmÀ1), which indicates that it is a better
acceptor. Dfpe (R = F) is the strongest p acceptor in the series
(n(CO) = 2120 cmÀ1). We hypothesized that if the sole reason
for higher reactivity of 1 over other small bite angle derived
PdII complexes was the greater p-acceptor properties of
dfmpe (R = CF3), then there should be an even greater
reactivity predicted for the R = F (dfpe) substituted phos-
phine PdII complex, since it is the better p-acceptor. However,
we predict a 2 kcalmolÀ1 greater energy barrier for the
reductive elimination of PhCF3 from [(R2PCH2CH2PR2)PdII-
(CF3)(Ar)] with R = F than for R = CF3 at the B3LYP level of
theory (and 1.4 kcalmolÀ1 at B3LYP-D3).[25] This suggests
that the electron-withdrawing properties of dfmpe can only in
part be responsible for its success, and that its electrostatic
repulsion with the “to-be-eliminated” CF3 (as described
above), together with low steric demand, is a crucial compo-
nent of its reactivity.[26]
In conclusion, this study presents an example of how
computational explorations can allow for the design of ligands
that trigger reactions which are not self-evident and may upon
first inspection contrast the generally accepted trends. We
report the first synthesis of a {PdII(Ph)(CF3)} complex with
a bidentate trifluoromethylphosphine ligand, and demon-
strate its high reactivity towards the reductive elimination of
PhCF3. Kinetic studies revealed the activation parameter
DH° = 27.9 Æ 1.6 kcalmolÀ1. The key design principles in the
ligand are: 1) the low steric demand of P(CF3)2 (large
substituents destabilize the TS); 2) electrostatic repulsion of
P(CF3)2 with the “to-be-eliminated” groups, which results in
reactant destabilization (and hence energy-barrier lowering);
and 3) the electron-withdrawing properties of CF3. This study
is a proof-of-principle of a successful computational reactivity
design and underlines the distinct properties of the CF3 group
in organometallic reactivity. Our future efforts are directed at
exploring these effects in catalysis.[27]
Figure 4. Reductive elimination from [(dfmpe)PdII(Ph)(CF3)] complex
1 in toluene in the presence of one equivalent of ligand.
analysis of the kinetic data gave an activation free enthalpy of
DH° = 27.9 Æ 1.6 kcalmolÀ1 for the reductive elimination of
PhCF3 from complex 1 in toluene. This value is slightly above
the computationally predicted barrier at the ONIOM level of
theory (see the Supporting Information for alternative
computational methods, namely activation barriers at DFT-
D3, M06L, M062X and solvation corrected values).
Within error margins, the observed reactivity of complex
1 is similar to that of the Xantphos-derived PdII complex and
exceptionally greater than that of analogous small bite angle
phosphine ligands, such as the completely ineffective dppe.
This highlights the special properties imposed by the CF3
substituents. Perfluoroalkyl phosphine ligands have previ-
ously been ascribed to have excellent p-acceptor proper-
ties,[23] and electron-poor metal centers have been found to
favor reductive elimination.[7b,24] To examine whether this is
also the predominant factor triggering facile reductive
elimination in our designed complex 1, we calculated the
carbonyl stretching frequencies of bis-CO-bound Pd com-
Received: January 25, 2014
Published online: && &&, &&&&
Keywords: computational chemistry · ligand effects ·
.
palladium · reductive elimination · trifluoromethylation
[1] a) A. McNally, C. K. Prier, D. W. C. MacMillan, Science 2011,
366; c) M. W. Kanan, M. M. Rozenman, K. Sakurai, T. M.
1413; e) Y. Chen, A. S. Kamlet, J. B. Steinman, D. R. Liu, Nat.
[2] For recent examples, see a) C. Mollar, M. Besora, F. Maseras, G.
b) R. B. Bedford, N. J. Gower, M. F. Haddow, J. N. Harvey, J.
Angew. Chem. Int. Ed. 2014, 53, 1 – 5
ꢀ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
3
These are not the final page numbers!