.
Angewandte
Communications
sufficiently active catalysts for hydroformylation of internal
alkenes;[2,9] and 2) the diamidodiindolylmethane pocket,
which can strongly bind to the carboxylate group.[13] The
rhodium ligand complex, [Rh(1)(acac)] (acac = acetylaceto-
nate), which is the precursor to the active hydroformylation
catalyst, was easily obtained by mixing a CD2Cl2 solution of
1 and [Rh(CO)2(acac)]. NMR titration experiments for
[Rh(1)(acac)] confirmed that the benzoate anion is strongly
bound within the pocket of 1 (Ka @ 105 mꢀ1, in CD2Cl2).
Molecular modeling (DFT, BP86, SV(P)) shows that, indeed,
the active form of the catalyst, [Rh(1)(CO)(H)], can bind the
model substrate deprotonated 2-vinylbenzoic acid (2a) in
a ditopic fashion (Figure 1a) with the double bond coordi-
the formation of the typical a-aldehyde product usually
formed in the hydroformylation of vinyl arenes.
As predicted by the model, hydroformylation of 2-vinyl-
benzoic acid (2a) by the Rh/1 catalyst leads to exclusive
formation of the b-aldehyde 3a, 2-(3-oxopropane)-benzoic
acid, and the reaction is 100% chemo- and regioselective
(Scheme 2a). Moreover, the activity of the catalyst is high
Scheme 2. Hydroformylation of a) 2a and b) 4 with the Rh/1 catalyst.
Product yields were determined by NMR spectroscopy and GC
analysis. DIPEA=diisopropylethylamine.
(TOF = 57 hꢀ1) under mild reaction conditions (308C, 20 bar
of CO/H2, 1:1). To demonstrate that the supramolecular
interactions between the substrate and the ligand are crucial
to obtain the selectivity, a series of control experiments with
substrates devoid of this functionality were carried out.
Hydroformylation of various styrene derivatives, with elec-
tron-withdrawing and electron-donating groups which cannot
bind in the pocket of the Rh/1 catalyst, showed typical
selectivity for a-aldehyde products, with only 3–10% of b-
aldehydes formed (see Table S3 in the Supporting Informa-
tion). The methyl ester of 2a (4), which is the closest in terms
of electronic effects but is unable to bind in the pocket, gives
only 5% of the b-aldehyde 5 and 95% of the a-aldehyde 6
(Scheme 2b), a sharp contrast to the 100% selectivity for the
b-aldehyde obtained for 2a. Moreover, the ester 4 reacts more
slowly than the acid analogue 2a (TOF = 11.7 hꢀ1 versus
57 hꢀ1) under the same reaction conditions. From the
selectivity and activity of the reactions with 2a and 4, one
can estimate the effect of substrate binding on relative
reaction rates for formation of the a- and b-aldehyde
products. Substrate preorganization accelerates the formation
of the b-aldehyde by a factor of 60, while the rate of the a-
aldehyde formation is slowed down by more than a factor 100.
These experiments clearly confirm that the high activity and
the unusual regioselectivity displayed by Rh/1 are a result of
substrate binding within the pocket of 1. For comparison,
hydroformylation of 2a with typical hydroformylation cata-
lysts, that is, rhodium complexes based on monodentate
(PPh3, P(OPh)3) or bidentate [xantphos and dppp; (xant-
phos = 4,5-bis(diphenylphosphino)-9,9-dimethylxanthene,
dppp = 1,3-bis(diphenylphosphino)propane)] ligands, under
otherwise similar reation conditions, show no or only
moderate activity (TOF < 9.5 hꢀ1) and typical regioselectivity
for the a-aldehyde.[16] Furthermore, the rhodium catalyst
containing a binol phosphite ligand without the covalently
Figure 1. Calculated reaction pathway (DFT, BP86, SV(P)) of the
regioselectivity-determining hydride-migration step in the hydroformy-
lation of 2-vinylbenzoic acid (2a) by the Rh/1 catalyst. The binol
moieties are included in calculations but omitted in the picture for
clarity. For full computational details, see the Supporting Information.
binol=1,1’-bi-2-naphthol.
nated to the metal center, while the carboxylate is held in the
binding pocket.[14] The carboxylate interaction severely
restricts the movement of the alkene moiety at the metal
center, and consequently the double bond can rotate only in
the direction of the hydride migration transition state (DGꢀ =
15.8 kJmolꢀ1; Figure 1b) which leads to the b-phenylalkyl
rhodium complex (Figure 1c).[15] The rotation towards the
transition state which leads to the a-phenylalkyl rhodium
complex is effectively blocked (DGꢀ = 71.6 kJmolꢀ1),[16] and
the usual stable p-benzyl intermediate cannot be formed
while the carboxylate of the substrate is bound in the pocket.
Therefore, this product can only be formed if either the
carboxylate leaves the binding site, or a different conformer
of the catalyst–substrate complex is formed. The complex
having inverted positions of the carbonyl and hydride such
that the formation of the a-aldehyde product is favored, is
much higher in energy (DG = 15.1 kJmolꢀ1), and also goes
through a much higher energy transition state (DGꢀ =
40.2 kJmolꢀ1).[16] Consequently, according to these calcula-
tions the bifunctional substrate binding effectively prevents
2
ꢀ 2013 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2013, 52, 1 – 6
These are not the final page numbers!