reaction with tributylamine was performed, which also resulted in
cleaved axial ligands formed by losing one butyl group.
when [Rh(cod)Cl]2 reacts with the deuterated corrole13 and NEt3
in dry CD2Cl2, the product also fails to exhibit a decreased and/or
a broadened peak at 26.02 ppm. This indicates that the corrole
NHs are also not the source of the N–H proton of the coordinated
NHEt2. We speculate that the dealkylation of NEt3 may involve
b-proton elimination. In order to examine this, trimethylamine
(which cannot undergo b-elimination), quinuclidine (which has
b-hydrogens pointing away from the active center) and
N,N-dimethylbutylamine were used as base. In these examples,
no cleavage is observed. It is noted, however, the absence of
rearrangement with these amines is consistent with the lower
steric hindrance of these ligands in addition to the changes in the
b-proton availability noted. The use of 2,2,6,6-tetramethylpiperi-
dine results in extensive decomposition. The aromatic nitrogen
base, 2-picoline gives a mixture of compounds similar to those
obtained with 2,6-lutidine; 2,6-difluoropyridine gives rise to
decomposition, while both 2,6-dichloropyridine and 2,6-di-tert-
butyl-4-methylpyridine do not react.
When a similar reaction is carried out in MeOH using 2,6-
lutidine as the base, a mixture of three compounds is isolated in a
1
yield of about 65–80%. The H NMR spectrum of this mixture
exhibits no upfield shift for the methyl groups of 2,6-lutidine,
expected of protons in close proximity to the corrole ring. 1H
NMR spectroscopy indicates that the first fraction is bis(4-
methylpyridine) RhIII corrole 4, the second fraction is a RhIII
corrole with 3-methylpyridine and 4-methylpyridine as the two
axial ligands 5, and the third fraction appears to be the bis(3-
methylpyridine) RhIII corrole 6. The structure of 5 was confirmed
using 1H COSY NMR spectroscopy. The structure of 6 was
confirmed by an X-ray analysis (Fig. 1).{ It should be noted that
the ratio of the three compounds varies in different runs of this
reaction. It is obvious that both of the methyl groups from 2,6-
lutidine have been cleaved, and one of them rearranged within the
pyridine moiety at 3- or 4-positions randomly.
The activation of C–C and C–N bonds in these axial ligands
likely follows different mechanisms. It is plausible that cleavage of
a C–N bond involves b-H elimination and C–C bond cleavage
involves the insertion of RhIII into the C–C bond as evidenced by
other transition metal catalyzed C–C bond activations in
homogenous systems.14,15 We believe that steric factors play
important roles in both rearrangements. As indicated above, C–C
and C–N bond activation does not occur with all bulky nitrogen
bases. Some bases do not react (2,6-dichloropyridine and 2,6-di-
tert-butyl-4-methylpyridine) and some decompose (2,2,6,6-
tetramethylpiperidine, 2,6-difluoropyridine). We tentatively
interpret the results as follows: in the metalation process with a
suitable, sterically bulky base, a five-coordinated intermediate
should form. This tentative 5-coordinated RhIII corrole would be a
16-electron species requiring one more axial ligand to attain a full-
shell configuration; however, steric strain induced by the bulk of
the base, along with the limited flexibility of the corrole macrocycle
leads to dealkylation of the axial base (NEt3 and 2,6-lutidine).
Competition may exist between cleavage and degradation of
the 5-coordinated intermediate. Therefore in some cases
decomposition is observed (2,2,6,6-tetramethylpiperidine, 2,6-
difluoropyridine). In the case of exceptionally bulky bases
(2,6-dichloropyridine and 2,6-di-tert-butyl-4-methylpyridine), the
formation of this five-coordinated intermediate may not be
accessible and starting material is recovered. It is interesting to
note that this reaction occurs at room temperature and that RhIII
corrole adducts adopt an octahedral coordination sphere with
both axial bases undergoing rearrangement.
When [Rh(cod)Cl]2 is mixed with TPFC in CH2Cl2 or MeOH in
the absence of any base, no reaction is detected. When
[Rh(cod)Cl]2 is treated with NEt3 in CH2Cl2 or 2,6-lutidine in
MeOH, without any free base corrole, again, no cleavage is
observed. It seems that the activation of the C–C and C–N bonds
occurs after insertion of the metal into the corrole. When the
reaction is carried out under anaerobic conditions in the presence
of either NEt3 or 2,6-lutidine, the color of the solution does not
change from green to red (as observed for the reactions carried out
in air) and a new brownish green compound is isolated. The nature
of this new compound is currently under investigation. It appears
that the oxidation of RhI to RhIII is essential for the cleavage
process to proceed. When [Rh(cod)Cl]2 reacts with TPFC in
CD2Cl2 in the presence of dry NEt3, the 1H NMR spectrum of the
product does not show broadening or a decrease in the peak
intensity at 26.02 ppm. These results suggest that the N–H proton
of the axial NHEt2 does not arise from the solvent. Alternatively,
We thank Mr Xiangjin Xie, Dr Xavier Ottenwaelder and
Dr Xiaoping Wang for providing X-ray crystallographic data.
James P. Collman,* Hong J. H. Wang, Richard A. Decreau,
Todd A. Eberspacher and Christopher J. Sunderland
Department of Chemistry, Stanford University, Stanford, CA 94305,
USA. E-mail: jpc@stanford.edu
Notes and references
{ Crystal data for 6: C55H36F15N6Rh, M 5 1168.81, orthorhombic, space
˚
group P212121. a 5 8.6618(15) A, b 5 16.427(3) A, c 5 33.971(6) A,
˚
˚
3
Fig. 1 The core of molecule 6 drawn with 50% probability ellipsoids. All
hydrogen atoms of the disordered 3-methylpyridine groups and the C6F5
groups have been omitted for clarity.
˚
a 5 90.00u, b 5 90.00u, n 5 90.00u, V 5 4833.7(14) A , Z 5 4,
m (Mo-Ka) 5 0.46 mm21, F(000) 5 2352.0. The structure was solved by
direct methods, and refined as a racemic twin. One of the 3-methylpyridine
2498 | Chem. Commun., 2005, 2497–2499
This journal is ß The Royal Society of Chemistry 2005
Collman, James P.
