Structural and oxo-transfer reactivity differences of hexacoordinate and pentacoordinate (nitro) (tetraphenylporphinato)cobalt(III) derivatives

Article abstract of DOI:10.1021/ic001442e

The oxo-transfer catalyst (nitro)(pyridyl)cobalt(III) tetraphenylporphyrin has been reinvestigated by substitution of the distal pyridine ligand with 4-N,N-dimethylaminopyridine and 3,5-dichloropyridine. Differences in their structures and in the reactivity of the compounds toward catalytic secondary oxo transfer were investigated by FT-IR and UV-visible spectroscopy, cyclic voltammetry, X-ray diffraction, semiempirical calculations, and reactions with alkenes in dichloromethane solution. Very modest differences in the hexacoordinate compounds' structures were predicted and observed, but the secondary oxo-transfer reactivity at the nitro ligand varies markedly with the basicity of the pyridine ligand and the position of the coordination equilibrium. Oxo transfer occurs rapidly through the pentacoordinate species (nitro)cobalt(III) tetraphenylporphyrin that is generated by dissociation of the pyridine ligand and therefore is strongly related to the Hammett parameters of these nitrogenous bases. The reactive pentacoordinate species CoTPP(NO₂) can be generated in solution by addition of lithium perchlorate to (py)CoTPP(NO₂) by Lewis acid-base interactions or more simply by using the weaker Lewis base Cl₂py instead of py as the distal ligand. In contrast to pentacoordinate (nitro)iron porphyrins, disproportionation reactions of CoTPP(NO₂) compound are not evident. This pentacoordinate derivative, CoTPP(NO₂), is reactive enough to stoichiometrically oxidize allyl bromide in minutes. Preliminary catalytic oxidation reaction studies of alkenes also indicate the involvement of both radical and nonradical oxo-transfer steps in the mechanism, suggesting formation of a peroxynitro intermediate in the reaction of the reduced CoTPP(NO) with O₂.

Full text of DOI:10.1021/ic001442e

Inorg. Chem. 2001, 40, 4217-4225
4217
Structural and Oxo-Transfer Reactivity Differences of Hexacoordinate and Pentacoordinate
(Nitro)(tetraphenylporphinato)cobalt(III) Derivatives
John Goodwin,*,† Rosa Bailey,‡ William Pennington,‡ Roger Rasberry,† Tony Green,†
Shira Shasho,† May Yongsavanh,† Valerie Echevarria,† Jessica Tiedeken,† Christina Brown,†
George Fromm,† Stephanie Lyerly,§ Natalie Watson,§ Amy Long,§ and Nick De Nitto§
Department of Chemistry and Physics, Coastal Carolina University, P.O. Box 261954,
Conway, South Carolina 29526-6054, Department of Chemistry, Clemson University,
Clemson, South Carolina 29634-1905, and The South Carolina Governor’s School for
Mathematics and Science, Hartsville, South Carolina 29550
ReceiVed December 22, 2000
The oxo-transfer catalyst (nitro)(pyridyl)cobalt(III) tetraphenylporphyrin has been reinvestigated by substitution
of the distal pyridine ligand with 4-N,N-dimethylaminopyridine and 3,5-dichloropyridine. Differences in their
structures and in the reactivity of the compounds toward catalytic secondary oxo transfer were investigated by
FT-IR and UV-visible spectroscopy, cyclic voltammetry, X-ray diffraction, semiempirical calculations, and
reactions with alkenes in dichloromethane solution. Very modest differences in the hexacoordinate compounds’
structures were predicted and observed, but the secondary oxo-transfer reactivity at the nitro ligand varies markedly
with the basicity of the pyridine ligand and the position of the coordination equilibrium. Oxo transfer occurs
rapidly through the pentacoordinate species (nitro)cobalt(III) tetraphenylporphyrin that is generated by dissociation
of the pyridine ligand and therefore is strongly related to the Hammett parameters of these nitrogenous bases.
The reactive pentacoordinate species CoTPP(NO2) can be generated in solution by addition of lithium perchlorate
to (py)CoTPP(NO2) by Lewis acid-base interactions or more simply by using the weaker Lewis base Cl2py
instead of py as the distal ligand. In contrast to pentacoordinate (nitro)iron porphyrins, disproportionation reactions
of CoTPP(NO2) compound are not evident. This pentacoordinate derivative, CoTPP(NO2), is reactive enough to
stoichiometrically oxidize allyl bromide in minutes. Preliminary catalytic oxidation reaction studies of alkenes
also indicate the involvement of both radical and nonradical oxo-transfer steps in the mechanism, suggesting
formation of a peroxynitro intermediate in the reaction of the reduced CoTPP(NO) with O2.
Introduction
for industrial catalysis in the late 1970s and early 1980s, and
was promising because of its catalytic activity for a wide range
of oxidation reactions. Several studies of this compound dealt
with a variety of mechanisms corresponding to the range of
reactions, and these were proposed to include bimolecular oxo
transfer to activated olefins, intramolecular oxo-transfer reactions
promoted by Lewis acids, and nitro exchange reactions.2,3
However, the catalyst had low turnover numbers and required
cocatalysts for oxidation of alkenes. No systematic structural
variations of the cobalt porphyrin compound for tuning its oxo-
transfer reactivity or detailed examinations of the mechanism
of Lewis acid activation have been presented. Now, through
study of a series of derivatives of the parent compound that
varied the basicity of the distal pyridine ligand, the importance
of the Lewis acid in promoting dissociation of this distal pyridine
ligand in oxo-transfer reactivity has been clarified. The penta-
coordinate nitro complex has been found to be much more
reactive than the hexacoordinate pyridyl analogues to the extent
that this catalyst is for the first time able to promote dioxygen
activation and oxidation of unactivated alkenes. Reactivity with
alkenes indeed appears to be limited to this pentacoordinate
species. This observation parallels a recent report of extreme
oxo-transfer reactivity of pentacoordinate (nitro)iron(III) picket-
fence porphyrin prepared in solution by Munroe and Scheidt.4
Secondary oxo-transfer reactions from metal-coordinated
nitrite occur with formation of reduced metal nitrosyl complexes.
In some of these reduced nitrosyl products, the NO ligand can
then react with dioxygen to be oxidized back to the nitro ligand
with concomitant reoxidation of the metal center. This reaction
sequence has been an intriguing prospect for dioxygen activation
and industrial oxo-transfer catalysis for over two decades.1,2 The
advantages of this approach are primarily in avoiding radical
mechanisms that reduce specificity of reaction and stability of
the catalysts by autoxidation and secondly in the use of dioxygen
as the ultimate oxidant rather than organic peroxides or hydrogen
peroxide.
The complex (nitro)(pyridyl)(tetraphenylporphinato)cobalt-
(III), (py)CoIIITPP(NO2), was studied2 in detail and patented
† Coastal Carolina University.
‡ Clemson University.
§ Students from The South Carolina Governor’s School for Mathematics
and Science.
(1) Cotton, F. A.; Wilkinson, G. AdVanced Inorganic Chemistry, 5th ed.;
Wiley: New York, 1988; pp 1281-1282 and references therein.
(2) (a) Tovrog, B. S.; Diamond, S. E.; Mares, F. J. Am Chem. Soc. 1979,
101, 270-272. (b) Tovrog, B. S.; Mares, F.; Diamond, S. E. J. Am.
Chem. Soc. 1980, 102, 6616-6618. (c) Tovrog, B. S.; Diamond, S.
E.; Mares, F.; Stalkiewicz, A. J. Am. Chem. Soc. 1981, 103, 3522-
3526. (d) Diamond, S. E.; Mares, F.; Szalkiewicz, A.; Muccigrosso,
D. A.; Solar, J. P. J. Am. Chem. Soc. 1982, 104, 4266-4268. (e)
Ercolani, C.; Paoletti, A. M.; Pennesi, G.; Rossi, G. J. Chem. Soc.,
Dalton Trans. 1991, 1317-21.
(3) Andrews, M. A.; Chang, C.-T.; Cheng, C.-W. Organometallics 1985,
4, 268-274.
10.1021/ic001442e CCC: $20.00 © 2001 American Chemical Society
Published on Web 07/14/2001

4218 Inorganic Chemistry, Vol. 40, No. 17, 2001
Goodwin et al.
(3,5-Dichloropyridyl)(nitro)cobalt(III) Tetraphenylporphyrin,
(Cl2py)CoIIITPP(NO2). This derivative was prepared only by the
Andrews method using AgNO2 oxidation of CoIITPP under N2
atmosphere in the presence of at least 10-fold excess of 3,5-
dichloropyridine, Cl2py (Aldrich). The compound was purified by silica
gel chromatography as described earlier. Crystallization of the porphyrin
mixture in the presence of excess Cl2py from CHCl3 by evaporation or
by vapor diffusion of hexane into toluene solutions resulted in a mixture
of crystals. A red-purple parallelepiped crystal of (Cl2py)CoIIITPP(NO2)‚
1/2(CHCl3) was selected for X-ray data collection. Anal. Calcd for
C49H31N6O2Cl2Co: C, 67.98; H, 3.609; N, 9.708. Found: C, 66.79; H,
4.45; N, 8.10. UV-vis: λmax ) 431 nm. IR: νNO 1424 cm-1, νNO
However, unlike the observations for the iron derivative, this
pentacoordinate (nitro)cobalt system is apparently stable with
respect to disproportionation reactions.
Experimental Section.
General Methods. Spectroscopic Methods. Infrared spectra were
recorded as KBr disks using a Bomem Michaelson FT-IR system.
Solution UV-visible spectra were recorded under nitrogen atmosphere
using screw-capped 1-cm path length cells with a Hewlett-Packard 8453
diode array spectrometer or in a Vacuum-Atmospheres Glovebox with
an Ocean Optics PC1000 fiber optics spectrometer controlled by
OOIBASE 1.5 or OOIBASE32 operating software. Reflectance spectra
of porphyrins adsorbed on silica gel were collected with the fiber optics
spectrometer using an Ocean Optics reflectance probe. Proton NMR
spectra were recorded in CDCl3 or CD2Cl2 (Aldrich) solution using a
Varian EM360 magnet that was controlled with an Anasazi EFT60
spectroscopy system. Elemental analyses were carried out by Galbraith
Laboratories.
Synthesis of Porphyrin Derivatives. The (nitro)(pyridyl)cobalt(III)
tetraphenylporphyrin derivatives were all prepared by the stoichiometric
reaction of CoIITPP (Aldrich or Midcentury Chemicals) with AgNO2
(Acros Organics, 99%) in the presence of pyridine or its derivatives in
refluxing acetonitrile (Sigma) under nitrogen atmosphere according to
the method of Andrews.3 All materials, except for dichloromethane
solvent used for extraction in these syntheses, were used as received.
It was important to remove the amylene stabilizer from this solvent
prior to its use by washing with sulfuric acid, sodium hydroxide, and
water before drying with calcium chloride and distilling from phos-
phorus pentoxide under nitrogen.5 The parent compound, (nitro)-
(pyridyl)cobalt(III) tetraphenylporphyrin, was also prepared by air
oxidation of CoIITPP(NO) in toluene in the presence of excess pyridine.6
The compounds were purified with column chromatography on silica
gel (Fisher 35-60 mesh, type 60A) eluting first with toluene followed
by a 10% solution of methanol in toluene. The second eluting orange-
red porphyrin fraction contained the desired products.
(Pyridyl)(nitro)cobalt(III) Tetraphenylporphyrin, (py)CoIIITPP-
(NO2). The previous synthetic procedure of Andrews3 et al. was
followed, except for purification using column chromatography. The
synthetic reactions were carried out under nitrogen using standard
Schlenk techniques, but the compound was then handled in the air.
FTIR: νNO2 1425 cm-1, νNO2 1309 cm-1, and δO-N-O 814 cm-1. UV-
vis: λmax 436 nm. This compound was also made by the original method
of Tovrog et al.2 by air oxidation of CoIITPP(NO) in toluene with excess
pyridine. Crystals of the pyridyl solvate (py)CoIIITPP(NO2)‚1/2(pyridine)
were grown by vapor diffusion of hexane into ∼1 mM solutions of
CoIITPP(NO) originally dissolved under nitrogen atmosphere in ∼1:1
toluene/methanol mixtures with pyridine at ∼0.1 M. These solutions
were allowed to react with air as crystallization took place over 2 weeks.
A purple plate was selected for X-ray diffraction analysis.
2
2
1309 cm-1, and δO-N-O 817 cm-1
.
Solution Phase (Nitro)(tetraphenylporphinato)cobalt(II), CoTPP-
(NO2). Solution preparation of CoTPP(NO2) was carried out by
reactions of (Cl2py)CoTPP(NO2) with lithium perchlorate in CH2Cl2.
This species was also observed in equilibrium with (Cl2py)CoTPP-
(NO2) and (py)CoTPP(NO2) in CH2Cl2 solution at low concentrations.
UV-vis: λmax ) 417 nm. It has also been previously reported by
Andrews as part of a mixture in the solid state with assigned infrared
bands νsNO of 1283 cm-1 and νaNO of 1453 cm-1
.
2
2
(3,5-Dichloropyridyl)(nitro)(tetra-r,r′,r′′,r′′′-(o-pivalamido-
phenyl)porphinato)cobalt(III), (Cl2py)CoIIITpivPP(NO2). This picket-
fence porphyrin derivative was prepared only by the Andrews method
using AgNO2 oxidation of CoIITpivPP under N2 atmosphere in the
presence of at least 10-fold excess of 3,5-dichloropyridine, Cl2py
(Aldrich). The compound was purified by silica gel chromatography.
Porphyrins Adsorbed on Silica Gel. (Nitro)cobaltTPP was adsorbed
onto silica gel (Sigma Type G with CaSO4 binder, 10-40 µm) by low-
pressure evaporation of CH2Cl2 solutions of (Cl2py)CoTPP(NO2) alone
and in the presence of LiClO4. Mass proportions of porphyrin to silica
gel were roughly 1:50. CoTPP(NO) was adsorbed onto silica gel (1:10
proportions) by evaporation immediately after preparation in toluene
solution by Schlenk techniques. The electronic spectrum before
evaporation was taken under nitrogen atmosphere in the glovebox with
an Ocean Optics PC1000 fiber optics spectrometer. The adsorbed
CoTPP(NO) was redissolved with CH2Cl2 under nitrogen to produce
an identical electronic spectrum with Soret position at 414 nm,
consistent with the assignment of CoTPP(NO). Reflectance electronic
spectra of the nitro derivatives were recorded using an Ocean Optics
reflectance probe. Regardless of the method of preparation, the Soret
peak appeared at 428 nm.
X-ray Crystallography. (py)CoIIITPP(NO2)‚1/2(pyridine). Data was
collected with a Rigaku AFC7R diffractometer using graphite-mono-
chromated Mo KR radiation (λ ) 0.71073 Å) at a room temperature
of 22 ( 1 °C. The structure was solved by direct methods and refined
using full-matrix least-squares on F2 (SHELXTL PLUS).
(Cl2py)CoIIITPP(NO2)‚1/2(CHCl3). Data was collected using a
Rigaku AFC8/Mercury CCD using graphite-monochromated Mo KR
radiation (λ ) 0.71073 Å) at a room temperature of 22 ( 1 °C. The
structure was solved by direct methods and refined using full-matrix
least-squares on F2 (SHELXTL PLUS).
(3,5-Lutidine)(nitro)cobalt(III) Tetraphenylporphyrin, (lut)-
CoIIITPP(NO2). This compound was prepared from CoTPP with
AgNO2 oxidation in the presence of a stoichiometric quantity of 3,5-
lutidine (Aldrich) in refluxing acetonitrile under nitrogen. Anal. Calcd
for C51H39N6O2Co: C, 74.08; H, 4.75; N, 10.16. Found: C, 74.41; H,
4.48; N, 8.45. UV-vis: λmax 436 nm. IR: νNO 1425 cm-1, νNO 1307
Electrochemistry. Cyclic voltammetry was performed in a nitrogen-
filled glovebox (Vacuum Atmospheres) using a EG&G VersaStat
system and a three-electrode arrangement of a Cypress Systems
platinum minielectrode, a nonaqueous (Ag/AgNO3-CH3CN) reference
electrode (Bioanalytical Systems), and a platinum wire counter
electrode. Solutions were prepared in CH2Cl2 purified as described
above. UV-visible spectra were checked immediately prior to elec-
trochemistry experiments to ensure that no decomposition had occurred.
The supporting electrolyte, tetrabutylammonium hexafluorophosphate,
TBAH (Aldrich), was recrystallized from ethyl acetate three times and
dried at 110 °C in a vacuum oven for 5 days.5 Its concentration in the
electrochemistry was 0.10 M. A scan rate of 20 mV/s over the range
of 0.0 to +2.0 V vs Ag/AgNO3-CH3CN was used for all measure-
ments. Ferrocene was used as an internal standard. Solutions were
initially prepared as 1 mM in porphyrin, tested for purity by UV-
visible spectroscopy in the glovebox, and immediately examined by
cyclic voltammetry. Generation of the pentacoordinate nitro complex
in solution was carried out by saturating the test solution of (Cl2py)-
2
2
cm-1, and δO-N-O 815 cm-1. This derivative was previously synthesized
by oxidation of CoIITPP(NO) in the presence of 3,5-lutidine and studied
by X-ray diffraction methods.6
(4-N,N-Dimethylaminopyridyl)(nitro)cobalt(III) Tetraphenylpor-
phyrin, (Me2Npy)CoIIITPP(NO2). This compound was prepared from
CoTPP with AgNO2 oxidation in the presence of a stoichiometric
quantity of 4-N,N-dimethylaminopyridine (Aldrich) in refluxing aceto-
nitrile under nitrogen. Anal. Calcd for C51H38N7O2Co: C, 72.93; H,
4.56; N, 11.67. Found: C, 71.90; H, 4.53; N, 12.58. UV-vis: λmax
)
.
438 nm. IR: νNO 1427 cm-1, νNO 1310 cm-1, and δO-N-O 813 cm-1
2
2
(4) Munro, O. Q.; Scheidt, W. R. Inorg. Chem. 1998, 37, 2308-2316.
(5) Armarego, W. L. F.; Perrin, D. D. Purification of Laboratory
Chemicals, 4th ed.; Butterworth-Heinemann: Oxford, 1996.
(6) Kaduk, J. A.; Scheidt, W. R. Inorg. Chem. 1974, 13, 1875-1880.

(Nitro)(tetraphenylporphinato)cobalt(III) Derivatives
Inorganic Chemistry, Vol. 40, No. 17, 2001 4219
Table 1. Summary of Measured and Predicted Nitro Ligand Structural Features
measd O-N-O
measd NO
bond lengths, Å
[av] (predicted)
measd Co-Nnitro
bond lengths, Å
(predicted)
measd Co-Npy
bond lengths, Å
(predicted)
Hammett
params
bond angle, deg
[av] (predicted)b
(Cl2py)CoTPP(NO2)
(py)CoTPP(NO2)
σm ) +0.64a
119.7(4)
(120.66)
1.215(4)
1.220(4)
(1.223)
1.357(16)
1.245(12)
1.079(17)
1.197(15)
[1.220(16)]
(1.223)
1.155(5)
(1.224)
NA
(1.224)
(1.220)
(1.221)
(1.222)
1.223(3)
(1.224)
1.227(3)
(1.224)
1.182(14)
1.207(14)
(1.223)
1.912(3)
(1.877)
1.952(3)
(1.956)
σp ) 0
σm ) 0
118.9(9)
121.6(13)
122.5(8)
119.3(11)
[120.6(21)]
(120.39)
115.4(6)
(120.53)
NA
(120.52)
(124.41)
(122.09)
(121.10)
119.8(4)
(120.09)
120.4(4)
(119.96)
115.4(11)
(120.92)
1.920(4)
(1.878)
2.032(4)
(1.961)
(lut)CoTPP(NO2)c
σm ) -0.10
σp ) -0.80a
1.948(4)
(1.878)
NA
2.036(4)
(1.958)
NA
(1.955)
NA
(Me2Npy)CoTPP(NO2)
(1.878)
(1.832)
(1.844)
(1.862)
1.898(4)
(1.875)
1.917(4)
(1.876)
1.897(11)
(1.869)
CoTPP(NO2)
(F5py)CoTPP(NO2)
(ClO4)CoTPP(NO2)
(1-MeIm)CoTpivPP(NO2)
σ ) +2.60
NA
(2.414)
1.995(4)
(1.955)
2.091(4)
(1.970)
2.044(10)
(2.064)
(1,2-Me2Im)CoTpivPP(NO2)
(pip)CoTPP(NO2)
a Hammett values were calculated according to the equation σ ) log(Ka/Ka0) where Ka is for the identically substituted benzoic acid derivative.
These pKa’s were obtained through the result of ACD/pKa DB v 4.0 that was accessed through the ACD/I-Lab Service. The pKa of 3,5-dichlorobenzoic
acid ) 3.56, the pKa of 4-N,N-dimethylaminobenzoic acid ) 5.00, the pKa of 3,6-dimethylbenzoic acid ) 4.30, and the pKa of pentafluorobenzoic
acid ) 1.60. These were all compared to the pKa of benzoic acid of 4.202. b Predicted by semiempirical calculations of equilibrium geometries
using the PM3 basis sets. c lut ) 3,5-lutidine. The X-ray crystal structure of this derivative was reported by Kaduk and Scheidt.4
CoTPP(NO2) with LiClO4. No corrections to the current were made
for increased ionic strength due to this addition.
nitrogen atmosphere. Further reactions with pyridine additions were
also carried out under nitrogen atmosphere.
Computational Modeling. Semiempirical calculations were carried
out using PM3 basis sets with PC Spartan Plus (Wavefunction, Inc.).
Equilibrium geometries and vibrational frequencies of the nitro ligands
were determined for all of the hexacoordinate (py)CoTPP(NO2)
derivatives and for the pentacoordinate CoTPP(NO2) derivative. No
corrections were made to the calculated vibrational frequencies.
Equilibrium Measurements. Equilibrium of association of the
respective pyridine ligands with the (nitro)cobalt tetraphenylporphyrins
was studied by UV-visible spectrophotometry in CH2Cl2 solution by
preparing a series of solutions with constant porphyrin concentration,
but with varying concentration of the pyridine ligand. Data were
analyzed for both the single equilibrium of association of the pyridine
ligand and for the limiting molar absorptivities of the equilibrium
species by use of the PC/DOS version of the SPECDEC program.7
Eight wavelengths ranging from 360 to 465 nm and six solution
concentrations of the pyridine ligand were used for each experiment.
Linear baseline corrections were applied to the spectral data.
Reaction Studies. The stoichiometric oxidation of triphenylphos-
phine, PPh3, that results in formation of triphenylphosphine oxide and
(nitrosyl)cobalt(II) tetraphenylporphyrin as described by Tovrog2a was
used initially to compare oxo-transfer reactivity of the derivatives. These
experiments were carried out under nitrogen atmosphere to allow
observation of the reduction reaction of the porphyrin complex without
the complication of ensuing oxidation by air. Solutions of porphyrins
with and without added pyridine derivative in solution were freshly
prepared under nitrogen, and solid triphenylphosphine was added and
dissolved immediately. Comparisons in oxo-transfer reactivity were also
made by anaerobic addition of allyl bromide to solutions of porphyrins
in dichloromethane solution.
Catalytic oxidation reactions with alkenes were qualitatively studied
by allowing solutions of (Cl2py)CoTPP(NO2) (∼1 mM) in CH2Cl2 (∼1
mL) to react at room temperature with excess (1 mL) cyclohexene
(Acros), allyl bromide (Acros), and styrene (Acros, stabilized with 10-
15 ppm p-tert-butylcatechol). Reactions were carried out in an oxygen-
rich atmosphere in vials capped with toy latex balloons filled with
dioxygen gas attached and were allowed to proceed either in the dark
or under fluorescent laboratory lighting. Samples were prepared with
and without addition of lithium perchlorate to the solutions. Catalysts
adsorbed on silica gel were reacted with neat liquid alkenes with the
same type of oxygen-filled balloon. The products of these reactions
were determined by GC-MS using a Hewlett-Packard 6890 GC system
with a 5793 ion selective detector.
Results Section
Computational Modeling. Results for the calculations of the
equilibrium geometries by semiempirical methods using PM3
basis are tabulated in Table 1. In this table are listed measured
and calculated nitro ligand NO bond lengths, Co-Nnitro and Co-
Npy bond lengths, and nitro ONO bond angles for pyridine
derivatives prepared in this work. Also listed are calculated
results for the proposed pentacoordinate nitro species and the
hypothetical pentafluoropyridyl derivative. While both the
calculated and measured nitro ligand parameters are quite
consistent for the pyridyl derivatives actually prepared in this
work, complete removal of the pyridine ligand in the calculations
gives a significantly larger ONO bond angle of 124° compared
to the 120° angles predicted for the pyridyl derivatives. The
semiempirical model of the hypothetical pentafluoropyridyl and
perchlorato derivatives gave intermediate values of 122° and
121° for the respective ONO bond angles and a large increase
in the Co-Npyr bond length from roughly 2 to 2.4 Å for the
pentafluoropyridyl derivative.
Potential disproportionation reactions4 were also monitored under
nitrogen atmosphere by UV-visible spectroscopy in solutions of
porphyrins prepared from freshly made solid samples. Lithium per-
chlorate additions were made immediately prior to data collection under
(7) Atkins, C. E.; Park, S. E.; Blaszek, J. A.; McMillin, D. R. Inorg. Chem.
1984, 23, 569-572.

4220 Inorganic Chemistry, Vol. 40, No. 17, 2001
Goodwin et al.
Table 2. Crystallographic Data for (py)CoTPP(NO2)‚0.5(py) and
(Cl2py)CoTPP(NO2)‚0.5(CHCl3)
Table 3. Bond Distances and Angles of the Cobalt Coordination
Site for (py)CoTPP(NO2)‚1/2(py) and
(Cl2py)CoTPP(NO2)‚1/2(CHCl3)
chem formula
C49H33N6O2Co‚
0.5(C5H5N)
836.29
C49H31N6O2Cl2Co‚
0.5(CHCl3)
925.35
(A) (py)CoTPP(NO2)‚1/2(py)
fw
Distances (Å)
space group
a (Å)
b (Å)
c (Å)
â (deg)
V (Å3)
Z
P21/n (No. 14)
13.231(3)
23.566(3)
14.537(2)
101.02(1)
4449(1)
4
P21/n (No. 14)
13.3698(9)
23.216(3)
15.1300(3)
105.613(3)
4523.1(5)
4
Co(1)-N(1)
Co(1)-N(4)
Co(1)-N(3)
O(1A)-N(1)
O(1B)-N(1)
O(1C)-N(1)
1.920(4)
1.964(3)
1.977(4)
1.357(16)
1.245(12)
1.254(13)
Co(1)-N(5)
Co(1)-N(6)
Co(1)-N(2)
O(2A)-N(1)
O(2B)-N(1)
O(2C)-N(1)
1.959(3)
1.974(4)
2.032(4)
1.079(17)
1.197(15)
1.22(2)
T (°C)
D
22 ( 1
22 ( 1
Angles (deg)
calcd (g cm-3
)
1.25
1.36
N(1)-Co(1)-N(5)
N(5)-Co(1)-N(4)
N(5)-Co(1)-N(6)
N(1)-Co(1)-N(3)
N(4)-Co(1)-N(3)
N(1)-Co(1)-N(2)
N(4)-Co(1)-N(2)
N(3)-Co(1)-N(2)
88.50(16) N(1)-Co(1)-N(4)
89.59(14) N(1)-Co(1)-N(6)
90.43(15) N(4)-Co(1)-N(6)
90.79(16) N(5)-Co(1)-N(3)
90.05(15) N(6)-Co(1)-N(3)
178.63(17) N(5)-Co(1)-N(2)
89.91(16) N(6)-Co(1)-N(2)
90.24(16)
90.76(16)
179.00(16)
179.20(15)
89.95(15)
90.14(15)
89.09(16)
λ (Å)
0.71073
4.3
0.065
0.71073
6.3
0.059
µ (cm-1
)
R1 (Fo )a
2
2
wR2 (Fo )b
0.213
0.193
2
a R1 ) ∑||Fo| - |Fc||/∑|Fo| for I > 2σ(I). b wR2 ) {∑[w(Fo
-
Fc2)2]/∑[w(Fo )2]}1/2 for all data.
2
90.57(16) O(2B)-N(1)-O(1B) 118.9(9)
O(2C)-N(1)-O(1C) 121.6(13) O(2A)-N(1)-O(1A) 118.7(12)
O(2A)-N(1)-Co(1) 125.0(10) O(2B)-N(1)-Co(1) 122.5(8)
O(2C)-N(1)-Co(1) 119.3(11) O(1B)-N(1)-Co(1) 116.7(6)
O(1C)-N(1)-Co(1) 118.2(6)
(B) (Cl2py)CoTPP(NO2).1/2(CHCl3)
Distances (Å)
O(1A)-N(1)-Co(1) 116.0(7)
Co(1)-N(1)
Co(1)-N(6)
Co(1)-N(3)
O(1)-N(1)
1.912(3)
Co(1)-N(5)
Co(1)-N(4)
Co(1)-N(2)
O(2)-N(1)
1.952(3)
1.955(3)
2.044(3)
1.220(4)
1.952(3)
1.959(3)
1.215(4)
Angles (deg)
N(1)-Co(1)-N(5)
N(5)-Co(1)-N(6)
N(5)-Co(1)-N(4)
N(1)-Co(1)-N(3)
N(6)-Co(1)-N(3)
90.83(13) N(1)-Co(1)-N(6)
90.86(13)
89.57(13)
89.77(12) N(1)-Co(1)-N(4)
90.77(12) N(6)-Co(1)-N(4) 179.30(13)
91.87(14) N(5)-Co(1)-N(3) 177.10(13)
89.13(12) N(4)-Co(1)-N(3)
90.30(12)
89.84(13)
90.59(13)
119.7(4)
Figure 1. ORTEP plot of (py)CoTPP(NO2) at the 30% probability
level.
N(1)-Co(1)-N(2) 179.31(14) N(5)-Co(1)-N(2)
N(6)-Co(1)-N(2)
N(3)-Co(1)-N(2)
88.97(12) N(4)-Co(1)-N(2)
87.45(13) O(1)-N(1)-O(2)
Crystal Structures. (py)CoTPP(NO2)‚1/2(py). The parent
pyridine derivative crystallized in the monoclinic P21/n space
group with four porphyrin molecules per unit cell. The unit cell
dimensions were a ) 13.231(3) Å, b ) 23.566(3) Å, and c )
14.537(2) Å; â ) 101.02(1)°. The crystallographic data is
summarized in Table 2. Disorder in the nitro ligand was modeled
by four equivalent oxygen atoms at half-occupancy. The
molecular structure of the porphyrin is illustrated at the 30%
probability level in the ORTEP diagram in Figure 1. Bond
lengths and angles are listed in Table 3A. Nitro ligand bond
lengths and angles are compared with those determined by
semiempirical modeling in Table 1.
O(1)-N(1)-Co(1) 119.6(3)
O(2)-N(1)-Co(1) 120.7(3)
C(5)-N(2)-C(1)
118.7(4)
C(5)-N(2)-Co(1) 120.6(3)
C(1)-N(2)-Co(1) 120.4(3)
C(6)-N(3)-Co(1) 126.6(2)
C(11)-N(4)-C(14) 106.4(3)
C(14)-N(4)-Co(1) 126.6(2)
C(16)-N(5)-Co(1) 127.2(2)
C(24)-N(6)-C(21) 105.8(3)
C(21)-N(6)-Co(1) 127.0(2)
C(6)-N(3)-C(9)
105.8(3)
C(9)-N(3)-Co(1) 126.9(2)
C(11)-N(4)-Co(1) 126.6(3)
C(16)-N(5)-C(19) 105.7(3)
C(19)-N(5)-Co(1) 126.9(2)
C(24)-N(6)-Co(1) 127.1(2)
N(2)-C(1)-C(2)
121.1(4)
values are listed in Table S1 in the Supporting Information.
The slight variation of these positions shows trends with the
Hammett parameters determined for these pyridine derivatives.
(Cl2py)CoTPP(NO2). The pyridyl derivative of CoTPP(NO2)
prepared with the most weakly coordinating pyridine ligand,
(Cl2py)CoTPP(NO2), also crystallized in the monoclinic P21/n
space group with four porphyrin molecules per unit cell. The
unit cell dimensions were a ) 13.3698(9) Å, b ) 23.216(3) Å,
and c ) 15.130(3) Å; â ) 105.613(3)°. The crystallographic
data is summarized in Table 2. The molecular structure of the
porphyrin is illustrated in the ORTEP diagram at the 30%
probability level in Figure 2. Bond lengths and angles are in
Table 3B. Nitro ligand bond lengths and angles are compared
with those determined by semiempirical modeling in Table 1.
Spectroscopy. The infrared spectra of (py)CoTPP(NO2) and
of its reduced analogue CoTPP(NO) have been used to monitor
the formation and to assess the nature of these compounds in
reaction studies.2a In the series of pyridine derivatives, only
slight differences in the positions of the characteristic bands of
the nitro ligand νaNO , νsNO and δO-N-O were observed. These
The electronic spectrum of (nitro)(pyridyl)cobalt(III) por-
phyrin has not been reported previously. The position of the
Soret peaks of the pyridyl derivatives was also found to vary
slightly with the basicity of the pyridine ligand according to
Hammett parameters as reported in Table S1. These values
ranged from 432 nm for the least basic pyridine derivative, Cl2-
py, to 438 nm for the most basic derivative, Me2Npy. The
similarity of the position of the Soret peaks for CoTPP (411
nm), CoTPP(NO) (414 nm),8 and the proposed pentacoordinate
species, CoTPP(NO2) (417 nm), led to confusion about the
identity of this latter species. Recent thin-film experiments by
Kurtikyan et al. have provided evidence for the formation of
CoTPP(NO2) in the solid state by two methods: (1) direct
(8) Richter-Addo, G. B.; Hodge, S. J.; Yi, G. -B.; Khan, M. A.; Ma, T.;
Van Caemelbecke, E.; Guo, N.; Kadish, K. M. Inorg. Chem. 1996,
35, 6530-6538.
2
2

(Nitro)(tetraphenylporphinato)cobalt(III) Derivatives
Inorganic Chemistry, Vol. 40, No. 17, 2001 4221
Figure 4. Electronic spectra of partially dissociated (Cl2py)CoTPP-
(NO2) (‚‚‚) in CH2Cl2 under nitrogen and the same solution after
addition of solid LiClO4 (saturated solution) (s). Spectra were recorded
at times of 1 and 3 min. The solution was then saturated with Cl2py
and immediately converted to the spectrum of fully associated (Cl2-
py)CoTPP(NO2) (- - - ).
SPECDEC program. The value for the association of the
pyridine ligand, found by the same method, was based upon
the spectra shown in Figure S1 in the Supporting Information.
The value of the equilibrium constant for the pyridine ligand
was found to be 1.08 × 108 with ø2 ) 1.29 × 10-4. None of
the other pyridyl derivatives show the higher energy band
attributed to CoTPP(NO2) even at low concentration, so no
attempts were made to measure their stability constants. This
behavior is consistent with their greater basicities as reflected
in their Hammett values.
Figure 2. ORTEP plot of (Cl2py)CoTPP(NO2) at the 30% probability
level.
The reactions with the lithium ion with (Cl2py)CoTPP(NO2)
in CH2Cl2 was followed by UV-visible spectroscopy as shown
in Figure 4. As noted previously, at solution concentrations in
the 1 × 10-5 M range, the Cl2py ligand is largely dissociated.
The figure shows the peaks at 431 nm assigned to the associated
(Cl2py)CoTPP(NO2) and the peak at 417 nm assigned to CoTPP-
(NO2). Upon addition of lithium perchlorate, the Cl2py is lost
from the complex within 10 min as shown in the figure. Addition
of a large excess of Cl2py to the same solution results in
complete re-formation of the complex instantly as shown by
the single peak at 432 nm represented by the dashed line.
Electrochemistry. Cyclic voltammetry of the series of
derivatives is more sensitive to the variation of the pyridine
ligand coordination equilibrium than the static structural results
given by crystal structures or spectroscopy. The cyclic voltam-
mogram for the most stable derivative, (Me2Npy)CoTPP(NO2),
is shown in Figure S2 in the Supporting Information and
provides three clear quasi-reversible couples. The half-wave
potentials of all the derivatives reported in volts vs the Ag/
AgNO3(CH3CN) reference electrode are listed in Table 4. Upon
addition of excess LiClO4 to the solution of (Cl2py)CoTPP-
(NO2), the pattern of three quasi-reversible processes quickly
changes to a single irreversible oxidation within the range 0.4-
2.0 V as shown in Figure S3 in the Supporting Information.
Reactions. In the reaction of (Cl2py)CoTPP(NO2) with Ph3P
(50 mM) with added LiClO4, the electronic spectra in Figure
S4 (Supporting Information) show rapid conversion to three
products with new Soret peak positions at 414, 447, and 467
nm. The peak position at 414 nm is again ambiguous, but may
be attributed to the CoTPP(NO) species since its formation may
be followed by IR spectroscopy. The other peaks at 447 and
467 nm appear on the same rapid time scale. Further addition
of Ph3P increases the intensity of the peak at 467 while
decreasing the intensity of the peak at 447 nm. These two peaks
are proposed to be a (triphenylphosphino)(nitro)cobalt(III)
complex, (Ph3P)CoTPP(NO2), and the bis(triphenylphosphino)-
cobalt(III) complex, respectively.
Figure 3. Electronic spectra of solutions of partially dissociated
(pyridyl)(nitro)cobalt(III) porphyrin complexes in CH2Cl2 and during
titration of excess pyridine ligand, used in determination of equilibrium
constants for pyridine association: (Cl2py)CoTPP(NO2) (8.78 × 10-6
M with Cl2py concentrations ranging from 8.78 × 10-6 M to 2.5 ×
10-5 M).
reaction of NO2 gas with sublimed CoTPP9 and reaction of O2
gas with CoTPP(NO)10 prepared in the solid state by reaction
of NO gas with CoTPP. The reported electronic spectrum that
has its Soret position at ∼422 nm is similar to our observations
of the CoTPP(NO2) in CH2Cl2 solution.
Coordination Equilibria. At low concentrations (∼10-6
M
in CH2Cl2), both (Cl2py)CoTPP(NO2) and (py)CoTPP(NO2)
derivatives show two Soret peaks, even after careful purification
of the porphyrins by column chromatography. Initially the
second, higher energy band in the 414-417 nm region was
assigned as contaminant CoTPP (411 nm) or CoTPP(NO) (414
nm) arising from incomplete reaction or decomposition of the
nitro derivative. Figure 3 shows the electronic spectrum of this
equilibrium mixture generated by dissolving (Cl2py)CoTPP-
(NO2) in CH2Cl2 under nitrogen atmosphere, and the resulting
spectra as Cl2py is titrated into the solution, the porphyrin
concentration being held constant. The isosbestic behavior of
these spectra implies that a clean conversion from a proposed
pentacoordinate CoTPP(NO2) species to (Cl2py)CoTPP(NO2)
is observed during these additions. The value of the equilibrium
constant at 25 °C for association of the Cl2py ligand was 6.94
× 106 with a value of ø2 ) 7.2 × 10-4 reported by the
(9) Kurtikyan, T. S.; Stepanyan, T. G.; Gasparyan, A. V. Russ. J. Coord.
Chem. 1997, 23, 563-567.
(10) Kurtikyan, T. S. Russ. J. Coord. Chem. 1999, 25, 28-30.
Electronic spectra collected over the course of the reaction
of triphenylphosphine (2 mM) with (Me2Npy)CoTPP(NO2) (1

4222 Inorganic Chemistry, Vol. 40, No. 17, 2001
Goodwin et al.
Table 4. Electrochemical and Stability Constant Effects of the Distal Pyridine Ligand
a
E1/2
E′1/2
[E′1/2
(∆Ep)
E′′1/2
[E′′1/2]
(∆Ep)
b
Hammett
params
[E1/2
]
]
(∆Ep)c
K
(Cl2py)CoTPP(NO2)
(py)CoTPP(NO2)d
σm ) +0.64
0.71
1.07
1.28
6.94 × 106
[0.95]
(120)
0.69
[0.93]
(180)
0.67
[0.91]
(140)
0.63
[1.31]
(190)
0.90
[1.14]
(160)
1.01
[1.25]
(220)
0.92
[1.16]
(160)
[1.52]
(160)
1.21
[1.45]
(220)
1.27
[1.51]
(220)
1.11
[1.35]
(100)
σp ) 0
1.08 × 108
(3,5-lut)CoTPP(NO2)
(Me2Npy)CoTPP(NO2)
σm ) -0.10
σp ) -0.80
h
h
[0.87]
(100)
1.24e
[1.48]
“CoTPP(NO2)” d
CoTPP(NO)
NA
NA
E
1/2(+/0)f,g
E1/2(+/2+)f,g
1.37f
1.00f
1.01g
1.25g
a All values determined in 0.10 M TBAH, reported in volts vs Ag/AgNO3 in CH3CN. Ferrocene as an internal standard gave an E1/2 of 0.240 V
with this electrode. ∆Ep for this reference peak was found to be 160 mV. No corrections were made for junction potentials. All measurements made
on compounds without added pyridine ligand, except for Cl2py, with 0.1 M ligand added. b Values corrected to SCE. c ∆Ep values given in mV.
d Obtained by reacting the solution of (Cl2py)CoTPP(NO2) with saturated LiClO4. e Irreversible oxidation peak potential. f Values and assignments
reported by Kini et al.16 g Values and assignments reported by Richter-Addo et al.8 h Coordination equilibria were not observable.
× 10-5 M) with added Me2Npy (50 mM) also show formation
of peaks assignable to (Ph3P)CoTPP(NO2) and CoTPP(NO).
Direct comparison of oxo-transfer reactivity with triphenylphos-
phine is complicated by the substitution chemistry with tri-
phenylphosphine.
2-oxopropionic acid, and 1,2-dibromocyclohexane. A similar
range of products was obtained for the reaction carried out in
the dark. The products are illustrated in Table 5.
Reactions that were carried out with silica gel adsorbed
CoTPP(NO2) and neat cyclohexene gave much simpler product
distributions with roughly the same rate of loss of reactant.
Regardless of the method of synthesis of the catalyst, the
products obtained were cyclohexen-2-ol, cyclohexen-2-one,
cyclohexene oxide, and bi-2-cyclohexen-1-yl, with appearance
of the last species occurring only after several days of reaction.
The proportions of products in the first 2 days showed that the
sum of the cyclohexene oxide and cyclohexen-2-one was
roughly equal to the total amount of the alcohol produced.
Catalytic oxidation products obtained with silica gel adsorbed
CoTPP(NO) from neat cyclohexene and O2 were the same as
those obtained from the adsorbed (nitro)cobalt porphryin
catalysts.
The (Cl2py)CoTpivPP(NO2)-catalyzed oxygenation reactions
with cyclohexene and allyl bromide provided product distribu-
tions similar to those obtained with the unhindered porphyrin
catalysts. However, for the reaction with styrene, Virtually no
reactiVity was obserVed for the hindered (dichloropyridyl)-
(nitro)cobalt(III) picket-fence porphryin.
The electronic spectra showing the reaction of (Cl2py)CoTPP-
(NO2) with allyl bromide (0.5 M) are shown in Figure S5 in
the Supporting Information. In this reaction in CH2Cl2 under
N2, no additional lithium perchlorate was added. The rapid
partial formation of the peak at 414 nm is consistent with the
formation of CoTPP(NO) from free CoTPP(NO2) with a slower
reaction taking place as Cl2py dissociates from the complex
leaving the reactive CoTPP(NO2) available for rapid oxidation
of the allyl bromide. A similar procedure was used to compare
the oxo-transfer reactivity of the fully complexed (Me2Npy)-
CoTPP(NO2) with allyl bromide. Under anaerobic conditions
and with excess Me2Npy (0.010 M) in solution, there is no
detectable reaction of (Me2Npy)CoTPP(NO2) with allyl bromide
in dichloromethane solution monitored over the course of 3 days.
Reactions under oxygen atmosphere have qualitatively shown
a wide range of catalytic oxidation reactions, depending upon
the conditions of the experiment as summarized in Table 5.
Room temperature catalytic oxidation of a large excess of
cyclohexene (1 mL added to 1 mL of 1 mM [(Cl2py)CoTPP-
(NO2)] in 1 mL of CH2Cl2) exposed to fluorescent room lighting
produced 2-cyclohexen-1-ol, 3-cyclohexen-1-ol, 1,2-cyclohex-
anediol, cyclohexene oxide, 2-cyclohexen-1-one, 1,2-cyclohex-
anedione, and bi-2-cyclohexen-1-yl after 1 day with roughly
half of the cyclohexene remaining, corresponding to a catalytic
turnover of ∼6000 (cyclohexene/cobalt) and a TOF of ∼5
cyclohexene‚cobalt-1‚min-1. Roughly the same results were
observed with this reaction carried out in the dark. Oxidation
of styrene with light exposure under the same reaction conditions
gives benzaldehyde, styrene oxide, R-methylstyrene, benzene
acetaldehyde, phenylethyl alcohol, 1-phenyl-1,2-ethanediol, and
benzoylformic acid. A similar distribution of products was
observed for the reaction carried out in the dark. Oxidation
reactions of allyl bromide in the light under similar conditions
produced a wide range of products including 1,3-dibromopro-
panol, 2,3-dimethyl-2,3-butanediol, the ethyl ester of 3-bromo-
Discussion Section
Table 1 summarizes the predicted structural effects of the
distal pyridine ligands on the nitro ligand. The variation in the
basicity of the distal ligand results in only small changes in the
N-O bond lengths and O-N-O bond angles. Exceptions
appear where the very weak Lewis base pentafluoropyridine or
the weakly coordinating perchlorate ion are used as the distal
ligand. Even so, a correlation of these small changes with
Hammett parameters for the pyridine ligands is evident from
examination of the predictions.
The two crystal structures of this study are the fifth and sixth
structures of (nitro)cobalt(III) porphyrins to be reported. The
cobalt coordination site and nitro ligand bond angles and
distances of two early structures of piperidine and 3,5-lutidine
derivatives of CoTPP(NO2) and of 1-methylimidazole and 1,2-
dimethylimidazole derivatives of CoTpivPP(NO2) are listed

(Nitro)(tetraphenylporphinato)cobalt(III) Derivatives
Inorganic Chemistry, Vol. 40, No. 17, 2001 4223
Table 5. Products of Alkene Oxidation by O2 as Catalyzed by (Cl2py)CoTPP(NO2) and (Cl2py)CoTpivPP(NO2)a
a Determined by GC-MS analysis of solutions after 24 h at room temperature.
along with those obtained here in Table 1. The nitro bond angles
and distances for the lutidine and piperidine structures are
noticeably smaller than those of any of the other derivatives.
But problems associated with the ligand disorder, different
temperatures of data collection, and low-resolution data col-
lection make detailed comparison of nitro bond angles and
lengths unreliable.
those of the pentacoordinate and the other hexacoordinate
pyridyl derivative values as shown in the table. For electronic
spectra, only the measured Soret peak positions were compared.
These show variation from 432 nm for the weakest base (Cl2-
py) to 438 nm for the strongest base (Me2Npy). The loss of the
sixth coordination site results in a shift of the Soret peak to
417 nm. It is noteworthy that, when CoTPP(NO2) is deposited
onto silica gel surfaces by a variety of methods, the consistently
observed Soret position of 428 nm is between the values of the
weakest pyridine base (432 nm) and the pentacoordinate species
in solution (417 nm). A weak bonding interaction with the silica
gel may give rise to this effect. Soret peak positions of 422 nm
for the pentacoordinate nitro complex observed in its preparation
by Kurtikyan are also between the Soret positions observed for
the Cl2py complex and the solution phase CoTPP(NO2) species.
The pentacoordinate nitro complex demonstrates much greater
secondary oxo-transfer reactivity than has been reported earlier.
The existence of this species has been proposed earlier by
Andrews and in the solid state by Kurtikyan, but its reactivity
has not been recognized. In the presence of lithium perchlorate
in large excess, it is conceivable that the perchlorate ion is
weakly coordinated with the metal center as we observed in
our study of the (nitro)iron(III) picket-fence porphyrin ana-
logue.11 Oxidation reactions using activation by lithium per-
chlorate as a cocatalyst were originally proposed to take place
by an interaction with nitrite ligand through an NO2 oxygen
atom.2c But the spectral changes represented in Figure 4 more
strongly support pyridine removal in the presence of Li+ and
addition of the pyridine ligand when it is added in excess.
“Quenching” of oxo-transfer reactions by addition of excess
There are also minor trends in the observed spectroscopic
measurements that correlate with predictions from semiempirical
calculations as a function of ligands’ basicities. Nitro ligand
asymmetrical vibrational frequencies, νa, vary only slightly over
the range of 1424-1427 cm-1 according to the order of
Hammett parameters. The calculated vibrational frequencies,
while a poor match for the observed frequencies due to the
approximations of the PM3 semiempirical model, are consistent
with these small changes, ranging from 1812.58 to 1810.43
cm-1, Andrews has reported the IR spectrum of the pentaco-
ordinate species in a mixture with the acetonitrile complex. The
asymmetrical stretching frequency, νa, of NO2 was reported to
be 1453 cm-1 compared to the 1424-1427 cm-1observed for
the hexacoordinate species in this work. The calculated value
for the pentacoordinate complex was determined to be 1856
cm-1 by the semiempirical model consistent in the observed
increase in vibrational energy. Both calculated and observed
symmetrical stretching frequencies show a decrease in vibra-
tional energies with the loss of the pyridine ligand. The observed
values for the pyridine complexes range from 1307 to 1310
cm-1, but without the pyridine ligand, the value is 1283 cm-1
.
The semiempirical model predicts a range of 1701.04 to 1702.65
cm-1 for the pyridine complexes and 1673.79 cm-1 for the
respective symmetric vibrational frequencies. Vibrational fre-
quency values for the hypothetical pentafluoropyridyl and
perchlorato derivatives were also determined and are between
(11) Frangione, M.; Fangione, M.; Port, J.; Baldiwala, M.; Judd, A.; Galley,
J.; DaVega, M.; Linna, K.; Caron, L.; Anderson, E.; Goodwin, J. Inorg.
Chem. 1997, 36, 1904-1911.

4224 Inorganic Chemistry, Vol. 40, No. 17, 2001
Goodwin et al.
pyridine has been discussed earlier for the pyridine complexes.6
We now interpret this effect as a result of promoting the
formation of the unreactive hexacoordinate (py)CoTPP(NO2)
species and not as a result of formation of (py)2CoTPP. For the
reactions with triphenylphosphine, it is apparent that, with a
reducing agent this strong, there is little difference in the
reactivity of the hexacoordinated nitro complexes, and that
ligand substitution by the triphenylphosphine ligand is a faster
process than oxo transfer. Even in the presence of excess (Me2-
Npy) ligand, both ligand substitution and the eventual formation
of the CoTPP(NO) product were observed.
The differences in reactivity of the pentacoordinate and
hexacoordinate species are most evident in the reactions with
alkenes. The Me2Npy complex has been found not to react with
any of the alkenes investigated in this project, even without
additional Me2Npy in CH2Cl2 solution. The equilibrium loss
of the pyridine ligand indicated by the coordination equilibrium
experiments can be promoted by the addition of the lithium ion
which competes for the pyridine ligand. In using more strongly
coordinating pyridine and higher concentrations of complexes,
the reactions were not observed until the pentacoordinate species
was generated by adding lithium perchlorate.
Since pyridine derivatives of (nitro)cobalt(III) porphyrin have
been commonly prepared by O2 oxidation of (nitrosyl)cobalt(II)
pophyrins in the presence of the nitrogenous base, the inter-
mediacy of the (nitrosyl)cobalt(II) species in the catalysis is
suggested. Our observation that the same catalytic oxidation
reactions occur for cyclohexene with oxygen with the silica-
adsorbed CoTPP(NO) species strongly supports this species as
an intermediate and indeed that the preparation of these solid
phase catalysts could be greatly simplified with reactions of
NO gas with solid adsorbed Co(II) porphyrins. Further mecha-
nistic interpretation of this result follows later in this section.
In removal of the pyridine ligand by addition of Li+, the
electrochemical behavior changes dramatically. Previous reports
of the electrochemistry of CoTPP(NO) complexes suggests that
the metal-centered oxidation for both series occurs at the extreme
positive potential observed. The pyridine ligand dependence of
all of the reversible couples listed in Table 5 is emphasized by
the trend in all the half-wave potentials and the complete change
in the electrochemical behavior upon the removal of the ligand.
The identity of the irreversible oxidation wave is unknown at
this time, but according to the trends expected with the basicity
of the ligands, the potentials of all of the couples will be shifted
to positive extremes. Assuming that this is the case, then the
observed irreversible process for CoTPP(NO2) has been shifted
by nearly +0.5 V by the loss of the Cl2py ligand that associates
with an equilibrium constant of 6.94 × 106.
Figure 5. Proposed catalytic cycle for the oxidation of alkenes
involving both bimolecular nonradical and radical oxo-transfer steps,
through the formation of a peroxynitro intermediate that arises from
the interaction of dioxygen with the reduced CoTPP(NO) form.
pentacoordinate nitro derivative, CoTPP(NO2), dioxygen, and
ozone. These results show a 114 kcal/mol thermodynamic
advantage in the formation of the products CoTPP(NO2) and
ozone from dioxygen and the peroxynitro derivative. Related
to this is Kurtikyan’s observation11 of the conversion of
amorphous solid state CoTPP(NO) to CoTPP(NO2) by reaction
with gas phase O2. Although those authors did not describe a
mechanism for the reaction of the immobilized complex with
dioxygen, the formation of a peroxynitro intermediate and its
decomposition by further reaction with O2 to form ozone and
the product nitro complex is a rather simple explanation.
The dramatic difference in reactivity between styrene and the
other alkenes with oxygen in the presence of (Cl2py)CoTpivPP-
(NO2) suggests that bimolecular oxo transfer is involved in the
initial oxo-transfer step, since this is limited by the steric
interactions of the picket-fence structure and the styrene
molecule. The nitro ligand preferentially coordinates in the
sterically hindered “pocket” and is only available for bimolecular
oxo transfer with small molecules. A bimolecular oxo transfer
has been invoked in nonradical mechanisms of epoxidation.
Therefore, two distinct oxo donors are envisioned in the
catalytic cycle, consistent with the mechanism proposed by
Basolo for the reaction of O2 with cobalt nitrosyls to form nitro
complexes. This mechanism includes the formation of the
intermediate peroxynitro species which, in the absence of
reducing agents, reacts with other nitrosyl complexes in a
comproportionation reaction to form two molecules of the nitro
complex.12 If the initial oxo-transfer reaction from the nitro
derivative occurs by a bimolecular mechanism, as suggested
by the lack of reactivity between styrene and CoTpivPP(NO2),
then the further progress of the reaction by a radical pathway,
not having the same steric constraints, would be ascribable to
the proposed peroxynitro intermediate.
The extreme shift in oxidation potential confers a much
greater reactivity toward secondary oxo transfer as observed in
the oxidation of alkenes that were previously reported to require
“activation” by thallium or palladium cocatalysts, or simply nitro
exchange to the palladium cocatalyst. The formation of alcohols
by a C-H insertion and the coupled bi-2-cyclohexen-1-yl
product suggests formation of O radicals as part of the catalytic
oxidation mechanism.
The proposed mechanism involving the two distinct oxo-
transfer steps is depicted in Figure 5. The radical mechanism
could be promoted by interaction of excess dioxygen with the
peroxynitro intermediate to form ozone. In Castro’s study of
ClFeTPP(NO2), the formation of ozone from O2 was confirmed
as a route to the ensuing radical oxidation. Our semiempirical
models included calculation of the heats of formation for the
pentacoordinate peroxynitro derivative, CoTPP(NO,O2), the
The decomposition of peroxynitrite by methemoglobin is
currently of interest and related to the generation of oxygen
radicals and catalytic regeneration of the nitro species in this
study. Several groups of authors13-15 have debated the decom-
position mechanism based upon UV-visible and EPR data and
(12) Clarkson, S. G.; Basolo, F. Inorg. Chem. 1973, 12, 1528-1534.
(13) Kahn, A. U.; Koviacic, D.; Kolbanovsky, A.; Desai, J.; Frenkel, K.;
Geacintov, N. E. Proc. Natl. Acad. Sci. U.S.A. 2000, 97, 2984-2989.
(14) Merenyi, G.; Link, J.; Czapski, G.; Goldstein, S. Proc. Natl. Acad.
Sci. U.S.A. 2000, 97, 8216.
(15) Martinez, G. R.; Di Mascio, P.; Bonini, M. G.; Augusto, O.; Briviba,
K.; Sies, H.; Maurer, P.; Rothlisberger, U.; Herold, S.; Kippenol, W.
H. Proc. Natl. Acad. Sci. U.S.A. 2000, 97, 10307.

(Nitro)(tetraphenylporphinato)cobalt(III) Derivatives
Inorganic Chemistry, Vol. 40, No. 17, 2001 4225
thermochemical and quantum mechanical calculations. Khan et
al. reported formation of nitrosylhemoglobin through interaction
of methemoglobin and peroxynitrite, suggesting that, upon
protonation, peroxynitrite produced singlet oxygen and the
nitroxyl anion that went on to react with methemoglobin. Other
authors14 countered this argument, claiming instead that forma-
tion of NO2 and OH radicals was thermodynamically favored,
and that the observation of the nitrosyl complex was in error.
Our experiments suggest that coordination of peroxynitrite by
an oxidized metal center may indeed promote generation of oxo
radicals and a reactive reduced nitro complex that can easily
form a nitrosyl product, consistent with all of these observations.
Increased reactivity toward oxo transfer in the iron analogue
FeTPP(NO2) makes these compounds susceptible to dispropor-
tionation. Figures 3 and S1 show the apparent stability of the
equilibrium mixtures of CoTPP(NO2) with (py)CoTPP(NO2) and
(Cl2py)CoTPP(NO2) without excess pyridine ligand. The spec-
trum of the original equilibrium mixture in each figure was
unchanged over a 2-week period. Addition of excess pyridine
or dichloropyridine to the equilibrium mixtures after this period
of time resulted in generation of the spectra attributable to the
fully coordinated derivatives.
transfer is lost with apparent generation of oxo radicals by way
of the proposed peroxynitro complex intermediate that forms
from reaction of dioxygen and the (nitrosyl)cobalt(II) porphyrin.
Further study of this intermediate and its involvement in the
overall oxo-transfer reactivity of the pentacoordinate cobalt(III)
porphyrins is being pursued.
Acknowledgment. This work was funded in part by the
Horry County Higher Education commission’s support of
Coastal Carolina University’s Academic Enhancement Grants
Program. Funding was also provided by the Southeast Regional
Education Board for faculty (J.G.) travel to Clemson University,
the South Carolina Governor’s School for Science and Math-
ematics for support of high school research interns, and the
Camille and Henry Dreyfus Foundation (W.P.) for support of
the X-ray Crystallography workshop at Clemson University.
Funding was also provided by the SC EPSCoR program for
summer research support (V.E. and J.T.) and by the Cathcart
Smith Foundation and the Department of Chemistry and Physics
of Coastal Carolina University. Thanks are due to David Evans
(CCU) for valuable conversations about organic reaction
chemistry.
Conclusions
Supporting Information Available: Crystallographic data in CIF
format. Cyclic voltammograms of (Me2Npy)CoTPP(NO2), (Cl2py)-
CoTPP(NO2), and “CoTPP(NO2)” prepared by adding LiClO4 to the
latter solution (Figures S1 and S2). Electronic spectra of (Cl2py)CoTPP-
(NO2) in CH2Cl2 under nitrogen after additions of solid Ph3P (Figure
S3). Electronic spectra of (Cl2py)CoTPP(NO2) in CH2Cl2 under nitrogen
after addition of allyl bromide (Figure S4). Spectroscopic effects of
the distal ligand (Table S1). This material is available free of charge
via the Internet at http://pubs.acs.org.
The similarity of structures and electrochemical behavior
suggests that the hexacoordinate pyridine derivatives share a
similar lack of reactivity toward oxo-transfer reactions, with
differences attributable to the extent of the pyridine ligand
dissociation. Oxidation strength increases substantially with
removal of the pyridine ligand, although specificity of the oxo
(16) Kini, A. D.; Washington, J.; Kubiak, C. P.; Morimoto, B. H. Inorg.
Chem. 1996, 35, 6904-6906.
IC001442E