X-ray structures and homolysis of some alkylcobalt(III) phthalocyanine complexes

Article abstract of DOI:10.1021/ic051078p

The first crystallographic data for σ-bonded alkylcobalt(III) phthalocyanine complexes are reported. A single-crystal X-ray structure of CH₃CH₂Co^IIIPc (Pc = dianion of phthalocyanine) reveals that the solid consists of centrosymmetric face-to-face dimers in which the CH₃CH₂Co^IIIPc units retain their square pyramidal geometry. The structure appears to be the first one reported for a five-coordinate RCo(III)(chelate) complex with an electron-deficient equatorial system. The Co-C bond in CH₃CH₂Co^IIIPc (2.031(5) A) is the longest found in five-coordinate RCo ^III(chel) complexes (R = simple primary alkyl group). Another X-ray study demonstrates that CH₃Co^IIIPc(py) has a distorted octahedral geometry with axial bonds of very similar length to those in methylcobalamin. The axial bonds are shorter than those in its octaethylporphyrin analogue, in accordance with a weaker trans axial influence in six-coordinate complexes containing an electron-deficient phthalocyanine equatorial ligand. A different trend has been observed for five-coordinate RCo^III(chel) complexes: electron-rich equatorial systems seem to make the Co-C axial bond shorter. Kinetic data for the homolysis of RCo ^IIIPc complexes (R = Me, Et) in dimethylacetamide are also reported. Homolysis of ethyl derivatives is faster. The Co-C bond dissociation energies (BDEs) for the pyridine adducts of the methyl and the ethyl derivative are 30 ± 1 and 29 ± 1 kcal/mol, respectively. The BDE for CH ₃CoPc(py) is considerably lower than that for MeCbl despite the very similar lengths of the axial bonds in the two complexes. The results of this work do not support any correlation between the Co-C bond length and the bond strength as defined by BDE.

Full text of DOI:10.1021/ic051078p

Inorg. Chem. 2005, 44, 9902−9913
X-ray Structures and Homolysis of Some Alkylcobalt(III) Phthalocyanine
Complexes
Wlodzimierz Galezowski* and Maciej Kubicki
Department of Chemistry, Adam Mickiewicz UniVersity, Grunwaldzka 6, 60-780 Poznan, Poland
Received June 29, 2005
The first crystallographic data for
σ
-bonded alkylcobalt(III) phthalocyanine complexes are reported. A single-crystal
dianion of phthalocyanine) reveals that the solid consists of centrosymmetric
X-ray structure of CH3CH2CoIIIPc (Pc
)
face-to-face dimers in which the CH3CH2CoIIIPc units retain their square pyramidal geometry. The structure appears
to be the first one reported for a five-coordinate RCo(III)(chelate) complex with an electron-deficient equatorial
system. The Co
(R
simple primary alkyl group). Another X-ray study demonstrates that CH3CoIIIPc(py) has a distorted octahedral
−
C bond in CH3CH2CoIIIPc (2.031(5) Å) is the longest found in five-coordinate RCoIII(chel) complexes
)
geometry with axial bonds of very similar length to those in methylcobalamin. The axial bonds are shorter than
those in its octaethylporphyrin analogue, in accordance with a weaker trans axial influence in six-coordinate complexes
containing an electron-deficient phthalocyanine equatorial ligand. A different trend has been observed for five-
coordinate RCoIII(chel) complexes: electron-rich equatorial systems seem to make the Co
Kinetic data for the homolysis of RCoIIIPc complexes (R
Me, Et) in dimethylacetamide are also reported. Homolysis
of ethyl derivatives is faster. The Co C bond dissociation energies (BDEs) for the pyridine adducts of the methyl
and the ethyl derivative are 30 1 and 29 1 kcal/mol, respectively. The BDE for CH3CoPc(py) is considerably
lower than that for MeCbl despite the very similar lengths of the axial bonds in the two complexes. The results of
−C axial bond shorter.
)
−
±
±
this work do not support any correlation between the Co
−C bond length and the bond strength as defined by BDE.
Introduction
none of these has as an electron-deficient8 equatorial ligand.
This work provides the first structural data for such complex.
As demonstrated by recent equilibrium studies of trans axial
ligation, also by cyclic voltammetry, alkyl-CoIIIPc complexes
clearly can be categorized as electron-poor models of
methylcobalamin.9 The electrophilic nature of the five-
coordinate CH3CoIIIPc species has been the rationale behind
its exceptional SN2 methyl-transfer reactivity. Consequently,
the considerable Lewis acidic strength of the Co center seems
to leave little hope for obtaining a solid consisting of the
five-coordinate units, especially since reasonable solubility
of these compounds can only be attained in binding solvents.
However, phthalocyanines are known for their strong pro-
pensity for π-stacking, which may prevent trans axial ligation
in the solid.10 The geometry of the coordination sphere in
Very few five-coordinate alkylcobalt(III) complexes have
been structurally characterized so far.2-7 Not surprisingly,
* To whom correspondence should be addressed. E-mail: wlodgal@
amu.edu.pl.
(1) Abbreviations: Pc, dianion of phthalocyanine; MeCbi+, methylcobi-
namide; MeCbl, methylcobalamin; OEP, dianion of 2,3,7,8,12,13,-
17,18-octaethylporphyrin; NPc, dianion of naphthalocyanine; TPP,
dianion of 5,10,15,20-tetraphenylporphyrin; TTP, dianion of 5,10,-
15,20-tetratolylporphyrin; DH, monoanion of dimethylglyoxime; (DO)-
(DOH)pn, anion of N2,N2′-propanediylbis(2,3-butanedione 2-imine-
3-oxime); salen, dianion of bis(salicydene)ethylenediamine; saloph,
dianion of bis(salicylidene)-o-phenylenediamine; TC, tropocoronand;
acacen, dianion of bis(acetylacetone)ethylenediimine; py, pyridine;
N-MeIm, N-methylimidazole; DMA, dimethylacetamide; DMSO,
dimethyl sulfoxide; TEMPO, 2,2,6,6-tetramethyl-1-piperidinyloxy.
(2) Bru¨ckner, S.; Calligaris, M.; Nardin, G.; Randaccio, L. Inorg Chim.
Acta 1969, 3, 308.
(3) Summers, M. F.; Marzilli, L. G.; Bresciani-Pahor, N.; Randaccio, L.
J. Am. Chem. Soc. 1984, 106, 4478.
(4) Parker, W. O., Jr.; Bresciani-Pahor, N.; Zangrando, E.; Randaccio,
L.; Marzilli, L. G. Inorg. Chem. 1985, 24, 3908.
(5) Jaynes, B. S.; Ren, T.; Masschelein, A.; Lippard, S. J. Am. Chem.
Soc. 1993, 115, 5589.
(6) Summers, J. S.; Petersen, J. L.; Stolzenberg, A. M. J. Am. Chem. Soc.
1994, 111, 7189.
(7) Cao, Y.; Stolzenberg, A. M.; Petersen, J. L. Inorg. Chim. Acta 1997,
263, 139.
(8) Cobalamins seem to be a natural reference point here.
(9) Galezowski, W. Inorg. Chem. 2005, 44, 1530.
(10) In solution, even in nonbinding solvents, the complexes are monomeric
(see ref 9).
9902 Inorganic Chemistry, Vol. 44, No. 26, 2005
10.1021/ic051078p CCC: $30.25
© 2005 American Chemical Society
Published on Web 11/10/2005

Alkylcobalt(III) Phthalocyanine Complexes
five-coordinate (or pseudo-five-coordinate) RCoIIIPc units in
a face-to-face dimer is of interest with regard to the SN2
reactivity of alkylcobalt(III) phthalocyanine complexes.
balamine chemistry being an obvious reason. Excessively
large bond dissociation energies (BDEs) and the resulting
inconveniently slow rates may have been another reason
for the scarcity of data on the homolytic cleavage of the
Co-(primary alkyl) systems, with simple primary alkyls such
as methyl or ethyl. In addition, â-hydrogen-containing alkyls,
such as ethyl, could have been avoided because of possible
in-cage â-H elimination. For the methyl-cobalt system, apart
from BDE values for the homolysis of MeCbl18 and its
popular model, CH3Co(DH)2(py),19 and some kinetic data
on CH3Co[TC-(4,4)],5 there is probably only one report on
the homolysis of σ-bonded methylcobalt(III) porphyrinic
model20 and one regarding the thermolysis of CH3CoIIIPc
under aerobic conditions.21
The structural data for σ-bonded alkyl-Co(III) octaeth-
ylporphyrins reported by Stolzenberg et al.6,7 set the stage
for interesting comparisons with their phthalocyanine con-
geners. While phthalocyanine complexes have the most
electron-deficient equatorial arrangement among porphyrinic
models, the octaalkylporphyrin compounds are just the
opposite. Octaalkylporphyrins (H2P) are ∼15 pKa units less
acidic than phthalocyanines.11 Another striking feature of
phthalocyanine complexes, consistent with their electron
deficiency, are their redox potentials, which are considerably
less negative than those for octaalkylporphyrins. For instance,
Co(II)/Co(I) potentials are -0.2 or -0.35 12 and -0.95 V13
vs SCE for phthalocyanine and octaethylporphyrin com-
pounds, respectively. Bond length and strength of the axial
system in alkylcobalt(III) complexes are a central problem
in bioinorganic chemistry. It appears that long Co-C bonds
coincide with a long trans Co-Nax bond. This unique feature
of the alkyl-Co-N axial system, known as inverse trans
influence, has recently gained much interest. Molecular
mechanics studies,14 as well as DFT calculations,15 aimed
at reproducing the structural data, and with less success, the
BDEs have been mostly conducted on corrinoids having
various axial ligands. The cis influence of the equatorial
system has received less attention. Riordan and Halpern noted
that, in closely related organoiron porphyrin complexes, the
metal-carbon bond dissociation is faster when the equatorial
ligand is less basic.16 Similar conclusions were derived by
Qiu and Sawyer from their electrochemical studies on iron
and cobalt porphyrins.17 No correlation of these observations
with the bond distances could be attempted because the
needed structural data are not available. A comparison of
the structural data for five-coordinate RCo(OEP) and RCoPc
complexes with equatorial ligands of sharply contrasting
basicity, as well as pyridine adducts with CH3CoPc and
CH3Co(OEP), should provide some insight into the scale of
the possible cis electronic influence.
A variety of alkylcobalt model complexes have been
synthesized and extensively studied. The phthalocyanine
model of MeCbl is unique in its outstanding methyl donor
ability. It is interesting to see if this atypical reactivity will
be reflected to any appreciable extent in the structural
parameters of the coordination sphere as well in the Co-C
bond dissociation energies. As follows from Marcus theory,
small BDEs should contribute to lowering the intrinsic free-
+
22
energy barrier for methyl (CH3 ) transfer.
Hence, a
comparison of the BDE values for inactive methyl donors,
such as methylcobaloxime or MeCbl, and reactive CH3CoPc
is of interest.
The alkyl group effect (Me > Et) has been demon-
strated to be a characteristic feature of the SN2 reactions of
RCoIIIPc complexes,9 in (anticipated) contrast to Et > Me,
for the homolytic cleavage. In this work, the faster homolysis
of CH3CH2CoIIIPc than that of CH3CoIIIPc is documented.
The homolysis data are also of interest because of the
possible correlation between structural parameters of the axial
system and the rates of homolysis. Rigid porphyrinic
models,23,24 with nonbulky axial alkyl groups, in which steric
trans influence is minimal and electronic effects can be more
easily observed, have not attracted as much attention as they
deserve. Further studies of these systems seem warranted.
Experimental Section
Materials. CH3CoIIIPc and CH3CH2CoIIIPc were prepared as
described earlier,9 and their solutions were handled at dim light.
All chemicals were purchased from Aldrich. Pyridine and toluene
were purified by standard methods. Double-distilled N-methylimi-
dazole was used as received. Dimethylacetamide (DMA) was
purified and handled as described elsewhere.9
Crystal Growth. X-ray quality crystals of CH3CH2CoIIIPc
were obtained by ethanol vapor infusion into a DMA solution of
the complex, which had been flash chromatographed on a SiO2-
packed column using DMA as the eluant. All manipulations were
Most of the literature data on the homolysis of the Co-C
bond relate to bulky axial alkyls, relevance to adenosylco-
(11) Stuzhin, P. A.; Khelevina, O. G.; Berezin, B. D. In Phthalocyanines.
Properties and Applications; Leznoff, C. C., Lever A. B. P., Eds.;
Wiley-VCH: New York, 1996; Vol. 4, Chapter 2.
(12) Lever, A. B. P.; Milaeva, E. R.; Speier, G. In Phthalocyanines.
Properties and Applications; Leznoff, C. C., Lever A. B. P., Eds.;
Wiley-VCH: New York, 1993; Vol. 3, Chapter 1, p 47.
(13) Zhou, D.-L.; Gao, J.; Rusling, J. F. J. Am. Chem. Soc. 1995, 117,
1127.
(14) Sirovatka, J. M.; Rappe´, A. K.; Finke, R. G. Inorg. Chim. Acta. 2000,
300-302, 545. (b) Brown, K. L.; Marques, H. M. J. Inorg. Biochem.
2001, 83, 121.
(15) Jensen, K. P.; Ryde, U. J. Phys. Chem. A 2003, 107, 7539. (b) Jensen,
K. P.; Ryde, U. THEOCHEM 2002, 585, 239. (c) Jensen, K. P.; Sauer,
S. P. A.; Liljefors, T.; Norrby, P.-O. Organometallics 2001, 20, 550.
(d) Friendorf, M.; Kozlowski, P. M. J. Am. Chem. Soc. 2004, 126,
1928. (e) Andriuniow, T.; Zgierski, M. Z.; Kozlowski, P. M. J. Am.
Chem. Soc. 2001, 123, 2679. (f) Andriuniow, T.; Zgierski, M. Z.;
Kozlowski, P. M. J. Phys. Chem. B 2000, 104, 10921. (g) Do¨lker, N.;
Maseras, F.; Lledo´s, A. J. Phys. Chem. B 2003, 107, 306.
(16) Riordan, C. G.; Halpern, J. Inorg. Chim. Acta 1996, 243, 19.
(17) Qiu, A.; Sawyer, D. T. J. Porphyrins Phthalocyanines 1997, 1, 125.
(18) Martin, B. D.; Finke, R. G. J. Am. Chem. Soc. 1992, 114, 585.
(19) Toscano, P. J.; Seligson, A. L.; Curran, M. T.; Skrobutt, A. T.;
Sonnenberger, D. C. Inorg. Chem. 1989, 28, 166.
(20) Fukuzumi, S.; Miyamoto, K.; Suenobu, T.; Van Caemelbecke, E.;
Kadish, K. M. J. Am. Chem. Soc. 1998, 120, 2880.
(21) Galezowski, W.; Ibrahim, P. N.; Lewis, E. S. J. Am. Chem. Soc. 1993,
115, 8660.
(22) See eq 12 in Marcus, R. A. J. Phys. Chem. A 1997, 101, 4072.
(23) Geno, M. K.; Halpern, J. J. Am. Chem. Soc. 1987, 109, 1238.
(24) Kaplan, W. A.; Scott, R. A.; Suslick, K. S. J. Am. Chem. Soc. 1990,
112, 1283.
Inorganic Chemistry, Vol. 44, No. 26, 2005 9903

Galezowski and Kubicki
Table 1. Yields (%) of the Products of Thermolysis of RCoIIIPc in
DMA
performed in dim light, but the presence of air was irrelevant.
CH3CoIIIPc(py) crystals were obtained by slow evaporation of a
mixed toluene/pyridine solution that had been chromatographed on
a SiO2-packed column using the toluene/pyridine solution as the
eluant. Numerous attempts to grow crystals of CH3CoIIIPc,
CH3CoIIIPc(DMA), CH3CH2CoIIIPc(py), or CH3CH2CoIIIPc(N-
MeIm) suitable for X-ray investigations proved to be unsuccessful.25
X-ray Crystallography. Dark blue crystals of CH3CH2CoIIIPc
and CH3CoIIIPc(py) were used for structure determination. Crystal,
data collection, and refinement parameters are summarized in
Table 2.
[py] [TEMPO]
R
(M)
(M)
CH4 C2H4 CH3CH3 CoIPc-
Me 60 h, 100 °C
Me 70 h, 100 °C 0.7
Me 70 h, 100 °C 0.7
0
0
0
0.02
0
0
0
0
0.02
0.02
0.02
0.02
100
92
0.92
0
0
0
2
2.3
0.35
0
75
0
Et 24 h, 60 °C
Et 6h, 80 °C
Et 15h, 70 °C
Et 15h, 70 °C
Et 15h, 70 °C
Et 15h, 70 °C
Et 15h, 60 °C
Et 6h, 80 °C
0
0
0
0.7
0
0.7
0.7
0.7
0
0
0
0
0
0
0
0
65
66
69
69
21
22
20
19
28
+ n.d.a
30
27
28
+ n.d.a
+ n.d.a
7.5
0
0
0
0
0.4
0.2
0.3
0.2
X-ray data collections were carried out with graphite-monochro-
mated Mo KR radiation (λ ) 0.71073 Å) using a KUMA KM4CCD
diffractometer26 equipped with an Oxford Cryosystems low-
temperature device. Data were corrected for Lorentz-polarization
and absorption effects.27 The structures were solved by direct
methods with SHELXS9728 and refined by full-matrix least-squares
against F2 using SHELXL97.29 All non-hydrogen atoms were
refined with anisotropic displacement parameters; hydrogen atoms
were placed in calculated positions and refined as a riding model,
and the isotropic displacement parameters for the hydrogen atoms
were set at 1.2× the Ueq values for the appropriate carrier atoms.
The ethyl group in CH3CH2CoIIIPc is disordered over two alternative
positions, with site occupation factors of 0.60(1) and 0.40(1). In
the crystal structure of CH3CoIIIPc(py), the disordered molecule of
solvent-pyridine was found; this molecule shows no specific
interactions with the CH3CoIIIPc(py) molecules. This solvent
molecule was refined isotropically with one common Uiso value
and subjected to weak geometric constraints.
Kinetic Measurements. Rates for homolysis of RCoIIIPc com-
plexes were measured spectroscopically in dilute solutions of the
organocobalt substrate under limiting trapping conditions. First-
order rate law was obeyed by the kinetic data for the disappearance
of the organocobalt substrate. The end-point method was used in
most cases. The end points of these slow reactions were determined
by anaerobic photolysis. The homolysis rates were zero-order in
[TEMPO] over the concentration range of 0.005-0.1 M. Poor
solubility of the CoIIPc product sets certain limitations on the kinetic
measurements. If the disappearance of the substrate is to be followed
spectrophotometrically, low concentrations of RCoPc (<2 × 10-5
M) are required, otherwise precipitation of CoIIPc may occur.
Kinetic measurements in toluene are not viable unless an N-donor
ligand is added, which not only prevents precipitation of the
CoIIPc product but also eliminates cleavage of the phthalocyanine
system.
a Observed but not quantified.
quenched to room temperature in a 1 cm cell, and after equilibra-
tion at 25 °C, a UV-visible spectrum in an HP 8452A spectrom-
eter was taken. In the absence of an N-donor, a small volume of
0.5 M sodium cyanide was added to the samples to turn unreacted
RCoIIIPc and the CoIIPc product into RCoPc(CN)- and CoIIIPc(CN)2-,
respectively, and the spectrum was recorded. The rates calculated
from the spectra of the original samples or from those with cyanide
added agreed well, but the latter were more precise. Spectral
changes at several wavelengths, usually the maxima of the
difference spectra, were used for calculations (see Figure S2). (In
the absence of added N-donor, they were 332, 646, and 664 nm,
or when cyanide was added to the sample, they were 348, 428,
508 (methyl), 528 nm (ethyl), and 682 nm; see Figure S4.) The
rate constants determined at various wavelengths agreed within
10%. The averaged values are reported.
In some cases, at 60 and 70 °C, the kinetic measurements were
also performed in spectrophotometric cell under Ar. The results
were consistent with those of sampling. Control runs in air in the
absence of TEMPO yielded similar rates, although aerobic thermoly-
ses of CH3CH2CoIIIPc in DMSO-d6 also produced small amounts
of CH3CH2OOCoIIIPc, as suggested by the appearance of ethyl reso-
1
nances in the H NMR spectrum: δ -1.31 (3H, t, 7 Hz), -0.883
(2H, q, 7 Hz), downfield shifted from those of CH3CH2CoPc.9 (A
similar downfield shift was observed upon oxygen insertion into
the Co-C bond in alkylcobalt(OEP) complexes.)30
Product Studies. Product studies were carried out in (i) DMA,
(ii) DMA in the presence of pyridine, and (iii) DMA in the presence
of pyridine and radical-trap TEMPO. Two exemplary procedures
are described below and the results are collected in Table 1. GC
analysis was performed as described earlier.9
In kinetic runs, usually, a deaerated solution of RCoIIIPc,
containing TEMPO (0.02 M) and an N-donor ligand (where
applicable) in a 25 mL flask was heated under Ar in an oil-
circulating bath. The temperature was maintained constant within
( 0.1 °C. Periodically, an aliquot from the reaction mixture was
CH3CoIIIPc Thermolysis in DMA in the Presence of 0.7 M
Pyridine. A solution of CH3CoIIIPc (12 mg, 0.02 mmol) and
pyridine (300 µL, 0.7 M) in 10 mL of DMA was sealed in a vial
under Ar. In the course of a 70 h incubation at 100 °C, the solution
turned yellow-green. The gas phase over the solution was subjected
GC analysis for light hydrocarbons. Methane (92%) and ethane
(2.3%) were found. A small portion of the yellow-green solution
was transferred under Ar into a 0.1 mm cell, and a visible spectrum
of CoIPc- was obtained (75% yield).
CH3CH2CoIIIPc Thermolysis in DMA in the Presence of
TEMPO. A solution of CH3CH2CoIIIPc (12 mg, 0.02 mmol) and
TEMPO (27 mg, 0.173 mmol) in 10 mL of DMA was sealed in a
vial under Ar. In the course of a 15 h incubation at 70 °C, a nearly
colorless solution was obtained in which a fine dark solid was
suspended. A GC analysis for light hydrocarbons yielded ethylene
(21%) and ethane (0.4%).
(25) Growing an X-ray quality crystal of a phthalocyanine complex is not
an easy task, as can be seen from the scarcity of structural data.
Difficulties were encountered in earlier intense efforts to grow a
suitable CH3CoIIIPc crystal from a THF solution (ref 21), where the
solid consisted prevalently of six-coordinate CH3CoIIIPc(THF) mol-
ecules, but a good refinement could not be achieved because the THF
was lost during the X-ray irradiation. Clearly the stronger ethyl labilizer
and protic medium, in which the strength of O-donor ligands is
considerably reduced, allowed dimeric CH3CH2CoIIIPc to be obtained.
(26) CrysAlisCCD, version 171; Oxford Diffraction Poland Sp.: Wroclaw,
Poland, 2003.
(27) Blessing, R. H. J. Appl. Crystallgr. 1989, 22, 396.
(28) Sheldrick, G. M. Acta Crystallogr. 1990, A46, 467.
(29) Sheldrick, G. M. SHELXL-97, Program for the Refinement of Crystal
Structures; University of Go¨ttingen: Go¨ttingen, Germany, 1997.
(30) Kendrick, M. J.; Al-Akhdar, W. Inorg. Chem. 1987, 26, 3792.
9904 Inorganic Chemistry, Vol. 44, No. 26, 2005

Alkylcobalt(III) Phthalocyanine Complexes
Table 2. Experimental Data for the X-ray Diffraction Studies of
Table 3. Selected and Average Bond Lengths (Å) and Angles (deg) for
CH3CH2CoIIIPc and CH3CoIIIPc(py)
CH3CH2CoIIIPca
CH3CH2CoPc
Co-C(11)
Co-Niso
Niso-Ca
Ca-Naza
Ca-Cb
2.031(3)
1.914(3)
1.376(5)
1.327(3)
1.449(6)
1.387(5)
C(11)-C(12A)
C(11)-C(12B)
Co-Ct
1.428(7)
1.428(12)
0.086(1)
1.391(3)
1.378(4)
1.395(5)
CH3CH2CoPc
(100K)a
(ambient
temperature)a CH3CoPc(py)
Cb-Cc
formula
C34H21CoN8
C34H21CoN8
C38H24CoN9‚
1/2C6H5N
705.14
Cc-Cd
Cd-Cd
Cb-Cb
fw
600.52
600.52
cryst syst
space group
a (Å)
b (Å)
c (Å)
triclinic
P1h
triclinic
P1h
monoclinic
P21/c
10.530(2)
14.977(3)
20.736(4)
90
104.36(3)
90
3168.1(1)
4
Co-C(11)-C(12A)
Co-C(11)-C(12B)
Niso-Co-C(11)
cis-Niso-Co-Niso
Co-Niso-Ca
119.0(3)
120.8(5)
92(2)
89.9(3)
126.7(2)
127.7(2)
106.4(1)
Ca-Naza-Ca
122.1(2)
121.8(2)
110.5(1)
106.3(3)
121.2(2)
117.4(4)
121.4(4)
Naza-Ca-Cb
Niso-Ca-Cb
Ca-Cb-Cb
Cb-Cb-Cc
Cb-Cc-Cd
Cc-Cd-Cd
10.290(2)
11.003(2)
12.728(2)
78.81(1)
72.65(1)
69.08(1)
1278.6(4)
2
10.425(1)
11.153(1)
12.805(1)
78.466(8)
72.39(1)
69.28(1)
1311.7(2)
2
R (deg)
â (deg)
γ (deg)
V (Å3)
Z
Niso-Ca-Naza
Ca- Niso- Ca
N-Co-C(11)-C(12A) -10.2(5) N-Co-C(11)-C(12B) -76.7(9)
a Data for the ambient temperature structure.
F
calc (g/cm3)
temp (K)
1.560
100(1)
3.33-25.00
8237
1.521
1.479
100(1)
3.38-25.00
15858
ambient
3.51-22.50
12697
θ range (deg)
reflns collected
independent refls
4407 (0.047)
4587 (0.059)
5542 (0.033)
(Rint
)
GOF on F2
1.038
1.010
1.13
R1 (I > 2σ(I))
0.0544
0.1035
0.0901
0.1028
0.0458
0.0653
0.0864
0.0967
0.0589
0.0805
0.1351
0.1468
1.321 and
-1.023
all data
wR2 (I > 2 σ (I))
all data
largest diff peak
0.468 and -0.410 0.545 and
and hole (e Å3)
-0.349
a Two different specimens were used to determine the low and ambient
temperature structures.
Figure 2. Perspective view of the (CH3CH2CoPc)2 central fragment
perpendicular to the coordination planes.
Figure 1. Side view of (CH3CH2CoPc)2. The dashed lines link cobalt
atoms with their Niso′ counterparts.
Results and Discussion
Description of Structures. (CH3CH2CoIIIPc)2 The crystal
consists of centrosymmetric dimers (Figure 1). The cobalt
atom of CH3CH2CoIIIPc is five-coordinate and has a distorted
square-pyramidal coordination geometry (Figure 3). The
equatorial ligand planes of the two CH3CH2CoPc molecules
are parallel with a 1.913 Å shift in the Co-Niso direction
from the eclipsed configuration (Figure 2). The distance
between the two mean planes of four isoindole atoms (Niso)
is 3.21(3) Å, while the separation between benzo groups of
one molecule and the N4 plane of the other varies from 2.99
to 3.6 Å. The interplanar separation is thus characteristic of
π stacking. Packing forces and intermolecular interactions
cause some peripheral deformations in the essentially planar
Figure 3. ORTEP drawing and numbering scheme for CH3CH2CoPc. The
displacement ellipsoids are drawn at the 50% probability level, and the
hydrogen atoms are depicted as spheres with arbitrary radii. For clarity,
only one orientation of the ethyl group is shown.
phthalocyanine ligand. The most out-of-N4-plane benzo
group (dihedral angle of ∼7°) is bent toward an ethyl group
of another “dimer”. The cobalt atom of one phthalocyanine
faces an isoindole N atom of the other, Niso′, at a separation
of 3.303 Å, an arrangement different from that of the â form
of CoIIPc, where an azamethine atom of the neighboring
32
molecule at a 3.21931 or 3.154 Å distance is the closest
contact. The straight line determined by Co and Niso′ is
Inorganic Chemistry, Vol. 44, No. 26, 2005 9905

Galezowski and Kubicki
Table 4. Selected and Average Bond Lengths (Å) and Angles (deg) for
CH3CoIIIPc(py)
Co-C(11)
Co-Niso
Niso-Ca
Ca-Naza
Ca-Cb
1.993(4)
1.922(3)
1.373(4)
1.322(5)
1.453(4)
1.391(6)
1.389(4)
Co-Nax
2.177(3)
-0.0052(14)
1.381(5)
1.394(6)
1.342(3)
Co-Ct
Cc-Cd
Cd-Cd
Nax-C2 (py)
C2-C3 (py)
C3-C4 (py)
Cb-Cb
1.379(3)
1.374(4)
Cb-Cc
Niso-Co-C(11)
cis-Niso-Co-Niso
Niso-Co-Nax
89.8(7)
90.0(2)
90.1(8)
Ca-Naza-Ca
121.5(1)
122.2(4)
110.0(3)
106.4 (3)
121.3 (6)
117.4 (4)
121.4 (3)
Naza-Ca-Cb
Niso-Ca-Cb
Ca-Cb-Cb
Cb-Cb-Cc
Cb-Cc-Cd
Cc-Cd-Cd
Co- Niso-Ca
Niso-Ca-Naza
Ca-Niso-Ca
126.4 (3)
127.8 (2)
107.2 (2)
Figure 4. ORTEP drawing and numbering scheme for CH3CoPc(py). The
displacement ellipsoids are drawn at the 50% probability level, and the
hydrogen atoms are depicted as spheres with arbitrary radii.
C2py-Nax-C6py
Nax-C2py-C3py
C2py-C3py-C4 py
117.1(3)
123.1(1)
119.2(1)
py rotationa
py cantb
C3py-C4py-C5 py
29.2
88.8(1)
118.4(4)
a
Dihedral angle between the pyridine plane and an Niso-Co-Nax plane.
b Dihedral angle between pyridine plane and the mean plane of the four
Niso atoms.
inclined to the plane of four Niso atoms at an angle of 87.7°.
The four isoindole nitrogen atoms that form the basal plane
are coplanar within ( 0.004 Å. The cobalt atom is displaced
0.0857(14) Å from the mean (Niso)4 plane toward the ethyl
group. The average Co-Niso bond length, 1.916(3) Å, is vir-
tually identical to that found in the â form of CoIIPc.33 The
cis Niso-Co-Niso angles average 89.9(3)°. The axial Co-C
bond is 2.031(3) Å long and is tipped slightly, 2.3°, from
the normal to the mean plane of four Niso atoms. The axial
ethyl group is disordered (see the Experimental Section).
There are two arrangements in which the CH2-CH3 bond
and a Co-Niso bond are nearly eclipsed (torsion angles
Niso-Co-C(11)-C(12) ) 10.5(6) and 76.2 (10)°). The
corresponding Co-C(11)-C(12) angles are 119.2(4) and
121.1(5)°, respectively. The disorder did not disappear when
the structure was determined at 100 K. The torsion angle
Niso-Co-C(11)-C(12) of 10.5° is similar to the analogous
angle in CH3CH2Co(OEP).7 The average bond lengths and
angles of the phthalocyanine macrocycle are typical.
atoms that form the basal plane are coplanar within (0.0056
Å. The cobalt atom is displaced by at most a trivial amount,
0.0054(12) Å, from this plane toward the nitrogen atom of
the axial pyridine ligand. The Co-C bond is 1.990(3) Å long
and is perpendicular to the plane. The average Co-Niso bond
length is 1.920(4) Å, and that of Co-Nax is 2.174(2) Å. The
Niso-Co-Niso angles depart trivially from the idealized 90°.
The plane of the pyridine ring is inclined at an angle of
88.8° with respect to the equatorial nitrogen mean plane and
is rotated 29.1° from an Niso-Co-Nax plane. The latter angle
is significantly smaller than 45°, the value corresponding to
the orientation with minimal nonbonded contacts between
the pyridine ortho hydrogen atoms and the phthalocyanine
core. It is close to the 28.6(2)° value found in CoIIPc(py)2,
where nonbonded interactions are unlikely because of the
very long Co-Nax bonds,34 and it is larger than the 14.3°
rotation angle in relevant CH3Co(OEP)(py). The average
bond lengths and angles of the phthalocyanine macrocycle
are unremarkable.
Five-Coordinate Structure of CH3CH2CoIIIPc in Solid
Form. The pseudo-five-coordinate structure of relatively
electron-deficient CH3CH2CoIIIPc can be observed in the
solid only because the attracting interaction between two
large aromatic systems is stronger than the binding energy
of trans axial O-donor ligand in the protic medium (the
solvents in which the crystals are grown). Packing effects
may also favor the dimeric structure of CH3CH2CoIIIPc in
the solid, since in dilute solutions, in nonbinding solvents,
the complex exists in monomeric form,9 which attests to the
weakness of Co-Niso′ interaction in the dimer. The weakness
of the interaction between the Co atom and Niso′ is also
evident from (i) the long Co-Niso′ distance, which is ca.
1.2 Å larger than the 2.05 Å Co(III)-N(sp2) coordination
distance calculated by Little and Ibers,35 (ii) the negligible
0.014 Å displacement of the Niso′ atom toward facing Co,
and (iii) the fact that the cobalt remains significantly
Thermal motion was held responsible for the extremely
short CH3-CH2 distance of 1.39(1) Å observed for the other-
wise normal ethyl group in the ambient temperature structure
of the six-coordinate complex CH3CH2Co(saloph)(py).3 Two
independent determinations of the CH3CH2CoIIIPc structure,
one of them carried out at 100 K and the other at ambient
temperature, gave essentially identical results. This makes
the thermal motion explanation for the shortening of the
C(11)-C(12) bond in CH3CH2CoIIIPc less plausible. Also,
the shortening of the CH3-CH2 distance in other σ-bonded
ethylcobalt(III) complexes appears to be real. For instance,
should thermal motion be the only reason for the apparent
shortening of the CR-Câ bond in the propyl group of
n-PrCo[TC-(3,3)],4 then the Câ-Cγ bond would be shorter
than CR-Câ one, which is not the case (Table 5).
CH3CoIIIPc(py). The cobalt atom of CH3CoIIIPc(py) has
distorted octahedral geometry. The four isoindole nitrogen
(31) Mason, R.; Williams, G. A.; Fielding, P. E. J. Chem. Soc., Dalton
Trans. 1979, 676.
(32) Williams, G. A.; Figgis, B. N.; Mason, R.; Mason, S. A.; Fielding, P.
E. J. Chem. Soc., Dalton Trans. 1980, 1688.
(33) Figgis, B. N.; Kucharski, E. S.; Reynolds, P. A. J. Am. Chem. Soc.
1989, 111, 1683.
(34) Janczak, J.; Kubiak, R. Inorg. Chim. Acta 2003, 342, 64.
(35) Little, R. G.; Ibers, J. A. J. Am. Chem. Soc. 1974, 96, 4440.
9906 Inorganic Chemistry, Vol. 44, No. 26, 2005

Alkylcobalt(III) Phthalocyanine Complexes
Table 5. Comparison of the Geometry of Five-Coordinate R-Cobalt(III) Complexes with Primary Alkyl Axial Group R
complex
Co-C (Å)
CR-Câ (Å)
Câ-Cγ (Å)
Co-C-C (deg)
Co-Neqa (Å)
Co-Ctc (Å)
ref
CH3CH2CoPcb
2.031(3)
1.988(2)
1.963(6)
1.976(7)
1.957(13)
1.973 (6)
1.940 (8)
1.952 (18)
1.428(7) 1.428(12)
1.503(3)
1.496(9)
1.49(1)
-
-
-
-
-
120.8(5) 119.0(3)
114.2(1)
116.7(4)
117.7(5)
-
-
-
-
1.914(3)
1.970(8)
1.890(4)
1.894(9)
1.868(5)
1.966(14)
1.90g
0.086(1)
0.103(1)
0.213(1)
0.185(1)
this work
CH3CH2Co(OEP)
CH3CH2Co[TC-(3,3)]
n-PrCo[TC-(3,3)]
CH3CoIII(saloph)d
CH3CoIII(OEP)
-
7
5
5
3
6
5
2
-
1.53(1)
-
-
-
-
0.105(2)
0.094 (3)e 0.105(3)f
0.107(1)
CH3Co[TC-(4,4)]
CH3Co(acacen)
1.873(12)
0.12
a Average value. b In this table ambient temperature data are given, but the less well-resolved low-temperature structure is not significantly different.
c Displacement of Co atom from the least-squares plane of the four equatorial donors. d Molecule B. e Molecule 1. f Molecule 2. g Average value for a
molecule where two pairs of different length Co-Neq bonds are observed.
displaced toward the ethyl group from the mean N4 plane,
while in six-coordinate complexes it is virtually in-plane.
The Co-Ct displacement in CH3CH2CoIIIPc is merely
0.017(2) Å smaller than it is in its strictly monomeric OEP
congener. In the X-ray structures of some RCoIII(chel)dimers,
considerably shorter distances between Co and a facing donor
(usually oxygen atom, O′) have been found, such as the
2.34 Å Co-O′ distance in (EtCo(salen))236or the 2.435 Å
Co-O′ distance in one of CH3Co(saloph) dimers,9 and the
donor atoms are visibly displaced from their equatorial
moiety toward the Co atom. These cannot be regarded truly
five-coordinate complexes.
Further discussion on this point would not be necessary
if it were not for the fact that the weak axial interactions in
â-MIIPc have been held responsible for changes in the ground
electronic state between the monomeric tetragonal molecules
and the â phase,37,38 where each cobalt atom interacts with
two Naza atoms at distant (∼3.2 Å) axial positions. It should
be noted that in present case we have only one interaction
of this type per one CH3CH2CoPc molecule. Furthermore,
the interaction is probably weaker than those in the â form.
According to Figgis et al., in a CoIIPc molecule, a consider-
able positive charge (+1.88) resides on the cobalt atom, and
Naza is more negatively charged than Niso.37 Face-to-face
interactions and charge distribution over the planar system
in CH3CH2CoPc are not expected to be very different. In
the â form and in the CoIIPc dimer, van der Waals forces
are the main contributor to binding energy, while electrostatic
interactions are believed to determine the orientation of the
two planes.39 Therefore, the preference for Naza over Niso
atoms in the â conformer, also present in the structure of
CoIINPc(py),40 can be accounted for by both the larger
Co-Co distance and the increased negative charge on
Naza. The fact that this preference is not manifested in
(CH3CH2CoPc)2 and in the R conformer of CoIIPc,41 where
there is no N atom in the axial position, underlines the
weakness of this type of interactions.42 Even though the
CH3CH2CoPc molecules in the solid are not separated as
ideal, and it would be unwise to disregard the interaction
between Co and Niso′ completely, it seems justified to view
the structure as essentially five-coordinate.
Comparison of the Structures. Comparisons with other
σ-bonded alkyl-Co(III) phthalocyanine complexes are not
possible because none have been structurally characterized
prior to this report. Despite intense efforts, a complete series
of structures needed for the investigation of structural effects
relevant to homolytic Co-C bond disruption, such as five-
coordinate RCoIIIPc, RCoIIIPc(L) and CoIIPc(L),43 is not
available. Also the alkyl group effect on the structure cannot
be directly studied in either the RCoIIIPc or RCoIIIPc(L)
series.
Factors influencing the length (and strength) of Co-C
bond have gained much interest recently, and a great deal
of theoretical work has been done to explain the so-called
inverse trans effect. On the other hand, it should be
recognized that alkylcobalt chemistry may exemplify a
fundamental rule of chemistry that the length of a single bond
between a given pair of atoms is remarkably independent of
the identity of a molecule in which these atoms reside.49
Small variations in the Co-C distances, in particular for
complexes with same axial alkyl group, should be considered
from this perspective, especially since the accuracy of the
X-ray determinations is often low compared to the observed
changes in bond lengths.
CH3CH2CoIIIPc appears to be the most electron-deficient
five-coordinate alkylcobalt(III) model among the few struc-
turally characterized complexes of this type. The Co-C bond
in CH3CH2CoIIIPc, 2.031(5) Å, is the longest found in five-
(42) It should be also noted that in the structure of MgPc(H2O)(py)2 dimer
N atoms of the phthalocyanine moiety are not even the closest contacts
of the central metal atoms. Fischer, M. S.; Templeton, D. H.; Zalkin,
A.; Melvin, C. J. Am. Chem. Soc. 1971, 93, 2622.
(43) No X-ray structures of strictly five-coordinate CoIIPc(N-donor)
complexes are known.
(44) Bigotto, A.; Zangrando, E.; Randaccio, L. J. Chem. Soc., Dalton Trans.
1976, 96.
(36) Calligaris, M.; Minichelli, D.; Nardin, G.; Randaccio, L. J. Chem.
Soc. A 1971, 2722.
(37) Reynolds, P. A.; Figgis, B. N. Inorg. Chem. 1991, 30, 2294.
(38) Liao, M.-S.; Watts, J. D.; Huang, M.-J. Inorg. Chem. 2005, 44, 1941
and references therein.
(39) Heuts, J. P. A.; Schipper, E. T. M. W.; Piet, P.; German, A. L. J.
Mol. Struct. 1995, 333, 39 and references therein.
(40) Gacho, E. H.; Naito, T.; Inabe, T.; Fukuda, T.; Kobayashi, N. Chem.
Lett. 2001, 260.
(41) Ballirano, P.; Caminiti, R.; Ercolani, C.; Maras, A.; Orru, M. A. J.
Am. Chem. Soc. 1998, 120, 12798.
(45) Calligaris, M.; Nardin, G.; Randaccio, L. Inorg. Nucl. Chem. Lett.
1972, 8, 477.
(46) Randaccio, L.; Furlan, M.; Geremia, S; Slouf, M.; Srnova, I.; Toffoli,
D. Inorg. Chem. 2000, 38, 3403.
(47) Zou, X.; Brown, K. L. Inorg. Chim. Acta 1998, 267, 305.
(48) Ouyang, L.; Rulis, P.; Ching, W. Y.; Nardin, G.; Randaccio, L. Inorg.
Chem. 2004, 43, 1235.
(49) The difference in the bond distance between probably extreme cases
of i-PrCo(DH)2(py) (Marzilli, L. G.; Toscano, P. J.; Randaccio, L.;
Bresciani-Pahor, N.; Calligaris, M. J. Am. Chem. Soc. 1979, 101, 6754)
and CF3Cbl (Table 6) appears to be significant, 0.22 Å.
Inorganic Chemistry, Vol. 44, No. 26, 2005 9907

Galezowski and Kubicki
Table 6. Bond Distances for the Coordination Sphere of Selected
Apart from electronic cis influence, there is another poten-
tial reason for the elongated Co-C bond in the phthalocya-
nine complex: it seems that the Co atom is held tighter
(lower Co-Ct value) by the rigid phthalocyanine equatorial
ligand, which forms shorter bonds to the central metal atom
than those formed by the porphyrin ligands. (The length of
Co-Neq bonds in phthalocyanine complexes are closer to
those formed by more flexible ligands (Table 6).) The
reduced displacement of the cobalt atom from the equatorial
ligand plane, Co-Ct, may force expansion of the Co-C bond
to avoid nonbonded interactions of the alkyl group with the
equatorial moiety. However, as a dominant factor, this is
incompatible with the fact that in the OEP model the sum
of the Co-C and Co-Ct distances is still 0.026 Å less than
that in the phthalocyanine complex (Table 5).
Short Co-C bonds in five-coordinate tropocoronand
models (Table 5), 0.068 Å shorter in CH3CH2Co[TC-(3,3)]
than those in CH3CH2CoPc, are striking. The tropocoronand
compounds must be extremely electron-rich, as shown by
the lack of any trans-axial ligation, even by good N-donor
ligands,5 and large negative Co(II)/Co(I) potentials.54 Despite
their short length, the Co-C bonds in the tropocoronand
models are not particularly strong.5
In the presence of two axial ligands, including a strong
alkyl labilizer, a strong electron-donating equatorial system
may exert a destabilizing effect causing an increase in both
axial bond distances. In practice, the lengths of relatively
weak Co-Nax bonds are far more sensitive than the Co-C
distances to changes in the equatorial system (Table 6). It
seems that in methylcobalt porphyrinic complexes the steric
interactions of (nonbulky) axial N-donor ligands with the
equatorial moiety are of secondary importance. Hence, the
Co-Nax distance in CH3Co(OEP)(py), larger than those in
MeCbl55 and CH3CoPc(py), is likely to reflect, mostly, a
stronger electron donation from OEP than that from Pc or
corrin.56 The Co-Nax bond length could be considered to
be a measure of the combined inductive effects of the axial
alkyl ligand and the equatorial system. This superposition
of trans and cis influences would appear to be similar in
MeCbl and CH3CoPc(py), in which both axial bond distances
are very similar (Table 6). We could try to use the Co-Nax
distances in complexes with same alkyl and N-donor axial
ligands to characterize the electron deficiency of the equato-
rial system. By this criterion, phthalocyanine falls between
the Costa-type models and the porphyrins (Table 6), an order
consistent with that of the formation constants for CH3CoIII-
(chel)(py) complexes.9 However, the Co-Npy bond length
in CH3Co(acacen)(py), thought to be one of the most
electron-rich B12 models,51 does not fit into the trend.
The long Co-Nax bond in CH3CoPc(py), 2.174(2) Å,
compared to 2.068(3) Å in CH3CoIII(DH)2(py) (Table 6),
CH3CoIII(chel)(py) Complexes and Some Alkylcobalamins
Co-C
(Å)
Co-Nax
(Å)
BDE
(kcal/mol)
complex
ref
CH3CoIII(DH)2(py)
1.998 (2)
2.068 (3) 44
33 ( 2a
[CH3Co(DO)(DOH) 2.003 (3)
pn(py)]PF6
2.106 (3)
4
CH3CoIIIPc(py)
CH3CoIIIOEP(py)
CH3Co(acacen)(py)
MeCbl
1.990 (3)
2.018 (12) 2.214 (9)
1.99 (1)
1.979 (4)
1.878 (12) 2.049 (1) 47
2.030 (3) 2.237 (3) 48
2.174 (2) this work 30 ( 1
6
2.16 (1)
45
2.162 (4) 46
37 ( 3b
CF3Cbl
AdoCbl
31.5 ( 1.3c
a Ref 19. b Ref 18. c Ref 48.
coordinate complexes with a simple primary alkyl axial group
(Table 5). It is as long as the bonds in CH3(CO)CH2Co-
(TPP), 2.028(3) Å,50 or (i-Pr)Co(saloph)‚1.5H2O, 2.031(8)
Å.51 Such an elongated Co-C bond distance in a 16-electron
complex with an “electron-deficient” equatorial system
(electron density outflow from the axial system) is perhaps
consistent with increased electrophilic character of the R
carbon, in accordance with the outstanding SN2 reactivity
of the five-coordinate RCoIIIPc complexes.9
As is apparent from Table 5, the short CH3-CH2 bond
and the large Co-C-C angle of ∼120° are in the range of
values characteristic of alkylcobalt(III) complexes. The
simultaneous decrease of the CR-Câ bond distance and
opening of the Co-C-C angle, with a virtually unchanged
Co-C bond length, were observed by Marzilli et al. upon
increasing the bulk of the trans axial ligand in ethyl-
cobaloxime models.52 It is interesting that, in CH3CH2CoPc,
the largest angle coincides with the shortest CR-Câ bond
of those for the five-coordinate structures shown in Table 5.
In view of the long Co-C bond in the phthalocyanine model,
it seems unlikely that the difference in the structural
parameters for the ethyl group in complexes containing
equatorial ligands of similar geometry, CH3CH2CoPc and
CH3CH2Co(OEP), is caused by steric interactions.
The Co-C bond lengths have been claimed to be insensi-
tive to the nature of the equatorial ligand but dependent on
the bulk and electron-donor properties of the axial alkyl
ligand.51 This rule seems to be followed by the six-coordi-
nate complexes as seen, for instance, in Table 6, but the first
part of it is not quite observed by few structurally charac-
terized five-coordinate complexes, where the Co-C bonds
tend to be shorter in the electron-rich models than they
are in the electron-poor models (Table 5). The Co-C bond
in CH3CH2CoIIIPc is 0.043(4) Å longer than that in
CH3CH2CoIII(OEP). When the small variance in the Co-C
bond distances is taken into account, especially among
complexes with same axial alkyl group, the difference seems
to be significant.53 As an example, it is comparable to the
Co-C bond elongation in AdoCbl, calculated for drastic
compression of the trans Co-Nax bond.14
(54) Doerrer, L. H.; Bautista, M. T.; Lippard, S. J. Inorg. Chem. 1997, 36,
3578.
(55) The Dmbz ligand causes some upward folding of the corrin ring but
with minimal effect on the Co-C bond when the ligand is CH3 (ref
51b).
(56) The elongation of the Co-C bond in CH3Co(OEP) on pyridine binding
is real but because of the lower resolution of the CH3Co(OEP)py
structure the extent of this expansion is less certain.
(50) Kastner, M. E.; Scheidt, W. R. J. Organomet. Chem. 1978, 157, 109.
(51) Marzilli, L. G.; Summers, M. F.; Bresciani-Pahor, N.; Zangrando, E.;
Charland, J.-P.; Randaccio, L. J. Am. Chem. Soc. 1985, 107, 6880.
(52) Marzilli, L. G.; Summers, M. F.; Zangrando, E.; Bresciani-Pahor, N.;
Randaccio, L. J. Am. Chem. Soc. 1986, 108, 4830.
(53) The comparison is meaningful because of the similar steric require-
ments of the two complexes, at least in the vicinity of the Co center.
9908 Inorganic Chemistry, Vol. 44, No. 26, 2005

Alkylcobalt(III) Phthalocyanine Complexes
Scheme 1. Thermolysis of the Co-C Bond in CH3CH2CoIIIPc
agrees well with the observation that RCoIIIPc (N-donor)
compounds are kinetically labile for complexes with an
electron-deficient equatorial system.9,57 The increased lability
of the long Co-Nax bonds has been observed by Marzilli
et al.58
Contradiction Between Cis Electronic Influences in
Five- and Six-Coordinate Complexes. It should be em-
phasized that, while the Co-C bond in CH3CH2CoIIIPc is
longer than that in CH3CH2CoIII(OEP), both axial bonds
in CH3CoIIIPc(py) are shorter than those in its OEP ana-
logue. The seemingly contradictory trends in the cis effect
on the length of the Co-C bond in the five- and six-
coordinate complexes are noteworthy, especially in view of
the similar steric requirements of the OEP and Pc com-
plexes, at least in the vicinity of the coordination center. The
changes in the Co-C bond distance in the three five-
coordinate ethylcobalt(III) compounds of Table 5 parallel
the order of Co(II)/Co(I) redox potentials for the correspond-
ing CoII(chel): -0.35,12 -0.95,13 and ca. -1.6 V59 vs SCE
for the phthalocyanine, octaethylporphyrin, and tropocoro-
nand complexes, respectively.
The inverse trend between the Co-C bond length and the
equatorial ligand basicity observed for five-coordinate
CH3CH2CoIII(chel) complexes is based on a few well-
resolved structures. More structural data in support of this
proposal are certainly needed. Unfortunately, because of large
uncertainties of the Co-C distances in some of the five-
coordinate CH3CoIII(chel) compounds of Table 6 and the lack
of data for a structure with a strongly electron-deficient
equatorial ligand, the trend, if existent, hardly could be
observed in the methyl-Co(III) system.
As was pointed out by one of reviewers for this work, the
trend may seem counterintuitive. Indeed, if we view the
Co-C bond as formed by an “electrostatic” interaction of
Co3+ with a negatively charged alkyl ligand, then the strong
alkyl donor should be pulled more tightly in a complex with
a less-donating equatorial system. However, it has been noted
on many occasions that the Co-C bond is a normal covalent
bond, as witnessed by homolytic cleavage of the bond, and
the cobalt should not be considered to be true Co3+.60 The
fact that nucleophilic attack on the C atom is possible, at
least in some cases, suggests that the axial system is electron
deficient,61 and there is evidence that the five-coordinate form
is the reactive species in these SN2 reactions.9 The hypothesis
of the longer Co-C bonds in five-coordinate complexes with
less electron-donating equatorial ligands is viable if the
Co-C bond is considered as essentially covalent. In five-
coordinate complexes with strong equatorial donors, an
electron-density shift into the axial bond would be possible
making it shorter. The trend would not be held in six-
coordinate complexes, where the trans axial ligand exerts a
destabilizing influence (4 electrons total in the axial system).
The Co-Nax bonds in the five-coordinate porphyrinic
complexes of Co(II) do not follow the trend described above.
The Co-Npy distances are remarkably invariant in CoII(TPP-
(CF3)4)(py),62 2.166(3) Å, CoIITPP(3,5-lutidine),63 2.166 Å,
and CoII(NPc)(py), 2.167(9) Å.40 A shorter Co-Nax distance
was expected in CoII(TPP(CF3)4)(py) because of the presence
of electron-withdrawing CF3 groups,62 and steric interactions
in this severely distorted molecule were invoked to explain
why it was not. The longest bond in compounds of this type,
2.191(2) Å, was found in CoII(OEP)(4-Me2N-py), where,
notably, no substantial steric interactions are present.6 Thus,
it appears that in 17-electron five-coordinate complexes of
2
Co(II) with an unpaired electron in the 3dz orbital, the trend
is not held: strong electron-donating equatorial ligands make
the axial bond, if anything, longer.
Homolysis of the Co-C Bond in RCoIIIPc. When
CH3CoIIIPc is thermolyzed in DMA in the absence of
TEMPO, regardless of the presence of pyridine ligand,
methane is evolved almost quantitatively (Table 1). Appar-
ently there is a sufficient amount of H-donors in the medium.
As expected, the fate of the methyl radical is different in
the presence of a radical trap such as TEMPO: low yields
of methane show that the scavenging by ca. 0.02 M TEMPO
is much faster than the interception by H-donors. At this
TEMPO concentration, the trapping of free (out-of-cage)
radicals usually exhibits saturation.64 Thermolysis of
CH3CH2CoIIIPc yields a â-H-elimination product, ethylene,
in 69 and 21% yield in the absence and presence of the
TEMPO trap, respectively (Table 1), consistent with ethylene
being produced in a cage reaction (Scheme 1).
(57) Galezowski, W. Inorg. Chem. 2005, 44, 5483.
(58) Parker, W. O., Jr.; Zangrando, E.; Bresciani-Pahor, N.; Marzilli, P.
A.; Randaccio, L.; Marzilli, L. G. Inorg. Chem. 1988, 27, 2170.
(59) Reported as approximately -2.1 V vs Fc+/Fc in THF in ref 52. A
typical shift between SCE and Fc+/Fc of 0.5 V is assumed (see ref
12).
(60) Stadlauber, E. A.; Holland, R. J.; Lamm, F. P.; Schrauzer, G. N.
Bioinorg. Chem. 1974, 4, 67. (b) A rather extreme view that
alkylcobalt(III) complexes have considerable Co(I) character has been
presented in Andriuniow, T.; Zgierski, M. Z.; Kozlowski, P. M. J.
Phys. Chem. B 2000, 104, 10921 and also in (c) De Ridder, D. J. A.;
Zangrando, E., Bu¨rgi, H.-B. J. Mol. Struct. 1996, 474, 63.
(61) The Mulliken charges calculated for the MeCbl-like structure in Jensen,
K. P.; Ryde, U. J. Am. Chem. Soc. 2003, 125, 13970 are -0.01 on
CH3 and +0.54 on Co.
The yields of ethylene found in this study are fairly large
compared to some other systems,65 but the thermolysis of
cyclopentylmethyl-B12 produced still larger amounts of the
(62) Terazono, Y.; Patrick, B. O.; Dolphin, D. H. Inorg. Chim. Acta 2003,
346, 265.
(63) Scheidt, W. R.; Ramanuja, J. A. Inorg. Chem. 1975, 14, 2643.
(64) Stolzenberg, A. M.; Cao, Y. J. Am. Chem. Soc. 2001, 123, 9078.
(65) Tsou, T.-T.; Loots, M.; Halpern, J. J. Am. Chem. Soc. 1982, 104, 623
and references therein.
Inorganic Chemistry, Vol. 44, No. 26, 2005 9909

Galezowski and Kubicki
Table 7. Rates for Thermolysis of CH3CoIIIPc, kobs (10-6 s-1), in
DMAa
ligand added
temp (°C)
none
pyridineb
N-MeImc
60
70
80
90
100
110
1.8 ( 0.2
6.7 ( 0.3
22.4 ( 0.4
110 ( 10
280 ( 20
2.7 ( 0.1
11.5 ( 0.4
35 ( 1
6.3 ( 0.3d
13.2 ( 0.3
38 ( 2
160 ( 5
425 ( 20
a [TEMPO] ) 0.02 M b [py] ) 2 M. c [N-MeIm] ) 0.5 M. d In toluene.
Table 8. Rates for Thermolysis of CH3CH2CoIIIPc(L), kobs (10-6 s-1),
in DMAa
Figure 5. Spectral changes upon homolysis of CH3CoIIIPc(py) in DMA
(80 °C) at 0, 6, 15, and 24 h. [py] ) 2 M.
ligand added
temp (°C)
none
pyridineb
N-MeImc
45 ( 2
55
60
65
70
80
90
30 ( 2
62 ( 2
130 ( 6
250 ( 10
880 ( 40
23 ( 2
110 ( 10
450 ( 30
1350 ( 50
160 ( 10
510 ( 20
a [TEMPO] ) 0.02 M. b [py] ) 2 M. c [N-MeIm] ) 0.5 M.
A comparison of the rates for the homolytic cleavage of
the Co-C bond in five- and six-coordinate species would
be of great interest. Unfortunately, homolysis of five-
coordinate RCoIIIPc species in toluene is not viable because
of a decay process yielding an unknown orange product.70
This unwanted process seems to be independent of the
presence of air. The addition of a good ligand, such as
pyridine prevents the bleaching, but then six-coordinate
species prevail.9 Ironically, spectral changes observed in the
course of thermolysis of dilute solutions of RCoIIIPc (Figure
5) are similar to those of SN2 reactions in dilute solutions of
the organocobalt substrate, where the immediate CoIPc-
product is not observed because of the presence of trace
oxidants.9 Furthermore, thermolysis of concentrated solutions
of the organocobalt substrate, in the absence of the TEMPO
trap, often yields CoIPc- (Table 1), the formation of which
could be generally viewed as an indication of heterolytic
methyl transfer. In the presence of TEMPO, no trace of
CoIPc- is observed.71 Therefore GC analysis of the gas phase,
in the absence of TEMPO, is of great help in distinguishing
between the homolytic and displacement mechanisms un-
ambiguously.
Figure 6. Effect of pyridine on the rates of thermolysis of CH3CH2CoPc
in DMA (60 °C), [TEMPO] ) 0.02 M. The curve was calculated for Kpy
) 12.5 M-1, kpy ) 6.2 × 10-5 s-1, and kDMA ) 2.3 × 10-5 s-1
.
â-H-elimination product.66 In the absence of an effective
radical trap, in concentrated solutions of CH3CH2CoIIIPc,
recombination of the ethyl free radicals with CoIIPc clearly
plays a role, leading to an increased yield of ethylene. It
appears that the putative second step of â-H elimination,
hydrogen atom transfer from the caged ethyl radical, is
extremely rapid, only about 4 times slower than separation
of the geminate radical pair. Such fast â-H eliminations are
not unprecedented and have been shown to be consistent with
a nonconcerted mechanism via a cobalt hydride,67 an
intermediate that was frequently postulated but rarely
detected. The CoIPc- product observed when concentrated
CH3CH2CoIIIPc is thermolyzed anaerobically in the absence
of TEMPO agrees well with this mechanism, since in DMA,
because of the extremely low basicity of CoIPc-, HCoIPc
would readily dissociate into H+ and CoIPc-.68,69
As depicted in Figure 6, ligation by pyridine accelerates
the homolysis, although the effect is very moderate. The
curve shown in Figure 6 was fitted using an independently
measured stability constant for CH3CH2CoIIIPc(py).9 In
kinetic runs performed in the presence of pyridine or
N-MeIm, the concentration of the ligand was large to enforce
saturation regime.
Kinetic data for homolysis of RCoPc are collected in
Tables 7 and 8. The rates were measured over a range of
temperatures to determine activation parameters and bond
dissociation energies (BDEs) that are listed in Table 9. Eyring
plots are available in the Supporting Information. Under
limiting trapping conditions, the kobs for disappearance of
CH3CoPc is simply the rate constant k1 for homolysis
(formation of free radicals). The k1 values for the ethyl
(70) A reaction similar to the formation of a yellow decay product in DMF
observed by others: Rollmann, D.; Iwamoto, R. T. J. Am. Chem. Soc.
1968, 90, 1455.
(71) The reduction potential of TEMPO is -0.15 V vs SCE: Finke, R.
G.; Smith, B. L.; Mayer, B. J.; Molinero, A. A. Inorg. Chem. 1983,
25, 3677 and references therein. This does suggest an easy oxidation
of CoIPc- by the trap. Oxidative properties of TEMPO were also
discussed in Samsel, E.; Kochi, J. J. Am. Chem. Soc. 1986, 108, 4790.
(66) Kim, S. H.; Chen, H. L.; Feilchenfeld, N.; Halpern, J. J. Am. Chem.
Soc. 1988, 110, 3120.
(67) Garr, C. D.; Finke, R. G. J. Am. Chem. Soc. 1992, 114, 10440.
(68) In ref 21, we demonstrated that there is no measurable H2 evolution
on addition of excess CF3COOH to a CoIPc- solution.
(69) Electron transfer from the ethyl radical to Co(II)Pc followed by
deprotonation of ethyl cation is also conceivable.
9910 Inorganic Chemistry, Vol. 44, No. 26, 2005

Alkylcobalt(III) Phthalocyanine Complexes
Table 9. Activation Parameters for Homolytic Cleavage of the Co-C
decreases the intrinsic free energy barrier for methyl
transfer.22 However, since the variation in the BDEs is not
very large, this is not going to be the dominant factor.
A comparison of the homolysis of Pc and OEP complexes,
in light of their nearly identical steric requirements, would
be of interest. Unfortunately, kinetic data for the thermolysis
of CH3Co(OEP)(py) or CH3CH2Co(OEP)(py) are not acces-
sible, possibly because of inconveniently slow rates.30,76 The
BDEs for the C6H5CH2CoIII(OEP) complexes are fairly large,
29.3 kcal/mol for C6H5CH2Co(OEP)PBu3,23 similar to those
of the RCoPc(py) complexes, whereas the activation enthal-
pies of benzyl derivatives are usually several kilocalories
per mole lower than those of normal primary alkyl com-
plexes. Consequently, it appears that a 0.03 Å shorter Co-C
bond in CH3CoIIIPc(py) is probably weaker (that is, its BDE
is smaller) than the longer bond in CH3CoIII(OEP)(py), which
is against the notion that a shorter bond should be stronger.
On the basis of results obtained with complexes containing
flexible equatorial arrangements, the lack of correlation
between the Co-C bond length or other properties of this
bond in the ground-state, such as the stretching Co-C
frequency and BDE, was pointed out by Marzilli et al. two
decades ago.77 As this study demonstrates, these claims are
also valid for complexes with more rigid equatorial arrange-
ments, where trans-steric influence is minimized.
Effect of the Equatorial Ligand on BDE. The electro-
chemical investigations by Qiu and Sawyer suggest an
increase in the free energy of Co-C bond dissociation,
described as -∆GBF, with the electron density of the
porphyrin ring.17 This is in agreement with the results of
homolytic cleavage studies on organoiron porphyrin com-
plexes, the behavior of which parallels that of RCoIII(chel)
complexes.78 The following order for the homolyses of (five-
coordinate) PhFe(P) complexes in mesitilene was found:
TPP > TTP > OEP. This is in agreement with the FeIII/FeII
potentials or porphyrin basicities, in other words, the BDE
increases with the equatorial ligand basicities. Riordan and
Halpern concluded that the cis-electronic influence visibly
changes the BDEs in a manner analogous to that of the trans
electronic effect observed in phosphine adducts with rigid
benzyl-CoIII(OEP), where the trans-steric influence is offset,
or in complexes with flexible equatorial system RCoIII(DH)2-
(X-py) complexes,79 where a series of isosteric trans ligands
are used. In each case, the “purely electronic effect” seems
to be the same, the Co-C bond dissociation rates decrease
with increasing ligand basicity. Hence, the homolyses of
RCoPc complexes would reasonably be faster than those of
their OEP analogues (not reported, but estimated to be very
slow). There are, however, observations that do not fit into
this scheme. Since Pc precedes TPP in the order of equatorial
ligands, one could predict a slower homolysis of RCo(TPP)
Bond in RCoIIIPc Complexes in DMA
q
q
a
axial
∆H1
∆S1
DCo-R
temp range
R
ligand (kcal/mol) (cal mol-1 K-1) (kcal/mol)
(°C)
Me DMA 32.6 ( 0.8
10 ( 2
16 ( 3
8 ( 3
32 ( 1
31 ( 1
30 ( 1
29 ( 1
70-110
60-90
60-100
55-80
Et
Me py
Et py
DMA
32 ( 1
31 ( 1
30.5 ( 0.3
13 ( 1
a Co-C bond dissociation energy, calculated as recommended in refs
74 and 20.
system (Scheme 1), used for determination of activation
parameters for homolysis (and the BDE), were derived from
kobs ) k1 + ko, where ko is the rate constant for the forma-
tion of ethylene, and from the yields of ethylene (Table 1),
which under limiting trapping conditions correspond to
ko/(ko+k1).72
It happens that in this particular case the yields of ethylene
depend little on the temperature.73 Consequently, ko ≈ 0.2kobs
q
and k1 ≈ 0.8kobs. As a result, ∆H1 ≈ ∆Hobsq, and the value
of ∆S1q is about 0.4 cal mol-1 deg-1 lower than the value of
q
∆Sobs . The BDE values in a moderate viscosity solvent DMA
q
were calculated as (∆H1 -1) ( 1 kcal/mol.74,20
Homolysis of the ethyl derivative is faster than that of
CH3CoPc, in contrast to the SN2 alkyl transfers from these
complexes.9 Also the activation enthalpies (and BDEs) listed
in Table 9 seem to be lower for CH3CH2CoPc, although the
difference is comparable to the experimental error. The
activation enthalpies for the methyl system are at least 6
kcal/mol higher than those determined for SN2 methyl
transfers from CH3CoPc, while the activation entropies for
the two reactions are not very different; this does not allow
for an easy distinction of the displacement and homolytic
processes.9
The previously determined activation parameters for the
thermolysis of the methyl system under aerobic conditions21
are generally in agreement with the BDEs obtained in this
study in the presence of TEMPO. Apparently, in dilute
solutions of organocobalt substrate, alkyl radical trapping
by O2 is quite effective.
The BDEs of Table 9 should be compared with those
for other models with similar primary alkyl ligands, some
of which are shown in Table 6. The Co-C bond in
CH3CoPc(py) appears to be weaker than the same length
bond in MeCbl (BDE ) 37 ( 318 or 36 ( 4 kcal/mol),75
MeCbi+ (37 ( 3 kcal/mol),75 or CH3Co(DH)2(py) (33.1 (
1.6 kcal/mol).19 Hence, despite the virtually equal Co-C
distances in these complexes, there are visible differences
in the BDEs. Conversely, the Co-C bond in CH3CoPc(py)
is of similar strength or weaker than that in AdoCbl, whereas
in the latter both axial bonds are significantly longer.
It is noteworthy that the BDE for CH3CoPc is smaller than
those for inactive methyl donors, methylcobaloxime and
MeCbl. As follows from Marcus theory, a small BDE
(76) In fact, electrochemical studies by Qiu and Sawyer (ref 17) suggest
extremely slow rates for homolysis of n-BuCoIII(porphyrin) (-∆GBF
> 36 kcal/mol).
(77) Nie, S.; Marzilli, P. A.; Marzilli, L. G.; Yu, N.-T. J. Am. Chem. Soc.
1990, 112, 6084 and references therein.
(78) Riordan, C. G.; Halpern, J. Inorg. Chim. Acta 1996, 243, 19.
(79) Ng, F. T. T.; Rempel, G. L.; Halpern, J. J. Am. Chem. Soc. 1982,
104, 621.
(72) Halpern, J. Pure Appl. Chem. 1983, 33, 1059.
(73) The constant ko/k1 ratio was observed by Halpern and co-workers,
see ref 66.
(74) Garr, C. D.; Finke, R. G. Inorg. Chem. 1993, 32, 4412.
(75) Hung, R. R.; Grabowski, J. J. J. Am. Chem. Soc. 1999, 121, 1359.
Inorganic Chemistry, Vol. 44, No. 26, 2005 9911

Galezowski and Kubicki
complexes than that of their phthalocyanine congeners.
Contrary to that prediction, the BDEs of the alkyl-CoIIITPP
complexes are considerably, ∼10 kcal/mol, lower.80 Steric
interactions in the TPP system could be held accountable
for this, but the very similar BDEs for the methyl and butyl
derivatives seem to refute such explanation. In a further twist,
the Co-C bond dissociation in tropocoronand complexes,
in which the equatorial ligand is strongly basic and the Co-C
bond is 0.068 Å shorter, for the ethyl derivative, than that
in CH3CH2CoIIIPc, occurs under very mild conditions.5 At
present, it would be hard to reconcile all the facts regarding
the cis influence of the equatorial ligand, and it is unlikely
that this could be done, at least in some cases, by invoking
the varied steric interactions of the (nonbulky) axial alkyl
group with the equatorial arrangement.
Effect of the Strength of the Co-Nax Bond on Homoly-
sis. Faced with a lack of correlation of BDE to the Co-C
distance in the ground state, Marzilli et al. concluded that
the only clear correlation to Co-C BDE values is the
Co-Nax bond length3,81 and also that transition state effects
dominate the Co-C cleavage.77 Consequently, a comparison
of the parameters of the Co-Nax bond in the ground state
and in the homolysis product, the latter believed to bear
similarity to the late transition state of the homolytic process,
is of interest.82 The Co-Nax bond in CH3CoIIIPc(py), 2.174-
(2) Å, is within 1 σ of being as long as that in CoIINPc(py),
2.167(9) Å,40 and it is plausible that the Co-Nax bond
in the latter would not be much different from that in
CoIIPc(py). The Co-Nax bond lengths in the CH3CoIII-
(OEP)(py) and CoII(OEP)(4-Me2Npy) couple, 2.214(9) and
2.191(2) Å, respectively, are the same within 2 σ. The crystal
structure of cob(II)alamin shows a relatively short Co-Nax
bond length of 2.131 Å,83 that is, ∼0.03 Å shorter than in
MeCbl.46 A simplistic approach, neglecting the fact that bond
energy may be distributed over many modes and disregarding
the ∼0.1 Å displacement of cobalt from the N4 plane on
formation of the five-coordinate Co(II) structure, would
indicate that there is no strong stabilization of the transition
state by a shorter (stronger) Co-Nax bond in the transition
state than in the ground state that could account for a
significant base-on effect. An analysis based on a comparison
of the formation constants of N-donor adducts with the
organocobalt(III) substrate and the Co(II) product, the latter
representing properties of the transition state, has been
described in detail by Stolzenberg and Summers.84 One order
of magnitude larger association constants of N-donors with
Co(II) than RCo(III) complexes have been found for steri-
cally hindered porphyrins, such as those studied by Stolzen-
berg, and a similar relationship also stems from earlier data
on the ligation of TPP complexes (stability constants for
CH3CH2Co(TPP)py and Co(TPP)py in CH2Cl2 are 8.985 and
48086 M-1, respectively). Hence, the phthalocyanine com-
plexes are expected to exhibit similar behavior. Unfortu-
nately, the formation constant for tetragonal CoIIPc with
N-donors cannot be conveniently measured because of the
π stacking. However, the stability constants for the ring-
substituted derivatives that do not self-associate but prob-
ably have very similar ligand binding properties as plain
CoIIPc87 appear to be larger than those for CH3CoIIIPc: the
association constants of pyridine with CoII(t-Bu4Pc)88 and
CH3CoIIIPc9 in CH2Cl2 are 30 200 and 3630 M-1, respec-
tively; this value is, again, ∼1 order of magnitude in favor
of Co(II) complexes. Differences of this magnitude cor-
respond to a 1-2 kcal/mol reduction in the free-energy
barrier for Co-C bond dissociation in RCoIII(chel)L com-
plexes relative to that of five-coordinate RCoIII(chel) com-
plexes. Consequently, in the absence of trans-steric influence,
no strong base-on effect is predicted. Also, no significant
variation in the rates of homolysis should be expected upon
changing the trans ligand, L, in a RCoIII(chel)(L) complex,
exactly as observed in this work upon replacing very weakly
basic DMA with pyridine or N-MeIm (Tables 6 and 7) and
elsewhere in alkylcobalt and organoiron porphyrins.78,89
Conclusions
A long Co-C bond in coordinatively unsaturated
CH3CH2CoPc, containing an equatorial arrangement known
for its electron deficiency, may indicate the electrophilic
character of the alpha carbon, in accordance with the en-
hanced alkyl (cation) transfer reactivity of the phthalocyanine
model of MeCbl. Not only the alkyl cation transfer but also
the homolytic rupture of the Co-C bond in phthalocyanine
models are faster than those in their alkylcorrinoid, OEP,
and cobaloxime analogues. In accordance with the Marcus
theory, the relatively low BDE for CH3CoPc contributes to
faster rates for methyl transfer from this complex. The
differences in the rates of homolysis between CH3CoPc(py)
and MeCbl are not reflected in the structural parameters of
the axial systems, which are virtually identical for the two
complexes. The faster homolysis of the Pc model than the
OEP model is consistent with the value of the BDE
increasing with the basicity of the equatorial ligand, but this
rule fails for TPP and tropocoronand alkylcobalt(III) com-
plexes. The results of this study reinforce the earlier claim
of Marzilli el al. that, even in rare cases where measurable
differences in the Co-C bond length are observed, the order
of bond strength as defined by BDE cannot be safely
q
(80) However, because of the atypical strongly negative ∆SBD value, the
(85) Kadish, K. M.; Han, B. C.; Endo, A. Inorg. Chem. 1991, 30, 4502.
(86) Kadish, K. M.; Botomley, L. A.; Beroiz, D. Inorg. Chem. 1978, 17,
1124.
(87) This is suggested by the similar ligand binding properties of the CH3-
CoPc and CH3Co[PcR4] complexes, R ) tert-butyl or neo-pentoxyl.
Galezowski, W. Unpublished results.
(88) Kimura, M.; Sakaguchi, A.; Ohtha, K.; Hanabusa, K.; Shirai, H.;
Kobayashi, N. Inorg. Chem. 2003, 42, 2821.
(89) It should be noted, however, that it has been claimed that the BDE
for benzyl-Co(OEP)(L) varies with the basicity of the trans ligand
within several kcal/mol (see ref 23).
value of ∆GBDq is only ca. 4 kcal/mol lower than that for CH3CoIIIPc.
(81) Summers, M. F.; Toscano, P. J.; Bresciani-Pahor, N.; Nardin, G.;
Randaccio, L.; Marzilli, L. G. J. Am. Chem. Soc. 1983, 105, 6259.
(82) In an enzymatic environment, in the absence of diffusion step, there
is probably no transition state. Brown, K. L.; Marques, H. M.
THEOCHEM 2005, 714, 209.
(83) Berendsen, H. J. C.; Postma, J. P. M.; Van Gunsteren, W. F.; DiNicola,
A.; Haak, J. R. J. Chem. Phys. 1984, 81, 3684. (b) Kra¨utler, B.; Keller,
W.; Kratky, C. J. Am. Chem. Soc. 1989, 111, 8936.
(84) Stolzenberg, A. M.; Summers, J. S. Inorg. Chem. 2000, 39, 1518.
9912 Inorganic Chemistry, Vol. 44, No. 26, 2005

Alkylcobalt(III) Phthalocyanine Complexes
predicted from the structural parameters of the axial system
in the ground state. For rigid porphyrinic complexes with
nonbulky axial alkyl groups, this lack of correlation is not
likely to be caused by variable steric interactions. A num-
ber of observations such as equal length Co-Nax bonds in
CH3CoIIIPc(py) and CoIINPc(py) and moderate ratios of the
association constants of N-donor ligands with RCoIII(chel)
and CoII(chel) suggest a weak base-on effect on the Co-C
BDEs. The weak dependence of the BDEs on the identity
of trans ligand observed in this work for phthalocyanine
complexes and elsewhere for iron and cobalt porphyrins is
in agreement with above statement.
Acknowledgment. Support of this work through a grant
from the Polish State Committee for Scientific Research
(Grant 7 T09A 07421) is gratefully acknowledged.
Supporting Information Available: X-ray crystallographic
files (in CIF format) for CH3CH2CoPc (ambient temperature),
CH3CH2CoPc (100 K), and CH3CoPc(py). These files have also
been deposited with Cambridge Crystallographic Data Centre,
CCDC Nos. 234305, 260988, and 23436, respectively. Exemplary
kinetic traces for homolytic cleavage of the Co-C bond in
CH3CH2CoPc measured at several wavelengths and corresponding
end-point plots and a figure showing the Eyring plots for homolyses
of CH3CoPc, CH3CH2CoPc, and their pyridine adducts, in DMA.
This material is available free of charge via the Internet at
http://pubs.acs.org.
The trend of decreasing Co-C bond distance in five-
coordinate RCoIII(chel) complexes with increasing basicity
of the equatorial system, tentatively established in this work,
needs more experimental support.
IC051078P
Inorganic Chemistry, Vol. 44, No. 26, 2005 9913