Structural, NMR, and EPR Studies of S = 1/2 and S = 3/2 Fe(III) Bis(4-Cyanopyridine) Complexes of Dodecasubstituted Porphyrins

Article abstract of DOI:10.1021/ic035010q

The NMR and EPR spectra for three complexes, iron(III) octamethyltetraphenylporphyrin bis(4-cyanopyridine) perchlorate, [FeOMTPP(4-CNPy)₂]ClO₄, and its octaethyl- and tetra-β,β′-tetramethylenetetraphenylporphyrin analogues, [FeOETPP(

Full text of DOI:10.1021/ic035010q

Inorg. Chem. 2004, 43, 757−777
1
3
Structural, NMR, and EPR Studies of S ) /2 and S ) /2 Fe(III)
Bis(4-Cyanopyridine) Complexes of Dodecasubstituted Porphyrins
Liliya A. Yatsunyk and F. Ann Walker*
Department of Chemistry, UniVersity of Arizona, Tucson, Arizona 85721-0041
Received August 27, 2003
The NMR and EPR spectra for three complexes, iron(III) octamethyltetraphenylporphyrin bis(4-cyanopyridine) perchlorate,
[FeOMTPP(4-CNPy)2]ClO , and its octaethyl- and tetra-â,â′-tetramethylenetetraphenylporphyrin analogues, [FeOETPP-
4
(4-CNPy)2]ClO and [FeTC TPP(4-CNPy)2]ClO , are presented. The crystal structures of two different forms of [FeOETPP-
4
6
4
(4-CNPy)2]ClO and one form of [FeOMTPP(4-CNPy)2]ClO are also reported. Attempts to crystallize [FeTC TPP-
4
4
6
(4-CNPy)2]ClO were not successful. The crystal structure of [FeOMTPP(4-CNPy)2]ClO reveals a saddled porphyrin
4
4
core, a small dihedral angle between the axial ligand planes, 64.3°, and an unusually large tilt angle (24.4°) of one of
the axial 4-cyanopyridine ligands with respect to the normal to the porphyrin mean plane. There are 4 and 2 independent
molecules in the asymmetric units of [FeOETPP(4-CNPy)2]ClO crystallized from CD Cl2/dodecane (1−4) and CDCl3/
4
2
cyclohexane (5−6), respectively. The geometries of the porphyrin cores in 1−6 vary from purely saddled to saddled with
15% ruffling admixture. In all structures, the Fe−N distances (1.958−1.976 Å) are very short due to strong nonplanar
p
distortion of the porphyrin cores, while the Fe−N distances are relatively long (∼2.2 Å) compared to the same distances
ax
1
in S ) /2 bis(pyridine)iron(III) porphyrin complexes. An axial EPR signal is observed (g ) 2.49, g ) 1.6) in frozen
|
solutions of both [FeOMTPP(4-CNPy)2]ClO and [FeTC TPP(4-CNPy)2]ClO at 4.2 K, indicative of the low spin (LS, S )
4
6
4
1/2), (dyzdxz)4(dxy)1 electronic ground state for these two complexes. In agreement with a recent publication (Ikeue, T.;
Ohgo, Y.; Ongayi, O.; Vicente, M. G. H.; Nakamura, M. Inorg. Chem. 2003, 42, 5560−5571), the EPR spectra of [FeOETPP-
(4-CNPy)2]ClO are typical of the S ) 3/2 state, with g values of 5.21, 4.25, and 2.07. A small amount of LS species with
4
g ) 3.03 is also present. However, distinct from previous conclusions, large negative phenyl-H shift differences δm − δo
and δm − δp in the 1H NMR spectra indicate significant negative spin density at the meso-carbons, and the larger than
1
expected positive average CH shifts are also consistent with a significant population of the S ) 2 Fe(II), S ) /2
2
porphyrin π-cation radical state, with antiferromagnetic coupling between the metal and porphyrin unpaired electrons.
This is the first example of this type of porphyrin-to-metal electron transfer to produce a partial or complete porphyrinate
radical state, with antiferromagnetic coupling between metal and macrocycle unpaired electrons in an iron porphyrinate.
The kinetics of ring inversion were studied for the [FeOETPP(4-CNPy)2]ClO complex using NOESY/EXSY techniques
4
and for the [FeTC TPP(4-CNPy)2]ClO complex using DNMR techniques. For the former, the free energy of activation,
6
4
∆G , and rate of ring inversion in CD Cl2 extrapolated to 298 K are 63(2) kJ mol-1 and 59 s-1, respectively, while for
q
2
the latter the rate of ring inversion at 298 K is at least 4.4 × 107 s-1, which attests to the much greater flexibility of the
TC TPP ring. The NMR and EPR data are consistent with solution magnetic susceptibility measurements that show S
6
3
) /2 in the temperature range from 320 to 180 K for [FeOETPP(4-CNPy)2]+, while both [FeOMTPP(4-CNPy)2]+ and
3
1
[FeTC TPP(4-CNPy)2]+ change their spin state from S ) /2 at room temperature to mainly LS (S ) /2) upon cooling
6
to 180 K.
Introduction
(pKa(BH+) < 3) have an admixed intermediate-spin, IS, state
(S ) 3/2, 5/2),3 and even in one case a thermal spin equilibrium
Iron(III) porphyrinate complexes with very weakly basic
ligands have been investigated for some time, and a variety
of magnetic and structural behavior has been reported. On
the basis of previous studies of [FeOEP(L)2]+ with various
pyridines, it was found that complexes with strongly basic
pyridines (pKa(BH+) > 8) are low spin, LS, with the
(dxy)2(dxz,dyz)3 electronic ground state,1,2 while on the other
hand, complexes with very weakly basic pyridine ligands
1
5
(S ) /2 T S ) /2).4 For example, [FeOEP(3-ClPy)2]ClO4
(pKa(PyH+) ) 2.83) gives both IS and spin equilibrium
states, depending on the crystal lattice,4 and [FeOEP(3,5-
(1) Innis, D.; Soltis, S. M.; Strouse, C. E. J. Am. Chem. Soc. 1988, 110,
5644-5650.
(2) Walker, F. A.; Reis, D.; Balke, V. L. J. Am. Chem. Soc. 1984, 106,
6888-6898.
(3) Scheidt, W. R.; Osvath, S. R.; Lee, Y. J.; Reed, C. A.; Shaevitz, B.;
Gupta, G. P. Inorg. Chem. 1989, 28, 1591-1595.
(4) Safo, M. K.; Scheidt, W. R.; Gupta, G. P.; Orosz, R. D.; Reed, C. A.
Inorg. Chim. Acta 1991, 184, 251-258.
* To whom correspondence should be addressed. E-mail: awalker@
u.arizona.edu.
10.1021/ic035010q CCC: $27.50 © 2004 American Chemical Society
Published on Web 12/23/2003
Inorganic Chemistry, Vol. 43, No. 2, 2004 757

Yatsunyk and Walker
Cl2Py)2]ClO4 is in a discrete admixed IS state.3 When
[FeTMP(L)2]+ and [FeTPP(L)2]+ complexes are considered,
all of them are LS, regardless of the basicity of the axial
pyridine, and specifically, complexes with weakly basic
pyridines are LS with the (dxz,dyz)4(dxy)1 electronic ground
state.5,6
pyrrole-â positions and little or no spin density at the
porphyrin meso-carbons,13-15 whereas in the case of the less
common ground state, (dxz,dyz)4(dxy)1, negligible spin delo-
calization occurs to the pyrrole-â positions and large spin
delocalization occurs to the porphyrin meso positions.14-16
The reason for the latter is believed to be that the half-filled
dxy orbital can participate in strong porphyrin f Fe π
donation from the filled 3a2u(π) porphyrin orbital if and only
if the porphyrinate ring ruffles strongly, leading to at least a
15° twist of nitrogen pz orbitals away from the heme normal
so that there is a significant pz component in the xy plane.6
It is thus porphyrin f Fe π donation that is believed to
stabilize the ruffled conformation of porphyrin complexes
that have the (dxz,dyz)4(dxy)1 electron configuration; this
mechanism is not available to low spin d6 Fe(II) porphyri-
nates, and thus the same ligands that stabilize the (dxz,dyz)4-
(dxy)1 electron configuration and have very ruffled porphy-
rinate cores and perpendicular axial ligand planes for low
spin Fe(III) have planar porphyrinate cores and parallel axial
ligand planes for low spin Fe(II).17
Octaalkyltetraphenylporphyrins such as OETPP are of
great interest because they combine the peripheral substit-
uents from both OEP and TPP, which results in a new
geometry (FeOEPL2+ complexes are usually close to planar,
the TPP analogues are S4-ruffled and the OETPP analogues
are for the most part saddled18) and, therefore, different
spectroscopic properties. Investigation of highly nonplanar
S4-saddled (OETPP) and S4-ruffled (TiPrP) iron(III) porphy-
rin complexes with weakly coordinated axial ligands (THF,
4-CNPy, Py) has resulted in three recent communications
by Nakamura et al.19-21 that have been interpreted as showing
that these complexes have very pure intermediate spin state.
The major reasons were suggested to be (1) the short Fe-
Np bond length commonly observed in highly deformed
porphyrin complexes; and (2) the weak coordinating ability
of the axial ligands. The first point was supported by
obtaining the crystal structure of [FeTiPrP(THF)2]ClO4. The
average Fe-Np distance of 1.967(12) Å is significantly
shorter than the same distances in the other bis(tetrahydro-
furan)iron(III) porphyrin complexes (1.994, 2.006, and 2.016
Å for [FeOEP(THF)2]ClO4, [FeTEtP(THF)2]ClO4, and [FeTPP-
(THF)2]ClO4, respectively).22-24 In the case of both strongly
The structures of [FeTPP(4-CNPy)2]ClO4 and [FeTMP-
(L)2]ClO4 with L ) 3-ClPy and 4-CNPy have strongly S4-
ruffled distortion of the porphyrin core.5,6 The axial ligands
are held in perpendicular planes over the meso positions (the
observed æ angle7 ranges from 29° to 48°).5,6 It is interesting
to note that [FeTPP(4-CNPy)2]ClO4 has the most extensively
ruffled porphyrin core among any of the bis(pyridine) iron-
(III) porphyrinates, which is manifested by a large deviation
of the meso-carbons from the mean plane ((0.55 Å), as well
as by strong twisting of the pyrrole rings.6 EPR spectra of
the Fe(III)TMP complexes with strongly and moderately
basic pyridines (4-Me2NPy, 4-NH2Py, and 3-EtPy) and even
with 3-ClPy consist of single-feature “large gmax” signals with
g values ranging from 3.40 to as low as 2.89, which are
indicative of the low spin (dxy)2(dxz,dyz)3 electronic ground
state.5 On the other hand, the EPR spectra of bis(4-
cyanopyridine) iron(III) complexes with various porphyrin
ligands, TMP, TPP, TnPrP, TcPrP, TiPrP, and OMTPP,
exhibit axial EPR spectra with g > 2 > g|.5,6,8-10 The
analysis of these g values in the “proper axis system” of
Taylor11 indicates a shift in electronic ground state from
(dxy)2(dxz,dyz)3 to a predominantly (dxz,dyz)4(dxy)1 configuration
as the basicity of the pyridine is lowered. The change in
electronic ground state was attributed to the stabilization of
the filled dxz,dyz orbitals of the iron by (dπ)Fe f L(π*) back-
donation due to the π-acceptor ability of the 4-CNPy ligand.6
In addition, and probably more important, the low basicity
of this pyridine toward the proton (pKa(PyH+) ∼ 1.15,12) leads
to its being an extremely weak σ-donor to Fe(III) as
compared to the porphyrin ring, which can also stabilize the
dxz,dyz orbitals relative to the in-plane dxy metal orbital.
The two different electron configurations of low-spin
Fe(III) porphyrinates result in different patterns of spin
delocalization, as shown by the proton NMR spectra of
representative complexes: in the case of the more common
ground state, (dxy)2(dxz,dyz)3, there is large spin density at the
(5) Safo, M. K.; Gupta, G. P.; Watson, C. T.; Simonis, U.; Walker, F.
A.; Scheidt, W. R. J. Am. Chem. Soc. 1992, 114, 7066-7075.
(6) Safo, M. K.; Walker, F. A.; Raitsimring, A. M.; Walters, W. P.; Dolata,
D. P.; Debrunner, P. G.; Scheidt, W. R. J. Am. Chem. Soc. 1994, 116,
7760-7770.
(7) The æ angle is defined as the angle between the projection of planar
axial ligand onto the porphyrin mean plane and the closest NP-Fe-
NP vector. It can also be defined as the dihedral angle between the
axial ligand plane and the plane formed by Fe, two opposite Ns of
the porphyrin core, and two ligand nitrogens that are coordinated to
Fe(III). It is expected to be close to 45° for ideal ruffled geometry of
the porphyrin core and 0-10° for ideal saddled porphyrin cores with
perpendicular orientation of axial ligands in both cases.
(13) La Mar, G. N.; Walker, F. A. J. Am. Chem. Soc. 1973, 95, 1782-
1790.
(14) Walker, F. A. Proton NMR and EPR Spectroscopy of Paramagnetic
Metalloporphyrins. In The Porphyrin Handbook; Kadish, K. M., Smith,
K. M., Guilard, R., Eds.; Academic Press: San Diego, CA, 2000; Vol.
5, Chapter 36, pp 81-183.
(15) Walker, F. A. Inorg. Chem. 2003, 42, 4526-4544.
(16) La Mar, G. N.; Bold, T. J.; Satterlee, J. D. Biochim. Biophys. Acta
1977, 498, 189-207.
(17) Safo, M. K.; Nesset, M. J. M.; Walker, F. A.; Debrunner, P. G.;
Scheidt, W. R. J. Am. Chem. Soc. 1997, 119, 9438-9448.
(18) Yatsunyk, L. A.; Carducci, M. D.; Walker, F. A. J. Am. Chem. Soc.
2003, 125, 15986-16005.
(19) Ikeue, T.; Saitoh, T.; Yamaguchi, T.; Ohgo, Y.; Nakamura, M.;
Takahashi, M.; Takeda, M. Chem. Commun. 2000, 1989-1990.
(20) Ikeue, T.; Ohgo, Y.; Yamaguchi, T.; Takahashi, M.; Takeda, M.;
Nakamura, M. Angew. Chem., Int. Ed. 2001, 40, 2617-2620.
(21) Ohgo, Y.; Ikeue, T.; Nakamura, M. Inorg. Chem. 2002, 41, 1698-
1700.
(8) Ikeue, T.; Ohgo, Y.; Saitoh, T.; Nakamura, M.; Fujii, H.; Yokoyama,
M. J. Am. Chem. Soc. 2000, 122, 4068-4076.
(9) Ikeue, T.; Ohgo, Y.; Saitoh, T.; Yamaguchi, T.; Nakamura, M. Inorg.
Chem. 2001, 40, 3423-3434.
(10) Ikeue, T.; Ohgo, Y.; Ongayi, O.; Vicente, M. G. H.; Nakamura, M.
Inorg. Chem. 2003, 42, 5560-5571.
(11) Taylor, C. P. S. Biochim. Biophys. Acta 1977, 491, 137-149.
(12) Hansch, C.; Leo, A.; Taft, R. W. Chem. ReV. 1991, 91, 165-195.
(22) Masuda, H.; Taga, T.; Osaki, K.; Sugimoto, H.; Yoshida, Z. I.; Ogoshi,
H. Inorg. Chem. 1980, 19, 950-955.
758 Inorganic Chemistry, Vol. 43, No. 2, 2004

4-Cyanopyridine Complexes of Saddled Iron Porphyrinates
S4-saddled [FeOETPP(THF)2]ClO4 and S4-ruffled [FeTiPrP-
(THF)2]ClO4, S ) /2 electronic states are observed, on the
techniques. The conclusions reached, on the basis of detailed
1- and 2-dimensional H NMR spectroscopy and careful
3
1
basis of NMR, EPR, Mo¨ssbauer, and magnetic data.19 In 1H
NMR spectroscopic studies, the IS state exhibits fairly large
downfield shifts of the pyrrole-â protons (or methylene
protons in the case of the OETPP porphyrin core), indicating
considerable π-symmetry spin density at the â-pyrrole
carbons. In the EPR spectra, the signals around g ∼ 4 and 2
are also consistent with the presence of the IS state. Further,
the IS state is supported by the magnetic moment determined
by liquid (Evans) and solid state (SQUID) methods. The
latter resulted in an almost constant magnetic moment of
3.80 ( 0.09 µB and 3.90 ( 0.10 µB at 50-300 K for
analysis of the chemical shifts and their temperature depen-
dence, thus differ markedly from those presented previously.
Molecular structures have been used in the present work to
help arrive at an understanding of the NMR results, and these
serve to strengthen the case of the current presentation.
Experimental Section
Synthesis. (OETPP)FeCl, (OMTPP)FeCl, and (TC6TPP)FeCl
were synthesized by previously reported methods.18 The conversion
of (OMTPP)FeCl to (OMTPP)FeClO4 was carried out by the
procedure developed earlier in this laboratory:25 72 mg (0.088
mmol) of (OMTPP)FeCl was dissolved in a small amount of freshly
distilled THF (∼5 mL) and heated; before reflux started, 7.3 mL
of 0.012 M solution of AgClO4 in THF (0.088 mmol of AgClO4)
was added at one time. Caution! Perchlorate salts are potentially
explosiVe when heated or shocked. Handle them in milligram
quantities with care. The mixture was refluxed for 15 min, cooled
to ambient temperature, and filtered using a fine or medium glass
frit filter. An equal amount of freshly distilled toluene was added
then, and solvents were removed under vacuum at 50 °C. The
residue was dissolved in a small amount of boiling toluene, and an
equal volume of boiling heptane was added. After 1 day, starlike
crystals formed; they were filtered, washed with cold heptane, and
dried under vacuum at 60 °C for 4 h. (OETPP)FeClO4 and
(TC6TPP)FeClO4 were prepared in a similar fashion. In all three
cases, the crystalline material contained a large amount of toluene.
Therefore, before NMR samples were prepared, the crystals were
dissolved in CH2Cl2, the solvent was removed under reduced
pressure, and the residue was kept under vacuum overnight in order
to remove any traces of toluene.
19
[FeOETPP(THF)2]ClO4 and [FeTiPrP(THF)2]ClO4,20 re-
spectively. When a stronger-coordinating ligand is consid-
ered, namely, 4-CNPy, results interpreted as showing a novel
1
3
spin crossover, LS (S ) /2) T IS (S ) /2), were observed
for [FeOETPP(4-CNPy)2]ClO4,20 and later for [FeOMTPP-
(4-CNPy)2]ClO4 and a related octaalkyltetraphenylporphy-
rinate,10 and a pure LS state with a (dxz,dyz)4(dxy)1 electron
configuration for [FeTiPrP(4-CNPy)2]ClO4, even though in
all cases short Fe-Np distances are expected.19 Therefore,
it appears that the binding ability of the axial ligand is an
important factor in tuning the ground state of strongly
distorted porphyrins. 4-CNPy is an interesting ligand, because
when combined with various porphyrin cores it can result
in LS or IS complexes.5,6,10,20
We have further investigated the NMR, EPR, and solution
magnetic susceptibility of [FeOETPP(4-CNPy)2]ClO4 and
related compounds with different S4-saddled porphyrin cores,
namely [FeOMTPP(4-CNPy)2]ClO4 and [FeTC6TPP-
(4-CNPy)2]ClO4, in order to gain a better understanding of
the factors that influence the ground state and electronic
properties of these complexes. Just at the time this paper
was submitted, a paper was published in this journal that
included NMR and EPR investigations of the same iron(III)
porphyrinates.10 Although some of the basic conclusions
reached in these two studies are the same, others differ
markedly, and some of the experimental results differ as well.
The major focus of this research is the correlation of solution
spectroscopic data with the structural parameters obtained
from X-ray crystallography. In the course of these investiga-
tions, we have obtained evidence that [FeOETPP(4-CNPy)2]-
ClO4 is not a pure IS complex, but rather a new electron
configuration for iron porphyrinates is observed, an S ) 2
(OETPP)FeClO4. 1H NMR (CD2Cl2, 23 °C, 300 MHz, δ, ppm)
13.0 (8H, o), 9.8 (4H, p), 7.2 (8H, m), 0.6 (24H, CH3); the
methylene protons are too broad to be detected at ambient
temperatures but are expected in the downfield region.
(OMTPP)FeClO4. 1H NMR (CD2Cl2, 22 °C, 300 MHz, δ, ppm)
60.6 (24H, CH3), 12.0 (8H, o), 9.44 (4H, p), 7.6 (8H, m).
(TC6TPP)FeClO4. 1H NMR (CD2Cl2, 23 °C, 300 MHz, δ, ppm)
89.8 (16, CH2(R)), 11.4 (8H, o), 9.2 (4H, p), 7.7 (8H, m), 1.7 (16H,
CH2(â)). ESI: [FeTC6TPP]+ m/z ) 884, [35ClO4]- m/z ) 99, and
[37ClO4]- m/z ) 101. In all three samples, ClO4- is not bound (or
weakly bound) to Fe(III); therefore, only one peak is observed for
each phenyl-ortho, and -meta as well as CH3 groups in OETPP
and OMTPP and CH2(â) groups in TC6TPP. Because elemental
analysis was not carried out on these compounds, the solvent content
of the crystalline solids is not known, although the heating to 60°
in a vacuum for 4 h should in principle have removed solvent
molecules.
[FeOETPP(4-CNPy)2]ClO4 samples were prepared by adding a
solution of 5 mg (0.005 mmol) of (OETPP)FeClO4 in 0.5 mL of
CD2Cl2 to 1.6 mg (0.015 mmol) of 4-cyanopyridine in a 5 mm
NMR tube to produce a complex concentration of 10 mM, with 10
mM free 4-CNPy if the complex is fully formed. The mixture was
heated in a warm water bath for 5 min. There was no evident color
change in the greenish-brown solution. Since no free (F) 4-CNPy
peaks were observed in the proton NMR spectra at temperatures
as low as -70 °C (203 K), another 0.0015 mmol of 4-CNPy was
added. With this amount of axial ligand (total [4-CNPy] ) 60 mM),
free 4-CNPy peaks were observed below -30° (243 K), indicating
1
Fe(II) center antiferromagnetically coupled to an S ) /2
3
porphyrin π-cation radical, yielding an overall S )
/
2
complex that can only be differentiated from intermediate
spin (IS) Fe(III) by careful analysis of the proton NMR shifts,
while both EPR spectra and magnetic susceptibility measure-
3
ments are indicative only of the overall spin, S ) /2. The
kinetics of inversion of the saddled porphyrinate complexes
have also been investigated by DNMR and/or NOESY/EXSY
(23) Ogho, Y.; Saitoh, T.; Nakamura, M. Acta Crystallogr., Sect. C 1999,
55, 1284-1286.
(24) Chen, L.; Yi, G.-B.; Wang, L.-S.; Dharmwardana, U. R.; Dart, A. C.;
Khan, M. A.; Richter-Addo, G. B. Inorg. Chem. 1998, 37, 4677-
4688.
(25) Nesset, M. J. M., Ph.D. Dissertation, University of Arizona, 1994.
Inorganic Chemistry, Vol. 43, No. 2, 2004 759

Yatsunyk and Walker
that ligand exchange (dissociation of coordinated ligand and binding
of free 4-CNPy) is fast at higher temperatures. NMR data were
acquired over 5° temperature intervals from 243 to 183 K, and at
the end a greater than 10-fold excess of axial ligand was added
protons in OETPP and TC6TPP complexes, or methyl protons in
the OMTPP complex). The relaxation delays in 2D experiments
were set so that the total recycle time was longer or equal to the T1
of the para-phenyl protons, which are typically the slowest relaxing
protons of the complex, or to the average of the T1’s of the free
ligand protons. NOESY/ROESY spectra were processed using the
Felix 2000 software package (Molecular Simulations). Data were
zero-filled to twice the original data size in both dimensions before
applying Gaussian apodization. After Fourier transformation, base-
line correction was done by detecting the baseline points using the
FLATT procedure,26 and fitting the points to an nth-order polyno-
mial (n can vary from 1 to 5). COSY and DQF-COSY data were
processed using standard VNMR software. Zero-filling and un-
shifted sine bell apodization functions were applied before each of
the two Fourier transforms followed by baseline.
Solution magnetic susceptibilities were determined as a function
of temperature by the method of Evans27,28 in the range from 308
to 180 K with 5 K increments in specially constructed Wilmad
coaxial tubes. Measurements were carried out on 10-14 mM
porphyrin solutions in CD2Cl2 that contained approximately 1% of
Me4Si and 150-250 mM 4-CNPy and the same concentration of
Me4Si and 4-CNPy in the reference (outer) tube. The molar
magnetic susceptibility, øM, for a long cylindrical sample oriented
parallel to the magnetic field, B, is calculated according to the
following equation28
1
and the H 1D NMR experiments were repeated over the same
temperature range. This was done in order to ensure that the amount
of axial ligand in the sample was high enough for full complexation
and that the chemical shifts and ligand exchange observed were
due only to the specific binding properties of the 4-CNPy and not
to its concentration. Chemical shifts, relaxation times, T1, of all
resonances, and the temperature at which free and bound 4-CNPy
peaks appeared in the spectra were essentially the same with excess
ligand.
In the same manner, [FeOMTPP(4-CNPy)2]ClO4 was prepared
starting from (OMTPP)FeClO4. The color of the complex changed
from greenish-brown to cherry red upon addition of 4-CNPy. All
1H 1D experiments were repeated 3 times with increasing ligand
concentration: 29, 48, and 82 mM 4-CNPy. The chemical shift of
the porphyrin CH3 at 295 K changed from 65.9 to 62.6 ppm upon
going from the lowest to highest ligand concentration. Chemical
shifts of all other resonances also changed slightly. However, when
the temperature was lowered below 213 K, all three samples showed
very similar chemical shifts for all resonances. Finally, [TC6TPP-
(4-CNPy)2]ClO4 was prepared by dissolving 2 mg of (TC6TPP)-
FeClO4 and 2 mg of 4-CNPy in 0.3 mL of CD2Cl2 to produce a
sample having [TC6TPPFeClO4] ) 6.7 mM and total [4-CNPy] )
21 mM. Free ligand peaks appeared below 213 K, indicating fast
ligand exchange above this temperature. After NMR data were
ø0‚MW(solute)
∆ν
øM
)
+ 3000
(1)
4π‚ν0‚C
MW(solvent)
1
acquired, a 10-fold excess of axial ligand was added, and the H
1D experiments were repeated. Relaxation times were measured
again and showed no dependence upon axial ligand concentration
for all peaks except for free 4-CNPy, for which T1 increased, with
increase in ligand concentration. In this case, the free ligand peaks
appeared at 223 K and below. All the chemical shifts were the
same for both samples at temperatures below 233 K.
On the basis of the differential effect of ligand concentration,
the binding constants of 4-CNPy to the three iron(III) porphyrinates
decrease in the order OETPP > TC6TPP > OMTPP; however, no
attempt was made to accurately measure the binding constants.
EPR spectra were recorded on a Bruker ESP-300E EPR
spectrometer operating at 9.4 GHz, with 100 kHz field modulation
and equipped with Oxford Instruments ESR 900 continuous flow
helium cryostat. Spectra were obtained in frozen CD2Cl2 solutions
and as crystalline samples at 4.2 K. Microwave frequencies were
measured using a Systron-Donner frequency counter.
where ø0 is the diamagnetic molar susceptibility of the pure solvent
(-46.6 × 10-6 for CD2Cl2), ν0 is the spectrometer frequency in
Hz; ∆ν is the difference in chemical shifts between the Me4Si
signals in the inner and outer tubes of the sample, in Hz; C is the
concentration of porphyrin in mol/L. The value of øM thus obtained
is then corrected for diamagnetic contributions from the porphyrin
anion, axial ligands, and iron using Pascal’s constants,29,30 and
finally, the effective magnetic moment, µeff, is calculated according
to
µeff ) 2.84 ø ′T
(2)
x
M
where øM′ is the corrected molar susceptibility. The accuracy of
the results obtained depends on the accurate knowledge of
concentration and the accurate determination of ∆ν and is never
better than 5%. Because it is possible that at least one solvent
molecule (CH2Cl2) was present in the crystals of each of the iron
porphyrinate complexes of this study, the concentrations of the
complexes could be 9-10% (for one) or 18-20% (for two
molecules of CH2Cl2) less than expected on the basis of the formula
weight of the compounds themselves, meaning that the magnetic
moments of the complexes are likely on the order of 5-10% larger
than shown in Figure 5. The total diamagnetic corrections for
[FeOETPP(4-CNPy)2]+, [FeOMTPP(4-CNPy)2]+, and [FeTC6TPP-
(4-CNPy)2]+ are -618.3 × 10-6, -523.4 × 10-6, and -591.7 ×
10-6, respectively.
1H NMR spectra were acquired using a Varian Unity-300
spectrometer operating at a 1H 299.964 MHz and equipped with a
broad-band inverse probe and variable-temperature unit. All spectra
were recorded in CD2Cl2 in the temperature range 180-308 K and
referenced to the residual solvent peak at 5.32 ppm. In the case of
[FeOETPP(4-CNPy)2]ClO4, other solvents were used as well:
CDCl3 (δr ) 7.24 ppm) in the temperature range from 303 to 330
K and C2D2Cl4 (δr ) 5.91 ppm) in the range from 303 to 363 K.
One-dimensional spectra were collected using the standard one-
pulse experiment with maximum spectral bandwidth of 40 kHz, a
block size of 80 K data points, an approximate 90° pulse, and 100
Porphyrin ring inversion in [FeOETPP(4-CNPy)2]ClO4 is an
example of a symmetrical two-site exchange with equal population
1
transients. Homonuclear H 2D spectra were acquired in States
mode at a number of temperatures between 184 and 308 K using
standard pulse sequences with 512 complex points in the directly
detected dimension and 128 t1 increments. The mixing time in
NOESY/ROESY experiments was set to the average T1 of the
protons that are closest to the paramagnetic center (usually ortho-
phenyl-H, the protons of coordinated 4-CNPy, the methylene
(26) Gu¨ntert, P.; Wu¨thrich, K. J. Magn. Reson. 1992, 96, 403-407.
(27) Evans, D. F. J. Chem. Soc. 1959, 2003-2005.
(28) Sur, S. K. J. Magn. Reson. 1989, 82, 169-173.
(29) Mabbs, F. E.; Machin, D. J. Magnetism and Transition Metal
Complexes; Chapman and Hall: London, 1973.
(30) Kahn, O. Molecular Magnetism; VCH: New York, 1993.
760 Inorganic Chemistry, Vol. 43, No. 2, 2004

4-Cyanopyridine Complexes of Saddled Iron Porphyrinates
of the two forms.31 It was studied using NOESY/EXSY experiments
in the temperature range from 243 to 308 K in CH2Cl2 and from
313 to 333 K in CDCl3. Mixing times, τm’s, were set on the order
of the T1 of the methylene protons, namely from 3 to 10 ms,
depending on temperature. The rate constants were calculated using
the following equation
in random orientation and examined on a Bruker SMART 1000
CCD detector X-ray diffractometer at 170(2) K. All measurements
utilized graphite monochromated Mo KR radiation (λ ) 0.71073
Å) with a power setting of 50 kV and 40 mA. Final cell constants
and complete details of the intensity collection and least squares
refinement parameters for all complexes are summarized in Table
1.
In two cases, a total of 3736 frames at 1 detector setting covering
0° < 2θ < 60° were collected, having an ω scan width of 0.2° and
an exposure time of 20 s per frame. In the case of [FeOETPP-
(4-CNPy)2]ClO4 from methylene chloride/cyclohexane, a total of
7372 frames at two detector settings with 2θ ) -17.50 and -33.5°,
covering 0° < 2θ < 51°, were collected having an ω scan width
of 0.2° and an exposure time of 20 s per frame. The frames were
integrated using the Bruker SAINT software package’s narrow
frame algorithm.33 Initial cell constants and an orientation matrix
for integration were determined from reflections obtained in three
orthogonal 5° wedges of reciprocal space.
All structures were solved using SHELXS in the Bruker
SHELXTL (Version 6.0) software package.34 Refinements were
performed using SHELXL, and illustrations were made using XP.34
Solution was achieved utilizing direct (or in some cases Patterson)
methods followed by Fourier synthesis. Hydrogen atoms were added
at idealized positions, constrained to ride on the atom to which
they are bonded and given thermal parameters equal to 1.2 or 1.5
times Uiso of that bonded atom. Although solvents of crystallization
were deuterated, they were refined as CHCl3 and CH2Cl2 and are
thus listed as such in Table 1. Empirical absorption and decay
corrections were applied using the program SADABS.35 Scattering
factors and anomalous dispersion were taken from International
Tables (Vol C, Tables 4.2.6.8 and 6.1.1.4.). The details of the
individual structure determinations are provided in the Supporting
Information.
Idiag
kex
)
(3)
(Icross + Idiag) × τm
where Idiag and Icross are the intensities (volumes) of the diagonal
and cross-peaks, respectively, and τm is the mixing time used in
the NOESY experiment.31,32 The volumes of the diagonal and cross-
peaks, measured using Varian VNMR software, were averaged
before being used for calculations. All experiments were done 2-3
times at a given temperature with different mixing times to obtain
higher accuracy.
Porphyrin Ring Inversion in [FeTC6TPP(4-CNPy)2]ClO4. For
two-site chemical exchange that is fast on the NMR time scale
(above the coalescence temperature, Tc, where only the averaged
chemical environment of two exchanging species can be detected),
the modified Bloch equations can be simplified to the expression31
π(∆ν)2
2(W* - W0)
kex
)
(4)
where ∆ν is the difference in the chemical shift (in Hz) between
two exchanging species extrapolated to temperatures above Tc, W*
is the width at half-height of the exchange-broadened line, and W0
is the inherent line width at half-height (usually it is measured at
high temperatures where chemical exchange is extremely fast on
the NMR time scale or at very low temperatures in the absence of
chemical exchange). The difference in the denominator indicates
how much the NMR line was broadened due to the presence of
chemical exchange only, eliminating the influence of NMR
parameters such as line broadening due to increased solvent
viscosity, miscalibration of the pulses, truncation of FIDs, and
others. In the case of diamagnetic molecules, all values in eq 4,
except W*, remain constant with temperature. However, for
paramagnetic molecules not only W* but also W0 and ∆ν change
with temperature, and all these changes should be taken into
account.
Results
[FeOMTPP(4-CNPy)2]ClO4 was obtained in one crystal-
line form from the CDCl3/cyclohexane solvent system.
[FeOETPP(4-CNPy)2]ClO4 was obtained in two different
crystalline forms, from CD2Cl2/dodecane and CDCl3/cyclo-
hexane. We were not successful in obtaining the crystal
structure of [FeTC6TPP(4-CNPy)2]ClO4: two attempts at
growing crystals of this complex resulted in two structures
of 5-coordinate (TC6TPP)Fe+ with ClO4- and NO3- as axial
ligands. These structures are not discussed here. Atomic
coordinates, a table of complete bond length and angles,
anisotropic thermal parameters, and hydrogen coordinates
and complete torsion angles for all complexes in this study
are listed in Tables S1-S15 in the Supporting Information.
Molecular Structure of [FeOMTPP(4-CNPy)2]ClO4.
The molecular structure of [FeOMTPP(4-CNPy)2]ClO4 is
displayed in Figure 1 together with the numbering scheme
for the crystallographically unique atoms. The porphyrin is
nonplanar and adopts a predominantly saddled conformation
with some ruffling admixture. This conclusion was drawn
upon close examination of the formal diagram and the linear
The rate constants thus obtained were used for the determination
of the activation parameters, ∆Hq and ∆Sq, of ring inversion. The
dependencies of ln(kexh/kBT) on inverse temperature were plotted
in Origin, and least squares linear regression fitting was applied.
The slope and intercept of the linear fit were assigned to ∆Hq and
∆Sq, respectively, according to the Eyring equation
∆Hq ∆Sq
kexh
ln
) -
+
(5)
kBT
RT
R
where h is Planck’s constant, kB is Boltzmann’s constant, T is
temperature (K), and R is the gas constant.
Structure Determination. General. Crystals were grown by
liquid diffusion methods. In all cases two solvent systems were
used: (1) methylene chloride and dodecane; (2) chloroform and
cyclohexane. Crystals of each complex were mounted on glass fibers
(31) Ernst, R. R.; Bodenhausen, G.; Wokaun, A. Principles of Nuclear
Magnetic Resonance in One and Two Dimensions; Clarendon Press:
Oxford, U.K., 1992; Chapter 9.
(32) Shokhirev, N. V.; Shokhireva, T. Kh.; Polam, J. R.; Watson, C. T.;
Raffii, K.; Simonis, U.; Walker, F. A. J. Phys. Chem. A 1997, 101,
2778-2786.
(33) SAINT Reference Manual Version 6.0; Bruker AXS Inc.: Madison,
WI, 2002.
(34) SHELXTL Reference Manual Version 6.0; Bruker AXS Inc.: Madison,
WI, 2002.
(35) Sheldrick, G. M. SADABS, version 2.3; Universita¨t Go¨ttingen:
Go¨ttingen, Germany, 2002.
Inorganic Chemistry, Vol. 43, No. 2, 2004 761

Yatsunyk and Walker
Table 1. Experimental Details Together with Crystal Data for [FeOMTPP(4-CNPy)2]ClO4, and [FeOETPP(4-CNPy)2]ClO4
[FeOMTPP(4-CNPy)2]-
[FeOETPP(4-CNPy)2]-
[FeOETPP(4-CNPy)2]-
ClO4‚CHCl3
ClO4‚1.5CH2Cl2
ClO4‚2CHCl3
empirical formula
fw
FeN8C65H53O4Cl4
1207.80
FeN8C73.5H71O4Cl4
1328.03
FeN8C74H70O4Cl7
1439.38
T, K
170(2)
170(2)
170(2) K
wavelength, Å
cryst syst
space group
a, Å
b, Å
c, Å
0.71073
monoclinic
P21/n
14.7529(16)
23.115(3)
17.3580(19)
90, 98.505(3), 90
5854.2(11)
4
1.370
0.497
2500
0.33 × 0.25 × 0.15
1.48-27.03°
-18e he 18
-29e ke 29
-22e le 22
68928
0.71073
monoclinic
P21
13.5691(10)
42.461(3)
24.0093(18)
90, 100.699(2), 90
13592.6(17)
8
1.298
0.434
5552
0.37 × 0.30 × 0.20
1.29-22.01°
-14e he 14
-44e ke 44
-25e le 25
106355
0.71073 Å
triclinic
P1h
14.3386(4)
21.7212(5)
24.1113(6)
101.250(1), 99.050(1), 97.391(1)
7174.4(3)
R, â, γ, deg
V, Å3
Z
4
D (calcd), g/cm3
abs coeff., mm-1
F(000)
1.333
0.525
2988
0.27 × 0.20 × 0.17
1.15-23.50°.
-16e he 16
-24e ke 24
-27e le 27
103862
103862 [Rint ) 0.0000]
98.3%
TWINABS
0.9160 and 0.8712
103862/2270/1750
0.650
cryst dimensions, mm3
θ limits
limiting indices
reflns utilized
indep reflns
completeness
abs corr
max and min transm
data/restraints/params
GOF on F2
12736 [Rint ) 0.0925]
99.6%
SADABS
0.9292 and 0.8532
12736/12/773
1.007
33209 [Rint ) 0.1154]
99.54%
SADABS
0.9182 and 0.8558
33209/355/3350
1.026
final R indices
[I > 2σ(I)]
R1 ) 0.0555
wR2 ) 0.1207
R1 ) 0.1142
wR2 ) 0.1458
0.562; -0.426 e/Å3
0.071 e/Å3
R1 ) 0.0707
wR2 ) 0.1675
R1 ) 0.1452
wR2 ) 0.2106
0.973; -0.583 e/Å3
0.078 e/Å3
R1 ) 0.0660
wR2 ) 0.0956
R1 ) 0.2480
wR2 ) 0.1519
0.690; -0.608 e/Å3
0.101 e/Å3
R indices (all data)
largest diff peak and hole
RMS difference density
deviation of the porphyrinato core (Figure 2). The average
deviations for adjacent â-carbons from the 24-atom mean
plane are (1.16 and (0.96 Å, indicating some twisting of
the pyrrole rings, consistent with a small ruffling component
in the molecular structure. This is supported by the deviation
of the meso-carbons, which are above and below the
porphyrin mean plane by (0.24 Å. The dihedral angle, ∆æ,
between the axial ligand planes is 64.3(1)°. However, it
should be taken into account that one ligand is severely tilted
by 24.4° with respect to the normal to the porphyrin plane,
and this interferes with measurements of ƾ. This is,
probably, one of the largest deviations of axial ligands from
the porphyrin normal observed in bis(pyridine or imidazole)
iron(III) porphyrinate complexes, and it appears to be due
to crystal packing forces. In contrast, the tilt of the other
ligand is only 1.6°. The dihedral angles æ between the axial
ligand plane and the plane defined by the closest NP-Fe-
NP vector and N5 and N7 (nitrogens coordinated to Fe) of
the axial ligands are 10.4(1)° and 35.4(1)° for the two ligands,
respectively, or -10.4(1)° and +54.6(1)° angles from the
same NP-Fe-NP vector. While the first angle is typical for
saddled porphyrins with perpendicular orientation of axial
pyridine ligands, the second angle shows substantial deviation
and is more typical of ruffled structures. The dihedral angles
between the phenyl ring planes and the porphyrin mean plane
are 44.6°, 53.7°, 44.2°, and 49.53°. In saddled porphyrins
the phenyl rings are rotated toward the porphyrin mean plane
in order to avoid steric interactions with the â-pyrrole
substituents. Therefore, the value of the dihedral angle
between the phenyl ring planes and porphyrin mean plane
can be used as indication of the degree of saddled distortion;
the smaller the dihedral angle between the phenyl planes and
the porphyrin mean plane the greater the saddled distortion.
The high degree of saddled conformation is associated with
the large dihedral angles between a pair of the adjacent
pyrrole rings (38.6°) and between a pair of opposite pyrrole
rings (56.5°). However, these are smaller than the same
angles for either of the [FeOETPP(4-CNPy)2]ClO4 structures
discussed below.
The average values for chemically equivalent bond
distances and angles with estimated standard deviations are
shown in Figure 2. The average Fe-NP distance is 1.973(2)
Å, and two Fe-Nax distances are 2.215(3) and 2.282(3) Å.
The short Fe-NP distance is an indication of the strong
nonplanar distortion of the porphyrin core, and the long Fe-
3
Nax distance is consistent with the overall S ) /2 state of
the complex (consisting of some degree of S ) 2 Fe(II) and
the same degree of S ) 1/2 porphyrin radical, antiferromag-
netically coupled) observed in solution for [FeOMTPP(4-
CNPy)2]ClO4 (see below).
Molecular Structure of [FeOETPP(4-CNPy)2]ClO4 from
Methylene Chloride/Dodecane. There are 4 unique mol-
ecules in the asymmetric unit of [FeOETPP(4-CNPy)2]ClO4.
The molecular structure of 1 is shown as an ORTEP diagram
(Figure 3) along with the numbering scheme for the
crystallographically unique atoms; all other porphyrins have
similar basic features. The numbering schemes start with 2,
3, and 4 for porphyrins 2, 3, and 4, respectively. The
762 Inorganic Chemistry, Vol. 43, No. 2, 2004

4-Cyanopyridine Complexes of Saddled Iron Porphyrinates
Figure 2. (a) Formal diagram of [FeOMTPP(4-CNPy)2]ClO4, showing
the displacement of the atoms in units of 0.01 Å, from the mean plane of
the 24-atom core. The orientations of the axial ligands with the closest Fe-
Np vector and selected bond lengths and angles are also shown. The numbers
given in parentheses are the estimated standard deviations. (b) Linear
representation of the deviation of the macrocycle atoms from the mean
porphyrin plane.
molecules in [FeOETPP(4-CNPy)2]ClO4, and the results are
displayed in Table 3. According to the NSD method, the
nonplanarity of the porphyrin core can be described in terms
of displacements along the lowest-frequency normal coor-
dinates of the porphyrin macrocycle in order to quantify the
contribution from different types of distortion (saddled,
ruffled, etc.). There are six out-of-plane displacements along
the lowest frequency vibrational modes of the porphyrin
core: B2u(saddle), B1u(ruffle), A2u(dome), Eg(x) (wave(x)),
Eg(y) (wave(y)), and A1u(propeller).36,37 As seen from the
results in Table 3, all molecules display predominantly
saddled conformations with large B2u parameter. The B1u
parameter (ruffling) is small for all molecules except 3, where
it is equal to 0.6489. This results in a 14% ruffling component
in the otherwise saddled structure of 3. The ruffling contribu-
tion is still not as high as in the structure of [FeOMTPP-
(4-CNPy)2]ClO4, where it is equal to 17%. The contributions
from the other types of nonplanar distortion are either small
or absent.
Figure 1. (a) ORTEP diagram of the porphyrin macrocycle of [FeOMTPP-
(4-CNPy)2]ClO4 with the numbering scheme for the unique atoms: top
view. Strong tilting of one of the axial ligands can be seen. (b) ORTEP
plot of [FeOMTPP(4-CNPy)2]ClO4 showing arrangement and numbering
scheme for axial ligands: edge-on view. Porphyrin core geometry is saddled
with a substantial (17%) degree of ruffling reflected in the twisting of pyrrole
rings. Thermal ellipsoids are shown at 50% probability. Hydrogen atoms
have been omitted for clarity.
deviation of core atoms from the 25-atom mean porphyrin
plane for all four molecules is displayed in Figure 4, together
with the axial ligand orientations. The linear representation
of the same deviations is shown in Figure S1A of the
Supporting Information. All four molecules adopt a pre-
dominantly saddled conformation, with alternating pyrrole
rings pointing up and down and meso-C’s being close to or
in the porphyrin mean plane. There are some important
differences in the molecular structures of the four porphyrin
molecules. Molecule 2 is purely saddled with an average
deviation of the meso-C’s from the porphyrin mean plane
of 0.015 Å (see Table 2), while for all other complexes the
average deviations of the meso-C’s are more substantial:
0.078(9), 0.225(11), and 0.103(11) Å for 1, 3, and 4,
respectively. One can see that molecule 3 has the highest
ruffling component of all four molecules. Normal-coordinate
structural decomposition, NSD, calculations were carried
out36,37 using the core coordinates of the four porphyrin
In molecules 3 and 4, the axial ligands are in close to
perpendicular planes (see Figure 4), with actual dihedral
angles, ∆æ, of 84.2° and 87.3°, respectively. But molecules
1 and 2 have substantially smaller dihedral angles of 79.4°
and 67.1°. The latter is one of the smallest angles observed
for saddled iron porphyrinates with planar aromatic ligands,
(36) (a) Jentzen, W.; Ma, J.-G.; Shelnutt, J. A. Biophys. J. 1998, 74, 753-
763. (b) Jentzen, W.; Song, X.-Z.; Shelnutt, J. A. J. Phys. Chem. B
1997, 101, 1684-1699.
(37) http://jasheln.unm.edu/jasheln/content/nsd/NSDengine.
Inorganic Chemistry, Vol. 43, No. 2, 2004 763

Yatsunyk and Walker
87.8° in 1, and 84.0° and 83.0° in 2. But in the molecular
structure of 3 and 4, substantial deviation occurs: the
corresponding dihedral angles are 84.2° and 74.0° in 3 and
87.0° and 78.5° in 4. Similar values to those in 3 are observed
in [FeTMP(4-CNPy)2]ClO4.5
The average dihedral angles between the porphyrinato core
and the phenyl rings are relatively small, 40.1°, 40.6°, 41.1°,
and 42.2° in 1, 2, 3, and 4, respectively, and indicate a high
degree of saddled distortion. For ruffled [FeTPP(4-CNPy)2]-
ClO4, these angles are much higher, with the average value
being 61.8°.6 The other evidence of the highly saddled
distortion in 1-4 is the angle between adjacent and opposite
pyrrole rings. The saddled conformation is associated with
a large dihedral angle between the opposite pyrrole rings.
Angles of 67.0°, 66.0°, 68.1°, and 66.1° are observed for 1,
2, 3, and 4, respectively. Angles between adjacent pyrrole
rings are also large: 46.0°, 44.8°, 46.9°, and 45.0° for 1, 2,
3, and 4, respectively.
The average values for selected bond distances and angles
with estimated standard deviation for 1-4 as well as for 5-6
([FeOETPP(4-CNPy)2]ClO4 from CDCl3/cyclohexane) are
included in Figure 4. Average Fe-NP distances are relatively
short: 1.966(8), 1.967(8), 1.961(8), and 1.958(8) Å for 1,
2, 3, and 4, respectively, indicating a large deformation of
the porphyrin core. In contrast, the Fe-Nax distances are long,
2.191(9) and 2.199(9), 2.167(9) and 2.214(11), 2.228(8) and
2.240(9), and 2.175(9) and 2.214(10) Å for 1, 2, 3, and 4,
respectively. These distances correlate well with the same
distances in [FeOMTPP(4-CNPy)2]ClO4, 5 and 6, and are
3
typical for S ) /2 porphyrinatoiron(III) complexes.
Molecular Structure of [FeOETPP(4-CNPy)2]ClO4 from
Chloroform/Cyclohexane. There are two independent mol-
ecules, 5 and 6, on general positions in the asymmetric unit.
The molecular structure of 5 is displayed in the ORTEP
diagram, Figure S2, together with the numbering scheme for
crystallographically unique atoms. The relative orientation
of the axial 4-CNPy ligands is close to perpendicular, as
observed in molecules 1-4 of the other [FeOETPP(4-
CNPy)2]+ crystalline form. One of the eight ethyl groups
points in the direction opposite from expected, suggesting a
low barrier for ethyl group rotation. Apart from that,
structures 5 and 6 share all the basic features of structures
1-4. Average values for the chemically equivalent bond
distances and angles, deviations of the core atoms from the
porphyrin mean plane, and the orientation of the axial
pyridine ligands are shown in the formal diagram (Figure
4). From Figure 4 and also from the linear display of the
core atom deviations (Figure S1B), one can see that both 5
and 6 adopt close to purely saddled conformation with a
small ruffling admixture, which is reflected in the slight
twisting of the pyrrole rings and in the deviation of the meso-
carbons from the mean plane (0.233(5) and 0.088(5) Å for
5 and 6, respectively). However, 5 is more distorted from
purely saddled geometry than 6. Adjacent pyrrole â-carbons
are above and below the porphyrin mean plane by (1.156-
(5), (1.337(5) Å, and (1.183(5), (1.257(5) Å with differ-
ences of 0.181 Å and 0.074 Å for 5 and 6, respectively. NSD
calculations37 (see Table 3) showed that 5 is overall
Figure 3. (a) ORTEP diagram of porphyrin macrocycle in [FeOETPP-
(4-CNPy)2]ClO4: top view. Only porphyrin molecule 1 is shown. All the
other porphyrins have similar molecular structures and numbering schemes
except that all numbers start with 2, 3, and 4 for porphyrins 2, 3, and 4,
respectively. (b) ORTEP diagram of the porphyrin 1 together with the
numbering scheme for the axial ligands: edge-on view. The porphyrin core
in 1 adopts close to ideal saddled geometry. Thermal ellipsoids enclose
50% probability, and hydrogens are omitted for clarity.
with the notable exception of paral-[FeOMTPP(1-MeIm)2]+,
in which the ligand plane dihedral angle is 19.5°.18 In all
four molecules, 1-4, the axial ligand planes are fairly close
to eclipsing the NP-Fe-NP axis. The æ angles are 16° and
5° in 1, 10° and 15° in 2, 22° and 22° in 3, and 16° and 15°
in 4. Only in the case of 3 are the angles between the NP-
Fe-NP axis and the projection of the axial ligands onto the
porphyrin mean plane somewhat large. This larger angle
contributes to the fact that the molecular structure of 3 has
a substantial degree of ruffling. In comparison, [FeTMP(4-
CNPy)2]ClO45 and [FeTPP(4-CNPy)2]ClO46 complexes adopt
a ruffled geometry, are much less distorted from planarity,
and have axial ligands in close to perpendicular planes
located over the meso carbons. This also holds true for Fe-
(III)TMP structures with other weakly basic pyridines such
as 3-ClPy.5
The 4-CNPy ligands in 1 and 2 are nearly perpendicular
to the porphyrin core with the dihedral angles between the
ligand planes and porphyrin mean plane equal to 88.8° and
764 Inorganic Chemistry, Vol. 43, No. 2, 2004

4-Cyanopyridine Complexes of Saddled Iron Porphyrinates
Figure 4. Formal diagram of [FeOETPP(4-CNPy)2]ClO4 molecules 1-6 showing the displacement of the atoms, in units of 0.01 Å, from the mean plane
of the 25-atom core of each. The orientations of the axial ligands with the closest NP-Fe-NP vector and selected bond lengths and angles are also shown.
Inorganic Chemistry, Vol. 43, No. 2, 2004 765

Yatsunyk and Walker
Table 2. Comparison of Structural Parameters for Complexes in This Study
angle æ between
N-Fe-N axis and
ligand planes, deg
dihedral
angle,
ƾ, deg
av dihedral
angles of
phenyls, deg
compd
Fe-NP, Å
Fe-Nax, Å
av |∆Cm|, Å
(0.240
av |∆Câ|, Å
[FeOMTPP(4-CNPy)2]ClO4‚CHCl3
1.973(2)
2.215(2)
2.282(3)
2.191(9)
2.199(9)
2.167(9)
2.214(11)
2.228(8)
2.240(9)
2.175(9)
2.214(10)
2.203(4)
2.269(5)
2.262(4)
2.267(5)
(1.16;
(0.96
(1.23;
(1.31
(1.25;
(1.27
(1.16;
(1.34
(1.23;
(1.30
(1.16;
(1.34
(1.18;
(1.26
-10.4; 54.6
-5.4; 74.0
+15.2; 79.9
-22.0; 67.6
-15.6; 75.5
-14.5; 67.4
+1.7; 89.9
64.3(1)
79.4(4)
67.1(4)
84.2(3)
87.3(5)
78.6(2)
88.2(2)
48.0
40.1
40.6
41.1
42.2
42.9
42.2
[FeOETPP(4-CNPy)2]ClO4‚CH2Cl2, 1
[FeOETPP(4-CNPy)2]ClO4‚CH2Cl2, 2
[FeOETPP(4-CNPy)2]ClO4‚CH2Cl2, 3
[FeOETPP(4-CNPy)2]ClO4‚CH2Cl2, 4
[FeOETPP(4-CNPy)2]ClO4‚CHCl3, 5
[FeOETPP(4-CNPy)2]ClO4‚CHCl3, 6
1.966(8)
1.967(8)
1.961(8)
1.958(8)
1. 975(4)
1.976(4)
(0.078
(0.015
(0.225
(0.103
(0.233
(0.088
Table 3. Normal-Coordinate Structural Decomposition (NSD)36,37 of Distortion Modes of the Complexes of This Study
B2u,
saddle
B1u,
ruffle
A2u,
dome
Eg(x),
Eg(y),
A1u,
av |∆Cm|,
av |∆Câ|,
ruf/sum
%
compd
wave(x) wave(y) propeller
sum
Å
Å
[FeOMTPP(4-CNPy)2]ClO4‚CHCl3
3.2044 0.6815 0.0442
0.0580
0.0355
0.0251
0.1324
0.0551
0.0706
0.0266
0.0683
0.0440
0.0431
0.0632
0.0410
0.0211
0.0929
0.0206
0.0218
0.0338
0.0276
0.0226
0.0066
0.0084
4.077
(0.24
(1.16; (0.96
(1.23; (1.31
(1.25; (1.27
(1.16; (1.34
(1.23; (1.30
(1.16; (1.34
(1.18; (1.26
17.0
5.0
0.7
14.0
6.5
14.6
6.0
[FeOETPP(4-CNPy)2]ClO4‚CH2Cl2, 1 3.8142 0.2236 0.0299
[FeOETPP(4-CNPy)2]ClO4‚CH2Cl2, 2 3.8141 0.0296 0.0180
[FeOETPP(4-CNPy)2]ClO4‚CH2Cl2, 3 3.7623 0.6489 0.0425
[FeOETPP(4-CNPy)2]ClO4‚CH2Cl2, 4 3.8204 0.2806 0.1003
4.1690
3.9637
4.6769
4.3200
4.5263
4.1726
(0.078
(0.15
(0.225
(0.103
(0.233
(0.088
[FeOETPP(4-CNPy)2]ClO4‚CHCl3, 5
[FeOETPP(4-CNPy)2]ClO4‚CHCl3, 6
3.7632 0.6603 0.0045
3.6813 0.2555 0.1079
somewhat more distorted from planarity (the sum of all
vibrational coefficients, 4.5263, is larger than the corre-
sponding sum for complex 6, 4.1726) and has a higher degree
of ruffled distortion: 15% for 5 versus 6% for 6.
of the 1-4 structures. The Fe-Nax distances, 2.203(4), 2.269-
(5) Å and 2.262(4), 2.267(5) Å for 5 and 6, respectively,
are similar to each other and to the same distances in the
molecular structures of 1-4 (see Table 2).
Axial ligand plane orientations in both 5 and 6 are close
to perpendicular. In 5, the dihedral angle between axial ligand
planes, ∆æ, is 78.6(2)°, forming 14.5(2)° and 22.7(2)° angles
to the closest NP-Fe-NP vector for ligand 1, L1 (N15), and
ligand 2, L2 (N17), respectively. Similar values are observed
for the structure of 6, which has ∆æ of 88.2(2)°. In this case,
the ligands are only 1.7(2)° and 0.1(2)° away from the closest
NP-Fe-NP bond for L1 (N25) and L2 (N27), respectively.
These, together with the small tilt of the axial ligands (89.7-
(1)° and 85.7(2)° for L1 and L2, respectively), create a very
symmetrical environment for the Fe atom in 6. In 5, the axial
ligands are more tilted with respect to the porphyrin mean
plane, forming 82.3(1)° and 78.3(1)° dihedral angles for L1
and L2, respectively.
The observed average dihedral angles of the phenyls are
42.9° and 42.2° for 5 and 6, respectively. The acute dihedral
angles of phenyls are an indication of a severe saddled
distortion of the porphyrin core. The average CPorph-CPh
distances are 1.483(7) and 1.485(7) Å for 5 and 6, respec-
tively.
Solution magnetic susceptibility studies for all com-
plexes were carried out using the Evans method28,29 in CD2Cl2
in the temperature range from +35 to -93 °C with
increments of 5 °C. The effective magnetic moment, µeff, of
a methylene chloride solution of [FeOETPP(4-CNPy)2]ClO4
changes slightly with temperature according to the following
linear dependence: µeff ) (3.87 ( 0.01) - (1.8 ( 0.1) ×
10-3 T µB; R ) 99.0%, where R is the correlation coefficient
(Figure 5 and Supporting Information Figure S3). It is
characteristic of the S ) 3/2 state. SQUID magnetic suscep-
tibility data above 200 K20 support this observation. On the
other hand, the effective magnetic moments of both [FeOMT-
PP(4-CNPy)2]ClO4 and [FeTC6TPP(4-CNPy)2]ClO4 change
drastically with temperature. In the case of [FeOMTPP(4-
CNPy)2]ClO4, at all temperatures above 260 K, µeff ) 3.56
µB, somewhat lower than expected for the S ) 3/2 state (3.87
µB); however, this value depends on the assumption that there
are no solvent molecules in the solid starting material,
(OMTPP)FeClO4 (unfortunately elemental analysis was not
done on these complexes, which were, however, heated in a
vacuum for 4 h before use). If there were one CH2Cl2 per
formula unit, the µeff would be 3.73 µB, while if there were
two it would be 3.91 µB. Ikeue and co-workers obtained a
very similar temperature dependence, but with a limiting
value of µeff of ∼4.3 µB.10 This value seems high, as might
be expected if there were some evaporation of the CD2Cl2
solvent during the measurements. Hence, there are probably
inaccuracies in both measurements, meaning that the values
of these solution measurements of µeff are probably only good
The average dihedral angle between the planes of adjacent
pyrrole rings and the average dihedral angle between the
opposite pyrrole rings are higher for 5 as compared to 6
(47.06° vs 43.67° for adjacent and 68.03° and 64.45°, for
opposite) reflecting a higher degree of saddle distortion in
5. In fact, 6 has the smallest saddled distortion compared to
all other structures of [FeOETPP(4-CNPy)2]+.
The average Fe-NP bond distances for 5 and 6 (1.975(4)
and 1.976(4) Å, respectively) are slightly longer than in any
766 Inorganic Chemistry, Vol. 43, No. 2, 2004

4-Cyanopyridine Complexes of Saddled Iron Porphyrinates
(F) and bound (L) 3,5-H are observed (not shown), implying
a high rate of ligand exchange (dissociation of bound ligand
and binding of free 4-CNPy to OMTPPFe(III)) even at this
low temperature. This is an unusual result, because in most
iron(III) dodecasubstituted complexes with various pyridines
and imidazoles, ligand exchange was undetectable on the
NMR time scale below 203 K.38,42 In the same NOESY
experiment, NOE cross-peaks between phenyl protons are
observed (o-m and m-p). Peak assignment was based on
2D COSY and NOESY results, as well as on the integration
of peaks in the 1D spectra and proton T1 values. Peak
assignments for two different temperatures, 193 and 180 K,
together with relaxation times, T1’s, at 180 K are presented
in Supporting Information Table S16, and chemical shifts
of the protons of the complex are given at 243, 213, and
183 K in Table 4, together with those reported in ref 10. All
T1’s (Supporting Information Table S16) are very short due
to rapid chemical exchange between the spin states and the
ligand on-off (dissociation/association) process.
Proton NMR Studies of [FeOETPP(4-CNPy)2]ClO4.
Proton NMR studies were carried out in CD2Cl2 (from 180
to 303 K), CDCl3 (from 303 to 330 K), and C2D2Cl4 (from
303 to 363 K). Results are presented here for [FeOETPP-
(4-CNPy)2]ClO4 in CD2Cl2 unless otherwise stated. An
example of the 1D 1H spectrum, at 203 K, is shown in Figure
6. Chemical shifts and relaxation times at 233 and 193 K
are presented in Supporting Information Table S17, and
chemical shifts at 243, 213, and 183 K are presented in Table
4, along with those of other complexes of this study. In Table
4, the chemical shifts are also compared to those from refs
10 and 20, and it can be seen that several resonances (δpyrr
for CH2(in) and δo) are quite different for the two investiga-
tions of this complex. We will return to this point in the
Discussion section. The temperature dependence of chemical
shifts is displayed as a Curie plot in Supporting Information
Figure S4B.
Figure 5. Temperature dependence of the effective magnetic moments,
µ
eff for [FeOMTPP(4-CNPy)2]ClO4, [FeOETPP(4-CNPy)2]ClO4, and [FeTC6-
TPP(4-CNPy)2]ClO4 in CD2Cl2 by the Evans method. In the case of
[FeOETPP(4-CNPy)2]ClO4, the results are shown for experiments that were
run on two different days. Slight inconsistency of data can be explained by
some evaporation of the solvent during storage.
to (10%. Despite this, the important point is that both studies
show that the magnetic moment increases with temperature
until it reaches a plateau characteristic of an S ) 3/2 complex
at approximately 260 K. At lower temperatures the spin state
decreases, but a pure LS state cannot be reached over the
temperature range accessible to the liquid-state magnetic
susceptibility measurements (>180 K). In the case of
[FeTC6TPP(4-CNPy)2]ClO4, above 230 K the complex is in
transition from IS, S ) 3/2, to LS, S ) 1/2, and below 230 K
the pure LS state is observed with its characteristic µeff of
1.95 ( 0.05 µB.
Proton NMR Studies of [FeOMTPP(4-CNPy)2]ClO4.
1
An example of the 1D H NMR spectrum of [FeOMTPP-
(4-CNPy)2]ClO4, at 203 K, is shown in Figure 6, together
with the spectra for two other complexes of this study. Note
that the spectra of the three complexes have different
chemical shift scales due to the large difference in chemical
shift ranges among the complexes. NMR experiments were
carried out in the temperature range from 295 to 180 K. The
stability of the bis(4-CNPy) complex of (OMTPP)FeClO4
is extremely low, even in the presence of a large excess of
4-CNPy, and the kinetics of ligand exchange are extremely
rapid. Thus, there are no free (F) or bound (L) 4-CNPy peaks
in its NMR spectrum at 203 K, and the large downfield shift
of the methyl resonances indicates that the majority of the
molecules are in the S ) 3/2 state. At all temperatures above
198 K, no clear evidence of ligand binding occurs, and the
temperature dependence of the chemical shifts for all protons
is not linear (Supporting Information Figure S4A). Only
below 198 K do broad free and bound 4-CNPy peaks begin
to appear. Below that point, the temperature dependence of
the chemical shifts becomes linear but does not extrapolate
to the diamagnetic values at T-1 ) 0 for all resonances. In
2D NOESY experiments carried out at 183 K (20 ms mixing
time), strong chemical exchange cross-peaks between free
Because of rapid ligand exchange kinetics, the free (F)
and bound ligand (L) peaks can be found only below 248
K. Even so, all resonances except CH2(in) show linear
temperature behavior on the Curie plot up to 330 K (in
CD2Cl2 and CDCl3), leading us to believe that at higher
temperatures (from 243 to 330 K), despite rapid ligand ex-
change in the presence of 40 mM excess 4-CNPy, the por-
phyrin complex is bis(4-CNPy) ligated. Also in support of
this conclusion is the fact that the proton NMR spectra of the
perchlorate complex, in the absence of added 4-CNPy, are
extremely broad and very different in appearance from the
spectra in the presence of 4-CNPy; the chemical shifts of
(38) Yatsunyk, L. A. Ph.D. Dissertation, University of Arizona, 2003.
(39) Shokhirev, N. V.; Walker, F. A. J. Phys. Chem. 1995, 99, 17795-
17804.
(40) http://www.shokhirev.com/nikolai/programs/prgsciedu.html.
(41) Yatsunyk, L. A.; Walker, F. A. Manuscript in preparation. The
experiments for (OETPP)FeCl were done at three different magnetic
field strengths, 300, 500, and 600 MHz, and in two different solvents,
CD2Cl2 and C2D2Cl4. All data can be described by single linear
dependence. Thus, a solvation effect is not present.
(42) Ogura, H.; Yatsunyk, L.; Medforth, C. J.; Smith, K. M.; Barkigia, K.
M.; Renner, M. W.; Melamed, D.; Walker, F. A. J. Am. Chem. Soc.
2001, 123, 6564-6578.
Inorganic Chemistry, Vol. 43, No. 2, 2004 767

Yatsunyk and Walker
Figure 6. Proton NMR spectra of (a) [FeOETPP(4-CNPy)2]ClO4, (b) [FeOMTPP(4-CNPy)2]ClO4, and (c) [FeTC6TPP(4-CNPy)2]ClO4 obtained in CD2Cl2
at 203 K. Note that each spectrum has a different scale since large differences in chemical shifts are observed for the three complexes. No free or bound
4-CNPy peaks are observed in the case of [FeOMTPP(4-CNPy)2]ClO4 due to the rapid kinetics of ligand exchange. [FeOETPP(4-CNPy)2]ClO4 is in the S
) 3/2 state with high downfield shifts for both ligand and CH2(out) resonances and following the order for phenyl-H δo > δp > δm. [FeOMTPP(4-CNPy)2]-
3
3
ClO4 is in transition from S ) /2 to LS with S ) /2 being predominant and is characterized by a large downfield shift of CH3 (though smaller than for
CH2(out)) and the same order of phenyl resonances as in [FeOETPP(4-CNPy)2]ClO4. [FeTC6TPP(4-CNPy)2]ClO4 is in the LS state with upfield shifts for
L 2,6 and 3,5-H and δm > δp > δo order of phenyl resonances. Signal assignment: o, m, p correspond to ortho-, meta-, and para-phenyl-H; F and L indicate
free and bound ligand peaks; s represents CH2Cl2.
768 Inorganic Chemistry, Vol. 43, No. 2, 2004

4-Cyanopyridine Complexes of Saddled Iron Porphyrinates
Table 4. Chemical Shifts for Selected Resonances in [FeOETPP(4-CNPy)2]ClO4, [FeOMTPP(4-CNPy)2]ClO4, and [FeTC6TPP(4-CNPy)2]ClO4, in
CD2Cl2
a
porphyrin
δpyrr
δL3,5-H
δL2,6-H
δm
273 K
δp
δo
δm - δp
-7.5
δm - δo
-12.2
OETPP
45.7, 14.0
(s)b
s
s
5.0
12.5
(10.4)b
12.5c
10.3
(7.3)b
10.5c
s
17.2
(6.8)b
(14.4)b
45.8, 17.2c
67.8
5.0c
6.8
14.1c
-7.5c
-3.5
-9.1c
-6.7
OMTPP
TC6TPP
s
s
s
s
13.5
(76.0)b
65.4c
(9.8)b
6.6c
(12.9)b
13.8c
-3.9c
s
-7.2c
s
51.3
s
s
(94.7,87.4)b
62.1c
(10.2,9.6)b
sc
(9.2)b
8.2c
(13.3,10.6)b
c
?
s
s
243 K
4.6
OETPP
OMTPP
TC6TPP
50.9, 14.2
(s)b
66.0
s
103.4
s
13.2
18.3
-8.6
-3.8
+3.8
-13.7
-6.9
+5.8
(6.8)b
(10.4)b
10.4
(14.4)b
13.6
65.8
6.7
(76.0)b
21.5
(9.8)b
(7.3)b
7.3
(12.9)b
5.3
s
s
11.1
(94.7, 87.4)b
(10.2, 9.6)b
(9.2)b
(13.3, 10.6)b
213 K
4.1
OETPP
OMTPP
TC6TPP
57.5, 14.5
(s)b
74.9
117.9
49.4d
-27.8
13.9
19.7
-9.8
-1.7
+5.4
-15.6
-3.6
+8.6
(6.6)b
(10.8)b
9.5
(15.4)b
11.3
53.9
38.1d
-5.6
7.7
(87.9)b
10.4
(10.1)b
12.5
(∼7.2)b
(13.6)b
3.9
7.1
(110.4, 101.5)b
(10.2, 8.5)b
(9.5)b
(14.4, 11.0)b
183 K
3.3
OETPP
OMTPP
TC6TPP
66.1, 14.8
(49.6, 92.6)b
67.6, 21.7c,e
31.4
86.4
14.9
-9.9
137.0
2.9d
14.8
21.3
-11.5
-18.0
(∼6.3)b
(11.4)b
15.0c,e
7.5
(∼16)b
14.8c,e
6.8
3.3c,e
-11.7c,e
+2.9
-11.5c,e
+3.5
10.3
(114, 94.4)b
30.2c,e
(10.5)b
(∼7.1)b
7.4c,e
7.0
(14.6)b
10.6c,e
6.4c,e
+3.2c,e
+6.7
+4.2c,e
+10.8
6.8
-35.5
13.7
2.9
(123.7, 113.7)b
13.3c,e
(10.2, 9.3)b
(no m-H)c,e
(9.7)b
6.6c,e
(15.6, 11.5)b
4.0c,e
(no m-H)c,e
(no m-H)c,e
a δpyrr indicates CH2(out) and CH2(in), in that order, for OETPP, CH2(R) for TC6TPP, and CH3 for OMTPP. b Chemical shifts for perchlorate complex,
in CD2Cl2. c Taken from ref 10. d Calculated from the Curie plot (Figure S3A). e Reference 10, average of chemical shifts at 193 and 173 K.
the perchlorate complexes at four temperatures are included
in Table 4. Intercepts at T-1 ) 0 are close to diamagnetic
values for the porphyrin CH3 and phenyl (o, m, and p) reso-
nances, indicating that the S ) 3/2 state is the only spin state
accessible to this complex. In line with this, use of the 2-level
temperature-dependent fitting program created by Dr. Nikolai
V. Shokhirev in this laboratory39,40 indicated that the tem-
perature-dependence was best represented by a 1-level fit;
i.e., there is no thermally accessible excited state.
Measurements of spin-lattice relaxation times, T1’s, were
done at all temperatures. All resonances fit into one of two
groups according to their T1’s. Protons of CH2(out), CH2(in),
porphyrin CH3, phenyl-o, L 3,5-H, and L 2,6-H belong to
the first group, with very short T1 values (1-10 ms). In fact,
the T1 of L 2,6-H was too short to obtain accurate measure-
ments. Protons of phenyl-m and -p belong to the second
group, with substantially longer T1’s (17-75 ms). In both
groups, T1 decreases either linearly or nearly so with
decreasing temperature, except in the case of L 3,5-H, where
a strongly curved decrease of T1 is observed (see Supporting
Information Figure S5). Relaxation times for F 2,6-H and F
3,5-H increase with decreasing temperature due to the
common fact that in the absence of chemical exchange the
T1’s of small molecules increase as the temperature is
lowered.
Even at high temperatures two distinct resonances can be
observed for the methylene protons of the ethyl groups, and
both COSY and NOESY/EXSY spectra show the geminal
and/or chemical exchange coupling of these protons; an
example of chemical exchange coupling is shown in Figure
7. For an adjacent pair of ethyl groups, one resonance, CH2-
(out), is due to the methylene protons that point away from
each other and toward the phenyl rings, and the other, CH2-
(in), is due to the protons that point toward each other. These
peaks are good probes for studying ring inversion in
[FeOETPP(4-CNPy)2]ClO4 using NOESY/EXSY techniques
by measuring the volumes of diagonal and cross-peaks
between these two methylene resonances. An example of
the downfield part of the NOESY spectrum (288 K, 3 ms
mixing time) is shown in Figure 7. Ring inversion is
detectable on the NMR time scale only above 243 K and
becomes very slow below this temperature.
Determination of the Rates of Porphyrin Ring Inver-
sion for [FeOETPP(4-CNPy)2]ClO4. NOESY/EXSY data
were obtained from experiments run on the complex contain-
ing the highest concentration of 4-CNPy, 82 mM, from 243
to 330 K in intervals of 5 K in CD2Cl2 and in CDCl3. The
diagonal- and cross-peak volumes of the methylene peaks
were used to obtain rates of ring inversion and activation
parameters for this process (see Experimental Section). Apart
Inorganic Chemistry, Vol. 43, No. 2, 2004 769

Yatsunyk and Walker
Figure 8. Eyring plot of the kinetic data obtained for the porphyrin ring
inversion process in [FeOETPP(4-CNPy)2]ClO4 in CD2Cl2 and CDCl3.
Calculated values of ∆Hq ) 32 and 35 kJ mol-1 and ∆Sq ) -104 and
-110 J mol-1 K-1 were determined for the ring inversion process in CD2-
Cl2 and CDCl3, respectively.
Figure 7. Downfield part of NOESY/EXSY spectrum of [FeOETPP(4-
CNPy)2]ClO4 at 288 K in CD2Cl2. It was acquired with a spectral width of
20 kHz, 512 × 128 complex points, 80 transients per t1 increment, a 3 ms
mixing time, and 50 ms relaxation delay between scans. The spectrum was
processed after application of Gaussian apodization (14 ms and 3 ms in the
1st and the 2nd dimension, respectively). Note that the spectral window is
substantially broader than required by the spread of the proton resonances;
this is due to the fact that the CH2(out) peak is at the edge of the spectral
window. It is attenuated due to resonance offset effect, and no accurate
measurements of its volume can be obtained. Therefore, the spectral width
was increased and transmitter offset was set to the position exactly between
the two methylene peaks to ensure their equal excitation by the 90° 1H
pulse. Volumes of cross-peaks were measured within the boxes shown, and
the average of the two equivalent cross-peaks was used.
-110(12) J mol-1 K-1, ∆Gq
) 68(7) kJ mol-1, and
298
kex(298) ) 8.6 s-1. The free energy of activation at 298 K is
similar to that of other OETPPFe(III) complexes with planar
aromatic axial ligands (55(3), 60(1), and 56(3) kJ mol-1 at
298 K for [FeF20OETPP(4-Me2NPy)2]Cl, [FeOETPP-
(1-MeIm)2]Cl, and [FeOETPP(4-Me2NPy)2]Cl, respec-
tively).38,41 However, the entropy and enthalpy of activation
differ substantially. ∆Hq is lower than that for any other
OETPPFe(III) complex, which is probably due to the long
Fe-Nax bonds (2.167-2.240 Å) in [FeOETPP(4-CNPy)2]-
ClO4 (vide supra) as compared to those in [FeOETPP-
(1-MeIm)2]Cl, (1.976(3) and 1.978(3) Å)42 and [FeOETPP-
(4-Me2NPy)2]Cl (1.984(5) and 2.099(12) Å).42 Molecular
mechanics calculations of the barrier of porphyrin ring
inversion would be required in order to rationalize the results
obtained.
from CE cross-peaks, strong NOE cross-peaks of negative
phase (opposite to the phase of both diagonal peaks and CE
cross-peaks) were observed between diastereotopic methyl-
ene protons over the temperature range of the NOESY
experiments. This negative intensity decreases the volume
of CE cross-peaks substantially in the case of slow ring
inversion and results in underestimated rates of ring inver-
sion. For this reason, points obtained for 243 and 253 K were
not used for the calculation of kinetic parameters. Using peak
volumes, the rate of ring inversion was calculated according
to eq 3, and the results are presented in the form of an Eyring
plot in Figure 8. Data obtained for two different solvents,
CD2Cl2 and CDCl3, show a small solvation effect. This is
unlike the case of (OETPP)FeCl,38,41 where no solvation
effect was observed and data for two different solvents were
described by one linear fit. However, an influence of solvent
(CD2Cl2 vs C2D2Cl4) on ring inversion activation parameters
was observed for (F20OETPP)FeCl.38,41 According to the
results obtained for [FeOETPP(4-CNPy)2]ClO4 in CD2Cl2
solution, the following activation parameters were deter-
mined: enthalpy, ∆Hq ) 32.0(8) kJ mol-1, entropy, ∆Sq )
The solvation effect is reflected not only in the activation
parameters, but also in the relaxation times of protons and
their chemical shifts in the different solvents, CD2Cl2, CDCl3,
and C2D2Cl4 (see Supporting Information Table S18). The
latter solvent was used in an attempt to obtain the rates of
ring inversion at high temperatures, but due to weak binding
of 4-CNPy and therefore extremely short relaxation times,
T1’s, NOESY results were not meaningful under these
conditions. T1 relaxation times for protons in [FeOETPP(4-
CNPy)2]ClO4 decrease with an increasing number of Cl
atoms in the solvent molecule. However, there is no direct
dependence of chemical shifts on the nature of the solvent;
the values of chemical shifts differ by e2.5 ppm among the
different solvents.
From 2D data,43 it was found that ligand exchange
becomes slow on the NMR (NOESY/EXSY) time scale
below 203 K. The NOE crossover point for the NOESY
spectra occurs at 223 K; below this temperature the NOE
cross-peaks are positive (in phase with the diagonal peaks).
As the temperature decreases, the NOE intensity increases:
-104(3) J mol-1 K-1, the free energy of activation, ∆Gq
298
) 63(2) kJ mol-1, and the rate of ring inversion at 298 K,
kex(298) ) 59 s-1. The same parameters in CDCl3 solvent
are slightly different: ∆Hq ) 35(4) kJ mol-1, ∆Sq
770 Inorganic Chemistry, Vol. 43, No. 2, 2004
)

4-Cyanopyridine Complexes of Saddled Iron Porphyrinates
for example, for methylene protons at 203 K the average
volume of the NOE cross-peaks is 1.2% of the volume of
the diagonal peak, while at 183 K it is 3.7%.44
is much lower than in the case of the [FeTC6TPP(4-Me2-
NPy)]Cl complex, where ligand exchange becomes too slow
to detect on the NMR time scale at 213 K.38,41 Such a
difference can be attributed to weaker basicity and binding
ability of 4-CNPy as compared to 4-Me2NPy. NOE cross-
peaks are observed in the NOESY spectra between the
phenyl-m and -p, and between CH2(R) and CH2(â). Only
weak NOE cross-peaks are observed between the phenyl-m
and -o due to the extremely short relaxation time, T1, of the
phenyl-o protons (see Supporting Information Table S19).
Determination of the Rates of Porphyrin Ring Inver-
sion for [FeTC6TPP(4-CNPy)2]ClO4. Besides rapid ligand
exchange, [FeTC6TPP(4-CNPy)2]ClO4 undergoes fast ring
inversion that is accompanied by ligand rotation. The
presence of one CH2(R) peak in all NMR spectra indicates
that ring inversion is very rapid at all temperatures. In the
limiting case of slow ring inversion, two resonances are
expected for the diastereotopic CH2(R) protons of this
porphyrinate complex (CH2(R)-up for methylene protons that
point toward the axial ligand and CH2(R)-down for the
protons that point away from the axial ligand). However,
lowering the temperature results only in broadening of the
CH2(R) peak: at 218 K the width of this peak is 31 Hz, but
at 180 K it has broadened to 74 Hz. NMR spectra were taken
every 3-5 K between these two temperatures. The data
obtained can be used only for very rough estimation of the
rate of ring inversion, since usually the CH2(R) peak broadens
to thousands of hertz, as in the case of (TC6TPP)FeCl and
[FeTC6TPP(4-Me2NPy)2]Cl,38,41 before it splits into two
resonances as the temperature is lowered further. Temper-
ature dependence of the rate of ring inversion in [FeTC6-
TPP(4-CNPy)2]ClO4 was estimated using the standard pro-
cedure of line shape analysis according to eq 4.45 The
temperature dependences of each component in eq 4, ∆ν,
W*, and W0, have to be taken into account. Since W0 is
proportional to the rotational correlation time of the molecule
(τc), which is in turn proportional to solvent viscosity (η/T),
the line width of the CH2(R) resonance was plotted against
1000η/T and the analytical expression for the inherent line
width, W0, (line width not influenced by chemical exchange)
was obtained from linear fitting of the nearly flat, high
temperature part of the plot (from 218 to 200 K): W0 )
24.66 + 2380η/T. Extrapolation of this line into the regime
where ring inversion becomes measurable on the NMR time
scale (below 198 K) allows us to calculate the increase in
line width, W* - W0, that is the due to the chemical exchange
only, excluding the effect of increased solvent viscosity
which slows molecular motion and also contributes to line
broadening. However, since the sample cannot be cooled to
the temperature where two different resonances for CH2(R)
are observed, the difference in chemical shift, ∆ν, cannot
be measured. We have assumed that the minimum ∆ν must
be at least as large as the maximum line width (74 Hz) at
180 K (in reality values of ∆ν at least as high as 1000 Hz
could be expected). If larger values of ∆ν are considered,
the calculated rate constant would simply be scaled by the
Proton NMR studies of [FeTC6TPP(4-CNPy)2]ClO4
were carried out over the temperature range from 295 to 180
K. It was found that the stability of the bis(4-CNPy) complex
is relatively low and chemical exchange of ligands is rapid,
so free (F) and bound (L) 4-CNPy peaks appear only below
223 K in the presence of 7.6 mM excess 4-CNPy (21 mM
total). As the temperature is lowered, binding of 4-CNPy to
(TC6TPP)FeClO4 becomes more favorable, and the exchange
rate decreases. Finally, at 183 K, ligand exchange becomes
undetectable on the NMR time scale in the NOESY experi-
ments (there are no exchange cross-peaks between the bound
and free ligand). Even at 183 K, however, all peaks in the
spectra remain fairly broad.
An example of the 1D 1H spectra for [FeTC6TPP-
(4-CNPy)2]ClO4 is shown in Figure 6. Chemical shifts and
relaxation times for all peaks at 213 and 183 K are presented
in Supporting Information Table S19, and chemical shifts
at 273, 243, 213, and 183 K are presented in Table 4, where
they are compared to those of the other two complexes of
this study and to those reported previously for a closely
related complex, [FeTBTXP(4-CNPy)2]+.10 As can be seen,
there is a large difference in the R-CH2 shifts and smaller
differences in shifts for o-H and p-H resonances for the
complex of this study and that reported previously.10 The
possible reasons for these differences will be considered
below in the Discussion section. The temperature dependence
of the chemical shifts is displayed in the form of a Curie
plot (Supporting Information Figure S4C). Due to low
complex stability, at relatively high temperatures all lines
in the Curie plot are curved. Only the low temperature part
of the graph (below 203 K), when the axial ligand is fully
bound and the complex is fully low spin, is described by a
straight line in all cases except for CH2(R). Even so,
diamagnetic intercepts (T-1 ) 0) are only observed for the
CH2(â) and phenyl-p protons. Detailed analysis of the
temperature dependence is presented in the Discussion
section.
The spin-lattice relaxation times, T1’s, were measured at
all temperatures below 263 K. From 263 to 203 K, T1 values
increased for all peaks due to stronger ligand binding, and
then, from 203 to 183 K, T1 decreased for all peaks that are
close to the paramagnetic center (see Supporting Information
Table S19). The latter is the usual behavior for protons in
paramagnetic complexes in the absence of chemical ex-
change.
Two-dimensional experiments were done in the temper-
ature range from 233 to 183 K. In the NOESY spectra at all
temperatures, two sets of CE cross-peaks are observed: F-L
3,5-H and F-L 2,6-H, indicating fast ligand exchange, which
becomes NMR-undetectable below 183 K. This temperature
(43) Only NOESY experiments were carried out since the spectral window
was too broad for the performance of any spin-lock experiments such
as ROESY or TOCSY.
(44) The NOE effect was measured as the following ratio: V(CH2(out)-
cross)/V(CH2(out)diagonal).
(45) Polam, J. R.; Shokhireva, T. Kh.; Raffi, K.; Simonis, U.; Walker, F.
A. Inorg. Chim. Acta 1997, 263, 109-117.
Inorganic Chemistry, Vol. 43, No. 2, 2004 771

Yatsunyk and Walker
[FeTC6TPP(4-CNPy)2]CO4 both have typical axial-type EPR
spectra with g ) 2.49, g| ) 1.6 (Supporting Information
Figure S7), characteristic of the LS Fe(III) (dxz,dyz)4(dxy)1
ground state;5,6,8,10 these g-values have previously been
reported as 2.52 and 1.82 for [FeOMTPP(4-CNPy)2]+ and
2.49 and unresolved for [FeTBTXP(4-CNPy)2]+.10
Discussion
Spin States of the Complexes. The EPR data indicate that
at 4.2 K both [FeOMTPP(4-CNPy)2]ClO4 and [FeTC6TPP-
(4-CNPy)2]ClO4 complexes are low spin (LS) with axial EPR
signals (g ) 2.49, g| ) 1.60) that are characteristic of the
(dxz,dyz)4(dxy)1 electronic ground state. Similar EPR spectra
are observed for bis(4-cyanopyridine)iron(III) complexes
with various porphyrinates, TMP (g ) 2.53, g| ) 1.56),5
TPP (g g 2.62, g| e 0.92),6 TnPrP (g ) 2.46, g| ) 1.68),8
TcPrP (g ) 2.49, g| ) 1.58),8 TiPrP (g ) 2.41, g| ) 1.79),8
OMTPP (g ) 2.52, g| ) 1.82),10 and TBTXP (g ) 2.49,
g| not observed).10 Iron(III) porphyrinate complexes with
other weak σ-basicity or π-acceptor axial ligands (Py,
3-CNPy, CN-, t-BuNC) display axial EPR spectra, as
well.5,8-10,46 Utilizing the proper axis definition of Taylor11
(gzz ) -g|, gyy ) -gxx ) g ), the crystal field splittings for
bis(4-cyanopyridine)iron(III) complexes were calculated from
the EPR data. The rhombicity, V/λ, is always 0.0 due to
degeneracy of the filled dxz, dyz orbitals, and the tetragonality,
∆/λ, is -3.12, -3.12, -2.92, -1.72, -3.50, -3.05, and
-4.26 for OMTPP, TC6TPP (both this work), TMP,5 TPP,6
TnPrP,8 TcPrP,8 and TiPrP,8 respectively. The absolute value
of the tetragonality, |∆/λ|, for representative bis(t-BuNC) (a
good π-acceptor) complexes is substantially higher (6.93-
11.3).9,46-48 The relatively small values of the tetragonality
for bis(4-CNPy)iron(III) complexes indicates a much smaller
splitting of the former t2g orbitals (the dxy orbital is only 2.9-
4.3λ higher than dxz,dyz) and a possible switch to the other
LS electron configuration, (dxy)2(dxz,dyz)3. On the other hand,
the large difference in the energy between the dxy and dxz,dyz
orbitals of the bis(t-BuNC) porphyrin complexes ensures the
(dxz,dyz)4(dxy)1 ground state at all temperatures. On the basis
of the EPR data for [FeOMTPP(4-CNPy)2]ClO4 and
[FeTC6TPP(4-CNPy)2]ClO4, we have also calculated the
values of the mixing coefficients a, b, and c for the wave
functions dxz, dyz, and dxy, respectively: a ) b ) 0.1712,
and c ) 0.9577; thus c2 ) 0.9172 and a2 + b2 + c2 ) 0.9758,
indicating the need for an orbital reduction factor, k, of 1.025.
As a result, the ground states of [FeOMTPP(4-CNPy)2]ClO4
and [FeTC6TPP(4-CNPy)2]ClO4 have 94% dxy character at
4.2 K. It is noteworthy that the ∑g2 for bis(4-CNPy) iron-
(III) porphyrin complexes (14.58-15.2)5,6,9 is far from the
theoretical value of 16, indicating partial quenching of orbital
angular momentum in the (dxz,dyz,)4(dxy)1 ground state con-
figuration. However, the g-values reported by Ikeue et al.10
Figure 9. 4.2 K EPR spectra of [FeOETPP(4-CNPy)2]ClO4 (a) in a frozen
CD2Cl2 solution, (b) as a polycrystalline sample obtained from CD2Cl2/
dodecane, and (c) as a polycrystalline sample obtained from the CDCl3/
cyclohexane. The g values of 5.21, 4.25, and 2.07 are characteristic of the
intermediate spin state, S ) 3/2, while g ) 3.03 suggests LS with (dxy)2-
(dxz,dyz)3 configuration. All spectra have the same basic features with
different intensity of signals.
constant quantity of (∆ν/74)2, and as a result, our assumption
would not affect ∆Hq, although it will affect ∆Sq and kex.
With the assumption that ∆ν ) 74 Hz, the temperature
dependence of kex was obtained in the range from 193 to
180 K, and an Eyring plot was constructed (Supporting
Information Figure S6). Linear fitting of the data resulted in
the following kinetic parameters: ∆Hq ) 24(4) kJ mol-1,
∆Sq ) -20(21) J mol-1 K-1, ∆Gq ) 30(10) kJ mol-1,
298
and the minimum rate, kex, of 3.8 × 107 s-1 at 298 K. For
comparison, ambient temperature rates of ring inversion in
(TC6TPP)FeCl and [FeTC6TPP(4-Me2NPy)2]Cl are 4.4 × 107
and 2.8 × 107 s-1, respectively.38,41 Substantially smaller
broadening of the CH2(R) peak in [FeTC6TPP(4-CNPy)2]-
ClO4 as compared to the same peak for the chloride or bis-
(4-Me2NPy) analogues38,41 indicates that the rate of ring
inversion in the bis(4-CNPy) complex should be at least as
high as 4.4 × 107 s-1. This is about a million times faster
than for [FeOETPP(4-CNPy)2]ClO4 in CD2Cl2 at the same
temperature (vide supra).
EPR Studies of Polycrystalline and Solution Samples
of [FeOETPP(4-CNPy)2]ClO4, [FeOMTPP(4-CNPy)2]-
ClO4, and [FeTC6TPP(4-CNPy)2]ClO4. The EPR spectra
of a frozen solution and two polycrystalline samples of
[FeOETPP(4-CNPy)2]ClO4 are shown in Figure 9. All three
spectra have the same basic features, with g ) 5.21, 4.25,
3.03, and 2.07. Ikeue and co-workers report somewhat
different g-values of this complex of 4.28, 3.80, and 2.08.20
The signals with g values larger than 4, plus the value near
2.0, are characteristic of the intermediate (S ) 3/2) spin state,
while the g ) 3.03 peak is typical of low spin (LS)
ferrihemes. However, it is not accompanied by other features
indicative of LS Fe(III) and, thus, may indicate a single
feature “large gmax” signal similar to those observed for other
lower-basicity pyridine complexes of TMPFe(III) with
unusually low gmax values.5 [FeOMTPP(4-CNPy)2]ClO4 and
(46) Yatsunyk, L. A.; Walker, F. A. Submitted for publication.
(47) Walker, F. A.; Nasri, H.; Turowska-Tyrk, I.; Mohanrao, K.; Watson,
C. T.; Shokhirev, N. V.; Debrunner, P. G.; Scheidt, W. R. J. Am.
Chem. Soc. 1996, 118, 12109-12118.
(48) Simonneaux, G.; Schu¨nemann, V.; Morice, C.; Carel, L.; Toupet, L.;
Winkler, H.; Trautwein, A. X.; Walker, F. A. J. Am. Chem. Soc. 2000,
122, 4366-4377.
772 Inorganic Chemistry, Vol. 43, No. 2, 2004

4-Cyanopyridine Complexes of Saddled Iron Porphyrinates
3
for [FeOMTPP(4-CNPy)2]+ (2.52, 1.82) lead to ∑g2
)
freezing the S ) /2 state polycrystalline sample of [FeO-
16.013, indicating no quenching of orbital angular momen-
tum. These values are different from the ones obtained in
this study (2.49, 1.60, ∑g2 ) 14.960).
ETPP(4-CNPy)2]ClO4 in an EPR tube to 4.2 K leads to spin
change for some molecules and as a result, a “large gmax
”
signal, especially, taking into account that no other geometry
changes, except shortening of Fe-Nax bonds, is required.
Of course, one would have expected that the LS state of
[FeOETPP(4-CNPy)2]ClO4 should have the (dxz,dyz,)4(dxy)1
rather than the (dxy)2(dxz,dyz,)3 electronic configuration;
however, neither we nor others20 have observed an axial
signal (g ∼ 2.6-2.5, g| ∼ 1.6) in the EPR spectrum of
[FeOETPP(4-CNPy)2]ClO4, but rather, we observed the
“large gmax” signal. Finally, another argument in support of
the (dxy)2(dxz,dyz,)3 ground state is that the saddled deviation
of [FeOETPP(4-CNPy)2]ClO4 is the highest among OETPP,
OMTPP, and TC6TPP complexes and probably one of the
highest observed thus far, and as was pointed out earlier,9
saddled complexes appear to resist switching to the
(dxz,dyz,)4(dxy)1 ground state. The EPR spectrum for [FeO-
ETPP(4-CNPy)2]ClO4 was reported previously with some-
what different g values of 4.28, 3.80, and 2.08.20
To summarize, it is shown in this work and by others5,6,9,10
that the stability of the (dxz,dyx)4(dxy)1 state in bis(4-CNPy)
iron(III) complexes increases in the order OETPP , TPP
< TMP e OMTPP ∼ TC6TPP ∼ TcPrP < TnPrP < TiPrP.
On the basis of 4.2 K EPR spectra, we also suggest that
the LS ground state in [FeOETPP(4-CNPy)2]+, observed for
a minor population of molecules, has the (dxy)2(dxz,dyz)3
configuration. Hence, EPR and solution magnetic suscepti-
bility measurements are in good agreement with each other.
1H NMR Studies of [FeOETPP(4-CNPy)2]ClO4, [Fe-
OMTPP(4-CNPy)2]ClO4, and [FeTC6TPP(4-CNPy)2]ClO4.
The first observation made in the NMR studies was that the
binding constant of 4-CNPy decreases in the order OETPP
> TC6TPP > OMTPP. Free ligand peaks are observed below
-30, -50, and -75 °C for OETPP, TC6TPP, and OMTPP,
respectively, in the presence of a high molar excess of
4-CNPy. In accord with that, ligand exchange becomes slow
on the NMR time scale at -70, -90, and much below -93
°C for OETPP, TC6TPP, and OMTPP, respectively. Despite
this, we were able to obtain crystals of [FeOMTPP(4-CNPy)2]-
ClO4, but any attempts to obtain crystals of [FeTC6TPP-
(4-CNPy)2]ClO4 resulted in the five-coordinate perchlorate
or nitrate complexes. In general, we found it difficult to
obtain crystals of [FeTC6TPP(L)2]+ with any of the axial
ligands, and only one structure, of [FeTC6TPP(1-MeIm)2]Cl,
has been reported.18
The solution magnetic susceptibility data for [FeOMTPP-
(4-CNPy)2]ClO4 and [FeTC6TPP(4-CNPy)2]ClO4 at the low-
est temperatures accessible in the liquid state are in good
agreement with the EPR results. In the case of [FeTC6TPP-
(4-CNPy)2]ClO4, a pure LS, S ) 1/2, state occurs below 240
K with µeff ) 1.95 ( 0.05 µB, indicative of little orbital
contribution and consistent with the (dxz,dyz,)4(dxy)1 ground
state detected in the 4.2 K EPR spectra. Even though in the
[FeOMTPP(4-CNPy)2]ClO4 case the pure LS state cannot
be achieved over the temperature range accessible to the
solution magnetic measurements, on the basis of the magnetic
data (Figure 5) we can predict that this complex adopts a
predominantly LS state at temperatures below 150-140 K
and, as a result, exhibits an axial EPR signal at 4.2 K. As
the temperature is raised, the population of the S ) 3/2 state
increases, and above 260 K, [FeOMTPP(4-CNPy)2]ClO4
3
exists exclusively in the S ) /2 state with µeff ) 3.56 (
3
0.36 µB. An S ) /2 state for [FeTC6TPP(4-CNPy)2]ClO4,
however, is never achieved, and even at 320 K, this complex
1
3
is still in spin equilibrium (S ) /2 T S ) /2). Transition
3
from the LS to the S ) /2 state is attributed to the weak
ligand field strength of the 4-CNPy ligand, which places both
[FeOMTPP(4-CNPy)2]ClO4 and [FeTC6TPP(4-CNPy)2]ClO4
3
1
very close to the spin crossover point (S ) /2 T S ) /2).
On the other hand, [FeOETPP(4-CNPy)2]ClO4 has a
completely different type of EPR spectrum at 4.2 K, with
the following g values for the solid samples (Figure 9): 5.21,
4.25, 3.03, and 2.07. Both the CD2Cl2 frozen solution and
polycrystalline samples display the same type of EPR signal,
but with different relative intensities of the peaks (Figure
9). The EPR signals with g larger than 4 and around 2 point
3
to the S ) /2 state, but g ) 3.03 is likely a “large gmax
”
signal characteristic of the LS (dxy)2(dxz,dyz,)3 ground state,
which is different from the LS electronic ground state of
either the OMTPP or TC6TPP complexes. There are several
arguments that support this suggestion. First, the g ) 3.03
value is very similar to the unusually low gmax values of 2.89,
3.07, and 3.17 observed in the EPR spectra of [FeTMP-
(3-EtPy)2]ClO4, [FeTMP(3-ClPy)2]ClO4, and [FeTMP-
(2-MeHIm)2]ClO4, respectively.5 Therefore, there must be,
especially in the frozen solution where the g ) 3.03 signal
is the most intense, some molecules of [FeOETPP(4-CNPy)2]-
ClO4 in the LS state. Low temperature SQUID measurements
of solid samples do indicate that [FeOETPP(4-CNPy)2]ClO4
The second observation relates to the data of Table 4,
where it can be seen that while the chemical shifts of
[FeOMTPP(4-CNPy)2]+ at 273 and 183 K are within the
deviations expected on the basis of calibration of the
temperature controllers of two different NMR spectrometers,
those for [FeOETPP(4-CNPy)2]+ (this work and refs 10 and
20) and [FeTC6TPP(4-CNPy)2]+ (this work) or [FeTBTXP-
(4-CNPy)2]+ (ref 10) differ dramatically for some of the
resonances. Notably, for [FeOETPP(4-CNPy)2]+, while the
chemical shift of CH2(out) is fairly similar at all common
temperatures, that of CH2(in) differs by 3 ppm at 273 K and
6.9 ppm at 183 K. Marked differences in the chemical shifts
1
3
is a mixture of the S ) /2 and /2 states below 200 K and
that the population of the LS (S ) 1/2) state increases as the
temperature is lowered.20 Therefore, it is not surprising that
in frozen solution at 4.2 K a detectable amount of LS (S
)1/2) molecules of [FeOETPP(4-CNPy)2]ClO4 is observed
in the EPR spectra. Second, the presence of the g ) 3.03
signal in the EPR spectra of the polycrystalline sample can
be justified by the fact that [FeOETPP(Py)2]ClO4 undergoes
spin transition upon lowering the temperature from 298 to
180 K and then to 80 K.21 Therefore, it is possible that
Inorganic Chemistry, Vol. 43, No. 2, 2004 773

Yatsunyk and Walker
of the ortho-phenyl protons are also observed (Table 4). The
most likely explanation of these differences between the
currently reported shifts and those reported previously10 is
that the previous study incorrectly assigned the two types
of resonances, CH2(in) and o-H. It is hard to imagine how
this error could have been made, when simple COSY and
NOESY/EXSY experiments available on all modern NMR
spectrometers allow ready assignments of the resonances
involved. Thus, the assignments of the present work are
unambiguously the correct ones.
smaller spin delocalization to the meso-carbons,46 even
though the EPR spectra indicate just as pure a (dxz,dyz)4(dxy)1
ground state for these 4-CNPy complexes as in the cases of
1
ruffled iron porphyrinates.5,6,8,10 When the H NMR results
are considered for [FeOMTPP(4-CNPy)2]ClO4, δm - δp and
δm - δo are equal to -1.7 and -3.6 ppm at 213 K, and
+2.9 and +3.5 ppm at 183 K (see Table 4). In comparison,
the highly ruffled [FeTPP(4-CNPy)2]+ complex has much
larger values of δm - δp and δm - δo of +5.7 and +14.2
ppm at 213 K.15,16 Smaller δm - δp and δm - δo values, as
compared to the same values in complexes with the (dxz,dyz)4-
(dxy)1 ground state stabilized by a ruffled porphyrin core, are
indicative of poorer spin delocalization to the meso-carbons
due to the purely saddled conformation expected for [FeOMT-
PP(4-CNPy)2]ClO4 in solution on the basis of its crystal
structure. A substantial degree of ruffling is required for good
interaction of the dxy orbital with the porphyrin a2u(π) orbital.6
In the case of [FeTC6TPP(4-CNPy)2]ClO4, δm - δp and δm
- δo are equal to +5.4 and +9.5 ppm at 213 K and +6.7
and +10.8 ppm at 183 K. These differences are larger than
in the case of [FeOMTPP(4-CNPy)2]ClO4, pointing to higher
(+) spin density at the meso-carbons, probably due to a sig-
nificant ruffling of the TC6TPP core. Although we were not
able to obtain the structure of [FeTC6TPP(4-CNPy)2]ClO4,
structures of other complexes, [FeTC6TPP(1-MeIm)2]Cl18 and
(TC6TPP)FeCl,38 suggest ∼0.6:0.4 of saddled versus ruffled
components are certainly possible in the core geometry.
In comparison, for [FeOETPP(4-CNPy)2]ClO4 at any
temperature, the order of the phenyl shifts is δo > δp > δm
(Table 4). This is the same order as for [FeOMTPP-
(4-CNPy)2]ClO4 at temperatures higher than 198 K (Figure
6). The phenyl shift differences δm - δp and δm - δo are
large and negative at all temperatures (Table 4), indicating
a fairly large negative spin density at the meso-carbons (i.e.,
the opposite sign as compared to the spin density on the
metal).15 The values of δm - δp and δm - δo, extrapolated
to 298 K (-7.3 and -11.6 ppm, respectively), are somewhat
smaller in magnitude (70% and 90%, respectively) but
opposite in sign to those for the (dxz,dyz)4(dxy)1 electron
configuration complex [FeTPP(t-BuNC)2]+ 49 and of com-
parable magnitude but opposite in sign to those for [FeTPP-
(4-CNPy)2]+ 16 (128% and 82%, respectively), where there
is significant spin delocalization to the 3a2u(π) orbital. These
values are about 17% and 24% of those observed for the
HS Fe(III) porphyrin radical [TPP‚FeCl]+,50,51 and 33% and
42% of those observed for the IS (S ) 3/2) Fe(III) corrolate
radical, [TPCorrFeCl],52 suggesting, in analogy to what was
found in the latter case, a partial S ) 2 Fe(II), S ) 1/2 π-cation
radical resonance structure, with antiferromagnetic coupling
between metal and porphyrin unpaired electrons; antiferro-
magnetic coupling would thus result from the presence of
For [FeTC6TPP(4-CNPy)2]+ it can also be seen from Table
4 that there are marked differences in the chemical shifts of
the R-CH2 protons from those reported for [FeTBTXP-
(4-CNPy)2]+.10 These two compounds differ only in the fact
that the latter compound has two meta-OCH3 groups on each
phenyl ring, and while these substituents might cause minor
shifts in the o-H and p-H resonances, it seems unlikely that
they would dramatically affect the ring conformation of the
cyclohexenyl rings and thus the chemical shift of the R-CH2
resonance from that observed in the present study of
[FeTC6TPP(4-CNPy)2]+. Nevertheless, at 273 K the chemical
shift of the complex of the previous study is nearly 10 ppm
larger than that of the present study, and even at 183 K the
chemical shifts of the R-CH2 protons in the two studies differ
by 6.5 ppm (Table 4). We have included in Table 4 the
chemical shifts of the perchlorate complexes of the three
3
porphyrinates of this study, all of which have S ) /2
electronic states throughout the temperature range studied.
While the direction of shift of the R-CH2 proton resonance
of ref 10 is that expected if not enough axial ligand had been
added, the authors used 64 equiv of axial ligand, which, at
10 mM concentration of iron porphyrinate, should have been
enough to ensure full complexation, especially at the lowest
temperatures. Thus, we do not have an explanation for the
differences in chemical shifts of the R-CH2 protons of these
two complexes.
In this work, 1H NMR and solution magnetic susceptibility
data were acquired over the same temperature range, and
the results correlate extremely well. As discussed above,
3
[FeOETPP(4-CNPy)2]ClO4 is in the S ) /2 state at all
temperatures from ambient down to 180 K, [FeOMTPP-
(4-CNPy)2]ClO4 is in the S ) 3/2 state from ambient temper-
ature to 240 K and then the LS state becomes populated as
the temperature decreases, and, finally, [FeTC6TPP(4-CNPy)2]-
ClO4 is in the LS state at any temperature below 230 K.
In the case of the (dxz,dyz)4(dxy)1 electronic ground state,
the dπ orbitals are filled, but the dxy unpaired electron can
engage in spin delocalization to the porphyrinate ring if it is
ruffled, and such ruffling is quite extreme in most of the
complexes with the (dxz,dyz)4(dxy)1 electron configura-
tion.5,6,9,10,46-48 Such spin delocalization results in a large
positive spin density at the meso-carbons, which leads to a
large and positive difference in the chemical shifts of the
phenyl protons (protons attached directly to the phenyl rings),
δm - δp, and δm - δo, or to the following order of chemical
shifts: δm > δp > δo.15 In the case of saddled porphyrins,
meso-phenyl shifts are much smaller because of the reduced
possibility of ruffling of the porphyrinate ring, and, therefore,
(49) Simonneaux, G.; Hindre, F.; Le Plouzennec, M. Inorg. Chem. 1989,
28, 823-825.
(50) Phillippi, M. A.; Goff, H. M. J. Am. Chem. Soc. 1982, 104, 6026-
6034.
(51) Gans, P.; Marchon, J.-C.; Reed, C. A.; Regnard, J.-R. NouV. J. Chim.
1981, 203-204.
(52) Cai, S.; Licoccia, S.; D’Ottavi, C.; Paolesse, R.; Nardis, S.; Bulach,
V.; Zimmer, B.; Shokhireva, T. Kh.; Walker, F. A. Inorg. Chim. Acta
2002, 339C, 171-178.
774 Inorganic Chemistry, Vol. 43, No. 2, 2004

4-Cyanopyridine Complexes of Saddled Iron Porphyrinates
2
2
an unpaired electron in the dx -y orbital of HS iron(II), which
has the proper symmetry to interact with the 3a2u(π) orbital
of the porphyrin ring when the latter is saddled.53 This
suggestion is further supported by arguments presented in
the following four paragraphs.
ration complex [FeOETPP(4-Me2NPy)2]Cl: If one dπ un-
paired electron gives rise to an average isotropic shift of 8.1
- 2.3 ) 5.8 ppm for this porphyrinate, then two would be
expected to give rise to an isotropic shift of about double
this amount, or a chemical shift of 2 × 5.8 + 2.3 ) 13.9
ppm. The third unpaired electron of this IS state is expected
Sakai and co-workers have pointed out that there are two
possible electron configurations for S ) /2 iron porphyri-
3
2
to be in the dz orbital, which will not contribute to the
3
1
1
2
2
1 54
2
2
isotropic shifts of the porphyrin protons (although it certainly
will contribute to those of the axial ligand protons in this
6-coordinate complex, as discussed below). However, instead
of an average chemical shift of the CH2 protons of 13.9 ppm
at 243 K, the observed average shift is 32.6 ppm. A similar
(somewhat larger) average CH2 chemical shift is shown by
nates, those being (dxz,dyz) (dxy) (dz ) and (dxy) (dxz,dyz) (dz ) .
Qualitatively, the main difference expected in the NMR
spectra is that the latter should have larger pyrrole-substituent
shifts than the former (due to having two dπ unpaired
electrons rather than one), and zero spin density at the meso
positions (due to the absence of an unpaired electron in the
dxy orbital), while the former might have spin density at those
positions if the porphyrinate ring is ruffled.15 While it is
indeed true that these two electron configurations are both
possible and that the latter should have larger spin density
at the pyrrole positions than the former (and we considered
this possibility as the explanation of the large pyrrole-CH2
shifts observed and discussed below before noticing the
negative phenyl shift differences δm - δp and δm - δo),
neither of these electron configurations can explain the
negative phenyl-H shift differences observed (or the excep-
tionally large pyrrole-CH2 shifts observed for [FeOETPP-
(4-CNPy)2]ClO4 discussed below and shown in Table 4).
Thus, these negative phenyl-H shift differences indicate that
something quite different is involved, i.e., spin transfer from
the porphyrinate ring to the iron to create a partial π-cation
radical, and further suggest the need for DFT calculations
to be carried out on this system15,55 to determine the extent
of electron transfer from porphyrinate to iron in this complex.
3
5
the S ) /2, /2 complex OETPPFeCl (44 ppm at 243 K38),
which has been shown to be largely HS (90-96%),56 with
the downfield-shifted average CH2 resonance thus resulting
from σ-spin delocalization to the â-carbons arising from
nearly full occupation of the dx -y orbital.15 In the present
2
2
3
case, where the overall spin state is S ) /2, the downfield-
shifted average CH2 resonance is fully consistent with a
significant degree of internal reduction of the metal by the
porphyrinate to create some degree of S ) 2 Fe(II) antifer-
1
romagnetically coupled to an S ) /2 porphyrin radical, as
was concluded above from the phenyl-H shift pattern.
For the [FeOETPP(4-CNPy)2]ClO4 complex, L 3,5-H and
L 2,6-H resonances are characterized by large downfield
shifts of 86.4 and 137.0 ppm at 183 K, respectively. These
large downfield shifts are totally consistent with either the
S ) 2 Fe(II) state antiferromagnetically coupled to a porphy-
rinate radical, or the S ) 3/2 Fe(III) state, with half-occupation
2
of the dz orbital in either case, which allows significant σ-spin
delocalization to the axial ligands, and downfield shifts.15
Thus, although the NMR spectra of [FeOETPP-
(4-CNPy)2]ClO4 have previously been explained in terms of
3
The S ) /2 state is characterized by large upfield shifts
for directly bound pyrrole protons,15 which corresponds to
large downfield shifts for protons of methyl or methylene
groups directly attached to the pyrrole â-carbons.15 This is
observed in the case of the [FeOETPP(4-CNPy)2]ClO4
complex by average CH2 shifts of 32.6 and 40.5 ppm at 243
and 183 K, respectively, and a positive slope of the
temperature dependence of the CH2(out) chemical shift on
the Curie plot and small temperature dependence of the CH2-
(in), both of which extrapolate to close to the average
diamagnetic CH2 shift (2.3 ppm) at infinite temperature
(Supporting Information Figure S4B), indicating no thermally
accessible excited state is involved. A closer look at the
3
a simple S ) /2 iron porphyrinate,9,10,20 we find that the
chemical shifts of this complex are characterized by (i) large
and negative phenyl-H shift differences δm - δp and δm -
δo, (ii) larger average CH2 shifts than expected for the pres-
ence of either one or two unpaired electrons in π-symmetry
d-orbitals, and (iii) large positive shifts of the 2,6-H and
3,5-H of the axial 4-CNPy ligands. The first (i) is indicative
of negative spin density at the meso-carbons, the second (ii)
2
2
is consistent with at least partial occupation of the dx -y
orbital, and the third (iii) is consistent with occupation of
2
the dz orbital. These three facts lead to an overall electron
average CH2 shifts of [FeOETPP(4-CNPy)2]ClO4, as com-
pared to those of the S )
2
2
1
(1-x)
2
1
2
2
configuration of the metal of (dxy) (dxz,dyz) (dz ) (dx -y
)
/
complex [FeOETPP-
2
and of the porphyrin ring of (OETPP)(2-x)- with antiferro-
magnetic coupling of metal and ligand fractional electrons
x. Whether x ) 1.0 or less than 1.0 cannot be determined
without additional experiments (for example magnetic Mo¨ss-
bauer studies) and theoretical, especially DFT, calculations.
In comparison to [FeOETPP(4-CNPy)2]ClO4, the chemical
shifts for the methyl and R-methylene protons in the
corresponding OMTPP and TC6TPP complexes have strong
negative temperature dependences, which indicate a change
(4-Me2NPy)2]Cl (+8.1 and +7.8 ppm at 243 and 183 K,
respectively38), both in CD2Cl2, shows that the 4-CNPy
complex has anomalously large pyrrole-CH2 shifts for a
complex expected to have two dπ unpaired electrons for the
3
2
2
1
2
S ) /2 state with electron configuration (dxy) (dxz,dyz) (dz ) ,
1
as compared to the S ) /2 (dxy)2(dxz,dyz)3 electron configu-
(53) Buisson, G.; Deronzier, A.; Due´e, E.; Gans, P.; Marchon, J.-C.;
Regnard, J.-R. J. Am. Chem. Soc. 1982, 104, 6793-6795.
(54) Sakai, T.; Ohgo, Y.; Ikeue, T.; Takahashi, M.; Takeda, M.; Nakamura,
M. J. Am. Chem. Soc. 2003, 125, 13028-13029.
(55) Zakharieva, O.; Schu¨nemann, V.; Gerdan, M.; Licoccia, S.; Cai, S.;
Walker, F. A.; Trautwein, A. X. J. Am. Chem. Soc. 2002, 124, 6636-
6648.
(56) Schu¨nemann, V.; Gerdan, M.; Trautwein, A. X.; Haoudi, N.; Mandon,
D.; Fischer, J.; Weiss, R.; Tabard, A.; Guilard, R. Angew. Chem., Int.
Ed. 1999, 38 (21), 3181-3183.
Inorganic Chemistry, Vol. 43, No. 2, 2004 775

Yatsunyk and Walker
3
from the S ) /2 state to LS as the temperature is lowered.
(-9.9 and -35.5 ppm at 183 K for L 3,5- and L 2,6-,
respectively). The NMR results obtained for [FeOMTPP-
(4-CNPy)2]ClO4 are in good agreement with the above
data: L 3,5-H and 2,6-H have strong negative slope in the
Curie plot (Figures S4A and S8A), indicating transition from
S ) 3/2 (with downfield shifts for L) to S ) 1/2 (with upfield
shifts for L) ground states as the temperature decreases.
In summary, as for the simple [FeTPP(4-CNPy)2]+ com-
plex,16 the LS state of the OMTPP and TC6TPP complexes
with (dxz,dyz)4(dxy)1 electron configuration is characterized by
large and positive values of the phenyl-H shift differences
δm - δp and δm - δo due to large (+)π spin density at the
meso-carbons. A small spin density at the pyrrole-â positions,
on the other hand, results in chemical shifts for methyl or
methylene protons in ORTPP complexes that are usually in
the diamagnetic region. Ligated 4-CNPy has large upfield
shifts for both of its proton resonances. Therefore, while EPR
spectroscopy is a good probe for the electronic ground state
at low temperatures, NMR spectroscopy is very useful in
defining the spin state of the complex at ambient conditions,
and for determining the nature of spin state behavior, whether
it is due to a thermally accessible excited state or a
thermodynamic equilibrium between spin states, or both.
Structures of the Complexes. The crystal structure data
for [FeOMTPP(4-CNPy)2]ClO4 and [FeOETPP(4-CNPy)2]-
ClO4 were obtained at 170 K. According to the solution
magnetic measurements at this temperature, [FeOETPP-
In the pure LS state, with the (dxz,dyz)4(dxy)1 configuration,
spin delocalization to the pyrrole-â position is negligible,14,15
and as a result, the chemical shift of CH2(R) in [FeTC6TPP-
(4-CNPy)2]ClO4 is in the diamagnetic region at low tem-
peratures (6.8 ppm at 183 K). The methyl resonance in
[FeOMTPP(4-CNPy)2]ClO4 is still shifted downfield (δ )
31.4 ppm at 183 K), indicating significant population of the
3
S ) /2 state, even at this temperature. The nature of the S
) 3/2 state for these two complexes cannot be deduced with
certainty because it is not possible to achieve fully the S )
3/2 state. However, it should be noted that above 198 K the
order of phenyl-H shifts of the [FeOMTPP(4-CNPy)2]ClO4
complex is the same as for [FeOETPP(4-CNPy)2]ClO4,
suggesting that this complex also has some contribution from
the S ) 2 Fe(II) state antiferromagnetically coupled to an S
1
) /2 porphyrin π-cation radical.
Analysis of the temperature dependences of the resonances
using our fitting program that includes the contributions of
one or more thermally accessible excited states39 shows that,
despite the complications due to other temperature-dependent
processes going on in the solution (ligand exchange kinetics,
porphyrin ring inversion kinetics, hindered ethyl rotation at
low temperatures), the phenyl-H resonances of [FeOETPP-
(4-CNPy)2]ClO4 show linear 1-level temperature dependence,
while the chemical shifts of both [FeOMTPP(4-CNPy)2]ClO4
and [FeTC6TPP(4-CNPy)2]ClO4 below 198 K have temper-
ature dependences that indicate an S ) 1/2 ground state and
a thermally accessible S ) 3/2 excited state, with the excited-
state being 650 cm-1 above the ground state for the former
and 900 cm-1 for the latter. The temperature-dependent fits
for these two complexes are shown in Supporting Information
Figure S8A,B; in both cases, only the data below 198 K were
used in the 2-level fit. The excited states in each case are
much higher in energy than would be predicted on the basis
of the full temperature dependence observed for these com-
plexes, and the chemical shifts of all protons above 198 K
clearly do not fit the 2-level plots. This is because of the
3
(4-CNPy)2]ClO4 exists as an S ) /2 state complex but
[FeOMTPP(4-CNPy)2]ClO4 is largely LS with the (dxz,dyz)4-
(dxy)1 electronic ground state; however, in the structures of
all three complexes the Fe-Nax distances of about 2.2 Å
are indicative of either the IS state or the HS Fe(II), S ) 1/2
cation radical with antiferromagnetic coupling. For the LS
complexes of OMTPP and OETPP with various strong
basicity pyridines and imidazoles (4-Me2NPy, 1-MeIm, and
2-MeHIm)18,42 (also see Table 2), the Fe-Nax distances are
on average 2.0 Å, about 0.2 Å shorter than in the present
3
5
case. For the admixed S ) /2, /2 spin state of [FeOEP-
(3,5-Cl2Py)2]ClO4, the average Fe-Nax distance is even
longer (2.347(4) Å) than for S ) 3/2 complexes.3 Thus, there
appears to be a correlation between the axial ligand bond
length and the spin state of the complex: axial bond length
increases as the spin state changes from the LS state (Fe-
Nax ∼ 2.0 Å) through the S ) 3/2 state (Fe-Nax ∼ 2.2 Å) to
the HS state (Fe-Nax > 2.3 Å).
1
onset of a thermodynamic equilibrium between the S ) /2
and S ) 3/2 states in each case, as was observed for the depen-
dence of magnetic moment on temperature, Figure 5, which
provides a much lower energy route for the change in spin
state and therefore the observed temperature dependence of
1
the H NMR shifts. Thus, at higher temperatures than 198
K, first the OMTPP complex and then the TC6TPP complex
become involved in thermodynamic equilibria between the
The question arises as to why the [FeOMTPP(4-CNPy)2]-
ClO4 complex is in the S ) 3/2 state rather than the LS state
in the crystal lattice used for structure determination. It should
be noted that all three sets of crystals were grown at room
temperature, where [FeOMTPP(4-CNPy)2]ClO4 and [FeO-
1
3
S ) /2 and the S ) /2 spin states, which dominate the
chemical shifts of both of these complexes above 198 and
223 K, respectively. Similar behavior was found for the alkyl
peroxide complex [FeTPP(OCH3)(OO-t-Bu)]-, where a
thermodynamic equilibrium between the (dxy)2(dxz,dyz)3 and
(dxz,dyz)4(dxy)1 electron configurations and, thus, likely also
a planar and a ruffled porphyrinate ring was observed.57
The axial ligand 2,6-H and 3,5-H resonances in the pure
LS [FeTC6TPP(4-CNPy)2]ClO4 are shifted strongly upfield
3
ETPP(4-CNPy)2]ClO4 complexes exist in the S ) /2 spin
state (Figure 5). Freezing of the samples on the diffractometer
was done in a cold stream of nitrogen over the course of a
few seconds; therefore, it is quite likely that the room temper-
ature geometry was trapped and preserved at 170 K. In the
case of [FeOMTPP(4-CNPy)2]ClO4, it is possible that a spin
transition may be seen if the sample is cooled slowly and to
lower temperatures. A spin transition was detected in
(57) Rivera, M.; Caignan, G. A.; Astashkin, A. V.; Raitsimring, A. M.;
Shokhireva, T. Kh.; Walker, F. A. J. Am. Chem. Soc. 2002, 124, 6077-
6089.
776 Inorganic Chemistry, Vol. 43, No. 2, 2004

4-Cyanopyridine Complexes of Saddled Iron Porphyrinates
[FeOETPP(Py)2]ClO4 by cooling the sample from 298 to 180
K and then to 80 K.21 Most of the bonds contracted at lower
temperatures, but the shortening of the axial bonds was the
most prominent. The average Fe-Nax bond length decreased
from 2.201(3) to 2.041 A and then to 1.993(3) Å.21 This
to the previous study,10 it has been shown herein that the
NMR spectra (phenyl-H shifts) indicate that the actual
electronic state involves a significant contribution from an
S ) 2 Fe(II) center antiferromagnetically coupled to an S )
1/2 porphyrinate π-cation radical; however, the amount of
this contribution has not been quantified. On the other hand,
as also concluded previously,10 [FeOMTPP(4-CNPy)2]ClO4
and [FeTC6TPP(4-CNPy)2]ClO4 complexes exist as an equi-
librium mixture of S ) 3/2 and LS spin states above 180 and
240 K, respectively, and at lower temperatures (4.2 K), they
adopt a low spin (dxz,dyz)4(dxy)1 configuration. For two of the
complexes, the rates of porphyrin ring inversion/axial ligand
rotation could be measured, by NOESY/EXSY methods for
[FeOETPP(4-CNPy)2]ClO4 and by dynamic NMR methods
(DNMR) for [FeTC6TPP(4-CNPy)2]ClO4. The results show
that [FeOETPP(4-CNPy)2]ClO4 has a fairly low activation
enthalpy (32 kJ mol-1) and very negative activation entropy
(-104 J mol-1 K-1), leading to a rate constant of 59 s-1 in
CD2Cl2 at 298 K. In comparison, for [FeTC6TPP(4-CNPy)2]-
ClO4 the activation enthalpy is even lower (24 kJ mol-1),
the estimated activation entropy is less negative (at least -20
J mol-1 K-1), and the estimated rate constant at the same
temperature in the same solvent is at least 4.4 × 107 s-1,
about a million times faster than for [FeOETPP(4-CNPy)2]-
ClO4. These values attest to the much greater flexibility of
the TC6TPP ring. In the solid state, the [FeOETPP(4-CNPy)2]-
ClO4 and [FeOMTPP(4-CNPy)2]ClO4 complexes are mostly
saddled with close to perpendicular ligand orientation. Fe-
3
indicates that the spin state changed from S ) /2 to LS in
the course of the X-ray experiments. In the solution magnetic
susceptibility experiments, [FeOMTPP(4-CNPy)2]ClO4 be-
haves in the same manner as observed for [FeOETPP(Py)2]-
ClO4 in the solid state: the magnetic moment of both com-
plexes decreases from ∼3.6 µB at 300 K to ∼2.8 µB at 200
K.21 It would therefore be interesting to try variable tem-
perature X-ray crystallographic experiments on [FeOMTPP-
(4-CNPy)2]ClO4. In contrast, [FeOETPP(4-CNPy)2]ClO4 is
3
in the S ) /2 state at all temperatures achievable by X-ray
experiments.
All complexes studied in this research have a predomi-
nantly saddled porphyrin core with estimated ruffling
admixture ranging from 1% to 17% (see Table 3). Of course,
the saddled distortion is stronger in any OETPP complex as
compared to OMTPP, because the former has bulkier groups
on the pyrrole â-positions. Stronger saddling is reflected in
smaller dihedral angles of the phenyls, and a larger deviation
of pyrrole â-C’s from the mean porphyrin plane. In the
structure of [FeOMTPP(4-CNPy)2]ClO4, the average dihedral
angle of the phenyls is 48.0°, while in any of the [FeOETPP-
(4-CNPy)2]ClO4 structures it is not higher than 43°. The
highest deviation of the porphyrin core atoms in [FeOMTPP-
(4-CNPy)2]ClO4 is 1.21 Å, which is substantially smaller
than the highest deviation in [FeOETPP(4-CNPy)2]ClO4, 3,
1.40 Å. A saddled distortion of the porphyrin core should
require perpendicular orientation of the axial ligands; how-
ever in the case of [FeOMTPP(4-CNPy)2]ClO4 and [FeO-
ETPP(4-CNPy)2]ClO4, 2, much smaller dihedral angles are
observed, 64.3° and 67.1°, respectively. To our knowledge
these are the smallest observed “perpendicular” angles in
octaalkyltetraphenylporphyrins with pyridine axial ligands,
but these small angles are probably made possible by the much
N
ax bond lengths (∼2.2 Å) indicate the S ) 3/2 state in both
cases. In order to reach the spin transition for the [FeOMTPP-
(4-CNPy)2]ClO4 case, lower temperatures are required than
were used for X-ray crystallography (170 K).
Acknowledgment. The support of the National Institutes
of Health, Grant DK-31038 (F.A.W.), the National Science
Foundation, Grant CHE9610374 (X-ray diffractometer grant),
and the University of Arizona Molecular Structure Labora-
tory in this research are gratefully acknowledged. Dr. Michael
D. Carducci, X-ray Facility specialist, Department of Chem-
istry, University of Arizona, is thanked for his help in struc-
ture refinement, and Dr. Charles F. Campana, from Bruker,
is thanked for his help with the absorption corrections and
refinement of the twinned structure of [FeOETPP(4-CNPy)2]-
ClO4 from CDCl3/cyclohexane. Dr. Sheng Cai is thanked
for his help in processing magnetic susceptibility data, Dr.
3
longer Fe-Nax bond lengths of the S ) /2 Fe(III) centers.
Due to the strongly saddled distortions, the Fe-NP
distances are short (see Table 2). In general, OMTPP
complexes should have longer distances than OETPP but
shorter than TC6TPP.18 In this case, the Fe-NP distances in
[FeOMTPP(4-CNPy)2]ClO4 are longer than those in [FeO-
ETPP(4-CNPy)2]ClO4 structures 1-4 but almost the same
as the Fe-NP distances in 5 and 6.
1
Nikolai V. Shokhirev for fitting the H NMR shifts to the
Conclusions. Broadly speaking, most of the conclusions
of this study are in agreement with those of the previous
study:10 [FeOETPP(4-CNPy)2]ClO4 exists as a complex with
2-level temperature dependent fitting program, TDFw, and
Professor Mario Rivera for providing very helpful comments.
This paper was revised when F.A.W. was on Sabbatical leave
at the Institute of Physics at the University of Lu¨beck with
support from an Alexander von Humboldt Senior Research
Award and is dedicated to Professor and Rector Alfred X.
Trautwein on the occasion of his 63rd birthday.
3
an overall spin state of S ) /2 at any temperature. This is
reflected in the NMR, EPR, and magnetic susceptibility
measurements, but the numerical data differ somewhat, as
discussed above. At 4.2 K it also exhibits a low spin (S )
1/2) EPR signal (g ) 3.03), suggesting that it is possible for
some molecules to have the (dxy)2(dxz,dyz)3 electron config-
uration in frozen solution. However, there is no evidence in
Supporting Information Available: Details of the structure
determinations, Figures S1-S8, and Tables S1-S19.This material
is available free of charge via the Internet at http://pubs.acs.org.
1
the temperature dependence of the H NMR shifts for an S
) 1/2 ground state for the majority of molecules. In contrast
IC035010Q
Inorganic Chemistry, Vol. 43, No. 2, 2004 777