Continuous Wave and Pulse EPR and ENDOR Study of Oxygenated Cobalt(II) Heme Model Systems

Article abstract of DOI:10.1021/jp993668c

Continuous wave (CW) and pulse electron paramagnetic resonance (EPR), and electron nuclear double resonance (ENDOR) techniques were used to study frozen solutions of (oxyCo)TPP(L) (TPP = tetraphenylporphyrin, L = pyridine, 1-methylimidazole). By using a combination of CW EPR at X-, Q-, and W-band and Davies-ENDOR at Q-band, the g and cobalt hyperfine matrices and their corresponding principal axes were, for the first time, determined in detail. The variation of the g values as a function of temperature was followed by Q-band CW EPR. The existing ambiguity about the assignment of one of the g principal axes to the O-O bound direction was analyzed in detail. By studying the different proton interactions with Mims-ENDOR at X-band, a considerable solvent effect was found. With hyperfine sublevel correlation (HYSCORE) at X-band and two-pulse ESEEM (electron spin-echo envelope modulation) at S-band, the hyperfine and nuclear quadrupole interactions of the nitrogen nuclei of the axial ligand and the TPP ligand were determined. All the magnetic parameters were analyzed as a function of the geometrical and electronic structure of the complexes.

Full text of DOI:10.1021/jp993668c

J. Phys. Chem. B 2000, 104, 2919-2927
2919
Continuous Wave and Pulse EPR and ENDOR Study of Oxygenated Cobalt(II) Heme
Model Systems
Sabine Van Doorslaer and Arthur Schweiger*
Physical Chemistry Laboratory, Swiss Federal Institute of Technology, 8092 Zurich, Switzerland
ReceiVed: October 14, 1999; In Final Form: January 10, 2000
Continuous wave (CW) and pulse electron paramagnetic resonance (EPR), and electron nuclear double
resonance (ENDOR) techniques were used to study frozen solutions of (oxyCo)TPP(L) (TPP ) tetraphen-
ylporphyrin, L ) pyridine, 1-methylimidazole). By using a combination of CW EPR at X-, Q-, and W-band
and Davies-ENDOR at Q-band, the g and cobalt hyperfine matrices and their corresponding principal axes
were, for the first time, determined in detail. The variation of the g values as a function of temperature was
followed by Q-band CW EPR. The existing ambiguity about the assignment of one of the g principal axes to
the O-O bound direction was analyzed in detail. By studying the different proton interactions with Mims-
ENDOR at X-band, a considerable solvent effect was found. With hyperfine sublevel correlation (HYSCORE)
at X-band and two-pulse ESEEM (electron spin-echo envelope modulation) at S-band, the hyperfine and
nuclear quadrupole interactions of the nitrogen nuclei of the axial ligand and the TPP ligand were determined.
All the magnetic parameters were analyzed as a function of the geometrical and electronic structure of the
complexes.
1. Introduction
O-O bond directions. Hori et al.8,9 repeated these measurements
at room temperature (RT) and 77 K. A comparison of the RT
For several decades, the dioxygen storage and transport
function of the iron-containing heme proteins, hemoglobin (Hb)
and myoglobin (Mb), has been the subject of many spectroscopic
studies. However, the analysis of the electronic and geometric
structure is very difficult because of the fast autoxidation rates
of these proteins and the fact that they are diamagnetic and
therefore EPR-silent (EPR ) electron paramagnetic resonance).
It has been found that many of the planar Co(II) complexes
can reversibly form 1:1 adducts with molecular oxygen (for
reviews see refs 1 and 2). These adducts are further related to
the oxygen-carrying heme proteins since they possess a coor-
dinated nitrogen base in the sixth coordination site of the metal
ion. Both the oxygenated and the deoxygenated Co(II) adducts
are paramagnetic and can thus be studied by EPR. Because of
their unique features, these adducts have become of special
interest as models for natural heme systems.1,2 In particular,
cobalt(II) porphyrins are used as model systems because of their
structural analogy with the porphyrin fragment of Hb and Mb,
and the fact that the autoxidation is generally slower than in
the case of the native proteins. The chemical substitution of
ferrous protoporphyrin IX by cobaltous porphyrin in Hb and
Mb3 opened the way for studying both the oxy and deoxy
species of the proteins with EPR and electron nuclear double
resonance (ENDOR) spectroscopy.4-9
EPR and X-ray data of oxyCoMb revealed that the principal
axis associated with the g value closest to the free electron value
ge is approximately oriented along the O-O bond. In a later
work,9 they made the same assignment for the 77 K data.
Subsequently, several authors assigned for different oxygenated
Co(II) complexes the smallest g value13-16 or the middle g value
to the O-O bond direction.17 This ambiguity has severe
consequences for the interpretation of all the other magnetic
parameters.
Moreover, in all studies done on frozen solutions of oxygen-
ated cobalt(II) porphyrin systems, the g and ACo matrices could
not be determined accurately and simulations of the continuous
wave EPR (CW EPR) spectra are found to be hardly convincing.
This is due to the fact that the two matrices are orthorhombic
and non-coaxial. The frozen solution CW EPR spectra recorded
at X-band show strongly overlapping features, which complicate
the evaluation of the parameters.
In this work, we report on a CW and pulse EPR and ENDOR
study of frozen toluene solutions of oxygenated cobalt tetraphen-
ylporphyrin(pyridine), (oxyCo)TPP(py) (1a), and oxygenated
cobalt tetraphenylporphyrin(1-methylimidazole), (oxyCo)TPP-
(1-MeIm) (1b). The two complexes have previously been
investigated by means of CW EPR18 and three-pulse electron
spin-echo envelope modulation (ESEEM) spectroscopy.17,19 In
these studies, the ambiguity in the assignment of the g principal
axes is not resolved and the reported data for g and ACo are not
convincing.
Although a number of EPR and ENDOR studies have been
carried out on oxygenated synthetic cobalt systems,1,2 many
problems are still left unsolved. The most pronounced ambiguity
is the one concerning the assignment of the g principal axes to
the dioxygen bond direction. Some of the authors1,10,11 ascribed
the principal axis with the largest g value to the O-O bond
direction in accordance with EPR studies on O2- defects in alkali
halide single crystals.12 In oxycobalt myoglobin (oxyCoMb)
single crystals Dickinson and Chien6,7 found at 77 K two oxy
species and also assigned the largest g principal value to the
By means of CW EPR at X-, Q-, and W-band frequencies in
combination with Davies-ENDOR at Q-band, the g and cobalt
hyperfine matrixes at low temperature are determined and the
temperature dependence of the g principal values is investigated.
The assignment of the g principal axes is discussed in detail.
10.1021/jp993668c CCC: $19.00 © 2000 American Chemical Society
Published on Web 03/10/2000

2920 J. Phys. Chem. B, Vol. 104, No. 13, 2000
Van Doorslaer and Schweiger
The hyperfine interactions of the surrounding protons are
determined by means of Mims-ENDOR at X-band. A pro-
nounced solvent effect is observed. Finally, the hyperfine and
nuclear quadrupole interactions of the pyridine (or 1-methylimi-
dazole) binding nitrogen and of the porphyrin nitrogens are
evaluated by means of two-dimensional HYSCORE at X-band
in combination with two-pulse ESEEM at S-band. The HY-
SCORE technique is found to give more information than three-
pulse ESEEM experiments.17,19 The observed interactions are
interpreted in terms of internuclear distances and spin density
distributions and are compared with values from the literature
and from our recent study on CoTPP(py).20
Matched HYSCORE.24 The second and third π/2 pulse were
replaced by matched pulses of length 88 ns. All other parameters
were the same as those for unmatched HYSCORE.
The Q-band CW EPR and pulse ENDOR spectra were
recorded on a home-built spectrometer (mw frequency, 35.39
GHz) equipped with a liquid Helium cryostat from Oxford Inc.
and a Bruker ENDOR ER 5106 QTE probehead. The CW EPR
spectra were recorded with a modulation amplitude of 0.2 mT
and an mw power of 10 mW. All pulse ENDOR measurements
were conducted at a repetition rate of 300 Hz and a temperature
of 15 K.
Davies-ENDOR.25 The spectra were measured with the
sequence π-T-π/2-τ-π-τ-echo, with mw pulse lengths of
tπ/2 ) 100 ns and tπ ) 200 ns and time intervals τ ) 240 ns
and T ) 17 µs. An rf π pulse of variable frequency νrf and
length 15 µs was applied during time T.
The W-band CW EPR spectra were measured on a Bruker
ELEXSYS 680 spectrometer (mw frequency, 94 GHz) equipped
with cooling systems from Oxford Inc. and Cryogenics and a
critically coupled Bruker TetraFlex probehead. An mw power
of 1.78 mW, a modulation amplitude of 0.5 mT, and a
modulation frequency of 100 kHz were used.
2. Materials and Methods
2.1. Sample Preparation. (Tetraphenylporphyrinato)cobal-
tate(II), CoTPP, from Aldrich was used without further purifica-
tion. Pyridine and 1-methylimidazole were purchased from Fluka
(pro analysis). As a solvent, toluene (Fluka, puriss., absolute,
over molecular sieves) was used. Deuterated toluene-d8 (>99.6%
purity) was obtained from Cambridge Isotope Laboratories
(CIL). Deuterated pyridine (py-d5) (> 99% purity) was pur-
chased from CIBA, and 15N-labeled pyridine ([15N]py) (>98%
purity) was obtained from CIL. The porphyrin complex was
dissolved in toluene containing 10 mM of pyridine or 1-meth-
ylimidazole. The final concentration of the CoTPP(py) and
CoTPP(1-MeIm) complex was taken as 1 mM. After the
components were mixed, the solution was transferred to an EPR
tube and left open to air to obtain the oxygenated complexes.
15N-labeled (>99% purity) cobaltic tetraphenylporphyrin
chloride, CoIII([15N]TPP)Cl was obtained from Porphyrin Prod-
ucts, Inc. To reduce CoIII([15N]TPP)Cl to the cobaltous state, it
was dissolved in degassed CH2Cl2 in a concentration of 14 mM
and then mixed with an equal volume of a solution of Na2S2O4
in degassed H2O (57 mM) for about 1 h. The CH2Cl2 phase
was separated from the aqueous phase and vacuum-distilled.
The remaining CoII[15N]TPP was then treated in the same way
as described above.
The S-band pulse EPR spectra were recorded on a home-
built spectrometer (mw frequency, 3.78 GHz).
Two-Pulse ESEEM. The spectra were measured with the
sequence π/2-τ-π-τ-echo, with pulse lengths of 20 and 40
ns. Time τ was varied from 260 to 6240 ns in steps of 20 ns.
2.3. Theory. Oxygenated Co(II) complexes contain one
unpaired electron which is mainly centered on the oxygen
moiety. The spin Hamiltonian in frequency units can be written
as
âe
H ) B0gS + SACoI + Hnucl
(1)
h
The CW EPR spectra are dominated by the electron Zeeman
interaction (first term) and the hyperfine interaction between
7
the unpaired electron and the nuclear spin (I ) /2) of cobalt
(second term). Hnucl describes the interactions with the sur-
rounding nitrogen nuclei and the protons, which can be observed
with ESEEM and ENDOR.
2.2. Equipment. The X-band CW EPR spectra were recorded
at 80 K on a Bruker ESP 300 spectrometer (microwave (mw)
frequency, 9.48 GHz), equipped with a liquid nitrogen cryostat.
An mw power of 160 mW, a modulation amplitude of 0.05
mT, and a modulation frequency of 100 kHz were used. The
X-band pulse EPR and ENDOR spectra (15 K throughout) were
recorded on a Bruker ESP380 spectrometer (mw frequency, 9.71
GHz) equipped with a liquid Helium cryostat from Oxford Inc.
The magnetic field was measured with a Bruker ER 035M NMR
gaussmeter. In all pulse EPR and ENDOR measurements a
repetition rate of 1 kHz was used. The observer positions were
carefully chosen to scan through all the molecular orientations
that contribute to the CW EPR spectrum.21
Mims-ENDOR.22 The spectra were recorded with the mw
pulse sequence π/2-τ-π/2-T-π/2-τ-echo, with pulse lengths
of 16 ns. Time τ was varied from 96 to 600 ns in steps of 8 ns.
A selective radio frequency (rf) π pulse of variable frequency
νrf and length 10 µs was applied during the time interval T.
The rf increment was set to 50 kHz.
The hyperfine interactions of the cobalt nucleus and the
protons are best determined with pulse ENDOR.26 HYSCORE,23
a two-dimensional experiment in which a mixing π pulse creates
correlations between the nuclear coherences in two different
electron spin (mS) manifolds, is found to be the most appropriate
method to study the magnetic parameters of the ligand nitrogens.
For an S ) 1/2, I ) 1/2 system (e.g. 15N), the nuclear transition
frequencies in the two mS manifolds are given by
2
A
2
B 2 1/2
2
νR(â)
)
( νI
+
(2)
[
(
)
( )
]
with the nuclear Zeeman frequency νI ) -gnânB0/h. A and B
describe the secular and pseudosecular part of the hyperfine
coupling. In the HYSCORE experiment, the correlations
between the nuclear transitions lead to cross-peaks (νR, νâ) and
(νâ, νR) in the 2D plots. The interpretation of HYSCORE spectra
HYSCORE (Hyperfine Sublevel Correlation).23 The ex-
periments were carried out with the pulse sequence π/2-τ-
π/2-t1-π-t2-π/2-τ-echo with pulse lengths tπ/2 ) 24 ns and
tπ ) 16 ns. The time intervals t1 and t2 were varied from 96 to
8272 ns in steps of 16 ns. Three τ values (96, 176, and 344 ns)
were used to remove the blind spots. An eight-step phase cycle
was used to eliminate unwanted echo contributions.
of disordered S ) /2, I ) /2 systems has been discussed by
1
1
several authors.27,28
The spin Hamiltonian of an S ) 1/2, I ) 1 system (e.g. 14N)
can be described in terms of the g matrix, the hyperfine matrix
A, and the nuclear quadrupole tensor Q. The principal val-
ues Qx, Qy, and Qz of the traceless Q tensor are usually ex-
pressed by the quadrupole coupling constant K ) e2qQ/4h

Oxygenated Cobalt(II) Heme Model Systems
J. Phys. Chem. B, Vol. 104, No. 13, 2000 2921
Figure 1. Structure of (oxyCo)TPP(py) (1a) and (oxyCo)TPP(1-MeIm)
(1b) and orientation of the g, ACo principal axes and the principal axes
of AN and QN of the sixth ligand with respect to the molecular frame
(x, y, z).
and the asymmetry parameter η, with Qx ) -K(1 - η), Qy )
-K(1 + η), and Qz ) 2K. In the case of exact cancellation (|A|
≈ 2|νI|), the effective field experienced by the nucleus in one
of the two mS manifolds vanishes. The ESEEM frequencies
within this manifold are therefore close to the nuclear quadrupole
resonance (NQR) frequencies29
ν0 ) 2Kη, ν- ) K(3 - η), ν+ ) K(3 + η)
(3)
1
HYSCORE spectra of disordered S ) /2, I ) 1 systems have
been discussed by Dikanov et al.30 The spectra are dominated
by the cross-peaks between the double-quantum (DQ) frequen-
cies31
2
1/2
a
2
νRD,Qâ ) 2 ( νI + K2(3 + η2)
(4)
[(
)
]
Figure 2. CW EPR spectra of (oxyCo)TPP(py) taken at 80 K. (e,
experimental spectra; s, simulations): (a) X-band; (b) Q-band; (c)
W-band.
where a is the hyperfine coupling at a particular observer
position.
2.4. Data Manipulation and Simulations. The data were
processed with the program MATLAB 5.1 (The MathWorks,
Inc., Natick, MA). The time traces of the HYSCORE spectra
were baseline corrected with a third-order polynomial, apodized
with a Hamming window, and zero filled. After 2D Fourier
transformation, the absolute-value spectra were calculated. Both
the Mims-ENDOR and HYSCORE spectra were recorded at
different τ values and added together to eliminate τ dependent
blind spots. The two-pulse ESEEM time traces were baseline
corrected with an exponential function, apodized with a Ham-
ming window, and zero filled. After 1D Fourier transformation,
the absolute-value spectra were calculated. To get rid of dead
time dependent line distortions in the absolute-value spectra, a
cross-term averaging procedure32 was used.
The CW EPR, Davies-ENDOR, and two-pulse ESEEM
spectra were simulated using MAGRES.33 Simulations of the
HYSCORE spectra were done with the program TRYSCORE.34
To obtain good starting values for the simulations of the
Figure 3. Q-band Davies-ENDOR spectra of (oxyCo)TPP(1-MeIm)
taken at different observer positions: (a) experimental spectra; (b)
simulations.
1
HYSCORE spectra of I ) /2 systems, the methods of Po¨ppl
and Kevan27 and Dikanov and Bowman28 were used.
these spectra. The spectral features corresponding to the three
principal g values overlap in the X-band spectrum but are nicely
separated in the W-band EPR spectrum.
3. Results
3.1. g and ACo Matrices. The X-band CW EPR spectrum of
a frozen solution of (oxyCo)TPP(py) at 80 K is shown in Figure
2a. From this spectrum, the principal values of the g and ACo
matrixes and their relative orientation are difficult to determine.
To facilitate the analysis, also the Q-band (Figure 2b) and
W-band (Figure 2c) CW EPR spectra were measured. Due to
the large g strain and second-order effects in the hyperfine
matrix, the cobalt hyperfine splittings are no longer resolved in
To determine the principal values of the ACo matrix and the
relative orientation of the principal axes with respect to the g
principal axes, cobalt Davies-ENDOR spectra at Q-band were
measured at different observer positions. Figure 3a shows some
of the recorded spectra for (oxyCo)TPP(1-MeIm). The signals
are weak and the cobalt nuclear quadrupole interaction is not

2922 J. Phys. Chem. B, Vol. 104, No. 13, 2000
Van Doorslaer and Schweiger
TABLE 1: Principal Values of the g and ACo Matrices for
(oxyCo)TPP(py) (1a) and (oxyCo)TPP(1-MeIm) (1b),
Derived from CW EPR Spectra at X-, Q-, and W-Band
Frequencies and from Q-Band Davies-ENDOR Spectraa
1a
1b
1a
1b
ACy o ( ( 1.0 MHz)
ACz o (( 1.0 MHz)
g
x′ ((0.0005)
y′ ((0.0005)
2.0020
1.9827
2.0705
2.0029
1.9836
-21.4 -21.5
-22.7 -22.5
g
gz′ ((0.0010)
2.0729 â1 ((5°)
-52.5
64
65
ACx o ( ( 1.0 MHz)
-53.0
a The Euler angles for the g principal axes are R1 ) 0 ( 5° and γ1
) 0 ( 5°; â1 is given in the table. The Euler angles for the ACo principal
axes are R2 ) â2 ) γ2 ) 0 ( 5° (see Figure 1).
Figure 5. X-band Mims-ENDOR spectra at observer position I (see
Figure 2a): (a) (oxyCo)TPP(py-d5) in toluene-d8; (b) (oxyCo)TPP(py)
in toluene-d8; (c) (oxyCo)TPP(py) in normal toluene.
Figure 4. Variation of the g values of (oxyCo)TPP(1-MeIm) as a
function of temperature, measured with Q-band CW EPR.
resolved. From the ENDOR line width, a maximum nuclear
quadrupole coupling of 130 kHz is estimated.
with the cobalt ENDOR lines, only Mims-ENDOR, which has
been found to be less sensitive to the strong cobalt couplings,
has been used.
The g and ACo principal values and the Euler angles with
respect to the molecular frame defined in Figure 1 are given in
Table 1. The signs of the cobalt hyperfine values will be
discussed later. The corresponding simulated EPR and ENDOR
spectra are shown in Figures 2 and 3b. For the ENDOR
simulations a line width of 1.8 MHz was assumed, which
accounts for the unresolved nuclear quadrupole interactions.
Although the values presented in Table 1 are the ones that
produced the best simulations, it can be seen from Figure 3
that the ENDOR simulations are still not perfect. This is due to
the fact that the nuclear quadrupole interaction is not taken into
account directly. Although this interaction is small, it will have
an influence both on the intensity and the position of the cobalt
ENDOR lines. These effects cannot be simulated by means of
a broader ENDOR line width. Surprisingly, for both complexes
the cobalt hyperfine interactions are found to be equal within
the experimental error.
Figure 5a shows the X-band Mims-ENDOR spectrum of
(oxyCo)TPP(py-d5) in toluene-d8 taken at observer position I
(see Figure 2a). The observed couplings are due to the protons
of the TPP ligand. A splitting of 0.4 MHz ((0.1 MHz) is found
for all observer positions in the EPR spectrum. For some
observer positions, the peaks show weak shoulders with a
maximal splitting of 0.7 ( 0.1 MHz (Figure 5a). Figure 5b
shows the corresponding Mims-ENDOR spectrum of (oxyCo)-
TPP(py) in toluene-d8. The additional splitting of 1.0 MHz
observed in this spectrum can be ascribed to protons of the
pyridine ligand. This splitting varies from 1.0 to 1.2 MHz when
the observer position is changed. Figure 5c shows the Mims-
ENDOR spectrum of (oxyCo)TPP(py) in normal toluene. A
strong peak at the nuclear Zeeman frequency νH is observed,
representing a large number of distant solvent protons. The broad
peak with a maximal width of 3.4 MHz varies between 2.6 and
3.4 MHz as a function of the observer position.
The Mims-ENDOR spectra of (oxyCo)TPP(1-MeIm) in
toluene-d8 (not displayed) show splittings of 0.5 ( 0.1 MHz,
which are assigned to the protons of TPP. In addition, weaker
signals with a splitting of 1.3 ( 0.1 MHz are observed that can
be assigned to the protons of 1-MeIm. The corresponding Mims-
ENDOR spectra of the complex in toluene show besides a strong
peak at νH, a broad peak with a width between 2.9 and 3.4 MHz
symmetric to νH.
It has been reported that upon cooling from 220 K to
approximately 80 K, the CW EPR spectrum of oxygenated
cobalt porphyrin complexes dissolved in toluene changes
drastically.17 This is traced back to a reduced mobility of the
O2 moiety and of the whole complex at lower temperature. To
check this behavior for the complexes under study, the tem-
perature dependence of the g values has been studied. The
change of the g values as a function of temperature can best be
determined using W-band CW EPR. Since this requires a large
amount of liquid helium (supercon sweep), the temperature
dependence of the Q-band CW EPR spectra has been measured.
Figure 4 shows for (oxyCo)TPP(1-MeIm) the change of the
three g principal values and of the calculated isotropic value
giso ) (gx′ + gy′ + gz′)/3 as a function of temperature. The same
trend was found for (oxyCo)TPP(py).
3.3. Interaction with the Surrounding Nitrogens. The
interactions of the nitrogens have been investigated with ESEEM
techniques. In particular, HYSCORE at different observer
positions is well-suited for the determination of the magnetic
parameters of nitrogen. Figure 6a shows the X-band HYSCORE
spectrum of (oxyCo)TPP(py) at observer position II (see Figure
2a). To distinguish between the peaks of the pyridine and the
porphyrin nitrogens, we also measured the corresponding
3.2. Interaction with the Surrounding Protons. Proton
interactions can be measured with the Davies- and the Mims-
ENDOR scheme. Since, at X-band, the proton signals overlap

Oxygenated Cobalt(II) Heme Model Systems
J. Phys. Chem. B, Vol. 104, No. 13, 2000 2923
Figure 7. Simulated nitrogen HYSCORE spectra of (oxyCo)TPP(py)
at observer position II: (a) 14N nucleus of pyridine; (b) 15N nuclei of
the porphyrin ring; (c) 14N nuclei of the porphyrin ring and 15N nucleus
of pyridine. The arrows indicate the cross-peaks assigned to the 15N
nucleus of pyridine.
Figure 6. X-band nitrogen HYSCORE spectra at observer position II
(see Figure 2a). (a) (oxyCo)TPP(py): The arrows indicate the DQ cross-
peaks of the interaction with the 14N of the pyridine ligand. (b) (oxyCo)-
[15N]TPP(py): The arrows indicate the peaks assigned to the 15N nuclei
of the porphyrin ligand. (c) (oxyCo)TPP([15N]py): The arrows indicate
the cross-peaks assigned to the 15N nucleus of pyridine.
TABLE 2: Principal Values of the Hyperfine Matrix and
Nuclear Quadrupole Interactions of the Nitrogen of the Base
of (oxyCo)TPP(py) (1a) and (oxyCo)TPP(1-MeIm) (1b),
Derived from X-band HYSCORE and S-Band Two-Pulse
ESEEM Spectraa
HYSCORE spectrum of (oxyCo)[15N]TPP(py) (Figure 6b) and
the matched HYSCORE spectrum of (oxyCo)TPP([15N]py)
(Figure 6c).
The weak signals of 15N in (oxyCo)TPP([15N]py) in the (-,+)
quadrant could only be observed with matched HYSCORE
(arrows in Figure 6c). The corresponding 14N signals of
(oxyCo)TPP(py) are very strong (Figure 6a). The occurrence
of the peaks in the (-,+) quadrant indicates that |A| > 2|νN|.
The positions of DQ cross-peaks (arrows in Figure 6a) depend
only slightly on the observer position, indicating that the
hyperfine coupling is dominated by the isotropic part. Since
1a
1b
ANx ((0.1 MHz)
ANy ((0.1 MHz)
ANz ((0.1 MHz)
|e2qQ/h| ((0.05 MHz)
η ((0.02)
3.4
3.4
3.7
2.95
0.23
3.2
3.2
3.8
2.25
0.10
a The Euler angles for both AN and QN are R3 ) â3 ) γ3 ) 0 ( 5°
(see Figure 1).
1
for I ) /2 spins nearly isotropic hyperfine couplings cause
shallow modulations, the corresponding 15N signals are hardly
visible. In the case of 14N, the deep modulations are caused by
the nuclear quadrupole interaction.
7c) are in good agreement with the experiment (arrows in Figure
6c). For (oxyCo)TPP(1-MeIm), the same procedure was used
(parameters see Table 2).
Figure 7a shows the simulation of the 14N HYSCORE
spectrum of the pyridine nitrogen (observer position II), which
should be compared with Figure 6b. The 15N porphyrin
interactions (arrows) are not considered in the simulation. The
parameters are given in Table 2. Taking into account that the
simulation was done with ideal pulses and that the TRYSCORE
program involves certain simplifications,34 the agreement with
the experiment is very good. The corresponding parameters for
the 15N nucleus were calculated by using the relation
The 15N peaks of the porphyrin are observed in the (+,+)
quadrant (arrows in Figure 6b). This implies that the coupling
with these nitrogens is weak (i.e. |A| < 2|νN|). The hyperfine
interaction consists of a considerable anisotropic part, which
causes the strong intensity of the 15N peaks. At S-band, the
hyperfine coupling will approximately be in exact cancellation
(|A| ≈ 2|νN|), which allows for a straightforward determination
of the nuclear quadrupole interaction (see theory). The combina-
tion of S-band and X-band HYSCORE spectra of (oxyCo)TPP-
([15N]py) would represent an ideal data set for the determination
of the nitrogen parameters. However in our experiment, the
mixing π pulse at S-band was found to be too weak to correlate
the nuclear coherences of the two mS manifolds. We therefore
used two-pulse ESEEM at S-band to measure the sharp features
gN(14N)
gN(15N)
|A1i 5N| ) |
A1i 4N
|
(5)
with i ) x, y, z. The simulated 15N cross-peaks (arrows in Figure

2924 J. Phys. Chem. B, Vol. 104, No. 13, 2000
Van Doorslaer and Schweiger
dioxygen bond axis is assumed to be parallel to the principal
axis of the largest g value.1,2
Comparison of the (oxyCo)Mb data with B12rO2 in a single
11
crystal of vitamin B12b sheds a totally different light on the
subject. For B12rO2, one data set at high temperature (HT, 313
K) with gH1 T ) 2.049, g2HT ) 2.023, and gH3 T ) 2.009, and one
data set at low temperature (LT, 113 K) with gL1T ) 2.089, g2LT
) 2.013, and gL3T ) 1.993 with different principal axes
directions are observed. EPR measurements at various inter-
mediate temperatures reveal a smooth change from gLi T to giHT
(i ) 1, 2, 3). This change in g values is explained by an
increasing thermal motion of the dioxygen fragment (wobbling
around the low-temperature O-O bond direction combined with
an oscillation). The similarities with (oxyCo)Mb are obvious.
The finding that in (oxyCo)Mb the traces of the gRT, gI, and gII
matrices are approximately the same, and at room temperature
only one data set is found, suggests a thermal motion of the O2
fragment. Moreover, the fact that the gRT principal axes
directions are approximately between those of gI and gII, and
the temperature dependence of the g values suggest a jumping
between positions I and II with an additional wobbling at higher
temperatures. The study on vitamin B12rO211 showed that the g
values and the orientation of the principal axes changed rapidly
in the temperature range 230-310 K. This implies that a
comparison of the EPR data at 300 K and the X-ray data at
258 K may lead to wrong conclusions. Moreover, the assump-
tion that the g value along the O-O bond direction should
remain the same, whereas all other parameters should drastically
change with increasing temperature, is very unlikely. The
Figure 8. S-band two-pulse ESEEM spectrum of (oxyCo)TPP(py) (1a)
and (oxyCo)TPP(1-MeIm) (1b), derived from CW EPR spectra at X-,
Q-, and W-band frequencies and from Q-band Davies-ENDOR spectra.
that correspond to the NQR frequencies ν0, ν-, and ν+ (Figure
8a). The peaks around 4-5 MHz are due to the pyridine nitrogen
and to protons.
It is difficult to get exact magnetic parameters for the pyrrole
nitrogens, because the interactions are weak and the four
nitrogens are not equivalent. Moreover, simulations show that
for coinciding AN and QN principal axes only the DQ cross-
peaks are observed, in contradiction to the experiment. The
computed hyperfine and nuclear quadrupole interactions are
collected in Table 3. The four nitrogens are assumed to be
geometrically equivalent with the Az axes pointing approximately
in the direction of the dioxygen fragment and the axes of the
largest Q value pointing along the N-Co bond. The errors in
the angles are rather large and an inequivalence of the four
nitrogens cannot be ruled out.
11
temperature dependence of the g matrix of vitamin B12rO2
shows that the smallest gHT value smoothly changes to the
smallest gLT value. For (oxyCo)Mb, this would mean that the
O-O bond direction taken at room temperature along g3RT
coincides with the gI3 (g3II) axis at 77 K. This is, however, not
possible, since, from the 17O hyperfine data of (oxyCo)Mb at
77 K,6 it is known that the gI3 (gI3I) axis is parallel to the lobe of
the π* orbital containing the unpaired electron. All these
arguments are in favor of the assignment of the largest g value
to the O-O bond direction, in accordance with earlier studies.1,2
Furthermore, Dickinson and Symons35 reconsidered their data
of (oxyCo)Mb and found that correspondence of the O-O
direction with g ) 2.00 presented them with severe difficulties.
In their opinion, the only reasonable interpretation of their data
and those of Hori8,9 could be obtained by taking gmax along the
O-O bond. Finally, the angles â1 for (oxyCo)TPP(py) and
(oxyCo)TPP(1-MeIm) are in good agreement with X-ray studies
on (oxyCo)Mb and cobalt porphyrins containing axially bound
nitric oxide.36,37
Figure 7b shows the simulated 15N pyrrole spectrum (compare
with Figure 6b, arrows) and Figure 7c gives the corresponding
14
spectrum of the
N pyrrole nitrogens (see Figure 6c). The
simulation of the two-pulse ESEEM spectrum (Figure 8a) is
shown in Figure 8b. The proton interactions are not considered,
which explains the different shapes of the signal in the 4-5
MHz region. The observed NQR signals of the pyrrole nitrogens
are nicely reproduced. For (oxyCo)TPP(1-MeIm), the same
procedure was used (parameters, see Table 3).
4. Discussion
4.1. g and ACo Matrices. For the assignment of the g
principal axes with respect to the O-O bond direction, the work
of Hori et al.8,9 on single crystals of (oxyCo)Mbs has to be
discussed in more detail. For single crystals of (oxyprotoCo)-
Mb and (oxymesoCo)Mb the authors found one set of g values
at room temperature (300 K) and two sets at 77 K. At room
temperature, the observed g values for (oxymesoCo)Mb are
gR1 T ) 2.065, g2RT ) 2.009, and g3RT ) 2.000, with the g3RT
principal axis (g value closest to ge) close to the O-O bond
direction as derived from X-ray data at 258 K. For the two forms
at 77 K the g values g1I ) 2.072, g2I ) 2.024, g3I ) 1.975 and gI1I
) 2.080, gI2I ) 2.005, g3II ) 1.987 are reported with the gI1 and
gI1I axes including an angle of approximately 90°. In agreement
with the room-temperature data, the principal axes of the g
values closest to ge (gI and gII) are taken along the O-O bond
According to Figure 4, the g values of (oxyCo)TPP(py) and
(oxyCo)TPP(1-MeIm) vary only slightly in the temperature
range 80-140 K (the systematic error for the temperature range
is about 10 K). From 140 to 180 K, a large change in the g
values is observed, indicating an increased mobility of the O-O
group and the whole molecule. Since the g matrix remains
orthorhombic, the O-O fragment does not freely rotate around
the Co-O bond. In a case of fast rotation of the O-O group
around the Co-O bond, but slow motion of the molecule, the
g matrix would be axial.17 Above 180 K, the O-O group and
the molecule as a whole are rotating rapidly, giving rise to an
isotropic EPR spectrum, in accordance with the melting point
of toluene (178 K). The slight decrease of giso with increasing
temperature can be explained by a weakening of the Co-O bond
at higher temperatures. Finally, Figure 4 shows clearly that, in
solution, upon increased mobility of the O-O group at higher
2
2
direction in contrast with all previous assignments where the

Oxygenated Cobalt(II) Heme Model Systems
J. Phys. Chem. B, Vol. 104, No. 13, 2000 2925
TABLE 3: Principal Values and Principal Axes Directions of the Hyperfine Matrix and Nuclear Quadrupole Tensor of the
Four Nitrogens of the Porphyrin Ring of (oxyCo)TPP(py) (1a]) and (oxyCo)TPP(1-MeIm) (1b), Obtained through Simulation
from X-Band HYSCORE Spectra and S-band Two-Pulse ESEEM Spectra
ANx′′
ANy′′
ANz′′
aNiso
|e2qQ/h|
((0.2 MHz) ((0.2 MHz) ((0.2 MHz) ((0.2 MHz) ((0.1 MHz) η ((0.2) R4 ((20°) R5 ((20°) â4 ((20°) â5 ((20°) γ4 ((20°) γ5 ((20°)
1a
1b
-1.3
-1.3
-1.3
-1.3
-1.2
-1.2
-1.2
-1.2
-1.3
-1.3
-1.3
-1.3
-1.2
-1.2
-1.2
-1.2
-0.4
-0.4
-0.4
-0.4
-0.3
-0.3
-0.3
-0.3
-1.0
-1.0
-1.0
-1.0
-0.9
-0.9
-0.9
-0.9
1.8
1.8
1.8
1.8
1.7
1.7
1.7
1.7
0.3
0.3
0.3
0.3
0.5
0.5
0.5
0.5
45
135
225
315
45
135
225
315
45
135
225
315
45
135
225
315
45
45
45
45
45
45
45
45
90
90
90
90
90
90
90
90
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
temperature, the g values change, but their order stays the same.
It is then very hard to imagine that in single crystals of (oxyCo)-
Mb increased thermal motion of the dioxygen fragment would
change the order of the three g values. This is again an argument
in favor of the largest g value along the O-O bond.
pairing model), an upper limit for (R′′)2 of 0.06 is found. Since
the individual values of f, h, and l may not be derived from eq
7, f or h can actually be negative implying direct interaction
rather than spin polarization. In fact, for some oxygenated
Co(II) complexes, f + h was found to be negative.1,11 Thus, the
spin-pairing model appears to be incomplete to describe the
cobalt hyperfine structure. Moreover, the cobalt hyperfine
interaction is virtually independent of the axial base. For base-
free (oxyCo)TPP, however, significantly larger (absolute) values
are found.1
For both complexes the cobalt hyperfine interactions are
found to be equal within the experimental errors. The electronic
structure of oxygenated Co(II) complexes is usually discussed
- 10
in terms of the superoxide formulation (CoIIIO2
)
or by the
spin-pairing model.1,38
In the spin-pairing model, it is assumed that the two molecular
orbitals (MO’s) of the Co-O2 fragment
Upon addition of a nitrogen base to CoTPP, the unpaired
2
electron resides in the 3dz (Co) + pσ(N) orbital. The 3dyz (3dxz)
orbital combine with the pπ(N) orbital, but the energy shift is
ψ1 ) R′dz2 + 4sγ + âπ*(x′) + R′′′dxz
2
smaller than the one for the 3dz orbital. If spin polarization of
2
the 3dz orbital plays an important role for the oxygenated
ψ2 ) R′′dyz + ꢀπ*(y′)
(6)
complexes, a change of the base strength would have a
significant influence on the cobalt hyperfine interaction, which
is not observed. The difference in the cobalt hyperfine interaction
for (oxyCo)TPP and (oxyCo)TPP(B) (B ) nitrogen base) seems
mostly to be due to the change in the 3dyz level, spin-polarization
affect the cobalt hyperfine interaction. The doubly occupied MO,
ψ1, describes the overlap of a π* (x′) orbital of O2 with the
2
cobalt 3dz , 3dxz, and 4s orbitals. The second MO, ψ2, contains
the unpaired electron. Since the two MO’s are close in energy,38
2
of the 3dz orbital plays only a minor role for this interaction.
ψ2 can polarize ψ1, resulting in a negative spin density in the
Two recent studies report on the g and ACo matrices of
(oxyCo)TPP(1MeIm) in toluene.16,17 The first study16 involves
the investigation of (oxyCo)[(o-R)TPP](1MeIm) (where R )
-H, -NHCOC(CH3)3, -NHCOCH3, -NHCONHC6H5, an
ortho substitutent on one of the four meso phenyls of TPP) in
different solvents. The g and ACo matrices were assumed to be
coaxial, which is shown to be incorrect. Accordingly, their
results for (oxyCo)TPP(1MeIm) in toluene deviate considerably
from ours. For the different systems, the authors calculated the
3d orbital spin density of cobalt from the ACo principal values.
In view of the present study, these data should be revised. In
the second work,17 the g and ACo matrices where determined
from X- and Q-band CW EPR spectra. Although the authors
included the non-coaxiality of the g and ACo principal axes,
there is still a discrepancy with our results (overestimation of
AxCo and underestimation of AyCo and ACz o). The larger experi-
mental data set used in the present work (CW EPR at three
microwave frequencies in combination with ENDOR) justifies
our results. Moreover, the simulation shown in ref 17 do not
match the experimental spectra.
2
3dz , 3dxz, and 4s orbitals.
The principal values of the cobalt hyperfine matrix are then
found to be
Ax ) -1/2T| + aCisoo + 2/7(f - 2l - h)
Ay ) -1/2T| + aCisoo + 2/7(f + l + 2h)
Az ) + T| + aCisoo + 2/7(-2f + l - h)
(7)
where T| ) 3 MHz is the direct point dipole-dipole interaction
between the unpaired electron in the π*(y′) orbital and the cobalt
nuclear spin38 and f ) PFOUCo-O(R′)2, h ) PFOUCo-O(R′′′)2,
2
2
and l ) P(R′′) describe the indirect spin polarization of dz
and dxz and the direct hyperfine contribution from dyz, respec-
tively. UCo-O is the spin-polarization constant, FO is the spin
density on the adjacent oxygen, and P ) geâegnân r-3 3d ) 600
MHz.38
Hoffman et al.,10 on the other hand, proposed a mechanism
for the description of the anisotropic cobalt hyperfine interaction
in a MO (ψ1) involving the antibonding π*(y′) oxygen orbital
and the cobalt 3dyz orbital (π back-bonding). In eq 7, this implies
that f ) h ) 0. This assumption leads to (R′′)2 ≈ 0.05 for both
complexes, which essentially means that almost complete
electron transfer from cobalt to oxygen has taken place
(superoxo formulation). The model does, however, not fully
describe the anisotropy of the cobalt hyperfine matrix.
The spin-pairing model does not allow for a large transfer of
the unpaired electron to the oxygen. Equation 7 leads to f + h
) 6.8 MHz and l + h ) 36.2 MHz for both complexes under
study. If f and h are taken positive (which is expected in a spin-
4.2. Proton Interactions. For several ferrous prophyrin
systems, the P50 value of oxygen binding was found to decrease
with increasing polarity of the solvent.39 The same effects have
been observed for cobaltous porphyrins.2 This demonstrates the
role of the solvent polarity in governing dioxygen affinity. For
both complexes under study, the largest proton hyperfine
coupling can be assigned to toluene protons. This proves the
interaction between the solvent molecules and the dioxygen
moiety.
Assuming aiso ) 0, the coupling of 0.4-0.7 MHz for the
TPP protons of both complexes corresponds to a distance of

2926 J. Phys. Chem. B, Vol. 104, No. 13, 2000
Van Doorslaer and Schweiger
0.5-0.6 nm. The protons of the porphyrin ring and the ortho-
protons of the phenyl groups fall within this range. The dioxygen
is clearly not stabilized by a hydrogen bound to the ortho-proton
of one of the phenyl groups. When the ortho-protons are
replaced by electron withdrawing amide groups (picket fence
porphyrins), an amide proton on the picket fence is believed to
interact with the dioxygen, explaining the increased oxygen
affinity of these complexes.16,40-42
the broad single-quantum peaks, which are usually lost in three-
pulse ESEEM experiments due to the instrumental dead time.
In the HYSCORE spectra, these peaks are clearly seen, so that
the assignment of the axes can easily be done through
simulation. The orientation of the QN tensor suggests that the
nuclear quadrupole coupling is mainly governed by the pσ
population and not by the population in the pyridine π system.
In the latter case, the largest Q value (in absolute value) would
be found perpendicular to the Co-N direction.
For (oxyCo)TPP(py) a maximum splitting of 1.0-1.2 MHz
is observed for the pyridine protons, a coupling which is
probably due to the ortho-protons of pyridine. For CoTPP(py),
a maximum splitting of 7.1 MHz was observed for these
protons.20 The decrease of the proton hyperfine coupling reflects
the electron withdrawal by the dioxygen molecule, which is also
manifested in the cobalt hyperfine coupling. For (oxyCo)TPP-
(1-MeIm), an analogous maximal splitting of 1.3 ( 0.1 MHz
was found for the 1-MeIm protons.
For (oxyCo)TPP(1-MeIm), the hyperfine data of the coordi-
nating 14N of 1-MeIm are very close to those observed in
(oxyCo)TPP(py) (aiso ) 3.4 MHz, FN ) 0.0022, and r ) 0.32
( 0.04 nm), in agreement with the finding that the cobalt
hyperfine interaction is similar in both complexes. The differ-
ence in the 14N QN tensor is, however, significant. For free
imidazole the NQR data |e2qQ/h| ) 3.220 MHz and η ) 0.119
are reported.45 For 1-MeIm, similar values are expected. Both
|e2qQ/h| and η are smaller than the ones for the free pyridine,46
as is also reflected in our data. Furthermore, the coordination
of 1-MeIm to the Lewis acid (oxyCo)TPP reduces the value of
|e2qQ/h| in accordance with the observations for (oxyCo)TPP-
4.3. Nitrogen Interactions. For (oxyCo)TPP(py), the ob-
served 14N hyperfine interaction of the pyridine ligand consists
of an isotropic part aiso ) 3.5 MHz and a dipolar part (-0.1,-
0.1,+0.2) MHz. From aiso, a spin density on the nitrogen of FN
) 0.0023 (FN ) aiso/a0 with a0 ) 1538.22 MHz 43) is obtained,
which is about a factor of 10 smaller than is found for oxygen
free CoTPP(py).20 This again agrees with the withdrawal of the
unpaired electron by the oxygen. From the dipolar part, a
distance r ) 0.39 ( 0.06 nm between the nitrogen nucleus and
the unpaired electron can be derived.44
For the pyridine 14N nuclear quadrupole interaction, we
evaluate the values |e2qQ/h| ) 2.95 MHz and η ) 0.23. Hsieh
et al.46 found from nuclear quadrupole resonance experiments
the values |e2qQ/h| ) 4.584 MHz and η ) 0.396 for the free
pyridine molecule. These authors also observed that upon
coordination of pyridine with a Lewis acid, the electric field
gradient at the nitrogen nucleus decreases. For CoTPP(py),
|e2qQ/h| decreases to 3.4 MHz,20 and upon oxygenation, this
value decreases even further. Hsieh et al.46 found a linear relation
between |h/e2qQ| and η. Using their simplified equation, η )
0.23 corresponds with |e2qQ/h| ) 3.05 MHz, which agrees
nicely with our findings.
(py). Using three-pulse ESEEM, Lee et al.16 found for the 14
N
interaction aiso ) 3.54 ( 0.06 MHz, r ) 0.34 ( 0.02 nm, |e2qQ/
h| ) 2.39 ( 0.07 MHz, and η ) 0.88 ( 0.13. The difference
from our results can again be ascribed to the shortcomings of
the applied technique (1D experiment and only one π value). It
should be noticed that the error limits given by Lee et al.16 were
clearly underestimated.
Until now, the remote nitrogen of 1-MeIm has not been
discussed. On the basis of the data we found for the binding
nitrogen of 1-MeIm, we expect the hyperfine interaction with
the remote nitrogen to be very small. For free imidazole the
NQR data for the amino nitrogen |e2qQ/h| ) 1.391 MHz and η
) 0.930 are reported.45 Simulations of HYSCORE spectra using
an isotropic hyperfine interaction of 100 kHz and the nuclear
quadrupole data from the free imidazole showed two broad
diagonal peaks at ≈1.2 and ≈2.3 MHz in the (+,+) quadrant.
In the HYSCORE spectra of (oxyCo)TPP(1-MeIm), such peaks
were indeed observed, confirming that the hyperfine interaction
with the remote nitrogen is small. Slight changes in the nuclear
quadrupole parameters had virtually no influence on the data.
For the interaction with the pyridine nitrogen in (oxyCo)-
TPP(py), Magliozzo et al.19 derived from three-pulse ESEEM
the parameters Axx ) 2.9 MHz, Ayy ) 3.4 MHz, Azz ) 3.9 MHz,
|e2qQ/h| ) 2.9 MHz, η ) 0.2, R ) 60°, â ) γ ) 0°, with the
Euler angles R, â, and γ, describing the relations between the
principal axes of AN and QN. The orthorhombicity of the AN
matrix and the noncoaxiality of the AN and QN axes could not
The hyperfine interactions with the porphyrin nitrogens of
(oxyCo)TPP(py) ((oxyCo)TPP(1-MeIm)), can be split into an
isotropic part with aiso ) -1 (-0.9) MHz and a dipolar part
with (-0.3, -0.3, +0.6) MHz. The dipolar part corresponds to
a distance r ) 0.27 ( 0.04 nm between the nitrogen and the
unpaired electron. The orientation of the hyperfine principal axes
supports again the assumption that an almost complete transfer
of the unpaired electron from cobalt to the dioxygen has taken
place, as suggested by Hoffman et al.10 A comparison of the
parameters of the pyrrole nitrogens of CoTPP(py) and (oxyCo)-
TPP(py) shows that the electron withdrawal by the oxygen has
a strong influence on the hyperfine interaction. For CoTPP(py)
the nitrogen orbital involved in the MO containing the unpaired
electron is an sp2-type hybrid; hence, there is direct spin density
in the nitrogen 2s orbital. This results in the positive aiso value.
For (oxyCo)TPP(py), we found a negative aiso value of about
-1 MHz for the pyrrole nitrogens, indicating that the s-spin
density is mainly due to polarization effects. This is in agreement
with the ENDOR data of N,N′-ethylenebis(acetylacetonatimi-
be confirmed by our measurements. However, the shallow 15
N
modulations do not allow for such a large anisotropy. The
inconsistency with our results may originate from the fact that
the authors measured the three-pulse ESEEM spectrum only at
one observer position (approximately position I, Figure 2a) and
for only one τ value (blind spots). Because of the small g
anisotropy, they averaged in the powder simulations over all
possible magnetic field orientations. Our measurements show,
however, a distinct orientation selectivity. The authors had to
assume an orthorhombic hyperfine matrix to account for some
of the spectral features, which in fact result from poor orientation
selection. We could nicely simulate the HYSCORE spectra at
different observer positions by using the parameter set given in
Table 2.
Magliozzo et al.19 also had difficulties in figuring out whether
the largest principal axis of the QN tensor is aligned along the
Co-N bond direction or along the 2pz orbital of the pyridine
nitrogen. Information about this direction can be derived from
2
nato)cobalt(II), Co(acacen).47,48 For Co(acacen) a | A,yz ground
state is assumed and negative aiso values of about -3.5 MHz
are reported for the binding nitrogen. This situation resembles

Oxygenated Cobalt(II) Heme Model Systems
J. Phys. Chem. B, Vol. 104, No. 13, 2000 2927
(2) Jones, R. D.; Summerville, D. A.; Basolo, F. Chem. ReV. 1979,
the case of (oxyCo)TPP(py), where the unpaired electron resides
in a MO to which the 3dyz orbital contributes (see eq 6).
In contrast to the hyperfine data, the nuclear quadrupole
parameters of the pyrrole nitrogens change only slightly upon
oxygenation of the CoTPP(py) complex (for CoTPP(py): |e2qQ/
h| ) 1.8 MHz; η ) 0.55 20). This is because the hyperfine
interaction is determined by the distribution of the unpaired
electron, whereas the nuclear quadrupole tensor is determined
79, 139.
(3) Hoffman, B. M.; Petering, D. H. Proc. Natl. Acad. Sci. U.S.A. 1970,
67, 637.
(4) Ikeda-Saito, M.; Brunori, M.; Yonetani, T. Biochim. Biophys. Acta
1978, 533, 173, and further references cited therein.
(5) Hu¨ttermann, J.; Stabler, R. In Electron Magnetic Resonance of
Disordered Systems; Yordanov, N. D., Ed.; World Scientific: Singapore,
1989; pp 127-148, and further references cited therein.
(6) Dickinson, L. C.; Chien, J. C. W. Proc. Natl. Acad. Sci. U.S.A.
1980, 77, 1235.
(7) Chien, J. C. W.; Dickinson, L. C. Proc. Natl. Acad. Sci. U.S.A.
1972, 69, 2783.
(8) Hori, H.; Ikeda-Saito, M.; Yonetani, T. Nature 1980, 288, 501.
(9) Hori, H.; Ikeda-Saito, M.; Yonetani, T. J. Biol. Chem. 1982, 257,
3636.
14
by the electric field gradient at the N nucleus generated by
the surrounding charges. It is interesting to note that the nuclear
quadrupole parameters of the binding nitrogens of Co(acacen)
resemble those reported here for the pyrrole nitrogens (N1,
|e2qQ/h| ) 1.79 MHz, η ) 0.69; N2, |e2qQ/h| ) 1.84 MHz, η
) 0.59 47). This may again be ascribed to the electronic
similarities as described above.
(10) Hoffman, B. M.; Diemente, D. L.; Basolo, F. J. Am. Chem. Soc.
1970, 92, 61.
(11) Jo¨rin, E.; Schweiger, A.; Gu¨nthard, Hs. H. J. Am. Chem. Soc. 1983,
105, 4277.
(12) Zeller, H. R.; Ka¨nzig, W. HelV. Phys. Acta 1967, 40, 845.
(13) Lee, H. C.; Peisach, J.; Dou, Y.; Ikeda-Saito, M. Biochemistry 1994,
44, 7609.
(14) Lee, H. C.; Ikeda-Saito, M.; Yonetani, T.; Magiozzo, R. S.; Peisach,
J. Biochemistry 1992, 31, 7274.
(15) Lee, H. C.; Peisach, J.; Tsuneshige, A.; Yonetani, T. Biochemistry
1995, 34, 6883.
(16) Lee, H. C.; Scheuring, E.; Peisach, J.; Chance, M. R. J. Am. Chem.
Soc. 1997, 119, 12201.
(17) Bowen, J. H.; Shokhirev, N. V.; Raitsimring, A. M.; Buttlaire, D.
H.; Walker, F. A. J. Phys. Chem. B 1997, 101, 8683.
(18) Walker, F. A. J. Am. Chem. Soc. 1970, 92, 4235.
(19) Magliozzo, R. S.; McCracken, J.; Peisach, J. Biochemistry 1987,
26, 7923.
(20) Van Doorslaer, S.; Bachmann, R.; Schweiger, A. J. Phys. Chem.
1999, 103, 5446.
(21) Rist, G.; Hyde, J. J. Chem. Phys. 1970, 52, 4633.
(22) Mims, W. B. Proc. R. Soc. London 1965, 283, 452.
(23) Ho¨fer, P.; Grupp, A.; Nebenfu¨hr, H.; Mehring, M. Chem. Phys.
Lett. 1986, 132, 279.
(24) Jeschke, G.; Rakhmatullin, R.; Schweiger, A. J. Magn. Reson. 1998,
131, 261.
5. Conclusion
Frozen toluene solutions of two oxygenated Co(II) heme
model systems, (oxyCo)TPP(py) and (oxyCo)TPP(1-MeIm), are
studied by means of different one- and two-dimensional EPR
and ENDOR methods at microwave frequencies between 3.7
and 95 GHz. The use of these methods together with isotopic
substitutions allows a detailed analysis of these systems.
(a) The g and ACo matrices and the direction of the principal
axes are determined by CW EPR at X-, Q-, and W-bands and
Davies-ENDOR at Q-band. The change of the g values as a
function of temperature is studied by means of Q-band CW EPR.
The ambiguity in the assignment of the g principal axes to the
dioxygen bond direction is discussed in detail. Strong arguments
in favor of the assignment of gmax along the O-O bond are
given. On the basis of the interpretation of the cobalt hyperfine
data, we found that the compounds studied in this work are
adequately described by the CoIIIO2- model rather than the spin-
pairing model.
(25) Davies, E. R. Phys. Lett. 1974, A47, 1.
(26) Gemperle, C.; Schweiger, A. Chem. ReV. 1991, 91, 1481.
(27) Po¨ppl, A.; Kevan, L. J. Phys. Chem. 1996, 100, 3387.
(28) Dikanov, S. A.; Bowman, M. K. J. Magn. Reson. 1995, A116, 125.
(29) Flanagan, H. L.; Singel, D. J. J. Chem. Phys. 1987, 87, 5606.
(30) Dikanov, S. A.; Xun, L.; Karpiel, A. B.; Tyryshkin, A. M.; Bowman,
M. K. J. Am. Chem. Soc. 1996, 118, 8408.
(b) The proton couplings (intra- and extramolecular) are
investigated using Mims-ENDOR at X-band. A considerable
solvent effect is observed.
(c) By means of HYSCORE at X-band and two-pulse ESEEM
at S-band, the hyperfine and nuclear quadrupole interactions of
the nitrogen of the axial ligand and the nitrogens in the porphyrin
ring are investigated. The hyperfine interaction with the axial
ligand nitrogen is found to be predominately isotropic. The
nuclear quadrupole tensor points along the Co-N bond direc-
tion. The influence of the base is discussed in detail. For the
pyrrole nitrogens, AN and QN are not coaxial. The hyperfine
interactions have a considerable anisotropic contribution arising
from the dipolar interaction with the unpaired electron on the
dioxygen fragment, and the isotropic hyperfine coupling is found
to be negative. A change of the axial ligand has no significant
influence on this interaction.
(31) Dikanov, S. A.; Tsvetkov, Yu. D.; Bowman, M. K.; Astashkin, A.
V. Chem. Phys. Lett. 1982, 90, 149.
(32) Van Doorslaer, S.; Sierra, G.; Schweiger, A. J. Magn. Reson. 1999,
136, 152.
(33) Keijzers, C. P.; Reijerse, E. J.; Stam, P.; Dumont, M. F.; Gribnau,
M. C. M. J. Chem. Soc., Faraday Trans. 1 1987, 83, 3493.
(34) Szosenfogel, R.; Goldfarb, D. Mol. Phys. 1998, 95, 1295.
(35) Dickinson, L. C.; Symons, M. C. R. Chem. Soc. ReV. 1993, 12,
387.
(36) Brucker, E. A.; Olson, J. S.; Phillips, G. N.; Dou, Y.; Ikeda-Saito,
M. J. Biol. Chem. 1996, 271, 25419.
(37) Richter-Addo, G. B.; Hodge, S. J.; Yi, G.-B.; Khan, M. A.; Ma,
T.; Van Caemelbecke, E.; Guo, N.; Kadish, K. M. Inorg. Chem. 1996, 35,
6530.
(38) Trovog, B. S.; Kitko, D. J.; Drago, R. S. J. Am. Chem. Soc. 1976,
98, 5144.
(39) Momenteau, M.; Reed, C. A. Chem. ReV. 1994, 94, 659.
(40) Mispelter, J.; Momenteau, M.; Lavalette, D.; Lhoste, J.-M. J. Am.
Chem. Soc. 1983, 105, 5165.
(41) Gerothanassis, I. P.; Momenteau, M.; Loock, B. J. Am. Chem. Soc.
1989, 111, 7006.
In this paper, the potential of EPR and ENDOR techniques
at different microwave frequencies is clearly demonstrated.
Acknowledgment. This research has been supported by the
Swiss National Science Foundation. Dr. Gunnar Jeschke and
Prof. Dr. W. Spiess of the Max Plank Institut in Mainz,
Germany, are gratefully thanked for the W-band measurements.
Dr. Igor Gromov is kindly thanked for his helpful assistance
with the Q-band spectrometer. We are grateful to Dr. R.
Bachmann for helpful discussions.
(42) Jameson, G. B.; Drago, R. S. J. Am. Chem. Soc. 1985, 107, 3017.
(43) Koh, A. K.; Miller, D. J. At. Data Nucl. Data Tables 1985, 33,
235.
(44) Hurst, G. C.; Henderson, T. A.; Kreilick, R. W. J. Am. Chem. Soc.
1985, 107, 7294.
(45) Garcia, M. L. S.; Smith, J. A. S.; Bavin, P. M. G.; Ganellin, C. R.
J. Chem. Soc., Perkin Trans. 2, 1983, 1391.
(46) Hsieh, Y.-N.; Rubenacker, G. V.; Cheng, C. P.; Brown, T. L. J.
Am. Chem. Soc. 1977, 99, 1384.
(47) Rudin, M.; Schweiger, A.; Gu¨nthard, Hs. H. Mol. Phys. 1982, 46,
1027.
(48) Rudin, M.; Schweiger, A.; Gu¨nthard, Hs. H. Mol. Phys. 1982, 47,
171.
References and Notes
(1) Smith, T. D.; Pilbrow, J. R. Coord. Chem. ReV. 1981, 39, 295.