The remarkable axial lability of iron(III) corrole complexes

Article abstract of DOI:10.1021/ja044090+

Flash photolysis of nitrosyl tris(aryl)corrolate complexes of iron(III), Fe(Ar₃C)(NO) (Ar₃C^3- = 5,10,15-tris(4-nitro- phenyl)corrolate (TNPC^3-), 5,10,15-tris(phenyl)corrolate (TPC ^3-) or 5,10,15-tris(4-tolyl)-corrolate (H₃TTC ^3-)) leads to NO labilization. This is followed by the rapid reaction of NO with Fe^III(C) to regenerate the starting complex. The second-order rate constants for the back reactions (k_NO) were determined to be many orders of magnitude faster than the corresponding reactions of ferric porphyrin complexes and indeed are reminiscent of the very large values seen for those of the corresponding ferrous porphyrin analogues. These data are interpreted in terms of the strongly electron-donating character of the trianionic corrolate ligand and the likely triplet electronic configuration of the iron(III) complex. These reduce the affinity of the metal centers to Lewis bases to the extent that axial ligands bind very weakly or not at all. This property is illustrated by the nearly identical k_NO values (～10⁹ M^-1 s^-1 at 295 K) recorded for the back reaction of Fe^III(TNPC) with NO after flash photolysis of Fe(TNPC)(NO) in toluene solution and in THF solution. Softer Lewis bases have a somewhat greater effect; for example, studies in 1:9 (v:v) acetonitrile:toluene and 1:9 pyridine:toluene gave k_NO values decreased ～33% and ～85%, respectively, but these both remain > 10⁸ M^-1 s^-1. The potential roles of Lewis bases in controlling the dynamics of NO addition to Fe(TNPC) in toluene was investigated in greater detail by determining the rates as a function of pyridine concentration over a wide range (10^-4 to 2.5 M). These data suggest that, while a monopyridine complex, presumably Fe(TNPC)(py), is readily formed (K ≈ 10⁴ M), this species is about one-sixth as reactive as Fe(TNPC) itself. It appears that a much less reactive bis(pyridine) complex also is formed at high [py] but the equilibrium constant is quite small (<1 M^-1).

Full text of DOI:10.1021/ja044090+

Published on Web 04/14/2005
The Remarkable Axial Lability of Iron(III) Corrole Complexes
Crisjoe A. Joseph and Peter C. Ford*
Contribution from the Department of Chemistry and Biochemistry, UniVersity of California,
Santa Barbara, Santa Barbara, California 93106-9510
Received September 28, 2004; E-mail: ford@chem.ucsb.edu
-
Abstract: Flash photolysis of nitrosyl tris(aryl)corrolate complexes of iron(III), Fe(Ar
3
C)(NO) (Ar
3
C3
)
3-
3-
5
,10,15-tris(4-nitro-phenyl)corrolate (TNPC ), 5,10,15-tris(phenyl)corrolate (TPC ) or 5,10,15-tris(4-tolyl)-
corrolate (H TTC )) leads to NO labilization. This is followed by the rapid reaction of NO with Fe (C) to
3
3-
III
regenerate the starting complex. The second-order rate constants for the back reactions (kNO) were
determined to be many orders of magnitude faster than the corresponding reactions of ferric porphyrin
complexes and indeed are reminiscent of the very large values seen for those of the corresponding ferrous
porphyrin analogues. These data are interpreted in terms of the strongly electron-donating character of
the trianionic corrolate ligand and the likely triplet electronic configuration of the iron(III) complex. These
reduce the affinity of the metal centers to Lewis bases to the extent that axial ligands bind very weakly or
9
-1 -1
not at all. This property is illustrated by the nearly identical kNO values (∼10 M
s
at 295 K) recorded
III
for the back reaction of Fe (TNPC) with NO after flash photolysis of Fe(TNPC)(NO) in toluene solution
and in THF solution. Softer Lewis bases have a somewhat greater effect; for example, studies in 1:9 (v:v)
acetonitrile:toluene and 1:9 pyridine:toluene gave kNO values decreased ∼33% and ∼85%, respectively,
8
-1 -1
but these both remain >10 M
s . The potential roles of Lewis bases in controlling the dynamics of NO
addition to Fe(TNPC) in toluene was investigated in greater detail by determining the rates as a function
-4
of pyridine concentration over a wide range (10 to 2.5 M). These data suggest that, while a monopyridine
4
complex, presumably Fe(TNPC)(py), is readily formed (K ≈ 10 M), this species is about one-sixth as
reactive as Fe(TNPC) itself. It appears that a much less reactive bis(pyridine) complex also is formed at
high [py] but the equilibrium constant is quite small (<1 M-1).
Introduction
sively studied the formation and reactivities of the nitrosyl
complexes of various metallo-porphyrins including several ferri-
Nitric oxide (NO, nitrogen monoxide) has an established
repertoire in mammalian biology with well-known roles in the
cardiovascular, neurological, and immune response systems.
A key component of the mammalian biochemistry of nitric oxide
involves the interactions between NO and metal centers. The
and ferro-heme proteins and models. In the present investigation,
we turn to the NO reactivity with an iron(III) center in a related
π-unsaturated macrocyclic tetrapyrrole ligand, namely, the
1
III
2
corrole complexes, Fe (C),
photochemical and thermal reactions of metal nitrosyl complexes
also draw interest from the possibilities that such species may
prove useful in strategies for the therapeutic delivery of NO to
(
4) (a) Hoshino, M.; Ozawa, K.; Seki, H.; Ford, P. C. J. Am. Chem. Soc. 1993,
1
15, 9568-9575. (b) Hoshino, M.; Maeda, M.; Konishi, R.; Seki, H.;
Ford, P. C. J. Am. Chem. Soc. 1996, 118, 5702-5707. (c) Laverman,
L. E.; Hoshino, M.; Ford, P. C. J. Am. Chem. Soc. 1997, 119, 12663-
3
specific tissue sites. Accordingly, this laboratory has been
1
2664. (d) Lorkovic, I. M.; Miranda, K. M.; Lee, B.; Bernhard, S.;
systematically investigating the reactions controlling the forma-
tion of metal nitrosyl complexes, as well as the subsequent
redox, nitrosation, and other reactions in which these species
Schoonover, J. R.; Ford, P. C. J. Am. Chem. Soc. 1998, 120, 11674-
1
1683. (e) Lorkovic, I. M.; Ford, P. C. Inorg. Chem. 1999, 38, 1467-
473. (f) Lorkovic, I. M.; Ford, P. C. J. Am. Chem. Soc. 2000, 122, 6516-
1
6517. (g) Laverman, L. E.; Wanat, A.; Oszajca, J.; Stochel, G.; Ford,
P. C.; van Eldik, R. J. Am. Chem. Soc. 2001, 123, 285-293. (h)
Laverman, L. E.; Ford, P. C. J. Am. Chem. Soc. 2001, 123, 11614-11622.
4
5
participate. Among such studies, we and others have exten-
(i) Kurtikyan, T. S.; Martirosyan, G. G.; Lorkovic, I. M.; Ford, P. C. J.
(
(
(
1) (a) Ignarro, L. J., Ed. Nitric Oxide, Biology and Pathology; Academic Press,
San Diego, 2000. (b) Richter-Addo, G. B., Legzdins, P., Burstyn, J., Eds.
Chem. ReV. 2002, 102, 857-1270.
Am. Chem. Soc 2002, 124, 10124-10129. (j) Leal, F. A.; Lorkovic, I.
M.; Ford, P. C.; Lee, J.; Chen, L.; Torres, L.; Khan, M. A.; Richter-Addo,
G. B. Can. J. Chem. 2003, 81, 872-881. (k) Fernandez, B. O.; Ford,
P. C. J. Am. Chem. Soc. 2003, 125, 10510-10511. (l) Patterson, J C.;
Lorkovic, I. M.; Ford, P. C. Inorg. Chem. 2003, 42, 4902-4908. (m)
Fernandez, B. O.; Lorkovic, I. M.; Ford, P. C. Inorg. Chem. 2004, 43,
5393-5402.
2) (a) Traylor, T. G.; Sharma, V. S. Biochemistry 1992, 31, 2847 (b) Radi, R.
Chem. Res. Toxicol. 1996, 9, 828. (c) Ford, P. C.; Laverman, L. E.;
Lorkovic, I. M. AdV. Inorg. Chem. 2003, 54, 203-257.
3) For examples: (a) Davies, K. M.; Wink, D. A.; Saavedra, J. E.; Keefer, L.
K. J. Am. Chem. Soc 2001, 123, 5473-5481. (b) Feelisch, M., Stamler, J.
S., Eds.; Methods in Nitric Oxide Research; J. Wiley & Sons: West Sussex,
England, 1996; Chapter 7, pp 71-115. (c) Bourassa, J.; DeGraff, W.; Kudo,
S.; Wink, D. A.; Mitchell, J. B.; Ford, P. C. J. Am. Chem. Soc. 1997, 119,
(5) For example: (a) Moore, E. G.; Gibson, Q. H. J. Biol. Chem. 1976, 251,
2788. (b) Rose, E. J.; Hoffman, B. M. J. Am. Chem. Soc. 1983, 105, 2866.
(c) Hoshino, M.; Laverman, L.; Ford, P. C. Coord. Chem. ReV. 1999, 187,
75-102 and references therein. (d) Cheng, L.; Richter-Addo, G. B. In The
Porphyrin Handbook, Kadish, K. M., Smith, K. M., Guilard, R., Eds.;
Academic Press: San Diego, CA, 2000; Vol. 4, Chapter 33. (e) Wyllie, G.
R. A.; Scheidt, W. R. Chem. ReV. 2002, 102, 1067-1090. (f) Franke, A.;
Stochel, G.; Jung, C.; van Eldik, R. J. Am. Chem. Soc. 2004, 126, 4181-
4191.
2
853-2862. (d) Namiki, S.; Arai, T.; Fujimori, K. J. Am. Chem, Soc. 1997,
1
19, 3840-3841. (e) Megson, I. L.; Webb, D. J. Expert Opinion on
InVestigational Drugs 2002, 11, 587-601. (f) Wang, P. G.; Xian, M.; Tang,
X.; Wu, X.; Wen, Z.; Cai, T.; Janczuk, A. J. Chem. ReV. 2002, 102, 1091-
1
134.
10.1021/ja044090+ CCC: $30.25 © 2005 American Chemical Society
J. AM. CHEM. SOC. 2005, 127, 6737-6743
9
6737

A R T I C L E S
Joseph and Ford
further purification. Nitric oxide (99%, Aire Liquide) was purified by
passage through a stainless steel column containing Ascarite II (Thomas
Scientific), attached via O-ring seal (Viton) to a greaseless vacuum
line.
The free base tris(aryl)corroles 5,10,15-tris(4-nitro-phenyl)corrole-
(H
3
3
TNPC), 5,10,15-tris(phenyl)corrole (H TPC), and 5,10,15-tris(4-
tolyl)corrole (H TTC) were prepared via the reaction of freshly distilled
3
pyrrole (Aldrich) with the appropriate substituted benzaldehyde in
refluxing acetic acid according to published procedures for the synthesis
of corroles. These compounds were purified by chromatography of
Figure 1. A. Free base corrole. B. Fe(Ar3C)(NO) (Ar ) phenyl,
4
-nitrophenyl, or 4-methylphenyl).
9
the crude reaction mixtures on an initial silica column using CH
as the elutant and then a second silica column using CH Cl :hexanes
10:1) as the elutant. The ligands were characterized and shown to be
2 2
Cl
Corroles differ from the porphyrin analogues by the absence
2
2
of a meso methine carbon (Figure 1A), a skeletal perturbation
that has a significant impact on the ligand properties of these
(
1
pure by HNMR and mass spectroscopic techniques. The iron com-
plexes Fe(Ar C)(NO) were formed via procedures reported previously
for the preparation of other Fe(Ar
FeCl ‚4H O (300 mg, 1.5 mmol) and the free base corrole (50 mg,
6
macrocycles. The most obvious difference is that a corrole has
3
three ionizable N-H protons; thus, when coordinated to metal
7a
3
C)(NO) complexes. A mixture of
3-
centers, the corrolate ligand is tri-anionic (C ), while the
2
2
2-
analogous porphyrinato ligand is dianionic (P ). The higher
ionic charge and the smaller cavity of the corrolate macro-
cycle stabilizes metals in higher oxidation states; thus, stable
∼76 µmol) in pyridine:methanol (1:2) was refluxed for 3 h under
dinitrogen, after which a 1.0 mL aliquot of saturated aqueous solution
2
of NaNO was added to the hot solution, and the system was refluxed
7
for an additional 30 min. The solution was then cooled to 0 °C, and
the resulting solid was collected by filtration and washed with water.
This solid product was then dissolved in dichloromethane and then
purified by chromatography on an alumina column (Brockman Activity
tri- and tetravalent metal complexes are common. These
8
“
contracted porphyrins” are receiving renewed interest owing
9
to more accessible synthetic procedures; however, reactivity
studies remain limited. The present study probes the kinetics
of the formation of ferric nitrosyl complexes of the type
Fe(Ar3C)(NO) (Figure 1B) from Fe(Ar3C) and NO in several
different solvent systems.
III) using CH
to formation of Fe(Ar
after recrystallization from CH
ized and shown to be pure by IR, H NMR, and mass spectrometry
techniques.11 The detailed syntheses of these compounds will be
reported elsewhere.
2
Cl
2
:hexane (1:3) as the elutant. The procedure leading
C)(NO) complexes typically gave yields of ∼80%
Cl :hexane (1:2). These were character-
3
2
2
1
Described here is a flash photolysis investigation of a repre-
sentative ferric corrolate complex of nitric oxide, Fe(TNPC)-
NO) (I) (Figure 1B, TNPC3- is the 5,10,15-tris(4-nitro-phenyl)-
(
Flash Photolysis Experiments. Solutions of known concentrations
of NO and of Fe(Ar C)(NO) were prepared by vacuum transfer
3
techniques on all-glass vacuum lines using flasks of known volume
and procedures designed to prevent contamination of the solution with
x
other NO impurities. The “pump-probe” flash photolysis apparatus has
been described previously.4d The excitation source (“pump”) was a
Continuum NY-61 Nd:YAG pulse laser operating in either the
frequency tripled (355 nm) or frequency doubled mode (532 nm) with
an energy of ∼10 mJ/pulse. Transient absorption changes were
monitored with monochromatic probe light beam focused into the
sample (at a right angle to the excitation beam) through a double grating
monochromator (SPEX model 1680) and onto a photomultiplier tube
corrolate ligand), in various solutions containing excess NO (in
most cases). Photolysis of I leads to NO labilization and the
formation of Fe(TNPC) (II), which subsequently undergoes
reaction with NO to regenerate I as the system relaxes back to
equilibrium. The dynamics of this process were investigated in
various media using time-resolved optical spectroscopy. A key
observation is that the back reaction of NO with II (eq 1) is
many orders of magnitude faster than the analogous reaction
III
with ferric porphyrinato analogues Fe (P) with NO. These
results and their possible interpretations are presented here.
(RCA IP28). The temporal response (25 shot averages) was recorded
using a digital oscilloscope (Tektronix TDS 540) linked to a desktop
computer. Plots of intensity versus time were converted to absorbance,
and these curves were fit using Igor Pro Carbon software (Wavemetrics).
UV-vis spectra were recorded using a Hewlett-Packard model
8452A diode array spectrophotometer. The solutions were prepared and
Experimental Section
contained in a custom-made glass flask of measured volume fused to
a quartz spectrophotometer cuvette and equipped with a coldfinger and
with a high-vacuum stopcock for connection to a vacuum line.
Materials. Solvents were distilled under nitrogen. Spectrochemical
grade toluene (Burdick & Jackson) was distilled from sodium, tetra-
hydrofuran (THF) from benzophenone and sodium, and acetonitrile
from calcium hydride. Pyridine was stored over molecular sieves. Other
organic compounds (Aldrich) were used as purchased without any
Results and Discussion
The optical absorption spectrum of Fe(TNPC)(NO) (I) in
toluene solution displays a broad band at λmax 392 nm (log ꢀ )
(
6) (a) Vogel, E.; Will, S.; Tilling, A. S.; Neumann, L.; Lex, J.; Bill, E.;
Trautwein, A. X.; Wieghardt, K. Angew. Chem., Int. Ed. Engl. 1994, 33,
(9) (a) Paolesse, R.; Nardis, S.; Sagone, F.; Khoury, R. G. J. Org. Chem. 2001,
66, 550-556. (b) Gross, Z.; Galili, N.; Simkhovich, L.; Saltsman, I.;
Botoshansky, M.; Bl a¨ ser, D.; Boese, R.; Goldberg, I. Org. Lett. 1999, 1,
599-602. (c) Mahammed, A.; Gray, H. B.; Meier-Callahan, A. E.; Gross,
Z. J. Am. Chem. Soc. 2003, 125, 1162-1163.
7
31-735. (b) Autret, M.; Will, S.; Caemelbecke, E. V.; Lex, J.; Gissel-
brecht, J.-P.; Gross, M.; Vogel, E.; Kadish, K. M. J. Am. Chem. Soc. 1994,
16, 9141-9149.
1
(7) (a) Simkhovich, L.; Goldberg, I.; Gross, Z. Inorg. Chem. 2002, 41, 5433-
5
439. (b) Steene, E.; Wondimagegn, T.; Ghosh, A. J. Phys. Chem. B. 2001,
(10) Gouterman, M. In The Porphyrins; Dolphin, D., Ed.; Academic Press: New
York, 1978; Vol. III, Chapter 1.
1
05, 11406-11413. (c) Meier-Callahan, A. E.; Di Bilio, A. J.; Simkhovich,
L.; Mohammed, A.; Goldberg, I.; Gray, H. B.; Gross, Z. Inorg. Chem.
(11) For example the spectroscopic properties for Fe(TNPC)(NO) are: 1H NMR
2
001, 40, 6788-6793. (d) Cai, S.; Licoccia, S.; Walker, F. A. Inorg. Chem.
3
(CDCl ): δ 7.37 (d, 2H, J ) 5 Hz), 7.55 (d, 2H, 5 Hz), 7.76 (d, 2H, 4 Hz),
2
001, 40, 5795-5798. (e) Ramdhanie, B.; Zakharov, L. N.; Rheingold, A.
7.85 (dd, 1H, 8 Hz), 7.90 (dd, 1H, 8 Hz), 8.04 (d, 4H, 8 Hz), 8.11 (d, 2H,
+
R.; Goldberg, D. P. Inorg. Chem. 2002, 41, 4105-4107.
8) Sessler, J. L.; Weghorn, S. J. In Expanded, Contracted & Isomeric
Porphyrins; Baldwin, J. E., Ed.; Tetrahedron Organic Chemistry Series,
Vol. 18; Pergamon: New York, 1997; pp 11-120.
4 Hz), 8.45 (dd, 2H, 8 Hz), 8.51 (d, 4H, 8 Hz). MS(FAB): m/z 744 (M );
714 (M+ - NO). FTIR: νno ) 1778 cm (KBr), 1785 cm (CHCl
).
3
4 -1 -1
-1
-1
(
UV-vis (toluene): λmax 392 nm (ꢀ ) 6.0 × 10 M cm ); 538 nm (7.8
3
-1
-1
× 10 M cm ).
6738 J. AM. CHEM. SOC.
9
VOL. 127, NO. 18, 2005

Axial Lability of Iron(III) Corrole Complexes
A R T I C L E S
been unsuccessful in preparing analytically pure samples of the
NO-free Fe(C) complexes owing to the ready oxidation of
these species. To address this question, a species believed to
be Fe(TNPC) was prepared in situ as follows in a glass apparatus
equipped with vacuum line adaptors and fused to a 1-cm path
length quartz spectrophotometer cell.
-
5
Initially, a solution of I (1.8 × 10 M) was prepared in
toluene (5.0 mL) at a concentration where the visible range
absorption bands are appropriately on-scale and the spectrum
was recorded (Supporting Information Figure S-1a). We had
noted previously that solutions of I in pyridine slowly undergo
loss of the coordinated NO when exposed to room light. To
exploit this, the toluene was removed from the above solution
under vacuum carefully to avoid the loss of any of the resulting
solid I, which was then redissolved in pyridine (5.0 mL). The
resulting solution was then exposed to room light and the
spectrum periodically recorded until no further change was
noted, and then the spectrum was recorded again (Figure S-1b).
The liquid pyridine was then removed by evacuating the flask
to give a dry, solid residue. The latter was redissolved in toluene
(5.0 mL) to give the same concentration as that of the original
solution, and the optical spectrum was recorded (Figure S-1b).
The difference between the two is given in Figure 2b. To
confirm that there was no loss of the Fe(TNPC) from the system,
the resulting toluene solution was then exposed to NO (∼200
Torr) and the optical spectrum recorded again. This spectrum
Figure 2. (a) Transient difference spectrum recorded using a CCD detector
at a time delay of 30 ns from the flash and a 50 ns gate following 532 nm
flash photolysis of a toluene solution of Fe(TNPC)(NO). (b) Difference
spectra between I and Fe(TNPC) (generated in situ, see text) both at 1.8 ×
(Figure S-1a) is identical to the one recorded initially; thus, there
was no significant loss of material resulting from the sequential
manipulations.
-
5
1
0
M in toluene.
Notably, the difference spectrum shown in Figure 2b is quali-
tatively very similar to that in Figure 2a, the major perturbation
being the much larger negative peak corresponding to the Soret
type band of I at 392 nm. However, the latter difference can be
attributed to the decreased sensitivity of the CCD camera at
the shorter wavelengths since the spectrograph grating is blazed
at 500 nm. Thus, a linear correspondence between the intensities
of the positive and negative peaks in the difference spectra
would not be expected. However, when temporal spectral studies
were carried out using the point-by-point mode (decay curves
measured at 5 nm intervals), the transient bleaching at 392 nm
matched that expected according to Figure 2b)
4
.78) and another band centered at 538 nm (3.89). The spectra
of Fe(TPC)(NO) (III) (strong bands at 390 nm (4.85) and 534
nm (3.97)) and Fe(TTC)(NO) (IV) (390 nm (4.86) and 532 nm
(3.97)) are quite similar. These bands can be assigned to ππ*
transitions largely localized on the corrole macrocycle and are
analogous to the Soret and Q-bands seen in the optical spectra
of porphyrinato complexes.10 In the infrared spectra, the NO
stretching frequencies (νNO) of I, III, and IV are 1778, 1766,
and 1762 cm , respectively, in KBr with the trend in accord
with the expected electron donor properties of the ring substit-
uents (-NO2 < -H < -CH3).
-
1
Flash photolysis of toluene solutions of each of these
Fe(C)(NO) complexes using the excitation wavelengths (λex)
55 or 532 nm led to transient bleaching of bands corresponding
to the parent compounds and the appearance of strong transient
absorbances. These transient spectral changes underwent rapid
Figure 3 illustrates the temporal decay of the transient
absorbance at 470 nm following 355 nm excitation of a
deaerated toluene solution of Fe(TNPC)(NO). These ∆Abs vs
time data fit well to a second-order process consistent with NO
photolabilization from I followed by the back reaction of II
with NO to reform I indicated by eq 1. Similar transient decays
were seen following 532 nm excitation. When the reaction was
carried out under various NO partial pressures PNO, the decay
traces were faster and could be fit as simple exponentials to
give the pseudo-first-order rate constants kobs. Plots of kobs vs
[NO] (Figure 4) over the concentration range 0.5-3.0 mM
proved to be linear with y-intercepts near zero according to the
expected rate law expressed in eq 2. The slope of this line is
therefore the second-order rate constant of recombination, kNO,
3
[
NO]-dependent decay back to the original spectra as described
in more detail below. Figure 2a displays the difference spectrum
recorded with a CCD detector and displaying transient absor-
bances at 460 and 593 nm following 532 nm flash photolysis
of I in toluene solution.
Experience with similar behavior of the Fe(P)(NO) ana-
logues4
a,h,12
as well as the observation of [NO]-dependent decay
rates suggests that the flash photolysis experiment induces
reversible NO labilization as indicated in eq 1. This could be
confirmed by comparing the spectral differences between the
nitrosyl reactant Fe(C)(NO) and the presumed Fe(C) product
to the difference spectrum recorded; however, we have as yet
8
-1 -1
and the value in toluene solution is 9.0 (( 0.5) × 10 M
s
at 295 K.
d[II]
( dt ) )
-kNO[NO][II]
(2)
(
12) Lim, M. D.; Lorkovic, I. M.; Wedeking, K.; Zanella, A. W.; Works, C. F.;
Massick, S. M.; Ford, P. C. J. Am. Chem. Soc. 2002, 124, 9737-9743.
J. AM. CHEM. SOC.
9
VOL. 127, NO. 18, 2005 6739

A R T I C L E S
Joseph and Ford
argon or by subjecting to a vacuum. Thus, it is obvious that the
rate constants of spontaneous NO dissociation (koff) are very
small and that the equilibrium constant for eq 3 is quite large.
Again, this behavior is much more analogous to that seen for
3
h
ferro-hemes than for the ferri-heme analogues.
kNO
III
Fe (C) + NO y z Fe(C)(NO)
(3)
koff
The marked reactivity of II toward NO is further demonstrated
by probing the back reaction after 532 nm flash photolysis of
I in 295 K tetrahydrofuran over a range of [NO]. Our
expectation was that in a stronger donor solvent, the reaction
of II with NO would be slowed owing to the occupation of the
III
axial sites of Fe (TNPC) by the THF. For comparison, the
activation parameters measured for the reaction of NO with the
III
ferric porphyrinato complex Fe (TPPS) (TPPS ) tetra(4-
sulfonato-phenyl)porphyrinato) in aqueous solution clearly
demonstrated that formation rate of the metal nitrosyl complex
was dominated by the lability of the axial ligands (in that case,
Figure 3. Transient absorbance decay at 470 nm following 355 nm flash
photolysis of I in toluene. The illustrated fit is to a second-order process.
3
c,h
waters).
determined for the NO reaction with II in THF was 10 ((1) ×
0 M s , a value essentially identical to that determined in
However, this expectation was not met; the kNO
8
-1 -1
1
toluene solution.
In the same context, flash photolysis studies of I were also
carried out in the mixed solvents 9:1 (v:v) toluene:acetonitrile
and 9:1 (v:v) toluene:pyridine at 295 K over a range of PNO. In
the former case, the effect of the added acetonitrile on the rate
of NO recombination with II was relatively small with kNO )
8
-1 -1
6
(
.0 (( 0.6) × 10 M s , about 33% smaller than in toluene.
This value may be an underestimate, given that it was
determined with the assumption that NO solubility in the mixed
solvent is the same as that in toluene. Since NO solubility in
neat CH3CN is about 50% that in toluene, NO solubility in the
mixed solvent might be somewhat lower than that in neat
toluene.) On the other hand, the effect of 10% added pyridine
was more significant. In that case, the kNO is a factor of 6 smaller
Figure 4. Plots of kobs values (under excess NO) vs [NO] in pyridine and
in 9:1 (v:v) toluene:pyridine following 532 nm excitation.
Analogous flash experiments were also carried out with the
corrolate complexes Fe(TTC)(NO) (III) and Fe(TPC)(NO) (IV)
in 295 K toluene solutions with a range of concentrations of
8
-1 -1
(
1.4 (( 0.1) × 10 M s ) than that found in neat toluene
under these conditions (Figure 4). Thus, of these various donor
solvents, only pyridine seems to have a significant effect,
although this value of kNO remains dramatically larger than that
found with the ferri-heme analogue.
-
3
added NO up to 3 × 10 M. In each case, transient absorption
generated by the 532 nm flash photolysis decayed back to
baseline exponentially, and plots of the resulting kobs vs [NO]
were linear. The kNO values determined from these plots for
An interesting comparison can be drawn to the solvent effects
on the kNO values determined in a kinetics study of the NO
reaction with the Ru(III) complexes Ru(salen)(Cl)(Sol) (salen
8
-1
the flash photolyses of III and IV were 5.9 ((1.3) × 10 M
-
1
8
-1 -1
s
and 5.8 (( 0.9) × 10 M s , respectively, similar to the
)
N,N′-bis(salicylidine)ethylenediamine dianion) generated by
kNO determined in the flash photolysis of I under similar
conditions.
flash photolysis of Ru(salen)(Cl)(NO) in different solvents (Sol)
1
4
at ambient temperature. In that study, the respective rates in
THF and in CH3CN were found to be ∼8 orders of magnitude
smaller and ∼11 orders of magnitude smaller than those found
These values for kNO seem surprisingly large. They are several
5
-1 -1
orders of magnitude larger than that (4.2 × 10 M s )
recorded for an analogous Fe(III) porphyrinate species,
Fe(TPP)(NO2), formed by flash photolabilization of NO from
Fe(TPP)(NO2)(NO) (TPP ) tetraphenylporphyrinato) in am-
bient temperature toluene solution.11 Indeed, the much higher
kNO for NO reaction with II, Fe(TTC), and Fe(TPC) are much
closer to those recorded for the ferrous porphyrinato complex
-4
7
-1 -1
in toluene (kNO ) 0.22, 4.7 × 10 , and 3.7 × 10 M
s
,
respectively). Although, like Fe(III), the Ru(III) complexes have
5
d
electronic configurations, a key difference is that the
III
electronic configuration of the Ru (salen)(Cl)(Sol) intermediate
is low spin and the complex is hexacoordinate; thus, the
reactivity with NO is dominated by the lability of the coordi-
nated solvent.
9
-1 -1
13
Fe(TPP) (kNO ) 5.2 × 10 M
s
in benzene). Furthermore,
the nitrosyl complexes Fe(TNPC)(NO), Fe(TTC)(NO), and
Fe(TPC)(NO) are all moderately stable in the solid state and in
aerated solution, and there are no observable changes in their
optical spectra when solutions are deaerated by entraining with
(13) Hoshino, M.; Kogure, M. J. Phys. Chem. 1989, 93, 5478-5484.
(
14) (a) Works, C. F.; Ford, P. C. J. Am. Chem. Soc. 2000, 122, 7592-7593.
(
b) Works, C. F.; Jocher, C. J.; Bart, G. D.; Bu, X.; Ford, P. C. Inorg.
Chem. 2002, 41, 3728.
6740 J. AM. CHEM. SOC.
9
VOL. 127, NO. 18, 2005

Axial Lability of Iron(III) Corrole Complexes
A R T I C L E S
Scheme 1
Figure 5. Plot of kobs-1 vs log [py] for the relaxation rates following 532
nm excitation of I in 295 K toluene:pyridine solutions under a constant
PNO of 110 Torr. The squares are the experimental points, while the circles
are simulated data based on the model described by Scheme 1 with the
and IR spectra. The apparently low value of K0 indicates that
the metal center in the Fe(III) corrolate complexes Fe(C)(NO)
is much less electrophilic than in analogous Fe(III) porphyrinate
complexes as evidenced by the typically hexacoordinate nature
of Fe(P)(L)(NO) compounds. This difference is also consistent
4
8
following values K0 ) 0, K1 ) 1 × 10 M, K2 ) 0.2 M, kNO(1) ) 9 × 10
-
1
-1
8
-1 -1
M
s
and kNO(2) ) 1.7 × 10 M
s .
-1
with the measurably lower frequency of νNO (1778 cm in KBr)
for I relative to that of a porphyrinato compound such as
To explore further the effect of pyridine on NO recombina-
tion, flash photolysis experiments were conducted with toluene
solutions of Fe(TNPC)(NO) under various pyridine concentra-
-
1
15
-1
Fe(TPP)(NO2)(NO) (1872 cm in benzene, 1868 cm in
16
-
5
KBr ). The reluctance for I to coordinate an axial ligand would
also be suggested by the structure of the tris(pentafluorophenyl)-
corrolate complex Fe(TFPC)(NO) reported by Gross and co-
tions ranging from 8 × 10 M to 2.5 M while maintaining a
fixed PNO at 110 Torr. (This would give a constant [NO] of
1
.77 ( 0.01 mM, if the nitric oxide solubility in the toluene/
7
a
workers, which shows the metal ion deviating from the N4
plane of the corrolate ligand by 0.45 Å toward the nitrosyl.
On the other hand, the dynamics of the reactive intermediate
II are indeed affected by the added pyridine in a manner that
suggests the formation of both mono- and bis(pyridine) adducts
of this species. Logical equilibria and NO addition reactions
are mapped in Scheme 1 to provide the framework for
discussion. In this scenario, the equilibrium constant K0 for the
reaction of I with pyridine is considered to be quite small as
indicated by the failure to observe significant changes in either
the optical or IR spectra of I when toluene was replaced by
pyridine as the solvent. Flash photolysis of I in the toluene/
pyridine mixed solvents is proposed to lead to NO labilization
to form Fe(TNPC). Once II is formed via NO labilization from
I, reaction with py would also give Fe(TNPC)(py) and presum-
ably Fe(TNPC)(py)2 in concentrations depending on the py
concentration. If it is assumed that these steps are fast relative
to relaxation to the nitrosyl species (however, see below) and
that there is no direct reaction of NO with Fe(TNPC)(py)2, kobs
is predicted by eq 4, where kNO(1) is the rate constant for the
reaction of NO with II and kNO(2) is that for the reaction of
NO with Fe(TNPC)(py). Notably the value of kNO(1) (9.0 ×
pyridine mixture remains the same as that in neat toluene). The
decay curves of the transient absorbances detected after 532
nm excitation appeared to fit exponential functions (but see
below) and give kobs(ap) values (the parenthetical “ap” is
introduced to indicate these to be “apparent” kobs values) that
decreased as the pyridine concentration was increased. For
6
-1
example, under these conditions kobs(ap) ) 1.59 × 10 s at
5
-1
-4
[
×
py] ) 0 M, 3.27 × 10 s at [py] ) 8.0 × 10 M and 2.42
5
-1
10 s at [py] ) 1.1 M. However, the behavior is not a
simple function of [py]. This is illustrated by Figure 5, which
is a plot of kobs(ap)-1 vs log[py]. In general terms, the kobs(ap)
-
4
are little affected up to ∼2 × 10 M [py], then decrease
-
3
somewhat sharply up to ∼10 M [py] with an overall change
of about a factor of 5. The kobs(ap) values then essentially plateau
5
-1
-3
-1
at ∼3 × 10 s over the [py] range 10 to 10 M before
dropping again with further increases in [py].
Notably, there is virtually no change in the optical spectrum
of I under these conditions (<5% reduction in the intensity of
the band at 392 nm) between neat toluene and 9:1 (v:v) toluene:
pyridine). Indeed, the spectrum of I in neat pyridine is virtually
the same as that in neat toluene if measured under an NO
atmosphere (see Supporting Information Figure S-2). This was
demonstrated by an experiment where a solution of I (1.8 ×
8
-1 -1
1
0 M
s
at 295 K) is already known from the experiments
in toluene solution under various NO pressures (Figure 4).
-
5
1
0
M) was prepared in toluene and the spectrum recorded;
then the solvent was removed under vacuum and the solid
redissolved in the same volume of pyridine. Under argon, there
were moderate changes in the optical spectrum as the solutions
were being prepared and handled in room light, but once NO
(
kNO(1) + kNO(2)K [py])[NO]
1
kobs
)
(4)
2
1
+ K [py] + K K [py]
1 1 2
On the basis of this model, inspection of the experimental data
presented in Figure 5 suggested that K1 should fall into the range
10 -10 M and that the product kNO(2)[NO] should have a
(200 Torr) was introduced to the closed vessel, the optical
spectrum attributed to I was fully restored. Furthermore, in the
infrared spectrum there is but a very modest shift of the
frequency of the NO stretching band (νNO) for I between
3
4
∼
5
-1
value of about 3.0 × 10 s (the plateau) under these conditions.
8
-
1
The latter conclusion would give a kNO(2) value of 1.7 × 10
solutions in neat chloroform (1785 cm ) and in neat pyri-
-
1 -1
-
1
M
s
for the reaction of NO with Fe(TNPC)(py), about 5-fold
dine (1778 cm , broader). Thus, coordination of pyridine to
Fe(TNPC)(NO) either has a very small equilibrium constant K0
or is not manifested by significant changes in the electronic
(
15) Settin, M. F.; Fanning, J. C. Inorg. Chem. 1988, 27, 1431-1435.
(16) Yoshimura, T. Inorg. Chem. Acta 1984, 83, 12-21.
J. AM. CHEM. SOC.
9
VOL. 127, NO. 18, 2005 6741

A R T I C L E S
Joseph and Ford
smaller than kNO(1) for the analogous reaction of II. Since the
second stage of reaction inhibition does not come into play until
[py] is nearly 1 M, the association constant K2 for formation of
Fe(TNPC)(py)2 must be 0.5 M or smaller. Attempts to fit the
kobs(ap) values to this model give the curve displayed in Figure
5
1
based on values of kNO(1) and kNO(2) noted above plus K1 )
.0 × 10 M and K2 ) 0.2 M. Notably, this fit is quite good at
4
the higher pyridine concentrations but marginal (at best) for
low [py]. Nonetheless, the concept that, coordination of I to
one pyridine is facile but coordination to two pyridines is much
less so, is consistent with earlier work by Vogel et al. who
reported the isolation and structure of the monopyridine iron-
6
a
III
Figure 6. Proposed splitting of d-orbitals in Fe (C).
(III) complex Fe(OEC)(py) (OEC ) octaethylcorrolate) under
conditions that would have yielded bis(pyridine) adducts for
porphyrin analogues.17 It should be noted, however, that the
bis(pyridine) ferric corrolate Fe(TDCC)(py)2 (TDCC ) tris-
less reactive toward NO than is II, but the latter apparently does
not react with NO on the time scale of these experiments. For
comparison, the iron(II) porphyrinato analogue Fe(TPP) binds
to pyridine to form Fe(TPP)(py) and Fe(TPP)(py)2 with respec-
(
2,6-dichlorophenyl)corrolate) has been isolated from pyridine/
7
a
3
4
-1
hexane/benzene solutions and structurally characterized.
tive K1 and K2 values of 1.1 × 10 and 3.3 × 10 M , and
13
The fundamental problem in attempting to analyze the
reaction dynamics of the relaxation processes after the flash
photolysis of I in mixed pyridine/toluene solutions in terms of
this model is the stated assumption that the equilibria between
II and the mono- and bis(pyridyl) adducts are established rapidly
relative to the decay of the transients by reaction with NO. Since
it is likely that the reaction of py with II will have a rate constant
no greater than kNO(1), the relaxation of the systems will involve
several competing and sequential processes on comparable time
scales at the lowest concentrations of pyridine. At higher [py],
the original assumption will have greater validity. For example,
the rate constant for the relaxation of II and the monopyridine
adduct to come into equilibrium will be krel ) k2[py] + k-2. If
Fe(TPP)(py)2 is not reactive with NO. With the much larger
K2 value, the rate of NO recombination with Fe(TPP) in 10%
pyridine would be reduced by more than 4 orders of magnitude
(relative to pure toluene). The much smaller effect seen here
for II demonstrates quantitatively that ligands in the sixth
coordination site of ferric corrolates are but weakly bound, an
attractive feature in catalyst design. Such an impact on axial
ligand binding has been noted by Mohammed et al.9c in a
proposed mechanism for aerobic oxidations catalyzed by
chromium(III) corrolates Cr(Ar3C), where the first step was
suggested to be a fast preequilibrium between hexa- and
pentacoordinate Cr(III), allowing the latter to react with O2.
Thus, the present flash photolysis kinetics studies demonstrate
9
-1 -1
3-
k2 ) ∼10 M s , i.e., comparable to kNO(1), this system will
approach equilibrium significantly faster than the reaction of
II with NO only when [py] > [NO], which is fixed in these
experiments at 1.77 mM.
that the exceptional electron-donor nature of the TNPC ligand
suppresses the formation of strongly bound hexacoordinate
1
complexes. Although H NMR studies clearly show I to be
II
+
IV
-
diamagnetic (whether best represented as Fe (NO ), Fe (NO ),
III
In this context, a closer examination of the apparent single-
exponential fits of the transient decays following flash photolysis
of I revealed the following. The single-exponential fits are
or ferromagnetically coupled Fe (NO) will be left for future
debate), Fe(III) corrolato complexes analogous to II have been
3
6a,18
described as having a /2 spin state.
While there is undoubt-
-4
excellent at very low [py] (<10 M), where decay is dominated
edly electronic delocalization and covalency between the
macrocycle σ- and π- and the metal d-orbitals, meaningful
insight can be drawn from a simple d-orbital diagram of a
tetragonal complex with a very strong equatorial ligand field
and a weak axial field (Figure 6). According to this diagram,
-
2
simply by the reaction of II with NO, and at high [py] (>10
M), where the equilibration between II and the pyridine
complexes is faster than regeneration of I. Better fits of the
decay curves at intermediate pyridine concentrations were
-
kat
3
2
obtained when a dual exponential function (Abs(t) ) Aa e
the /2 spin state would give a half-occupied, σ-antibonding dz
-
kbt
4
+
Ab e
+ B) was used with ka fixed at a value equal to
orbital. Electronic configurations such as high-spin d or low-
spin d metal complexes and the ligand field excited states of
d metal complexes with a single electron in a dz orbital
7
kNO(1)[NO] and with Aa, Ab, and kb allowed to iterate to find
the best fits. A key observation is that the weighting factor Ab
for the second term increases significantly with increasing
pyridine concentration while the rate constant diminishes. For
6
2
invariably are extremely labile to axial ligand substitution, owing
1
9
to the resulting strong Jahn-Teller distortion. Given that
-
4
20
6
examples, for [py] ) 2.3 × 10 M, the ratio Ab/Aa was ∼0.4,
Fe(TPP) has been reported to have a triplet spin state, this d
-4
2
but at 7.9 × 10 M, that ratio has shifted to ∼2.5. This behavior
is qualitatively consistent with the model described by eq 4,
with the proviso that the dynamics at low [py] reflect the
competition for I among the various reactive species present,
including NO and py.
Fe(II) species would also display a singly occupied dz or-
bital, while analogous ferric complexes such as Fe(TPP)(Cl)
2
1
have sextet, high-spin configurations. We attribute that the
(
18) Licoccia, S.; Paci, M.; Paolesse, R.; Boschi, T. J. Chem. Soc., Dalton Trans.
1991, 401.
Based on this analysis, II binds to py to form mono- and
(19) (a) Jordan, R. B. Reaction Mechanisms of Inorganic and Organometallic
Systems, 2nd ed., Oxford University Press: New York, 1998. (b) Bergkamp,
M. A.; Brannon, J.; Magde, D.; Watts, R. J.; Ford, P. C. J. Am. Chem.
Soc. 1979, 101, 4549-4554.
4
bis(pyridine) complexes with equilibrium constants ∼10 M and
0
.2 M, respectively. The mono(pyridine) adduct is only 5-fold
(
20) (a) Boyd, P. D. W.; Buckingham, D. A.; McMeeking, R. F.; Mitra, S. Inorg.
Chem. 1979, 18, 3585-3591. (b) Misspelter, J., Momenteau, M.; Lohste,
J. M. J. Chem. Phys. 1980, 72, 1003-1010.
(
17) Nesset, M. J. M.; Shokhirev, N. V.; Enemark, P. D.; Jacobson, S. E.; Walker,
F. A. Inorg. Chem. 1996, 35, 5188-5200;
(21) Reed, C. A.; Guiset, F. J. Am. Chem. Soc. 1996, 118, 3281-3282.
6742 J. AM. CHEM. SOC.
9
VOL. 127, NO. 18, 2005

Axial Lability of Iron(III) Corrole Complexes
A R T I C L E S
lower reactivity of the ferric porphyrins toward NO, relative to
coordinated in the axial positions not only in hydrocarbon
solvents such as toluene but also in stronger Lewis base solvents
such as THF and acetonitrile. Furthermore, unlike Fe(TPP),
II
III
III
Fe (P) and Fe (C), to the greater Lewis acidity of the Fe (P)
center, leading to stronger solvent coordination at the axial sites.
5
The low-spin 4d Ru(salen)(Cl)(sol) complexes formed after
which has been shown to form a bis(pyridine) complex Fe-
flash photolysis of the nitrosyl complex Ru(salen)(Cl)(NO) in
different solvents described above would be expected to bind
axial donor ligands even more strongly and thus would be much
less reactive in donor solvents than are any of the Fe(II) or
Fe(III) compounds discussed here.
13
(
TPP)(py)2, the reactivity of II toward NO is only modestly
suppressed by high concentrations of pyridine, since formation
of the hexacoordinate analogue Fe(TNPC)(py)2 has a much
smaller equilibrium constant.
Ongoing studies in this laboratory with the iron corrolate
complexes are focused on evaluating quantitatively how the
electronic and steric properties of representative corrolate ligands
define the reactivities of coordinated metal centers and how these
compare to the representative porphyrinato complexes.
Notably, the bis(pyridine) complex Fe(TFPC)(py)2 has been
reported7 to have a low-spin (S ) /2) electronic configuration,
consistent with the apparent low reactivity of Fe(TNPC)(py)2
toward NO as represented in Scheme 1.
a
1
In summary, we have used flash photolysis studies to generate
III
ferric corrolate complexes of the type Fe (Ar3C) and to
Acknowledgment. This research was supported by the
National Science Foundation (CHE-0352650).
probe for the first time the quantitative dynamics of such
species in ligand addition, in this case, reaction with nitric oxide.
The corrolate complex Fe(TNPC) (II) is much more reactive
toward NO than is an analogous ferric porphyrinato complex
Fe(TPP)(NO2) in toluene solution. Indeed II shows a kNO for
reactivity toward NO similar to that seen for the ferrous por-
phyrin analogue Fe(TPP) under comparable conditions. This
suggests that II is pentacoordinate or only very weakly
Supporting Information Available: Figures S-1 and S-2
showing spectra of Fe(TNPC)(NO) in different solvents before
and after photolysis. This material is available free of charge
via the Internet at http://pubs.acs.org.
JA044090+
J. AM. CHEM. SOC.
9
VOL. 127, NO. 18, 2005 6743