β-Octafluorocorroles

Article abstract of DOI:10.1021/ja021158h

The one-pot corrole synthesis first reported by the Gross and Paolesse groups appears to have evolved into a remarkably general and predictable self-assembly based synthetic reaction. Gross's solvent-free procedure (refs 8 and 9) has proven particularly effective in our hands and, in fact, more general than originally claimed. In earlier work (ref 17), we showed that the reaction works for a variety of aromatic aldehyde starting materials and was not limited to relatively electron-deficient aldehydes, as reported by Gross and co-workers. Here, we show that the pyrrole component is also variable in that 3,4-difluoropyrrole undergoes oxidative condensation with four different p-X-substituted benzaldehydes to yield the corresponding β -octafluoro-meso-tris(para-X-phenyl)corroles (X = CF₃, H, CH ₃, and OCH₃). Further, we have prepared the Cu and FeCl derivatives of the β-octafluorocorrole ligands. The XPS nitrogen 1s ionization potentials of these fluorinated ligands are some 0.7 eV higher than those of the corresponding β-unfluorinated ligands. The oxidation half-wave potentials of the Cu and FeCl complexes of the fluorinated corroles are also positively shifted by 300-400 mV relative to their β -unsubstituted analogues, demonstrating the strongly electron-deficient character of the fluorinated ligands. ¹H NMR spectroscopy suggests that like their β-unfluorinated counterparts, the new β -octafluorinated triarylcorroles act as substantially noninnocent ligands, i.e., exhibit corrole π-cation radical character, in the FeCl complexes. Quantitatively, however, NMR spectroscopy and DFT calculations indicate that the β-octafluorinated corroles are somewhat less noninnocent (i.e., carry less radical character) than their β-unfluorinated counterparts in the FeCl complexes. Temperature-dependent ¹⁹F NMR spectroscopy suggests that the Cu octafluorocorroles have a thermally accessible paramagnetic excited state, which we assign as a Cu(II) corrole π-cation radical. We have previously reported that the electronic absorption spectra, particularly the Soret absorption maxima, of high-valent transition metal triarylcorroles are very sensitive to the nature of the substituents in the meso positions. In contrast, the Soret absorption maxima of free-base triarylcorroles are not particularly sensitive to the nature of the meso substituents. This scenario also holds for the fluorinated corroles described here. Thus, although the four free-base fluorinated triarylcorroles exhibit practically identical Soret absorption maxima, the Soret bands of the Cu derivatives of the same corroles red-shift by approximately 35 nm on going from the p-CF₃ to the p-OCH₃ derivative.

Full text of DOI:10.1021/ja021158h

Published on Web 12/05/2003
â-Octafluorocorroles
Erik Steene, Abhishek Dey,† and Abhik Ghosh*
Contribution from the Department of Chemistry, UniVersity of Tromsø, N-9037 Tromsø, Norway
Received September 6, 2002; E-mail: abhik@chem.uit.no
Abstract: The one-pot corrole synthesis first reported by the Gross and Paolesse groups appears to have
evolved into a remarkably general and predictable self-assembly based synthetic reaction. Gross’s solvent-
free procedure (refs 8 and 9) has proven particularly effective in our hands and, in fact, more general than
originally claimed. In earlier work (ref 17), we showed that the reaction works for a variety of aromatic
aldehyde starting materials and was not limited to relatively electron-deficient aldehydes, as reported by
Gross and co-workers. Here, we show that the pyrrole component is also variable in that 3,4-difluoropyrrole
undergoes oxidative condensation with four different p-X-substituted benzaldehydes to yield the corre-
sponding â-octafluoro-meso-tris(para-X-phenyl)corroles (X ) CF3, H, CH3, and OCH3). Further, we have
prepared the Cu and FeCl derivatives of the â-octafluorocorrole ligands. The XPS nitrogen 1s ionization
potentials of these fluorinated ligands are some 0.7 eV higher than those of the corresponding â-unfluorinated
ligands. The oxidation half-wave potentials of the Cu and FeCl complexes of the fluorinated corroles are
also positively shifted by 300-400 mV relative to their â-unsubstituted analogues, demonstrating the strongly
electron-deficient character of the fluorinated ligands. 1H NMR spectroscopy suggests that like their
â-unfluorinated counterparts, the new â-octafluorinated triarylcorroles act as substantially noninnocent
ligands, i.e., exhibit corrole π-cation radical character, in the FeCl complexes. Quantitatively, however,
NMR spectroscopy and DFT calculations indicate that the â-octafluorinated corroles are somewhat less
noninnocent (i.e., carry less radical character) than their â-unfluorinated counterparts in the FeCl complexes.
Temperature-dependent 19F NMR spectroscopy suggests that the Cu octafluorocorroles have a thermally
accessible paramagnetic excited state, which we assign as a Cu(II) corrole π-cation radical. We have
previously reported that the electronic absorption spectra, particularly the Soret absorption maxima, of
high-valent transition metal triarylcorroles are very sensitive to the nature of the substituents in the meso
positions. In contrast, the Soret absorption maxima of free-base triarylcorroles are not particularly sensitive
to the nature of the meso substituents. This scenario also holds for the fluorinated corroles described here.
Thus, although the four free-base fluorinated triarylcorroles exhibit practically identical Soret absorption
maxima, the Soret bands of the Cu derivatives of the same corroles red-shift by approximately 35 nm on
going from the p-CF3 to the p-OCH3 derivative.
Introduction
macrocycles, the N-confused porphyrins and particularly the
corroles have exhibited an extensive coordination chemistry.2,5,14
Long known as the classic one-pot route to porphyrins,1 the
oxidative condensation of pyrrole and aromatic aldehydes2 has
also been found to yield N-confused porphyrins,3-6 sapphyrins,7
corroles,8-11 and various expanded porphyrins.12,13 Of these
In particular, the trianionic corrole ligands appear to stabilize a
wide variety of high-valent transition metal ions.15 In the late
1990s, the Gross and Paolesse groups reported one-pot syntheses
of meso-triarylcorroles.8,10 Gross reported a solvent-free pro-
cedure for the synthesis of corroles with electron-withdrawing
meso substituents such as meso-tris(pentafluorophenyl)corrole,
meso-tris(2,6-dichlorophenyl)corrole, and meso-tris(heptafluo-
robutyl)corrole.16 In our hands,17 Gross’s procedure turned out
to be more general than originally claimed and yielded meso-
triarylcorroles from a variety of aromatic aldehydes as starting
† Current address: Department of Chemistry, Stanford University,
Stanford, CA 94305.
(1) Lindsey, J. S.; Schreiman, I. C.; Hsu, H. C.; Kearney, P. C.; Marguerettaz,
A. M. J. Org. Chem. 1987, 52, 827.
(2) For a minireview on pyrrole-aldehyde condensations viewed as self-
assembly processes, see: Ghosh, A. Angew. Chem., Int. Ed., in press.
(3) Chmielewski, P. J.; Latos-Grazynski, L.; Rachlewicz, K.; Glowiak, T.
Angew. Chem., Int. Ed. Engl. 1994, 33, 779.
(4) Furuta, H.; Asano, T.; Ogawa, T. J. Am. Chem. Soc. 1994, 116, 767.
(5) Chmielewski, P. J.; Latos-Grazynski, L. Inorg. Chem. 1997, 36, 840.
(6) Geier, G. R., III; Haynes, D. M.; Lindsey, J. S. Org. Lett. 1999, 1, 1455.
(7) Chmielewski, P. J.; Latos-Grazynski, L.; Rachlewicz, K. Chem. Eur. J.
1995, 1, 68.
(8) Gross, Z.; Galili, N.; Saltsman, I. Angew. Chem., Int. Ed. 1999, 38, 1427.
(9) Gross, Z.; Galili, N.; Simkhovich, L.; Saltsman, I.; Botoshansky, M.; Bla¨ser,
D.; Boese, R.; Goldberg, I. Org. Lett. 1999, 599.
(10) Paolesse, R.; Mini, S.; Sagone, F.; Boschi, T.; Jaquinod, L.; Nurco, D. J.;
Smith, K. M. Chem. Commun. 1999, 14, 1307.
(11) Paolesse, R.; Nardis, S.; Sagone, F.; Khoury, R. G. J. Org. Chem. 2001,
66, 550.
(12) Shin, J.-Y.; Furuta, H.; Yoza, K.; Igarashi, S.; Osuka, A. J. Am. Chem.
Soc. 2001, 123, 7190.
(13) Shimizu, S.; Shin, J.-Y.; Furuta, H.; Ismael, R.; Osuka, A. Angew. Chem.,
Int. Ed. 2003, 42, 78.
(14) Nilsen, H. J.; Ghosh, A. Acta Chem. Scand. 1998, 52, 827.
(15) Erben, C.; Will, S.; Kadish, K. M. In The Porphyrin Handbook; Kadish,
K. M., Smith, K. M., Guilard, R., Eds.; Academic Press: San Diego, 2000;
Vol. 2, p 233.
(16) Simkhovich, L.; Goldberg, I.; Gross, Z. J. Inorg. Biochem. 2000, 80, 235.
(17) Wasbotten, I. H.; Wondimagegn, T.; Ghosh, A. J. Am. Chem. Soc. 2002,
124, 8104.
9
16300
J. AM. CHEM. SOC. 2003, 125, 16300-16309
10.1021/ja021158h CCC: $25.00 © 2003 American Chemical Society

â-Octafluorocorroles
A R T I C L E S
regarded as a strongly electron-deficient version of corrole,
serves as a more innocent ligand, compared with corrole. Thus,
the questionsin fact, the one that partially led to our undertaking
this projectsarises as to whether an electron-deficient corrole
ligand such as a â-octafluorocorrole should generally serve as
a noninnocent ligand like ordinary corroles or act as a relatively
innocent one like corrolazine.27 On the basis of NMR spectros-
copy, electrochemical measurements, and DFT calculations we
conclude that the â-octafluorocorroles reported here, like the
â-unsubstituted corroles studied previously,20-22,24-26 indeed act
as noninnocent ligands, i.e., exhibit corrole π-cation radical
character, in the FeCl complexes.
Figure 1. Free-base â-octafluoro-meso-tris(p-X-phenyl)corrole (X ) OCH3,
CH3, H, and CF3) ligands prepared in this work.
Experimental Section
materials. We wished to determine whether the pyrrole fragment
could also be varied in this reaction, and based on the fact that
3,4-difluoropyrrole has been successfully condensed with certain
aromatic aldehydes to yield â-octafluoro-meso-tetraarylporphy-
rins18 and, very recently, â-perfluorinated expanded porphy-
rins,13 we proceeded to investigate the reactivity of 3,4-
difluoropyrrole vis-a`-vis the one-pot solvent-free corrole synthesis.
This turned out to be a successful experiment, and we report
here the synthesis of a series of four â-octafluoro-tris(p-X-
phenyl)corroles (X ) CF3, H, CH3, and OCH3) (Figure 1).19
The one-pot corrole synthesis may be viewed as a seven-
component self-assembly process, involving three aldehydes and
four pyrrole units, and it is gratifying that it seems to be robust
with respect to variations in both the aldehyde and pyrrole
components.2
Corroles generally exhibit significantly lower oxidation
potentials than analogous porphyrin derivatives.15 For example,
the first half-wave potential for oxidation of Sn(OEC)Cl is 0.67
V compared to 1.36 V (vs SCE) for Sn(OEP)Cl2 (OEC )
â-octaethylcorrole and OEP ) â-octaethylporphyrin).15 An
important question considered in the recent literature concerns
how corroles, with such low oxidation half-wave potentials, can
stabilize high-valent transition metal ions. Using NMR spec-
troscopy, DFT calculations, as well as electrochemical argu-
ments, Walker and co-workers20-23 and we24-26 have shown
that the corrole ligand is noninnocent, i.e., exhibits corrole
π-cation radical character, in many high-valent transition metal
complexes. In contrast, recent DFT calculations by Tangen and
Ghosh27 suggest that the corrolazine28 ligand, which may be
General Comments. All samples for NMR analysis were prepared
in 5 mm NMR tubes using CDCl3 as solvent. A small drop of
trifluoroacetic acid was added to all free-base corrole NMR samples
to protonate the corrole and thereby to generate a symmetrical
macrocycle comparable to the metal complexes. Unless otherwise
mentioned, 1H and 19F NMR spectra were recorded with a Varian 400
MHz spectrometer at 298 K. The 1H NMR spectra were referenced to
residual CHCl3 (δ 7.24 ppm), whereas the 19F chemical shifts were
referenced to trifluoroacetic acid (δ -76.2 ppm) for the free-base
corroles and to 2,2,2-trifluoroethanol-d3 (δ -77.8 ppm) for the Cu and
FeCl corrole derivatives.
Cyclic voltammetry was carried out using an EG&G Model 263A
Potensiostat with a three-electrode system consisting of a glassy carbon
working electrode, a platinum wire counter electrode, and a saturated
calomel reference electrode (SCE). Tetra(n-butyl)ammonium perchlorate
(TBAP), recrystallized from ethanol and dried in a desiccator for at
least one week, was used as the supporting electrolyte. Dichloromethane,
distilled from calcium hydride and stored over molecular sieves, was
used as the solvent for the cyclic voltammetry experiments. The
reference electrode was separated from the bulk solution by a fritted-
glass bridge filled with the solvent/supporting electrolyte mixture and
all half-wave potentials reported here are referenced to the SCE. Pure
argon was bubbled though solutions containing the metallocorroles for
at least 2 min prior to the cyclic voltammetry experiments and the
solutions were also protected from air by an argon blanket during the
experiments.
X-ray photoelectron spectra were acquired at room temperature with
a Physical Electronics Quantum 2000 spectrometer, equipped with a
hemispherical analyzer, and 350 W of monochromatized Al KR
radiation. Sample preparation consisted of rubbing a tiny speck of the
corrole into a thin film on gold foil. The films, which appeared as a
colored sheen on gold, were so thin that they showed no evidence of
charging in the course of X-ray bombardment. Flooding the samples
with low-energy electrons did not change the positions or shapes of
the XPS peaks, again proving the absence of sample charging. The
binding energies reported here were externally referenced to the Au
4f7/2 binding energy (84.0 eV) of the gold substrate and were
reproducible to (0.1 eV. A full analysis of the XPS results will be
presented elsewhere; in this paper, we will only discuss the XPS of
the free-base fluorinated corroles.
(18) Woller, E. K.; DiMagno, S. G. J. Org. Chem. 1997, 62, 1588.
(19) We should note that with pentafluorobenzaldehyde and 3,4-difluoropyrrole
we did not succeed in isolating corrole products under the exact conditions
of Gross’s solvent-free procedure. Mass spectrometric analysis of the
products suggested the presence of a number of expanded porphyrins,
although we did not fully characterize these products in this work. As
mentioned in the text, Osuka and co-workers have obtained perfluorinated
expanded porphyrins from the condensation of pentafluorobenzaldehyde
and 3,4-difluoropyrrole under modified Lindsey conditions. However, while
this paper was under review, Chang and co-workers reported that increasing
the reaction time and temperature in Gross’s solvent-free procedure does
allow the successful isolation of perfluorinated meso-triphenylcorrole, with
pentafluorobenzaldehyde and 3,4-difluoropyrrole as starting materials: Liu,
H.-Y.; Lai, T.-S.; Yeung, L.-L.; Chang, C. K. Org. Lett. 2003, 5, 617-
620.
UV-visible spectra were recorded with an HP 8453 spectropho-
tometer in CH2Cl2. Column chromatography was performed on Matrex
35-70 µ silica from Millipore. Elemental Analyses were conducted
by Analytische Laboratorien (Prof. Dr. H. Malissa und G. Reuter
GmbH), Lindlar, Germany.
Synthesis of 3,4-Difluoropyrrole. 3,4-Difluoropyrrole was synthe-
sized using a slightly modified version of the procedure reported by
DiMagno and co-workers.29 The difference relative to the published
(20) Cai, S.; Walker, F. A.; Licoccia, S. Inorg. Chem. 2000, 39, 3466.
(21) Zakharieva, O.; Schuenemann, V.; Gerdan, M.; Licoccia, S.; Cai, S.; Walker,
F. A.; Trautwein, A. X. J. Am. Chem. Soc. 2002, 124, 6636.
(22) Cai, S.; Licoccia, S.; D’Ottavi, C.; Paolesse, R.; Nardis, S.; Bulach, V.;
Zimmer, B.; Shokhireva, T. K.; Walker, F. A. Inorg. Chim. Acta 2002,
339, 171.
(23) Walker, F. Inorg. Chem. 2003, 42, 4526.
(24) Steene, E.; Wondimagegn, T.; Ghosh, A. J. Phys. Chem. B 2001, 105,
11 406. Addition/Correction: J. Phys. Chem. B 2002, 106, 5312.
(25) Steene, E.; Wondimagegn, T.; Ghosh, A. J. Inorg. Biochem. 2002, 88, 113.
(26) Ghosh, A.; Steene, E. J. Inorg. Biochem. 2002, 91, 423.
(27) Tangen, E.; Ghosh, A. J. Am. Chem. Soc. 2002, 124, 8117.
(28) Ramdhanie, B.; Stern, C. L.; Goldberg, D. P. J. Am. Chem. Soc. 2001,
123, 9447.
9
J. AM. CHEM. SOC. VOL. 125, NO. 52, 2003 16301

A R T I C L E S
Steene et al.
procedure was that solvent was removed from the final dichloromethane
extract of 3,4-difluoropyrrole by blowing a gentle stream of dry N2
over the solution at room temperature for approximately 2 h. Rotary
evaporation was avoided to prevent loss of the volatile 3,4-difluoro-
pyrrole.
Cl2 and then 2% MeOH in CH2Cl2 as eluent. On the basis of NMR
analysis, the product form the column appeared to be the iron chloride
and small amounts of the iron µ-oxo dimer corrole complex. We
obtained the pure iron chloride by extracting a solution of the crude
product in CH2Cl2, i.e., the eluate from the column chromatography,
three times with 2 M HCl and then removing the CH2Cl2 by rotary
evaporation. Bubbling of HCl gas through a solution of the crude
product can be used as an alternative to extraction for all the analogues
except for the p-CF3 analogue for which this treatment seems to result
in demetalation of the macrocycle. Yields: 55-85%.
Characterization. The compounds were characterized by electronic
absorption, 1H NMR, and 19F NMR spectroscopies, MALDI-TOF mass
spectrometry, and elemental analysis. Elemental analyses of all but two
of the compounds, viz. Cu[F8T(p-CF3-P)C] and Cu[F8T(p-CH3-P)C],
were within 0.5% of the theoretical values. However, even the two
recalcitrant compounds were judged to be pure on the basis of their 1H
NMR, 19F NMR, and MALDI-TOF mass spectra, as indicated below.
The following describes the characterization of the new compounds
reported here. The 1H and 19F NMR spectra are also included as
Supporting Information.
H3[F8T(p-CF3-P)C]. UV-vis: λmax (log ꢀ/(M-1 cm-1)) 404 nm
(5.09), 503 nm (3.91) 540 nm (4.13), 632 nm (4.00). 1H NMR (in the
presence of a small drop of CF3COOH): δ 8.28 (d, J ) 8.0 Hz, 4H,
5,15-o- or m-phenyl), 8.19 (d, J ) 8.0 Hz, 2H 10-o- or m-phenyl),
8.00 (d, J ) 8.0 Hz, 4H, 5,15-o- or m-phenyl), 7.95 (d, J ) 8.0 Hz,
2H, 10-o- or m-phenyl). 19F NMR (in the presence of a small drop of
CF3COOH): δ -62.12 (s, 3F, 10-p-CF3), -62.19 (s, 6F, 5,15-p-CF3),
-147.1 (s, 2F, â-pyrrolic), -148.1 (s, 2F, â-pyrrolic), -148.8 (s, 2F,
â-pyrrolic), -158.8 (s, 2F, â-pyrrolic). MS (MALDI-TOF, major
isotopomer): M+ ) 873.96 (expt.), 874.10 (calcd.). Elemental analy-
sis: 54.75%C (calcd. 54.93%), 1.92%H (calcd. 1.73%), 6.29%N (calcd.
6.41%).
H3[F8TPC]. UV-vis: λmax (log ꢀ/(M-1 cm-1)) 403 nm (4.99), 495
nm (3.99), 537 nm (4.02), 633 nm (4.00). 1H NMR (in the presence of
a small drop of CF3COOH): 7.69-7.33 (m, 15H, all phenyl protons).
19F NMR (in the presence of a small drop of CF3COOH): δ -145.3
(s, 2F, â-pyrrolic), -148.1 (s, 2F, â-pyrrolic), -149.1 (s, 2F, â-pyrrolic),
-155.6 (s, 2F, â-pyrrolic). MS (MALDI-TOF, major isotopomer):
M+ ) 670.02 (expt.), 670.14 (calcd.). Elemental analysis: 66.03%C
(calcd. 66.27%), 2.91%H (calcd. 2.71%), 8.23%N (calcd. 8.36%).
General Synthesis of the Free Base â-Octafluoro-meso-triaryl-
corroles. A 1.5-g portion of basic alumina was placed in an open 50
mL round-bottomed flask with a stirring bar, 3,4-difluoropyrrole (5.92
mmol), an aromatic aldehyde (5.92 mmol), and 1 mL CH2Cl2. The
flask was slowly heated to 70 °C with stirring while the CH2Cl2
evaporated, and the temperature was then maintained for 4 h. Thereafter,
heating was discontinued, 80-100 mL CH2Cl2 was added, solid residues
were dislodged from the walls of the flask with a spatula, the mixture
was swirled, and the solids were brought into solution as quickly and
as much as possible. DDQ (2.96 mmol) was then added to the solution
and the mixture was stirred for an hour. The mixture was then filtered
with a Bu¨chner funnel. The filtrate was evaporated to dryness and the
dry material was subjected to flash chromatography. The mobile phases
used, which varied somewhat for the different corroles synthesized,
are indicated below. The corroles eluted were best detected by their
UV-visible spectra. All the free-base â-octafluoro-meso-triarylcorroles
prepared were also found to be strongly fluorescent. Yields: 6-8%.
Flash Chromatography of â-Octafluoro-meso-tris(p-trifluoro-
methylphenyl)corrole, H3[F8T(p-CF3-P)C]. The crude corrole was first
passed through a short column (∼10 cm) of silica gel with CH2Cl2 as
eluent. The product from this column was then passed through a longer
column (∼22 cm) of silica gel with n-hexane/CH2Cl2, 3:1 initially and
then 2:1, as eluents.
Flash Chromatography of â-Octafluoro-meso-triphenylcorrole,
H3[F8TPC]. The crude corrole was first passed through a short column
(∼10 cm) of silica gel with CH2Cl2 and then 3% MeOH in CH2Cl2 as
eluents. The product from this column was then passed through a longer
column (∼22 cm) of silica gel with n-hexane/CH2Cl2, 3:1 initially and
then 2:1, as eluents.
Flash Chromatography of â-Octafluoro-meso-tris(p-methylphen-
yl)corrole, H3[F8T(p-CH3-P)C]. The crude corrole was first passed
through a short column (∼10 cm) of silica with CH2Cl2 and then 3%
MeOH in CH2Cl2 as eluents. The product from this column was then
passed through a longer column (∼22 cm) of silica gel with n-hexane/
CH2Cl2, 3:1initially and then 2.5:1, as eluents.
Flash Chromatography of â-Octafluoro-meso-tris(p-methoxy-
phenyl)corrole, H3[F8T(p-OCH3-P)C]. The crude corrole was first
passed through a short column (∼10 cm) of silica gel with CH2Cl2 as
eluent. The product from this column was then passed through a longer
column (∼22 cm) of silica gel with successively 1:1, 1:2, and 1:3
n-hexane/CH2Cl2 as eluents.
H3[F8T(p-CH3-P)C]. UV-vis: λmax (log ꢀ/(M-1 cm-1)) 405 nm
(5.03), 503 nm (3.94), 540 nm (4.10), 637 nm (4.09). 1H NMR (in the
presence of a small drop of CF3COOH): δ 7.10-7.70 (m, 12H, all
phenyl protons), 2.65 (s, 9H, 5,10,15-p-CH3). 19F NMR (in the presence
of a small drop of CF3COOH): δ -145.7 (s, 2F, â-pyrrolic), -148.4
(s, 2F, â-pyrrolic), -149.5 (s, 2F, â-pyrrolic), -155.8 (s, 2F, â-pyrrolic).
MS (MALDI-TOF, major isotopomer): M+ ) 712.26 (expt.), 712.19
(calcd.). Elemental analysis: 67.60%C (calcd. 67.42%), 3.57%H (calcd.
3.39%), 7.80%N (calcd. 7.86%).
H3[F8T(p-OCH3-P)C]. UV-vis: λmax (log ꢀ/(M-1 cm-1)) 410 nm
(5.13), 505 nm (3.98), 544 nm (4.18), 642 nm (4.23). 1H NMR (in the
presence of a small drop of CF3COOH): δ 8.12 (s, 4H, 5,15-o- or
m-phenyl), 8.05 (s, 2H, 10-o- or m-phenyl), 7.32 (d, 6H, 5,15-o- or
m-phenyl and 10-o- or m-phenyl), 4.05 (s, 9H, 5,10,15-p-OCH3). 19F
NMR (in the presence of a small drop of CF3COOH): δ -148.2 (s,
2F, â-pyrrolic), -150.8 (s, 2F, â-pyrrolic), -152.5 (s, 2F, â-pyrrolic),
-159.0 (s, 2F, â-pyrrolic). MS (MALDI-TOF, major isotopomer):
M+ ) 760.35 (expt.), 760.17 (calcd.). Elemental analysis: 62.99%C
(calcd. 63.16%), 3.30%H (calcd. 3.18%), 7.19%N (calcd. 7.37%).
General Synthesis of Cu â-Octafluoro-meso-triarylcorroles. Free-
base â-octafluoro-meso-triarylcorrole (∼0.040 g) was dissolved in 25
mL pyridine at room temperature, five equivalents of Cu(OAc)2‚4H2O
were added, and the mixture was stirred for 15 min. The pyridine was
removed by rotary evaporation and the pure product was obtained by
flash chromatography of the residue on silica gel using 2:1 n-hexane/
CH2Cl2 as eluent. Yields: 85-98%.
General Synthesis of FeCl â-Octafluoro-meso-triarylcorroles.
Free-base â-octafluoro-meso-triarylcorrole (∼0.040 g) was dissolved
in 25 mL methanol and heated to reflux. Five equivalents of FeCl2‚
4H2O were added, and the reaction was monitored by UV-visible
spectroscopy. After the solution had been heated for 5-30 min and no
more changes occurred in the UV-visible spectra, heating was
discontinued. The methanol was removed by rotary evaporation and
the product was dissolved in CH2Cl2 and extracted three times with 2
M aqueous HCl. The solvent was again removed by rotary evaporation
and the crude product was chromatographed on silica gel using CH2-
Cu[F8T(p-CF3-P)C]. UV-vis: λmax (log ꢀ/(M-1cm-1)) 401 nm
1
(4.92). H NMR: δ 7.69 (d, J ) 8.0 Hz, 4H, 5,15-o- or m-phenyl),
7.64 (d, J ) 8.0 Hz, 2H 10-o- or m-phenyl), 7.57 (d, J ) 8.0 Hz, 4H,
5,15-o- or m-phenyl), 7.40 (d, J ) 8.0 Hz, 2H, 10-o- or m-phenyl). 19
F
NMR: δ -63.27 (s, 6F, 5,15-p-CF3), -63.30 (s, 3F, 10-p-CF3), -143.9
(s, 2F, â-pyrrolic), -145.5 (s, 2F, â-pyrrolic), -146.5 (s, 2F, â-pyrrolic),
(29) Woller, E. K.; Smirnov, V. V.; DiMagno, S. G. J. Org. Chem. 1998, 63,
5706.
9
16302 J. AM. CHEM. SOC. VOL. 125, NO. 52, 2003

â-Octafluorocorroles
A R T I C L E S
-152.9 (s, 2F, â-pyrrolic). MS (MALDI-TOF, major isotopomer):
M+ ) 934.10 (expt.), 934.01 (calcd.).
1
Cu[F8TPC]. UV-vis: λmax (log ꢀ/(M-1 cm-1)) 409 nm (4.91). H
NMR: δ 7.57 (t, J ) 6.8 Hz, 2H, 10-m-phenyl or 5,15-p-phenyl), 7.50
(d, J ) 6.0 Hz, 4H 5,15-o-phenyl), 7.44-7.34 (m, 7H, 5,15-m-phenyl,
10-p-phenyl and 5,15-p-phenyl or 10-o- or m-phenyl), 7.24 (s, 2H, 5,15-
p-phenyl or 10-o- or m-phenyl; overlaps with CHCl3 peak). 19F NMR:
δ -144.9 (s, 2F, â-pyrrolic), -146.7 (s, 2F, â-pyrrolic), -147.7 (s,
2F, â-pyrrolic), -154.7 (s, 2F, â-pyrrolic). MS (MALDI-TOF, major
isotopomer): M+ ) 730.23 (expt.), 730.05 (calcd.). Elemental analy-
sis: 60.95%C (calcd. 60.79%), 2.19%H (calcd. 2.07%), 7.48%N (calcd.
7.66%).
Cu[F8T(p-CH3-P)C]. UV-vis: λmax (log ꢀ/(M-1 cm-1)) 380 nm
(4.73, sh.), 421 nm (4.97). 1H NMR: δ 7.40 (d, J ) 7.6 Hz, 4H, 5,15-
o- or m-phenyl), 7.24 (d, 2H, 10-o- or m-phenyl; overlaps with CHCl3
peak), 7.20 (d, J ) 7.6 Hz, 4H, 5,15-o- or m-phenyl), 7.14 (d, J ) 7.2
Hz, 2H, 10-o- or m-phenyl), 2.38 (s, 6H, 5,15-p-CH3), 2.34 (s, 3H,
10-p-CH3). 19F NMR: δ -145.2 (s, 2F, â-pyrrolic), -147.2 (s, 2F,
â-pyrrolic), -148.2 (s, 2F, â-pyrrolic), -155.3 (s, 2F, â-pyrrolic). MS
(MALDI-TOF, major isotopomer): M+ ) 771.99 (expt.), 772.09
(calcd.).
Cu[F8T(p-OCH3-P)C]. UV-vis: λmax (log ꢀ/(M-1 cm-1)) 380 nm
(4.78), 436 nm (4.85), 535 nm (4.24, broad). 1H NMR: δ 7.53 (d, J )
8.0 Hz, 4H, 5,15-o- or m-phenyl), 7.40 (d, J ) 8.4 Hz, 2H, 10-o- or
m-phenyl), 6.93 (d, J ) 8.0 Hz, 4H, 5,15-o- or m-phenyl), 6.88 (d,
J ) 8.0 Hz, 2H, 10-o- or m-phenyl), 3.88 (s, 6H, 5,15-p-OCH3), 3.86
(s, 3H, 10-p-OCH3). 19F NMR: δ -145.5 (s, 2F, â-pyrrolic), -147.6
(s, 2F, â-pyrrolic), -148.9 (s, 2F, â-pyrrolic), -155.8 (s, 2F, â-pyrrolic).
MS (MALDI-TOF, major isotopomer): M+ ) 820.11 (expt.), 820.08
(calcd.). Elemental analysis: 58.26%C (calcd. 58.51%), 2.78%H (calcd.
2.58%), 6.66%N (calcd. 6.82%).
Fe[F8T(p-CF3-P)C]Cl. UV-vis: λmax (log ꢀ/(M-1cm-1)) 353 nm
(4.78). H NMR: δ 19.9 (s, 2H, 5,15-o-phenyl), 19.0 (s, 2H, 5,15-o-
1
Figure 2. (a) H and (b) 19F NMR spectra of Cu[F8T(p-CF3-P)C].
1
phenyl), 18.0 (s, 2H, 10-o-phenyl), -1.16 (s, 2H, 5,15-m-phenyl), -1.42
(s, 2H, 5,15-m-phenyl), -2.14 (s, 2H, 10-m-phenyl). 19F NMR: δ -32.2
(s, 2F, â-pyrrolic), -40.5 (s, 2F, â-pyrrolic), -47.0 (s, 2F, â-pyrrolic),
-59.9 (s, 2F, â-pyrrolic), -69.2 (s, 3F, 10-p-CF3), -71.5 (s, 6F, 5,15-
p-CF3). MS (MALDI-TOF, major isotopomer): M+ - Cl ) 927.25
(expt.), 927.01 (calcd.). Elemental analysis: 50.14%C (calcd. 49.90%),
1.46%H (calcd. 1.26%), 5.96%N (calcd. 5.82%).
analysis: 56.55%C (calcd. 56.59%), 2.64%H (calcd. 2.49%), 6.41%N
(calcd. 6.60%).
DFT Calculations. DFT calculations on FeCl corrole and â-octa-
fluorocorrole were carried out using Slater-type valence triple-ú plus
polarization basis sets, the VWN local functional, the Perdew-Wang
1991 gradient corrections, a spin-unrestricted formalism, a fine mesh
for numerical integration of matrix elements, full geometry optimiza-
Fe[F8TPC]Cl. UV-vis: λmax (log ꢀ/(M-1 cm-1)) 355 nm (4.73).
1H NMR: δ 20.9 (s, 2H, 5,15-o-phenyl), 20.1 (s, 2H, 5,15-o-phenyl),
18.8 (s, 2H, 10-o-phenyl), 15.5 (s, 2H, 5,15-p-phenyl), 13.0 (s, 1H,
10-p-phenyl), -2.16 (s, 2H, 5,15-m-phenyl), -2.39 (s, 2H, 5,15-m-
phenyl), -3.29 (s, 2H, 10-m-phenyl). 19F NMR: δ -29.8 (s, 2F,
â-pyrrolic), -44.3 (s, 2F, â-pyrrolic), -52.9 (s, 2F, â-pyrrolic), -58.3
30
tions with Cs symmetry constraints, and the ADF program system.
Similar calculations with C2V symmetry constraints and full geometry
optimizations were also carried out for three electronic states of copper
octafluorocorrole: (a) the S ) 0 Cu(III) state, (b) a 3A2 Cu(II) corrole
3
b1 radical state, and (c) a B1 Cu(II) corrole a2 radical state.
(s, 2F, â-pyrrolic). MS (MALDI-TOF, major isotopomer): M+
-
Results and Discussion
Cl ) 723.51 (expt.), 723.05 (calcd.). Elemental analysis: 58.74%C
(calcd. 58.56%), 2.19%H (calcd. 1.99%), 7.19%N (calcd. 7.38%).
Fe[F8T(p-CH3-P)C]Cl. UV-vis: λmax (log ꢀ/(M-1 cm-1)) 360 nm
NMR Spectroscopy. Figure 2 shows the room-temperature
1H and 19F NMR spectra of Cu[F8T(p-CF3-P)C]. The 1H NMR
spectrum exhibits four doublets (ratio ) 2:2:1:1, total of 12
protons) in the aromatic region which correspond to the ortho-
and meta-protons of the symmetry-equivalent 5- and 15-aryl
groups (two doublets of 4 protons each) and of the 10-aryl group
(two doublets of 2 protons each). The 19F NMR spectrum
exhibits two peaks around -63 ppm (ratio ) 2:1, total of 9
fluorines) corresponding to the para-CF3 groups on the sym-
metry-equivalent 5- and 15-aryl groups and on the 10-aryl group,
respectively. Four peaks also occur in the -143 to -153 ppm
(ratio ) 1:1:1:1, total 8 fluorines) range corresponding to the
1
(4.78). H NMR: δ 22.0 (s, 2H, 5,15-o-phenyl), 21.2 (s, 2H, 5,15-o-
phenyl), 19.7 (s, 2H, 10-o-phenyl), -2.98 (s, 2H, 5,15-m-phenyl), -3.16
(s, 2H, 5,15-m-phenyl), -4.25 (s, 2H, 10-m-phenyl), -5.08 (s, 3H,
10-p-CH3), -8.12 (s, 6H, 5,15-p-CH3). 19F NMR: δ -29.6 (s, 2F,
â-pyrrolic), -46.8 (s, 2F, â-pyrrolic), -55.5 (s, 2F, â-pyrrolic), -57.4
(s, 2F, â-pyrrolic). MS (MALDI-TOF, major isotopomer): M+
-
Cl ) 765.54 (expt.), 765.10 (calcd.). Elemental analysis: 59.96%C
(calcd. 59.99%), 2.90%H (calcd. 2.64%), 6.76%N (calcd. 7.00%).
Fe[F8T(p-OCH3-P)C]Cl. UV-vis: λmax (log ꢀ/(M-1cm-1)) 367 nm
1
(4.80). H NMR: δ 23.2 (s, 2H, 5,15-o-phenyl), 22.7 (s, 2H, 5,15-o-
phenyl), 20.9 (s, 1H, 10-o-phenyl), 20.7 (s, 1H, 10-o-phenyl), -2.09
(s, 4H, 5,15-m-phenyl), -3.36 (s, 1H, 10-m-phenyl), -3.41 (s, 1H,
10-m-phenyl), 2.34 (s, 3H, 10-p-OCH3), 1.86 (s, 6H, 5,15-p-OCH3).
19F NMR: δ -27.8 (s, 2F, â-pyrrolic), -50.3 (s, 2F, â-pyrrolic), -57.7
(s, 4F, â-pyrrolic; two overlapping peaks). MS (MALDI-TOF, major
isotopomer): M+ - Cl ) 813.17 (expt.), 813.08 (calcd.). Elemental
(30) The ADF program system was obtained from Scientific Computing and
Modeling, Department of Theoretical Chemistry, Vrije Universiteit, 1081
HV Amsterdam, The Netherlands. For details of basis sets, grids for
numerical integration, etc., the reader is referred to the program manual
obtainable from this source.
9
J. AM. CHEM. SOC. VOL. 125, NO. 52, 2003 16303

A R T I C L E S
Steene et al.
observed in the 1H NMR spectra of a copper â-octaalkylcorrole
by E. Vogel, M. Gross and co-workers,32 who attributed their
observation to a thermally accessible copper(II) corrole π-cation
radical excited state. We suggest that a similar scenario holds
for the copper octafluorocorroles reported here. The fact that
only two of the four â-19F peaks broaden significantly with
temperature may reflect the fact that only two of the four
symmetry-unique â-positions of a corrole b1-type radical carry
significant spin density.33 We also investigated the electronic
absorption and Soret-resonant Raman spectra34 of the Cu
octafluorocorroles as a function of temperature, but these
remained essentially unchanged at low temperatures, suggesting
that the diamagnetic Cu(III) ground state is the dominant species
between -40 °C and 40 °C.
Figure 4 shows the room-temperature 1H and 19F NMR
spectra of Fe[F8TPC]Cl. The other three FeCl complexes, viz.
Fe[F8T(p-CF3-P)C]Cl, Fe[F8T(p-CH3-P)C]Cl, and Fe[F8T(p-
OCH3-P)C]Cl, exhibit similar spectra (see Supporting Informa-
tion Figures S3-S15). The general pattern of peaks observed
1
in the H NMR spectra of these complexes is similar to what
has previously been observed for their â-unsubstituted ana-
logues22,24 except that the â-proton peaks obviously are missing
1
for the present compounds. All peaks in the H NMR spectra
were assigned based on COSY experiments (see Figures S4,
S8, S11, and S14 in the Supporting Information). Note that the
ortho (as well as the meta) protons of the 10-aryl group in all
these complexes should have given rise to two peaks in the
spectra, as is the case for the ortho and meta protons of the
symmetry-equivalent 5- and 15-aryl groups. However, for the
10-aryl group, only one peak for the ortho protons and one for
the meta protons are observed for the p-CF3, p-H, and p-CH3
FeCl derivatives, because these peaks are not resolved at 400
MHz. In contrast, for the p-OCH3 FeCl derivative, the peaks
are only partially overlapping and there are two peaks both for
the ortho and for the meta protons of the 10-aryl group. Similar
overlapping peaks for the two symmetry-distinct ortho and meta
protons in the 10-aryl group are also observed for the FeCl
complexes of â-unfluorinated meso-triarylcorroles.22,24
Figure 3. 19F NMR spectra of Cu[F8T(p-CF3-P)C] at various temperatures.
â-fluorines, two of which are seen to couple in the 19F-19F
COSY spectrum (see Supporting Information Figure S28). As
might be expected, the 1H and 19F NMR spectra of Cu[F8TPC],
Cu[F8T(p-CH3-P)C], and Cu[F8T(p-OCH3-P)C] (see Support-
ing Information Figures S15-S27) are qualitatively similar to
those of Cu[F8T(p-CF3-P)C].
1
Table 1 summarizes the isotropic shifts (δiso) of all H and
19F peaks as well as their assignments for the four FeCl
â-octafluoro-meso-triarylcorroles reported in this paper, along
with analogous data for some other relevant FeCl meso-
triarylcorroles. The substantial isotropic shifts seen for the
phenyl protons of the â-fluorinated complexes are qualitatively
similar to those seen for their â-unsubstituted analogues,22,24
The 1H and 19F NMR spectra of the Cu corroles are
qualitatively very similar to those of the protonated free-base
corroles (see Supporting Information Figures S36-S45).31 Not
a great deal of comment is warranted on the latter spectra except
for a couple of points. Thus, neither the free bases nor the Cu
complexes exhibit 19F peaks that are doublets, which implies
very small 19F-19F J-couplings in both classes of compounds.
This is consistent with the small J-couplings (3-5 Hz) observed
for the â-protons of the Cu derivatives of â-unfluorinated
triarylcorroles.17
The 19F NMR spectra of the Cu octafluorocorroles exhibit a
remarkable temperature dependence: for each Cu complex, two
of the four â-19F peaks broaden significantly with temperature
between -40 and +40 °C, as shown for Cu[F8T(p-CF3-P)C]
in Figure 3, which is quite unlike the behavior of the protonated
free bases. The temperature-dependent 19F NMR behavior of
the Cu octafluorocorroles is reminiscent of similar behavior
(32) Will, S.; Lex, J.; Vogel, E.; Schmickler, H.; Gisselbrecht, J.-P.; Haubtmann,
C.; Bernard, M.; Gross, M. Angew. Chem., Int. Ed. Engl. 1997, 36, 357.
For a new, relevant article, see: Bruckner, C.; Brinas, R. P.; Krause Bauer,
J. A. Inorg. Chem. 2003, 42, 4495.
(33) Ghosh, A.; Wondimagegn, T.; Parusel, A. B. J. J. Am. Chem. Soc. 2000,
122, 5100.
(34) These Resonance Raman (RR) spectra can be found in Figures S32-S35
of the Supporting Information. Because these spectra have not been
assigned, we refrain from describing them in detail. We only note, however,
that an intense high-frequency marker band at approximately 1487-1492
cm-1 for the â-unsubstituted copper triarylcorroles17 appears to downshift
to 1461-1467 cm-1 for the Cu â-octafluorocorroles studied here and to
1458-1462 cm-1 for Cu â-octabromocorroles.17 We speculate that this
vibration has Câ-Câ stretching character which would indicate that
â-octafluorination and â-octabromination have
a comparable bond-
lengthening influence on the Câ-Câ bonds. Because â-octafluorination
and â-octabromination may be expected to induce different degrees of
nonplanarity on the corrole ring, the similarity of this frequency for the
two classes of halogenated copper corroles may indicate that the halogena-
tion-induced downshift of this vibration reflects a direct bond-lengthening
effect of the â-halogen substituents on the Câ-Câ bonds rather than a
nonplanarity effect.
(31) The line widths of the 1H and 19F NMR peaks of the protonated free-base
corroles are sensitive to the amount of acid present in the NMR samples.
We have not investigated this phenomenon in depth but have empirically
found that the sharpest spectra are obtained with only a small drop of added
trifluoroacetic acid. Excess acid leads to peak broadening in these spectra.
9
16304 J. AM. CHEM. SOC. VOL. 125, NO. 52, 2003

â-Octafluorocorroles
A R T I C L E S
Table 1. 1H and 19F NMR Isotropic Shiftsa (ppm) for Different FeCl Meso-triarylcorroles
isotropic shifts (ppm)a
5,15-p-CF /
10-p-CF /
3
3
compound
Fe[F8T(p-CF3-P)C]Clb CDCl3, 12.2(2H), 10.3(2H) -8.64(2H), -9.61(2H)
298 K 11.4(2H) -8.88(2H)
CDCl3, 13.4(2H), 11.3(2H) -9.67(2H), -10.8(2H) 7.98(2H) 5.46(1H)
298 K 12.6(2H) -9.90(2H)
Fe[F8T(p-CH3-P)C]Clb CDCl3, 14.6(2H), 12.3(2H) -10.3(2H), -11.5(2H)
298 K 13.8(2H) -10.4(2H)
Fe[F8T(p-OCH3-P)C]Clb CDCl3, 16.0(2H), 13.7(1H), -9.29(4H) -10.6(1H),
conditions 5,15-o-H
10-o-H
5,15-m-H
10-m-H
5,15-p-H
10-p-H
CH /OCH
CH /OCH
3
â-pyrrole H/F
3
3
3
-8.25(6F) -5.88(3F) 115(2F), 107(2F),
100(2F), 87.3(2F)
Fe[F8TPC]Clb
119(2F), 104(2F),
95.6(2F), 90.2(2F)
-10.5(6H) -7.44(3H) 119(2F), 102(2F),
93.4(2F), 91.4(2F)
-2.02(6H) -1.52(3H) 122(2F), 99.2(2F),
91.8(4F)
298 K 15.5(2H) 13.5(1H)
CDCl3, 16.3(2H), 14.3(2H) -9.45(2H), -10.5(1H),
rt 15.1(2H) -9.65(2H) -10.6(1H)
CDCl3, 17.6(2H), 15.5(2H) -10.2(2H), -11.3(1H), 12.0(2H) 9.68(1H)
rt 16.4(2H) -10.5(2H) -11.5(1H)
CDCl3, 18.5(2H), 16.4(2H) -10.8(2H), -12.0(1H),
-10.6(1H)
Fe[T(p-CF3-P)C]Cl24
Fe[TPC]Cl24
-2.15(2H), -14.0(2H),
-14.5(2H), -47.5(2H)
-1.59(2H), -13.4(2H),
-14.5(2H), -48.7(2H)
Fe[T(p-CH3-P)C]Cl24
-14.3(6H) -12.0(3H) -1.85(2H), -12.9(2H),
-14.8(2H), -48.9(2H)
rt
17.2(2H)
-11.0(2H) -12.4(1H)
Fe[T(p-OCH3-P)C]Cl22 CD2Cl2, 18.9(2H), 17.0(1H), -11.3(2H), -12.6(1H),
293 K 17.7(2H) 16.8(1H) -11.5(2H) -13.0(1H)
-1.9(2H), -12.6(2H),
-21.5(2H), -45.3(2H)
-2.2(2H), -14.2(2H),
Fe[T(p-NO2-P)C]Cl22
CD2Cl2, 14.6(2H), 12.8(2H) -9.5(2H), -10.5(1H),
303 K 13.6(2H) -9.8(2H) -10.6(1H)
-15.4(2H), -46.1(2H)
a The isotropic shifts were calculated by subtracting the corresponding diamagnetic shifts from the chemical shifts. Average 1H diamagnetic shifts of 7.7
ppm for Cu[T(p-CF3-P)C], 7.6 ppm for Cu[TPC], 7.5 ppm for Cu[T(p-CH3-P)C], 7.6 ppm for Cu[F8T(p-CF3-P)C], 7.5 ppm for Cu[F8TPC], 7.3 ppm for
Cu[F8T(p-CH3-P)C], and 7.2 ppm for Cu[F8T(p-OCH3-P)C] were used for the pyrrole and phenyl 1H peaks because not all these peaks have been assigned
for neither the copper complexes, nor the iron complexes. For the same reason, average 19F diamagnetic shifts of -147.2 ppm for Cu[F8T(p-CF3-P)C],
-148.5 ppm for Cu[F8TPC], -149.0 ppm for Cu[F8T(p-CH3-P)C], and -149.5 ppm for Cu[F8T(p-OCH3-P)C] were used for the pyrrole 19F peaks. b This
work.
the meso-phenyl protons alternate in sign around the phenyl
ring (e.g., p-H and o-H positive, m-H negative) in the new
octafluorinated complexes. These signs are opposite to the ones
in Fe[(TPP)(t-BuNC)2]+, which have been shown to have large
positive π spin density on the meso carbons.35,36 Hence, in FeCl
â-octafluoro-meso-triarylcorroles, as well as in their â-unsub-
stituted analogues, there is substantial negative π spin density
on the meso positions of the corrole ligand, which in turn
indicates antiferromagnetic coupling between a corrolate π-cat-
ion radical and an S ) 3/2 iron center. Table 1 also shows that
the isotropic shifts for the FeCl â-octafluoro-meso-triarylcorroles
are somewhat smaller than for the corresponding â-unsubstituted
analogues, indicating somewhat less radical character on the
corrole macrocycle. The electron-deficient fluorinated corroles
are thus somewhat less noninnocent than “ordinary” corroles.
Gratifyingly, the FeCl octafluorocorroles also yielded broad
but distinctly discernible 19F NMR signals (Figure 4b). Until
now, the only significant 19F NMR study of relevant paramag-
netic compounds is one by Yatsunyk and Walker37 on iron(III)
meso-pentafluorophenyl porphyrins and corroles, but this study
necessarily focused on only phenyl-bound fluorines. We were
therefore intrigued by the possibility that the present FeCl
corroles, being â-fluorinated, might exhibit very large 19F
isotropic shifts and thereby advance our understanding of 19F
NMR chemical shifts in paramagnetic compounds in general.
Although a complete treatment of this topic is outside the scope
of this paper, a few interesting observations pertaining to this
issue may be made. If we choose the mean of the chemical
shifts of the â-fluorines in the copper complexes as the reference
(for calculating the isotropic shifts), then the isotropic shifts of
the â-fluorines in the corresponding FeCl complexes span a
1
Figure 4. (a) H and (b) 19F NMR spectra of Fe[F8TPC]Cl.
range of 87-122 ppm, as shown in Table 1. While signifi-
(35) Simonneaux, G.; Hindre, F.; Le Plouzennec, M. Inorg. Chem. 1989, 28,
indicating the presence of substantial radical character at the
meso positions. As noted by Walker and co-workers for a
number of other FeCl triarylcorroles,22,23 the isotropic shifts of
823.
(36) Walker, F. A. In The Porhyrin Handbook; Kadish, K. M., Smith, K. M.,
Guilard, R., Eds.; Academic Press: San Diego, 2000; Vol. 5, p 81.
(37) Yatsunyk, L.; Walker, F. A. Inorg. Chim. Acta 2002, 337, 266.
9
J. AM. CHEM. SOC. VOL. 125, NO. 52, 2003 16305

A R T I C L E S
Steene et al.
Table 2. Half-Wave Potentials, E1/2 (V vs SCE) and Hammett F
Values (mV), for Cu, MnCl, FeCl and FeOFe Triarylcorroles in
CH2Cl2 Containing 0.1 M TBAP
compound
X subst.
3σ
E
1/2ox
E
1/2red
Fox Fred
ref
Cu[T(p-X-P)C]
OCH3 -0.80 0.65 -0.24 95 68
17
CH3
H
-0.51 0.70 -0.23
0.76 -0.20
1.62 0.89 -0.08
0
CF3
Cu[Br8T(p-X-P)C] OCH3 -0.80 1.10
0.04 58 86
17
CH3
H
CF3
-0.51 1.12
1.14
1.62 1.24
0.07
0.12
0.25
0
Cu[F8T(p-X-P)C]
OCH3 -0.80 1.06
0.17 67 77 this work
CH3
H
-0.51 1.12
0.18
0.22
0.35
0
1.15
CF3
CH3
H
1.62 1.24
Mn[T(p-X-P)C]Cl
Fe[T(p-X-P)C]Cl
-0.51 0.97
0.07 82 77
0.09
0.23
0.03 74 77
0.05
0.19
24
24
24
Figure 5. Cyclic voltammograms of the Cu â-octafluoro-meso-tris(p-X-
phenyl)corroles (X ) OCH3, CH3, H, and CF3) prepared in this work.
0
1.03
CF3
CH3
H
1.62 1.15
-0.51 1.02
0
1.07
CF3
1.62 1.18
{Fe[T(p-X-P)C]}2O CH3
-0.51 0.59 -0.35 114 109
0.64 -0.31
1.62 0.83 -0.12
H
CF3
0
Fe[F8T(p-X-P)C]Cl OCH3 -0.80 1.34
0.47 49 68 this work
CH3
H
CF3
-0.51 1.38
1.40
1.62 1.47
0.48
0.49
0.63
0
X-ray photoelectron spectra (XPS, Figure 7) of the free bases
which show that the core ionization potentials of the protonated
and unprotonated nitrogens of the fluorinated corroles are about
38,39
0.5 eV upshifted, relative to free-base octamethylcorrole
and 0.7 eV upshifted, relative to free-base tetraphenylporphy-
rin.40,41 A detailed description of the XPS of various corrole
derivatives prepared in our laboratory will be presented else-
where.
Figure 6. Cyclic voltammograms of the FeCl â-octafluoro-meso-tris(p-
X-phenyl)corroles (X ) OCH3, CH3, H, and CF3) prepared in this work.
cant and indeed larger than those observed for Fe(TPFPC)Cl
(TPFPC ) meso-tris(pentafluoropenyl)corrole),22 these shifts are
by no means enormous, indicating that the highly electronegative
fluorine atoms carry only relatively modest spin densities, i.e.,
relatively little of the π-cation radical character of the corrole
macrocycles delocalizes onto the fluorines. The DFT calcula-
tions described later in this paper appear to contribute some
additional insights into this issue.
In our previous work,24,26 we have proposed that the oxidation
half-wave potentials of metallocorroles serve as a probe of
electronic character of the corrole ligand. This proposal was
based on the observation that FeCl, MnCl, SnCl, and CrO
corrole complexes generally exhibit higher oxidation half-wave
potentials than analogous Cu, SnPh, FePh, and Fe-O-Fe
complexes. We proposed that this indicates that the corrole
ligands are substantially noninnocent (i.e., exhibit corrole
π-cation radical character) in the former complexes while they
are more innocent in the latter group of complexes.26 The
complexes reported here exhibit the same trend. Thus, the half-
wave oxidation potentials of the FeCl â-octafluoro-meso-
triarylcorroles are 230-270 mV higher than those of the
corresponding Cu complexes, suggesting that in the FeCl
â-octafluoro-meso-triarylcorrole complexes the corrole ligand
is noninnocent. As before, we freely admit that this interpretation
of the electrochemical results is speculative, but it is in
agreement with the conclusions drawn on the basis of the NMR
results.
Electrochemistry and XPS. Figures 5 and 6 present the
cyclic voltammograms of the new complexes described in this
paper. Table 2 presents the half-wave potentials for one-electron
oxidation and reduction of the Cu and FeCl â-octafluoro-meso-
triarylcorroles, along with those of certain other relevant
complexes. Despite the somewhat limited data set of four
compounds, calculated Hammett F values of substituent effects
on the half-wave potentials are included in Table 2. The
corresponding Hammet plots are shown in Figures S1 and S2
of the Supporting Information.
Table 2 shows that the half-wave potentials for one-electron
oxidation of the Cu â-octafluoro-meso-triarylcorroles are posi-
tively shifted by approximately 400 mV, compared to their
â-unsubstituted analogues. Interestingly, this positive shift due
to â-octafluorination is identical to that previously observed for
â-octabromination.17 However, the Cu octafluorocorroles exhibit
slightly higher half-wave potentials for reduction, i.e., are
somewhat easier to reduce, than Cu octabromocorroles.17
A closer examination of the data in Table 2 allows us to
engage in further speculation, vis-a`-vis the electronic character
of the fluorinated metallocorroles. In particular, we note that
(38) Licoccia, S.; Paolesse, R. Struct. Bonding 1995, 84, 71.
(39) Zanoni, R.; Boschi, T.; Licoccia, S.; Paolesse, R.; Tagliatesta, P. Inorg.
Chim. Acta 1988, 145, 175.
(40) Gassman, P. G.; Ghosh, A.; Almlo¨f, J. J. Am. Chem. Soc. 1992, 114, 9990.
(41) Ghosh, A.; Moulder, J.; Bro¨ring, M.; Vogel, E. Angew. Chem., Int. Ed.
2001, 113, 431.
The relatively high oxidation half-wave potentials of the
fluorinated corroles are also consistent with the nitrogen 1s
9
16306 J. AM. CHEM. SOC. VOL. 125, NO. 52, 2003

â-Octafluorocorroles
A R T I C L E S
Figure 7. (a) Carbon 1s and (b) nitrogen 1s XPS of H3[F8T(p-CF3-P)C]. Note that the C 1s peak due to the CF3 carbons is well-resolved at 292.5 eV. All
of the â-octafluorocorroles also exhibit a resolved peak at 286.8 eV, which we assign to the fluorinated â-carbons.
Figure 9. UV-visible spectra of the Cu â-octafluoro-meso-tris(p-X-
phenyl)corroles (X ) OCH3, CH3, H, and CF3) prepared in this work.
Figure 8. UV-visible spectra of the free-base â-octafluoro-meso-tris(p-
X-phenyl)corroles (X ) OCH3, CH3, H, and CF3) prepared in this work.
the difference in the oxidation half-wave potentials between the
FeCl and Cu corrole complexes (E1/2(ox)(FeCl) - E1/2(ox)(Cu))
decreases as the corrole ligand becomes more electron deficient.
Thus, [E1/2(ox)(FeCl) - E1/2(ox)(Cu)] decreases in the following
order: T(p-CH3-P)C (315 mV) > TPC (308mV) > T(p-CF3-
P)C (288 mV) > F8T(p-OCH3-P)C (280 mV) > F8T(p-CH3-
P)C (260 mV) > F8TPC (250 mV) > F8T(p-CF3-P)C (230
mV). In other words, the oxidation half-wave potentials of the
FeCl and Cu corrole complexes approach one another with
increasing electron deficient character of the corrole ligand. We
propose that this reflects the fact that a very electron-deficient
ligand is always relatively innocent, regardless of the coordi-
nated metal ion. Accordingly, for ligand-centered oxidation of
such a ligand, different metal complexes should exhibit similar
oxidation half-wave potentials.
Figure 10. UV-visible spectra of the FeCl â-octafluoro-meso-tris(p-X-
phenyl)corroles (X ) OCH3, CH3, H, and CF3) prepared in this work.
The Hammett F () (1/3)(dE1/2/dσ)) of 67 mV for the
oxidation half-wave potentials of the Cu â-octafluoro-meso-
triarylcorroles is considerably smaller than that for the corre-
sponding unfluorinated Cu corroles, but similar in magnitude
to that for Cu â-octabromo-meso-triarylcorroles (Table 2). As
in the case of the Cu octabromocorroles,17 we propose that the
small Hammett F in this case reflects a HOMO switch among
the three fluorinated Cu corroles studied, i.e., the more electron-
rich Cu[F8T(p-OCH3-P)C], Cu[F8T(p-CH3-P)C], and Cu[F8-
TPC] may have a b1 HOMO, whereas the more electron
deficient Cu[F8T(p-CF3-P)C] may have an a2 HOMO. It may
be recalled that the corrole a2 and b1 HOMOs crudely resemble
the a1u and a2u HOMOs of porphyrins, respectively.
Electronic Absorption Spectra. The electronic absorption
spectra of the free-base, Cu, and FeCl â-octafluorocorrole
derivatives are shown in Figure 8, 9, and 10, respectively. The
Soret maxima are listed in Table 3, along with those of other
relevant corrole derivatives. As observed for free-base meso-
triarylcorroles which are not fluorinated at the â positions,17,24
the Soret maxima of free-base â-octafluoro-meso-triarylcorroles
are not particularly substituent-dependent. In contrast, the Soret
maxima of the Cu â-octafluoro-meso-triarylcorroles red-shift
strongly with increasing electron-donating character of the para
substituents on the aryl group. The Soret maxima of the FeCl
â-octafluoro-meso-triarylcorroles also red-shift somewhat in
9
J. AM. CHEM. SOC. VOL. 125, NO. 52, 2003 16307

A R T I C L E S
Steene et al.
Table 3. “Soret” Band Maxima (nm) in CH2Cl2 for Some Relevant
Triarylcorroles
substituent (X)
corrole
p-OCH
p-CH
p-H p-CF
F
5
ref
3
3
3
H3[T(p-X-P)C]
Cu[T(p-X-P)C]
421
433
417
418
453
441
421
389
405
417 417 408
413 407 406
439 436 442
433 423
411 402
385 384
17
17
17
24
24
Cu[Br8T(p-X-P)C] 468
Mn[T(p-X-P)C]Cl
Fe[T(p-X-P)C]Cl
{Fe[T(p-X-P)C]}2O
24
H3[F8T(p-X-P)C]
Cu[F8T(p-X-P)C]
410
403 404
this work
this work
this work
436, 380 421, 380 (sh) 409 401
360 355 353
Fe[F8T(p-X-P)C]Cl 367
response to meso substituents. As shown in Table 3, the
substituent sensitivity seen for Cu â-octafluoro-meso-triaryl-
corroles has also been observed for Cu, FeCl, and MnCl
complexes of â-unsusbstituted meso-triarylcorroles and also for
Cu â-octabromo-meso-triarylcorroles. This phenomenon thus
appears to be a fairly general feature of high-valent transition
metal corrole complexes. We have ascribed these red shifts to
significant charge transfer (CT) character in one or more
electronic transitions in the Soret region.17,24 DFT/CI and time-
dependent DFT(PW91/TZP) calculations on gallium corrole
(C2V) and copper corrole (C2V) revealed a number of CT
transitions of high oscillator strength in the Soret region of Cu
corrole, but not in the case of Ga corrole.33 The CT transitions
in the case of Cu corrole involve excitation into the LUMO,
Figure 11. Spin-unrestricted DFT(PW91/TZP) spin populations for FeCl
â-octafluorocorrole and its â-unsubstituted analogue. Also shown are the
spin populations of the Cu(II) b1 and a2 corrole radical states of Cu
â-octafluorocorrole.
2
2
which is of b2 symmetry and has predominant Cu(dx -y
)
examination of the high-lying occupied MOs revealed an MO
character. The low-energy transitions of Ga corrole, in contrast,
are all π-π* transitions.
2
with a metal(dz )-corrole(“b1”) overlap. This overlap is clearly
facilitated by the significant displacements of 0.39 Å of the iron
from the N4 plane of the corrole in the optimized geometry. As
in the case of the â-unsubstituted analogue,24 these results are
consistent with an electronic structure involving a corrole π
cation radical antiferromagnetically coupled to an intermediate
spin (S ) 3/2) Fe(III) center. The DFT calculations thus support
the conclusions drawn from the electrochemical and NMR
results.
DFT Calculations. Spin-unrestricted DFT (PW91/TZP)
calculations on Cu â-octafluorocorrole shows that the S ) 1
Cu(II) b1 and a2 radical states are only 0.11 and 0.19 eV,
respectively, higher in energy than the S ) 0 Cu(III) state. These
results are qualitatively consistent with the temperature-depend-
ent 19F spectra of the Cu corroles studied, which indicate the
presence of a thermally accessible paramagnetic exited state.
Perhaps not surprisingly, the low energies of the Cu(II) radical
states are also consistent with what we found for â-unsubsituted
Cu corrole, although this could not have been predicted a priori.
In other words, â-octafluorination does not significantly effect
the relative energetics of the Cu(III) versus the Cu(II) radical
states.
Spin-unrestricted DFT (PW91/TZP) calculations on FeCl
â-octafluorocorrole reveal a spin density profile (Figure 11)
similar to the one calculated for the â-unsubstituted analogue.24
Thus, the Fe and Cl atoms carry 1.91 and 0.21 R spins and the
â-carbons carry a total of 0.12 R spins, respectively. In contrast,
the meso carbons carry a total of 0.24 â spins and the nitrogens
a total 0.11 â spins. The spatial distribution of the â spins
coincides precisely with the shape of the b1 HOMO of the
corrole ligand,33 which crudely resembles the a2u HOMO of
porphyrins. The overall spin density profile of this molecule,
in particular the significant spatial separation of the R and â
spins, is further highly characteristic of antiferromagnetically
coupled spin systems.21,33,42,43 The antiferromagnetic coupling
in FeCl â-octafluorocorrole seems to involve the metal center
on one hand and the corrole b1 HOMO on the other hand. An
A comparison of the spin density profile of the FeCl
â-octafluorocorrole with the â-unsubstituted counterpart reveals
that the meso carbons and nitrogens in the octafluorocorrole
complex carry only about 80% of the â-spin found at the same
positions in the â-unsubstituted complex (Figure 11). In other
words, the octafluorocorrole ligands in the FeCl complexes are
substantially noninnocent, but less so than their â-unsubstituted
counterparts. This is consistent with the smaller isotropic shifts
1
observed in the H NMR of the Fe[F8T(p-X-P)C]Cl (X )
OCH3, CH3, H, and CF3) series compared to the Fe[T(p-X-
P)C]Cl (X ) CH3, H, and CF3) series (Table 1).
Finally, we make an attempt to relate the calculated spin
density profile of FeCl â-octafluorocorrole with the 19F NMR
spectra of the FeCl corroles studied. Figure 11 shows that all
four symmetry-distinct â-fluorines carry small, positive spin
densities, whereas the carbon-nitrogen skeleton of the corrole
carries a net negative spin density. The fluorine spin densities
(43) It should be noted that the PW91 functional used by us and the B3LYP
functional used in ref 21 give qualitatively similar but, quantitatively,
significantly different spin populations for FeCl corroles. In the absence
of higher-level calculations of these results, we cannot say whether PW91
or B3LYP provide a better description of reality. For a review of calibration
of PW91 and B3LYP calculations on biologically relevant transition metal
complexes against CASPT2 and CCSD(T) calculations, see: Ghosh, A.;
Taylor, P. R. Curr. Opin. Chem. Biol. 2003, 7, 113.
(42) Ghosh, A.; Vangberg, T.; Gonzalez, E.; Taylor, P. J. Porphyrins Phtha-
locyanines 2001, 5, 345.
9
16308 J. AM. CHEM. SOC. VOL. 125, NO. 52, 2003

â-Octafluorocorroles
A R T I C L E S
1
are too small to be taken at face value, but efforts are underway
to verify the overall spin density profile with B3LYP and
CASSCF calculations. Nevertheless, the fact that all the fluorine
spin densities are small44 and similar may be consistent with
the observation that all the fluorines in a given FeCl octafluo-
rocorrole derivative exhibit a narrow range of moderate isotropic
shifts.
metal compounds and of noninnocent ligands. Thus, H NMR
spectroscopy suggests that the new â-octafluorinated triaryl-
corroles act as substantially noninnocent ligands, i.e., exhibit
corrole π-cation radical character, in the FeCl complexes.
Quantitatively, however, NMR spectroscopy and DFT calcula-
tions indicate that the â-octafluorocorrole ligands are less
noninnocent than the corresponding â-unsubstituted triarylcor-
role ligands in their FeCl complexes. We have argued, albeit
less rigorously, that our electrochemical results also support this
picture.
Conclusions
The results presented in this paper lead to the following
conclusions:
(4) It is gratifying that the FeCl complexes exhibit broad but
distinctively discernible 19F NMR peaks. The peaks exhibit
significant but by no means enormous isotropic shifts which
may reflect that the highly electronegative fluorine atoms carry
only relatively small spin densities, i.e., relatively little of the
π-cation radical character of the corrole macrocycles delocalizes
onto the fluorines.
(5) The 19F NMR of the Cu complexes exhibit two relatively
broad peaks that sharpen at lower temperature, and we propose
that this is indicative of a thermally accessible Cu(II) corrole
π-cation radical excited state. There is thus a tendency toward
noninnocent corrole ligands for the copper complexes studied,
even though a diamagnetic Cu(III) ground state is the over-
whelmingly dominant species between -40 and 40 °C.
(6) We have previously reported that the electronic absorption
spectra, particularly the Soret absorption maxima, of high-valent
transition metal triarylcorroles are very sensitive to the nature
of the substituents at the meso positions. In contrast, the Soret
absorption maxima of free-base triarylcorroles are not particu-
larly sensitive to the nature of the meso substituents. The same
scenario holds for the fluorinated corroles described here. Thus,
while the three free-base fluorinated triarylcorroles exhibit
practically identical Soret absorption maxima, the Soret bands
of the Cu derivatives of the same corroles red-shift by
approximately 35 nm on going from the p-CF3 to the p-OCH3
derivatives. Theoretical calculations suggest that these red shifts
result from significant charge-transfer character in one or more
transitions in the Soret region.
(1) The one-pot corrole synthesis first reported by the Gross8,9
and Paolesse10,11 groups in the late 1990s has evolved into a
remarkably general and predictable self-assembly based syn-
thetic reaction. In our hands, Gross’s solvent-free procedure
proved particularly effective and, in fact, more general than
originally claimed. Thus, we showed that the reaction worked
for a variety of aromatic aldehyde starting materials17 and was
not limited to relatively electron-deficient aldehydes, as in the
original reports by Gross and co-workers.8,9 Here, we have
shown that the pyrrole component is also variable in that 3,4-
difluoropyrrole undergoes oxidative condensation with four
different aromatic aldehydes to yield â-octafluorocorroles.19 We
note that a variety of different substituted pyrroles have been
synthesized in recent years such as 3,4-dimethoxypyrrole45,46
and 3,4-bis(methylthio)pyrrole47 and these have also been
converted to the corresponding â-octasubstituted porphyrins.47-49
We speculate that one-pot corrole syntheses should also be
possible with a wide variety of different substituted pyrroles.
Given that this synthesis may be viewed a seven-component
self-assembly process, it is gratifying that it is robust with respect
to significant variations in the electronic character of both the
aldehyde and presumably also the pyrrole component.
(2) Electrochemistry and XPS have demonstrated the strongly
electron-deficient character of the â-octafluorocorrole ligands.
Thus, the XPS nitrogen 1s ionization potentials of these ligands
are some 0.7 eV higher than those of the corresponding
â-unfluorinated ligands. The oxidation half-wave potentials of
the Cu and FeCl complexes of the fluorinated corroles are also
positively shifted by 300-400 mV relative to their â-unsub-
stituted analogues.
Acknowledgment. This work was supported by the VISTA
program of Statoil (Norway) as well as by the Norwegian
Research Council. We thank Prof. F. Ann Walker for her
detailed comments on the manuscript, a number of constructive
suggestions related to the NMR studies, and a preprint of ref
21.
(3) Various measurements on the compounds reported here
throw considerable light on the nature of high-valent transition
(44) In an early DFT study (Ghosh, A. J. Am. Chem. Soc. 1995, 117, 4691) of
porphyrins, one of us noted that meso- and â-halogen substituents exert
quite different electronic effects. In particular, meso-halogen substituents
strongly stabilize porphyrin A2u radicals, presumably by absorbing some
spin density from the meso positions. A similar situation might be operative
for meso-fluorinated corroles, which are unknown so far.
(45) Merz, A.; Schwarz, R.; Schropp, R. AdV. Mater. 1992, 4, 409.
(46) Merz, A.; Meyer, T. Synthesis 1999, 94.
Supporting Information Available: DFT optimized Cartesian
coordinates and 1H NMR, 1H1H COSY, 19F NMR, and
resonance Raman spectra (PDF). This material is available free
of charge via the Internet at http://pubs.acs.org.
(47) Sugiura, K.-i.; Kumar, M. R.; Chandrashekar, T. K.; Sakata, Y. Chem. Lett.
1997, 291.
(48) Merz, A.; Schropp, R.; Lex, J. Angew. Chem., Int. Ed. Engl. 1993, 32,
293.
(49) Merz, A.; Schroppe, R.; Doetterl, E. Synthesis 1995, 7, 795.
JA021158H
9
J. AM. CHEM. SOC. VOL. 125, NO. 52, 2003 16309