High-resolution NMR spectroscopic trends and assignment rules of metal-free, metallated and substituted corroles

Article abstract of DOI:10.1002/mrc.1389

Major advances over the last few years have facilitated the synthesis of a large variety of meso-only substituted corroles that display interesting catalytic, therapeutic and photophysical properties. This work is the first to study extensively the NMR sp

Full text of DOI:10.1002/mrc.1389

MAGNETIC RESONANCE IN CHEMISTRY
Magn. Reson. Chem. 2004; 42: 624–635
Published online in Wiley InterScience (www.interscience.wiley.com). DOI: 10.1002/mrc.1389
High-resolution NMR spectroscopic trends and
assignment rules of metal-free, metallated and
substituted corroles
Yael S. Balazs, Irena Saltsman, Atif Mahammed, Elena Tkachenko, Galina Golubkov,
Joshua Levine† and Zeev Gross∗
Department of Chemistry and Institute of Catalysis Science and Technology, Technion—Israel Institute of Technology, Haifa 32000, Israel
Received 19 November 2003; Revised 19 February 2004; Accepted 20 February 2004
Major advances over the last few years have facilitated the synthesis of a large variety of meso-only
substituted corroles that display interesting catalytic, therapeutic and photophysical properties. This work
is the first to study extensively the NMR spectral characteristics of both metallated and non-metallated
triarylcorroles in various organic solvents and provide guidelines for easy and reliable assignments of
1
1D H spectra from trends of J coupling constants and chemical shifts. An excellent correlation is found
3
between C C bond lengths derived from J(H,H) values and experimental lengths determined by x-ray
crystallography of the same molecules. The nuclear Overhauser effect provides a robust 1D 1H NMR
tool for determining the selectivity of electrophilic substitutions. Variable-temperature NMR and isotopic
labelling reveal a single preferred tautomerization state and unsymmetric ring orientations at −70 ◦C.
The b-pyrrole protons demonstrate long-range heteronuclear couplings with the coordination core (15N)
and with the ortho-19F nuclei of the meso-carbon aryl rings. In sum, application of multinuclear magnetic
resonance to corroles and their metal complexes, through the compilation of chemical shifts and J couplings
and the recognition of trends therein, provides basic information essential to reliable spectral assignments.
Additionally, the conclusions drawn about the structures of corroles and the electron densities at various
positions of the corrole macrocycle resulting from the application of high-resolution NMR techniques are
of importance to an in-depth understanding of the molecular interactions and processes of this relatively
new and rapidly expanding class of compounds. Copyright  2004 John Wiley & Sons, Ltd.
KEYWORDS: NMR; 1H NMR; 13C NMR; 15N NMR; chemical shifts; coupling constants; triarylcorroles; metal complexes; ring
rotations; tautomerization.
tetraarylporphyrins are constantly utilized.5 In particular, the
metal complexes of 5,10,15-tris(pentafluorophenyl)corrole
(Scheme 1) were shown to be very efficient catalysts for
atom (oxygen) and group (carbene, nitrene) transfer to
organic substrates.6 In fact, for the latter reactions the corrole
metal complexes are significantly more efficient than anal-
ogous porphyrins.6c,e In addition, water-soluble derivatives
of 5,10,15-tris(pentafluorophenyl)corroles have almost unex-
plored potential for detection and selective damage to cancer
cells and novel photophysical properties of non-transition
metal corroles were recently disclosed.7,8
The two major structural peculiarities of corroles rel-
ative to porphyrins are the presence of three rather than
two NH protons in the coordination core and their lower
symmetry. Great emphasis was devoted to the first fea-
ture, particularly for the stabilization of metal ions in
high oxidation states,9 whereas the lower symmetry was
ignored. For example, although N-substituted corroles were
reported as early as 1965,10 the fact that these molecules
are chiral was not appreciated until recently.11 A dif-
ferent aspect of triarylcorrole peculiarities is the highly
INTRODUCTION
For decades, the chemistry of corroles was almost entirely
limited to derivatives with fully alkylated ˇ-pyrrole
positions,1 with only three examples of meso-only substi-
tuted corroles,2 defined as substitutions occurring at C-5,
C-10 or C-15 [Fig. 1(a)]. This situation changed dramatically
in 1999 with the disclosure of the first facile methodolo-
gies for the synthesis of 5,10,15-triarylcorroles from simple
aldehydes and pyrrole.3 About 100 new corroles substi-
tuted only at the three meso-carbon atoms have now been
reported4 with the consequent opportunities for extensive
investigations of corroles in the many applications for which
ŁCorrespondence to: Zeev Gross, Department of Chemistry and
Institute of Catalysis Science and Technology, Technion –Israel
Institute of Technology, Haifa 32000, Israel.
E-mail: chr10zg@techunix.technion.ac.il
†Summer student; present address: 121 Kings Point Road, Great
Neck, New York 11024, USA.
Contract/grant sponsor: Israel Science Foundation; Contract/grant
number: 368/00.
Contract/grant sponsor: Petroleum Research Fund; Contract/grant
number: PRF 36550-AC3.
Copyright  2004 John Wiley & Sons, Ltd.

High-resolution NMR trends of corroles 625
Scheme 1
19F atoms of the meso-carbon substituents, which reveal fur-
ther information about the structures of corroles and electron
densities at various positions of the macrocycle.
selective substitution of the ˇ-pyrrole protons, in contrast
to meso-CH substitution of ˇ-pyrrole-alkylated corroles,12
a formidable task for porphyrins and phthalocyanines.13
For example, even a bis-sulfonated phthalocyanine that
was separated from mono- and multisulfonated products
was shown to be a mixture of at least eight isomers.14
The situation for triarylcorroles would only be better if
the four different ˇ-pyrrole carbon atoms displayed sig-
nificantly different reactivities. Otherwise, the number of
possible products would be exceedingly large, up to 140
(Scheme 1).
Extensive NMR work has been done investigating the
tautomerization of porphyrins, which involves a double-
proton transfer between four nitrogen atoms, and has been
characterized in great detail in both the liquid and solid
states, where rate constants and kinetic isotope effects
are conserved.19 In corroles, the presence of three rather
than two NH protons in the coordination core and lower
symmetry leave open many questions concerning the spe-
cific details of the tautomerization mechanism. In the
molecular structures of the two metal-free triarylcorroles
that are reported in the literature, H3(tpfc) and H3(tpc)
[tpfc D tris(pentafluorophenyl)corrole and tpc D the tri-
anion of 5,10,15 triphenylcorrole],3b,20 unsymmetric ring
orientations are observed and the only nitrogen atom
that is not protonated is on the porphyrin-like side of
the macrocycle.21,22 We report here spectral characteristics
of variable-temperature solution NMR studies of 5,10,15-
tris(pentafluorophenyl)corroles, highlighting some of the
influences of tautomerization and aryl ring rotations.
In previous publications, we have shown that the periph-
eral carbon atoms of 5,10,15-tris(pentafluorophenyl)corrole
and its metal complexes can indeed be modified with high
selectivity. This includes the reduction of only one of its
double bonds8b and selective replacement of one to three of
its ˇ-pyrrole H atoms by either SO2Cl,15 SO3H,16 CHO17,18 or
NO2.17 A major task in these investigations was the structural
assignment of the products, as even a single substitution
may result in four different isomers. This appeared to be
a major challenge since the NMR spectra of corroles had
never seriously been analysed. What is more, a survey
of corrole-related publications reveals cases of vague and
even wrong spectral assignments.3,4,9c These shortcomings
initiated the present research, in which the spectra of 5,10,15-
tris(pentafluorophenyl)corrole and several of its diamagnetic
metal complexes were examined. Significant characteristics
in terms of chemical shifts and J coupling constants were
revealed and related to structural findings from x-ray crys-
tallography. In particular, we found that the resonance of
C21H appears deshielded and J coupling constants between
vicinal 1H atoms on pyrrole rings A and D are invariably
smaller than those on rings B and C (Fig. 1). Empirical eluci-
dation of these two simple facts results in a straightforward
assignment of all other hydrogen and carbon atoms in the
corresponding spectra, which is a great aid for determin-
ing the structures of new corroles and a prerequisite for
analysing the spectra of paramagnetic metal complexes. We
also present long-range spin–spin interactions of ˇ-pyrrole-
1H with the 15N atoms of the coordination core and specific
EXPERIMENTAL
Chemicals and syntheses
All reagents were purchased from commercial sources and
used as received unless noted otherwise. Except for isotope-
labelled compounds and complexes, the synthetic details
of the preparation of 5,10,15-tris(pentafluorophenyl)corrole
(H3(tpfc)) and all the metal complexes used here have been
reported previously.3,6,8
Preparation of 21,22,23,24-tetra-15N-5,10,15-
tris(pentafluorophenyl)corrole, 1-15N4
The 15N-labelled corrole was prepared via the standard procedure
for H3(tpfc),3 using 15N-pyrrole (Cambridge Isotopes Laborato-
ries, 98%C 15N, 0.5 g, 7.3 mmol), 2,3,4,5,6-pentafluorobenzaldehyde
(1.44 g, 7.3 mmol) and alumina (1.5 g). MS ꢀDCIꢀꢁ: m/z 799.9.
Copyright  2004 John Wiley & Sons, Ltd.
Magn. Reson. Chem. 2004; 42: 624–635

626
Y. S. Balazs et al.
(a)
(b)
C6F5
Hd
100
Hd
8
13C
12
9
110
11
10
Hc
7
13 Hc
B
C
120
6
N
N
14
130
13C5
13
15 C
16
Ga
C6F5
C6F5
140
N
4
3
N
A
150
D
17
1
19
Hb
Hb
2
18
δCppm
Ha
Ha
δH ppm
9.5 9.0 8.5 8.0
(c)
(d)
1Ha
1Hc 1Hb 1Hd
1Ha
13
13
90
C
C
10
100
110
120
130
5,15
1Hb
1Hd
1Hc
13
C
a-d
13C`
140
2
2
2
2
9.2 9.1 9.0 8.9 8.8 8.7 8.6
δHppm
δCppm
δHppm
9.2 9.1 9.0 8.9 8.8 8.7 8.6
Figure 1. The 5,10,15-tris-13C-labelled gallium(III) corrole complex: (a) the formal numbering of carbon atoms and pyrrole rings;
symmetric ˇ-pyrrolic 1H positions have been assigned letters Ha –Hd; (b) 2D 1Hf13Cg one-bond correlation spectrum (HMQC); (c) 2D
1Hf13Cg multiple-bond correlation spectrum (HMBC); (d) the corresponding 1D 1H NMR spectrum with each doublet assigned.
Preparation of 5,10,15-tris-13C-5,10,15-
under reduced pressure provided practically pure (by NMR)
2,3,4,5,6-pentafluorobenzaldehyde-carbonyl-13C (0.61 g, 3.1 mmol).
tris(pentafluorophenyl)corrole (1-13C3)
19F NMR (CDCl3):
υ
D ꢀ143.44 (m, 1F; p-F), ꢀ144.5 (m,
2F; o-F), ꢀ160.37 (m, 2F; m-F); 1H NMR (CDCl3): υ D 10.24
and its metal complexes
Preparation of 2,3,4,5,6-pentafluorobenzoyl-carboxy-13C chloride
2,3,4,5,6-Pentafluorobenzoic-carboxy-13C acid (2 g, 9.4 mmol) was
dissolved in thionyl chloride (7 ml) and the solution was heated
at reflux for 24 h. Evaporation of the excess thionyl chloride
provided a colourless oil (1.73 g, 7.5 mmol) that was found by
NMR spectroscopy to be almost pure. The crude product was used
for the next step without further purification. 19F NMR ꢀCDCl3ꢁ:
υ D ꢀ135.75 (m, 2F; o-F), ꢀ143.37 (m, 1F; p-F), ꢀ159.06 (m, 2F; m-F);
IR (film), ꢂ (cmꢀ1) 1800 ꢀC Oꢁ.
[d, JꢀC,Hꢁ
ꢀC Oꢁ.
D
192.3 Hz, 1H; CHO]; IR (film) ꢂ ) 1720
(cmꢀ1
Preparation of
5,10,15-tris-13C-5,10,15-tris(pentafluorophenyl)corrole
The standard procedure for H3(tpfc) was applied, starting with pyr-
role (0.5 g, 7.3 mmol), 2,3,4,5,6-pentafluorobenzaldehyde-carbonyl-
13C (1.44 g, 7.3 mmol) and alumina (1.5 g). The resulting corrole
(0.19 g, 0.2 mmol) was examined by 13C NMR and estimated to
contain at least 97% enrichment of meso-13C carbon atoms. The
corresponding gallium(III) and (triphenylphosphine)cobalt(III) com-
plexes were prepared in analogy with the published procedures for
the non-labelled corrole.22
Preparation of 2,3,4,5,6-pentafluorobenzyl-methylene-13C alcohol
2,3,4,5,6-Pentafluorobenzoyl-carboxy-13C chloride (1.73 g, 7.5 mmol)
and THF (6 ml) were placed in a flask fitted with a stirrer, separating
funnel, low-temperature thermometer and nitrogen inlet and outlet.
°
NMR measurements
The flask was cooled to ca ꢀ75 C and a THF solution of lithium tri-
Sample concentrations ranged from 8 to 13.5 ꢀmg mlꢀ1ꢁ.
Chemical shift referencing of the 1H spectra was relative
to the residual 1H signal of the deuterated solvent, tak-
ing chloroform D 7.24 ppm, benzene D 7.15 ppm, toluene
ꢀC1HD2ꢁ D 2.09 ppm, and acetonitrile D 1.93 ppm. Calibra-
tion of 13C spectra was also based on the solvent signal, with
CDCl3 D 77.0 ppm and C6D6 D 128.0 ppm. In the case of
19F chemical shifts, solvent-based automatic referencing was
used (relative to the lock signal). The 15N chemical shifts
were referenced using the IUPAC /% value24 relative to
the residual solvent 1H signal, calibrated as given.
tert-butoxyaluminohydride (15 mmol in 25 ml of THF, prepared as
described in the literature,23 method D) was added over a period of
1 h, avoiding any major rise of temperature. After stirring for 2 h at
°
ꢀ75 C, the cooling bath was removed and the contents were poured
into crushed ice. THF was evaporated from the reaction mixture
and several extractions with chloroform provided practically pure
2,3,4,5,6-pentafluorobenzyl-methylene-13C alcohol (0.89 g, 4.5 mmol).
19F NMR (CDCl3): υ D ꢀ144.74 [dd, 3JꢀF,Fꢁ D 21 Hz, 4JꢀF,Fꢁ D
5.4 Hz, 2F; o-F], ꢀ154.40 [t, 3JꢀF,Fꢁ D 19.6 Hz, 1F; p-F], ꢀ162.14 [td,
3JꢀF,Fꢁ D 20.7 Hz, 4JꢀF,Fꢁ D 7.6 Hz, 2F; m-F]; 1H NMR (CDCl3):
υ D 4.76 [dd, JꢀC,Hꢁ D 147.9 Hz, JꢀH,Hꢁ D 4.7 Hz, 2H; ꢀCH2ꢀ], 2.01
(br s, 1H; OH); IR (film), ꢂ (cmꢀ1) 3400 (OH).
Preparation of 2,3,4,5,6-pentafluorobenzaldehyde-carbonyl-13C
A
solution of 2,3,4,5,6-pentafluorobenzyl-methylene-13
C alcohol
The data presented in Table 1 were collected at room tem-
perature on a Bruker Avance 600 solution NMR spectrometer
with a TXI probe equipped with z-gradients. The 1D 1H
spectrum was acquired with an 8.4 kHz spectral width and
32K real points and processed without zero-filling or win-
(0.89 g, 4.5 mmol, 5 ml CH2Cl2) was added in one portion
(slightly exothermic reaction) to a reaction vessel, loaded with
˚
molecular sieves (4 A, 4.8 g), pyridinium chlorochromate (4.85 g,
°
22.5 mmol) and CH2Cl2 (35 ml) and first stirred at 0 C under an
inert atmosphere for 30 min. The reaction mixture was stirred
for 30 min at room temperature and chromium salts were
removed by filtration under a nitrogen atmosphere through a
10 cm layer of silica gel. Evaporation of the colourless filtrate
13
dow functions. 1Hf Cg HMQC25 spectra were acquired with
pulsed field gradients (PFGs): 4 transients, 32 dummy scans,
Copyright  2004 John Wiley & Sons, Ltd.
Magn. Reson. Chem. 2004; 42: 624–635

High-resolution NMR trends of corroles 627
Table 1. The unsubstituted gallium corrole complexa: 1H and 13C chemical shifts (ppm)
Ha (2,18) D 9.20 (d, 2H, 3.9 Hz) Hb (3, 7) D 8.75 (d, 2H, 3.9 Hz) Hc (7, 13) D 8.89 (d, 2H, 4.4 Hz) Hd (8,12) D 8.67(d, 2H, 4.4 Hz)
C-2 D C-18 D 118.6
C-3 D C-17 D 127.9
C-7 D C-13 D 125.8
C-8 D C12 D 124.5
C-10 D 92.8 (s, 1H)
C(1,19 or 4,16) D 136.7, 138.2
C(6,14 or 9,12) D 141.2, 143.8
C-5 D C-15 D 98.5 (s, 2H)
a Ga(tpfc)(py) in C6D6.
1024 ð 128 acquisition points and garp decoupling. Spectral
achieved with multiple low-power, on-resonance pulses
(50 ms each) lasting a total of 3.5 s. On- and off-resonance
excitation was interleaved: 8 transients per cycle for a total
of 16 cycles. Spectra were processed using 1 Hz of line
broadening and no zero filling. The processing and manipu-
lation of the difference spectra were done on a PC (Compaq,
Pentium IV, OS Windows 2000) using the freeware program
Spinworks, version 1.3 (K. Marat, University of Manitoba,
ftp://davinci.chem.umanitoba.ca/pub/marat/
widths were 4.195 kHz (1H) and 12.071 kHz (13C). Multiple-
13
bond 1Hf Cg correlations were acquired with PFG HMBC,26
a low-pass J-filter and no decoupling during acquisition:
8 transients, 32 dummy scans and 1024 ð 128 acquisition
points. Spectral widths were 4.195 kHz (1H) and 10.563 kHz
(13C). Magnitude mode Fourier transforms were performed
with zero filling to 2048 ð 512 points and application of a
sine window function, in both cases.
The 1H chemical shifts and J couplings presented in
Tables 2 and 3 were measured at room temperature on
a Bruker Avance 500 spectrometer, unless noted other-
wise. Standard acquisition parameters were spectral width
6500 Hz, 16K real points, 32 transients and 2 dummy
scans. Standard processing was with zero filling to 32K
real points and Fourier transformation without applied
window functions. Statistical graphs were generated using
Origin 6.1 (Origin Laboratories, Northampton, MA, USA;
http://www.Originlab.com).
1D NOE difference spectra27 were acquired on a Bruker
AC 200 instrument with a 5 mm double-frequency probe
with 2H lock. Acquisition was with spectral widths of
2300 Hz using 4K real points. Selective 1H saturation was
SpinWorks/).
1H spectra with 19F inverse-gated decoupling (Waltz-16)
were acquired on a Bruker Avance 300 spectrometer with a
1
2
5 mm H/19F/31P/13C QNP probe and H lock. The spectra
shown were acquired with spectral widths of 330 Hz, 4K
real points, 4 transients and 2 dummy scans then Fourier
transformed after zero-filling to 16K real points. Larger
spectral widths were used for chemical shift referencing.
Variable-temperature studies on the Bruker Avance
500 equipped with a bbo probe and 2H lock used a
BVT3000 Eurotherm unit (10 kHz spectral widths and 32K
real points). Cooling was achieved with the boil-off gas
from liquid nitrogen, or on a Bruker Avance 300 with a
1H/19F/31P/13C QNP probe and 2H lock (spectral widths
Table 2. Calculated vs. experimental coupling constants, 3J(H,H)
3J(H,H) (Hz)
Experimental

Complex
H3(tpfc)b
C—C bond
Rꢃꢄ (nm)
Calculateda
3.98
Jcalc ꢀ Jexpt
C2—C3
C17—C18
C7—C8
C12—C13
C2—C3
C17—C18
C7—C8
C12—C13
C2—C3
C17—C18
C7—C8
C12—C13
C2—C3
C17—C18
C7—C8
C12—C13
C2—C3
0.1383
0.1374
0.1359
0.1343
0.1383
0.1374
0.1359
0.1343
0.1376
0.1383
0.1358
0.1358
0.1371
0.1368
0.1357
0.1354
0.1377
0.1362
0.1353
0.1356
4.0
4.5
4.3
4.7
4.1
4.5
4.4
4.6
4.6
4.8
0.0
C0.4
ꢀ0.3
C0.2
C0.2
C0.1
ꢀ0.1
C0.1
ꢀ0.3
ꢀ0.1
4.87
3.98
4.87
3.95
4.64
4.27
4.72
4.27
4.75
H3(tpfc)
Ga(tpfc)(Py)
Rh(tpfc)(PPh3)
Fe(tpfc)(NO)b
C17—C18
C7—C8
C12—C13
a 3JꢀH,Hꢁ D ꢀ322.6Rꢃꢄ C 48.45 (Rꢃꢄ D average x-ray structure bond length).28
b Dissolved in C6D6 (all others dissolved in CDCl3).
Copyright  2004 John Wiley & Sons, Ltd.
Magn. Reson. Chem. 2004; 42: 624–635

628
Y. S. Balazs et al.
Table 3. 1H NMR chemical shifts (ppm) and 3J(H,H) couplings (Hz) of unsubstituted metal-free corrolesa
Corrole
Solvent
1Ha (2, 18)
1Hc (7, 13)
1Hb (3, 17)
1Hd (8, 12)
3Jab
3Jcd
H3(tpfc)b
H3(tpfc)c
H3(tpfc)
H3(tdClc)
H3(tdClc)
H3(tpc)
Toluene-d8
C6D6
CDCl3
CDCl3
C6D6
8.72
8.74
9.10
8.81
8.53
8.53
8.56
8.75
8.48
8.51
8.35
8.37
8.27
8.26
4.3
4.0
4.3
4.4
4.2
4.7
4.5
4.7
4.8
4.8
8.57 (2d overlapping, 4H)
8.33 (2d overlapping, 4H)
8.40
8.52
8.41
8.36 š 0.05
8.20
8.55
8.36
CDCl3
Toluene-d8
8.90
8.66
8.85
8.56
4.3
4.2
4.6
4.8
H3(tdFc)d
Metal-free mean valuese
8.78 š 0.07
8.61 š 0.05
8.42 š 0.04
4.24 š 0.05
4.70 š 0.05
a All peaks correspond to (d, 2H), unless specified otherwise.
b
°
Splittings become resolved at 300 MHz and 55 C.
c At 300 MHz and room temperature there is no resolution of the 1Hb.
d
°
Splittings become resolved at 300 MHz and 70 C.
e Also see Fig. 3(a).
of 9000 Hz and 4K real points). Data were zero-filled with
2–4ð real points and Fourier transformed (without window
functions). Temperatures were regulated to within 0.1 K.
High temperatures were calibrated with ethylene glycol and
low temperatures with methanol.28
from the symmetry-enforced equivalence of C-5 and C-15.
1
13
To assign the remaining NMR chemical shifts, 2D Hf Cg
HMQC [Fig. 1(b)] and HMBC [Fig. 1(c)], were performed.
The geometry and aromaticity of the gallium(III) corrole
complex were conducive to observation of three- and even
four-bond 1H, 13C couplings in the HMBC spectrum. A
high signal-to-noise ratio of the cross peaks was observed
Total correlation spectroscopy (2D TOCSY29) was per-
formed on a Bruker Avance 500 with /2ꢅB1 D 21.9 kHz,
2.5 ms trim pulses and a 75 ms mixing period (mlev-
17). States-TPPI mode was used for acquisition with 16
dummy scans, 4 transients, 1 kHz spectral widths in both
dimensions, 1024 ð 64 acquisition points and 1024 ð 1024
1
between the ˇ-pyrrolic Hs and the selectively 13C-labelled
meso carbon atoms due to the isotope enrichment.
3
1H, 1H coupling constants: 3Jab < Jcd
15
processed points applying a sine window function. 1Hf Ng
Experimental average bond lengths determined from x-
ray crystallographic measurements are compared with the
experimental 3J(H,H) coupling constants from 1H NMR
spectroscopy on the same compounds (Table 2) to test the
appropriateness of applying a standard textbook equation for
estimating 3J(H,H) values in five-membered rings to corroles.
A linear relationship is indeed observable between the
magnitude of the C C bond length and the 3J(H,H) coupling
constant [3JꢀH,Hꢁ D ꢀ322.6Rꢃꢂ C 48.45, where Rꢃꢂ is the
carbon–carbon bond length28]. In other words, the inherently
longer HaC CHb bond lengths between carbon atoms
correspond to the smaller 1H–1H splittings, while the larger
3J(H,H) values arise from the shorter HcC CHd bonds.
In metal-free corroles, the two scalar coupling constants
have mean values (šstandard error) of 3Jab D 4.24 š 0.05 Hz
HMQC spectra were collected on the same spectrometer:
16 dummy scans, 1 transient, 2048 ð 128 acquisition points
and garp decoupling. Spectral widths were 10.000 kHz (1H)
and 7.605 kHz (15N). A magnitude mode fourier transform
was performed after processing to 1024 ð 512 points with
application of a sine window function. The 1D 15N spec-
trum shown was acquired with a spectral width of 7.576 kHz
and 8K real points. Data were processed with zero-filling
to 32K real points, 5 Hz line broadening and exponential
multiplication.
RESULTS AND DISCUSSION
Total spectral assignment
3
Lower-case letters a–d were sequentially assigned to the ˇ-
and Jcd D 4.70 š 0.05 Hz, as summarized in Table 3. In
1
1
pyrrolic Hs and correlated with the 1D H NMR spectrum
of a gallium(III) corrole with three site-specific 13C enriched
quaternary carbon atoms at the meso-positions [C-5, C-10,
C-15; Figure 1(a) and (d)]. Owing to the C2v symmetry, the
eight ˇ-pyrrolic 1Hs (¾8–9 ppm) appear as four doublets of
equal peak area. Notably, the doublet of 1Ha is narrower
than the other resonance peaks, as discussed separately
Table 4, the J-couplings of 13 diamagnetic metal corrole
3
complexes are summarized. The mean values are Jab
D
4.2 š 0.1 Hz and 3Jcd D 4.6 š 0.1 Hz, excluding the binuclear
Ru complex whose 1Hc and 1Hd resonances are broad. In both
tables 3Jab <3 Jcd in all cases, except Fe(tpfc)NO where 3Jab
D
3Jcd. There is a smaller overall range of 3J values measured for
metal-free corroles compared with the metallated complexes,
probably due to the difference in sample sizes: n D 7
(Table 3) compared with n D 12–13 (Table 4). Furthermore,
1
under dynamic processes. The through-bond H–1H scalar
interactions between 1Ha and 1Hb (3Jab) and between 1Hc and
1Hd (3Jcd) are not of equal magnitude, as discussed under 1H,
1H coupling constants below.
C
C bond lengths, and correspondingly 1H, 1H coupling
constants, are expected to be more sensitive to the identity of
the metal ion than to peripheral substitutions.
1
Table 1 summarizes the observed H and 13C chemical
shifts. Assignment of the 13C chemical shifts was straight-
forward, based on the observed 2 : 1 peak ratios expected
On average, Jab is smaller than Jcd by 0.41 š 0.08 Hz.
3
3
Accurately measuring such small coupling constants in a 1D
Copyright  2004 John Wiley & Sons, Ltd.
Magn. Reson. Chem. 2004; 42: 624–635

High-resolution NMR trends of corroles 629
Table 4. 1H NMR chemical shifts (ppm) and 3J(H,H) couplings (Hz) of unsubstituted metallated corroles
X
Axial ligands
Solvent
1Ha (2, 18) 1Hc (7, 13)
1Hb (3, 17) 1Hd (8, 12)
3Jab
3Jcd
Co
PPh3
PPh3
ꢀPyꢁ2
Py
C6D6
CDCl3
C6D6
C6D6
CDCl3
C6D6
C6D6
C6D6
C6D6
8.66
8.68
9.32
9.20
9.15
9.21
9.07
8.94
7.78
8.78
9.23
8.85
8.62
8.42
8.32
9.06
8.89
8.83
8.92
8.86
8.78
7.40
8.40
8.85
8.75
8.28
8.07
8.83
8.75
8.70
8.78
8.67
8.52
7.45
8.23
8.78
8.59
8.08
8.22
8.87
8.67
8.61
8.70
8.66
8.59
7.31
8.13
8.68
8.51
7.96
4.6
4.5
4.3
3.9
4.0
4.1
4.3
4.2
4.7
4.4
4.1
3.8
4.2
5.0
4.7
4.8
4.4
4.5
4.6
4.7
4.7
Coa
Coa
Ga
Ga
Al
Py
ꢀPyꢁ2
(OH)
Cl
Ge
Sna
Feb,c
NO
4.7
4.6
4.6
Rh
Mn
Pd
PPh3
ꢁ Ne
ꢀOHꢁ2
None
CDCl3
CD3CN
C6D6
4.0
[(tpfc)Ru]2
Metallated mean valuesf
Toluene-d8
8.19 (broad) 7.94 (broad)
8.6 š 0.1 8.4 š 0.1
Not resolved
4.6 š 0.1g
8.8 š 0.1
8.4 š 0.1
4.2 š 0.1
a The relative shielding of 1Hb (3, 17) (ppm) and 1Hd (8, 12) (ppm) is switched ꢀ1Hd >1Hbꢁ.
b The relative shielding of 1Hc (7, 13) (ppm) and 1Hb (3, 17) (ppm) is switched ꢀ1Hb >1Hcꢁ.
3
c Note that 3JabꢀHzꢁ D Jcd (Hz), coupled pairs may still be resolved with 2D 1H–1H COSY.
d Extra splittings are present due to 31P–1H scalar interactions.
e Nitrido axial ligand.
f Also see Fig. 3(b).
g Excluding [(tpfc)Ru]2 owing to unresolved 1Hc and 1Hd resonances.
1H NMR spectrum is feasible. Where instrumental resolution
is problematic, a 2D J-resolved spectrum can be used. These
general, yet robust, trends of 3J(H,H) values provide a useful
tool for direct assignments of 1Ha –1Hb and 1Hc –1Hd pairs in
a 1D H NMR spectrum, ꢀ3Jab < Jcdꢁ. Further assignments
of 1H NMR resonances follow readily upon identification
of 1Ha as the least shielded (left-most) resonance at room
temperature, as discussed under 1H chemical shifts.
[Fig. 2(a)]. There is some frequency preference for 1Hd > 1Hb
as seen by the mean resonance values, but the Hb position
1
is variable. The mean chemical shifts (šstandard error) from
Fig. 2(a) are 1Ha (H-2,H-18) D 8.78 š 0.07, 1Hc (H-7,H-13)
D 8.61 š 0.05, 1Hb (H-3,H-17) D 8.36 š 0.05, 1Hd (H-8,H-
1
3
1
12) D 8.42 š 0.04. Two examples of 1D H NMR spectra of
metal-free corroles are presented in Fig. 3.
1
In the metallated corroles, Ha resonances are still well
Checking the experimental 3J(H,H) coupling constants
for solvent effects on scalar spin-spin couplings gives variable
results. For example, Ga(tpfc)(py) shows no appreciable
differences in 3J(H,H) between CDCl3 and C6D6, whereas
Co(tpfc)(PPh3) shows a (relatively) dramatic solvent effect
between benzene-d6 and chloroform-d [3J(H,H) is reduced
by half]. This apparently reflects a greater tendency for
association of solvent molecules to the latter complex,
reflected also by the fact the two complexes crystallize with
and without additional solvent molecules, respectively.8a,4h
separated from the other peaks [Fig. 2(b)]. The degeneracy of
1Hb and 1Hd chemical shift values is apparent from Fig. 2(b).
Relative peak positions can be assigned on the basis of
3J(H,H) coupling with Ha or Hc, respectively. The 1D H
NMR spectrum of Fig. 1(c), where 1Ha > 1Hc > 1Hb > 1Hd, is
typical of the general trend for metallated corroles. The mean
chemical shifts from the box plot of Fig. 2(b) are 1Ha (H-2,H-
1
1
1
1
1
18) D 8.8 š 0.1, Hc (H-7,H-13) D 8.6 š 0.1, Hb (H-3,H-17)
D 8.4 š 0.1, 1Hd (H-8,H-12) D 8.4 š 0.1.
The effects of solvent on chemical shifts vary (Tables 3
1
and 4). For example, the H chemical shifts of Ga(tpfc)(py)
1H chemical shifts
are the same in either CDCl3 or C6D6, whereas Co(tpfc)(PPh3)
shows a (relatively) dramatic solvent effect on going from
benzene (typical chemical shifts) to chloroform, where the
relative positions of 1Hb and 1Hd are exchanged. As already
proposed in the previous section, a most likely explanation is
that association of solvent molecules to the cobalt ion affects
the distribution of electron density in the macrocycle.
Trends of 1H NMR peak positions are illustrated by box plots
of chemical shifts assigned to the resonances of 1Ha –1Hd for
four different metal-free corroles and 11 different metallated
corroles (Fig. 2). The left-most peak is assigned to 1Ha
[consistently smaller 3J(H,H) value]. As seen in Fig. 2, the
1
chemical shift of peak Ha appears well separated from the
other ˇ-pyrrolic 1H peaks. The splitting of 1Ha by 1Hb leads
to a direct assignment of the latter. The peak immediately to
the right of 1Ha is assigned to 1Hc. The remaining peak 1Hd
can be identified by its through-bond coupling to 1Hc (3Jcd is
always larger than 3Jab).
In the metal-free corroles, the chemical shifts of the
ˇ-pyrrolic 1Hs are affected by the aryl groups. This is
illustrated in Fig. 3: the 1H chemical shifts of the meso-
C6F5-substituted H3(tpfc) are more dispersed than those
of the meso-C6F2-substituted H3(tdfc). Shielding effects due
to the more electronegative aryl groups are smallest on
the most distant ˇ-pyrrolic proton: 1Ha (υ D 0.03 ppm),
In the metal-free corroles the chemical shift trend is
that 1Ha
>
1Hc > 1Hb
³
1Hd, with fairly good resolution
Copyright  2004 John Wiley & Sons, Ltd.
Magn. Reson. Chem. 2004; 42: 624–635

630
Y. S. Balazs et al.
(a)
(b)
1Hd
1Hb
1Hc
1Ha
1Hd
1Hb
1Hc
1Ha
9.25 9.00 8.75 8.50 8.25 8.00
H (ppm)
10.0 9.5 9.0 8.5 8.0 7.5 7.0
H (ppm)
δ
δ
Figure 2. 1H NMR experimental chemical shift distributions. Statistical box plots of (a) unsubstituted metal-free
tris(pentafluorophenyl)corroles and (b) unsubstituted metallated corrole complexes. The percentile distributions of the 1H chemical
shifts are marked as follows: ð D 1 and 99%,
D 5 and 95%, large box D 25 and 75%, box mid-line D 50% and the mean is
indicated by the small box.
1Ha
1Hc
1Hd
1Hb
(d)
(c)
(b)
(a)
(b)
8.8
8.8
8.7
8.6
8.6
8.5
8.5
8.4
8.3
8.3
δ
H (ppm)
*
*
1HC=O
(a)
10.4 10.2 10.0 9.8 9.6 9.4 9.2 9.0 8.8 8.6
8.7
8.4
δH (ppm)
δH (ppm)
Figure 3. Influence of 19F spins on ˇ-pyrrolic 1H resonance
peaks. 1H NMR spectra of (a) meso-C6H3F2-substituted
H3(tdfc) and (b) meso- C6F5-substituted H3(tpfc) in toluene-d8.
Spectra were taken at a 1H resonance frequency of
Figure 4. 1H NMR 1D and NOE difference spectra of a
monoformylated gallium(III) corrole complex. (a) The 500 MHz
spectrum resolves the unique chemical shifts of each of the six
ˇ-pyrrolic 1H doublets. (b) The 200 MHz spectrum shows
some overlap between doublets due to smaller chemical shift
resolution at lower field. (c) The 200 MHz 1D NOE difference
spectrum with selective irradiation at the ˇ-pyrrolic 1H singlet
(9.66 ppm). (d) An internal control NOE difference spectrum
with irradiation at the formyl proton only gives an NOE
enhancement at the ˇ-pyrrolic singlet.
3
°
300.13 MHz and a temperature of 74 C. Jab D 4.3 Hz,
3Jcd D 4.8 Hz. The additional splittings in 1Hc and 1Hd (¾1 Hz)
arise from 19F–1H interactions as described in the text and
demonstrated in Fig. 8.
1Hb (υ D ꢀ0.11 ppm), 1Hc (υ D ꢀ0.05), and 1Hd
(υ D ꢀ0.08 ppm). The spectra in Fig. 3 were recorded
°
at elevated temperatures (74 C at 300 MHz), the optimal
macrocycle. First, synthetically, substitutions are more likely
to occur at position 3,17, the site of highest electron density
where the excess negative charge behaves as a nucleophile
most prone to carbocation attack.17 Second, structurally, the
proximity between 1H2 and 1H18 is sufficiently close to give
rise to a detectable through-space spin–spin interaction,
observed as a signal enhancement via the nuclear Overhauser
effect.27
conditions for well-resolved spectra (see below).
Substituted corroles
The symmetry reduction on substitution of a singleˇ-pyrrolic
hydrogen atom of corroles by another functional group
results in an NMR spectrum consisting of seven unique
1
peaks. This is exemplified in the 1D H NMR spectrum of
a monoformylated gallium corrole complex in chloroform-d
[Fig. 4(a)], which consists of two singlets [1HC O (1H) and a
ˇ-pyrrolic 1H (1H)] and six doublets (1H each). The chemical
shift resolution of the six ˇ-pyrrolic 1H doublets depends on
the magnetic field strength [Fig. 4(b)], solvent effects and the
extent of electronic perturbation by the substituent on the
distant ˇ-pyrrolic hydrogens. In most cases of overlapping
resonance lines a 2D COSY resolved the coupled 1H pairs.
The task of identifying a substitution site benefits from
two inherent properties of the corrole-like portion of the
In the 1D NMR spectrum of the monoformylated
gallium corrole complex (Fig. 4), the multiplicity state of
1
the ˇ-pyrrolic H singlet is immediately identifiable as the
macrocyclic 1H nearest the substitution site, due to the
loss of 3J(H,H) splittings. In Fig. 4(c), the NOE difference
spectrum after selective irradiation at the frequency of the
ˇ-pyrrolic singlet shows a large negative signal at the point
of irradiation and two small positive peaks at 10.50 and
9.11 ppm due to through-space NOE signal enhancements.
Copyright  2004 John Wiley & Sons, Ltd.
Magn. Reson. Chem. 2004; 42: 624–635

High-resolution NMR trends of corroles 631
Table 5. 1H chemical shifts (ppm) and 3J(H,H) couplings (Hz) of substituted corroles
X
Ligand
Py
Substitution
NO2
Solvent
CDCl3
1H-2
1H-3a
—
1H-17
1H-18
1Hc(7, 13)
1Hd(8, 12)
Gab
9.67 (s)
8.42 (4.4, d)
9.03 (4.4, d)
8.71 (4.7, d)
8.63 (4.7,d)
8.47 (4.7,d)
8.38 (4.7,d)
Gab
Ga
Py
Py
NO2
CHO
C6D6
CDCl3
CDCl3
9.59 (s)
9.66 (s)
9.94 (s)
—
—
—
8.49 (4.1,d)
8.73 (4.1,d)
8.93 (s)
8.76 (4.1,d)
9.11 (4.1,d)
—
8.67 (4.5,d)
8.64 (4.7,d)
8.76 (4.5,d)
8.67 (4.8,d)
8.60 (4.7,d)
8.55 (4.7,d)
8.61 (4.6,d)
8.34 (5.0,d)
8.42 (5.0,d)
8.22 (5.0,d)
8.48 (4.5,d)
8.52 (4.8,d)
8.31 (m)
Ga
ꢀPyꢁ2
ꢀCHOꢁ2 asym
Gab
Cob
Cob
ꢀPyꢁ2
PPh3
PPh3
ꢀNO2ꢁ2 sym
ꢀSO2piꢁ2 symc
ꢀSO2piꢁ2 asymc
CDCl3
C6D6
C6D6
9.55 (s)
9.06 (s)
9.68 (s)
—
—
—
—
—
8.72 (s)
9.55 (s)
9.06 (s)
—
8.37 (4.6,d)
8.21 (5.0,d)
8.29 (5.0,d)
8.20 (5.0,d)
a Substitution site in all cases.
b Compounds with an x-ray structure where the substitution site is verified.
c pi D Piperidine
3
The 9.11 ppm doublet has a relatively small J(H,H) value
(4.1 Hz; Table 5), consistent with an assignment as either
1Ha or 1Hb. Since only position 1Ha can satisfy the NOE
(g)
(f)
condition of a doublet close in space to the irradiated singlet
peak [Fig. 1(c)], the site of substitution must be at position
(e)
C-3. As an internal control, selective irradiation of the formyl
1
1
group H frequency results in an NOE enhancement at H-
2 [Fig. 4(d)]. The monoformylated gallium corrole complex
described above did not readily crystallize to x-ray quality
crystals and the assignment of the substitution site could
only be elucidated using solution-state NMR spectroscopy.
We can assign the remaining peaks from the 3Jab splitting
of 1H-18 due to an interaction with 1H-17. The remaining
doublets must be due to 1Hc and 1Hd. According to the
observed trends in chemical shift distributions, 1Hc is
assigned to the least shielded of the remaining peaks and is J-
coupled to 1Hd. The remaining 1Hc –1Hd pair can be similarly
assigned. There are seven different substituted corroles listed
in Table 5, of which five yielded crystals for x-ray diffraction
(d)
(c)
(b)
(a)
0
-1
-2
-3
-4
-5
-6
δ
H (ppm)
Figure 5. Temperature dependence of NH tautomerization in
H3(tpfc) at 500.13 MHz 1H frequency. (a) Natural abundance
1
structures. In all cases the solution-state 1D H NMR NOE
spectra revealed the same substitution site provided by the
crystallographic work. Table 6 provides the corresponding
13C chemical shift assignments of the substituted corroles
derived from 2D HMQC experiments.
metal-free tris(pentafluorophenyl)corrole in toluene-d8 at
N
ꢀ72 C with three distinct 1H resonances of equal peak area.
°
(b) Uniformly-15N-labelled compound at ꢀ69 C, splitting the
°
1HN peaks into doublets. The remaining spectra of
[U-15N]-H3(tpfc) were taken at (c) ꢀ35, (d) ꢀ12 and (e) 0.1 C
°
Assignment of low-temperature H3(tpfc) spectra
The1HN spectralregionofthemetal-freetris(pentafluorophe-
nyl)corrole in toluene-d8 is shown in Fig. 5. At low tem-
°
and at 300.13 MHz at (f) C22 and (g) C74 C.
of this 1H between two different nitrogens of the macrocycle,
one strong and one weak (15N—1H Ð Ð Ð15 N).
°
peratures (ꢀ72 C) and 500.13 MHz resonance frequency,
a single preferred tautomeric state of the NH hydrogens
is evident from three distinct 1HN shifts of equal peak
area [Fig. 5(a)]. With uniform 15N enrichment it is possible
to resolve the 1H, 15N through-bond couplings [Fig. 5(b)].
The two most shielded peaks are each split into dou-
blets by the 15N label: υH D ꢀ6.25 ppm, 1JꢀN,Hꢁ D 94 Hz;
υH D ꢀ6.34 ppm 1JꢀN,Hꢁ D 95 Hz. Most interesting is the
least shielded of the 1HN peaks, which is split into a doublet
of doublets (dd), υH D ꢀ0.09 ppm 1JꢀN,Hꢁ D 70 and 19 Hz.
Such a dd pattern could be accounted for by unequal sharing
The ˇ-pyrrolic 1H region of the metal-free tris(pentafluo-
rophenyl)corrole in toluene-d8 is shown in Fig. 6. At low
°
temperature (ꢀ77 C) and 500.13 MHz resonance frequency,
the resonances of the ˇ-pyrrolic 1H nuclei resolve into seven
unique, although relatively broad peaks (ꢂ1/2 D 14 Hz). The
peak areas correspond to six signals of 1H each and a single
signal (8.6 ppm) of 2H, accounting for all eight 1H nuclei on
the peripheral carbon atoms. The chemical shift distribution
has also increased relative to the room temperature spectra.
Copyright  2004 John Wiley & Sons, Ltd.
Magn. Reson. Chem. 2004; 42: 624–635

632
Y. S. Balazs et al.
Table 6. 13C chemical shifts (ppm) of substituted corrolesa
X
Ligand
Py
Substitution
NO2
Solvent
CDCl3
13C-2
115.5
13C-17
124.3
13C-18
117.6
13Cc(7, 13)
13Cd(8, 12)
Gab
127.3
129.3
128.7
130.4
128.7
129.6
130.7
131.6
132.5
132.0
126.6
132.1
126.9
130.8
124.1
126.7
124.1
126.1
126.0
127.0
126.9
126.5
126.4
131.1
Gab
Ga
Py
Py
NO2
CHO
C6D6
CDCl3
CDCl3
115.5
120.9
126.7
127.0
126.7
128.9
117.6
117.9
—
Ga
ꢀPyꢁ2
ꢀCHOꢁ2 asym
Gab
Cob
Cob
ꢀPyꢁ2
PPh3
PPh3
ꢀNO2ꢁ2 sym
ꢀSO2piꢁ2 symc
ꢀSO2piꢁ2 asymc
CDCl3
C6D6
C6D6
114.5
125.1
125.7
—
—
124.7
114.5
125.1
—
a 13C assignments based on cross peaks with 1H in HMQC experiment, therefore no 13C-3 assignments.
b Compounds with an x-ray structure where the substitution site is verified.
c pi D Piperidine.
Further clues for assignment of the low-temperature spec-
trum can be gleaned from the shielding effect of protonation
of the nitrogen atoms, extrapolated to a general affect on
each of the pyrrole rings. The shielding order of the 15N
1Hc
1Hd 1Hb
1Ha
1
1
15
peaks correlates with the H peaks [2D Hf Ng HMQC of
the uniformly 15N enriched H3(tpfc), Fig. 7(b)]. The least
shielded nitrogen resonance (ꢀ170.2 ppm) lacks a through-
(f)
1
bond H cross peak. This is confirmed by the 1D 15N NMR
(e)
spectrum acquired without decoupling where the left-most
peak is essentially a singlet [Fig. 7(c)]. The linewidth of this
peak (ꢂ1/2 D 40 Hz) and the signal-to-noise ratio preclude
observation of any weak coupling, as was observed in the
1H spectrum. The structure proposed in Fig. 7(d) suggests
that the frequencies of the nitrogens peaks would follow the
order B < C < A < D. Consequently, such labels were used
to label the nitrogen peaks in Fig. 7(b).
(d)
(c)
(b)
(a)
Assignment of the low-temperature ˇ-pyrrolic 1H
resonances of the metal-free tris(pentafluorophenyl)corrole
is proposed based on a number of assumptions: (i) the non-
protonated nitrogen atom is on the porphyrin-like side of
the macrocycle (ring B) as in the x-ray crystal structure of the
metal-free tris(pentafluorophenyl)corrole3b; (ii) the shielding
effects observed due to protonation of the nitrogen atoms can
be generalized to the pyrrole ring as a whole (i.e. shielding
follows B < C < A < D); (iii) the metal-free chemical shift
correlations of the ˇ-pyrrolic 1H resonances still hold at low
temperatures, namely υa > υb and υc > υd [Fig. 2(a)]. Starting
with the 1Ha —1Hb and 1Hc —1Hd pairings from the TOCSY
spectrum, taking into account the relative chemical shifts,
υa > υb and υc > υd, and finally adding the shielding effect
from the protonation state of the nitrogen generalized on to
the pyrrole ring, results in the labelling of the peaks shown
in Fig. 7(a). The complete assignment of the ˇ-pyrrolic 1H
resonances is listed in Table 7.
9.1 9.0 8.9 8.8 8.7 8.6 8.5 8.4 8.3 8.2 8.1 8.0
δH(ppm)
Figure 6. Temperature dependence of tris(pentafluorophenyl)
ring rotations. 500.13 MHz 1H spectra of (a) natural abundance
°
H3(tpfc) in toluene-d8 at ꢀ77 C, showing seven distinct
(relatively broad, ꢂ1/2 D 14 Hz) ˇ-pyrrolic 1H chemical shifts,
with the area under the 8.6 (ppm) peak twice that of the other
1H peaks; [U-15N]-H3(tpfc) at (b) ꢀ46, (c) ꢀ35 and (d) 0.1 C;
°
°
and natural abundance H3(tpfc) at (e) C23 C and (f) at
°
300.13 MHz and C45 C.
To assign the (unresolved) through-bond coupled pairs
at low temperature, a 2D TOCSY spectrum was acquired
[Fig. 7(a)]. The vicinal 1H pairs are highlighted with dot-
ted lines. A conservative assumption made for assignment
purposes is that the 1Ha chemical shifts account for the unre-
solved peak (1H-2 ³ 1H-18) of 2H relative area. The justifica-
tion is the relative distance from 1Ha to the pentafluorophenyl
rings; any changes in chemical shift due to differences in ring
positions at low temperatures would be smallest at 1Ha.
Dynamic processes of metal-free
5,10,15-tris(pentafluorophenyl)corrole
As seen in the previous section, two separate dynamic pro-
cesses occurring in metal-free tris(pentafluorophenyl)corrole
Copyright  2004 John Wiley & Sons, Ltd.
Magn. Reson. Chem. 2004; 42: 624–635

High-resolution NMR trends of corroles 633
Figure 7. Low-temperature (ꢀ76 C) 2D NMR experiments establishing through-bond connectivities. (a) 2D 1H–1H TOCSY
spectrum of the natural abundance metal-free tris(pentafluorophenyl)corrole. Dotted lines are drawn to highlight the through-bond
1H–1H connectivities. The ring and chemical shift assignments are labelled according to arguments outlined in the text. (b) An
expansion plot of the 2D 1H–15N HMQC (15N decoupled) spectrum of uniformly 15N-labelled H3(tpfc). An independently acquired 1D
15N spectrum (single excitation pulse with 1H decoupling) showing all four 15N resonances is drawn along the left axis. Along the
top, a 1D 1H projection of the 2D HMQC spectrum shows the three 1HN resonances (measured from 12.0 to ꢀ8.0 ppm and showing
no other peaks). (c) Single excitation pulse 15N spectrum (no decoupling). (d) A single tautomerization state with a weakly shared
hydrogen and variations in aryl ring conformations drawn as a guide to the arguments used for chemical shift assignments as given
in Table 7.
°
in toluene-d8 can be observed under the influence of variable-
temperature 1D 1H NMR spectroscopy: NH tautomerization
and pentafluorophenyl ring rotations. The existence of a sin-
gle preferred tautomeric state of the NH hydrogens is evident
Figure 5 shows that with increasing temperature the rate
of HN tautomerization is no longer slow on the NMR time-
scale. The 1HN peaks broaden and eventually coalesce around
°
°
room temperature [22 C at 300.13 MHz; Fig. 5(f)]. At 74 C, a
single, very broad peak is again visible, with a chemical shift
of ꢀ3.6 ppm [Fig. 5(g)].
N
1
by three distinct H shifts of equal peak area [Fig. 5(a)]. A
reduction in the rate of pentafluorophenyl ring rotations is
held responsible for the asymmetry of the corrole leading
Figure 6 shows that as the solution temperature increases,
the rate of the aryl ring rotations increase. A coalescence
1
to the resolution of the ˇ-pyrrolic H resonances into seven
°
unique peaks [Fig. 6(a)]. The presence of different preferred
conformations for each of the pentafluorophenyl rings,
destroying the symmetry of the macrocycle, is supported
by x-ray diffraction studies of metal-free corroles.3b,20,21
point can be seen around ꢀ35 C at 500 MHz. To reach
the fast exchange limit, where symmetry among ˇ-pyrrolic
hydrogen atoms is restored, giving rise to four narrow dou-
blets, it is necessary to raise the temperature to >30 C
°
Table 7. 1H and 15N NMR chemical shiftsa (ppm) and J(N, H) couplings (Hz) of metal-free [U-15N]-H3(tpfc) at
low temperaturesb
Pyrrole ring A
Pyrrole ring B
Pyrrole ring C
Pyrrole ring D
1Ha-2
8.62
8.32
1Hc-7
8.93
8.89
1Hd-12
8.48
8.52
1Hb-17
7.95
8.62
1Hb-3
15N
1Hd-8
15N
1Hc-13
15N
1Ha-18
15N
ꢀ254.6
ꢀ170.2
ꢀ206.6
ꢀ259.7
1HN
1J(N, H)
nJ(N, H)
ꢀ6.25
1HN
1J(N, H)
nJ(N, H) ꢀ15Nꢁ
—
—
1HN
1J(N, H)
nJ(N, H)
ꢀ0.09
(d) 70
(d) 19
1HN
1J(N, H)
nJ(N, H)
ꢀ6.34
c
c
(d) 94
(d) 95
d
d
—
(m)
—
a The assumptions used in making the assignments are given in the text.
1
b 1
15
°
°
H spectra at ꢀ72 C, N spectra at ꢀ75 C. All chemical shift calibrations to toluene ꢀC HD2ꢁ D 2.09
( D 10.136767 for 15N; IUPAC, 2001).
c No 1H signal
d Not observed.
Copyright  2004 John Wiley & Sons, Ltd.
Magn. Reson. Chem. 2004; 42: 624–635

634
Y. S. Balazs et al.
and decrease the field strength (300.13 MHz). In Fig. 6(f)
CONCLUSION
3
°
(45 C and 300.13 MHz), it is possible to resolve the J(H,H)
splittings of all four ˇ-pyrrolic 1H peaks, although some
source of broadening is still apparent. The additional broad-
enings can be attributed to long-range 6J(H,F) spin–spin
couplings.
The recent advances in synthetic methodologies for substi-
tuted corroles are providing opportunities to investigate the
unique chemical properties of corrole derivatives, with impli-
cations on a variety of fields. This paper is the first to identify
the correlations between 3J couplings and chemical shifts
fundamental for the identification of NMR spectral lines
and prerequisite for NMR analyses of basic structural and
dynamic properties of corroles and their metal complexes.
Analysis of the coupling constants and resonance frequencies
With 19F decoupling, the 1Hc, 1Hd, and 1Hb peaks are
significantly sharpened [Fig. 8(a)]. The 1Ha peaks (separated
from the 19F nuclei by seven bonds) are unaffected by the 19
F
decoupling, presumably owing to their relative distance. This
1
explains why, generally, the Ha doublet is narrower than
1
the other resonance peaks in tris(pentafluorophenyl)corroles.
The splittings of the 1Hc and 1Hd doublets into a triplet
of the vicinal H nuclei on the pyrrolic rings show that the
3Jab <3 Jcd and υa > υc > (υb, υd) behaviours are robust both in
metal-free and metallated corrole complexes. Comparisons
between calculated and experimental NMR coupling con-
6
due to J(H,F) couplings of ¾1 Hz can be seen in Fig. 8(b).
That the o-19F spins alone are sufficient to account for the
entire line shape effect can be confirmed by comparison
with Fig. 3(d), which displays the spectrum of a corrole with
meso-2,6-difluorophenyl substituents.
stants underscore the basic dependence of 3J(H,H) on C
C
bond lengths. Extensive tables of J-couplings and chemi-
1
cal shifts are also reported, including complete H and 13C
Figure 8(c) and (d) show that the 1H pairs in the corrole-
like and the porphyrin-like portions of the macrocycle
display different 3J(H,N) coupling patterns, as seen in the
1H spectra of 15N enriched metal-free corrole. In Fig. 8(c)
(19F-decoupled), which contains only 3J(H,H) and 3J(H,N)
splittings, each of the 3J(H,H) doublets is further split
chemical shift assignments of a gallium(III) corrole complex,
using a combination of specific labelling and sophisticated
high-resolution NMR spectroscopic methods.
With the reliable assignment of 1H NMR peaks, 1D NOE
difference spectra were shown to provide a relatively fast and
reliable solution-state method for determining the position
at which a site-directed substitution has occurred, a finding
of particular significance for instances when x-ray quality
crystals cannot be readily obtained. This study also paves
the way for investigations of paramagnetic metallocorroles
where the isotropic shifts cannot be analysed without prior
knowledge of the diamagnetic contributions. An a priori
determination of C C bond lengths via NMR measurements
also provides a lead for predicting reactivity of the various
carbon atoms. For example, the NMR data for H3(tpfc)
suggests that the C2 C3 and C17 C18 bonds participate
significantly more in the aromatic conjugation than the
C7 C8 and C12 C13 bonds and, indeed, hydrogenation of
H3(tpfc) proceeds with absolute selectivity on the latter.8b On
the other hand, electrophilic substitution proceeds selectively
on the C2 carbon atom, which according to DFT calculations
has the largest coefficient in the highest occupied molecular
orbital (HOMO).17 The same calculations also reveal a nodal
plane bisecting the ˇ-pyrrole carbon atoms in the HOMO
of corroles. Accordingly, it seems that small HC CH J-
coupling constants (and hence larger bond lengths) reflect
large coefficients on the carbon atoms, which is ‘translated’
into enhanced nucleophilicity of C2 C3 and C17 C18
relative to C7 C8 and C12 C13 bonds in corroles.
into doublets by the 15N nuclei [3J(H,N) of 1Ha –1HN
D
N
N
N
1
1
1Hb –1H D 3.4 Hz; Hc –1H D Hd –1H D 2.4 Hz). In the
3
complex spectra in Fig. 8(d), where all J(H,H), 6J(H,F), and
1
1
3J(H,N) splittings are present, Ha is still unaffected, Hb is
somewhat broadened and 1Hc and 1Hd are further split, each
peak into a ¾1 Hz triplet.
Figure 8. 1H NMR spectra demonstrating spin–spin
interactions on ˇ-pyrrolic hydrogen atoms. The natural
At elevated temperatures, studies of meso-C6F2-substi-
tuted H3(tdfc), and comparisons with the meso-C6F5-
substituted H3(tpfc), have demonstrated the influence of
the pentafluorophenyl ring motions on the 1H chem-
ical shifts and linewidths. Long-range couplings were
deconvoluted and identified by heteronuclear decoupling
experiments. Near-frozen solutions of triarylcorroles were
observed to display similar properties to crystalline metal-
free tris(pentafluorophenyl)corroles, namely one preferred
tautomeric state and unsymmetric ring orientations. It is
expected that further low-temperature studies at very high
°
abundance metal-free tris(pentafluorophenyl)corrole at 74 C,
(a) 19F-decoupled spectrum containing only 3J(H,H) splittings
(3JAC D 4.3 Hz, 3JBD D 4.8 Hz); (b) direct 1H excitation
spectrum showing the effects of 19F spins on ˇ-pyrrolic 1
H
linewidths and splittings. Additional splittings caused by the
pyrrolic 15N are apparent in 1H NMR spectra of a
[U-15N]-H3(tpfc) at 76 C: (c) 19F-decoupled spectrum and
°
(d) direct 1H excitation. All spectra were taken at 300.13 MHz
resonance frequency.
Copyright  2004 John Wiley & Sons, Ltd.
Magn. Reson. Chem. 2004; 42: 624–635

High-resolution NMR trends of corroles 635
7. Aviezer D, Cotton S, David M, Segev A, Khaselev N, Galili N,
Gross Z, Yayon A. Cancer Res. 2000; 60: 2973.
magnetic fields could possibly provide additional struc-
tural information on preferred conformations of corroles
and related phenomena.
8. (a) Bendix J, Dmochowski IJ, Gray HB, Mahammed A,
Simkhovich L, Gross Z. Angew. Chem., Int. Ed. Engl. 2000; 39:
4048; (b) Mahammed A, Gross Z. J. Inorg. Biochem. 2002; 88: 305.
9. (a) Gross Z. J. Biol. Inorg. Chem. 2001; 6: 733; (b) Erben, C, Will,
S, Kadish, KM. In The Porphyrin Handbook, vol. II, Kadish KM,
Smith, KM, Guilard R (eds). Academic Press: New York, 2000;
233; (c) Steene E, Wondimagegn T, Ghosh A, J. Phys. Chem. B
2001; 105: 11406; (d) Steene E, Wondimagegn T, Ghosh A, J.
Phys. Chem. B 2002; 106: 5312; (e) Yatsunyk L, Walker FA. Inorg.
Chim. Acta 2002; 337: 266; (f) Simkhovich L, Goldberg I, Gross Z.
Inorg. Chem. 2002; 41: 5433.
Acknowledgements
This research was supported by the Israel Science Foundation (grant
No. 368/00) and the Petroleum Research Fund (ACS, PRF 36550-
AC3). We are grateful to Dr H. Gottlieb (Bar Ilan University) for the
14.1 T solution-state measurements of the 13C-enriched Ga corrole
complex. We thank the Gilbert Foundation Magnetic Resonance
Facility in the Institute of Catalysis, Department of Chemistry,
Technion.
10. Johnson AW, Kay IT. J. Chem. Soc. Abstr. 1965; 1620.
11. (a) Gross Z, Galili N. Angew. Chem., Int. Ed. Engl. 1999; 38: 2366;
(b) Simkhovich L, Iyer P, Goldberg I, Gross Z. Chem. Eur. J. 2002;
8: 2595.
12. Paolesse R, Jaquinod L, Senge MO, Smith KM. J. Org. Chem. 1997;
62: 6193.
13. (a) Bartoli JF, Battioni P, De Foor WR, Mansuy D. J. Chem. Soc.,
Chem. Commun. 1994; 1: 23; (b) Ozette K, Leduc P, Palacio M,
Bartoli J-F, Barkigia KM, Fajer J, Battioni P, Mansuy D. J. Chem.
Soc. 1997; 119: 6442; (c) Garcia-Ortega H, Ribo JM. J. Porphyrins
Phthalocyanines 2000; 4: 564.
14. Bishop SM, Khoo BJ, MacRobert AJ, Simpson MSC, Phllips D,
Beeby A. J. Chromatogr. 1993; 646: 345.
15. Mahammed A, Goldberg I, Gross Z. Org. Lett. 2001; 3: 3443.
16. Gross Z, Mahammed A. J. Porphyrins Phthalocyanines 2002; 6: 553.
17. Saltsman I, Mahammed A, Goldberg I, Tkachenko E, Botoshan-
sky M, Gross Z. J. Am. Chem. Soc. 2003; 124: 7420.
18. Paolesse R, Nardis S, Venanzi M, Mastroianni M, Russo M,
Fronczek FR, Vicente MGH. Chem. Eur. J. 2003; 9: 1192.
19. (a) Braun J, Schlabach M, Wehrle B, Ko¨cher M, Vogel E,
Limbach HH. J. Am. Chem. Soc. 1994; 116: 6593; (b) Schlabach M,
Limbach HH, Bunnenberg E, Shu AYL, Bo-Ragnar T, Djerassi C.
J. Am. Chem. Soc. 1993; 115: 4554.
20. Paolesse R, Marini A, Nardis S, Froiio A, Mandoj F, Nurco DJ,
Prodi L, Montalti M, Smith KM. J. Porphyrins Phthalocyanines
2003; 7: 25.
REFERENCES
1. (a) Sessler JL. In Expanded, Contracted and Isomeric Porphyrins,
Sessler JL, Weghorn SJ (eds). Pergamon Press: Oxford, 1997; 11;
(b) Paolesse R. In The Porphyrin Handbook, vol. II, Kadish KM,
Smith KM, Guilard R (eds). Academic Press: New York, 2000;
201.
2. (a) Loim NM, Grishko EV, Pyshnograeva NI, Vorontsov EV,
Sokolov VI. Izv. Akad Nauk., Ser. Khim. 1994; 5: 925; (b) Rose E,
Kossanyi A, Quelquejeu M, Soleihavoup M, Duwavran F,
Bernard N, Lecas A. J. Am. Chem. Soc. 1996; 118: 1567; (c) Tse MK,
Zhang Z, Mak TCW, Chan KS. Chem. Commun. 1998; 1199.
3. (a) Gross Z, Galili N, Saltsman I. Angew. Chem., Int. Ed. Engl.
1999; 38: 1427; (b) Gross Z, Galili N, Simkhovich L, Saltsman I,
Botoshansky M, Blaser D, Boese R, Goldberg I. Org. Lett. 1999;
1: 599; (c) Paolesse R, Jaquinod L, Nurco DJ, Mini S, Sagone F,
Boschi T, Smith KM. Chem. Commun. 1999; 1307.
4. (a) Cho W-S, Lee C-H. Tetrahedron Lett. 2000; 41: 697; (b)
Gryko DT. Chem. Commun. 2000; 2243; (c) Brin˜as RP, Bru¨ckner C.
SynLett. 2001; 3: 442; (d) Gryko DT, Jadach K. J. Org. Chem. 2001;
66: 4267; (e) Ka J-W, Cho W-S, Lee C-H. Tetrahedron Lett. 2000; 41:
8121; (f)Paolesse R, Nardis S, Sagone F, Khoury RG. J. Org. Chem.
2001; 66: 550; (g) Asokan CV, Smeets S, Dehaen W. Tetrahedron
Lett. 2001; 42: 4483; (h) Simkhovich L, Goldberg I, Gross Z. Org.
Lett. 2003; 5: 1241; (i) Gryko DT, Koszarna B. Org. Biomol. Chem.
2003; 1: 350; (j) Decreau RA, Collman JP. Tetrahedron Lett. 2003;
44: 3323.
21. Simkhovich L, Goldberg I, Gross Z. J. Inorg. Biochem. 2000; 80:
235.
22. Mahammed A, Giladi I, Goldberg I, Gross Z. Chem. Eur. J. 2001;
7: 4259.
23. Brown HC, Surba RBC. J. Am. Chem. Soc. 1958; 80: 5377.
24. Harris RK, Becker ED, De Menezes SMC, Goodfellow R,
Granger P. Pure Appl. Chem. 2001; 73: 1795.
5. (a) Sessler JL. In Expanded, Contracted, and Isomeric Porphyrins,
Sessler JL, Weghorn SJ (eds.) Pergaman Press: Oxford, 1997;
429; (b) Montanari F, Casella L (eds). Metalloporphyrin Catalyzed
Oxidations. Kluwer: Dordrecht, 1994; (c) Sternberg ED,
Dolphin D, Bruckner C. Tetrahedron 1998; 54: 4151.
25. (a) Bax A, Griffey RH, Hawkins BL. J. Magn. Reson. 1983; 55: 301;
(b) Bax A, Subramanian S. J. Magn. Reson. 1986; 67: 565.
26. (a) Hurd RE, John BK. J. Magn. Reson. 1991; 94: 648; (b) Parella T,
Sanchez-Ferrando F, Virgili A. J. Magn. Reson. A 1995; 112: 241.
27. (a) Neuhaus, D, Williamson, M. The Nuclear Overhauser Effect in
Structural and Conformational Analysis. VCH: Weinheim, 1989;
(b) Neuhaus D. J. Magn. Reson. 1983; 53: 109; (c) Kinns M,
Sanders JKM. J. Magn. Reson. 1984; 56: 518.
6. (a) Gross Z, Simkhovich L, Galili N. Chem. Commun. 1999; 599;
(b) Gross Z, Golubkov G, Simkhovich L. Angew. Chem., Int.
Ed. Engl. 2000; 39: 4045; (c) Simkhovich L, Mahammed A,
Goldberg I, Gross Z. Chem. Eur. J. 2001; 7: 1041; (d) Golubkov G,
Bendix J, Gray HB, Mahammed A, Goldberg I, DiBilio AJ,
Gross Z. Angew. Chem., Int. Ed. Engl. 2001; 40: 2132; (e)
Simkhovich L, Gross Z. Tetrahedron Lett. 2001,; 42: 8089; (f)
Simkhovich L, Goldberg I, Gross Z. J. Porphyrins Phthalocyanines
2002; 6: 439; (g) Eikey RA, Khan SI, Abu-Omar MM. Angew.
Chem., Int. Ed. Engl. 2002; 41: 3592; (h) Liu HY, Lai TS, Yeung LL,
Chang CK. Org. Lett. 2003; 5: 617; (i) Mahammed A, Gray HB,
Meier-Callahan AE, Gross Z. J. Am. Chem. Soc. 2003; 125: 1162.
28. Gu¨nther, H. NMR spectroscopy: Basic Principles, Concepts, and
Applications in Chemistry (2nd edn). Wiley: New York, 2001.
29. Bax A, Davis DG. J. Magn. Reson. 1985; 65: 355.
Copyright  2004 John Wiley & Sons, Ltd.
Magn. Reson. Chem. 2004; 42: 624–635