Mono-aromatic ring-fused versus adjacently di-aromatic ring-fused tetraazaporphyrins: Regioselective synthesis and their spectroscopic and electrochemical properties

Article abstract of DOI:10.1021/ja012382u

A method which preferentially produces adjacently di-aromatic ring-substituted tetraazaporphyrins (TAPs) has been developed, and their electrochemical and spectroscopic properties have been studied and compared with those of the corresponding series of mo

Full text of DOI:10.1021/ja012382u

Published on Web 06/14/2002
Mono-Aromatic Ring-Fused versus Adjacently Di-Aromatic
Ring-Fused Tetraazaporphyrins: Regioselective Synthesis
and Their Spectroscopic and Electrochemical Properties
Nagao Kobayashi* and Takamitsu Fukuda
Contribution from the Department of Chemistry, Graduate School of Science, Tohoku UniVersity,
Sendai 980-8578, Japan
Received October 17, 2001
Abstract: A method which preferentially produces adjacently di-aromatic ring-substituted tetraazaporphyrins
(TAPs) has been developed, and their electrochemical and spectroscopic properties have been studied
and compared with those of the corresponding series of mono-aromatic ring-fused TAPs. Mono-aromatic
ring-fused TAPs show a split Q-band, and the splitting energy increases with increasing size of the aromatic
ring. In addition, for the split Q-bands, the relative intensity of the band at longer wavelength decreases
with increasing molecular size of the fused aromatics, compared with the shorter wavelength band. In the
di-aromatic ring-fused TAPs, this kind of splitting is not seen, and only a shift of the band is observed. The
intensity and band position of the split or unsplit Q-bands are quantitatively evaluated by simultaneous
band deconvolution analysis, using both electronic absorption and magnetic circular dichroism spectra.
The preparation of these TAP compounds has made it possible to adjust the Q-band position in a stepwise
manner between ca. 600 and 750 nm. The first reduction and oxidation potentials of the TAP ring shift
negatively with increasing number and size of the fused aromatics. The extent of the shift is found to be
very small for the LUMOs but significant for the HOMOs. These spectroscopic and electrochemical properties
are almost perfectly reproduced by molecular orbital calculations within the framework of the Pariser-
Parr-Pople approximation. In particular, a small variation of the LUMO level and large destabilization of
the HOMO level on ring expansion are rationalized from the extent of stretch of molecular orbitals: i.e.,
since the LUMOs are localized in the central TAP moiety irrespective of the molecular size, while the HOMOs
have appreciable coefficients even over the fused aromatics, the HOMO level destabilizes while the LUMO
level remains constant with increasing molecular size. In one CoTAP derivative, CoIII/II and the first ligand
oxidation couples occur experimentally at the same potential.
Introduction
mainly because of the difficulty in preparing precursors and
the low solubility of the resultant TAPs. In this respect,
Hoffman’s group developed the chemistry of TAPs intensively,
particularly in the fields of synthesis and X-ray crystallographic
analysis.12 However, reports on spectroscopic analysis and
electrochemical properties have been much less frequent. To
Porphyrins and phthalocyanines (Pc’s) are a very versatile
class of compound. Since their structural confirmation in the
early part of the 20th century,1-3 numerous papers have been
published, both in basic academic and in applied fields.4-10
However, research on tetraazaporphyrins (TAPs), which are
considered to be structural intermediates between normal
porphyrins and Pc’s,11 become popular only about a decade ago,
(12) (a) Michel, S. L. J.; Goldberg, D. P.; Stern, C.; Barrett, A. G. M.; Hoffman,
B. M. J. Am. Chem. Soc. 2001, 123, 4741. (b) Lee, S.; White, A. J. P.;
Williams, D. J.; Barrett, A. G. M.; Hoffman, B. M. J. Org. Chem. 2001,
66, 461. (c) Ehrlich, L. A.; Skrdla, P. J.; Jarrell, W. K.; Sibert, J. W.;
Armstrong, N. R.; Saavedra, S. S.; Barrett, A. G. M.; Hoffman, B. M.
Inorg. Chem. 2000, 39, 3963. (d) Bellec, N.; Montalban, A. G.; Williams,
D. B. G.; Cook, A. S.; Anderson, M. E.; Feng, X.; Barrett, A. G. M.;
Hoffman, B. M. J. Org. Chem. 2000, 65, 1774. (e) Montalban, A. G.; Jarrell,
W.; Riguet, E.; McCubbin, Q. J.; Anderson, M. E.; White, A. J. P.;
Williams, D. J.; Barrett, A. G. M.; Hoffman, B. M. J. Org. Chem. 2000,
65, 2472. (f) Nie, H.; Stern, C. L.; Barrett, A. G. M.; Hoffman, B. M.
Chem. Commun. 1999, 703. (g) Nie, H.; Barrett, A. G. M.; Hoffman, B.
M. J. Org. Chem. 1999, 64, 6791. (h) Montalban, A. G.; Steinke, J. H. G.;
Anderson, M. E.; Barrett, A. G. M.; Hoffman, B. M. Tetrahedron Lett.
1999, 40, 8151. (i) Andersen, K.; Anderson, M.; Anderson, O. P.; Baum,
S.; Baumann, T. F.; Beall, L. S.; Broderick, W. E.; Cook, A. S.; Eichhorn,
D. M.; Goldberg, D.; Hope, H.; Jarrell, W.; Lange, S. J.; McCubbin, Q. J.;
Mani, N. S.; Miller, T.; Montalban, A. G.; Rodriguez-Morgade, M. S.;
Lee, S.; Nie, H.; Olmstead, M. M.; Sabat, M.; Sibert, J. W.; Stern, C.;
White, A. J. P.; Williams, D. B. G.; Williams, D. J.; Barrett, A. G. M.;
Hoffman, B. M. J. Heterocycl. Chem. 1998, 35, 1013 and references therein.
(j) Forsyth, T. P.; Williams, D. B. G.; Montalban, A. G.; Stern, C. L.;
Barrett, A. G. M.; Hoffman, B. M. J. Org. Chem. 1998, 63, 331.
(1) Robertson, J. M.; Linstead, R. P.; Dent, C. E. Nature 1935, 135, 506.
(2) Linstead, R. P.; Robertson, J. M. J. Chem. Soc. 1936, 1195.
(3) Crute, M. B. Acta Crystallogr. 1959, 12, 24.
(4) Porphyrins and Metalloporphyrins; Smith, K. M., Ed.; Elsevier: Amster-
dam, 1975.
(5) The Porphyrins; Dolphin, D., Ed.; Academic Press: New York, 1978; Vols.
1-7.
(6) The Porphyrin Handbook; Kadish, K. M., Smith, K. M., Guilard, R., Eds.;
Academic Press: New York, 1999.
(7) Lever, A. B. P. AdV. Inorg. Chem. Radiochem. 1965, 7, 27.
(8) PhthalocyaninessProperties and Applications; Leznoff, C. C., Lever, A.
B. P., Eds.; VCH: New York, 1989 (Vol. 1), 1992 (Vol. 2), 1993 (Vol. 3),
and 1996 (Vol. 4).
(9) The Phthalocyanines; Moser, F. H., Thomas, A. H., Eds.; CRC Press: Boca
Raton, FL, 1983; Vols. 1 and 2.
(10) (a) De la Torre, G.; Claessens, C. G.; Torres, T. Eur. J. Org. Chem. 2000,
2821. (b) PhthalocyaninessChemistry and Functions; Shirai, H., Koba-
yashi, N., Eds.; IPC: Tokyo, 1997.
(11) (a) Kobayashi, N. In ref 8, Vol. 2, Chapter 3. (b) Kobayashi, N. In ref 6,
Vol. 2, Chapter 13.
9
10.1021/ja012382u CCC: $22.00 © 2002 American Chemical Society
J. AM. CHEM. SOC. 2002, 124, 8021-8034
8021

A R T I C L E S
Kobayashi and Fukuda
ideal for the purpose of the present study. As described below,
the finding of preferential formation of adjacently disubstituted
TAP species has made this kind of study possible.
For the same reason described in the preceding paper
(Kobayashi, N.; Miwa, H.; Nemykin, V. N., preceding paper
in this issue), we chose cobalt as the central metal. Although
zinc complexes are desirable from the standpoint of MO
calculations and availability of both NMR and fluorescence
spectroscopy, the naphtho-fused compounds were particularly
unstable and could not be purified. On the other hand, low yields
of nickel derivatives prevented the preparation of a series of
this compound.
Experimental Section
(i) Measurements. The 400 MHz 1H NMR and IR spectral
measurements were made with JEOL GSX-400 and Shimadzu FTIR-
8100M spectrometers, respectively. Mass spectra were measured with
a Hitachi M-2500S spectrometer (EI mass) and with a JEOL JMS-
HX110 mass spectrometer (FAB mass) using m-nitrobenzyl alcohol
(NBA) as a matrix. Electronic absorption spectra were measured with
a Hitachi U-3410 spectrophotometer. Magnetic circular dichroism
(MCD) spectra were run on a Jasco J-725 spectrodichrometer using a
toluene solution of ca. 10-5-10-6 mol/L, with a Jasco electromagnet
that produced a magnetic field of up to 1.09 T. The magnitude of this
was expressed in terms of molar ellipticity per tesla, [θ]M/deg mol-1
dm3 cm-1 T-1
.
Cyclic voltammetry data were collected with a Hokuto Denko HA-
501 potentiostat connected to a Graphtec WX1200 XY recorder.
Differential pulse voltammetry data were recorded with a Yanaco
P-1100 polarographic analyzer connected to a Watanabe WX 4401 XY
recorder. Electrochemical experiments were performed under purified
nitrogen gas. A glassy carbon working electrode (area ) 0.07 cm2)
and a Pt wire counter electrode were employed. The reference electrode
was Ag/AgCl, corrected for junction potentials by being referenced
internally to the ferrocenium/ferrocene (Fc+/Fc) couple. In the solution
used, i.e., in o-dichlorobenzene (o-DCB, Nacalai Tesque, HPLC grade)
containing 0.1 mol/L tetrabutylammonium perchlorate (TBAP), the Fc+/
Fc couple was observed at 0.51 ( 0.02 V vs Ag/AgCl. For spectro-
electrochemical measurements, an optically transparent thin-layer
electrode (OTTLE) cell of path length 1 mm was employed using a Pt
minigrid as both working and counter electrodes13 and a TBAP
concentration of 0.3 mol/L.
Figure 1. Structures and abbreviations of the compounds in this study.
our knowledge, no one has yet reported these kinds of properties
on a series of TAP compounds of varying size. In one respect,
this is due to the difficulty in controlling the structure of TAP
compounds. Since low symmetrical TAPs can be obtained only
by the mixed condensation of two kinds of maleonitrile or of a
maleonitrile and aromatic ortho dinitriles, at least six types of
product can possibly be formed, and the separation of these is
not necessarily easy. In another respect, there is the difficulty
in preparing a series of aromatic ortho dinitriles of a similar
structure as one of the two starting precursors. To extract
properties intrinsic to a series of compounds, we need to use
similar compounds with similar substituents, and we hope they
have sufficient solubility in a common solvent. For the TAP
compounds, it was not easy to overcome these two difficulties.
In this study, we have chosen diphenylmaleonitrile and 3,6-
diphenylphthalonitrile, 2,3-dicyano-1,4-diphenylnaphthalene, or
2,3-dicyano-1,4-diphenylanthracene as precursors. Octaphenyl-
ated TAPs with D4h symmetry generally have low solubility
themselves, but we considered that eight phenyl groups may
prevent the aggregation of low symmetrical TAPs obtained by
a mixed condensation reaction, which is important for spectro-
scopic and theoretical studies. In addition, as shown in Figure
1, the phenyl groups in the compounds are at the same positions,
precluding ambiguities in analysis, which makes our compounds
(ii) Molecular Orbital Calculations. MO calculations were per-
formed by the ZINDO/S Hamiltonian in the HyperChem R.5.1
program14a and the Pariser-Parr-Pople (PPP) Hamiltonian described
previously.14b The structures of Pc analogues for calculations were
constructed by using X-ray structural data of standard Pc,15 naphthalene,16a
and anthracene16b and by making the rings perfectly planar and adopting
the C2V symmetry, and therefore essentially the same as those we used
in our previous papers.17 The central metal was assumed as zinc for
the ZINDO/S calculations and nil for the PPP calculations. The choice
(13) Kobayashi, N.; Nishiyama, Y. J. Phys. Chem. 1985, 89, 1167.
(14) (a) HyperChem R.5.1 Pro; Hypercube Inc., Gainesville, FL, 1997. (b)
Tokita, S.; Matsuoka, K.; Kogo, Y.; Kihara, K. Molecular Design of
Functional Dyes-PPP Method and Its Application; Maruzen: Tokyo, 1990.
(15) Robertson, J. M.; Woodward, I. J. Chem. Soc. 1937, 219. Barrett, P. A.;
Dent, C. E.; Linstead, R. P. J. Chem. Soc. 1936, 1719. Brown, C. J. J.
Chem. Soc. A 1968, 2488, 2494. Kirner, J. F.; Dow, W.; Scheidt, D. R.
Inorg. Chem. 1976, 15, 1685.
(16) (a) House, H. O.; Koepsell, D. G.; Campbell, W. J. J. Org. Chem. 1972,
37, 1003. (b) Fillers, J. P.; Ravichsandran, K. G.; Abdalmuhdi, I.; Chang,
C. K. J. Am. Chem. Soc. 1986, 108, 417.
(17) (a) Gouterman, M. J. Mol. Spectrosc. 1961, 6, 138. (b) Kobayashi, N.;
Konami, H. In PhthalocyaninessProperties and Applications; Leznoff, C.
C., Lever, A. B. P., Eds.; VCH: New York, 1996; Vol. 4, Chapter 13. (c)
Kobayashi, N.; Konami, H. J. Porphyrins Phthalocyanines 2001, 5, 233.
(d) Kobayashi, N.; Ishizaki, T.; Ishii, K.; Konami, H. J. Am. Chem. Soc.
1999, 121, 9096.
9
8022 J. AM. CHEM. SOC. VOL. 124, NO. 27, 2002

Aromatic Ring-Fused Tetraazaporphyrins
A R T I C L E S
of configuration was based on energetic considerations, and all singly
excited configurations up to 9 eV (72 590 cm-1) were included.
[21,26-Diphenylnaphtho[b]-7,8,12,13,17,18-hexaphenyl-5,10,15,20-
tetraazaporphyrinato(2-)]cobalt(II), CoNpAP, and [21,26,71,76-
Tetraphenyldinaphtho[b,g]-12,13,17,18-tetraphenyl-5,10,15,20-tet-
raazaporphyrinato(2-)]cobalt(II), CoNp2AP. As for the above
CoBzAP and CoBz2AP, these were synthesized from diphenylmale-
onitrile (330 mg, 1.43 mmol) and 2,3-dicyano-1,4-diphenylnaphthalene
(372 mg, 1.13 mmol). The crude reaction product was purified using
a PLC plate on silica with toluene as eluent. Two fractions with Rf )
0.85 (CoNp2AP) and 0.65 (CoNpAP) were collected and recrystallized
from dichloromethane/methanol, to give 23 mg (3.3%) of CoNpAP as
a blue solid and 82 mg (11%) of CoNp2AP as a green solid. Mass
(FAB) (m/z): 1080 (M+ + 1 for CoNpAP), 1180 (M+ + 1 for CoNp2-
AP). Anal. Found: C, 80.19; H, 4.24; N, 10.43. Calcd for C72H44N8Co
(CoNpAP): C, 80.06; H, 4.11; N, 10.37. Anal. Found: C, 80.56; H,
4.38; N, 9.34. Calcd for C80H48N8Co (CoNp2AP): C, 81.41; H, 4.10;
N, 9.49.
[21,28-Diphenylanthraco[b]-7,8,12,13,17,18-hexaphenyl-5,10,15,20-
tetraazaporphyrinato(2-)]cobalt(II), CoAnAP, and [21,28,71,78-
Tetraphenyldianthraco[b,g]-12,13,17,18-tetraphenyl-5,10,15,20-tet-
raazaporphyrinato(2-)]cobalt(II), CoAn2AP. To 5 mL of 1-hexanol
containing ca. 100 mg (14 mmol) of lithium was added a mixture of
finely powdered diphenylmaleonitrile (180 mg, 0.78 mmol) and 2,3-
dicyano-1,4-diphenylanthracene (250 mg, 0.66 mmol), and the solution
was stirred at 150 °C for 1 h. Most of the solvent was removed by
passing nitrogen gas over the solution at this temperature, and then 30
mL of DMF and anhydrous cobalt(II) chloride (1.4 g, 11 mmol, 31
equiv) were added, and the stirring continued for 30 min more at this
temperature. After cooling, the solution was poured into 200 mL of
water, and the resultant precipitate was collected by filtration and
washed with water and methanol. The crude reaction product was
purified using a PLC plate on silica with toluene as eluent, and two
portions with Rf ) 0.83 (CoAn2AP) and 0.68 (CoAnAP) were collected.
Recrystallization from dichloromethane/methanol produced the desired
CoAnAP (16 mg, 3.9%) as a blue solid and CoAn2AP (40 mg, 8.7%)
as a green solid. Mass (FAB) (m/z): 1130 (M+ + 1 for CoAnAP),
1280 (M+ + 1 for CoAn2AP). Anal. Found: C, 80.45; H, 4.44; N,
10.00. Calcd for C76H46N8Co (CoAnAP): C, 80.77; H, 4.10; N, 9.91.
Anal. Found: C, 81.98; H, 4.33; N, 8.65. Calcd for C88H52N8Co (CoAn2-
AP): C, 82.55; H, 4.09; N, 8.75.
(iii) Synthesis. All new low symmetrical TAPs in Figure 1 have
been obtained by mixed condensation between diphenylmaleonitrile18
and either 3,6-diphenylphthalonitrile,19a 2,3-dicyano-1,4-diphenylnaph-
thalene,19b,c or 2,3-dicyano-1,4-diphenylanthracene. Yields of the TAP
derivatives were calculated on the basis of the total nitrile mixed.
2,3-Dicyano-1,4-diphenyl-1,4-oxide-2,3-dihydroanthracene. To 10
mL of dichloromethane containing 6.3 g (20 mmol) of 1,3-diphenyl-
naphtho[2,3-c]furan20 was added fumaronitrile stepwise with stirring
until the solution changed from red to yellow. The solvent was
evaporated and the residue recrystallized from methanol, to give 7.2 g
(90%) of the desired product, mp 214-217 °C. Mass (EI) (m/z): 398
(M+). IR (KBr): 2338, 2365 cm-1 (CN). 1H NMR (CDCl3, 400
MHz): δ 3.69 (d, 1H), 3.97 (d, 1H), 7.50-7.68 (m, 10H), 7.73 (d,
2H), 7.78-7.82 (m, 2H), 7.88 (d, 2H). 13C NMR (CDCl3, 100 MHz):
δ 18.2, 58.3, 71.9, 83.7, 126.4, 127.5, 128.0, 128.1, 128.6, 129.8, 130.9,
131.6, 132.6, 141.2, 147.0, 147.1. Anal. Found: C, 84.36; H, 4.58; N,
6.99. Calcd for C28H18N2O: C, 84.40; H, 4.55; N, 7.03.
2,3-Dicyano-1,4-diphenylanthracene. The above compound (7.2
g, 18 mmol) was added to 5 mL of concentrated sulfuric acid and stirred
for 10 min at 0 °C. The reaction mixture was poured into ice and stirred
for a further 20 min, and then the precipitated solid was filtered off
and washed with water and methanol. Recrystallization from CH2Cl2
gave 6.5 g (94%) of the desired compound. mp > 320 °C. Mass (EI)
1
(m/z): 380 (M+). IR (KBr): 2226 cm-1 (CN). H NMR (CDCl3, 400
MHz): δ 7.55-7.58 (m, 6H), 7.63-7.66 (m, 6H), 7.91-7.94 (dd, 2H),
8.34 (s, 2H). 13C NMR (CDCl3, 100 MHz): δ 108.5 (13C-CN), 116.1
(C-13CN), 128.2, 128.6, 128.8, 129.1, 129.7, 129.9, 130.0, 133.3, 135.6,
149.1. Anal. Found: C, 88.41; H, 4.22; N, 7.37. Calcd for C28H16N2:
C, 88.40; H, 4.24; N, 7.36.
[2,3,7,8,12,13,17,18-Octaphenyl-5,10,15,20-tetraazaporphyrinato-
(2-)]cobalt(II), CoTAP. This compound was prepared in 50% yield
according to the literature.18 Mass (FAB) (m/z): 980 (M+ + 1). Anal.
Found: C, 77.70; H, 4.92; N, 11.08. Calcd for C64H40N8Co: C, 78.44;
H, 4.11; N, 11.43.
[21,24-Diphenylbenzo[b]-7,8,12,13,17,18-hexaphenyl-5,10,15,20-
tetraazaporphyrinato(2-)]cobalt(II), CoBzAP, and [21,24,71,74-Tet-
raphenyldibenzo[b,g]-12,13,17,18-tetraphenyl-5,10,15,20-tetraaza-
porphyrinato(2-)]cobalt(II), CoBz2AP. To 3 mL of 1-hexanol
solution containing ca. 20 mg (2.9 mmol) of lithium was added a
mixture of finely powdered diphenylmaleonitrile (299 mg, 1.30 mmol)
and 3,6-diphenylphthalonitrile (314 mg, 1.12 mmol), and the solution
was stirred at 140 °C for 1 h. Most of the solvent was removed by
passing nitrogen gas over the solution at this temperature, and then 20
mL of DMF and anhydrous cobalt(II) chloride (2.9 g, 22 mmol, 36
equiv) were added, and the stirring was continued for a further 30 min
at this temperature. After cooling, the solution was poured into 200
mL of water, and the resultant precipitate was collected by filtration
and washed with water and methanol. After drying, the crude product
was repeatedly imposed on a silica gel column followed by a preparative
thin-layer (PLC, silica) plate using carbon tetrachloride/toluene (1:1
v/v) as eluent, and the portions with Rf ) 0.50 (CoBz2AP) and 0.33
(CoBzAP) were collected. Recrystallization from dichloromethane/
methanol produced the desired CoBzAP (7.7 mg, 1.2%) and CoBz2AP
(54 mg, 8.3%) as a blue solid. Mass (FAB) (m/z): 1030 (M+ + 1 for
CoBzAP), 1080 (M++ 1 for CoBz2AP). Anal. Found: C, 77.77; H,
5.13; N, 10.04. Calcd for C68H42N8Co‚H2O (CoBzAP): C, 77.93; H,
4.23; N, 10.69. Anal. Found: C, 79.89; H, 4.33; N, 10.29. Calcd for
C72H44N8Co (CoBz2AP): C, 80.06; H, 4.11; N, 10.37.
[21,28-Diphenylanthraco[b]-7,8,12,13,17,18-hexaphenyl-5,10,15,20-
tetraazaporphyrinato(2-)]nickel(II), NiAnAP, and [21,28,71,78-Tet-
raphenyldianthraco[b,g]-12,13,17,18-tetraphenyl-5,10,15,20-tetraaza-
porphyrinato(2-)]nickel(II), NiAn2AP. These compounds were obtained
using nickel(II) chloride instead of cobalt(II) chloride in the preparation
1
of CoAnAP and CoAn2AP, to confirm the structure from H NMR
and to investigate the spectroelectrochemistry of ring-reduced and
-oxidized species. The yield of blue colored NiAnAP and reddish brown
NiAn2AP was 4 mg (0.9%) in both cases, this low yield being
attributable to reduction in the purification process due to low solubility.
Mass (FAB) (m/z): 1129 (M+ + 1 for NiAnAP), 1279 (M+ + 1 for
NiAn2AP). Anal. Found: C, 80.87; H, 4.12; N, 9.74. Calcd for
C76H46N8Ni (NiAnAP): C, 80.79; H, 4.10; N, 9.92. Anal. Found: C,
82.09; H, 4.27; N, 8.69. Calcd for C88H52N8Ni (NiAn2AP): C, 82.57;
H, 4.09; N, 8.75. 1H NMR (CDCl3, 400 MHz): for NiAnAP, δ 7.07-
7.15 (m, 2H), 7.27-7.31 (dd, 4H), 7.38-7.50 (m, 26H), 7.60-7.62
(d, 4H), 7.84-7.89 (m, 4H), 7.93-7.96 (m, 2H), 8.06-8.13 (m, 4H),
8.53 (s, 2H); for NiAn2AP, δ 7.10-7.55 (m, 40H), 7.73-7.76 (m, 4H),
7.85-7.92 (m, 4H), 8.35 (s, 2H), 8.61 (s, 2H).
Results and Discussion
(i) Synthesis. As described above, from the condensation
reaction using two kinds of ortho dinitriles in a 1:1 ratio, only
adjacently di-aromatic-ring fused (AABB type) and mono-
aromatic ring-fused (ABBB type) species were obtained, and
the preferential formation of the former was confirmed (roughly
(18) Cook, A. H.; Linstead, R. P. J. Chem. Soc. 1937, 929.
(19) (a) Kobayashi, N.; Ashida, T.; Osa, T. Chem. Lett. 1992, 1567, 2031. (b)
Kobayashi, N.; Ashida, T.; Osa, T.; Konami, H. Inorg. Chem. 1994, 33,
1735. (c) Kobayashi, N.; Togashi, M.; Osa, T.; Ishii, K.; Yamauchi, S.;
Hino, H. J. Am. Chem. Soc. 1996, 118, 1073.
(20) Cava, M. P.; VanMeter, J. P. J. Org. Chem. 1969, 34, 538.
9
J. AM. CHEM. SOC. VOL. 124, NO. 27, 2002 8023

A R T I C L E S
Kobayashi and Fukuda
Figure 2. 400 MHz 1H NMR spectra of NiAnAP (left) and NiAn2AP (right) in CDCl3. In the spectra of NiAn2AP, a small amount of NiAnAP was
intentionally added in order to increase the solubility. Signals from the solvent are marked X.
3-7 times). The key factor in the synthesis was the use of
lithium at the first stage, so that the subsequent addition of
transition metals produced the above results. If transition metals
were used from the beginning, no AABB type compounds would
be obtained.19 From experiments carried out under various
conditions, we have explained the above results as follows.
Lithium acts as a strong template, so that initially two dinitriles
containing two protruding phenyl groups such as 3,6-diphen-
ylphthalonitrile react to form a half Pc,21 after which it will
become harder for another 3,6-diphenylphthalonitrile to link to
this unit due to steric hindrance of the phenyl groups. In
particular, lithium is the hardest metal in the soft-hard concept
(i.e., small volume but high charge density), so that when two
dinitrile molecules, say 3,6-diphenylphthalonitrile, are connected
by lithium to form a half Pc, these molecules are considered to
take a folded conformation around lithium, thereby making the
attack of the third dinitrile difficult. Of course, at higher
temperature, the third dinitrile may be able to link to this half
Pc,22 but by controlling the reaction temperature, it is possible
to stop further addition of the sterically hindered diphenylphtha-
lonitrile. Diphenylmaleonitrile then reacts at slower rate to form
a half Pc, because of the absence of a unit to keep the dinitrile
moiety flat (in phthalonitrile, for example, two nitrile units are
Figure 3. Electronic absorption (bottom) and MCD spectra (top) of
NiAnAP (solid lines) and NiAn2AP (broken lines) in toluene. Note the
kept flat by the presence of a benzene ring, which makes
coordination to template metals easier, thereby facilitating the
formation of the Pc skeleton). In this study, we used diphenyl-
maleonitrile as one of the precursors because of the easy
accessibility,18 and therefore octaphenyltetraazaporphyrin was
also formed. However, because of its low solubility in various
solvents, this compound was easily removed during column
chromatography.
encircled magnification factor.
show a split and unsplit Q-band, respectively, and the Q-band
position of the latter should appear at longer wavelength than
the former. Indeed, the anticipated spectra were observed, and
the corresponding MCD spectra were also consistent with the
structure, i.e., Faraday B-term type and A-term type, respec-
tively.24 Although the structure and spectral shape relationship
was as expected, of particular note is the large intensity
difference between NiAnAP and NiAn2AP. As indicated by the
encircled magnification factor in the figure, the intensity of
NiAn2AP is much higher than that of NiAnAP. However, this
is also consistent with the MO calculation result, that the
intensity of the Q-band of TAP increases markedly with fusing
of large aromatic molecules.17a-c The small absorption peaks
seen at 446 and 485 nm can be ascribed to transitions in
anthracene-centered orbitals.17b,25
1
Figure 2 shows H NMR spectra of NiAnAP and NiAn2AP
specially prepared for this purpose. As expected for the
structures of these compounds, the protons at the center of
anthracene moiety appear as a singlet for NiAnAP and two
separate singlets for NiAn2AP, confirming the desired struc-
tures.23
(ii) Electronic Absorption and MCD Spectroscopy. Figure
3 shows the absorption and MCD spectra of NiAnAP and
NiAn2AP recorded in order to look for any correlation with the
above NMR results. According to the symmetry-adapted
perturbation (SAP) theory,17a,b NiAnAP and NiAn2AP should
Figure 4A and B displays electronic absorption and MCD
spectra, respectively, of ABBB type cobalt species in toluene.
(21) Nolan, K. J. M.; Hu, M. J.; Leznoff, C. C. Synlett 1997, 593.
(22) Kobayashi, N.; Fukuda, T.; Ueno, K.; Ogino, H. J. Am. Chem. Soc. 2001,
123, 10740.
(24) Stillman, M. J.; Nyokong, T. In PhthalocyaninessProperties and Applica-
tions; Leznoff, C. C., Lever, A. B. P., Eds.; VCH: New York, 1989; Vol.
1, Chapter 3.
(23) The structure of CoAn2AP has also been confirmed by X-ray crystal-
lography.22
(25) Miwa, H.; Kobayashi, N. Chem. Lett. 1999, 1303.
9
8024 J. AM. CHEM. SOC. VOL. 124, NO. 27, 2002

Aromatic Ring-Fused Tetraazaporphyrins
A R T I C L E S
of the fused aromatics. Of the split Q-bands, the band at shorter
wavelength shows little or no change in its position, while that
at longer wavelength shifts to the red with a concomitant
decrease in relative intensity. If we consider the center of the
split Q-band as the position of the Q-band of the ABBB type
species, the shifts from the Q-band of CoTAP are 680, 1130,
and 1710 cm-1 for CoBzAP, CoNpAP, and CoAnAP, respec-
tively, based on band deconvolution analysis. Thus, the energy
differences (∆E) for CoTAP-CoBzAP, CoBzAP-CoNpAP,
and CoNpAP-CoAnAP are 680, 450, and 580 cm-1, respec-
tively. The Soret band position shifts to shorter wavelength on
going from CoTAP to CoBzAP, but then it shifts to longer
wavelength with increasing molecular size; i.e., the energy
difference between CoBzAP and CoNpAP is 580 cm-1, while
that between CoNpAP and CoAnAP is 790 cm-1. In this way,
∆E values in the Soret band are not particularly small compared
with those in the Q-band. All compounds show a curvature at
the foot of the longer wavelength tail of the Soret band,
suggesting there is at least one other transition in this region.
The MCD spectra show dispersion type curves in the Q-band
region, roughly corresponding to the split absorption peaks.
These are therefore interpreted as a superimposition of Faraday
B-terms.26 However, since the splitting is not large enough to
be observed as separate peaks, particularly for CoBzAP and
CoNpAP, they look apparently like Faraday A-terms (pseudo
A-term).27 Similarly to the case with normal TAPs and Pc’s
with D4h symmetry,24 the intensity in the Soret band region is
much weaker than that in the Q-band, reflecting the ∆L (angular
momentum) ) (1 nature of the former compared with ∆L )
(9 of the latter.
Figure 5A and B shows spectra for adjacently disubstituted
AABB type compounds. With increasing size of the fused
aromatics, the Q-band shifts to longer wavelength without
appreciable splitting. The position of the Q-band lies always at
longer wavelength, while the Soret band lies at shorter
wavelength than in the corresponding monosubstituted ABBB
type species in Figure 4A. The shift of the Q-band from the
Q-band of CoTAP is approximately twice as large as that in
the case of the corresponding monosubstituted ABBB type
compounds, as has been anticipated by the SAP method.17a,b
For example, the ∆E values of CoTAP-CoBz2AP and CoTAP-
CoNp2AP are 1390 and 2230 cm-1, respectively, while ∆E )
680 and 1130 cm-1 for CoTAP-CoBzAP and CoTAP-
CoNpAP, respectively. One or two small curvatures are discern-
ible at the foot of the red tail of the Soret band envelope. The
MCD spectra of all compounds are dispersion type Faraday
A-terms corresponding to the Q absorption peaks, suggesting
the presence of degenerate excited states. In the Soret band
region, there appear to be several transitions since curves are
not simple and appear to be a superimposition of several Faraday
A-terms.
The absorption and MCD data of all the compounds in this
study are summarized in Table 1. As can be seen in this table
and Figures 4A and 5A, it becomes possible to adjust the main
absorption band (i.e., Q-band) of TAPs stepwise over the wide
region ranging ca. 600-750 nm. In the case of ABBB type
compounds, the bands are broad, but for AABB type com-
Figure 4. (A) Electronic absorption and (B) MCD spectra of CoTAP,
CoBzAP, CoNpAP, and CoAnAP in toluene. Note the encircled magnifica-
tion factor in (B). The Gaussian lines drawn by broken lines indicate a
split Q00-band by band deconvolution analysis.
(26) Tajiri, A.; Winkler, J. Z. Naturforsch. 1983, 38a, 1263.
(27) Kaito, A.; Nozawa, T.; Yamamoto, T.; Hatano, M.; Orii, Y. Chem. Phys.
Lett. 1977, 52, 154.
As exemplified by the above NiAnAP, all these complexes show
split Q-bands, and the splitting increases with increasing size
9
J. AM. CHEM. SOC. VOL. 124, NO. 27, 2002 8025

A R T I C L E S
Kobayashi and Fukuda
Deconvolution of the spectroscopic envelope can give reliable
results if both the absorption and MCD spectra are fitted with
the same or very similar band centers and bandwidth parameters.
In fact, to date, many absorption and MCD spectra have been
analyzed by band deconvolution, and this has been successful
in analyzing transitions that had been postulated by MO
calculations or in finding a new degenerate excited state.28 We
have therefore employed this method in order to obtain
quantitative information. Figure 6 shows an example of band
deconvolution analysis for the Q-band of CoAnAP. In this
figure, the two obtained Q00-bands are indicated by solid lines.
Deviation of the simulated spectrum from that obtained experi-
mentally is very small, as shown at the bottom of the figure.
The Q-bands of the other compounds have been similarly
analyzed, with the obtained parameters summarized in Table
2, and the Q00 components are added in Figures 4 and 5 as
broken lines. In the case of monosubstituted ABBB type
compounds shown in Figure 4A and B, two components are
required in both the absorption and MCD spectra. In the case
of absorption spectra of Figure 4A, splitting of the Q-band
increases and the relative intensity of the component at longer
wavelength decreases with increasing size of the fused aromat-
ics. For adjacently disubstituted AABB type compounds, only
one component is required in the Q-band region (Figure 5). The
Q-band MCD is expressed by a single Faraday A-term. Since
Gaussian curves are used for the deconvolution analysis,
parameters of each envelope are accurately obtained (Table 2).
For example, absorption coefficients, band positions, oscillator
strengths, and so forth are numerically expressed, and these are
used in the comparison of the results of the MO calculations.
(iii) Electrochemistry. Redox potentials have been reported
for many Pc’s29 and porphyrins.30 In particular, the first
oxidation and reduction potentials are important for finding the
relationship between the absorption bands lying at the longest
wavelength. Figure 7 displays the cyclic voltammograms of
cobalt complexes in o-DCB containing 0.1 mol/L TBAP, while
the data are collected in Table 3 and Figure 8. The oxidation
potential of CoBzAP was read from the differential pulse
voltammogram, and the second oxidation potential of CoNp2-
AP was not obtained because of the instability of the first
oxidation product. TAP derivatives fused with naphthalene rings
such as some types of naphthalocyanine31 and tribenzo-
mononaphthoTAP32 are experimentally known to be unstable
with respect to oxidation, and the above data are consistent with
this. From the difference between the various redox potentials,
all processes except for the first oxidation of CoTAP and
CoBzAP are one-electron processes. The first oxidation curve
of CoTAP was found to be a superimposition of two oxidation
curves, since two peaks appeared at higher scans, of which the
more positive peak became more prominent at higher scan rates.
(28) Some examples: Nyokong, T.; Gasyna, Z.; Stillman, M. J. Inorg. Chem.
1987, 26, 1087. Browett, J. R.; Stillman, M. J. J. Am. Chem. Soc. 1988,
110, 3633. Ough, E.; Nyokong, T.; Kreber, K. A.; Stillman, M. J. Inorg.
Chem. 1988, 27, 2724. Ough, E. A.; Stillman, M. J. Inorg. Chem. 1994,
33, 573. Mack, M. J.; Stillman, M. J. J. Am. Chem. Soc. 1994, 116, 1292.
Kobayashi, N.; Fukuda, T.; Lelie`vre, D. Inorg. Chem. 2000, 39, 3632.
(29) Lever, A. B. P.; Milaeva, E. R.; Speier, G. In PhthalocyaninessProperties
and Applications; Leznoff, C. C., Lever, A. B. P., Eds.; VCH: New York,
1993; Vol. 3.
Figure 5. (A) Electronic absorption and (B) MCD spectra of CoTAP,
CoBz2AP, CoNp2AP, and CoAn2AP in toluene. Note the encircled
magnification factor in (B). The Gaussian lines shown by broken lines
indicate a Q00-band obtained by band deconvolution analysis.
(30) Kadish, K. M.; Royal, G.; Caemelbecke, E. V.; Gueletti, L. In The Porphyrin
Handbook; Kadish, K. M., Smith, K. M., Guilard, R., Eds.; Academic
Press: New York, 1999; Chapter 59.
(31) Kobayashi, N.; Higashi, Y.; Osa, T. Chem. Lett. 1994, 1813.
(32) Kobayashi, N.; Kondo, R.; Nakajima, S.; Osa, T. J. Am. Chem. Soc. 1990,
112, 9640.
pounds, the Q-band width does not change significantly from
species to species.
9
8026 J. AM. CHEM. SOC. VOL. 124, NO. 27, 2002

Aromatic Ring-Fused Tetraazaporphyrins
A R T I C L E S
Table 1. Absorption and MCD Data in Toluene
b
compound
absorptiona
MCD
CoTAP
354 (2.69)
615 (3.83)
344 (5.09)
634 (5.48)
351 (5.56)
635 (5.01)
361 (4.69)
654 (3.68)
566shc (1.42)
374 (-0.23)
598 (1.04)
337 (-0.06)
658 (-2.45)
328 (0.04)
587sh (1.29)
362 (-0.23)
601 (0.87)
417 (0.03)
628 (-1.98)
587 (1.66)
564 (0.93)
CoBzAP
CoNpAP
CoAnAP
574sh (2.36)
655sh (4.61)
585sh (2.72)
687 (3.35)
609sh (2.67)
726 (1.88)
612sh (1.37)
364 (-0.33)
615 (1.50)
430 (-0.04)
640 (0.66)
430 (0.04)
684 (-1.80)
486 (0.08)
681sh (-0.65)
722 (-0.89)
367 (-0.25)
656 (1.80)
330 (0.11)
690 (1.12)
CoBz2AP
CoNp2AP
338 (5.39)
676 (7.27)
312 (5.74)
434sh (1.30)
712 (5.61)
355 (7.83)
490 (1.48)
748 (7.82)
361 (5.68)
598sh (2.66)
729 (3.66)
358 (12.1)
673 (5.44)
613sh (3.12)
397sh (-0.16)
685 (-4.25)
359 (-0.24)
727 (-2.93)
606 (1.87)
637 (1.27)
344 (5.06)
639sh (2.46)
CoAn2AP
NiAnAP
NiAn2AP
425sh (2.69)
669sh (3.06)
357 (-0.49)
498 (0.19)
760 (-3.83)
357 (-0.46)
450 (-0.10)
606 (1.13)
315 (1.13)
457 (-1.70)
723 (4.52)
388 (0.10)
663 (1.26)
424 (-0.07)
725 (1.75)
446 (1.49)
645 (5.82)
387 (0.08)
490 (0.14)
410 (-0.05)
527 (0.03)
485 (2.87)
745 (16.8)
357 (-1.02)
498 (0.26)
755 (-8.33)
389 (1.48)
663 (2.63)
a λ/nm (10-4 ꢀ/dm3 mol-1 cm-1). b λ/nm (10-5 [θ]M/deg dm3 mol-1cm-1 T-1). c sh denotes shoulder.
rates or at lower sample concentrations is due to the nonaggre-
gated monomer species. Interestingly, oxidation of aggregates
always occurs at less positive potentials than that of nonaggre-
gated species.34
Voltammograms of cobalt TAPs, Pc’s, and naphthalocyanines
generally differ from those of the zinc, copper, and nickel
complexes, in showing redox couples due to cobalt oxidation
and reduction. In this study, the assignment of the couples is
based on the previous data29 on redox couples of TAPs and
Pc’s, as well as spectroelectrochemical data to be treated in a
later section. As a result, the first and second reductions of all
the compounds in o-DCB were assigned to the CoII/I and the
first ligand reduction (L-2/-3) couples,35 respectively, while the
first and second oxidations were assigned to the ligand first
oxidation and CoIII/II couples, respectively, except in the case
of CoTAP. The assignment of the two oxidation couples of
CoTAP could not be performed, even by the spectroelectro-
chemical method, due to the low solubility. Accordingly, the
result of highly soluble tetra-tert-butylated CoTAP36 in o-DCB
was employed, and then the first oxidation was tentatively
assigned to the CoIII/II couple and the second to the ligand first
oxidation. These assignments match well with the series of data,
since all redox couples of the monosubstituted ABBB type (left-
hand side in Figure 8) appear at more positive potentials than
those of the corresponding adjacently disubstituted AABB type
species (right-hand side in Figure 8). This means that the fused
aromatics function as electron-releasing groups. Accordingly,
the first ligand reduction of AABB type species is more difficult
than that of the ABBB type species, but conversely the first
ligand oxidation of the former is easier than that of the latter
because of the higher electron density in the central TAP moiety
of the former. The first ligand reduction of CoTAP occurs, for
Figure 6. Simultaneous band deconvolution analysis of the Q absorption
and MCD bands of CoAnAP. Split Q00-bands are shown by bold solid lines,
while the experimental curves are shown as thin solid lines. The bottom
part shows the deviation of the simulated absorption spectrum from the
experimental absorption spectrum.
It is generally known that oxidized species aggregate more easily
than neutral species,33 and since it takes time for aggregation
to occur, the peak that becomes more prominent at higher scan
(34) Kadish, K. M. Recent Prog. Inorg. Chem. 1986, 34, 435. Kobayashi, N.;
Miwa, H.; Isago, H.; Tomura, T. Inorg. Chem. 1999, 38, 479.
(35) The standard oxidation potential of the Pc and TAP anion is represented
by Pc(-2) and TAP(-2). Thus, if L indicates a general ligand, the first
oxidized and first reduced species will be represented by L(-1) and
L(-3), respectively.
(33) Fuhrhop, J. H.; Wasser, P.; Riesner, D.; Mauzeral, D. J. Am. Chem. Soc.
1972, 94, 7996. Song, H.; Reed, C. A.; Scheidt, W. R. J. Am. Chem. Soc.
1989, 111, 6865. Schulz, C. Z.; Song, H.; Mislanker, A.; Orosz, R. D.;
Reed, C. A.; Debrunner, P. G.; Scheidt, W. R. Inorg. Chem. 1997, 36,
406.
(36) Kobayashi, N.; Nakajima, S.; Osa, T. Chem. Lett. 1992, 2415.
9
J. AM. CHEM. SOC. VOL. 124, NO. 27, 2002 8027

A R T I C L E S
Kobayashi and Fukuda
Table 2. Band Fitting Parametersa of the Q-Band Used for Deconvolution Analysis
-1
compd
ν/cm
λ/nm-1
ꢀmax
∆ν/cm-1
ꢀ
D
0
f
MCD int.b
type
ꢀΜ
103A (B )
103B /D
A /D
0
n
1
0
0
0
1
0
CoTAP
16329
612
39134
970
2474
7.58
0.17
-144009
-50010
-262869
129930
-160405
160975
A
B
B
B
B
B
B
B
A
B
A
B
A
B
1.92 × 103
12.6 × 103
1.66
-0.96
-6.30
-0.83
-4.65
2.02
-3.40
2.77
CoBzAP
CoNpAP
CoAnAP
CoBz2AP
CoNp2AP
CoAn2AP
15265
16029
14499
15901
13811
15419
14936
655
624
690
629
724
649
670
36615
41611
30728
37894
18305
31060
74657
1047
1020
991
1083
1004
1052
937
2674
2820
2235
2748
1418
2256
4986
8.19
8.63
6.84
8.41
4.34
6.91
0.18
0.20
0.14
0.19
0.08
0.15
0.32
-5.81
2.67
-3.55
3.55
-38.1
17.5
-23.3
23.3
-83593
74503
-1.96
1.64
-12.9
-2.97
1.56
10.8
15.3
-269354
-121379
-211222
-100770
-262892
-193741
3.67 × 103
24.1 × 103
1.58
1.93
1.37
-2.46
-16.1
-1.05
-1.23
-1.72
14102
13409
709
746
53496
73298
1036
814
4182
4739
12.8
14.5
0.25
0.27
3.76 × 103
24.7 × 103
-15.7
-2.39
3.02 × 103
19.8 × 103
-24.9
-3.79
a D (dipole strength) ) ꢀ /326.6. f is the oscillator strength. ꢀΜ 0 (n ) 0) and ꢀΜ 1 (n ) 1) are the zeroth and first moments of the MCD B- and
0
A-terms, respectively. A1 ) ꢀΜ 1/152.5 and B0 ) ꢀΜ 0/152.5. b MCD intensity is based on molar ellipticity per tesla.
Figure 7. Cyclic voltammograms of the cobalt complexes in this study in o-DCB containing 0.1 mol/L TBAP.
the same reason, at a slightly more positive potential than those
of ABBB and AABB type species. Plots in Figure 8 also indicate
that the first ligand reduction potentials among the ABBB type
species do not change much from species to species. This is
true also for AABB type species, although the potential is
slightly more negative than that of the ABBB type species,
indicating that the LUMO level does not sensitively change.
However, it is seen that the change in the first ligand oxidation
potential is more marked than that in the first ligand reduction
potentials. In addition, since the oxidation becomes easier with
increasing molecular size, it is concluded that the HOMO
destabilizes significantly with enlargement of the π system. This
is consistent with the hitherto predicted17a-c and accumulated37
data and is reproduced by MO calculations described in the
following section. The potential difference between the first
ligand oxidation of CoTAP and CoBzAP or CoBz2AP may still
appear large; however, there is a reason for this. The cobalt
center is first oxidized to CoIII, and this positive charge
delocalizes to some extent into the TAP ligand, thereby making
oxidation of the ligand more difficult.38 The first ligand
oxidations of CoBzAP and CoIII/II coincidentally occur experi-
mentally at the same potential, due to the positive shift of the
first ligand oxidation potential.
(iv) Molecular Orbital Calculations. To enhance our
understanding of the electronic spectra, we have performed MO
(38) A similar reason has been adopted to explain the relatively negative first
reduction potentials of cobalt Pc derivatives compared with those in nickel,
copper, and zinc derivatives.29 In this case, however, the first ligand
reduction of Pc appears after the cobaltII/I couple. Accordingly, partial
negative charge from the reduced cobalt delocalizes to the ligand, thereby
making the first ligand reduction more negative or difficult.
(37) Kobayashi, N.; Nakajima, S.; Osa, T. Inorg. Chim. Acta 1993, 210, 131.
9
8028 J. AM. CHEM. SOC. VOL. 124, NO. 27, 2002

Aromatic Ring-Fused Tetraazaporphyrins
A R T I C L E S
Table 3. Redox Potential Data (vs Fc+/Fc) in o-DCB Containing
0.1 M TBAPa
couple
E
1/2/V ∆E /mV
couple
E
1/2/V ∆E /mV
p
p
CoTAP
CoIIIL(-1)/CoIIIL(-2) 0.94 130
CoIIIL(-2)/CoIIL(-2)
0.68
CoIIL(-2)/CoIL(-2) -0.81 120
CoIL(-2)/CoI(-3)
-1.77 110
CoBzAP
CoIIIL(-1)/CoIIIL(-2) 0.48 350 CoIIIL(-1)/CoIIL(-1)
CoIIIL(-2)/CoIIL(-2) 0.48 350 CoIIL(-1)/CoIIL(-2)
CoBz2AP
0.50 170
0.25 165
CoIIL(-2)/CoIL(-2) -0.85 210 CoIIL(-2)/CoIL(-2) -0.87 150
CoIL(-2)/CoI(-3)
-1.85 210 CoIL(-2)/CoI(-3)
CoNp2AP
0.48 120 CoIIIL(-1)/CoIIL(-1)
-1.97 150
CoNpAP
CoIIIL(-1)/CoIIL(-1)
CoIIL(-1)/CoIIL(-2)
not obsd
0.32 120 CoIIL(-1)/CoIIL(-2)b 0.09
CoIIL(-2)/CoIL(-2) -0.86 120 CoIIL(-2)/CoIL(-2) -0.86 110
CoIL(-2)/CoI(-3)
-1.83 105 CoIL(-2)/CoI(-3)
-1.96 110
CoAnAP
CoAn2AP
CoIIIL(0)/CoIIIL(-1)
CoIIIL(-1)/CoIIL(-1)
CoIIL(-1)/CoIIL(-2)
0.75 180 CoIIIL(0)/CoIIIL(-1)
0.50 200 CoIIIL(-1)/CoIIL(-1)
0.28 155 CoIIL(-1)/CoIIL(-2)
0.70 160
0.31 135
0.01 140
CoIIL(-2)/CoIL(-2) -0.86 160 CoIIL(-2)/CoIL(-2) -0.88 130
CoIL(-2)/CoI(-3)
NiAnAP
-1.81 170 CoIL(-2)/CoI(-3)
NiAn2AP
-2.01 130
Figure 9. Partial energy diagram for the (pyrrole proton-) deprotonated
ABBB and AABB type species. Orbitals marked by / are either naphthalene-
or anthracene-centered orbitals. See the text concerning orbital 20b1 of
L(-1)/L(-2)
L(-2)/L(-3)
L(-3)/L(-4)
0.46 140 L(0)/L(-1)
-1.24 120 L(-1)/L(-2)
-1.56 115 L(-2)/L(-3)
L(-3)/L(-4)
0.57 130
0.20 130
-1.45 140
-1.81 140
Np2AP2-
.
It requires short computation times, and a great amount of data
have been accumulated in the past (calculations on macrocycles
slightly overestimate transition energies).14b The results de-
scribed here are mostly based on the PPP method, unless
otherwise noted.
a ∆Ep is the potential difference between cathodic and anodic peak
potentials at a sweep rate of 30 mV/s. L, ligand. b Data from differential
pulse voltammogram.
Partial MO energy diagrams for four frontier orbitals of
(pyrrole proton-) deprotonated ABBB and AABB type species
are shown in Figures 9 and 10, respectively, with calculated
transition energies, oscillator strengths (f), and configurations
summarized in Table 4. As described in Figure 9, all the
HOMOs and LUMOs+1 belong to the a2, while the LUMOs
belong to the b1 irreducible representations, respectively. In the
same figure, the orbitals marked by / represent the naphthalene-
or anthracene-centered orbitals. These orbitals, however, barely
contribute to the Q-band, as seen in the configurations in Table
4. Accordingly, in considering a four-orbital model40,41 in which
the optical spectrum can be interpreted in terms of two HOMOs
and LUMOs, these orbitals were not taken into account. In the
case of Np2AP, the 20b1 orbital might be considered to be, from
the size of the coefficients, a naphthalene-centered orbital, but
it was taken as the HOMO-1 since it had fair coefficients over
the central TAP moiety, and this orbital contributed substantially
to the Q-band (Table 4). Hence, we can consider that all the
HOMOs-1 belong to b1 symmetry. In the C2V point group, the
z-axis is an in-plane axis, and therefore all one-electron
transitions become allowed transitions, at least in one direction.
Figure 8. Electrochemically obtained redox data of cobalt derivatives in
o-DCB. L-2/-3 and L-1/-2 correspond to the ligand first reduction and
oxidation, respectively.
(39) Ishii, K.; Kobayashi, N.; Matsuo, T.; Tanaka, M.; Sekiguchi, A. J. Am.
Chem. Soc. 2001, 123, 5356. Mack, J.; Kobayashi, N.; Leznoff, C. C.;
Stillman, M. J. Inorg. Chem. 1997, 36, 5624.
(40) Gouterman, M. In The Porphyrins; Dolphin, D., Ed.; Academic Press: New
York, 1978; Vol. III, Part A, Chapter 1. Gouterman, M. J. Chem. Phys.
1959, 30, 1139.
(41) To our knowledge, the spectra of porphyrin or Pc derivatives can be
explained by Gouterman’s four-orbital model except for one case, which
was explained by a five-orbital model including two LUMOs and three
HOMOs (Kobayashi, N.; Sasaki, N.; Konami, H. Inorg. Chem. 1997, 36,
5674).
calculations within the framework of the PPP approximation
and ZINDO/S Hamiltonian. The ZINDO/S method is more
recent than the PPP method and has parameters for some metals,
but the computation time is long, and the calculations generally
underestimate transition energies.39 The PPP method has been
known for a long time and needs no parameters for the metals.
9
J. AM. CHEM. SOC. VOL. 124, NO. 27, 2002 8029

A R T I C L E S
Kobayashi and Fukuda
Figure 10. Four frontier orbitals of the (pyrrole proton-) deprotonated ABBB and AABB type species. Note that LUMO(1) and LUMO(2) of TAP2- are
degenerate, while the anthracene-centered orbitals in AnAP2- (i.e., HOMO-1), An2AP2- (HOMO-1), and An2AP2- (HOMO-2) are excluded, since their
contribution to the Q-band is small.
In the preceding section, it was found experimentally that
the LUMO energy remained almost constant while the HOMO
energy destabilized with increasing size of the fused aromatics.
As seen in Figure 9, this is almost perfectly reproduced,
confirming the reliability of the calculations (note in particular
that the potential change of the L-2/-3 couple in Figure 8, i.e.,
the first ligand reduction, correlates perfectly with the change
of the LUMO energies in Figure 9, while the potential change
of the L-1/-2 couple in Figure 8, i.e., the first ligand oxidation,
correlates nicely with the change of the HOMO energies in
Figure 9). Although this phenomenon concerning the HOMO
and LUMO levels on ring-expansion has been noted several
times in the past,17a-c,37 no one has succeeded in explaining it.
Indeed, this is a strange phenomenon, taking into account that
the energies of both the HOMO and LUMO levels of normal
hydrocarbons change to an almost comparable extent with
changing size of the π system.42 The resolution of this question
is given reasonably, however, for the first time from the careful
discernment of the MOs drawn in Figure 10. As typically seen
for a series of ABBB type species, the HOMOs spread over
the whole molecule, while the LUMOs are localized in the
central TAP moiety and have only small coefficients over the
fused aromatics. Accordingly, the LUMO energy is essentially
fixed, irrespective of the size of the π system, while the HOMO
energy destabilizes with increasing molecular size. Similar
localization is seen for the HOMO-1 levels (Figure 10), and
indeed the energy does not change sensitively with size (Figure
9). The LUMO+1 levels, on the other hand, have coefficients
over the whole molecule, and therefore they destabilize with
increasing molecular size (Figure 9), as for the HOMO. In the
cases of the disubstituted AABB type species, the above trend
is less obvious, but still the coefficients over the fused aromatic
moiety are larger for the HOMO than the LUMO levels, so
that the change of the energy level is larger for the HOMO
levels than for the LUMO levels. In addition, specifically for
the AABB type species, the increase in magnitude of the
coefficients of the LUMO and LUMO+1 levels is very similar
over each molecule, so that the energy shifts of the LUMO and
LUMO+1 levels on ring expansion are very similar and thus
remain almost degenerate. Thus, we can reasonably explain the
shift of MO levels on the basis of the magnitude of their
coefficients. The explanation and interpretation used here for
the behavior of the HOMO and LUMO levels based on the
coefficient of their molecular orbitals appear valid when applied
to other macrocyclic systems.
Figure 11 shows the calculated absorption spectra. For
monosubstituted ABBB type species, split Q-bands are calcu-
lated between 577 and 637 nm, shorter than experimentally
observed (624-724 nm). This is due, in part, to the large
substituent effect of the eight phenyl groups (estimated to be
(42) For example, in our calculations using the ZINDO/S Hamiltonian on
benzene, naphthalene, and anthracene, the energy of the HOMOs changed
from -9.825 to -7.025, and further to -5.514 eV, while that of the LUMOs
changed from 8.064 to 5.963, and further to 4.656 eV in the order of the
above compounds.
9
8030 J. AM. CHEM. SOC. VOL. 124, NO. 27, 2002

Aromatic Ring-Fused Tetraazaporphyrins
A R T I C L E S
Table 4. Calculated Transition Energies, Oscillator Strength (f), and Configurations for Deprotonated TAP Derivativesa
energy/eV
/nm
f
configurationsb
energy/eV
/nm
f
configurationsb
TAP2-
2.10
2.10
3.22
589 0.14 13f15 (67%) 12f14 (33%)
589 0.14 13f14 (67%) 12f15 (33%)
385 0.30
8f15 (37%) 11f15 (33%) 12f15 (10%)
13f14 (10%)
3.22
3.85
3.85
3.98
3.98
385 0.30
8f14 (37%) 11f14 (33%) 12f14 (10%)
13f15 (10%)
322 1.83 12f14 (31%) 11f14 (29%) 13f15 (13%)
12f15 (11%) 11f15 (10%)
322 1.83 12f15 (31%) 11f15 (29%) 13f14 (13%)
12f14 (11%) 11f14 (10%)
311 1.18
8f15 (38%)
12f15 (10%)
8f14 (38%)
12f14 (10%)
8f14 (19%) 11f15 (15%)
311 1.18
8f15 (19%) 11f14 (15%)
BzAP2-
Bz2AP2-
2.04
2.15
3.17
3.38
3.73
3.74
3.82
4.00
609 0.30 15f16 (76%) 14f17 (23%)
577 0.21 15f17 (69%) 14f16 (30%)
2.04
2.08
3.31
3.32
3.56
3.65
609 0.41 17f18 (78%) 16f19 (22%)
595 0.42 17f19 (79%) 16f18 (20%)
374 0.35 14f18 (50%) 15f19 (18%) 16f18 (16%)
373 0.37 15f18 (52%) 14f19 (20%) 16f19 (13%)
348 0.14 17f20 (85%)
340 1.07 16f19 (31%) 15f18 (17%) 11f19 (16%)
17f20 (11%) 17f18 (10%)
339 1.22 16f18 (34%) 12f19 (22%) 14f18 (16%)
11f18 (12%)
391 0.41 11f16 (66%) 14f16 (12%) 15f17 (11%)
367 0.29 13f17 (67%) 14f17 (14%)
332 1.56 13f16 (42%) 14f16 (34%) 15f17 (12%)
332 0.81 10f17 (40%) 14f17 (27%) 13f17 (17%)
325 1.39 13f16 (51%) 14f16 (21%) 11f16 (11%)
310 1.37 10f17 (36%) 14f17 (31%) 11f17 (13%)
3.66
3.88
319 0.69 11f19 (35%) 12f18 (16%) 14f19 (14%)
16f19 (13%) 15f18 (11%)
3.89
4.00
319 1.33 12f19 (38%) 16f18 (28%) 11f18 (13%)
310 0.77 14f19 (56%) 15f18 (13%)
NpAP2-
Np2AP2-
1.98
2.15
3.18
3.43
3.45
625 0.38 17f18 (80%) 16f19 (20%)
577 0.33 17f19 (72%) 16f18 (27%)
390 1.21 12f18 (44%) 16f18 (27%) 17f19 (15%)
362 0.28 14f19 (66%) 16f19 (17%)
359 1.02 12f18 (34%) 15f19 (25%) 17f20 (17%)
16f18 (13%)
1.95
2.06
3.26
3.29
3.49
3.56
3.62
3.72
3.76
3.99
4.06
4.07
634 0.62 21f22 (82%) 20f23 (13%)
603 0.61 21f23 (84%) 20f22 (11%)
381 0.96 19f23 (46%) 17f22 (15%) 21f26 (13%)
377 0.78 20f22 (36%) 16f22 (36%) 17f23 (10%)
355 0.75 16f22 (32%) 20f22 (25%) 18f23 (22%)
348 0.15 21f26 (73%) 19f23 (20%)
343 0.63 21f25 (46%) 20f23 (22%) 17f22 (14%)
333 0.38 21f25 (43%) 20f23 (35%) 17f22 (11%)
330 0.71 19f22 (64%)
311 0.41 18f23 (41%) 20f22 (17%)
306 0.12 17f23 (65%) 18f23 (11%)
305 0.69 16f23 (40%) 18f22 (23%)
3.62
3.66
3.75
3.93
4.06
342 0.14 15f18 (70%)
339 0.23 17f20 (60%) 14f18 (26%)
330 0.11 14f18 (71%) 17f20 (21%)
315 1.95 16f19 (49%) 15f18 (22%)
305 0.73 15f19 (57%) 16f18 (25%)
AnAP2-
An2AP2-
1.95
2.14
3.04
637 0.39 19f20 (80%) 17f21 (18%)
579 0.41 19f21 (71%) 17f20 (25%)
408 1.29 17f20 (27%) 13f20 (21%) 19f21 (18%)
18f21 (17%)
376 0.20 18f21 (35%) 13f20 (23%) 15f20 (23%)
360 0.20 16f21 (66%) 18f20 (10%) 17f21 (10%)
342 0.10 19f23 (72%) 16f20 (16%)
329 2.15 17f20 (30%) 16f20 (21%) 18f21 (15%)
13f20 (12%)
1.91
2.02
2.95
3.08
3.14
3.29
3.43
3.54
3.66
3.69
649 0.76 25f27 (83%) 22f26 (13%)
614 0.73 25f26 (85%) 22f27 (12%)
420 0.40 24f26 (65%) 22f26 (11%)
402 0.44 25f29 (74%)
394 0.90 24f27 (43%) 23f26 (18%) 22f27 (15%)
377 0.27 21f27 (34%) 24f26 (14%) 22f26 (13%)
362 0.25 24f27 (39%) 23f26 (26%) 22f27 (13%)
351 0.22 25f30 (44%) 23f27 (23%) 21f27 (13%)
339 0.54 22f26 (30%) 24f28 (19%) 21f27 (13%)
336 0.51 22f27 (35%) 24f29 (21%) 23f28 (12%)
20f27 (10%)
3.30
3.45
3.62
3.77
3.84
4.00
4.30
323 0.68 17f21 (38%) 18f22 (31%) 16f21 (15%)
310 1.42 18f22 (51%) 17f21 (18%)
289 0.15 15f21 (51%) 13f21 (42%)
3.86
321 1.69 24f29 (28%) 23f28 (16%) 23f26 (14%)
22f27 (10%)
3.87
4.04
320 0.95 23f27 (25%) 24f28 (25%) 23f29 (21%)
307 0.47 20f26 (71%) 21f27 (12%)
a Excited states with less than 4.1 eV and f > 0.10 are shown. b Orbital numbers 13, 15, 17, 19, 17, 21, and 25 are HOMOs of TAP2-, BzAP2-, NpAP2-
,
AnAP2-, Bz2AP2-, Np2AP2-, and An2AP2-, respectively. In NpAP2- and Np2Ap2-, the MOs numbered 15, 20 and 18, 24, and 25 are naphthalene-centered
orbitals in the above compounds order. Similarly, in AnAP2- and An2AP2-, MOs 18, 22 and 23, 24, 28, and 29 are anthracene-centered orbitals, respectively.
ca. 1260-1300 cm-1 19b,33,43
at shorter wavelength are estimated at 577, 577, and 579 nm
without much variation, in the order of BzAP2-, NpAP2-, and
AnAP2-, respectively, while those at longer wavelengths are
)
Of the split Q-band, the transitions
calculated at 609, 625, and 637 nm, in the above order,
reproducing the trend observed experimentally (Figure 4A). The
calculated splitting energies are 910, 1330, and 1570 cm-1, also
showing close resemblance to experiment (ca. 760, 1400, and
1610 cm-1, respectively). Experimentally, the relative intensity
of the longer wavelength component to the shorter wavelength
(43) Luk’yanets, E. A. Electronic Spectra of Phthalocyanines and Related
Compounds; NIOPIK: Moscow, 1989.
9
J. AM. CHEM. SOC. VOL. 124, NO. 27, 2002 8031

A R T I C L E S
Kobayashi and Fukuda
Figure 11. Calculated, within the framework of the PPP approximation, absorption spectra for the (pyrrole proton-) deprotonated species. Broken lines
indicate the Q-band obtained using the ZINDO/S Hamiltonian.
component was found to decrease from 0.90 to 0.74 and further
to 0.53 (Table 2) on going from CoBzAP to CoNpAP and
further to CoAnAP. Although this trend is reproduced, the
reproducibility of the absolute value is not good (1.43, 1.15,
and 0.95, respectively). For all ABBB type species, more than
90% of the Q-band is composed of the four one-electron
transitions from the HOMO and HOMO-1 to LUMO and
LUMO+1 (Table 4). The Q-bands to the shorter wavelength
are composed of the HOMOfLUMO+1 and the HOMO-
1fLUMO transitions. Since neither the HOMO-1 nor LUMO
shows significant energy variation on ring expansion, their
energy gaps are similar, being 5.28, 5.28, and 5.27 eV for
BzAP2-, NpAP2-, and AnAP2-, respectively. On the other hand,
both the HOMO and LUMO+1 energy levels increase at a
similar rate, so that their energy gap shows little variation on
ring expansion, i.e., 4.50, 4.44, and 4.37 eV in the above order.
Accordingly, the one-electron excitation energy constituting the
shorter wavelength Q-band remains almost constant, irrespective
of the size of the ligand. The Q-bands to longer wavelength are
composed of the HOMOfLUMO and the HOMO-1fLU-
MO+1 transitions. The energy gap between the HOMO and
LUMO decreases, while that between the HOMO-1 and
LUMO+1 increases with increasing size of the π system. The
magnitude of the contribution of a certain one-electron transition
to an observed transition can be estimated by the magnitude of
the product of the electronic transition moment and the
coefficient of the one-electron transition in the wave functions
of the excited states. For the HOMOfLUMO transition, the
values were 2.22 for both BzAP and NpAP and 2.12 for AnAP,
respectively. By calculating similarly for the HOMO-1fLU-
MO+1 transition, values of 0.92, 0.81, and 0.76 were obtained
in the order of the above compounds. Accordingly, the contribu-
tion of the HOMOfLUMO transition to the longer wavelength
Q-band is larger. Hence, it is concluded that the Q-band-shift
to longer wavelength is a result of a reduction of energy between
the HOMO and LUMO.
From the configuration in Table 4, the Soret bands of BzAP
and NpAP may correspond to the bands at 332 and 367 nm,
and 330 and 362 nm, respectively. However, since the Soret
band spectra are composed of a superimposition of broad bands
(Figure 4A), comparison with the calculation results is difficult.
In the case of AnAP2-, transitions arising from the anthracene
moiety overlap, so that it is even more difficult to determine
which transition corresponds to the Soret band.
Next, we consider the spectra of adjacently disubstituted
AABB type species. As shown in Table 4, the estimated splitting
of the Q-band is smaller than that in the ABBB type species
(390, 810, and 880 cm-1 for Bz2AP2-, Np2AP2-, and An2AP2-
,
respectively), while the intensities of the two bands are very
close, i.e., 0.42 and 0.41, 0.61 and 0.62, and 0.73 and 0.76,
respectively, in order of the above compounds. In Figure 5, the
Q00-band appeared degenerate and could be fitted using a single
Gaussian curve, while the Q MCD band was fitted as a Faraday
A-term. However, the calculation results here indicate that the
Q-bands split slightly, and therefore that the Q state is nearly
degenerate (according to the results on the basis of ZINDO/S
Hamiltonian, the above values (390, 810, and 880 cm-1) were
improved to ca. 50, 200, and 300 cm-1, respectively (see also
broken lines in Figure 11)).44 However, in the cases of AABB
type species, the energies of the LUMO and LUMO+1 are very
similar, so that energies composed of the HOMOfLUMO+1
and the HOMO-1fLUMO transitions and the HOMOfLUMO
and the HOMO-1fLUMO+1 transitions become nearly equal,
giving rise to the pseudo-degeneracy of the Q-band.45 The
(44) From the symmetry-adapted perturbation (SAP) method,17a,b only a shift
of the Q-band is expected without detectable energy splitting. However,
this is understood within the precision of the first-order perturbation (note
that the SAP is a first-order perturbation). A Q-band split predicted by the
MO calculations is not surprising since the PPP calculation was based on
the actual C2V symmetry and included much higher perturbation terms than
the SAP method. Since the results of MO calculations showed splitting of
the Q-band, the dispersion type Q-band MCD in Figure 5B is actually a
pseudo Faraday A-term, which is produced by the superimposition of closely
lying, interacting Faraday B-terms.24
9
8032 J. AM. CHEM. SOC. VOL. 124, NO. 27, 2002

Aromatic Ring-Fused Tetraazaporphyrins
A R T I C L E S
Figure 12. Development of the electronic absorption and MCD spectra
with time during the oxidation of CoAn2AP (top) at 0.70, NiAn2AP (middle)
at 0.90, and CoAnAP (bottom) at 0.90 V vs Ag/AgCl in o-DCB (0.3 M
TBAP) to ring-oxidized cation radicals. The directions of spectroscopic
changes are shown by the bold arrows. The MCD spectrum of the starting
neutral species is shown as a broken line, and that of the electrolyzed species
is shown as a solid line. Note the encircled magnification factor for the
MCD spectrum of oxidized species.
Figure 13. Development of the electronic absorption and MCD spectra
with time during the reduction of CoAn2AP (top) at -1.00, NiAn2AP
(middle) at -1.30, and CoAnAP (bottom) at -1.00 V vs Ag/AgCl in o-DCB
(0.3 M TBAP) to CoI species (CoAn2AP and CoAnAP) or a ring-reduced
anion radical (NiAn2AP). The directions of spectroscopic changes are shown
by the bold arrows. The MCD spectrum of the starting neutral species is
shown as a broken line, and that of the electrolyzed species is shown as a
solid line. Note the encircled magnification factor for the MCD spectrum
of reduced species.
destabilization of the HOMO on ring expansion is the cause of
the red shift of the Q-band.
Thus, as described above, the characteristics of the Q-band
of both ABBB and AABB type species can be rationalized in
the level of molecular orbitals.
(v) Spectroelectrochemistry. Spectroelectrochemistry has
been performed in order to facilitate the assignment of some of
the redox couples and to illustrate the characteristics of the
spectra of electrolyzed species of low symmetrical TAPs.
Figures 12 and 13 show the development of the spectra during
the first oxidation and reduction, respectively. In general, nickel
TAPs and Pc’s do not show couples due to nickel within the
potentials applied. Accordingly, the spectra of nickel species
can be used as exemplary data for the ring-oxidized or -reduced
species. During the course of the oxidation of NiAn2AP (Figure
12, middle), the Q-band lost intensity, and instead, broad bands
appeared at the longer wavelength side, at ca. 800-1050 and
1200-1500 nm. MCD showed a broad negative band corre-
sponding to the 800-1050 nm band (positive Faraday B-term).
These broad bands are produced by dimerization of cationic
radical species.46 The Soret band shifted to the red and lost
intensity slightly on oxidation. The spectroscopic change of
CoAn2AP on the first oxidation (Figure 12, top) is similar to
that seen for NiAn2AP, indicating that the oxidation is ligand-
centered. However, a different behavior was observed in the
(46) (a) Stillman, M. J. In Phthalocyanines: Properties and Applications;
Leznoff, C. C., Lever, A. B. P., Eds.; VCH: New York, 1993; Vol. 3,
Chapter 5. (b) Nevin, W. A.; Liu, W.; Greenberg, S.; Hempstead, M. R.;
Marcuccio, S. M.; Melnik, M.; Leznoff, C. C.; Lever, A. B. P. Inorg. Chem.
1987, 26, 891.
(45) Ballhausen, C. J. Introduction to Ligand Field Theory; McGraw-Hill: New
York, 1962.
9
J. AM. CHEM. SOC. VOL. 124, NO. 27, 2002 8033

A R T I C L E S
Kobayashi and Fukuda
region between the Soret band and the Q-band, where a small
band developed with electrolysis and an associated negative
Faraday A MCD-term was observed. A negative Faraday A
MCD-term is expected when a hole generated by electron
excitation has appropriate angular momentum.47 Accordingly,
this negative A-term might be produced by a transition from
the accidentally degenerate ligand orbital to cobalt (LMCT) or
by a transition from the degenerate cobalt orbitals to the ligand
orbital (MLCT). Of the two possibilities, however, the spectra
of CoAnAP support the latter possibility. This is because, as
seen in the bottom of Figure 12, the first oxidation of this species
is also ligand-centered, and a deformed negative Faraday A-term
was observed, corresponding to the new absorption between
the Q-band and the Soret band. Since it is very hard to consider
the presence of degenerate orbitals in a monosubstituted ABBB
type ligand, the possibility of the LMCT may be ruled out. Since
negative Faraday A-term type MCD spectra were observed
anyway for the ABBB and AABB type species, it is conjectured
that the ligand field directly surrounding the central metal is
close to D4h, irrespective of the lower symmetry of the ligand.
The spectroscopic changes on reduction of NiAn2AP (Figure
13, middle) are completely different from those observed for
CoAn2AP and CoAnAP (Figure 13 top and bottom), suggesting
that the reduction of the latter two is occurring at the cobalt, as
observed for CoPc in the same solvent.46,48 With the progress
of reduction, the Q-band and the Soret band shifted to the blue
and red, respectively, while a new peak developed at ca. 490
nm for CoAn2AP. In the case of CoIIPc, the Q-band shifts to
the red with the formation of CoIPc, and a band which has been
assigned as an MLCT develops at ca. 480 nm.46b The above
spectroscopic changes of CoAn2AP are, in this way, different
from those of CoPc. Accordingly, it is premature to give a
coherent interpretation on the spectra of the reduced cobalt
species.
spectroscopies, electrochemistry, and spectroelectrochemistry,
together with molecular orbital calculations.
(i) The shift and splitting (ABBB type species) or nonsplitting
(AABB type species) of the Q absorption band upon ring
expansion by fusion of aromatic molecules have been quanti-
tatively analyzed experimentally by adopting simultaneous band
deconvolution of the electronic absorption and MCD spectra.
In ABBB type species, the Q-band shifts and splits with fusion
of the aromatic rings, the extent of which increases with
increasing size of the fused aromatics. Of the split Q-band, the
intensity of the longer wavelength component decreases with
increasing size of the fused aromatics. In AABB type species,
the Q-band shifts without splitting due to the near degeneracy
of the two LUMOs. We are able to adjust the Q-band position
between ca. 600 and 750 nm in a stepwise manner by preparing
a series of CoTAP compounds.
(ii) The redox potentials of ABBB type species appear at
slightly more positive potentials than those of the corresponding
AABB type species, indicating that the fused aromatic rings
function as electron-releasing groups. Although the first ligand
reduction potentials of these species vary slightly, the first ligand
oxidation potentials show higher sensitivity to the size of the
macrocycles, while the potential shifts negatively with increasing
molecular size.
(iii) Both the spectroscopic results and the changes in redox
potentials have been reproduced well by molecular orbital
calculations within the framework of the PPP approximation.
In particular, from consideration of the delocalization of the
molecular orbitals obtained from the calculations, a reasonable
explanation has been given to the long-standing question that
the LUMO level remains almost constant while the HOMO level
destabilizes with increasing size of the macrocycle. Here, since
the HOMO is spreading over the whole moiety while the LUMO
is localized in the central TAP core moiety, irrespective of the
size of the molecule, the HOMO destabilizes while the LUMO
energy is essentially fixed on ring expansion.
Conclusions
We have succeeded in finding a method to obtain adjacently
disubstituted AABB type TAPs preferentially in a mixed
condensation reaction of two types (A and B) of ortho dinitriles.
The reaction is started in the presence of lithium, and the later
addition of a transition metal leads to the adjacently disubstituted
species. A series of these compounds and the corresponding
monosubstituted ABBB type compounds have been character-
ized by electronic absorption and magnetic circular dichroism
Acknowledgment. This research was supported by the Daiwa
Anglo-Japanese Foundation and Shorai Foundation for Science
and Technology to N.K., and a Grant-in-Aid for JSPS fellows
to T.F. (No. 13008393).
Supporting Information Available: Mass spectroscopic cal-
culated and observed isotope distribution patterns of all com-
pounds in this study (PDF). This material is available free of
charge via the Internet at http://pubs.acs.org.
(47) Michl, J. J. Am. Chem. Soc. 1978, 100, 6801; 6812.
(48) Kobayashi, N.; Lam, H.; Nevin, W. A.; Janda, P.; Leznoff, C. C.; Lever,
A. B. P. Inorg. Chem. 1990, 29, 3415.
JA012382U
9
8034 J. AM. CHEM. SOC. VOL. 124, NO. 27, 2002