Low symmetrical phthalocyanine analogues substituted with three crown ether voids and their cation-induced supermolecules

Article abstract of DOI:10.1021/ja950929x

Unsymmetrical phthalocyanine analogues with three 15-crown-5 ether voids at the 3,4-positions (MtNIL, MtBZ, MtNAP) (Mt = Zn, Cu) have been synthesized and characterized. Their dimerization is induced by addition of some cations, particularly Rb²⁺ and K⁺. Cofacial dimer formation in the presence of these cations proceeds in a two-step three-stage process, as indicated by absorption and emission spectroscopy. These cofacial species have a highly specific C(2v) eclipsed configuration providing well-defined dimeric species for spectroscopic analysis. The ESR spectra of the cation-induced dimeric copper derivatives show axial symmetry and may be analyzed in terms of interplanar separation of 4.2 ?. CuNAP alone forms a cofacial dimer even in the presence of Na⁺ or Cs⁺, and the estimated interplanar distances are 4.1-4.2 ? without depending on the size of cations. The ¹H NMR spectra of zinc dimers are consistent with a cofacial configuration. Magnetic circular dichroism (MCD) spectra of C(2v) type monomers can be interpreted as the superimposition of Faraday B-terms. Upper excited state (Soret, S₂) emission is observed for all zinc mononuclear species, and the quantum yield of S₁ emission is smaller than that of zinc phthalocyanine containing four crown units (ZnCRPc), suggesting that the lowering of molecular symmetry produces a decrease in quantum yield. Fluorescence decay of S₁ and S₂ emissions can be analyzed by mono- and biexponential fits, respectively. The zero field splitting (zfs) parameters, D, of the excited triplet (T₁) states of the monomers estimated from time-resolved EPR (TREPR) technique decrease in the order ZnNIL, ZnBZ, and ZnNAP, qualitatively indicating delocalization of excited π-electrons over additionally fused benzo and naphtho rings. A remarkable decrease in D value on dimerization is interpreted as an indication of delocalization of π-electrons over two macrocycles. Molecular orbital (MO) calculations within the framework of the Pariser-Parr-Pople (PPP) approximation succeed to reproduce the splitting, intensity, and the relative position of the Q absorption band of unsymmetrical monomers.

Full text of DOI:10.1021/ja950929x

J. Am. Chem. Soc. 1996, 118, 1073-1085
1073
Low Symmetrical Phthalocyanine Analogues Substituted with
Three Crown Ether Voids and Their Cation-Induced
Supermolecules
Nagao Kobayashi,* Mika Togashi,† Tetsuo Osa,† Kazuyuki Ishii,‡
Seigo Yamauchi,‡ and Hiroaki Hino§
Contribution from the Department of Chemistry, Pharmaceutical Institute, Institute for Chemical
Reaction Science, Tohoku UniVersity, Sendai 980-77, Japan, and School of Business
Administration, Ishinomaki Senshu UniVersity, Ishinomaki 986, Japan
ReceiVed March 21, 1995. ReVised Manuscript ReceiVed September 15, 1995X
Abstract: Unsymmetrical phthalocyanine analogues with three 15-crown-5 ether voids at the 3,4-positions (MtNIL,
MtBZ, MtNAP) (Mt ) Zn, Cu) have been synthesized and characterized. Their dimerization is induced by addition
of some cations, particularly Rb2+ and K+. Cofacial dimer formation in the presence of these cations proceeds in
a two-step three-stage process, as indicated by absorption and emission spectroscopy. These cofacial species have
a highly specific C2V eclipsed configuration providing well-defined dimeric species for spectroscopic analysis. The
ESR spectra of the cation-induced dimeric copper derivatives show axial symmetry and may be analyzed in terms
of interplanar separation of 4.2 Å. CuNAP alone forms a cofacial dimer even in the presence of Na+ or Cs+, and
1
the estimated interplanar distances are 4.1-4.2 Å without depending on the size of cations. The H NMR spectra
of zinc dimers are consistent with a cofacial configuration. Magnetic circular dichroism (MCD) spectra of C2V type
monomers can be interpreted as the superimposition of Faraday B-terms. Upper excited state (Soret, S2) emission
is observed for all zinc mononuclear species, and the quantum yield of S1 emission is smaller than that of zinc
phthalocyanine containing four crown units (ZnCRPc), suggesting that the lowering of molecular symmetry produces
a decrease in quantum yield. Fluorescence decay of S1 and S2 emissions can be analyzed by mono- and biexponential
fits, respectively. The zero field splitting (zfs) parameters, D, of the excited triplet (T1) states of the monomers
estimated from time-resolved EPR (TREPR) technique decrease in the order ZnNIL, ZnBZ, and ZnNAP, qualitatively
indicating delocalization of excited π-electrons over additionally fused benzo and naphtho rings. A remarkable
decrease in D value on dimerization is interpreted as an indication of delocalization of π-electrons over two
macrocycles. Molecular orbital (MO) calculations within the framework of the Pariser-Parr-Pople (PPP)
approximation succeed to reproduce the splitting, intensity, and the relative position of the Q absorption band of
unsymmetrical monomers.
Introduction
cyanines and porphyrins appears when they form cation-induced
cofacial dimers; the former exhibit eclipsed configurations, while
the latter dimers are always staggered because of their structural
limitation. Accordingly, it may not be particularly interesting
to prepare porphyrins containing three crown ether units.
Phthalocyanines with three crown ether units are, on the other
hand, more attractive since aromatic compounds of varying size
can be fused to the fourth position. That is, it may be possible
to realize an eclipsed configuration of C2V type phthalocyanine
analogues. The properties of mononuclear C2V type phthalo-
cyanine analogues per se have scarcely been elucidated, although
they seem to be important in fields, for example, using lasers
such as photodynamic cancer therapy and optical discs. Ac-
cordingly, if in addition, we can show the possibility of cofacial
dimer formation of such low symmetrical phthalocyanine
analogues, it may further invoke the attention of many research-
ers. Also, such research seems important to understand the
nature of phthalocyanines themselves. From such respect, we
report here the synthesis and solution properties of mono-
substituted C2V type phthalocyanines with three 15-crown-5 ether
voids (Figure 1). As shown below, spectroscopic properties of
these lower symmetrical Pc derivatives are often quite different
from those of Pcs with D4h symmetry, but their cation-induced
cofacial dimer formation was confirmed by several types of
spectroscopy. The results are compared with those of four
crown ether-linked phthalocyanine (MtCRPc) systems.1
There has been considerable interest in the properties of the
crown ether substituted phthalocyanines (Pcs)1 and porphyrins.2
One of the most important differences between these phthalo-
† Pharmaceutical Institute.
‡ Institute for Chemical Reaction Science.
§ School of Business Administration.
X Abstract published in AdVance ACS Abstracts, January 15, 1996.
(1) Some representative ones: (a) Kobayashi, N.; Nishiyama, Y. J. Chem.
Soc., Chem. Commun. 1986, 1462. (b) Koray, A. R.; Ahsen, V.; Bekaroglu,
O. J. Chem. Soc., Chem. Commun. 1986, 932. (c) Sielcken, O. E.; Van de
Kuil, L. A. Drenth, W.; Nolte, R. J. M. J. Chem. Soc., Chem. Commun.
1986, 1232. (d) Sirlin, C.; Bosio, L.; Simon, J.; Ahsen, V.; Yilmazer, E.
Bekaroglu, O. Chem. Phys. Lett. 1987, 139, 362. (e) Kobayashi, N.; Lever,
A. B. P. J. Am. Chem. Soc. 1987, 109, 7433. (f) Sielcken, O. E.; van Tilborg,
M. M.; Roks, M. F.; Hendriks, R.; Drenth, W.; Nolte, R. J. M. J. Am. Chem.
Soc. 1987, 109, 4261. (g) Ahsen, V.; Yilmazer, E.; Gurek, A.; Gul, A.;
Bekaroglu, O.; Gebze, T. HelV. Chim. Acta 1988, 189, 2533. (h) Ahsen,
V.; Yilmazer, E.; Ertas, M.; Bekaroglu, O. J. Chem. Soc., Dalton Trans.
1988, 401. (i) Gasyna, Z.; Kobayashi, N.; Stillman, M. J. Chem. Soc., Dalton
Trans. 1989, 2397. (j) Sielcken, O. E.; Vandekuil, L. A.; Drenth, W.;
Schoonman, J.; Nolte, R. J. M. J. Am. Chem. Soc. 1990, 112, 3086. (k)
Kobayashi, N.; Opallo, M.; Osa, T. Heterocycles 1990, 30, 389. (l) Gurek,
A.; Ahsen, V.; Gul, A.; Bekaroglu, O. J. Chem. Soc., Chem. Commun. 1991,
3367. (m) Miyamoto, R.; Yamauchi, S.; Kobayashi, N.; Osa, T.; Ohba, Y.;
Iwaizumi, M. Coord. Chem. ReV. 1994, 132, 57.
(2) Some representive ones: Kobayashi, N.; Osa, T. Heterocycles 1981,
15, 675. Thanabal, V.; Krishnan, V. J. Am. Chem. Soc. 1982, 104, 3463.
Thanabal, V.; Krishnan, V. Inorg. Chem. 1982, 21, 3606. Richardson, N.
M.; Sutherland, I. O.; Camilleri, P.; Page, J. A. Tetrahedron Lett. 1985,
26, 3739. Van Willigen, H.; Chandrashekar, T. K. J. Am. Chem. Soc. 1986,
108, 709. Gunter, M. J.; Johnston, M. R. Tetrahedron Lett. 1990, 31, 4801.
0002-7863/96/1518-1073$12.00/0 © 1996 American Chemical Society

1074 J. Am. Chem. Soc., Vol. 118, No. 5, 1996
Kobayashi et al.
containing one or two 15-crown-5 ether unit(s). In the subsequent
preparative TLC using methanol and CH2Cl2-pyridine as the eluting
solvents, the second green band was always collected, and after drying
the residue was dissolved in CH2Cl2 and washed with water. The
desired Pc analogues with three 15-crown-5 ether voids were collected
from the organic layer. Elemental analyses and fast atom bombardment
(FAB) mass data were satisfactory (Table 1). 500 MHz 1H NMR data
of monomeric zinc complexes in CDCl3 (δ) are as follows. ZnNIL:
3.76 (s, 12H, CH2), 3.83 (s, 12H, CH2), 4.09 (s, 12H, CH2), 4.55 (s,
4H, CH2), 4.61 (s, 8H, CH2), 7.55 (s, 2H, CH), 7.67 (s, 4H, CH), 8.32
(s, 4H, CH), 8.58 (s, 2H, CH), 8.79 (s, 4H, CH). ZnBz: 3.89 (s, 12H,
CH2), 3.94 (s, 12H, CH2), 4.19 (s, 12H, CH2), 4.44 (s, 12H, CH2),
7.20-8.12 (m, 18H, CH). ZnNAP: 3.16-4.66 (m, 48H, CH2), 6.92-
8.70 (m, 20H, CH).
(iii) Computational Method. The deprotonated H2NIL, H2BZ, and
H2NAP (i.e., NIL2-, BZ2-, and NAP2-) structures were constructed
by using standard phthalocyanine X-ray structural data7 and by making
the ring perfectly planar and adopting either D4h (BZ) or C2V (NIL and
NAP) symmetry. Molecular orbital (MO) calculations were performed
for the (pyrrole proton-) deprotonated dianionic species within the
framework of the Pariser-Parr-Pople (PPP) approximation,8a where
semiempirical parameters recommended in a recent book8b were
employed. These were atomic valence state ionization potentials of
11.16 (carbon), 20.21 (central nitrogen), and 14.12 eV (imino nitrogen),
together with atomic valence state electron affinities of 0.03 (carbon),
5.32 (central nitrogen), and 1.78 eV (imino nitrogen). The central
nitrogen atoms were assumed to be equivalent, supplying 1.5 electrons
each to the π-system. In addition, σ polarizability was taken into
account according to Hammond.8c Resonance integrals were taken to
be -2.48 (âCN) and -2.42 eV (âCC).8b Two-center electron repulsion
integrals were computed by the method of Mataga and Nishimoto.8d
The choice of configuration was based on energetic considerations, and
all singly excited configurations up to 48 393 cm-1 were included.
Figure 1. Structures and abbreviations of compounds appearing in
this study.
Experimental Section
(i) Measurements. Electronic spectra were recorded with a Shi-
madzu VU-250 spectrophotometer, and magnetic circular dichroism
(MCD) measurements were made with a JASCO J-720 spectrodichrom-
eter equipped with a JASCO electromagnet that produced magnetic
fields up to 1.53 T (T ) tesla) with parallel and then antiparallel fields.
Its magnitude was expressed in terms of molar ellipticity per tesla,
[Θ]M/104 deg mol-1 dm3 cm-1 T-1. Fluorescence and excitation spectra
were recorded with a Shimadzu RF-500 spectrofluorimeter with
appropriate filters to eliminate scattered light. Fluorescence quantum
yield (ΦF) was determined by the use of H2Pc and ZnPc (ΦF ) 0.60
and 0.30, respectively)3a and quinine sulfate in 1 N H2SO4 (ΦF ) 0.55
at 296 K)3b,c as standards. Data were obtained by a comparative
calibration method with use of the same excitation wavelength and
absorbance for ZnNIL, ZnBZ, and ZnNAP and the calibrants and the
same emission energies. Fluorescence decay curves were obtained at
20 °C by a Horiba NAES-550 series, using combinations of glass filters
and a monochrometer for monitoring the emission. The lifetimes were
determined from the decay curves by the use of the least-squares
method. All sample solutions for fluorescence measurements were
Results and Discussion
(i) Cation Complexation Leading to Cofacial Dimers.
Phthalocyanines with four crown-ether voids (MtCRPcs) are
known to dimerize by encapsulating cations such as K+ and
Ca2+. In these cases, the dimerization proceeds in a two-step
three-stage process,1e and the first dimer is noncofacial while
the last stage dimer is a rigidly cofacial eclipsed D4h species.
In electronic absorption spectroscopy, this was monitored as a
decrease of the Q band peak of monomer (around 660-700
nm) and a concomitant increase of absorption intensity to shorter
wavelength side (ca. 620-640 nm, the so-called dimer peak).
A similar phenomenon was observed for MtNIL, MtBZ, and
MtNAP in the present study. An example is shown in Figure
2, where the so-called monomer peak at 677 nm loses its
intensity, while a new peak appears at 634 nm and develops,
and the Soret band shifts to shorter wavelength with the increase
of [Rb+]. Considering that the blue-shift in the Q and Soret
absorption bands in the Pc dimers and oligomers can be
explained by an excitonic interaction9 and that the EPR data
for copper complexes in the presence of K+ and Rb+ will
indicate the presence of two copper atoms in very close
proximity, the final stage spectra can be ascribed to cofacial
dimers. Moreover, the lack of a new peak to the longer
1
purged with argon before measurement. 500 MHz H NMR spectra
were recorded with a Jeol GX-500 spectrometer using CDCl3 or CD2-
Cl2 containing a small amount of CD3OD as solvents. EPR spectra
were collected with a Varian E4 spectrometer of samples in chloroform
containing ca. 20% methanol at 77 K. The microwave frequencies
were monitored by a Takedariken TR 5501 frequency counter with a
TR-5023 frequency converter. TREPR spectra were observed for the
lowest triplet states at 77 K and 0.5 µs after the laser pulse. The samples
of 1 × 10-4 M in ethanol-chloroform (1:1 v/v) were excited at 600-
650 nm by a Spectra Physics MOPO-710 broad band OPO laser pumped
with a Spectra Physics GCR-170-10 Nd:YAG laser.
(ii) Materials. The so-called dicyanobenzo-15-crown-5 (2 equiv)1e
and diphenylmaleonitrile,4 3,6-diphenylphthalonitrile,5 or 2,3-dicyano-
1,4-diphenylnaphthalene6 (1 equiv) were fused in the presence of zinc-
or copper acetate (0.5 equiv) at ca. 250-260 °C for 20-30 min. After
cooling, the residue was washed well with water and methanol, and
separation of the mixture was carried out using columns and preparative
TLC on basic alumina. First, columns were eluted with CH2Cl2 and
acetone to remove unreacted starting materials and Pc analogues
(7) 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. E.; Dow, W.; Scheidt, D. R.
Inorg. Chem. 1976, 15, 1685.
(8) (a) Pariser, R.; Parr, R. G. J. Chem. Phys. 1953, 21, 466, 767. Pople,
J. A. Trans. Faraday Soc. 1953, 46, 1375. (b) Tokita, S.; Matsuoka, K.;
Kogo, Y.; Kihara, K. Molecular Design of Functional Dyes-PPP Method
and Its Application; Maruzen, Tokyo, 1990. (c) Hammond, H. Theo. Thim.
Acta 1970, 18, 239. (d) Mataga, N.; Nishimoto, K. Z. Phys. Chem. (Frankfurt
am Main) 1957, 13, 140.
(3) Seybold, P. G.; Gouterman, M. J. Mol. Spectrosc. 1969, 31, 1.
(4) Cook, A. H.; Linstead, R. P. J. Chem. Soc. 1937, 929.
(5) Mikhalenko, S. A.; Gladyr, S. A.; Luk’yanets, E. A. J. Org. Chem.
USSR (Engl. Transl.) 1972, 8, 341.
(6) Kovshev, E. I.; Luk’yanets, E. A. J. Gen. Chem. USSR (Engl. Transl.)
1972, 42, 1584.
(9) Dodsworth, E. S.; Lever, A. B. P.; Seymour, P.; Leznoff, C. C. J.
Phys. Chem. 1985, 89, 5698.

Phthalocyanine Analogues
J. Am. Chem. Soc., Vol. 118, No. 5, 1996 1075
Table 1. FAB Mass and Elemental Analytical Data
anal.(%)
calcd
H
found
H
compd
yield (%)
molecular formula
Mf for Mass
m/za
Mf
C
N
C
N
ZnNIL
CuNIL
ZnBZ
CuBZ
ZnNAP
CuNAP
2.27
1.78
2.65
4.98
3.30
2.57
C64H64O15N8Zn
C64H64O15N8Cu
C68H66O15N8Zn
C68H66O15N8Cu
C72H68O15N8Zn
C72H68O15N8Cu
1249.85
1248.00
1299.87
1298.02
1349.87
1348.03
1250
1248
1301
1299
1350
1348
1250.64
1248.81
1300.70
1298.87
1350.76
1348.93
61.46
61.55
62.79
62.88
64.02
64.12
5.16
5.17
5.11
5.12
5.07
5.08
8.96
8.97
8.61
8.63
8.30
8.31
61.27
61.04
62.67
62.48
63.61
63.76
5.37
5.39
5.40
5.29
5.38
5.34
8.73
8.66
8.42
8.27
8.05
7.99
a Parent ion peak.
[dimer].13 A region of slope 2 is the region where the
dimerization process is taking place. Indeed, as revealed for
ZnBZ at the bottom of Figure 3, a region with slope ) ca. 2.0
continues untill [Rb+]/[Pc analogues] ) ca. 0.5, indicating that
the monomer-dimer conversion proceeds with a very high
formation constant (K for eq 1 with n)1, (5.5 ( 2.0) × 109 L2
mol-2) until two Pc-analogue units bind one cation. After this
point, the slope of the plots approaches zero, especially beyond
the region of [Rb+]/[Pc analogues] > 1.5, revealing that
dimerization is not occurring in this region. Taking the above
information into account, it is easily conceivable that the linear
Pc analog-cation-Pc analog complex formed in 0 < [Rb+]/[Pc
analog] < 0.5 range transforms into a cofacial dimer by further
encapsulation of cations. Differing from the cases of MtCRPcs,
however, clear isosbestic points were not observed. This may
be due to the presence of several types of linear dimer, since
when a cation is trapped by two crown ether units belonging to
two different Pc analogues, several isomers are conceivable
depending on which crown ether is used.
Figure 2. Change of absorption spectrum of CuNIL by the addition
of CH3COORb to 3 mL of a CHCL3 solution in a 10-mm cell. The
CH3COORb in CHCl3-MeOH (9:1 v/v) was added with a microsy-
ringe; 30 µL were added in total. Arrows indicate the dirrection of
spectroscopic change.
(ii) Electron Paramagnetic Resonance (EPR). Figure 4
shows the ESR spectra of copper derivatives in the presence or
absence of cations in CHCl3-MeOH (4:1 v/v) at 77 K. EPR
spectra of Cu(II) ions have been intensively studied because of
their relative simplicity due to d9 configuration.14 The spectra
in the absence of any cations (for example curve a) are
structureless and seem to be those of somewhat aggregated
species.15 Addition of excess sodium ions ([Na+]/[Pc analog]
) ca. 10-50) led to the development of more structural signal
(curve b), basically similar to monomeric copper porphyrins
and phthalocyanines. In this case, the Na+ ions are trapped in
the cavity of the crown ether moiety, and the resulting positive
charge appears to keep the copper atoms separated and a clear,
highly resolved EPR spectrum is obtained.
Addition of Rb+ or K+ ions to the solution of copper
complexes induces dramatic changes in the EPR spectrum
(curves c-f). In the g ) 2 region, nitrogen super-hyperfine
coupling is not seen but two strong perpendicular transitions
accompanying the characteristic seven-line pattern due to two
equivalent Cu(II) ions (I ) 3/2) coupled together are yielded.16
In the half-field region (∆M ) (2), seven equally spaced lines
(90-115 G separation) are observed. These data provide
unmistakable evidence for the formation of a symmetric dimeric
molecule, and the absence of any EPR signal originating from
a monomeric copper derivative can be attested to a high dimer
formation constant. It seems difficult, in most cases, particularly
wavelength side of the monomer peak suggests theoretically10
that the spectrum at the final stage is that of a cofacial and
eclipsed configurated species.
In order to see the absorption intensity-cation concentration
relationship, Figure 3 was obtained from experimental data such
as those in Figure 2. In all cases, Rb+ and K+ are much more
effective than Na+ and Cs+ with respect to the amount by which
the intensity is altered. This is attributed to the size of cations.1
K+ and Rb+ ions are sandwiched between two crown ether units,
while Na+ resides within the ether ring, and Cs+ ion may be
too large to be sandwiched.11 The change almost saturates until
[Rb+] or [K+]/[MtNIL, MtBz, or MtNAP] values reach roughly
2, in particular the change of the monomer peak is most sharp
in the region of the above ratio of 0-1, indicating that
encapsulation probably continues until the available sites are
saturated, i.e. until two macrocyclic units bind three cations in
a rigid eclipsed dimer.
Assuming the monomer-dimer equiliblium (eq 1) for the Rb+
(or K+)-triggered
K
nK+ + 2monomer T dimer K+
(1)
n
spectroscopic change, the specific monomer and dimer con-
centrations were calculated by the method of West and Pearce12
from the changes of intensity at the monomer peak, as a function
of [K+]. With use of this procedure, proof of dimer formation
is obtained by plotting the calculated [monomer] against
(13) Nevin, W. A.; Liu, W.; Lever, A. B. P. Can. J. Chem. 1987, 26,
891.
(14) Manoharan, P. T.; Rogers, M. T. In Electron Spin Resonance of
Metal Complexes; Yen, T. F., Ed.; Plenum: New York, 1979.
(15) Skorobogaty, A.; Smith, T. D; Dougherty, G.; Pilbrow, R. J. Chem.
Soc., Dalton Trans. 1985, 651 and several references cited therein.
(16) Smith, T. D.; Pilbrow, J. R. Coord. Chem. ReV. 1974, 13, 173. Eaton,
S. S.; More, K. M.; Sawant, B. M.; Boymel, P. M.; Eaton, G. R. J. Magn.
Reson. 1983, 52, 435. Thomson, C. Q. ReV. Chem. Soc. 1968, 22, 45. Kottis,
P.; Lefebure, R. J. Chem. Phys. 1963, 39, 393.
(10) Gouterman, M.; Holten, D.; Lieberman, E. Chem. Phys. 1977, 25,
139.
(11) The cavity size of 15-crown-5 is estimated from CPK models to be
1.7-2.2 Å. The ionic diameters of the cations used in this paper are Na+
1.90 Å, K+ 2.66 Å, Rb+ 2.96 Å, and Cs+ 3.38 Å.
(12) Pearce, S.; West, W. P. J. Phys. Chem. 1965, 69, 1894.

1076 J. Am. Chem. Soc., Vol. 118, No. 5, 1996
Kobayashi et al.
Figure 3. Dependence of absorbance of ZnBZ, CuBZ, CuNIL, and CuNAP on cation concentration for the so-called monomer peak (mp) and
cofacial dimer peak (dp). Experiments were conducted as described in Figure 2. The wavelengths of mp and dp are shown in each figure, and the
initial intensity without cation is set at unity. Curves going up with [cation] represent changes at dimer peaks and should be referred to the right-
axis, while those going down are changes at monomer peaks and should be referred to the left-axis. Cations; Na+ (b), Cs+ (0), K+ (O), and Rb+
(4). Plots of log[ZnBZ] versus log([ZnBZ]2(Rb+)3) for ZnBZ in CHCl3 solution are shown at the botom. Experimental data such as those appeared
in Figure 2 were analyzed with a computer program based on the approximation method of Pearce and West.12 A line with a theoretical slope of
2 is added to show that dimerization is occurring in this region. An open arrow indicates the position of [Rb+]/[ZnBZ] ) 1.5. The inset shows the
dependence of monomer and/or dimer concentration of ZnBZ on [Rb+]/[ZnBZ].
in the presence of Rb+ or K+, to envision the presence of linear
oligomers in which Pcs are cofacial but joined to additional
Pcs above or below the plane of the crown ether substituents
(Scheme 1). For, if such species are present, a more complex
specrum would be generated due to the superimposition of
signals from “monomeric coppers” in Pcs pointed by arrows as
shown in Figure 4 curves a and b and those from dimeric
coppers as shown typically in curves e and f. In reality,
however, such a spectrum is not recorded except for CuNAP
plus saturated Na+ system (curve h).
The behavior of CuNAP is slightly different from that of
CuNIL and CuBZ. The latter two barely dimerize by the
addition of Na+ or Cs+. However, in the presence of a large
excess of Na+ ([Na+]/[CuNAP] > ca. 1000) or Cs+ ion ([Cs+]/
[CuNAP] ) ca. 100-150), CuNAP produces an EPR spectrum
characteristic of a symmetric dimeric molecule (curves g and
h). Although a trace of nitrogen super-hyperfine structure due
to the monomeric molecule is still seen in the g ) 2 region in
the presence of saturated Na+ (curve h), this high tendency for
cofacial dimerization of CuNAP compared with CuNIL and
CuBZ may be attributable to its large flat π-system. Generally,
larger flat π-systems have higher tendency toward cofacial
aggregation due to more intense π-π hydrophobic interaction
as seen typically in the comparison between phthalocyanines
and naphthalocyanines.17 The difference between CuBZ and
CuNAP is only one benzene ring. Accordingly, it is interesting
to note such a big difference between the two systems.
The EPR parameters for these spectra are collected in Table
(17) Tai, S.; Hayashi, N. J. Chem. Soc., Perkin Trans. 1991, 1275. Tai,
S.; Hayashi, N.; Katayose, M. Prog. Org. Coatings 1994, 24, 323.

Phthalocyanine Analogues
J. Am. Chem. Soc., Vol. 118, No. 5, 1996 1077
Figure 4. EPR spectra of copper complexes under various conditions in CHCl3-MeOH (4:1 v/v) at 77 K. (a) CuBZ, (b) CuBZ + Na+, (c) CuNIL
+ Rb+, (d) CuBZ + Rb+, and CuNAP with (e) Rb+, (f) K+, (g) Cs+, and (h) Na+. [Na+]/[CuBZ] ) ca. 100, [Rb+]/[Cu complex] ) ca. 5-10,
[K+]/[CuNIL] ) 10, [Cs+]/[CuNAP] ) ca. 120-130, [Na+]/[CuNAP] ) ca. 1000, and [Cu complex]/M ) ca. 0.001.
Table 2. Magnetic Parameters of Cation-Induced CuNIL, CuBZ, and CuNAP Cofacial Dimersa
compound
g1
g2
A1,G
Ah,G
D1,G
D2,G
freq,MHz
Cu-Cubdist, Å
[CuNIL]2(Rb+)3
[CuBZ]2(Rb+)3
[CuNAP]2(Rb+)3
[CuNIL]2(K+)3
[CuNIL]2(Cs+)3
[CuNIL]2(Na+)3
2.049
2.057
2.050
2.051
2.151
2.155
2.153
2.142
2.148
2.142
105
101
106
101
106
98
98
114
102
89
104
92
369
352
344
335
424
418
414
424
421
417
9177
9176
9175
9178
9180
9176
4.131
4.154
4.166
4.126
4.139
4.149
2.053
355
a See Figure 4e for definition of parameters. b Calculated via r3 ) (1.39 × 104) × g2/D2, where D is in G and r is in Å.
Scheme 1
presence and absence of cations. Figure 5 shows those of
ZnNIL. The spectrum of the mononuclear species in the
absence of cations (Figure 5B) is relatively simple and could
be assigned as shown, based on the integrated ratios of protons
and taking the ring current anisotropy of the azamacrocyclic
core into account. Of crown ether protons, those of 1 and 1′,
closest to the macrocyclic core, appeared at two positions (δ )
4.55 and 4.61 ppm, respectively) because of chemical inequiva-
lency due to the presence of phenyl rings. Comparing the
positions of the aliphatic proton resonance in ZnNIL (ca. 3.8-
4.6 ppm) and ZnCRPc (ca. 3.7-4.3 ppm),1e the ring current in
the former appears to be a little larger than the latter.
Qualitatively, this phenomenon might be explained by a
“descent-in-symmetry” model.19 It is known that the inner 16-
membered ring contributes most to the aromaticity of porphy-
rinic compounds. If we consider that the porphyrinic structure
is made by deforming an idealized 16-membered cyclic polyene
with the D16h symmetry, the ring current in porphyrins appears
to be smaller than that in polyene due to “bumps” in the ring,
such as acute corners and the presence of heteroatoms. Since
the ring current seems to decrease with increasing of interference
along this polyene, the ring current of ZnNIL appears to be
larger than that in ZnCRPc. In accordance with this argument,
it is already suggested that the ring current in naphthalocyanine
is smaller than in Pcs.20 That is, regarding that fused benzene
rings are interferences to the ring current of polyene, there are
three benzenes in ZnNIL while there are four in ZnCRPc.
The addition of RbCl to a solution of ZnNIL ([Rb+]/[ZnNIL]
2, having been evaluated by standard procedures.18 The
calculated Cu-Cu distance is 4.12-4.17 Å and is not sensitive
to the sizes of cations added. Since the diameter difference
between Na+ and Cs+ ions is 1.48 Å,11 the above insensitivity
of the Cu-Cu distance with the cation size suggests that two
crown ether units sandwiching cations are not necessarily in
parallel with each other (they may be opened outward looking
from the center of the macrocycles particularly when larger
cations are trapped).
(iii) 1H Nuclear Magnetic Resonance (NMR) Spectra. The
1H NMR spectra of zinc complexes were recorded in both the
(18) (a) Chikira, M.; Yokoi, H.; Isobe, T. Bull. Chem. Soc. Jpn. 1974,
47, 2208. (b) Chikira, M.; Kon, H.; Hawley, R. A.; Smith, K. M. J. Chem.
Soc., Dalton Trans. 1979, 246 and some references cited therein. (c)
Chasteen, N. D.; Belford, R. N. Inorg. Chem. 1970, 9, 169.
(19) Ceulemans, A.; Oldenhof, W.; Wahrand, C. G.; Vanquickenborne,
L. G. J. Am. Chem. Soc. 1986, 108, 1155.

1078 J. Am. Chem. Soc., Vol. 118, No. 5, 1996
Kobayashi et al.
Figure 5. 500 MHz proton NMR spectra of (A) [ZnNIL]2(Rb+)3 and
(B) ZnNIL in CD2Cl2. Curves (C) are expansion of (A).
Figure 6. Proposed structures for the cation-induced dimers of MtNIL,
MtBZ, and MtNAP. Solid circles indicate cations such as Rb+ and K+.
Although not well represented, the phenyl rings would be tilted from
the macrocyclic plane.
) ca. 2-10) increases the complexity in the whole region
(Figure 5A and C), and the signals spread out. These data are
consistent with cofacial dimer formation. Aromatic proton
resonances shift upfield slightly and become sharper, due
probably to the restricted rotaion of the O-CH2CH2-O groups
or to the greater rigidity imposed by encapsulation of the cations.
While the aliphatic protons are either unshifted (protons 2, 3,
and 4) or shifted to downfield only slightly (protons 1 and 1′).
These resonance shifts are consistent with the prediction by
calculation on the Pc ring current effect,21 assuming the
interplanar distance is ca. 4.1-4.2 Å.
As shown in the above “Materials” section, the NMR signals
of ZnBZ and ZnNAP were not well resolved compared with
those of ZnNIL even in the absence of cations. It is conceivable
that these compounds are, to some extent, in aggregated forms
in the NMR concentration range, as suggested by the ESR signal
in the absence of cations (concentrations for ESR and NMR
a line width of NMR signals would be broader than that of
monomers due to the presence of various types of linear dimers.
However, the line width of dimers in the presence of Rb+
(Figure 5A) is much sharper than in its absence (Figure 5B).
(ii) It is already well-known and substantiated that the ESR
spectra such as those shown in Figure 4 curves c-h are not
produced unless two equivalent copper(II) ions are in very close
proximity.18a,b In particular, the seven-line pattern in the half-
field (∆M ) (2) region is not observed if one copper skews
from directly above the other.
(iv) Electronic Absorption and MCD spectra. In Figures
7 and 8 are shown the electronic absorption and MCD spectra
of ZnNIL and ZnNAP, and Table 3 summarizes the UV-visible
absorption and MCD data of all compounds in this study. The
electronic absorption spectra of tribenzotetraazaporphyrins22 and
tribenzomononaphthotetraazaporphyrins22b,23 have seldomly been
reported and their MCD spectra are, of course, not found in the
literature. The absorption spectrum of monomeric ZnNIL
(Figure 7) is quite different from that of general MtPcs in that
four-peak Q bands are seen at 686, 638, 625, and 582 nm due
to the lowering of molecular symmetry to C2V. From the
interpeak energy difference (∆Ep), the bands at 686 and ca. 638
nm (∆Ep ) 1117 cm-1) are considered to belong to one couple
and those at 625 and 582 nm (∆Ep ) 1182 cm-1) to a second
couple. The ∆Ep difference of ca. 1100-1200 cm-1 is
appropriate as the spacing of vibrational bands in Pc com-
measurements are approximately the same order, i.e., 10-4
-
10-3 M). In the presence of Rb+ ([Rb+]/[ZnBZ] ) ca. 6);
however, each singlet ether proton resonance of ZnBZ splits
into triplets, suggesting and consistent with the formation of
cofacial dimers (Figure 6).
One referee pointed out that the NMR shown in Figure 5A
might be consistent with a dimeric species containing only one
cation opposite the two phenyl substituents. We would rather
like to deny this possibility for the following two reasons: (i)
If two ligands have only one cation, the fixation of the dimer
structure is evidently impossible. Under such a circumstance,
(20) Janson, T. R.; Kane, A. R.; Sullivan, J. F.; Knox, K.; Kenney, M.
E. J. Am. Chem. Soc. 1969, 91, 5210. Kane, A. R.; Yalman, R. G.; Kenney,
M. E. Inorg. Chem. 1968, 7, 2558. Koyama, T.; Suzuki, T.; Hanabusa, K.;
Shirai, H.; Kobayashi, N. Inorg. Chim. Acta 1994, 218, 41.
(22) (a) Elvidge, J. A.; Linstead, R. P. J. Chem. Soc. 1955, 3536. (b)
Kobayashi, N.; Kondo, R.; Nakajima, S.; Osa, T. J. Am. Chem. Soc. 1990,
112, 9640.
(23) Ikeda, Y.: Konami, H, Hatano, M: Mochizuki, K. Chem. Lett. 1992,
763.
(21) Johnson, C. E.; Bovey, F. A. J. Chem. Phys. 1958, 29, 1012.

Phthalocyanine Analogues
J. Am. Chem. Soc., Vol. 118, No. 5, 1996 1079
intensifies with an increasing number of alkoxy groups on Pc
periphery,1e,i this moiety may be ascribed to a transition
involving ether oxygen lone pairs, such as n-π*. The spectro-
scopic behavior of CuNIL resembles that of ZnNIL.
The electronic spectrum of ZnNAP is different from that of
ZnNIL in the Q band region but close to that of ZnPc. In this
way, the extent of departure from D4h symmetry seems small
in this compound. However, the effect of fusion of a naph-
thalene ring can be seen in the Q band region. In fact, the band
to the longest wavelength with a peak at 698 nm is not
symmetrical and suggests that at least one weak band is buried
in the longer wavelength side of this peak. Although not shown,
this phenomenon was more clearly detected in the spectra of
CuNAP. In this case, a shoulder appeared at 710 nm with ꢀ )
125 800, together with an intense (ꢀ ) 137 800) peak at 699
nm (Table 3). The splitting energy of this Q0-0 band in ZnNAP
(roughly ca. 280 cm-1) is much smaller than that in ZnNIL
(ca. 1420 cm-1). Another interesting characteristic on the Q
band is that, of the split two peaks, the one at lower energy is
stronger for MtNIL while weaker for MtNAP. The large
difference in the splitting of the Q bands and the difference in
the relative intensity of the split Q bands observed for MtNIL
and MtNAP will be reproduced later by MO calculations. The
apparently unsplit Soret (B1 + B2) and N bands appeared at
351 and 296 nm, respectively, and they are very close to those
of ZnBZ (353 and 292 nm). From the standpoint of symmetry,
these bands also should contain two bands each. The lack of
splitting in these bands implies that the splitting energies in the
Soret and N bands are also small, as in the Q band.
Figure 7. UV-visible absorption (bottom) and MCD (top) spectra of
ZnNIL monomer (solid lines), ZnNIL-Rb+-ZnNIL linear dimer (dotted
lines), and [ZnNIL]2(Rb+)3 cofacial dimer (broken lines) in CHCl3.
The spectra of ZnBZ and CuBZ are very close to those of
MtPcs with D4h symmetry, since their π-conjugated system has
D4h symmetry. However, because of the low symmetrical
distribution of substituent groups on Pc periphery, the Q0-0
bands of these compounds are less symmetrical with respect to
their peak positions, and in the case of CuBZ, a weak shoulder
is seen to the red (685 nm) of the main Q0-0 band at 679 nm
(Table 3, the splitting energy ) ca. 130 cm-1).
pounds.24 The apparently intense peak at 638 nm is then
considered to be the Q0-1 vibrational peak which was
strengthened by the overlap of a peak at 625 nm, and the weak
582 nm peak is the Q0-11 vibrational component of the band at
625 nm. Reflecting a big splitting of the Q0-0 band (1423
cm-1), the total Q bandwidth of ZnNIL (ca. 3500 cm-1) is much
broader than that of usual ZnPc (for example, ca. 2600 cm-1
in ZnCRPc).1e In the Soret (or B1 + B2) band region, four
bands are detected at 364 (sh), 353, 303, and 273 nm, in addition
to a small curvature between ca. 430 and 470 nm. In the case
of general MtPcs with D4h symmetry, the B1 and B2 bands
appear at ca. 300-400 nm1i,25 with separations of ca. 2000-
7000 cm-1, together with a weaker band at ca. 260-300 nm
named the N band.24a Thus, a shoulder at 364 and a peak at
353 nm appear to be resulted of the complex superimposition
of four bands generated by the splitting of both the B1 and B2
bands, while two peaks at 273 and 303 nm are considered to
correspond to the split components of the N band. The 303
nm peak is too small to assign to the split components of the
B2 band. Because of the low intensity, the assignment of the
curvature between ca. 430 and 470 nm is ambiguous. If we
take into account that the B1 band of NiCRPc was found at
410 nm,1i this moiety may correspond to a split component of
B1 band. However, since the alkoxy group-substituted Pcs
commonly show a curvature in this region, and since this band
On cofacial dimer formation ([Rb+]/[ligand] ) ca. 4-6), the
band due to the monomer species decreases in intensity, while
there is growth in the band due to the cofacial dimeric species.9,10
That is, the Q0-0 band of monomers decreased and concomi-
tantly a broad new peak developed to the blue, as shown in
Figure 2 for CuNIL system. This phenomenon can be explained
qualitatively as a result of the interaction among excited states
as follows.10 Namely, if we have nondegenerate dipolar allowed
states X, Y on moiety A interacting with a similar pair X′, Y′
on moiety B, the nature of exciton coupling depends on the
detailed dimer geometry. However, to a first approximation,
the four states interact in pairs. Thus, the four resulting dimeric
states are roughly of the form X ( X′ and Y ( Y′ except
coefficients. In the case of a cofacial eclipsed dimer, the
expected geometry would have the transition dipoles of two
pairs of states, say X and X′ and Y and Y′, perpendicular to
the vector R between the centers; the allowed exciton state may
then be unshifted or blue shifted. This accounts for the presence
of two bands in each of the Q, B1, B2, and N bands, although
the splitting of bands is not clearly seen in the actual spectra.
The exciton splitting energies estimated as twice the difference
between the Q band absorption of dimers and the average energy
of the split Q bands of monomers are 567, 749, 2176, 1924,
2445, and 2497 cm-1 for ZnNIL, CuNIL, ZnBZ, CuBZ, ZnNAP,
and CuNAP, respectively, indicating that the exciton energy
increases with increasing molecular size. Since the center-
center distance of these dimers is the same, this fact further
indicates that transition dipoles become larger with increasing
(24) (a) Stillman, M. J.: Nyokong, T. Phthalocyanines-Properties and
Applications; Leznoff, C. C., Lever, A. B. P., Eds.; VCH Publishers:
Weinheim, Germany, New York, 1989; Chapter 3. (b) Luk’yanets, E. A.
Electronic Spectra of Phthalocyanines and Related Compounds; Tch-
erkassy: Moscow, 1989. (c) Kobayashi, N. Phthalocyanines-Properties and
Applications; Leznoff, C. C., Lever, A. B. P., Eds.; VCH Publishers:
Weinheim, Germany, New York, 1993; Vol. 2, Chapter 3. (d) Kobayashi,
N.; Nakajima, S.; Osa, T. Inorg. Chim. Acta 1993, 210, 131.
(25) Nyokong, T.; Gasyna, Z.; Stillman, M. J. Inorg. Chem. 1987, 26,
1087. Ough, E.; Nyokong, T.; Creber, K. A. M.; Stillman, M. J. Inorg.
Chem. 1988, 27, 2724.

1080 J. Am. Chem. Soc., Vol. 118, No. 5, 1996
Kobayashi et al.
Table 3. Electronic Absorption and MCD Data in Chloroform
electronic absorption l/nm(10-4ꢀ/M-1 cm-1
)
MCD l/nm (10-5[Θ]M/deg dm3 mol-1 cm-1 T-1
)
compd
ZnNIL
686(8.50)
364(5.57)sh
641(5.08)
638(5.73)
353(6.47)
593(2.66)
625(5.22)sh
303(2.85)
344(6.28)
582(1.76)
273(3.28)
272(4.34)
691(-7.61)
307(-0.14)
654(-5.62)
335(1.01)
687(-7.58)
352(-0.48)
651(-2.22)
287(0.18)
622(6.68)
288(0.31)
626(3.46)
305(-0.04)
673(3.96)
321(0.33)
614(1.61)
372(-1.35)
344(1.12)
366(-0.95)
612(2.03)
323(0.25)
362(-0.77)
632(3.41)
351(-0.52)
620(2.19)
634(2.62)
358(-0.29)
348(-0.97)
323(0.37)
[ZnNIL]2(Rb+)3
CuNIL
593(4.32)
281(0.30)
657(-0.39)
287(0.24)
677(6.72)
292(3.45)
634(3.54)
641(2.85)
276(3.40)
332(3.25)
624(2.51)sh
287(3.29)
353(4.48)
342(3.61)
616(1.75)
640(2.93)
351(5.97)
345(5.53)
631(3.94)
415(2.05)
338(4.08)
274(3.47)
292(2.39)
290(2.28)
341(3.81)
337(3.03)
296(4.17)
293(4.45)
418(2.25)
339(6.62)
[CuNIL]2(Rb+)3
ZnBZ
348(-0.28)
678(9.22)
279(2.60)
673(2.73)
278(2.76)
685(6.25)sh
274(4.50)
698(2.36)sh
271(4.47)
698(11.53)
613(1.67)
637(3.98)
679(6.85)
677(2.83)
628(2.59)
649(4.27)
691(-13.2)
675(9.69)
616(2.67)
651(-2.38)
617(2.10)
659(-1.02)
670(2.12)
638(2.31)
333(0.43)
[ZnBZ]2(Rb+)3
CuBZ
693(-2.84)
368(-0.32)
694(-7.37)
324(0.42)
700(-3.07)
350(-0.37)
712(-14.9)
360(-0.55)
712(-1.99)
333(0.36)
678(0.14)
330(0.43)
678(5.24)
[CuBZ]2(Rb+)3
ZnNAP
681(0.30)
323(0.37)
698(11.57)
334(0.42)
676(-1.69)
284(0.12)
700(10.81)
293(0.24)
637(2.69)
272(-0.07)
[ZnNAP]2(Rb+)3
CuNAP
694(3.57)
710(12.58)sh
342(5.59)
693(5.19)
297(6.12)
699(13.78)
299(6.44)
650(5.80)
716(-15.84)
636(3.24)
275(-0.09)
350(-0.52)
259(0.05)
322(0.51)
[CuNAP]2(Rb+)3
695(-2.06)
290(0.19)
molecular size of constituent monomers. In accordance with
this result, the Q band intensifies in the order tetraazaporphyrin,
Pc, naphthalocyanine, and anthracocyanine.24c,d In the cases
of ZnCRPc and CuCRPc,1e these values were 1954 and 1910
cm-1, respectively. The Soret band also shifts to the blue with
concomitant decrease in absorption coefficient. Although the
shift is only 3-11 nm in nm unit, this is not small in cm-1 unit
(ca. 260-910 cm-1).
MCD spectra of monomeric species, i.e., in the absence of
cations, can be interpreted at least theoretically as the super-
imposition of Faraday B-terms,26 showing minima and maxima
corresponding roughly to the absorption maxima, because no
degenerate states are included (however, MtBZs show a Faraday
A-term-like pattern due to the D4h symmetry of the chro-
mophore). The dispersion type (the A-term-like) MCD curves
corresponding to the Q0-0, Soret, and N bands are produced as
the superimposition of adjacent interacting B-terms. This is
most clearly seen in the spectrum of ZnNIL (Figure 7). MCD
troughs are observed at 691, 372, and 307 nm corresponding
to the absorption peaks at 686, 364 (sh), and 303 nm,
respectively, while MCD peaks are at 622, 344, and 288 nm
associated with the absorption peaks at 622, 353, and 273 nm,
respectively. The curvature seen in 400-500 nm of the
spectrum of ZnNIL and in 370-500 nm of that of ZnNAP
indicates that there are indeed weak transitions in these regions.
These are perhaps associated with either the split components
of the B1 band in MtPcs with D4h symmetry or transitions
involving ether oxygen lone pairs, as mentioned for the spectrum
of ZnNIL.
By the addition of Rb+ required to make cofacial dimers,
the MCD band also shifted to the blue with a concomitant
decrease in intensity. A small negative MCD trough detected
at the Q band to the longest wavelength suggests that a small
amount of monomeric species may be still present in the solution
(not explicit for ZnNIL but clearly seen for ZnNAP at 712 nm).
The positions of the inflection points of dispersion curves are
close to the absorption peaks of the main Q, B1 + B2, and N
bands. However, these are perhaps being produced as the
superimposition of Faraday B-terms of opposite signs, although
Figure 8. UV-visible absorption (bottom) and MCD (top) spectra of
ZnNAP monomer (solid lines) and [ZnNAP]2(Rb+)3 cofacial dimer
(broken lines) in CHCl3.
no one can rule out the possibility of the Faraday A-terms
completely. According to Kaito et al.27 who constructed the
“apparent” A-term using band shape functions assuming the
damped oscillator model, the dispersion type curve could easily
be produced corresponding to the seemingly single broad
absorption peak when the two bands of opposite signs were
separated by 1000 cm-1 and the bandwidth at half height (Γ)
(27) Kaito, A.; Nozawa, T; Yamamoto, T.; Hatano, M; Orii, Y. Chem.
Phys. Lett. 1977, 52, 154.
(26) Tajiri, A.; Winkler, J. Z. Naturforsh. 1983, 38a, 1263.

Phthalocyanine Analogues
J. Am. Chem. Soc., Vol. 118, No. 5, 1996 1081
Table 4. Quantum Yields (φF) and Lifetimes (τ) of ZnNIL, ZnBz,
and ZnNAP in Deaerated Chloroform at Room Temperaturea
S1
S2
τ1/ns(%)
compd
φF
τ/ns
103φF
τ2/ns(%)
ZnNIL
ZnBZ
ZnNAP
0.09
0.16
0.09
2.23
3.12
3.46
8.87
6.54
8.49
5.94(29.4)
3.28(48.5)
3.31(27.5)
19.4(70.6)
25.5(51.5)
26.2(72.5)
a Emission wavelengths for recording S1 and S2 excitation spectra
were 715, 705, and 725 nm and 460, 460, and 470 nm, respectively, in
the order ZnNIL, ZnBZ, and ZnNAP. Because of solution instability,
φF values of S2 emission were obtained using the first spectra recorded
after deaeration.
compounds exhibited the so-called S1 and S2 emission, as have
been found commonly for alkoxy group-substituted Pcs.1e,30 The
S2 emission is much broader than the S1 emission, and the band
shape is similar to that observed for tetraneopentoxy group-
substituted Pcs.30 In each case, the Stokes shift of the S1
emission is very small, while that of the S2 emission is generally
large. In the case of the S1 emission of ZnNIL, the correspond-
ing excitation spectrum indicates that the emission occurs mainly
from the lower energy component of the split Q band, while
the two peaks (at 702 and 709 nm) in the S1 excitation spectrum
of ZnNAP substantiate experimentally that the Q band of this
species is split and therefore that the molecular symmetry of
ZnNAP is lower than D4h.
(29) 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 center and bandwidth parameters. In fact, to date many
absorption and MCD spectra have been analyzed by band deconvolution
and it has been successful in analysing transitions which had been postulated
by MO calculations or in finding a new degenerate excited state (some
representative ones are refs 25, 28, and follows: (a) Browett, W. R.; Gasyna,
Z.; Stillman, M. J. J. Am. Chem. Soc. 1988, 110, 3633. (b) Gasyna, Z.;
Kobayashi, N.; Stillman, M. J. J. Chem. Soc., Dalton Trans. 1989, 2397.
(c) Ough, E. A.; Stillman, M. J.; Creber, K. A. Can. J. Chem. 1993, 71,
1898. (d) Ough, E. A.; Stillman, M. J. Inorg. Chem. 1994, 33, 573). It is
important to underscore the fact that both sets of spectra (that is, the
associated absorption and MCD spectra recorded from the same solution)
must be fitted with the same set of parameters. Of course band deconvolution
has to be carefully performed to the extent that removal of only one of the
fitted bands results in fits that clearly do not realistically account for all
the intensity in one or the other of the two spectral envelopes. Accordingly,
we made a program so that the fitting procedures are based upon the
assumption that the bandwidths and band centers will be identical in both
the absorption and MCD spectra. As introduced in previous papers (refs
25, 28, and above), whenever an A-term was used, a B-term was also added
in, and this B-term had exactly the same parameters as the A-term. This
home made program uses least-squares and Simplex iteration routines to
obtain the best fit between calculated and experimental data by varying the
band centers and bandwidths of the Gaussian-shaped bands that are assumed
to make up the spectra. Also, we took the recently identified result into
account that the bandwidth generally becomes larger the higher the energy
of the band.28 When A-terms dominate the MCD spectrum, the band
assignment is straightforward. However, since in the present case, both the
ground and excited states are nondegenerate, the spectra were deconvoluted
entirely with Gaussian-shaped B-terms. Analysis was performed for dimeric
spectra of copper complexes, since their EPR spectra in the presence of
Rb+ clearly indicated that they exist as dimers. For CuNIL, CuBZ, and
CuNAP, 23, 22, and 22 components are required, respectively. The
“apparent” A-term-like, dispersion curves are reproduced at ca. 15 670 and
29 860 cm-1 for CuNIL, 15 690 and 29 650 cm-1 for CuBZ, and 15 080
and 29 530 cm-1 for CuNAP, supporting approximate red-shifting with
increasing molecular size. The inversion of the Q band position between
CuNIL and CuBZ would be ascribed to the large splitting of the Q band of
the former. We dared to deconvolute the MCD of the above derivatives
introducing A-terms also. In this case, A-terms are required at 15 720 (band
half width ) 508 cm-1) and 29 919 (1456) cm-1 for CuNIL, 15 505 (899)
and 29 857 (1328) cm-1 for CuBZ, and 15 181 (893) and 29 506 (1287)
cm-1 for CuNAP. As reiterated in the text, A-terms are not present
theoretically. However, if this were the case, this then means that the MCD
spectra of cofacial C2V species are insensitive to the molecular symmetry
of constituting monomers. We believe that no one can deny this possibility
completely because the spectroscopic shape is similar among species, and
the spectra contain little structures (see for example the Q band region).
(30) Kobayashi, N.; Lam, H.; Nevin, W. A.; Leznoff, C. C.; Koyama,
T.; Monden, A.; Shirai, H. J. Am. Chem. Soc. 1994, 116, 879.
Figure 9. Fluorescence emission (solid lines) and excitation spectra
(broken lines) of ZnNIL (top), ZnBZ (middle), and ZnNAP (bottom)
in CHCl3. In each species, there is no intensity relationship between
S1 and S2 emissions. Excitation wavelengths and emission wavelengths
used to record excitation spectra are shown. ΦF is quantum yield. The
absorbance at the excitation wavelengths was always less than 0.05.
was set at 2000 cm-1, i.e., when the separation of the two bands
are a half of Γ. In this particular case, the Γ of the resultant
absorption band became roughly ca. 3000 cm-1 and the energy
difference between MCD trough and peak (∆E(trough-to-peak))
reached ca. 1600 cm-1. This means that the “apparent” A-term
is of course produced from two B-terms of opposite signs, if
their separation is less than half of the bandwidth at half height.
In the case of usual MtPc, the Γ values of monomers are roughly
ca. 700-1500 cm-1 in the Q band and 1200-2000 cm-1 in
the Soret band regions (in general, the bandwidth of the spectra
of aromatic compounds broadens the higher the energy of the
bands).28 In our experimental spectra of cofacial dimers, Γ and
∆E(trough-to-peak) values in the Q band region are ca. 1200-
1400 and 680-1300 cm-1, while those in the Soret band are
ca. 3300-4800 and 2100-2500 cm-1, respectively. If we admit
that the Q and Soret bands of C2V compounds theoretically
involve two and four transitions (the split B1 and B2 compo-
nents), respectively, the above data strongly suggest that Faraday
A-term-like MCD spectra can be produced as the superimposi-
tion of B-terms in the spectra of MtNIL and MtNAP.29
(v) Emission Spectra. We measured the fluorescence
emission and excitation spectra of three zinc complexes. The
spectra are shown in Figure 9, and the obtained values of
quantum yield (φF) and lifetimes (τ) are listed in Table 4. All
(28) Mack, J.; Stillman, M. J. J. Am. Chem. Soc. 1994, 116, 1292.

1082 J. Am. Chem. Soc., Vol. 118, No. 5, 1996
Kobayashi et al.
Figure 11. Intensity decrease of S1 emission of Zn complexes upon
addition of Na+ (b) or Rb+ (O) in CHCl3. The intensity in the absence
of cation was set at unity.
Figure 10. S2 fluorescence decays of ZnNIL and ZnNAP in deaerated
CHCl3 and their typical biexponential fit. For lifetimes (τ), see Table
4.
upon cation addition (Figure 3). Accordingly, in the latter
region, rigid eclipsed dimers are present and a weak, constant
residual fluorescence is due to the small amount of monomer
species in equilibrium with the cofacial species, indicating that
cofacial species do not emit or emit very weakly, if any.34
The effect of Na+ is of course less than Rb+, but it is still
effective. Quenching occurs efficiently in the 0 < [Na+]/
[ligand] < 1 region, and above [Na+]/[ligand] ) ca. 1.5 or 2
the intensity becomes almost constant. The residual intensity
has an excitation spectrum identical with that of monomer
species and therefore is perhaps attributable to some amount of
monomer species in equiliblium with cofacial dimers. The
larger residual intensity in Na+ added systems compared to Rb+
systems would imply that monomers are more abundant in the
former systems. Although not shown, the residual intensity
changes only slightly in the 2 < [Rb+] or [Na+]/[ligand] < 20
region, indicating that further oligomerization does not occur
in this region. Also, it is interesting to note that the relative
residual fluorescence intensity at a fixed [Na+]/[Zn complex]
becomes smaller with increasing area of the π-systems. For
example, at [Na+]/[Zn complex] ) ca. 2, the values are 0.57,
0.50, and 0.42 for ZnNIL, ZnBZ and ZnNAP, respectively.
(vi) Time-Resolved EPR (TREPR) Spectra. TREPR
spectra could be observed for all the zinc complexes and typical
examples are shown in Figure 12. These spectra come from
polarized excited triplet (T1) states produced by anisotropic
intersystem crossing (isc) (Pi; i ) x, y, z) from S1. The
polarization pattern of A/E, A at lower stationary fields (h/x, y,
z) and E at higher stationary fields (A and E denote absorption
and emission of microwaves, respectively), is the same as those
of other zinc Pc and porphyrin (Por) complexes.1m,35 The
spectra were simulated with Pi’s and zero field splitting (zfs)
parameters, D and E, under consideration of random orientations
of molecules in glass. The simulated spectra (g ) 2.0) and the
obtained values are shown in Figure 12 and Table 5. It was
found that both a monomer and the dimer contributed to the
observed spectra of ZnNIL and ZnNAP, but the dimer spectrum
was clearly separated from the monomer one for ZnBZ. The
D value () -3z/2; z is the direction toward an out of Pc plane)
of the monomer becomes smaller in the order of ZnNIL, ZnBZ,
and ZnNAP. For the E value () |x - y|/2; a measure of
anisotropy within the ring plane), a slightly larger value was
obtained for ZnBZ. On dimerization, remarkable decrease in
D was observed.
Concerning the quantum yield of S1 fluorescence emission,
it is known already that it decreases with lowering of molecular
symmetry.31 Indeed, the φF values are fairly small compared
with φF () 0.30) of ZnPc without substituent groups.3 Although
the π-conjugated system of ZnBZ is approximated by D4h
symmetry, its φF value is already about half of that of ZnPc
(0.30), suggesting that the substituent groups on ZnBZ affects
the effective symmetry.
Knowledge on S2 emission of Pcs is still very limited.
According to two papers so far published,30,32 φF values of S2
emission are of the order of 10-4 and in order to obtain a good
fit of the decay curves, double-exponential functions are always
required. This holds in the present systems also. Figure 10
shows fluorescence decay of ZnNIL and ZnNAP and their
biexponential fits. Two components, one with τ ) subnano-
second and the other with τ ) 19-27 ns were required. Since
the latter values are fairly larger for states with a singlet
multipticity,33 the S2 emission may be originating in ligand-
centered triplet states, such as π-π*.
According to hitherto accummulated data on porphyrins,34
porphyrin dimers emit much more weakly than monomers, and
since the dimer formation by the addition of certain cations was
suggested in our systems also, we pursued the S1 emission
intensity as a function of [Rb+] and [Na+]. The results are
shown in Figure 11. Addition of Rb+ to zinc monomers leads
to quenching, and its behavior corresponds almost exactly to
that in the absorption study. Quenching is particularly marked
in 0 < [Rb+]/[ligand] < 1 region, while above [Rb+]/[ligand]
) ca. 1.5 or 2, the emission intensity becomes almost constant,
supporting, and consistent with two macrocycles approaching
in the former region and that cofacial dimers in the latter region
emit most weakly. Note here again that the dimerization is the
process in which two macrocycles are being linked by a cation,
so the 0 < [Rb+]/[ligand] < 1 region corresponds to this process,
while the [Rb+]/[ligand] > ca. 1.5 or 2 region is the region
where all crown ether voids are saturated with Rb+ or K+, as
mentioned in detail in the explanation of the absorbance change
(31) Kobayashi, N.; Ashida, T.; Hiroya, K.; Osa, T. Chem. Lett. 1992,
1567: Kobayashi, N.; Ashida, T.; Osa, T. Chem. Lett. 1992, 2031.
Kobayashi, N.; Ashida, T.; Osa, T.; Konami, H. Inorg. Chem. 1994, 33,
1735.
(32) Muralidharan, S.; Ferraudi, G.; Patterson, L. K. Inorg. Chim. Acta
1982, 65, L235.
(33) Generally, subnanosecond lifetimes have been reported for the
fluorescence of Pcs: Menzel, E. R.; Rieckhoff, K. E.; Voigt, E. M. Chem.
Phys. Lett. 1972, 13, 604. Brannon, J. H.; Magde, D. J. Am. Chem. Soc.
1980, 102, 62.
(34) In the porphyrin system, cofacial dimer units emit more weakly
than monomer: Kagan, N. E.; Mauzerall, D.; Merrifield, R. B. J. Am. Chem.
Soc. 1977, 99, 5486. Chang, C. K.; Kuo, M.-S.; Wang, C.-B. J. Heterocycl.
Chem. 1977, 14, 943. Chang, C. K. J. Heterocycl. Chem. 1977, 14, 1285.
Guckel, F.; Schweiser, D.; Collman, J. P.; Bencosme, S.; Evitt, E.; Sessler,
J. Chem. Phys. 1984, 86, 161. Mialoco, C.; Giannotti, A.; Maillard, P.;
Momenteau, M. Chem. Phys. Lett. 1984, 112, 87.
The zfs parameter D reflects anisotropic distribution of
unpaired electrons towards an out of plane axis, z, with respect
to that in the ring plane (x, y). Therefore the decreases in D
are qualitatively interpreted by delocalization of the unpaired
(35) (a) Ishii, K.; Yamauchi, S.; Ohba, Y.; Iwaizumi, M.; Hirota, N.;
Maruyama, K.; Osuka, A. J. Phys. Chem. 1994, 98, 9431. (b) Akiyama,
K.; Tero-Kubota, S.; Ikegami, Y. Chem. Phys. Lett. 1991, 185, 65. (c) van
Dorp, W. G.; Schoemaker, W. H.; Soma, M.; van der Waals, J. H. Mol.
Phys. 1975, 30, 1701.

Phthalocyanine Analogues
J. Am. Chem. Soc., Vol. 118, No. 5, 1996 1083
Here EX (exciton) and CR (charge resonance) denote an
excitation localized and delocalized type, respectively. Super-
script 3 means the excited triplet state. For the sandwich type
parallel dimer as DEX ) DM (M; monomer),1m,35a,39 an extent
of excitation delocalization over both rings, b2, is obtained from
the equations
DDM ) a2DEX + b2DCR
(3)
(4)
b2 ) (DM - DDM)/(DM - DCR
)
Using the observed D values of the monomer (DM ) 0.698
GHz) and the dimer (DDM ) 0.555 GHz) and the calculated D
value for the CR state (DCR ) -0.182 GHz) under a point
charge approximation,1m,35a,39 b2 was obtained as 0.16. This
value is very similar to that reported for a crown-ether bridged
zinc porphyrin (0.16)1m and is larger than those obtained for
several kinds of aromatic ring-bridged zinc porphyrins.35a This
indicates that the crown-ether bridges play an apparent role for
an interplanar interaction in the dimer.
The predominant isc to Tz (Pz . Px, Py) is interpreted by
participation of the d orbitals (dxz, dyz) of the Zn atom in the
T1(π-π*) state,35c which presents striking contrast to those for
metal-free Pc and MgPc (Px, Py . Pz).1m,35b
Figure 12. Time-resolved EPR spectra of ZnNIL (top), ZnBZ (middle),
and [ZnBZ]2(K+)3 (bottom) observed at 77 K in EtOH-CHCl3 (1:1
v/v) with their simulations. The spectra were taken at 0.5 µs after laser
excitation at 640 nm.
(vii) Molecular Orbital (MO) Calculations. The spectra
of metallophthalocyanines with D4h symmetry have most widely
been described by Gouterman’s four-orbital model40 in which
the optical spectrum can be interpreted in terms of a1u highest
occupied molecular orbital (HOMO), the a2u second HOMO,
and the doubly degenerate eg lowest unoccupied molecular
orbitals (LUMOs). Unlike the situation with porphyrins, the
aza nitrogens at the meso positions and fused benzenes separate
the a1u and a2u orbitals to dramatically increase the requirement
for intensity mixing between the lowest two singlet excited
states, Q and B1, so that the intense Q and B1 bands generally
appear at around 660-700 and 320-360 nm.41,42 The next band
called the B2 band24a appears ca. 2000-7000 cm-1 above the
B1 band, while the next higher band called the N band with
lower intensity appears at around 250-300 nm. This feature
is shown in Figure 13. Although the Q band is relatively pure,
other bands are mixtures of several configurations, and can not
be described by a one-electron description. Therefore, the
transitions shown in this figure represent a first-order assign-
ment.24a,40
With the lowering of molecular symmetry to C2V, the eg
orbitals split into two nondegenerate b1 and b2 orbitals. As
shown in Figure 13, transitions from the a1((4), a2((4), a1((2),
and another a1((2) orbitals to b1((5) and b2((5) orbitals
become allowed in this symmetry (angular momenta in paren-
theses). In order to correlate the transitions in Figure 13 and
to enhance our interpretation of the electronic absorption (and
MCD) spectra of monomer species, MO calculations have been
performed for the (pyrrole proton) deprotonated dianionic
species (i.e., NIL2-, BZ2-, NAP2-,and, in addition ANT2-, the
dianion of monoanthracotribenzoporphyrin) within the frame-
Table 5. Zfs Parameters Obtained from the Simulation
compd
D/GHz
E/GHz
Px:Py:Pz
ZnNIL
ZnBZ
(ZnBZ)2
ZnNAP
0.743
0.698
0.555
0.690
0.138
0.208
0.135
0.130
0:0.l:0.9
0:0.3:0.7
0:0.3:0.7
0:0.15:0.85
Table 6. Calculated Zfs Using a Half Point Charge Approximation
compd
D/GHz
E/GHz
ZnNIL
ZnBZ
ZnNAP
0.609
0.605
0.563
0.131
0.125
0.126
π-electrons over additionally fused benzo and naphtho rings in
the monomers and over another monomer unit in the dimer. In
order to check these changes more quantitatively, we have first
calculated the zfs of T1 monomer by using HOMO and LUMO
of Pc analogues under a half point charge approximation.36,37
The results were summarized in Table 6. It is found that the
amounts of the changes in the D values are consistent with our
consideration. The larger E value observed for ZnBZ may be
interpreted in terms of a pseudo-Jahn-Teller distortion,38
because the LUMO and the next LUMO are considered to be
close in energy for ZnBZ.36
For the dimer, we have already proposed a novel method for
qualitative analysis by using the D value as the following: on
the bases of the valence bond (VB) method any wave function
of the T1 dimer (DM) is described by a linear combination of
two types of excited states as
(39) Yamauchi, S.; Konami, H.; Hatano, M.; Ohba, Y.; Iwaizumi, M.
Mol. Phys. 1994, 83, 335.
(40) Gouterman, M. In The Porphyrins; Dolphin, D., Ed.; Academic
Press: New York, 1978; Vol. III, Part A, pp 1-165. Gouterman, M. J.
Chem. Phys. 1959, 30, 1139.
(41) Schaffer, A. M.; Gouterman, M.; Davidson, E. R. Theor. Chim. Acta
1973, 30, 9.
3ΦDM ) a3ΦEX + b3ΦCR
(2)
(42) McHugh, A. J.; Gouterman, M.; Weiss, C. Theor. Chim. Acta 1972,
24, 346. Dedieu, A.; Rohmer, M. M.; Veillard, A. AdV. Quantum. Chem.
1982, 16, 43. Orti, E.; Bredas, J. L.; Clarisse, C. J. Chem. Phys. 1990, 92,
1228. Rosa, A.; Baerends, E. J. Inorg. Chem. 1992, 31, 4717. Ishikawa,
N.; Ohno, O.; Kaizu, Y. J. Phys. Chem. 1993, 97, 1004. Henrikson, A.;
Roos, B.; Sundobom, M. Theor. Chim. Acta 1972, 27, 303.
(36) LCAO coefficients obtained in the MO calculations in the following
section (section vii) were used.
(37) Yamauchi, S.; Hirota, N.; Higuchi, J. J. Phys. Chem. 1988, 92, 2129.
(38) Chan, I. Y.; van Dorp, W. G.; Schaafsma, T. J.; van der Waals, J.
H. Mol. Phys. 1971, 22, 741.

1084 J. Am. Chem. Soc., Vol. 118, No. 5, 1996
Kobayashi et al.
Table 7. Calculated Transition Energies, Oscillator Strengths (f),
and Configurations for Deprotonated Pc Derivativesa
configurationsb
energy/eV(nm)
f
NIL2-
1.92892 (643) 0.70 19f20(87%) 18f21(13%)
1.99413 (622) 0.63 19f21(84%) 18f20(16%)
3.51648 (353) 1.10 18f20(30%) 17f20(27%) 16f20(27%)
3.70328 (335) 1.34 18f20(38%) 16f20(35%) 17f20(11%)
3.80931 (325) 2.11 18f21(63%) 13f21(12%) 19f20(11%)
3.99365 (310) 0.30 12f20(67%) 19f23(16%)
4.15704 (298) 0.16 15f20(40%) 13f21(29%) 16f21(22%)
4.46007 (278) 0.17 19f25(45%) 17f21(15%) 16f21(12%)
BZ2-
1.87357 (662) 0.84 21f22(87%) 20f23(12%)
1.87357 (662) 0.84 21f23(87%) 20f22(12%)
3.73470 (332) 2.19 20f23(67%) 13f23(12%) 21f22(11%)
3.73470 (332) 2.19 20f22(67%) 13f22(12%) 21f23(11%)
4.11100 (302) 0.25 16f22(23%) 21f27(20%) 13f22(15%)
16f23(14%) 21f26(12%)
4.11100 (302) 0.25 16f23(23%) 21f26(20%) 13f23(15%)
16f22(14%) 21f27(12%)
4.24330 (292) 0.27 21f27(64%) 16f22(13%)
4.24330 (292) 0.27 21f26(64%) 16f23(13%)
Figure 13. (A) Frontier molecular orbitals in D4h symmetry involved
in the major π-π* absorption transitions with energies lower than ca.
38 000 cm-1. The order of the orbital is based on theoretical models
for ZnPc(-2) species24a,25 and the results in the present calculations.
The Q, B1, and B2 transitions shown are those anticipated through
application of Gouterman’s four-orbital LCAO model40 and from
analysis of MCD data of Zn- and MgPc.24a,25 The arrows indicate
allowed transitions that give rise to bands observed in the 300-800
nm region of the absorption and MCD spectra. Each of the state
generated will be degenerate, and MCD A-terms are observed for each
transition. Numbers with ( indicate orbital angular momenta. (B)
Frontier molecular orbitals in C2V symmetry involved in the major π-π*
transitions with energies lower than ca. 38 000 cm-1 adapted from
Schaffer et al.41 No account has been taken of changes in the energies
of the molecular orbitals as a result of different symmetry from (A).
The possible ∆ML ) (1 transitions are shown based on application of
Gouterman’s four-orbital LCAO model40 as shown for (A). The orbitals
shown are labeled for the C2V symmetry and were obtained from the
corresponding D4h labeling through the use of correlation tables.
NAP2-
1.81979 (681) 0.90 23f24(89%) 21f25(11%)
1.84198 (673) 0.99 23f25(88%) 21f24(11%)
3.52917 (351) 1.72 22f25(46%) 21f24(38%)
3.76872 (329) 2.06 21f25(71%)
3.85608 (322) 0.54 22f25(43%) 21f24(28%)
4.19372 (296) 0.21 23f29(36%) 20f24(24%) 15f24(14%)
4.23624 (293) 0.15 20f24(66%)
4.71329 (263) 0.61 22f26(57%) 15f25(18%) 14f24(16%)
ANT2-
1.78867 (693) 0.90 25f26(89%)
1.83323 (676) 1.11 25f27(89%)
work of the PPP approximation. Parameters are the same as
those introduced in a previous publication,30 which have been
proven to be reliable. For example, for Pc2-, the calculations
reveal the Q and Soret transitions to be at 662 and 332 nm,
which are in good agreement with previously reported experi-
mental data.24 In the cases of NIL2- and NAP2- with C2V
symmetry, the degenerate pairs of orbitals comprising the lowest
and fourth lowest unoccupied and the third highest occupied
levels in BZ2- (i.e., D4h symmetry, Figure 13(A)) split into two
levels each, and therefore it is anticipated that the spectra
become more complex. However, as shown below, the MO
calculations have succeeded in reproducing the spectroscopic
characteristics mentioned in (iv), and further gave information
which appear useful in interpreting the experimental spectra.
The main results of the calculations are summarized as follows
(see also Table 7). (i) The Q band split into two (as shown
typically in Figure 7), but the splitting in NIL2- (526 cm-1) is
larger than NAP2- (237 cm-1). In ZnNIL and ZnNAP, the
values are 1423 and 282 cm-1, respectively, and in CuNIL and
CuNAP these are 808 and 340 cm-1, respectively. This is
consistent with the previously accummulated fact that the
substituent effect is larger the smaller the size of the parent
molecules.24b (ii) As judged from the comparison of oscillator
strengths in Table 7, of the two split bands, the band to lower
energy (a1((4) f b1((5)) is stronger for NIL2- but weaker
for NAP2-. This inversion occurs at BZ2-, because the intensity
of the degenerate two Q bands of BZ2- is the same (see Table
7). In order to confirm this tendency, we calculated also the
MOs of the monoanthraco analogue, i.e., ANT2-, and the results
are included in Table 7. As seen, the relative strength of the
second Q band (a1((4) f b2((5)) to the first Q band increases
in the order NIL2- (0.90), BZ2- (1.00), NAP2- (1.10), and
ANT2- (1.23), i.e., with the increase of molecular size of fused
aromatics. Such a tendency has not been reported to date,
3.23727 (383) 0.58 24f27(63%) 25f29(16%)
3.73149 (332) 2.10 23f26(60%) 24f27(11%)
3.82358 (324) 2.21 23f27(43%) 25f30(17%) 24f28(12%)
3.85049 (322) 0.33 25f30(60%) 24f28(19%)
4.06174 (305) 0.15 25f31(51%) 17f26(35%)
4.14686 (299) 0.39 22f26(28%) 25f31(26%) 17f26(24%)
19f28(12%)
4.23818 (293) 0.15 25f33(36%) 20f27(27%) 21f26(20%)
4.64845 (267) 0.22 19f27(40%) 17f27(38%)
a Excited states with less than 5.0 eV and f greater than 0.15 are
shown. b Orbital numbers 19, 21, 23, and 25 are HOMOs of NIL2-
,
BZ2-, NAP2-, and ANT2-, respectively. In NAP2-, the MOs numbered
22 (the second HOMO) and 26 (the third LUMO) are naphthalene-
centered orbitals. Similarly, in ANT2-, the second HOMO (MO number
24) and the third LUMO (MO number 28) are anthracene-centered
orbitals. These naphthalene and anthracene-centered orbitals are not
included in the energy diagram of Figure 13B.
although, of the split Q bands of other tribenzotetraazaporphy-
rins, the band to the longer wavelength (a1((4) f b1((5)) is
always stronger, consistent with the above results.22a (iii) The
Q band energy, defined as the mid-energy between the split Q
bands, shifts to longer wavelength in the order NIL2-, (632 nm),
BZ2- (662 nm), and NAP2- (676 nm), in accord with the
experimental order. For example, experimental values for zinc
complexes are 661, 683, and 705 nm for ZnNIL, ZnBZ, and
ZnNAP, respectively (values for ZnBZ and ZnNAP are the
average values of a Q MCD peak and trough, i.e., two B terms
of opposite sign). (iv) In a one-electron description, the Q band
to lower energy corresponds to a transition from the HOMO to
the LUMO (a1((4) f b1((5)), while that to higher energy
corresponds to one from the HOMO to the second LUMO
(a1((4) f b2((5)), also in accordance with Figure 13. Thus,
for example, in ZnNIL the bands at 686 and 625 nm, respec-
tively, correspond to the above transitions. However, their

Phthalocyanine Analogues
J. Am. Chem. Soc., Vol. 118, No. 5, 1996 1085
“purity” increases with increasing molecular size as suggested
from their configuration. Taking the instance of the Q band to
the higher energy, the percentage of the HOMO to the second
LUMO transition increases from 84% of NIL2- to 87% of BZ2-
and further to 88% of NAP2- and 89% of ANT2-. The situation
in the Q band to the lower energy is the same. (v) The features
in the shorter wavelengths are very complex. The split B1 bands
are no longer described in a one-electron description, since they
the first HOMOs in NIL2-, BZ2-, and NAP2- are predicted at
-7.7496, -7.4869, and -7.3728 eV, respectively, while the
energy of the first LUMOs in these systems are close to one
another (-3.9378, -3.7442, and -3.7447 eV, respectively).
Concluding Remarks
Monomers and cation-induced dimers of low symmetrical Pc
analogues substituted with three crown ether voids have been
analyzed by various spectroscopies including electronic absorp-
tion, magnetic circular dichroism, fluorescence emission, NMR,
EPR, and time-resolved EPR. Electronic absorption spectra of
monomers have been reproduced by molecular orbital calcula-
tions within the framework of the approximation treating only
π-electrons. From the change of absorption and emission
spectra, it was found that the dimerization is a two-step three-
stage process, and the final stage cofacial dimers are eclipsed
appear to be mixtures of several configurations. For NIL2-
,
two bands are estimated to lie at 353 and 335 nm, and transitions
to the LUMO from the second, third, and fourth HOMOs appear
to contribute comparably. Interestingly, contribution of a
transition from the second HOMO to the second LUMO (orbital
number 18 to 21) is not included in this band, although this
band is allowed (Figure 13(B)). For NAP2-, two bands are
calculated at 351 and 329 nm. For the 351 nm band, transitions
from the second HOMO to the second LUMO and the third
HOMO to the LUMO are the main contributions (note that the
second HOMO (MO number 22) and the third LUOMO (MO
number 26) are naphthalene-centered MOs). The 329 nm band
corresponds to one of the split B1 components and relatively
pure and strong. Judging from the similarity of configurations,
the bands at 351 and 322 nm may constitute one set, plausibly
the split two components of either B1 or B2 band. In 260-
360 nm of the spectra of NIL2- and NAP2-, six transitions are
recognized when f greater than 0.15 are listed (Table 7). (vi)
The lower energy shift of the Q band on going from MtNIL,
MtBZ, and MtNAP (Mt ) Zn or Cu) may be largely explained
as a destabilization of the HOMOs on ring expansion, since
C2V type with a center-center separation of 4.1-4.2 Å without
depending on the diameters of the cations used. Analysis of
time-resolved EPR spectra showed delocalization of excited
π-electrons over monomer chromophores and two macrocycles
in the dimers. The monomers showed both S1 and S2 emissions
whose quantum yields and lifetimes are smaller than those of
the corresponding symmetrical D4h species.
Acknowledgment. We thank Dr. H. Konami for his help in
MO calculations.
JA950929X