Gallium phthalocyanines: Structure analysis and electroabsorption study

Article abstract of DOI:10.1021/jp961231o

Electroabsorption has been measured with thin films of several (phthalocyaninato)galliums. It has been observed that the lowest energy absorption peak of most of the compounds studied is sensitive to the applied electric field. The shape of the electroabsorption spectrum is well reproduced by the second derivative of the corresponding absorption, indicating that the excited state has the character of a charge transfer state. It has been found that a clear correlation exists between the magnitude of the electromodulation and the sensitivity for the charge carrier generation. Crystal structure analysis of two compounds, hydroxygallium phthalocyanine type XI and chlorogallium phthalocyanine type II, made by Rietveld analysis, clarified that the molecules are stacked in pairs. The implication of the dimer structure is discussed.

Full text of DOI:10.1021/jp961231o

J. Phys. Chem. B 1997, 101, 13-19
13
Gallium Phthalocyanines: Structure Analysis and Electroabsorption Study
Kazuo Yamasaki*
Digital Xerography Laboratory, Fuji Xerox Co., Ltd., Takematsu, Minamiashigara-shi,
Kanagawa 250-01, Japan
Okimasa Okada, Kazuki Inami, and Kozo Oka
Foundation Research Laboratory, Fuji Xerox Co., Ltd., Takematsu, Minamiashigara-shi,
Kanagawa 250-01, Japan
Masahiro Kotani* and Hideo Yamada
Faculty of Science, Gakushuin UniVersity, Mejiro, Tokyo 171, Japan
ReceiVed: April 30, 1996; In Final Form: August 14, 1996X
Electroabsorption has been measured with thin films of several (phthalocyaninato)galliums. It has been
observed that the lowest energy absorption peak of most of the compounds studied is sensitive to the applied
electric field. The shape of the electroabsorption spectrum is well reproduced by the second derivative of the
corresponding absorption, indicating that the excited state has the character of a charge transfer state. It has
been found that a clear correlation exists between the magnitude of the electromodulation and the sensitivity
for the charge carrier generation. Crystal structure analysis of two compounds, hydroxygallium phthalocyanine
type XI and chlorogallium phthalocyanine type II, made by Rietveld analysis, clarified that the molecules are
stacked in pairs. The implication of the dimer structure is discussed.
∆ꢀG(F) ) -mG‚F - 1/2F‚pGF
∆ꢀE(F) ) -mE‚F - 1/2F‚pEF
(1)
(2)
Introduction
Phthalocyanines have been the subject of intensive study.
They are extremely stable pigments and exhibit many interesting
photochemical and photophysical properties. In addition to a
wide variety of central metals, there are also many polymorphs.
Some of them are efficient photocarrier generating pigments
while others are not.1 From an application point of view
phthalocyanines are especially important since they belong to
a few compounds which have a sensitivity in the deep red in
the spectral region and can be excited with laser diodes. At
the same time they pose an interesting challenge of how to
rationalize the charge carrier generation with photons of such a
low energy.
Unfortunately, many of the polymorphs of phthalocyanines
are difficult to grow in large enough crystals for single crystal
analysis. However, the structure of such a crystal could be
obtained from powder X-ray diffraction data by employing
Rietveld analysis2,3 and Monte Carlo methods. The analyses
produced plausible crystal structures for several kinds of
phthalocyanine crystals.4-7 Carrier generation by light absorp-
tion or, in other words, photoionization in a condensed phase
is of fundamental importance in photophysics of solids, as well
as of liquids, and has been the subject of intensive study for a
considerable time. In a discussion of photocarrier generation,
especially with low-energy excitation, an important issue is the
possible involvement of a charge transfer (CT) state.
Generally, a CT state has a small absorption cross section
and, as a result, is not easy to be observed directly. Very often
a CT transition is mixed with absorption which has much larger
absorption cross sections. The electromodulation technique
provides a sensitive means for detecting such a transition.8,9
In the presence of an electric field F, the energy of a molecule
in its ground state ꢀG and that in the excited state ꢀE are shifted
as
where m’s are the dipole moments and p’s are the polarizabili-
ties. Accordingly, the shift of the transition energy is given by
∆(ꢀE - ꢀG) ) -∆m‚F - 1/2F‚∆pF
This shift influences the absorbance at energy ꢀ ) ꢀE - ꢀG as
(3)
∂A
1 ∂2A
∆A )
∆ꢀ +
(∆ꢀ)2
( )
∂ꢀ
(
)
∂ꢀ2
2
∂A
)
{-∆m‚F - 1/2F‚∆pF} +
∂ꢀ
1 ∂2A
1
2
2
-∆m‚F - F‚∆pF (4)
(
){
}
∂ꢀ2
2
If the system under study is an ensemble of randomly oriented
molecules (or crystallite)
m‚F ) 0
may be assumed, since the orientation of both mG and mE with
respect to the field F is random. The term proportional to F4
can be neglected when ∆m is of a significant magnitude. In
this case only terms proportional to F2 remain. An isotropic
average over the randomly oriented dipole moments gives
1 ∂A
1 ∂2A
∆A )
∆pF2 +
(∆m)2F2
(5)
( )
(
)
∂ꢀ2
2 ∂ꢀ
6
X Abstract published in AdVance ACS Abstracts, November 15, 1996.
In this simplified model the molecular polarizability is assumed
S1089-5647(96)01231-X CCC: $14.00 © 1997 American Chemical Society

14 J. Phys. Chem. B, Vol. 101, No. 1, 1997
Yamasaki et al.
for the peak profile description and March function for the
correction of preferred orientation.19 The error function, which
is defined as the square of the difference between observed XRD
intensities and the calculated ones, was minimized by the Monte
Carlo method. Optimized were unit cell parameters, coordinates
of the center of mass of the molecules with respect to the crystal
axes, and Euler angles which determine the orientation of the
molecules. The molecules were considered to be rigid. Other
adjustable parameters were the half-width ω of the Lorentz
function for the peak profile and the preferred orientation
parameter r of the March function. The isotropic temperature
factors were fixed in our calculations because their influences
on XRD patterns were small. At the end of the minimization
cycle, the R factor
Figure 1. Synthesis and polymorphic conversion of gallium phthalo-
cyanines.
|Fobs(θi) - Fcalc(θi)|
∑
i
to be isotropic. Equation 5 predicts that an electroabsorption
spectrum should reproduce the first derivative of the absorption
spectrum if the first term is dominant (i.e., the second-order
Stark effect). If, on the other hand, the excited state has a large
dipole moment, the second term becomes important, and
accordingly the electroabsorption spectrum should give the
second derivative of the absorption spectrum. This is expected
to be the case when the final state of the transition is a CT
state, which has a large dipole moment. There have been some
reports on electroabsorption of phthalocyanine compounds.10-14
Here we report on electroabsorption of thin films of a few
polymorphs of (phthalocyaninato)galliums (hereafter GaPc’s).
Crystal structure analysis of two compounds, hydroxygallium
phthalocyanine type XI (HOGaPc-XI) and chlorogallium
phthalocyanine type II (ClGaPc-II), made by Rietveld analysis,
is presented. Daimon et al. reported that a polymorph of
hydroxygallium phthalocyanine (HOGaPc) exhibits a high
sensitivity for photocarrier generation.15
R )
|Fobs(θi)|
∑
i
was evaluated to assess the accuracy of the fit, where F(θi) is
the square root of the intensity at equidistant 2θ values between
5° and 35°.
Electroabsorption Measurement. Electroabsorption mea-
surements were carried out on a thin film of GaPc particles
dispersed in a copolymer of vinyl chloride and vinyl acetate
(VMCH from Union Carbide Co.). Two different types of cell
structures were employed. The structure used for measuring
electroabsorption spectrum was composed of a dispersed layer
of GaPc (50 wt %) on top of a N-(methoxymethyl)nylon
(LM5003 from DIC Co.) layer formed on a NESA glass plate.
The thickness of the GaPc layer was about 0.2 µm, and that of
the N-(methoxymethyl)nylon layer was about 5 µm. A field
dependence of the modulation was measured with a cell of single
layer structure, in which the concentration of GaPc was reduced
to 5 wt %. The GaPc pigments studied were ClGaPc-II,
HOGaPc-V, HOGaPc-XI, µ-oxogallium phthalocyanine
((GaPc)2O), and methoxygallium phthalocyanine (CH3OGaPc).
We have constructed an extrastable unit for absorption
measurements with single-beam arrangement. Light transmis-
sion was measured with and without an applied dc electric field.
The light from a tungsten halogen lamp was monochromatized
with a 25 cm monochromator (Nikon P250). The intensity of
the light transmitted through a sample was detected with a low-
noise Si photodiode (Hamamatsu S1137-16BQ), amplified with
a low noise preamplifier (NF, LI-76) and measured with a 61/2
digit multimeter. The monochromator was set to a wavelength,
a square pulse of a voltage was applied for 250 ms, and light
transmission was measured. The transmission at 0 V was then
measured during the next 250 ms. This was repeated 100 times
in 25 s, and the results were averaged on a PC. The
monochromator was then advanced to the next wavelength. This
averaging procedure, combined with a stable mechanical
construction careful selection of a light source and a low noise
photodiode, enabled us to attain a resolution in absorbance (∆A)
ranging from 2 × 10-5 to 1 × 10-4, depending on the
wavelength region. All measurements were performed in
ambient atmosphere and at room temperature.
Experiments
Synthesis of GaPc. Chlorogallium phthalocyanine (ClGaPc)
was synthesized by a modification of the method described in
ref 16 in which phthalodinitrile and gallium trichloride were
heated in 1-chloronaphthalene at 200 °C for 24 h. The product
pigments were filtered and purified by washing with N,N-
dimethylformamide (DMF) and methanol and then dried under
reduced pressure to give ClGaPc type I (ClGaPc-I). ClGaPc-
II was obtained after grinding and milling ClGaPc-I with benzyl
alcohol at room temperature for 24 h. HOGaPc type I
(HOGaPc-I) was prepared in the following procedure:17 Cl-
GaPc-I was dissolved in concentrated sulfuric acid. HOGaPc-I
precipitated when the solution was cast into water and neutral-
ized with aqueous ammonia. HOGaPc type V (HOGaPc-V)
was prepared by milling HOGaPc-I with DMF in a vessel with
glass beads at room temperature for 24 h. HOGaPc-XI was
obtained by changing the solvent on milling from DMF to
xylene. The pigments were isolated by centrifugation and then
dried under reduced pressure.
Crystal Structure Analysis. The powder crystals were
pressed onto a standard glass holder and mounted on an X-ray
diffractometer (Rigaku RU-200B). The sealed tube with a Cu
target was operated at 40 kV and 30 mA. The Cu KR radiation
(λ ) 1.54 Å) was monochromatized with a curved graphite
crystal. Step-scan X-ray data up to 2θ ) 40.0° were collected
with the speed of 0.5 s/step and the step length of 0.01° in 2θ.
The densities of the pigments were measured at 25 °C using a
density gradient tube.18
Results
The structure parameters and final R factors of HOGaPc-V,
HOGaPc-XI, and ClGaPc-II are shown in Table 1. The result
of the structure analysis for HOGaPc-V has been published in
ref 4. The atomic coordinates for HOGaPc-XI and ClGaPc-II
are shown in Tables 2 and 3. The observed XRD patterns, the
The refinement procedure for the Rietveld analysis has been
reported in previous papers.5,7 We used the Lorentz function

Gallium Phthalocyanines
J. Phys. Chem. B, Vol. 101, No. 1, 1997 15
TABLE 1: Crystallographic Data of HOGaPc-V,
HOGaPc-XI, and ClGaPc-II
HOGaPc-V4
HOGaPc-XI
ClGaPc-II
crystal system
space group
Z
a (nm)
b (nm)
triclinic
P1h
monoclinic
P21/c
4
1.663
2.632
0.716
90.00
129.93
90.00
1.66
triclinic
P1h
2
2
1.152
1.275
0.889
95.87
96.00
69.45
1.64
1.293
1.273
0.773
107.05
102.48
91.72
1.77
c (nm)
R (deg)
â (deg)
γ (deg)
density (g/cm3)
(measured)
R factor
(1.61)
0.195
(1.66)
0.279
(1.67)
0.216
TABLE 2: Positional Parameters of HOGaPc-XI with
Respect to the Triclinic Axes
a
b
c
a
b
c
Ga
O
N
N
N
N
N
N
N
N
C
C
C
C
C
C
C
C
C
C
C
0.136
0.026 0.647
C
C
C
C
C
C
C
C
C
C
C
C
C
C
C
C
C
C
C
C
C
-0.119 -0.155
-0.132 -0.108
-0.041 -0.082
0.126 -0.065
-0.025 -0.037
0.697
0.582
0.659
0.868
0.587
0.215
0.215
0.075
0.143 -0.001 0.420
0.267 0.009 0.957
0.077 -0.025 0.713
0.003
0.193
Figure 2. Powder XRD pattern observed yobs(θi) of HOGaPc-XI,
calculated pattern ycalc(θi), and their difference.
0.059 0.467
0.094 0.711
-0.132
-0.175
-0.285
-0.350
-0.305
-0.195
-0.025
-0.091
0.386
0.115
0.070
0.065
0.107 -0.067
0.152 -0.067
0.157
0.106
0.038
0.240
0.189
0.159
0.177
0.228
0.260
0.108
0.137
0.228 -0.071 1.031
-0.105 -0.009 0.412
0.041
0.375
0.142 0.411
0.079 1.027
0.615 -0.021 1.640
0.571 -0.067 1.641
0.462 -0.076 1.461
0.400 -0.038 1.282
0.443
0.552
0.294 -0.035 1.083
0.360 0.034 1.082
0.057 -0.101 0.842
0.072 -0.147 0.961
-0.017 -0.174 0.886
0.075
0.370
0.370
0.877
0.953
0.837
0.655
0.576
0.690
0.865
0.585
0.006 1.281
0.016 1.459
0.399
0.310
0.211
0.195
0.284
0.295
0.143
TABLE 3: Positional Parameters of ClGaPc-II with
Respect to the Triclinic Crystal Axes
a
b
c
a
b
c
Ga
Cl
N
N
N
N
N
N
N
N
C
C
C
C
C
C
C
C
C
C
C
0.228
0.002 0.853
C
C
C
C
C
C
C
C
C
C
C
C
C
C
C
C
C
C
C
C
C
0.650
0.679
0.617
0.521
0.435
0.354 1.378
0.265 1.440
0.165 1.379
0.158 1.248
0.066 1.160
0.274 -0.066 0.614
0.169
0.357
0.147 0.837
0.104 1.044
0.256 -0.100 1.017
0.069 -0.055 0.812
0.354 -0.104 1.137
0.352 -0.201 1.188
0.427 -0.242 1.304
0.403 -0.340 1.328
0.302 -0.403 1.234
0.255 -0.360 1.122
0.251 -0.262 1.099
0.194 -0.193 0.995
0.038 -0.155 0.824
-0.077 -0.180 0.730
-0.149 -0.274 0.700
-0.253 -0.274 0.594
-0.282 -0.188 0.530
-0.213 -0.094 0.561
-0.109 -0.092 0.666
-0.015 -0.012 0.721
Figure 3. Powder XRD pattern observed yobs(θi) of ClGaPc-II,
calculated pattern ycalc(θi), and their difference.
0.330
0.277 0.981
0.436 -0.025 1.200
0.090 -0.221 0.902
-0.018
0.083 0.688
0.160 0.748
0.269 0.727
0.322 0.645
0.427 0.650
0.488 0.750
0.433 0.827
0.326 0.818
0.249 0.882
0.208 1.055
0.244 1.182
0.345 1.243
0.067
0.062
-0.018
0.004
0.105
0.183
0.165
0.226
0.384
0.492
0.552
Figure 4. Molecular stacking in HOGaPc-V crystal. Two molecules
encircled are most adjacent to each other and constitute a pair.
calculated ones, and their differences are shown in Figures 2
and 3. The R factor for HOGaPc-XI is relatively large as
compared with that of HOGaPc-V and ClGaPc-II. This is
caused by inadequate peak profile function. While diffraction
peaks observed are asymmetric in shape, the Lorentz function
used for peak profile description is symmetric. The R factor is
very sensitive to peak profile, although the calculated peak
positions and intensities which are used in determining the
crystal structure are in good agreement with the observed ones.
The molecular stacking in the crystals is shown in Figure 4
for HOGaPc-V, in Figure 5 for HOGaPc-XI, and in Figure 6
for ClGaPc-II. In all these crystals the molecules are arranged

16 J. Phys. Chem. B, Vol. 101, No. 1, 1997
Yamasaki et al.
at 12 700 cm-1 than at the higher energy peak at 14 800 cm-1
,
although in the absorption spectrum the lower energy peak has
a smaller absorption coefficient. This indicates that the dipole
moment associated with the final state of the transition A is
larger than that of the transition B (see eq 5).
In HOGaPc-V, the ordinary absorption spectrum has two
absorption bands, one around 11 780 cm-1 (849 nm) and the
other around 15 630 cm-1 (640 nm). The small peak which is
seen at 14 500 cm-1 (690 nm) gains its intensity relative to other
peaks on dilution and is considered to be due to the monomer
dissolved in the matrix polymer. The peak at 690 nm is
observed also with HOGaPc-XI, CH3OGaPc, and (GaPc)2O
samples. The electroabsorption spectrum of HOGaPc-V exhibits
two negative peaks, one around 12 200 cm-1 (820 nm) and the
other around 14 300 cm-1 (700 nm). The electromodulation at
14 300 cm-1 may not be associated with the monomer peak at
14 500 cm-1 (690 nm), since modulation is observed around
14 500 cm-1 with other samples which have an absorption peak
at this energy. The electroabsorption spectrum is reproduced
by the second derivative of two Gaussians centered at 12 200
Figure 5. Molecular stacking in HOGaPc-XI crystal. Two molecules
encircled are most adjacent and form a pair.
and 14 300 cm-1
.
HOGaPc-XI and (GaPc)2O exhibit an electroabsorption spec-
trum having a peak at about 11 100 cm-1 (900 nm) and another
at 13 900 cm-1 (720 nm). It is the compound which has
absorption in the longest wavelength region among this group
of compounds. CH3OGaPc shows only a small electroabsorp-
tion signal. Figure 9 summarizes the electric field dependence
of the magnitude of the electromodulation. It clearly shows
that the modulations are proportional to the square of the applied
field, the expected behavior of a CT state described in eq 5.
Figure 6. Molecular stacking in ClGaPc-II crystal. Two molecules
encircled are the most adjacent molecules which constitute a pair.
in pairs. Figure 7 shows the molecular overlap within the pairs,
projected onto the average molecular plane. It could be argued
that molecules with their Ga-O (or Ga-Cl) bonds sticking to
each other may also be regarded as a pair. Actually, when
viewed down along the Ga-O (or Ga-Cl) bonds, these
molecules overlap much less, compared to a pair shown in
Figure 7.4
Discussion
Most of GaPc’s studied exhibit two absorption peaks, while
the solution spectrum has only one prominent peak at 14 500
cm-1 (690 nm). This splitting must reflect the strength of the
intermolecular interaction. [Mizuguchi et al. accounted for the
solid state spectrum of titanyl phthalocyanine type II in forms
of molecular distortion without considering any intermolecular
interaction.14 However, as they pointed out, it is quite probable
that the distortion in the crystal lattice is the result of
intermolecular interaction and that this is automatically included
in their calculation.] An easy comparison of the intermolecular
interaction is difficult, since the molecules are not planar.
However, if we compare the distances between the centers of
gravity of the molecules in a pair, they seem to be correlated
with the magnitude of the spectral splitting: they are 4.51 Å in
HOGaPc-XI (which exhibits the largest splitting of 4500 cm-1),
4.62 Å in HOGaPc-V (3700 cm-1), and 5.42 Å in ClGaPc-II
(2300 cm-1). Unfortunately, the crystal structures of CH3-
OGaPc and (GaPc)2O are not known so far.
Figure 8 shows an ordinary absorption spectrum and an
electroabsorption spectrum for ClGaPc-II, HOGaPc-V, HOG-
aPc-XI, (GaPc)2O, and CH3OGaPc. The ordinary absorption
spectrum of ClGaPc-II has two peaks, one at 12 700 cm-1 and
the other at 14 800 cm-1. It can be decomposed into two
overlapping Gaussians (A, which is centered at 12 700 cm-1
,
and B, centered at 14 800 cm-1). The electroabsorption
spectrum of ClGaPc-II exhibits two negative peaks, one around
12 700 cm-1 (790 nm) and the other around 14 800 cm-1 (680
nm). The spectrum can be decomposed into a superposition of
the second derivatives of A and B. This seems to indicate that
the final states of these transitions are of a CT character. The
amplitude of the modulation is larger at the lower energy peak
Figure 7. Molecular arrangement in a pair in (a) HOGaPc-V, (b) HOGaPc-XI, and (c) ClGaPc-II projected down to the average molecular plane
of phthalocyanine.

Gallium Phthalocyanines
J. Phys. Chem. B, Vol. 101, No. 1, 1997 17
Figure 8. Ordinary absorption spectrum (drawn out line) and electroabsorption spectrum (dotted line) of (a) ClGaPc-II, (b) HOGaPc-V, (c) HOGaPc-
XI, (d) CH3OGaPc, and (e) (GaPc)2O dispersive films. The applied electric field is about 5.0 × 105 V/cm. In (a) thin dotted line represents the
second derivative of the two Gaussian peaks which fit the ordinary absorption spectrum.
A common feature in the molecular arrangement found in
the three structures analyzed here is that molecules are stacked
in pairs with their centers staggered along the diagonal direction
of the molecules (Figure 10). This type of staggered overlap
is also common in all highly photosensitive phthalocyanine
pigments whose crystal structures have been determined:
Titanyl phthalocyanine phase I (TiOPc-I) and phase II (TiOPc-
II),20 Y-type TiOPc (Y-TiOPc),7 vanadyl phthalocyanine phase
II (VOPc-II),21 and x-form metal-free phthalocyanine (x-H2-
Pc).5 The sensitivity of HOGaPc-V is close to that of Y-TiOPc,
which is known as an extremely photosensitive crystal. ClGaPc-
II and HOGaPc-XI also have a photosensitivity comparable to
TiOPc-II. The structural findings support the proposal22 that
the specific molecular arrangement (as in Figure 10) within the
pair, or in a one-dimensional stack (as in x-H2Pc), is essential
in realizing a high photosensitivity of phthalocyanine crystals.
Figure 11 shows the relationship between the magnitude of
the modulation at the peak position normalized to the peak

18 J. Phys. Chem. B, Vol. 101, No. 1, 1997
Yamasaki et al.
(6b)
| + ) â|mm* + R|CT
where R2 + â2 ) 1. The dipole moments of the mixed states
are expressed by the following formulas:
∆m- ) -|m|- ≈ â2mCT
∆m+ ) +|m|+ ≈ R2mCT
(7a)
(7b)
where mCT is the dipole moment of the CT state. The
absorption, which can be modulated, is assumed to have a
Gaussian shape with a width Γ. Substituting eqs 7 and 8 for
eq 5, we obtain eq 9
2
ꢀ - ꢀ0
A(ꢀ) ) A(ꢀ ) exp -
(8)
(9)
{ (
) }
0
Γ
Figure 9. Amplitude of the electroabsorption signal at wavelengths
indicated versus the square of the electric field applied to the thin films
of several GaPc’s.
(â2mCT)2F2
(∆m)2F2
∆A
)
ꢀ0
)
(
)
Γ2
Γ2
A
when the second term of eq 5 is dominant. Structure analysis
with X-rays shows that the molecules are arranged in dimer
structures and that intermolecular distances are approximately
the same in the crystals studied. Hence, it may be reasonable
to assume that CT states in all GaPc’s studied have dipole
moments of comparable magnitude, mCT. The width Γ appears
to be the same in different Pc’s, judging from the observed
widths of the longest wavelength absorption peaks of GaPc’s
which are approximately the same. Hence, from eq 9 the
magnitude of electromodulation should be proportional to (â2)2.
The results shown in Figure 11 seem to indicate that the
efficiency of the photocarrier generation is indeed closely related
to the amount of the CT character admixed in the excited state.
The magnitude of â should be determined by the energy
difference of a CT state and of the lowest excited state.10 In
this context it is interacting to note that HOGaPc-XI, which
seems to have the largest intermolecular interaction in the pair,
exhibits a lower efficiency of carrier generation compared to
HOGaPc-V and ClGaPc-II. A CT state may be located close
to the lowest excited state in HOGaPc-V (and to a lesser extent
in ClGaPc-II), whereas the strong intermolecular interaction
pushes down the lowest excited state so that it is situated much
lower than the CT state and, as a result, the mixing is small.
It is customary to model the carrier generation as a two-step
process,23 as shown in Figure 12a. In the first step the system
is brought, by absorption of light, to the lowest singlet excited
state which undergoes nonradiative transition to a CT state which
ionizes by the applied electric field. Popovic studied the effect
of an applied electric field upon the fluorescence of Pc’s.24 He
observed that the fluorescence was indeed quenched by applying
an electric field which he interpreted in terms of two competing
channels, i.e., the fluorescence and the field-assisted transforma-
tion into a CT state. A high field favors the formation of a CT
state, or the dissociation of it, but since these are secondary
processes occurring after an optical transition, this should not
influence the absorption. With the present method, however,
one directly observes the influence of the external field upon
the state which itself is the final state of an optical transition.
Hence, we propose that a CT state, directly produced by light
absorption, is the intermediate for carrier generation: a model
shown in Figure 12b.
Figure 10. Molecular overlapping characteristic of (a) highly photo-
sensitive pigments and (b) low-photosensitivity pigments.
Figure 11. Relationship between the quantum yield of the photocarrier
generation and the magnitude of the modulation ∆A normalized to the
absorbance A at the peak position.
absorbance and the quantum yield of the photocarrier generation,
obtained through xerographic gain measurements.1 The two
quantities show a linear correlation: The larger the modulation
is, the higher is the sensitivity. Saito et al. studied electroab-
sorption of titanyl phthalocyanines12 and also concluded that
there is a correlation between the magnitude of electromodu-
lation and the photocarrier generation efficiency.
We consider the excited states of a molecular pair to be
described by a mixture of a locally excited state |mm* and a
CT state|CT . The mixed states are represented by
Generation of a CT state itself does not necessarily mean
the generation of charge carriers. A CT state is a pair of charged
molecules bound with Coulombic force. The charges are not
free. An additional energy is needed in order for the charges
|- ) R|mm* - â|CT
(6a)
and

Gallium Phthalocyanines
J. Phys. Chem. B, Vol. 101, No. 1, 1997 19
This type of spontaneous ionization could also be involved
in the primary step in photosynthesis. It has been known since
the elucidation of the crystal structure of photosynthetic unit in
rhodobacterium by Deisenhofer et al.28 that the first act of charge
separation occurs in the special pair which lies in the center of
the photosynthetic unit. It is a pair of (seemingly identical)
bacteriochlorophyll molecules, and recombination of charges
which should be very fast, is efficiently suppressed by an
arrangement of molecules which conduct electrons and holes
to both ends of the protein structure which is embedded in the
membrane. It is conceivable that in Pc’s, as well as in those
molecules mentioned above, the molecular environment provides
with a large enough polarization energy which favors an ion
pair state against a neutral (excited) dimer.
In conclusion we have studied structural and spectral features
of some photoconductive GaPc’s. Rietveld analysis has revealed
that in ClGaPc-II and HOGaPc-XI Pc molecules are arranged
in dimeric units with a specific molecular overlap which we
believe is characteristic of most highly photosensitive Pc’s. In
all GaPc crystals studied, except for CH3OGaPc, which dose
not belong to highly sensitive pigments, the longest wavelength
absorption peak can be modulated by external electric field
which indicates the CT nature of the final state of the transition.
The amount of the CT character has been shown to correlate
with the efficiency of photocarrier generation.
Figure 12. Schematic representation of charge carrier generation in
which (a) a CT state is produced indirectly following the photoexcitation
to singlet Frenkel exciton state (S*) and (b) the photoexcitation
populates directly a CT state, which dissociates into charge carriers.
The energy shift of the CT state due to external electric field is reflected
in the electroabsorption while in (a) the absorption should be essentially
insensitive to the electric field.
Acknowledgment. The authors are grateful to K. Daimon
and M. Iijima for pigment preparation and to R. Igarashi for
useful discussions.
References and Notes
to separate. A classical discussion based on that first proposed
by Lyons relates the energy of the charge carries with ionization
energy and electron affinity of the material as well as with
electrostatic polarization energy associated with the presence
of a pair of charges in a condensed phase:
(1) Low, K.-Y. Chem. ReV. 1993, 93, 449.
(2) Rietveld, H. M. J. Appl. Cryst. 1969, 2, 65.
(3) Malmros, B. G.; Thomas, J. O. J. Appl. Cryst. 1977, 10, 7.
(4) Inami, K.; Oka, K.; Daimon, K. J. Imaging Technol. 1995, 39, 298.
(5) Oka, K.; Okada, O. J. Imaging Technol. 1993, 37, 13.
(6) Okada, O.; Oka, K. Iijima, M. Jpn. J. Appl. Phys. 1993, 32, 3556.
(7) Oka, K.; Okada, O.; Nukada, K. Jpn. J. Appl. Phys. 1992, 31, 2181.
(8) Abbi, S. C.; Hanson, D. M. J. Chem. Phys. 1974, 60, 319.
(9) Sebasitian, L.; Weiser, G.; Baessler, H. Chem. Phys. 1981, 61, 125.
(10) Tokura, Y.; Koda, T.; Iyechika, Y.; Kuroda, H. Chem. Phys. Lett.
1983, 102, 174.
(11) Yoshida, H.; Tokura, Y.; Koda, T. Chem. Phys. 1986, 109, 375.
(12) Saito, T.; Kobayashi, T.; Suzuki, S.; Iwayanagi, T. J. Phys. Chem.
1993, 97, 8026.
(13) Ono, Y.; Kotani, M. To be published.
(14) Mizuguchi, J.; Rihs, G.; Karfunkel, H. R. J. Phys. Chem. 1995,
E ) I - A - P - C
(10)
where I is the ionization potential, A is the electron affinity, P
is the polarization energy, and C is the Coulombic energy.25 It
has been known that in many organic materials P has a
magnitude on the order of 1.5 eV.26
At this point it is interesting to note that there is a certain
similarity between the dimeric structure found in highly
photosensitive Pc’s and systems which are known to undergo
spontaneous ionization on photoexcitation.
In solution photochemistry photoinduced electron transfer is
known to occur when, for example, bianthryl is photoexcited
in a sufficiently polar solvent.27 This compound is composed
of a pair of identical (anthracene) molecules connected together
with a covalent bond. It shows absorption spectrum similar to
that of anthracene monomer. The neutral excited state which
is generated by photoabsorption, however, rapidly relaxes into
a charge-separated state, in which one anthracene moiety is
positively charged while the other is negatively charged. This
is only possible when the stabilization provided by polarization
induced in the solvent environment is sufficiently large, so that
the condition
99, 16217.
(15) Daimon, K.; Nukada, K.; Sakaguchi, Y.; Igarashi, R. In Proceedings
of IS&T’s Tenth International Congress on AdVances in Non-Impact Printing
Technologies; Melnyk, A., Ed.; The Society for Imaging Science and
Technology: Springfield, VA, 1994; p 215.
(16) Linsky, J. P.; Paul, T. R.; Nohr, R. S.; Kenney, M. E. Inorg. Chem.
1980, 19, 13131.
(17) Colaitis, D. Bull. Soc. Chem. Fr. 1962, 1962, 23.
(18) Linderstrom-Lang, K. Nature 1939, 139, 713.
(19) Dollase, W. D. J. Appl. Crystallogr. 1986, 19, 267.
(20) Hiller, W.; Stra¨hle, J. Z. Kristallogr. 1982, 159, 173.
(21) Ziolo, R. F.; Griffiths, C. H.; Troup, C. H. J. Chem. Soc., Dalton
Trans. 1980, 2300.
(22) Oka, K.; Okada, O. J. Imaging Sci. Technol. 1993, 37, 607.
(23) Noolandi, J.; Hong, K. M. J. Chem. Phys. 1979, 70, 3233.
(24) Popovic, Z. Chem. Phys. 1984, 86, 311 and references therein.
(25) Lyons, L. E. J. Chem. Soc. 1955, 5001.
(26) Sato, N.; Inokuchi, H.; Silinsh, E. A. Chem. Phys. 1987, 115, 269.
(27) Barbara, P. F.; Jarzeba, W. AdV. Photochem. 1990, 15, 1.
(28) Deisenhofer, J.; Epp, O.; Miki, K.; Huber, R.; Michel, H. J. Mol.
Biol. 1982, 158, 257; Nature 1985, 318, 618.
I - A - P - C < 0
(11)
is fulfilled.