Effects of benzoannulation and α-octabutoxy substitution on the photophysical behavior of nickel phthalocyanines: A combined experimental and DFT/TDDFT study

Article abstract of DOI:10.1021/ic061524o

The photophysical properties of a group of Ni(II)-centered tetrapyrroles have been investigated by ultrafast transient absorption spectrometry and DFT/TDDFT methods in order to characterize the impacts of α-octabutoxy substitution and benzoannulation on the deactivation pathways of the S ₁(π,π*) state. The compounds examined were NiPc, NiNc, NiPc(OBu)₈, and NiNc(OBu)₈, where Pc = phthalocyanine and Nc = naphthalocyanine. It was found that the S₁(π,π*) state of NiNc(OBu)₈ deactivated within the time resolution of the instrument (200 fs) to a vibrationally hot T₁(π,π*) state. The quasidegeneracy of the S₁(π,π*) and ³(d_z2,d_x2-y2) states allowed for fast intersystem crossing (ISC) to occur. After vibrational relaxation (ca. 2.5 ps), the T₁(π,π*) converted rapidly (ca. 19 ps lifetime) and reversibly into the ³LMCT(π,d_x2-y2) state. The equilibrium state, so generated, decayed to the ground state with a lifetime of ca. 500 ps. Peripheral substitution of the Pc ring significantly modified the photodeactivation mechanism of the S₁(π,π*) by inducing substantial changes in the relative energies of the S₁(π,π *), ³(d_π,d_x2-y2), ³(d _z2,d_x2-y2), T₁(π,π*), and ^1,3LMCT(π,d_x2-y2) excited states. The location of the Gouterman LUMOs and the unoccupied metal level (d_x2-y2) with respect to the HOMO is crucial for the actual position of these states. In NiPc, the S₁(π,π*) state underwent ultrafast (200 fs) ISC into a hot (d,d) state. Vibrational cooling (ca. 20 ps lifetime) resulted in a cold (d_z2,d_x2-y2) state, which repopulated the ground state with a 300 ps lifetime. In NiPc(OBu)₈, the S₁(π,π *) state deactivated through the ³(d_z2,d _x2-y2), which in turn converted to the ³LMCT(π,d _x2-y2) state, which finally repopulated the ground state with a lifetime of 640 ps. Insufficient solubility of NiNc in noncoordinating solvents prevented transient absorption data from being obtained for this compound. However, the TDDFT calculations were used to make speculations about the photoproperties.

Full text of DOI:10.1021/ic061524o

Inorg. Chem. 2007, 46, 2080−2093
Effects of Benzoannulation and
r-Octabutoxy Substitution on the
Photophysical Behavior of Nickel Phthalocyanines: A Combined
Experimental and DFT/TDDFT Study
Alexandra V. Soldatova, Junhwan Kim,‡ Xinzhang Peng,‡ Angela Rosa,*,§ Giampaolo Ricciardi,*,§
Malcolm E. Kenney,*,‡ and Michael A. J. Rodgers*,
Center for Photochemical Sciences and Department of Chemistry, Bowling Green State UniVersity,
Bowling Green, Ohio 43403, Department of Chemistry, Case Western ReserVe UniVersity,
CleVeland, Ohio 44106, and Dipartimento di Chimica, UniVersita` della Basilicata,
Via N. Sauro 85, 85100 Potenza, Italy
Received August 11, 2006
The photophysical properties of a group of Ni(II)-centered tetrapyrroles have been investigated by ultrafast transient
absorption spectrometry and DFT/TDDFT methods in order to characterize the impacts of -octabutoxy substitution
and benzoannulation on the deactivation pathways of the S1( π*) state. The compounds examined were NiPc,
NiNc, NiPc(OBu)8, and NiNc(OBu)8, where Pc phthalocyanine and Nc naphthalocyanine. It was found that
the S1( π*) state of NiNc(OBu)8 deactivated within the time resolution of the instrument (200 fs) to a vibrationally
hot T1(
(ISC) to occur. After vibrational relaxation (ca. 2.5 ps), the T1(
R
π
,
)
)
π
,
3
2
2
2
π,π*) state. The quasidegeneracy of the S1(π,π*) and (dz ,dx y ) states allowed for fast intersystem crossing
-
π
,
π*) converted rapidly (ca. 19 ps lifetime) and
reversibly into the 3LMCT(
π
,dx y ) state. The equilibrium state, so generated, decayed to the ground state with a
2
2
-
lifetime of ca. 500 ps. Peripheral substitution of the Pc ring significantly modified the photodeactivation mechanism
3
2
2
2
2
2
of the S1(
T1(
π*), and 1,3LMCT(
level (dx y ) with respect to the HOMO is crucial for the actual position of these states. In NiPc, the S1(
π,π*) by inducing substantial changes in the relative energies of the S1(π,
π*), 3(dπ,dx y ), (dz ,dx y ),
- -
2
2
π
,
π,dx y ) excited states. The location of the Gouterman LUMOs and the unoccupied metal
-
2
2
π
,
π*) state
-
underwent ultrafast (200 fs) ISC into a hot (d,d) state. Vibrational cooling (ca. 20 ps lifetime) resulted in a cold
2
2
2
(dz ,dx y ) state, which repopulated the ground state with a 300 ps lifetime. In NiPc(OBu)8, the S1(
π
,
π*) state
-
3
3
2
2
2
2
2
deactivated through the (dz ,dx y ), which in turn converted to the LMCT(
π,dx y ) state, which finally repopulated
-
-
the ground state with a lifetime of 640 ps. Insufficient solubility of NiNc in noncoordinating solvents prevented
transient absorption data from being obtained for this compound. However, the TDDFT calculations were used to
make speculations about the photoproperties.
Introduction
within the therapeutic window where tissue absorption and
scattering is minimal, metallophthalocyanines have been
employed as photodynamic and photothermal sensitizers for
tumor therapy and other medical applications, including
photodiagnostics.3-5
The nature of the coordinated metal plays a central role
in determining the excited-state dynamics of MPcs. For
instance, metallophthalocyanines containing metal ions with
Metallophthalocyanines (MPcs) have a range of potential
applications in high-tech fields, including molecular electron-
ics, nonlinear optics, liquid crystals, and photovoltaic solar
cells.1,2 Moreover, owing to their intense near-IR absorptions
* To whom correspondence should be addressed. E-mail: rosa@unibas.it
(A.R.), rg010sci@unibas.it (G.R.), malcolm.kenney@case.edu (M.E.K.),
rodgers@bgnet.bgsu.edu (M.A.J.R.).
Bowling Green State University.
‡ Case Western Reserve University.
(3) Ferna´ndez, D. A.; Awruch, J.; Dicelio, L. E. Photochem. Photobiol.
1996, 63, 784.
§ Universita` della Basilicata.
(1) Wagner, R. W.; Lindsey, J. S.; Seth, J.; Palaniappan, V.; Bocian, D.
F. J. Am. Chem. Soc. 1996, 118, 3996.
(2) Phthalocyanines: Properties and Applications; Leznoff, C. C., Lever,
A. B. P., Eds.; VCH Publishers: New York, 1990-1996; Vols. 1-4.
(4) Lawrence, D. S.; Whitten, D. G. Photochem. Photobiol. 1996, 64,
923.
(5) Szacilowski, K.; Macyk, W.; Drzawiecka-Matuszek, A.; Brindell, M.;
Stochel, G. Chem. ReV. 2005, 105, 2647.
2080 Inorganic Chemistry, Vol. 46, No. 6, 2007
10.1021/ic061524o CCC: $37.00
© 2007 American Chemical Society
Published on Web 02/15/2007

Photophysical BehaWior of Nickel Phthalocyanines
d0 or d10 configurations (closed-shell) are generally fluores-
cent, showing significant quantum yields of long-lived triplet
states.6,7 On the other hand, complexes with open-shell
central metals, such as Ni(II), Co(II), Fe, and others,8-10 are
nonluminescent. In these complexes, emission is not observed
because rapid radiationless decay of the primary excited
S1(π,π*) state into low-lying (d,d) and/or MLCT/LMCT
states occurs, from whence the ground-state surface can
readily be reached. In these cases, rapid conversion of
electronic energy into vibrational modes of the ground state
is achieved, generating highly localized thermal effects that,
in tissue environments, can lead to cell death.10
Recently, Ni(II)5,9,14,18,23,27,32,36-octabutoxy-2,3-
naphthalocyanine, NiNc(OBu)8,11 has been found to meet
the conditions for effective photothermal photosensitization,
namely, it shows efficient photon absorption (ꢀ845nm ∼ 2 ×
105 M-1 cm-1) in the near-IR region where light penetrates
tissues more effectively and the photogenerated excited
states decay rapidly. There are now indications that this
compound, being efficiently taken up by amelanotic mela-
noma cells, is able to effectively destroy such cells by
photothermal sensitization without any requirement of mo-
lecular oxygen.12,13
Scheme 1. Synthetic Routes Used for NiNc and NiNc(5,36-OBu)8
investigated using femtosecond transient absorption spec-
trometry and density functional (DFT) and time-dependent
DFT (TDDFT) methods in an attempt to independently
characterize the effects of the R-octabutoxy substitution and
benzoannulation on the nature and rates of the deactivation
pathways.
Experimental Section
A. Materials. The solvents toluene (99.5+%, spectrophotometric
grade, Aldrich) and 1-chloronaphthalene (90%, technical grade,
Aldrich) were used without further purification. NiPc was purchased
from Aldrich Chemical Co. NiPc(OBu)8 was prepared as previously
described.10 NiNc and NiNc(OBu)8 were synthesized as follows
(Scheme 1).
An understanding of the factors governing the efficiency
of the photothermal behavior of NiNc(OBu)8 and related
compounds requires knowledge of their excited-state dynam-
ics. A combined experimental and theoretical investigation
of the photophysics of the Ni(II)1,4,8,11,15,18,22,25-oc-
tabutoxyphthalocyanine, NiPc(OBu)8, in toluene solution was
published recently.10 In that compound, the primary excited
1. NiNc. The synthesis used for NiNc is a conventional
cyclization and has been reported briefly before.14 A mixture of
2,3-dicyanonaphthalene (950 mg), NiCl2 (250 mg), and (NH4)2-
MoO4 (2 mg) was heated (250 °C) for 3 h. The reaction product
was extracted (Soxhlet extractor) with benzene for 20 h, CH3OH
for 6 h, and dimethylformamide for 60 h. It then was heated (80
°C) with aqueous NaOH (10%, 20 mL) for 1 h, heated (80 °C)
with aqueous HCl (10%, 20 mL) for 1 h, dissolved in H2SO4
(concentrated, 6 mL), and recovered by dilution of the solution
with H2O (30 mL). Finally, it was washed (H2O, acetone), vacuum
dried (100 °C), and weighed (400 mg, 39%). UV-vis (1-
chloronaphthalene) λmax, nm: 770. The compound is a green solid.
It is insoluble in CH2Cl2, dimethylformamide, toluene, and hexanes.
The insolubility of this compound in organic solvents precludes
its purification by chromatography.
2. NiNc(5,36-OBu)8. Under Ar, a refluxing solution of H2Nc-
(5,36-OBu)8 (134 mg), 1,8-diazabicyclo[5.4.0]undec-7-ene (0.02
mL), and 1-butanol (3 mL) was treated with 10 portions of
Ni(CH3CO2)2‚4H2O (154 mg) over 3 days, diluted with a solution
of CH3OH and H2O (1:1, 6 mL), and filtered. The solid was
chromatographed (Al2O3(III), toluene), vacuum dried (room tem-
perature), and weighed (88 mg, 63%). UV-vis (toluene) λmax, nm
(log ꢀ): 844 (5.5). NMR (CDCl3): δ 8.93 (m, 8H, 1,4-Nc H), 7.86
(m, 8H, 2,3-Nc H), 5.03 (t, 16H, OCH2), 2.20 (m, 16H, OCH2CH2),
1.65 (m, 16H, OC2H4CH2), 1.03 (t, 24H, OC3H6CH3). HRMS-
ESI-TOF m/z: [M]+ calcd for C80H88N8O858Ni, 1346.6079; found,
1346.6073. The compound is a green solid. It is soluble in
CH2Cl2, dimethylformamide, and toluene and is slightly soluble in
hexanes.
3
2
2
2
S1(π,π*) state deactivates via lower-lying (dz ,dx -y ) and
3
2
2
LMCT(π,dx -y ) states, from whence ground-state repopu-
lation occurs with a lifetime of 640 ps. The deactivation
mechanism proposed for NiPc(OBu)8 does not necessarily
apply to NiNc(OBu)8, as the chemical changes in the
macrocycle are expected to have significant impact on the
3
2
2
2
excited-state pattern and energies. For instance, the (dz ,dx -y ),
which, in NiPc(OBu)8, was found to intervene only at the
early stage of the decay process, might not be involved at
all in the radiationless decay of NiNc(OBu)8 due to the
significant energy lowering (ca. 200 meV) of the primary
excited S1(π,π*) state.11
In the present work, the spectral and dynamic behaviors
of NiPc, NiNc, NiPc(OBu)8, and NiNc(OBu)8 have been
(6) Rihter, B. D.; Kenney, M. E.; Ford, W. E.; Rodgers, M. A. J. J. Am.
Chem. Soc. 1993, 115, 8146.
(7) Kobayashi, N.; Ogata, H.; Nonaka, N.; Luk’yanets, E. A. Chem.s
Eur. J. 2003, 9, 5123.
(8) Ishii, K.; Kobayashi, N. In The Porphyrin Handbook; Kadish, K. M.,
Smith, K. M., Guilard, R., Eds.; Academic Press: New York, 2003.
(9) Lecomte, C.; Rohmer, M. M.; Benard, M. In The Porphyrin Handbook;
Kadish, K. M., Smith, K. M., Guilard, R., Eds.; Academic Press: New
York, 2003.
(10) Gunaratne, T. C.; Gusev, A. V.; Peng, X.; Rosa, A.; Ricciardi, G.;
Baerends, E. J.; Rizzoli, C.; Kenney, M. E.; Rodgers, M. A. J. J. Phys.
Chem. A 2005, 109, 2078.
(11) Busetti, A.; Soncin, M.; Reddi, E.; Rodgers, M. A. J.; Kenney, M. E.;
Jori, G. J. Photochem. Photobiol., B 1999, 53, 103.
(12) Camerin, M.; Rello, S.; Villaneuva, A.; Ping, X.; Kenney, M. E.;
Rodgers, M. A. J.; Jori, G. Eur. J. Cancer 2005, 41, 1203.
(13) Camerin, M.; Rodgers, M. A. J.; Kenney, M. E.; Jori, G. Photochem.
Photobiol. Sci. 2005, 4, 251.
B. UV-Visible Absorption Spectra. The ground-state electronic
absorption spectra were recorded at room temperature on a Varian
Cary 50 Bio (Varian Corporation) UV-visible single-beam spec-
trophotometer using a 10 or 2 mm path length quartz cuvette.
(14) Mikhalenko, S. A.; Luk’yanets, E. A. Zh. Obshch. Khim. 1969, 39,
2554-2558; J. Gen. Chem. USSR (Engl. Transl.) 1969, 39, 2495.
Inorganic Chemistry, Vol. 46, No. 6, 2007 2081

Soldatova et al.
Table 1. Selected Bond Distances (Å) and Bond Angles (deg)
Calculated for NiPc, NiNc, NiPc(OMe)8, and NiNc(OMe)8. X-ray Data
Are in Italics
C. Ultrafast Pump-Probe Measurements. Details of the
pump-probe instrument for ultrafast transient absorption measure-
ments at the Ohio Laboratory for Kinetic Spectrometry at Bowling
Green State University have been described previously.15 The recent
improvements to enhance signal-to-noise characteristics have been
published elsewhere.16 For the present measurements, the main part
(95%) of the fundamental output beam (95 fs, 1 kHz, 800 nm)
from an amplified, mode-locked Ti:sapphire laser (Hurricane,
Spectra-Physics) was used to excite the sample or to pump an
optical parametric amplifier (OPA 800, Spectra-Physics) to generate
excitation pulses in the 330-845 nm region. The pump intensity
was attenuated to have an energy of 1-6 µJ per pulse at the sample.
The linear polarization of the pump beam was set at the magic
angle (54.7°) with respect to that of the probe beam in order to
eliminate any temporal components arising from molecular rota-
tions. The transient absorption signal was monitored with the white-
light continuum probe in the spectral range of 460-780 nm or with
the extended to the near-IR continuum in the region of 800-1100
nm. The instrument response time of the ultrafast spectrometer was
ca. 200 fs. The absorption spectra of the solutions were measured
before and after the experiment to check for possible sample
degradation.
NiPc
D4h
NiNc
D4h
NiPc(OMe)8 NiNc(OMe)8
D2d
D2d
Ni-Np
1.912
1.917
1.900
1.906
1.887(6)a
1.384
1.929(7)b
1.384
1.878(4)c
1.382
CR-Np
1.381
1.453
1.430
1.320
1.368
1.379(12)
1.452
1.456(7)
1.402
1.292(10)
1.318
1.320(9)
1.376(6)
1.453
1.455(7)
1.406
1.397(7)
1.319
1.320(6)
1.363
CR-Câ
1.451
1.416
1.317
Câ-Câ
CR-Nb
Co-O
1.361(6)
CR-Np-CR
CR-Nâ-CR
O‚‚‚O
106.9
110.6(14)
121.3
107.2
107.5
107.0(4)
121.2
120.1(4)
3.793
3.910(2)
107.6
122.2
3.865
121.4
119.2(6)
CMe-O-Co-Câ
(CR-Np-Np-CR)ad
177.4
12.5
16.2(4)
63.0
11.3
d
0.0
0.0
a X-ray data for NiPcI; from ref 25. b From ref 28. c X-ray data for
NiPc(OBu)8; from ref 10. d Dihedral angle (deg) between adjacent pyrrole
ring planes.
D. Quantum Chemical Calculations. All calculations have been
performed with the ADF (Amsterdam Density Functional) suite of
programs, release 2004.01.17,18 The calculations made use of the
local density approximation (LDA) functional of Vosko-Wilk-
Nusair (VWN)19 plus the generalized gradient approximation
(GGA), employing Becke’s20 gradient approximation for exchange
and Perdew’s21 approximation for correlation (BP). The excitation
energies were calculated using time-dependent density functional
theory (TDDFT). In the ADF implementation,22,23 the solution of
the TDDFT response equations proceeds in an iterative fashion
starting from the usual ground-state or zeroth-order Kohn-Sham
(KS) equations.23 For these, one needs an approximation to the usual
static exchange-correlation (xc) potential, Vxc(r). After the ordinary
KS equations have been solved, the first-order density change has
to be calculated from an iterative solution to the first-order KS
equations. In these first-order equations, an approximation is needed
to the first functional derivative of the time-dependent xc potential,
Vxc(r, t), with respect to the time-dependent density F(r′, t′), the
so-called xc kernel. For the xc kernel, we used the adiabatic local
density approximation (ALDA). In this approximation, the time
dependence (or frequency dependence, referring to the Fourier-
transformed kernel) is neglected, and one simply uses the dif-
ferentiated static LDA expression. In our case, the Vosko-Wilk-
Nusair parametrization was used.19 For the exchange-correlation
potentials, which appear in the zeroth-order KS equations, we
employed the same GGA as that in the DFT calculations.
To facilitate the calculations, the butyl substituents were replaced
by methyl groups.
In all of the DFT and TDDFT calculations, the all-electron ADF
TZ2P basis set, which is an uncontracted, triple-ú STO basis set
with one 3d and one 4f polarization function for C, N, and O atoms,
one 2p and one 3d polarization function for H, and a triple-ú 3d,
4s basis with one 4p and one 4f function for Ni, was used.
The vertical absorption energies, Eva, have been evaluated at the
ground-state optimized geometry. The adiabatic energies, Eadia, have
been obtained according to the expression
Eadia ) Eve + ∆E
where Eve is the vertical emission energy, which is calculated at
the TDDFT level using the relaxed excited-state geometry. The
∆E term accounts for the change in energy of the ground state
upon deformation to the relaxed geometry of the excited state (for
a schematic definition of the calculated energies, see Figure 10 in
ref 10).
Geometry optimizations were performed for the ground and
selected triplet excited states of the complexes of the series. Open-
shell DFT calculations for the triplet states were performed within
a spin-unrestricted formalism, and spin contamination, monitored
by the expectation value of S2, was found to be negligible. The
optimized ground-state and excited-state structures were verified
to be true minima by frequency calculations.
(15) Nikolaitchik, A. V.; Korth, O.; Rodgers, M. A. J. J. Phys. Chem. A
1999, 103, 7587.
(16) Pelliccioli, A. P.; Henbest, K.; Kwag, G.; Carvagno, T. R.; Kenney,
M. E.; Rodgers, M. A. J. J. Phys. Chem. A 2001, 105, 1757.
(17) Amsterdam Density Functional Program; Theoretical Chemistry, Vrije
Universiteit: Amsterdam, The Netherlands. http://www.scm.com.
(18) te Velde, G.; Bickelhaupt, F. M.; Baerends, E. J.; Fonseca Guerra,
C.; van Gisbergen, S. J. A.; Snijders, J. G.; Ziegler, T. J. Comput.
Chem. 2001, 22, 931.
Results and Discussion
A. Ground-State Molecular and Electronic Structure.
In this section, the molecular and electronic structure effects
of introducing benzo rings and alkoxy groups at the R-posi-
tion both separately and together into the basic Pc structure
are examined.
The most relevant geometrical parameters calculated for
NiPc, NiNc, NiPc(OMe)8, and NiNc(OMe)8 are gathered in
Table 1 and compared to the available experimental data.
The molecular structure of NiPc has been investigated
(19) Vosko, S. H.; Wilk, L.; Nusair, M. Can. J. Phys. 1980, 58, 1200.
(20) Becke, A. Phys. ReV. A 1988, 38, 3098.
(21) Perdew, J. P. Phys. ReV. B 1986, 33, 8822 (Erratum: Phys. ReV. B
1986, 34, 7406).
(22) van Gisbergen, S. J. A.; Snijders, J. G.; Baerends, E. J. Comput. Phys.
Commun. 1999, 118, 119.
(23) van Gisbergen, S. J. A.; Snijders, J. G.; Baerends, E. J. J. Chem. Phys.
1995, 103, 9347.
2082 Inorganic Chemistry, Vol. 46, No. 6, 2007

Photophysical BehaWior of Nickel Phthalocyanines
Figure 1. Top view (left) and side view (right) of the DFT-optimized
molecular structures of NiPc(OMe)8 and NiNc(OMe)8.
Figure 2. Energy level scheme for NiPc, NiNc, NiPc(OMe)8, and NiNc-
(OMe)8. In the numbering of the MOs, the core electrons (C, N, O, 1s; Ni,
1s-2p) are not included. For the sake of comparison with NiPc and NiNc,
the unoccupied metal orbitals of NiPc(OMe)8 and NiNc(OMe)8, 22b2 and
extensively both experimentally24-26 and theoretically,26,27
and a planar D4h structure is, by now, well-established for
this complex. According to the crystallographic data by Araki
et al.,28 the linearly benzoannulated analogue, NiNc, also
possesses a planar D4h structure. Experimental and theoretical
structural data indicate that linear benzoannulation induces
a small, although not negligible, lengthening of the Ni-Np
distance (cf. Table 1).
2
2
27b2, respectively, are referred to as dx -y MOs.
slightly smaller than that computed for NiPc(OMe)8 (11.3°
vs 12.5°). A ruffled structure of S4 symmetry, character-
ized by an up-down arrangement of the methyl groups, was
also theoretically explored. However, the S4 structure was
found to be ∼12 kcal/mol less stable than the D2d-saddled
structure.
For the R-butoxy-substituted systems, X-ray studies of
NiPc(OBu)8 have shown that introduction of eight butoxy
groups at the R-positions induces a remarkable saddling of
the phthalocyanine ring,10 with the dihedral angle between
adjacent indole rings averaging 16.2(4)°. DFT studies on the
model complex NiPc(OMe)8 also pointed to a D2d-saddled
conformation of the macrocycle (see Figure 1) both in the
gas phase and in solution.10 Therefore, it has been argued
that the saddling distortion is not merely dictated by packing
requirements; rather, it originates from intrinsic electronic
factors such as the necessity to minimize the steric hindrance
between the lone pairs of the facing oxygen atoms, while
preserving an efficient nickel-phthalocyanine interaction.
Indeed, as can be inferred from Table 1, the Ni-Np distance
shortens significantly going from the planar NiPc to the
saddled butoxy derivative.
Unfortunately, there are no X-ray structural data available
for NiNc(OBu)8. In spite of the numerous attempts, we were
not able to grow crystals of this complex suitable for
structural investigation. Geometry optimization of the model
complex NiNc(OMe)8 without symmetry constraint afforded
the D2d-saddled structure of Figure 1. This structure, which
corresponds to a stable energy minimum (all-positive har-
monic frequencies), shows a degree of saddling that is only
It is apparent from the DFT-calculated structures of NiPc-
(OMe)8 and NiNc(OMe)8 in Figure 1 that the most striking
effect of benzoannulation is the displacement of the methyl
groups from the plane of the corresponding indole ring to
relieve steric hindrance between the CMe-H and Cγ-H σ
bonds. The torsion angle CMe-O-Co-Câ, which provides
a measure of the displacement of the CMe atoms from the
indole plane, is 63.0° in NiNc(OMe)8 and 177.4° in NiPc-
(OMe)8, where the CMe atoms are nearly coplanar with the
benzo rings.
The lengthening of the Co-O bond distance going from
NiPc(OMe)8 to NiNc(OMe)8, see Table 1, is indicative of a
less effective conjugation between the oxygen lone pairs and
the macrocycle π system in the naphthalocyanine. The
decreased π conjugation can be traced to the tilting of the
methyl groups that makes the interaction between the oxygen
lone pairs and the macrocycle π system less favorable.
The highest-occupied and the lowest-unoccupied ground-
state, one-electron levels of NiPc, NiNc, NiPc(OMe)8, and
NiNc(OMe)8 are shown in Figure 2.
In our choice of the axes, the x and y axes bisect the indole
rings in the planar D4h structures of NiPc and NiNc and point
to the meso-nitrogen atoms in the saddled D2d structures of
NiPc(OMe)8 and NiNc(OMe)8.
Considering, first, NiPc as a point of reference for the
series, among the one-electron levels in Figure 2, one may
recognize the π G-2a1u and the π* G-7eg (G ) Gouterman)
orbitals and the metal 3d orbitals. In the virtual spectrum,
there is the 3dx -y2 (11b1g), which is composed of 3dx -y and
Np lone pairs in almost equal amount and pushed up by
(24) Robertson, J. M. J. Chem. Soc. 1936, 1195.
(25) Schramm, C. J.; Scaringe, R. P.; Stojakovic, D. R.; Ibers, J. A.; Marks,
T. J. J. Am. Chem. Soc. 1980, 102, 6702.
(26) Mastryukov, V.; Ruan, C.-y.; Fink, M.; Wang, Z.; Pachter, R. J. Mol.
Struct. 2000, 556, 225.
(27) Rosa, A.; Ricciardi, G.; Baerends, E. J.; van Gisbergen, S. J. A. J.
Phys. Chem. A 2001, 105, 3311.
2
2
2
(28) Morishige, K.; Araki, K. J. Chem. Soc., Dalton Trans. 1996, 4303.
Inorganic Chemistry, Vol. 46, No. 6, 2007 2083

Soldatova et al.
Figure 3. Visual representation of the HOMO and the pair of Gouterman LUMOs for NiPc, NiNc, NiPc(OMe)8, and NiNc(OMe)8.
2
2
antibonding between the 3dx -y and the Np lone pairs. The
and 15a2 are upshifted by ∼1.8 eV with respect to the
equivalent orbitals (5eg and 2b1u) in NiPc (these orbitals are
too low in energy to show in Figure 2).
2
highest occupied 3d levels are the 3dz (13a1g), which is an
almost pure metal orbital, and the 3dπ (6eg). The 6eg is
heavily (41%) mixed with the Np-based π orbitals of the
phthalocyanine ring. (For an extensive discussion of the
electronic structure of NiPc, see refs 27, 29.) According to
the level pattern in Figure 2, the outstanding difference
between NiP and NiNc is the destabilization of the HOMO
(3a1u). The pushing up effect of the annulated benzo rings
is in accordance with the antibonding between the Pc ring
system and the annulated benzo rings, which is apparent in
the plot of the NiNc HOMO in Figure 3. The fused benzo
rings also have some pushing up effect on the G-eg(9eg)
orbitals. The two degenerate 9eg MOs are, however, less
destabilized than the HOMO because each component of the
degenerate pair feels the antibonding effect of only two of
the four fused benzo rings. As for the metal states, apart
from a modest stabilization of the Ni-3dx -y due to the slight
lengthening of the Ni-Np bond distance, their energy and
composition change very little upon benzoannulation.
In a previous paper,10 it was pointed out that the electron-
releasing methoxy groups cause a destabilization of all of
the levels in NiPc(OMe)8. MOs with large amplitude on the
ortho-carbons of the benzo ring are most affected, owing to
the antibonding between the Co-2pz and the oxygen lone pairs
of the methoxy groups. This is the case for the 17b1 (HOMO)
(cf. the plot of this orbital for NiPc(OMe)8 in Figure 3), the
37e (HOMO-1), and the 15a2 (HOMO-2). Notably, the 37e
When comparing the one-electron levels of NiNc and
NiNc(OMe)8 in Figure 2, it is apparent that R-alkoxy
substitution of the Nc ring does not induce the generalized
destabilization of the levels observed upon R-alkoxy sub-
stitution of the Pc ring. This can be explained primarily by
the tilting of the methyl groups (see Figure 1) lowering the
charge transfer from the electron-releasing alkoxy groups to
the macrocycle. According to the Hirshfeld charge analysis,
the oxygen atoms carry more negative charge in NiNc(OMe)8
than in NiPc(OMe)8 (0.1061 vs 0.0906 electrons). The level
scheme in Figure 2 shows that the most relevant electronic
effect of R-alkoxy substitution of the Nc ring consists of
the upshift of the 3a1-derived 21b1 (HOMO) induced by the
antibonding (Figure 3) between the Nc-3a1u, which has large
amplitude (28%) on the ortho-carbons (Co), and the oxygen
lone pairs of the methoxy groups whose contribution to this
orbital is approximately 6%. The R-methoxy substitution also
lifts two low-lying NiNc levels, the 7eg and 3b1u (not shown
in the diagram in Figure 2), which become the HOMO-1
(19a2) and the HOMO-2 (46e) of NiNc(OMe)8. The empty
G orbitals, the 47e, although having appreciable amplitude
(∼10%) on the ortho-carbons, are destabilized very little due
to the minor (∼1%) contribution of the oxygen lone pairs.
The important consequence is that the energy gap between
the HOMO and the pair of unoccupied Gouterman orbitals
diminishes considerably going from NiNc to NiNc(OMe)8.
2
2
(29) Rosa, A.; Baerends, E. J. Inorg. Chem. 1994, 33, 584.
2084 Inorganic Chemistry, Vol. 46, No. 6, 2007

Photophysical BehaWior of Nickel Phthalocyanines
Table 3. Excitation Energies (eV), Composition, and Character of the
Lowest Excited States of NiPc and NiNc
state
composition (%)
character
NiPc
MLCT
MLCT
π,π*
Eva
Eadia
13A2g
13B2g
11Eu
6eg f 7eg (100)
6eg f 7eg (100)
2a1u f 7eg (93)
6eg f 7eg (100)
6eg f 11b1g (99)
13a1g f 11b1g (100)
2a1u f 11b1g (100)
2a1u f 11b1g (100)
2a1u f 7eg (100)
2.08
2.02
1.98
1.98
1.77
1.62
1.47
1.45
1.37
13A1g
13Eg
MLCT
dπ,dx -y
1.56 (13B3g)a
1.34b
2
2
13B1g
11B1u
13B1u
13Eu
dz ,dx -y
LMCT
LMCT
π,π*
2
2
2
1.30b,c
1.29b
1.30 (13Au)d
NiNc
π,π*
dπ,dx -y
13B2g
13Eg
3a1u f 5b2u (97)
8eg f 15b1g (89)
3a1u f 9eg (94)
17a1g f 15b1g (100)
3a1u f 9eg (100)
3a1u f 15b1g (100)
3a1u f 15b1g (100)
1.85
1.71
1.61
1.58
1.10
1.05
1.04
1.52(13B3g)a
2
2
11Eu
π,π*
13B1g
13Eu
dz ,dx -y
π,π*
1.30b
2
2
2
1.05 (13Au)e
0.90b,c
11B1u
13B1u
LMCT
LMCT
0.89b
Figure 4. Normalized ground-state absorption spectra of NiPc, NiNc,
NiPc(OBu)8, and NiNc(OBu)8 in 1-chloronaphthalene at room temperature.
a The energy values refer to the Jahn-Teller-distorted rectangular D2h
structures. b The energy values refer to the structures optimized under the
D4h symmetry constraint. c Computed at the optimized geometry of the
corresponding triplet. d The energy value refers to the Jahn-Teller-distorted
C2h structure. e The energy value refers to the Jahn-Teller-distorted Ci
structure.
Table 2. Vertical Excitation Energies and Oscillator Strengths (f)
Computed for the Optically Allowed Excited States Responsible for the
Q Band
TDDFT
composition
(%)
Expt
Eva
(eV/nm)
E
(nm)
state
f
NiNc, 11E for the R-alkoxy-substituted systems). The
calculated excitation energies are in satisfactory agreement
with the experiment, with the Q(0,0) λmax being underesti-
mated by no more than 0.13 eV (NiPc). However, the
discrepancy between theory and experiment varies from one
compound to another such that the Q-band maxima of NiNc
and NiPc(OMe)8, which are only 30 nm apart, are predicted
at nearly the same energy. The composition of the BP/ALDA
solution vectors, in terms of the major one-electron MO
transitions reported in Table 2, indicates that, in all com-
plexes, the Q state is largely derived from promotion of one
electron from the HOMO to the degenerate pair of unoc-
cupied Gouterman MOs. The increasingly reduced energy
gap between these orbitals along the series (see Figure 2)
fits in with the observed shift of the Q(0,0) band to longer
wavelengths. In the energy regime of the Q bands, a second,
intense excited state, the 31E, is predicted for the R-alkoxy-
substituted systems. This state, computed at 2.07 eV (599
nm) in NiPc(OMe)8 and at 1.98 (627 nm) in NiNc(OMe)8,
contributes to the broadening of the blue side of the Q-band
system.
2. Optically Silent Excited States Below the S1(π,π*)
State. In order to provide a theoretical basis for the
interpretation of the deactivation mechanism of the S1(Q)
state along the investigated series, we have examined the
lowest singlet and triplet excited states of NiPc, NiNc, NiPc-
(OMe)8, and NiNc(OMe)8. The vertical absorption energies
(Eva) calculated for the whole set of singlet and triplet states
up to ∼0.3 eV above the S1(Q) state are gathered in Tables
3 and 4, together with the adiabatic energies (Eadia) calculated
for selected states. In Figure 5, these selected excited states
are plotted for all investigated compounds, for the sake of
comparison.
NiPc
NiNc
11Eu 2a1u f 7eg (93)
11Eu 3a1u f 9eg (94)
1.98/627 0.588 670a
1.61/770 0.742 770a
b
NiPc(OMe)8 11E
17b1 f 38e (91) 1.60/775 0.513 740,a 734c
11B2 37e f 38e (92)
2.02/614 0.012
15a2 f 38e (92) 2.07/599 0.357
21b1 f 47e (94) 1.38/898 0.658 845,a 845d
19a2 f 47e (86) 1.98/627 0.302
31E
NiNc(OMe)8 11E
31E
11B2 46e f 47e (84)
2.01/617 0.007
a Data taken in 1-chloronaphthalene at room temperature; this work.
b Theoretical data from ref 10. c Data taken in toluene at room temperature;
ref 10. d Data taken in toluene at room temperature; this work.
From the level scheme in Figure 2, it may be inferred that
the relative position of the Gouterman LUMOs and that of
the unoccupied metal level (dx -y ) with respect to the HOMO
changes significantly along the series. This is crucial for the
photophysical properties (vide infra).
2
2
B. Excited States. 1. Ground-State Absorption Spectra.
The normalized ground-state absorption spectra of the
complexes, in the energy window of interest, are displayed
in Figure 4. The spectra show a red shift of the intense Q(0,0)
band upon either benzoannulation (100 nm shift) or octa-
butoxy substitution (70 nm shift). The Q(0,0) band of NiNc-
(OBu)8, where the effects of the benzoannulation and
R-butoxy substitution are both operative, is red shifted by
∼170 nm relative to that of NiPc. Looking at the overlaid
spectra in Figure 4, it is apparent that R-butoxy substitution
of both the Pc and Nc rings also induces a substantial
broadening of the Q bands, a feature that can be linked to
the conformational flexibility of the peripheral butoxy groups.
An interpretation of the spectral changes observed along
the series is provided by TDDFT calculations of the lowest
optically allowed excited states. According to the vertical
absorption energies (Eva) and oscillator strengths gathered
in Table 2, the intense Q(0,0) band is invariably assigned to
the lowest optically allowed excited state (11Eu for NiPc and
Considering first NiPc, the TDDFT results in Table 3
indicate that five dark states lie vertically below the S1(π,π*)
Inorganic Chemistry, Vol. 46, No. 6, 2007 2085

Soldatova et al.
Table 4. Excitation Energies (eV), Composition, and Character of the
Lowest Excited States of NiPc(OMe)8 and NiNc(OMe)8
state
composition (%)
character
Eva
Eadia
a
NiPc(OMe)8
23A2
13B1
23B2
23E
37e f 38e (100)
37e f 38e (100)
37e f 38e (100)
37e f 22b2 (78)
36e f 22b2 (22)
22a1 f 22b2 (98)
17b1 f 38e (91)
17b1 f 22b2 (100)
17b1 f 22b2 (100)
17b1 f 38e (100)
π,π*
π,π*
π,π*
1.88
1.86
1.83
1.83
2
2
LMCT/dπ,dx -y
13B2
11E
dz ,dx -y
π,π*
1.69
1.60
1.28
1.27
1.15
1.39b
2
2
2
11A2
13A2
13E
LMCT
LMCT
π,π*
1.15b,c
1.13b
1.09 (13B2)d
NiNc(OMe)8
dz ,dx -y
LMCT/dπ,dx -y
13B2
23E
27a1 f 27b2 (100)
46e f 27b2 (86)
45e f 27b2 (13)
21b1 f 47e (94)
21b1 f 47e (100)
21b1 f 27b2 (100)
21b1 f 27b2 (100)
1.63
1.61
1.31b
2
2
2
2
2
11E
π,π*
π,π*
LMCT
LMCT
1.38
0.92
0.84
0.84
13E
0.88 (13B2)d
0.67b,c
11A2
13A2
0.66b
a Data from ref 10. b The energy values refer to the structure optimized
under the D4h symmetry constraint. c Computed at the optimized geometry
of the corresponding triplet. d The energy values refer to the Jahn-Teller-
distorted C2V structures.
Figure 5. Excited-state diagram for NiPc, NiNc, NiPc(OMe)8, and NiNc-
(OMe)8. The (d,d) states are indicated with green lines, and LMCT states
are indicated with red lines.
geometries of the 3(d,d) and 3LMCT states are characterized
by a considerable lengthening of the Ni-Np distance and a
sizable expansion of the macrocycle core. These geometrical
changes are directly related to occupation of the 11b1g, which
is a strongly σ-antibonding, Ni-Np orbital.29 As for the 13Eu
(π,π*) excited state, it JT distorts into a very slightly
“stepped” C2h structure. However, as can be inferred from
the structural parameters gathered in Table S1, this state fails
to show significant geometrical relaxation. Consistent with
the minor conformational relaxation, the adiabatic excitation
energy of the JT-distorted 13Au (π,π*) is only 0.07 eV (see
Table 3) lower than the vertical absorption energy. Because
of the different relaxation, the minima of the relaxed energy
state. They are the 13Eg, 13B1g, 11B1u, 13B1u, and 13Eu, which
appear vertically at 1.77, 1.62, 1.47, 1.45, and 1.37 eV,
respectively. These are the most plausible candidates as
deactivation channels for the initially excited S1(π,π*) state.
According to the BP/ALDA solution vectors in Table 3,
the 13Eg and 13B1g both have (d,d) character, being described
2
2
2
by the 6eg f 11b1g (dπ f dx -y ) and 13a1g f 11b1g (dz f
d
3
2
2
x -y ) transitions, the 1 Eu is the normally emissive T1(π,π*)
state, and the 11,3B1u are pure 2a1u f 11b1g LMCT states.30-35
As inferred from the data in Table 3, the adiabatic energies
of the LMCT and (d,d) states are significantly decreased
relative to the corresponding vertical absorption energies,
1,3
surfaces of the 13B1g (dz ,dx -y ), 1 B1u (LMCT), and
13Au(π,π*) (T1 in Figure 5) lie at nearly the same energy.
A gap of ∼0.2 eV separates this manifold of excited states
2
2
2
particularly that of the 13B1g (dz ,dx -y ) (0.28 eV). This fits
in with the relaxed geometry of these states (note that the
doubly degenerate 13Eg state Jahn-Teller (JT) distorts into
a “rectangular”,36,37 D2h structure), showing significant
changes with respect to the ground-state geometry, as can
be seen in Table S1, which collects the optimized geometries
of the lowest triplet excited states of NiPc. The relaxed
2
2
2
from the upper lying 13B3g (dπ,dx -y ) state (see the excited-
state diagram in Figure 5).
2
2
According to TDDFT results, other excited states may
become involved in the decay of the primary excited S1(π,π*)
state, besides the ones just discussed. These are the 13A2g,
13B2g, and 13A1g MLCT states arising from the 6eg f 7eg
(dπ f π*) transition, which lie vertically at nearly the same
energy as the S1(π,π*). These states are not expected to
undergo significant conformational relaxation; therefore, they
are likely to intervene only at the very early stage of the
S1(π,π*) decay.
(30) What we call LMCT transitions are not really ligand-to-metal CT
transitions, the classification in terms of LMCT used here being merely
based on the character of the orbitals involved in the transitions. In
fact, a thorough analysis of the changes in Mulliken population
accompanying these transitions reveals that no net charge transfer
occurs from the ligand to the metal, but only a reorganization of the
electronic density occurs, which happens quite frequently in the so-
called LMCT states of transition-metal complexes (for an extensive
discussion on this topic, see ref 31). Therefore, the well-known failure
of TDDFT in predicting the excitation energies of charge-transfer states
correctly (see refs 31-35) does not occur here, and the TDDFT
energies calculated for the LMCT states can be trusted.
(31) Rosa, A.; Ricciardi, G.; Gritsenko, O.; Baerends, E. J. Struct. Bonding
2004, 112, 49.
(32) Casida, M. E.; Gutierrez, F.; Guan, J.; Gadea, F.-X.; Salahub, D.;
Daudey, J.-P. J. Chem. Phys. 2000, 113, 7062.
(33) Dreuw, A.; Weisman, J. L.; Head-Gordon, M. J. Chem. Phys. 2003,
119, 2943.
(34) Dreuw, A.; Head-Gordon, M. J. Am. Chem. Soc. 2004, 126, 4007.
(35) Gritsenko, O.; Baerends, E. J. J. Chem. Phys. 2004, 112, 655.
(36) Cory, M. G.; Hirose, H.; Zerner, M. C. Inorg. Chem. 1995, 34, 2969.
(37) Nguyen, K. A.; Pachter, R. J. Chem. Phys. 2003, 118, 5802.
The excited-state diagram in Figure 5 clearly indicates that,
3
in substituted NiPcs, the (dz ,dx -y ), 1,3LMCT, and T1(π,π*)
are the most plausible candidates as deactivation channels
for the initially excited S1(π,π*) state. The relative energies
of these states vary significantly along the series due to the
change in the relative position of the Gouterman LUMOs
2
2
2
2
2
and the unoccupied metal level (dx -y ) with respect to the
HOMO (see Figure 2). Thus, it can be anticipated that the
deactivation pathway of the S1(π,π*) state might differ from
one compound to another.
2086 Inorganic Chemistry, Vol. 46, No. 6, 2007

Photophysical BehaWior of Nickel Phthalocyanines
The excited-state pattern of NiNc reveals that benzo-
annulation of the phthalocyanine ring, while having very little
impact on the energy of the 3(d,d) states, causes a downward
shift of the S1(π,π*), T1(π,π*), and 1,3LMCT states, resulting
from the upshift of the HOMO (Figure 2). The important
3
2
2
consequences are that (i) the (dπ,dx -y ) becomes nearly
degenerate with the S1 (it lies adiabatically ∼0.1 eV below
the S1) and, hence, might be involved only in the early stage
3
2
2
2
of the deactivation process and (ii) the (dz ,dx -y ) is no longer
degenerate with the T1(π,π*) and 1,3LMCT states. Notably,
the 1,3LMCT states undergo a particularly large stabilization;
they are actually the lowest excited states due to the
concomitant upshift of the HOMO and downward shift of
2
2
the Ni-dx -y (see Figure 2).
The upshift of the HOMO results in the downward shift
of the S1(π,π*), T1(π,π*), and 1,3LMCT states going from
NiPc to its R-methoxy derivative. It is seen that R-methoxy
substitution also modifies the energy of the metal states
3
3
2
2
2
2
2
(Figure 5). These (dπ,dx -y ) and (dz ,dx -y ) lie at a somewhat
higher energy in NiPc(OMe)8 than in NiPc and in NiNc,
which is caused by the saddling-induced upshift of the Ni-
dx -y orbital. In turn, the 1,3LMCT states end up slightly
above the T1(π,π*).
2
2
As inferred from the excited-state diagram in Figure 5,
the most conspicuous effect of R-methoxy substitution of
the Nc ring is the stabilization of the S1(π,π*), T1(π,π*),
and 1,3LMCT states, resulting from the upshift of the HOMO
going from NiNc to NiNc(OMe)8 (see Figure 2). It is worth
noting that, owing to the huge downward shift of the S1-
(π,π*) in NiNc(OMe)8, the minimum of the relaxed energy
3
2
2
2
surface of the (dz ,dx -y ) is located immediately below this
state and, hence, is expected to play a role, if any, only at
the early stage of the deactivation mechanism. Just as in
NiPc(OMe)8, the (dπ,dx -y ) lies vertically well above the
S1(π,π*). Therefore, even though the relaxation energy would
be considered, it is unlikely that this state might play a role
in the deactivation of the S1(π,π*).
Figure 6. Transient absorption behavior of NiPc in 1-chloronaphthalene
solution photoexcited at 625 nm. (a) Cooling of the vibrationally hot (d,d)
state; (b) decay of the (d,d) state. Inset: kinetic profile at 675 nm.
3
2
2
C. Photophysics. 1. Transient Absorption Experi-
ments: Spectral Observations and Dynamic Properties.
Photoexcitation experiments were carried out for all members
of the series except NiNc, which had insufficient solubility
in all solvents, thus prohibiting the recording of useful
transient absorption data.
1.a. NiPc. The excited-state deactivation mechanism of
NiPc has not been fully studied heretofore. An early
investigation of the excited-state dynamics of NiPc by
Millard and Green38 reported only the relaxation of the first
excited singlet state, with a lifetime of 2.8 ps. These authors
suggested that the S1 state decayed to “a variety of possible
states of different spin multiplicity”.38 Their single probe
wavelength measurements were unable to detect intermediate
species involved in the deactivation sequence. To clarify this
issue, transient absorption experiments with excitation at 625
nm were carried out on NiPc (20 µM) in 1-chloronaphtha-
lene, a noncoordinating solvent in which the complex shows
acceptable solubility. Figure 6 shows spectral profiles at a
Figure 7. Transient absorption spectra of NiPc(OBu)8 in 1-chloronaph-
thalene photoexcited at 625 nm corresponding to three species participating
in the deactivation pathway. Inset: time profile at 750 nm.
series of delay times after 625 nm excitation. The earliest
time spectrum in Figure 6 shows no indication of significant
positive absorption in the spectral region at the blue side of
the ground-state Q bands. This is markedly different from
the situation with NiPc(OBu)810 (also see Figure 7) and other
MPc and MNc systems6,15,16 where absorptions at the blue
(38) Millard, R. R.; Green, B. I. J. Phys. Chem. 1985, 89, 2976.
Inorganic Chemistry, Vol. 46, No. 6, 2007 2087

Soldatova et al.
side of the Q bands have been assigned to transitions
localized on the respective π systems, namely, S1 f Sn or
T1 f Tn. Thus, the first observed transient (FOT) is, most
likely, not the S1(π,π*) state. The earliest time transient
absorption spectra (Figure 6a) have an asymmetric derivative-
like shape and experience a blue shift with time, finally
evolving at 31 ps to a spectrum having symmetric derivative-
like shape (the 50 ps spectrum in Figure 6b). This latter is
typical of a “classic” (d,d) state spectrum, as observed in
nominally planar nickel porphyrins,39-41 NiOMTP (OMTP
) octamethylthioporphyrin),42 and NiPc(OBu)8.10 The blue
shift and spectral narrowing of the early spectral feature have
been previously observed in other nickel tetrapyrrole
systems10,39-43 and have been ascribed to vibrational relax-
ation and conformational readjustment within the state. In
NiPc, this cooling shows biphasic kinetics (Figure 6b, inset);
the fast process with a 1.4 ps lifetime is probably relaxation
within the molecule (IVR), whereas the slower one (τ ∼ 18
ps) represents thermal losses to solvent oscillators. After
vibrational cooling, the (d,d) state deactivates to the ground
state, as shown by isosbestic behavior at 680 nm (Figure
6b), with a lifetime of 300 ps.
Figure 8. Evolution of the transient absorption spectra of NiNc(OBu)8 in
toluene (ca. 30 µM). The energy of the pump was 6 µJ/pulse.
1.b. NiPc(OBu)8. For the sake of comparison with NiPc,
the photophysical behavior of NiPc(OBu)8 has been rein-
vestigated in 1-chloronaphthalene. Figure 7 shows three
spectral profiles taken at different post-pulse delays after 625
nm excitation. Comparison with previous results in toluene10
reveals that the spectral and temporal characteristics are
substantially the same as those in toluene.
1.c. NiNc(OBu)8. Ultrafast pump-probe absorption spec-
trometry employing excitation wavelengths of 330 (B band),
845 (Q-band maximum), and 800 (vibrational peak of the Q
band) nm was used to investigate the excited-state behavior
of NiNc(OBu)8. The transient species formed were probed
in the visible and near-IR region of the spectrum. Ultrafast
experiments were conducted either in toluene or in 1-chlo-
ronaphthalene. As no significant solvent effect was detec-
ted, only the results obtained in toluene are here reported
and discussed. Figure 8 shows overlaid transient absorption
spectra recorded at different delay times after excitation of
NiNc(OBu)8 with 800 nm pulses. Figure 9 shows tran-
sient absorption spectra of the excited-state species probed
in the near-IR region. No dependence of the transient spectral
and temporal characteristics on excitation wavelength was
observed.
Figure 9. Transient absorption spectra of NiNc(OBu)8 in toluene photo-
excited at 800 nm in near-IR region. The energy of the pump was 5 µJ/
pulse.
with a maximum at 650 nm and two negative bands at 475
and 750 nm due to the ground-state depopulation.
The transient absorption spectra in the near-IR region
(Figure 9) were dominated by a strong negative absorption
band due to the ground-state depopulation in the region of
the ground-state Q-band maximum, with a small positive
absorption band to the red at around 900 nm. Figure 10
shows the temporal profiles of the transient absorption
displayed by NiNc(OBu)8 taken at four spectral regions. The
decay of the positive absorption signal as well as the recovery
of the negative signal were well-fitted by a three-exponential
function to a zero baseline, providing three lifetimes.
Table S4 summarizes the kinetic data for probe wave-
lengths ranging from 475 to 920 nm for the compound. As
can be inferred from this table, the lifetime of the fastest
component and its relative contribution to the overall decay
showed strong wavelength dependence, varying from ca. 0.3
The spectral feature formed within the instrument response
time (ca. 200 fs) is characteristic of excited states localized
on the π system of phthalocyanine and naphthalocyanine
macrocycles,6,10,15,16,44-47 namely, a broad positive absorption
(39) Rodriguez, J.; Holten, D. J. Chem. Phys. 1989, 91, 3525.
(40) Rodriguez, J.; Kirmaier, C.; Holten, D. J. Am. Chem. Soc. 1989, 111,
6500.
(41) Rodriguez, J.; Kirmaier, C.; Holten, D. J. Chem. Phys. 1991, 94, 6020.
(42) Rosa, A.; Ricciardi, G.; Baerends, E. J.; Zimin, M.; Rodgers, M. A.
J.; Matsumoto, S.; Ono, N. Inorg. Chem. 2005, 44, 6609.
(43) Drain, C. M.; Kirmaier, C.; Medforth, C. J.; Nurco, D. J.; Smith, K.
M.; Holten, D. J. Phys. Chem. 1996, 100, 11984.
(44) Gunaratne, T. C.; Kennedy, V. O.; Kenney, M. E.; Rodgers, M. A. J.
J. Phys. Chem. A 2004, 108, 2576.
(45) Fournier, M.; Pepin, C.; Houde, D.; Ouellet, R.; van Lier, J. E.
Photochem. Photobiol. Sci. 2004, 3, 120.
(46) Aoudia, M.; Cheng, G.; Kennedy, V. O.; Kenney, M. E.; Rodgers,
M. A. J. J. Am. Chem. Soc. 1997, 119, 6029.
(47) Firey, P. A.; Ford, W. E.; Sounik, J. R.; Kenney, M. E.; Rodgers, M.
A. J. J. Am. Chem. Soc. 1988, 110, 7626.
2088 Inorganic Chemistry, Vol. 46, No. 6, 2007

Photophysical BehaWior of Nickel Phthalocyanines
Figure 10. Kinetic profiles of the transient absorption signal at various probe wavelengths for NiNc(OBu)8 in toluene after photoexcitation at 800 nm.
to ca. 3.5 ps. A plot of this first lifetime versus wavelength
is shown in Figure 11.
three-exponential expression to obtain satisfactory fits to the
time profiles. The main positive absorption band, λmax ) 650
nm, decayed with minimal spectral changes, save for the
development of a poorly resolved peak at 730 nm on the
red side of the major absorption band. Leaving aside a
consideration of the very early, wavelength-dependent kinetic
component for the moment, what is observed is that the
absorption growth at 730 nm has a similar lifetime to the
second decay component at shorter wavelengths. This
suggests that the 730 nm species is generated from the 650
nm transient. However, it is clearly not a simple situation
because, having been produced, the 730 nm species decays
to the baseline with the same lifetime as observed for the
third component of the decay at all other wavelengths,
indicating that the two species have a common lifetime. This
can be understood if the 730 and 650 nm transients are in
dynamic equilibrium, and the 19 ps growth lifetime is the
rate at which the equilibrium is established, and the 500 ps
decay, at all wavelengths, is the decay of the equilibrium
state. A major requirement for the equilibrium condition is
that the two states participating are close in energy. Turning
to Table 4, it is seen that candidate states for satisfying this
condition are the 13B2 (π,π*) and the nearly degenerate 1,3A2
(LMCT) pair. Although the calculated energy difference is
The intermediate lifetime, although longer, showed a
similar wavelength dependence (Table S4); the slowest
lifetime component was essentially wavelength independent
at ca. 500 ps. Figure 12 illustrates the kinetic behavior of
the transient absorption at 730 nm, where a characteristic
spectral shoulder arises at later delay times.
2. Interpretation of the Photophysical Behavior. The
experimental and theoretical results described previously can
be employed to elucidate the dynamics of the excited states
of the compounds studied and to comment on the separate
and combined effects of R-butoxy substitution and benzo-
annulation. The photophysical behavior of NiNc(OBu)8, not
reported before, will be considered first.
2.a. The Photophysical Behavior of NiNc(OBu)8. The
salient features of the transient spectra observed immediately
after pump-probe excitation of NiNc(OBu)8 are as shown
in Figure 8. There is a broad, almost featureless positive ∆A
region stretching from ca. 500 to ca. 740 nm, with negative
∆A regions outside this range. During the first 15 ps or so,
the regions of negative ∆A underwent slight spectral shifts
to the red, after which no further shifting was observed.
Throughout the spectral range, it was necessary to use a
Inorganic Chemistry, Vol. 46, No. 6, 2007 2089

Soldatova et al.
phosphorescence emission at 0.93 eV.6 In addition, TDDFT
results on PdNc(OMe)8 allocated T1(π,π*) vertically at 0.95
eV,48 almost isoenergetic with that for T1(π,π*) in NiNc-
(OBu)8, calculated here to be 0.92 eV (Table 4). Taking this
energy argument into account, it would not be unreasonable
to conclude that the behavior of the 3(π,π*) in NiNc(OBu)8
would be similar to that exhibited by the Pd complex.
However, the value of the 500 ps lifetime measured here
for the decay of the 650 nm transient in NiNc(OBu)8 seems
much too short for the spin-forbidden repopulation of the
ground state from the π-localized triplet excited state, and
only electronic factors, involving metal orbitals, that are
missing in the Pd complex can explain the short T1(π,π*)
lifetime. Indeed, in NiNc(OBu)8, the proximity of the LMCT
and the T1(π,π*) states, as predicted by the calculations,
supports the proposal that these states are in dynamic
equilibrium, allowing an understanding of why the lifetime
is less than expected. Under conditions where the reactions
participating in the equilibrium are proceeding much more
rapidly than the decays of the individual participating states,
the lifetime of the equilibrium state is a linear combination
of the two individual lifetimes, weighted by the mole fraction
contributions of each.49
Figure 11. Upper panel: first lifetime/wavelength dependence. Lower
panel: two transient absorption spectral cuts taken at 2 and 10 ps for NiNc-
(OBu)8 in toluene.
In summary, it is considered most likely that photoexci-
tation within the Q band of NiNc(OBu)8 generates the 11E
3
(π,π*) state, which converts to the (π,π*) state within the
instrument response time (ca. 200 fs). This π-localized triplet
state in our experiment is seen as the first observed transient
(FOT). This state rapidly (ca. 19 ps lifetime) and reversibly
3
1
converts into the LMCT or the LMCT state, and the
equilibrium state, so generated, decays to the ground state
with a lifetime of ca. 500 ps. That the intersystem crossing
(ISC) process 11E f 13E (13B2) occurs so rapidly (within
the instrument response time) is certainly highly unusual;
metallophthalocyanines with p-block metal centers show ISC
rate constants in the 1 ns-1 range.16,44 The most likely
explanation of this is that ISC is facilitated by the 13B2
2
2
2
(dz ,dx -y ) state. This state has, indeed, a calculated adiabatic
energy of 1.31 eV, some 70 meV above the 11E (Q) state.
Such an energy difference is within the confidence limits of
the calculations, and it is not unlikely that these states are
close to degenerate and, thus, interact strongly, providing
fast intersystem crossing.
Figure 12. Kinetic profiles of the transient absorption signal at 730 nm
for NiNc(OBu)8 in toluene after photoexcitation at 800 nm.
∼0.2 eV, there is enough uncertainty in the calculations for
the states to be actually closer than this.
Taking as a working hypothesis that the FOT (650 nm) is
the 13B2 (π,π*) state and the 730 nm daughter is one of the
1,3A2 (LMCT) pair, and that the two are in a dynamic
equilibrium, it is instructive to seek other evidence. First,
Returning to the kinetic data, we have argued that the
second and third exponential components correspond, re-
3
3
spectively, to the conversion of the (π,π*) to the LMCT
or 1LMCT state and to the latter’s decay to the ground state;
it remains to consider the earliest component. Considering
that the FOT is identified as the T1(π,π*) that must be formed
very rapidly from S1(π,π*), it must initially carry excess
vibrational energy, and given that the spectral shape did not
change during this decay event, it is not unreasonable to
assign the first exponential component to the cooling of the
state. In NiNc(OBu)8, the first and second decay components
3
2
2
2
unlike the case of NiPc(OMe)8, the B2 (dz ,dx -y ) state has
a calculated energy that is too high for it to be involved in
the deactivation of the S1(π,π*) state produced by excitation
in the Q band. Moreover, no tell-tale absorption pattern of a
(d,d) state was seen in the experimental spectra (see, e.g.,
the NiPc spectra in Figure 6). The spectral shape of the FOT,
namely, a broad positive absorption at around 650 nm, is
reminiscent of the transient absorption signal arising from
the T1 f Tn transition of the Pd(II)-octabutoxynaphthalo-
cyanine studied earlier.6 The Pd analogue, however, has a
triplet lifetime of hundreds of nanoseconds and showed
(48) Soldatova, A. V.; Kim, J.; Rosa, A.; Ricciardi, G.; Kenney, M. E.;
Rodgers, M. A. J. Manuscript in preparation.
(49) Birks, J. B. Photophysics of Aromatic Molecules; Wiley: London,
1970.
2090 Inorganic Chemistry, Vol. 46, No. 6, 2007

Photophysical BehaWior of Nickel Phthalocyanines
a 1-chloronaphthalene solution, where the three spectral
3
3
2
2
2
signatures of the S1 (1.3 ps), (dz ,dx -y ) (7.4 ps), and LMCT
states (220 ps) are closely similar to those observed in toluene
solution.10 However, in the case of NiPc (Figure 6), the
deactivation pathway appears to be quite different, allowing
an interesting comparison to be drawn between the early
processes in deactivation of the S1 states of NiPc and NiPc-
(OBu)8. The octabutoxy derivative (Figure 7) shows clear
evidence of a S1 f Sn transition in the 600 nm region with
a lifetime of 1.9 ps; the unsubstituted compound shows no
such spectral evidence. In fact, the FOT is a hot (d,d) state.
Thus, NiPc, as NiNc(OBu)8 considered previously, exhibits
an ultrafast ISC process out of the photogenerated Q state,
probably induced by the proximity of the 13A1g MLCT state.
The TDDFT results presented in Table 3 indicate that the
3
2
2
next state below is the 1 Eg (dπ,dx -y ) state, lying adiabati-
cally at 1.56 eV. Below this state, there are four states lying
Figure 13. A schematic explanation of the wavelength dependence of
the first lifetime (see text).
3
2
2
2
adiabatically very close in energy, namely, the (dz ,dx -y
)
state at 1.34 eV, the nearly degenerate 11,3B1u LMCT pair
close to 1.30 eV, and the 13Au (π,π*) at 1.30 eV. It must be
stated, however, that the ordering of these states may differ
on account of the uncertainty in the calculated energies. What
is clear from the experimental data is that the state that
precedes the recovery to the ground-state surface is a cooled
(d,d) state, and the lowest-lying of these (Table 3) is the
showed significant wavelength dependence (see Figure 11
for the early component). Interestingly, as the figure shows,
at the two crossings between regions of negative and positive
absorption, the decay lifetimes are shortest; in the region
between, the dependence is bell shaped. This can be
understood as a manifestation of intramolecular vibrational
relaxation from high-lying vibrational quantum states where
the wave functions are such that the oscillators have a high
probability of being near the classical turning points.50 The
reason for the bell-shaped nature of the lifetime wavelength
dependence (Figure 11) must arise from the detailed nature
of the upper and lower potential energy surfaces, and a
simplistic schematic explanation is presented in Figure 13.
There, it is seen that upper and lower electronic curves are
drawn to be slightly offset, such that upward transitions from
the two turning points, labeled A and B, occur at different
energies and correspond to the two lifetime minima in Figure
11. It is likely that a similar explanation holds for the second
temporal component.
13B1g (dz ,dx -y ). In all likelihood then, the immediate
precursor of the ground state is the 13B1g state with a lifetime
2
2
2
3
of some 300 ps, similar to that of (d,d) states in some Ni
porphyrins.39-42 Whether a vibrationally hot version of this
state, or a hot 13Eg state, is the one that is observed as the
FOT is not clear. Both of these states, being of (d,d) nature,
will have absorption spectra that are dominated by the π
system and only affected in a minor way by the metal
electronic configuration. The UV-vis transient absorption
spectra in Figure 6 are insufficiently detailed to provide a
resolution of this question.
The large destabilization of the HOMO in NiPc(OMe)8
due to an antibonding interaction with the oxygen lone pairs
considerably lowers the energy of the LMCT states, remov-
2.b. Effects of the Octabutoxy Substitution and Benzo-
annulation on the Deactivation of the Excited States. A
comparison of the photophysical behaviors of NiPc and
NiPc(OBu)8 allows the effects of octabutoxy substitution to
be understood. In the report on the relaxation of the NiPc-
(OBu)8 excited states,10 the experimental and theoretical
results have shown that the S1 state, generated upon
photoexcitation into the Q band, deactivates through the
3
2
2
2
ing their quasidegeneracy with the (dz ,dx -y ) state that
occurs in NiPc. This makes possible the population of the
3LMCT state observed in the transient absorption experiment.
Concerning the effect of the benzoannulation of the Pc
ring on the deactivation of the S1(π,π*), we are not able to
draw definitive conclusions, as insufficient solubility of NiNc
in noncoordinating solvents prevented us from obtaining
transient absorption data on this compound. However, on
the basis of TDDFT calculations and transient absorption
data obtained for NiNc(OBu)8, some projections can be
3
3
2
2
2
(dz ,dx -y ) state, which in turn, converts to the LMCT state
that finally repopulates the ground-state surface. Figure 7
represents time-resolved absorption data for NiPc(OBu)8 in
3
2
2
2
made. For example, the (dz ,dx -y ) does not appear in
photoexcited NiNc(OBu)8 during the ultrafast transient
absorption experiment because of its near-degeneracy with
the S1 state. Nevertheless, it might play a more distinct role
in the excited-state deactivation in the NiNc complex since,
in that complex, this near-degeneracy is removed (Figure
5), and the energy gap between S1 and the (d,d) state is more
like that in the NiPc(OBu)8 complex, where the (d,d) state
(50) The optical absorption spectrum of an excited vibronic state arises
from the sum of all possible Franck-Condon transitions to vibrational
levels in a higher electronic state. When the origin state is vibrationally
cold, transitions originating at the equilibrium geometry are favored
because the wave function has maximum amplitude there. However,
when the origin state is hot, upward transitions originating closer to
the classical turning points become increasingly probable. By their
very nature, the vibrational states with high quantum numbers have a
higher density of vibrational oscillators below them than do states
with lower quantum numbers, and they will therefore cool faster,
resulting in different wavelength regions having different cooling
lifetimes, as observed here.
3
2
2
was evident. Moreover, both the T1(π,π*) and LMCT(π,dx -y
)
Inorganic Chemistry, Vol. 46, No. 6, 2007 2091

Soldatova et al.
Figure 14. Proposed schematic diagrams for the excited-state relaxation pathways of NiPc, NiNc, NiPc(OBu)8, and NiNc(OBu)8 in 1-chloronaphthalene.
would most likely play a role in the deactivation mechanism,
as in the case in NiNc(OBu)8.
Figure 14 concisely summarizes the photophysical proper-
ties of the metallotetrapyrroles investigated herein.
3(d,d) state, which cools within 20 ps. This cooled (d,d) state,
probably 13B1g (dz ,dx -y ), repopulates the ground state with
a lifetime of 300 ps.
2
2
2
Previous and present results for NiPc(OBu)8 indicate that,
in this complex, the S1(π,π*) state deactivates through the
Conclusions
3
*
2
2
2
(dz ,dx -y ), which is no longer degenerate with the T1(π,π )
1,3
3
2
2
2
2
2
and LMCT(π,dx -y ). The (dz ,dx -y ), in turn, converts to
The excited-state spectral and dynamic behavior of NiNc-
(OBu)8, a very promising photothermal sensitizer, has been
investigated by ultrafast transient absorption spectrometry
and interpreted in light of DFT/TDDFT theoretical studies.
It is found that the initially formed S1(π,π*) state deactivates,
within the time resolution of the instrument (200 fs), to the
vibrationally hot T1(π,π*) state. The quasidegeneracy of the
3
2
2
the LMCT(π,dx -y ), which finally repopulates the ground-
state surface within 640 ps. Insufficient solubility of NiNc
in noncoordinating solvents prevented us from obtaining
transient absorption data. However, on the basis of TDDFT
calculations and transient absorption data obtained for NiNc-
(OBu)8, it can be anticipated that (i) the (dz ,dx -y ) that does
not appear spectroscopically during the ultrafast transient
absorption experiment in the NiNc(OBu)8 complex due to
its near-degeneracy with the S1 state might participate in the
excited-state deactivation in NiNc and (ii) both the T1(π,π*)
3
2
2
2
3
2
2
2
S1(π,π*) and (dz ,dx -y ) allows for fast intersystem crossing
to occur. After vibrational relaxation (ca. 2.5 ps), the T1(π,π*)
converts rapidly (ca. 19 ps lifetime) and reversibly into the
3
2
2
LMCT(π,dx -y ) state. The so generated equilibrium state
3
2
2
and LMCT(π,dx -y ) states are likely to play a role in the
deactivation mechanism.
decays to the ground state with a lifetime of ca. 500 ps.
With the aim to independently characterize the impact of
R-octabutoxy substitution and benzoannulation of the phthalo-
cyanine ring on the deactivation pathway of the S1(π,π*)
state, the ground- and excited-state properties and the
photophysical behavior of NiPc, NiNc, and NiPc(OBu)8 have
also been examined. Ultrafast experiments and DFT/TDDFT
calculations have consistently shown that peripheral substitu-
tion of the Pc ring modifies the photodeactivation mechanism
of the S1(π,π*) by inducing substantial changes in the relative
In this study, toluene and chloronaphthalene have been
employed as solvents to avoid coordinating reactions at the
metal site. It can be questioned as to whether this is a relevant
model medium for biological media. The compounds are
exceedingly lipophillic, and in cellular and tissue studies,
they are necessarily administered via liposomes.10 In vivo,
the compounds reside in hydrophobic organelles, and for this
reason, it would appear that toluene can be regarded as a
not-inappropriate environment for photophysical studies.
The work described here lends credence to the argument
that a comprehensive understanding of the deactivation
mechanisms of metallotetrapyrrole excited states, particularly
when the metal is a first-row transition metal, can only be
properly gained when the natures and energies of the
3
3
*
2
2
2
2
2
energies of the S1(π,π*), (dπ,dx -y ), (dz ,dx -y ), T1(π,π ),
and 1,3LMCT(π,dx -y ) excited states. The actual position of
these states has been found to markedly depend on the
location of the Gouterman LUMOs and the unoccupied metal
2
2
2
2
level (dx -y ) with respect to the HOMO. In NiPc, the
S1(π,π*) state undergoes ultrafast (<200 fs) ISC into a hot
2092 Inorganic Chemistry, Vol. 46, No. 6, 2007

Photophysical BehaWior of Nickel Phthalocyanines
spectroscopically silent states lying between the photo-
generated state and the ground state are known. Such
information is available only through theoretical approaches,
and this report adds to the growing evidence10,27,42,51 that the
combination of DFT and TDDFT is capable of achieving
considerable accuracy in excitation energy calculations for
large molecules such as MTPs. This knowledge of the nature
and energies of the low-lying excited states along the
relaxation pathways has proven extremely useful for inter-
preting the results of transient absorption experiments and
to assist in understanding the differences in the observed
deactivation rates from compound to compound.10,42
were supported, in part, by NIH Grant CA 91027 and by an
instrumentation grant from the Hayes Investment Foundation
(Ohio Board of Regents). A.V.S. thanks the McMaster
Foundation at BGSU for a predoctoral fellowship. A.R. and
G.R. wish to thank the Italian MIUR (Ministero dell’
Istruzione, dell’ Universita` e della Ricerca) and the Universita`
della Basilicata, Italy, for a grant (Grant 2003038084_002).
Grateful thanks are expressed to Dr. Eugene Danilov for the
technical assistance with ultrafast instrumentation.
Supporting Information Available: Optimized geometrical
parameters for the lowest triplet excited states of NiPc, NiNc, and
NiNc(OMe)8 (Tables S1-S3). Lifetimes and relative amplitudes
of the transient absorption dynamics of NiNc(OBu)8 in toluene after
800 nm excitation (Table S4). This material is available free of
charge via the Internet at http://pubs.acs.org.
Acknowledgment. The photophysical studies conducted
at the Ohio Laboratory for Kinetic Spectrometry (BGSU)
(51) Rogers, J. E.; Nguyen, K. A.; Hufnagle, D. C.; McLean, D. G.; Su,
W.; Gossett, K. M.; Burke, A. R.; Vinogradov, S. A.; Pachter, R.;
Fleitz, P. A. J. Phys. Chem. A 2003, 107, 11331.
IC061524O
Inorganic Chemistry, Vol. 46, No. 6, 2007 2093