Synthesis, transient absorption, and transient resonance Raman spectroscopy of novel electron donor-acceptor complexes: [5,15-bis[(4'- nitrophenyl)ethynyl]-10,20-diphenylporphinato]copper(II) and [5-[[4'- (dimethylamino)phenyl]ethynyl]-15-[(4-n

Article abstract of DOI:10.1021/ja964436j

We report the synthesis, transient absorption, FT Raman, resonance Raman, time-resolved resonance Raman, and transient resonance Raman spectra of pseudo-D(2h) symmetric [5,15-bis[(4'-nitrophenyl)ethynyl]-10,20- diphenylporphinato]copper(II) (I) and electr

Full text of DOI:10.1021/ja964436j

12578
J. Am. Chem. Soc. 1997, 119, 12578-12589
Synthesis, Transient Absorption, and Transient Resonance
Raman Spectroscopy of Novel Electron Donor-Acceptor
Complexes: [5,15-Bis[(4′-nitrophenyl)ethynyl]-10,20-
diphenylporphinato]copper(II) and [5-[[4′-(Dimethylamino)phen-
yl]ethynyl]-15-[(4′′-nitrophenyl)ethynyl]-10,20-diphenyl-
porphinato]copper(II)
Steven M. LeCours,†,§ Charles M. Philips,† Julio C. de Paula,*,‡ and
Michael J. Therien*,†
Contribution from the Department of Chemistry, UniVersity of PennsylVania,
Philadelphia, PennsylVania 19104-6323, and Department of Chemistry, HaVerford College,
370 Lancaster AVenue, HaVerford, PennsylVania 19041-1392
ReceiVed December 27, 1996. ReVised Manuscript ReceiVed May 16, 1997X
Abstract: We report the synthesis, transient absorption, FT Raman, resonance Raman, time-resolved resonance Raman,
and transient resonance Raman spectra of pseudo-D2h symmetric [5,15-bis[(4′-nitrophenyl)ethynyl]-10,20-diphe-
nylporphinato]copper(II) (I) and electronically asymmetric [5-[4′-(dimethylamino)phenyl]ethynyl]-15-[(4′′-nitrophenyl)-
ethynyl]-10,20-diphenylporphinato]copper(II) (II), which bears both electron-releasing and electron-withdrawing groups
conjugated directly to the porphyrin periphery. The spectroscopic results suggest extensive electronic communication
between the 5- and 15-arylethynyl groups and the porphyrin core. Relative to the parent compound, (tetraphe-
nylporphinato)copper(II) (CuTPP), the arylethynyl substituents increase the lifetime of the excited trip-multiplet
states. CuTPP, as well as compounds I and II, however, shows similar solvent-dependent dynamics: the trip-
multiplet lifetimes are longer in a noncoordinating solvent such as benzene than in a coordinating solvent such as
THF. This behavior is consistent with the existence of a quenching state whose effect is more pronounced upon
coordination of solvent. The time-resolved resonance Raman spectrum of compound II shows features commonly
associated with the relatively long-lived triplet excited states of copper(II) porphyrins. The transient resonance Raman
spectrum of a short-lived excited state present in both compounds I and II is characterized by marked shifts in the
nitro and porphyrin stretching frequencies relative to that observed for the ground states of both (4-nitrophenyl)-
ethyne and (tetraphenylporphinato)copper(II). We interpret these results for the compounds I and II as arising from
(i) a short-lived excited state present at early time that possesses enhanced porphyrin-to-nitro charge-transfer character
with respect to the ground state and (ii) a longer-lived excited state deriving from this initially probed charge-
transfer state that is largely porphyrin localized.
Introduction
systems with unique photophysical properties.7 Figure 1 shows
examples of such structures and illustrates a new approach to
engineering D-A complexes: pseudo-D2h symmetric [5,15-bis-
[(4′-nitrophenyl)ethynyl]-10,20-diphenylporphinato]copper(II)
(I) and an electronically asymmetric version of the [5,15-bis-
(arylethynyl)-10,20-diphenylporphinato]metal structural motif,
[5-[[4′-(dimethylamino)phenyl]ethynyl]-15-[(4′′-nitrophenyl)-
ethynyl]-10,20-diphenylporphinato]copper(II) (II). There are
three key design features in these complexes. First, the [5,15-
bis(arylethynyl)-10,20-diphenylporphinato]metal core,8 synthe-
sized via metal-mediated cross-coupling of substituted aryl-
One of the goals of the developing field of molecular
electronics is the design of donor-acceptor (D-A) complexes
capable of light-induced charge separation over long distances
on the molecular scale.1-3 For example, the fastest possible
light-addressable electronic communication between nanoscale
circuitry requires that the interconnects access very large
molecular transition dipoles. As a result, ultrafast charge-
transfer systems will likely figure prominently in proposed
photo-driven molecular-scale rectifiers, transistors, and switches,4
new technologies focusing on light-modulated data storage and
retrieval,5 as well as novel computing architectures and ultras-
mall memory devices.6
(4) (a) Aviram, A.; Ratner, M. A. Chem. Phys. Lett. 1974, 29, 277-
283. (b) Miller, J. S. AdV. Mater. 1990, 2, 378-379. (c) Ebbesen, T. W.
New J. Chem. 1991, 15, 191-198. (d) O’Neil, M. P.; Niemczyk, M. P.;
Svec, W. A.; Gosztola, D.; Gaines, G. L., III; Wasielewski, M. R. Science
1992, 257, 63-65.
(5) (a) Emmelius, M.; Pawlowski, G.; Vollmann, H. W. Angew. Chem.,
Int. Ed. Eng. 1989, 28, 1445-1471. (b) Parthenopoulos, D. A.; Rentzepis,
P. M. Science 1989, 245, 843-845. (c) Hunter, S.; Kiamilev, F.; Esener,
S.; Parthenopoulos, D. A.; Rentzepis, P. M. Appl. Opt. 1990, 29, 2058-
2066.
(6) (a) Hopfield, J. J.; Onuchic, J. N.; Beratan, D. N. Science 1988, 241,
817-820. (b) Hopfield, J. J.; Onuchic, J. N.; Beratan, D. N. J. Phys. Chem.
1989, 93, 6350-6357.
(7) (a) Lin, V. S.-Y.; DiMagno, S. G.; Therien, M. J. Science 1994, 264,
1105-1111. (b) Lin, V. S.-Y.; Therien, M. J. Chem.-Eur. J. 1995, 1, 645-
651. (c) Goll, J. G.; Moore, K. T.; Ghosh, A.; Therien, M. J. J. Am. Chem.
Soc. 1996, 118, 8344-8354.
Toward these ends, we have developed new porphyrin-based
* To whom correspondence should be addressed.
† University of Pennsylvania.
‡ Haverford College.
§ Present address: Mobil Technology Company, Paulsboro Technology
Center, Paulsboro, NJ 08066.
X Abstract published in AdVance ACS Abstracts, December 15, 1997.
(1) (a) Carter, F. L. Molecular Electronic DeVices I; Decker: New York,
1982. (b) Carter, F. L. Molecular Electronic DeVices II; Decker: New York,
1987.
(2) Lehn, J. M. Angew. Chem., Int. Ed. Engl. 1988, 27, 89-112.
(3) Burland, D. M.; Miller, R. D.; Walsh, C. A. Chem. ReV. 1994, 94,
31-75.
S0002-7863(96)04436-8 CCC: $14.00 © 1997 American Chemical Society

NoVel Electron Donor-Acceptor Complexes
J. Am. Chem. Soc., Vol. 119, No. 51, 1997 12579
small decreases in absorbance intensity (1-2%) after 5 min and
moderate photobleaching (5-6%) after 45 min of continuous irradiation.
Resonance and FT Raman Spectroscopies. All Raman spectra
(FT, resonance, and time-resolved resonance) were acquired as
described previously.12,13 The ground state resonance Raman spectra
were recorded with the same instrument and geometries as the transient
resonance Raman spectra except that a continuous-wave Liconix 2040
HeCd laser (excitation wavelength 441.6 nm, 25 mW) was used. The
FT Raman spectra were obtained with a Nicolet Raman 950 spectrom-
eter with a Nd:YVO4 laser (1064 nm, 300 mW) for excitation. Solvent
bands were digitally subtracted.
Transient Resonance Raman Spectroscopy. One-laser transient
Raman spectra were obtained with pulsed excitation at 436 nm (5 ns,
10 Hz), with instrumentation described previously.13 High-power
spectra were obtained at 16 or 28 mW while low-power spectra were
obtained by placing neutral-density filters in front of the high-power
beam to attenuate the power (1.2 or 2.1 mW). The low-intensity spectra
seen in Figures 7B and 8B were acquired by defocusing an ap-
proximately 10 mW beam and signal-averaging the data over several
hours in order to obtain transient Raman data in which ground state
modes dominated the spectrum. All spectra were recorded at room
temperature in the back-scattering geometry. Spectral subtractions and
data handling were performed with the methodology described by de
Paula et al.13 to ensure data quality and reproducibility. Band maxima
reported in this paper were reproducible to <2 cm-1 under the
conditions employed. Compounds I and II were examined by FT
Raman and ground state electronic absorption spectroscopy and thin-
layer chromatography (TLC) before and after transient resonance Raman
spectroscopy experiments were performed; in all cases, no significant
decomposition was observed. While we did observe some irreversible
photochemistry of compound I during the transient absorption measure-
ments (see the Results), the back-scattering geometry of the Raman
measurements coupled with the higher sample concentrations used in
these experiments appear to have minimized the effect even in spite of
the fact that photon fluxes are typically higher in Raman experiments
than in absorption experiments.
Two-laser time-resolved Raman spectra were obtained with the same
spectrograph and detector described previously.13 The pump wave-
length of λpump ) 436 nm (tightly focused, 5 mW, 10 Hz) was generated
as described above. The probe beam at λprobe ) 532 nm was the second
harmonic of a second Continuum Surelite laser (defocused, 20 mW,
10 Hz). The beams were made to excite the same rectangular spot on
the sample. Raman scattering was collected in the back-scattering
geometry at room temperature. Both lasers were triggered externally
by a Stanford Research Systems DG535 digital delay generator. The
time delay between the pulses was ∆t ) 10 ns.
Electronic Structure Calculations. Frontier orbital energies for
compounds I and II were determined by the ZINDO method with
standard INDO-1 semiempirical parameters.14 Molecular structure files
were created in the following manner: zinc analogs of compounds I
and II were first constructed with D2h or Cs symmetry, respectively, in
which the central zinc metal atoms were assigned a dsp2 (square planar)
geometry. For computational simplicity, ZINDO-optimized geometrical
structures were obtained in which the dihedral angles of the 5- and
15-arylethynyl phenyl moieties were initially adjusted to 0° and the
10- and 20-meso-phenyl groups were fixed at 90° with respect to the
porphyrin least-squares plane. The convergence criteria for these
restricted Hartree-Fock (RHF) self-consistent field (SCF) calculations
required the root mean square difference in the elements of the density
matrix to be below 0.000 001 on two successive SCF cycles. Once an
optimized (porphinato)zinc(II) structure was obtained, the central
porphyrin zinc atom was replaced by a copper(II) center (dsp2
hybridization; square planar geometry) and a ZINDO energy calculation
was performed. Because the d9 copper(II) electronic configuration has
a ground state doublet spin multiplicity, a restricted open-shell Hartree-
Fock (ROHF) SCF calculation was run (compound I, CI ) 18;
Figure1. [5,15-Bis[(4′-nitrophenyl)ethynyl]-10,20-diphenylporphinato]-
copper(II) (I) and [5-[[4′-(dimethylamino)phenyl]ethynyl]-15-[(4′′-
nitrophenyl)ethynyl]-10,20-diphenylporphinato]copper(II) (II).
ethynyl moieties to meso-halogenated porphyrin templates,9 has
highly polarized, singly degenerate excited states.8,10 Second,
electronic coupling between the porphyrin core and its substit-
uents is optimized by the cylindrically π symmetric ethynyl
linker. Because steric factors do not preclude a coplanar
arrangement of the aromatic ring systems along the C2 molecular
axis of highest conjugation, a highly extended π system is
effectively created, leading to the expectation that the arylethynyl
phenyl unit and the polarizable porphyrin macrocycle will
communicate electronically in both the ground and excited
states. Finally, augmented electronic coupling between the
(porphinato)copper(II) core and the 5- and 15-meso substituents
is expected to enhance charge transfer to the electron-withdraw-
ing NO2 group.
We present transient absorption and transient resonance
Raman spectra of I and II. Our results indicate that (i) there is
extensive electronic coupling between the arylethynyl and
porphyryl moieties, as predicted above, (ii) this electronic
coupling in the short-lived, initially probed excited state exceeds
that present in the ground state, (iii) this short-lived excited state
with enhanced charge-transfer character evolves into a trip-
multiplet state, (iv) the excited trip-multiplet states for com-
pounds I and II have substantially longer lifetimes with respect
to that reported for the parent compound, (tetraphenylporphi-
nato)copper(II) (CuTPP), and (v) both I and II show solvent-
dependent excited state dynamics similar to that observed for
CuTPP: the trip-multiplet lifetimes are longer in a noncoordi-
nating solvent such as benzene than in a coordinating solvent
such as THF. We conclude that, on the basis of their
photophysical properties, compounds I and II show promise as
structural motifs in the development of molecular electronic
devices.
Experimental Section
Transient Absorption Spectroscopy. Ground state electronic
spectra were recorded on an OLIS UV/vis/near-IR spectrophotometry
system that is based on the optics of a Cary 14 spectrophotometer.
The transient absorption experimental setup housed at the University
of Pennsylvania Regional Laser and Biotechnology Laboratory has been
described previously.11 The samples were prepared in distilled, dry
solvents, and degassed by three freeze-pump-thaw cycles. Ground
state electronic spectra of both compounds were recorded before and
after laser irradiation to assess photostability. While compound II
showed only slight or no photobleaching (<2%) irrespective of solvent
even after prolonged laser irradiation (t > 45 min), compound I showed
(8) LeCours, S. M.; DiMagno, S. G.; Therien, M. J. J. Am. Chem. Soc.
1996, 118, 11854-11864.
(9) (a) DiMagno, S. G.; Lin, V. S.-Y.; Therien, M. J. J. Am. Chem. Soc.
1993, 115, 2513-2515 . (b) DiMagno, S. G.; Lin, V. S.-Y.; Therien, M. J.
J. Org. Chem. 1993, 58, 5983-5993.
(12) de Paula, J. C.; Walters, V. A.; Nutaitis, C.; Lind, J.; Hall, K. J.
Phys. Chem. 1992, 96, 10591-10594.
(13) de Paula, J. C.; Walters, V. A.; Jackson, B. A.; Cardozo, K. J. Phys.
Chem. 1995, 99, 4373-4379.
(10) LeCours, S. M.; Jahn, L.; Therien, M. J. Manuscript in preparation.
(11) Papp, S.; Vanderkooi, J. M.; Owen, C. S.; Holtom, G. R.; Phillips,
C. M. Biophys. J. 1990, 58, 177-186.
(14) ZINDO Software provided by: CAChe Scientific (Beaverton,
Oregon).

12580 J. Am. Chem. Soc., Vol. 119, No. 51, 1997
LeCours et al.
compound II, CI ) 12). The convergence criteria were identical to
those used for geometry optimization. The results of these calculations
were pictorially generated (tabulated) utilizing an isosurface value of
0.04.
Materials. All compounds were handled using methods and
protocols described previously.9
(5,15-Dibromo-10,20-diphenylporphinato)zinc(II), (4-Substituted
phenyl)acetylenes, and [5-[[4′-(Dimethylamino)phenyl]ethynyl]-15-
[(4′′-nitrophenyl)ethynyl]-10,20-diphenylporphinato]copper(II). p-
(Dimethylamino)phenyl)acetylene and (p-nitrophenyl)acetylene were
synthesized similarly to literature methods.15 The syntheses of (5,15-
dibromo-10,20-diphenylporphinato)zinc(II),9 [5,15-bis[(4′-nitrophenyl)-
ethynyl]-10,20-diphenylporphinato]zinc(II),8 and [5-[[4′-(dimethylamino)-
phenyl]ethynyl]-15-[(4′′-nitrophenyl)ethynyl]-10,20-diphenylpor-
phinato]copper(II)16 have been previously described.
[5,15-Bis[(4′-nitrophenyl)ethynyl]-10,20-diphenylporphinato]cop-
per(II) (I). [5,15-Bis[(4′-nitrophenyl)ethynyl]-10,20-diphenylporphi-
nato]zinc(II)8 was placed in a 125-mL Erlenmeyer flask and dissolved
with a 1:1 mixture of THF-CHCl3 (60 mL). To the green porphyrin
solution was added dropwise a 2 M HCl solution diluted in 1:1 THF-
2-propanol. After 5 min, the reaction mixture was poured into a 250-
mL separatory funnel containing a 1 M KOH solution (30 mL). The
mixture was shaken and the organic layer collected and washed once
more with distilled water. The organic solvents were then removed
by flash evaporation, and the crude free base porphyrin was used
without further purification.
5,15-Bis[(4′-nitrophenyl)ethynyl]-10,20-diphenylporphyrin (9.3 mg,
12 mmol) was added to a 250-mL round bottom flask containing toluene
(100 mL) and cupric acetate dihydrate (110 mg, 551 mmol). After
being refluxed for 8 h, the reaction mixture was placed in a 250-mL
separatory funnel and washed twice with 1 M ammonium hydroxide
solution (30 mL) and once with water. The toluene layer was
evaporated to a 15-mL total volume and chromatographed on silica
gel using toluene as the eluant. A green band was collected and
recrystallized to give compound I, isolated as a dark green solid (5.1
mg, 51%). Vis (THF): 450, 636.
Figure 2. Electronic spectra of (porphinato)copper complexes recorded
in THF: (A) compound I; (B) compound II; (C) CuTPP.
Results
Ground State Electronic Spectroscopy. Compared to that
of the parent compound, CuTPP, the ground state electronic
spectra of complexes I and II exhibit red shifted, broader B-
(also called Soret) and Q-type absorption bands at about 450
nm and 630-650 nm, respectively (Figure 2), consistent with
inclusion of the arylethynyl moieties in the π frameworks of I
and II. Studies of electronically symmetric [5,15-bis(arylethyn-
yl)-10,20-diphenylporphinato]zinc(II) complexes demonstrate
that a large share of the oscillator strength of both the B and Q
transitions is polarized along the C2 molecular axis of highest
conjugation.8,10
Transient Absorption Spectroscopy. Figures 3 and 4 show
the transient absorption spectra at various time delays for
compounds II and I, respectively, in THF and benzene. Table
1 lists the excited state lifetimes for I and II and compares them
to data for the CuTPP parent compound.
The spectra are dominated by a transient bleach nearly
coincident with the location of the ground state Soret band
(∼450 nm) and by transient absorption bands on both sides of
this bleach. The spectra recorded in THF and benzene for II
(parts A and B, respectively, of Figure 3) are characterized by
symmetrical recovery of all transient features and by two
(15) (a) Eastmond, R.; Walton, D. R. M. Tetrahedron 1972, 28, 4591-
4599. (b) Westmijze, H.; Vermeer, P. Synthesis 1979, 390-392. (c)
Takahashi, S.; Kuroyama, Y.; Sonogashira, K.; Hagihara, N. Synthesis 1980,
627-630. (d) Mesnard, D.; Bernadou, F.; Miginiac, L. J. Chem Res. (S)
1981, 270-271. (e) Zhang, Y.; Wen, J. Synthesis 1990, 727-728. (f) Park,
K. M.; Schuster, G. B. J. Org. Chem. 1992, 57, 2502-2504.
(16) (a) LeCours, S. M.; Guan, H.-W.; DiMagno, S. G.; Wang, C. H.;
Therien, M. J. J. Am. Chem. Soc. 1996, 118, 1497-1503. (b) Priyadarshy,
S.; Therien, M. J.; Beratan, D. N. J. Am. Chem. Soc. 1996, 118, 1504-
1510.
Figure 3. Transient absorption spectrum of compound II recorded at
various time delays: (A) THF, 15 f 50 ns; (B) benzene, 15 f 400
ns.
solvent-independent isosbestic points at 396 and 483 nm. The
Soret bleach recovers and the transient absorption decays with
indistinguishable, monoexponential kinetics.
Unlike the transient absorption spectra for compound II, those

NoVel Electron Donor-Acceptor Complexes
J. Am. Chem. Soc., Vol. 119, No. 51, 1997 12581
Table 1. Excited State Triplet Lifetime Data of I and II Obtained from Transient Absorption Experimentsa
triplet state lifetime (ns)
compound Ib
absorption decay
compound IIb
solvent
Soret recovery
absorption decay
Soret recovery
CuTPPc
35
benzene
77 ( 6 (75%)
45 ( 2
92 ( 3d
82.3 ( 0.9
(7 ( 2) × 102 (25%)
5.3 ( 0.6 (97%)
(6 ( 1) × 102 (3%)
THF
5.8 ( 0.4e
8.3 ( 0.5
7.2 ( 0.1
0.035f
a Experimental conditions: pump wavelength 630 nm, power 150 mW, T ) 25 °C. All the time-resolved data could be fit as either single
exponential (compound I) or biexponential (compound II) decays over a minimum of 5 lifetimes of the fast component. Analyses of the goodness
of the data fit corresponded to minimum values of ø2 of 0.995 and 0.9994 for the biexponential and single exponential decay processes, respectively.
Wavelengths chosen for monitoring transient absorption and bleaching were those that corresponded to the maximal extinction coefficient. b Rate
constants reported ( their standard errors. The percentage of respective components for the biexponential decay processes is given in parentheses.
c See ref 28a. d The data are better fit to a biexponential decay (ø2 ) 0.9993) with τ1 ) 79 ( 7 ns (93%) and τ2 ) 610 ( 1100 ns (7%). e The data
fit a single exponential decay after utilizing a nonzero baseline correction. f In THF, CuTPP decays with two time constants, 35 and 190 ps; the
former time constant has been assigned to the decay of the triplet state while the nature of the state giving rise to the latter decay has not been
identified.25,28
Figure 4. Transient absorption spectrum of compound I recorded at
various time delays: (A) THF, 15 f 300 ns; (B) benzene, 15 f 900
ns.
Figure 5. FT Raman spectra of (porphinato)copper complexes I and
obtained for I (Figure 4) are more complex and lack well-defined
II recorded in THF solution. (A) compound I; (B) compound II
isosbestic points. Moreover, compound I’s absorption decay
(excitation wavelength 1064 nm, laser power 300 mW, T ) 25 °C).
kinetics are not single exponential in nature. With increasing
time, the Soret bleach feature observed in benzene (Figure 4B)
appears to red shift by 3 nm, flatten, and eventually split. The
the FT Raman spectra of compounds I and II (λex ) 1064 nm)
while parts A and B of Figure 6 display their respective
Soret bleach recovery (λ ) 455.6 nm) for compound I in
resonance Raman spectra (λex ) 441.6 nm). Despite key
benzene is monoexponential, but its lifetime (44.7 ( 2.0 ns)
differences in electronic and geometric structure between the
differs from that of the fast component of the biexponential
[5,15-(arylethynyl)porphinato]copper(II) complexes (compounds
absorption decay (76.7 ( 6.3 ns) evaluated at λ ) 508.1 nm.
I and II) and CuTPP, a number of striking similarities exist in
In THF (Figure 4A), the fast components of the absorption decay
their ground state FT Raman, resonance Raman, and transient
(5.26 ( 0.56 ns) and the Soret bleach recovery (5.82 ( 0.39
Raman spectra (Vide infra). These spectral correspondences
have allowed us to assign qualitatiVely the mode compositions
for these novel chromophores in lieu of performing a normal
coordinate analysis.
ns) are indistinguishable. However, the ground state does not
recover completely even after t > 1 µs. Also, other spectral
features appear in the 470-490 nm region of the spectrum over
the 100-300 ns time domain. The complex nature of compound
The bands around 1600 cm-1 (phenyl stretch), 1560 cm-1
I’s transient absorption spectra suggests the irreversible forma-
(ν2; CâCâ), 1364 cm-1 (ν4; CRN, CRCâ), and 1230-1240 cm-1
tion of a photoproduct and is consistent with the partial
photobleaching of I that occurs over the course of the transient
(ν1; CmesoCPh)17 are assigned readily by analogy with the
vibrational spectrum of CuTPP. Primarily due to their frequen-
cies, two other nonporphyrinic peaks can also be assigned: the
absorption study (see the Experimental Section).
FT and Resonance Raman Spectroscopy. Figure 5 shows

12582 J. Am. Chem. Soc., Vol. 119, No. 51, 1997
LeCours et al.
Figure 6. Resonance Raman spectra of the ethynyl and aromatic/porphyrin regions for compounds I and II recorded in THF solution: (A) compound
I; (B) compound II (excitation wavelength 441.6 nm, laser power 25 mW, T ) 25 °C).
ethynyl stretch at around 2195 cm-1 and the nitro stretch18 at
1337-1343 cm-1
of the (dimethylamino)phenyl moiety (Figure 6B).21 The 1585
cm-1 band probably arises from ν11 (CRCmeso).19,20
.
Many of the remaining bands in Figures 5 and 6 are not
observed in the Soret-excited resonance Raman spectra of
CuTPP.19 As expected for symmetries lower than D4h, many
modes gain considerable intensity in compounds I and II, giving
rise to more complicated spectra. On the basis of the normal
coordinate analyses for CuTPP by Czernuszewicz et al.20 and
Atamian et al.,19 we tentatively assign the bands at 1498-1507
and 1537-1538 to ν12 (CâCâ) and ν20 (CRCmeso), respectively.
The main difference between the vibrational spectra of
compounds I and II is that the resonance Raman spectrum of
the latter has two additional peaks at 1611 and 1585 cm-1. The
1611 cm-1 band clearly derives from the phenyl C-C stretch
Transient Resonance Raman Spectroscopy of Compound
I. Figure 7 shows the transient resonance Raman spectra for
compound I in the aromatic/porphyrin region, obtained with
436 nm excitation (pulse width 5 ns; repetition rate 10 Hz).
With the exception of frequency shifts of several vibrational
bands, the spectrum of Figure 7A, obtained with high laser
power, has many features in common with the Soret-excited
resonance Raman spectrum of the ground state (Figure 6A).
These data are consistent with the spectrum of a single state at
high laser powers; this situation is somewhat unique in transient
Raman spectra of metalloporphyrins, where typically the high-
power spectrum shows distinct contributions from both the
ground state and one or more excited states (see de Paula et
al.13 and references cited therein). It was necessary to defocus
the laser pulse substantially in order to obtain the spectrum in
Figure 7B, which resembles the FT Raman spectrum of
compound I (Figure 5A). These observations indicate that the
high-power spectrum of Figure 7A arises from an excited state
of compound I.
The minor differences between the resonance Raman spec-
trum of Figure 6A and the transient resonance Raman spectrum
of Figure 7B are probably due to one or more of three effects.
First, the spectrum in Figure 7B has a low signal-to-noise ratio,
which makes the determination of peak frequencies difficult.
Second, the spectra in Figures 6A and 7B were obtained at
different excitation wavelengths (441.6 in Figure 6A vs 436
nm in Figure 7B). It is likely that the Soret band of compound
I (Figure 2) is inhomogeneously broadened due to (i) the
(17) Note: this peak has been labeled as both a ν1 (CmesoCphenyl
)
stretch13,20 and the C mode;19 we have elected to use the ν1 convention.
Our choice of assignment is partly based on the behavior of this mode upon
formation of excited states (Vide infra). In CuTPP, the ν1 mode has a
modestly small downshift of 2-4 cm-1 upon photoexcitation. Two other
modes we considered as plausible alternatives to the ν1 assignment for the
1228 cm-1 band were ν22 (CRN) and ν31 (CâH in-plane deformation). While
the ν22 mode would be expected to downshift, the ν31 mode would be
expected to be relatively insensitive to photoexcitation. The ν31 mode has
b2g symmetry in both D4h and D2h point groups and so would be expected
to be Raman active. The ν22 mode, though formally Raman inactive (a2g
symmetry) in the D4h point group, becomes the Raman allowed (b1g) in
D2h symmetry porphyrins. On the basis of this reasoning, we suggest that
the 1228 cm-1 band arises from either ν1 or ν22. Regardless, whether this
mode is ν1 or in fact ν22 does not impact our general analysis of the transient
Raman spectrum of compound I.
(18) Akins, D. L.; Guo, C.; Zhu, H.-R. J. Phys. Chem. 1993, 97, 3974-
3977.
(19) Atamian, M.; Donohoe, R. J.; Lindsey, J. S.; Bocian, D. F. J. Phys.
Chem. 1989, 93, 2236-2243.
(20) Czernuszewicz, R. S.; Macor, K. A.; Li, X.-Y.; Kincaid, J. R.; Spiro,
T. G. J. Am. Chem. Soc. 1989, 111, 3860-3869.
(21) LeCours, S. M.; de Paula, J. C.; Therien, M. J. Unpublished results.

NoVel Electron Donor-Acceptor Complexes
J. Am. Chem. Soc., Vol. 119, No. 51, 1997 12583
obtained at 436 nm in the interpretation of one-laser transient
Raman spectra.
The low- and high-power transient Raman spectra of the
ethynyl stretching frequency region of compound I (data not
shown)23 show that the ethynyl stretch occurs at the same
frequency (2195 cm-1) as it does in both the FT Raman (Figure
5A) and ground state resonance Raman (Figure 6A) spectra.
Although the scaled difference spectrum for the porphyrin/
aromatic region of compound I shows no unusual spectral
features (data not shown), the analogous scaled difference
spectrum in the ethynyl region23 shows a trough at the peak of
the ethynyl stretching frequency and two peaks symmetrically
disposed around the trough. These data suggest that the ethynyl
stretching band becomes slightly broader at high laser powers.
Performing a multipoint fit to force a flat baseline for I’s low-
and high-power transient Raman spectra and taking a scaled
difference still evinces a broadened ethynyl mode at high laser
power.23 While power broadening is generally thought to arise
from systems with either unrelaxed low-frequency inner sphere
or solvent vibrations, or nonlinear resonance processes,24 given
the time scale of the transient experiment and the modestly-
low-power densities employed in this study, it is unlikely that
in this instance power broadening derives from unrelaxed
molecular vibrations. On the other hand, nonlinear resonance
scattering occurs typically at higher powers than used in this
study. For example, the resonance Raman spectrum of â-car-
otene in isopentane exhibits nonlinear resonance scattering at
power levels approaching 1 × 109 W/cm2,24b whereas the peak
power level in our experiments was 2 × 107 W/cm2. Although
it may be attractive to hypothesize that such an effect may be
in accord with the established nonlinear optical properties of
this class of chromophores,16 because the extent of broadening
of the ethynyl band is minor and its mechanism unclear from
these experiments, these data will not be analyzed further in
the Discussion. While we note that this will not impact our
interpretation of the excited state dynamics of compound I (Vide
infra), we provide below justification for the apparent lack of
a frequency shift in this band upon electronic excitation.
Transient Resonance Raman Spectra of Compound II.
The low- and high-power transient resonance Raman spectra
for compound II are shown in Figure 8. In spite of the lower
symmetry of compound II (Cs), its transient resonance Raman
spectra are very similar to those observed for compound I.
However, three differences observed in the high-intensity
spectrum are worthy of mention (Figure 8A). First, the
additional feature at 1607 cm-1 observed in the high-intensity
spectrum is related to the phenyl C-C stretch of the (dimethyl-
amino)phenyl moiety (Vide supra). Second, the band at 1450
cm-1 in Figure 8A has little or no intensity in either the ground
state resonance Raman or low-laser-power transient resonance
Raman spectrum, but possesses moderately high intensity in
the high-power spectrum. Finally, we tentatively assign the
band at 1509 cm-1 to the ν12 (CâCâ) mode, which has been
identified in the resonance Raman spectrum of CuTPP+ (Vide
infra).20
Figure 7. Transient resonance Raman spectra of the aromatic/porphyrin
region of compound I in THF obtained with pulsed excitation at 436
nm: (A) high-intensity spectrum (power 28 mW); (B) low-intensity
spectrum (laser power , 1 mW).
disparity in oscillator strength between the extensively split x-
and y-polarized transitions, (ii) the coupling of other electronic
(e.g., CT) states to the normally porphyrin-localized S2 state,
(iii) spectral contributions from ground state conformations that
differ with respect to the torsional angle between the arylethynyl
phenyl ring and the porphyrin macrocycle, and (iv) vibronic
coupling.8,16 Such spectral complexity will likely result in the
observation of vibrational signatures in the Raman experiment
that are strongly dependent upon the excitation wavelength
within the Soret envelope.22 Finally, Figure 7B may contain a
contribution from the excited state that gives rise to the spectrum
in Figure 7A. We justify this conclusion with the observation
that the spectrum in Figure 7B has a broad feature in the 1300-
1385 cm-1 range, whereas the excited state (Figure 7A) and
pure ground state spectra (Figures 5A and 6A) show only two
major peaks in the same region. For these reasons, we will (i)
interpret only the most intense bands in Raman spectra obtained
under low-power pulsed excitation and (ii) use only data
(22) The discrepancies between data at 441.6 and 436 nm suggest
complex excitation profile dependence for these modes, similar to what
has been observed previously for (octaethylporphinato)nickel(II) (NiOEP)
(Alden, R. G.; Crawford, B. A.; Doolen, R.; Ondrias, M. R.; Shelnutt, J.
A. J. Am. Chem. Soc. 1989, 111, 2070-2072). The excitation profile for a
number of modes in NiOEP was consistent with the existence of multiple
conformational forms in solution that differed in their degree of porphyrin
ring ruffling. Likewise, our data on compound I may be explained by the
existence of a distribution of macrocycle conformations in solution. For
example, conformers of compound I may exist in solution that differ with
respect to their arylethynyl phenyl-to-porphyrin dihedral angles. The ambient
temperature electronic spectrum of I (Figure 2) shows that the B band region
is broader than that observed for CuTPP; while this effect likely derives
primarily from the extensive splitting between the Bx and By states (see
refs 8 and 16b), conformational heterogeneity would influence the degree
of absorption band broadening as well. Consistent with this, it is important
to note that the nitro band in the FT Raman experiment is approximately
twice as broad as comparably intense peaks. We conclude that changing
the excitation wavelength from 441.6 (Figure 6A) to 436 nm (Figure 7B)
in the resonance Raman experiments likely selects populations of chro-
mophores that are not structurally isomorphous and thus produce similar
but nonsuperimposible spectra. This is congruent with the low-intensity
transient resonance Raman spectrum of Figure 7B which exhibits a large
number of overlapping bands.
The scaled transient Raman difference spectrum for the
ethynyl region of compound II (data not shown)23 possesses
power-broadened features similar to those observed for com-
pound I. Also as observed for compound I, the ethynyl
stretching band does not shift upon electronic excitation.
Time-Resolved Resonance Raman Spectrum of Com-
pound II. We were not able to observe reproducible features
(23) These data are available as Supporting Information.
(24) (a) Carroll, P. J.; Brus, L. E. J. Am. Chem. Soc. 1987, 109, 7613-
7616. (b) Carroll, P. J.; Brus, L. E. J. Chem. Phys. 1987, 86, 6584-6590.
(c) Dick, B.; Hochstrasser, R. M. J. Chem. Phys. 1984, 81, 2897-2906.

12584 J. Am. Chem. Soc., Vol. 119, No. 51, 1997
LeCours et al.
Figure 8. Transient resonance Raman spectra of the aromatic/porphyrin
region of compound II in THF obtained with pulsed excitation at 436
nm: (A) high-intensity spectrum (power 16 mW); (B) low-power
spectrum (laser power , 1 mW). Inset: Time-resolved resonance
Raman spectrum of compound II in piperidine (λpump ) 436 nm, λprobe
) 532 nm, and ∆t ) 10 ns).
Figure 9. Frontier molecular orbitals for compound I. Electron density
localized on the 10- and 20-meso-phenyl moieties is not shown. Orbital
symmetries and the calculated energies are noted on the right. The half-
filled HOMO, which is largely metal dx -y in character, is not depicted.
This orbital is orthogonal to the porphyrin, ethyne, and (arylethynyl)-
phenyl π-symmetry orbitals, and hence does not play a significant role
in the analysis of the data reported herein.
in the two-laser time-resolved resonance Raman spectrum of
compound I, possibly due to its photoinstability (Vide supra).
While the time-resolved Raman spectra of compound II
generally exhibited low signal-to-noise ratios, it should be noted
that (i) two-laser nanosecond time-resolved resonance Raman
spectra of (porphinato)copper(II) complexes are notoriously
weak and (ii) the data obtained in our experiments is of quality
comparable to that of published benchmark spectra for such
species.25 We observe two bands at 1342 and 1362 cm-1 when
conditions are optimized for probing the species with a transient
2
2
density resides in one or both of the aryl moieties linked to the
macrocycle via the ethynyl group. Our results indicate the
following.
(i) The HOMO - 1 and HOMO - 2 are very close in
energy.8 This is due to the electron-withdrawing effect of the
nitrophenylethynyl moiety, which stabilizes the HOMO - 1
with respect to the HOMO - 2.8 In CuTPP the separation
between the highest filled a2u and a1u orbitals is large enough
to result in low-energy excited triplet states that are due to the
(a2ueg) configuration.25,27-29 By contrast, we cannot be certain
if, for example, the first excited triplet state of compound I will
be formed by removal of the au (a1u in D4h symmetry) or b1u
absorption feature at 520-530 nm (λpump ) 436 nm, λprobe
532 nm, and ∆t ) 10 ns, inset of Figure 8).
)
Electronic Structure of Compounds I and II. In order to
provide a framework in which to rationalize the results of our
spectroscopic investigations, we performed electronic structure
calculations (ZINDO; see the Experimental Section).14 Figures
9 and 10 display the frontier orbitals (FOs) of compounds I
and II, respectively, along with their calculated relative energies
and symmetries.
(27) (a) Antipas, A.; Dolphin, D.; Gouterman, M.; Johnson, E. C. J. Am.
Chem. Soc. 1978, 100, 7705-7709. (b) Spellane, P. J.; Gouterman, M.;
Antipas, A.; Kim, S.; Liu, Y. C. Inorg. Chem. 1980, 19, 386-391. (c) Kim,
D.; Holten, D.; Gouterman, M. J. Am. Chem. Soc. 1984, 106, 2793-2798.
(d) Asano, M.; Kaizu, Y.; Kobayashi, H. J. Chem. Phys. 1988, 89, 6567-
6576. (e) Yan, X.; Holten, D. J. Phys. Chem. 1988, 92, 5982-5986. (f)
Rodriguez, J.; Kirmaier, C.; Holten, D. J. Am. Chem. Soc. 1989, 111, 6500-
6506. (g) Asano-Someda, M.; Kaizu, Y. J. Photochem. Photobiol. A 1995,
87, 23-29. (h) Liu, F.; Cunningham, K. L.; Uphues, W.; Fink, G. W.;
Schmolt, J.; McMillin, D. R. Inorg. Chem. 1995, 34, 2015-2018. (i)
Cunningham, K. L., McNett, K. M.; Pierce, R. A., Davis, K. A.; Harris, H.
H.; Falck, D. M.; McMillin, D. R. Inorg. Chem. 1997, 36, 608-613.
(28) (a) Kruglik, S. G., Apanasevich, P. A., Chirvony, V. S., Kvach, V.
V. and Orlovich, V. A. J. Phys. Chem. 1995, 99, 2978-2995. (b) Asano-
Someda, M.; Sato, S.; Aoyagi, K.; Kitagawa, T. J. Phys. Chem. 1995, 99,
13800-13807.
Both Figures 9 and 10 suggest that, in the region of the
porphine macrocycle, the HOMO - 1, HOMO - 2, LUMO,
and LUMO + 1 are related to the a2u, a1u, and doubly degenerate
eg FOs of square-planar porphyrins.26 Unlike CuTPP, whose
low-energy electronic states may be explained by four porphy-
rin-based FOs, compounds I and II have a fifth orbital, the
LUMO + 2, which is nearly identical in energy with respect to
the eg-derived LUMO + 1, where the majority of the electron
(25) Jeoung, S. C.; Kim, D.; Cho, D. W.; Yoon, M. J. Phys. Chem. 1995,
99, 5826-5833.
(26) Gouterman, M. In The Porphyrins; Dolphin, D., Ed.; Academic
Press: London, 1978; Vol. III, pp 1-165.

NoVel Electron Donor-Acceptor Complexes
J. Am. Chem. Soc., Vol. 119, No. 51, 1997 12585
by the observation that the ethynyl stretches in compounds I
and II are downshifted 25-40 cm-1 relative to the ethynyl
frequencies typically observed in the Raman spectra of simple
diarylacetylenes.30 Ground state electronic coupling is less
strong between the central (porphinato)copper(II) unit and the
nitro substituent of the arylethynyl phenyl ring: the HOMO -
1 and HOMO - 2 do not exhibit appreciable electron density
at the nitro group, and consequently, the frequency of the nitro
band in both I and II is downshifted by only 5-10 cm-1 with
respect to that observed in the FT Raman spectrum of (4-
nitrophenyl)acetylene (1347 cm-1; data not shown).
Our data also suggest that electronic coupling between the
(porphinato)copper(II) core and the arylethynyl moieties is
different in the ground and excited singlet states. Table 2 shows
that the relative degree of enhancement of the Raman modes
in compounds I and II changes upon electronic excitation in
the Soret band. The nitro band is more intense than the ethynyl,
ν2 (CâCâ), or ν4 (CRN, CâCâ) band in the resonance Raman
spectrum, but is generally weaker than all of these modes in
the FT Raman spectrum. Furthermore, the ethynyl stretching
mode is weaker than the porphyrin bands in the resonance
Raman spectrum, but stronger than the porphyrin bands in the
FT Raman spectrum.
The data suggest that electronic excitation in the Soret band
results in a substantial redistribution of electron density; clearly
the nitro group is more strongly coupled to the porphyrin in
the initially-prepared singlet excited state with respect to the
ground state. By the same token, it appears that the ethynyl
group is not coupled more strongly in the singlet state(s)
accessible through Soret excitation than in the ground state.
These observations are consistent with our electronic structure
calculations, which show a significant increase in electron
density at the nitro group in the LUMO and LUMO + 2, but
predict less pronounced changes in electron density at the
ethynyl group when the LUMOs are populated at the expense
of the HOMOs.
Figure 10. Frontier molecular orbitals for compound II. Electron
density localized on the 10- and 20-meso-phenyl moieties is not shown.
Orbital symmetries and the calculated energies are noted on the right.
The half-filled HOMO, which is largely metal dx -y in character, is
not depicted.
2
2
Dynamics of Excited States Observed by Transient
Absorption Spectroscopy. We discuss our data on compounds
I and II in the context of theoretical and spectroscopic studies
performed on their parent compound, CuTPP.13,25,27-29 In
summary, it is known that (i) efficient intersystem crossing
induced by the paramagnetic d9 Cu(II) center shortens the sing-
doublet excited state lifetime to <350 fs, (ii) the initial sing-
doublet excited state decays to lower energy trip-doublet (2T)
and trip-quartet (4T) states, whose luminescence properties are
both solvent and temperature dependent, (iii) the interaction
(a2u in D4h symmetry) orbital (see Figure 9). We must rely on
spectroscopic evidence for such an assignment.
(ii) The LUMO and LUMO + 2 have considerable electron
density at the nitro moiety; for this reason, we predict that some
of the low-energy excited states have porphyrin-to-nitro charge-
transfer (CT) character.
(iii) The LUMO + 1 has negligible electron density in the
arylethynyl region; thus, we predict that electron promotion from
either the HOMO - 2 or HOMO - 1 to the LUMO + 1 results
in excited states that are localized largely within the porphyrin
macrocycle.
2
2
between the half-filled dx -y metal orbital and the empty
porphyrin eg* orbitals results in a metal-to-ring (ligand) (d, π*)
charge-transfer (MLCT) state, (iv) promotion of an electron from
the porphyrin-centered a2u orbital to the singly occupied copper
2
2
2
dx -y defines a ring-to-metal (π, d) LMCT state, and (v) a dz -
Discussion
2
2
to-dx -y (d, d) excited state can have energies similar to those
of the (π, π*) and CT states.
Electronic Coupling in Arylethynyl-Substituted Porphy-
rins. Our Raman spectral data (Figures 5 and 6) and electronic
structure calculations (Figures 9 and 10) provide evidence for
electronic coupling between the arylethynyl and (porphinato)-
copper(II) moieties in the ground state of compounds I and II.
The HOMO - 1 orbitals of compounds I and II indicate
delocalization of electron density between the porphyrin mac-
rocycle and the ethynyl groups. This conclusion is supported
Two models have been proposed to explain the dynamics of
2
the T-4T manifold. On the basis of the decrease of the trip-
multiplet lifetime in the presence of coordinating solvents, Kim
et al.27c proposed that coordination lowers the energy of an
excited state that quenches the 2T-4T manifold. Indeed,
transient resonance Raman spectroscopy has shown that coor-
dination of THF or piperidine results in the formation of an
excited state that has spectral features characteristic of either a
CT state13,25 or a (d, d) state.28a However, Asano-Someda and
Kaizu27g and Cunningham et al.27i explain modulations in the
(29) (a) Zerner, M.; Gouterman, M. Theor. Chim. Acta (Berlin) 1966, 4,
44-63. (b) Roos, B.; Sundbom, M. J. Mol. Spectrosc. 1970, 36, 8-25. (c)
Case, D. A.; Karplus, M. J. Am. Chem. Soc. 1977, 99, 6182-6194. (d)
Shelnutt, J. A.; Straub, K. D.; Rentzepis, P. M.; Gouterman, M.; Davidson,
E. R. Biochemistry 1984, 23, 3946-3954. (e) Stavrev, K.; Zerner, M. C.
Chem. Phys. Lett. 1995, 233, 179-184.
(30) In Raman/Infrared Atlas of Organic Compounds, 2nd ed.; Schrader,
B., Ed.; VCH Publishers: New York, 1989.

12586 J. Am. Chem. Soc., Vol. 119, No. 51, 1997
LeCours et al.
Table 2. Relative Degree of Vibrational Mode Enhancements Observed for Compounds I and II in the FT and Resonance Raman Spectraa
ratio of intensities
compound
experiment
ν2/ν4
NO2/ν2
NO2/ν4
NO2/ethyne
ν2/ethyne
ethyne/ν4
ν2/PhB
PhB/ν4
I
I
II
II
resonance Raman
FT Raman
resonance Raman
FT Raman
1.79
1.48
0.77
0.72
1.72
0.83
2.63
0.41
3.12
1.22
2.00
0.29
3.12
0.52
3.03
0.26
1.79
0.63
1.16
0.64
1.00
2.34
0.67
1.12
1.25
1.20
2.08
2.22
1.44
1.24
0.37
0.32
a Mode compositions are described in the text and listed in Table 3; PhB refers to the arylethynyl phenyl ring (see Figure 1).
excited state lifetime of CuTPP by invoking a vibronic inter-
action that broadens the excited state potential energy surface
and enhances radiationless decay processes. Presumably,
coordination of a solvent molecule to either the ground state
(as in the case of piperidine) or the excited state (as in the case
of THF) promotes the vibronic interaction through a doming
distortion that pulls the Cu2+ ion out of the porphine plane.27i
The transient absorption spectra of compounds I and II
(Figures 3 and 4) show features that are qualitatively similar to
those observed in triplet-triplet absorption spectra of (porphi-
nato)copper(II) complexes.27c,e,f,28a Hence, we assign the tran-
sient absorption features in the 300-400 and 500-550 nm
a vibronic interaction or direct quenching by a short-lived
excited state determines the relaxation time of the 2T-4T
manifold. Coordination of a solvent molecule in either the
ground or excited state enhances the relaxation. The observation
2
that quenching of the T-4T manifold is less effective in our
donor-acceptor complexes than in CuTPP may be a conse-
quence of the substantial electronic coupling that exists between
the porphyrin core and the meso-arylethynyl moieties. In order
to determine how the electronic coupling between these groups
in the triplet state compares with that observed for the ground
state, we examined the time-resolved resonance Raman spec-
troscopy of these species.
2
regions to electronic excitation of the T-4T states of com-
Dynamics of Excited States Observed by Time-Resolved
Resonance Raman Spectroscopy. Two-laser time-resolved
resonance Raman spectroscopy on the nanosecond time scale
probes the 2T-4T manifold of copper(II) porphyrins selec-
tively.25,28 Two distinct features are apparent in the time-
resolved resonance Raman spectrum of compound II (Figure
pounds I and II.
The triplet state lifetimes in compounds I and II are
approximately 80 ns in noncoordinating benzene and are longer
by a factor of 2-3 relative to the triplet state lifetime of CuTPP
under similar experimental conditions.27c,e,f These data alone
do not allow us to attribute the increase in lifetime to a specific
substituent effect because McMillin and co-workers27h,i have
shown that the triplet state lifetimes of arene-ring-substituted
CuTPPs do not vary systematically with the electronic properties
of the phenyl substituent. In order to interpret the dynamics of
the 2T-4T manifold in our donor-acceptor complexes, we
examined the influence of solvent on the transient absorption
spectra. As noted in Table 1, the triplet lifetimes for I and II
in coordinating THF are shorter than those that are observed in
benzene; it is important to recognize, however, that the triplet
lifetimes for I and II in THF are longer by a factor of 150-
230 with respect to that observed for CuTPP in this solvent.26a
In any event, while (porphinato)copper(II) species having an
a1u HOMO, such as (octaethylporphinato)copper(II) (CuOEP),
typically have longer triplet lifetimes with respect to analogous
species having an a2u HOMO, such as CuTPP, it is interesting
to note the lifetimes of CuOEP and CuTPP differ only by a
factor of 4 in THF.28a While the origin of the enhanced lifetimes
for I and II is unclear at present,31 it is likely either (i) the
energy gap between I and II’s triplet manifold and the (d, d)
and CT quenching states that have been identified for CuTPP
is larger than in this benchmark chromophore or (ii) different
quenching states determine I and II’s lifetimes in THF. A
possible origin for this behavior may derive from the π-con-
jugating meso-arylethynyl moieties, which likely affect the
Lewis acidity of Cu2+, potentially impacting the axial ligand
affinity of these species in their ground and/or excited states.
This hypothesis is consistent with the fact that I and II’s trip-
multiplet lifetimes in benzene are long with respect to those
observed in THF, suggesting the importance of a quenching
state whose effect is more pronounced upon coordination of
solvent.
2
8, inset). Namely, upon formation of the T-4T manifold (i)
ν4 does not shift (relative to the FT Raman spectrum of the
ground state) and (ii) the nitro band upshifts by 6 cm-1 (relative
to the FT Raman spectrum of the ground state). The behavior
of ν4 is consistent with a (π, π*) state that mixes with a CT
state (Vide infra).25 The upshift in the nitro band may result
from the rotation of the nitrophenyl group out of the plane of
the porphyrin upon formation of the triplet excited state. In
any event, these data suggest that II’s triplet state possesses a
similar degree of CT character relative to the ground state.
Clearly, the resonance Raman data (Table 2, Figure 6) indicate
that the initially prepared sing-doublet excited state features a
substantial redistribution of electron density with respect to the
ground state (Figure 5). Moreover, because the nitro group,
for example, is much more strongly coupled in the initially
prepared excited state, it suggests that this state exhibits
substantial CT character. Because these data argue that
electronic excitation into the B band at 441.6 nm should produce
a state with different characteristics than were observed in both
the time-resolved Raman and transient absorption experiments,
we sought to probe states with lifetimes shorter than that of the
2T-4T manifold.
Dynamics of Excited States Observed by Transient
Resonance Raman Spectroscopy. It has been shown that
(porphinato)copper(II) excited states with short lifetimes, such
as CT or (d, d) states, may be probed by one-laser, transient
resonance Raman spectroscopy.13,25,32 Our transient Raman
strategy for compounds I and II is similar to that used by de
Paula et al.13 and Jeoung et al.25 to study quenching states in
CuTPP, which utilized a laser wavelength within the ground
state Soret band and relatively far away from the absorption
spectrum of the 2T-4T manifold. Given the long triplet lifetime
and the fact that it absorbs weakly at 436 nm, due to the
Qualitative similarities in the transient absorption properties
of our donor-acceptor complexes and of CuTPP allow us to
use a single model to explain their excited state dynamics. Either
(32) Additional examples include the following: (a) Findsen, E. W.;
Shelnutt, J. A.; Ondrias, M. R. J. Phys. Chem. 1988, 92, 307-314. (b)
Crawford, B. A.; Ondrias, M. R. J. Phys. Chem. 1989, 93, 5055-5061. (c)
Strahan, G. D.; Lu, D.; Tsuboi, M.; Nakamoto, K. J. Phys. Chem. 1992,
96, 6450-6457. (d) Jeoung, S. C.; Kim, D.; Cho, D. W., Yoon, M.; Ahn,
K.-H. J. Phys. Chem. 1996, 100, 8867-8874.
(31) Enhanced triplet lifetimes of similar magnitude have been observed
in other coordinating solvents. LeCours, S. M.; Phillips, C. M.; Therien,
M. J. Unpublished results.

NoVel Electron Donor-Acceptor Complexes
J. Am. Chem. Soc., Vol. 119, No. 51, 1997 12587
Table 3. Raman Ground and Excited State Band Assignments for Compound II
vibrational frequencies (cm-1
)
ground state
excited state
Fourier
transform Raman
resonance
Raman (441.6 nm)
resonance
Raman (436 nm)
transient
resonance Raman (436 nm)
mode descriptiona
ν1
NO2
1238
1337
1235
1338
f
1227
1324
1335
ν4 (CRN, CRCâ)
1362
b
1560
1359
1538
1568
1355
f
1566
f
f
1348
1542
1571
ν
20 (CRCmeso
)
ν2 (CRCmeso, CâCâ)
(CPhBCPhB
ethyne
)
(1605, 1596)c
(2181, 2196)d
(1611, 1599)c
2194e
(1607, 1598)c
2192e
a Tentative assignments. Mode descriptors are taken from ref 19 unless otherwise indicated. The NO2 and ethyne modes were assigned by
reference to previously studied benchmark compounds (see ref 30). b Weak or not observed. c Two frequencies observed. The first value in parentheses
corresponds to C-C stretching of the (dimethylamino)phenyl moiety and the second to the nitrophenyl moiety. d Two frequencies observed. The
first value in parentheses corresponds to the ethyne directly linked to the (dimethylamino)phenyl moiety and the second to the nitrophenyl moiety.
e Only the frequency of the ethyne directly linked to the nitrophenyl moiety is observed. f Assignment not made due to spectral congestion and/or
large spectral shift.
substantial transient bleach observed at this wavelength (Figures
3 and 4), any states that absorb strongly at 436 nm and have
lifetimes as short as 100 ps can be probed selectively in the
transient Raman experiment.
amended. Jeoung et al.25 have shown that the time-resolved
Raman spectra of the 2T-4T states of CuOEP and CuTPP show
very minor, or no, shifts in ν4. This was seen as a result of
mixing between (π, π*) and CT states in the formation of the
trip-multiplet states.
Results from our electronic structure calculations provide a
basis for interpretation of both time-resolved and transient
Raman spectra. Given the similarities between the frontier
orbitals of compounds I and II and those of D4h porphyrins,
we may use the arguments put forth by de Paula et al.12,13 to
predict shifts in the vibrational spectrum of the porphyrin moiety
upon formation of (π, π*) states. Considering compound II,
for example, we observe (Figure 10) that in the HOMO - 1
(a2u in D4h symmetry) the CRN interaction is slightly antibonding
in character; in the LUMO and LUMO + 1, however, the CRN
interactions are clearly net antibonding. Hence, promotion from
the HOMO - 1 to either the LUMO or LUMO + 2 results in
weakening (downward frequency shifts) of the CRN bonds.
Because the CâCâ interaction is essentially nonbonding in
character in the LUMO and LUMO + 1 while antibonding in
the HOMO - 2, an excited state configuration involving these
orbitals (analogous to the (a1ueg) configuration in D4h symmetry)
leads to strengthening of the CâCâ bonds. Similar arguments
lead to the conclusion that the CRCâ and CRN bonds are
correspondingly weakened. Figure 10 also suggests that modes
composed of NO2 stretches should shift to lower frequencies
upon promotion of electrons from the HOMO - 1 or HOMO
- 2 to either the empty LUMO or LUMO + 2; such excited
configurations correspond to porphyrin-to-nitro CT states.
Because such CT states are predicted to exist for both
compounds I and II, we must also consider the effect of partial
oxidation and reduction of the porphyrin ring on their transient
Raman spectra. Czernuszewicz et al.20 have determined that
cation radicals of copper(II) porphyrins with a2u character, such
as CuTPP+, show downshifts in ν4 and ν2 relative to the neutral
species. On the other hand, cation radicals with a1u character,
such as CuOEP+, show downshifts in ν4 and upshifts in ν2
relative to the neutral species. Data for anion radicals of copper-
(II) porphyrins are not available, but Bocian and co-workers19,33
have studied ZnOEP- and ZnTPP-. They find that reduction
of ZnTPP results in a small (4 cm-1) downshift in ν4 and in a
splitting of ν2 with one component upshifted by only 1 cm-1
and the other downshifted by 14 cm-1. Reduction of ZnOEP
We now apply the principles described above to the transient
Raman data obtained for electronically asymmetric compound
II (Figure 8). The transient resonance Raman spectrum of
compound II indicates that (i) ν4 and the nitro band downshift
14 and 13 cm-1, respectively, relative to the FT Raman spectrum
of the ground state and (ii) ν2 upshifts 11 cm-1 (relative to the
FT Raman spectrum of the ground state). The behavior of ν2,
ν4, and nitro bands is thus consistent with an electronically
excited state that features a porphyrin core with cation radical
character. Moreover, it suggests that this excited state species
is formed by promotion of an electron from an a1u-like orbital
(HOMO - 2, a′′ symmetry, Figure 10).
The vibrational properties of the state probed via the transient
Raman experiment at 436 nm are distinctly different from those
observed for the 2T-4T state. Our spectral assignments,
summarized in Table 3, are as follows.
(i) Our electronic structure calculations predict that compound
II should exhibit a slight decrease in bond order for both ethynyl
moieties in the excited state with respect to the ground state
(Figure 10). We do not observe this effect in the transient
Raman spectra;23 this may be due to power broadening effects
that may mask minor spectral shifts occurring in the ethynyl
region.
(ii) The bands at 1598 and 1607 cm-1 (Figure 8A) are
probably C-C stretching modes of the phenyl groups in the
(nitrophenyl)ethynyl and [(dimethylamino)phenyl]ethynyl moi-
eties, respectively. The fact that these bands do not shift in
frequency upon electronic excitation is consistent with our
ZINDO calculations (Figure 10), which show that the HOMO
- 2, HOMO - 1, and LUMO + 1 do not have significant
electron density in these phenyl groups. Likewise, while the
LUMO and LUMO + 2 do exhibit significant electron density
on the (nitrophenyl)ethynyl moieties, the arene ring-localized
orbitals are essentially nonbonding in character.
(iii) The band at 1571 cm-1 (Figure 8A), assigned as ν2
(CâCâ) of the excited state, upshifts by 11 cm-1 relative to the
ground state spectrum obtained at 436 nm (Figure 5B). It is
important to emphasize that in contrast, the ν2 mode of CuTPP
downshifts by 20 cm-1 in the LMCT excited state and by 3
cm-1 in the MLCT or (d, d) excited state.13
does not shift ν2 and downshifts ν4 by 7 cm-1
.
In systems where both (π, π*) and CT states are close in
energy, such as CuOEP and CuTPP, the predictions for
frequency shifts upon formation of (π, π*) states must be
(33) Perng, J.-H.; Bocian, D. F. J. Phys. Chem. 1992, 96, 4804-4811.
(iv) The broad band at 1542 cm-1 is consistent with the ν20

12588 J. Am. Chem. Soc., Vol. 119, No. 51, 1997
LeCours et al.
Table 4. Resonance Raman Vibrational Band Assignments for
Compound I
vibrational frequencies (cm-1
)
excited state
transient
resonance resonance resonance
ground state
Fourier
mode
transform
Raman (441.6 nm) (436 nm)
Raman
Raman
Raman
(436 nm)
descriptiona
ν1
1234
1342
1364
b
1241
1343
1365
1537
b
1563
1597
2195
1247
1336
c
1228
1323
1347
1537
1550
1568
1596
2195
NO2
ν4 (CRN, CRCâ)
Figure 11. Summary of the proposed excited state dynamics for
compounds I and II.
ν20 (CRCmeso
ν20 (CRCmeso
)
)
c
c
b
ν2 (CâCâ,CRCmeso,)
1560
1595
2195
1566
1598
d
(CPhBCPhB
)
The main difference in the vibrational properties of the excited
state probed in the transient Raman experiment between
compounds I and II lies in the behavior of the nitro band. While
the nitro band for compound II is located at 1337 cm-1 in the
FT Raman spectra, the analogous nitro stretching frequency for
compound I is located at 1342 cm-1. Compound II’s 1337
cm-1 band lies 11 cm-1 to lower frequency with respect to the
analogous band in (4-nitrophenyl)acetylene. Because the loca-
tion for the nitro band in the excited state is similar for both
compounds (approximately 1323 cm-1; see Figures 7A and 8A),
we suggest that more porphyrin-to-nitro CT character is mixed
into the ground state for compound II, consistent with the
expectation that the electron-releasing [(dimethylamino)phenyl]-
ethynyl moiety should make II’s porphyrin macrocycle more
electron rich than I’s (Figures 9 and 10). Furthermore, while
both compounds I and II show N-O antibonding character (data
not shown) in the HOMO - 1, Mulliken population analysis
of this orbital shows that II has more N- and O-atom-centered
electron density in this orbital with respect to I, congruent with
the FT Raman data. Finally, since the shifts of all bands in the
excited state with respect to the ground state are similar for
both compounds, this signifies that the extent of charge
delocalization from the porphyrin to the nitro substituent must
be similar for both I and II.
ethyne
a Tentative assignments. Mode descriptors are taken from ref 19.
The NO2 and ethyne modes were assigned by reference to previously
studied benchmark compounds (see ref 30). b Weak or not observed.
c Not readily assignable due to spectral congestion or low intensity.
d Experiment not performed.
(CRCmeso) mode observed in CuTPP.19 There may be more than
one peak buried within this envelope.
(v) Two low-intensity bands appear at 1290 and 1450 cm-1
which are probably due to ν27 and ν28 in the excited state. Both
of these modes were also observed in the transient resonance
Raman spectrum of CuTPP in THF.13
(vi) The band at 1348 cm-1 is assigned as ν4 (CRN) of the
excited state and is downshifted by 7-14 cm-1 relative to the
ground state spectrum, consistent with the expectation based
on our electronic structure calculation. In CuTPP, the ν4 mode
downshifts by 24 cm-1 in the LMCT state and by 4 cm-1 in
the MLCT or (d, d) state.13
(vii) The nitro band downshifts by 13 cm-1 upon formation
of the excited state at 436 nm.
(viii) The band at 1227 cm-1 is probably ν1 of the excited
state and is downshifted by 11 cm-1 relative to the ground state
spectrum.34 The ν1 mode of CuTPP also downshifts upon
electronic excitation, albeit by only 3-4 cm-1 13
.
Conclusions
We conclude that high-power pulsed excitation at 436 nm
probes only one excited electronic state of compound II, whose
vibrational spectrum differs from the spectrum of the ground
state by an upshift of ν2 and downshifts in ν4 and nitro bands.
Furthermore, the Raman spectroscopic signature of this state
does not resemble that of the triplet state, which was probed in
the time-resolved Raman experiment. The behavior of ν2 and
ν4 is consistent with removal of electron density from a
porphyrin-based molecular orbital with a′′ symmetry (a1u in the
D4h point group). The downshift in the nitro mode is consistent
with occupancy of either empty a′ orbital. We also suggest
that the filled a′′ orbital is higher in energy in this state than
the filled a′ orbital, in contrast to the results of our ZINDO
calculations carried out for the ground state electronic structure
(Figure 10).
A model for the excited state dynamics of compounds I and
II is presented in Figure 11, which is based upon the results
obtained from the FT, resonance, time-resolved resonance, and
transient resonance Raman spectroscopic experiments, as well
as the data obtained from our transient absorption studies. The
resonance and FT Raman data suggest that excitation into the
2
Soret band produces an excited state, denoted as S1 (1CT) in
Figure 11, that has enhanced charge-transfer character with
respect to the ground state (2S0). The transient absorption
experiments and time-resolved Raman data evince that the triplet
manifolds of compounds I and II bear many features in common
2
with the T-4T states of CuTPP; most importantly, neither
experiment suggests that the long-lived triplet states of I and
II have significant CT interactions involving the porphyrin and
the pendant arylethynyl groups. Because the initially prepared
excited state (2S1) has augmented porphyrin-to-(nitrophenyl)-
Nearly all of the points raised above for compound II apply
to the transient resonance Raman spectrum of compound I
(Figures 7 and 8, Table 4). Namely, the upshift in ν2 and
downshifts in both ν4 and the nitro mode are consistent with
formation of a porphyrin-to-nitro CT state where electron density
is removed from the au orbital (a1u in D4h symmetry).
2
ethynyl CT character with respect to that observed for S0, we
argue that the state we probe in the transient resonance Raman
3
experiment is best formulated as a CT state, since it bears so
many vibrational signatures in common with that which would
2
be expected for S1, on the basis of our electronic structure
(34) We considered two alternative assignments for the 1227 cm-1
band: ν22 (CRN) and ν31 (CâH in-plane deformation). While the ν22 mode
would be expected to downshift, the ν31 mode would be expected to be
relatively insensitive to photoexcitation. On the basis of this prediction, we
suggest that the 1228 cm-1 band arises from either ν1 or ν22. However,
whether the band is due to ν1 or ν22 does not influence significantly our
analysis of the transient Raman spectrum of compound I.
calculations. Congruent with the short (τ < 350 fs) lifetime of
2
2S1 for CuTPP, we propose that S1 for compounds I and II
rapidly decays, but in our case, to a 3CT state. The 3CT state,
though probed via transient Raman methods, must be relatively
short-lived as well (τ < 5 ns) because no evidence for this state

NoVel Electron Donor-Acceptor Complexes
J. Am. Chem. Soc., Vol. 119, No. 51, 1997 12589
is present in the transient absorption and time-resolved Raman
experiments.
tion (Grant CHE-9530707). M.J.T. is indebted to the National
Institutes of Health (Grant GM 48130-01A1.4) and the Office
of Naval Research (Grant 96-1-0725) for their research support,
and to the Searle Scholars Program (Chicago Community Trust),
the Arnold and Mabel Beckman Foundation, E. I. du Pont de
Nemours, and the National Science Foundation for Young
Investigator Awards, as well as to the Alfred P. Sloan and
Camille and Henry Dreyfus Foundations for research fellow-
ships. Transient triplet-triplet absorption spectra were obtained
at the NIH-supported Regional Laser and Biotechnology
Laboratory (RLBL) located at the University of Pennsylvania.
The authors thank Dr. Ranjit Kumble, Professor Valerie A.
Walters, and Mr. John Vrettos for helpful discussions.
Studies of [5,15-bis[(4′-nitrophenyl)ethynyl]-10,20-diphen-
ylporphinato]copper(II) and [5-[[4′-(dimethylamino)phenyl]eth-
ynyl]-15-[(4′′-nitrophenyl)ethynyl]-10,20-diphenylporphinato]-
copper(II) suggest important ways to tune the photophysical
properties of copper(II) porphyrins. At least in compound II,
production of the electronically excited CT state leads to a
substantial change in the dipole moment of the molecule. We
conclude that compounds I and II are important members of a
new class of chromophores that can undergo long-range, fast,
charge migration upon electronic excitation. Given the polariz-
able nature of the porphyrinic core, the large oscillator strength
of porphyrin-centered transitions, and the fact that these charge-
transfer states may be formed very efficiently with pulsed laser
excitation, molecules based on this structural motif represent a
new direction in the design of chromophores relevant to
optoelectronic devices and materials.16
Supporting Information Available: Transient resonance
Raman spectra of the ethynyl region of compounds I and II in
THF obtained with pulsed excitation at 436 nm (6 pages). See
any current masthead page for ordering and Internet access
instructions.
Acknowledgment. J.d.P. gratefully acknowledges funding
from the Kresge Foundation and the National Science Founda-
JA964436J