Metal-to-ligand charge-transfer excited-states in binuclear copper(I) complexes. Tuning MLCT excited-states through structural modification of bridging ligands. A resonance Raman study

Article abstract of DOI:10.1039/a909121a

The resonance Raman spectra of the ground- and lowest excited-states for a series of binuclear copper(I) complexes with bridging ligands based on 2,3-di-(2-pyridyl)quinoxaline have been measured. Analyses of the ground-state resonance Raman spectra show strong enhancement of an inter-ring stretching mode and quinoxaline-based modes; consistent with the Frank-Condon state having bonding changes about the copper(I) centre and on the quinoxaline ring system. The electronic absorption data also suggest the initially formed MLCT-state is localised towards the quinoxaline ring system. The excited-state resonance Raman spectroscopy shows similar spectral features for complexes which have substituent changes at the pyridyl rings of the bridging ligand. They show spectra associated with the radical anion of the bridging ligand. This is consistent with the formation of an excited-state which is metal-to-ligand charge-transfer in nature and where the radical anion is localised on the quinoxaline ring system of the bridging ligand. The Royal Society of Chemistry 2000.

Full text of DOI:10.1039/a909121a

View Article Online / Journal Homepage / Table of Contents for this issue
Metal-to-ligand charge-transfer excited-states in binuclear
copper(I) complexes. Tuning MLCT excited-states through
structural modification of bridging ligands. A resonance
Raman study
Mark R. Waterland,† Amar Flood and Keith C. Gordon*
Department of Chemistry, University of Otago, PO Box 56, Dunedin, New Zealand.
E-mail: kgordon@alkali.otago.ac.nz
Received 12th July 1999, Accepted 18th November 1999
The resonance Raman spectra of the ground- and lowest excited-states for a series of binuclear copper() complexes
with bridging ligands based on 2,3-di-(2-pyridyl)quinoxaline have been measured. Analyses of the ground-state
resonance Raman spectra show strong enhancement of an inter-ring stretching mode and quinoxaline-based modes;
consistent with the Frank–Condon state having bonding changes about the copper() centre and on the quinoxaline
ring system. The electronic absorption data also suggest the initially formed MLCT-state is localised towards the
quinoxaline ring system. The excited-state resonance Raman spectroscopy shows similar spectral features for
complexes which have substituent changes at the pyridyl rings of the bridging ligand. They show spectra associated
with the radical anion of the bridging ligand. This is consistent with the formation of an excited-state which is
metal-to-ligand charge-transfer in nature and where the radical anion is localised on the quinoxaline ring system
of the bridging ligand.
Introduction
The metal-to-ligand charge-transfer (MLCT) states of transi-
tion metal complexes with polypyridyl ligands show consid-
erable promise as charge-transfer sensitisers in solar energy
conversion schemes.1 Incorporating a ligand with two chelation
sites into the coordination sphere allows systematic construc-
tion of large supramolecular assemblies, capable of acting as
antennae in these energy conversion schemes.2 A rational study
of the properties of these large multi-nuclear assemblies begins
with an investigation of the smaller units such as mono- and
binuclear complexes. This study investigates the lowest MLCT-
state of some binuclear copper() systems that incorporate the
bridging ligand 2,3-di(2-pyridyl)quinoxaline (dpq) and some of
its simple derivatives. The complexes have interesting prop-
erties; firstly, the copper() centre provides a unique mechanism
for controlling the photophysics in the excited-state. In the
MLCT-state the copper() centre is formally oxidised to the
copper() oxidation state. In coordinating solvents (e.g. meth-
anol) this MLCT-state is rapidly quenched by exciplex form-
ation with the copper() centre.3 Methyl groups can be used
to block the approach of the solvent quencher molecule and
thus extend the MLCT excited-state lifetime. Hence, a series of
ligands with a methyl group at the 6Ј-position of the pyridyl
ring have been utilised in this study (Fig. 1). Secondly, the
addition of methyl and other groups alters the electronic energy
of the π-accepting properties of the bridging ligand. In these
binuclear systems the energy of the π-acceptor MO is an
important factor in controlling the extent of metal–metal
communication.4 The addition of electron-withdrawing and
-donating moieties provides a means for fine-tuning this degree
of communication.
Fig. 1 Ligands and numbering scheme: RЉ = RЈ = H, 1; RЉ = H,
RЈ = benzo, 2; RЉ = H, RЈ = CH3 3; RЉ = H, RЈ = Cl, 4; RЉ = CH3,
RЈ = H, 5; RЉ = CH3, RЈ = benzo, 6; RЉ = CH3, RЈ = CH3, 7; RЉ = CH3,
RЈ = Cl, 8.
systems. Secondly, to characterise the excited-state created after
relaxation from the FC-state.
This paper presents the resonance Raman spectra of a series
of binuclear copper() systems containing the dpq ligand. The
effects of substitutions at the pyridyl and quinoxaline rings of
the bridging ligands will be discussed. Resonance Raman
spectroscopy provides a detailed probe of the nuclear geometry
changes that occur in the FC region immediately following
photoexcitation. Broad unstructured absorption bands, that are
typical of MLCT transitions, are indicative of systems that
undergo so-called ‘short-time’ dynamics5 on the excited-state
surface. When only a single electronic excited-state contributes
to the scattering tensor, the resonance Raman intensities may
be simply related to excited-state geometry displacements in the
‘short-time’ regime.6 Modes that are able to mimic the nuclear
geometry changes that occur upon electronic excitation will be
selectively enhanced when the frequency of the laser radiation
field approaches resonance with the electronic transition.7,8 The
relative intensities of the observed bands may be related to the
normal coordinate displacement (∆) on going from the ground-
to excited-state.
The aims of this study were two-fold. Firstly, to establish the
nature of the Frank–Condon (FC) state into which the excited
electron is transferred upon MLCT excitation and how this
is altered by substituents at the pyridyl and quinoxaline ring
Resonance Raman scattering, in the condensed phase, is
physically a short-time effect (≈10Ϫ15 s) and as such provides
information for only the FC region on the excited-state surface.
Pulsed excitation may be used to create a significant excited-
† Present address: Department of Chemistry, Kansas State University,
Manhattan, Kansas, USA.
J. Chem. Soc., Dalton Trans., 2000, 121–127
This journal is © The Royal Society of Chemistry 2000
121

View Article Online
state population that can generate Raman scattering and thus
serve as a probe for the populated excited-state. For the pulse
duration (approx. 10 ns) that we employ, the majority of the
scattering from the excited-state comes from the thermally-
equilibrated excited (THEXI)9 state. In this paper we also
present time-resolved resonance Raman (TR3) spectra and
compare structural changes evident in the FC region with the
geometry in the THEXI-state.
[(PPh3)2Cu(dmpb)Cu(PPh3)2]ؒ[BF4]2
(Found C, 63.34; H, 4.65; N, 3.05; F, 7.53. Calculated C, 63.66;
H, 4.55; N, 3.05; F, 8.25%): yield = 85%.
(6ؒ[BF4]2ؒ1.5CH2Cl2).
[(PPh3)2Cu(dmpqMe2)Cu(PPh3)2]ؒ[BF4]2 (7ؒ[BF4]2ؒ0.5CH2-
Cl2). (Found C, 65.84; H, 4.86; N, 3.21; F, 8.41. Calculated C,
65.50; H, 4.71; N, 3.23; F, 8.76%): yield = 81%.
[(PPh3)2Cu(dmpqCl2)Cu(PPh3)2]ؒ[BF4]2
(8ؒ[BF4]2ؒ0.5CH2-
Cl2). (Found C, 62.45; H, 4.39; N, 3.21; F, 7.90. Calculated C,
62.64; H, 4.26; N, 3.16; F, 8.56%): yield = 79%.
Experimental
Synthesis
The ligands were synthesised by literature methods.10 In a
typical procedure, one equivalent of the appropriately substi-
tuted 1,2-diaminobenzene (2.1 mmol) was added to a solution
of 500 mg (2.1 mmol) of 2,2Ј-(6-methylpyridyl) in ethanol (100
mL) and refluxed for 12 h in the absence of light. The volume
was reduced and sufficient water added to initialise precip-
itation. The solution was then heated until dissolution
occurred. The product that separated upon cooling was isolated
by vacuum filtration and washed several times with an ice-cold
water–ethanol mixture (50:50 v:v) and recrystallised from a
water–ethanol mixture of the same composition.
Binuclear copper() complexes of each ligand were prepared
from the copper() precursor, bis(acetonitrile)bis(triphenylphos-
phine)copper() tetrafluoroborate [Cu(PPh3)2(MeCN)2]ؒ[BF4]
and the appropriate ligand, following the method of Yam and
Lo.11 One equivalent of bridging ligand (50 mg, 0.17 mmol) was
added to a stirred, degassed solution of dichloromethane (20
mL) containing 2 mol equivalents (266 mg, 0.34 mmol) of
[Cu(PPh3)2(MeCN)2]ؒ[BF4] under an argon atmosphere. A deep
red colour formed immediately. The solution was stirred for a
further 10 min then evaporated to dryness. The complex was
dissolved into 3 mL of dichloromethane and crystallised via
diethyl ether diffusion. The purity of the complexes [(PPh3)2-
Cu(dpq)Cu(PPh3)2]ؒ[BF4]2, 1ؒ(BF4)2, [dpq = 2,3-di(2-pyridyl)-
quinoxaline], [(PPh3)2Cu(dpb)Cu(PPh3)2]ؒ[BF4]2, 2ؒ(BF4)2,
{dpb = 2,3-di(2-pyridyl)benzo[g]quinoxaline}, [(PPh3)2Cu(dpq-
Me2)Cu(PPh3)2]ؒ[BF4]2, 3ؒ(BF4)2, [dpqMe2 = 6,7-dimethyl-2,3-
di(2-pyridyl)quinoxaline] and [(PPh3)2Cu(dpqCl2)Cu(PPh3)2]ؒ
[BF4]2, 4ؒ(BF4)2, [dpqCl2 = 6,7-dichloro-2,3-di(2-pyridyl)quin-
oxaline] was determined by microanalysis.
Physical measurements and calculations
Either a continuous-wave Spectra-Physics Model 166 argon-ion
laser or an Optlectra helium–neon laser was used to generate
ground-state resonance Raman (GSRR) scattering. Band-pass
filters removed the Arϩ plasma emission lines from the laser
output. Typically, the laser output was adjusted to give 25 mW
at the sample. The incident beam and the collection lens were
arranged in a 135Њ back-scattering geometry to reduce Raman
intensity reduction by self-absorption.12 An aperture-matched
lens was used to focus scattered light through a narrow band
line-rejection (notch) filter (Kaiser Optical Systems) and a
quartz wedge (Spex) and onto the entrance slit of a spectro-
graph (Spex 750M). The collected light was dispersed in the
horizontal plane by a 1800 grooves mmϪ1 holographic diffrac-
tion grating and detected by a liquid nitrogen-cooled 1152-
EUV CCD controlled by an ST-130 controller and CSMA
2.3b(v.2) software (Princeton Instruments).
The frequency-tripled output from a Nd:YAG (Continuum
Surelite I) pulsed laser (5–7 ns pulse width, operating at 10 Hz)
was used to generate stimulated Raman scattering13 for the
excited state resonance Raman (ESRR) experiments. The 355
nm output from the Nd:YAG laser was focused into a 10 cm
cell filled with spectrophotometric grade acetonitrile. A quartz
Pellin–Broca prism was used to separate the co-linear partially
depleted pump beam and four Stokes orders of stimulated
Raman scattering. The second Stokes output from the Raman
shifter was directed onto the sample in the same manner as the
continuous-wave Raman experiments. Typically, a 355 nm, 7–30
mJ pump pulse was used, generating 0.5–2.5 mJ of the second
Stokes-shifted line at 448.2 nm. The beam diameter at the
sample was 300 µm.
2,3-Di(6-methyl-2-pyridyl)quinoxaline (dmpq). δH (200 MHz,
solvent CDCl3, standard SiMe4) 8.20 (dd, 2H), 7.79 (dd, 2H),
7.67 (d, 4H), 7.08 (dd, 2H), 2.28 (s, 6H) (Found C, 76.78;
H, 5.22; N, 17.80. Calculated C, 76.90; H, 5.16; N, 17.94%):
yield = 34%.
Wavelength calibration was performed with a combination
of solvent Raman lines14 and emission lines from the argon-ion
plasma. Calibrations were reproducible to within 1–2 cmϪ1
.
Spectra were obtained with a resolution of 5 cmϪ1. For reson-
ance Raman spectra generated with λex = 514.5, 488, 457.9 and
448.2 nm, no attempt was made to correct the Raman inten-
sities for self-absorption or for the spectrometer response.
Resonance Raman spectra generated with λex = 632.8 nm were
corrected by comparing solvent band intensities with those
from the literature.14
2,3-Di(6-methyl-2-pyridyl)benzo[g]quinoxaline (dmpb). δH
(200 MHz, solvent CDCl3, standard SiMe4) 8.79 (s, 2H), 8.13
(dd, 2H), 7.82 (d, 2H), 7.71 (t, 2H), 7.59 (dd, 2H), 7.10 (d, 2H),
2.27 (s, 6H) (Found C, 79.36; H, 5.23; N, 15.36. Calculated C,
79.53; H, 5.01; N, 15.46%): yield = 40%.
Spectrophotometric grade dichloromethane (Aldrich) was
used as received. Freshly prepared samples were used through-
out. Sample concentrations were typically of the order of 5–10
mM. When using pulsed excitation, samples were degassed by
bubbling with argon for 10 min prior to irradiation. Sample
integrity following irradiation with the pulsed laser was checked
by recording UV/Vis and GSRR spectra.
6,7-Dimethyl-2,3-di(6-methyl-2-pyridyl)quinoxaline (dmpq-
Me2). δH (200 MHz, solvent CDCl3, standard SiMe4) 7.96 (s,
2H), 7.65 (t, 4H), 7.08 (dd, 2H), 2.53 (s, 6H), 2.29 (s, 6H)
(Found C, 77.50; H, 5.75; N, 16.55. Calculated C, 77.62; H,
5.92; N, 16.46%): yield = 29%.
PM3 calculations were carried out using the PC Spartan Plus
6,7-Dichloro-2,3-di(6-methyl-2-pyridyl)quinoxaline
(dmpq-
(version 1.5) software package.15
Cl2). δH (200 MHz, solvent CDCl3, standard SiMe4) 8.34 (s,
2H), 7.71 (t, 4H), 7.12 (dd, 2H), 2.27 (s, 6H) (Found C, 62.92;
H, 3.47; N, 14.68. Calculated C, 63.01; H, 3.70; N, 14.70%):
yield = 18%.
Results and discussion
Electronic absorption data
[(PPh3)2Cu(dmpq)Cu(PPh3)2]ؒ[BF4]2
(Found C, 65.60; H, 4.66; N, 3.39; F, 8.40. Calculated C, 65.20;
H, 4.55; N, 3.28; F, 8.91%): yield = 70%.
(5ؒ[BF4]2ؒ0.5CH2Cl2).
Electronic absorption data for the complexes are presented in
Table 1. The complexes exhibit a visible absorption near 480 nm
(εmax ≅ 3000 MϪ1 cmϪ1) and a UV absorption near 370 nm
122
J. Chem. Soc., Dalton Trans., 2000, 121–127

View Article Online
Table 1 Electronic absorption data for [(PPh3)2Cu(BL)Cu(PPh3)2]2ϩ
systems
plex22 also assigned a band at 1320 cmϪ1 as having inter-ring
stretch character.
For 2 and 6 a number of modes are enhanced in the GSRR
spectra; these lie at 1626, 1590, 1562, 1515, 1468, 1399, 1320
and 1277 cmϪ1 for 2. The spectrum of phenazine23 shows bands
at 1556, 1479, 1403 and 1280 cmϪ1, thus most of the bands in 2
may be assigned as phenazine-based. The residual bands at
1515 and 1320 cmϪ1 are assigned as pyridine (ν4) and inter-ring
stretches (ν9). A number of bands (i.e. ν1) are also present at
high wavenumber, these are assigned as pyridyl in nature. The
spectrum of 6 is very similar to that of 2 [Fig. 2(b)], with an
extra band at 1395 cmϪ1. This extra band may be due to a
lowering of symmetry in the complex as a consequence of the
steric crowding about the copper() centres, brought about by
the methyl groups and the phenazine ring.
The electronic spectra for the complexes show broad struc-
tureless absorptions at about 480 and 370 nm. The pattern of
band enhancements observed from the resonance Raman
scattering, as a function of excitation wavelength, may provide
some insight into the types of transitions present.24 GSRR
data were obtained for each of the complexes with 457.9, 488
and 514.5 nm excitation wavelengths. These span the MLCT
transition present in each of the complexes. For the complexes
reported herein, the GSRR spectra show the presence of 2
transitions across the excitation wavelengths used. The changes
in band enhancements are exemplified by 4 (Fig. 3). This shows
the enhancement of bands at 1591, 1545, 1515, 1451, 1352,
1323 and 1286 cmϪ1 with 457.9 nm excitation but only the 1545,
1451, 1352, 1323 and 1286 cmϪ1 are strongly enhanced with
514.5 nm excitation. The wavenumbers for the observed spec-
tral bands are presented in Table 2, with an indication of their
relative enhancements at differing excitation wavelengths. The
general pattern of behaviour is that the higher wavenumber
modes that are pyridine-based (ν2, ν4) are more strongly
enhanced with blue excitation and the 1450–1470 (ν6) and 1320
(ν9) cmϪ1 bands become more enhanced with longer wavelength
excitation. These enhancements patterns are consistent with the
presence of two MLCT transitions which have differing
acceptor π* bridging ligand MOs; that at lower energy having
significant wavefunction amplitude at the quinoxaline ring,
hence enhancing the quinoxaline modes, and the second having
greater amplitude over the pyridyls and thereby enhancing the
pyridine modes.
For the complexes with 6-methyl substituted bridging
ligands, the pyridine-based modes are much less enhanced than
for the corresponding nonsubstituted complexes. This is con-
sistent with the methyl groups polarising the acceptor MO of
the bridging ligand away from the pyridine groups. This polar-
isation of MOs has been observed previously for substituted
bipyridine ligands in resonance Raman and TRIR studies.25,26
Depolarisation ratio measurements showed the scattering to
be consistent with A-term scattering (i.e. ρ = 1/3).27 A-term
scattering is also consistent with a resonance with a strongly
allowed electronic transition.
In the low wavenumber region, a band is prominent at 527
cmϪ1 for 4 (Fig. 3), with weaker features close to 630 cmϪ1. All 8
complexes show similar wavenumbers for bands in this region.
Such low wavenumber vibrations are typical of out-of-plane
ligand vibrations and may also contain significant metal-to-N
character.22 However, they are often not strongly enhanced
because of the so-called missing mode effect.6 Their large
enhancements indicate substantial displacements from the
ground-state in terms of ligand out-of-plane and metal–ligand
bond length changes. This is consistent with a change in
coordination number upon photoexcitation of a copper()
complex, as discussed by McMillin and others.3
Compound
λmax(CH2Cl2)/nm (ε × 10Ϫ4/MϪ1 cmϪ1
)
1ؒ(BF4)2
2ؒ(BF4)2
3ؒ(BF4)2
4ؒ(BF4)2
5ؒ(BF4)2
6ؒ(BF4)2
7ؒ(BF4)2
8ؒ(BF4)2
366 (0.866), 494 (0.365)
404 (0.754), 535 (0.257)
372 (1.08), 488 (0.380)
378 (1.12), 517 (0.232)
376 (0.638), 481 (0.292)
409 (0.691), 515 (0.362)
379 (0.729), 464 (0.304)
383 (0.940), 482 (0.223)
(εmax ≅ 10 000 MϪ1 cmϪ1). The visible absorption band shifts to
longer wavelength as the reduction potential of the ligand
decreases16 and is absent in the spectra of the corresponding
ligand and copper precursor. On this basis this band is assigned
to a MLCT transition. The possibility of a ligand-to-ligand
charge-transfer (LLCT) between PPh3 → BL occurring in the
visible region is ruled out. The related complex [(PPh3)2-
Ag(phen)]ϩ, which has an LLCT transition, shows no visible
absorptions.17 In each of the complexes, a band at ca. 370 nm
is observed. This is attributed to a π → π* transition of the
ligand; similar absorption bands are observed for the free
ligand species.
The relative energies of the absorption maxima for the
complexes are consistent with the reduction potentials for these
species as measured by Yam and Lo11 and ourselves.18
Complexes containing ligands that are more easily reduced
show red-shifted MLCT absorptions. Consequently the
energies of the π-accepting orbitals for the ligands follow the
order dpqMe2 > dpq > dpb > dpqCl2, with a similar pattern
observed for the ligands with RЉ = Me, although at slightly
higher energy.
Resonance Raman data
The ground state resonance Raman (GSRR) spectra show
about 8–10 bands within the 1100–1650 cmϪ1 region. The data
are presented in Fig. 2 and Table 2, with labels for the vibra-
tional modes. The pattern of band enhancements and
wavenumber shifts on going through the series of complexes
is complicated, but a number of empirical relationships are
evident. The ν2 band is assigned as pyridyl in nature, since it lies
at the appropriate wavenumber for the 8a mode of pyridine.19
Furthermore, a study of the related dimer [{Ru(bpy)2}2-
(µ-dpq)]4ϩ (bpy = 2,2Ј-bipyridine) by Wertz et al.20 revealed a
close correspondence between the GSRR spectrum of [{Ru-
(bpy)2}2(µ-dpq)]4ϩ and the surface-enhanced Raman scattering
(SERS) spectrum of quinoxaline. The only mode that did not
correlate with the SERS spectrum was at 1596 cmϪ1, which they
attributed to a pyridine-based mode. The fact that it is weakly
enhanced in the reported spectra suggests the acceptor MO
of the MLCT transition points away from the pyridine. The
wavenumber of the ν2 mode is increased when the pyridine ring
is methyl substituted, 1 to 5 and 2 to 6. The ν3 mode is assigned
as a delocalised vibration involving the motion of atoms on
both ring systems. The wavenumber of the ν3 mode is decreased
on going to the Cl quinoxaline ring substitution and it is also
shifted with methyl substitution at the pyridyl. The ν4 mode is
assigned as pyridine-based, as it is unshifted with substitution
at the quinoxaline rings, i.e. 1 through 4, but does shift to higher
wavenumber with methyl substitution at the pyridine ring. The
ν6 mode is assigned as a quinoxaline ring stretch, decreasing
in wavenumber as the electron density in the ring is reduced,
4 → 1 → 3. The ν9 mode is assigned as an inter-ring stretch
on the basis of the wavenumber value.18 An analysis of a series
of polypyridyl and phenyl compounds21 assigned bands in the
region 1200–1350 cmϪ1 as inter-ring stretches. A detailed
normal coordinate analysis of the related [Ru(bpy)3]2ϩ com-
Attempts were made to measure the steady-state and time-
resolved emission spectra for the binuclear copper systems.
All systems exhibited very weak emission, the systems with
RЉ = Me were slightly stronger emitters than the systems with
J. Chem. Soc., Dalton Trans., 2000, 121–127
123

View Article Online
Fig. 2 Ground state resonance Raman spectra (457.9 nm continuous-wave excitation, 15 mW at sample) of 10 mM solution of complexes in
CH2Cl2: (a)(i) 1; (a)(iii) 5; (b)(i) 2; (b)(iii) 6; (c)(i) 3; (c)(iii) 7; (d)(i) 4; (d)(iii) 8. Excited-state resonance Raman spectra (448.2 nm pulsed excitation,
2 mJ per pulse, 10 Hz) of 10 mM solutions of complexes in CH2Cl2: (a)(ii) 1*; (a)(iv) 5*; (b)(ii) 2*; (b)(iv) 6*; (c)(ii) 3*; (c)(iv) 7*; (d)(ii) 4*; (d)(iv) 8*.
The asterisk denotes the excited-state of the complexes.
Table
2
Raman bands (cmϪ1
) observed for [(PPh3)2Cu(BL)-
RЉ = H. Only an upper bound could be determined for the
excited-state lifetimes, τ < 50 ns, the lifetimes were too short to
be resolved reliably with the available apparatus.
The excited-state resonance Raman (ESRR) spectra of the
complexes show striking differences from the ground-state
spectra (Fig. 2). Studies were carried out at 448.2 nm. This
wavelength is resonant with the ground-state MLCT transition,
thus ensuring efficient pumping to the excited state. Coupled
Cu(PPh3)2]2ϩ systems (CH2Cl2 solution)
Mode
Label
1
2
3
4
5
6
7
8
1
2
3
4
5
6
7
8
9
1626a
1629a
1595a 1590a 1594 1591
1599 1598a
1602a
1556 1545b 1563 1566a 1557 1551
1515a 1515a 1520a 1519 1520a 1523a
with the fact that it is resonant with the absorption of
1555
1562
Ϫ 20,28
1514a 1515
1491b
dpq
,
it is therefore possible to establish an excited-state
᝽
1473
1486
1483
population and probe it using resonance Raman scattering
with single-colour pulsed excitation.29 Studies were carried out
using varying photon fluxes up to pulse energies where the
absorbed photon:molecule ratio was 3:130 and where no
residual ground-state scattering was observed.
1468
1468b 1469 1451b
1470 1462 1463
1399b
1403
1390 1406 1405 1387a
1364
1319
1290
1363 1352b 1351 1352
1329 1323b 1325b
1320
1341 1330
10
1277b 1291 1286b
1288
The spectrum of 1 shows excited-state bands at high pulse
energy; no residual ground-state is present in this spectrum, as
evidenced by the absence of ground-state bands at 1555, 1514,
a Preferentially enhanced with 457.9 nm excitation. b Preferentially
enhanced with 514.5 nm excitation.
124
J. Chem. Soc., Dalton Trans., 2000, 121–127

View Article Online
1340 cmϪ1, barely shifted from its position in the GSRR spec-
tra, and some features in the 1560–1590 cmϪ1 region (Table 3).
In general, the spectra appear similar for complexes which
differ only by their 6Ј-methyl substituents. For example, the
ESRR spectra of 1 and 5 show bands at similar wavenumber
(1586 and 1565 cmϪ1 for 1*, 1587 and 1563 cmϪ1 for 5*). This
suggests that the THEXI-state is similar in both cases and this
would be consistent with occupation of a ligand π* MO that is
strongly polarised to the quinoxaline ring system. The radical
anion spectrum of dpq has been measured spectroelectro-
chemically for 1 and related rhenium() complexes.28 The
spectrum of [Re(CO)3Cl(dpq)]Ϫ shows strong bands at 1586
and 1565 cmϪ1, coincident with those observed for 1*. However,
these differ significantly from the spectrum observed for
[Re(CO)3Cl(dpq)Re(CO)3Cl]Ϫ; this shows bands at 1604, 1562
and 1327 cmϪ1, with a stronger enhancement of the low
wavenumber mode.32 This finding was rationalised as being due
to polarisation of charge to the pyridyl rings in the binuclear
rhenium() complex, caused by the strong charge-stabilising
effect of the two Re() centres. The [Cu(PPh3)2]ϩ moiety is well
known to be less stabilising of charge and is unable to distort
the charge localisation in the manner observed for the binuclear
rhenium() complexes.
The radical anion feature observed at about 1580 cmϪ1
appears to be a quinoxaline-based mode from the fact that it
shifts with quinoxaline substitution, but not 6-methyl group
substitution at the pyridyl. A shift to higher wavenumber upon
reduction of a ligand is by no means unusual.33 We propose that
the 1570–1590 cmϪ1 band is the ν3 mode. This would corre-
spond to shifts of about 20–30 cmϪ1, such shifts have been
reported in the radical anion spectra of other polypyridyl
ligands. The very small shifts in the wavenumber for the inter-
ring stretch on going from the ground-state to the excited-state
are indicative of small bond changes at the C2–C2Љ or C3–C2Ј
linkages upon formation of the radical anion in the MLCT
excited-state. This is consistent with previous computational
and experimental studies.18
This interpretation of the excited-state Raman data is
consistent with electron spin resonance experiments on [{Ru-
(bpy)2}2(µ-dpq)]4ϩ that show the unpaired electron localised to
the quinoxaline ring of the dpq bridging ligand.19 That the
excited-state electron distribution is independent of the nature
of the substituent at the pyridyl ring may be the result of a
geometry change following photoexcitation.
Fig. 3 Ground-state resonance Raman spectra of 4 in CH2Cl2 solution
(10 mM concentration, 20 mW at sample): (i) λexc = 632.8 nm (60 mM
solution); (ii) λexc = 514.5 nm; (iii) λexc = 488 nm; (iv) λexc = 457.9 nm.
Table
3
Raman bands (cmϪ1
) observed for [(PPh3)2Cu(BL)Cu-
(PPh3)2]2ϩ systems with 448.2 nm pulsed excitation (CH2Cl2 solution)
1*
2*
3*
4*
5*
6*
7*
8*
1586
1565
1591
1561
1514
1588
1551
1588
1587
1563
1594
1563
1516
1588
1555
1586
1565
1529
1488
1467
1484
1462
1475
1399
1471
1328
1467
1404
1460
1337
1463
1383
1362
1325
1287
1351
1324
1318
1294
1325
1318
1278
1236
1328
The GSRR data provide an insight into the state(s) initially
populated by MLCT excitation at differing wavelengths. These
data suggest that with λexc = 457.9 nm the transition populates a
π* MO on the ligand that has significant pyridyl character, as
the λexc is tuned to the red a lower energy transition becomes
resonant which terminates across the quinoxaline ring. One
surprising result of this is that the GSRR spectra at long λexc
show strong enhancement of the inter-ring stretch mode (ν9),
yet the ESRR spectra shows a band unshifted from the corre-
sponding ground-state ν9 in all of the complexes studied. The
GSRR data suggest significant distortion of the inter-ring
bond, consistent with the population of a π* MO with ampli-
tude at the C2–C2Љ and C3–C2Ј linkages (Fig. 1), yet the ESRR
data suggest little change in bond order across those linkages.
There are a number of possible reasons for this observation:
(1) The relative enhancement of different modes, e.g. ν9 versus
ν3, may oscillate with excitation wavelength under certain con-
ditions. Primary among these is a low damping factor (Γ) for
the resonant transition. It is possible that the apparent strong
enhancement of ν9 is a consequence of the oscillation rather
than indicating a large ∆.34,35 This hypothesis may be tested by
measuring the spectrum under pre-resonance conditions,8 as
under such conditions Γ is large.35 The spectrum of 4 measured
with 632.8 nm excitation shows the enhancement of 3 modes
[Fig. 3(i)] in the 1000–1600 cmϪ1 region. The strongly enhanced
modes are the same as those enhanced at 514.5 nm excitation,
1364 and 1319 cmϪ1. The excited-state of 1, denoted 1*, has
bands at 1586, 1565, 1488 and 1467 cmϪ1, and a weak band at
1325 cmϪ1. This is in agreement with previous studies.19,31
The ESRR spectrum of 5 also possesses no ground-state, as
evidenced by the absence of the 1351 cmϪ1 band. The bands of
5* are very similar in wavenumber to those of 1* (see Table 3).
The ESRR spectra of 2 and 6 are similar to each other; no
residual ground-state scattering is observed as evidenced by
the absence of the 1625–1629 cmϪ1 band observed for both
complexes with 457.9 nm excitation. The spectrum of 6* shows
greater enhancement of the ν3 mode at 1563 cmϪ1 than does
that of 2* (1561 cmϪ1).
The ESRR spectra of 3 and 7 reveal excited-state features at
1588 and about 1553 cmϪ1 (1551 cmϪ1 for 3, and 1555 cmϪ1
for 7), these differ from those of 1, 2, 5 and 6 but are the same
for 3 and 7. 3* also shows bands at 1471 and 1328 cmϪ1 while 7*
has bands at 1460 and 1337 cmϪ1. The spectrum of 3* [Fig. 2(c)]
has no residual ground-state, as the ground-state band at 1291
cmϪ1 is absent from the spectrum. The 7* spectrum [Fig. 2(c)] is
also entirely excited-state, as the ground-state 1520 cmϪ1 band
is absent.
The ESRR spectra of 4 and 8 show a common excited-state
band at 1588 cmϪ1 for 4 and 1586 cmϪ1 for 8.
The general appearance of the ESRR spectra is similar for all
of the complexes; each shows an inter-ring band at about 1320–
J. Chem. Soc., Dalton Trans., 2000, 121–127
125

View Article Online
suggesting that the Γ for the transition is large. This means the
enhanced ν9 mode indicates a significant ∆ along that normal
mode.
(2) The FC-state probed in the GSRR spectra with 514.5 nm
excitation has a different electronic configuration from the
THEXI-state. Previous semi-empirical calculations on the
ligand dpq in the conformation in which it is found in 1 showed
the presence of two low-lying UMOs which differed signifi-
cantly in their wavefunction amplitude at the inter-ring link-
ages. If the MLCT transition resonant in the GSRR spectra
terminated on the π* ligand MO, resulting in a significant inter-
ring bonding change, then the ν9 mode would be enhanced, but
if the THEXI-state had the excited electron in the MO with a
non-bonded character across the inter-ring then there would be
very little perturbation in the wavenumber of the ν9 mode.
(3) The resonance enhancement from the THEXI-state results
in the enhancement of a vibration associated with one of the
inter-ring bonds that has been relatively unchanged by popu-
lation of the MLCT-state. This may occur because the MLCT
excited-state renders the molecule asymmetric from C2 sym-
metry because only one copper() site is oxidised. As a result
of this one of the sides of the ligand is flattened out. X-Ray
crystallographic data suggest the dihedral angle for a copper()
dpq complex is ca. 25Њ,36 whereas that for 1 is ca. 32Њ.20 In order
to understand these potential changes, a series of calculations
were performed.
Fig. 4 Plot of the variation in bond lengths (r) for the C2–C2Љ and
C3–C2Ј linkages (Fig. 1) versus dihedral angle (as defined by N1Љ–C2Љ–
C2–N1) for dpq and its dianion. ᭡ r(C2–C2Љ) dpq; ᮀ r(C3–C2Ј) dpq;
᭝ r(C2–C2Љ) dpq2Ϫ; ᭜ r(C3–C2Ј) dpq2Ϫ
.
For these calculations, three approximations to the reported
crystal structure of 1 were made. The Cu moieties were omitted
due to a limitation of confidence with the PM3 method beyond
the first row elements. In a study of ruthenium() complexes,
calculations of the electron localisation on the bipyridyl ligands
was modelled with substituted pyridine moieties. The validity
of this simplification was that the metal moiety interacts with
the ligand predominately through the σ-bonding system,
whereas the photophysics are dominated by the π-system of
the ligand.37 This argument should be more valid given that the
Cu() moiety has a smaller influence on the electronic nature of
the ligand than a Ru() moiety.38
Three structural forms of dpq were studied: 0/30-dpq, 15/30-
dpq and 30/30-dpq. The numbers correspond to constraints
placed on the dihedral angles between each of the pyridyl rings
and the quinoxaline ring system, as shown in Fig. 1 (defined as
the dihedral angle between N1Љ–C2Љ–C2–N1). The 30/30-dpq
form represents constraints associated with coordination of
two [Cu(PPh3)2]ϩ moieties. The 0/30-dpq is a limiting represen-
tative form of the constraints placed upon the dihedral angle by
coordination of a Cu() and a Cu() moiety, respectively. The
large change in dihedral angle was chosen to both bracket the
limiting forms of dpq and to aid in visualisation. The 15/30-dpq
form was selected as another data point in the quantum chem-
ical experiment.
To model the electronic form of dpq in the excited-state, the
dianion form of dpq was studied: dpq2Ϫ. Calculations on the
radical anion were avoided due to problems associated with
open shell configurations.39 Although the dianion represents the
addition of two electrons to the LUMO, this configuration
brackets the actual configuration. In regard to the present
study, the presence or absence of spectral shifts are the primary
concern and the calculations were not undertaken to provide a
detailed quantitative understanding of the spectral changes
observed.
Geometry optimisation and single point energy calculations
were carried out to determine the inter-ring bond distance and
the form of the molecular orbitals on 0/30-dpq, 15/30-dpq, 30/
30-dpq and the dianion forms, respectively, 0/30-dpq2Ϫ, 15/30-
dpq2Ϫ, 30/30-dpq2Ϫ. The 0/30-dpq and 15/30-dpq forms of the
neutral ligand were studied to assess any inter-ring bond
length changes independent from the addition of 2 electrons.
The inter-ring bond lengths (r) for the 3 conformations of dpq
and dpq2Ϫ are shown in Fig. 4.
There are 2 striking results from these calculations: (a) The
inter-ring bond length r(C2–C2Љ) is significantly shortened by
the flattening of the dihedral angle when electron density is
added to the LUMO. (b) The other side of the dpq has little
change in inter-ring bond length r(C3–C2Ј).
It should be noted that the shortening of r(C2–C2Љ) is not
observed for the neutral dpq with variation in dihedral angle
(Fig. 4).
For the GSRR experiment, enhancement of the inter-ring
results from the changes in the bond length associated with the
FC-state, i.e. dihedral angle 30Њ (Fig. 4). However the ESRR
variant will enhance modes associated with its resonance trans-
ition; that is for the ligand in its THEXI conformation. This has
r(C2–C2Љ) significantly shortened and r(C3–C2Ј) slightly
lengthened from the FC-state. The enhancement pattern will
result from geometry changes for the resonant radical anion
transition, which is π → π* in nature. If this transition results
in a change in r(C3–C2Ј) then the ν9 mode would be enhanced;
however the wavenumber of the ν9 mode would be similar to
that for the ground-state because the bond length is almost the
same as in dpq.
(4) The ESRR experiment may be probing a different mode
from that of the GSRR spectroscopy and this new mode is
coincident with the ν9 vibration. This seems unlikely to occur
for all of the complexes observed.
The most likely explanation for the observed spectral
changes, notably the enhancement of the ν9 mode in the GSRR
spectra and its appearance in the ESRR spectra at the same
wavenumber, are given by options (2) and (3). The observation
of strong enhancement of low wavenumber out-of-plane modes
(e.g. 527 cmϪ1, Fig. 3) indicates a strong distortion of the ligand
upon photoexcitation. This is supportive of option (3).
Conclusions
Analysis of the resonance Raman intensities in the ground-state
has provided a qualitative description of the geometry changes
following photoexcitation. The wavenumbers of the bands
observed in the excited-state have provided similar information
for the THEXI-state structure. The GSRR spectra reveal the
presence of two MLCT chromophores in the visible region. The
higher energy absorbing chromophore has an acceptor orbital
with significant polarisation toward the pyridyl rings, the lower
126
J. Chem. Soc., Dalton Trans., 2000, 121–127

View Article Online
wavenumbers for these standards are available at http://chemistry.
energy transition appears more polarised to the quinoxaline.
The THEXI-state appears to have an electronic configuration
in which the acceptor orbital is predominantly quinoxaline in
nature. Quantum chemical calculations suggest the inter-ring
bonds connecting the pyridyl to quinoxaline rings are shortened
to differing extents, depending on the flattening of the ligand.
This offers an explanation for the observation of an unshifted ν9
mode in the ESRR spectra.
ohio-state.edu/~rmccreer/shift.html. Methods for correcting Raman
spectra are given at http://chemistry.ohio-state.edu/~rmccreer/
intensty/intensty.html. The Raman spectrum of cyclohexane
obtained from our instrument was compared to literature values and
corrected where appropriate. This correction factor was applied to
the GSRR spectrum of 4 at 632.8 nm excitation.
15 PC SPARTAN Plus version 1.5, Wavefunction, Inc., 18401 Von
Karman Ave., Suite 370, Irvine, CA 92612, 1998.
16 M. R. Waterland, K. C. Gordon and A. K. Burrel, manuscript in
preparation.
17 D. J. Casadonte, Jr. and D. R. McMillin, J. Am. Chem. Soc., 1987,
109, 331.
18 K. C. Gordon, A. H. R. Al-Obaidi, P. M. Jayaweera, J. J. McGarvey,
J. F. Malone and S. E. J. Bell, J. Chem. Soc., Dalton Trans., 1996,
1591.
19 F. R. Dollish, W. G. Fateley and F. F. Bentley, Characteristic Raman
Frequencies of Organic Compounds, Wiley, New York, 1974.
20 J. B. Cooper, D. B. MacQueen, J. D. Petersen and D. W. Wertz,
Inorg. Chem., 1990, 29, 3701.
21 J. R. Schoonover, P. Chen, W. D. Bates, R. B. Dyer and T. J. Meyer,
Inorg. Chem., 1994, 33, 793; R. A. McNichol, J. J. McGarvey,
A. H. R. Al-Obaidi, S. E. J. Bell, P. M. Jayaweera and C. G. Coates,
J. Phys. Chem., 1995, 99, 12268.
22 D. P. Strommen, P. K. Mallick, G. D. Danzer, R. S. Lumpkin and
J. R. Kincaid, J. Phys. Chem., 1990, 94, 1357; P. K. Mallick, G. D.
Danzer, D. P. Strommen and J. R. Kincaid, J. Phys. Chem., 1988,
92, 5628.
23 T. J. Durnick and S. C. Wait, Jr. J. Mol. Spectrosc., 1972, 42, 211.
24 S. M. Scott, A. K. Burrell, P. A. Cocks and K. C. Gordon, J. Chem.
Soc., Dalton Trans., 1998, 3679; J. Sherborne, S. M. Scott and
K. C. Gordon, Inorg. Chim. Acta, 1997, 260, 199; J. Sherborne
and K. C. Gordon, Asian J. Spectrosc., 1998, 2, 137.
25 G. D. Danzer, J. A. Golus and J. R. Kincaid, J. Am. Chem. Soc.,
1993, 115, 8643.
26 K. M. Omberg, G. D. Smith, D. A. Kavaliunas, P. Chen, J. A.
Treadway, J. R. Schoonover, R. A. Palmer and T. J. Meyer, Inorg.
Chem., 1999, 38, 951; P. Chen, K. M. Omberg, D. A. Kavaliunas,
J. A. Treadway, R. A. Palmer and T. J. Meyer, Inorg. Chem., 1997,
36, 954.
Acknowledgements
Support from the New Zealand Lottery Commission and the
University of Otago Research Committee for the purchase of
the Raman spectrometer is gratefully acknowledged. M. R. W.
thanks the John Edmond postgraduate scholarship and the
Shirtcliffe fellowship for support for Ph.D. research and The
Royal Society of New Zealand for the award of a Beginning
Scientist Award. A. F. thanks the Alliance Group Postgraduate
Scholarship award.
References
1 M. K. Nazeeruddin, A. Kay, I. Rodicio, R. Humphrey-Baker,
E. Mueller, P. Liska, N. Vlachopoulos and M. Gratzel, J. Am. Chem.
Soc., 1993, 115, 6382; K. Kalyanasundaram and M. Gratzel, Coord.
Chem. Rev., 1998, 177, 347; J. E. Moser, P. Bonnote and M. Gratzel,
Coord. Chem. Rev., 1998, 177, 245; C. A. Bignozzi, J. R. Schoonover
and F. Scandola, Prog. Inorg. Chem., 1997, 44, 1.
2 V. Balzani, M. Gomez-Lopez and J. F. Stoddart, Acc. Chem. Res.,
1998, 31, 405; M. Venturi, S. Serroni, A. Juris, S. Campagna and
V. Balzani, Top. Curr. Chem., 1998, 197, 193; V. Balzani, S.
Campagna, G. Denti, A. Juris, S. Serroni and M. Venturi, Acc.
Chem. Res., 1998, 31, 26; R. Ziessel, M. Hissler, A. El-Ghayoury
and A. Harriman, Coord. Chem. Rev., 1998, 180, 1251; R. Ziessel
and A. Harriman, Coord. Chem. Rev., 1998, 171, 331; F. Scandola,
R. Argazzi, C. A. Bignozzi, C. Chiorboli, M. T. Indelli and M. A.
Rampi, in Supramolecular Chemistry, ed. V. Balzani and L. DeCola,
Kluwer Academic Publishers, Netherlands, 1992, p. 235; L. De Cola
and P. Belser, Coord. Chem. Rev., 1998, 177, 301; M. D. Ward, C. M.
White, F. Barigelletti, N. Armaroli, G. Calogero and L. Flamigni,
Coord. Chem. Rev., 1998, 171, 481.
3 M. K. Eggleston, D. R. McMillin, K. S. Koenig and A. J.
Pallenberg, Inorg. Chem., 1997, 36, 172; M. T. Miller, P. K. Gantzel
and T. B. Karpishin, J. Am. Chem. Soc., 1999, 121, 4292;
M. Ruthkosky, C. A. Kelly, M. C. Zaros and G. J. Meyer, J. Am.
Chem. Soc., 1997, 119, 12004.
4 J. H. Elias and R. S. Drago, Inorg. Chem., 1972, 11, 415; R. W
Callahan, G. M. Brown and T. J Meyer, Inorg. Chem., 1975, 14,
1443; C. Creutz, Prog. Inorg. Chem., 1983, 30, 1.
27 D. P. Strommen, J. Chem. Educ., 1992, 69, 803; R. J. H. Clark and
T. J. Dines, Angew. Chem., Int. Ed. Engl., 1986, 25, 131.
28 M. R. Waterland, T. J. Simpson, K. C. Gordon and A. K. Burrell,
J. Chem. Soc., Dalton Trans., 1998, 185.
29 P. G. Bradley, N. Kress, B. A. Hornberger, R. F. Dallinger and W. H.
Woodruff, J. Am. Chem. Soc., 1981, 103, 7441; P. A. Mabrouk and
M. S. Wrighton, Inorg. Chem., 1986, 25, 526; K. C. Gordon and J. J.
McGarvey, Chem. Phys. Lett., 1989, 162, 117.
30 At 2 mJ pulseϪ1 with a spot size of 300 µm, sample penetration of
2 mm and sample concentration of 1 × 10Ϫ2 mol LϪ1 the number
of molecules in the irradiated volume is less than 1015, the number of
photons per pulse is 4.5 × 1015.
31 K. C. Gordon and J. J. McGarvey, Chem. Phys. Lett., 1990, 173,
5 S.-Y. Lee and E. J. Heller, J. Chem. Phys., 1979, 71, 4777; D. J.
Tannor and E. J. Heller, J. Chem. Phys., 1982, 77, 202; E. J. Heller,
R. L. Sundberg and D. J. Tannor, J. Phys. Chem., 1982, 86, 1822.
6 A. B. Myers, Excited Electronic State Properties from Ground-State
Resonance Raman Intensities, in Laser Techniques in Chemistry, ed.
A. B. Myers and T. R. Rizzo, Wiley, New York, 1995.
443.
32 T. J. Simpson and K. C. Gordon, Inorg. Chem., 1995, 34, 6323.
33 G. Buntinx, O. Poizat and N. Leygue, J. Phys. Chem., 1995, 99, 2343.
34 J. I. Zink and K.-S. K. Shin, Adv. Photochem., 1991, 16, 119.
35 J. I. Zink and K.-S. K. Shin, Inorg. Chem., 1989, 28, 4358.
36 K. V. Goodwin, W. T. Pennington and J. D. Petersen, Acta Crys-
tallogr., Sect. C, 1990, 46, 898.
7 M. Tsuboi and A. Y. Hirakawa, Science, 1975, 188, 359.
8 S. K. Doorn and J. T. Hupp, J. Am. Chem. Soc., 1989, 111, 1142;
V. Petrov, J. T. Hupp, C. Mottley and L. C. Mann, J. Am. Chem.
Soc., 1994, 116, 2171.
37 N. H. Damrauer, B. T. Weldon and J. K. McCusker, J. Phys. Chem.
A, 1998, 102, 3382.
38 S. M. Scott, A. K. Burrell, P. A. Cocks and K. C. Gordon, J. Chem.
Soc., Dalton Trans., 1998, 3679; S. M. Scott, A. K. Burrell and K. C.
Gordon, J. Chem. Soc., Dalton Trans., 1998, 2873; S. M. Scott,
K. C. Gordon and A. K. Burrell, Inorg. Chem., 1996, 35, 2452.
39 Preliminary calculations on the radical anion reveal high values for
9 A. W. Adamson, J. Chem. Educ., 1983, 60, 797.
10 H. A. Goodwin and F. Lions, J. Am. Chem. Soc., 1959, 81, 6415.
11 V. W.-W. Yam and K. K.-W. Lo, J. Chem. Soc., Dalton Trans., 1995,
499.
12 D. B. Shriver and B. R. Dunn, Appl. Spectrosc., 1974, 28, 319.
13 R. P. Van Duyne and K. D. Parks, Chem. Phys. Lett., 1980, 76, 196;
C. A. Grant and J. L. Hardwick, J. Chem. Educ., 1997, 74, 318.
14 The ASTM subcommittee on Raman spectroscopy has adopted
eight materials as Raman shift standards (ASTM E 1840). The band
the spin expectation parameter; 〈S2〉 = 1.5201 (0/30-dpq Ϫ), 1.3936
᝽
(15/30-dpq Ϫ) and 0.8531 (30/30-dpq Ϫ). See ref. 38.
᝽
᝽
Paper a909121a
J. Chem. Soc., Dalton Trans., 2000, 121–127
127