Spectroscopic and electrochemical studies of a series of copper(I) and rhenium(I) complexes with substituted dipyrido[3,2-a:2′,3′-c]-phenazine ligands

Article abstract of DOI:10.1039/a706110b

Copper(I) and rhenium(I) complexes with the ligand dipyrido[3,2,-a:2′,3′-c]phenazine (dppz) and a number of substituted analogues have been synthesized. Their spectroscopic and electrochemical properties have been studied. It is found that the lowest-energy transition for the complexes is metal-to-ligand charge transfer (MLCT) in nature. This has a low ε value. The resonance Raman spectra for the complexes show groups of bands that shift with substitution at the ligand and groups that remain unchanged in wavenumber. Electrochemical reduction of the complexes resulted in the formation of the ligand radical anion species for all but one system. This was confirmed by UV/VIS spectroelectrochemistry. Using resonance Raman spectroelectrochemistry marker bands have been identified for the radical anion species. The excited states of the complexes were studied by excited-state electronic absorption and time-resolved resonance Raman techniques. The former spectra are ambiguous as to the nature of the lowest excited state; however, the latter spectra confirm that this state is ligand-centred for complexes of dppz and its 11-methyl derivative. Complexes with the 11-nitro derivative appear to form excited states that are MLCT in nature.

Full text of DOI:10.1039/a706110b

View Article Online / Journal Homepage / Table of Contents for this issue
Spectroscopic and electrochemical studies of a series of copper(I)
and rhenium(I) complexes with substituted dipyrido[3,2-a:2Ј,3Ј-c]-
phenazine ligands†
Mark R. Waterland,a Keith C. Gordon,*,a John J. McGarvey*,b and Pradeep M. Jayaweerab
a Chemistry Department, University of Otago, PO Box 56, Dunedin, New Zealand
b School of Chemistry, The Queen’s University of Belfast, David Keir Building, Belfast,
UK BT9 5AG
Copper() and rhenium() complexes with the ligand dipyrido[3,2,-a:2Ј,3Ј-c]phenazine (dppz) and a number of
substituted analogues have been synthesized. Their spectroscopic and electrochemical properties have been
studied. It is found that the lowest-energy transition for the complexes is metal-to-ligand charge transfer (MLCT)
in nature. This has a low ε value. The resonance Raman spectra for the complexes show groups of bands that shift
with substitution at the ligand and groups that remain unchanged in wavenumber. Electrochemical reduction of
the complexes resulted in the formation of the ligand radical anion species for all but one system. This was
confirmed by UV/VIS spectroelectrochemistry. Using resonance Raman spectroelectrochemistry marker bands
have been identified for the radical anion species. The excited states of the complexes were studied by excited-state
electronic absorption and time-resolved resonance Raman techniques. The former spectra are ambiguous as to the
nature of the lowest excited state; however, the latter spectra confirm that this state is ligand-centred for complexes
of dppz and its 11-methyl derivative. Complexes with the 11-nitro derivative appear to form excited states that are
MLCT in nature.
Metal complexes with ligands based on dipyrido[3,2-a:2Ј,3Ј-c]-
LC state
phenazine (dppz) are of interest because of their interactions
with DNA,1 their molecular switching properties2 and their
Relaxation
potential in solar energy conversion schemes.3 In most solar
MLCT state
energy conversion schemes, based on transition-metal com-
plexes, solar energy is captured through a charge-separation
scheme, such as excitation to the metal-to-ligand charge-
transfer, MLCT, state.4 The advantage of using dppz-based sys-
tems in such schemes is that the overlap between the metal dπ
and ligand π* orbitals is small,5 thus reducing the back
electron-transfer rate. However the poor overlap between these
orbitals results in a very low oscillator strength for the MLCT
transition reducing the efficiency of directly forming the
charge-separated state. A possible solution to this is to use the
strongly allowed ligand-centred, LC, transitions of the complex
to sensitize the formation of the charge-separated MLCT state
as depicted in Scheme 1. For this strategy to be successful the
LC states must have a large ε in the solar spectral region, the
lowest excited state must be MLCT in nature and the photo-
physics of the complex must lead to population of the MLCT
state.
The excited-state properties of complexes with dppz as a
ligand have been studied by a number of groups. The nature of
the lowest excited state for these systems falls into two categor-
ies: (1) for complexes containing ruthenium() and osmium()
evidence from excited-state absorption, electronic emission and
time-resolved resonance Raman spectroscopies suggests a
lowest excited state of MLCT character;6 (2) for rhenium()
complexes with dppz the nature of the lowest excited state is
somewhat ambiguous. Schanze and co-workers7 have shown
that [Re(CO)3(dppz)(Mepy)]ϩ (where Mepy is 4-methyl-
pyridine) has an LC lowest excited state. The emission spectrum
from this complex shows vibrational structure characteristic of
hν
Ground state
Scheme 1
such states. A detailed analysis of the quenching properties
of the complex revealed the presence of an MLCT state
lying close in energy to the lowest LC state. Time-resolved
infrared spectroscopy has been used to probe the nature of the
lowest excited state of [Re(CO)3(dppz)(PPh3)]ϩ.8 The infrared
spectrum provided definitive evidence for an LC lowest excited
state. Furthermore the same paper reports that the emission
spectrum of [Re(CO)3(dppz)(PPh3)]ϩ and the corresponding
chloro complex appears to be of MLCT origin at room
temperature.
In the foregoing studies it is clear that for rhenium() com-
plexes with dppz the LC and MLCT states lie very close in
energy. Our study seeks to use substitution at the dppz ligand to
alter the levels of the LC and MLCT states. In this regard the
properties of a number of copper() and rhenium() complexes
with substituted dipyrido[3,2-a:2Ј,3Ј-c]phenazine ligands have
been investigated. The advantages of using such complexes are
ϩ
as follows. (i) The Cu(PPh3)2 and Re(CO)3Cl moieties do not
possess chromophores in the visible region, thus the spec-
troscopy of the complexes is more readily interpreted. (ii)
Copper() and rhenium() offer interesting contrasts in photo-
physical behaviour. The excited-state properties of copper()
complexes are strongly affected by the substituent groups on
the ligands used.9 For rhenium() systems the photophysical
properties, such as excited-state lifetime, depend more on the
energy-gap law than on steric effects.10 (iii) The metal moieties
have different π-acid properties perturbing the ligand to differ-
ing extents when complexed.
† Supplementary data available: electronic and resonance Raman spec-
tra. For direct access see http://www.rsc.org/suppdata/dt/1998/609,
otherwise available from BLDSC (No. SUP 57350, 8pp) or the RSC
Library. See Instructions for Authors, 1998, Issue 1 (http://www.rsc.org/
dalton).
J. Chem. Soc., Dalton Trans., 1998, Pages 609–616
609

View Article Online
The substituents on the dipyrido[3,2-a:2Ј,3Ј-c]phenazine lig-
ands are electron-withdrawing NO2 and Cl groups or electron-
donating methyl groups.
dt), 9.25 (2 H, dt), 8.21 (1 H, d), 8.08 (1 H, s), 7.76 (3 H, m) and
2.70 (3 H, s) (Found: C, 77.29; H, 3.95; N, 19.12. Calc.: C,
77.01; H, 4.08; N, 18.91%); yield 55%.
L3: δH(200 MHz, solvent CDCl3, standard SiMe4) 9.54 (2 H,
m), 9.30 (2 H, m), 9.20 (1 H, d), 8.64 (1 H, dd), 8.44 (1 H, d) and
7.80 (2 H, m) (Found: C, 65.73; H, 2.98; N, 21.28. Calc.: C,
66.05; H, 2.77; N, 21.40%); yield 48%.
N
N
N
N
N
N
N
N
L4: δH(200 MHz, solvent CDCl3, standard SiMe4) 9.48 (2 H,
dd), 9.28 (2 H, dd), 8.42 (2 H, s) and 7.77 (2 H, dd) (Found: C,
61.34; H, 2.06; N, 15.77. Calc.: C, 61.56; H, 2.30; N, 15.95%);
yield 56%.
CH3
L1
L2
Synthesis of complexes
Copper() complexes were formed by the reaction of [Cu-
(PPh3)2(MeCN)2]BF4 with the appropriate ligand in dichloro-
methane solution under an argon atmosphere. The reactions
proceeded rapidly at room temperature forming the coloured
complex within seconds of mixing the two reagents. The com-
plexes were purified by recrystallization from methanol:
[CuL1(PPh3)2]BF4 1 (Found: C, 68.07; H, 4.60; N, 5.51. Calc.: C,
N
N
N
N
N
N
N
N
Cl
Cl
NO2
L3
L4
67.74; H, 4.21; N, 5.85%), yield 70%; [CuL2(PPh3)2]BF4
2
+
(Found: C, 67.79; H, 4.23; N, 6.11. Calc.: C, 68.00; H, 4.36; N,
5.76%), yield 76%; [CuL3(PPh3)2]BF4 3 (Found: C, 64.35; H,
4.03; N, 7.36. Calc.: C, 64.71; H, 3.92; N, 6.99%), yield 73%;
[CuL4(PPh3)2]BF4 4 (Found: C, 62.86; H, 3.88; N, 5.29. Calc.: C,
63.13; H, 3.73; N, 5.45%), yield 60%.
N
N
(Ph3P)2Cu
N
N
The rhenium complexes were formed by the addition of
[Re(CO)5Cl] to a solution of appropriate ligand in dry meth-
anol under reflux conditions. The reacting mixture was kept
under an argon atmosphere for ca. 8 h. At this time it was pale
yellow and contained the complex as an orange-yellow precipi-
tate. This was filtered off while hot and washed with ether. It
was found that if the mixture was allowed to cool residual lig-
and would coprecipitate with the complex.
1
N
N
N
N
Cl(OC)3Re
[ReL1(CO)3Cl] 5: δH(200 MHz, solvent CDCl3, standard
SiMe4) 9.87 (2 H, dd), 9.47 (2 H, dd), 8.46 (2 H, dd) and 8.04 (4
H, m) (Found: C, 42.84; H, 1.54; N, 9.67. Calc.: C, 42.90; H,
1.71; N, 9.53%); yield 78%.
5
[ReL2(CO)3Cl] 6: δH(200 MHz, solvent CDCl3, standard
SiMe4) 9.82 (2 H, dt), 9.45 (2 H, dt), 8.32 (1 H, d), 8.20 (1 H, s),
8.00 (2 H, dd), 7.88 (1 H, dd) and 2.64 (3 H, s) (Found: C, 43.71;
H, 1.80; N, 9.29. Calc.: C, 43.89; H, 1.84; N, 9.31%); yield 72%.
[ReL3(CO)3Cl] 7: δH(200 MHz, solvent CDCl3, standard
SiMe4) 9.86 (2 H, dd), 9.52 (2 H, dd), 9.38 (1 H, d), 8.80 (1 H,
dd), 8.63 (1 H, d) and 8.09 (2 H, m) (Found: C, 39.73; H, 1.10;
N, 11.37. Calc.: C, 39.84; H, 1.34; N, 11.07%); yield 84%.
[ReL4(CO)3Cl] 8: δH(200 MHz, solvent CDCl3, standard
SiMe4) 9.79 (2 H, dd), 9.48 (2 H, dd), 8.60 (2 H, s) and 8.04 (2
H, dd) (Found: C, 38.41; H, 1.06; N, 8.83. Calc.: C, 38.40; H,
1.23; N, 8.53%); yield 75%.
Experimental
Materials
The compounds [Re2(CO)10], 1,10-phenanthroline, o-phenylene-
diamine, 2,3-diaminotoluene, 1,2-diamino-3-nitrobenzene and
1,2-diamino-4,5-dichlorobenzene were obtained commercially
and used without further purification. The complex [Cu-
(PPh3)2(MeCN)2]BF4 was prepared by the reaction of [Cu-
(MeCN)4]BF4 with 2 equivalents of PPh3 in acetonitrile solu-
tion,11 [Re(CO)5Cl] from [Re2(CO)10] by the method of Schmidt
et al.12 For electrochemical and spectroscopic measurements
solvents of spectroscopic grade were used. These were further
purified by distillation and stored over 5 Å molecular sieves.
The supporting electrolytes used in the electrochemical meas-
urements were tetrabutylammonium perchlorate and hexa-
fluorophosphate. These were further purified by repeated
recrystallizations from ethanol–water and ethyl acetate–diethyl
ether respectively.13
Physical measurements
A Perkin-Elmer Lambda-19 spectrophotometer was used for
collection of electronic absorption spectra. This was calibrated
with a Ho2O3 filter and spectra were run with a 2 nm resolution.
Cyclic voltammograms were obtained from argon-purged
degassed solutions of compound (ca. 1 m) with 0.1  concen-
tration of NBu4ClO4 or NBu4PF6 present. The electrochemical
cell consisted of a 1.6 mm diameter platinum working electrode
embedded in a Kel-F cylinder with a platinum auxiliary elec-
trode and a saturated potassium chloride calomel reference
electrode. The potential of the cell was controlled by an EG&G
PAR 273A potentiostat with model 270 software. The NMR
spectra were recorded using a Varian 200 MHz spectrometer.
Raman scattering was generated using continuous wave
(Spectra-Physics model 166 argon ion) and nanosecond pulsed
(Continuum Nd:YAG Surelite I-10) lasers. The sample was
held in a spinning NMR tube, or optically transparent thin-
layer electrode (OTTLE) cell, and the scattering collected in a
Ligand synthesis
Ligands were prepared by the Schiff-base condensation of 1,10-
phenanthroline-5,6-dione with the appropriate diamino com-
pound in ethanol under reflux. The best yields were obtained if
the diamino compound was present in ca. 20% excess.14 The
dione was prepared by oxidation of 1,10-phenanthroline
following the method of Gillard et al.15
L1: δH(200 MHz, solvent CDCl3, standard SiMe4) 9.65 (2 H,
dd), 9.28 (2 H, dt), 8.36 (2 H, dd), 7.93 (2 H, dd) and 7.80 (2 H,
dd) (Found: C, 76.21; H, 3.69; N, 19.51. Calc.: C, 76.58; H, 3.57;
N, 19.85%); yield 40%.
L2: δH(200 MHz, solvent CDCl3, standard SiMe4) 9.59 (2 H,
610
J. Chem. Soc., Dalton Trans., 1998, Pages 609–616

View Article Online
135Њ backscattering geometry. The irradiated volume was
imaged into a Spex 750M spectrograph using a two-lens
arrangement.16 The spectrograph was equipped with an 1800 g
mmϪ1 holographic grating which provided a dispersion of 0.73
nm mmϪ1. The Raman photons were detected using a Princeton
Instruments liquid-nitrogen-cooled 1152-EUV charge-coupled
detector (CCD) controlled by a Princeton Instruments ST-130
controller; CSMA v2.4 software (Princeton Instruments) was
used to control the CCD and spectra were analysed and pre-
pared for presentation using GRAMS/32 (Galactic Industries
Corp.) software. Spectral windows were approximately 18 nm
wide and were calibrated using emission lines from a neon lamp
or from an argon-ion laser. The calibrations were checked by
measuring the Raman band wavenumber positions for a known
solvent.17 For the data reported herein the calibrations were
accurate to ca. 1 cmϪ1. Rayleigh and Mie scattering from the
sample were attenuated using a Notch filter (Kaiser Optical
Systems Inc.) of appropriate wavelength. A polarization
scrambler was placed in front of the spectrograph entrance slit.
A 150 µm slit width was used on the spectrograph and this gave a
resolution of ca. 6 cmϪ1 with 450 nm excitation.
For the time-resolved resonance Raman measurements 448.3
nm was employed as an excitation wavelength. This was gener-
ated by stimulated Raman scattering18 through acetonitrile,
contained in a 10 cm cell, using the 354.7 nm third harmonic of
the Nd:YAG pulsed laser. It was found that with 50 mJ per
pulse of 354.7 nm light, 3.5 mJ per pulse of 448.3 nm was
obtained. The pulse energy of the 448.3 nm beam was attenu-
ated by lowering the 354.7 nm pump energy.
The electronic absorption spectra of reduced species were
measured using an OTTLE cell with a platinum grid as the
working electrode.19 For Raman spectra of reduced species a
similar cell was employed. Initial measurements found that the
signal-to-noise ratios of the Raman spectra were reduced
because of reflection off the platinum grid. This problem was
alleviated by removing a portion of the centre of the grid (ca.
2 × 4 mm) and aligning the laser to irradiate the solution in that
region.
and rhenium() complexes with electron-withdrawing ligands L3
and L4 are more readily reduced relative to the complexes with
ligand L1. The pattern of shifts in reduction potential with sub-
stitution is unsurprising in that the electron-withdrawing
groups shift EoЈ to more positive values. Binding of the ligands
to copper() has almost no effect on the EoЈ values, however the
rhenium complexes are easier to reduce than the ligands alone.
The differences observed between differing rhenium complexes
is similar to that observed between the corresponding free
phenazines. Therefore the reductions are assigned as ligand
based.
The electronic absorption data for the ligands and complexes
are presented in Table 2. The absorption spectra of the ligands
show π* ← π transitions in the 370 nm region. Stronger
ligand-centred transitions are observed at wavelengths shorter
than 300 nm. The electronic spectra of the complexes show
visible absorptions tailing out to 400–500 nm. These visible
shoulders extend further into the red with complexes which are
easier to reduce.
The Raman spectra of the ligands, L1–L3, are complicated by
strong emission signals even at excitation wavelengths in the
red, i.e. 632.8 nm. It is possible to observe the surface-enhanced
Raman scattering (SERS) by measuring the scattering from a
roughened silver electrode.22 The SERS spectra of L1–L4 are
shown in Fig. 1 (band positions are in Table 3) with the reson-
ance Raman spectrum of L4 in CH2Cl2, generated at 363.8 nm.
The SERS spectra are similar, showing strong bands at about
1580 and 1600 cmϪ1. The resonance Raman spectrum of L4
shows a number of bands at similar wavenumbers to those of
the SERS spectrum, however the strong features at 1600 cmϪ1
are absent. This suggests that these features are enhanced
through interaction with the silver electrode surface. The wave-
numbers of observed bands in the spectra of the ligands and
complexes are listed in Table 3. The resonance Raman spectra
of the rhenium() complexes are shown in Fig. 2.
Table 1 Electrochemical data for ligands and complexes
Excited-state absorption (ESA) difference spectra were
measured using apparatus described previously.20
Compound
EoЈa/V
Ϫ1.28
Compound
EoЈa/V
Ϫ1.05
L1
L3
L2
Ϫ1.30
L4
Ϫ0.93
Results
1
Ϫ1.25
Ϫ1.31
Ϫ1.04
Ϫ0.93
5
6
7
8
Ϫ1.01
Ϫ1.04
Ϫ0.95
Ϫ0.83
The compounds are readily soluble in organic solvents. They
appear stable in solution for prolonged periods. The rhenium()
complexes are less soluble than their copper() counterparts but
are also stable in solution.
Electrochemical data for the ligands, copper() and rhe-
nium() complexes in dichloromethane are presented in Table 1.
The methyl group has little effect on the position of the first
reduction potential in the metal complexes. However, copper()
2b
3
4
a Reduction potentials are given versus SCE. Solvent used CH2Cl2, sup-
porting electrolyte NBu4ClO4 or NBu4PF6. All waves are fully revers-
ible, unless indicated, showing a ratio of currents for the anodic and
cathodic peaks of unity and peak separation between anodic and cath-
odic waves of ca. 60 mV.21 b Does not show reversible electrochemistry.
Table 2 Electronic absorption data for ligands and complexes and their reduced species
λmax/nm (CH2Cl2)* (10Ϫ4 × ε/Ϫ1 cmϪ1
)
Compound
Parent species
Reduced species
L1
L2
L3
L4
1
270 (4.37), 367 (1.13), 379 (1.29)
273 (4.70), 365 (1.05), 385 (1.22)
297 (4.31), 306 (3.84), 371 (1.61), 389 (1.31)
272 (5.39), 370 (1.54), 391 (2.13)
277 (5.5), 364 (0.86)
281 (8.35), 369 (1.64)
285 (6.35), 368 (1.90), 386 (1.45) (sh)
278 (12.0), 369 (2.50), 388 (2.74)
280 (5.00), 360 (1.38)
285 (9.71), 324 (2.03) (sh), 371 (2.01), 387 (2.05)
285 (5.67), 296 (5.66), 320 (1.84) (sh), 359 (1.63)
283 (9.97), 322 (2.07) (sh), 370 (2.36), 375 (2.38), 389 (2.21)
267 (3.84), 361 (1.08), 487 (0.494), 569 (0.245)
263 (3.08), 297 (2.21), 364 (0.635), 484 (0.223)
297 (4.31), 358 (1.68), 387 (sh), 413 (sh), 483 (0.75), 513 (0.59)
264 (2.65), 374 (0.632), 483 (0.257)
270 (3.8), 335 (1.31), 587 (0.45)
272 (7.85), 365 (1.12), 384 (1.32)
267 (4.43), 355 (2.58), 405 (1.48), 488 (1.23), 520 (1.13), 802 (0.717)
266 (7.97), 335 (3.75), 370 (1.96), 391 (2.01), 609 (1.06)
258 (3.51), 297 (2.79), 336 (2.18), 584 (0.945)
260 (5.10), 297 (3.87), 347 (3.58), 601 (1.38)
284 (3.78), 357 (2.85), 406 (1.50), 491 (1.11), 523 (1.09)
260 (5.74), 294 (3.88), 338 (4.27), 394 (2.00), 458 (0.920), 604 (1.61)
2
3
4
5
6
7
8
* Except for L2 (MeOH).
J. Chem. Soc., Dalton Trans., 1998, Pages 609–616
611

View Article Online
The changes in the electronic absorption spectra of com-
pounds L1–L4 through electrochemical reduction are presented
in Table 2. The electrochemical reduction of L1 to (L1)Ϫ has
been reported by Kaim and co-workers.23 They observed the
bleach of the ground-state features at 379 and 359 nm with new
bands at 572, 545 and 450 nm upon reduction. The spectrum of
(L1)Ϫ generated in CH2Cl2 differs somewhat in terms of band
intensities from that reported by Kaim. However we obtained
identical spectral changes to those reported when using dmf as
solvent. For L1–L4 the changes in the spectra with reduction
have well defined isosbestic points. The electronic spectra of
the reduced ligands show visible absorptions that are assigned
as π* ← π transitions of the radical anion. The presence of
long-wavelength absorptions is characteristic of radical anion
polypyridyl ligands.24
The wavelengths of the absorption bands observed in the
reduced species of complexes 1–8, are presented in Table 2.
Well defined isosbestic points are observed for each reduction,
indicating the conversion from one species into another, or a
number of other species with minimum side reactions. For all
complexes except 2 the observed changes are fully reversible.
The electronic spectra of all the reduced complexes except 2Ϫ
show absorption features similar to those of the corresponding
reduced ligands. For 2 application of a potential to a solution
of the complex gives a species with an electronic spectrum
identical to that of free L2. This strongly suggests that the
reduction of 2 is occurring at the copper() site, thus leading to
demetallation of the complex. Such electrochemistry is very
common among copper() systems,25 indeed the behaviour of
the other copper complexes, 1, 3 and 4 is unusual in that they
show reduction at the ligand.26
Fig. 1 Raman spectra of ligands: (a) L1, (b) L2, (c) L3, (d) L4.
λexc = 514.5 nm, 20 mW. Scattering measured from the surface of a
roughened silver electrode in methanolic solution (for L1–L3), CH2Cl2
solution for L4. (e) L4, λexc = 363.8 nm, 8 mW, CH2Cl2 solution
Strong emission from the reduced products of L1–L4 makes it
impossible to measure the resonance Raman spectra for these
species. However, it is possible to observe the surface-enhanced
Fig. 2 Upper trace: ground-state resonance Raman spectrum in CH2Cl2 solution (1 m), λexc = 457.9 nm continuous wave, laser power 20 mW for
complexes 5 (a), 6 (b), 7 (c) and 8 (d). Middle trace: corresponding time-resolved resonance Raman spectrum, λexc = 448.3 nm pulsed excitation, laser
energy 3 mJ per pulse. Lower trace: resonance Raman spectrum of electrochemically reduced 5 (a) or 8 (d ) in CH2Cl2 solution (1 m), λexc = 514.5 nm
continuous wave, laser power 20 mW; or surface-enhanced resonance Raman spectrum of ligand L2 (b) or L3 (c) in methanol reduced at roughened
silver electrode, λexc = 514.5 nm, 20 mW
612
J. Chem. Soc., Dalton Trans., 1998, Pages 609–616

View Article Online
Table 3 Wavenumbers (cmϪ1) of bands observed in Raman spectra for ligands and complexes
L1 a
L2 a
L3 a
L4 b
L4 c
1d
2d
3d
4d
5d
6d
7d
8d
1601s
1574s
1602
1578
1604
1579
1546
1495
1580 (br)
1614w
1597w
1613vw
1594vw
1503vw
1487vw
1443m
1412s
1592w
1599w
1579w
1601w
1582w
1501w
1492w
1450m
1418s
1595w
1576w
1496s
1472
1495m
1468m
1443m
1402m
1477m
1446s
1401s
1495m
1406s
1491s
1449m
1407m
1487m
1450s
1404s
1443m
1408m
1456
1447
1427
1401
1413s
1324m 1324w
1402s
1351w
1336m
1400s
1342
1321
1333s
1317m
1346w
1316s
1315s
1316s
1316s
1314s
1317s
1315s
1266m
1264w
1264m
a The SERS spectrum in methanolic solution. b The SERS spectrum in CH2Cl2 solution. c λexc = 363.8 nm, 8 mW, CH2Cl2 solution. d λexc = 457.9 nm,
20 mW, CH2Cl2 solution. s = Strong, m = medium, w = weak, v = very, br = broad.
Table 4 Wavenumbers (cmϪ1) for bands observed in the resonance
Raman spectra of reduced ligands and complexes
Table 5 Wavenumbers (cmϪ1) for bands observed in the time-resolved
resonance Raman spectra of complexes in CH2Cl2 solution
(L1)Ϫ a
(L2)Ϫ a
(L3)Ϫ a
(L4)Ϫ a
5Ϫ b
8Ϫ b
1
2
3
5
6
7
1633s
1636s
1592w
1591m
1597vw
1577vw
1529w
1589m
1520w
1480w
1462w
1592m
1532m
1518m
1481m
1462m
1340w
1591s
1524s
1496w
1466w
1423w
1388w
1329s
1584s
1581s
1586s
1505m
1453w
1491w
1456w
1479w
1448w
1493w
1441m
1403s
1357m
1316m
1277w
1486m
1440s
1408m
1345m
1315s
1495vw
1405s
1441m
1410s
1491m
1449m
1348w
1316s
1408s
1389w
1302m
1264m
1363m
1361m
1336vw
1297w
1278w
1318m
1293m
1278m
a The SERRS spectrum in methanol reduced at roughened silver
electrode. λexc = 514.5 nm, 20 mW. b Resonance Raman spectrum in
CH2Cl2 solution. λexc = 514.5 nm, 20 mW.
features at 1278 and 1297 cmϪ1 appear in the pulsed spectrum.
Fig. 2(c) shows the time-resolved spectrum of 7. The pulsed
spectrum shows one feature (1529 cmϪ1) not observed in the
ground-state spectrum generated at 457.9 nm.
resonance Raman scattering (SERRS) of the reduced ligands at
a roughened silver electrode. The wavenumbers for bands
observed in the SERRS spectra of (L1)Ϫ–(L4)Ϫ are given in
Table 4. The spectra of (L1)Ϫ and (L2)Ϫ are similar in the high-
wavenumber region, i.e. around 1600 cmϪ1. The spectrum of
(L3)Ϫ is significantly different with no 1630 cmϪ1 feature. This
may be caused by the nitro group interacting with the surface of
the roughened electrode. Such interactions are well known for
silver and gold electrode surfaces.27 This normally results in
increased enhancement of Raman scattering for the nitro vibra-
tion, relative to other vibrations of the molecule, and small
shifts in the wavenumber of the band, relative to scattering from
a bulk sample. It is possible to generate the spectra of 5Ϫ and 8Ϫ
in solution [Fig. 2(a) and 2(d)] due to the low emission from
these species. Band positions are listed in Table 4.
The wavenumbers of bands observed in the resonance
Raman spectra using 448 nm pulsed excitation for complexes 1–
3 and 5–7 are listed in Table 5. No excited-state features are
observed in the time-resolved spectra of 4 or 8. The 448 nm
excitation is in resonance with electronic transitions of the
ground and excited state for all of the complexes. Thus it is
possible to generate the time-resolved resonance Raman spectra
of the excited states using the single-colour pump-probe
protocol.28
The pulsed spectrum for complex 8 has no new features com-
pared to the ground state. As the pulse energy is increased the
ground-state bands diminish relative to the solvent feature at
1423 cmϪ1
.
Discussion
The majority of studies of complexes with dppz and related
ligands have focused on the interaction of the systems with
DNA.1 A number of studies have examined the detailed spec-
troscopic properties of the complexes.5,23 This study represents
an examination of the spectroscopic, electrochemical and
excited-state properties of a series of complexes in which the
ligand is altered by substitution with electron-donating and
-withdrawing groups.
A number of questions concerning the use of such ligands
need to be addressed. (i) How does binding to the metal alter
the ground-state properties? (ii) How do the complexes behave
when reduced, i.e. their spectroelectrochemistry? This is an
important point because it provides spectral information on the
ligand-radical anion species which is present in the MLCT
excited state. (iii) How do the complexes behave when photo-
excited?
The resonance Raman spectra for the rhenium complexes
with 448 nm pulsed excitation are shown in Fig. 2. Resonance
Raman scattering generated with 448 nm pulsed excitation for 5
shows similar bands to the ground-state spectrum generated
with 457.9 nm excitation. However, a number of important dif-
ferences are observed: (i) the ground-state band at 1314 cmϪ1 is
almost completely bleached in the pulsed spectrum; (ii) a new
band at 1278 cmϪ1 is observed in the pulsed spectrum, assigned
to the excited state of 5. Fig. 2(b) shows the time-resolved spec-
trum of 6. This differs from the ground-state resonance Raman
spectrum in a number of ways: ground-state features at 1264,
1317, 1582 and 1601 cmϪ1 are absent from the pulsed spectrum
and that at 1418 cmϪ1 appears shifted to 1408 cmϪ1. New
Ground-state properties
Theoretical treatments of compound L1 suggest that the dppz
ligand has three MOs which lie close in energy:23 (i) the lowest-
lying MO is termed b1(phz), and is localized on the phenazine
(phz) portion of the ligand; (ii) the b1(ψ)29 orbital, localized at
the 2,2Ј-bipyridine (bpy) portion of the ligand; (iii) the a2(χ)
orbital also localized at the bpy portion of the ligand.29 These
calculations show that the amplitude of the wavefunction for
the LUMO [b1(phz)] is very small at the chelating nitrogens.
Thus the oscillator strength and ε for the MLCT transitions
of complexes with L1 would be expected to be small. The
J. Chem. Soc., Dalton Trans., 1998, Pages 609–616
613

View Article Online
electronic spectra of 1–8 show weak absorptions tailing to 450–
500 nm depending on the complex. The ε for these features are
ca. 1000 Ϫ1 cmϪ1. The absorptions extend further into the red
in complexes with electron-withdrawing ligands such as L3.
Furthermore, the rhenium() complexes absorb further into the
red than do the copper() counterparts. They are also easier to
reduce suggesting the π* ligand acceptor orbital is at lower
energy. These patterns of behaviour are consistent with MLCT
transitions.
A number of studies report the resonance Raman spectra of
L1 and its ruthenium() complexes.30 The other ligands and
their complexes have not been reported to date. The resonance
Raman spectrum of [Ru(dmb)2L1]2ϩ (dmb = 4,4Ј-dimethyl-2,2Ј-
bipyridine), generated using 496.5 nm excitation, has been
reported.6 The following bands were assigned to dppz: 1181,
1304, 1404, 1446, 1469, 1488, 1569 and 1596 cmϪ1. The reson-
ance Raman spectrum of [Ru(phen)2L1]2ϩ (phen = 1,10-
phenanthroline), generated using 457.9 nm excitation, showed
dppz bands at 1189, 1266, 1311, 1361, 1407, 1475, 1496, 1577
Fig. 3 Electronic absorption spectrum of complex 1 in CH2Cl2. Inset:
∆A spectrum of transient electronic absorption signals for the excited
state of 1 generated by 355 nm pulsed (7 ns duration) excitation
normal modes of vibration of the ligand and the nature of the
interaction with the surface.34 There is therefore a need for
caution in relating SERRS spectra to those of species in the
solution phase. However, for two of the complexes it was pos-
sible to generate a resonance Raman spectrum of the reduced
species. These spectra of 5Ϫ and 8Ϫ [Fig. 2(a) and 2(d)] are
similar, showing two prominent bands for the radical anion
species. These lie at ca. 1362 and 1583 cmϪ1. For 5Ϫ the 1581
cmϪ1 feature is close to the 1589 cmϪ1 band observed in the
SERRS of (L1)Ϫ. A band at 1367 cmϪ1 is also observed in the
SERRS spectrum although it is of low intensity. The SERRS
spectra for (L1)Ϫ contain the bands observed for 5Ϫ but at
rather different intensities. The strongest band in the SERRS
spectrum lies at 1633 cmϪ1 and is presumably due to some
surface-binding phenomenon. (The SERS spectrum of 1,10-
and 1600 cm
described as a dominant feature. In the second study the most
prominent features lie at 1475, 1496, 1577 and 1600 cmϪ1
.
Ϫ1 29 In the former study the band at 1404 cmϪ1 was
.
The resonance Raman spectra of complexes 1–8 differ from
those reported for ruthenium complexes in that only the dppz
ligand is involved in the resonant chromophore(s). For all of
the complexes reported herein the higher wavenumber bands at
1600 cmϪ1 are only weakly enhanced and the strongest Raman
bands lie at ca. 1400 and at ca. 1315 cmϪ1. The band at ca. 1400
cmϪ1 shifts significantly with substitution of the ligand and
this behaviour is consistent with assignment to a dominant
ring stretch involving the entire ligand framework. The 1315
cmϪ1 band is wavenumber-insensitive to substitution.
A
phenanthroline-based vibration, such as that proposed by
Campos-Valette and co-workers,31 is consistent with this
behaviour. This is a vibration that only involves motion of the
atoms bound close to the metal. Complexes containing the
nitro group, 3 and 7, show bands at 1333 and 1346 cmϪ1 respect-
ively. These are assigned as the nitro stretch on the basis of their
frequency.32
Resonance Raman spectroscopy provides some insight into
where an excited electron may be directed in an MLCT tran-
sition by virtue of the fact that the modes most strongly
enhanced are associated with the resonant chromophore.33 The
strong enhancement of the 1400 cmϪ1 band and the enhance-
ment of the nitro vibration in the spectra of complexes 3 and 7
is consistent with an MLCT from the transition metal to
b1(phz). The strong enhancement of the phen-based mode at
1300 cmϪ1 suggests that an MLCT transition involving the
‘bpy-based’ MOs may also be resonant at 457.9 nm.
phenanthroline shows a high-frequency mode at 1625 cmϪ1
See ref. 34.)
.
It is clear that the SERRS spectra give only a qualitative
signature for the radical anion species. It is pleasing that the
spectra of the two reduced complexes are similar as might be
expected. The spectra of complex 8Ϫ, and the corresponding
ligand SERRS spectrum of (L4)Ϫ both show strong bands at
1586 and 1584 cmϪ1 respectively. These provide spectral
marker bands for the radical anion species. Further evidence
that these are radical anion features comes from the time-
resolved resonance Raman spectra of [Ru(phen)2(dppz)]2ϩ
.
The lowest excited state for this compound is assigned as an
MLCT state in which the dppz is reduced, thus the spectra
Ϫ
show features associated with the dppz
radical anion.
᝽
Bands are observed at 1575, 1457 and 1366 cmϪ1, which lie
close to the radical anion features identified in our spectrum
of 5Ϫ.
Spectroelectrochemistry
Excited-state properties
For all of the ligands and complexes, except 2, reduction is
based on the ligand, thus a radical anion is formed. In the case
of the ligands the spectra observed shift considerably with sub-
stitution, exemplified by those of (L3)Ϫ and (L2)Ϫ (Table 2). This
is consistent with population of a b2(phz) MO. On binding to a
metal centre there are some small changes in the electronic spec-
tra of the reduced species. The species (L3)Ϫ, 3Ϫ and 7Ϫ all show
absorptions at approximately 360, 410, 490 and 520 nm. This
suggests the metal has little effect on the MO occupied by the
reducing electron in the reduced species.
In attempting to model the Raman spectral signatures of the
radical anion species of ligands L1–L4 SERRS spectra were
collected because conventional Raman spectra were obscured
by strong emission from the reduced ligands. The SERRS
spectra may be significantly perturbed from solution-phase
spectra as evidenced by the SERS and resonance Raman spec-
tra of L4 (Fig. 2). Different enhancement patterns and wave-
number shifts in band positions may occur depending on the
The λmax for the ESA spectra for all systems (Table 6) that could
be measured are remarkably similar. A typical spectrum, that of
complex 1, is shown in Fig. 3. The similarity between the spec-
tra suggests that all of the observed excited states are ligand-
centred in nature. The dynamics of the excited states are some-
what more ambiguous. The lifetimes of the ligands are long, up
to 10 µs, whereas the complexes with these ligands have rather
short lifetimes. For L2 τ ≈ 2.5 µs, but for 2 it is 1.7 µs and
for 6 it is only 0.34 µs. This suggests that the metal is having a
significant perturbation on the excited state. The most striking
example of this is with the nitro-ligand, L3. The lifetime for the
ligand excited state is 10 µs, those of the copper and rhenium
complexes are ca. 6 ns.
The upshot of this is that the ESA spectra alone are insuffi-
cient to provide a clear idea as to the nature of the lowest
excited state in these systems. Time-resolved resonance Raman
spectroscopy offers a more incisive probe of the excited-state
614
J. Chem. Soc., Dalton Trans., 1998, Pages 609–616

View Article Online
Conclusion
Table 6 Excited-state lifetimes for ligands and complexes in degassed
CH2Cl2 at room temperature
This is the first study in which the nature of the lowest excited
state for copper() and rhenium() complexes with dppz-type
ligands in which substituents on the ligands are changed has
been carried out. The presence of electron-withdrawing groups
appears to lower the energy of the MLCT state so that it dom-
inates the THEXI state population. Without such groups pres-
ent the lowest excited state is LC in nature. The determination
of the nature of the excited state is not straightforward. Time-
resolved vibrational spectroscopy, in this case resonance
Raman spectroscopy, appears to be the least ambiguous tool
for probing the excited-state nature.
We have found that when the MLCT state is lowest in energy
its lifetime is very short. This clearly limits the utility of such
complexes in homogeneous photocatalytic systems.37 However,
they may be of use in heterogeneous systems, such as liquid-
junction solar cells. In such cells the dye material is known
rapidly to photoinject into the semiconductor conduction
band38 thereby permitting dye materials with short lifetimes to
operate successfully.
Compound
τ/µs
3(300)
2.5
10
10
0.16
1.7
<0.006
n.s.
0.04
0.34
<0.006
n.s.
ESA, λmax/nm
L1*
L2
L3
L4
1
460
465
460
462
460
475
—
2
3
4
—
5
475
470
—
6
7
8
—
n.s. = No signal observed.
* Ref. 7 reports a biphasic decay with two components; the lifetime of
the stronger signal is 3 µs and that of the second decay is given in
parentheses.
nature. The spectra for the complexes with L1 show distinct
bands associated with their respective excited states. The life-
times of these states are long with respect to the probe laser
pulse and, assuming the absorption of the sample is suf-
ficiently high, efficient population of the state should be read-
ily achievable. The spectral signatures observed bear no resem-
blance to the resonance Raman spectra of the ligand radical
anion species (L1)Ϫ, as observed either in the SERRS spectrum
of the ligand or the Raman OTTLE spectrum for the rhe-
nium() complex, 5. The simplest explanation for this finding is
that the thermally equilibrated excited (THEXI) states for 1 and
5 do not possess any radical anion character and are LC in
nature. This finding is consistent with emission temperature
studies carried out on 5 by Meyer and co-workers.6 For the
rhenium() complex 6 with ligand L2, the 448 nm pulsed spec-
trum clearly shows excited-state features not present in the
ground-state spectrum. The excited-state features are similar to
those observed for complexes with L1; they do not correspond
well with the SERRS spectrum of (L2)Ϫ or the Raman spectra
of the reduced rhenium complexes 5Ϫ and 8Ϫ [Fig. 2(a) and
2(d)] recorded for related ligands. The time-resolved spectrum
strongly suggests that the THEXI state for 6 has no radical
anion character and is LC based.
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
Shirtcliffe fellowship for support for Ph.D. research. This work
was supported, in part, by the New Zealand Public Good
Science Fund (Contract number UOO-508).
References
1 C. M. Dupureur and J. K. Barton, Inorg. Chem., 1997, 36, 33;
T. K. Schoch, J. L. Hubbard, C. R. Zoch, G.-B. Yi and M. Soerlie,
Inorg. Chem., 1996, 35, 4383; R. E. Homlin and J. K. Barton, Inorg.
Chem., 1995, 34, 7; Y. Jenkins, A. E. Friedman, N. J. Turro and
J. K. Barton, Biochemistry, 1992, 31, 10 809; V. W.-W. Yam,
K. K.-W. Lo, K.-K. Cheung and R. Y.-C. Kong, J. Chem. Soc.,
Dalton Trans., 1997, 2067.
2 J.-C. Chambron and J.-P. Sauvage, Chem. Phys. Lett., 1991, 182,
603; A. E. Friedman, J.-C. Chambron, J.-P. Sauvage, N. J. Turro and
J. K. Barton, J. Am. Chem. Soc., 1990, 112, 4960.
3 M. Graetzel, Coord. Chem. Rev., 1991, 111, 167.
4 M. K. Nazeeruddin, A. Kay, I. Rodicio, R. Humphrey-Baker,
E. Mueller, P. Liska, N. Vlachopoulos and M. Graetzel, J. Am.
Chem. Soc., 1993, 115, 6382.
Complexes with the nitro-substituted ligand, L3, behave dif-
ferently from the aforementioned. The excited-state resonance
Raman spectra of the copper complex 3 shows a feature at 1591
cmϪ1 which is of similar wavenumber to that observed in the
SERRS spectrum of (L3)Ϫ. Furthermore a band is observed at
1346 cmϪ1 which is close to that seen in the Raman spectra of
the reduced rhenium complexes. The time-resolved data suggest
that the THEXI state of 3 has significant radical anion charac-
ter. The spectra of 7 taken with 448 nm pulsed irradiation
show few changes from that of the ground state measured using
457.9 nm excitation [Fig. 2(c)]. The lifetime measurements
showed 7 to have a lowest excited-state lifetime less than ca. 6
ns. It would be extremely difficult to populate completely such a
state using the 10 ns pulsed laser in this study although partial
population of the state is possible.35 The 1529 cmϪ1 band
observed in the pulsed spectrum may be due to population of
the excited state. It is interesting that a band at 1524 cmϪ1 is
present in the SERRS spectrum of (L3)Ϫ. The other features
5 J.-C. Chambron and J.-P. Sauvage, Nouv. J. Chim., 1985, 9, 527;
E. Amouyal, A. Homsi, J.-C. Chambron and J.-P. Sauvage, J. Chem.
Soc., Dalton Trans., 1990, 1841.
6 J. R. Schoonover, W. D. Bates and T. J. Meyer, Inorg. Chem., 1995,
34, 6421.
7 H. D. Stoeffler, N. B. Thornton, S. L. Temkin and K. S. Schanze,
J. Am. Chem. Soc., 1995, 117, 7119.
8 J. R. Schoonover, G. F. Strouse, R. B. Dyer, W. D. Bates, P. Chen
and T. J. Meyer, Inorg. Chem., 1996, 35, 273.
9 D. R. McMillin, R. E. Gamache, jun., J. R. Kirchoff and A. A. Del
Paggio, Copper Coordination Chemistry: Biochemical and Inorganic
Perspectives, eds. K. D. Karlin and J. Zubieta, Adenine Press, New
York, 1983, p. 233.
10 J. V. Caspar and T. J. Meyer, J. Phys. Chem., 1983, 87, 952.
11 P. F. Barron, J. C. Dyason, L. M. Engelhardt, P. C. Healy and A. H.
White, Aust. J. Chem., 1985, 38, 261.
12 S. P. Schmidt, W. C. Trogler and F. Basolo, Inorg. Synth., 1990, 28,
161.
13 H. O. House, E. Feng and N. P. Peet, J. Org. Chem., 1971, 36, 2371.
14 J. E. Dickeson and L. A. Summers, Aust. J. Chem., 1970, 23, 1023.
15 R. D. Gillard, R. E. E. Hill and R. Maskill, J. Chem. Soc. A, 1970,
1447.
that may be expected for (L3)Ϫ would lie at 1329 and 1591 cmϪ1
,
if one assumes a similar spectral signature to that of the MLCT
of 3, unfortunately ground-state bands are present in this
region and may obscure transient features. With the ground and
transient state present together in the spectrum it is more dif-
ficult to interpret what is observed. Furthermore, the need to
use high pulse energies to generate the transient spectrum in 7
may lead to non-linear effects which complicate the spectrum.36
16 D. P. Strommen and K. Nakamoto, Laboratory Raman
Spectroscopy, Wiley, New York, 1984.
17 J. R. Ferraro and K. Nakamoto, Introductory Raman Spectroscopy,
Academic Press, San Diego, 1994.
18 C. A. Grant and J. L. Hardwick, J. Chem. Educ., 1997, 74, 318.
19 A. Babaei, P. A. Connor, A. J. McQuillan and S. Umapathy,
J. Chem. Educ., 1997, 74, 1200.
J. Chem. Soc., Dalton Trans., 1998, Pages 609–616
615

View Article Online
20 R. A. McNicholl, J. J. McGarvey, A. H. R. Al-Obaidi, S. E. J. Bell,
P. M. Jayaweera and C. G. Coates, J. Phys. Chem., 1995, 99, 12 268.
21 A. J. Bard and L. R. Faulkner, Electrochemical Methods,
Fundamentals and Applications, Wiley, New York, 1980.
22 M. Fleishmann, P. J. Hendra and A. J. McQuillan, Chem. Phys.
Lett., 1974, 26, 123.
30 C. G. Coates, L. Jaquet, J. J. McGarvey, S. E. J. Bell, A. H. R.
Al-Obaidi and J. M. Kelly, Chem. Commun., 1996, 35.
31 R. E. Clavijo, F. Mendizabul, W. Zamudio, R. Baraona, G. Diaz and
M. M. Campos-Valette, Vib. Spectrosc., 1996, 12, 37.
32 F. R. Dollish, W. G. Fateley and F. F. Bentley, Characteristic Raman
Frequencies of Organic Compounds, Wiley, New York, 1974.
33 R. J. H. Clark and T. J. Dines, Angew. Chem., Int. Ed. Engl., 1986,
25, 131.
23 J. Fees, W. Kaim, M. Moscherosch, W. Mateis, J. Klima, M. Krejcik
and S. Zalis, Inorg. Chem., 1993, 32, 166.
24 T. Shida, Electronic Absorption Spectra of Radical Ions, Elsevier,
Amsterdam, 1988.
25 A. H. R. Al-Obaidi, K. C. Gordon, J. J. McGarvey, S. E. J. Bell and
J. Grimshaw, J. Phys. Chem., 1993, 97, 10 942; J.-P. Sauvage, J. Am.
Chem. Soc., 1988, 111, 7791.
26 M. A. Masood and P. S. Zacharias, J. Chem. Soc., Dalton Trans.,
1991, 111; S. M. Scott, K. C. Gordon and A. K. Burrell, Inorg.
Chem., 1996, 35, 2452.
34 T. C. Strekas and P. S. Diamondopolous, J. Phys. Chem., 1990, 94,
1986.
35 K. C. Gordon and J. J. McGarvey, Chem. Phys. Lett., 1989, 162,
117.
36 P. J. Carroll and L. E. Brus, J. Am. Chem. Soc., 1987, 109, 7613.
37 K. Kalyanasundaram, Photochemistry and Photophysics of
Polypyridine and Porphyrin Complexes, Academic Press, London,
1992.
27 R. Holze, Electrochim. Acta, 1991, 36, 1523; P. Gao and M. J.
Weaver, J. Phys. Chem., 1989, 93, 6205.
38 J. M. Rehm, G. L. McLendon, Y. Nagasawa, K. Yoshihara, J. Moser
and M. Graetzel, J. Phys. Chem., 1996, 100, 9577.
28 R. F. Dallinger and W. H. Woodruff, J. Am. Chem. Soc., 1979, 101,
4391; P. G. Bradley, N. Kress, B. A. Hornberger, R. F. Dallinger and
W. H. Woodruff, J. Am. Chem. Soc., 1981, 103, 7441.
29 L. E. Orgel, J. Chem. Soc., 1961, 3683.
Received 20th August 1997; Paper 7/06110B
616
J. Chem. Soc., Dalton Trans., 1998, Pages 609–616