Spectroscopy and electrochemical properties of a homologous series of acetylacetonato and hexafluoroacetylacetonato cyclopalladated and cycloplatinated complexes

Article abstract of DOI:10.1039/b804478c

A photophysical, electrochemical and computational study has been performed on a homologous series of cyclometallated Pd(ii) (1a-f) and Pt(ii) (2a-f) complexes of general formula [(C,N)M(O,O)]; (H(C,N) = azobenzene, 2-phenylpyridine, benzo[h]quinoline; M = Pd, Pt; H(O,O) = acetylacetone, hexafluoroacetylacetone). Experimental and computational data have shown the strong influence exerted by electronegativity of the ancillary ligand on the frontier orbitals of the complexes, such an effect being enhanced for the Pt(ii) species.

Full text of DOI:10.1039/b804478c

View Article Online / Journal Homepage / Table of Contents for this issue
PAPER
www.rsc.org/dalton | Dalton Transactions
Spectroscopy and electrochemical properties of a homologous series of
acetylacetonato and hexafluoroacetylacetonato cyclopalladated and
cycloplatinated complexes†
Mauro Ghedini,*a Teresa Pugliese,a Massimo La Deda,a Nicolas Godbert,a Iolinda Aiello,a Mario Amati,b
Sandra Belviso,b Francesco Lelj,*b Gianluca Accorsi*c and Francesco Barigellettic
Received 17th March 2008, Accepted 9th May 2008
First published as an Advance Article on the web 23rd June 2008
DOI: 10.1039/b804478c
A photophysical, electrochemical and computational study has been performed on a homologous series
of cyclometallated Pd(II) (1a–f) and Pt(II) (2a–f) complexes of general formula [(C,N)M(O,O)];
(H(C,N) = azobenzene, 2-phenylpyridine, benzo[h]quinoline; M = Pd, Pt; H(O,O) = acetylacetone,
hexafluoroacetylacetone). Experimental and computational data have shown the strong influence
exerted by electronegativity of the ancillary ligand on the frontier orbitals of the complexes, such an
effect being enhanced for the Pt(II) species.
The available literature data suggests that the lack of room
Introduction
temperature luminescence for cyclopalladated compounds is likely
Square-planar cyclometallated palladium and platinum complexes
due to a rapid non-radiative deactivation of the excited levels,
are currently studied for purposes which include the preparation
as a consequence of population of metal-centered states (MC),
geometrically distorted with respect to their ground states.3a Thus,
in order to better understand the properties exhibited by square-
planar complexes which feature both a metallacycle and an
ancillary chelating ligand, particular attention has to be deserved
to the metal-mediated fine interplay among states originating from
the molecular orbitals of the two chelated rings. Moreover, an
accurate HOMO and LUMO description, in terms of atomic
orbital composition, absolute energy, and relative energy gap,
affords a set of parameters which are connected to relevant photo-
physical properties (i.e. luminescence, photorefractivity or charge
photogeneration) these complexes display.1e Thus ultimately, such
studies could be of help for the design of new ligands and the
tuning of specific desired properties of the complexes they form.
Given the potential of these compounds as photo-active
materials, we have performed an experimental and theoretical
comparative study on some homologous series of cyclometallated
Pd(II) and Pt(II) complexes. This paper provides an account of the
synthesis, spectroscopic and electrochemical characterisation, and
the computational results on the electronic structure of the acety-
lacetonato (acac) and hexafluoroacetylacetonato (hfacac) Pd(II)
and Pt(II) complexes (Chart 1). In this series of complexes, the
cyclometallated ligands are azobenzene (HAzo), 2-phenylpyridine
(HPhPy) and benzo[h]quinoline (HBzQ), (1a–f, 2a–f,).
of bio-active molecules and the synthesis of new photo-active
materials whose applications as advanced materials require species
with luminescent, or strictly related, properties.1 Luminescent
palladium complexes have received until now less attention than
platinum complexes;2 however, several investigations are presently
focused on such complexes.3
In this context, we have recently reported efficient lumine-
scent cyclopalladated Nile Red (9-diethylamino-5H-benzo[a]-
phenoxazine-5-one) complexes, comprising as ancillary lig-
ands deprotonated acetylacetone or hexafluoroacetylacetone
(1,1,1,5,5,5-hexafuoro-2,4-pentanendionato(1−)), which show the
highest emission quantum yield ever reported for cyclopalladated
complexes (12% and 50% in cyclohexane respectively).4 For these
complexes, the lower energy excited states are ligand-centered
(LC) and are not influenced by competing non-emissive excited
states located at the metal center. In contrast, in a recent
study performed on cyclometallated Pd(II) complexes of general
formula [(C,N)Pd(O,O)] (H(C,N) = azobenzene, 2-phenylpyridine
or benzo[h]quinoline; H(O,O) = acetylacetone or hexafluoroacety-
lacetone), on the bases of photophysical and cyclovoltammetric
data, we drew the conclusion that the frontier molecular orbitals
HOMO and LUMO could be mainly centered on the acetylaceto-
nato and on the cyclopalladated ligands respectively.5
aCentro di Eccellenza CEMIF.CAL, LiCryL and LASCAMM, CR-INSTM
Unita` della Calabria, Dipartimento di Chimica, Universita` della Calabria,
I-87036, Arcavacata di Rende (CS), Italy. E-mail: m.ghedini@unical.it;
Fax: +39 0984 492066; Tel: +39 0984 492062
Results and discussion
bLaMI and LASCAMM, CR-INSTM Unita` della Basilicata, Diparti-
mento di Chimica, Universita` della Basilicata, I-85100, Potenza, Italy.
E-mail: francesco.lelj@unibas.it; Fax: +39 0971 202246; Tel: +39 0971
202223
Electrochemical study
The electrochemical properties of the complexes were examined
using cyclic voltammetry (CV), and the redox data are reported
in Table 1. All of the electrochemical potentials were measured
relative to an internal ferrocene reference (Cp2Fe/Cp2Fe+). All
complexes show a single irreversible oxidation wave between +1.33
and +0.34 V. Such irreversibility has been observed by us in the
cIstituto per la Sintesi Organica e la Fotoreattivita` ISOF-CNR, Via
P. Gobetti, I-40129, Bologna, Italy. E-mail: accorsi@isof.cnr.it
† Electronic supplementary information (ESI) available: Synthetic proce-
dures and full characterizations of all studied compounds, spectroscopic
characterization in various solvents and theoretical calculations. See DOI:
10.1039/b804478c
This journal is
The Royal Society of Chemistry 2008
Dalton Trans., 2008, 4303–4318 | 4303
©

View Article Online
Chart 1
free ligands BzQ, Azo and PhPy. Pt(II) complexes are oxidized
measurements in the case of all the studied complexes. The
performed computations are able to reproduce the experimental
trend for the HOMO and LUMO energies of all the compounds.
The high electronegativity of the fluorinated groups seems to
greatly influence the HOMO energy of all complexes and, mostly
in the case of PhPy-based and BzQ-based complexes, also the
LUMO energy. The computed HOMO–LUMO gap is always
overestimated by the computations, a fact possibly associated to
the lack of solvation effects in the computations. This explanation
can be extended to the inaccuracy in the computed HOMO
energy. The LUMO energies are in better agreement with the
computations when the ancillary ligand is hfacac. This result will
be discussed below.
Fig. 2 shows the HOMO and LUMO orbitals in terms of
the AO contribution in the case of 1a, 2a, 1e, and 1d. It is
possible to see that, in all the compounds, the HOMO is mostly
localised on the cyclometallated ring, though the contribution of
the ancillary ligand is significant. A comparison of the 1a and
2a HOMOs highlights that the localisation is more evident in
Pt-based complexes than in Pd-based ones. This feature has been
found in all the studied complexes (the interested reader is referred
to the ESI† for the HOMO in terms of the AO contribution of
the remaining compounds), whereas the LUMO shows an uneven
behaviour. From Fig. 2, in 1a and 2a the LUMO is principally
localised on the cyclometallated ligand. This is found in all the
complexes with an acac ancillary ligand. In contrast, in all the
more easily than their corresponding Pd(II) analogues (ca. 200–
300 mV negative shift). The introduction of the high electroneg-
ative fluorinated groups influences the oxidation potentials of all
complexes. A typical shift of ca. 400–600 mV at higher potentials
is observed for all complexes with respect to their non-fluorinated
analogues. All complexes have shown a reduction wave between
−1.2 and −2.4 V. Complexes of 2-phenylpyridine have shown a
reversible or quasi-reversible reduction, whereas all complexes
of benzo[h]quinoline have a tendency for irreversible behaviour.
The introduction of the high electronegative CF3 groups on the
azobenzene complexes lead to the irreversibility of the reduction
wave. According to our measurements, the free ligands Azo, BzQ
and PhPy always show irreversible CV reductions. The nature of
the metal center (Pd(II) or Pt(II)) has a weak influence on reduction
potentials but slightly larger when electronegativity of the ligand is
increased (shift of ca. 10–70 mV for acetylacetonate and ca. 100–
200 mV for hexafluoroacetylacetonates). Reduction potentials are
strongly affected by the electronegativity of the ancillary ligand,
a significant positive shift of ca. 800–900 mV being observed
for the hfacac complexes 1e–f and 2e–f with respect to their
acac analogues 1b–c and 2b–c respectively. Only in the case of
azobenzene as cyclometallated ligand (1a,1d with respect to 2a,2d)
is this shift clearly less pronounced (ca. 170 mV).
Table 1 and Fig. 1 further compare the Kohn–Sham orbital
energies evaluated in vacuo against those deduced from CV
Table 1 Electrochemical and computational data for 1a–f and 2a–f
Complexes
Oxidation EOx/V a
Reduction ERed /V a
HOMO/eV e
LUMO/eV e
HOMO/eV f
LUMO/eV f
1a
1b
1c
1d
1e
1f
2a
2b
2c
2d
2e
2f
+0.82b
+0.68b
+0.55b
+1.33b
+1.18b
+0.95b
+0.65b
+0.34b
+0.38b
+0.98b
+0.67b
+0.57b
−1.33c
−2.49d
−2.20b
−1.50b
−1.63c
−1.35b
−1.40c
−2.40c
−2.19b
−1.23b
−1.49c
−1.53d
−5.68
−5.58
−5.35
−6.13
−5.95
−5.75
−5.45
−5.14
−5.18
−5.78
−5.47
−5.38
−3.74
−2.31
−2.60
−3.65
−3.17
−3.45
−3.40
−2.40
−2.60
−3.57
−3.31
−3.27
−6.34
−5.97
−5.87
−7.02
−6.78
−6.59
−6.19
−5.77
−5.69
−6.92
−6.59
−6.45
−2.74
−1.64
−1.84
−3.33
−3.11
−3.12
−2.62
−1.66
−1.86
−3.44
−3.24
−3.25
a Potentials were obtained in dry dimethyl formamide solution using tetrabutylammonium hexafluorophosphate (0.1 M) as supporting electrolyte; values
are reported relative to Cp2Fe/Cp2Fe+, using a glassy carbon working electrode vs. Ag wire as a pseudoreference electrode. b Irreversible. c Reversible.
d Quasi-reversible. e Values calculated using −4.8 eV for ferrocene. f Our computations at the MPW1PW91/SDD level of approximation.
4304 | Dalton Trans., 2008, 4303–4318
This journal is
The Royal Society of Chemistry 2008
©

View Article Online
Fig. 1 Comparison of the experimental (dashed line) and computed (solid line) HOMO and LUMO energies of the studied complexes.
Fig. 2 Computed HOMO and LUMO Kohn–Sham orbitals of complexes 1a, 2a, 1e and 1d.
complexes with hfacac (e.g. 1e in Fig. 2), the LUMO is mostly
The change in LUMO localization and energy lowering of the
LUMO is found in all the compounds based on the BzQ and PhPy
cyclometallated ligand, and agrees with the experimental strong
influence of the ancillary ligand electronegativity on the reduction
peak.
localised on the ancillary ligand. One exception has been found in
1d (Fig. 2), whose LUMO also shows a large contribution from
the cyclometallated ring (see ESI for a visual description of the
LUMO in all the complexes).
The change in LUMO localisation can be related to a definite
shift in the experimental LUMO energy when the ancillary ligand
is changed from acac to hfacac. This is particularly evident
comparing 2c and 2f. Fig. 3 shows that the LUMO and LUMO + 1
orbitals of 2c are mostly localised on the cyclometallated ring and
ancillary ligand, respectively. When the acac ligand is replaced by
hfacac (2f compound), the LUMO and LUMO + 1 energy order
is reversed. This fact can be directly explained as a consequence
of the electronegativity of the fluorine atoms, which induces a
large stabilization of all the fragment orbitals of the hfacac ligand.
In the azo-based complexes, the hfacac electronegativity is not
enough to produce such an evident change. Fig. 1 shows that the
LUMO energy is more negative in acac-based complexes 1a and
2a and less sensitive to the fluorine introduction than in the other
complexes. As mentioned above, such a LUMO is localised on the
cyclometallated ligand (as in all the acac-containing complexes),
which contains twice as many electronegative atoms (two nitrogen
atoms) than the other ligands, i.e. PhPy and BzQ. Thus, the
replacement of acac with hfacac leads to a smaller change in
the LUMO energy. In fact, in the case of 1a and 1d, we do not
This journal is
The Royal Society of Chemistry 2008
Dalton Trans., 2008, 4303–4318 | 4305
©

View Article Online
Fig. 3 The exchange of LUMO and LUMO + 1 orbitals by replacing the acac ligand with the hfacac ligand (from 2c to 2f). The HOMO and LUMO
orbitals of 2c are lowered in energy by hfacac. The LUMO + 1 orbitals of 2c undergoes a much greater energy lowering, so as to become the LUMO in
2f. A similar behaviour has been observed in all the complexes apart from azo-based complexes.
observe the sharp change in LUMO localisation passing from
acac to hfacac. This explains the lower influence of the ancillary
ligand electronegativity on the experimental reduction peak in CV
experiments of azo-based complexes.
2d. Such trends in the HOMO–LUMO gap have been reproduced
by DFT computations; furthermore, they have been evidenced by
UV-Vis spectroscopy (see below).
Another point inferred from Fig. 2 is the larger contribution of
the central metal to the HOMO rather than to the LUMO. This
finding explains the greater importance of the central metal
nature on the computed and the experimental HOMO energy in
comparison of its relevance on the LUMO.
In particular, concerning the HOMO composition, the Pt(II)
atomic orbitals (mostly a dp orbital) seems to be significantly more
important than the Pd(II) atomic orbitals in Pd-based complexes.
The fact that Pt(II) atomic orbitals show a larger contribution in
the HOMO is in line with the less negative HOMO energy in the
cases of Pt(II)-based complexes. In fact, the interaction between
the central metal and the cyclometallated ligand is anti-bonding,
thus it is a destabilising interaction.
From the above discussions, the replacement of acac with hfacac
in complexes based on BzQ and PhPy has the most dramatic effect
on the LUMO energy. As a consequence, a reduction in HOMO–
LUMO gap is induced by the fluorine electronegativity of hfacac.
The HOMO–LUMO reduction spreads in range from 0.25 eV to
0.58 eV passing from acac-based compounds 1b, 2b, 1c and 2c to
the respective hfacac-based compounds 1e, 2e, 1f and 2f. The azo-
based complexes are different; their HOMO–LUMO gap increases
by 0.54 eV passing from 1a to 1d and by 0.16 eV passing from 1d to
Photophysical properties
Absorption and low-temperature (77 K) emission spectra were
recorded for all complexes of the Pd(II) and Pt(II) series, and, in all
cases, no emission was detected at room temperature. The relevant
data are collected in Table 2. Absorption maxima, kabs, and extinc-
tion coefficients, e are similar to those of other cyclometallated
complexes reported in the literature.6 Representative absorption
spectra (those for 2d,e) are shown in Fig. 4. In order to assess the
nature of the various absorption bands, the solvatochromism of
each member of the series 1 and 2 was investigated by recording
spectra in solvents spanning a large range of Reichardt’s polarity
parameter7 (e.g., cyclohexane, ET = 0.006, dichloromethane, ET =
0.309; dimethyl formamide, ET = 0.404; methanol, ET = 0.762).
However, no substantial band shift was evidenced (Table S2 in
ESI). In particular, the high energy (k < 360 nm), more intense
absorption bands do not show any appreciable solvatochromism,
suggesting the assignment to p–p* mainly ligand-centered (1LC)
transitions. As a consequence of the ligand–metal orbital mixing,
these transitions are red-shifted with respect to what happens for
free ligand absorption. Also for the absorption bands falling in
the range 360–500 nm (with e between 2700 and 6600 M−1 cm−1,
4306 | Dalton Trans., 2008, 4303–4318
This journal is
The Royal Society of Chemistry 2008
©

View Article Online
Table 2 Photophysical results obtained in dichloromethane solution
Absorption (RT)
Emission (77 K)a
Complex
kabs /nm (e/M−1 cm−1
)
kem/nm
s/ls
b
b
1a
1b
1c
1d
1e
1f
2a
2b
2c
2d
2e
2f
250 (28300,sh), 311 (23852), 354 (18187), 410 (8850,sh), 490 (5190,sh)
259 (63168), 305 (30360,sh), 316 (33649), 360 (10700,sh)
240 (41500,sh), 297 (21608), 315 (13900,sh), 370(3250,sh), 393 (4814)
248 (16200,sh), 280 (8430,sh), 330 (16183), 365 (13200,sh), 385 (10658), 445 (4759), 480 (3230,sh)
261 (23087), 300(8900,sh), 315 (10182), 340 (8400,sh), 358 (5100,sh)
250 (54200,sh), 280 (33500,sh), 281 (36033), 315 (22700,sh), 365 (10400,sh), 390 (9200,sh)
252 (17450), 300 (10448), 351 (9855), 395 (6300,sh), 448 (4255)
249 (22810), 276 (16046), 315 (7750,sh), 330 (6900,sh), 363 (4724), 400 (2200,sh)
246 (52633), 255 (38700,sh), 319 (19949), 370 (9300,sh), 420 (7405)
246 (22274), 282 (9809), 340 (13800,sh), 364 (15478), 388 (13936), 430 (5150,sh), 502 (2705)
241 (24268), 251 (23300), 265 (19000,sh), 310 (7100,sh), 323 (7113), 366 (6585), 390 (4780,sh)
243 (55689), 290(18400,sh), 305 (19820), 320 (15300,sh), 382 (14330), 410 (11550,sh)
461
138.0
500
2130.0
b
b
462
83.0
b
b
b
b
480
7.3
500
26.0
b
b
546
552
1.7
2.9, 1.0c
a Highest energy peak, excitation performed at 337 nm for both emission spectra and lifetimes. b No emission detected. c Dual exponential emission.
the spectra. Our computations neither take into account explicitly
SO interactions, nor couple electronic and vibronic transitions,
and thus they can be particularly inaccurate in computing band
intensities in case of large geometry deformations associated
to the electronic transitions. On the other hand, inspection of
composition of the states involved in the transitions in terms
of one-electron MO transitions and analysis of the computed
harmonic vibration normal modes can give some hint on the origin
of some of these disagreements.
All the spectra are characterized by four major features of
decreasing intensity: an high energy region (H), a middle and low
energy region (M) and (L) respectively, and a tail at lower energy
(T) that extends into the visible spectrum. Fig. 5 and 6 compare
the UV-Vis spectra of 1c,1f and 2c,2f in cyclohexane solution. In
Fig. 4 Absorption spectra of 2d and 2e recorded in dichloromethane
all the reported spectra, the results of our TD-DFT computations
solution.
are included for comparison.
According to the computations, transitions that are uniquely
Table 2), which are not observed in the cyclometallating ligand
localized on the benzoquinoline ligand are pushed toward the
precursors (Table S1 in ESI), the registered absorption features in
high energy region of the spectrum, beyond band H, so that only
the solvents employed do not show the relevant solvatochromism
(see ESI). In principle, these low energy bands could be due to (i)
some of their tails fall under band H. In case of 1c, the onsets of
band M and band L are at 28000 cm−1 and 24000 cm−1 respectively.
metal-to-ligand charge transfer (1MLCT) transitions and possibly
The presence of the hfacac ligand in 1f shifts the onset of band M
include some ligand-to-ligand character (1LLCT, from metallated
to heterocontaining ligand fragments), or to (ii) perturbed ligand
localized (1LC) transitions involving the metallacycle.3b,8 However,
the solvatochromic studies did not allow clear-cut conclusions
about the nature of the low-energy absorption bands for the
complexes of both 1 (Pd(II)) and 2 (Pt(II)) series.
to higher energy by 500 cm−1. In the case of Pt(II) complexes 2c
and 2f, the hfacac ligand moves the onset of band M and band L
from 28000 cm−1 to 29000 cm−1 and from 22000 to 22500 cm−1
respectively.
Band H is due to high energy singlet states that starts from
singlet S20 and S19 for 1c and 1f respectively. In addition, the first
transitions in the case of 2c and 2f are S15 and S18. Being deep in the
UV region, band H is not very interesting as regards luminescence
properties and will not be further discussed.
Absorption spectra: the benzo[h]quinoline complexes. We will
thoroughly discuss the 1(c,f) and 2(c,f) series of molecules due to
their photophysical properties and novelty. Further details and
decomposition in terms of the Gaussian functions of this group
of molecules are reported in Table 3. All the decomposed spectra
are reproduced quite well as concerns the excitation energies even
though the number of Gaussian functions is well below the number
of computed transitions. On the other hand, the intensities do
not show the same level of agreement. In fact, in all the instances
where vibronic and spin-orbit (SO) coupling play a relevant role in
determining the transition intensities and structure of the spectra,
our computational results give only a qualitative indication of their
relative oscillatory strength especially in the low energy region of
1c. In the case of 1c, the main features of band M, as
represented by the fitting Gaussians g17, g15, g14, g13 and g11,
well match the computed transitions. g17 takes into account of S19
and S20 and half of its contribution belongs to band H. g15 includes
a group of transitions mainly represented by S16. g14 corresponds
to S12, g13 to S10 and g11 to S6. Transition toward S10 is the most
important excitation in determining the intensity of band M. Its
composition is substantially a p–p* LC transition with MLCT
character which involves charge flow toward the cyclometallated
This journal is
The Royal Society of Chemistry 2008
Dalton Trans., 2008, 4303–4318 | 4307
©

View Article Online
Table 3 Fitting functions and TD-DFT results for 1c, 1f, 2c, and 2f
Decomposition
TD-DFT
State
Function
m˜(max)/cm−1
k(max)/nm
C/cm−1
emax
m˜(max)/cm−1
f
1c
Band H
S40
S39
S38
S37
S36
S35
S34
S33
S32
S31
S30
S29
S28
S27
S26
S25
S24
S23
S22
S21
S20
46515
46511
46138
45590
45123
44783
44744
44128
43711
43310
43263
42769
42740
42207
42095
41736
41423
40599
40095
39076
38725
0.0159
0.0003
0.0510
0.3253
0.0002
g21
44126.85
227
1804.46
20592.23
<0.0001
0.0599
0.0002
0.0283
0.0998
0.0028
0.0308
0.0002
0.0502
0.0013
0.0016
0.0009
0.0550
0.0424
0.0001
0.0617
g20
g19
41226.77
40275.05
243
248
1038.27
1165.13
5907.29
8263.92
g18
g17
39435.29
37939.93
254
264
605.91
2195.35
5080.22
Band M
1223.50
S19
S18
S17
S16
S15
T20
S14
T19
S13
S12
T18
T17
S11
T16
S10
S9
T15
S8
T14
T13
S7
S6
S5
38150
37387
36704
36584
35632
35601
35524
35397
35056
34640
34225
34085
33455
33374
33088
32934
32689
32361
31915
31643
31589
31385
31049
30854
30368
30319
30077
29156
28862
0.0792
0.0001
0.0005
0.0332
0.0001
g16
g15
37260.52
36228.58
268
276
504.04
828.84
323.84
2850.49
0.0086
0.0009
0.0345
g14
34624.49
289
1110.19
4604.85
0.0022
g13
g12
33100.24
32409.51
302
309
929.09
356.60
4877.74
238.59
0.0857
<0.0001
0.0242
0.0004
0.0743
<0.0001
g11
g10
31538.73
30430.29
317
329
996.52
570.04
3807.72
1056.08
T12
S4
T11
T10
S3
0.0097
g9
g8
29448.59
29245.89
340
342
326.27
980.64
250.83
1222.51
0.0183
T9
Band L
Band T
g7
g6
g5
g4
27241.46
26656.72
25996.42
24983.34
367
375
385
400
988.32
207.81
447.17
561.35
871.14
58.99
897.34
914.11
S2
T8
S1
T7
T6
T5
T4
27267
26465
26024
26144
25657
25656
23474
0.0018
0.0471
g3
g2
24604.50
24137.28
406
414
264.12
1126.95
730.29
125.21
g1
21598.67
463
3094.94
238.80
T3
T2
T1
23163
22954
19928
1f
Band H
g20
41808.08
239
239.19
1133.97
S25
S24
41235
41037
0.0636
<0.0001
4308 | Dalton Trans., 2008, 4303–4318
This journal is
The Royal Society of Chemistry 2008
©

View Article Online
Table 3 (Contd.)
Decomposition
Function
TD-DFT
State
m˜(max)/cm−1
39885.39
39438.49
k(max)/nm
251
C/cm−1
250.72
253.56
emax
678.08
m˜(max)/cm−1
f
g19
g18
S23
S22
S21
S20
40501
40211
39516
39099
0.0359
<0.0001
0.1490
0.0005
254
1226.94
Band M
g17
37659.58
266
265.54
815.95
S19
S18
S17
S16
S15
S14
T20
T19
S13
S12
S11
T18
S10
T17
S9
39030
36602
36121
35267
34955
34567
34522
33227
32856
32631
32567
32407
31852
31671
31324
30725
30649
30453
30267
30091
29907
29622
29433
29297
29165
0.0242
0.0028
0.0365
0.0169
0.0039
0.1017
g16
g15
36538.84
35269.52
274
284
273.68
283.53
692.33
868.80
g14
33921.66
295
294.80
763.03
<0.0001
0.0007
0.0648
g13
g12
32528.23
31360.32
307
319
307.43
318.87
834.86
704.01
0.0017
0.0001
g11
30258.24
330
330.49
1253.22
T16
T15
S8
0.0027
0.0051
S7
T14
T13
S6
T12
T11
S5
<0.0001
g10
29358.94
341
340.61
919.79
0.0244
Band L
g9
g8
27707.75
26412.44
361
379
360.91
378.61
943.94
562.78
S4
S3
27235
27172
26820
26544
25924
24742
23599
22880
0.0636
0.0492
T10
T9
T8
T7
T6
S2
g7
g6
25547.52
25071.39
391
399
391.43
398.86
367.28
245.25
g5
25026.23
400
399.58
553.66
<0.0001
Band T
g4
g3
23975.14
21189.23
417
472
417.10
471.94
1001.47
2519.68
T5
T4
T3
S1
T2
T1
22360
22210
22064
21807
20187
17166
0.0090
g2
g1
18642.97
16495.68
536
606
536.40
606.22
967.96
1681.59
2c
Band H
g21
42069.23
238
1735.69
30902.20
S30
S29
S28
S27
S26
S25
S24
S23
S22
S21
S20
S19
S18
S17
S16
S15
43559
42584
42163
42093
41818
41813
41170
40345
40170
39335
39199
38864
38735
37728
37651
37448
0.2230
0.0893
0.0050
0.0005
0.0035
0.0016
0.0411
<0.0001
0.0686
<0.0001
0.0264
0.0002
0.0095
0.0209
0.0459
0.0006
g20
g19
g18
40136.09
39265.29
38563.38
249
255
259
1006.83
585.78
654.09
37115.40
9741.36
15713.47
g17
37638.21
266
931.91
11429.66
10546.51
Band M
S14
S13
S12
37271
35915
35795
0.0001
<0.0001
0.0709
g16
35834.57
279
1023.58
This journal is
The Royal Society of Chemistry 2008
Dalton Trans., 2008, 4303–4318 | 4309
©

View Article Online
Table 3 (Contd.)
Decomposition
Function
TD-DFT
State
m˜(max)/cm−1
k(max)/nm
C/cm−1
emax
m˜(max)/cm−1
f
T20
T19
S11
S10
T18
S9
T17
T16
T15
T14
S8
T13
T12
S7
T11
T10
S6
35336
34483
33834
33772
33557
33422
33249
33177
33075
32895
32796
32680
32468
32285
32258
31447
30657
30641
30120
29674
29325
g15
34734.00
288
1050.43
3052.14
0.0029
0.0048
<0.0001
g14
g13
33467.26
31660.06
299
316
1143.14
1147.65
10011.50
10428.02
0.0441
0.0285
0.0004
0.1100
g12
g11
30398.65
29876.77
329
335
910.18
485.20
8836.86
2527.24
S5
T9
T8
S4
g10
29144.61
343
532.24
7172.28
0.1190
Band L
g9
g8
g7
g6
g5
g4
28447.42
27117.96
26067.67
25774.78
25574.25
24015.91
352
369
384
388
391
416
593.67
1112.07
459.97
254.63
480.46
1089.73
3195.64
4642.38
3165.67
1176.52
1111.85
2887.97
S3
S2
T7
T6
T5
S1
T4
T3
27891
27083
28329
27855
26178
24233
24096
22936
0.0149
0.0018
0.0344
g3
22738.99
440
520.74
2138.05
Band T
g2
g1
21288.79
18708.50
470
535
424.41
7558.94
153.84
101.44
T2
T1
21645
19231
2f
Band H
g21
g20
g19
42657.04
40432.17
40171.37
234
247
249
2213.83
1123.51
372.01
431.73
285.90
40.36
S24
S23
S22
S21
S20
S19
42276
41897
40688
39270
39231
39035
0.1540
0.0121
0.0672
<0.0001
0.0704
0.0619
g18
38373.38
261
774.32
80.44
Band M
g17
36939.54
271
955.67
115.65
S18
S17
S16
S15
T20
S14
S13
S12
T19
S11
S10
T18
T17
T16
T15
T14
S9
38775
37839
37140
37109
35717
35575
35027
33982
33594
33322
33148
32950
32358
31706
31505
31064
30961
30685
30653
30526
30364
29885
0.0597
<0.0001
0.0019
0.0036
0.0002
0.0409
0.0067
g16
g15
35029.75
33803.83
285
296
1272.73
597.77
131.42
64.57
g14
32565.85
307
880.97
106.26
0.1076
0.0008
g13
g12
g11
31020.95
30185.60
29882.98
322
331
335
1500.02
328.88
835.87
99.63
3.72
17.36
0.0697
0.0127
0.0028
T13
S8
T12
S7
T11
Band L
T10
S6
T9
29295
28743
28546
<0.0001
4310 | Dalton Trans., 2008, 4303–4318
This journal is
The Royal Society of Chemistry 2008
©

View Article Online
Table 3 (Contd.)
Decomposition
Function
TD-DFT
State
m˜(max)/cm−1
k(max)/nm
C/cm−1
emax
m˜(max)/cm−1
f
g10
g9
g8
27876.60
26839.28
25788.60
359
373
388
1092.77
500.07
881.85
77.98
17.54
69.89
S5
T8
S4
T7
T6
S3
28179
27268
25980
25617
25281
25113
0.0075
0.0876
g7
24356.29
411
851.12
56.63
0.0757
Band T
g6
g5
g4
23448.70
22430.96
21813.99
426
446
458
522.91
356.72
439.00
21.76
4.02
3.80
T5
S2
T4
T3
S1
T2
T1
22498
21969
21338
19961
19484
19007
15744
<0.0001
g3
g2
g1
20944.17
19523.53
17504.14
477
512
571
762.84
654.15
5504.88
2.87
0.65
0.58
0.0176
ligand. S6 has a very similar character p–p* LC transition with
larger MLCT character which involves charge flow toward both
the ligands.
S2 modifies the pattern of the triplet state moving them toward red
to 23327, 25092, 25511, 25889 and 26326 cm−1. The analysis of the
composition of the triplet states (see Table S19 in ESI) indicates
that only T5–T7 include monoelectronic transitions with significant
metal contribution. Hence, we suggest that g2 takes into account
for T5 and g3 for the set T4–T7 respectively. This trend suggests
that triplet states up to T7 lie below S1.
Two further very low intensity Gaussians g6 (26657 cm−1)
and g10 (29449 cm−1) take into account T8 (26465 cm−1), the
triplet corresponding to S2, and T10. This last state contains
two monoelectronic excitation from HOMO − 5 and HOMO − 6
toward the ML dxy–r* LUMO + 3 (see Fig. S18 in ESI).
We can conclude that, in 1c, the more evident experimental
features can be assigned to LC transitions. Band L is principally
associated to the BzQ ligand, and band M has a similar character,
but it shows a relatively larger participation of the acac ligand.
This computational result is in line with the observed low
solvatochromism of bands M and L of the spectrum that has been
discussed before. The d–d transitions and MLCT pure transitions
are computed to be weaker bands than the p–p* LC transitions.
In the low energy region of band M, the main features are
represented by g10 and g8. The latter corresponds to an electronic
transition toward S3. It is assigned by our computations to a p–
p*-symmetry MLCT from an orbital containing both the central
metal and acac to the cyclometallated ligand and it is computed
at 29156 cm−1 (343 nm). S4 results from a d–d transition (from
−1
2
2
2
the dz * to dx *). The TD-DFT result, 30368 cm (329 nm),
−y
compares very well with the fitting g10 at 30430 cm−1 (329 nm).
Its computed oscillator strength is not as large as the LC p–p*
toward S6 and the other LC transitions we have encountered in
this complex.
Band L presents two main peaks at 25974 cm−1 (384 nm)
and 24691 cm−1 (406 nm). In the decomposed spectrum they
correspond to Gaussian functions g2-g6; g5 (25996 cm−1, 385 nm)
and g6 (27241 cm−1, 367 nm) can be easily assigned to S1
(26024 cm−1, 384 nm). A further Gaussian function (g7) can
be assigned to S2 (27267 cm−1, 367 nm). S2 is an almost pure
MLCT characterized by a HOMO − 1 to LUMO mono-electronic
transition. The broadness of the band suggests that this transition
is enhanced by vibronic coupling in spite of a very low computed
intensity. In particular from the computed vibrational spectrum,
low energy vibrations, e.g., at 39, 154, 255, 293 cm−1, involve
the central metal and the first coordination sphere. In fact, the
2c. The two main features of 2c (bottom of Fig. 5) are band M
(peak at 30864 cm−1, 322 nm) as for 1c and band L (peak at
26525 cm−1, 377 nm).
The high energy part of the band M is described by Gaussians
g13, g14 and g16, which can be assigned to the most intense
transitions due to S7, S8 and S12. In addition, the low energy
part is mostly produced by electronic transitions toward S4
(29324 cm−1) and S5 (303641 cm−1) (g10 and g12 respectively),
which substantially parallel the 1c transitions toward S6 and S10.
Gaussians g4–g8 belong to band L that, in the low energy part,
shows a similar shape to 1c, albeit of different origin. The broad
Gaussian g8 can be assigned, as in case of 1c, to S2 (27083 cm−1),
˚
puckering of the Pd(II) atom out of plane by 0.1 A, within
2RT at room temperature increases the oscillator strength of this
transition more than fivefold.
As far as it concerns g2–g4, they give rise to the 24691 cm−1
(406 nm) peak. In this case it is not possible to assign this
peak and the corresponding Gaussian functions to any of singlet
transitions. It is worth noting that using less than three Gaussian
functions prevents any attempt to achieve a satisfactory fit of
the spectrum in this region. On the other hand, the computed
triplet manifold shows a pattern which closely resemble that one
described by the set of Gaussian functions. In particular, the
triplet manifold shows four transitions, T4–T7, which have energies
(23474, 25656, 25657, 26144 and 26465 cm−1) that are close to the
fitting Gaussian functions 24137, 24604, 24983 and 26657 cm−1.
Furthermore, the same perturbation that increases the intensity of
2
a low intensity MLCT r–p* transition from the metal dz *,
HOMO − 1, to the LUMO orbital localized on BzQ with some
contribution from a tail of the S3.
The energy of S2 transition is substantially unmodified in the
two complexes, 1c and 2c, as is also shown by the difference in
the (HOMO − 1)-LUMO gap that moves the transition of only
6 nm. Also in this case metal puckering increases the intensity of
the transition by a factor of three.
This journal is
The Royal Society of Chemistry 2008
Dalton Trans., 2008, 4303–4318 | 4311
©

View Article Online
Fig. 5 Absorption spectra of 1c (top) and 2c (bottom) recorded in cyclohexane solution and TD-DFT computations. Black circles represent the
experimental data and the line the fitted spectrum by the reported Gaussian functions. The computed Sn electronic vertical transitions are reported as a
vertical line, whose length is the associated oscillator strength and, with a rhombus on top. Computed triplet transitions are reported with a vertical line
of fixed length and with a circle on top. The position of kmax of the Gaussian function is indicated by a heavy line with a triangle on top. The absorption
spectrum of 1c; the absorption spectrum of 2c.
Taking as reference 1c and the CV results it is possible to
evaluate the energy of the first transition, which is almost a pure
HOMO–LUMO mono electronic excitation. Being 0.17 eV, the
experimental difference between the HOMO–LUMO gaps of 1c
and 2c, it is fair to assume that in the latter case the S1 energy
should be moved by 1371 cm−1 toward red, i.e., to 24626 cm−1, in
good agreement with the computed S1 energy at 24233 cm−1. This
results suggests that the S1 transition, mostly a p–p*-symmetry
4312 | Dalton Trans., 2008, 4303–4318
This journal is
The Royal Society of Chemistry 2008
©

View Article Online
Fig. 6 Absorption spectra of 1f (top) and 2f (bottom) recorded in cyclohexane solution and TD-DFT computations. Black circles represent the
experimental data and the line the fitted spectrum by the reported Gaussian functions. The computed Sn electronic vertical transitions are reported as a
vertical line, whose length is the associated oscillator strength and, with a rhombus on top. Computed triplet transitions are reported with a vertical line
of fixed length and with a circle on top. The position of kmax of the Gaussian function is indicated by a heavy line with a triangle on top.
MLCT from one metal dij* to the LUMO (see Fig. 3 for a graphical
description of the orbitals), should be assigned to g4 (24016 cm−1,
416 nm).
The feature at 25907 cm−1 (386 nm), is due to the tail of g8 and g4
overlapped to a group of Gaussians g7–g5 and g3 whose origin can
be tracked back to the same SO vibronic enhanced ground state
to triplet transitions due to the computed out-of-plane vibrations
from 29 to 457 cm−1 involving the heavy atom as in case of 1c.
We can conclude that the electronic spectrum of 2c is very
similar to that of 1c: the two main bands considered have to be
mostly assigned to p–p* of the cyclometallated ring, with minor
intervention of the ancillary ligand on band M. On the other hand,
it is worth noting that S1 moves below T5.
1f. Fig. 6 shows the experimental spectrum of 1f.
Dalton Trans., 2008, 4303–4318 | 4313
This journal is
The Royal Society of Chemistry 2008
©

View Article Online
Band M shows a very intense peak at 34014 cm−1 (293 nm) and
comprises Gaussians g10–g17. All the kmax values of the fitting
Gaussian functions agree quite well with the computed transitions.
Table 3 reports the association of each Gaussian fitting function to
groups of computed transitions. In particular, excitations toward
S11, S14, S16, S17 and S19 seems to be responsible for the features
of the high energy part of this band. The computed value of S14
energy (34567 cm−1, 289 nm) is larger than the g14 kmax (33921 cm−1,
295 nm) value, and this transition appears to be the main feature
of the M band.
S11 (32567 cm−1, 307 nm) mostly results from CT from the
cyclometallated ring to hfacac, whereas S14 is a LC transition of
the cyclometallated ring. S16 and S17 have MLCT character, the
former involving both ligands whereas the latter has an intraligand
character and only one monoelectronic transition with a full
MLCT which contributes 33% (see Table S25 in ESI). S19 is an
IL transition localized on the hfacac ligand, with a small MLCT
contribution from the metal to the hfacac ligand.
The lower energy part of the band M is characterized by two
main Gaussian functions, g10 and g11. These can be associated to
a group of transitions within the singlet manifold, S5–S10, which
are vibronic-enhanced, and possibly have some contribution from
triplets T11–T17. In particular, T12 is pure metal d–r to ligand–p
vibrationally enhanced transition.
The band L shape apparently shows a high similarity to the
same band of 1c and can be fitted by five Gaussian functions,
g5–g9.
Unlike 1c, the most relevant computed transitions are singlet
S3 and S4, assigned to g8 and g9, although these last two are less
close in energy then the computed transitions. S3 is described as
a CT from the cyclometallated ring to the hfacac ligand. It is
predominantly a monoelectronic excitation from the HOMO − 1
to the LUMO. It has not a clear corresponding feature in 1c (S6 of
1c is more indicated as a reference feature); however, its location
in the low energy part of the spectrum has to be associated to
the LUMO energy stabilisation induced by the fluorine atoms and
consequent energy gap reduction between occupied orbitals and
the LUMO as shown by CV data (Table 1 and Fig. 1).
Interestingly, the S4 composition, in terms of monoelectronic
transitions, corresponds to the S1 excitation of 1c (see Table S11
in ESI). It mostly involves a mono-electronic excitation from the
HOMO to the LUMO + 1 (as shown in Fig. 3 or ESI for a graphical
description in the case of 2c). The experimental band associated
with this LC transition of the cyclometallated ring (with small
MLCT) is only marginally displaced (blue shifted of 8 nm, as
mentioned above) and shows the same shape as for 1c.
the larger separation between T7 and T8 from T9 and T10 induced
by the same perturbation.
2f. Fig. 6 shows the experimental and computed electronic
spectrum of 2f.
Band M is produced by excitations in the range S7–S18.
Excitation toward S9 and S11 give the more relevant contribution
in the low energy region. S9 is a CT from the metal and the
cyclometallated ligand to hfacac. In addition, S11 is a LC transition
on the cyclometallated ligand and analogue to the excitations
toward S4 and S5 in 1c; thus band M assignment in terms of
molecular moieties involved is akin to 1c.
As in 1f, 2f band L is due to two close-in-energy electronic
transitions toward S3 and S4 and a transition toward S5 (g10)
that in 1f belonged to band M. Also in this case, the excitation
toward S3 is a p–p*-symmetry LLCT from the cyclometallated
ring to hfacac, and excitation toward S4 is the LC transition of
the cyclometallated ring. S5 is a MLCT due to excitation from
HOMO − 3 (mainly a p dij* metal orbital) to the LUMO, which is
now localised on the hfacac ligand (Fig. 3).
Band T: a comparison between 1c, 2c, 1f and 2f. A detailed
analysis of band T in the complexes studied shows that it has many
features that can be tracked back to the effect on the electronic
structure by the substitution of the coordinating metal and the
peripheral substitution of H by F as well.
Furthermore, whereas in 1c and 2c S1 belonged to band M (even
in the case of 2c it is bathochromically shifted), this is no more true
in the case of the last two molecules 1f and 2f, where this transition
is moved to band T and becomes more evident in 2f (Fig. 6 and
inset). This result is suggested by the TD-DFT computations and
is strongly supported by the CV data.
Thereduction ofthe HOMO–LUMO gaps between 1c–f and 2c–
f, related to the discussed effect of the fluorine substitution in the
acac ligand, as evaluated by CV measurements (Table 1) amount
to 0.45 eV for the former and 0.48 eV in the latter case (Table 1).
Furthermore, the S1 state is an almost pure monoelectronic
HOMO → LUMO transition in all the studied molecules as shown
by TD-DFT computations. Hence it is possible to evaluate the
shift of the S1 in case of 1f and 2f from the S1 energies of 1c
and 2c that amount to 3629 and 3870 cm−1 respectively. These
data locate the S1 state well inside band T at 22367 (447 nm)
and 20361 cm−1 (491 nm) respectively and compare well with the
computed TD-DFT values of 21807 (459 nm) and 19486 cm−1
(513 nm). Furthermore, TD-DFT results suggest that the oscillator
strength of these transitions are much smaller in case of 1f and 2f
than in case of their non fluorinated counterpart 1c and 2c.
On the other hand, we can then suggest that band T in the case
of 1c and 2c is associated to low-absorbing triplet states (Fig. 5
and Table 3).
The lowest energy peak at 25062 cm−1 shows a behaviour akin
to that in 1c, as can be also seen by the decomposition of the
spectrum. In fact, three Gaussian functions are necessary for an
accurate description of this feature. They can be assigned to a
group of triplet states, T7–T10.
In fact, the same perturbation of the geometry of the complex
previously described shows that they gain a large contribution of
configurations characterized by monoelectronic transitions from
This band is (Fig. 5 and inset) very broad and of low intensity
and can be ascribed to T2 and T1. Excitations related to T2 involve
only 22% of contribution coming from a mono-electronic one
2
characterized by pure metal dz to dxy orbitals and 32% of a
2
metal dz (HOMO − 2) to M–BzQ p* ligand orbitals or from M–
HOMO to LUMO + 3 of p–p type whose contribution is 50%
in T2, suggesting a vibronic contribution to the intensity of these
transitions. In fact, the puckering of the Pd atom out of plane
reduces the T2 energy and adds some mono-electronic excitation
between pseudo aꢀ (r-type) and aꢀꢀ (p-type) orbitals switching on
BzQ p-ligand orbital to metal dxy (LUMO + 3).
Furthermore, the fitting function g4, giving rise to the low
energy feature of 1c band L, is split into two fitting functions
g5 and g6 in the case of 1f; this effect can also be traced back to
4314 | Dalton Trans., 2008, 4303–4318
This journal is
The Royal Society of Chemistry 2008
©

View Article Online
the SO interaction. Further contribution can be given by T3 and
T4.
This assignment is more clearly evident in case of 2c (Fig. 5
and inset) where this band is mainly due to the same Gaussian
functions assigned to T1 and T2.
In addition, in the case of 1f and 2f, the very low intensity
band T is due to a combination of low-absorbing singlets, S1
and S2, and to the three triplets T3–T5. In particular, we suggest,
by comparison with the 1c spectrum, that S1 is assigned to the
Gaussian g3 (20944 cm−1) and that g4 and g5 could be assigned to
T4 and vibronically enhanced S2 whereas g2 and g1 take account
of T1 and T2. The higher electronegativity of the hfacac ligand
leads to some changes in the band shape, but induces a drastic
change in the spectral position and nature of S1. This behaviour
can be observed even in the PhPy-based complexes (see ESI for
more information).
Phenyl pyridine and azobenzene complexes. The computational
assignment of UV-Vis spectra of PhPy-based complexes can
be traced back to the previous discussion about BzQ-based
complexes. The interested reader can achieve more information
from the ESI; here we only underline some main points of
similarity between 1b, 2b, 1e and 2e and, respectively, the analogues
1c, 1f, 2c and 2f:
(a) The low-energy part of the spectrum (k >300 nm) in
cyclohexane solution is characterised by two main absorption
bands (referred to as band L and band M as above).
(b) In 1b and 2b, band L is clearly assigned to the HOMO–
LUMO transition, which is a LC transition on the cyclometallated
ligand with small perturbation due to CT from the central metal
to the cyclometallated ligand.
(c) In 1e and 2e, band L results from an overlap of the same
electronic transition of point (b) and a new transition involving a
LLCT from the cyclometallated ring to the hfacac ligand.
(d) In 1e and 2e the HOMO–LUMO transition is a low-intensity
CT from the cyclometallated ring to the hfacac ligand which is red-
shifted in comparison to band L and visible as a long tail of the
same band.
Fig. 7 Absorption spectra of (a) 1a and (b) 1d recorded in cyclohexane
solution and TD-DFT computations. The computed electronic vertical
transition wavelengths are reported as a vertical line whose length is the
associated oscillator strength and the numeric label indicates the excited
(Sn) state.
transition on the cyclometallated ligand, with a small perturbation
of MLCT involving the same ligand as acceptor. The most intense
computed transition consists of the excitation toward S5 in 1a and
toward S6 in 1d (Fig. 7). Also in this case, the main contribution
to such a transition is a LC excitation on the cyclometallated ring.
The interested reader is referred to the ESI for a detailed
description of the computed spectra in these complexes and for the
analogues 2a and 2d. Here, we only note that the considerations
regarding 1a and 1d can be extended to 2a and 2d.
(e) Band M is mainly due to LC transitions of the cyclometal-
lated ring and the ancillary ligand. Also in this case a MLCT is
present as a perturbation.
In addition, our computations allow an easy explanation of the
lack of solvatochromism of the main features. Furthermore, a neat
change of energy and characteristics of S1 is present when acac is
replaced by hfacac.
As already pointed out, according to our computations, Azo-
based complexes do not show the neat change in LUMO energy
passing from the acac-based 1a and 2a to the hfacac-based 1d
and 2d. Accordingly, no long tails of the first intense band (band
L) have been observed when hfacac replaces acac in 1a and 2a.
This fact confirms the previously discussed assignment of band T
tail at longer wavelengths, and is in good agreement with the CV
experimental results.
Apart from this difference, the basic assignment discussed for
BzQ-based and PhPy-based complexes can be extended to the
Azo-based ones. Fig. 7 reports the UV-Vis spectra of 1a and 1d
recorded in cyclohexane. In these complexes, a low-energy band
is present that clearly corresponds to the discussed band L. In
all the Azo-based compounds, this band can be assigned to a LC
Luminescence properties. The luminescence properties, as ob-
served at 77 K in dichloromethane (Table 2), provide some useful
hints about the nature of the low-lying emitting levels. On general
grounds, it is known that for Pd(II) and Pt(II) cyclometallated
3
complexes, LC emissions observed at 77 K exhibit narrow and
resolved emission spectra and, by contrast, 3MLCT ones are
broader and less resolved.2a,3a In addition, luminescence lifetimes
are usually much longer for 3LC emissions than for 3MLCT cases.
From Table 2, one can notice that the luminescence lifetimes
of the Pd(II) complexes exhibit systematically longer lifetimes
than the analogous complexes of Pt(II). This is likely due to
a sum of different effects. Firstly, the emitting states for the
Pd(II) cases are likely to display a purer triplet character than
for the Pt(II) counterparts, owing to the difference in spin–orbit
coupling constant for the two metal centers, nPd = 1504 cm−1 vs.
nPt = 4481 cm−1.9 Secondly, for the Pt(II) complexes, the emission
3
3
includes a larger degree of MLCT (and ILCT) contributions
against 3LC contributions. In fact, as already discussed, the metal–
ligand interaction is of an antibonding character (see ESI and
the “Electrochemical study” section) and thus, a larger metal
This journal is
The Royal Society of Chemistry 2008
Dalton Trans., 2008, 4303–4318 | 4315
©

View Article Online
contribution means higher-energy orbitals and easier oxidation
processes.
Of particular relevance is the effect of the fluorine substitution
in hfacac. In fact, the main effect of the substitution is to move
to lower energy the MO strongly localized on the hfacac, so that
in this series of derivatives the LUMO is localized on this moiety
of the molecule. This is different from the acac ligand, where the
LUMO is mainly localized on the cyclometallated ligand. (Azo)-
containing compounds do not show any emission due to the pres-
ence of the diazo bond, whereas the benzoquinoline derivatives,
with the exception of the 2c complex, show comparable emission
behaviour to the already known phenylpyridine derivatives.
Furthermore, the electron-withdrawing effect of the CF3 groups
results in shorter-lived and lower-lying emitting states, exhibiting
broader spectra with respect to the non-fluorinated counterparts.
For the sake of comparison, Fig. 8 displays emission profiles for
1b, 1e, 2b, and 2e. From the figure, one sees that the electron-
withdrawing effect of the CF3 groups (1e and 2e) results in
shorter-lived and lower-lying emitting states, exhibiting broader
spectra with respect to the unfluorinated counterparts (1b and 2b,
respectively). This further suggests that the observed emission of
both Pd(II) and Pt(II) fluorinated complexes includes a larger 3CT
character than for the unfluorinated ones. In this respect, assuming
that the triplet states are mainly represented by monoelectronic
excitation from HOMO to LUMO, it is easily seen in the case of
1e from Fig. 2 and 1b from Fig. S9 (in the ESI) that the former
shows a large contribution of hfacac in LUMO whereas in the case
of 1b both HOMO and LUMO have a large contribution from the
cyclometallated moiety with almost missing contribution from
acac in the LUMO.
Experimental
Synthesis
All commercially available chemicals were purchased from Aldrich
Chemical Co. and were used without further purification. IR spec-
tra (KBr pellets) were recorded on a Perkin-Elmer Spectrum One
1
FT-IR spectrometer equipped for reflectance measurements. H-
NMR spectra were recorded on a Bruker WH-300 spectrometer
in CDCl3 solutions, with TMS as internal standard. Elemental
analyses were performed with a Perkin-Elmer 2400 analyzer
CHNS/O.
The Pd(II) complexes 1a–f were synthesised as previously
reported from their corresponding acetate-bridged complexes,4
whereas the Pt(II) analogues 2a–f were synthesised from the
dinuclear chloro-bridged cycloplatinated parents by a reported
procedure10,11 with the following modifications: the cyclometal-
lated chloride bridged dimer intermediates were prepared using
strictly one equivalent of cyclometallating ligand, and the bridge
intermediate was, after isolation, treated with Tl(acac)12 or
Na(hfacac)5 to yield 2b–c and 2e–f respectively, in good overall
yield. The previously described procedure is not successful when
using azobenzene as the cyclometallated ligand; consequently, for
the synthesis of 2a and 2d, the chloride-bridged platinum complex
intermediate was obtained via the corresponding platinum allyl.13
Fig. 8 Luminescence spectra (and lifetimes) of the indicated complexes
observed in 77 K dichloromethane, kexc = 337 nm.
Preparation of Pt azobenzene chloride-bridged dimer. Azoben-
3
zene (0.60 mmol, 0.11 g) was added to a solution of [(g -C4H7)Pt(l-
A possible explanation of the lack of emission in Azo-based
complexes might be tracked back to the large geometry deforma-
tion in the first singlet excited state. In fact, the N–N elongation,
Cl)]2 (0.30 mmol, 0.17 g) in chloroform (15 mL) The mixture was
heated at reflux for 35 h, cooled to room temperature and filtered
off; the dark-brown solid was dried under vacuum to give the
˚
◦
e.g. 0.056 A in the case of 2a, was the largest geometry deformation
pure product in 72% yield (0.18 mg). Mp. 300 C. Anal. Calcd.
evaluated by TD-DFT computations in the case of the S1 excited
state. This behavior can be explained on the ground of monoelec-
tronic excitations, ending (in all cases) mainly into the LUMO,
which shows a large N–N antibonding character (see ESI).
for C24H18N4Cl2Pt2 (%): C, 34.97; N, 6.80; H, 2.18. Found (%): C,
34.92; N, 6.84; H, 2.22. IR (KBr, cm−1): 1458, 1360, 1303, 760, 714,
578. 1H NMR data are not available because of the low solubility
in common solvents.
Preparation of Pt 2-phenylpyridine chloride-bridged dimer. 2-
Phenylpyridine (1.20 mmol, 0.50 g) in 2-ethoxyethanol (9 mL) was
added to a solution of K2PtCl4 in water (3 mL). The mixture was
heated at 80 ◦C for 48 h in an inert gas atmosphere. The green solid
obtained was filtered, washed with water (5 mL) and methanol
(15 mL) and air-dried. Yield 75% (0.96 g). Mp. 280 ◦C. Anal.
Calcd. for C22H16N2Cl2Pt2 (%): C, 34.33; N, 3.64; H, 2.80. Found
(%): C, 34.00; N, 3.35; H, 2.25. IR (KBr, cm−1): 3102, 3017, 1610,
Conclusions
A comparison between complexes of Pd(II) and Pt(II) with different
cyclometallated ligands and with acac and hfacac, by means of the
selective perturbation of different moieties of this kind of complex
together with TD-DFT computations, has allowed us to gain a
deeper insight into their electronic structures as investigated by
CV, UV-Vis emission and measurements of the decay times of the
low-energy-emitting electronic states.
1
1491, 746, 690. H NMR (300 MHz, DMSO-d6, 298 K, TMS),
4316 | Dalton Trans., 2008, 4303–4318
This journal is
The Royal Society of Chemistry 2008
©

View Article Online
ppm: 9.49 (d, J = 5.7 Hz, J(195Pt–H) = 29 Hz, 2H), 8.30–8.07 (m,
4H), 7.77 (d, J = 7.59 Hz, 2H,), 7.53–7.49 (m, 4H), 7.20–7.11 (m,
4H).
xc functional.15 The geometry optimizations were performed by
applying the Stuttgart/Dresden ECP basis set16 on Pd and Pt
atoms, and the Dunning/Huzinaga valence double-f (D95V) basis
set17 on C, N, O, F and H atoms (default “SDD” Gaussian98 key-
word). Default gradient and displacement thresholds were used for
the geometry optimization convergence criteria. All the reported
geometries are relative minima of the potential energy surface
(electronic energy in the Born–Oppenheimer approximation), as
confirmed by the analytical computation of the Hessian matrix at
the same level of approximation.
Time-dependent density functional theory (TD-DFT)18 was
applied to compute the excitation wavelengths, oscillator strengths
and associated excited state percentage composition in terms of
monoelectronic excitations between occupied and virtual Kohn–
Sham orbitals. Such percentage values were computed by doubling
the squared coefficient associated with the monoelectronic excita-
tion as printed in the Gaussian 98 output file. TD-DFT calcula-
tions were performed by using the same level of approximation
used in the structure optimization.
According to the updates to the Gaussian03 manual,19 in
Gaussian98: “the original implementation of the mPW exchange
functional was not consistent with the equations in the paper:15
the local scaling factor was applied in computing the non-local
correction. The version of mPW in Gaussian03 corrects this error,
but since the parameters in the original paper were optimized with
the incorrect functional, the correct one cannot reproduce the
original results. . ..”
Preparation of Pt benzo[h]quinoline chloride-bridged dimer.
This complex was prepared following a similar procedure to that
used for the preparation of 2-phenylpyridine chloride-bridged
dimer. Yellow solid, yield 56% (0.77 g). Mp. 280 ◦C. Anal. Calcd.
for C26H16N2Cl2Pt2 (%): C, 38.16; N, 3.42; H, 1.97. Found (%): C,
37.99; N, 3.31; H, 1.91. IR (KBr, cm−1): 3048, 1623, 1453, 741,
706. 1H NMR (300 MHz, DMSO-d6, 298 K, TMS), ppm: 9.69 (d,
with broad 195Pt satellites, J = 5.8 Hz, 2H), 8.74 (d, J = 8.1 Hz,
2H), 8.39 (d, J = 7.55 Hz, 2H), 7.97–7.86 (m, 6H), 7.81 (d, J =
8.78 Hz, 2H), 7.76–7.52 (m, 2H).
Synthesis of 2a. A typical preparation is reported for complex
2a; all analogous complexes, 2b–f, were prepared following the
same synthetic procedure. Colour, yield, melting point, elemental
analyses, IR and 1H NMR data are reported in the ESI.
A suspension of thallium acetylacetonate (0.27 mmol, 0.08 g)
in dichloromethane (5 mL) was added to a suspension of the
dinuclear chloro-bridged Pt(II) complex (0.13 mmol, 0.11 g,)
dissolved in dichloromethane (15 mL). The resulting mixture
was stirred for 180 h at room temperature. The reaction was
monitored by TLC and, after completion, the reaction mixture was
filtered through Celite and the solvent was removed under reduced
pressure. Recrystallization of the crude product from chloroform–
methanol solution afforded 2a as brown crystalline solid. Yield
70% (0.09 g).
In order to check the effect of this modification, in the case
of BzQ-based complexes the computations were also performed
with Gaussian03. Geometries were unmodified within 10−4 A. All
˚
Electrochemical study
the reported computations of S0 and excited states properties (by
TD-DFT) have also been confirmed.
Excited-state geometries were obtained by the Turbomole 5.9
suite of programs.20 The Kohn–Sham orbitals were drawn by the
program Molekel4.3.21
Cyclic voltammetry data were measured with IR compensation
using an Epsilon electrochemical analyser. The experiments were
carried out with 3 mL of a ca. 10−3 M solution of compound
at a scan rate of 100 mV s−1. Potentials were measured using
a glassy carbon working electrode, a platinum wire as counter-
electrode and a Ag wire as pseudoreference electrode. Potentials
were finally corrected versus ferrocene/ferrocenium+ by adding
ferrocene as an internal standard to the studied solution after
the experiment. Oxidations and reductions of the complexes
were observed in dry N,Nꢀ-dimethyl formamide using tetra(n-
butyl)ammonium hexafluorophosphate (0.1 M) as supporting
electrolyte, and under nitrogen atmosphere.
Acknowledgements
This research was supported by the Italian Ministero
dell’Istruzione, dell’Universita` e della Ricerca (MiUR) through
FIRB (RBNE01P4JF) and PRIN (2005035277_005) and Centro
di Eccellenza CEMIF.CAL (CLAB01TYEF) grants.
Photophysical characterization
Notes and references
UV-Vis absorption spectra were obtained with a Perkin-Elmer
lambda 900 UV/vis spectrometer. Luminescence spectra and life-
times of the complexes at 77 K were obtained with a Perkin-Elmer
LS 50B luminescence spectrometer, equipped with a Hamamatsu
R-928 photomultiplier tube. The samples were placed within
capillary tubes immersed in liquid nitrogen and the experiments
were performed by applying both a temporal delay and gating in
order to exclude fast scattering from the apparatus; the excitation
source was set at 337 nm in all cases.
1 (a) W. Paw, S. D. Cummings, M. A. Mansour, W. B. Connick, D. K.
Geiger and R. Eisenberg, Coord. Chem. Rev., 1998, 171, 125; (b) M.
Hissler, J. E. McGarrah, W. B. Connick, D. K. Geiger, S. D. Cummings
and R. Eisenberg, Coord. Chem. Rev., 2000, 208, 115; (c) A. Diez, J.
Fornies, A. Garcia, E. Lalinde and M. T. Moreno, Inorg. Chem., 2005,
44, 2443; (d) J. Dupont, C. S. Consorti and J. Spencer, Chem. Rev., 2005,
105, 2527; (e) M. Ghedini, I. Aiello, A. Crispini, A. Golemme, M. La
Deda and D. Pucci, Coord. Chem. Rev., 2006, 250, 1373; (f) K. Godula
and D. Sames, Science, 2006, 312, 67; (g) I. Omae, J. Organomet. Chem.,
2007, 692, 2608.
2 (a) J. A. G. Williams, Top. Curr. Chem., 2007, 281, 205; (b) P. Shao, Y.
Li and W. Sun, J. Phys. Chem. A, 2008, 112, 1172.
3 (a) M. Maestri, V. Balzani, C. Deuschel-Cornioley and A. Von
Zelewsky, in: Advanced Photochemistry, ed. D. Volman, G. Hammond
and D. Neckers, John Wiley & Sons, New York, 1992, pp. 1–68; (b) M.
Ghedini, I. Aiello, M. La Deda and A. Grisolia, Chem. Commun., 2003,
2198; (c) S. M. Borisov, A. S. Vasylevska, C. Krause and O. S. Wolfbeis,
Adv. Funct. Mater., 2006, 16, 1536.
Computational methods
The Gaussian98 package (revision A11)14a and Gaussian03 (revi-
sion D02)14b have been used for all the computations. Density
functional theory was applied using the MPW1PW91 hybrid
This journal is
The Royal Society of Chemistry 2008
Dalton Trans., 2008, 4303–4318 | 4317
©

View Article Online
4 (a) M. La Deda, M. Ghedini, I. Aiello, T. Pugliese, F. Barigelletti and
G. Accorsi, J. Organomet. Chem., 2005, 690, 857; (b) M. W. Cooke and
G. S. Hanan, Chem. Soc. Rev., 2007, 36, 1466.
5 T. Pugliese, N. Godbert, M. La Deda, I. Aiello and M. Ghedini, Chem.
Phys. Lett., 2005, 410, 201.
6 (a) L. Chassot, E. Mu¨ller and A. von Zelewsky, Inorg. Chem., 1984, 23,
4249; (b) M. Maestri, D. Sandrini, V. Balzani, L. Chasson, P. Jolliet and
A. von Zelewsky, Chem. Phys. Lett., 1985, 122, 375; (c) L. Chassot and
A. von Zelewsky, Inorg. Chem., 1987, 26, 2814; (d) P. Jolliet, M. Gianini,
A. von Zelewsky, G. Bernardinelli and H. Stoeckli-Evans, Inorg. Chem.,
1996, 35, 4883.
Johnson, W. Chen, M. W. Wong, J. L. Andres, C. Gonzalez, M. Head-
Gordon, E. S. Replogle and J. A. Pople, GAUSSIAN 98 (Revision
A.11), Gaussian, Inc., Pittsburgh, PA, 2001; (b) M. J. Frisch, G. W.
Trucks, H. B. Schlegel, G. E. Scuseria, M. A. Robb, J. R. Cheeseman,
J. A. Montgomery, Jr., T. Vreven, K. N. Kudin, J. C. Burant, J. M.
Millam, S. S. Iyengar, J. Tomasi, V. Barone, B. Mennucci, M. Cossi, G.
Scalmani, N. Rega, G. A. Petersson, H. Nakatsuji, M. Hada, M. Ehara,
K. Toyota, R. Fukuda, J. Hasegawa, M. Ishida, T. Nakajima, Y. Honda,
O. Kitao, H. Nakai, M. Klene, X. Li, J. E. Knox, H. P. Hratchian,
J. B. Cross, V. Bakken, C. Adamo, J. Jaramillo, R. Gomperts, R. E.
Stratmann, O. Yazyev, A. J. Austin, R. Cammi, C. Pomelli, J. Ochterski,
P. Y. Ayala, K. Morokuma, G. A. Voth, P. Salvador, J. J. Dannenberg,
V. G. Zakrzewski, S. Dapprich, A. D. Daniels, M. C. Strain, O. Farkas,
D. K. Malick, A. D. Rabuck, K. Raghavachari, J. B. Foresman, J. V.
Ortiz, Q. Cui, A. G. Baboul, S. Clifford, J. Cioslowski, B. B. Stefanov,
G. Liu, A. Liashenko, P. Piskorz, I. Komaromi, R. L. Martin, D. J.
Fox, T. Keith, M. A. Al-Laham, C. Y. Peng, A. Nanayakkara, M.
Challacombe, P. M. W. Gill, B. G. Johnson, W. Chen, M. W. Wong, C.
Gonzalez and J. A. Pople, GAUSSIAN 03 (Revision D.02), Gaussian,
Inc., Wallingford, CT, 2004.
7 C. Reichardt, in Solvents and Solvent Effects in Organic Chemistry, 2nd
edn, VCH, Weinheim, 1998.
8 (a) M. Ghedini, D. Pucci, G. Calogero and F. Barigelletti, Chem. Phys.
Lett., 1997, 297, 2198; (b) R. C. Evans, P. Douglas and C. J. Winscom,
Coord. Chem. Rev., 2006, 250, 2093.
9 M. Montalti, A. Credi, L. Prodi and M. T. Gandolfi, in Handbook of
Photochemistry, 3rd edn, CRC Press/Taylor & Francis, Boca Raton,
2006, p. 617.
10 (a) J. Brooks, Y. Babayan, S. Lamansky, P. I. Djurovich, I. Tsyba, R. Bau
and M. E. Thompson, Inorg. Chem., 2002, 41, 3055; (b) J. L. Pratihar,
B. S. Shee, P. Pattanayak, D. Patra, A. Bhattachryya, V. G. Puranik,
C. H. Hung and S. Chattopadhyay, Eur. J. Inorg. Chem., 2007, 4272.
11 C. Damm, G. Israel, T. Hegmann and C. Tschierske, J. Mater. Chem.,
2006, 16, 1808.
15 C. Adamo and V. Barone, J. Chem. Phys., 1998, 108, 664.
16 A. Nicklass, M. Dolg, H. Stoll and H. Preuss, J. Chem. Phys., 1995,
102, 8942 and references cited therein.
17 T. H. Dunning, Jr. and P. J. Hay, in Modern Theoretical Chemistry, ed.
H. F. Schaefer, III, Plenum, New York, 1976, vol. 3, p. 1–28.
18 (a) M. E. Casida, ‘Time dependent density functional response theory
for molecules’, in Recent Advances in Density Functional Methods, Pt.1,
ed. D. P. Chong, World Scientific, Singapore, 1995, vol. 1, p. 155;
(b) M. E. Casida, C. Jamorski, K. C. Casida and D. R. J. Salahub,
Chem. Phys., 1998, 108, 4439.
19 http://www.gaussian.com/g_tech/g03_rel.htm#updatedman, as ac-
cessed on April 2008.
20 R. Ahlrichs, M. Ba¨r, M. Ha¨ser, H. Horn and C. Ko¨lmel, Chem. Phys.
Lett., 1989, 162, 165; F. Furche and R. Ahlrichs, J. Chem. Phys.,
2002, 117, 7433; F. Furche and R. Ahlrichs, J. Chem. Phys., 2004, 121,
12772.
12 N. J. Thompson, J. L. Serrano, M. J. Baena and P. Espinet, Chem.–
Eur. J., 1996, 2, 214.
13 M. Ghedini, D. Pucci, A. Crispini and G. Barberio, Organometallics,
1999, 18, 2116.
14 (a) M. J. Frisch, G. W. Trucks, H. B. Schlegel, G. E. Scuseria, M. A.
Robb, J. R. Cheeseman, V. G. Zakrzewski, J. A. Montgomery, Jr., R. E.
Stratmann, J. C. Burant, S. Dapprich, J. M. Millam, A. D. Daniels,
K. N. Kudin, M. C. Strain, O. Farkas, J. Tomasi, V. Barone, M.
Cossi, R. Cammi, B. Mennucci, C. Pomelli, C. Adamo, S. Clifford,
J. Ochterski, G. A. Petersson, P. Y. Ayala, Q. Cui, K. Morokuma,
P. Salvador, J. J. Dannenberg, D. K. Malick, A. D. Rabuck, K.
Raghavachari, J. B. Foresman, J. Cioslowski, J. V. Ortiz, A. G. Baboul,
B. B. Stefanov, G. Liu, A. Liashenko, P. Piskorz, I. Komaromi, R.
Gomperts, R. L. Martin, D. J. Fox, T. Keith, M. A. Al-Laham,
C. Y. Peng, A. Nanayakkara, M. Challacombe, P. M. W. Gill, B. G.
21 P. Fluckiger, H. P. Luthi, S. Portmann and J. Weber, Molekel 4.3, Swiss
Center for Scientific Computing, Manno, Switzerland, 2000–2002; S.
Portmann and H. P. Luthi, ‘MOLEKEL: An Interactive Molecular
Graphic Tool’, Chimia, 2000, 54, 766.
4318 | Dalton Trans., 2008, 4303–4318
This journal is
The Royal Society of Chemistry 2008
©