Synthesis, electrochemistry, UV/VIS/NIR spectroelectrochemistry and ZINDO calculations of a dinuclear ruthenium complex of the tetraoxolene bridging ligand 9-phenyl-2,3,7-trihydroxy-6-fluorone

Article abstract of DOI:10.1039/b002563l

Reaction of 9-phenyl-2,3,7-trihydroxy-6-fluorone (H₃L) with [Ru(bipy)₂(H₂O)₂]²⁺ affords the complex [{Ru(bipy)₂}₂(μ-L)]⁺ (3⁺) which was isolated as its hexafluorophosphate salt; in this complex a {Ru(bipy)₂}²⁺ fragment is coordinated to each dioxolene-like terminus of the bridging ligand [L]^3-. Voltammetric experiments in MeCN show the presence of three reversible one-electron redox couples at -0.36, 0.00 and +0.53 V vs. ferrocene/ferrocenium, which means that the complex is part of a four-membered redox chain spanning the oxidation states 3⁺ to 3⁴⁺. A spectroelectrochemical study in MeCN at -30 °C reveals a complicated series of electronic spectra in the different oxidation states which include some intense transitions in the near-IR region of the spectrum. The spectra could be partly assigned with the assistance of ZINDO calculations, which show that extensive mixing between the metal-centred and bridging-ligand-centred orbitals occurs; metal-to-ligand charge-transfer, ligand-to-ligand charge- transfer and intra-ligand transitions, amongst others, could be identified.

Full text of DOI:10.1039/b002563l

View Article Online / Journal Homepage / Table of Contents for this issue
Synthesis, electrochemistry, UV/VIS/NIR spectroelectrochemistry
and ZINDO calculations of a dinuclear ruthenium complex of the
tetraoxolene bridging ligand 9-phenyl-2,3,7-trihydroxy-6-Ñuorone
Anita M. Barthram and Michael D. Ward*
School of Chemistry, University of Bristol, CantockÏs Close, Bristol, UK BS8 1T S.
E-mail: mike.ward=bristol.ac.uk
Received (in Cambridge, UK) 29th March 2000, Accepted 11th May 2000
Published on the Web 14th June 2000
Reaction of 9-phenyl-2,3,7-trihydroxy-6-Ñuorone (H L) with [Ru(bipy) (H O) ]2` a†ords the complex
3
2
2
2
[MRu(bipy) N (l-L)]` (3`) which was isolated as its hexaÑuorophosphate salt; in this complex a MRu(bipy) N2`
2 2
2
fragment is coordinated to each dioxolene-like terminus of the bridging ligand [L]3~. Voltammetric experiments in
MeCN show the presence of three reversible one-electron redox couples at [0.36, 0.00 and ]0.53 V vs.
ferrocene/ferrocenium, which means that the complex is part of a four-membered redox chain spanning the
oxidation states 3` to 34`. A spectroelectrochemical study in MeCN at [30 ¡C reveals a complicated series of
electronic spectra in the di†erent oxidation states which include some intense transitions in the near-IR region of
the spectrum. The spectra could be partly assigned with the assistance of ZINDO calculations, which show that
extensive mixing between the metal-centred and bridging-ligand-centred orbitals occurs; metal-to-ligand
charge-transfer, ligand-to-ligand charge-transfer and intra-ligand transitions, amongst others, could be identiÐed.
Complexes of the chelating dioxolene ligand series
([catecholate]2~, [semiquinonate]~ and quinone; hereafter
abbreviated as cat, sq and q, respectively) have been studied
extensively because of their electrochemical and spectroscopic
behaviour.1h3 A major focus of this interest is in the internal
charge distribution of the complexes, because extensive mixing
of metal-based and ligand-based orbitals means that oxidation
state assigments for metal and ligand are not always clear and
“redox isomerismÏ can even occur as two forms interchange.4,5
describe here the preparation, electrochemical and spectro-
scopic properties of [MRu(bipy) N L]n` (n \ 1È4), where H L
2 2
3
is 9-phenyl-2,3,7-trihydroxy-6-Ñuorone. Despite its obvious
appeal as a bis-bidentate bridging ligand there is only a single
Lever et al. have shown that in the [Ru(bipy) (OO)]n`
2
series [where “OOÏ denotes a dioxolene ligand without
specifying oxidation state; n \ 0, 1, 2] the ruthenium centre is
in the ]2 oxidation state throughout, with the di†erent oxi-
dation states arising from ligand-based redox processes
between the catecholate, semiquinonate and quinone forms.6
In the Ru(II)Èsemiquinone and Ru(II)Èquinone forms the com-
plexes display intense, low-energy Ru(II) ] OO MLCT tran-
sitions.6 We have been interested in linking several of these
[Ru(bipy) (OO)]n` units together by use of suitable bridging
2
ligands which contain two or more dioxolene binding sites
linked together.7h10 The resultant complexes display a large
number of well-separated redox processes largely centred on
the bridging ligand, whose spectroelectrochemical properties
are exceptionally rich. For example complex 1 (Fig. 1) under-
goes four reversible one-electron interconversions linking Ðve
redox states in which the bridging ligand changes in steps
from the fully reduced bis-catecholate to the fully oxidised bis-
quinone; in this case the “bis-semiquinoneÏ state (as shown in
Fig. 1) is diamagnetic due to formation of a double bond
between the “semiquinoneÏ units and adoption of a planar,
quinonoidal structure.8 Likewise the trinuclear complex 2 can
exist in seven redox states from the tris-catecholate to the tris-
quinone.10 The very strong MLCT transitions in the near-IR
region displayed by these complexes in some oxidation states
but not others makes them of potential interest as dyes for
electro-optic switching in the near-IR region of the spec-
trum.11h13
Fig. 1 Structures of complexes 1–3. Complexes 1 and 2 are drawn
with each dioxolene fragment in the semiquinone oxidation level,
which is how the complexes are isolated; complex 3 is drawn in the
fully reduced state (as isolated).
We have extended our investigations to complexes with
related bridging ligands based on two dioxolene sites, and
DOI: 10.1039/b002563l
New J. Chem., 2000, 24, 501È504
501
This journal is ( The Royal Society of Chemistry and the Centre National de la Recherche ScientiÐque 2000

View Article Online
report of its use this way, in a dinuclear complex of Ni(II).14
Although it has two dioxolene-like termini, the single carbon
atom spacer linking the two rings suggest that the behaviour
of this complex will be fundamentally di†erent from that of 1
and 2, but will have more in common with a recently-reported
bis-catecholate ligand in which a methine unit links the two
binding sites.15,16
Experimental
General details
Instrumentation used for routine spectroscopic and electro-
chemical measurements, and for low-temperature UV/VIS/
NIR spectroelectrochemical studies, has been described
earlier.17 H L was obtained commercially (Aldrich) and used
3
as received; [Ru(bipy) Cl ] É 2H O was prepared according to
2
2
2
the published method.18 ZINDO calculations were performed
on a CAChe workstation.19
Fig. 3 Cyclic and square-wave voltammograms of 3` in MeCN,
containing 0.1 M Bu NPF as base electrolyte, using a Pt-bead
4
6
working electrode and a scan rate of 0.1 V s~1.
Synthesis of [{Ru(bipy) } L] [PF ] [3(PF )]
2 2
6
6
To a suspension of [Ru(bipy) Cl ] É 2H O (0.500 g, 0.96
2
2
2
mmol) in aqueous ethanol (50 cm3, 1 : 1) was added AgNO
3
Cyclic and square-wave voltammetry of 3(PF ) in MeCN
(0.347 g, 2.04 mmol) and the mixture was heated to reÑux
for 30 min to give red solution containing
[Ru(bipy) (H O) ]2`. This was Ðltered to remove the AgCl.
H L (0.152 g, 0.48 mmol) and KOH (0.10 g, excess) were then
added and the mixture was heated to reÑux for a further 2 h.
After removal of the EtOH under reduced pressure, addition
of aqueous NH PF precipitated the complex as a dark blue
powder; this was Ðltered o† and dried. Final puriÐcation by
column chromatography on alumina using MeCN : toluene
(3 : 2, v/v) as eluent gave the product as a blue-black solid in
60% yield. ES-MS: m/z 1290 (M`), 1145 (M [ PF )`, 572
6
(Fig. 3) reveals three reversible, one-electron (*E \ 60È70
a
p
mV) processes at E \ [0.36, 0.00 and ]0.53 V vs. the
1@2
2
2
2
ferrocene/ferrocenium couple (Fc/Fc`). In addition there is a
3
completely irreversible oxidation (no return wave) at ]1.30 V
vs. Fc/Fc`, and some overlapping poorly-resolved irreversible
reductions at potentials more negative than [2 V vs. Fc/Fc`.
By analogy with the properties of related Ru(II)/dioxolene/
bipyridyl complexes such as dinuclear 1 and mononuclear
4
6
[Ru(bipy) (OO)]n`,6,8 it is likely that: (i) some or all of the
2
three reversible redox processes are centred on the bridging
6
ligand; (ii) the irreversible wave at high positive potential is a
metal-centred Ru(II)/Ru(III) process; and (iii) the irreversible
waves at high negative potential are reductions of the terminal
(M [ PF )2`.
6
Recrystallisation for elemental analysis produced a material
whose analysis was consistent with aerial oxidation to give
[32`(PF ) ], in agreement with its electrochemical behaviour
(see Results and discussion). Found: C, 50.0; H, 2.6; N, 7.8.
bipyridyl ligands. We note that the complex [MNi(CTH)N (l-
2
6 2
L)]` (CTH \ 5,7,7,12,14,14-hexamethyl-1,4,8,11-tetraazacy-
clotetradecane) shows only two redox couples in this region,
at [0.38 and ]0.40 V vs. Fc/Fc`, of which only the Ðrst was
fully reversible and was assigned to a redox-centred L3~/L2~
couple.14 The presence of three redox processes in the same
Required for C
7.8%.
H
F
N O P Ru : C, 49.4; H, 2.9; N,
59 41 12 5 2
8
2
Results and discussion
range for 3(PF ) suggests that at least one of them is formally
6
metal-centred.
Synthesis and electrochemical properties of 3(PF )
Whereas 3` is diamagnetic, the Ðrst oxidation product
6
32` (formed by treatment of 3` with ferrocenium hexa-
The complex 3(PF ), which is an intense blue-black colour,
was simply prepared by reaction of the H L with two equiva-
lents of [Ru(bipy) (H O) ]2` under basic conditions, followed
by treatment with NH PF and chromatographic puriÐ-
cation. The 1H NMR spectrum contains numerous overlap-
ping peaks in the aromatic region of the spectrum, but it is
clear from the number of environments that the complex has
twofold symmetry such that the two metal termini are equiva-
lent. Despite the apparent asymmetry of the coordinated
ligand it is easy to see how delocalistion renders the two sites
equivalent (Fig. 2). The fact that the complex is diamagnetic
conÐrms that it is in the fully reduced form with the bridging
ligand carrying a 3[ charge.
6
Ñuorophosphate) is paramagnetic, giving
a broad EPR
3
spectrum at 100 K (frozen CH Cl /thf glass) which shows an
2
2
2
2
2
4
inÑection at g \ 1.99 with an additional feature on the low-
Ðeld side at g \ 2.10 (Fig. 4). As 0.5 and then 1.0 equivalents
of ferrocenium hexaÑuorophosphate were added to the sample
this signal grew in intensity; addition of excess ferrocenium
hexaÑuorophosphate resulted in the signal diminishing in
6
Fig. 2 Delocalisation in [L]3~ which renders the two chelating sites
Fig. 4 EPR spectrum of 32` (generated by oxidation of 3` with fer-
equivalent.
rocenium hexaÑuorophosphate) at 100 K in a CH Cl /thf glass.
2
2
502
New J. Chem., 2000, 24, 501È504

View Article Online
intensity. This indicates that the doubly-oxidised complex 33`
starts to form and is diamagnetic.
Simulation of the electronic spectrum using the ZINDO
method showed that the broad transition at 714 nm (predicted
j
\ 445 nm, predicted e \ 109 000 dm3 mol~1 cm~1)¤ con-
max
tains at least 10 closely-spaced components involving all of
these orbitals [HOMO ] LUMO, HOMO [ 1 ] LUMO;
HOMO ] LUMO ] 1; HOMO ] LUMO ] 2 etc.] [see Fig.
6(a)]. The HOMO ] LUMO and related transitions are
therefore charge-transfer in origin, from a delocalised metal/
bridging ligand orbital to the bipyridyl p* levels. In simple
terms these are analogous to the near-overlapping
Ru[d(p)] ] bipy MLCT and catecholate ] bipy LLCT tran-
Spectroelectrochemical study on 3(PF )
6
We have performed a UV/VIS/NIR spectroelectrochemical
study on 3`, spanning the four oxidation levels 3` (fully
reduced, the form in which the complex is isolated) to 34` for
the dinuclear complex core, in MeCN at [30 ¡C; the results
are shown in Fig. 5 and Table 1. All processes were found to
be fully reversible, as shown by (i) the presence of clean isos-
bestic points for the overlaid spectra recorded during each
redox conversion, and (ii) the complete regeneration of the
spectrum of the starting material by re-reduction back to 3`
at the end of the experiment.
sitions of mononuclear [Ru(bipy) (cat)] which occur at ca. 620
2
and 725 nm respectively.8 In addition, transitions such as
HOMO ] LUMO ] 2, involving the lowest p* level of the
bridging ligand L, have a combination of Ru ] L MLCT and
intra-ligand p ] p* characters. In agreement with this, the
For 3`, apart from the usual p ] p* transitions of the bipy
ligands in the UV region, the most prominent feature of the
spectrum is an intense transition at 714 nm (e \ 55 000 dm3
mol~1 cm~1). The ZINDO molecular orbital calculation on
3` showed that the HOMO ([8.03 eV) contains substantial
contributions from both Ru atoms [20% total, using orbitals
of d(p) symmetry] and the four terminal oxygen atoms of the
bridging ligand [30% total, using orbitals of p(p) symmetry]
with the remainder being delocalised over the skeleton of the
bridging ligand. Just below this in energy ([8.6 to [9 eV) lie
a closely-spaced set of similar orbitals which likewise contain
contributions from both metals and bridging ligand. This sug-
gests that the d(p) orbitals on each metal and the highest-
energy Ðlled orbitals of the bridging ligand have strongly
mixed to the extent that a localised description of the orbitals
as metal-centred or ligand-centred is unsatisfactory. Both the
LUMO of 3` ([2.57 eV), as well as LUMO ] 1 ([2.50 eV)
are simply bipy-centred p* orbitals, as expected; however at
about the same energy is the lowest p* level of the bridging
ligand L3~ (LUMO ] 2, [2.40 eV).
spectroscopically simpler complex [MNi(CTH)N (l-L)]` has a
2
strong transition at 588 nm which was assigned as a p ] p*
transition of the bridging ligand, and such a transition is one
component of this manifold for 3(PF ).
6
On oxidation to 32`, there are two signiÐcant changes: the
transition at 714 nm in 3` is shifted to lower energy (882 nm,
with a shoulder at 748 nm), and a new transition appears at
ca. 1650 nm (e \ 12 000 dm3 mol~1 cm~1). If we try to explain
the behaviour in simple terms by analogy with the model
compound series [Ru(bipy)(OO)]n`, we expect that a formally
ligand-centred oxidation will result in a low-lying SOMO just
above the metal-centred d(p) levels Mcf. oxidation of
[Ru(bipy) (cat)] to [Ru(bipy) (sq)]`N, resulting in a low-
2
2
energy Ru ] sq MLCT transition which occurs at 880 nm in
[Ru(bipy) (sq)]`.6,8 On this basis we could reasonably assign
2
the 882 nm transition of 32` to an MLCT transition from Ru
to the bridging ligand SOMO, subject to the caveats outlined
Fig. 5 Electronic spectra of (i) 3`, (ii) 32`, (iii) 33` and (iv) 34` from
the spectroelectrochemical experiment (MeCN, [30 ¡C).
Fig. 6 Predicted electronic spectra (using the ZINDO method) of the
diamagnetic complexes (a) 3` and (b) 33`, showing the positions of
the individual components of each absorption envelope.
Table 1 Electronic spectra of complex 3 in its di†erent oxidation
states (MeCN, [30 ¡C)
Complex
3`
j
/nm (10~3e/dm3 mol~1 cm~1)
max
¤ The electronic spectral maxima as predicted by the ZINDO method
for this series of complexes are consistently too high in energy (e.g. for
3`, the 714 nm transition is predicted at 445 nm; for 33`, the 1237 nm
transition is predicted at 1010 nm). However the numbers, relative
energy ordering and approximate intensities of the transitions are
consistent with what was observed, and the ZINDO calculations
therefore give a reasonable qualitative picture of the behaviour of the
complexes.
714(55), 582(sh), 407(23), 354(24), 294(120),
245(99)
1655(12), 882(36), 748(21), 517(24), 427(24),
292(150), 245(99)
1237(41), 1137(sh), 765(18), 545(sh), 465(26),
346(sh), 290(130), 244(100)
1100(sh), 775(54), 406(20), 285(87), 246(84)
32`
33`
34`
New J. Chem., 2000, 24, 501È504
503

View Article Online
above regarding the limited accuracy of this localised descrip-
tion. The higher-energy transitions will include the
Ru[d(p)] ] bipy MLCT transitions which are still expected.
There are various possibilities for the broad near-IR tran-
sition at ca. 1650 nm. Its position and intensity are consistent
with an inter-valence charge-transfer, either (i) from Ru(II) to
Ru(III) following a metal-centred oxidation,20 or (ii) from an
electron-rich to an electron-poor end of the bridging ligand
following a ligand-centred oxidation which is localised at one
Conclusion
The new complex 3` has rich electrochemical behaviour,
being the Ðrst member of a four-membered redox series which
spans 3` to 34`. A UV/VIS/NIR spectroelectrochemical
study showed the presence of intense transitions in the visible
and near-IR region of the spectrum in all four oxidation
states. These could be assigned to a limited extent with the
help of ZINDO calculations on the complex in its dia-
magnetic oxidation states, which indicated that strong mixing
between metal and ligand-centred orbitals results in delocal-
isation of the frontier orbitals, such that simple assignment of
individual redox processes as metal- or ligand-based is not
appropriate. The intense transitions in the near-IR region, and
their strong variation with oxidation state over a four-
membered redox series, make this compound of interest as an
electrochromic near-IR dye.10h13
end.21,22 In addition the spectrum of [MNi(CTH)N (l-L)]2`, in
2
which the bridging ligand has been oxidised to (L2~)~, shows a
transition at 1660 nm (e B 3000 dm3 mol~1 cm~1) which was
assigned to an intra-ligand transition involving the SOMO. It
is likely that the true situation is a combination of these, with
the low-energy transition being an “end-to-endÏ charge-
transfer from the non-oxidised Ru/dioxolene fragment to the
oxidised one; it will therefore have both intra-ligand and
charge-transfer character.
Acknowledgements
We thank the EPSRC for a project studentship (A.M.B.).
Further oxidation to 33` results in the further red-shift of
the ““Ru(II) ] LÏÏ MLCT transition to 1237 nm (e \ 41 000
dm3 mol~1 cm~1), and the reduction in intensity of the
weaker transition at 1650 nm that was apparent for 32`. Since
this oxidation state is diamagnetic we could again carry out a
ZINDO calculation to determine the molecular orbitals and
help assign the electronic spectrum. The intense 1237 nm tran-
References
1
2
3
C. G. Pierpont and C. W. Lange, Prog. Inorg. Chem., 1994, 41,
331.
A. B. P. Lever, H. Masui, R. A. Metcalfe, D. J. Stufkens, E. S.
Dodsworth and P. R. Auburn, Coord. Chem. Rev., 1993, 125, 317.
S. I. Gorelsky, E. S. Dodsworth, A. B. P. Lever and A. A. Vlcek,
Coord. Chem. Rev., 1998, 174, 469.
sition (predicted j \ 1010 nm; predicted e \ 140 000 dm3
max
mol~1 cm~1)¤ is the HOMO ([11.70 eV) ] LUMO ([8.83
eV) transition, and we can draw an analogy with the mono-
4
5
A. S. Attia and C. G. Pierpont, Inorg. Chem., 1998, 37, 3051.
O. S. Jung, D. H. Jo, Y. A. Lee, B. J. Conklin and C. G. Pierpont,
Inorg. Chem., 1997, 36, 19.
nuclear complex [Ru(bipy) (q)]2` in which there is a Ru ] q
2
MLCT transition into the low-lying orbital of the quinone
6
M. Haga, E. S. Dodsworth and A. B. P. Lever, Inorg. Chem.,
1986, 25, 447.
which is now empty following two ligand-centred oxidations.6
However examination of the form of the HOMO and LUMO
of 33` shows that this transition has very little MLCT charac-
ter, as these frontier orbitals contain comparable contribu-
tions from metal and bridging ligand. The principal
contributions to the HOMO are 36% from the Ru d(p)
orbitals, with the remaining 64% evenly distibuted across the
bridging ligand; for the LUMO the relative contributions are
25% from the metals and 75% from the bridging ligand. Thus
the HOMO ] LUMO transition has only slight MLCT char-
acter; it is more accurately a p ] p* transition between two
“metal ] bridging ligandÏ delocalised orbitals. We note also
the presence of residual absorbance at around 1650 nm in the
spectrum of 33` which is not accounted for by the ZINDO
calculation [Fig. 6(b)]; the cause of this is not clear at the
moment.
Finally, oxidation to 34` results in replacement of the
1237 nm transition of 33` by one at much higher energy, at
775 nm (e \ 54 000 dm3 mol~1 cm~1) with a low-energy
shoulder at ca. 1100 nm. Assuming that the electron is
removed from the HOMO of 33` this is a highly delocalised
transition which cannot be simply described as metal- or
ligand-centred. By analogy with the spectrum of 32` we
assume that the relatively weak, lowest-energy transition at
1100 nm involves the singly-occupied orbital, and that the
much more intense transition at 775 nm is a charge-transfer
transition of some sort.
7
8
M. D. Ward, Inorg. Chem., 1996, 35, 1712.
L. F. Joulie, E. Schatz, M. D. Ward, F. Weber and L. J. Yellow-
lees, J. Chem. Soc., Dalton T rans., 1994, 799.
9
A. M. Barthram, R. L. Cleary, J. C. Je†ery, S. M. Couchman and
M. D. Ward, Inorg. Chim. Acta, 1998, 267, 1.
10 A. M. Barthram, R. L. Cleary, R. Kowallick and M. D. Ward,
Chem. Commun., 1998, 2695.
11 J. Fabian, H. Nakazumi and M. Matsuoka, Chem. Rev., 1992, 92,
1197.
12 (a) R. J. Mortimer, Chem. Soc. Rev., 1997, 26, 147; (b) R. J. Morti-
mer, Electrochim. Acta, 1999, 44, 2971.
13 N. C. Harden, E. R. Humphrey, J. C. Je†ery, S.-M. Lee, M. Mar-
caccio, J. A. McCleverty, L. H. Rees and M. D. Ward, J. Chem.
Soc., Dalton T rans., 1999, 2417.
14 A. Dei, D. Gatteschi and L. Pardi, Inorg. Chim. Acta, 1991, 189,
125.
15 D. A. Shultz, S. H. Bodnar, R. K. Kumar and J. W. Kampf, J.
Am. Chem. Soc., 1999, 121, 10664.
16 D. A. Shultz and S. H. Bodnar, Inorg. Chem., 1999, 38, 591.
17 S.-M. Lee, R. Kowallick, M. Marcaccio, J. A. McCleverty and
M. D Ward, J. Chem. Soc., Dalton T rans., 1998, 3443.
18 B. P. Sullivan, D. J. Salmon and T. J. Meyer, Inorg. Chem., 1978,
17, 3334.
19 ZINDO version 4.0.2 in the CAChe system, Oxford Molecular,
Oxford, 1998.
20 M. D. Ward, Chem. Soc. Rev., 1995, 121.
21 N. C. Fletcher, T. C. Robinson, A. Behrendt, J. C. Je†ery, Z. R.
Reeves and M. D. Ward, J. Chem. Soc., Dalton T rans., 1999,
2999.
22 P. R. Auburn and A. B. P. Lever, Inorg. Chem., 1990, 29, 2551.
504
New J. Chem., 2000, 24, 501È504