Mono- and Dinuclear Ruthenium Carbonyl Complexes with Redox-Active Dioxolene Ligands: Electrochemical and Spectroscopic Studies and the Properties of the Mixed-Valence Complexes

Article abstract of DOI:10.1021/ic034579o

The mononuclear complex [Ru(PPh₃)₂(CO) ₂(L¹)] (1; H₂L¹ = 7,8-dihydroxy-6-methoxycoumarin) and the dinuclear complexes [{Ru(PPh ₃)₂(CO)₂}₂(L₂)][PF ₆] {[2][PF₆]; H₃L² = 9-phenyl-2,3,7-trihydroxy-6-fluorone} and [{Ru(PBu₃) ₂-(CO)₂}₂(L³)] (3; H ₄L³ = 1,2,3,5,6,7-hexahydroxyanthracene-9,10-dione) have been prepared; all complexes contain one or two trans,cis-{Ru(PR ₃)₂(CO)₂} units, each connected to a chelating dioxolene-type ligand. In all cases the dioxolene ligands exhibit reversible redox activity, and accordingly the complexes were studied by electrochemistry and UV/vis/NIR, IR, and EPR spectroscopy in their accessible oxidation states. Oxidation of 1 to [1]⁺ generates a ligand-centered semiquinone radical with some metal character as shown by the IR and EPR spectra. Dinuclear complexes [2]⁺ and 3 show two reversible ligand-centered couples (one associated with each dioxolene terminus) which are separated by 690 and 440 mV, respectively. This indicates that the mixed-valence species [2] ²⁺ has greater degree of electronic delocalization between the ligand termini than does [3]⁺, an observation which was supported by IR, EPR, and UV/vis/NIR spectroelectrochemistry. Both [2]²⁺ and [3]⁺ have a solution EPR spectrum consistent with full delocalization of the unpaired electron between the ligand termini on the EPR time scale (a quintet arising from equal coupling to all four ³¹P nuclei); [3]⁺ is localized on the faster IR time scale (four CO vibrations rather than two, indicative of inequivalent {Ru(CO)₂} units) whereas [2]²⁺ is fully delocalized (two CO vibrations). UV/vis/NIR spectroelectrochemistry revealed the presence of a narrow, low-energy (2695 nm) transition for [3]⁺ associated with the catecholate → semiquinone intervalence transition. The narrowness and solvent-independence of this transition (characteristic of class III mixed-valence character) coupled with evidence for inequivalent {Ru(CO) ₂} termini in the mixed-valence state (characteristic of class II character) place this complex at the class II-III borderline, in contrast to [2]²⁺ which is clearly class III.

Full text of DOI:10.1021/ic034579o

Inorg. Chem. 2003, 42, 7887−7896
Mono- and Dinuclear Ruthenium Carbonyl Complexes with Redox-Active
Dioxolene Ligands: Electrochemical and Spectroscopic Studies and the
Properties of the Mixed-Valence Complexes
,§
Andrew P. Meacham,† Kathryn L. Druce,† Zo1e R. Bell,† Michael D. Ward,*,†,‡ Jerome B. Keister,* and
,|
A. B. P. Lever*
School of Chemistry, UniVersity of Bristol, Cantock’s Close, Bristol BS8 1TS, U.K.,
Department of Chemistry, UniVersity at Buffalo, State UniVersity of New York,
Buffalo, New York 14260, and Department of Chemistry, York UniVersity,
4700 Keele Street, Toronto, Ontario, M3J 1P3, Canada
Received May 28, 2003
The mononuclear complex [Ru(PPh3)2(CO)2(L1)] (1; H L1 ) 7,8-dihydroxy-6-methoxycoumarin) and the dinuclear
2
complexes [{Ru(PPh3)2(CO)2}2(L2)][PF ] {[2][PF ]; H L2 ) 9-phenyl-2,3,7-trihydroxy-6-fluorone} and [{Ru(PBu3)2-
6
6
3
(CO)2}2(L3)] (3; H L3 ) 1,2,3,5,6,7-hexahydroxyanthracene-9,10-dione) have been prepared; all complexes contain
4
one or two trans,cis-{Ru(PR ) (CO)2} units, each connected to a chelating dioxolene-type ligand. In all cases the
3 2
dioxolene ligands exhibit reversible redox activity, and accordingly the complexes were studied by electrochemistry
and UV/vis/NIR, IR, and EPR spectroscopy in their accessible oxidation states. Oxidation of 1 to [1]+ generates a
ligand-centered semiquinone radical with some metal character as shown by the IR and EPR spectra. Dinuclear
complexes [2]+ and 3 show two reversible ligand-centered couples (one associated with each dioxolene terminus)
which are separated by 690 and 440 mV, respectively. This indicates that the mixed-valence species [2]2+ has
greater degree of electronic delocalization between the ligand termini than does [3]+, an observation which was
supported by IR, EPR, and UV/vis/NIR spectroelectrochemistry. Both [2]2+ and [3]+ have a solution EPR spectrum
consistent with full delocalization of the unpaired electron between the ligand termini on the EPR time scale (a
quintet arising from equal coupling to all four 31P nuclei); [3]+ is localized on the faster IR time scale (four CO
vibrations rather than two, indicative of inequivalent {Ru(CO)2} units) whereas [2]2+ is fully delocalized (two CO
vibrations). UV/vis/NIR spectroelectrochemistry revealed the presence of a narrow, low-energy (2695 nm) transition
for [3]+ associated with the catecholate f semiquinone intervalence transition. The narrowness and solvent-
independence of this transition (characteristic of class III mixed-valence character) coupled with evidence for
inequivalent {Ru(CO)2} termini in the mixed-valence state (characteristic of class II character) place this complex
at the class II−III borderline, in contrast to [2]2+ which is clearly class III.
Introduction
behavior2 which can lead to unusual properties for their
complexes. The presence of ligand-based redox activity in
the complexessas exemplified by the catecholate (cat)/
semiquinone (sq)/quinone (q) redox seriesscombined with,
in many cases, metal-based redox activity means that
dioxolene complexes can have exceptionally rich redox and
spectroscopic behavior. For example, in Pierpont’s “redox
isomers” based on first-row transition-metal complexes, the
proximity of metal-based and ligand-based redox orbitals
Chelating 1,2-dioxolenes1 have attracted much attention
as ligands recently, in part because of their noninnocent
* Authors to whom correspondence should be addressed. Fax: (+44)
117 9290509 (M.D.W.). E-mail: m.d.ward@sheffield.ac.uk (M.D.W.);
blever@yorku.ca (A.B.P.L.); keister@buffalo.edu (J.B.K.).
† University of Bristol.
‡ Present address: Department of Chemistry, University of Sheffield,
Sheffield, U.K., S3 7HF.
§ University of Buffalo.
| University of York.
(2) Ward, M. D.; McCleverty, J. A. J. Chem. Soc., Dalton Trans. 2002,
(1) Pierpont, C. G.; Lange, C. W. Prog. Inorg. Chem. 1994, 41, 331.
275.
10.1021/ic034579o CCC: $25.00 © 2003 American Chemical Society
Published on Web 11/05/2003
Inorganic Chemistry, Vol. 42, No. 24, 2003 7887

Meacham et al.
Scheme 1. Structural Formulas of the New Complexes
are expected to lead to extensive ligand-centered redox
activity. Further, the combination of intense colors and
multiple oxidation states suggests the possibility of using
the complexes as switchable dyes.
Although examples of dinuclear complexes with bis-
dioxolene bridging ligands based on {Ru(bpy)2}2+ and {Os-
(bpy)2}2+ termini are known,5 organometallic analogues of
the type we describe here have received scant attention.
Inclusion of the CO and phosphine units (instead of, for
example, bpy units) as coligands facilitates the spectroscopic
studies of the complexes in their different oxidation states.
The CO stretches in the IR spectra provide a convenient
method for assessing changes in electron density at each
metal center by IR spectroelectrochemistry, and the spin-
active 31P nuclei result in simple hyperfine coupling patterns
in the EPR spectra of the paramagnetic oxidation states which
also affords useful information. Accordingly we describe here
IR, EPR, and UV/vis/NIR spectroelectrochemical studies on
the complexes which provide an interesting picture of how
the different structures of the dinucleating bridging ligands
[L2]3- and [L3]4- result in different electron-delocalization
properties in their mixed-valence states.
allows for an unusual type of switching behavior.3 In
addition, complexes of dioxolenes containing Ru(II) or Os(II)
termini with polypyridine or ammine coligands have been
subjected to extensive theoretical and spectroscopic studies
which have provided insights into the charge distribution;4
many of these complexes also show pronounced near-IR
electrochromism arising from metal-to-dioxolene charge-
transfer transitions.5
In this paper we describe the synthesis, redox, and
spectroelectrochemical properties of some organometallic
ruthenium complexes containing {Ru(PR3)2(CO)2} units
attached to mononucleating and bridging dioxolene ligands,
as shown in Scheme 1. Mononuclear complex 1 is based on
7,8-dihydroxy-6-methoxycoumarin (H2L1), dinuclear com-
plex [2][PF6] is based on 9-phenyl-2,3,7-trihydroxy-6-
fluorone (H3L2), and dinuclear complex 3 is based on
1,2,3,5,6,7-hexahydroxyanthracene-9,10-dione (H4L3, also
known as rufigallol). A variety of mononuclear complexes
of the type [Ru(PR3)2(CO)2(OO)] (where “OO” denotes a
generic catecholate-type ligand) have been investigated, and
show reversible ligand-centered redox activity;6,7 we are now
extending this area to the study of dinuclear analogues. In
complexes of [L2]3- and [L3]4-, which have two dioxolene-
like termini, redox processes associated with each terminus
Results and Discussion
Synthesis and Characterization. The complexes were
readily prepared by reaction of Ru3(CO)12, the appropriate
phosphine, and the dioxolene ligand in refluxing toluene
which afforded a deeply colored solution in each case.
Complexes 1 and 3 are neutral and were isolated by
chromatography of the crude reaction mixture; complex [2]-
[PF6] is a monocation (due to the presence of only three
dissociable protons in H3L2) and was accordingly isolated
as its hexafluorophosphate salt. In all cases, two CO stretches
were observed in the IR spectrum, confirming their mutual
cis orientation in approximate local C2V symmetry. A single
signal in the 31P NMR spectrum in every case is also
consistent with the trans-(PR3)2, cis-(CO)2 arrangement of
ancillary ligands which occurs in other mononuclear com-
plexes of this type.7 The formulations of the complexes were
further confirmed by mass spectrometry and elemental
analyses. Although complex [2][PF6] appears to be asym-
metric, with inequivalent dioxolene-like binding sites (car-
rying 2- and 1- charges), the picture in Scheme 1 represents
only one extreme canonical form and delocalization renders
the two termini equivalent,5c as shown by the NMR and IR
spectra. Complex 3 was easily characterized by comparison
of spectroscopic data to the 1,2,3-trihydroxyanthraquinone
monometallic analogue;6a an alternative 1,2- and 5,6-complex
geometry can be excluded because of the low-field shift of
the hydroxyl resonance, which indicates hydrogen-bonding
to the central quinone carbonyl.
(3) (a) Pierpont, C. G. Inorg. Chem. 2001, 40, 5727. (b) Pierpont, C. G.
Coord. Chem. ReV. 2001, 219, 415. (c) Pierpont, C. G. Coord. Chem.
ReV. 2001, 216, 99. (d) Attia, A. S.; Pierpont, C. G. Inorg. Chem.
1998, 37, 3051. (e) Jung, O. S.; Jo, D. H.; Lee, Y. A.; Conklin, B. J.;
Pierpont, C. G. Inorg. Chem. 1997, 36, 19.
(4) (a) Gorelsky, S. I.; Dodsworth, E. S.; Lever, A. B. P.; Vlcek, A. A.
Coord. Chem. ReV. 1988, 174, 469. (b) da Silva, R. S.; Gorelsky, S.
I.; Dodsworth, E. S.; Tfouni, E.; Lever, A. B. P. J. Chem. Soc., Dalton
Trans. 2000, 4078. (c) Lever, A. B. P.; Masui, H.; Metcalfe, R. A.;
Stufkens, D. J.; Dodsworth, E. S.; Auburn, P. R, Coord. Chem. ReV.
1993, 125, 317. (d) Venegas-Yazigi, D.; Mirza, H.; Lever, A. B. P.;
Lough, A. J.; Costamanga, J.; Vega, A.; LaTorre, R. Inorg. Chem.
2000, 39, 141.
(5) (a) Joulie´, L. F.; Schatz, E.; Ward, M. D.; Weber, F.; Yellowlees, L.
J. J. Chem. Soc., Dalton Trans. 1994, 799. (b) Barthram, A. M.; Cleary,
R. L.; Kowallick, R.; Ward, M. D. Chem. Commun. 1998, 2695. (c)
Barthram, A. M.; Ward, M. D. New J. Chem. 2000, 24, 501. (d)
Barthram, A. M.; Reeves, Z. R.; Jeffery, J. C.; Ward, M. D. J. Chem.
Soc., Dalton Trans. 2000, 3162.
(6) (a) Churchill, M. R.; Keil, K. M.; Gilmartin, B. P.; Schuster, J. J.;
Keister, J. B.; Janik, T. S. Inorg. Chem. 2001, 40, 4361. (b) Churchill,
M. R.; Keil, K. M.; Bright, F. V.; Pandey, S.; Baker, G. A.; Keister,
J. B. Inorg. Chem. 2000, 39, 5807.
(7) (a) Connelly, N. G.; Manners, I.; Protheroe, J. R. C.; Whiteley, M.
W. J. Chem. Soc., Dalton Trans. 1984, 2713. (b) Girgis, A. Y.; Sohn,
Y. S.; Balch, A. L. Inorg. Chem. 1975, 14, 2327.
7888 Inorganic Chemistry, Vol. 42, No. 24, 2003

Ruthenium Carbonyl Complexes with Dioxolene Ligands
+0.76 V vs ferrocene/ferrocenium. A third process at +1.52
V, near the limit of the solvent window, was detected by
square-wave voltammetry, but it did not appear to be
reversible by cyclic voltammetry. Again, we assign these as
ligand-centered, with the first two being analogous to the
cat/sq couple of 1, with one being associated with each
binding site. It is perhaps not strictly accurate to call these
“cat/sq” processes, since the binding sites in [2]+ are not
wholly catecholate-like in that they carry a charge of -1.5
(rather than -2) and have a formal C-O bond order of
greater than 1 due to the contribution of a single CdO
component (Scheme 1). However [L2]3- may be considered
as catecholate-like because it is in its fully reduced,
diamagnetic form, with the two reversible processes resulting
first in a semiquinone-like ligand radical species, and then a
diamagnetic quinone-like form (Scheme 2). The third couple
we assume to be also ligand-centered; the related complex
[{Ru(bpy)2}2(µ-L2)]+ also underwent three ligand-centered
redox processes which were all chemically reversible.5c An
important question relating to complex [2]+ is whether the
mono-oxidized radical form [2]2+, which may be considered
as a ligand-centered mixed-valence species, is fully delo-
calized (resulting in equivalent Ru centers) or is localized
(resulting in spectroscopically distinct Ru centers); this will
be addressed later. The separation of 690 mV between the
first two redox processes means that the comproportionation
constant for the mixed-valence state, Kc, is 7 × 1011, which
is in the domain indicative of strongly interacting, class III
(fully delocalized) mixed-valence states.9
In complex 3 the two binding termini of the bridging
ligand are catecholate-like units, and accordingly (by analogy
with 1) we may expect a reversible cat/sq couple and an
irreversible sq/q couple associated with each of the two
metal-dioxolene fragments. In CH2Cl2 the cyclic voltam-
mogram shows two reversible one-electron processes at
-0.07 and +0.37 V vs ferrocene/ferrocenium, which we
assign as cat/sq couples associated with the two Ru-
catecholate substructures (the two sq/q couples are not
apparent, presumably occurring at potentials beyond the limit
of the solvent). The separation of 440 mV between these
couples indicates a strong electronic interaction between the
two sites (Kc for the mixed-valence state is 4 × 107), although
the interaction is significantly weaker than that observed in
complex [2]+, indicative of a reduced degree of delocalization
of the unpaired electron in the mixed-valence state of 3.
IR and EPR Studies. The IR and EPR spectroscopic
properties of 1 in its neutral and oxidized forms provide a
useful reference for comparison with the mixed-valence
dinuclear complexes [2]2+ and [3]+. Complex 1 could be
oxidized to [1]+, either chemically (using ferrocenium
hexafluorophosphate) or electrochemically (in a KBr OTTLE
cell). On oxidation, the two CO stretching vibrations of the
carbonyl ligands, which start at 1978 and 2042 cm-1, move
to 2018 and 2072 cm-1 respectively, an average shift of 35
cm-1 to higher energy which is consistent with a slight
Figure 1. Structure of complex 1.
The crystal structures of complexes 1 and [2][PF6]‚3CH2-
Cl2 have been determined; see Figures 1 and 2 and Tables
1-3. For 1 (Figure 1), the trans-(PR3)2, cis-(CO)2 arrange-
ment of ligands is apparent, and the O(12)-C(12) and
O(21)-C(21) separations of 1.34 Å each are in agreement
with a fully reduced catecholate ligand coordinated to a Ru-
(II) center.8
In the structure of complex [2][PF6]‚3CH2Cl2, the coor-
dination environment around each of the metal centers of
the complex cation is essentially the same as in 1 (compare
Tables 2 and 3). One significant difference, however, is that
the C-O distances in the bridging ligand associated with
the four donor atoms O(11), O(24), O(17), and O(18) are
uniformly shorter in [2]+ (average, 1.302 Å) than the
corresponding C-O distances in 1 (1.34 Å). This arises from
the partial double bond character due to the presence of
(formally) one double and three single C-O bonds which
are rendered equivalent by delocalization; there is no
localization of single and double C-O bonds on the basis
of the structural data. There is disorder in the bridging ligand
involving the position of the phenyl substituent [C(141)-
C(146)] and the oxygen atom O(21) which is discussed in
more detail in the Experimental Section.
Electrochemical Studies. The cyclic voltammogram of
1 in CH2Cl2 (see Table 4) showed a reversible, one-electron
couple at -0.17 V vs ferrocenium/ferrocene and a second
process at +0.86 V which is irreversible (the return wave is
of lower intensity than the outward wave, and there is a small
product wave at +0.52 V). Square-wave voltammetry
showed the two processes to be of equal intensity, and by
analogy with the behavior of previously described complexes
of this nature,7 these are assigned as successive cat/sq and
sq/q processes centered on the dioxolene ligand; the separa-
tion of about 1 V between the two couples and the
irreversibility of the second couple are both characteristic
of complexes of this type, as are the redox potentials.7
The dinuclear complex [2][PF6] displayed under the same
conditions two reversible one-electron couples at +0.07 and
(8) (a) Pierpont, C. G.; Buchanan, R. M. Coord. Chem. ReV. 1981, 38,
45. (b) Haga, M.; Dodsworth, E. S.; Lever, A. B. P.; Boone, S. R.;
Pierpont, C. G. J. Am. Chem. Soc. 1986, 108, 7413. (c) Boone, S. R.;
Pierpont, C. G. Inorg. Chem. 1987, 26, 1796. (d) Haga, M.; Isobe,
K.; Boone, S. R.; Pierpont, C. G. Inorg. Chem. 1990, 29, 3795.
(9) (a) Ward, M. D. Chem. Soc. ReV. 1995, 121. (b) Creutz, C. Prog.
Inorg. Chem. 1983, 30, 1. (c) Hush, N. S. Coord. Chem. ReV. 1985,
64, 135.
Inorganic Chemistry, Vol. 42, No. 24, 2003 7889

Meacham et al.
Figure 2. Structure of the complex cation of [2][PF6]‚3CH2Cl2.
Table 1. Crystallographic Data for Complexes 1 and [2][PF6]‚3CH2Cl2
Table 3. Selected Bond Lengths (Å) and Angles (deg) for
[2][PF6]‚3CH2Cl2
complex
empirical formula
fw
wavelength, Å
space group
a, Å
b, Å
1
[2][PF6‚3CH2Cl2
C98H75Cl6F6O9P5Ru2
2080.27
0.71073
P1h
14.380(4)
16.759(5)
21.153(8)
85.70(3)
87.55(2)
74.64(2)
4901(3)
2
1.410
0.619
0.0603
0.1305
0.1078
0.1467
C48H36O7P2Ru2
887.78
0.71073
P1h
12.0966
Ru(1)-C(31)
Ru(1)-C(41)
Ru(1)-O(11)
Ru(1)-O(24)
Ru(1)-P(100)
Ru(1)-P(200)
O(11)-C(11)
O(24)-C(24)
1.868(6)
1.874(6)
2.062(4)
2.074(3)
2.4150(17)
2.4163(17)
1.289(7)
1.305(6)
Ru(2)-C(61)
Ru(2)-C(51)
Ru(2)-O(17)
Ru(2)-O(18)
Ru(2)-P(400)
Ru(2)-P(300)
O(18)-C(18)
O(17)-C(17)
1.878(6)
1.881(7)
2.065(3)
2.073(3)
2.4016(17)
2.4076(17)
1.310(6)
1.305(6)
13.2288
15.691
c, Å
R, deg
103.585(11)
90.756(15)
116.967(8)
2154.4(5)
2
1.369
0.488
R1 ) 0.0261
wR2 ) 0.0558
R1 ) 0.0392
wR2 ) 0.0580
â, deg
C(31)-Ru(1)-C(41)
C(31)-Ru(1)-O(11)
C(41)-Ru(1)-O(11)
C(31)-Ru(1)-O(24)
C(41)-Ru(1)-O(24)
O(11)-Ru(1)-O(24)
C(31)-Ru(1)-P(100)
C(41)-Ru(1)-P(100)
O(11)-Ru(1)-P(100)
O(24)-Ru(1)-P(100)
C(31)-Ru(1)-P(200)
C(41)-Ru(1)-P(200)
O(11)-Ru(1)-P(200)
O(24)-Ru(1)-P(200)
95.2(2)
172.3(2)
92.5(2)
C(61)-Ru(2)-C(51)
C(61)-Ru(2)-O(17)
C(51)-Ru(2)-O(17)
96.0(2)
90.31(19)
173.7(2)
169.9(2)
93.9(2)
79.82(14)
91.34(17)
91.47(17)
87.87(11)
90.35(11)
90.42(17)
89.74(17)
90.73(11)
87.68(11)
γ, deg
vol, Å3
Z
calcd density, Mg/m3
92.10(19) C(61)-Ru(2)-O(18)
172.6(2) C(51)-Ru(2)-O(18)
µ(mm-1
)
80.18(14) O(17)-Ru(2)-O(18)
90.76(17) C(61)-Ru(2)-P(400)
91.65(17) C(51)-Ru(2)-P(400)
88.91(11) O(17)-Ru(2)-P(400)
87.24(11) O(18)-Ru(2)-P(400)
91.28(17) C(61)-Ru(2)-P(300)
91.73(17) C(51)-Ru(2)-P(300)
88.59(11) O(17)-Ru(2)-P(300)
89.10(11) O(18)-Ru(2)-P(300)
final R indices
[I > 2σ(I)]
final R indices
(all data)
Table 2. Selected Bond Lengths (Å) and Angles (deg) for 1
Ru(1)-C(41)
Ru(1)-C(31)
Ru(1)-O(12)
C(12)-O(12)
C(18)-O(18)
1.8737(18)
1.8840(l9)
2.0528(12)
1.343(2)
Ru(1)-O(21)
Ru(1)-P(100)
Ru(1)-P(200)
C(21)-O(21)
C(18)-O(19)
2.0722(11)
2.4009(6)
2.4297(6)
1.3438(19)
1.380(2)
P(100)-Ru(1)-P(200) 175.88(5) P(400)-Ru(2)-P(300) 177.76(5)
Table 4. Electrochemical Dataa
redox potentials [E1/2, V vs Fc/Fc+]
1.213(2)
C(41)-Ru(1)-C(31)
C(31)-Ru(1)-O(12)
C(31)-Ru(1)-O(21)
C(41)-Ru(1)-P(100)
O(12)-Ru(1)-P(100)
C(41)-Ru(1)-P(200)
O(12)-Ru(1)-P(200)
93.31(7)
173.72(6)
92.34(6)
89.88(6)
87.18(4)
90.49(6)
91.19(4)
C(41)-Ru(1)-O(12)
92.90(6)
C(41)-Ru(1)-O(21) 174.04(6)
complex
Ab
Bc
O(12)-Ru(l)-O(21)
81.48(4)
C(31)-Ru(1)-P(100) 91.94(6)
O(21)-Ru(1)-P(100) 91.82(4)
C(31)-Ru(1)-P(200) 89.64(6)
O(21)-Ru(1)-P(200) 87.66(4)
1
-0.17
+0.07, +0.76
-0.07, +0.37
+0.86
+1.52
[2]+
3
a Measurements made at a Pt-bead working electrode in CH2Cl2
P(100)-Ru(1)-P(200) 178.347(18)
n
containing 0.1 M Bu4NPF6 as base electrolyte; scan rate was 0.2 V s-1
.
b Fully reversible catecholate/semiquinone interconversions. c Irreversible
semiquinone/quinone interconversions.
decrease in electron density at the metal center. This is a
relatively small shift, in agreement with the oxidation being
a largely ligand-centered cat/sq couple; wholly metal-
centered oxidations of carbonyl complexes result in shifts
of ca.100 cm-1 of the CO stretching vibrations to higher
energy.10,11 In addition, the carbonyl stretching vibration
associated with the coumarin ester unit moves from 1686 to
1725 cm-1; oxidation of the catecholate unit to semiquinone
decreases its ability to act as an electron-rich substituent to
this ester group, and accordingly this ligand-centered CO
vibration also moves to higher energy.
The EPR spectrum of chemically generated [1]+ (Figure
3a) also confirms the substantially ligand-centered nature of
the oxidation. The g value of 2.0013 is close to that expected
for organic radicals, and the hyperfine coupling constants
to the metal and phosphine ligands are small. The most
obvious feature of the hyperfine coupling is the triplet, which
has a coupling constant AP of 22 G, arising from coupling
to two equivalent 31P nuclei. (That this coupling involves
(10) (a) Anderson, K. M.; Connelly, N. G.; Llamas-Rey, E.; Orpen, A. G.;
Paul, R. L. Chem. Commun. 2001, 1734. (b) Connelly, N. G.; Hicks,
O. M.; Lewis, G. R.; Orpen, A. G.; Wood, A. J. J. Chem. Soc., Dalton
Trans. 2000, 1637.
(11) Atwood, C. G.; Geiger, W. E. J. Am. Chem. Soc. 2000, 122, 5477.
7890 Inorganic Chemistry, Vol. 42, No. 24, 2003

Ruthenium Carbonyl Complexes with Dioxolene Ligands
Scheme 2. Sketches of the Oxidized Forms of [L2]3- and [L3]4- a
Figure 4. (a, top) IR spectra recorded during oxidation of [2]+ to [2]2+ in
an OTTLE cell (CH2Cl2, 243 K). (b, bottom) Temperature dependence of
the spectrum of the mixed-valence complex [2]2+ between 233 and 223 K.
CO vibrations are observed in the mixed-valence state, and
that the shift from the neutral form is approximately half of
that observed following oxidation of 1 to [1]+, both indicate
that the odd electron in [2]2+ is fully delocalized across the
bridging ligand at 243 K, resulting in two equivalent termini
with the positive charge evenly shared between them on the
fast IR time scale. Cooling of the sample, however, resulted
in the two peaks becoming four; this occurs between 233
and 223 K, and the spectral changes (Figure 4b) were fully
reversible over several heating/cooling cycles. The reduction
in temperature is slowing the electron delocalization rate
down to the extent where it becomes slow on the IR time
scale at 223 K. The spectrum that is observed at 223 K is
far from being a fully localized spectrum, in which two CO
vibrations would be unshifted (at the same position as in
[2]+) whereas two would be shifted by ca. 30-40 cm-1 (cf.
the shift observed on oxidation of 1 to [1]+), but it does show
the transition from fully delocalized on the IR time scale at
233 K to the onset of localization by 223 K.
a Only one canonical form is represented for each oxidation state.
Further oxidation to [2]3+ at 243 K results in a similar
further shift to higher energy of the carbonyl peaks, to 2030
and 2076 cm-1, an average shift of an additional 17 cm-1;
these values for [2]3+ are quite similar to those seen for [1]+
because of the presence of one additional positive charge
associated with each Ru-dioxolene terminus in each case.
In keeping with the fully delocalized nature of [2]2+ at
temperatures above 233 K, as shown by the IR spectra, its
isotropic EPR spectrum at room temperature (cf. Figure 3b)
shows a quintet at g ) 2.00 due to hyperfine coupling to
four equivalent phosphorus nuclei; given that EPR has a
slower time scale than IR (ca. 10-8 vs 10-13 s), this was to
be expected. The coupling constant AP of 13.5 G is much
less than that observed in [1]+, as is the coupling to the
ruthenium nuclei (ARu ) 1.9 G), both of which are consistent
with the fact that greater delocalization of the electron will
reduce the magnitude of its coupling to each spin with which
it interacts. To a first approximation, the difference in these
coupling constants between a mononuclear complex and an
Figure 3. X-band EPR spectra (in CH2Cl2) of the ligand-centered radical
species (a) [1]+ and (b) [2]2+ (the EPR spectrum of [3]+ is very similar to
that of [2]2+).
the axial phosphine ligands and not the semiquinone radical
H atoms was shown by Connelly et al., who observed a
similar EPR spectrum on oxidation of [Ru(PPh3)2(CO)2-
(Cl4cat)], which has no H atoms on the dioxolene ligand).7a
Closer inspection also reveals small satellites on each
component of the triplet, which are the outermost lines of
the sextet arising from coupling to those spin-active Ru nuclei
5
which have I ) /2, from which a coupling constant ARu of
4 G can be determined.
Complex 2 was studied by IR spectroelectrochemistry at
243 K in the three accessible oxidation states, with charges
of +1, +2 (radical), and +3. The two CO stretching
vibrations are at 1988 and 2046 cm-1 in [2]+; on oxidation
to [2]2+ they move to 2010 and 2063 cm-1, respectively, an
average shift of ca.19 cm-1 (Figure 4a). The facts that two
Inorganic Chemistry, Vol. 42, No. 24, 2003 7891

Meacham et al.
somewhere between completely delocalized and completely
localized, indicative of class II behavior, with the rate of
electron hopping between the termini in the mixed-valence
state being slow on the IR time scale (<1013 s-1) at room
temperature; this decreased delocalization compared to [2]2+
(which is fully delocalized down to 233 K) is in agreement
with the electrochemical data described above. Further
oxidation to [3]2+ results in the four CO stretches being
replaced by two, at 2012 and 2063 cm-1 (cf. the mononuclear
semiquinone complex [1]+), as the complex regains its 2-fold
symmetry with each ligand terminus becoming a semi-
quinone unit.
Figure 5. IR spectra of 3 (dashed line) and [3]+ (solid line) in CH2Cl2 at
room temperature.
Since the mixed-valence species [3]+ is localized on the
IR time scale, it was of particular interest to examine this
species by EPR spectroscopy to see if the delocalization
could be “bracketed” by the different time scales of the two
methods. The EPR spectrum of [3]+ does indeed show a
quintet centered at g ) 2.006, arising from equal coupling
to all four 31P nuclei, with AP ) 12.1 G and ARu ) 2 G.
These coupling constants are very similar to those seen for
[2]2+ and about half the magnitude of those seen for [1]+
(the appearance of this spectrum is almost identical to that
of [2]2+, Figure 3b). It is clear that the unpaired electron in
[3]+ is delocalized on the EPR time scale despite being
localized on the IR time scale, such that its exchange rate
between the two dioxolene sites is >108 s-1 but <1013 s-1
at room temperature. (The same applies to [2]2+ but only at
temperatures below 223 K; at higher temperatures it is fully
delocalized by both EPR and IR methods.) Similar behavior
has been described recently for a mixed-valence dinuclear
Mo(I)/Mo(II) complex, where the unpaired electron showed
equal coupling to both Mo nuclei in the EPR spectrum, but
the IR spectrum showed the presence of two sets of vibrations
associated with distinct Mo(I) and Mo(II) termini.13 Geiger
and co-workers have also described how the dinuclear mixed-
valence complex [(fulv){Mn(CO)2}2]+ (fulv ) fulvalenediyl)
is delocalized according to EPR spectroscopy but localized
according to IR spectroscopy.11
UV/Vis/NIR Spectroelectrochemical Studies. All three
complexes were examined by UV/vis/NIR spectroelectro-
chemistry in CH2Cl2 using an OTTLE cell thermostated at
243 K; the redox interconversions examined were fully
chemically reversible, as shown by the presence of clean
isosbestic points during the interconversions in all cases.
(i) Spectra of Mononuclear Complex 1. The spectrum
of 1 (Figure 6, Table 5) shows three discernible transitions,
with clear absorption maxima at 432 and 279 nm, and a
shoulder at about 375 nm. On oxidation to [1]+ these
transitions all change slightly; the transition at ca. 430 nm
is diminished in intensity, and new transitions appear at 341,
535, and 582 nm. In addition, at lower energy, a weak
transition appears at ca. 1250 nm which is not shown in the
figure.
analogous dinuclear complex should be a factor of 2, in
reasonable agreement with what we observe.
In complex 3 the two CO stretching vibrations are at 1965
and 2030 cm-1 with almost equal intensities. One-electron
oxidation, either chemically (ferrocenium hexafluorophos-
phate) or electrochemically (OTTLE cell) resulted in these
being replaced by four new peaks at 2055 (m), 2040 (vs),
1998 (m), and 1980 (m) cm-1 (Figure 5). The presence of
four peaks indicates that the complex is no longer symmetric,
with the oxidation being localized at one terminus more than
the other, as observed for [2]2+ below 233 K. However, if
the semiquinone site were completely localized at one
terminus, we would expect to seesby analogy with the
results abovestwo CO peaks that scarcely move, associated
with the catecholate terminus, and two which move by about
35 cm-1 on average, associated with the semiquinone
terminus. In fact, we see one pair of peaks (1980 and 2040
cm-1) shifted by an average of 12 cm-1, and the other (1998
and 2055 cm-1) shifted by an average of 29 cm-1. Further-
more, the relative intensities of the peaks are unexpected.
Complex 3 displays symmetric and antisymmetric stretches
of roughly equal intensities, typical of cis-Ru(CO)2 fragments
with OC-Ru-CO angles ca. 90°; for [3]+, three of the
absorptions are comparable and with slightly lower intensity
than for 3, but the 2040 cm-1 band is 4-5 times more intense
than the others. This should be a textbook example relating
the inter-carbonyl angle to the ratio of intensities of the
symmetric and antisymmetric stretches.12 However, this
cannot be the explanation here for the variations in intensities
because the OC-Ru-CO angle would have to be unreason-
ably small. A possible explanation for the greater intensity
of the 2040 cm-1 band is that the four CO stretches are
coupled. Assuming approximate C2V symmetry (not D2h
because the metal termini are no longer equivalent), the four
CO stretching vectors combine to give a totally symmetric
A1 mode (expected to be of highest frequency), an asym-
metric (with respect to the molecular center) A1 mode (next
highest in frequency), and two lower frequency B1 modes.
The asymmetric A1 mode has the same symmetry as those
vibrations in the Ru-(rufigallolate)-Ru plane which would
accompany intervalence charge transfer, which might allow
for an increase in intensity for this mode alone compared to
the other three. Complex [3]+ is apparently behaving
Partial assignments for these electronic spectra can be
made on the basis of ZINDO calculations on the complexes,
(12) Cotton, F. A.; Wilkinson, G. AdVanced Inorganic Chemistry, 5th ed.;
Wiley-Interscience: New York, 1988; pp 1035-1037.
(13) Wlodarczyk, A.; Maher, J. P.; McCleverty, J. A.; Ward, M. D. J. Chem.
Soc., Dalton Trans. 1997, 3287.
7892 Inorganic Chemistry, Vol. 42, No. 24, 2003

Ruthenium Carbonyl Complexes with Dioxolene Ligands
Figure 6. Electronic spectra recorded during oxidation of 1 to [1]+ in an
OTTLE cell (CH2Cl2, 243 K; the spectrum of the fully oxidized form is
shown as a dashed line for clarity).
Table 5. Results of UV/Vis/NIR and IR Spectroelectrochemical
Studies (CH2Cl2, 243 K)a
complex ν(CO), cm-1
λ
max, nm (10-3ꢀ, dm3 mol-1 cm-1
)
1
1978, 2042
2018, 2072
279 (32), 432 (10)
[1]+
283 (28), 341 (6.0), 433 (7.9), 535 (2.2),
582 (2.9), ≈1250 (0.2)
[2]+
1988, 2046
2010, 2063
2030, 2076
272 (38), 599 (48), 899 (0.4)
[2]2+
[2]3+
272 (38), 626 (33), 899 (4.8), 1995 (0.2)
294 (34), 630 (34), 670 (13), 820 (16),
1348 (4.3), 1714 (16)
Figure 7. Electronic spectra of [2]+ (‚‚‚) [2]2+ (- - -), and [2]3+ (s)
3
1965, 2030
2055, 2040,
1998, 1980
2012, 2063
360 (92), 430 (sh), 449 (14), 547 (10)
288 (18), 342 (12), 398 (39), 464 (13),
556 (6.3), 682 (6.1), 1118 (4.4), 2695 (16)
294 (18), 521 (10), 802 (11)
recorded during an OTTLE experiment (CH2Cl2, 243 K).
[3]+
most effectively with the sq π-system is expected to be the
most intense and also to have the highest energy. The
observed transitions at 582, 535, and 433 nm agree well with
this assignment which is supported by the ZINDO calcula-
tion. At higher energy, the new 341 nm transition has a
mixture of Ru f sq MLCT and sq-based π f π* character,
and the intense band at 283 nm consists of several compo-
nents of which the most significant are sq-centered π f π*,
and PPh3 f sq ligand-to-ligand charge transfer.
(ii) Dinuclear Complex 2. The electronic spectra of [2]+
and its one-electron and two-electron oxidized forms [2]2+
and [2]3+ are in Figure 7 (see also Table 5). Most of the
spectral assignments proposed below follow from the above
analysis of 1/[1]+ and are again based on ZINDO calcula-
tions, using the geometry provided by the crystal structure
for [2]+. For the singly and doubly oxidized forms, the
ZINDO calculations were performed on DFT-optimized
geometries in which (for simplicity) the PPh3 ligands were
replaced by PMe3.
In the fully reduced state [2]+ the electronic spectrum is
dominated by an intense transition at 599 nm, which is
analogous to the lowest-energy transition of 1 (Figure 6) in
that it is mostly a ligand-centered π f π* transition with a
small amount of Ru[d(π)] f [bridging ligand (π*)] MLCT
character. On oxidation to [2]2+, two new features develop
in the low-energy part of the spectrum: a relatively weak
transition at 899 nm and a broad, much lower energy
transition at ca. 2000 nm. The former of these does not have
any strong counterpart in the computed spectrum, although
the calculations do indicate a weak absorbance in this region
ascribable to Ru[d(π)] f [bridging ligand(SOMO)] MLCT
[3]2+
a Suggested assignments for the electronic spectra are given in the main
text.
using the geometry provided by the X-ray structure (for 1)
and the DFT-minimized geometry (for [1]+). Because of the
low symmetry there is extensive orbital mixing and only the
dominant contributions are given; a more detailed analysis
will be the subject of a future paper. For 1 the HOMO is a
π-orbital on the catecholate ligand, with a small admixture
of metal d(π) character; the LUMO is a catecholate-centered
π* orbital. The 432 nm transition of 1 is therefore mostly
catecholate-centered π f π* with a small Ru[d(π)] f
catecholate (π*) MLCT component. The high-energy shoul-
der on this, and the more intense transition in the UV region
at 279 nm, both consist of a similar mixture of (higher-
energy) catecholate-centered π f π* and Ru[d(π)] f
catecholate (π*) transitions, with the former dominating.
For the oxidized complex [1]+, the calculation confirms
that (as expected) the SOMO is almost entirely (96%) on
the dioxolene ligand, i.e., it is correct to call it a semiquinone
(sq) species, and the low-energy, weak NIR transition at 1250
nm is an internal semiquinone-centered transition from a
filled π orbital to the SOMO.14 For a Ru(II)-sq complex in
low symmetry it is expected that there can be three Ru[d(π)]
f sq(SOMO) MLCT transitions, one from each of the
ruthenium d(π) orbitals; the metal orbital which overlaps
(14) Haga, M.; Dodsworth, E. S.; Lever, A. B. P. Inorg. Chem. 1986, 25,
447.
(15) This footnote was deleted during revision.
Inorganic Chemistry, Vol. 42, No. 24, 2003 7893

Meacham et al.
transitions originating from the three inequivalent d(π)
orbitals. The 2000 nm transition is an internal π f π*
transition (HOMO f SOMO) associated with the bridging
ligand radical and is well reproduced by the calculation (the
analogous transition in [1]+ occurred at ca.1300 nm). The
main transition in the visible region, at 599 nm for [2]+, is
reduced in intensity and slightly red-shifted to 626 nm, and
is again largely bridging-ligand centered π f π* in character.
The higher-energy, weaker transitions around 400 nm for
[2]2+ are also largely of bridging-ligand centered π f π*
character, originating from lower-energy filled π orbitals.
According to the calculations no transitions in the UV or
visible region for [2]2+ involve the PMe3 or CO groups, but
in the “real” complex (with axial PPh3 ligands) mixing of
frontier orbitals with the P-phenyl rings is likely, and this
may account for some of the discrepancies, most notably
the fact that the peak at 899 nm is not well reproduced in
the calculations, being predicted to have a much lower
intensity than it actually does.
On further oxidation to [2]3+, the two low-energy transi-
tions are blue-shifted and become more intense (Figure 7).
Attempts to match these transitions to calculated spectra were
complicated by the fact that the calculation showed that the
two lowest energy states are a singlet and a triplet, and that
they are very close in energy, so the ground state could
actually have triplet character with the doubly oxidized
bridging ligand being a diradical. We could find no evidence
for this in the EPR spectrum, with the EPR spectrum of [2]3+
being featureless, although such negative evidence is not
conclusive. In view of this ambiguity we prefer not to attempt
a detailed analysis of the spectra but just point out that, in
view of the assignment of the spectrum of [2]2+, transitions
of bridging-ligand centered π f π* character and Ru[d(π)]
f (bridging ligand) MLCT character are expected in this
region.
(iii) Dinuclear Complex 3. The electronic spectra associ-
ated with oxidation of 3 to [3]+ and then [3]2+ are in Figure
8 (see also Table 5). The lowest-energy transition of 3 at
547 nm is directly analogous to the lowest-energy transitions
of 1 and 2, at 432 and 599 nm respectively; ZINDO
calculations, based on a DFT-optimized geometry using axial
PMe3 ligands in place of PPh3, showed that this has
principally bridging-ligand based π f π* character with a
small amount of Ru[d(π)] f (bridging ligand) MLCT
character. The two higher-energy transitions at 449 and 360
nm have similar provenance.
Oxidation to [3]+ results in appearance of two lower-
energy transitions at 682 and 1118 nm, both being principally
bridging-ligand based π f π* transitions which, as in the
spectra of [1]+ and [2]2+, have become red-shifted compared
to the parent reduced complexes; a higher-energy series of
transitions between 300 and 500 nm are also bridging-ligand
based π f π* transitions with some additional MLCT
character involving the bridging ligand SOMO. In addition,
a broad, intense transition appears at low energy in the
spectrum of [3]+ at 2695 nm (3710 cm-1; ꢀ ≈ 2 × 104 M-1
cm-1). Such a feature is not present in any mononuclear
complexes of this nature such as [1]+,6 and has the
Figure 8. Electronic spectra of 3 (s), [3]+ (- - -), and [3]2+ (‚‚‚) recorded
during an OTTLE experiment (CH2Cl2, 243 K). The inset showing the IVCT
transition for [3]+ is plotted in a linear energy scale.
characteristic position and appearance of an intervalence
charge-transfer (IVCT) transition. This implies that the first
oxidation of 3 is localized at one terminus of the bridging
ligand, which can therefore be described has having distinct
catecholate and semiquinone termini, i.e., a mixed-valence
species; such a transition is consistent with a catecholate f
semiquinone IVCT transition between the two termini of the
bridging ligand.16 The absorption is asymmetric, narrower
on the low-energy side; this is most likely due to overlap of
a narrow (width ca. 800 cm-1) band near 3700 cm-1 with
two weaker and broader (width ca. 2000 cm-1) bands at
higher energy.18 The position of the absorption maximum is
very slightly solvent-dependent, being blue-shifted in more
polar solvents (THF, 3690 cm-1; dichloromethane, 3710
cm-1; acetone, 3800 cm-1; acetonitrile, 3840 cm-1), in
keeping with its directional charge-transfer nature.
Meyer and co-workers have reviewed the class II to class
III transition in mixed-valence complexes and have defined
the new category class II-III.17 The experimental criteria
for class II-III mixed-valence compounds include the
presence of narrow, solvent-independent, low-energy IVCT
absorptions (typically associated with class III complexes),
(16) (a) Lahlil, K.; Moradpour, A.; Bowlas, C.; Nenou, F.; Cassoux, P.;
Bonvoisin, J.; Launay, J.-P.; Dive, G.; Dahareng, D. J. Am. Chem.
Soc. 1995, 117, 9995. (b) Lambert, C.; No¨ll, G. J. Am. Chem. Soc.
1999, 121, 8434.
(17) Demandis, K. D.; Hartshorn, C. M.; Meyer, T. J. Chem. ReV. 2001,
101, 2655.
(18) This interpretation of the band at 3710 cm-1 was provided by Dr.
Thomas J. Meyer (private communication).
7894 Inorganic Chemistry, Vol. 42, No. 24, 2003

Ruthenium Carbonyl Complexes with Dioxolene Ligands
combined with experimental evidence of localized (class II)
electronic structure, such as the low symmetry of the CO
stretches observed here. An IVCT transition at this energy
for a class II mixed-valence state should be Gaussian and
symmetric with a ∆v1/2 value of about 3000 cm-1 [from ∆v1/2
) (2310Eop)1/2].
The narrowness and solvent-insensitivity of the IVCT band
can be explained by solvent averaging, which requires that
electron transfer be rapid compared to the rate of reorientation
of local solvent dipoles (0.2-10 ps).17 Accordingly the rate
of electron transfer is greater than ca. 1011 s-1, while still
being slower than the IR time scale (1013 s-1), which allows
us to bracket the electron-transfer rate in a narrower range
than we could get from comparison of the EPR and IR
spectra (electron-transfer rate > 108 s-1 but < 1013 s-1). The
behavior we see for [3]+ is therefore consistent with class
II-III behavior as described by Meyer;17 a more detailed
analysis of the spectroscopic properties of [3]+ and related
complexes will be the subject of further studies.
vis/NIR spectroelectrochemical experiments were performed
to measure the electronic spectra of these complexes in all
accessible oxidation states; these spectra were partially
assigned with the aid of ZINDO calculations. Mixed-valence
complex [3]+ shows a low-energy transition in its electronic
spectrum ascribable to a catecholate f semiquinone inter-
valence charge transfer between inequivalent termini (cf. the
IR spectra data); the narrowness of the IVCT band indicates
solvent averaging and suggests that the electron-transfer rate
is actually greater than 1011 s-1 (borderline class II-III
behavior). No such transition is present for [2]2+. These
results are in agreement with the smaller separation between
the two redox processes of 3 compared to [2]+.
Experimental Section
General Details. The ligands H2L1 and H3L2 were purchased
from Aldrich and used as received; rufigallol (H4L3) was prepared
according to a literature procedure.19 Ru3(CO)12 was obtained from
Strem.
On further oxidation to [3]2+, the IVCT transitions and
the 1118 nm ligand-centered transitions of [3]+ disappear
(Figure 8) and are replaced by an intense transition at 810
nm. The ZINDO calculations on doubly oxidized [3]2+
highlight the same problem that occurs for calculations of
[2]3+, viz., that there are singlet and triplet states very close
in energy such that the ground state could be either. Actually
the triplet ground state gives a much better prediction of the
observed electronic spectrum than does the singlet state; both
confirm the presence of a near-IR transition at ca. 1000 nm,
having a mixture of bridging-ligand centered π f π* and
some MLCT character, but if the ground state is assumed to
be a singlet the predicted intensity of this transition is
enormously over-estimated. The predicted triplet spectrum
in contrast gives a reasonable intensity for this transition.
Given the ambiguity in the nature of the ground state, further
analysis of this spectrum is inappropriate.
Physical and Spectroscopic Methods. UV/vis/NIR spectroelec-
trochemical measurements were performed in a home-built OTTLE
cell mounted in the sample compartment of a Perkin-Elmer Lambda
19 spectrometer as described previously;20 distilled CH2Cl2 was
the solvent in every case, and all measurements were carried out
at -30°C (to minimize evaporation of the solvent from the cuvette
during the experiment). For all of the transformations described,
clean isosbestic points were obtained and the spectra of the starting
materials could be regenerated with no significant changes by
reversing the electrolysis. IR spectroelectrochemical measurements
were carried out using a homebuilt OTTLE cell based on a KBr
microcavity cuvette (Spectratech), with a Pt-gauze working elec-
trode, a Pt wire counter electrode, and an Ag wire coated with AgCl
as the reference; again, distilled CH2Cl2 was used as solvent and
the measurements were carried out at -30 °C using a Bruker IFS-
25 spectrometer. Electrochemical measurements were carried out
as described previously, using a standard three-electrode cell;5
ferrocene was used as an internal standard, and all potentials are
quoted vs the ferrocenium/ferrocene couple. X-band EPR spectra
were recorded on either an IBM/Bruker ER-200 SRC spectrometer
(Buffalo) or a Bruker ESP-300E spectrometer (Bristol), in CH2Cl2
solutions at room temperature using a microwave power of 20 mW;
spectral parameters were determined by comparison of experimental
spectra with spectra simulated using Bruker WINEPR Simfonia
(ver. 1.25) software. Other instrumentation used for routine
measurements on 1 and [2][PF6] carried out in Bristol have been
described before.5 Measurements on complex 3 at Buffalo used
the following equipment: infrared spectra, a Nicolet Magna 550
spectrophotometer; NMR spectra (in CDCl3), Varian Associates
Gemini 300 or VXR-400S instruments; UV/visible spectra (300-
800 nm), a Hewlett-Packard 8452A diode array spectrophotometer.
Syntheses. (a) [Ru(CO)2(PPh3)2(L1)], 1. A mixture of Ru3(CO)12
(140 mg, 0.22 mmol), PPh3 (380 mg, 1.46 mmol), and H2L1 (140
mg, 0.71 mmol) in dry toluene (30 cm3) under N2 was heated to
reflux for 12 h to give a red solution. After removal of the solvent
in vacuo, the crude material was purified by column chromatog-
raphy (alumina, CH2Cl2:MeOH, 99:1 v/v). The second (major)
orange band was the product: yield, 0.302 g (52%). FABMS: m/z
888 (M+), 860 (M+ - CO), 626 (M+ - PPh3). 31P{1H} NMR: δ
Conclusions
Mononuclear complex 1 and dinuclear complexes [2]+ and
3 undergo reversible oxidations (one for 1, two each for 2
and 3) which are formally catecholate/semiquinone-type
processes associated with the dioxolene bridging ligands. For
3, the separation between the two successive processes (440
mV) is substantially less than in [2]+ (690 mV), indicative
of a weaker electronic couplingsand hence less delocaliza-
tion in the mixed-valence statesfor 3 compared to [2]+. For
these dinuclear complexes, mono-oxidation generates a
ligand-centered “mixed-valence” radical for which the extent
of delocalization could be studied by IR and EPR spectros-
copy. EPR spectra indicate that both [2]2+ and [3]+ are
delocalized on the EPR time scale; however IR spectra
indicate that whereas [2]2+ is also delocalized on the IR time
scale at 243 K, [3]+ is localized with spectroscopically
distinct {Ru(CO)2} termini, with evidence for coupling
between the two pairs of CO vibrations. Accordingly the
rate of exchange of the odd electron between the two termini
of [3]+ can be bracketed, as greater than 108 s-1 (EPR time
scale) but less than 1013 s-1 (IR time scale) at 243 K. UV/
(19) Grimshaw, J.; Haworth, R. D. J. Chem. Soc. 1956, 4224.
(20) Lee, S.-M.; Kowallick, R.; Marcaccio, M.; McCleverty, J. A.; Ward,
M. D. J. Chem. Soc., Dalton Trans. 1998, 3443.
Inorganic Chemistry, Vol. 42, No. 24, 2003 7895

Meacham et al.
17.0 (PPh3). IR (CH2Cl2): 2042, 1978, 1686 cm-1. Anal. Calcd
for C48H36O7P2Ru: C, 64.9; H, 4.1. Found: C, 64.5; H, 4.0.
(b) [{Ru(CO)2(PPh3)2}2(L2)][PF6], (2)[PF6]. A mixture of Ru3-
(CO)12 (100 mg, 0.16 mmol), PPh3 (380 mg, 1.45 mmol), and H3L2
(120 mg, 0.357 mmol) in dry toluene (30 cm3) under N2 was heated
to reflux for 15 h to give a deep blue solution. After removal of
the solvent, the crude material was purified by column chroma-
tography (alumina, CH2Cl2: MeOH, 95:5 v/v). After removal of
traces of unreacted Ru3(CO)12, a deep blue band eluted, which was
collected and evaporated to dryness. The blue complex was
redissolved in the minimum amount of MeOH and precipitated as
its hexafluorophosphate salt by addition of aqueous KPF6. The
suspension was extracted several times with CH2Cl2; the combined
organic extracts were dried over MgSO4 and evaporated to dryness
to give pure (2)[PF6], which was finally recrystallized from CH2-
Cl2/hexanes. Yield: 0.309 g (45%). FABMS: m/z 1680 (M -
PF6)+, 1417 (M - PF6 - PPh3)+, 1391 (M - PF6 - PPh3 - CO)+,
1157 (M - PF6 - 2PPh3)+. 31P{1H} NMR: δ 21.3 (PPh3), -144.2
(septet, PF6-). IR (CH2Cl2): 2046, 1988, 1607 cm-1. Anal. Calcd
for C95H69F6O9P5Ru2‚CH2Cl2: C, 60.3; H, 3.7. Found: C, 60.1;
H, 3.7.
and refined without positional constraints; hydrogen atoms were
constrained to ideal geometries and refined with fixed isotropic
displacement parameters. For 1 the solution and refinement were
straightforward and presented no problems. The structural deter-
mination of [2][PF6]‚3CH2Cl2 was complicated by the presence of
disorder involving the bridging ligand and some of the solvent
molecules. In the picture shown in Figure 2, the phenyl substituent
of the bridging ligand [C(141)-C(146)] is directed into the page
and the central oxygen atom O(21) is at the front, with a lattice
CH2Cl2 molecule nearby; the disorder involves all of these
components swapping over such that the region either side of the
center of the bridging ligand is occupied by a partial phenyl ring
and a partial CH2Cl2 molecule. Atoms O(21) and C(14) were
likewise mutually disordered. This could be successfully resolved,
but the disordered C atoms had to be refined with isotropic thermal
parameters; H atoms were not included on these C atoms.
Calculations. The INDO/S-derived predicted spectra and mo-
lecular orbital descriptions were based on DFT optimized structures
and will be discussed in depth elsewhere.23 These data were
obtained using the INDO/S method in the HYPERCHEM 5.1
program (Hypercube Inc., Gainesville, FL). The DFT calculations
presented were carried out using the GAUSSIAN 98 program24
using Becke’s three-parameter hybrid functional25 with the LYP
correlation functional26 (B3LYP) and an effective core potential
basis set LanL2DZ.27 For further details see ref 28.
(c) [{Ru(CO)2(PBu3)2}2(L3)], 3. A solution of Ru3(CO)12 (104
mg, 0.162 mmol), PBu3 (240 µL, 0.95 mmol), and rufigallol (81
mg, 0.27 mmol) in dry toluene (20 cm3) was heated at reflux under
an argon atmosphere for 7 h. The resulting red solution was
evaporated to dryness, and the residue was applied as a dichlo-
romethane solution to a silica gel preparative TLC plate. Elution
with 4% ethyl acetate in dichloromethane gave a red-brown colored
band, which was extracted with ethyl acetate. Recrystallization from
Acknowledgment. Financial support from EPSRC and
Johnson Matthey (U.K.) is gratefully acknowledged. We
thank Dr. Thomas J. Meyer of Los Alamos National
Laboratory for helpful comments concerning the spectros-
copy of the mixed-valence species [3]+, and Dr. John Maher
of the University of Bristol for assistance with the EPR
spectroscopy.
1
methanol gave brown crystals. Yield: 212 mg, 61%. H NMR
(CDCl3): δ 13.15 (s, 2 H), 7.02 (s, 2 H), 1.74 (m, 12 H), 1.46 (br,
12 H), 1.38 (m, 12 H), 0.90 (t, 18 H, J 7 Hz). 13C{1H} NMR
(CDCl3): 197.9 (br t, 4 C, JPC 11 Hz), 186.9 (2 C), 167.1 (2 C),
154.6 (2 C), 153.8 (2 C), 123.6 (2 C), 110.1 (2 C), 110.0 (2 C),
25.6 (12 C), 24.6 (t, 12 C, JPC 6 Hz), 22.9 (t, 12 C, JPC 12 Hz),
13.9 (12 C) ppm. 31P{1H} NMR (CDCl3): δ 19.11 (PBu3). IR (CH2-
Cl2): 2030 vs, 1965 s cm-1. Anal. Calcd for C66H112O12P4Ru2: C,
55.7; H, 7.9. Found: C, 55.7; H, 8.0.
Supporting Information Available: Crystallographic data in
CIF format. This material is available free of charge via the Internet
at http://pubs.acs.org.
IC034579O
Crystallography. Suitable crystals were mounted on a Bruker
SMART-CCD (for 1) or APEX (for [2][PF6]‚3CH2Cl2) diffracto-
meter equipped with graphite-monochromatized Mo KR radiation.
Crystallographic measurements were carried out at 173 K (for 1)
or 100 K (for [2][PF6]‚3CH2Cl2); details of the crystal, data
collection, and refinement parameters are summarized in Table 1,
and selected structural parameters are collected in Tables 2 and 3.
After integration of the raw data and merging of equivalent
reflections, an empirical absorption correction was applied based
on comparison of multiple symmetry-equivalent measurements.21
The structures were solved by direct methods and refined by full-
matrix least squares on weighted F2 values for all reflections using
the SHELX suite of programs.22 All non-hydrogen atoms were
assigned anisotropic displacement parameters (apart for a few
disordered atoms in the structure of [2][PF6]‚3CH2Cl2, see below)
(23) Lever, A. B. P.; Keister, J. B.; Ward, M. D., in preparation.
(24) Frisch, M. J.; Trucks, G. W.; Schlegel, H. B.; Scuseria, G. E.; Robb,
M. A.; Cheeseman, J. R.; Zakrzewski, V. G.; Montgomery, J. A., Jr.;
Stratmann, R. E.; Burant, J. C.; Dapprich, S.; Millam, J. M.; Daniels,
A. D.; Kudin, K. N.; Strain, M. C.; Farkas, O.; Tomasi, J.; Barone,
V.; Cossi, M.; Cammi, R.; Mennucci, B.; Pomelli, C.; Adamo, C.;
Clifford, S.; Ochterski, J.; Petersson, G. A.; Ayala, P. Y.; Cui, Q.;
Morokuma, K.; Malick, D. K.; Rabuck, A. D.; Raghavachari, K.;
Foresman, J. B.; Cioslowski, J.; Ortiz, J. V.; Stefanov, B. B.; Liu, G.;
Liashenko, A.; Piskorz, P.; Komaromi, I.; Gomperts, R.; Martin, R.
L.; Fox, D. J.; Keith, T.; Al-Laham, M. A.; Peng, C. Y.; Nanayakkara,
A.; Gonzalez, C.; Challacombe, M.; Gill, P. M. W.; Johnson, B.; Chen,
W.; Wong, M. W.; Andres, J. L.; Gonzalez, C.; Head-Gordon, M.;
Replogle, E. S.; Pople, J. A. GAUSSIAN-98, revision A.7; Gaussian
Inc.: Pittsburgh, PA, 1998.
(25) Becke, A. D. J. Chem. Phys. 1993, 98, 5648.
(26) Lee, C.; Yang, W.; Parr, R. G. Phys. ReV. B 1988, 37, 785.
(27) (a) Dunning, T. H., Jr.; Hay, P. J. In Modern Theoretical Chemistry;
Schaefer, H. F., III, Ed.; Plenum: New York, 1976; Vol. 3, p 1. (b)
Hay, P. J.; Wadt, W. R. J. Chem. Phys. 1985, 82, 270. (c) Hay, P. J.;
Wadt, W. R. J. Chem. Phys. 1985, 82, 284. (d) Hay, P. J.; Wadt, W.
R. J. Chem. Phys. 1985, 82, 299.
(21) Sheldrick, G. M. SADABS: A program for absorption correction with
the Siemens SMART system; University of Gottingen: Go¨ttingen,
Germany, 1996.
(22) SHELXTL program system, version 5.1; Bruker Analytical X-ray
Instruments Inc.: Madison, WI, 1998.
(28) Gorelski, S. I.; Lever, A. B. P.; Ebadi, M. Coord. Chem. ReV. 2002,
230, 97.
7896 Inorganic Chemistry, Vol. 42, No. 24, 2003