Paramagnetic effects on the NMR spectra of "diamagnetic" ruthenium (bis-phosphine)(bis-semiquinone) complexes

Article abstract of DOI:10.1021/ic802302v

Ligand substitution on cis-Ru(PPh₃)2(1,2-O₂C ₆H₄)₂ gives cis-RuL¹L ²(1,2-O₂C₆H₄)2 (L¹ = PPh₃, L² = P(OPh)₃, PBu₃; L ¹ = L² = PBu₃, P(OMe)₃). Syntheses of cis-Ru(PPh₃)₂(3,4-O₂C₆H ₂(5-OH)CO₂Me)₂ and cis-Ru(PPh₃) ₂(AGSQ)₂ (AGSQ = the semiquinone derived from 1,2,3-trihydroxyanthracene-9, 10-dione) are also reported. The upfield chemical shifts and line broadening of the semiquinone 4,5-proton resonances in the NMR spectra indicate that these complexes, while having no detectable magnetic moments, have a weak, temperature-, ligand- and solvent-variable residual paramagnetism, not previously recognized in this series. Density functional theory (DFT) calculations predict a low-lying triplet state, about 14 kJ/mol (0.15 eV) above the singlet ground state. The paramagnetic effects on the NMR spectra are attributed to singlet-triplet equilibria. Temperature dependence of the proton resonances of the semiquinone rings of c/s-Ru(PPh₃) ₂(1,2-O₂C₆H₄)₂ was used to calculate the singlet-triplet free energy difference as 17.5-18.0 kJ/mol in toluene.

Full text of DOI:10.1021/ic802302v

5504 Inorg. Chem. 2009, 48, 5504–5511
DOI: 10.1021/ic802302v
Paramagnetic Effects on the NMR Spectra of “Diamagnetic” Ruthenium
(bis-phosphine)(bis-semiquinone) Complexes
Boris Le Guennic,† Tavon Floyd, Brandon R. Galan, Jochen Autschbach,* and Jerome B. Keister*
Department of Chemistry, University at Buffalo, State University of New York, Buffalo, New York 14260-3000.
† Present address: University of Lyon, Laboratoire de Chimie, UMR 5182 CNRS - Ecole Normale
ꢀ
ꢀ
Superieure de Lyon, 46 allee d’Italie, 69007 Lyon, France
Received December 1, 2008
Ligand substitution on cis-Ru(PPh3)2(1,2-O2C6H4)2 gives cis-RuL1L2(1,2-O2C6H4)2 (L1 = PPh3, L2 = P(OPh)3, PBu3;
L1 = L2 = PBu3, P(OMe)3). Syntheses of cis-Ru(PPh3)2(3,4-O2C6H2(5-OH)CO2Me)2 and cis-Ru(PPh3)2(AGSQ)2
(AGSQ = the semiquinone derived from 1,2,3-trihydroxyanthracene-9,10-dione) are also reported. The upfield
chemical shifts and line broadening of the semiquinone 4,5-proton resonances in the NMR spectra indicate that these
complexes, while having no detectable magnetic moments, have a weak, temperature-, ligand- and solvent-variable
residual paramagnetism, not previously recognized in this series. Density functional theory (DFT) calculations predict a
low-lying triplet state, about 14 kJ/mol (0.15 eV) above the singlet ground state. The paramagnetic effects on the NMR
spectra are attributed to singlet-triplet equilibria. Temperature dependence of the proton resonances of the
semiquinone rings of cis-Ru(PPh3)2(1,2-O2C6H4)2 was used to calculate the singlet-triplet free energy difference
as 17.5-18.0 kJ/mol in toluene.
Introduction
H2O, prepared by reaction of HRu(sal)(PPh3)3 (sal = salicy-
late) with triethylamine and tetrachlorocatechol.3 On the other
hand Pierpont et al. reported the crystal structure of cis-Ru
The electronic structures and redox states of metal com-
plexes of o-quinone, semiquinone, and catecholate ligands
have been of interest for over 30 years.1 The variety of
possible metal and ligand redox states and the resulting
electrochemical and magnetic properties make these systems
quite complicated. In complexes containing two or more
semiquinone ligands spin exchange coupling between the
ligands and with paramagnetic metal ions can result in highly
variable magnetic behavior.
Complexes Ru(PPh3)2(semiquinone)2 (Figure 1), where
the semiquinone ligands are derived from catechol (SQ),
4-t-butylcatechol, 3,5-di-t-butylcatechol (DBSQ), and tetra-
chlorocatechol (Cl4SQ), have been previously reported to
be diamagnetic and have been characterized as RuII(SQ)2
with strong electronic coupling between the semiquinone
ligands.2,3 These complexes have been prepared as both cis
and trans isomers, but the factors determining which isomer
is preferred have not been identified. Chakravorty et al.
(PPh3)2(Cl4SQ)2 CH2Cl2, prepared by reaction of triethyla-
3
mine and tetrachlorocatechol with RuCl2(SQ)(PPh3)2.2
We re-investigated the syntheses and properties of these
bis(semiquinone) complexes, with the goal of preparing
one-dimensional polymers based upon a repeat unit
RuL2(RG-4H) where RG = rufigallol (1,2,3,5,6,7-hexahy-
droxyanthacene-9,10-dione). In the course of this investi-
gation we observed unexpected NMR behavior which
suggested a previously unrecognized paramagnetic compo-
nent. These observations and the syntheses of phosphine
ligand substitution products are reported here.
Experimental Section
RuCl2(PPh3)3,4 Ru(PPh3)2(3,5-DBSQ)2,2 Ru(PPh3)2(SQ)2,3
and anthragallol5 were prepared by previously reported
procedures. Catechol, 3,5-di-t-butylcatechol, and methyl
3,4,5-trihydroxybenzoate were obtained from Aldrich.
P(OMe)3, P(OPh)3, and PBu3 were obtained from Aldrich
and were distilled under reduced pressure before use.
reported the crystal structure of trans-Ru(PPh3)(Cl4SQ)2 2
3
Physical Methods of Characterizations. 1H and 13C NMR
spectra were obtained on Varian Associates Gemini 300 or
Varian Associates VXR-400S instruments. Measurements of
*To whom correspondence should be addressed. E-mail: keister@
buffalo.edu (J.B.K.), jochena@buffalo.edu (J.A.).
(1) (a) Pierpont, C. G.; Lange, C. W. Prog. Inorg. Chem. 1994, 41, 331.
(b) Pierpont, C. G. Coord. Chem. Rev. 2001, 216-217, 99. (c) Pierpont, C. G.
Coord. Chem. Rev. 2001, 219-221, 415.
(2) Bhattacharya, S.; Pierpont, C. G. Inorg. Chem. 1991, 30, 1511–1516.
(3) Bag, N.; Lahiri, G. K.; Basu, P.; Chakravorty, A. J. Chem. Soc.,
Dalton Trans. 1992, 113–117.
(4) Hallman, P. S.; Stephenson, T. A.; Wilkinson, G. Inorg. Synth. 1970,
12, 237–40.
(5) Seuberlich, C. Ber. 1877, 10, 38.
r
2009 American Chemical Society
pubs.acs.org/IC
Published on Web 04/28/2009

Article
Inorganic Chemistry, Vol. 48, No. 12, 2009 5505
the solution was evaporated to dryness, and the residue was
separated by TLC on silica eluting with 4% ethyl acetate in
toluene. One blue band was extracted with ethyl acetate and
evaporated to give 44.5 mg, 81%, of the product. Anal. Calcd
for C48H38O7P2Ru: C, 64.79; H, 4.30. Found: C, 63.95; H, 4.36.
1H NMR (CDCl3): 7.4-6.6 (m, 30 H), 6.33 (d, 1 H, J 8 Hz),
6.21 (d, 1 H, J 8 Hz), 6.19 (d, 1 H, J 8 Hz), 5.98 (d, 1 H J 8 Hz),
5.52 (br, 1 H), 5.19 (br, 1 H), 4.88 (br, 1 H), 4.76 (br, 1 H)
ppm. 31P{1H} NMR (CDCl3): 111.57 (d, J 67 Hz), 37.44 (d, J 67
Hz) ppm.
cis-Ru(PPh3)2(3,5-DBSQ)2. This complex was prepared
as reported previously.2 NMR spectra were recorded for
comparison.
1H NMR (CDCl3): 7.3-7.0 (m, ca. 30 H), symmetrical isomer
- 6.83 (d, 2 H, J 8 Hz), 6.40 (d, J 2 Hz), 1.38 (s), 0.72 (s);
unsymmetrical isomer - 6.77 (br), 6.63 (br), 1.44 (s), 1.43 (s),
1.20 (s), 0.86 (s) ppm. 31P{1H} NMR: symmetrical - 32.07 (s);
unsymmetrical - 31.85 (br), 29.65 (br) ppm. Symmetrical/
unsymmetrical ratio = 0.7.
Figure 1. Structures of complexes and semiquinone ligands.
relaxation times, T1, were done by the inversion recovery
method. 31P NMR spectra were recorded on the VXR-400S
instrument, and chemical shifts are reported relative to
o-phosphoric acid. UV/visible spectra (300-800 nm) were
recorded using a Hewlett-Packard 8452A diode array spectro-
photometer.
cis-Ru(PPh3)2(3,4-O2C6H2(5-OH)CO2Me)2. A mixture
of RuCl2(PPh3)3 (214 mg, 0.223 mmol), methyl 3,4,5-trihydroxy-
benzoate (84 mg, 0.46 mmol), and acetone (10 mL) was placed in
a 50 mL Schlenk flask, equipped with magnetic stir bar, under
an argon atmosphere. Triethylamine (50 μL) was added, and
the color of the solution immediately began to turn blue.
The solution was stirred overnight. The next day solvent was
removed by rotary evaporation. The residue was applied as a
dichloromethane solution to a silica TLC plate. Elution with
10% ethyl acetate in toluene gave a black band, followed by two
blue bands, which were extracted with ethyl acetate. The top
blue band was one of the two symmetrical isomers. The second
blue band was a mixture of the unsymmetrical isomer and the
other symmetrical isomer. Band 1: 20 mg, 9.3% yield. 1H NMR
(CDCl3): 7.3-7.0 (m, 30 H), 6.12 (s, 2 H, T1 0.32 s), 5.27 (s, 2 H,
T1 0.58 s), 3.96 (s, 6 H, T1 0.75 s) ppm. 31P{1H} NMR:
13.79 ppm. Anal. Calcd for C52H42O10P2Ru: C, 63.09; H,
4.28. Found: C, 62.86; H, 4.22. Band 2: 34 mg 15% yield.
1H NMR (CDCl3): 7.3-7.0 (m, 30 H), 6.61 (s), 6.58 (s), 6.15 (s),
5.99 (s), 5.35 (s), 4.00 (s), 3.98 (s), 3.92 (s) ppm. 31P{1H} NMR
(CDCl3): 29.21 (s, OPPh3), 21.22 (s), 18.26 (d, J 31 Hz), 13.80
(d, J 31 Hz) ppm.
Electrochemistry. The electrochemical procedures were
conducted as described previously.6 All potentials are reported
relative to the ferrocene/ferrocenium couple as 0 V.
cis-Ru(PPh3)2(SQ)2. This compound has been previously
reported but was assumed to be the trans isomer.3 The NMR
data were not determined, and these indicate that only the
1
cis isomer is present in solution. H NMR (CDCl3): 7.3-7.0
(m, 30 H), 6.54 (d, 2 Ha, Jad 8 Hz, T1 1.1 s), 6.49 (d, 2 Hb, Jbc 8 Hz,
T1 0.93 s), 5.49 (“t”, 2 Hc, Jbc 8 Hz, Jcd 8 Hz, T1 0.52 s), 4.91 (br t,
2 Hd, Jad 8 Hz, Jcd 8 Hz, T1 0.10 s) ppm. 31P{1H} NMR (CDCl3):
23.53 ppm. 13C{1H} NMR (CDCl3): 179.5, 165.0, 134.7 (o- or m-
C and i-C, PPh3), 132.9, 129.7 ( p-C, PPh3), 128.1 (o- or m-C,
PPh3), 121.9, 118.0, 114.4, 54.0 (CH2Cl2) ppm.
cis-Ru(PBu3)2(SQ)2. A solution of Ru(PPh3)2(SQ)2 (31 mg,
0.037 mmol) and PBu3 (20 μL) in 0.8 mL deuteriochloroform
was monitored by 31P NMR. After 3 days no starting mate-
rial remained, and product mixture contained 42% Ru(PPh3)
(PBu3)(SQ)2 (31P{1H} NMR (CDCl3): 22.8 (d), 8.0 (d) ppm,
J 36 Hz) and 58% Ru(PBu3)2(SQ)2. The latter was separated
by thin layer chromatoraphy (TLC) on silica eluting with
toluene.
Ru(PBu3)2(SQ)2. Anal. Calcd for C36H62O4P2Ru: C, 59.90;
H, 8.66. Found: C, 59.92; H, 8.41. 1H NMR (CDCl3): 6.88 (d,
2 H, J 7 Hz), 6.85 (d, 2 H, J 7 Hz), 6.37 (t, 2 H, J 7 Hz), 6.05 (br t,
2 H, J 7 Hz), 1.6 (m. 12 H), 1.35 (m, 24 H), 0.96 (m, 18 H) ppm.
31P{1H} NMR (CDCl3): 8.23 ppm.
cis-Ru(PPh3)2(AGSQ)2. The same procedure but using
anthragallol gave the product as a brown solid in 22%
yield. The product was recrystallized from dichloromethane/
methanol. UV/vis (dichloromethane): 298 (26000), 352 (15000),
422 (12400), 592 (4300) nm (ε, M-1 cm-1). FAB MS (102Ru):
1134 M+, 880 (M-AGSQ)+. Anal. Calcd for RuP2C64H42O10:
C, 67.78; H, 3.73. Found: C, 67.26; H, 4.12. 1H NMR (CDCl3):
11.92 (s), 11.56 (s), 11.25 (s), 10.95 (s), 8.5-6.5 (m), 5.78 (s),
5.44 (s) ppm. 31P{1H} NMR (CDCl3): -40.9 (br), -45.6 (br),
-66.4 (br), -69.2 (br) ppm.
cis-Ru(P(OMe)3)2(SQ)2. A solution of Ru(PPh3)2(SQ)2
(112 mg, 0.133 mmol) and P(OMe)3 (80 μL, 0.68 mmol) in
25 mL dichloromethane was stirred for 4 days. The solution was
evaporated to dryness, and the residue was separated by TLC
on silica eluting with 8% ethyl acetate in toluene. Three
blue bands eluted, in order: Ru(PPh3)2(SQ)2 (2 mg, 2%), Ru
(PPh3)(P(OMe)3)(SQ)2 (1 mg, 1%), and Ru(P(OMe)3)2(SQ)2
(37 mg, 49%). Anal. Calcd for C18H26O10P2Ru: C, 38.24; H,
4.64. Found: C, 38.79; H, 4.68. 1H NMR (CDCl3): 6.79 (d, 2 H,
J 8 Hz), 6.59 (d, 2 H, J 8 Hz), 5.62 (br “t”, J 8 Hz), 5.53 (br t, 2 H,
J 8 Hz), 3.63 (br, 18 H) ppm. 31P{1H} NMR (CDCl3):
130.26 ppm.
Carbonylation of cis-Ru(PPh3)2(AGSQ)2. A solution of
20.0 mg of cis-Ru(PPh3)2(AGSQ)2 in 10 mL THF was heated at
70-80 °C in a pressure bottle under 2 atm of carbon monoxide.
After 3 h the solution was evaporated to dryness, and the residue
was separated by preparative thin layer chromatography on
silica, eluting with 4% ethyl acetate in dichloromethane. The
bright green band was extracted with ethyl acetate, and the
solvent was removed with vacuum. After transfer to a tared
vial as a dichloromethane solution, evaporation yielded
19
12.3 mg (75%) Ru(CO)2(PPh3)2(AG-2H) (AG-2H refers to
the catecholate dianion derived from anthragallol). The product
was characterized by comparison of the IR and 1H and 31P
NMR spectra to those of an authentic sample.
cis-Ru(PPh3)(P(OPh)3)(SQ)2. A solution of Ru(PPh3)2-
(SQ)2 (52 mg, 0.062 mmol) and P(OPh)3 (140 mg, 0.45 mmol)
in 10 mL of dichloromethane was stirred under argon. After 4 days
Computational Details
Density functional theory (DFT) computations were
performed with the Amsterdam Density Functional program
(6) Churchill, M. R.; Keil, K. M.; Bright, F. B.; Pandey, S.; Baker, G. A.;
Keister, J. B. Inorg. Chem. 2000, 39(25), 5807–5816.

5506 Inorganic Chemistry, Vol. 48, No. 12, 2009
Le Guennic et al.
package (ADF).7 All calculations employed the scalar
relativistic zeroth-order regular approximation (ZORA)
computational model. Therefore, the discussion is based on
the gas phase calculations.
8,9
and the revised Perdew-Burke-Ernzerhof (revPBE) func-
tional.10,11 Geometries were optimized using the valence
triple-ζ polarized Slater-type basis TZP optimized for ZORA
calculations (ZORA/TZP) from the ADF basis set database.
Default settings were applied except for the geometry
convergence (gradient convergence of 1 ꢀ 10-4) and the
numerical integration accuracy (parameters were set as 5.0
for “accint” and 6.0 for “accsph”. The NMR computations
were carried out with these integration grids, too). For the
PH3 model complex, the optimized closed shell ground state
and the excited triplet state structure were confirmed by
frequencies calculations. Apart from very low frequency
imaginary modes (around 20 cm-1) for the PH3 rotations,
all harmonic vibrations had real positive frequencies.
The PH3 hindered rotation wavenumber is within the accu-
racy of the numerical differentiation used to compute the
normal modes and inconsequential for the results since these
ligands would undergo quasi free rotations at the experimen-
tal temperatures. Singlet and triplet excitation energies were
subsequently calculated by using time-dependent DFT
(TDDFT) linear response theory,12-14 as well as with a
Δ-SCF approach.15 Here, we have employed the doubly
polarized ZORA basis TZ2P for Ru and the TZP basis
for all other atoms. For the TDDFT computations an inte-
gration accuracy parameter of 8.0 was used which roughly
corresponds to the number of significant figures for the
integration of the electron density. In general, the results
obtained from TDDFT and Δ-SCF to calculate excitation
energies were in reasonable agreement with each other. To
check for possible influence from the solvent, additional
optimizationsofthesingletgroundstateandthelowesttriplet
state were performed for the PH3 model system, using
the “COSMO” continuum solvation model with dielectric
constants for toluene, chloroform, and acetone. A reduction
of the singlet-triplet energy splitting from 0.163 to 0.135 eV
when going from gas phase to the most polar solvent model,
acetone, was calculated. However, this solvent model does
nottakeintoaccountspecifichydrogenbondingeffectswhich
might be important to describe the overall effect from the
solvent, and the solvent-polarity induced changes do not
exceed the effects from changing the R-groups on the
phosphine ligands nor do they exceed other remaining
uncertainties in the excitation energies caused by the
Results
Using a modification of procedures reported in the litera-
ture2 we found that reaction of RuCl2(PPh3)3 with 2 equiv
of the appropriate catechol and excess NEt3 in acetone gives
the corresponding cis-Ru(PPh3)2(semiquinone)2 in good
yields.
The NMR characterizations of complexes previously
prepared were not fully reported. Chakravorty et al. prepared
Ru(PPh3)2(SQ)2, among other semiquinone complexes, but
did not report NMR data.3 Bhattacharya and Pierpont noted
that the NMR spectrum of Ru(PPh3)2(DBSQ)2 contains
three sets of t-Bu resonances, in a 2:1.3:1 ratio, characterized
as cis-isomers, and did not comment on the other DBSQ
resonances or the 31P NMR data.2
The reactions of catechol, triethylamine, and either
RuCl2(PPh3)3 (Pierpont’s conditions2) or HRu(salicylate)
(PPh3)3 (Chakravorty’s conditions3) in acetone give the
same product, Ru(PPh3)2(SQ)2. The NMR spectra for
Ru(PPh3)2(SQ)2 clearly indicate only the cis isomer in
solution. The 31P NMR spectrum displays a single resonance.
The 13C{1H}NMR spectrum displays six resonances for the
semiquinone ligands, as expected for the C2 symmetric cis
geometry and inconsistent with the trans geometry. The 1H
NMR spectrum displays four resonances for the SQ ligands,
as expected for the cis isomer. However, the appearance of
the spectrum is unexpected. The 4,5-proton resonances are
shifted upfield, and all four SQ proton resonances have
different line-widths, with the 4- and 5-proton resonances
broader than the 3- and 6-proton resonances. In deuterio-
chloroform the four SQ proton resonances appear at 6.54
(br d, T1 1.1 s), 6.49 (d, T1 0.93 s), 5.49 (t, T1 0.52 s), and 4.91
(br t, T1 0.10 s) ppm (Figure 2). Correlation spectroscopy
(COSY) shows the coupling of the lowest field doublet to
the highest field triplet. The T1 measurements found a
much shorter relaxation time for the most upfield resonance.
This suggests that there is some residual paramagnetism
associated with the complex.
Other phosphine complexes can be prepared by ligand
substitution. We have found that cis-Ru(PBu3)2(SQ)2 can be
readily prepared by PBu3 substitution on cis-Ru(PPh3)2(SQ)2
at room temperature. The 4- and 5-proton resonances of this
complex arenot shifted as far upfield and are sharper than the
corresponding resonances for cis-Ru(PPh3)2(SQ)2. The 31P
{1H} NMR spectrum of the intermediate Ru(PPh3)(PBu3)
(SQ)2 displays a pair of doublets with a coupling constant of
39 Hz. Ligand exchange was also used to prepare cis-
Ru(P(OMe)3)2(SQ)2 and cis-Ru(PPh3)(P(OPh)3)(SQ)2.
The complex cis-Ru(PPh3)2(3,5-DBSQ)2 (DBSQ = di-
t-butylsemiquinone) has three possible isomers because
of the orientations of the unsymmetrical DBSQ ligands.
The NMR spectra are most consistent with the presence
of two of the three isomers, one symmetrical (two t-butyl
resonances and a single 31P signal) and the other unsymme-
trical (four t-butyl resonances and two broad doublet 31P
resonances); under these conditions the resonance for
free PPh3 is sharp, as expected since phosphine exchange
occurs over a period of hours. Again the broadness of the
ring proton resonances suggests a previously unrecognized
paramagnetism.
(7) Baerends, E. J.; Autschbach, J.; Berces, A.; Bo, C.; Boerrigter, P. M.;
Cavallo, L.; Chong, D. P.; Deng, L.; Dickson, R. M.; Ellis, D. E.; Fan, L.;
Fischer, T. H.; Guerra, C. F.; van Gisbergen, S. J. A.; Groeneveld,
J. A.; Gritsenko, O. V.; Gruning, M.; Harris, F. E.; van den Hoek, P.;
Jacobsen, H.; van Kessel, G.; Kootstra, F.; van Lenthe, E.; Osinga,
V. P.; Patchkovskii, S.; Philipsen, P. H. T.; Post, D.; Pye, C. C.; Ravenek,
W.; Ros, P.; Schipper, P. R. T.; Schreckenbach, G.; Snijders, J. G.; Sola, M.;
Swart, M.; Swerhone, D.; Velde, G. t.; Vernooijs, P.; Versluis, L.; Visser, O.;
van
Wezenbeek,
E.;
Wiesenekker,
G.;
Wolff,
S. K.; Woo, T. K.; Ziegler, T. Amsterdam Density Functional
program (ADF); Vrije Universiteit: Amsterdam, 2004; URL:
http://www.scm.com.
(8) van Lenthe, E.; Baerends, E. J.; Snijders, J. G. J. Chem. Phys. 1993, 99,
4597–4610.
(9) Lenthe, E. v.; Ehlers, A.; Baerends, E. J. J. Chem. Phys. 1999, 110(18),
8943–8953.
(10) Perdew, J. P.; Burke, K.; Ernzerhof, M. Phys. Rev. Lett. 1996, 77,
3865–3868.
(11) Hammer, B.; Hansen, L. B.; Norskov, J. K. Phys. Rev. B 1999, 59,
7413–7421.
(12) van Gisbergen, S. J. A.; Snijders, J. G.; Baerends, E. J. Comput. Phys.
Commun. 1999, 118, 119–138.
(13) van Gisbergen, S. J. A.; Fonseca-Guerra, C.; Baerends, E. J. J.
Comput. Chem. 2000, 21(16), 1511–1523.
(14) Autschbach, J.; Ziegler, T. Coord. Chem. Rev. 2003, 238/239, 83–126.
(15) Ziegler, T.; Rauk, A.; Baerends, E. J. Theor. Chim. Acta 1977, 43,
261–271.

Article
Inorganic Chemistry, Vol. 48, No. 12, 2009 5507
Figure 2. 1H NMR spectra of Ru(PPh3)2(SQ)2 in deuteriochloroform (upper, ꢀ marks co-crystallized dichoromethane) and d6-benzene (lower).
Fox and Pierpont have reported a similar phenomenon for
the 1H NMR spectra of Pd(DBSQ)2.16 Temperature-depen-
dent chemical shifts, larger for resonances associated with the
4,5-ring positions, were attributed to weak SQ-SQ exchange
which results in low-level paramagnetism, even though
the compound has no paramagnetism detectable by magnetic
susceptibility measurements. A plot of the chemical shift of
the 5-t-Bu group versus 1/T was termed “reasonably linear.”
It was observed that SQ-SQ exchange was much greater
for the analogous Pt complex, which does not display
temperature-dependent spectra. Temperature-dependent
spectra for diamagnetic Cr(DBSQ)3 also suggested that this
complex had low-level paramagnetism.
found that the 4,5-ring proton resonances shift upfield
by 0.73 and 0.95 ppm as the temperature is increased from
-10 °C to +50 °C. This study also revealed a significant
solvent effect on the chemical shifts and line-widths of the
4,5-ring proton resonances, with shifts for the highest field
proton resonance at 6.14 ppm (sharp triplet) in d6-benzene
at 21°C (Figure 2) compared with 4.91 ppm (broad)
in deuteriochloroform, 5.68 ppm in d2-dichloromethane,
5.36 ppm in 5% d4-methanol/95% d2-dichloromethane,
5.06 ppm in 15% d4-methanol/85% d2-dichloromethane
(the chemical shifts are the same in 30% d4-methanol/70%
d2-dichloromethane), 5.88 ppm (sharp triplet) in carbon
tetrachloride, and 5.83 ppm (sharp triplet) in d6-acetone.
The solvent shifts do not correlate with polarity (e.g.,
E values17 for benzene (34.5), carbon tetrachloride (32.5),
We therefore examined the temperature dependence of the
1H NMR chemical shifts of the SQ protons of cis-Ru-
(PPh3)2(SQ)2. Initial experiments in d8-toluene solution
(17) (a) Reichardt, C.; Dimroth, K. Fortschr. Chem. Forsch. 1968, 11, 1.
(b) Fowler, F. W.; Katritzky, A. R.; Rutherford, R. J. D. J. Chem. Soc., B
1971, 460.
(16) Fox, G. A.; Pierpont, C. G. Inorg. Chem. 1992, 31, 3718–3723.

5508 Inorganic Chemistry, Vol. 48, No. 12, 2009
Le Guennic et al.
31P NMR resonances were unexpectedly broad and shifted
upfield. To further support the identification of the product
as Ru(PPh3)2(AGSQ)2, a sample was heated at 70-80 °C in
THF under 2 atm of carbon monoxide, yielding the known
complex Ru(PPh3)2(CO)2(AG-2H)19 (74% yield after chro-
matography). This carbonylation product, in combination
with the mass spectral data, supports the proposed ligand
environment.
The cyclic voltammogram of Ru(PPh3)2(AGSQ)2 displays
two nearly reversible 1-e oxidations at +0.42 V (ΔEp 60 mV)
and -0.42 V (74 mV), a nearly reversible 1-e reduction at
-1.15 V (ΔEp 64 mV) and a nearly irreversible 1-e reduction
at -1.8 V (ΔEp 170 mV). This compares with two reversible
oxidations (+0.41 and -0.045 V) and two reductions (-1.05
and -1.79 V) reported for Ru(PPh3)2(DBSQ)2.
The unusual NMR behavior of these complexes prompted
us to undertake a computational study to determine if a low-
energy triplet state could account for the unexpected NMR
behavior.
Figure 3. Plot of chemical shifts of highest field proton resonance of the
SQ rings for Ru(PPh3)2(SQ)2 versus the Kamlet-Taft solvent parameter
R for the solvent (toluene (0), carbon tetrachloride (0), acetone (0.08),
dichloromethane (0.13), chloroform (0.20)).
acetone (42.2), chloroform (39.1), and dichloromethane
(41.1)), but do correlate reasonably well with the Kamlet-
Taft solvent parameter R,18 which is associated with the
hydrogen bond donor ability of the solvent (Figure 3).
The order of upfield shift of the 4,5-proton resonances for
the RuL2(SQ)2 complexes, L2 = (PBu3)2 < (P(OMe)3)2 <
(PPh3)2 < (PPh3)(P(OPh)3), does not correlate in any
straightforward fashion with the sigma donor/pi acceptor
properties of L. However, the broadness of the NMR
resonances also depends on the semiquinone ligands, with
the line width for the semiquinone aromatic proton reso-
nances of cis-Ru(PPh3)2(DBSQ)2 less than that for cis-
Ru(PPh3)2(SQ)2.
The reaction of anthragallol, RuCl2(PPh3)3, and triethyla-
mine produced a brown crystalline solid, characterized as
cis-Ru(PPh3)2(AGSQ)2, where AGSQ is the semiquinone
derived from anthragallol (1,2,3-trihydroxyanthracene-
9,10-dione). The FAB mass spectrum contains the molecular
ion and a fragmentation pattern consistent with the composi-
tion Ru(PPh3)2(AGSQ)2. The 1H NMR spectrum contained
unresolved aromatic proton resonances and four hydroxyl
resonances suggesting a mixture of isomers. The AGSQ
ligand most likely chelates through the 2- and 3-oxygens, as
was shown previously for Ru(PPh3)2(CO)2(AG-2H) (AG-2H
refers to the catecholatedianion derivedfrom anthragallol).19
In this coordination mode, as for cis-Ru(PPh3)2(DBSQ)2,
one would expect three possible isomers, two symmetrical
with a single 31P and hydroxyl proton resonance each and a
single unsymmetrical isomer with two 31P and two hydroxyl
proton resonances. Consistent with this, the NMR spectrum
displays very broad 31P resonances at -40.9 (br), -45.6 (br),
-66.4 (br), -69.2 (br) ppm, but these are far upfield from
the resonances for cis-Ru(PPh3)2(SQ)2, cis-Ru(PPh3)2
(3,4-O2C6H2(5-OH)CO2Me)2, and cis-Ru(PPh3)2(DBSQ)2.
This observation suggests a higher degree of residual para-
magnetism for this complex. However, magnetic susceptibil-
ity measurements were unable to detect any paramagnetism
at 300 K in the solid state. The Evans method was also
unsuccessful for detecting paramagnetism of a saturated
solution in dichloromethane, but the low solubility of the
complex is problematic for this method.
Our focus has been on the lowest triplet excitation energy
to determine a trend among the series of complexes. Calcula-
tions were performed on cis-Ru(PH3)2(SQ)2, cis-Ru
(PPh3)2(SQ)2, cis-Ru(PMe3)2(SQ)2, cis-Ru(P(OMe)3)2(SQ)2,
and cis-Ru(PMe3)2(DHAQ)2 (DHAQ = 2,3-oxo-1,4-dihy-
droxyanthracene-9,10-dione, chosen as a surrogate for Ru
(PPh3)2(AGSQ)2). For the vertical excitations, an analysis
of the TDDFT density response vector showed that the
lowest triplet excitation corresponds in all cases to a simple
HOMO-LUMO transition (99% contribution to the transi-
tion density). We compared the TDDFT results to those
obtained with a Δ-SCF approach to test for some potential
shortcomings of the approximate exchange-correlation ker-
nels applied20 (in the present work we used the adiabatic local
density approximation (ALDA)21 kernel).
The calculations (both TDDFT and Δ-SCF) clearly
demonstrate that the series of complexes has very low lying
excited triplet states. Based on the C2 point group of the
complexes that was adopted in the computations, the
calculated lowest triplet state is of B symmetry because of a
HOMO (b-symmetry)-to-LUMO (a-symmetry) transition.
The vertical triplet excited states (Table 1) are only 0.43 to
0.37 eV above the ground state (Δ-SCF data. For TDDFT
we obtain 0.45 to 0.39 eV). The lowest singlet excited states
are considerably higher, between 0.89 to 0.97 eV (Δ-SCF.
In TDDFT: 0.95 to 1.25 eV). The agreement between
Δ-SCF and TDDFT is quite reasonable, in particular
for the triplet energies, which might be taken as an indication
that the computations do not suffer substantial errors.
Further details about the nature of the triplet states can
be found in the Supporting Information. The molecular
orbitals (MOs) indicate a significant interaction of the
singly occupied semiquinone ligand orbitals with the
metal center and with each other. The ground state is a closed
shell singlet state in which the lower lying linear combination
of the mainly ligand-centered singly occupied orbitals
is doubly occupied. The lowest triplet state corresponds
to the triplet diradical, with some involvement of the
metal center, while the open shell singlet diradical electronic
The spectroscopic characterization of product as Ru-
(PPh3)2(AGSQ)2 was ambiguous, particularly since the
(18) (a) Taft, R. W.; Kamlet, M. J. J. Am. Chem. Soc. 1976, 98, 2886.
(b) Lagalante, A. F.; Jacobson, R. J.; Bruno, T. J. J. Org. Chem. 1996, 61,
6404.
(19) Churchill, M. R.; Keil, K. M.; Gilmartin, B. P.; Schuster, J. J.;
Keister, J. B.; Janik, T. S. Inorg. Chem. 2001, 40, 4361–4367.
(20) Casida, M. E.; Gutierrez, F.; Guan, J.; Gadea, F.-X.; Salahub, D. R.;
Daudey, J.-P. J. Chem. Phys. 2000, 113(17), 7062–7071.
(21) Bauernschmitt, R.; Ahlrichs, R. Chem. Phys. Lett. 1996, 256, 454–
464.

Article
Inorganic Chemistry, Vol. 48, No. 12, 2009 5509
Table 1. Δ-SCF Vertical Triplet Excitation Energies for Model Complexes, and Some Experimental NMR Characteristics of the Full Systems
complex
calcd ET - EG (eV) (vertical)
complex
δ for highest field 4,5-proton resonance (CDCl3)
Ru(PH3)2(SQ)2
0.430
0.423
0.413
0.403
0.369
Ru(PMe3)2(SQ)2
Ru(P(OMe)3)2(SQ)2
Ru(PPh3)2(SQ)2
Ru(PBu3)2(SQ)2
Ru(P(OMe)3)2(SQ)2
Ru(PPh3)2(SQ)2
6.05, sharp 31P resonances
5.53, sharp 31P resonances
4.91, sharp 31P resonances
NA - very broad 31P resonances
Ru(PMe3)2(DHAQ)2
Ru(PPh3)2(AGSQ)2
structure as described by the lowest TDDFT and Δ-SCF
excitations is significantly higher in energy.
The vertical singlet-triplet energy gap (Table 1) varies in
a way consistent with the observed NMR behavior. The
Table 2. Calculated Adiabatic Triplet Excitation Energies (eV)
E(optimized triplet) - E(optimized singlet) (eV)
Ru(PH3)2(SQ)2
Ru(PMe3)2(SQ)2
Ru(PPh3)2(SQ)2
Ru(POMe3)2(SQ)2
Ru(PMe3)2(DHAQ) 2
0.163
0.132
0.147
0.170
0.159
singlet-triplet gap decreases as Ru(PH3)2(SQ)2
Ru(PMe3)2(SQ)2 > Ru(P(OMe)3)2(SQ)2 > Ru(PPh3)2(SQ)2 >
Ru(PMe3)2(DHAQ)2, and the NMR spectra suggest increas-
ing paramagnetic effects as Ru(PBu3)2(SQ)2 < Ru(P
(OMe)3)2(SQ)2 < Ru(PPh3)2(SQ)2 < Ru(PPh3)2(AGSQ)2.
We have subsequently optimized the geometries of the
Δ-SCF triplet states and obtained considerable stabilization
(Table 2). The optimized triplet states are now less than
0.2 eV above the ground state.
>
between diamagnetic, square planar and paramagnetic,
tetrahedral species.27 The temperature dependence of the
chemical shifts was used to determine ΔG° for the singlet-
triplet equilibrium and the hyperfine coupling constants in
the paramagnetic complexes.
In the triplet state the NMR resonances will be strongly
affected by the electron spin, mainly because of the contact
shift. If we consider only the contact hyperfine interaction,
the contact shift δC is given by eq 1 below, where a is the
isotropic hyperfine coupling constant, β and βN are the Bohr
and nuclear magnetons, respectively, g and gN are the
electronic and nuclear g-values, respectively, and (2S+1) is
the electronic spin multiplicity:28,29
Discussion
Complexes containing two or more semiquinone ligands
can display different magnetic properties, depending on
the degree of electronic coupling between the ligands and
between the ligands and unpaired electrons of the metal
center.22 For example, Cr(Cl4SQ)3 is paramagnetic at room
temperature and the magnetic moment drops to near-zero
at low temperature,35 whereas Cr(DBSQ)3 appears to be
diamagnetic at room temperature but with residual para-
magnetism causing broadening of the NMR resonances.16
Neutral complexes cis- and trans-RuL2(SQ)2, where L =
nitrogen donor, phosphine, or CO, have been studied by
several research groups. All complexes have been reported
to be diamagnetic. The complexes Ru(bpy)(SQ)2, and trans-
RuL2(SQ)2, where L = nitrogen donor, were proposed to
have a fully delocalized mixed-valence state best described as
RuIIIL2(catecholate)(semiquinone).23-25 On the other hand
cis-Ru(CO)2(3,6-DBSQ)2 and cis-Ru(CO)2(PhenoxSQ)2
were characterized as RuII(semiquinone)2 redox states with
antiferromagnetic coupling giving rise to diamagnetism and
sharp NMR resonances.26 The complexes Ru(PPh3)2(SQ)2
were describedas diamagnetic with redox statesbest regarded
as RuII(semiquinone)2.2,3
gβ aSðS þ 1Þ
δC
¼
ð1Þ
gNβN
3kT
The hyperfine coupling constants in semiquinone radical
anions have been reported.30,31 Although the values are
somewhat dependent on the cation and solvent, typical
values are about -0.5 G for the 3,6-hydrogens and -4 G
for the 4,5-hydrogens. Since the unpaired electron spin
density in the semiquinone radical is greater at the 4- and
5-positions than at the 3- and 6-positions, this accounts for
the shorter T1’s for the 4- and 5-protons of the semiquinone
ligands. The calculated contact shift the 4,5-H resonance of
the triplet would be upfield by about 300 ppm at 300 K,
assuming a hyper-fine coupling constant a of -4 G and S =
1/2 for a semiquinone radical anion. The observed shift for
Ru(PPh3)2(SQ)2 is upfield on the order of 1 ppm or less. If the
observed shift is due to the population-weighted, exchange
averaged shift of singlet and triplet, one can estimate that the
We propose that the unusual NMR spectral behavior of
cis-RuL2(semiquinone)2 complexes is due to thermally ac-
cessible triplet states. The diamagnetism of the ground state
arises from antiferromagnetic coupling of the semiquinone
ligands, which are ferromagnetically coupled in the triplet
state. The upfield shifts of the proton resonances as the
temperature increases are explained by the greater triplet
population at higher temperatures.
(27) (a) Phillips, W. D.; Benson, R. E. J. Chem. Phys. 1960, 33, 607.
(b) Benson, R. E.; Eaton, D. R.; Josey, A. D.; Phillips, W. D. J. Am. Chem.
Soc. 1961, 83, 3714–3716. (c) Eaton, D. R.; Josey, A. D.; Benson, R. E.;
Phillips, W. D.; Cairns, T. L. J. Am. Chem. Soc. 1962, 84, 4100–4106. (d)
Eaton, D. R.; Josey, A. D.; Phillips, W. D.; Benson, R. E. Mol. Phys.. 1962, 5,
407. (e) Eaton, D. R.; Josey, A. D.; Phillips, W. D; Benson, R. E. J. Chem.
Phys.. 1962, 37, 347. (f ) Holm, R. H.; Chakravorty, A.; Dudek, G. O. J. Am.
Chem. Soc. 1964, 86, 379–387. (g) Eaton, D. R.; Phillips, W. D. J. Chem.
Phys. 1965, 43, 392.
The magnetic and NMR spectral properties of Ni(II)
complexes have previously been analyzed in terms of equilibria
(28) Carrington, A.; McLachlan, A. D. Introduction to Magnetic Reso-
nance with Applications to Chemistry and Chemical Physics; Harper & Row:
New York, 1967.
(22) Pierpont, C. G.; Attia, A. S. Collect. Czech. Chem. Commun. 2001,
66, 33–51.
(29) Moon, S.; Patchkovskii, S. In Calculation of NMR and EPR Para-
meters. Theory and Applications; Kaupp, M., Buhl, M., Malkin, V. G., Eds.;
Wiley-VCH: Weinheim, 2004; pp 325-338.
(23) Auburn, P. R.; Dodsworth, E. S.; Haga, M.; Liu, W.; Nevin, W. A.;
Lever, A. B. P. Inorg. Chem. 1991, 30, 3502.
(24) Lever, A. P. B.; Auburn, P. R.; Dodsworth, E. S.; Haga, M.; Liu, W.;
Melnik, M.; Nevin, W. A. J. Am. Chem. Soc. 1988, 110, 8076.
(25) Bhattacharya, S.; Pierpont, C. G. Inorg. Chem. 1991, 30, 1511.
(26) Bhattacharya, S.; Pierpont, C. G. Inorg. Chem. 1994, 33, 6038.
(30) Langgard, M.; Spanget-Larsen, J. J. Mol. Struct., THEOCHEM
1998, 431, 173–180.
(31) Felix, C. C.; Sealy, R. C. J. Am. Chem. Soc. 1982, 104, 1555–1560.

5510 Inorganic Chemistry, Vol. 48, No. 12, 2009
Le Guennic et al.
equilibrium constant is about 3.3 ꢀ 10-3, or the triplet state is
about 14 kJ/mol higher in energy than the singlet state. This
estimate is in good agreement with the optimized singlet-
optimized triplet energy gap of 14.2 kJ/mol (0.147 eV).
The temperature dependence of the chemical shifts is thus
due to both the temperature dependence of the contact shift
(eq 1), which decreases with increasing temperature, and
the equilibrium population of the triplet (S = 1, triply
degenerate), which increases with temperature. Combining
the two, the chemical shift is given by eq 2:27,32
gβe aSðS þ 1Þ
δðppmÞ ¼ δsinglet þ 106
f3 þ
gNβN
kT
Figure 4. Temperature dependence of chemical shifts of protons at
4- and 5-positions (curves [1] and [2]) of the SQ rings for Ru(PPh3)2(SQ)2
in d8-toluene.
-1
expððEtriplet -EsingletÞ=RTÞg
ð2Þ
where g is the electronic g value for the triplet (assumed to be
2), δsinglet is the chemical shift (ppm) of the resonance in
the singletstate, a is the hyperfine coupling constant (negative
sign) for the unpaired electron with the proton in the triplet
state, and (Etriplet - Esinglet) is the optimized singlet-optimized
triplet energy gap (approximately equal to ΔG° since ΔV°
and ΔS ° should be small). The result is not linear with either
temperature-1 or ln(temperature).
semiquinone rings, compared to the singlet state. The fact
that the chemical shift of the most upfield ring proton
resonance is about the same in 15% d4-methanol/85%
d2-dichloromethane and in 30% d4-methanol/70% d2-di-
chloromethane suggests that saturation of the hydrogen
bonding is achieved in the former solvent mixture and
perhaps in deuteriochloroform since the chemical shift in
that solvent is almost the same.
The theoretical calculations allow us to predict the chemi-
cal shifts for singlet states as 6.7-7.0 ppm for H4 and H5.
Plots [1] and [2] in Figure 4 display the experimentally
determined (in d8-toluene) chemical shifts of the H4 and H5
proton resonances of cis-Ru(PPh3)2(SQ)2, with the curves
representing the best fits to eq 2, witha, (Etriplet - Esinglet), and
the value of the singlet chemical shift as adjustable para-
meters. The best fits give δsinglet = 6.94(0.02) and 6.90(0.02)
ppm, (Etriplet - Esinglet) = 17.5(0.3) and 18.0(0.5) kJ/mol and
a = -1.7(0.2) and -1.5(0.2) G (error limits are one standard
deviation). Additional calculations were performed which
estimated the hyperfine coupling constants for the 4- and 5-H
nuclei in the triplet states of the complexes as -0.8 to -1.3 G
(-2.2 to -3.6 MHz). The calculated curves are clearly
consistent with the observed data, and the best-fit parameters
for (Etriplet - Esinglet) and a are in good agreement with the
theoretically calculated values. Of course, chemical shifts in
clearly diamagnetic species are subject to solvent and tem-
perature variations, but the NMR behavior observed for
these complexes is best explained by exchange between
diamagnetic and paramagnetic complexes.33 The singlet-
triplet energy gap appears to be affected by the phosphine
ligand, by the semiquinone ring substituents, and by hydro-
gen bond donor solvents.
The degree of antiferromagnetic coupling also depends on
the semiquinone ligands. Qualitatively, the NMR spectra
suggest that the contribution of the triplet state increases in
the series DBSQ < SQ < AGSQ. This observation is
consistent with the semiquinone ligand dependence of the
magnetic properties of Fe(SQ)3 complexes.35 Strong antifer-
romagnetic exchange interaction between the SQ-based
unpaired electrons and the FeIII unpaired d electrons would
create a triplet ground state with two unpaired electrons.
Weak exchange interaction gives rise to thermally accessible
excited states of higher spin multiplicity. Thus, the magnetic
moments of these complexes are temperature- and ligand
dependent. A study of the magnetic susceptibility of Fe
(semiquinone)3 found the relative degree of intramolecular
antiferromagnetic coupling to decrease as 3,5-DBSQ >
Cl4SQ > phenSQ; this trend was attributed to reduction of
unpaired electron density on the oxygens (because of electron
withdrawing ring substituents or increased delocalization in
a
larger aromatic ring), causing reduced antiferro-
magnetic interaction between the ligands through the metal
bridge. The same argument can be made for Ru(PPh3)2
(semiquinone)2.
Finally, the singlet-triplet equilibrium also depends on
the phosphine ligands. The order of upfield shift of the 4,5-
proton resonances for the cis-RuL2(SQ)2 complexes, L2 =
(PBu3)2 < (P(OMe)3)2 < (PPh3)2 < (PPh3)(P(OPh)3) does
not correlate in any straightforward fashion with the sigma
donor/pi acceptor properties of L since the order of sigma
donor/pi capability decreases in the series PBu3 > PPh3 >
P(OMe)3. However, the calculated vertical singlet-triplet
The solvent effects suggest that the triplet state is stabilized
relative to the ground state by hydrogen bonding.34 Hydro-
gen bonding between the solvent and the oxygen atoms of the
semiquinone ligands might shift electron density away from
the metal onto the semiquinone rings, stabilizing the triplet
state which may have greater electron localization in the
energy gap decreases slightly as Ru(PMe3)2(SQ)2
>
(32) Some previous analyses (ref 27) of NMR shifts due to diamagnetic-
paramagnetic equilibria replace the isotropic hyperfine coupling constant
a with a/2S. Equation (2) is as used in ref 27(g).
Ru(P(OMe)3)2(SQ)2 > Ru(PPh3)2(SQ)2; the optimized
singlet-optimized triplet gaps are all small, and the small
differences in energies for the various complexes may already
(33) Indeed, analysis of the published16 temperature dependence of
chemical shift data for Pd(DBSQ)2 according to eq 2 gives a much superior
fit to the data (see Supporting Information) than does a fit to inverse
temperature.
(34) Hyperfine coupling constants are also affected by solvent but the
(35) Buchanan, R. M.; Kessel, S. L.; Downs, H. H.; Pierpont, C. G.;
variations are not large.
Hendrickson, D. N. J. Am. Chem. Soc. 1978, 100, 7894.

Article
Inorganic Chemistry, Vol. 48, No. 12, 2009 5511
be exceeding the accuracy limits of the calculations.
Differential solvation may be a factor accounting for the
differences among these complexes.
Colorado. J.A. acknowledges support from the National
Science Foundation (CHE 0447321).
Supporting Information Available: Computational results.
Variable temperature H NMR spectra of cis-Ru(PPh3)2(SQ)2
1
Acknowledgment. We acknowledge research contribu-
tions from undergraduate students Tzu-chuan Cheng and
Daisuke Ishio and helpful discussions with Dr. Frank V.
Bright and Dr. David F. Watson of the University at
Buffalo, and Dr. Cortlandt Pierpont of the University of
in d8-toluene between -10 and +50 °C. Details for fitting the
temperature dependence of the chemical shifts to eq 2 and
analysis of the published16 temperature dependence of chemical
shift data for Pd(DBSQ)2. This material is available free of
charge via the Internet at http://pubs.acs.org.