Electronic structure of square planar bis(benzene-1,2-dithiolato)metal complexes [M(L)2]z (z = 2-, 1- 0; M = Ni, Pd, Pt, Cu, Au): An experimental, density functional, and correlated ab initio study

Article abstract of DOI:10.1021/ic0507565

The three diamagnetic square planar complexes of nickel(II), palladium(II), and platinum(II) containing two S,S-coordinated 3,5-di-tert-butylbenzene-1,2- dithiolate ligands, (L^Bu)^2-, namely [M^II(L ^Bu)₂]^2-, have been synthesized. The corresponding paramagnetic monoanions [M^II(L^Bu)(L ^Bu?)]^- (S = 1/2) and the neutral diamagnetic species [M^II(L^Bu?)₂] (M = Ni, Pd, Pt) have also been generated in solution or in the solid state as [N(n-Bu)₄][M ^II(L^Bu)(L^Bu?)] salts. The corresponding complex [Cu^III(L^Bu)₂] has also been investigated. The complexes have been studied by UV-vis, IR, and EPR spectroscopy and by X-ray crystallography; their electro- and magnetochemistry is reported. The electron-transfer series [M(L^Bu)₂] ^2-,-,0 is shown to be ligand based involving formally one (L ^Bu?)^- π radical in the monoanion or two in the neutral species [M^II(L^Bu?)₂] (M = Ni, Pd, Pt). Geometry optimizations using all-electron density functional theory with scalar relativistic corrections at the second-order Douglas-Kroll-Hess (DKH2) and zeroth-order regular approximation (ZORA) levels result in excellent agreement with the experimentally determined structures and electronic spectra. For the three neutral species a detailed analysis of the orbital structures reveals that the species may best be described as containing two strongly antiferromagnetically interacting ligand radicals. Furthermore, multiconfigurational ab initio calculations using the spectroscopy oriented configuration interaction (SORCI) approach including the ZORA correction were carried out. The calculations predict the position of the intervalence charge-transfer band well. Chemical trends in the diradical characters deduced from the multiconfigurational singlet ground-state wave function along a series of metals and ligands were discussed.

Full text of DOI:10.1021/ic0507565

Inorg. Chem. 2005, 44, 5345−5360
Electronic Structure of Square Planar Bis(benzene-1,2-dithiolato)metal
Complexes [M(L)2]z (z
, 1 , 0; M Ni, Pd, Pt, Cu, Au): An
Experimental, Density Functional, and Correlated ab Initio Study
)
2−
−
)
Kallol Ray, Thomas Weyhermu1ller, Frank Neese,* and Karl Wieghardt*
Max-Planck-Institut fu¨r Bioanorganische Chemie, Stiftstrasse 34-36,
D-45470 Mu¨lheim a. d. Ruhr, Germany
Received May 12, 2005
The three diamagnetic square planar complexes of nickel(II), palladium(II), and platinum(II) containing two S,S-
2
coordinated 3,5-di-tert-butylbenzene-1,2-dithiolate ligands, (LBu
)
-, namely [MII(LBu)2]2-, have been synthesized. The
corresponding paramagnetic monoanions [MII(LBu)(LBu•)]- (S
(M
corresponding complex [CuIII(LBu)2]- has also been investigated. The complexes have been studied by UV
and EPR spectroscopy and by X-ray crystallography; their electro- and magnetochemistry is reported. The electron-
)
1/2) and the neutral diamagnetic species [MII(LBu•)2]
)
Ni, Pd, Pt) have also been generated in solution or in the solid state as [N(n-Bu)4][MII(LBu)(LBu•)] salts. The
−vis, IR,
transfer series [M(LBu)2]2
or two in the neutral species [MII(LBu•)2] (M
theory with scalar relativistic corrections at the second-order Douglas
is shown to be ligand based involving formally one (LBu
Ni, Pd, Pt). Geometry optimizations using all-electron density functional
Kroll Hess (DKH2) and zeroth-order regular
)
• - π radical in the monoanion
-,-,0
)
−
−
approximation (ZORA) levels result in excellent agreement with the experimentally determined structures and electronic
spectra. For the three neutral species a detailed analysis of the orbital structures reveals that the species may best
be described as containing two strongly antiferromagnetically interacting ligand radicals. Furthermore, multicon-
figurational ab initio calculations using the spectroscopy oriented configuration interaction (SORCI) approach including
the ZORA correction were carried out. The calculations predict the position of the intervalence charge-transfer
band well. Chemical trends in the diradical characters deduced from the multiconfigurational singlet ground-state
wave function along a series of metals and ligands were discussed.
Introduction
stituted derivative benzene-1,2-dithiolate, (L)2-, have been
used predominantly, but recently, Sellmann et al.6 have
introduced the sterically more demanding 3,5-di-tert-butyl-
benzene-1,2-dithiolate ligand, (LBu)2-, and investigated its
coordination chemistry with nickel.7
Since their discovery in 1963, square planar bis(benzene-
1,2-dithiolato)metal complexes1-5 have received considerable
attention due to their interesting and often unusual electronic
structures. Toluene-3,4-dithiolate, (LMe)2-, and the unsub-
It has been established2 that the square planar complexes
[M(L)2]2-, [M(L)2]-, and [M(L)2]0 (M ) Ni, Pt) form a three-
membered electron-transfer series where the neutral species
and the dianionic forms are diamagnetic with S ) 0 ground
* Authors to whom correspondence should be addressed. E-mail:
wieghardt@mpi-muelheim.mpg.de (K.W.); neese@mpi-muelheim.mpg.de
(F.N.).
(1) Reviews: (a) Schrauzer, G. N. Transition Met. Chem. 1968, 4, 299.
(b) McCleverty, J. A. Prog. Inorg. Chem. 1968, 10, 49. (c) Holm, R.
H.; O’Connor, M. J. Prog. Inorg. Chem. 1971, 14, 241. (d) Eisenberg,
R. Prog. Inorg. Chem. 1970, 12, 295. (e) Gray, H. B. Transition Met.
Chem. 1965, 1, 240. (f) Dithiolene Chemistry; Stiefel, E. I., Ed.; Prog.
Inorg. Chem. 2004, 52.
(2) Gray, H. B.; Billig, E. J. Am. Chem. Soc. 1963, 85, 2019.
(3) Williams, R.; Billig, E.; Waters, J. H.; Gray, H. B. J. Am. Chem. Soc.
1966, 88, 43.
(4) Baker-Hawkes, M. J.; Billig, E.; Gray, H. B. J. Am. Chem. Soc. 1966,
88, 4870.
1
states whereas the monoanions are paramagnetic (S ) /2).
In contrast, the monoanion [Cu(LMe)2]- possesses an S ) 0
ground state2 as does [Au(L)2]-,3 whereas the corresponding
dianions [Cu(LMe)2]2- and [Au(L)2]2- are paramagnetic (S
) 1/2) as is the neutral species [Au(LBu)2]0.8 Throughout the
(6) (a) Sellmann, D.; Freyberger, G.; Eberlein, R.; Bo¨hlen, E.; Huttner,
G.; Zsolnai, L. J. Organomet. Chem. 1987, 323, 21. (b) Sellmann,
D.; Ka¨ppler, O. Z. Naturforsch. 1987, 42b, 1291.
(5) Stiefel, E. I.; Waters, J. H.; Billig, E.; Gray, H. B. J. Am. Chem. Soc.
1965, 87, 3016.
(7) Sellmann, D.; Binder, H.; Ha¨usinger, D.; Heinemann, F. W.; Sutter,
J. Inorg. Chim. Acta 2000, 300-302, 829.
10.1021/ic0507565 CCC: $30.25
Published on Web 06/23/2005
© 2005 American Chemical Society
Inorganic Chemistry, Vol. 44, No. 15, 2005 5345

Ray et al.
Chart 1. Ligands and Complexes
past 40 years there has been a debate on the nature of the
electronic structure of these complexes. The question of
ligand- vs metal-centered redox activity has led to diverse
(and sometimes mutually exclusive) descriptions of the
electronic structures of the members of an electron-transfer
series.
As a recent example from the year 2000, consider the
series [Ni(LBu)2]0,-,2- all of which have been characterized
by X-ray crystallography (they are square planar) and by
spectroscopy (UV-vis, EPR, XPS).7 The authors concluded
that “benzene-1,2-dithiolate(2-) and its (LBu)2- derivative
are normal 1,2-dithiolate(2-) ligands.” This implies the
occurrence of NiII, NiIII, and NiIV in the di- and monoanion
and in the neutral species, respectively. This interpretation
had already been refuted5 in 1965, where the radical character
of the ligand in [NiII(LMe•)(LMe)]- and [NiII(LMe•)2]0 had been
emphasized. In a recent correlated ab initio and density
functional theoretical (DFT) investigation of diamagnetic
ISQ •-
[Ni(LNISQ)2], where (LN
)
represents the o-diiminoben-
zosemiquinonate(1-) π radical monoanion, we have shown9a
that this species has ∼77% diradical character (from DDCI
method). By using a broken symmetry DFT approach for
the calculation of the electronic structure of the present sulfur
containing [Ni(L•)2]0 (S ) 0) species, its diradical character
has been estimated to be negligible (∼0%).10 This is in
agreement with previous DFT calculations.11,12 We have now
performed elaborate multireference post-Hartree-Fock ab
initio calculations on the neutral complexes [M(L•)2]0 (M )
Ni, Pd, Pt) using the recently developed spectroscopy
oriented configuration interaction (SORCI) method,13 which
is well suited for diradical species and will be used below
to discuss the trends in the ground-state singlet wave
functions along the series of complexes investigated.
density functional theoretical calculations on [CuIII(pds)2]-
have clearly established the presence of a CuIII (d8, S ) 0)
ion (pds2- represents the ligand pyrazine-2,3-dithiolate(2-)).16
Also, the electronic structure of [Au(L)2]- has been
studied17 by ab initio calculations, and it was shown that
this species is best described as a AuIII (d8, S ) 0) species
with two closed-shell benzene-1,2-dithiolate(2-) ligands. The
one-electron oxidation yielding neutral [Au(L)2]0 was
proposed to be predominantly ligand-centered yielding
[AuIII(L•)(L)] T [AuIII(L)(L•)]. In this work we have
synthesized the square planar bis(3,5-di-tert-butylbenzene-
1,2-dithiolato)metal complexes of palladium, platinum, and
copper shown in Chart 1. Those of nickel7 and gold8 have
been reported previously.
Similarly, for diamagnetic [Cu(L)2]- an electronic struc-
ture of (i) [CuIII(L)2]- containing a CuIII ion (d8, S ) 0 in a
square planar ligand environment) and two closed-shell
dianions (L)2- has been proposed2-4 or, alternatively, of (ii)
[CuII(L•)(L)]- has been proposed, where a monoanionic
ligand π radical (L•)- has been suggested to be intramo-
lecularly antiferromagnetically coupled to a CuII (d9) ion.14
It may be noted that the Goodenough-Kanamori rules for
intramolecular spin exchange coupling between a π radical
Experimental Section
The ligand 3,5-di-tert-butylbenzene-1,2-dithiol, H2[LBu], and the
nickel complex [Ni(LBu•)2] (1) have been prepared as described in
refs 6 and 7.
[N(n-Bu)4][Ni(LBu)(LBu•)] (1a). To a solution of the ligand
H2[LBu] (0.65 g; 2.6 mmol) in absolute ethanol (10 mL) was added
a solution of potassium metal (240 mg) in absolute ethanol (5 mL)
under an argon-blanketing atmosphere. To this solution was added
solid NiCl2‚6H2O (0.31 g; 1.3 mmol) and [N(n-Bu)4]Br (0.41 g,
1.3 mmol) at ambient temperature. The resulting reddish solution
was stirred for 30 min in the presence of air whereupon a greenish
residue precipitated which was collected by filtration. The crude
material was recrystallized from a CH2Cl2/ethanol mixture (1:1 v:v)
yielding shining green microcrystals of 1a (0.78 g; 73%). Anal.
Calcd for C44H76S4NNi: C, 65.57; H, 9.50; N, 1.74. Found: C,
65.8; H, 9.4; N, 1.8. Electrospray mass spectrum (ESI) (CH2Cl2
solution): m/z ) 564.03 (neg ion mode) [Ni(LBu)2]- and 242.3 [N(n-
Bu)4]+ (pos ion mode).
1
2
II
2
orbital and a (dx -y ) orbital of Cu would predict a fer-
romagnetic coupling yielding an S ) 1 ground state15 for
the latter description because the two magnetic orbitals would
be strictly orthogonal in a square planar CuS4 unit. Recent
(8) Ray, K.; Weyhermu¨ller, T.; Goossens, A.; Craje´, M. W. J.; Wieghardt,
K. Inorg. Chem. 2003, 42, 4082.
(9) (a) Herebian, D.; Wieghardt, K.; Neese, F. J. Am. Chem. Soc. 2003,
125, 10997. (b) Herebian, D.; Bothe, E.; Neese, F.; Weyhermu¨ller,
T.; Wieghardt, K. J. Am. Chem. Soc. 2003, 125, 9116.
(10) Bachler, V.; Olbrich, G.; Neese, F.; Wieghardt, K. Inorg. Chem. 2002,
41, 4179.
(11) Lauterbach, C.; Fabian, J. Eur. J. Inorg. Chem. 1999, 1995.
(12) Weber, J.; Daul, C.; von Zelewsky, A.; Goursot, A.; Penigault, E.
Chem. Phys. Lett. 1982, 88, 78.
[N(n-Bu)4]2[Ni(LBu)2] (1b). To a degassed solution of the ligand
H2[LBu] (0.25 g; 1.0 mmol) and [N(n-Bu)4]Br (0.32 g; 1.0 mmol)
(13) Neese, F. J. Chem. Phys. 2003, 119, 9428.
(14) Sawyer, D. T.; Srivatsa, G. S.; Bodini, M. E.; Schaefer, W. P.; Wing,
R. M. J. Am. Chem. Soc. 1986, 108, 936.
(15) Chaudhuri, P.; Verani, C. N.; Bill, E.; Bothe, E.; Weyhermu¨ller, T.;
Wieghardt, K. J. Am. Chem. Soc. 2001, 123, 2213.
(16) Ribas, X.; Dias, J. C.; Morgado, J.; Wurst, K.; Molins, E.; Ruiz, E.;
Almeida, M.; Veciana, J.; Rovira, C. Chem.-Eur. J. 2004, 10, 1691.
(17) Schiødt, N. C.; Sommer-Larsen, P.; Bjørnholm, T.; Folmer-Nielsen,
M.; Larsen, J.; Bechgaard, K. Inorg. Chem. 1995, 34, 3688.
5346 Inorganic Chemistry, Vol. 44, No. 15, 2005

Bis(benzene-1,2-dithiolato)metal Complexes
in absolute methanol (10 mL) was added NaOMe (0.24 g, 4 mmol),
NiCl2‚6H2O (0.12 g; 0.5 mmol), and an argon atmosphere (glove-
box). The solution was stirred for ∼12 h at ambient temperature
during which time an extremely air-sensitive reddish-brown powder
of 1b precipitated. Yield: 0.29 g (54%). Anal. Calcd for C60H112S4N2-
Ni: C, 68.74; H, 10.77; N, 2.67. Found: C, 68.9; H, 10.6; N, 2.5.
ESI mass spectrum (CH2Cl2 solution): m/z ) 564.03 [Ni(LBu)2]-
(oxidized dianion) and 242.3 [N(n-Bu)4]+.
[PdII(LBu•)2] (2) and [PtII(LBu•)2] (3). These complexes have been
generated electrochemically by one-electron oxidation at a con-
trolled, fixed potential of +0.2 V vs Fc+/Fc at ambient temperature
of a CH2Cl2 solution (5 mL; 0.10 M [N(n-Bu)4]PF6) containing 2a
or 3a in 1 mM concentration.
[N(n-Bu)4][PdII(LBu)(LBu•)] (2a). To a solution of the ligand
H2[LBu] (0.13 g; 0.5 mmol) in dimethylformamide (2 mL) was
added dropwise a solution of [N(n-Bu)4]2PdCl6 (0.20 g; 0.25 mmol)
in CH3CN (9 mL) under an Ar-blanketing atmosphere. After
addition of 0.5 mL of ethanol the solution was heated to 40 °C for
30 min during which time the color of the solution changed from
dark brown to light green. The solution was allowed to cool to 20
°C and was kept under a continuous flow of Ar gas for 3 h. Shining,
light green microcrystals precipitated, which were collected by
filtration. Yield: 0.12 g (54%). Anal. Calcd for C44H76S4NPd: C,
61.85; H, 8.29; N, 1.64. Found: C, 62.0; H, 8.4; N, 1.6. ESI mass
spectrum (pos and neg ion mode CH2Cl2 solution): m/z ) 610
[Pd(LBu)2]- and 242.3 [N(n-Bu)4]+.
[PPh4]2[PdII(L)2] (2b*). To an absolutely anaerobic solution of
the ligand H2[L] (0.28 g; 2 mmol) and trimethylamine (0.4 mL) in
dry methanol (30 mL) were added PdCl2 (0.18 g; 1.0 mmol) and
[PPh4]Br (0.85 g; 2.0 mmol). The resulting solution was stirred at
20 °C for 12 h, after which time bright orange microcrystals had
precipitated which were collected by filtration under an Ar-
blanketing atmosphere. Recrystallization from CH3CN yielded dark
orange cubes suitable for X-ray diffraction. Anal. Calcd for
C60H48S4P2Pd: C, 67.63; H, 5.26. Found: C, 67.6; H, 5.1. ESI mass
spectrum (neg and pos ion mode; CH2Cl2 solution): m/z ) 386
[Pd(L)2]- (oxidized species) and 343.3 [PPh4]+.
[AuIII(LBu)(LBu•)] (4) and [AuIII(LBu)2] [N(n-Bu)4] (4a). These
complexes have been synthesized as described in ref 8.
[N(n-Bu)4][CuIII(LBu)2] (5). Sodium metal (0.069 g; 3.0 mmol)
was added to a solution of [N(n-Bu)4]Br (0.16 g; 0.5 mmol) and
the ligand H2[LBu] (0.25 g; 1.0 mmol) in absolute ethanol (10 mL).
A solution of CuCl2‚2H2O (0.085 g; 0.5 mmol) in ethanol (8 mL)
was added dropwise to the above solution at 20 °C. The resulting
red solution was stirred for 0.5 h at 20 °C in the presence of air,
during which time a green precipitate formed which was collected
by filtration. The crude product was recrystallized from CH2Cl2
yielding light green microcrystals. Yield: 0.35 g (85%). Anal. Calcd
for C44H76S4NCu: C, 65.59; H, 9.45; N, 1.73. Found: C, 65.4; H,
9.4 N, 1.8. ESI mass spectrum (pos. and neg. ion mode; CH2Cl2
solution): m/z ) 567.2 [Cu(LBu)2]- and 242.3 [N(n-Bu)4]+. By
using [As(CH3)(C6H5)3]Br instead of [N(n-Bu)4]Br in the above
reaction, the complex [As(CH3)(C6H5)3][CuIII(LBu)2] (5′) was
obtained. Single crystals suitable for X-ray crystallography were
grown from a CH2Cl2/C2H5OH mixture (1:1 v:v).
X-ray Crystallographic Data Collection and Refinement of
the Structures. A light violet single crystal of 2a, an orange red
crystal of 2b*, and a light green specimen of 5′ were coated with
perfluoropolyether, picked up with a glass fiber, and mounted in
the nitrogen cold stream of the diffractometer. Intensity data were
collected on a Nonius Kappa-CCD diffractometer equipped with a
Mo-target rotating-anode X-ray source and a graphite monochro-
mator (Mo KR, λ ) 0.710 73 Å). Final cell constants were obtained
from least-squares fits of all integrated reflections. Crystal faces
of 2a and 2b* were determined, and the corresponding intensity
data were corrected for absorption using the Gaussian-type routine
embedded in XPREP.18 The data set for 5′ was left uncorrected.
Crystallographic data of the compounds are listed in Table 1. The
Siemens ShelXTL18 software package was used for solution and
artwork of the structure, and ShelXL9719 was used for the
refinement. The structures were readily solved by direct and
Patterson methods and subsequent difference Fourier techniques.
All non-hydrogen atoms were refined anisotropically, and hydrogen
atoms were placed at calculated positions and refined as riding
atoms with isotropic displacement parameters. Compound 5′
crystallizes in space group P1. No additional symmetry was found,
but crystals appeared to be racemically twinned. The tert-butyl
groups bound to C(4) and C(24) were found to be disordered. Two
split positions for each of the four carbon atoms were refined with
occupancies in a 57:43 ratio. Chemically equal C-C distances were
restrained to have similar values using the SADI Instruction of
ShelXL, and equal anisotropic displacement parameters were refined
for corresponding C atoms.
[N(n-Bu)4]2[Pd(LBu)2] (2b). To a solution of the ligand H2[LBu]
(0.25 g; 1.0 mmol), NaOMe (0.24 g, 4 mmol), and [N(n-Bu)4]Br
(0.32 g; 1.0 mmol) in methanol (10 mL) was added PdCl2 (0.88 g;
0.5 mmol) under strictly anaerobic conditions in a glovebox. The
solution was stirred at 20 °C for 12 h, during which time an
extremely air-sensitive reddish-brown powder precipitated. Yield:
0.30 g (54%). Anal. Calcd for C60H112S4N2Pd: C, 65.74; H, 10.30;
N, 2.56. Found: C, 65.9; H, 10.4; N, 2.4. ESI mass spectrum (pos
and neg ion mode; CH2Cl2 solution): m/z ) 610 [Pd(LBu)2]- and
242.3 [N(n-Bu)4]+.
[N(n-Bu)4][PtII(LBu)(LBu•)] (3a). This compound has been
prepared as described above for 2a by using [N(n-Bu)4]2PtCl6 as
starting material. Yield: 62%. Anal. Calcd for C44H76S4NPt: C,
56.08; H, 8.14; N, 1.49. Found: C, 56.3; H, 8.3; N, 1.6. ESI mass
spectrum (pos and neg ion mode; CH2Cl2 solution: m/z ) 700.1
[Pt(LBu)2]- and 242.3 [N(n-Bu)4]+.
[N(n-Bu)4]2[Pt(LBu)2] (3b). To a solution of the ligand H2[LBu]
(0.25 g; 1.0 mmol), NaOMe (0.24 g, 4 mmol), and [N(n-Bu)4]Br
(0.32 g; 1.0 mmol) in methanol (10 mL) was added PtCl2 (0.13 g;
0.5 mmol) under strictly anaerobic conditions (glovebox; N2). The
solution was stirred at 20 °C for 12 h, after which time an extremely
air-sensitive red-brown powder had formed which was collected
by filtration. Yield: 0.36 g (62%). Anal. Calcd for C60H112S4N2Pt:
C, 60.82; H, 9.53; N, 2.36. Found: C, 61.0; H, 9.6; N, 2.3. ESI
mass spectrum (pos and neg ion mode; CH2Cl2 solution): m/z )
700.1 [Pt(LBu)2]- (oxidized species) and 242.3 [N(n-Bu)4]+.
Physical Measurements. Electronic spectra of complexes and
spectra of the spectroelectrochemical measurements were recorded
on a Perkin-Elmer Lambda 9 spectrophotometer (range: 200-2000
nm). Cyclic voltammograms and coulometric measurements were
performed with an EG&G potentiostat/galvanostat. X-band EPR
spectra were recorded with a Bruker ESR 300 spectrometer. Infrared
spectra were measured as KBr disks (or in CH2Cl2 solution; NaCl
windows) on a Perkin-Elmer FFIR spectrophotometer 2000.
Calculations. All calculations in this work were performed with
the electronic structure program ORCA.20 All electron DFT-based
geometry optimizations were carried out using the BP86 func-
(18) ShelXTL V.5; Siemens Analytical X-ray Instruments Inc.: Madison,
WI, 1994.
(19) Sheldrick, G. M. ShelXL97; Universita¨t Go¨ttingen: Go¨ttingen, Ger-
many, 1997.
Inorganic Chemistry, Vol. 44, No. 15, 2005 5347

Ray et al.
Table 1. Crystallographic Data for 2a, 2b*, and 5′
param
2a
2b*
5
chem formula
fw
space group
a, Å
b, Å
c, Å
R, deg
â, deg
γ, deg
V, Å
C44H76NPdS4
853.70
C60H48P2PdS4
1065.56
P21/n, No. 14
11.0521(6)
14.7262(8)
14.9977(8)
90
95.13(1)
90
2431.2(2)
2
C47H58AsCuS4
889.63
C2/c, No. 15
26.438(2)
10.1521(8)
18.098(2)
90
102.79(1)
90
4737.0(7)
4
P1, No. 1
8.3444(4)
9.5921(4)
15.6799(9)
83.80(1)
83.23(1)
66.76(1)
1142.6(1)
1
Z
T, K
F
200(2)
1.197
100(2)
1.456
100(2)
1.293
calcd, g cm-3
reflcns collcd/2θmax
unique reflcns/I > 2σ(I)
no. of params/restr
µ(Mo KR), cm-1
27 273/47.62
3629/2931
236/0
5.96
0.0363/1.040
0.0647
18 765/55.0
5496/4305
304/0
6.60
0.0408/1.045
0.0768
17 544/61.2
10 758/9767
503/141
14.10
0.0471/1.021
0.1112
R1a/goodness of fitb
wR2c (I < 2σ(I))
resid density, e Å-3
+0.26/-0.31
+0.49/-0.43
+1.37/0.59
a Observation criterion: I > 2σ(I). R1 ) ∑||Fo| - |Fc||/∑|Fo|. b GooF ) [∑[w(Fo2 - Fc )2]/(n - p)]1/2
.
c wR2 ) [∑[w(Fo2 - Fc )2]/∑[w(Fo )2]]1/2, where
2
2
2
w ) 1/σ2(Fo ) + (aP)2 + bP and P ) (Fo + 2Fc )/3.
2
2
2
tional.21 Since this work is concerned with heavy transition metal
complexes, scalar relativistic corrections were included using the
second-order Douglas-Kroll-Hess (DKH2) method.22 In the
geometry optimizations the one-center approximation was used
which eliminates DKH2 contributions to the analytic gradients. In
the context of the zeroth-order regular approximation (ZORA),23a
it was shown that the one-center approximation only introduces
minor errors into the final geometries.23 Large uncontracted
Gaussian basis sets were used at the metal center which were
derived from the well-tempered basis sets of Huzinaga.24 For the
remaining atoms the all-electron polarized triple-ê (TZVP)25a,b
Gaussian basis sets of the Ahlrichs group were used in an
uncontracted form to allow for a distortion of the inner shell orbitals
in the presence of the relativistic potential.
multicenter mean-spin-orbit coupling (SOC) operator has recently
been implemented into ORCA,28c we have not used it here, due to
the limitations of the underlying Breit-Pauli approach for heavier
metals such as Pt. Instead, the effective nuclear charge model of
Koseki was employed.28d,e However, since we have performed all-
electron calculations using scalar relativistic corrections, the ef-
fective nuclear charges for Pd, Pt, and Au needed to be adjusted.
From comparison between calculated SOC matrix elements with
effective one-electron SOC constants the values Zeff(Pd) ) 14.3,
Zeff(Pt) ) 22.5, and Zeff(Au) ) 27.5 were determined which lead
to únd values of ∼1300 cm-1 (Pd), ∼3400 cm-1 (Pt), and ∼4500
cm-1 (Au), respectively. All other effective nuclear charges had
their default values. TD-DFT calculations were carried out accord-
ing to ref 28f.
The property calculations at the optimized geometries were done
with the B3LYP functional.26 In this case the same basis sets were
used but the scalar-relativistic ZORA method was used since in
this formalism magnetic properties are more readily formulated.27a
For the ZORA calculations, the method of van Wu¨llen27b is
implemented in ORCA. Care was taken to ensure accurate
numerical integration in the presence of steep basis functions.
Details of the methods used to calculate the g-tensor and metal
hyperfine couplings can be found in refs 28a,b. Although a
Ab initio calculations of the optical spectra of the neutral [M(L)2]
complexes with M ) Ni, Pd, and Pt were also undertaken with the
ORCA program and utilized the recently developed spectroscopy
oriented configuration interaction (SORCI) formalism.13 Scalar-
relativistic ZORA corrections were considered in the calculations.
For the first row transition atoms, Wachters29 basis sets were used
at the metal center, contracted triple-ê (TZVP)25a,b basis sets for
the sulfurs, nitrogens, and oxygens, and SV(P) basis sets for the
remaining atoms. For second row transition metals, the TZVP basis
sets of Ahlrichs and May were used.25c For Pt we have recontracted
the well-tempered basis set used in the geometry optimization in
an atomic ZORA calculation as (34s24p19d13f) f [16s11p11d4f].
The selection threshold Tsel was 10-6 Eh, the prediagonalization
threshold Tpre ) 10-6, and the natural orbital selection threshold
Tnat ) 10-5. The calculations were started with spin-restricted
orbitals from BP86 DFT calculations. The reference space was
CAS(2,2) for the neutral complexes.
(20) Neese, F. Orca-an ab initio, DFT and semiempirical Electronic
Structure Package, version 2.4, revision 16; Max-Planck Institut fu¨r
Bioanorganische Chemie: Mu¨lheim, Germany, Nov 2004.
(21) (a) Becke, A. D. J. Chem. Phys. 1988, 84, 4524. (b) Perdew, J. P.
Phys. ReV. B 1986, 33, 8522.
(22) Hess, B. A.; Marian, C. M. In Computational Molecular Spectroscopy;
Jensen, P., Bunker, P. R., Eds.; John Wiley & Sons: New York, 2000;
p 169 ff.
(23) (a) van Lenthe, E.; Snijders, J. G.; Baerends, E. J. J. Chem. Phys.
1996, 105, 6505. (b) van Lenthe J. H.; Faas, S.; Snijders J. G. Chem.
Phys. Lett. 2000, 328, 107.
(24) (a) Huzinaga, S.; Miguel, B. Chem. Phys. Lett. 1990, 175, 289. (b)
Huzinaga, S.; Klobukowski, M. Chem. Phys. Lett. 1993, 212, 260.
(25) (a) Scha¨fer, A.; Horn, H.; Ahlrichs, R. J. Chem. Phys. 1992, 97, 2571.
(b) Scha¨fer, A.; Huber, C.; Ahlrichs, R. J. Chem. Phys. 1994, 100,
5289. (c) Ahlrichs, R.; May, K. Phys. Chem. Chem. Phys. 2000, 2
943-945.
Natural population analysis (NPA)30a-c was done through an
interface of ORCA to the Gennbo program version 5.0.30d Isocon-
tour plots were done with the ChemBats3D program.31
(28) (a) Neese, F. J. Chem. Phys. 2001, 115, 11080. (b) Neese, F. J. Chem.
Phys. 2003, 118, 3939. (c) Neese, F. J. Chem. Phys. 2005, 122, 034107.
(d) Koseki, S.; Gordon, M. S.; Schmidt, M. W.; Matsunaga, N. J.
Phys. Chem. 1995, 99, 12764. (e) Koseki, S.; Schmidt, M. W.; Gordon,
M. S. J. Phys. Chem. A 1998, 102, 10430. (f) Neese, F.; Olbrich, G.
Chem. Phys. Lett. 2002, 362, 170.
(26) (a) Lee, C. Yang, W. Parr, R. G. Phys. ReV. B 1988, 37, 785. (b)
Becke, A. D. J.Chem. Phys. 1993, 98, 5648.
(27) (a) van Lenthe E.; van der Avoird, A,; Wormer, P. E. S. J. Chem.
Phys. 1998, 108, 4783. (b) van Wu¨llen, C. J. Chem. Phys. 1998, 109,
392.
(29) Wachters, A. J. H. J. Chem. Phys. 1970, 52, 1033.
5348 Inorganic Chemistry, Vol. 44, No. 15, 2005

Bis(benzene-1,2-dithiolato)metal Complexes
Table 2. Redox Potentials of [M(LBu)2]z (z ) 2-, 1-, 0) Complexes
in CH2Cl2 Solution (0.10 M [N(n-Bu)4]PF6) at 20 °C
complex
E11/2, V vs Fc+/Fc
E21/2, V vs Fc+/Fc
1a
2a
3a
4aa
5
-0.23
-1.10
-0.92
-1.10
-2.18
-1.44
-0.22
-0.27
0.17
0.17 (quasi reversible)
a Reference 8.
Results
Synthesis. The reaction of disodium benzene-1,2-dithi-
olates, Na2[LBu] or Na2[L], with MCl2 (M ) Ni, Pd, Pt) in
absolute ethanol under strictly anaerobic conditions in the
ratio 2:1 yields the square planar, diamagnetic dianions
[MII(LBu)2]2- or [MII(L)2]2- (M ) Ni, Pd, Pt) of which the
bis(tetra-n-butylammonium) salts have been isolated: [N(n-
Bu)4]2[M(LBu)2] (M ) Ni (1b), Pd (2b), Pt (3b)). These salts
are very sensitive toward dioxygen both in the solid state
and in solution. Thus, it has been possible to isolate [N(n-
Bu)4][M(LBu)(LBu•)] salts (M ) Ni (1a), Pd (2a), Pt (3a))
from solutions of 1b-3b exposed to air. These monoanions
[M(LBu)(LBu•)]- are also square planar but paramagnetic with
Figure 1. Cyclic voltammograms of 1a (top), 2a (middle), and 3a (bottom)
in CH2Cl2 solution (0.10 M [N(n-Bu)4]PF6. Conditions: scan rate 200 mV
s-1 at 25 °C (glassy carbon working electrode; ferrocene (Fc) internal
standard).
1
an S ) /2 ground state.
Further one-electron oxidation of 1a, 2a, or 3a either
chemically with 1 equiv of ferrocenium hexafluorophosphate
or electrochemically via controlled potential coulometry (see
below) affords light green, solutions of the square planar,
diamagnetic, neutral complexes [MII(LBu•)2] (M ) Ni (1),7
Pd (2), Pt (3)). These compounds possess an S ) 0 ground
state. Sellmann et al.7 had prepared 1 by oxidation of 1a
with iodine in acetone solution as solid material and
determined its crystal structure by X-ray crystallography.
Similarly, solid light-green [N(n-Bu)4][CuIII(LBu)2] (5) has
been synthesized by the reaction of CuCl2·2H2O with 2 equiv
of Na2[LBu] in ethanol in the presence of air and addition of
[N(n-Bu)4]Br. This material is diamagnetic (S ) 0). The salt
[As(CH3)(C6H5)3][CuIII(LBu)2] (5′) has been obtained as single
crystals suitable for X-ray crystallography. The synthesis and
Figure 2. Cyclic voltammograms of 5 (top) and 4a (bottom) in CH2Cl2
solution (0.10 M [N(n-Bu)4]PF6). Conditions are as in Figure 1.
redox potentials. Coulometric measurements established that
each of the monoanions in 1a, 2a, and 3a, [M(LBu)(LBu•)]-,
undergoes a reversible one-electron oxidation and a reversible
one-electron reduction. Thus, the three-membered electron-
transfer series consisting of [MII(LBu)2]2-, [MII(LBu)(LBu•)]-,
and [MII(LBu•)2]0 (M ) Ni, Pd, Pt) is clearly established.
Similar results have been reported for the series containing
the unsubstituted benzene-1,2-dithiolate or its toluene-3,4-
dithiolate analogues.3
The results in Figure 1 and Table 2 for 1a, 2a, and 3a
clearly show that the one-electron oxidation and the one-
electron reduction must involve orbitals which display
1
spectroscopic properties of [AuIII(LBu)(LBu•)] (S ) /2) (4)
and [AuIII(LBu)2]- (S ) 0) (4a) have been described in ref 8.
Electro- and Spectroelectrochemistry. Cyclic voltam-
mograms of complexes 1a, 2a, 3a, and 5 have been recorded
at 25 °C in CH2Cl2 solutions containing 0.10 M [N(n-Bu)4]-
PF6 as supporting electrolyte and 0.1-1.0 mM complex by
using a glassy carbon working electrode. Ferrocene was used
as an internal standard; redox potentials are referenced versus
the ferrocenium/ferrocene (Fc+/Fc) couple. Table 2 sum-
marizes the results.
predominantly ligand character because the potentials E1
1/2
and E21/2 are nearly independent of the nature of the central
metal ion Ni, Pd, or Pt.
In contrast, the CVs of complexes 4a and 5 shown in
Figure 2 display both a reversible (4a) and a quasi-reversible
(5) one-electron oxidation at the same redox potential of
+0.17 V for the [MIII(LBu•)(LBu)]0/[MIII(LBu)2]- couple, eq
1, but the corresponding one-electron reductions at -2.18
V for 4a8 and -1.44 V for 5 differ by 0.74 V indicating the
metal-centered character of these reductions, eq 2.
Figure 1 shows the CVs of 1a, 2a, and 3a recorded at a
scan rate of 200 mV s-1 at 25 °C. Each compound displays
two reversible one-electron transfer waves at nearly identical
[MIII(LBu)2]- y-e
\
+e
z [MIII(LBu)(LBu•)]0
(1)
(2)
(30) (a) Reed, A. E.; Weinhold, F. J. Chem. Phys. 1983, 78, 4066. (b)
Reed, A. E.; Weinstock, R. B.; Weinhold, F. J. Chem. Phys. 1985,
83, 735. (c) Reed, A. E.; Curtiss, L. A.; Weinhold, F. Chem. ReV.
1988, 88, 899. (d). Glendening, E. D.; Badenhoop, J. K.; Reed, A. E.;
Carpenter, J. E.; Bohmann, J. A.; Morales, C. M.; Weinhold F. NBO
5.0; Theoretical Chemistry Institute, University of Wisconsin: Madi-
son, WI, 2001; http://www.chem.wisc.edu/∼nbo5.
[MII(LBu)2]- y+e
\
z [MII(LBu)2]2-
-e
Figure 3 displays the infrared spectra (KBr disks) of solid
1 and 1a,b in the range 500-1500 cm-1. It is quite
remarkable that the tetra-n-butylammonium salts 1a,b exhibit
(31) ChemBats3d, version 6.0; CambridgeSoft.com, 2000.
Inorganic Chemistry, Vol. 44, No. 15, 2005 5349

Ray et al.
Figure 3. Infrared spectra of solid 1 (black), 1a (green), and 1b (red) in
the range 500-1500 cm-1 (KBr disks). The spectra of the tetra-n-
butylammonium salts of 1a,b were recorded.
Table 3. ν(CdS•) Stretching Mode of Complexes
complex
ν(CdS•), cm-1
complex
ν(CdS•), cm-1
1
1099
1114
not obsd
1082
1102
3a
3b
4a
4aa
5
1109
not obsd
1085
not obsd
not obsd
not obsd
1a
1b
2
2a
2b
3
not obsd
1088
6
Figure 4. Infrared spectra of 2 (red), 2a (blue), and 2b (black) and 3
(red), 3a (blue), and 3b (black) in CH2Cl2 solution (0.10 M [N(n-Bu)4]-
PF6) at -25 °C. Both neutral forms 2 and 3 and the dianions 2b and 3b
were generated electrochemically.
different spectra: 1a displays a frequency at 1114 cm-1,
which is absent in the corresponding spectrum of 1b. In the
spectrum of neutral 1 the former frequency is again observed
at 1099 cm-1 (Table 3), whereas the band at 1029 cm-1 has
only gained intensity in comparison with the spectra of 1a,b.
As we have shown previously for 4 and 4a, the former
displays this band at 1085 cm-1 and is absent in the spectrum
of 4a.8 We have pointed out that this band is a marker for
Table 4. Electronic Spectra of Complexes in CH2Cl2 Solution
complex T, °C
λ )
max, nm (ꢀ, 104 M-1 cm-1
1a
1a
1ba
2a
2a
2ba
3a
-25 309 (4.6), 348 (1.4), 378 (0.6), 600 (br, 0.2), 840 (3.0)
20 303 (2.8, 389 (1.3), 698 (sh, 0.2), 890 (1.5)
-25 310 (2.9), 398 (0.75)
-25 379 (1.2), 620 (0.8), 850 (5.0)
20 340 (7.7), 408 (0.1), 894 (0.5), 1140 (2.5)
-25 350 (0.9), 392 (0.6)
-25 320 (0.6), 390 (0.5), 802 (3.9)
20 302 (1.5), 383 (0.3), 732 (0.2), 900 (1.9)
-25 300 (1.5), 385 (0.25)
20 400 (0.6), 433 (0.34), 1023 (0.22), 1452 (2.7)
20 380 (0.1), 415 (0.02), 618 (0.01)
20 343 (1.2), 398 (4.0)
the presence of S,S-coordinated (LBu• -
) radicals in square
planar coordination compounds containing the 3,5-di-tert-
butyl-benzene-1,2-dithiolato ligands.
3a
3ba
4b
To make a comparative study of the infrared spectra of 2,
2a,b, 3, and 3a,b we have generated the neutral and dianionic
forms in CH2Cl2 solutions (0.10 M [N(n-Bu)4]PF6) by
coulometry from 2a and 3a, respectively, and recorded their
solution infrared spectra. These are displayed in Figure 4,
and the ν(CdS•) frequencies are summarized in Table 3. It
is immediately clear that 2, 2a, 3, and 3a exhibit ν(CdS•)
bands but 2b and 3b do not. Thus, 2a and 3a may best be
described as [MII(LBu•)(LBu)]- and 2 and 3 as [MII(LBu•)2]
(M ) Pd, Pt), whereas 2b and 3b are [MII(LBu)2]2- (M )
Pd, Pt) species.
4ab
5
a Generated electrochemically at -25 °C. b Reference 8.
1a (top), 2a (middle), and 3a (bottom) together with their
electrochemically generated one-electron-oxidized species
1-3 and their one-electron-reduced forms 1b, 2b, and 3b.
The assignment of the major absorption features follows in
analogy to the assignments made in refs 9a,b and will be
corroborated below though electronic structure calculations.
According to our interpretation, the most salient features
of these spectra are the following: (a) The neutral species
1-3 display a very intense (>104 M-1 cm-1) ligand-to-ligand
charge-transfer band (LLCT) at 840, 850, and 802 nm,
respectively. (b) Similarly, the monoanions 1a, 2a, and 3a
display an intense (>104 M-1 cm-1) intervalence charge
transfer band (IVCT) at 890, 1140, and 900 nm, respectively
(the intensity of these IVCT bands of the monoanions is
approximately half of that observed for the LLCT bands of
the neutral species). (c) The dianions 1b, 2b, and 3b do not
absorb significantly (>102 M-1 cm-1) above 400 nm. (d)
All species exhibit an intense (>104 M-1 cm-1) ligand-to-
Interestingly, the spectrum of the tetra-n-butylammonium
salt of 5 does not display such a ν(CdS•) band. In this case
also no π radical anions are present and the monoanion
should be described as [CuIII(LBu)2]-.
The present IR results on [M(LBu)2]n-, [M(LBu)(LBu•)]m-
,
and [M(LBu•)2] complexes establish that the presence of one
or two π radical anions enhances the intensity of the ν(CdS•)
stretching mode to an extent where it is observed as a band
of medium intensity in the infrared spectrum.
The electronic spectra of all complexes have been recorded
in CH2Cl2 solutions in the range 300-1500 nm; the results
are summarized in Table 4. Figure 5 exhibits the spectra of
5350 Inorganic Chemistry, Vol. 44, No. 15, 2005

Bis(benzene-1,2-dithiolato)metal Complexes
Figure 6. X-band EPR spectrum of 3a in CH2Cl2 solution at 10 K.
Conditions: frequency, 9.6355 GHz; modulation, 10 G; power, 504 W.
The X-band rhombic EPR spectrum of (AsPh4)[NiII(LBu)-
(LBu•)] in CHCl3/DMF solution has been reported:7 g1 ) 2.18;
g2 ) 2.04; g3 ) 2.01. By using a 61Ni-enriched sample,
3
hyperfine coupling with the 61Ni (I ) /2) nucleus was
observed: A1(61Ni) ) 131(15), A2(61Ni) < 17, and A3(61Ni)
) 57(10) mT. This has been interpreted7 as a strong
indication for “considerable metal contribution to the SOMO”.
The electronic structure calculations described below provide
metal contributions to the SOMO on the order of 30-35%
depending on the method of population analysis. Similar
values have been reported for (NR4)[Ni(S2C2(CN)2)2]32 and
(NR4)[Ni(xylene-1,2-dithiolate)2].33 The spectrum of [Ni(L)2]-
is also very similar.2-5
The X-band EPR spectrum of 3a in CH2Cl2 solution at
10 K is shown in Figure 6. A rhombic signal is observed:
g ) 2.21, 2.06, and 1.80. Hyperfine coupling to the 195Pt
(I ) /2; 33% abundance) with Axx ) 320 MHz, Ayy ) 278
MHz, and Azz ) 227 MHz is recorded indicating significant
metal spin density to be present.
Figure 5. Electronic spectra of 1 (green), 1a (red), and 1b (blue) (top), of
2 (black), 2a (red), and 2b (blue) (middle), and of 3 (brown), 3a (red), and
3b (blue) in CH2Cl2 solutions (0.10 M [N(n-Bu)4]PF6) at -25 °C. The
neutral forms 1-3 were generated electrochemically as have been the
dianionic forms 1b, 2b, and 3b.
3
In contrast, the corresponding spectrum for 2a at 10 K
metal charge-transfer bands (LMCT) and intraligand transi-
tions <400 nm. It is noteworthy that [AuIII(LBu)(LBu•)]0
displays an IVCT band at 1452 nm (ꢀ ) 2.7 × 104 M-1
cm-1) which is absent in [AuIII(LBu)2]-.8 The electronic
spectrum of 5 in CH2Cl2 solution, on the other hand, does
not display a prominent feature in the near-infrared which
could be assigned to a LLCT or IVCT transition but instead
shows an absorption band at 398 nm (ꢀ ) 4 × 104 M-1
cm-1). On the basis of the calculations presented below, it
will be argued that this transition is a LMCT band similar
to that observed for the dianions 1b, 2b, and 3b. In agreement
with the results from IR spectroscopy, the monoanion is
therefore a genuine Cu(III) species [CuIII(LBu)2]-.
5
does not display 105Pd (I ) /2; 25% abundance) hyperfine
coupling. An isotropic signal at giso ) 2.02 is observed. It is
noted that the EPR spectrum of [AuIII(LBu•)(LBu)] (4) also
does not display hyperfine coupling to 197Au (I ) 3/2, 100%
natural abundance).8 However, before the conclusion that the
metal spin density must be highest for 3a, it should be
recalled that the nuclear g-value of 195Pt (gN ) 1.219) is
roughly 5 times larger than that of 105Pd (gN ) -0.256) and
12 times larger than that of 197Au (gN ) 0.097). Thus, the
hyperfine coupling constants (HFCs) of the central Pd might
have been detected if the metal spin density were comparable
to the Pt case whereas in the Au case the metal HFC splitting
is likely to buried within the line width of the experiment.
X-ray Crystallography. Sellmann et al.7 reported the
crystal structures of [NiII(LBu•)2] (1), [AsPh4][NiII(LBu•)(LBu)]
(1a), and [AsPh4]2[NiII(LBu)2]‚2Et2O (1b). The results are
summarized in Figure 7. The structures determined at 293
K are of good quality. The experimental errors of the C-C
Magnetic Properties and EPR Spectra. Complexes 1-3,
4a, and 5 are diamagnetic; they possess an S ) 0 ground
1
state as was judged from SQUID magnetometric and H
NMR measurements. The same is true for the species 1b,
2b, and 3b, which are also diamagnetic. In contrast,
complexes 1a, 2a, and 3a are paramagnetic. The temperature-
independent (50-300 K) effective magnetic moments are
observed in the range 1.7-1.9 µB for 1a, 2a, and 3a
(32) Maki, A. H.; Edelstein, M.; Davison, A.; Holm, R. H. J. Am. Chem.
Soc. 1964, 86, 4580.
(33) Kirmse, R.; Stach, J.; Dietzsch, W.; Steinecke, G.; Hoyer, E. Inorg.
Chem. 1980, 19, 2679.
1
indicative of an S ) /2 ground state, respectively.
Inorganic Chemistry, Vol. 44, No. 15, 2005 5351

Ray et al.
Table 5. Bond Distances (Å) in the Anions of 2a, 2b*, and 5′
5
2a
2b*
C1-S1
C2-S2
C1-C2
C2-C3
C3-C4
C4-C5
C5-C6
C1-C6
M-S1
1.751(3)
1.750(3)
1.404(4)
1.402(4)
1.375(4)
1.400(4)
1.389(4)
1.435(4)
2.2560(8)
2.2633(8)
1.758(3)
1.764(3)
1.400(4)
1.406(4)
1.382(4)
1.395(4)
1.385(4)
1.404(4)
2.3009(7)
2.3087(7)
1.769(4)
1.751(4)
1.398(6)
1.386(6)
1.382(7)
1.389(7)
1.404(7)
1.409(6)
2.170(1)
2.160(1)
1.770(4)
1.767(5)
1.394(6)
1.415(5)
1.366(8)
1.405(8)
1.404(5)
1.419(6)
2.163(1)
2.173(1)
Figure 7. Selected average bond distances (Å) for square planar complexes
[Ni(LBu)2] (red), [AsPh4][Ni(LBu)2] (green), and [AsPh4]2[Ni(LBu)2] (black)
from ref 7. The average error for C-C and C-S bond distances is (0.01
Å (3σ).
and C-S bond distances are of the order (0.01 Å (≡3σ).
Therefore, we wish to argue that the observed decrease of
the average C-S bond length on going from 1b (average
1.765 Å) to 1a (average 1.745 Å) and to 1 (average 1.725
Å) reflects a real ligand-centered stepwise oxidation of two
closed-shell aromatic dianions (LBu)2- in 1b to one such
ligand and an additional π radical monoanion (LBu•)- in 1a
to two such radicals in the neutral species 1. The oxidation
state of the central nickel ion remains +II (d8, SNi ) 0)
throughout the electron-transfer series. In particular, dia-
magnetic square planar 1 cannot contain a Ni(IV) (d6) ion
and two (LBu)2- ligands since an S ) 1 ground state as is
observed for isoelectronic [CoIII(LBu)2]- 34 would prevail
according to simple ligand field considerations. The six C-C
bonds of the six-membered ligand rings are nearly equidistant
in 1b (average 1.40 Å) but display weak quinoid-like
distortions in 1a and more pronounced so in 1. This is
excellent evidence that the HOMO in 1b possesses ligand
character with little metal d-orbital contributions.
M-S2
dianion in 2b* (middle), and the monoanion in 5′ (bottom).
All three species possess a planar MS4 polyhedron. The
tertiary butyl groups in 2a and 5′ are in a trans-position
relative to each other.
As expected, the C-S bonds in the dianion of 2b* are
long at average 1.76 ( 0.01 Å and the six C-C bonds are
equidistant at 1.395 ( 0.01 Å. Therefore, the closed-shell,
dianionic form of the ligand, (L)2-, is coordinated to a Pd(II)
ion (d8, S ) 0) as in square planar [(bpy)PdII(L)]0 38 and
[(bpy)PtII(L)]0.39
In the monoanion of 2a the C-S bonds are slightly shorter
at average 1.75 ( 0.01 Å and the six C-C bonds display
Interestingly, in similar complexes containing the unsub-
stituted benzene-1,2-dithiolate ligands as in [N(CH3)4]2[Ni-
(L)2] the average C-S bond at 1.762(8) Å is long and the
six C-C distances are within experimental error equidistant
(average 1.40 Å).35 In contrast, in two structure determina-
tions of salts containing the monoanion [Ni(L)2]- the C-S
bonds are always shorter at average 1.74 Å and the six C-C
bonds display two alternating shorter ones and four longer
ones (quinoid-like distortion).36 Similar results have been
reported for the monoanion [Ni(LMe•)(LMe)]-.37
In this work we have determined the structures of [N(n-
Bu)4][PdII(LBu•)(LBu)] (2a), [PPh4]2[PdII(L)2] (2b*), and [AsPh4]-
[CuIII(LBu)2] (5′) at 100 K by X-ray crystallography. Table
5 summarizes important bond lengths; Figure 8 shows the
structures of the monoanion in crystals of 2a (top), the
(34) Ray, K.; Begum, A.; Weyhermu¨ller, T.; Piligkos, S.; van Slagaren, J.;
Neese, F.; Wieghardt, K. J. Am. Chem. Soc. 2005, 127, 4403-4415.
(35) Sellmann, D.; Fu¨nfgelder, S.; Knoch, F.; Moll, M. Z. Naturforsch.
1991, 46b, 1601.
(36) Mahadevan, C.; Seshasayee, M.; Kuppusamy, P.; Manoharam, P. T.
J. Crystallogr. Spectrosc. Res. 1985, 15, 305.
(37) (a) Ochocki, J.; Chaudhuri, P.; Driessen, W. L.; de Graaf, R. A. G.;
Hulsbergen, F. B.; Reedijk, J. Inorg. Chim. Acta 1990, 167, 15. (b)
Kamenicek, J.; Valach, F.; Kratochvil, B.; Zak, Z. Pol. J. Chem. 2002,
76, 483. (c) Robertson, N.; Parsons, S.; Awaga, K.; Fujita, W. Cryst.
Eng. Commun. 2000, 22.
(38) Cocker, T. M.; Bachman, R. E. Inorg. Chem. 2001, 40, 1550.
(39) Connick, W. B.; Gray, H. B. J. Am. Chem. Soc. 1997, 119, 11620.
Figure 8. Structures of the monoanions in crystals of 2a (top) and 5′
(bottom) and of the dianion in crystals of 2b*.
5352 Inorganic Chemistry, Vol. 44, No. 15, 2005

Bis(benzene-1,2-dithiolato)metal Complexes
signs of a quinoid-like distortion. The two alternating C-C
bonds C5-C6 at 1.389(4) Å and C3-C4 at 1.375(4) Å
are shorter than the other four C-C bonds at average 1.41
( 0.01 Å. The structure is very similar to that of
[Ni(LBu•)(LBu)]- 7 and is in accord with the notion that one-
electron oxidation from the dianion to the corresponding
monoanion (1b f 1a; 2b f 2a; 3b f 3a) is a predominantly
ligand-centered process.
The structure of the monoanion in crystals of 5′ on the
other hand clearly shows that both ligands are in the closed-
shell dianionic form (LBu)2- rendering the central copper ion
trivalent (CuIII; d8, S ) 0). The average C-S bond length is
long at 1.764 ( 0.01 Å, and the C-C distances do not show
a quinoid-like distortion. This interpretation is in contrast to
Sawyer et al.,14 who reported a crystal structure of [N(n-
Bu)4][Cu(LMe)2] which is of rather low quality with an error
of (0.1 Å for the C-S bond. These authors concluded from
electrochemical results that the monoanion possesses a
[CuII(LMe•)(LMe)]- electronic structure. Two other structure
determinations of [PPh4][Cu(L)2] and [PPh4][Cu(LMe)2] also
have too large errors of ∼0.02 and 0.03 Å, respectively, for
the C-S bond distances to determine unambiguously the
oxidation level of the ligands.40
It is instructive to study the electronic structure of the
uncoordinated ligand (L)2- and its planar dimer {(L)2}4- first.
The MO diagrams obtained from BP86 calculations are
shown in Figure 9: the resulting MO’s may be divided into
out-of-plane (π, a2, and b2) and in-plane (σ, a1, and b1)
orbitals. The redox-active MO is 2b2, which is the HOMO
for (L)2- and the SOMO for (L•)-. The 2b2 MO is bonding
between C4 and C5 but antibonding between C3 and C4 and
C5 and C6 and also between the sulfur atoms and the adjacent
carbon atoms C1 and C6, respectively. Thus, upon oxidation
of the aromatic closed-shell dianion (L)2- to the semiquinon-
ate form (L•)- the C4-C5 bond expands slightly whereas
the C3-C4 and C5-C6 and both C-S bonds shrink. The
effect on the six-membered ring is small and inside the
normal experimental error limit of a single-crystal X-ray
structure determination, where an error of (0.01 Å is
considered to be normal. In contrast, the C-S bond distance
changes are outside the experimental error limits. The
calculated spin density of the H(L•) radical in Figure 9 shows
a large spin density at the deprotonated sulfur atom and
limited delocalization of spin density on the six-membered
phenyl ring. These results are in stark contrast to genuine
oxygen-containing semiquinonates where delocalization of
spin density onto the ring is much more pronounced.41
Calculations. 1. Ligands. Due to the localization of spin
density at the sulfur atom the two-electron-oxidized form of
(L)2- is calculated to be unstable against intramolecular
disulfide bond formation. The calculated structure of L is
shown in Figure 9 (bottom).
Figure 9. Top: MO diagram for benzene-1,2-dithiolate(2-) and its
hypothetical planar dimer. Bottom: Bond distances (Å) from a geometry
optimization (BP86, DFT, TZVP basis for sulfur atoms and SV(P) for the
remaining atoms) on benzene-1,2-dithiol (H2L) and its deprotonated one-
electron oxidized form (HL) and the fully deprotonated two-electron-
oxidized form (L). The spin density distribution of the radical HL (S )
1/2) is shown below.
their respective symmetry only the 1b2g orbital can undergo
a symmetry-allowed π interaction with a dxz orbital of a
central transition metal ion. The extent of this mixing is of
course dependent on the relative energies of the interacting
orbitals. Similarly, the symmetric and antisymmetric com-
binations of the 2a2 MOs give rise to the 1au and 1b3g MOs
where again only the 1b3g can π- interact with the dyz metal
orbital. Finally, the antisymmetric combination of the 1b1
MOs gives rise to a 1b1g orbital which can interact with the
dxy metal orbital.
2. [M(L)(L•)]z Species with S ) 1/2 Ground State (M )
Ni, Pd, Pt, z ) 1-; Au, z ) 0). (a) Geometries. As shown
in Table 6 the optimized calculated structures of paramag-
netic complexes [M(L)(L•)]z (M ) NiII, PdII, PtII, z ) 1-;
AuIII, z ) 0) are in excellent agreement with the experimental
findings. The slight overestimation of the M-S bond
distances is typical for present day DFT functionals. How-
ever, the metrical parameters for the dithiolate ligands are
For the ligand dimer {(L)2}4-, the positive and negative
combinations of the HOMO (2b2) of two L2- give rise to
the 1b1u and 1b2g orbitals (Figure 9), respectively. Due to
(41) (a) Pierpont, C. G.; Lange, C. W. Prog. Inorg. Chem. 1994, 41, 331.
(40) Mao-Chun, H.; Zhi-Ying, H.; Xin-Jiaw, L.; Ling-Hong, W.; Ham-
(b) Pierpont, C. G.; Buchanan, R. M. Coord. Chem. ReV. 1981, 38,
45.
Qin, L. J. Struct. Chem. 1991, 10, 237.
Inorganic Chemistry, Vol. 44, No. 15, 2005 5353

Ray et al.
Table 6. Experimental and Calculated (in Parentheses) Metrical Parameters for the Complexes in Å Obtained from Scalar Relativistic DKH2-BP86DFT
Methods Using Large Uncontracted Gaussian Bases at the Metal Center and Uncontracted All Electron Polarized Triple-ê (TZVP) Gaussian Bases for
the Remaining Atoms
complex
M-S
C-S
C1-C2
C2-C3
C3-C4
C4-C5
C5-C6
C1-C6
[Ni(L)2]2- a
[Ni(L)2]- a
[Ni(L)2]0 a
[Pd(L)2]2-
[Pd(L)2]- a
[Pd(L)2]0
2.173 (2.210)
2.146 (2.181)
2.126 (2.158)
2.301 (2.358)
2.263 (2.325)
(2.300)
1.762 (1.770)
1.747 (1.763)
1.727 (1.744)
1.761 (1.773)
1.751 (1.763)
(1.744)
1.426 (1.407)
1.425 (1.407)
1.429 (1.412)
1.419 (1.408)
1.435 (1.410)
(1.414)
1.398 (1.399)
1.387 (1.390)
1.375 (1.382)
1.385 (1.397)
1.389 (1.388)
(1.381)
1.390 (1.402)
1.412 (1.405)
1.422 (1.412)
1.395 (1.401)
1.400 (1.405)
(1.412)
1.383 (1.399)
1.367 (1.390)
1.373 (1.382)
1.382 (1.397)
1.375 (1.388)
(1.381)
1.402 (1.407)
1.391 (1.407)
1.409 (1.412)
1.406 (1.408)
1.402 (1.410)
(1.414)
1.401 (1.405)
1.407 (1.413)
1.419 (1.424)
1.400 (1.403)
1.404 (1.417)
(1.421)
[Pt(L)2]2-
[Pt(L)2]-
(2.318)
(2.288)
(2.261)
(1.770)
(1.756)
(1.739)
(1.408)
(1.408)
(1.411)
(1.397)
(1.389)
(1.381)
(1.402)
(1.406)
(1.412)
(1.397)
(1.388)
(1.381)
(1.408)
(1.408)
(1.411)
(1.405)
(1.416)
(1.419)
[Pt(L)2]0
[Cu(L)2]- a
[Au(L)2]-
[Au(L)2]0
2.168 (2.217)
2.310 (2.338)
2.300 (2.317)
1.768 (1.774)
1.764 (1.776)
1.735 (1.758)
1.419 (1.408)
1.397 (1.404)
1.402 (1.408)
1.404 (1.398)
1.386 (1.393)
1.374 (1.385)
1.405 (1.402)
1.392 (1.400)
1.415 (1.406)
1.366 (1.390)
1.382 (1.393)
1.384 (1.385)
1.415 (1.405)
1.402 (1.404)
1.420 (1.408)
1.394 (1.402)
1.397 (1.4 06)
1.406 (1.411)
a Experimental values for complexes containing 3,5-di-tert-butyl-1,2-benzenedithiolate ligand.
very accurately reproduced by the calculations with the
typical error in computed bond lengths not exceeding ∼0.02
Å.
the spin up and the spin down MOs are plotted in order of
increasing energy. Since this bonding scheme does not
change qualitatively for all [M(L)(L•)]z species under con-
sideration here, it is equally applicable for the Ni, Pd, and
Pt counterparts. The composition of important MOs is
summarized in Table 7.
The calculated average C-S bond length of 1.76 Å
corresponds to the arithmetic average of calculated C-S
bond distances at 1.79 Å for complexes containing only
closed-shell (L)2- dianions and at 1.73 Å for species
containing only (L•)- radicals. Thus, as expected, the
complexes behave like class III delocalized ligand mixed-
valent systems [M(L•)(L)]z T [M(L)(L•)]z in these calcula-
tions.
(b) Bonding. For the MO description of these paramag-
netic complexes within the D2h point group, we choose the
z axis to be along the normal of the complex; the x axis is
then along the long axis of the complex and the y axis along
the short axis as shown in Figure 10. The qualitative bonding
scheme derived from the spin-unrestricted B3LYP DFT
calculation on the [AuIII(L)(L•)]0 species is also shown where
The ground-state electronic configuration of the [M(L•)-
(L)]z species is predicted to be
(1ag)2(2ag)2(1b3g)2(1b2g)2(1au)2(2b3g)2(1b1u)2(2b2g)1(1b1g)0
2
The ground state B2g is in agreement with earlier results
from extended Hu¨ckel calculations on the [Ni(L•)(L)]-
complex.10-12 The same ground state has been proposed for
[M(L)(L•)]- (M ) Ni, Pd, Pt) complexes from an analysis
of their EPR spectra.2-5
The bonding scheme in Figure 10 reveals that in general
four metal d orbitals are lower in energy as compared to the
ligand-based 1b2g and 1b3g orbitals shown in Figure 9 which
undergo symmetry-allowed π interactions with metal d
orbitals. This is consistent with recent calculations for the
[NiII(S2C2Me2)2]- complex.42 In these and our present
2
2
calculations four doubly occupied orbitals, namely 1ag(dx -y ),
2
2ag(dz ), 1b3g(dyz), and 1b2g(dxz) are predominantly of metal-d
character and, hence, the physical oxidation state of the
metals is best represented as d8 MII (Ni, Pd, Pt) and AuIII
ions.
The LUMO of these complexes is the antibonding
combination of the metal dxy and the ligand 1b1g fragment
orbitals. The overlap between these two orbitals is favorable
and provides an efficient pathway for ligand-to-metal σ
electron donation. The SOMO features mainly the π* 2b2g
MO of the ligand which mixes with the metal dxz orbital.
This interaction accounts for metal-to-ligand π electron
donation in these paramagnetic complexes.
(c) Population Analysis. Table 8 summarizes the charge
and spin populations at the respective metal ion resulting
Figure 10. Kohn-Sham MOs and energy scheme for [Au(L)(L•]0 from
a spin-unrestricted ZORA-B3LYP DFT calculation.
(42) Szilagyi, R. K.; Lim, B. S.; Glaser, T.; Holm, R. H.; Hedman, B.;
Hodgson, K. O.; Solomon, E. I. J. Am. Chem. Soc. 2003, 125, 9158.
5354 Inorganic Chemistry, Vol. 44, No. 15, 2005

Bis(benzene-1,2-dithiolato)metal Complexes
Table 7. Composition of Selected Molecular Orbitals of [M(L)2]z Complexes (%) As Obtained from the Scalar Relativistic ZORA-B3LYP DFT
Calculations Using Large Uncontrated Gaussian Bases at the Metal and Uncontracted All Electron Polarized Triple-ê (TZVP) Gaussian Bases for the
Remaining Atoms
complex
MO
M(ndyz)
M(ndxz)
M(ndxy)
S(3pz)
S(3px,y
)
C(2pz)
C(2px,y)
[Ni(L)2]2-
2b3g
2b2g
1b1g
2b3g
2b2g
1b1g
2b3g
2b2g
1b1g
2b3g
2b2g
1b1g
2b3g
2b2g
1b1g
2b3g
2b2g
1b1g
2b3g
2b2g
1b1g
2b3g
2b2g
1b1g
2b3g
2b2g
1b1g
2b3g
2b2g
1b1g
2b3g
2b2g
1b1g
2b3g
2b2g
1b1g
71
17
36
8
10
52
50
46
43
41
40
39
39
41
40
33
33
33
40
43
48
46
48
48
38
42
43
58
55
55
2
5
3
3
3
4
1
2
2
4
5
5
[Ni(L)2]-
[Ni(L)2]0
[Pd(L)2]2-
[Pd(L)2]-
[Pd(L)2]0
[Pt(L)2]2-
[Pt(L)2]-
[Pt(L)2]0
55
49
44
33
28
42
33
29
17
11
9
30
48
10
14
34
26
29
21
16
30
24
20
11
9
34
51
12
16
37
52
18
17
44
56
19
21
43
56
23
25
36
52
20
17
44
56
19
18
44
56
23
22
[Cu(L)2]-
[Au(L)2]-
[Au(L)2]0
53
58
30
27
58
64
30
26
51
61
38
30
8
Table 8. Comparison of the Charge and Spin Populations at the Metal
Ion Resulting from a Natural Population Analysis of the One-Electron
Density of the Ground State Obtained from Scalar Relativistic
ZORA-B3LYP DFT Calculationsa
density of 0.06 at the gold(III) ion. This species carries the
spin density of one unpaired electron in a predominantly
ligand-based 2b2g orbital with only 8% (from Lo¨wdin
population) metal d character.
complex nd electrons (n+1)s electrons nd spin metal oxidn state
In contrast, the spin density at the central nickel ion in
[Ni(L)(L•)]- is larger at 0.31 indicating an enhanced Ni(III)
character in this species. The corresponding spin densities
at the central palladium and platinum ion in [MII(L)(L•)]-
complexes are between those of the nickel and gold case at
0.16 and 0.21, respectively.
[Ni(L)2]2-
[Ni(L)2]-
[Ni(L)2]0
[Pd(L)2]2-
[Pd(L)2]-
[Pd(L)2]0
[Pt(L)2]2-
[Pt(L)2]-
[Pt(L)2]0
9.13
9.01
8.99
9.24
9.16
9.08
9.07
8.97
8.92
8.92
8.99
8.90
0.51
0.53
0.54
0.43
0.47
0.48
0.74
0.76
0.78
0.49
0.63
0.65
0.00
0.31
0.00
0.00
0.16
0.00
0.00
0.21
0.00
0.00
0.00
0.05
NiII
NiII
NiII
PdII
PdII
PdII
PtII
PtII
PtII
(d) Calculation of EPR Parameters. Table 9 compares
the calculated EPR g-values with the experimental ones for
the [M(L)(L•)]z complexes of Ni, Pd, and Pt with z ) 1-
and of Au with z ) 0. Table 10 shows the individual isotropic
spin-dipolar and spin-orbit coupling contributions to the
metal HFC tensor.
[Cu(L)2]-
[Au(L)2]-
[Au(L)2]0
CuIII
AuIII
AuIII
a Spin restricted calculations are reported for the neutral [M(L)2]0 species
with M ) Ni, Pd, and Pt.
from a natural population analysis30a-c of the one-electron
density of the ground state obtained from scalar reletavistic
ZORA-B3LYP DFT calculations. The calculated nd popula-
tions for the complexes [M(L•)(L)]- (M ) Ni, Pd, Pt) and
[Au(L•)(L)]0 are consistent with the nd8 valence electron
configuration. The charge excess over the formal nd8
configuration arises from the population of the otherwise
unpopulated ndxy orbital (almost single occupation) due to
strong σ donation from the b1g ligand fragment orbital. The
neutral gold species [AuIII(L•)(L)] possesses a very low spin
As shown above, all of these species possess a common
2B2g ground state with an unpaired electron in the predomi-
nantly ligand-based 2b2g orbital. Since this SOMO transforms
“gerade” upon inversion, it can mix with the out-of-plane
metal dxz orbital whereby it acquires some metal character
(Table 7). This metal character gives rise to sizable first-
order dipolar metal hyperfine coupling. It also enhances the
spin-polarization of the metal core which is responsible for
the isotropic Fermi contact contribution to the hyperfine
coupling constant. Furthermore, the 2B2g ground state readily
Inorganic Chemistry, Vol. 44, No. 15, 2005 5355

Ray et al.
Table 9. Experimental {for [M(LBu)(LBu•)]n Complexes} and (in Parentheses) Calculated {for [M(L)(L•)]n Complexes} EPR Parameters Obtained from
the Scalar Relativistic ZORA-B3LYP DFT Methods Using Large Uncontracted Gaussian Bases at the Metal and Uncontrated All Electron Polarized
Triple-ê (TZVP) Gaussian Bases for the Remaining Atomsa
complex
gx
gy
gz
Ax,b MHz
Ay,b MHz
Az,b MHz
Wx,c G
Wy,c G
Wz,c G
[NiII(L)(L )]-
[PdII(L)(L )]-
[PtII(L)(L )]-
[AuIII(L)( L)]0
2.05 (2.07)
2.03 (2.04)
2.06 (2.06)
2.07 (2.04)
2.18 (2.19)
2.03 (2.05)
2.21 (2.13)
2.03 (2.03)
2.01 (2.02)
2.01 (1.98)
1.83 (1.91)
1.91 (1.95)
- (-25)a
- (32)a
- (-37)a
7
75
30
30
30
104
36
8
54
44
44
- (11)a
- (37)a
- (11)a
320 (-302)
- (-16)a
278 (-210)
227 (-129)
- (-17)a
- (-21)a
62
a Fermi contact, dipolar, and metal spin-orbit contributions to the hyperfine coupling are included in the calculations.
Table 10. Calculated Contributions of the Isotropic Fermi-Contact, Anisotropic Dipolar, and Second-Order Orbital Terms to the Hyperfine Coupling
Parameters in the [M(L)(L•)]n Complexes As Obtained from the Scalar Relativistic ZORA-B3LYP DFT Methods Using Large Uncontracted Gaussian
Bases at the Metal and Uncontracted All Electron Polarized Triple-ê (TZVP) Gaussian Bases for the Remaining Atoms
A
Dip, MHz
AOrb, MHz
ATot., MHz
complex
Aiso, MHz
x
y
z
x
y
z
x
y
z
[NiII(L)(L)]-
[PdII(L)(L)]-
[PtII(L)(L)]-
[AuIII(L)(L)]0
31
26
-355
-22
-45
-13
272
3
79
22
23
-5
-31
-11
127
2
-11
-2
-219
3
-78
-11
122
6
-37
-4
99
-25
11
-302
-16
32
37
-210
-21
-37
11
-129
-17
3
mixes with relatively low-lying d-d excited states thereby
giving rise to a sizable orbital angular momentum which
contributes to the hyperfine coupling as well as to the g-shifts
in the EPR spectra. From Table 10 it is obvious that all three
contributions are of comparable magnitude in the [M(L)-
(L•)]z complexes but partially cancel each other. Experimental
HFC numbers are only available for the Pt complex where
the calculations appear to give a reasonable result. The
remaining errors are still sizable as expected from the
complexity of the task (note that the sign of the HFC values
is not known from experiment). As pointed out above, in
[AuIII(L•)(L)]0 there is little mixing (8%) of the Au dxz orbital
in the 2b2g SOMO because the d manifold is low-lying in
energy due to the large effective nuclear charge of AuIII.
Together with the rather small nuclear moment of the 197Au
nucleus, HFC’s of only -16, -21, and -17 MHz are
calculated which would fall well within the line width of
the experiment.8 There is also little angular momentum in
the ground-state wave function, and thus, the calculated and
observed g-shifts are small and reflect the organic radical
character of the ground state.
[NiII(L)(L•)]- represents the case of the highest back-
bonding interaction (Zeff of Ni(II) being the lowest in energy
in the present series) with a metal character of 34% in the
SOMO (31% d-orbital spin population from NPA). The
calculated g-shift is relatively large at ∆g ∼ 0.15 in full
agreement the experimental results. Calculated 61Ni hyperfine
splittings of -25, +32, and -37 MHz are also in qualitative
agreement with values reported for related complexes (NR4)-
[Ni(S2C2(CN)2)2] (-15, 38, -12 MHz)32 and (NR4)[Ni-
(xylene-1,2-dithiolate)2] (9, 46, 6 MHz).33
nents are somewhat under- and overestimated, respectively,
probably reflecting the limitations of the simple semiem-
pirical spin-orbit operator used in this study. The metal
contribution to the 2b2g SOMO is 24% (21% d-orbital spin
population from NPA).
(d) Optical Spectra. The optical electronic spectra of
[M(LN)2]n and [M(LN,O)2]n complexes, where M ) Ni, Pd,
and Pt, n ) 0 and 1-, and LN represent an o-phenylenedia-
mide or an o-amidophenolate (LN,O) ligand, have been studied
experimentally and by DFT calculations.9 The analysis
showed that the very intense absorption maximum in the
near-infrared (NIR) region of the monoanions corresponds
to an intervalence charge-transfer transition (IVCT) of the
type (L)M(L•) T (L•)M(L). In the spectra of the present
benzene-1,2-dithiolato monoanionic complexes 1a-3a these
intense IVCT are also observed (Figures 5 and 11). Their
positions and intensities are dependent on the nature of the
central metal ion (Table 4).
For the paramagnetic [M(L)(L•)]z complexes there are 4
spin and electric dipole allowed transitions from the seven
doubly occupied MOs in Figure 10 into the singly occupied
2b2g and virtual 1b1g orbitals expected. The low-energy 1b1u
f 2b2g and 1au f 2b2g transitions are IVCTs in origin with
2
the former expected to be more intense (x-polarized B1u
state) than the latter (y-polarized 2Au state). The third allowed
1au f 1b1g transition (ligand-to-metal charge transfer (LMCT)
in character) leads to two 2B3u and one 4B3u states. Note that
in the excited state there are three unpaired electrons which
can spin couple to two doublets and one quartet state. The
2
transitions to the B3u states are allowed in z-polarization.
However, the second 2B3u component corresponds to a double
excitation and should have no intensity on its own. The two
2B3u states may interact, and the “dark” component may
borrow intensity from the “bright” one. As pointed out
previously,9 a single determinant description for the com-
plexes is an oversimplification and not appropriate for such
a complicated spin-coupling system in the excited states.
Therefore, we will limit the description of the electronic
structure of these complexes to the two IVCT bands in the
time-dependent DFT calculations. This appears even more
For [Ni(LBu•)(LBu)]- 61Ni HFCs of A1(61Ni) ) 16, A2(61Ni)
< 5, and A3(61Ni) ) 40 MHz have been observed7 in good
agreement with the calculated ones for [Ni(L•)(L)]-. The
calculated EPR parameters for [PdII(L)(L•)]- agree also well
with experiment; a small metal contribution of 21% of the
SOMO (16% d-orbital spin population from NPA) gives rise
to only small HFCs. For the [PtII(L)(L•)]- species the
calculated 195Pt hyperfine splittings are in acceptable agree-
ment with experiment, but the calculated gy and gz compo-
5356 Inorganic Chemistry, Vol. 44, No. 15, 2005

Bis(benzene-1,2-dithiolato)metal Complexes
trend of decreasing energy and increasing intensity on going
from the Ni to the Pt, to the Pd, and, finally, to the Au
species. The origin of this behavior may be traced back to
the relative energies of the donor and the acceptor orbitals
of the IVCT states. The main reason for the splitting of the
energies of the 1b1u and the 2b2g orbitals is attributed to the
strength of the interaction of the ligand b2g combination with
the metal dxz orbital. Since this interaction is antibonding, it
raises the energy of the 2b2g orbital with respect to the 1b1u.
An increase of the energy of the metal d orbitals will bring
them closer in energy to the ligand orbitals, and superex-
change through the central metal ion will be more effective
yielding an increased destabilization of the 2b2g level with
respect to the 1b1u and 1au levels. A blue shift of the IVCT
bands results. In the present paramagnetic complexes 1a, 2a,
3a, and 4 the variation of the effective nuclear charge
combined with relativistic effects at the respective central
metal ion increases the energy of the nd orbitals in the
following order: Au < Pd < Pt < Ni. Accordingly, the
calculated energies of the IVCT bands increase on going
from 4 to 2a, to 3a, and to 1a in excellent agreement with
experiment (Figure 11).
3. Diamagnetic [M(L)2]z Species (M) Ni, Pd, Pt, z )
2-; Cu, Au, z ) 1-). (a) Geometry. The calculated
structures for [M(L)2]z [M ) Ni, Pd, Pt (z ) 2-); M ) Cu,
Au (z ) 1-)] feature long C-S bond distances at an average
of 1.77 Å. The C-C bond lengths within the phenyl rings
are essentially equivalent at an average distance of 1.403 Å.
This indicates the presence of two closed-shell benzene-1,2-
dithiolate(2-) ligands in each complex rendering the oxida-
tion state +II for the Ni, Pd, and Pt complexes and +III for
the Au species (nd8, SMetal ) 0).
(b) Bonding. Upon one-electron reduction of the [M(L•)-
(L)]z species the additional electron enters the 2b2g orbital
which becomes the HOMO of these reduced species. The
calculated ground-state electronic configuration is therefore
the following: (1ag)2(2ag)2(1b3g)2(1b2g)2(1au)2(2b3g)2(1b1u)2-
(2b2g)2(1b1g)0. This reduction process is accompanied by an
increase in the metal character of the 2b2g orbital (Table 7),
but the composition of the LUMO remains unaltered. The
natural population analysis for the diamagnetic [M(L)2]z
complexes reveals nearly identical nd populations as has been
established for their paramagnetic counterparts [M(L•)(L)]z
(Table 8). It is therefore appropriate to describe these
[M(L)2]z species as NiII, PdII, PtII, CuIII, and AuIII complexes
Figure 11. Experimental and calculated absorption spectra for [M(L)-
(L•]z species. The experimental values are for complexes 4, 2a, 3a, and 1a
(from top to bottom) containing 3,5-di-tert-butylbenzene-1,2-dithiolate
ligands.
justified as the components of the 1au f 1b1g transition are
expected to appear at rather high energies. They do not
appear in the calculated first 25 (energy range 5000-25 000
cm-1) states for these complexes.
The room-temperature absorption spectra of 1a, 2a, 3a,
and 4 in CH2Cl2 have been subjected to a Gaussian
deconvolution (20 000-5000 cm-1). The results are sum-
marized in the Supporting Information Table S1. In each
case the spectrum is dominated by two intense transitions
labeled 1 and 2 in Figure 11 followed by some weaker
transitions which most probably correspond to d-d transi-
tions. Bands 1 and 2 are assigned to the IVCT transitions
1b1uf 2b2g and 1au f 2b2g, respectively. The experimental
transition energies are found to be in reasonable agreement
with the calculated ones given in Table 11. Both are
dependent on the nature of the central metal ion and show a
Table 11. Analysis of the Intervalence Charge Transfer Bands in the [M(L)(L•)]z Complexes Following Gaussian Deconvolution of the Experimental
Data (in CH2Cl2 Solution for [M(LBu)(LBu•)]z Complexes) Shown in Figure 11 Combined with the Results of the Scalar Relativistic ZORA-B3LYP
TD-DFT Calculation in Vacuum Using Large Uncontracted Gaussian Bases at the Metal and Uncontracted All Electron Polarized Triple-ê (TZVP)
Gaussian Bases for the Remaining Atoms
energy, cm-1
oscillator strength
complex
transition
expt
calcd
expt
calcd
[AuIII(L)(L)]
1b1u f 2b2g
1au f 2b2g
1b1u f 2b2g
1au f 2b2g
1b1u f 2b2g
1au f 2b2g
1b1u f 2b2g
1au f 2b2g
7 112
7 800
8 874
10 260
10 570
11 220
12 905
14 358
15 732
13 819
14 017
0.122
0.028
0.131
0.044
0.105
0.023
0.065
0.025
0.277
0.019
0.236
0.031
0.292
0.032
0.182
0.023
[PdII(L)(L)]-
[PtII(L)(L)]-
[NiII(L)(L)]-
9 750
11 320
12 300
11 231
12 372
Inorganic Chemistry, Vol. 44, No. 15, 2005 5357

Ray et al.
not affected on going from the dianionic to the monoanionic
and to the neutral form.
(c) Ab initio Calculations. It has recently been shown9
that the neutral, diamagnetic square planar complex [NiII-
(LNISQ)2], where (LNISQ)- represents the o-diiminobenzosemi-
quinonate(1-) π radical, is best viewed as a central low-
spin Ni(II) ion (d8, SNi ) 0) coordinated by two benzo-
semiquinonate(1-) ligands (singlet diradical). This means
that a single determinant closed-shell description is not an
appropriate starting point for a quantitative description of
the system. As we show here, the same is true for complexes
[MII(L•)2]0 (M ) Ni, Pd, Pt). A description on a quantitative
level using correlated ab initio electronic structure methods
has been developed previously9 and is used here.
Figure 12. Electronic spectrum of 5 in CH2Cl2 solution at 25 °C (solid
line, experimental spectrum; broken line, calculated spectrum): (a) hyperfine
splitting experimentally not observed; (b) sign of the hyperfine coupling
parameters not determined experimentally; (c) experimental line width.
We use the spectroscopy oriented configuration interaction
(SORCI)13 method to compute the electronic properties of
the [MII(L•)2]0 (M ) Ni, Pd, Pt) complexes. The SORCI
formalism is not restricted to species where a single
determinant is dominating the wave function but is a genuine
multiconfigurational ab initio method. On the basis of a CAS-
(2,2) reference space, which is the minimum required to
properly describe a diradical situation with two unpaired
electrons, the SORCI method explicitly includes up to
quadruple excitations from the closed-shell determinant.
These high excitations are required to properly capture the
differential dynamic correlation effects. This is particularly
important to describe the balance between the ground-state
determinant and the contributions of the (1b1u)2(2b2g)0 f
(1b1u)0(2b2g)2 double excitation properly. The relative weights
of these two configurations determine the amounts of the
ionic and neutral character in the ground-state wave function
which is used to estimate the diradical character from9
with an nd8 electron configuration (SM ) 0) and two closed-
shell benzene-1,2-dithiolate ligands. The DFT calculations
on the [CuIII(L)2]- species are in excellent agreement with
those reported for [CuIII(pdt)2]-, where pdt2- represents the
dianion of pyrazine-2,3-dithiol.16 These and the previous
calculations clearly rule out an electronic structure as in [CuII-
(L)(L•)].14
(c) Optical Spectra. Due to the double occupancy of the
2b2g HOMO there is only one spin and electric dipole allowed
1au f 1b1g transition in these [M(L)2]z species. This is a
LMCT transition to the vacant dxy orbital which is to be
expected at a higher energy (∼25 000 cm-1). Since intense
transitions in the near-infrared region are absent, the expected
d-d transitions cannot borrow intensity from them and,
hence, they are not observed in the experimental spectra.
The calculated absorption spectra for the [CuIII(L)2]-
species is found to be in reasonable agreement with the
experiments (Figure 12). The band at 363 nm (experimental
398 nm) is ligand-to-metal charge transfer (LMCT) in
character. The other band at 343 nm (experimental 303 nm)
is probably an intraligand n f π* ligand transition. The fact
that no intense IVCT bands are observed >400 nm is a clear
indication that the electronic structure is compatible only with
an electronic structure as in [CuIII(LBu)2]-. Note the difference
between the spectra of the monoanions [Cu(LBu)2]- in Figure
12 and [Ni(LBu•)(LBu)]- in Figure 5 (top).
4. Diamagnetic, Neutral [M(L•)2] Species (M ) Ni, Pd,
Pt). (a) Geometry from DFT. The DFT-calculated geom-
etries of the diamagnetic neutral molecules [MII(L•)2] are in
excellent agreement with the notion that the metal ions are
divalent (nd8, S ) 0) and S,S-coordinated to two equivalent
1,2-dithiosemiquinonate(1-) ligands. The average C-S bond
length at 1.74 Å is short, and the six-membered rings exhibit
each two alternating relatively short C-C bonds (C3-C2;
C4-C5) at 1.382 Å and four longer ones at an average 1.415
Å. The overall effect of this quinoidal distortion is small.
(b) Bonding from DFT. In the [MII(L•)2]0 species (M )
Ni, Pd, Pt) the 2b2g orbital is then the LUMO and its metal
dxz character is smaller than in the corresponding mono- and
dianions (Table 7). The calculated ground state is as
follows: (1ag)2(2ag)2(1b3g)2(1b2g)2(1au)2(2b3g)2(1b1u)2(2b2g)0-
(1b1g)0. The natural population analysis for the [MII(L•)2]0
species shows again that the nd8 electron configuration is
2
2
c0 cd
% diradical ) 200
(3)
2
c
2 + cd
x
0
2
2
where c0 and cd are the weights of the closed-shell
determinant and the double excitation in the final CI wave
function. Furthermore, the SORCI method allows for the
calculation of more accurate transition energies for the
(1b1u)2(2b2g)0 f (1b1u)1(2b2g)1 single and (1b1u)2(2b2g)0 f
(1b1u)0(2b2g)2 double excitations in [M(L•)2] complexes with
M ) Ni, Pd, and Pt than is possible within a DFT framework.
Third, the singlet-triplet energy gaps (-2JGS) can also be
calculated on the basis of pure spin wave functions which
in turn give the estimation of the ground-state effective
exchange coupling constants JGS(Hex ) -2JGSSa‚Sb) in the
above complexes.
Since we are dealing with heavy metals, scalar relativistic
effects were included in the calculations using the ZORA
method. The results of the scalar relativistic ZORA-SORCI
calculations for the [M(L•)2] species with M ) Ni, Pd, and
Pt are summarized in Table 12. To put the present results
into perspective, we have also calculated the corresponding
quantities for the [Ni(LNISQ)2] and [Ni (LOSQ)2] complexes
5358 Inorganic Chemistry, Vol. 44, No. 15, 2005

Bis(benzene-1,2-dithiolato)metal Complexes
Table 12. Calculated (ZORA-SORCI) Absorption Spectra of the [M(L•)2], [M(LNISQ)2], and [M(LOISQ)2] Species and Their Comparison with the
Experiments
LLCT energy, cm-1
103fosc
complex
singlet-triplet gap, cm-1
% diradical
calcd
expt
calcd
expt
[Ni(L)2]0
9734
5799
10423
5019
1397
32
50
30
48
78
10 266
12 851
15 716
13 584
13 516
11 508a
11 326a
12 669a
12 660b
10 890c
403
464
635
611
318
135a
265a
208a
287b
∼64c
[Pd(L)2]0
[Pt(L)2]0
[Ni(LNISQ)2]0
[Ni(LOSQ)2]0
a Experimental values obtained in CH2Cl2 solutions for complexes involving the 3,5-di-tert-butyl-1,2-benzenedithiolate ligands. b From ref 9. c Measured
in hexane (Lange, C. W.; Pierpont, C. G. Inorg. Chim. Acta 1997, 263, 21). The oscillator strength was estimated from the published extinction coefficient
by assuming the same band shape as in ref 9 for the corresponding nitrogen complex.
using the same methods and basis sets.43 The singlet-triplet
separation is found to be larger in the [M(L•)2] complexes
as compared to that in the [M(LNISQ)2] and [M(LOSQ)2]
complexes, which is reflected in the lower diradical character
in the former complexes. On the basis of our previous
estimates of the singlet-triplet gap in [M(LNISQ)2] of ∼3100
cm-1, the present calculations appear to overestimate the
singlet-triplet separation by ∼0.2 eV. The chemical trends
from the present calculations are nevertheless believed to
be reliable.
The results obtained for the composition of the singlet
ground-state wave function and the diradical character are
compatible with the comparative capability of the ligands
in the two cases to stabilize the unpaired π electrons by
delocalization over the phenyl rings. This delocalization is
determined by the semiquinone character of the ligands,
which is related to the capability of the N and S atoms to
form double bonds with the phenyl rings. This double bond
forming tendency increases on going from S to N donor
atoms. The weak double bond forming tendency of S atoms
is also reflected in the relatively large accumulated spin
density (53%, Figure 9) at the S atoms of (L•)-.
The intense near-IR transition in a large variety of metal
dithiolene complexes had been interpreted earlier as simple
HOMO-LUMO transitions.11,12 As has been explained
previously,9 this interpretation is valid only if the ground-
state wave functions for the complexes contain equal
contributions from the ionic and the neutral terms, i.e., for
complexes with a diradical index of zero. However, in the
present case, the diradical index for the [M(L•)2] complexes
is found to be too high (Table 12) to assign their intense
near-IR transitions as simple HOMO-LUMO transitions.
They should rather be interpreted as LLCT transitions.
The SORCI calculations show increasing diradical char-
acter and a decreasing singlet-triplet (-2JGS) energy gap
on going from the platinum to the nickel and to the palladium
species (Table 12). Thus, the results of the present analysis
show that the antiferromagnetic interaction between radicals
through the central palladium is weaker than through nickel,
whereas the interaction through the central platinum is the
strongest in the series. Similar trends were also obtained for
the [M(LNISQ)2] species.9
The experimental absorption spectrum of the [M(L•)2]
species with M ) Ni, Pd, and Pt between 5000 and 20 000
cm-1 together with a Gaussian fit is shown in the Supporting
Information Figure S1. The results are summarized in the
Supporting Information Table S2. It is evident that more than
one absorption band is required to fit the experimental
spectrum in each case. However, the spectrum is strongly
dominated by an intense transition LLCT band (11 508 cm-1
for Ni, 11 326 cm-1 for Pd, 12 669 cm-1 for Pt) in each
case, followed by some weaker bands. As explained earlier,
these low-intensity bands probably correspond to a compli-
cated series of d-d transitions with varying spin coupling
to the open-shell ligands; they are not taken care of in the
present SORCI calculations.
As evident from Table 12, the calculated transition energies
are found to be in reasonable agreement with the experi-
ments. However, the intensities are unfortunately not well
reproduced in the calculations. It is particularly puzzling that
the Pt species experimentally has the LLCT intensity between
the Ni and Pd case while from qualitative considerations12
and the ab initio calculations it is expected to have maximum
intensity in the whole series.
Conclusions
The electronic structures of square planar complexes
containing the diamagnetic dianions 1b, 2b, and 3b, the
paramagnetic monoanions 1a, 2a, and 3a, and, finally, the
diamagnetic neutral compounds 1-3 have been elucidated
experimentally and by density functional theory and cor-
related ab initio methods (SORCI). The most salient features
may be summarized as follows:
[MII(L)2]2- (M ) Ni, Pd, Pt). These diamagnetic square
planar dianions consist of a divalent metal ion (Ni, Pd, Pt)
and two S,S-coordinated, closed-shell benzene-1,2-dithiolate-
(2-) ligands.
(43) The results of the structure optimization using the BP86/ZORA method
for [Ni (LOSQ)2] resulted in excellent agreement with the X-ray
structure of [Ni(3,6-di-tert-butyl-o-benzosemiquinone)2] reported by
Abakumov et al. (Abakumov, G. A.; Cherkasov, V. K.; Bubnov, M.
P.; Ellert, O. G.; Rakitin, Y. V.; Zakharov, L. N.; Struchkov, Y. T.;
Saf’yanov, Y. N. Bull. Russ. Acad. Sci. 1992, 41, 1813. Translated
from: IzV. Acad. Nauk, Ser. Khim. 1992, 10, 2315) with all metal-
ligand and intraligand distances reproduced within 0.02 Å of their
experimental values. The compound was reported to be diamagnetic
up to 470 K.
[MII(L)(L•)]- (M ) Ni, Pd, Pt). These square planar,
1
paramagnetic (S ) /2) monoanions consist of a divalent
metal ion (Ni, Pd, Pt; nd8, SM ) 0) and one S,S-coordinated
benzene-1,2-dithiolate(2-) ligand as well as one S,S-
coordinated benzene-1,2-dithiosemiquinonate(1-) π radical
anion. This ligand mixed valency is of class III (delocalized).
An intense IVCT band in the NIR is observed in all three
Inorganic Chemistry, Vol. 44, No. 15, 2005 5359

Ray et al.
cases. The 2b2g MO is the SOMO; it contains a small M(ndxz)
metal character of 34% for the Ni, of 21% for the Pd, of
24% for the Pt, and of only 8% for the [AuIII(L)(L•)]0
complex.
the investigated systems. Similarly, the calculated singlet-
triplet gaps increase from ∼1400 cm-1, to ∼5000 cm-1, and
to ∼9700 cm-1 along the same series. The electronic spectra
of 1-3 display very intense LLCT bands in the NIR region.
[M(L)2]- (M ) CuIII, AuIII). Both complexes consist of
a trivalent metal ion with nd8 configuration and two S,S-
coordinated benzene-1,2-dithiolate(2-) ligands. One electron
oxidation yields the mixed-valence neutral species [CuIII(L)-
(L•)] and [AuIII(L•)(L)], respectively. In contrast, the one-
electron reductions are metal-centered and yield [CuII(L)2]2-
[MII(L•)2] (M ) Ni, Pd, Pt). These square planar
diamagnetic neutral compounds consist of a divalent metal
ion (Ni, Pd, Pt; nd8, SM ) 0) and two S,S-coordinated 1,2-
dithiosemiquinonate(1-) π radical anions. The two spins are
intramolecularly, strongly antiferromagnetically coupled via
a superexchange mechanism mediated through the diamag-
netic metal(II) ion (nd8, SM ) 0). Thus, these species possess
32%, 50%, and 30% singlet diradical character, respectively,
as has been deduced from spectroscopy oriented configura-
tion interaction ab initio methods with inclusion of scalar
relativistic corrections. At the same level of approximation,
the diradical character9 in the series of square planar neutral
bis(dioxolene)nickel(II) complexes decreases on going from
two O,O-coordinated semiquinonate(1-) ligands (∼78%),
to two N,N-coordinated o-diiminebenzosemiquinonate(1-)
ligands (∼48%), to two S,S-coordinated benzene-1,2-dithi-
olate(1-) ligands (32%). It should be noted that the values
of the diradical character apparently depend rather sensitively
on the details of the theoretical method as we have found
significantly larger values (∼75-80%) for the [Ni(LNISQ)2]0
system previously using somewhat different theoretical
methods.9a However, the diradical character is not a physical
observable and its significance is consequently in that it
provides insight into the chemical trends along a series. In
addition it provides a feeling for the open-shell character of
1
1
(S ) /2) and [AuII(L)2]2- (S ) /2).
Acknowledgment. We are grateful to the Fonds der
Chemischen Industrie for financial support. The development
of ab initio methods for open-shell systems was financially
supported within the DFG priority program 1137 (“Molecular
Magnetism”), which is gratefully acknowledged. K.R. thanks
the Max-Planck Society for a stipend.
Supporting Information Available: Figure S1, showing the
Gaussian deconvolution of the experimental absorption spectra of
the [M(LBu•)2] species with M ) Ni, Pd, and Pt between 5000 and
20 000 cm-1, Tables S1 and S2, containing the results of the
Gaussian deconvolution of the absorption spectra of the [M(LBu•)-
(LBu)]z and [M(LBu•)2] complexes, respectively, in CH2Cl2 solutions,
and CIF files, containing the complete listings of the crystal-
lographic details, atom coordinates, bond lengths and angles,
thermal displacement parameters, and calculated positional param-
eters for complexes 2a, 2b*, and 5′. This material is available free
of charge via the Internet at http://pubs.acs.org.
IC0507565
5360 Inorganic Chemistry, Vol. 44, No. 15, 2005