Square planar vs tetrahedral coordination in diamagnetic complexes of nickel(II) containing two bidentate π-radical monoanions

Article abstract of DOI:10.1021/ic040117e

The reaction of three different 1-phenyl and 1,4-diphenyl substituted S-methylisothiosemicarbazides, H₂[L^1-6], with Ni(OAc) ₂·4H₂O in ethanol in the presence of air yields six four-coordinate species [Ni(L^1-6.)₂] (1-6) where (L ^1-6.)^1- represent the monoanionic π-radical forms. The crystal structures of the nickel complexes with 1-phenyl derivatives as in 1 reveal a square planar structure trans-[Ni(L^1-3.)2], whereas the corresponding 1,4-diphenyl derivatives are distorted tetrahedral as is demonstrated by X-ray crystallography of [Ni(L^5.)₂] (5) and [Ni(L^6.)₂] (6). Both series of mononuclear complexes possess a diamagnetic ground state. The electronic structures of both series have been elucidated experimentally (electronic spectra magnetization data). The square planar complexes 1-3 consist of a diamagnetic central Ni(II) ion and two strongly antiferromagnetically coupled ligand π-radicals as has been deduced from correlated ab initio calculations; they are singlet diradicals. The tetrahedral complexes 4-6 consist of a paramagnetic high-spin Ni^II ion (S_Ni = 1), which is strongly antiferromagnetically coupled to two ligand π-radicals. This is clearly revealed by DFT and correlated ab initio calculations. Electrochemically, complexes 1-6 can be reduced to form stable, paramagnetic monoanions [1-6]^- (S = 1/2). The anions [1-3] ^- are square planar Ni(II) (d,⁸ S_Ni = 0) species where the excess electron is delocalized over both ligands (class III, ligand mixed valency). In contrast, one-electron reduction of 4, 5, and 6 yields paramagnetic tetrahedral monoanions (S = 1/2). X-band EPR spectroscopy shows that there are two different isomers A and B of each monoanion present in solution. In these anions, the excess electron is localized on one ligand [Ni^II(L^4-6.)(L^4-6)]^- where (L ^4-6)^2- is the closed shell dianion of the ligands H ₂[L^4-6] as was deduced from their electronic spectra and broken symmetry DFT calculations. Oxidation of 1 and 5 with excess iodine yields octahedral complexes [Ni^II(L^1,ox)₂I ₂] (7), [Ni^II(L^1,ox)₃](I ₃)₂ (8), and trans-[Ni^II(L^5,ox) ₂(I₃)₂] (9), which have been characterized by X-ray crystallography; (L^1-6,ox) represent the neutral, two-electron oxidized forms of the corresponding dianions (L^1-6)^2-. The room-temperature structures of complexes 1, 5, and 7 have been described previously in refs 1-5.

Full text of DOI:10.1021/ic040117e

Inorg. Chem. 2005, 44, 3636−3656
Square Planar vs Tetrahedral Coordination in Diamagnetic Complexes
of Nickel(II) Containing Two Bidentate -Radical Monoanions
π
Sebastien Blanchard, Frank Neese, Eberhard Bothe, Eckhard Bill, Thomas Weyhermu1ller, and
Karl Wieghardt*
Max-Planck-Institut fu¨r Bioanorganische Chemie, Stiftstrasse 34-36,
D-45470 Mu¨lheim an der Ruhr, Germany
Received October 18, 2004
-
The reaction of three different 1-phenyl and 1,4-diphenyl substituted S-methylisothiosemicarbazides, H2[L1 6], with
-
-6• 1-
Ni(OAc)2
‚
4H2O in ethanol in the presence of air yields six four-coordinate species [Ni(L1 6•)2] (1
−
6) where (L1
)
represent the monoanionic
π
-radical forms. The crystal structures of the nickel complexes with 1-phenyl derivatives
-
as in 1 reveal a square planar structure trans-[Ni(L1 3•)2], whereas the corresponding 1,4-diphenyl derivatives are
distorted tetrahedral as is demonstrated by X-ray crystallography of [Ni(L5 )2] (5) and [Ni(L6 )2] (6). Both series of
•
•
mononuclear complexes possess a diamagnetic ground state. The electronic structures of both series have been
elucidated experimentally (electronic spectra magnetization data). The square planar complexes 1
diamagnetic central Ni(II) ion and two strongly antiferromagnetically coupled ligand -radicals as has been deduced
from correlated ab initio calculations; they are singlet diradicals. The tetrahedral complexes 4 6 consist of a
paramagnetic high-spin NiII ion (SNi
1), which is strongly antiferromagnetically coupled to two ligand -radicals.
This is clearly revealed by DFT and correlated ab initio calculations. Electrochemically, complexes 1 6 can be
reduced to form stable, paramagnetic monoanions [1
6]- (S 1/2). The anions [1 3]- are square planar Ni(II) (d,8
SNi 0) species where the excess electron is delocalized over both ligands (class III, ligand mixed valency). In
contrast, one-electron reduction of 4, 5, and 6 yields paramagnetic tetrahedral monoanions (S
1/2). X-band EPR
−3 consist of a
π
−
)
π
−
−
)
−
)
)
spectroscopy shows that there are two different isomers A and B of each monoanion present in solution. In these
-
-
-6 2-
anions, the excess electron is localized on one ligand [NiII(L4 6•)(L4 6)]- where (L4
)
is the closed shell dianion
of the ligands H2[L4 6] as was deduced from their electronic spectra and broken symmetry DFT calculations. Oxidation
-
of 1 and 5 with excess iodine yields octahedral complexes [NiII(L1,ox)2I2] (7), [NiII(L1,ox)3](I3)2 (8), and trans-[NiII-
(L5,ox)2(I3)2] (9), which have been characterized by X-ray crystallography; (L1 6,ox) represent the neutral, two-electron
-
oxidized forms of the corresponding dianions (L1
)
-. The room-temperature structures of complexes 1, 5, and 7
-6 2
have been described previously in refs 1 5.
−
Introduction
The doubly deprotonated forms of the neutral molecules,
(Lx)2-, could, in principle, act as bidentate dianionic ligands
in transition metal ion complexes generating five-membered
chelate rings.
Revenco and Gerbeleu et al.1-5 have reported in a series
of elegant papers on the coordination chemistry of S-methyl-
1-phenyl-isothiosemicarbazides and of the sterically more
demanding S-methyl-1,4-diphenyl-isothiosemicarbazides with
nickel(II) and cobalt(III) ions. Scheme 1 displays the
structures of these ligands and gives their abbreviations.
* To whom correspondence should be addressed. E-mail: wieghardt@
mpi-muelheim.mpg.de.
(1) (a) Simonov, Y. A.; Revenco, M. D.; Bourosh, P. N.; Gerbeleu, N.
V.; Vyrtosu, N. I.; Indrichan, K. M.; Chumakov, Y. M.; Bel’skii, V.
K. Dokl. Akad. Nauk SSSR 1988, 298, 378-82. (b) Revenco, M. D.;
Simonov, Y. A.; Vyrtosu, N. I.; Gerbeleu, N. V.; Bourosh, P. N.;
Bel’skii, V. K.; Indrichan, K. M. Zh. Neorg. Khim. 1988, 33, 2049.
However, such species have not been observed to date,
but their oxidized monoanions which are paramagnetic
3636 Inorganic Chemistry, Vol. 44, No. 10, 2005
10.1021/ic040117e CCC: $30.25
© 2005 American Chemical Society
Published on Web 04/16/2005

Diamagnetic Complexes of Nickel(II)
Scheme 1. Ligands and Abbreviations
Scheme 2. Syntheses of 1 and 2 According to Ref 2
π-radicals (Srad ) 1/2), namely, (Lx)1-•, have been identified
in diamagnetic, square planar [NiII(L1•)2] (1).1 These π-radical
anions have also been identified by X-ray crystallography
in [CoIII(L1•)2I]5 and in [FeIII(L1•)2X] (X ) Cl, CH3S).6
Interestingly, square planar 1 can be oxidized by iodine
yielding paramagnetic, octahedral trans-[NiII(L1,ox)2I2] (7) (S
) 1) the crystal structure of which has been determined2 at
ambient temperature (Scheme 2). We have discovered here
that oxidation of 1 can also afford paramagnetic [NiII(L1,ox)3]-
(I3)2 (8), which we have characterized structurally.
Scheme 3. Complexes of This Worka
In all of these complexes, the two C-N and the N-N
bond lengths of each five-membered chelate ring are
characteristic for the oxidation level of the ligand as either
(Lx•)1- or (Lx,ox)0. For example, in complexes with (Lx•)1-
π-radicals the N-N bond is rather short at 1.34 ( 0.01 Å
but significantly shorter at 1.24 ( 0.01 Å in complexes
containing the (Lx,ox) form. Thus, high-quality X-ray crystal-
lography is an efficient tool to assign the given ligand
oxidation level. Since the structures of 1 and 7 (Scheme 3)
have been determined at room temperature only, we decided
to redetermine their structures by using new data sets
collected at 100(2) K. The estimated standard deviations are
significantly smaller.
a (a) Ground state; (b) s. pl. ) square planar geometry at NiII; (c) tetr.
) tetrahedral geometry at NiII; (d) oct. ) octahedral geometry at NiII.
Most importantly, Revenco and Gerbeleu et al.3,4 reported
in two papers the synthesis and crystal structures of two such
nickel(II) complexes containing each two π-radical monoan-
ions of the ligands H2(L5) and phenylazothiomethyl-(2-
methoxy-phenylimino)methane shown in Figure 1. Both
OCH3•
Ph
complexes, [NiII(L5•)2] (5) and [NiII(L
)2] are diamag-
Figure 1. Diamagnetic tetrahedral complexes of 5 from ref 3 and
[NiII(LPOhCH3•)2] from ref 4.
(2) Revenco, M. D.; Vyrtosu, N. I.; Gerbeleu, N. V.; Simonov, Y. A.;
Bourosh, P. N.; Sobolev, A. N. Zh. Neorg. Khim. 1988, 33, 2353.
(3) Simonov, Y. A.; Battaglia, L. P.; Corradi, A. B.; Pelosi, G.; Revenco,
M. D.; Gerbeleu, N. V. Acta Crystallogr. 1991, C47, 1826.
(4) Bourosh, P. N.; Revenco, M. D.; Timco, L. A.; Simonov, Y. A. Zh.
Neorg. Khim. 1999, 44, 1258.
(5) Revenco, M. D.; Gerbeleu, N. V.; Simonov, Y. A.; Bourosh, P. N.;
Vyrtosu, N. I.; Sobolev, A. N.; Malinovskii, T. I. Dokl. Akad. Nauk
SSSR 1988, 300, 127.
netic, but they possess a distorted tetrahedral N4-donor set.
Thus, the spin coupling yielding the S ) 0 ground state must
be different from that in the corresponding square planar
complexes. The authors imply that this structural difference
is a consequence of the presence of a second N-phenyl group
in these two ligands, which is suggested to lead to an
(6) Blanchard, S.; Bill, E.; Weyhermu¨ller, T.; Wieghardt, K. Inorg. Chem.
2003, 43, 2324.
Inorganic Chemistry, Vol. 44, No. 10, 2005 3637

Blanchard et al.
Scheme 4. Ligand Oxidation Products
sodium nitrite, phenyl isothiocyanate, and Ni(OAc)2‚4H2O were
purchased from Aldrich. All chemicals were used without further
purification. The synthesis of S-methyl-1-phenyl-isothiosemicar-
bazide hydroiodide H2[L1]‚HI has been described elsewhere.6
p-Nitrophenylhydrazine hydrochloride was prepared according to
published literature.10
S-Methyl-1-p-nitrophenyl-isothiosemicarbazide Hydroiodide,
H2[L2]‚HI. Starting from p-nitrophenylhydrazine hydrochloride and
ammonium thiocyanate, 1-p-nitrophenyl-isothiosemicarbazide was
obtained in 82% yield according to a published procedure.11 EI-
MS: m/z ) 212 for {C7H8N4O2S}+. 1H NMR (400 MHz, DMSO-
d6): δ(ppm) ) 6.71 (d, J ) 9.6 Hz, 2H, o-HArNN), 7.67 (br s,
1H, C(S)NH2), 7.94 (br s, 1H, C(S)NH2), 8.10 (d, J ) 9.6 Hz, 2H,
m-HArNN), 9.06 (s, 1H, ArNH), 9.54, (s, 1H, ArNNH); 13C NMR
(400 MHz, DMSO-d6): δ(ppm) ) 111.2 (CH, o-ArNN), 125.8 (CH,
p-ArNN), 138.8 (Cq, p-ArNN), 154.0 (Cq, ArNN), 182.5 (Cq, Cd
S).
increased steric bulk as compared to those complexes with
two ligands having each only a single N-phenyl group
(H2[L1-3]).
1-p-Nitrophenyl-isothiosemicarbazide (5.24 g, 24.6 mmol) was
suspended in absolute ethanol (150 mL) and methyl iodide (1.53
mL, 3.49 g, 24.6 mmol) was added. After heating of the dark red
solution for 2 h, the solvent was removed by evaporation and the
orange powder was suspended in dichloromethane (20 mL). The
precipitate was filtered off and washed with dichloromethane (2.5
mL). Yield: 7.36 g (85%). EI-MS: m/z ) 226 for {C8H11N4O2S}+.
As was observed in the case of S-methyl-1-phenyl-isothiosemicar-
To the best of our knowledge, there are only two other
tetrahedral complexes with an S ) 0 ground state containing
a central nickel(II) ion and two π-radical monoanions known.
Pierpont et al.7 reported the structure of tetrahedral, diamag-
netic [NiII(PhenoxSQ)2] (PhenoxBQ ) 2,4,6,8-tetra-tert-
butylphenoxazin-1-one and (PhenoxSQ)1-• is its one-electron
reduced semiquinone form), and Hazell8 reported the tetra-
hedral structure of diamagnetic bis(N,N-diethyl-phenyla-
zothioformamide)nickel(II), [Ni{(C6H5)NNCSN(C2H5)2}2],
where the corresponding Zn(II) complex possesses an S )
1 ground state9 verifying the π-radical anion character of
the coordinated ligands.
Our present results indicate clearly that the oxidation of
diamagnetic, tetrahedral [Ni(L5•)2] (5) with excess iodine
yields paramagnetic, octahedral trans-[NiII(L5,ox)2(I3)2] (9),
where the two ligands (L5,ox)0 are coordinated in trans-
position relative to each other with a planar NiN4 moiety.
Thus, in this case no or only weak steric repulsion between
the ligands appears to exist.
We have synthesized here a series of new S-methyl-1-
phenyl-isothiosemicarbazones H2[L2], H2[L3] and also their
1,4-diphenyl derivatives H2[L4-6] (Scheme 1) and prepared
the diamagnetic nickel(II) complexes 1-6 shown in Scheme
3. From spectroscopic characterizations and X-ray structure
determinations, it is shown that complexes 1, 2, and 3 are
square planar molecules, whereas in 4, 5, and 6 the nickel-
(II) ion is in a tetrahedral environment. Here we also present
density functional theoretical (DFT) and correlated ab initio
calculations on these neutral square planar and tetrahedral
complexes and their monoreduced forms.
1
bazide hydroiodide,6 the H NMR (400 MHz, DMSO-d6, 300 K)
spectrum corresponds to that of two tautomers in equilibrium, in a
85/15 ratio (M ) major isomer, m ) minor isomer): δ(ppm) )
2.48 (s, 0.45 H, SMe(m)), 2.72 (s, 2.55 H, SMe(M), 6.92 (d, 2H,
J ) 5.2 Hz, o-HArNN), 8.17 (d, 2H, J ) 5.2 Hz, m-HArNN), 9.48
(s, 0.85 H, ArNHN(M)), 9.57 (br s, NH), 9.60 (s, 0.15 H, ArNHN-
(m)), 11.39 (br s, NH). 13C NMR (400 MHz, DMSO-d6, 300 K):
δ(ppm) ) 12.7 (CH3, SMe(m)), 13.3 (CH3, SMe(M)), 112 (CH,
o-ArNN), 125.7 (CH, p-ArNN), 140.0 (Cq, p-ArNN(M)), 140.3
(Cq, p-ArNN(m)), 152.0 (Cq, ArNN(M)), 152.2 (Cq, ArNN(m)),
170 (Cq, CdS).
S-Methyl-1-p-methoxyphenyl-isothiosemicarbazide Hydroio-
dide, H2[L3]‚HI. Using the same procedure as for 1-p-nitrophe-
nylisothiosemicarbazide, we obtained a ca. 80:20 mixture of the
tautomers in the 1- and 2-positions of p-methoxyphenyl-isothi-
osemicarbazide. These two products could not easily be separated;
therefore, we performed the methylation step on the mixture.
p-Methoxyphenyl-isothiosemicarbazide (5.6 g, 27.7 mmol) was
suspended in absolute ethanol (150 mL), and then methyliodide
was added (1.73 mL, 3.95 g, 27.7 mmol) and the mixture was heated
to reflux for 2 h. After cooling of the sample to room temperature,
a white powder was filtered off. Recrystallization from dichlo-
romethane at -30 °C yielded 3.55 g (37%) of pure pale yellow
H2[L3]‚HI. ESI-MS (MeOH; pos. ion): m/z ) 212 for {C9H14N3-
OS}+. Coalescence processes in variable temperature 1H NMR (400
MHz, DMSO-d6, 300-340 K) experiments prove that two tau-
tomers are in equilibrium in solution, with a 80:20 ratio at 300 K;
1H NMR (400 MHz, DMSO-d6, 300 K, M ) major isomer only,
m ) minor isomer only): δ(ppm) ) 2.49 (s, 0.6 H, SMe(m)), 2.68
(s, 2.4 H, SMe(M)), 3.65 (s, 0.6 H, OMe(m)), 3.68 (s, 2.4 H, OMe-
(M)), 6.76 (m, 2H, o-HArNN), 6.86 (m, 2H, m-HArNN), 8.15 (s,
0.8 H, ArNHN(M)), 8.27 (s, 0.2 H, ArNHN(m)), 9.43 (s, 0.8 H,
CSNH2(M)), 9.55 (s, 0.8 H, CSNH2(M)), 11.14 (s, 0.8 H, ArNNH-
(M)). 13C NMR (400 MHz, DMSO-d6, 300 K): δ(ppm) ) 12.5
(CH3, SMe(m)), 13.2 (CH3, SMe(M)), 55.34 (CH3, OMe), 114.5
Finally, we have isolated and characterized two different
ligand oxidation products of the ligands H2[L5] and H2[L6]
shown in Scheme 4.
Experimental Section
p-Methoxyphenylhydrazine hydrochloride, phenylhydrazine hy-
drochloride, ammonium thiocyanate, methyl iodide, p-nitroaniline,
(7) Whalen, A. K.; Bhattacharya, S.; Pierpont, C. G. Inorg. Chem. 1994,
33, 347.
(8) Hazell, R. G. Acta Chem. Scand. 1976, A30, 322.
(9) Jensen, K. A.; Bechgaard, K.; Pedersen, C. T. Acta Chem. Scand. 1972,
26, 2913.
(10) Pilgram, K. H. Synth. Comm. 1985, 15, 697.
(11) Pyl, T.; Scheel, K. H.; Beyer, H. J. Prakt. Chem. 1963, 4(20), 255.
3638 Inorganic Chemistry, Vol. 44, No. 10, 2005

Diamagnetic Complexes of Nickel(II)
(CH, m-ArNN(M)), 114.8 (CH, m-ArNN(m)), 115.1 (CH, o-ArNN),
139.5 (Cq, ArNN), 154.2 (Cq, p-ArNN), 170.3 (Cq, C(SMe)(M)),
176.5 (Cq, C(SMe)(m)).
equilibrium at 300 K in a 2:1 ratio): δ(ppm) ) 2.68 and 2.70 (s,
SMe), 3.72 and 3.76 (OMe), 6.84 (br s), 6.94 (d, J ) 8.8 Hz), 7.02
(d, J ) 8.8 Hz), 7.39 (d, J ) 7.4 Hz), 7.42-7.60 (m). 13C NMR
(400 MHz, CD3OD, 300 K): δ(ppm) ) 14.0 and 14.2 (SMe), 56.1
(OMe), 115.7, 115.9, 116.6, 117.2, 127.8, 128.4, 130.3, 130.6,
130.8, 131.6, 136.0, 140.2, 156.8, 157.0, 172.9 and 177.2 (Cq-S).
S-Methyl-1-p-nitrophenyl-4-phenyl-isothiosemicarbazide Hy-
droiodide, H2[L5]‚HI. The same procedure as for H2[L4]‚HI was
applied starting with p-nitrophenylhydrazine hydrochloride (1.89
g, 10 mmol) and phenyl isothiocyanate (1.35 g, i.e., 1.2 mL, 10
mmol). 2.61 g of 1-p-nitrophenyl-4-phenyl-isothiosemicarbazide
was isolated as a pale yellow powder (Yield 90%). EI-MS: m/z )
288 for {C13H12N4O2S}+. 1H NMR (500 MHz, DMSO-d6, 300 K):
δ(ppm) ) 6.82 (d, J ) 9 Hz, o-ArNN), 7.14 (t, J ) 7.3 Hz, 1H,
p-ArNC), 7.30 (t, J ) 7.3 Hz, 2H, m-ArNC), 7.45 (d, J ) 7.3 Hz,
2H, o-ArNC), 8.14 (d, J ) 9 Hz, 2H, m-ArNN), 9.19 (s, 1H, NH),
9.90 (br s, 1H, NH), 9.95 (br, s, 1H, NH). 13C NMR (500 MHz,
DMSO-d6, 300 K): δ(ppm) ) 111.6 (CH, m-ArNN), 125.1 (CH,
p-ArNC), 125.7 (CH, o-ArNC), 125.8 (CH, m-ArNN), 128.0 (CH,
m-ArNC), 139.1 (Cq, ArNC and NO2Ar), 154.1 (Cq, ArNN), 181.5
(Cq, CS). The methylation was performed on 2.5 g of 1-p-
nitrophenyl-4-phenyl-isothiosemicarbazide (8.7 mmol) with 1.36
g methyl iodide (0.59 mL, 9.5 mmol). Evaporation of the ethanol
solution yielded 3 g of dark red oily S-methyl-1-p-nitrophenyl-4-
phenyl-isothiosemicarbazide, which was used without further
S-Methyl-1-phenyl-4-phenyl-isothiosemicarbazide Hydroio-
dide, H2[L4]‚HI. The starting material 1-phenyl-4-phenyl-isothi-
osemicarbazide was synthesized by using a slight modification of
a published procedure.12 Phenylhydrazine hydrochloride (7.2 g, 50
mmol) was suspended in 20 mL of absolute ethanol. Phenyl-
isothiocyanate (6.8 g, i.e., 6.0 mL, 50 mmol) and pyridine (5 mL)
were added. The slurry was heated to reflux for 2 h. After cooling
of the mixture to 20 °C, distilled water (5 mL) was added. The
white precipitate was filtered off, washed twice with ethanol (5
mL), and dried under vacuum to yield 8.73 g of 1-phenyl-4-phenyl-
isothiosemicarbazide (72%). EI-MS: m/z ) 243 for {C13H13N3S}+.
1H NMR (500 MHz, DMSO-d6, 300 K): δ(ppm) ) 6.79 (d, J )
8 Hz, 2H, o-ArNN), 6.83 (d, J ) 7.5 Hz, 1 H, p-ArNN), 7.12 (t,
J ) 7.5 H, 1H, p-ArNC), 7.23 (t, J ) 8 Hz, 2H, m-ArNN), 7.29 (t,
J ) 7.5 Hz, 2H, m-ArNC), 7.55 (d, J ) 8 Hz, 2H, o-ArNC), 8.08
(br, s, 1H, NH), 9.70 (s, 1H, NH), 9.82 (br, s, 1H, NH). 13C NMR
(500 MHz, DMSO-d6, 300 K): δ(ppm) ) 113.0 (CH, o-ArNN),
119.8 (CH, p-ArNN), 124.7 (CH, p-ArNC), 125.0 (CH, o-ArNC),
127.9 (CH, m-ArNC), 128.8 (CH, m-ArNN), 139.2 (Cq, ArNC),
148.0 (Cq, ArNN), 181.1 (Cq-CS). To 1-phenyl-4-phenyl-isothi-
osemicarbazide (3 g, 12.3 mmol) suspended in 50 mL absolute
ethanol was added methyl iodide (1.75 g, 0.77 mL, 12.3 mmol).
The mixture was heated to reflux for 2 h yielding a pale yellow
solution, which was cooled to 20 °C. The solvent ethanol was
removed by evaporation. Yield: 4.87 g of yellow oily S-methyl-
1-phenyl-4-phenyl-isothiosemicarbazide hydroiodide, which was
used without further purification. EI-MS: m/z ) 257 for
1
purification. EI-MS: m/z ) 302 for {C14H14N4O2S}+. H NMR
(250 MHz, DMSO-d6, 300 K): δ(ppm) ) 2.67 (s, 3H, SMe), 6.85-
7.00 (m, 3H), 7.24-7.34 (m, 3H), 7.40-7.51 (m, 4H), 8.55 (s, br,
NH). 13C NMR (400 MHz, DMSO-d6, 300 K): δ(ppm) ) 13.92
(CH3), 111.4, 111.5, 125.4, 125.6, 127.0, 127.8, 128.9, 138.4, 151.4.
[NiII(L1•)2] (1).1 This complex was synthesized using a slight
modification of the literature procedure.1 To a pale yellow solution
of S-methyl-1-phenylisothiosemicarbazide hydroiodide (2 g, 6.36
mmol) in ethanol (40 mL) was added Ni(OAc)2‚4H2O (793 mg,
3.18 mmol) in the presence of air. Upon addition of Et3N (2.65
mL, 19.1 mmol), a darkening of the solution occurred. After 2 h
of stirring at room temperature, a dark green precipitate was filtered
off and washed with ethanol to yield 880 mg of 1 (66% yield).
X-ray quality crystals were grown by slow evaporation of the
solvent from a concentrated solution of 1 in ethyl acetate. 1H NMR
(400 MHz, CDCl3, 300 K): δ(ppm) ) 2.45 (s, 6H, SCH3), 5.86
(br s, 2H, NH), 7.43 (m, 6H, m-ArH and p-ArH), 7.97 (d of m,
1
{C14H15N3S}+. By analogy with the former cases, the H NMR
(400 MHz, CD3OD, 300 K) spectrum shows the presence of two
tautomers in equilibrium at 300 K in a 7:3 ratio; δ(ppm) ) 2.74
and 2.75 (s, SMe), 6.94 (d, J ) 8.4 Hz and t), 7.05 (t, J ) 7.4 Hz),
7.13 (d, J ) 7.8 Hz), 7.30 (t, J ) 7.8 Hz), 7.37-7.57 (m), 7.59-
7.65 (m). 13C NMR (400 MHz, CD3OD, 300 K): δ(ppm ) 14.4
and 14.7 (SMe), 114.7, 115.2, 122.8, 123.1, 127.7, 128.2, 130.18,
130.23, 130.3, 130.5, 130.7, 131.5, 135.5, 135.7, 146.6, 172.9 and
177.1 (Cq-S).
S-Methyl-1-p-methoxyphenyl-4-phenyl-isothiosemicarba-
zide Hydroiodide, H2[L6]‚HI. The same procedure as for H2[L4]‚
HI was applied starting with p-methoxyphenylhydrazine hydro-
chloride (5 g, 28.6 mmol) and phenyl isothiocyanate (3.87 g, i.e.,
3.4 mL, 28.6 mmol). A total of 6.28 g of 1-p-methoxyphenyl-4-
phenyl-isothiosemicarbazide was isolated as a white powder (Yield
80%). EI-MS: m/z ) 273 for {C14H15N3OS}+. 1H NMR (500 MHz,
DMSO-d6, 300 K): δ(ppm) ) 3.67 (s, 3H, MeO), 6.73 (d, J ) 8.5
Hz, 2H, o-ArNN), 6.84 (d, J ) 8.5 Hz, 2H, m-ArNN), 7.11 (t, J )
7.5 Hz, 1H, p-ArNC Hz, 2H, o-ArNC), 7.77 (br s, 1H, NH), 9.65
(s, 1H, NH), 9.79 (s, 1H, NH). 13C NMR (500 MHz, DMSO-d6,
300 K): δ(ppm) ) 55.3 (CH3, OMe), 114.3 (CH, m-ArNN), 114.5
(CH, o-ArNN), 124.5 (CH, p-ArNC), 124.8 (CH, o-ArNC), 127.8
(CH, m-ArNC), 139.1 (Cq, ArNC), 141.5 (CQ, ArNN), 153.5 (Cq,
MeOAr), 180.8 (Cq, CS). The methylation was performed on 3 g
of 1-p-methoxyphenyl-4-phenyl-isothiosemicarbazide (11.0 mmol)
with 1.55 g methyl iodide (0.68 mL, 11.0 mmol). Evaporation of
the ethanol solvent and precipitation in dichloromethane at -30
°C yielded 4.29 g of S-methyl-1-p-methoxyphenyl-4-phenyl-isothi-
osemicarbazide hydroiodide as a white powder (Yield: 95%). EI-
Ms: m/z ) 287 for {C15H17N3OS}+. 1H NMR (400 MHz, CD3OD,
300 K) (By analogy with the former cases, two conformers in
3
4H, J ) 6.8 Hz, o-ArH). 13C NMR (400 MHz, CDCl3, 300 K):
δ(ppm) ) 15.7 (CH3, SMe), 125.0 (CH, o-Ar), 127.0 (CH, p-Ar),
129.3 (CH, m-Ar), 156.2 (Cq, Ar), 164.5 (Cq, CSMe).
[NiII(L2•)2] (2). The same procedure as described for 1 was used
by starting with S-methyl-1-p-nitrophenyl-isothiosemicarbazide
hydroiodide, H2[L2]‚HI, (708 mg, 2 mmol) to yield 470 mg (92%)
of dark red crystalline 2. EI-MS: m/z ) 506 for {C16H16N8NiS2}+.
Anal. Calcd for C16H16N8NiS2: C, 37.89; H, 3.18; N, 22.09; Ni,
1
11.57. Found: C, 38.04; H, 3.24; N, 21.91; Ni, 11.53. H NMR
(400 MHz, CDCl3, 300 K): δ(ppm) ) 2.51 (s, 6H, SCH3), 5.99
(br s, 2H, NH), 8.16 (d, 4H, J ) 9.2 Hz, HAr), 8.29 (d, 4H, J )
9.2 Hz, HAr).
[NiII(L3•)2] (3). The same procedure as described for 1 was used
by starting with S-methyl-1-p-methoxyphenyl-isothiosemicarbazide
hydroiodide, H2[L3]‚HI, (820 mg, 2.5 mmol) to yield 490 mg (82%)
of dark red microcrystalline 3. EI-MS: m/z ) 476 for {C18H22N6-
NiO2S2}+. Anal. Calcd for C18H22N6NiO2S2: C, 45.30; H, 4.65;
N, 17.61; Ni, 12.30. Found: C, 45.38; H, 4.63; N, 17.59; Ni, 12.32.
1H NMR (400 MHz, CDCl3, 300 K): δ(ppm) ) 2.44 (s, 6H, SCH3),
3.87 (s, 6H, OCH3), 5.85 (br s, 2H, NH), 6.95 (d, 4H, J ) 9.0 Hz,
m-HArNN), 8.03 (d, 4H, J ) 9.0 Hz, o-HArNN).
(12) Wu, M. T.; Waksmunski, F. S.; Hof, D. R.; Fisher, M. H.; Egerton,
J. R.; Patchett, A. A. J. Pharm. Sci. 1977, 66, 1150.
Inorganic Chemistry, Vol. 44, No. 10, 2005 3639

Blanchard et al.
[NiII(L4•)2] (4).1b The same procedure as for 1 was used by
starting with S-methyl-1-phenyl-4-phenyl-isothiosemicarbazide hy-
droiodide, H2[L4]‚HI, (1.5 g, 3.9 mmol) to yield 880 mg (79%) of
dark red crystalline 4. EI-MS: m/z ) 568 for {C28H26N6NiS2}+.
Anal. Calcd for C28H26N6NiS2: C, 59.07; H, 4.60; N, 14.76; Ni,
trans-[NiII(L5,ox)2(I3)2]‚2CHCl3 (9‚2CHCl3). To a purple solution
of 5 (132 mg; 0.2 mmol) in CHCl3 (5 mL) was added at 40 °C a
solution of I2 (155 mg; 0.6 mmol) in CHCl3 (30 mL) with stirring.
The color of the solution turned brownish. After 20 min the solution
was allowed to cool to 20 °C and the CHCl3 was allowed to slowly
evaporate. X-ray quality of brown-black crystals of 9‚2CHCl3 were
1
10.31. Found: C, 58.88; H, 4.76; N, 14.67; Ni, 10.32. H NMR
(500 MHz, CDCl3, 300 K): δ(ppm) ) 2.30 (s, 6H, SCH3), 7.03 (t,
4H, J ) 7.8 Hz, m-HArNN), 7.22 (d, 2H, J ) 7.4 Hz, p-HArNC),
7.29 (t, 4H, J ) 7.6 Hz, m-HArNC), 7.50 (d, 4H, J ) 7.1 Hz,
o-HArNC), 7.95 (t, 2H, J ) 7.4 Hz, p-HArNN), 9.52 (d, 4H, J )
8.1 Hz, o-HArNN). 13C NMR (400 MHz, CDCl3, 300 K): δ(ppm)
) 15.3 (CH3, SMe), 120.0 (CH, o-ArNN), 122.8 (CH, p-ArNN),
123.8 (CH, o-ArNC), 125.9 (CH, p-ArNC), 128.7 (CH, m-ArNC),
133.8 (CH, m-ArNN), 151.9 (Cq, ARNC), 157.2 (Cq, CSMe), 167.3
(Cq, ArNN).
obtained (280 mg, 80% yield). Anal. Calcd for C30.3H26.3Cl6.9I6.4N8-
NiO4S2 (9‚(2.3CHCl3)‚(0.2I2)):C, 20.84; H, 1.52; N, 6.42; Ni, 3.36;
I, 46.51. Found: C, 20.66; H, 1.45; N, 6.59; Ni, 3.43; I, 46.58.
(L6,ox) (10). Solid 6 (315 mg, 0.5 mmol) was dissolved in
chloroform (10 mL) and heated to 40 °C. Then a solution of I2
(127 mg, 0.5 mmol) in chloroform (20 mL) was added dropwise.
The initially red purple solution rapidly turned brownish. It was
taken to dryness, redissolved in dichloromethane (5 mL), and
pentane (20 mL) was added. A brown precipitate was filtered off,
and the filtrate was left for slow evaporation to yield 80 mg (28%)
of X-ray quality reddish crystals of (L6,ox) (10). EI-MS: m/z )
[NiII(L5•)2] (5).3 The same procedure as for 1 was used by starting
with S-methyl-1-p-nitrophenyl-4-phenyl-isothiosemicarbazide hy-
droiodide, H2[L5]‚HI, (1.5 g, 3.5 mmol) to yield 930 mg (80%) of
dark red crystalline 5. X-ray quality crystals were grown from a
concentrated solution of 5 in ethyl acetate by slow evaporation of
the solvent. EI-MS: m/z ) 658 for {C28H24N8NiO4S2}+. Anal.
Calcd for C28H24N8NiO4S2: C, 51.00; H, 3.67; N, 16.99; Ni, 8.90.
Found: C, 50.81; H, 3.76; N, 17.11; Ni, 9.14. 1H NMR (400 MHz,
CDCl3, 300 K): δ(ppm) ) 2.34 (s, 6H, SCH3), 7.31 (m, 6H,
m-HArNC and p-HArNC), 7.53 (m, 4H, o-HArNC), 7.91 (ABA′
pattern, 3J ) 9.2 Hz, 4J ) 2.68 Hz, m-HArNN), 9.52 (ABA′ pattern,
285 for {C15H15N3OS}+. Anal. Calcd for C15H15.4N3O1.2S ((Lox MeO)‚
Ph
0.2(H2O)): C, 62.35; H, 5.37; N, 14.54. Found: C, 62.45; H, 5.47;
N, 14.39. 1H NMR (400 MHz, CDCl3, 300 K): δ(ppm) ) 2.49 (s,
3H, SMe), 3.86 (s, 3H, OMe), 6.93 (d, 2H, J ) 8.8 Hz, m-HArNN),
7.04 (d, 2H, J ) 7.6 Hz, o-HArNC), 7.12 (t, 1H, J ) 7.6 Hz,
p-HArNC), 7.30 (t, 2H, J ) 7.6 Hz, m-HArNC), 7.75 (d, 2H, J )
8.8 Hz, o-ArNN). 13C NMR (400 MHz, CDCl3, 300 K): δ(ppm)
) 13.8 (CH3, SMe), 55.7 (CH3, OMe), 114.5 (CH, m-ArNN), 123.2
(CH, o-ArNC), 124.3 (p-ArNC), 126.6 (CH, o-ArNN), 128.2 (CH,
m-ArNC), 145.9 (Cq, ArNN), 148.1 (Cq, ArNC), 164.1 (Cq,
p-ArNN), 166.0 (Cq, CSMe).
4
3J ) 9.2 Hz, J ) 2.68 Hz, o-HArNN). 13C NMR (400 MHz,
CDCl3, 300 K): δ(ppm) ) 15.6 (CH3, SMe), 120.6 (CH, o-ArNN),
123.5 (CH, o-ArNC), 126.8 (CH, p-ArNC), 129.1 (CH, m-ArNC),
129.6 (CH, p-ArNN), 140.3 (Cq, p-ArNN), 151.3 (Cq, ArNC), 162.1
(Cq, CSMe), 170.7 (Cq, ArNN).
3-Methylsulfanyl-1-(4-nitrophenyl)-1,4-dihydrobenzo[1,2,4]-
triazine (11). This compound was isolated in an attempt to
synthesize the zinc complex [Zn(L5•)2]: To a suspension of H2-
[L5]‚HI (430 mg, 1 mmol) in ethanol (10 mL), Zn(OAc)2‚2H2O
(110 mg, 0.5 mmol) and triethylamine (140 µL, 1 mmol) were
added. Upon stirring of the sample in the air, a dark color quickly
developed. After 10 min of stirring, 2 more equivalents of
triethylamine (280 µL, 2 mmol) were added and the mixture was
stirred for an additional 2 h. Upon concentration of the solution by
slow evaporation (2 days) of the solvent ethanol a dark microc-
rystalline material formed, which was filtered off yielding 160 mg
of 11 (53% yield). Slow evaporation of a concentrated dichlo-
romethane solution afforded X-ray quality single crystals. EI-MS:
m/z ) 300 for {11}+ (C14H12N4O2S). 1H NMR (400 MHz, DMSO-
d6, 300 K): δ(ppm) ) 2.52 (s, 3H, SMe), 6.80 (d, 1H, J ) 7.6 Hz,
o-Hbenzotriazine), 7.00 (m, 2H, m-Hbenzotriazine and m′-Hbenzotriazine), 7.19
(d, 1H, J ) 7.8 Hz, o′-Hbenzotriazine), 7.46 (d, 2H, J ) 8 Hz,
o-HArNN), 8.11 (d, 2H, J ) 8 Hz, m-HArNN), 9.75 (s, 1H, NH).
13C NMR (400 MHz, CDCl3, 300 K): δ(ppm) ) 13.4 (CH3, SMe),
112.8 (CH, o-ArNN), 114.6 (CH, o-Arbenzotriazine), 116.5 (CH, o′-
Arbenzotriazine), 123.3 (CH, m′-Hbenzotriazine), 125.5 (CH, m-ArNN),
[NiII(L6•)2]‚0.5H2O (6). The same procedure as described for 1
was used by starting with S-methyl-1-p-methoxyphenyl-4-phenyl-
isothiosemicarbazide hydroiodide, H2[L6]‚HI, (1.5 g, 3.6 mmol) to
yield 880 mg (79%) of dark red crystalline 6. X-ray quality crystals
were grown by slow evaporation of the solvent from a concentrated
solution of 6 in ethyl acetate. EI-MS: m/z ) 628 for {C30H30N6-
NiO2S2}+ {[Ni(LPhMeO*)2]+}. Anal. Calcd for C30H31N6NiO2.5S2:
C, 56.44; H, 4.89; N, 13.16; Ni, 9.19. Found: C, 56.42; H, 4.76;
1
N, 13.13; Ni, 9.28. H NMR (400 MHz, CDCl3, 300 K): δ(ppm)
) 2.26 (s, 6H, SCH3), 6.61 (d, 4H, J ) 8.2 Hz, m-HArNN), 7.22
(t, 2H, J ) 7.1 Hz, p-HArNC), 7.29 (t, 4H, J ) 7.6 Hz, m-HArNC),
7.48 (d, 4H, J ) 7.4 Hz, o-HArNC), 9.57 (d, 4H, J ) 8.8 Hz,
o-HArNN). 13C NMR (400 MHz, CDCl3, 300 K): δ(ppm) ) 15.2
(CH3, SMe), 55.5 (CH3, OMe), 119.3 (CH, m-ArNN), 121.3 (CH,
o-ArNN), 124.1 (CH, o-ArNC), 125.6 (CH, p-ArNC), 128.6 (CH,
m-ArNC), 151.9 (Cq, ArNC), 154.4 (Cq, ArNN), 155.4 (Cq, CSMe),
162.4 (Cq, ArNN).
[NiII(L1,ox)2I2]‚2CHCl3 (7‚2CHCl3).2 To a green solution of 1
(230 mg, 0.55 mmol) in chloroform (6 mL) at 40 °C was added
dropwise a solution of I2 (140 mg, 0.55 mmol) in chloroform (3
mL). The color of the solution changed to brownish. It was then
cooled to 25 °C. The crystalline material was filtered off. Yield:
445 mg. Anal. Calcd for C18H20Cl6I2N6NiS2: C, 23.76; H, 2.22;
N, 9.24; Ni, 6.45; I, 27.9. Found: C, 23.75; H, 2.14; N, 9.31; Ni,
6.35; I, 27.52. X-ray quality crystals of unsolvated [Ni(Lox)2I2] (7)
were obtained by slow evaporation of the solvent from a 1/3 toluene
and 2/3 dichloromethane solution.
125.7 (m-Arbenzotriazine), 127.6 (Cq, Arbenzotriazine), 134.5 (Cq, Arbenzo
-
triazine), 138.3 (Cq, ArNO2), 148 (Cq, ArNN), 156.3 (Cq, CSMe).
X-ray Crystallographic Data Collection and Refinement of
the Structures. A dark green single crystal of 1, dark red crystals
of 5, 7, 9, 10 and 11, and black crystals of 6 and 8 were coated
with perfluoropolyether and mounted in the nitrogen cold stream
of a Nonius Kappa-CCD diffractometer equipped with a Mo-target
rotating-anode X-ray source and a graphite monochromator (Mo-
KR, λ ) 0.71073 Å). Final cell constants were obtained from least-
squares fits of a subsets of several thousand strong reflections. Data
collections were performed by taking frames at 1.0° in ω (0.8° in
case of 8, 1.5° for 11). Crystal faces of 1 and 6-9 were determined,
and the corresponding intensity data were corrected for absorption
using the Gaussian-type routine embedded in XPREP.13 Data sets
[NiII(L1,ox)3](I3)2 (8). An X-ray quality single crystal of 8 was
obtained from slow evaporation of the solvent from the crude filtrate
obtained upon synthesis of solvated 7 (see above). Every attempt
to synthesize this molecule in a rational fashion has been unsuc-
cessful.
3640 Inorganic Chemistry, Vol. 44, No. 10, 2005

Diamagnetic Complexes of Nickel(II)
Table 1. Crystallographic Data for 1 and 5-11
1
5
6
7
chem formula
fw
C16H18N6NiS2
417.19
C28H24N8NiO4S2
659.38
C30H30N6NiO2S2
629.43
C16H18I2N6NiS2
670.99
space group
a, Å
b, Å
P1h, No. 2
9.9844(3)
10.0508(3)
10.2507(3)
81.478(6)
62.793(6)
89.716(6)
902.45(5)
2
P21/c, No. 14
11.4764(4)
19.1692(6)
13.0665(4)
90
91.62(1)
90
2873.4(2)
4
Pbcn, No. 60
11.8076(4)
13.7892(4)
17.6810(6)
90
90
90
2878.8(2)
4
P21/c, No. 14
8.6347(3)
14.3456(5)
9.4459(4)
90
111.467(4)
90
1088.89(7)
2
c, Å
R, deg
â, deg
γ, deg
V, Å
Z
T, K
100(2)
100(2)
100(2)
100(2)
F calcd, g cm-3
reflns collected/2Θmax
unique reflns/I > 2σ(I)
no. of params/restr
µ(Mo KR), cm-1
R1a goodness of fit b
wR2c (I > 2σ (I))
residual density, eÅ-3
1.535
1.524
1.452
2.047
27647/61.9
5661/5338
235/1
13.18
0.0231/1.057
0.0550
74991/62.0
9116/8256
390/0
8.76
0.0282/1.047
0.0655
40004/62.08
4604/3571
188/0
8.59
0.0399/1.038
0.0748
28577/61.92
3444/3210
129/0
39.32
0.0148/1.105
0.0340
+0.45/-0.37
+0.39/-0.25
+0.42/-0.43
+0.53/-0.40
8
9
10
11
chem formula
Fw
space group
a, Å
b, Å
c, Å
C24H27I6N9NiS3
1357.84
Pbca, No. 61
15.5006(6)
16.9618(6)
29.2707(10)
90
90
90
7695.8(5)
8
C30H26Cl6I6N8NiO4S2
1659.52
C15H15N3OS
285.36
C14H12N4O2S
300.34
P1h, No. 2
9.7733(6)
11.4035(6)
11.5154(8)
71.14(1)
75.65(1)
76.89(1)
1161.43(12)
1
Pn, No. 7
11.8164(6)
9.8640(4)
24.5392(12)
90
91.97(1)
90
2858.5(2)
8
P21/c, No. 14
14.9324(10)
11.9210(8)
7.4166(6)
90
95.02(1)
90
1315.2(2)
4
R, deg
â, deg
γ, deg
V, Å
Z
T, K
100(2)
100(2)
100(2)
100(2)
F calcd, g cm-3
reflns collected/2Θmax
unique reflns/I > 2σ(I)
no. of params/restr.
µ(Mo KR), cm-1
R1a/ goodness of fitb
wR2c (I > 2σ (I))
residual density, eÅ-3
2.344
2.373
1.326
1.517
41986/60.0
11158/8578
401/0
55.11
0.0488/1.016
0.1088
27801/62.04
7381/7069
290/30
48.84
0.0291/1.096
0.0666
28776/52.0
11006/8597
722/2
2.25
0.0609/1.019
0.1106
20117/60.0
3825/2907
190/0
2.57
0.0482/1.019
0.0949
+2.36/-3.35
+3.29/-2.46
+0.62/-0.34
+0.37/-0.34
2
2
2
a Observation criterion: I > 2σ(I). R1 ) ∑||Fo| - |Fc||/∑|Fo|. b GOF ) [∑[w(Fo2 - Fc )2]/(n-p)]1/2
.
c wR2 ) [∑[w(Fo2 - Fc )2]/∑[w(Fo )2]]1/2 where w
2
2
2
) 1/σ2(Fo ) + (aP)2 + bP, P ) (Fo + 2Fc )/3.
-
5, 10, and 11 were not corrected for absorption. Crystallographic
data of the compounds are listed in Table 1. The Siemens
ShelXTL13 software package was used for solution and artwork of
the structure, and ShelXL9714 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. Hydrogen atoms attached to carbon
atoms were placed at calculated positions and refined as riding
atoms with isotropic displacement parameters. Protons bound to
nitrogen atoms in 1, 7 and 8 were located from the difference map
and refined with restrained bond distances (ShelX SADI instruction)
and thermal parameters. One of the two I3- anions in 8 was found
to be disordered. Two positions in a 91:9 ratio were refined with
equal thermal parameters of corresponding I-atoms (ShelX EADP
instruction). Structure of 10 was first thought to be centrosymmetric
(space group P21/n), but disorder problems led to the conclusion
that the structure is more accurately described in space group Pn.
The “disorder problems” vanished completely refining four inde-
pendent molecules, one of them having a different conformation
of the methoxy group. The chloroform molecule in the asymmetric
unit of 9 is disordered. Two split positions were refined with
constrained C-Cl and Cl-Cl distances. A total of 30 constraints
was used, but residual electron density suggests that the disorder
of the molecule, residing in a I3 walled solvent channel, is even
more complicated, but a third split position could not be refined
yielding a stable solution.
Physical Measurements. Electronic spectra of complexes and
corresponding spectra from spectroelectrochemical measurements
were recorded on an HP 8452 A diode array spectrophotometer
(range: 200-1700 nm). Cyclic voltammograms and coulometric
experiments were performed with an EG & G potentiostat/
galvanostat. Variable field (1-7 T), variable temperature (4-300
K) magnetization data were recorded on a SQUID magnetometer
(MPMS Quantum Design). The experimental magnetic susceptibil-
ity data were corrected for underlying diamagnetism by use of
tabulated Pascal’s constants. X-band EPR spectra were recorded
on a Bruker ESP 300 spectrometer.
Calculations. All calculations reported in this paper were done
with the program package ORCA.15 As will be further discussed
in the text, the geometry optimizations were carried out at the BP86
and B3LYP levels of DFT.16 These functionals have proved in many
applications their ability to reliably predict structures of transition
(13) ShelXTL V.5, Siemens Analytical X-ray Instruments, Inc.; Madison,
WI, 1994.
(14) Sheldrick, G. M. ShelXL97; University of Go¨ttingen, 1997.
Inorganic Chemistry, Vol. 44, No. 10, 2005 3641

Blanchard et al.
metal complexes. However, in the present case they yield distinctly
different results, which made a comparison of both approaches
necessary. The all-electron Gaussian basis sets used were those
reported by the Ahlrichs group.17 Accurate triple-ê valence basis
sets with one set of polarization functions on the all atoms were
used (TZV(P)). The auxiliary basis sets used to fit the electron
density were taken from the Turbomole library (Basis sets were
obtained from the ftp server of the quantum chemistry group
at the university of Karlsruhe (Germany) under ftp://ftp.
chemie.uni-karlsruhe.de/pub/basen/) and were chosen to match the
orbital basis.
The SCF calculations were always of the spin-polarized type
and were tightly converged (10-7 Eh in energy, 10-6 Eh in the
density change and 10-6 in maximum element of the DIIS error
vector). Single point calculations with the B3LYP functional were
carried out at the optimized geometries to predict EPR and
absorption spectral parameters. In the calculation of the EPR, the
same basis sets were used as in the geometry optimization, except
for the metal basis which was the triply polarized “Core Properties”
(CP(PPP)) basis.18a Ab initio calculations of the optical spectra were
also undertaken with the ORCA program and utilized the recently
developed spectroscopy oriented configuration interaction (SORCI)
formalism.18b The calculations were started from closed-shell
B3LYP orbitals and the thresholds used were Tsel ) 10-6 Eh, Tpre
) 10-2, and Tnat ) 10-5. The basis set used in the SORCI
calculation was reasonably large and based on Ahlrichs’s triple-ú
basis augmented with 2p and 1f polarization functions on the central
nickel, two sets of d-functions on C,N and S and one set of
p-functions on H, all taken from the TurboMole library (see above).
ity measurements (see below). Complexes 1-3 exhibit a
single ν(N-H) stretching frequency at 3354 cm-1 for 1, 3368
cm-1 for 2, and 3371 cm-1 for 3. In contrast, complexes
4-6 do not display this mode.
Oxidation of square planar 1 with iodine in chloroform
yields octahedral trans-[Ni(L1,ox)2I2]‚2CHCl3 (7‚2CHCl3).2 In
one instance, using the conditions given in the experimental
section for the synthesis of 7, we obtained single crystals of
octahedral [NiII(L1,ox)3](I3)2 (8). The preparation of 8 is as
yet not rationally reproducible. Temperature-dependent
magnetic susceptibility measurements on a solid sample of
7‚2CHCl3 in an external magnetic field of 1.0 T revealed
that the molecular species possesses an S ) 1 ground state.
From the fit of the data of additional measurements at 4 and
7 T shown in Figure S1 (Supporting Information) the
following spin Hamiltonian parameters for 7 were ob-
tained: S ) 1; giso ) 2.14 ((0.02), |D| ) 4.0 ( 1 cm-1.
Note that when a 3-fold excess of iodine was used for the
oxidation, the octahedral trans-[NiII(L1,ox)2(I3)2] was isolated.2
Oxidation of tetrahedral [NiII(L5•)2] with an equivalent of
iodine resulted in the isolation in ca. 30% yield of trans-
[NiII(L5,ox)2(I3)2] (9). Using a 3-fold excess iodine, the yield
was improved to 80% of isolated 9. The strong electron
withdrawing NO2 group probably renders the (L5•)- anion
harder to oxidize, thus favoring formation of coordinated (I3)-
anions over the oxidation of another [Ni(L5•)2] molecule.
Fitting of temperature-dependent magnetic susceptibility
measurements on a solid sample of 9‚2HCCl3 (Figure S2,
Supporting Information) in external fields of 1, 4, and 7 T
revealed that the molecule possesses an S ) 1 ground state
and leads to the following spin Hamiltonian parameters: giso
) 2.08 and |D| ) 6.0 ( 0.5 cm-1.
Results
Synthesis of Ligands and Complexes. To evaluate the
influence of electronic and steric effects on the structure of
their nickel complexes, we used the two series of ligands
shown in Scheme 1: on one hand, the S-methyl-1-phenyl-
isothiosemicarbazide series H2[L1-3] with p-phenyl substit-
uents H, NO2, OCH3, and, on the other hand, the bulkier
1,4-diphenyl series H2[L4-6] where the 1-phenyl group again
carries an H, NO2, or OCH3 substituent in the para-position.
The corresponding neutral, highly colored bis(ligand)-
nickel(II) complexes 1-6 (Scheme 3) were synthesized by
reaction of 2 equiv of the respective ligand H2[L1-6]‚HI
dissolved in ethanol in the presence of excess triethylamine
and air with 1 equiv of [NiII(CH3CO2)2]‚4H2O. Dark pre-
cipitates of microcrystalline powders were obtained in good
to excellent yields. The EI mass spectra of 1-6 each display
a molecular ion peak {M}+ in agreement with their formula-
tion as [NiII(L1-6•)2]. Complexes 1 and 5 have been prepared
previously.1,3 Species 1-6 are diamagnetic (S ) 0) as was
judged from their 1H NMR spectra and magnetic susceptibil-
Oxidation of tetrahedral [NiII(L6•)2] (6) with iodine in
chloroform affords red single crystals of the metal free,
oxidized ligand (L6,ox 0
) (10) in 28% yield (Scheme 4).
Attempts to prepare this molecule directly from H2[L6]‚HI
by oxidation with I2 led to complicated mixtures of organic
compounds, which we did not investigate further.
An attempt to synthesize the analogous tetrahedral zinc
complexes [Zn(L6•)2] and [Zn(L5•)2] starting from H2L6‚HI
and H2L5‚HI, respectively, and Zn(OAc)2‚2H2O led to the
isolation of (L6,ox 0
) and the heterocyclic compound 11
(Scheme 4). The latter is an isomer of (L5,ox)0, resulting from
a (2+4) cycloaddition of one C-C double bond of the
4-phenyl ring with the NdC-NdN imine-hydrazone frag-
ment, followed by rearomatization via a 1,3-H shift.
Figure 2 shows the electronic spectra of square planar
complexes 1-3 and of their tetrahedral analogues 4-6. The
spectra of 1-3 are dominated by a very intense (ꢀ ∼ 1.5 ×
104 M-1 cm-1) spin and dipole allowed ligand-to-ligand
charge-transfer band (LLCT) at 700-800 nm. This absorp-
tion maximum is typical for square planar nickel(II) com-
plexes containing two π-radical monoanionic ligands.19,20
(15) Neese, F. ORCA, an ab Initio Density Functional and Semiempirical
Electronic Structure Program Package, Version 2.3, Revision 07; Max-
Planck-Institut fu¨r Bioanorganische Chemie: Mu¨lheim, Germany,
November 2003.
(16) (a) Becke, A. D. J. Chem. Phys. 1988, 84, 4524. (b) Perdew, J. P.
Phys. ReV. 1986, 33, 8522. (c) Lee, C.; Yang, W.; Parr, R. G. Phys.
ReV. B. 1988, 37, 785. (d) Becke, A. D. J. Chem. Phys. 1993, 98,
5648.
(17) (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,
5829.
(18) (a) Neese, F. Inorg. Chim. Acta 2002, 337, 181. (b) Neese, F. J. Chem.
Phys. 2003, 119, 9428.
(19) Herebian, D.; Bothe, E.; Neese, F.; Weyhermu¨ller, T.; Wieghardt, K.
J. Am. Chem. Soc. 2003, 125, 9116 and references therein.
(20) (a) Herebian, D.; Wieghardt, K.; Neese, F. J. Am. Chem. Soc. 2003,
125, 10997. (b) Ghosh, P.; Bill, E.; Weyhermu¨ller, T.; Neese, F.;
Wieghardt, K. J. Am. Chem. Soc. 2003, 125, 1293.
3642 Inorganic Chemistry, Vol. 44, No. 10, 2005

Diamagnetic Complexes of Nickel(II)
Figure 2. Electronic spectra of square planar complexes 1, 2, and 3 (top)
and of tetrahedral species 4, 5, and 6 (bottom) in CH2Cl2 solution.
Significantly, this LLCT absorption is not observed in the
spectra of tetrahedral complexes 4-6.
The spectrum of octahedral 7 in CH2Cl2 displays a weak
absorption maximum at 716 nm (ꢀ ∼ 1.8 × 103 M-1 cm-1),
and stronger ones at 450sh (1.0 × 104), 380 (2.4 × 104).
The latter bands are similar to those observed for the ligand
10, (L6,ox).
Crystal Structures. The crystal structures of 1 and 5-11
have been determined by X-ray crystallography; in all cases
data sets have been collected at 100(2) K. The structures of
1 and 5 determined at 298 K have been described previ-
ously.1,3 Agreement between these and the present structure
determinations is excellent, but the average esd’s of N-N,
C-N, and C-C distances at 298 K is three to four times
larger than those obtained from the 100 K data sets here.
Thus, the accuracy of these redetermined structures of 1 and
5 has been significantly improved. Crystallographic data are
summarized in Table 1; important bond distances are given
in Table 2.
Figure 3 shows the structures of the neutral molecules [NiII-
(L1•)2], [NiII(L5•)2], and [NiII(L6•)2] in crystals of 1, 5, and 6,
respectively. In all three cases, the C-N, N-N, and C-S
bond distances of the five-membered chelate rings Ni-N1-
N2-C1-N3 are identical within experimental error (3σ ∼
(0.006 Å): average N1-N2 at 1.342 Å, C1-N2 at 1.346 Å,
C1-N3 at 1.324 Å, and C1-S at 1.752 Å. This is a clear
indication that the oxidation level of the ligand is the same
in 1, 5, and 6 irrespective of the N-substituents or the
geometry at the nickel(II) ion (square planar vs tetrahedral).
In accordance with previous authors,1-3 we conclude that
the ligands are monoanionic π-radicals (L1•)1- in 1, (L5•)1-
in 5, and (L6•)1- in 6. This has also been shown for iron and
Figure 3. Structures of the neutral molecules in crystals of 1, 5, and 6,
respectively.
cobalt complexes containing two π-radical anions derived
from S-methyl-1-phenyl-isothiosemicarbazide.5,6
The two best planes defined each by the five atoms of the
five-membered chelate rings in 1, 5, and 6 have dihedral
angles of 12°, 85°, and 71°, respectively. Thus, the coordina-
Inorganic Chemistry, Vol. 44, No. 10, 2005 3643

Blanchard et al.
Table 2. Selected Bond Distances (Å) and Angles (deg) of Complexes
complex 1
complex 7
Ni-N10
1.8387(9)
1.8434(9)
1.8518(8)
1.8524(8)
1.417(1)
1.350(1)
1.339(1)
1.322(1)
98.41(4)
81.51(4)
81.51(4)
99.20(4)
C6-N7
1.418(1)
1.350(1)
1.339(1)
1.319(1)
1.752(1)
1.800(1)
1.751(1)
1.802(1)
169.23(4)
176.64(4)
Ni-I1
2.7523(2)
N8-C9
1.437(2)
1.276(2)
1.736(1)
1.805(2)
Ni-N30
N7-N8
Ni-N10
2.021(1)
2.113(1)
1.427(2)
1.267(2)
180.0
104.85(5)
75.15(5)
C9-N10
C9-S11
S11-C12
Ni-N27
Ni-N7
N8-C9
C9-N10
C9-S11
S11-C12
C29-S31
S31-C32
N10-Ni-N30
N27-Ni-N7
Ni-N7
C6-N7
C26-N27
N27-N28
N28-C29
C29-N30
N10-Ni-N27
N30-Ni-N27
N10-Ni-N7
N30-Ni-N7
N7-N8
N10*-Ni-N10
N10*-Ni-N7
N10-Ni-N7
complex 8
Ni-N30
Ni-N50
Ni-N10
C6-N7
N7-N8
N8-C9
N27-N28
2.019(5)
2.030(4)
2.031(5)
1.427(7)
1.254(7)
1.446(7)
1.272(6)
Ni-N7
2.144(5)
2.168(5)
2.180(4)
1.277(7)
1.722(6)
1.798(6)
1.265(6)
Ni-N27
Ni-N47
C9-N10
C9-S11
S11-C12
N47-N48
complex 5
Ni-N10
1.8905(9)
1.8970(9)
1.9070(9)
1.9200(9)
1.333(1)
1.407(1)
1.465(1)
1.355(1)
1.750(1)
1.336(1)
C6-N7
1.433(1)
1.326(1)
1.352(1)
1.749(1)
1.807(1)
1.432(1)
1.325(1)
1.407(1)
Ni-N40
N7-C8
Ni-N37
C8-N9
Ni-N7
C8-S20
S20-C21
C36-N37
N37-C38
N40-C41
complex 9
N9-N10
Ni-N10
2.068(2)
2.099(2)
2.7611(4)
1.736(3)
1.443(3)
3.1193(5)
C6-N7
N7-C8
C8-N9
N9-N10
S17-C18
I32-I33
1.440(4)
1.290(4)
1.440(4)
1.259(3)
1.811(3)
2.7808(4)
N10-C11
C14-N17
C38-N39
C38-S50
N39-N40
N10-Ni-N40
N10-Ni-N37
N40-Ni-N37
N10-Ni-N7
Ni-N7
Ni-I31
C8-S17
N10-C11
I31-I32
122.19(4)
N40-Ni-N7
N37-Ni-N7
130.84(4)
124.96(4)
I33-I32-I31
N10-Ni-N7
174.65(1)
75.14(9)
N10*-Ni-N7
104.86(9)
124.92(4)
80.20(4)
80.39(4)
compound 10a
C6-N7
N7-C8
C8-N9
S19-C20
1.421(6)
1.262(5)
1.435(5)
1.799(5)
C8-S19
N9-N10
N10-C11
1.773(4)
1.258(5)
1.413(5)
complex 6
Ni-N10
1.870(1)
1.913(1)
1.418(2)
1.328(2)
1.346(2)
1.758(2)
123.94(8)
134.90(6)
80.68(6)
N9-N10
N10-C11
S19-C20
1.343(2)
1.411(2)
1.806(2)
Ni-N7
C6-N7
compound 11b
N7-C8
C1-C6
1.401(2)
1.404(2)
1.382(2)
1.420(2)
1.375(2)
C8-N9
C8-S20
S20-C21
1.289(2)
1.753(2)
1.803(2)
C8-N9
C6-N7
N7-C8
N9-N10
N10-C11
C8-S19
N10-Ni-N10*
N7-Ni-N10*
N10-Ni-N7
N10*-Ni-N7*
N10-Ni-N7*
N7-Ni-N7*
80.68(6)
134.90(6)
109.28(8)
a Data for one coordinated (L1,ox) are given. b Data for one of four crystallographically independent molecules are given.
tion geometry of the four nitrogen donors around the nickel-
(II) ions is best described as slightly distorted square planar
in 1 and distorted tetrahedral in 5 and 6. This dihedral angle
is 0° in a completely planar bis(chelate)nickel complex and
90° in an ideal tetrahedron.
Scheme 5. Geometric Isomers
The two N,N-coordinated bidentate S-methyl-1-phenyl-
isothiosemicarbazide ligands (irrespective of their oxidation
level) can adopt a cis- or trans-configuration relative to each
of the complexes with a square planar MN4 polyhedron as
shown in Scheme 5. Experimentally, only trans-configurated
square planar species have been characterized to date as in
1 (and probably also 2 and 3). It is now important to realize
that in distorted tetrahedral complexes (4, 5, and 6) two
geometrical isomers, namely, A and B in Scheme 5, exist if
the dihedral angle, θ, between the two five-membered chelate
rings is not 90° as in 5 and 6. Both isomers A and B possess
C2 symmetry, but the intramolecular distance between the
nitro groups in 5 (or the two methoxy groups in 6) differ in
both isomers if θ * 90°. It is conceivable that A and B easily
interconvert in solution.
The crystal structure of octahedral trans-[NiII(L1,ox)2I2]‚
2CHCl3, 7‚2CHCl3, has been determined using a room-
temperature data set.2 Nonsolvated single crystals of 7 were
grown by slow evaporation of a CH2Cl2/toluene mixture of
7. Its crystal structure has been determined at 100(2) K. The
new structure is significantly more accurate than the previous
determination2 but does not reveal structural differences of
the neutral complex. Figure 4 exhibits the structures of 7
and 8. The two five-membered chelate rings Ni-N1-N2-
C1-N3 in 7 and three in the dication in crystals of 8 have
The average Ni-N bond distances in square planar 1 are
shorter at av. 1.85 Å than in the corresponding tetrahedral
species 5 and 6 at 1.90 Å, which reflects the difference of
ionic radii between a diamagnetic square planar Ni(II) ion
(0.63 Å) and a paramagnetic tetrahedral one (0.69 Å).21
(21) Shannon, R. D. Acta Crystallogr. 1976, A32, 751.
3644 Inorganic Chemistry, Vol. 44, No. 10, 2005

Diamagnetic Complexes of Nickel(II)
Figure 5. Structure of neutral molecule trans-[NiII(L5,ox)2(I3)2] in crystals
of 9.
Figure 4. Structures of the neutral molecule in crystals of 7 (top) and of
the dication in crystals of 8 (bottom).
identical N-N, C-N, C-S bond distances, which clearly
differ from those in 1, 5, and 6. Remarkably, the average
N1-N2 bond length at 1.265 Å is significantly shorter than
in 1, 5, and 6; it indicates the presence of an NdN double
bond. Similarly, the average C1-N3 bond length at 1.277 Å
indicates the presence of a CdN double bond, whereas the
C1-N2 bond at 1.438 Å is representative of a C-N single
bond. Thus, the ligands in 7 and 8 are the diamagnetic,
neutral form as in 10.
Figure 5 shows the structure of the neutral molecule trans-
[NiII(L5,ox)2(I3)2] in crystals of 9. The two 1,4-diphenyl ligand
derivatives (L5,ox) are coordinated in trans-position relative
to each other, the NiN4 moiety being completely planar. The
two neutral ligands (L5,ox)0 are each N,N-bound in a fashion
that maximizes the intramolecular distance between the two
nitro substituents (they are in trans-position with respect to
each other). The phenyl group of one (L5,ox) ligand and the
nitrophenyl group of the other are opposite to each other in
the neutral molecule; the dihedral angle between these two
phenyl rings is 7.8° and the distance between the centroids
of both rings is 3.465 Å indicative of a weak π-stacking.
Thus, there appears to be no steric hindrance between the
two ligands when coordinated in a “planar” manner in 9.
The average Ni-N distance in octahedral at 2.083 Å is longer
than in square planar 1 at 1.847 Å. Simple geometrical
considerations then yield a ring-to-ring π-stacking distance
of 2.8-2.9 Å in a hypothetical square planar [Ni(L5•)2]
complex. This π-stacking distance may be too short and may
Figure 6. Structures of neutral molecules in crystals of 10 (top) and 11
(bottom).
explain why [Ni(L5•)2] (5) favors a tetrahedral structure, but
the situation is borderline. In {[Pd(bpy)(B•)](PF6)}2 the ring-
to-ring distance of the π-radical of o-phenylenediamine,
(B•)1-, is only 2.92 Å.22
Figure 6 displays the crystal structures of neutral 10 and
11, both of which do not contain a transition metal ion. 10
Inorganic Chemistry, Vol. 44, No. 10, 2005 3645

Blanchard et al.
Table 4. Redox Potentials of Complexesa
E11/2 (0/-1), E12/2 (-1/-2)
complex Epox, Vb
Epred, Vc
Vd
Vd
Kd
e
119
2
3
0.05 -0.24
-1.24
-0.75
-1.29
-1.31
-0.85
-1.43
-1.865
-0.97
-2.01
-1.64
-1.01
-1.80
2.6 × 10-5
2.4 × 10-2
1.0 × 10-6
3.6 × 10-3
6.6 × 10-2
2.1 × 10-3
0.22
-0.03
0.07
0.0
0.15, 0.26
0.10
4
5f
6
0.37
0.15
0.04 -0.15
a Referenced vs the Fc+/Fc couple. b Peak potential for oxidation. c peak
potential for the corresponding reduction(s). d Half-wave potentials for the
0/-1 (and -1/-2) couple. e Disproportionation constant according to eqs
3
f
1 and 2. E1/2 (-2/-3) is observed at -2.02 V. Conditions: scan rate 100
mV s-1; 20 °C; CH2Cl2 solution (0.20 M [(n-Bu)4N]PF6).
HArNN) positions opposite the hydrazine substituent for the
three uncoordinated H2[L1-3]‚HI ligands and the six com-
plexes 1-6. We note that the chemical shift values of the
meta-protons in the uncoordinated ligand and the corre-
sponding nickel complex are similar and only slightly shifted
upon complexation to lower fields by 0.16 ppm (R ) H),
0.12 (R ) NO2), and 0.09 ppm (R ) OCH3). In contrast, in
the same series the signals of the ortho protons increase by
1.18 ppm (R ) H), 1.24 ppm (R ) NO2), and 1.17 ppm (R
) OCH3). Such a variation cannot originate from simple
solvent effects; it confirms, instead, the presence of coordi-
nated species in solution. Moreover, this rather drastic change
indicates an important electronic modification in the nature
of ortho-C-H bonds. We attribute this change to the ligand
oxidation on going from the uncoordinated, closed-shell
ligands to the coordinated radicals in [Ni(LR•)2].
Interestingly, the o-HArNN proton signals of the three
tetrahedral [Ni(LR•Ph)2] complexes are much more deshielded
with 9.52 ppm for 4, and 5, and 9.57 ppm for 6 than those
of the square planar complexes [Ni(LR•)2] with 7.43 for 1,
8.29 ppm for 2, and 6.95 ppm for 3. Thus, the tetrahedral
and square planar complexes exhibit significantly different
signals in solution at ambient temperature. No additional
peaks or coalescence processes were observed in the
spectrum of [Ni(L1•)2] (1) recorded in CD2Cl2 in the
temperature range 300 to 215 K. Thus, the tetrahedral and
square planar complexes are not in equilibrium; they remain
intact in their given coordination polyhedron even in solution.
Electro- and Spectroelectrochemistry. The electrochem-
istry of complexes 1-6 has been studied in CH2Cl2 solutions
containing 0.2 M [(n-Bu)4N]PF6 as supporting electrolyte
by cyclic voltammetry (cv). Ferrocene was used as internal
standard; all redox potentials are referenced versus the
ferrocenium/ferrocene (Fc+/Fc) couple, and the results are
summarized in Table 4. In general, the complex concentration
was in the millimolar range, but low solubility of 2 and 3
enforced lower concentrations at ∼0.1 mM for 3 and ∼10
µM for 2. The cv of 1 has been reported previously.23
As shown in Figure 8, the complexes display very similar
electrochemical behavior irrespective of the coordination
polyhedron (square planar 1-3 or tetrahedral 4-6): all
undergo a two-electron oxidation in the range -0.03 to 0.40
1
Figure 7. 400 MHz H NMR spectra of complexes 1-6 (CDCl3) at 300
K in the range 5.8-10 ppm.
1
Table 3. 400 MHz H NMR Spectra of Free Ligands and Complexes
speciesa
δ(o-HArNN proton)b
δ(m-HArNN proton)c
H2 L1‚HI
6.79
6.92
6.76
7.97
8.16
8.03
9.52
9.52
9.57
7.27
8.17
6.86
7.43
8.29
6.95
7.03
7.91
6.61
H2 L2‚HI
H2 L3‚HI
1
2
3
4
5
6
a Spectra of free ligands were measured in CD3OD and those of
complexes 1-6 in CDCl3 at 300 K. b Chemical shift values are given in
ppm for the ortho hydrogens of the 1-phenyl group of the ligand. c Chemical
shift values are given in ppm for the meta hydrogens of the 1-phenyl group
of the ligand.
represents (L6,ox)0, whereas 11 is a cyclic isomer of (L5,ox)0.
The short N-N bond at 1.262(2) Å in 10 represents a double
bond, whereas the same bond in 11 at 1.420(5) Å is typical
for a single bond in agreement with the formula shown in
Scheme 4.
NMR Studies of 1-6. Since all [Ni(Lx•)2] complexes (1-
6) possess a diamagnetic ground-state irrespective of the
presence of a tetrahedral coordination polyhedron as in 4-6
or a square planar one as in 1-3, the NMR spectra have
been investigated to explore the possibility of an equilibrium
between the two forms in solution. The 1-D and 2-D NMR
spectra of 1 and 4-6 were recorded and analyzed. Com-
plexes 2 and 3 are quite insoluble, and, therefore, only their
1H NMR spectra were recorded.
Figure 7 shows the 1H NMR spectra of 1-6 in the region
5.8 to 10 ppm at 300 K. In all cases only a single set of
signals for the two coordinated ligands is observed indicating
that both ligands are magnetically equivalent on the NMR
time scale. A single broad N-H signal is observed for the
square planar complexes at 5.88 ppm for 1, 5.99 ppm for 2,
and 5.85 ppm for 3; 4-6 do not display such a signal. Table
3 summarizes the values of the chemical shift, δ, of the
1-phenyl protons in the ortho (o-HArNN) and meta (m-
(22) Ghosh, P.; Begum, A.; Herebian, D.; Bothe, E.; Hildenbrand, K.;
Weyhermu¨ller, T.; Wieghardt, K. Angew. Chem., Int. Ed. 2003, 42,
563.
(23) The cyclic voltammogram of 1 in acetonitrile has been reported:
Tarazewska, J.; Revenco, M. D.; Timco, L. Pol. J. Chem. 1997, 71,
253.
3646 Inorganic Chemistry, Vol. 44, No. 10, 2005

Diamagnetic Complexes of Nickel(II)
Table 5. Electronic Spectra in CH2Cl2 Solution
complexa
λ )
max, nm (104 ꢀ, M-1 cm-1
1
710(1.5), 465sh(0.5), 370sh(1.0), 310(2.6)
1120(101), 390(1.1), 305(2.5)
405(1.8), 295(2.4)
795(1.2), 385(2.5), 290(1.7)
1385(1.2), 755(1.2), 480(2.6)
760(3.1), 555(2.0), 475(2.2), 365(1.0), 255(1.9)
720(1.5), 465sh(0.8), 405(1.2), 315(1.7), 296sh(1.6)
1115(1.0), 390(1.1), 305(2.1)
900(0.1), 745(0.3), 495(1.7), 381sh(1.1), 305(2.5),
240(1.8)
905(0.1), 745(0.3), 490(1.1), 300(2.2), 240(2.0)
910(0.2), 700sh(0.3), 600sh(0.4), 475(0.6), 345sh(2.3),
320(2.3), 243(2.2)
[1]-
[1]2-
2
[2]-
[2]2-
3
[3]-
4
[4]-
[4]2-
5
800(0.5), 545(1.5), 390(3.7), 265(2.4)
665(1.6), 495(2.7), 410(2.3)
710(4.1), 650(3.9), 595(5.1), 555sh(3.7), 470sh(2.0),
360(1.6), 305(1.9)
[5]-
[5]2-
6
750(0.3), 495(2.1), 405(1.3), 310(2.3), 265(2.0)
905(0.2), 705(0.3), 480(1.0), 315(2.6), 255(2.2)
650sh(0.3), 485sh(0.7), 360(2.9), 330(2.8), 250(2.2)
715(0.2), 435sh(0.9), 340(2.5)
[6]-
[6]2-
7
Figure 8. Cyclic voltammograms of complexes 1-6 in CH2Cl2 solution
[0.20 M [N(n-Bu)4]PF6 supporting electrolyte, glassy carbon working
electrode, 20 °C, ferrocene (Fc) internal standard, scan rate 100 mV s-1].
[9]
[10]
790(0.5), 535(2.2), 385(4.8), 265(3.4)
495(0.2), 358(2.1)
a The mono- and dianions were generated electrochemically in CH2Cl2
solutions (0.20 M [N(n-Bu)4]PF6).
V and a two-electron reduction in the range -0.2 to 0.0 V.
These waves have irreversible character (∆Ep ∼ 300 mV)
and have not been investigated further. In addition, all
complexes display two successive, reversible one-electron
reduction waves. Only complex 5 exhibits an additional
reversible one-electron reduction at E31/2 ) -2.02 V, which
we assign to a metal-centered reduction NiII/NiI of the couple
[NiII(L5)2]2-/[NiI(L5)2]3-.
The redox potentials within the series of square planar
complexes 1-3 clearly vary as a function of the nature of
the p-substituent of the 1-phenyl group: the electron-
withdrawing properties of the nitro group in 2 makes the
neutral species more difficult to oxidize (by ∼0.2 V) and
easier to reduce (by 0.4 or 0.9 V) than the corresponding
unsubstituted species 1.
For the methoxy electron-donating substituted species 3
the opposite trends are observed in comparison with the data
for unsubstituted 1. Similar trends hold for the tetrahedral
complexes 4-6.
The disproportionation constant (corresponding to a com-
proportionation constant Kc ) Kd-1) for the equilibrium
between the monoanionic forms of complexes 1-6 and their
neutral and dianionic forms eq 1 can be calculated by use
of eq 2.
a ligand mixed valency of class III. In contrast, Kd for the
tetrahedral species 4, 5, and 6 is larger by at least 3 orders
of magnitude. We take this as an indication that the
corresponding mixed valent monoanions possess a more
localized electronic structure [NiII(Lx,red)(Lx•)]- (class I or II).
Controlled potential electrolysis at -20 °C of CH2Cl2
solutions of the neutral complexes 1-6 (0.20 M [(n-Bu)4N]-
PF6) at appropriately fixed negative potentials allowed the
generation of stable solutions containing the corresponding
mono- or dianions, for which it has been possible to record
their electronic spectra (Table 5). An example for 1, [1]-,
and [1]2- is shown in Figure 9. It is remarkable that the
ligand-to-ligand charge-transfer band (LLCT) at 717 nm (ꢀ
) 1.5 × 104 M-1 cm-1) of 1 is shifted to 1122 nm (ꢀ )
1.06 × 104 M-1 cm-1) upon one-electron reduction to the
monoanion and disappears completely upon further one-
electron reduction yielding the dianion. This is excellent
evidence that both reductions are ligand-centered processes,
eq 3.19,20
[NiII(L1•)2] w-ex [NiII(L1)(L1•)]- w+-ee
\ \
x [NiII(L1)2]2- (3)
+e
Kd
\
Similarly, the LLCT band at 722 nm (ꢀ ) 1.5 × 104 M-1
cm-1) for 3 shifts to 1115 nm (ꢀ ) 1.0 × 104 M-1 cm-1) in
the monoanion (Figure S3, Supporting Information). For the
nitro complex 2 (Figure 9), the shift of the LLCT band is
observed from 795 nm (ꢀ ) 1.1 × 104 M-1 cm-1) to 1385
nm (ꢀ ) 1.25 × 104 M-1 cm-1). The spectrum of [2]- also
displays an intense band at 756 nm (ꢀ ) 1.15 × 104 M-1
cm-1), which increases in intensity to ꢀ ) 3.45 × 104 M-1
cm-1 upon generating the dianion [2]2- while the LLCT band
at 1400 nm disappears. We propose that the 756 nm band
corresponds to a π-π* transition of the dianionic form of
the nitro ligand H2[L3]. In the N,N-coordinated dianion the
strong electron acceptor nitro substituent is in the para
2 [NiII(Lx•)(Lx,red)]- w
x [NiII(Lx,red)2]2- + [Ni(Lx•)2]0 (1)
Kd ) exp (nF(E21/2 - E11/2)/RT ) exp((E21/2 - E11/2)/0.059)
(2)
The results are summarized in Table 4. A small value of Kd
(or a large one for Kc) as for square planar 1 and 3 indicates
a relatively large electronic stabilization of the corresponding
square planar mixed valent monoanion, i.e., this species is
probably a class III mixed valent species where the unpaired
electron is delocalized over both ligands. It is unclear to us
why Kd of 2 is so large since the electronic spectrum of its
monoanion is very similar to that of [1]- and [3]-, indicating
Inorganic Chemistry, Vol. 44, No. 10, 2005 3647

Blanchard et al.
Figure 10. Electronic spectra of electrochemically generated mono- and
dianions of 4 (top) and 5 (bottom) in CH2Cl2 solutions (0.20 M [N(n-Bu)4]-
PF6).
Figure 9. Electronic spectra of electrochemically generated mono- and
dianions of 1 (top) and 2 (bottom) in CH2Cl2 solutions (0.2 M [N(n-Bu)4]-
PF6).
tetrahedral containing localized, N,N-coordinated mono- and
dianionic ligands and central Ni(II) ions, [NiII(Lx•)(Lx)]- (x
) 4-6).
position of an aromatic phenyl ring with a strong electron
donating hydrazido function: the molecule displays the
characteristic behavior of an organic dye. These results
clearly indicate that the monoanions of 1-3 are also square
planar species.
Electron Spin Resonance Spectra of the Monoanions
[NiII(Lx•)(Lx)]- (x ) 1-6). The monoanionic forms of
complexes 1-6, namely, square planar or tetrahedral
[NiII(Lx•)(Lx)]- species, have been generated electrochemi-
cally from the neutral forms 1 - 6 in CH2Cl2 solutions
containing 0.2 M [(n-Bu)4N]PF6. All of them possess an S
The tetrahedral complexes 4-6 display this LLCT band
neither in the neutral nor in their monoanionic forms,
suggesting that these complexes remain tetrahedral upon
reduction. The spectral changes observed are shown in Figure
10 for 4 upon reduction to [4]- and [4]2-. It is noted that
these reductions proceed with retention of isosbestic points,
which indicates the clean formation of the mono- and
dianionic forms from neutral 4-6.
Figure 10 displays the electronic spectra of the nitro
complexes 5, [5]-, and [5]2-. In contrast to the results for 4
and 6 (Figure S4, Supporting Information), the evolution of
spectra upon one- and two-electron reduction is dominated
by an increase of absorption in the visible region. As has
been discussed for the spectral changes of 2 upon reduction,
this increase in absorption can be understood as a result of
π-π* transition within the N,N-coordinated ligand dianion
[L5]2- (nitro-hydrazide push-pull mechanism). This would,
at the same time, indicate that both reductions of tetrahedral
4-6 are ligand-centered processes. The lack of such LLCT
bands for the monoanions indicates that they are also
1
) /2 ground state and display characteristic X-band EPR
spectra as shown in Table 6.
The EPR spectra of the monoanionic forms of complexes
1-3 exhibit quasi- rhombic signals that are very similar to
those reported for monoanionic species of [NiL2] where L
represents a dioxolene type ligand.24-26 Figure 11 shows the
spectrum of [1]-; those for [2]- and [3]- are shown in Figure
S5 (Supporting Information). The relative large g-anisotropy
of these square planar monoanions has been traced25,26 to
the nature of the SOMO involved. This SOMO has been
shown to be ligand-centered but with symmetry allowed,
significant mixing of a d metal orbital yielding a ground state
with sizable orbital angular momentum. This mechanism
(24) Chaudhuri, P.; Verani, C. N.; Bill, E.; Bothe, E.; Weyhermu¨ller, T.;
Wieghardt, K. J. Am. Chem. Soc. 2001, 123, 2213.
(25) Herebian, D.; Bothe, E.; Bill, E.; Weyhermu¨ller, T.; Wieghardt, K. J.
Am. Chem. Soc. 2001, 123, 10012.
(26) Sun, X.; Chun, H.; Hildenbrand, K.; Bothe, E.; Weyhermu¨ller, T.;
Neese, F.; Wieghardt, K. Inorg. Chem. 2002, 41, 4296.
3648 Inorganic Chemistry, Vol. 44, No. 10, 2005

Diamagnetic Complexes of Nickel(II)
Table 6. X-band EPR Spectra of Electrochemically Generated
Monoanions of Complexes 1-6 in CH2Cl2 Solution (0.2 M
[(n-Bu)4N]PF6) at -20 °C
liquid N2+ pentane/
freezing liquid N2
process
(%)
process
(%)
complex
gx
gy
gz
giso
[1]-
[2]-
[3]-
2.081 2.011 2.004 2.032
2.067 2.027 2.008 2.034
2.080 2.018 2.009 2.063
2.077 2.014 1.994 2.028
square planar
[Ni(Lcalcd)2]- a
[4]- subspecies a 2.265 2.157 2.008 2.143
72
28
65
35
85
15
58
42
subspecies b
[5]- subspecies a 2.31 2.17 2.01 2.163
subspecies b 2.22 2.12 2.03 2.12
[6]- subspecies a 2.257 2.160 2.011 2.143
2.215 2.133 2.043 2.130
63
37
subspecies b
2.198 2.137 2.043 2.126
2.205 2.198 2.037 2.147
[6]- a
a From DFT calculations (see text).
Figure 12. X-band EPR spectrum of electrochemically generated [4]- in
CH2Cl2 solution (0.2 M [N(n-Bu)4]PF6) at 80 K. (a) Spectrum of shock
frozen sample in a liquid n-heptane/N2 mixture and (b) in liquid N2. Two
subspectra a with gx ) 2.265, gy ) 2.157, gz ) 2.008 and b with gx )
2.215, gy ) 2.133, gz ) 2.043.
in an equilibrium. The EPR spectra are shown in Figures
S6 and S7 (Supporting Information), respectively.
We note that the g-anisotropy of the monoanions [4]-,
[5]-, and [6]- (irrespective of their configuration a or b) is
significantly larger than is observed for the square planar
monoanions [1]-, [2]-, or [3]-. This rules out the possibility
that the monoanions [4]-, [5]- and [6]- exist in a square
planar configuration in solution. It would appear that both
configurations a and b are in a distorted tetrahedral
environment.
Figure 11. X-band EPR spectrum of electrochemically generated [1]- in
CH2Cl2 solution (0.2 M [N(n-Bu)4]PF6) at 10 K.
requires a square planar configuration of the monoanions of
1-3. No resolved hyperfine coupling to the nitrogen donor
atoms has been observed. The g values are given in Table
6.
The solid-state structures of 5 and 6 have revealed an
interesting facet. In both structures, the NiN4 polyhedron is
that of a severely distorted tetrahedron. The two ligand planes
Figure 12 exhibits the X-band EPR spectrum of a frozen
CH2Cl2 solution of electrochemically generated [4]- at 80
K. To our surprise, this spectrum could only be satisfactorily
simulated by taking into account two subspectra a and b.
Both spectra display rather similar but not identical gx, gy,
gz values (Table 6). The ratio of species a/b was found to
be 58:42 when the -20 °C electrochemically generated
solution of [4]- was shock-frozen in a liquid n-pentane/liquid
dinitrogen bath. Upon changing the freezing procedure by
using pure liquid dinitrogen, the resulting frozen solution
contains the two species a and b in a ratio 72:28. These
experiments clearly show that species a and b are in a
thermodynamic equilibrium in solution with rapid transfor-
mation rates. From biological studies it is well-known that
the heat transfer rates in liquid N2 differ markedly from those
in liquid n-pentane/N2 (it is slow in the former and faster in
the latter process).
defined each by the five-membered chelate Ni-N1-N2-
C1-N3 form in 5 and 6 a dihedral angle of 84° and 65° (and
not 90° as is required for an ideal tetrahedron). Thus, in a
distorted tetrahedral complex containing two identical but
asymmetric bidentate ligands the two geometrical configura-
tions A and B shown in Scheme 5 exist. We propose that
these two isomers are in equilibrium in the [4]-, [5]-, and
[6]-; they can easily and rapidly be interconverted.
Calculations. In this section, a detailed picture of the
electronic structures of the neutral [NiII(Lx•)2] complexes and
of their corresponding monoanions will be developed. In an
attempt to dissect influences of electronic from steric effects,
we calculated a truncated (small) model that is labeled [Ni-
1-
(L•calcd)2] where (L•calcd
)
is (HN-NdC(SH)NH•)1-. The
It is quite remarkable that [5]- and [6]- display the same
behavior. In each case two isomers, namely, a and b, exist
coordination geometry of the complex was assumed to be
tetrahedral and/or square planar. Then we calculated the
Inorganic Chemistry, Vol. 44, No. 10, 2005 3649

Blanchard et al.
lengths in both calculated geometries of [NiII(L1•)2] are very
similar and very close to those of the experimental structure
(Table 2). In particular, the N1-N2, N2-C3, and C3-N4
bond lengths are in both cases in excellent agreement with
the experimental findings and characteristic of a radical
monoanion form of the ligand. Finally, the B3LYP calcula-
tions, with dNi-N ) 1.891 Å, gave a slightly longer average
Ni-N bond distance than BP86 calculations (dNi-N ) 1.862
Å). They are both longer than the experimental value at 1.847
Å. Both functionals tend also to overestimate the dihedral
angle between the two NCNNNi metallacycles, with angles
of 19° and 22°, respectively, as compared to an experimental
value of 12°.
The B3LYP calculated geometry of the truncated model,
[NiII(L•calc)2], displays the main characteristics of the experi-
mental findings. However, the two NCNNNi metallacycles
are coplanar, suggesting that the twist found in the experi-
mental structure is related to steric interactions involving the
1-phenyl rings. This decrease of the dihedral angle in the
truncated model is associated with a decrease of the Ni-N
average bond distance from 1.891 Å in [NiII(L1•)2] (1) to
1.861 Å in [NiII(L•calc)2]. It has been shown previously for
related complexes20 that the metal character of the 3bg MO
(see electronic structure below) indicates that there is
significant metal-to-ligand π-back-donation. The difference
in Ni-N bond length is therefore likely to be related to a
larger π-back-bonding interaction from the nickel to the
ligand in a true square planar environment as compared to a
distorted one.
For the square planar monoanions, [NiII(L1)2]- and
[NiII(Lcalc)2]-, the geometries were optimized in spin-
unrestricted S ) 1/2 calculations. The calculated N1-N2 and
C3-N4 bond lengths increase as compared to that of the
corresponding neutral molecules, while the N2-C3 bond
length decreases (see Table 4). These variations are in
agreement with a ligand-centered reduction. It is important
to note that in the optimized structures of the square planar
monoanions the bond distances within the two metallarings
are identical. This indicates delocalization of the added
electron over the two ligands. These results can be explained
by considering the LUMO of the neutral molecules (3bg-
(π*)) shown in Figure 14. This orbital, corresponding to the
antisymmetric combinations of the SOMO of the free
π-radical anions, is indeed mainly ligand centered with a
small metal contribution. Upon reduction, it becomes singly
occupied. The antibonding character of this orbital between
N1 and N2 and also C3 and N4 and bonding character
between N2 and C3 thus explains the observed bond length
modifications. This also explains the increase of the average
Ni-N bond distance from 1.840 Å to 1.864 Å because
population of this orbital reduces the metal π-back-bonding
donation. This MO is metal-to-ligand antibonding in char-
acter. Thus, the calculations are in agreement with a ligand-
centered reduction in the case of the square planar complexes.
Figure 13. Calculated bond distances in the free ligands H2[Lcalcd],
[L•calcd
]
1-, and [Lcoaxlcd]0. The HOMO of H2[Lcalcd] is also shown (it is the
ox
1-
SOMO in [L•calcd
]
and the LUMO in [Lcalcd]0).
structures of 1, 5, and 6 and the corresponding monoanionic
forms thereof.
It is educational to calculate the structure of the ligand in
three different oxidation levels first: the optimized geom-
etries of (H2N-NdC(SH)NH2) (H2Lcalcd); its anionic radical
(HN-NdC(SH)NH•)1-, (L•calcd)1-, and of the neutral oxi-
ox
calcd
dized molecule HNdN-C(SH)dNH, (L ), were eluci-
dated. The results of the B3LYP optimizations are shown in
Figure 13 together with the redox-active orbital, which is
ox
the LUMO of (L ), the SOMO for (L•calcd)1-, and the
calcd
HOMO for H2Lcalcd. This orbital is antibonding between the
two hydrazine nitrogens, N1-N2, but bonding between
N2-C3 and antibonding between C3-N4. Thus, upon reduc-
tion from Lox to H2[Lcalcd] the N1-N2 and C3-N4 bond
lengths increase while that of N2-C3 decreases. Gratifyingly,
these calculated bond distances for the radical anion agree
nicely with those determined experimentally in 1, 5, and 6
ox
calcd
and for L
with those in 7, 8, 9, and 10.
1. Calculated Geometries of Square Planar Complexes.
Consistent with the notion that the square planar complexes
[NiII(L•calc)2] and [NiII(L1•)2] (1) contain two radical ligands,
the ground-state geometries for these two neutral molecules
were optimized in the high-spin state (S ) 1) and in the
corresponding broken symmetry (BS) diamagnetic Ms ) 0
state.
Using the BP86 functional, we found that the BS state
does not exist; all calculations converged to a closed shell
state. In contrast, using the B3LYP functional, the BS
solutions exist and they are lower in energy than the
corresponding high-spin states. Such discrepancy between
the two functionals has previously been noticed.20 However,
although describing different electronic structures, the bond
The dihedral angle between the two NCNNNi metallacycles
3650 Inorganic Chemistry, Vol. 44, No. 10, 2005

Diamagnetic Complexes of Nickel(II)
overlap within the spin-coupled pair, the stronger the
electronic coupling between the spin-carrying fragments. The
mutual overlap of these two ligand-centered orbitals is 0.40,
which correlates with a strong antiferromagnetic interaction
between the two organic radicals. This exchange interaction
parameter is best estimated through Yamaguchi’s formula
(eq 4,28 Hˆ HDvV ) - 2JSˆASˆB) and is calculated to be ∼ -1660
cm-1.
The electronic structure of the [Ni(L1•)2] complex is
essentially similar to that of the truncated model, with an
overlap of the two ligand-centered π-orbitals of 0.63 leading
to an antiferromagnetic interaction between the two ligand
radicals of about -1800 cm-1.
EHS - EBS
S2 HS - S2
J ) -
(4)
BS
Note that the spin-expectation value is directly related to
the corresponding orbital overlap eq 5:
NR - Nâ NR - Nâ
S2
)
+ 1 + Nâ - nRnâ|SRâ|2
∑
BS
i
i
ii
(
)(
)
2
2
i
(5)
Figure 14. Optimized bond distances (Å) for square planar [Ni(L•calcd)2]
and qualitative MO diagram derived from DFT calculations at the B3LYP/
TZV(P) level.
where NR,â are the number of spin-up and spin-down
electrons, niR,â are the individual spin-up and spin-down
occupation numbers, and the sum over i extends over the
corresponding orbital pairs with SiiR,â being the spatial overlap
in [Ni(L1•)2] increases slightly upon reduction, whereas in
[NiII(Lcalc)2]- these two metallacycles remain to be coplanar.
2. Electronic Structure of the Square Planar Com-
plexes. The electronic structure of the SP nickel complexes
is very similar to that of other Ni(II) square planar complexes
containing two ligand radicals.20 In the case of the truncated
model, the overall geometry possesses C2h symmetry, which
will be used for the nomenclature of the orbitals, the z axis
being chosen perpendicular to the plane of the molecule, the
x axis being along the Ni-N1 direction (Figure 14). A
qualitative MO diagram, obtained from the BS B3LYP
calculation, is shown in Figure 14. Four doubly occupied
MOs are mainly centered on the nickel ion, while the empty
LUMO+1 orbital is a Ni dxy orbital. Thus, in this complex,
the nickel ion is best described as a low-spin d8 system. The
HOMO and the LUMO orbitals correspond to the symmetric
and antisymmetric combinations of the SOMO orbitals of
the free π-radical anions (L•calc)- ligand, respectively. The
LUMO displays ∼8% Ni-3d character.
R,â
integral for the ith pair. In the strong interaction limit Sii
f 1 the BS solution represents a pure spin state. In the weak
interaction limit SiiR,â f 0 and the BS state represents a mixed
spin state that requires a multideterminant wave function in
a more concise electronic structure description (see below).
However, eq 4 involving the spin expectation values neatly
takes care of the two interaction extremes and has been
shown to be a reasonable approximation over the whole
domain of interaction strengths.28 The strength of the
R,â 2
antiferromagnetic interaction is reflected in the value |Sii | .
3. SORCI Calculations of the Truncated Square Planar
Model. The relatively small size of the truncated square
planar model allows the application of a correlated ab initio
method for a more rigorous description of the multiconfigu-
rational ground state of diradical complexes. In the present
work, we have used the recently developed SORCI method.18b
This method can be thought of as a simplified version of
the multireference difference dedicated configuration interac-
tion (DDCI3) approach of Malrieu, Caballol, and co-
workers,29 which has shown excellent accuracy in the
application to magnetically coupled transition metal dimers.30
For details of the methodology and a discussion about the
necessary active space, we refer to our previous work.20 In
the present SORCI calculations, we have used a slightly
For the analysis of the spin coupling, it is advantageous
to change to an alternative set of MOs, i.e., the corresponding
orbitals.27 This set of orbitals shows most clearly the involved
spin couplings since it will transform the orbitals in the spin-
up and the spin-down sets such that the most similar possible
pairs result. In practice, this means that one finds spin-up
and spin-down pairs with an overlap of essentially unity
(within 10-3-10-4) and one spin-coupled pair with an
overlap between 0 and unity. For the latter, the larger the
(28) (a) Yamaguchi, K.; Takahara, Y.; Fueno, T. In Applied Quantum
Chemistry; Smith, V. H., Eds.; Reidel: Dordrecht, 1986; p 155. (b)
Soda, T.; Kitagawa, Y.; Onishi, T.; Takano, Y.; Shigeta, Y.; Nagao,
H.; Yoshioka, Y.; Yamaguchi, K. Chem. Phys. Lett. 2000, 319, 223.
(29) (a) Miralles, J.; Daudey, J. P.; Caballol, R. Chem. Phys. Lett. 1992,
198, 555. (b) Miralles, J.; Castell, O.; Caballol, R.; Malrieu, J. P. Chem.
Phys. Lett. 1993, 172, 33.
(27) (a) Amos, A. T.; Hall, G. G. Proc. R. Soc. Ser. A 1961, 263, 483. (b)
King, H. F.; Stanton, R. E.; Kim, H.; Wyatt, R. E.; Parr, R. G. J.
Chem. Phys. 1967, 47, 1936. (c) Neese, F. J. Phys. Chem. Solids 2004,
65, 781.
Inorganic Chemistry, Vol. 44, No. 10, 2005 3651

Blanchard et al.
larger than the minimal CAS(2,2) reference space required
in the square planar case to be able to describe the three
competing singlet, triplet, and quintet spin states. This
demanded a CAS(6,5) active space. In agreement with the
experimental results, the SORCI method predicts a singlet
ground state for the square planar geometry and a coupling
constant of JSORCI ) -1353 cm-1 in reasonable agreement
with the DFT results (-1600 cm-1). The diradical character
estimated through eq 4 of ref 20 was found to be ∼33%.
The quintet state was predicted considerably higher at 13380
cm-1 for the square planar geometry.
calculations performed on the truncated model complex [Ni-
(Lcalc )2] with the BP86 functional starting from a tetrahedral
geometry converged back to the square planar closed shell
state obtained before. Thus, these calculations will not be
further discussed.
In contrast, all the B3LYP calculations led to BS Ms ) 0
states lower in energy than the corresponding high-spin S
) 2 states. In all cases, the calculated N1-N2, N2-C3 and
C3-N4 bond lengths are in excellent agreement with the
presence of two ligand radicals coordinated to a NiII center
ion. Moreover, the experimental trends in bond length
variations as a function of the geometry are accurately
predicted by these calculations, as shown by the lengthening
of the average Ni-N and N2-C3 bond lengths together with
the shortening of the N1-N2 distances on going from a
square planar geometry to the tetrahedral one.
•
4. Calculations on the Monoanions. For the monoanions
of square planar complexes 1-3, the EPR spectra are very
similar to those of the monoreduced form of other Ni(II)
square planar species containing two ligands in a radical
form:19 they display an unusually large g-anisotropy for
organic centered radicals. On the basis of our previous DFT
calculations,26 this g-anisotropy can be understood by
considering the SOMO of these monoreduced species: upon
reduction, one electron populates the LUMO of the neutral
complexes. Note that in this case, no broken symmetry
The calculated dihedral angle between the two NNCNNi
metallocycles is slightly twisted at 82° in the tetrahedral [Ni-
•
(Lcalc )2] case, where in an ideal tetrahedron this angle would
be 90°. In the case of [Ni(L5•)2], this dihedral angle is
accurately reproduced with a calculated value of 86° as
compared to an experimental one of 85°, while for [Ni(L6•)2],
a larger deviation is observed with a calculated value of 85°
compared to the experimental value of 71°.
1
calculation is necessary and the BP86 and B3LYP S ) /2
open shell calculations give quite similar electronic structures.
The SOMO orbital possesses bg symmetry and is mainly
ligand-centered (π*). However, because of that symmetry,
it can mix with the out-of-plane dxz and dyz orbitals of the
Ni(II) center. Thus, the SOMO possesses significant metal
character leading to a ground state with sizable orbital angular
momentum due to mixing with low lying d-d excited states
of the same symmetry. The latter explains the observed
g-anisotropy for the monoanions of 1-3. Indeed, the
calculated principal values of the g-tensor, 2.077, 2.014, and
1.99 for the reduced square planar [Ni(Lcalc)(Lcalc•)]-, and
2.058, 2.006, and 2.001 for [Ni(L1)(L1•)]-, are in excellent
agreement with the experimental ones (Table 6).
As the SOMO is ligand centered, one could expect to
observe hyperfine interactions of the electronic spin with the
nitrogen nuclear spins. However, no hyperfine was detected
in the experimental spectra. The calculated nitrogen hyperfine
couplings for the truncated [Ni(Lcalc)(Lcalc•)]- model are 30,
2, and 2 MHz for N1, 20, 1, and 1 MHz for N4 , and 4 MHz,
-2 MHz and -2 MHz for N3, along the g-tensor axis. It is
very likely that these weak hyperfine couplings are hidden
in the line width of the transitions.
It is worth noticing that in the case of [Ni(L5•)2] where
we performed a B3LYP geometry optimization starting from
a square planar environment around the nickel ion, the
calculations converged back to a tetrahedral geometry at the
Ni(II) center. Thus, in agreement with the NMR studies
where only one isomer was detected in solution, there is no
minimum for a (distorted) square planar isomer.
•
6. Electronic Structure of Tetrahedral [Ni(Lcalc )2], [Ni-
(L5•)2], [Ni(L6•)2] Complexes. In the tetrahedral cases, the
B3LYP calculations show that the lowest energy spin-
unrestricted solution is of Ms ) 0 BS type. The results are
similar for all three calculations performed. We will therefore
focus on the analysis of the electronic structure of tetrahedral
•
[Ni(Lcalc )2].
A qualitative bonding scheme derived from the B3LYP
BS Ms ) 0 calculation of tetrahedral [Ni(Lcalc•)2] is shown
in Figure 15. Five occupied orbitals of mainly nickel in
character are identified. Three of these orbitals (not shown),
two from the e set and one from the t2 set, are found in the
spin-up as well as the spin-down manifold and are therefore
doubly occupied. The other two nickel-based orbitals origi-
nate from the t2 set and occur only in the spin-up manifold.
These orbitals are thus singly occupied with parallel spins.
This orbital occupation pattern defines a high-spin Ni(II)
configuration (SNi ) 1) at the metal center. In addition to
these two metal-centered orbitals, two ligand-centered orbit-
als are identified in the spin-down manifold that are not
populated in the spin-up manifold, thus leading to the
observed overall MS ) 0 BS state. These two orbitals
correspond mainly to two symmetry-adapted combinations
of the SOMO of the two ligand radicals. Thus, the basic
electronic structure description features a high-spin Ni(II)
ion that is strongly antiferromagnetically coupled to two
5. Optimized Geometries of the Neutral Tetrahedral
Species. Assuming that the tetrahedral complexes are featur-
ing a high-spin Ni(II) ion and two ligand radicals led us to
optimize the ground-state geometry for both high-spin S )
2 and BS Ms ) 0 states.
Again, in the case of the BP86 functional no BS solutions
exist for the calculations on [Ni(L5•)2] and [Ni(L6•)2]. All
calculations converge to closed shell states. Moreover, the
(30) (a) Calzado, C. J.; Sanz, J. F.; Malrieu, J. P.; Ilas, F. Chem. Phys.
Lett. 1999, 307, 102. (b) Calzado, C. J.; Malrieu, J. P. Chem. Phys.
Lett. 2000, 317, 404. (c) Calzado, C. J.; Sanz, J. F.; Malrieu, J. P. J.
Chem. Phys. 2000, 112, 5158. (d) Calzado, C. J.; Cabrero, J.; Malrieu,
J. P.; Caballol, R. J. Chem. Phys. 2002, 116, 2728. (e) Calzado, C. J.;
Cabrero, J.; Malrieu, J. P.; Caballol, R. J. Chem. Phys. 2002, 116,
3985. (f) Castell, O.; Caballol, R. Inorg. Chem. 1999, 38, 668.
3652 Inorganic Chemistry, Vol. 44, No. 10, 2005

Diamagnetic Complexes of Nickel(II)
7. SORCI Calculations on the Truncated Tetrahedral
Model. The SORCI calculations on the tetrahedral species
[NiII(L•calc)] were carried out analogously to the calculations
on the square planar structure. For the tetrahedral species a
CAS(4,4) reference space was used, which is the minimum
needed to describe the first singlet, triplet, and quintet states.
The four orbitals are two metal-based t2 derived MOs and
two ligand orbitals and were generated by a spin-restricted
B3LYP calculation on the quintet state. The calculations are
again in agreement with experiment in predicting a singlet
ground state with the next triplet state lying at 2751 cm-1.
From this, a JSORCI of -1357 cm-1 is obtained which is,
again, in excellent agreement with the B3LYP prediction.
The first quintet state is calculated at 10541 cm-1, which is
slightly larger than predicted by the spin-coupling model for
two S ) 1 units (6JSORCI ) 8142 cm-1).31
Figure 15. Optimized bond distances (Å) of tetrahedral [NiII(L•calcd)2] and
qualitative MO diagram of the corresponding orbitals of magnetic pairs
calculated at the B3LYP/TZV(P) level.
8. Optimized Geometries of the Monoreduced Tetra-
hedral Species. B3LYP DFT calculations of the monore-
duced forms of these tetrahedral species were also performed.
Upon reduction in silico of the tetrahedral [Ni(L•calc)2] model,
no minimum for a monoanion with tetrahedral geometry can
be found. The structure converges back to square planar.
ligand-centered π-radicals as has originally been proposed
by the authors of ref 7.
1
This is also confirmed by the corresponding orbital
analysis, where one finds two pairs of orbitals with an overlap
between 0 and 1. Both spin-coupled (“magnetic”) pairs are
formed between a metal centered t2 orbital and a ligand-
centered orbital of the corresponding symmetry, as depicted
in Figure 15. The mutual overlaps of these two orbitals for
each pair are 0.50 and 0.57, respectively, which indicates in
both cases a strong antiferromagnetic interaction. The
exchange interaction parameter is calculated to be -1450
cm-1.
However, a B3LYP BS Ms ) /2 optimized tetrahedral
geometry was found for the monoreduced species of [Ni-
(L6•)2].
In this case, the tetrahedral geometry is conserved, as
indicated by the dihedral angle between the two NNCNNi
metallacycles of 78° for [Ni(L6)(L6•)]-. Lengthening of the
average N1-N2 and C3-N4 bond distances with the
corresponding shortening of the N2-C3 is indicative of a
ligand-centered reduction. In contrast to the square planar
monoreduced species where the additional charge is fully
delocalized over the two ligands, the one-electron reduction
process appears here to be much more localized on one
ligand: the corresponding C-N and N-N bond distances
of the two ligands in [Ni(L6)(L6•)]- are no longer identical
(see Figure 16). One ligand features a long N1-N2 bond at
1.381 Å, a short N2-C3 bond at 1.302 Å, and a long C3-
N4 at 1.355 Å, which are characteristic of a fully reduced
(L6)2- ligand, whereas the other ligand displays bond lengths
characteristic of the (L6•)- radical form of the ligand.
9. Electronic Structure of the Monoreduced Tetrahe-
dral Species [Ni(L6)(L6•)]-. The lowest energy, spin-
unrestricted solution for the monoreduced tetrahedral species
is a BS Ms ) 1/2 solution. As described for the neutral case,
we identify five occupied orbitals that are mainly metal in
character. Three of these are found in both the spin-up and
In the present case, the result of the corresponding orbital
analysis is that there is a strong antiferromagnetic exchange
• -
interaction between the (Lcalc ) radical ligands and the central
nickel via a π-overlap pathway. The calculated strong
exchange interaction is consistent with the observed ground-
state spin of St ) 0 with no indication of thermal population
of the higher lying S ) 1 or 2 states up to room temperature.
However, since the precise values of J are not known from
experiment, the accuracy of the calculation cannot be
checked.
The calculated energy difference between the square planar
(lowest) and tetrahedral form of [NiII(L•calcd)2] is only ∼4
kcal/mol: obviously, the change in geometry from square
planar to tetrahedral in this diradical containing Ni(II)
complexes is electronically not strongly disfavored. Note that
in this instance steric interactions between the two ligands
are not involved. This is different for [Ni(L5•)2] and [Ni-
(L6•)2] where square planar forms do not exist in silico. The
electronic structures of these two species are very similar to
that of tetrahedral [Ni(L•calcd)2] described above. The overlaps
of 0.49 and 0.47 for 5 and 0.52 and 0.50 for 6 of the two
corresponding orbital pairs lead to calculated coupling
constants, J, of -950 and -1080 cm-1, respectively.
(31) The singlet wave function is quite complicated having 62% contribution
of the closed shell configuration and 11% from double excitations in
the active space, 2% contribution from the configuration with four
open shells and still 2% contribution from the quadruple excitation
within the active space. Thus, the first-order interacting space of this
calculation included up to hextuple excitations relative to the closed
shell configuration, which shows that a multireference treatment is
hardly avoidable in predicting the ground-state properties of such
complicated spin-coupled molecules from correlated ab initio methods.
The only state with a simple configurational structure is the quintet
state. The population analysis of the correlated density shows indeed
1.8 unpaired electrons on the nickel center and two unpaired electrons
delocalized over the ligands, which is in support of the DFT results.
Inorganic Chemistry, Vol. 44, No. 10, 2005 3653

Blanchard et al.
Figure 17. Structure and bond distances (Å) of a tetrahedral nickel(II)
and a square planar platinum(II) complex from refs 8 and 33, respectively.
As it has been discussed previously,32 it is possible to
calculate reasonable g-tensors for antiferromagnetically
coupled dimers from DFT calculations. The main idea of
this approach is to determine the g-tensors corresponding to
the high-spin and broken symmetry states and then to extract
from these calculations hypothetical site-g-tensors. Coupling
the latter by using the usual vector coupling model then
yields a g-tensor reflecting the properties of the antiferrro-
magnetic state. Applying this methodology, we calculated
the g-tensor for [Ni(L6)(L6•)]-, which has principal values
of gx ) 2.037, gy ) 2.198, and gz ) 2.205, which leads to
giso ) 2.147. This tensor displays a large anisotropy, in
1
agreement with a metal-centered S ) /2 species. The giso
Figure 16. Optimized geometry of tetrahedral [NiII(L6)(L6•)]- and
qualitative MO diagram of the magnetic orbitals derived from B3LYP/
TZV(P) calculations.
value is surprisingly close to that of the a subspecies
identified in the spectra of electrochemically generated
[Ni(L6)(L6•)]- but compares also well with that of the b
subspecies (Table 6). Compared to the experimental data for
these two subspecies, the calculated g-tensor is more axial.
spin-down manifolds and are thus doubly occupied, while
two are only populated in the spin-up manifold. Thus, upon
reduction of the complexes the oxidation state of the metal
center did not change: it is still best described as a high-
spin nickel(II) ion (S ) 1). In addition, one finds one ligand-
centered orbital which is only occupied in the spin-down
manifold. This overall occupation pattern leads to the Ms )
1/2 BS state. Thus, the basic electronic description of the
reduced species features a high-spin Ni(II) strongly antifer-
romagnetically coupled to a single localized ligand-centered
π-radical.
However, the dihedral angle between the two NNCNNi
metallacycles is an important parameter for the g-anisotropy
which has not yet been determined experimentally. The
present results may be compatible with our proposal that the
two subspecies a and b observed experimentally correspond
to an equilibrium between two different distorted tetrahedral
isomers.
Discussion
This description is confirmed by the corresponding orbital
analysis of [Ni(L6)(L6•)]- shown in Figure 16. It is obvious
from this figure that the unpaired electron is localized at the
metal center. Moreover, in the magnetic pair, the spin density
on the organic part is localized on one ligand only. In
conjunction with the bond length differences between the
two ligands in the optimized geometries, this clearly indicates
that this monoanionic complex possesses two ligands in
different oxidation states, namely, the fully reduced dianionic
(L6)2- closed shell form and the π-radical monoanionic (L6•)-
form, with little electronic delocalization of the extra electron
from the fully reduced form to the monoreduced one.
Thus, all calculations are in agreement with ligand-
centered reduction processes leading to St ) 1/2 complexes,
the doublet state resulting from the antiferromagnetically
coupling of a high-spin Ni(II) with one ligand radical.
For complexes 1-3 spectroscopic evidence such as an
intense LLCT band, the X-ray structure determination of 1,
and the theoretical calculations clearly establish that the
complexes are diamagnetic square planar nickel(II) com-
plexes with two ligands in their π-radical monoanionic form.
The diamagnetic ground state of these complexes originates
from a strong antiferromagnetic interaction between the two
radical ligands. These species are singlet diradicals as has
been determined for bis(o-phenylenediamine radical)nickel-
(II) complexes.19,20 From correlated ab initio calculations,
we conclude that they have ∼33% diradical character.
(32) (a) Slep, L. D.; Mijovilovich, A.; Meyer-Klaucke, W.; Weyhermu¨ller,
T.; Bill, E.; Bothe, E.; Neese, F.; Wieghardt, K. J. Am. Chem. Soc.
2003, 125, 15554. (b) Sinnecker, S.; Neese, F.; Noodleman, L.; Lubitz,
W. J. Am. Chem. Soc. 2004, 126, 2613.
3654 Inorganic Chemistry, Vol. 44, No. 10, 2005

Diamagnetic Complexes of Nickel(II)
Table 7. Optimized Bond Distances (Å) of Square Planar and Tetrahedral Neutral and Monoanionic Complexes
a
complex
method
av. N1-N2
av. N2-C3
av. C3-N4
av. Ni-N
θ, deg
square planar geometry
[Ni(L•calcd)2]
[NiII(L1•)2]
B3LYP, BS
BP86
B3LYP
exp
B3LYP
BP86
B3LYP
1.324
1.348
1.337
1.343
1.362
1.385
1.379
1.341
1.346
1.337
1.339
1.310
1.318
1.306
1.324
1.336
1.319
1.321
1.349
1.344
1.1.380
1.861
1.862
1.891
1.847
1.864
1.880
1.889
0
22
19
12
0
29.4
22.5
[Ni(Lcalcd)2]1-
[Ni(L1)2]1-
tetrahedral geometry
[Ni(L•calcd)2]
[NiII(L5•)2]
B3LYP, BS
BP86
exp
B3LYP, BS
BP86
exp
1.316
1.342
1.333
1.327
1.342
1.343
1.364
1.341
1.352
1.346
1.344
1.305
1.345
1.325
1.321
1.340
1.328
1.966
1.891
1.904
1.977
1.903
1.891
82
53.5
85
85
65
[NiII(L6•)2]
1.346
71
tetrahedral geometry
1.323
[Ni(L5)2]1-
[Ni(L6)2]1-
BP86
1.370
1.354
1.908
55
B3LYP, BSb
BP86
1.365
1.336
1.381
1.327
1.339
1.302
1.352
1.320
1.355
1.939
2.025
1.952
B3LYP, BSb
a Labeling see Figure 13. b Two inequivalent ligands (L6•)1- and (L6)2-
.
In contrast, complexes 5 and 6 adopt a tetrahedral
geometry. Again, the bond lengths within the ligands clearly
indicate the presence of two π-radical monoanions in each
complex. A similar tetrahedral structure has been deduced
for 4 from spectroscopy. DFT calculations performed on
these tetrahedral complexes clearly establish that the metal
ion is nickel(II) in a high-spin configuration (S ) 1) and
that the diamagnetism of these three molecules results from
a strong antiferromagnetic coupling with two ligand radicals.
Electronic spectroscopic studies on the reduced monoan-
ionic species [1]-, [2]-, and [3]- indicate that in these
complexes the square planar NiIIN4 unit is retained. An
intense LLCT band at 1100-1400 nm is indicative of
complete delocalization of the excess electron over both
ligands.
In the dianionic species [1]2-, [2]2-, and [3]2- the above
SOMO becomes a ligand-centered HOMO. Thus, the square
planar neutral, monoanionic, and dianionic forms of com-
plexes 1, 2, and 3 contain each a diamagnetic nickel(II) ion
(d,8 S ) 0) in a square planar ligand field and (a) two
monoanionic π-radicals and (b) a class III mixed valent
ligand system [(L1-3•)1-M(L1-3)2-]1- T [(L1-3)2-M(L1-3•)1-]1-,
and (c) two closed-shell dianionic ligands, respectively.
In stark contrast, complexes 4, 5, and 6 containing 1,4-
diphenyl derivatives of the S-methyl-isothiosemicarbazides
are distorted tetrahedral. The electronic structures of these
neutral complexes must be described as species with a central
high-spin nickel(II) ion (d,8 S ) 1) and two N,N-coordinated
π-radical monoanions. The radical spins couple strongly
intramolecularly in an antiferromagnetic fashion with the two
t2 metal d orbitals each containing an unpaired electron (S
) 0). DFT calculations have shown that the energetic
difference between a tetrahedral and a square planar ar-
rangement of the ligands is small (∼4 kcal mol-1) if no steric
hindrance comes into play.
g-anisotropy than that observed for the square planar
analogues [1]-, [2]-, and [3]-. In addition, there are always
two subspectra observed for [4]-, [5]-, and [6]-, which led
us to propose two tetrahedral geometric isomers A and B
(Scheme 5) for each of these species. The electronic spectra
of the tetrahedral monoanions lack the intense MVCT band
> 1100 nm observed for the square planar monoanions. This
observation and the results of the DFT calculations allow
us to conclude that the excess electron is localized on one
ligand monoanion. Thus, their electronic structure is best
described as tetrahedral [NiII(L4-6•)(L4-6)]- where a high-
spin Ni(II) ion (d,8 S ) 1) couples antiferromagnetically with
a ligand π-radical. The structure of the dianions [Ni(L4-6)2]2-
remains unclear due to the lack of a crystal structure
determination.
Revenco and Gerbeleu et al. had shown1-5 that oxidation
of 1 with iodine affords 7 an octahedral complex of NiII with
two neutral ligands (L1,ox) and two in trans-position relative
to each other coordination iodide ligands. We have discov-
ered that the analogous oxidation of tetrahedral 5 yields
complex 9, which again contains two oxidized ligands (L5,ox
)
and two in trans-position coordinated (I3)- ligands. Appar-
ently, the two organic ligands in the NiN4 plane do not
experience steric hindrance (repulsion). To the contrary,
intramolecular π-stacking between the ligands is observed.
This experimental result seems at first glance to contradict
the notion that the tetrahedral NiN4 polyhedron in complexes
4-6 is the result of steric hindrance between two 1,4-
diphenyl-isothiosemicarbazide derivatives in a planar NiN4
coordination mode. DFT calculations have shown that in
sterically unhindered 1-phenyl systems (1-3) the planar
geometry is more stable by 4 ( 3 kcal mol-1 than the
tetrahedral. Thus, the energetic difference based on electronic
effects alone between the two geometries is small. It is
interesting to note in this context that the structure of bis-
(N,N-diethylphenylazothioformamide)nickel(II)8 is tetrahe-
dral, whereas the very similar platinum(II) complex bis-
Interestingly, the paramagnetic monoanions [4]-, [5]-, and
[6]- (S ) 1/2) display EPR spectra with a significantly larger
Inorganic Chemistry, Vol. 44, No. 10, 2005 3655

Blanchard et al.
(methyl-N′-phenyldiazenecarbodithioato)platinum(II) shown
in Figure 17 is planar.33 Both complexes have the same five-
square planar [NiII(L5•)2] may become as short as ∼2.8 Å
where this interaction may be repulsive rather than attractive.
membered chelate rings M-S-C-N-N. Here, of course
the larger crystal field stabilization energy of a diamagnetic
Pt(II) ion (d,8 S ) 0) favors the square planar arrangement
over a tetrahedral one. As pointed out previously, the
π-stacking interaction between the phenyl rings in a putative
Acknowledgment. This work was financially supported
by the Fonds der Chemischen Industrie.
Supporting Information Available: Figures S1-S7 and X-ray
crystallographic files for compounds 1 and 5-11. This material is
available free of charge via the Internet at http://pubs.acs.org.
(33) Dessy, G.; Fares, V. Acta Crystallogr. 1980, B36, 2266.
IC040117E
3656 Inorganic Chemistry, Vol. 44, No. 10, 2005