Paramagnetic NMR investigations of high-spin nickel(II) complexes. Controlled synthesis, structural, electronic, and magnetic properties of dinuclear vs mononuclear species

Article abstract of DOI:10.1021/ja010342k

New dissymmetric tertiary amines (N₃SR) with varying N/S donor sets have been synthesized to provide mono- and dinuclear complexes. Acetate ions are used to complete the octahedral coordination sphere around nickel(II) atom(s). The facile conversion of mononuclear to dinuclear systems can be controlled to produce either mono- or dinuclear complexes from the same ligand. The dinuclear complex a(BPh₄)₂ ([Ni₂(N₃SSN₃)(OAc)₂] (BPh₄)₂) has been characterized in the solid state by X-ray diffraction techniques as solvate: a(BPh₄)₂·1/2[5(CH₃OH)· (CH₃CN)·(CH₃CH₂OH)]. The two Ni atoms are six-coordinated and bridged by a disulfide group and two bidentate acetates. Magnetic susceptibility reveals a weak ferromagnetic exchange interaction between the two Ni atoms with J = 2.5(7) cm^-1. UV-vis studies suggest that the six coordinated structure persists in solution. The ¹H NMR spectrum of a(BPh₄)₂ exhibits sharp significantly hyperfine shifted ligand signals. A complete assignment of resonances is accomplished by a combination of methods: 2D-COSY experiments, selective chemical substitution, and analysis of proton relaxation data. Proton isotropic hyperfine shifts are shown to originate mainly from contact interactions and to intrinsically contain a small J-magnetic coupling and/or zero-field splitting contribution. A temperature dependence study of longitudinal relaxation times indicates that a very unusual paramagnetic Curie dipolar mechanism is the dominant relaxation pathway in these weakly ferromagnetically spin-coupled dinickel(II) centers. The mononuclear nickel(II) analogue exhibits extremely broader ¹H NMR signals and only partial analysis could be performed. These data are consistent with a shortening of electronic relaxation times in homodinuclear compounds with respect to the corresponding mononuclear species.

Full text of DOI:10.1021/ja010342k

J. Am. Chem. Soc. 2001, 123, 8053-8066
8053
Paramagnetic NMR Investigations of High-Spin Nickel(II)
Complexes. Controlled Synthesis, Structural, Electronic, and Magnetic
Properties of Dinuclear vs Mononuclear Species
Catherine Belle,*,† Catherine Bougault,*,† Marie-The´re`se Averbuch,† Andre´ Durif,†
Jean-Louis Pierre,† Jean-Marc Latour,‡ and Laurent Le Pape‡
Contribution from the Laboratoire d’Etudes Dynamiques et Structurales de la Se´lectiVite´ (LEDSS, UMR
CNRS 5616), UniVersite´ Joseph Fourier, BP 53X, 38041 Grenoble Cedex 9, France, and the SerVice de
Chimie Inorganique et Biologique (SCIB, UMR CNRS 5046), DRFMC/CEA-Grenoble, 17 rue des Martyrs,
38054 Grenoble Cedex 9, France
ReceiVed February 7, 2001. ReVised Manuscript ReceiVed May 1, 2001
Abstract: New dissymmetric tertiary amines (N3SR) with varying N/S donor sets have been synthesized to
provide mono- and dinuclear complexes. Acetate ions are used to complete the octahedral coordination sphere
around nickel(II) atom(s). The facile conversion of mononuclear to dinuclear systems can be controlled to
produce either mono- or dinuclear complexes from the same ligand. The dinuclear complex a(BPh4)2 ([Ni2(N3-
SSN3)(OAc)2](BPh4)2) has been characterized in the solid state by X-ray diffraction techniques as solvate:
a(BPh4)2‚1/2[5(CH3OH)‚(CH3CN)‚(CH3CH2OH)]. The two Ni atoms are six-coordinated and bridged by a
disulfide group and two bidentate acetates. Magnetic susceptibility reveals a weak ferromagnetic exchange
interaction between the two Ni atoms with J ) 2.5(7) cm-1. UV-vis studies suggest that the six-coordinated
structure persists in solution. The 1H NMR spectrum of a(BPh4)2 exhibits sharp significantly hyperfine shifted
ligand signals. A complete assignment of resonances is accomplished by a combination of methods: 2D-
COSY experiments, selective chemical substitution, and analysis of proton relaxation data. Proton isotropic
hyperfine shifts are shown to originate mainly from contact interactions and to intrinsically contain a small
J-magnetic coupling and/or zero-field splitting contribution. A temperature dependence study of longitudinal
relaxation times indicates that a very unusual paramagnetic Curie dipolar mechanism is the dominant relaxation
pathway in these weakly ferromagnetically spin-coupled dinickel(II) centers. The mononuclear nickel(II)
1
analogue exhibits extremely broader H NMR signals and only partial analysis could be performed. These
data are consistent with a shortening of electronic relaxation times in homodinuclear compounds with respect
to the corresponding mononuclear species.
Introduction
center, it is of fundamental interest to understand the mecha-
nisms, which control substrate conversion or inhibition pathways
of these enzymes. In that respect, NMR techniques can provide
a wealth of unique information on the electronic, magnetic, and
molecular properties of the metal-binding sites, and thus allow
extraction of crucial mechanistic data.8,9
Contact and dipolar interactions between the considered
nucleus and the metal induce hyperfine shifts of NMR signals
and shortening of nuclear relaxation times. High-resolution
NMR can nevertheless be successfully performed and data can
be interpreted in cases where electronic relaxation times of the
metal are short enough, usually 10-11-10-12 s, as in low-spin
Growing interest, in the past decade, has been directed toward
nickel as a biologically and spectroscopically relevant metal.
To date, natural occurrence of nickel(II) has been evidenced in
six types of enzymes: ureases, where it catalyzes the hydrolysis
of urea, and hydrogenases, methyl coenzyme M reductase,
carbon monoxide dehydrogenase, acetyl coenzyme A synthase,
and nickel superoxide dismutase, where it is known to present
redox activity.1 A detailed picture of the active site of two of
these enzymes, [Ni-Fe] hydrogenase2,3 and urease,4-7 has
emerged through recent X-ray crystallographic studies. In
addition to elucidation of the coordination structure at the metal
* To whom correspondence should be addressed. E-mail: Catherine.
Belle@ujf-grenoble.fr. Catherine.Bougault@ujf-grenoble.fr. FAX: (33) 4
76 51 48 36.
(3) Davidson, G.; Choudhury, S. B.; Gu, Z. J.; Bose, K.; Roseboom,
W.; Albracht, S. P. J.; Maroney, M. J. Biochemistry 2000, 39, 7468-7479
and references therein.
† LEDSS: Laboratoire d’Etudes Dynamiques et Structurales de la
Se´lectivite´.
(4) Jabri, E.; Carr, M. B.; Hausinger, R. P.; Karplus, P. A. Science 1995,
268, 998-1004.
‡ SCIB: Service de Chimie Inorganique et Biologique.
(1) For recent reviews, see: (a) Cammack, R.; van Vliet, P. In Bioinor-
ganic Catalysis; Reedjik, J., Bouwman, E., Eds.; Marcel Dekker: New York,
1999; pp 231-268. (b) Ermler, U.; Grabarse, W.; Shima, S.; Goubeaud,
M.; Thauer, R. K. Curr. Opin. Struct. Biol. 1998, 8, 749-758. (c) Maroney,
M. J.; Davidson, G.; Allan, C. B.; Figlar, J. Struct. Bonding 1998, 92, 1-65.
(d) Ragsdale, S. W. Curr. Opin Chem. Biol. 1998, 2, 208-215.
(2) Volbeda, A.; Garcin, E.; Piras, C.; de Lacey, A. L.; Fernandez, V.
M.; Hatchikian, E. C.; Frey, M.; Fontecilla-Camps, J. C. J. Am. Chem. Soc.
1996, 118, 12989-12996.
(5) Pearson, M. A.; Michel, L. O.; Hausinger, R. P.; Karplus, P. A.
Biochemistry 1997, 36, 8164-8172.
(6) Benini, S.; Rypnievski, W. R.; Wilson, K. S.; Miletti, S.; Ciurli, S.;
Mangani, S. Structure 1999, 7, 205-216.
(7) Benini, S.; Rypniewski, W. R.; Wilson, K. S.; Ciurli, S.; Mangani,
S. J. Biol. Inorg. Chem. 1998, 3, 268-273.
(8) Bertini, I.; Turano, P.; Vial, A. Chem. ReV. 1993, 93, 2833-2932.
(9) See: Biological Magnetic Resonance, NMR of Paramagnetic Mol-
ecules Proteins; Berliner, L. J., Reuben, J., Eds.; Plenum Press: New York,
1993; Vol. 12.
10.1021/ja010342k CCC: $20.00 © 2001 American Chemical Society
Published on Web 07/28/2001

8054 J. Am. Chem. Soc., Vol. 123, No. 33, 2001
Belle et al.
FeIII 10 and high-spin CoII samples.8 Electronic relaxation times
for high-spin nickel(II) are typically 10-10-10-12 s. In the most
favorable case (i.e, pseudotetrahedral or high-spin pentacoor-
dinated nickel(II)), reasonably sharp and well-resolved NMR
signals are observed and structural information can be re-
trieved.11,12 For instance, NiII-substitution of ZnII, CuII, and FeII
was advantageously introduced for NMR studies of a variety
of metalloproteins.13-17 In octahedral nickel(II) monomers,
electronic relaxation times are usually longer and observed
resonances are consequently broader.18,19 When more than one
paramagnetic center is present in the same molecule, the
situation may however be completely different. Recent NMR
studies on copper(II) homodimers showed in fact that nuclear
relaxation enhancement and relatively sharp signals may be
observed in dinuclear species, while mononuclear analogues
gave rise to broad 1H NMR resonances. Studies on model
complexes lead to suspecting the spin-coupling interactions
between paramagnetic centers are responsible for the more
efficient overall electronic relaxation and the subsequent sharp-
ness of NMR signals.20,21 The real origin of this behavior is
nevertheless far from clear.22,23 Systematic studies on a large
variety of homodimers with an extended range of J-values are
then necessary. To date, NMR properties of compounds
containing octahedral nickel(II) remain largely unexplored and,
Scheme 1
implications of the J spin-coupling constant and intrinsic zero-
field splitting of each nickel(II) on the observed enhanced
nuclear relaxation are discussed. The present approach demon-
1
strates that H NMR is a viable tool for studying octahedral
nickel(II) dimeric species in solution and may be relevant when
applied to wild-type or engineered dinuclear nickel metallo-
proteins.
Experimental Section
General Methods. Commercial reagents were used as obtained
without further purification. Solvents were purified by standard methods
before use. Elemental analyses were performed by the CNRS Mi-
croanalysis Laboratory of Lyon, France. Mass spectra were recorded
on a Nermag R 1010C apparatus. Fast-atom bombardment (FAB) in
the positive mode were obtained with the spectrometer equipped with
a M scan (Wallis) atom gun (8 kV, 20 mA). IR (KBr disks) spectra
were recorded on a Nicolet 550 FT-IR spectrometer. UV-visible spectra
were obtained using a Perkin-Elmer Lambda 2 spectrophotometer
operating in the range 200-1100 nm with quartz cells. Extinction
1
to our knowledge, no H NMR data have been reported on
dimeric species, while both ferromagnetic and antiferromagnetic
(large and weak) J-coupling constants have been reported.
1
Correlation of the structural, magnetic, and H NMR features
may then provide a fundamental insight into the description and
understanding of the relaxation mechanisms in paramagnetic
clusters.
coefficients ꢀ are given in M-1‚cm-1 1H NMR spectra of organic
.
compounds were recorded on a BRUKER AC 200 spectrometer, at 25
°C. Chemical shifts in ppm were referenced to residual solvent peaks.
Syntheses: Ligand Syntheses. Experimental details relative to the
syntheses of 6, 7, and 9 and to physicochemical properties of
compounds 1 to 4 and 6 to 9 (elemental analysis, 1H and 13C NMR, as
well as mass spectrometry data) are provided as Supporting Information.
1: Under an inert atmosphere, a solution of thiosalicylic acid (1 g,
6.29 mmol) in 20 mL of dry THF was added to a suspension, in dry
THF (5 mL), of NaH (60% in oil, 500 mg, 12.58 mmol), previously
washed with dry pentane. The mixture was cooled with an ice bath
during the addition. The suspension was stirred subsequently for 1 h
at room temperature. A solution of 3-bromopropionitrile in THF (20
mL) was then added dropwise, and the mixture was heated and refluxed
for 12 h under vigorous stirring. After the mixture was cooled, the
THF was removed in vacuo. The residue was melted in water and 4 N
HCl was added to the mixture up to precipitation of a white solid. A
small amount of dichloromethane was added to this suspension and
the product was isolated by filtration, washed with pentane, and dried
under vacuum. 1 was obtained as a white powder (1.24 g, 5.99 mmol)
with 95% yield and used in the following steps without further
purification. An analytical sample was obtained after recrystallization
from dichloromethane/acetone: mp 185 °C.
2: A solution of ethyl chloroformate (0.25 mL, 2.61 mmol) in
distilled THF (5 mL) was added at -5 °C to a mixture of acid 1 (0.5
g, 2.40 mmol), NEt3 (0.3 mL, 2.16 mmol), and distilled THF (15 mL).
The mixture was allowed to stir for 1 h at -5 °C. After filtration, the
filtrate was treated with an excess of NaBH4 in H2O (10 mL) at 0 °C.
The mixture was allowed to warm to room temperature and was stirred
for 12 additional hours. The suspension was further acidified with 4 N
HCl to pH 1, and the solvent was removed under reduced pressure.
The residue was taken up with water and extracted three times with
ethyl acetate. The organic extracts were neutralized with a solution of
NaHCO3 and washed with H2O and brine. The obtained organic layer
was dried over Na2SO4 and reduced in vacuo to yield the pure product
2 as a pale yellow oil (396 mg, yield 85%) which solidified upon
standing at -20 °C.
In the present study, synthesis of nonsymmetric tertiary
amines (N3SR, Scheme 1) and preparation of mono- and
dinuclear octahedral nickel(II) complexes are extensively de-
scribed. One of the latter, the a(BPh4)2 complex ([Ni2(N3SSN3)-
(OAc)2](BPh4)2), is fully characterized. Assignment of proton
resonances can be completely achieved using 2D NMR meth-
odology and T1 relaxation measurements with the help of X-ray
crystal data and magnetic susceptibility results. The possible
(10) For a review, see: La Mar, G. N.; de Ropp, J. S. NMR Methodology
for Paramagnetic Proteins; pp 1-78 in ref 9.
(11) Sacconi, L.; Mani, F.; Bencini, A. In ComprehensiVe Coordination
Chemistry; Wilkinson, S. R., Gillard, R. D., McCleverty, J. A., Eds.;
Pergamon Press: Oxford, 1987; Vol. 5, pp 56-60.
(12) Baidya, N.; Olmstead, M. M.; Whitehead, J. P.; Bagyinka, C.;
Maroney, M. J.; Mascharak, P. K. Inorg. Chem. 1992, 31, 3612-3619.
(13) Moratal Mascarell, J. M.; Martinez Ferrer, M.-J.; Jimenez, H. R.;
Donaire, A.; Castells, J.; Salgado, J. J. Inorg. Biochem. 1992, 45, 231-
243.
(14) (a) Donaire, A.; Salgado, J.; Moratal, J.-M. Biochemistry 1998, 37,
8659-8673. (b) Salgado, J.; Kalverda, A. P.; Diederix, R. E. M.; Canters,
G. W.; Moratal, J. M.; Lawler, A. T.; Dennison, C. J. Biol. Inorg. Chem
1999, 4, 457-467.
(15) Fernandez, C. O.; Sannazzaro, A. I.; Diaz, L. E.; Vila, A. J. Inorg.
Chim. Acta 1998, 273, 367-371.
(16) Moura, I.; Teixeira, M.; LeGall, J.; Moura, J. J. G. J. Inorg. Biochem.
1991, 44, 127-139.
(17) Luchinat, C.; Ciurli, S. NMR of Polymetallic Systems in Proteins;
pp 357-420 in ref 9.
(18) Rosenfield, S. G.; Berends, H. P.; Gelmini, L.; Stephan, D. W.;
Mascharak, P. K. Inorg. Chem. 1987, 26, 2792-2797.
(19) Dietz, C.; Heinemann, F. W.; Kuhnigk, J.; Kru¨ger, C.; Gerdan, M.;
Trautwein, A. X.; Grohmann, A. Eur. J. Inorg. Chem. 1998, 1041-1049.
(20) Brink, J. M.; Rose, R. A.; Holz, R. C. Inorg. Chem. 1996, 35, 2878-
2885.
(21) Asokan, A.; Varghese, B.; Manoharan, P. T. Inorg. Chem. 1999,
38, 4393-4399.
(22) Murthy, N. N.; Karlin, K. D.; Bertini, I.; Luchinat, C. J. Am. Chem.
Soc. 1997, 119, 2156-2162.
3: The alcohol 2 was added (1 g, 5.18 mmol) to a suspension of
MnO2 (5.3 g, 51.80 mmol) in dry dichloromethane (50 mL). The
(23) Holz, R. C.; Bennett, B.; Chen, G.; Ming, L.-J. J. Am. Chem. Soc.
1998, 120, 6329-6335.

Paramagnetic NMR InVestigations of High-Spin Ni(II) Complexes
J. Am. Chem. Soc., Vol. 123, No. 33, 2001 8055
mixture was heated and refluxed for 2 h. After cooling, filtration over
Celite, and evaporation in vacuo, the crude product was isolated as a
yellow solid. Recrystallization in diethyl ether/hexane afforded 0.95 g
(97%) of white solid 3: mp 64 °C.
argon. The mixture was stirred for 2 h and acidified with 4 N HCl.
The resulting solution was evaporated in vacuo, and ethanol (30 mL)
was added to the residue. After 12 h at 0 °C, the suspension was filtered.
The orange filtrate was evaporated in vacuo to provide the solid
trihydrochloride salt of the deprotected thiolate from 5 (310 mg). The
corresponding free base was recovered by neutralizing the crude product
to pH 6 with NaHCO3 in aqueous solution, followed by extraction with
CH2Cl2. The organic extracts were combined and dried over anhydrous
Na2SO4. Solvent removal in vacuo afforded 200 mg of yellow oil 10.25
1H NMR (CDCl3) δ 8.49 (d, 2H, J ) 4.2 Hz, Py-H1), 7.60 (m, 5H,
Ar-H), 7.35 (m, 1H, Ar-H), 7.13 (m, 4H, Ar-H), 3.86 (s, 2H, NCH2-
ArSH), 3.80 (s, 4H, NCH2Py), 1.26 (s br, -SH); 13C NMR (CDCl3) δ
158.9, 148.7, 148.4, 137.7, 137.6, 129.9, 128.0, 127.7, 126.3, 123.5,
122.0, 59.5, 56.7; mass spectrum (CI, NH3), m/z 322 [M + H]+. To an
argon-purged solution of 10 (50 mg, 0.156 mmol) in methanol (3 mL)
was added Ni(OAc)2‚4H2O (40 mg, 0.158 mmol) in methanol (3 mL).
The green solution was stirred for 1 h and a solution of NaBPh4 (106
mg, 0.310 mmol) in methanol (6 mL) was added to induce precipitation
of the product as a powder. Filtration and recrystallization by vapor
diffusion of ethanol into an acetonitrile solution of this solid afforded
crystalline material, which was filtered and dried in vacuo. a(BPh4)2
(65.7 mg) was isolated with 43% overall yield. This compound was
shown to be identical (UV-vis, IR, 1H NMR, mass spectra) to a(BPh4)2
obtained in method A.
4: A 2.00 g sample of 3 (10 mmol) was dissolved in 125 mL of
MeOH and 0.91 mL of 2-(aminomethyl)pyridine (8.73 mmol) was
added to the reaction vessel. The solution was refluxed and stirred for
2 h. After the mixture was cooled, the resulting Schiff’s base was
reduced with NaBH4 (0.85 g, 21 mmol) and the mixture was stirred
for 12 h. Then, 4 N HCl was added to the ice-cold suspension and the
solvent was removed under reduced pressure. The residue was taken
up with water and extracted with dichloromethane. Then the aqueous
extract was neutralized with a NaHCO3 solution and extracted three
times with portions of dichloromethane. The combined organic extracts
were dried over sodium sulfate. The solvent was removed under reduced
pressure affording the yellow oil 4 (2.4 g, 8.54 mmol, 98% yield). The
product was used without any purification. Compound 8 was prepared
with 51% overall yield, according to this procedure. Compound 7 was
obtained from 2-formaldoxime-6-methylpyridine, using modifications
of a previously reported procedure.24
5: A solution of 4 (1.08 g, 3.83 mmol) in dry methanol (50 mL)
containing 1.1 mL of glacial acetic acid was added to a solution of
pyridine-2-carboxaldehyde (442 µL, 4.6 mmol) in 50 mL of dry
methanol. NaBH3CN (1 g, 15.32 mmol) was added to the reaction
mixture under a nitrogen atmosphere, and the suspension was stirred
for 3 days. After this period, the pH of the ice-cold reaction mixture
was adjusted to 2.0 by careful addition of concentrated HCl, and the
mixture was stirred for an additional hour. The solution was reduced
to dryness under reduced pressure, and the resulting residue was
redissolved in 30 mL of water. The aqueous solution was washed two
times with dichloromethane, and the pH of the solution was then
adjusted to about 8 by addition of NaHCO3. The resulting slightly basic
solution was then extracted four times with dichloromethane. The
extracts were combined, washed with brine, and dried over anhydrous
Na2SO4. The solvent was removed in vacuo to give 1.23 g of crude
product as a brown oil. The later was purified by chromatography on
neutral alumina and eluted with ethyl acetate/cyclohexane (2:1,v /v) to
Preparation of Mononuclear Complex m(BPh4). The PF6 salt of
this complex was prepared according to an essentially similar procedure
(see Supporting Information for details).
m(BPh4). To a degassed solution of cyanoethyl-protected thiolate 5
(105 mg, 0.28 mmol) in methanol (5 mL) was added Ni(OAc)2‚4H2O
(70 mg, 0.28 mmol) in methanol (7 mL). After the mixture was stirred
for 1 h in argon atmosphere, a solution of NaBPh4 (96 mg, 0.28 mmol)
in degassed methanol (5 mL) was added. A gray precipitate formed
instantly. The suspension was cooled at -20 °C for 12 h and the
precipitate was removed by filtration. The filtrate was concentrated
and cooled to -20 °C and after 1 week a violet powder was recovered
by filtration, which after drying in vacuo afforded 119 mg (52% yield)
of m(BPh4). This material was stable both in the solid state at room
temperature and in acetone solution at -20 °C for months. Anal. Calcd
for C48H44N4O2SNiB: C, 71.13; H, 5.47; N, 6.91; S, 3.96; Ni, 7.24; B,
1.33. Found: C, 70.20; H, 5.53; N, 6.90; S, 3.97; Ni, 6.53; B, 1.30.
Mass spectrum (FAB, matrix NBA), m/z 491 [M - BPh4]+; IR (KBr,
cm-1) 3047.5, 2975, 1735, 2252, 1607, 1540 (νas(COO)), 1487.5, 1426
(νs(COO)), 1240, 735; UV-vis (acetone, λmax, nm (ꢀ, M-1‚cm-1)) 334
(148), 550 (10), 941 (16).
1
give 1.07 g (75% yield) of oil 5 (2.87 mmol). H NMR (CDCl3) δ
8.51 (d, 2H, J ) 4.1 Hz, Py-H1), 7.60 (m, 4H, Ar-H), 7.2 (m, 6H,
Ar-H), 3.98 (s, 2H, NCH2ArSR), 3.94 (s, 4H, NCH2Py), 3.06 (t, 2H, J
) 7.2 Hz, -SCH2-), 2.54 (t, 2H, J ) 7.2 Hz, -CH2CN); 13C NMR
(CDCl3) δ 159.3, 148.7, 137.7, 136.1, 132.4, 131.2, 129.6, 127.7, 126.7,
122.0, 121.6, 117.8, 54.3, 51.3, 29.7, 17.8; mass spectrum (EI), m/z
375 [M]+. Anal. Calcd for C22H22N4S: C, 70.56; H, 5.92; N, 14.96; S,
8.56. Found: C, 70.66; H, 5.94; N, 14.91; S, 8.58. Compounds 6 and
9 were prepared according to this procedure.
Preparation of Dinuclear Complex a(BPh4)2. Syntheses (according
to method A) and physicochemical properties (elemental analysis, UV-
vis, IR, as well as mass spectrometry data) of PF6 salts of the dinuclear
complexes, a(PF6)2, b(PF6)2, and c(PF6)2, are provided as Supporting
Information.
Method A. To a solution of 5 (110 mg, 0.28 mmol) in methanol (2
mL) was added a solution of Ni(OAc)2‚4H2O (93 mg, 0.37 mmol) in
methanol (7 mL). The blue solution was stirred at room temperature
and progressively turned green over 1 day. After an additional day, a
solution of NaBPh4 (116 mg, 0.34 mmol) in methanol (3 mL) was added
to precipitate the crude product as a powder. Filtration and recrystal-
lization by vapor diffusion of ethanol into an acetonitrile solution of
this solid afforded pale-green crystals of X-ray quality. This crystalline
material was dried in vacuo and afforded a(BPh4)2 (68.7 mg, yield
32%). Anal. Calcd for C90H82N6O4S2Ni2B2: C, 71.27; H, 5.58; N, 5.54;
S, 4.23; Ni, 7.74; B, 1.43. Found: C, 71.13; H, 5.41; N, 5.46; S, 4.02;
Ni, 7.65; B, 1.42. Mass spectrum (FAB, matrix NBA), m/z 1195 [M -
BPh4]+; IR (KBr, cm-1) 3042, 2968, 1605, 1580 (νas(COO)), 1472,
1423 (νs(COO)), 703; UV-vis (acetone, λmax, nm (ꢀ, M-1‚cm-1)) 400
(sh), 575 (35), 966 (46).
Magnetization Data. Magnetization experiments have been per-
formed on a Quantum Design MPMS superconducting quantum
interference device (SQUID) magnetometer operating at four magnetic
fields, 0.5, 1, 2.5, and 5 T, over the temperature range 4-300 K. The
powdered sample a(BPh4)2 (11.7 mg, 7.72 µmol) was contained in a
kel F bucket, whose contribution to the magnetic susceptibility was
evaluated independently and subtracted from the sample data. The molar
susceptibility corresponds to the resulting magnetization per magnetic
field unit per mole. The diamagnetic contributions were estimated from
Pascal’s constants at -905 × 10-6 cm3‚mol-1
.
X-ray Data Collection and Crystal Structure Determination.
Crystal data, together with details of diffraction experiment and
refinement, are summarized in Table 1S (Supporting Information).
Single crystals of complex a(BPh4)2 were obtained as [Ni2S2N6-
C42O4H42]‚2(C24BH20)‚1/2solvate, where solvate ) [5(CH3OH)‚(CH3-
CN)‚(CH3CH2OH)] and Mw ) 1638.48. A yellowish triclinic prism
(0.22 × 0.21 × 0.19 mm3) was mounted on a Enraf-Nonius CAD4
diffractometer operating at 293 K with Mo KR radiation (λ ) 0.7107
Å) monochromated by a graphite plate. The compound is triclinic, P1h,
with a ) 18.128(7) Å, b ) 20.224(7) Å, c ) 26.221(9) Å, R ) 101.30-
(2)°, â ) 104.07(2)°, γ ) 102.37(3)°, V ) 8791(6) Å3, Z ) 4, Dx )
1.238 g‚cm-3, and µ ) 0.533 mm-1. 17557 reflections were collected
in the range 4.2° e 2θ e 44.2° leading to a set of 16972 independent
reflections. The diffraction data were corrected for Lorentz and
Method B. A solution of potassium tert-butoxide (200 mg, 1.69
mmol) in distilled THF (30 mL) was added slowly (1 h) to a stirred
solution of 5 (300 mg, 0.80 mmol) in distilled THF (30 mL) under
(25) An alternate synthesis of 10 was reported (Burth, R.; Stange, A.;
Scha¨fer, M.; Vahrenkamp, H. Eur. J. Inorg. Chem. 1998, 1759-1764).
(24) Fuentes, O.; Paudler, W. W. J. Org. Chem. 1975, 40, 1210-1213.

8056 J. Am. Chem. Soc., Vol. 123, No. 33, 2001
Belle et al.
Table 1. Main Interatomic Distances (Å) and Bond Angles (deg)
in the Two Dicationic [Ni2(C42H42N6O4S2)]2+ Entities a1 and a2 from
Complex a(BPh4)2
using methanol (188-283 K) and ethylene glycol (283-378 K). The
1
spectral window for H NMR data acquisition was adjusted for each
collected temperature. A typical spectrum consisted of 16K data points,
512 scans on the whole spectral bandwidth, and a 500 ms relaxation
delay. An exponential weighting function (lb ) 10 Hz) was used during
processing. Enhancement of the resolution in the [-20 ppm, 100 ppm]
spectral region was achieved using 32K data points, 512 scans, and a
1-5 Hz line broadening exponential function on the corresponding 48
kHz bandwidth. Chemical shifts (in ppm) were referenced to the residual
solvent peak at 2.04 and 5.32 ppm for CD3COCD3 and CD2Cl2,
respectively. Inversion-recovery pulse sequence [180°-τ-90°-AQ]
was set up with 36 variable delay values ranging from 10 µs to 200
ms. The 90° pulse (7.3 µs) was calibrated at 293 K and dictated a
maximum 34 kHz spectral bandwidth for nonselective T1 measurements.
Acquisition windows were consequently adjusted to avoid peak folding
in the spectral region of interest. Relaxation time calculations involved
a monoexponential fitting procedure of the intensities vs τ values for
each selected resonance. A typical magnitude COSY spectrum was
collected with 512 t1 blocks of 4096 t2 data points and 2048 scans
over a 32 kHz bandwidth with a recycle time of 163 ms. Zerofilling to
1024 data points was performed in the t1 dimension prior to Fourier
transformation and experimental data were processed using unshifted
sinebell apodization functions in both dimensions. Standard least-
squares fit procedures were used to process data showing the temper-
ature dependence of observed chemical shifts or relaxation rates.
first entity (a1)
second entity (a2)
Ni1-Ni2
4.243(1)
2.527(2)
2.067(4)
1.981(4)
2.088(4
2.056(4)
2.054(5)
2.073(2)
2.524(2)
2.050(4)
2.010(3)
2.098(4)
2.051(5)
2.051(5)
87.2(1)
93.7(1)
93.8(1)
82.0(2)
84.0(1)
95.2(2)
90.0(2)
88.2(2)
99.7(2)
81.4(2)
91.5(2)
94.0(2)
173.4(1)
167.7(2)
171.1(2)
91.7(2)
82.1(2)
79.8(2)
98.9(2)
87.6(1)
95.0(1)
94.0(1)
85.3(1)
93.3(2)
87.2(2)
89.2(2)
96.4(1)
169.5(2)
172.0(1)
168.3(2)
Ni3-Ni4
4.299(2)
2.517(2)
2.042(4)
2.006(3)
2.100(4)
2.028(5)
2.043(5)
2.080(2)
2.513(2)
2.055(4)
1.986(4)
2.091(4)
2.039(5)
2.074(5)
85.6(1)
94.3(1)
94.1(1)
84.7(1)
173.1(1)
95.9(2)
89.7(2)
166.8(2)
90.7(2)
170.3(2)
91.9(2)
82.1(2)
80.0(2)
97.9(2)
93.8(2)
89.3(2)
166.2(2)
170.5(2)
91.7(2)
95.2(2)
80.0(2)
81.7(2)
99.6(2)
83.5(1)
94.9(2)
94.5(1)
89.7(2)
86.6(1)
94.1(1)
173.2(2)
Ni1-S1
Ni3-S3
Ni1-O1
Ni3-O5
Ni1-O2
Ni3-O6
Ni1-N1
Ni3-N7
Ni1-N2
Ni3-N8
Ni1-N3
Ni3-N9
S1-S2
S4-S3
Ni2-S2
Ni4-S4
Ni2-O3
Ni4-O7
Ni2-O4
Ni4-O8
Ni2-N4
Ni4-N10
Ni2-N5
Ni4-N11
Ni2-N6
Ni4-N12
S1-Ni1-O1
S1-Ni1-O2
S1-Ni1-N1
N1-Ni1-N3
S1-Ni1-N3
O1-Ni1-O2
O1-Ni1-N1
O1-Ni1-N2
N2-Ni1-N3
N1-Ni1-N2
O2-Ni1-N2
O2-Ni1-N3
S1-Ni1-N2
O1-Ni1-N3
O2-Ni1-N1
O4-Ni2-N6
N4-Ni2-N5
N4-Ni2-N6
N5-Ni2-N6
S2-Ni2-O3
S2-Ni2-O4
S2-Ni2-N4
S2-Ni2-N5
O4-Ni2-N5
O3-Ni2-N6
O3-Ni2-N4
O3-Ni2-O4
O4-Ni2-N4
S2-Ni2-N6
O3-Ni2-N5
S3-Ni3-O5
S3-Ni3-O6
S3-Ni3-N7
S3-Ni3-N8
S3-Ni3-N9
O5-Ni3-O6
O5-Ni3-N7
O5-Ni3-N8
O5-Ni3-N9
O6-Ni3-N7
O6-Ni3-N9
N7-Ni3-N8
N7-Ni3-N9
N8-Ni3-N9
O6-Ni3-N8
O7-Ni4-N11
O7-Ni4-N12
O8-Ni4-N10
O8-Ni4-N11
O8-Ni4-N12
N10-Ni4-N11
N10-Ni4-N12
N11-Ni4-N12
S4-Ni4-N12
O7-Ni4-O8
S4-Ni4-N10
O7-Ni4-N10
S4-Ni4-O7
S4-Ni4-O8
S4-Ni4-N11
Results and Discussion
Ligands Preparations. A series of nonsymmetric tertiary
amines N3SR (Scheme 1) are synthesized from thiosalicylic acid
and first require protection of the thiolate function to avoid
oxidation (Figure 1). The cyanoethyl-protecting group is used
here: it is more easily introduced and removed than the largely
used tert-butyl group and has the virtue of easier workup
conditions and higher yields.27-29 On the other hand, it is known
that metal ions such as copper(II)30 or nickel(II)31,32 may cleave
the S-tert-butyl linkage present in a multidonor protected ligand
by refluxing the protected ligand with metallic salts in alcohol
solution. This approach with the cyanoethyl group is then
applied here.
The cyanoethyl-protected thiolate compound 1 is prepared
by alkylation with 3-bromopropionitrile in refluxing THF (95%
yield). Then, reaction with ethyl chloroformate and NEt3
followed by reduction with NaBH433 gives the alcohol 2 in 85%
yield. Conversion of 2 to the desired aldehyde 3 is accomplished
by oxidation with MnO2 in dichloromethane (97% yield). Upon
addition of the 2-(aminomethyl)pyridine or primary amine 7 to
the aldehyde 3, and reduction, the corresponding nonsymmetric
amines 4 and 8 are obtained via the procedure outlined in Figure
1. The iminium ion resulting from the condensation of the
secondary amine 4 with pyridine-2-carboxaldehyde or 6-meth-
ylpyridine-2-carboxaldehyde (or from the condensation of amine
8 with 6-methylpyridine-2-carboxaldehyde) is reduced using
NaBH3CN.34 The desired ligands 5 (75% yield), 6 (67% yield),
polarization effects, but not for absorption. Crystal structure was solved
by direct methods26a using the texsan software.26b All non-hydrogen
atoms were refined with anisotropic thermal parameters. Almost all
hydrogen atoms were located by difference Fourier syntheses, but not
refined. All the components of the solvate have high thermal motions.
Owing to the great number of parameters to be refined, the solvate
components were only refined in the early steps of the crystal structure
determination and then fixed. Final cycle refinement, including 1909
parameters, converged to R(F) ) 0.059 and Rw(F) ) 0.060 for 11586
F > 1σ(F), S ) 1.973, ∆/σmax ) 0.01, ∆Fmax ) 0.44 e‚Å-3, and ∆Fmin
) -0.41 e‚Å-3. Structure overlays and comparisons, identification of
pseudosymmetry axes, interatomic distances, as well as dihedral angles
were obtained from Insight II software,26c where the crystallographic
.pdb file could be easily retrieved.
(27) Nikiforov, T. T.; Connolly, B. A. Tetrahedron Lett. 1992, 33, 2379-
2382.
(28) Dahl, B. H.; Bjergårde, K.; Sommer, V. B.; Dahl, O. Acta Chem.
Scand. 1989, 43, 896-901.
(29) Svenstrup, N.; Rasmussen, K. M.; Hansen, T. K.; Becher, J.
Synthesis 1994, 809-812.
NMR Data. A Varian Unity 400 spectrometer, equipped with an
extended-duration liquids Variable Temperature Control unit and a 50-L
dewar filled with liquid nitrogen, was used for collection of both 1D
and 2D NMR spectra at the required temperature. A temperature
calibration curve was recorded on the temperature range 188-378 K
(30) Becher, J.; Toftund, H.; Olesen, P. H. J. Chem. Soc., Chem.
Commun. 1983, 740-742.
(31) Dutton, J. C.; Fallon, G. D.; Murray, K. S. Chem. Lett. 1990, 983-
986.
(32) Bouwman, E.; Henderson, R. K.; Powell, A. K.; Reedijk, J.; Smeets,
W. J. J.; Spek, A. L.; Veldman, N.; Wocadlo, S. J. Chem. Soc., Dalton
Trans. 1998, 3495-3499.
(33) Ramsamy, K.; Olsen, R. K.; Emery, T. Synthesis 1982, 42-43.
(34) Borch, R. F.; Bernstein, M. D.; Durst, D. H. J. Am. Chem. Soc.
1971, 93, 2897-2904.
(26) (a) Altomare, A.; Cascarano, A.; Giacovazzo, G.; Guagliardi, A. J.
Appl. Crystallogr. 1993, 26, 343-350. (b) TEXAN, Single-Crystal Structure
Analysis Software, Version 1.7; Molecular Structure Corporation: The
Woodlands, TX, 1995. (c) Insight II Software, Version 98.0, run on a IBM
Risc Graphic Station; Molecular Simulations Inc.: San Diego, 1998.

Paramagnetic NMR InVestigations of High-Spin Ni(II) Complexes
J. Am. Chem. Soc., Vol. 123, No. 33, 2001 8057
Figure 1. Synthesis of the ligands 5, 6, and 9.
Figure 2. Mono- and dinuclear nickel(II) complexes prepared from ligand 5.
and 9 (43% yield) are then isolated. It is worth noting that the
same procedure with 3 and 2,2-bis(aminomethyl)-6-methylpy-
ridine did not yield the desired tripod 9.
upon complexation with the nickel salts and a subsequent
oxidation process. NiII-S(disulfide) linkages are relatively
rare: some mononuclear35-37 NiII disulfide and bridging bis-
(µ-disulfide)38 compounds have been reported. a(BPh4)2 is one
of the two known examples39 of µ-disulfide, bis(µ-acetato)
dinuclear NiII complexes. The six-coordinated structure of a-
(BPh4)2 persists in solution as demonstrated by the electronic
spectral properties: a(BPh4)2 displays, in acetone solution, three
broad absorption/shoulder bands with maxima at λmax (nm, (ꢀ,
M-1‚cm-1)) 400 (sh), 575 (35), and 966 (46). These bands fall
in the ranges 1430-770, 910-500, and 530-370 nm, respec-
tively, assigned to the three spin-allowed transitions from the
3A2g(F) ground state to the next excited triplet states (3T1g(P),
Syntheses of the Nickel(II) Complexes. Different routes to
the preparation of mono- and dinuclear nickel(II) complexes
starting from ligand 5 are summarized in Figure 2, together with
the complexes’ nomenclature used in the rest of this work.
Dimeric Complexes: Method A. The NiII complexes are
synthesized by mixing stoichiometric quantities of Ni(OAc)2‚
4H2O to a solution of the thio-protected ligand 5 in methanol,
then the obtained blue solution is exposed to air at room
temperature. A progressive color change from pale-blue to pale-
green is observed over the period of a day. Upon stirring over
an additional day and subsequent addition of NaBPh4, the crude
product precipitates. Filtration and recrystallization, by vapor
diffusion of ethanol into an acetonitrile solution of this solid,
affords pale-green crystals of X-ray quality. This crystalline
material is dried in vacuo and affords pure a(BPh4)2 ([Ni2(N3-
SSN3)(OAc)2](BPh4)2). Its formulation as a dinuclear complex
is confirmed by X-ray crystal structure determination (vide infra)
of the solvate a(BPh4)2‚1/2[5(CH3OH)‚(CH3CN)‚(CH3CH2OH)].
The two metal ions are six-coordinated and bridged by a
disulfide group and by two additional bidentate acetates.
3
3T1g(F), and T2g(F)) in octahedral NiII d8 complexes.40
Dimeric Complexes: Method B. Alternatively (Figure 2),
reaction of 5 with potassium tert-butoxide in THF affords, after
workup, the pure free thiolate ligand 10.25 Reaction of 10 with
(35) Warner, L. G.; Ottersen, T.; Seff, K. Inorg. Chem. 1974, 13, 2529-
2534.
(36) Warner, L. G.; Kadooka, M. M.; Seff, K. Inorg. Chem. 1975, 14,
1773-1778.
(37) Riley, P. E.; Seff, K. Inorg. Chem. 1972, 11, 2993-2999.
(38) Fox, S.; Stibrany, R. T.; Potenza, J. A.; Knapp, S.; Schugar, H. J.
Inorg. Chem. 2000, 39, 4950-4961.
(39) Handa, M.; Mikuriya, M.; Okawa, H. Chem. Lett. 1989, 1663-
Formation of this disulfide-bridged dinuclear nickel(II) complex
involves in situ cleavage of the protecting cyanoethyl group
1666.
(40) Reference 11 pp 46-52.

8058 J. Am. Chem. Soc., Vol. 123, No. 33, 2001
Belle et al.
Ni(OAc)2‚4H2O in methanol under a nitrogen atmosphere and
subsequent addition of NaBPh4 induces the precipitation of a
green powder. A crystalline material is obtained after purifica-
tion. Spectroscopic and analytic data, in the solid state as well
as in solution, proved this material to be identical to the former
one prepared from method A. This study shows that complex
a(BPh4)2 could be obtained independently by two reaction
pathways and gives evidence for the advantage of method A,
which avoids isolation of the deprotected ligand prior to reaction
with the nickel salts.
To investigate the spectroscopic properties of this complex
(in particular the 1H NMR assignment procedure described
below), a series of dinuclear complexes were synthesized,
including the PF6 salt of the previous dimer, a(PF6)2, and
complexes bearing methyl-substituted ligands 9 (b(PF6)2) and
6 (c(PF6)2).
Monomeric Complexes. To further explore the reactivity of
the protected thiolate ligands with nickel salts, we have now
focused on the isolation and characterization of compounds prior
to the dimerization process. When the blue solutions obtained
after the reaction of 5 with Ni(OAc)2‚4H2O in methanol are
stored at -20 °C and after NaBPh4 or NH4PF6 are added,
complex m(BPh4) or m(PF6) precipitates. The FAB spectra of
these products show a peak at m/z 491 for the two complexes
that are consistent with [(N3SR)Ni(OAc)]+ species. Elemental
analysis and spectroscopic data also suggest an overall complex
to anion 1:1 stoichiometry, thus leading to the formula [(N3-
SR)Ni(OAc)]BPh4 (or PF6). The infrared spectra of these
complexes measured in KBr disks show two strong features at
ca. 1540 (1540) and 1426 (1420) cm-1 for m(BPh4) (m(PF6)),
assigned to νas(COO) and to νs(COO), respectively. The νas-
(COO) features are similar to the values (range 1510-1550
cm-1) described for metal complexes with bidentate nonbridging
carboxylates and significantly differ from the values associated
with a monodentate coordination mode (1600-1725 cm-1). The
νs(COO) features are also consistent with a bidentate nonbridg-
Figure 3. ORTEP plot of the dicationic unit a ) [Ni2(C42H42N6O4S2)]2+
showing only one (a1) of the two independent entities a1 and a2 with
atom labels and numbering scheme. Hydrogens are omitted for clarity.
Ellipsoids are drawn at the 35% probability level.
ing carboxylate coordination mode (range 1450-1470 cm-1
)
described for such complexes.41 These properties then suggest
that m(BPh4) and m(PF6) complexes have one carboxylate
group ligated to the nickel(II) in a bidentate mode. UV-visible
spectroscopic data of dinuclear complexes a(BPh4)2 and a(PF6)2
and of mononuclear nickel(II) complexes m(BPh4) and m(PF6)
are highly similar. λmax (nm, (ꢀ, M-1‚cm-1)) for m(BPh4): 334
(148), 550 (10), 941 (16). λmax (nm, (ꢀ, M-1‚cm-1)) for m-
(PF6): 408 (sh), 592 (6), 966 (4). This suggests an octahedral
environment for the nickel center in mononuclear complexes
m(BPh4) and m(PF6). These bands are approximately half of
the intensity of the corresponding bands in dimeric complexes,
which is consistent with the presence of only one nickel per
m(BPh4) and m(PF6) complex. Unfortunately attempts to
characterize these materials by single-crystal X-ray studies have
Figure 4. 1H NMR spectra at 298 K of (A) the mononuclear complex
m(BPh4) in acetone-d6 and of (B) the dinuclear complex a(BPh4)2 in
CD2Cl2. Solutions are 10-2 M in complex and spectra are recorded at
a repetition rate of 2 s-1. Resolution is enhanced in the upfield portion
of the spectra through acquisition on a limited 48 kHz bandwidth.
1
been unsuccessful so far. Nevertheless, H NMR spectra (see
ligands, depending only on the synthetic strategy. These ligands
then offer unique opportunities in coordination chemistry.
Description of a(BPh4)2 Crystal Structure. The complex
a(BPh4)2 was isolated from methanol and recrystallized from
acetonitrile/ethanol. The unit cell contains two crystallographi-
cally independent [Ni2(C42H42N6O4S2)](BPh4)2‚1/2solvate enti-
ties, where solvate stands for five methanol, one ethanol, and
one acetonitrile solvent molecules. Each cationic unit, one of
which (a1) is shown in Figure 3, consists of two Ni atoms
bridged by two acetate ions and a disulfide group. The
hexacoordination sphere of each nickel atom is completed by
one amino- and two pyridyl-nitrogen atoms. Selected bond
distances and angles are reported in Table 1 and show no
Figure 4 below) afford a unique tool for discriminating between
the two species. The solid compounds are air stable for months
at room temperature, as well as in solution (acetone at -20
°C). Solutions in alcohol or acetonitrile can be handled in air
at room temperature for short periods, but after some time, the
dimeric cation is formed. The facile conversion of the mono-
nuclear to the dinuclear systems can be controlled and both
mono- and dinuclear complexes can be produced from the same
(41) (a) Nakamoto, K. Infrared and Raman Spectra of Inorganic and
Coordination Compounds, 4th ed.; Wiley: New York, 1986; pp 231-233.
(b) Blake, A. J.; Danks, J. P.; Fallis, I. A.; Harrison, A.; Li, W.; Parsons,
S.; Ross, S. A.; Whittaker, G.; Schro¨der, M. J. Chem. Soc., Dalton Trans.
1998, 3969-3976.

Paramagnetic NMR InVestigations of High-Spin Ni(II) Complexes
J. Am. Chem. Soc., Vol. 123, No. 33, 2001 8059
significant difference between the two nickel dimers. Overlaying
the two dimeric structures a1 (Figure 3) and a2 (see Supporting
Information) with the Insight II software26c furthermore leads
to a high degree of homology between the two crystallographic
entities. In fact, the two structures mainly differ on pyridyl
aromatic ring orientation, which infers slight interatomic elonga-
tions and angle distortions due to steric constraints. Superimpos-
ing Ni1, O1, O2, N1, N2, and N3 atoms in a1 with respectively
Ni4, O7, O8, N10, N11, and N12 in a2 yields to a 0.0338(1)
root-mean-square deviation for the whole molecule distances.
This emphasizes the lack of significant differences between the
two dimers. Superposition of Ni1 and Ni3 in the same fashion
gives rise to a very similar, although slightly higher, 0.0353(1)
root-mean-square value. This closely relates the two nickel
centers in each dimer, which can be emphasized through the
definition of a pseudo-C2 axis passing through the midpoint of
the Ni‚‚‚Ni and of disulfide bridge. The X-ray crystal structure
of a(BPh4)2 can then be considered as highly symmetrical
despite the lack of a true crystallographic symmetry element.
toethanol-inhibited urease (average Ni-S distance 2.36 Å7) falls
in the same range. The Ni-N distances (2.028(5)-2.100(4) Å)
are characteristic of octahedral high-spin NiII 47 and are indeed
longer than in square-planar NiII complexes.45a The weaker
Ni-N interactions for tertiary amine (2.051(5)-2.100(4) Å)
relative to pyridine amine (2.028(5)-2.074(5) Å) in a1 and a2
are typical.
Magnetic Susceptibility Studies. The magnetization proper-
ties of a(BPh4)2 have been studied to evaluate the magnetic
interactions between the two nickel atoms. The temperature
dependence of the product of the molar susceptibility by
temperature is illustrated in the Supporting Information (Figure
2S). The data are roughly constant with øT ∼ 2.54 cm3‚K‚mol-1
when the temperature is lowered from 300 to 100 K. This value
is close to that expected (2.42 cm3‚K‚mol-1) for two uncoupled
S ) 1 spins arising from nickel ions with g ) 2.2. At lower
temperatures, øT shows an increase up to 3.3 cm3‚K‚mol-1
(value at 7 K) at 0.5 T. Such a behavior is characteristic of a
ferromagnetically coupled nickel pair. The decrease observed
at very low temperatures is essentially due to saturation effects,
but also to moderate zero-field splitting, as can be seen on a
magnetization versus âH/kBT representation (where H is the
applied magnetic field).
Inside each dimer, the Ni atoms are separated by 4.243(1) Å
in a1 and 4.299(2) Å in a2, which is typical of the intermetallic
II
distances measured in octahedral Ni2 -S(sulfide) complexes
(4.220(3)38 and 4.134(2) Å39). The observed S-S bond lengths
in a(BPh4)2 (2.073(2) Å for a1 and 2.080(2) Å for a2) also fall
in the 2.015(6)-2.108(6) Å range mentioned by Fox et al.38
They are nevertheless slightly longer than the 2.056(3) Å value
in a similar dinuclear octahedral NiII compound39 with an
aliphatic disulfide bridge and to the 2.039(7) or 2.040(3) Å
values in mononuclear complexes with nonbridging disulfide
groups.37 It can be noted that the Ni-S-S-Ni atoms are not
coplanar, with a torsion angle of 72.67(7)° in a1 and 69.75(7)°
in a2. The bridging acetate ligands adopt a typical “syn-anti”
conformation, with Ni-O distances ranging from 1.981(4)-
2.010(3) Å to 2.042(4)-2.067(4) Å. These features have also
been observed in other bis(µ-acetato)-dinuclear NiII paramag-
netic complexes.39,42-44
The following equations, based on the Hamiltonian Hˆ , were
used to fit the magnetic data:
Hˆ ) -2JSˆ1.Sˆ2 + âHB.g˜.Sˆ
(1)
2J
kBT
6J
kBT
exp
+ 5 exp
2J
3g2
4
øT )
+ 2TIP‚T
(2)
6J
kBT
1 + 3 exp
+ 5 exp
kBT
The first term represents the contribution of the coupled pair
with its g factor and its exchange interaction (J). The last term
represents the contribution of the temperature independent
paramagnetism (TIP) of each nickel atom. The best least-squares
fit was obtained using the data at the four magnetic fields
together, in the valid temperature domain, and led to the
following values: J ) 2.5(7) cm-1, g ) 2.19(6), TIP ) 200-
The coordination geometry around each nickel is a distorted
octahedron. Nevertheless, angles involving the Ni atoms show
little deviation from the theoretical angles, emphasizing the
quasiregularity of the polyhedra. As a matter of fact, the largest
deviations are encountered for O7-Ni4-N12 (166.2°, 13.8°
deviation) and N4-Ni2-N6 (79.8°, 10.2° deviation). With
respect to distances between the metal center and bound
heteroatoms, the Ni-S bonds (2.524(2) and 2.527(2) Å for a1,
and 2.513(2) and 2.517(2) Å for a2) are elongated; they are
indeed much longer than the other Ni-O or Ni-N distances,
which all fall in the range 1.981(4)-2.100(4) Å. These results
are however consistent with the Ni-S distances observed in
high-spin six-coordinated NiII complexes bearing disulfide
ligands, including monodentate disulfide (2.470(5),35 2.456(2),36
and 2.472(2) Å37) and bridging disulfide (2.503(3)39 and 2.519-
(4)-2.565(4) Å38). In paramagnetic Ni-SR2 complexes, the
reported Ni-S bond lengths are in the range 2.32-2.52 Å45 as
in bridging alkyl- or arylsulfide (2.518(1)-2.541(1) Å18 and
2.499(4)-2.482(4) Å46). Interestingly, the Ni-S in 2-mercap-
(200) × 10-6 cm3‚mol-1
.
More sophisticated equations were used including zero-field
splitting, inter-dimer exchange interaction, and paramagnetic
impurities.48 However, in the temperature range used for the
fitting procedure, zero-field splitting and inter-dimer effects (the
distance between two Ni atoms of two distinct dimers is dmin
)
8.699 Å) should be very small and will not significantly change
the obtained J value. Moreover, the saturation effects, which
are important at low temperature, will be superimposed to these
interactions and eq 2 is no more valid.
Only one nickel(II) dimer with a disulfide and two acetate
bridges has been structurally and magnetically characterized so
far.39 A Curie-Weiss law was observed. However, the tem-
perature domain explored was limited (100-300 K) and
therefore the small ferro- or antiferromagnetic interaction, if
existing, could not be measured. A nickel(II) dimer with two
disulfide bridges recently has been structurally and magnetically
characterized by Fox et al.38 An antiferromagnetic interaction
was described with J ) -13.0(2) cm-1. This complex presents
some similarities to a(BPh4)2: Ni‚‚‚Ni distances (4.220 Å for
(42) Volkmer, D.; Ho¨rstmann, A.; Griesar, K.; Haase, W.; Krebs, B.
Inorg. Chem. 1996, 35, 1132-1135.
(43) Buchanan, R. M.; Mashuta, M. S.; Oberhausen, K. J.; Richardson,
J. F. J. Am. Chem. Soc. 1989, 111, 4497-4498.
(44) Chaudhuri, P.; Ku¨ppers, H.-J.; Wieghardt, K.; Gehring, S.; Haase,
W.; Nuber, B.; Weiss, J. J. Chem. Soc., Dalton Trans. 1988, 1367-1370.
(45) (a) Bohle, D. S.; Zafar, A.; Goodson, P. A.; Jaeger, D. A. Inorg.
Chem. 2000, 39, 712-718. (b) Cha, M.; Gatlin, C. L.; Critchlow, S. C.;
Kovacs, J. A. Inorg. Chem. 1993, 32, 5868-5877 and references therein.
(46) Baidya, N.; Olmstead, M.; Mascharak, P. K. Inorg. Chem. 1991,
30, 929-937.
(47) Mikuriya, M.; Handa, M.; Shigematsu, S.; Funaki, S.; Fujii, T.;
Okawa, H.; Toriumi, K.; Koshiba, T.; Terauchi, H. Bull. Chem. Soc. Jpn.
1993, 66, 1104-1110.
(48) Ginsberg, A. P.; Martin, R. L.; Brookes R. W.; Sherwood, R. C.
Inorg. Chem. 1972, 11, 2884-2889.

8060 J. Am. Chem. Soc., Vol. 123, No. 33, 2001
Belle et al.
1
Figure 5. (A) H 1D and (B) magnitude COSY spectra of a(PF6)2 in CD2Cl2 at 293 K. The spectrum was obtained using 4096 t2 and 512 t1 data
points. Nonshifted sinebell apodization function and zerofilling to 4K t2 × 1K t1 data points prior to Fourier transform were used during data
processing. No symmetrization was performed on the resulting 2D spectrum, due to noise in the t1 dimension. Pyridyl and thiophenolic spin systems
are shown. Part C reproduces the ligand numbering scheme used for signal assignment.
Fox’s complex and an average value of 4.271 Å for a(BPh4)2),
a bridging disulfide. Nevertheless the overall symmetry of each
complex is rather different. A better orbital overlap in the Fox
dimer, which is favorable to a superexchange pathway, and
consequently, to an antiferromagnetic interaction, may be
responsible for the sign and intensity difference observed in
the J values. Further studies such as theoretical calculations
should be useful for understanding the magnetostructural
correlations in disulfide-bridged bis-nickel(II) complexes.
1H NMR Resonance Assignment Strategy for a(BPh4)2.
The high similarity between the two dimers a1 and a2 identified
in the crystallographic unit cell is expected to yield to the
detection of a unique dimer in solution through 1H NMR studies.
Furthermore, the presence of a pseudo-C2 axis in each dimer’s
crystal structure suggests that the complex may adopt a 2-fold
symmetry in dichloromethane solution. This is fully consistent
with the identification of 16 resolved resonances only, on the
1H NMR spectrum of a(BPh4)2 in CD2Cl2 at 298 K (Figure
4B). As a starting point for assignment, integration of every
resolved signal can be considered, thus allowing the unambigu-
ous identification of the µ-acetato methyl group resonance at
25.59 ppm. This assignment is furthermore confirmed through
acetate vs chloride substitution (data not shown).
1
1H NMR General Features. H NMR spectra of dinuclear
a(BPh4)2 and mononuclear m(BPh4) complexes, illustrated in
Figure 4, show striking differences. The monomeric species m-
(BPh4) (Figure 4A) present few and extremely broad reso-
nances: a maximum of seven resonances are detected and line
widths up to 16 000 Hz are observed. In the case of the a(BPh4)2
dimer (Figure 4B), 16 very well resolved isotropically shifted
COSY experiments are particularly adapted to mapping spin-
spin coupling networks in diamagnetic samples and cross-peak
intensities are dictated by the importance of the nuclear scalar
J-coupling constant value. In paramagnetic complexes, however,
its success can severely be hampered by the following factors:
intrinsic T2 relaxation time,10 spectral width and related 90° pulse
efficiency, and presence of diamagnetic species. Taking these
limitations into account, a magnitude COSY spectrum with
decent signal-to-noise ratio (Figure 5) could be obtained on a
34 kHz window for a a(PF6)2 sample in CD2Cl2 at 293 K with
2048 scans. It must be noted that chemical shift and resonance
line width of the complex cation a are essentially insensitive to
anion substitution, which appears to be essential. The upfield
1
resonances can be identified on the H NMR spectrum on a
300 ppm spectral window with 2000 Hz maximum line width.
1H NMR features such as smaller hyperfine shifts and larger
line width could thus be considered as the spectroscopic
signature for high-spin nickel(II) monomers vs high-spin nickel-
(II) dimers, at least in a mixed N/S/O coordination mode. Taking
advantage of the versatility of ligands N3SR (Scheme 1),
resonances assignment strategies, impact of magnetic coupling
on the 1H NMR spectral properties of high-spin nickel(II)
compounds, and relaxation mechanisms are addressed in the
following sections.

Paramagnetic NMR InVestigations of High-Spin Ni(II) Complexes
J. Am. Chem. Soc., Vol. 123, No. 33, 2001 8061
Table 2. NMR Parameters in CD2Cl2 at 293 K and Crystal Structure Data for the Dimeric Species a(BPh4)2
resonance assignment
δ (ppm)a
∆ν1/2 (Hz)b
Ni1‚‚‚H (Å)c
Ni2‚‚‚H (Å)c
angle (deg)d
T1exp (ms)e
T1calc (ms)f
Py1+2
H1
155.03
171.78
52.04
16.30
51.20
48.38
14.15
48.38
∼ 5.32i
15.33
0.09
12.36
137.80
245.82
112.28
∼7.5j
0.09i
1484
2027
238
60
554
204
55
3.14g
3.12h
5.07g
5.71g
4.87g
5.05h
5.71h
4.89h
4.90
6.48
6.61
5.28
3.69
3.72
3.86
3.07
3.27
3.06
4.20
6.01g
5.77h
7.76g
8.60g
7.95g
7.80
9.06
8.66
5.83
7.98
8.79
7.85
7.01
7.49
6.45
5.66
7.41
4.93
4.56
0.16
0.14
3.09
7.57
2.75
3.13
8.00
3.13
0.15g
0.15h
2.55g
5.16g
2.05g
2.50h
5.27h
2.14h
1.65
Py1
Py2
Thio
H2
H3
H4
H2
H3
H4
H1
H2
204
44
43
99
1069
1790
1045
1092
10.35
11.41
3.88
0.58
0.60
0.62
0.17
9.29
H3
H4
11.41f
3.20
CH2 Heq
CH2 Hax
OAc
Py transg
Py cish
Thio
Py transg
Py cish
Thio
CH3
161.30
151.13
167.73
79.16
89.51
49.10
0.40
0.42
0.51
0.13
0.20
0.13
0.55
59.29
25.99
351
1.10
a Observed chemical shifts, in ppm, referenced to residual solvent resonance at 5.32 ppm. b Line width of each resolved resonance. c Average
distances obtained from the four slightly nonequivalent proton positions in the crystal structure unit cell. These distances are measured within the
Insight II program. Ni1 points out the Ni atom closer to the considered proton. d Angles considered are Ni1-N-C-H angles and are measured
within the Insight II program. e T1 values obtained from the monoexponential fitting procedure of the inversion-recovery data on 36 array values.
A comparison of three independently acquired data sets yields a maximal 2.5% error on the reported values. f T1 values calculated on the basis of
a model taking into account a predominant dipolar relaxation mechanism, arising from two paramagnetic NiII S ) 1 centers with negligible magnetic
interactions. The relaxation rates (T1calc-1) are calculated on the basis of crystal structure data (giving the rNi-H values) with respect to that of a
-6
reference resonance (T1ref-1), according to the following equation: T1calc-1 ) T1ref-1[(rNi1-H-6 + rNi2-H-6)/(rNi1-H-6 + rNi2-H )ref] (see ref 50). The
Thio-H3 proton is used as a reference in this calculation, with T1ref ) 11.41 ms, rNi1-H ) 6.61 Å, and rNi2-H ) 8.79 Å. A change in the reference
implies different T1 values, but highly similar qualitative agreement between the experimental and calculated values. This fitting procedure gives
better results than other described methods, involving the average distance between the metal centers and the proton of interest in a point dipole
approximation.20, 22 g Structural data measured for the pyridyl ring in a respectively trans position with respect to the thiophenolic anchoring on the
metal. h Structural data measured for the pyridyl ring in a respectively cis position with respect to the thiophenolic anchoring on the metal. i Resonances
identified on the basis of a temperature study that can solve fortuitous overlap. j Proposed chemical shift value based on spectral comparison with
a(PF6)2.
portion of the spectrum exhibits a three-spins system (15.36,
12.75, and 0.12 ppm) that unambiguously assigns the corre-
sponding protons to one of the aromatic rings. Two additional
similar spin-spin connectivity patterns, involving the resonances
at 52.05, 51.24, and 16.42 ppm (48.40 and 14.27 ppm,
respectively), are outlined in the downfield portion of the
spectrum. The resonance at 48.40 ppm integrates for two protons
and results from coincidental degeneracy, which can be raised
with solvent change (to DMSO for example). These spin systems
then account for parts of all three aromatic rings. It must be
noted that the Ni-S bond (average distance 2.520 Å) is expected
to be weaker than the Ni-N bond (average distance 2.064 Å).
Assuming a major contribution of the contact shift to the overall
hyperfine shift, like in nickel(II) monomers,49 the most upfield
shifted three-spin system can consequently tentatively be
assigned to the thiophenolic ring protons. Furthermore, spin-
spin coupling to aromatic H1 may occur beyond detection due
to the proximity of these protons to the nickel center and
enhanced resulting relaxation properties. This leads to the
following specific assignment: resonances at 16.42, 14.27, and
0.12 ppm correspond respectively to Py1-H3, Py2-H3, and Thio-
H3. The subsequent H3-to-H2 and H3-to-H4 coupling schemes
then result in the observed connectivities. These assignments
are transposable to a(BPh4)2 and are summarized in Table 2 (it
must be noted that labels 1 and 2 are arbitrarily proposed for
one or the other of the pyridyl rings).
Figure 6. The substitution of Py-H1s in a(PF6)2 by methyl
groups in b(PF6)2 results in the disappearance of the 1200 Hz
broad resonances at 171.43 and 155.43 ppm (Figure 6A) and
appearance of two new resonances, integrating for three protons
each, at -8.11 and -10.25 ppm (Figure 6B). The magnetic
inequivalence of the two pyridyl H1s resonances originates in
differences in contact and dipolar shifts for the pyridyl in the
cis vs trans position with respect to the thiophenolic ring. The
Py-H1 signal detected at 171.43 ppm in a(PF6)2 probably
corresponds to the most upfield-shifted methyl resonance at
-10.25 ppm in b(PF6)2, but no specific assignment can
nevertheless be proposed at the present stage. The present partial
assignment is nevertheless confirmed by the detection of both
sets of resonances in the dissymmetrically substituted c(PF6)2
complex (Figure 6C). Furthermore, four resonances in a 1:1:
1:1 ratio are detected in the later case in the methyl upfield
region. Such observation is consistent with the nonselective
formation of four stereoisomers upon complexation: the cis-
cis, trans-trans, cis-trans, and trans-cis isomers, cis or trans
referring to the position of the methylated ring with respect to
the thiophenolic groups.
Specific chemical substitution at the Py-H1 position also
provides additional evidence to former assignments. Comparison
of COSY spectra collected under the same conditions (see
Supporting Information), as well as 1D-temperature and relax-
ation studies, for a(PF6)2, b(PF6)2, and c(PF6)2 clearly identifies
the resonances at 15.36, 12.75, and 0.12 ppm in a(PF6)2 to
respectively Thio-H2, Thio-H4, and Thio-H3, as well as the
presence of a fast relaxing resonance under the residual CD2-
Cl2 solvent peak at 5.32 ppm to Thio-H1. Present assignments
are then transposed to a(BPh4)2 and reported in Table 2.
Further assignments cannot be made without specific chemical
substitution, or interpretation of proton longitudinal relaxation
measurements in correlation with distances between Hs and the
nickel centers. Directed chemical substitution, especially at the
key Py-H1 position, is considered first and is monitored in
(49) Reference 11 pp 56-60.

8062 J. Am. Chem. Soc., Vol. 123, No. 33, 2001
Belle et al.
cos2 θ
cos2 (120 + θ)
δax
δeq
)
)
4.4 × 10-1 for thiophenolic methylene
4.0 × 10-2 for trans-pyridyl methylene
9.7 × 10-5 for cis-pyridyl methylene
(3)
where θ states for the Ni-N-C-Hax dihedral angle, δax and
δeq, are the observed chemical shift for axial- and equatorial-
like methylene protons CH2 Hax and Heq, and trans and cis
designate the position of the pyridyl ring bound to the methylene
of interest with respect to the thiophenolic anchor on the nickel
center. According to the above results and relaxation data in
Table 2, Thio CH2 Heq is expected to show the smaller isotropic
hyperfine shift and the longer T1 value. It is therefore assigned
to the resonance at 112.28 ppm (Figure 4B). Furthermore, this
resonance remains unshifted upon chemical substitution on the
pyridyl ring (Figure 6A,B), which is in complete agreement with
the present assignment. According to the δax/δeq ratio, Thio CH2
Hax (Ni‚‚‚H distance ) 3.06 Å) is expected to resonate at 49.40
ppm and to present a 0.13 ms T1 relaxation time. The detected
peak at 59.29 ppm, with T1 ) 0.17 ms, is then a very likely
candidate. Assuming that the pyridyl methylene CH2 Hax protons
fall in the diamagnetic window, the above criteria lead to the
identity of the resonances at 245.82 and 137.80 ppm respectively
as cis- and trans-Py-CH2 Heq. Their corresponding geminal
partners are then expected to resonate at 5.51 and 0.02 ppm
and to present 0.13 and 0.20 ms longitudinal relaxation times.
Interestingly, the 1H NMR spectrum of complex a(PF6)2 at 293
K, reported in Figure 6A, shows a resonance at 8.58 ppm (T1
) 0.15 ms) and a temperature study of a(BPh4)2 reveals an
additional resonance at 0.09 ppm at 263 K, for which chemical
shift and the slope of the δ ) f(1/T) plot are consistent with its
assignment to cis-Py-CH2 Hax. The good agreement between
these observations and the above predictions is considered strong
evidence in favor of the assignment of pyridyl CH2 Hax.
Complete assignment and spectroscopic properties (observed
chemical shift, longitudinal relaxation times, and line width)
of each resonance are reported in Table 2 for complex a(BPh4)2.
These results have been obtained through a combination of 2D
methods, directed chemical substitution, and analysis of T1 data,
as well as the chemical shift temperature dependence study.
Similar strategy yields the complete assignment of a(PF6)2, b-
(PF6)2, and c(PF6)2 1H NMR resonances that is provided in the
Supporting Information.
Figure 6. Portions of the 1H NMR spectra in CD2Cl2 at 293 K of the
dinuclear complexes bearing different chemical substitutions on the
ligand: (A) X1 ) X2 ) H1, a(PF6)2; (B) X1 ) X2 ) CH3, b(PF6)2;
(C) X1 ) H1 and X2 ) CH3, c(PF6)2, according to the ligand numbering
scheme shown in part D. The solvent residual signal at 5.32 ppm is
labeled S on each spectrum. All solutions are 5 × 10-3 to 10-2 M in
dinuclear complex.
At this point, pyridyl and thiophenolic ring assignments,
together with acetate identification, leave four resolved signals
in the downfield spectral window of the a(BPh4)2 spectrum (at
245.82, 137.80, 112.28, and 59.29 ppm). These resonances must
consequently arise from methylene protons. According to
literature,50,51 geminal methylene protons involved in the chelate
ring experience significantly different hyperfine shift (contact
shift) and relaxation times (dipolar relaxation). Equatorial-like
protons (i.e. methylene protons located further away from NiII
with a dihedral angle Ni-N-C-H closer to 180°) are conse-
quently expected to show longer relaxation times and greater
downfield hyperfine shift with respect to their geminal partners
CH2 Hax. Three of the unassigned resonances with large
downfield isotropic hyperfine shifts (>100 ppm) and enhanced
0.7-0.9 ms T1 and 0.3-0.6 ms T2 relaxation times are then
considered likely candidates. More detailed specific assignment
of these protons implies T1 measurement analysis, in correlation
with the crystal structure considerations reported in Table 2.
Using these data, the expected δax/δeq ratio is calculated for
each methylene group, according to the following empirical
formula proposed by Ho and Reilley:51
1H NMR Resonance Assignment Strategy for m(PF6) or
m(BPh4). The above strategy assignment remains inefficient
for the monomer species m(PF6) or m(BPh4), due to extreme
T1 and especially T2 relaxation times. Nevertheless, a tentative
assignment can be proposed on the basis of T1 comparisons
between the monomer and the dimer, providing a conservative
geometry around the nickel center, as suggested by other
spectroscopic methods except NMR. The resonances at 15.26
(T1 ) 6.99 ms) and 11.41 ppm (T1 ) 9.85 ms) must arise from
aromatic H3 protons and/or Thio-H2, while signals at 48.85
(T1 ) 2.27 ms) and 10.07 ppm (T1 ) 2.90 ms) rather correspond
to aromatic H2 or H4 protons (Figure 4A). On the other hand,
resonances at 158.54 (T1 ) 0.18 ms) and -2.96 ppm (T1 )
0.55 ms) may be respectively assigned to methylene groups (or
aromatic H1 protons) and the acetate group. Observed T1 values
are overall slightly shorter than the corresponding values in the
(51) Holm, R. H.; Hawkins, C. J. NMR of paramagnetic molecules:
Principles and applications; La Mar, G. N., Horrocks, W. DeW., Jr., Holm,
R. H., Eds; Academic Press: New York, 1973; pp 243-332 and references
therein.
(50) Ming, L.-J.; Jang, H. G.; Que, L., Jr. Inorg. Chem. 1992, 31, 359-
364.

Paramagnetic NMR InVestigations of High-Spin Ni(II) Complexes
J. Am. Chem. Soc., Vol. 123, No. 33, 2001 8063
for example, Py1+2-H1, Py- and Thio-CH2s), thus suggesting
small deviations to the Curie law.
Such deviations can be understood considering the presence
of one (or more) low-lying excited state(s), where the differences
in energy between ground and excited levels are of the order
of thermal energy kBT. In the particular case of a(BPh4)2, each
nickel center lays in a pseudo-octahedral environment of four
3
different ligand atoms. The ligand field theory leads to a A2g
orbitally nondegenerate ground state for each nickel atom and
the spin levels may be split to a small extent through the zero-
field splitting D, usually on the order of 0-60 cm-1 in
mononuclear complexes.52 Such contribution also stands for
dinuclear species, although the magnitude of the zero-field
splitting may differ from the mononuclear ones. According to
Kurland and McGarvey,53 this implies, for a spin S ) 1 when
D , kBT, a T-1 but also additional T-2 temperature dependence
of the hyperfine chemical shift. In the dinuclear complexes
studied, the two nickel atoms are furthermore ferromagnetically
spin-coupled to each other and the resulting J coupling constant
is an additional source for energy level splitting. If J is on the
order of a few cm-1, the S′ ) 0, 1, 2 energy levels (S′ is the
total dimer spin-state) are expected to be significantly populated
at room temperature. According to a very elegant study by
Shokheriev and Walker,54 this also implies an additional T-2
contribution to fitting expressions of observed chemical shift.
To target the potential importance of J and/or D in dinuclear
nickel(II) complexes and deviations to Curie law, observed
chemical shifts versus the inverse temperature are processed
using a second-order polynomial fitting expression. No signifi-
cant improvement of fit quality can be observed from this
procedure and correlation coefficients remain rather insensitive
to the type of fitting expression retained. Nevertheless, very
good fits, reported in Table 3, can be obtained in the latter
procedure that give intercepts identical to the expected diamag-
netic shift. T-1 and T-2 coefficients are obtained with a 10%
accuracy and the T-2 coefficient for each proton contains crucial
information on J and D parameters. Nevertheless, in the case
of the series of dinuclear high-spin nickel(II) studied, deviations
to the strict Curie law are too small on the temperature range
studied to determine precisely either the nature (J and/or D) or
the exact value of these parameters (it can only be estimated to
be on the order of a few cm-1 on the basis of a b′/a′ analysis
from Table 3). This observation is in complete agreement with
the plateau observed from magnetic susceptibility data øT )
f(T) between 188 and 298 K (see Supporting Information).
Similar behavior and results are obtained for b(PF6)2 and c-
(PF6)2, as reported in the Supporting Information.
Figure 7. Observed chemical shift dependence vs reciprocal temper-
ature for resonances of a(BPh4)2 in CD2Cl2. Signs and lines show
respectively experimental and second-order polynomial square fit data
for all assigned resonances, except the cis-pyridyl CH2 Heq signal. (+)
Filled lines emphasize the behavior of pyridyl ring and methylene
protons, (() dotted lines the behavior of the thiophenolic ring and
methylene protons, and (∆) broken lines the behavior of acetate protons.
dinuclear species. Enhancement of the monomer vs dimer
relaxation rates is even more sensitive for the transversal
relaxation rates. As an example, the resonance at 158.54 ppm
shows the same T1 value as Py-H1 in the dimeric species, while
the line width of the corresponding peaks differs by a factor of
8. These observations are consistent with the presence of a
relaxation cooperative effect in dimeric vs monomeric com-
pounds. This effect is similar to results previously reported on
copper(II) model complexes.22
Hyperfine Shifts in Nickel(II) Dimers. Chemical shift of
Relaxation Properties in a(BPh4)2. This dimer presents a
small ferromagnetic coupling constant (J ) 2.5 cm-1) as
emphasized through SQUID measurements and NMR temper-
ature-dependence studies. Relaxation rates, T1-1, can conse-
1
every H NMR dimer resonance is significantly temperature
sensitive. Experimental data over the temperature range 188-
298 K are reported in Figure 7 for a(BPh4)2 in CD2Cl2, showing
a typical Curie behavior. Fits of reasonably good quality
(correlation coefficients greater than 0.985 for all resonances
and greater than 0.999 in most cases) can be obtained using a
linear Curie law and are reported in Table 3. Slopes highly
correlate with the magnitude and sign of the hyperfine shift in
each case (δhf ) δ - δdia, where δ is the observed chemical
shift). A comparison of the slopes for the two methylene geminal
partners (16.5 × 103 ppm.K for Thio-CH2 Hax and 38.3 × 103
ppm.K for Thio-CH2 Heq) furthermore emphasizes the absence
of correlation between the magnitude of the slope and the Ni‚
‚‚H distances (respectively 3.06 and 3.86 Å). This is a strong
evidence in favor of a highly predominant contact contribution
to chemical shift. On the contrary, some of the intercepts
significantly differ from the expected8 diamagnetic shift (see,
quently be handled in the point dipole approximation20,22 and
are expected55,56 to be proportional to the sum rNi1‚‚‚H
+
-6
rNi2‚‚‚H-6, where Ni1 is the closest metal atom to the considered
proton H. Crystal structure data and Thio-H3 longitudinal
relaxation time are used in this calculation. Results are reported
(52) Reference 11, pp 52-55.
(53) Kurland, R. J.; McGarvey, B. R. J. Magn. Reson. 1970, 2, 286-
301.
(54) Shokhirev, N. V.; Walker, F. A. J. Phys. Chem. 1995, 99, 17795-
17804.
(55) Banci, L.; Bertini, I.; Luchinat, C. In Nuclear and electron
relaxation; VCH Publishers: New York, 1991; pp 91-122 and 143-162.
(56) If the correlation time in a homodimer is considered to be the same
-1
for each spin-level, T1 in the coupled system is equal to half that in the
uncoupled system (see ref 53).

8064 J. Am. Chem. Soc., Vol. 123, No. 33, 2001
Belle et al.
Table 3. Temperature Dependence of Chemical Shifts and Relaxation Rates for a(BPh4)2 in CD2Cl2
Curie plotb
polynomial plotc
T1-1 slopedd
b
c
c
d
δ
δint
a × 10-3
δdia
a′ × 10-3
b′ × 10-5
(ppm‚K2)
R × 10-3
resonance assignment
(ppm)a
(ppm)b
(ppm‚K)
(ppm)c
(ppm‚K)
(ms-1‚K2) (R)
Py1+2
H1
155.03
171.78
52.04
16.30
51.20
48.38
14.15
48.38
∼ 5.32e
15.33
0.09e
12.36
137.80
245.82
112.28
∼7.5f
0.09e
-2.85
-0.23
5.85
6.85
4.86
5.51
6.14
5.51
7.09
44.7
50.3
13.5
2.8
13.6
12.5
2.3
12.5
-0.5
2.5
-2.3
1.6
42.2
75.5
38.3
-5.1
16.5
8.49
8.49
7.29
7.29
7.60
7.29
7.29
7.60
7.35
7.35
7.35
7.35
3.80
3.80
3.86
3.80
3.86
45.5
41.8
12.8
2.6
12.2
11.7
1.8
11.6
-0.6
2.2
-2.0
1.4
37.3
70.0
27.5
-7.1
15.6
6.4
3.6
0.8
0.2
1.6
1.0
0.6
1.2
0.2
h
h
Py1
Py2
Thio
H2
H3
H4
H2
H3
H4
H1
H2
26.1(0.991)
10.6(0.982)
28.7(0.987)
26.1(0.996)
10.1(0.976)
26.1(0.996)
e
7.9(0.988)
e
20.8(0.981)
h
h
h
e
6.86
7.88
6.86
0.2
-0.3
0.2
H3
H4
CH2 Heq
CH2 Hax
OAc
Py transg
Py cisg
Thio
Py transg
Py cisg
Thio
CH3
-6.53
-3.54
-18.64
17.53
2.50
5.8
11.0
12.6
1.1
1.2
h
59.29
25.99
2.16
7.0
2.50
6.2
0.9
26.1(0.991)
a Observed chemical shifts, in ppm, at 293 K. b Fit of the chemical shift variations (in ppm) with temperature (in K), according to the Curie law:
δobs) δint + a/T, where δint is the intercept at infinite temperature and a is the Curie slope. The correlation coefficient is greater than 0.99 in each
single case. c Fit of the chemical shift variations (in ppm) with temperature (in K), according to a polynomial function: δobs ) δdia+ a′/T + b′/T2,
where δdia is the observed diamagnetic shift in the free ligand. The correlation coefficient is greater than 0.99 in each single case. They do not
significantly differ from the values obtained in the case of the Curie law. d Fit of the relaxation rate T1 (ms-1) with the temperature (in K),
-1
-1
according to the following expression: T1 ) R/T2. The correlation coefficients are given within parentheses. T1 values of CH2 Hax and CH2 Heq
are very difficult to follow with temperature, as relaxation times get lower with temperature. Fitting results are not given due to the small number
of available data. e Fortuitous overlap at 293 K, which can be solved through temperature change. f Resonance buried under the BPh4- peak. g Methylene
group bound to the pyridyl ring fixed to the metal in a trans (respectively cis) configuration with respect to the thiophenolic group. h T1 values are
too short when temperature is decreased. Low reproducibility of T1 measurement forbid curve fitting.
in Table 2, which shows good agreement with experimental T1
values, especially for methylene protons. In the case of aromatic
ring protons, increased differences are observed, which can arise
from slight discrepancies between the solid-state and solution
structures of the dinuclear complex (especially ring flipping)
or from electron spin-delocalization and contact interactions
within the ring. Nevertheless, qualitative agreement is substantial
within a specified aromatic ring and thus confirms, if needed,
proposed assignments. It can be noted that distinction of the
two pyridyl rings (cis or trans with respect to the thiophenolic
anchor on the nickel atom) remains hazardous, due to high
similarity of measured T1 values for protons occupying equiva-
lent positions on the two different rings and despite clear distinct
parameters in the solid state. These results are indications in
favor of pyridyl rings flipping around an equilibrium position
in solution.
Whether temperature is also a relevant factor is addressed
next. Decreasing the temperature of a a(BPh4)2 solution in
dichloromethane induces an increase in the longitudinal relax-
ation rates. This observation can only be explained if electronic
spin-relaxation overcomes all relaxation mechanisms and if the
Figure 8. Observed relaxation rate vs reciprocal temperature for the
resonance assigned to Py1-H2 for a(BPh4)2 in CD2Cl2. The result of a
second-order polynomial least-squares fit, Τ1-1(ms-1) ) a + b/T +
c/T2 (T in K), is shown with the solid line: a ) 0.06 ms-1, b ) -10
K‚ms-1, c ) 2.5 × 10-3 K2‚ms-1, and correlation coefficient ) 0.995.
nuclear spins interact with a time-averaged electron magnetic
moment. This rather unusual behavior is quantitatively repro-
duced in Figure 8 in the particular case of Py1-H2 for 188 e T
(K) e 298. Least-squares fit procedures are carried out using
alternatively linear or second-order polynomial regression. A
mathematically and physically acceptable result is obtained for
at 293 K. These data are then strong evidence in favor of a
dominant Curie longitudinal relaxation mechanism, where,
according to Clementi and Luchinat:57
-1
a T-2 fit of the relaxation rate T1 (correlation coefficient )
0.991 and slope ) 26081 ( 180 K2‚ms-1). When applied to
a(BPh4)2 protons for which the number of T1 data are numerous
enough, this procedure gives the results reported in Table 3.
Interestingly, the slope of detected relaxation rate vs the square
of the inverse temperature is related to the distance between
the proton of interest and the nickel centers, and it is highly
consistent with what has been shown for calculated T1 values
2
2
2
2
µ0 γn ge µB
3τc
1 + ωn τc2
2
-1
T1
)
f(J,D)
(4)
6
2
2
( )
5 4π
[
]
r (3kBT)
f(J,D) being almost temperature independent when J, D , kBT
and τc ) τr-1 + τm-1. If Curie behavior has been supported
-1

Paramagnetic NMR InVestigations of High-Spin Ni(II) Complexes
J. Am. Chem. Soc., Vol. 123, No. 33, 2001 8065
by experimental data in the case of transverse relaxation,
complex a(BPh4)2 is, to our knowledge, the first example of
this type for longitudinal relaxation. This pattern is nevertheless
common to the whole series of high-spin nickel(II) dimers, as
b(PF6)2 and c(PF6)2 present the same characteristics (see
Supporting Information).
in the monomer would then render enhanced transversal
relaxation rates and would be consistent with the observations.
Relaxation times in the present series of high-spin nickel(II)
dimers are sensitive to the counterion nature. Replacement of
-
-
the electronrich BPh4 by the smaller PF6 anion in a(BPh4)2
induces an overall relaxation time lengthening of the corre-
sponding protons by a factor of 1.1 to 1.6, in dichloromethane
solution (T1 values are provided in the Supporting Information).
When Py-H1 protons are substituted by methyl groups such as
in the case of b(PF6)2, comparison of the T1 data also shows a
uniform decrease in the observed relaxation rates (T1 values
given in the Supporting Information). On the other hand, when
the behavior of the observed chemical shifts vs inverse tem-
perature is compared for a(BPh4)2 and b(PF6)2, the T-1 and
T-2 slopes are similar in the two complexes. The J-coupling
constant value, which is mainly dependent on the bridges
geometry, is not expected to be highly modified through pyridyl
chemical substitution. On the contrary, fluctuations in geometry
around the nickel atom and intermolecular interactions may
affect the zero-field splitting of each individual nickel atom and
consequently the zero-field splitting of the total S′ ) 2 and 1
spin-states. The observations are then consistent with a dominant
J-dependence on chemical shift and a dominant D-dependence
on relaxation times in the nickel(II) dimers. This emphasizes
the possible role of zero-field splitting magnitude in the detection
of sharper resonances in the case of dimers vs monomers. Both
parameters are nevertheless in the case of the studied dimers
very small and these proposed explanations must be applied to
a larger number of dinuclear high-spin nickel(II) complexes.
NMR and Magnetic Properties of High-Spin Nickel(II)
Complexes. Despite extremely short longitudinal relaxation
times (the longer T1 for a(BPh4)2 is 11.41 ms at 293 K),
resonances in dinuclear high-spin nickel(II) complexes have all
been identified. This results from a combination of 2D NMR,
specific site-directed chemical substitution, as well as relaxation
and chemical shift temperature dependence data. Comparison
of calculated and experimental T1 values shows that solid-state
structure persists in solution and allows the dipolar relaxation
mechanism to be established as a major pathway. The assign-
ment strategy developed here for a(BPh4)2 applies successfully
in the case of b(PF6)2 and c(PF6)2 and is most probably general
to every dimeric species containing two weakly ferromagneti-
cally coupled high-spin nickel(II)s.
The same strategy nevertheless appears to remain hopeless
in the case of high-spin nickel(II) monomers. This is due to
intrinsic line widths, which are about 1 order of magnitude
sharper in the dinuclear nickel(II) complexes than in the
equivalent monomer. An analogous situation has been described
for dinuclear vs mononuclear copper(II) complexes, indepen-
dently of the sign and size of the magnetic coupling between
the two metal centers.20,22 In the case of weakly coupled
dinuclear copper(II) complexes, where relaxation mechanisms
are dominated by dipolar contribution,22 the sharpness of NMR
signals is due to shorter electronic relaxation times in dimeric
vs monomeric species. Among other factors, the proximity of
the two metal ions, flexibility around them due to ligand
structure, and electronic delocalization are suggested to be
responsible for this observation. Modulation of zero-field
splitting of the total spin-state S′ ) 1 in the coupled systems is
considered, according to the authors,22 a likely source of
electron-relaxation. In the case of homodimers containing high-
spin nickel(II) in a mixed N/O/S environment, the present study
shows that the predominant longitudinal relaxation mechanism
occurs via a Curie dipolar mechanism. T2 relaxation times in
the dimer show similar behavior with temperature as T1, i.e.,
increasing transversal relaxation rates with decreasing temper-
ature, thus suggesting an analogous relaxation mechanism. The
electron contribution to nuclear relaxation is then given by the
interaction between the time-averaged electron magnetic mo-
ment and the nucleus. Amazingly, experimental T1 values do
not significantly differ for high-spin nickel(II) complexes, when
going from the dimer to the monomer. On the other hand,
transversal relaxation rates are 1 order of magnitude longer in
the monomer m(BPh4) than in the dimer a(BPh4)2. These results
suggest then different contributions to T2 relaxation pathways
in the dimer vs the monomer. The possibility that the signals
are narrower in the dimer due to magnetic coupling is unlikely,
because the small ferromagnetic coupling allows all spin-states
to be largely populated in the temperature range of interest.
Implications of intrinsic zero-field splitting, arising indepen-
dently from each nickel atom, and overall S′ ) 2, 1, or 0 zero-
field splitting can be considered next. Magnetic anisotropy in
the dimer, which presents a pseudo-C2 axis, is expected to differ
from the anisotropy in the corresponding monomer that contains
a distorted octahedral nickel atom. A stronger zero-field splitting
Conclusion
The aim of controlling the preparation of six-coordinated
nickel(II) amine-thioether complexes has been achieved by
using the cyanoethyl-protected thiolate in the N3SR ligand series.
Reaction with Ni(OAc)2‚4H2O in alcohol allows the simple and
controlled preparation of both mono- and dinuclear nickel(II)
complexes. This synthetic strategy has provided a well-
characterized µ-disulfide dinuclear NiII complex a(BPh4)2. A
second complex m(BPh4) has been isolated from the dinuclear
complex synthesis and spectroscopically characterized as a
mononuclear NiII octahedral complex. Our data demonstrate,
1
for the first time, that H NMR spectra can be easily obtained
for weakly ferromagnetically spin-coupled dinuclear octahedral
NiII centers. In this particular case, despite very short relaxation
times, a complete assignment strategy has been performed. This
allowed further establishment of the main relaxation pathways
through temperature-dependence studies. Analysis of chemical
shift behavior vs temperature demonstrated the occurrence of
the weak spin-spin coupling constant and/or zero-field splitting
in these complexes. The precise determination of these param-
eters is hazardous in this case, due to the relative insensitivity
of magnetic susceptibility in the accessible temperature range.
Nevertheless, the established procedure, as well as the NMR
feasibility, would then open large perspectives for the study of
an increasing number of high-spin nickel(II) dimers. Applying
it to strongly antiferromagnetically coupled systems would
provide essential data, to get further insight into understanding
the relaxation mechanisms in dimers and to compare it to the
one in monomers. From the structural point of view, these results
also suggest that 1H NMR in NiII high-spin species may be used
for wild-type or engineered dinuclear nickel metalloproteins to
probe the local coordination environment around the metallic
center.
(57) Clementi, V.; Luchinat, C. Acc. Chem. Res. 1998, 31, 351-361.

8066 J. Am. Chem. Soc., Vol. 123, No. 33, 2001
Belle et al.
Acknowledgment. The authors are grateful to Dr. Ste´phane
Me´nage and Dr. Serge Gambarelli for their interest in this work
and fruitful discussions.
) f(T) for a(BPh4)2 (Figure 2S); 1H COSY spectrum of b(PF6)2
and c(PF6)2 (Figures 3S and 4S); resonance assignment,
relaxation properties, and dependence in temperature of a(PF6)2,
b(PF6)2 (Table 2S), and c(PF6)2 (Table 3S); Curie law deviations
due to D and J; and linear least-squares fit of experimental
relaxation rate vs reciprocal temperature of Py1-H2 resonance
for a(BPh4)2 in CD2Cl2 (PDF). An X-ray crystallographic file
for a(BPh4)2 (CIF). This material is available free of charge
via the Internet at http://pubs.acs.org.
Supporting Information Available: Experimental details
relative to the syntheses of 6, 7, and 9 and detailed characteriza-
tion of compounds 1 to 4 and 6 to 9; syntheses and physico-
chemical properties of the complexes a(PF6)2, b(PF6)2, c(PF6)2,
and m(PF6); crystallographic experimental details for a(BPh4)2
(Table 1S) and an ORTEP plot showing the entity a2 from
complex a(BPh4)2 (Figure 1S); magnetic susceptibility data øT
JA010342K