Autoionization of homogeneous nickel(II) diphosphane hydrogenation catalysts. An NMR study and crystal structures of [Ni(o-MeO-dppe)I2] and [Ni(o-MeO-dppe)2](PF6)2

Article abstract of DOI:10.1021/ic0006955

The synthesis of a number of nickel(II) complexes containing the didentate phosphane ligand 1,2-bis(di(o-methoxyphenyl)phosphino)ethane (o-MeO-dppe) is reported. Two types of complexes have been synthesized, i.e., the mono(chelate) complex (1) of the general formula [Ni(o-MeO-dppe)X₂] (where X = Cl, Br or I) and the bis(chelate) complex (2) of the general formula [Ni(o-MeO-dppe)₂] Y₂ (where Y = PF₆ or trifluoroacetate (TFA)). These complexes have been characterized using electronic absorption and NMR spectroscopy. The structures of the mono(chelate) complex [Ni(o-MeO-dppe)I₂] (1c) and of the bis(chelate) complex [Ni(o-MeO-dppe)₂](PF₆)₂ (2e) have been determined by X-ray crystallography. [Ni(o-MeO-dppe)I₂] crystallizes in the monoclinic space group P2₁/c with Z = 4, a = 12.1309(1) A, b = 16.5759(3) A, c = 17.6474(2) A, β = 119.3250(10)°. [Ni(o-MeO-dppe)₂](PF6)₂ crystallizes in the monoclinic space group C2/c with Z = 4, a = 22.5326(3) A, b = 13.6794(2) A, c = 21.7134(3) A, β = 107.1745(7)°. In both structures the nickel ion is in a square-planar geometry with a NiP₂I₂ and NiP₄ chromophore, respectively. Using ¹H and ³¹P{¹H} NMR spectroscopy the behavior of the complexes in various solvents has been studied. It appears that in solution these nickel complexes are involved in an autoionization equilibrium: 2[Ni(o-MeO-dppe)X₂] ? [Ni(o-MeO-dppe)₂]²⁺ + [ NiX₄ ]^2-. The ionized complex (3) consists of a cationic unit in which a nickel atom is surrounded by two didentate phosphane ligands, and an anionic unit that stoichiometrically consists of a nickel atom and four anions. The position of the autoionization equilibrium is highly dependent on the anion and the solvent used. In a polar solvent in combination with weakly coordinating anions only the ionized complex is observed, whereas in an apolar solvent in combination with coordinating anions only the mono(chelate) complex occurs. A comparison of the behavior of o-MeO-dppe with its unsubstituted analogue dppe in combination with nickel(II) acetate using ³¹P{¹H} NMR spectroscopy shows that the latter is more readily oxidized.

Full text of DOI:10.1021/ic0006955

Inorg. Chem. 2001, 40, 2073-2082
2073
Autoionization of Homogeneous Nickel(II) Diphosphane Hydrogenation Catalysts. An NMR
Study and Crystal Structures of [Ni(o-MeO-dppe)I2] and [Ni(o-MeO-dppe)2](PF6)2
Ingrid M. Angulo,† Elisabeth Bouwman,*,† Martin Lutz,‡ Wilhelmus P. Mul,*,§ and
Anthony L. Spek‡
Leiden Institute of Chemistry, Gorlaeus Laboratories, Leiden University, PO Box 9502,
2300 RA Leiden, The Netherlands, Shell Research and Technology Center Amsterdam,
Shell International Chemicals B.V., PO Box 38000, 1030 BN Amsterdam, The Netherlands, and
Department of Crystal and Structural Chemistry, Bijvoet Center for Biomolecular Research,
Utrecht University, Utrecht, The Netherlands
ReceiVed June 27, 2000
The synthesis of a number of nickel(II) complexes containing the didentate phosphane ligand 1,2-bis(di(o-
methoxyphenyl)phosphino)ethane (o-MeO-dppe) is reported. Two types of complexes have been synthesized,
i.e., the mono(chelate) complex (1) of the general formula [Ni(o-MeO-dppe)X2] (where X ) Cl, Br or I) and the
bis(chelate) complex (2) of the general formula [Ni(o-MeO-dppe)2]Y2 (where Y ) PF6 or trifluoroacetate (TFA)).
These complexes have been characterized using electronic absorption and NMR spectroscopy. The structures of
the mono(chelate) complex [Ni(o-MeO-dppe)I2] (1c) and of the bis(chelate) complex [Ni(o-MeO-dppe)2](PF6)2
(2e) have been determined by X-ray crystallography. [Ni(o-MeO-dppe)I2] crystallizes in the monoclinic space
group P21/c with Z ) 4, a ) 12.1309(1) Å, b ) 16.5759(3) Å, c ) 17.6474(2) Å, â ) 119.3250(10)°. [Ni(o-
MeO-dppe)2](PF6)2 crystallizes in the monoclinic space group C2/c with Z ) 4, a ) 22.5326(3) Å, b ) 13.6794(2)
Å, c ) 21.7134(3) Å, â ) 107.1745(7)°. In both structures the nickel ion is in a square-planar geometry with a
NiP2I2 and NiP4 chromophore, respectively. Using 1H and 31P{1H} NMR spectroscopy the behavior of the complexes
in various solvents has been studied. It appears that in solution these nickel complexes are involved in an
autoionization equilibrium: 2[Ni(o-MeO-dppe)X2] H [Ni(o-MeO-dppe)2]2+ + [“NiX4”]2-. The ionized complex
(3) consists of a cationic unit in which a nickel atom is surrounded by two didentate phosphane ligands, and an
anionic unit that stoichiometrically consists of a nickel atom and four anions. The position of the autoionization
equilibrium is highly dependent on the anion and the solvent used. In a polar solvent in combination with weakly
coordinating anions only the ionized complex is observed, whereas in an apolar solvent in combination with
coordinating anions only the mono(chelate) complex occurs. A comparison of the behavior of o-MeO-dppe with
its unsubstituted analogue dppe in combination with nickel(II) acetate using 31P{1H} NMR spectroscopy shows
that the latter is more readily oxidized.
Introduction
silylation, and hydrocyanation.7 Recently, we reported that
selected nickel(II) diphosphane complexes are able to hydro-
genate homogeneously 1-octene to n-octane.8 Nickel(II) com-
plexes of the o-methoxy-substituted didentate phosphane ligands9
1,2-bis(di(o-methoxyphenyl)phosphino)ethane (o-MeO-dppe) or
1,3-bis(di(o-methoxyphenyl)phosphino)propane (o-MeO-dppp)
(see Scheme 1) show catalytic hydrogenation activity in contrast
to nickel(II) complexes containing the nonsubstituted didentate
phosphane ligands 1,2-bis(diphenylphosphino)ethane (dppe) or
1,3-bis(diphenylphosphino)propane (dppp), which do not show
any catalytic hydrogenation activity at all.8 The origin of the
effect of the o-methoxy group in the diphosphane ligands
o-MeO-dppe and o-MeO-dppp compared to dppe and dppp on
the hydrogenation activity could be steric and/or electronic. The
Although nickel is widely used as a heterogeneous catalyst
for hydrogenation reactions (Raney nickel), homogeneous
hydrogenation catalysts containing nickel are scarcely found in
the literature.1-6 In fact, it is disputable whether the reported
nickel complexes yield really homogeneous hydrogenation
catalysts or that catalytic hydrogenation activity is due to
formation of colloidal (heterogeneous) nickel particles. Yet,
homogeneous nickel catalysts are known for related reactions
like isomerization, oligomerization, polymerization, hydro-
* To whom correspondence should be addressed. E-mail: bouwman@
chem.leidenuniv.nl and pim.w.p.mul@opc.shell.com.
† Leiden University.
‡ Utrecht University.
§ Shell International Chemicals B.V.
(7) Applied Homogeneous Catalysis with Organometallic Compounds;
Cornils, B., Herrmann, W. A., Eds.; VCH Publishers: New York,
1996; Sections 2.3, 2.4, 2.5, and 3.2.14.
(1) Emken, E. A.; Frankel, E. N.; Butterfield, R. O. J. Am. Oil Chem.
Soc. 1966, 43, 14.
(2) Itatani, H.; Bailar, J. C., Jr. J. Am. Chem. Soc. 1967, 89, 1600.
(3) Abley, P.; McQuillin, F. J. Discuss. Faraday Soc. 1968, 46, 31.
(4) Henrici-Olive´, G.; Olive´, S. J. Mol. Catal. 1975/76, 1, 121.
(5) Thangaraj, T.; Vancheesan, S.; Rajaram, J.; Kuriacose, J. C. Indian J.
Chem., Sect. A 1980, 19, 404.
(6) Chatterjee, D.; Bajaj, H. C.; Halligudi, S. B.; Bhatt, K. N. J. Mol.
Catal. 1993, 84, L1.
(8) Angulo, I. M.; Kluwer, A. M.; Bouwman, E. Chem. Commun. 1998,
2689.
(9) Abbreviations: TFA ) trifluoroacetate, dppe ) 1,2-bis(diphenylphos-
phino)ethane, dppp ) 1,3-bis(diphenylphosphino)propane, o-MeO-
dppe ) 1,2-bis(di(o-methoxyphenyl)phosphino)ethane, o-MeO-dppp
) 1,3-bis(di(o-methoxyphenyl)phosphino)propane, p-MeO-dppp )
1,3-bis(di(p-methoxyphenyl)phosphino)propane.
10.1021/ic0006955 CCC: $20.00 © 2001 American Chemical Society
Published on Web 03/24/2001

2074 Inorganic Chemistry, Vol. 40, No. 9, 2001
Angulo et al.
showing cationic and anionic complex fragments, respectively. Samples
of nickel complexes were dissolved in methanol and introduced by
infusion in a methanol flow. Infrared spectra were recorded with a
Perkin-Elmer Paragon 1000 IR spectrophotometer equipped with a
Golden Gate ATR device, using the reflectance technique (4000-300
cm-1). Solid-state electronic absorption spectra were obtained on a
Perkin-Elmer Lambda 900 spectrophotometer using the diffuse reflec-
tance technique, with MgO as a reference. The electronic absorption
spectra in solution were taken using 1 cm path length cells at ambient
temperature, with the used solvent as the reference. The 1H NMR and
31P{1H} NMR spectra were recorded on a Varian VXR-200 spectrom-
eter at 200.06 and 80.98 MHz, respectively, or on a Bruker WM-300
spectrometer at 300.13 and 121.50 MHz, respectively. The high-pressure
NMR experiments were conducted on a Varian INOVA-400 spectrom-
Scheme 1
Scheme 2
1
eter operating at 399.97 MHz for H NMR and at 161.91 MHz for
31P{1H} NMR. The 31P and 1H chemical shifts were recorded in δ units
relative to triphenylphosphane (-6 ppm) and the deuterated solvent
fact that nickel acetate in combination with a para analogue of
the methoxy-substituted ligands (p-MeO-dppp) does not show
any catalytic hydrogenation activity8 indicates that the positive
effect of the o-methoxy groups on the hydrogenation activity
is predominantly steric.
1
(CD2Cl2, 5.24 ppm; CD3OD, 3.35 ppm), respectively. The H NMR
and 31P{1H} NMR data of the ligand and the nickel complexes in
CD2Cl2 are collected in Table 1.
Syntheses. All chemicals were reagent grade and were used as
received or synthesized as described below. Reactions were carried out
under an atmosphere of dry argon or dry nitrogen, unless stated
otherwise, using standard Schlenk techniques. Tris(o-methoxyphenyl)-
phosphane was synthesized according to a literature procedure.13
Generally the complexes were obtained in yields of 30-50%; no
attempts have been undertaken to optimize the yields.
A hypothesis to explain the difference in catalytic activity of
nickel complexes containing o-MeO-dppe versus dppe concerns
the autoionization equilibrium depicted in Scheme 2. Two
mono(chelate) complexes (1) react to form an ionized complex
(3), which consists of a bis(chelate) cationic unit and an anionic
unit formally containing one nickel atom and four anions. This
autoionization reaction was already observed by Jarrett et al.,10
while studying the solid-state and solution chemistry of Ni(II)
complexes containing the ligand dppe. It is assumed that the
resulting ionized complex (3) is unable to act as a catalyst in
the hydrogenation of 1-octene. This assumed catalytic inactivity
is rationalized by the lack of available sites for coordination of
the reactants in the cationic unit, whereas the anionic unit in 3,
containing “naked” nickel (i.e., not containing a didentate
phosphane ligand), is only expected to yield an active catalyst
at higher temperatures, when colloidal nickel particles are
formed.11
It was surmised that the o-methoxy groups could prevent the
formation of the inactive ionized complex due to their bulkiness,
hampering the complexation of a second didentate phosphane
ligand to the nickel ion. Therefore, a study was undertaken to
investigate the nature of species in solution for both o-methoxy-
containing ligands o-MeO-dppe and o-MeO-dppp. This manu-
script describes the results of the NMR investigation on the
autoionization equilibrium of nickel complexes containing the
didentate phosphane ligand o-MeO-dppe. The results of the
complementary investigation concerning o-MeO-dppp as the
ligand will be reported elsewhere,12 as it showed quite different
behavior which is not directly related with that described in
this manuscript.
1,2-Bis(di(o-methoxyphenyl)phosphino)ethane (o-MeO-dppe). This
synthesis was performed according to the procedure earlier described
for the synthesis of 1,3-bis(di(o-methoxyphenyl)phosphino)propane.14
The synthesis was performed under an argon atmosphere. It is important
that absolutely no air or water is present during the reaction. A 250
mL round-bottom flask was equipped with an argon inlet, a condenser,
an ammonia inlet, and a magnetic stirrer bar. While the flask was cooled
with an acetone/dry ice mixture, 100 mL of dry ammonia was
condensed into the flask. Traces of water were removed from the
ammonia by addition of a small piece of sodium metal (10 mg). To
the liquid ammonia was added 60 mmol (1.4 g) of sodium metal through
the condenser. After 15 min 30 mmol (10.6 g) of tris(o-methoxyphenyl)-
phosphane was slowly added through the condenser to the blue solution,
followed by 25 mL of dry and peroxide-free THF. The blue mixture
turned deep orange/red within 1 h of stirring. After 5 h of stirring,
while the temperature was kept between -78 and -50 °C, 30 mmol
(1.6 g) of anhydrous ammonium chloride was added. After 15 min 15
mmol (1.5 g) of 1,2-dichloroethane in 15 mL of THF was added. The
temperature was allowed to rise to room temperature, and the ammonia
was allowed to evaporate overnight. The remaining part of the synthesis
of the ligand could be performed in air. The solvent was removed in
vacuo, and the residue was dissolved in dichloromethane, to which an
aqueous solution of ammonium chloride was then added. The organic
layer was separated, dried with sodium sulfate, and filtered. The organic
solvent was removed in vacuo. The white/yellow oil was treated with
methanol, resulting in a white precipitate. The white solid was collected
by filtration and washed with methanol. Drying in vacuo gave 6.3 g
(12.2 mmol), 81% yield; mp 180-4 °C. Anal. Calcd for C30H32O4P2
(fw ) 518.5): C, 69.49; H, 6.22. Found: C, 68.15; H, 6.15. MS
(ESI): m/z 519 [M + H]+.
Experimental Section
[Ni(o-MeO-dppe)Cl2] (1a). Method A. Ni(OAc)2‚4H2O (50 mg,
0.20 mmol) was mixed with an equimolar amount of o-MeO-dppe (104
mg, 0.20 mmol) in 20 mL of ethanol. The mixture was refluxed for
31/2 hours, and a clear orange solution was obtained. After the solution
was partly cooled, 0.5 mL of hydrochloric acid was added (1 M in
Et2O, 0.5 mmol). After several minutes an orange precipitate had
formed. The solid was collected by filtration and dried in air.
Method B. NiCl2‚6H2O (370 mg, 1.57 mmol) was dissolved in 10
mL of ethanol and mixed with o-MeO-dppe (520 mg, 1.00 mmol)
dissolved in 10 mL of chloroform. Immediately a dark orange solution
General Procedures. Melting points were determined on a Bu¨chi
apparatus and are uncorrected. Elemental analyses were performed on
a Perkin-Elmer 2400 series II analyzer. Mass spectra were obtained
with a Finnigan MAT TSQ-700 mass spectrometer equipped with a
custom-made electrospray interface (ESI). Spectra were collected by
constant infusion of the analyte dissolved in methanol/water with 1%
HOAc. Nickel complexes were analyzed by electrospray mass spec-
trometry using a Micromass Quattro triple-quadrupole mass spectrom-
eter. Mass spectra were collected both in positive and in negative mode,
(10) Jarrett, P. S.; Sadler, P. J. Inorg. Chem. 1991, 30, 2098.
(11) Kelber, C. Ber. Dtsch. Chem. Ges. 1917, 50, 1509.
(12) Angulo, I. M.; Bouwman, E.; Lok, S. M.; Lutz, M.; Mul, W. P.; Spek,
A. L. Eur. J. Inorg. Chem., in press.
(13) Brandsma, L.; Verkruijsse, H. D. Synth. Comm. 1990, 20, 2273.
(14) Budzelaar, P. H. M.; van Doorn, J. A.; Meijboom, N. Recl. TraV. Chim.
Pays-Bas 1991, 110, 420.

Nickel(II) Diphosphane Hydrogenation Catalysts
Inorganic Chemistry, Vol. 40, No. 9, 2001 2075
Table 1. 1H NMR and 31P{1H} NMR Data of o-MeO-dppe and Nickel Complexes in CD2Cl2 at 25 °C
1H NMR
-C6H4
-OCH3
3.64 (s, 12H)
-CH2CH2-
2.01 (t, 4H)
31P{1H} NMR
-31.67
o-MeO-dppe
6.78 (t, 8H), 6.95 (m, 4H),
7.22 (t, 4H)
[Ni(o-MeO-dppe)Cl2] (1a)
[Ni(o-MeO-dppe)Br2] (1b)
[Ni(o-MeO-dppe)I2] (1c)
[Ni(o-MeO-dppe)2](TFA)2 (2d)
6.90 (d, 4H), 6.96 (t, 4H),
7.48 (t, 4H), 8.13 (m, 4H)
6.88 (d, 4H), 6.98 (t, 4H),
7.46 (t, 4H), 8.13 (m, 4H)
6.88 (d, 4H), 6.97 (t, 4H),
7.43 (t, 4H), 8.06 (m, 4H)
5.77 (m, 2H), 6.49 (t, 2H),
6.69 (d, 2H), 6.93 (d, 2H),
7.23 (m, 4H), 7.72 (t, 2H),
8.58 (m, 2H)
3.56 (s, 12H)
3.56 (s, 12H)
3.57 (s, 12H)
2.39 (d, 4H)
2.35 (d, 4H)
2.39 (d, 4H)
58.24
66.03
75.63
3.42 (s, 6H), 3.47 (s, 6H) 2.23 (d, 2H), 2.83 (m, 2H) 52.70
[Ni(o-MeO-dppe)2](PF6)2 (2e)
5.75 (m, 2H), 6.48 (t, 2H),
6.67 (d, 2H), 6.91 (d, 2H),
7.21 (m, 4H), 7.70 (t, 2H),
8.59 (m, 2H)
3.41 (s, 6H), 3.46 (s, 6H) 2.23 (d, 2H), 2.83 (m, 2H) 54.40, -143.9
(septet, PF6)
[Ni(o-MeO-dppe)(TFA)2] (1d) or
[Ni(o-MeO-dppe)2][Ni(TFA)4] (3d)
a
52.70, 53.76
a In CD2Cl2 a mixture of 1d and 3d is observed, see Table 4 and discussion.
was obtained, and an orange solid precipitated within a few minutes.
The solid was collected by filtration and dried in air.
extinction coefficients (L mol-1 cm-1)): νmax 34 000 cm-1 (34 200),
29 600 cm-1 (16 000), and 28 000 cm-1 (shoulder).
[Ni(o-MeO-dppe)(TFA)2] (1d) or [Ni(o-MeO-dppe)2][Ni(TFA)4]
(3d). Ni(OAc)2‚4H2O (72 mg, 0.29 mmol) and an equimolar amount
of o-MeO-dppe (150 mg, 0.29 mmol) were mixed in 50 mL ethanol.
This mixture was brought to reflux temperature, and after 16 h the
bright orange solution was cooled. After the mixture was concentrated
to approximately 5 mL under vacuum, 45 mL of toluene was added
and the mixture brought to reflux temperature for 15 min to afford an
orange/red solution. When the reaction mixture was cooled to 40 °C,
0.07 mL of trifluoroacetic acid was added (0.92 mmol). After removal
of half of the solvent under vacuum, an orange precipitate was formed
overnight, which was isolated by filtration and washed twice with
hexane. Anal. Calcd for C34H32F6NiO8P2 (1d) (fw ) 803.25) or
C68H64F12Ni2O16P4 (3d) (fw ) 1606.49): C, 50.84; H, 4.02 Found: C,
48.52; H, 3.80. Diffuse reflectance electronic absorption of the solid:
three weak signals were observed at ν 8500, 13 500, and 20 800 cm-1
(shoulder). Following this procedure, it is inevitable that some physical
mixture may be obtained or some impurities are present, explaining
the rather poor elemental analysis. The compound, however, is essential
for the NMR experiments described in this manuscript.
X-ray Crystal Structure Determinations of [Ni(o-MeO-dppe)I2]
(1c) and Ni[(o-MeO-dppe)2](PF6)2 (2e). Intensities were measured on
a Nonius KappaCCD diffractometer with rotating anode (Mo KR, λ )
0.710 73 Å) at 150 K. The structures were solved with Patterson
methods (DIRDIF-9715) and refined with the program SHELXL-9716
against F2 of all reflections up to a resolution of (sin θ/λ)max ) 0.65
Å-1. Non-hydrogen atoms were refined freely with anisotropic displace-
ment parameters, and hydrogen atoms were refined as rigid groups.
The drawings, structure calculations, and checking for higher symmetry
were performed with the program PLATON.17 Further experimental
details are given in Table 2.
Anal. Calcd for C30H32Cl2NiO4P2 (fw ) 648.13): C, 55.60; H, 4.98.
Found: C, 54.95; H, 4.55. Diffuse reflectance electronic absorption of
the solid: νmax 28 100 and 21 200 cm-1. Electronic absorption in CH2Cl2
(molar extinction coefficients (L mol-1 cm-1)): νmax 31 900 cm-1
(19 700) and 21 300 cm-1 (1200).
[Ni(o-MeO-dppe)Br2] (1b). NiBr2 (220 mg, 1.00 mmol) was
dissolved in 10 mL of ethanol, and o-MeO-dppe (520 mg, 1.00 mmol)
was dissolved in 10 mL of chloroform. The solution containing NiBr2
was filtered before it was added to the ligand solution. When both
solutions were mixed, a dark red solution was obtained. After several
minutes a dark red solid precipitated from the solution. The solid was
collected by filtration and dried in air. Anal. Calcd for C30H32Br2NiO4P2
(fw ) 737.03): C, 48.89; H, 4.38. Found: C, 49.66; H, 4.48. Diffuse
reflectance electronic absorption of the solid: νmax 30 100 and 20 900
cm-1
.
[Ni(o-MeO-dppe)I2] (1c). Ni(OAc)2‚4H2O (83 mg, 0.33 mmol) was
dissolved in 10 mL of ethanol, and o-MeO-dppe (172 mg, 0.33 mmol)
was dissolved in 10 mL of chloroform. When both solutions were
mixed, an orange solution was obtained. Then 0.08 mL of concentrated
hydrogen iodide (50% in water, 0.64 mmol) was added and a purple
solution was obtained. After some days purple crystals precipitated from
the solution. These crystals were suitable for X-ray diffraction. Anal.
Calcd for C30H32I2NiO4P2 (fw ) 831.03): C, 43.36; H, 3.88. Found:
C, 43.17; H, 3.53. Diffuse reflectance electronic absorption of the
solid: νmax 28 200 and 19 200 cm-1
.
[Ni(o-MeO-dppe)2](TFA)2 (2d). Ni(OAc)2‚4H2O (130 mg, 0.52
mmol) was mixed with o-MeO-dppe (530 mg, 1.02 mmol) in a mixture
of 10 mL of dichloromethane and 1 mL of methanol. When a clear
orange solution was obtained, 0.1 mL of trifluoroacetic acid was added
(1.3 mmol). The solvent was partly evaporated and diethyl ether was
added, resulting in a yellow precipitate. The solvent was decanted, and
the solid was dried in air. Anal. Calcd for C64H64F6NiO12P4‚CH2Cl2
(fw ) 1406.71): C, 55.50; H, 4.73 Found: C, 55.01; H, 4.59. Diffuse
The crystal structure of 2e contains large voids (520 Å3) filled with
disordered solvent molecules (methanol and dichloromethane). Their
contribution to the structure factors was secured by back-Fourier
transformation (program PLATON,17 CALC SQUEEZE, 123 e-/unit
cell).
NMR Studies. Both 1H NMR and 31P{1H} NMR spectra of the
above-mentioned complexes were taken in the solvents methanol-d4,
reflectance electronic absorption of the solid: νmax 28 900 cm-1
.
[Ni(o-MeO-dppe)2](PF6)2 (2e). Ni(OAc)2‚4H2O (125 mg, 0.50
mmol) was dissolved in 15 mL of ethanol, and o-MeO-dppe (570 mg,
1.10 mmol) was dissolved in 15 mL of chloroform. When both solutions
were mixed, a red solution was obtained. Then NH4PF6 (160 mg, 0.98
mmol) dissolved in 5 mL of ethanol was added and an orange solution
was obtained. After several days an orange solid precipitated from the
solution. Recrystallization from a methanol/dichloromethane mixture
yielded crystals which were suitable for X-ray diffraction. Anal. Calcd
for C60H64F12NiO8P6‚(CHCl3)2 (fw ) 1624.43): C, 45.84; H, 4.10.
Found: C, 46.05; H, 3.90. Diffuse reflectance electronic absorption of
the solid: νmax 28 000 cm-1. Electronic absorption in CH2Cl2 (molar
(15) Beurskens, P. T.; Admiraal, G.; Beurskens, G.; Bosman, W. P.; Garcia-
Granda, S.; Gould, R. O.; Smits, J. M. M.; Smykalla, C. The DIRDIF97
program system. Technical Report of the Crystallography Laboratory;
University of Nijmegen: Nijmegen, The Netherlands; 1997
(16) Sheldrick, G. M. SHELXL97. Program for crystal structure refinement;
University of Go¨ttingen: Go¨ttingen, Germany, 1997.
(17) Spek, A. L. PLATON. A multipurpose crystallographic tool; Utrecht
University: Utrecht, The Netherlands, 2000.

2076 Inorganic Chemistry, Vol. 40, No. 9, 2001
Angulo et al.
Table 2. Crystal Data and Structure Refinement Details for
[Ni(o-MeO-dppe)I2] (1c) and [Ni(o-MeO-dppe)2](PF6)2 (2e)
Scheme 3
1c
2e
formula
C30H32I2NiO4P2
C60H64NiO8P4‚2PF6 +
solvent
fw
831.01
1385.64a
space group
a (Å)
b (Å)
c (Å)
â (deg)
V (Å3)
Z
P21/c (No. 14)
12.1309(1)
16.5759(3)
17.6474(2)
119.3250(10)
3093.82(7)
4
1.784
2.76
0.0278/0.0362
0.0672/0.0703
C2/c (No. 15)
22.5326(3)
13.6794(2)
21.7134(3)
107.1745(7)
6394.34(15)
4
synthesized. Using a ligand-to-metal ratio of two and more
weakly coordinating anions, bis(o-MeO-dppe) complexes of the
type [Ni(o-MeO-dppe)2]Y2 (Y ) trifluoroacetate (TFA) (2d)
or PF6 (2e)) were obtained. Finally, an attempt was undertaken
to synthesize the mono(chelate) complex [Ni(o-MeO-dppe)-
(TFA)2] (1d). As this complex contains the weakly coordinating
anion TFA, the formulation of this compound could also be
given as [Ni(o-MeO-dppe)2][Ni(TFA)4] (3d), being the ionized
complex (see below). In order to synthesize this complex,
nickel(II) acetate and o-MeO-dppe were mixed in a 1:1 ratio in
the polar solvent ethanol, leading to the ionized complex [Ni(o-
MeO-dppe)2][Ni(OAc)4]. With the aim to form complex 1d most
of the ethanol was evaporated and replaced by the apolar solvent
toluene, before trifluoroacetic acid was added to replace the
acetate anions.
F
calc (g cm-3
)
1.439a
µ (mm-1
)
0.54a
0.0340/0.0400
0.0935/0.0974
R1b (obs/all reflns)
wR2c (obs/all reflns)
a Derived values do not contain the contribution of the disordered
solvent molecules. b R1 ) ∑||Fo| - |Fc||/∑|Fo|. c wR2 ) [∑w(Fo
-
+
2
Fc2)2/∑w(Fo )2]1/2, w ) 1/[σ2(Fo ) + (AP)2 + BP] where P ) (Fo
2
2
2
2Fc2)/3, A ) 0.0355 (1c) or 0.0530 (2e), and B ) 2.6536 (1c) or 4.8452
(2e).
dichloromethane-d2, acetone-d6, and toluene-d8. All deuterated solvents
were reagent grade and were used as received. For the high-pressure
NMR experiments, the sapphire NMR tube was pressurized with 10
bar of dihydrogen gas.
For the in situ experiments Ni(OAc)2‚4H2O (5.0 mg, 0.02 mmol)
and the ligand, o-MeO-dppe (9.1 mg, 0.02 mmol) or dppe (7.0 mg,
0.02 mmol), were mixed in the NMR tube. The nickel salt and the
ligand were weighed into an NMR tube, which was then evacuated
and filled with nitrogen three successive times. Under a nitrogen stream
the deuterated solvent was added. The solution was analyzed when,
sometimes after filtration, a clear solution was obtained.
For the analysis of a spent catalyst, a hydrogenation experiment was
carried out. Ni(OAc)2‚4H2O (25 mg, 0.1 mmol) and o-MeO-dppe (57
mg, 0.11 mmol), o-MeO-dppp (59 mg, 0.11 mmol), or dppp (45 mg,
0.11 mmol) were mixed in 10 mL of methanol and 10 mL of
dichloromethane. To the clear solution was added 7.5 mL of 1-octene
(50 mmol), and the solution was loaded into the autoclave. The
hydrogenation experiment was performed at 50 bar of hydrogen pressure
and 50 °C. After 30 min, the solution was extracted from the autoclave
into a round-bottom flask under an inert atmosphere. The solvent was
evaporated, and the resulting solid was dissolved in deuterated methanol
for NMR analysis.
Characterization of the Synthesized Complexes in the
Solid Phase. The infrared spectra of all synthesized nickel(II)
complexes are very similar to each other and also to that of the
free ligand o-MeO-dppe. Unfortunately, because of the lack of
differences, infrared spectroscopy cannot be used to distinguish
a mono(chelate) complex (1) from a bis(chelate) complex (2).
The diffuse reflectance absorption spectra of both types of
complexes proved more useful, as these show some clear
differences. In the absorption spectrum of a mono(chelate)
complex [Ni(o-MeO-dppe)X2] (1) only one band at around
20 000 cm-1 is observed. Strong bands at higher energies are
attributed to charge-transfer transitions. The band around 20 000
1
1
cm-1, which is ascribed to the A1 f B2 transition, is typical
for a square-planar nickel complex containing one didentate
phosphane ligand.18 In the absorption spectrum of a bis(chelate)
complex [Ni(o-MeO-dppe)2]Y2 (2) only charge-transfer bands
are observed. It is known that the bands characteristic of the
NiP4 chromophore should lie at higher energies than those for
the NiP2X2 chromophore.19 The bands due to NiP4 apparently
fall in the region obscured by strong charge-transfer absorptions.
As no other bands are observed in the absorption spectra of
these bis(chelate) complexes, they are tentatively ascribed
square-planar geometry.
Results and Discussion
Syntheses. The ligand o-MeO-dppe was prepared using the
method based on the earlier reported reductive splitting of tris(o-
methoxyphenyl)phosphane (P(o-MeO-Ph)3) using sodium in
ammonia.14 It is important that absolutely no water is present
during this reductive splitting. Even traces of water lead to an
incomplete splitting of P(o-MeO-Ph)3 and contamination of
the resulting product with starting material P(o-MeO-Ph)3,
which cannot be removed. The 31P{1H} NMR spectrum of the
ligand in CD2Cl2 shows a singlet at -31.7 ppm. The bridging
ethylene protons appear as a virtual triplet, due to coupling with
the two phosphorus atoms, at 2.0 ppm in the 1H NMR spectrum,
which further shows a singlet at 3.64 ppm, corresponding to
the protons of the o-methoxy groups. With the obtained ligand
a number of nickel(II) complexes were synthesized varying the
anions and with various ligand-to-metal ratios. Although it is
known that in general phosphane ligands are sensitive toward
oxidation, the obtained complexes appeared to be stable in air.
An overview of the synthesized compounds is given in Scheme
3. With halide anions, mono(o-MeO-dppe) complexes of the
type [Ni(o-MeO-dppe)X2] (X ) Cl (1a), Br (1b), or I (1c)) were
In an attempt to synthesize the mono(chelate) complex [Ni(o-
MeO-dppe)(TFA)2] (1d), an orange complex was obtained, for
which elemental analysis showed the metal-to-ligand ratio to
be 1. Although this ratio is as expected for the desired mono-
(chelate) complex 1d, it also corresponds to the ionized complex
3d. In the diffuse reflectance absorption spectrum of the obtained
orange solid only three weak bands, tentatively assigned to an
octahedral complex,19 are observed, apart from the strong
charge-transfer bands. For 1d the band corresponding to the
mono(chelate) complex with an NiP2O2 chromophore is ex-
pected to lie at a higher energy than that of [Ni(o-MeO-dppe)-
Cl2], according to the established spectrochemical series.19 It is
even possible that this band will disappear under the charge-
(18) van Hecke, G. R.; Horrocks, W. DeW., Jr. Inorg. Chem. 1966, 5, 1968.
(19) Lever, A. B. P. Inorganic Electronic Spectroscopy; Elsevier: Am-
sterdam, 1983; Section 6.2.

Nickel(II) Diphosphane Hydrogenation Catalysts
Inorganic Chemistry, Vol. 40, No. 9, 2001 2077
Table 3. Selected Interatomic Distances, Angles, and Torsion Angles for [Ni(o-MeO-dppe)I2] (1c) and [Ni(o-MeO-dppe)2](PF6)2 (2e). A
Comparison Is Made with the Analogous Complexes with the Unsubstituted Ligand dppe, [Ni(dppe)Cl2]‚CH2Cl2 (Form B) (I)20,21 and
[Ni(dppe)2](NO3)2 (II)23 a,b
[Ni(dppe)Cl2] (I)
(form B)
[Ni(o-MeO-dppe)I2] (1c)
X ) I
Ni(o-MeO-dppe)2](PF6)2 (2e)
Distances (Å)
X ) Cl
[Ni(dppe)2](NO3)2 (II)
Ni1-X2
2.5442(4)
2.5214(5)
2.1789(9)
2.1776(7)
3.773(2)
3.636(2)
2.7913
2.195(2)
2.205(2)
2.157(2)
2.145(2)
Ni1-X3
Ni1-P4
Ni1-P4
2.2732(5)
2.2724(5)
3.9300(14)
3.9317(14)
2.7299
2.261(3)
2.256(3)
Ni1-P7
Ni1-P7
Ni1-O27
Ni1-O37
Ni1-H16
Ni1-H46
Ni1-O17
Ni1-O47
Ni1-H26
Ni1-H36
-
2.745
2.770
3.112
2.625
2.7786
2.6532
Angles (deg)
X2-Ni1-X3
X2-Ni1-P4
X2-Ni1-P7
X3-Ni1-P4
X3-Ni1-P7
P4-Ni1-P7
92.88(1)
91.29(2)
177.77(3)
175.56(2)
89.35(3)
86.48(3)
95.47(6)
89.01(6)
175.06(7)
175.19(6)
88.68(6)
86.93(6)
P7-Ni1-P4
P7-Ni1-P4a
96.61(1)
96.75
179.35(2)
180.00
P7-Ni1-P7a
P4-Ni1-P4a
83.34(2)
83.44(2)
83.25(12)
83.25(12)
Torsion Angles (deg)
Ni1-P4-C11-C12
Ni1-P4-C11-C12
Ni1-P4-C21-C22
Ni1-P7-C31-C32
Ni1-P7-C41-C42
P4-C5-C6-P7
165.5(2)
-70.9(3)
-66.0(3)
176.3(2)
48.5(2)
174.88
-85.66
-78.87
172.91
47.85
72.44(14)
68.26
Ni1-P4-C21-C22
Ni1-P7-C31-C32
Ni1-P7-C41-C42
P4-C5-C5a-P4a
P7-C6-C6a-P7a
-164.72(13)
-170.96(13)
74.66(15)
-55.15(13)
-51.85(14)
-166.46
-174.83
73.12
30.82(15)
30.82(15)
a Although the atom numbering and symmetry of the complexes [Ni(dppe)Cl2]‚CH2Cl2 (form B) (I) and [Ni(dppe)2](NO3)2 (II) are not the same
as those in our compounds, the values listed in the table are arranged in such a way that the like atoms are compared. b Values for [Ni(dppe)Cl2]‚CH2Cl2
(form B) (I) and [Ni(dppe)2](NO3)2 (II) which could not be found in the report were calculated, without standard deviations, with the program
PLATON17 using the atomic coordinates as retrieved from CCDC (I, FUJXUD; II, VASCIB).
transfer bands. Thus, unfortunately, even electronic absorption
spectroscopy is not helpful to assign unambiguously either
formulation 1d or 3d to the isolated compound, as for either of
the two compounds also the occurrence of a slight octahedral
distortion can be explained. In 1d this distortion could be caused
by a complex in which the TFA anions are coordinated in a
didentate fashion to the nickel ion, whereas in 3d the anionic
unit of the ionized complex could be the source of the bands
assigned to an octahedral complex.
Crystal Structure of [Ni(o-MeO-dppe)I2] (1c). A molecular
plot of the structure showing 50% probability ellipsoids is given
in Figure 1. Selected interatomic distances and angles are given
in Table 3. The present structure of [Ni(o-MeO-dppe)I2] (1c)
largely resembles that of the related structure of [Ni(dppe)Cl2]‚
CH2Cl2 (I).20,21 In Table 3 the relevant bond distances and angles
are compared, whereas in Figure 2a the two compounds are
compared in a graphical representation. The coordination
geometry around the nickel(II) center in both complexes is
essentially square-planar. In both structures the dihedral angle
between the X-Ni-X and P-Ni-P planes does not differ
considerably from zero (1.53° in 1c and 3.30° in I). The Ni-P
distances in 1c are slightly longer than those in I. The bite angle
P-Ni-P is essentially the same in the two compounds. It is
possible to distinguish the aryl rings as oriented either axially
(rings 1 and 4) or equatorially (rings 2 and 3) with respect to
the five-membered chelate ring. This orientation of the aryl rings
is essentially the same as in I, as can be seen in Figure 2a.The
o-methoxy groups of the axial aryl rings in 1c are pointing away
from the central nickel atom (torsion angles Ni1-P4-C11-
Figure 1. Displacement ellipsoid plot (50% probability) of [Ni(o-MeO-
dppe)I2] (1c) with atom-labeling scheme. Hydrogen atoms are omitted
for clarity.
C12 ) 165.5(2)° and Ni1-P7-C41-C42 ) 176.3(2)°), in
contrast to the o-methoxy groups of the equatorial aryl rings,
which are more pointing toward the nickel atom (torsion angles
Ni1-P4-C21-C22 ) -70.9(3)° and Ni1-P7-C31-C32 )
-66.0(3)°). However, the corresponding Ni‚‚‚O distances of
3.773(2) Å (O27) and 3.636(2) Å (O37) are too large to consider
an electrostatic interaction. In Figure 1 it can be seen that the
o-hydrogen atoms of the axial aryl groups are pointing toward
the nickel atom. The Ni‚‚‚H distances of 2.79 Å (H16) and 2.78
Å (H46) indicate the presence of either agostic interactions or
Ni‚‚‚H-C hydrogen bonds, a phenomenon which has been
(20) Spek, A. L.; van Eijck, B. P.; Jans, R. J. F.; van Koten, G. Acta
Crystallogr. 1987, C43, 1878.
(21) Busby, R.; Hursthouse, M. B.; Jarrett, P. S.; Lehmann, C. W.; Malik,
K. M. A.; Phillips, C. J. Chem. Soc., Dalton Trans. 1993, 3767.

2078 Inorganic Chemistry, Vol. 40, No. 9, 2001
Angulo et al.
Figure 3. Displacement ellipsoid plot (50% probability) of the cationic
unit of [Ni(o-MeO-dppe)2](PF6)2 (2e) with atom-labeling scheme.
Hydrogen atoms are omitted for clarity. Symmetry operation: 1 - x,
y, 0.5 - z.
Figure 2. (a) Crystal structures of [Ni(o-MeO-dppe)I2] (1c) (wire) and
of [Ni(dppe)Cl2] (I) (tube) drawn such that the Ni and P atoms from
the two structures are coincident. (b) Crystal structures of [Ni(o-MeO-
dppe)2](PF6)2 (2e) (wire) and of [Ni(dppe)2](NO3)2 (II) (tube) drawn
such that the Ni and P atoms from the two structures are coincident
(for clarity only one ligand of the bis(chelate) complexes is depicted).
(c) Crystal structures of 1c and of 2e (for clarity only one ligand is
depicted) drawn such that the Ni and P atoms from the two structures
are coincident. In all representations the perspective has the depicted
P atoms away from the viewer, and the two other not depicted
coordinating atoms toward the viewer.
whereas the o-methoxy groups of the equatorial aryl rings are
more pointing toward the nickel atom (torsion angles Ni1-
P4-C11-C12 ) 72.44(14)° and Ni1-P7-C41-C42 )
74.66(15)°). However, the corresponding Ni‚‚‚O distances of
3.9300(14) Å (O17) and 3.9317(14) Å (O47) are too large to
consider an electrostatic interaction. In Figure 3 it can be seen
that the o-hydrogen atoms of the axial aryl groups are pointing
toward the nickel atom. Again the Ni‚‚‚H distances of 2.73 Å
(H26) and 2.65 Å (H36) indicate the presence of either agostic
interactions or Ni‚‚‚H-C hydrogen bonds.22
Comparison of the Structures [Ni(o-MeO-dppe)I2] (1c)
and [Ni(o-MeO-dppe)2](PF6)2 (2e). In Figure 2c the orienta-
tions of the ligands in the structures of 1c and 2e are compared.
To facilitate the comparison of both structures only one of the
ligands in 2e is shown. It can be seen that in both structures
the ligand is oriented in the same way around the central nickel
atom, with alternating axial and equatorial orientation of the
aryl rings. The ligands in the bis(chelate) complex 2e are less
tightly bound than in the mono(chelate) complex 1c, which
follows from the slightly shorter Ni-P bonds in the latter. The
bite angle P-Ni-P in the bis(o-MeO-dppe) complex (83.34°)
is somewhat smaller than the one in the mono(o-MeO-dppe)
complex (86.48°). The steric crowding caused by the presence
of two ligands in the coordination sphere of the nickel ion in
the bis(o-MeO-dppe) complex forces the ligands to adopt a
slightly different geometry.
observed before for square-planar d8-ML4 metal complexes.22
In the crystal packing intermolecular stacking of aryl rings is
observed.
Crystal Structure of [Ni(o-MeO-dppe)2](PF6)2 (2e). A
molecular drawing of the structure of the complex cation
showing 50% probability ellipsoids is given in Figure 3. Selected
interatomic distances and angles are given in Table 3. The
structure of [Ni(o-MeO-dppe)2](PF6)2 (2e) resembles that of the
related structure of [Ni(dppe)2](NO3)2 (II).23 However, the
symmetry in 2e is different from that in II. The cation in II is
centrosymmetric around the nickel(II) center, whereas in the
cation of 2e an exact, crystallographic 2-fold symmetry axis
bisecting the ligands is present. In Table 3 the relevant bond
distances and angles are compared, whereas in Figure 2b the
two compounds are compared in a graphical representation. In
both complexes the coordination geometry around the nickel(II)
center is essentially square-planar. In the cationic unit of 2e
two aryl rings can be considered to be oriented axially (rings 2
and 3) with respect to the five-membered chelate ring, and two
aryl rings are oriented equatorially (rings 1 and 4) as is the case
in II. However, in the cationic unit of 2e the axial aryl rings
are lying on opposite sides of the five-membered chelate ring,
whereas in the cation of II the axial phenyl rings are lying on
the same side of the five-membered chelate ring, as can be seen
in Figure 2b. As in the crystal structure of [Ni(o-MeO-dppe)I2],
the o-methoxy groups of the axial aryl rings are pointing away
from the central nickel atom (torsion angles Ni1-P4-C21-
C22 ) -164.72(13)° and Ni1-P7-C31-C32 ) -170.96(13)°),
Characterization of a Mono(chelate) and a Bis(chelate)
1
Complex in Solution. The chemical shifts in H NMR and
31P{1H} NMR in CD2Cl2 for the six complexes and the free
1
ligand are summarized in Table 1. The H NMR spectra in
CD2Cl2 of [Ni(o-MeO-dppe)2](PF6)2 (2e), as typical for a bis-
(chelate) complex, and of [Ni(o-MeO-dppe)Cl2] (1a), as typical
for a mono(chelate) complex, are shown in Figure 4. The sharp
1H NMR signals confirm that the complexes are diamagnetic
and that the square-planar geometry around the central nickel
atom is retained in solution. This is also observed in the
electronic spectra of 1a and 2e taken in CH2Cl2, which are
similar to the diffuse reflectance electronic spectra of the solids.
However, as in solution the charge-transfer bands shift to
somewhat higher energies, now for 2e a band at 29 600 cm-1
with a shoulder at 28 000 cm-1 is observed.
(22) 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.
(23) Williams, A. F. Acta Crystallogr. 1989, C45, 1002.

Nickel(II) Diphosphane Hydrogenation Catalysts
Inorganic Chemistry, Vol. 40, No. 9, 2001 2079
Scheme 4
attached to the axial and to the equatorial aryl rings, respectively.
Finally, a doublet at 2.2 ppm and a multiplet at 2.8 ppm
corresponding to the ethylene bridge are observed, as the two
hydrogen atoms on each of the carbon atoms in the bridge are
not equivalent.
1
In contrast to the complicated H NMR spectrum of 2e, the
1H NMR spectrum of 1a is surprisingly simple. In the spectrum
depicted in Figure 4b, four different signals corresponding to
the hydrogen atoms in the aryl groups, one signal corresponding
to the methoxy groups, and one signal corresponding to the
ethylene bridge are observed. However, on the basis of the
crystal structure of the related compound 1c, a 1H NMR
spectrum rather similar to that for 2e would be expected as also
in this compound the four aryl rings are equivalent two by two
(see Figure 1 and Figure 2c). The observed spectrum can be
explained by assuming the occurrence of a rapid interchange
on the NMR time scale of the equatorial and axial aryl groups
of two possible rotamers, as depicted in Scheme 4, thus making
all four aryl groups of the ligand spectroscopically equivalent.
In one rotamer the four aryl groups are arranged as is depicted
in Figure 1; in the other rotamer the aryl groups which were
oriented axially with respect to the five-membered chelate ring
have become equatorial and the equatorial aryl groups have
become axial by twisting the ligand about the midpoint of the
ethylene bridge while keeping the phosphorus positions fixed.
This twist is accompanied by simultaneous rotation of the phenyl
groups around the P-C bond, making all four methoxy groups
and the o-hydrogen atoms mutually equivalent. In the region
6.9-8.1 ppm four signals corresponding to the aryl protons are
found. The signal corresponding to the o-hydrogens in 1a is
broadened and shifted downfield to 8.1 ppm. This downfield
Figure 4. (a) 1H NMR spectrum of [Ni(o-MeO-dppe)2](PF6)2 in CD2Cl2
at 25 °C as a typical example of a bis(chelate) complex. (b) H NMR
spectrum of [Ni(o-MeO-dppe)Cl2] in CD2Cl2 at 25 °C as a typical
example of a mono(chelate) complex. Signals arising from solvent or
impurities are indicated with an asterisk.
1
By virtue of the presence of methoxy groups in the ligand, it
is possible to distinguish a mono(chelate) complex from a
1
1
bis(chelate) complex using H NMR spectroscopy. In the H
NMR spectrum of 2e, depicted in Figure 4a, eight different
signals corresponding to the hydrogen atoms in the aryl groups,
two different signals corresponding to the methoxy groups, and
two different signals corresponding to the ethylene bridge are
observed. This confirms the presence of two types of aryl rings,
equatorial and axial, in agreement with the crystal structure of
2e (see Figure 3). In solution all four equatorial rings are
equivalent, as are all four axial rings. The signal corresponding
to one type of o-hydrogen in 2e is shifted far downfield to 8.6
ppm. This large downfield shift compared to the free ligand is
ascribed to an interaction of the o-hydrogen atoms of the axial
1
shift is also observed in the H NMR spectrum of 2e and can
be attributed to an interaction of the o-hydrogen atoms of the
2
axial aryl groups with the filled dz orbital of the nickel atom.
The downfield shift for 1a is somewhat smaller than the one
observed for 2e, which can be rationalized with the fluxional
behavior of the ligand, leading to a time-averaged signal. Also
in the case of 1a the Ni‚‚‚H-C interaction is best described as
a hydrogen bond.24 At 3.6 ppm the signal for the time-average
equivalent methoxy groups is observed, and at 2.4 ppm a virtual
doublet for the four time-average equivalent hydrogen atoms
in the ethylene bridge is present, caused by inequivalent coupling
with the two phosphorus atoms.
2
aryl groups with the filled dz orbital of the nickel atom. This
indicates that the short Ni‚‚‚H distances which are observed in
the crystal structure of 2e are retained in solution. The large
downfield shift of these protons in the 1H NMR spectrum
indicates that the observed Ni‚‚‚H interaction in the solid state
is better described as a hydrogen bond, rather than as an agostic
interaction.24 The signal corresponding to the other type of
o-hydrogen atom in 2e is shifted upfield to 5.8 ppm. This large
upfield shift is likely to be caused by the shielding of the
hydrogen atoms H16 and H46 (connected to C16 and C46,
respectively) by the electron clouds of aryl rings 4 and 1,
respectively (see Figure 3). At 3.4 and 3.5 ppm signals of the
methoxy groups are found, corresponding to the methoxy groups
The differences observed in the NMR spectra of the mono-
(chelate) and the bis(chelate) are not merely quantitative. It is
to be expected that at lower temperatures the NMR spectra of
the mono(chelate) complex will show the decoalescence of the
signals.12 Raising the temperature for the bis(chelate) complex
does not result in the coalescence of the signals, but instead
results in the oxidation of the ligand to the bis(phosphane) oxide.
However, in a HH NOESY NMR spectrum exchange peaks
were observed, indicating that also the bis(chelate) complex is
involved in a dynamic process like the mono(chelate) complex.
This process is much faster on the NMR time scale for the
mono(chelate) complex than for the bis(chelate) complex due
to steric congestion in the latter.
(24) Yao, W.; Eisenstein, O.; Crabtree, R. H. Inorg. Chim. Acta 1997, 254,
105.

2080 Inorganic Chemistry, Vol. 40, No. 9, 2001
Angulo et al.
Table 4. Summary of the NMR Results in the Study of the Autoionization Equilibrium Using o-MeO-dppe. Formation of Mono(ligand) or
Bis(ligand) Complexes at 25 °C
a
CD3OD
acetone-d6
CD2Cl2
acetone-d6/toluene-d8
[Ni(o-MeO-dppe)Cl2] (1a)
[Ni(o-MeO-dppe)Br2] (1b)
[Ni(o-MeO-dppe)I2] (1c)
[Ni(o-MeO-dppe)2](TFA)2 (2d)
[Ni(o-MeO-dppe)2](PF6)2 (2e)
[Ni(o-MeO-dppe)(TFA)2] (1d) or
[Ni(o-MeO-dppe)2][Ni(TFA)4] (3d)
Ni(OAc)2‚4H2O:o-MeO-dppe ) 1
Ni(OAc)2‚4H2O:dppe ) 1
b
b
b
bis
bis
bis
b
b
b
c
mono
mono
mono
bis
c
c
c
monod
c
c
bis
mono/bis
mono/bis
(3:1)e
mono
monog
mono
(1:1)e,f
bis
c
c
c
c
bisg
a Acetone and toluene were mixed in a 1:4 ratio. b Compound is not soluble. c Not measured. d Free ligand was observed as well. e Measured 10
min after the addition of the solvent. f This ratio changes to 3:1 after 3 days. g Determined using electrospray mass spectroscopy.
For both the mono(chelate) and the bis(chelate) complex only
one signal is observed in 31P{1H} NMR, indicating that in each
metal complex the phosphorus atoms are equivalent. The
difference in chemical shift between free and coordinated ligand,
∆δ, is 84-107 ppm, as expected for five-membered chelate
rings, which induce a strong deshielding effect.25 For the TFA
compound, it is not possible to assign either a mono(chelate)
complex (1d) or a bis(chelate) complex (3d) using 31P{1H}
NMR, as the difference in chemical shift between the two
complexes is very small (see Table 1).
Finally, when the orange solid is dissolved in a medium of
intermediate polarity (acetone or dichloromethane), a mixture
of mono- and bis(chelate) species is observed, whereby the
timespan in which equilibrium is reached depends on the solvent
used.
In acetone the mono-to-ionized ratio gradually changes in 3
days from 1:1 to 3:1, whereas in dichloromethane a 3:1 ratio is
observed instantaneously and this ratio does not change in time.
As in acetone the mono-to-ionized ratio gradually changes in
favor of the mono(chelate) complex, it is likely that the solid
which is initially dissolved largely consists of the ionized
complex 3d. This ionized complex is to some extent stabilized
in the more polar solvent acetone and, therefore, more slowly
converted to the mono(chelate) complex than in the more apolar
dichloromethane.
Influence of the Solvent on the Autoionization Equilibri-
um. An NMR study of the six complexes was initiated to
determine the influence of the polarity of the solvent (methanol
ꢀ ) 32.6, acetone ꢀ ) 20.7, dichloromethane ꢀ ) 9.1, and an
acetone/toluene mixture ꢀtoluene ) 2.4)26 on the nature of the
species present in solution. The results of this NMR study are
presented in Table 4. For the complexes [Ni(o-MeO-dppe)X2]
(1), where X is a strongly coordinating anion (Cl, Br, or I),
only the mono(chelate) complex is observed in dichloromethane.
Unfortunately, 1H NMR spectra in methanol and acetone could
not be obtained, as these complexes are insoluble in these
solvents. The 1H NMR spectra of these complexes in an acetone/
toluene mixture were not taken, as no differences with the
spectra obtained in CD2Cl2 are expected. For the complexes
[Ni(o-MeO-dppe)2]Y2 (2), where Y is a more weakly coordinat-
ing anion (TFA or PF6), only the bis(chelate) complex is
observed in dichloromethane, just as in the more polar solvent
methanol. In an apolar acetone/toluene (1:4) mixture, the mono-
(chelate) complex and free, uncoordinated, ligand can be
observed; however, the solubility of the complex [Ni(o-MeO-
dppe)2](TFA)2 (2d) in this acetone/toluene mixture is very low.
The intended mono(chelate) complex [Ni(o-MeO-dppe)-
(TFA)2] (1d) appeared to be the most suitable compound to
study the autoionization equilibrium. Upon dissolution of the
orange solid in the polar solvent methanol, a yellow solution is
obtained, and the typical pattern for a bis(o-MeO-dppe) complex
is observed. This means that the autoionization equilibrium in
this solvent lies completely to the right and the complex is
So, it appears that there is a direct relation between the
polarity of the solvent and the position of the autoionization
equilibrium. In a polar solvent only the ionized complex is
present, whereas in an apolar solvent only the mono(chelate)
complex is observed. The influence of the anion on the position
of the autoionization equilibrium is discussed below.
Catalyst Mixtures Containing o-MeO-dppe or dppe as the
Ligand. The crystal structure of 2e and the NMR studies
described above show that the presence of the o-methoxy group
on the phenyl ring does not prevent the formation of an ionized
complex. It is of interest to know whether in the in situ prepared
mixtures for the catalytic hydrogenation reactions autoionization
occurs as well. Therefore, some in situ prepared mixtures,
containing nickel(II) acetate and the ligand o-MeO-dppe or dppe,
1
were analyzed using both H and 31P{1H} NMR spectroscopy
(see Table 4). As the mono(dppe) and bis(dppe) complexes
cannot readily be distinguished using NMR spectroscopy, the
nature of the species present in solution was determined using
electrospray mass analysis. In methanol, the solvent that is used
in the reported catalytic hydrogenation experiments,8 only the
ionized complex is observed for both ligands, whereas in
dichloromethane only the mono(chelate) complex proved to be
present in both cases. This result indicates that, apart from the
solvent, also the anion has a considerable influence on the
position of the autoionization equilibrium as for the TFA
complex 1d in chloroform a mixture of mono- and bis(chelate)
species is present. The stronger are the coordinating abilities
of the anion, the more the autoionization equilibrium will be
shifted toward the mono(chelate) complex in a solvent of
intermediate polarity.
1
present as 3d. The H NMR spectrum of an ionized complex
appears to be identical to that of a bis(ligand) complex. It seems
that the presumably paramagnetic anionic unit of the ionized
1
complex has no influence on the H NMR spectrum of the
diamagnetic cationic unit. Upon dissolution of the orange solid
in the apolar acetone/toluene mixture, an orange solution is
obtained and the typical pattern of a mono(o-MeO-dppe)
complex is observed, indicating that the equilibrium now lies
completely to the left and that the compound is present as 1d.
In the case of o-MeO-dppe the influence of methanol on the
autoionization equilibrium was investigated. Even when only
5 drops (∼2%) of CD3OD were added to a solution of nickel(II)
acetate and the ligand in CD2Cl2, approximately 30% of the
mono(chelate) complex was converted to the ionized complex
(25) Garrou, P. E. Chem. ReV. 1981, 81, 229.
(26) Burger, K. Organic Reagents in Metal Analysis; Pergamon Press:
Oxford, 1973; Table 59.

Nickel(II) Diphosphane Hydrogenation Catalysts
Inorganic Chemistry, Vol. 40, No. 9, 2001 2081
Table 5. 31P{1H} NMR Data (recorded at 25 °C) of the
Experiments Mimicking the Catalytic Conditions Using
o-MeO-dppe and dppe as the Ligands
o-MeO-dppe
dppe
54.4 (s)
Ni(OAc)2‚4H2O:ligand ) 1:1
in CD3OD
+ 1 h at 50 °C
52.5 (s)
52.5 (s)
54.4 (s), 34.6 (s)
(ratio 1:2)
55.4 (s)
55.3 (s), 35.4 (s)
(ratio 1:1)
a
+ 10 bar of H2 pressure
+ 10 bar of H2 pressure and
30 min at 50 °C
53.5 (s)
53.5 (s)
spent catalyst in CD3OD
52.5 (s)
a Not measured.
dppe (in a 1:1 ratio) in methanol. Subsequently, the solvent was
evaporated and the remaining solid was dissolved in di-
chloromethane. It appeared that the ionized complex was indeed
converted to the mono(o-MeO-dppe) complex, as indicated by
1H NMR spectroscopy. From these experiments it appears that
the autoionization equilibrium is fully reversible.
Figure 5. 1H NMR spectrum of nickel(II) acetate mixed with o-MeO-
dppe (1:1) in CD2Cl2 at 25 °C after the addition of 5 drops of CD3OD.
Signals arising from solvent or impurities are indicated with an asterisk.
The identity of the anionic part of the ionized complex (3)
was investigated by electrospray mass spectroscopy in the
negative mode. Electrospray mass analysis of the in situ formed
complex using Ni(OAc)2‚4H2O and o-MeO-dppe in methanol
showed the presence of OAc-, [Ni(OAc)3]-, and a very small
amount of [Ni2(OAc)5]-. In view of these findings it is probably
more correct to represent the ionized complex (3) as [Ni(o-
MeO-dppe)2][Ni(OAc)3](OAc).
Probing the Working Hypothesis. From the studies de-
scribed above it appears that the differences in catalytic
hydrogenation activity of nickel(II) acetate in combination with
dppe or o-MeO-dppe cannot be rationalized by the autoioniza-
tion hypothesis. Therefore, in situ prepared catalytic mixtures
containing nickel(II) acetate and either of these ligands were
further studied in order to try to get some insight into the origin
of the positive effect of the o-methoxy group on the catalytic
activity. The catalyst, nickel(II) acetate and the ligand dissolved
in methanol, was heated to 50 °C under dihydrogen pressure of
10 bar and analyzed using 31P{1H} NMR spectroscopy. The
results are presented in Table 5. Under all conditions in the
case of o-MeO-dppe only the ionized complex was observed,
also in a spent catalyst after a catalytic hydrogenation reaction.
However, in the 31P{1H} NMR spectra of the solutions
containing the ligand dppe in some cases a peak at 34.6 ppm
appears besides the peak at 54.4 ppm. LC-MS analysis showed
the peak at 34.6 ppm to originate from the fully oxidized dppe
ligand [bis(phosphane oxide)]. This signal appears when the
solution is heated, even under reducing conditions, indicating
that the dppe ligand is oxidized under the catalytic hydrogenation
conditions.
New Hypothesis. Oxidation of the ligand may account for
the difference in activity between catalysts containing the ligands
o-MeO-dppe and dppe. This oxidation was expected to be an
important factor in the case of complexes containing the ligands
o-MeO-dppp and dppp, as with these ligands autoionization to
bis(chelate) complexes is not observed at all.12 Therefore, the
spent catalyst after a hydrogenation reaction, employing a
catalyst based on either of the two ligands, was analyzed using
31P{1H} NMR spectroscopy. In the case of o-MeO-dppp no
oxidized ligand was observed, whereas dppp was oxidized to
the bis(phosphane oxide). Monitoring the fate of dppp in CDCl3
at room temperature, it appeared that ligand is fully oxidized
in 6 h in the presence of nickel(II) acetate, whereas in the
absence of nickel(II) acetate the ligand is hardly oxidized over
the same period of time. Moreover, as all experiments were
Figure 6. Electronic absorption spectra of the in situ formed complex
using Ni(OAc)2‚4H2O and o-MeO-dppe (1:1) in a dichloromethane/
methanol mixture (4:1) recorded at 25 °C under Ar. Spectra recorded
during 3 h after dissolution, at 3, 5, 10, 15, 30, 45, 60, 120, and 180
minutes.
(see Figure 5).27 In a 1:1 mixture of methanol and dichloro-
methane, which appeared to be the most suitable for catalytic
hydrogenation reactions,28 only the ionized complex can be
observed in NMR.
Some Aspects of the Autoionization Equilibrium. Elec-
tronic absorption spectroscopy was used to prove that the
mono(chelate) complex is converted to the bis(chelate) complex
in the autoionization equilibrium. When o-MeO-dppe and
nickel(II) acetate are mixed in a 1:1 ratio in a 4:1 mixture of
dichloromethane and methanol, the conversion of mono(chelate)
complex into ionized complex can be followed in time, as is
shown in Figure 6. In 2 h the absorption at 450 nm, corre-
sponding to the mono(chelate) complex, disappears, while an
absorption at 340 nm and a shoulder at 360 nm, corresponding
to the bis(chelate) complex, appear after 15 min and increase
in intensity with time. An isosbestic point is observed at 420
nm, indicating that one species is converted into the other, and
the color of the solution slowly changes from orange to yellow.
The question arises whether the ionized complex can be
reconverted to the mono(chelate) complex. Therefore, an ionized
complex was generated using nickel(II) acetate and o-MeO-
(27) Surprisingly, the NMR spectra do not show the presence of the acetate
anions. In the case of a bis(chelate) complex this can be explained by
the formation of paramagnetic [Ni(OAc)3]- or [Ni(OAc)4]2- species,
which are not visible in the NMR. In the case of mono(chelate)
complexes we assume that this is caused by intermediate exchange of
the coordinating acetate ions with the methanol solvent.
(28) Angulo, I. M.; Bouwman, E.; Lok, S. M.; Quiroga, V. F. In preparation.

2082 Inorganic Chemistry, Vol. 40, No. 9, 2001
Angulo et al.
carried out in an inert atmosphere the oxidation seems to be
related to the nickel(II) center, as already observed by Jarrett
et al.10 It was reported earlier that Pd(OAc)2 is able to oxidize
tertiary phosphanes.29 We now believe that in our case also the
Ni(OAc)2 oxidizes the tertiary phosphane ligands dppe and dppp,
whereas the ligands o-MeO-dppe and o-MeO-dppp are not
oxidized.
Prevention of oxidation of the phosphorus atom by the
o-methoxy group could again be due to either a steric or
electronic effect. We have reported8 that when using p-MeO-
dppp9 no catalytic hydrogenation activity is observed. 31P{1H}
NMR spectroscopy shows that also in this case the ligand is
rapidly oxidized. These two observations imply that the preven-
tion of oxidation in the o-methoxy-substituted ligands is merely
steric; the nearby presence of an o-methoxy group shields the
phosphorus atom to which it is attached from oxidation.
Although the o-methoxy groups in o-MeO-dppe do not
prevent autoionization, still the derived nickel complexes yield
active hydrogenation catalysts. The performance of the o-MeO-
dppe containing nickel(II) catalysts is rather sensitive toward
variations in solvents and anions, which is presumably due to
the occurrence of the autoionization equilibrium.28 The catalytic
activity of this system may (partly) be explained by the presence
of the apolar substrate 1-octene. The addition of this apolar
substrate in a significant amount will lower the overall polarity
of the reaction mixture and thereby will shift the autoionization
equilibrium in favor of the mono(chelate) complex. As the nickel
complexes containing the ligand o-MeO-dppe are involved in
an autoionization equilibrium under the conditions used in
catalytic hydrogenation reactions, it may well be that only a
small part of the available nickel complexes is involved in
hydrogenation catalysis. It is expected that when autoionization
could be suppressed, more active hydrogenation catalysts are
to be obtained.
Conclusion
It has become clear that the o-methoxy substituent on the
phenyl groups in the ligand o-MeO-dppe does not prevent
autoionization of the nickel complexes coordinated by this
ligand. In a polar solvent in combination with weakly coordinat-
ing anionssthe conditions used in the catalytic hydrogenation
of 1-octenesonly the ionized complex is observed, whereas in
an apolar solvent in combination with coordinating anions only
the mono(chelate) complex occurs. Complexes containing the
ligands o-MeO-dppp and dppp do not show autoionization to
bis(chelate) complexes.12 It is shown that the role of the methoxy
substituents in o-MeO-dppe and o-MeO-dppp may be that these
prevent oxidation of the ligand. This oxidation appeared to be
rapid in the case of the unsubstituted ligands dppe and dppp,
whereas it is relatively slow or nonexistent for o-MeO-dppe
and o-MeO-dppp.
Acknowledgment. This research has been financially sup-
ported by the Council for Chemical Science of the Netherlands
Organization for Scientific Research (CW-NWO). We thank Dr.
J. Reedijk and Dr. E. Drent for fruitful discussions. We thank
A. B. van Oort (Shell) for measuring the high-pressure NMR
spectra, Dr. W. Genuit (Shell) for measuring MS and LC-MS
spectra, and V. F. Quiroga (Leiden) for measuring electronic
absorption spectra.
Supporting Information Available: X-ray crystallographic files
in CIF format for the complexes [Ni(o-MeO-dppe)I2] and [Ni(o-MeO-
dppp)2](PF6)2. This material is available free of charge via the Internet
at http://pubs.acs.org.
(29) Amatore, C.; Carre´, E.; Jutand, A.; M’Barki, M. A. Organometallics
1995, 14, 1818.
IC0006955