Applying the macrocyclic effect to smaller ring structures. N,N'-dimethyl-3,7-diazabicyclo[3.3.1]nonane nickel(0) complexes

Article abstract of DOI:10.1021/ja971037v

The two N donor atoms in the tertiary diamine N,N'-dimethyl-3,7-diazabicyclo[3.3.1]nonane (dabn, C₉H₁₈N₂) are ideally positioned in the bicyclic structure for chelation to a metal center. This feature was utilized to synthesize unusual diamine nickel(0)-ethene and -ethyne complexes, which represent limiting cases of the Pearson hard-soft acid-base concept. Thus, the reaction of Ni(C₂H₄)₃ with dabn affords yellow TP-3 (C₉H₁₈N₂)Ni(C₂H₄) (1) (dec. 0°C) in which the ethene ligand displays extreme high-field NMR shifts at δ(H) 0.27 and δ(C) 20.4 and an exceptionally small coupling constant ¹J(CH) = 142 Hz. Reaction of 1 with butadiene yields the red mononuclear T-4 complex (C₉H₁₈(N)₂)Ni(η²-C₄H₆)₂ (2a) in solution, from which the dinuclear derivative {(C₉H₁₈N₂)Ni(η²-C₄H₆,)}₂(μ-η²,η^2-C₄H₆) (2) (dec. 20°C) is isolated. Complexes 2 and 2a are more soluble than 1 and thus better suited for further reactions. When ethyne is added to a solution of 2 or 2a at -78°C, the yellow TP-3 complex (C₉H₁₈N₂)Ni(C₂H₂) (3) (dec. -30°C) is obtained. The ethyne ligand of 3 exhibits the lowest IR C≡C stretching frequency (1560 cm^-1) and by far the smallest NMR coupling constant ¹J(CH) = 178 Hz yet reported for a mononuclear nickel(0)-ethyne complex. In addition, Ni(CO)₄ reacts with dabn to yield orange T-4 (C₉H₁₈N₂)Ni(CO)₂ (4). The results demonstrate that tertiary diamines, which are hard Lewis bases and which a priori are expected to coordinate poorly to the soft Lewis acid Ni(0), may be supported in such a coordination by the stabilizing principle of preorganization and consequently act as very powerful donor ligands.

Full text of DOI:10.1021/ja971037v

7992
J. Am. Chem. Soc. 1997, 119, 7992-7999
Applying the Macrocyclic Effect to Smaller Ring Structures.
N,N′-Dimethyl-3,7-diazabicyclo[3.3.1]nonane Nickel(0)
Complexes†
Karl-Josef Haack, Richard Goddard, and Klaus-Richard Po1rschke*
Contribution from the Max-Planck-Institut fu¨r Kohlenforschung,
D-45466 Mu¨lheim an der Ruhr, Germany
ReceiVed April 1, 1997X
Abstract: The two N donor atoms in the tertiary diamine N,N′-dimethyl-3,7-diazabicyclo[3.3.1]nonane (dabn, C9H18N2)
are ideally positioned in the bicyclic structure for chelation to a metal center. This feature was utilized to synthesize
unusual diamine nickel(0)-ethene and -ethyne complexes, which represent limiting cases of the Pearson hard-soft
acid-base concept. Thus, the reaction of Ni(C2H4)3 with dabn affords yellow TP-3 (C9H18N2)Ni(C2H4) (1) (dec. 0
°C) in which the ethene ligand displays extreme high-field NMR shifts at δ(H) 0.27 and δ(C) 20.4 and an exceptionally
1
small coupling constant J(CH) ) 142 Hz. Reaction of 1 with butadiene yields the red mononuclear T-4 complex
(C9H18N2)Ni(η2-C4H6)2 (2a) in solution, from which the dinuclear derivative {(C9H18N2)Ni(η2-C4H6)}2(µ-η2,η2-
C4H6) (2) (dec. 20 °C) is isolated. Complexes 2 and 2a are more soluble than 1 and thus better suited for further
reactions. When ethyne is added to a solution of 2 or 2a at -78 °C, the yellow TP-3 complex (C9H18N2)Ni(C2H2)
(3) (dec. -30 °C) is obtained. The ethyne ligand of 3 exhibits the lowest IR CtC stretching frequency (1560 cm-1
)
and by far the smallest NMR coupling constant 1J(CH) ) 178 Hz yet reported for a mononuclear nickel(0)-ethyne
complex. In addition, Ni(CO)4 reacts with dabn to yield orange T-4 (C9H18N2)Ni(CO)2 (4). The results demonstrate
that tertiary diamines, which are hard Lewis bases and which a priori are expected to coordinate poorly to the soft
Lewis acid Ni(0), may be supported in such a coordination by the stabilizing principle of preorganization and
consequently act as very powerful donor ligands.
Introduction
complexes, (C7H13N)Ni(C2H4)2 reacts with CO to yield pure
(C7H13N)Ni(CO)3 (dec. > 0 °C).3
According to the Pearson hard-soft acid-base (HSAB) concept
of acids and bases, amines are hard bases and Ni(0) is a soft
acid,1 and thus, the coordination of amines to Ni(0) is predicted
to be unfavorable. This is expected to be particularly the case
when Ni(0) is coligated by a nonactivated alkene such as ethene.
We were therefore surprised to find that one ethene ligand in
Following from these studies on (amine)Ni(C2H4)2 type
complexes, we were interested in the question as to whether
complexes of the inverse stoichiometry (amine)2Ni(C2H4) would
also be stable and whether an ethyne derivative (amine)2Ni(C2H2)
could be prepared. In these complexes, the charge transferred
to Ni(0) by two nitrogen donor atoms has to be accommodated
by back-bonding to a single π-ligand. Since these complexes
are expected to be inherently labile, we embarked on a project
to study ways of stabilizing them, and the results of this work7
are reported here.
2
Ni(C2H4)3 can be readily displaced by an amine [e.g., NH3,
Et2NH, quinuclidine (C7H13N)] to yield trigonal-planar (TP-3)
complexes of the type (amine)Ni(C2H4)2. An X-ray analysis
of (C7H13N)Ni(C2H4)2 (dec. 0 °C) reveals a long Ni-N bond
(2.07 Å).3 This type of complex must be distinguished from
the more stable Strohmeier type complexes, such as (C7H13N)-
Cr(CO)5,4 in which the carbonyl ligands induce a fairly large
electron withdrawal from the metal(0) center, rendering the latter
somewhat harder and consequently more suited for complexing
a hard amine ligand. This tendency is also seen in the reaction
of Ni(CO)4 with liquid NH3 at 20 °C (sealed tube), which affords
a mixture of the presumed complexes (NH3)Ni(CO)3 and
(NH3)2Ni(CO)2 in solution, which upon isolation decompose
at -60 °C.5,6 In accord with the increased stability of the CO
Results
According to current knowledge, it does not appear likely
that any (amine)2Ni(C2H4) complex with two monodentate
amine ligands is stable enough to be synthesized by classical
wet chemistry. An ab initio MO-SCF calculation8 on the
simplest homolog (NH3)2Ni(C2H4) predicts the molecule to be
thermodynamically stable. This complex may possibly be
formed as a transient species, e.g., by the reaction of a
Ni(0)(NH3)2 matrix9 with ethene. However, all attempts to
† Part VII of our series Amine-Nickel(0) Complexes. For part VI see ref
15.
X Abstract published in AdVance ACS Abstracts, August 1, 1997.
(1) (a) Pearson, R. G. J. Am. Chem. Soc. 1963, 85, 3533. (b) Pearson,
R. G. J. Chem. Educ. 1968, 45, 581; 1968, 45, 643. (c) Parr, R. G.; Pearson,
R. G. J. Am. Chem. Soc. 1983, 105, 7512.
(6) A 1:1 mixture of Ni(CO)4 and tmeda is yellow at 20 °C and orange
at 30 °C but no CO evolves. After some pentane has been added, no product
crystallizes at -78 °C. This and other observations suggest that [(tmeda)-
Ni(CO)2] cannot be isolated. Moreover, it is expected that [(tmeda)-
(2) Fischer, K.; Jonas, K.; Wilke, G. Angew. Chem. 1973, 85, 620; Angew.
Chem., Int. Ed. Engl. 1973, 12, 565.
(3) Kaschube, W.; Po¨rschke, K.-R.; Bonrath, W.; Kru¨ger, C.; Wilke, G.
Angew. Chem. 1989, 101, 790; Angew. Chem., Int. Ed. Engl. 1989, 28,
772.
(4) Aroney, M. J.; Clarkson, R. M.; Hambley, T. W.; Pierens, R. K. J.
Organomet. Chem. 1992, 426, 331. Choi, M.-G.; Brown, T. L. Inorg. Chem.
1993, 32, 1548.
Ni(CO)2] if formed will readily disproportionate into [Ni(tmeda)2]2+
-
[Ni6(CO)12]2-, cf., the characterization of (NH3)Ni(CO)3 and (NH3)2Ni(CO)2
in ref 5.
(7) (a) Haack, K.-J. Diplomarbeit, Universita¨t Du¨sseldorf, 1992. (b)
Haack, K.-J.; Po¨rschke, K.-R.; Schro¨der, W. Euchem Conference: Nitrogen
Ligands in Organometallic Chemistry and Homogeneous Catalysis, OC 11;
Alghero, Italy, May 10-15, 1992.
(8) Åkermark, B.; Almemark, M.; Almlo¨f, J.; Ba¨ckvall, J.-E.; Roos, B.;
Støgård, A° . J. Am. Chem. Soc. 1977, 99, 4617.
(5) Behrens, H.; Zizlsperger, H. J. Prakt. Chem., 4. Reihe 1961, 14, 249.
S0002-7863(97)01037-8 CCC: $14.00 © 1997 American Chemical Society

3,7-Diazabicyclo[3.3.1]nonane Nickel(0) Complexes
J. Am. Chem. Soc., Vol. 119, No. 34, 1997 7993
synthesize and isolate the compound have so far been unsuc-
cessful. In the hope of introducing some stability by the chelate
effect, we scrutinized various structural types of bidentate
amines for their suitability as ligands for Ni(0).
at Ni(0), but the tmeda ligand influence is too weak to deactivate
the π-coordinated ethyne and prevent its polymerization.
(b) 1,4-Diazacyclohexane, 1,5-Diazacyclooctane, and Their
N,N′-Dimethyl-Substituted Derivatives. We then looked at
the monocyclic diamines, 1,4-diazacyclohexane (piperazine),
1,5-diazacyclooctane (daco) and their N,N′-dimethyl-substituted
derivatives, where the conformational freedom of the ligand is
restricted. For (N,N′-dimethyl) piperazine the boat conformation
necessary for a chelating coordination (A) is energetically
unfavorable as compared with the chair conformation due to
the torsional strain of the neighboring methylene groups.18a
(a) Tetramethylethylenediamine (tmeda). When tmeda is
added to a pentane solution of Ni(C2H4)3 at -78 °C, the yellow
color intensifies, but no diamine complex is observed. We
assume that the dinuclear complex (C2H4)2Ni(µ-Me2NC2H4-
NMe2)Ni(C2H4)2 is in equilibrium with tmeda and Ni(C2H4)3.
In (C2H4)2Ni(µ-Me2NC2H4NMe2)Ni(C2H4)2, the bridging tmeda
ligand is part of two (amine)Ni(C2H4)2 type moieties. The
species [(tmeda)Ni(C2H4)], containing a chelating tmeda ligand,
is not observed in the solution.3 The same is true for N,N,N′,N′-
tetramethyl-1,3-propanediamine. Thus, it would appear that for
acyclic diamines with N atoms in the 1,4- or 1,5-positions, no
Ni(0)-ethene complex of the type (amine)2Ni(C2H4) with a
chelating diamine ligand is isolable. A likely explanation is
that the π-acceptor strength of ethene is not large enough to
make Ni(0) sufficiently electrophilic to bind two tertiary amine
N atoms when their coordination is merely supported by an
open-chain-ligand chelate effect.10,11
Hence, the required energy to attain the boat conformation has
to be gained from the formation of strong metal-nitrogen bonds,
which is only the case when the diamine is coordinated to metal
ions. Thus, {MeN(C2H4)2NMe}PdCl218b has been isolated, but
in general (N,N′-dimethyl) piperazine complexes are rare.18c,d
Several complexes are known in which daco chelates to metal
ions.19 Coordination results in the formation of two metalla-
cyclic six-membered rings, as well as the original eight-
membered daco ring. If both six-membered rings attained the
chair conformation (B), then the torsional strain in each ring is
the least, but the 3,7-methylene groups interfere with each other
in the eight-membered ligand. Consequently, the chair/chair
conformation cannot be realized by the daco ligand. If a chair/
boat combination (C) is present, then the transannular interaction
of the methylene groups in the eight-membered ligand ring is
avoided, but the boat-shaped six-membered ring has an increased
energy due to the torsional strain caused by four C-H bonds
being eclipsed to N-C bonds. (For the same reason a
hypothetical boat/boat conformation can be dismissed here.)
Nevertheless, the chair/boat and an energetically similar chair/
semiboat conformation have been observed in certain com-
plexes.19 It follows from these arguments that the daco molecule
can in fact only coordinate to a metal center in a chair/boat or
chair/twisted-boat conformation. Both have a higher energy
than the uncoordinated flexible diamine.
The situation is somewhat different when the coligands have
stronger π-accepting properties than ethene. For example, the
reaction of the tmeda/Ni(C2H4)3 system with butadiene results
in the isolation of the dinuclear complex {(tmeda)Ni(η2-
C4H6)}2(µ-η2,η2-C4H6) (dec. -40 °C), and (tmeda)Ni(η2-C4H6)2
is observed in solution.12 When methyl acrylate is employed,
the complex (tmeda)Ni(H2CdCHCOOMe)2 (dec. 110 °C) can
be isolated.13 Interestingly, tmeda/Ni(C2H4)3 reacts with form-
aldehyde to yield (tmeda)Ni(C2H4)(CH2O) (dec. -15 °C), in
which the joint coordination of chelating tmeda and ethene to
Ni(0) is facilitated by the additional H2CdO ligand.14 In these
tmeda complexes, which contain π-ligands of moderate acceptor
strength (butadiene, H2CdCHCOOMe, H2CdO), the 18 e Ni(0)
center is tetrahedrally (T-4) coordinated. The 16 e TP-3 Ni(0)
complexes with chelating tmeda are only known for strong
π-acceptor ligands such as stilbene, maleic acid anhydride,
diphenylacetylene, tetrafluoroethene, or benzophenone, of which
(tmeda)Ni(C2F4)15 and (tmeda)Ni(OdCPh2)16 are structurally
characterized examples. A corresponding ethyne complex
[(tmeda)Ni(C2H2)] is not accessible, for as soon as an appropri-
ate synthesis is attempted, ethyne polymerizes instantaneously
at -100 °C, thereby reacting as if “naked nickel” were present.17
For ethyne, the electron withdrawal from Ni(0) by the ethyne
ligand is possibly sufficient to stabilize the tmeda coordination
We have, however, not been able to synthesize a correspond-
ing (amine)2Ni(C2H4) type complex with piperazine, daco, or
their N,N′-dimethyl-substituted derivatives. If we assume that
the expected R3N-Ni(0) coordinative bonds are inherently
weak, then the energy gained by forming these bonds is clearly
too low to compensate for the energy required by the diamine
ligands to attain the required conformation for chelation. This
also explains why the monocyclic diamines piperazine and daco,
although coordinating to certain metal ions, do not coordinate
to neutral Ni(0) in combination with ethene as coligand.20
(9) Ball, D. W.; Hauge, R. H.; Margrave, J. L. High Temp. Sci. 1988,
25, 95.
(10) Hancock, R. D. Prog. Inorg. Chem. 1989, 37, 187.
(11) In our experience, the species [(tmeda)Ni(0)] [145086-32-4] and
[(tmeda)Ni(cod)] [121719-01-5], which have often been proposed as
intermediates, but without experimental evidence,11a,18d are not formed to
any significant extent in the system tmeda/Ni(0)/cod. This is indicated by
the fact that when Ni(cod)2 (500 mg) is dissolved in undiluted tmeda (25
mL) at 40 °C, it recrystallizes unchanged at -30 °C (430 mg; 86%). (a)
Wenschuh, E.; Zimmering, R. Z. Anorg. Allg. Chem. 1989, 570, 93.
Walther, D.; Bra¨unlich, G.; Kempe, R.; Sieler, J. J. Organomet. Chem. 1992,
436, 109. Fischer, R.; Walther, D.; Kempe, R.; Sieler, J.; Scho¨necker, B.
J. Organomet. Chem. 1993, 447, 131. Kempe, R.; Sieler, J.; Walther, D.;
Reinhold, J.; Rommel, K. Z. Anorg. Allg. Chem. 1993, 619, 1105.
(12) (a) Schro¨der, W.; Po¨rschke, K.-R. J. Organomet. Chem. 1987, 322,
385. (b) Schro¨der, W. Diplomarbeit, Universita¨t Bonn, 1986.
(13) Kaschube, W.; Po¨rschke, K.-R.; Seevogel, K.; Kru¨ger, C. J.
Organomet. Chem. 1989, 367, 233.
(14) Schro¨der, W.; Po¨rschke, K.-R.; Tsay, Y.-H.; Kru¨ger, C. Angew.
Chem. 1987, 99, 953; Angew. Chem., Int. Ed. Engl. 1987, 26, 919.
(15) Kaschube, W.; Schro¨der, W.; Po¨rschke, K.-R.; Angermund, K.;
Kru¨ger, C. J. Organomet. Chem. 1990, 389, 399.
(16) Blum, K. Dissertation (K. Jonas), Universita¨t Bochum, 1978.
(17) Po¨rschke, K.-R.; Mynott, R.; Angermund, K.; Kru¨ger, C. Z.
Naturforsch., B: Anorg. Chem., Org. Chem. 1985, 40, 199.
(18) (a) Eliel, E. L.; Allinger, N. L.; Angyal, S. J.; Morrison, G. A.
Conformational Analysis; Wiley Interscience: Washington, DC, 1981; p
250. (b) Hassel, O.; Pedersen, B. F. Proc. Chem. Soc. 1959, 394. (c)
Farhangi, Y.; Graddon, D. P. Austr. J. Chem. 1974, 27, 2103. On the basis
of thermodynamic data, the authors postulate a 1:1 complex of chelating
N,N′-dimethylpiperazine with HgI2 but also note a “steric misfitting of the
bidentate base with the mercury atom”. (d) Eisch, J. J.; Hallenbeck, L. E.;
Han, K. I. J. Am. Chem. Soc. 1986, 108, 7763. The authors suggest a
displacement of a cod ligand from Ni(cod)2 by N,N′-dimethylpiperazine or
tmeda, which appears to be highly unlikely in the light of our results.
(19) Musker, W. K. Coord. Chem. ReV. 1992, 117, 133 and literature
cited therein.

7994 J. Am. Chem. Soc., Vol. 119, No. 34, 1997
Haack et al.
Table 1. 1H and 13C NMR Data of Uncoordinated dabn and the dabn Ligand of the Nickel(0) Complexes 1-4a
δ(H)
δ(C)
NCH3
NCHeqH
NCHHax
NCH3
2.09
Ctert
H
CH2
1.43
NCH2
60.6
Ctert
H
CH2
30.0
dabn
2.61
2.23
1.84
46.5
30.6
2J(HH) -11.0
3J(HH) 2.8
3.59
2J(HH) -11.0
3J(HH) 4.6
2.38
3J(HH) 3.3
1.50
1
2.48
1.92
61.7
55.7
30.6
32.4
2J(HH) -11.4
3J(HH) e 1
3.83/3.04
3.55
2J(HH) -11.4
3J(HH) 2.6
2.43/2.11
2.42
3J(HH) 3.3
1.56
1.53
2a
3
2.16
2.60
1.98
1.99
63.0/62.2
62.5
50.8
56.3
30.8
30.6
34.3
32.3
2J(HH) -11.3
3.37
2J(HH) -11.3
2.56
4
2.72
2.05
1.61
65.6
54.5
31.8
33.2
2J(HH) -10.8
2J(HH) -10.8
a Coupling constants are in hertz. For instrument and temperature specifications, see the Experimental Section.
It follows that a diamine ligand suitable for coordination to
the [Ni(0)(C2H4)] fragment can perhaps be derived from the
1,5-diazacyclooctane skeleton in its chair/chair conformation
(B) by substituting the methylene protons, which exert the
repulsive transannular interaction in the eight-membered ring,
with a bridging methylene group. We therefore turned our
attention to derivatives of 3,7-diazabicyclo[3.3.1]nonane (bis-
pidine), which contains a rigid molecular scaffold and provides
an apparently optimal arrangement of the N atoms for chelating
a metal center.
(c) N,N′-Dimethyl-3,7-diazabicyclo[3.3.1]nonane (dabn,
C9H18N2). The bispidine skeleton21-23 was synthesized by a
double Mannich reaction of N-methyl-4-piperidone, formalde-
hyde, and methylamine to afford N,N′-dimethyl-3,7-diazabicy-
clo[3.3.1]nonan-9-one.24,25 The bispidinone itself, which forms
two conformers,26 does not coordinate to the [Ni(0)(C2H4)] frag-
ment. Wolff-Kishner reduction (hydrazine/triethyleneglycol)
yields a mixture of dabn24 and several byproducts, from which
dabn was separated by rectification.
feature of six-membered rings and the σ-electron delocalization
of the N atom lone pairs into the antiperiplanar C-H bonds.28
It is worth noting that, in addition to the C-H stretching bands
in the usual region of the dabn IR spectrum (2960, 2930, 2880,
2839 cm-1), intense Bohlmann bands are observed at lower
frequencies (2773, 2735, 2689 cm-1). These Bohlmann bands
can be likewise attributed to the NCHHax protons antiperiplanar
to the N atom lone pairs.29 A similar feature is found in the
Raman spectrum. The observations are in agreement with the
assumption that the N atom lone pairs formally take up axial
positions, pointing to a common hypothetical coordination center
and leaving the N-Me substituents in equatorial positions. As
will be shown below, for (dabn)Ni complexes, the Bohlmann
bands are substantially reduced while a strong band near 2800
cm-1 occurs. Thus, coordinative constraint of the nitrogen lone
pairs weakens the delocalization into the antiperiplanar C-H
bonds.
The various results are best summarized by the statement that
liquid or dissolved dabn is present in a flattened chair/chair
conformation with the N-Me substituents in the equatorial
positions (D).24,27a,30 It becomes evident that in dabn, due to
The conformation of dabn has been determined on the basis
of its 1H24,27a and 13C27b NMR spectra (the latter in comparison
with the 13C NMR data of the bispidine derivatives (-)-sparteine
and R-isosparteine), its dipole moment,24 MO calculations,27b,d
and EI-MS27c and photoelectron (PE)27b spectra. In addition,
1
we have measured its high-resolution H NMR spectrum and
1
the IR and Raman (RA) spectra. In the H NMR spectrum
(Table 1), the signals of the NCH protons follow the rule
δ(NCH2) > δ(NCH3) with inequivalent geminal protons
NCHeqHax. Assuming a chair conformation of both six-mem-
bered rings (see below), the protons NCHHax are shielded by
0.4 ppm more than the protons NCHeqH, due to both a general
its particular bicyclic structure, the amine functions are almost
ideally positioned for chelation and that this substrate is by far
the best choice for attempting an otherwise unfavorable
coordination of a diamine to the [Ni(0)(C2H4)] fragment.
(d) N,N′-Dimethyl-3,7-diazabicyclo[3.3.1]nonane-Nickel(0)
Complexes. Following the arguments developed above,
Ni(C2H4)3 was reacted with dabn to yield the isolated (dabn)-
Ni(0)-ethene complex 1. From 1, the (dabn)Ni(0)-butadiene
(2, 2a) and (dabn)Ni(0)-ethyne (3) complexes are accessible.
(20) We did not study the combination of (N,N′-dimethyl-substituted)
daco and activated alkenes with nickel(0). We expect the situation to be
similar as that for the combination of tmeda with such alkenes, see above.
(21) Mannich, C.; Mohs, P. Ber. Dtsch. Chem. Ges. 1939, 63, 608.
(22) 3,7-Diazabicyclo[3.3.1]nonane, the parent bispidine, is prepared from
pyridine-3,5-dicarbonic acid in a multistep synthesis in fair yield. (a) Stetter,
H.; Merten, R. Chem. Ber. 1957, 90, 868. Stetter, H.; Hennig, H. Chem.
Ber. 1955, 88, 789. (b) Stetter, H. Angew. Chem. 1962, 74, 369; Angew.
Chem., Int. Ed. Engl. 1962, 1, 286.
(23) For a detailed review on bispidine and related compounds, see:
Jeyaraman, R.; Avila, S. Chem. ReV. 1981, 81, 149.
(24) Douglass, J. E.; Ratliff, T. B. J. Org. Chem. 1968, 33, 355.
(25) Smissman, E. E.; Ruenitz, P. C.; Weis, J. A. J. Org. Chem. 1975,
40, 251.
(28) (a) Hamlow, H. P.; Okuda, S.; Nakagawa, N. Tetrahedron Lett. 1964,
2553. (b) Bohlmann, F.; Schumann, D.; Schulz, H. Tetrahedron Lett. 1965,
173.
(26) (a) For IR evidence, see: Haack, K.-J.; Seevogel, K.; Po¨rschke,
K.-R. Unpublished results. (b) For conformational features of the bispi-
dinone ring system, see: Black, D. St C.; Craig, D. C.; Horsham, M. A.;
Rose, M. J. Chem. Soc., Chem. Commun. 1996, 2093 and references cited
therein.
(27) (a) Zefirov, N. S.; Rogozina, S. V. Tetrahedron 1974, 30, 2345.
(b) Livant, P.; Roberts, K. A.; Eggers, M. D.; Worley, S. D. Tetrahedron
1981, 37, 1853. (c) Ruenitz, P. C.; Smissman, E. E.; Wright, D. S. J.
Heterocycl. Chem. 1977, 14, 423. (d) Chakrabarty, M. R.; Ellis, R. L.;
Roberts, J. L. J. Org. Chem. 1970, 35, 541.
(29) (a) Bohlmann, F. Chem. Ber. 1958, 91, 2157. (b) Lambert, J. B.;
Keske, R. G.; Carhart, R. E.; Janovich, A. P. J. Am. Chem. Soc. 1967, 89,
3761.
(30) Ab initio calculations on the parent bispidine (NH instead of NMe)
suggest the chair/chair conformation to be present in 80% in the gas phase.
For dabn, a flattened chair/chair conformation is assumed to predominate
on the basis of the PE spectrum and various MO calculations; a concentration
of 99.8% is predicted by a MINDO/3 calculation. It may be concluded
that the predominance of the chair/chair conformer of the bispidine skeleton
is supported by the N-CH3 substituents.27b

3,7-Diazabicyclo[3.3.1]nonane Nickel(0) Complexes
J. Am. Chem. Soc., Vol. 119, No. 34, 1997 7995
Table 2. Ethene and Ethyne Ligand NMR and IR Data of the (dabn)Ni(0) Complexes 1 and 3 and of Reference Compounds
δ(H)
δ(C)
1J(CH) [Hz]
ν(CdC) [cm-1
]
ν(CtC) [cm-1
]
C2H4
5.30
2.42
1.74
1.59
0.27
123.5
64.9
31.6
26.7
20.4
2.3
156
154
151
146
142
129
1623a
36 b
(ArNdCHCHdNAr)Ni(C2H4)2
(iPr2PC2H4PiPr2)Ni(C2H4)42
(bpy)Ni(C2H4)37
1530
1510
(dabn)Ni(C2H4) (1)
(tmeda)2(LiC2H5)2Ni(C2H4)55a
-0.83
C2H2
2.40
6.41
7.29
4.46
4.65
71.9
122.1
123.8
114.0
118.7
249
212
202
197
178
1974a
1630
1598
(Ph3P)2Ni(C2H2)43a
(iPr2PC2H4PiPr2)Ni(C2H2)42
(tBuNdCHCHdNtBu)Ni(C2H2)59
(dabn)Ni(C2H2) (3)
1560
a Gas phase; Raman. b Ar ) C6H3-2,6-iPr2.
The (dabn)Ni(0)-CO complex 4 was prepared from dabn and
Ni(CO)4.
(C9H18N2)Ni(C2H4) (1). The reaction of Ni(C2H4)3 with dabn
in diethyl ether at -10/-30 °C results in displacement of two
ethene ligands from Ni(0) and precipitation of yellow micro-
crystalline 1 in 83% yield (eq 1). Below -30 °C the reaction
that a fast low-temperature reaction in homogeneous solution
is required for a successful synthesis of labile Ni(0)-ethyne
complexes.33
Spectroscopical Properties. In the EI mass spectrum of 1
the molecular ion and the fragment [(dabn)Ni]+ are detected,
but the overall spectrum essentially corresponds to that of
uncoordinated dabn. The IR spectrum (KBr) is dominated by
vibrations of dabn coordinated to Ni(0). Additional bands which
we assign to the ethene ligand34 [ν(CdC) + δs(CH2) ) 1510
cm-1] are comparatively strongly shifted by complexation. In
1
the H NMR spectrum of 1, the dabn ligand signals (Table 1)
are all shifted to low field as compared with those of
uncoordinated dabn, and this shift is strongest for the NCHeqH
(∆δ ) -1.0 ppm) and NCH3 (∆δ ) -0.4 ppm) protons. Such
changes are less pronounced in the 13C NMR spectrum (Table
1), for which the low-field shift is strongest for NCH3 (∆δ )
-9 ppm). In contrast to dabn, the ethene ligand proton and
carbon resonances of δ(H) 0.27 and δ(C) 20.4 lie at extremely
high field, and the coupling constant 1J(CH) ) 142 Hz is
extraordinary low [typical range for Ni(0)-ethene complexes
is 155 - 151 Hz; for examples and exceptions to the rule, see
Table 2].
Unfortunately, it has not been possible to obtain crystals
suitable for an X-ray analysis, so the exact coordination is not
known. On the basis of spectroscopic data, 1 probably contains
a TP-3 Ni(0) center coordinated by the dabn N atoms and the
ethene ligand with its C atoms in the coordination plane (planar
conformation). Although no information is available about
possible rotation of the ethene ligand about the bond axis to
the nickel atom, we expect the bonding to be static as is found
for other L2Ni-alkene complexes.35
is very slow. Complex 1 is insoluble in diethyl ether or pentane
and dissolves sparingly in THF or toluene at -30 °C. In
solution, the product slowly decomposes above this temperature,
whereas the solid is stable to about 0 °C. It is therefore
important for a successful synthesis that the product rapidly
precipitates from the solution.
Chemical Properties. The reactivity of 1 toward CO, 1,5-
cyclooctadiene (cod), ethene, and ethyne has been studied. A
yellow suspension of complex 1 in diethyl ether or THF reacts
at -30 °C with 4 equiv of CO by displacing all ligands from
Ni(0) to yieldsvia the intermediate (C9H18N2)Ni(CO)2 (4; see
later)sa colorless solution of Ni(CO)4. Similarly, when cod is
added to a suspension of 1 at -30 °C, the complex slowly
converts into the yellow precipitate of Ni(cod)2 (identified by
its IR spectrum).31 These facile reactions that afford typical
Ni(0) complexes indicate that, in complex 1, (a) Ni is indeed
in the oxidation state 0 and (b) the coordination of the ligands
to Ni(0) is relatively weak. However, the exchange of the ethene
ligand with uncoordinated ethene is rather slow. This has been
shown for a THF-d8 solution of 1 saturated with ethene, which
Complex 1 is the only (diamine)Ni(0)-alkene complex
containing a nonactivated alkene ligand (ethene) that has been
isolated so far. It is important to distinguish this complex
from (N-donor)Ni(0)-ethene complexes such as T-4 (2,6-
1
exhibits in the H NMR spectrum (80 MHz) at -80 °C sharp
and at -20 °C somewhat broadened but separate signals for
uncoordinated (δ 5.30) and coordinated ethene (δ 0.27; see
below). When a suspension of 1 in diethyl ether is treated in
an autoclave with 55 bar of ethene at 20 °C for several hours,
no reaction is observed, i.e., the formation of a nickelacyclo-
pentane [(C9H18N2)Ni(C4H8)], a catalytic ethene dimerization
(butene isomers, cyclobutane), or oligo/polymerization does not
occur. Instead, after discharging the autoclave, complex 1 is
partially recovered. It appears that the otherwise thermally labile
complex 1 is stabilized by additional ethene.32 Furthermore,
complex 1 suspended in THF or diethyl ether reacts only very
slowly with ethyne at -78 °C. Most of the ethyne polymerizes
in time, although in solution (THF-d8) the formation of the
Ni(0)-ethyne complex 3 (see below) has been detected by 1H
NMR. The incomplete formation of complex 3 from 1 is
attributed to the low solubility and hence low reactivity of 1 at
-78 °C. These results correspond to our general observation
36
iPr2C6H3NdCHCH)NC6H3-2,6-iPr2)Ni(C2H4)2 and TP-3
(bpy)Ni(C2H4) (bpy ) 2,2′-bipyridine).37a In these compounds
the hybridization of the N atoms in the ligands is sp2, bringing
the ligands into the category of soft base. In contrast, dabn
with its sp3-hybridized N atoms must be considered a hard base.
The ethene ligand NMR data of the (N-donor)Ni(0)-ethene
(31) Bogdanovic, B.; Kro¨ner, M.; Wilke, G. Liebigs Ann. Chem. 1966,
699, 1.
(32) It cannot be ruled out that the dabn ligand is reversibly displaced
from Ni(0) when complex 1 is pressurized with ethene to form Ni(C2H4)3
(which is itself stabilized by ethene).
(33) Po¨rschke, K.-R. J. Am. Chem. Soc. 1989, 111, 5691.
(34) Nakamoto, K. Infrared and Raman Spectra of Inorganic and
Coordination Compounds, 3rd ed.; Wiley: New York, 1978; p 383.
(35) Po¨rschke, K.-R.; Pluta, C.; Proft, B.; Lutz, F. Z. Naturforsch., B:
Anorg. Chem., Org. Chem. 1993, 48, 608.
(36) Bonrath, W.; Po¨rschke, K.-R.; Michaelis, S. Angew. Chem. 1990,
102, 295; Angew. Chem., Int. Ed. Engl. 1990, 29, 298.

7996 J. Am. Chem. Soc., Vol. 119, No. 34, 1997
Haack et al.
complexes (Table 2) indicate an increase in back-bonding from
Ni(0) to the ethene ligand in the sequence 1,4-diazabutadiene
< bpy37b,c < dabn.
Hence, it appears that dabn is the by far the strongest N-donor
ligand to nickel(0), which is consistent with the fact that its
sp3-hybridized N atoms are simple σ-donors and lack a miti-
gating π-acceptor ability.
{(tmeda)Ni(η2-C4H6)}2(µ-η2,η2-C4H6) already mentioned (the
latter decomposes at -40 °C and is thus markedly less stable
than 2). As detailed NMR studies have shown, these dinuclear
bpy and tmeda complexes dissociate in solution (THF, toluene)
into the mononuclear T-4 complexes (bpy)Ni(η2-C4H6)2 and
(tmeda)Ni(η2-C4H6)2, respectively, to which 2a is in turn
related.39 It appears that, in the N-donor/Ni(0)/butadiene system,
the mononuclear butadiene complexes represent the basic
coordination mode of Ni(0) in solution, whereas the dinuclear
complexes separate from the solutions due to their lower
solubility.
{(C9H18N2)Ni(η2-C4H6)}2(µ-η2,η2-C4H6) (2) and (C9H18N2)-
Ni(η2-C4H6)2 (2a). A yellow suspension of 1 in diethyl ether
reacts at -30 °C with butadiene by displacement of the ethene
ligand to afford a red solution of the (nonisolated) mononuclear
complex 2a (for NMR characterization, see below). Small dark
red cubes of complex 2 slowly separate from this solution at
-78 °C (eqs 2 and 3). Crystalline 2 decomposes at 20 °C with
NMR Characterization of 2a. The 1H and 13C NMR spectra
of a THF-d8 solution of 2 (-80 °C) display relative signal
intensities dabn:butadiene of 1:2. The butadiene ligands show
six proton and four carbon signals of equal intensities for one
type of coordinated and one type of uncoordinated CdC bond.
The dabn ligand (Table 1) exhibits seven proton and five carbon
signals, indicating that the symmetry of the complex is reduced
to C2 (as compared with the more symmetric 1, 3, and 4). The
3
coupling constant of the butadiene dCH- protons of J(HH)
) 10.5 Hz indicates a trans conformation with respect to the
central single bond.40 The spectra are consistent with a T-4
Ni(0) center coordinated by the dabn N atoms and two
equivalent, singly bound, single-trans butadiene ligands. The
substituted C atoms of the coordinated butadiene CdC bonds
are chiral due to the complexation41 and necessarily of the same
configuration (R,R or S,S) consistent with the C2 symmetry of
the complex. Thus, the spectra are assigned to the racemate
2a (only one enantiomer is depicted in eqs 2-4).
No exchange of the coordinated and uncoordinated butadiene
CdC bonds of 2a takes place at -80 °C on the NMR time
scale. The NMR spectra give no information regarding a
possible rotation of the coordinated CdC bonds about the
coordination axis to the Ni atom. When the temperature is
raised, all signals broaden and a dynamic process, which has
not been investigated further, occurs. In the 80 MHz 1H NMR
spectrum, the signals are broad up to 0 °C, at which temperature
2a decomposes.
The low-temperature 1H and 13C NMR resonances of the dabn
NCH2 and NMe groups and the butadiene ligands of 2a almost
coincide with those of (tmeda)Ni(η2-C4H6)2.12 Bearing in mind
that the composition and chemical properties of the respective
dabn (2, 2a) and tmeda complexes are also alike, it can be
concluded that tmeda and dabn, if coordinated to Ni(0), will
exhibit similar electronic effects.
(C9H18N2)Ni(C2H2) (3). When a red slurry of the (dabn)-
Ni(0)-butadiene complex 2 in diethyl ether is exposed to excess
ethyne at -78 °C, 2 dissolves and a yellow precipitate of 3 is
formed in moderate yield (42%) (eq 4). Solid 3 decomposes
above -30 °C. The complex is almost insoluble in pentane or
diethyl ether and dissolves poorly in THF or toluene at -78
°C. The solutions are only stable below -60 °C. Complex 3
is thus far the only reported Ni(0)-ethyne complex that is
exclusively stabilized by an amine ligand.
elimination of butadiene. At -78 °C, the complex dissolves
poorly in diethyl ether or pentane but quite well in THF and
toluene, thereby reforming 2a. The red solutions are stable to
about 0 °C and produce a black precipitate (metallic nickel)
above this temperature.
In the EI mass spectrum of 2, the largest observable ion is
[(dabn)Ni(C4H6)]+, which fragments by cleaving the butadiene
ligand to afford [(dabn)Ni]+. In the IR spectrum (KBr), various
bands in addition to the dabn modes due to (inequivalent)
butadiene ligands are observed. On the basis of the composition
of the complex (elemental analysis), its properties, and the
analogy to similar systems, the dinuclear structure of complex
2 is beyond doubt. In 2 both Ni atoms are T-4 coordinated by
the bidentate dabn ligand with a singly bound butadiene ligand
and a butadiene ligand bridging the Ni centers. Compound 2
is analogous to the bpy complex {(bpy)Ni(η2-C4H6)}2(µ-η2,η2-
C4H6), for which a dinuclear structure has been confirmed by
an X-ray crystal structure analysis,38 and to the tmeda derivative
Spectroscopic Properties and Structure. In the IR spectrum
of 3, the sharp CtC stretching band of the ethyne ligand appears
(39) The equilibrium between the dinuclear complexes, containing 1.5
butadiene ligands per Ni atom, and the mononuclear species is complicated
because the mononuclear complexes contain 2 butadiene ligands per Ni
(37) (a) Mayer, W.; Po¨rschke, K.-R.; Kru¨ger, C. Unpublished results
(1978). (b) It follows from the 1H and 13C NMR data of (bpy)Ni(C2H4)
(Table 2), from the similar constitutions and NMR data of the (dabn, tmeda,
bpy)Ni(0)-butadiene complexes, and from the very small coupling constant
1J(CH) < 190 Hz for the complexes (bpy)Ni(HCtCR) [(bpy)Ni(HCtCH)
is not stable] that bpy is a stronger donor ligand for Ni(0) than are
phosphanes. Po¨rschke, K.-R.; Bonrath, W.; Michaelis, S. Unpublished
results. (c) Anisotropy effects of the bpy ligand on the chemical shifts of
the ethene ligand of (bpy)Ni(C2H4) are presumably small.
12
atom. For the systems, tmeda/Ni(0)/C4H6 and dabn/Ni(0)/C4H6, neither
[(tmeda)Ni(C4H6)] nor [(dabn)Ni(C4H6)] were identified by high-resolution
NMR, but further weak diamine signals were detected. At present the
dissociation equilibrium is tentatively described by the equation: 2{(di-
amine)Ni(η2-C4H6)}2(µ-η2,η2-C4H6) f 3(diamine)Ni(η2-C4H6)2 + [(di-
amine)Ni]. It is tempting to suggest that the species [(diamine)Ni] may be
a dimer [(diamine)NidNi(diamine)] with a d10(Ni)-d10(Ni) bond.
(40) Hobgood, R. T.; Goldstein, J. H. J. Mol. Spectrosc. 1964, 12, 76.
(41) Paiaro, G. Organomet. Chem. ReV. Sect. A 1970, 6, 319.
(38) Mayer, W.; Wilke, G.; Benn, R.; Goddard, R.; Kru¨ger, C. Monatsh.
Chem. 1985, 116, 879.

3,7-Diazabicyclo[3.3.1]nonane Nickel(0) Complexes
J. Am. Chem. Soc., Vol. 119, No. 34, 1997 7997
at significantly lower wave number (1560 cm-1) than those of
mass spectrometer (EI, 70 eV, 50 °C), the complex vaporizes
without decomposition. In the spectrum, the molecular ion (268,
1%) is detected and 2-fold loss of CO affords the base peak
[(dabn)Ni]+ (212).
analogous bis(phosphane)nickel(0)-ethyne complexes, e.g.,
(iPr2PC2H4PiPr2)Ni(C2H2) (1598 cm-1 42
(Table 2). In the
)
1H and 13C NMR spectra (-80 °C) of 3, the dabn ligand
(Table 1) gives rise to five proton and four carbon signals,
which are (on average) at slightly lower field than those
observed for the (dabn)Ni(0)-ethene complex 1. The ethyne
ligand displays signals at δ(H) 4.65 and δ(C) 118.7 with a
For dissolved 4, the 1H and 13C NMR data of the dabn ligand
(Table 1) resemble those of the (dabn)Ni(0) complexes 1-3.
The 13C resonance of the CO ligand is found at δ ) 196.0,
which is, surprisingly, not much different from that of
Ni(CO)4.44,45 In the IR spectrum of 4 (pentane), two strong
ν(CO) bands are observed at 1984 (A1) and 1896 cm-1 (B1),
which are at a lower wave number than for other L2Ni-
(CO)2 complexes [e.g., (tBu2PC2H4PtBu2)Ni(CO)2 1994, 1935
cm-1].35
1
coupling constant of J(CH) ) 178 Hz. In order to assess the
information the ethyne resonances provide, one has to recall
that for mononuclear bis(phosphane)nickel(0)-ethyne com-
plexes, e.g., (iPr2PC2H4PiPr2)Ni(C2H2) [δ(H) 7.29; δ(C) 123.8;
1J(CH) ) 202 Hz],42 the 1H and 13C resonances are at low field
relative to those of uncoordinated ethyne (Table 2). This shift
is reversed43 for dinuclear complexes in which ethyne is ligated
by two nickel(0) moieties, e.g., {(iPr2PC2H4PiPr2)Ni}2(µ-C2H2)
Thus, we conclude that in complex 4 a T-4 Ni(0) center is
coordinated by dabn and two CO ligands. The structure of 4
is related to those of the (dabn)Ni(0)-butadiene complexes 2
and 2a. Complex 4 may be considered as the first well-
characterized complex of the type (diamine)Ni(CO)2, particularly
since (NH3)2Ni(CO)2 has not been isolated in a pure form (see
the Introduction) and the hypothetical [(tmeda)Ni(CO)2] is not
accessible in a similar reaction.6
1
[δ(H) 5.52; δ(C) 86.3; J(CH) ) 188 Hz].42 The coupling
constant 1J(CH) ) 249 Hz, as attributed to uncoordinated
ethyne, decreases to about 200 Hz upon complexation of the
ethyne molecule to a single bis(phosphane)nickel(0) moiety (as
verified by us for more than 30 compounds). A further,
moderate reduction down to 187 Hz is observed only when the
ethyne molecule is ligated to two bis(phosphane)nickel(0)
moieties (verified by us for approximately 10 compounds).43
Thus, although the 1H and 13C chemical shifts observed for the
ethyne ligand in complex 3 appear to be unexceptional, the
coupling constant 1J(CH) ) 178 Hz is extremely low and is by
far the smallest for any known Ni(0)-ethyne complex (Table
2). We conclude from the IR and NMR data that, in complex
3, TP-3 Ni(0) is coordinated by the dabn N atoms and one
ethyne ligand and that an exceedingly large charge transfer from
Ni(0) into an antibonding orbital of the ethyne ligand occurs.
(C9H18N2)Ni(CO)2 (4). Complexes 1-3 react with CO (-30
°C) by the exchange of all ligands to give Ni(CO)4. This raised
the question as to whether complex 4, which is conceived as
an intermediate in these reactions, is stable or not. When the
neat colorless liquids of Ni(CO)4 (in excess) and dabn are mixed
at 20 °C, the mixture changes color to orange and CO is evolved.
After a short period, pure microcrystalline orange needles of
the (dabn)Ni(0)-CO complex 4 separate (85%) (eq 5). On the
other hand, when 4 is exposed to CO at -30 °C the dabn ligand
is readily displaced to form Ni(CO)4.
Discussion
The aim of this work was to prepare (diamine)Ni(0)-ethene
and -ethyne complexes, and this was eventually achieved using
the bicyclic diamine, N,N′-dimethyl-3,7-diazabicyclo-[3.3.1]nonane
(dabn). In this paper, we describe our search for a suitable
ligand and the preparation and properties of four novel (di-
amine)Ni(0) complexes containing ethene (1), butadiene (2, 2a),
ethyne (3), and CO (4) using this ligand. Prior to this study,
no (diamine)Ni(0) complex with ethene and ethyne as the parent
alkene and alkyne was known. It was also found that the dabn
complexes 2 and 2a (dec. 20 °C) containing butadiene are
markedly more stable than the already known tmeda derivatives
(dec. -40 °C). Similarly, the rather stable CO complex 4 (dec.
130 °C) was preceded only by reports of the highly labile and
poorly characterized (NH3)2Ni(CO)2 (dec. -60 °C). All previ-
ous attempts to prepare complexes such as (tmeda)Ni(C2H4),
(tmeda)Ni(C2H2), and (tmeda)Ni(CO)2 have been unsuccessful.
Dabn is currently unique among hard chelating diamine
ligands in its ability to ligate the soft Ni(0) in combination not
only with stronger π-acceptors but also with nonactivated
alkenes or alkynes. Preliminary studies have shown that neither
the bispidine derivatives 3,7-diazabicyclo[3.3.1]nonan-9-one nor
the commercially available (-)-sparteine coordinate to the
[Ni(0)(C2H4)] fragment. (-)-Sparteine, which forms stable
complexes with metal ions,46 prefers the chair/boat conforma-
tion,47 and this conformation must be converted into a chair/
chair conformation before chelating coordination is possible. It
would appear that the energy required for this conformational
change is too large to be compensated by the energy gained
upon forming two NfNi(0)(C2H4) coordinative bonds. Per-
haps we should mention here that we have not yet tried
the symmetrical R-isosparteine, which is a stronger binding
ligand than (-)-sparteine due to its rigid chair/chair conforma-
tion.48,49
As expected, complex 4 is thermally more stable than the
Ni(0)-alkene complexes 1-3. The solid may be kept indefi-
nitely at 20 °C. When the complex is heated to 50 °C, the color
slowly changes to red, but a rapid decomposition of the crystals
takes place only at 130 °C. Also, under the conditions of the
(42) Po¨rschke, K.-R. Angew. Chem. 1987, 99, 1321; Angew. Chem., Int.
Ed. Engl. 1987, 26, 1288.
(44) The 13C NMR signal of Ni(CO)4 is at δ 191.6. For L-Ni(CO)3
and L2Ni(CO)2 complexes, the resonance of the carbon monoxide ligand is
shifted to lower field, and the magnitude of the low-field shift generally
follows the donor strength of the ligand L; e.g., (Me3P)Ni(CO)3: 196.6,44a
Li+[CH3-Ni(CO)3]-: 209.5,44b (tBu2PC2H4PtBu2)Ni(CO)2: 204.8.35 (a)
Bodner, G. M. Inorg. Chem. 1975, 14, 1932. (b) Kleimann, W.; Po¨rschke,
K.-R.; Wilke, G. Chem. Ber. 1985, 118, 323 and literature cited therein.
(45) The changes in the ligand NMR data of T-4 Ni(0) complexes can
be relatively small. This is due to the fact that back-bonding from Ni(0)
into acceptor orbitals on the π-ligand is significantly smaller than for TP-3
Ni(0) complexes. Po¨rschke, K.-R.; Mynott, R.; Kru¨ger, C.; Romao, M. J.
Z. Naturforsch., B: Anorg. Chem., Org. Chem. 1984, 39, 1076.
(43) The changes in the chemical shift may be explained by a formal
change in the hybridization of the ethyne carbon atoms sp f sp2 f sp3
upon coordination of the triple bond to one and two nickel atoms,
respectively. Thus, in the first instance, the protons and carbon atoms
become olefinic due to the limiting structure of a metallacyclopropene
(downfield shift), while upon formal coordination of this metallacyclopro-
pene to a second metal center (limiting structure of a dimetallabicyclobutane)
they become aliphatic (upfield shift). The steady decrease of the coupling
constant is also in agreement with the anticipated change in hybridization
(hydrocarbons: sp, 249 Hz; sp2, 156 Hz; sp3, 125 Hz).33

7998 J. Am. Chem. Soc., Vol. 119, No. 34, 1997
Haack et al.
One feature favoring coordination of the dabn ligand to Ni(0)
is the arrangement of the donor N atoms in the molecule. Not
only is the diamine structurally preorganized but the N atoms
are also ideally prepositioned for coordination to the metal.50
The principles of preorganization and prepositioning are es-
sential components of the well-developed concept of the
macrocyclic effect.51,52 Moreover, the species formed when
dabn is bound to Ni(0) is structurally highly organized.
Coordination of dabn to the Ni(0) center results in the formation
of two additional six-membered rings and turns the dabn
molecule with its semi adamantane type structure into one with
a rigid full adamantane type structural motif.53
In order to assess the influence dabn exerts to Ni(0) relative
to other donor ligands, it is instructive to look at the properties
of the coligated π-acceptor ligands. In the T-4 (dabn)Ni(0)-
butadiene complexes, the electronic influence of bicyclic dabn
is similar to that of open-chain tmeda, which only coordinates
to Ni(0) in combination with π-ligands of moderate (e.g.,
butadiene) to high acceptor strength. No assessment of the
donor strength of tmeda has so far been made, and there are no
phosphane complexes of the type (R2PC2H4PR2)Ni(η2-C4H6)2
known to which 2a and (tmeda)Ni(η2-C4H6)2 could be com-
pared. For the T-4 (dabn)Ni(0)-carbonmonoxide complex 4,
the NMR CO resonance [δ(C) 196.0] is almost of the same
magnitude as for Ni(CO)4, and it occurs at a much higher field
than expected (∼210).44 No reasonable explanation can be
given at present. In contrast, in the IR spectrum of 4 the CO
stretching absorptions appear at very low wave numbers,
indicating a high degree of back-bonding from Ni(0) to the
acceptor ligand. This effect is even stronger in the TP-3
(dabn)Ni(0)-ethene and -ethyne complexes 1 and 3 (see Table
2). For both compounds the CsH, CdC, and CtC stretching
vibrations occur at lower wave numbers than for any other
common Ni(0)-ethene or -ethyne complex. Furthermore, in
the NMR spectra, the ethene ligand protons and carbon atoms
of 1 are strongly shielded [δ(H) 0.27; δ(C) 20.4], and the
resonances almost correspond to those expected for a saturated
hydrocarbon. The effect on the chemical shift is less obvious
for the ethyne protons and carbon atoms of 3, but for both ethene
and ethyne ligands the coupling constants of 1J(CH) ) 142 Hz
(1) and 178 Hz (3) are exceptionally low.
These findings indicate that the dabn ligand coordinated to
Ni(0) induces an exceedingly strong charge transfer from Ni(0)
into an antibonding π-ligand orbital.54 According to the IR and
NMR data, the charge transfer induced is markedly stronger
for the dabn ligand than for a diphosphane ligand or for bpy37b
(and even more so than for phosphite type and 1,4-diazabuta-
diene ligands). The donor effect of the dabn ligand is
intermediate between that of diphosphane (or bpy) ligands and
carbanionic ligands (Table 2), of which the latter represent the
strongest donor ligands known.55
The very large ligand effect of the diamine toward Ni(0),
observed spectroscopically for complexes 1, 3, and 4, is
consistent with ab initio MO-SCF calculations on the hypotheti-
cal amine and phosphane TP-3 Ni(0)-ethene complexes
(H3N)2Ni(C2H4) and (H3P)2Ni(C2H4).8 These indicate a charge
distribution in (H3N)2Ni(C2H4) of +0.11 e (NH3), +0.58 e (Ni),
-0.78 e (C2H4; calculated CdC bond length 1.45 Å), and in
(H3P)2Ni(C2H4) of +0.07 e (PH3), +0.35 e (Ni), -0.48 e (C2H4;
the calculated CdC bond length of 1.41 Å agrees well with
experimental data). The data confirms that the hard amine is a
stronger donor toward Ni(0) than the soft phosphane, which
can also act as an acceptor. Furthermore, the data show that,
although in both compounds the negative charge on the ethene
ligand is mainly compensated by the positive charge at the nickel
center, the polarization of the Ni(0)-C2H4 coordinative bond
is considerably larger in (H3N)2Ni(C2H4) than in (H3P)2Ni(C2H4).
In terms of the Pearson HSAB concept, the hard amine ligand
forces the soft nickel atom to transfer charge to the π-ligand,
causing Ni(0) to become less soft and hence able to coordinate
the hard amine ligand more strongly.56
(46) (a) For Li, see: Aratani, T.; Gonda, T.; Nozaki, H. Tetrahedron
1970, 26, 5453. Byrne, L. T.; Engelhardt, L. M.; Jacobsen, G. E.; Leung,
W.-P.; Papasergio, R. I.; Raston, C. L.; Skelton, B. W.; Twiss, P.; White,
A. H. J. Chem. Soc., Dalton Trans. 1989, 105. Marsch, M.; Harms, K.;
Zschage, O.; Hoppe, D.; Boche, G. Angew. Chem. 1991, 103, 338; Angew.
Chem., Int. Ed. Engl. 1991, 30, 321. Hoppe, I.; Marsch, M.; Harms, K.;
Boche, G.; Hoppe, D. Angew. Chem. 1995, 107, 2328; Angew. Chem., Int.
Ed. Engl. 1995, 34, 2158. (b) For Mg, see: Fraenkel, G.; Cottrell, C.;
Ray, J.; Russell, J. J. Chem. Soc., Chem. Commun. 1971, 273. Fraenkel,
G.; Appleman, B.; Ray, J. G. J. Am. Chem. Soc. 1974, 96, 5113. (c) For
Co(II), Ni(II), Cu(II), Zn(II), Mn(III), and Fe(III), see: Mason, S. F.;
Peacock, R. D. J. Chem. Soc., Dalton Trans. 1973, 226. (d) For Pd(II),
see: Togni, A.; Rihs, G.; Pregosin, P. S.; Ammann, C. HelV. Chim. Acta
1990, 73, 723. (e) For Zn, see: Motevalli, M.; O’Brien, P.; Robinson, A.
J.; Walsh, J. R.; Wyatt, P. B. J. Organomet. Chem. 1993, 461, 5.
(47) Bohlmann, F.; Zeisberg, R. Chem. Ber. 1975, 108, 1043 and
literature cited therein.
(48) Przybylska, M.; Barnes, W. H. Acta Crystallogr. 1953, 6, 377.
(49) For Mg, see: Kageyama, H.; Miki, K.; Kai, Y.; Kasai, N.; Okamoto,
Y.; Yuki, H. Acta Crystallogr. B 1982, 38, 2264. Okamoto, Y.; Suzuki,
K.; Kitayama, T.; Yuki, H.; Kageyama, H.; Miki, K.; Tanaka, N.; Kasai,
N. J. Am. Chem. Soc. 1982, 104, 4618.
(50) A change in the hybridization of the donor upon complex formation
must be taken into account when choosing an optimal geometry for a
potential ligand.
Conclusion
Hancock and Martell have predicted that high leVels of
preorganization are not confined to macrocycles or cryptates
and can just as easily occur in chelating ligands,52b but no
experimental evidence was given. Our study has verified this
presumption, since the coordination of a diamine to the
[Ni(0)(C2H4)] fragment is possible only when the principles of
preorganization (ligand) and prepositioning (donor atoms) are
applied to the diamine ligand. These principles, which were
developed from studies with macrocyclic ligands and thus are
collectively known as macrocyclic effect, aresas shown
heresalso valid for common and medium-sized rings and bi-
or multicyclic ring systems.57 The denotation macrocyclic effect
with its implication of large rings is therefore too narrow an
expression, and since it is the position of the donor atoms and
(54) (a) Dewar, M. J. S. Bull. Soc. Chim. Fr. 1951, C71. (b) Dewar, M.
J. S.; Ford, G. P. J. Am. Chem. Soc. 1979, 101, 783. (c) Chatt, J.;
Duncanson, L. A. J. Chem. Soc. 1953, 2939.
(55) (a) Po¨rschke, K.-R.; Jonas, K.; Wilke, G. Chem. Ber. 1988, 121,
1913. (b) Jonas, K.; Kru¨ger, C. Angew. Chem. 1980, 92, 513; Angew.
Chem., Int. Ed. Engl. 1980, 19, 520. Jonas, K. AdV. Organomet. Chem.
1981, 19, 97 and literature cited therein.
(51) (a) Cabbiness, D. K.; Margerum, D. W. J. Am. Chem. Soc. 1969,
91, 6540. (b) Hinz, F. P.; Margerum, D. W. J. Am. Chem. Soc. 1974, 96,
4993; Inorg. Chem. 1974, 13, 2941.
(52) (a) Izatt, R. M.; Bradshaw, J. S.; Nielsen, S. A.; Lamb, J. D.;
Christensen, J. J. Chem. ReV. 1985, 85, 271. (b) Hancock, R. D.; Martell,
A. E. Comments Inorg. Chem. 1988, 6, 237 and literature cited therein.
(53) Although illustrative, this fact alone does not explain the stability
of the complexes because both (-)-sparteine and N,N′-dimethyl-3,7-
diazabicyclo[3.3.1]nonan-9-one fail to form a nickel(0)-ethene complex.
The stability of complexes 1-4 is based on the optimal geometry of free
dabn, which is superior to that of the above-mentioned reference diamines
with a bispidine skeleton (leaving R-isosparteine out of consideration). It
is, however, conceivable that a different bicyclic or multicyclic structure
would also give a satisfactory arrangement of the N atoms.
(56) For a detailed treatment of this matter, see: Pearson, R. G. Inorg.
Chem. 1988, 27, 734.
(57) (a) Examples of common and medium-sized rings in coordination
chemisty: 1,5-cyclooctadiene (cod), 1,4,7-triazacyclononane, 1,5,9-cy-
clododecatriene isomers. Example of bicyclic rings: norbornadiene. Ex-
amples of multicyclic rings: sparteine, R-isosparteine. (b) Earlier we used
the term macrocyclic effect to explain the stability of certain cod complexes36
and the term cyclic effect with reference to Ni(cdt).35

3,7-Diazabicyclo[3.3.1]nonane Nickel(0) Complexes
J. Am. Chem. Soc., Vol. 119, No. 34, 1997 7999
liquor by means of a capillary, washed twice with pentane, and dried
in vacuo; yield 400 mg (83%). The solid decomposes rapidly at 50
°C and slowly at 0 °C (decoloration to grey). The complex is very
sensitive to oxygen. IR (KBr): 2990, 2980 (ν C-H), 1510 [ν(CdC)
+ δs(CH2)], 1115 (Fw CH2), 865 cm-1 (Fr CH2). EI-MS (70 eV, 50
°C): m/e (%) 240 (M+, <1), 212 ([M - C2H4]+, <1), 154 ([dabn]+,
90), 58 ([Me2NdCH2]+, 100), 42 ([H2CdNdCH2]+, 40). 1H NMR
(400 MHz, -30 °C): δ 0.27 (s, 4H, C2H4); for dabn resonances see
Table 1. 13C NMR (100.6 MHz, -30 °C): δ 20.4 (2C, 1J(CH) ) 142
Hz, C2H4); for dabn resonances see Table 1. Anal. Calcd for
C11H22N2Ni (241.0): C, 54.82; H, 9.20; N, 11.62; Ni, 24.36. Found:
C, 54.81; H, 9.24; N, 11.68; Ni, 24.31.
the direction of the lone-pairs which is important, we would
suggest that the term prepositioning effect is a more suitable
alternative.
In view of the interesting ligand properties of dabn reported
here (increased binding and electron donation), we expect that
dabn and related molecules will lead to novel complexes with
unusual stability and metals with unusual oxidation states.58
Experimental Section
The complexes were handled under an atmosphere of argon with
Schlenk type glassware in order to exclude oxygen and moisture.
Microanalyses were performed by Microanalytisches Labor Dornis und
Kolbe, Mu¨lheim. 1H NMR spectra were recorded at 80, 200, and 400
MHz and 13C NMR spectra were at 50.3, 75.5, and 100.6 MHz on
Bruker WP-80, AM-200, AX-300, and WH-400 instruments (relative
to internal tetramethylsilane). For the NMR spectra, solutions of the
compounds in THF-d8 were used unless otherwise indicated. Solutions
of thermolabile complexes were prepared under an argon atmosphere
at the temperature at which the spectra were recorded. IR spectra were
taken on a Nicolet 7199 FT-IR system, Raman were on a Coderg-LRT
800 instrument (excitation: argon lines at 488.0 and 514.5 nm), and
{(C9H18N2)Ni(η2-C4H6)}2(µ-C4H6) (2). Butadiene (2 mL) was
added to a suspension of 1 (482 mg, 2.00 mmol) in diethyl ether (20
mL) at -40 °C. The color of the mixture changed to red, and the
mixture was stirred at -40 °C until all 1 was dissolved. The solution
was briefly warmed to 20 °C, and pentane (20 mL) was added. In the
course of 12 h, dark red crystals separated at -78 °C. The product
was recrystallized from diethyl ether/pentane and dried under vacuum
at -30 °C; yield 310 mg (53%). IR (KBr): 3065, 2995 (ν C-H),
1580 (ν CdC, uncoordinated), 1190, 655 cm-1
. EI-MS (70 eV, 0
°C): m/e (%) 266 ([(dabn)Ni(C4H6)]+, <1), 212 ([(dabn)Ni]+, 2), 154
1
([dabn]+, 100), 58 ([Me2NdCH2]+, 30). For H and 13C NMR data,
2
EI mass spectra were on a Finnigan MAT 8200. Ni(C2H4)3 was
prepared from Ni(cdt),31 as published (cdt ) trans,trans,trans-1,5,9-
cyclododecatriene).
see that for 2a. Anal. Calcd for C30H54N4Ni2 (588.2): C, 61.26; H,
9.25; N, 9.52; Ni, 19.96. Found: C, 61.35; H, 9.26; N, 9.48; Ni, 19.88.
(C9H18N2)Ni(η2-C4H6)2 (2a). This complex was observed when 2
was dissolved in THF-d8. 1H NMR (400 MHz, -80 °C) δ 5.49 (m,
2H, dCH-uncoord), 4.80 (dd, 2H, dCHZHuncoord), 4.52 (dd, 2H, dCHHE-
,uncoord), 3.45 (m, 2H, dCH-coord), 1.87 (d, 2H, dCHHE,coord), 1.59 (d,
2H, dCHZHcoord); for dabn resonances see Table 1. The assignment
of the butadiene proton signals is in accordance with a COSY 2D-
NMR experiment. 13C NMR (100.6 MHz, -80 °C): δ 145.6 (2C,
-CHduncoord), 97.9 (2C, dCH2,uncoord), 62.2 (2C, -CHdcoord), 48.8 (2C,
dCH2,coord); for dabn resonances see Table 1.
(C9H18N2)Ni(C2H2) (3). A red suspension of 2 (588 mg, 1.00 mmol)
in diethyl ether (10 mL) was exposed to ethyne (48 mL, 2 mmol) at
-78 °C. Within a few minutes, the suspension turned yellow. The
precipitate was isolated from the mother liquor, washed twice with
pentane at -78 °C, and dried under vacuum at -78 °C; yield 200 mg
(42%); dec. as a solid >-30 °C, dissolved >-60 °C (THF). IR (KBr,
-80 °C): 3010, 2980 (ν tCH), 1560 (ν CtC), 900, 630 cm-1 (γCCH).
1H NMR (200 MHz, -80 °C): δ 4.65 (s, 2H, C2H2); for dabn
resonances, see Table 1. 13C NMR (75.5 MHz, -80 °C): δ 118.7
(2C, 1J(CH) ) 178 Hz, C2H2); for dabn resonances, see Table 1. Anal.
Calcd for C11H20N2Ni (239.0): C, 55.28; H, 8.44; N, 11.72; Ni, 24.56.
Found: C, 55.20; H, 8.78; N, 11.60; Ni, 24.51.
N,N′-Dimethyl-3,7-diazabicyclo[3.3.1]nonan-9-one. The synthesis
was carried out according to the published procedure24 but on a pilot-
plant scale (30 L reaction vessel) because of the poor yield: 1-methyl-
4-piperidone (12 mol; Aldrich), N-methylamine (12 mol; Fluka),
paraformaldehyde (24 mol), acetic acid (24 mol), methanol (20 L);
concentric tube column rectification; yield 60 g (3%); colorless solid;
mp 44 °C; bp 53 °C (0.07 mbar); C9H16N2O (168.2) [14789-54-9]. IR
(neat or THF): 1740, 1706 cm-1 (ν CdO). Raman: 1734, 1713 cm-1
(ν CdO). EI-MS (70 eV, 40 °C): m/e (%) 168 (M+, 14), 58
([Me2NdCH2]+, 100), 42 ([H2CdNdCH2]+, 69). 1H NMR (400 MHz,
27 °C): δ 2.95 (dd, 4H, Heq), 2.69 (dd, 4H, Hax), 2.39 (m, 2H, CtertH),
2.22 (s, 6H, CH3). 13C NMR (50.3 MHz, 27 °C): δ 211.7 (1C, CdO),
61.3 (4C, CH2), 47.1 (2C, CtertH), 44.8 (2C, NCH3).
1-(Methoxymethyl)-N,N′-dimethyl-3,7-diazabicyclo[3.3.1]nonan-
9-one. The substance is a byproduct of the synthesis described above
(not mentioned in ref 24); yield 56 g (2.2%); C11H20N2O2 (212.3); bp
62 °C (0.07 mbar). IR (neat): 1730 (ν CdO), 1110 cm-1 (CH2OCH3).
EI-MS (70 eV, 40 °C): m/e (%) 212 (M+, 20), 58 ([Me2NdCH2]+,
100), 42 ([H2CdNdCH2]+, 25). 1H NMR (200 MHz, CDCl3, 27 °C):
δ 3.39 (2H, CH2O), 3.31 (3H, OCH3), 2.27 (6H, NCH3), 3.10, 2.97,
2.74, 2.50 (each dd, 2H, NCHaHb and NC′HaHb), 2.54 (m, 1H, CtertH).
13C NMR (50.3 MHz, CDCl3, 27 °C): δ 213.9 (1C, CdO), 74.4 (1C,
CH2O), 60.8, 64.0 (each 2C, NCH2 and NC′H2), 59.4 (1C, OCH3), 50.1
(1C, Cquat), 46.5 (1C, CtertH), 44.9 (2C, NCH3).
(C9H18N2)Ni(CO)2 (4). Ni(CO)4 (1.0 g, excess) was mixed with
dabn (771 mg, 5.00 mmol) at 20 °C in the absence of a solvent. CO
evolved from the initially colorless mixture, and an orange crystalline
product was formed within 1 h. Excess Ni(CO)4 was evaporated under
vacuum to afford the analytically pure product; yield 1.14 g (85%).
EI-MS (70 eV, 50 °C): m/e (%) 268 (M+, 1), 240 ([M - CO]+, 48),
212 ([M - 2CO]+, 100), 154 ([dabn]+, 24). IR: (pentane) 1984 (CO,
A1), 1896 cm-1 (B1); (THF) 1971, 1879 cm-1; (KBr) 1970, 1880 cm-1
N,N′-Dimethyl-3,7-diazabicyclo[3.3.1]nonane. The compound was
synthesized as published24 from N,N′-dimethyl-3,7-diazabicyclo[3.3.1]-
nonan-9-one (30 g, 180 mmol); concentric tube column rectification;
yield 16.5 g (60%); C9H18N2 (154.2) [14789-33-4]; bp 54 °C (28 mbar);
20
nD ) 1.4902. EI-MS (70 eV, 40 °C): m/e (%) 154 (M+, 23), 124
([M - 2CH3]+, 16), 58 ([Me2NdCH2]+, 77), 42 ([H2CdNdCH2]+,
1
(broad). For H (400 MHz, 27 °C) and 13C (50.3 MHz, 27 °C) NMR
1
100). For H (400 MHz, 27 °C) and 13C (50.3 MHz, 27 °C) NMR
data, see the text and Table 1. Anal. Calcd for C11H18N2NiO2
(269.0): C, 49.12; H, 6.75; N, 10.42; Ni, 21.82. Found: C, 49.07; H,
6.78; N, 10.46; Ni, 21.69.
data, see Table 1.
(C9H18N2)Ni(C2H4) (1). To a solution of Ni(C2H4)3, prepared from
Ni(cdt) (466 mg, 2.00 mmol) and ethene in diethyl ether (20 mL) at 0
°C, is added a solution of dabn (0.4 mL; excess) in diethyl ether (10
mL) at -10 °C. Over the course of 30 min, a yellow microcrystalline
precipitate forms. At -30 °C the product is separated from the mother
Acknowledgment. We thank Dr. Klaus Seevogel for the IR
studies and Dr. Graham Russell for helpful discussions. Further
thanks are due to the Fonds der Chemischen Industrie for
financial support, including an FCI stipend granted to K.-J.H.
(58) Since the reaction used to prepare dabn only has a yield of ∼2%,
we would welcome a more efficient synthesis of dabn.
(59) Michaelis, S. Ph.D. Dissertation, Universita¨t Bochum, 1991.
JA971037V