Monomeric cyclopentadienylnickel methoxo and amido complexes: Synthesis, characterization, reactivity, and use for exploring the relationship between H-X and M-X bond energies

Article abstract of DOI:10.1021/ja971829p

Reactive monomeric amino and methoxo complexes, Cp*Ni(PEt₃)NHTol and Cp*Ni(PEt₃)OMe, have been synthesized and fully characterized. The former is the first monomeric 18-electron nickel amide to be synthesized and the latter is the first structurally characterized monomeric nickel methoxide complex. The amido complex Cp*Ni(PEt₃)NHTol reacts with various Bronsted acids (HX) to produce complexes of the type Cp*Ni(PEt₃)X (X = NHAr, OR, Osilica, SR), and compounds with hydridic hydrogens to give the hydridonickel complex Cp*Ni(PEt₃)H. The polarity of Ni-N and Ni-O bonds is also demonstrated by reactions with alkali metal salts and trimethylsilyl chloride, and by the crystallographic and NMR characterization of phenol adducts of Cp*Ni(PEt₃)OTol. The phosphine ligands in Cp*Ni(PEt₃)X (X = OTol, SAr) compounds exchange with PMe₃ through an associative mechanism; the rate increases with the electronegativity of X. The thermodynamics of reactions interconverting Cp*Ni(PEt₃)X + HX' and Cp*Ni(PEt₃)X' + HX have been analyzed using solution equilibrium studies and calorimetry. Instead of being 1:1, the correlation between H-X and M-X bond energies shows a marked preference for nickel binding to more electronegative ligands. This preference is not specific to nickel: examples of similar thermodynamic preferences occur throughout transition metal chemistry. These results may be attributable to a large electrostatic component in the bonding between Ni and X. This qualitative E-C model explains the reactivity, thermodynamics, and phosphine exchange rates of this series of nickel complexes, and may be general to many metal-ligand bonds.

Full text of DOI:10.1021/ja971829p

12800
J. Am. Chem. Soc. 1997, 119, 12800-12814
Monomeric Cyclopentadienylnickel Methoxo and Amido
Complexes: Synthesis, Characterization, Reactivity, and Use
for Exploring the Relationship between H-X and M-X Bond
Energies
Patrick L. Holland,† Richard A. Andersen,*,† Robert G. Bergman,*,†
Jinkun Huang,‡ and Steven P. Nolan‡
Contribution from the Department of Chemistry, UniVersity of California,
Berkeley, California 94720, and the Department of Chemistry, UniVersity of New Orleans,
New Orleans, Louisiana 70148
ReceiVed June 4, 1997X
Abstract: Reactive monomeric amido and methoxo complexes, Cp*Ni(PEt3)NHTol and Cp*Ni(PEt3)OMe, have
been synthesized and fully characterized. The former is the first monomeric 18-electron nickel amide to be synthesized
and the latter is the first structurally characterized monomeric nickel methoxide complex. The amido complex Cp*Ni-
(PEt3)NHTol reacts with various Brønsted acids (HX) to produce complexes of the type Cp*Ni(PEt3)X (X ) NHAr,
OR, Osilica, SR), and compounds with hydridic hydrogens to give the hydridonickel complex Cp*Ni(PEt3)H. The
polarity of Ni-N and Ni-O bonds is also demonstrated by reactions with alkali metal salts and trimethylsilyl chloride,
and by the crystallographic and NMR characterization of phenol adducts of Cp*Ni(PEt3)OTol. The phosphine ligands
in Cp*Ni(PEt3)X (X ) OTol, SAr) compounds exchange with PMe3 through an associative mechanism; the rate
increases with the electronegativity of X. The thermodynamics of reactions interconverting Cp*Ni(PEt3)X + HX′
and Cp*Ni(PEt3)X′ + HX have been analyzed using solution equilibrium studies and calorimetry. Instead of being
1:1, the correlation between H-X and M-X bond energies shows a marked preference for nickel binding to more
electronegative ligands. This preference is not specific to nickel: examples of similar thermodynamic preferences
occur throughout transition metal chemistry. These results may be attributable to a large electrostatic component in
the bonding between Ni and X. This qualitative E-C model explains the reactivity, thermodynamics, and phosphine
exchange rates of this series of nickel complexes, and may be general to many metal-ligand bonds.
Introduction
the factors controlling the nature of M-N and M-O bonds are
still not well understood.
Late transition metal compounds with anionic nitrogen and
oxygen ligands are important intermediates in industrial1,2 and
biological3 processes. While many new complexes of this type
have been synthesized recently by our group4-7 and others,8-21
In this work, new examples of monomeric nickel amido and
alkoxo complexes are presented. Their reactivities are governed
by the high polarity of their Ni-X bonds. In most cases, the
products of the reactions have been crystallographically char-
acterized; the structural details are presented separately with a
study of effects on the ancillary pentamethylcyclopentadienyl
(Cp*) ligand.22 In addition to studying the reactivity of these
compounds toward X-H bonds, we have used equilibrium
studies and substituent effect correlations to gain insight into
the electronic nature of the Ni-X bond. These results run
counter to the current model for predicting late metal-X bond
energies and we offer a more general qualitative model based
on E-C theory. The new model follows from trends observed
in the reactivity, stability, and thermodynamic preferences of
all covalent bonds between late metals and ligands, and it shows
promise of further understanding of how electronic effects
change M-X bond energies.
† University of California, Berkeley.
‡ University of New Orleans.
X Abstract published in AdVance ACS Abstracts, December 15, 1997.
(1) Roundhill, D. M. Chem. ReV. 1992, 92, 1.
(2) Bryndza, H. E.; Tam, W. Chem. ReV. 1988, 88, 1163.
(3) For leading references, see: Holm, R. H.; Kennepohl, P.; Solomon,
E. I. Chem. ReV. 1996, 96, 2239, and other articles in the same issue.
(4) Bergman, R. G. Polyhedron 1995, 14, 3227.
(5) Kaplan, A. W.; Bergman, R. G. Organometallics 1997, 16, 1106.
(6) For related chemistry of a late-metal fluoride, see: Veltheer, J. E.;
Burger, P.; Bergman, R. G. J. Am. Chem. Soc. 1995, 117, 12478.
(7) Holland, P. L.; Andersen, R. A.; Bergman, R. G. J. Am. Chem. Soc.
1996, 118, 1092.
(8) Kawataka, F.; Kayaki, Y.; Shimizu, I.; Yamamoto, A. Organome-
tallics 1994, 13, 3517.
(9) Brunet, J.-J.; Commenges, G.; Neibecker, D.; Phillippot, K.; Rosen-
berg, L. Inorg. Chem. 1994, 33, 6373.
(10) Koo, K.; Hillhouse, G. L. Organometallics 1995, 14, 4421.
(11) VanderLende, D. D.; Abboud, K. A.; Boncella, J. M. Inorg. Chem.
1995, 34, 5319.
(12) Bennett, M. A.; Jin, H.; Li, S.; Rendina, L.; Willis, A. C. J. Am.
Chem. Soc. 1995, 117, 8335.
(13) Grushin, V. V.; Alper, H. A. Organometallics 1996, 15, 5242.
(14) Driver, M. S.; Hartwig, J. F. J. Am. Chem. Soc. 1996, 118, 7217.
(15) Mann, G.; Hartwig, J. F. J. Am. Chem. Soc. 1996, 118, 13109.
(16) Woffe, J. P.; Wagaw, S.; Buchwald, S. L. J. Am. Chem. Soc. 1996,
118, 7215.
Results
Synthesis of a Nickel Triflate. The synthesis of Cp*Ni-
(PEt3)X complexes proceeded from the conveniently prepared
(20) Canty, A. J.; Jin, H.; Roberts, A. S.; Skelton, B. W.; White, A. H.
Organometallics 1996, 15, 5713.
(17) Galanski, M.; Keppler, B. K. Inorg. Chem. 1996, 35, 1709.
(18) Li, J. J.; Li, W.; Sharp, P. R. Inorg. Chem. 1996, 35, 604.
(19) Bu¨cken, K.; Koelle, U.; Pasch, R.; Ganter, B. Organometallics 1996,
15, 3095.
(21) Klein, H.-F.; Bickelhaupt, A.; Lemke, M.; Sun, H.; Brand, A.; Jung,
T.; Ro¨hr, C.; Flo¨rke, U.; Haupt, H.-J. Organometallics 1997, 16, 668.
(22) See the accompanying paper. Holland, P. L.; Smith, M. E.; Andersen,
R. A.; Bergman, R. G. J. Am. Chem. Soc. 1997, 119, 12815.
S0002-7863(97)01829-5 CCC: $14.00 © 1997 American Chemical Society

Monomeric CpNi Methoxo and Amido Complexes
J. Am. Chem. Soc., Vol. 119, No. 52, 1997 12801
Scheme 1
Figure 1. ORTEP diagram of one of the two independent molecules
of 1a in the unit cell, using 50% probability surfaces. Selected distances
and angles for both molecules: Ni-N ) 1.903(5), 1.903(5) Å; Ni-
Cp(centroid) ) 1.763, 1.762 Å; Cp-Ni-P ) 139.97, 142.08°; Cp-
Ni-N ) 126.92, 126.27°; P-Ni-N ) 92.9(2), 91.4(2)°; P-Ni-N-C
(dihedral) ) 74.8(5), 72.3(5)°.
alkyl complex Cp*Ni(PEt3)Me (Scheme 1).22 This methyl
complex was resistant to protonation of the methyl ligand by
anilines or phenols, but the stronger acid TolSH reacted with
Cp*Ni(PEt3)Me to give the thiolate complex Cp*Ni(PEt3)STol
(5), and triflic acid reacted with Cp*Ni(PEt3)Me to give the
red triflate complex Cp*Ni(PEt3)OTf. Although this triflato-
nickel complex has been fully characterized, it was resistant to
isolation in good yield, so crude Cp*Ni(PEt3)OTf was usually
used in subsequent syntheses.
In dichloromethane solution, the triflato complex readily
reacts with Lewis bases such as THF, p-toluidine, or acetonitrile,
as indicated by a change in color from red to yellow. Over a
period of days, Cp*Ni(PEt3)OTf or its base adducts decompose
in THF solution to give the bis(phosphine) complex [Cp*Ni-
(PEt3)2][OTf] as the only isolable product (Scheme 1). When
1 equiv of PEt3 is added to Cp*Ni(PEt3)OTf, the bis(phosphine)
complex can be isolated in 61% yield.23 The X-ray crystal
structure of [Cp*Ni(PEt3)2][OTf] shows that the triflate ligand
is outer sphere.22
delocalization of the nitrogen lone pair into the aromatic ring.
The torsion angles P-Ni-N-C are 72° and 75°, indicating that
the filled nitrogen p orbital of the amido ligand avoids interaction
with the HOMO of the metal fragment, which is perpendicular
to the Ni-P-X plane.26 In a series of analogous complexes
Cp*Ni(PEt3)X (X ) TolNH, TolO, TolS, PhCH2), the geom-
etries about the metal center and X ligand are quite similar,22
suggesting that steric effects, not π interactions, control the
molecular conformation.
The resonances for the aryl protons in the proton NMR
spectrum of 1a were broad at room temperature; the signals
sharpened to two doublets at higher temperatures, and decoa-
lesced to a complicated series of signals at lower temperatures
(Tc ≈ -42 °C; ∆Gq ) 11 kcal/mol).27,28 This phenomenon
was accompanied by decoalescence of the signals corresponding
to the PEt3 methylene protons, which become inequivalent in
the low-temperature regime, indicating that this fluxionality
corresponds to hindered rotation about the Ni-N bond.
Thermal decomposition of 1a over 1 d at room temperature
in benzene-d6 solution, as monitored by 1H NMR spectroscopy,
produces a mixture of TolNH2, [Cp*Ni(µ-NHTol)]2, and the
new nickel(0) tetramethylfulvene complex (C5Me4CH2)Ni(PEt3)2
(2) (Scheme 1). No intermediates are observed in the 31P{1H}
NMR spectra of decomposing solutions of 1a. However, this
decomposition may be envisioned as proceeding through
deprotonation of the Cp* ligand by the amido ligand to produce
TolNH2 and (C5Me4CH2)Ni(PEt3), which could abstract a
phosphine ligand from another molecule of 1a to produce 2
and a second intermediate, Cp*Ni(NHTol), which quickly
dimerizes to [Cp*Ni(µ-NHTol)]2, even at low temperature.7
It has not been possible to isolate 2 cleanly from the mixtures
obtained by decomposition of 1a. Deprotonation of [Cp*Ni-
(PEt3)2][OTf] with LDA, though, can be used to give 2 in an
isolable form (Scheme 1). Complex 2 has a characteristic AB
Synthesis of a Monomeric Nickel Amide. Crude Cp*Ni-
(PEt3)OTf was treated with LiNHTol (Tol ) p-C6H4Me);
workup afforded Cp*Ni(PEt3)NHTol (1a) in typical yields of
50-60% from the methylnickel complex. The product displays
1
a characteristic resonance in its H NMR spectrum at δ -1.5
ppm for the nitrogen-bound proton. The solid-state structure
of 1a was determined by X-ray crystallography; an ORTEP
diagram is shown in Figure 1. There are two crystallographi-
cally independent, but structurally similar, molecules in the
asymmetric unit (bond distances and angles to Ni are within
0.01 Å and 3°, respectively). The nickel-nitrogen distances
are both 1.903(5) Å; this distance is longer than those in the
low-coordinate nickel amides prepared by Power,24,25 but shorter
than that in the only crystallographically characterized square-
planar terminal amidonickel complex.11 Notably, the nitrogen
atoms are planar in 1a; this geometry presumably arises from
(23) Jolly, P. W. η-Cyclopentadienyl Complexes. In ComprehensiVe
Organometallic Chemistry, 1st ed.; Wilkinson, G., Ed.; Pergamon: New
York, 1982; Vol. 6, p 189.
(24) Hope, H.; Olmstead, M. M.; Murray, B. D.; Power, P. P. J. Am.
Chem. Soc. 1985, 107, 712.
(25) Bartlett, R. A.; Chen, H.; Power, P. P. Angew. Chem., Int. Ed. Engl.
1989, 28, 316.
(26) Albright, T. A.; Burdett, J. K.; Whangbo, M.-H. Orbital Interactions
in Chemistry; Wiley: New York, 1985.
(27) Legzdins, P.; Jones, R. H.; Phillips, E. C.; Lee, V. C.; Trotter, J.;
Einstein, F. W. B. Organometallics 1991, 10, 986.
(28) Thomas, W. A. Annu. ReV. NMR Spectrosc. 1968, 1, 43.

12802 J. Am. Chem. Soc., Vol. 119, No. 52, 1997
Holland et al.
Table 1. Equilibrium Constants for Amide/Amine Exchangesa
Scheme 2
Ar ) XC6H4
Keq (est. error)
∆G° (kcal/mol)
p-NMe2 (1b)
p-OMe (1c)
p-Me (1a)
p-F (1d)
p-CF3 (1e)
p-Ac (1f)
0.21 (0.02)
0.50 (0.05)
1
2.2 (0.3)
400 (50)
2 (1) × 103
1.1 (0.2)
+ 0.9 (0.1)
+ 0.4 (0.1)
0
-0.5 (0.1)
-3.5 (0.1)
-4.5 (0.3)
-0.1 (0.1)
m-NMe2 (1g)
a Keq values at room temperature for the reactions in eq 1 were
determined by integration of 1H and 31P{1H} NMR spectra: see
Experimental Section for details.
pattern in its 31P{1H} NMR spectrum, showing that the
phosphorus atoms are inequivalent. A molecule very similar
to 2, (C5Me4CH2)Pd(PMe3)2, has been crystallographically
characterized, and it binds through the exo double bond.29 The
NMR data of 2 also substantiate exo binding: the resonance
1
for the methylene protons in the H NMR spectrum has an
upfield chemical shift of δ 2.05 ppm, and the coupling pattern
is consistent with coupling of these protons to both phosphorus
atoms.
Solutions containing other arylamido complexes could be
generated from 1a by reaction with the appropriately substituted
aniline. The equilibrium constants (Table 1) for the equilibria
1
indicated in eq 1 were measured by H NMR spectroscopy in
Reactions of 1a with Polar Substrates. Complex 1a quickly
reacts with methanol to form an equilibrium mixture of 1a, 3,
methanol, and TolNH2; the equilibrium constants derived from
these reactions were erratic, but always lay between 0.1 and
10.30 However, when isolated 3 was mixed with p-toluidine, a
value of 0.1 was reproducibly derived for Keq as shown in eq
2.
Cp*Ni(PEt3)NHTol + XC6H4NH2 yK
z
eq
C6D6
1a
Cp*Ni(PEt3)NHC6H4X
+ TolNH2 (1)
1b: X ) p-NMe2 1e: X ) p-CF3
1c: X ) p-OMe 1f: X ) p-Ac
1d: X ) p-F
1g: X ) m-NMe2
Keq = 0.1
(2)
Cp*Ni(PEt3)NHTol + MeOH
Cp*Ni(PEt3)OMe + TolNH2
benzene-d6 solution and were independent of the concentrations
of reagents used. The ability to isolate para-substituted aryl-
amidonickel complexes 1b-g from this reaction was often
hampered by small equilibrium constants and/or volatility
differences between the anilines; this could be overcome by
using lithium amides in reactions analogous to that used to
synthesize 1a, or from the methoxide as described below.
Synthesis of a Monomeric Nickel Methoxide. When a
methanolic solution of Cp*Ni(PEt3)Me was treated with triflic
acid, followed by addition of NaOMe, the new methoxide
complex Cp*Ni(PEt3)OMe (3) could be isolated in a yield of
53% (Scheme 2). Complex 3 is even more unstable than 1a,
decomposing to 2 in benzene solution within a few hours at
room temperature. The X-ray crystal structure of 3 is presented
in the accompanying paper.
The reactivity of 3 is similar to that of 1a. As a useful reagent
for HX exchange, methoxide 3 has an advantage over amide
1a in that methanol is more volatile than toluidine. In this way,
formation of the substituted arylamido species 1b-g from 3
was easily pushed to completion by evaporation of methanol.
However, the lability of the methoxide ligand hinders its
isolation in reproducibly high yields. For this reason, we have
devoted more time and effort to the reactions of 1a.
C6D6
3
1a
O
OH
Cp*Ni(PEt3)NHTol +
Cp*Ni(PEt3)(acac) + TolNH2
(3)
1a
Cp*Ni(acac) + PEt3
The reaction of 1a with 2,4-pentanedione, which exists
primarily as its enol tautomer at room temperature,31 also
resulted in proton exchange. The product, the previously
characterized complex Cp*Ni(η2-acac)PEt3, exists in equilibrium
with Cp*Ni(acac) and free PEt3,32 so the products were verified
by comparison of the 1H NMR spectrum of the mixture to those
of the known compounds. Thus, the triethylphosphine ligand
of 1a can be labilized by addition of a chelating alcohol.
More acidic alcohols, when added to 1a, gave complete
conversion to p-toluidine and the corresponding oxygen-bound
nickel complex. For example, addition of 1.00 equiv of p-cresol
gave only p-toluidine and the new red cresolate complex Cp*Ni-
(PEt3)OTol (4), as shown by 1H NMR and 31P{1H} NMR
spectroscopy (eq 4, Scheme 2). Solution calorimetry of this
reaction at 30 °C in benzene solution showed that ∆H303K
)
-9.4 ( 0.2 kcal/mol. The thermodynamics of this reaction
could also be analyzed by spectroscopic techniques. Addition
of a large excess (>500 equiv) of p-toluidine to 4 gave a small
All attempts to synthesize a hydroxo species cleanly have
failed, although a product consistent with this formulation may
be observed by NMR spectroscopy after adding water to a
sample of 1a. However, the Ni-OH group could not be
amount of 1a visible by H and 31P{1H} NMR spectroscopy.
1
1
detected by H NMR spectroscopy, and it was impossible to
(30) Presumably, the problems with methanol addition lie in the difficulty
of completely freeing methanol from trace water.
(31) Hush, N. S.; Livett, M. K.; Peel, J. B.; Willett, G. D. Aust. J. Chem.
1987, 40, 599.
identify the putative hydroxide conclusively.
(29) Werner, H.; Crisp, G. T.; Jolly, P. W.; Kraus, H.-J.; Kru¨ger, C.
Organometallics 1983, 2, 1369.
(32) Smith, M. E.; Andersen, R. A. J. Am. Chem. Soc. 1996, 118, 11119.

Monomeric CpNi Methoxo and Amido Complexes
J. Am. Chem. Soc., Vol. 119, No. 52, 1997 12803
Keq
Scheme 3
Cp*Ni(PEt3)OTol + TolNH2 y
z
C6D6 or THF
4
Cp*Ni(PEt3)NHTol + TolOH (4)
1a
The equilibrium constant for eq 4 was determined by 1H NMR
spectroscopy to be roughly 10-6 in C6D6 and roughly 4 × 10-5
in THF-d8; it was independent of variations in the large amount
of p-toluidine added. Because of the difficulty in integrating
NMR spectra with a large excess of TolNH2, UV/vis spectros-
copy was used to verify this equilibrium constant. This method
confirmed that Keq(C6H6) ≈ 4 × 10-6 and Keq(THF) ) (4 ( 2)
× 10-5; the values in THF solution showed no systematic
variation over a 100-fold range of nickel and toluidine concen-
trations. Although the errors calculated for the equilibrium
constants are a large fraction of the values themselves, ∆G°
(proportional to the logarithm of this number) can be determined
much more accurately than this implies. Thus, spectroscopic
values of ∆G°(C6H6) ) 7.5 ( 0.8 kcal/mol and ∆G°(THF) )
6.0 ( 0.3 kcal/mol can be derived for eq 4.
The red cresolate complex 4 is much more stable than its
amide analogue 1a. In its crystal structure,22 the Ni-ligand
distances are very similar to those in 1a, underlining the isosteric
character of the OTol and NHTol ligands. The Ni-O distance
of 1.889(2) Å is slightly shorter than the Ni-N bond in 1a,
consistent with the difference in covalent radii of N and O. The
Ni-O distance in 4 is very similar to that in 3 (1.877(2) Å),
and so this distance does not change substantially upon
substitution of an aryl for an alkyl substituent on the oxygen
atom. However, the aryl C-C distances indicate delocalization
of negative charge into the ring, with the C(ring)-C(ipso)
distances (av 1.404 Å) longer than the other C(ring)-C(ring)
distances (av 1.388 Å), and the C-O distance of 1.330(3) Å
being almost as short as that calculated for free LiOPh.33 Thus,
the ground-state geometry of 4 indicates that the Ni-O bond
is very polarized.
of the thiolate ligand in 5 with p-CF3-substituted thiophenol
(eq 6) had an equilibrium constant of only 30, indicating that
the thiolate exchanges are less sensitive to substituent effects
than the amide exchanges above (eq 1).
Cp*Ni(PEt3)STol + HSC6H4CF3 yK
z
eq
C6D6
5
Cp*Ni(PEt3)SC6H4CF3 + HSTol (6)
5-F3
Even Brønsted acids as weak as anilines reacted with 1a; as
described above, the equilibrium constants for these reactions
were measured by 1H NMR spectroscopy. Less acidic amines,
such as alkylamines or ammonia, either did not react or caused
decomposition of 1a. In addition, sterically hindered acids, such
as diphenylamine and tert-butyl alcohol, did not react with 1a;
these results are readily understandable in view of the steric
hindrance of the metal center as seen in the crystal structure of
1a.
Electrophilic organometallic reagents exchanged the amide
ligand in 1a for an alkyl ligand. For example, addition of 1
equiv of methyllithium (as an ether solution) to 1a caused
quantitative conversion to Cp*Ni(PEt3)Me (Scheme 3). The
polar Zr-Me bond in Cp2ZrMe2 also reacted to give some of
the methylnickel complex, but substantial decomposition also
occurred. When a solution of 1a was treated with a solution
of 1 equiv of KCH2Ph, the solution turned green and the
complex Cp*Ni(PEt3)CH2Ph could be isolated in 67% yield
(Scheme 3). Its X-ray structure22 shows that the benzyl ligand
is in a typical η1 configuration.35
Silica is another acidic hydroxy compound that could react
with 1a to form Ni-O bonds with loss of p-toluidine. Indeed,
silica turned red after treatment with 1a, indicative of a oxygen-
bound Cp*Ni(PEt3) species. The nickel could be removed from
the silica in the form of a p-nitrophenolate complex by use of
the stronger acid p-nitrophenol. This series of observations is
consistent with the production of a complex of the type Cp*Ni-
(PEt3)(Osilica) (eq 5).34
Sources of chloride and bromide invariably transformed 1a
into Cp*Ni(PEt3)X (X ) Cl, Br). Thus, LiCl and LiBr, when
added in large excess to ether solutions of 1a, gave partial
conversion to the nickel chloride and bromide complexes,22
respectively. The more convenient chloride source Me3SiCl
reacted with 1a to give quantitative conversion to Cp*Ni(PEt3)-
"Cp*Ni(PEt3)(Osilica)" + TolNH2 (5)
Cp*Ni(PEt3)NHTol + HOsilica
1a
HO
NO2
Cp*Ni(PEt3)OC6H4NO2
1
Cl (as judged by H NMR spectroscopy). This product was
p-Thiocresol reacted quantitatively with 1a or 4 to produce
the green-brown nickel thiolate complex Cp*Ni(PEt3)STol (5)
(Scheme 2). This reaction was not reversible, for no achievable
amount of p-cresol added to 5 caused observable resonances of
4 to appear in the 1H or 31P{1H} NMR spectra of the solutions.
Consistent with this observation, solution calorimetry at 30 °C
in benzene showed that reactions between TolSH and 1a and 4
are exothermic by 23.0 ( 0.2 and 14.0 ( 0.1 kcal/mol,
respectively. The crystal structure of 5 has features very similar
to those of the arylamide and cresolate complexes.22 Exchange
isolated in 76% yield and fully characterized (Scheme 3).
Other H-X bonds reacted with the Ni-N bond of 1a to give
a nickel hydride species: the cleanest reaction of this type was
with Me3SiH (Scheme 3). The product, Cp*Ni(PEt3)H (6),
could be obtained in a yield of 97% from 1a. Production of 6
from methoxide 3 also proceeded in quantitative yield (as
determined by NMR spectroscopy). However, in each of these
reactions, it was important to remove residual Me3SiCl from
the commercial Me3SiH, because reaction with chlorosilane to
produce Cp*Ni(PEt3)Cl (see above) is faster than the reaction
(33) Kremer, T.; Schleyer, P. v. R. Organometallics 1997, 16, 737.
(34) Our group has addressed this issue with other transition metal
complexes. See: Meyer, T. Y.; Woerpel, K. A.; Novak, B. M.; Bergman,
R. G. J. Am. Chem. Soc. 1994, 116, 10290.
(35) η3-Benzyl complexes of nickel are known: Ca´mpora, J.; Gutie´rrez,
E.; Poveda, M. L.; Ruiz, C.; Carmona, E. J. Chem. Soc., Dalton Trans.
1992, 1769. Carmona, E.; Paneque, M.; Poveda, M. L. Polyhedron 1989,
8, 285.

12804 J. Am. Chem. Soc., Vol. 119, No. 52, 1997
Holland et al.
Figure 2. The proposed transition state for aryloxide/phenol exchange.
with the silane. Hydride 6 could also be produced from 1a
using 9-BBN (9-borabicyclononane dimer), but this transforma-
tion was not as clean as the silane reaction.
No clean reactions were observed between 1a and less polar
bonds such as the C-C π bonds in acetylenes or ethylene or
the H-H bond in dihydrogen. Substrates with activated C-H
bonds such as MeCN, CH2(CN)2, and acetone were observed
to react with 1a to give products with NMR spectra consistent
with alkylnickel complexes, but these reactions occurred on the
same time scale as decomposition of 1a, and products were not
isolated.
Figure 3. ORTEP diagram of the crystal structure of 4‚HOXyl, using
50% probability surfaces.
Table 2. Selected Bond Distances and Angles in 4‚HOXyla
atoms
distance (Å)
atoms
angle (°)
Ni-Cp
1.755
Cp-Ni-P
140.7
125.4
Hydrogen Bonding to 4. If more than 1 equiv of cresol
was used in the production of cresolate complex 4 from amido
complex 1a, we were not able to remove the excess cresol, even
by crystallization. Unfortunately, we could not successfully
characterize the hydrogen-bonded adduct Cp*Ni(PEt3)OTol‚
HOTol (4‚HOTol, Scheme 2) by X-ray crystallography, but
NMR experiments gave results characteristic of strong hydrogen
Ni-P
2.1696(8)
1.909(2)
2.126(3)
2.199(3)
2.190(3)
2.080(3)
2.153(3)
1.350(3)
1.401(4)
1.393(4)
1.395(4)
1.391(4)
1.393(4)
1.401(4)
1.363(3)
1.399(4)
1.376(4)
1.397(4)
1.383(4)
1.387(4)
1.411(4)
Cp-Ni-O(1)
Ni-O(1)
P-Ni-O(1)
Ni-O(1)-C(17)
96.11(6)
122.9(2)
123.2(2)
119.4(2)
121.0(2)
121.6(3)
117.4(2)
121.6(2)
121.1(2)
117.4(2)
121.4(2)
121.2(2)
123.9(2)
115.4(2)
119.1(3)
121.4(3)
118.8(3)
121.7(3)
118.3(3)
120.6(3)
120.6(3)
120.3(3)
119.5(3)
122.2(2)
Ni-C(1)
Ni-C(2)
O(1)-C(17)-C(18)
O(1)-C(17)-C(22)
C(17)-C(18)-C(19)
C(18)-C(19)-C(20)
C(19)-C(20)-C(21)
C(20)-C(21)-C(22)
C(17)-C(22)-C(21)
C(18)-C(17)-C(22)
C(19)-C(20)-C(23)
C(21)-C(20)-C(23)
O(2)-C(24)-C(25)
O(2)-C(24)-C(29)
C(24)-C(25)-C(26)
C(25)-C(26)-C(27)
C(26)-C(27)-C(28)
C(27)-C(28)-C(29)
C(24)-C(29)-C(28)
C(25)-C(24)-C(29)
C(24)-C(25)-C(30)
C(26)-C(25)-C(30)
C(24)-C(29)-C(31)
C(28)-C(29)-C(31)
Ni-C(3)
Ni-C(4)
Ni-C(5)
O(1)-C(17)
C(17)-C(18)
C(18)-C(19)
C(19)-C(20)
C(20)-C(21)
C(21)-C(22)
C(17)-C(22)
O(2)-C(24)
C(24)-C(25)
C(25)-C(26)
C(26)-C(27)
C(27)-C(28)
C(28)-C(29)
C(24)-C(29)
1
bonding. Thus, the H chemical shift of the resonance for the
hydroxylic proton in cresol added to 4 shifted from 6.11 ppm
(1.5 equiv of cresol) to 5.29 ppm (3 equiv of cresol) upon
addition of additional cresol. This observation is indicative of
rapid exchange between free and hydrogen-bonded cresol.36 The
infrared spectrum of 4‚HOTol has a very broad peak from
3500-2300 cm-1, indicating weakening of the H-O bond of
cresol.
The dynamic behavior of 4‚HOTol was examined using
1
variable-temperature H NMR spectroscopy. When a sample
of 4‚HOTol was heated, the resonances corresponding to the
cresol and cresolate eventually broadened and coalesced at 95
( 5 °C. This corresponds to a fluxional process (∆Gq ) 18.5
( 0.2 kcal/mol)27,28 that renders the cresol and cresolate
equivalent; presumably, the symmetric transition state for this
process resembles Figure 2.37
a Estimated standard deviations are given in parentheses.
In order to characterize the hydrogen bonding in this type of
compound crystallographically, 4 was mixed with 1 equiv of
2,6-dimethylphenol to give Cp*Ni(PEt3)OTol‚HOXyl (4‚
HOXyl). This complex crystallized in space group P21/c, and
refinement proceeded well enough to locate the bridging
hydrogen atom and to refine its position and isotropic thermal
parameter. An ORTEP diagram of the molecular structure is
shown in Figure 3, and important bond distances and angles
are listed in Table 2. The O-O distance is 2.627(3) Å, which
is in a typical range for alkoxide-alcohol hydrogen bonds.38
The O(cresol)-H distance is 0.76(3) Å, the O(phenolate)-H
distance is 1.89(3) Å, and the O-H-O angle is 166(4)°.
Several bond distances in the nickel cresolate portion of the
molecule change from those in 422 in response to the addition
of the hydrogen-bonded phenol. For the most part, these appear
to be steric in origin, resulting from the larger effective size of
the X ligand (X ) OTol‚HOXyl rather than OTol). Thus, the
Ni-(Cp centroid) distance increases from 1.755 to 1.768 Å,
the Ni-P distance increases from 2.1618(6) to 2.1696(8) Å,
and the O-Ni-L (L ) Cp centroid, PEt3) angles expand from
125.4° and 92.99(5)° to 126.5° and 96.11(6)°, respectively. The
slight lengthening of the Ni-O bond from 1.889(2) Å in 4 to
1.909(2) Å in 4‚HOXyl can also be understood in this context.
Interestingly, the C(aryl)-O distance in the cresolate ligand
lengthens from 1.330(3) to 1.350(3) Å, and the C(ring)-
C(ipso)-C(ring) angle expands from 116.9(2)° to 117.4(2)°,
indicating that delocalization of electrons from the electron-
rich oxygen into the aromatic ring of the cresolate decreases
upon coordination of the 2,6-dimethylphenol.33
When an excess of p-cresol was added to a C6D6 solution of
thiocresolate 5 in an attempt to determine Keq, the color of the
solution changed from green-brown to red. It seemed possible
that this behavior could be due to hydrogen bonding, because
thiolate ligands have recently been shown to be weak hydrogen-
bond acceptors.39 However, NMR and UV/vis analysis showed
that there was no new complex in solution; this behavior appears
to be associated with the polarity of the solvent (red in more
(36) Joesten, M. D.; Schaad, L. J. Hydrogen Bonding; Marcel Dekker:
New York, 1974.
(37) Simpson, R. D.; Bergman, R. G. Organometallics 1993, 12, 781.
(38) Cambridge Structural Database: Allen, F. H.; Davies, J. E.; Galloy,
J. J.; Johnson, O.; Kennard, O.; Macrae, C. F.; Mitchell, E. M.; Mitchell,
G. F.; Smith, J. M.; Watson, D. G. J. Chem. Inf. Comput. Sci. 1991, 31,
187.
(39) Kapteijn, G. M.; Grove, D. M.; Smeets, W. J. J.; Kooijman, J.; Spek,
A. L.; van Koten, G. Inorg. Chem. 1996, 35, 534.

Monomeric CpNi Methoxo and Amido Complexes
J. Am. Chem. Soc., Vol. 119, No. 52, 1997 12805
a
Table 3. Half-lives of the Phosphine Exchange Reaction in C6D6
X
T (°C)
approximate t1/2
OTol
NHTol
STol
CH2Ph
Me
25
25
45
75
75
25
<2 min
15 min
6 h
1 d
2 d
H
<5 min
a Half-lives for equilibration of the reaction in eq 7 were based on
1
monitoring the reaction by H and 31P{1H} NMR spectroscopy.
Figure 5. Concentration dependence of the pseudo-first-order rate
constants for the reactions Cp*Ni(PEt3)SAr + PMe3 f Cp*Ni(PMe3)-
SAr + PEt3 at 305 K. Dashed line, diamonds: Ar ) p-Tol. Solid line,
triangles: Ar ) p-C6H5CF3.
(Figure 5). Thus, an electron-withdrawing group in the aryl
ring of 5 does not substantially affect the rate of phosphine
exchange.
Discussion
Syntheses of New (Pentamethylcyclopentadienyl)nickel
Complexes. Generation of a nickel triflato species from Cp*Ni-
(PEt3)Me22 has made it possible to prepare unsupported amido
and methoxo complexes of nickel(II). Each of these classes of
compounds is rare. Very few monomeric amido complexes of
group 10 transition metals were known until recently,2,43,44 but
a flurry of recent activity has centered around palladium
amides.14,45-50 Monomeric, unsupported nickel amides have
been synthesized recently by the Hillhouse10,51 and Boncella11
groups; both groups utilized a square-planar nickel(II) core. The
amido complexes reported here, like those of Boncella, are prone
to ligand loss and dimerization.7 In this case, dimerization to
give [Cp*Ni(µ-NHR)]2 could be avoided by using the trieth-
ylphosphine rather than the trimethylphosphine ligand; the PEt3
adducts were isolable and stable for a reasonable amount of
time in solution at room temperature. Attempts to synthesize
analogous (alkylamido)nickel complexes were unsuccessful,45
since the strong base deprotonated the Cp* ligand to give the
nickel(0) complex 2.
Figure 4. Concentration dependence of the pseudo-first-order rate
constant for the reaction Cp*Ni(PEt3)OTol + PMe3 f Cp*Ni(PMe3)-
OTol + PEt3 at 235 K.
polar solvents; green in nonpolar solvents) as well as the polarity
of the thiolate group itself (e.g., 5 is green in nonpolar solvents;
5-CF3 is red under all conditions). This is reminiscent of color
differences found in CpNi(PR′3)SR complexes described some
time ago.40-42
Exchange of Triethylphosphine with Trimethylphosphine.
The lability of phosphine groups in the various Cp*Ni(PEt3)X
complexes vary drastically with the nature of X. Estimates of
the half-lives of the reaction of Cp*Ni(PEt3)X with PMe3 (∼1
equiv) are given in Table 3. In each case, Keq was estimated to
be between 1 and 5, but in most cases this could not be measured
accurately because of decomposition at long reaction times.
Kinetic studies of this reaction were pursued with X ) OTol
(4) and X ) STol (5) in the presence of large excesses of PMe3,
in order to isolate the forward rate constant k1. Repetition of
Late-metal methoxo complexes are even more rare, because
â-hydride elimination can be a facile pathway for such
compounds.2,43 However, 3 is coordinatively and electronically
saturated, disfavoring this pathway.52,53 Although a few mon-
omeric nickel methoxide complexes have been reported, their
characterization would be considered incomplete by today’s
standards.54-58 Complex 3, on the other hand, has been fully
Cp*Ni(PEt3)X + PMe3 yk1z Cp*Ni(PMe3)X + PEt3 (7)
k-1
Keq ) k1/k-1
kobs ) k1 + k-1
the experiment under such “flooding” conditions showed that
the pseudo-first-order rate constants were linearly dependent
on [PMe3]. Graphs of kobs vs PMe3 concentration are shown in
Figure 4 (X ) OTol) and Figure 5 (X ) STol). These data
can be used to calculate rate constants of k235K ) (2.2 ( 0.1)
× 10-3 M-1 s-1 for 4 and k305K ) (1.37 ( 0.04) × 10-4 M-1
s-1 for 5; each rate law is of the type rate ) k[Cp*Ni(PEt3)X]-
[PMe3].
In order to examine the effect of an electron-withdrawing X
group on the rate of this reaction, Cp*Ni(PEt3)(p-SC6H4CF3)
(5-F3) was synthesized. The rate constant for phosphine
exchange (k305K ) (1.40 ( 0.05) × 10-4 M-1 s-1) is the same
as for its unsubstituted analogue, within experimental error
(43) Fryzuk, M. D.; Montgomery, C. D. Coord. Chem. ReV. 1989, 95,
1.
(44) Cowan, R. L.; Trogler, W. C. J. Am. Chem. Soc. 1989, 111, 4750.
(45) Villanueva, L. A.; Abboud, K. A.; Boncella, J. M. Organometallics
1994, 13, 3921.
(46) Driver, M. S.; Hartwig, J. F. J. Am. Chem. Soc. 1995, 117, 4708.
(47) Louie, J.; Paul, F.; Hartwig, J. F. Organometallics 1996, 15, 2796.
(48) Kim, Y.-J.; Choi, J.-C.; Osakada, K. J. Organomet. Chem. 1995,
491, 97.
(49) Widenhoefer, R. A.; Buchwald, S. L. Organometallics 1996, 15,
2755.
(50) Widenhoefer, R. A.; Buchwald, S. L. Organometallics 1995, 14,
3534.
(51) Matsunaga, P. T.; Hess, C. R.; Hillhouse, G. L. J. Am. Chem. Soc.
1994, 116, 3665.
(52) Hartwig, J. F. J. Am. Chem. Soc. 1996, 118, 7010.
(53) â-Hydrogen elimination can proceed through other pathways in
coordinatively saturated complexes. See: Ritter, J. C. M.; Bergman, R. G.
J. Am. Chem. Soc. 1997, 119, 2580.
(54) Birkenstock, U.; Bo¨nnemann, H.; Bogdanovic, B.; Walter, D.; Wilke,
G. AdV. Chem. Ser. 1968, 70, 250.
(40) Cooke, J.; Green, M.; Stone, F. G. A. J. Chem. Soc., A 1968, 170.
(41) Davidson, J. L.; Sharp, D. W. A. J. Chem. Soc., Dalton Trans. 1972,
107.
(42) Sato, M.; Yoshida, T. J. Organomet. Chem. 1972, 39, 389.

12806 J. Am. Chem. Soc., Vol. 119, No. 52, 1997
Holland et al.
characterized, including X-ray crystallography; this is the first
crystallographically characterized monomeric nickel methoxide
compound.
carboxamides, and (b) the bulky weak acids tert-butyl alcohol
and diphenylamine are also inert to 1a. However, the presence
or absence of ion pairs (either contact or solvent-separated) is
more difficult to establish. It seems reasonable that decomposi-
tion of 1a to 2 is mediated by an ion pair; this is consistent
with the higher solution stability of the phenolate and thio-
phenolate analogues 4 and 5, in which X is less basic.
Another way to rule out the presence of free or solvent-
separated ion pairs is to confirm that the “anion” does not
exchange between metals in solution. Recent kinetic studies
on rhenium alkoxide compounds used a double-labeling experi-
ment to show that alkoxide groups do not exchange between
metal centers.37,71 Literature kinetic and conductance evidence
also argues against the dissociation of late transition metal-N
and -O bonds into free ions in nonpolar solvents.2,4 However,
the polarization of this bond gives rise to substantial reactivity.
The extremely strong hydrogen bond formed between 4 and
phenols, as well as the metrical changes between 4 and
4-HOXyl discussed above, are also a testament to the amount
of electron density localized on the X atom in the ground state
of aryloxonickel complexes.
Nickel hydrido complexes are important intermediates in
organometallic59-61 and biological processes,62 but few such
species have actually been isolated.62-65 The easy isolation and
characterization of 6 thus serves as an example of steric
protection of this functionality. The very well-behaved crystal
structure determination of 6 makes it possible to characterize
the Ni-H bond length as 1.46(3) Å,66 and the large range of
other Cp*Ni(PEt3)X structures allows us to characterize this as
a distance characteristic of a coordination number between 4
and 6.22
Reactivity of an Unsupported Ni-N Bond. The nickel-
nitrogen bond in 1a reacts with polarized Xδ--Yδ+ bonds to
place the more electronegative group on nickel and the less
electronegative group on nitrogen. This is general for elec-
tronegative X in XH (X ) OR, Osilica, NHR, SR), substrates
with electropositive Y in YH (X ) BR2, SiR3), organometallics,
and trimethylsilyl chloride. Less polar bonds (H-H, C-H) do
not react with 1a faster than it decomposes.
Thermochemistry Wia Exchange Reactions. The equilibria
observed between 1 and cresol, methanol, and other arylamides
provide information about relative Ni-O and Ni-N bond
energies in the Cp*Ni(PEt3)X system.
The use of exchange reactions to derive the relative bond
dissociation energies of late metal-ligand bonds has been
employed most extensively in work by Bryndza and Bercaw.72
In this work, the series of equilibria in eqs 8 and 9 were
evaluated. Assuming that replacement of X1 with X2 does not
The decomposition of 1a takes place by deprotonation of
another acid, the coordinated pentamethylcyclopentadiene ligand.
The reducing ligand environment of phosphine and alkene
undoubtedly favors the reduction of Ni(II) to Ni(0), as does the
large effective concentration of an intramolecular C-H bond.
Our group has previously observed decomposition of an amide
complex by Cp* deprotonation;67 interestingly, similar processes
have been used to functionalize Cp* rings in other late transition
metal systems.68,69
The reactivity of 1a indicates that the Ni-N bond is strongly
polarized, with an electrophilic nickel atom and a nucleophilic
nitrogen atom. Although the X-ray crystal structure shows the
presence of a covalent Ni-N bond, it is possible that 1a in
solution is in equilibrium with an ionic tautomer through which
it reacts. Recently, the reactivity of a ruthenium hydroxide has
been rationalized by invoking such a pathway.5
The presence of free ions in either of these reactions seems
unlikely, because the reactions proceed rapidly in nonpolar
solvents such as benzene and pentane, in which formation of
free ions is difficult.70 In the case of 1a, we have also made
the following observations that are inconsistent with the presence
of free amide anions: (a) no reaction is observed with esters or
Keq
Cp*Ru(PMe3)2X1 + HX2 y z Cp*Ru(PMe3)2X2 + HX1 (8)
X ) OH, OMe, CH2C(O)Me, NHPh, NMePh, NPh2
Keq
(DPPE)Pt(Me)X1 + HX2 y z (DPPE)Pt(Me)X2 + HX1 (9)
X ) OH, OMe, CCPh, CH2C(O)Me, NHPh
perturb the bond strengths within [M], X1, or X2,73,74 and
assuming that ∆S for the reaction is zero,75 the values of BDE-
(H-X1), BDE(H-X2), and Keq can be used to derive the
difference between the M-X1 and M-X2 bond strengths. By
finding Keq for a large series of X ligands for which the H-X
bond energies are known, a scale of relative M-X bond energies
can be devised.
(55) Griffith, W. P.; Lewis, J.; Wilkinson, G. J. Chem. Soc. 1959, 1775.
(56) Griffith, W. P.; Lewis, J.; Wilkinson, G. J. Chem. Soc. 1961, 775.
(57) Feltham, R. D. Inorg. Chem. 1964, 3, 116.
(58) Bhaduri, S.; Johnson, B. F. G.; Matheson, T. W. J. Chem. Soc.,
Dalton Trans. 1977, 561.
(59) Emad, A.; Rausch, M. D. J. Organomet. Chem. 1980, 191, 313.
(60) Mu¨ller, U.; Keim, W.; Kru¨ger, C.; Betz, P. Angew. Chem., Int. Ed.
Engl. 1989, 28, 1011.
(71) Unfortunately, we have not been able to isolate pure samples of
ethyltetramethylcyclopentadienyl analogues of the nickel amido complexes
in order to perform crossover experiments here.
(72) Bryndza, H. E.; Fong, L. K.; Paciello, R. A.; Tam, W.; Bercaw, J.
E. J. Am. Chem. Soc. 1987, 109, 1444.
(61) Seligson, A. L.; Cowan, R. L.; Trogler, W. C. Inorg. Chem. 1991,
30, 3371.
(62) James, T. L.; Cai, L.; Muetterties, M. C.; Holm, R. H. Inorg. Chem.
1996, 35, 4148 and references therein.
(63) Srivastava, S. C.; Bigorgne, B. J. Organomet. Chem. 1969, 18, P30.
(64) Saito, T.; Nakajima, M.; Kobayashi, A.; Sasaki, Y. J. Chem. Soc.,
Dalton Trans. 1978, 482.
(65) Darensbourg, M. Y.; Ludwig, M.; Riordan, C. G. Inorg. Chem. 1989,
28, 1630.
(66) Note that bonds to hydrogen derived from X-ray crystallography
are systematically short: Stout, G. H.; Jensen, L. H. X-Ray Structure
Determination: A Practical Guide, 2nd ed.; Wiley: New York, 1989.
(67) Glueck, D. S.; Bergman, R. G. Organometallics 1990, 9, 2862.
(68) Wei, C.; Aigbirhio, F.; Adams, H.; Bailey, N. A.; Hempstead, P.
D.; Maitlis, P. M. Chem. Comm. 1991, 883.
(69) Gloaguen, B.; Astruc, D. J. Am. Chem. Soc. 1990, 112, 4607.
(70) An outer-sphere phenolate has been observed in a ruthenium
complex, but this was aided by extensive hydrogen bonding to excess
phenol. Burn, M. J.; Fickes, M. G.; Hollander, F. J.; Bergman, R. G.
Organometallics 1995, 14, 137.
(73) Benson, S. W.; Buss, J. H. J. Chem. Phys. 1958, 29, 546.
(74) This assumption has been questioned subsequently by Bryndza and
Bercaw: Bryndza, H. E.; Domaille, P. J.; Paciello, R. A.; Bercaw, J. E.
Organometallics 1989, 8, 379. They found that changing X in Cp*Ru-
(PMe3)2X changed the barrier for phosphine exchange by >10 kcal/mol
(since exchange proceeded through a dissociative mechanism, one can
estimate ∆Hqexchange ) BDERu-P), with N and O ligands giving the weakest
Ru-P bond and C and H ligands giving the strongest Ru-P bond. In the
system described here, the Ni-P bond lengths imply a similar trend (see
ref 22). However, their phosphine dissociation was aided by π-bonding in
the unsaturated intermediate; no such mechanism is possible for the
complexes here (see ref 7). Since we have no way to accurately evaluate
the Ni-P bond strengths, this effect is ignored here. Note that this does
not invalidate our estimates of bond energy differences, because bond
energies are defined to include changes in ancillary bonding (see footnote
34 in ref 87).
(75) Since translational entropy is conserved, ∆S should be small in
systems in which two particles are converted into two particles. See: Minas
de Piedade, M. E.; Simo˜es, J. A. M. J. Organomet. Chem. 1996, 518, 167.
Entropic factors in our system will be discussed more thoroughly below.

Monomeric CpNi Methoxo and Amido Complexes
J. Am. Chem. Soc., Vol. 119, No. 52, 1997 12807
Figure 7. Plot of relative Ni-N bond energy versus H-N bond
energy;78 slope ) 1.9. Relative Ni-N bond energies were calculated
as RT ln Keq + BDE (H-N) - 92.0.80
Figure 6. Hammett plot comparing log Keq for the exchange in eq 1
to Hammett σp-; F ) 3.4.
In the study of Bryndza and Bercaw, the equilibrium constants
for the reactions in eqs 8 and 9 were found to be near unity for
exchanges between carbon (CCPh, CH2C(O)Me), nitrogen
(NHPh, NMePh), and oxygen (OH, OMe) ligands.76 Thus,
given the assumptions stated above, the bond dissociation of
the M-X bond scales 1:1 with the bond dissociation of the
H-X bond for these examples, as a graph of H-X bond energy
vs M-X bond energy has a slope of one.77,78 Since the H-X
bond energies for these X groups span 50 kcal/mol, this was
taken as an indication that this correlation is general over that
large range of energies. This correlation implied a model for
bonds between late transition metals and anionic ligands in
which metal fragments bind in the same way as protons, and
thus the late metal auxiliaries “behave like large hydrogens”.2
We will refer to this model as the “1:1 model”, and once again
note its important aspects: (a) a good correlation between D(H-
X) and D(M-X), and (b) these quantities scale 1:1. In the data
presented here, a systematic deviation from the 1:1 model is
observed for the exchange of para-substituted arylamines for
arylamido complexes of nickel. In this study, electron-
withdrawing amines faVored nickel binding, and electron-
donating amines faVored proton binding. The best correlation
with Hammett or Taft parameters was found for σp-, and this
correlation held for para substituents from NMe2 (σp- ) -0.32)
to Ac (σp- ) +0.87).79 The Hammett correlation is shown in
Figure 6; F ) 3.4. Thus, the formation of a nickel amide from
an aniline is controlled by the same effects that control the
deprotonation of phenols (the standard reaction used for defining
σ- parameters). These results cannot be explained in the context
of the 1:1 model: in that model, electronic differences of the
ligand should affect H-X and M-X bonds equally, and the
equilibrium constant should still be unity. This may also be
stated as a relationship between BDE(H-NHAr)80 and the BDE-
(Ni-NHAr) that can be derived from these numbers, the Keq
values, and the entropy and group additivity assumptions
discussed above. A diagram of our data in the style of ref 72
is shown in Figure 7. There is a reasonable correlation between
H-X and M-X bond energies, but instead of a slope of unity,
the slope is 1.9. Although our analysis was necessarily done
over a smaller H-X BDE range, this is as large a range possible
within a series of ligands with such similar substituents.
The exchange between the amide ligand in 1a and the
cresolate ligand in 4 (eq 4) shows another equilibrium constant
that is unexpectedly large; this corresponds to ∆G°THF ) 6.0
( 0.3 kcal/mol and ∆G°benzene ) 7.5 ( 0.8 kcal/mol. We were
interested in determining whether this could be due to a large
entropic effect. In the study of the Pt and Ru systems,72 very
different HX acids were used, and the entropic contribution was
very small (in the absence of steric effects); it would be
surprising if isosteric and electronically similar HX exchanges
like the ones discussed here caused larger entropic changes.
Unfortunately, limitations imposed by signal/noise ratios,
solubility of the amine, and the thermal sensitivity of 1a made
it impossible to accurately gauge ∆S° for this reaction from
the temperature dependence of the equilibrium constant. An
independent assessment of ∆H(303 K) ) -9.4 ( 0.2 kcal/mol
for eq 4 using solution calorimetry in benzene, however,
confirmed that the large equilibrium constant is purely enthalpic.
Thus, both calorimetric methods and equilibrium methods
indicate strongly that the relative bond energies are not in
accordance with the 1:1 model. Importantly, this substantial
deviation was found not for a reaction with drastic differences
between the X ligands, but for one where the steric and
electronic properties of the organic fragment were carefully
chosen to be the same.
(76) The deviation of one example (NPh2) was shown to be largely due
to a steric (entropic) effect. Large deviations were also observed for X
ligands that bind through a second-row atom or can participate in
backbonding.
(77) The 1:1 correlation has sometimes been generalized to BDE(M-N)
< BDE(M-O). This should only follow if BDE(H-N) < BDE(H-O), which,
although true much of the time, is far from universal (e.g., H2N-H ) 107
kcal/mol; PhO-H ) 87 kcal/mol). Thus, such generalizations must be
qualified. However, the present work shows that there is some truth to the
generalization that BDE(M-NHR) < BDE(M-OR), because in this study,
with LnM and R (and, fortuitously, H-X BDE) held constant, the oxygen
ligand binds more strongly than the nitrogen ligand.
(78) In this kind of graph, deviations from 1:1 behavior are dampened
by the fact that both axes contain BDE(H-X). Thus, a systematic deviation
from 1:1 behavior (such as that in ref 81) can appear to give a H-X/M-X
graph with a slope nearly equal to 1.0.
(79) log (Keq) had a poorer correlation with σp, because σp(CF3) >
σp(Ac). Thus, the deprotonation of benzoic acids is a poorer model for Ni-
N/H-N exchange, as are other Hammett substituent constants.
We believe that the inequivalence of the equilibrium and
calorimetric values of ∆H(benzene) is due to hydrogen bonding
between 4 and HOTol in the mixture produced in the equilibra-
tion experiments. When a double equilibrium made up of eq 4
and an equilibrium representing formation of 4‚HOTol was
mathematically modeled, the results confirmed that hydrogen
(80) ArNH-H bond energies in DMSO solution are given in: Bordwell,
F. G.; Zhang, X.-M.; Cheng, J.-P. J. Org. Chem. 1993, 58, 6410. Some of
the BDE values used here were extrapolated from the correlations between
σ- and the N-H bond energies found there.

12808 J. Am. Chem. Soc., Vol. 119, No. 52, 1997
Holland et al.
bonding with a reasonable strength of 5-6 kcal/mol would cause
the apparent equilibrium constant to vary from the actual
equilibrium constant for eq 4 (as deduced from the calori-
metrically determined ∆H value and the assumption that ∆S ≈
0) by the amount observed.
hydrogen bonding acts to make the apparent equilibrium
constant closer to one than the true equilibrium constant. Thus,
a metal fragment for which the 1:1 model was previously
believed to hold is not immune to the thermodynamic preference
for oxygen found here.
The literature also substantiates the preference for isosteric
oxygen over nitrogen ligands, relative to their H-X counter-
parts. For example, phenols displace arylamides and alkoxides
quantitatively in other series of complexes.37,87,88 Some early
metal systems also follow these trends. When aryl ligands are
exchanged in an ingeniously devised chelating scandium system,
the carbon atom of the aryl ring with the para-CF3 substituent
binds preferentially.89 The bond energy orders in early metal
systems studied by Marks show that M-O bonds are dispro-
portionately (relative to a 1:1 correlation) strong with respect
to M-N bonds.90 Also, density functional calculations predict
that (D([M]-OH) - D(H-OH)) - (D([M]-NH2) - D(H-
NH2)) ) 9 kcal/mol for [M] ) (CO)4Co and Cl3Ti.91 Taken
together, these results suggest that some factor causes a
preference for oxygen over nitrogen ligands in metal binding
for both early and late transition metals; this will be addressed
more fully below.
We have shown that the 1:1 model does not hold when only
the heteroatom is changed in isosteric ligands or when the para
substituent in an arylamido ligand is varied. Thus, we sought
a model in which σ or π effects cause transition metals to favor
aryl substituents over alkyl substituents, oxygen over nitrogen,
and electron-withdrawing substituents over electron-donating
ones, relative to hydrogen.
A tempting explanation for the deviation found from the 1:1
H-X/M-X BDE model lies in repulsive nonbonded interactions
between filled d orbitals on the nickel atom and the p orbital(s)
on the X atom.92 Nitrogen is a stronger π donor than oxygen,
both in organic and organometallic93,94 systems, so amido
complexes would be destabilized relative to aryloxo complexes,
as observed. More extensive delocalization of the electrons in
the nitrogen p-orbital into the aryl ring in the electron-
withdrawing amides would explain the Keq values for amido/
amine exchange (eq 1). However, several observations are
inconsistent with this explanation. First, the exchange in eq 1
was repeated using m-(dimethylamino)aniline; for this reaction,
Keq ) 1.1 ( 0.2. The small difference between the equilibrium
constants found with a m-dimethylamino substituent (strongly
electron-withdrawing) and a p-dimethylamino substituent (strongly
electron-donating) suggests that resonance effects do not play
a dominant role. In addition, the geometry of 1a (as shown in
the crystal structure and discussed above) is such that pπ-dπ
repulsion is minimized by steric forces. Finally, the good
Hammett correlation with σp- implies that Ni-NHAr/H-NHAr
exchange is affected by the same factors as Na-OAr/H-OAr
exchange (the standard for σp- parameters), i.e., the interaction
A New Model for Thermochemistry and Reactivity.
Deviations from the 1:1 correlation between BDE(H-X) and
BDE(M-X) are not unique to this nickel auxiliary. A series
of para-substituted phenoxorhenium complexes have been
exchanged with the corresponding phenols by the Mayer group;
as here, electron-withdrawing groups stabilized the M-O bond
relative to the H-O bond.81 Their equilibrium constants
correlated well with Hammett σ (F ) 0.71), although fewer very
electron-deficient substituents were used in comparison with
our study. At that time, the O-H bond energies for these
phenols were not known, but a plot of BDE (H-O) vs BDE
(Re-O) using recently published solution O-H bond energy
data gives a straight line with a slope of 1.07.78,82 Thus, a
systematic deviation from the 1:1 model, albeit a smaller
deviation, is observed with a third-row metal and oxygen rather
than nitrogen ligands. In other late-metal systems, exchanges
of platinum glycolate complexes,83 phenolatoiridium com-
plexes,34 and rhenium amido complexes84 have electronic effects
similar to those observed in this nickel system. However, the
only complexes that have systematically incorporated a large
number of substituents with little steric change, besides the
rhenium alkoxides discussed above, are the arylrhodium com-
plexes Cp*Rh(PMe3)(Ar)(H) (Ar ) 3,5-C6H3X2), a series for
which the corresponding C-H bond energies are not known.85
As a result, a similar plot for these exchanges cannot be
constructed, but it is evident that a distinct preference for
electron-withdrawing X on the metal is present, consistent with
a deviation from the 1:1 model like the one in the nickel system.
The exchange of 1a and methanol with 3 and amine was the
only exchange in which the prediction of the 1:1 model was
upheld; the equilibrium constant of 0.1 for eq 3 is only an order
of magnitude different from the Keq of 2.3 for the analogous
exchange with Cp*Ru(PMe3)2X compounds.72 However, the
ruthenium study did not include the OTol ligand; it included
alkoxide and arylamide ligands, but no examples of aryloxide
or alkylamide complexes for a more systematic comparison. In
the nickel system, 1a or 3 was completely converted to 4 to
within our detection limits when 1 equiv of cresol was added.
This deviation cannot be explained by the large difference in
dissociation energy between the MeO-H bond (104 kcal/mol)
and the ArO-H bond (∼87 kcal/mol)86 because, according to
the 1:1 model, this should be compensated by a much larger
Ni-OMe than Ni-OAr bond strength.
In order to compare the large equilibrium constant for NHTol/
OTol exchange at nickel to the (DPPE)Pt(Me)X system for
which the 1:1 model was demonstrated, we have investigated
an analogous exchange at Pt. Through 1H NMR analyses like
those used for the nickel experiments, we have found that the
equilibrium constant with X1 ) TolO and X2 ) TolNH in eq 9
is approximately 4 × 10-4. As above, this deviation from the
1:1 model cannot be caused by hydrogen bonding, because
(87) Hartwig, J. F.; Andersen, R. A.; Bergman, R. G. Organometallics
1991, 10, 1875.
(88) Glueck, D. S.; Winslow, L. J. N.; Bergman, R. G. Organometallics
1991, 10, 1462.
(89) Bulls, A. R.; Bercaw, J. E.; Manriquez, J. M.; Thompson, M. E.
Polyhedron 1988, 7, 1409.
(90) Schock, L. E.; Marks, T. J. J. Am. Chem. Soc. 1988, 110, 7701. A
reviewer has pointed out that these bond energies are undoubtedly affected
by constructive π-interaction between M and X.
(91) Ziegler, T.; Tschinke, V.; Versluis, L.; Baerends, E. J.; Ravenek,
W. Polyhedron 1988, 7, 1625. Although absolute bond energies derived
from calculations are usually not highly accurate, differences in bond
energies are more trustworthy.
(92) Caulton, K. G. New J. Chem. 1994, 18, 25.
(93) Darensbourg, D. J.; Klausmeyer, K. K.; Reibenspies, J. H. Inorg.
Chem. 1996, 35, 1529; 1535.
(81) Erikson, T. K. G.; Bryan, J. C.; Mayer, J. M. Organometallics 1988,
7, 1930.
(82) ArO-H bond energies in DMSO solution are given in: Bordwell,
F. G.; Cheng, J.-P. J. Am. Chem. Soc. 1991, 113, 1736.
(83) Andrews, M. A.; Voss, E. J.; Gould, G. L.; Klooster, W. T.; Koetzle,
T. F. J. Am. Chem. Soc. 1994, 116, 5730.
(84) Simpson, R. D.; Bergman, R. G. Organometallics 1992, 11, 4306.
(85) Selmeczy, A. D.; Jones, W. D.; Osman, R.; Perutz, R. N.
Organometallics 1995, 14, 5677.
(86) Gas-phase BDEs: McMillen, D. F.; Golden, D. M. Annu. ReV. Phys.
Chem. 1982, 33, 493.
(94) Espinet, P.; Bailey, P. M.; Maitlis, P. M. J. Chem. Soc., Dalton
Trans. 1979, 1542.

Monomeric CpNi Methoxo and Amido Complexes
J. Am. Chem. Soc., Vol. 119, No. 52, 1997 12809
between Ni and N in complexes 1 is mostly σ in character like
the interaction between Na and O in NaOAr. Thus, a pπ-dπ
repulsion model may not be the best explanation for our data.
Instead, we propose that the ground-state energy differences
in these nickel complexes are associated with substantial charge
polarization in the late M-X bond. This polarization simply
corresponds to the significant electronegativity difference
between transition metals and N or O and is not primarily π in
character. A similar model based on electronegativities has been
advanced to explain bond energy correlations in early transition
metal and lanthanide complexes.95 In this qualitative model,
factors that stabilize the partial negative charge on the N or O
atom will stabilize the metal complex. Thus, in the energies
of metal complexes relative to their H-X analogues, the more
electronegative oxygen will be favored over nitrogen, delocal-
izing aryl groups will be favored over alkyl groups, and electron-
withdrawing substituents will be favored.
The fact that metal complexes are more stable with a more
stable anion presumably reflects a large component of electro-
static bonding in metal-X bonds. This is conveniently
described using the E-C model of Drago,96 in which one would
state that the transition metal has a greater inclination to bind
electrostatically than covalently to an anionic ligand. A ∆EX
- ∆CX substituent constant analysis of log (Keq) for the nickel
amido/amine exchanges described above (eq 1) gave sensitivity
factors of dE ) -18.4 and dC ) 1.84 (R2 ) 0.985).97 Although
the linear regression analysis used only the six points for the
para substituents, the substituents varied widely in their ∆CX/
∆EX ratios, so the ratio dC/dE ) -0.1 indicates that electrostatic
factors play a much larger role in determining the bond energy
difference than covalent factors.98 However, as argued above,
this does not necessarily imply that the Ni-N bond is ionic or
that full heterolysis of this bond occurs during the reactions of
complexes 1. The signs of dE and dC also demonstrate that
electrostatically, electron-withdrawing groups favor stronger
Ni-N than H-N bonds, but covalently, electron-donating
groups favor stronger Ni-N than H-N bonds. In this nickel
system, electrostatic effects dominate and so a preference for
electron-withdrawing groups on nickel is observed overall.
Importantly, the electrostatic character of the Ni-N bond
found in this model is consistent with the reactivity of 1a toward
polarized bonds, and the large amount of Ni+ OTol- character
in the crystal structure of 4 discussed above. Thus, a simple
model (based ultimately on the ideas of Pauling) can explain
both the ground-state and reactivity properties of this class of
molecules.
Scheme 4
advanced above replaces the need for the pπ-dπ model, one
might suggest that many of the phenomena ascribed to non-
bonded pπ-dπ repulsions simply reflect the electrostatic
component of M-X bonds.92,99 When polarization effects can
explain the data, “Ockham’s razor” dictates that there is no
reason to add additional variables (here, π-symmetry repulsion)
to the model. Thus, the simple E-C model presented here can
rationalize thermochemical data for saturated complexes with
no need to postulate additional forces. For unsaturated
complexes, though, metal-ligand π interactions clearly affect
the bonding substantially, as shown by many workers.92 Indeed,
in a recent example of systematic alkoxide exchanges at an
unsaturated vanadium center, a stability order opposite to ours
was found.100
Phosphine Exchange Rates Also Reflect Ni-X Polariza-
tion. The triethylphosphine ligand also takes part in exchange
reactions (Scheme 4), and the rates depend on the nature of X.
In the series with isosteric X ligands the rate trend is OTol >
NHTol > STol > CH2Ph; the nickel atoms with the most
electronegative X ligands exchange phosphine ligands the most
rapidly. This order does not correspond to π-donation ability
(NHTol > OTol), a trend that has been observed for exchanges
of phosphine ligands in 18-electron phosphine complexes.74,92
In several cases, the nickel complexes were sufficiently stable
to excess PMe3 that flooding experiments could be used to
determine the overall rate law. With X ) OTol (4) or SAr (5,
5-F3), the reaction had a first-order dependence on both [Cp*Ni-
(PEt3)X] and [PMe3], indicating that the incoming phosphine
is coordinated to the metal in the rate-determining transition
state. Scheme 4 shows a reasonable mechanism that involves
a 20-electron Cp*NiL2X intermediate; this mechanism is
consistent with the second-order rate law and the high rate for
phosphine exchange in the case of the extremely small hydride
ligand. Although this intermediate violates the “18-electron
rule,” an isolable Cp*NiXL2 complex, Cp*Ni(η2-acac)(PEt3),
has recently been crystallographically characterized, demonstrat-
ing that nickel complexes similar to this intermediate are
accessible.32,101
The rates of phosphine substitution are consistent with the
polarized Ni-X model presented above. Thus the nickel
cresolate, which has the most electrophilic nickel atom, reacts
most rapidly, and the nickel alkyl complexes, which have less
electrophilic Ni atoms, react most slowly. In an attempt to
substantiate the electrostatic explanation for the rate ordering,
the rate of the phosphine exchange for the electron-withdrawing
thiolate 5-F3 was measured. The rate constant was equivalent
to that for the p-methyl-substituted thiolate, showing that this
variation did not affect the rate! One can tentatively explain
We do not wish to suggest that destabilizing nonbonding π
interactions between metals and nitrogen or oxygen lone pairs
do not exist. However, the evidence above suggests that this
force is not the major one responsible for the preference of
oxygen over nitrogen ligands, at least in this system. Further,
in view of the success with which the electrostatic model
(95) Marks, T. J.; Gagne, M. R.; Nolan, S. P.; Schock, L. E.; Seyam, A.
M.; Stern, D. Pure Appl. Chem. 1989, 61, 1665.
(96) (a) Drago, R. S. Applications of Electrostatic-CoValent Models in
Chemistry; Surfside: Gainesville, FL, 1994. Drago’s theory only applies
to enthalpy changes; the same zero-entropy assumptions applied above must
be made to apply an E-C analysis to our system. Note tht the analysis
here reflects Drago’s T and R terms, which undoubtedly are incorporated
into our dE because they contribute to ionicity. A more complete analysis
of M-X energies (and a similar criticism of the 1:1 model) in the context
of ECT theory may be found in: (b) Drago, R. S.; Wong, N. M.; Ferris, D.
C. J. Am. Chem. Soc. 1992, 114, 91.
(99) Mayer, J. M. Comments Inorg. Chem. 1988, 8, 125.
(100) Thorn, D. L.; Harlow, R. L.; Herron, N. Inorg. Chem. 1996, 35,
547.
(101) Two referees have questioned the necessity of a 20-electron
intermediate, instead favoring intermediates with loose ionic bonding (i.e.,
[Cp*Ni(PEt3)(PMe3)]+‚‚‚[X]-) or a slipped Cp ring. Although we cannot
rule out either intermediate on the basis of our data, the conclusive
characterization of Cp*Ni(η2-acac)(PEt3) as a three-legged piano stool with
a symmetric, η5-Cp in ref 32 (as well as the more stereotypical example,
nickelocene, which has two symmetric η5-Cp rings) leads us to believe
that a simple associative mechanism is quite reasonable, the 18-electron
“rule” (for which there are many exceptions) notwithstanding.
(97) Drago, R. S. Inorg. Chem. 1995, 34, 3543. This gives a result that
is the two-dimensional analogue of a Hammett correlation; instead of a F
value, one obtains dC (the sensitivity to covalent factors) and dE (the
sensitivity to electrostatic factors).
(98) The dC/dE ratio also accounts for the good fit found with σp
-
(dC/dE ) -0.05) but not with σ (dC/dE ) +0.10).

12810 J. Am. Chem. Soc., Vol. 119, No. 52, 1997
Holland et al.
this result by postulating that the Ni-S bond has much less
electrostatic character than the Ni-N and Ni-O bonds, and is
thus less susceptible to electronic changes in the ligand. This
suspicion was supported by the small para-substituent depen-
dence of the equilibrium constant for protic exchange of thiolate
ligands (eq 6).
phosphine ligands.110,111 This reflects a switch from more
electrostatic bonding in M-X to more covalent bonding in
M-L, and is reflected in the large CB/EB ratios found for
phosphine ligands.112 In the terms of an E-C model, this means
that transition metals also have a substantial ability to bind
covalently when the ligand is not receptive toward electrostatic
bonding. In the E-C analysis of the nickel amido/amine
exchanges above, the covalent terms did favor electron-donating
groups, even though this was obscured by the larger electrostatic
effect. Thus, in a qualitative sense, this verifies that E and C
parameters can be used to understand the two completely
different series of bond strength trends.
It was not feasible to examine the rate laws of the M-X/
H-X′ exchanges (eqs 1 and 4) above because of thermal
instability, detection limits, and rapid rates. However, the
mechanism of the phosphine exchange reactions may be used
to speculate on the mechanism of proton exchange. First, an
incoming HX′ compound could form a transient hydrogen bond
between its proton and the nucleophilic nickel-bound X atom
(as in 4‚HOAr). This could be accompanied by coordination
of X′ to the nickel atom; coordination is well-known to lower
the pKa of R-hydrogen atoms. Our mechanistic study of
phosphine exchange indicates that a coordination site is avail-
able, resulting in a symmetric intermediate, analogous to that
in Figure 2, that can break either H-X bond to either revert to
starting materials or continue to product. This “σ-bond me-
tathesis” mechanism is analogous to those proposed for ex-
changes of alkoxide ligands at rhenium.37,81
The high stability of thiolate complexes can also be explained
in terms of a qualitative E-C model. When aryl thiols were
added to 1a or 4, there was complete, irreversible conversion
to the thiolate complex 5. The values of ∆H for these reactions
C6D6
Cp*Ni(PEt3)OTol + TolSH
8
4
Cp*Ni(PEt3)STol + TolOH (10)
5
Conclusions: Generality of a Qualitative E-C Model.
The 1:1 model is a qualitatively correct model that is useful for
estimating relative M-X bond energies, particularly when the
H-X bond dissociation energies span a wide range. However,
deviations from this generalization exist when H-X bond
energies are varied systematically over a small range by the
use of very similar ligands. Thus, the arguments presented here
indicate that the 1:1 model, in which the proton and the transition
metal bind to X with similar covalent vs electrostatic contribu-
tions, should be amended to incorporate more electrostatic
contribution in metal-N and metal-O bonds than in H-N and
H-O bonds.
The electrostatic modification to the 1:1 model may be general
to all bonds between electronically saturated metals and anionic
ligands, to differing degrees. In a series of exchanges at
rhenium, a similar preference for electron-withdrawing X,
though muted, is observed.81 Because the electronegativity of
the third-row metal is not much lower than that of hydrogen,
the electrostatic bonding tendency does not differ as much
between a proton and the transition metal. In another series of
exchanges, platinum was shown to favor more electron-donating
alkoxides relative to sodium alkoxides.102 Thus, consistent with
our model, when presented with a choice between a more
electrostatically binding metal center, Na(I), and a less elec-
trostatically binding metal, Pt(II), electron-withdrawing groups
favor the former and electron-donating groups the latter. These
qualitative arguments follow the E-C model of Drago, which
can also be used quantitatively.103 The addition of π-symmetry
repulsive effects to the model is not necessary to qualitatively
explain the data in these electronically saturated complexes,
although π effects may eventually be evaluated from deviations
from a quantitative E-C model.96
were found to be -23.0 kcal/mol and -14.0 kcal/mol, respec-
tively. This preference for a second-row main-group element,
consistent with results by Bryndza72 and Glueck113 for the
(DPPE)Pt(Me)X system, has generally been attributed to soft/
soft effects. A similar argument, using a qualitative E-C model,
can rationalize these observations equally well. Bonds from
second-row elements to metals should have a larger contribution
from covalent bonding (CB), independent of the electrostatic
trend.96 In this case, the covalent trend (favoring electron-
donating substituents) should be able to compete more ef-
fectively with the electrostatic trend (favoring electron-
withdrawing substituents); this is indeed observed in the lower
sensitivity of thiol/thiolate equilibria to substituent variation (eq
6). In another result suggesting the separation of E (favoring
more electronegative ligating atoms) and C (favoring second-
row over first-row ligating atoms) variables, the (DPPE)Pt(Me)X
system prefers thiolate over phosphide ligands, and both of these
are preferred over C, N and O ligands.113 Hopefully, future
work will elaborate the range of applicability of this qualitative
E-C model in the chemistry of M-X complexes of the late
transition metals.
Experimental Section
General. All manipulations were carried out in oven-dried glassware
either using standard Schlenk techniques or in a Vacuum Atmospheres
circulating inert atmosphere glove box. Celite and silica were dried
overnight at 200 °C under dynamic vacuum. Pentane, diethyl ether,
hexanes, hexamethyldisiloxane, toluene, C6H6, C6D6, THF, and THF-
d8 were dried over Na/benzophenone. Methylene chloride and CD2-
Cl2 were dried over calcium hydride. Methanol was distilled from
(105) Doherty, N. M.; Bercaw, J. E. J. Am. Chem. Soc. 1985, 107, 2670.
(106) Kurosawa, H.; Ikeda, I. J. Organomet. Chem. 1992, 428, 289.
(107) Kurosawa, H.; Miki, K.; Kasai, N.; Ikeda, I. Organometallics 1991,
10, 1607.
(108) Rix, F. C.; Brookhart, M.; White, P. S. J. Am. Chem. Soc. 1996,
118, 2436.
Although exchanges of M-X and H-X (X ) anionic ligand)
favor electron-withdrawing groups on the metal, exchanges of
neutral ligands (L) favor electron donating groups bound to M,
in metal complexes of arene ligands,104 olefin ligands105-109 and
(109) Burger, B. J.; Santarsiero, B. D.; Trimmer, M. S.; Bercaw, J. E. J.
Am. Chem. Soc. 1988, 110, 3134.
(102) Dockter, D. W.; Fanwick, P. E.; Kubiak, C. P. J. Am. Chem. Soc.
1996, 118, 4846.
(103) See ref 96. We hope that others will carry out quantitative E-C
analyses that test the generality of the notions presented here. Such studies
should be performed on more robust systems than these nickel complexes,
so that entropy changes can be evaluated, and so that decomposition does
not limit the choice of ligands.
(110) Luo, L.; Li, C. B.; Cucullu, M. E.; Nolan, S. P. Organometallics
1995, 14, 1333.
(111) The opposite order can be observed when back-bonding is
important, as in pyridines: Kulig, J.; Lenarcik, B.; Rzepka, M. Pol. J. Chem.
1987, 61, 735. However, more recently, it was found that electron-donating
substituents cause stronger pyridine binding in B12 models: Garr, C. D.;
Sirovatka, J. M.; Finke, R. G. Inorg. Chem. 1996, 35, 5912.
(112) Drago, R. S.; Joerg, S. J. Am. Chem. Soc. 1996, 118, 2654.
(113) Glueck, D. S., Nolan, S. P. Unpublished information.
(104) Nolan, S. P.; Martin, K. L.; Stevens, E. D.; Fagan, P. J.
Organometallics 1992, 11, 3947.

Monomeric CpNi Methoxo and Amido Complexes
J. Am. Chem. Soc., Vol. 119, No. 52, 1997 12811
NaBH4 and stored over activated 4Å molecular sieves. p-Toluidine,
p-anisidine, and p-H2NC6H4NMe2 were sublimed prior to use. m-H2-
NC6H4F and p-H2NC6H4CF3 were dried over activated 4 Å sieves.
p-Cresol and p-H2NC6H4C(O)CH3 were dried as benzene solutions over
activated 4 Å molecular sieves, then filtered, and freeze-dried. m-H2-
NC6H4NMe2 was isolated from its dihydrochloride by extracting with
diethyl ether and aqueous KOH, drying over MgSO4, and distilling
from CaCl2. Commercial Me3SiH (Petrarch) as stored over NaOH or
KOH to remove Me3SiCl, which was present to the extent of 5-10%,
Celite (1 cm), reduced in volume to 25 mL, and slowly cooled to -80
°C. This gave large blue crystals of Cp*Ni(PEt3)NHTol (1.15 g, 55%
yield). 1H NMR (C6D6): δ 6.95 (d, 2, J ) 8 Hz, Tol aryl), 6.78 (d, 2,
J ) 8 Hz, Tol aryl), 2.32 (s, 3, Tol methyl), 1.54 (d, 15, J ) 1.4 Hz,
Cp*), 1.10 (m, 6, P(CH2CH3)3), 0.87 (m, 9, P(CH2CH3)3), -1.48 (s, 1,
NHTol). 13C{1H} NMR (C6D6): δ 157.9 (Tol quaternary), 129.5 (Tol
methine), 118.0 (Tol methine), 117.6 (Tol quaternary), 100.5 (C5Me5),
20.8 (Tol methyl), 14.2 (d, J ) 20 Hz, P(CH2CH3)3), 10.4 (P(CH2CH3)3),
7.7 (C5Me5). 31P{1H} NMR (C6D6): δ 25.1. IR (Nujol): νNH ) 3333-
(vw), 3323(vw). λmax ) 610 nm (ꢀ (THF) ) 1600 L/mol cm). Anal.
Calcd: C, 66.05; H, 9.16; N, 3.35. Found: C, 66.01; H, 9.13; N, 3.17.
Complexes 1b, 1c, 1g. These compounds were not isolated; they
were identified by their 1H and 31P NMR spectra in equilibration
experiments. Cp*Ni(PEt3)NH(p-C6H4NMe2) (1b). 1H NMR (C6D6):
δ 2.69 (s, 3, NMe2), 1.57 (d, 15, J) 1.4 Hz, Cp*), -1.67 (s, 1, NH).
31P{1H} NMR (C6D6): δ 25.1. Cp*Ni(PEt3)NH(p-C6H4OMe) (1c). 1H
NMR (C6D6): δ 3.54 (s, 3, NMe2), 1.55 (overlaps w/SM Cp*, Cp*),
-1.67 (s, 1, NH). 31P{1H} NMR (C6D6): δ 25.2. Cp*Ni(PEt3)NH(m-
C6H4NMe2) (1g). 1H NMR (C6D6): δ 2.87 (s, 3, NMe2), 1.56 (d, 15,
J) 1.4 Hz, Cp*), -1.30 (s, 1, NH). 31P{1H} NMR (C6D6): δ 25.2.
Cp*Ni(PEt3)NH(p-C6H4F) (1d). This amide was synthesized by a
method analogous to that used for Cp*Ni(PEt3)NHTol, giving a blue
microcrystalline solid. 1H NMR (C6D6): δ 6.87 (vt, 2, J apparent ) 9
Hz, aryl), 6.59 (m, 2, aryl), 1.48 (d, 15, J ) 1.4 Hz, Cp*), 1.06 (m, 6,
P(CH2CH3)3), 0.83 (m, 9, P(CH2CH3)3), -1.54 (s, 1, NHTol). 13C-
{1H} NMR (C6D6): δ 156.8 (ipso quaternary), 152.5 (d, JCF ) 240
Hz, para quaternary), 117.1 (aryl methine), 115.0 (d, JCF ) 20 Hz,
aryl methine), 100.6 (C5Me5), 14.2 (d, J ) 20 Hz, P(CH2CH3)3), 10.3
(P(CH2CH3)3), 7.6 (C5Me5). 31P{1H} NMR (C6D6): δ 25.2. 19F{1H}
NMR (C6D6): δ -138.1. Anal. Calcd: C, 62.59; H, 8.36; N, 3.32.
Found: C, 62.62; H, 8.45; N, 3.28.
1
as shown by H NMR spectroscopy. Lithium amides were prepared
by adding a slight deficiency of n-butyllithium or methyllithium (in
hexanes or diethyl ether, respectively) to a hexanes or toluene solution
of the appropriate amine, washing with hexanes or toluene, and
removing volatile materials in Vacuo. Benzylpotassium was prepared
by a known method.114 Cp*Ni(acac)115 was synthesized according to
published procedures,32 as was Cp*Ni(PEt3)Me.22 All other starting
materials were of reagent grade and were purified according to standard
procedures, as necessary.116 1H, 13C{1H}, and 31P{1H} NMR spectra
were recorded on a Bruker AMX-400 spectrometer (100.6 MHz for
carbon, 162.1 MHz for phosphorus). Chemical shifts are reported in
parts per million (δ), coupling constants are reported in Hertz (Hz)
and integrations are reported in number of protons. Unless otherwise
indicated, 13C{1H} NMR resonances are singlets. Chemical shifts are
referenced internally to the C6D6 solvent peaks at δ 7.15 (1H) and 128.0
(13C) ppm, the THF-d8 solvent peaks at δ 3.58 (1H) and 67.4 (13C)
ppm, or the CD2Cl2 solvent peaks at δ 5.31 (1H) and 53.8 (13C) ppm.
In some cases, DEPT was used to assign the 13C{1H} NMR resonances
as CH3, CH2, CH, or quaternary. IR spectra were recorded on a Mattson
Galaxy Series FTIR 3000 spectrophotometer, as Nujol mulls on NaCl
plates, as pressed KBr pellets, or as solutions between NaCl plates in
a solution cell, and the frequencies are reported in cm-1. UV/vis spectra
were recorded at 298 K on solutions in quartz cuvettes (sealed to Kontes
vacuum adapters), using a HP 8450A UV/vis spectrophotometer
equipped with a temperature-control unit. Elemental analyses were
performed at the UCB Microanalysis Facility; in the products of
reactions where 1a was a starting material, nitrogen was not detected
(<0.1%).
Cp*Ni(PEt3)OTf. In the glovebox, a solution of trifluoromethane-
sulfonic (triflic) acid (0.13 mL, 1.5 mmol) in CH2Cl2 (7 mL) was added
dropwise to a stirred solution of Cp*Ni(PEt3)Me (469 mg, 1.43 mmol)
in CH2Cl2 (15 mL), causing a change in color from green to burgundy.
After 30 min of stirring, the volatile materials were removed in Vacuo.
The residue was extracted into diethyl ether (20 mL) and filtered to
give a burgundy solution. The volume of the solution was reduced to
14 mL, and the solution was cooled to -30 °C, giving clumps of red
crystals of Cp*Ni(PEt3)OTf (261 mg, 39% yield). Collection of further
crops of crystals was generally unsuccessful. 1H NMR (CD2Cl2): δ
1.50 (m, 6, P(CH2CH3)3), 1.41 (s, Cp*), 1.21 (m, 9, P(CH2CH3)3).
13C{1H} NMR (CD2Cl2): δ 105.1 (C5Me5), 14.2 (d, J ) 30 Hz, P(CH2-
CH3)3), 10.2 (P(CH2CH3)3), 7.9 (C5Me5). 31P{1H} NMR (CD2Cl2): δ
23.0. 19F{1H} NMR (CD2Cl2): δ -76.7. IR (Nujol): 2750 (w), 1535
(w), (1460-1370 obscured by Nujol), 1346 (m), 1288 (s), 1259 (s),
1223 (s), 1160 (s), 1106 (w), 1027 (s), 944 (w), 765 (m). Anal.
Calcd: C, 44.28; H, 6.56. Found: C, 44.52; H, 6.64.
Cp*Ni(PEt3)NH(p-C6H4CF3) (1e). A solution of p-trifluorometh-
ylaniline (26 mg, 0.16 mmol) in diethyl ether (5 mL) was mixed with
a solution of 3 (40.4 mg, 0.118 mmol) in diethyl ether (5 mL). The
solution turned from orange-red to purple over ∼30 s. After the mixture
was stirred for 10 min, the volatile materials were removed in Vacuo
to give a quantitative yield of Cp*Ni(PEt3)NH(p-C6H4CF3). This
material was crystallized from pentane before use to give spectroscopi-
1
cally pure material. In the H NMR spectrum of this complex, the
signals for the aryl group are broad due to fluxionality at room
temperature; these peaks coalesced upon raising and decoalesced upon
lowering the temperature. 1H NMR (room temperature, C6D6): δ 7.5
(br, 2, aryl), 6.8 (br, 1, aryl), 6.3 (br, 1, aryl), 1.39 (d, 15, J ) 1.5 Hz,
Cp*), 0.98 (m, 6, P(CH2CH3)3), 0.78 (m, 9, P(CH2CH3)3), -0.52 (s, 1,
NH). 13C{1H} NMR (C6D6, 345 K): δ 163.4 (aryl), 126.1 (aryl), amide
ligand quaternary carbons not observed, 101.3 (C5Me5), 14.8 (d, J )
20 Hz, P(CH2CH3)3), 10.2 (P(CH2CH3)3), 7.5 (C5Me5). 31P{1H} NMR
(C6D6): δ 25.6. 19F{1H} NMR (C6D6): δ -58.4.
Cp*Ni(PEt3)NH(m-C6H4C(O)CH3) (1f). A mixture of p-aminoac-
etophenone (8.5 mg, 0.063 mmol) and 3 (16.2 mg, 0.047 mmol) was
dissolved in diethyl ether (10 mL). The color of the solution changed
from red to purple over 1 min. Volatile materials were removed in
Vacuo, leaving a purple residue that was extracted with 1:1 pentane/
ether (10 mL) and filtered. The solution was concentrated to 2 mL;
pentane (5 mL) was added, and the solution was cooled to -30 °C,
giving a purple microcrystalline solid. This solid was spectroscopically
pure except for 0.1 equiv of excess p-aminoacetophenone (which is
apparently not volatile enough for removal). Since this amine would
be a component of equilibrium mixtures formed from the addition of
TolNH2 to 1f, this was simply accounted for in Keq calculations. 1H
NMR (C6D6): δ 8.14 (br s, aryl), 7.75 (br s, aryl), 6.84 (br s, aryl),
6.26 (br s, aryl), 2.39 (s, 3, C(O)Me), 1.38 (d, 15, J ) 1.5 Hz, Cp*),
0.03 (s, 1, NH). 31P{1H} NMR (C6D6): δ 25.7.
Cp*Ni(PEt3)NHTol (1a). To a stirred solution of Cp*Ni(PEt3)Me
(1.64 g, 5.01 mmol) in CH2Cl2 (50 mL) was added a solution of triflic
acid (0.52 mL, 5.89 mmol) in CH2Cl2 (25 mL). This caused a color
change from pea-green to burgundy, as well as moderate effervescence.
After the mixture was stirred for 15 min at ambient temperature, the
volatile materials were removed in Vacuo over several hours. The crude
triflato complex was dissolved in THF (50 mL), and a solution of
LiNHTol (692 mg, 6.12 mmol) in THF (30 mL) was added at 0 °C,
giving an immediate color change to dark green. After the mixture
was stirred for 10 min at 0 °C, the volatile materials were completely
removed in Vacuo at room temperature. The residue was extracted
into pentane (4 × 20 mL); the resulting solution was filtered through
[Cp*Ni(PEt3)2][OTf]. A red solution of Cp*Ni(PEt3)OTf was
prepared by adding a solution of triflic acid (103 mg, 0.686 mmol) in
dichloromethane (20 mL) to a green solution of Cp*Ni(PEt3)Me (210
mg, 0.642 mmol) in dichloromethane (50 mL) at -78 °C. Trieth-
ylphosphine (0.10 mL, 0.68 mmol) was added to this solution via
syringe. The solution was warmed to room temperature, and the
solution turned orange over 12 h. After this time, all volatile materials
were removed in Vacuo. The brown residue was extracted with THF
(10 mL) and filtered. The orange filtrate was concentrated to 7 mL, 4
(114) Schlosser, M.; Hartmann, J. Angew. Chem., Int. Ed. Engl. 1973,
12, 508.
(115) Bunel, E. E.; Valle, L.; Manriquez, J. M. Organometallics 1985,
4, 1680.
(116) Perrin, D. D.; Armarego, W. L. F. Purification of Laboratory
Chemicals, 3rd ed.; Pergamon: New York, 1988.

12812 J. Am. Chem. Soc., Vol. 119, No. 52, 1997
Holland et al.
mL of diethyl ether was added, and the solution was cooled to -30
°C. Two crops of brown crystals were obtained in this way in a total
yield of 226 mg (61%). 1H NMR (CD2Cl2): δ 1.68 (t, 15, J ) 1.6
Hz, Cp*), 1.63 (m, 6, P(CH2CH3)3), 1.09 (m, 9, P(CH2CH3)3). 13C-
{1H} NMR (THF-d8): δ 106.6 (C5Me5), 13.6 (t, J ) 14 Hz, P(CH2-
CH3)3), 11.4 (P(CH2CH3)3), 8.3 (C5Me5). 31P{1H} NMR (CD2Cl2): δ
23.75. 19F NMR (THF-d8): δ -75.3. IR (Nujol): 2288 (w), 1455
(s), 1378 (s), 1351 (m), 1266 (s), 1224 (sh), 1146 (s), 1136 (m), 1030
(s), 944 (m), 757 (m), 714 (m), 705 (m), 673 (w), 636 (m), 615 (w).
Anal. Calcd. for C23H45F3NiO3P2S: C, 47.69; H, 7.83. Found: C,
47.91; H, 7.87.
(η2-C5Me4CH2)Ni(PEt3)2 (2). A solution of lithium diisopropyl-
amide (38 mg, 0.35 mmol) in THF (5 mL) was added to a solution of
[Cp*Ni(PEt3)2][OTf] (179 mg, 0.309 mmol) in THF (15 mL), causing
a slight color change from yellow-brown to yellow-orange. After the
mixture was stirred for 15 min, the volatile materials were removed in
Vacuo. The orange residue was extracted into pentane (1 mL) and
filtered. Crystallization from this solvent did not occur, so the solution
was extracted into (Me3Si)2O (1.5 mL) and cooled to -35 °C, giving
2 as an orange solid (64 mg, 48% yield). The solubility of this
compound was similar to that of contaminants, so this solid was of
∼90% purity, sufficient for spectroscopic characterization. 1H NMR
(C6D6): δ 2.13 (s, 6, fulvene Me), 2.08 (d, 6, J ) 3.0 Hz, fulvene
Me), 2.05 (dd, 2, J ) 5.9, 1 Hz, exo CH2), 1.35 (m, 6, P(CH2CH3)3),
1.09 (m, 6, P(CH2CH3)3), 0.93 (m, 9, P(CH2CH3)3), 0.84 (m, 9,
P(CH2CH3)3). 13C{1H} NMR (C6D6, for assignments, compare to the
Pd analogue29): δ 128.5 (dd, J ) 4, 2 Hz), 126.2 (d, J ) 3 Hz), 83.0
(d, J ) 10 Hz), 31.5 (d, J ) 17 Hz, methylene), 19.1 (dd, J ) 17, 3
Hz, methylene), 15.3 (dd, J ) 17, 2 Hz, methylene), 12.5 (s), 12.0 (s),
8.8 (s), 8.3 (s). 31P{1H} NMR (C6D6): δ 21.6 (d, J ) 40 Hz), 14.1 (d,
J ) 40 Hz). IR (Nujol): 2731 (w), 1523 (m), 1375 (s), 1336 (s), 1252
(m), 1178 (m), 1149 (w), 1034 (s), 997 (m), 943 (m), 885 (m), 760 (s),
710 (s), 663 (m), 617 (m). High-resolution EI-MS (found/calcd for
NiP2C22H44): 431.226 643/431.225 972, 430.223 262/430.222 618,
429.231 032/429.230 530, 428.227 093/428.227 176.
pure THF; however, the silica did not return to white. The combined
washings were dried in vacuo to give an orange solid (25.8 mg, 85:15
1
ratio of nickel phenolate to phenol by H NMR spectroscopy; 90%
yield of nickel phenolate). For Cp*Ni(PEt3)OC6H4NO2: 1H NMR
(C6D6): δ 8.24 (d, 2, J ) 7.2 Hz, aryl), 6.81 (d, 2, J ) 7.2 Hz, aryl),
1.21 (d, 15, J ) 1.5 Hz, Cp*), 0.93 (m, 6, P(CH2CH3)3), 0.82 (m, 9,
P(CH2CH3)3). 31P{1H} NMR (C6D6): δ 21.9.
Cp*Ni(PEt3)OTol (4). Complex 1a (355 mg, 0.850 mmol) and
p-cresol (92.1 mg, 0.852 mmol) were placed in a flask. Diethyl ether
(10 mL) was added, and the solution rapidly changed color from blue-
purple to deep red. After the mixture was stirred for 15 min at ambient
temperature, the volatile materials were removed in Vacuo and the
residue was exposed to high vacuum (∼1 mTorr) overnight. The red
residue was extracted into pentane (12 mL), filtered, reduced in volume
to 8 mL, and cooled to -30 °C. This gave red crystals of Cp*Ni-
(PEt3)OTol (299 mg, 84% yield). 1H NMR (C6D6): δ 7.17 (d, 2, J )
8 Hz, Tol aryl), 7.04 (d, 2, J ) 8 Hz, Tol aryl), 2.31(s, 3, Tol methyl),
1.43 (s, 15, J ) 1.4 Hz, Cp*), 1.13 (m, 6, P(CH2CH3)3), 0.96 (m, 9,
P(CH2CH3)3). 13C{1H} NMR (C6D6): δ 166.5 (Tol quaternary), 129.4
(Tol methine), 121.8 (Tol methine), 120.6 (Tol quaternary), 101.5 (C5-
Me5), 20.7 (Tol methyl), 14.0 (d, J ) 20 Hz, P(CH2CH3)3), 10.3
(P(CH2CH3)3), 7.6 (C5Me5). 31P{1H} NMR (C6D6): δ 21.8. λmax: 540
nm (ꢀ (THF) ) 725 L/mol cm). Anal. Calcd: C, 65.90; H, 8.90.
Found: C, 65.93; H, 9.04.
Cp*Ni(PEt3)STol (5). A mixture of 1a (26.0 mg, 0.0622 mmol)
and p-thiocresol (8.7 mg, 0.070 mmol, 1.1 equiv) was dissolved in
diethyl ether (5 mL). The color of the solution rapidly changed from
blue to brown. This solution was exposed to vacuum overnight to
remove solvent, p-toluidine, and excess p-thiocresol. The residue was
extracted with pentane (5 mL) and filtered. The green-brown filtrate
was concentrated to 1 mL and cooled to -30 °C, giving green blocks
of 5. Collection of an additional crop of crystals from hexamethyldi-
siloxane (1 mL) gave a total yield of 22 mg (81%). 1H NMR (C6D6):
δ 7.84 (d, 2, J ) 8 Hz, Tol aryl), 6.90 (d, 2, J ) 8 Hz, Tol aryl),
2.17(s, 3, Tol methyl), 1.62 (d, 15, J ) 1.3 Hz, Cp*), 1.32 (m, 6, P(CH2-
CH3)3), 0.84 (m, 9, P(CH2CH3)3). 13C{1H} NMR (C6D6): δ 144.5 (Tol
quaternary), 132.6 (Tol methine), 129.7 (Tol quaternary), 128.3 (Tol
methine), 101.5 (C5Me5), 20.9 (Tol methyl), 15.1 (d, J ) 20 Hz, P(CH2-
CH3)3), 10.6 (P(CH2CH3)3), 7.9 (C5Me5). 31P{1H} NMR (C6D6): δ
25.0. IR (Nujol): 1893 (w), 1880 (w), 1713 (w), 1639 (w), 1596 (s),
1349 (s), 1246 (s), 1157 (s), 1083 (s), 1029 (s), 1001 (sh), 944 (m),
833 (w), 804 (s), 762 (s), 715 (s), 671 (m), 629 (m). Anal. Calcd: C,
63.46; H, 8.57. Found: C, 63.38; H, 8.67.
Cp*Ni(PEt3)SC6H4CF3 (5-F3). To a solution of Cp*Ni(PEt3)Me
(31.2 mg, 0.095 mmol) in diethyl ether (2 mL) was added a solution
of p-(trifluoromethyl)thiophenol (21.7 mg, 0.122 mmol) in diethyl ether
(1 mL). Over a few min, the solution turned from green to rust-orange.
The solution was heated to 45 °C for 33 h, after which time it was
dark red. The volatile materials were removed in Vacuo, and the brown-
red residue was warmed to 60 °C under vacuum to remove excess thiol.
Finally, the residue was extracted with pentane (4 mL), filtered,
concentrated to 1 mL, and cooled to -30 °C, giving red-brown crystals
(23.3 mg, 50%). The high solubility of the product hampered attempts
to isolate further crystals. 1H NMR (C6D6): δ 7.82 (d, 2, J ) 8 Hz,
Tol aryl), 7.28 (d, 2, J ) 8 Hz, Tol aryl), 1.51 (s, 15, Cp*), 1.21 (m,
6, P(CH2CH3)3), 0.75 (m, 9, P(CH2CH3)3). 31P{1H} NMR (C6D6): δ
25.1. 19F NMR (C6D6): δ -60.96. Anal. Calcd: C, 56.46; H, 7.00.
Found: C, 56.35; H, 7.13.
Cp*Ni(PEt3)H (6). Complex 1a (520 mg, 1.24 mmol) was dissolved
in benzene (15 mL) and placed in a 50 mL reaction tube. The solution
was frozen and the headspace was evacuated. Trimethylsilane (505.7
mL at 666 Torr, 18 mmol) was condensed into the tube. The tube
was then closed and warmed to room temperature for 1 d; over this
time, the color changed slowly from blue to brown. The volatile
materials were removed in Vacuo, and the tube was placed under
dynamic vacuum (∼10-4 Torr) overnight. The brown residue was
extracted with pentane (10 mL), filtered, concentrated to 1 mL in Vacuo,
and cooled to -30 °C. Two crops of brown crystals were collected
for a total yield of 377 mg (97%). 1H NMR (C6D6): δ 2.06 (s, 15, no
coupling resolved, Cp*), 1.16 (m, 6, P(CH2CH3)3), 0.88 (m, 9,
P(CH2CH3)3), -21.4 (d, 1, J ) 105 Hz, Ni-H). 13C{1H} NMR
(C6D6): δ 96.8 (C5Me5), 20.5 (d, J ) 30 Hz, P(CH2CH3)3), 11.8
Cp*Ni(PEt3)OMe (3). A solution of triflic acid in methanol (3.7
mL, 0.35 M) was added to a stirred solution of Cp*Ni(PEt3)Me (197
mg, 0.602 mmol) in methanol (20 mL). This caused the color to change
from green to burgundy. A solution of NaOMe, prepared by adding
MeOH (10 mL) to elemental sodium (42.4 mg), was added to the
burgundy solution of nickel triflate. The volatile materials were
removed from the resulting red solution in Vacuo. The red-brown
residue was extracted with 20 mL of pentane, filtered through a pad of
Celite, and concentrated to 5 mL in vacuo. Hexamethyldisiloxane (1
mL) was added, and the solution was filtered again and cooled to -30
°C. Three crops of red-orange crystals were collected in a total yield
of 109 mg (53%). 1H NMR (C6D6): δ 3.32 (s, 3, OCH3), 1.61 (d, 15,
J ) 1.3 Hz, Cp*), 1.28 (m, 6, P(CH2CH3)3), 1.05 (m, 9, P(CH2CH3)3).
13C{1H} NMR (C6D6): δ 99.9 (C5Me5), 57.9 (OMe), 13.6 (d, J ) 20
Hz, P(CH2CH3)3), 10.3 (P(CH2CH3)3), 8.0 (C5Me5). 31P{1H} NMR
(C6D6): δ 25.0. IR (C6D6): 2749 (m), 1065 (s), 1034 (m), 764 (m),
720 (m). Anal. Calcd: C, 59.51; H, 9.69. Found: C, 59.59; H, 9.70.
Generation of “Cp*Ni(PEt3)OH”. A solution of 1 (2 mg, 0.005
mmol) in C6D6 (0.6 mL) was placed in an NMR tube that was capped
with a septum. Degassed water (1.0 µL, 0.056 mmol) was added via
syringe, and the mixture was shaken vigorously for 3 min. The solution
1
turned to an orange-yellow color; H NMR spectroscopy showed one
new product, along with 1 equiv of TolNH2, excess water (δ 5.3 ppm
(br)), and a small amount (∼5%) of the starting amido complex. For
“Cp*Ni(PEt3)OH”: 1H NMR (C6D6): δ 1.41 (s, 15, Cp*), 1.24 (m, 6,
P(CH2CH3)3), 0.99 (m, 9, P(CH2CH3)3), OH not detected. 31P{1H}
NMR (C6D6): δ 24.3.
Adsorption and Removal of [Cp*Ni(PEt3)] on Silica. A solution
of 1a (25.4 mg, 0.0607 mmol) in diethyl ether (4 mL) was passed
through a column of dried silica in a disposable pipet (5 cm), causing
the silica to turn red. The column was rinsed with diethyl ether (3
mL) and THF (3 mL). The eluent was dried in Vacuo, leaving 7.0 mg
of a tan residue (>90% toluidine by 1H NMR spectroscopy; theoretical
yield ) 6.5 mg; >95% yield of TolNH2). A solution of p-nitrophenol
(9.2 mg, 1.1 equiv) in THF (1 mL) was passed through the column of
red silica three times; after this treatment the solution was orange-red,
and the silica remained tan. The column was rinsed with 2 mL of

Monomeric CpNi Methoxo and Amido Complexes
J. Am. Chem. Soc., Vol. 119, No. 52, 1997 12813
(P(CH2CH3)3), 8.4 (C5Me5). 31P NMR{1H} (C6D6): δ 38.0 (s). IR
(C6D6): 1894 (vs), 1468 (m), 1414 (m), 1375 (m), 1249 (m), 1034 (s),
766 (s), 722 (s), 627 (m). Anal. Calcd: C, 61.38; H, 9.98. Found:
C, 61.26; H, 10.05.
several hours. Although this equilibration rate was typical for most
substituted anilines, the equilibration time for the aryloxide was only
a few minutes, while the equilibration took several days in the case of
Cp*Ni(PEt3)NH(p-C6H4CF3) + TolNH2. The ratios of the amines and
1
amides were determined by integration of one-pulse H NMR spectra
Cp*Ni(PEt3)Cl. Chlorotrimethylsilane (0.15 mL, 1.2 mmol, 11
equiv) was added to a solution of 1a (46 mg, 0.11 mmol) in diethyl
ether (5 mL). Over 2 h, the solution turned from blue to red. Volatile
materials were removed in Vacuo, and the residue was redissolved in
ether (10 mL), filtered, concentrated to 2 mL, and cooled to -30 °C.
This gave small red blocks (29 mg, 76%) of analytical purity. 1H NMR
(C6D6): δ 1.45 (s, 15, Cp*), 1.33 (m, 6, PCH2CH3), 0.96 (m, 9,
PCH2CH3). 13C{1H} NMR (C6D6): δ 102.6 (Cp* quaternary), 15.0
(d, J ) 30 Hz, P(CH2CH3)3), 10.4 (P(CH2CH3)3), 8.1 (C5Me5). 31P-
{1H} NMR (C6D6): δ 22.95. IR (C6D6): 2964 (s), 2910 (s), 2875
(m), 1462 (w), 1415 (m), 1377 (m), 1347 (w), 1330 (w), 1249 (w),
1035 (s), 764 (s), 720 (s), 554 (s). Anal. Calcd for C16H30ClNiP: C,
55.29; H, 8.70. Found: 55.25; H, 8.85.
Cp*Ni(PEt3)CH2Ph. A red solution of KCH2Ph (21.3 mg, 0.164
mmol) in THF (4 mL) was added to a blue solution of 1a (65 mg,
0.155 mmol) in THF (4 mL). The mixture immediately turned green-
brown. After the mixture was stirred for 1 h, the volatile materials
were removed in Vacuo. The residue was extracted with 8 mL pentane
and filtered through a pad of Celite. The green filtrate was determined
to be >98% pure by 1H and 31P NMR spectroscopy. Analytically pure
material could be obtained by crystallization from pentane. 1H NMR
(C6D6): δ 7.39 (d, 2, J ) 7 Hz, Ph), 7.11 (t, 2, J ) 8 Hz, Ph), 6.90 (m,
1, Ph), 1.75 (d, 15, J ) 0.8 Hz, Cp*), 1.37 (d, 2, J ) 6 Hz, Ni-CH2),
0.98 (m, 6, P(CH2CH3)3), 0.75 (m, 9, P(CH2CH3)3). 13C{1H} NMR
(C6D6): δ 153.6 (Ph quaternary), 129.7 (Ph methine), 127.5 (Ph
methine), 122.5 (Tol methine), 98.7 (C5Me5), 13.9 (d, J ) 20 Hz, P(CH2-
CH3)3), 10.7 (P(CH2CH3)3), 7.7 (C5Me5), 3.9 (d, J ) 20 Hz, Ni-CH2).
31P{1H} NMR (C6D6): δ 29.4. IR (Nujol): 1592 (m), 1157 (m), 1035
(s), 1025 (s), 763 (s), 750 (s), 715 (s), 699 (s). Anal. Calcd: C, 68.51;
H, 9.25. Found: C, 68.44; H, 9.49.
and of 31P NMR spectra; the error was estimated from the variance of
the ratios using different peaks. For this equilibrium, the final relative
ratio [1a]: [Cp*Ni(PEt3)NHAr] ) [TolNH2]: [m-H2NC6H4NMe2] was
1:1.55 ((0.05):2.2 ((0.1); thus, Keq ) 1.1 ( 0.1. Errors were estimated
from the range of measured Keq values. A list of equilibration times
and Keq values for the various equilibria is in the Supporting Informa-
tion.
Equilibrium Measurements by UV/Vis Spectroscopy. The ex-
tinction coefficients (at 500, 538, 608, and 690 nm) for 1a, 4, and
TolNH2 were derived from successive dilution of solutions having
roughly the molarity used in the equilibrium experiments. Linear plots
verified that Beer’s Law holds in the concentration regime used. In
the actual equilibration experiments, measured volumes of solutions
of 4 and TolNH2 were mixed, and the absorbances at 500, 538, 608,
and 690 nm were monitored. These approached equilibrium values
over a few hours, and eventually began dropping, presumably due to
decomposition. Equilibrium constants were calculated from the equi-
librium absorbances at each wavelength monitored; however, reproduc-
ible values were found only using the high-wavelength absorbance
values (608 and 690 nm). The equations used to derive equilibrium
constants (A ) absorbance) at each wavelength were
[1a]eq ) (A - (ꢀ4[4]o) - (ꢀTolNH [TolNH2]o))/(ꢀ1a - ꢀ4) (11)
2
Keq ) [1a]eq2/([TolNH2]o)([4]o - [1a]eq)
(12)
Thus, the change in the small absorbance of the organic molecules in
the visible region was ignored.
The experiment was repeated using several different [4]o and
[TolNH2]o in order to gauge reproducibility. The concentrations and
Keq values used may be found in the Supporting Information. Errors
were estimated from the variance in Keq values derived from eqs 11
and 12, as the standard deviation (∼70% confidence).
(DPPE)Pt(Me)(O(p-C6H4Me)). A solution of p-cresol (5.6 mg, 52
µmol) in CH2Cl2 (2 mL) was added to a solution of (DPPE)Pt(Me)-
117
(OH)‚C6H6 (35.2 mg, 50.0 µmol) in CH2Cl2 (1 mL) and stirred for
1.5 d. The volatile materials were removed in Vacuo, and the residue
was treated with benzene and then evaporated several times to remove
water. This residue was dissolved in CH2Cl2/ether; vapor diffusion of
pentane into this solution gave pale yellow crystals of (DPPE)Pt(Me)-
(O(p-C6H4Me)) (24 mg, 67% yield). 1H NMR (THF-d8): δ 7.98 (m,
4, Ph), 7.75 (m, 4, Ph), 7.45 (m, 6, Ph), 7.32 (m, 6, Ph), 6.63 (d, 2, J
) 8.4 Hz, Tol), 6.57 (d, 2, J ) 8.4 Hz, Tol), 2.45 (m, 2, PCH2CH2P),
2.24 (m, 2, PCH2CH2P), 2.09 (s, 3, Tol), 0.42 (dd, 3, J ) 2.9, 7.8 Hz,
Pt-Me). 31P{1H} NMR (THF-d8): δ 45.14 (JPt-P ) 1764 Hz), 39.24
(JPt-P ) 3811 Hz).
Phosphine Exchange Kinetics. In the glovebox, a solution of
Cp*Ni(PEt3)X (2 mg) in C6D6 (0.5 mL) was placed in a J. Young NMR
tube and degassed. An aliquot of trimethylphosphine at known pressure
in a bulb of known volume was condensed into the NMR tube at 77
K. The sample was placed in an NMR probe (300 MHz) at the
appropriate temperature (X ) OTol: 235 ( 3 K;119 X ) STol, SC6H4-
CF3: 31.7 ( 0.1 °C), and one-pulse spectra were acquired every 2
min (X ) OTol), 3 min (X ) OTol), or 6 min (X ) SAr) for several
hours; in each case, the reaction was monitored over at least 3 half-
lives. The spectra were integrated, and the integrations for the Cp*
methyl resonances (X ) SAr) or the tolyl methyl resonances (X )
OTol) were plotted and fit using KaleidaGraph. Error limits are derived
from Kaleidagraph’s curve-fitting procedures. Phosphine concentrations
were derived from integration of the NMR spectra, and were corrected
for the volume of trimethylphosphine.
(DPPE)Pt(Me)(NH(p-C6H4Me)). This uses the method of Bryndza
et al.72 A flask was charged with LiNH(p-C6H4Me) (17 mg, 150 µmol),
(DPPE)Pt(Me)(Cl)118 (78 mg, 121 µmol) and a stir bar and capped with
a vacuum adapter. The flask was cooled to -196 °C, THF (8 mL)
was condensed into the flask, and the yellow solution was slowly
warmed to room temperature. After the mixture was stirred for 10
min at room temperature, the volatile materials were removed in Vacuo.
In the glovebox, the residue was extracted with benzene (6 mL), filtered
through a Celite pad, and the solvent was evaporated to leave a fluffy
yellow powder of (DPPE)Pt(Me)(NH(p-C6H4Me)) (81 mg, 94% yield).
This material was not completely pure, but was used for spectral
comparisons only. 1H NMR (THF-d8): δ 7.83 (m, 4, Ph), 7.76 (m, 4,
Ph), 7.44 (m, 6, Ph), 7.35 (m, 6, Ph), 6.38 (d, 2, J ) 8.2 Hz, Tol), 6.22
(d, 2, J ) 8.3 Hz, Tol), 2.39 (m, 2, PCH2CH2P), 2.19 (m, 2, PCH2-
CH2P), 2.00 (s, 3, Tol), 0.54 (dd, 3, J ) 4.8, 7.6 Hz, Pt-Me). 31P{1H}
NMR (THF-d8): δ 44.95 (JPt-P ) 1729 Hz), 44.69 (JPt-P ) 3145 Hz).
X-ray Crystal Structure of 1a. General procedures for crystal-
lography on the CAD4 system at Berkeley have been given previ-
ously;120 parameters specific to this data set are given in Table 4.
Automatic peak search and indexing procedures yielded a monoclinic
reduced primitive cell. Inspection of the Niggli values revealed a
potential C-centered orthorhombic unit cell, but inspection of the data
showed only monoclinic symmetry. Inspection of the intensity
standards revealed a reduction of 6.6% of the original intensity; the
data were corrected for this decay. Inspection of the azimuthal scan
data showed a variation Imin/Imax ) 0.72 for the average curve.
However, the curves clearly showed that the crystal was not well-
oriented at the time of the collection. An empirical correction was
made to the data on the basis of the combined differences of Fobs and
Equilibrium Measurements by NMR Spectroscopy. In a typical
experiment, 1a (10.7 mg, 25.6 µmol) and m-H2NC6H4NMe2 (4.7 mg,
35 µmol) were weighed into a vial. This mixture was dissolved in
C6D6 (0.5 mL), and the resulting blue solution was placed in a J. Young
NMR tube. The solution was monitored by 1H and 31P NMR
spectroscopy, which indicated that equilibrium had been reached in
Fcalc following refinement of all atoms with isotropic thermal parameters
(Tmax ) 1.20, Tmin ) 0.87, no θ dependence). Inspection of the
(119) Although the absolute temperature is not known to an accuracy
of better than (3 K, the temperature was maintained to (0.1 K within
each kinetic run using 4.
(117) Appleton, T. G.; Bennett, M. A. Inorg. Chem. 1978, 17, 738.
(118) Clark, H. C.; Jablonski, C. R. Inorg. Chem. 1975, 14, 1518.
(120) Dobbs, D. A.; Bergman, R. G. Inorg. Chem. 1994, 33, 5329.

12814 J. Am. Chem. Soc., Vol. 119, No. 52, 1997
Table 4. Crystal and Data Parameters for 1a and 4‚HOXyl
Holland et al.
Table 5. Calorimetric Enthalpies for Reaction and Solution in
C6H6 at 30 °C
compound
1a
4‚HOXyl
complex
solvent + reactant
∆H (kcal/mol)
crystal dimensions (mm)
0.25 × 0.27 × 0.45 0.32 × 0.19 × 0.10
24 (26° < θ < 30°) 8752 with
I > 10σ(I)
1a
1a
1a
4
C6H6
6.5 (0.2)
-2.9 (0.2)
-16.5 (0.2)
4.6 (0.2)
reflections for cell
determination
a (Å)
b (Å)
c (Å)
C6H6 + TolOH
C6H6 + TolSH
C6H6
16.639(3)
16.648(3)
16.615(3)
91.615(3)
4600.5(25)
P21/c (no. 14)
8
11.4752(3)
13.7416(3)
18.8274(5)
99.734(1)
2926.1(1)
P21/c (no. 14)
4
4
C6H6 + TolSH
-9.4 (0.1)
â (deg)
V (Å3)
the organometallic complex was placed in a Wilmad screw-capped
NMR tube fitted with a septum, and C6D6 was subsequently added.
The solution was titrated with a solution of the ligand of interest by
injecting the latter in aliquots through the septum with a microsyringe,
followed by vigorous shaking. The reactions were monitored by 31P
and 1H NMR spectroscopy and were found to be rapid, clean, and
quantitative under experimental calorimetric conditions. These condi-
tions are necessary for accurate and meaningful calorimetric results
and were satisfied for all organometallic reactions investigated.
space group
Z
µ
calc (Mo KR, cm-1
)
9.2
7.43
T (°C)
-102
N/A
-141
30
frame time (s)
no. of reflections
6246
11986
no. of unique reflections
min, Tmax
no. observations (I > 3σ(I)) 4059
no. variables
5987
1.20, 0.87
4384 (Rint ) 4.3%)
0.990, 0.834
3332
T
469
459
Calorimetric Measurement for Reaction between 1a and TolOH.
The mixing vessels of the Setaram C-80 were cleaned, dried in an oven
maintained at 120 °C, and then taken into the glovebox. A 20 mg
sample of 1a was accurately weighed into the lower vessel, closed,
and sealed with 1.5 mL of mercury. Four milliliters of 1 equiv stock
solution of cresol (26 mg of cresol in 20 mL of C6H6) was added and
the remainder of the cell was assembled, removed from the glovebox
and inserted in the calorimeter. The reference vessel was loaded in an
identical fashion with the exception that no organonickel complex was
added to the lower vessel. After the calorimeter had reached thermal
equilibrium (about 2 h) the reactants were mixed by inverting the
calorimeter, and the vessel was then allowed to return to equilibrium.
Then, the vessels were removed from the calorimeter, taken into the
glovebox, opened, and analyzed using 1H NMR spectroscopy. Conver-
sion to 4 was found to be quantitative under these reaction conditions.
The enthalpy of reaction, -2.9 ( 0.2 kcal/mol, represents the average
of five individual calorimetric determinations. The enthalpy of solution
of 1a was then subtracted from this value to obtain a value of -9.4 (
0.3 kcal/mol for an enthalpy of reaction with all species in solution.
R; Rw
0.047, 0.056
0.082
1.50
0.030, 0.039
0.040
1.43
R (including zeros)
goodness of fit indicator
systematic absences indicated uniquely space group P21/c. Hydrogen
atoms were assigned idealized locations and values of Biso approximately
1.25 times the Beq of the atoms to which they were attached. The amino
hydrogens were located on a difference Fourier map prior to their
inclusion in calculated positions. All hydrogens were included in
structure factor calculations, but not refined. Final agreement factors
are as follows: R ) 4.7%, wR ) 5.6%, and GOF ) 1.50; using all
unique data, R ) 8.15%. The largest peak in the final difference Fourier
map had an electron density of 0.52 e-/Å3, and the lowest excursion
-0.15 e-/Å3.
X-ray Crystal Structure of 4‚HOXyl. General procedures for
crystallography on the SMART/CCD system with integration and
solution using SAINT and teXsan, respectively, are given in the
accompanying paper;22 parameters specific to this data set are given in
Table 4. Red plates were obtained by cooling a pentane solution
Calorimetric Measurement of Enthalpy of Solution of 1a in C6H6.
This was performed by using a similar procedure as the one described
above with the exception that no TolOH was added to the reaction
cell. This enthalpy of solution, representing the average of five
individual determinations, was measured as 6.5 ( 0.2 kcal/mol.
containing equimolar amounts of 4 and HOXyl to -30 °C.
semiempirical ellipsoidal absorption correction (Tmax ) 0.990, Tmin
A
)
0.834) was applied to the raw data. Equivalent reflections were
averaged. The non-hydrogen atoms were refined anisotropically. The
hydrogen atoms were located in Fourier syntheses and their positions
were refined in the final stages of refinement, using fixed isotropic
thermal parameters of approximately 1.2 times that of the atom to which
they were attached. Because of the high quality of the data set, it was
possible to refine the hydrogen-bonding hydrogen atom isotropically;
the value of Biso ) 3.3 for this atom shows that this model behaved
well. Final agreement factors are as follows: R ) 3.0%, Rw ) 3.9%,
and GOF ) 1.43; using all unique data, R ) 4.0%, Rw ) 4.2%.
Calorimetry. Only materials of high purity as indicated by IR and
NMR spectroscopies were used in the calorimetric experiments.
Calorimetric measurements were performed using a Calvet calorimeter
(Setaram C-80) which was periodically calibrated using the TRIS
reaction121 or the enthalpy of solution of KCl in water.122 The
experimental enthalpies for these two standard reactions compared very
closely to literature values. This calorimeter has been previously
described123 and typical procedures are described below. Experimental
enthalpy data (Table 5) are reported with 95% confidence limits.
NMR Titrations. Prior to every set of calorimetric experiments
involving a new ligand, an accurately weighed amount ((0.1 mg) of
Acknowledgment. R.G.B. is grateful for financial support
from the National Institutes of Health (grant no. GM-25459)
and from the Arthur C. Cope Fund administered by the
American Chemical Society, and S.P.N. thanks the National
Science Foundation (CHE- 963116) for support. The structure
of 1a was solved by F. J. Hollander, director of the UCB
CHEXRAY facility. The authors also thank Henry Bryndza and
John Bercaw for helpful discussions, and David Glueck for
sharing unpublished information.
Supporting Information Available: Data from equilibration
experiments, and tables of crystal structure determination data
for 1a and 4‚HOXyl (10 pages). See any current masthead
page for ordering and Internet access instructions.
JA971829P
(123) (a) Nolan, S. P.; Hoff, C. D.; Landrum, J. T. J. Organomet. Chem.
1985, 282, 357. (b) Nolan, S. P.; Lopez de la Vega, R.; Hoff, C. D. Inorg.
Chem. 1986, 25, 4446.
(121) Ojelund, G.; Wadso¨, I. Acta Chem. Scand. 1968, 22, 1691.
(122) Kilday, M. V. J. Res. Natl. Bur. Stand. (U.S.) 1980, 85, 467.