A π-donor spectrochemical series for X in (Me5C5)2TiX, and β-agostic interactions in X = Et and N(Me)Ph

Article abstract of DOI:10.1021/ja953272o

The EPR and electronic spectra of d¹-bent metallocene compounds of the type Cp*₂TiX, where X is halide, alkoxide, amide, alkyl, or hydride and Cp* = Me₅C₅, have been studied. Several of these compounds are new, and those with X = N(Me)H and F were characterized by X-ray crystallography. The crystal structure of Cp*₂TiN(Me)H showed that the N(Me)H ligand lies on the plane defined by Cp*(centroid)-Ti-Cp*(centroid). This is the sterically most unfavorable conformation but allows maximum Ti-N π-bonding. The anisotropic frozen solution EPR spectra were analyzed by the method used by Petersen and Dahl for the d¹-metallocenes, Cp₂VX₂, which gives g(x), g(y), and g(z). Although the values of g(x) and g(z) are relatively constant throughout the series, the value of g(y) varies with the π-donor ability of X. The π-donor series is N(Me)H ~ NH₂ ~ OMe > OPh ~ F > N(Me)Ph > CI > Br > I > H. Among the known alkyls, the π-donor ability was Et > Me > n-Pr ~ CH₂CMe₃ > CH₂C₆H₅, which is rationalized, in part, by a β-agostic interaction in the case of Et. The β-agostic interaction in Cp*₂TiEt and in Cp*₂TiN(Me)Ph was investigated by variable-temperature EPR spectroscopy giving an enthalpy and entropy for the agostic interaction. For Cp*₂TiEt the parameters for the agostic interaction are ΔH° = -1.93(3) kcal/mol and AS° = -6.3(2) eu, and for Cp*₂TiN(Me)Ph, ΔH* = -1.5(1) kcal/mol and ΔS° = -7.9(5) eu.

Full text of DOI:10.1021/ja953272o

J. Am. Chem. Soc. 1996, 118, 1719-1728
1719
A π-Donor Spectrochemical Series for X in (Me5C5)2TiX, and
â-Agostic Interactions in X ) Et and N(Me)Ph
Wayne W. Lukens, Jr., Milton R. Smith III, and Richard A. Andersen*
Contribution from the Department of Chemistry, Chemical Sciences DiVision, Lawrence Berkeley
Laboratory, UniVersity of California, Berkeley, California 94720
ReceiVed September 25, 1995X
Abstract: The EPR and electronic spectra of d1-bent metallocene compounds of the type Cp*2TiX, where X is
halide, alkoxide, amide, alkyl, or hydride and Cp* ) Me5C5, have been studied. Several of these compounds are
new, and those with X ) N(Me)H and F were characterized by X-ray crystallography. The crystal structure of
Cp*2TiN(Me)H showed that the N(Me)H ligand lies on the plane defined by Cp*(centroid)-Ti-Cp*(centroid).
This is the sterically most unfavorable conformation but allows maximum Ti-N π-bonding. The anisotropic frozen
solution EPR spectra were analyzed by the method used by Petersen and Dahl for the d1-metallocenes, Cp2VX2,
which gives gx, gy, and gz. Although the values of gx and gz are relatively constant throughout the series, the value
of gy varies with the π-donor ability of X. The π-donor series is N(Me)H ≈ NH2 ≈ OMe > OPh ≈ F > N(Me)Ph
> Cl > Br > I > H. Among the known alkyls, the π-donor ability was Et > Me > n-Pr ≈ CH2CMe3 > CH2C6H5,
which is rationalized, in part, by a â-agostic interaction in the case of Et. The â-agostic interaction in Cp*2TiEt and
in Cp*2TiN(Me)Ph was investigated by variable-temperature EPR spectroscopy giving an enthalpy and entropy for
the agostic interaction. For Cp*2TiEt the parameters for the agostic interaction are ∆H° ) -1.93(3) kcal/mol and
∆S° ) -6.3(2) eu, and for Cp*2TiN(Me)Ph, ∆H° ) -1.5(1) kcal/mol and ∆S° ) -7.9(5) eu.
Introduction
Cp*2TiX complexes (X is a monodentate, one-electron ligand
such as a halide, amide, alkoxide, or alkyl group and Cp* is
Me5C5) appear to be ideal for the study of ligand-to-metal
π-bonding. They are monomeric, unlike [Cp2TiX]2 in which
the electrons are coupled. They have a single electron in the
a1 orbital making electronic spectroscopy simple, unlike Cp2-
VX in which the a1 and b2 orbitals are singly occupied.1-3 The
trivalent decamethyltitanocenes have an empty b2 orbital avail-
able for π-bonding, unlike Cp2MX2 in which the b2 orbital is
used for σ-bonding.
The best known bonding model for bent metallocenes is due
to Lauher-Hoffmann; Figure 1 shows the metallocene orbitals
with two different coordinate systems.4-9 The coordinate
system used here, due to Petersen and Dahl, is more convenient
2
because this coordinate system minimizes the mixing the dz
and dx -y orbitals.10,11 This metallocene bonding model is
supported by the work of Petersen and Dahl in which single-
2
2
* Address correspondence to this author at the Department of Chemistry.
X Abstract published in AdVance ACS Abstracts, January 15, 1996.
(1) Green, J. C.; Payne, M. P.; Teuben, J. H. Organometallics 1983, 2,
203-210.
(2) Ko¨hler, F. H.; Hofmann, P.; Prossdorf, W. J. Am. Chem. Soc. 1981,
103, 6359-6367.
(3) Curtis, C. J.; Smart, J. C.; Robbins, J. L. Organometallics 1985, 4,
1283-1286.
(4) Lauher, J. W.; Hoffmann, R. J. Am. Chem. Soc. 1976, 98, 1729-
1742.
(5) Brintzinger, H. H.; Bartell, L. S. J. Am. Chem. Soc. 1970, 92, 1105-
1107.
Figure 1. Qualitative MO diagram for a bent metallocene (after Lauher
and Hoffmann).4 The broken arrows labeled 1a1 f 2a1, ∆Exy, and ∆Exz
show the observed transitions.
(6) Brintzinger, H. H.; Lohr, L. L.; Wong, K. L. T. J. Am. Chem. Soc.
1975, 97, 5146-5155.
crystal EPR spectroscopy shows that the single electron in (η5-
MeC5H4)2VCl2 and Cp2VS5 occupies an orbital that is perpen-
dicular to the plane formed by the metal and the two Cp
(7) Petersen, J. L.; Lichtenberger, D. L.; Fenske, R. F.; Dahl, L. F. J.
Am. Chem. Soc. 1975, 97, 6433-6441.
(8) Fieselmann, B. F.; Stucky, G. D. J. Organomet. Chem. 1977, 137,
43-54.
2
centroids and is largely of dz parentage (in this coordinate
system).10,11 In Cp*2TiX, the unpaired electron resides in the
(9) Albright, T. A.; Burdett, J. K.; Whangbo, M. H. Orbital Interactions
in Chemistry; Wiley: New York, 1985.
2
low-lying a1 orbital which is largely dz . The empty b2 orbital
(10) Petersen, J. L.; Dahl, L. F. J. Am. Chem. Soc. 1975, 97, 6416-
6422.
(11) Petersen, J. L.; Dahl, L. F. J. Am. Chem. Soc. 1975, 97, 6422-
6433.
can interaction with the pz orbital of the X ligand to form a
π-bond and a π-antibond; the latter is the b2 orbital. Thus, the
energy of the 1a1 f b2 transition depends directly upon the
0002-7863/96/1518-1719$12.00/0 © 1996 American Chemical Society

1720 J. Am. Chem. Soc., Vol. 118, No. 7, 1996
Lukens et al.
8êk2b′ 2
2
2
π-donor ability of X. By comparing the energy of this transition
to that of a complex with a σ-only ligand, Cp*2TiH, the amount
of destabilization of b2 due to π-bonding can be quantified. A
combination of optical and EPR spectra of these complexes can
be used to rank ligands according to the strength of their
π-interaction with the Cp*2Ti fragment.
x
2êk (a′ 3 - b′)
gy ) g0 -
;
gz ) g0 -
(3)
∆Exz
∆Exy
2
ky
b2kz2
x
a′
b′
1
a
3
)
1 +
- 1
;
k2 )
;
a′ 2 + b′ 2 ) 1
(
)
[
]
b
x
k 2
b′ 2
3
x
z
The parameters a′, b′, and k2 no longer have a direct physical meaning,
but allow the transition energies to be determined if the covalency is
anisotropic. If the covalency actually is isotropic (ky2 ) kz2), then a′
and b′ are the same as a and b. From the relative energies of the b1
and b2 orbitals, it is likely that ky2 g k2 g kz2, so b will be slightly
larger than b′. If gy, ∆Exz, gz, and k2 are known, eq 3 can be used to
obtain ∆Exz.
Equation 3 helps assign the EPR spectra of the Cp*2TiX complexes.
Since ∆Exz (1a1 f b2) is much smaller than ∆Eyz or ∆Exy and since a
will be much greater than b (in (MeCp)2VCl2, a2/b2 ) 20), gy will be
the g component with the smallest value. In addition, since ∆Exz
changes depending upon the π donor ability of X, gy will also change
greatly among the complexes. Since b is much smaller than a, gz will
have the largest value and be close to g0. The middle component of
the EPR spectra will be gx.
Theory
For a Cp*2TiX complex, three d-d absorptions are expected: 1a1
f b2, 1a1 f b1, and 1a1 f 2a1 in order of increasing energy (see Figure
1). The 1a1 f a2 transition should be similar in energy to the 1a1 f
b1 and 1a1 f 2a1 absorptions, but the 1a1 f a2 transition is electric
dipole forbidden and should be weak or unobserved. Of the three
absorptions, two should be to higher energy, 1a1 f 2a1 and 1a1 f b1,
and one should be much lower in energy, 1a1 f b2. Of the two higher
energy transitions, the 1a1 f 2a1 transition will have greater intensity
because this orbital is σ-antibonding toward X and the 1a1 f 2a1
transition should have some charge transfer character.12 The 1a1 f b1
transition should be less intense but somewhat similar in energy. The
b1 orbital is Ti-Cp* antibonding and Ti-X π-antibonding if the X
ligand has a filled orbital of b1 symmetry capable of acting as a π-donor
(e.g. the nitrogen lone pair of Cp*2TiN(Me)Ph). The 1a1 f b2 transition
will be much lower in energy.
Results
The EPR spectra are closely related to the electronic spectra. As
shown by McGarvey, the deviation of the gi values from g0 is due to
coupling of excited states into the ground state as shown in eq 1 where
The syntheses of the titanium complexes were straightfor-
ward. Teuben has shown that Cp*2TiCl is a useful synthon for
the preparation of Cp*2TiX complexes by chloride metathe-
sis.15,16 This synthetic route was used to prepare additional
examples of Cp*2TiX where X ) F, N(Me)H, OMe, or OPh,
all potential π-donor ligands. The brown-purple methoxide and
phenoxide and the lilac colored methylamide are soluble in
hexane from which they were crystallized. The N-H stretching
frequency of the methylamide is a sharp, low-intensity feature
found at 3370 cm-1 in the solid state. The salt elimination
metathesis did not yield the simplest amide, Cp*2TiNH2, cleanly;
however, this amide was prepared from Cp*2TiMe and ammonia
in hexane from which it was crystallized. The infrared spectrum
shows a single, sharp N-H absorption at 3437 cm-1 in the solid
state.
0|Li|n n|Li|0
En - E0
gi ) g0 - 2ê
(1)
∑
n
i is x, y, or z; ê is the spin-orbit coupling constant, En - E0 is the
difference in energy of the orbitals, and the sum is over all orbitals
containing d-character.13 For bent metallocenes, Petersen and Dahl have
shown that the relationship of the g values to the energies of the excited
states is as shown in eq 2, where ê is the spin-orbit coupling constant
2
2
2
2
x
x
2êkx (a 3 + b)
2êky (a 3 - b)
gx ) g0 -
;
gy ) g0 -
;
∆Eyz
∆Exz
The fluoride, Cp*2TiF, was synthesized in two steps. First,
the difluoride, Cp*2TiF2, was prepared by the method used by
Lappert to make Cp2TiF2, the reaction of Cp*2TiMe2 with BF3‚
OEt2 in diethyl ether.17 Curiously, the difluoride is almost
identical in color and solubility to Cp*2TiMe2; consequently,
the reaction proceeded with little color change. Reduction of
the difluoride with potassium-graphite18 gave Cp*2TiF as green
crystals from hexane. Recently, Cp*2TiF has also been prepared
by the reaction of Cp*2TiCl with Me3SnF.19
The solid-state structure of Cp*2TiN(Me)H is shown in Figure
2, and is almost identical to that of the amide Cp*2TiNH2.20
Useful bonding parameters are listed in Table 1. The most
interesting aspect of the crystal structure is the orientation of
the methylamide ligand which adopts the least sterically
favorable conformation. In Cp*2TiNH2, the amide group adopts
a similar conformation. The methylamide group lies just slightly
out of the plane formed by the titanium atom and the two ligand
centroids with a Cp1-Ti-N-C21 torsion angle of 13.5°. The
8êkz2b2
∆Exy
gz ) g0 -
(2)
for Ti(III), 154 cm-1, g0 is 2.002 (the value of g for a free electron),
∆Eyz, ∆Exy, and ∆Exz are the energies of the excited states of dyz, dxy,
and dxz character relative to 1a1, that is, the a2, b1, and b2 orbitals,
2
2
2
respectively, and a and b are the coefficients of dz and dx -y in the
ground state, Ψ ) a|dz + b|dx -y . The k2 terms have been added to
the original equations to account for covalency.10,11 Since ∆Eyz is not
measured, only the last two relationships of eq 2 will be used. The
use of eq 2 implicitly assumes that only spin-orbit coupling to the
unoccupied d-orbitals is important. Since the d-orbitals are involved
in bonding, some low-lying orbitals will also have d-character and could
potentially affect the value of g. However, since these orbitals are
much further in energy from 1a1 and contain little d-orbital character,1
they are not expected to change g to any great extent.
2
2
2
Because it is not possible to derive the values of a, ky2, and kz2 from
only the last two relationships of eq 2, the assumption that ky2 ) kz2
)
k2 is made implying that the covalency is isotropic. Since these orbitals
are involved in bonding to the Cp* versus X ligands, this assumption
might not be true.14 Substituting the parameters k2, a′, and b′ for ky2,
kz2, a, and b gives the following equation (eq 3) which is applicable
whether the covalency is anisotropic (ky2 * kz2) or isotropic (ky2 ) kz2).
(15) Luinstra, G. A.; ten Cate, L. C.; Heeres, H. J.; Pattiasina, J. W.;
Meetsma, A.; Teuben, J. H. Organometallics 1991, 10, 3227-3237.
(16) Pattiasina, J. W.; Heeres, H. J.; van Bolhuis, F.; Meetsma, A.;
Teuben, J. H. Organometallics 1987, 6, 1004-1010.
(17) Druce, P. M.; Kingston, B. M.; Lappert, M. F.; Spalding, T. R.;
Srivastava, R. C. J. Chem. Soc. A 1969, 2106-2110.
(18) Schwindt, M. A.; Lejon, T.; Hegedus, L. B. Organometallics 1990,
9, 2814-2819.
(19) Herzog, A.; Liu, F.; Roesky, H. W.; Demsar, A.; Keller, K.;
Noltemeyer, M.; Pauer, F. Organometallics 1994, 13, 1251-1256.
(20) Brady, E.; Lukens, W.; Telford, J.; Mitchell, G. Acta. Crystallogr.
1994, C51, 558-560.
(12) Lever, A. B. P. Inorganic Electronic Spectroscopy, 2nd ed.;
Elsevier: New York, 1984.
(13) McGarvey, B. R. In Transition Metal Chemistry, a Series of
AdVances; Carlin, R. L., Ed.; Marcel Dekker, Inc.: New York, 1966; Vol.
3, pp 90-201.
(14) As noted by one of the reviewers of the earlier manuscript.

A π-Donor Spectrochemical Series in (Me5C5)2 TiX
J. Am. Chem. Soc., Vol. 118, No. 7, 1996 1721
Figure 3. ORTEP drawing of (Me5C5)2TiF with 50% probability
thermal ellipsoids.
Table 2. Selected Distances (Å) and Angles (deg) in Cp*2TiF
Figure 2. ORTEP drawing of (Me5C5)2TiN(Me)H with 50% prob-
ability thermal ellipsoids.
molecule 1
molecule 2
Table 1. Selected Distances and Angles in Cp*2TiN(Me)H
Ti1-F1
1.845(4)
2.38(2)
2.06
2.06
1.44.1
107.3
108.5
Ti2-F2
1.838(4)
2.38(2)
2.05
2.05
145.6
106.3
108.0
Ti1- Cring
Ti1-Cp1
Ti1-Cp2
Ti2- Cring
Ti2-Cp3
Ti2-Cp4
distances, Å
angles, deg
Ti-N
1.955(5)
2.084
2.094
2.41(2)
0.77(7)
1.446(8)
Cp1-Ti-Cp2
141.7
13.5
110.4
107.9
105 (5)
110 (5)
144.9 (5)
Cp1-Ti1-Cp2
Cp1-Ti1-F1
Cp2-Ti1-F1
Cp3-Ti2-Cp4
Cp3-Ti2-F2
cp4-Ti2-F2
Ti-Cp1
Ti-Cp2
Ti- Cring
N-H1
Cp1-Ti-N-C21
N-Ti-Cp1
N-Ti-Cp2
C21-N-H1
Ti-N-H1
N-C21
and 1.838(4) Å. However, as seen in Figure 3, the fluorine
atoms have large thermal parameters making the bond lengths
seem shorter. The bonds lengths corrected for the thermal
motion using the root-mean-square displacements are 1.860 and
1.855 Å, respectively.23 The corrected bond distances are 0.5
Å shorter that the Ti-Cl distance of 2.363(1) Å in Cp*2TiCl16
in agreement with the size difference between chloride and
fluoride.24 Like Cp*2TiN(Me)H, the rest of the structure of
Cp*2TiF is similar to the other known Cp*2TiX struc-
tures.15,16,20,22
The EPR spectra of the new compounds and several known
Cp*2TiX compounds15,16,25 were measured as methylcyclohex-
ane solutions at room temperature and as frozen glasses. The
EPR results are listed in Table 3. For Cp*2TiBr and Cp*2TiI,
the EPR parameters were obtained from the simulated spectra.
Like the spectra of Cp*2TiBr and Cp*2TiI,25 the spectra of
Cp*2TiH and Cp*2TiF display ligand hyperfine coupling at low
temperature. The EPR spectra and simulations for some of these
complexes are shown in Figure 4. The spectra were assigned
as outlined above.
The electronic spectra of several Cp*2TiX were measured at
room temperature as 10-2 M solutions in methylcyclohexane.
The spectrum of Cp*2TiOMe is shown Figure 5. In addition,
the spectra of Cp*2TiEt and Cp*2TiN(Me)Ph at -78 °C are
shown in Figure 6 along with the least-squares fits used to obtain
the peak positions. The spectra were fit using a sum of Gaussian
peaks. The energies of the peaks determined in this way are
listed in Table 4. In the visible region, two peaks are present
for all complexes: a more intense peak at higher energy and a
less intense peak at lower energy. In the near-infrared, weak
transitions are observed for some of the compounds. The energy
of the near-IR absorption varies from 5630 cm-1 for Cp*2TiF
Ti-N-C21
steric interaction of the amide methyl group with the Cp* ligand
bends the amide group “down” opening the Ti-N-Me angle
to 145° rather than the 120° expected for an sp2-hybridized
nitrogen atom. In Cp*2TiNH2, the Ti-N-H angle is only 126°.
Stabilization of the nitrogen lone pair by interaction with the
empty b2 orbital is presumably the reason the methylamide
ligand adopts this conformation. A similar explanation was
given for the orientation of the methylamide ligand in the solid-
state structure of Cp*2Hf(H)N(Me)H.21
In contrast to the orientation of the amide group in Cp*2TiN-
(Me)H, the crystal structure of Cp*2TiN(Me)Ph22 shows that
the N-methylanilide ligand is perpendicular to the Cp*2Ti
fragment with a Cp(centroid)-Ti-N-Me torsion angle of ca.
90° preventing the nitrogen lone pair from acting as a π-donor
to the empty b2 orbital. In Cp*2TiNH2 and Cp*2TiN(Me)H,
the conformation of the amide group relative to Cp*2Ti implies
maximum Ti-N π-bonding while in Cp*2TiN(Me)Ph, the
conformation of the amide group implies minimal π-bonding.
The Ti-N bond distances are consistent with this hypothesis.
In Cp*2TiNH2 and Cp*2TiN(Me)H, the Ti-N bond distances
are 1.944(2) and 1.955(2) Å, respectively, while the Ti-N bond
distances of Cp*2TiN(Me)Ph is 2.054(2) Å. Other than the
orientation of the amide ligand and the short Ti-N bond length,
the crystal structure is unremarkable. The other structural
features are similar to related crystallographically characterized
Cp*2TiX compounds.15,16,20,22
The crystal structure analysis of Cp*2TiF revealed two
crystallographically independent but virtually identical molecules
in the asymmetric unit, one of which is shown in Figure 3. The
important bond parameters for both independent molecules are
listed in Table 2. The Ti-F bond lengths are short at 1.845(4)
(23) Dunitz, J. X-Ray Analysis and the Structure of Organic Molecules;
Cornell University Press: Ithaca, New York, 1979.
(24) Pauling. L. The Nature of the Chemical Bond; Cornell University
Press: Ithaca, New York, 1960.
(25) Mach, K.; Raynor, B. J. Chem. Soc., Dalton Trans. 1992, 683-
688.
(21) Hillhouse, G. L.; Bulls, A. R.; Santarsiero, B. D.; Bercaw, J. E.
Organometallics 1988, 7, 1309-1312.
(22) Feldman, J.; Calabrese, J. C. J. Chem. Soc, Chem. Commun. 1991,
1042-1044.

1722 J. Am. Chem. Soc., Vol. 118, No. 7, 1996
Lukens et al.
Table 3. EPR Data for Cp*2TiX Compounds (Ligand Hyperfine
Coupling Constant in Parentheses in MHz)
a
b
compd
Cp*2TiH
gave
g
gzc
gx
gy
d
1.916 1.997(39) 1.981
1.780
Cp*2TiI
1.939 1.941 1.997
1.948 1.948 1.996
1.973(36) 1.852
Cp*2TiCH2Ph
1.978
1.985
1.984
1.870
1.880
1.881
Cp*Ti(η6-H2CC5Me4) 1.950 1.954 1.997
Cp*2TiCH2CMe3
Cp*2TiBr
1.951 1.952 1.998
1.953 1.953 1.996(12) 1.980(21) 1.883
Cp*2Ti(n-Pr)
Cp*2TiCl
Cp*2TiN(Me)Ph
Cp*2TiN(Et)Ph
Cp*2TiMe
Cp*2TiNMe2
Cp*2TiF
Cp*2TiEt
Cp*2TiOPh
Cp*2TiOMe
Cp*2TiNH2
Cp*2TiN(Me)H
1.953 1.955 1.998
1.956 1.957 1.999
1.958 1.972 1.999
1.955 1.958 1.998
1.958 1.959 1.998
1.962 1.967 1.998
1.984
1.984
1.981
1.980
1.981
1.979
1.884
1.889
1.937
1.895
1.898
1.924
1.938
1.974
1.945
1.956
1.962
1.965
1.972 1.973 1.998(37) 1.982
1.972 1.985 2.000
1.974 1.976 1.999
1.977 1.979 1.999
1.979 1.980 1.998
1.980 1.981 1.998
1.982
1.983
1.981
1.981
1.980
b
a The averaged g-values in solution at room temperature. g )
1/3(gx + gy + gz). c The anisotropic g-values from frozen solutions.
d Unobserved.
Figure 6. Electronic spectra (dotted lines) and least-squares fits (solid
lines) of Cp*2TiEt (a) and Cp*2TiN(Me)Ph (b) at 77 K.
Table 4. Electronic Transitions of Cp*2TiX Complexes in cm-1
(The Extinction Coefficient, L cm-1 mol-1, Is in Parentheses)
1a1 f
1a1 f
T (in K) 1a1 f 2a1 b2 (∆Exz) b1 (∆Exy)
Cp*2TiH
Cp*2TiI
295 20976(131)
295 16065(135)
295 20203(173)
18272(69)
14610(51)
16017(42)
17816(180)
15190(29)
17342(63)
15023(40)
15426(59)
15893(121)
17595
Cp*2TiCH2Ph
Cp*Ti(η6-Me4C5CH2) 295 23000(180)
Cp*2TiCH2CMe3
Cp*2Ti(n-Pr)
Cp*2TiBr
295 20340(182)
295 21702(189)
295 17260(131)
295 18118(110)
295 19465(247)
77 19850
295 21781(170)
295 20826(122)
77 19828
Figure 4. EPR spectra (solid lines) and simulations (dotted lines) of
(Me5C5)2TiH (a), (Me5C5)2TiF (b), (Me5C5)2TiBr (c), and (Me5C5)2TiI
(d) in a methylcyclohexane glass at ca. 70 K.
Cp*2TiCl
Cp*2TiN(Me)Ph
Cp*2TiN(Me)Ph
Cp*2TiMe
Cp*2TiEt
Cp*2TiEt
16665(50)
15895(21)
16550
8643
Cp*2TiF
295 23231(167) 57722(23) 17124(29)
295 19596(134) 6544(29) 15980(51)
295 19607(128) 7800(21) 16155(47)
Cp*2TiOPh
Cp*2TiOMe
Cp*2TiNH2
Cp*2TiN(Me)H
295 20369(90)
7942(6) 15422(34)
295 19593(114) 8180(8) 15159(40)
Discussion
As noted previously, the unpaired electron occupies the 1a1
2
2
2
orbital which is largely dz with some dx -y character. This
orbital is nonbonding and interacts only weakly with the X
ligand in Cp*2TiX and is, therefore, not expected to be sensitive
to the σ-donor ability of X.4 The LUMO, b2, is mainly dxz and
is close to 1a1 in energy in the absence of π-effects. When X
is a π-donor, b2 acts as a π-acceptor and is destabilized as the
extent of ligand π-donation increases. By comparing the energy
of the 1a1 f b2 transition for a series of complexes to the energy
of this transition of a complex with a σ-only ligand, Cp*2TiH,
the relative strength of the π-interaction in these complexes can
be determined. Since b2 is π-antibonding, the actual strength
of the π-interactions is somewhat less than the energy deter-
Figure 5. Electronic spectrum of Cp*2TiOMe in methylcyclohexane
at 298 K.
to 8220 cm-1 for Cp*2TiN(Me)H. No near-infrared absorption
was observed for many of the compounds presumably because
it was too low in energy. The spectra were assigned as outlined
above.

A π-Donor Spectrochemical Series in (Me5C5)2 TiX
J. Am. Chem. Soc., Vol. 118, No. 7, 1996 1723
Table 6. Calculated Values for Cp*2TiX Complexes
Table 5. Calculated Values of a′, b′, and k2 from the Electronic
and EPR Spectra
∆Exz
relative to
b′
a′
k2
b′
a′ ∆Exz(calc) ∆Exz(obs) Cp*2TiH
Cp*2TiF
0.29
0.25
0.26
0.30
0.30
0.96
0.97
0.97
0.95
0.95
0.64
0.60
0.58
0.56
0.54
Cp*2TiH
Cp*2TiI
Cp*2TiCH2C6H5
0.57 0.82
0.36 0.93
0.40 0.92
447
1537
1601
1794
2157
2096
1923
2852
4870
2410
5622
6445
7995
7479
8217
8695
0
1090
1154
1348
1710
1649
1476
2405
4423
1963
5175
5998
7549
7032
7770
8248
Cp*2TiOPh
Cp*2TiOMe
Cp*2TiNH2
Cp*2TiN(Me)H
Cp*Ti(η6-Me4C5CH2) 0.39 0.92
Cp*2TiCH2CMe3
Cp*2Ti(n-Pr)
Cp*2TiBr
Cp*2TiCl
Cp*2TiN(Me)Ph
Cp*2TiMe
0.32 0.95
0.34 0.94
0.38 0.92
0.27 0.96
0.28 0.96
0.33 0.94
0.29 0.96
0.25 0.97
0.26 0.97
0.30 0.95
0.30 0.95
0.33 0.94
Cp*2TiF
5738
6563
7700
7633
8180
8460
Cp*2TiOPh
Cp*2TiOMe
Cp*2TiNH2
Cp*2TiN(Me)H
Cp*2TiEta
a Not used in determining the k2 relationship.
in Cp*2TiH. A potential problem exists, however, in that Lauher
and Hoffmann have predicted that the hydride ligand does not
lie on the x-axis.4 This distortion increases the value of ∆Exz
for Cp*2TiH from a true σ-only value since the σ-orbital of the
hydride ligand will interact with the b2-orbital. However, more
recent calculations suggest that the hydride ligand does lie along
the x-axis.28 It should be noted that the CH2CMe3 ligand of
Cp*2TiCH2CMe3 does not lie on the x-axis,15 and the destabi-
lization of the b2 orbital in this complex is 550 cm-1 greater
than the destabilization of b2 in Cp*2TiCH2C6H5. Presumably,
the less sterically demanding benzyl ligand does lie on the x-axis.
The relatively high value of b′ for this complex is disturbing
since this observation implies that the bonding in Cp*TiH is
different from that in the other metallocenes.
Figure 7. Observed covalency versus ligand electronegativity for some
Cp*2TiX complexes. X is given next to the corresponding data point.
mined in this way. Thus, this method overemphasizes the
π-bonding capability of X.
In the compounds for which gy, gz, ∆Exz, and ∆Exy are
observed, the values of a′, b′, and k2 can be calculated directly.
The results are given in Table 5. The values of a′ and b′ vary
only slightly among these complexes, and the values of a′ 2/
b′ 2 are slightly smaller than those seen by Petersen and Dahl
for Cp2VX2 complexes,10,11 but are similar to those calculated
for Cp2TiS5.7 The fact that a′ and b′ have the same sign shows
A geometric distortion, e.g. bending of the hydride ligand,
could be responsible for the difference in bonding suggested
by the high value of b′ for Cp*2TiH. Its structure shows that
the hydride ligand does indeed lie on the x-axis of the molecule.
However, the metallocene angle ( Cp*-Ti-Cp*), 152°, is
much greater than that in other decamethyltitanocenes.29 The
larger metallocene angle in Cp*2TiH is the geometric distortion
responsible for the higher value of b′ in this complex. As the
2
that 1a1 resides mainly in the yz plane (the dz orbital is
2
compressed along the x axis, see the dz orbital in Figure 1). In
contrast, Petersen and Dahl found that in the Cp2VX2 complexes,
a and b had opposite signs, so that 1a1 is mainly in the xz
plane.10,11 However, for the related Cp2VCO, the ratio a2/b2 is
about the same as in Cp2VX2, but the signs of a and b are the
same as they are for Cp*2TiX complexes.26 This apparent
contradiction was explained by noting that in Cp2VCO, as in
the trivalent decamethyltitanocenes, the change in sign of b
reflects a decrease in electron density along the x axis which
minimizes a destabilizing interaction with the σ-bonding orbital
of the ligand.
2
2
2
metallocene angle increases, the dz and dx -y orbitals become
closer in energy and will interact more strongly, giving 1a1 more
dx -y character relative to the other complexes.4 The larger
metallocene angle in Cp*2TiH makes ∆Exz somewhat smaller
than it would be if the metallocene angle were the same as in
a normal Cp*2TiX complex. Consequently, using ∆Exz of
Cp*2TiH as the anchor for the π-bonding ability of X will, again,
tend to overestimate the π-bond strength.
2
2
Unlike the values of a′ and b′, the value of k2 changes with
the ligand. The greater the ligand electronegativity,27 the higher
the value of k2. As seen in Figure 7, this relationship is roughly
linear. The less electronegative ligands have a more covalent
interaction with the titanium center, decreasing k2 for the
unpaired electron.
The transition energy ∆Exz was calculated using the g values
from the EPR spectra, the observed value of ∆Exy, and k2 values
estimated using the linear relationship shown in Figure 7. The
results, along with the values of a′ and b′, are listed in Table 6.
The values of ∆Exz calculated from the EPR spectra agree fairly
well with those obtained from the near-infrared spectra.
The amount of destabilization of b2 caused by the π-donor
ligand is determined by comparing ∆Exz to the value of ∆Exz
Among the halides, the trend in π-bond strengths is F > Cl
> Br > I. This trend has been observed previously in other
analyses of bonding in bent metallocenes and has been attributed
to strong overlap between the p-orbitals of the halide and the
d-orbitals of the bent metallocene fragment.30,31 Furthermore,
fluoride is a fairly good π-donor, only slightly weaker than
phenoxide. The trend in π-bond strengths agrees with the trend
seen in octahedral Cr(III) complexes.12
A potential problem exists in the analysis of the halides since
the observed spin-orbit coupling could increase due to ligand
(28) Bierwagen, E. P.; Bercaw, J. E.; Goddard, W. A., III J. Am. Chem.
Soc. 1994, 116, 1481-1489.
(29) Lukens, W. W.; Matsunaga, P. T.; Andersen, R. A. To be
resubmitted.
(30) Hunter, J. A.; Lindsell, W. E.; McCullough, K. J.; Parr, R. A.;
Scholes, M. L. J. Chem. Soc., Dalton Trans. 1990, 2145-2153.
(31) Asaro, M. A.; Cooper, S. R.; Cooper, N. J. J. Am. Chem. Soc. 1986,
108, 5187-5193.
(26) Razuvaev, G. A.; Abakumov, G. A.; Cherkasov, V. K. Russ. Chem.
ReV. 1985, 54, 1235-1259.
(27) Inamoto, N.; Masuda, S. Chem. Lett. 1982, 1003-1006.

1724 J. Am. Chem. Soc., Vol. 118, No. 7, 1996
Lukens et al.
Figure 9. An equilibrium between agostic and anagostic conformers.
most likely by a â-agostic N-methyl group rather than a
γ-agostic phenyl group since Cp*2TiCH2Ph shows no evidence
of an agostic interaction. The presence of a â-N-methyl agostic
interaction is supported by weak infrared absorptions at 2570
and 2620 cm-1 and by the published crystal structure.22 In
Cp*2TiN(Me)Ph, the N(Me)Ph ligand is planar and lies in the
plane between the Cp* ligands. The Ti-N-CMe angle is
110.8(2)° while the Ti-N-Cipso angle is 131.6(1)°; this
geometry is consistent with a â-agostic interaction. On the other
hand, in the n-butyl isocyanide adduct, the Ti-N-CMe angle
is 121.2(4)° and the Ti-N-Cipso angle is 125.6(4)°.22 The
geometry of the base adduct is not consistent with a â-agostic
interaction. The lack of an agostic interaction is expected since
the base adduct has no low-lying, empty orbitals available to
form an agostic interaction.
For the equilibrium shown in Figure 9, K ) [anagostic]/
[agostic] ) exp(-∆H°/RT + ∆S°/R) where ∆H° and ∆S° are
the enthalpy and entropy of the anagostic conformation relative
to the agostic conformation, respectively. Based on the as-
sumption that the observed gave value represents the weighted
average of the gave values of the agostic and anagostic forms,
then the formula for the observed gave value is given in eq 4
where gagostic is the gave value for the agostic conformation,
ganagostic is the gave value for the anagostic conformation, and
∆H° and ∆S° are as defined earlier.
Figure 8. Destabilization of “b2” by a â-agostic ethyl ligand. Note
that the actual symmetry of the complex is Cs or C1; however, the
symmetry labels for C2V are given for consistency.
character in the metallocene d-orbitals. This effect would
decrease gy by increasing ê rather than by decreasing b2 making
the heavier halides seem like poorer π-donors than they actually
are. For Cp*2TiI with êI ) 5069 cm-1 32
, this effect would be
greatest, but since 3gave - gy (that is, gx + gz) is approximately
the same for all of these compounds, ê does not differ greatly
from that of the other complexes. In addition, based upon the
small values observed for the ligand hyperfine coupling, only
a small amount of ligand character is present in the 1a1 orbital.33
This observation is in agreement with PES studies on Cp2 VX
in which the amount of ligand character in the b2 orbitals was
found to be tiny.1
1
gobs
)
(gagostic + Kganagostic)
K + 1
K ) e-(∆H°-T∆S°)/RT
(4)
A more interesting observation is the high value of gy for
Cp*2TiEt. The magnitude of gy implies that the ethyl ligand is
a stronger π-donor than N(Me)H. The ethyl group in Cp*2TiEt
is thought to be â-agostic based upon the observation of low-
frequency C-H stretching absorption in its infrared spectrum.15
As shown in Figure 8, a â-agostic interaction will raise the
energy of b2 in much the same way as a π-interaction. In
addition to the high value of gy, at 77 K, the 1a1 f b2 transition
can be observed directly at 8640 cm-1. While this energy would
seem to indicate that the electronic contribution to the agostic
interaction is about 8000 cm-1, this is not true. If the σ-bond
of the ethyl group in Cp*2TiEt moves off of the x-axis, b2 will
be destabilized by the σ-bond in addition to the agostic
interaction, and the electronic contribution to the agostic
Fitting the gave values from the variable-temperature EPR
spectra of these complexes with eq 4, using the average of the
g components observed in the frozen glass spectra for gagostic
will give the values of the unknowns, ∆H°, ∆S°, and ganagostic
,
.
Plots of giso versus T for Cp*2TiEt and Cp*2N(Me)Ph are shown
in Figure 10. For Cp*2TiEt, the variable-temperature EPR data
from -98 to 68 °C yield ∆H° ) 1.93(3) kcal/mol, ∆S° ) 6.3(2)
eu, and ganagostic ) 1.9570(7). For Cp*2TiN(Me)Ph from -58°
to 105 °C, the values are ∆H° ) 1.5(1) kcal/mol, ∆S° ) 7.9(5)
eu, and ganagostic ) 1.9545(3). The data are for three separate
runs for each complex and assume an error of 1 × 10-4 in gave
(based upon σ(gave) for spectra acquired at the same tempera-
ture).
interaction will be considerably less than 8000 cm-1
.
The entropy difference between the agostic and anagostic
molecules is the same in both cases. The entropy difference
can be estimated by R ln(σ) where σ is the product of the
symmetry numbers of the anagostic molecule versus the agostic
molecule.34 Assuming all of the ligands are freely rotating, then
the anagostic molecule has C2V symmetry (σ ) 2) while the
agostic molecule has Cs symmetry (σ ) 1). Additionally, in
the agostic molecule, a 3-fold methyl rotation and a 2-fold Ti-
Et or Ti-N(Me)Ph rotation are being stopped. The symmetry
difference σ is 3 × 2 × 2 and R ln(σ) is 4.9 in rough agreement
with the observed value of ∆S°.
Curiously, the gave value determined from the room temper-
ature EPR spectrum of Cp*2TiEt is quite different from g ,
the average of the g components recorded in the frozen glass
spectrum. This observation, along with the inability to observe
the 1a1 f b2 transition at room temperature, led us to postulate
that an equilibrium between agostic and anagostic forms of
Cp*2TiEt is present, Figure 9.
In addition to Cp*2TiEt, gave and g are quite different in
Cp*2TiN(Me)Ph. The N(Me)Ph ligand also seems to be agostic,
(32) Moore, C. E. Atomic Energy Levels As Derived From the Analyses
of Optical Spectra. National Bureau of Standards, 1958.
(33) Weltner, W., Jr. Magnetic Atoms and Molecules; Dover Publications,
Inc.: New York, 1983.
(34) Lowry, T. H.; Richardson, K. S. Mechanism and Theory in Organic
Chemistry, 3rd.; Harper & Row: New York, 1987.

A π-Donor Spectrochemical Series in (Me5C5)2 TiX
J. Am. Chem. Soc., Vol. 118, No. 7, 1996 1725
Figure 11. Steric congestion in a â-agostic alkyl complex of Cp*2Ti.
it moves off the x-axis. In Cp*2TiN(Me)Ph, in which N(Me)-
Ph lies on the x-axis, b2 is destabilized by 6 kcal/mol in the
agostic conformation.
In both complexes, the enthalpy difference, ∆H°, is quite a
bit smaller than the destabilization of the b2 orbitals. The
destabilization of b2 reflects only the electronic contribution to
the agostic bond. The electronic contribution is greater than
the net interaction since it does not reflect destabilization of
the agostic conformation due to steric crowding or strain caused
by bending the ligand. The net enthalpy, ∆H°, of the agostic
bond is also much smaller than in agostic bonds in other
complexes. For example, in (Cy3P)2(CO)3W (Cy ) cyclohexyl)
one of the PCy3 ligands has a γ-agostic interaction with the
tungsten center. The strength of the agostic interaction is
estimated to be 16 kcal/mol.35 Additionally, theoretical calcula-
tions on the molecule Ti(Et)Cl3(dmpe) (dmpe ) 1,2-bis-
(dimethylphosphino)ethane) show that the agostic form is 12.4
kcal/mol lower in energy than the anagostic conformation.36,37
In comparing this energy to that of the Cp*2TiX complexes, it
is important to note that the strength of the agostic interaction
in a titanium(III) complex is expected to be weaker than that
for an analogous titanium(IV) complex since titanium(IV)
complex is more electrophilic or Lewis acidic and is expected
to form a stronger agostic bond. In addition, the Cp*2Ti
environment is more sterically demanding than either of these
two examples (the Cy3P ligand is large, but its bulk is well
away from the tungsten center).
Figure 10. Observed gave versus temperature for Cp*2TiEt (a) and
Cp*2TiN(Me)Ph (b). The different symbols are for independent data
runs, and the line is a least-squares fit to eq 4.
The ganagostic values for Cp*2TiEt and Cp*2TiN(Me)Ph are
very similar to the gave values of Cp*2TiMe (1.958) and
Cp*2TiN(Et)Ph (1.955) as expected. The ganagostic values can
be used to estimate the gy value for the anagostic form of the
molecules by assuming that gz and gx are the same in both
conformations. For Cp*2TiEt and Cp*2TiN(Me)Ph, the gy
values for the anagostic conformation are 1.890 and 1.884,
respectively. If the optical spectra are known for the agostic
and anagostic conformers, the change in the 1a1 f b2 energy
between them can be estimated. This energy gives the electronic
contribution to the agostic bond.
The fact that the agostic interaction in Cp*2TiEt is so weak
explains why the other alkyl complexes of Cp*2Ti do not form
â-agostic bonds. As shown in Figure 11, the substituent on
the â-carbon atom will have an unfavorable steric interaction
with the Cp* ligand. If the steric repulsions in the â-agostic
conformations of other alkyl groups are more than 2 kcal/mol
greater than in Cp*2TiEt, the agostic conformation will be
unfavorable relative to the anagostic conformation. The lack
of R- or γ-agostic interactions can be rationalized in a similar
manner.
From the values of ∆H° and ∆S°, at 20 °C the constants for
the equilibrium in Figure 9 for Cp*2TiEt and Cp*2TiN(Me)Ph
are 0.87 and 4.1, respectively. The equilibrium constants help
to explain why no 1a1 f b2 transition is observed for Cp*2TiEt
or Cp*2TIN(Me)Ph at room temperature. For Cp*2TiN(Me)-
Ph, most of the molecules have no agostic interaction and for
Cp*2TiEt approximately 45% have no agostic interaction.
While the transferability of the solution data to a frozen glass
is uncertain, at 77 K, the equilibrium constants for Cp*2TiEt
and Cp*2TiN(Me)Ph are 8 × 10-5 and 3 × 10-3, respectively.
Spectra acquired at that temperature are expected to be due only
to the agostic species. The spectra of Cp*2TiEt and Cp*2TiN-
(Me)Ph in toluene-d8 at 77 K are shown in Figure 6. Unfor-
tunately, the 1a1 f b2 transition for Cp*2TiN(Me)Ph was not
observed due to the presence of C-H or C-D stretching
overtones from the solvent.
In conclusion, electronic and EPR spectroscopy can be used
to rank the X ligands of Cp*2TiX in terms of π-donating ability.
The π-donor series is similar to that seen in other systems. No
evidence is seen for R- or γ-agostic interactions in the trivalent
decamethyltitanocene alkyl complexes, but â-agostic interactions
can be observed for Cp*2TiEt and Cp*2TiN(Me)Ph. The
equilibrium between the agostic and anagostic conformations
of these molecules can be examined using variable-temperature
EPR spectroscopy, and the equilibrium constant is found to be
0.9 and 4 at room temperature for Cp*2TiEt and Cp*2TiN(Me)-
Ph, respectively, and ca. 10-4 at 77 K.
As noted earlier, the combined electronic and EPR data can
be used to estimate the electronic contribution to the agostic
bond. For Cp*2TiEt, b2 is destabilized by 18 kcal/mol in the
agostic conformation, but some of the destabilization is likely
the result of the σ-bond of the ethyl ligand moving off of the
x-axis. This is shown by the observation that at 77 K, the 1a1
f 2a1 transition is 1000 cm-1 lower in energy than at room
temperature and 2000 cm-1 lower than in Cp*2TiMe indicating
that the ethyl group is interacting more weakly with the 2a1 as
(35) Gonzales, A. A.; Zhang, K.; Nolan, S. P.; Lopez de la Vega, R. L.;
Murkerjee, S. L.; Hoff, C. D.; Kubas, G. J. Organometallics 1988, 12,
2429-2435.
(36) Dawoodi, Z.; Green, M. L. H.; Mtetwa, V. S. B.; Prout, K. J. Chem.
Soc., Chem. Commun. 1982, 12, 802-803.
(37) Munakata, H.; Ebisawa, Y.; Takashima, Y.; Wrinn, M. C.; Sheiner,
A. C.; Newsam, J. M. Catal. Today 1995, 23, 403-408.

1726 J. Am. Chem. Soc., Vol. 118, No. 7, 1996
Lukens et al.
38
Cp*2TiF2. Cp*2TiMe2 (1.15 g, 3.30 mmol) was dissolved in 70
mL of diethyl ether, and BF3‚OEt2 (0.96 g, 6.8 mmol) was added slowly
using a syringe. The yellow solution became orange. After the mixture
was stirred for 12 h, the solvent was removed under reduced pressure
and the residue was heated at 90 °C under dynamic vacuum for 4 h to
remove MeBF2. The yellow solid residue was dissolved in 100 mL of
hexane and the solution was filtered. The volume of the filtrate was
reduced to ca. 20 mL. Cooling to -80 °C gave orange needles (1.1 g,
93%). Mp ) 207-208 °C. 1H NMR δ 1.82 (s). IR 2720 (w), 1165
(w), 1065 (w), 1020 (m), 810 (w), 725 (w), 635 (w), 610 (m), 580 (s),
565 (s), 545 (m), 440 (s), 390 (m) cm-1. MS (M)+ m/z (calc, found)
354 (10, 11), 355 (12, 12), 356 (100, 100), 357 (30, 29), 358 (11, 11),
359 (2, 2). Anal. Calcd for C20H30TiF2: C, 67.4; H, 8.48. Found:
C, 67.4; H, 8.59.
Experimental Section
All reactions and manipulations were carried out in an inert
atmosphere using standard Schlenk and drybox techniques. Ammonia
and methylamine were dried over sodium at -78 °C and distilled
immediately prior to use. The lithium salts of the amides were prepared
by the addition of the amine to a solution of n-butyllithium in hexane.
The lithium salts of the alkoxides were prepared by treating lithium
metal with the alcohol in hexane. Cp*2TiBr, Cp*2TiI, Cp*2TiCl;16 Cp*2-
TiCH2C6H5, Cp*2TiCH2CMe3, Cp*2Ti(n-Pr), Cp*2TiEt, Cp*2TiMe,
Cp*2TiH;15 Cp*2TiN(Me)(Ph);22 Cp*2TiMe2;38 and KC818 were prepared
by literature methods. Infrared spectra were recorded on a Nicolet 5DX
FTIR spectrometer or a Perkin-Elmer 283 spectrometer as Nujol mulls
between CsI or KBr plates. 1H NMR spectra were measured on a JEOL
FX90Q FT NMR spectrometer operating at 89.56 MHz or on a home-
built FT NMR spectrometer (Cryomagnet Systems 5.9 T magnet
interfaced to a Nicolet 1180 computer) operating at 250.80 MHz at
the Berkeley Department of Chemistry NMR facility. Chemical shifts
were referenced to tetramethylsilane (δ ) 0) with positive values at
lower field. Melting points were measured on a Thomas-Hoover
melting point apparatus in sealed capillaries and are uncorrected. EPR
spectra were measured as solutions or frozen glasses in either
methylcyclohexane or 2-methyltetrahydrofuran using a Varian E-12
spectrometer. For the frozen glass spectra, an Oxford Instruments ESR-
10 liquid helium cryostat was used. For the variable-temperature
studies, a Wilmad Version 1 variable-temperature apparatus was used;
the temperature was determined using a thermocouple referenced to a
0 °C ice bath. The microwave frequency was measure using an EIP-
548 microwave frequency counter and the magnetic field was measure
using a Varian E-500 NMR gaussmeter. Spectra were digitized using
UNPLOTIT. EPR simulations were done using the program ABVG.39
Optical spectra were recorded as 10-3 to 10-2 M solutions in
methylcyclohexane using matched quartz cells and a Cary 17 spectro-
photometer controlled by a PC. For the low-temperature spectra a
quartz dewar of liquid nitrogen was used, and the samples were
contained in an EPR tube surrounded by a copper mask with a 2 mm
wide by 7 mm high window. Spectra were fit as sums of Gaussian
curves using the program Horizon.40 Electron impact mass spectra were
recorded by the mass spectroscopy laboratory at the University of
California, Berkeley. Elemental analyses were performed by the
analytical laboratories at the University of California, Berkeley.
Cp*2TiF. A slurry a KC8 (0.21 g, 2.0 mmol) in 20 mL of
tetrahydrofuran was added by cannula to a solution of Cp*2TiF2 (0.67
g, 1.9 mmol) dissolved in 10 mL of tetrahydrofuran. The solution
immediately turned dark green. After the mixture was stirred for 3 h,
the solvent was removed under reduced pressure. The solid residue
was suspended in 100 mL of hexane. The dark green suspension was
filtered, and the volume of the filtrate was reduced to ca. 10 mL.
Cooling to -20 °C gave dark green crystals (0.40 g, 63%). Mp )
201-203 °C. IR 2720 (w), 1165 (w), 1065 (w), 1025 (m), 805 (w),
725 (w), 635 (w), 610 (w), 570 (s), 450 (s), 415 (w), 395 (w) cm-1
.
MS (M)+ m/z (calc, found) 335 (11, 16), 336 (12, 18), 337 (100, 100),
338 (30, 31), 339 (11, 12). Anal. Calcd for C20H30TiF: C, 71.1; H,
8.96. Found: C, 70.9; H, 8.94.
Cp*2TiOCH3. A mixture of Cp*2TiCl16 (0.50 g, 1.4 mmol) and
LiOCH3 (0.06 g, 1.6 mmol) was suspended in 40 mL of tetrahydrofuran.
The solution was warmed to 70 °C for 3 h during which time the
solution turned red-orange. The suspension was allowed to cool to
room temperature. After the mixture was stirred for 12 h, the solvent
was removed under reduced pressure, and the solid residue was
suspended in 50 mL of hexane. The red-purple suspension was filtered,
and the volume of the filtrate was reduced to ca. 1 mL. Cooling to
-20 °C produced brown plates (0.33 g, 67%). Mp ) 135-150 °C.
IR 2790 (s), 2720 (w), 1270 (w), 1150 (s), 1075 (m), 1025 (m), 800
(w), 760 (m), 725 (w), 660 (w), 620 (w), 550 (m), 500 (m), 420 (m)
cm-1. MS (M)+ m/z (calc, found) 347 (11, 4), 348 (12, 5), 349 (100,
100), 350 (31, 12), 351 (12, 5). Anal. Calcd for C21H33OTi: C, 72.2;
H, 9.52. Found: C, 71.7; H, 9.65.
Cp*2TiNH2. Approximately 300 mL of gaseous NH3 at room
temperature (ca. 13 mmol) was vacuum transferred to a solution of
Cp*2TiMe (1.20 g, 3.60 mmol) in hexane at -78 °C. The solution
was allowed to warm to room temperature and the evolved gasses were
periodically vented. After gas evolution had ceased, the solution was
filtered and the volume of the filtrate was reduced to ca. 5 mL. Cooling
at -80 °C gave dark crystals (0.70 g, 58%). It was necessary to dry
the ammonia over sodium for at least 1 h at -78 °C to obtain
spectroscopically and analytically pure material. Mp ) 193-196 °C.
IR: 3437 (m), 2721 (m), 1535 (s), 1491 (s), 1023 (s), 802 (w), 634
(s), 626 (m), 616 (s), 598 (s), 486 (s), 431 (s), 395 (m) cm-1. MS
(M)+ m/z (found) 352 (100), 353 (30). Anal. Calcd for C20H32NTi:
C, 71.8; H, 9.65; N, 4.20. Found: C, 71.7; H, 9.74; N, 4.14.
Cp*2TiN(Me)H. A mixture of Cp*2TiCl16 (0.71 g, 2.0 mmol) and
LiN(Me)H (0.09 g, 2.4 mmol) was suspended in 30 mL of diethyl ether.
After the mixture was stirred for 12 h, the solvent was removed under
reduced pressure, and the solid was suspended in 50 mL of hexane
forming a lilac-colored solution. The solution was filtered, and the
volume of the filtrate was reduced to ca. 5 mL. Cooling to -20 °C
gave green crystals (0.32 g, 46%). Mp ) 202-205 °C. IR 3360 (w),
2765 (m), 2725 (w), 1405 (m), 1160 (w), 1083 (s), 1037 (m), 1010
(m), 790 (w), 711 (w), 617 (w), 535 (m), 494 (s), 419 (s), 378 (m)
cm-1. MS (M)+ m/z (calc, found) 347 (12, 44), 348 (100, 100), 349
(31, 32), 350 (12, 11). Anal. Calcd for C21H34NTi: C, 72.4; H, 9.84;
N, 4.02. Found: C, 73.0; H, 9.89; N, 4.04.
Cp*2TiOC6H5. A mixture of Cp*2TiCl16 (0.50 g, 1.4 mmol) and
LiOC6H5 (0.16 g, 1.6 mmol) was dissolved in 30 mL of tetrahydrofuran.
The solution immediately became purple-red. After the mixture was
stirred for 10 h, the solvent was removed under reduced pressure, and
the residue was suspended in 50 mL of hexane. The purple suspension
was filtered, and the volume of the filtrate was reduced to ca. 10 mL.
Cooling to -20 °C gave big purple-brown crystals (0.48 g, 83%). Mp
) 202-207 °C. IR: 2720 (w), 2610 (w), 1615 (w), 1585 (s), 1565
(m), 1485 (s), 1310 (s), 1160 (s), 1065 (w), 1020 (w), 995 (m), 880
(s), 750 (s), 695 (m), 630 (w), 620 (m), 605 (w), 520 (w), 430 (m),
405 (w), 360 (m) cm-1. MS (M)+ m/z (calc, found) 409 (10, 20), 410
(13, 22), 411 (100, 100), 412 (36, 48), 413 (13, 17), 414 (3, 4). Anal.
Calcd for C26H35OTi: C, 75.9; H, 8.57. Found: C, 76.3; H, 8.59.
Cp*2TiN(Et)Ph. A mixture of Cp*2TiCl16 (0.50 g, 1.4 mmol) and
LiN(Et)Ph (0.20 g, 1.6 mmol) were suspended in 40 mL of diethyl
ether. After the mixture was stirred for 3 h, the solvent was removed
under reduced pressure. The black solid residue was suspended in 30
mL of hexane, and the mixture was filtered. The volume of the filtrate
was reduced to ca. 3 mL. Cooling to -20 °C gave small black crystals
(0.36 g, 58%). Mp ) 174-181 °C. IR 2720 (w), 1370 (s), 1355 (w),
1340 (w), 1305 (s), 1280 (s), 1185 (m), 1135 (w), 1090 (m), 1020 (s),
980 (s), 850 (m), 775 (s), 745 (s), 695 (s), 635 (m), 535 (m), 485 (w),
450 (s), 420 (m), 405 (s), 370 (m), 345 (s) cm-1. MS (M)+ m/z (calc,
found) 437 (13, 45), 438 (100, 100), 439 (39, 37), 440 (14, 13), 441
(2, 3). Anal. Calcd for C28H40NTi: C, 76.7; H, 9.19; N, 3.19.
Found: C, 7.60; H, 9.28; N, 3.36.
(38) Bercaw, J. E.; Marvich, R. H.; Bell, L. G.; Brintzinger, H. H. J.
Am. Chem. Soc. 1972, 94, 1219.
(39) Daul, C.; Schlapfer, C. W.; Mohos, B.; Ammeter, J.; Gamp, E.
Comp. Phys. Commun. 1981, 21, 385-395.
(40) Guenther, D.; Jaffe, A. Horizon; 1.3.0 ed.; Blue Star Software,
Inc.: Honolulu, 1992.
Cp*2TiN(Me)Ph.22 IR (not previously reported) 3075 (w), 3055
(w), 2720 (w), 2620 (w), 2570 (w), 1585 (s), 1555 (m), 1390 (s), 1190
(m), 1165 (m), 1050 (w), 1030 (m), 990 (s), 855 (w), 825 (s), 755 (s),
705 (m), 630 (w), 545 (w), 470 (w), 420 (m), 350 (m) cm-1
.

A π-Donor Spectrochemical Series in (Me5C5)2 TiX
J. Am. Chem. Soc., Vol. 118, No. 7, 1996 1727
X-ray Crystal Structure Determination for Cp*2TiF. Dark green
crystals of the fluoride were grown by cooling a saturated hexane
solution to -30 °C. A large, blocky single crystal was selected and a
block-shaped piece measuring 0.30 × 0.40 × 0.45 mm was cut from
one corner. The crystal was mounted on the end of a 0.4 mm diameter
quartz capillary with a drop of Paratone N. The crystal was transferred
to an Enraf-Nonius CAD-4 diffractometer41 and cooled to -115 °C
under a cold stream of nitrogen gas previously calibrated by a
thermocouple placed in the sample position. Automatic peak search
and indexing procedures indicated that the crystal possessed a primitive
monoclinic cell and yielded the cell parameters. The cell parameters
and data collection parameters are given in the supporting information.
hydrogen scattering factors were corrected for both the real and
imaginary components of anomalous dispersion.45
Inspection of the residuals ordered in the ranges of sin(θ/λ), |Fo|,
and parity and values of the individual indexes showed no trends. Five
reflections had anomalously high values of wδ2, and were weighted to
zero toward the end of the refinement. The largest positive and negative
peaks in the final difference Fourier map have electron densities of
0.66 and -0.17 e-/Å3.
X-ray Crystal Structure Determination for Cp*2TiN(Me)H.
Rose-colored crystals of the amide were grown by cooling a saturated
hexane solution to -30 °C. A blade-shaped single crystal was selected
and a roughly pyramidal chunk measuring 0.37 × 0.40 × 0.40 mm
was cut from the tip. The crystal was mounted on the end of a 0.3
mm diameter glass capillary with a drop of Paratone N. The crystal
was transferred to an Enraf-Nonius CAD-4 diffractometer41 and cooled
to -130 °C under a cold stream of nitrogen gas previously calibrated
by a thermocouple placed in the sample position. The crystal was
centered in the beam. Automatic peak search and indexing procedures
indicated that the crystal possessed a primitive orthorhombic cell and
yielded the cell parameters. The cell parameters and data collection
parameters are given in the supporting information.
The 5178 raw intensity data were converted to structure factor
amplitudes and their esds by correction for scan speed, background,
and Lorentz-polarization effects.41,42 Inspection of the intensity
standards showed a sudden intensity loss of 12% between hours 6 and
7. The data collected after hour 7 was corrected for a 12% loss in
intensity. The 226 systematic absences (h,0,l) [l odd], (0,k,0) [k odd],
the 191 redundant data (0,k,l) [l < 0], and the 225 data collected
between hours 6 and 7 were then rejected yielding 4536 unique data
of which 3188 possessed Fo > 3σ(Fo). Azimuthal scan data showed a
difference of Imin/Imax ) 0.81; however, the absorption curves were
asymmetric. No empirical absorption correction was applied. The
systematic absences indicated that the space group was P21/c.
The 1514 raw intensity data were converted to structure factor
amplitudes and their esds by correction for scan speed, background,
and Lorentz-polarization effects.41,42 Inspection of the intensity
standards showed a smooth, slightly curved decay of 14% over the
data collection. The data were corrected for a linear decay of 14%.
The 20 systematic absences (h,0,0) [h odd], (0,k,0) [k odd], and (0,0,l)
[l odd] were then rejected yielding 1494 unique data of which 1323
possessed Fo > 3σ(Fo). Azimuthal scan data showed a difference of
Imin/Imax ) 0.85. An empirical absorption correction was applied. The
systematic absences indicated that the space group was P212121.
The cell volume indicated that 8 molecules were present in the unit
cell. The titanium atom positions for the two independent molecules
were obtained by solving the Patterson map with the program
SHELXS86.43 The remaining heavy atom positions were obtained by
successive Fourier searches and cycles of refinement. The heavy atom
structure was refined by standard least-squares techniques. The heavy
atoms were refined isotropically, and the hydrogen positions were then
calculated based upon idealized bonding geometry and assigned thermal
parameters equal to 1.3 Å2 larger than the carbon atom to which they
were connected. A Gaussian absorption correction, DIFABS, was then
used. The heavy atoms were then refined anisotropically. The fluorine
atoms of both molecules appear to be either moving or disordered. A
final difference Fourier map showed no additional atoms in the
asymmetric unit. Examination of intermolecular close contacts (<3.5
Å) showed that the molecules were monomeric.
The cell volume indicated that 4 molecules were present in the unit
cell. The titanium atom position was obtained by solving the Patterson
map. Refinement on the titanium position led to the titanium becoming
non-positive definite. However, a difference Fourier search yielded
most of the other heavy atom positions. Refinement of these positions
followed by a Fourier search yielded the remaining positions. The
heavy atom structure was refined by standard least-squares and Fourier
techniques. The heavy atoms were refined anisotropically, and a
difference Fourier search showed that almost all of the hydrogen atoms
could be located. The amide hydrogen was included in the refinement
with an isotropic thermal parameter and it behaved normally. The other
hydrogen positions were then calculated based upon idealized bonding
geometry and assigned thermal parameters equal to 1.3 Å2 larger than
the carbon atom to which they were connected. The non-amide
hydrogen positions were included in the structure factor calculations
but not refined by least squares. At the end of the refinement, the
enantiomer was changed, and the structure refined again. The
refinement was very slightly worse, so the enantiomer was changed
back to the original one, and the structure was rerefined. A final
difference Fourier map showed no additional atoms in the asymmetric
unit. Examination of intermolecular close contacts (<3.5 Å) showed
that the molecule was a monomer.
The final residuals for 397 variables refined against the 3188 unique
data with Fo > 3σ(Fo) were R ) 6.90%; Rw ) 9.11%, and GOF )
2.054. The R value for all data (including unobserved reflections) was
9.91%. The quantity minimized by the least-squares refinements was
w(|Fo| - |Fc|)2, where w is the weight given to a particular reflection.
The p-factor,44 used to reduce the weight of intense reflections, was
set to 0.03 initially, but later changed to 0.07. The analytical form of
the scattering factor tables for neutral atoms was used and all non-
(41) Calculations were performed on a DEC Microvax II using locally
modified Molen software operating under the Micro-VMS operating system.
(42) The data reduction formulas are
ω
Lp
2
ω
Lp
Fo
)
(C - 2B)
σo(Fo ) ) (C + 4B)1/2
2
2
Fo (Fo )1/2
ωo(F) ) [Fo2 + σo(Fo )]1/2 - Fo
2
The final residuals for 212 variables refined against the 1315 unique
data with Fo > 3σ(Fo) were R ) 4.83%, Rw ) 6.22%, and GOF )
2.04. The R value for all data (including unobserved reflections) was
5.58%. The quantity minimized by the least-squares refinements was
w(|Fo| - |Fc|)2, where w is the weight given to a particular reflection.
The p-factor, used to reduce the weight of intense reflections, was set
to 0.03 initially, but later changed to 0.05.44 The analytical form of
the scattering factor tables for neutral atoms were used and all non-
hydrogen scattering factors were corrected for both the real and
imaginary components of anomalous dispersion.45
where C is the total count of the scan, B is the sum of the two background
counts, ω is the scan speed used in deg/min, and
sin (2θ)(1 + cos2(2θm))
1 + cos2(2θm) - sin2(2θ)
1
Lp
)
is the correction for Lorentz and polarization effects for a reflection with
scattering angle 2θ and radiation monochromatized with a 50% perfect single
crystal monochrometer with scattering angle 2θm.
(43) Sheldrick, G. M. In Crystallographic Computing 3; Sheldrick, G.
M., Goddard, R., Eds.; Oxford University Press: London, 1985; pp 175-
189.
Inspection of the residuals ordered in the ranges of sin(θ/λ), |Fo|,
and parity and values of the individual indexes showed no trends. Eight
reflections had anomalously high values of w∆2, and were weighted to
zero toward the end of the refinement. The largest positive and negative
(44) R ) (∑||Fo| - |Fc||)/∑|Fo|, wR ) [∑(|Fo| - |Fc|)2/∑wFo ]1/2, and
2
GOF ) [∑(|Fo| - |Fc|)2/(no - nv)]1/2, where no is the number of observations
and nv is the number of variable parameters, and the weights were given
by w ) 1/σ2(Fo) and σ(Fo ) ) [σo (Fo ) + (pF2)2]1/2, where σ2(Fo) is
2
2
2
2
calculated as above from σ(Fo ) and where p is the factor use to lower the
weight of intense reflections.
(45) Cromer, D. T.; Waber, J. T. In International Tables for X-Ray
Crystallography; Kynoch Press: Birmingham, England, 1974; Vol. IV.

1728 J. Am. Chem. Soc., Vol. 118, No. 7, 1996
Lukens et al.
peaks in the final difference Fourier map have electron densities of
0.43 and -0.58.
Supporting Information Available: Tables of bond dis-
tances and angles, tables of positional parameters, and tables
of thermal parameters for Cp*2TiN(H)Me and Cp*2TiF and a
plot of the electronic spectra of the trivalent decamethylti-
tanocenes (17 pages). This material is contained in many
libraries on microfiche, immediately follows this article in the
microfilm version of the journal, can be ordered from the ACS,
and can be downloaded from the Internet; see any current
masthead page for ordering information and Internet access
instructions. Tables of structure factors are available from the
author.
Acknowledgment. The authors would like to thank Dr. F.
J. Hollander for assistance in the crystal structure analysis and
Dr. Norman Edelstein for helpful discussion about EPR
spectroscopy. W.L. would like to thank the National Science
Foundation for a graduate fellowship. M.S. thanks the Miller
Institute for a postdoctoral fellowship. This work was partially
supported by the Director, Office of Energy Research, Office
of Basic Energy Sciences, Chemical Sciences Division of the
U.S. Department of Energy under Contract No. DE-AC03-
76SF00098.
JA953272O