Titanocene borane σ-complexes

Article abstract of DOI:10.1021/ja9900757

The chemistry of titanocene bisborane complexes Cp₂Ti(HBcat')₂ (1a-g) (HBcat' = catecholborane or a substitued catecholborane) and monoborane complexes Cp₂Ti(HBcat')(L) (2-4) (L = PMe₃, PhSiH₃, or PhCCPh) is reported. These complexes are unusual σ-complexes. The B-H bond in the catecholborane of 1 acts as a two-electron-donor ligand. The 4-tert-butyl version 1a was studied in depth and underwent ligand substitution reactions with PMe₃, CO, PhSiH₃, and PhCCPh. The products of the reaction of 1a with PMe₃ and PhSiH₃ are the novel monoborane σ-complexes Cp₂Ti(HBcat')(PMe₃) (2a; HBcat' = HBO₂C₆H₃-4-t-Bu) and Cp₂Ti(HBcat')(PhSiH₃) (3; HBcat' = HBO₂C₆H₃-4-t-Bu), in which the catecholborane remains a two-electron- donating ligand. Reaction with CO formed Cp₂Ti(CO)₂. Reaction with PhCCPh formed Cp₂Ti(HBcat')-(PhCCPh) (4; HBcat' = HBO₂C₆H₃-4-t-Bu), which was observed in solution and reductively eliminated the vinyl boronate ester (Ph)(Bcat')C=C(Ph)(H). The rates for the reactions of 1a with these substrates showed a first-order dependence on the concentration of 1a and a zero-order dependence on the concentrations of both the departing HBcat' and the incoming ligand. The substitution reaction proceeded at the same rate ((3.8 ± 0.3) x 10^-4) regardless of the identity of the incoming ligand. The entropy of activation was +30 ± 5 eu. These data are consistent with a dissociative substitution mechanism for the reaction of 1a with these substrates. The ΔH(+) value of 25 ± 3 kcal mol^-1 for these reactions provides an upper limit for the strength of the borane-metal interaction. Electronic effects on the reaction rate support a bonding model involving back-donation from titanium to the borane, and the unusual steric effects allow a proposal for the geometric changes that occur upon formation of the transition state.

Full text of DOI:10.1021/ja9900757

J. Am. Chem. Soc. 1999, 121, 5033-5046
5033
Titanocene Borane σ-Complexes
Clare N. Muhoro, Xiaoming He, and John F. Hartwig*
Contribution from the Department of Chemistry, Yale UniVersity, P.O. Box 208107,
New HaVen, Connecticut 06520-8107
ReceiVed January 8, 1999
Abstract: The chemistry of titanocene bisborane complexes Cp2Ti(HBcat′)2 (1a-g) (HBcat′ ) catecholborane
or a substitued catecholborane) and monoborane complexes Cp2Ti(HBcat′)(L) (2-4) (L ) PMe3, PhSiH3, or
PhCCPh) is reported. These complexes are unusual σ-complexes. The B-H bond in the catecholborane of 1
acts as a two-electron-donor ligand. The 4-tert-butyl version 1a was studied in depth and underwent ligand
substitution reactions with PMe3, CO, PhSiH3, and PhCCPh. The products of the reaction of 1a with PMe3
and PhSiH3 are the novel monoborane σ-complexes Cp2Ti(HBcat′)(PMe3) (2a; HBcat′ ) HBO2C6H3-4-t-Bu)
and Cp2Ti(HBcat′)(PhSiH3) (3; HBcat′ ) HBO2C6H3-4-t-Bu), in which the catecholborane remains a two-
electron-donating ligand. Reaction with CO formed Cp2Ti(CO)2. Reaction with PhCCPh formed Cp2Ti(HBcat′)-
(PhCCPh) (4; HBcat′ ) HBO2C6H3-4-t-Bu), which was observed in solution and reductively eliminated the
vinyl boronate ester (Ph)(Bcat′)CdC(Ph)(H). The rates for the reactions of 1a with these substrates showed a
first-order dependence on the concentration of 1a and a zero-order dependence on the concentrations of both
the departing HBcat′ and the incoming ligand. The substitution reaction proceeded at the same rate ((3.8 (
0.3) × 10-4) regardless of the identity of the incoming ligand. The entropy of activation was +30 ( 5 eu.
These data are consistent with a dissociative substitution mechanism for the reaction of 1a with these substrates.
The ∆Hq value of 25 ( 3 kcal mol-1 for these reactions provides an upper limit for the strength of the borane-
metal interaction. Electronic effects on the reaction rate support a bonding model involving back-donation
from titanium to the borane, and the unusual steric effects allow a proposal for the geometric changes that
occur upon formation of the transition state.
Introduction
variants of alkane σ-complexes are commonly referred to as
agostic complexes. These complexes are generally more stable
because the X-H donor possesses another point of attachment
to the metal, and many examples of agostic complexes are
known.14
σ-Complexes have been invoked as intermediates in the
oxidative addition of X-H bonds to transition metals. These
addition reactions are important because they provide the initial
step for activation of unreactive substrates. This enhancement
of reactivity by coordination of the X-H σ-bond is an important
step in catalytic hydrogenation15,16 and hydrosilation16-18 and
for stoichiometric processes such as C-H activation19-23 and
σ-bond metatheses.24,25 R-Agostic interactions in metal alkyl
Transition metal σ-complexes are a remarkable type of
coordination complex created by intermolecular binding of a
substrate to a metal center through an X-H σ-bond. Stable
σ-complexes are limited to examples where X is H or Si, and
numerous examples of dihydrogen and silane complexes have
been reported.1-3 Transition metal alkane σ-complexes are
highly unstable,4 but have been detected in the gas phase,5 in
matrix isolation,6,7 in solution by direct NMR spectroscopic
observation of continuously irradiated solutions8 or by indirect
mechanistic probes such as isotopic labeling and kinetic isotope
effects,9-12 and in the solid state when supported by ac-
companying hydrophobic interactions.13 The intramolecular
(14) Brookhart, M.; Green, M. L. H.; Wong, L.-L. Prog. Inorg. Chem.
1988, 36, 1-124.
(1) Crabtree, R. H. Angew. Chem., Int. Ed. Engl. 1993, 32, 789.
(2) Heinekey, D. M.; Oldham, W. J. Chem. ReV. 1993, 93, 913.
(3) Schubert, U. AdV. Organomet. Chem. 1990, 30, 151.
(4) For a review of transition metal alkane complexes see: Hall, C.;
Perutz, R. N. Chem. ReV. 1996, 96, 3125.
(15) Halpern, J. Science 1982, 217, 401.
(16) Collman, J. P.; Hegedus, L. S.; Norton, J. R.; Finke, R. G. Principles
and Applications of Organotransition Metal Chemistry, 2nd ed.; University
Science Books: Mill Valley, CA, 1987.
(17) Chalk, A. J.; Harrod, J. F. J. Am. Chem. Soc. 1965, 87, 16.
(18) Sommer, L. H.; Lyons, J. E.; Fujimoto, H. J. Am. Chem. Soc. 1969,
91, 2670.
(19) Crabtree, R. H. Chem. ReV. 1995, 95, 987.
(20) Hill, C. L. ActiVation and Functionalization of Alkanes; John Wiley
and Sons: New York, 1989.
(5) Brown, C. E.; Ishikawa, Y.-I.; Hackett, P. A.; Rayner, D. M. J. Am.
Chem. Soc. 1990, 112, 2530.
(6) Turner, J. J.; Burdett, J. K.; Perutz, R. N.; Poliakoff, M. Pure Appl.
Chem. 1977, 49, 271.
(7) Rest, A. J.; Whitwell, I.; Graham, W. A. G.; Hoyano, J. K.; McMaster,
A. D. J. Chem. Soc. 1987, 1181.
(8) Geftakis, S.; Ball, G. E. J. Am. Chem. Soc. 1998, 120, 9953.
(9) Buchanan, J. M.; Stryker, J. M.; Bergman, R. G. J. Am. Chem. Soc.
1986, 108, 1537-1550.
(10) Gould, G. L.; Heinekey, D. M. J. Am. Chem. Soc. 1989, 111, 5502.
(11) Bullock, R. M.; Headford, C. E. L.; Hennessy, K. M.; Kegley, S.
E.; Norton, J. R. J. Am. Chem. Soc. 1989, 111, 3897.
(12) Bullock, R. M. In Transition Metal Hydrides; Dedieu, A., Ed.;
VCH: New York, 1992; p 263.
(21) Arndtsen, B. A.; Bergman, R. G.; Mobley, T. A.; Petersen, T. H.
Acc. Chem. Res. 1995, 28, 154.
(22) Shilov, A. E.; Shul′pin, G. B. Chem. ReV. 1997, 97, 2879.
(23) For a discussion on experiments using isotope effects to determine
the intermediacy of alkane complexes in C-H activation see: Bullock, R.
M. In Transition Metal Hydrides; Dedieu, A., Ed.; VCH: New York, 1992;
pp 263-307.
(24) Thompson, M. E.; Baxter, S. M.; Bulls, A. R.; Burger, B. J.; Nolan,
M. C.; Santarsiero, B. D.; Schaefer, W. P.; Bercaw, J. E. J. Am. Chem.
Soc. 1987, 109, 203.
(13) Evans, D. R.; Drovetskaya, T.; Bau, R.; Reed, C. A.; Boyd, P. D.
W. J. Am. Chem. Soc. 1997, 119, 3633.
10.1021/ja9900757 CCC: $18.00 © 1999 American Chemical Society
Published on Web 05/15/1999

5034 J. Am. Chem. Soc., Vol. 121, No. 21, 1999
Muhoro et al.
complexes of coordinatively unsaturated electrophilic metal
centers have also been used to explain stereochemical and rate
data for olefin insertion reactions during polymerization of
alkenes.26-31
We previously reported that titanocene dimethyl is an efficient
catalyst for the hydroboration of alkenes with catecholborane.32
This hydroboration was likely to proceed by a mechanism that
is distinct from that of late transition metal hydroboration
catalysts.33-38 Our initial investigations focused on the identi-
fication of the catalyst and revealed that the reaction of the
titanocene dimethyl with catecholborane was the activating step
since no reaction was observed between titanocene dimethyl
and alkene at room temperature on the time scale of the catalytic
process. These studies led to the discovery of a new class of
σ-complex with unusual coordination of two catecholborane
molecules to titanium through borane B-H bonds.39,40 The
reaction chemistry of these novel bisborane complexes initiated
fundamental studies of titanium borane complexes that contain
this new type of metal-ligand interaction. We report here the
synthesis of bisborane complex Cp2Ti(HBcat-4-t-Bu)2 (1a) and
its reactions with PMe3, PhSiH3, and PhCCPh that give
complexes of the general formula Cp2Ti(HBcat-4-t-Bu)(L)
where L ) PMe3, PhSiH3, and PhCCPh. This set of compounds
raises questions concerning the borane bonding mode, factors
controlling the stability of the compounds, and the mechanism
of borane substitution that are also addressed in detail here.
Figure 1. ORTEP of 1g.
of titanocene dimethyl and 3 equiv of 4-tert-butylcatecholborane
in pentane at -30 °C for 8 h resulted in the precipitation of
Cp2Ti(HBcat′)2 (1a) as a yellow solid in 70% yield. Methane
and B-methyl-4-tert-butylcatecholborane were also generated
1
in this reaction as determined by H NMR spectroscopy. At
room temperature, compound 1a is stable enough to be handled
as a solid, but is unstable in solution. Complex 1a was used as
a precursor to new titanocene monoborane complexes as
described in the next section. Five other analogues (1b-f) of
complex 1a were prepared in 11-83% yields by identical
procedures using 3,5-di-tert-butylcatecholborane (1b), 4-meth-
ylcatecholborane (1c), 4-chlorocatecholborane (1d), 4-meth-
ylthiocatecholborane (1e), and 3-fluorocatecholborane (1f)
instead of 4-tert-butylcatecholborane. The σ-complex containing
unsubstituted catecholborane (1g) was also prepared. However,
this analogue was insoluble in aromatic solvents after isolation.
Thus, 1g was characterized by solution NMR spectroscopy
during its formation. Importantly, highly dilute reaction solutions
deposited single crystals of 1g that were suitable for X-ray
diffraction, permitting structural characterization of these new
compounds.
Results
Synthesis of 1. A group of new titanocene σ-complexes with
unusual coordination of neutral catecholborane through the B-H
σ-bond was prepared (eq 1). The syntheses of these complexes
are presented first. Because their spectroscopic data are not
straightforward, spectral information will be presented in a
separate section after the synthetic routes to the different
compounds.
pentane
The connectivity of titanocene bisborane complexes was
revealed by the X-ray crystallographic study of the parent
bisborane complex 1g. An ORTEP drawing is provided in Figure
1. Acquisition parameters for 1g are provided in Table 1, and
selected bond distances and angles are provided in Table 2. The
Ti-B bond length was 2.335(5) Å. This distance is longer than
the bond length in metallocene boryl complexes of tungsten
(2.23 Å),41 tantalum (2.263 Å),42 and niobium (2.295 Å),41
despite the smaller size of titanium, and is indicative of a Ti-B
bond order of less than 1. The B-Ti-B bond angle of 55° was
too small for the complex to be a Ti(IV) bisboryl compound.43
Indeed, the two hydrides were located in the difference map,
One member of this class of σ-complexes is Cp2Ti(HBcat′)2
(1a), within which HBcat′ ) HBO2C6H3-4-t-Bu. The reaction
(25) Zeigler, T.; Folga, E. J. Am. Chem. Soc. 1993, 115, 636.
(26) Brintzinger, H.; Beck, S.; Leclerc, M.; Stehling, U.; Roll, W. Stud.
Surf. Sci. Catal. 1994, 89, 193.
1
and their presence was also determined by H NMR spectros-
(27) Clawson, L.; Soto, J.; Buchwald, S. L.; Steigerwald, M. L.; Grubbs,
R. H. J. Am. Chem. Soc. 1985, 107, 3377.
(28) Cotter, W. D.; Bercaw, J. E. J. Organomet. Chem. 1991, 417, C1.
(29) Grubbs, R. H.; Coates, G. W. Acc. Chem. Res. 1996, 29, 85.
(30) Leclerc, M. K.; Brintzinger, H. H. J. Am. Chem. Soc. 1995, 117,
1651.
(31) Piers, W. E.; Bercaw, J. E. J. Am. Chem. Soc. 1990, 112, 9406.
(32) He, X.; Hartwig, J. F. J. Am. Chem. Soc. 1996, 118, 1696.
(33) Beletskaya, I.; Pelter, A. Tetrahedron 1997, 53, 4957.
(34) Dorigo, A. E.; Schleyer, P. v. R. Angew. Chem., Int. Ed. Engl. 1995,
34, 115.
(35) Westcott, S. A.; Blom, H. P.; Marder, T. B.; Baker, R. T. J. Am.
Chem. Soc. 1992, 114, 8863.
(36) Evans, D. A.; Fu, G.; Anderson, B. A. J. Am. Chem. Soc. 1992,
114, 6679.
(37) Burgess, K.; van der Donk, W. A.; Westcott, S. A.; Marder, T. B.;
Baker, R. T.; Calabrese, J. C. J. Am. Chem. Soc. 1992, 114, 9350.
(38) Burgess, K.; Ohlmeyer, M. J. Chem. ReV. 1991, 91, 1179.
(39) Hartwig, J. F.; Muhoro, C. N.; He, X.; Eisenstein, O.; Bosque, R.;
Maseras, F. J. Am. Chem. Soc. 1996, 118, 10936.
(40) Muhoro, C. N.; Hartwig, J. F. Angew. Chem., Int. Ed. Engl. 1997,
36, 1510.
(41) Hartwig, J. F.; DeGala, S. R. J. Am. Chem. Soc. 1994, 116, 3661.
(42) Lantero, D. R.; Motry, D. H.; Ward, D. L.; Smith, M. R. J. Am.
Chem. Soc. 1994, 116, 10811.
(43) Lauher, J. W.; Hoffmann, R. J. Am. Chem. Soc. 1976, 98, 1729.

Titanocene Borane σ-Complexes
J. Am. Chem. Soc., Vol. 121, No. 21, 1999 5035
Table 2. Selected Bond Angles and Distances in 1g
Table 1. Acquisition Parameters for 1g
empirical formula
formula weight
color of crystal
crystal system
space group
unit cell dimensions
C29H28B2O4Ti
510.06
atom
atom
atom
angle (deg)
bond
distance (Å)
B1
H1
B1
Ti
Ti
Ti
Ti
O1
Cp1
Ti
Ti
Ti
B1
H1
B1
B1
B1
Ti
B1A
H1A
H1
H1
B1
O1
O2
O2
Cp2
53.8(2)
117(2)
32(1)
47(2)
101(2)
125.4(3)
126.0(3)
108.5(4)
145.0(1)
Ti-B
2.335(5)
1.74(4)
1.25(3)
2.11
2.44(1)
2.40(1)
pale yellow
monoclinic
P21/m (no. 11)
a ) 10.116(3) Å
b ) 10.077(2) Å
c ) 12.222(1) Å
2
Ti-H
B-H
B‚‚‚B
Cp1-Ti
Cp2-Ti
â ) 94.3710(10)°
Rw ) 0.064
Z value
goodness of fit
final R indices
1.92
R ) 0.057
Scheme 1
copy (vide infra). The Ti-H distance was 1.74(4) Å, and the
B-H bond length was 1.25(3) Å, typical of bridging titanocene
hydrides.44 These data are consistent with 1g being a Ti(II)
complex in which both hydrogen and boron have a bonding
interaction with the metal center while preserving B-H bond
character.
The most striking feature of this structure is the unusual
geometry for a boron atom that has bonding interactions with
four other atoms. The sum of the three angles at boron that do
not include the hydride is 359.9°. This value indicates that the
titanium, two oxygen atoms, and boron all lie in the same plane,
that the catecholborane is drastically distorted, and that the boron
is far from tetrahedral. Calculations of the geometry of Cp2Ti-
[HB(OH)2]2 by density functional theory provided a similar
unusual geometry.39,40 The hydrides of this model compound
were located in a position similar to those identified in the
Fourier difference map of 1g, and the boron atom in the model
compound also adopted a geometry in which the titanium, two
oxygen atoms, and boron were coplanar.
Syntheses of 2-4: Ligand Substitution Reactions of 1.
The reactivity of 1a included diverse ligand substitution
reactions as shown in Scheme 1. The catecholborane ligand in
1a was displaced by CO, PMe3, PhSiH3, and PhCCPh to form
titanocene dicarbonyl and new compounds 2-4, respectively,
as presented below. The reaction of compound 1a at -30 °C
with CO gave Cp2Ti(CO)2 in 88% yield and HBcat′ in 93%
yield. No intermediate corresponding to the product of substitu-
tion of one HBcat′ ligand, Cp2Ti(CO)(HBcat′), was observed.
Further, there was no reaction between Cp2Ti(CO)2 and 1a to
form Cp2Ti(CO)(HBcat′).
The reaction of compound 1a with PMe3 at -30 °C in toluene
resulted in the formation of 2a in 84% yield. However, attempts
to isolate 2a from this reaction mixture were unsuccessful. The
product generated by this method was short-lived in the reaction
medium at room temperature and decomposed to unidentified
products. Attempts to isolate other analogues of 2a from reaction
mixtures of PMe3 and corresponding complexes 1f,g were also
unsuccessful. Thus, the new route in eq 2 that generates no
toluene at 0 °C generated a maroon solution containing Cp2-
Ti(HBcat)(PMe3) (2b). Cooling a solution of 2b in toluene to
-30 °C led to the precipitation of spectroscopically pure 2b in
46% yield. Complex 2b was unstable in solution at room
temperature, but showed little degradation as a solid at room
temperature over a few hours. The compound was stable as a
solid for several weeks at -30 °C.
The phosphine borane complex 2b, containing an unsubsti-
tuted catecholate substituent, was also generated by displacement
of the coordinated silane in Cp2Ti(H2SiPh2)(PMe3)44 by cat-
echolborane (eq 2). Reaction of Cp2Ti(H2SiPh2)(PMe3) with 1
equiv of catecholborane in the presence of roughly 5 equiv of
excess silane45 formed 2b in 56% yield while leaving 22% of
Cp2Ti(H2SiPh2)(PMe3) unreacted. Reaction of 2b with silane
led to the formation of Cp2Ti(H2SiPh2)(PMe3) in 26% yield
while leaving 55% of 2b unreacted. These two experiments
indicate that an equilibrium between the two complexes exists
and that this equilibrium favors 2b and free silane. Due to
chemistry that competes with establishment of the equilibrium,
we have not found conditions that allow for a quantitative
analysis of this equilibrium.
The connectivity of compound 2a was revealed by X-ray
diffraction studies of the analogue Cp2Ti(HBcat-3-F)(PMe3) (2c)
prepared from the 3-fluoro bisborane complex 1f and Cp2Ti-
(PMe3)2. An ORTEP diagram is provided in Figure 2. Acquisi-
tion parameters for 2c are provided in Table 3, and selected
bond distances and angles are provided in Table 4. The
H-Ti-B bond angle of only 36° was too small to formulate
2c as a Ti(IV) boryl hydride complex.43 Further, the Ti-B bond
length of 2.267(6) Å was, again, comparable to or longer than
those in metallocene boryl complexes of W, Ta, and Nb, despite
reaction byproducts and involves an exact ligand stoichiometry
was devised to allow for isolation of monoborane phosphine
complexes. The reaction of compound 1g and Cp2Ti(PMe3)2 in
(44) Spaltenstein, E.; Palma, P.; Kreutzer, K. A.; Willoughby, C. A.;
Davis, W. M.; Buchwald, S. L. J. Am. Chem. Soc. 1994, 116, 10308.
(45) Because Cp2Ti(H2SiPh2)(PMe3) is unstable in the absence of added
silane, excess silane was added to the sample before addition of borane.

5036 J. Am. Chem. Soc., Vol. 121, No. 21, 1999
Muhoro et al.
Scheme 2
Scheme 3
Figure 2. ORTEP of 2c.
Table 3. Acquisition Parameters for 2c
the trends in bond lengths may be attributed to differences in
electron density at the titanium center in the two complexes.
The basicity of the phosphine ligand would create a more
electron-rich titanium center in complex 2c. As a result, the
structure of 2c would lie further toward oxidative addition of
catecholborane to Ti than does that of 1g.
empirical formula
formula weight
color of crystal
crystal system
space group
unit cell dimensions
C19H23BFO2PTi
392.05
dark red
triclinic
P1h
a ) 8.12190(10) Å
b ) 10.70550(10) Å
c ) 11.66170(10) Å
2
1.024
R ) 0.0638
R ) 112.1880(10)°
â ) 94.3710(10)°
γ ) 98.4890(10)°
Reaction of 1a with phenylsilane in toluene at -5 °C for 5
h resulted in the formation of a yellow-green solution. Evapora-
tion of toluene at -5 °C, followed by washing of the resulting
yellow solid with cold pentane, afforded Cp2Ti(HBcat′)(PhSiH3)
(3; HBcat′ ) HBO2C6H3-4-t-Bu) in 65% yield as a spectro-
scopically pure material. Compound 3 is an unusual example
of a complex ligated by two different types of X-H σ-bonds.
Complex 1a reacted with 2 equiv of diphenylacetylene to
form Cp2Ti(PhCCPh)(HBcat′) (4; HBcat′ ) HBO2C6H3-4-t-Bu)
and the vinyl boronate ester Ph(H)CdC(Ph)Bcat′. The alkyne
complex 4 was unstable at -30 °C and could not be isolated.
As will be discussed in detail later, the vinyl boronate ester
Ph(H)CdC(Ph)Bcat′ and 1a were formed by the decomposition
of 4 if excess HBcat′ was present. Cp2Ti(PhCCPh) was the
titanium product at these low temperatures if excess dipheny-
lacetylene was present. Compound 4 was also prepared from
the titanocene alkyne complex Cp2Ti(PhCCPh). The reaction
of 1 equiv of HBcat′ with Cp2Ti(PhCCPh) at -78 °C for 10
Z value
goodness of fit
final R indices
Rw ) 0.1631
Table 4. Selected Bond Angles and Distances in 2c
atom
atom
atom
angle (deg)
bond
distance Å)
B1
Ti
H1
B1
Ti
Ti
O1
Cp1
Ti
B1
Ti
H1
H1
P
36(2)
44(2)
110(2)
74.6(1)
127.4(3)
126.1(4)
106.2(4)
136.2 (1)
Ti-B
2.267(6)
1.61(5)
1.35(5)
2.522(2)
2.04(1)
2.05(1)
Ti-H1
B-H1
Ti-P
Ti
P
B1
B1
B1
Ti
O1
O2
O2
Cp2
Cp1-Ti
Cp2-Ti
the smaller size of Ti.41,42,46-52 The Ti-H bond distance of 1.61
Å is typical of bridging titanium hydrides.44 These data again
suggest the presence of a partial bond order among titanium,
hydrogen, and boron in complex 2c. Again the geometry at
boron is remarkable and nearly identical to that in 1g. The sum
of the angles at boron that do not include the hydride is 359.7°.
Thus, the boron again lies in the same plane as the titanium
and two catecholate oxygens, and the geometry about this atom
is far from tetrahedral.
It is informative to make comparisons between the structure
of this compound and that of Cp2Ti(HBcat)2 (1g). The Ti-B
bond length in 2c of 2.267(6) Å was shorter than that in 1g
(2.335(5) Å). The B-H bond length in 2c (1.35(5) Å) was
longer than that in 1g (1.25(3) Å), while the Ti-H bond length
in 2c (1.61(5) Å) was shorter than the distance in 1g (1.74(4)
Å). While bearing in mind the ambiguity in hydride positions,
1
min gave 4 in quantitative yield, as determined by H NMR
spectroscopy.
The bisborane σ-complexes were found to readily exchange
borane ligand. For example, addition of catecholborane to a
solution of 1c in toluene-d8 at -35 °C generated free 4-meth-
ylcatecholborane and precipitated 1g from solution (Scheme 2).
Similarly, reaction between 10 equiv of DBcat-4-Me and 1c at
-10 °C led to the observation of deuterium in the hydride
position, as determined by 2H NMR spectroscopy of the reaction
mixture. The monoborane phosphine complexes behaved analo-
gously. For example, addition of 3-fluorocatecholborane to a
solution of 2b generated a mixture of 2b and 2c while addition
of B-deuteriocatecholborane (DBcat) afforded a mixture of 2b
and 2b-d1 (Scheme 3). The phosphine ligand in 2b was also
readily exchanged. Trimethylphosphine-d9 reacted with 2b to
generate a mixture of 2b and its deuterated analogue 2b-d9.
Spectroscopic Characterization. Compounds 1-4 were
(46) Hartwig, J. F.; He, X. Angew. Chem., Int. Ed. Engl. 1996, 35, 315.
(47) For example, the M-Cl bond length in Cp*2Ti-Cl (2.363 Å) is
shorter than the M-Cl distances in Cp′2NbCl2 and Cp′2TaCl2 of 2.439-
2.483 Å.
(48) Miao, F. M.; Prout, K. Cryst. Struct. Commun. 1982, 11, 269.
(49) Prout, K.; Cameron, T. S.; Forder, R. A.; Critchley, S. R.; Denton,
B.; Rees, G. V. Acta Crystallogr., B 1974, 30, 2290.
(50) McCamley, A.; Miller, T. J. Acta Crystallogr., C 1994, 50, 33.
(51) Labella, L.; Chernega, A.; Green, M. L. H. J. Organomet. Chem.
1995, C18, 485.
1
1
characterized by H, 31P{1H}, 11B, and H{11B} NMR spec-
troscopy at -30 °C. The bisborane complexes 1a-g showed
1
11B NMR signals between 45.7 and 46.3 ppm. The H NMR
spectra of these complexes contained characteristic broad
hydride resonances between -5.37 and -5.85 ppm. The number
and ratio of peaks in the Cp region varied depending on the
number of substituents on the catechol aromatic ring. For
(52) Pattasina, J. W.; Heeres, H. J.; van Boulhuis, F.; Meetsma, A.;
Teuben, J. H. Organometallics 1987, 6, 1004.

Titanocene Borane σ-Complexes
J. Am. Chem. Soc., Vol. 121, No. 21, 1999 5037
Figure 3. Isomers of complex 1a.
Figure 5. Hydride region of the NOESY spectrum of 3.
typically located upfield of those for 3-coordinate boron.53 For
example, the doublet in the 11B NMR spectrum of free HBcat
is shifted from δ 29 to δ 10.5 upon coordination by Et3N to
form the Lewis acid-base adduct HBcat‚Et3N.39,54 Thus, the
upfield shift in the 11B NMR spectrum of 3 may indicate an
1
increase in coordination number. The H NMR spectrum of
complex 3 displayed two hydride signals at δ -4.99 and -6.79
in addition to two Cp singlets at 4.99 and 4.95 ppm. The two
uncoordinated silane hydrogens were observed as a single
resonance at 5.00 ppm at -30 °C, but were resolved into two
resonances at 5.13 and 5.11 ppm at -80 °C. A 1H{11B} NMR
spectrum of 3 showed two sharp hydride signals. A NOESY
spectrum of 3 was obtained with a mixing time of 300 ms at
-80 °C, at which the coordinated silane hydrogens do not
undergo site exchange on the NMR time scale. The NOESY
spectrum (Figure 5) showed NOE interactions with the titanium-
borane hydride and the titanium-bound silane hydride, but not
with the uncoordinated silane hydrides. An NOE interaction was
also observed between the titanium-silane hydride and the
uncoordinated silane hydrogens. A 2H NMR spectrum of Cp2-
Ti(DBcat)(PhSiH3), synthesized using 1g-d2 was obtained at
-30 °C. This spectrum revealed a resonance at 4.9 ppm
corresponding to the chemical shift of the uncoordinated silane
hydrogens, in addition to resonances at -4.98 and -6.78 ppm
for the coordinated hydrides, indicating that exchange of the
borane hydride with the silane hydrides occurs on the laboratory
time scale at -30 °C. Complex 3 showed one sharp band at
2066 cm-1 in the IR spectrum for uncoordinated Si-H, along
with broader bands between 1739 and 1840 cm-1 corresponding
to the hydrides coordinated to titanium.
NMR spectroscopic characterization of alkyne complex 4 was
conducted on a sample generated in situ by the addition of 1
equiv of HBcat′ to a solution of Cp2Ti(PhCCPh) in toluene-d8
at -78 °C. The 11B NMR spectrum of 4 contained a resonance
at 29.8 ppm, again upfield of those for 1a-g and 2a-c. The
1H NMR spectrum of 4 contained a hydride signal at -1.59
ppm. This signal sharpened when decoupled from the boron
nucleus, showing the presence of a B-H interaction. The
deuterated analogue 4-d1 was synthesized using DBcat′. Spec-
troscopic analysis of 4-d1 by 2H NMR revealed a single
resonance at -1.59 ppm.
Figure 4. Isomers of complex 1b.
1
example, the H NMR spectrum of 1g, containing no substit-
uents on the catechol aromatic ring, displayed a single Cp
resonance. However, the spectrum of 1a, which contains a tert-
butyl group at the 4-position in the aromatic ring of the catechol,
revealed three Cp resonances in an approximate ratio of 1:1:2.
These signals corresponded to equal ratios of the Cs (two Cp
resonances) and C2 (one Cp resonance) isomers of 1a (Figure
3). These data indicated that 1a exists as two isomers with
different orientations of the tert-butyl groups of the catechol.
Other analogues of 1a with a single substituent on the catechol
ring (1c-f) displayed a similar set of resonances in a 1:1:2 ratio
in the Cp region, again due to equal amounts of the two isomers
1-Cs and 1-C2. However, compound 1b, which contains two
1
bulky tert-butyl groups on the catechol ring, displayed a H
NMR spectrum containing three peaks in a ratio of 1:1:5,
indicating that the ratio of the Cs isomer to the C2 isomer was
1:2.5. The two hydride resonances in 1a and 1c-f were not
resolved, but the hydride resonances for the two isomers of 1b
were resolved and were located at -5.37 and -6.15 ppm. We
propose that the C2 isomer of 1b is the dominant isomer because
the tert-butyl groups on the catecholate can be staggered (Figure
4). Infrared vibrations of the hydrides of 1a were observed at
1722 and 1603 cm-1, and the other bisborane derivatives showed
bands between these values. Complex 1a-d2 displayed no
vibrations in this region but instead showed bands in the region
between 1250 and 1160 cm-1
.
The NMR spectra of compounds 2a-c were nearly identical,
and tert-butyl-substituted compound 2a is presented as an
informative example. The 11B NMR spectrum contained a single
resonance at 64.0 ppm. The 31P{1H} NMR spectrum consisted
of a sharp singlet at 29.3 ppm. The 1H NMR spectrum displayed
a broad signal at -9.7 ppm in addition to two doublets at 5.03
ppm (J(H,P) ) 2.70 Hz) and 5.01 ppm (J(H,P) ) 2.40 Hz) for
the Cp ligands. The IR spectrum of 2b contained a single broad
band at 1650 cm-1 while the spectrum of 2b-d1 displayed a band
Mechanistic Studies of Ligand Substitution Reactions of
1a. Kinetic studies were conducted to investigate the mechanism
of the reactions described above of compound 1a with trim-
ethylphosphine, carbon monoxide, phenylsilane, and dipheny-
at an appropriate value of 1184 cm-1
.
(53) No¨th, H.; Wrackmeyer, B. Nuclear Magnetic Resonance Spectros-
copy of Boron Compounds, 1st ed.; Springer-Verlag: New York, 1978; p
460.
(54) Goetze, R.; No¨th, H.; Pommerening, H.; Sedlak, D.; Wrackmeyer,
B. Chem. Ber. 1981, 114, 1884.
Silane borane complex 3 displayed a signal at 37.1 ppm in
the 11B NMR spectrum. This signal lies significantly upfield of
those for 1a-g and 2a-c. Signals for 4-coordinate boron are

5038 J. Am. Chem. Soc., Vol. 121, No. 21, 1999
Muhoro et al.
Table 6. Values of kobs for Substitution Reactions by PMe3 for
Various σ-Complexes, Cp2Ti(HBcat′)2
Figure 6. ln[[1a(t)]/[1a(0)]] vs time (s) for the reaction of PMe3
with 1a.
To test potential bonding models and transition state structures
for displacement reactions, we determined the effect of the
borane steric and electronic properties on the rate of substitution
by phosphine. To this end, a series of titanocene borane
σ-complexes were synthesized using different catecholboranes.
The rate constants for the substitution of the boranes by PMe3
were measured for each of 1a-e and are presented in Table 6.
In general, more electron-withdrawing substituents on the
aromatic ring of the catecholborane ligand decreased the rate
of substitution. For example, the electron-poor 4-chloro complex
1d underwent substitution 3 times slower than electron-rich
4-methyl complex 1c.55 The 4-methylthio complex 1e, which
is slightly more electron-rich than complex 1d but less electron-
rich than 1c, underwent substitution slightly faster than did 1d
and about 2 times slower than did 1c. Surprisingly, the rate of
the reaction decreased with increased steric bulk of the borane
ligands. Di-tert-butyl compound 1b underwent ligand substitu-
tion at the slowest rate. Complex 1a with only one tert-butyl
ligand on the catechol ring reacted 3 times faster than did 1b,
while 4-methyl complex 1c reacted 5 times faster than did 1b.
Figure 7. Eyring plot for the reaction of PMe3 with 1a.
Table 5. Values for the Observed Rate Constants at Various
a
Concentrations of HBcat′ and PMe3
[PMe3],
mmol
kobs
[PMe3],
mmol
kobs
[HBcat′] (×104, s-1
)
[HBcat′] (×104, s-1
)
0.0568
0.0568
0.0568
0.0284
0.0568
0.0862
3.57
3.29
3.23
0.0473
0.0785
0.1421
0.0473
0.0473
0.0473
3.25
3.82
3.74
aAll measurements were made at -30 °C with [1] ) 0.009 57 M.
Carbon monoxide displaced both of the borane ligands in
complex 1a to form Cp2Ti(CO)2. The rate of this reaction was
(3.7 ( 0.6) × 10-4 s-1. The reaction rate was measured using
concentrations of added HBcat′ of 0.0474-0.237 M and CO
pressures of 0.5-3.0 atm. The reaction of 1a with CO displayed
a first-order dependence on 1a and a zero-order dependence on
both borane and CO. This absence of reaction order in the
incoming or departing ligand is identical to that observed for
the reaction of 1a with PMe3.
lacetylene. With the exception of the studies on kobs vs T, all
kinetic studies were conducted using 9.6 mM solutions of 1a
in deuterated toluene at -30 °C. The rates of the reactions were
measured by monitoring the disappearance of the cyclopenta-
1
dienyl H NMR signal of 1a.
Trimethylphosphine displaced one borane ligand in titanocene
bisborane 1a to form titanocene phosphine borane complex 2a.
The reaction rate was measured using HBcat′ concentrations
ranging from 0.0284 to 0.0852 M and PMe3 concentrations
ranging from 0.0474 to 0.142 M. The observed rate constants
are shown in Table 5. The rate of this substitution reaction
showed a first-order dependence on the concentration of
bisborane complex as determined by a linear first-order plot
(Figure 6). The rate constants for reactions containing no borane
or 0.00284-0.0852 M added borane and the rate constants for
solutions containing 0.0474-0.142 M added phosphine were
3.5 × 10-4 s-1 with a standard deviation of only 0.2 × 10-4
s-1. The constant values of kobs with varied concentrations of
HBcat′ and PMe3 demonstrated that the reaction was zero order
in both borane and phosphine concentrations. Rate constants
were measured over the temperature range of -15 to -40 °C.
The Eyring plot (Figure 7) provided the activation parameters
∆Hq ) 24.7 ( 2.8 kcal mol-1, ∆Sq ) 29.9 ( 5.4 eu, and ∆Gq
) 17.4 ( 2.8 kcal mol-1 at 233 K.
The rate constant for the reaction of 1a with PhSiH3 to form
silane-borane complex 3 was (4.2 ( 0.1) × 10-4 s-1. The
reaction showed a first-order rate dependence on the concentra-
tion of 1a. Linear first-order plots were observed with no added
HBcat′ and with only 1 equiv of added silane, suggesting that
the reaction was again zero order in both departing and incoming
ligands. The kobs value similar to those of the reaction of 1a
with PMe3 and CO and the absence of an order in HBcat′ and
(55) The electronic properties of the borane can be systematically
perturbed by changing the identity of the substituent on the aromatic ring
of the catecholborane. Hammett substituent constants for the respective
benzoic acids were used to estimate the electronic properties of the boranes,
bearing in mind that the substituents have different σ parameters when
located in the metal and para positions. The electrophilicity of the boranes
should decrease in the trend HBcat-4-Cl > HBcat-4-SMe > HBcat > HBcat-
4-Me. Connors, K. A. Chemical Kinetics: the study of reaction rates in
solution, 1st ed.; VCH: New York, 1990; p 480.

Titanocene Borane σ-Complexes
J. Am. Chem. Soc., Vol. 121, No. 21, 1999 5039
Figure 9. The two dominant orbital interactions in bisborane σ-com-
plexes reveal that the titanocene fragment and the boranes have correct
symmetry for overlap. Importantly, the borane LUMO possesses
significant p-orbital character.
selection of the limiting structure that most closely resembles
each of the observed species will be presented in this section.
The observed 11B NMR shifts of 1a-g located between 45
and 46 ppm are downfield of free catecholborane, indicating
an interaction of the boron with the titanium metal center.41,42,56
The 1H NMR spectra of 1a displayed a broad hydride signal at
-5.7 ppm, which sharpened upon decoupling from boron. The
upfield shift of this signal was indicative of a d2 metal center,
and the sharpened {11B}1H NMR signal suggested a B-H
interaction. The IR bands of 1a-g between 1599 and 1722 cm-1
were reduced from that for catecholborane (ν ) 2660 cm-1),
indicating a reduction in B-H bond order, as are νX-H values
in agostic and σ-complexes. All these data suggest that 1B is
the structural extreme most closely resembling the observed
complex 1a.
Figure 8. Possible structures for complexes 1-4.
PhSiH3 indicate that the reaction proceeded by a similar
mechanism.
The reaction of 1a with diphenylacetylene formed the alkyne
addition product 4. This reaction showed a first-order rate
dependence on the concentration of 1a. The reaction rate was
measured at concentrations of PhCCPh from 0.0474 to 0.237
M and concentrations of HBcat′ from 0.0189 to 0.0474 M. The
observed rate constant for these reactions was (3.8 ( 0.1) ×
-4
10
s-1. The consistency of the kobs values at varied
concentrations of PhCCPh and HBcat′ showed that the reaction
was zero order in HBcat′ and PhCCPh.
Complex 4 was unstable in solution at -30 °C and formed
the alkyne complex Cp2Ti(PhCCPh) in the presence of diphe-
nylacetylene. In this reaction the diphenylacetylene in 4 was
hydroborated and eliminated as the vinyl boronate ester Ph-
In collaborative work, we had previously reported extended
Hu¨ckel calculations on Cp2Ti(HB(OH)2)2, a model for com-
pound 1a. These studies indicated that there is some electron
density between the two boron atoms. However, the B-B
distance of 2.11 Å, which is 0.23-0.25 Å longer than those in
(H)CdC(Ph)Bcat′. The observed rate constant for this hydrobo-
-4
ration reaction was (5.5 ( 0.3) × 10
s-1
.
polynuclear B2H6 complexes,57,58 suggests that interaction
2-
We investigated the steric effects of the borane ligands on
the rates of this hydroboration process. The alkyne-borane
complexes Cp2Ti(PhCCPh)(Bcat-4-Me) and Cp2Ti(PhCCPh)-
(Bcat-3,5-di-t-Bu) were synthesized by reacting Cp2Ti(PhCCPh)
with 4-methylcatecholborane and 3,5-di-tert-butylcatecholbo-
rane, respectively. The observed rate constants for the formation
of vinyl boronate esters from these complexes were (2.83 (
0.03) × 10-4 s-1 for formation of Ph(H)CdC(Ph)Bcat-4-Me
and (16.8 ( 0.2) × 10-4 s-1 for formation of Ph(H)CdC(Ph)-
Bcat-3,5-di-t-Bu. Our results indicate that vinyl boronate ester
formation occurred faster from complexes containing more
sterically demanding boranes.
between the two boron atoms is small.
In addition, these calculations provided a description of the
bonding interactions in the borane complexes (Figure 9). As
one might expect from other σ-complexes, these studies
indicated that the LUMO of the titanocene fragment and the
HOMO of the borane, which contains the antisymmetric
combination of the B-H σ-bonding orbitals, have the proper
symmetry for overlap.39 Further, this interaction involving
donation of electron density from the X-H σ-bond is supple-
mented by back-donation from the metal HOMO into the ligand
LUMO, as it is for most σ-complexes. However, the ligand
LUMO in this case is a boron p-orbital rather than a σ*-orbital
as it is in other dihydrogen, silane, or alkane complexes. This
distinction is important because the boron p-orbital is lower in
energy than a σ*-orbital, creating a closer energy match with
the metal d-orbitals and a stronger σ-complex than many
previous alkane, dihydrogen, or silane complexes. Further, this
back-bonding can be used to rationalize the unusual geometry
at boron we observed by X-ray diffraction. Maximum overlap
Discussion
Spectroscopic and Structural Studies. An understanding
of the structures of the novel coordination complexes 1-4 was
one focus of our study. Spectroscopic studies indicated that each
complex exhibited some degree of partial bonding and that the
overall structural description is most accurately depicted as a
resonance hybrid. NMR properties of the complexes were
employed to determine the limiting structures that most closely
resembled the observed structure. Possible structures of com-
plexes 1-4 are presented in Figure 8, and the rationale for
(56) Hartwig, J. F.; Huber, S. J. Am. Chem. Soc. 1993, 115, 4908.
(57) Kaez, H. D.; Fellmann, W.; Wilkes, G. R.; Dahl, L. F. J. Am. Chem.
Soc. 1965, 87, 2755.
(58) Andersen, E. L.; Felhner, T. P. J. Am. Chem. Soc. 1978, 100, 4606.

5040 J. Am. Chem. Soc., Vol. 121, No. 21, 1999
Muhoro et al.
of the metal HOMO with the borane LUMO will exist when
the titanium, boron, and catecholate oxygens are coplanar.
Scheme 4
The phosphine monoborane complexes 2a-c could be
described by the two structural extremes 2A and 2B in Figure
8 or the isomeric 2c. The ambiguity of the hydride positions in
the X-ray crystallographic study makes spectroscopic studies
more capable than X-ray crystallography of distinguishing 2a
from 2b. The 11B NMR spectra of compounds 2a-c all showed
a signal near 64 ppm located downfield of free catecholborane
and consistent with both structures 2A and 2B. Further, the sharp
signals in the 31P NMR spectra of 2a-c and the doublet
cyclopentadienyl resonance for 2a that collapsed to a singlet
upon decoupling from phosphorus showed that the phosphine
was bound to the metal center and not the quadrupolar boron
nucleus. The 1H NMR spectra of 2a-c also contained a broad
resonance between -9.4 and -9.8 ppm, which sharpened upon
decoupling from boron. The downfield shift contrasts the sharp
upfield signal that would be expected for a Ti(IV) hydrido boryl
complex because related complexes show sharp hydride sig-
nals.41,42 These data reveal the presence of a B-H interaction
in compounds 2a-c and indicate that the metal center is best
described as a Ti(II) d2 center and not a Ti(IV) d0 center. The
broad band at 1650 cm-1 in the IR spectrum of 2b indicated a
weakened B-H bond, but the small B-Ti-H angle and the
short B-H distance support a partial B-H bonding interaction.
Thus, limiting structure 2B most closely resembles complexes
2a-c.
activation barriers (5-10 kcal mol-1).59 Thus, the deuterium
label may scramble from the borane to the silane through a
bidentate-monodentate equilibrium (Scheme 4) in which rota-
tion of the monodentate borohydride would cause the label to
be exchanged into the silane. Once incorporated into the silane
hydride position, the deuterium label may exchange with the
uncoordinated silane hydrogen positions by a turnstile rotation
similar to that in agostic complexes.60
The structure of alkyne adduct 4 was deduced by NMR
spectroscopy. Four possible structures for 4 are presented in
Figure 8. The 11B NMR spectrum of 4 consisted of a resonance
1
at 29.4 ppm, while the H NMR spectrum contained a broad
hydride peak at -1.59 ppm. This hydride signal sharpened when
decoupled from boron, indicating a bonding interaction between
boron and hydrogen. On the basis of these data, 4A and 4B,
which would show hydride resonances in the vinyl region of
η2-coordinated vinyl boronate esters61 and in the region of Ti-
(IV) hydrides,62,63 respectively, can be eliminated as possible
structures of 4. The 11B NMR resonance of 29.4 ppm, which is
nearly equivalent to that of catecholborane, and the broad
hydride signal at δ -1.59, which is slightly downfield of the
resonances of other borane complexes (1-3) and upfield of the
resonances of d0 metallocene borohydrides,64 suggest that the
structure of 4 results from nearly equivalent contributions from
a σ-bonded borane compound (4C) and an η2-vinyl hydride
complex (4D).61 This conclusion is consistent with the reactivity
of 4. Rapid exchange of the borane ligand with free borane
would be expected from a complex with structure 4C, while
spontaneous elimination of H(Ph)CdC(Ph)(Bcat′) would be
expected from a complex such as 4D. Both processes were
observed.
Mechanistic Studies. Kinetic studies were conducted to
investigate the mechanism of ligand substitution by PMe3. The
mechanisms of these reactions were investigated for two
principal reasons. First, the mechanism of the substitution
reactions would probe whether the B-H bond of the borane
was acting as a simple two-electron, dative ligand. If 1a-e
contained an anionic H2B2cat′2 ligand rather than two neutral
borane ligands, simple borane displacement would be unlikely
to occur. Second, if the mechanism of substitution were
dissociative, then we could obtain experimental values for the
enthalpy of activation. In general, the enthalpies of activation
Complex 3 is the product from the substitution of one borane
ligand in 1a by PhSiH3. A structure for 3 in which the hydrides
occupy the endo positions and the boryl and silyl groups occupy
the exo positions is unreasonable because this structure would
maximize steric interaction between the ligand aryl groups and
the Cp ligands. Three possible structures for compound 3, in
which unfavorable interaction is minimized, are presented in
Figure 8, and spectroscopic methods were again used to
distinguish between them. The 11B NMR spectrum of 3
1
contained a signal at 37 ppm, while the H NMR spectrum
revealed two broad signals at -4.98 and -6.78 ppm. Structure
3A, in which boron is 4-coordinate and anionic, can be ruled
out as the dominant resonance contributor by the 11B NMR
chemical shift that is downfield of free HBcat′. Two-dimensional
1H NMR spectroscopy was used to differentiate between
structures 3B and 3C. NOE interactions were observed between
the bridging borane hydride (HA) and the bridging silane hydride
(HB) and between the silane uncoordinated hydrogens (HC, HD)
and HB. No interactions were observed between HA and either
HC or HD. These data indicate that 3 exists as the regioisomer
3C and that potential residual turnstile rotation of the silane
does complicate the interpretation of the NOE experiment. Two
sharp infrared bands (ν ) 2051 and 2095 cm-1), which were
attributed to the vibrations of the uncoordinated silane hydro-
gens, HC/HD, along with broad bands between 1754 and 1863
cm-1 corresponding to the silane and borane hydrides also
supported structure 3C. The 1754 and 1863 cm-1 values are
close to those of other titanium-silane complexes that possess
weakened X-H bonds.44 Although structure 3C appears to be
the dominant resonance form of compound 3, the 11B NMR
chemical shift that is upfield of that for compounds 1a-g and
2a-c implies that there is a significant contribution from 3A
to the overall structure.
(59) Marks, T. J.; Kolb, J. R. Chem. ReV. 1977, 77, 263.
(60) Brookhart, M.; Green, M. L. H. J. Organomet. Chem. 1983, 250,
395.
(61) Binger, P.; Sandmeyer, F.; Kru¨ger, C.; Kuhnigk, J.; Goddard, R.;
Erker, G. Angew. Chem., Int. Ed. Engl. 1994, 33, 197.
(62) Bercaw, J. E.; Brintzinger, H. H. J. Am. Chem. Soc. 1972, 94, 1219.
(63) Bennet, J. L.; Wolczanski, P. T. J. Am. Chem. Soc. 1994, 116, 2179.
(64) Davies, N.; James, B. D.; Wallbridge, M. G. H. J. Chem. Soc. A
1969, 2601.
The site exchange of the hydrogen atoms in 3 was revealed
by 2H NMR spectroscopy and corroborates the substantial
borohydride character in 3. Metallocene borohydride complexes
are known to undergo hydrogen exchange processes with low

Titanocene Borane σ-Complexes
J. Am. Chem. Soc., Vol. 121, No. 21, 1999 5041
Scheme 5
Our kinetic studies indicated that the ligand substitution
reaction has a first-order dependence on the concentration of
1a and no dependence on the concentration of either borane or
PMe3. These data are inconsistent with the conventional ring
slip pathway.67 Instead, these data support the mechanism
involving initial dissociation of the catecholborane ligand,
followed by rapid association of PMe3 to form 2a. Thus, k2 .
k-1 in eq 5, and HBcat′ dissociation is irreversible and rate
determining in these reactions.
The reaction rates of dissociative ligand substitutions with
initial irreversible ligand dissociation are independent of the
nature of the incoming ligand. Our results are consistent with
this requirement. Complex 1a reacted with trimethylphosphine,
carbon monoxide, phenylsilane, and diphenylacetylene with
equal rate constants. These data confirm our conclusions based
on the dependence of rate on the concentrations of HBcat′ and
PMe3.
in dissociative ligand substitution reactions are similar to the
bond dissociation energies of the ligand-metal bond. From the
values of ∆Hq, we could provide an upper limit on the strength
of the Ti-(HBcat) σ-bond. Two possible pathways for ligand
substitution in 1a are presented in Scheme 5. The associative
pathway depicts a mechanism involving an η5-η3 ring slip of
the Cp ligands followed by addition of a ligand. The second
pathway involves dissociation of a ligand to form a coordina-
tively unsaturated intermediate, which is rapidly trapped by the
incoming ligand.
Associative ligand substitution reactions of coordinatively
saturated metallocenes commonly proceed via ring slip mech-
anisms.65,66 Such mechanisms occur less frequently in first-row
transition metal complexes where the small sizes of the metal
centers disfavor associative processes, and weak metal-ligand
bonds favor dissociative processes. An initial ring slip followed
by coordination of PMe3 would result in the formation of an
η3-cyclopentadienyl complex. This complex could then undergo
borane dissociation and resume the η5-cyclopentadienyl coor-
dination mode to generate 2a. The rate of the reaction by this
mechanism may be expressed as in eqs 3 and 4. The rate of a
reaction by this pathway would be first order in 1a and PMe3
and zero order in borane.
Finally, a ∆Sq value that is positive is commonly observed
for dissociative substitution because of the lengthening of the
M-L bond during formation of the transition state, and
accompanying increase in the degrees of freedom.68 The entropy
of activation for the reaction of 1a with PMe3 was 30 ( 5 eu,
consistent with a dissociative mechanism. The ∆Hq value of
25 ( 3 kcal mol-1 provided an upper limit for the Ti-(B-H)
bond dissociation energy which was intermediate in strength
between those of known silane (∼30 kcal mol-1) and dihydrogen
(∼13 kcal mol-1) σ-complexes of other metal systems.2,3 The
similar energy of borane and silane σ-complexes is consistent
with our observation of an equilibrium between phosphine
borane complex 2a and Cp2Ti(H2SiPh2)(PMe3).
Kinetic studies were conducted on a series of titanocene
borane σ-complexes to investigate the effect of the steric and
electronic properties of the borane ligand on the substitution
reaction. This study was performed by measuring the rate
constants of substitution for different σ-complexes synthesized
from boranes with different substituents on the catecholate ring
(Table 6). Because the mechanism was dissociative, a trend in
the strength of the Ti-(HBcat) interaction as a function of
electronic factors could be established from rate constants. The
observed trend is more likely to result from changes in the
strength of the Ti-(HBcat) interaction than perturbations of the
free H-B bond strength since BDEs of boranes change only
slightly with variation of substituents.69 It should be noted that
the metal-borane BDEs may be different from enthalpies of
activation because of the significant reorganization detected by
the anomalous steric effects. This phenomenon, however, does
not preclude comparisons of activation parameters as long as
the steric factors are relatively constant. Thus, we conducted
our studies on electronic effects using substituents of similar
size.
Dissociative mechanisms are common for ligand substitution
in coordinatively saturated first-row transition metal complexes.
In such a mechanism, initial dissociation of one borane ligand
from 1a would occur to generate a monoborane intermediate,
which would be rapidly trapped by PMe3 to form 2a (Scheme
5). The rate expression for this mechanism is shown in eq 5. If
the rate of phosphine coordination is much faster than the rate
of borane recoordination, then the overall reaction rate would
have a concentration dependence on only 1a. However, if the
rate of phosphine coordination is much slower than the rate of
borane recoordination, then the rate of the overall reaction would
have a first-order dependence on the concentrations of 1a and
PMe3 and would have an inverse first-order dependence on the
concentration of HBcat′.
Our results indicated that complexes with electron-withdraw-
ing substituents on the aromatic ring of the catecholborane
underwent substitution reactions more slowly than did those with
more electron-donating substituents. This observation suggests
that σ-borane complexes are stabilized by increasing the degree
of electron back-donation from titanium. This result is consistent
with the importance of back-bonding between the titanium
HOMO and the borane LUMO used above to rationalize the
(67) A less conventional ring slip mechanism involving an intramolecular
and irreversible rate-determining transformation from an η5 to an η3
cyclopentadienyl complex is less consistent with the large positive entropy
of activation and steric effects described below.
(65) Palmer, G. T.; Basolo, F.; Kool, L. B.; Rausch, M. D. J. Am. Chem.
Soc. 1986, 108, 4417.
(66) For an early example with monocyclopentadienyl compounds see:
Schuster-Woldan, H. G.; Basolo, F. J. Am. Chem. Soc. 1966, 88, 1657.
(68) Atwood, J. D. Inorganic and Organometallic Reaction Mechanisms;
Cole Publishing Co.: Monterey, CA, 1985; p 17.
(69) Rablen, P. R.; Hartwig, J. F.; Nolan, S. P. J. Am. Chem. Soc. 1994,
116, 4121.

5042 J. Am. Chem. Soc., Vol. 121, No. 21, 1999
Muhoro et al.
Scheme 6
Figure 10. Proposed mechanism for structural rearrangement of
catecholborane ligands prior to dissociation.
stability and geometry of the borane σ-complexes. A more
electrophilic borane would create a greater back-bonding
interaction and a stronger overall metal-borane bonding
interaction. In addition, the greater stability we observed of the
phosphine-borane complex 2a, relative to that of Cp2Ti(H2-
SiPh2)(PMe3), is likely to result from the same back-bonding
interaction. The metal will interact more strongly with the borane
LUMO.
The steric effects on the substitution reactions were unusual.
Classic dissociative ligand substitution reactions typically
proceed faster with larger departing ligands. Contrary to this
usual trend, we found that increasing steric bulk on the catechol
resulted in a decrease in the rate of substitution by PMe3. These
observations are more common for an associative mechanism.
However, the reproducibility of numerous kinetic experiments,
the independence of the reaction rate on the nature of the
incoming ligand, and the values of the activation parameters
clearly show that the mechanism for the borane substitution
reactions is dissociative.
borane ligand in the alkyne-borane complex 4 presumably has
a significantly weakened B-H bond. This property may activate
the borane ligand in 4 toward addition across an uncoordinated
alkyne. In this case, the ligated diphenylacetylene would remain
on the metal to form the metal-alkyne product. This mechanism
would provide a first-order rate dependence on the concentra-
tions of both diphenylacetylene and 4 and a zero-order
dependence on the concentrations of borane.
The unusual steric effects can be explained by the structural
changes in Figure 10. The free catecholborane product is planar.
Thus, the departing borane will be flatter in the transition state
than it is in the ground state. The flattening of the catecholborane
requires that its aromatic ring swing toward that of the
nondissociating borane. The initial titanium product is [Cp2Ti-
(HBcat-4-R)] which could be formulated as either a Ti(IV)
hydrido boryl or Ti(II) monoborane σ-complex. The geometry
of this intermediate is not simple to predict. To form a boryl
hydride complex, the boryl or hydride group would presumably
relocate to the side of the titanocene wedge formerly occupied
by the dissociating ligand. However, in a monoborane complex
intermediate, the location of the borane ligand may be relatively
unchanged because Cp2ML complexes that are low spin d2 are
predicted to be nonsymmetric.43 Thus, it is not clear whether
rearrangements of the remaining borane would exacerbate the
steric effects of the flattening of the departing borane. However,
the transition state is likely to have greater steric interactions
in either case than would the ground state because of the
geometrical changes of the departing borane. Complexes with
more structurally hindered boranes would undergo dissociation
more slowly.
Diphenylacetylene reacted with 1a to form the alkyne
complex 4. The rate of this substitution reaction was the same
as those for borane substitution by other incoming ligands.
However, the product 4 was unstable in solution at -30 °C.
Compound 4 reacted further with diphenylacetylene to generate
Cp2Ti(PhCCPh) and vinyl boronate ester. This transformation
of 4 to Cp2Ti(PhCCPh) and vinyl boronate ester is related to
the mechanism of catalytic hydroboration of alkynes by Ti(II)
complexes.32 To study the reaction of 4 with diphenylacetylene,
the alkyne complex 4 was independently prepared by the
reaction of Cp2Ti(PhCCPh) with catecholborane.
In an alternative mechanism, the ligated alkyne in 4 could
undergo electrophilic attack by external HBcat′, particularly in
reactions containing excess borane. This reaction would release
the vinyl boronate ester product and generate the Ti(IV) hydrido
boryl or Ti(II) monoborane complex. The fate of this intermedi-
ate would depend on the reagents in solution. In the presence
of borane, 1a would be generated; in the presence of alkyne, 4
would be generated. In either case, this pathway would show a
first-order rate dependence on the concentration of borane and
4, and a zero-order dependence on alkyne.
Finally, it might be expected from the proposed structure of
4 that reductive elimination of vinyl boronate ester would be
facile. A simple reductive elimination involving the hydride and
the Ti-C bond would form the vinyl boronate ester and release
the Cp2Ti fragment, which would rapidly react with alkyne to
form Cp2Ti(PhCCPh). This mechanism would provide a zero-
order dependence of the reaction rate on both borane and alkyne
concentration and a simple first-order dependence on the
concentration of 4.
Our kinetic studies indicated that the formation of vinyl
boronate ester from 4 proceeded with a first-order dependence
on the concentration of 4 and a zero-order dependence on the
concentrations of both borane and alkyne. These results
eliminated initial external attack by either borane or alkyne onto
4 as possible mechanisms for the reaction. Therefore, the most
plausible mechanism for the hydroboration of alkyne in 4 is,
indeed, the intramolecular reductive elimination.
It should be noted, however, that a zero-order dependence
on HBcat′ and alkyne could also result from a combination of
borane dissociation followed by external attack by HBcat′ on
the unsaturated intermediate (Scheme 7). Borane dissociation
from 4 is reasonable because 4′ is a resonance contributor to
the overall structure of the alkyne borane complex. However,
dissociation of borane and reaction of free borane with Cp2Ti-
(PhCCPh) by addition of the B-H bond across the Ti-C bond
Three potential mechanisms for the hydroboration reaction
are shown in Scheme 6. One possible mechanism may involve
external attack by diphenylacetylene on coordinated HBcat′. The

Titanocene Borane σ-Complexes
J. Am. Chem. Soc., Vol. 121, No. 21, 1999 5043
Figure 12. Stabilizing interactions in monoborane complexes.
for formation of new monoborane complexes Cp2Ti(HBcat′)-
(L) by coordination of L.
Although it is tempting to compare the stability of the
different complexes Cp2Ti(HBcat′)(L) as a function of the nature
of L, the stabilities of complexes 1, 3, and 4 are likely to be
influenced by interactions that provide some degree of intramo-
lecular interaction with boron as shown in Figure 12. A small
boron-boron interaction in 1, a substantial hydrogen-boron
interaction in 3, and a strong B-C interaction in 4 would
complement the σ-coordination of the B-H bond to stabilize
the compounds. Thus, a simple correlation can be made between
the electronic nature of the ligand and the stability of the Cp2-
Ti(HBcat′)(L) complex only in the cases where L is CO or
PMe3. In these cases, the stability of the monoborane σ-complex
Cp2Ti(HBcat′)(L) appears to depend on the π-acidity and
σ-basicity of the ligand L. The monoborane complex of the
strong σ-donor Cp2Ti(HBcat)(PMe3) (2b) is generated by a
conproportionation reaction between Cp2Ti(PMe3)2 and Cp2Ti-
(HBcat)2 and is stable in solution at 0 °C. The bisborane
complex is less stable and decomposes more rapidly at 0 °C in
the absence of added borane. Finally complex Cp2Ti(HBcat′)-
(CO) is merely a reactive intermediate in the CO substitution
reactions. Attempts to prepare this compound by conpropor-
tionation between Cp2Ti(CO)2 and Cp2Ti(HBcat′)2 gave no
reaction. Further, only Cp2Ti(CO)2 and Cp2Ti(HBcat′)2 were
observed throughout the reaction of CO with Cp2Ti(HBcat′)2
and when less than 2 equiv of CO was added to Cp2Ti(HBcat′)2.
From the observed trend, it appears that π-acidic ligands such
as carbon monoxide destabilize the σ-complex. Such ligands
would compete with the borane for electron density on the metal
that provides back-donation into boron and stabilizes the
complex. Thus, the strong σ-donating and weak π-accepting
ligand PMe3 provides a more stable complex. This trend is
supported by the electronic effects of substitution on the borane
ligand. The rates for borane dissociation decreased with
increased borane electrophilicity. Because ∆H and ∆Hq are
closely related in dissociative substitution, the slower reaction
of the complexes of electrophilic boranes is most likely due to
their increased thermodynamic stability.
Figure 11. Qualitative free energy diagram for borane exchanges in
4.
Scheme 7
would generate the metallacycle 4′′, which is identical to the
starting complex 4′. Clearly such a mechanism involving
sequential dissociation and reassociation of borane ligand would
be nonproductive. However, dissociation of catecholborane
followed by its addition across the C-C bond of the coordinated
alkyne would be consistent with the kinetic data. This mecha-
nism cannot be ruled out, but would be a highly unusual type
of X-H addition process.
In an attempt to confirm the intramolecular reductive elimina-
tion mechanism, a crossover experiment was performed in which
4-methylcatecholborane was added to a solution of 4, and the
identity of the vinyl boronate ester product was determined.
Instead of providing information on the mechanism of the vinyl
boronate ester as designed, these experiments revealed a rapid
borane dissociation from 4. Both 4-methyl- and 4-tert-butyl-
substituted vinyl boronate ester products and both 4-methyl-
and 4-tert-butyl-substituted borane alkyne complexes were
observed in solution. Further, borane exchange occurred faster
than elimination of vinyl boronate ester product. A qualitative
free energy diagram for the borane exchanges and intermolecular
reductive elimination for the two alkyne complexes is presented
in Figure 11. The two alkyne complexes coexist in equilibrium
with a barrier to interconversion that is smaller than that for
reductive elimination of the vinyl boronate ester products. As
noted in the discussion of the spectroscopic features of 4, this
chemistry supports a structure of 4 that is a combination of a
vinyl hydride and an alkyne borane complex.
Experimental Section
General Considerations. Unless otherwise noted, all manipulations
were conducted using standard Schlenk techniques or in an inert
1
atmosphere glovebox. H NMR spectra were obtained on a GE QE
300 MHz, GE Ω 300 MHz, or Bruker AM-500 MHz Fourier transform
2
spectrometer. 11B, 31P, and H NMR spectra were obtained on the Ω
300 MHz spectrometer operating at 96.38, 121.65, and 46.13 MHz,
respectively. 1H NMR spectra were recorded relative to residual
protiated solvent. 11B NMR and 31P NMR spectra were recorded in
units of parts per million relative to BH3‚Et2O and 85% H3PO4,
respectively, as external standards.
Conclusion
Unless specified otherwise, all reagents were purchased from
commercial suppliers and used without further purification. Cp2Ti-
(PMe3)2,70 Cp2TiMe2,71 Cp2Ti(PhCCPh),72 Cp2Ti(H2SiPh2)(PMe3),44
Our studies on titanocene bisborane complexes demonstrate
that catecholborane can behave as a simple two-electron-
donating ligand. The stability of these complexes is attributed
in part to the favorable match of the titanocene orbitals with
the π-accepting orbitals of the borane ligand. The catecholborane
ligand dissociates from the metal, and this dissociation allows
(70) Kool, L. B.; Rausch, M. D.; Alt, H. G.; Herberhold, M.; Thewalt,
U.; Wolf, B. Angew. Chem., Int. Ed. Engl. 1985, 24, 394.
(71) Erskine, G. H.; Wilson, D. A.; McCowan, J. D. J. Organomet. Chem.
1976, 114, 119.

5044 J. Am. Chem. Soc., Vol. 121, No. 21, 1999
Muhoro et al.
4-chlorocatechol,73 4-methylthiocatechol,74 and B-deuterio-4-methyl-
catecholborane75 were prepared using literature procedures. All catechols
were dried using a Dean-Stark apparatus prior to reaction. Protiated
solvents were refluxed and distilled from purple solutions containing
sodium benzophenone. Deuterated solvents were dried similarly, but
were collected by vacuum transfer. Reaction yields that were obtained
by 1H NMR spectroscopy were determined using ferrocene as the
internal standard.
Preparation of Catecholboranes. All the catecholboranes were
prepared using a modification of a literature preparation for catecholbo-
rane (1,3,2-benzodioxaborole),76 substituting BH3‚SMe2 for BH3‚THF.
To a stirred ether solution of the catechol was added BH3‚SMe2 via
syringe. Effervescence was observed, and the resulting clear tan-colored
solution was stirred at room temperature for 4 h. Solvent was removed
in vacuo, and the product was vacuum distilled at 0.5 Torr and 50-60
°C to give the catecholborane as a clear colorless oil.
4-tert-Butylcatecholborane. Product obtained in 62% yield (9.9 g).
1H NMR (C6D6): δ 7.17 (d, J ) 1.5 Hz, 1H), 6.94 (d, J ) 8.4 Hz,
1H), 6.87 (dd, J ) 8.4 Hz, 1.5 Hz, 1H), 4.40 (br q, JH-B ) 186 Hz,
1H), 1.15 (s, 9H). 11B NMR (C6D6): δ 29.4 (d, JH-B) 186 Hz). IR
(Nujol, cm-1): 2659 (s, νH-B), 1486 (s), 1431 (s), 1281 (s), 1250 (s),
1222 (s), 1130 (s), 866 (s), 812 (s).
4-Methylcatecholborane. Product obtained in 56% yield (3.0 g).
1H NMR (C6D6): δ 6.86 (d, J ) 8.2 Hz, 1H), 6.77 (s, 1H), 6.53 (d, J
) 8.2 Hz, 1H), 4.0 (br q, JH-B ) 207 Hz, 1H), 2.02 (s, 3H). 11B NMR
(C6D6): δ 29.5 (d, JH-B) 207 Hz). IR (Nujol, cm-1): 2656 (s, νH-B),
1311 (s), 1281 (s), 1247 (s), 1168 (s), 1134 (s), 859 (s), 802 (s).
1.5 Hz, 2H), 5.42 (s, 10H), 1.66 (s, 18H), 1.12 (s, 18H), -6.15 (br s,
2H). (Cs isomer) δ 6.92 (s, 2H), 6.73 (s, 2H), 5.46 (s, 5H), 5.42 (s,
5H), 1.58 (s, 18H), 1.21 (s, 18H), -5.37 (br s, 2H). 11B NMR (C7D8,
-10 °C): δ 46.3. IR (Nujol, cm-1): 1665 (br, m), 1621 (m), 1592 (s),
1512 (s), 1408 (s), 1287 (s), 1072 (s), 828 (s), 758 (s).
Cp2Ti(HBcat-4-Me)2 (1c). Yield: 83% (833 mg). 1H NMR (C7D8,
-10 °C, two isomers): δ 6.85 (dd, J ) 8.0, 4.1 Hz, 4H), 6.76 (d, J )
4.1 Hz, 4H), 6.35 (d, J ) 8.0, 4H), 5.42 (s, 5H), 5.41 (s, 10H), 5.40 (s,
5H), 1.97 (s, 6H), 1.96 (s, 6H), -5.45 (br s, 4H). 11B NMR (C7D8,
-10 °C): δ 45.8. IR (Nujol, cm-1): 1650 (br, m), 1611 (m), 1481 (s),
1431 (s), 1234 (s), 1101 (s), 1099 (s), 1000 (s), 828 (s), 750 (s).
Cp2Ti(HBcat-4-Cl)2 (1d). Yield: 67% (49.0 mg). 1H NMR (C7D8,
-10 °C, two isomers): δ 6.95 (d, J ) 1.6 Hz, 2H), 6.94 (d, J ) 1.6
Hz, 2H), 6.62 (d, J ) 5.4 Hz, 2H), 6.59 (d, J ) 5.412, 2H), 6.50 (dd,
J ) 5.4 Hz, 1.6 Hz, 2H), 6.48 (dd, J ) 4.7 Hz, 2.3 Hz, 2H), 5.30 (s,
5H), 5.30 (s, 10H), 5.29 (s, 5H), -5.80 (br s, 4H). 11B NMR (C7D8,
-10 °C): δ 45.8. IR (Nujol, cm-1): 1695 (br, m), 1599 (m), 1477
(br), 1232 (s), 1210 (s), 1113 (s), 825 (s).
Cp2Ti(HBcat-4-SMe)2 (1e). Yield: 29% (90 mg). 1H NMR (C7D8,
-10 °C, two isomers): δ 6.78 (d, J ) 3.6 Hz, 2 H), 6.75 (d, J ) 3.6
Hz, 2H), 6.57 (d, J ) 7.7 Hz, 2H), 6.54 (d, J ) 7.7 Hz, 2H), 5.40 (s,
5H), 5.39 (s, 10H), 5.38 (s, 5H), -5.83 (br s, 4H), the most downfield
aromatic resonance was obscured by the residual protiated toluene
solvent. 11B NMR (C7D8, -10 °C): δ 45.7. IR (Nujol, cm-1): 1651
(br, m), 1599 (m), 1231 (s), 1078 (s), 880 (s).
Cp2Ti(HBcat-3-F)2 (1f). Yield: 76.5% (168 mg). 1H NMR (C7D8,
-10 °C, two isomers): δ 6.62 (m, 4H), 6.49 (m, 4H), 6.32 (m, 4H),
5.33 (s, 5H), 5.33 (s, 10H), 5.313 (s, 5H), -5.8 (br s, 4H). 11B NMR
(C7D8, -10 °C): δ 45.7. IR (Nujol, cm-1): 1672 (br, m), 1626 (m),
1492 (s), 1435 (s), 1270 (s), 1232 (s), 1048 (m), 1002 (s), 849 (s), 769
(s), 723 (s).
Cp2Ti(HBcat)2 (1g). Yield: 94% (1.77 g). Due to the insolubility
of 1g in aromatic solvents, NMR data were obtained by reacting a
solution of Cp2TiMe2 (2.0 mg, 0.0097 mmol) in toluene-d8 with
catecholborane (3.8 mg, 0.031 mmol) in an NMR tube at -10 °C and
obtaining NMR data prior to the precipitation of 1g from solution. 1H
NMR (C7D8, -10 °C): δ 6.95 (m, 4H), 6.57 (m, 4H), 5.40 (s, 10H),
-5.85 (br s, 2H). 11B NMR (C7D8, -10 °C): δ 45.0. IR (Nujol, cm-1):
1683 (br, m), 1612 (m), 1485 (s), 1459 (s), 1249 (s), 1217 (s), 1130
(s), 1027 (s), 823 (s), 803 (s).
3,5-Di-tert-butylcatecholborane. Product obtained in 67% yield (3.2
1
g). H NMR (C6D6): δ 7.19 (s, 1H), 7.15 (s, 1H), 4.3 (br q, J
)
H-B
197 Hz, 1H), 1.461 (s, 9H), 1.223 (s, 9H). 11B NMR (C6D6): δ 29.8
(d, JH-B ) 197 Hz). IR (Nujol, cm-1): 2650 (s, νH-B), 1505 (s), 1402
(s), 1327 (s), 1232 (s), 1138 (s), 979 (s), 865 (s), 824 (s).
4-Chlorocatecholborane. Product obtained in 12% yield (129 mg).
1H NMR (C6D6): δ 6.92 (d, J ) 2.2 Hz, 1H), 6.71 (dd, J ) 8.6, 2.2
Hz, 1H), 6.57 (d, J ) 8.6, 197 Hz, 1H), 4.3 (br q, J
) 197 Hz,
H-B
1H). 11B NMR (C6D6): δ 30.2 (d, JH-B ) 197 Hz). IR (Nujol, cm -1):
2661 (s, νH-B), 1276 (s), 1236 (s), 1157 (s), 1129 (s), 882 (s), 807 (s).
4-Methylthiocatecholborane. Product obtained in 44% yield (382
1
mg). H NMR (C6D6): δ 7.03 (s, 1H), 6.79 (s, 1H), 6.79 (s, 1H), 4.5
(br q, JH-B) 196 Hz, 1H), 1.916 (s, 3H). 11B NMR (C6D6): δ 29.5 (d,
J H-B ) 196 Hz). IR (Nujol, cm-1): 2660 (s, νH-B), 1599 (s), 1422 (s),
1292 (s), 1272 (s), 1146 (s).
Cp2Ti(DBcat)2 (1g-d2). Yield: 58% (333 mg). Due to the insolubility
of 1g-d2 in aromatic solvents, NMR data were obtained by reacting a
solution of Cp2TiMe2 (2.0 mg, 0.0097 mmol) in toluene with B-
deuteriocatecholborane (3.7 mg, 0.031 mmol) in an NMR tube at -10
°C and obtaining NMR data prior to the precipitation of 1g-d2 from
3-Fluorocatecholborane. Product obtained in 46.4% yield (2.54 g).
1H NMR (C6D6): δ 6.61 (d, J ) 7.2 Hz, 1H), 6.46-6.54 (m, 2H),
4.40 (br q, JH-B ) 205 Hz, 1H). 11B NMR (C6D6): δ 29.2 (d, J H-B
)
2
solution. H NMR (C7H8, -10 °C): δ -5.9. 11B NMR (C7H8, -10
205 Hz). IR (Nujol, cm-1): 2664 (m, νH-B), 1634 (s), (1506 (s), 1338
(s), 1173 (s), 1023 (s), 873 (s), 766 (s).
°C): δ 45.0. IR (Nujol, cm-1): 1238 (s), 1193 (s), 1105 (m), 1032
(m), 985 (s), 865 (m), 826 (s), 749 (s).
Preparation of Titanocene Bisborane Complexes. All the ti-
tanocene bisborane complexes were prepared by adding the catecholbo-
rane to solutions of Cp2TiMe2 in 20 mL of pentane that were cooled in
a -30 °C freezer. The resulting mixture was quickly placed back into
the freezer (-30 °C) and left overnight. Pale yellow solids precipitated
and were collected by vacuum filtration. The products were washed
with cold pentane and then dried under reduced pressure.
Generation of Cp2Ti(HBcat-4-t-Bu)(PMe3) (2a) in Situ. Into an
NMR tube were weighed 1a (10.0 mg, 0.0189 mmol) and ferrocene
(approximately 3 mg). The tube was sealed with a septum and immersed
in a dry ice-acetone bath. To the tube was added 0.6 mL of toluene-
d8, and a 1H NMR spectrum was obtained at 0 °C. The tube was
immersed again in the dry ice-acetone bath, and PMe3 (1.96 µL, 0.0189
mmol) was added by syringe. The tube was shaken and allowed to
remain at 0 °C for 3 h. A 1H NMR spectrum at 0 °C indicated that 2a
1
Cp2Ti(HBcat-4-t-Bu)2 (1a). Yield: 63% (1.6 g). H NMR (C7D8,
-10 °C, two isomers): δ 7.05 (d, J ) 2 Hz, 2H), 7.03 (d, 2 Hz, 2H),
6.87 (d, J ) 7.8 Hz, 2H), 6.85 (d, J ) 7.2 Hz, 2H), 6.58 (dd, J ) 7
Hz, 2 Hz, 2H), 6.57 (dd, J ) 8 Hz, 2 Hz, 2H) 5.48 (s, 5H), 5.46 (s,
10H), 5.44 (s, 5H), 1.10 (s, 18H), -5.67 (br s, 4H). 11B NMR: (C7D8,
-10 °C): δ 46.0. IR (Nujol, cm-1): 1722 (br, m), 1603 (s), 1493 (s),
1423 (s), 1255 (s), 861 (s), 828 (s).
1
was formed in 84% yield. H NMR (C7D8, 0 °C): δ 7.18 (d, J ) 1.7
Hz, 1H), 6.86 (d, J ) 8.5 Hz, 1H), 6.79 (dd, J ) 8.5, 1.7 Hz, 1 H),
5.03 (d, JH-P ) 2.7 Hz, 5H), 5.01 (d, JH-P ) 2.7 Hz, 5H), 1.26 (s,
9H), 0.75 (d, J ) 6.0 Hz, 9H), -9.7 (br s, 1H). 11B NMR (C7D8, 0
°C): δ 64.0. 31P{1H} NMR: δ 29.3.
Preparation of Cp2Ti(HBcat)(PMe3) (2b). Into a side-arm flask
were weighed Cp2Ti(PMe3)2 (395 mg, 1.20 mmol) and 1g (500 mg,
0.947 mmol). The flask was cooled to -30 °C and then charged with
3 mL of cold (-30 °C) toluene. This solution was stirred for 90 min
at 0 °C. The resulting dark maroon solution was then kept at -30 °C
for 2 d, after which time a dark pink solid precipitated from solution.
The dark maroon supernatant was removed, and the solid was dried in
vacuo to give 413 mg (46.4%) of Cp2Ti(HBcat)(PMe3). 1H NMR (C7D8,
0 °C): δ 7.09 (m, 2H), 6.79 (m, 2H), 5.02 (d, JH-P ) 3.0 Hz, 10H),
0.75 (d, JH-P ) 6.3 Hz, 9H), -9.8 (br s, 1H). 11B NMR (C7D8, 0 °C):
1
Cp2Ti(HBcat-3,5-di-t-Bu)2 (1b). Yield: 74% (394 mg). H NMR
(C7D8, -10 °C): (C2 isomer) δ 6.95 (d, J ) 1.5 Hz, 2H), 6.81 (d, J )
(72) Shur, V. B.; Burlakov, V. V.; Vol′pin, M. E. J. Organomet. Chem.
1988, 347, 77.
(73) Kabalka, G. W.; Reddy, N. K.; Narayana, C. Tetrahedron Lett. 1992,
33, 865.
(74) Cooksey, C., J.; Land, E. J.; Riley, P. A. Org. Prep. Proced. Int.
1996, 28, 463.
(75) Ma¨nnig, D.; No¨th, H. J. Chem. Soc., Dalton Trans. 1985, 1689.
(76) Brown, H. C.; Gupta, S. K. J. Am. Chem. Soc. 1971, 93, 1816.

Titanocene Borane σ-Complexes
J. Am. Chem. Soc., Vol. 121, No. 21, 1999 5045
(-5 °C, C7H8) δ 64; 2H NMR (-5 °C, C7H8) δ -9.9; IR (Nujol, cm-1
1591 (s), 1485 (s), 1426 (s), 1349 (m), 1237 (m), 1207 (m), 1184 (br,
m), 950 (m), 800 (s).
Reaction of 2b with Phenylsilane. Into an NMR tube was weighed
9.6 mg (0.026 mmol) of 2b, and the tube was sealed with a septum.
The tube was cooled to -78 °C, and a solution of phenylsilane (0.26
mmol, 48 mg) in 0.6 mL of toluene-d8 was added to the tube by syringe.
The tube was quickly immersed into the precooled probe (-5 °C) of
the NMR spectrometer for analysis.
δ 64.2. 31P{1H} NMR (C7D8, 0 °C): δ 29.3.13C{1H} NMR (C7D8, -5
°C): δ 153.09, 120.69, 110.10, 96.69, 20.37 (this resonance is obscured
)
by toluene, but was identified by a H-13C correlation experiment;
1
the coupling constant was not determined). IR (Nujol, cm-1): 1650
(br, m), 1485 (s), 1426 (s), 1355 (s), 1237 (s), 1207 (s), 1004 (s), 960
(s), 846 (s), 806 (s).
Preparation of Cp2Ti(HBcat-3-F)(PMe3) (2c). Compound 2c was
prepared in 40% yield (13.7 mg) using a procedure similar to that for
1
the preparation of 2b. H NMR (C7D8, 0 °C): δ 6.75 (m, 1H), 6.53
(m, 2H), 4.97 (d, JH-P ) 3.0 Hz, 10H), 0.70 (d, JH-P ) 6.3 Hz, 9H),
-9.4 (br s, 1H). 11B NMR (C7D8, 0 °C): δ 64.9. 31P{1H} NMR (C7D8,
0 °C): δ 27.9.
Reaction of Cp2Ti(H2SiPh2)(PMe3) with HBcat. Into an NMR tube
was weighed 20 mg (0.046 mmol) of Cp2Ti(H2SiPh2)(PMe3). The tube
was sealed with a septum. The tube was cooled to -78 °C, and a
solution of diphenylsilane (85 mg, 0.46 mmol) in 0.4 mL of toluene-
d8 was added by syringe. A solution of catecholborane (5.5 mg, 0.046
mmol) in 0.2 mL of toluene-d8 was then added. The tube was quickly
immersed into the precooled probe (-5 °C) of the NMR spectrometer
for analysis.
Preparation of Cp2Ti(PhSiH3)(HBcat′) (3; HBcat′ ) HBO2C6H3-
4-t-Bu). A side-arm flask was charged with 1a (300 mg, 0.568 mmol),
sealed with a septum, and cooled to -78 °C. Approximately 10 mL of
toluene was then added to the flask via syringe. With stirring, the yellow
slurry was allowed to warm to -5 °C, and phenylsilane (100 µL, 0.810
mmol) was added to the mixture by syringe. Stirring was continued
for 6 h, over which time the solution became yellow-green. Solvent
was completely removed from the mixture under vacuum at -5 °C to
afford a yellow powder. The product was collected by vacuum filtration,
washed with cold pentane, and dried under vacuum (103.4 mg, 39%).
1H NMR (C7D8, -80 °C): 7.68 (d, J ) 6.3 Hz, 2H), 7.29 (s, 1H), 7.14
(m, 3H) 7.19 (d, J ) 7.8 Hz, 1H), 6.79 (d, J ) 7.8 Hz, 1H), 5.13 (br
s, 1H), 5.11 (br s, 1H), 4.99 (s, 5H), 4.94 (s, 5H), 1.29 (s, 9H), -4.99
(br s, 1H), -6.78 (br s, 1H). 11B NMR (C7D8, 0 °C): 37.1. IR (Nujol,
cm-1): 2066 (s), 1739 (s), 1773 (s), 1252 (s), 1213 (s), 1024 (s), 800
(s).
Kinetic Studies of the Reaction of PMe3 with 1a-e. Into an NMR
tube was weighed 0.0057 mmol of the appropriate titanocene borane
σ-complex, and the tube was sealed with a septum. The NMR tube
was then cooled to -78 °C by immersing it into a dry ice-acetone
bath, and 0.6 mL of toluene-d8 was added to the NMR tube by syringe.
The tube was shaken to dissolve all of the solid without allowing the
sample to warm. The tube was then reimmersed into the dry ice-
acetone bath, and 1 equiv of PMe3 was added by syringe. The tube
was quickly shaken and introduced into the precooled (-30 °C) probe
of the NMR spectrometer. Single-pulse experiments were performed
every 90 s over at least 3 half-lives using an automated program. The
experiment was repeated at -15, -20, -25, -35, and -40 °C for
values of the rate constants to be used in the Eyring plot. Rate
measurements were performed by measuring the integrals of the
cyclopentadienyl peaks of the titanocene starting material and product
by 1H NMR spectroscopy at -30 °C. For studies of the dependence of
the reaction rate on HBcat′ and PMe3 concentrations, the appropriate
amounts of phosphine and 4-tert-butylcatecholborane were added by
syringe to an NMR tube containing 0.0057 mmol of 1a before the tube
was introduced into the NMR spectrometer. Measurements were made
with 0.0945 mmol (5 equiv) of added 4-tert-butylcatecholborane and
0.0284, 0.0471, and 0.0853 mmol of added PMe3. Measurements were
also made with 0.0341 mmol (6 equiv) of PMe3 and 0.0170, 0.0341,
and 0.0517 mmol of added 4-tert-butylcatecholborane.
Reaction of 1a with CO. Into a vial was weighed 3.0 mg (0.0057
mmol) of 1a. The vial was cooled to -30 °C. To this vial was added
0.6 mL of cold toluene-d8, and the resulting yellow solution was kept
at -30 °C for 5 min. The solution was then quickly transferred to an
NMR pressure tube, and the tube was sealed. The tube was returned to
the freezer for another 5 min, after which time it was rapidly removed
from the drybox and fully immersed into a Dewar flask containing a
dry ice-acetone mixture. The tube was then connected to a tank of
carbon monoxide using a Swage-lock adapter without removing it from
the dry ice-acetone bath. One atm of CO gas was added to the tube,
and the tube was sealed again. The tube was then shaken and rapidly
introduced into the precooled NMR spectrometer. Rate measurements
were made by monitoring the Cp resonances in 1a and the product,
Cp2Ti(CO)2. The experiment was repeated with 3 atm of CO gas to
study the dependence of the reaction rate on CO concentration. Single-
pulse experiments were performed every 90 s over at least 3 half-lives
at -30 °C.
Preparation of Cp2Ti(PhSiH3)(DBcat′). This compound was
prepared in 74% yield (214.7 mg) using the method described for the
synthesis of 3 and DBcat′. 2H NMR: δ -6.78, -4.98, 4.90. 11B
NMR: δ 37.0. IR (Nujol, cm-1): 2093 (s), 2053 (s), 1556 (s), 1466
(s), 1380 (m), 1299 (m), 1220 (m), 1100 (s), 780 (s), 600 (s).
Generation of Cp2Ti(PhCCPh)(HBcat′) (4; HBcat′ ) HBO2C6H3-
4-t-Bu) in Situ. A solution of Cp2Ti(PhCCPh) (5.0 mg, 0.014 mmol)
in approximately 0.6 mL of toluene-d8 was placed into an NMR tube,
and the tube was sealed with a septum. The tube was cooled to -78
°C, and 4-tert-butylcatecholborane (2.5 µL, 0.014 mmol) was added
by syringe. The tube was then quickly introduced into the precooled
1
probe of the NMR spectrometer for analysis. H NMR (C7D8, -60
°C): δ 7.07 (m, 6H), 7.03 (d, J ) 1.9 Hz, 1H), 6.89 (d, J ) 8.1 Hz,
1H), 6.83 (m, 4H), 6.67 (dd, J ) 8.1, 1.9 Hz, 1H), 5.44 (s, 5H), 5.43
(s, 5H), 1.25 (s, 9H), -1.59 (br s, 1H). 11B NMR (C7D8, -30 °C): δ
29.8. 13C{1H} NMR (C7D8, -60 °C): δ 207.7, 153.8, 151.8, 141.9,
141.1, 140.1, 129.8, 128.6, 127.9, 126.2, 115.9, 107.9, 34.5, 31.8. (five
aromatic carbons are obscured by toluene solvent).
Reaction of 1c with B-Deuterio-4-methylcatecholborane. Into an
NMR tube was weighed 10 mg (0.022 mmol) of 1c, and the tube was
sealed with a septum. The tube was cooled to -78 °C, and a solution
of B-deuterio-4-methylcatecholborane (31.6 mg, 0.22 mmol) in 0.6 mL
of toluene was added into the tube by syringe. The tube was quickly
immersed into the precooled probe (-5 °C) of the NMR spectrometer
for analysis. Spectroscopic analyses of 1c-d2: 11B NMR (-5 °C, C7H8)
2
δ 46; H NMR (-5 °C, C7H8) δ -5.5.
Reaction of 1c with Catecholborane. Into a vial were weighed 10
mg (0.022 mmol) of 1c and 1 equiv of ferrocene as an internal standard.
Catecholborane (52 mg, 0.44 mmol) in 0.6 mL of toluene-d8 was added
into the tube by syringe. A yellow solid (1g) was deposited. A 1H NMR
spectrum was obtained of the supernatant to determine the yield of
extruded 4-methylcatecholborane.
Ligand Substitution Reactions of 2b. Into an NMR tube was
weighed 10 mg (0.027 mmol) of 2b, and the tube was sealed with a
septum. The tube was cooled to -78 °C, and a solution of reagent in
0.6 mL of toluene-d8 was added into the tube by syringe. The tube
was quickly immersed into the precooled probe (-5 °C) of the NMR
spectrometer for analysis. Experiments were performed using trimeth-
ylphosphine-d9, B-deuteriocatecholborane, and 3-fluorocatecholborane
as reagents to give Cp2Ti(HBcat)(PMe3-d9), Cp2Ti(DBcat)(PMe3), and
2c, respectively. Spectroscopic analyses are as follows. Cp2Ti(HBcat)-
(PMe3-d9): 11B NMR (-5 °C, C7H8) δ 64; 2H NMR (-5 °C, C7H8) δ
0.80; 31P NMR (-5 °C, C7H8) δ 26.5. Cp2Ti(DBcat)(PMe3): 11B NMR
Reaction of 1a with PhSiH3. Into an NMR tube was weighed 3.0
mg (0.0057 mmol) of 1a, and the tube was sealed with a septum. The
NMR tube was then cooled to -78 °C by fully immersing it into a dry
ice-acetone bath, and 0.6 mL of toluene-d8 was added to the NMR
tube by syringe. The tube was shaken to dissolve all of the solid without
allowing it to warm. The tube was then reimmersed into the dry ice-
acetone bath, and 1.4 µL (0.011 mmol) of PhSiH3 was added by syringe.
The tube was quickly shaken and introduced into the precooled (-30
°C) probe of the NMR spectrometer. Rate measurements were
performed by monitoring the integrals of the cyclopentadienyl peaks
of the titanocene starting material and product by 1H NMR spectroscopy
at -30 °C. Single-pulse experiments were performed every 90 s over
at least 3 half-lives.

5046 J. Am. Chem. Soc., Vol. 121, No. 21, 1999
Muhoro et al.
unweighted and weighted agreement factors of R ) 5.7% and Rw
6.4%. The crystallographic data for 1g were deposited in the Cambridge
Crystallographic Data Centre under reference code TUBSUE as part
of a previous communication.39
Reaction of PhCCPh with 1a. Into an NMR tube were weighed
3.0 mg (0.0057 mmol) of 1a and 3.0 mg (0.017 mmol) of dipheny-
lacetylene, and the tube was sealed with a septum. The NMR tube was
then cooled to -78 °C by fully immersing it into a dry ice-acetone
bath, and 0.6 mL of toluene-d8 was added to the NMR tube by syringe.
The tube was shaken to dissolve all of the solid without allowing it to
warm. The NMR tube was introduced into the precooled (-30 °C)
probe of the NMR spectrometer. Rate measurements were performed
by monitoring the integrals of the cyclopentadienyl peaks of the
)
Crystallographic Analysis of 2c. Dark red prism-like single crystals
of 2c that were suitable for X-ray diffraction studies were obtained by
cooling a concentrated toluene solution of 2c to -30 °C for several
days. X-ray data for 2c were measured at 153 K with a Mo-target X-ray
tube (λ ) 0.710 73 Å). Crystal data are provided in Tables 3 and 4. A
total of 1321 frames were collected with a scan width of 0.3° in ω and
an exposure time of 30 s/frame. The frames were integrated with the
Siemens SAINT software package using a narrow-frame integration
algorithm. The integration of the data using a triclinic unit cell yielded
a total of 5507 reflections to a maximum 2θ angle of 45°, 2329 of
which were independent and 1943 of which were greater than 4σ(F).
The final cell constants are based upon the refinement of the XYZ-
centroids of 3685 reflections above 2σ(I). The data were corrected for
absorption using the SADABS program with minimum and maximum
transmission coefficients of 0.564 and 0.928, respectively. The structure
was solved and refined using the Siemens SHELXTL (version 5.0)
software package, using space group P1h, with Z ) 2. Final anisotropic
full-matrix least-squares refinement on F2 converged at R1 ) 6.38%,
wR2 ) 16.31%, and a goodness of fit of 1.024. The hydrogen atoms
were located and refined. The crystallographic data for 2c were
deposited in the Cambridge Crystallographic Data Centre with code
CCDC-100125 as part of a previous communication.40
1
titanocene starting material and product by H NMR spectroscopy at
-30 °C. Single-pulse experiments were performed every 90 s over at
least 3 half-lives.
Kinetic Studies of the Formation of Vinyl Boronate Ester from
4. Into a vial was weighed 5.0 mg (0.014 mmol) of Cp2Ti(PhCCPh),
and 0.5 mL of toluene-d8 was added. The solution was transferred to
an NMR tube, and the tube was sealed with a septum. The solution
was then immersed into a dry ice-acetone mixture and allowed to cool.
To the NMR tube was then added 0.042 mmol of the appropriate
catecholborane by syringe. The tube was rapidly shaken and introduced
into the precooled probe (-30 °C) of the NMR spectrometer. Rate
measurements were performed by monitoring the integrals of the vinylic
protons of the vinyl boronate ester product. Single-pulse experiments
were performed every 90 s over at least 3 half-lives.
Crystallographic Analysis of 1g. Pale yellow single crystals of 1g
that were suitable for X-ray diffraction studies were obtained by cooling
a highly dilute solution of Cp2TiMe2 and catecholborane to -30 °C
for several days. The X-ray intensity data were measured at 195 K
with graphite monochromatic Mo KR radiation. Crystal data are
provided in Tables 1 and 2. For Z ) 2 and FW ) 510.06, the calculated
density is 1.361 g/cm3. The space group was determined to be P21/m
(no. 11). The structure was solved by direct methods using the SHELXS
software package. The non-hydrogen atoms were refined anisotropically.
The hydrogen atoms were observed in the difference map, and H1 was
isotropically refined. The hydrogens are included in idealized positions,
except for H1 which was included in the difference map position. The
final cycle of full matrix least squares refinement was based on 1552
observed reflections (I > 3.00σ(I)) and 178 variable parameters and
converged (largest parameter shift was 0.00 times its esd) with
Acknowledgment. This work was supported by the National
Science Foundation. We thank the Dreyfus Foundation for a
Camille Dreyfus Teacher-Scholar Award. J.F.H. is an Alfred
P. Sloan Research Fellow. In addition, we thank Susan De Gala
for the X-ray structural analysis of 1g and Charles Campana at
Siemens for the X-ray analysis of 2c. We are grateful to Odile
Eisenstein for her insightful analysis on several aspects of this
work.
JA9900757