Synthesis and crystal structure of some novel titanocene phosphido compounds by P-H activation in the presence of hydrosilanes

Article abstract of DOI:10.1021/ja960657d

The reaction of dimethyltitanocene (Cp₂TiMe₂) with PhSiH₃ in hexane in the presence of Cy₂PH (Cy = cyclohexyl) is photoinitiated to give paramagnetic Cp₂Ti(SiH₂Ph)(PHCy₂), characterized in solution by EPR spectroscopy (a(H) of phosphine = 4.8 G, a(P) = 26.9 G, g(iso) = 1,9950). Under the same conditions, a similar reaction with Ph₂SiH₂ gives a crystalline solid, shown by an X-ray structure determination to be [(Cp₂Ti)₂(μ-H)(μ-PCy₂)] (1). Under continuous photolysis Cp₂TiMe₂ reacts with CyPH₂, in the presence of either PhSiH₃ or Ph₂SiH₂, to give crystals of [Cp₂Ti(μ-PHCy)]₂ (2). rac-[(EBTHI)Ti(μ-H)]₂ (EBTFII = ethylene-1,2-bis(η⁵-4,5,6,7-tetrahydro-1-indenyl)) reacts with PhPH₂ to give crystals of the complex (EBTHI)Ti(P₂Ph₂) (4). In this reaction, excess PhPH₂ is polymerized by dehydrocoupling to a mixture of linear and cyclic oligophosphanes. The diphosphane is produced as an equal mixture of the two diastereoisomers. Due to precipitation of the highly insoluble 4, the catalytic reaction eventually ceases. The mechanisms of these reactions are discussed in terms of the oxidative addition of phosphines to Cp₂Ti(II), conproportionation of Cp₂Ti(II) and Cp₂Ti(IV) to Cp₂Ti(III), and σ-bond metethesis reactions of P-H, Ti-H, and Ti-P compounds.

Full text of DOI:10.1021/ja960657d

J. Am. Chem. Soc. 1997, 119, 5307-5313
5307
Synthesis and Crystal Structure of Some Novel Titanocene
Phosphido Compounds by P-H Activation in the Presence of
Hydrosilanes
Shixuan Xin,† Hee Gweon Woo,†,‡ John F. Harrod,*,† Edmond Samuel,*,§ and
Anne-Marie Lebuis†
Contribution from the Chemistry Department, McGill UniVersity, Montreal, Canada H3A 2K6,
and Laboratoire de Chimie Organome´tallique de l’ENSCP (URA 403 CNRS),
11 rue P. et M. Curie, 75005 Paris, France
ReceiVed February 28, 1996X
Abstract: The reaction of dimethyltitanocene (Cp2TiMe2) with PhSiH3 in hexane in the presence of Cy2PH (Cy )
cyclohexyl) is photoinitiated to give paramagnetic Cp2Ti(SiH2Ph)(PHCy2), characterized in solution by EPR
spectroscopy (a(H) of phosphine ) 4.8 G, a(P) ) 26.9 G, giso ) 1.9950). Under the same conditions, a similar
reaction with Ph2SiH2 gives a crystalline solid, shown by an X-ray structure determination to be [(Cp2Ti)2(µ-H)(µ-
PCy2)] (1). Under continuous photolysis Cp2TiMe2 reacts with CyPH2, in the presence of either PhSiH3 or Ph2SiH2,
to give crystals of [Cp2Ti(µ-PHCy)]2 (2). rac-[(EBTHI)Ti(µ-H)]2 (EBTHI ) ethylene-1,2-bis(η5-4,5,6,7-tetrahydro-
1-indenyl)) reacts with PhPH2 to give crystals of the complex (EBTHI)Ti(P2Ph2) (4). In this reaction, excess PhPH2
is polymerized by dehydrocoupling to a mixture of linear and cyclic oligophosphanes. The diphosphane is produced
as an equal mixture of the two diastereoisomers. Due to precipitation of the highly insoluble 4, the catalytic reaction
eventually ceases. The mechanisms of these reactions are discussed in terms of the oxidative addition of phosphines
to Cp2TiII, conproportionation of Cp2TiII and Cp2TiIV to Cp2TiIII, and σ-bond metathesis reactions of P-H, Ti-H,
and Ti-P compounds.
Introduction
reactive species, and thus contributing to the diversity of
catalytic properties, makes the isolation and definitive identifica-
tion of reaction intermediates difficult. Consequently, the
definition of reaction pathways in titanocene chemistry is a
considerable challenge.
Among the cyclopentadienyl compounds of the transition
metals, metallocenes of group 4 exhibit a particularly wide range
of activities covering various stoichiometric and catalytic
reactions.1-11 Although there are many similarities between the
chemistries of the three metallocenes of this group, titanocene
derivatives exhibit a distinctive and rich chemistry in that all
of the common oxidation states II, III, and IV are represented
by an important number of isolated and well-characterized
compounds. In particular, the more or less easy access to the
lower oxidation states is due to the relatively low Ti(IV)/Ti-
(III) reduction potential, compared to the other members of the
group, and also to the tendency of Ti(III) to disproportionate.12
This particularity, while increasing the number of possible
The recent implication of these metallocenes in Si-H bond
activation opened a new and important area of interest.3-10
Parallel to the development of useful catalytic reactions, a
number of Ti(III) compounds with Ti-Si bonds have been
isolated and fully characterized.4,13-16 Their active role as
catalyst intermediates was supported by trapping them in a stable
form with tertiary phosphines (e.g., [Cp2TiIII(PMe3)(SiH2Ph)]).14
The isolation and characterization of Cp2Ti(PMe3)Ph2SiH2) by
Buchwald and co-workers,16 under conditions similar to those
which produced the latter silylphosphine complex, illustrates
the sensitive dependence of the product on reaction conditions.
* Author To whom correspondence should be addressed. Phone: 514-
398-6911. FAX: 514-398-3797. E-mail: harrod@omc.lan.mcgill.ca.
† McGill University.
‡ Present address: Department of Chemistry, Chonnam University,
Chonnam, Republic of Korea.
The successful use of tertiary phosphines to stabilize non-
bridging Ti-Si bonds raises the question as to whether a primary
or secondary phosphine would behave in the same way, or
whether a competitive P-H bond activation would intervene,
leading to different types of products. This question has been
partially addressed by Stephan and co-workers,17a,b who obtained
a phosphido-bridged dimer with the system Cp2ZrCl2/Mg/PCy2H
§ ENSCP.
X Abstract published in AdVance ACS Abstracts, May 1, 1997.
(1) Brintzinger, H. H.; Fischer, D.; Mu¨lhaupt, R.; Rieger, B.; Waymouth,
R. M. Angew. Chem., Int. Ed. Engl. 1995, 34, 1143.
(2) Willoughby, C. A.; Buchwald, S. L. J. Am. Chem. Soc. 1994, 116,
11703.
(3) Broene, R. D.; Buchwald, S. L. J. Am. Chem. Soc. 1993, 115, 12659.
(4) (a) Aitken, C.; Harrod, J. F.; Samuel, E. J. Am. Chem. Soc. 1986,
108, 4059. (b) Tilley, T. D. Acc. Chem. Res. 1993, 26, 22.
(5) Xin, S.; Aitken, C.; Harrod, J. F.; Mu, Y.; Samuel, E. Can. J. Chem.
1990, 68, 471.
(6) Xin, S.; Harrod, J. F. J. Organomet. Chem. 1995, 499, 181.
(7) Carter, M. B.; Schiott, B.; Gutie´rrez, A.; Buchwald, S. L. J. Am.
Chem. Soc. 1994, 116, 11670.
(13) Gauvin, F.; Harrod, J. F.; Samuel, E.; Britten, J. J. Am. Chem. Soc.
1992, 114, 1489.
(14) Samuel, E.; Mu, Y.; Harrod, J. F.; Dromzee, Y.; Jeannin, Y. J. Am.
Chem. Soc. 1990, 112, 3435.
(15) Xin, S.; Harrod, J. F.; Samuel, E. J. Am. Chem. Soc. 1994, 116,
11563.
(8) Xin, S.; Harrod, J. F. Can. J. Chem. 1995, 73, 999.
(9) Grasczyk, T.; Harrod, J. F. J. Polym. Sci., Part A 1994, 32, 3183.
(10) Rousset, C. J.; Fanwick, P. E.; Negishi, E. J. Am. Chem. Soc. 1991,
113, 8564.
(11) Petasis, N. A.; Fu, D.-K. J. Am. Chem. Soc. 1993, 115, 7208.
(12) Battaglia, L. P.; Nardelli, M.; Pelizzi, C. J. Organomet. Chem. 1983,
259, 301.
(16) Spaltenstein, E.; Palma, P.; Kreutzer, K. A.; Willoughby, C. A.;
Davis, W. M.; Buchwald, S. L. J. Am. Chem. Soc. 1994, 116, 10308.
(17) (a) Ho, J.; Dick, D. G.; Stephan, D. W. Organometallics 1991, 10,
3001. (b) Ho, J.; Hou, Z.; Drake, R. J.; Stephan, D. W. Organometallics
1993, 12, 3145. (c) Ho, J.; Stephan, D. W. Inorg. Chem. 1994, 33, 4595.
(d) Ho, J.; Breen, T. L.; Ozarowski, A.; Stephan, D. W. Inorg. Chem. 1994,
33, 865.
S0002-7863(96)00657-9 CCC: $14.00 © 1997 American Chemical Society

5308 J. Am. Chem. Soc., Vol. 119, No. 23, 1997
Xin et al.
where previously Cp2Zr(PMe3)2 had been obtained with PMe3.18
On the other hand, extensive studies by several groups have
shown that a range of interesting phosphine, phosphido, and
phosphinidine complexes of group 4 metallocenes can be
prepared16,19-25 by classically known methods which consist of
reacting Li or K salts of dialkylphosphides with Cp2MIV
dihalides, or with Cp2MII species generated in situ.
In this paper we report on a new approach to the synthesis
of P-ligated titanocene derivatives with primary and secondary
phosphines in the presence of silanes as reducing agents. This
approach has also led to the discovery of a catalytic dehydro-
coupling reaction of PhPH2 to give Ph2P2H2 and higher
oligophosphines.
Figure 1. ORTEP41 diagram of [Cp2Ti]2(µ-H)(µ-PCy2) (1) at 30%
probability. Bond lengths (Å): Ti(1)-Ti(2) 3.456(4); Ti(1)-H 2.082;
Ti(2)-H 2.018; Ti(1)-P 2.643(5); Ti(2)-P 2.643(4). Bond angles
(deg): Ti(1)-P-Ti(2) 81.7; Ti(1)-P-C(1) 117.2; H-Ti(1)-P 81;
H-Ti(2)-P 82; Ti(1)-H-Ti(2) 115.
Results and Discussion
It has been shown previously that the light-assisted reaction
of Cp2TiMe2 with primary silanes proceeds vigorously with
evolution of methane and hydrogen and the formation of
strongly paramagnetic solutions which, in the presence of tertiary
phosphines, yield crystalline titanocene(III) silyl(phosphine)
compounds.14 Since the P-H bond dissociation energy is
similar to that of Si-H, it was of interest to investigate whether
a primary or secondary phosphine would influence the course
of the reaction by direct or indirect involvement of the P-H
fragment of the phosphine. At the outset it was established
experimentally that these phosphines do not react significantly,
either thermally or photochemically, with Cp2TiMe2 alone.
Reaction of Cy2PH with Cp2TiMe2 in the Presence of
Silanes. (a) Synthesis and Structure of [(Cp2Ti)2(µ-H)(µ-
PCy2)] (1). The reaction of DMT with PhSiH3 in hexane/
toluene solution in the presence of Cy2PH under UV-visible
light is accompanied immediately by gas evolution and the
formation of a dark violet solution from which, unlike a number
of tertiary phosphine reactions14,26a no solid compound can be
isolated. This is probably because the Cy groups confer high
solubility on the anticipated product, Cp2Ti(SiH2Ph)(PHCy2).
Formation of this product was supported by monitoring the
appearance of its intense EPR spectrum, consisting of a doublet
of pseudo-triplets. This signal can be reliably attributed to
coupling of the unpaired electron on Ti with one P nucleus and
two slightly inequivalent hydrogen nuclei. The hyperfine pattern
is temperature dependent in the range 220-290 K, which
indicates that molecular tumbling is not completely averaged
within this temperature range, most probably because of the
bulky Cy groups.
deuterium substitution, the residual coupling can be confidently
attributed to hyperfine interaction with the hydrogen atom of
the phosphine ligand. It is interesting to note that no scrambling
occurs between silyl deuterium and hydrogen on phosphorus.
Formation of a compound such as Cp2Ti(SiHDPh)(PDCy2)
would give a completely different spectral pattern so that it can
be safely assumed that the spectrum is due to Cp2Ti(SiD2Ph)-
(PHCy2). It should be mentioned here that observation of
hyperfine interactions with hydrogen of a phosphine ligand is
very rare in paramagnetic transition metal compounds.
In order to isolate a crystalline reaction product, we sought
to reduce the solubility by using Ph2SiH2 instead of PhSiH3.
Under the same conditions as above, well-shaped dark-violet
crystals of [(Cp2Ti)2(µ-H)(µ-PCy2)] (1) were deposited. An
X-ray structure determination of 1 revealed the structure shown
in Figure 1.
Compound 1 can be formally considered to be the association
product of either titanococene(III) hydride with a titanocene-
(III) phosphide or a titanocene(IV) (hydrido)phosphide with
titanocene(II). This type of ambivalence was previously
discussed in relation to the structure of isoelectronic Si-H
bridged titanocene dimers.4a
1 is almost insoluble in hexanes and is very poorly soluble
in toluene. However, measurement at 253 K in the latter solvent
showed that the compound is EPR silent, which confirms its
diamagnetism at this temperature. The same sample, after being
held at room temperature for ca. 30 min, exhibits a strong signal
consisting of a singlet (gav ) 1.9808; a(Ti) ) 9.43 G). In frozen
toluene at 140 K, the g tensor anisotropy is cleanly resolved
(g1 ) 2.0020; g2 ) 1.9853; g3 ) 1.9539). These results show
that 1 dissolves at low temperature as a diamagnetic dimer, but
decomposes at room temperature to give a paramagnetic Ti-
(III) monomer. The singlet observed by EPR at 253 K cannot
be attributed to either of the two possible Ti(III) components
of the dimer, namely Cp2TiH or Cp2TiPCy2. The exact
formulation of this species is not possible because of the absence
of any ligand hyperfine interaction. However, the non-axial
symmetry of the frozen solution g tensor and the parameters
show that it is a monomeric titanocene(III) entity of relatively
simple structure, in agreement with previous data on this class
of compounds.26b
When the reaction is performed with PhSiD3, the spectrum
simplifies to a doublet of doublets and can only be due to
coupling of the unpaired electron with one P nucleus (a(P) )
26.9 G; 1.0 G ) 0.1 mT) and with one hydrogen atom (a(H) )
4.8 G). Assuming the usual 6-fold reduction in the hyperfine
coupling to the two hydrogen atoms of the silane resulting from
(18) Payne, R.; Hachgenei, J.; Fritz, G.; Fenske, D. Naturforsch. 1986,
41b, 1535.
(19) (a) Baker, R. T.; Calabrese, J. C.; Harlow, R. L.; Williams, I. D.
Organometallics 1993, 12, 830. (b) Fermin, M. C.; Stephan, D. W. J. Am.
Chem. Soc. 1995, 117, 12645.
(20) Isslieb, K.; Hackert, H. Naturforsch. 1966, 21b, 519.
(21) Isslieb, K.; Willie, G.; Krech, F. Angew. Chem., Int. Ed. Engl. 1972,
11, 527.
(22) Kopf, H.; Voigtlander, R. Chem. Ber. 1981, 114, 2731.
(23) Hey, E.; Bott, S. G.; Atwood, J. L. Chem. Ber. 1988, 121, 561.
(24) Benac, B. L.; Jones, R. A. Polyhedron 1989, 8, 1774.
(25) (a) Dick, D. G.; Stephan, D. W. Can. J. Chem. 1991, 69, 1146. (b)
Dick, D. G.; Stephan, D. W. Organometallics 1991, 10, 2811. (c) Drake,
R. J.; Stephan, D. W. Organometallics 1993, 12, 3145.
(26) (a) Woo, H. G.; Harrod, J. F.; He´nique, J.; Samuel, E. Organome-
tallics 1993, 12, 2883. (b) Mach, K.; Barrie Raynor, J. J. Chem. Soc.,
Dalton Trans. 1992, 683, and references therein.
The structure of the mixed hydride-phosphide-bridged com-
pound 1 is unique among early transition metal phosphides. It
is, on the other hand, isostructural with isoelectronic mixed
hydride-silyl-bridged titanocene complexes reported earlier.4 The
bond distances within the Ti2(µ-X)(µ-H) unit are indistinguish-

NoVel Titanocene Phosphido Compounds
J. Am. Chem. Soc., Vol. 119, No. 23, 1997 5309
able for X ) (µ-SiHPh)(µ-H) and Cy2P. The angles appear
superficially to be different, but when the large standard
deviations for angles involving H atoms are taken into account,
they are also indistinguishable.
(b) Synthesis and Structure of [Cp2Ti(µ-PHCy)]2 (2). The
reactions of DMT with CyPH2 in the presence of either PhSiH3
or Ph2SiH2 proceed in the same manner, but differently from
either of the reactions of Cy2PH described above. They take
place under sustained photolysis, as opposed to the photoini-
tiation required in the case of tertiary phosphines. The color
of the solution is light brown in contrast to the very dark violet
color observed in the latter case, and gas evolution is rather
sluggish. After a while, dark-brown microcrystals of [(Cp2Ti)-
(µ-PHCy)]2 (2) begin to deposit on the walls of the Schlenk
tube facing the light source.
Compound 2, formally a Ti(III) dimer, gave no EPR spectrum
in toluene solution, probably due to strong antiferromagnetic
coupling between the two unpaired electrons on the two Ti
atoms. However, the dimeric structure seems to be labile in
the presence of strong donors, as is often the case with Ti(III)
dimers.17a,25b Thus PMe3 reacts with 2 in either THF or toluene
to give a solution that exhibits an intense EPR signal at room
temperature consisting of a doublet (a(P) ) 20.0 G, a(Ti) )
10.5 G, g ) 1.9867) identical with that previously assigned to
Cp2Ti(CH2PMe2).14 If the solution is left standing at room
temperature for a while, another signal develops of slightly lower
intensity, composed of a doublet of doublets (a(P) ) 16 G; a(P)
) 3.6 G; g ) 1.9828). This signal is tentatively assigned to
the phosphide phosphine complex Cp2Ti(PHCy)(PMe3). The
splitting of titanocene(III) phosphide dimers by tertiary alkyl-
phosphines has been reported by Dick and Stephan, and the
structure of one such product, Cp2Ti(PPh2)(PMe3), was con-
firmed by X-ray crystallography.25b The low value of the
hyperfine coupling to the phosphide P was attributed to the
nonplanarity of the phosphorus coordination in this type of
compound. In the monomeric complex Cp2Ti(PCy2), the
phosphide group is planar and the hyperfine coupling constant
is larger.27
Figure 2. Thermal ellipsoid plot of [Cp2Ti(µ-P(H)Cy)]2 (2).
Figure 3. ORTEP diagram of a molecule of complex 4, at 30%
probability. Bond lengths (Å): Ti-P 2.525(2); P-P′ 2.174(3). Bond
angles (deg): P-Ti-P′ 51.00(6); Ti-P-C(11) 117.8(2); Ti-P-P′
64.50(3); P-P′-C(11) 107.2(2); C(11)-P-P′-C(11)′ 133.9(5).
The extremely thin plates of 2 were marginally adequate for
single-crystal X-ray structure determination, even collecting the
X-ray data with a rotating anode source. The structure, shown
in Figure 2, was solved with the use of the best data set, and
this confirmed that 2 is a titanocene dimer containing Ti2P2
bridging units. This formulation is also in agreement with the
elemental analysis. Despite extensive disorder in the P-Cy
fragment and in the Cp rings, the resolved structure clearly
supports the conclusion that 2 is the titanocene(III) phosphide
dimer. This conclusion is based mainly on the fact that the
phosphorus coordination is clearly pyramidal with respect to
its Ti2Cy ligand set, since the hydrogen atom on the phosphine
was not observed.
contain a Ti-Si bond. Pursuing these investigations further,
we also recently described the preparation of rac-[(EBTHI)-
TiH]2 (3; EBTHI ) ethylene-1,2-bis(η5-4,5,6,7-tetrahydro-1-
indenyl), a paramagnetic dimer, by reaction of [EBTHI]TiMe2
with primary and secondary silanes.15 Similar reactions carried
out in the presence of phosphines were not conclusive except
for PhPH2. Addition of 10 equiv of this phosphine to a freshly
prepared solution of 3 in toluene at 40-50 °C gave well-formed,
diamond-shaped, green crystals of rac-[ethylene-1,2-bis(η5-
4,5,6,7-tetrahydro-1-indenyl)(1,2-diphenyldiphosphene)titanium-
(IV)] (4) in high yield. The product is diamagnetic and insoluble
in the common nonhalogenated solvents, a factor that impeded
the study of its NMR properties and its reactivity in solution.
The absence of a vibrational band in the 2100-2350-cm-1
region in its FT-IR spectrum is indicative of the absence of a
P-H or a Ti-H bond.
Single-crystal X-ray structural analysis confirmed that 4 is a
C2-symmetric molecule with one 1,2-diphenyldiphosphene
ligand coordinated to an [EBTHI]Ti unit. The unit cell contains
two pairs of racemic molecules. The molecular structure of 4
(Figure 3) shows that the [EBTHI]Ti fragment retains the same
structure as the precursor molecule 3.15 The most notable
feature of the structure is the opposite configurations at the C(1)
and P atoms of each molecule, i.e., an R,R (EBTHI)Ti unit is
bound to an S,S diphenyldiphosphene ligand, and vice versa.
The molecule shown in Figure 4 is the R,R,S,S molecule.
Reaction of rac-[(EBTHI)Ti(µ-H)]2 (3) with PhPH2. Syn-
thesis and X-ray Structure of rac-[(EBTHI)TiP2Ph2] (4). In
a previous report we showed that substituting the (η5-Cp) ring
with (η5-indenyl) results in quite different chemistry,13 producing
a mixed valence Ti(III)/Ti(IV) hydride dimer that does not
(27) Baker, R. T.; Whitney, J. F.; Wreford, J. S. Cited in ref 22b.
(28) Harrod, J. F.; Mu, Y.; Samuel, E. Can. J. Chem. 1992, 70, 2980.
(29) We have previously reported that triphenylgermane undergoes a slow
σ-bond metathesis with DMT under certain conditions, to give a Ti(IV)
germylmethyl complex. Under other conditions it undergoes the rapid
branched chain reaction to produce reduced titanocene species. An
equivalent of the methylgermane or the methylsilane is always produced
when a titanocene(IV) is reduced by a germane or silane. Aitken, C.;
Harrod, J. F.; Malek, A.; Samuel, E. J. Organomet. Chem. 1988, 349, 285.
Harrod, J. F.; Malek. A.; Rochon, F. D.; Melanson, R. Organometallics
1987, 6, 2117.

5310 J. Am. Chem. Soc., Vol. 119, No. 23, 1997
Xin et al.
Scheme 1 Proposed Mechanism for the Formation of 1 and
2
•
Figure 4. 31P{H} NMR spectrum of the mother liquor from the reaction
of 3 with phenylphosphine. Inset: Proton-coupled 31P NMR spectrum
of the diphenyldiphosphine.
silanes is the formation of a Cp2TiIV(silylmethyl) complex,29
which then forms titanocene(II) as a reactive intermediate,
possibly by reductive elimination of Si-Me. Moreover, in the
course of the present work it has been shown that Cp2Ti(PMe3)2
reacts cleanly with phenylsilane to give Cp2Ti(SiH2Ph)(PMe3),
detected by its characteristic EPR spectrum.14 With this
experimental evidence in mind, a mechanism for the formation
of the products observed in the reactions of PhSiH3 and Ph2-
SiH2 with DMT in the presence of Cy2PH is shown in Scheme
1. The silylphosphine complex a is formed through steps A,
B, and C. When R ) H, the reaction does not go beyond step
C and a is the only product identified. On the other hand, when
R ) Ph, a is unstable and the reaction proceeds further through
sequence D. A plausible explanation is that excessive steric
encumbrance, resulting from the presence of a second phenyl
group on the silicon, facilitates decomposition of the silylphos-
phine complex a, possibly by homolysis of the Ti-Si bond.
This homolysis gives Cp2TiII(PCy2H), which then undergoes
intramolecular oxidative addition to produce the titanocene(IV)
complex, b. The latter then captures a titanocene(II) molecule
to give the observed product, 1.
The role of light is an important factor in all of the reactions
reported in this study. The reaction with Cy2PH described
above, in common with most reactions of DMT with silanes,
requires photoinitiation, which leads to the rapid, autocatalytic
reduction of DMT to lower oxidation states. In contrast, the
necessity for continuous photolysis in the reaction with CyPH2,
and the different product obtained, suggests that a different
mechanism is operating in this case. In a complex, multi-step
reaction of this kind, photochemistry may intervene at various
stages. But given our present ignorance of many details of the
overall reaction process, other more complex pathways could
very likely be involved. However, we suggest that compound
a is formed in the same manner (R′ ) H), which then undergoes
a photoinduced dehydrosilation, sequence E, and dimerization
to give 2.
The Ti-P bond distance of 2.525(2) Å is shorter than those
seen in related Ti(III) species [Cp2Ti(µ-PR2)]2 (Ti-Pav
)
2.618(3) Å, R ) Me,18 Et25a), Cp2Ti(PPh2)(PMe3) (Ti-P )
2.658(3) Å)23b or [Cp2Ti(PPh2)]- (Ti-Pav ) 2.69(1) Å),25b but
it is similar to the fulvalene-bridged Ti(III) dimer [CpTi(µ-
PH2)]2(µ-C10H8)(Tiav ) 2.513 (5) Å).17b This is consistent with
its Ti(IV) character. The P-P bond distance is 2.174(3) Å,
which is longer than in the anionic complex [Cp2Zr(PPh)2Br]-
(5, P-P ) 2.145(3) Å),17c and slightly longer than in the
vanadium complex Cp2V(PPh)2 (6, P-P ) 2.160(4) Å).17c These
data suggest a stronger Ti-P bond and a weaker P-P bond
compared to those in the related ZrIVP2 and VIVP2 complexes,
and are consistent with the 16-electron count about Ti in 4
compared to the 18-electron Zr in 5 and 17-electron V in 6.
The P-Ti-P bond angle of 51.00(6)° in 4 is similar to the
P-V-P angle in 6 (50.3(1)°) and the P-Mo-P angle (49.9-
(2)°) in Cp2Mo(PH)2,30 but larger than the P-Zr-P in 5
(46.6°),17c typical of such MP2 three-membered rings. The
P-P-C(11) bond angle of 107.2(2)° and the C(11)-P-P′-
C(11)′ torsion angle of 133.6(3)° in 4 are consistent with a cyclo-
propane-type, rather than a cyclopropene-type, bonding mode.
The structure of 4 is unique for a number of reasons: (i) it
is the first example of a TiIVP2 three-membered-ring complex
and the first diphosphanato complex possessing a chiral ansa
ligand; (ii) the highly strained three-membered P-P-Ti ring
is normally not expected to be preferred over the less strained
four-membered, bimetallic ring product in this type of reaction;
and (iii) a ligand-directed chirality induction is clearly indicated
on the configuration of the diphosphanato ligand. Although
polyphosphanato ligands usually adopt the trans orientation of
their organic substituents to minimize their mutual repulsion,
the dependence of the absolute configuration of the ligand on
the configuration of the [EBTHI]Ti fragment is a prominent
feature of 4. A similar effect has recently been observed in the
structure of (EBTHI)ZrBz2 (Bz ) benzyl).31
Reaction Mechanisms. Formation of 1 and 2. In earlier
publications we have discussed at some length the kinds of
species that are generated when dimethyltitanocene is reacted
with organosilanes, either in the presence or absence of
phosphines.26,28 Based on the present and earlier results it is
reasonable to assume that the initial reaction of Cp2TiMe2 with
The Catalytic Dehydrocoupling of PhPH2 and the Forma-
tion of 4. Previous reports describing the formation of
Cp2MIVPn complexes (M ) Zr, Hf, V; n ) 1-3) involved the
intramolecular elimination of PH2R from a Cp2MIV(PRH)2
complex, with the concurrent formation of a metal phosphin-
idene as the key intermediate.25 A Ti(IV) phosphinidene species
was also suggested in the formation of a fulvalene-bridged Ti-
(III) phosphide dimer.25b Unlike their zirconocene analogs, such
Ti(IV) phosphinidene species have never been detected. If 2
is a phosphide dimer, one may ask why there is no further
(30) (a) Green, J. C.; Green, M. L. H.; Morris, G. E. J. Chem. Soc.,
Chem. Commun. 1974, 212. (b) Cannillo, E.; Coda, A.; Prout, K.; Daran,
J. C. Acta Crystallogr. 1977, B33, 2608.
(31) Jordan, R. F.; Lapointe, R. E.; Baenziger, N.; Hinch, G. D.
Organometallics 1990, 9, 1539.

NoVel Titanocene Phosphido Compounds
J. Am. Chem. Soc., Vol. 119, No. 23, 1997 5311
Scheme 2 Proposed Mechanism for the Formation of 4 and the Catalytic Dehydrocoupling of PhPH2
dehydrogenation of the CyPH bridging ligand to give either the
phosphinidine dimer or its decomposition products. This, or
some similar process, clearly occurs in the reaction of PhPH2
with 3. Analogs of 4, with less sterically demanding Cp ligands,
should presumably be able to capture a Cp2TiII moiety to give
a bridged phosphinidene dimer, but this apparently does not
occur.
formation of the highly insoluble (RP)5 are known to occur
spontaneously under these conditions.
A group of signals located in the region of +17 to +22 ppm
which are further complicated in the proton-coupled spectrum
are suggestive of some higher oligophosphines with P-H end
groups, but an attempt to separate the products by GC was not
successful. This is not surprising given the tendency of
oligophenylphosphines to disproportionate and undergo oxida-
tion. However, the high yield and low solubility of 4, together
with the 31P NMR data, suggest that the reaction of PhPH2 with
3 initially proceeds via a catalytic dehydrocoupling pathway to
give 4 and higher phenylphosphines. The catalytic reaction
eventually ceases due to loss of catalyst through precipitation
of the highly insoluble complex 4.
Scheme 2 accounts for the experimental observations by using
a conventional intermolecular σ-bond metathesis loop for the
dehydrocoupling process and an intramolecular loop for produc-
tion of cyclophosphanes and 4. A σ-bond metathesis reaction
could occur with either a Ti(IV) or a Ti(III) hydride, while a
classic oxidative addition/reductive elimination process could
involve Ti(II) and Ti(IV) species. Such a mechanism can also
explain the production of a nearly equal distribution of meso-
and rac-1,2-diphenyldiphosphine. In the σ-bond metathesis
process, a free PhPH2 can approach a Ti-P bond from either
side. If the phosphine approaches the Ti-P bond from one
side it will give rac diphosphine, while a meso diphosphine is
produced if the PhPH2 approaches the Ti-P bond from the other
side. Spontaneous racemization of either diphosphine is very
unlikely under the present reaction conditions because of the
high inversion energy (∆G* ∼ 22 kcal/mol).33b
An equilibrium between the dimer and monomer of 3 has
been observed in its EPR spectra both at room temperature and
in frozen toluene.15 Slow decomposition of 3 to liberate
hydrogen gas when it is dissolved in organic solvents has also
been observed at room temperature. These observations suggest
the possibility of disproportionation of 3 to (EBTHI)TiII and
(EBTHI)TiIVH2, and the latter may further reductively eliminate
dihydrogen to generate more of the former. In the presence of
PhPH2, any, or all, of the three possible oxidation states of
titanium in this system may catalyze the dehydrocoupling
reaction to give the observed cyclic and linear oligophosphines.
The reaction of 3 with an excess of PhPH2 to form 4 is
accompanied by a steady evolution of H2. The 31P NMR
spectrum of the reaction mother liquor is shown in Figure 4. In
addition to the peaks due to the starting material, PhPH2 (-123
ppm), some new peaks appear in the spectrum. The 31P{1H}
spectrum shows two signals which exhibit symmetrical AA′XX′
coupling patterns in the corresponding proton-coupled spectrum.
These two signals of nearly equal intensity are the meso (-66
ppm) and rac (-70 ppm) diastereomers of 1,2-diphenyldiphos-
phine (PhHPPHPh).32 A sharp peak at -3.5 ppm in the 31P-
{1H} spectrum shows no change in the proton-coupled spectrum,
suggesting a cyclic phosphine with an even number of P atoms
and the phenyl groups in the all-trans configuration. A likely
candidate is all trans (PhP)6, which has previously been
synthesized by noncatalytic methods, which also give the all-
trans isomer.33 This result appears to conflict with the recent
report of catalytic dehydrocoupling of primary phosphines by
an anionic hydridozirconocene complex, where the reported
products were cyclopentaphosphines.34 However, the latter
reactions were carried out at much higher temperature (120 °C)
in dioxane. The disproportionation of oligophosphines and the
It is not unexpected that (EBTHI)TiII would have a different
reactivity from Cp2TiII, in view of the radically altered steric
and electronic properties of EBTHI compared to the Cp ligand.
It is believed that the reason for the phosphine dehydrocoupling
activity of the EBTHI complex and the absence of observable
activity for the simple titanocenes probably lies in its lesser
tendency to form an inactive bridged dimer. The enhancement
of catalytic activity for analogous silane dehydrocoupling
reactions by increasing the steric bulk of the Cp ligand has also
been attributed to suppression of inactive dimer formation.4b
Attempts at dehydrogenative cross-coupling of Ph2PH, or
PhPH2, with PhSiH3 mediated by Cp2TiMe2 were unsuccessful.
After 24 h the 1H NMR spectra only showed resonances
assignable to unreacted PhSiH3 and Ph2PH, or PhPH2 with a
(32) (a) Dupont, T. J.; Mills, J. L. Inorg. Chem. 1973, 12, 2487. (b)
Albrand, J. P.; Robert, J. B. J. Chem. Soc., Chem. Commun. 1976, 876.
(33) (a) Baudler, M.; Carlsohn, B.; Koch, D.; Medda, P. K. Chem. Ber.
1978, 111, 1210. (b) Albrand, J. P.; Gagnaire, D. J. Am. Chem. Soc. 1972,
94, 8630.
(34) Fermin, M. C.; Stephan, D. W. J. Am. Chem. Soc. 1995, 117, 12645.

5312 J. Am. Chem. Soc., Vol. 119, No. 23, 1997
Table 1. Crystallographic Information
Xin et al.
1
2
4
chemical formula
formula wt
cryst system
space group
Z
C32H43Ti2P
554.43
orthorhombic
Pnma
4
1176
C32H44Ti2P2
586.41
monoclinic
P21/c
2
620
C32H34TiP2
528.47
orthorhombic
Pbcn
4
1112
F(000)
a (Å)
b (Å)
c (Å)
10.130(2)
16.290(4)
16.902(6)
10.329(5)
16.871(9)
8.382(4)
102.44(2)
1426.4(12)
9.796(4)
15.829(6)
16.757(6)
b deg
V (Å3)
2788.9(14)
1.320
2598(2)
1.351
dcalc
wavelength (Å)
1.54178
55.34
0.71073
6.84
0.71073
1.90
m (cm-1
)
trans. range
0.37-1.00
0.50, 0.30, 0.25
120
0.139
0.233
0.8329-1.00
0.40, 0.21, 0.02
45
0.90-1.00
cryst size (mm)
max 2θ (deg)
R
0.50, 0.28, 0.15
50
0.1546
0.1313
0.076
Rw
0.140a
S
1.050
8408
2335 (0.139)
2110
3
156
21
1.078; -0.397
1.009
0.901
5214
2614 (0.119)
2287
3
160
0
no. of reflcns measd
no. of unique reflcns (Rint)
no. used in LS
observation criteria (s)
no. of variables
no. of restraints
last D F map (e/Å3)
1790 (0.056)
1762
2
220
255
0.24; -0.23
0.282; -0.300
a Value based on F2.
small amount of PhMeSiH2. P-Si cross-coupling products were
not observed. It is probable that in these reactions, no catalytic
chemistry is observed due to the formation of stable, unreactive
titanocene(III) phosphide products.
Experimental Section
All reactions were performed under argon with use of conventional
Schlenk apparatus and methods. Manipulation of air-sensitive crystals
was performed in a Braun Lab-Master 130 glovebox with O2 and H2O
maintained at less than 0.5 ppm. Solvents were dried and purged of
36
oxygen by standard methods. Cp2TiMe2 and compound 315 were
Conclusion
prepared by methods described in the literature. All phosphines were
purchased from the Aldrich Chemical Co. and used as received.
Organosilanes were prepared by reduction of the corresponding
chlorosilanes, purchased from Hu¨ls America, with LiAlH4.37
Combustion analyses were performed by Oneida Research Services,
Whitesboro, NY. NMR spectra were recorded on a Varian Unity 500
FT spectrometer, operating at 200.33 MHz for 31P. EPR experiments
were recorded on a Bruker ER 220D spectrometer. Mass spectra were
recorded on a KRATOS MS25RFA spectrometer equipped with a
KRATOS DS90 data system. Photolyses were performed with a
HANAU low-pressure mercury-vapor lamp cooled inside a Pyrex
cooling jacket.
Reaction of Cp2TiMe2 and Cy2PH in the Presence of PhSiH3.
Cp2TiMe2 (50 mg; 0.25 mmol) was dissolved in 4 mL of hexane and
0.12 mL (1 mmol) of PhSiH3 and 0.1 mL (0.5 mmol) of PCy2H was
added. The light-orange solution was photolyzed for a period of 5
min until it suddenly darkened in color to become deep violet. This
color change was accompanied by vigorous gas evolution that lasted
for about 15 min. No crystals separated, even after a long time at 20
or -30 °C. The EPR spectrum at room temperature exhibited a strong
signal consisting of a doublet of pseudo-triplets (see text), strongly
indicating the formation of Cp2TiIII(SiH2Ph)(PCy2H). The same reaction
performed with PhSiD3 instead of PhSiH3 gave an EPR signal consisting
In this work we show that the reactions of dimethyltitanocene
with secondary and primary organophosphines resemble in many
respects those of the isoelectronic silanes. These reactivity
studies led to the isolation and structural characterization of
several novel Ti-P complexes. Two major differences between
silanes and phosphines are the inability of the phosphines to
effect the rapid reduction of titanocene(IV) and the failure of
the phosphines to undergo facile catalytic dehydrocoupling under
most circumstances. Both of these effects are likely due to the
ability of the phosphines to coordinate to key intermediates,
thus blocking further reaction. The absence of nonbonding
electron pairs from silanes (or other group 4 hydrides) precludes
these inhibitory effects. However, prereduction of the Cp2TiMe2
by a silane permits reactions of the phosphines, leading to
products (compounds 1 and 2) that are isostructural and
isoelectronic with those previously isolated from Cp2TiMe2/
silane reactions.
The diphosphene complex 4, although without precedent in
titanocene chemistry, is analogous to similar compounds that
have been characterized for zirconocene and other early transi-
tion metal metallocene derivatives. The silicon analogues are
unknown, for early transition metals, although disilenes are
known for later transition metals.35 The discovery of 4, as an
apparent participant in a dehydrocoupling cycle, keeps open the
possibility that, under the right conditions, a silicon analogue
might be isolated. We are continuing our explorations in this
direction.
of
a doublet of doublets indicating the formation of Cp2-
TiIII(PCy2H)(PhSiD2) (see text).
Reaction of Cp2TiMe2 and Cy2PH in the Presence of Ph2SiH2.
The reaction was conducted under exactly the same conditions as above,
but with Ph2SiH2 instead of PhSiH3 (same molar ratios of reactants).
Shortly after gas evolution had ceased, a dark-violet crystalline
compound began to settle. After 2 h, the solution was decanted and
the crystals were washed twice with hexane and dried to yield 30 mg
(35) (a) Zarate, E. A.; Tessier-Youngs, C. A.; Youngs, W. J. J. Am. Chem.
Soc. 1988, 110, 4068. (b) Pham, E. K.; West, R. Organometallics 1990,
9, 1517. (c) Heyn, R. H.; Tilley, T. D. J. Am. Chem. Soc. 1992, 114, 1917.
(36) Samuel, E.; Rausch, M. D. J. Am. Chem. Soc. 1973, 95, 6264.
(37) Benkeser, R. A.; Landesman, H.; Foster, D. J. J. Am. Chem. Soc.
1952, 74, 648.

NoVel Titanocene Phosphido Compounds
J. Am. Chem. Soc., Vol. 119, No. 23, 1997 5313
(25%) of 1. Anal. Found: C, 66.56; H, 7.60; P, 5.70; Ti, 16.45.
Calcd: C, 69.33; H, 7.76; P, 5.59; Ti, 17.29.
coupling products were not observed. This reaction was repeated with
other phosphines and silanes with the same result.
Reaction of Cp2TiMe2 and CyPH2 in the Presence of PhSiH3.
Cp2TiMe2 (50 mg; 0.25 mmol) was dissolved in 4 mL of hexane and
0.12 mL (1 mmol) of PhSiH3 and 0.1 mL (0.8 mmol) of CyPH2 was
added. The light-orange solution was photolyzed. The color gradually
changed to light brown and became darker, without appreciable gas
evolution. After a period of 30 min, dark crystals began to deposit on
the wall of the Schlenk tube facing the light source. The solution was
decanted and the crystals collected. Submitting the supernatant solution
to a second photolysis resulted in the deposition of a new crop of
crystals. Three such operations yielded 30 mg of 2. Yield: 40%.
Microanalysis did not give satisfactory results. Attempts at purification
of the compound by recrystallization were unsuccessful due to its rapid
decomposition in solution. However, some of the directly obtained
crystals appeared to be suitable for X-ray diffraction studies, and a
structure determination was undertaken (vide infra).
Crystal Structures. The crystallographic data for compounds 1,
2, and 4 are listed in Table 1. For 1 and 4, data sets were collected on
a Rigaku AFC6S diffractometer. The structures were solved by direct
methods and refined by full-matrix least-squares with use of SHELX93.38
Data reduction was performed by using the NRCVAX system of
crystallographic software for structure 1,39 SHELX for 2, and the
TEXSAN crystallographic software package of Molecular Structure
Corporation for 4.40 For 1, the molecule sits on a mirror plane passing
through both Ti atoms and the bridging P atom plane. The Cp rings
on both titaniums were found to be disordered over two orientations at
0.60 and 0.40 occupancies. These were modeled by using rigid C5H5
groups with independent isotropic thermal parameters for carbons. The
remaining non-H atoms were refined anisotropically. The bridging
hydride H was located on a difference map and held fixed. Other
hydrogen atoms were introduced at their ideal positions.
Synthesis of 4. (a) In Situ Preparation. A solution of MeLi (0.40
mL of 1.4 M; 0.56 mmol) was added to a solution of rac-[EBTHI]-
TiCl2 (96 mg; 0.25 mmol) dissolved in 2 mL of dry THF in a 15-mL
Schlenk tube under argon. The mixture was stirred for 12 h, and the
solvent was removed in Vacuo. The residue was extracted with dry
hexanes until the extract was practically colorless (5 × 1 mL) and was
cannula-filtered under argon to another 15-mL Schlenk tube. The
combined extract was evaporated to dryness under reduced pressure,
and 5 mL of dry toluene was added to dissolve the yellow residue.
Phenylmethylsilane (0.070 mL; 0.50 mmol) was added to the toluene
solution, and after the solution changed color to dark green or blue,
phenylphosphine (0.27 mL, 2.5 mmol) was syringed in and the mixture
was heated to 40-50 °C for 7 h until the gas evolution stopped. The
mixture was then left at room temperature for another 40 h. The
colorless supernatant liquid was decanted, and the green crystalline
material was washed with cold toluene (2 × 3 mL) and dried under
vacuum (127.6 mg; yield 86%).
For 4, the molecule sits on a 2-fold axis. Non-hydrogen atoms were
refined anisotropically, except the phenyl ring on phosphorus, which
was refined as a rigid C6H5 group with independent isotropic thermal
parameters for carbons. Hydrogen atoms were introduced at their ideal
positions. Refinement was based on |F|.
Compound 2 diffracted very poorly. Data were collected on a
Siemens P4 diffractometer equipped with a rotating anode generator.
The compound is a dimer about an inversion center. All atoms except
Ti are conformationally disordered, occupancies for each orientation
are 50%. This amount of disorder explains the weak diffraction. The
Cp rings were defined as rigid groups. Isotropic thermal parameters
were restrained to be equal within standard uncertainties. The two
P-C6H11 fragments were restrained to have similar geometries.
Acknowledgment. We wish to thank Dr. James Britten for
collecting the X-ray data for 2. The technical assistance of
Mme. J. He´nique is gratefully acknowledged. We thank
NSERC of Canada and Fonds FCAR du Que´bec for financial
support for this work.
(b) By Reaction of Phenylphosphine with 3. Complex 4 can also
be synthesized by dissolving 3 in benzene at 40-50 °C followed by
addition of 10 equiv of phenylphosphine. This method gave 4 in
isolated yields ranging from 90 to 95%. Anal. Calcd for C32H34P2Ti:
C, 72.73; H, 6.49; P, 11.72. Found: C, 72.87; H, 6.06; P, 11.90. MW
calcd for C32H34P2Ti: 528.47. Measured: MS (CI) 529 (M+ + H);
MS (EI) 528 (M+°). IR (KBr pellet): 3073 m, 3044 m, 2921 vs 2853
s, 1635 br w, 1575 s, 1560 w, 1472 s, 1438 vs, 1439 s, 1419 m, 1375
w, 1261 s, 1240 m, 1112 br vs, 1026 s, 902 w, 804 vs, 738 vs, 697 vs,
Supporting Information Available: Tables of crystal-
lographic data for compounds 1, 2, and 4 (34 pages). See any
current masthead page for ordering and Internet access
instructions.
JA960657D
567 w, 481 s cm-1
.
(38) (a) Sheldrick, G. M. SHELXS86 Structure solution program,
University of Go¨ttingen, 1985. (b) Sheldrick, G. M. SHELXL93 Program
for the refinement of crystal structures, University of Go¨ttingen, 1993.
(39) Gabe, E. J.; LePage, Y.; Charland, J. P.; Lee, F. L.; White, P. S. J.
Appl. Crystallogr. 1989, 22, 384.
(40) TEXSAN-TEXRAY Structure Analysis Package, Molecular Struc-
ture Corporation.
Attempted dehydrogenative Cross-Coupling Reaction of Ph2PH
and PhSiH3 in the Presence of Cp2TiMe2. PhSiH3 (0.25 mL, 2.0
mmol) and Ph2PH (0.28 mL, 1.6 mmol) were added to Cp2TiMe2 (42
mg, 0.2 mmol). The mixture was allowed to stir under ambient
fluorescent room lighting. After 24 h, the solution turned brown-purple.
The 1H NMR spectrum showed resonances assignable to unreacted
PhSiH3 and Ph2PH with a small amount of PhMeSiH2. P-Si cross-
(41) Johnson, C. K., ORTEPII Report ORNL-5138; Oak Ridge National
Laboratory, 1976.