Carbon-carbon bond formation via oxidative-addition processes of titanium(II) reagents with π-bonded organic substrates. Reactivity modifications by Lewis acids and Lewis bases Part 22. Organic chemistry of subvalent transition metal complexes

Article abstract of DOI:10.1016/S0022-328X(00)00921-9

A series of titanium(II) derivatives, TiE₂, was prepared by alkylative reduction of TiE₄ by two equivalents of n-butyllithium in THF at -78 to 25°C (E = Cl, F, OBuⁿ, OPrⁱ and 0.5 NPh-CH₂-CH₂-NPh). The LiE by-product could usually be removed by THF evaporation and dissolution of the TiE₂ into toluene. All such TiE₂ derivatives were shown to effect the epimetallation and oligomerization of olefins, acetylenes and carbonyl derivatives in varying degrees. Particularly pertinent were the isolation and chemical reactions of titanium(II) isoproxide, the postulated intermediate in the Kulinkovich synthesis of cyclopropanols from ethyl Grignard reagents and organic esters, as well as an intermediate in many allied reactions developed by the Sato group. The findings of the present study corroborate completely the foregoing hypothesis that titanium(II) isopropoxide is the key intermediate in such novel reactions in organic synthesis. Furthermore, Ti(OPr^l)₂ can be prepared readily in a relatively pure state and has been found to react with 1-alkenes, alkynes and ketones by epimetallation at 25°C to form three-membered titanacycles, which can be utilized in organic synthesis. Finally, the ease with which such TiE₂ derivatives epimetallate unsaturated organic substrates has been shown to be decreased by the steric demands of E and by the coordination of Lewis bases or donor solvent to the titanium(II) center. Lewis acids, on the other hand, greatly increase the rate of epimetallation by TiE₂. A dramatic illustration of this effect is in the action of TiCl₂·Me₂AlCl on unsaturated hydrocarbons, wherein the polymerization of ethylene and of 1-alkene and the cyclotrimerization of alkynes are found to occur at room temperature.

Full text of DOI:10.1016/S0022-328X(00)00921-9

Journal of Organometallic Chemistry 624 (2001) 229–238
www.elsevier.nl/locate/jorganchem
Carbon–carbon bond formation via oxidative-addition processes
of titanium(II) reagents with p-bonded organic substrates.
Reactivity modifications by Lewis acids and Lewis basesꢀ
Part 22. Organic chemistry of subvalent transition metal complexes
John J. Eisch *, John N. Gitua, Peter O. Otieno, Xian Shi
Department of Chemistry, The State Uni6ersity of New York at Binghamton, Binghamton, NY 13902-6016, USA
Received 16 October 2000; accepted 27 November 2000
Dedicated to Professor Jean F. Normant on the occasion of his 65th birthday
Abstract
A series of titanium(II) derivatives, TiE2, was prepared by alkylative reduction of TiE4 by two equivalents of n-butyllithium in
THF at −78 to 25°C (E=Cl, F, OBun, OPri and 0.5 NPh–CH2–CH2–NPh). The LiE by-product could usually be removed by
THF evaporation and dissolution of the TiE2 into toluene. All such TiE2 derivatives were shown to effect the epimetallation and
oligomerization of olefins, acetylenes and carbonyl derivatives in varying degrees. Particularly pertinent were the isolation and
chemical reactions of titanium(II) isoproxide, the postulated intermediate in the Kulinkovich synthesis of cyclopropanols from
ethyl Grignard reagents and organic esters, as well as an intermediate in many allied reactions developed by the Sato group. The
findings of the present study corroborate completely the foregoing hypothesis that titanium(II) isopropoxide is the key
intermediate in such novel reactions in organic synthesis. Furthermore, Ti(OPri)2 can be prepared readily in a relatively pure state
and has been found to react with 1-alkenes, alkynes and ketones by epimetallation at 25°C to form three-membered titanacycles,
which can be utilized in organic synthesis. Finally, the ease with which such TiE2 derivatives epimetallate unsaturated organic
substrates has been shown to be decreased by the steric demands of E and by the coordination of Lewis bases or donor solvent
to the titanium(II) center. Lewis acids, on the other hand, greatly increase the rate of epimetallation by TiE2. A dramatic
illustration of this effect is in the action of TiCl2·Me2AlCl on unsaturated hydrocarbons, wherein the polymerization of ethylene
and of 1-alkene and the cyclotrimerization of alkynes are found to occur at room temperature. © 2001 Elsevier Science B.V. All
rights reserved.
Keywords: Titanium(II) derivatives; Alkylative reduction; Oxidative-addition processes; Lewis acids and bases
unfilled metal d-orbital of proper symmetry, supple-
mented by dp–pp back-bonding from a filled d-orbital
1. Introduction
to a substrate empty p*-orbital, again with net positive
overlap. Depending upon the bonding contributions of
the s- and p-components, the resulting structure can
vary from that of a weak p-complex with a slightly
perturbed p-bond (1a) to a metallacyclopropane (1b)
wherein the p-bond, based upon the C–C separation,
has largely been weakened [2,3], and the Mt valence has
increased formally by 2. Judging whether 1a or 1b is the
more important resonance contribution must be de-
cided upon on the basis of spectroscopic and single-
crystal X-ray data for the individual complex. Such
The interaction of subvalent transition metal centers
with p-bonded organic substrates has become a hall-
mark of both structure and reactivity among organ-
otransition metal compounds [1]. The orbital
description of the metal-p-bonded substrate interaction
in the Dewar–Chatt–Duncanson model involves s-
bonding by a filled p-orbital of the substrate with an
ꢀ For the previous Part 21, see J. Organomet. Chem. 617-618
(2001) 148.
* Corresponding author.
0022-328X/01/$ - see front matter © 2001 Elsevier Science B.V. All rights reserved.
PII: S0022-328X(00)00921-9

230
J.J. Eisch et al. / Journal of Organometallic Chemistry 624 (2001) 229–238
structural assessments have been carried out both for
the early transition metal complex, [(CH3)5Cp]2Ti·
H2CꢀCH2 [4] and the late transition metal complex,
BpyNi·Ph–CꢁC–Ph [5].
The purpose of the present study, therefore, has been
to examine the facility with which divalent titanium
salts form such titanacyclopropane intermediates with
p-bonds, such as CꢀC, CꢁC and CꢀO, as a function of
the groups primarily bonded on titanium (Cl, OPri, F
and NR2) as well as a function of any added Lewis acid
or any Lewis basic solvent.
In the last two decades Group 4 metal complexes,
especially those formally with titanocene(II) and zir-
conocene(II) character, have been shown to form di-
verse p-complexes with benzynes, cycloalkynes and
alkynes but with relatively few alkenes. Carbon-carbon
couplings, cyclizations and oligomerizations of great
importance in organic synthesis have issued rewardingly
from such investigations [6,7].
2. Results
However, the most noteworthy discovery of the last
ten years has been the observation that even ordinary
subvalent derivatives of Group 4 metals, not involving
any cyclopentadienyl ligands, can readily form a wide
variety of complexes with simple alkenes or alkynes [8].
This seminal observation was made by Kulinkovich
and coworkers in 1989, who reported that ethylmagne-
sium bromide reacted with carbocyclic acid esters in the
presence of catalytic amounts of titanium(IV) isopro-
poxide to provide high yields of cyclopropanols (Eq.
(1)). It was proposed that the intermediate Ti(OPri)2
generated then complexed with the ethylene from the
Grignard reagent to form reactive intermediate 2 [9].
2.1. Preparation of titanium(II) deri6ati6es by
alkylati6e reduction
The most general, convenient and reliable method for
preparing titanium(II) compounds, TiE2 (8), is the
treatment of a solution or suspension of TiE4 (6) in
anhydrous tetrahydrofuran (THF) under argon with
two equivalents of n-butyllithium at −78°C [12]. Slow
warming of the reaction mixture to 25°C causes the
intermediate 7 to decompose with the formation of
invariably black 8 (E=Cl, OPri, F and one-half of
NPhCH2CH2NPh). The by-product LiE can usually be
separated from the reagent 8 by evaporating all of the
THF under reduced pressure with warming. Such treat-
ment also serves to disrupt any complex between TiE2
and 1-butene (Eq. (2)). The residue is extracted with
toluene, which dissolves the TiE2 and leaves the LiE
which can be filtered off (except for LiOPri, which is
soluble). The toluene extract contains TiE2 either as the
bis-tetrahydrofuran complex (E=Cl, F) or as the sol-
(1)
Numerous subsequent reactions involving analogs of
the original eponymous Kulinkovich reagent 2 have
been studied with substituted alkenes and alkynes both
by the Kulinkovich group in Russia [10] and after 1995
by the Sato group in Japan [11], leading to the great
enrichment of modern organic synthesis.
vate-free
TiE2
(E=OPri
or
one-half
of
NPhCH2CH2NPh). For routine use as a reagent, the
TiE2–2LiE mixture can be used directly.
In our own research with subvalent Group 4 metal
reagents and catalysts, the observed cyclotrimerizations
of alkynes and the stereoselective racemic bimolecular
reduction of methyl phenyl ketone led us to propose in
1995 that metallacyclopropenes 3 and 4 were appealing
intermediates in such possible two-electron transfers
(TET) or oxidative addition processes [12]. Our subse-
quent studies on the synthesis of metallocene dihalides
by the reductive dimerization of fulvenes by Group 4
dihalides [13] and the reduction and dimerization of
either alkynes or ketones by LiCrH2 [14,15] have
strengthened our confidence in the presence of such
intermediates. Particular support for the presence of 5
in the reduction of Ph2CꢀO by LiCrH2 is that workup
with D2O produced almost a 90% yield of Ph2CD–OD
[15].
(2)
These LiE-free complexes of TiE2 can be obtained
and analyzed for the amount of Ti, E and complexed
THF according to the methods described previously
[12].
The preparation of TiE2 (8) can alternatively be
conducted in THF at −78°C with ethyl-, n-propyl- or
allylmagnesium chloride in diethyl ether and the reac-
tion mixture brought to 25°C. In this procedure there is
evidence that the TiE2 and MgE2 form stable complexes
with each other [16].
Finally, the preparation of TiE2 (8) can be under-
taken in toluene, in order to avoid any complexation
with a donor solvent such as THF. In this situation a

J.J. Eisch et al. / Journal of Organometallic Chemistry 624 (2001) 229–238
231
cooled solution or suspension of TiE4 in toluene at
−78°C is treated with two equivalents of butyllithium
in hexane, one equivalent of butyl(ethyl)magnesium in
heptane (Akzo-Nobel) or two equivalents of trimethyla-
luminum in hexane. In each case the reaction mixture is
brought to 25°C and allowed to stir for 20 h to
complete the decomposition to black TiE2. Especially
with Bu2TiF2, stirring and warming are essential to
produce TiF2. For the TiE2 reagents prepared from
TiE4 with BuLi or with BuMgEt, the by-product LiE or
MgE2 remains admixed and probably weakly com-
plexed with the TiE2.
OPri) by alkylative reduction of Ti(OPri)4 with n-butyl-
lithium (Eq. (4)). The Ti(OPri)2 was soluble in toluene
1
and gave H- and 13C-NMR signals for the isopropyl
1
group, which were broadened to obscure H hyperfine
splitting, indicative of paramagnetic titanium(II). (Col-
laborative ESR studies of such titanium(II) derivatives
are currently underway with Professor David C.
Doetschman of this Department.) Significant by their
absence were any absorptions due to complexed THF
or 1-butene.
The reaction of TiCl4 with Me3Al is noteworthy (Eq.
(3)).
(4)
Removal of the toluene left 8 as a black solid, which
can be readily redissolved in THF. Bubbling ethylene
into the latter solution, the subsequent addition of
benzonitrile (one-half equivalent based on 8) and heat-
ing the reaction mixture gave upon subsequent hydroly-
sis 88% of ethyl phenyl ketone (12). This outcome is
pertinent corroboration of the view that Ti(OPri)2 has
added to the p-bond of ethylene to form 10, a process
that can be termed epimetallation by analogy with the
epoxidation of alkenes in the Prilezhaev reaction [18].
Insertion of benzonitrile into a C–Ti bond of 10 would
readily lead to 11, whose hydrolysis generates 12.
In a modification of this procedure, proof for the
existence of the remaining Ti–C bond in 11 was ob-
tained by passing CO2 into the reaction mixture result-
ing from 10 and benzonitrile. Evidence that CO2 had
inserted into 11 to form 11a came from the observation
that subsequent hydrolysis of 11a produced 3-benzoyl-
propanoic acid (11b) in a non-optimized 30% overall
yield.
To our knowledge, this sequence of reactions is the
first unequivocal proof of the intermediacy of Ti(OPri)2
in the reactions discovered and developed by the
Kulinkovich and the Sato groups. Furthermore, the
Ti(OPri)2 is then shown to add to a separately added
olefin to form 10, both of whose two distinct Ti–C
bonds can then be identified by individual chemical
trapping (Scheme 1).
In a similar manner, it can be shown that the
Ti(OPri)2 as initially generated in THF (Eq. (4)) imme-
diately coordinates with the 1-butene arising from the
alkylative reduction (Eq. (2)) to form 13: addition of
benzonitrile to the THF-solution of Ti(OPri)2 yields 14
and subsequent hydrolysis provided a 50% yield of
n-butyl phenyl ketone (15) (Scheme 2). Any concern
that some or all of 15 could have resulted from the
addition of any residual n-butyllithium to benzonitrile
was completely allayed by finding that one equivalent
of BuLi in THF, when allowed to warm from −78 to
+25°C overnight (typical reaction period employed to
prepare TiE2 in Eq. (2)), was completely destroyed by
attack on THF [19].
(3)
By elemental analysis the black TiCl2 precipitate was
found to be complexed with only one equivalent of
Me2AlCl; the other equivalent of Me2AlCl produced
was found free in solution [16]. The 27Al-NMR spec-
trum (CDCl3) of 9 showed a peak at 101.5 ppm,
consistent with a tetracoordinate aluminate anion (cf.
the AlCl−4 signal at 103.3 ppm in [Cp2Ti–
C(SiMe3)ꢀCPhMe]+[AlCl4]− [17]); the 1H- and 13C-
NMR spectra had signals at 1.37 and 14.2 ppm,
respectively. Such evidence is consistent with a poly-
meric structure of TICl+ centers bridged by [Cl–
AlMe2–Cl]−.
2.2. Addition of titanium(II) deri6ati6es, TiE2 (8), to
y-bonds in6ol6ing carbon
2.2.1. To CꢀC bonds
The interaction of Grignard reagents with catalytic
amounts of titanium(IV) isopropoxide (6, E=OPri) has
led to remarkable and useful organic chemical transfor-
mations, which have been rationalized by Kulinkovich
[10] and by Sato [11] as proceeding via titanacyclo-
propanes of type 2 (Eq. (1)). Our efforts reported here
have been directed to verifying this hypothesis by the
preparation of titanium(II) isopropoxide (TiE2 (8), E=
Scheme 1.

232
J.J. Eisch et al. / Journal of Organometallic Chemistry 624 (2001) 229–238
greater stability of 10 over 13 most likely arises from
the lesser frontal steric strain in 10 between the isopro-
poxy groups and the hydrogen substituents on the
C=C bond (cf. 16) (Scheme 3).
Attempts to detect adducts of 1-butene with either
TiCl2 or TiF2 in the THF solutions in which they are
prepared (Eq. (2)) have thus far been unsuccessful.
Possibly relevant is that these titanium(II) halides form
THF complexes of the type, TiX2·2THF. Such com-
plexation may be firmer than that achievable with
1-butene. By contrast, Ti(OPri)2 can be isolated free of
any complexed THF and thus is sterically capable of
complexing with 1-butene readily. Moreover, in prelim-
inary experiments the titanium(II) diamide (17) does
show the capability of epimetallating CꢀC and CꢀO
bonds (cf. infra) [20]. Like Ti(OPri)2, titanium(II)
derivative 17 fails to form an isolable THF complex
and thus is capable of adding to CꢀC and CꢀO
linkages.
Scheme 2.
2.2.2. To CꢁC bonds
Diphenylacetylene (18) reacts readily either with iso-
lated Ti(OPri)2 (one equivalent), which is free of 1-
butene or with Ti(OPri)2 as generated in THF where
LiOPri, 1-butene and adduct 13 are still present (Eq.
(2)). Hydrolytic work-up yields cis-stilbene (20) as the
major product (90% of product), together with minor
amounts of (E,E)-1,2,3,4-tetraphenyl-1,3-butadiene 22
and hexaphenylbenzene (5%). The origin of these hy-
drocarbons can be attributed to the epimetallation ad-
duct 19 and to the insertion of 18 into 19 to form 21
(Scheme 4).
Scheme 3.
Finally, when the foregoing reaction mixture contain-
ing 13 (Scheme 2) is directly treated with ethylene gas,
without prior removal of the THF and the 1-butene
(13“8), and then with benzonitrile, hydrolytic work-
up yields only ketone 12 and none of ketone 15. This
result demonstrates that titanacyclopropane 13 is in
equilibrium with Ti(OPri)2 (8) and 1-butene and that 8
bonds more firmly with ethylene in forming 10. The
This reaction scheme parallels the previously pre-
sented mechanism by which TiCl2, especially in toluene
Scheme 4.

J.J. Eisch et al. / Journal of Organometallic Chemistry 624 (2001) 229–238
233
Table 1
heated at reflux. Hydrolytic work-up gave the reduction
products given in Table 1 [20].
Products from the reaction of titanium(II) derivatives, TiE2, with
benzophenone (24) (2:1) and subsequent hydrolysis (%)
These results can be understood in terms of the
reaction scheme proposed previously for the McMurry-
like transformation of 24 into tetraphenylethylene (28) by
TiCl2 in THF [12,14]. In Scheme 5 the epimetallation of
24 by TiE2 would first generate 25, which could undergo
insertion of 24 to produce pinacolate salt 26 or could
simply dimerize directly to pinacolate salt 27. Further
reaction of 27 with TiE2 would finally provide 28 with
the elimination of E2Ti–O–TiE2 (29). Hydrolysis of 25
and 27 would give rise to benzhydrol (30) and benzpina-
col (31), respectively. In light of previous steric effects
noted for Ti(OPri)2 (cf. supra), it is readily apparent that
the epimetallated intermediate 25, where E=OBun, OPr,i
or one-half of NPh–CH2–CH2–NPh, would be greatly
sterically hindered from either autodimerization to 27 or
insertion of 24 to yield 26. Accordingly, it would accu-
mulate in the reaction mixture and lead to 30. Similar
steric interference would retard the conversion of 27 to
28 by TiE2 (Scheme 5). On the other hand, the failure to
detect benzhydrol in reactions of benzophenone with
TiCl2 or TiF2 supports the conclusion that 25 (E=Cl,
F) is too reactive to persist in solution.
TiE2
Ph2CH–OH
(30)
(Ph2C–OH)2
(31)
Ph2CꢀCPh2
(28)
TiCl2
TiF2
100
100
30
B1
25
Ti(OBun)2
Ti(OPri)2
Ti(II) amide
17
50
95
60
20
B1
15
Scheme 5.
2.3. Effect of Lewis bases on the reacti6ity of TiE2
deri6ati6es
The retarding effect of Lewis bases on the epimetallat-
ing reaction of TiE2 derivatives is especially evident with
TiCl2. When prepared in THF, it can be separated from
the LiCl as a complex, TiCl2·2THF (32), soluble in
toluene. If the reductive alkylation shown in Eq. (2) is
carried out directly in toluene, the TiCl2 forms a partial
solution with the LiCl by-product (33). Toward dipheny-
lacetylene (18) in refluxing solvent, 33 proved to be much
more reactive in the cyclotrimerization than 32 (Scheme
6). Although the reflux temperatures of the solvents
(THF at 66°C and toluene at 111°C) might be considered
as responsible for the differences in the conversion of 18
into 23, temperature and time were not decisive factors.
Further prolongation of time of reaction for 18 and 32
changed the conversion to 22 and 23 only slightly.
Similar reactivity differences between 33 and 32 were
shown in the catalytic cyclotrimerization of 5-decyne (34)
(Eq. (5)), and the cyclodimerization of 1,1-diphenyl-
ethylene (36) (Eq. (6)).
Scheme 6.
and in the absence of THF, effects the cyclotrimerization
of 18 and other alkynes [12,14]. What is significantly
different in the reaction of Ti(OPri)2 with 18 is the minor
amount of hexaphenylbenzene (23), the usual cy-
clotrimeric product of such reactions, and the preponder-
ance of 20 and hence 19 as a product. In the reactions
of TiCl2 not even traces of an olefin like 20 can be
detected. We propose that the predominance of 19 and
the absence of 23 can be ascribed to the steric effect of
isopropoxy groups in 19 and in 21, which would hinder
the insertion of 18 into the C–Ti bond of 19 and impede
the Diels–Alder addition of 18 to 21 that leads to the
production of 23.
2.2.3. To CꢀO bonds
The titanium(II) reagents, TiCl2, TiF2, Ti(OBun)2,
Ti(OPri)2 and Ti(II) amide 17, were prepared in THF by
treating TiE4 with two equivalents of n-butyllithium, first
at −78°C and then at gentle reflux until black reaction
mixtures resulted. The TiE2 reagent was then treated with
0.5 equivalents of benzophenone (24) and the mixture
(5)
(6)

234
J.J. Eisch et al. / Journal of Organometallic Chemistry 624 (2001) 229–238
It should be noted that the foregoing comparisons of
aluminum alkyl co-catalyst such as MAO, and to cy-
clotrimerize diphenylacetylene (18) to hexaphenylben-
zene (23) at 25°C suggests 9 furnishes [TiCl]+ (39) as
the potent, electrophilic epimetallating agent to the
carbon-carbon p-bond, thereby generating a three-
membered titanium(IV) cation capable of rapidly in-
serting further alkene or alkyne units and leading to
polymer or cyclotrimers (Scheme 7) [16].
the reactivity of the reagents, 32 (TiCl2·2 THF dissolved
in THF), and 33 (TiCl2+2LiCl suspended in toluene)
suffer from the heterogeneity of 33 in toluene, a factor
that by itself should render reagent 33 less active than
reagent 32, which is homogeneous in THF solution.
Despite this hampering influence, therefore, it is all the
more striking that in every case reagent 33 has proved
to be more reactive than reagent 32. This finding makes
indisputable the retarding influence of THF on the
reactivity of TiCl2.
3. Discussion
2.4. Effect of Lewis acids on the reacti6ity of TiE2
deri6ati6es
The epimetallation of CꢀC, CꢁC and CꢀO linkages
by subvalent transition metal compounds having at
least one unshared electron pair, like :TiE2, :CrCl and
:ZrCl2, bears a striking parallel to carbene additions to
the same substrates (Scheme 8). In a recent analysis
these similarities have induced us to term such
reagents as TiE2, (Me5C5)2Ti and Bpy·Ni transition
metal carbenoids [4]. The analogy with carbenes has
been extended to consider whether such transition
metal carbenoids would further exist either as bent
singlet complexes (40) or linear triplet complexes (41).
The reactivity of the 1:1 complex of TiCl2 and
Me2AlCl (9), prepared as shown in Eq. (3) illustrates
clearly the accelerating action of Me2AlCl on the
epimetallation of CꢀC and CꢁC bonds by TiCl2. Al-
though the structure of the amorphous, apparently
polymeric 9 has not yet been determined, its black color
supports the presence of a Ti(II) salt and the 27Al-
NMR signal at 101.5 ppm indicates the presence of
tetracoordinate aluminum. We therefore suggest that 9
has a chain polymeric structure of Ti(II)Cl cationic
centers bridged by Cl2AlMe2 anions (38).
It has been suggested that the singlet form 40 would
be more likely to undergo oxidative additions with
p-bonds while 41 would be more responsive to s-bond
insertions.
Such a structure would be similar to the known
helical polymeric chain structure of [Me3Sn]+[AlCl4]−
[21]. The observed ability of 9 to polymerize ethylene
and alpha-olefins at 25°C in toluene, unaided by an
Finally, with reference to the electronic character of
ligand E, electron-withdrawing groups (EWG) should
make such carbenoids electrophilic, accepting p-elec-
tron density into unfilled d-orbitals (42). Conversely,
electron-donating groups (43) should foster the nucle-
ophilicity and hence the release of the unshared electron
pair into the p*-molecular orbital of the p-bond [2,3].
In designating ligands to the Mt carbenoid center as
either electron-withdrawing or electron-donating, we
consider that such electron-attraction or release would
be operative along the s Mt–E bond. With this
premise, then, we judge that the Taft |* parameter
would better reflect the electronic effect, rather than the
Hammett | parameter, which measures electronic at-
traction or release operative via p-bonding. Only with
NR2 ligands, for which |* values are not obtainable,
do we propose that its strongly negative |-value (ca.
−0.8) causes it to donate electron density via p-bond-
ing [22].
Scheme 7.
Scheme 8.

J.J. Eisch et al. / Journal of Organometallic Chemistry 624 (2001) 229–238
235
(7)
From a preparative standpoint in organic synthesis,
it is noteworthy that the Ti(OPri)2 necessary for the
many synthesis applications discovered by the groups
of Kulinkovich and of Sato can be made most readily
and conveniently from Ti(OPri)2 according to Eq. (2)
and can be freed of any 1-butene or THF contami-
nants. Its adducts with alkenes, alkynes, carbonyl
derivatives and allylic halides are then directly accessi-
ble (Scheme 9).
Scheme 9.
From the present study it is clear that an electrophilic
carbene 43 will be deactivated in its epimetallating
reactivity by Lewis bases that occupy necessary coordi-
nation sites, as with TiCl2·2THF (32), which reacts
much more slowly with alkynes or with alkenes than
TiCl2 in toluene (Scheme 6; Eqs. (5) and (6)). On the
other hand, an electrophilic carbene will be activated by
Lewis acids towards epimetallation, which acids both
open up a coordination site and confer a partial posi-
tive charge on the transition metal. This would best
explain the enhanced epimetallating and polymerizing
action of TiCl2·Me2AlCl (9) on alkenes and alkynes
(Scheme 7), the higher reactivity readily being ascrib-
able to the generation of the more electrophilic car-
benoid cation, [TiCl]+.
4. Conclusions
A series of titanium(II) derivatives of the type, TiCl2,
TiF2, Ti(OPri)2 and Ti(NPh–CH2–CH2–NPh) have
been generated by the alkylative reduction of the corre-
sponding titanium(IV) derivative by two equivalents of
n-butyllithium in THF at −78°C. The titanium(II)
derivatives can usually be obtained free of the LiE
by-product and any complexed 1-butene. The TiCl2 and
TIF2 remained complexed with THF (2 mol), while
Ti(OPri)2 and Ti(NPh–CH2–CH2–NPh) can be ob-
tained free of THF.
Moreover, the failure of Ti(OPri)2 to complex with
THF, apparently on steric grounds, represents another
way by which the electrophilic character and an avail-
able coordination site can be preserved for the
epimetallation of CꢀC, CꢁC and CꢀO linkages even in
THF solution.
Nucleophilic carbenes of type 43 should be especially
prone to reaction with carbonyl linkages, where the
unshared electron pair in a d-orbital can be shared with
a p*-orbital of the CꢀO group of (44). Coordination of
Lewis bases should enhance the nucleophilicity of 43,
so long as the unshared electron pair on Mt is still
sterically accessible for coordination to the substrate.
All Ti(II) derivatives can effect the epimetallation
and oligomerization of CꢀC, CꢁC and CꢀO derivatives
with varying facility. Lewis bases retard such reactions
and Lewis acids, such as Me2AlCl, can greatly acceler-
ate such epimetallations.
Steric factors in such derivatives as Ti(OPri)2 and
Ti(NPh–CH2–CH2–NPh) can greatly hinder the coor-
dination with Lewis bases as well as with p-bonds in
unsaturated substrates.
Finally, the hypothesis has been corroborated that
Ti(OPri)2 forms key titanacyclopropane complexes with
olefins in the Ti(OPri)4-catalyzed reactions of Grignard
reagents with esters in the Kulinkovich synthesis of
cyclopropanols. As a bonus in convenience, it has been
found that the Ti(OPri)2 necessary for the Kulinkovich
reaction and related reactions uncovered by Sato can be
readily generated from Ti(OPri)4 and n-butyllithium in
THF.
A final structural feature that can enhance the nucle-
ophilicity of 43 is the presence of negative charge. This
factor seems to explain why the:CrH2 anion adds read-
ily and extensively to benzophenone (24) at −78°C in
THF [15]. The presence of the epimetallated adduct,
chromaoxacyclopropane 45, is corroborated by the iso-
lation of dideuteriated benzhydrol 46 in 90% yield (Eq.
(7)). Had a hydrochromation of 24 by LiCrH2 ensued
instead, then the benzhydrol isolated would have been
Ph2CH–OD.
5. Experimental
5.1. Instrumentation, analysis and starting reagents
All reactions were carried out under a positive pres-
sure of anhydrous, oxygen-free argon. All solvents em-
ployed with organometallic compounds were dried and

236
J.J. Eisch et al. / Journal of Organometallic Chemistry 624 (2001) 229–238
distilled from a sodium metal-benzophenone ketyl mix-
ture prior to use [23]. The IR spectra were recorded
with a Perkin–Elmer instrument, model 457 and sam-
ples were measured either as mineral oil mulls or as
KBr films. The NMR spectra (1N and 13C) were
recorded with a Bruker spectrometer, model EM-360,
and tetramethylsilane (Me4Si) was used as the internal
standard. The chemical shifts reported are expressed in
the l-scale and in parts per million (ppm) from the
reference Me4Si signal. The GC/MS measurements and
analyses were performed with a Hewlett–Packard GC
5890/Hewlett–Packard 5970 mass-selective-detector in-
strument. The gas chromatographic analyses were car-
ried out with a Hewlett–Packard instrument, model
5880, provided with a 6 ft. OV-101 packed column or
with a Hewlett–Packard instrument, model 4890, hav-
ing a 30 m SE-30 capillary column, respectively. Melt-
ing points were determined on a Thomas–Hoover
Unimelt capillary melting point apparatus and are
uncorrected.
plete the conversion of the brown mixture (Bun2TiF2) to
the black TiF2.
Such TiE2 reagents in THF, still containing the LiE
by-product, can be used directly for epimetallation and
oligomerization reactions of unsaturated organic sub-
strates. In fact, the results obtained with benzophenone
and given in Table 1 were obtained with such reagents.
When the LiE by-product is to be removed, one can
do so by the procedure used to isolate pure TiCl2·2THF
and previously reported [12]. This procedure applied to
isolation of titanium(II) propoxide is a given as a model
method for separating LiE from the other foregoing
TiE2 derivatives. Thus, the previously obtained black
solution of titanium(II) isopropoxide in THF was sub-
jected to reduced presence with warming at 40°C, until
all volatiles (THF, 1-butene) were removed. The residue
black solid was successively slurried with 3×20 ml
portions of anhydrous, deoxygenated toluene and
filtered. Evaporation of the toluene filtrate gave the
Ti(OPri)2 as a black solid admixed with LiOPri. Disso-
lution of a sample of Ti(OPri)2 in THF-d8 gave an
1H-NMR spectrum showing an intense, but broad (0.4
ppm) signal centered at 1.2 ppm (–CH(CH3)2), and a
weak, broadened (0.3 ppm) signal at 3.7 ppm (–
CH(CH3)2), expected for the isoproxyl group. Note-
worthy is the absence of any signals ascribable to THF
or to 1-butene.
In a similar manner, the lithium salt-free TiF2 was
shown to have coordinated THF (probably two equiva-
lents) and the Ti(OBun)2 and Ti(NPh–CH2–CH2–NPh)
shown not to have coordinated THF.
Because of the solubility of titanium(II) isopropoxide
in toluene, the reagent can be prepared directly in this
solvent. Thus, to a solution of Ti(OPri)4 (5.95 ml, 20
mmol) dissolved in 30 ml of anhydrous, deoxygenated
toluene was slowly added at −78°C n-butyllithium (25
ml, 1.6 M). The initially brown-colored solution turned
black over 18 h of rising to 20°C. The solution of
Ti(OPri)2 may be used directly.
The TiCl4, TiF4, Ti(OPri)4, Ti(OBun)4, ethyl (n-bu-
tyl)magnesium, trimethylaluminum and n-butyllithium
were obtained from commercial sources. The titaniu-
m(IV) bis(1,2-ethane-dianilide), Ti(NPhCH2CH2CH2
NPh)2 was prepared in toluene by heating two equiva-
lents of 1,2-dianilinoethane with four equivalents of
n-butyllithium, and then adding one equivalent of TiCl4
to the resulting sodium salt, Li2(NPh–CH2–CH2NPh).
The slurry was stirred at reflux and the LiCl filtered
from the hot toluene solution. The cooled toluene
solution deposited colorless Ti(NPh–CH2–CH2–NPh)2
in 75% yield [20].
5.2. Preparation of titanium(II) deri6ati6es, TiE2, by
alkylati6e reduction of titanium(IV) deri6ati6es
5.2.1. With n-butyllithium
The preparation of the titanium(II) derivatives, TiE2,
where E=Cl, F, OPri, OBun and 0.5 NPh–CH2–CH2–
NPh, from the TiE4 derivative was conducted
analogous to the published preparation of TiCl2 in
THF solution containing the two equivalents of LiCl
by-product [12]. Given below is the model procedure
for the preparation of titanium(II) isopropoxide, which
was followed for all other titanium(II) derivatives as
well.
5.2.2. With trimethylaluminum
The alkylative reduction of a titanium(IV) derivative
with either trimethylaluminum in toluene or ethyl(-
butyl)magnesium in heptane has thus far been applied
only to TiCl4. A detailed procedure employing Me3Al is
given below.
To a solution of titanium(II) isopropoxide (5.95 ml,
20 mmol) dissolved in anhydrous, deoxygenated THF
(30 ml) at −78°C was added slowly n-butyllithium (25
ml in hexane, 40 mmol) whereupon the solution turned
light brown and then brown. The reaction mixture was
stirred overnight (18 h) as the reaction temperature was
allowed to rise to room temperature. At this point, a
black solution of titanium(II) isopropoxide was ob-
tained. In the analogous preparation of TiF2 it was
necessary to warm the THF reaction mixture to com-
To 15.0 ml of 1.0 M TiCl4 (15 mmol) in toluene and
80 ml of anhydrous, deoxygenated toluene at −78°C
were slowly added 15 ml of 2.0 M Me3Al (30 mmol) in
toluene over 90 min. The resulting reaction mixture was
allowed to warm to 20°C over 24 h to produce a black
suspension. The black suspension was filtered off (3.1 g)
1
and the H-NMR spectrum of the light yellow filtrate
showed the presence of one equivalent (ca. 15 mmol) of
uncomplexed (CH3)2AlCl (27Al-NMR) signal at 175
ppm.

J.J. Eisch et al. / Journal of Organometallic Chemistry 624 (2001) 229–238
237
The black solid was analyzed for titanium and for
chlorine as previously published for TiCl2 [12]. (Found:
Cl, 50.49; Ti, 22.50. C2H6Cl3AlTi Calc.: Cl, 50.37; Ti,
22.68%). The black solid (9) dissolved in dry, deoxy-
genated CDCl3 gave the following NMR signals: 27Al,
that was extracted with 3×20 ml portions of aqueous
NaHCO3. The combined aqueous extracts were aci-
dified with aqueous 6 N HCl and the liberated car-
boxylic acid taken up in ether. The isolated acid,
3-benzoylpropanoic acid (11b), was obtained in a non-
1
1
+101.5 ppm, H, 1.37 ppm and 13C, 14.2 ppm.
optimized 30% yield. H-NMR (CDCl3) l 2.8 (2H, t),
3.3 (2H, t), 7.4–7.6 (8H, m), 7.9–8.0 (2H, m); 13C-
NMR (CDCl3) l 28.0, 33.2, 128, 128.5, 133.2, 136.4,
198.1, 198.2.
5.3. Reactions of titanium(II) deri6ati6es prepared in
THF with benzophenone
5.6. Reaction of titanium(II) isopropoxide in THF with
benzophenone
The following titanium(II) derivatives were prepared
on a 20-mmol scale in THF by alkylative reduction
with n-butyllithium (Section 5.2): TiCl2, TiF2,
Ti(OBun)2 Ti(OPri)2 and Ti(II) amide 17. To such solu-
tions were added 1.82g (10 mmol) of benzophenone.
The reaction mixture was heated under reflux for 24 h
and a hydrolysis with aqueous 3 N HCl ensued. Usual
separation and drying of the organic layer, solvent
removal and a combined NMR spectral and chromato-
graphic analysis yielded the results compiled in Table 1.
The reaction of 20 mmol of titanium(II) isopropoxide
in THF with 0.5 equivalents (1.82 g, 10 mmol) of
benzophenone gave, after 24 h at 25°C and hydrolytic
work-up, 1.65 g of product, which by NMR spectral
analysis was shown to consist of 95% cis-stilbene, 5% of
remaining benzophenone and B1% of benzpinacol and
tetraphenylethylene.
5.7. Reactions of the 1:1 complex, TiCl2·Me2AlCl (9)
5.4. Reaction of titanium(II) isopropoxide in THF with
benzonitrile
5.7.1. With ethylene
The reaction mixture of 9 with the Me2AlCl by-
product resulting from admixing 1.0 mmol of TiCl4
with 2.0 mmol of Me3Al in 100 ml of toluene was
filtered to separate 9. Black solid 9 was resuspended in
100 ml of toluene and treated with ethylene gas at 25°C
and at 40 psi for 1 h. Usual hydrolytic work-up yielded
7.10 g of polyethylene, representing an activity of 54.6
To the black solution of titanium(II) isopropoxide in
THF was added benzonitrile (1.02 ml, 10 mmol) at
25°C. The solution was stirred at 25°C for 12 h. This
period was followed by hydrolysis with aqueous 3 N
HCl, extraction of the organic products in Et2O (3×30
ml) and neutralization of the combined organic extracts
with NaHCO3. The extracts were dried with anhydrous
MgSO4 and the removal of the solvent under reduced
pressure gave a thick brown liquid. Yield of the crude
g Ti−1 h−1 bar−1
.
1
product=0.82 g. From the H-NMR spectrum, 50% of
n-butyl phenyl ketone was found to result.
5.7.2. With 1-hexene
Black 9 (1 mmol) suspended in 150 ml of cyclohexane
was treated with 10 mmol of 1-hexene for 48 h at room
temperature. Usual hydrolytic work-up and removal of
volatiles under reduced pressure left 31 g (78%) of clear,
5.5. Reaction of titanium(II) isopropoxide in THF with
ethylene and benzonitrile
1
pale yellow viscous liquid, whose H-NMR spectrum in
toluene-d8 had two broad signals at 0.9–1.00 ppm and
at 1.30–1.40 ppm in an area ratio of 3:7, consistent
with the presence of a polyalkane.
Into 20 mmol of titanium(II) isopropoxide in THF at
25°C was passed an excess of dry ethylene gas. Then
1.05g (10 mmol) of benzonitrile was introduced and the
resulting reaction mixture was stirred at room tempera-
ture for 12 h. Hydrolytic work-up and NMR spectral
analysis showed that 88% of ethyl phenyl ketone had
resulted and no trace of n-butyl phenyl ketone could be
detected.
In a modification of the foregoing procedure, after
the reaction between titanium(II) isopropoxide (20
mmol) and benzonitrile (10 mmol) had been permitted
to react at 25°C for 10 h, the reaction mixture was
cooled to −78°C and treated with anhydrous, gaseous
CO2 for 2 h. Usual hydrolytic work-up with aqueous 3
N HCl and diethyl ether produced an organic extract
5.8. Reaction of titanium(II) isoproxide with
diphenylacetylene
To 20 mmol of titanium(II) isopropoxide in THF
was added one equivalent of diphenylacetylene (3.56 g,
20 mmol) and the resulting solution stirred at 25°C for
24 h. Usual hydrolytic work-up and NMR spectral
analysis revealed that 50% of the alkyne was recovered,
45% of the product was cis-stilbene and the remaining
5% was a mixture of (E,E)-1,2,3,4-tetraphenyl-1,3-buta-
diene and hexaphenylbenzene.

238
J.J. Eisch et al. / Journal of Organometallic Chemistry 624 (2001) 229–238
(Eds.), Comprehensive Organometallic Chemistry, vol. 3, Perga-
mon, Oxford, 1982, pp. 1–88.
5.9. Reaction of titanium(II) chloride with 5-decyne
[2] M.J.S. Dewar, Bull. Soc. Chim. Fr. 18 (1951) C79.
[3] J. Chatt, L.A. Duncanson, J. Chem. Soc. (1953) 2939.
[4] J.J. Eisch, J. Organomet. Chem. 617–618 (2000) 148.
[5] J.J. Eisch, X. Ma, K.I. Han, J.N. Gitua, C. Kru¨ger, Eur. J.
Inorg. Chem. (2001) 77.
[6] S.L. Buchwald, R.B. Nielsen, Chem. Rev. 88 (1988) 1047.
[7] E. Negishi, Acc. Chem. Res. 20 (1987) 65.
[8] B. Breit, J. Prakt. Chem. 342 (2000) 211.
Two parallel reactions were carried out: (1) 20 mmol
of TiCl2·2 THF (32) in 80 ml of anhydrous THF was
heated at reflux with 40 mmol of 5-decyne. By intermit-
tent sampling and GC analysis after hydrolysis it was
found that a 95% yield of hexa-n-butylbenzene was
attained after 60 h; (2) 20 mmol of TiCl+2LiCl,
prepared as a suspension in 80 ml of toluene, was
allowed to react at reflux with 40 mmol of 5-decyne.
Similar analytical sampling showed that a 95% yield of
trimer was attained after 12 h.
[9] O.G. Kulinkovich, S.V. Sviridov, D.A. Vasilevski, T.S. Prityck-
aja, Zh. Org. Khim. 25 (1989) 2244.
[10] O.G. Kulinkovich, A.d. Meijere, Chem. Rev. 100 (2000) 2789.
[11] F. Sato, H. Urabe, S. Okamoto, Chem. Rev. 100 (2000) 2835.
[12] J.J. Eisch, X. Shi, J. Lasota, Z. Naturforsch. 50b (1995) 342.
[13] J.J. Eisch, X. Shi, F.A. Owuor, Organometallics 17 (1998) 5219.
[14] J.J. Eisch, X. Shi, J.R. Alila, S. Thiele, Chem. Ber./Recueil 130
(1997) 1175.
Acknowledgements
[15] J.J. Eisch, J.R. Alila, Organometallics 19 (2000) 1211.
[16] J.J. Eisch, S.I. Pombrik, X. Shi, S.C. Wu, Macromol. Sym. 89
(1995) 221.
[17] J.J. Eisch, A.M. Piotrowski, S.K. Brownstein, E.J. Gabe, F.L.
Lee, J. Am. Chem. Soc. 107 (1985) 7219.
[18] D. Swern, in: R. Adams (Ed.), Organic Reactions, vol. 7, Wiley,
New York, 1953, pp. 378–433.
[19] P.O. Fregene, unpublished studies of this Laboratory, 2000.
[20] S. K.H. Thiele, unpublished studies of this Laboratory, 1995.
[21] J.J. Eisch, K. Mackenzie, H. Windisch, C. Kru¨ger, Eur. J. Inorg.
Chem. (1999) 153.
The foregoing research was conducted with the finan-
cial support of the US National Science Foundation, a
Senior Scientist Award to J.J.E. from the Alexander
von Humboldt Stiftung of Bonn, Germany and the
Solvay Corporation of Brussels, Belgium. Furthermore,
we are indebted to Drs S.I. Pombrik and S.K.H. Thiele
for valuable orienting experiments.
References
[22] J. Shorter, Q. Rev. (Lond.) 24 (1970) 443.
[23] J.J. Eisch, Organometallic Syntheses, vol. 2, Academic Press,
New York, 1981, pp. 3–37.
[1] D.M.P. Mingos, in: G. Wilkinson, F.G.A. Stone, E.W. Abel
.