Direct epimetallation of π-bonded organic substrates with titanium(II) isopropoxide: Intermediacy of biradical, oligomeric titanium(II) reagents

Article abstract of DOI:10.1002/ejic.200501074

Extensive EPR experiments show that a titanium-containing molecular triplet state is formed in solution by the reaction of two equivalents of butyllithium with one equivalent of titanium(IV) isopropoxide. At higher concentrations this product, titanium(II) isopropoxide, admixed with two equivalents of lithium isopropoxide, is accompanied by the formation of a variety of nontitanium-containing side products. The powder EPR spectrum of the molecular triplet state in frozen solution is consistent with an asymmetric molecular chain of three Ti centers on which the unpaired electron centers are three metal atoms apart. Dilution experiments show that at lower concentrations, where the nontitanium-containing side products have dissipated, the intensity of the molecular triplet spectrum varies approximately linearly with concentration. Thus there is no evidence that the observed triplet molecule is one component in a series of concentration-dependent oligomerization steps. The bulky isopropoxy substituents and the coordination of the isopropoxide anions from the LiOiPr present appear to prevent closure of the Ti₃ centers into an equilateral triangular diamagnetic structure. This steric hindrance, operative at the terminal diisopropoxytitanium centers and preventing closure to a ring, seems not to be observed with TiCl₂·2THF, which is diamagnetic and may thus be expected to exist as an equilateral triangular cluster of three units of TiCl₂·2THF, a structural model currently under further investigation. The smaller steric demand of the chloro and THF units would seem to permit octahedral coordination about each Ti center in such an equilateral trigonal array of Ti₃ atoms. Chemical reactions carried out individually with diphenylacetylene, cis-stilbene or cis-stilbene oxide and titanium(II) diisopropoxide provide stoichiometric and stereochemical evidence that the attacking titanium(II) reagent is in fact the trimeric biradical. The role of the lithium isopropoxide byproduct in fostering the course of the previously reported SET reactions of titanium(II) isopropoxide and in determining the detailed structure of the open-chain Ti trimer biradical has been explicated. Wiley-VCH Verlag-GmbH & Co. KGaA, 2006.

Full text of DOI:10.1002/ejic.200501074

FULL PAPER
DOI: 10.1002/ejic.200501074
Direct Epimetallation of π-Bonded Organic Substrates with Titanium(II
)
Isopropoxide: Intermediacy of Biradical, Oligomeric Titanium(II) Reagents[‡]
John J. Eisch,*[a] John N. Gitua,[a] and David C. Doetschman[a][b]
Keywords: Titanium() / EPR determination of biradicals / Titanium() isopropoxide trimer / Epititanation / Nonstereo-
chemical TiII additions
Extensive EPR experiments show that a titanium-containing
molecular triplet state is formed in solution by the reaction
of two equivalents of butyllithium with one equivalent of tita-
nium(IV) isopropoxide. At higher concentrations this product,
titanium(II) isopropoxide, admixed with two equivalents of
lithium isopropoxide, is accompanied by the formation of a
variety of nontitanium-containing side products. The powder
EPR spectrum of the molecular triplet state in frozen solution
is consistent with an asymmetric molecular chain of three Ti
centers on which the unpaired electron centers are three
metal atoms apart. Dilution experiments show that at lower
concentrations, where the nontitanium-containing side prod-
ucts have dissipated, the intensity of the molecular triplet
spectrum varies approximately linearly with concentration.
Thus there is no evidence that the observed triplet molecule
is one component in a series of concentration-dependent
oligomerization steps. The bulky isopropoxy substituents and
the coordination of the isopropoxide anions from the LiOiPr
present appear to prevent closure of the Ti3 centers into an
equilateral triangular diamagnetic structure. This steric hin-
drance, operative at the terminal diisopropoxytitanium cen-
ters and preventing closure to a ring, seems not to be ob-
served with TiCl2·2THF, which is diamagnetic and may thus
be expected to exist as an equilateral triangular cluster of
three units of TiCl2·2THF, a structural model currently under
further investigation. The smaller steric demand of the chloro
and THF units would seem to permit octahedral coordination
about each Ti center in such an equilateral trigonal array of
Ti3 atoms. Chemical reactions carried out individually with
diphenylacetylene, cis-stilbene or cis-stilbene oxide and tita-
nium(II) diisopropoxide provide stoichiometric and stereo-
chemical evidence that the attacking titanium(II) reagent is
in fact the trimeric biradical. The role of the lithium isopro-
poxide byproduct in fostering the course of the previously
reported SET reactions of titanium(II) isopropoxide and in de-
termining the detailed structure of the open-chain Ti trimer
biradical has been explicated.
(© Wiley-VCH Verlag GmbH & Co. KGaA, 69451 Weinheim,
Germany, 2006)
sary to perform individual evaluations of the structural pa-
rameters of adduct 3, if isolated, or of its observed chemical
reactions.[2–4]
Introduction
The complexation of subvalent transition-metal reagents
(LmMtn, 1) with various C=C, CϵC, C=O, or C=N linkages
(as in 2) has led to adducts 3 that may be considered as π
complexes 3a, where the metal has undergone little change
in oxidation number (Mtn). On the other hand, the metal
center can have released significant electron density to the
π*-orbital of the organic substrate with the formation of
three-membered metallocycles 3b, and the metal may
thereby have a higher oxidation number (Mtn+2) [Equa-
tion (1)]. To determine which of the resonance structures 3a
and 3b is the better single structure of adduct 3, it is neces-
(1)
The process involved in the formation of adduct 3 has
been termed epimetallation, because the metal reagent 1 is
added with the rupture of a π-bond.[2] If reagent 1 is gener-
ated in a separate step and then added to 2, this process
may be termed direct epimetallation (Scheme 1, path a).
More recently, however, we have found that a precursor di-
alkyl derivative of higher oxidation number, R2Mtn+2Lm (4),
can itself transfer the LmMtn unit to the unsaturated sub-
strate 2 with concomitant loss of the R groups (Scheme 1,
path b). This latter process has accordingly been termed
transfer-epimetallation.[5a, 6]
[‡] Organic Chemistry of Subvalent Transition-Metal Complexes,
38. Part 37: Ref.[1]
[a] State University of New York at Binghamton, Department of
Chemistry,
P. O. Box 6000, Binghamton, NY 13902-6000, U.S.A.
Fax: +1-607-777-4865
E-mail: jjeisch@binghamton.edu
[b] Professor of Chemistry & Director, Regional Center for Pulsed
EPR & Photochemical Studies, State University of New York
at Binghamton, Department of Chemistry,
P. O. Box 6000, Binghamton, NY 13902-6000, U.S.A.
1968
© 2006 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Eur. J. Inorg. Chem. 2006, 1968–1975

Epimetallation of π -Bonded Organic Substrates with Titanium() Isopropoxide
FULL PAPER
A most likely reason for the greater epititanation activity
of Bu2TiL2 (10 or 11) over that of TiL2 (5 or 7) may well
be the more associated character of TiL2. In the solid state,
it should be noted that the TiCl2 crystal lattice consists of
layers of an infinite plane of Ti atoms flanked on both sides
by infinite planes of chlorine atoms. It is understandable
that considerable activation energy would be required to
generate monomeric TiL2 units, even from a far less associ-
ated form of these titanium() salts. As produced from
2 equiv. of butyllithium with 1 equiv. of TiCl4 or Ti(OiPr)4,
the resulting TiL2 remains admixed with 2 equiv. of either
LiCl or Li(OiPr)2. The separation of pure TiCl2 or Ti(OiPr)2,
free of Li salts, for analytical or molecular weight measure-
ments, is a difficult task. We have been able to isolate and
analyze successfully TiCl2·2THF, free of LiCl, but we have
been uniformly unsuccessful in separating Ti(OiPr)2 from
LiOiPr, since both components are freely soluble in pure
THF or in hexane.
Hindered from determining the degree of association of
pure Ti(OiPr)2 (7) in solution, we have availed ourselves of
another interesting physical property of 7. In contrast with
the diamagnetic character of TiCl2 (5) in THF over a range
of temperatures below 25 °C, we have found that Ti(OiPr)2
with its byproduct LiOiPr in heptane/toluene solution is
pronouncedly paramagnetic and displays an EPR triplet
signal characteristic of a biradical. The detailed study of
such an EPR spectrum now permits us to estimate the de-
gree of association of such Ti(OiPr)2 units.[7] Extrapolating
from data on alkali metal complexes of TiCl2, such as
Na2Ti3Cl8, we now suggest a possible similarity in the asso-
ciation of TiCl2 units in THF.
Scheme 1.
It is noteworthy to observe that the epititanation of ethyl-
ene by titanium() diisopropoxide (7) is widely thought to
lead to 1,1-diisopropoxy-1-titanacyclopropane, postulated
as the key intermediate in the Kulinkovich cyclopropanol
synthesis.[5]
A most unusual and unexpected observation made in
comparing such epimetallations with titanium reagents is
that the direct epititanations with preformed TiL2 (L = Cl,
OiPr)2 are much slower than transfer-epititanations with
Bu2TiL2.[2,5] For example, preformed TiCl2 (5) in THF can
achieve no epititanation of diphenylacetylene (6) whatso-
ever at 25 °C; similarly, Ti(OiPr)2 (7) in THF reacts with 6
over 24 h to provide only 31% of the epititanation product
9, presumed to stem from adduct 8, which was detected
through D2O cleavage as 9 [Equation (2)]. In considering
this low reactivity of TiCl2 or Ti(OiPr)2 in direct epititana-
tions, it is still noteworthy that some moderate activity is
shown by Ti(OiPr)2 at 25 °C, while TiCl2 is completely un-
reactive. As will be observed later in this study, this distinct
difference in the chemical reactivity of Ti(OiPr)2 vs. TiCl2
is also reflected in their respective EPR activity.
Results and Discussion
The EPR Investigation of the 1:2 Mixture of Titanium(II
Diisopropoxide (7) and Lithium Isopropoxide in 1:1
Solutions of Heptane/Toluene
)
(2)
The initial CW EPR investigation of diluted samples of
such mixtures, where the concentration of 7, formally con-
sidered as the monomer, was successively 20.0, 7.0, 2.0, and
0.7 m, revealed a broad, strong line at g = 1.944, a sharp
line at g = 1.964, and a complex set of lines around g =
1.98. Subsequent examination of this flash-frozen sample at
80 K gave a spectrum with a central set of features between
In sharp contrast, however, when either Bu2TiCl2 (10)
or Bu2Ti(OiPr)2 (11) was generated in THF at –78 °C and
0.9 equiv. of diphenylacetylene (6) was then added, subse-
quent warming to 25 °C and treatment with D2O gave a
91–99% yield of cis-[D1,2]stilbene (9). This outcome clearly
demonstrates that 10 or 11 achieves a rapid epititanation of
6 at or below 25 °C. The facility of such transfer-epititan-
ation has been rationalized by the formation of an octahe-
dral transition state 12 with 6, in which the new, incipient
C–Ti bonds of the epititanated product are formed as the
butyl–titanium bonds are thereby predisposed to homolytic
cleavage [Equation (3)].
3100 and 3800 Gauss (microwave frequency,
ν
=
9.36947 GHz) and a weak feature at 1690 Gauss (micro-
wave frequency, ν = 9.36583 GHz). The experiment was re-
peated with the spin-echo technique at 25 K, which con-
firmed the CW EPR results. In both cases, the central re-
gion appeared to comprise weak wing features at 3150 and
3750 Gauss, shoulders at 3125 and 3575 Gauss, and a peak
at 3460 Gauss, relative to ν = 9.4181 GHz. After refine-
ments in sample preparation had been made, the 3150 and
3575-Gauss features were eliminated. Please note the g-
value representation of the central region of the spectrum
as presented in Figure 1, as the derivative of EPR absorp-
(3) tion with respect to the field. An integral (absorption vs.
Eur. J. Inorg. Chem. 2006, 1968–1975
© 2006 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.eurjic.org
1969

J. J. Eisch, J. N. Gitua, D. C. Doetschman
FULL PAPER
field) of the EPR spectrum in this region is shown in Fig- propoxide concentrations of 20.0, 7.0, 2.0, and 0.7 m were
ure 2. The low-field part of the EPR spectrum is shown in done with successively greater refinements in the pro-
Figure 3.
cedures. The sharp line at g = 1.964 observed at 20.0 and
7.0 m was largely eliminated with this improved pro-
cedure. The complex set of weak lines around g = 1.98 ini-
tially observed at all concentrations except 0.7 m was elim-
inated by means of the improved technique at all concentra-
tions except 20.0 m. The intensity of the triplet signal var-
ies approximately in a linear dependence with the concen-
tration, once the complex at g = 1.98 disappears upon di-
lution.
Analysis of the Observed EPR Spectra of Titanium
Diisopropoxide (7)
The central part of the frozen solution spectrum, shown
in integral form in Figure 2, is characteristic of a powder
pattern of a molecular triplet state. The so-called half-field
feature shown in the frozen solution spectrum of Figure 3
is located at a field consistent with a weakly allowed ∆ms =
2 transition between the levels of a triplet state under the
influence of the dipole–dipole interaction of the unpaired
triplet-state electrons.[8] The parameters determined with
the following triplet-state spin Hamiltonian [Equation ()]
are given in Table 1.
Figure 1. CW EPR derivative spectrum of a 0.7 m frozen 1:1 hex-
ane/toluene solution of titanium isopropoxide at 50 K as a function
of g value in the g = 2 region.
2
2
2
H = H·g·S + DSz + E(Sx – Sy )
(4)
Table 1. Spin Hamiltonian parameters obtained from a least-
squares fit of the parameters in Equation (4) to the titanium diiso-
propoxide frozen solution in Figure 2. (The standard deviations σ
in the parameters are also presented. The absolute signs of D and
E are unknown but their relative signs are opposite.).
Figure 2. CW EPR absorption spectrum of a 0.7 m frozen 1:1
hexane/toluene solution of titanium isopropoxide at 50 K, experi-
ment and theory at microwave frequency 9.54971 GHz. (The nor-
mal derivative output of the EPR spectrometer has been numeri-
cally integrated.).
Value
σ
D (10–4 cm–1 hc)
81.7
ϯ23.1
1.9293
1.9449
1.9304
18.7
1.8
0.2
0.0002
0.0002
0.0002
0.5
E (10–4 cm–1 hc)
gxx
gyy
gzz
∆H (Gauss)
The g value (1.944) of the broad strong line in the solu-
tion spectrum (not shown) is the same as that of the triplet
molecule. Thus the solution spectrum is believed to be the
motionally averaged spectrum of the powder pattern of the
triplet molecule observed at lower temperatures (See Fig-
ure 2). The relatively smaller g-shift of the complexes giving
the variable set of lines at g = 1.98 suggests that these are
radicals free of a Ti center. In the most highly refined di-
lution spectra it appears that the only highly concentration-
dependent spectrum is that of the non-Ti-containing com-
plex. It would appear that the triplet spectrum associated
with Ti at g = 1.944 is present at all concentrations and its
Figure 3. CW EPR derivative spectrum of a 0.7 m frozen 1:1 hex-
ane/toluene solution of titanium isopropoxide at 50 K as a function
of g value in the g = 4 region.
Dilution experiments were done in order to try to under- spectrum does not change appreciably with concentration.
stand the relationships between the several features ob- Its intensity scales approximately with concentration, once
served in the CW EPR solution spectrum (not shown) and the complex at g = 1.98 disappears upon dilution.
to consider the possibility that various oligomers of titani-
Ignoring the E value for the moment, the D value alone
um() isopropoxide were being observed. Four sets of di- corresponds to the dipolar interaction of two point-electron
lution experiments at nominal monomeric titanium() iso- spins on titanium atoms at a distance of 0.668 nm.[9] Since
1970
www.eurjic.org
© 2006 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Eur. J. Inorg. Chem. 2006, 1968–1975

Epimetallation of π -Bonded Organic Substrates with Titanium() Isopropoxide
FULL PAPER
this Ti–Ti separation is more than twice the titanium atom a greater average triplet electron separation than the
separation in the [Ti3] cluster found in Na2Ti3Cl8, the un- Յ0.2996 nm value in the [Ti3]6+ cluster. The nonzero E
paired electrons are clearly not on the same Ti center or value rules out a symmetric trigonal structure, such as was
unlikely even on adjacent Ti centers. The nonzero E value found for the [Ti3]6+ cluster, for which E = 0 by symmetry.
rules out a symmetric trigonal (equilateral) array, such as Thus, with all factors considered, the structural model most
the [Ti3] cluster[10] for which E = 0 by symmetry.
in accord with the EPR data would appear to be 7c, where
Quantitative estimates of the triplet component(s) pres- θ approaches 180°, and the individual Ti–Ti bonds are
ent in samples with 0.7, 2.0, 7.0, and 20.0 m Ti content lengthened to greater than 0.2996 nm by the steric repulsion
gave an average of 0.28 0.14 triplet spins per Ti atom. Al- of the isopropoxy groups attached to adjacent titanium cen-
though within one standard deviation such a value is not ters (7e, L = OiPr) (See Figure 4). The same steric bulk of
consistent with triplet spins on the same Ti atom (1.0) or the isopropoxy groups would weaken any coordination of
on adjacent Ti atoms (0.50). However, this value is in agree- THF units to Ti centers, as well as hindering spin-pairing
ment with three Ti atoms per triplet molecule (0.33). Of and closure to the symmetrical [Ti3] cluster 7f [Equa-
course, this value is also consistent with four (0.25), five tion (5)].
(0.20), six (0.17), or seven (0.14) Ti atoms per triplet mole-
cule. This outcome supports the conclusion that a determi-
nation of the number of Ti atoms per triple molecule, mea-
sured with a precision typical of such experiments, cannot
by itself be expected to determine the specific number of Ti
atoms per Ti cluster.
(5)
Comparison of the EPR Data for Titanium(II
)
Diisopropoxide (7) with Various Structural Models
A continuum of titanium() isopropoxide oligomers
come under consideration (7a–7d, arrow = relative orienta-
tion of electron spin). From the D and E values, monomer
7a and diamagnetic trimer 7d can be excluded from further
consideration. From the D values alone the Ti–Ti distance
in 7b (0.2996 nm) is too short, and even a linear trimer of
7c (θ = 180°) falls short (0.599 nm) of the calculated experi-
mental value of 0.668 nm (see Figure 4). The linear tetra-
mer of 7 would lead to a D value of 0.8988; however, the
substantial value of E is not predicted by a range of models
reproducing the value of D, including one with parallel p-
type orbitals simulated by point spins on terminal Ti atoms,
a model with a twist angle between the terminal Ti p-type
orbitals, and a model with in-plane p-type point spins and
a Ti–Ti–Ti bond angle θ = 105°. This failure in prediction
suggests that there may also be one-center dipole–dipole
contributions from different orbitals for different spin con-
tributions or spin-orbit contributions to the zero-field pa-
rameters of Equation (4). The D value and the nonzero E
value suggest a structure differing fundamentally in two re-
spects from that of the [Ti3]6+ cluster found by Hinz et al.
in Na2Ti3Cl8 at low temperatures.[10] The D value suggests
Figure 4. Logarithmic plot of the D values as a function of point-
spin–point-spin distance, according to a simple dipole-dipole
model placing the spins on terminal Ti atoms. On this plot, the
experimentally measured value of D is indicated in relation to mul-
tiples of the Ti–Ti internuclear distance in the Ti3 cluster reported
by Hinz and co-workers.[10]
The tendency of divalent titanium centers to form tri-
[10]
meric Ti clusters, as in the solid state of Na2Ti3Cl8
and
as here in solutions of Ti(OiPr)2 (7), encourages us to sug-
gest that diamagnetic TiCl2·2THF (5·2THF) may exist in
solution as the cyclic symmetrical structure 7f, (L = Cl), in
which each Ti center, because of the smaller, electronegative
chlorine ligand, could further accommodate two THF li-
gands and thereby attain octahedral coordination. The pos-
sible existence of cyclic and open-chain trimeric clusters of
7e and 7f (L = Cl) will be probed by EPR measurements of
TiCl2·2THF conducted at higher temperatures.[11]
Structure of Titanium(II) Diisopropoxide (7) Implied by Its
Chemical Reactions
In both [Ti(OiPr)2]3 and [TiCl2·2THF]3, it is readily ap-
parent that the monomeric TiL2 unit, necessary for the epi-
Eur. J. Inorg. Chem. 2006, 1968–1975
© 2006 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.eurjic.org
1971

J. J. Eisch, J. N. Gitua, D. C. Doetschman
FULL PAPER
titanation of unsaturated organic ligands [Equation (1)],
would require considerable activation energy to be liberated
from its trimer, and thus such direct epititanations [Equa-
tion (1)] would understandably be slower than transfer-epit-
itanations [Equation (3)].
Moreover, certain reactions of Ti(OiPr)2 (7) with organic
substrates, as reported in this study, follow pathways consis-
tent with the participation mainly of the trimeric biradical
7c. For example, with 1 equiv. of diphenylacetylene (6) and
2 equiv. of 7, the addition reaction leading to cis-stilbene
(15) upon hydrolysis proceeded to the extent of 30% at
25 °C. When a similar reaction mixture was heated at re-
flux, the reaction went only to an extent of 74%, despite
the presence of an excess of 7. Now, if the 2 equiv. of 7 is
actually present as trimer 7c, then the 0.67 equiv. of 7c
could react with only 0.67 equiv. of 6, close to the observed
0.74 equiv. of 6 consumed. Thus, we conclude that the ad-
dition reaction to 6 should be described as involving mostly
attack by 7c [Equation (6)] to form 14. Hydrolysis would
then yield 15 or with D2O, 1,2-dideuterio-cis-stilbene (9),
as is observed.
Scheme 2.
Scheme 3.
(6)
that of the similar TiCl2/2LiCl reagent (5) on 6 [Equa-
tion (2) and related text]. The greater reactivity of 7 over 5
has been ascribed to the electron-withdrawal effect of the
isopropoxy groups already bonded in 7, as well as any iso-
propoxy anions (from the LiOiPr) that may also coordinate
with the Ti(OiPr)2 with the formation of complex titani-
um() anions, such as [Ti(OiPr)4]2–.[12]
Two other reactions giving evidence for the participation
of biradical 7c directly are those involving cis-stilbene (15)
itself and cis-stilbene oxide (20). Regardless of how long
and at what temperature the reaction proceeded and
whether 7 contained LiOiPr or not, the main reaction of a
1:2 equiv. ratio of 15 and 7, as revealed by hydrolysis, was
isomerization to trans-stilbene (17), rather than addition to
form 18 (detected as deuteriolysis product 19). Workup
with D2O produced 1,2-dideuteriobibenzyl (19), consistent
with its formation from organotitanium precursor 18.
Again, we suggest that the 0.67 equiv. of trimer 7c is opera-
tive, as shown in Scheme 2. The trans-stilbene could be
formed directly from 16 by reeliminating 7c or from 18 by
stepwise Ti–C bond homolysis. In the conversion of 16 into
17, the greater thermodynamic stability of 17 over 15 (about
91:9) should favor the formation of the trans isomer.
In a similar manner, the deoxygenation of cis-stilbene ox-
ide (20) by 7 to give the thermodynamic mixture of trans-
and cis-stilbenes (17 and 15) in a 91:9 ratio is consistent
with the stepwise attack of biradical 7c on 20 to form an
acyclic biradical intermediate 21. Biradical 21 could exist
sufficiently long to permit rotation about the C–C bond
before the second Ti–C bond undergoes induced homolysis
and thereby produces the thermodynamic mixture of 15
and 17 (Scheme 3).
A final previously studied reaction of 7 that clearly in-
volves the participation of radical intermediates in SET
(Single Electron Transfer) processes and implicates the by-
product LiOiPr as an essential coreactant, is the reaction of
7 with aromatic nitriles (22) or geminal dichlorides (23),
as shown in Scheme 4. The hydrolysis products, 24 and 25
respectively, had individually resulted from a reductive iso-
propylation. Since Ti(OiPr)2 containing no LiOiPr bypro-
duct (only LiCl) was unable to bring about such a reaction,
the necessary SET reagent was proposed to be 26, a com-
plex of LiOiPr with the supposed monomer Ti(OiPr)2.[12]
Scheme 4.
The first indication that the isopropoxy ligand exerts an
With the new structural insight into the nature of 7 of-
activating effect on titanium(), over that observed with the fered by the present study, we would now propose that tri-
chloro ligand, was the faster direct epimetallating action of meric 7c or 7d undergoes coordination with one or more
the Ti(OiPr)2/2LiOiPr system on diphenylacetylene (6) over equivalents of LiOiPr to form anionic coordination com-
1972
www.eurjic.org
© 2006 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Eur. J. Inorg. Chem. 2006, 1968–1975

Epimetallation of π -Bonded Organic Substrates with Titanium() Isopropoxide
FULL PAPER
plex(es), such as 27 [Equation (7)]. The coordination of the
Finally, chemical reactions carried out individually with
LiOiPr with adjacent Ti atoms in 7d or terminal Ti atoms diphenylacetylene, cis-stilbene, or cis-stilbene oxide and ti-
in 7c would tend to force the Ti trimer into the open-chain tanium() isopropoxide provide stoichiometric and stereo-
dianionic biradical 27. Therefore, we can deduce from this chemical evidence that the attacking titanium() reagent is
chemical evidence that a mixture of 7 and LiOiPr most in fact the trimeric biradical 7c, rather than some mono-
likely exists in solution as one or more coordination com- meric form of titanium() isopropoxide. The previously re-
plexes of the type depicted in 27.
ported SET reductive isopropylation of aromatic nitriles
and geminal dichlorides is reinterpreted in light of this new
evidence on the structure of titanium() isopropoxide, and
the essential role of lithium isopropoxide in shaping the de-
tailed structure of the trimeric titanium biradical is expli-
cated.
Experimental Section
Starting Materials, Reaction Conditions, and Instrumentation
The starting TiCl4 and Ti(OiPr)4 employed were of 98% or greater
purity as commercially available, and the n-butyllithium used was
as a 1.60- solution in hexane. All reactions, transfers, and separa-
tions were carried out under a positive pressure of anhydrous, oxy-
gen-free argon. All solvents, mainly tetrahydrofuran, heptane, and
toluene, employed with these air- and moisture-sensitive reagents
were dried and distilled from a mixture of sodium metal and benzo-
(7)
A corollary deduction of these considerations is that
highly pure titanium() diisopropoxide, absolutely free of
coordinating anions, such as from LiOiPr or LiCl, and per-
haps even free of coordinating solvents such as THF, might
well exist at lower temperatures as the diamagnetic trimer
7d. This possibility is currently under investigation.[11]
1
phenone ketyl prior to use.[13] The H- and 13C NMR spectra were
recorded with a Bruker spectrometer, model EM-360, with tet-
ramethylsilane (Me4Si) as the internal standard. The chemical
shifts reported are expressed on the δ-scale in parts per million
(ppm) from the Me4Si reference signal.
The electron paramagnetic spectra (EPR) measurements were per-
formed with a Bruker EPR spectrometer, model ESP580 X-band,
equipped with a split-ring ER4AA9 resonator and were detected
with 100-kHz field modulation and phase-sensitive detection. Tem-
peratures below room temperature were achieved with an Oxford
Instruments CF935 cryostat and an ITC502 temperature controller.
Two types of EPR experiments were performed on solutions of the
divalent titanium derivatives in THF, heptane, or 1:1 (v/v) heptane/
toluene mixtures: (1) CW EPR scans of various widths (200–
1000 Gauss) were performed at room temperature (20–25 °C),
mainly about fields corresponding to g = 2, and similar CW EPR
experiments were performed with solutions flash-frozen at 50 K or
80 K; and then (2) a Hahn-echo detected, pulsed EPR spectrum
was acquired at 20 K. Quantitative estimates of the concentrations
of paramagnetic components present in the samples were obtained
by comparison with the spectrum of a freshly prepared 1.032-m
standard solution of 1,1-diphenyl-2-picrylhydrazyl.
Conclusions
In summary, the EPR experiments show that a Ti-con-
taining molecular triplet state is formed in solution from
the reaction of 2 equiv. of butyllithium with 1 equiv. of tita-
nium() isopropoxide. At high concentrations this product,
titanium() isopropoxide, admixed with 2 equiv. of lithium
isopropoxide, is accompanied by the formation of a variety
of non-Ti-containing side products. The powder EPR spec-
trum of the molecular triplet state in frozen solution is con-
sistent with an asymmetric molecule or chain of three Ti
centers on which the unpaired electron centers are most
probably three centers apart. Dilution experiments show
that at lower concentrations, where the non-Ti-containing
side products have dissipated, the molecular triplet spec-
trum intensity varies approximately linearly with concentra-
tion. Thus there is no evidence that the observed triplet
molecule is one component in a series of concentration-de-
pendent oligomerization steps. Coordination of this open-
chain, three-titanium cluster with the lithium isopropoxide
byproduct would, for steric reasons, tend to keep such
anions in their open-chain biradical form. This steric hin-
drance for the terminal diisopropoxytitanium centers clos-
ing to a ring seems not to be observed with TiCl2·2THF,
which is diamagnetic and may thus exist as an equilateral
triangular cluster of three units of TiCl2·2THF. The smaller
steric demand of the chloro and THF units may permit
octahedral coordination about each Ti center in such a tri-
gonal array of Ti3 atoms.
Preparation of Titanium(II) Isopropoxide (7, containing LiOiPr): In
a typical procedure on a 10-mmol scale, a solution of titanium()
isopropoxide (1.28 g, 10 mmol) in THF (40 mL) and cooled to
–78 °C was slowly treated with n-butyllithium (12.5 mL, 1.6 ) in
hexane (20 mmol) in an argon atmosphere. Stirring the light-brown
mixture for 2 h at this temperature and then warming to 25 °C
over 10 h gave a black, completely soluble mixture of titanium()
diisopropoxide (7) containing lithium isopropoxide (2 equiv.). This
THF solution measured directly gave a strong EPR signal (cf. Re-
sults) and was employed for reactions with diphenylacetylene (6),
with cis-stilbene (15) and with cis-stilbene oxide (20).[12]
Upon removal of all volatiles from THF solutions of 7 and redis-
solving the residue in heptane or 1:1 mixtures of heptane and tolu-
ene (v/v), black complete solutions of the 1:2 mixtures of 7 and
LiOiPr resulted. For EPR studies, solutions of 7 in a 1:1 heptane/
Eur. J. Inorg. Chem. 2006, 1968–1975
© 2006 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.eurjic.org
1973

J. J. Eisch, J. N. Gitua, D. C. Doetschman
FULL PAPER
toluene mixture were prepared individually at concentrations of
20.0, 7.0, 2.0, and 0.7 m. That the THF had been completely re-
moved was indicated by 1H NMR spectra of these hydrocarbon
solutions of 7, which showed no 1H triplets of THF at δ = 3.72
and 1.84 ppm. Instead, only broad 1H signals of the isopropoxy
group were evident at around 4.0 and 1.2 ppm. These hydrocarbon
solutions also displayed strong EPR signals.
erization of cis-stilbene: 50% of trans-stilbene (17), 15% of biben-
zyl, and 35% of the starting 15.
(3) A reaction identical to run (1), except that a 3-h reflux period
was included, yielded 64% of trans-stilbene (17), 27% of bibenzyl,
and 9% of the starting 15.
(4) One-half of the reaction mixture in run (3) was worked up with
D2O (min. 98%). The cis- and trans-stilbenes were found to contain
no deuterium, but the bibenzyl was found to be 1,2-dideuteriated
(PhCHD-CHDPh, 19).
Preparation of Titanium(II) Diisopropoxide (7, free of LiOiPr): To
a stirred solution of titanium() chloride (1.89 g, 10 mmol) in THF
(40 mL), prepared and maintained under an argon atmosphere at
–78 °C, in order to avoid cleavage of the THF ring by the TiCl4,
was added slowly with stirring nBuLi (2 equiv.) in hexane (12.5 mL,
1.6 , 20 mmol), and the resulting mixture was stirred for 30 min.
The di-n-butyltitanium() dichloride thereby formed was then
treated with isopropyl alcohol (2 equiv., 1.20 g, 20 mmol), also at
–78 °C. The resulting mixture was warmed to 25 °C over 10 h, and
the dark brown suspension of the proposed titanium() dichloride
diisopropoxide was cooled again to –78 °C. Finally, a further ali-
quot of nBuLi (2 equiv.) was added. A subsequent warming of this
last reaction product gave a black solution of titanium() diisopro-
poxide from which the white LiCl slowly precipitated. Concentra-
tion and cooling of the solution permitted about 1.6 g (94%) of
the LiCl to be removed. Thus the yield of titanium() seems to
have been almost quantitative.[12]
Diphenylacetylene (6)
(1) A solution of 6 (5.0 mmol, 890 mg) and 7 (10 mmol, containing
LiOiPr) in THF (40 mL) was heated at reflux for 12 h and then
subjected to the usual hydrolytic workup. Flash column chromatog-
raphy of the product by using a 60:1 (v/v) hexane/THF mixture
showed the presence of 74% of cis-stilbene (15), 5% of bibenzyl,
1% of hexaphenylbenzene, and 20% of the starting 6.
(2) A reaction mixture identical with the foregoing was stirred for
1
24 h at 25 °C. Hydrolysis and a similar H NMR analysis revealed
the presence of only 30% of cis-stilbene (15) and 5% of bibenzyl,
the rest being unreacted 6. A deuteriolytic workup produced exclu-
sively cis-1,2-dideuterio-cis-stilbene (9, with deuterons at the vinyl
carbons).
(3) A reaction mixture treated as in run (2), except that 7 did not
contain any LiOiPr, gave again 31% of 15 and 4% of bibenzyl,
essentially in the same product ratio as in section (2).
The careful examination of such titanium() diisopropoxide solu-
tions, which contain little LiOiPr or LiCl, will have to await further
refinements in mixing the reactants and filtering the solutions to
remove suspected paramagnetic particles in suspension. It will be
of great interest to learn whether solutions of absolutely pure titani-
um() diisopropoxide show any EPR activity. Such information
will be valuable in determining what the role of LiOiPr in influenc-
ing the triplet character of trimeric clusters of [Ti(OiPr)2]3, com-
mingled with LiOiPr.
cis-Stilbene Oxide (20): A mixture of 20 (5.0 mmol) and 7
(10 mmol) in THF (30 mL) was allowed to react for 12 h at 25 °C.
1
Hydrolytic workup and H NMR analysis of the flash-chromatog-
raphy eluate obtained with a 60:1 (v/v) mixture of hexane/THF
showed the presence of 91% of trans-stilbene (17) and 9% of its
cis isomer (15).
Preparation of Titanium(II) Chloride (5): In a procedure analogous
to the foregoing, a stirred solution of titanium() chloride (1.89 g,
10 mmol) in THF (40 mL) was treated at –78 °C under argon with
n-butyllithium (20 mmol) in hexane. Further stirring at –78 °C and
warming to 25 °C yielded a gray-black suspension of titanium()
chloride (5) and lithium chloride as a 1:2 mixture. The resulting
black solution of 5 with suspended LiCl showed no EPR signal,
either at 25 °C or at lower temperatures.
As already has been published,[14] titanium() chloride (5) can be
isolated free of LiCl from the suspension in THF by removing the
THF and all volatiles under reduced pressure in argon. The residue
is then slurried with toluene, and the suspension is filtered to re-
move all the LiCl. The filtrate is concentrated and chilled, in order
to deposit black crystals, whose analysis corresponds to a 1:2 ratio
of Ti/Cl and a Ti analysis agreeing with the empirical formula of
TiCl2·THF (5). Solutions of 5 in toluene again showed no EPR
signal either at room temperature or at 80 K.
Acknowledgments
The authors are grateful to the Boulder Scientific Company, Mead,
Colorado, for the general support of their research with transition-
metal alkyls and to the Alexander von Humboldt Foundation for
a Senior Scientist Award to the first-cited author, renewed in 2005
for studies in the laboratory of Professor W. A. Herrmann, Institute
of Inorganic Chemistry, Technische Universität München, Munich,
Germany.
Furthermore, they thank the Regional Facility for Pulsed EPR and
Photochemistry at Binghamton University, established with
National Science Foundation support (CHE9512274), for its ser-
vices in performing the EPR experiments.
[1] J. J. Eisch, S.Dutta, J. N. Gitua, Organometallics 2005, 24,
6291–6294.
Reactions of Titanium(II) Isopropoxide with cis-Stilbene (15)
[2] J. J. Eisch, X. Ma, K. I. Han, J. N. Gitua, C. Krüger, Eur. J.
Inorg. Chem. 2001, 77–88.
[3] J. J. Eisch, J. Organomet. Chem. 2001, 617–618, 148–157.
[4] J. J. Eisch, J. N. Gitua, P. O. Otieno, X. Shi, J. Organomet.
Chem. 2001, 624, 229–238.
[5] a) J. J. Eisch, J. N. Gitua, Organometallics 2003, 22, 24–26; b)
J. J. Eisch, A. A. Adeosun, J. N. Gitua, Eur. J. Org. Chem. 2003,
4721–4727.
(1) The reaction of 15 (5.0 mmol, 0.89 mL) with 7 (10 mmol, con-
taining LiOiPr) in THF (30 mL) for 24 h gave a product (880 mg)
after hydrolytic workup with aqueous HCl (6 ) and extraction of
the organic products into ethyl ether, drying of the ether with
Na2SO4, and removal of volatiles by rotary evaporation. Analysis
1
of the product by H NMR spectroscopy showed the presence of
[6] J. J. Eisch, J. N. Gitua, Organometallics 2003, 22, 4172–4174.
[7] A. Carrington, A. D. McLachlan, Introduction to Magnetic
Resonance, Wiley, New York, 1979.
[8] J. H. Van der Waals, M. S. DeGroot, J. Chim. Phys. 1964,
1643–1654.
35% of trans-stilbene (17), 25% of bibenzyl, and 40% of the start-
ing 15.
(2) A reaction identical to that given above, except that 7 did not
contain any LiOiPr, produced a somewhat greater amount of isom-
1974
www.eurjic.org
© 2006 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Eur. J. Inorg. Chem. 2006, 1968–1975

Epimetallation of π -Bonded Organic Substrates with Titanium() Isopropoxide
FULL PAPER
[9] S. P. McGlynn, T. Azumi, M. Hinoshite, Molecular Spec-
troscopy of the Triplet State, Prentice-Hall, Englewood.
[10] D. J. Hinz, G. Meyer, T. Dedecke, W. Urland, Angew. Chem.
Int. Ed. Engl. 1995, 34, 71–73.
[11] J. J. Eisch, D. C. Doetschman, unpublished studies in progress.
[12] J. J. Eisch, J. N. Gitua, Eur. J. Inorg. Chem. 2002, 3091–3096.
[13] J. J. Eisch, Organometallic Synthesis, vol. 2, Academic Press,
New York, N. Y., 1981, ch. 1.
[14] J. J. Eisch, X. Shi, J. Lasota, Z. Naturforsch., Teil B 1995, 50,
342–350.
Received: December 1, 2005
Published Online: March 20, 2006
Eur. J. Inorg. Chem. 2006, 1968–1975
© 2006 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.eurjic.org
1975