Versatile zirconium reductants and carbon-carbon coupling agents selectively accessible from the 2:1 molar aggregate of n-butyllithium and zirconium(IV) salts

Article abstract of DOI:10.1002/ejoc.201100307

In previous studies of transition metal alkyls the 2:1 molar aggregate of n-butyllithium and zirconium(IV) salts has been found to react both with benzylic hydrocarbons and aromatic carbonyl derivatives in diverse and useful ways. In the present study the reactions of the aggregates, 2nBuLi· ZrE₄ (E = Cl, OEt), with benzaldehyde have involved carbometallation, hydrometallation and reductive dimerization (paths 1-3) in THF and were selectively achievable by temperature control alone. First, at -78 °C benzaldehyde underwent carbolithiation to give upon hydrolysis 1-phenyl-1-pentanol. However, short-term reaction times and prompt D ₂O quenching revealed that with Zr(OEt)₄ both benzaldehyde and 1-phenyl-1-pentanol were deuteriated, consistent with the presence of a phenyl(lithioxy)carbene intermediate. The observed dimerization of benzaldehyde to benzyl benzoate by lithium 2,2,6,6-tetramethylpiperidide is also consistent with such a phenyl(lithioxy)carbene intermediate. Second, at 25 °C the 2nBuLi·ZrE₄ aggregate reduced benzaldehyde exclusively to benzyl alcohol, which observation is consistent with the formation of the hydrozirconating agent, H₂ZrE₂. Third, heating the aggregate at reflux and subsequent reaction with benzaldehyde produced solely the reduced dimer, 1,2-diphenyl-1,2-ethanediol with high stereoselectivity: E = Cl, rac/meso of 93:7 and E = OEt, rac/meso of 100:0. The proposed mechanism involves the formation of ZrE₂, the epizirconation of benzaldehyde and the insertion of the second benzaldehyde into the zirconaoxacyclopropane under steric control. Finally, the high selectivity in hydrozirconation and reductive dimerization shown by 2nBuLi·ZrE₄ appears at this time to be superior to that attainable with analogous titanium or hafnium aggregates. The aggregate 2 nBuLi·ZrE₄ (E = Cl, OEt) in THF undergoes highly selective reactions with benzaldehyde controllable by temperature: 1) at -78 °C carbolithiation gives 1-phenyl-1-pentanol; 2) at 25 °C benzyl alcohol is exclusively formed; and 3) after reflux reductive dimerization to the glycol by ZrE₂ takes place with high selectivity (rac/meso ≈ 95:5). Copyright

Full text of DOI:10.1002/ejoc.201100307

FULL PAPER
DOI: 10.1002/ejoc.201100307
Versatile Zirconium Reductants and Carbon–Carbon Coupling Agents
Selectively Accessible from the 2:1 Molar Aggregate of n-Butyllithium and
Zirconium(IV) Salts[‡]
John J. Eisch,*[a] John N. Gitua,[b] and Kun Yu[a]
Keywords: Zirconium / Zirconation / Hydrozirconation / Lithium / Lithiation / Carbenes / Dimerization / Reduction
In previous studies of transition metal alkyls the 2:1 molar
aggregate of n-butyllithium and zirconium(IV) salts has been
found to react both with benzylic hydrocarbons and aromatic
carbonyl derivatives in diverse and useful ways. In the pres-
ent study the reactions of the aggregates, 2nBuLi·ZrE4 (E =
Cl, OEt), with benzaldehyde have involved carbometallation,
hydrometallation and reductive dimerization (paths 1–3) in
THF and were selectively achievable by temperature control
alone. First, at –78 °C benzaldehyde underwent carbolithi-
ation to give upon hydrolysis 1-phenyl-1-pentanol. However,
phenyl(lithioxy)carbene intermediate. Second, at 25 °C the
2nBuLi·ZrE4 aggregate reduced benzaldehyde exclusively to
benzyl alcohol, which observation is consistent with the for-
mation of the hydrozirconating agent, H2ZrE2. Third, heating
the aggregate at reflux and subsequent reaction with benzal-
dehyde produced solely the reduced dimer, 1,2-diphenyl-1,2-
ethanediol with high stereoselectivity: E = Cl, rac/meso of
93:7 and E = OEt, rac/meso of 100:0. The proposed mecha-
nism involves the formation of ZrE2, the epizirconation of
benzaldehyde and the insertion of the second benzaldehyde
short-term reaction times and prompt D2O quenching re- into the zirconaoxacyclopropane under steric control. Finally,
vealed that with Zr(OEt)4 both benzaldehyde and 1-phenyl-
1-pentanol were deuteriated, consistent with the presence of
a phenyl(lithioxy)carbene intermediate. The observed dimer-
ization of benzaldehyde to benzyl benzoate by lithium
2,2,6,6-tetramethylpiperidide is also consistent with such a
the high selectivity in hydrozirconation and reductive dimer-
ization shown by 2nBuLi·ZrE4 appears at this time to be su-
perior to that attainable with analogous titanium or hafnium
aggregates.
such as toluene (9), diphenylmethane (10) and 1,1,2,2-tetra-
phenylethane (14).[2]
Introduction
To the dismay but the ultimate delight of the research
chemist, admixing a given set of starting reagents can often
lead to two or more different products, whose formation
is individually dependent on the experimental variables of
temperature, time, concentration and ratio of reactants. We
have encountered such an instructive and versatile competi-
tion of kinetic and thermodynamic processes with zirconi-
um(IV) chloride (1) or zirconium(IV) ethoxide (2) and n-
butyllithium (3) in a 1:2 equiv. ratio in THF. In previous
studies we have observed that such a combination of rea-
gents can lead to the various important transformations de-
picted in Scheme 1. On the right side are shown the three
reactions (paths 1–3) applicable to the typical aromatic car-
bonyl derivative, benzaldehyde (4); on the left side in paths
4–6 are shown reactions found with benzylic hydrocarbons,
Finally, because the 2:1 aggregate of n-butyllithium (3)
with Zr(OEt)2 (2) in THF can lithiate benzylic hydro-
carbons, such as toluene at 0 °C (path 4 in Scheme 1), we
have been prompted to search for the possible involvement
of lithiation at the carbonyl group in the reaction of benzal-
dehyde (4) at low temperatures. In such a possible lithiation
(path 7 in Scheme 1) the initially formed lithium derivative
15a would be more stable as phenyl(lithioxy)carbene 15b,
with the anionic charge on oxygen.[3]Such carbenes are
known to be immediately trapped by ambient RLi or car-
bonyl reagents.[4] As part of the quest for likely routes to
carbenes like 15b, we have also explored the behavior of
benzaldehyde (4) toward hindered bases, such as lithium
2,2,6,6-tetramethylpiperidide (LTMP).[5]
In the present study we have focused on the individual
reactions of benzaldehyde (4), given in paths 1, 2, 3 and 7,
in order to determine the specific lithium or zirconium rea-
gent responsible for each transformation and to achieve, if
possible, each individual transformation in high yield and
with great selectivity. We have chosen to defer further study
of the reactions of the 2:1 n-butyllithium-ZrCl4 and
-Zr(OEt)4 aggregates with benzylic hydrocarbons 9, 10 and
14 in paths 5–7, because such reactions are complex, involv-
[‡] Organic Chemistry of Subvalent Transition Metal Complexes,
47. Part 46: Ref.[1]
[a] Department of Chemistry, State University of New York at
Binghamton,
P. O. Box 6000, Binghamton, NY, 13902-6000, USA
Fax: +1-607-777-4865
E-mail: jjeisch@binghamton.edu
[b] Chemistry Department, Drake University,
207 University Avenue, Des Moines, Iowa 50311, USA
E-mail: john.gitua@drake.edu
Eur. J. Org. Chem. 2011, 3523–3530
© 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
3523

J. J. Eisch, J. N. Gitua, K. Yu
FULL PAPER
Scheme 1. All reactions were worked up with H2O or D2O. The individual products shown from paths 1–6 were in fact found but in
widely varying yields, which were not optimized (cf. text).
ing thermal, photochemical and/or iron-promoted reaction though some complexation between 3 and 2 may have oc-
steps and often causing unusual side products to domi- curred (cf. infra), the aggregate still underwent the same
nate.[6]
chemical reaction of reagent 3 alone and added quantita-
tively to benzaldehyde (4) to yield upon hydrolysis the ex-
pected 1-phenyl-1-pentanol (5) [Equation (1)].[8,9] The 2:1
aggregate of 3 and 1 gave identical results.
Results and Discussion
General Observations on Reactions of the 2:1 Aggregate of
n-Butyllithium (3) and Zirconium(IV) Chloride (1) or
Zirconium(IV) Ethoxide (2) with Benzaldehyde (4)
After preliminary reaction runs with the 2:1 aggregates
of 3 with 1 or with 2 in THF at various temperatures, it
was determined that path 1 was solely followed at –78 °C,
path 2 was completely selective at 25–30 °C, and path 3 was
the only reaction course after the reaction mixture was pre-
heated at reflux (Scheme 1).
Possible Role of the Phenyl(lithioxy)carbene (15b) in the
Reaction of n-Butyllithium (3) and Zirconium(IV) Ethoxide
with Benzaldehyde in THF at –78 °C
The 2:1 aggregates of 3 with 1 or with 2 in THF existed
as a suspension or solution, which was pale yellow at
–78 °C, orange at 25 °C and dark brown after several hours
at reflux. As individual reactions of these three with benzal-
dehyde (4) have shown, the benzaldehyde samples were
quantitatively transformed under these individual condi-
tions into 1-phenyl-1-pentanol (5), benzyl alcohol (6) and
the 1,2-diphenyl-1,2-ethanediols, respectively, which diols
were preponderantly the rac-isomer (7) over the meso-iso-
mer (8). Unless otherwise stated, the 2:1 aggregates of 3
with 1 or with 2 in THF gave very similar results when
admixed with benzaldehyde (4) under the stated conditions
and at the given temperature. Therefore, the observed reac-
tions will be depicted and discussed with the 2:1 n-butyllith-
ium–Zr(OEt)4 aggregate. Where there were any significant
differences between 1 and 2, a word of explanation will be
offered.
Since two equivalents of n-butyllithium are involved in
the reaction with benzaldehyde, it is possible that competing
with straightforward carbolithiation [4Ǟ5a, Equation (1)],
known for over 80 years,[8,9] might be the lithiation of 4 at
the carbonyl group (15b) (Scheme 2). There is rich pre-
Formation of the 2:1 Aggregate of n-Butyllithium (3) and
Zirconium(IV) Ethoxide (2) at –78 °C and Reaction with
Benzaldehyde (4)
Admixing n-butyllithium (3) with Zr(OEt)4 (1) in THF
at –78 °C seems to cause no profound chemical change. Al- Scheme 2.
3524
www.eurjoc.org
© 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Eur. J. Org. Chem. 2011, 3523–3530

Versatile Zirconium Reductants and C–C Coupling Agents
cedent for the alpha-addition of organolithium reagents to Such alkoxide coordination may also be operative in en-
carbenes, as shown by the transformation of 15b to 5b.[4] hancing the lithiating activity of organolithium reagents by
One such example is illustrated in Scheme 3.[10]
potassium alkoxides.[13]
Lithiation of Benzaldehyde (4) at the Carbonyl Group by
Lithium 2,2,6,6-Tetramethylpiperidide (LTMP)
The acidity of the aldehydic H–C bond of 4 can be esti-
mated by the following polarographic measurements of pKa
values: benzene, 37; ethylene 36 and the formyl proton of
methyl formate, 35.[14a] Accordingly, the pKa of benzalde-
hyde (4) should be certainly smaller than that of benzene,
whose H–C bonds can be lithiated in THF and especially
in TMEDA. Hence, deprotonation of benzaldehyde was at-
tempted with LTMP, a strong deprotonating base with little
nucleophilic tendency to attack the carbonyl group because
of steric hindrance. Indeed, when benzaldehyde is treated
with LTMP in hexane, it is dimerized readily and upon hy-
drolysis yields benzyl benzoate. This process can be satisfac-
torily explained by the lithiation of benzaldehyde (4) by
LTMP to produce carbene 15b and its carbene-like addition
to a second molecule of 4. Opening of the cyclic intermedi-
ate would provide the lithium derivative (16a) of benzyl
benzoate (16) (Scheme 4).[14b,14c]
Scheme 3.
If 5b is sufficiently stable in THF, it might persist until
workup with D2O and dideuterio 5c might be isolable. But
deprotonation of THF, producing ethylene and the lithium
enolate of acetaldehyde,[11] might instead produce O-deu-
terio-5d. In the following section an attempt has been de-
scribed for generating carbene 15b in a straightforward
manner.
In order to trap intermediates, such as 5b and 15b, reac-
tion mixtures of a 2:1 ratio of n-butyllithium and Zr(OEt)4
with subsequently added benzaldehyde were quenched with
D2O (98%) at different stages: a) the reaction mixture
warmed finally to 25 °C gave only 5d and no 5c; b) the reac-
tion mixture promptly brought to 25 °C likewise gave only
5d; and c) the reaction mixture mixed for 15 min at –78 °C
and promptly quenched with D2O gave 95% of 5 and 5%
of benzaldehyde (4). Alcohol 5 contained about 13% of 5c
and the benzaldehyde had about 2% of 4 as C6H5–CD=O
(4a). Thus about 15% of 4 was lithiated to 15b (Scheme 2)
and most of 15b converted to 5b.
Scheme 4.
It is noteworthy that the 2:1 aggregate of n-butyllithium
(3) and zirconium(IV) chloride (1) showed no sign of lithiat-
ing benzaldehyde. This is consistent with the similar failure
of this aggregate to lithiate toluene or diphenylmethane be-
tween –78 and 0 °C (path 4, Scheme 1), a reaction achieved
by 2 in the nBuLi–Zr(OEt)4 aggregate.[12]
For this reason, we suggest that the 2:1 nBuLi-Zr(OEt)4
aggregate at –78 °C has not yet undergone the intermetallic
ligand exchange depicted in Scheme 5. Rather the C–Li
bond in 3 is further polarized by a chelating coordination
of 2 with 3 as proposed in complex 2–3 of Equation (2).
Formation of the 2:1 Aggregate of n-Butyllithium (3) and
Zirconium(IV) Chloride (1) at 25 °C and Subsequent
Reaction with Benzaldehyde (3)
When the 2:1 aggregate of nBuLi (3) and ZrCl4 (1) was
brought to 25 °C, stirred and recooled to –78 °C (path 2),
addition of benzaldehyde (4) and subsequent workup now
gave a quantitative yield of only benzyl alcohol (6), indica-
tive that transmetallation had in fact taken place to produce
di-n-butylzirconium(IV) dichloride (17) and what is typical
of transition metal alkyls, 17 was in equilibrium with zirco-
nium(IV) dichloride dihydride 17a and 1-butene. The zirco-
nium reagent 17a had then performed a monomeric hydro-
zirconation of benzaldehyde (4) to yield 19 and only benzyl
alcohol (6) upon hydrolysis (Scheme 5). The generation of
the hydrozirconating agent H2ZrCl2 (17a) in this way thus
proves to be more convenient and less costly than the prep-
aration of the commercially available Cp2ZrHCl or
Schwartz reagent.[15] Zirconium(IV) dichloride dihydride
(17a) has been successfully used thus far for the hydrozir-
conation of 1-alkenes, alkynes, carbonyl derivatives, imines,
nitriles and epoxides.[16,17] The 2:1 aggregate of 3 and 2 gave
identical results.
Eur. J. Org. Chem. 2011, 3523–3530
© 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.eurjoc.org
3525

J. J. Eisch, J. N. Gitua, K. Yu
FULL PAPER
second benzaldehyde into the sterically more accessible side
of the C–Zr bond of 20 to yield 21.[18] An alternative reac-
tion involving the coupling of zirconium radical intermedi-
ates (20a) seems less satisfactory. Slow coupling of 20a radi-
cals would favor the formation of the meso-dimer for steric
reasons, while rapid coupling would tend to favor compar-
able amounts of rac- and meso-dimers (Scheme 7).
Scheme 5.
Status of Reductions of Organic Compounds by Low-Valent
Titanium Reagents
Heating the 2:1 Aggregate of n-Butyllithium (3) and
Zirconium(IV) Chloride (1) at Reflux in THF and Reaction
Thereafter with Benzaldehyde (4)
The pioneering and revolutionary studies of Karl Ziegler
and co-workers with early transition metal salts (TiCl4,
ZrCl4 and VCl4) in combination with aluminum alkyls in
the 1950s led to the discovery of highly active catalysts for
the oligomerization and/or polymerization of olefins and
diolefins.[19–21] Thereafter, early transition metals were in-
tensively studied as reagents for organic transformations,
with the result that in 1973 three research groups reported
independently on the novel reductive dimerization of car-
bonyl derivatives to olefins by various subvalent titanium
reagents[22–24] [Equation (3)].
Finally, when the foregoing solution of nBu2ZrCl2 (17)
in THF is allowed to reflux for several hours and benzalde-
hyde then added (path 3), the only products obtained upon
hydrolysis were the 1,2-diphenyl-1,2-ethanediols, in essen-
tially quantitative yield and as a 93:7 ratio of the racemic
and meso isomers, 7, and 8. When the solution of
nBu2Zr(OEt)2 (18) was treated in an analogous way, only
the rac-isomer was formed. Thus in this overall reaction the
reductive dimerization of benzaldehyde was achieved with
very high (nBu2ZrCl2, 17) to exclusive [nBu2Zr(OEt)2, 18]
stereoselectivity (Scheme 6). The most likely steps in path 3
are the reductive elimination of butyl groups from n-
BuZrE2 to zirconium salt ZrE2 (19), the epizirconation of
benzaldehyde by ZrE2 to form 20 and the insertion of the
The transformation depicted in Equation (3), which has
come to be termed the McMurry reaction after one of its
discoverers, is a most impressive and versatile method for
the construction of a wide variety of open-chain and ring
carbon skeletons in high yields.[25,26] Unfortunately, the na-
ture of various subvalent titanium reductants used for the
McMurry reaction have largely remained shrouded in un-
certainty: in almost all applications, neither the exact oxi-
dation state(s) of the titanium reductants in solution nor
their mechanism(s) of reaction can be confidently specified,
even after almost 40 years. Problems in formulating a mo-
lecular interpretation of this reaction are not difficult to
identify: (1) the titanium reductant is generated in a hetero-
geneous medium; (2) the presence of more than one subval-
ent state [Ti0 and TiII] often cannot be ruled out; (3) resid-
ual reductant used to generate subvalent titanium (LiAlH4,
Mg, Al, Zn) may play a role; and (4) the Lewis acid formed
from generating subvalent titanium (LiCl, MgCl2, AlCl3,
ZnCl2) can and almost surely does influence the reducing
character of the titanium reagent.
Scheme 6.
Therefore, if greater mechanistic insight into the course
of the McMurry reaction was to be achieved or if superior
reagents were to be found for carrying out such reductive
couplings on a wider variety of C=O, C=N, C=S and CϵN
derivatives, low-valent transition metal complexes of Ti and
other early transition metals were required that would have
the following attributes: (1) a soluble metal reagent; (2) the
metal present in one well-defined oxidation state; (3) the
absence of any residual, secondary reductant (main group
metal or metal hydride); and (4) absence of any possibly
disturbing Lewis acid (main group metal halide).
Scheme 7.
3526
www.eurjoc.org
© 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Eur. J. Org. Chem. 2011, 3523–3530

Versatile Zirconium Reductants and C–C Coupling Agents
Because of investigations undertaken by our research revealed that the unreacted benzaldehyde was partially de-
group in the 1990s, reductants of low-valent early transition uteriated at the carbonyl as C6H5–CD=O and that the
metal salts, largely fulfilling the aforementioned criteria, alcohol 5 contained some deuterium as in 5c. This is clear
were found to be readily accessible by the reductive alky- evidence that the Zr(OEt)4 in the 2 nBuLi·Zr(OEt)4 aggre-
lation of high-valent starting metal salts with organo- gate had accelerated the lithiation of benzaldehyde.
lithium, -magnesium or -aluminum reagents in a variety of
Second, the reaction of either aggregate, 1 and 3 or 2 and
nonprotic polar or nonpolar media. Such a reductive alkyl- 3, with benzaldehyde at 25 °C was identical: after hydrolysis
ation can best be exemplified by the prototypical generation a quantitative yield of benzyl alcohol was obtained. This
of titanium(II) chloride reported in 1995[27] (Scheme 8). outcome is perfectly consistent with transmetallation occur-
This convenient, high-yielding preparation provided TiCl2 ring within the original aggregate to yield nBu2ZrE2 and 2
in a form most useful for exploring the reduction of many LiE (E = Cl, OEt) and the facile beta-elimination of the
organic substrates.[2,18,28]
former to provide the useful hydrozirconating agent,
H2ZrE2.
Third, refluxing the former hydrozirconating agent in
THF led to the complete elimination of its butyl groups
and the formation of ZrCl2 or Zr(OEt)2, respectively. Reac-
tion with benzaldehyde and subsequent hydrolysis now led
exclusively to the 1,2-diphenyl-1,2-ethanediol system.
Moreover, the stereoselectivity for the racemic diol 7 over
the meso-diol 8 was most gratifying: ZrCl2 produced a
rac/meso ratio of 93:7 and Zr(OEt)2 yielded solely the rac-
diol 7. Therefore, by careful temperature control alone the
2:1 aggregate 2BuLi·ZrE2 can function as a highly selective
hydrozirconating agent or reductive coupling reagent to-
ward benzaldehyde. As previous work has already shown,
such reductions are readily applicable to other carbonyl, ep-
oxide and acetylenic and olefinic systems.[16,17]
Scheme 8.
Exploring the possible scope of this reaction has led to
successful preparation of analogous complexes of ZrCl2,
HfCl2, VCl and CrCl. In addition, the use of an excess of
n-butyllithium with VCln (n = 3 or 4) or CrCl3 in THF has
led to the stoichiometric formation of low-valent hydridic
anionic salts of V(I) or Cr(I), as depicted in Scheme 9.[30,31]
Given the demonstrated versatility of the zirconium rea-
gents discussed here, one might wonder how the titanium
and hafnium derivatives compare with those of zirconium.
Current studies indicate that pure samples of nBu2TiE2 are
difficult to prepare, because such samples tend to decom-
pose into TiCl2 rapidly above 0 °C.[16b] Secondly, samples
of TiCl2 in THF can readily bring about the reductive cou-
pling of benzaldehyde at 25 °C but with distinctly less
stereoselectivity: a typical rac/meso ratio observed is
70:30.[33] On the other hand, the nBu2HfE2 can be readily
generated at 25 °C but it has proved to be mostly thermally
stable to decomposition. Thus, in reactions of such a rea-
gent with ketones, reduction to the monomeric alcohol pre-
dominates and little coupling to the diol occurs.[16b] Thus it
appears that in terms of reactivity and selectivity in such
reductions, zirconium reagents will remain superior to those
of titanium and hafnium.
Scheme 9.
Evaluation of the relative reducing powers of these neu-
tral and anionic reagents has been the ongoing focus of our
investigations and the role of zirconium reductants are ad-
dressed in this present report.
It should be noted that empirical combinations of low-
valent transition metal salts and a second component,
either as reductant or as Lewis base, continue to appear in
the literature but no decisive evidence is presented concern-
ing the oxidation state of the transition metal actually
operative in such reductions.[32]
Experimental Section
Summary and Conclusions
Instrumentation, Analysis and Starting Reagents: The n-butyllith-
ium (1.6 m in hexane), and the zirconium(IV) chloride and zirconi-
um(IV) ethoxide (both anhydrous and of 97% purity) were used as
received from Aldrich Chemical Company. All reactions were car-
ried out under a positive pressure of anhydrous, oxygen-free argon.
All solvents employed with organometallic compounds were dried
and distilled from a sodium metal-benzophenone ketyl mixture
prior to use.[34] The IR spectra were recorded with a Perkin–Elmer
instrument (model 457), and samples were measured either as min-
eral oil mulls or as KBr films. The NMR spectra (1H and 13C)
Careful investigation of the reactions between benzalde-
hyde (4) and the 2:1 molar aggregates of n-butyllithium (3)
with either zirconium(IV) chloride (1) or zirconium(IV)
ethloxide (2) in THF as a function of temperature has es-
tablished the following selectivities in reaction. First, at
–78 °C the aggregates react with benzaldehyde, followed by
hydrolysis, to give quantitatively the same product produced
by n-butyllithium alone, namely 1-phenyl-1-pentanol (5).
But in the case of the 2:1 aggregate of 3 and 2 the reaction
with benzaldehyde, followed by prompt quench with D2O, were recorded with a Bruker spectrometer (model EM-360) and
Eur. J. Org. Chem. 2011, 3523–3530
© 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.eurjoc.org
3527

J. J. Eisch, J. N. Gitua, K. Yu
FULL PAPER
tetramethylsilane (Me4Si) was used 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 GC/MS mea-
1) Stirring for Two Hours at –78 °C Followed by Addition of Benzal-
dehyde (4) and Subsequent Workup with H2O: After 2 h at –78 °C
the white suspension was treated with 210 mg (0.20 mL, 2.0 mmol)
surements and analyses were performed with a Hewlett–Packard of benzaldehyde (4) (freshly distilled under argon), whereupon the
GC 5890/Hewlett–Packard 5970 mass-selective detector instru-
ment. The gas chromatographic analyses were carried out with a
Hewlett–Packard instrument (model 5880) provided with a 2-m
OV-101 packed column or with a Hewlett–Packard instrument
(model 4890) having a 30-m SE-30 capillary column. Melting
points were determined on a Thomas-Hoover Unimelt capillary
melting point apparatus and are uncorrected.
reaction mixture turned pale yellow. After stirring at –78 °C for an
additional 2 h and warming to 25 °C over 24 h the orange reaction
solution was quenched with 15 mL of water. Hydrolytic workup
with ether extraction, drying the organic extract with anhydrous
Na2SO4 and solvent removal by rotary evaporation gave 35 mg of
1
liquid (≈ 100%), which by H and 13C NMR analysis proved to be
solely 1-phenyl-1-pentanol (5). Such spectra were superimposable
[35]
with the published spectrum in CDCl3 and lacked any trace of
Reaction of Lithium 2,2,6,6-Tetramethylpiperidide (LTMP) with
Benzaldehyde (4)
characteristic 1H absorptions for benzyl alcohol (6), such as the
peak at δ = 5.80 ppm, or for rac- and meso-1,2-diphenyl-1,2-ethane-
diol (7 and 8) at δ = 4.57 and 4.75 ppm, respectively.
Generation of Lithium 2,2,6,6-Tetramethylpiperidide (LTMP): To
10 mL of anhydrous, deoxygenated hexane at –78 °C were intro-
duced 1.00 mL (5.8 mmol) of 2,2,6,6-tetramethylpiperidine
(Ն 99%) and then 3.60 mL (5.8 mmol) of 1.6 m n-butyllithium in
hexane. The resulting orange solution became a yellow suspension
of LTMP and was kept at 0 °C for 3 h.
2) Stirring for Two Hours at –78 °C, Followed by Addition of Benzal-
dehyde (4) and Subsequent Treatment with D2O: a) Normal pro-
cedure: Repetition of the procedure exactly as given in section 1),
except that D2O (98%) was substituted for H2O in the workup,
again gave solely 1-phenyl-1-pentanol (5) as the only product. But
(5) contained a deuteron in the hydroxy group (1H, 2H NNHR
peaks) and none at the C1 site (5a). b) Inverse addition: In this
procedure the mixture of the specified amounts of zirconium(IV)
ethoxide (2) and benzaldehyde in THF at –78 °C was treated with
the two equivalents of n-butyllithium. After 15 min. of stirring at
–78 °C the pale yellow suspension was warmed up to 25 °C over
2 h. The mixture then was treated with D2O in THF. Subsequent
workup showed the presence of only CH3CH2CH2CH2–CHPh–OD
(5b). c) Inverse addition with prompt quench with D2O at –78 °C: As
in procedure 2b the same mixture of 2 and 4 in THF at –78 °C was
treated with two equivalents of n-butyllithium. After the resulting
yellow suspension was stirred for 15 min, it was promptly quenched
at –78 °C with a solution of D2O (98%) in THF. Usual hydrolytic
workup led to a mixture of 1-phenyl-1-pentanol and 5% remaining
benzaldehyde, as shown by integrating the C-H peaks at 10.0 and
at 4.6 ppm in the 1H NMR spectrum. Furthermore, integrating the
methyne H of 5 vs. the methyl H of 5 gave a ratio of 0.87:3.0,
compared with the expected 1.0:3.0 ratio of undeuteriated 5. These
results are consistent with the presence of about 13% of deuterium
on the methyne C, as in 5c: CH3CH2CH2CH2–CD(OH)–C6H5. Ad-
ditional confirmation for the presence of 5c was obtained from the
2H NMR spectrum of the reaction mixture: a small but distinct
peak was found at 4.6 ppm, where the methyne C-D peak would
occur. Finally, the same 2H NMR spectrum displayed a further
peak at δ = 10.0 ppm, which observation shows that the unreacted
4 was also partially deuteriated on the carbonyl carbon (4a).
Reaction of LTMP in Hexane with Benzaldehyde: To a solution of
6.00 mL (60 mmol) of benzaldehyde (freshly distilled under argon)
in 50 mL of toluene and kept at 0 °C was added the previously
prepared hexane solution of LTMP. The resulting orange solution
was left overnight at 25 °C and then hydrolyzed with water and
ether. The separated organic layer was dried with anhydrous
Na2SO4 and all volatiles removed by evaporation under reduced
1
pressure to yield 5.3 g of an orange liquid. The H NMR spectral
analysis of this liquid showed the presence of 52% (13 mmol) of
benzyl benzoate (16) (1H peak at 5.34 ppm), 7% (3.4 mmol) of
benzyl alcohol (6) (1H peak at 4.6 ppm) and 41% of benzaldehyde.
The identity of 6 and 16 were confirmed by individual separation
by column chromatographic on silica gel with hexane elution and
1
comparison with the H and 13C NMR spectra of authentic sam-
ples.[35]
The original aqueous layer from the reaction workup was evapo-
rated to about one-half its volume and then acidified with 1 n aque-
ous HCl. Cooling of the solution led to the deposition of 310 mg
(2.5 mmol.) of benzoic acid (IR and mp). The conclusion is that
the benzyl alcohol and benzoic acid most likely arose from the
saponification of 16 in the strongly alkaline aqueous solution prior
to workup.
Reactions of the 2:1 Aggregate of Zirconium(IV) Ethoxide (2) with
n-Butyllithium (3)
General Reaction Procedure and Hydrolytic Workup: All glassware
and needles for the transfer and reaction of liquid samples were
dried in an oven at 120 °C and while cooling were thoroughly
flushed with argon. The standard reaction apparatus was a 125-
mL, two-necked Schlenk flask, provided with a Teflon®-coated stir-
bar and having the argon inlet on one neck and a rubber septum
on the other neck. Then 30 mL of anhydrous, deoxygenated tetra-
hydrofuran (cf. supra) was introduced and cooled to –78 °C in a
solid CO2/acetone bath. Thereupon a tared 2.00 mmol sample
(560 mg) of zirconium(IV) ethoxide (2) (Ϯ0.10 mmol and corrected
for its 97% purity) in a glass-stoppered vial under argon was emp-
tied into the flask in the stream of emerging argon. To the reaction
suspension stirred magnetically at –78 °C were added dropwise by
gastight syringe 2.6 mL of 1.60 m n-butyllithium (3) in hexanes
(4.10 mmol). The resulting white suspension was subjected to stir-
ring at different temperatures for given times before the benzalde-
hyde (4) was introduced. All such reactions were conducted under
the diffuse fluorescent lighting of the laboratory provided during
the daytime.
3) Stirring While Allowing to Warm to 25 °C, Followed by Addition
of Benzaldehyde (4) at –78 °C: The white suspension, generated at
–78 °C, was warmed to 25 °C and then stirred for 1 h. The suspen-
sion turned into a pale yellow solution. The reaction mixture was
recooled to –78 °C and treated with 210 mg (0.20 mL, 2.0 mmol)
of freshly distilled 4. After warming to 25 °C over 24 h the orange
reaction mixture was hydrolyzed and worked up as in section1. The
pale yellow organic residue (400 mg) consisted of solely benzyl
1
alcohol (6) (100%) with some THF by H and 13C NMR analyses.
No characteristic absorptions for 1-phenyl-1-pentanol (5) or for
rac- or meso-1,2-dihenyl-1,2-ethanediol (7 or 8) could be detected.
An identical reaction run to the foregoing was carried out, except
that the workup quench employed D2O (98%). Again benzyl
alcohol (6) was the only product, containing a deuteron only in the
2
hydroxy group and none at the Cα of 6 (PhCH2–OD) (1H and H
NMR spectra).
3528
www.eurjoc.org
© 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Eur. J. Org. Chem. 2011, 3523–3530

Versatile Zirconium Reductants and C–C Coupling Agents
4) Stirring While Allowing to Warm to 25 °C, Followed by Heating
at Reflux and then Addition of Benzaldehyde (4) at 25 °C: A reaction
run identical to that in section 3 and also brought to 25 °C for 1 h
was then heated at reflux for 4 h. The resulting dark brown solution
was cooled to 25 °C and treated with 1.95 mmol of freshly distilled
4. After stirring for 24 h the brown mixture was hydrolyzed in the
[4] W. Kirmse, Carbene Chemistry, Academic Press, New York,
1964, p. 302; for reactions with organometallics, see: p. 40–41;
p. 92; p. 185–188.
1
usual manner. The pale organic liquid (450 mg) was shown by H
and 13C NMR spectroscopy to consist of THF and solely of rac-
1,2-diphenyl-1,2-ethanediol (7). No spectral absorptions character-
istic of 1-phenyl-1-penatnol (5) or benzyl alcohol (6) or meso-1,2-
dipihenyl-1,2-ethanediol (8) were detectable.
[5] J. J. Eisch, A. A. Adeosun, unpublished studies (March 2002)
of the interaction of benzaldehyde with lithium 2,2,6,6-tet-
ramethylpiperidine (LTMP) leading to the formation of benzyl
benzoate. These preliminary experiments have now been repro-
duced, expanded and are being published here for the first time.
[6] In a previous publication (J. J. Eisch, S. Dutta, Organometallics
2004, 23, 4181–4183) and in the doctoral dissertation (S. Dutta,
2005) upon which this publication was based, the principal
product accompanying the 1,1,2,2-tetraphenylethane (14) in
path 5 was mistakenly identified as tetraphenylethene. Indeed,
reinvestigation and careful analysis of this reaction by Mr. Kun
Yu of this laboratory has shown unambiguously that the major
product formed in 85% yield is in fact 5,5-diphenyllpentanol.[7]
Prior studies have indicated that this alcohol arises from THF
and diphenylmethyllithium by an electron-transfer cleavage of
the solvent: a) J. J. Eisch, J. Org. Chem. 1963, 28, 707–710; b)
H. Gilman, B. J. Gaj, J. Org. Chem. 1963, 28, 1725–1727.
[7] J. J. Eisch, K. Yu, Organometallics, 2011, accepted, under final
review.
Reactions of Zirconium(IV) Chloride (1) with n-Butyllithium (3): Re-
action runs were identical to the three sets of experimental condi-
tions employed above for 2.00 mmol of zirconium(IV) ethoxide (2)
with 4.10 mmol of n-butyllithium (3) in THF, except that
2.00Ϯ0.10 mmol of zirconium(IV) chloride (1) was substituted for
2. Details of the individual procedure and the hydrolytic workup
were exactly parallel to the corresponding numbered sections
above.
1) Stirring for Two Hours at –78 °C, Followed by Addition of Benzal-
dehyde (4): The isolated organic residue consisted of essentially
only 1-phenyl-1-pentanol with traces of THF.
An identical reaction run to the foregoing was carried out, except
that the workup-quench employed D2O (98%). As above, 1-phenyl-
1-pentanol (5) was the sole product, which contained a deuteron
only in the hydroxy group and none at the C1 of 5 (1H and 2H
NMR spectra).
[8] G. Wittig, M. Leo, Ber. Dtsch. Chem. Ges. 1930, 63, 943.
[9] H. Gilman, F. Breuer, J. Am. Chem. Soc. 1934, 56, 1127.
[10] J. J. Eisch, J. E. Galle, J. Organomet. Chem. 1988, 341, 293–313.
[11] A. Rembaum, S. P. Siao, N. Indictor, J. Polym. Sci. 1962, 56,
17.
[12] J. J. Eisch, S. Dutta, Organometallics 2004, 23, 4181–4183.
2) Stirring While Allowing to Warm to 25 °C, Followed by Addition
of Benzaldehyde (4) at –78 °C: As carried out and worked up as in
section 3 above, solely benzyl alcohol (6) was obtained in almost a
quantitative yield.
[13] M. Schlosser, J. Organomet. Chem. 1967, 8, 9.
[14] a) K. P. Butin, I. P. Beletskaya, A. N. Kashin, O. A. Reutov, J.
Organomet. Chem. 1967, 10, 197. b) The authors acknowledge
that basic reagents can cause aromatic aldehydes to dimerize
to benzoins or to disproportionate in a Tishchenko- or Canniz-
zaro-like manner. Our choice of the lithiated carbene interme-
diate 15b stems from the most recent work on formaldehyde
and its tautomeric hydroxycarbene (ref.[3]). With this excellent
precedent, we did not consider it necessary to try to reconcile
our results here with the excellent findings of Seyferth and co-
workers on acyllithium reagents (D. Seyferth, R. M. Weinstein,
R. C. Hui, W.-L. Wang, C. M. Archer, J. Org. Chem. 1991, 56,
5768–5773; D. Seyferth, R. M. Weinstein, R. C. Hui, W.-L.
Wang, C. M. Archer, J. Org. Chem. 1992, 57, 5620–5629). c)
The suggestion of a referee that the reaction mixture of Scheme
4 be worked up with D2O to learn whether 16 would be mono-
deuteriated was considered by the authors. However, previous
work has shown that LTMP at 0 °C in hexane can readily lithi-
ate 16 to produce 16a, which result destroys any probative
value to the test.
3) Stirring While Allowing to Warm to 25 °C, Followed by Heating
at Reflux and then Addition of Benzaldehyde (4) at 25 °C: As carried
out and worked up as in section 4 above, the product consisted of
THF and solely of a 93:7 mixture of racemic and meso-1,2-di-
phenyl-1,2-ethanediol (7 and 8).
Acknowledgments
The authors are grateful to Dr. John M. Birmingham for the en-
couragement and to the Boulder Scientific Company, Mead, Colo-
rado for financial support of this research on transition metal alkyl
compounds. In addition, the Alexander von Humboldt Foundation
has extended a Senior Scientist Award to the first-cited author,
which grant permitted fruitful visits and discussions with Professor
Günther Wilke at the Max Planck Institut für Kohlenforschung,
Mülheim (Ruhr), Germany. Special thanks go to Dr. A. A. Adeo-
sun for his initial and most helpful studies on the generation of
oxycarbenes.
[15] J. Schwartz, Angew. Chem. 1976, 88, 402; Angew. Chem. Int.
Ed. Engl. 1976, 15, 333.
[16] a) J. J. Eisch, F. A. Owuor, X. Shi, Organometallics 1999, 18,
1583–1585; b) X. Shi, Doctoral Dissertation, State University
of New York at Binghamton, May 1996.
[17] In the initial report on the generation of ZrCl2 at 25 °C in
THF (ref.[16a]) we mistakenly concluded that all Bu2ZrCl2 or
H2ZrCl2 had decomposed with loss of butane, 1-butene and
dihydrogen. Subsequent heating at reflux was then shown nec-
essary to complete the conversion to ZrCl2. We therefore judge
that both Bu2ZrCl2 and H2ZrCl2 are precursors to the final
ZrCl2.
[1] J. J. Eisch, Inorg. Chim. Acta 2010, 364, 3–9.
[2] J. J. Eisch, A. A. Adeosun, S. Dutta, P. O. Fregene, Eur. J. Org.
Chem. 2005, 2657–2670. This review offers a comprehensive
survey of such ramified reactions.
[3] Recent studies of the high-vacuum flash pyrolysis (HVFP) of
glyoxylic acid (a) leads, via immediate matrix isolation (Ar,
[18] J. J. Eisch, X. Shi, J. R. Alila, S. Thiele, Chem. Rev./Recueil
1997, 130, 1175–1187.
11 K) to hydroxymethylene (c), which with a half-life of 2 h at [19] K. Ziegler, E. Holzkamp, H. Breil, H. Martin, Angew. Chem.
11 K rearranges to formaldehyde (c); P. R. Simmonett, W. D.
Allen, E. Matyus, A. G. Csaszar, Nature 2008, 453, 906–909.
1955, 67, 426.
[20] K. Ziegler, Angew. Chem. 1964, 76, 545.
Eur. J. Org. Chem. 2011, 3523–3530
© 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.eurjoc.org
3529

J. J. Eisch, J. N. Gitua, K. Yu
FULL PAPER
[21]
[27] J. J. Eisch, X. Shi, J. Lasota, Z. Naturforsch., Teil B 1995, 50,
342.
[28] J. J. Eisch, J. Organomet. Chem. 2001, 617–618, 148–157.
[29] J. J. Eisch, J. N. Gitua, D. C. Doetschman, Eur. J. Inorg. Chem.
2006, 1968–1975.
[30] J. J. Eisch, P. O. Fregene, D. C. Doetschman, Eur. J. Org. Chem.
2008, 2825–2835.
[31] J. J. Eisch, P. O. Fregene, Eur. J. Org. Chem. 2008, 4482–4492.
[32] Illustrative of such most useful but strictly empirical searches
for transition metal reductants is the study of the reductant
system of Fe(CO)5–HMPA: T. T. Vasil’eva, O. V. Chakhov-
skaya, A. B. Terent’ev, K. A. Kochetkov, Russian J. Org. Chem.
2005, 41, 1553. Moreover, in this publication the authors refer
to previously reported empirical reductants involving V, Ti, Mn
and Ni.
[33] J. J. Eisch, W. Liu, unpublished research.
[34] J. J. Eisch, Organometallic Syntheses, vol. 2, Academic Press,
New York, 1981, p. 3–37.
[35] Website Name: Spectral Database for Organic Compounds.
Website Organized by National Institute of Advanced Indus-
trial Science and Technology (AIST), Japan. Free Website:
http://riodb01.ibase.aist.go.jp/sdbs/cgi-bin/cre_index.cgi
Received: March 3, 2011
J. Borr Jr., Ziegler–Natta Catalysts and Polymerizations, Aca-
demic Press, New York, 1979.
T. Mukaiyama, T. Sato, J. Hanna, Chem. Lett. 1973, 1041.
[22]
[23]
S. Tyrlik, I. Wolochowicz, Bull. Soc. Chim. Fr. 1973, 2147.
[24] J. E. McMurry, M. P. Fleming, J. Am. Chem. Soc. 1974, 96,
4708.
[25] J. E. McMurry, Chem. Rev. 1989, 89, 1513.
[26] a) J. E. McMurry, Acc. Chem. Res. 1983, 16, 405; b) Reductive
couplings on a broad spectrum of carbonyl and azomethine
derivatives have been the subject of many recent investigations,
exemplified by the following: A. Fürstner, A. Hupperts, A.
Ptock, E. Janssen, J. Org. Chem. 1994, 59, 5215; A. Fürstner,
A. Hupperts, J. Am. Chem. Soc. 1995, 117, 4468; A. Fürstner,
G. Seidel, B. Gabor, C. Kopiske, C. Krüger, R. Mynott, Tetra-
hedron 1995, 51, 8875; A. R. Katritzky, J. Li, J. Org. Chem.
1997, 62, 238; J. Gao, M.-Y. Hu, J.-X. Chen, S. Yuan, W.-X.
Chen, Tetrahedron Lett. 1993, 1617; D.-Q. Shi, J.-X. Chen, W.-
Y. Chai, W.-X. Chen, T.-Y. Kao, Tetrahedron Lett. 1993, 2963;
J.-X. Chen, W.-Y. Chai, J.-L. Zhu, J. Gao, W.-X. Chen, T.-Y.
Kao, Synthesis 1993, 87; R. Dams, M. Malinowski, I. Westdop,
H. Geise, J. Org. Chem. 1981, 46, 2407; Y. Okude, S. Hirano,
T. Hiyama, H. Nozaki, J. Am. Chem. Soc. 1977, 99, 3179; A.
Fürstner, N. Shi, J. Am. Chem. Soc. 1996, 118, 12349.
Published Online: May 26, 2011
3530
www.eurjoc.org
© 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Eur. J. Org. Chem. 2011, 3523–3530