Metalation of 2,4,6-trimethylacetophenone using organozinc reagents: The role of the base in determining composition and structure of the developing enolate

Article abstract of DOI:10.1021/om800658s

The new mixed lithium-zinc enolate compounds [(TMEDA)₂Li ₂Zn{OC(=CH₂)Mes}₄] (2) and [{TMP(H) } ₂Li₂Zn{OC(=CH₂)Mes}₄] (3) were prepared by reaction of the sterically demanding ketone 2,4,6- trimethylacetophenone (1) with the all-amido homoleptic zincate [LiZn(TMP) ₃] (TMP = 2,2,6,6-tetramethylpiperidide). X-ray crystallographic studies revealed that these compounds adopt a trinuclear Li... Zn... Li chain arrangement with enolate O bridges. In contrast, the metalation of 1 with heteroleptic [(TMEDA)LiZn(TMP)Me2] afforded the dimeric lithium enolate [(TMEDA)₂Li₂ {OC(=CH2)Mes}₂] (4) as a crystalline solid, which has been characterized in the solid state by X-ray crystallography, and Me₂Zn·TMEDA and TMP(H) as coproducts, showing that the dimethylamido zincate behaves as an amide base. The homoleptic zinc enolate [(TMEDA)Zn {OC(=CH₂)Mes }₂] (5) was obtained by reaction of 1 with the zinc amide Zn(TMP)₂, and its structure was determined by X-ray crystallography. 5 adopts a rarely observed monomeric arrangement where the two enolate groups bind terminally to the zinc. New enolates 2-5 have also been characterized by ¹H, ¹³C, and ⁷Li NMR spectroscopy in C₆D₆ solution. DFT studies of the metalation of 1 by Zn(TMP)₂ and Et₂Zn revealed that the former amide has a much greater kinetic basicity than the latter alkyl reagent.

Full text of DOI:10.1021/om800658s

5860
Organometallics 2008, 27, 5860–5866
Metalation of 2,4,6-Trimethylacetophenone Using Organozinc
Reagents: The Role of the Base in Determining Composition and
Structure of the Developing Enolate
David R. Armstrong, Allison M. Drummond, Liam Balloch, David V. Graham, Eva Hevia,*
and Alan R. Kennedy
WestCHEM, Department of Pure and Applied Chemistry, UniVersity of Strathclyde,
Glasgow, U.K., G1 1XL
ReceiVed July 11, 2008
The new mixed lithium-zinc enolate compounds [(TMEDA)2Li2Zn{OC(dCH2)Mes}4] (2) and
[{TMP(H)}2Li2Zn{OC(dCH2)Mes}4] (3) were prepared by reaction of the sterically demanding ketone
2,4,6-trimethylacetophenone (1) with the all-amido homoleptic zincate [LiZn(TMP)3] (TMP ) 2,2,6,6-
tetramethylpiperidide). X-ray crystallographic studies revealed that these compounds adopt a trinuclear
Li · · · Zn · · · Li chain arrangement with enolate O bridges. In contrast, the metalation of 1 with heteroleptic
[(TMEDA)LiZn(TMP)Me2] afforded the dimeric lithium enolate [(TMEDA)2Li2{OC(dCH2)Mes}2] (4)
as a crystalline solid, which has been characterized in the solid state by X-ray crystallography, and
Me2Zn · TMEDA and TMP(H) as coproducts, showing that the dimethylamido zincate behaves as an
amide base. The homoleptic zinc enolate [(TMEDA)Zn{OC(dCH2)Mes}2] (5) was obtained by reaction
of 1 with the zinc amide Zn(TMP)2, and its structure was determined by X-ray crystallography. 5 adopts
a rarely observed monomeric arrangement where the two enolate groups bind terminally to the zinc.
New enolates 2-5 have also been characterized by 1H, 13C, and 7Li NMR spectroscopy in C6D6 solution.
DFT studies of the metalation of 1 by Zn(TMP)2 and Et2Zn revealed that the former amide has a much
greater kinetic basicity than the latter alkyl reagent.
sterically demanding ketone 2,4,6-trimethylacetophenone (1)
prepared by the reaction of 3 molar equiv with the sodium or
potassium trisamido zincate [MZn(HMDS)3] (M ) Na, K;
HMDS ) 1,1,1,3,3,3-hexamethyldisilazide).7 Zinc enolates can
be prepared generally by three different routes: these are
insertion of activated zinc metal into an R-halogen-substituted
ketone,8 deprotonation of a ketone by an organolithium reagent
such as LDA followed by metathesis with ZnX2,9 and direct
metalation of a ketone with a suitable reactive organozinc
reagent.10 The main drawback of the third approach is the low
kinetic reactivity of most of the common commercially available
organozinc reagents such as the alkyl Me2Zn or the Grignard
equivalent EtZnCl due to the strong carbophilic character of
zinc, which reduces the leaving ability of attached C-nucleo-
philes.11 An alternative to this problem is to employ zinc amide
reagents due to the presence of the more reactive, easier-cleaved
Zn-N bonds in these compounds.12 Thus, although known for
a few years, only very recently Zn(TMP)2 (TMP ) 2,2,6,6-
tetramethylpiperidide) has been reported as a highly efficient
Introduction
Selective deprotonation of ketones to afford metal enolate
intermediates is a cornerstone reaction in synthesis. Anionic
enolates play an essential role in a rich variety of fundamental
organic reactions involving the formation of new C-C bonds
such as Michael additions, acylations, alkylations, or aldol
additions.1 Among this family of compounds zinc enolates are
attracting special interest, as they combine a poweful nucleo-
philicity with a greater functional group tolerance than analogous
alkali-metal enolates.2 Thus, zinc enolates can participate in
additions to carbonyl compounds,3 transition-metal-catalyzed
reactions with carbon electrophiles,4 or catalyzed acrylate
polymerizations.5 However, despite their many synthetic ap-
plications only a select few zinc enolate structures have been
elucidated. The first structural elucidation of a series of zinc
enolates derived from simple ketones was reported by Hagadorn
as recently as 2006.6 The following year we reported the first
examples of mixed alkali-metal-zinc enolates derived from the
* Corresponding author. E-mail: eva.hevia@strath.ac.uk.
(1) Williard, P. G. In ComprehensiVe Organic Synthesis; Trost, B. M.,
Flemming, I., Ed.; Pergamon: Oxford, 1990; Vol. 1, Chapter 1.
(2) Nakamura, E. In Organometallics in Synthesis. A Manual, 2nd ed.;
Schlosser, M., Ed.; Wiley: Chichester, 2002, Chapter 5.
(3) Hansen, M. H.; Bartlett, P. A.; Heathcock, C. H. Organometallics
1987, 6, 2069.
(4) (a) Nakamura, E.; Kuwajima, I. Tetrahedron Lett. 1986, 27, 83. (b)
Nakamura, E.; Sekiya, K.; Kuwajima, I. Tetrahedron Lett. 1987, 28, 337.
(c) Tamaru, Y.; Ochiai, H.; Nakamura, T.; Yoshida, Z. Angew. Chem., Int.
Ed. Engl. 1987, 26, 1157. (d) Sekiya, K.; Nakamura, E. Tetrahedron Lett.
1988, 29, 5155. (e) Kinoshita, N.; Kawabata, T.; Tsubaki, K.; Bando, M.;
Fuji, K. Tetrahedron 2006, 62, 1756.
(7) Baillie, S. E.; Hevia, E.; Kennedy, A. R.; Mulvey, R. E. Organo-
metallics 2007, 26, 204.
(8) (a) Dekker, J.; Budzelaar, P. H. M.; Boersma, J.; van der Kerk,
G. L. M.; Spek, A. L. Organometallics 1984, 3, 1403. (b) Hama, T.; Culkin,
D. A.; Hartwig, J. F. J. Am. Chem. Soc. 2006, 128, 4976.
(9) (a) van der Steen, F. H.; Jastrzebski, J. T. B. H.; van Koten, G.
Tetrahedron Lett. 1986, 27, 83. (b) van der Steen, F. H.; Kleijn, H.; Spek,
A. L.; van Koten, G. Chem. Commun. 1990, 503.
(10) (a) Hlavinka, M. L.; Hagadorn, J. R. Organometallics 2005, 24,
4116. (b) Hlavinka, M. L.; Hagadorn, J. R. Tetrahedron Lett. 2006, 47,
5049.
(11) Armstrong, D. R.; Clegg, W.; Dale, S. H.; Garcia-Alvarez, J.;
Harrington, R. W.; Honeyman, G. H.; Kennedy, A. R.; Mulvey, R. E. Chem.
Commun. 2008, 187.
(12) Hlavinka, M. L.; Greco, J. F.; Hagadorn, J. R. Chem. Commun.
2005, 5304.
(5) Garner, L. E.; Zhu, H.; Hlavinka, M. L.; Hagadorn, J. R.; Chen,
E. Y.-X. J. Am. Chem. Soc. 2006, 128, 14822.
(6) Hlavinka, M. L.; Hagadorn, J. R. Organometallics 2006, 25, 3501.
10.1021/om800658s CCC: $40.75
2008 American Chemical Society
Publication on Web 10/15/2008

Metalation of 2,4,6-Trimethylacetophenone
Organometallics, Vol. 27, No. 22, 2008 5861
Scheme 1
of 2, whose stoichiometry is different from that in the reaction
mixture, could be explained by a disproportionation process
whereby an intermediate similar to the Na/Zn or K/Zn enolates
is initially formed but subsequently leads to 2 via elimination
of Zn(enolate)2 (eq 1). An equilibrium between the three enolate
species could be ocurring in solution, which could be favoring
the precipitation (crystallization) of 2 due to its lower solubility
in hexane solution.
(TMEDA)2Li2Zn2(enolate)6 h
(TMEDA)2Li2Zn(enolate)4 + Zn(enolate)2
(1)
Recently Mongin has demonstrated that the metalation activity
of [LiZn(TMP)3] is dramatically dependent on the Lewis basicity
of the solvent system employed.15a Therefore, we next carried out
the reaction using neat hexane as a solvent in the absence of
TMEDA. Colorless crystals of the mixed-metal enolate 3 (Scheme
1) were obtained where the lithium centers are now solvated by
the secondary amine TMP(H), the coproduct of this metalation
reaction. These results show not only that the metalation of 1 is
not TMEDA-activated but also that the suggested disproportion-
ation process in eq 1 is not favored nor induced by the presence
of the diamine, though a donor solvent (even a poor sterically
restricted one such as TMP(H) in 3) is needed to support the lithium
coordination.
reagent to metalate a wide range of substrates, including ketones,
under very mild conditions.13 Alkali-metal zincates containing
bulky amido ligands have also revealed themselves as highly
selective reagents for deprotonation reactions.14 Herein we
explore the reactivity of the homoleptic “inorganic” zincate
[LiZn(TMP)3]15 with 1 and report the synthesis and character-
ization of the dilithium-monozinc tetraanionic enolates
[(TMEDA)2Li2Zn-
{OC(dCH2)Mes}4] (2) and [{TMP(H)}2Li2Zn{OC(dCH2)-
Mes}4] (3), which are to the best of our knowledge the first
reported examples of lithium-zinc enolate compounds. In
addition the structure of the first simple monomeric zinc enolate
[(TMEDA)Zn{OC(dCH2)Mes}2] (5), obtained by reaction of
Zn(TMP)2 with 1, has been elucidated. The reaction of the
heteroleptic zincate [(TMEDA)LiZn(TMP)Me2] toward 1 was
also studied, affording the homometallic lithium enolate
[(TMEDA)2Li2{OC(dCH2)Mes}2] (4). A comparative theoreti-
cal study of the energies involved in the reactions of 1 with
Zn(TMP)2 and ZnEt2 was carried out, from which it can be
surmised that the formation of these new zinc enolate com-
pounds is dictated by the kinetics of the reaction.
For comparative purposes the reaction of 1 with the hetero-
leptic zincate [(TMEDA)LiZn(TMP)Me2]17 was subsequently
performed. This mixed-metal reagent belongs to the family of
dialkyl(TMP) zincates, which have been found to be excellent
chemoselective bases for direct zincation of a broad range of
aromatic substrates.14 So far, all of the metalated intermediates
isolated and structurally characterized from such reactions have
shown that these zincates behave overall as monoalkyl bases.18
The reaction of 1 with the dimethylamido zincate afforded an
insoluble white solid that dissolved in toluene on the addition
of an extra molar equivalent of TMEDA. On cooling, this
solution deposited colorless crystals of the zinc-free lithium
enolate 4. In order to gain more understanding of this reaction,
isolated crystals of the starting material [(TMEDA)LiZn(T-
MP)Me2] were dissolved in deuterated benzene solution and a
stoichiometric amount of 1 was added. A white solid precipi-
tated. The 1H NMR analysis of the filtrate revealed that
Me2Zn · TMEDA and TMP(H) were the other products of the
reaction. This allowed us to suggest the reaction pathway shown
in Scheme 2, where the zincate behaves as an amido base to
form the putative intermediate [(TMEDA)LiZn(OR)Me2] (A),
which then disproportionates into the insoluble unsolvated
lithium enolate and the zinc alkyl complex Me2Zn · TMEDA.
This reaction was also attempted with 3 molar equiv of 1;
however, the lithium enolate 4 was the only metalated product
obtained, and unreacted ketone and Me2Zn · TMEDA could be
Results and Discussion
Synthesis. New mixed metal enolate 2 was prepared accord-
ing to the reaction sequence shown in Scheme 1. First the
trisamido zincate [LiZn(TMP)3] was prepared in situ by reaction
of 3 molar equiv of the lithium amide with ZnCl2 in ether
solution.15 After removal of deposited LiCl by filtration, the
addition of 3 equivalents of 1 and 1 equiv of TMEDA produced
compound 2 as colorless crystals in reasonably good isolated
yields (35-40%: note that the maximum possible yield from
this reaction is 50%; see below). The [Li2Zn(enolate)4] stoichi-
ometry of 2 contrasts with previous results found for the reaction
of related tris(amido) zincates [MZn(HMDS)3]16 (M ) Na, K)
with 1, which afforded the relevant inverse crown compounds
[M2Zn2(enolate)6],7 having the same stoichiometry (1M:1Zn:3
enolate) as that employed in the reaction mixture. The formation
1
observed in the H NMR spectrum of the filtrate. Thus, these
results permit a comparison to be made between the reactivity
of the all-amido zincate [LiZn(TMP)3] and the dialkyl(amido)
zincate [(TMEDA)LiZn(TMP)Me2], showing that for the former
the three basic “arms” (or ligands) can react with 1, whereas
for the latter only one of the “arms” (the amide ligand) is
(13) Hlavinka, M. L.; Hagadorn, J. R. Organometallics 2007, 26, 4105.
(14) For comprehensive reviews of alkali-metal zincates and their
applications in metalation reactions see: (a) Mulvey, R. E. Organometallics
2006, 25, 1060. (b) Mulvey, R. E.; Mongin, F.; Uchiyama, M.; Kondo, Y.
Angew. Chem., Int. Ed. 2007, 46, 3802.
(15) (a) Seggio, A.; Lannou, M. I.; Chevallier, F.; Nobuto, D.; Uchiyama,
M.; Golhen, S.; Roisnel, T.; Mongin, F. Chem.-Eur. J. 2007, 13, 9982.
(b) Seggio, A.; Chevallier, F.; Vaultier, M.; Mongin, F. J. Org. Chem. 2007,
72, 6602. (c) L’Helgoual’ch, J.-M.; Seggio, A.; Chevallier, F.; Yonehara,
M.; Jeanneau, E.; Uchiyama, M.; Mongin, F. J. Org. Chem. 2008, 73, 177.
(16) The same synthetic methodology could not be employed to prepare
a lithium analogue since Li(HMDS) and Zn(HMDS)2 fail to form a mixed-
metal base.
(17) Graham, D. V.; Hevia, E.; Kennedy, A. R.; Mulvey, R. E.
Organometallics 2006, 25, 3297.
(18) For two exceptions where [LiZn(TMP)tBu2] behaves as a dual alkyl-
amido base see: (a) Clegg, W.; Dale, S. H.; Harrington, R. W.; Hevia, E.;
Honeyman, G. W.; Mulvey, R. E. Angew. Chem., Int. Ed. 2006, 43, 2374.
(b) Kondo, Y.; Morey, J. V.; Morgan, J. C.; Naka, H.; Nobuto, D.; Raithby,
P. R.; Uchiyama, M.; Wheatley, A. E. H. J. Am. Chem. Soc. 2007, 129,
12734.

5862 Organometallics, Vol. 27, No. 22, 2008
Armstrong et al.
Scheme 2
reactive enough to metalate 1 and the methyl groups remain
coordinated to zinc and act only as spectators. This difference
in reactivity illustrates the concept of the lower kinetic reactivity
of the Zn-C bonds in comparison to Zn-N bonds. In addition,
this is the first example of a metalation reaction using a lithium
TMP-dialkyl zincate where the metalated intermediates have
been characterized and show that the mixed-metal reagent has
behaved exclusively as an amido base with the concomitant loss
of TMP(H).19 The completely different nature of the metalated
products obtained from these reactions should also be mentioned.
Whereas 2 is a mixed lithium-zinc enolate, where each enolate
anion is bridged between the two distinct metals (see below),
in 4 the enolate ligands are bonded solely to lithium. This could
have a pronounced effect when more complicated ketones
containing functionalization are employed since a lithium enolate
such as 4 will possess a very limited functional group tolerance
in comparison with a mixed lithium-zinc enolate such as 2 or
3. Moreover, the nature of the enolate nucleophile, whether it
be lithium- or zinc-bonded, can have a dramatic influence if
the metalated intermediate is to be involved in subsequent
organic transformations such as transition metal cross-coupled
reactions.
are connected through enolate ligands. Compound 3 adopts the
same structural motif with the lithium centers solvated by a
molecule of the secondary amine TMP(H)20 instead of being
chelated by TMEDA ligands as in 2. Unfortunately due to a
large amount of motion within the TMP(H) ligands, the structure
is disordered, which prevents further discussion of the bond
lengths and bond angles. The molecule in 2 is centrosymmetric,
containing two orthogonal fused {LiOZnO} four-membered
rings. Both rings are planar, as evidenced by the sum of the
internal angles (360.00°), and the arrangement of the three
metals is virtually linear, as shown by the Li · · · Zn · · · Li angle
[179.99(2)°]. Each metal center in 2 is tetracoordinated in a
distorted tetrahedral geometry. Each lithium center is bound to
two oxygens and two nitrogens from the enolate and TMEDA
ligands, respectively, whereas zinc is bound to four oxygens
from the enolate ligands. The Li-O bond distance [1.954(2)
Å] is within the average range found for other lithium enolates
such as the dimer [(TMEDA)2Li2{OC(dCHCH3)OtBu}2] (aver-
age Li-O length 1.925 Å).21 The Zn-O bond length [1.9542(9)
Å] is similar to those found in the homometallic ketone enolates
[Br2Zn2{O(dCH2)Mes}2(dmf)2] [1.954 Å]22 and [MeLZn2{OC-
(dCMe2)iPr}2] (MeL2- is a bis(amidoamine)ligand) [1.979 Å].6
The linear metal motif of 2 contrasts with the inverse crown
cyclic structures found for the related mixed-metal trianionic
zincates [Na2Zn2{OC(dCH2)Mes}6{OC(CH3)Mes}2] (6) and
[K2Zn2{OC(dCH2)Mes}6(CH3Ph)2] (7)7 (Figure 3).
As aforementioned, Hagadorn has recently reported the
application of Zn(TMP)2 as an effective metalating reagent under
mild conditions.13 Thus, following this methodology homome-
tallic enolate 5 could be prepared by reaction of 2 molar equiv
of 1 with Zn(TMP)2 in the presence of TMEDA at room
temperature (Scheme 2). Here, TMEDA seems to merely aid
crystallization of 5 since the neutral zinc amide can metalate 1
in an almost quantitative yield in the absence of the diamine,
as shown when the reaction was carried out in C6D6 solution
1
and monitored by H NMR.
Solid-State Structures. The molecular structures of com-
pounds 2-5 were successfully characterized by X-ray diffraction
studies. Figures 1 and 2 show that of mixed-metal enolates 2
and 3, respectively.
The structure of 2 can be described as a ion-contacted
aggregate of two oppositely disposed [Li(TMEDA)]+ cations
coordinated to the tetrahedral anion Zn(OR)42- giving rise to a
trimetallic Li · · · Zn · · · Li chain arrangement in which the metals
(19) The preferred kinetic amido basicity toward alkyl basicity for TMP-
dialkyl zincates has been the subject of several theoretical studies; see for
example: (a) Uchiyama, M., M.; Matsumoto, Y.; Nobuto, D.; Furuyama,
T.; Yamaguchi, K.; Morokuma, K. J. Am. Chem. Soc. 2006, 128, 8748. (b)
Uchiyama, M.; Matsumoto, Y.; Usui, S.; Hashimoto, Y.; Morokuma, K.
Angew. Chem., Int. Ed. 2007, 46, 926. (c) Garcia, F.; McPartlin, M.; Morey,
J. V.; Nobuto, D.; Kondo, Y.; Naka, H.; Uchiyama, M.; Wheatley, A. E. H.
Eur. J. Org. Chem. 2008, 644.
Figure 1. Molecular structure of 2 showing selected atom labeling
and with hydrogen atoms omitted for clarity.

Metalation of 2,4,6-Trimethylacetophenone
Organometallics, Vol. 27, No. 22, 2008 5863
Figure 4. Molecular structure of 4 showing selected atom labeling
and with hydrogen atoms omitted for clarity.
Figure 2. Molecular structure of 3 showing selected atom labeling
and with hydrogen atoms and disorder component omitted for
clarity.
Figure 3. ChemDraw representations of [K2Zn2{OC(dCH2)-
Mes}6(CH3Ph)2] (7) and [(TMEDA)2Na2Mg{OC(dCH2)Mes}4] (8).
Figure 5. Molecular structure of 5 showing selected atom labeling
and with hydrogen atoms omitted for clarity.
To the best of our knowledge 2 represents the first example
of a mixed-metal lithium zinc enolate and the first tetraanionic
zincate characterized where the anionic ligands around zinc are
enolates. The closest precedent to 2, apart from the Na-Zn and
K-Zn enolates aforementioned, is the Na-Mg enolate
[(TMEDA)2Na2Mg{OC(dCH2)Mes}4] (8),23 which also adopts
an approximately linear Na · · · Mg · · · Na arrangement [Na · · ·
Mg · · · Na angle 167.70(5)°], where the metals are connected
by four enolate bridges.
solvated enolate23 derived from 1, which is structurally similar
to 4. In this compound the sodium center forms additional
interactions due to π-engaging with the olefinic carbons of the
enolate ligands. These types of interactions are absent in 4 (the
closest Li · · · C contact is 2.906(4) Å, which is too long to
suggest any significant π-interaction) probably due to the harder
character, smaller size, and thus easier satisfied coordination
of lithium in comparison to that of sodium.24
Lithium enolate 4 exhibits a dimeric centrosymmetric struc-
ture forming a {LiOLiO} four-membered ring where each
lithium is externally solvated by a molecule of the chelating
diamine TMEDA (Figure 4). The four-membered ring adopts
an almost planar disposition (sum of the internal angles is
357.38°). Having a distorted tetrahedral geometry, the lithium
coordination is made up of two nitrogens from the TMEDA
ligand and two oxygens from the enolate ligands. This structural
motif has been previously reported by Seebach for other lithium
ester enolates solvated by TMEDA.21 The Li-O distance in 4
[1.879(2) Å] is modestly shorter than that found in the mixed-
metal lithium-zinc enolate 2 [1.954(2) Å]. Previously we have
reported the structure of the congeneric sodium-TMEDA-
Zinc enolate 5 adopts a simple monomeric structure where
zinc occupies a distorted tetrahedral geometry bonded to two
nitrogens and two oxygens from the TMEDA and the enolate
ligands, respectively (Figure 5). Surprisingly,to the best of our
knowledge this is the first example of a homoleptic zinc enolate
that has been characterized in the solid state. As previously
mentioned, the first examples of zinc enolates derived from
simple ketones isolated and structurally characterized were
reported by Hagadorn.6 These compounds exhibit a dimeric
arrangement where each zinc binds to two bridging enolate
ligands and a sterically hindered bidentate bis(amidoamine)
ligand. The same group has recently reported a new series of
ketone enolates of formula [Br2Zn2{OC(dCR2)R′}2(dmf)2]
(where dmf is dimethylformamide), which also adopt dimeric
structures.22 Thus 5 is also the first simple monomeric zinc
enolate to be structurally characterized. As expected, the Zn-O
bond distances in 5 [1.8952(16) Å] are shorter than those found
in the aforementioned dimeric compounds (Zn-O bond lengths
(20) A search of the CCDB showed there are only 19 examples of
compounds where TMP(H) acts as a donor ligand, and of those only two
contain lithium: (a) Garcia-Alvarez, J.; Hevia, E.; Kennedy, A. R.; Klett,
J.; Mulvey, R. E. Chem. Commun. 2007, 23, 2402. (b) Braun, U.; Habereder,
H.; Noth, H. Eur. J. Inorg. Chem. 2004, 18, 3629.
(21) Seebach, D.; Amstutz, R.; Laube, T.; Schweizer, W. B.; Dunitz,
J. D. J. Am. Chem. Soc. 1985, 107, 5403.
(22) Greco, J. F.; McNevin, M. J.; Shoemaker, R. K.; Hagadorn, J. R.
Organometallics 2008, 27, 1948.
(23) Hevia, E.; Henderson, K. W.; Kennedy, A. R.; Mulvey, R. E.
Organometallics 2006, 25, 1778.
(24) For examples of metal-π-arene interactions see: (a) Gokel, G. W.;
De Wall, S. L.; Meadows, E. S. Eur. J. Org. Chem. 2000, 2967. (b) Forbes,
G C.; Kennedy, A. R.; Mulvey, R. E.; Roberts, B. A.; Rowlings, R. B.
Organometallics 2002, 21, 5115.

5864 Organometallics, Vol. 27, No. 22, 2008
Armstrong et al.
Table 1. Key Bond Lengths (Å) and Bond Angles (deg) within the
Structures of 2, 4, and 5
than those found for the mixed-metal Li-Zn enolates (5.07,
4.10 ppm for 2 and 5.00, 4.18 ppm for 3). In addition, the
relative positioning of the TMEDA signals in the H NMR
1
For 2
spectrum25 shows that in compounds 2 and 5 the chelating
diamine remains coordinated to the metal (lithium in 2 and zinc
in 5), suggesting that the structures found in the solid state for
these compounds should be maintained in benzene solution. In
contrast when the lithium enolate 4 is dissolved in C6D6 solution,
the TMEDA ligands appear as free, which suggests a different
structure in solution than in the solid state, where the lithium
centers are probably interacting with the arene solvent instead
of the chelating diamine. For compound 3, the chemical shifts
of the corresponding TMP(H) resonances (1.40, 1.10, 0.85, and
0.15 ppm) are slightly further upfield to those found for the
free amine (1.53, 1.22, 1.06, and 0.30 ppm) in the same
deuterated solvent, indicating that TMP(H) remains coordinated
to the lithium atoms.
Li1*-O1
1.954(2)
1.9546(9)
2.145(3)
85.52(13)
89.72(14)
89.70(5)
O1-C2
1.3443(17)
1.331(2)
1.497(2)
Zn1-O1
C1-C2
Li1-N1
C2-C3
N1-Li1-N1*
O1***-Li1-O1**
O1*-Zn1-O1
O1*-Zn1-O1**
Zn1-O1-Li1
C5-C4-C3
C5-C4-C9
C9-C4-C3
90.29(7)
119.14(18)
119.59(18)
121.25(14)
118.22(5)
For 4
Li1-O1
Li1-O1*
Li1-N2
Li1-N1
O1*-Li1-O1
N1-Li1-N2
1.879(4)
1.923(4)
2.146(5)
2.236(5)
93.74(19)
82.70(16)
O1-C8
C7-C8
C8-C9
1.325(3)
1.338(4)
1.508(4)
O1*-Na1-N1 107.58(13)
Li1-O1-Li1*
C8-O1-Li1
84.95(19)
148.75(18)
For 5
Zn1-O1
1.8952(16) C1-C2
1.9019(15) C2-C3
2.1036(19) O2-C13
2.0902(14) C12-C13
1.338(3)
1.498(3)
1.336(2)
1.324(4)
1.504(3)
108.39(8)
114.59(7)
101.81(7)
Zn1-O2
As recently reported, the difference between the chemical
shifts in the 13C NMR spectrum of the C-O and dCH2
resonances of the enolato ligand can be considered a good
indicator of the degree of polarization of the CdC bonds and
the nucleophilicity of the enolate ligand.6 Table 3 summarizes
the values found for the new enolates 2-5 and compares them
with those previously reported for the related enolates derived
from 1, namely, homometallic [(TMEDA)2Na2{O(dCH2)-
Mes}2]23 and [(toluene)4K4{O(dCH2)Mes}4]26 and the hetero-
metallic Na-Mg enolate [(TMEDA)2Na2Mg{OC(dCH2)-
Mes}4]23 in the same deuterated solvent. The ∆δ values for the
Li-Zn enolates 2 and 3 are remarkably lower (by an average
of 9.10 ppm) than that found in the homometallic lithium enolate
4. This can be rationalized in terms of the reduced polarization
of the M-O bonds in 2 and 3 (as previously discussed).
Following the same trend, a comparison between the ∆δ values
of the homometallic lithium (4), sodium, and potassium enolates
shows that the more electropositive the metal, the more polarized
are the M-O bonds. The mixed-metal Na-Mg enolate, isos-
tructural with 2 as previously mentioned, possesses a higher
value of ∆δ probably due to the combined effect of magnesium
(with a less covalent character than zinc) and sodium (more
electropositive than lithium). Surprisingly zinc enolate 5 has a
∆δ value significantly greater than those found for Li-Zn
enolates 2 and 3 and closer to the lithium enolate 4. This is
consistent with the monomeric structure of 5, where the terminal
enolate ligands form more polarized (and therefore reactive)
Zn-O bonds, in comparison with 2 and 3, where all the enolate
ligands are acting as bridges.
Zn1-N1
Zn1-N2
O1-C2
1.327(3)
129.72(7)
87.64(8)
C13-C14
O1-Zn1-O2
N2-Zn1-N1
O1-Zn1-N2
O2-Zn1-N2
O1-Zn1-N1
O2-Zn1-N1
106.66(10)
Table 2. 1H NMR Chemical Shifts (δ in ppm) for the Enolate
Ligands in Compounds 2-5 in C6D6 Solution
compound
δ(m-H)
δ(dCHH′)
δ(o-CH3)
δ(p-CH3)
2
3
4
5
6.79
6.77
6.84
6.94
5.07, 4.10
5.00, 4.18
4.10, 3.83
4.61, 4.16
2.77
2.76
2.72
2.79
2.21
2.18
2.21
2.27
Table 3. Selected 13C NMR Chemical Shifts (δ in ppm) for the
Enolate Ligands in C6D6 Solution
compound
δ(C-O)
δ(dCH2)
∆δ
2
3
4
5
163.76
162.75
167.53
166.71
165.17
169.94
171.23
86.10
85.34
80.86
83.14
83.91
79.60
74.58
77.66
77.41
86.67
83.57
81.26
90.34
96.65
[(TMEDA)2Na2Mg{O(dCH2)Mes}4]
[(TMEDA)2Na2{O(dCH2)Mes}2]
[(toluene)4K4{O(dCH2)Mes}4]
range from 1.952(5) to 2.047(2) Å) or in the mixed-metal
lithium-zinc enolate 2 [1.9542(9) Å].
Solution-State Studies. The new enolate compounds 2-5
show good solubilities in arene solvents, and therefore their 1H,
7Li, and 13C NMR spectra were recorded from C6D6 solutions.
Tables 2 and 3 compare the most distinct chemical shifts in the
1H and 13C NMR spectra respectively of compounds 2-5. As
observed from Table 2 the chemical shifts of the relevant signals
for the ortho and para methyl groups in compounds 2-5 show
little change in the 1H NMR spectrum, while the aromatic and
olefinic protons show significant variations. Thus, the aromatic
protons in the mixed-metal Li-Zn enolates 2 and 3 (6.79 and
6.77 ppm, respectively) appear further upfield than those
corresponding to the monometallic lithium enolate 4 (6.84 ppm).
This can be attributed largely to the more electropositive nature
of lithium, which increases the polarization of the Li-O bonds.
Therefore the metal-oxygen bond will be more polarized in 4,
where each enolate is bound to two lithium atoms, whereas in
2 and 3 the ligands bind to lithium and the more covalent zinc.
Surprisingly the most deshielded aromatic protons are those
corresponding to the homometallic zinc enolate 5 (6.94 ppm)
probably due to the fact that the two enolate ligands are terminal,
bound solely to a zinc. For the olefinic protons, the resonances
of the lithium enolate 4 are further upfield (4.10 and 3.83 ppm)
Theoretical Studies. As alluded to previously, dialkyl zinc
reagents are rarely used in deprotonative metalations due to their
low kinetic basicity as a consequence of the high covalency of
the Zn-C bonds. An efficient synthetic alternative is the use
of zinc amides, as Zn-N bonds are presumed to be easier to
break. In order to shed some light on the kinetic/thermodynamic
reactivity of zinc reagents, we carried out a theoretical study
using DFT calculations at the B3LYP/6-311G** level. The
reactions of Et2Zn and Zn(TMP)2 with 2 equiv of 1 to afford
the enolate [Zn{O(dCH2)Mes}2] were modeled in a two-step
mechanism, where initially one molecule of the ketone coor-
dinates to the organozinc reagent (CP1) followed by deproto-
nation to afford the alkyl-enolate or amide-enolato intermediate
(25) Andrews, P. C.; Barnett, N. D. R.; Mulvey, R. E.; Clegg, W.;
O’Neil, P. A. O.; Barr, D.; Cowton, L.; Dawson, A. J.; Wakefield, B. J. J.
Organomet. Chem. 1996, 518, 85.
(26) He, X.; Noll, B. C.; Beatty, A.; Mulvey, R. E.; Henderson, K. W.
J. Am. Chem. Soc. 2004, 126, 7444.

Metalation of 2,4,6-Trimethylacetophenone
Organometallics, Vol. 27, No. 22, 2008 5865
Figure 6. Energetics of the modeled reactions of 2 equiv of ketone 1 with ZnEt2 (blue) or Zn(TMP)2 (green). Energy changes are shown
in kcal/mol.
[RZn(OR)] via a transition state (TS1). The second equivalent
of the ketone then coordinates to this intermediate (CP2), and
the second metalation takes place to yield the desired zinc
enolate Zn(OR)2 and the relevant coproduct (ethane or the amine
TMPH) via a second transition state, TS2 (Figure 6). As
observed, from a strictly thermodynamic point of view Et2Zn
should be a better base since the overall reaction to form the
bis(enolato) product is exothermic by -6.01 kcal/mol. However
this reaction sequence involves two distinct transition states (TS1
and TS2) with extremely high activation barriers (36.99 and
42.27 kcal/mol, respectively). On the other hand the reaction
of 1 with the zinc amide Zn(TMP)2 is overall endothermic
(+9.89 kcal/mol); however the activation energies of the
transition states TS1 and TS2 are considerably lower (19.83
and 13.52 kcal/mol). In addition the reaction of Zn(OR)2 with
TMEDA to afford enolate 5 is exothermic by 39.14 kcal, which
makes the reaction of Zn(TMP)2 with 1 favored not only
kinetically but also thermodynamically (-29.25 kcal/mol). Thus
these calculations illustrate the greater kinetic basicity of
Zn(TMP)2 in comparison with ZnEt2, which makes the former
an efficient and versatile base for accomplishing ketone
metalations.
metalate 1 to yield the simple monomeric zinc enolate
[(TMEDA)Zn{OC(dCH2)Mes}2] (5). A theoretical study of the
reactions of 1 with Zn(TMP)2 and ZnEt2 quantifies the greater
kinetic reactivity of the Zn-N bonds in comparison with Zn-C
bonds.
Experimental Section
General Procedures. All reactions were performed under a
protective argon atmosphere using standard Schlenk techniques.
Hexane and toluene were dried by heating to reflux over sodium
benzophenone ketyl and distilled under nitrogen prior to use. nBuLi
and Me2Zn, were purchased from Aldrich Chemicals and used as
27
received. Zn(TMP)2 and [(TMEDA)LiZn(TMP)Me2]17 were
prepared following literature procedures. NMR spectra were
recorded on a Bruker DPX 400 MHz spectrometer, operating at
400.13 MHz for 1H, 150.32 MHz for 7Li, and 100.62 MHz for
13C.
X-ray Crystallography. Single-crystal diffraction data were
measured with a Nonius Kappa CCD diffractometer (compounds
2 and 5) or an Oxford Diffraction Xcalibur S instrument (samples
3 and 4) using graphite-monochromated Mo KR radiation (λ )
0.71073 Å). The structures were refined by full-matrix least-squares
and against F2 to convergence using the SHELXL-97 program.28
Specific crystallographic data and refinement parameters are given
in Table 4.
Conclusions
The metalation of the highly sterically demanding ketone 1
has been investigated with three different organozinc reagents.
When the homoleptic inorganic zincate [LiZn(TMP)3] is em-
ployed, the mixed-metal lithium-zinc enolates [(L)2Li2Zn{O(d
CH2)Mes}4] (L) TMEDA, 2; TMPH, 3) were obtained where
the three amido groups of the base react with the ketone. In
contrast to this ligand-efficient process, when the heteroleptic
alkyl-amido zincate [(TMEDA)LiZn(TMP)Me2] reacts with 1,
only the amido group is basic enough to metalate 1, affording
the lithium enolate [(TMEDA)2Li2{O(dCH2)Mes}2] (4) and
releasing Me2Zn · TMEDA and TMP(H) as coproducts. Enolates
2-4 exhibit similar structural motifs, where the metals are
connected through enolate bridges. Zn(TMP)2 can successfully
Synthesis of [(TMEDA)2Li2Zn{OC(dCH2)Mes}4] (2). To a
solution of LiTMP [6 mmol, prepared in situ by reaction of BuLi
(3.75 mL of a commercial 1.6 M solution in hexane, 6 mmol) with
TMPH (1.01 mL, 6 mmol)] in hexane cooled at 0 °C was added
ZnCl2 (2 mL of a commercial 1 M solution in ether, 2 mmol). A
white precipitate (LiCl) was formed instantaneously. The reaction
mixture was allowed to stir at room temperature overnight and
filtered through Celite, and TMEDA was added (0.3 mL, 2 mmol),
(27) Rees, W. S.; Just, O.; Schumann, H.; Weimann, R. Polyhedron
1998, 17, 1001.
(28) Sheldrick, G. M. SHELXL-97, program for crystal structure
refinement; University of Go¨ttingen: Germany, 1997.

5866 Organometallics, Vol. 27, No. 22, 2008
Armstrong et al.
Table 4. Key Crystallographic and Refinement Parameters for Compounds 2-5
2
3
4
5
empirical formula
mol wt
cryst syst
space group
C56H84Li2N4O4Zn
956.52
orthorhombic
Fddd
C62H90Li2N2O4Zn
1006.61
orthorhombic
Fddd
C34H58Li2N4O2
568.72
orthorhombic
Pnn2
C28H42N2O2Zn
504.01
monoclinic
P21
a/Å
b/Å
c/Å
17.8650(5)
24.3175(7)
21.4188(7)
90
11911.6(6)
8
20.948(5)
20.964(5)
27.849(8)
90
12230(6)
8
16.4657(8)
12.1496(6)
8.7833(5)
90
1757.11(16)
2
9.4749(3)
12.6708(3)
11.6753(3)
98.4680(10)
1386.39(7)
2
ꢀ/deg
V/Å3
Z
T/K
123(2)
173(2)
150(2)
123(2)
indep reflns
goodness-of-fit on F2
R1 [I > 2σ(I)]
wR2
3409
1.061
0.0342
0.0804
3912
0.982
0.0454
0.1186
2280
0.865
0.0395
0.0840
5511
1.026
0.0283
0.0650
affording a pale yellow solution. Then 1.00 mL of 1 (6 mmol) was
introduced, and the mixture was allowed to stir for 30 min, affording
a white solid. Toluene (5 mL) was added at this stage, and the
mixture was gently heated until all the white solid had dissolved,
affording a pale orange solution. Allowing this solution to cool
slowly to room temperature produced a crop of colorless crystals
6.84 (s, 2H, m-H Mes), 4.10 (s, br, 1H, CHH′), 3.83 (s, br, 1H,
CHH′), 2.72 (s, 6H, o-CH3, Mes), 2.26 (s, CH2, TMEDA), 2.21 (s,
3H, p-CH3, Mes), 2.09(s, CH3, TMEDA). 13C{1H} NMR (100.63
MHz, 25 °C, d6-benzene): 167.53 (OCdCH2), 142.38 (Cipso, Mes),
135.19 (Cortho, Mes), 128.70 (Cpara, Mes), 127.64 (Cmeta, Mes), 80.86
(OCdCH2), 57.96 (CH2, TMEDA), 45.79 (CH3, TMEDA), 20.87
(p-CH3, Mes), 20.68 (o-CH3, Mes). 7Li NMR (155.50 MHz, C6D6
reference LiCl in D2O at 0.00 pppm): 0.57.
Synthesis of [(TMEDA)Zn{OC(dCH2)Mes}2] (5). 1 (0.67 mL,
4 mmol) was added to a solution of freshly prepared Zn(TMP)2
(0.62 g, 2 mmol) in hexane (10 mL). The solution changed from
colorless to pale yellow. TMEDA (0.3 mL, 2 mmol) was then
introduced, affording a white suspension. Toluene was introduced
at this stage (2 mL), and the mixture was gently heated until all
the solid had dissolved, affording a pale orange solution. Allowing
this solution to cool slowly at room temperature afforded a crop of
colorless crystals (0.35 g, 42% isolated crystalline yield, NMR
analysis of the filtrate showed the reaction is almost quantitative).
1H NMR (400 MHz, 25 °C, d6-benzene): 6.94 (s, 4H, m-H Mes),
4.61 (s br, 2H, CHH′), 4.16 (s br, 2H, CHH′), 2.79 (s, 12H, o-CH3,
Mes), 2.27 (s, 6H, p-CH3, Mes), 2.03 (s, 12H, CH3, TMEDA), 1.51
(s, 4H, CH2, TMEDA). 13C{1H} NMR (100.63 MHz, 25 °C, d6-
1
(0.68 g, 36%). H NMR (400 MHz, 25 °C, d6-benzene): 6.79 (s,
8H, m-H Mes), 5.07 (s br, 4H, CHH′), 4.10 (s br, 4H, CHH′), 2.77
(s, 24H, o-CH3, Mes), 2.21 (s, 12H, p-CH3, Mes), 1.50 (s, 8H, CH2
TMEDA). 13C{1H} NMR (100.63 MHz, 25 °C, d6-benzene): 163.76
(OCdCH2), 142.49 (Cipso, Mes), 136.35 (Cortho, Mes), 135.31 (Cpara
,
Mes), 127.88 (Cmeta, Mes), 86.10 (OCdCH2), 56.27 (CH2 TMEDA),
45.20 (CH3 TMEDA), 20.84 (p-CH3, Mes). 20.16 (o-CH3, Mes).
7Li NMR (155.50 MHz, C6D6 reference LiCl in D2O at 0.00 pppm):
0.07.
Synthesis of [{TMP(H)}2Li2Zn{OC(dCH2)Mes}4] (3). Com-
pound 3 was prepared following the procedure described for 2 with
the exclusion of TMEDA. A white solid was obtained. Addition
of toluene and gently heating afforded a pale orange solution, which
deposited crystals after cooling slowly at room temperature (0.42
g, 46%). 1H NMR (400 MHz, 25 °C, d6-benzene): 6.77 (s, 8H,
m-H Mes), 5.00 (s br, 4H, CHH′), 4.18 (s br, 4H, CHH′), 2.76 (s,
24H, o-CH3, Mes), 2.18 (s, 12H, p-CH3, Mes), 1.40 (s br, 4H, Hγ,
TMPH), 1.10 (s br, 8H, Hꢀ, TMPH), 0.85 (s br, 24H, CH3, TMPH),
0.14 (s br, 2H, NH, TMPH). 13C{1H} NMR (100.63 MHz, 25 °C,
benzene): 166.71 (OCdCH2), 142.38 (Cipso, Mes), 135.64 (Cortho
,
Mes), 134.94 (Cpara, Mes), 128.71 (Cmeta, Mes), 83.14 (OCdCH2),
55.87 (CH2, TMEDA), 45.65 (CH3, TMEDA), 20.45 (p-CH3, Mes),
20.11 (o-CH3, Mes).
d6-benzene): 162.75 (OCdCH2), 142.20 (Cipso, Mes), 135.84 (Cortho
,
Mes), 135.48 (Cpara, Mes), 128.36 (Cmeta, Mes), 85.34 (OCdCH2),
49.68 (CR, TMPH), 36.06 (Cꢀ, TMPH), 31.08 (CH3, TMPH), 20.25
Acknowledgment. We thank the Royal Society (Univer-
sity Research Fellowship to E.H.) and the Faculty of Science,
University of Strathclyde (starter grant to E.H.), for their
generous sponsorship of this research. We also thank
Professor R. E. Mulvey for helpful discussions.
7
(p-CH3, Mes), 20.12 (o-CH3, Mes), 17.33 31.08 (Cγ, TMPH). Li
NMR (155.50 MHz, C6D6 reference LiCl in D2O at 0.00 pppm):
0.59.
Synthesis of [(TMEDA)2Li2{OC(dCH2)Mes}2] (4). A 0.36 g
(1 mmol) amount of [(TMEDA)LiZn(TMP)Me2] was dissolved in
hexane. 1 (0.33 mL, 1mmol) was then introduced. A white solid
precipitated instantaneously. TMEDA (0.15 mL, 1 mmol) was
added, affording a colorless solution, which was placed in the
freezer at -28 °C. A crop of colorless crystals were deposited
overnight (0.19 g, 67%). 1H NMR (400 MHz, 25 °C, d6-benzene):
Supporting Information Available: CIF file giving crystal data,
computational details, and NMR spectra are available free of charge
via the Internet at http://pubs.acs.org.
OM800658S