Condensation of β-diester titanium enolates with carbonyl substrates: A combined DFT and experimental investigation

Article abstract of DOI:10.1002/chem.200901595

The condensation of dialkyl β-diesters with various aldehydes promoted by TiCl₄ has been studied by DFT approaches and experimental methods, including NMR, IR and UV/ Vis spectroscopy. Various possible reaction pathways have been investigated a

Full text of DOI:10.1002/chem.200901595

FULL PAPER
DOI: 10.1002/chem.200901595
Condensation of b-Diester Titanium Enolates with Carbonyl Substrates: ACHTUNGTRENNUNG
A Combined DFT and Experimental Investigation
Alessandro Marrone,[a] Andrea Renzetti,[a] Paolo De Maria,[a] Stꢀphane Gꢀrard,[b]
Janos Sapi,[b] Antonella Fontana,*[a] and Nazzareno Re*[a]
Dedicated to the Centenary of the Italian Chemical Society
Abstract: The condensation of dialkyl
b-diesters with various aldehydes pro-
moted by TiCl4 has been studied by
DFT approaches and experimental
methods, including NMR, IR and UV/
Vis spectroscopy. Various possible reac-
tion pathways have been investigated
and their energy profiles evaluated to
find out a plausible mechanism of the
reaction. Theoretical results and exper-
imental evidence point to a three-step
mechanism: 1) Ti-induced formation of
the enolate ion; 2) aldol reaction be-
tween the enolate ion and the alde-
hyde, both coordinated to titanium;
and 3) intramolecular elimination that
leads to a titanyl complex. The present-
ed mechanistic hypothesis allows one
to better understand the pivotal role of
titanium(IV) in the reaction.
Keywords: aldol reaction · density
functional calculations
· elimina-
tion · Knoevenagel condensation ·
titanium
Introduction
The formation of carbon–carbon bonds is one of the most
common organic reactions. Among the synthetic routes de-
veloped to achieve this goal, an aldol reaction is a corner-
stone and has been subjected to considerable investigation,
especially in terms of diastereo- and enantiocontrol.[1–4]
Metal ions are particularly useful in catalyzing this reaction
because they can activate both the nucleophile and the elec-
trophile by metal complexation (Scheme 1). Indeed, the co-
ordination of carbonyl compounds to metal centres increas-
es their electrophilicity, thus accelerating the reaction with
nucleophiles such as enolates. At the same time, the com-
Scheme 1. Metal ions catalyze an aldol reaction through activation of
both the nucleophile and the electrophile by metal complexation.
plexation of carbonyl compounds bearing a-hydrogen atoms
allows the generation of the corresponding enolates under
mild reaction conditions.
[a] Dr. A. Marrone, Dr. A. Renzetti, Prof. P. D. Maria, Prof. A. Fontana,
Prof. N. Re
Dipartimento di Scienze del Farmaco
Universitꢀ degli Studi “G. d’Annunzio” di Chieti-Pescara
Via dei Vestini, 66100 Chieti (Italy)
Fax : (+39)0871-3554614
Titanium(IV) derivatives are effective catalysts for several
reactions.[5] In particular, the TiCl4–Et3N reagent system is
widely used to generate titanium enolates for applications in
aldol condensations[6] and related reactions.[7] This method is
simpler than others based on transmetalation from alkali-
metal enolates,[8] silyl enol ethers[9] or acylation of a titano-
cene methylene complex,[10] because in this way titanium
enolates are directly generated in the reaction mixture.[11]
Recently, Renzetti et al.[12] have developed a one-pot trimo-
lecular condensation of isobutyraldehyde, indole and di-
E-mail: fontana@unich.it
nre@unich.it
[b] Dr. S. Gꢁrard, Prof. J. Sapi
Facultꢁ de Pharmacie, ICMR, UMR CNRS 6229
Universitꢁ de Reims Champagne-Ardenne
51 rue Cognacq-Jay, 51096 Reims Cedex (France)
Supporting information for this article is available on the WWW
under http://dx.doi.org/10.1002/chem.200901595.
Chem. Eur. J. 2009, 15, 11537 – 11550
ꢂ 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
11537

However, neither experimental
nor theoretical studies have
ever been performed to ex-
plain the role of titanium in
this reaction. In the present
work, we investigated the
Scheme 2. TiCl4-promoted trimolecular condensation.
mechanism of the condensa-
tion of three dialkyl b-diesters
methyl malonate in the presence of TiCl4 and Et3N
(Scheme 2).
with various aldehydes promoted by TiCl4 (steps I–IV) by a
combination of experimental techniques and DFT calcula-
tions. The elucidation of the trimolecular condensation reac-
tion mechanism would allow the experimental procedure to
be tuned and might shed light on the mechanism of analo-
gous reactions.
This process belongs to the class of multicomponent reac-
tions, which are particularly useful in combinatorial chemis-
try to prepare libraries of compounds having a common
scaffold.[13] Scheme 3 reports the initially proposed mecha-
nism[12] for this trimolecular reaction. Titanium tetrachloride
forms a complex with dimethyl malonate (Scheme 3, step I).
The complexation increases the acidity of the a-hydrogen
atoms because of the electron-withdrawing effect of the
electron-poor titanium(IV). Thus the diester can be easily
deprotonated by a weak base (triethylamine) and converted
into the enolate ion, which is the reactive species (Scheme 3,
step II). The enolate reacts with the aldehyde (Scheme 3,
step III) and generates the Knoevenagel adduct K after the
protonation of the resulting alkoxide and the elimination of
a water molecule (Scheme 3, step IV). The Michael addition
of indole to K followed by deprotonation leads to the con-
densation product CP (Scheme 3, steps V and VI). Accord-
ing to this mechanism, titanium promotes both the Knoeve-
nagel condensation (steps I–IV) and the Michael addition
(steps V–VI). The Knoevenagel condensation promoted by
TiCl4 and Et3N has been known for almost forty years. This
reaction was first performed by Lehnert between aldehydes
or ketones and a wide range of methylene active com-
pounds.[14] It has been used to synthesise some natural prod-
ucts[15] and its stereochemistry has also been studied.[16,17]
Results and Discussion
Formation of the enolate
Experimental evidence: To gain an insight into the reaction
mechanism, NMR spectroscopic studies were undertaken on
three different diesters: dimethyl malonate, diethyl malo-
nate and diisopropyl malonate. A first set of experiments
confirmed the formation of the enolate for all of the three
compounds (Scheme 3, steps I and II). The addition of one
molar equivalent of TiCl4 to a solution of diethyl malonate
in CD2Cl2 at 258C resulted in a dramatic change of the
13C NMR spectrum. The a-carbon atom showed the most
significant change in chemical shift, passing from d=41.4 to
39.1 ppm (Figure 1, compare spectra a and b). Addition of
one molar equivalent of triethylamine caused a new shift in
the signal to d=75.6 ppm and resulted in the formation of a
titanium enolate (Figure 1c), according to previous evidence
on similar enolates.[18] Quench-
ing with a solution of DCl in
D2O resulted in the tautomer-
ACHUTNGRENiNUG saACHTUNGTRENNtUGN ion to the deuterated ester
(results not shown). The cou-
pled 13C NMR spectrum con-
firmed the formation of the
enolate. As a matter of fact,
the a-carbon atom provided a
doublet rather than a triplet on
addition of one molar equiva-
lent of triethylamine (Table 1,
entries 5 and 6). Moreover, the
ACTHNUTRGNEUNG
large increase of 1J(13C,H) on
passing from the ester to the
enolate (from about 130 to
170 Hz) points to a change in
the Ca hybridisation. Both ob-
servations are consistent with
deprotonation.[16] A similar be-
haviour was observed for the
other two diesters (Table 1, en-
Scheme 3. Mechanism for the TiCl4-promoted trimolecular condensation proposed in ref. [12]. R=Me, Et, iPr;
R’=H, Ph, iPr.
11538
www.chemeurj.org
ꢂ 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2009, 15, 11537 – 11550

b-Diester Titanium Enolates
FULL PAPER
Figure 2. Titration plot of diethyl malonate (at the initial concentration
of 0.41m in CD2Cl2) with TiCl4 (T=25.08C, n=mol number).
Table 2. Apparent formation constant of the TiCl4–diethyl malonate
complex at various initial concentrations of ester (DEM) (in CD2Cl2 at
T=25.08C).
Figure 1. 13C NMR spectra of diethyl malonate a) pure; b) after the addi-
tion of TiCl4 (1.0 equiv); c) after the addition of TiCl4 (1.0 equiv) and
Et3N (1.0 equiv) (solvent: CD2Cl2).
A
Kapp [mꢀ1
]
R2 (Points)
8.24ꢃ10ꢀ2
2.23ꢃ10ꢀ1
4.10ꢃ10ꢀ1
6.77ꢃ102
1.10ꢃ103
3.02ꢃ103
0.71213 (10)
0.99311 (11)
0.99083 (13)
Table 1. Enolate formation tests for dimethyl malonate, diethyl malonate
and diisopropyl malonate upon addition of TiCl4 and Et3N (T=25.08C).
Entry
Species
d(Ca) [ppm]
Multiplicity (J [Hz])
1
2
3
4
5
6
7
8
9
dimethyl malonate[a]
+TiCl4 (1 equiv)
+Et3N (1 equiv
41.4
39.1
76.2
41.4
39.1
75.6
42.4
38.3
70.8
t (132.3)
t (133.4)
d (170.9)
t (132.3)
t (133.4)
d (170.9)
t (132.0)
t (132.0)
d (172.0)
periments, concentrations should be replaced with activities
to obtain true equilibrium constants.
diethyl malonate[b]
+TiCl4 (1 equiv)
+Et3N (1 equiv)
diisopropyl malonate[a]
+TiCl4 (1 equiv)
+Et3N (1 equiv)
Theoretical calculations: The formation of titanium enolate
was investigated in detail by a DFT approach that consid-
ered the simplest dimethyl malonate (Scheme 4). An octahe-
[a] The solvent is CDCl3. [b] The solvent is CD2Cl2.
tries 1–3 and 7–9). The 1H NMR spectrum also showed a
shift of signals upon addition of TiCl4 and Et3N. However,
the peaks of the enolate were broad and partly overlapped,
thus making their attribution difficult. The formation of the
enolate ion could be followed visually as well. Actually, the
starting ester is uncoloured; its complex with TiCl4 is yellow
and the titanium enolate is dark red.
We further investigated the TiCl4–diethyl malonate com-
plex to elucidate its structure and stability in solution. The
corresponding Job plot[19] for a constant [diethyl malonate+
Scheme 4. Formation of titanium enolates.
TiCl4] concentration of 0.82m, shows a maximum at cTiCl
=
4
0.5 (Figure S1 in the Supporting Information). This evidence
indicates that the complex has a 1:1 stoichiometry at the in-
vestigated concentration of substrate.[20] The apparent for-
dral geometry was assumed for the initial complex formed
by this b-diester with TiCl4 on the basis of previous experi-
1
mation constant Kapp of the complex, obtained by H NMR
mental investigations.[22] This [TiCl
4ACHTUNGTERN{NUG (O=COCH3)2CH2}]
spectroscopic titration[21] (Figure 2), indicates a highly stable
complex. Three different NMR spectroscopic titrations
point to a strong dependence of Kapp on the initial concen-
tration of diester (Table 2). This is not surprising because in
concentrated solutions, such as those used in the present ex-
complex was optimised and the comparison of the calculat-
ed geometry with the X-ray data was used to assess the ac-
curacy of the employed level of theory. Because no X-ray
structure is available for the complex, we considered the di-
ethyl malonate analogue, the geometry of which has been
Chem. Eur. J. 2009, 15, 11537 – 11550
ꢂ 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
11539

A. Fontana, N. Re et al.
characterised by X-ray diffraction[22] and is expected to be
very close to that of [TiCl4A{(O=COCH3)2CH2}].
The calculated geometrical parameters are in good agree-
ment with the experimental X-ray data (Table S1 in the Sup-
porting Information), with most bond lengths and bond
enolate 2 is thermodynamically favoured over the neutral
complex 2’, probably due to the energy required for the
chloride dissociation in the latter case. These results indicate
that the deprotonation of complex 1 is almost complete and
occurs through step I.a, which leads to the titanium ate eno-
late. However, a small amount of the neutral enolate is ex-
pected to be present at the equilibrium, as its free energy is
not exceedingly higher (11.1 kcalmolꢀ1). The results of theo-
retical calculations are in agreement with experimental evi-
dence. Indeed, if the reaction mixture coming from the addi-
tion of triethylamine to the titanium complex of diethyl mal-
onate is treated with an excess of AgSbF6, AgCl precipitates
and the solution changes colour; it passes from dark red to
light orange. This result suggests the presence of a equilibri-
um [Eq. (1)], although shifted towards the titanium ate eno-
late.
CTHUNGTRENNUNG
ꢀ
angles within 0.05 ꢄ and 58. Only the calculated Ti O bond
lengths are slightly overestimated by approximately 0.10 ꢄ,
whereas most of the remaining bond lengths differ by less
than 0.05 ꢄ from the experimental values. Recent DFT cal-
culations[23] also showed a similar overestimate of the Ti O
ꢀ
bond lengths in TiCl4 complexes with carbonyl compounds
(0.05–0.10 ꢄ) at the same level of theory. However, the
ꢀ
error on the Ti O bond length at this level of theory is
quite low and is not expected to affect the energy profiles of
the considered reactions that occur in the second-coordina-
tion sphere of titanium.
The formation of the enolate ion by deprotonation with
Et3N has been investigated for both the [TiCl4ACTHUNTGRNEUNG{(O=
COCH3)2CH2}] complex and the free dimethyl malonate to
point out the role played by the metal centre. We calculated
the proton-exchange enthalpies and free energies between
ꢀ
*
ðtitanium ate enolateÞ)ꢀðtitanium neutral enolateÞ þ Cl
ð1Þ
The addition of AgSbF6 favours the dissociation of chloride
ions from the ate enolate and shifts the equilibrium towards
the neutral trichlorotitanium enolate (Scheme 5). The pre-
triethylamine and both the metal complex and the free di
ACHTUNGERTNeNUNG s-
ACHTUNGTRENNUNGter. Results (Table S2 in the Supporting Information) show
the pivotal role played by solvation due to dichloromethane.
Indeed, in the gas phase the proton-exchange free energies
are positive in both cases, but that of the complex is about
50 kcalmolꢀ1 lower than that for the free diester, thus indi-
cating that the metal coordination strongly increases the Ha
acidity of dimethyl malonate. The inclusion of solvation re-
duces the proton-exchange free energies by approximately
80–100 kcalmolꢀ1, as the reaction implies the formation of
ionic products from neutral reactants and thus it is favoured
only in a polar solvent. The calculated free energies in
CH2Cl2 were ꢀ15.4 and 19.7 kcalmolꢀ1 for the complexed
and the free diester, respectively. These results indicate that
the dicarbonyl compound can be deprotonated in dichloro-
methane solution only if it is complexed with TiCl4.
Actually, the deprotonation of the metal complex 1 may
lead to the formation of either the anionic tetrachlorotitani-
um enolate 2 (titanium ate enolate; Scheme 4, step I.a), as
assumed in the calculations above, or the neutral trichloroti-
tanium enolate 2’, if a chloride ion dissociates from the com-
plex (Scheme 4, step I.b).
Scheme 5. Equilibrium between the tetrachloro- and the trichlorotitani-
um enolate of diethyl malonate.
cipitation of AgCl from tetrachlorotitanium enolates has
been previously observed in several aldol-type reactions.[11h,i]
Moreover, calculated 13C NMR spectroscopic shifts for the
CO and the Ca on passing from the ate to the neutral eno-
late (Table S4 in the Supporting Information) are roughly
consistent with the experimentally observed shifts (Table S5
in the Supporting Information).
The calculated geometries of the ate and the neutral tita-
nium enolates are reported in Table S3 in the Supporting In-
Practical implications: An in-depth knowledge of the mech-
anism of the investigated trimolecular condensation is cru-
cial for tuning the experimental procedure. Because the
onset of the entire process is the formation of the enolate
ion, any strategy that favours the formation of the enolate
may improve the reaction yields. Firstly, the order of addi-
tion of the reactants is important. If reactants are mixed si-
multaneously, the enolate ion does not form properly and
the yield of the condensation product A is low (Scheme 6).
In this case, a secondary reaction takes place leading mainly
to the bis-indolic derivative B. The best procedure is to mix
the reactants in the following order: 1) dimethyl malonate
plus TiCl4, 2) Et3N, 3) isobutyraldehyde and 4) indole. The
ꢀ
formation. Deprotonation causes a lengthening of the Ti Cl
ꢀ
bonds and a shortening of the Ti O bonds of about 0.2 ꢄ
with respect to the [TiCl
{(O=COCH3)2CH2}] complex. De-
protonation also leads to minor effects: 1) a shortening of
ꢀ
the Ca Cester bond and a lengthening of the C=O bond, and
2) a flattening of the boat-like conformation of the chelating
diesters in the [TiCl4ACHTNUGTRNE{UGN (O=COCH3)2CH2}] complex, which re-
flects the change in the Ca hybridisation from sp3 to sp2 and
the conjugation within the enolate form.
The calculated free energies for the formation of the tita-
nium enolates 2 and 2’ in dichloromethane are ꢀ15.4 and
ꢀ4.3 kcalmolꢀ1, respectively (see Table 3 below). The ate
11540
www.chemeurj.org
ꢂ 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2009, 15, 11537 – 11550

b-Diester Titanium Enolates
FULL PAPER
TiCl4 without molecular sieves
in his Knoevenagel condensa-
tions.[15] The limited role of
molecular sieves in the investi-
gated reactions indicates that
another species that traps
water is present in the reaction
mixture, probably a titanium
species.
Scheme 6. Dependence of the trimolecular reaction yield on the experimental procedure.
Theoretical results also sug-
gest a different pathway for
base has to be added to the reaction mixture containing the
diester and TiCl4 to allow the formation of the enolate. In
the absence of the diester, the base reduces titanium(IV) to
this condensation reaction. The nucleophilic attack of [TiCl4-
AHCTUNGTRENNUNG
step III) was simulated by linear transit calculations in
which the distance between the two carbon atoms involved
in the formation of the new bond (i.e., the Ca of the enolate
and the carbonyl carbon of the aldehyde) was gradually re-
duced from 2.15 to 1.60 ꢄ. These calculations indicate that
the electronic energy of the enolate–aldehyde system in-
creases monotonically by about 25 kcalmolꢀ1 on decreasing
titaniumACHTUNGTRENNUNG(III) and undergoes irreversible reactions in the
presence of aldehydes.[24] Secondly, one must wait enough
time (30 min at 08C) before adding the base to allow the
complexation of all of the diester and to favour its deproto-
nation. This procedure is general because it also works with
linear carbon acids different from b-diesters and heterocy-
cles other than indole (Scheme 7).[12]
ꢀ
the C C distance (Figure S2 in the Supporting Information).
Accordingly, any attempt to obtain an energy minimum of
the enolate–aldehyde covalent adduct led to the separation
ꢀ
of the two molecules. The C C bond formation is thus ther-
modynamically unfavoured, that is, the aldol adduct is not
stable under these conditions. A possible explanation for
this result is that the metal fragment reduces the nucleophi-
licity of the coordinated enolate by electron withdrawal and
that the carbonyl carbon of the aldehyde is not electrophilic
enough to be attacked by the enolate ion. As a matter of
fact, Brønsted acid catalysis is generally advised for aldol re-
actions.[1–3]
Scheme 7. Variation of the reactants in the trimolecular condensation.
Het=substituted indoles and furans; R1 =alkyl, aryl; R2 =Me, Et, iPr;
R3 =COOR2, COMe, PO
ACTHNUTRGEN(UGN OEt)2, NO2.
Formation of the Knoevenagel adduct by means of elimina-
tion of a water molecule
Formation of the Knoevenagel adduct by means of forma-
tion of a titanyl chloride
In the mechanism originally proposed for the considered tri-
molecular reaction (Scheme 2), the Knoevenagel condensa-
tion was assumed to involve the nucleophilic addition of the
titanium enolate to the free aldehyde, thereby leading to an
alkoxide species which, after protonation, eliminates a water
molecule (Scheme 3, steps III and IV). However, some ex-
perimental evidence cast some doubts on the proposed steps
of this condensation. Indeed, it is not clear how alkoxide is
protonated in a basic environment. The elimination of a
water molecule (Scheme 3, step IV) in the reaction mixture
is also unlikely because water reacts vigorously with titani-
um tetrachloride according to the Equation (2):
Theoretical thermodynamics: The presence of TiCl4 in the
reaction medium suggests a possible method to enhance the
electrophilicity of the aldehyde by its coordination to the
metal centre of the titanium enolate complex, followed by
its intramolecular reaction with the adjacent enolate ligand.
Although the hexacoordinated anionic titanium enolate 2 is
coordinatively saturated, the aldehyde can easily replace
either one of the carbonyl groups of the malonate or a chlo-
ride ion (Scheme 8, steps II.a and II.b, respectively). Howev-
er, the pentacoordinated neutral titanium enolate 2’ is al-
ready coordinatively unsaturated and can straightforwardly
coordinate the aldehyde to give the corresponding octahe-
dral complex 3’ (Scheme 8, step II.c).
TiCl4 þ 2 H2O ! TiO2 þ 4 HCl
ð2Þ
The elimination of water in the presence of titanium(IV)
would still be possible if water were trapped by molecular
sieves added into the reaction vessel. However, the trimolec-
ular condensation of indole, dimethyl malonate and iso-
Table 3 reports the calculated enthalpies and free ener-
gies—in the gas phase and in solution—for the coordination
of dimethyl malonate with three aldehydes, that is, formal-
dehyde (R=H), benzaldehyde (R=Ph) and isobutyralde-
hyde (R=iPr). Steps II.a and II.b, which involve the ate
enolate complex, are endothermic, with DH298 values in so-
lution in the range of 14–18 and 1–9 kcalmolꢀ1, respectively.
On the other hand, step II.c, which involves the neutral eno-
ACHTUNGTRENNUNGbutyrACHTUNGTRENNUNGaldehyde promoted by TiCl4 takes place also in the ab-
sence of molecular sieves, even if the yield of the reaction
decreases (the yield is 81% in the presence of molecular
sieves and 42% in their absence). Lehnert himself used
Chem. Eur. J. 2009, 15, 11537 – 11550
ꢂ 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
11541

A. Fontana, N. Re et al.
hyde coordination implies the loss of a translational degree
of freedom.
The subsequent aldol reaction within complexes 3 and 3’
was investigated by addressing the intramolecular nucleo-
philic attack of the Ca of the enolate to the carbonyl carbon
of the aldehyde (Scheme 9 and Table 3, steps III.a and
III.b).
Scheme 8. Possible pathways for aldehyde coordination to titanium.
Table 3. Calculated reaction enthalpies and free energies in the gas
phase and dichloromethane for the steps I–IV of the Knoevenagel con-
densation. Values are in kcalmolꢀ1
.
Scheme 9. Proposed pathways for the intramolecular aldol reaction and
Newman projections of 4 and 4’.
For the simplest formaldehyde substrate, the reaction is
exothermic with reaction enthalpies in solution of ꢀ17.3 and
ꢀ5.9 kcalmolꢀ1 for steps III.a and III.b, respectively. For the
more sterically hindered isobutyraldehyde and benzalde-
hyde substrates, the corresponding processes III.a and III.b
are less exothermic or slightly endothermic. Indeed, the en-
thalpies in solution of processes III.a and III.b are ꢀ3.3 and
7.7 kcalmolꢀ1 for R=Ph, or ꢀ5.7 and ꢀ4.4 kcalmolꢀ1 for
R=iPr, respectively. Notice that for all aldehydes the intra-
molecular aldol reaction is more favourable for 3 (step III.a)
than for 3’ (step III.b), probably due to the higher ring
strain in 4’.
Indeed, products 4 and 4’ have quite different structures:
the former is a metallacycle and the latter is a metallabicy-
cle. The metallacycle unit in 4 assumes a chair-like confor-
mation. For R=H, the uncoordinated ester group of the
product occupies an equatorial position. For R=Ph and R=
iPr, the alkyl group and the carboxymethyl group may
assume two different conformations: equatorial–equatorial,
which corresponds to the anti-aldol product, and axial–equa-
torial, which corresponds to the syn-aldol (non-Evans) pro-
duct.[6a,11g–i] However, the formation of the anti adduct is pre-
vented by the steric clashes between the axial group and the
TiCl4 metal moiety. For these reasons, in the case of benzal-
dehyde and isobutyraldehyde the geometry could be opti-
mised only for the syn adduct (see Scheme 9).
Step
R
ACHTNGUTRENNUG CATHUNGRTEN(GNUN sol) DDGsolv
DH298(g) DG298(g) DH298(sol) DG298
I.a
I.b
II.a
II.a Ph
II.a iPr
II.b
II.b Ph
II.b iPr
II.c
II.c
II.c
III.a
III.a Ph
III.a iPr
III.b
III.b Ph
III.b iPr
IV.a
H, Ph, iPr
H, Ph, iPr 108.5
H
64.2
64.5
107.2
23.1
32.1
24.0
44.0
42.7
49.6
1.3
ꢀ15.7
ꢀ15.4
ꢀ4.3
25.7
35.6
27.5
10.4
16.6
18.5
ꢀ79.9
ꢀ111.5
2.6
ꢀ2.9
11.7
14.7
12.5
35.1
32.7
39.6
ꢀ9.3
ꢀ11.7
ꢀ4.8
ꢀ8.5
1.3
14.3
18.2
16.0
1.5
6.6
8.5
3.5
3.6
H
ꢀ33.6
ꢀ26.1
ꢀ31.1
ꢀ2.0
H
ꢀ11.3
ꢀ6.2
ꢀ4.3
ꢀ17.3
ꢀ3.3
ꢀ5.7
ꢀ5.9
7.7
ꢀ4.4
ꢀ13.6
ꢀ17.8
ꢀ18.3
ꢀ15.6
ꢀ17.0
ꢀ14.3
ꢀ0.7
Ph
iPr
H
0.1
7.0
5.5
5.4
0.5
7.4
ꢀ6.4
ꢀ15.2
ꢀ1.5
ꢀ9.0
ꢀ3.6
10.4
ꢀ1.3
ꢀ18.5
ꢀ21.1
ꢀ22.5
ꢀ18.4
ꢀ20.9
ꢀ18.3
ꢀ8.8
3.1
ꢀ4.6
3.2
0.0
1.8
13.5
4.4
ꢀ9.0
H
ꢀ0.4
10.7
1.3
ꢀ5.5
ꢀ3.1
ꢀ5.7
H
99.9
95.4
93.1
107.0
108.2
110.0
95.0
92.0
88.5
104.2
104.3
106.1
ꢀ113.6
ꢀ113.2
ꢀ113.1
ꢀ122.6
ꢀ125.2
ꢀ124.4
IV.a Ph
IV.a iPr
IV.b
IV.b Ph
H
IV.b iPr
late, is exothermic, with DH298 in solution in the range from
ꢀ12 to ꢀ4 kcalmolꢀ1. These results are not surprising, be-
cause the coordination of the aldehyde to the pentacoordi-
nated neutral enolate occurs without the cleavage of any
bond. As expected, the entropic effect in steps II.a and II.c
is unfavourable by approximately 11 kcalmolꢀ1, as the alde-
The final elimination that leads to the Knoevenagel
ꢀ
adduct requires the breaking of the aldehyde C O bond.
On the basis of the high oxophilicity of titanium,[25] we as-
sumed that the aldehyde oxygen leaves as a titanyl group
(Ti=O, Scheme 10). The titanyl group has a double-bond
character,[26] so it can be formed only if a chloride ion leaves
11542
www.chemeurj.org
ꢂ 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2009, 15, 11537 – 11550

b-Diester Titanium Enolates
FULL PAPER
A (I.a–II.a–III.a–IV.a, solid line) and B (I.a–II.b.–III.b–IV.b
or I.b–II.c–III.b–IV.b, dashed line), leading respectively to 5
and 5’. Indeed, the two alternative routes of pathway B are
equivalent because 2 and 2’ are in equilibrium. In both path-
ways the overall reaction is highly exothermic in solution,
with DH values of ꢀ23.7 and ꢀ25.9 kcalmolꢀ1, respectively
(Figure 3). Similar values are observed for formaldehyde
and benzaldehyde. To see which is the most favoured path-
way, we studied the kinetics of the Knoevenagel condensa-
tion.
Theoretical kinetics: We investigated the kinetics of the al-
dehyde coordination to the titanium enolate (steps II.a–
II.c). The reaction II.c is a simple ligand association in
which the aldehyde binds to the trichlorotitanium enolate 2’.
This process is expected to have a small barrier mainly re-
sulting from the metal-fragment reorganisation. The coordi-
nation of the aldehyde with the titanium ate enolate 2
occurs through a ligand-exchange process in which either an
ester carbonyl group (II.a) or a chloride ion (II.b) is re-
placed with the carbonyl oxygen of the aldehyde. We as-
sumed a dissociative mechanism for this ligand exchange, in
which titanium loses the leaving ligand before binding to the
incoming one. The energy barriers of steps II.a and II.b
were first estimated by linear transit calculations in which
the distances between titanium and the leaving ligand (O=
Cester for II.a and Cl for II.b) are increased from the value
they assume in 2 (2.0 and 2.4 ꢄ respectively) to a value
(4.5 ꢄ) large enough to consider the ligand uncoordinated
(Figure S3 in the Supporting Information). The results indi-
cate that the energy barriers of steps II.a and II.b are 23.8
and 20.7 kcalmolꢀ1, respectively, in solution.
Scheme 10. Possible pathways for the formation of the Knoevenagel
adduct.
the metal and the coordination number of titanium decreas-
es from six to five.
Large and positive gas-phase enthalpies are calculated for
steps IV.a and IV.b (Table 3), thereby underlining the high
energy cost associated with the cleavage of three bonds, that
ꢀ
ꢀ
ꢀ
is, C H, C O and Ti Cl. This energy is only partially count-
ꢀ
ꢀ
ACHTUNGTRENNUNGerACHTUNGTRENNUNGbalanced by the strengthening of the Ti O and Caldehyde
Ca bonds and by the base protonation. However, the forma-
tion of the ions Et3NH+ and Clꢀ gives rise to large and neg-
ative solvation contributions that overcome the energy costs,
thus making both processes exothermic and exoergonic in
solution. The calculated reaction enthalpies and free ener-
gies for steps IV.a and IV.b are not significantly affected by
the nature of R, probably implying that substituent steric ef-
fects are negligible, as confirmed by the calculated geome-
tries. The titanium complexes 5 and 5’ maintain the pyrami-
dal-square coordination geometry, with apical oxygen.
Figure 3 reports the enthalpy diagram for the Knoevena-
gel condensation of isobutyraldehyde and dimethyl malo-
nate in solution. Two pathways are possible for this reaction:
We then investigated the kinetics of reactions III.a and
III.b. The aldol reactions are characterised by relatively low
activation enthalpies of 6–16 kcalmolꢀ1, and activation free
energies of 8–19 kcalmolꢀ1, in
solution (Table 4). The three
aldehydes show
a different
trend of reactivity. In pa-
th III.a, the activation enthal-
pies follow both the hindrance
and the electronic effect order
(i.e., iPrꢁPh>H), with form-
aldehyde being the most reac-
tive one. In path III.b, the acti-
vation enthalpies follow the
order Ph>H>iPr with iso-
ACHUTNGRENbNUG utyrACHTUNGTNERaNUGN ldehyde being unexpect-
edly the most reactive. The
lower reactivity of 3’ with R=
Ph could be due to the conju-
gation between the carbonyl
group and the phenyl ring,
which reduces the electrophi-
licity of the aldehyde substrate.
To gain further insight into the
influence of steric effects on
Figure 3. Calculated enthalpy diagrams for the Knoevenagel condensation of isobutyraldehyde and dimethyl
malonate in CH2Cl2. Solid line (c): path A. Dashed line (a): path B.
Chem. Eur. J. 2009, 15, 11537 – 11550
ꢂ 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
11543

A. Fontana, N. Re et al.
Table 4. Calculated activation enthalpies and free energies in the gas
phase and dichloromethane for the aldol reaction (step III). Values are in
This difference could explain why the isobutyrraldehyde is
more reactive in III.b rather than in III.a.
Two different mechanisms have been considered for the
elimination reactions: an E1cB-like mechanism (route IV.a)
and an E2-like mechanism (route IV.b). Route IV.a occurs in
two steps (Scheme 11a): 1) deprotonation of Ca in 4 by Et3N
kcalmolꢀ1
.
Step
III.a
R
DH°(g)
DG°(g)
DH
(sol)
DG
(sol)
DDG°
solv
H
10.7
13.7
17.2
11.2
16.8
9.0
12.0
16.0
14.3
14.1
20.1
12.1
7.1
13.0
13.3
10.3
16.0
6.3
8.4
15.3
10.4
13.2
19.2
9.4
ꢀ3.6
ꢀ0.7
ꢀ3.9
ꢀ0.9
ꢀ0.8
ꢀ2.7
Ph
iPr
H
Ph
iPr
III.b
the kinetics of these aldol reactions, we analyzed in detail
the geometries of the calculated transition states for R=iPr
in the two paths, TSIIIa and TSIIIb (Figure 4). They are both
late transition states and reflect essentially the typical Zim-
merman–Traxler cyclic structure.[27] In both transition states
the isopropyl group is close to one of the ester groups. How-
ever, in TSIIIa this ester group is not coordinated to the
metal and can exert a steric hindrance towards the isopropyl
group. In TSIIIb the ester group is held far away by the metal
coordination, which leads to a less hindered transition state.
Scheme 11. Proposed mechanisms for the Knoevenagel condensation in
CH2Cl2: a) step IV.a and b) step IV.b.
(the strongest base in the reaction medium) with formation
of the corresponding carbanion I; 2) formation of the C=Ca
ꢀ
double bond with the breaking of the Caldehyde O bond, for-
mation of a titanyl group and dissociation of a chloride ion.
The intermediate I is relatively stable due to the delocalisa-
tion of the negative charge on the proximal ester group
(Scheme 11a), thus making the deprotonation of 4 a rela-
tively favourable process. Route IV.b is characterised by a
concerted proton transfer from Ca to Et3N, the formation of
a titanyl complex and the release of a chloride ion from the
metal complex. The deprotonation of 4’ (Scheme 11b) is an
unfavourable process because the corresponding carbanion
lacks of any resonance stabilisation. Figure 4 reports the op-
timised geometries of the intermediate I and the transition
states TSIVa1, TSIVa2 and TSIVb. The discussion will be limited
to the most representative structures with R=iPr because
the nature of R did not significantly affect the geometry of
the calculated structures.
The geometry of TSIVa1 is that expected for a typical
proton exchange from the acidic Ca to the basic nitrogen of
Et3N. The proton transfer occurs along the direction of the
Figure 4. Calculated transition state geometries a) of the transition states
TSIIIa and TSIIIb for the paths III.a and III.b and b) of the intermediate
and the transition states TSIVa1, TSIVa2 and TSIVb for the paths IV.a and
IV.b. Interatomic distances and bond angles are reported in Angstrom
[ꢄ] and degrees [8], respectively.
ꢀ
Ca H bond, thus minimising the steric repulsion between
the base and the equatorial R group. The geometry of TSIVa2
ꢀ
ꢀ
is characterised by the incipient breaking of C O and Ti Cl
ꢀ
ꢀ
bonds and the formation of Ti O and C Ca p bonds. The
11544
www.chemeurj.org
ꢂ 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2009, 15, 11537 – 11550

b-Diester Titanium Enolates
FULL PAPER
energy imbalance between the breaking of two strong bonds
and the formation of two weak bonds suggests a high-
energy TSIVa2. Nevertheless, TSIVa2 is a late transition state
and benefits from the stabilisation associated with the incipi-
ent formation of two monocharged and highly stabilised
anions. The relatively high value of 2.177 ꢄ for the length of
the Knoevenagel condensation is carried out in the presence
of a chloride scavenger, such as AgSbF6, the reaction is
forced to follow pathway B (I.a–II.b–III.b–IV.b or I.b–II.c–
III.b–IV.b, Figure 3) and, due to the high activation enthalpy
foreseen for step IV.b, the Knoevenagel adduct hardly
forms. As a consequence, the final trimolecular condensa-
tion product is obtained only in traces (Table S7 in the Sup-
porting Information, entry 1).
Calculations indicate that the solvent plays an important
role in the overall Knoevenagel condensation, with DDGsolv
values being essential to make the condensation exothermic
(Table 3). Experimental evidence also supports these theo-
retical results, thus showing that the yield of the trimolecular
condensation product increases by increasing the solvent po-
larity (14% in toluene and 81% in dichloromethane).[12]
The formation of titanyl
ꢀ
the breaking C O bond confirms the late transition state.
The molecular structure of TSIVb is formed by three mo-
lecular fragments: the basic Et3N that accepts the proton
from the acidic Ca, the incipient titanyl complex and the dis-
sociating chloride ion. The calculated distances were 1.561
ꢀ
ꢀ
ꢀ
ꢀ
for N H, 1.480 for Ca H, 1.783 for C O, 1.724 for Ti O
ꢀ
and 2.773 ꢄ for Ti Cl . Figure 5 reports the energy profiles
calculated for routes IV.a and IV.b. The activation enthalpies
corresponding to transition states TSIVa1 and TSIVa2 are quite
chloride has been hypothesised
in various reactions promoted
by titanium(IV) that formally
imply the elimination of
a
water molecule;[7d,26] these in-
clude the Knoevenagel con-
densation performed by Leh-
nert.[21d]
The formation of titanyl
chloride has practical implica-
tions on the reaction condi-
tions that are also consistent
with the experimental evi-
dence. The fact that TiCl4 is
consumed during the Knoeve-
nagel condensation explains
why the trimolecular reaction
is not catalytic and requires a
stoichiometric
amount
of
TiCl4. Unfortunately, to the
best of our knowledge, there is
no reagent able to recover
TiCl4 from TiOCl2, so there is
no way to make this reaction
Figure 5. Calculated enthalpy paths for the processes IV.a and IV.b of isobutyraldehyde in CH2Cl2. R.C.=reac-
tion coordinate.
low, 11.0 and 11.8 kcalmolꢀ1, respectively, thereby suggest-
ing that IV.a is a kinetically facile process. On the other
hand, the activation enthalpy corresponding to TSIVb is
39.9 kcalmolꢀ1 and is therefore a kinetically unfavoured pro-
cess.
In summary, reaction IV.a is kinetically much more fav-
oured than reaction IV.b, so the Knoevenagel condensation
takes place through pathway A (steps I.a–II.a–III.a–IV.a,
Figure 3).
catalytic. One possible drawback of the proposed mecha-
nism is that it requires two molar equivalents of base where-
as only one is experimentally used. However, because the
final b elimination (step IV.a) is a highly favoured process,
any base may easily deprotonate compound 4, thus explain-
ing why the whole Knoevenagel condensation requires only
one molar equivalent of Et3N.
To get further experimental evidence for the formation of
titanyl complexes in the last step of the Knoevenagel con-
densation (b elimination), we performed some analytical
and synthetic investigations.
Experimental evidence: Experimental evidence agrees with
the picture emerging from theoretical calculations, thereby
indicating two possible pathways for the Knoevenagel con-
densation: one passing through the fully chlorinated com-
plex 3 (pathway A) and one passing through complex 3’, in
which titanium has lost a chloride ion (pathway B), the
latter being significantly kinetically unfavoured. Indeed, if
NMR spectra: Three separate experiments were performed
on the Knoevenagel adduct of benzaldehyde. In the first
one, we recorded the H NMR and the 13C NMR spectra of
1
the purified Knoevenagel adduct Ky (Tables 5 and 6,
entry 1). In the second experiment, we prepared the adduct
Chem. Eur. J. 2009, 15, 11537 – 11550
ꢂ 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
11545

A. Fontana, N. Re et al.
Table 5. 1H NMR spectra of benzylidene malonate Ky.
(Ky), its complex with TiCl4 (TiCl4–Ky) and the correspond-
ing adduct formed in the reaction mixture. The most intense
absorbance bands are reported in Table S10 in the Support-
ing Information. The UV/Vis profile of the adduct formed
in the reaction mixture and that of TiCl4–Ky are dominated
by a broad band at l=350 nm (Table S10 in the Supporting
Information, entries 11 and 7, respectively). This band is
probably due to the presence of the metal because it is not
detected in the free Knoevenagel adduct Ky.
The spectra of TiCl4–Ky and the adduct in the reaction
mixture show some differences as well. The UV region of
the former spectrum is characterised by one band at 281 nm
and one at 350 nm, whereas the latter is characterised by
four bands at 245, 290, 332 and 350 nm. These results indi-
cate that the two complexes are different from each other,
which is in agreement with the formation of the titanyl com-
plex 5 in the latter mixture.
Entry
d(H2)
[ppm]
d
G
)
dACHTUNGTRENNUNG
E
E
ACHTUNGTRENNUNG
1
2
3
Ky
7.78
7.36–
7.45
7.41
(br)
7.45–
7.58
3.85 and 3.84 (6H)
3.82 (6H)
Ky in the reaction mix- 7.77
ture
Ky+TiCl4 (1.0 equiv)
(br)
8.62
4.27 (3H) and 4.01
(3H)
Table 6. 13C NMR spectra of benzylidene malonate Ky.
Entry
d(C1),
d(C2)
[ppm]
d(C3),
d(C4)
[ppm]
d
U
dACHTNURGTNEG(UN CAr)
[ppm]
Time-dependent density functional theory (TDDFT) cal-
culations were carried out on free Ky, TiCl4–Ky and 5 in so-
lution. These calculations allowed us to assign the absorb-
ance bands observed in the experimental UV/Vis spectra to
specific electronic transitions (Table S10 in the Supporting
Information). The observed bands for the free Knoevenagel
adduct at 281 nm (calcd: 253, 277 and 284 nm) and 220 nm
(calcd: 215 nm) are due to p!p* transitions involving the
aromatic moiety with only slight contributions due to
n(O)!p* transitions from the ester oxygen lone pairs. A
band at 281 nm (calcd: 289 nm) is also detected in the
TiCl4–Ky spectrum, but is assigned to an n(Cl)!p* transi-
tion from the chlorine lone pairs of the TiCl4 moiety. The
two bands observed for the reaction mixture at 290 and
245 nm (calcd: 278 and 246 nm, respectively) are assigned to
1
2
3
Ky
143.3, 125.9 167.5, 164.8 53.0
129.2–
133.2
129.2–
131.2
Ky in the reac-
tion mixture
Ky+TiCl4
135.0, 128.6 168.2, 168.7 55.3
134.5, 132.4 171.4, 176.0 57.9, 56.8 129.5–
134.5
(1.0 equiv)
directly in the NMR spectroscopy tube by adding TiCl4 and
Et3N to benzaldehyde and dimethyl malonate (see the Ex-
perimental Section; Tables 5 and 6, entry 2). In the third ex-
periment, we recorded the NMR spectra of the pure adduct
Ky in the presence of one equivalent of TiCl4 (Tables 5 and
6, entry 3). We found different H NMR and 13C NMR spec-
1
tra for each of the three experiments. In particular, the
Knoevenagel adduct prepared in the NMR spectroscopy
tube showed NMR spectroscopic signals shifted with respect
to those of the pure adduct (Tables 5 and 6; compare en-
tries 1 and 2). These results highlight that the Knoevenagel
adduct is bound to titanium in the reaction mixture. Howev-
er, the spectra of this complexed adduct are different from
those of the complex TiCl4–Ky (Tables 5 and 6; compare en-
tries 2 and 3). The only explanation for these results is that
titanium is not present as a tetrachloride complex in the re-
action mixture. This result is consistent with the formation
of the titanyl complex 5. The same behaviour was observed
for the Knoevenagel adduct of isobutyraldehyde (Tables S8
and S9 in the Supporting Information).
n
G
corresponds to the n(Cl)!p* transition involving the chlor-
ine lone pairs, the same observed for the peak at 281 nm
of the TiCl4–Ky adduct. The signal at 245 nm is due to
nN
ence supports the formation of complex 5. TDDFT calcula-
tions also allowed us to assign the broad band at 330–
350 nm observed for TiCl4–Ky and for the adduct in the re-
action mixture. For both complexes the most intense transi-
tion is calculated at 330 nm with less intense peaks around
this value, which is in agreement with the broad experimen-
tal spectral shape. However, the peak at 330 nm involves
different transitions in the two cases: d!p* and d!d* for
the TiCl4–Ky adduct and n
G
UV/Vis spectra: UV/Vis absorbance studies were performed
to further support the hypothesis of the formation of 5.
Indeed, titanium coordination is expected to affect the UV/
Vis profile of the Knoevenagel adduct through the electron-
ic transitions involving metal d orbitals. This is particularly
true for benzaldehyde complexes because the presence of
an aromatic group conjugated with the first coordination
shell of the metal is expected to give some UV/Vis signals
that are diagnostic of the metal coordination. Analogously
to NMR spectroscopic measurements, we recorded the UV/
Vis spectra of the free Knoevenagel adduct of benzaldehyde
transition at 424 nm calculated for 5, although weak, is re-
sponsible for the yellow colour of the reaction mixture; no
band is experimentally observed in this region of the spec-
trum, probably due to the low concentration used in the ex-
periment.
IR spectra: IR spectroscopy provided further evidence of
the formation of a titanyl complex. The adducts Ky and Kz
in the reaction mixture show two peaks at around 1000 cmꢀ1
that are absent in the pure adduct and in its complex with
TiCl4 (Tables S11 and S12 in the Supporting Information,
11546
www.chemeurj.org
ꢂ 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2009, 15, 11537 – 11550

b-Diester Titanium Enolates
FULL PAPER
entries 1–3). These signals can be assigned to the titanyl
group Ti=O, based on the analogy with the spectra of simi-
lar titanyl complexes.[26,28a] If the complex Ky in the reaction
mixture is treated with water, thus causing the hydrolysis of
titanium(IV), the same signals around 1000 cmꢀ1 are ob-
served (Table S11 in the Supporting Information, entry 4),
thus supporting their attribution to the titanyl group.
has to be formed for the Knoevenagel adduct to undergo
the Michael addition.
Conclusion
The titanium-promoted Knoevenagel condensation of di-
methyl malonate and three different aldehydes has been in-
vestigated by theoretical and experimental approaches. A
plausible three-step mechanism has been proposed and each
step is assisted by the presence of titanium: 1) the initial de-
protonation of diethyl malonate is favoured by the acidify-
ing effect due to the coordination of titanium(IV) to the die-
ster; 2) the following aldol condensation is promoted by the
coordination of titanium(IV) to both the reagents, the eno-
late and the aldehyde; 3) the intramolecular elimination of
water is feasible because of the formation of a titanyl com-
plex. The results of this study are very interesting because
they would allow for the tuning of the experimental proce-
dure of the trimolecular condensation, thereby reducing the
formation of secondary products and improving the reaction
yield. Moreover, the information provided by DFT calcula-
tions, such as the structures of transition states and inter-
mediates, can be used to design stereocontrolled TiCl4–
Et3N-promoted trimolecular condensation. In the light of
these results, we are currently investigating synthetic ap-
proaches based on the stereochemical implications of this
model.
Synthesis and reaction of the proposed intermediate 5: One
way to investigate a reaction mechanism is to isolate a possi-
ble intermediate from the reaction mixture and to use it for
the preparation of the final product under the same experi-
mental conditions. In our case, the intermediate 5 is unstable
ꢀ
in air because of the presence of reactive Ti Cl bonds and
therefore could not be isolated. Nevertheless, we could pre-
pare 5 directly using a different procedure. The further addi-
tion of indole, under the same experimental conditions, al-
lowed to obtain the final trimolecular condensation product.
If isobutylidene malonate Kz is treated with one equiva-
lent of indole in CH2Cl2 in the absence of TiCl4, no reaction
takes place (Scheme 12a). If Kz is treated with one equiva-
Experimental Section
Computational details: All calculations were carried out at density func-
tional theory level by using the Jaguar 6.0 program.[29] Geometry optimi-
sation has been performed by using the B3LYP functional[30,31] with the
6-31G** basis set for main groups elements (C, O, N, H) and by the Los
Alamos LACV3P[32–34] basis for Ti; this includes relativistic effective core
potentials. This basis set will be thereafter referred to as LACV3P**. The
geometry of reactants, intermediates, products (local minima) and transi-
tion states (saddle points) of reaction obtained at this level of theory
have been used to calculate the reaction path in both the gas and solution
phase. Intrinsic reaction coordinate (IRC) calculations[35,36] have been fur-
ther carried out to assess the configuration of reactants and products that
are connected through the involved TS. Subsequent vibrational frequency
calculations based on analytical second derivatives at the B3LYP/
LACV3P** level of theory have been carried out to confirm the nature
of the local minima and transition state and to compute the zero-point-
energy (ZPE) and vibrational entropy corrections at 298.15 K. Solvation
free energies were evaluated by a self-consistent reaction field (SCRF)
approach[37,38] based on accurate numerical solutions of the Poisson–
Boltzmann equation.[39, 40] Solvation calculations were carried out on the
gas-phase geometry at the B3LYP/LACV3P** level of theory by employ-
ing a dielectric constant of 8.93 and a solvent radius of 2.67 for dichloro-
methane. The activation and reaction enthalpies in solution have been es-
timated by adding the solvation free energy to the corresponding gas-
phase enthalpies. Their variations are expected to be reliable since the
solvation entropy variations are likely to be negligible. UV/Vis spectra
have been calculated for benzylidene malonate, its complex with TiCl4
and the hypothesised titanyl complex 5’ by using Gaussian 03 program
package.[41] Geometry optimisations have been first performed at B3LYP/
LACV3P** level of theory in a vacuum, then time-dependent DFT calcu-
lations at the same level of theory have been carried out to calculate the
Scheme 12. Reaction of isobutylidene malonate Kz and indole under var-
ious reaction conditions.
lent of TiCl4 in CH2Cl2, the solution becomes yellow and the
NMR spectrum of the adduct changes (Tables S8 and S9 in
the Supporting Information), thereby suggesting again the
formation of a TiCl4–Kz complex. Nevertheless, upon fur-
ther addition of one equivalent of indole, the condensation
product is not obtained and a complex mixture of com-
pounds is recovered (Scheme 12b). Because neither the pure
Knoevenagel adduct nor its complex with TiCl4 afford the
final product under the adopted reaction conditions, it can
be inferred that neither of them is the intermediate of the
trimolecular condensation. On the contrary, if TiCl4–Kz is
ꢀ
treated with one equivalent of water, only two of the Ti Cl
bonds are expected to hydrolyse, thus leading to 5. Upon
further addition of one equivalent of Et3N and one equiva-
lent of indole, the condensation product forms in 33% yield
(Scheme 12c). These results indicate that titanyl complex 5
Chem. Eur. J. 2009, 15, 11537 – 11550
ꢂ 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
11547

A. Fontana, N. Re et al.
electronic transitions with the corresponding oscillator strengths for a
total number of 30 states. The analyzed spectrum region for the three
Knoevenagel adducts ranged from 250 to 600 nm, in which the most in-
tense and informative signals are expected for the considered species.
The 13C chemical shifts for the free diethyl malonate, its complex with
TiCl4 and the corresponding ate and neutral enolates have been calculat-
ed at the same level of theory by using the approach implemented in the
Jaguar code.[42]
on silica gel (eluent petroleum ether/ethyl acetate 9:1), then put in the
refrigerator. Product
(300 MHz, CD2Cl2): d=3.84 (s, 3H; OCH3), 3.85 (s, 3H; OCH3), 7.36–
7.45 (m, 5H), 7.78 ppm (s, 1H; H2); 13C NMR (75 MHz, CDCl3): d=53.0
(OCH3); 125.9 and 143.3 (C1 and C2); 129.2–133.2 (CAr); 164.8 and
167.5 ppm (C3 and C4).
6
crystallised spontaneously (36%). 1H NMR
Synthesis of Ky in the NMR spectroscopy tube: A solution of dimethyl
malonate in deuterated solvent (500 mL, 4.38ꢃ10ꢀ1 m) was prepared in an
NMR spectroscopy tube. The tube was cooled in an ice-water bath. TiCl4
(2.19ꢃ10ꢀ4 mol, 1.0 equiv), Et3N (2.19ꢃ10ꢀ4 mol, 1.0 equiv) and benzal-
dehyde (2.19ꢃ10ꢀ4 mol, 1.0 equiv) were added in sequence. The mixture
was left in the bath for 3 h before the 1H NMR and 13C NMR spectra of
the solution were recorded at 25.08C.
General remarks: All solvents were dried and purified by standard proce-
dures prior to use. Melting points were determined using a Reichert
Thermovar hot-stage apparatus. Reactions were monitored with Fluka
TLC aluminium sheets (Kieselgel 60F254) and flash chromatography was
carried out with silica gel 60 (230–400 mesh ASTM) supplied by Fluka.
1H NMR (300 MHz) and 13C NMR (75 MHz) spectra were acquired
using a Varian AC 300 spectrometer at (25.0ꢂ0.1)8C. UV/Vis spectra
were acquired using a Cary 100 Bio spectrophotometer at (25.0ꢂ0.1)8C
with 1 mm cuvettes. IR spectra were measured using a Varian spectro-
photometer (4000–400 cmꢀ1). The solutions in CH2Cl2 were prepared in a
glove box under a nitrogen atmosphere and placed between two KBr
discs. The adherence of the liquid film to the discs prevented the hydroly-
sis of the complex.
Complexation of Ky with TiCl4: Titanium tetrachloride (3.63ꢃ10ꢀ5 mol,
1.0 equiv) was added to
a
solution of Ky (8.00 mg, 3.63ꢃ10ꢀ5 mol,
1
1.0 equiv) in deuterated solvent (500 mL). H NMR and 13C NMR spectra
of the mixture were recorded before and after the addition of TiCl4.
Dimethyl isobutylidenemalonate (Kz):[44] Dimethyl malonate (8.76 mmol,
1.0 equiv) was added to a solution of TiCl4 (1.0 equiv) in of dry dichloro-
methane (20 mL) at 08C while stirring under a nitrogen atmosphere, in
the presence of 3 ꢄ molecular sieves. After 30 min, triethylamine was
added (1.0 equiv) to the resulting yellow suspension and the mixture
turned immediately dark red. Isobutyraldehyde (1.0 equiv) was added
and the mixture was stirred for 3 h at 08C. On addition of a 1m HCl
aqueous solution (20 mL), two phases formed. The aqueous layer was
separated, the organic phase was dried over MgSO4 and evaporated to
dryness under reduced pressure. The crude residue was purified by flash
chromatography on silica gel (eluent petroleum ether/ethyl acetate 9:1),
then distilled at reduced pressure to afford 7 as an uncoloured liquid
(76%). 1H NMR (300 MHz, CD2Cl2): d=1.05 (d, J3,4 =J3,5 =6.6 Hz, 1H;
H4 and H5), 2.65 (dst, J2,3 =10.8 Hz, J3,4 =J3,5 =6.6 Hz, 1H; H3), 3.73 (s,
3H; OCH3), 3.78 (s, 3H; OCH3), 6.78 ppm (d, J2,3 =10.8 Hz, 1H; H2);
13C NMR (75 MHz, CD2Cl2): d=22.1 (C4 and C5), 29.8 (C3), 52.5
(OCH3), 52.6 (OCH3), 126.1 (C1), 156.1 (C2), 164.8 and 166.3 ppm (C6
and C7).
Enolate formation test, general procedure: A solution of carbon acid
(1.0 equiv) in deuterated solvent (700 mL) was prepared. TiCl4
(1.0 equiv), Et3N (1.0 equiv) and DCl/D2O 35% w/w (3.0 equiv) were
added in sequence. 1H NMR, 13C NMR and coupled 13C NMR spectra of
the solution were acquired after each addition. The concentrations of the
carbon acids were 8.75ꢃ10ꢀ2 m in CDCl3 for dimethyl malonate and di-
isopropyl malonate and 4.10ꢃ10ꢀ1 m in CD2Cl2 for diethyl malonate.
NMR spectroscopic titrations: An NMR spectroscopy tube was filled
with a solution of the host diester, H, in a deuterated solvent (CDCl3 or
CD2Cl2) at three different concentrations ([H]=8.24ꢃ10ꢀ2, 2.23ꢃ10ꢀ1
and 4.10ꢃ10ꢀ1 m). TiCl4 [i.e., the guest G of the following Eq. (3)] was
added to the diester [i.e., the host H of the following Eq. (3)] solution in
the NMR spectroscopy tube with increasing volumes from 0 to 2 equiv
(relative to the diester). Volume and concentration changes were taken
into account in the calculation. The binding constants were determined
by nonlinear regression of Equation (3):
Synthesis of Kz in the NMR spectroscopy tube: A solution of dimethyl
malonate in deuterated solvent (500 mL, 4.38ꢃ10ꢀ1 m) was prepared in an
NMR spectroscopy tube. The tube was cooled in an ice-water bath. TiCl4
(2.19ꢃ10ꢀ4 mol, 1.0 equiv), Et3N (2.19ꢃ10ꢀ4 mol, 1.0 equiv) and isobutyr-
aldehyde (2.19ꢃ10ꢀ4 mol; 1.0 equiv) were added in sequence. The mix-
ture was left in the bath for 3 h before the 1H NMR and 13C NMR spec-
tra of the solution were recorded at 25.08C.
rffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi
ꢀ
ꢁ
þ ½Gꢃ0 þ ½Hꢃ0 2ꢀ4½Gꢃ0½Hꢃ
1
K þ ½Gꢃ0 þ ½Hꢃ0 ꢀ
1
K
dobsd ꢀ d0
¼
0 ðdINF ꢀ d0Þ
2½Gꢃ0
ð3Þ
Complexation of Kz with TiCl4: Titanium tetrachloride (3.63ꢃ10ꢀ5 mol,
1.0 equiv) was added to
a
solution of Kz (8.00 mg, 3.63ꢃ10ꢀ5 mol,
1
1.0 equiv) in deuterated solvent (500 mL). H NMR and 13C NMR spectra
in which the binding constant K=[complex]/([H]eq[G]eq), dobsd is the ex-
perimentally measured chemical shift after each addition of titrant, d0 is
the chemical shift of a nucleus in the host molecule, dINF is the chemical
shift of a nucleus in the host–guest complex, [H]0 is the initial concentra-
tion of the host and [G]0 the added concentration of the guest.
of the mixture were recorded before and after the addition of TiCl4.
UV/Vis spectra of the free Knoevenagel adduct (Ky), its complex with
TiCl4 ACHTNUTRGNEU(GN TiCl4–Ky) and the complex in the reaction mixture: (See
Table S10 in the Supporting Information). Three solutions in CH2Cl2
were prepared having a concentration of 2.00ꢃ10ꢀ3 m for each compound
and their UV/Vis spectrum was recorded: 1) Ky, 2) Ky+TiCl4, 3) dimethyl
malonate+TiCl4+Et3N+benzaldehyde.
Job plots: Several solutions of host and guest compounds were prepared,
at a total concentration of 8.21ꢃ10ꢀ1 m in CDCl3 and with host molar
ratio varying from 0.02 to 1. 1H NMR spectra of the mixtures were re-
corded and the chemical shifts were analysed by the Job method,[19]
modified for NMR spectroscopic data.
Test with AgSbF6: AgSbF6 (182.7 mg, 5.32ꢃ10ꢀ4 mol, 1.7 equiv) was
added to a solution of the titanium enolate of diethyl malonate in
CD2Cl2 previously obtained. 1H and 13C NMR spectra were acquired
before and after the addition of AgSbF6.
Acknowledgements
Dimethyl benzylidenemalonate (Ky):[43] Dimethyl malonate (8.76 mmol;
1.0 equiv) was added to a solution of TiCl4 (1.0 equiv) in dry dichlorome-
thane (20 mL) at 08C while stirring under nitrogen atmosphere, in the
presence of 3 ꢄ molecular sieves. After 30 min, triethylamine was added
(1.0 equiv) to the resulting yellow suspension and the mixture turned im-
mediately dark red. Benzaldehyde (1.0 equiv) was added and the mixture
was stirred for 3 h at 08C. On addition of a 1m HCl aqueous solution
(20 mL), two phases formed. The aqueous layer was separated, the or-
ganic phase was dried over MgSO4 and evaporated to dryness under re-
duced pressure. The crude residue was purified by flash chromatography
This work was carried out with the support from the University of Chieti.
[1] R. Mahrwald, D. A. Evans, Modern Aldol Reactions,Wiley-VCH,
Weinheim, 2004.
[2] a) G. Casiraghi, F. Zanardi, G. Appendino, G. Rassu, Chem. Rev.
2000, 100, 1929–1972; b) T. D. Machajewski, C. H. Wong, Angew.
Chem. 2000, 112, 1406–1430; Angew. Chem. Int. Ed. 2000, 39, 1352–
1374; c) R. Mahrwald, Curr. Org. Chem. 2003, 7, 1713–1723; d) P.
Arya, H. Qin, Tetrahedron 2000, 56, 917–947.
11548
www.chemeurj.org
ꢂ 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2009, 15, 11537 – 11550

b-Diester Titanium Enolates
FULL PAPER
[3] a) C. H. Heathcock, Science 1981, 214, 395–400; b) D. A. Evans,
J. M. Takacs, L. R. McGee, M. D. Ennis, D. J. Mathre, J. Bartroli,
Pure Appl. Chem. 1981, 53, 1109–1127; c) D. A. Evans, J. V. Nelson,
T. R. Taber, Top. Stereochem. 1982, 13, 1–115; d) S. Masamune, W.
Choy, J. S. Petersen, L. R. Sita, Angew. Chem. 1985, 97, 1–31;
Angew. Chem. Int. Ed. Engl. 1985, 24, 1–31; e) M. Braun, H. Sacha,
J. Prakt. Chem. 1993, 335, 653–668; f) Modern Synthetic Methods
(Ed.: R. Scheffold), Thieme, Stuttgart, 1992.
[13] a) J. Zhu, H. Bienaymꢁ, Multicomponent Reactions, Wiley-VCH,
Weinheim, 2005; b) L. Weber, Curr. Med. Chem. 2002, 9, 1241–
1253; c) A. Ulaczyk-Lesanko, D. G. Hall, Curr. Opin. Chem. Biol.
2005, 9, 266–276; d) R. W. Armstrong, A. P. Combs, P. A. Tempest,
S. D. Brown, T. A. Keating, Acc. Chem. Res. 1996, 29, 123–131; e) I.
Ugi, Acta Polytech. Scand. Chem. 1997, 244, 64–66; f) M. A. Miro-
nov, QSAR Comb. Sci. 2006, 25, 423–431.
[14] a) W. Lehnert, Tetrahedron Lett. 1970, 11, 4723–4724; b) W. Leh-
nert, Tetrahedron 1972, 28, 663–666; c) W. Lehnert, Tetrahedron
1973, 29, 635–638; d) W. Lehnert, Tetrahedron 1974, 30, 301–305;
e) W. Lehnert, Synthesis 1974, 667–669.
[4] a) S. Saito, H. Yamamoto, Acc. Chem. Res. 2004, 37, 570–579; b) H.
Yamamoto, K. Futatsugi, Angew. Chem. 2005, 117, 1958–1977;
Angew. Chem. Int. Ed. 2005, 44, 1924–1942.
[5] a) I. Marek, Titanium and Zirconium in Organic Synthesis, Wiley-
VCH, Weinheim, 2002; b) M. Schlosser, Organometallics in Synthe-
sis Wiley, New York, 2002; c) M. T. Reetz, J. Westermann, R. Stein-
bach, B. Wenderoth, R. Peter, R. Ostarek, S. Maus, Chem. Ber.
1985, 118, 1421–1440; d) T. Mukaiyama, Angew. Chem. 1977, 89,
858–866; Angew. Chem. Int. Ed. Engl. 1977, 16, 817–826; e) A.
[15] a) W. Oppolzer, I. Rodriguez, Helv. Chim. Acta 1993, 76, 1275–
1281; b) G. Cardillo, L. Gentilucci, M. Gianotti, R. Perciaccante, A.
Tolomelli, J. Org. Chem. 2001, 66, 8657–8660.
[16] O-Titanation has also been observed in the titanium(IV) enolates of
some alkyl phosphonoacetates. See: M. T. Reetz, M. von Itzstein, J.
Organomet. Chem. 1987, 334, 85–90.
´
Fꢅrstner, B. Bogdanovic, Angew. Chem. 1996, 108, 2582–2609;
[17] M. T. Reetz, R. Peter, M. von Itzstein, Chem. Ber. 1987, 120, 121–
122.
[18] T. Rossi, C. Marchioro, A. Paio, R. J. Thomas, P. Zarantonello, J.
Org. Chem. 1997, 62, 1653–1661.
[19] a) P. Job, Ann. Chim. 1928, 9, 113–203; b) R. Sahai, G. L. Loper,
S. H. Lin, H. Eyring, Proc. Natl. Acad. Sci. USA 1974, 71, 1499–
1503; c) V. M. S. Gil, N. C. Oliveira, J. Chem. Educ. 1990, 67, 473–
478; d) Z. D. Hill, P. MacCarty, J. Chem. Educ. 1986, 63, 162–167.
[20] This result is valid only at the concentration used in the experiment.
At other concentrations, other species could be predominant. See:
a) R. W. Ramette, J. Chem. Educ. 1963, 40, 71–72; b) H. E. Bent,
C. L. French, J. Am. Chem. Soc. 1941, 63, 568–572.
Angew. Chem. Int. Ed. Engl. 1996, 35, 2442–2469; f) B. Weidmann,
D. Seebach, Angew. Chem. 1983, 95, 12–26; Angew. Chem. Int. Ed.
Engl. 1983, 22, 31–45; g) S. Sana, Synlett 2002, 0364–0365; h) R.
Mahrwald, J. Prakt. Chem. 1999, 341, 595–599; i) R. Mahrwald, B.
Gꢅndogan, J. Am. Chem. Soc. 1998, 120, 413–414.
[6] a) D. A. Evans, D. L. Rieger, M. T. Bilodeau, F. Urpi, J. Am. Chem.
Soc. 1991, 113, 1047–1049; b) R. Annunziata, M. Cinquini, F. Cozzi,
P. G. Cozzi, E. Consolandi, Tetrahedron 1991, 47, 7897–7910;
c) D. A. Evans, F. Urpi, T. C. Somers, J. S. Clark, M. T. Bilodeau, J.
Am. Chem. Soc. 1990, 112, 8215–8216; d) M. A. Walker, C. H.
Heathcock, J. Org. Chem. 1991, 56, 5747–5750; e) C. R. Harrison,
Tetrahedron Lett. 1987, 28, 4135–4138; f) R. Annunziata, M. Cinqui-
ni, F. Cozzi, A. Lombardi Borgia, J. Org. Chem. 1992, 57, 6339–
6342.
[7] a) Y. Tanabe, A. Makita, S. Funakoshi, R. Hamasaki, T. Kawakusu,
Adv. Synth. Catal. 2002, 344, 507–510; b) Y. Tanabe, N. Matsumoto,
S. Funakoshi, N. Manta, Synlett 2001, 1959–1961; c) A. K. Ghosh, J.-
H. Kim, Tetrahedron Lett. 2002, 43, 5621–5624; d) Y. Tanabe, N.
Manta, R. Nagase, T. Misaki, Y. Nishii, M. Sunagawa, A. Sasaki,
Adv. Synth. Catal. 2003, 345, 967–970.
[21] L. Fielding, Tetrahedron 2000, 56, 6151–6170.
[22] a) P. Sobota, S. Szafert, T. Lis, J. Organomet. Chem. 1993, 443, 85–
91; b) H. J. Kakkonen, J. Pursiainen, T. A. Pakkanen, M. Ahlgrꢁn, J.
Organomet. Chem. 1993, 453, 175–184.
[23] L. Cavallo, S. Del Piero, J.-M. Ducꢁrꢁ, R. Fedele, A. Melchior, G.
Morini, F. Piemontesi, M. Tolazzi, J. Phys. Chem. C 2007, 111, 4412–
4419.
[24] A. Clerici, N. Pastori, O. Porta, Tetrahedron Lett. 2004, 45, 1825–
1827.
[8] a) M. T. Reetz, R. Peter, Tetrahedron Lett. 1981, 22, 4691–4694;
b) C. Siegel, E. R. Thornton, J. Am. Chem. Soc. 1989, 111, 5722–
5728; c) M. Riediker, R. O. Duthaler, Angew. Chem. 1989, 101, 488–
490; Angew. Chem. Int. Ed. Engl. 1989, 28, 494–495; d) R. O.
Duthaler, P. Herold, W. Lottenbach, K. Oertle, M. Riediker, Angew.
Chem. 1989, 101, 490–491; Angew. Chem. Int. Ed. Engl. 1989, 28,
495–497.
[25] a) F. A. Cotton, G. Wilkinson, C. A. Murillo, M. Bochmann, Ad-
vanced Inorganic Chemistry, Wiley, New York, 1999; b) N. N. Green-
wood, A. Earnshow, Chemistry of the Elements, Elsevier, Amster-
dam, 2008.
[26] a) A. Z. Von Feltz, Z. Anorg. Allg. Chem. 1965, 338, 147–154;
b) A. Z. Von Feltz, Z. Anorg. Allg. Chem. 1965, 338, 155–162;
c) A. M. Stolzenberg, G. S. Haymond, Inorg. Chem. 2002, 41, 300–
308; d) G. D. Stucky, M. L. F. Phillips, T. E. Gier, Chem. Mater. 1989,
1, 492–509.
[9] E. Nakamura, J. Shimada, Y. Horiguchi, I. Kuwajima, Tetrahedron
Lett. 1983, 24, 3341–3342.
[10] a) S. H. Bertz, G. Dabbagh, C. P. Gibson, Organometallics 1988, 7,
563–565; b) J. R. Stille, R. H. Grubbs, J. Am. Chem. Soc. 1983, 105,
1664–1665.
[27] H. E. Zimmerman, M. D. Traxler, J. Am. Chem. Soc. 1957, 79, 1920–
1923.
[28] a) J. W. Kabalka, Y. Ju, Z. Wu, J. Org. Chem. 2003, 68, 7915–7917;
b) Z. Li, W.-H. Sun, X. Jin, C. Shao, Synlett 2001, 1947–1949.
[29] Jaguar, version 6.0, Schrçdinger, LLC. New York, 2005.
[30] A. D. Becke, J. Chem. Phys. 1993, 98, 5648–5652.
[31] C. T. Lee, W. T. Yang, R. G. Parr, Phys. Rev. 1988, 37, 785–789.
[32] P. J. Hay, W. R. Wadt, J. Chem. Phys. 1985, 82, 270–283.
[33] W. R. Wadt, P. J. Hay, J. Chem. Phys. 1985, 82, 284–298.
[34] P. J. Hay, W. R. Wadt, J. Chem. Phys. 1985, 82, 299–310.
[35] C. Gonzalez, H. B. Schegel, J. Chem. Phys. 1989, 90, 2154–2161.
[36] C. Gonzalez, H. B. Schegel, J. Chem. Phys. 1990, 92, 5523–5527.
[37] J. Tomasi, M. Persico, Chem. Rev. 1994, 94, 2027–2094.
[38] C. J. Cramer, D. G. Truhlar, Chem. Rev. 1999, 99, 2161–2200.
[39] D. J. Tannor, B. Marten, R. B. Murphy, R. A. Friesner, D. Sitkoff, A.
Nicholls, M. N. Ringnalda, ; W. A. Goddard III, B. Honig, J. Am.
Chem. Soc. 1994, 116, 11875–11882; W. A. Goddard III, B. Honig, J.
Am. Chem. Soc. 1994, 116, 11875–11882.
[11] a) T.-H. Yan, C.-W. Tan, H.-C. Lee, H.-C. Lo, T.-Y. Huang, J. Am.
Chem. Soc. 1993, 115, 2613–2621; b) Y. Yoshida, R. Hayashi, H. Su-
mihara, Y. Tanabe, Tetrahedron Lett. 1997, 38, 8727–8730; c) R.
Mahrwald, B. Costisella, B. Gꢅndogan, Tetrahedron Lett. 1997, 38,
4543–4544; d) G. Koch, O. Loiseleur, D. Fuentes, A. Jantsch, K.-H.
Altmann, Org. Lett. 2002, 4, 3811–3814; e) S. Figueras, R. Martꢆn, P.
Romea, F. Urpꢆ, J. Vilarrasa, Tetrahedron Lett. 1997, 38, 1637–1640;
f) R. Mahrwald, B. Costisella, B. Gꢅndogan, Synthesis 1998, 262–
264; g) D. A. Evans, J. S. Clark, R. Metternich, V. J. Novack, G. S.
Sheppard, J. Am. Chem. Soc. 1990, 112, 866–868; h) M. T. Crim-
mins, B. W. King, E. Tabet, J. Am. Chem. Soc. 1997, 119, 7883–
7884; i) M. T. Crimmins, B. W. King, E. A. Tabet, K. Chaudhary, J.
Org. Chem. 2001, 66, 894–902; j) V. A. Brunet, D. OꢇHagan,
A. M. Z. Slawin, J. Fluorine Chem. 2007, 128, 1271–1279; k) R.
Nagase, N. Matsumoto, K. Hosomi, T. Higashi, S. Funakoshi, T.
Misaki, Y. Tanabe, Org. Biomol. Chem. 2007, 5, 151–159.
[40] B. Marten, K. Kim, C. Cortis, R. A. Friesner, R. B. Murphy, M. N.
Ringnalda, D. Sitkoff, B. Honig, J. Phys. Chem. 1996, 100, 11775–
11788.
[12] A. Renzetti, E. Dardennes, A. Fontana, P. De Maria, J. Sapi, S.
Gꢁrard, J. Org. Chem. 2008, 73, 6824–6827.
Chem. Eur. J. 2009, 15, 11537 – 11550
ꢂ 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
11549

A. Fontana, N. Re et al.
[41] Gaussian 03, Revision C.02, M. J. Frisch, G. W. Trucks, H. B. Schle-
gel, G. E. Scuseria, M. A. Robb, J. R. Cheeseman, Jr. J. A. Montgo-
mery, Jr., T. Vreven, K. N. Kudin, J. C. Burant, J. M. Millam, S. S.
Iyengar, J. Tomasi, V. Barone, B. Mennucci, M. Cossi, G. Scalmani,
N. Rega, G. A. Petersson, H. Nakatsuji, M. Hada, M. Ehara, K.
Toyota, R. Fukuda, J. Hasegawa, M. Ishida, T. Nakajima, Y. Honda,
O. Kitao, H. Nakai, M. Klene, X. Li, J. E. Knox, H. P. Hratchian,
J. B. Cross, V. Bakken, C. Adamo, J. Jaramillo, R. Gomperts, R. E.
Stratmann, O. Yazyev, A. J. Austin, R. Cammi, C. Pomelli, J. W.
Ochterski, P. Y. Ayala, K. Morokuma, G. A. Voth, P. Salvador, J. J.
Dannenberg, V. G. Zakrzewski, S. Dapprich, A. D. Daniels, M. C.
Strain, O. Farkas, D. K. Malick, A. D. Rabuck, K. Raghavachari,
J. B. Foresman, J. V. Ortiz, Q. Cui, A. G. Baboul, S. Clifford, J. Cio-
slowski, B. B. Stefanov, G. Liu, A. Liashenko, P. Piskorz, I. Komaro-
mi, R. L. Martin, D. J. Fox, T. Keith, M. A. Al-Laham, C. Y. Peng,
A. Nanayakkara, M. Challacombe, P. M. W. Gill, B. Johnson, W.
Chen, M. W. Wong, C. Gonzalez, J. A. Pople, Gaussian, Inc., Wall-
ingford CT, 2004.
[42] Y. Cao, M. D. Beachy, D. A. Braden, L. A. Morrill, M. N. Ringnalda,
R. A. Friesner, J. Chem. Phys. 2005, 122, 224116–224126.
[43] R. M. Wilson, A. Hengge, J. Org. Chem. 1987, 52, 2699–2707.
[44] D. A. White, J. Chem. Soc. Perkin Trans. 1 1976, 1926–1930.
Received: June 11, 2009
Published online: September 23, 2009
11550
www.chemeurj.org
ꢂ 2009 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2009, 15, 11537 – 11550