Fluorine... and π... alkali metal interactions control in the stereoselective amide enolate alkylation with fluorinated oxazolidines (fox) as a chiral auxiliary: An experimental and theoretical study

Article abstract of DOI:10.1002/chem.200701604

The α-alkylation of amide enolates by using a pseudo-C₂ symmetry trans 4-phenyl-2-trifluoromethyloxazolidine (trans-Fox) as a chiral auxiliary occurs with an extremely high diastereoselectivity (>99% de). The origin of this excellent stereocontrol was investigated by an experimental and theoretical (DFT) study. With this trans chiral auxiliary, both F...metal and π...metal interactions compete to give the same diastereomer through Re face alkylation of the enolate. A 5.5 kcal mol^-1 energy difference found between the Re face and the Si face attack transition states is consistent with the complete diastereoselectivity that has been experimentally achieved. On the other hand, in the case of the cis chiral auxiliary (cis-Fox) the competition between the F...metal and π...metal interactions is unfavourable to the diastereoselectivity. In this case, the Re face and the Si face attack transition states were found to be nearly isoenergetic (0.3 kcal mol^-1 difference), which is in good agreement with the very low diastereoselectivity observed.

Full text of DOI:10.1002/chem.200701604

FULL PAPER
DOI: 10.1002/chem.200701604
Fluorine··· and p···Alkali Metal Interactions Control in the Stereoselective
Amide Enolate Alkylation with Fluorinated Oxazolidines (Fox) as a Chiral
Auxiliary: An Experimental and ACHTRETUNG heoretical Study
Gjergji Sini,*[a] Arnaud Tessier,[b] Julien Pytkowicz,[b] and Thierry Brigaud*[b]
Abstract: The a-alkylation of amide
enolates by using a pseudo-C2 symme-
try trans 4-phenyl-2-trifluoromethylox-
azolidine (trans-Fox) as a chiral auxili-
ary occurs with an extremely high dia-
stereoselectivity (>99% de). The
origin of this excellent stereocontrol
was investigated by an experimental
and theoretical (DFT) study. With this
trans chiral auxiliary, both F···metal
and p···metal interactions compete to
give the same diastereomer through Re
face alkylation of the enolate.
A
On the other hand, in the case of the
cis chiral auxiliary (cis-Fox) the compe-
tition between the F···metal and
p···metal interactions is unfavourable
to the diastereoselectivity. In this case,
the Re face and the Si face attack tran-
sition states were found to be nearly
isoenergetic (0.3 kcalmolÀ1 difference),
which is in good agreement with the
very low diastereoselectivity observed.
5.5 kcalmolÀ1 energy difference found
between the Re face and the Si face
attack transition states is consistent
with the complete diastereoselectivity
that has been experimentally achieved.
Keywords: alkali metals · asymmet-
ric synthesis · density functional
calculations · fluorine · transition
states
Introduction
auxiliaries for the stereoselective alkylation of amide eno-
lates.[4] The trans 4-phenyl-2-trifluoromethyloxazolidine
(trans-Fox) proved to be an outstanding chiral auxiliary for
the synthesis of enantiopure a-substituted aldehydes, car-
boxylic acids and b-substituted alcohols with a very efficient
recovery of the chiral auxiliary (Scheme 1).
The use of fluorinated nitrogen heterocycles as chiral aux-
iliaries is mostly documented in the literature with highly
fluorinated compounds for asymmetric fluorous catalysis,[5]
Owing to the specific properties of the fluorine atom, orga-
nofluorine compounds are finding an increasing interest
mainly in the area of pharmaceuticals, agrochemicals and
material sciences.[1] For these developments, the discovery of
new chiral fluorine-containing synthons plays a crucial
role.[2] During the ten last years, fluoral-based trifluoro-
ACHTREUNGmethylated chiral oxazolidines (Fox) have emerged as pow-
erful building blocks for the stereoselective synthesis of vari-
ous chiral trifluoromethyl amino compounds.[3] In addition,
we recently reported their use as high-performance chiral
[a] Dr. G. Sini
Laboratoire LPPI, UniversitØ de Cergy-Pontoise
5 Mail Gay-Lussac, Neuville sur Oise
95031 Cergy-Pontoise cedex(France)
Fax: ( +33)134-257-071
E-mail: Gjergji.Sini@u-cergy.fr
[b] Dr. A. Tessier, Dr. J. Pytkowicz, Prof. Dr. T. Brigaud
Laboratoire “Synth›se Organique SØlective et
Chimie OrganomØtallique“ (SOSCO), UMR CNRS 8123
UniversitØ de Cergy-Pontoise, 5 Mail Gay-Lussac
Neuville sur Oise, 95031 Cergy-Pontoise cedex(France)
Fax: ( +33)134-257-071
E-mail: thierry.brigaud@u-cergy.fr
Supporting information for this article is available on the WWW
under http://www.chemeurj.org/ or from the author.
Scheme 1. Synthetic applications of trans 4-phenyl-2-trifluoromethyloxa-
zolidine (trans-Fox) chiral auxiliary.
Chem. Eur. J. 2008, 14, 3363 – 3370
ꢀ 2008 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
3363

but the fluorine-containing group is rarely at the chiral
centre. The use of fluorinated oxazolidinones with a per-
fluoroalkyl chain[6] or a difluorobenzyl group[7] as chiral aux-
iliaries was recently reported. In the first case, the role of
the perfluoroalkyl group is mainly to facilitate the chiral
auxiliary recovery and in both cases the stereoselectivity
does not seem to be oriented by interactions between the
fluorine atoms and metals.
Fluorine···metal interactions, in a larger sense, have al-
ready been reported in the literature.[8] As they have been
frequently evocated to explain specific reactivity and unusu-
al diastereoselectivities of reactions involving metal contain-
ing fluorinated compounds,[9] we suspected the crucial role
of such interactions to explain the extremely high diastereo-
selectivity of our reaction. To our knowledge, this would be
the first application of the fluorine···metal interaction con-
trol with a fluorinated chiral
Results and Discussion
Synthesis: The same high level of diastereoselectivity was
achieved with Li, K and Na enolates (Table 1, entries 1–3).
The alkylation of trans-1 was completely diastereoselec-
tive with benzyl bromide and with ethyl iodide (entry 4).
Therefore the calculations were done with the less complex
ethyl iodide. Moreover, we showed that the diastereoselec-
tivity was not affected by the nature of the solvent. The
same excellent reactivity and diastereoselectivity was ach-
ieved in different solvents such as THF and toluene (en-
tries 3–5). However the benzylation reaction failed in
hexane because of the very low solubility of the enolate in
this non polar solvent. Moreover, the addition of 1,3-dime-
thyltetrahydro-2-pyrimidinone (DMPU) or N,N,N’,N’-tetra-
methylethylenediamine (TMEDA) may be expected to min-
auxiliary in asymmetric synthe-
sis.
The study of the origin of the
Table 1. Diastereoselective alkylation reactions of N-acyl oxazolidine trans-1.
p-facial diastereoselection in
enolate alkylation reactions is a
current topic in organic synthe-
sis. Its rationalization has been
approached in the literature by
both experimental and theoreti-
cal studies.[10] Therefore, with
our chiral fluorinated oxazoli-
dines (Fox) in hand, we carried
out a theoretical study to gain
more insight into the origin of
the extremely high diastereose-
lectivity of the corresponding
amide enolates alkylation.
Entry
Base
LiHMDS[b]
KHMDS
NaHMDS
NaHMDS
NaHMDS
NaHMDS
NaHMDS
LiHMDS
Solvent
RX
Product Yield [%][a]
Ref
[c]
1
2
3
4
5
6
7
8
THF
THF
BnBr
BnBr
BnBr
EtI
BnBr
BnBr
BnBr
BnBr
(S)-2a (82)
(S)-2a (63)
(S)-2a (85)
(S)-2b (78)
(S)-2a (58)
(S)-2a (77)
(S)-2a (90)
(S)-2a (87)
[c]
[c]
THF
THF[d,e]
[c,f]
[f]
toluene
THF/DMPU[g]
THF/TMEDA[h]
THF/TMEDA[i]
[f]
[f]
[f]
[a] Yield of isolated product. [b] HMDS: hexamethyldisilazane. [c] Ref. [4]; [d] The reaction in toluene gave
(S)-2b as a single diastereomer. [e] No reaction occurred in hexane. [f] This study. [g] THF/DMPU: 4:1.
[h] THF/TMEDA: 4:1. [i] THF/TMEDA: 3:2.
imise coordination of the alkali metal with electron rich
groups such as the enolate, the fluorine atoms or the phenyl
group and to cause a decrease of the diastereoselectivity. In
fact, intriguingly, this made no change and the diastereose-
lectivity remained unaffected (entries 6–8). This suggests
that the geometry of the alkylation transition structures in-
volves strong molecular interactions, which are not distur-
bed by the nature of the solvent.
Abstract in French: Lꢀauxiliaire chiral fluorØ de pseudosymØ-
trie C2 trans-Fox (4-phenyl-2-trifluoromethyloxazolidine)
permet lꢀa-alkylation des Ønolates dꢀamides avec une diastØ-
rØosØlectivitØ exceptionnellement haute (>99% de). Lꢀorigine
de cet excellent stØrØocontrôle a ØtØ ØtudiØe de faÅon expØri-
mentale et thØorique (DFT). Avec lꢀauxiliaire chiral trans-
Fox, des interactions F···mØtal et p···mØtal sont en compØtition
mais conduisent au mÞme diastØrØoisom›re via lꢀalkylation
de lꢀØnolate par la face Re. Une diffØrence dꢀØnergie de
5,5 kcalmolÀ1 entre les Øtats de transitions correspondant à
lꢀattaque face Re et à lꢀattaque face Si est cohØrente avec la
totale diastØrØosØlectivitØ observØe expØrimentalement. Par
As we have previously reported,[4] the benzylation of the
cis-1 oxazolidine amide enolate occurred with a very low
diastereoselectivity (Scheme 2). In order to have an experi-
mental reference for the theoretical study, the alkylation re-
contre, avec lꢀauxiliaire chiral de configuration cis ACTHRE(UNG cis-Fox),
la compØtition entre les interactions F···mØtal et p···mØtal est
nØfaste à la diastØrØosØlectivitØ. Dans ce cas, la diffØrence dꢀØ-
nergie entre les Øtats de transitions correspondant aux atta-
ques face Re et face Si est tr›s faible (0,3 kcalmolÀ1) expli-
quant la tr›s faible diastØrØosØlectivitØ expØrimentalement ob-
servØe.
Scheme 2. Diastereoselective alkylation reactions of N-acyl oxazolidine
cis-1.
3364
www.chemeurj.org
ꢀ 2008 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2008, 14, 3363 – 3370

Stereoselectivity Control in Amide Enolate Alkylation
FULL PAPER
action of cis-1 was repeated with ethyl iodide. This reaction
on the same Re face. Another enolate (trans-E3) presenting
an F···Na interaction with the sodium on the Si face of the
enolate is 1.9 kcalmolÀ1 higher in energy. The geometrical
parameters of trans-E1 and trans-E2 and the strongly posi-
tive Na charges clearly suggest strong electrostatic type
Na···CF3 and Na···p(Ph) interactions.[16] It could be conclud-
ed at this stage that the two interactions appear to be of
comparable strength. These two interactions force the trans-
E1 and the trans-E2 enolate plane to preferentially adopt a
nearly coplanar position with respect to the oxazolidine
mean plane. Considering for example the trans-E1 case
(Figure 1), it is clear that the presence of the sodium atom
close to the Re face of the enolate will subsequently orien-
tate the ethyl iodide approach by the same face through
I···Na interaction giving rise to the weakly bonded complex
trans-RC1 (Figure 1).
This last complexis stabilized by both the Na···I interac-
tion and a hydrogen type interaction between ethyl iodide
and the enolate p-system. The role of the Na···CF3 interac-
tion seems to be twofold: rigidifying the structure of the
enolate and orientating the ethyl iodide approach. The same
could be said for the Na···p(Ph) interaction in the case of
the trans-E2 enolate.
To compare different reaction profiles three possible tran-
sition structures were located: trans-TS1 (F···Na interaction,
Re face attack), trans-TS2 (p···Na interaction, Re face
attack) and trans-TS3 (F···Na interaction, Si face attack)
(Figure 1).[17] The trans-TS1 and trans-TS2 structures corre-
sponding to the Re-face alkylation of the enolate give the
same (S)-2b product (Figure 1). The results show that trans-
TS2 is 2.4 kcalmolÀ1 higher than trans-TS1, hence only the
reaction profiles including trans-TS1 and trans-TS3 will be
compared hereafter.
also proceeded with
a very poor diastereoselectivity
(Scheme 2). The theoretical study of this reaction was also
carried out, to give an explanation of this low diastereose-
lectivity and to understand the differences with the trans-1
case.
The contrasting results obtained from the trans-Fox and
the cis-Fox chiral auxiliaries are apparently related to the
trans or cis relative configurations of the CF3 and the phenyl
group. Each of them can play a double role: a steric one
and a metal-attractive one. The steric role of the phenyl
group is well known and largely used to explain stereoselec-
tive and chemoselective effects. Moreover, numerous experi-
mental and theoretical studies provide evidence for the pres-
ence in many chemical structures of cation···p interactions
of alkali metals with arenes.[11,9a] Most of them underline the
presence of this type of interactions in the gas phase and
solid-state structures, as well as in biological systems. To our
knowledge, their implication in transition-state structures
has never been clearly demonstrated. On the other hand,
the trifluoromethyl group has been rarely used in the design
of chiral auxiliaries. Although the steric hindrance of the tri-
fluoromethyl group is well known,[12] CF3···alkali metal inter-
actions have been rarely reported. Only a few theoretical
studies show that metal···F close contacts are able to provide
rigidification of reaction intermediates and to stabilize spe-
cific transition states.[9a,h] Consequently, it could be suspected
that Na···F and Na···p(Ph) interactions and their potential
competition played a central role in our reactions. To our
knowledge, these interactions have not yet been exploited in
the chiral auxiliary strategy for the synthesis of enantiopure
compounds.
In the following theoretical study we examine the exact
role of the trifluoromethyl and the phenyl groups and we
propose an explanation of their contribution to the diaste-
reoselectivity of the reaction using the trans- and the cis-Fox
chiral auxiliaries. To shed light on these questions, density
functional calculations (DFT) were carried out by using
The reaction free energy profiles corresponding to the
ethylation of the Re or the Si enolate faces giving (S)-2b or
(R)-2b are shown in Figure 2.[18]
An energy comparison between the trans-TS1 and the
trans-TS3 transition states shows that the trans-TS1 is the
most stable one by 5.5 kcalmolÀ1, which is in very good
agreement with the high diastereoselectivity experimentally
observed for this reaction (>99% de).[19] This result is also
coherent with the simple idea that the best direction for the
SN2 attack in these reactions corresponds to the enolate
face on which the sodium atom is located. On the other
hand, the trans-TS2 transition structure presenting a p···Na
interaction is 2.4 kcalmolÀ1 higher in energy than trans-TS1.
Consequently, it should be supposed that the reaction path-
way giving rise to trans-TS2 also contributes to the result of
the reactions. Fortunately, this small energy difference has
no consequence on the stereochemical issue of the reaction.
Actually, because of the pseudo-C2 symmetry of the trans-
Fox chiral auxiliary, both transition states give the same (S)-
2b product. It can be concluded at this stage that the alkali
metal···F and the alkali metal···p(Ph) close contacts give rise
to competitive transition states so both interactions should
be considered to explain the diastereoselectivity of the reac-
tions.
GAUSSIAN98 program[13] at B3LYP/6-311+G
ACHTRE(UGN 2d,p)//
B3LYP/6-31G* level followed by polarisable conductor con-
tinuum model calculations (CPCM)[14] for the solvent effect
(see computational details for further information).
Trans-Fox DFT study: The trans-1 sodium enolates, some of
the weakly bonded enolate···ethyl iodide reactant complexes
(RC) formed during the ethyl iodide approach together with
the corresponding transition states (TS) were calculated and
are shown in Figure 1. The ethylation of the trans-1 enolate
giving 2b is discussed first. As it is commonly assumed for
amide enolates, the first step of this reaction is the deproto-
nation of trans-1 to give the resulting Z enolates in order to
minimize the A-1;3 interactions.[15] Among the different pos-
sible enolates, the trans-E1 and trans-E2 structures
(Figure 1) are the most stable, the latter is only
0.9 kcalmolÀ1 higher in energy. Trans-E1 presents an F···Na
interaction with the sodium on the Re face of the enolate
and trans-E2 presents a p···Na interaction with the sodium
Chem. Eur. J. 2008, 14, 3363 – 3370
ꢀ 2008 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
3365

T. Brigaud, G. Sini et al.
Figure 1. Optimised structures for trans-1 enolates ethylation. Distances are given in Š (*=mean value for Na···Ph). The relative free energies [kcal
molÀ1] are given in parentheses (§ =also contains the energy of the isolated C2H5I species).
trans case, three transition states noted cis-TS1, cis-TS2 and
cis-TS3 were calculated. The cis-TS1 transition state is
4.7 kcalmolÀ1 higher in energy than the others, so only cis-
TS2 and cis-TS3 should be considered for the explanation of
the poor diastereoselectivity of the reaction. The cis-TS2
transition state is stabilised by a Na···p(Ph) interaction and
corresponds to the Re-face ethylation (Figure 3), whereas
Figure 2. Relative free energies [kcalmolÀ1] corresponding to the ethyla-
tion reaction of enolates trans-E1 (Re-face alkylation) and trans-E3 (Si
face-alkylation).
Cis-Fox DFT study: This last conclusion furnishes an ex-
planation of the very low diastereoselectivity achieved in
the cis-1 enolate alkylation (Scheme 2). Actually, as in the
Figure 3. Optimised transition structures for the cis-1 enolates ethylation.
Distances are given in Š (*mean value for Na···Ph distance). The relative
free energies [kcalmolÀ1] are given in parentheses.
3366
www.chemeurj.org
ꢀ 2008 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2008, 14, 3363 – 3370

Stereoselectivity Control in Amide Enolate Alkylation
FULL PAPER
cis-TS3 is stabilised by a Na···CF3 interaction and corre-
sponds to the Si-face alkylation (Figure 3).
The key result is that cis-TS2 and cis-TS3 are nearly iso-
ACHTREUNG
energetic, the latter being only 0.3 kcalmolÀ1 lower in
energy (Figure 4). In other words, the transition states based
Scheme 3. Alkylation reactions of valinol-based N-acyl oxazolidines
trans-4 and cis-4.
to the isopropyl group do not provide a positive contribu-
tion to the diastereoselectivity.
Figure 4. Relative free energies [kcalmolÀ1] corresponding to the ethyla-
The contribution of the trifluoromethyl group was also es-
timated theoretically[21] in the case of a trans-oxazolidine
chiral auxiliary having a methyl group in place of the tri-
fluoromethyl one. The transition states corresponding to the
Re-face and Si-face attacks are in this case separated by
2.9 kcalmolÀ1 in favour of the Re-face attack (see Figure 2
in the Supporting Information for the corresponding struc-
tures). Compared to the 5.5 kcalmolÀ1 calculated for the
trans-Fox chiral auxiliary, this result suggests the important
role of the trifluoromethyl group in the control of the dia-
stereoselectivity.
tion reaction of cis-1.
on Na···CF3 and Na···p(Ph) interactions and giving rise to
(R)-2b and (S)-2b products respectively, are in competition.
A mixture of R and S products is therefore suggested by the
calculations, which is in good agreement with the nearly
equimolar diastereomeric mixture obtained experimentally.
Steric and electronic factors: It is worth highlighting at this
point that despite the EtI···Ph proximity the cis-TS3 transi-
tion state is found to be 1.2 kcalmolÀ1 lower than the trans-
TS1. From a sterical point of view, one should expect that
trans-TS1 would be more stable (Figure 5), suggesting, in
turn, that the steric hindrance of the phenyl group is not a
determining factor for these reactions.
Conclusion
Regrouping now the results of the alkylation reactions using
the trans- and the cis-Fox chiral auxiliaries we can say that,
as expected, the Na···CF3 and Na···p(Ph) interactions play a
central role in these reactions: rigidifying the reactant geo-
metries and orientating the ethyl iodide approach leading to
the corresponding precursor complexes and transition states.
Moreover, the disappointing experimental results obtained
with the valinol-based Fox and the theoretical results calcu-
lated with an unfluorinated oxazolidine suggest that both
the CF3 and the phenyl groups participate to achieve a good
diastereoselectivity. It follows that the relative structural po-
sition between the trifluoromethyl and the phenyl groups
becomes crucial for the diasteroselective control of these re-
actions. Two metal attractor groups disposed in a pseudo-C2
symmetry (trans-Fox) give rise to the same product with a
complete diastereoselectivity. On the other hand, when the
chiral auxiliary presents two badly disposed metal attractor
sites (pseudo-C2v symmetry), the diastereoselective control
of these reactions is ruined by competitive interactions.
Based on these conclusions, we are currently investigating
the design of novel fluorinated chiral auxiliaries.
Figure 5. Comparison of trans-TS1 and cis-TS3. Distances are given in Š.
To evaluate the contribution of the p coordination of the
phenyl group in the diastereoselectivity, the enolate alkyla-
tion reaction was experimentally performed with trans-4 and
cis-4
valinol-based
trifluoromethylated
oxazolidines
(Scheme 3).[20] Replacing the phenyl group by an isopropyl
group caused a decrease in the diastereoselectivity. This
result suggests that the electronic properties of the phenyl
group are playing an important role in the control of the
diastereoselectivity. Moreover, the steric effects associated
Chem. Eur. J. 2008, 14, 3363 – 3370
ꢀ 2008 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
3367

T. Brigaud, G. Sini et al.
À77.84 ppm (m, 3 F); MS: m/z (%): 301 (M+, 52), 273 (5), 217 (15), 203
Experimental Section
(27), 188 (56), 148 (36), 120 (31), 104 (24), 85 (51), 57 (100).
3b (minor diastereomer): GC: Rt =9.47 min; 1H NMR (250 MHz,
CDCl3): d=0.52 (m, 3H), 0.88 (t, 3H, 3J=7.4 Hz), 1.26–1.68 (m, 2H),
2.31 (m, 1H), 4.11 (m, 1H), 4.62 (m, 1H), 5.09 (m, 1H), 6.02 (m, 1H),
7.18–7.39 ppm (m, 5H); 19F NMR (235.35 MHz, CDCl3): d=À77.63 ppm
General information: Commercial reagents were purchased from Al-
drich, Acros or Avocado and used as received. Trifluoroacetaldehyde-
ethylhemiacetal was generously offered by Central Glass Company. All
alkylation reactions were performed under argon atmosphere with oven-
dried glassware fitted with rubber septa. Ether and THF were distilled
under nitrogen from sodium/benzophenone prior to use. Flash chroma-
tography was performed by using SDS 60A, (40–63 mm.) silica gel. Thin
layer chromatography was performed on precoated aluminium sheets
(MACHEREY-NAGEL ALUGRAM SIL/G 0.2 mm). They were visual-
ized under a 254 nm UV light. 1H NMR, 19F NMR and 13C NMR spectra
3
(d, 3 F, JH-F =5.4 Hz); MS: m/z (%): 301 (M+, 52), 273 (4), 217 (14), 203
(27), 188 (56), 148 (35), 120 (29), 104 (22), 85 (49), 57 (100).
Computational methods: The calculations in this study were carried out
at the DFT (B3LYP) and MP2 theory levels, using the GAUSSIAN98
program.[13] During the geometry optimisations the internal 6-31G* basis
set was used for all the atoms except iodine for which the Lanl2DZ ECP
and valence shell basis set was chosen[22a–c] augmented by one polariza-
tion function (z=0.289).[22d,e] The same basis set for I atom was used af-
terwards for higher level energy comparisons. To estimate the error in-
duced by this last limitation, test calculations were done at B3LYP/aug-
cc-pvdz//B3LYP/6-31G* level which resulted with a change of the energy
difference by only 0.2 kcalmolÀ1 (the aug-cc-pvdz-pp basis for I and aug-
cc-pvdz basis for Na were chosen from reference 22f).
were recorded by using
a
JEOL ECX-400 (1H: 400 MHz; 19F:
376.2 MHz; 13C: 100.5 MHz). Chemical shift values (d) are reported in
ppm downfield from Me4Si (d=0.0 ppm), C6F6 (d=À164.9 ppm) or
CDCl3 as internal standard (d=77.0 ppm). Data are reported as follows:
chemical shift (d=ppm), multiplicity (s=singlet, d=doublet, t=triplet,
m=multiplet), integration, coupling constant (Hz). GC and low resolu-
tion mass spectra were performed by means of a Hewlett Packard GC
6890 coupled with a Hewlett Packard MSD 5973 apparatus (capillary
column HP-5mS, 30 m”250 mm, He as vector gas, method: 708C during
2 min, 208C/min climb rate and 2508C during 15 min).
Geometries: Different conformations were considered for the oxazoli-
dine cycle (with H at the N atom instead of an acyle group) containing
the phenyl and CF3 groups in trans and cis positions. The most stable
conformers were used afterwards as starting geometries during the opti-
misations of enolates, reactant complexes and the transition states. All
the geometry optimisations were followed by frequency calculations
showing no imaginary frequencies for the minimum structures and only
one imaginary frequency for the transition state structures. Further analy-
sis of the vibrational modes associated to the imaginary frequencies
showed correspondence between the transition state structures and the
reactants (products).
Experimental procedures: Experimental procedures and analytical data
for oxazolidines trans-1, (S)-2a, (S)-2b, cis-1 and 3a were reported in our
previous communication.[4]
Procedure for the benzylation reaction of trans-1 in toluene: The oxazoli-
dine trans-1 (0.321 g, 1.19 mmol) was dissolved in toluene (10 mL) under
argon atmosphere. The solution was cooled down to À788C and
NaHMDS was added dropwise (1.12 mL, 2m in THF, 2.24 mmol). The re-
action mixture was stirred for 1.5 h at this temperature and the benzyl-
bromide (0.269 mL, 2.24 mmol) was added slowly. The reaction mixture
was stirred for 2 additional hours at À788C, quenched with a saturated
NH4Cl solution (15 mL) and extracted with dichloromethane (2x30 mL).
The combined organic layers were dried over MgSO4, evaporated under
reduced pressure and the resulting crude mixture was purified by filtra-
tion through a short pad of silica gel (cyclohexane/ethyl acetate: 90/10).
(S)-2a (0.247 g, 58%) was obtained as a single diastereomer. No traces
of the (R)-2a epimer were detected by GC. Spectral data of (S)-2a were
identical to those reported in our preceding report.[4]
Solvent effect: The influence of the solvent on the reactions (THF in this
case) was considered by carrying out single point energy calculations in
the frame of CPCM continuum model.[14] The PCM continuum model[23]
was also tested but the results shown in the Table 2 and discussed here-
Table 2. Energy differences [kcalmolÀ1] between some representative
[b]
transition states.[a] a=6-311+G
A
the experimen-
tal result in this case is indirectly deduced. The “0” value is based on the
fact that a mixture of 56% (S) and 44% (R) products was obtained. See
the text hereafter for more details.
In a similar manner, the benzylation of the sodium enolate of trans-1
(0.325 g, 1.19 mmol) in a THF/DMPU (10:1 mL) gave 2a (0.332 g, 77%)
as a single diastereomer. The benzylation of the sodium enolate of trans-
1 (0.325 g, 1.19 mmol) in a THF/TMEDA (8:2 mL) mixture gave 2a
(0.389 g, 90%) as a single diastereomer. The benzylation of the lithium
enolate of trans-1 (0.325 g, 1.19 mmol) in a THF/TMEDA (6:4 mL) mix-
ture gave 2a (0.375 g, 87%) as a single diastereomer. In any case, no
traces of the (R)-2a epimer were detected by GC of the corresponding
crude mixtures.
PCM
CPCM
B3LYP/a[a] MP2/a[a] B3LYP/a[a] MP2/a[a] Exp[b]
DEtrans-TS3–trans-TS1
DEtrans-TS2–trans-TS1
DEcis-TS2–cis-TS3
4.3
1.6
À2.0
7.1
0.1
À2.4
5.5
2.4
0.3
7.7
À0.1
À1.4
–
–
“0”
after suggest that the CPCM model gives in this case better agreement
with experimental result. The effect of CPCM calculations on the energy
differences between transition states is less than 1 kcalmolÀ1 which seems
coherent with the small dipole moments of these species. These results
are also in line with the identical experimental results obtained with the
same reactions carried out in solvents of different polarity (THF and tol-
uene).[24]
ACHTREUNG(2R,4R)-2-Trifluoromethyl-3-(-2-methylbutanoyl)-4-phenyloxazolidines
(3b): To a solution of amide cis-1 (0.360 g, 1.32 mmol) in THF (10 mL)
under argon atmosphere at À788C was added dropwise a solution of
NaHMDS (1.25 mL, 2m in THF, 2.5 mmol). The reaction mixture was
stirred for 2 h at À788C and ethyl iodide (0.2 mL, 2.5 mmol) was added.
The reaction mixture was stirred for 3 additional hours at À788C and the
reaction mixture was quenched with a saturated NH4Cl solution (15 mL)
extracted with diethyl ether (2”30 mL) and dichloromethane (30 mL).
The combined organic layers were dried over MgSO4, evaporated under
reduced pressure and the resulting crude mixture was purified by filtra-
tion through a short pad of silica gel (cyclohexane/ethyl acetate, 90:10).
3b (0.178 g, 45%) was isolated as a 56:44 mixture of diastereomers and a
fraction of unreacted starting material cis-1 (0.145 g, 36%) was recov-
ered.
B3LYP versus MP2 methods: Different studies have shown that DFT
methods in general, and B3LYP in particular, fail in determining some
thermochemistry parameters such as reaction enthalpies so MP2/6-31+
G* or MP2/6-31+G*//B3LYP/6-31+G* level studies are recommende-
d.[25a,b] It has also been shown that B3LYP fails to describe the M···p in-
teractions, although it does well in the case of electrostatic M+n···p inter-
actions.[25c] To test the reliability of our model chemistry we carried out
3b (major diastereomer): GC: Rt =9.52 min; 1H NMR (250 MHz,
CDCl3): d=0.88 (t, 3H, 3J=7.4 Hz), 1.09 (d, 3H, 3J=6.7 Hz), 1.26–1.68
(m, 2H), 2.31 (m, 1H), 4.11 (m, 1H), 4.62 (m, 1H), 5.09 (m, 1H), 6.02
(m, 1H), 7.18–7.39 ppm (m, 5H); 19F NMR (235.35 MHz, CDCl3): d=
some test calculations at B3LYP/6-311+G
ACHTRE(UNG 2d,p)//B3LYP/6–1G* and
MP2/6-311+G(2d,p)//B3LYP/6-31G* levels. The energy differences be-
AHCTREUNG
tween some representative reactants and transition states are shown in
Table 2.
3368
www.chemeurj.org
ꢀ 2008 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2008, 14, 3363 – 3370

Stereoselectivity Control in Amide Enolate Alkylation
FULL PAPER
The nature of these species is described in the paper and it is not impor-
tant to detail here. It could be simply said that trans-TS2 and cis-TS2 spe-
cies involve Na···p(Ph) interactions while the others contain Na···F inter-
actions. The results in the table indicate non negligible changes between
B3LYP versus MP2 methods. As it is shown in the paper these differen-
ces do not affect the chemical interpretations and the agreement with the
experimental results. However, these differences remain important for
discussing the competing physical phenomena inside these structures
(comparison between Na···p(Ph) and Na···F interactions). As DFT and
MP2 methods are known to respectively underestimate and overestimate
the electronic correlation experimental values were necessary for com-
parison. In the absence of other experimental thermochemistry parame-
ters, we considered that the mixture of 56%(S) versus 44%(R) products
obtained in the case of cis-1 giving 3 reactions (Scheme 2) could serve as
a good indication for comparing the corresponding transition states. Ac-
tually, in the case of this reaction only two transition states were found to
be in competition, each presenting only one competing conformational
isomer (other isomers differing on the oxazolidine cycle conformation
were found to be roughly 3 kcalmolÀ1 higher in energy). Therefore, in
view of the experimental result an energy difference of roughly 0 kcal
molÀ1 could be supposed to be found between the two competing transi-
tion states (cis-TS3 and cis-TS2 in this case). The results (Table 1, third
line) show that CPCM/B3LYP and CPCM/MP2 approaches describe this
reaction correctly. Both results seem to be different by less than 1 kcal
molÀ1 roughly from the supposed experimental value of 0 kcalmolÀ1, with
the (S)/(R) ratio being slightly underestimated by the B3LYP result and
slightly overestimated by the MP2 one. For the sake of smaller computa-
[1] a) R. D. Chambers, Fluorine in organic chemistry, Oxford, Blackwell
Boca Raton, CRC Press, 2004; b) T. Hiyama, Organofluorine Com-
pounds, Chemistry and applications, (Ed: H. Yamamoto), Springer-
Verlag, Berlin Heidelberg, 2000; c) P. Kirsch, Modern fluoroorganic
chemistry: synthesis, reactivity, applications, Wiley-VCH, Weinheim,
2004; d) Organofluorine Chemistry, Principles and Commercial Ap-
plications, (Eds R. E. Banks, B. E. Smart, J. C. Tatlow), Plenum:
New York, 1994.
[2] a) Fluorine-containing synthons, (Ed. V. A. Soloshonok), American
Chemical Society, Washington, DC, 2005; b) Enantiocontrolled Syn-
thesis of Fluoro-Organic Compounds, (Ed. V. A. Soloshonok), Wiley,
Chichester, UK, 1999; c) Asymmetric Fluoroorganic Chemistry: Syn-
thesis, Applications, and Future Directions, (Ed. P. V. Ramachan-
dran), ACS Symp. Ser. No. 746, American Chemical Society, Wash-
ington, DC, 2000; d) Organofluorine Chemistry (Ed. K. Uneyama),
Blackwell, 2006.
[3] a) A. Ishii, K. Higashiyama, K. Mikami, Synlett 1997, 1381–1382;
b) N. Lebouvier, C. Laroche, F. Huguenot, T. Brigaud, Tetrahedron
Lett. 2002, 43, 2827–2830; c) F. Gosselin, A. Roy, P. D. OꢁShea, C.
Cheng, R. P. Volante, Org. Lett. 2004, 6, 641–644; d) W. C. Black,
C. I. Bayly, D. E. Davis, S. Desmarais, J.-P. Falgueyret, S. Leger, C. S.
Li, F. MassØ, D. J. McKay, J. T. Palmer, M. D. Percival, J. Robichaud,
N. Tsou, R. Zamboni, Bioorg. Med. Chem. Lett. 2005, 15, 4741–
4744; e) T. Billard, B. R. Langlois, J. Org. Chem. 2002, 67, 997–
1000; f) J. Legros, F. Meyer, M. Coliboeuf, B. Crousse, D. Bonnet-
Delpon, J.-P BØguØ, J. Org. Chem. 2003, 68, 6444–6446; g) F. Hu-
guenot, T. Brigaud, J. Org. Chem. 2006, 71, 2159–2162; h) F. Hugue-
not, T. Brigaud, J. Org. Chem. 2006, 71, 7075–7078.
tional cost the B3LYP/6-311+GACHTREU(NG 2d,p)//B3LYP/6-31G* approach was
then adopted.
[4] A. Tessier, J. Pytkowicz, T. Brigaud, Angew. Chem. 2006, 118, 3759–
3763; Angew. Chem. Int. Ed. 2006, 45, 3677–3681.
Energies: The energy values given in the paper and used for the discus-
sions correspond to the free energies obtained at CPCM/B3LYP/6-311+
[5] F. Fache, New J. Chem. 2004, 28, 1277–1283.
GACHTREUNG(2d,p)//B3LYP/6-31G* level. The corrections for the free energies are
[6] a) J. E. Hein, P. G. Hultin, Synlett 2003, 635–638; b) J. E. Hein, J.
Zimmerman, M. P. Sibi, P. G. Hultin, Org. Lett. 2005, 7, 2755–2758;
c) J. E. Hein, P. G. Hultin, Tetrahedron: Asymmetry 2005, 16, 2341–
2347.
[7] Santos Fustero, J. Piera, J. F. Sanz-Cervera, P. Bello, N. Mateu, J.
Fluorine Chem. 2007, 128, 647–653.
[8] a) H. Takemura, S. Nakashima, N. Kon, M. Yasutake, T. Shinmyozu,
T. Inazu, J. Am. Chem. Soc. 2001, 123, 9293–9298; b) B. Goldfuss,
P. V. R. Schleyer, S. Handschuh, F. Hampel, J. Organomet. Chem.
1998, 552, 285–292; c) H. Plenio, Chem. Rev. 1997, 97, 3363–3384;
d) H. Plenio, ChemBioChem, 2004, 5, 650–655; e) ref[1b] pp. 129–
135; f) ref[2a] pp. 399–411; g) ref[2d] pp. 139–172.
[9] a) N. J. R. van Eikema Hommes, P. von R. Schleyer, Angew. Chem.
1992, 104, 768–771; Angew. Chem. Int. Ed. Engl. 1992, 31, 755–758;
b) Y. Morizawa, A. Yasuda, K. Uchida, Tetrahedron Lett. 1986, 27,
1833–1836; c) T. Hanamoto, T. Fuchikami, J. Org. Chem. 1990, 55,
4969–4971; d) N. Shinohara, J. Haga, T. Yamazaki, T. Kitazume, S.
Nakamura, J. Org. Chem. 1995, 60, 4363–4374; e) T. Yamazaki, M.
Ando, T. Kitazume, T. Kubota, M. Omura, Org. Lett. 1999, 1, 905–
908; f) K. Tamura, T. Yamazaki, T. Kitazume, T. Kubota, J. Fluorine
Chem. 2005, 126, 918–930; g) H. Ito, A. Saito, T. Taguchi, Tetrahe-
dron: Asymmetry 1998, 9, 1989–1994; h) Y. Itoh, M. Yamanaka, K.
Mikami, J. Am. Chem. Soc. 2004, 126, 13174–13175; i) M. Shimizu,
T. Fujimoto, H. Minezaki, T. Hata, T. Hiyama, J. Am. Chem. Soc.
2001, 123, 6947–6948; j) M. Shimizu, T. Fujimoto, L. Xinyu, H.
Minezaki, T. Hata, T. Hiyama, Tetrahedron 2003, 59, 9811–9823;
k) T. Ritter, M. W. Day, R. H. Grubbs, J. Am. Chem. Soc., 2006, 128,
11768–11769.
taken at the experimental temperature (195.15 K) and are scaled by a
factor of 0.9804 to take into account the systematic errors in the frequen-
cy values.
However, it has been evocated[26] that DS° values contain errors, owing
to the low frequency intermolecular vibrational modes and the neglect of
their anharmonic character. The energy differences between equivalent
species (between transition states or between enolates) vary by less than
1 kcalmolÀ1, owing to the entropic effects, meaning that the agreement
with the experimental results is not affected. On the other hand, the acti-
vation free energy values (DG° with respect to the reactants) are approx-
imately 8 kcalmolÀ1 greater than the corresponding DH°. The DG°
values could consequently contain errors, which we can not estimate
owing to a lack of experimental thermochemistry parameters.
Despite this, we prefer presenting the free energy values because of the
more realistic shape of the reaction profiles. Although the entropy prob-
lem does not modify the comparison between the competing transition
states and the corresponding conclusions, one should be prudent with the
absolute values of the activation parameters given in the Figure 2 and
Figure 4.
Finally, the basis set superposition error (BSSE) was calculated for all
the transition state structures and the reactant complexes by the counter-
poise approach.[27] The BSSE values were found to vary between 1.1–
1.2 kcalmolÀ1 and are included in the relative energy values presented
here.
[10] a) K. Ando, N. S. Green; Y. Li, K. N. Houk, J. Am. Chem. Soc. 1999,
121, 5334–5335; Y. Li, K. N. Houk, J. Am. Chem. Soc. 1999, 121,
5334–5335; b) K. Ando, J. Am. Chem. Soc. 2005, 127, 3964–3972;
c) Y. Itkuta, S. Tomoda, Org. Lett. 2004, 6, 189–192; d) A. I.
Meyers, M. A. Seefeld, B. A. Lefker, J. F. Blake, P. G. Williard, J.
Am. Chem. Soc. 1998, 120, 7429–7438; e) C. Aydillo, G. JimØnez-
OsØs, J. H. Busto, J. M. Peregrina, María M. Zurbano, A. Avenoza,
Chem. Eur. J. 2007, 13, 4840–4848; f) G. JimØnez-OsØs, C. Aydillo,
J. H. Busto, M. M. Zurbano, J. M. Peregrina, A. Avenoza, J. Org.
Acknowledgements
We thank the French ministry of research for awarding a research fellow-
ship to A. T., Central Glass Company for the generous gift of trifluoro-
ACHTREUNGacetaldehyde hemiacetal and financial support and the SIR of Cergy-
Pontoise University for the calculation facilities. G.S. thanks HØl›ne
GØrard (L.C.T., UniversitØ Paris 6, France) for valuable help for ELF
calculations and discussions.
Chem. Eur. J. 2008, 14, 3363 – 3370
ꢀ 2008 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
3369

T. Brigaud, G. Sini et al.
Chem. 2007, 72, 5399–5402; g) I. Soteras, O. Lozano, A. Gómez-
EsquØ, C. Escolano, M. Orozco, M. Amat, J. Bosch, F. J. Luque, J.
Am. Chem. Soc. 2006, 128, 6581–6588; h) S. D. Bull, S. G. Davies,
M.-S. Key, R. L. Nicholson, E. D. Savory, Chem. Commun. 2000,
1721–1722; i) S. D. Bull, S. G. Davies, A. C. Garner, D. Kruchinin,
M.-S. Key, P. M. Roberts, E. D. Savory, A. D. Smith, J. E. Thomson,
Org. Biomol. Chem. 2006, 4, 2945–2964; j) S. D. Bull, S. G. Davies,
A. C. Garner, A. L. Parkes, P. M. Roberts, T. G. R. Sellers, A. D.
Smith, J. A. Tamayo, J. E. Thomson, R. J. Vickers, New J. Chem.
2007, 31, 486–495; k) J. Mulzer, O. Langer, M. Hiersemann, J. W.
Bats, J. Buschmann, P. Luger, J. Org. Chem. 2000, 65, 6540–6546;
l) F. Andersson, E. Hedenstrçm, Tetrahedron: Asymmetry 2006, 17,
1952–1957.
[19] Different observations could help to understand this energy differ-
ence but it seems to be mainly controlled by the following: the
trans-TS3 structure shows that ethyl iodide approaches the enolate
face opposite to the CF3 group. This trans-TS3 transition state suf-
fers then from a weaker Na···F attractive interaction than trans-TS1
(dNa···F =2.367 Š for trans-TS3 versus 2.288 Š for trans-TS1,
Figure 1) and a stronger repulsive F···Oenolate interaction (dF···O
=
2.846 Š versus 2.974 Š respectively, Figure 1). Moreover, to opti-
mise the F···Oenolate and F···Na interactions the CF3 group in trans-
TS3 adopts a nearly eclipsed conformation with respect to the oxa-
zolidine cycle (the dihedral F-C-C-Ocycle angle is À24.18) as com-
pared to the unstrained stagered conformation for trans-TS1 (the F-
C-C-Ocycle angle is À60.98). We point out that this nearly eclipsed
conformation is absent in trans-E3 enolate. This unfavourable situa-
tion for trans-TS3 seems to be reinforced by the Et-I···Ph proximity.
As a conclusion, the trans-TS1 structure takes advantage of all these
electronic and steric factors.
[11] a) J. Sunner, K. Nishizawa, P. Kebarle, J. Phys.Chem 1981, 85, 1814–
1820; b) D. A. Dougherty, Science, 1996, 271, 163–168; c) J. Wout-
ers, Protein Sci. 1998, 7, 2472–2475; d) S. L. De Wall, L. J. Barbour,
G. W. Gokel, J. Am. Chem. Soc. 1999, 121, 8405–8406; e) V.
Ryzhov, R. C. Dunbar, J. Am. Chem. Soc. 1999, 121, 2259–2268;
f) R. C. Dunbar, J. Phys. Chem. A 2002, 106, 7328–7337; g) J. Siva-
guru, R. B. Sunoj, T. Wada, Y. Origane, Y. Inoue, V. Ramamurthy, J.
Org. Chem. 2004, 69, 6533–6547.
[20] trans-4 and cis-4 were obtained by our standard acylation proce-
dure[4] from the CF3-oxazolidines prepared from (S)-valinol and tri-
fluoroacetaldehyde methyl hemiacetal.[3]
[21] Because of the hydrolytical instability of the chiral auxiliary, this re-
action could not be performed experimentally.
[12] a) M. Schlosser, D. Michel, Tetrahedron 1996, 52, 99–108; b) D.
OꢁHagan, H. S. Rzepa, Chem. Commun. 1997, 645–652.
[22] a) P. J. Hay, W. R. Wadt, J. Chem. Phys. 1985, 82, 270–283; b) idem
page 284–298; c) idem, page 299–310; d) A. Hçllwarth, M. Bçhme,
S. Dapprich, A. W. Ehlers, A. Gobbi, V. Jonas, K. F. Kçhler, R. Steg-
mann, A. Veldkamp, G. Frenking, Chem. Phys. Lett. 1993, 208, 237–
240; e) A. W. Ehlers, M. Bçhme, S. Dapprich, A. Gobbi, A. Hçll-
warth, V. Jonas, K. F. Kçhler, R. Stegmann, A. Veldkamp, G. Frenk-
ing, Chem. Phys. Lett. 1993, 208, 111–114; f) K. A. Peterson, D.
Figgen, E. Goll, H. Stoll, M. Dolg, J. Chem. Phys. 2003, 119, 11113–
11123.
[23] a) S. Miertus, J. Tomasi, Chem. Phys. 1981, 55, 117–129; b) S. Mier-
tus, J. Tomasi, Chem. Phys. 1982, 65, 239–245; c) M. Cossi, V.
Barone, R. Cammi, J. Tomasi, Chem. Phys. Lett. 1996, 255, 327–335.
[24] a) the dielectric constant values for these two solvents could seem
not very different (e=7.58 for THF as compared to e=2.379 for tol-
uene). However, it has been shown (references b and c hereafter)
that the change in the solvent effect is very rapid for small dielectric
constants. The difference in polarity between THF and toluene
could consequently bee sufficient for testing the solvent effect in
these reactions. b) G. Sini, M. Bellassoued, N. Brodie, Tetrahedron
2000, 56, 1207–1215; c) B. Bogdanov, T. B. McMahon, Int. J. Mass
Spectrom. 2005, 241, 205–223.
[13] Gaussian 98 (Revision A.7), M. J. Frisch, G. W. Trucks, H. B. Schle-
gel, G. E. Scuseria, M. A. Robb, J. R. Cheeseman, V. G. Zakrzewski,
J. A. Montgomery, Jr., R. E. Stratmann, J. C. Burant, S. Dapprich,
J. M. Millam, A. D. Daniels, K. N. Kudin, M. C. Strain, O. Farkas, J.
Tomasi, V. Barone, M. Cossi, R. Cammi, B. Mennucci, C. Pomelli,
C. Adamo, S. Clifford, J. Ochterski, G. A. Petersson, P. Y. Ayala, Q.
Cui, K. Morokuma, D. K. Malick, A. D. Rabuck, K. Raghavachari,
J. B. Foresman, J. Cioslowski, J. V. Ortiz, B. B. Stefanov, G. Liu, A.
Liashenko, P. Piskorz, I. Komaromi, R. Gomperts, R. L. Martin,
D. J. Fox, T. Keith, M. A. Al-Laham, C. Y. Peng, A. Nanayakkara,
C. Gonzalez, M. Challacombe, P. M. W. Gill, B. G. Johnson, W.
Chen, M. W. Wong, J. L. Andres, M. Head-Gordon, E. S. Replogle,
J. A. Pople, Gaussian, Inc., Pittsburgh, PA, 1998.
[14] V. Barone, M. Cossi, J. Phys. Chem. A 1998, 102, 1995–2001.
[15] D. Evans, J. M. Takacs, Tetrahedron Lett. 1980, 21, 4233–4236.
[16] The NPA charge on Na atom for trans-E1 structure is 0.90. ELF cal-
culations and total electronic density surfaces on this structure
(Figure 1, Supporting Information) confirm the nearly pure electro-
static nature of the Na···F interaction. For more details about ELF
method see; B. Silvi, A. Sevin, Nature 1994, 371, 683–686.
[17] The common characteristic of the three transition states is that the
I···C distances are 16–21% longer than in the reactants, while the
C···C distances are approximately 78% longer than the C-C distance
in the final product (S)-2b. These transition states are consequently
much closer to the reactants than to the products and can be consid-
ered as early type transition states.
[25] a) C. E. Check, T. M. Gilbert, J. Org. Chem. 2005, 70, 9828–9834;
b) L. M. Pratt, N. V. NguyÞn, B. Ramachandran, J. Org. Chem. 2005,
70, 4279–4283; c) C. Aleman, D. Curco, J. Casanovas, Phys. Rev. E
2005, 72, 026704.
[26] a) A. A. Mohamed, F. Jensen, J. Phys. Chem. A 2001, 105, 3259–
3268; b) B. Bogdanov, T. B. McMahon, Int. J. Mass Spectrom. 2005,
241, 205–223.
[18] The calculated DS° values contain errors owing to the low frequency
intermolecular vibrational modes and the neglect of their anharmon-
ic character. Although this problem does not modify the comparison
between the competing transition states and the corresponding con-
clusions, one should be prudent with the absolute values of the acti-
vation parameters given in the Figures 2 and 4. (for more informa-
tion see computational details)
[27] S. F. Boys, F. Bernardi, Mol. Phys. 1970, 19, 553–566.
Received: October 10, 2007
Published online: February 18, 2008
3370
www.chemeurj.org
ꢀ 2008 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2008, 14, 3363 – 3370