The role of the counterion and of silicon chelation in the selective mono(trifluoromethylation) of tartaric acid derived 1,4-diketones

Article abstract of DOI:10.1002/chem.201100551

The addition of trifluoromethyl(trimethyl)silane (TFMTMS) to tartaric acid derived aromatic 1,4-diketones was reported to be highly stereoselective but also chemoselective towards a monoaddition product. This chemoselectivity remained unexplained. Complementary experiments and monitoring of the reaction by electrospray ionization mass spectrometry (ESI-MS) show that: i) the countercation (tetrabutylammonium (TBA⁺) or Cs⁺) associated to the initiator and the charged intermediates in the chain reaction plays a crucial role and ii) in the case of a tetrabutylammonium salt initiator, the second addition on the preformed monoadduct does occur but the resulting bis(trifluoromethyl)carbinolate is unable to evolve through the chain-transfer process and it exhibits an intramolecular Si-O interaction.

Full text of DOI:10.1002/chem.201100551

DOI: 10.1002/chem.201100551
The Role of the Counterion and of Silicon Chelation in the Selective
Mono(trifluoromethylation) of Tartaric Acid Derived 1,4-Diketones
Dominique Harakat, Fabien Massicot, Jean Nonnenmacher, Fabienne Grellepois, and
Charles Portella*[a]
Abstract: The addition of trifluorome-
thyl(trimethyl)silane (TFMTMS) to
tartaric acid derived aromatic 1,4-dike-
tones was reported to be highly stereo-
selective but also chemoselective to-
wards a monoaddition product. This
trospray ionization mass spectrometry
(ESI-MS) show that: i) the countercat-
ion (tetrabutylammonium (TBA+) or
Cs+) associated to the initiator and the
charged intermediates in the chain re-
action plays a crucial role and ii) in the
case of a tetrabutylammonium salt ini-
tiator, the second addition on the pre-
formed monoadduct does occur but the
resulting bis(trifluoromethyl)carbino-
late is unable to evolve through the
chain-transfer process and it exhibits
Keywords: ketones
· mass spec-
chemoselectivity
remained
unex-
trometry · reaction mechanisms ·
silicon · trifluoromethylation
plained. Complementary experiments
and monitoring of the reaction by elec-
À
an intramolecular Si O interaction.
Introduction
vantage of this methodology and of this reagent lies in the
stabilization of the TFM anion within a silicate species,
which prevents its decomposition into difluorocarbene.[4]
The need for only a catalytic amount of the nucleophilic ac-
tivator is in accordance with this chain mechanism.
We recently reported the preparation of enantiopure a-
alkoxy-a-TFM aldehydes from l-tartaric acid derived dike-
tones[5] or ketoamides.[6] The first step was the fluoride-in-
duced[5,6] or carbonate-induced[6] diastereoselective and che-
moselective addition of TFMTMS (Scheme 2). If the selec-
The nucleophilic trifluoromethylation of carbonyl com-
pounds
by
using
trifluoromethyl(trimethyl)silane
(TFMTMS; Ruppert–Prakash reagent)[1,2] is a widely ap-
plied methodology to have access to trifluoromethyl (TFM)
derivatives,[3] an important class of organic chemicals. This
reaction proceeds through a chain mechanism: the initiation
is induced by nucleophilic activation of TFMTMS and the
chain transfer proceeds by reaction of the intermediate
TFM-carbinolate with TFMTMS (Scheme 1). The key ad-
Scheme 2. Monotrifluoromethylation of tartaric acid derived diketones or
ketoamides. TMS: trimethylsilyl.
tive monotrifluoromethylation of ketoamides was rather ex-
pected, we found the highly selective monoaddition on the
diketone intriguing. Indeed, the bisphenone derivative 1 was
converted, under tetrabutylammonium fluoride trihydrate
(TBAF) or tetrabutylammonium difluorotriphenylsilicate
(DFTPSi; DeShong reagent[7]) initiation,[8] into the corre-
sponding ketocarbinol silyl ether 2, even with an excess of
TFMTMS (Scheme 3, Table 1). In contrast, the bisadduct 3
was obtained by using CsF as the initiator, in accordance
with previous results for simple diketones.[9,10] We were keen
on investigating this reaction more deeply to understand the
reasons for such selectivity. Electrospray ionization mass
spectrometry (ESI-MS) is a powerful tool for reactivity stud-
ies.[11,12] We made use of this analytical technology, in addi-
Scheme 1. Chain mechanism of trifluoromethylation with CF3SiMe3.
[a] Dr. D. Harakat, Dr. F. Massicot, Dr. J. Nonnenmacher,
Dr. F. Grellepois, Prof. C. Portella
Institut de Chimie Molꢀculaire de Reims (UMR 6229)
Universitꢀ de Reims Champagne-Ardenne - CNRS
UFR Sciences, BP 1039, 51687 REIMS Cedex 2 (France)
Fax : (+33)326-91-31-66
E-mail: charles.portella@univ-reims.fr
Supporting information for this article is available on the WWW
under http://dx.doi.org/10.1002/chem.201100551.
10636
ꢁ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2011, 17, 10636 – 10642

FULL PAPER
mediates 4 and 5 and the corresponding m/z figures (for the
anion species). As regards a possible subsequent reaction of
2 with TFMTMS, the situation seems not so clear. With CsF
as the initiator, we can assume that a similar chain reaction
occurs (Scheme 5, path a). With TBAF or DFTPSi, both
Scheme 3. The selectivity of the trifluoromethylation of tartaric acid de-
rived diketone 1 varied according to the fluoride salt initiator.
Table 1. Selectivity of trifluoromethylation of tartaric acid derived dike-
tone 1 versus fluoride salt initiator.[a]
Equivalents of Initiator
CF3SiMe3
Reaction
Yield of 2
[%] (de [%]) 3 [%]
Yield of
conditions[b]
1–3
1–3
1
TBAF·3H2O
THF, 08C or 90 (96)
RT, 0.33 h
0
À
TBA+Ph3SiF2 CH2Cl2, 08C, 99 (92–96)
0
0.5–2 h
CsF
CsF
DME,
90 (40)
9 (0)
0
À408C, 3 h
DME, RT,
12 h
3
54[c]
[a] Taken from reference [5]. [b] THF: tetrahydrofuran; RT: room tem-
perature; DME: 1,2-diemthoxyethane. [c] Ratio of diastereoisomers: C2
symmetrical product/C2 symmetrical product/C1 symmetrical product 16/
1
84/0, as determined by H NMR spectroscopy.
tion to complementary chemical experiments, to revisit our
reaction. We report herein this study, which highlights the
role of the cation and of the silyl moiety.
Scheme 5. Reaction of monoadduct 2 with CF3SiMe3.
Results and Discussion
bringing a tetrabutylammonium counterion, the pathway is
obviously different. Our basic hypothesis took into account
the fact that intermediate 6, resulting from the trifluorome-
thylation of 2 would bear both a silyl ether moiety and an
alkoxide moiety (Scheme 5). Owing to the high oxophilicity
of silicon, a strong intramolecular interaction, not to say a
bond, could induce a particular behavior of this intermedi-
ate, which also depends on the countercation. Thus, an Si-
chelated species would be favored under TBAF or DFTPSi
initiation, which would prevent the chain-addition process
from continuing (Scheme 5, path b). We have attempted to
assess such an hypothesis by two approaches: complementa-
ry chemical experiments and monitoring of reactions by
ESI-MS under various conditions.
Before detailing the study on bisphenylketone 1, we have to
disclose some side results observed during the preparative
aspect of this study. When the monoTFM adduct derived
from the corresponding bis(ethylketone) was treated with
TFMTMS under TBAF activation, the second addition took
place, even if its diastereoselectivity was lower.[13] In con-
trast, adduct 2 was recovered after a new treatment under
the same conditions. Thus, the conjugated character and/or
the steric environment of the remaining carbonyl group af-
fects its reactivity. The reaction giving monoadduct 2 does
not deserve particular comment with regard to the chain
mechanism, which is depicted in Scheme 4, including inter-
Chemical experiments: Bis(trifluoro)methylation occurred
under CsF initiation and the reaction with CsF needed
DME as a solvent, so we first checked that using this solvent
with TBAF did not modify the chemoselectivity for mono-
addition on 1.
To gain insights into the possible role of the TMS group
in stopping the reaction at the monoadduct stage, the O-
allyl ketoether analogue 8 was prepared from the corre-
sponding amido ether[6] and treated with TFMTMS under
similar conditions (TBAF, THF). We were unable to isolate
the bisTFM derivative 9, even after adding an excess of re-
agents (Scheme 6).[14] This observation means that chelation
of silicon, if there is any, is not the only parameter responsi-
Scheme 4. First addition of CF3SiMe3 onto diketone 1.
Chem. Eur. J. 2011, 17, 10636 – 10642
ꢁ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
10637

C. Portella et al.
Scheme 6. Tentative reaction of CF3SiMe3 with a ketoether.
ble for the chemoselectivity of the monoaddition. Given
that the nature of the fluoride salt is important to control
the chemoselectivity of the addition on the O-silylated mon-
oadduct 2, the reaction of the O-allyl analogue 8 was per-
formed under CsF initiation in DME and, indeed, the
second addition took place to give the bisTFM derivative 9
in 87% yield after isolation (two diastereomers in a ratio of
31:69). Once again, a control experiment with TBAF in
DME enabled us to exclude any role of the solvent. These
experiments led us to conclude that the nature of the coun-
terion is a decisive factor in stopping the reaction after mon-
oaddition (TBA+) or allowing the second addition to take
place (Cs+), whatever the substituent on the oxygen atom.
The influence of the size of the cation associated to the fluo-
ride ion has already been observed in the case of TFMTMS
addition on sulfinimides,[15] in which case tetramethylammo-
nium fluoride proved to be more effective than TBAF. In a
reaction of similar type, fluoride-induced (phenylsulfonyl)di-
fluoromethylation of carbonyl compounds from the parent
TMS reagent was effective by using DFTPSi as an initiator
for aldehydes, but CsF was needed for ketones.[16] The
countercation is the only difference in the first example,[15]
which seems indicative of the influence of its size, even if
the hydrated nature of TBAF may also be of importance
with the strong Lewis base character of sulfinimides. In the
second example,[16] both the cation (TBA+ or Cs+) and the
nature of the fluoride species (fluoride ion or the bulky
DFTPSi) are different and their relative influence is not so
clear. Let us remember that DFTPSi and TBAF gave the
same results and selectivity in our case, which indicates that
the chemoselectivity is closely linked to the TBA+ counter-
cation. Nevertheless, we cannot rule out silicon chelation in
the case of further reaction on compound 2.
Scheme 7. Does
a
pseudosymmetrical intermediate exist during the
second addition step? Conditions: CF3SiMe3 (1.1 equiv), TBAF
(0.05 equiv), THF, À208C.
We thus propose that, if any chelation takes place, then
intermediate 6 (resulting from trifluoromethylation of the
monoadduct 2) evolves, under TBAF or DFTPSi initiation,
À
via a chelated form of type 6’, with an Si O intramolecular
interaction that is weaker than a covalent bond (Scheme 5).
The possible occurrence of such an intermediate cannot be
confirmed by chemical means, which justifies the following
ESI-MS study.
ESI-MS study: The chemical approach afforded some in-
sight into the mechanism of these transformations based on
reaction products. We have undertaken an ESI-MS study in
order to have access to the postulated intermediates. Sam-
ples of crude reaction medium were directly introduced into
the mass spectrometer ion source, with the instrument oper-
ating in the negative ESIÀ ion mode (all of the trifluorome-
thylated intermediate species are negatively charged). The
key parameters for this study are the reaction time, initial
stoichiometry of TFMTMS, and nature of the fluoride salt
initiator.
The difluorosilicate (DFTPSi) salt was chosen as represen-
tative of the tetraalkylammonium-type initiators allowing a
selective monoaddition and was compared to caesium fluo-
ride. Selected experiments were presented in Figure 1–4. Di-
ketone 1 was first treated with a substoichiometric amount
of TFMTMS to approach the intermediates involved in the
very first addition process depicted in Scheme 4. As exhibit-
ed in Figure 1, the ESI-MS spectra contain a peak corre-
sponding to the nonsilylated intermediate 4 (m/z 379),[18] but
no trace of the silylated compound 5 (m/z 521) was detected
(Scheme 4). This means that the chain-transfer step is very
fast and that the reaction kinetics are controlled by the step
from 4 to 5. On the other hand, the absence of a signal at
m/z 521 is of importance to make easier further interpreta-
tion and discrimination between 5 and the isomeric species
6/6’.
In order to assess the degree of possible Si chelation in
the intermediate, we reasoned that covalent bonding be-
tween the oxygen and silicon atoms should give a pseudo-
symmetrical intermediate I and release the trifluoromethide
anion from one or the other TFM carbinolate moiety
(Scheme 7). Either of the keto silyl ethers 10 and 11 should
give a mixture of both of them. Compounds 10 and 11 were
prepared[17] and treated under similar conditions with the
TFMTMS/TBAF system. In each case, only the starting ma-
terial was recovered, which rules out the occurrence of the
covalently chelated intermediate I. Similar results were ob-
tained with analogues in which the methoxyphenyl group
was replaced by an ethyl group.
We next examined the behavior of the monoadduct 2 with
TFMTMS under DFTPSi initiation, reaction conditions
under which no apparent reaction occurred (Scheme 5,
path b). Small peaks at m/z 379 (Figure 2) correspond to de-
10638
www.chemeurj.org
ꢁ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2011, 17, 10636 – 10642

Selectivity in Mono(trifluoromethylation) of Diketones
FULL PAPER
À
Figure 1. ESI-MS spectra for diketone 1 and CF3SiMe3 (0.8 equiv) with initiation by TBA+Ph3SiF2
.
silylation of starting material, which is not surprising owing
to the possibility of TMS transfer to fluoride or alkoxide
species. A peak appears at m/z 521[18] and disappears after
25 min. This is in accordance with an addition occurring to
give an intermediate adduct that, instead of initiating a
chain process, decomposes by releasing CF3 (see below)
À
À
Figure 2. ESI-MS spectra for monoadduct 2 and CF3SiMe3 (1.6 equiv) with initiation by TBA+Ph3SiF2
.
Chem. Eur. J. 2011, 17, 10636 – 10642
ꢁ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
10639

C. Portella et al.
Figure 3. ESI-MS spectra for diketone 1 and CF3SiMe3 (3 equiv) with initiation by CsF.
and the starting compound 2, until total consumption of the
TFMTMS.
count this observation as a decisive argument, owing to the
qualitative character of this study and the very low relative
intensity of the peak at m/z 69.
At this stage, it is still difficult to comment more accurate-
ly on the nature of this intermediate. The next experiment is
more enlightening on that point. The spectra for the moni-
toring of the reaction of diketone 1 with an excess of
TFMTMS initiated with CsF are shown in Figure 3. Inter-
mediate 5 was undetected during the first addition, and in-
termediate 7 (m/z 663), its counterpart in the second addi-
tion process, was also undetected (Scheme 4 and Scheme 5,
path a). More interestingly, the lack of a signal at m/z 521
under these conditions in which bisadduct 3 was formed via
intermediate 6 means that such a signal under DFTPSi ini-
tiation could correspond to the chelated intermediate 6’
(Scheme 5, path b).
We have attempted in a last experiment to corroborate
the previous observations by monitoring the reaction of di-
ketone 1 with an excess of TFMTMS under DFTPSi initia-
tion (Figure 4). The peak at m/z 379 is present at the very
beginning of the reaction and the peak at m/z 521 was de-
tected from the sample taken at 30 min. At such a concen-
tration of TFMTMS, the first addition takes place rapidly
and, as expected, the degenerated chain process through in-
termediate 6’ was observed after completion of the first ad-
dition.
Concluding remarks
The chemical investigation disclosed that the counterion as-
sociated firstly to the initiator and then to the intermediates
is of crucial importance to direct the selectivity towards the
monoTFM product 2 (with TBA+) or the bisTFM product 3
(with Cs+). The ESI-MS study informed us of an intramo-
À
lecular Si O association in the intermediate, which results
from a second addition in the case of a TBA+ counterion,
with subsequent collapse of this intermediate towards the
monoadduct 2. The failure of the reaction of TFMTMS with
ketoether 8 under TBA salt initiation, in contrast to its ach-
ievement under CsF activation, indicated that the chelation
of silicon is not the cause but a consequence of the difficulty
of the TBA intermediate 6 to undergo the chain-transfer
step. These experiments confirm the sensitivity of the
second addition to steric hindrance, as already mentioned at
the beginning of this discussion. The much higher ionic
volume of TBA+ (0.2951 nm3)[20] compared to that of Cs+
(0.0206 nm3)[20,21] and its weaker solvation could explain
both the lower chain reactivity of intermediate 6/6’ and the
effective releasing of steric pressure to give back compound
2.
Complementary observations: Remarkably, in most of the
ESI-MS experiments, a mass peak at m/z 69 corresponding
to the TFM anion[18] was observed.[19] Moreover, a concomi-
tance between the appearance of this peak and the disap-
pearance of the peak at m/z 521 in the tentative second ad-
dition on the monoadduct seems to corroborate the hypoth-
Experimental Section
À
esis of the release of CF3 from the intermediate 6/6’. Even
if this observation is worthy of note considering the very
low stability of the TFM anion, we have not taken into ac-
Procedures for chemical experiments and structural data of compounds
8–11 are given in the Supporting Information.
10640
www.chemeurj.org
ꢁ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2011, 17, 10636 – 10642

Selectivity in Mono(trifluoromethylation) of Diketones
FULL PAPER
À
Figure 4. ESI-MS spectra for diketone 1 and CF3SiMe3 (3 equiv) with initiation by TBA+Ph3SiF2
.
All electrospray ionization mass spectrometry experiments (MS and
HRMS) were carried out by using a hybrid tandem quadrupole/time-of-
flight (Q-TOF) instrument, equipped with a pneumatically assisted elec-
trospray (Z-spray) ion source (Micromass, Manchester, UK) operated in
negative mode (ESIÀ). Nitrogen was used as the nebulizer gas, and the
samples were introduced into the ESI source through a needle at a flow
rate of 5 mLminÀ1. The main conditions were: capillary voltage 3000 V,
cone voltage 40 V, source temperature 808C, and desolvation tempera-
ture 1108C. The relative peak intensities observed in the spectra are in
arbitrary units, with 100% assigned to the highest peak in the mass
domain considered.
S. Mizuta, H. Kawai, Tetrahedron: Asymmetry 2008, 19, 2633–2644;
h) C. Portella, F. Grellepois, F. Massicot, J. Nonnenmacher, Chim.
Oggi 2009, 27, 50–53.
[4] D. J. Burton, Z. Y. Yang, Tetrahedron 1992, 48, 189–275.
[5] F. Massicot, N. Monnier-Benoit, N. Deka, R. Plantier-Royon, C. Por-
tella, J. Org. Chem. 2007, 72, 1174–1180.
[6] J. Nonnenmacher, F. Massicot, F. Grellepois, C. Portella, J. Org.
Chem. 2008, 73, 7990–7995.
[7] C. J. Handy, Y. F. Lam, P. DeShong, J. Org. Chem. 2000, 65, 3542–
3543.
[8] Initiation by using dried tetramethylammonium fluoride (TMAF)
and
tetrabutylammonium
difluorotriphenylstannate
(TBA+
ESI-HRMS of intermediates observed: intermediate 4: m/z calcd for
C20H18F3O4À: 379.1157 [MÀ]; found: 379.1150; intermediate 6’: m/z calcd
for C24H27F6O4SiÀ: 521.1583 [MÀ]; found: 521.1580; trifluoromethyl
anion: m/z calcd for CF3À: 68.9952 [MÀ]; found: 68.9958.
Ph3SnF2À; Gingras reagent) worked as well, even if the latter gave
lower yield and diastereoselectivity (see reference [5]).
[9] R. P. Singh, J. M. Leitch, B. Twamley, J. M. Shreeve, J. Org. Chem.
2001, 66, 1436–1440.
[10] Monotrifluoromethylation was reported in the particular case of
rigid-cage diketones, due to a subsequent transannular cyclization
´
´
leading to an O-silyl ketal species: J. Romanski, G. Mloston, ARKI-
VOC 2007, 179–187.
Acknowledgements
[11] For a monograph on ESI-MS, see: Electrospray Ionization Mass
Spectrometry, Fundamentals, Instrumentation, and Applications (Ed.:
R. B. Cole), Wiley, New York, 1997.
This work has been supported by CNRS, Rꢀgion Champagne–Ardenne
and Ville de Reims.
[12] For a recent review on the investigation of chemical reactions by
using this technique, see: L. S. Santos, Eur. J. Org. Chem. 2008, 235–
253.
[13] The isolated monoadduct, treated with TFMTMS (1.1 equiv) and
TBAF (0.05 equiv) in THF at À408C, was converted after 4 h into
the bisaddition product in 75% yield (75:25 d.r.). In contrast, the
direct formation of the bisadduct by treating the diethyl ketone with
an excess of TFMTMS failed.
[14] The reaction was performed by using initially 1.05 equivalents of
TFMTMS and 0.2 equivalent of TBAF in THF at 08C, followed by
the addition of the same amounts of TFMTMS and initiator 8 h
later.
[1] I. Ruppert, K. Schlich, W. Volbach, Tetrahedron Lett. 1984, 25,
2195–2198.
[2] G. K. S. Prakash, R. Krishnamurti, G. A. Olah, J. Am. Chem. Soc.
1989, 111, 393–395.
[3] a) G. K. S. Prakash, A. K. Yudin, Chem. Rev. 1997, 97, 757–786;
b) R. P. Singh, J. M. Shreeve, Tetrahedron 2000, 56, 7613–7632;
c) G. K. S. Prakash, M. Mandal, J. Fluorine Chem. 2001, 112, 123–
131; d) J.-A. Ma, D. Cahard, Chem. Rev. 2004, 104, 6119–6146;
update 1: Chem. Rev. 2008, 108, PR1–PR43; e) B. R. Langlois, T.
Billard, S. Roussel, J. Fluorine Chem. 2005, 126, 173–179; f) J.-A.
Ma, D. Cahard, J. Fluorine Chem. 2007, 128, 975–996; g) N. Shibata,
Chem. Eur. J. 2011, 17, 10636 – 10642
ꢁ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemeurj.org
10641

C. Portella et al.
[15] G. K. S. Prakash, M. Mandal, J. Am. Chem. Soc. 2002, 124, 6538–
6539.
[20] Y. Marcus, H. D. B. Jenkins, L. Glasser, J. Chem. Soc. Dalton Trans.
2002, 3795–3798.
[16] C. Ni, J. Hu, Tetrahedron Lett. 2005, 46, 8273–8277.
[17] Diketones 10 and 11 were prepared by aryl magnesium condensa-
tion with the corresponding N-methoxy-N-methyl- or N,N-dimethy-
lamidoketone (see the Supporting Information).
[21] This value does not take into account solvation by DME, but the
gap with TBA+ is high enough to keep this comparison significant.
[18] The structures of all intermediates considered in this ESI-MS study
were confirmed by ESI-HRMS.
[19] To the best of our knowledge, the trifluoromethyl anion has never
been observed in condensed medium by mass spectrometry.
Received: February 18, 2011
Revised: May 23, 2011
Published online: August 11, 2011
10642
www.chemeurj.org
ꢁ 2011 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Chem. Eur. J. 2011, 17, 10636 – 10642