Highly enantioselective organocatalytic trifluoromethyl carbinol synthesis-a caveat on reaction times and product isolation

Article abstract of DOI:10.1021/ja302511t

Aldol reactions with trifluoroacetophenones as acceptors yield chiral α-aryl, α-trifluoromethyl tertiary alcohols, valuable intermediates in organic synthesis. Of the various organocatalysts examined, Singh's catalyst [(2S)-N-[(1S)-1-hydroxydiphenylmethyl

Full text of DOI:10.1021/ja302511t

Article
pubs.acs.org/JACS
Highly Enantioselective Organocatalytic Trifluoromethyl Carbinol
SynthesisA Caveat on Reaction Times and Product Isolation
Nongnaphat Duangdee,† Wacharee Harnying,† Giuseppe Rulli,‡ Jorg-M. Neudorfl,† Harald Groger,*,‡,§
̈
̈
̈
and Albrecht Berkessel*,†
†Department of Chemistry (Organic Chemistry), University of Cologne, Greinstrasse 4, 50939 Cologne, Germany
‡Department of Chemistry and Pharmacy (Organic Chemistry II), University of Erlangen-Nurnberg, Henkestr. 42, 91056 Erlangen,
̈
Germany
§Department of Chemistry (Organic Chemistry I), Bielefeld University, Universitatsstrasse 25, 33615 Bielefeld, Germany
̈
S
* Supporting Information
ABSTRACT: Aldol reactions with trifluoroacetophenones as
acceptors yield chiral α-aryl, α-trifluoromethyl tertiary alcohols,
valuable intermediates in organic synthesis. Of the various
organocatalysts examined, Singh’s catalyst [(2S)-N-[(1S)-1-
hydroxydiphenylmethyl-3-methylbutyl]-2-pyrrolidinecarbox-
amide] was found to efficiently promote this organocatalytic
transformation in a highly enantioselective manner. Detailed
reaction monitoring (19F-NMR, HPLC) showed that, up to
full conversion, the catalytic transformation proceeds under
kinetic control and affords up to 95% ee in a time-independent
manner. At longer reaction times, the catalyst effects racemization. For the product aldols, even weak acids (such as ammonium
chloride) or protic solvents, can induce racemization, too. Thus, acid-free workup, at carefully chosen reaction time, is crucial for
the isolation of the aldols in high (and stable) enantiomeric purity. As evidenced by 19F-NMR, X-ray structural analysis, and
independent synthesis of a stable intramolecular variant, Singh’s catalyst reversibly forms a catalytically inactive (“parasitic”)
intermediate, namely a N,O-hemiacetal with trifluoroacetophenones. X-ray crystallography also allowed the determination of the
product aldols’ absolute configuration (S).
work described herein was on the use of trifluoromethyl
ketones 1 as electrophiles, which allows the preparation of
chiral β-trifluoromethyl-β-hydroxy ketones of type 3 (Scheme
1, bottom) by aldol addition of acetone (2).
In this paper, we describe the development of a practical and
highly enantioselective aldol reaction providing β-trifluoro-
methyl-β-hydroxy ketones of type 3 with up to 95% ee (and
known absolute configuration). The development of the
method hinges on our careful reaction analysis (revealing
surprising effects of “standard” sample preparation protocols)
and a detailed mechanistic study which, inter alia, identified a
“parasitic” adduct formation between the organocatalyst and
the trifluoromethyl ketone substrate.
A literature search for trifluoromethyl ketones as acceptors in
asymmetric organocatalytic aldol reactions yielded only a few
results. In 2005, Zhang and co-workers reported the first
example of a proline-catalyzed asymmetric aldol addition of
methyl ketones to aryl trifluoromethyl ketones, affording β-
trifluoromethyl-β-hydroxy ketones in satisfactory yields, but
with only moderate enantioselectivity of up to 64% ee.5 A few
years later, the use of L-proline bearing a bulky tris-
1. INTRODUCTION
Organofluorine compounds are of ever increasing importance
as pharmaceuticals, agrochemicals, functional materials, or
catalysts.1 This development calls for a new synthetic
methodology for the stereoselective introduction of fluorine
or fluorine containing substructures. Over the past decade,
organocatalysis has emerged as a highly efficient synthetic tool
for enantioselective transformations.2 Among those, the
asymmetric organocatalytic aldol reaction still serves as one
of the most powerful carbon−carbon bond forming methods,
providing access to β-hydroxycarbonyl compounds in an
enantioselective fashion.3 We recently reported that the
enantioselective organocatalytic aldol addition of acetone to
aldehydes can be combined, in a one-pot procedure, with the
diastereoselective enzymatic reduction of the aldols’ keto-
function. As a result, all four stereoisomeric 1,3-diols can be
produced at will, with diasteriomeric ratio (dr) typically >20:1
and virtually perfect enantiomeric purity (ee >98%; Scheme 1,
top).4 We are currently aiming at the development of a related
approach providing chiral 1,3-diols carrying a trifluoromethyl
substituent at the quaternary chiral center. In general, the
switch from aldehydes to ketones as aldol acceptors still
represents a challenge in organic synthesis, and accordingly,
only few examples are known in this field. The focus of the
Received: March 21, 2012
© XXXX American Chemical Society
A
dx.doi.org/10.1021/ja302511t | J. Am. Chem. Soc. XXXX, XXX, XXX−XXX

Journal of the American Chemical Society
Article
Scheme 1. (Top) Combined Organo- and Biocatalytic
Reaction Sequence for the Synthesis of Chiral 1,3-Diols from
Aldehydes (ref 4); (Bottom) Acetone (2) Addition to
Trifluoroacetophenones 1 (This Paper), Affording tert-
Aldols 3 with a CF3-Group
Table 1. Aldol Addition of Acetone (2, Reagent/Solvent) to
2,2,2-Trifluoroacetophenone (1a) to Afford 3a in the
Presence of 10 mol % of Various Catalysts
additive (10
mol %)
time
(h)
conversion
ee 3a
b
a
entry catalyst
(%)
(%)
1
2
3
4
5
6
7
4
5
6
7
8
8
8
3
6
quant
quant
quant
80
36
60
93
5
c
c
3
168
168
168
168
57
3
TFA
65
2
(trimethylsilyl)silyloxy group at position 4 of the pyrrolidine
ring was described by Yuan and Zhang.6 Here, α,β-unsaturated
trifluoromethyl ketones acted as acceptor. With a catalyst
loading of 5 mol % at −20 °C, the 1,2-addition of acetone
proceeded to give the aldol adducts in good yield and up to
91% ee, albeit at long reaction times of 3−7 days. Very recently,
in the course of our study, Nakamura et al. have developed an
improved enantioselective aldol reaction of acetone with
trihalomethyl ketones catalyzed by the TFA salt of N-(8-
quinolinesulfonyl)prolinamide.7 β-Trifluoromethyl-β-hydroxy
ketones were obtained in good yields and with up to 90%
enantiomeric excess, employing 10 mol % of the catalyst at −20
°C over 48 h. Absolute product configuration was not
determined. For a practical process, a further increase in
enantioselectivity, at low catalyst loading, short reaction time,
and preferably at room temperature is desirablein other
words, improvement of both the catalyst activity and
enantioselectivity. Herein, we describe the practical and highly
enantioselective preparation of β-trifluoromethyl-β-hydroxy
ketones 3 using readily accessible Singh’s catalyst.8,9 Detailed
kinetic and spectroscopic analyses of this aldol reaction turned
out to be the key to success and additionally identified reaction
intermediates and a “parasitic” catalyst-substrate adduct.
TfOH
70
2
a
Conversion was determined by 19F NMR of the reaction mixture.
The reaction mixture was quenched with sat. NH4Cl before analysis,
and the ee of product 3a was determined by HPLC using a Daicel
b
c
Chiralcel OD-H column. The reaction was performed in a NMR tube
and directly monitored by 19F NMR.
were catalytically highly active, too, and gave 3a quantitatively
in a short reaction time, albeit with much lower asymmetric
induction compared to the case of 6 (entries 1 and 2). Chiral
primary amine catalysts were also examined. With the cis-
DACH-derived organocatalyst 712 and 9-amino-9-deoxyepiqui-
nine (8),13 low catalytic activity and poor enantiocontrol were
observed (entries 4 and 5). The slow conversion may be due to
the sluggish generation of the enamine intermediate. We found
that the addition of acid in combination with 8 accelerated the
reaction, but the product enantiomeric excess remained
negligible (entries 6 and 7).
In summary, compound 6 proved to be the catalyst of choice
for promoting the enantioselective addition of acetone to
trifluoroacetophenone (1a) at room temperature. As Singh’s
catalyst 6 has been shown to perform well in combined
asymmetric organo- and biocatalytic reaction sequences
(Scheme 1),4 our current results open a perspective for the
chemoenzymatic synthesis of 1,3-diols harboring a trifluoro-
methyl carbinol. This extension of our current work shall be
addressed in future studies. A further advantage of organo-
catalyst 6 is that it can readily be prepared on large laboratory
scale in a one-pot procedure, starting from cheap proline and
the corresponding leucine ester.9
2. RESULTS AND DISCUSSION
2.1. Catalyst Screening. Ideally, the aldol addition of
acetone (2) to trifluoromethyl ketones such as trifluoroaceto-
phenone (1a) should proceed at room temperature and
without additional solvents. In our catalyst screening, the
reaction was monitored by TLC and 19F NMR, using 0.2 mmol
of trifluoroacetophenone (1a) in an excess of acetone-d6 (35
equiv) and in the presence of 10 mol % of the catalyst (relative
to 1a). The results are summarized in Table 1.
Among the catalysts tested, we identified the proline amide 6
(“Singh’s catalyst”) as the most effective one: the desired aldol
adduct 3a was obtained in excellent yield and with high
enantioselectivity of 93% ee, over a period of only 3 h (entry
3).10 Both proline (4) and the N-tosylated proline amide 511
2.2. Kinetic Investigations on the Direct Aldol
Addition of Acetone to 1a in the Presence of Catalyst
6. We have recently reported the significant influence of kinetic
versus thermodynamic control on the enantioselectivity of the
organocatalytic aldol reaction of acetone with 3-chlorobenzal-
dehyde in the presence of catalyst 6.4b In that study, we found
that, at 5 mol % catalyst loading, the reaction rapidly reached
B
dx.doi.org/10.1021/ja302511t | J. Am. Chem. Soc. XXXX, XXX, XXX−XXX

Journal of the American Chemical Society
Article
Figure 1. Kinetic investigations on the effect of solvents (A), of additives (B), and of quenching methods (C) on the aldol reaction of acetone to
trifluoroacetophenone (1a) in the presence of 10 mol % of the catalyst 6. For A and B, the reaction mixture was quenched with sat. NH4Cl before
HPLC analysis of the enantioselectivity.
equilibrium (<30 min), thus switching from kinetic to
thermodynamic control, leading to a depletion in enantiomeric
purity of the product. This earlier observation prompted us to
investigate the formation of the aldol adduct 3a and its
enantiomeric excess as a function of time. For this purpose, the
conversion of 1a was monitored by in situ 19F NMR.
Additionally, the enantiomeric ratio of the product 3a/ent-3a
was determined by chiral HPLC, after a “mini workup” of
aliquots of the reaction mixture. This mini workup consisted in
quenching the sample with sat. NH4Cl, and extraction of the
product 3/ent-3 with Et2O. Figure 1 summarizes the results: In
the left column, the conversion of starting trifluoroacetophe-
none 1a is plotted as a function of reaction time. In the middle
column, the enantiomeric excess of the product aldol 3a is
given as a function of reaction time. The graphs in the right
column depict the enantiomeric excess of the product 3a as a
function of conversion of the starting 1a.
2.2.1. Effect of Solvents. Solvent effects on the addition of
acetone to trifluoroacetophenone (1a) are summarized in the
top row (A) of Figure 1. As indicated by the blue line, with 10
mol % of the catalyst 6 in neat acetone-d6 (35 equiv), the
reaction equilibrium was reached in ca. 2.5 h (quantitative
conversion, 93% ee).14 At extended reaction times, thermody-
namic control results in a slow decrease of product
enantiomeric excess. The most astonishing feature of the
reaction course is, however, the increasing enantiomeric excess
of the product 3a in the initial phase of the reaction, i.e. over
the course of the reaction under kinetic control (<2.5 h).
Theoretically, one would have expected time-independent ee in
this regime. This remarkable phenomenon was similarly
observed when the reaction was run in other solvents, such
as CD2Cl2, CDCl3, or toluene-d8 (10 equiv of acetone-d6 were
used in these cases). All reactions studied occurred in the
homogeneous phase, except for the reaction in toluene-d8, for
C
dx.doi.org/10.1021/ja302511t | J. Am. Chem. Soc. XXXX, XXX, XXX−XXX

Journal of the American Chemical Society
Article
Figure 2. Effect of catalyst (6) loading on the aldol addition of acetone (2) to trifluoroacetophenone (1a), shown as relationships of product-based
conversion vs time (left), the enantioselectivity vs time (middle), and the enantioselectivity vs conversion (right).
which we observed a slightly turbid solution. Consequently,
enantioselective amplification based on phase behavior, such as
solid-solute equilibria in certain amino acid mediated aldol
reactions,15 could be excluded.
In our search for an explanation for this unusual effect, we
first considered and evaluated the possibility of a kinetically
induced amplification,16 via either an autocatalytic17 or an
autoinductive18 process in which the reaction product itself acts
as a catalyst or is involved in the formation of a more selective
catalyst. By addition of the aldol product 3a (90% ee) to the
mixture of 1a and acetone in the absence of the catalyst 6, no
(further) formation of the aldol adduct 3a occurred. This
indicates that the effect is not due to autocatalysis.
2.2.2. Effect of Additives [Middle Row (B) of Figure 1]. We
next examined whether a catalyst-product adduct (6−3a) may
form in the course of the reaction and may act as a “secondary”
catalyst of higher enantioselectivity. When 10 mol % of 3a
(88% ee) was added, right at the beginning, to the reaction
mixture containing 10 mol % of the catalyst 6, the optical purity
of the product 3a still increased gradually [compare blue (no
additive) and pink (with additive) lines in the middle row (B)
of Figure 1]. This result excludes enantioselective auto-
induction by the reaction product in this transformation.
In principle, a time course such as the blue line in Figure 1,
middle row (B), could also be rationalized by a less selective
catalyst (or racemization catalyst) being present at the onset of
the reaction and being consumed/deactivated in the course of
the reaction. We excluded this possibility by running our
standard aldol reaction (in acetone-d6) to completion and
adding the substrate 1a once again. No effect on the ee/time-
course resulted (graph not shown).
It has been demonstrated that Singh’s catalyst (6) is
compatible with aqueous media;4,8 in fact, water increases the
rate and enantioselectivity of aldol reactions catalyzed by 6.8
We therefore investigated the effect of water as an additive in
our reaction. Upon addition of water (10 mol %), the reaction
became faster and reached completion within 1.5 h, at slightly
decreased enantiopurity of the product [90% ee; Figure 1, row
B; compare green and blue lines]. Note, however, that, in the
presence of trifluoroacetophenone (1a), water in the reaction
mixture is scavenged in the form of its hydrate, the gem-diol 9
(as evidenced by 19F NMR analysis). As revealed by control
experiments, this species accelerates the reaction but does not
account for the observed increase in ee over time.19 A similar
effect was also observed when 10 mol % of 1,1,1,3,3,3-
hexafluoro-2-propanol (HFIP) was added. In this case,
conversion was complete within 20 min, while the enantiomeric
excess decreased slightly to 87% ee [at complete conversion
(Figure 1, row (B), black line]. It is evident from the graphs
that HFIP, as a mildly acidic (pKa 17.2 in DMSO)20 additive,
enhances the rate of the reaction in both the kinetic and
thermodynamic regimes. For the latter, this corresponds to an
increased rate of racemization. With this in mind, we turned to
investigate the effect of the mildly acidic (NH4Cl) aqueous
workup.
2.2.3. Effect of Workup/Product Isolation. Typically,
aliquots withdrawn from the reaction mixture were quenched
with aqueous NH4Cl prior to HPLC analysis. Alternatively,
they were either diluted (1% iPrOH in n-hexane) and injected
directly or filtered through a short pad of silica gel (same
solvent mixture) prior to analysis. Figure 1, row C, summarizes
the results (compare red and blue lines). Upon omission of the
aqueous NH4Cl quenching, the measured enantiomeric excess
of the product 3a (93% ee) becomes time independent; that is,
the initial increase in ee has vanished. This clearly indicates that
even the mildly acidic workup procedure previously used
effected racemization of the aldol product 3a, in particular at
low conversion. When the reaction reaches equilibrium, the
acidic quenching has much less or no influence on the
enantiopurity of the product, as shown by the superposition of
the blue and red lines (Figure 1, row C). Clearly, at low
conversion, the substrate (i.e., aldol 3a) to catalyst ratio is
highest.21 The above results raise an important caveat for the
handling of β-hydroxyketones of type 3a (and other organic
reaction products), which are prone to racemization in the
presence of traces of remaining catalyst and/or acids.
2.3. Effect of Catalyst Loadings. We furthermore
investigated the effect of catalyst loading (6, 1−20 mol %)
on the aldol addition to trifluoroacetophenone (1a) in acetone-
d6 (“no-quench” ee-analysis). The results are illustrated in
Figure 2. As expected, increasing catalyst loading results in
increased reaction rate. The initial rates are roughly propor-
D
dx.doi.org/10.1021/ja302511t | J. Am. Chem. Soc. XXXX, XXX, XXX−XXX

Journal of the American Chemical Society
Article
tional to the catalyst concentration; the reaction thus shows
first order kinetics in catalyst. Higher catalyst concentrations at
the same time effect an earlier onset of the thermodynamic
regime, i.e. racemization. From a practical point of view, a
catalyst loading of 1−2 mol % proved optimal, as the kinetic
regime extended to at least 60 h, during which the product
enantiomeric excess remained virtually unchanged (Figure 2,
blue and green lines). In contrast, with catalyst loadings of 5−
20 mol %, equilibrium was reached quickly [0.5, 3, and 8 h for
catalyst loadings of 20, 10, and 5 mol %, respectively], after
which the ee value gradually eroded. In summary, the best
results with respect to yield, product enantiopurity, and
reaction time were achieved with just 2 mol % of the catalyst
6: the reaction was complete in 18 h to give a quantitative yield
of the aldol 3a with 94% ee. By lowering the catalyst loading to
1 mol %, the reaction took longer time for completion (up to
45 h), while there was no significant increase in enantiomeric
excess (95% ee).
structure of the adduct (S,S,S)-10 by X-ray diffraction analysis,
as shown in Scheme 2. The carbinolamine 10 results from the
Scheme 2. X-ray Crystal Structure of (S,S,S)-10 (Top); in
Solution, 10 Reverts to 1a and 6; Subsequent Reaction with
Acetone Gives Rise to the Aldol Adduct 3a (Bottom)
2.4. Characterization of a Substrate−Catalyst Adduct
Formed in the Reaction of Acetone with Trifluoroace-
tophenone (1a) in the Presence of the Catalyst 6. In the
course of in situ 19F NMR monitoring of the reaction between
1a and acetone, in the presence of the catalyst 6, we noted the
transient appearance of small resonances of unknown species at
δ = −76.3 and −77.2 ppm (Figure 3, highlighted by the green
box) and even smaller peaks in the region of δ = −81 and −82
ppm.
nucleophilic addition of the pyrrolidine-NH of 6 to the
carbonyl group of 1a. In principle, two diastereomeric
carbinolamines can result from the addition of 6 to 1a
[(S,S,S)-10 and (R,S,S)-10]. The crystal structure of (S,S,S)-10
indicates that there is an intramolecular hydrogen bond
between the quaternary OH group and the carbonyl oxygen
atom. This hydrogen bond may well account for the preferred
crystallization of the (S,S,S)-diastereomer. Note that similar
hydrogen bonding interactions have been invoked in the
stabilization of tetrahedral carbinolamine intermediates formed
in the early stages of the catalytic cycle of iminium−enamine
catalysis, involving proline or derivatives thereof.2c To the best
of our knowledge, 10 represents the first spectroscopically and
X-ray crystallographically characterized catalyst−substrate
adduct of this type.
19F NMR analysis of (S,S,S)-10 in toluene-d8 shows two
small peaks at δ = −76.4 and −77.5 ppm, together with a large
peak of trifluoroacetophenone (1a) at −72.2 ppm (Figure 4,
trace b). Apparently in solution, the equilibrium of the
reversible 1a−6 addition lies far on the side of dissociation.
We assign the two weak resonances at δ = −76.4 and −77.5
ppm to the two diastereomers of 10. When acetone was used as
solvent, the formation of the aldol adduct 3a was additionally
observed (Figure 4, trace c). In fact, the NMR spectrum
obtained from dissolution of 10 in acetone-d6 closely resembles
that of the reaction mixture containing 1a and a catalytic
amount of 6 (Figure 4, compare traces c and a). Consequently,
both diastereomers of intermediate 10 are formed, in small
amount, from the reversible addition of the chiral amine
catalyst 6 to the trifluoroketone 1a. In the presence of acetone,
the released catalyst 6 promotes the aldol addition of the
former to 1a, resulting in a shift of equilibrium between 1a and
the carbinolamine 10 (Scheme 2). Once 1a was completely
consumed to give 3a, the resonances of the intermediate 10
disappeared. As mentioned before, in an acetone solution of 10
Figure 3. 19F NMR spectra of the reaction mixture of 1a and acetone-
d6 in the presence of 10 mol % of the catalyst 6 as a function of time
(blue lines a−f) and of a 1:1 mixture of 1a and 6 in CD2Cl2 (red line
g).
These signals completely vanished when the reaction was
complete (Figure 3, trace f). Clearly, the formation of these
intermediates must originate from trifluoroacetophenone (1a),
as it is the only source of 19F. No signal in the above-mentioned
regions was detected in the spectrum of 1a alone in acetone-d6.
However, upon mixing 1a and the catalyst 6 in a ratio of 1:1 in
CD2Cl2, we could again observe two small peaks between δ =
−76 and −78 ppm (Figure 3, trace g). This suggests that the
species giving rise to those resonances result from the
interaction of trifluoroacetophenone (1a) with Singh’s catalyst
6.
To our delight, slow evaporation of a solution of 6 and 1a
(1:1) in DCM furnished crystals suitable for X-ray analysis.
Thus, we were able to establish, for the first time, the molecular
E
dx.doi.org/10.1021/ja302511t | J. Am. Chem. Soc. XXXX, XXX, XXX−XXX

Journal of the American Chemical Society
Article
Avoiding contact with acids and low temperatures is crucial for
the long-time storage of rac-11.
As shown in Figure 4, trace d, the quaternary CF3 group of
11 resonates at δ = −76.0 ppm, which is almost identical to the
lower-field resonance of the carbinolamines 10. The model
compound rac-11 furthermore enabled us to examine the
catalytic activity of such carbinolamine adducts. To this end, we
mixed rac-11 (10 mol %) and trifluoroacetophenone (1a) in
acetone-d6. No aldol adduct 3a was formed under these
conditions, indicating the catalytic inactivity of rac-11. We
extrapolate from this that the “real” carbinolamines 10 are
catalytically inactive, too.
2.5. Substrate Scope of the Reaction. Besides the parent
compound 1a, five substituted trifluoroacetophenones 1b−f
were reacted with acetone (neat) in the presence of 2 mol % of
catalyst 6. In several cases, commercially available trifluoro-
acetophenones contained traces of trifluoroacetic acid, as
evidenced by 19F NMR. To achieve reproducible and
comparable results in the aldol reactions, it was crucial to
remove these impurities by storing the starting materials over
MgSO4/NaHCO3 and filtering through neutral Al2O3 prior to
use. The aldol reactions were conducted on 2−5 mmol scale,
and they were monitored by TLC. Upon completion, the
products 3a−f were isolated from the reaction mixture, without
aqueous workup, by short column chromatography. The results
are summarized in Table 2.
Figure 4. 19F NMR spectra of (a) the reaction mixture of 1a and 10
mol % of 6 in acetone-d6 at t = 10 min; (b) 10 in toluene-d8; (c) 10 in
acetone-d6; (d) rac-11 in acetone-d6; (e) a 1:1 mixture of 3a and 6 in
acetone-d6; (f) a 1:1 mixture of 3a and 6 in DMSO-d6.
(and also in the course of the aldol reaction), we furthermore
observed two weak 19F-NMR resonances in the region δ = −81
and −82 ppm (Figure 4, traces a and c). These resonances
could also be generated by mixing the catalyst 6 with the aldol
product 3a, in particular in DMSO-d6 as solvent (Figure 4,
traces e and f). In the absence of further data, we tentatively
assign these resonances to catalyst−product adducts, presum-
ably of the iminium or enamine type.22
Table 2. Enantioselective Addition of Acetone to Various
Aryl Trifluoromethyl Ketones 1a−f in the Presence of 6 (2
mol %) to Afford the Aldols 3a−f
To further support our assignment of the transient 19F NMR
resonances at δ = −76.4 and −77.5 ppm, we synthesized a
cyclic, and thus more stable, carbinolamine mimic of 10:
namely the iso-indolinol rac-11. As depicted in Scheme 3, by
Scheme 3. Synthesis of rac-11 and Its X-ray Structure
a
b
c
d,e
20
entry
3
R
time (h)
yield (%)
ee (%)
[α]D
f
1
3a
3b
3c
3d
3e
3f
H
20 (18)
25 (24)
42 (40)
20 (19)
10 (9)
91
90
92
97
90
92
92 (93)
94 (94)
95 (95)
84 (84)
94 (94)
92 (93)
+24
+24
+24
+29
+23
+23
2
3
4
5
p-Me
p-OMe
o-OMe
p-F
f
6
p-Cl
24 (18)
a
Unless otherwise noted, the reaction was carried out on the 2 mmol
scale (substrates 1a−f) in dry acetone (35 equiv), in the presence of
b
catalyst 6 (2 mol %), at room temperature. In parentheses: reaction
time for completion in acetone-d6 (0.2 mmol scale); monitoring by 19
F
c
NMR (see details in SI). Isolated yields by column chromatography.
d
The ee value of the isolated products was determined by HPLC
e
analysis using a Daicel Chiralcel OD-H column. Values in parentheses
refer to crude product, before column chromatography. The reaction
f
was performed on 5 mmol scale.
starting from 2-bromobenzaldehyde (12), rac-11 could be
synthesized in a three step sequence comprising reductive
amination,23 trifluoroacetylation of the resulting amine 13, and
final cyclization of 14 via lithium−halogen exchange. We noted
that the carbinolamine rac-11 is prone to dehydration, affording
the iso-indole 15 (see SI for spectra of 15). Nevertheless, pure
iso-indolinol rac-11 could be obtained by crystallization. The
We were delighted to find that, under these conditions, the
desired products 3a−f could be isolated in excellent yields of
90−97% and with high enantiopurity of typically 92−95% ee
[with the exception of o-methoxy trifluoroacetophenone (3d),
84% ee; Table 2, entry 4],24 which were only marginally lower
than that measured from the reaction mixtures (for the reaction
profiles of the substrates 3b−f; see Supporting Information).
With regard to reaction rates, the following trend was found: p-
F (3e) ≫ unsubstituted (3a) ≈ o-OMe (3d) ≈ p-Cl (3f) ≈ p-
1
structure of rac-11 was established by H and 13C NMR, and
additionally confirmed by X-ray diffraction analysis (Scheme 3).
F
dx.doi.org/10.1021/ja302511t | J. Am. Chem. Soc. XXXX, XXX, XXX−XXX

Journal of the American Chemical Society
Article
Figure 5. X-ray crystal structures of rac-3f (a) and (S)-3f (b).
Me (3b) ≫ p-OMe (3c). This trend roughly reflects the
expected electrophilicity of the ketone substrates.
in which the quaternary OH group of one molecule forms an
intermolecular hydrogen bond with the carbonyl oxygen atom
of another molecule. In line with the structural differences
described above, the two materials rac-3f and (S)-3f exhibit
significantly different melting points of 82 and 42 °C,
respectively. The predominantly formed configuration of the
other products 3a−e is, in analogy, assumed to be (S) as well
[when using (S,S)-configurated Singh-catalyst 6].
2.6. Absolute Configuration of the Aldol Adducts: To
clarify the absolute configuration of the aldol adducts,25 we
attempted the crystallization of heavy-atom containing 3f and
the determination of its absolute configuration by anomalous
X-ray scattering. In our first experiments, we employed optically
active 3f (92% ee) and attempted crystallization from protic
solvents, such as EtOH. We were surprised to find that the
crystals grown under these conditions were those of the
racemate (rac-3f). X-ray crystallographic analysis furthermore
revealed that rac-3f crystallizes in the form of cyclic hydrogen-
bonded (R+S)-dimers, as shown in Figure 5a. We next
attempted the crystallization of highly enantiomerically
enriched 3f directly after purification, by evaporation of the
solvent under reduced pressure. To our delight, crystals of
enantiomerically pure 3f, suitable for X-ray structural analysis,
could be grown at low temperature (0 °C). The absolute
configuration was determined to be (S), as shown in Figure 5b.
This material crystallized in a noncentrosymmetric space group
(P1) with a Flack parameter26 of 0.02(6), indicating that the
crystals of (S)-3f are indeed enantiomerically pure. Importantly,
X-ray diffraction (Figure 5) revealed that (S)-3f and rac-3f
show distinct intermolecular hydrogen bonding interactions.
Whereas enantiomerically pure (S)-3f forms an endless, zigzag-
patterned ribbon, rac-3f exists in the crystal as a heterochiral
dimer, with a cyclic hydrogen bonding network consisting of
two bifurcated OH-hydrogen bonds. Interestingly, the specific
hydrogen bonding patterns observed here are reminiscent of
what we previously reported for the racemic and the
enantiopure forms of 1-phenyl-2,2,2-trifluoroethanol
(PhTFE). In the crystal structures of the latter, (R)-PhTFE
crystallizes in endless zigzag chains, whereas the racemic
PhTFE forms cyclic (2R+2S)-tetramers.27 In addition, the
crystal packing of (S)-3f reveals no intramolecular hydrogen
bond between the carbonyl oxygen atom and the OH group,
which is found in the case of rac-3f. Instead, there are
significant intermolecular hydrogen-bonding interactions be-
tween three molecules of (S)-3f found in the asymmetric unit,
3. CONCLUSIONS
On the basis of reaction analysis and the resulting mechanistic
insight, we have developed a high-yielding and enantioselective
aldol addition of acetone to aryl trifluoromethyl ketones,
proceeding at room temperature. Employing Singh’s catalyst 6
at low loading of 2 mol %, high yields (90−97%) of the
resulting β-trifluoromethyl-β-hydroxy ketones 3a−f were
obtained under proper reaction conditions, with very high
enantiomeric excesses (84−95%). To achieve these numbers, it
was crucial to identify (and to avoid) conditions that induce
racemization of the aldol adducts 3: protic solvents, including
aqueous solution, the presence of acid, and high catalyst
loadings of ≥5 mol %. This study also demonstrated the
importance of careful sample preparation for reaction analysis.
Moreover, we characterized the structure of the “parasitic” (i.e.,
catalytically incompetent) intermediate 10, which is formed
reversibly from the amine catalyst 6 and the aldol acceptor
component, i.e. ketone 1a. Further studies in our laboratory
aim at the combination of the organocatalytic aldol process
described herein with subsequent biocatalytic reduction, to
generate the corresponding chiral 1,3-diols in stereochemically
defined form.
4. EXPERIMENTAL SECTION
4.1. General. 19F NMR spectra were measured on a Bruker Avance
400 instrument; 1H and 13C NMR spectra were measured on a Bruker
Avance 400, a Bruker DRX 500, or a Bruker Avance II 600 instrument
at 25 °C and are referenced to the solvent used. IR spectra were
recorded on a Shimadzu IR Affinity-1 FT-IR spectrometer. HPLC
analyses were performed on a Merck Hitachi LaChrom HPLC
G
dx.doi.org/10.1021/ja302511t | J. Am. Chem. Soc. XXXX, XXX, XXX−XXX

Journal of the American Chemical Society
Article
instrument. Optical rotations were measured on a Perkin-Elmer 343
scale to afford 3c as a colorless viscous liquid (0.483 g, 92%, 95% ee);
20
1
Plus polarimeter. Melting points were determined on a Buchi
Rf = 0.37 (acetone/hexane 1:3); [α]D +23.6 (c 1.0, CHCl3); H
NMR (600 MHz, CDCl3): δ = 7.50 (d, J = 8.6 Hz, 2H, Ar−H), 6.93
(d, J = 8.8 Hz, 2H, Ar−H), 5.41 (br. s, 1H, OH), 3.83 (s, 3H, OCH3),
3.36 (d, J = 17.1 Hz, 1H, CHH), 3.18 (d, J = 17.1 Hz, 1H, CHH), 2.22
(s, 3H, CH3) ppm; 13C NMR (150 MHz, CDCl3): δ = 209.1 (CO),
159.9, 129.4, 127.5, 127.4−121.7 (q, CF3), 113.8, 76.1−75.5 (q,
OCCF3), 55.2 (CH3), 45.1 (CH2), 32.1 (CH3) ppm; 19F NMR (376
̈
apparatus and are uncorrected.
L-Proline (99%, BioChemica AppliChem), DL-proline (99%, Sigma
Aldrich), 1,1,1,3,3,3-hexafluoro-2-propanol (HFIP, 99%, ABCR GmbH
& Co. KG), deuterated acetone (99.8 %D, H2O+D2O < 0.02%,
Euriso-Top), dry acetone (99.8%, extra dry, AcroSeal, Acros Organics)
were purchased from the suppliers indicated and used as received.
2,2,2-Trifluoroacetophenone (1a, 99%, Sigma Aldrich), 4'-chloro-
2,2,2-trifluoroacetophenone (1f, 99%, Sigma Aldrich), 4'-fluoro-2,2,2-
trifluoroacetophenone (1e, Fluorochem), 4'-methoxy-2,2,2-trifluoroa-
cetophenone (1c, Fluorochem), 2'-methoxy-2,2,2-trifluoroacetophe-
none (1d, Fluorochem) and 4'-methyl-2,2,2-trifluoroacetophenone
(1b, Fluorochem) were stored over MgSO4/NaHCO3, and filtered
through neutral Al2O3 prior to use.
4.2. General Procedure for Kinetic Investigations: Monitor-
ing by 19F NMR. To the aryl trifluoromethyl ketone 1 (0.2 mmol)
was added a solution of catalyst 6 (7.3 mg, 0.02 mmol, 10 mol %) in
deuterated acetone (0.5 mL, 35 equiv). When a different solvent was
used for the reaction, acetone-d6 (0.14 mL, 10 equiv) was mixed with
that solvent to a total volume of 0.5 mL. The reaction mixture was
transferred to a NMR tube and immediately monitored by 19F NMR.
The product-based conversion was determined from the ratio of the
integral of the product signal in relation to the sum of integrals of all
signals (substrate 1, product, and side-products).
4.3. General Procedure for Kinetic Investigations: Monitor-
ing by HPLC. The reaction mixture was prepared as described for 19F
NMR analysis. At intervals, aliquots were withdrawn and prepared for
chiral HPLC analysis by one of the two following methods: (1)
addition to saturated aqueous ammonium chloride and extraction with
Et2O prior to analysis; (2) filtration though a short pad of silica gel,
eluting with 1% i-PrOH in n-hexane.
4.4. General Procedure for the Synthesis of 3a−f:
Preparative Scale. A solution of the catalyst 6 (0.0367 g, 0.1
mmol) in dry acetone (12.5 mL) was added to the aryl trifluoromethyl
ketone 1 (5 mmol). The reaction mixture was stirred at room
temperature and monitored by TLC. Upon completion, the reaction
mixture was diluted with hexane (12.5 mL) and purified immediately
by short column chromatography on silica gel (acetone/hexane) to
give the aldol adducts (S)-3. See the Supporting Information for the
spectra and HPL-chromatograms of the aldol products.
̃
MHz, CDCl3): δ = −80.7 ppm. IR (ATR): ν = 3417, 3007, 2962,
2916, 2841, 1705 (s), 1612, 1538, 1514, 1463, 1419, 1363, 1334, 1301,
1246, 1153 (s), 1152, 1058, 1029, 983, 912, 831, 800, 738, 713 cm−1.
5,5,5-Trifluoro-4-hydroxy-4-(2-methoxyphenyl)pentan-2-one
(3d). This was prepared following the general procedure on a 2 mmol
scale to afford 3d as a white solid (0.509 g, 97%, 84% ee); Rf = 0.36
(acetone/hexane 1:3); mp 64−74 °C; [α]D20 +29.4 (c 1.0, CHCl3); 1H
NMR (600 MHz, CDCl3): δ = 7.66 (d, J = 7.9 Hz, 1H, Ar−H), 7.40−
7.33 (m, 1H, Ar−H), 7.05 (t, J = 7.6 Hz, 1H, Ar−H), δ = 6.96 (d, J =
8.3 Hz, 1H, Ar−H), 5.66 (br s, 1H, OH), 3.90 (s, 3H, OCH3), 3.84 (d,
J = 16.7 Hz, 1H, CHH), 3.13 (d, J = 16.7 Hz, 1H, CHH), 2.19 (s, 3H,
CH3) ppm; 13C NMR (150 MHz, CDCl3): δ = 207.8 (CO), 157.1,
130.5, 130.2, 127.6−121.9 (q, CF3), 124.0, 121.3, 112.1, 76.7−76.1 (q,
OCCF3), 55.8 (OCH3), 46.3 (CH2), 31.1 (CH3) ppm; 19F NMR (376
̃
MHz, CDCl3): δ = −81.1 ppm; IR (ATR): ν = 3498, 3084, 3010,
2976, 2839, 1718 (s), 1600, 1587, 1490, 1465, 1435, 1402, 1361, 1342,
1246, 1228, 1170, 1155 (s), 1132, 1107, 1060, 1045, 1016, 987, 916,
754, 721, 702, 624 cm−1; Anal. Calcd for C12H13F3O3: C, 54.96; H,
5.00. Found: C, 55.22; H, 5.43.
4-(4-Fluorophenyl)-5,5,5-trifluoro-4-hydroxypentan-2-one (3e).
This was prepared following the general procedure on a 2 mmol
scale to afford 3e as a colorless viscous liquid (0.450 g, 90%, 94% ee);
20
1
Rf = 0.57 (acetone/hexane 1:3); [α]D +22.9 (c 1.0, CHCl3); H
NMR (600 MHz, CDCl3): δ = 7.57 (dd, J = 8.5, 5.3 Hz, 2H, Ar−H),
7.10 (t, J = 8.6 Hz, 2H, Ar−H), 5.53 (br s, 1H, OH), 3.33 (d, J = 17.2
Hz, 1H, CHH), 3.23 (d, J = 17.2 Hz, 1H, CHH), 2.24 (s, 3H, CH3)
ppm; 13C NMR (150 MHz, CDCl3): δ = 208.9 (CO), 163.7−162.1
(d, Ar−CF), 133.3, 128.1, 128.1, 127.2−121.5 (q, CF3), 115.5, 115.4,
76.0−75.4 (q, OCCF3), 45.0 (CH2), 32.0 (CH3) ppm; 19F NMR (376
MHz, CDCl3): δ = −80.5 (CF3), −113.1 (−113.2) (Ar−F) ppm. IR
̃
(ATR): ν = 3417, 3078, 3010, 2922, 1707 (s), 1604, 1510, 1419, 1363,
1334, 1234, 1157 (s), 1132, 1095, 1055, 993, 835, 736 cm−1.
4-(4-Chlorophenyl)-5,5,5-trifluoro-4-hydroxypentan-2-one (3f).
This was prepared following the general procedure on a 5 mmol
scale to afford 3f as a white semisolid (1.227 g, 92%, 92% ee); Rf =
0.52 (acetone/hexane 1:3); mp 42−44 °C (for rac-3f; mp 82−85 °C);
5,5,5-Trifluoro-4-hydroxy-4-phenylpentan-2-one (3a). This com-
pound was prepared following the general procedure on a 5 mmol
scale to afford 3a as a colorless viscous liquid which solidified as a
white solid upon standing at 0 °C (1.056 g, 91%, 92% ee); Rf = 0.60
(acetone/hexane 1:3); mp 39−41 °C (for rac-3a; mp 55−57 °C);
20
1
[α]D +22.8 (c 1.0, CDCl3); H NMR (600 MHz, CHCl3): δ = 7.53
(d, J = 8.5 Hz, 2H, Ar−H), 7.40 (d, J = 8.6 Hz, 2H, Ar−H), 5.52 (s,
1H, OH), 3.33 (d, J = 17.3 Hz, 1H, CHH), 3.23 (d, J = 17.3 Hz, 1H,
CHH), 2.24 (s, 3H, CH3) ppm; 13C NMR (150 MHz, CDCl3): δ =
208.8 (CO), 136.1, 135.0, 128.7, 127.6, 127.1−121.4 (q, CF3),
76.0−75.5 (q, OCCF3), 44.9 (CH2), 32.0 (CH3) ppm; 19F NMR (376
20
1
[α]D +24.1 (c 1.0, CHCl3); H NMR (600 MHz, CDCl3): δ = 7.60
(d, J = 7.6 Hz, 2H, Ar−H), 7.44−7.38 (m, 3H, Ar−H), 5.50 (s, 1H,
OH), 3.39 (d, J = 17.2 Hz, 1H, CHH), 3.23 (d, J = 17.2 Hz, 1H,
CHH), 2.21 (s, 3H, CH3) ppm; 13C NMR (150 MHz, CDCl3): δ =
209.0 (CO), 137.5, 128.8, 128.5, 127.4−121.7 (q, CF3), 76.3−75.7
(q, OCCF3), 45.1 (CH2), 32.0 (CH3) ppm; 19F NMR (376 MHz,
̃
MHz, CDCl3): δ = −80.4 ppm. IR (ATR): ν = 3398, 3095, 3008,
2918, 1705 (s), 1598, 1492, 1417, 1363, 1332, 1244, 1159 (s), 1136,
1093, 1055, 1014, 991, 914, 827, 758, 732 cm−1.
̃
CDCl3): δ = −81.1 ppm. IR (ATR): ν = 3462, 3280, 2910, 1714, 1691
4.5. Synthesis of rac-11. N-Methyl-2-bromobenzylamine (13).
To a mixture of 2-bromobenzaldehyde (1.850 g, 10 mmol) and 3 Å
molecular sieves (2 g) in dry ethanol (4 mL) was added CH3NH2 (4
mL, 33% solution in ethanol, 32 mmol) under argon. The reaction
mixture was stirred for 15 h at room temperature. The resulting
mixture was cooled down to 0 °C, and NaBH4 (0.567 g, 15 mmol) was
added. The reaction mixture was then stirred at room temperature for
24 h. Water (10 mL) was added, and the molecular sieves were filtered
off. The filtrate was extracted with ethyl acetate (3 × 20 mL). The
combined organic layers were washed with water and brine and dried
over MgSO4. After removal of the solvent, the crude product was
obtained as a clear pale yellow liquid. Purification by flash column
chromatography on SiO2 (ethyl acetate/n-hexane) afforded the desired
(s), 1456, 1411, 1365, 1344, 1199, 1159 (s), 1136, 1060, 991, 906, 761,
700, 675, 640 cm−1.
5,5,5-Trifluoro-4-hydroxy-4-p-tolylpentan-2-one (3b). This com-
pound was prepared following the general procedure on a 2 mmol
scale to afford 3b as a white solid (0.443 g, 90%, 94% ee); Rf = 0.50
(acetone/hexane 1:3); mp 56−58 °C; [α]D20 +24.1 (c 1.0, CHCl3); 1H
NMR (600 MHz, CDCl3): δ = 7.48 (d, J = 8.0 Hz, 2H, Ar−H), 7.23
(d, J = 8.1 Hz, 2H, Ar−H), 5.42 (s, 1H, OH), 3.39 (d, J = 17.1 Hz, 1H,
CHH), 3.21 (d, J = 17.1 Hz, 1H, CHH), 2.38 (s, 3H, CH3), 2.21 (s,
3H, CH3) ppm; 13C NMR (150 MHz, CDCl3): δ = 209.1 (CO),
138.7, 134.5, 129.2, 126.7, 127.4−121.7 (q, CF3), 76.3−75.7 (q,
OCCF3), 45.1 (CH2), 32.1 (CH3), 21.0 (CH3) ppm; 19F NMR (376
̃
MHz, CDCl3): δ = −80.5 ppm. IR (ATR): ν = 3444, 3394, 3035,
1
product 13 as a clear colorless liquid (1.911 g, 96% yield); H NMR
2927, 1714, 1701 (s), 1516, 1411, 1361, 1253, 1219, 1193, 1155 (s),
1138, 1109, 1047, 1022, 979, 920, 817, 806, 731, 715 cm−1.
(600 MHz, CDCl3): δ = 7.56 (dd, J = 8.0 Hz, 0.8 Hz, 1H, Ar−H), 7.39
(dd, J = 7.6 Hz, 1.4 Hz, 1H, Ar−H), 7.32−7.27 (m, 1H, Ar−H), 7.14
(td, J = 7.8 Hz, 1.6 Hz, 1H, Ar−H), 3.85 (s, 2H, NCH2), 2.47 (s, 3H,
5,5,5-Trifluoro-4-hydroxy-4-(4-methoxyphenyl)pentan-2-one
(3c). This was prepared following the general procedure on a 2 mmol
H
dx.doi.org/10.1021/ja302511t | J. Am. Chem. Soc. XXXX, XXX, XXX−XXX

Journal of the American Chemical Society
Article
NCH3), 1.51 (bs, 1H, NH) ppm; 13C NMR (150 MHz, CDCl3): δ =
139.1, 132.8, 130.3, 128.6, 127.4, 124.0, 55.8 (CH2), 35.9 (CH3) ppm;
(2) For reviews, see: (a) Berkessel, A.; Groger, H. Asymmetric
̈
Organocatalysis: From Biomimetic Concepts to Applications in Asymmetric
Synthesis; Wiley-VCH: Weinheim, 2005. (b) Dalko, P. I., Ed.
Enantioselective Organocatalysis; Wiley-VCH: Weinheim, 2007. (c) Mel-
chiorre, P.; Marigo, M.; Carlone, A.; Bartoli, G. Angew. Chem., Int. Ed.
2008, 47, 6138−6171.
(3) Review: Trost, B. M.; Brindle, C. S. Chem. Soc. Rev. 2010, 39,
1600−1632.
(4) (a) Baer, K.; Kraußer, M.; Burda, E.; Hummel, W.; Berkessel, A.;
̃
IR (ATR): ν = 3319 (br), 3059, 2931, 2843, 2791, 1591, 1566, 1487,
1438, 1352, 1199, 1130, 1103, 1024, 823, 744 (s), 655 cm−1.
N-Methyl-1-(trifluoromethyl)isoindolin-1-ol (rac-11). To a sol-
ution of 13 (0.600 g, 3 mmol) in dry THF (0.6 mL) was added TFAA
(0.6 mL, 4.2 mmol), in a dropwise manner, with ice cooling (5−10
°C). The mixture was stirred in the ice bath for an additional 1 h. After
removal of THF and TFAA under vacuum, compound 14 was
obtained, and it was used in the next step without further purification.
To a solution of 14 in dry THF (12 mL) was added n-Buli (3.5 mL,
2.5 M in hexane, 8.7 mmol) in a dropwise manner at −78 °C under Ar.
The reaction mixture was stirred at the same temperature for 1 h and
quenched with water (5 mL) at −78 °C. The mixture was immediately
extracted with diethyl ether (3 × 15 mL), and the combined extracts
were dried over MgSO4. After removal of the solvent, and upon
storage at 0 °C, the desired product rac-11 separated from the yellow
oily crude material as a pale yellow solid (0.101 g, 16% yield); mp 82−
84 °C; 1H NMR (600 MHz, acetone-d6): δ = 7.51 (d, J = 7.7 Hz, 1H,
Ar−H), 7.41 (d, J = 7.4 Hz, 1H, Ar−H), 7.35 (t, J = 7.4 Hz, 2H, Ar−
H), 5.62 (s, 1H, OH), 4.24 (d, J = 13.4 Hz, 1H, CHH), 4.02 (d, J =
13.4 Hz, 1H, CHH), 2.71 (s, 3H, NCH3) ppm; 13C NMR (150 MHz,
acetone-d6): δ = 140.2, 138.6, 129.3, 127.8−122.1 (q, CF3), 127.2,
123.7, 122.3, 92.7−92.1 (q, OCCF3), 57.8 (CH2), 32.5 (CH3) ppm;
Groger, H. Angew. Chem., Int. Ed. 2009, 48, 9355−9358. (b) Rulli, G.;
̈
Duangdee, N.; Baer, K.; Hummel, W.; Berkessel, A.; Groger, H. Angew.
̈
Chem., Int. Ed. 2011, 50, 7944−7947.
(5) Qiu, L.-H.; Shen, Z.-X.; Shi, C.-Q.; Liu, Y.-H.; Zhang, Y.-W. Chin.
J. Chem. 2005, 23, 584−588.
(6) Zhang, D.; Yuan, C. Tetrahedron 2008, 64, 2480−2488.
(7) Hara, N.; Tamura, R.; Funahashi, Y.; Nakamura, S. Org. Lett.
2011, 13, 1662−1665.
(8) (a) Raj, M.; Maya, V.; Ginotra, S. K.; Singh, V. K. Org. Lett. 2006,
8, 4097−4099. (b) Maya, V.; Raj, M.; Singh, V. K. Org. Lett. 2007, 9,
2593−2595. (c) Vishnumaya, M. R.; Singh, V. K. J. Org. Chem. 2009,
74, 4289−4297.
(9) Berkessel, A.; Harnying, W.; Duangdee, N.; Neudolfl, J.-M.;
̈
Groger, H. Org. Process Res. Dev. 2012, 16, 123−128.
̈
(10) Lowering the reaction temperature did not significantly affect
product enantiopurity: when the reaction was carried out at −20 °C
and at 0 °C, the product 3a was obtained in 94% ee after 50 h and in
94% ee after 10 h, respectively.
19F NMR (376 MHz, acetone-d6): δ = −75.9 ppm; IR (ATR): ν
̃
=
3062 (br), 3045, 2962, 2848, 1614, 1595, 1471, 1458, 1354, 1280,
1259, 1157 (s), 1147 (s), 1132 (s), 1085, 1006, 916, 765, 742, 673,
638 cm−1.
CCDC 869533−CCDC 869536 contain the supplementary
crystallographic data for this paper [compounds (S,S,S)-10, rac-11,
rac-3f, and (S)-3f]. These data can be obtained free of charge from
The Cambridge Crystallographic Data Centre via www.ccdc.cam.ac.
uk/data_request/cif.
(11) Berkessel, A.; Koch, B.; Lex, J. Adv. Synth. Catal. 2004, 346,
1141−1146.
(12) Berkessel, A.; Ong, M.-C.; Nachi, M.; Neudorfl, J.-M.
̈
ChemCatChem 2010, 2, 1215−1218.
(13) Brunner, H.; Bugler, J.; Nuber, B. Tetrahedron: Asymmetry 1995,
̈
6, 1699−1702.
(14) We observed no significant kinetic isotope effect when switching
from acetone to acetone-d6.
ASSOCIATED CONTENT
■
(15) (a) Klussmann, M.; Iwamura, H.; Mathew, S. P.; Wells, D. H.,
Jr.; Pandya, U.; Armstrong, A.; Blackmond, D. G. Nature 2006, 441,
621−623. (b) Hayashi, Y.; Matsumoto, M.; Yamagushi, J.; Yonehara,
S.; Matsumoto, Y.; Shoji, M.; Hashizume, D.; Koshino, H. Angew.
Chem., Int. Ed. 2006, 45, 4593−4597. For a review, see:
(c) Blackmond, D. G. Philos. Trans. R. Soc. B 2011, 366, 2878−2884.
(16) (a) Blackmond, D. G. Angew. Chem., Int. Ed. 2009, 48, 386−390.
For reviews on kinetic aspects of nonlinear effects in asymmetric
catalysis, see: (b) Blackmond, D. G. Acc. Chem. Res. 2000, 33, 402−
411. (c) Blackmond, D. G. Tetrahedron: Asymmetry 2010, 21, 1630−
1634.
(17) (a) Soai, K.; Shibata, T.; Morioka, H.; Choji, K. Nature 1995,
378, 767−768. (b) Shibata, T.; Morioka, H.; Hayase, T.; Choji, K.;
Soai, K. J. Am. Chem. Soc. 1996, 118, 471−472. (c) Shibata, T.; Choji,
K.; Hayase, T.; Aizu, Y.; Soai, K. Chem. Commun. 1996, 1235−1236.
(d) Kawasaki, T.; Matsumura, Y.; Tsutsumi, T.; Suzuki, K.; Ito, M.;
Soai, K. Science 2009, 324, 492−495.
(18) (a) Alberts, A. H.; Wynberg, H. J. Am. Chem. Soc. 1989, 111,
7265−7266. (b) Danda, H.; Nishikawa, H.; Otaka, K. J. Org. Chem.
1991, 56, 6740−6741. (c) Mathew, S. P.; Iwamura, H.; Blackmond, D.
G. Angew. Chem., Int. Ed. 2004, 43, 3317−3321. (d) Iwamura, H.;
Mathew, S. P.; Blackmond, D. G. J. Am. Chem. Soc. 2004, 126, 11770−
11771.
(19) The effect of the gem-diol 9 was studied by its in situ generation,
by preincubation of a 1:1 mixture of trifluoroacetophenone and water
(10 mol%) in acetone for 1 h (formation of 9 evidenced by 19F NMR
analysis). The subsequent addition of the catalyst 6 and the substrate
1a gave reaction profiles that were identical to those obtained by
simply adding water (10 mol%) to the reaction mixture.
S
* Supporting Information
HPLC, NMR (1H, 13C, 19F) and IR data of the aldols 3a−f, of
the catalyst−substrate adduct 10, and of the model compound
rac-11 (including synthetic intermediate 13 and iso-indole 15);
additional kinetic data and reaction monitoring by 19F-NMR;
X-ray data of 3f, rac-3f, 10, and rac-11. This material is available
free of charge via the Internet at http://pubs.acs.org.
AUTHOR INFORMATION
■
Corresponding Author
harald.groeger@uni-bielefeld.de; berkessel@uni-koeln.de
Notes
The authors declare no competing financial interest.
ACKNOWLEDGMENTS
■
This work was supported by the Deutsche Forschungsgemein-
schaft (priority program “Organocatalysis”, SPP 1179) and by
the Fonds der Chemischen Industrie. Membership in the
COST Action no. CM0905 Organocatalysis (ORCA) is
gratefully acknowledged.
REFERENCES
■
(1) For reviews, see: (a) Filler, R., Kobayashi, Y., Yagupolskii, L. N.,
Eds. Organofluorine Compounds in Medicinal and Biochemical
Applications; Elsevier: Amsterdam, 1993. (b) Hiyama, T., Ed.
Organofluorine Compounds: Chemistry and Applications; Springer:
New York, 2000. (c) Kirsch, P. Modern Fluoroorganic Chemistry:
Synthesis, Reactivity, Application; Wiley-VCH: Weinheim, 2004.
(d) Purser, S.; Moore, P. R.; Swallow, S.; Gouverneur, V. Chem. Soc.
Rev. 2008, 37, 320−330.
(20) Christ, P.; Lindsay, A. G.; Vormittag, S. S.; Neudorfl, J.-M.;
̈
Berkessel, A.; O’Donoghue, A. C. Chem.Eur. J. 2011, 17, 8524−
8528.
(21) The Singh-catalyst 6 is active in aqueous mediumfor reviews,
see: (a) Raj, M.; Singh, V. K. Chem. Commun. 2009, 6687−6703.
I
dx.doi.org/10.1021/ja302511t | J. Am. Chem. Soc. XXXX, XXX, XXX−XXX

Journal of the American Chemical Society
Article
(b) Mase, N.; Barbas, C. F., III. Org. Biomol. Chem. 2010, 8, 4043−
4050.
(22) (a) Schmid, M. B.; Zeitler, K.; Gschwind, R. M. Angew. Chem.,
Int. Ed. 2010, 49, 4997−5003. (b) Schmid, M. B.; Zeitler, K.;
Gschwind, R. M. J. Am. Chem. Soc. 2011, 133, 7065−7074. (c) Schmid,
M. B.; Zeitler, K.; Gschwind, R. M. J. Org. Chem. 2011, 76, 3005−
3015. (d) Schmid, M. B.; Zeitler, K.; Gschwind, R. M. Chem. Sci. 2011,
2, 1793−1803.
(23) Iwai, Y.; Kita, K.; Matsushira, Y.; Yamauchi, A.; Kihara, M.
Chem. Pharm. Bull. 2002, 50, 441−443.
(24) According to a recent report (see ref 7), a series of β-
trifluoromethyl-β-hydroxy ketones and one sample of β-nonafluoro-
butyl-β-hydroxy ketones were obtained in up to 95% yield and 92% ee
on a reaction scale of 0.1 mmol.
(25) In ref 7, the absolute configuration of the product β-
trifluoromethyl-β-hydroxy ketones was assigned to be (S), in analogy
to the case of a β-trichloromethyl-β-hydroxy ketone for which the
absolute configuration was determined by X-ray diffraction analysis.
(26) Flack, H. Acta Crystallogr., Sect. A 1983, 39, 876−881.
(27) Berkessel, A.; Adrio, J. A.; Huttenhain, D.; Neudorfl, J. M. J. Am.
̈
̈
Chem. Soc. 2006, 128, 8421−8426.
J
dx.doi.org/10.1021/ja302511t | J. Am. Chem. Soc. XXXX, XXX, XXX−XXX