Diamine-catalyzed addition of ZnEt2to PhC(O)CF3: Two mechanisms and autocatalytic asymmetric enhancement

Article abstract of DOI:10.1002/chem.201403648

NMR spectroscopic studies of the catalytic addition reaction of ZnEt2 to PhC(O)CF 3 in the presence of three very efficient catalysts [TMEDA, tBuBOX, and L ; where L is a chiral diamine synthesized from optically pure (R,R)-1,2-diphenylethylenediamine and

Full text of DOI:10.1002/chem.201403648

DOI: 10.1002/chem.201403648
Full Paper
&
Asymmetric Catalysis
Diamine-Catalyzed Addition of ZnEt2 to PhC(O)CF3:
Two Mechanisms and Autocatalytic Asymmetric Enhancement
Mercedes Calvillo-Barahona,[a] Juan A. Casares,[a] Carlos Cordovilla,[a] Miroslav N. Genov,[b]
Jesffls M. Martínez-Ilarduya,*[a] and Pablo Espinet*[a]
Dedicated to Pierre Braunstein for his life contribution to make chemistry in Europe a home
Abstract: NMR spectroscopic studies of the catalytic addi-
tion reaction of ZnEt2 to PhC(O)CF3 in the presence of three
very efficient catalysts [TMEDA, tBuBOX, and L; where L is
a chiral diamine synthesized from optically pure (R,R)-1,2-di-
phenylethylenediamine and (S)-2,2’-bis-(bromomethyl)-1,1’-
binaphthalene] reveal large differences in their behavior. For
the ligands TMEDA and tBuBOX, the catalysis shows no un-
usual features and proceeds via [(NÀN)Zn(Et){OC(CF3)(Et)Ph}].
For NÀN=L, the observation of autocatalytic asymmetric en-
hancement during the catalysis, and unusual inverse concen-
tration dependence on the reaction rate, indicate the partici-
pation of an additional novel catalytic cycle that goes
through a dinuclear intermediate containing one ZnEt2 and
one ZnEt fragment connected by NÀN and OR bridges. Inter-
estingly, the 19F NMR signals of the main product of the re-
action ([Zn(Et){OC*(CF3)(Et)Ph}]2) allowed us to assess the
enantioselectivity of the processes in situ without the assis-
tance of chiral chromatography.
Introduction
make trifluoromethylated molecules more bioavailable, more
lipophilic, and more chemically/metabolically stable. These
unique properties, together with their binding affinity and se-
lectivity at the molecular level, are responsible for much of the
activity shown by many trifluoromethylated drugs with thera-
peutic action and by many crop protection reagents.
The preparation of chiral molecules bearing a trifluoromethyl
group (Figure 1)[1,2] has been the subject of extensive research
because the properties and reactivities of such compounds
make them suitable for numerous applications in different
fields such as agrochemicals and pharmaceuticals.[3] The pres-
ence of a CF3 moiety affects the properties of organic mole-
cules because the electronegativity and size of the group can
The asymmetric addition of dialkylzinc compounds ZnR2 to
aldehydes and ketones is a powerful methodology with which
to access optically active secondary and tertiary alcohols.[4]
However, direct alkylation of trifluoromethylketones[5] with
ZnEt2 (and higher alkyl groups) is very challenging because of
extensive formation of the undesired reduction product upon
addition of the organometallic reagent (Scheme 1).[6]
Figure 1. Representative compounds containing CF3.
Scheme 1. Competing addition vs. reduction reaction in the alkylation of tri-
fluoromethylketones.
[a] M. Calvillo-Barahona, Prof. Dr. J. A. Casares, Dr. C. Cordovilla,
Dr. J. M. Martꢀnez-Ilarduya, Prof. Dr. P. Espinet
IU CINQUIMA/Quꢀmica Inorgꢁnica
Facultad de Ciencias, Universidad de Valladolid
Paseo Belꢂn 5, 47011 Valladolid (Spain)
Fax: (+34) 983423013
Several examples of catalytic enantioselective addition of
ZnEt2 to trifluoromethylketones,[6,7] with high chemoselectivity
and up to 61% enantioselectivity at 196 K were achieved
through the use of the chiral ligand tBuBOX (tBuBOX=2,2’-iso-
propylidenebis[(4S)-4-tert-butyl-2-oxazoline]). In those studies,
an efficient but not asymmetric version of the reaction, cata-
lyzed by TMEDA (TMEDA=N,N,N’,N’-tetramethylethylenedi-
amine), was also developed. More recently, we have reported[8]
the enantioselective addition of ZnR2 (R=Me, Et) to 2,2,2-tri-
E-mail: jmi@qi.uva.es
espinet@qi.uva.es
[b] Dr. M. N. Genov
Sealife Pharma GmbH, Technopark I/Geb.B/EG
3430 Tulln (Austria)
Supporting information for this article is available on the WWW under
http://dx.doi.org/10.1002/chem.201403648 or from the author.
Chem. Eur. J. 2014, 20, 14800 – 14806
14800
ꢀ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Paper
diamines and the reactants and products. This study lead us to
propose a more complex catalytic cycle for the bulkier ligand L
and, in a practical aspect, to improve the chiral–ligand econo-
my in the enantioselective addition of ZnEt2 to PhC(O)CF3.
Results and Discussion
Identification of intermediates in the reactions
Monitoring the reaction of ZnEt2 (in [D8]toluene solution) with
PhC(O)CF3 (1), in the presence of the nonchiral ligand TMEDA
(10% relative to 1), at 244 K (for the detailed procedure see
the Supporting Information) by using 19F NMR spectroscopy,
revealed the formation of two main fluorinated products, as
shown in Scheme 3: [Zn(Et){OC(CF3)(Et)Ph}]2 (2), and
[(TMEDA)Zn(Et){OC(CF3)(Et)Ph}] (3a). The major product 2,
which is a dimer with trifluoromethyl carbinol bridges, can be
transformed quantitatively into the monomer 3a by addition
of TMEDA.
Figure 2. Molecular NÀN catalysts.
fluoroacetophenone (PhC(O)CF3) and several related ketones
by using the bulky chiral diamines L and ent-L (Figure 2) de-
rived from (R,R)- or (S,S)-1,2-diphenylethylenediamine and (S)-
or (R)-2,2’-bis-(bromomethyl)-1,1’-binaphthalene.[9] These li-
gands allowed us to prepare both enantiomers of the products
with excellent yield (up to 99%) and the highest enantioselec-
tivity reported so far (up to 92%), which was achieved by
working at 213 K. A drawback of the use of our chiral diamines
was the low addition rates observed, compared with those ob-
tained with tBuBOX, even for loadings of 10% ligand.
Some clues for the mechanism of these addition reactions
were provided by Hevia and co-workers in a nonchiral model
(TMEDA), showing that stoichiometric reactions of PhC(O)CF3
with [ZnR2(TMEDA)] produce alkylation (R=Me, Et) or reduc-
tion (R=tBu) products [(TMEDA)Zn(R){OC(CF3)(R/H)Ph}], and
their corresponding Zn(alkyl)(alkoxide) polymers.[10] These ob-
servations support the proposed mechanism shown in
Scheme 2.
As discussed below, preliminary experiments with the chiral
ligand L revealed the necessity of studying additional models
to explain the special features of the enantioselective catalysis
with L; this prompted us to explore in more detail the reac-
Scheme 3. Synthesis of alkyl(alkoxide) compounds 2 and 3a.
Consistent with the dimeric structure proposed for 2, the 19F
NMR spectrum (Figure 3) displays two singlets of very similar
intensity, corresponding to the three expected stereoisomers
of similar stability: two enantiomers (2RR, 2SS; ca. 25% each,
with the same chemical shift), and the meso diastereomer
(2RS; ca. 50%). Higher aggregates would be expected to show
very complex spectra, as observed for [Zn4(tBu)2{OC-
(CF3)(H)Ph}6],[10] and these are not observed. Compound 2 was
also prepared in independent experiments by the reaction of
PhC(OH)(CF3)(Et) (4) (either racemic or enantiopure) with ZnEt2
in [D8]toluene. When the reaction was carried out with either
(R)- or (S)-PhC(OH)(CF3)(Et) (4R or 4S) only one singlet was ob-
served in both 19F NMR spectra, which corresponds to 2RR or,
respectively, 2SS and allows for unambiguous assignment of
the two singlets as shown in Figure 3. The chemical shift of
this singlet coincided with the chemical shift of the most
shielded signal observed when racemic 4 was used in the syn-
thesis.
1
tions with TMEDA, tBuBOX, and L by using H and 19F NMR
spectroscopic methods to clarify the interaction between these
Monitoring the reaction of 4 with ZnEt2 by 19F NMR spectros-
copy showed that, during the reaction, broadening of the sig-
nals took place due to exchange of the alkoxide fragment {OC-
(CF3)(Et)Ph} between the dimer 2 and 4, as confirmed by a 19F
Scheme 2. Proposed cycle for alkylation of PhC(O)CF3 using TMEDA as cata-
lyst.
Chem. Eur. J. 2014, 20, 14800 – 14806
www.chemeurj.org
14801
ꢀ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Paper
sion that the enantioselectivity
of the processes can be estimat-
ed fairly accurately by 19F NMR
spectroscopy, without the assis-
tance of chiral GC analysis.
In the case of reactions cata-
lyzed by tBuBOX, the monomeric
[(tBuBOX)Zn(Et){OC*(CF3)(Et)Ph}]
(3b) can be detected by 19F
NMR spectroscopic analysis,
along with the signals of 2. The
defined chirality of the ligand
tBuBOX gives rise to diastereo-
mers:
[(tBuBOX)Zn(Et){S-[OC-
(CF3)(Et)Ph]}] 3bS, and the corre-
sponding 3bR. To unequivocally
assign their signals, 2SS (pre-
pared by reaction of the enan-
tiomerically
pure
PhC(OH)-
Figure 3. 19F NMR spectrum for the reaction of 1 and ZnEt2, using TMEDA (10%) as catalysts, in [D8]toluene at
244 K.
(CF3)(Et) (4S) with ZnEt2 in
[D8]toluene) was treated with
tBuBOX (1:2 ratio; Scheme 4).
EXSY experiment conducted at 293 K. Only when a full conver-
sion of 4 was achieved does the 19F NMR spectrum of the solu-
tion show two sharp singlets of similar intensity, which were
1
assigned to compounds 2. H NMR spectra were less informa-
tive because the signals of the ethyl groups of the two afore-
mentioned stereoisomers partially overlap. Nevertheless, these
spectra clearly show two distinct sets of resonances for the
ethyl groups directly linked to zinc atoms and those bonded
to carbon. In addition, an exchange process between 2 and
free ZnEt2 of the ethyl groups directly linked to zinc atoms
1
could be established by variable-temperature H NMR experi-
ments.
Scheme 4. Synthesis of 2SS and 3bS.
The enantioselective addition reaction of 1 with ZnEt2 in the
presence of a chiral ligand (NÀN: tBuBOX or L), under the
same conditions indicated for TMEDA, also reveals the forma-
tion of the dinuclear complexes [Zn(Et){OC(CF3)(Et)Ph}]2 (2) in
the 19F NMR spectra. In the two cases, the different abundan-
ces of the (R)- or (S)-[OC(CF3)(Et)Ph] fragments, generated in
the enantioselective addition, result in different intensity of the
signals of the groups of diastereomers observed (see the Sup-
porting Information). Considering that the statistical distribu-
tion of R and S fragments in the mixture is given by 2RR/2SS/
2RS=(%R)2:(%S)2:2(%R)(%S), the ratio of the integrals of the
singlets observed for the isomers of 2 can be correlated to the
enantioselectivity of the reaction, determined by GC analysis of
the tertiary alcohol obtained after hydrolysis (Table 1). The
good correlation of the values in Table 1 supports the conclu-
Analysis of the corresponding 19F NMR spectrum shows only
the more intense singlet of the two observed in the catalytic
reaction with tBuBOX (corresponding to the less shielded fluo-
rine), which allowed us to assign the compound to 3bS. In
contrast, in the case of reactions catalyzed by L, signals of the
expected [(L)Zn(Et){OC(CF3)(Et)Ph}] (3c) could not be detected,
suggesting that L is a much less effective coordinating ligand
than tBuBOX or TMEDA. Models suggest that the hindrance
between the two halves of L will make it difficult to coordinate
as a chelate to the same Zn atom. This is further supported by
the discussion on equilibria that follows.
Stoichiometric reactions
Table 1. Observed and calculated ratio of stereoisomers in 2
The catalytic cycle shown in Scheme 2, applied to NÀN ligands,
takes the shape of the transformations depicted in Equa-
tions (1)–(3). However, we have found that the behavior of the
three ligands differs substantially, which requires some consid-
erations.
Ligand
ee [%]
(confign)
(2RR+2SS)/2RS ratio
from NMR from GC
tBuBOX
L
55 (S)
83 (S)
1.8:1
5.7:1
1.9:1
5.4:1
Chem. Eur. J. 2014, 20, 14800 – 14806
www.chemeurj.org
14802
ꢀ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Paper
Equation (1)
tion (3) in the reverse mode, when 2 mol TMEDA is added to 2,
compound 3a is produced stoichiometrically; with tBuBOX, an
The initial coordination of NÀN to ZnEt2 gives rise to [(NÀ equilibrium between 3b and 2 is produced, but under the
N)ZnEt2] (5a–c), which are stronger Et nucleophiles than ZnEt2.
The addition of an equimolar amount of TMEDA or tBuBOX to
a solution of ZnEt2 in [D8]toluene results in the quantitative for-
mation of [(N-N)ZnEt2] {NÀN=TMEDA (5a), tBuBOX (5b)}, which
were observed in the NMR spectra. In contrast, for NÀN=L, the
corresponding complex was not observed as such. However,
addition of ZnEt2 to the solution produced a shift of the sig-
nals (particularly those of L), suggesting that in this case Equa-
tion (1) corresponds to a fast dissociation equilibrium. This dif-
ferent behavior of L is easily understood considering that, be-
cause of excessive steric congestion, L coordinates in a mono-
dentate fashion and not as chelate, which makes L dissociation
require much lower activation energy than the dissociation of
chelating ligands. The coordination equilibrium constant for L
was calculated from the shift of the more shielded protons of
the CH2 groups of L, which afforded K=1.9 molÀ1 LÀ1 at 244 K
(compared with quantitative formation for TMEDA and
tBuBOX). Clearly, this equilibrium results in a much lower effec-
tive concentration of the active nucleophilic species [(L)ZnEt2]
(5c).
same conditions L does not produce detectable amounts of
the expected complex 3c.
Catalytic reactions: singular behavior of L
From the experiments discussed so far, the catalytic cycle
shown in Scheme 2 explains perfectly the behavior observed
for TMEDA and tBuBOX. These chelating ligands activate effi-
ciently the nucleophilicity of the Et groups in their complexes
and give rise to efficient fast reactions. Chelate complexes are
observed as starting materials or as intermediates.
In contrast, for ligand L, the putative complexes [LZnEt2] or
[LZn(Et){OC(CF3)(Et)Ph}] are not observed (although the forma-
tion of [LZnEt2] in equilibrium with free L can be deduced
from chemical shifts in the NMR spectra). Assuming that L can
act only in a monodentate mode towards a single Zn atom,
and that the concentration of complexed ZnEt2 must be very
small (much lower than for the chelating ligands), it is easy to
understand that a cycle similar to Scheme 2 (cycle A in
Scheme 5) should be quite inefficient. In fact, the addition re-
Equation (2)
The stoichiometric addition reaction of [(NÀN)ZnEt2] to 1 is
quantitative and fast for TMEDA and tBuBOX, affording [(NÀ
N)Zn(Et){OC(CF3)(Et)Ph}] (3a–b). However, this reaction does
not occur with L, and the expected complex 3c is not detect-
ed. This is an intriguing observation considering that the same
reaction works under catalytic conditions using lower L/Zn
ratios, as discussed later.
Scheme 5. Initial catalytic cycle for the addition reaction of 1 with ZnEt2 and
L. R=C(CF3)(Et)Ph.
Equation (3)
Finally, the rearrangement of intermediates 3a–b with ZnEt2 to
produce alkoxy product 2 and generate new alkylating reagent
5a–b in equilibrium occurs easily. This equation was studied
for 3a by using isolated complex; the equilibrium constant in
action catalyzed with L is initially very slow, but further catalyt-
ic studies showed that it was accelerated to achieve good rate
after an activation period. Our proposal is that cycle A is the
initial pathway, which follows the reaction during the activa-
tion period until faster cycle B in Scheme 6 takes the lead. Fur-
ther features of the catalysis with L are the observation of
enantioselective autoinduction,[12] and an inverse dependence
of the reaction rate on L loading. These highly unusual catalyt-
ic properties deserve an explanation.
[D8]toluene at 244 K was K=0.87 molÀ1/2 L1/2 [11]
For 3b the ten-
.
dency to produce 2 was much higher, so that it is produced
spontaneously even under the conditions of Equation (2) (i.e.,
in the absence of free ZnEt2). In other words, complex 3b re-
leases tBuBOX spontaneously. Since Equation (2) does not
work for L, this process could not be studied. Looking at Equa-
Chem. Eur. J. 2014, 20, 14800 – 14806
www.chemeurj.org
14803
ꢀ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Paper
very small because it depends on the amount of 2, produced
in the inefficient cycle. Although always low, the concentration
of 6 will progressively increase with time as cycle B proceeds,
producing an autocatalytic effect.
For a given concentration of 2 (at a given moment of the
catalysis) one might expect from Equation (4) that increasing
the loading of L should increase the concentration of 6 and
make the reaction faster. However, the opposite effect is ob-
served, as shown in Figure 4: Higher loading of L retards the
reaction. In fact, when loading of L reaches the proportion Zn/
L=1:1, the addition reaction does not proceed at all, as com-
mented in the study of stoichiometric reactions above. This ap-
parent paradox can be explained when we consider that the
predictions derived from consideration of the thermodynamic
equilibria in a static situation cannot be applied to a dynamic
system in which the concentration of 6 is not determined by
thermodynamics but corresponds to a steady-state concentra-
tion. Intermediate 6 contains bridging bonds that can be split
by extra L. If the formation of 6 is slow and its destruction by
additional L is faster, in some instances the increase in concen-
tration of L will diminish the steady-state concentration of the
catalyst 6, hence the reaction rate. The real system is far too
complex to describe mathematically considering the lack of
quantitative experimental data, but a simpler model with arbi-
trary data has been calculated as a proof of concept and is
given in the Supporting Information.
Scheme 6. Faster catalytic cycle operating after induction for the addition re-
action of 1 with ZnEt2 and L. R=C(CF3)(Et)Ph
Induction time and kinetics
According to the conventional mechanism presented in
Scheme 2, an increase in the catalyst (ligand) loading is expect-
ed to induce an increased reaction rate because it increases
the concentration of the reactive complex. This behavior has
been confirmed for the cases of TMEDA and tBuBOX, but kinet-
ic studies on reactions with L reveal a striking inverse depend-
ence of the reaction rate on catalyst loading (Figure 4).
The shapes of the curves in Figure 4 reveal an autocatalytic
process. For a better understanding, three successive catalytic
cycles were performed (Figure 5). The first catalytic run was
monitored by 19F NMR spectroscopy using PhC(O)CF3 (1),
ZnEt2, and L (1:1.2:0.02) in [D8]toluene at 244 K and the plot
showed the typical ‘S’-shape of autocatalysis. After completion,
an additional equivalent of PhC(O)CF3 and 1.2 equivalents of
ZnEt2 were added (second run) and, when the reaction was
Figure 4. Catalytic cycles for the addition reaction of 1 with ZnEt2 (1:1.2) and
different amounts of L. Starting concentrations: [1]=0.20 M, [ZnEt2]=0.24 M
(3, 6.5, and 12% L); [1]=0.10 M, [ZnEt2]=0.12 M (19% L).
Equation (4)
It is clear that cycle A (Scheme 5) cannot account for the kinet-
ic behavior of L. This observation prompted us to consider al-
ternative pathways involving active species with Zn/L ratios
higher than 1. Thus, we propose that the reaction of L, 2, and
ZnEt2 generates a small concentration of the dinuclear inter-
mediate 6 [Eq. (4)]. This opens the alternative pathway B
(Scheme 6), which turns out to be faster. This faster addition is
to be expected because the Et groups are more nucleophilic in
the tetracoordinate ZnEt2 fragment of 6 than in tricoordinate
5c. Note that initially the amount of 6 available for catalysis is
Figure 5. Variation of the concentration of PhC(O)CF3 (1) for the addition re-
action of 1, ZnEt2 and L (1:1.2:0.02) in three successive catalytic cycles after
the addition of 1 (1 equiv.) and ZnEt2 (1.2 equiv.).
Chem. Eur. J. 2014, 20, 14800 – 14806
www.chemeurj.org
14804
ꢀ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Paper
complete, another equivalent of PhC(O)CF3 and 1.2 equivalents
of ZnEt2 were added to the reaction mixture (third run). The
second and third runs showed a higher reaction rate and
a normal shape, confirming that the more active catalyst (6 in
Scheme 6) had been formed during the first run and remains
active in the subsequent runs. In other words, once catalyst 6
had been formed, the catalytic cycle B becomes kinetically
dominant.
tion S6 of the Supporting Information), indicating that the two
diastereomers of 6 are almost equally active. In sharp contrast
with the reaction in the absence of added alcohol 4, in these
three experiments the induction time was significantly reduced
(although not totally suppressed, showing that formation of 6
is slow and still being formed), and the reaction was complete
after 3 h instead of 20 h. The reaction produced (S)-PhC(OH)-
(CF3)(Et) (4S) as the major product after hydrolysis, showing
that the conformation of the bridging alcohol in 6 has little in-
fluence on the conformation of the product. The ee in the
newly formed alcohol (eeprod) depends mostly on the configura-
tion of L. The enantioselectivity found for the initial addition of
one equivalent of 4R, 4rac, or 4S, calculated discounting the
originally added chiral carbinol was eeprod =80.9, 86.1, and 91.1,
respectively, in all cases of the (S)-alcohol. Considering that
these figures come from just one run, in a normal catalysis
with small loads of catalyst and many turnovers, the effect on
enantioselectivity of feeding initially with chiral carbinol would
be negligible. Moreover, the accelerating effect of initial feed-
ing with alcohol requires a significant amount of this addition-
al reagent, which would eventually contaminate the product
unless it is the same alcohol being produced. Thus, currently,
this accelerating procedure has no clear synthetic application
and it is more convenient to use successive runs as indicated
in Figure 5.
Enantioselective autoinduction
Catalytic reactions were performed by using PhC(O)CF3 (1),
ZnEt2 (1:1.2), and a L catalyst loading as low as 2%, in
[D8]toluene at 253 K. Aliquots were periodically quenched and
the ee and chemical yields were determined, respectively, by
GC and 19F NMR analyses. A significant enantioselectivity en-
hancement was observed as the reaction proceeded, from
a moderate initial ee of 55% to a highly enantioselective value
of 88% when full conversion of the ketone was reached
(Figure 6). According to these results, the more active catalyst
Conclusions
The NMR study of the catalytic addition reaction of ZnEt2 with
PhC(O)CF3 in the presence of three very efficient catalysts
(TMEDA, tBuBOX, and L) reveals clear differences in their be-
havior. The 19F NMR signals of the main product, [Zn(Et){OC-
(CF3)(Et)Ph}]2, allow a fairly accurately estimation of the enan-
tioselectivity of the process for the chiral ligands. By far the
more enthralling behavior corresponds to the bulky ligand L.
The observation of an autocatalytic asymmetric enhancement
during the reaction and an unusual concentration dependence
on the reaction rate support the participation, for this ligand,
of a catalytic cycle additional to the one operating for the
other two ligands. This second cycle, which is favored by the
coordination preferences of L, explains the unique behavior of
ligand L and the increased efficiency of the process operating
via a dinuclear intermediate with bridging alkoxy and L
groups. The improved procedure reported here provides
a new synthetic record as it affords basically the same ee (93%
vs. the previous 92%) but at a higher temperature (244 vs.
213 K), thanks to reduction of the percentage of catalyst used,
which produces a stronger influence of the autocatalytic asym-
metric enhancement effect of ligand L.
Figure 6. Autocatalytic asymmetric enhancement in the reaction 1/ZnEt2/L
(1:1.2:0.02) in [D8]toluene at 253 K. Initial volume of the reaction: 5 mL; ini-
tial concentration of PhC(O)CF3: 0.144 M; volume (constant) of the aliquot
periodically extracted=0.2 mL.
6, formed during the reaction, is also more enantioselective
than the initial slow catalyst (5c in Scheme 5). This looks rea-
sonable considering that 6 has a more rigid structure that de-
fines better the space around Zn than the flexible and easy to
dissociate complex 5c.
As a matter of fact, catalyst 6 can show two diastereomers
and its chiral conformation is defined by the chirality of L,
which is fixed, and by the chiral conformation of the alkoxy
group (S or R), which is being created competitively during the
catalysis. To determine the effect of the chiral conformation of
the alkoxy group in the enantioselectivity of the products,
three different experiments were made to produce 6 directly
from their components (avoiding cycle A) and were monitored
by 19F NMR spectroscopy in [D8]toluene at 244 K: three NMR
tubes were charged, each with one equivalent of (S)-, (R)-, or
rac-PhC(OH)(CF3)(Et) (4), and a mixture of ZnEt2 and L was
added (2.4:0.02); after formation of the dimer 2 and complete
disappearance of the free alcohol 4 (ca. 20 min) one equivalent
of PhC(O)CF3 (1) was added. The kinetic profiles of the three
experiments overlapped almost perfectly (see plot in sec-
Experimental Section
Improved procedure for the enantioselective addition of
ZnEt2 to 2,2,2-trifluoroacetophenone
Diethylzinc (1.0 M in toluene, 0.58 mL, 0.58 mmol) was added to
a solution of L (7.4 mg, 0.0096 mmol, 2 mol%) in anhydrous tolu-
Chem. Eur. J. 2014, 20, 14800 – 14806
www.chemeurj.org
14805
ꢀ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim

Full Paper
Chem. 2007, 11, 127–157; e) M. Hatano, K. Ishihara, Synthesis 2008,
1647–1675; f) M. Shibasaki, M. Kanai, Chem. Rev. 2008, 108, 2853–2873;
g) M. R. Luderer, W. F. Bailey, M. R. Luderer, J. D. Fair, R. J. Dancer, M. B.
Sommer, Tetrahedron: Asymmetry 2009, 20, 981–998; h) R. Somanathan,
L. Z. Flores-López, R. Montalvo-Gonzµlez, D. Chµvez, M. Parra-Hake, G.
Aguirre, Mini-Rev. Org. Chem. 2010, 7, 10–22; i) M. Hatano, R. Gouzu, T.
Mizuno, H. Abe, T. Yamada, K. Ishihara, Catal. Sci. Technol. 2011, 1,
1149–1158; j) C. M. Binder, B. Singaram, Org. Prep. Proced. Int. 2011, 43,
139–208.
ene (1 mL) under argon at RT. The solution was stirred for 5 min,
then 2,2,2-trifluoroacetophenone (0.48 mmol) was added at 244 K
and this temperature was maintained. When the reaction was com-
plete, saturated ammonium chloride solution was added and the
mixture was extracted with diethyl ether. The solvents were re-
moved under reduced pressure and the crude product was puri-
fied by chromatography on silica gel (pentane/dichloromethane,
3:1) to give 1,1,1-trifluoro-2-phenylbutan-2-ol (95% yield; 93%ee)
as a colorless oil.
[5] C. B. Kelly, M. A. Mercadante, N. E. Leadbeater, Chem. Commun. 2013,
49, 11133–11148.
[6] K. Yearick, C. Wolf, Org. Lett. 2008, 10, 3915–3918 and references there-
in.
[7] K. Higashiyama, S. Sasaki, H. Kubo, T. Yamauchi, A. Ishii, M. Kanai, Japa-
nese Patent 200609692, April 13, 2006.
[8] a) M. Genov, J. M. Martínez-Ilarduya, M. Calvillo-Barahona, P. Espinet, Or-
ganometallics 2010, 29, 6402–6407; b) M. Calvillo-Barahona, C. Cordovil-
la, M. N. Genov, J. M. Martínez-Ilarduya, P. Espinet, Dalton Trans. 2013,
42, 14576–14582; c) P. Espinet, M. N. Genov, J. M. Martínez-Ilarduya, M.
Calvillo-Barahona Spanish Patent P200931196, December 17, 2009.
[9] T. Arai, M. Watanabe, A. Yanagisawa, Org. Lett. 2007, 9, 3595–3597
[10] E. Hevia, A. R. Kennedy, J. Klett, Z. Livingstone, M. D. McCall, Dalton
Trans. 2010, 39, 520–526.
Acknowledgements
The authors are grateful for financial support by the Spanish
MINECO (CTQ2013-48406-P) and the Junta de Castilla y León
(GR169 and VA256U13). C.C. acknowledges a Juan de la Cierva
grant from the MICINN.
Keywords: alkylation · asymmetric catalysis · autocatalysis ·
ligand effects · zinc
[11] In the equilibrium constant for Eq. (3), ZnEt2 is taken as monomeric ac-
cording to the recent X-ray determination, which shows a linear mono-
mer in the crystal: J. Bacsa, F. Hanke, S. Hindley, R. Odedra, G. R. Darling,
A. C. Jones, A. Steiner, Angew. Chem. 2011, 123, 11889–11891; Angew.
Chem. Int. Ed. 2011, 50, 11685–11687. Note that the value of K we are
defining is only for practical purposes. Clearly it is likely that in solution
ZnEt2 will contain coordinated ketone (in fact the chemical shift of the
ketone changes in the presence of Zn species), which we do not specify
for simplicity, but this will not change the nuclearity of the Zn species.
[12] a) H. Danda, H. Nishikawa, K. Otaka, J. Org. Chem. 1991, 56, 6740–6741;
b) E. F. Kogut, J. C. Thoen, M. A. Lipton, J. Org. Chem. 1998, 63, 4604–
4610; c) M. Szlosek, B. Figadere, Angew. Chem. 2000, 112, 1869–1871;
Angew. Chem. Int. Ed. 2000, 39, 1799–1801; d) M. De Rosa, M. R. Acocel-
la, A. Soriente, A. Scettri, Tetrahedron: Asymmetry 2001, 12, 1529–1531;
e) A. M. Costa, C. García, P. J. Carroll, P. J. Walsh, Tetrahedron 2005, 61,
6442–6446; f) J. M. Brown, I. Gridnev, J. Klankermayer, Top. Curr. Chem.
2008, 284, 35–65; g) K. Soai, T. Kawasaki, Catalytic Asymmetric Synthesis,
3edrd edWiley, 2010, pp. 891–930.
[1] a) J. A. Dale, D. L. Dull, H. S. Mosher, J. Org. Chem. 1969, 34, 2543–2549;
b) V. V. Grushin, Acc. Chem. Res. 2010, 43, 160–171 and references
therein; c) O. A. Tomashenko, V. V. Grushin, Chem. Rev. 2011, 111, 4475–
4521 and references herein.
[2] a) J. Nie, H.-C. Guo, D. Cahard, J.-A. Ma, Chem. Rev. 2011, 111, 455–529;
b) F. Tur, J. Mansilla, V. J. Lillo, J. M. Saµ, Synthesis 2010, 1909–1923 and
references herein; c) A. V. R. Madduri, S. R. Harutyunyan, A. J. Minnaard,
Angew. Chem. 2012, 124, 3218–3221; Angew. Chem. Int. Ed. 2012, 51,
3164–3167; d) R. Luo, K. Li, Y. Hu, W. Tang, Adv. Synth. Catal. 2013, 355,
1297–1302; e) V. Valdivia, I. Fernµndez, N. Khiar, Org. Biomol. Chem.
2014, 12, 1211–1214.
[3] a) A. M. Thayer, Chem. Eng. News 2006, 84, 15–24; b) C. Isanbor, D.
O’Hagan, J. Fluorine Chem. 2006, 127, 303–319; c) J.-P. BØguØ, D.
Bonnet-Delpon, J. Fluorine Chem. 2006, 127, 992–1012; d) K. L. Kirk, J.
Fluorine Chem. 2006, 127, 1013–1029; e) K. Müller, C. Faeh, F. Diederich,
Science 2007, 317, 1881–1886; f) The Pesticide Manual: A World Com-
pendium (16th ed.), version 3.2 (Ed. C. D. S. Tomlin), British Crop Produc-
tion Council, 2012.
[4] a) K. Soai, S. Niwa, Chem. Rev. 1992, 92, 833–856; b) L. Pu, H.-B. Yu,
Chem. Rev. 2001, 101, 757–824; c) M. Yus, D. J. Ramón, Pure Appl. Chem.
2005, 77, 2111–2119; d) M. Hatano, T. Miyamoto, K. Ishihara, Curr. Org.
Received: May 22, 2014
Published online on September 15, 2014
Chem. Eur. J. 2014, 20, 14800 – 14806
www.chemeurj.org
14806
ꢀ 2014 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim