Highly enantioselective addition of dialkylzinc to trifluoroacetophenones, catalyzed by 1,2-diamines. synthesis of key fragments of inhibitors of the enzyme 11β-HSD1 and kinetic analysis of the process

Article abstract of DOI:10.1021/om1007727

Chiral diamines derived from (R,R)- or (S,S)-1,2-diphenylethylenediamine and (S)- or (R)-2,2′-bis(bromomethyl)-1,1′-binaphthalene perform very well as catalysts in the enantioselective addition of ZnEt₂ and ZnMe₂ to trifluoroacetophenones, avoiding reduction products, in contrast to the poor results with amino alcohols. Excellent yields (up to 99%) and high enantioselectivity (up to 92%) are achieved with the best ligands. A kinetic study at low temperature (-37 °C) shows that the reduction reaction rate on ligandless ZnEt₂ is negligible (2 orders of magnitude slower) compared to the rate of addition reaction on [ZnR₂(N-N)]. Using this new procedure, reported fragments of inhibitors of enzyme 11β-HSD1 that are active against obesity and type 2 diabetes mellitus, as well as new unreported modified fragments of these bioactive molecules, were produced efficiently and enantioselectively.

Full text of DOI:10.1021/om1007727

6
402 Organometallics 2010, 29, 6402–6407
DOI: 10.1021/om1007727
Highly Enantioselective Addition of Dialkylzinc to Trifluoroacetophenones,
Catalyzed by 1,2-Diamines. Synthesis of Key Fragments of Inhibitors of the
†
Enzyme 11β-HSD1 and Kinetic Analysis of the Process
Miroslav Genov, Jes uꢀ sM. Martı
´
nez-Ilarduya, Mercedes Calvillo-Barahona, and Pablo Espinet*
IU CINQUIMA/Quı ´m ica Inorg aꢀ nica, Facultad de Ciencias, Universidad de Valladolid, E-47071 Valladolid, Spain
Received August 9, 2010
0
Chiral diamines derived from (R,R)- or (S,S)-1,2-diphenylethylenediamine and (S)- or (R)-2,2 -bis-
0
(bromomethyl)-1,1 -binaphthalene perform very well as catalysts in the enantioselective addition of ZnEt2
and ZnMe to trifluoroacetophenones, avoiding reduction products, in contrast to the poor results with
2
amino alcohols. Excellent yields (up to 99%) and high enantioselectivity (up to 92%) are achieved with the
best ligands. A kinetic study at low temperature (-37 °C) shows that the reduction reaction rate on
ligandless ZnEt is negligible (2 orders of magnitude slower) compared to the rate of addition reaction on
2
[
active against obesity and type 2 diabetes mellitus, as well as new unreported modified fragments of these
ZnR (N-N)]. Using this new procedure, reported fragments of inhibitors of enzyme 11β-HSD1 that are
2
bioactive molecules, were produced efficiently and enantioselectively.
2
-9
Introduction
active secondary and tertiary alcohols, respectively.
The
group of Toru and Shibata reported recently a nucleophilic
enantioselective trifluoromethylation of aryl ketones using a
Trifluoromethyl-substituted tertiary alcohols are important
substructures in many biologically active natural products and
1
0
Cinchona alkaloid/TMAF combination. Noyori’s group
developed some efficient addition reactions to conventional
1
in synthetic pharmaceutical compounds. For instance, the
5,11
chiral molecule p-(HO C)-C H -C(CF )(Me)OH is the head
2
aldehydes, catalyzed by the presence of amino alcohols.
Alternatively, the direct alkylation of trifluoromethyl ketones
6
4
3
of a family of products presently being investigated as inhibitors
of enzyme 11β-HSD1, against obesity and type 2 diabetes
mellitus. This compound is obtained at present by a modest-
with organometallic reagents is practical for ZnMe but fails
2
for higher alkyls (e.g., ZnEt ), due to extensive reduction via
2
1
b
yielding multistep process. Other slightly modified mole-
cules that might head new active series, such as p-(HO C)-
β-hydride elimination. The reduction reaction is initiated
by nucleophilic attack of the ethyl β-H to the carbonyl
carbon atom, which makes the problem particularly acute
2
1
2
C H -C(CF )(Et)OH, have not been reported so far.
6
4
3
The asymmetric addition of dialkylzincs to aldehydes and
ketones is today a powerful method of access to optically
for ketones with electron-withdrawing substituents, such as
3
1
trifluoromethyl ketones.
Only recently have more extensive studies facing the reduc-
14,15
tion problem in trifluoromethyl ketones been reported,
†
This paper is dedicated to Prof. Jos ꢀe Barluenga on occasion of his
7
0th birthday.
To whom correspondence should be addressed. E-mail: espinet@
qi.uva.es.
achieving very efficient additions catalyzed by TMEDA (1;
Figure 1) and other nonchiral chelating ligands. The screening
of a high number of chiral ligands (mostly diamines and bis-
oxazolines) attained the best result using TBOX (2; Figure 1) as
a chiral ligand: 37% enantioselectivity for the catalytic enantio-
selective addition of ZnEt2 to 1,1,1-trifluoroacetophenone,
almost without reduction byproduct. The enantioselectivity
increased to 61% at -78 °C. Ligand 3 (Figure 1) also catalyzes
*
(1) See for example: (a) Xue, Y.; Chao, E.; Zuercher, W. J.; Willson,
T. M.; Collins, J. L.; Redinbo, M. R. Bioorg. Med. Chem. 2007, 15, 2156–
2
166. (b) Julian, L. D.; Wang, Z.; Bostick, T.; Caille, S.; Choi, R.;
DeGraffenreid, M.; Di, Y.; He, X.; Hungate, R. W.; Jaen, J. C.; Liu, J.;
Monshouwer, M.; McMinn, D.; Rew, Y.; Sudom, A.; Sun, D.; Tu, H.; Ursu,
S.; Walker, N.; Yan, X.; Ye, Q.; Powers, J. P. J. Med. Chem. 2008, 51, 3953–
3
960. (c) Powers, J.; Degraffenreid, M.; Julian, L.; Kaizerman, J.; McMinn,
D.; Rew, Y.; Sun. D.; Yan, X.; Wang, Z. Patent WO 2007/145835 A2, 2007.
2) For a comprehensive review, see: Pu, L.; Yu, H.-B. Chem. Rev.
001, 101, 757–824 and references cited therein.
the addition of ZnEt to trifluoroacetophenone, giving the
2
(
2
(
3) Soai, K.; Niwa, S. Chem. Rev. 1992, 92, 833–856.
(10) Mizuta, S.; Shibata, N.; Akiti, S.; Fujimoto, H.; Nakamura, S.;
Toru, T. Org. Lett. 2007, 9, 3707–3710.
(4) Oguni, N.; Omi, T.; Yamamoto, Y.; Nakamura, A. Chem. Lett.
1
983, 841–842.
5) Kitamura, M.; Suga, S.; Kawai, K.; Noyori, R. J. Am. Chem. Soc.
986, 108, 6071–6072.
(11) Yamakava, M.; Noyori, R. J. Am. Chem. Soc. 1995, 117, 6327–
6335.
(
1
(12) (a) Coates, G. E.; Ridley, D. J. Chem. Soc. A 1966, 1064–1069.
(b) Arnott, G.; Hunter, R. Tetrahedron 2006, 62, 992–1000.
(13) Sasaki, S.; Yamauchi, T.; Kubo, H.; Kanai, M.; Ishii, A.;
Higashiyama, K. Tetrahedron Lett. 2005, 46, 1497–1500.
(14) Higashiyama, K.;Sasaki, S.; Kubo, H.; Yamauchi, T.; Ishii, A.;
Kanai, M. Japanese Patent 200609692, April 13, 2006.
(15) Yearick, K.; Wolf, C. Org. Lett. 2008, 10, 3915–3918 and refer-
ences cited therein.
(
(
6) Ramon, D.; Yus, M. Angew. Chem., Int. Ed. 2004, 43, 284–287.
7) Forrat, V. J.; Prieto, O.; Ramon, D. J.; Yus, M. Chem. Eur. J.
2006, 12, 4431–4445.
8) Garcia, C.; Larochelle, L. K.; Walsh, P. J. J. Am. Chem. Soc. 2002,
124, 10970–10971.
9) Jeon, S.-J.; Li, H.; Walsh, P. J. J. Am. Chem. Soc. 2005, 127,
16416–16425.
(
(
pubs.acs.org/Organometallics
Published on Web 11/03/2010
r 2010 American Chemical Society

Article
Organometallics, Vol. 29, No. 23, 2010 6403
Figure 1
Figure 3
Scheme 1
The reactions produced addition and reduction products,
depending on the conditions (Scheme 1). For ZnMe2 the
reactions are usually noticeably slower than for ZnEt and
Figure 2
2
desired 1,1,1-trifluoro-2-phenylbutan-2-ol in 97% yield but
only 6% ee.
can give only the addition product 9b. For the more nucleophilic
ZnEt the addition to give 9a competes, depending on the
15
2
We considered that higher enantioselectivity might be induced,
while still retaining reasonable reactivity, with chiral ligands
containing the same N-N skeleton as 3 but having bulkier
substituents at the nitrogen atoms that would better match the
ligand, with the reduction reaction to give 10. The results of
the preparative reactions are shown in Table 1, which includes
for comparison the best literature results so far (those of the
chiral ligand 2, entries 5 and 6). Table 1 also includes (in italics)
the results of the NMR-monitored reactions, discussed later.
The results collected in Table 1 show clearly that diamine
N-N chelating ligands are, with the notable exception of
ligand 5, much more effective than N-O amino alcohols, for
the addition of either ZnMe or ZnEt to trifluoroacetophe-
none. Among them, ligands 4 and 6 provide the best yields
and the best chemo- and enantioselectivities ever reported
for the reaction with ZnEt2.
In Table 1, different yields and enantiomeric excesses are
found for the addition product 9 in the reactions with N-N*,
whereas the reduction product 10, when observed, was
16
reagents. A patent on these synthetic results has been filed.
Results and Discussion
a). Catalytic Syntheses. The reactions of ZnR (R=Me, Et)
(
2
2
2
with trifluoroacetophenone were studied, either in the absence of
catalytic ligand or in the presence of nonchiral (1) orchiral(4-6)
N-N chelating ligands, using the ratio ZnR /trifluoroacetophe-
2
none/N-N = 1.2/1/0.1 (10 mol % of N-N relative to the
ketone; 8.3 mol % relative to Zn). Ligands 4 (and its enantiomer
ent-4), 5, and 6 (Figure 2) were synthesized from the commer-
cially available sources (1R,2R)- or (1S,2S)-1,2-diphenylethane-
always racemic. For comparison, in the absence of any ligand,
“
0
0
1
,2-diamine, (R)- or (S)-2,2 -dimethyl-1,1 -binaphthyl, and
23
ligandless” ZnEt2 reacts easily (although slowly) with trifluor-
17-19
1,2-bis(bromomethyl)benzene, using literature procedures.
Ligands 4 and 6 have been reported to be efficient in the
enantioselective nitro-aldol reaction.
The additions of ZnEt and ZnMe to trifluoroacetophenone
oacetophenone 11, giving almost exclusively the reduction pro-
duct 10 in 96% yield, plus a small proportion (3%) of racemic
9 (Table 1, entry 1). This suggests that, on ligandless ZnEt , β-H
2
elimination is much faster than addition. In the same line, in the
case of the slower reagent ZnMe (where reduction is excluded)
2
no product was detected after 24 h at room temperature (Table 1,
entry 2), showing that the expected addition of ligandless ZnMe2
1
7
2
2
were also checked with Peric ꢁa s’ type N-O amino alcohol
ligands (Figure 3). The commercial chiral amino alcohol 7,
which is remarkably efficient in addition reactions to alde-
20,21
hydes,
22
was chosen as standard. The amino alcohol 8, with
to ketone does not occur at a perceptible rate.
The addition reaction is activated in the presence of ligands.
the larger N substituents used in the N-N ligands 4-6, was
synthesized by us and tested for the reaction.
Thus, using 10% of the nonchiral ligand TMEDA (1) afforded
high addition reaction rates (for ZnEt ), as well as high addition
2
(
16) Espinet, P.; Genov, M. N.; Martı
Barahona, M. P200931196, 2009.
17) Arai, T.; Watanabe, M.; Yaganisawa, A. Org. Lett. 2007, 9,
595–3597.
´
nez-Ilarduya, J. M.; Calvillo-
vs reduction selectivity, although, obviously, with no enantios-
electivity (Table 1, entries 3 and 4).
(
Enantioselective results can be obtained using chiral N-N*
3
(
(
(
18) Maigfot, N.; Mazaleyrat, J.-P. Synthesis 1985, 317–320.
19) Mazaleyrat, J.-P. Tetrahedron: Asymmetry 1997, 8, 2709–2721.
20) Garcı
´
a-Delgado, N.; Fontes, M.; Peric ꢁa s, M. A.; Riera, A.;
ligands, from addition reactions occurring on ZnR (N-N*)
2
Verdaguer, X. Tetrahedron: Asymmetry 2004, 15, 2085–2090 and refer-
ences cited therein.
2
(23) By “ligandless” ZnR we mean molecules that are not coordi-
nated by N-N added ligand, although most likely the Zn center is
coordinated initially with molecules of the ketone reagent, forming
[ZnR (trifluoroacetophenone) ] complexes, and later with alcohol- or
2 n
alkoxy-coordinated products; both oxygen ligands are weaker coordi-
nating molecules than the N-N chelating ligands.
(
21) Genov, M.; Salas, G.; Espinet, P. J. Organomet. Chem. 2008,
017–2020.
22) See the Experimental Section and the Supporting Information
for details.
2
(

6
404 Organometallics, Vol. 29, No. 23, 2010
Genov et al.
Table 1. Results of the Addition of Diethyl- or Dimethylzinc to
a
Trifluoroacetophenone and its Para-Substituted Derivatives
30% yield and very low ee (about 4%) and reduction product
10 in 58% yield (Table 1, entry 10). This is in contrast with
the good results obtained with 4, 1, and 2 and is somewhat
c
9
ee (%)
10
b
c
c
closer to the result obtained on ligandless ZnEt . The reac-
2
entry ZnR
ZnEt
2
Z
L
T (°C) t (h) (%) (confign) (%)
tion showed some improvement at -37 °C (addition/ reduc-
tion = 36/60; Table 1, entry 10m) but was very slow and still
had very poor enantioselectivity (5% ee). The addition of
1
1
2
3
3
4
5
6
7
7
8
9
1
1
1
1
1
1
1
1
1
1
1
1
2
2
2
2
2
2
2
2
2
H
-37/rt 10
3
96
82
m
-37
rt
52
24
ZnMe
ZnEt
2
2
H
H
0
95
99
97
95
85
90
92
97
98
30
36
85
97
95
90
12
9
ZnMe catalyzed with 10% 5, however, showed noticeably
2
2
1
-37/rt
-37
2.5
4
n.o.
e
better results (85% yield, 29% ee, in 24 h), although it was
still far from the result with 4 (Table 1, entry 11 vs entry 9).
For ligand 6, the N substituents at one side of the ligand
are smaller than in 4, thus reducing the crowding in the complex
around the Zn center. This difference leads to a notable loss of
m
8
10
ZnMe
H
H
H
H
1
2
2
4
rt
d
d
e
e
ZnEt
ZnEt
ZnEt
2
2
2
-35
-78
0.5
2
51
61
n.o.
n.o.
8
-37/rt 10
82 (S)
84 (S)
92 (S)
83
m
-37
-60
15
48
5
n.o.
e
enantioselectivity in the reaction with ZnMe (83% ee for 4
2
ZnEt
ZnMe
ZnEt
2
H
H
H
4
4
5
versus 15% for 6), but not so much in the reactions with the
bulkier ZnEt (82% ee for 4 versus 75% for 6, at room tem-
2
2
-37/rt 24
-37/rt 10
0
2
4
58
2
0m
1
-37
50
5
29
60
perature) (Table 1, entries 12 and 13). The moderate loss of
enantioselectivity in the reaction with ZnEt using 6 instead
ZnMe
ZnEt
H
H
5
6
-37/rt 24
-37/rt 10
2
2
2
75 (S)
76 (S)
15
<1
<1
of 4 is somehow compensated by its improved chemoselec-
tivity (almost complete suppression of reduction at room
temperature), which rates the performance of 6 as excellent.
For the case of amino alcohol 7 as ligand, the main product
of the reaction with ZnEt (Table 1, entry 14) was the reduction
2
compound 10 (86%); only a 12% yield of 9a was obtained,
although with good enantioselectivity (82% ee). The reaction
with ZnMe was worse: only 9% yield in 9b and 28% ee after
2
24 h (Table 1, entry 15). With the amino alcohol 8, with larger N
substituents, the results were even more disappointing (Table 1,
2m
3
-37
25
ZnMe
ZnEt
ZnMe
ZnEt
2
2
H
H
H
H
H
6
7
7
8
8
4
4
4
4
4
4
4
4
4
4
-37/rt 24
-37/rt 24
-37/rt 24
-37/rt 10
-37/rt 24
-37/rt 24
-37/rt 24
-37/rt 24
-37/rt 24
-37/rt 24
-37/rt 24
4
2
82 (R)
28
86
94
5
6
2
6
5
7
ZnMe
ZnMe
ZnMe
ZnMe
ZnMe
ZnMe
ZnMe
ZnMe
ZnMe
ZnMe
ZnMe
2
2
2
2
2
2
2
2
2
2
2
15
85
90
95
95
95
95
95
97
97
99
rac
26
8
COO-t-Bu
COO-i-Pr
COOEt
COOMe
Br
9
26
44
0
1
60
70
2
3
Cl
Cl
76
81
entry 16): the reaction with ZnEt gave, after 10 h, 94% of
2
4
-65
48
reduction product and only 6% of the addition product 9b, with
5% ee. It seems that amino alcohols are not efficient for addition
reactions of ZnR on trifluoroacetophenones. Note that the
5
Me
Et
-37/rt 72
-37/rt 48
-37/rt 72
54
64
6
f
2
7
81
kind of complex intervening in the case of amino alcohols is very
different: an alkoxo-bridged dimer [ZnR(O-N)] formed by
a
ZnR
2
/trifluoroacetophenone/L = 1.2/1/0.1. Reactions were carried
(1 M solution in hexane), and in
(2 M solution in toluene). Concentration: 1 mmol of
2
out in 1/2.5 hexane/toluene with ZnEt
toluene with ZnMe
ketone in 2 mL of solvent. Except for entries 2, 4, 8 and entries nm, the
ketone was added to the ZnR
was removed. rt = room temperature. Determined by GC and HPLC
analysis: CHIRASIL-DEX CB (25 m ꢀ 0.25 mm ꢀ 0.25 mm) and
CHIRALPAK AD columns. Determination of the absolute configuration
for 9a was possible by comparison with literature data. Compound 10 is
racemic. Literature data.
,1,1-trifluoroacetophenone.
2
11,24,25
alcoholysis of the starting ZnR2.
2
b
Entries 18-24 in Table 1 collect reactions using ligand 4,
2
/L mixture at -35 °C, and then the cooling
ZnMe , and pharmacologically useful para-substituted ace-
2
c
26
tophenones as substrate. The yields are basically unaffected
in all cases. However, a clear influence of the para substituent
on the ee values is observed, which seems to be related to their
bulkiness. Thus, for the series of carboxylates RCO as
10
d
15
e
f
Not observed. The substrate is 3 ,5 -dimethyl-
0
0
2
1
substituents (Table 1, entries 18-21), the sequence of ee
values is t-Bu (26%) ≈ i-Pr (26%) <Et (44%) <Me (60%).
Moving to halide substituents (Table 1, entries 22-24), the
ee values are again lower than for trifluoroacetophenone
itself but higher than for the carboxylate-substituted mole-
cules and the order depends again, apparently, on the size of
the halogen: Br (70%)<Cl (76%). Interestingly, carrying out
the reaction at lower temperature (- 65 °C) the reaction of
p-Cl-C H -C(O)(CF ) affords the corresponding chiral alco-
enantioselective intermediates (see later) formed in situ (maxi-
mum 8.33%) by mixing ZnR and N-N* (10/1). Using 10% of
2
ligand 4, the reaction of ZnEt and trifluoroacetophenone
2
afforded the desired addition product 9a in 90% yield and
82% ee in 10 h (Table 1, entry 7; the same result with opposite
enantioselectivity was obtained with its enantiomer ent-4). A
certain amount of reduction product 10 (8%) was also formed.
At -60 °C (48 h) (Table 1, entry 8) the reaction was much more
chemoselective, and no reduction product 10 was observed,
which increased the yield in addition product to 97%; moreover,
the enantiomeric excess increased to 92%. These results out-
perform by a great deal the best result reported so far in the
literature (these were reported for ligand 2: Table 1, entries 5
6
4
3
hol p-Cl-C H -C(CF )(Me)OH in 95% yield and 81% ee,
6
4
3
which furnishes an appropriate starting head for further
functionalization.
Since all the para substituents in entries 18-24 are electron
withdrawing, the effect of electron-donating substituents
was also checked in entries 25-27, for the sake of complete-
15
and 6). The reaction with ZnMe catalyzed with 4 also worked
2
ness. The reactions were clearly slower, as expected, but
very well, and product 9b was formed cleanly in 98% yield and
83% ee (24 h, room temperature).
(
24) Kitamura, M.; Okada, S.; Suga., S.; Noyori, R. J. Am. Chem.
Ligand 4 has the diamine and the binaphthalene moieties
with opposite absolute configurations: S,R,R. Its diastereo-
isomer 5, where the diamine and the binaphthalene substit-
uents are all S, performs dramatically worse: the reaction
Soc. 1989, 111, 4028–4036.
(25) V ꢀa zquez, J.; Peric ꢁa s, M. A.; Maseras, F.; Lled oꢀ s, A. J. Org.
Chem. 2000, 65, 7303–7309.
(
26) Para substitution on the aryl of trifluoroacetophenone by a
synthetically active function is essential for easy modification of this
parent product to its biologically investigated derivatives.
with ZnEt is clearly slower and leads, after 10 h, to 9a in only
2

Article
Organometallics, Vol. 29, No. 23, 2010 6405
longer reaction times afforded good to excellent yields and
moderate to very good enantioselectivity.
(
leads to the following conclusions. (i) For ZnMe the reactions
b). Kinetic Study. An analysis of all the results in Table 1
2
under ligandless conditions are so inefficient that, in the presence
of catalytic amounts of ligand, all the addition products must
be formed on [ZnR (N-N)]. Any loss of enantioselectivity is to
2
be attributed to this complex, not to a competition with reac-
tions occurring on nonchiral ligandless ZnMe . (ii) For ZnEt
2
2
the activity under ligandless conditions is faster than for
ZnMe , affording less than 3% addition product. This poor rate
2
toward addition will be certainly negligible in front of the faster
processes occurring on [ZnEt (N-N)]. However, since the
2
reduction process is fairly efficient under ligandless conditions,
the question still arises whether the percentage of reduction
observed in the reactions is occurring on ligandless ZnEt , on
2
1
9
Figure 4. F NMR spectra of the reaction mixture for the
ligandless reaction (Table 1, entry 1) before (above) and after
(below) hydrolysis. Similar spectra are observed for the reac-
tions with ligands.
[
ZnEt (N-N)] complexes, or on both pathways, were they com-
2
petitive. Associated with this is the following question: are the
ligands showing lower activity completely coordinated to Zn, or
are they partially dissociated? Important dissociation might
seriously reduce the proportion of [ZnEt (N-N)] catalyst in
2
the mixture (in case of complete coordination the expected
proportion is 8.33% [ZnEt (N-N)] vs 91.66% ZnEt ), making
2
2
the ligandless ZnEt catalyst more abundant and more compet-
2
itive. In order to answer these questions (particularly for the
singular case of ligand 5), the reactions of trifluoroacetophenone
1
with ZnEt were monitored by F NMR at low tempera-
9
2
ture (-37 °C), with ligands 1 and 4-6, as well as under ligand-
less conditions, and their kinetic behavior was inspected.
In the course of these reactions many signals appear in the
19
F NMR spectra. They probably correspond to oligomeric
intermediates of different complexity (whether free or com-
plexed to Zn), which is supported by the fact that all these
signals collapse to only two products (9a and 10) upon hydro-
lysis (Figure 4).
Figure 5. Ligand influence in the rate of consumption of Ph-
27
3 2
C(O)-CF (11) in reactions with ZnEt with or without added
1
9
ligand (monitored by F NMR, in toluene, at -37 °C. ZnR /
2
trifluoroacetophenone/L=1.2/1/0.1. The initial concentration of
11 is 0.24 mmol/mL.
This complexity before hydrolysis prevents the kinetic mon-
itoring of the reactions through the products, but the reactions
could be easily studied, at -37 °C, through the decay of the
starting trifluoroacetophenone signal (Figure 5). The final pro-
portion of the two possible products (9a and 10) could be cleanly
integrated after hydrolysis, once the reaction was terminated. The
latter data are gathered in Table 1 (entries nm, data in italics).
The reaction rates of consumption of trifluoroacetophenone
follow the trend 1 > 4 ≈ 6 > 5 . no ligand. Assuming
bimolecular reactions, the rate of consumption of 11 should
obey a rate law composed of addition and reduction processes
28
details). This proves that (i) practically all the N-N ligand is
coordinated to Zn, thus keeping the concentration of ZnEt -
2
29
(
N-N) constant throughout the reaction, (ii) all the com-
plexes with N-N ligands react very quickly, whether for
addition or for reduction, in comparison to the ligandless ZnEt2,
so that the contribution of the latter to the products is negligible,
and (iii) consequently the addition/reduction rates and the
enantiomeric excesses observed, whether high or low, can be
attributed almost exclusively to the [ZnEt (N-N)] complex.
2
For the reaction in the absence of any L (no ligand, entry
23
m), the reaction rate should simplify to rate = k red-
on complex [ZnEt (N-N)] plus addition and reduction pro-
2
0
1
cesses on ligandless ZnEt . The kinetic behavior might be quite
complex if all these components were significant (eq 1).
2
[
ZnEt ][11]. The fitting of our experimental data for this case
2
to a second-order rate equation is reasonably good but shows
some irregularity that reveals higher complication. We con-
sider that it is the consequence of a competition of the
starting ketone 11 and their oxygen-containing products as
ligands for the coordination sphere of Zn, so that the
rate ¼ ðkad þ k Þ½ZnEt ðN - NÞꢁ½11ꢁ
red
2
þ ðk0 þ k0 Þ½ZnEt ꢁ½11ꢁ
ð1Þ
ad
red
2
However, for all the ligands (even for the apparently peculiar
ligand 5) the reaction rates fit to a pseudo-first-order law:
rate = (kad þ k )[ZnEt (N-N)][11] = k[11], where k =
“ligandless” ZnEt is changing its nature along the reaction,
2
as oxygen-containing products are being formed, especially
at the early stages of the reaction.
red
2
(
k þk )[ZnEt (N-N)] (see the Supporting Information for
ad
red
2
(
[ZnEt
29) In other words, the possible ligand-dissociation equilibrium of
(27) Some probable intermediates in the addition of dialkylzinc to
trifluoromethyl ketones have been characterized recently: Hevia, E.;
Kennedy, A. R.; Klett, J.; Livingstone, Z.; McCall, M. D. Dalton Trans.
2
(N-N)] is negligible, the concentration of this complex is con-
stant and essentially identical with the concentration of the ligand, and
the concentration of free N-N is unobservable. This does not mean that
2
010, 39, 520–526.
28) Note that [ZnEt
ZnEt throughout the reaction until the initial ZnEt
full consumption.
2
the N-N ligand stays on the same ZnEt center all the time. On the
(
2
L] remains constant to 8.33% of the initial
(91.66%) is close to
2
contrary, NMR experiments prove that, in a solution of [ZnEt -
2
2
(TMEDA)] þ TMEDA, there is fairly fast exchange between free and
complexed TMEDA.

6
406 Organometallics, Vol. 29, No. 23, 2010
Genov et al.
Table 2. Rate Constants Obtained from the Kinetic Experiments
(
needed to achieve high addition activity and high enantios-
electivity.
See Figure 4)
0
k
red
(M min )
The N-N diamine ligands, active for additions to ketones,
turn out to be complementary to the N-O amino alcohols,
active for additions to aldehydes. This different activity must
arise from different mechanisms, since the neutral diamines, not
having active hydrogens, are not likely to produce the kind of
alkoxy-amino Zn intermediates proposed for amino alcohols.
k
ad þ kred
-
1
-1
-1
-1
-1
entry
L
K (min
)
(M min
)
a
m
b
1
3
7
1
1
0.001 136(8)
m
1
4
5
6
0.014 87(10)
0.004 052(18)
0.001 528(5)
0.003 355(8)
0.619(4)
m
0.1688(8)
0.0637(2)
0.1398(3)
0m
2m
a
Experimental Section
ZnEt
2
/trifluoroacetophenone/L = 1/1/0.1. The reaction was carried
(1 M solution in hexane).
Assuming second order for the reaction.
out in 1/2.5 hexane/toluene with ZnEt
2
b
General Information. All reactions were performed under an
argon atmosphere. Toluene was dried over sodium and distilled.
Flash chromatography was performed on silica gel (silica gel 60,
30-400 mesh, Merck). Dimethylzinc (2 M in toluene), diethyl-
zinc (1 M in hexanes), and 1,2-bis(bromomethyl)benzene
Table 2 collects the rate constants associated with the
experiments in Figure 4, confirming that the reactions on the
N-N complexed ZnEt are, including the slowest ligand 5,
about 2 orders of magnitude larger than on ligandless ZnEt2.
This confirms that only the complex [ZnEt (N-N)] needs to
be considered to analyze the results.
Finally, it is noticeable that the ee values drop from 83% for
ligand 4 to 15% for ligand 6 in the reaction with ZnMe (entries
2
9 and 13), whereas the difference is much smaller in the reac-
tions with ZnEt (82% vs 75%, entries 7m and 12m). Also, the
addition/reduction ratio with ligand 5 changes dramatically
from ZnMe (85/0, entry 11) to ZnEt (18/81, entry 10). It
is also interesting that both the addition/reduction ratio and
the enantioselectivity improve at lower temperatures. These
observations are compatible with the idea that a tight steric
2
0
2
were purchased from Aldrich. (R)- and (S)-2,2 -bis(bromomethyl)-
0
1,1 -binaphthalene were purchased from TCI Europe. Analytical
gas chromatography was performed on a Hewlett-Packard 5890
2
Series II machine equipped with a CHIRASIL-DEX CB (25 mꢀ
0
.25 mmꢀ0.25 mm) capillary column. All NMR experiments were
performed on Bruker AV400, ARX300, and AC300 spectrometers.
Optical rotation was measured on a Perkin-Elmer 343 apparatus.
Mass spectra were recorded on an Agilent Technologies 5973
Network apparatus.
2
Synthesis of Amino Alcohol 8. Under an argon atmosphere in
dry THF (6 mL) were mixed 0.072 g (0.25 mmol) of (S)-2-amino-
2
2
3
0
0
1,1,2-triphenylethanol, 0.110 g (0.25 mmol) of (R)-2,2 -bis(bromo-
methyl)-1,1 -binaphthalene, and 0.07 mL (0.5 mmol) of triethyl-
0
amine. The reaction mixture was stirred at 70 °C for 24 h. Then
the solvent was removed and the crude product was purified
by column chromatography (8/1 hexanes/ethyl acetate) to
match of the N-N ligand and the ZnR reagent, producing a
2
lower entropy [ZnR (N-N)] intermediate, favors higher enan-
2
give 0.104 g (73%) of 8 as white crystals. R
f
= 0.38 (8/1
tioselectivity in the additions and hinders the reduction process.
1
hexanes/ethyl acetate). H NMR (400 MHz, CDCl ; δ (ppm)):
3
3
4
6
7
2 2
.33 (d, J=12.31 Hz, 2H, CH ); 3.62 (d, J=12.31 Hz, 2H, CH );
Conclusions
.83 (s, 1H, CH); 5.30 (s, 1H, OH); 6.55 (d, J=8.47 Hz, 2H, Ar);
.89 (t, J = 6.88 Hz, 1H, Ar); 7.00 (t, J = 7.94 Hz, 2H, Ar);
.19-7.24 (m, 6H, Ar); 7.35-7.39 (m, 8H, Ar); 7.41 (t, J=6.88
While amino alcohols are excellent ligands to catalyze the
addition of alkylzincs to aldehydes, they fail with trifluor-
oacetophenone. In contrast, chelating diamines are very
efficient, and we have developed a highly enantioselective
addition of ZnEt or ZnMe to trifluoroacetophenone. The
Hz, 2H, Ar); 7.74 (t, J=7.94 Hz, 4H, Ar); 7.87 (d, J=8.47 Hz,
H, Ar). Anal. Calcd: C, 88.86; H, 5.86; N, 2.47. Found:
2
2
5
C, 88.38; H, 5.79; N, 2.43. [R]D =-35° (c=0.7, CHCl ). Mp:
145 °C. MS (EI; m/z (%)) 384 (100), 279 (28), 263 (43), 182 (10),
118 (18), 105 (35), 91 (27), 77 (17), 51(8).
3
2
2
corresponding addition products (trifluoromethyl-substi-
tuted tertiary alcohols), are obtained in excellent yield and
good enantioselectivity at room temperature, using the
appropriate ligand. To the best of our knowledge, this is
the first report of addition of the less reactive ZnMe to
2
trifluoromethyl ketones. In the case of ZnEt , working at low
2
temperature (-60 °C) with ligand 4 totally quenches the
formation of the reduction product 10 and affords exclu-
sively the addition product in high yield (98%) and higher
enantioselectivity (92% ee). Alternatively, ligand 6 also
quenches the formation of reduction product at room tem-
perature, although with lower enantioselectivity (75% ee).
These are by far the best enantioselective additions to
trifluoromethyl ketones reported to date. The chiral alcohol
p-Cl-C H -C(CF )(Me)OH, which is synthetically useful as
2 2
Typical Procedure for the ZnEt and ZnMe Addition Reac-
tions. In a 50 mL Schlenk flask with a Young’s tap and a Teflon
stirring bar was introduced 0.025 mmol of the L ligand dissolved
in toluene (see Table 1). A 0.3 mmol solution of ZnEt or ZnMe
2 2
was added at -35 °C. The mixture was stirred at that tempera-
ture for 45 min. Then 0.25 mmol (35 mL) of trifluoroacetophe-
none were added, also at -35 °C to avoid the overheating
associated with this mixing, the cooling bath was removed,
and the mixture was allowed to slowly reach room temperature
and stirred for the time indicated in Table 1. Then the mix-
ture was carefully hydrolyzed with saturated NH
extracted with Et O, filtered through a short pad of silica
and analyzed with GC or HPLC. GC retention times: for 9a,
=18.7 min, t =19.9 min; for 10, t =27.1 min, t =28.3 min.
GC: 15 min at 100 °C, then 13 min at 120 °C, ramp 10 °C/min;
for 9b (Z = H), t =23.4 min, t
=24.4 min. GC: isotherm at 100 °C;
4
Cl solution,
2
t
1
2
1
2
6
4
3
1
2
a head of pharmacologically interesting derivatives, is obtained
in 95% yield and 81% ee.
The results of a kinetic monitoring of the reactions with
for 9b (Z = CN), t = 24 min, t = 25.2 min. GC: isotherm at
1
2
1
1 2
40 °C; for 9b (Z = Cl), t =15.7 min, t =16.2 min. GC: isotherm
at 125 °C. Compound 10 is racemic. HPLC retention times:
for 9b (Z=COOEt), t =14.7 min, t =16.5 min (hexane/i-PrOH
ZnEt support that both the addition and the reduction
2
1
2
reactions occur exclusively on [ZnEt (N-N)] complexes
for N-N = diamine, because their rates are at least 2 orders
of magnitude greater than on ligandless ZnEt2.
95/5, 0.5 mL/min); for 9b (Z=COO-t-Bu), t =15.3 min, t =
1 2
17.5 min (hexane/i-PrOH 96/4, 0.5 mL/min); for 9b (Z=COO-i-Pr),
2
Less selective results suggest that this reaction is particu-
larly delicate and a very good match between the sizes of the
ligand, the alkyls, and the substituents at the aryl ring is
(
30) Ma, X.; Da, C.-S.; Yi, L.; Jia, Y.-N.; Guo, Q.-P.; Che, L.-P.; Wu,
F.-C.; Wang, J.-R.; Li, W.-P. Tetrahedron: Asymmetry 2009, 20, 1419–
1424.

Article
Organometallics, Vol. 29, No. 23, 2010 6407
t
9
0
1
=16.7 min, t
b (Z=COOMe), t
.5 mL/min).
2
=18.1 min (hexane/i-PrOH 96/4, 0.5 mL/min); for
Consolider Ingenio 2010, Grant INTECAT, CSD2006-0003,
and from the Junta de Castilla y Le oꢀ n (Project GR169).
=23.1 min, t =24.6 min (hexane/i-PrOH 96/4,
1 2
Supporting Information Available: Figures and tables giving
H NMR spectra of ligands and products and data for the
kinetic studies. This material is available free of charge via the
Internet at http://pubs.acs.org.
1
Acknowledgment. Financial support is gratefully
acknowledged from the Spanish Direcci oꢀ n General
de Investigaci oꢀ n (DGI): Grant CTQ2007-67411/BQU;