2-Piperidino-1,1,2-triphenylethanol: A Highly Effective Catalyst for the Enantioselective Arylation of Aldehydes

Article abstract of DOI:10.1021/jo035824o

Here we report the use of 2-piperidino-1,2,2-triphenylethanol (5) as an outstanding catalyst for the ligand-catalyzed arylation of aldehydes. The use of 5 and a 2/1 mixture of Et₂Zn/Ph₂Zn provided the corresponding chiral diarylcarbinols with enantiomeric excess of up to 99% ee. The effect of temperature on the reaction enantioselectivity was studied and the inversion temperature (T_inv) was determined to be 10 °C for reaction with p-tolylaldehyde. Most remarkably, lowering the amount of catalyst (5) to 0.5 mol % still afforded excellent levels of enantiocontrol (93.7% ee). Kinetics of the catalyzed and uncatalyzed arylation of aldehydes was studied by means of in situ FT-IR. The background uncatalyzed addition rates to p-tolylaldehyde when using pure Ph₂Zn and Et₂Zn/Ph ₂-Zn (2/1) suggest that in the latter case a mixed zinc species forms (EtPhZn) minimizing the undesired nonselective addition. Formation of EtPhZn was modeled at the DFT calculation level. A four-center TS (TS-V) corresponding to the Et/Ph scrambling was localized along with two dimers (D-IV and D-VI). The model supports the hypothesis that Et/Ph exchange is a kinetically facile process. Gas evolution experiments during the formation of the active catalyst showed that the formation of an active site with a ONZn-Et (10) moiety is kinetically favored over ONZn-Ph (11). Finally, the phenyl transfer to benzaldehyde was modeled at the PM3(tm) level through anti and syn 5/4/4 tricyclic TS structures for both 10 and 11. The model could correctly predict the sense and selectivity of the overall process and predicted that 11 should be more selective than 10.

Full text of DOI:10.1021/jo035824o

2
-P ip er id in o-1,1,2-tr ip h en yleth a n ol: A High ly Effective Ca ta lyst
for th e En a n tioselective Ar yla tion of Ald eh yd es
Montserrat Fontes, Xavier Verdaguer, Llu ´ı s Sol a` , Miquel A. Peric a` s,* and Antoni Riera
Unitat de Recerca en S ´ı ntesi Asim e` trica (URSA-PCB), Parc Cient ´ı fic de Barcelona and Departament de
Qu ı´ mica Org a` nica, Universitat de Barcelona, J osep Samitier 1-5, E-08028 Barcelona, Spain
mapericas@pcb.ub.es
Received December 16, 2003
Here we report the use of 2-piperidino-1,2,2-triphenylethanol (5) as an outstanding catalyst for
the ligand-catalyzed arylation of aldehydes. The use of 5 and a 2/1 mixture of Et
the corresponding chiral diarylcarbinols with enantiomeric excess of up to 99% ee. The effect of
temperature on the reaction enantioselectivity was studied and the inversion temperature (Tinv
2 2
Zn/Ph Zn provided
)
was determined to be 10 °C for reaction with p-tolylaldehyde. Most remarkably, lowering the amount
of catalyst (5) to 0.5 mol % still afforded excellent levels of enantiocontrol (93.7% ee). Kinetics of
the catalyzed and uncatalyzed arylation of aldehydes was studied by means of in situ FT-IR. The
2 2 2
background uncatalyzed addition rates to p-tolylaldehyde when using pure Ph Zn and Et Zn/Ph -
Zn (2/1) suggest that in the latter case a mixed zinc species forms (EtPhZn) minimizing the undesired
nonselective addition. Formation of EtPhZn was modeled at the DFT calculation level. A four-
center TS (TS-V) corresponding to the Et/Ph scrambling was localized along with two dimers (D-
IV and D-VI). The model supports the hypothesis that Et/Ph exchange is a kinetically facile process.
Gas evolution experiments during the formation of the active catalyst showed that the formation
of an active site with a ONZn-Et (10) moiety is kinetically favored over ONZn-Ph (11). Finally, the
phenyl transfer to benzaldehyde was modeled at the PM3(tm) level through anti and syn 5/4/4
tricyclic TS structures for both 10 and 11. The model could correctly predict the sense and selectivity
of the overall process and predicted that 11 should be more selective than 10.
In tr od u ction
providing racemic product arylcarbinol that lowers the
overall system performance.3
Starting from the pioneering work of Fu, who reported
Enantiomerically pure diarylmethanols are the basis
of economically and therapeutically important medicines
such as (R)-neobenodine, (R)-orphenadrine, or (S)-carbi-
2
in 1997 the addition of salt-free Ph Zn to p-chloroben-
zaldehyde using a planar-chiral azaferrocene ligand with
1
noxamine. Besides the reduction of appropriate diaryl
4
moderate enantioselectivity, much effort has been dedi-
ketones, where the achievement of high enantioselectivi-
ties can become problematic when the two aryl groups
are similar in volume or electronic nature, the enanti-
oselective arylation of aldehyde substrates appears to be
the most promising alternative for the preparation of
cated to the generation of a practical system for the
highly enantioselective arylation of aromatic aldehydes.
Since then, two major approaches have been devised to
minimize the competing uncatalyzed process. On one
hand, Pu and co-workers found that carrying out the
addition reaction at low concentrations improved dra-
matically the enantioselectivity when using a chiral
2
these molecular systems. However, the development of
an enantioselective Ph Zn addition to aldehydes is a
2
challenging endeavor compared to the related and well-
studied diethylzinc addition. The main difficulty that
hampers the asymmetric aryl transfer is the background-
uncatalyzed addition that competes with the ligand-
catalyzed reaction. Diarylzinc species are several orders
of magnitude more reactive toward aldehydes than the
corresponding alkyl species (Et
such an enhanced reactivity gives rise to the undesired
direct addition of Ar Zn to the carbonyl compound
5
binaphthol ligand 1. Alternatively, Bolm and co-workers
have developed an effective reagent system that involves
the use of a Ph Zn/Et Zn mixture to suppress the
2 2
(3) (a) Bolm, C.; Hildebrand, J . P.; Muniz, K.; Hermanns, N. Angew.
Chem., Int. Ed. 2001, 40, 3284-3308. For leading references on aryl
transfer to ketones see: (b) Ram o´ n, D. J .; Yus, M. Tetrahedron 1998,
54, 5651-5666. (c) Ram o´ n, D. J .; Yus, M. Eur. J . Org. Chem. 2003,
2 2 2
Zn, Me Zn), and
Zn, Pri
2
9
4
745-2748. (d) J eon, J -.S.; Walsh, P. J . J . Am. Chem. Soc. 2003, 125,
544-9545. (e) Dosa, P. I.; Fu, G. C. J . Am. Chem. Soc. 1998, 120,
45-446. For general reviews on enantioselective addition of orga-
2
(
1) (a) Harms, A. F.; Nauta, W. T. J . Med. Pharm. Chem. 1960, 2,
7-77. (b) Sperber, N.; Papa, D.; Schwenk, E.; Sherlock, M. J . Am.
Chem. Soc. 1949, 71, 887-890.
2) For asymmetric hydrogenation and oxazaborolidine reduction of
nozincs to aldehydes see: (f) Soai, K.; Niwa, S. Chem. Rev. 1992, 92,
833-856. (g) Pu, L.; Yu, H.-B. Chem. Rev. 2001, 101, 757-824. (h)
Noyori, R.; Kitamura, M. Angew. Chem., Int. Ed. Engl. 1991, 30, 34-
48.
5
(
benzophenones see: (a) Ohkuma, T.; Koizumi, M.; Ikehira, H.;
Yokozawa, T.; Noyori, R. Org. Lett. 2000, 2, 659-662. (b) Corey, E. J .;
Helal, C. J . Tetrahedron Lett. 1996, 37, 5675-5678. (c) Corey, E. J .;
Helal, C. J . Tetrahedron Lett. 1995, 36, 9153-9156.
(4) Dosa, P. I.; Ruble, J . C.; Fu, G. C. J . Org. Chem. 1997, 62, 444-
445.
(5) (a) Huang, W.-S.; Hu, Q.-S.; Pu, L. J . Org. Chem. 1999, 64, 7940-
7956. (b) Huang, W.-S.; Pu, L. J . Org. Chem. 1999, 64, 4222-4223.
1
0.1021/jo035824o CCC: $27.50 © 2004 American Chemical Society
Published on Web 03/05/2004
2
532
J . Org. Chem. 2004, 69, 2532-2543

2
-Piperidino-1,1,2-triphenylethanol
6
undesired background reaction. Employing ferrocenyl
oxazoline 2 and rhenium-tricarbonyl oxazoline 3 ligands
a highly enantioselective aryl transfer was achieved for
a variety of arylaldehydes. The same catalytic system also
has been applied to the phenyl transfer to N-formyl-
imines.7 Whereas the Bolm methodology allows the
addition reactions to be performed at synthetically useful
concentrations within reasonable reaction times, the
achievement of high enantioselectivities relies on rather
complex ligands and requires their use in substantial
amounts. In contrast, Ha and co-workers, reported
recently that when using a biaryl derived amino-alcohol
TABLE 1. P h en yl Tr a n sfer Rea ction s a t Room
Tem p er a tu r e (a ld eh yd e con cen tr a tion 10 m M)
entry
solvent
yield (%)
ee (%)a
1
2
3
4
hexane
toluene
THF
Et2O
Et2O
57
29
2
73
50
10
78
48
13
29
b
5
2 2
the combination of Ph Zn and Et Zn provides a decreased
6
CH2Cl2
16
7
isolated yield of product carbinol. Despite these ad-
vances, a practical solution for the highly enantioselective
arylation of aldehydes involving the use of a readily
available ligand and low catalyst loading has still to be
developed.
a
Enantiomeric excess was determined by HPLC analysis of the
corresponding carbinol. b Reaction performed with no Et2Zn.
modular construction, which allows the easy modification
of their structures at any step of the synthesis. These
characteristics have been shown to be of fundamental
importance for the fine-tuning of catalytic activity and
enantioselectivity, and have allowed the ex novo design
of optimal ligands for a variety of important reactions.
Among these ligands, 2-piperidino-1,1,2-triphenyletha-
nol (5),8
d,11
available on the mol scale in both enantio-
meric forms through two simple operations from com-
mercially available triphenylethylene, occupies a singular
position. Not only does it exhibit a very high catalytic
activity in the ethylation of aldehydes of all structural
types, but it also shows a very high enantioselectivity at
low catalyst loading in these additions provided that the
carbon R to the formyl group bears at least two carbon
substituents.
The characteristics of 2-piperidino-1,1,2-triphenyl-
ethanol discussed above rendered this amino alcohol a
promising candidate ligand for the enantioselective ary-
lation of benzaldehydes. In particular, both the prefer-
ential recognition of one of the enantiotopic faces of the
substrate and a good level of suppression of the back-
ground reaction through its high catalytic activity were
anticipated. We wish to report in this paper how these
initial ideas have been elaborated and have led to the
establishment of 2-piperidino-1,1,2-triphenylethanol (5)
as a most practical ligand for the enantioselective aryl
In the past few years we have reported the synthesis
8
9
of new families of â-amino alcohol (5), â-amino thiol (4),
and bis(oxazoline) (6)10 ligands ultimately arising from
the regioselective and stereospecific ring opening of
epoxides obtained by Sharpless and J acobsen epoxida-
tions. The main advantages of these ligands are the broad
variety of structural types they can cover and their
2
transfer of Ph Zn to aldehydes.
Resu lts a n d Discu ssion
(
6) (a) Bolm, C.; Hermanns, N.; Hildebrand, J . P.; Muniz, K. Angew.
Chem., Int. Ed. 2000, 39, 3465-3467. (b) Bolm, C.; Kesselgruber, M.;
Hermanns, N.; Hildebrand, J . P.; Raabe, G. Angew. Chem., Int. Ed.
P h en yl Tr a n sfer u n d er High Dilu tion Con d ition s.
Pu and co-workers used the binol-derived catalyst 1
pretreated with 2.0 equiv of diethylzinc. The authors
assume that the zinc complex generated from the reaction
2
1
001, 40, 1488-1490. (c) Bolm, C.; Muniz, K. Chem. Commun. 1999,
295-1296. (d) Hermanns, N.; Dahmen, S.; Bolm, C.; Brase, S. Angew.
5a
Chem., Int. Ed. 2002, 41, 3692-3694. For the enantioselective aryl
transfer from Ar(OH)
2
/Et
2
Zn see: (e) Bolm, C.; Rudolph, J . J . Am.
Chem. Soc. 2002, 124, 14850-14851.
of 1 with diethylzinc is a better catalyst than the one
(
(
7) Ko, D.-H.; Kim, K. H.; Ha, D.-C. Org. Lett. 2002, 4, 3759-3762.
8) (a) Vidal-Ferran, A.; Moyano, A.; Peric a` s, M. A.; Riera, A. J . Org.
generated with Ph
solvent for each substrate and the use of additives
MeOH) proved to be critical to obtain high enantiose-
lectivity. We first investigated the addition of Ph Zn to
2
Zn. In addition, the right choice of
Chem. 1997, 62, 4970-4982. (b) Vidal-Ferran, A.; Bampos, N.; Moyano,
A.; Peric a` s, M. A.; Riera, A.; Sanders, J . K. M. J . Org. Chem. 1998,
3, 6309-6318. (c) J imeno, C.; Pasto, M.; Riera, A.; Pericas, M. A. J .
Org. Chem. 2003, 68, 3130-3138. (d) Sol a` , L.; Reddy, K. S.; Vidal-
Ferran, A.; Moyano, A.; Peric a` s, M. A.; Riera, A.; Alvarez-Larena, A.;
Piniella, J . F. J . Org. Chem. 1998, 63, 7078-7082. (e) Reddy, K. S.;
Sola, L.; Moyano, A.; Peric a` s, M. A.; Riera, A. J . Org. Chem. 1999, 64,
(
6
2
p-tolylaldehyde under high dilution conditions (10 mM)
using 10 mol % of 2-piperidino-1,1,2-triphenylethanol (5)
pretreated with an equimolar amount of Et Zn. The
2
results when performing the reaction in different solvents
are displayed in Table 1. The best enantioselectivities
3
969-3974. (f) Reddy, K. S.; Sola, L.; Moyano, A.; Peric a` s, M. A.; Riera,
A. Synthesis 2000, 2000, 165-176.
9) J imeno, C.; Moyano, A.; Pericas, M. A.; Riera, A. Synlett 2001,
155-1157.
10) Pericas, M. A.; Puigjaner, C.; Riera, A.; Vidal-Ferran, A.; Gomez,
M.; J imenez, F.; Muller, G.; Rocamora, M. Chem. Eur. J . 2002, 8,
164-4178.
(
1
(
(11) 2-Piperidino-1,1,2-triphenylethanol (5) is commercially available
in both enantiomeric forms.
4
J . Org. Chem, Vol. 69, No. 7, 2004 2533

Fontes et al.
SCHEME 1
evident. The addition reaction in toluene at room tem-
perature afforded the corresponding carbinol 7 in 96%
ee and 98% isolated yield (Table 2, entry 1). Decreasing
the reaction temperature to 0 °C had a slightly positive
effect on selectivity and provided 7 in 97% ee (Table 2,
entry 2). Hexane was a somewhat better solvent for this
process giving rise to the product with 97% and 98% ee
at room temperature and 0 °C, respectively (Table 2,
entries 3 and 4). Except for in diethyl ether, addition
reactions went to completion within 2 h even at 0 °C.
Concentration did not seem to be a critical issue with
this protocol. All reactions in Table 2 were performed at
an aldehyde concentration of 100 mM. However, an
aldehyde concentration of 250 mM in hexane afforded the
product alcohol in an excellent 95% ee. Under the present
set of conditions, the aryl transfer was completely selec-
tive over the ethylation process and no ethyl addition
product could be detected in the final reaction mixtures.
TABLE 2. Asym m etr ic P h en yl Tr a n sfer Rea ction s
Usin g a P h 2Zn /Et2Zn Mixtu r e
a
ee (%)b
entry
solvent
T (°C)
yield (%)
1
2
3
4
5
toluene
toluene
hexane
hexane
Et2O
rt
0
rt
0
0
rt
98
85
94
90
80
90
96
97
97
98
96
88
c
6
toluene
2
Finally, to check whether the excess of Et Zn was of any
a
Aldehyde concentrations approximately 100 mM. b Enantio-
benefit, a blank experiment lacking the additive was
carried out (Table 2, entry 6). Without diethylzinc an
extremely fast reaction took place and the product was
obtained in high yield but decreased optical purity (88%
meric excess was determined by HPLC analysis of the correspond-
ing carbinol. 1.2 equiv of Ph2Zn was used and no extra Et2Zn
was added.
c
2 2
ee). This confirmed the use of Ph Zn/Et Zn mixtures as
a clear advantage when using 5 as a catalyst.
(78% ee) could be observed when using diethyl ether as
solvent (entry 4). Hexane gave rise to slightly lower
selectivity but with higher conversion numbers (entry 1).
As reported, not pretreating the ligand with Et Zn had
2
With the purpose of determining the scope and utility
of the present system (Ph Zn/Et Zn and 5) for the
2 2
asymmetric arylation of aldehydes a series of substrates
with diverse steric and electronic properties was inves-
tigated (Table 3). It should be emphasized that the level
of enantiocontrol for all aromatic substrates is outstand-
ing. The corresponding diarylcarbinols are obtained in
optical purities in the range of 93% to 98% enantiomeric
excess. Hexane was used as the solvent of choice. Most
intriguingly, similar enantioselectivities were achieved
working either at 0 °C or room temperature (vide infra).
This could conveniently simplify the experimental pro-
cedure and represents an important improvement with
respect to previously reported catalytic systems. Ortho-
substituted benzaldehydes are good substrates and lead
to excellent enantiomeric excess (Table 3, entries 1, 4,
and 5). On the other hand, linear aliphatic aldehydes
provide poorer selectivities. Addition to n-heptanal yielded
the product alcohol in 63% ee (Table 3, entry 9). Increas-
ing the steric bulk of the carbon chain resulted in higher
enantioselectivities (Table 3, entries 10-12). Finally, a
significant solvent dependence was observed for R,â-
unsaturated aldehydes (Table 3, entries 13 and 14). In
the case of (E)-R-methylcinnamaldehyde, running the
reaction in toluene instead of hexane confers a jump in
enantiomeric excess from 87% to 94%. An assessment of
the data exposed in Table 3 shows that, since outstanding
results were obtained for aromatic and fully substituted
aliphatic aldehydes (Table 3, entries 12 and 14), it is
necessary to have no hydrogens R to the carbonyl to
obtain high enantioselectivity. This is in contrast with
the geometric requirements found for the ethylation
process when using 5 as a catalyst where a single
substituent R to the aldehyde is sufficient to reach high
ee. For example, cyclohexylcarbaldehyde provides 98%
a deleterious effect and the enantiomeric excess of the
resulting carbinol (7) dropped to 48% ee (entry 5).
Throughout the study utilizing high dilution conditions
the yields and conversion numbers were not satisfactory
and in most experiments starting material remained
unconsumed after extended reaction time (24 h) even at
room temperature.
2 2
P h en yl Tr a n sfer Usin g a P h Zn /E t Zn Mixt u r e.
The use of mixed zinc species in asymmetric catalysis is
a well-documented strategy when a valuable alkyl radical
is to be completely transferred.12 Knochel and co-workers
developed the bis[(trimethylsilyl)methyl]zinc 8 which, in
the presence of a diorganozinc compound (R
2
Zn), leads
1
3
to the formation of R-Zn-CH Si(CH ) (9). In the newly
2 3 3
formed organometallic complex the (trimethylsilyl)methyl
group acts as a nontransferable group in the ligand-
catalyzed addition to aldehydes. NMR experiments sup-
port the existence of mixed (â-silylalkyl)-Zn-alkyl species
in equilibrium with the starting symmetric zinc com-
pounds.13 Although no real evidence exists for the related
2 2
Ph Zn/Et Zn mixture, a similar equilibrium has been
postulated for the formation of the mixed EtPhZn
complex.6 Under these circumstances only the phenyl
group on Zn is transferred and the ethyl group would
act as a nontransferable radical.
a
We explored the use of an excess of Et
2
Zn (1.32 equiv)
along with 0.64 equiv of Ph Zn with p-tolylaldehyde as
2
a test substrate (Table 2). From the very beginning the
effectiveness of ligand 5 for the asymmetric phenyl
transfer in combination with the mixed zinc species was
(
12) (a) Wipf, P.; Ribe, S. J . Org. Chem. 1998, 63, 6454-6455. (b)
Lal o¨ e, E.; Srebnik, M. Tetrahedron Lett. 1994, 35, 5587-5590. (c) Lutz,
C.; Knochel, P. J . Org. Chem. 1997, 62, 7895-7898. (d) Traverse, J .
F.; Hoveyda, A. H.; Snapper, M. L. Org. Lett. 2003, 5, 3273-3275.
2
ee in the Et Zn addition compared to 60% ee (Table 3,
8d
entry 10) in the arylation reaction.
Comparison of the absolute configuration of the product
alcohols in Table 3 with the stereochemistry of â-amino
(
13) Berger, S.; Langer, F.; Lutz, C.; Knochel, P.; Mobley, T. A.;
Reddy, C. K. Angew. Chem., Int. Ed. Engl. 1997, 36, 1496-1498.
2
534 J . Org. Chem., Vol. 69, No. 7, 2004

2
-Piperidino-1,1,2-triphenylethanol
TABLE 3. Ar yla tion Rea ction w ith Differ en t Ald eh yd esa
a
Reactions were carried out in the presence of 10 mol % of 5 using a mixture of 0.64 equiv of Ph2Zn and 1.32 equiv of Et2Zn.
b
Enantiomeric excess was determined by chiral HPLC analysis.
alcohol 5 enabled us to ascertain that the present
of catalyst in the reaction mixture. Thus, the relationship
between enantiomeric excess and temperature was in-
vestigated at 1 mol % catalyst loading, between 0 and
arylation process follows the empirical rule that Noyori
1
4
established for the ethylation process. (R)-Piperidino-
,1,2-triphenylethanol exclusively catalyzed the phenyl
1
2 2
25 °C, for the Ph Zn/Et Zn addition to p-tolylaldehyde
addition to the Si face of the starting aldehyde.
(Figure 1). In both toluene and hexane a maximum
enantioselectivity was observed when the reaction was
performed around 10 °C. From that point, either raising
or lowering the temperature has a deleterious effect on
reaction selectivity. Such behavior can be rationalized
according to the isoinversion principle.15 When a chemical
Interested by the uncommon behavior we encountered
in this system in terms of selectivity and reaction
temperature (vide supra), we decided to examine it more
closely. We soon discovered that the temperature depen-
dence became more perceptible when reducing the amount
(
14) Noyori, R.; Kitamura, M. Angew. Chem., Int. Ed. Engl. 1991,
(15) Buschmann, H.; Scharf, H.-D.; Hoffmann, N.; Esser, P. Angew.
Chem., Int. Ed. Engl. 1991, 30, 477-515.
3
0, 49-69.
J . Org. Chem, Vol. 69, No. 7, 2004 2535

Fontes et al.
TABLE 4. Ar yla tion Rea ction s a t 10 °C Usin g 1.5 m ol %
of 5a
F IGURE 1. Enantioselectivity vs temperature in the arylation
of p-tolylaldehyde (5, 1 mol %).
a
Enantiomeric excess was determined by chiral HPLC analysis.
was obtained in enantiomeric excess one point higher
(96.7% ee) than that observed in toluene. Running the
reaction in hexane at room temperature, a higher tem-
perature than Tinv, was shown to be less effective,
although the more practical procedure with regard to
reaction conditions may pay off a slight decrease on
enantioselectivity. To exemplify the effectiveness of the
present catalytic system even at low catalyst loadings a
series of substrates were submitted to the phenyl addition
using 1.5 mol % of 5 at 10 °C (Table 4). Excellent yields
and enantioselectivities were obtained for a wide range
of aromatic aldehydes. Most remarkably, to our knowl-
edge the highest ever observed enantiomeric excess in
F IGURE 2. Enantioselectivity vs catalyst loading in the
arylation of 4-methylbenzaldehyde.
process shows for a selectivity property (enantio-, regio-,
or chemoselectivity) two distinct linear regions in the
Eyring graph, the temperature at which the inversion
occurs is called the inversion temperature (Tinv). In our
system and for the case of p-tolylaldehyde Tinv is ap-
proximately 10 °C. Other examples of similar tempera-
ture dependence in asymmetric catalysis can be found
in the oxazaborolidine-mediated borane reduction of
Ph
2
Zn addition to aldehydes (99% ee) was attained for
4
-phenylbenzaldehyde (Table 4, entry 2). In addition,
diarylcarbinols derived from asymmetric phenyl addition
to p-tolylaldehyde and o-tolylaldehyde (Table 4, entries
1 and 3) can be valuable intermediates in the asymmetric
synthesis of antihistaminic neobenodine and orphenadrine
drugs.1
1
6
prochiral ketones or in the thoroughly studied transi-
tion-metal-catalyzed hydrogenation.17 To know the inver-
sion temperature for the system at hand is of tremendous
importance in terms of process optimization. In our case,
performing the reaction at 10 °C allowed the catalyst
loading to be reduced to 1 mol % with practically no loss
of enantioselectivity (96.6% ee).
At hand there was a tremendously active catalytic
system, since with 1 mol % of 5, reactions were complete
within 2-3 h. This discovery encouraged us to further
investigate the limits of the arylation reaction in terms
of catalyst loading. The enantiomeric excess observed
versus the amount of catalyst is depicted in Figure 2. We
were pleased to find that working at the inversion
temperature (10 °C) and a catalyst loading of 0.5 mol %
still afforded excellent levels of enantiocontrol (93.7% ee).
Hexane was once again slightly more efficient a solvent
than toluene in the phenyl addition to p-tolylaldehyde.
At 1 mol % of catalyst loading (10 °C) the product alcohol
In Situ F T-IR Stu d y of th e Ar yla tion P r ocess.
Despite the efficiency of the present catalytic system,
little is known about the factors which make it so
effective. In the reaction mixture there are several zinc
2 2
species: Ph Zn, Et Zn, and the postulated PhEtZn, which
could, in principle, add to the starting aldehyde. A
working hypothesis is that diphenyl- and diethylzinc are
in equilibrium with phenyldiethylzinc and that the new
mixed alkylarylzinc, being less reactive than Ph Zn, helps
2
to suppress the background uncatalyzed arylation reac-
tion. However, there is no experimental evidence sup-
(16) (a) Puigjaner, C.; Vidal Ferran, A.; Moyano, A.; Peric a` s, M. A.;
Riera, A. J . Org. Chem. 1999, 64, 7902-7911. (b) Corey, E. J .; Bakshi,
R. K.; Shibata, S. J . Am. Chem. Soc. 1987, 109, 5551-5553.
(
17) (a) Ojima, I.; Kogure, T.; Yoda, N. J . Org. Chem. 1980, 45,
4
728-4735. (b) Sinou, D. Tetrahedron Lett. 1981, 22, 2987-2998.
2
536 J . Org. Chem., Vol. 69, No. 7, 2004

2
-Piperidino-1,1,2-triphenylethanol
background reaction when using pure Ph
2
Zn and the
Ph Zn/Et Zn (2/1) mixtures. The corresponding plots are
2
2
depicted in Figures 5 and 6. In the case of the mixed zinc
reagent (Figure 6) the recorded reaction rates were
extremely slow even at 30 °C. The half-life of the starting
2
0
aldehyde at 10 °C was estimated to be 5.7 h. From data
0
in Figure 6, initial rate constants k
v
were calculated by
adjusting the aldehyde absorbance (concentration) to a
second-order rate law expression as has been proposed
2
1,22
for the addition of Grignard reagents to ketones.
Linear least-squares fitting of calculated k against 1/T
permitted the estimation of activation energy (E ) at 70.1
kJ ‚mol (16.7 kcal‚mol ) for the background Ph Zn/
Et Zn addition process. In a similar fashion from data
in Figure 5, the half-life for the direct addition of pure
Ph Zn at 10 °C was 6.0 min and the E was evaluated to
i
a
-
1
-1
2
2
2
a
-
1
-1 23
be 13.8 kJ ‚mol (3.3 kcal‚mol ). This value is lower
than that found for the addition of 2-methoxyphenyl-
magnesium bromide to di-tert-butyl-ketone in THF (41.5
kJ ‚mol ), but is in the same range as the E
the reaction of heptylmagnesium bromide with CO (17.6
F IGURE 3. In situ FT-IR spectra of the reaction mixture.
porting this hypothesis.18 To shed light onto these
-
1
24
a
found for
questions we undertook a study of the arylation process
by means of in situ FT-IR technology.19 FT-IR allowed
the disappearance of the starting aldehyde to be moni-
tored through its CdO absorbance at different time
intervals. IR spectra were recorded with a ReactIR 1000
fitted with an immersible DiComp ATR probe. A repre-
sentative reaction spectrum series for the arlylation of
p-tolylaldehyde is shown in Figure 3.
We first focused on the rate evaluation of the catalyzed
versus the uncatalyzed addition reaction to p-tolylalde-
2 2 2
hyde (0.25 M) using either pure Ph Zn or the Ph Zn/Et -
Zn (1/2) mixture. Reactions were carried out in toluene
at 0 °C employing the standard reaction conditions. The
corresponding plots of absorbance versus time are shown
in Figure 4. With Ph Zn (0.27 M) alone the uncatalyzed
2
arylation reaction is surprisingly rapid at 0 °C and the
reaction went to completion within approximately 25
2
-
1
25,26
kJ ‚mol ).
These data support the hypothesis that
Zn
when mixing diphenylzinc and diethylzinc (1:2), Ph
2
is depleted from the reaction medium and that a new
kind of zinc complex forms that barely reacts with the
aldehyde in the absence of the catalyst, thus minimizing
the background reaction.
DF T Com p u ta tion a l Stu d y on th e F or m a tion of
th e Mixed Zin c Sp ecies. Diorganozinc compounds that
contain saturated alkyl groups always occur as monomers
in solution.2
7,28
The zinc in these compounds is unable to
attain coordination saturation through the formation of
alkyl bridges, as occurs in the lower dialkylmagnesium
compounds. Bridging alkyl groups between zinc and zinc
centers have only been postulated as transition species
in the exchange of alkyl groups.28 On the other hand,
alkynyl groups directly bound to zinc are known to form
stable zinc bridges between zinc atoms. Thus, lower
dialkylzinc compounds are volatile nonpolar liquids while
dialkynylzinc species are solids and have very low
solubilities in nonpolar organic solvents. In this context,
diarylzinc compounds are a middle ground case: while
2
min. The presence of 2.5 mol % of 5 with Ph Zn (0.27 M)
accelerated the addition reaction to the point that it
reached completion in 5 min. On the other hand, when
the Ph
the direct background addition was extremely slow at 0
C, almost negligible, while the 5 (2.5 mol %) catalyzed
2 2
Zn/Et Zn (0.16 and 0.33 M) mixture was employed
°
arylation proceeded fast enough to be over within 15 min.
Despite the rate enhancement observed when using 5 as
a catalyst, the background reaction when using pure
(
20) Concentration of starting materials: aldehyde (0.25 M), Et
0.33 M), and Ph Zn (0.16 M).
21) Concentration of ethylphenylzinc was estimated to be 0.32 M,
2
Zn
(
2
(
Ph
2
Zn is fast enough to successfully compete with the
based on the assumption that the equilibrium is completely shifted
toward the formation of the mixed zinc species.
catalyzed process and cause a decrease in selectivity as
can be observed in the product enantioselectivity (Table
(
22) (a) Ashby, E. C.; Laemmle, J .; Neumann, H. M. Acc. Chem. Res.
974, 7, 272-280. (b) Ashby, E. C.; Laemmle, J .; Neumann, H. M. J .
Am. Chem. Soc. 1972, 94, 4.
1
2
, Entry 6). In contrast, when using the Et
mixture the catalyzed arylation is slightly slower than
that with pure Ph Zn, but the rate difference with the
2 2
Zn/Ph Zn
(
2
23) Concentration of starting materials: Aldehyde (0.25 M) and
Zn (0.27 M).
Ph
2
(
24) Oki, M. Tetrahedron Lett. 1967, 8, 1785-1788.
background reaction is greatly increased.
The uncatalyzed addition rates were next investigated
at different reaction temperatures to fully describe the
(25) Yajima, K.; Kawashima, H.; Hashimoto, N.; Miyake, Y. J . Phys.
Chem. 1996, 100, 14936-14940.
(26) Addition of Grignard and alkyllithium reagents to aldehydes
should provide lower activation parameters than their corresponding
ketone analogues due to reduced steric hindrance. For secondary
deuterium kinetic isotope effects in the irreversible additions of allyl
reagents to benzaldehyde see: (a) Gajewski, J . J .; Bocian, W.; Brich-
ford, N. L.; Henderson, J . L. J . Org. Chem. 2002, 67, 4236-4240. For
kinetic studies on the amino-alcohol catalyzed dialkylzinc addition to
aldehydes see: (b) Buono, F.; Walsh, P. J .; Blackmond, D. G. J . Am.
Chem. Soc. 2002, 124, 13652-13653. (c) Blackmond, D. G. J . Am.
Chem. Soc. 1998, 120, 13349-13353.
(27) Knochel, P.; J ones, P. Organozinc Reagents; Oxford University
Press: Oxford, UK, 1999.
(28) Boersma, J . In Comprehensive Organometallic Chemistry;
Wilkinson, G., Stone, F. G. A., Abel, E. W., Eds.; Pergamon Press:
Oxford, UK, 1982; Vol. 2, pp 823-862.
(
18) During the preparation of this manuscript a theoretical DFT
study on the arylation of ethanal using a simplified achiral ligand
system was published, see: Rudolph, J .; Rasmussen, T.; Bolm, C.;
Norrby, P. O. Angew. Chem., Int. Ed. 2003, 42, 3002-3005.
(
19) For recent applications of in situ FT-IR see: (a) Evans, D. A.;
Scheidt, K. A.; J ohnston, J . N.; Willis, M. C. J . Am. Chem. Soc. 2001,
1
23, 4480-4491. (b) Sun, X.; Kenkre, S. L.; Remenar, J . L.; Gilchrist,
J . H.; Collum, D. B. J . Am. Chem. Soc. 1997, 119, 4765-4766. (c) Kainz,
S.; Brinkmann, A.; Leitner, W.; Pfaltz, A. J . Am. Chem. Soc. 1999,
1
21, 6421-6429. (d) Marchueta, M.; Olivella, S.; Sol a` , L.; Moyano, A.;
Peric a` s, M. A.; Riera, A. Org. Lett. 2001, 3, 3197-3200.
J . Org. Chem, Vol. 69, No. 7, 2004 2537

Fontes et al.
F IGURE 4. Comparison of the catalyzed vs the background process at 0 °C.
effective core potential (ECP) for zinc, while carbon atoms
are described using a 6-31G* basis set.2
9,30
As a first step we investigated the simple methyl
exchange between two molecules of Me Zn to determine
2
whether the method results were in agreement with the
physical properties of this compound. A four-center
symmetric transition state TS-II (Figure 7) was localized
in which the exchanging methyl groups are 2.27 Å away
from each metal center. From this point, geometry and
energy optimization led to a loosely bound dimer (D-I)
-
1
5
.93 kcal‚mol more stable than TS-II and only 1.53
1
-
kcal‚mol more stable than two monomer molecules.
Relative energies for the methyl exchange process (Table
5
) predict that alkyl (methyl) exchange is a kinetically
F IGURE 5. Background reaction with Ph
temperatures.
2
Zn at different
feasible process but that aggregation between dialkylzinc
species is not thermodynamically favored since entropic
factors should prevail over the small stabilizing effects
2
found for D-I with respect to monomeric Me Zn. This is
in agreement with experimental observations of Me
being a liquid with a low boiling point.
2
Zn
In the case of diphenylzinc, X-ray crystallography
studies show this compound exists as a dimer in the solid
state as has been demonstrated by Bickelhaupt and co-
workers.31 The structure of (Ph
2 2
Zn) was modeled at the
same theory level to further check whether density
functional theory could correctly predict the structure of
this class of compounds. The optimized structure for the
diphenylzinc dimer (D-III) is shown in Figure 7. The
main structural features of D-III fully coincide with the
reported solid-state X-ray structure. The central four-
membered ring is nonplanar with a ring folding along
the Zn-Zn axis of around 25°. The interactions between
2 2
F IGURE 6. Background reaction using the Et Zn/Ph Zn
mixture at different temperatures.
2
the two Ph Zn moieties were correctly predicted: the
distances Zn-C(aromatic) within the two monomers are
they are believed to exist in a monomeric form in solution,
Ph Zn is a solid with a melting point of 107 °C and
exhibits a decreased solubility compared to plain dialky-
lzinc compounds. To gain further insight into the
mechanism and the nature of the species formed when
mixing diphenylzinc and diethylzinc a computational
study was undertaken. All calculations were made using
density functional theory (DFT) with the B3LYP func-
tional. Geometry and energy optimization were run using
a LACVP* basis set that includes Hay and Wadt’s
2
2
.44 and 2.36 Å for the crystal structure and 2.51 and
2
.52 Å for the calculated model (D-III).
For the ethyl-phenyl exchange a four-center transition
state could be found (TS-V) in which the phenyl group
to be transferred is perpendicular to the four-center
2
8
(
(
(
29) Spartan 02′; Wavefuction, Inc.; Irvine, CA.
30) Hay, P. J .; Wadt, W. R. J . Chem. Phys. 1985, 82, 270-283.
31) Markies, P. R.; Schat, G.; Akkerman, O. S.; Bickelhaupt, F.
Organometallics 1990, 9, 2243-2247.
2
538 J . Org. Chem., Vol. 69, No. 7, 2004

2
-Piperidino-1,1,2-triphenylethanol
F IGURE 7.
TABLE 5. Com p u ted Absolu te a n d Rela tive En er gies of
Sp ecies In volved in th e Me2Zn In ter ch a n ge
the diphenylzinc dimer, D-III). Starting from the Et
Zn
D-III the resulting energy profile for the formation
of the mixed EtPhZn is depicted in Figure 8.
The most stable species in the reaction profile cor-
2
1
and /
2
absolute ∆Hf
relative ∆Hf
1
-
compd
(au)
(kcal‚mol )
2
D-I
TS-II
× Me2Zn
-290.784 516
-290.786 966
-290.777 507
1.53
0.00
5.93
1
respond to the starting materials (Et
2
2
Zn + / D-III) and
the symmetrical dimer formed by two molecules of
EtPhZn (D-VI). On the other hand, the mixtures of
discrete (Et
2
Zn + Ph
2
Zn) and (2 × EtPhZn) are energeti-
TABLE 6. Com p u ted Absolu te a n d Rela tive En er gies of
Sp ecies In volved in Et2Zn /P h 2Zn In ter ch a n ge
-1
cally disfavored by approximately 5.76 and 5.17 kcal‚mol .
Taking into account only the monomeric species, the
formation of EtPhZn is moderately favored by only 0.6
absolute ∆Hf
relative ∆Hf
1
-
compd
(au)
(kcal‚mol )
-
1
1/
kcal‚mol , while if we consider the formation of ag-
2D-III + Et2Zn
-752.872 412
-752.863 229
-752.869 430
-752.867 368
-752.872 111
-752.864 166
0.00
5.76
1.87
3.16
0.19
5.17
Ph2Zn + Et2Zn
D-IV
TS-V
D-VI
gregates (dimers) the process is slightly disfavored, ∆E
-1
)
0.19 kcal‚mol . From the FT-IR study (vide supra),
2 2
we can ascertain that when mixing Et Zn and Ph Zn in
a 2:1 ratio, diphenylzinc is not available in high concen-
trations in the reaction mixture since the rate of the
background addition reaction is much slower than when
2
× EtPhZn
using Ph
2
Zn alone. The present theoretical model (Figure
8
) indicates that alkyl-aryl interchange between differ-
ent zinc species is a kinetically facile process since the
transition state identified for the Et-Ph exchange TS-V
represents a low energy barrier. The energies calculated
in the gas phase for the different species predict the
formation of the mixed zinc species as a basically thermo
neutral process. This is not in agreement with the
experimental observations that point to an equilibrium
predominantly shifted toward the formation of EtPhZn.
Still, entropy and solvation factors which are not evalu-
ated in the present study could account for the prefer-
ential formation of EtPhZn. Diphenylzinc has been
calculated to be monomeric in hexane at concentrations
of 0.001-0.005 M, and its solubility has been calculated
to be 0.015 M in the same solvent.31 In the present
F IGURE 8. Energy profile for the formation of EtPhZn.
reaction core (Figure 7). The aromatic carbon atom
directly attached to both metal centers adopts a distorted
tetrahedral geometry with a Zn-C-Zn angle of 72 °.
Most surprisingly, the heat of formation found for TS-V
is lower than that of either of the starting materials
(Ph
2
2
Zn + Et Zn) or of the exchanged product ethylphen-
ylzinc (Table 6 and Figure 8). Perturbation and energy
optimization from TS-V led to the localization of dimers
D-IV and D-VI (Figure 7) with relative energies 1.87 and
-
1
2
arylation reactions and in the FT-IR studies Ph Zn was
used at much higher concentration (0.15-0.25 M) in
toluene. The experimental facts along with the calculated
2 2
model may suggest that mixing Et Zn and Ph Zn (2/1)
0
.19 kcal‚mol , respectively (Table 6). Both D-IV and
D-VI show higher interaction energy than the analogous
dimethylzinc aggregate D-I, as can be deduced from the
shorter Zn-C distances between monomers and bigger
distortion from linear geometry (although less than for
at a concentration between 0.15 and 0.25 M leads to the
J . Org. Chem, Vol. 69, No. 7, 2004 2539

Fontes et al.
F IGURE 9.
SCHEME 2
formation of EtPhZn, which in turn can exist in equilib-
rium between its monomer and its dimer or a higher
aggregate.
Two arguments are in favor of the phenyl group being
transferred to the aldehyde from a mixed species. On one
hand, the comparison of the experimentally determined
activation energies for the noncatalytic transfer of phenyl
from Ph
activation energy for the Ph/Et scrambling, suggest that,
under the Curtin-Hammet principle, if Ph Zn were
present at all in the reaction medium it would lead to
the rapid arylation from Ph Zn, which is not the case.
On the other hand, the complete transfer of the phenyl
groups when using a substoichiometric amount of diphe-
nylzinc can only be explained if transfer is taking place
from a mixed species arising from group exchange
2
Zn and EtPhZn, along with the DFT-calculated
F IGURE 10. PM3(tm) tricyclic transition states for catalyst
1
0.
2
ones with the hydroxyl groups of the ligand, and that the
2
2 2
active catalytic complex when using the Et Zn/Ph Zn
mixture is predominantly 10.34
Com p u ta tion a l Stu d y on th e En a n tioselective
P h en yl Tr a n sfer . On the basis of the preceding results
the asymmetric phenyl transfer from monomeric EtPhZn
catalyzed by 10 and 11 was modeled. The large size of
the transition state structures involved and computa-
tional time restrictions enforced the use of a semiempiri-
cal PM3(tm) method. Transition state structures calcu-
lated at the PM3 level reproduce the results of much
more time-consuming methods (DFT, MP2) and are able
2 2
between Et Zn and Ph Zn. In addition, differential
calorimetric studies performed by Nehl et al. on binary
mixtures of diorganozinc compounds are consistent with
an equilibrium completely shifted toward the mixed
species when using Ph
2
Zn.32
Exp er im en ta l Stu d ies on th e Na tu r e of th e Ca ta -
lytic Sp ecies. The active species in the â-amino-alcohol
catalyzed asymmetric dialkylzinc addition to aldehydes
is a five-member-ring chelate in which the Zn atom is
attached to both nitrogen and oxygen and to an alkyl
group (Figure 9). If only diethylzinc is used, the R group
8
c,35
to predict the sense of enantioselection in most cases.
Thus the stereochemical outcome of the reaction was
studied considering the four diastereomeric transition
state structures (anti-(S), syn-(S), anti-(R), syn-(R)) which,
according to the seminal work of Noyori, should deter-
mine the final selectivity.36 Conformational isomerism
coming from the rotation of the nontransferred ethyl
groups and the five-membered-ring chelate with the
ligand was thoroughly studied. The four most stable TS
structures located for each catalytic species 10 and 11
are depicted in Figures 10 and 11 along with their
corresponding relative energies. Vibration analysis re-
vealed imaginary stretching frequencies of 339i (Et-Anti-
on Zn will be ethyl (10), and when only Ph
Zn-Ph moiety should be part of the active catalyst (11).
When using a Ph Zn/Et Zn mixture a more complex
2
Zn is used a
2
2
situation arises and, in principle, either Zn-Et or Zn-
Ph moieties could form and serve as active catalytic
species. To respond to this question a simple experiment
was envisaged. Diethylzinc (0.64 mmol) and diphenylzinc
(0.64 mmol) were mixed together in toluene and allowed
to equilibrate for 15-30 min to allow for the formation
of the mixed EtPhZn. To this mixture the amino alcohol
(33) In a similar experiment, a preequilibrated mixture of Et
2
Zn (1
5
(1.0 mmol) was added in one portion and the volume
mmol) and Ph Zn (0.5 mmol) was treated with 1.5 mmol of 5. Upon
2
of generated gas was measured. If 10 were formed
exclusively an equal amount of benzene would be pro-
duced, while if the alkyl group on Zn were phenyl (11)
the evolution of 1.0 mmol of ethane should be observed
addition of the amino alcohol 11 mL of ethane was collected, which
corresponds to 0.5 mmol of Et Zn in excess, out of a maximum of 33
2
mL.
(34) The present experiment suggests that formation of 10 is
kinetically favored over 11. However, during the course of the reaction
scrambling between ethyl and phenyl on the Zn center of the catalyst
cannot be ruled out.
(Scheme 2). The gas evolution was measured with a gas
buret. Upon carrying out the experiment, 2 mL (0.09
mmol) of ethane was collected out of a theoretical
maximum of 22.4 mL (1.0 mmol).33 This result indicates
that the phenyl groups on zinc react faster than the alkyl
(35) (a) Kozlowski, M. A.; Dixon, S. L.; Panda, M.; Lauri, G. J . Am.
Chem. Soc. 2003, 125, 6614-6615. (b) Panda, M.; Phuan, P.-W.;
Kozlowski, M. C. J . Org. Chem. 2003, 68, 564-571. (c) Goldfuss, B.;
Steigelmann, M.; Khan, S. I.; Houk, K. N. J . Org. Chem. 2000, 65,
7
7-82. (d) Vazquez, J .; Peric a` s, M. A.; Maseras, F.; Lledos, A. J . Org.
Chem. 2000, 65, 7303-7309.
(32) Nehl, H.; Scheidt, W. R. J . Organomet. Chem. 1985, 289, 1-8.
(36) Yamakawa, M.; Noyori, R. Organometallics 1999, 18, 128-133.
2
540 J . Org. Chem., Vol. 69, No. 7, 2004

2
-Piperidino-1,1,2-triphenylethanol
Con clu sion s
In summary, the present studies provide a precise
picture of the different aspects controlling the stereo-
chemical outcome of the catalytic phenylation of alde-
hydes: The importance of background, noncatalytic
additions for the most usual reacting systems has been
assessed, and the reasons for the superior performance
of Ph Zn/Et Zn as the phenylation reagent thus have
2 2
been demonstrated. We have provided experimental
evidence in favor of the EtPhZn as the largely predomi-
nating phenyl delivering species when mixtures of Ph
2
Zn
and Et Zn of the usual composition (1:2) are employed,
2
and have also provided experimental evidence on the
nature of the active catalytic species when an amino
alcohol ligand is combined with this mixed diorganozinc
reagent.
The nature of the equilibria involved in the formation
of the mixed diorganozinc species and the importance of
the energy barriers associated with their establishment
have been addressed by theoretical methods with DFT
theory.
All together, the kinetic, experimental, and theoretical
studies reported here provide a comprehensive picture
of the different aspects involved in the most synthetically
useful phenylation reaction.
F IGURE 11. PM3(tm) tricyclic transition states for catalyst
1.
1
S) and 336i cm-1 (Ph-Anti-S) for the forming C-C bond
between the transferred aryl and the aldehyde. As occurs
in the ethylation process, there exists a considerable
energy gap between syn and anti TS structures, the latter
ones being more favored. In this context, reaction selec-
tivity should be determined exclusively on the basis of
Anti-S and Anti-R energy variations. To our delight, the
stereochemical course of the arylation reaction was
correctly predicted by the present model. Thus, when
employing (R)-2-piperidino-1,1,2-triphenylethanol 5, Anti-
S, the most stable TS structure would lead to the
experimentally observed (S) diarylcarbinols. This trend
was observed for both catalytic species 10 and 11. The
energy difference between TS structures, Et-Anti-S - Et-
From a more practical perspective, we have identified
a readily available, yet extremely active ligand for the
highly enantioselective phenylation of aromatic alde-
hydes showing promise for large-scale application and
have shown its suitability for a full range of substrates.
The enantioselectivity recorded with this ligand has been
rationalized through a theoretical PM3(tm) study of the
nature of the diasteromeric transition states involved in
the phenylation of benzaldehyde when 2-piperidino-1,1,2-
triphenylethanol is used as ligand in combination with
EtPhZn.
Application of this catalytic system to the phenylation
of other unsaturated systems is underway in our labo-
ratories.
-
1
Anti-R (3.0 kcal‚mol ) and Ph-Anti-S - Ph-Anti-R (5.5
-
1
Exp er im en ta l Section .
kcal‚mol ), is in agreement with the experimentally
observed enantioselectivity. Most interestingly, on the
basis of this model, the catalyst containing a Ph-Zn
moiety 11 should, in principle, be more selective than 10.
Experimentally, when performing the arylation exclu-
Ar yla t ion R ea ct ion s u n d er H igh Dilu t ion Con d i-
tion s: Typ ica l Exp er im en ta l P r oced u r e. To a Schlenk
flask under nitrogen were added solvent (25 mL), (R)-2-
piperidino-1,1,2-triphenylethanol (36 mg, 0.1 mmol), and di-
ethylzinc 1 M in hexane (0.025 mL, 0.025 mmol). The reaction
mixture was stirred at room temperature for 15 min and then
sively with Ph
ee, Table 2, entry 6), although this is still lower than that
obtained when using the Et Zn/Ph Zn mixture. We
2
Zn a notable selectivity is observed (88%
2
transferred via cannula to Ph Zn (55 mg, 0.25 mmol). Stirring
2
2
was continued for 20 min and then the aldehyde (0.25 mmol)
was added. The reaction mixture was stirred at room temper-
ature for 24 h and quenched with HCl (1 M, 15 mL). The
aqueous layer was extracted with hexane and the organic
believe that under these circumstances two different
trends coexist: an enhanced selectivity for aryl transfer
due to higher selectivity of catalyst 11 and a deleterious
uncatalyzed reaction that lowers the product enantio-
layers were combined and dried with Mg(SO
of the solvent in vacuo and purification by flash chromatog-
raphy (SiO , hexane, and hexane/ethyl acetate 98:2) gave the
4 2
) . Evaporation
2 2
meric excess. Thus, when using the Et Zn/Ph Zn mixture,
2
the benefits of shutting down the uncatalyzed process pay
off the use of a slightly less selective catalyst (10). The
present theoretical study shows that the amino alcohol-
catalyzed asymmetric aryl transfer can be conveniently
modeled at the PM3(tm) level by means of the energy
evaluation of anti and syn 5/4/4 tricyclic TS structures.
Moreover, the model of 10 and 11 as active catalysts and
EtPhZn as a phenyl source can correctly predict the sense
and the selectivity of the overall process.
secondary alcohol. Enantiomeric excess was determined by
HPLC.
Ar yla tion Rea ction s Usin g a Mixtu r e of Et
Typ ica l Exp er im en ta l P r oced u r e. Diphenylzinc (70 mg,
.32 mmol) was charged into a Schlenk flask in a drybox. The
flask was removed from the drybox and freshly distilled solvent
6 mL) was added followed by Et Zn 1 M in hexane (0.66 mL,
.66 mmol). The reaction mixture was stirred for 20 min at
2 2
Zn /P h Zn :
0
(
2
0
room temperature and then (R)-2-piperidino-1,1,2-triphenyl-
ethanol (18 mg, 0.05 mmol) was added. The resulting solution
J . Org. Chem, Vol. 69, No. 7, 2004 2541

Fontes et al.
).
was stirred for 20 min and then cooled to the desired temper-
ature. Next, the aldehyde (0.5 mmol) was added and the
reaction mixture was monitored by TLC. The reaction was
(S)-(3-Tolyl)p h en ylm eth a n ol.37 [R]
(silica gel, hexane/AcOEt, 3:1): 0.54. H NMR (400 MHz,
CDCl ): δ 2.17 (d, J ) 3.6 Hz, 1H), 2.33 (s, 3H), 5.81 (d, J )
3.6 Hz, 1H), 7.07-7.28 (m, 5H), 7.31-7.40 (m, 4H) ppm.
NMR (100 MHz, CDCl ): δ 21.7, 76.5, 123.8, 126.7, 127.4,
127.7, 128.5, 128.6, 128.7, 138.4, 144.0, 144.1 ppm. F NMR
D
-12.5 (c 0.60, CHCl
3
1
R
f
3
1
3
quenched by adding NH
with CH Cl
(3 × 10 mL). The organic layers were combined
and dried over MgSO and the solvent was evaporated in
vacuo. Flash chromatography of the crude product (SiO
4
Cl solution (10 mL) and was extracted
C
2
2
3
1
9
4
2
,
(376 MHz, CDCl
(major) ppm.
3
) Mosher esther derivative: -73.84, -73.96
hexane, and hexane/ethyl acetate 98:2) afforded the pure
secondary alcohol. Enantiomeric excess was determined by
chiral HPLC.
Ar yla tion Rea ction s Usin g 1.5 m ol % of Ca ta lyst. The
reactions were conducted using a Cooled Carousel Reaction
Station under argon atmosphere. To a solution of diphenylzinc
(S)-(2-Tolyl)p h en ylm eth a n ol.38
R
f
(silica gel, hexane/
AcOEt, 3:1): 0.53. IR (KBr): νmax 3240 (broad), 3075, 1603,
-
1
1
1491 cm . H NMR (400 MHz, CDCl
(s, 1H), 5.94 (s, 1H), 7.10-7.30 (m, 8H), 7.47-7.49 (m, 1H)
ppm. 13C NMR (100 MHz, CDCl
3
): δ 2.21 (s, 3H), 2.28
3
): δ 19.5, 73.4, 126.2, 126.4,
(
1
586 mg, 2.67 mmol) in toluene (50 mL) was added diethylzinc
M in hexane (5.40 mL, 5.40 mmol) at room temperature. The
reaction was stirred at this temperature for 20 min and then
2 mg of (R)-2-piperidino-1,1,2-trifeniletanol (0.062 mmol) was
127.2, 127.6, 127.7, 128.6, 130.6, 135.5, 141.5, 142.9 ppm.
HPLC: Daicel CHIRALCEL-OD. Hexane/i-PrOH 99:1, 1.0 mL/
min, λ ) 254 nm, t
CHIRALPACK-AD. Hexane/i-PrOH 98:2, 0.5 mL/min, λ ) 254
nm, t (S) ) 36.48 min, t (R) ) 38.55 min.
(S)-(2-Ch lor op h en yl)p h en ylm eth a n ol.
R
(S) ) 64.2 min, t ) 58.2 min. HPLC: Daicel
R
2
added. After the mixture had been stirred for 20 min, 6.6 mL
of the solution was transferred to each carousel tube and the
station was cooled to 10 °C by means of an immersion cooler
with a flexible cooling coil. The aldehyde was added (0.5 mmol)
and the resulting reaction was allowed to proceed for 4 h. The
R
R
3
9
R
f
(silica gel,
hexane/AcOEt, 3:1): 0.43. H NMR (400 MHz, CDCl ): δ 2.58
(s broad, 1H), 6.17 (s, 1H), 7.16-7.37 (m, 8H), 7.56-7.58 (m,
1H) ppm. 13C NMR (100 MHz, CDCl
1
3
3
): δ 72.7, 127.0, 127.2,
reactions were quenched by addition of NH
and were extracted with CH Cl
(3 × 10 mL). The organic
layers were combined and dried with Mg(SO . Evaporation
of the solvent under reduced pressure and purification by
column chromatography (SiO , hexane, and hexane/AcOEt 98:
) gave the desired secondary alcohol. Enantiomeric excess was
determined by chiral HPLC.
S)-(4-Tolyl)p h en ylm eth a n ol.
4
Cl solution (10 mL)
127.8, 128.1, 128.6, 128.8, 129.6, 132.6, 141.1, 142.3 ppm.
HPLC: Daicel CHIRALCEL-OD. Hexane/i-PrOH 90:10, 0.5
2
2
4
)
2
mL/min, λ ) 254 nm, t
R
(S) ) 19.8 min, t (R) ) 15.9 min.
R
40
(S)-(2-F lu or op h en yl)p h en ylm eth a n ol.
R
f
(silica gel,
1
hexane/AcOEt, 3:1): 0.50. H NMR (400 MHz, CDCl ): δ 2.47
3
2
2
(s, 1H), 6.14 (s, 1H), 6.99-7.00 (m, 1H), 7.13-7.17 (m, 1H),
7.23-7.36 (m, 4H), 7.39-7.41 (m, 2H), 7.48-7.53 (m, 1H) ppm.
1
3
6
a
3
C NMR (100 MHz, CDCl ): δ 70.2 (d, J CF ) 3 Hz), 115.5 (d,
(
R
f
(silica gel, hexane/
AcOEt, 3:1): 0.50. H NMR (400 MHz, CDCl ): δ 1.26 (s, 1H),
.32 (s, 3H), 5.79 (s, 1H), 7.13 (d, J ) 8 Hz, 2H), 7.24-7.37
1
J
1
CF ) 21.4 Hz), 124.4 (d, J CF ) 3.1 Hz), 126.5, 127.7, 127.8,
27.9, 128.6, 129.2 (d, J CF ) 8.4 Hz), 131.1 (d, J CF ) 13 Hz),
3
2
(
160.0 (d, J CF ) 244.6 Hz) ppm. 19F NMR (376 MHz, CDCl
):
1
3
3
m, 7H) ppm. C NMR (100 MHz, CDCl
26.7, 127.6, 128.6, 129.3, 137.4, 141.1, 144.1 ppm. MS (EI)
3
): δ 21.2, 76.2, 126.6,
1
18.84 ppm. HPLC: Daicel CHIRALCEL-OD. Hexane/i-PrOH
1
+
95:5, 0.5 mL/min, λ ) 270 nm, t
min.
R
(S) ) 26.5 min, t
S)-(4-Bip h en yl)p h en ylm eth a n ol.6a
R (silica gel, hexane/
f
R
(R) ) 23.1
m/e: 197 (M - H ), 183 (M - CH
3
), 121 (M - C
). HPLC: Daicel CHIRALCEL-OD. Hexane/
i-PrOH 90:10, 0.5 mL/min, λ ) 254 nm, t (S) ) 19.1 min, t (R)
6 5
H ), 91
+ +
7 7 6 5
(C H ), 77 (C H
(
R
R
1
AcOEt, 3:1): 0.42. H NMR (400 MHz, CDCl ): δ 2.30 (s, 1H),
5.87 (s, 1H), 7.23-7.37 (m, 4H), 7.40-7.45 (m, 6H), 7.54-7.57
m, 4H) ppm. 13C NMR (100 MHz, CDCl
): δ 76.2, 126.7, 127.1,
127.2, 127.4, 127.5, 127.8, 128.7, 128.9, 140.6, 140.9, 143.0,
43.9 ppm. HPLC: Daicel CHIRALCEL-OD. Hexane/i-PrOH
3
)
21.1 min.
(
S)-(2-Na p h th yl)p h en ylm eth a n ol.6a
f
R (silica gel, hexane/
(
3
AcOEt, 3:1): 0.47. IR (KBr): νmax 3349 (broad), 3051, 1600,
-
1
1
1
(
3
1
1
492 cm . H NMR (400 MHz, CDCl
3
): δ 2.34 (s, 1H), 5.99
1
s, 1H), 7.24-7.35 (m, 3H), 7.40-7.47 (m, 5H), 7.77-7.83 (m,
H), 7.88 (s, 1H) ppm. 1 C NMR (100 MHz, CDCl
3
94:6, 0.8 mL/min, λ ) 254 nm, t
min.
R
(S) ) 32.4 min, t
R)-(E)-1,3-Dip h en yl-2-p r op en ol.6b
R (silica gel, hexane/
f
R
(R) ) 29.8
3
): δ 76.4,
24.7, 125.0, 125.9, 126.2, 126.7, 127.7, 128.1, 128.3, 128.5,
32.9, 133.3, 141.1, 143.6 ppm. HPLC: Daicel CHIRALCEL-
(
1
AcOEt, 3:1): 0.50. H NMR (400 MHz, CDCl ): δ 2.35 (s broad,
H), 5.34 (d, J ) 7 Hz, 1H), 6.37 (dd, J ) 7 Hz, J ) 16 Hz, 1H),
.65 (d, J ) 16 Hz, 1H), 7.23-7.37 (m, 4H), 7.19-7.42 (m, 6H)
3
OD. Hexane/i-PrOH 90:10, 1.0 mL/min, λ ) 254 nm, t
6.0 min, t (R) ) 21.1 min.
S)-(4-Meth oxyp h en yl)p h en ylm eth a n ol.
hexane/AcOEt, 3:1): 0.31. IR (KBr): νmax 3350 (broad), 3050,
R
(S) )
1
6
1
R
6
a
(
f
R (silica gel,
1
3
ppm. C NMR (100 MHz, CDCl
28.0, 128.8, 128.9, 130.8, 131.8, 136.8, 143.1 ppm. HPLC:
Daicel CHIRALCEL-OD. Hexane/i-PrOH 90:10, 1.0 mL/min,
3
): δ 75.4, 126.6, 126.9, 127.4,
1
-
1 1
1
3
6
5
3
613, 1514 cm . H NMR (400 MHz, CDCl ): δ 2.18 (s, 1H),
.78 (s, 3H), 5.80 (s, 1H), 6.88-6.85 (m, 3H), 7.24-7.38 (m,
H) ppm. HPLC: Daicel CHIRALCEL-OD. Hexane/i-PrOH 95:
, 1.0 mL/min, λ ) 254 nm, t
λ ) 254 nm, t
(
R
(S) ) 15.9 min, t (R) ) 20.4 min.4
R
1
R)-(E)-2-Meth yl-1,3-d ip h en yl-2-p r op en ol. [R] -4.85
D
R
(S) ) 23.9, min, t (R) ) 22.7
R
1
(c 1.6 /CHCl
3
). R
f
(silica gel, hexane/AcOEt, 3:1): 0.47. H NMR
): δ 1.73 (d, J ) 1.2 Hz, 3H), 2.11 (s broad,
H), 5.27 (s, 1H), 6.77 (s broad, 1H), 7.21-7.23 (m, 1H), 7.28-
min.
(400 MHz, CDCl
3
6
b
(
R)-1-P h en ylh ep th a n ol.
R
f
(silica gel, hexane/AcOEt,
3
1
7
1
3
3
4
:1): 0.53. H NMR (400 MHz, CDCl ): δ 0.86 (t, J ) 6.8 Hz,
13
.37 (m, 7H), 7.42-7.44 (m, 2H) ppm. C NMR (100 MHz,
): δ 14.3, 79.7, 126.3, 126.7, 126.8, 127.8, 128.4, 128.7,
29.3, 137.7, 139.8, 142.3 ppm. HPLC: p-nitrobenzoate deriva-
tive: Daicel CHIRALCEL-OD. Hexane/i-PrOH 95:5, 0.5 mL/
min, λ ) 254 nm, t (S) ) 26.9 min, t (R) ) 29.8 min.
R)-2-Eth yl-1-p h en ylbu ta n ol. [R] -8.08 (c 0.26, CHCl
(silica gel, hexane/AcOEt, 3:1): 0.66. IR (KBr): νmax 3417
H), 1.21-1.43 (m, 8H), 1.64-1.83 (m, 2H), 1.97 (s broad, 1H),
.63 (t, J ) 6.6 Hz, 1H), 7.24-7.28 (m, 1H), 7.31-7.33 (m, 4H)
CDCl
3
1
1
3
ppm. C NMR (100 MHz, CDCl ): δ 14.3, 22.8, 26.0, 29.4, 32.0,
3
3
9.3, 74.9, 126.1, 127.7, 128.6, 145.2 ppm. HPLC: Daicel
CHIRALCEL-OD. Hexane/i-PrOH 95:5, 0.5 mL/min, λ ) 254
nm, t (S) ) 15.5 min, t (R) ) 14.6 min.
S)-(4-Ch lor op h en yl)p h en ylm eth a n ol.
hexane/AcOEt, 3:1): 0.48. IR (KBr): νmax 3260 (broad), 3050,
R
R
42
(
D
3
).
R
R
R
f
6
b
(
f
R (silica gel,
(
(
37) Guijarro, D.; Yus, M. Tetrahedron 2000, 56, 1135-1138.
-
1 1
1
5
600, 1493, cm . H NMR (300 MHz, CDCl
3
): δ 1.26 (s, 1H),
.81 (s, 1H), 7.25-7.35 (m, 9H) ppm. C NMR (75 MHz,
): δ 75.6, 126.5, 127.9, 128.6, 128.6, 128.7, 133.3, 142.2,
43.5 ppm. HPLC: acetate derivative: Daicel CHIRALCEL-
OD. Hexane/i-PrOH 99:1, 0.2 mL/min, λ ) 254 nm, t (S) )
7.9 min, t (R) ) 39.6 min. HPLC: p-nitrobenzoate deriva-
tive: Daicel CHIRALCEL-OD. Hexane/i-PrOH 98:2, 0.5 mL/
min, λ ) 254 nm, t (S) ) 44.1 min, t (R) ) 40.5 min.
38) Huang, W.-S.; Pu, L. Tetrahedron Lett. 1999, 41, 145-149.
1
3
(39) Zhao, G.; Li, X. G.; Wang, X. R. Tetrahedron: Asymmetry 2001,
CDCl
1
3
12, 399-403.
(
40) Ohkuma, T.; Koizumi, M.; Ikehira, H.; Yokozawa, T.; Noyori,
R. Org. Lett. 2000, 2, 659-662.
R
(
41) Nudelman, N. S.; Garcia, G. V. J . Org. Chem. 2001, 66, 1387-
3
R
9
4.
(42) Calas, M.; Calas, B.; Giral, L. Bull. Soc. Chim. Fr. 1973, 2079-
2086.
R
R
2
542 J . Org. Chem., Vol. 69, No. 7, 2004

2
-Piperidino-1,1,2-triphenylethanol
(
broad), 3050, 2964, 1451, 1017 cm-1. 1H NMR (400 MHz,
CDCl ): δ 0.85 (t, J ) 7 Hz, 3H), 0.90 (t, J ) 7 Hz, 3H), 1.15-
.23 (m, 1H), 1.24-1.33 (m, 2H), 1.40-1.61 (m, 2H), 1.75
broad s, 1H), 4.64 (d, J ) 6.4 Hz, 1H), 7.25-7.28 (m, 1H),
0.5 mL/min, λ ) 254 nm, t
S
) 20.4 min, t
R
R
) 31.4 min. GC:
3
â-DEX 120 Column, 120 °C: t
(S) ) 44.3 min, t (R) ) 45.5
R
1
(
7
1
min.
(S)-(3-Meth oxyp h en yl)p h en ylm eth a n ol.6a
R
(silica gel,
): δ 2.40
f
.31-7.34 (m, 4H) ppm. 13C NMR (100 MHz, CDCl
1.5, 20.8, 22.0, 48.0, 76.1, 126.7, 127.5, 128.5, 144.3 ppm.
): δ 11.1,
hexane/AcOEt, 3:1): 0.33. H NMR (400 MHz, CDCl
1
3
3
(broad s, 1H), 3.75 (s, 3H), 5.76 (s, 1H), 6.77-6.80 (m, 1H),
1
3
HPLC: Daicel CHIRALCEL-OD. Hexane/i-PrOH 93:7, 0.5 mL/
min, λ ) 254 nm, t (S) ) 11.1 min, t (R) ) 16.4 min.
R)-Cycloh exylp h en ylm eth a n ol. [R]
CHCl ). R (silica gel, hexane/AcOEt, 3:1): 0.61. H NMR (400
MHz, CDCl ): δ 0.88-1.28 (m, 5H), 1.36-1.39 (m, 1H), 1.56-
.68 (m, 3H), 1.75-1.81 (m, 2H), 1.97-2.00 (m, 1H), 4.36 (d,
6.91-6.93 (m, 2H), 7.20-7.36 (m, 6H) ppm. C NMR (100
MHz, CDCl ): δ 55.3, 76.2, 112.2, 113.1, 119.0, 126.6, 127.7,
128.6, 129.6, 143.8, 145.6, 159.8 ppm. HPLC: Daicel CHIRAL-
CEL-OD. Hexane/i-PrOH 95:5, 1.0 mL/min, λ ) 254 nm, t (S)
) 27.1 min, t (R) ) 42.0 min.
R
R
4
3
3
(
D
+17.0 (c 0.60,
1
3
f
R
3
R
1
1
3
J ) 7 Hz, 1H), 7.25-7.36 (m, 5H) ppm. C NMR (100 MHz,
CDCl ): δ 26.2, 26.3, 26.6, 29.0, 29.5, 45.2, 79.6, 126.8, 127.6,
28.4, 143.8 ppm. HPLC: Daicel CHIRALCEL-OD. Hexane/
i-PrOH 95:5, 0.5 mL/min, λ ) 254 nm, t (S) ) 14.8 min, t (R)
18.0 min.
R)-2,2-Dim eth yl-1-p h en ylp r op a n ol.6c
(silica gel, hex-
ane/AcOEt, 3:1): 0.62. IR (KBr): νmax 3424 (broad), 3050, 2954,
Ack n ow led gm en t. We gratefully thank DGI-MCYT
(BQU2002-02459) and DURSI (2001SGR50) for finan-
cial support. X.V. thanks the MCYT for a Ram o´ n y Cajal
research position. We thank Dr. Alex Comely for reading
the manuscript. We also thank Mr. Daniel Font for his
assistance with the in situ FT-IR experiments.
3
1
R
R
)
(
R
f
-
1 1
1
451, 1385 cm . H NMR (400 MHz, CDCl
.85 (s, 1H), 4.40 (s, 1H), 7.24-7.32 (m, 5H) ppm. C NMR
): δ 26.1, 35.8, 82.6, 127.4, 127.7, 127.8, 142.3
ppm. HPLC: Daicel CHIRALCEL-OD. Hexane/i-PrOH 98:2,
3
): δ 0.92 (s, 9H),
Su p p or tin g In for m a tion Ava ila ble: General experimen-
tal methods and Cartesian coordinates and energies for
molecular modeling compounds and transition states. This
material is available free of charge via the Internet at
http://pubs.acs.org.
13
1
(100 MHz, CDCl
3
(43) Langer, F.; Schwink, L.; Devasagayaraj, A.; Chavant, P.-Y.;
Knochel, P. J . Org. Chem. 1996, 61, 8229-8243.
J O035824O
J . Org. Chem, Vol. 69, No. 7, 2004 2543