Enantioselective ethylation of aromatic aldehydes catalysed by titanium(IV)-bis-BINOLate-2′,2″-propylether complexes: An inside view of the catalytic active species

Article abstract of DOI:10.1016/j.molcata.2010.04.003

¹H NMR, CD, ESI-MS and PM6 calculation were used to investigate the nature of the catalytic active species, using bis-BINOLate-2′, 2″-propylether-titanium complexes in ethylation of aldehdyes.

Full text of DOI:10.1016/j.molcata.2010.04.003

Journal of Molecular Catalysis A: Chemical 325 (2010) 91–97
Contents lists available at ScienceDirect
Journal of Molecular Catalysis A: Chemical
journal homepage: www.elsevier.com/locate/molcata
Enantioselective ethylation of aromatic aldehydes catalysed by
titanium(IV)–bis-BINOLate-2ꢀ,2ꢀꢀ-propylether complexes: An inside view of the
catalytic active species
Artur R. Abreua,b, Mirtha Lourenc¸ oa, Daniel Peralb, Mário T.S. Rosadoa, Maria E.S. Eusébioa,
Òscar Palaciosb, J. Carles Bayónb,1, Mariette M. Pereiraa,∗
a Departamento de Química, Universidade de Coimbra, Rua Larga, 3004-535 Coimbra, Portugal
b Departament de Química, Universitat Autònoma de Barcelona, Bellaterra, 08193 Barcelona, Spain
a r t i c l e i n f o
a b s t r a c t
A series of bis-BINOL-2ꢀ,2ꢀꢀ-propyl ethers with different substituents at the propyl ether bridge, were
investigated in the asymmetric titanium catalysed ethylation of arylaldehydes with Et2Zn, with con-
versions up to 99% and enantiomeric excesses up to 80%. Semiempirical PM6 calculations indicate that
the higher ability of the unsubstituted ligand to form chelated titanium complexes could be related to
its higher enantioselectivity. Catalytic experiments with partially optically enriched ligand put in evi-
dence a negative non-linear effect (−)-NLE that suggest the presence of two ligand molecule in the
active titanium species. Further catalytic data, together with 1H NMR and circular dichroism (CD) titra-
tions of the ligand with Ti(iPrO)4, as well as ESI-MS experiments, allow to propose a trinuclear species
[Ti3L2(OiPr)8] (L = dianion of (1R,1ꢀR)-2ꢀ,2ꢀꢀ-(propane-1,3-diyl)bis(oxy)di-1,1ꢀ-binaphthyl-2-ol) as respon-
sible for the catalytic asymmetric addition of Et2Zn to aldehydes. This catalytic species is only formed
with great Ti(OiPr)4/ligand molar ratios (ca. 8).
Article history:
Received 16 January 2010
Received in revised form 7 April 2010
Accepted 8 April 2010
Available online 24 April 2010
Keywords:
Asymmetric alkylation
Aldehyde
Titanium tetraisopropoxide
Diethylzinc
bis-BINOL-2ꢀ,2ꢀꢀ-ethers
Mechanism
© 2010 Elsevier B.V. All rights reserved.
1. Introduction
is difficult to establish owing to the kinetic lability of titanium
alkoxide ligands [5]. However, the structure in solution of these
The enantioselective addition of organometallic reagents to
aldehydes affords optically active secondary alcohols. This is a
very important reaction in organic synthesis, since chiral secondary
alcohols are components of many naturally occurring biologically
active compounds. These alcohols are also valuable intermediates
for the synthesis of other functionalities, such as halides, amides,
esters, and ethers [1].
In recent years, the enantioselective addition of organozinc
reagents to aldehydes has been extensively studied [2]. Numer-
ous catalytic systems have been developed for this reaction, often
providing excellent enantioselectivities [3].
Among the countless examples of metal complexes of BINOL as
asymmetric catalysts, a key role is played by the Ti(IV) complexes
family. In spite of their synthetic importance, there is little informa-
tion regarding the structure of these titanium complexes [4]. Only
[BINOLate)Ti(OiPr)2] and its derivatives are relatively well char-
acterized. However, the structure in solution of these complexes
complexes is difficult to establish owing to the kinetic lability of
titanium alkoxide ligands. Moreover, the characterization of active
catalysts is hampered by aggregation phenomena [6].
Catalysts containing partially hydrogenated BINOL ligands,
5,5ꢀ,6,6ꢀ,7,7ꢀ,8,8ꢀ-octahydro-1,1ꢀ-bi-2-naphthol (H8-BINOL) [7], and
5,6,7,8-tetrahydro-1,1ꢀ-bi-2-naphthol (H4-BINOL) [8], exhibited in
some cases better stereoselectivity than those obtained from the
corresponding BINOL catalysts [9]. This has been attributed to the
steric and electronic modulation produced by the saturated frag-
ment in the binaphthyl backbone [10].
It is well known that the accurate identification of the nature
and structure of enantioselective catalysts or catalytic precursors is
always a significant piece of information, since it may open the way
to the rational design of new and more efficient chiral auxiliaries.
In some special cases, the structural elucidation of the catalytic
species represents itself an important task from the spectroscopic
point of view.
We have recently reported the synthesis of a series of bis-
BINOL-2ꢀ,2ꢀꢀ-ether ligands and their catalytic evaluation in the
enantioselective ethylation of benzaldehyde. Among the bis-
BINOL-2ꢀ,2ꢀꢀ-ethers tested, those containing a propyl ether bridge
showed the best performance [11]. Herein we report the effect of
the substituents on the propyl ether bridge on the activity and
∗
Corresponding authors. Tel.: +351 239852080, fax: +351 239827703.
E-mail addresses: joancarles.bayon@uab.es (J.C. Bayón), mmpereira@qui.uc.pt
(M.M. Pereira).
1
Tel.: +34 935812889; fax: +34 935812477.
1381-1169/$ – see front matter © 2010 Elsevier B.V. All rights reserved.
doi:10.1016/j.molcata.2010.04.003

92
A.R. Abreu et al. / Journal of Molecular Catalysis A: Chemical 325 (2010) 91–97
selectivity of catalytic ethylation of aromatic aldehydes. It should
also be pointed out that this paper describes the first insight into the
mechanistic studies of catalytic alkylation of aldehydes using bis-
BINOLate–Ti complexes. So, an insight view of the catalytic active
species is described, using several approaches, including detailed
desired alkane ditosylate (1.3 mmol) in dry DMF was then slowly
added to the previous mixture (0 ◦C, 1 h). After the addition of
ditosylalkane was completed, the reaction was stirred for 6 h at
80 ◦C. After cooling the reaction mixture, water was added drop-
wise (0 ◦C) and the organic compound was extracted with CH2Cl2.
The organic layer was washed with water and a brine solution, and
the concentrated organic phase was purified by flash chromatog-
raphy (silica, CH2Cl2/n-hexane 2:1). The crude was recrystallized
from AcOEt/PriOH.
analysis of the catalytic data, spectroscopic characterization by 1
NMR, circular dichroism (CD) and ESI-MS of solution samples.
H
2. Experimental
1,3-bis[(R)-2ꢀ-(benzyloxy)-1,1ꢀ-binaphthyl-2-yloxy]propane 5a
was obtained with 85% yield and data is in agreement with our
previously described procedure [11].
2.1. General
(1R,1ꢀꢀR)-2ꢀ,2ꢀꢀ-(2,2-dimethylpropane-1,3-diyl)bis(oxy)bis(2-
(benzyloxy)-1,1ꢀ-binaphthyl 3a, was obtained in 85% yield, as a
white powder. 1H NMR (400 MHz, CDCl3) ı 7.83 (dd, J = 14.3, 6.4 Hz,
8H), 7.40–7.15 (m, 10H), 7.13–6.92 (m, 10H), 6.82 (d, J = 8.7 Hz,
16H), 4.81 (m, 4H), 3.19 (d, J = 8.3 Hz, 2H), 3.05 (d, J = 8.3 Hz, 2H),
0.18 (s, 6H). 13C NMR (101 MHz, CDCl3) ı 154.09, 154.00, 137.73,
134.27, 134.06, 129.38, 129.11, 129.03, 128.11, 127.90, 127.71,
127.23, 126.66, 126.22, 126.20, 125.55, 125.45, 123.77, 123.32,
121.09, 119.41, 116.10, 114.59, 73.45, 71.21, 35.71, 21.18; MS
(ESI): m/z = 843.3411 (M+Na), calcd. for C59H48O4Na+ 843.3445.
(iii) A BBr3 solution (1 M in CH2Cl2, 1.55 mL), was added to
a dry CH2Cl2 solution of 3a or 5a (0.94 mmol, 10 mL) cooled at
−78 ◦C and the reaction was stirred for 2 h at this temperature.
When the mixture reached room temperature, the organic phase
was washed with 2 M HCl(aq), and then dried over MgSO4. The
crude obtained by evaporation was purified by flash chromatog-
raphy (silica, CH2Cl2). The compounds were recrystallized from
toluene/n-hexane.
All catalytic and synthetic reactions were performed using stan-
dard Schlenk techniques, under N2 inert atmosphere. Glassware
was oven-dried. Solvents were purified by standard procedure and
reagents were used as received. 1H NMR and 13C NMR spectra
were recorded in CDCl3 solution on Bruker 250, 300 or 400 spec-
trometers. Chemical shifts are relative to SiMe4. Circular dichroism
(CD) measurements were gathered with a JASCO model J-715
spectrophotometer equipped with a computer (J-700 software,
JASCO). These measurements were carried out at a constant tem-
perature (25 ◦C) maintained by a Peltier PTC-351 apparatus (TE
Technology Inc., Traverse City, MI, USA). A high resolution ESI
mass spectrometer Bruker microTOFQ was used to characterize
the new ligands. Medium resolution spectra of catalytic solutions,
prepared dissolving Ti(OiPr)4 plus ligand in CH2Cl2/PriOH (10/1)
were recorded on a Brucker Squire 3000 equipped with an ESI
source. Conversion and chemoselectivity were obtained by gas
chromatography on an Agilent-6890, equipped with a capillary
HP5 column (30 m × 0.32 mm i.d., 0.25 ␮m film thickness, carrier
gas N2, F.I.D. detector). The enantiomeric excesses and absolute
configuration were measured with Konik-300C gas chromato-
graph equipped with ␤-cyclodextrin capillary column (Supelco
␤-Dex120, 30 m × 0.25 mm). The configuration of the alcohols was
determined by comparison with optically pure samples of (R) and
(S)-2-phenyl-2-propanol.
(1R,1ꢀR)-2ꢀ,2ꢀꢀ-(propane-1,3-diylbis(oxy))di-1,1ꢀ-binaphthyl-2-
ol) 5, obtained with 78% yield and data is in agreement with our
previously described procedure [11].
(1R,1ꢀꢀR)-2ꢀ,2ꢀꢀ-(2,2-dimethylpropane-1,3-diyl)bis(oxy)di-1,1ꢀ-
binaphthyl-2-ol 3 was obtained in 87% yield. 1H NMR (400 MHz,
CDCl3) ı 8.06–7.75 (m, 8H), 7.35 (t, J = 7.3 Hz, 2H), 7.29–7.07
(m, 10H), 6.97 (m, 6H), 4.76 (s, 2H), 3.22 (d, J = 8.2 Hz, 2H), 3.14
(d, J = 8.2 Hz, 2H), 0.32 (s, 6H). 13C NMR (101 MHz, CDCl3) ı
155.01, 151.04, 133.82, 130.70, 129.37, 129.19, 128.93, 128.16,
127.83, 127.11, 126.19, 124.90, 124.71, 123.89, 123.11, 117.33,
115.29, 114.99, 114.39, 35.58, 21.20. m.p. 101–102 ◦C. MS (ESI):
m/z = 663.2482 (M+Na), calcd. for C45H36O4Na+ 663.2506.
(iv) A solution of 5a (1.88 mmol) in CHCl3/MeOH (3:1) in the
presence of Pd/C 5% was submitted to 35 bar of H2 along 72 h at the
temperature of 35 ◦C. The reaction mixture was filtrated in celite,
and the residue washed with methanol. Then, the mixture was
evaporated under reduced pressure and the residue was purified
by silica gel column chromatography using dichloromethane
as eluent. The desired fraction was evaporated yielding
(1R,1ꢀꢀR)-2ꢀ,2ꢀꢀ-(propane-1,3-diylbis(oxy))bis(5,5ꢀ,6,6ꢀ,7,7ꢀ,8,8ꢀ-
octahydro-1,1ꢀ-binaphthyl-2-ol) as a white powder. Yield (80%).
Physical and spectroscopic data were in agreement with the
literature data [11].
2.2. Ligand synthesis
(1R,1ꢀꢀR)-2ꢀ,2ꢀꢀ-(2R,4R)-pentane-2,4-diylbis(oxy)di-1,1ꢀ-
binaphthyl-2-ol
1
(1S,1ꢀꢀS)-2ꢀ,2ꢀꢀ-(2R,4R)-pentane-2,4-
diylbis(oxy)di-1,1ꢀ-binaphthyl 2, were prepared according to
previously described procedure, and all the analytical data are in
good agreement with previously reported [12].
2.2.1. General procedure for the synthesis of ligands
bis-BINOL-2ꢀ,2ꢀꢀ-propyl ether and derivatives 3–5
(i) (R)-2ꢀ-(benzyloxy)-1,1ꢀ-binaphthyl-2-ol (R)-BnBINOL: The
compound was synthesized by slightly modified Mitsunobu reac-
tion [11–13]. To
a stirred solution of (R)-binaphthol (5.0 g,
17.5 mmol), PPh3 (4.59 g, 17.5 mmol) and benzyl alcohol (2.1 mL,
20 mmol) in dry THF (200 mL), a solution of diethyl azodicar-
boxylate (DEAD) (7.7 mL, 40% in toluene, 17.5 mmol) was added
dropwise. The reaction was kept with stirring at room temperature
during 48 h. Then, the mixture was evaporated under reduced pres-
sure. The residue was redissolved in dichloromethane and washed
with water and brine. After partial evaporation of the solvent at
reduced pressure, the residue was purified by preparative chro-
matography (silica, CH2Cl2:n-hexane, 1:1). After evaporation of
solvents, the solid was recrystallized from toluene/n-hexane, to
afford 5.73 g (87% yield) of the product as white solid. Physical and
spectroscopic data were in agreement with the literature data [12].
(ii) To a suspension of sodium hydride (160 mg, 60% in paraffin,
4 mmol) in dry dimethylformamide (DMF) (10 mL, 0 ◦C), a solu-
tion of (R)-2ꢀ-(benzyloxy)-1,1ꢀ-binaphthyl-2-ol (2.7 mmol) in dry
DMF (5 mL) was added dropwise along 30 min. A solution of the
2.3. General procedure for the catalytic reactions
Titanium tetraisopropoxide (120 ␮L, 0.40 mmol) was added via
syringe to the desired bis-BINOL-2ꢀ,2ꢀꢀ-ethers (5.0 × 10−2 mmol) in
the appropriate freshly dried solvent (2 mL), under N2 atmosphere
(15 min). To the resulting solution, diethylzinc (0.75 mL, 1.0 M in
hexane, 0.75 mmol) was added, followed by the addition of arylic
aldehyde (0.25 mmol). The reaction was kept at the appropriate
temperature for 5 h and quenched by adding with 2 M HCl(aq).
The aqueous layer was extracted with ethyl acetate and organic
phase was then evaporated. The crude was redissolved in pen-
tane, producing the precipitation of the ligand, which was removed

A.R. Abreu et al. / Journal of Molecular Catalysis A: Chemical 325 (2010) 91–97
93
Table 1
Asymmetric
catalysed
methyl groups only add steric hindrance to the bridge. Results with
ligands 4 and 5 are reported for comparative purposes.
Ligands 1, 2,4 and 5 were prepared accordingly to previously
described methods [11,12].
ethylation
Ti(OiPr)4
of
benzaldehyde
and
6
with
Et2Zn
by
bis-BINOL-2ꢀ,2ꢀꢀ-ether
ligands.a
Ligand
3 was synthesized using our previously synthetic
strategy [11]. Firstly (R)-BnBINOL was prepared from the reac-
tion of BINOL with benzyl alcohol via slightly modifications
of Mitsunobu reaction. The (R)-BnBINOL was coupled with
2,2-dimethylpropane-1,3-diyl-bis(4-methylbenzenesulfonate)
.
Entry
Ligand
T (◦C)
Conv.b (%)
eeb (%)
1
2
3
4
5
6
7
8
1
2
3
4
5
0
82
87
83
97
99
70
99
75
53 (R)
30 (S)
38 (R)
81 (R)
62 (R)
55 (R)
47 (R)
–
affording
(1R,1ꢀꢀR)-2ꢀ,2ꢀꢀ-(2,2-dimethylpropane-1,3-
3a in 85%
diyl)bis(oxy)bis(2-(benzyloxy)-1,1ꢀ-binaphthyl)
yield, followed by BBr3 benzyl deprotection. In this way
(1ꢀR,1ꢀR)-2ꢀ,2ꢀꢀ-(2,2-dimethylpropane-1,3-diyl)bis(oxi)di-1,1ꢀ-
0
−10
40
0
bibenzobenzen-2-ol
3
was obtained in 87% yield, after
chromatographic purification.
–
a
Reaction conditions: 0.25 mmol of benzaldehyde, 0.05 mmol of ligand,
0.75 mmol of Et2Zn and 0.4 mmol of Ti(OiPr)4 in 2 mL of toluene.
3.2. Catalytic asymmetric ethylation of aldehydes
b
Conversion in 1-phenylpropanol and steroselectivity after 5 h reaction.
Firstly, the performance of the ligands bis-BINOL-2ꢀ,2ꢀꢀ-ethers
1–5, on the titanium(IV) catalytic ethylation of benzaldehyde, was
studied and the results are summarized in Table 1.
Table 2
Asymmetric
des using
addition
Ti(OiPr)4
of
diethylzinc
presence
to
of
aromatic
ligands
aldehy-
and
5.a
Results in Table 1 show that methyl substituents on the propyl
ether bridge did not improve the performance of the ligands, since
both the conversion and the ee, at 0 ◦C, are lower with ligands 1–3
(Table 1, entries 1–4), when compared with ligand 5 (Table 1, entry
5). Furthermore, these results also show that it is possible to predict
the absolute configuration of the final alcohol depending essentially
from the configuration of the BINOL moiety, with a moderated con-
tribution of the stereogenic centers on the propyl ether bridge (53%
ee of (R)-phenylpropanol and 30% ee of (S)-phenylpropanol, Table 1,
entries 1 and 2, respectively).
in
the
4
.
Entry
Subs.
Ligand
Solv.
conv.b (%)
eec (%)
Best conversions (97 and 99%) and enantioselectivities (81 and
62%) were achieved when the reaction was carried out with ligands
4 and 5 (Table 1, entries 4 and 5), which do not have any substituent
in the propyl ether bridge. Furthermore, it should also be noticed
that the enantioselectivity of the ethylation reaction is dependent
from the temperature (−10, 0 and 40 ◦C), showing a maximum at
0 ◦C (Table 1, entries 5–7). These results are indicative that there are
more than one catalytic active species in equilibrium, in solution.
Since the titanium complexes of ligands 4 and 5 have showed
the best results in the catalytic alkylation of benzaldehyde, the cat-
alytic studies were extended to 2- and 3-chloro-benzaldehyde (7
and 8 in Table 2, entries 5–12) using these ligands. In order to
investigate the influence of the solvent polarity in the reaction per-
formance, toluene (PhMe) and dichloromethane (DCM) were used
as solvents. The results are collected in Table 2, and they show
that in all cases the conversion is higher in reactions carried out
in PhMe than those in DCM. However, contrary to the results pre-
viously described [5,6] there is only a small solvent effect on the
enantioselectivity of the reactions, regardless of the ligand or the
aldehyde used. Furthermore, a significant influence of the aldehyde
structure on the ee was observed, being the ee of the alkylation of
2-chloro-benzaldehyde 7 significantly lower than those obtained
with 3-chloro-benzaldehyde or benzaldehyde. From the overall
results, it is possible to conclude that the best results are obtained
with ligand 4 with aldehydes not substituted on the position 2
(Table 2, entries 1–4, 9–12).
1
2
6
6
4
5
PhMe
97
99
81
62
3
4
6
6
4
5
DCM
PhMe
DCM
PhMe
DCM
62
93
80
69
5
6
7
7
4
5
98
98
63
45
7
8
7
7
4
5
68
86
54
46
9
10
8
8
4
5
80
99
79
68
11
12
8
8
4
5
71
85
80
65
a
Reaction conditions: 0.25 mmol of aryldehyde, 0.05 mmol of ligand, 0.75 mmol
of Et2Zn and 0.4 mol of Ti(OiPr)4 in 2 mL of toluene.
b
% of conversion in alcohol after 5 h reaction.
% of enantiomeric excess in (R)-1-arylpropan-1-ol.
c
by filtration. The resulting solution was analyzed by GC, GC/MS
and GC equipped with a chiral column. The data are collected in
Tables 1 and 2.
3. Results and discussion
3.1. Ligand synthesis
In our previous work, we showed that bis-BINOL-2ꢀ,2ꢀꢀ-propyl
ether 5 was one of the most active and selective ligand in asym-
metric titanium catalysed ethylation of benzaldehyde [11]. In an
attempt to improve the catalytic performance of bis-BINOLate-
2ꢀ,2ꢀꢀ-propyl ether titanium(IV) complexes we enlarge these studies
investigating the behavior of ligands 1–3, Scheme 1. Ligands 1 and
2, incorporate methyl substituents on the propyl fragment, lead-
ing to two additional stereogenic centers, while with ligand 3 the
3.3. Analysis of the chelation ability of the ligands via PM6
computational calculations
In order to correlate the relative preference of ligands towards
the formation of a titanium chelate complex with their perfor-
mance as co-catalyst, a theoretical computational study was carried
out. The reaction enthalpies for the formation of two simplified
model complexes, Ti(L)(OiPr)2 and Ti2(L)(OiPr)6 (LH2 = bis-BINOL-

94
A.R. Abreu et al. / Journal of Molecular Catalysis A: Chemical 325 (2010) 91–97
Scheme 1.
Table 3
Calculated reaction enthalpies for reactions 1–3 in Scheme 3.
Entry
Ligand
Reaction 1
Reaction 2
Reaction 3
(kJ mol−1
)
(kJ mol−1
)
(kJ mol−1
)
1
2
3
4
5
1
2
3
4
5
15.0
2.1
−11.2
−9.2
−11.2
−13.4
−5.1
28.4
7.2
7.5
−18.7
−3.0
−6.2
−3.5
−7.7
These results indicate that the enthalpy for reaction
3
(ꢀH3 = ꢀH1 − ꢀH2), that converts one mol of binuclear species
in one mol of chelate complex plus one of Ti(OiPr)4, is strongly
dependent on nature of the ether bridge (Table 3). The enthalpic
contribution favors the ligand chelated species Ti(L)(OiPr)2 versus
the bridged one only in the case of ligands 4 (ꢀH3 = −6.2 kJ mol−1
)
and 5 (ꢀH3 = −7.7 kJ mol−1) (Table 3, entries 4 and 5).
For the rest of ligands studied, enthalpy favors the formation of
binuclear titanium with a bridging ligand.
Ligands 4 and 5 show exothermic processes for the formation of
the chelated complex from Ti(OiPr)4 and LH2 (ꢀH1 = −9.5 kJ mol−1
and ꢀH1 = −11.2 kJ mol−1, respectively), while for ligands 1 and 2
this process is endothermic, revealing steric restrictions in the for-
mation of the chelate probably due to the presence of methyl groups
in 1 and 3 positions of the propy ether bridge. Methyl groups at the
position-2 in ligand 3 seem not to interfere with the formation of
the chelate, since ꢀH1 for this ligand is nearly identical to that of
ligand 5 (Table 3, entries 3 and 5). However, the formation of the
chelate ligand 3-titanium complex is not favored with respect to
the ligand bridged species, since a very negative value of the ꢀH2
(−18.7 kJ mol−1) was calculated. In summary, this theoretical cal-
culations supports that the higher experimental enantioselectivity
achieved with ligands 4 and 5, correlates with the calculated higher
chelation tendency of these two ligands.
Scheme 2.
2ꢀ,2ꢀꢀ-ether ligands 1–5), containing a chelating or a bridging
ligand, respectively, were calculated and compared (Scheme 2).
The conformational flexibility of these ligands made the choice
of semiempirical quantum chemistry methods particularly ade-
quate for this task. The recently developed PM6 hamiltonian
[14], available in the MOPAC2009 software [15], which included
parameterization for titanium, was chosen. This software has been
successfully used in the simulation of a great variety of systems
[16].
Furthermore, the coordination of the ether fragments to the tita-
nium was also tested, but its unfeasibility was observed due to
steric constrains.
The main goal of this study was to evaluate the relative energy
of each of the two alternative (open or chelate) complex products.
An energetically favorable chelate for a given ligand is expected to
show a greater enantioselectivity.
The first step of the study was to explore the conforma-
tional space of the ligands in order to identify the relevant most
stable conformer or conformers with a significant population
and determine their heats of formation. Then, the same proce-
dure was carried out for both the bridging Ti2(L)(OiPr)6 and the
chelate Ti(L)(OiPr)2 species. The heats of formation of the con-
formers selected in the way described above were used in suitable
isodesmic reactions that account for the energetic differences in the
formation of the alternative complex products. Calculated reaction
enthalpies for reactions 1–3 are collected in Table 3.
3.4. The nature of catalytic species
In order to gain some insight into the nature of the catalytic
species, a number of techniques were tested, using the ligand 5,
as model. Firstly, the effect of the optical purity of the ligand on
the stereoselectivity of the asymmetric alkylation of benzaldehyde
was analyzed, using the simplified mathematical model developed
by Kagan and co-workers [17] for catalytic species containing two
chiral ligands, such as ML2 or (ML)2 complexes. In this case, a steady
state involving three possible complexes, MLRLR, MLSLS, and MLRLS

A.R. Abreu et al. / Journal of Molecular Catalysis A: Chemical 325 (2010) 91–97
95
Fig. 1. Graphic representation of the ee of 1-phenylpropanol against ee of ligand 5.
Fig. 3. 1H NMR titration of 5 with Ti(OiPr)4. Solvent CDCl3; T = 25◦C.
In an attempt to obtain complementary information about the
nature of the catalytic active species, a 1H NMR titration of the lig-
and with Ti(OiPr)4 was undertaken. Thus, to a CDCl3 solution of
ligand 5 (90 mM), successive amounts of Ti(OiPr)4 (0.5–4 equiv.)
were added and the 1H NMR spectra of the resulting solutions were
recorded. Selected spectra of the aliphatic region (ı = 3.1–5.1 ppm)
are presented in Fig. 3. The 1H NMR spectra of 5 shows a multi-
plet signal centered at 3.55 ppm assigned to the O–CH2CH2CH2–O
protons and a broad signal at 4.80 ppm assigned to the hydroxyl
groups of the ligand, in addition to a multiplet at 1.53 from
O–CH2CH2CH2–O protons. After addition of 0.5 equivalents of
Ti(OiPr)4, an evident decrease in the intensity of the signals at 3.55
and 4.80 ppm was observed, together with a downfield shift of both
signals. At the same time, three new signals rose at 3.92, 3.80 and
3.65 ppm. The first corresponds to free isopropanol, while the other
two are assigned to the diasteorotopic O–CH2 protons of the ligand
coordinated to the metal. The signal at 1.53 ppm is poorly sensi-
tive to the coordination of the ligand. After further addition of 0.5
equivalents of Ti(OiPr)4 (Ti/L = 1), the signals from free ligand 5 dis-
appeared, while a new broad signal rises at 3.15 ppm, which was
attributed to the OCH(CH3)2 isopropoxide coordinated to the metal.
The signals at 3.92, 3.80 and 3.65 ppm showed the same relative
integration, and were assigned as two equivalents of OCH(CH3)2
from free isopropoxide, and two pairs of diastereotopic protons
from the O–CH2-ether bridge Ti–bis-BINOLate complex.
Fig. 2. Enantiomeric excess of (R)-1-phenylpropanol plotted against the Ti(OiPr)4/5
molar ratio on the catalytic alkylation of benzaldehyde in toluene. Reaction condi-
tions are the same of Table 1. T = 0 ◦C.
is assumed, and the relation between the optical purity of the ligand
eeL, and that of the product of the reaction ee is shown in Eq. (1). In
this equation, eeo and g represent the enantiomeric excess achieved
with the homochiral ligand, and the relative rate of the hetero- over
the homochiral catalysts, respectively:
2 · eeo · eeL
ee =
(1)
1 + g + (1 − g) · ee2L
Fig. 1 shows the ee values of (R)-1-phenylpropanol obtained
from enantiomerically enriched mixtures of ligand 5 plotted
against the enantiomeric excess on the (R,R) isomer of this ligand
(eeL). The experimental values were fitted with a value of g = 4 in
Eq. (1).
After addition of another equivalent of Ti(OiPr)4 (Ti/L = 2), a
broadening of the overall spectra was observed, as well as a
decrease in the 3.15 ppm signal, probably due to the fast exchange
between the isopropoxide groups coordinated to the metal in the
Ti–bis-BINOLate complex and those of Ti(OiPr)4. Further addition
of Ti(OiPr)4 did not significantly change the spectra, except for
the expected increase in the isopropoxide signal of Ti(OiPr)4 at
4.40 ppm. The most relevant result from the NMR study is the
significant split of the signals of the diasteorotopic OCH2 protons
of the ligand that shows two clearly separated multiplets (ꢀı ca.
0.15 ppm) when it is coordinated to the titanium. This is consis-
tent with the formation of a complex with a chelated ligand that
somewhat freezes the ligand conformation. Therefore, both 1H
NMR analysis and PM6 calculations point out that ligand 5 form
a chelated titanium complex. Unfortunately, the fast isopropoxide
ligands exchange in the NMR time scale did not allow the analysis
of complexes formed at high Ti/ligand molar ratios.
It is well established that a catalytic system with g > 1 points
to a negative non-linear effect (−)-NLE, due to a greater reactiv-
ity of the meso (MLRLS) catalytic species versus the homochiral
one [18]. Furthermore, this NLE indicates that the active catalytic
species contains two ligands, in contrast with what was previously
reported for the BINOL/Ti(OiPr)4 catalyst that contains two tita-
nium ions bridged by a single BINOL ligand [9]. The activity of the
background reaction (see last entry in Table 1) does not have any
influence on the overall conclusion, since the amount of racemic
phenylpropanol obtained is constant in all the experiments.
The effect of the Ti/L molar ratio on the ee of the product of the
ethylation of benzaldehyde was also investigated and the results
are represented in Fig. 2. These results show that at [L] = 25 mM,
the ee of the product rises linearly with the Ti/L molar ratio, until
ca. 8, for which the maximum stereoselectivity was reached. After
this point, a smooth drop in the ee of the product was observed.
This result reveals that a relatively large molar excess of Ti(OiPr)4,
with regard to the ligand, is required to form a stereoselective
catalyst. The requirement of an excess of titanium with 2ꢀ,2ꢀꢀ-bis-
BINOL-propyl ether ligand, to obtain higher ee, is in agreement with
our recent results [11] and also the observations of others authors,
using 3ꢀ3ꢀꢀ-bis-BINOL [19] and also BINOL-type ligands [9].
To obtain conformational information, a titration of ligand 5
with Ti(OiPr)4 was carried out, but in this case circular dichroism
spectra (CD) of the resulting solutions were analyzed. It is well
established that CD of 1,1ꢀ-binaphthyl derivatives are extremely
sensitive to the conformation that they adopt in solution, and
in particular to the value of the dihedral angle ꢁ between the

96
A.R. Abreu et al. / Journal of Molecular Catalysis A: Chemical 325 (2010) 91–97
Fig. 4. CD spectra of the solutions of 5 and Ti(OiPr)4 at different molar ratios.
two naphthalene planes [6,10]. The evolution of CD spectra of a
dichloromethane solution of ligand 5 (9.0 × 10−2 mM) titrated with
Ti(OiPr)4 is shown in Fig. 4.
The CD spectrum of ligand 5 shows the characteristic intense
positive couplet, due to the excitation coupling between long-
axes polarized 1Bb transitions with a positive maximum at 325 nm
(Fig. 4, spectrum 1). After addition of 2 equiv. of Ti(OiPr)4, almost no
change in the shape and intensity of the band was observed (Fig. 4,
spectrum 2). After the subsequent additions of Ti(OiPr)4 an increase
in the intensity of the 325 nm band is observed, together with the
appearance of a new positive band at 350 nm (Fig. 4, spectrum 3).
The intensity of both bands steadily increases with the titanium
concentration of the solutions until a Ti/ligand molar ratio ca. 6–8
(Fig. 4, spectra 4 and 5). At higher concentrations of Ti(OiPr)4, a
broadening of the positive bands concomitantly with a decrease in
the intensity is observed (Fig. 4, spectrum 6). These results points
out to the existence of an equilibrium between species and one of
them should be the predominant at the same Ti/ligand molar ratio
(6–8) for which the best ee was achieved in the catalytic alkylation
of benzaldehyde. Thus, this stereoselective catalytic species must
be responsible for the band at 350 nm, as well as for the increase in
the intensity of the band at 325 nm.
Further information about the nature of the species formed at
different molar ratios of the ligand 5 and Ti(OiPr)4 was attained
analyzing solutions of Ti(OiPr)4 and ligand 5 by ESI-MS. At the
molar ratio Ti(OiPr)4/5 = 0.5, one major species was observed, with
a mass molecular cluster at 1267–1274 amu, and a 100% peak at
1269.2. This signal matches with the species [TiL2H]+ (II, Scheme 3),
where L represents the deprotonated dianionic form of ligand 5.
At Ti(OiPr)4/5 molar ratio equal to 1, besides the previous signal,
two other clusters were observed. One at 710–720 amu with a
100% peak at 717.1, that matches with [TiL(OiPr)]+, which likely
must arise from the species [(TiL(OiPr)2)] (IIIa, Scheme 3). The
other at 773.8–780.8 amu with a 100% peak at 777.8, is a defini-
tive evidence for the formation of a species containing two ligands
[Ti2L2(OiPr)4H2]2+, as required to explain the NLE described above
(IIIb, Scheme 3)
Scheme 3.
iments at [Ti]/[5] molar ratios higher than 1 show evidence that
polymerization occurs, preventing the full identification of species
IV.
4. Conclusion
These studies allow us to conclude that titanium (IV)-bis-
BINOLate-2ꢀ,2ꢀꢀ-propylether complexes (4 and 5) yield the most
stable Ti-chelates, leading to the best results in the ethylation of aryl
aldehydes (ca. 80, 69%, respectively). The experimental results were
corroborated by PM6 computacional calculations and 1H NMR,
which showed that ligands 4 and 5 have a higher chelating ten-
dency.
A number of other techniques were used to investigate the
nature of the catalytic species, formed in the reaction of ligand 5
with Ti(OiPr)4. The presence of two ligands in the catalytic active
and selective species was corroborated by the observation of a
negative NLE and by ESI-MS. Furthermore, experimental catalytic
results as well as CD experiments showed that the most enantios-
elective species is formed only at Ti/ligand molar ratios close to 8.
So, from the overall results a trinuclear complex [Ti3L2(OiPr)8] (IV
in Scheme 3) is proposed to be the most enantioselective catalytic
active species.
This cluster corresponds to a doubly charged species, as evi-
denced by the 0.5 amu increments between isotopic peaks.
Since the species IIIb contains two chiral ligands, it could pro-
duce a NLE. However, from the catalytic results ploted in Fig. 2,
the highest enantioselective species is only formed when the molar
ratio is 8. Therefore, we propose that the catalytic species producing
the highest ee’s is species IV (Scheme 3). This species can be formed
by reaction of the binuclear species IIIb in the presence of an excess
of Ti(OiPr)4. We assume that at low [Ti]/[5] molar ratio the equilib-
rium is shifted toward the bimetallic species IIIb. Finally, it should
be pointed out that a large excess of Ti(OiPr)4 is required to obtain
an high enantioselective species IV as the dominant one, like was
previously observed for BINOL studies [5,6]. Further ESI-MS exper-
The mechanistic overall results, especially the infor-
mation obtained from computational calculations related
to the chelating tendency of these type of ligands opens
the way for further studies in order to design the ideal
ligand structure to promote enantioselective catalytic reac-
tions
Acknowledgements
The authors thank financial support from Portuguese FCT
through project PTDC/QUI/66015/2006, Merquinsa (Barcelona,
Spain), and the Spanish DGI through project BIO2006-14420-

A.R. Abreu et al. / Journal of Molecular Catalysis A: Chemical 325 (2010) 91–97
97
CO2-02. ARA and DC thanks for a PhD grant to FCT (PhD grant
SFRH/BD/21314/2005) and UAB, respectively.
[10] (a) F.-Y. Zhang, C.-C. Pai, A.S.C. Chan, J. Am. Chem. Soc. 120 (1998) 5808–5809;
(b) T. Uemura, X. Zhang, K. Matsumura, N. Sayo, H. Kuobayashi, T. Otha, K.
Nozaki, H. Takaya, J. Org. Chem. 61 (1996) 5510–5516.
[11] A.R. Abreu, M.M. Pereira, J.C. Bayón, Tetrahedron 66 (2010) 743–
749.
References
[12] M. Takahashi, K. Ogasawara, Tetrahedron: Asymmetry
3130.
8 (1997) 3125–
[1] K. Soai, S. Niwa, Chem. Rev. 92 (1992) 833–856.
[13] (a) O. Mitsunobu, Synthesis (1981) 1–18;
[2] K.-H. Wu, H.-M. Gau, Organometallics 22 (2003) 5193–5200;
(b) K.-H. Wu, H.-M. Gau, Organometallics 23 (2004) 580–588;
(c) S.-H. Hsieh, H.-M. Gau, Chirality 18 (2006) 569–574.
[3] P.J. Walsh, Acc. Chem. Res. 36 (2003) 739–749.
[4] (a) T.J. Boyle, N.W. Eilerts, J.A. Heppert, F. Takuasagawa, Organometallics 13
(1994) 2218–2229;
(b) K. Mikami, M. Ueki, Y. Matsumoto, M. Terada, Chirality 13 (2001) 541–544.
[5] (a) J. Balsells, T.J. Davis, P. Carroll, P.J. Walsh, J. Am. Chem. Soc. 124 (2002)
10336–10348;
(b) K.M. Waltz, P.J. Carroll, P.J. Walsh, Organometallics 23 (2004) 127–134.
[6] G. Pescitelli, L. Di Bari, P. Salvadori, Organometallics 23 (2004) 4223–4229.
[7] D.J. Cram, R.C. Helgeson, S.C. Peacock, L.J. Kapan, L.A. Domeier, P. Moreau, K.
Koga, J.M. Mayer, Y. Chao, M.G. Siegel, D.H. Hoffman, G.D.Y. Sogah, J. Org. Chem.
43 (1978) 1930–1946.
(b) R.M.B. Carrilho, A.R. Abreu, G. Petıcz, J.C. Bayón, M.J.S.M. Moreno, L. Kollár,
M.M. Pereira, Chem. Lett. 8 (2009) 844–845.
[14] J.J.P. Stewart, J. Mol. Mod. 13 (2007) 1173–1213.
[15] MOPAC2009, J.J.P. Stewart, Stewart Computational Chemistry, Colorado
Springs, CO, USA, http://openmopac.net, 2009.
[16] (a) A. Alparone, V. Librando, Z. Minniti, Chem. Phys. Lett. 460 (2008) 151–
154;
(b) R.A. Kwiecien, M. Rostkowski, A. Dybala-Defratyka, P. Paneth, J. Inorg.
Biochem. 98 (2004) 1078–1086;
(c) A. Hermann, R.P. Krawczyk, M. Lein, P. Schwerdtfeger, Phys. Rev. A 76 (2007)
013202.
[17] D. Guillaneaux, S.-H. Zhao, O. Samuel, D. Rainford, H.B. Kagan, J. Am. Chem. Soc.
116 (1994) 9430–9439.
[18] (a) C. Girard, H.B. Kagan, Angew. Chem. Int. Ed. 37 (1998) 2922–2959;
(b) T. Satyanarayana, S. Abraham, H.B. Kagan, Angew. Chem. Int. Ed. 48 (2009)
456–494.
[8] X. Shen, H. Guo, K. Ding, Tetrahedron: Asymmetry 11 (2000) 4321–4327.
[9] (a) F.-Y. Zang, C.-W. Yip, R. Cao, A.S.C. Chan, Tetrahedron: Asymmetry 8 (1997)
585–589;
[19] T. Harada, K. Kanda, Y. Hiraoka, Y. Marutani, M. Nakatsugawa, Tetrahedron:
Asymmetry 15 (2004) 3879–3883.
(b) F.-Y. Zhang, A.S.C. Chan, Tetrahedron: Asymmetry 8 (1997) 3651–3655;
(c) M.T. Reetz, C. Merk, G. Naverfeld, J. Rudolph, N. Gibenow, R. Goddard, Tetra-
hedron Lett. 38 (1997) 5273–5276;
(d) A.S.C. Chan, F.-Y. Zhang, C.-W. Yip, J. Am. Chem. Soc. 119 (1997) 4080–4081.