Preparation of polystyrene beads with dendritically embedded TADDOL and use in enantioselective Lewis acid catalysis

Article abstract of DOI:10.1002/1522-2675(200201)85:1<352::AID-HLCA352>3.0.CO;2-D

A full account is given of the preparation and use of TADDOLates, which are dendritically incorporated in polystyrene beads (Scheme 1). A series of styryl-substituted TADDOLs with flexible, rigid, or dendritically branching spacers between the TADDOL core and the styryl groups (2-16 in number) has been prepared (5-7, 20, 21, 26 in Schemes 2-4 and Fig. 1-3). These were used as cross-linkers in styrene-suspension polymerization, leading to beads of ca, 400-μm diameter (Schemes 5 and 6, b). These, in turn, were loaded with titanate and used for the Lewis acid catalyzed addition of Et₂Zn to PhCHO as a test reaction (Scheme 6). A comparison of the enantioselectivities and degrees of conversion (both up to 99%), obtained under standard conditions, shows that these polymer-incorporated Ti-TADDOLates are highly efficient catalysts for this process (Table 1). In view of the effort necessary to prepare the novel, immobilized catalysts, emphasis was laid upon their multiple use. The performance over 20 cycles of the test reaction was best with the polymer obtained from the TADDOL bearing four first-generation Frechet branches with eight peripheral styryl groups (6, p-6, p-6·Ti(OⁱPr)₂): the enantioselectivity (Fig. 4), the rate of reaction (Fig. 5), and the swelling factor (Fig. 6) were essentially unchanged after numerous operations carried out with the corresponding beads of 400-μm diameter and a degree of loading of 0.1 mmol TADDOLate/g polymer, with or without stirring (Fig. 7). The rate with the dendritically polymer-embedded Ti-TADDOLate (p-6·Ti(OⁱPr)₂) was greater than that measured with the corresponding monomer, i.e., 6·Ti(OⁱPr)₂ (Fig. 8). Possible interpretations of this phenomenon are proposed. A polymer-bound TADDOL, generated on a solid support (by Grignard addition to an immobilized tartrate ester ketal) did not perform well (Scheme 4 and Table 2). Also, when we prepared polystyrene beads by copolymerization of styrene, a zero-, first-, or second-generation dendritic cross-linker, and a mono-styryl-substituted TADDOL derivative, the performance in the test reaction did not rival that of the dendritically incorporated Ti-TADDOLate ((p-6·Ti(OⁱPr)₂) (Scheme 7 and Fig. 10). Finally, we have applied the dendritically immobilized Cl₂ and (TsO)₂Ti-TADDOLate as chiral Lewis acid to preferentially prepare one enantiomer of the exo and the endo (3 + 2) cycloadduct, respectively, of diphenyl nitrone to 3-crotonoyl-1,3-oxazolidinone; in one of these reaction modes, we have observed an interesting conditioning of the catalyst: with an increasing number of application cycles, the amount of polymer-incorporated Lewis acid required to induce the same degree of enantioselectivity, decreased; the degrees of diastereo- and enantioselectivity were, again, comparable to those reported for homogeneous conditions (Fig. 9).

Full text of DOI:10.1002/1522-2675(200201)85:1<352::AID-HLCA352>3.0.CO;2-D

352
Helvetica Chimica Acta Vol. 85 (2002)
Preparation of Polystyrene Beads with Dendritically Embedded TADDOL
and Use in Enantioselective Lewis Acid Catalysis1)
by Holger Sellner2), P. Beat Rheiner2), and Dieter Seebach*
Laboratorium f¸r Organische Chemie der Eidgenˆssischen Technischen Hochschule, ETH-Hˆnggerberg,
CH-8093 Z¸rich
Dedicated to Professor Barry M. Trost on the occasion of his 60th birthday
A full account is given of the preparation and use of TADDOLates, which are dendritically incorporated in
polystyrene beads (Scheme 1). A series of styryl-substituted TADDOLs with flexible, rigid, or dendritically
branching spacers between the TADDOL core and the styryl groups (2 16 in number) has been prepared (5 7,
20, 21, 26 in Schemes 2 4 and Fig. 1 3). These were used as cross-linkers in styrene-suspension polymerization,
leading to beads of ca. 400-mm diameter (Schemes 5 and 6,b). These, in turn, were loaded with titanate and used
for the Lewis acid catalyzed addition of Et2Zn to PhCHO as a test reaction (Scheme 6). A comparison of the
enantioselectivities and degrees of conversion (both up to 99%), obtained under standard conditions, shows that
these polymer-incorporated Ti-TADDOLates are highly efficient catalysts for this process (Table 1). In view of
the effort necessary to prepare the novel, immobilized catalysts, emphasis was laid upon their multiple use. The
performance over 20 cycles of the test reaction was best with the polymer obtained from the TADDOL bearing
i
¬
four first-generation Frechet branches with eight peripheral styryl groups (6, p-6, p-6 ¥ Ti(O Pr)2): the
enantioselectivity (Fig. 4), the rate of reaction (Fig. 5), and the swelling factor (Fig. 6) were essentially
unchanged after numerous operations carried out with the corresponding beads of 400-mm diameter and a
degree of loading of 0.1 mmol TADDOLate/g polymer, with or without stirring (Fig. 7). The rate with the
dendritically polymer-embedded Ti-TADDOLate (p-6 ¥ Ti(OiPr)2) was greater than that measured with the
corresponding monomer, i.e., 6 ¥ Ti(OiPr)2 (Fig. 8). Possible interpretations of this phenomenon are proposed. A
polymer-bound TADDOL, generated on a solid support (by Grignard addition to an immobilized tartrate ester
ketal) did not perform well (Scheme 4 and Table 2). Also, when we prepared polystyrene beads by
copolymerization of styrene, a zero-, first-, or second-generation dendritic cross-linker, and a mono-styryl-
substituted TADDOL derivative, the performance in the test reaction did not rival that of the dendritically
incorporated Ti-TADDOLate ((p-6 ¥ Ti(OiPr)2) (Scheme 7 and Fig. 10). Finally, we have applied the dendriti-
cally immobilized Cl2 and (TsO)2Ti-TADDOLate as chiral Lewis acid to preferentially prepare one enantiomer
of the exo and the endo (3 ‡ 2) cycloadduct, respectively, of diphenyl nitrone to 3-crotonoyl-1,3-oxazolidinone;
in one of these reaction modes, we have observed an interesting conditioning of the catalyst: with an increasing
number of application cycles, the amount of polymer-incorporated Lewis acid required to induce the same
degree of enantioselectivity, decreased; the degrees of diastereo- and enantioselectivity were, again, comparable
to those reported for homogeneous conditions (Fig. 9).
1. Introduction. Binding chiral ligands to insoluble polymer supports is a field of
growing interest in current chemistry. Heterogeneous catalysis offers the advantages of
easy separation of the supported catalyst from the reaction mixture, possible recycling
as well as, in many cases, enhanced stability of the polymer-bound catalyst as compared
1
)
Preliminary communications: [1][2]; parts of the results have been mentioned in review articles, without
experimental detail [3 5].
Part of the Ph.D. theses of H.S., ETH Z¸rich, Dissertation No. 14145, 2001, and of P.B.R., ETH Z¸rich,
Dissertation No. 12773, 1998.
2
)

Helvetica Chimica Acta Vol. 85 (2002)
353
to its soluble analogue [6 14]. In this way, even sophisticated and expensive chiral
ligand systems might become attractive for industrial applications. However, more
often than not, the activity of supported catalysts is reduced with respect to their
soluble analogues used under homogeneous conditions, due to diffusion problems or to
the fact that the preferred conformation of the catalytic moiety cannot be adopted on
or in the polymer. Therefore, new methodologies for immobilization have to be found
that provide a catalytic activity of the polymer-bound catalyst similar to that obtained
under homogeneous conditions.
By far, the most frequently used strategy of heterogenization is covalent binding of
the catalyst or ligand of choice to a polymer support. In the case of organic polymer
supports, this can be achieved either by grafting the desired ligand onto a preformed
support containing reactive groups, or by copolymerization of a suitably functionalized
ligand with polymerizable monomers and a cross-linker (for the preparation of
insoluble polymers). The first approach is often preferred, since many suitable polymer
supports for all kinds of applications (e.g., Merrifield resins) are commercially
available. Furthermore, the approach of grafting is very convenient, especially for
chemists who are not familiar with polymer or solid-state chemistry, since the
immobilization process is performed by simply linking the ligand to the polymer resin
by means of well-known coupling reactions. In contrast, immobilization by copoly-
merization of the ligand with suitable vinylic monomers and a cross-linker, although
requiring more synthetic effort than immobilization by grafting, offers many more
possibilities for generating and controlling a specific environment around the ligands
within the polymer matrix. For example, polymeric materials with cavities of molecular
dimension around the catalytic centers can be obtained [15] (cf. the method of
−molecular imprinting×, developed by Wulff [16][17]). In most applications, polystyrene
is used as polymer support, i.e., a vinyl-substituted ligand is copolymerized with styrene
and divinyl benzene (DVB) as cross-linker (Scheme 1,a). As can be seen from this
schematic representation, in this approach a ligand is directly incorporated into the
cross-linked polystyrene resin, and this may result in reduced catalytic activity due to
steric congestion around the catalyst moieties.
To circumvent these problems, we have introduced a novel concept: the chiral
ligand to be immobilized is placed in the core of a polymerizable dendrimer, followed
by copolymerization of the latter with styrene (Scheme 1,b). In this approach, no
further cross-linking agent, such as DVB, is necessary, since the dendrimer itself acts as
cross-linker. The dendritic branches are thought to act as spacer units keeping the
−obstructing× polystyrene backbone away from the catalytic centers, leading to better
accessibility and, thus, to enhanced catalytic activity. Our work represents the first
example of dendrimers used as polymer cross-linkers. Quite recently, Reetz and Giebel
introduced an alternative concept for the preparation of dendritically cross-linked
materials: triflate-functionalized poly(propylene imine) dendrimers were intercon-
nected by the addition of Sc(OTf)3 to give a dendrimer cross-linked material, which was
used for the heterogeneous catalysis of various Sc Lewis acid mediated test reactions
[18].
Both approaches presented in Scheme 1 have been realized with the chiral
TADDOL ligands (TADDOL ˆ a,a,a',a'-tetraaryl-1,3-dioxolane-4,5-dimethanol) de-
veloped in our group [4][19]. Mono-styryl-substituted TADDOL ligands were

354
Helvetica Chimica Acta Vol. 85 (2002)
Scheme 1. a) Immobilization of Mono-Vinyl-Functionalized Chiral Ligands by Copolymerization with Styrene
and DVB. b) Novel Approach: Copolymerization of Dendritically Surrounded Chiral Ligands with Styrene.
a)
b)
Chiral
Ligand
*
Chiral
Ligand
*
+
+
+
Suspension
Suspension
Copolymerization
Copolymerization
Ph
Ph
Ph
Ph
Ph
Ph
Ph
Ph
Ph
Ph
Ph
Ph
Ph
Ph
Chiral
Ligand
*
Ph
Ph
Ph
Ph
Ph
Ph
Chiral
Ligand
*
Ph
Ph
Ph
Ph
Ph
Ph
Ph
Ph
Ph
Ph
Ph
Ph
Ph
Ph
Ph
Ph
Ph
Ph
Chiral
Ligand
*
Chiral
Ligand
*
Ph
Ph
Ph
Ph
Ph
Ph
Ph
Ph
Ph
Ph
Ph
Ph
Ph
Ph
Ph
Ph
Ph
Ph
Ph
Ph
Ph
less steric congestion around
the catalytic center
copolymerized with styrene and DVB as cross-linker to give, after loading with metal
salts, polymer-bound TADDOLate complexes of high catalytic activity [20]3). In 1997,
we reported for the first time on the copolymerization with styrene of styryl-
functionalized TADDOL dendrimers according to Scheme 1,b [1]. It turned out that
3
)
Besides our own work, TADDOL has also been immobilized by other groups [21].

Helvetica Chimica Acta Vol. 85 (2002)
355
the derived polymeric TADDOLates exhibit excellent catalytic activity with respect to
reproducibility of enantioselectivity and of degree of conversion during multiple use as
catalyst ligands in a test reaction [2]. Motivated by this success, we have also
immobilized the chiral ligands BINOL [22a e]4) and Salen [23] in this way to confirm
the potential of this new technique of immobilization, in a variety of different test
reactions.
In the present paper, we describe and summarize the results obtained with
TADDOL complexes immobilized in a cross-linking fashion in polystyrene. Applica-
tions of these new polymer-bound catalysts in various test reactions are presented, and
a comparison of the catalytic properties of this new class of supported TADDOLates
with those of the already existing systems is provided.
2. Preparation of Styryl-Substituted TADDOLs for Cross-Linking Copolymeriza-
tion with Styrene.
We decided to prepare TADDOL cross-linkers for the
copolymerization with styrene, in which the spacer units are attached to the para-
positions of the TADDOL Ph groups. We knew that modification of the TADDOL
moiety (with dendritic branches of up to the fourth generation) at these positions has
hardly any influence on the catalytic activity of the corresponding metal complexes in
¬
solution [24]. A linear benzyl-ether branch 2 [22] and dendritic Frechet branches [25] 3
and 4 of first and second generation [22] were coupled to the TADDOL core unit 1
[1][24] by etherification (Fig. 1). The TADDOL cross-linkers 5, 6, and 7 were obtained
in yields between 65 and 70% after purification by column chromatography (Fig. 2). As
a result of laborious optimization, we found that heating a suspension of hexol 1
(1 equiv.) and the corresponding benzyl bromide (exactly 4 equiv.) in the presence of
K2CO3 (4 equiv.) in acetone at 658 were the best conditions to reduce the amount of C1-
symmetrical by-products formed by excess coupling with a fifth branch to one of the
TADDOL OH groups [1][24]. Generally, special precautions had to be taken in order
to avoid spontaneous polymerization of the vinyl-substituted derivatives: crude
products were immediately purified by chromatography or recrystallization, and
stored in the refrigerator; only sufficiently dilute solution were heated.
In many cases in the literature, ligands to be immobilized are attached to the
polymer support via alkyl linkers [21c][26]. A flexible and long spacer moiety reduces
interactions between polymer backbone and catalytic site within the polymer, and
facilitates the adoption of an optimal conformation of the catalytically active
substructure. Furthermore, Itsuno et al. introduced the strategy of using flexible
poly(ethylene glycol)-derived cross-linkers for the preparation of polystyrene resins
[27][28]. The resulting materials exhibited much better swelling properties and
mechanical stabilities than polystyrene resins cross-linked with a comparable amount
of DVB [29][30]. Therefore, we have also prepared a TADDOL cross-linker with
flexible alkyl spacer units instead of the rather rigid benzyl ether moieties of
TADDOLs 5 7. The preparation of the corresponding benzyl bromide branch 14 is
presented in Scheme 2.
In addition to the flexible spacer moiety 14, a rather rigid elongation unit 19 of
comparable length (containing a 1,1'-biphenyl unit) was prepared (Scheme 3): starting
4
)
TADDOL has recently been successfully immobilized also on silica by grafting [21f].

356
Helvetica Chimica Acta Vol. 85 (2002)
HO
OH
O
O
OH
OH
O
O
O
Br
Br
OH
HO
1
2
3
O
O
O
O
O
O
Br
4
Fig. 1. TADDOL Core unit 1 [1][24] to which the spacer unit 2 [22] as well as first-generation dendritic branch 3
[22] and second-generation dendritic branch 4 [22] are coupled by etherification (K2CO3, acetone, 658)
from the 1,1'-biphenyl building block 15 [31], the bromide 19 was obtained according to
standard procedures, via the intermediates 16 18 (Scheme 3)5)6).
Coupling TADDOL core unit 1 with spacer units 14 and 19 under the usual
conditions afforded TADDOL cross-linkers 20 and 21 (Fig. 3).
The TADDOL cross-linkers described so far carry the spacer units on the aryl
moieties, while, in all previous approaches directed towards immobilization of
TADDOL on polystyrene or silica, the TADDOL moieties were attached to the
support via the acetal center [20][21]. Therefore, we decided to prepare also a
TADDOL cross-linker bearing two styryl groups at the acetal center (Scheme 4).
Starting from 4-vinylbenzaldehyde [20a], the ketal 24 of a benzophenone was obtained
5
)
)
During the coupling of 4-vinylbenzyl chloride with alcohol 16 in the presence of NaH as base, migration of
the silyl protecting groups occurred, so that the yield of the desired product was rather low.
Br/OH Exchange 18 ! 19 under Appel conditions (CBr4/PPh3) failed due to formation of an inseparable
product mixture. However, with PBr3 in Et2O, the desired bromide 19 was obtained in pure form without
formation of by-products, even though 18 was insoluble in Et2O.
6

Helvetica Chimica Acta Vol. 85 (2002)
357
O
O
O
O
O
O
O
O
O
O
O
O
O
O
OH
OH
OH
OH
O
O
O
O
O
O
O
O
O
O
5
6
O
O
O
O
O
O
O
O
O
O
O
O
O
O
O
O
OH
OH
O
O
O
O
O
O
O
O
O
O
O
O
O
O
7
Fig. 2. TADDOL Cross-linker 5, and first- and second-generation dendritic TADDOL cross-linkers 6 and 7

358
Helvetica Chimica Acta Vol. 85 (2002)
Scheme 2. Preparation of the Spacer Unit 14 Starting from the Mono-silylated Butane-1,4-diol 8 via the
Intermediates 9 13
Cl
OTBDMS
O
HO
RO
NaH, THF
68 %
8
9
R = TBDMS
TBAF, THF
97 %
10 R = H
OTBDPS
10, NaH
O
O
THF
X
58 %
Br
12 X = OTBDPS
13 X = OH
11
TBAF, THF
77 %
TBAF = Bu4NF
TBDMS = (t-Bu)Me2Si
TBDPS = (t-Bu)Ph2Si
CBr4, PPh3
THF, 90 %
14 X = Br
Scheme 3. Preparation of the 1,1'-Biphenyl Spacer Unit 19 Starting from Aldehyde 15 [31] via the Intermediates
16 18
OH
O
H
O
NaBH4
Cl
THF
quant.
NaH, THF
59 %
TBDPSO
TBDPSO
X
16
15
17 X = OTBDPS
18 X = OH
TBAF, THF
89 %
PBr3, Et2O
65 %
TBAF = Bu4NF
TBDPS = (t-Bu)Ph2Si
19 X = Br

Helvetica Chimica Acta Vol. 85 (2002)
359
in a three-step sequence. Transacetalization of 24 with (R,R)-diethyl tartrate at 08 in
the presence of BF3 ¥ OEt2 afforded diester 25 (without formation of polymeric by-
products), which was converted to the TADDOL derivative 26 by the addition of
PhMgBr.
3. Copolymerization of Cross-Linking TADDOLs with Styrene. TADDOL Cross-
linkers 5 7, 20, 21, and 26 were copolymerized with styrene, according to a procedure
¬
first proposed by Itsuno, Frechet, and co-workers [29]. Thus, a solution of the cross-
linking TADDOL, styrene, and a,a'-azobis(isobutyronitrile) (AIBN) in benzene/THF
was mixed with an aqueous phase containing poly(vinyl alcohol) (PVA) and heated
with constant slow overhead stirring for 48h at 90 8 (Scheme 5,a). During this process,
spherical beads (average diameter of 400 mm) of polystyrene with incorporated
O
O
O
O
O
O
O
O
OH
OH
O
O
O
O
O
O
20
O
O
O
O
O
O
OH
OH
O
O
O
O
21
Fig. 3. TADDOL Cross-linkers 20 and 21, prepared from the TADDOL core unit 1, and the spacer units 14 and
19, respectively

360
Helvetica Chimica Acta Vol. 85 (2002)
Scheme 4. Preparation of TADDOL 26 via the Intermediates 22 25
Na2Cr2O7
H2SO4
MgCl
OH
O
THF
80%
Et2O
72%
O
H
22
23
HO
HO
CO2Et
CO2Et
HC(OMe)3
TsOH
CO2Et
CO2Et
OMe
OMe
O
O
MeOH
95%
BF3 OEt2
AcDEt, 0°
79%
25
24
PhMgBr
O
O
OH
OH
THF
69%
26
TADDOLs were formed. The beads were isolated by filtering, washing, and drying, and
then collected by size with a set of sieves. Usually, the yield of the polymers was
quantitative so that the loading could be directly calculated from the relative amounts
of comonomers used (mmol TADDOL per g polymer).
In addition to copolymerizing preformed TADDOL derivatives with styrene
(Scheme 5,a) we attempted to generate polymer-bound TADDOL by solid-phase
synthesis (Scheme 5,b). For practical reasons, this might have been an attractive
alternative route, through a precursor that could be transformed (on the polymer) to a
variety of different polymer-bound TADDOL ligands. Thus, the diester 25 was
copolymerized with styrene to give polystyrene-bound diester p-25, which was then
treated with two different aromatic Grignard reagents to give the corresponding

Helvetica Chimica Acta Vol. 85 (2002)
361
Scheme 5. Immobilization of TADDOL a) by Copolymerization of Cross-Linking TADDOLs 5 7, 20, 21, and
26 with Styrene, b) by Copolymerization of the Diester 25 with Styrene, Followed by Grignard Addition, and c) by
Copolymerization of TADDOL 28 with Styrene and DVB. Conditions for suspension copolymerization:
benzene, THF/H2O, PVA; AIBN, 908, 48 h.
a)
Cross-linker
5, 6, 7, 20, 21 and 26
TADDOL-styrene
Suspension
copolymers
p-5, p-6, p-7, p-20,
p-21, and p-26
Copolymerization
+
b)
CO2Et
CO2Et
CO2Et
O
O
O
Suspension
PS
Copolymerization
O
CO2Et
p-25
25
PS = polystyrene
ArlMgBr
Arl
Arl
+
O
O
OH
PS
OH
Arl
Arl
p-26' Arl = Ph
c)
Ph
Ph
Ph
p-27 Arl = 2-naphthyl
O
O
H
OH
OH
Ph
TADDOL-styrene-
DVB copolymer
p-28
Suspension
28
Copolymerization
+
+
polymer-bound TADDOLs p-26' (the −prime× is added in order to emphasize that p-26'
was prepared in a way different from that leading to p-26) and p-27 (Scheme 5,b)7). In
line with observations by Mayoral and co-workers and Irrure et al. [21a d], who had
7
)
IR Spectroscopy of a sample of beads of p-26' showed that the signal of the CˆO ester group at n ˆ
1735 cmÀ1 had completely disappeared, indicating formation of the desired TADDOL moieties.

362
Helvetica Chimica Acta Vol. 85 (2002)
tried to generate polymer-bound TADDOLs by a comparable approach, it turned out
to be difficult to remove the magnesium salts after the reaction.
The best method was to first remove excess Grignard reagent by extensive washing
of the beads with THF, followed by stirring in a mixture of THF/1n HCl.
To compare the results achieved with polymers prepared with cross-linking
TADDOLs (following the approaches outlined in Scheme 5,a and b) with existing
polymer-bound versions of this ligand, polymer p-28 was prepared by copolymerization
with styrene and DVB of the mono-vinyl derivative 28 of TADDOL, according to a
procedure developed previously in our group (Scheme 5,c) [20].
4. Application of Polymer-Bound Ti-TADDOLate Complexes in Enantioselective
Additions of Et2Zn and Bu2Zn to PhCHO.
The addition of Et2Zn to PhCHO
catalyzed by (iPrO)2Ti-TADDOLate complexes offers several advantages, making it an
attractive test reaction for polymer-bound TADDOLs. Besides high enantioselectiv-
ities, with which 1-phenylpropan-1-ol is formed (up to 99%), and high conversions that
are generally observed under homogeneous conditions [32], the analysis is simple and
can be performed by capillary gas chromatography (CGC). Therefore, this reaction
allows for a fast interpretation of catalytic properties. Very importantly, the
Ti-TADDOLate-mediated addition of Et2Zn to PhCHO shows a linear correla-
tion between the enantiomer purity of the TADDOL employed and that of the alcohol
formed [32], indicating that the catalytically active species is monomeric, in contrast
to reactions with a so-called −nonlinear effect× (NLE) that are very likely to
proceed via nonmonomeric catalyst complexes [33]; of course such reactions are not
promising candidates for being tested with a polymer-bound catalyst, since aggrega-
tion of catalytic moieties is (usually) prevented in or on a polymer, due to site isola-
tion.
The reactions were generally performed in the presence of 0.2 equiv. of supported
Ti-TADDOLate catalyst (Scheme 6,a). The desired amount (0.2 equiv.) of polymer-
bound TADDOL p-5 p-7, p-20, p-21, p-26, or p-28 was suspended in toluene, followed
by evaporation of the solvent under high vacuum to remove traces of H2O. The beads
were resuspended in toluene, and an equimolar amount of Ti(OiPr)4 (0.2 equiv.) was
added for loading of the beads with titanate. The beads thereby adopted a yellow color
(Scheme 6,b and c). After stirring at room temperature for 14 h, the solvent was
evaporated under reduced pressure for removal of i-PrOH. The polymer-bound
Ti-TADDOLates p-5 ¥ Ti(OiPr)2 p-7 ¥ Ti(OiPr)2), p-20 ¥ Ti(OiPr)2, p-21 ¥ Ti(OiPr)2,
p-26 ¥ Ti(OiPr)2 or p-28 ¥ Ti(OiPr)2 were resuspended in toluene, Ti(OiPr)4 (1 equiv.)
as well as PhCHO (1 equiv.) were added, and the reaction mixture was cooled to À 208,
followed by addition of a 2m solution of Et2Zn (1.8equiv.) in toluene under Ar.
After 2 h, the solvent was removed by syringe, preventing contact of the poly-
mer-bound Ti-TADDOLates with air. The beads were washed several times
with toluene, the combined organic fractions were extracted with 1m aqueous HCl,
and the products were isolated from the organic phases. The conversions and
enantioselectivities were determined by GC on a chiral column with the crude products.
The polymer beads were dried, resuspended in the reaction solvent, and recharged with
substrates for a new run (Scheme 6,a). The conversions and enantioselectivities
observed in the Et2Zn-to-PhCHO addition with polymer-bound Ti-TADDOLates

Helvetica Chimica Acta Vol. 85 (2002)
363
Scheme 6. a) Multiple Use of Polymer-Bound Ti-TADDOLates in the Addition of Et2Zn to PhCHO. b) Polymer
Beads of p-6 (diameter ca. 400 mm) before (colorless) and after Loading with Titanate (yellow). c) Two Round-
Bottomed Flasks Containing a Suspension of Beads of p-6 Swollen in Toluene before (colorless) and after
Loading with Titanate (yellow).
p-5 ¥ Ti(OiPr)2 p-7 ¥ Ti(OiPr)2, p-20 ¥ Ti(OiPr)2, p-21 ¥ Ti(OiPr)2, p-26 ¥ Ti(OiPr)2, or p-
28 ¥ Ti(OiPr)2 are listed in Table 1.
In most cases the enantiomer ratios and conversions observed with these polymer-
bound Lewis acids were similar to those observed with Ti-TADDOLate complexes

364
Helvetica Chimica Acta Vol. 85 (2002)
Table 1. Conversions and Enantiomer Ratios in the Formation of 1-Phenylpropan-1-ol by Addition of Et2Zn to
PhCHO, Catalyzed by Polymer-Bound Ti-TADDOLates. If not stated otherwise, 20 mol-% of supported
catalyst was employed.
Polymer
Loading [mmol gÀ1
]
Conversion [%]
(S)/(R)
p-5 ¥ Ti(OiPr)2
0.25
0.10
84
97
54
71
7897: 3
92
99
78
quant.
97: 3
98: 2
91 : 9
95 : 5
a
p-6 ¥ Ti(OiPr)2
0.10
0.10
0.25
0.14
0.10
0.14
0.11
0.25
0.10
0.25
0.10
0.21
0.10
)
)
b
97: 3
98: 2
8
93 : 7
98: 2
98
94 : 6
99 : 1
97: 3
98: 2
c
p-7 ¥ Ti(OiPr)2
p-20 ¥ Ti(OiPr)2
p-21 ¥ Ti(OiPr)2
)
9 : 11
: 2
c
)
76
d
76
60
70
quant.
99
)
d
)
p-26 ¥ Ti(OiPr)2
p-28 ¥ Ti(OiPr)2
a) 2 mol-% of catalyst are used. b) 5 mol-% of catalyst are used. c) Reaction time 14 h. d) [PhCHO] ˆ 0.05m.
under homogeneous conditions (thus, with Ti-TADDOLate 28 ¥ Ti(OiPr)2 in solution:
(S)- and (R)-1-phenylpropan-1-ol were obtained in a 99 :1 ratio, with complete
conversion after 2 h). Importantly, polymers p-5 ¥ Ti(OiPr)2 p-7 ¥ Ti(OiPr)2, p-20 ¥
Ti(OiPr)2, p-21 ¥ Ti(OiPr)2, and p-26 ¥ Ti(OiPr)2, prepared by copolymerization of
cross-linking TADDOLs with styrene by novel approach led, in most cases, to similar
results as those obtained with DVB-cross-linked polymer p-28 ¥ Ti(OiPr)2 [20] (cf.
Scheme 5,c and Table 1). Generally, the catalytic performance of the polymer-
supported TADDOLates with respect to conversion and enantioselectivity increased
with decreasing loading (e.g., results with polymer p-6 ¥ Ti(OiPr)2 in Table 1). This can
be rationalized by the fact that, with our polymers, a lower degree of loading
corresponds to a lower degree of cross-linking, and that the accessibility of the catalytic
sites is better in the case of less highly cross-linked polymer resins. In the same context,
the lower enantioselectivities observed with p-7 ¥ Ti(OiPr)2 (of degrees of loading
similar to the other polymers) can be interpreted as resulting from the high degree of
cross-linking achieved with the second-generation TADDOL cross-linker 7 bearing as
many as 16 peripheral styryl groups. Interestingly, also polymer p-26 ¥ Ti(OiPr)2 leads to
very good results although two polystyrene chains are interconnected at the TADDOL
acetal center, close to the metal-bonding site.
As mentioned in the Introduction, one aim of employing heterogeneous catalysts is
to re-use them as often as possible in consecutive catalytic runs with as little as possible
loss of activity in order to justify the synthetic effort of their preparation (first the
synthesis of suitable monomers and then the copolymerization). Therefore, we have set
the main focus of our investigations on this issue. Polymers p-6 ¥ Ti(OiPr)2, cross-linked
with first-generation TADDOL dendrimer 6 with three different degrees of loading
were re-used in 20 consecutive runs for the addition of Et2Zn to PhCHO, according to
the protocol outlined in Scheme 6. As can be seen from Fig. 4,a, the polymer with a

Helvetica Chimica Acta Vol. 85 (2002)
365
loading of 0.10 mmol gÀ1 gave rise to a stable performance without any loss in
enantioselectivity over 20 cycles. When increasing the loading ( ! 0.14 mmol gÀ1
!
0.25 mmol gÀ1), a slight decrease of the enantiomer purity of the product alcohol was
encountered during recycling. The enantiomer purities of 1-phenylpropan-1-ol
obtained with polymers p-5 ¥ Ti(OiPr)2, p-20 ¥ Ti(OiPr)2, p-21 ¥ Ti(OiPr)2, and p-28 ¥
Ti(OiPr)2 with a degree of loading of 0.10 mmol gÀ1 are presented in Fig. 4,b. Whereas
p-5 ¥ Ti(OiPr)2 and p-28 ¥ Ti(OiPr)2 gave rise to stable enantioselectivities during
recycling, comparable to p-6 ¥ Ti(OiPr)2, polymers p-20 ¥ Ti(OiPr)2 (containing TAD-
DOL cross-linker 20 with flexible alkyl spacers) and p-21 ¥ Ti(OiPr)2 (containing
TADDOL cross-linker 21 with rigid 1,1'-biphenyl spacers) showed an erratic and
Fig. 4. Enantioselectivities for the formation of 1-phenylpropan-1-ol during recycling of a) polymer p-6 ¥
Ti(OiPr)2 with different degrees of loading and of b) polymers p-5 ¥ Ti(OiPr)2, p-20 ¥ Ti(OiPr)2, p-21 ¥ Ti(OiPr)2
and p-28 ¥ Ti(OiPr)2 with degrees of loading of 0.10 mmol gÀ1

366
Helvetica Chimica Acta Vol. 85 (2002)
unstable performance accompanied by an overall loss of the enantioselectivity with
which 1-phenylpropan-1-ol was formed during multiple use.
Besides maintaining a constant enantioselectivity, polymer-bound catalysts should
also be recyclable with no or only minor loss in reaction rate. The kinetics of the Et2Zn
to PhCHO additions can be easily followed by GC analysis of samples taken from the
reaction mixture. Some representative examples of reaction kinetics during recycling,
measured with different polymers of 0.10 mmol gÀ1 loading, are presented in Fig. 5.
Besides a constant performance with respect to enantioselectivity (cf. Fig. 4,a) polymer
p-6 ¥ Ti(OiPr)2, cross-linked with first-generation TADDOL dendrimer 6, also gave rise
to very stable reaction rates during recycling (Fig. 5,a). Within experimental error, the
reaction rate in the 20th run is identical to the one in the first run (!). All polymers p-5 ¥
Ti(OiPr)2, p-20 ¥ Ti(OiPr)2, and p-21 ¥ Ti(OiPr)2, prepared with nondendritic TADDOL
cross-linkers, showed a decline in reaction kinetics during multiple use. The reaction
rates measured with p-20 ¥ Ti(OiPr)2 constitute a representative example (Fig. 5,b):
with each recycling step, the conversion dropped step by step. The decline of the
reaction kinetics during re-use was even more pronounced, when we increased the
degree of loading of the polymers: the polymer pores get blocked more easily with
increasing degrees of cross-linking, thus leading to a faster decline in catalytic activity
(of both reaction rate and enantioselectivity, cf. Fig. 4,a). DVB-Cross-linked polymer
p-28 ¥ Ti(OiPr)2 had given rise to a constant performance with respect to enantiose-
lectivity during recycling (cf. Fig. 4,b), like p-6 ¥ Ti(OiPr)2. In contrast, however, the
reaction rates gradually dropped (Fig. 5,c), in pronounced contrast to polymer p-6 ¥
Ti(OiPr)2 (Fig. 5,a). This means that only with dendritically cross-linked polymer p-6 ¥
Ti(OiPr)2 can a constant performance with respect to enantioselectivity and reaction
rate be achieved. All other polymers investigated gave rise to a decline of the catalytic
properties during recycling.
What might be the origin of the excellent stability of the dendritically cross-linked
p-6 ¥ Ti(OiPr)2 in contrast to the other polymers? The majority of catalytic sites are
located in the interior of the polymer beads. Therefore, swelling of the polymer resin in
the reaction solvent is indispensable to allow access to the catalytic centers throughout
the polymer network. Thereby, the polymer chains move apart, and the reactants are
allowed to penetrate into the resin. To guarantee a long-term stability of a polymer-
bound catalyst, the swelling properties of the polymer have to remain unchanged
during recycling. We decided to measure swelling factors of our polymers in the
reaction solvent toluene prior to and after 20 recycling steps. The results are presented
in Fig. 6. It is obvious that, only in the case of dendritically cross-linked p-6, is a
constant swelling behavior encountered: although the original swelling factor (ca. 2.5 in
toluene) is the lowest of all polymers tested it remains unchanged during 20 recycling
steps (!). The swelling factors of all other polymers decline, which results in restricted
diffusion of substrates into the polymer resin; this explains the loss of catalytic activity
of these polymers during re-use. Especially polymer p-20 ¥ Ti(OiPr)2, containing
TADDOL cross-linker 20 with flexible alkyl spacers, underwent a most drastic loss in
swelling ability after 20 runs (from a factor of 4.0 to 1.7), so that only a fraction of
catalytic sites might still have been accessible for substrates, which accounts for the
pronounced decline in activity (cf. Fig. 4,b, and Fig. 5,b). It appears that the use of the
first-generation dendritic TADDOL cross-linker 6 leads to a polymer p-6 with a

Helvetica Chimica Acta Vol. 85 (2002)
367
Fig. 5. Reaction kinetics during recycling of polymer a) p-6 ¥ Ti(OiPr)2 b), p-20 ¥ Ti(OiPr)2, and c) p-28 ¥
Ti(OiPr)2 with degrees of loading of 0.10 mmol gÀ1

368
Helvetica Chimica Acta Vol. 85 (2002)
Fig. 6. Swelling factors in toluene of different polymers prior to use and after 20 recycling steps in the addition of
Et2Zn to PhCHO
very persistent and stable morphology, which explains its superior catalytic perform-
ance.
To learn more about the factors that influence the reaction kinetics observed with p-
6 ¥ Ti(OiPr)2 several additional experiments were performed (Fig. 7): i) It was shown
that the reaction rate depends on the degree of loading of the polymers: the higher
loaded (and, therefore, also the higher cross-linked) the polymer, the slower the rate
(Fig. 7,a), due to restricted diffusion of substrates. ii) A quantity of smaller beads has a
larger surface area than the same mass of larger beads, so that the total area for
diffusion of substrates into the polymer beads is larger when using smaller beads; if the
diffusion of substrates into the polymer is rate-limiting, the reaction rate is expected to
be faster with smaller beads; this was actually the case with polymer beads of p-6 ¥
Ti(OiPr)2: beads with a diameter of ca. 400 mm gave rise to a faster reaction rate than
beads with a diameter of ca. 800 mm (Fig. 7,b) with the same degree of loading of
0.10 mmol gÀ1. iii) Stirring the reaction mixture with different rpm did not have any
influence on the kinetics: an unstirred suspension gave the same reaction rate as that
stirred at 600 rpm (Fig. 7,c); this offers, of course, the possibility to perform the
reactions without stirring, so that abrasion of the beads is avoided.

Helvetica Chimica Acta Vol. 85 (2002)
369
Fig. 7. Dependence of the reaction kinetics with polymer p-6 ¥ Ti(OiPr)2 on a) the degree of loading, b) the
diameter of the polymer beads, and c) the stirring speed of the reaction mixture

370
Helvetica Chimica Acta Vol. 85 (2002)
Usually, the rate of a catalytic reaction in solution is expected to be faster than the
corresponding rate measured with a polymer-bound catalyst, due to hindered diffusion
of the substrates to the catalytic centers within the polymer. This was indeed found
when we compared the reaction rates caused by DVB-cross-linked polymer p-28 ¥
Ti(OiPr)2 with those observed when we used the corresponding monomer 28 ¥
Ti(OiPr)2 under homogeneous conditions (Fig. 8,a). However, when comparing the
kinetics of the dendrimer complex 6 ¥ Ti(OiPr)2 in solution and the dendritically cross-
linked polymer p-6 ¥ Ti(OiPr)2 under heterogeneous conditions, the opposite was the
case: the reaction rate measured with the polymer-bound catalyst was faster than that
under homogeneous conditions (!) (Fig. 8,b), a result that was reproduced several
times.
Fig. 8. Comparison of the reaction kinetics under homogeneous and heterogeneous conditions: a) 28 ¥ Ti(OiPr)2
and p-28 ¥ Ti(OiPr)2; b) 6 ¥ Ti(OiPr)2 and p-6 ¥ Ti(OiPr)2

Helvetica Chimica Acta Vol. 85 (2002)
371
One interpretation of this phenomenon might be that, in solution, the dendritic
arms of complex 6 ¥ Ti(OiPr)2 are wrapped around the TADDOL core, thus hindering
access of substrates to the catalytic center. When these arms are −fixed× in the
polymerization process, they are immobilized away from the TADDOL moiety, and a
better diffusion of reactants to and from the catalytic centers, although now located
within a polymer resin, is provided. A second possible explanation is based on the
principle of −molecular imprinting× [16][17]. TADDOL Cross-linker 6 is chiral (C2
symmetry), so that, in the copolymerization with styrene, chiral cavities within the
polymer matrix are likely to be generated. In an enzyme-like way, such chiral cavities
around the catalytic centers could reinforce the catalytic activity. In contrast, polymer
p-28, generated with achiral DVB as cross-linker, would contain an achiral polymer
backbone. Finally, the catalytic centers in p-6 ¥ Ti(OiPr)2 are located in a more polar
environment within the apolar polystyrene matrix due to the −oxygen-rich× dendritic
benzyl ether branches surrounding the TADDOL cores; this could lead to an
accumulation of the polar substrates (PhCHO, Et2Zn and Ti(OiPr)4) in these regions,
leading to higher local concentrations and, therefore, a higher reaction rate. This is not
the case with polymer p-28 ¥ Ti(OiPr)2 and may serve as a third rationale supporting a
higher reaction rate with polymer p-6 ¥ Ti(OiPr)2 (under heterogeneous conditions), as
compared to the monomer 6 ¥ Ti(OiPr)2 (under homogeneous conditions). Elemental
analysis of a sample of p-6 ¥ Ti(OiPr)2 indicated that 85% of the TADDOL sites are
complexed with Ti. In contrast, all dendritic TADDOL ligands 6 must be complexed
with Ti-atom under homogeneous conditions. Due to these facts, and since the
performance of the polymer-bound catalyst is higher than that of its soluble analogue,
the intrinsic activity of each TADDOLate moiety of the polymer-bound catalyst must
be higher than that of the soluble analogue (!).
As already mentioned, one aim of our work was to check whether catalytically
active TADDOL ligands could be generated by copolymerization of diester 25
followed by Grignard addition (cf. Scheme 5,b). The Ti-TADDOLate-mediated
addition of Bu2Zn to PhCHO was chosen as test reaction for polymers p-26 ¥
Ti(OiPr)2 and p-27 ¥ Ti(OiPr)2 [34] (Table 2). The reactions were performed in analogy
to the addition of Et2Zn to PhCHO. The Bu2Zn was prepared in situ by mixing BuLi
(solution in hexane) with ZnCl2 (solution in Et2O) in toluene and removing LiCl with a
syringe filter, prior to addition to the reaction mixture. For comparison, p-26 ¥
Ti(OiPr)2, prepared by copolymerization of TADDOL 26 with styrene, followed by
loading with titanate, was also tested in this reaction. It turned out that this polymer
gave rise to the highest catalytic activity. The enantiomer purity of 1-phenylpentan-1-ol
was only slightly lower than that observed in solution under identical conditions (99%
es [34]). Although the enantioselectivity achieved with p-26' ¥ Ti(OiPr)2 was only 2%
lower, the reaction was markedly slower. It was not possible, as in the case of the
addition of Et2Zn to PhCHO, to simply recycle the polymers by washing and reloading
with substrates. The polymer-bound Ti-TADDOLate complexes had to be hydrolyzed
with THF/1n HCl, followed by reloading with titanate, prior to each catalytic run.
Probably, the reaction of BuLi with ZnCl2 was not complete, and these reagents caused
inactivation of the polymer-bound Ti complex. Polymer p-26' ¥ Ti(OiPr)2 was used in
two additional runs, resulting in a decline in conversion, but not in enantioselectivity,
within experimental error (Table 2). When the test reaction was performed with

372
Helvetica Chimica Acta Vol. 85 (2002)
polymer p-27 ¥ Ti(OiPr)2, bearing naphthalen-2-yl groups at the TADDOL units,
1-phenylpentan-1-ol was obtained with a lower enantiomer purity in comparison to p-
26' ¥ Ti(OiPr)2, whereas, in solution, the analogous tetra(naphthalene-2-yl)-TADDOL
gave rise to higher selectivities than the corresponding tetraphenyl-TADDOL [34b].
This may mean that only a fraction of the available diester moieties has reacted with the
Grignard reagent to give polymer-bound TADDOL units. The catalyst was re-used
once more, resulting in a better selectivity but poorer conversion (Table 2). In
summary, catalytically active TADDOLate complexes can be obtained by this
approach. However, the performance of these catalysts is poorer than that of the
corresponding TADDOLates prepared by copolymerization of preformed TADDOL
ligands with styrene.
Table 2. Addition of Bu2Zn to PhCHO Catalyzed by Polymer-Bound Ti-TADDOLates. Polymer p-26 was
prepared by copolymerization of preformed TADDOL derivative 26 with styrene, whereas p-26' and p-27 were
prepared by copolymerization of diester 25 with styrene, followed by Grignard addition (cf. Scheme 5,b).
0.2 equiv.
OH
(S)
O
p-TADDOL•Ti(OiPr)2
+
+
1 equiv. ZnCl2
2 equiv. BuLi
H
1.2 equiv. Ti(OiPr)4
Et2O, toluene, –20°
p-TADDOL
(aryl group)
Catalytic cycle
Conversion [%]
(reaction time)
(S)/(R)
p-26 ¥ Ti(OiPr)2 (Ph)
p-26' ¥ Ti(OiPr)2 (Ph)
68(4 h)
96 : 4
94 : 6
95 : 5
94 : 6
89 : 11
93 : 7
1
2
3
1
2
67 (5 d)
61 (5 d)
43 (4 d)
70 (14 h)
40 (5 d)
p-27 ¥ Ti(OiPr)2
5. Application of Polymer-Bound Ti-TADDOLate Complexes in Enantioselective
1,3-Dipolar Cycloadditions of Diphenyl Nitrone to [(E)-But-2-enoyl]-oxazolidinone.
By the work presented in the previous section, it was demonstrated that copolymer-
ization of a dendritically surrounded TADDOL 6 with styrene affords a polymeric
reagent p-6, which, when loaded with titanate and employed in the enantioselective
addition of Et2Zn to PhCHO, gives rise to an unprecedented catalytic activity.
To demonstrate the usefulness of our dendritically cross-linked catalyst p-6, we
decided to perform 1,3-dipolar cycloadditions as another test reaction. J˘rgensen and
co-workers had successfully employed Ti-TADDOLates for the catalysis of such
reactions [35]. It had already been shown in our group that polymer-bound
TADDOLates of the type p-28 ¥ Ti(OiPr)2 can be used for the catalysis of the addition
of diphenyl nitrone to [(E)-but-2-enoyl]-oxazolidinone (Fig. 9), giving rise to similar
selectivities as in solution [20a]. Due to these results and to the fact that a linear
correlation was found between the enantiomer purities of the exo-cycloadducts 29 and
those of the TADDOLate employed, a transition state involving a single catalyst
moiety was proposed [20a].

Helvetica Chimica Acta Vol. 85 (2002)
373
Fig. 9. Reuse of polymers a) p-6 ¥ TiCl2 or b) p-6 ¥ (OTs)2 in the 1,3-dipolar cycloaddition of diphenyl nitrone to
[(E)-but-2-enoyl]-oxazolidinone

374
Helvetica Chimica Acta Vol. 85 (2002)
Polymer p-6 was loaded with titanate by the addition of a solution of TiCl2(OiPr)2 in
toluene [36] to the polymer beads suspended in toluene. A solution of diphenyl nitrone
and [(E)-but-2-enoyl]-oxazolidinone in toluene was added, and the mixture was
worked up after a reaction time of 40 h at 08. The cycloadducts were isolated, and the
conversion, the exo/endo selectivity, as well as the enantiomer ratio of exo-29 were
determined as described in [35a]. It turned out that high conversions could only be
achieved when 50 mol-% of polymer p-6 ¥ TiCl2 were used, whereas the reaction in
solution proceeded well in the presence of 10 mol-% of catalyst [20a][35a]. With
50 mol-% of p-6 ¥ TiCl2, the exo/endo selectivity (82 :18), the enantiomer purity of exo-
29 (75%), and the conversion (93%) were only slightly lower than in solution with
10 mol-% of tetraphenyl-TADDOLate (exo/endo(29) 90 :10, es(exo-29) 79%, con-
version 94% [35a]). Re-using the beads by simply washing (as in the addition of Et2Zn
to PhCHO) was not possible: the subsequent reaction proceeded very slowly and
nonselectively. The polymer-bound complexes had to be hydrolyzed by stirring the
beads in THF/1n HCl, followed by reloading with titanate. With this protocol, p-6 ¥
TiCl2 was reused six times without decline of the exo/endo ratio and of the
enantioselectivity. The conversions were between 80 and 96% in all cases (Fig. 9,a).
The following catalytic cycles were performed by steadily decreasing the amount of
catalyst by 10 mol-% from one run to the next. As can be seen from Fig. 9,a, the exo/
endo selectivity and the enantioselectivity remained on a constant level, within
experimental error, whereas the conversion in the runs with 30 and 20 mol-% of
catalyst decreased. However, the catalytic activity with respect to conversion could be
fully re-installed by subjecting the beads to a thorough hydrolysis procedure, prior to
run 10. These results indicate that the −quality× of the polymer-bound catalyst has
steadily increased during recycling: after a couple of runs in the presence of a larger
amount of p-6 ¥ TiCl2, it is possible to perform the cycloaddition with only 10 mol-% of
catalyst, which was not possible in the beginning. A very similar observation was made
by A. Heckel in our group with silica-supported TADDOLates, so that this peculiar
behavior cannot be attributed to the particular polymer support used [37]. After all, it
is very remarkable that the polymer-bound catalyst remained active over a period of 10
recycling steps, bearing in mind that hydrolysis and reloading with titanate took place
after each run, and that most of the catalytic sites are located inside the polymer beads.
The preferential formation of exo-cycloadducts 29 is induced, as shown above, by
Cl2Ti-TADDOLates. However, the formation of the corresponding endo-cycloadducts
is favored, when (TsO)2Ti-TADDOLates are employed [35b]. In solution, this reaction
gives rise to good results only when 50 mol-% of catalyst are used [35b]. Thus, polymer
p-6 (0.5 equiv.) was loaded with titanate by the addition of a solution of Ti(OTs)2-
(OiPr)2 in toluene [35b] to give p-6 ¥ Ti(OTs)2, according to the procedure described
for the preparation of p-6 ¥ TiCl2, and a solution of the starting materials was added.
After a reaction time of 40 h at room temperature, the reaction solution was separated
from the polymer beads, and pure endo-29 was isolated [35b]. The results observed
with p-6 ¥ Ti(OTs)2 (endo/exo 88 :12, es (endo-29) 93 :7, conversion 72%) were slightly
poorer than with tetraphenyl-TADDOLate in solution (endo/exo > 95 :5, es (endo-29)
97:3, conversion quantitative [35b]). Recycling was, again, possible only after
hydrolysis and reloading with titanate. The results are shown in Fig. 9,b. Within
experimental error, the diastereo- (endo/exo) and enantioselectivity remained un-

Helvetica Chimica Acta Vol. 85 (2002)
375
changed during four catalytic cycles, the best conversion was even obtained in the
fourth run. The fact that comparable results are obtained under homogeneous and
heterogeneous conditions suggests that this cycloaddition also proceeds via monomeric
TADDOLate complexes.
The results presented in this section demonstrate that the dendritically cross-linked
TADDOL polymers can also be successfully applied in 1,3-dipolar cycloadditions
giving rise to a catalytic performance as under homogeneous conditions8).
6. TADDOL Polymers with Achiral Dendrimers as Cross-Linkers. In the previous
sections, it was demonstrated that catalytically highly active TADDOL polymers can be
obtained by copolymerizing a dendritically surrounded TADDOL cross-linker with
styrene. However, the preparation of these cross-linkers, e.g., 6, is tedious and requires
quite a number of synthetic steps, including a chromatographic purification. Hence, the
question arose whether TADDOL polymers with a comparable activity could be
prepared in a simpler way, namely by copolymerizing an achiral and, therefore,
synthetically easily accessible, dendritic cross-linker with styrene and mono-styryl-
substituted TADDOL 28. Furthermore, we wanted to investigate whether a dendriti-
cally cross-linked polystyrene gives rise to special material properties, which account
for the observations made with p-6 ¥ Ti(OiPr)2. The dendritic cross-linkers 30, 31, and 32
of zeroeth, first, and second generation were prepared by etherification of 1,1,1-tris(4-
hydroxyphenyl)ethane with 4-vinylbenzyl chloride ( ! 30) or branches 3 ( ! 31) and 4
8
)
In 1999, J˘rgensen and co-workers reported on the 1,3-dipolar cycloaddition of diphenyl nitrone to alkyl
vinyl ethers using catalytic amounts of MeAl-BINOLates in solution [38]. In most cases, the exo-
cycloadducts, which were formed almost exclusively, were obtained in enantioselectivities of up to 96%. We
have already shown that cross-linking polymer-bound Al-BINOLates give rise to the same catalytic activity
as their homogeneous counterparts [22]. In an effort to implement this reaction with MeAl-TADDOLates,
we screened several TADDOL ligands and found that the best results in the addition of diphenyl nitrone to
tert-butyl vinyl ether were obtained, when 20 mol-% of tetra(naphthalen-2-yl)-TADDOL I (exo/endo
87:13, es(exo) 70%) were employed. The same results were achieved in the presence of 20 mol-% of a
polymer-bound TADDOLate prepared by copolymerization of TADDOL II with styrene and DVB. In
contrast to the reaction under homogeneous conditions, it was possible to increase the exo/endo selectivity
to 96 :4 by increasing the amount of polymer-bound catalyst to 50 mol-%, whereas the enantiomer purity of
the cycloadducts remained unchanged. The catalyst was recycled up to five times by hydrolysis followed by
reloading with AlMe3, giving rise to a stable performance. However, due to the rather modest
enantioselectivities, no further investigations were carried out with this reaction.
O
O
Ph
Ph
OtBu
OtBu
Ph
O
TADDOL•AlMe
N
N
OtBu
N
+
+
toluene
Ph
Ph
Ph
exo
endo
TADDOLs:
Arl
Arl
Arl
R1
R2
I R1 = R2 = Me
II R1 = styryl, R2 = H
O
OH
Arl =
OH
Arl
O

376
Helvetica Chimica Acta Vol. 85 (2002)
( ! 32) in the presence of K2CO3 and 18-crown-6 (Scheme 7,a). These reactions
proceeded cleanly without the formation of by-products, and all three compounds were
obtained in yields of ca. 90% after chromatography. So far, the best polymer p-6 with a
degree of loading of 0.10 mmol gÀ1 had been obtained by copolymerization of
dendrimer 6 (with eight peripheral styryl groups) with 80 equiv. of styrene. Therefore,
to achieve a comparable degree of cross-linking in polymers p-30, p-31, and p-32,
10 equiv. of comonomers styrene and TADDOL 28 per styryl end group of the
dendritic cross-linker were employed in the copolymerization process (Scheme 7,b).
The amount of TADDOL 28 was adjusted in such a way that the loading of the resulting
polymers was either 0.24 or 0.10 mmol gÀ1. For example, for the preparation of polymer
p-31 with a loading of 0.10 mmol gÀ1, 1 equiv. of dendrimer 31 was copolymerized with
59 equiv. of styrene and 1 equiv. of TADDOL 28 (total: 60 equiv. of mono-vinyl-
functionalized co-monomers per dendrimer 31 with six peripheral styryl groups).
Polymers p-30, p-31, and p-32 with a loading of 0.24 mmol gÀ1 were transformed
into the corresponding Ti complexes p-30 ¥ Ti(OiPr)2, p-31 ¥ Ti(OiPr)2, and p-32 ¥
Ti(OiPr)2 by the addition of Ti(OiPr)4, and employed for the catalysis of the addition
of Et2Zn to PhCHO, according to the protocol depicted in Scheme 6. Generally,
20 mol-% of supported Ti-TADDOLate was used. The enantiomer purities of
1-phenylpropan-1-ol during multiple uses are collected in Fig. 10. The enantioselec-
tivities (between 97 and 98%) in the first run are comparable for all three polymers,
and a loss in selectivity between 2 and 3% during 20 recycling steps is generally
observed. Whereas the selectivities of 1-phenylpropan-1-ol obtained with polymers
p-31 ¥ Ti(OiPr)2 and p-32 ¥ Ti(OiPr)2, containing first- and second-generation den-
drimers 31 and 32, respectively, as cross-linkers, are quite stable and reproducible,
recycling p-30 ¥ Ti(OiPr)2 gives rise to rather erratic results. All three polymers showed a
decline in reaction rate and swelling ability during recycling. Decreased loading (from
0.24 to 0.10 mmol gÀ1) did not increase the catalytic performance as it did with
p-6 ¥ Ti(OiPr)2.
In summary, it was not possible by this approach to obtain polymer-bound
TADDOLates with a comparable performance as p-6 ¥ Ti(OiPr)2. A dendritically cross-
linked polystyrene support itself is not sufficient to guarantee a long-term stability of
the polymer-bound catalyst. It seems that placing the TADDOL unit in the core of a
dendritic cross-linker would be necessary to obtain polymers with an outstanding
catalytic activity (constant swelling ability over many cycles!).
7. Summary and Conclusion. In the present paper, we have presented a new way to
immobilize chiral ligands on polystyrene: copolymerization of a TADDOL dendrimer
6 with styrene, followed by loading with titanate, affords polymer-bound Ti-
TADDOLate complexes with an unprecedented stability and activity during recycling.
Our approach is the first example of the use of dendrimers as cross-linkers in
polystyrene. Furthermore, we have demonstrated that these polymers p-6, in contrast
to all other polymers tested, maintain a constant swelling ability during recycling, which
might be the reason for their excellent performance. Surprisingly, p-6 ¥ Ti(OiPr)2 gives
rise to a faster reaction rate in the Et2Zn addition to PhCHO than its soluble precursor
6 ¥ Ti(OiPr)2. We have also shown that it is necessary to place the TADDOL ligand in
the core of the dendritic cross-linker; copolymerizing a mono-vinyl-substituted

Helvetica Chimica Acta Vol. 85 (2002)
377
Scheme 7. Dendrimers 30, 31, and 32 (a) Used as Polymer Cross-Linkers for the Copolymerization of Styrene
with TADDOL 28 (b)
a)
O
O
O
O
O
O
O
O
O
O
O
O
30
31
O
O
O
O
O
O
O
O
O
O
O
O
O
O
O
O
O
O
O
O
O
32
b)
Ph
Ph
Ph
O
O
H
OH
OH
Ph
Suspension
Copolymerization
TADDOL polymers
p-30, p-31 or p-32
28
dendrimers
+
+
30, 31 or 32

378
Helvetica Chimica Acta Vol. 85 (2002)
Fig. 10. Recycling of polymers p-30 ¥ Ti(OiPr)2, p-31 ¥ Ti(OiPr)2 ¥ Ti(OiPr)2 and p-32 ¥ Ti(OiPr)2 in the addition of
Et2Zn to PhCHO according to the conditions outlined in Scheme 6
TADDOL 28 with styrene and achiral dendrimers as cross-linkers affords a polymeric
reagent, which is less active during recycling than polymer p-6.
In the meantime, we have also successfully immobilized the chiral ligands BINOL
[22] and Salen [23] in a dendritically cross-linking fashion, in order to establish the
general applicability of this novel way of immobilization.
We gratefully acknowledge financial support by the Swiss National Science Foundation (project No. 2027-
048157 and 20-50674.97, and Chiral2, project No. 20-48157-96). The following companies provided chemicals:
Witco GmbH (Et2Zn), Chemische Fabrik Uetikon (diethyl tartrate). We thank Novartis Pharma AG, Basel, for
continuing financial support.
Experimental Part
1. General. Reagents: (i-PrO)2TiCl2 [35a] and 2m stock soln. of Et2Zn [32] were prepared according to
reported procedures. The compounds 1 [1][24], 2 4 [22], 11 [39], 15 [31], and 28 [20a] were synthesized
according to literature procedures. PhCHO was distilled prior to use. All other commercially available
chemicals were used as received from Fluka, Aldrich, or Acros. TLC: precoated silica gel 25 Durasil UV254 plates
(Macherey-Nagel); visualization by UV254 light or by development with phosphomolybdic acid solution
(phosphomolybdic acid (25 g), Ce(SO4)2 ¥ H2O (10 g), H2SO4 (conc., 60 ml), and H2O (940 ml)), followed by
heating with a heat gun. Flash column chromatography (FC): silica gel 60 (Fluka, 0.040 0.063 mm); N2-
pressure ca. 0.2 0.4×bar. Anal. HPLC: Waters HPLC system (Waters 515 HPLC Pump, Waters 484 tunable
absorbance detector, Waters automated gradient controller); Daicel Chiracel OD (Daicel Chemical Industries,
Ltd.; 4.6 Â 250 mm, 10 mm), eluent: hexane/i-PrOH 400 :1 to 9 :1; UV detection at l 254 nm. Cap. gas-
chromatography (CGC): Carlo Erba GC 8000; columns (Supelco): a) a-Dex (30 m  0.25 mm i.d.), b) b-Dex
(30 m  0.25 mm i.d.); injector temp. 2008, detector temp. 2258 (FID); carrier gas: H2. M.p.: B¸chi-510
apparatus with open capillaries, uncorrected. Optical rotation [a]rD:t:: Perkin-Elmer 241 polarimeter (10 cm, 1 ml
cell) at r.t.; p.a. solvents. IR Spectra: Perkin-Elmer 1600 FTIR; solns. in CHCl3; n in cmÀ1. NMR Spectra: Bruker
AMX-II-500 (1H: 500 MHz, 13C: 125 MHz), AMX-400 (1H: 400 MHz, 13C: 100 MHz), AMX-300 (1H: 300 MHz,
13C: 75 MHz), Varian XL-300 (1H: 300 MHz, 13C: 75 MHz), Gemini-300 (1H: 300 MHz, 13C: 75 MHz), or
Gemini-200 (1H: 200 MHz, 13C: 50 MHz); chemical shifts (d) in ppm downfield from TMS (d ˆ 0 ppm); J values
in Hz; spectra were recorded in CDCl3. MS: Hitachi-Perkin-Elmer RMU-6M spectrometer (EI), VGZAB2-

Helvetica Chimica Acta Vol. 85 (2002)
379
SEQ spectrometer (FAB; in a 3-nitrobenzyl-alcohol matrix (3-NOBA)), Finnigan MAT-TSQ 7000 spectrom-
eter (ESI), Bruker ReflexTM spectrometer (positive-ion mode) with N2 Laser (337 nm) (MALDI-TOF; in a 2,5-
dihydroxybenzoic acid (2,5-DHB), a 2-(4-hydroxy phenylazo)benzoic acid (HABA), a 2,4,6-trihydroxyaceto-
phenone (THA), or a dithranol matrix), Ion Spec Ultima 4.7 FT ion cyclotron resonance (ICR) spectrometer
(HR-MALDI; in a 2,5-DHB matrix); fragment ions in m/z with relative intensities (% of base peak) in
parentheses. Elemental analyses (C and H) were performed by the Microanalytical Laboratory of the
Laboratorium f¸r Organische Chemie, ETH-Z¸rich. The Ti analysis was performed by Analytische
Laboratorien, Prof. Dr. H. Malissa und G. Reuter GmbH, Lindlar, Germany.
2. Coupling of Benzyl Bromides with Hexol 1 or with 1,1,1-Tris(4-hydroxyphenyl)ethane to Give Dendritic
Cross-Linkers. General Procedure 1 (G P 1). K2CO3 (4 equiv.) was added to a soln. of hexol 1 [1, 24] (1 equiv.) or
1,1,1-tris(4-hydroxyphenyl)ethane (1 equiv.) in acetone, followed by addition of a soln. of the PhCH2Br
(4 equiv.) in acetone. After heating at 50 608 for 20 48h, the suspension was allowed to cool to r.t., and the
salts were filtered off and extensively washed with CH2Cl2. After evaporation of the solvent, the residue was
redissolved in CH2Cl2, and H2O was added. The org. layer was separated, and the aq. phase was extracted with
CH2Cl2, followed by drying of the combined org. phases (MgSO4), evaporation of the solvent, and isolation of
the product by FC.
3. Coupling of Alcohols with Benzyl Halides with NaH as a Base. General Procedure 2 (G P 2). A soln. of the
alcohol (1 3 equiv.) in THF was slowly added to a suspension of NaH (1 3 equiv.) in THF at 08. After
complete addition, the suspension was heated to 508 for 30 min and recooled to 08. Then, a soln. of the benzyl
halide (1 equiv.) in THF was slowly added, and the mixture was heated at 60 708. For workup, H2O was
carefully added at 08. After extraction (3 Â Et2O), drying of the combined org. phases (MgSO4), and
evaporation of the solvent, the resulting crude product was purified by FC.
4. Preparation of TADDOL Cross-Linkers 5 7 and 20, 21. (4R,5R)a,a,a',a'-Tetrakis{4-[3-(4-ethenylben-
zyloxy)benzyloxy]phenyl}-2,2-dimethyl-1,3-dioxolane-4,5-dimethanol (5). A suspension of hexol 1 [1,24]
(366 mg, 0.6 mmol), benzyl bromide 2 [22] (755 mg, 2.5 mmol), and K2CO3 (344 mg, 2.5 mmol) in acetone
(25 ml) was heated at 508 for 20 h, according to G P 1. Workup and FC (CH2Cl2 ! CH2Cl2/acetone 99.5 :0.5)
afforded 5 (580 mg, 66 %). Colorless foam. Rf (hexane/acetone 1:1) 0.55. [a]rD:t: ˆ À32.4 (c ˆ 1.0, CHCl3). IR
(CHCl3): 3360w, 3008m, 2879w, 1710s, 1606s, 1586s, 1509s, 1490m, 1450m, 1406w, 1372s, 1289m, 1264s, 1248s,
1087w, 1016s, 992w, 913w, 8 8w4, 8 31s. 1H-NMR (500 MHz, CDCl3): 1.05 (s, 2 Me); 3.97 (s, 2 OH); 4.48( s,
HÀC(4), HÀC(5)); 4.97, 5.02, 5.04, 5.05 (4s, 8 CH2O); 5.24 (dd, J ˆ 10.8, 0.9, 4 vinyl H); 5.74 (dd, J ˆ 17.7, 0.9, 2
vinyl. H); 6.68 6.74 ( m, 4 CHCH2); 6.82 (d, J ˆ 9.0, 4 arom. H (TADDOL)); 6.91 (d, J ˆ 8.9, 4 arom. H
(TADDOL)); 6.68 7.44 ( m, 32 arom. H, 8arom. H (TADDOL)). 13C-NMR (125 MHz, CDCl3): 27.22; 69.76;
69.79; 69.83; 77.60; 81.12; 109.23; 113.54; 113.82; 113.95; 114.10; 114.11; 114.30; 114.34; 114.40; 119.87; 120.02;
126.42; 126.44; 127.72; 128.89; 129.64; 129.73; 135.36; 136.44; 137.35; 137.36; 138.68; 138.70; 157.86; 157.96;
‡
159.02; 159.03. MALDI-TOF-MS (HABA): 1443.0 ([M ‡ Na] ). Anal. Calc. for C95H86O12 (1419.69): C 80.37, H
6.11; found: C 80.21, H 6.31.
(4R,5R)-a,a,a',a'-Tetrakis{4-[3,5-bis(4-ethenylbenzyloxy)benzyloxy]phenyl}-2,2-dimethyl-1,3-dioxolane-
4,5-dimethanol (6). According to G P 1, K2CO3 (0.32 g, 2.3 mmol) was added to a soln. of 1 [1, 24] (0.31 g,
0.58mmol) in acetone (20 ml), followed by addition of a soln. of 3 [22] (1.00 g, 2.3 mmol) in acetone (5 ml).
After heating under reflux for 40 h, the mixture was worked up, and the crude product was purified by FC
(CH2Cl2 ! CH2Cl2/acetone 99.5 :0.5) to give 6 (0.72 g, 64%). Colorless foam. Rf (hexane/acetone 2 :1) 0.36.
[a]Dr:t: ˆ À24.0 (c ˆ 1.00, CHCl3). IR (CHCl3): 3374w, 3008m, 1598s, 1509s, 1458m, 1372m, 1294m, 1158s, 1060m,
1017m, 913m, 8 31m. 1H-NMR (500 MHz, CDCl3): 1.05 (s, 2 Me); 3.92 (s, 2 OH); 4.49 (s , HÀC(4), HÀC(5));
4.92, 4.98, 4.99, 5.01 (4s, 12 CH2O); 5.23 (dd, J ˆ 11.8, 0.9, 8 vinyl H); 5.74 (dd, J ˆ 17.6, 0.8, 8 vinyl H); 6.52, 6.55
(2t, J ˆ 2.3, 4 arom. H); 6.62 (d, J ˆ 2.3, 4 arom. H); 6.67 (d, J ˆ 2.6, 4 arom. H); 6.62 6.73 (m, 8vinyl C HCH2);
6.80, 6.90 (2d, J ˆ 9.0, 8arom. H (TADDOL)); 7.23 ( d, J ˆ 7.0, 4 arom. H (TADDOL)); 7.25 7.55 (m, 32 arom.
H, 4 arom. H (TADDOL)). 13C-NMR (125 MHz, CDCl3): 27.26; 69.87; 69.89; 81.15; 101.56; 101.61; 106.36;
106.49; 109.27; 113.57; 114.11; 114.33; 126.42; 126.44; 127.55; 127.75; 128.87; 129.7; 135.44; 136.31; 136.33;
136.45; 137.37; 137.39; 138.69; 139.49; 139.52; 157.82; 157.92; 160.12; 160.13. MALDI-TOF-MS (dithranol):
‡
1971.2 ([M ‡ Na] ). Anal. calc. for C131H118O16 (1948.4): C 80.76 , H 6.10 ; found: C 80.75, H 6.11.
(4R,5R)-a,a,a',a'-Tetrakis(4-{3,5-bis[3,5-bis(4-ethenylbenzyloxy)benzyloxy]benzyloxy}phenyl)-2,2-di-
methyl-1,3-dioxolane-4,5-dimethanol (7). A suspension of 1 [1,24] (0.51 g, 0.95 mmol), benzyl bromide 4 [22]
(3.65 g, 4 mmol) and K2CO3 (0.525 g, 3.8mmol) in acetone (80 ml) was heated under reflux for 48h, according
to G P 1. After workup and purification of the crude product by FC (CH2Cl2/hexane 3 :1 ! CH2Cl2 ! CH2Cl2/
acetone 95 :5), 7 (2.56 g, 70%) was obtained. Colorless foam. Rf (hexane/acetone 2 :1) 036. [a]rD:t: ˆ À10.2 (c ˆ
1.00, CHCl3). IR (CHCl3): 3333w, 3008m, 2379w, 1596s, 1512m, 1456m, 1407w, 1372m, 1295m, 1157s, 1056m,

380
Helvetica Chimica Acta Vol. 85 (2002)
1017m, 991w, 913m, 8 33m. 1H-NMR (500 MHz, CDCl3): 1.02 (s, 2 Me); 4.10 (s, 2 OH); 4.44 (s , HÀC(4),
HÀC(5)); 4.88 4.96 (m, 28 CH O); 5.21 (dd, J ˆ 10.9, 0.7, 16 vinyl H); 5.71 (dd, J ˆ 17.6, 0.6, 16 vinyl H);
2
6.48 6.52 ( m, 12 arom. H); 6.57 6.71 (m, 16 CHCH2, 24 arom. H); 6.76, 6.87 (2d, J ˆ 8.9, 8 arom. H
(TADDOL)); 7.18( d, J ˆ 8.9, 4 arom. H (TADDOL)); 7.30 7.41 (m, 64 arom. H, 4 arom. H (TADDOL)).
13C-NMR (125 MHz, CDCl3): 27.26; 69.83; 69.93; 70.00; 81.18; 101.58; 101.66; 101.69; 106.23; 106.35; 106.45;
109.12; 113.55; 114.09; 114.28; 126.41; 126.58; 127.53; 127.74; 128.90; 129.73; 135.35; 136.30; 136.44; 137.33;
138.59; 139.24; 139.51; 157.75; 157.89; 160.02; 160.06; 160.11. MALDI-TOF-MS (THA/citrate): 3876.7 ([M ‡
‡
Na] ).
4-[(tert-Butyl)dimethylsiloxy]butan-1-ol (8). Imidazole (33.2 g, 488 mmol) and (t-Bu)Me2SiCl (36.8g,
244 mmol) were added to a soln. of butane-1,4-diol (20.0 g, 222 mmol) in DMF (300 ml) at 08. After stirring at
r.t. for 18h, DMF was evaporated under reduced pressure, Et 2O (300 ml) and H2O (400 ml) were added, the
org. phase was separated, and the aq. layer was extracted with Et2O (2 Â 300 ml). After drying the combined
org. phases (MgSO4), the crude product was subjected to FC (hexane/Et2O 1:1) to give 8 (18.0 g, 40 %).
1
Colorless oil. H-NMR (200 MHz, CDCl3): 0.04 (s, 2 Me); 0.89 (s, t-Bu); 1.25 (s, OH); 1.57 1.71 (m, 2 CH2),
3.60 3.75 (m, 2 CH2O).
1-({4-[(tert-Butyl)dimethylsilyloxy]butoxy}methyl)-4-ethenylbenzene (9). A soln. of 8 (18.0 g, 88 mmol) in
THF (50 ml) was added to a suspension of NaH (6.4 g, 264 mmol) in THF (400 ml) at 08, according to G P 2.
After heating the suspension at 508 for 30 min, the mixture was recooled to 08, and a soln. of 4-ethenylbenzyl
chloride (17.9 g, 88 mmol) in THF (50 ml) was slowly added. Then, the mixture was heated at 608 for 72 h.
Workup and FC (hexane/Et2O 2 :1 ! 1 :1) gave 9 (19.2 g, 68%). Yellow oil. Rf (hexane/CH2Cl2) 0.20. IR
(CHCl3): 3089w, 3005m, 2932s, 2856s, 1629w, 1512w, 1471m, 1406w, 1389w, 1361m, 1256s, 1088s, 1073s, 1006w,
990m, 938w, 911m, 8 46s, 8 32s. 1H-NMR (400 MHz, CDCl3): 0.04 (s, 2 Me); 0.89 (s, t-Bu); 1.57 1.71 (m, 2 CH2);
3.48( t, J ˆ 6.5, SiOCH2); 3.62 (t, J ˆ 6.3, PhCH2O); 4.49 (s, PhCH2O); 5.23 (dd, J ˆ 10.8, 0.9, 1 vinyl H); 5.73 (dd,
J ˆ 17.6, 1.0, 1 vinyl H); 6.71 (dd, J ˆ 17.6, 10.9, CH2CH); 7.29, 7.38(2 d, J ˆ 8.1 Hz, 4 arom. H). 13C-NMR
(100 MHz, CDCl3): 18.35; 25.97; 26.24; 29.52; 62.97; 70.26; 72.57; 113.67; 126.21; 127.83; 136.60; 136.87; 138.30.
‡
EI-MS: 320.2 (<0.05, M) , 118.1 (37), 117.1 (100), 116.1 (5), 115.1 (14), 95.6 (19), 91.1 (8), 75.0 (7), 73.1 (6),
57.1 (3).
4-(4-Ethenylbenzyloxy)butan-1-ol (10). A soln. of 9 (19.2 g, 60 mmol) in THF (400 ml) was treated with
Bu4NF ¥ 3 H2O (28.3 g, 90 mmol) at 08. After stirring at r.t. for 12 h, H2O (500 ml) and Et2O (500 ml) were
added, the org. phase was separated, and the aq. layer was extracted with Et2O (2 Â 500 ml). Drying of the
combined org. phases (MgSO4), evaporation of the solvent, and purification of the residue by FC (CH2Cl2/
acetone 5 :1 ! 3 :1) yielded 10 (12.0 g, 97%). Yellow oil. Rf (CH2Cl2/acetone 2 :1) 0.46. IR (CHCl3): 3629w,
3416m, 3008s, 2938s, 2864s, 1710m, 1629m, 1512m, 1447w, 1406m, 1447m, 1406m, 1361m, 1225s, 1217s, 1098s,
1049m, 1017w, 990m, 913m, 8 27m. 1H-NMR (400 MHz, CDCl3): 1.63 1.75 (m, 2 CH2); 2.25 (s, OH); 3.51, 3.64
(2t, J ˆ 5.9, 2 CH2O); 4.51 (s, PhCH2O); 5.23 (dd, J ˆ 11.1, 1.0, 1 vinyl H); 5.74 (dd, J ˆ 17.6, 0.9, 1 vinyl H); 6.71
(dd, J ˆ 17.6, 10.9, 1 CH2CH); 7.29, 7.39 (2d, J ˆ 8.1, 2 arom. H). 13C-NMR (100 MHz, CDCl3): 26.72; 30.17;
‡
62.74; 70.31; 72.79; 113.84; 126.29; 127.96; 136.52; 137.06; 137.72. EI-MS: 206.2 (14, M ), 134.1 (20), 133.1 (100),
118.1 (20), 117.1 (90), 116.1 (6), 115.1 (21), 105.1 (16), 91.1 (13), 77.1 (5). Anal. Calc. for C13H18O2 (206.28): C
75.69, H 8.79; found: C 75.18, H 8.57.
1-{[(tert-Butyl)diphenylsilylyloxy]methyl}-4-{[4-(4-ethenylbenzyloxy)butoxy]methyl}benzene (12). A soln.
of 10 (4.00 g, 19.4 mmol) in THF (25 ml) was slowly added to a suspension of NaH (1.41 g, 58.2 mmol) in THF
(100 ml) at 08, according to G P 2. After heating at 508 for 30 min, the mixture was recooled to 08, and a soln. of
11 [39] (9.40 g, 21.3 mmol) in THF (25 ml) was added, whereupon the suspension was heated at 708 for another
14 h. Workup and purification of the crude product by FC (CH2Cl2/hexane 3 :1) gave 12 (6.40 g, 58%). Yellow
oil. Rf (CH2Cl2/hexane 3 :1) 0.28. IR (CHCl3): 3074w, 3054w, 3008s, 2932m, 2859s, 1590w, 1512m, 1472m, 1428s,
1362s, 1302w, 1112s, 1088s, 1019m, 912m, 8 27s. 1H-NMR (400 MHz, CDCl3): 1.09 (s, t-Bu); 1.69 1.72 (m,
2 CH2); 3.47 3.40 (m, 2 CH2O); 4.48, 4.49 (2s, 2 PhCH2O); 4.76 (s, SiOCH2); 5.22 (dd, J ˆ 10.8, 0.8, 1 vinyl H);
5.73 (dd, J ˆ 17.6, 0.9, 1 vinyl H); 6.70 (dd, J ˆ 17.6, 10.9, CH2CH); 7.28 7.70 ( m, 18arom. H). 13C-NMR
(100 MHz, CDCl3): 19.31; 26.51; 26.52; 26.83; 65.36; 70.07; 70.10; 72.56; 72.72; 113.66; 126.00; 126.20; 127.50;
127.58; 127.70; 127.80; 129.66; 133.53; 135.52; 135.56; 136.58; 136.87; 137.19; 138.26; 140.37. EI-MS: 564.3 (<0.05,
‡
M ), 507.2 (3), 317.1 (4), 313.1 (3), 297.1 (3), 273.1 (2), 235.0 (5), 221.1 (3), 199.0 (23), 117.1 (100), 104.0 (16),
91.0 (6), 77.0 (3). Anal. calc. for C37H44O3Si (564.83): C 78.68, H 7.85; found: C 78.74, H 7.80.
(4-{4-[(4-Ethenylbenzyloxy)butoxy]methyl}phenyl)methanol (13). Bu4NF ¥ 3 H2O (7.00 g, 22.2 mmol) was
added to a soln. of 12 (6.25 g, 11.1 mmol) in THF (100 ml) at 08. After stirring at r.t. for 14 h, workup as
described for 10 gave 13 (2.80 g, 77%). Colorless oil. Rf (CH2Cl2/acetone 2 : 1) 0.32 and FC (CH2Cl2/acetone
10 :1). IR (CHCl3): 3602m, 3429m, 3008s, 2941m, 2862s, 1629m, 1512m, 1455w, 1406m, 1362m, 1089s, 1016m,

Helvetica Chimica Acta Vol. 85 (2002)
381
991m, 912s, 8 27m. 1H-NMR (400 MHz, CDCl3) 1.69 1.71 (m, 2 CH2); 1.76 (t, J ˆ 5.7, 1 OH); 3.46 3.49 (m,
2 CH2O); 4.47, 4.48(2 s, 2 PhCH2O); 4.66 (d, J ˆ 5.4, CH2OH); 5.22 (dd, J ˆ 10.8, 0.9, 1 vinyl H); 5.73 (dd,
J ˆ 17.6, 0.9, 1 vinyl H); 6.71 (dd, J ˆ 17.6, 10.9, 1 CH2CH); 7.27 7.38( m, 8arom. H). 13C-NMR (100 MHz,
CDCl3): 26.51; 65.15; 70.10; 70.13; 72.57; 113.70; 126.21; 127.04; 127.81; 127.85; 136.59; 136.89; 138.10; 138.24;
‡
140.19. EI-MS: 324.2 (<0.05, M ), 209.2 (2), 205.2 (9), 192.2 (2), 137.1 (24), 134.1 (11), 133.1 (100), 132.1 (4),
131.1 (6), 121.1 (27), 119.1 (12), 118.1 (11), 117.1 (65), 115.1 (9), 105.1 (8), 104.1 (5), 103.1 (5), 91.1 (18), 77.1
(10), 73.1 (15), 55.1 (8). Anal. calc. for C21H26O3 (326.4): C 77.27, H 8.03; found: C 77.28, H 8.16.
1-(Bromomethyl)-4-{[4-(4-ethenylbenzyloxy)butoxy]methyl}benzene (14). CBr4 (4.30 g, 12.4 mmol) and
PPh3 (3.30 g, 12.4 mmol) were added to a soln. of 13 (2.70 g, 8.3 mmol) in THF (100 ml) at 08. After stirring at 08
for 1 h and at r.t. for 12 h, the solvent was evaporated, and the residue was immediately subjected to FC (CH2Cl2/
hexane 2 :1 ! 4 :1) to afford 14 (2.88 g, 90%). Yellow oil. Rf (hexane/acetone 3 :1) 0.47. IR (CHCl3): 3008s,
2942m, 2861s, 1629m, 1512m, 1448w, 1406m, 1362s, 1091s, 1018m, 990m, 913s, 8 27m. 1H-NMR (400 MHz,
CDCl3): 1.69 1.72 (m, 2 CH2); 3.46 3.49 (m, 2 CH2O); 4.48 4.49 ( m, 2 PhCH2O, CH2Br); 5.23 (dd, J ˆ 10.9,
0.8, 1 vinyl H); 5.73 (dd, J ˆ 17.6, 0.8, 1 vinyl H); 6.71 (dd, J ˆ 17.7, 10.9, CH2CH); 7.25 7.39 (m, 8arom. H). 13C
NMR (100 MHz, CDCl3): 26.51; 33.36; 70.07; 70.29; 72.38; 72.58; 113.70; 126.21; 127.81; 127.91; 129.08; 136.59;
‡
136.97; 138.25; 139.09. EI-MS: 390.3 (<0.05), M ), 221.3 (6), 207.2 (10), 205.2 (17), 201.1 (12), 199.1 (13), 185.1
(17), 183.1 (18), 154.9 (4), 134.1 (11), 133.1 (100), 132.1 (4), 131.1 (6), 121.1 (2.4), 120.1 (5), 119.1 (3), 118.1 (9),
117.1 (57), 116.1 (3), 115.1 (10), 105.1 (12), 104.1 (27), 103.1 (7), 91.1 (10), 86.0 (10), 84.0 (14), 77.1 (4), 73.1 (6),
49.0 (4). Anal. calc. for C21H25O2Br (389.33): C 64.79, H 6.47; found: C 64.87, H 6.61.
4'-{[(tert-Butyl)diphenylsilyloxy]methyl}[1,1'-biphenyl]-4-methanol (16). NaBH4 (0.34 g, 9.0 mmol) was
added to a soln. of aldehyde 15 [31] (4.00 g, 8.9 mmol) in THF (100 ml) at 08. After stirring at r.t. for 20 h, H2O
(100 ml) was carefully added, and the mixture was extracted with Et2O (3 Â 200 ml). After drying of the
combined org. phases (MgSO4) and evaporation of the solvent 16 (4.30 g, quant.) was obtained as a slightly
yellow solid, which was pure according to 1H-NMR and directly used for further transformations without
purification. The 1H-NMR data corresponded to those in [31].
4-{[(tert-Butyldiphenylsilyloxy]methyl}-4'-[(4-ethenylbenzyloxy)methyl]-1,1'-biphenyl (17). A soln. of 16
(5.00 g, 11.1 mmol) in THF (30 ml) was added to a suspension of NaH (0.80 g, 33.2 mmol) in THF (90 ml) at 08,
according to G P 2. After heating at 508 for 30 min, the suspension was recooled to 08, and a soln. of 4-
ethenylbenzyl chloride (2.02 g, 13.3 mmol) in THF (30 ml) was added, followed by KI (5 mg). Then, the
resulting mixture was heated at 708 for 20 h. Workup according to G P 2and FC (hexane/CH2Cl2 3 :1 ! 1 :1)
afforded 17 (3.70 g, 59%). Yellow oil. Rf (hexane/Et2O 2 :1) 0.61. IR (CHCl3): 3072m, 3063m, 3008s, 2931s,
2858s, 1907w, 1826w, 1629w, 1589w, 1500s, 1472s, 1428s, 1376m, 1361m, 1308w, 1263w, 1112s, 1027w, 1006s, 939w,
912m, 8 26s. 1H-NMR (400 MHz, CDCl3): 1.11 (s, t-Bu); 4.58, 4.59, 4.82 (3s, 3 PhCH2O); 5.24 (dd, J ˆ 10.9, 0.9,
1 vinyl H); 5.76 (dd, J ˆ 17.6, 0.9, 1 vinyl H); 6.72 (dd, J ˆ 17.6, 10.9, CH2CH); 7.34 7.73 (m, 22 arom. H).
13C-NMR (100 MHz, CDCl3): 19.34; 26.86; 65.30; 71.81; 71.88; 113.81; 126.28; 126.42; 126.93; 127.11; 127.74;
128.03; 128.27; 129.71; 133.50; 135.60; 136.56; 137.04; 137.15; 137.88; 139.52; 140.21; 140.51. EI-MS: 568.5
‡
(<0.05, M ), 513.4 (11), 512.4 (32), 511.4 (78), 481.4 (5), 393.3 (7), 315.3 (5), 314.3 (25), 313.3 (100), 199.2
(32), 197.2 (7), 196.2 (5), 195.2 (25), 183.2 (6), 181.2 (19), 180.2 (42), 179.2 (8), 168.2 (7), 167.2 (21), 166.2 (6),
165.2 (13), 152.1 (4), 131.1 (3), 118.1 (11), 117.1 (55), 115.1 (6), 105.1 (8), 91.1 (7), 86.0 (19), 84.0 (30), 77.1 (8),
51.0 (12), 49.0 (34), 47.0 (5), 41.1 (6). Anal. calc. for C39H40O2Si (568.82): C 82.35, H 7.09; found: C 82.33, H 7.22.
4'-[(4-Ethenylbenzyloxy)methyl][1,1'-biphenyl]-4-methanol (18). Bu4NF ¥ 3 H2O (3.42 g, 10.9 mmol) was
added to a soln. of 17 (3.10 g, 5.5 mmol) in THF (120 ml) at 08. After stirring at r.t. for 12 h, workup as described
for the preparation of 10 and FC (CH2Cl2/acetone 10 :1) gave 18 (1.60 g, 89%). White solid. M.p. 124.0 126.08.
Rf (CH2Cl2/acetone 10 :1) 0.37. IR (CHCl3): 3667w, 3600m, 3008s, 2932m, 2860s, 1909m, 1702s, 1610s, 1500s,
1448w, 1384m, 1361m, 1263w, 1082s, 1039w, 1006s, 913w, 8 16m. 1H-NMR (400 MHz, CDCl3): 1.71 (t, J ˆ 5.8,
1 OH), 4.58, 4.59, 4.82 (3s, 3 PhCH2O); 4.74 (d, J ˆ 5.1, CH2OH); 5.24 (dd, J ˆ 10.9, 0.9, 1 vinyl H); 5.75 (dd, J ˆ
17.6, 1.0, 1 vinyl H); 6.72 (dd, J ˆ 17.6, 10.9, CH2CH); 7.34 7.61 (m, 12 arom. H). 13C-NMR (100 MHz, CDCl3):
65.13; 71.76; 71.91; 113.84; 126.29; 127.14; 127.29; 127.49; 128.03; 128.29; 136.56; 137.07; 137.40; 137.84; 139.92;
‡
140.23; 140.37. EI-MS: 331.2 (13), 330.2 (51, M ), 226.1 (6), 213.1 (25), 212.1 (6), 199.1 (63), 197.1 (100), 196.1
(15), 195.1 (15), 183.1 (10), 182.1 (5), 181.1 (11), 180.1 (21), 178.1 (6), 169.1 (11), 168.1 (17), 167.1 (64), 166.1
(9), 165.1 (28), 156.1 (5), 155.1 (40), 154.1 (8), 153.0 (12), 152.0 (19), 119.1 (11), 118.1 (79), 117.0 (74), 116.0
(6), 115.0 (19), 105.0 (11), 103.0 (6), 91.0 (16). Anal. calc. for C23H22O2 (330.42): C 83.61, H 6.71; found: C 83.42,
H 6.92.
4-(Bromomethyl)-4'-[(4-ethenylbenzyloxy)methyl]-1,1'-biphenyl (19). PBr3 (0.06 ml, 0.61 mmol) was
added to a suspension of 18 (0.20 g, 0.61 mmol) in Et2O (20 ml) at 08 and stirred at this temp. for 1 h. For
workup, H2O (20 ml) was added, and the mixture was extracted with CH2Cl2 (3 Â 50 ml). The combined org.