An improved heterogeneous asymmetric epoxidation of homoallylic alcohols using polymer-supported Ti(IV) catalysts

Article abstract of DOI:10.1016/S0957-4166(98)00408-X

Heterogeneous asymmetric epoxidation of homoallylic alcohols has been achieved using polymer-supported Ti(IV) catalysts with tert-butyl hydroperoxide as oxidant. The enantioselectivities and chemical yields obtained are significantly higher (ee values up to 80%) than with monomeric tartrate ligands.

Full text of DOI:10.1016/S0957-4166(98)00408-X

TETRAHEDRON:
ASYMMETRY
Pergamon
Tetrahedron: Asymmetry 9 (1998) 3895–3901
An improved heterogeneous asymmetric epoxidation of
homoallylic alcohols using polymer-supported Ti(IV) catalysts
Jaana K. Karjalainen, Osmo E. O. Hormi ∗ and David C. Sherrington
Department of Chemistry, University of Oulu, PO Box 333, FIN-90570 Oulu, Finland
Received 17 September 1998; accepted 8 October 1998
Abstract
Heterogeneous asymmetric epoxidation of homoallylic alcohols has been achieved using polymer-supported
Ti(IV) catalysts with tert-butyl hydroperoxide as oxidant. The enantioselectivities and chemical yields obtained
are significantly higher (ee values up to 80%) than with monomeric tartrate ligands. © 1998 Elsevier Science Ltd.
All rights reserved.
1. Introduction
Chiral epoxy alcohols are versatile building blocks of high synthetic utility which are widely used in
organic synthesis.1 Despite their potential value only very few reports on the enantioselective synthesis of
3,4-epoxy alcohols have been published,2,3 and none of these describe a highly efficient enantioselective
procedure. We have previously reported on the synthesis of linear and branched/crosslinked poly(tar-
trate)s and have shown these optically active ligands to be highly effective in asymmetric epoxidations
of trans- and cis-allylic alcohols with titanium tetraisopropoxide and tert-butyl hydroperoxide.4–6 Thus
we have been encouraged to examine the efficiency of these polymer-supported catalysts towards the
epoxidation of homoallylic alcohols. In the present paper we would like to report a much improved
method for the synthesis of chiral 3,4-epoxy alcohols.
2. Results and discussion
The branched/crosslinked polyester 3 was prepared by condensation polymerisation of L-(+)-tartaric
acid and 1,8-octanediol as shown in Scheme 1.5 The branching/crosslinking degree was determined as
described previously by 1H NMR.5 These polymer ligands containing varying degrees of branching were
∗
Corresponding author. E-mail: osmo.hormi@oulu.fi
0957-4166/98/$ - see front matter © 1998 Elsevier Science Ltd. All rights reserved.
PII: S0957-4166(98)00408-X

3896
J. K. Karjalainen et al. / Tetrahedron: Asymmetry 9 (1998) 3895–3901
tested with four different homoallylic alcohols (Tables 1–4). We have previously pointed out that a high
level of branching/crosslinking of the polymer ligand inhibited the enantioselective epoxidation of allylic
alcohols.5,6 However, in the case of homoallylic alcohols the effect was surprisingly much less significant.
Scheme 1. Preparation of branched/crosslinked poly(tartrate)s. Reagents and conditions: toluene-p-sulfonic acid (3 mass%), 20
mol% excess of diol 2, ca. 130°C, 4 days
The results summarised in Tables 1 and 2 allow us to make some interesting observations. There is no
significant effect on enantioselectivity of the branching of the polymer ligand. Even using the polymer
ligand having a degree of branching of 16% (Table 2, entry 4) still gave satisfactory enantioselectivity.
The enantioselectivities varied from 38–54% as determined by chiral HPLC. The literature procedure2
using monomeric L-(+)-dimethyl tartrate with titanium isopropoxide (Table 1, entry 1 and Table 2, entry
1) was followed, but the enantioselectivities as well as the chemical yields were both lower than using our
polymer-supported analogue. However, Sharpless et al.2 have reported slightly better enantioselectivities
using L-(+)-diethyl tartrate with titanium isopropoxide, but both the enantioselectivities and the chemical
yields were still lower than using our polymer-supported catalysts. The lack of substrate reactivity still
remains a problem. In some cases the rate of reaction was high (Table 1, entry 4, Table 2, entry 3) but
generally reaction times of up to 1 week were required to obtain 50% conversion. The reaction products
were also prone to rearrangement to form isomeric tetrahydrofuranols. In some cases about 10–20% of
uncharacterised by-products with long retention times were detected by GC or some non-GC-detectable
side products were apparent from a shortfall in the mass balance. The recovery of polymer catalyst by
simple filtration aids the isolation of products. The complex work-up required in the Sharpless procedure
is remarkably simplified and foaming emulsions are avoided. One of the attractive features of using these
polymer-supported catalysts is that a large excess of catalyst can be used to accelerate the reactions and
yet the work-up and isolation of product remains very simple. Indeed all the reactions were found to
proceed much better with substrate:Ti:tartrate ratios in the range 100:100:200 to 100:200:400, i.e. using
roughly stoichiometric levels of Ti and tartrate. However, the systems remain catalytic in that TBHP
needs to be activated by the Ti/tartrate complex.
Tables 3 and 4 give the results of the asymmetric epoxidation of two other homoallylic alcohols. In
these two cases great difficulties were observed using low molecular weight tartrate (DMT and DET).
With our polymer-supported ligands we managed to reduce somewhat the formation of side-products
in the epoxidation of 4-methyl-3-penten-1-ol. In the homogeneous reactions, titanium-assisted epoxide
opening occurs to give diols (Table 3). With this substrate both the enantioselectivity and the chemical
yield are better with the polymer tartrate than with the monomeric tartrate ligand. The reactions were also
relatively fast (up to 3 days) compared to the epoxidation of cis/trans-3-hexen-1-ols. The epoxidation of
3-buten-1-ol is problematic in terms of lack of substrate reactivity. The chemical yields were much lower
than with the other substrates and the isolation of the product was difficult using flash chromatography.
However, in this case the large excess of polymer ligand facilitated the conversion to 20% at −20°C in 2
weeks. Interestingly, the enantioselectivities obtained were higher than with any other substrates (up to
80%). This same feature was also observed with monomeric ligand.2

J. K. Karjalainen et al. / Tetrahedron: Asymmetry 9 (1998) 3895–3901
3897
Table 1
Asymmetric epoxidation of trans-3-hexen-1-ol 4 with TBHP in the presence of L-(+)-polyester 3 and
Ti(OPri)4 in CH2Cl2 at −20°C
Table 2
Asymmetric epoxidation of cis-3-hexen-1-ol 6 with TBHP in the presence of L-(+)-polyester 3 and
Ti(OPri)4 in CH2Cl2 at −20°C
The asymmetric epoxidation of homoallylic alcohols using polymer-supported catalyst has several
characteristic features when compared with the epoxidation of allylic alcohols.7 The reactivity and
enantioselectivity are markedly lower than with allylic alcohols. The lengthening of the carbon chain by
one methylene unit presumably results in a more a flexible Ti complex and hence poorer enantiocontrol.
However, the more rigid environment of the polymeric ligand seems to some extent to compensate, giving
significant improvement over simple tartrate esters.8 The absolute configuration of the homoallylic epoxy
alcohols produced are opposite to that of the allylic epoxy alcohols. This same feature has been observed
with the monomeric tartrate system.2

3898
J. K. Karjalainen et al. / Tetrahedron: Asymmetry 9 (1998) 3895–3901
Table 3
Asymmetric epoxidation of 4-methyl-3-penten-1-ol 8 with TBHP in the presence of L-(+)-polyester 3
and Ti(OPri)4 in CH2Cl2 at −20°C
Table 4
Asymmetric epoxidation of 3-buten-1-ol 10 with tBHP in the presence of L-(+)-polyester 3 and
Ti(OPri)4 in CH2Cl2 at −20°C
3. Conclusion
We have demonstrated a significantly improved synthesis of chiral 3,4-epoxy alcohols. The enantio-
selectivities obtained vary from 30 to 80% with reasonable chemical yields (up to 75%). A large excess
of polymer-supported catalyst can be used to improve reaction rates while avoiding complex work-up.
In some cases, the polymer-supported catalyst even reduces by-product formation, for example titanium-
assisted epoxide opening to give diols.

J. K. Karjalainen et al. / Tetrahedron: Asymmetry 9 (1998) 3895–3901
3899
4. Experimental
4.1. General
The 1H NMR and 13C NMR spectra were obtained on Bruker AM200 MHz or Bruker DPX400 MHz
spectrometers. Chemical shifts (δ) are given in ppm relative to TMS. Gas chromatographic (GC) analyses
were carried out on a Perkin–Elmer 8500 equipped with a FID employing a 25 m column packed with
fused silica (phase layer 0.25 µm). The enantiomeric composition of the epoxide was measured by HPLC
using chiral Daicel Chiralcel OB or Chiralpak AS columns (25 cm×0.46 cm) together with a Daicel Chi-
ralcel OB or Chiralpak AS 5 cm×0.46 cm pre-column (RI-detection, eluent: 9:1 n-hexane:isopropanol,
flow rate: 0.5 ml/min). High resolution GC–MS analyses were performed by the Mass Spectrometry
Labotarory of the University of Oulu, Finland. Optical rotations were determined on Perkin–Elmer
243 B polarimeters using a 1 cm3 capacity, 1 dm path length, quartz cell. Flash chromatography was
performed using Merck silica gel 60 (230–400 mesh) with diethyl ether/petroleum ether (40–60°C) in
various proportions as eluent. Absolute configurations were determined by comparison of the observed
rotation by polarimetry with the literature value.
4.2. Materials
Pre-activated and powdered 4 Å molecular sieves were available from Aldrich Chemical Co. and they
were used as received. Cooling was accomplished through the use of the an acetone–liquid nitrogen bath.
The dichloromethane used did not contain methanol and therefore was not distilled but was stored over
CaCl2. Aqueous 70% tert-butyl hydroperoxide (TBHP) was obtained from the Aldrich Chemical Co.
Reagents handled by syringe were measured by weight or by volume.
4.3. General procedure for asymmetric epoxidation of homoallylic alcohols
The literature procedure for the epoxidation of allylic alcohols was followed with minor
modifications.2 An oven-dried three-necked round-bottomed flask equipped with a magnetic stirbar,
nitrogen inlet, septum and bubbler was charged with 4 Å powdered, activated molecular sieves, polymer
ligand 3 and dry CH2Cl2. The flask was cooled to −20°C and Ti(OPri)4 (via syringe) was added
dropwise with stirring. The reaction mixture was stirred at −20°C and after about 1 hour, 3 equiv. of
TBHP in iso-octane was added with a syringe at a moderate rate. The resulting mixture was stirred at
−20°C for at least 1 hour. The substrate (dissolved in dry CH2Cl2) was added dropwise with a syringe,
being careful to maintain the reaction temperature between −15°C and −20°C. The mixture was stirred
for an additional 3–12 hours at between −15°C and −20°C. The reaction mixture was stored in a
freezer for up to 3 weeks. The reaction was monitored by gas chromatography (GC) using dodecane as
an internal standard. The polymer was filtered from the reaction mixture and washed thoroughly with
CH2Cl2. Work-up was then performed. The crude product was purified by flash chromatography [eluent:
petroleum ether (40–60°C):diethyl ether, 1:1-→1:2-→100% diethyl ether] and analysed by 1H NMR, GC
and HRMS. All compounds showed satisfactory spectroscopic and analytical data.
4.3.1. (+)-3(R),4(R)-Epoxyhexan-1-ol 5
The epoxy alcohol 5 was prepared on a 3–10 mmol scale and isolated according to the general
1
procedure. H NMR (200 MHz, CDCl3): δ=3.70 (t, J=6 Hz, 2H), 2.68–2.85 (m, 2H), 1.37–1.98 (m,

3900
J. K. Karjalainen et al. / Tetrahedron: Asymmetry 9 (1998) 3895–3901
4H), 0.93 (t, J=7 Hz, 3H). MS (EI, m/z, relative intensity) 115 (<1, M+−1), 99 (<1), 85 (100), 74 (15),
67 (20), 57 (99), 41 (67). HRMS (EI) calcd for C6H12O2: 116.0837, found (M+1) 117.0878.
4.3.2. (+)-3(R),4(S)-Epoxyhexan-1-ol 7
The epoxy alcohol 7 was prepared on a 3–5 mmol scale and isolated according to the general
1
procedure. H NMR (200 MHz, C6D6): δ=3.58 (t, J=6 Hz, 2H), 2.85–2.93 (m, 1H), 2.57–2.66 (m,
1H), 2.26 (br s, 1H), 1.12–1.65 (m, 4H), 0.82 (t, J=7 Hz, 3 H). MS (EI, m/z, relative intensity) 115 (<1,
M+−1), 97 (<1), 85 (100), 74 (10), 67 (10), 57 (45), 41 (28). HRMS (EI) calcd for C6H12O2: 116.0837,
found 116.0857.
4.3.3. (+)-4-Methyl-3(R),4-epoxypentan-1-ol 9
The epoxy alcohol 9 was prepared on a 5 mmol scale and isolated according to the general procedure.
1H NMR (200 MHz, CDCl3): δ=3.75 (m, 2H), 2.83 (dd, J=5, 8 Hz, 1H), 2.73 (s, 1H), 1.54–1.91 (m,
2H), 1.27 (s, 3H), 1.23 (s, 3H). MS (EI, m/z, relative intensity) 101 (1, M+−15), 85 (47), 71 (7), 69 (2),
59 (100), 57 (25), 43 (21). HRMS (EI) calcd for C6H12O2: 116.0837, found (M+1) 117.0936.
4.3.4. (+)-3(R),4-Epoxybutan-1-ol 11
The epoxy alcohol 11 was prepared on a 14 mmol scale and isolated according to the general
1
procedure. H NMR (200 MHz, CDCl3): δ=3.73 (t, J=6 Hz, 2H), 2.997–3.09 (m, 1H), 2.75 (t, J=5
Hz, 1H), 2.68 (s, 1H), 2.53 (dd, J=3, 5 Hz, 1H), 1.82–1.98 (m, 1H), 1.54–1.71 (m, 1H). MS (EI, m/z,
relative intensity) 87 (9, M+−1), 71 (2), 61 (3), 57 (100), 45 (17), 41 (12). HRMS (EI) calcd for C4H8O2:
88.0524, found (M+1) 88.0511.
Acknowledgements
JKK acknowledges financial support from the Neste Oy Foundation, Finland which made this work
possible. She also thanks Ms. Päivi Joensuu from the University of Oulu, Finland for performing the MS
measurements.
References
1. Rossiter, B. E. Asymmetric Synthesis; Academic Press: Orlando, FL, 1985; Vol. 5, pp. 193–246.
2. Rossiter, B. E.; Sharpless, K. B. J. Org. Chem. 1984, 49, 3707–3711.
3. In fact, when using Zr(OR)4 which has a longer metal–oxygen bond than the Ti complex, the enantioselectivity in the
epoxidation of cis-homoallylic alcohol is higher. Ikegami, S.; Katsuki, T.; Yamaguchi, M. Chem. Lett. 1987, 83–84.
4. (a) Canali, L.; Karjalainen, J. K.; Hormi, O.; Sherrington, D. C. J. Chem. Soc., Chem. Commun. 1997, 123–124 (b)
Karjalainen, J. K.; Hormi, O. E. O.; Sherrington, D. C. Molecules 1998, 3, 51–59.
5. Karjalainen, J. K.; Hormi, O. E. O.; Sherrington, D. C. Tetrahedron: Asymmetry 1998, 9, 1563–1575.
6. Karjalainen, J. K.; Hormi, O. E. O.; Sherrington, D. C. Tetrahedron: Asymmetry 1998, 9, 2019–2022.
7. Raifel’d, Y. E.; Vaisman, A. M. Russ. Chem. Rev. 1991, 60, 123–145.
8. In order to get further insight into the catalyst system, preliminary quantum chemical ab initio and DFT calculations on
the model molecules of the chiral linear poly(tartrate ester)s with the different length spacers between the tartrate residues
have been performed (structures shown in Ref. 4) (Hormi, O. E. O.; Pietilä, L. O.; Ahjopalo, L., unpublished results). An
analysis of conformers possible for polymers with the length of the methylene units in the polymer backbone ((CH2)n,
n=2, 6, 8, 12) revealed that intramolecular polymer–Ti complexes, where the tartaric units are adjacent in the chain, are
feasible for 6, 8, 12. When n=2, such regular rigid structures are not possible due to the geometric constraints imposed
by the short alkyl chain. Therefore, good enantioselectivity is expected only for the polymers with longer alkyl chains.

J. K. Karjalainen et al. / Tetrahedron: Asymmetry 9 (1998) 3895–3901
3901
The same feature was also observed according to the results shown in Ref. 4. In the light of these results we think that
the complex between the polymer and titanium is intramolecular. (a) Turbomole 96.0/4.0.0 User Guide, September 1996,
San Diego: Molecular Simulations, 1996. (b) Discover 96.0/4.0.0 User Guide, September 1996, San Diego: Molecular
Simulations, 1996. The pcff force field used in the calculations was modified to fit the torsional behaviour of some model
molecules at the MP2/6-31G(d) level of theory. Details of the modifications will be described elsewhere (Korpelainen, V.;
Ahjopalo, L.; Mannfors, B.; Pietilä, L. O. in preparation).