Scale-up studies for the asymmetric Julia-Colonna epoxidation reaction

Article abstract of DOI:10.1002/adsc.200404079

The asymmetric Julia-Colonna epoxidation reaction under triphasic/PTC conditions has been successfully scaled-up to a one-hundred gram substrate level providing the corresponding epoxy ketone with high optical purity (up to 97% ee) and in good yield (75-78%). The amount of polyamino acid catalyst could be reduced leading to a simplified reaction work-up. In order to minimise the overall reaction time baffled reactors with pitched-blade impellers have been employed.

Full text of DOI:10.1002/adsc.200404079

UPDATES
Scale-Up Studies for the Asymmetric Julia´ –Colonna Epoxidation
Reaction
Arne Gerlach,a,* Thomas Gellerb
a
Bayer Chemicals AG, Fine Chemicals Research & Development, 51368Leverkusen, Germany
Fax: (þ49)-214-309674144, e-mail: arne.gerlach.ag@bayer-ag.de
b
Bayer CropScience AG, Process Research, Alfred-Nobel-Str. 50, 40789 Monheim, Germany
Received: March 12, 2004; Accepted: July 19, 2004
´
Abstract: The asymmetric Julia–Colonna epoxida-
tion reaction under triphasic/PTC conditions has
been successfully scaled-up to a one-hundred gram
substrate level providing the corresponding epoxy
ketone with high optical purity (up to 97% ee) and
in good yield (75–78%). The amount of polyamino
acid catalyst could be reduced leading to a simplified
reaction work-up. In order to minimise the overall
reaction time baffled reactors with pitched-blade im-
pellers have been employed.
Scheme 1. Reaction conditions: Poly-l-leucine (pll), aqueous
H2O2/NaOH, tetrabutylammonium bromide (TBAB), tol-
uene, rt.
Keywords: amino acids; asymmetric catalysis; epoxi-
dation; heterogeneous catalysis; phase-transfer catal-
ysis
Scheme 2. Reaction conditions: 5 in THF, 3.4 equivs. phos-
gene (gaseous, addition over 6.5 h, at 22–338C), 16 h rt,
358C/80 mbar (ꢀTHF), then addition of n-hexane, 85%
yield.
´
The poly-a-amino acid-catalysed asymmetric Julia–Co-
lonna epoxidation has emerged as an efficient synthetic
method for the enantioselective functionalisation of
a,b-unsaturated ketones. Since its discovery in 1980,[1]
the heterogeneous catalyst and the oxidant/solvent sys-
tem has been optimised further so that now a broad sic/PTC conditions 4 could be obtained successfully
range of substrates containing a substituted enone sys- from enone 2[5] with 97%ee on a small scale.[6]
tem can be epoxidised in high yield and with high enan-
tioselectivity.[2]
In order to prepare larger quantities of epoxy ketone 4
we subsequently investigated the synthesis of the poly-
Recently, we have developed a new protocol for the a-amino acid catalyst, which is usually obtained by
´
Julia–Colonna epoxidation reaction under triphasic amine-induced polymerisation of the corresponding
conditions[3] employing our own polyamino acid cata- N-carboxyanhydride.[7] Direct phosgenation[8] of l-leu-
lyst. The addition of small amounts of tetrabutylammo- cine 5 in THF provided pure, crystalline l-leucine N-car-
nium bromide (TBAB) as a phase-transfer catalyst boxyanhydride (l-leu-NCA, 6) on a multi-100 g scale.[9]
(PTC) to the heterogeneous reaction mixture resulted
After some optimisation it turned out that the subse-
in a significant acceleration of the epoxidation reac- quent statistical polymerisation initiated by 1,3-diami-
tion.[4] With this new protocol we could obtain complete nopropane was preferably performed in toluene at
conversion of chalcone 1 into epoxy ketone 3 with high elevated temperature.[7d] This method leads to a range
chemo- and enantioselectivity (94% ee) within 7 mi- of chain lengths in the polymer product 7, although the
nutes at room temperature. In contrast to our new pro- average length can be controlled by varying the ratio
cedure less than 60% conversion has been observed of l-leu-NCA to 1,3-diamino propane.[7b, c]
without TBAB under otherwise identical epoxidation
conditions after 1.5 hours.
After complete consumption of l-leu-NCA 6 and ad-
dition of methanol to the reaction mixture the product
Some time ago we became interested in preparing polyleucine 7 has been isolated on a 100-g scale. This ma-
enantiomerically pure epoxy ketone 4 as a valuable syn- terial could be used directly without further purifica-
thetic intermediate. Using our newly developed tripha- tion[10] or pre-activation.[7d,11] Having established a relia-
Adv. Synth. Catal. 2004, 346, 1247–1249
DOI: 10.1002/adsc.200404079
ꢀ 2004 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
1247

Arne Gerlach, Thomas Geller
UPDATES
Scheme 3. Reaction conditions: 6 in toluene, addition of 0.0125 equivs. 1,3-diaminopropane, gradually heated to 808C (over
2 h), 1 h at 808C, then addition of methanol, 88% yield.
´
Table 1. Asymmetric Julia–Colonna epoxidation reaction of
the starting material. With 20 w/w % of poly-l-leucine 7
up to 20 hours (overnight) at room temperature were re-
quired to ensure full conversion (Table 1, Entry 1), al-
though epoxy ketone 4 was obtained in good optical pu-
rity. When the same reaction was performed in a sealed
glass autoclave providing efficient mixing (stirring rate
ca. 700 rpm) we could successfully reduce the reaction
time to only 12 hours (Table 1, Entry 2). In an attempt
to improve the enantioselectivity we lowered the reac-
tion temperature to 158C. However, the conversion af-
ter 20 hours was below 50% and the optical purity of 4
was slightly diminished (Table 1, Entry 3). Next, we
tried to reduce the amount of poly-l-leucine 7 employed
for the epoxidation (Table 1, Entries 4–6). To our sur-
prise even with only 5 w/w % pll the reaction was finish-
ed in 20 hours, although the enantiomeric excess drop-
ped to 92%. With such low pll catalyst loadings, reaction
work-up is dramatically simplified. The catalyst could
easily be recovered by filtration and re-used after wash-
ing.[13]
chalcone 2 as per protocol at room temperature.
[a]
Based on chalcone 2.
[b]
Determined by HPLC analysis.
Isolated yield of pure 4 after one crystallisation from to-
luene of the crude product.
Ascertained by chiral HPLC analysis of isolated 4 and
comparison to the racemic compound.
This reaction has been performed in a sealed autoclave
[c]
[d]
In conclusion, we have demonstrated that the asym-
´
metric Julia–Colonna epoxidation reaction under tri-
[e]
phasic/PTC conditions could be scaled-up successfully
to a 100-g substrate level. With only 10 w/w % of poly-
l-leucine catalyst epoxy ketone 4 could be obtained in
75% yield and 95.5% ee. Scaling-up the epoxidation re-
sulted in prolonged reaction times even when baffled re-
with efficient stirring.
[f]
Reaction temperature has been kept at 158C.
ble access to larger quantities of poly-l-leucine (pll, actors were used.
7),[12] the stage was set for scaling-up the epoxidation
of chalcone 2.
For our triphasic/PTC conditions the epoxidation rate
Experimental Section
was found to have a strong dependency on the stirring
speed. On a small scale highest reaction rates have
been obtained when the epoxidation was performed in
a 5-mL vial with maximum magnetic stirring (approxi-
mately 1250 rpm). For this scale-up study we have
used jacketed glass reactors and, in order to ensure effi-
cient radial and axial flow within the heterogeneous re-
action mixture, glass baffles at the reactor inner walls
have been employed in combination with pitched-blade
glass impellers at a stirring rate of 700 to 800 rpm.
Protocol for the Asymmetric Julia´ –Colonna
Epoxidation Reaction of Chalcone 2
All reactions have been performed in jacketed glass vessels
(250 mL, 1000 mL, 2000 mL) connected to a thermostat. To en-
sure efficient mixing the jacketed glass vessels were equipped
with four glass baffles at the reactor inner walls and pitched-
blade glass impellers. The stirring rate was usually kept at a
range of 700 to 800 rpm.
A 2000-mL jacketed glass vessel was successively charged
with tetrabutylammonium bromide (TBAB) (11 g, 0.032 mol,
10 mol %), poly-l-leucine (pll, 7; 10 g, 1.1 mmol, 0.35 mol %),
´
Scaling-up the Julia–Colonna epoxidation of 2 from a
0.5 g, 5.0 g, 25 g, 50 g to a 100 g substrate level under tri-
phasic/PTC conditions led to a significant increase of the
overall reaction time to allow complete consumption of toluene (1600 mL) and 5 M aqueous sodium hydroxide
1248ꢀ 2004 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
asc.wiley-vch.de
Adv. Synth. Catal. 2004, 346, 1247–1249

´
Scale-Up Studies for the Asymmetric Julia–Colonna Epoxidation
UPDATES
(82 mL, 0.410 mol, 1.3 equivs.). After cooling to 158C 30% hy-
References and Notes
drogen peroxide (162 mL, 1.57 mol, 5 equivs.) was added at
that temperature (slightly exothermic) and the resulting heter-
ogeneous mixture was warmed to 258C again. This mixture was
then kept at 258C for 1 hour before chalcone 2 (100 g,
0.318mol, 1.0 equiv.) was added as a solid. Subsequently, the
reaction was stirred for 20 hours in the dark at 258C. The reac-
tion mixture was diluted with ethyl acetate (1500 mL) and
quenched by addition of ice-cold 20% aqueous NaHSO3 solu-
tion (100 mL). A peroxide test of the aqueous layer was nega-
tive. After addition of water (1000 mL) a phase separation fur-
nished an organic layer containing epoxide 4 and an aqueous
layer containing insoluble poly-l-leucine. Poly-l-leucine was
then filtered out and washed with ethyl acetate (100 mL).
The combined organic layers were dried over sodium sulphate,
filtered and concentrated under vacuum to afford a beige solid.
This material was crystallised from toluene (300 mL) to furnish
epoxy ketone 4 as a white solid; yield: 79 g (75%). Analysis by
chiral HPLC indicated 95.5% ee.
´
[1] a) S. Julia, J. Masana, J. C. Vega, Angew. Chem. Int. Ed.
´
Engl. 1980, 19, 929–931; b) S. Julia, J. Guixer, J. Masana,
J. Rocas, S. Colonna, R. Annuziata, H. Molinari, J.
Chem. Soc. Perkin Trans. 1 1982, 1317–1324.
[2] a) C. Lauret, S. M. Roberts, Aldrichimica Acta 2002, 35,
47–51; b) M. J. Porter, S. M. Roberts, J. Skidmore, Bio-
org. Med. Chem. 1999, 7, 2145–2156.
[3] The reaction mixture observed by following the original
protocol[1] consists of an organic phase (toluene and
starting material/product), an inorganic aqueous phase
(sodium hydroxide and hydrogen peroxide) and a third
solid phase (insoluble polyamino acid catalyst).
[4] a) T. Geller, A. Gerlach, C. M. Krüger, H.-C. Militzer,
Chimia Oggi/Chemistry Today 2003, 21, 6–8; b) T. Gel-
ler, A. Gerlach, C. M. Krüger, H.-C. Militzer, Tetrahe-
dron Lett. 2004, 45, 5065–5067.
[5] E. P. Kohler, H. M. Chadwell, Org. Synth. Coll. Vol. I,
1941, 78 –8 0.
[6] Early screening experiments were conducted in magneti-
cally stirred vials (5 mL) using 75 mg (0.24 mmol) of
chalcone 2.
´
[7] a) M. E. Lasterra-Sanchez, U. Felfer, P. Mayon, S. M.
Preparation of Poly-l-leucine (pll) 7
Roberts, S. R. Thornton, C. J. Todd, J. Chem. Soc. Perkin
To a stirred suspension of l-leucine (5; 200 g, 1.51 mol, 1.0
equiv.) in THF (2000 mL) gaseous phosgene (507 g, 5.13 mol,
3.4 equivs.) was added continuously over 6.5 h. During this pe-
riod the reaction mixture became homogeneous and the reac-
tion temperature was allowed to rise to 338C. Stirring at ambi-
ent temperature was maintained for additional 16 h. Subse-
quently, most of the solvent THFand excess phosgene was dis-
tilled off at 358C/80 mbar. At this point n-hexane (1100 mL)
was added and a precipitate was formed. After stirring for
1 h the precipitate was filtered off and washed with n-hexane
(2ꢁ600 mL). Drying provided pure, crystalline l-leucine N-
carboxyanhydride (l-leu-NCA, 6); yield: 203 g (85%). Analyt-
ical data were in agreement with reported values.[8]
Freshly prepared l-leucine N-carboxyanhydride (l-leu-
NCA, 6; 100 g, 0.636 mol, 1.0 equiv.) was dissolved in anhy-
drous toluene (1500 mL) under argon at ambient temperature.
To this homogeneous solution was added the initiator 1,3-dia-
minopropane (0.589 g, 0.008 mol, 0.0125 equiv.). Subsequent-
ly, the reaction mixture was gradually heated to 808C over
2 h. During this period CO2 was liberated. Stirring was main-
tained for 1 h at 808C. After cooling down to 608C methanol
(1000 mL) was added. This heterogeneous reaction mixture
was stirred for 1 h at 608C before the solid was filtered off. Dry-
ing furnished poly-l-leucine (pll, 7); yield: 63 g (88%). The
product structure has been confirmed by MALDI-MS.
Trans. 1 1996, 343–348; b) W. Kroutil, M. E. Lasterra-
´
Sanchez, S. J. Maddrell, P. Mayon, P. Morgan, S. M. Rob-
erts, S. R. Thornton, C. J. Todd, M. Tüter, J. Chem. Soc.
Perkin Trans. 1 1996, 2837–2844; c) P. A. Bentley, W.
Kroutil, J. A. Littlechild, S. M. Roberts, Chirality 1997,
9, 198–202; d) S. Baars, K.-H. Drauz, H.-P. Krimmer,
S. M. Roberts, J. Sander, J. Skidmore, G. Zanardi, Org.
Process Res. Dev. 2003, 7, 509–513.
[8] a) F. Fuchs, Chem. Ber. 1922, 55, 2943; b) A. C. Farthing,
J. Chem. Soc. 1950, 3213–3217; c) W. H. Daly, D. Poche,
Tetrahedron Lett. 1988, 29, 5859–5862.
[9] Under the used conditions (see Scheme 2 and experi-
mental section) l-leu-NCA 6 was not accompanied by
a dark brown oil as described earlier.[7d]
[10] Further washings or grinding of the crude poly-l-leucine
7 had no effect on enantioselectivity or reaction rate.
[11] J. R. Flisak, K. J. Gombatz, M. M. Holmes, A. A. Jarmas,
I. Lantos, W. L. Mendelson, V. J. Novack, J. J. Remich, L.
Snyder, J. Org. Chem. 1993, 58, 6247–6254.
[12] This poly-l-leucine-1,3-diaminopropane catalyst is com-
mercially available from Fluka (product number: 93197).
[13] A single re-use of recovered poly-l-leucine catalyst 7 fur-
nished epoxy ketone 4 in 75% yield with 97.3% ee after
20 hours on a 100 g substrate scale.
Adv. Synth. Catal. 2004, 346, 1247–1249
asc.wiley-vch.de
ꢀ 2004 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
1249