Hydrolysis of Mono- and Diepoxyoctadecanoates by Alumina

Article abstract of DOI:10.1007/s11746-003-0792-z

In this study it is shown that the epoxide derivative of oleic acid, methyl 9, 10-epoxyoctadecanoate, is readily hydrolyzed to a diol when exposed to a commercial preparation of neutral alumina. Comparison of ¹H and ¹³C NMR spectra of the diol with those of standards showed that the product was the threo isomer. When methyl 9, 10-12, 13-diepoxyoctadecanoate was treated with alumina, a mixture of dihydroxytetrahydrofuran regioisomers, methyl 9, 12-epoxy-10, 13-dihydroxystearate and methyl 10, 13-epoxy-9, 12-dihydroxystearale, was obtained. These results show that alumina is an unsuitable support for epoxidation catalysts. However, alumina-catalyzed hydrolysis of fatty epoxides is an efficient way to synthesize polyhydroxy materials, and these materials are suitable for several industrial applications.

Full text of DOI:10.1007/s11746-003-0792-z

Hydrolysis of Mono- and Diepoxyoctadecanoates by Alumina
George J. Piazza*, Alberto Nuñez, and Thomas A. Foglia
USDA, ARS, Eastern Regional Research Center, Wyndmoor, Pennsylvania 19038
ABSTRACT: In this study it is shown that the epoxide deriva-
tive of oleic acid, methyl 9,10-epoxyoctadecanoate, is readily
hydrolyzed to a diol when exposed to a commercial prepara-
tion of neutral alumina. Comparison of 1H and 13C NMR spec-
tra of the diol with those of standards showed that the product
was the threo isomer. When methyl 9,10-12,13-diepoxyocta-
decanoate was treated with alumina, a mixture of dihydroxy-
tetrahydrofuran regioisomers, methyl 9,12-epoxy-10,13-dihy-
droxystearate and methyl 10,13-epoxy-9,12-dihydroxystearate,
was obtained. These results show that alumina is an unsuitable
support for epoxidation catalysts. However, alumina-catalyzed
hydrolysis of fatty epoxides is an efficient way to synthesize
polyhydroxy materials, and these materials are suitable for sev-
eral industrial applications.
MATERIALS AND METHODS
Chemicals. t-Butyl hydroperoxide (70%), oleic acid, and
methyl linoleate were from Sigma Chemical (St. Louis, MO).
Methyl ( )-threo-9,10-dihydroxyoctadecanoate and methyl
( )-erythro-9,10-dihydroxyoctadecanoate were purchased
from Larodan Fine Chemicals AB (Malmö, Sweden). Oxone
(2KHSO5·KHSO4·K2SO4), 1,1,1-trifluoroacetone, and alu-
minum oxide (activated, neutral, Brockmann 1, 150 mesh, 58Å)
were purchased from Aldrich (Milwaukee, WI). Trimethylsilyl
(TMS) derivatives were prepared using N,O-bis(trimethylsilyl)-
trifluoroacetamide from Pierce (Rockford, IL).
Preparation of fatty epoxides. Methyl 9,10-epoxyoctadec-
anoate 1 was prepared by the epoxidation of oleic acid with
t-butyl hydroperoxide by using immobilized peroxygenase as
the catalyst followed by methylation as described previously
(7). Methyl 9,10-12,13-diepoxyoctadecanoate 3 (Scheme 1)
was prepared as follows: Methyl linoleate (100 mg), 3 mL ace-
Paper no. J10561 in JAOCS 80, 901–904 (September 2003).
KEY WORDS: Alumina, methyl 9,10-12,13-diepoxyoctadec-
anoate, methyl 9,10-dihydroxyoctadecanoate, methyl 9,12-
epoxy-10,13-dihydroxyoctadecanoate, methyl 10,13-epoxy-
9,12-dihydroxyoctadecanoate, methyl 9,10-epoxyoctadec-
anoate.
Heterogeneous catalysts are easily recovered and reused in
continuous processes, and because of this, there is a continu-
ing search for ways to prepare heterogenized catalysts (1).
Epoxidation is an important industrial process, and epoxi-
dized FA derivatives are used as plasticizers and plastic stabi-
lizers (2). Recent research efforts have focused on the use of
transition metal complexes for epoxidation, and some of these
catalysts have been supported or heterogenized on alumina
(3,4). However, prior work showed that alumina is an effec-
tive catalyst for oxirane ring opening by oxygen- and nitro-
gen-containing nucleophiles (5,6). This raises the possibility
that heterogeneous catalysis by alumina might be effective in
promoting epoxide hydrolysis. Accordingly, we exposed
methyl 9,10-epoxyoctadecanoate and methyl 9,10-12,13-
diepoxyoctadecanoate to neutral alumina and found that the
epoxides are readily hydrolyzed to FA derivatives with hy-
droxy groups that are useful in a wide variety of applications.
Generally, dilute acid treatment is used for the hydrolysis of
epoxides. However, alumina is superior to such acid treatment
because it is easily removed from the product. Also, for in-
dustrial use, alumina is preferred because it is safe, low in
cost, and has a long shelf life.
*To whom correspondence should be addressed at USDA, ARS, ERRC, 600
E. Mermaid Lane, Wyndmoor, PA 19038. E-mail: gpiazza@arserrc.gov
SCHEME 1
Copyright © 2003 by AOCS Press
901
JAOCS, Vol. 80, no. 9 (2003)

902
G.J. PIAZZA ET AL.
tonitrile, 1,1,1-trifluoroacetone (300 µL), and 2 mL 40 mM MS (not derivatized): m/z (obs. fragment, intensity), 390
Na2EDTA were combined in a 50-mL glass-stoppered Erlen- ([M + 1 + 18 (H2O) + 41 (CH3CN)]+, 10.5%), 348 ([M + 18]+,
meyer flask. The flask was cooled in a 5°C water bath, and 195 14.7%), 331 ([M + 1]+, 100%); EI-MS: m/z (obs. fragment, in-
mg NaHCO3 and 0.7 g oxone were added. The reaction flask tensity), 187 ([HOCH(CH2)7COOCH3]+, 40.5%), 155 ([187 −
was shaken vigorously for 6 h. After the addition of 10 mL H2O, 32 (CH3OH)]+, 100%); APCI-MS (bis-TMS ether): m/z (obs.
the product was extracted with 50 mL diethyl ether. The ether fragment, intensity), 475 ([M + 1]+, 100%); EI-MS: m/z (obs.
fraction was washed with 2 × 20 mL H2O, dried with anhydrous fragment, intensity), 259 ([TMSOCH(CH2)7COOCH3]+,
Na2SO4, and the ether removed under a stream of nitrogen.
57.4%), 215 ([M − 259]+, 66.7%); 1H NMR: (400 MHz, C6H6,
Epoxide hydrolysis. Methyl 9,10-epoxyoctadecanoate 1 (5 δH) 1.16 (t, J = 6.8 Hz, 3H, CH3), 1.40–1.90 (m, 24H, CH2),
mg), neutral alumina (115 mg), and cyclohexane (0.6 mL) 2.38 (t, J = 7.4 Hz, 2H, 2-H2), 3.49 (bs, 2H, 9-H, and 10-H),
were placed in a 1.7-mL glass HPLC vial with a screw cap. 3.61 (s, 3H, COOCH3); 13C NMR: (100 MHz, C6H6, δC)15.0
The vial was attached to laboratory rotator (radius of 17 cm) (C-18), 23.8, 25.9, 26.6, 26.8, 30.0, 30.2, 30.5, 30.7, 30.8,
and rotated at 3.5 revolutions per minute. At the indicated 33.0, 34.8, 51.6 (COOCH3), 75.3 (C-9 and C-10), 174.3 (C-1);
time, the contents of the vial were filtered through a sintered (100 MHz, CDCl3, δC) 14.1 (C-18), 22.7, 24.9, 25.5, 25.6,
glass funnel and the alumina was washed sequentially with 5 29.0, 29.1, 29.3, 29.4, 29.5, 29.7, 31.9, 33.6, 34.1, 51.4
mL each of methanol, dichloromethane, and diethyl ether. (COOCH3), 74.5 (C-9 and C-10), 174.3 (C-1).
Hydrolysis of methyl 9,10-12,13-diepoxyoctadecanoate 3
(ii) Methyl 9,12-epoxy-10,13-dihydroxystearate (4) and
was conducted in an identical manner except that 230 mg methyl 10,13-epoxy-9,12-dihydroxystearate (5). Peak I:
neutral alumina was used. Alternatively, the epoxides 1 and 3 RT 5.9 min; Peak II: RT 9.4 min; IR (I): 3468 (hydroxy),
were hydrolyzed by placing them in a 0.6-mL mixture of 1738 (carbonyl), 1174 (C–O–C) cm−1; IR (II): 3388, 1738,
THF/water (3:2, vol/vol) containing 1% HClO4 and agitating 1171 cm−1; APCI-MS (not derivatized) (I): 404 ([M + 1 + 18
the mixture. At the indicated time, the product was partitioned + 41]+, 15.4%), 362 ([M + 18]+, 25.6%), 346 ([M + 2]+,
between 20 mL diethyl ether and 20 mL H2O. The ether layer 28.2%), 345 ([M + 1]+, 100%); APCI-MS (II): m/z (obs. frag-
was washed with 2 × 20 mL portions of 2% NaHCO3.
Chromatographic and instrumental methods. 1H and 13
NMR spectra were obtained as described previously (8).
ment, intensity), 362 ([M + 18]+, 16.0%), 346 ([M + 2]+,
28.0%), 345 ([M + 1]+, 100%); EI-MS (I): m/z (obs. frag-
ment, intensity), 308 ([M − 18 − 18]+, 0.9%), 295 ([M −
C
Epoxidized fatty methyl esters were analyzed by RP- 31(·OCH3) − 18]+, 1.1%), 277 ([M − 31 − 18 − 18]+, 0.9%),
HPLC. Solutions of the methyl esters in dichloromethane 187 ([HO=CHC6H12COOCH3]+, 100%), 157 ([M − 187]+,
were filtered through 13-mm, 0.45-µm syringe filters (PVDF; 23.7%), 155 ([187 − 32]+, 87.6%), 139 ([C8H16COOCH3 −
Scientific Resources, Eatontown, NJ). Dichloromethane was 32]+, 11.9%), 113 ([C7H12OH]+, 63.2%. EI-MS (II): m/z (obs.
removed under a stream of nitrogen, and the filtered products fragment, intensity), 308 ([M − 18 − 18]+, 1.3%), 295 ([M −
were dissolved in 1 mL isopropanol. Reaction mixtures were 31(·OCH3) − 18]+, 2.1%), 277 ([M − 31 − 18 − 18]+, 1.6%),
separated on two Symmetry 3.5 µm C18 reversed-phase 187 ([HO=CHC6H12COOCH3]+, 64.4%), 157 ([M − 187]+,
columns (150 × 2.1 mm and 50 × 2.1 mm) (Waters, Milford, 16.9%), 155 ([187 − 32]+, 100%), 139 ([C8H16COOCH3 −
MA) connected in series. Quantification of hydrolyzed epox- 32]+, 13.8%), 113 ([C7H12OH]+, 39.7%; APCI-MS (bis-TMS
ide products was made using a Varex MK III ELSD (Alltech, ether) (I): m/z (obs. fragment, intensity), 490 ([M + 2]+,
Deerfield, IL) operated at 55°C, and with N2 as the nebuliz- 46.7%), 489 ([M + 1]+, 100%), 417 ([M − TMSOH + 18 +
ing gas at a flow rate of 1.5 L/min. Mobile phase composition 1]+, 2.0%), 399 ([M − TMSOH + 1]+, 1.3%); APCI-MS (II):
and gradient were 0–5 min H2O/CH3CN (40:60, vol/vol); m/z (obs. fragment, intensity), 490 ([M + 2]+, 41.6%), 489
5–30 min H2O/CH3CN (40:60, vol/vol) to CH3CN (100); ([M + 1]+, 100%), 417 ([M − TMSOH + 18 + 1]+, 9.1%), 399
30–54 min CH3CN (100). The flow rate was 0.25 mL/min. ([M − TMSOH + 1]+, 2.6%); EI-MS (I): m/z (obs. fragment,
Products were characterized by HPLC with mass detection intensity), 315 ([M − C6H12OTMS]+, 2.6%), 259 ([M −
using EI-MS (Thermabeam Mass Detector; Waters) and at- TMSOC8H15COOCH3]+, 91.9%), 229 ([M − 259]+, 4.8%),
mospheric pressure chemical ionization (APCI) HPLC-MS 173 ([M − 315]+, 100%); EI-MS (II): m/z (obs. fragment, in-
(Micromass ZMD; Waters). The EI-MS detector was set to tensity), 315 ([M − C6H12OTMS]+, 1.1%), 259 ([M −
scan in the mass range of m/z 55–600 at 1000 amu/s and had TMSOC8H15COOCH3]+, 100%), 229 ([M − 259]+, 10.7%),
an ionization energy of 70 eV. Ionization source, nebulizer, 173 ([M − 315]+, 77.3%); NMR: (see Tables 1 and 2).
and expansion region temperatures were 200, 64, and 75°C,
respectively. When using the APCI-MS, the HPLC gradient
contained 0.1% formic acid, and the APCI-MS detector was
RESULTS AND DISCUSSION
set to scan in the mass range of m/z 150–550 at 400 amu/s. Methyl 9,10-epoxyoctadecanoate 1 was prepared from oleic
The corona, cone, and extractor voltages were 3700, 20, and acid. Monoepoxide 1 was exposed to alumina, and analysis
5 eV, respectively. The source and APCI heater temperatures by RP-HPLC showed that a single product was formed that
were 150 and 400°C, respectively.
had an RT of 18.2 min. The MS data and the 1H and 13C NMR
Structural confirmation of hydrolysis products. (i) Methyl spectra given in the Materials and Methods section show that
threo-9,10-dihydroxyoctadecanoate (2). Retention time (RT): a 1,2-diol was formed with hydroxy groups on C-9 and C-10.
18.2 min; IR: 3280 (hydroxy), 1702 (carbonyl) cm−1; APCI- Furthermore, comparison of the carbon and proton shifts with
JAOCS, Vol. 80, no. 9 (2003)

HYDROLYSIS OF MONO- AND DIEPOXYOCTADECANOATES
903
TABLE 1
1H NMR and 13C NMR of Methyl 9(12)-oxy-10,13-dihydroxystearate and Methyl
10(13)-Oxy-9,12-dihydroxystearatea (peak I, RT on RP-HPLC: 5.9 min)
Chemical shift
Number
Coupling
(400 MHz, C6D6, δH)
of protons
Appearance
Assignment
constant (Hz)
1.14 and 1.15
1.40–1.96
2.34 and 2.36
3.51
3.60
3.75
3H
22H
2H
1H
3H
1H
1H
t
m
t
m
s
CH3CH2–
–CH2–
–CH2COOCH3
–CH–OH
–COOCH3
–CH–OH
–HC(O)–CH–OH
(H-13 or H-9)
–HC(O)–CH–OH
(H-10 or H-12)
J = 7.0, J = 7.0
J = 7.6, J = 7.4
dt
d
J = 7.0, J = 2.9
J = 3.6
4.06
4.24
1H
td
J = 6.4, J = 9.2
13C NMR (100 MHz, C6D6, δC): 14.9 (CH3CH2–), 23.7, 23.7, 25.9, 25.9, 26.5, 26.7, 27.1, 27.2, 30.0, 30.1, 30.2,
30.6, 33.0, 33.1, 34.5, 34.7, 38.8, 51.6 (–COOCH3), [74.1, 74.8 (–CH–OH)], [81.2, 83.6 (–HC(O)–CH–OH)],
174.1 (–COOCH3). 13C NMR (100 MHz, CDCl3, δC): 14.1 (CH3CH2–), 22.7, 25.0, 25.4, 25.6, 26.1, 26.3, 29.0,
29.1, 29.2, 29.2, 29.6, 32.0, 32.1, 33.4, 34.2, 38.1, 51.4 (–COOCH3), [73.5, 74.1 (–CH–OH)], [80.2, 82.5
(–HC(O)–CH–OH)], 174.1 (–COOCH3).
aAbbreviation: RT, retention time.
those obtained from commercial preparations of methyl ( )- MS spectra of silylated Peak I and Peak II are at m/z 489 (M
threo- and erythro-9,10-dihydroxyoctadecanoate shows that + H+). The indicated M.W. are consistent with the formation
the threo isomer 2 was formed exclusively. Thus, anti addi- of a mixture of dihydroxytetrahydrofuran derivatives, methyl
tion occurred during oxirane ring opening. The time course 9(12)-oxy-10,13-dihydroxystearate 4, and methyl 10(13)-
of hydrolysis showed that nearly complete hydrolysis of the oxy-9,12-dihydroxystearate 5, derived from intramolecular
starting epoxide was obtained in less than 4 h.
cyclization (Scheme 1). EI-MS of the free and silylated de-
Methyl 9,10-12,13-diepoxyoctadecanoate 3 was prepared rivatives of both products shows that two regioisomers are
from methyl linoleate (Scheme 1). Diepoxide 3 was treated present in both HPLC fractions. In particular, in the silylated
with alumina, and the progress of the reaction was followed derivatives, the intense peak at m/z 259 (86.8%) clearly shows
by RP-HPLC. The peak corresponding to the diepoxide 3 (RT the presence of a hydroxyl group at C-9, whereas the intense
18.4 min) disappeared, and two product peaks arose at RT 5.9 peak at m/z 173 (100%) shows the presence of a hydroxyl
min (Peak I) and 9.4 min (Peak II). APCI-MS spectra of both group at C-13. The two product fractions were separated by
products were nearly identical, as were the spectra of their preparative TLC using a published method (9). Data from the
silylated derivatives (see the Materials and Methods section). 1H and 13C NMR spectra of the two fractions are shown in
The base peaks in the APCI-MS spectra for Peak I and Peak Tables 1 and 2. They indicate that both product fractions
II are at m/z 345 (M + H+), and the base peaks in the APCI- contained complex mixtures of regio- and stereoisomers. Our
TABLE 2
1H NMR and 13C NMR of Methyl 9(12)-oxy-10,13-dihydroxystearate and Methyl
10(13)-oxy-9,12-dihydroxystearate (peak II, RT on RP-HPLC: 9.5 min)a
Chemical shift
Number
Coupling
(400 MHz, C6D6, δH)
of protons
Appearance
Assignment
constant (Hz)
1.14
3H
22H
2H
1H
3H
1H
1H
m
m
t
m
s
CH3CH2–
–CH2–
–CH2COOCH3
–CH–OH
–COOCH3
–CH–OH
–HC(O)–CH–OH
(H-13 or H-9)
–HC(O)–CH–OH
(H-10 or H-12)
1.40–2.2
2.34 and 2.36
3.41
3.61
3.67
J = 7.4, J = 7.2
dt
d
J = 2.8, J = 6.8
J = 9.6
3.93
4.09
1H
m
13C NMR (100 MHz, C6D6, δC): 14.9 (CH3CH2–), 23.7, 25.9, 26.7, 26.9, 27.2, 27.3, 30.0, 30.1, 30.2, 30.3, 30.7,
32.8, 33.2, 34.8, 35.3, 39.5, 51.6 (–COOCH3), [72.5, 74.4 (–CH–OH)], [79.9, 80.0, 85.2 (–HC(O)–CH–OH)],
174.1 (–COOCH3). 13C NMR (100 MHz, CDCl3, δC): 14.0 (CH3CH2–), 22.6, 24.8, 24.9, 25.7, 25.9, 25.9, 26.1,
28.8, 29.0, 29.1, 29.3, 29.6, 31.7, 32.0, 34.1, 34.4, 38.8, 51.4 (–COOCH3), [71.6, 73.9, 74.0 (–CH–OH)], [79.1,
84.3, 84.3 (–HC(O)–CH–OH)], 174.3 (–COOCH3).
aSee Table 1 for abbreviation.
JAOCS, Vol. 80, no. 9 (2003)

904
G.J. PIAZZA ET AL.
sis by the alumina support should be considered as a contribut-
ing factor. Hydrolyzed fatty epoxides are useful for a variety
of applications such as polymers, high-temperature greases,
and boundary lubricants for aluminum metal working (14). In
those cases where hydrolyzed material is the desired product,
the most efficient process would be to conduct epoxidation in
the presence of alumina, and in this way obtain the hydrolyzed
material directly without isolating the intermediate epoxide.
ACKNOWLEDGMENTS
C. Frank Fox and Janine N. Brouillette provided technical assistance.
REFERENCES
1. Sheldon, R., and H. van Bekkum, Introduction, in Fine Chemi-
cals Through Heterogeneous Catalysis, edited by R.A. Sheldon
and H. van Bekkum, Wiley-VCH, Weinheim, 2001, pp. 1–11.
2. Lutz, J.T., Jr., Epoxidation, in Encyclopedia of Chemical Tech-
nology, Vol. 9, edited by M. Grayson, John Wiley & Sons, New
York, 1980, pp. 251–266.
3. Mandelli, D., M.C.A. van Vliet, U. Arnold, R.A. Sheldon, and
U. Schuchardt, Epoxidation of Alkenes with Hydrogen Perox-
ide Catalyzed by ReO4–SiO2·Al2O3 and ReO4–Al2O3, J. Mol.
Cat. A:Chem. 168:165–171 (2001).
4. Salavati-Niasari, M., F. Farzaneh, and M. Ghandi, Oxidation of
Cyclohexene with tert-Butylhydroperoxide and Hydrogen Per-
oxide Catalyzed by Alumina-supported Manganese(II) Com-
plexes, J. Mol. Cat. A: Chem. 186:101–107 (2002).
5. Posner, G.H., Organic Reactions at Alumina Surfaces, Angew.
Chem. Int. Ed. Engl. 17:487–496 (1978).
FIG. 1. Time course for the hydrolysis of methyl 9,10-12,13-diepoxyoc-
tadecanoate 3 (5 mg) with 230 mg alumina in 0.6 mL cyclohexane (●)
and 1% perchloric acid in 0.6 mL THF/H2O (3:2, vol/vol) (●).
results are similar to those obtained by Hammock’s group
(10,11) when they investigated products obtained from fatty
diepoxide hydrolysis using low levels of epoxide hydrolase;
they reported that the fraction with a short RT in RP-HPLC
contained side chains that were in the trans configuration
across the THF ring, whereas the fraction with a longer RT
contained side chains in the cis configuration. Weber et al.
(12) also have reported on the synthesis of methyl dihydroxy-
tetrahydrofuran octadecanoates from the treatment of epox-
ides of plant oils and epoxides of methylated fatty esters using
sulfuric acid-activated bleaching earth.
6. Harrak, Y., and M.D. Pujol, Mild Cleavage of Aliphatic Epox-
ides with Substituted Anilines on Alumina, Tetrahedron Lett.
43:819–822 (2002).
7. Piazza, G.J., A. Nuñez, and T.A. Foglia, Epoxidation of Fatty
Acids, Fatty Methyl Esters, and Alkenes by Immobilized Oat
Seed Peroxygenase, J. Mol. Cat. B: Enzymat. 21:143–151
(2003).
8. Piazza, G.J., T.A. Foglia, and A. Nuñez, Preparation of Fatty
Epoxy Alcohols Using Oat Seed Peroxygenase in Nonaqueous
Media, J. Am. Oil Chem. Soc. 76:551–555 (1999).
9. Bilyk, A., G.J. Piazza, R.G. Bistline, Jr. and M.J. Haas, Separa-
tion of Cholesterol, and Fatty Acylglycerols, Acids and Amides
by Thin-Layer Chromatography, Lipids 26:405–406 (1991).
10. Halarnkar, P.P., J. Nourooz-Zadeh, E. Kuwano, A.D. Jones, and
B.D. Hammock, Formation of Cyclic Products from the Diepox-
ide of Long-chain Fatty Esters by Cytosolic Epoxide Hydrolase,
Arch. Biochem. Biophys. 294:586–593 (1992).
11. Nourooz-Zedeh, J., T. Uematsu, B. Borhan, M.J. Kurth, and
B.D. Hammock, Characterization of the Cytosolic Epoxide Hy-
drolase-Catalyzed Hydration Products from 9,10:12,13-Diepoxy
Stearic Esters, Ibid. 294:675–685 (1992).
12. Weber, N., K. Vosmann, E. Fehling, K.D. Mukherjee, and D.
Bergenthal, Analysis of Hydroxylated Fatty Acids from Plant
Oils, J. Am. Oil Chem. Soc. 72:361–368 (1995).
13. Lambert J.-F., and M. Che, The Molecular Approach to Sup-
ported Catalysts Synthesis: State of the Art and Future Chal-
lenges, J. Mol. Cat. A: Chem. 162:5–18 (2000).
14. Hotten, B.W., Bidentate Organic Oxygen Compounds as
Boundary Lubricants for Aluminum, Lubr. Eng. 30:398–403
(1974).
Figure 1 shows the time course of hydrolysis of methyl
9,10-12,13-diepoxyoctadecanoate 3 with alumina and with a
low concentration (1%) of perchloric acid (HClO4). The con-
centration of perchloric acid was kept low to eliminate ester
hydrolysis during the study period. One can observe in Fig-
ure 1 that alumina hydrolysis was complete in 1.5 h, whereas
perchloric acid hydrolysis took somewhat longer. The prod-
ucts of perchloric acid hydrolysis were identical to the prod-
ucts obtained from alumina hydrolysis.
Our results show that unmodified alumina catalyzes the hy-
drolysis of epoxides, and therefore its use as a support for
epoxidation catalysts should be avoided. At this time, the
structure of the surface of alumina has not been completely
described (13). However, it is known that the surface contains
free hydroxyl groups, which may act as nucleophiles for the
promotion of hydrolysis. Posner (5) speculated that rate accel-
eration by alumina for nucleophile addition is provided by an
entropy of activation when a reactant and a reagent are ad-
sorbed close to each other. Also, the alumina may activate the
epoxide by forming a complex with aluminum atoms. In those
studies of transition metal epoxidation catalysts that have used
alumina as a catalyst support, an appreciable amount of the
product was found as diols, which are the hydrolyzed epoxide
products. The hydrolysis has been ascribed to the acidity of
the catalyst, but the results presented here show that hydroly-
[Received February 3, 2003; accepted May 20, 2003]
JAOCS, Vol. 80, no. 9 (2003)