Efficient oxidative cleavage of olefins to carboxylic acids with hydrogen peroxide catalyzed by methyltrioctylammonium tetrakis(oxodiperoxotungsto)phosphate(3-) under two-phase conditions. Synthetic aspects and investigation of the reaction course

Article abstract of DOI:10.1021/jo980481t

The oxidative cleavage of alkenes to carboxylic acids with 40% w/v aqueous hydrogen peroxide catalyzed by methyltrioctylammonium tetrakis(oxodiperoxotungsto)phosphate(3-) (1a) is reported to occur in high yields and selectivities under two-phase conditions in the absence of organic solvents. On the basis of a study of the reaction, two main reaction pathways leading to acids have been recognized, the first one involving the perhydrolysis and the second one the hydrolysis of the epoxide initially formed. The "perhydrolytic" reaction pathway appears to play a primary role in the oxidation of medium- and long-chain alkenes to acids, while it intervenes to a rather limited extent in the oxidation of arylalkenes and C₅-C₇ cycloalkenes. The occurrence of this pathway has been proved by the isolation of the intermediate β-hydroperoxy alcohols and their transformation into acids with H₂O₂ and la. The course of this transformation, involving an initial oxidation (to α-oxo hydroperoxide) or decomposition (to carbonyl compounds) of the β-hydroperoxy alcohol intermediate, is described. The primary oxidation products, α-hydroperoxy ketones, have been isolated in the case of internal β-hydroperoxy alcohols, whereas their presence has been evidenced with terminal β-hydroperoxy alcohols bearing a secondary hydroxy group. Hydrogen peroxide concentration appears to exert a remarkable influence on medium acidity, and its effects on the reaction efficiency are shown.

Full text of DOI:10.1021/jo980481t

7190
J . Org. Chem. 1998, 63, 7190-7206
Efficien t Oxid a tive Clea va ge of Olefin s to Ca r boxylic Acid s w ith
Hyd r ogen P er oxid e Ca ta lyzed by Meth yltr ioctyla m m on iu m
Tetr a k is(oxod ip er oxotu n gsto)p h osp h a te(3-) u n d er Tw o-P h a se
Con d ition s. Syn th etic Asp ects a n d In vestiga tion of th e
Rea ction Cou r se
Ermanno Antonelli, Rino D’Aloisio, Mario Gambaro, Tiziana Fiorani, and Carlo Venturello*
EniChem S.p.A., Centro Ricerche Novara Istituto G. Donegani, via G. Fauser 4, 28100 Novara, Italy
Received March 13, 1998
The oxidative cleavage of alkenes to carboxylic acids with 40% w/v aqueous hydrogen peroxide
catalyzed by methyltrioctylammonium tetrakis(oxodiperoxotungsto)phosphate(3-) (Ia ) is reported
to occur in high yields and selectivities under two-phase conditions in the absence of organic solvents.
On the basis of a study of the reaction, two main reaction pathways leading to acids have been
recognized, the first one involving the perhydrolysis and the second one the hydrolysis of the epoxide
initially formed. The “perhydrolytic” reaction pathway appears to play a primary role in the
oxidation of medium- and long-chain alkenes to acids, while it intervenes to a rather limited extent
in the oxidation of arylalkenes and C5-C7 cycloalkenes. The occurrence of this pathway has been
proved by the isolation of the intermediate â-hydroperoxy alcohols and their transformation into
acids with H2O2 and Ia . The course of this transformation, involving an initial oxidation (to R-oxo
hydroperoxide) or decomposition (to carbonyl compounds) of the â-hydroperoxy alcohol intermediate,
is described. The primary oxidation products, R-hydroperoxy ketones, have been isolated in the
case of internal â-hydroperoxy alcohols, whereas their presence has been evidenced with terminal
â-hydroperoxy alcohols bearing a secondary hydroxy group. Hydrogen peroxide concentration
appears to exert a remarkable influence on medium acidity, and its effects on the reaction efficiency
are shown.
In tr od u ction
with the oxidation of organic substrates.1b However,
there are few reports concerning the olefinic double bond
cleavage reaction using these oxidizing systems. They
mainly refer to the conversion of (cyclo)alkenes to (di)-
aldehydes in a nonaqueous medium.6 To our knowledge,
in the past decade only two papers described the use of
aqueous hydrogen peroxide in the presence of tris-
(cetylpyridinium) 12-tungstophosphate7 or tungstic acid8
catalysts under homogeneous conditions (t-BuOH as the
solvent) for the production of carboxylic acids from
alkenes. These procedures, however, are of limited
practical value, as they require very long reaction times
(24 h at 80 °C) and afford moderate-to-low yields of acids
with R-olefins.
The oxidative cleavage of alkenes to produce carboxylic
acids is a transformation widely used for preparative
purposes.1a It is commonly achieved by ozonolysis with
oxidative workup2 or by using oxometal reagents such
as permanganate3 or ruthenium tetroxide,4 the latter
preferably employed in catalytic amounts in combination
with a number of oxygen donors (NaOCl,5a-c NaIO4,5d,e
or peracetic acid5f,g) as primary oxidants.
The early-transition-metal/hydrogen peroxide systems
have attracted much interest in recent years in context
(1) Sheldon, R. A.; Kochi, J . K. Metal-Catalyzed Oxidation of Organic
Compounds; Academic: New York, 1981; (a) Chapter 9, pp 297-299;
(b) Chapter 3, pp 48-56; Chapter 9, pp 286-288; Chapter 12, pp 366-
368; Chapter 13, pp 387-393; (c) Chapter 3, pp 51-52.
(2) (a) Augustine, R. L., Ed. Oxidation; Dekker: New York, 1969;
Vol. 1, p 259. (b) Bayley, P. S. Ozonation in Organic Chemistry;
Academic Press: New York, 1978; Vol. 1.
(3) (a) Starks, C. M. J . Am. Chem. Soc. 1971, 93, 195. (b) Sam, D.
J .; Simmons, H. E. J . Am. Chem. Soc. 1972, 94, 4024. (c) Krapcho, A.
P.; Larson, J . R.; Eldridge, J . M. J . Org. Chem. 1977, 42, 3749. (d)
Lee, D. G.; Chang, V. S. J . Org. Chem. 1978, 43, 1532.
Our earlier studies9,10 on the tungstate-phosphate
(arsenate)-hydrogen peroxide system under acidic condi-
tions led us to isolate a new class of peroxotungsten
complexes with remarkable oxidizing properties, which
were identified as quaternary ammonium (phosphonium)
tetrakis(oxodiperoxotungsto)phosphates or -arsenates (3-)
(I).11,12 Indeed, when used in conjuction with hydrogen
(4) Lee, D. G.; van der Engh, M. In Oxidation in Organic Chemistry;
Trahanovsky, W. S., Ed.; Academic Press: New York, 1973; Part B, p
186.
(6) (a) Waldemann, H.; Schwerdtel, W.; Swodenk, W. Ger. Offen.
2,201,456, 1973 and 2,252,674, 1974 to Bayer A.-G.; Chem. Abstr. 1973,
79, 91679 and 1974, 81, 25083. (b) Nippon Oil Co. J pn. Kokai Tokkyo
Koho 82 95,921, 1982; Chem. Abstr. 1982, 97, 162366.
(7) Ishii, Y.; Yamawaki, K.; Ura, T.; Yamada, H.; Yoshida, T.;
Ogawa, M. J . Org. Chem. 1988, 53, 3587.
(8) Oguchi, T.; Ura, T.; Ishii, Y.; Ogawa, M. Chem. Lett. 1989, 857.
(9) (a) Venturello, C.; Alneri, E.; Lana, G. Ger. Offen. 3,027,349, 1981
(It. Appl. 24478/79); Chem. Abstr. 1981, 95, 42876. (b) Venturello, C.;
Alneri, E.; Ricci, M. J . Org. Chem. 1983, 48, 3831.
(10) (a) Venturello, C.; Ricci, M. Eur. Pat. Appl. 123,495, 1984 (It.
Appl. 20605/83); Chem. Abstr. 1985, 102, 78375. (b) Venturello, C.;
Ricci, M. J . Org. Chem. 1986, 51, 1599.
(5) (a) Kleybs, K. A.; Dubeck, M. U.S. Patent 3,409,649, 1968 to
Ethyl Corp.; Chem. Abstr. 1969, 70, 114575. (b) Wolfe, S.; Hasan, S.
K.; Campbell, J . R. J . Chem. Soc., Chem. Commun. 1970, 1420. (c)
Foglia, T. A.; Barr, P. A.; Malloy, A. J .; Costanzo, M. J . J . Am. Oil
Chem. Soc. 1977, 54, 870A. (d) Starks, C. M.; Washecheck, P. H. U.S.
Patent 3,547,962, 1970 to Continental Oil Co.; Chem. Abstr. 1971, 74,
140895. (e) Carlsen, H. J .; Katsuki, T.; Martin, V. S.; Sharpless, K. B.
J . Org. Chem. 1981, 46, 3936. (f) Washecheck, P. H. Ger. Offen. 2,046,-
034, 1971 to Continental Oil Co.; Chem. Abstr. 1971, 75, 35158. (g)
Merck, W.; Schreyer, G.; Weigert, W. Ger. Offen. 2,106,307, 1972 to
Degussa; Chem. Abstr. 1972, 77, 151485.
S0022-3263(98)00481-2 CCC: $15.00 © 1998 American Chemical Society
Published on Web 09/22/1998

Oxidative Cleavage of Olefins to Carboxylic Acids
J . Org. Chem., Vol. 63, No. 21, 1998 7191
isolated) carboxylic acids in good-to-high yields. Thus,
the commercially available even-numbered linear chain
R-olefins 1-octene, -decene, and -hexadecene were easily
converted into the respective odd-numbered alkanoic
acids of one less carbon in 80% yields (entries 1, 2, and
6). Likewise, styrene gave an 87% yield of benzoic acid
(entry 3). The oxidation of the unsymmetrical internal
olefins trans-2-octene and trans-4-decene (entries 4 and
5) produced hexanoic acid in 81% and 78% yields,
together with acetic acid (78%) and butanoic acid (80%),
respectively, while the oxidation of trans-5-decene gave
a 77% yield of pentanoic acid (entry 7). In a similar
manner, on oxidation of oleic acid (entry 8), pelargonic
(nonanoic) acid and azelaic (nonanedioic) acid were
obtained in 82% and 79% yields, respectively.
Cyclopentene, cyclohexene, and 1-methylcyclohexene
were cleaved readily under the reaction conditions, with
the first two substrates affording an 87% yield of glutaric
and adipic acid, respectively (entries 9 and 10), and the
third one an 85% yield of 6-oxoheptanoic acid (entry 11).
The oxidation of cycloheptene (entry 12) was somewhat
less satisfactory, with pimelic (heptanedioic) acid being
obtained in 74% yield. In this context, it is of interest
that unsaturated medium-sized rings (e.g., eight- and
ten-membered) gave poor or insignificant amounts of the
expected diacids. Thus, oxidation of cis-cyclooctene (entry
13) gave only 24% of suberic (octanedioic) acid, together
with minor amounts of other acidic components (for
details, see Table 1, footnote k). In the case of cis-
cyclodecene (entry 14), the yield of sebacic (decanedioic)
acid dropped to less than 1%, while 5-ketosebacic acid
(25%), together with its semialdehyde (6,10-dioxodecanoic
acid, 5-6%), was the major acidic product (for other acidic
components formed, see Table 1, footnote m). With both
substrates, several neutral products (mostly not further
investigated) were formed along with significant amounts
of heavy materials. 1,4-Cyclooctanedione was found to
be the major GC component of the neutral fraction in the
oxidation of cis-cyclooctene.
Diolefins also underwent oxidative cleavage, as il-
lustrated in Table 1 (entry 15) by the oxidation of 1,7-
octadiene. In this case, however, the isolated yield of the
expected adipic acid was considerably lower (60%) than
that obtained from the oxidation of cyclohexene.
It is noteworthy that the overoxidation product result-
ing from a loss of one carbon was generally formed (<5%)
together with the expected acid. In several cases (see
Table 1, footnote c), the acid of two less carbons was also
found.
II. Th e Cou r se of th e Rea ction . A study of the
reaction was undertaken to probe into the nature of the
oxidation pathway which leads to carboxylic acids from
olefins. Given the peroxidic character of catalyst Ia and
its known11-13 epoxidizing ability, the epoxidation of the
alkene had to be regarded as the early stage of the
reaction.
peroxide under two-phase conditions, complexes I, espe-
cially Ia , were found to promote numerous important
reactions, such as the epoxidation of simple, nonactivated
alkenes,11-13 as well as their direct dihydroxylation (via
epoxidation),14,15 the ketonization of secondary alco-
hols,16,17 and the oxidation of primary alcohols and
aldehydes to carboxylic acids.17
The list of the oxidative processes achieved with the
I/H2O2 system (and patented) also included the fission
of olefinic double bonds into carboxylic acids.18 In this
case, however, results were only partially satisfactory
from the synthetic point of view. In fact, while good
yields of acids could be obtained with substrates such as
styrene and cyclohexene, with acyclic olefins the reaction
took place with increasing difficulty as the chain length
increased beyond eight C atoms, even at a high catalyst-
to-substrate molar ratio (ca. 1:30).
Later on, we had the opportunity to reexamine the
aforementioned two-phase approach to olefin-to-carboxy-
lic acid oxidation in more detail. To our surprise, we
found that, on excluding the solvent (an aromatic or
chlorinated aliphatic hydrocarbon) previously employed
in our patent work,18 the reaction gave significantly
improved results with medium- and mostly long-chain
alkenes, even at a lower catalyst-to-substrate molar ratio.
On the basis of this finding, we could develop an efficient,
general procedure for the two-phase oxidation of olefins
to carboxylic acids with hydrogen peroxide catalyzed by
the very active complex Ia .
In the present article, we report the full details of the
synthetic method and the results of a complete investiga-
tion of the reaction course, which also enable us to
accommodate the new aspects observed in a general
mechanistic scheme.
Resu lts a n d Discu ssion
I. Syn th esis of Ca r boxylic Acid s fr om Olefin s
w ith Ia /H2O2. The oxidation of alkenes to carboxylic
acids is carried out in the absence of the solvent by
vigorously stirring a biphase mixture made up of the
appropriate olefin, catalyst Ia , and 40% w/v aqueous
hydrogen peroxide at 85 °C for 5-7 h. Catalyst/substrate
molar ratios of 1-1.2:100 and, usually, a 10% molar
excess of oxidant over the stoichiometric amount re-
quired19 are employed. Pertinent results are summarized
in Table 1.
As shown, a number of olefins are selectively cleaved
by the present method to give reasonably pure (when
In agreement with this view, the aforementioned7,8
tungsten-promoted oxidations of alkenes to acids with
aqueous hydrogen peroxide under homogeneous condi-
tions were proposed to follow a course which, apart from
the catalyst and reaction medium used, corresponds to
(11) Venturello, C.; D’Aloisio, R.; Ricci, M. Eur. Pat. Appl. 109,273,
1984 (It. Appl. 24154A/82); Chem. Abstr. 1984, 101, 191668.
(12) Venturello, C.; D’Aloisio, R.; Bart, J . C. J .; Ricci, M. J . Mol.
Catal. 1985, 32, 107.
(13) Venturello, C.; D’Aloisio, R. J . Org. Chem. 1988, 53, 1553.
(14) Venturello, C.; Gambaro, M. Eur. Pat. Appl. 146,374, 1985 (It.
Appl. 24203A/83); Chem. Abstr. 1985, 103, 195855.
(15) Venturello, C.; Gambaro, M. Synthesis 1989, 295.
(16) Venturello, C.; Gambaro, M.; Ricci, M. Eur. Pat. Appl. 232,-
742, 1987 (It. Appl. 19098A/86); Chem. Abstr. 1988, 108, 131028.
(17) Venturello, C.; Gambaro, M. J . Org. Chem. 1991, 56, 5924.
(18) Venturello, C.; Ricci, M. Eur. Pat. Appl. 122,804, 1984 (It. Appl.
20604A/83); Chem. Abstr. 1985, 102, 95256.
(19) Oxidation of trisubstituted and internal disubstituted olefins
to acids obeys a 1:3 and 1:4 substrate-to-hydrogen peroxide stoichi-
ometry, respectively. The molar ratio required becomes 1:5 with
monosubstituted terminal olefins as the coproduced formic acid can
be further oxidized to carbon dioxide.

7192 J . Org. Chem., Vol. 63, No. 21, 1998
Antonelli et al.
Ta ble 1. Oxid a tive Clea va ge of Olefin s to Ca r boxylic Acid s w ith H2O2 Ca ta lyzed by
Tetr a k is(oxod ip er oxotu n gsto)p h osp h a te Ia a
olefin
(50 mmol)
Ia
(mmol)
H2O2
(mmol)
time
(h)
yieldb
(%)
purityc
(%)
entry
product(s)
1
2
3
4
1-octene
0.5
0.6
0.5
0.5
275
275
275
220
6
6
5
6
heptanoic acid
nonanoic acid
benzoic acid
hexanoic acid
acetic acid
80
80
87
95
95
97e
97
1-decene
styrened
trans-2-octene
81
78f
77g
80g
80
5
trans-4-decene
0.5
220
6
hexanoic acid
butanoic acid
6
7
8
1-hexadecene
trans-5-decene
oleic acid
0.6
0.5
0.6
275
220
275
6
6
5
pentadecanoic acid
pentanoic acid
nonanoic acid
nonanedioic acid
glutaric acid
95
96
77
82g
79g
87
9
10
11
12
13
14
cyclopentene
0.6
0.6
0.5
0.6
0.3
0.3
230
220
175
230
137.5
137.5
7
6
6
7
7
7
97
cyclohexened
adipic acid
87
85
74
99h
96i
92
1-methylcyclohexened
cycloheptene
6-oxoheptanoic acid
heptanedioic acid
octanedioic acid
decanedioic acid
5-oxodecanedioic acid
6,10-dioxodecanoic acid
adipic acid
cis-cyclooctenej
cis-cyclodecenej
24g,k
<1g
25g,l,m
5-6g,n
60
15
1,7-octadienej
0.6
250
7
99
a
All reactions were performed without added solvent at 85 °C using 40% w/v aq H2O2, in a 10-37.5% molar excess over the stoichiometric
amount required.19 Unless otherwise noted, the yield (based on substrate charged) refers to the isolated product which, where necessary,
b
was purified by either distillation (entries 1, 2, and 6), elution on column (entries 9, 11, and 12), or recrystallization from acetonitrile
(entry 15) (for details, see Experimental Section). c Determined by GC after silylation with BSTFA (at 60 °C) or methylation with CH2N2
(for monocarboxylic acids). Unless otherwise stated, the product was mainly contaminated by the next lower acid. In entries 1, 2, 6, and
d
12, it also showed 0.4-0.7% contamination by the overoxidation product of two less carbons. Water (5 mL) was introduced into the
reactor together with the olefin and catalyst to achieve better control of the initial exothermicity during the hydrogen peroxide addition.
e Phenylacetic acid was the major byproduct. f Yield determined by GC (n-PrOH as an internal standard) on the combined aqueous solutions
(original aqueous layer, mother liquor from acid extraction, and washing waters; see Experimental Section) after dilution with acetone.
g
Yield determined by GC (internal standard) on the organic layer as such (entries 5 and 8) or the acid fraction (entries 13 and 14) after
h
silylation (for details, see Experimental Section). Refers to the major crop (containing 0.3% of inorganic residue). The minor crop (containing
i
1% of inorganic residue) was obtained 96% pure (see Experimental Section). In addition to the overoxidation product 5-oxohexanoic acid
(0.5%),89 additional contaminants were glutaric acid (0.4%), adipic acid (0.3%),90 and δ-acetoxypentanoic acid (1%) (GC-MS (as methyl
ester) m/z (CI) 174 (MH+, 100) (EI) M+ absent, 131, 114, 101, 59, 55, 43 (100)). In control experiments, they were proved to arise from
j
k
oxidation of the 6-oxoheptanoic acid. Reaction performed on 25 mmol of substrate. Minor acidic products obtained were adipic acid
(7%), heptanedioic acid (2.6%), 6-oxoheptanoic acid (4%), 4-oxooctanedioic acid (3%)91 and 4-hydroxyoctanedioic acid γ-lactone (2%) (GC-
MS (as methyl ester) m/z (CI) 187 (MH+, 100) (EI) M+ absent, 155, 126, 85 (100), 74, 59, 55). Could be isolated 95% pure from the acid
l
m
fraction by elution on column (2.8-cm diameter) of SiO2 (100 g; CHCl3/AcOH, 9.5:0.5; Rf 0.35). 4-Hydroxydecanedioic acid γ-lactone (3%)
was also formed (along with some succinic, glutaric, and adipic acids). GC-MS (as methyl ester) m/z (CI) 215 (MH+, 100) (EI) M+ absent,
183, 154, 141, 87, 85 (100), 74, 59, 55, 41. GC-MS (as methyl ester) m/z (CI) 215 (MH+, 100), (EI) 214 (M+, 0.2), 196, 186, 164, 158, 143,
n
99, 71, 55, 43, 42, 41 (100). Was converted to 5-oxodecanedioic acid on further oxidation with Ia /H2O2.
path b of Scheme 1. (This Scheme, which for the sake of
simplicity is limited to mono- and disubstituted olefins,
RCHdCHR1 where R ) alkyl, aryl and R1 ) H, alkyl,
will henceforth be referred to for the numbering of
reactants and products.)
The reaction is believed to involve the formation of a
1,2-diol intermediate 3 upon acid hydrolysis of epoxide
2, initially formed from olefin 1. Diol 3 is then oxidized
to R-hydroxy ketone (R-ketol) 4, which undergoes C-C
bond fission by hydrogen peroxide20 leading ultimately
to acids 8, according to the reaction sequence previously
reported by us10b for the tungsten-catalyzed oxidation of
vic-diols to carboxylic acids with hydrogen peroxide under
acidic conditions.
In principle, a similar pathway (path b of Scheme 1,
henceforth referred to as the “hydrolytic” reaction path-
way) might be plausible also in our two-phase process
for olefins whose epoxides are easily opened by acids to
yield diols (e.g., styrene and unsaturated five- to seven-
membered rings). Indeed, when put into contact with
water in a biphase system, catalyst Ia as a heteropoly-
metalate (in a peroxidic form) undergoes a partial
degradation with the generation of enough acidity in the
aqueous phase (determined as phosphoric and tungstic
acid) to promote hydrolysis of the cited epoxides (see
Experimental Section). However, this pathway as such
seemed inadequate to us to account for the observed
reactivity of medium- and long-chain alkenes (gC8) in
view of the fairly good stability to hydrolysis exhibited
by the corresponding epoxides under the same acidity
conditions.
(20) (a) Recently, Ishii et al.20b reported the two-phase oxidation of
internal 1,2-diols to R-diketones, via R-ketols, with 35% w/w H2O2 and
complex [π-C5H5N(CH2)15CH3]3{PO4[W(O)(O2)2]4} (PCWP) as the cata-
lyst using 1,2-dichloroethane (4 L per mol of substrate) as the solvent.
These authors also reported that R-diketones are cleaved to carboxylic
acids by H2O2 under these conditions. Accordingly, it was stated that
conversion of internal 1,2-diols into acids with PCWP/H2O2 takes place
through the formation of the corresponding R-diketones under biphasic
conditions. However, we have found that, under the two-phase condi-
tions adopted (lack of the solvent), even with internal 1,2-diols the
formation of carboxylic acids appears to proceed mainly through acid-
catalyzed oxidative cleavage of the intermediate R-ketols by H2O2. This
was proved by monitoring (by GC) the oxidation (60 °C) of 6-hydroxy-
5-decanone (4d ) (1 mmol) with Ia (0.01 mmol)/40% w/v H2O2 (3 mmol)
at low conversion (18%, 1 h). Pentanal (7d ), which forms in equimolar
amounts with pentanoic acid (8d ) from the oxidative cleavage of 4d ,
was found to be present in the reaction mixture in a molar ratio of
1:1.75 to 8d and of at least 20:1 to 5,6-decanedione. (b) Iwahama, T.;
Sakaguchi, S.; Nishiyama, Y.; Ishii, Y. Tetrahedron Lett. 1995, 36,
1523.
For this reason, we deemed it proper to address our
investigation primarily to this group of alkenes to
ascertain whether other reaction pathways leading to
carboxylic acids from the first-formed epoxides were at
work.

Oxidative Cleavage of Olefins to Carboxylic Acids
J . Org. Chem., Vol. 63, No. 21, 1998 7193
Sch em e 1a
a
â-Hydroxy aldehydes 4(I; R1 ) H) could not be detected (see ref 24). bCompound 5IIf was not formed (see text). cCompounds 6I
(R1 ) H) could not be detected (see text).
A. Med iu m - a n d Lon g-Ch a in Alk en es a s Su b-
str a tes. In conformity with above, we explored the Ia -
catalyzed oxidation of some model epoxyalkanes under
the reaction conditions (85 °C, molar ratio H2O2/epoxide
) 5:1) at low oxidant conversion, with a view to detecting
relevant intermediates. 1,2-Epoxyoctane (2a ) and -de-
cane (2b), among the terminal epoxyalkanes, and trans-
4,5- and -5,6-epoxydecane (2c and 2d ), among the inter-
nal ones, were the epoxides of choice.
3a and 3b, indicative of the hydrolytic reaction pathway
(Scheme 1, path b), were also found to form in consider-
able amounts (25% and 16%, respectively), together with
their primary oxidation products R-ketols 4IIa (9.5%) and
4IIb (7.5%).24 This result will be considered again for
discussion in section IVA.
After silylation with N,O-bis(trimethylsilyl)trifluoro-
acetamide (BSTFA), the peak of the aldehyde decreased
dramatically while two new major components (43%, on
the whole, in the first case and 50% in the second one),
in ratios of approximately 2.7:1 (for 2a ) and 2:1 (for 2b),
appeared neatly in GC.25 The characteristics of their
GC-mass spectra (i.e., the presence of two silyl groups
and the same molecular weight, 16 units higher than that
of the corresponding diol) suggested the formation of
regioisomeric C8 and C10 â-hydroperoxy alcohols.
(a ) In ter r u p ted Oxid a tion of Ep oxya lk a n es w ith
Ia /H2O2. F or m a tion of â-Hyd r op er oxy Alcoh ols. The
terminal epoxides 2a and 2b were examined first. In
both cases, when the reaction was stopped (after 30 min,21
with the epoxide conversion being almost quantitative
for 2a and 92-93% for 2b), GC analysis showed a
substantial presence of the aldehyde of one less carbon,
7a (25%) and 7b (33%),22 which was accompanied by a
small amount of the related final acid, 8a (4%) and 8b
(2.3%).23 Despite the aforementioned stability to hy-
drolysis of epoxides 2a ,b under the acidity conditions
generated by catalyst Ia when put into contact with
water in a two-phase system, the corresponding 1,2-diols
(24) R-Hydroxy aldehydes 4Ia and 4Ib could not be detected.
However, small amounts of the corresponding aldehydes of one less
carbon, suggesting an early formation of the former, were found (along
with R-ketols 4IIa and 4IIb) when diols 3a ,b were oxidized with Ia /
H2O2 under two-phase conditions at low conversion using CHCl3 as
the solvent (to minimize further oxidation of the R-ketol; cf. ref 20).
Thus, on reacting 1,2-decanediol (3b) (5 mmol) in CHCl3 (25 mL) for
5 h at reflux with Ia (0.05 mmol)/40% w/v H2O2 (25 mmol), a molar
ratio of 1-hydroxy-2-decanone to nonanal of about 11:1 was found (by
GC) at 22% conversion of the diol. Nonanoic acid was formed only in
negligible amounts (molar ratio C9 acid to C9 aldehyde ) 1:5).
(25) Other minor components were also detected by GC. Their GC-
mass spectra were consistent with the structures of â,â′-dihydroxyalkyl
(C8, C10) peroxides and ethers. They presumably result from further
reaction of, respectively, the â-hydroperoxy alcohols (see ahead in the
text) and vic-diols previously formed with the respective starting
epoxides.
(21) Five to six minutes for dropping hydrogen peroxide should be
added (see Experimental Section).
(22) With the temperature of the injector set at 300 °C (carrier gas,
He). At 250 °C, the formation of the aldehyde decreased by more than
50%. This suggested the decomposition of a parent compound occurring
in the gas chromatograph (see ahead in the text).
(23) The coproduced formic acid was revealed by ionic chromatog-
raphy of the water phase.

7194 J . Org. Chem., Vol. 63, No. 21, 1998
Antonelli et al.
The two components (nonsilylated) were revealed in
TLC as a single spot. Their isolation by column chro-
matography gave an oily product, whose 1H and 13C NMR
spectra confirmed it to be a mixture of 2-hydroperoxy-1-
octanol (5Ia ) or -decanol (5Ib) (major component) and
1-hydroperoxy-2-octanol (5IIa ) or -decanol (5IIb), respec-
tively, in the ratios indicated above. Attempts to obtain
specimens of the individual components from the mixture
were unsuccessful. Nevertheless, the formed regioiso-
mers 5IIa and 5IIb were compared with pure samples
separately obtained by the addition of alkaline hydrogen
peroxide to the appropriate epoxides (see Experimental
Section).
and a lower temperature. NMR data of the isolated
products thus obtained confirmed the identities of 5-hy-
droperoxy-4-decanol (5Ic) and 4-hydroperoxy-5-decanol
(5IIc), present in about a 1:1 ratio, and of 6-hydroperoxy-
5-decanol (5d ). These compounds were assigned the
erythro configuration on the basis of their reduction with
triphenylphosphine to the corresponding vic-diols.
It should be noted that when the parent alkenes 1a -d
were subjected to interrupted-oxidation experiments with
Ia /H2O2 in place of epoxides 2a -d under identical
conditions, qualitatively analogous results were obtained,
thus confirming the intermediacy of the epoxide in the
early stage of the reaction beyond any reasonable doubt.
In these experiments, however, when the reaction was
stopped (with the olefin conversion reaching 90-99%),
the value of the product ratio of diol + ketol versus
â-hydroperoxy alcohol was in general higher than that
found from the corresponding epoxide. This suggests
that a competitive tungsten-catalyzed epoxidation of
alkene 1 by the formed â-hydroperoxy alcohol 5, with
concomitant production of diol 3,28 occurs in our process,
the extent of the reaction depending on the nature of the
substrate and related intermediates (epoxide and â-hy-
droperoxy alcohol). For the sake of clarity, this reaction
has been omitted in Scheme 1.
(b ) P er h yd r olyt ic ver su s H yd r olyt ic R ea ct ion
P a th w a y. When caused to react with Ia /H2O2 under the
reaction conditions, the isolated â-hydroperoxy alcohols
5a -d were readily converted into the corresponding
carboxylic acids 8a -d . Thus, a reaction pathway leading
to acids 8 from olefins 1 via â-hydroperoxy alcohols 5
(Scheme 1, path a) appears to be operative in our process
for the group of substrates under study as an alternative
to the pathway via 1,2-diols 3 (Scheme 1, path b).
Clearly, hydrogen peroxide is directly involved in this
pathway (henceforth referred to as “perhydrolytic” reac-
tion pathway), being responsible for the formation of 5
by nucleophilic attack at the oxirane carbon of 2.
It should be observed that the easy formation of 5 in
our process by perhydrolysis of 2 indicates that peroxo-
complex Ia does not intervene on 2 as an “oxidation”
catalyst (oxidative cleavage of the epoxide to carbonyl
compounds) to any appreciable extent. Although early-
transition-metal peroxocomplexes have been reported to
effect such a cleavage catalytically (with H2O2 as the
terminal oxidant)36 in addition to stoichiometrically,36,37
the reactivity of this class of complexes toward epoxides
(to give carbonyl compounds) according to what is ob-
served with Ia appears far lower than that exhibited by
hydrogen peroxide (to give â-hydroperoxy alcohols) under
the reaction conditions adopted.
When injected (as a 0.5-2% ethereal solution) into the
gas chromatograph, the isolated mixed 2(1)-hydroperoxy-
1(2)-octanols 5a and -decanols 5b decomposed, forming
the previously observed aldehydes.27 â-Hydroperoxy
alcohols are known to undergo thermal27 as well as acid-
29,30
and base-catalyzed26 decomposition to carbonyl com-
pounds.
The interrupted oxidation of the internal epoxyalkanes
trans-2c,d with Ia /H2O2 showed a similar trend. After
1 h,21 the corresponding â-hydroperoxy alcohols 5c (I +
II)31 and 5d were found to be formed in an approximately
1:1.3 molar ratio32 to the respective 1,2-diol33/R-ketol34
pairs (where the R-ketol was the major component),
together with appreciable amounts of aldehydes 7c and
7d and related acids (8c and 8d ).
As compared with the terminal epoxyalkanes 2a ,b, the
rate of disappearance of the epoxide (as followed by GC)
was in these cases considerably slower.35a For instance,
whereas conversion of 1,2-epoxydecane (2b) reached over
90% after 30 min,21 conversion of trans-4,5- and -5,6-
epoxydecane (2c and 2d ) was less than 30% after 1 h
under the same conditions. This circumstance, combined
with the fact that the resulting â-hydroperoxy alcohols
5c,d are quite reactive under the reaction conditions (see
section III), prevented an adequate accumulation of the
desired intermediates, rendering their isolation rather
laborious. The inconvenience, however, could be over-
come by the use of more concentrated H2O2 (70% w/w)
(26) Ogata, Y.; Sawaki, Y.; Shimizu, H. J . Org. Chem. 1978, 43,
1760.
(27) Small amounts of vic-diols 3a ,b and R-ketols 4IIa ,b were also
formed.
(28) Mattucci, A. M.; Perrotti, E.; Santambrogio, A. J . Chem. Soc.,
Chem. Commun. 1970, 1198.
(29) Adam, W.; Rios, A. J . Chem. Soc., Chem. Commun. 1971, 822.
(30) Subramanyam, V.; Brizuela, C. R.; Soloway, A. H. J . Chem. Soc.,
Chem. Commun. 1976, 508.
(31) As silyl derivatives, 5Ic and 5IIc eluted together in GC. The
presence of the two regioisomers, however, could be detected from the
analysis of the GC-mass spectrum (see Experimental Section).
(32) The found value of the molar ratio (in favor of the 1,2-diol/R-
ketol pair) is misleading because of the high reactivity of â-hydroperoxy
alcohols 5c,d under the reaction conditions (see section III).
(33) (a) The formed diols were identified as erythro-4,5-decanediol
and meso-5,6-decanediol, respectively, by comparison with authentic
samples.33b (b) Abe, Y.; Matsumura, S.; Shibuya, Y. Yukagaku 1979,
28(4), 276.
(34) A nearly 1:1 regioisomeric mixture of 5-hydroxy-4-decanone
(4Ic) and 4-hydroxy-5-decanone (4IIc) was obtained from the epoxide
trans-2c, as determined by GC and GC-MS (after silylation with
BSTFA). Regioisomer 4Ic eluted first.
(35) (a) Long-chain R-epoxides have been reported to undergo ring
opening by nucleophiles (e.g., MeOH) faster than the internal ones
under homogeneous conditions.35b As amphiphiles, in the present two-
phase system, the former should be regarded as being more suitably
oriented than the latter to contact their polar head with the nucleo-
philes water and hydrogen peroxide. This might contribute to their
much faster hydrolysis/perhydrolysis. (b) Bischoff, M.; Zeidler, U.;
Baumann, H. Fette, Seifen, Anstrichm. 1977, 79(3), 131.
Nevertheless, catalyst Ia is important in this stage of
the reaction, providing electrophilic assistance to the
epoxide opening by hydrogen peroxide/water mainly as
a Bro¨nsted acid through the acidic species that arise, as
already mentioned, from its partial hydrolytic degrada-
tion. Indeed, ring opening of epoxyalkanes 2a -d failed
when carried out in the presence of aqueous hydrogen
peroxide (neutral) only.
Of course, between the hydrolysis and perhydrolysis
of the epoxyalkane a competition exists, which under the
two-phase conditions adopted is governed by the greater
(36) Bonchio, M.; Conte, V.; Di Furia, F.; Modena, G. J . Mol. Catal.
1991, 70, 159.
(37) Za´honi-Budo´, E.; Sima´ndi, L. Inorg. Chim. Acta 1987, 134, 25.

Oxidative Cleavage of Olefins to Carboxylic Acids
J . Org. Chem., Vol. 63, No. 21, 1998 7195
or lesser lipophilic character of the epoxide involved and
the nature of the acid catalyst, in addition to the
concentration of the hydrogen peroxide used and the
relative reactivity of the two nucleophiles, hydrogen
peroxide and water, intervening in the oxirane opening
(see section IVA). Accordingly, the perhydrolytic path-
way (path a) appears to be the major route for the present
two-phase oxidation of medium- and long-chain alkenes
to carboxylic acids. It can be expected, however, that
with lower aliphatic alkenes (<C8) the alternative hy-
drolytic pathway (path b) operates to a larger extent.
B. Ar yla lk en es a n d Cycloa lk en es a s Su bstr a tes.
In light of the results attained, we wondered whether the
perhydrolytic reaction pathway might play some role also
with arylalkenes and C5-C7 cycloalkenes, whose oxida-
tion to acids we at first assumed to follow the hydrolytic
reaction pathway on the basis of the known good-to-high
susceptibility to acid hydrolysis of the respective ep-
oxides.
To answer this question, the representative cyclohex-
ene and styrene oxides, 2e and 2f, were subjected to
interrupted-oxidation experiments with Ia /H2O2, as had
previously been done with epoxyalkanes 2a -d . In both
cases when the reaction was stopped (after 10 min),21
together with the expected 1,2-diol/R-ketol pair (trans-
3e/4e and 3f/4IIf, respectively) as main products (66%
for 2e and 78% for 2f on the whole, where the 1,2-diol
was the major component), GC analysis after silylation
revealed the formation of the corresponding â-hydroper-
oxy alcohols 5e (16%) and 5If (5%),38 which were identi-
fied by GC-MS.
Operating at ambient temperature (instead of 85 °C),
these compounds could be obtained in somewhat higher
yields (21-22%), which made their isolation easier.
NMR data of the isolated products confirmed the identi-
ties of 2-hydroperoxycyclohexanol (5e) and 2-hydroper-
oxy-2-phenylethanol (5If). The former compound was
assigned the trans configuration on the basis of its
reduction to glycol with triphenylphosphine. It is note-
worthy that, at variance with the terminal epoxides 2a ,b,
from perhydrolysis of 2f only the regioisomer bearing the
hydroperoxy group on the most-substituted carbon atom
(“abnormal” product)39 was obtained. The selective for-
mation of the “abnormal” isomer from terminal epoxides
and hydrogen peroxide has been observed in neutral29
or acidic28,30 nonaqueous media.
As expected, compounds 5e,f, like the other â-hydro-
peroxy alcohols examined, were easily converted upon
treatment with Ia /H2O2 into the corresponding carboxylic
acids 8e,f. These results thus indicate that also for the
above alkenes the perhydrolytic reaction pathway is
operative in our process, although to a modest extent
compared to the hydrolytic one, which remains highly
favored.
the poor results obtained with unsaturated medium rings
(e.g., eight- and ten-membered) as contrasted with the
good-to-excellent results obtained with unsaturated com-
mon rings (five- to seven-membered) (cf. section I). This
difference is undoubtedly to be attributed to the anoma-
lous behavior of the related intermediate 1,2-epoxides in
the ring-opening step. Indeed, transannular reactions,
which are known40 to occur upon the opening of medium-
sized ring cycloalkene oxides under acidic conditions,
readily take place in the presence of the acidity provided
by the catalyst hydrolysis and seriously compete with or
prevail on the direct nucleophilic attack by hydrogen
peroxide or water, thus making the obtainment of the
desired acids difficult and highly unselective.
III. Th e P er h ydr olytic Reaction P ath way. Tr an s-
for m a tion of th e â-Hyd r op er oxy Alcoh ol In ter m e-
d ia tes in to Ca r boxylic Acid s. As already mentioned,
â-hydroperoxy alcohols 5 are susceptible to decomposition
with the formation of carbonyl compounds.28-30 On the
other hand, bearing a hydroxy function, they also would
be expected to undergo oxidation in the presence of Ia /
H2O2, since this system is known to oxidize alcohols to
carbonyl and carboxylic compounds.17 Relative to the
study of the perhydrolytic pathway of our process, we
deemed it interesting to ascertain whether these modes
of reaction were both operative and to determine the
extent of each in the conversion of intermediates 5 into
carboxylic acids 8 with Ia /H2O2, reported in section II
for the isolated compounds 5a -f.
In one case (initial oxidation of the â-hydroperoxy
alcohol), transformation 5 f 8 implies the formation of
the carbonyl derivative 6. Decomposition of 6 to the
aldehyde and carboxylic acid followed by oxidation of the
former then affords the final acids 8 (Scheme 1, path a,
way a′). In the other case (initial decomposition of the
â-hydroperoxy alcohol), transformation 5 f 8 proceeds
through the oxidation of the resultant aldehydes 7
(Scheme 1, path a, way a′′).
An investigation of the two ways, a′ and a′′, potentially
involved in the above transformation was undertaken by
monitoring (by GC or HPLC) the progress of the conver-
sion into acids 8 of some â-hydroperoxy alcohols 5
significantly representative of the perhydrolytic pathway
of our process. Accordingly, compounds 5Ib (isolated
admixed with 5IIb), 5IIb (obtained pure apart), and 5d
were utilized for this study, as models of terminal (with
a primary (5Ib) or secondary (5IIb) hydroxy group) and
nonterminal (5d ) â-hydroperoxy alcohols. By “terminal”
or “nonterminal” â-hydroperoxy alcohol we mean the one
deriving from a terminal or an internal epoxide, respec-
tively. Referring to their initial step, way a′ and way a′′
will henceforth be concisely denoted as the “oxidation”
and “decomposition” ways, respectively.
A. Th e Oxid a tion a n d Decom p osition Wa ys (a ′
a n d a ′′). For a better understanding of the following part
the reader is referred to Schemes 2 and 3. These provide
a more detailed picture of ways a′ and a′′ (not included
in Scheme 1 for the sake of simplicity) relative to the
examined model â-hydroperoxy alcohols 5d and 5IIb,
respectively, which lent themselves to an easier or
cleaner determination of the products resulting from the
two ways. The third compound examined, 5Ib, is dis-
cussed later.
With regard to the cycloalkene-to-acid oxidation with
Ia /H2O2, some additional comment is appropriate about
(38) The molar ratio of 2-hydroperoxy-2-phenylethanol (5If) versus
1-phenyl-1,2-ethanediol (3f) + phenacyl alcohol (4IIf) changed from
1:15 to 1:5.5 when styrene (1f), rather than its oxide 2f, was subjected
to interrupted-oxidation experiments with Ia /H2O2 under the same
conditions. This is due to the fact that, starting from 2f, heating of
the substrate/catalyst Ia /water mixture prior to addition of hydrogen
peroxide according to the procedure adopted (cf. Experimental Section)
favors an early hydrolysis of the epoxide.
(39) For a definition of “normal” and “abnormal” product with regard
to the direction of ring opening by a nucleophile in unsymmetrical
epoxides, see: Parker, R. E.; Isaacs, N. S. Chem. Rev. 1959, 742 and
references therein.
(40) Cope, A. C.; Martin, M. M.; McKervey, M. A. Q. Rev. (London)
1966, 20, 119 and references therein.

7196 J . Org. Chem., Vol. 63, No. 21, 1998
Antonelli et al.
Sch em e 2. Cou r se of Wa y a ′ for â-Hyd r op er oxy
peroxide (40% w/v) and was accelerated by the addition
of the metal catalyst. In the presence of the latter,
decomposition took place (though very slowly and much
more consistently if the water phase was omitted) even
under milder conditions (35 °C) where thermolysis oc-
curred only to a negligible extent. Under these condi-
tions, the above-mentioned R-ketol and R-diketone byprod-
ucts, indicative of a radical pathway,42b,c were observed
as well.
Alcoh ol 5d
These observations suggest that the decomposition of
the R-hydroperoxy ketone intermediate promoted by
tungsten complex Ia , like the thermal one, proceeds
homolytically. The tungsten-catalyzed homolytic decom-
position of hydroperoxides has been reported.43 The
R-hydroperoxy ketone homolysis may be somewhat com-
plicated because of competitive acid-catalyzed heterolysis
due to both the organic acid, which accumulates during
the decomposition,42a and the Bro¨nsted acid activity of
Sch em e 3. Cou r se of Wa y a ′′ for â-Hyd r op er oxy
Alcoh ol 5IIba
catalyst Ia (see section IVA). Thermal, redox (Fe2+
,
Mn2+), and acid-promoted decompositions of R-hydrop-
eroxy ketones and related mechanisms have been previ-
ously investigated.42
Operating with the terminal â-hydroperoxy alcohol
5IIb in a way analogous to that followed for 5d , we were
able to evidence the formation of the corresponding
primary R-hydroperoxy ketone 6IIb. Rather notably,
compounds of this kind have not been hitherto observed.
At variance with the secondary R-hydroperoxy ketone
6d , detection of the primary 6IIb was possible by GC
(cold on-column injector) only with some difficulty be-
cause of its greater instability;44 however, this did not
prevent identification by GC-MS. On the contrary, this
compound could be easily revealed by HPLC, and liquid
chromatography/thermospray-mass spectrometry (LC/
TSP(MS) analysis confirmed the molecular weight. The
structure in any case was unambiguously established by
reduction with triphenylphosphine to the corresponding
R-ketol.
It is worthwhile that silylation with BSTFA (even in
the absence of pyridine) of 1-hydroperoxy-2-decanone
(6IIb) in the reaction mixture did not lead to the
corresponding silyl derivative to any appreciable extent
(as revealed by GC-MS analysis) but essentially to the
addition product at the aldehydic carbonyl group of the
R-keto aldehyde expected to form from decomposition of
the former.45 The reaction of nonenolizable carbonyl
compounds with silylating agents under nucleophilic
catalysis conditions is known.46
a
For 5d the R-ketol can arise directly from decomposition of
the â-hydroperoxy alcohol, in addition to the 1,2-diol oxidation.
(a ) â-Hyd r op er oxy Alcoh ols Bea r in g a Secon d a r y
Hyd r oxy Gr ou p . P r ed om in a n ce of Wa y a ′: F or m a -
tion of r-Hyd r op er oxy Keton es. Proceeding with the
nonterminal â-hydroperoxy alcohol 5d as indicated above,
we observed (by GC) a facile formation of the correspond-
ing R-hydroperoxy ketone 6d ,41 which was identified by
GC-MS. The yield of 6d was found to pass through a
maximum value (82% at 40% conversion of 5d ) and then
to decrease after the reaction had proceeded for a time,
with the concomitant progressive increase of the final
acid 8d . The R-hydroperoxy ketone structure was con-
firmed by NMR and IR spectra of the isolated product
as well as by its reduction with triphenylphosphine to
the corresponding R-ketol. When heated (85 °C, under
helium) in the presence of water, the isolated 6-hydrop-
eroxy-5-decanone (6d ) decomposed readily, affording
pentanal (7d ) and pentanoic acid (8d ) in nearly equimo-
lar amounts, along with small quantities of 6-hydroxy-
5-decanone (4d , 3%) and 5,6-decanedione (1.5%). The
latter compounds, like the aldehyde 7d , are easily
converted into the acid 8d under the reaction conditions
(Scheme 2).
The yield of the R-hydroperoxy ketone 6IIb (analyzed
as R-ketol), like that of 6d , passed through a maximum
(42) (a) Pritzkow, W. Chem. Ber. 1955, 58, 572. (b) Pinkus, A. G.;
Haq, M. Z.; Lindberg, J . G. J . Org. Chem. 1970, 35, 2555 (c) Sawaki,
Y.; Ogata, Y. J . Org. Chem. 1976, 41, 2340 (d) Sawaki, Y., Ogata, Y.
J . Am. Chem. Soc. 1978, 100, 856.
(43) Sheldon, R. A.; Van Dorn, J . A. J . Catal. 1973, 31, 427 and
references therein.
(44) This is in line with the fact that secondary R-hydroperoxy
ketones in turn exhibit thermal as well as hydrolytic lability much
higher than the tertiary ones.47b.
(45) Analogously, under the same conditions the silyl derivative of
the secondary R-hydroperoxy ketone 6d was found to gradually convert
into the corresponding R-diketone (see Experimental Section).
(46) (a) Birkofer, L.; Ritter, A. Angew. Chem., Int. Ed. 1965, 4, 417.
(b) Nakamura, E.; Shimizu, M.; Kuwajima, I. Tetrahedron Lett. 1976,
1699. (c) Furin, G. G.; Vyazankina, O. A.; Gostevsky, B. A.; Vyazankin,
N. S. Tetrahedron 1988, 44, 2675. (d) Kozyukov, V. P.; Kozyukov, Vik.
P.; Mironov, V. F. Zh. Obshch. Khim. 1982, 52(6), 1386; Chem. Abstr.
1982, 97, 163077. (e) Kuvajima, I.; Nakamura, E. Acc. Chem. Res. 1985,
18, 181.
Interestingly, the decomposition of 6d occurred much
more slowly if carried out in the presence of hydrogen
(41) In addition to detection as the silyl derivative (for details, see
Experimental Section), 6d could also be revealed as such by GC (cold
on-column injector), without appreciable decomposition. Its formation
during the progress of transformation 5d f 8d was quantitated by
GC as the R-ketol, after reduction in situ with triphenylphosphine.

Oxidative Cleavage of Olefins to Carboxylic Acids
J . Org. Chem., Vol. 63, No. 21, 1998 7197
value (78%) which, however, was obtained at a much
lower conversion of the â-hydroperoxy alcohol (17%
instead of 40%), owing to the slower oxidation of the
latter combined with the faster decomposition of the
resulting oxo derivative (see above).
To the best of our knowledge, these are the first
examples of R-hydroperoxy ketones obtained by metal-
catalyzed oxidation of â-hydroperoxy alcohols with hy-
drogen peroxide. R-Hydroperoxy ketones (secondary and
tertiary) are usually prepared by autoxidation of aliphatic
ketones at 100 °C in the absence of catalysts47 and at or
below 0 °C in the presence of strong bases.48
Following the progress of the transformation of 1-hy-
droperoxy-2-decanol (5IIb) into acids with Ia /H2O2, we
found the aldehyde of one less carbon, nonanal (7b), to
be formed together with 1-hydroperoxy-2-decanone (6IIb)
(6% at 17% conversion of 5IIb, corresponding to the
maximum yield (78%) for intermediate 6IIb, as men-
tioned above). Since the C9 aldehyde cannot come from
fragmentation of 6IIb, which leads to formaldehyde and
nonanoic acid, this result unequivocally shows that the
decomposition way a′′ is operative, together with the
oxidation way a′, for compound 5IIb (Scheme 3).
Evidence of way a′′ through the formation of the
aldehyde could also be gained indirectly for the nonter-
minal â-hydroperoxy alcohol 5d . It was based on the
value of the aldehyde-to-carboxylic acid molar ratio at
partial conversion of 5d , which was found to be higher
than unity; it would be expected to be less if only way a′
were operative, considering that the formed aldehyde
oxidizes under the reaction conditions. In fact, on
decomposition (way a′′) 6-hydroperoxy-5-decanol (5d )
gives rise to two molecules of C5 aldehyde while it affords
one molecule of C5 aldehyde and one of C5 acid on
oxidation, through cleavage of the resulting R-hydroper-
oxy ketone 6d (way a′). At 40% conversion of 5d ,
corresponding to the maximum yield (82%) for intermedi-
ate 6d (see above), the value of this ratio indicated the
presence of about 8% of pentanal ascribable to the
decomposition (4%) of 5d .
ary hydroperoxy group, the corresponding R-ketol 4d was
also formed.
The above findings, while proving the occurrence of
ways a′ and a′′ for both nonterminal (R, R1 ) alkyl) and
type II terminal (R ) alkyl, R1 ) H) â-hydroperoxy
alcohols 5 such as 5d and 5IIb, clearly point to the
predominance of way a′ for these classes of compounds.
On the basis of the maximum yield found with 5IIb and
5d for the R-hydroperoxy ketone intermediate (also
taking into account, on one hand, the associated cleavage
products and, on the other, the competitively produced
aldehyde, 1,2-diol, and R-ketol), we can estimate that for
the substrates in question the final carboxylic acids
pertaining to transformation 5 f 8 come from the
oxidation way a′ to the extent of at least 85%. This figure
is likely to actually be a little lower if one takes into
account that the intermediates formed by the alternative
way a′′ are in part oxidized during the experiment.
(b) â-Hyd r op er oxy Alcoh ols Bea r in g a P r im a r y
Hydr oxy Gr ou p. Com par able Con tr ibu tion of Ways
a ′ a n d a ′′. In light of these results, no doubt also with
type I terminal (R ) alkyl, R1 ) H) â-hydroperoxy
alcohols 5, such as 2-hydroperoxy-1-decanol (5Ib), ways
a′ and a′′ are both operative. In this case, however,
monitoring (by GC or HPLC) the progress of the trans-
formation of 5Ib into acids with Ia /H2O2, we were not
able to evidence the formation of the relative intermedi-
ate diagnostic of the oxidation way a′, R-hydroperoxy
aldehyde 6Ib. This may be due to the very high instabil-
ity of such compounds,50 which justifies the fact that, to
our knowledge, no authentic example of R-hydroperoxy
aldehyde has so far been reported. Moreover, it was also
impossible to distinguish the aldehyde formed (together
with the 1,2-diol and R-ketol) by the alternative way a′′,
through decomposition of 5Ib, since the same aldehyde
(other than formaldehyde) can also be generated by way
a′ through cleavage of 6Ib. This situation, while not
allowing us to give a precise picture of way a′ for the
compounds under study, also makes it rather difficult to
assess the relative importance of the two ways for these
compounds.
It should be pointed out that the aforementioned
formation (6%) of nonanal (7b), observed at low conver-
sion (17%) of 1-hydroperoxy-2-decanol (5IIb), was ac-
companied by that of 1,2-decanediol (3b, 4.5%) and
1-hydroxy-2-decanone (4IIb, 3%) (Scheme 3). The pres-
ence of the corresponding 1,2-diol 3d (in a very small
amount) and R-ketol 4d (3%) was likewise observed with
the â-hydroperoxy alcohol 5d (at 40% conversion). These
products, like the aldehyde, have to be regarded as
indicative of the decomposition of hydroperoxide com-
pounds, such as 5IIb and 5d ,49 and contribute to forma-
tion of the final carboxylic acid by way a′′.
However, if we consider the fact that from comparative
experiments primary alcohols appear to be over three
times less oxidizable than the secondary ones by Ia /H2O2
under the reaction conditions, we can also reasonably
expect 1-hydroxy-2-hydroperoxy derivatives 5I (R ) alkyl,
R1 ) H) to exhibit an analogous behavior toward oxida-
tion with respect to their 2-hydroxy-1-hydroperoxy iso-
mers 5II. Thence, we would expect the former to be more
available than the latter to be converted into acids by
the alternative decomposition way a′′, the stability of the
This view is supported by the results of blank decom-
position experiments with the above â-hydroperoxy al-
cohols under the reaction conditions. When heated (85
°C, under helium) admixed with catalyst Ia in the
presence of the aqueous phase but in the absence of the
oxidant hydrogen peroxide, compounds 5IIb and 5d gave
rise to the respective 1,2-diols 3b and 3d along with the
expected aldehydes. In the case of 5d , bearing a second-
(49) At variance with 5d , in the case of 5IIb bearing a primary OOH
group the found R-ketol cannot originate directly from decomposition
of the â-hydroperoxy alcohol (H-abstraction reaction). Its formation
has to be ascribed to oxidation of the previously formed 1,2-diol by
Ia /H2O2. In any case, it speaks for way a′′ as well. A priori, we cannot
exclude, for both 5IIb and 5d , that the found R-ketol may in part arise
also from decomposition of the corresponding R-hydroperoxy ketone
formed by the alternative way a′ (cf. text and Scheme 2 for the case of
5d ). However, since (i) the figures considered refer to experiments at
low/partial conversion, where the intermediate R-hydroperoxy ketone
formed is present to the extent of 78-82%, and (ii) decomposition of
the latter leads only to minor amounts of R-ketol (cf. text and Scheme
2 for the case of 6d ), the possible contribution of way a′ to the formation
of the found R-ketol (due to the decomposed R-hydroperoxy ketone) is
insignificant.
(47) (a) Sharp, D. B.; Patton, L. W.; Whitcomb, S. E. J . Am. Chem.
Soc. 1951, 73, 5600. (b) Sharp, D. B.; Whitcomb, S. E.; Patton, L. W.;
Moorhead, A. D. J . Am. Chem. Soc. 1952, 74, 1802.
(48) (a) Cubbon, R. C. P.; Hewlett, C. J . Chem. Soc. C 1968, 2978.
(b) Sawaki, Y.; Ogata, Y. J . Am. Chem. Soc. 1975, 97, 6983.
(50) Baader, W. J .; Bohne, C.; Cilento, G.; Dunford, H. B. J . Biol.
Chem. 1985, 260, 10217.

7198 J . Org. Chem., Vol. 63, No. 21, 1998
Antonelli et al.
two types of compounds under the reaction conditions
being nearly the same as ascertained in control experi-
ments with 5Ib and 5IIb.51 Thus, the contribution of way
a′′, estimated at about 15% for terminal (R ) alkyl, R1 )
H) â-hydroperoxy alcohols 5 of type II on the basis of the
data obtained for 5IIb, should reach much higher values
(around 50%) with analogous compounds of type I.
B. Med iu m Effect on Tr a n sfor m a tion 5 f 8. In
addition to the structural factors of â-hydroperoxy alco-
hols 5 (the presence of a primary rather than secondary
hydroxy group), the reaction medium certainly also
affects the extent to which way a′′ operates in the
reaction. This emerges from the results of blank decom-
position experiments with compounds 5IIb and 5d
performed under conditions different from the standard
ones (which implied the presence of both catalyst Ia and
water, as reported above and in the Experimental Sec-
tion). Indeed, whereas decomposition of 5IIb and 5d was
found to occur much faster when the aqueous phase was
omitted, on the contrary it took place much more slowly
in the presence of water alone (without the metal
catalyst) and even more slowly when water was replaced
with hydrogen peroxide (40% w/v). It thus turns out that
hydrogen peroxide exerts a stabilizing action on the
â-hydroperoxy alcohol, analogous to that exerted on the
R-hydroperoxy ketone. Furthermore, water retards de-
composition of the â-hydroperoxy alcohol by the metal
catalyst, likely by complexing with the latter.52 The
retarding effect of the aqueous medium should, of course,
be greater under the actual reaction conditions due to
the presence of hydrogen peroxide which, as a better
nucleophile than water,53 competes more efficiently with
the â-hydroperoxy alcohol for coordination sites on the
metal.
As to the nature of the decomposition of 5 under the
reaction conditions, the results of the relative blank
experiments suggest that a homolytic process (thermal
and metal-induced) is mainly involved, analogously to the
previously discussed decomposition of R-hydroperoxy
ketones. Also in this case, however, an alternative acid-
catalyzed heterolysis of 5 may occur to some extent under
the acidity conditions of the reaction (see section IVA).
C. Min or Rea ction s in Tr a n sfor m a tion 5 f 8:
F or m a tion of â-Hyd r oxyp er oxyh em ia ceta ls. As we
have seen, transformation 5 f 8 is the result of inter-
twining oxidation and decomposition reactions following
ways a′ and a′′. Actually, side reactions may further
complicate the course. Indeed, as results from blank
experiments, aldehyde 7 (which in transformation 5 f
8 originates from 5 by both way a′ and way a′′) can react
with the â-hydroperoxy alcohol itself to give â-hydrox-
yperoxyhemiacetal 9 (eq 1, where R2 and R3 ) alkyl and
H, alkyl, respectively, or vice versa and R4 ) H, alkyl
stand for R or R1 of Scheme 1). Adduct 9 then decom-
poses on heating into carboxylic acid 8 and R-hydroxy
carbonyl compound 4 with evolution of hydrogen (eq 2),
according to a pattern amenable to the previously re-
ported53 thermolysis of peroxyhemiacetals, leading to
acids and aldehydes.
As a matter of fact, after subjecting â-hydroperoxy
alcohols 5IIb and 5d to blank decomposition experi-
ments, appreciable amounts of the corresponding adducts
9 (R2, R4 ) H, R4 ) n-C8H17 and R2 ) H, R3, R4 ) n- C8H17
in the former case, R2, R3, R4 ) n-C4H9 in the latter) were
detected. Likewise, the carboxylic acids 8b and 8d
related to the expected aldehydes 7b and 7d invariably
formed, along with the above-mentioned 1,2-diols and
R-ketols. It has to be observed, however, that the
equilibrium of eq 1, as ascertained in control experiments,
is markedly shifted to the left in the presence of excess
hydrogen peroxide, which interacts with the aldehyde to
form adducts54 and the corresponding acid. It follows
that the actual contribution of the “â-hydroxyperoxy-
hemiacetal” route to the formation of carboxylic acids 8
(and associated precursors 4) from â-hydroperoxy alcohols
5 is negligible in our process.
IV. F a ctor s Con d ition in g th e Rea ction . In addi-
tion to the solvent mentioned at the beginning, another
factor influences the course of the reaction: the hydrogen
peroxide concentration. As to the perhydrolytic pathway,
the importance of this factor is readily understandable,
because the activity of hydrogen peroxide as a nucleophile
in the oxirane-opening step (to give the â-hydroperoxy
alcohol) depends on its concentration. Less apparent is
the fact that the oxirane opening by water, which
initiates the alternative hydrolytic pathway, is also
indirectly affected by the hydrogen peroxide concentra-
tion. This aspect will now be discussed.
A. Hyd r ogen P er oxid e Con cen tr a tion a n d Me-
d iu m Acid ity. Im p lica tion s for th e Oxir a n e-Op en -
in g Step . Studying the course of the oxidation of
medium- and long-chain alkenes to carboxylic acids with
Ia /H2O2 (section IIA), we found that the interrupted
oxidation of the intermediate epoxyalkanes under the
two-phase reaction conditions (performed on C8 and C10
model compounds), in addition to affording the respective
â-hydroperoxy alcohols (major product), also leads to
substantial amounts of the corresponding hydrolysis
products (1,2-diols and their primary oxidation products,
R-ketols).
The situation, however, is quite different if the above
reaction is carried out with 16% instead of 40% w/v
hydrogen peroxide, other conditions being equal. In this
case, the formation of the 1,2-diol (and thence of the
R-ketol) is considerably slowed, as is that of the â-hy-
droperoxy alcohol. The effect of the concentration of
hydrogen peroxide used (16% and 40% w/v) on the
hydrolysis of 1,2-epoxyoctane (2a ) and -decane (2b), when
the latter are put into contact under stirring with the
Ia /H2O2 two-phase system (procedure A), is shown in
Table 2 (runs 1-4).
(51) In comparative experiments carried out in the presence of
catalyst Ia and water (85 °C, 1 h, under helium), 5IIb was found to
decompose only ca. 1.2 times faster than 5Ib.
The observation that in the above experiments epoxy-
alkanes 2a ,b exhibit a different aptitude to hydrolysis
(52) Such a behavior is exhibited by coordinating solvents with
hydroperoxides, see: Sheldon, R. A.; Van Doorn, J . A.; Schram, C. W.
A.; De J ong, A. J . J . Catal. 1973, 31, 438.
(54) (a) Durham, L. J .; Wurster, C. F.; Mosher, H. S. J . Am. Chem.
Soc. 1958, 80, 332. (b) Durham, L. J .; Mosher, H. S. J . Am. Chem.
Soc. 1960, 82, 4537.
(53) Sander, E. G.; J encks, W. P. J . Am. Chem. Soc. 1968, 90, 6154.

Oxidative Cleavage of Olefins to Carboxylic Acids
J . Org. Chem., Vol. 63, No. 21, 1998 7199
Ta ble 2. Tr en d of th e Ia -Ca ta lyzed Hyd r olysis of
1,2-Ep oxyocta n e (2a ) a n d -d eca n e (2b) u n d er Tw o-P h a se
Con d ition s (w ith ou t Ad d ed Solven t) in th e P r esen ce of
Hyd r ogen P er oxid e of Differ en t Con cen tr a tion a
reaction requiring acid catalysis, as is the case for the
epoxide hydrolysis, is involved therein, it must feel the
effects of this situation, as shown by the following
comparative experiments (performed under homogeneous
conditions).
% conv 2 to 3b,c
% unconv 2b
% H2O2 proce-
run epoxide
w/v
dure
70 °C
85 °C
70 °C 85 °C
Three acidic (H3PO4, 0.8 mg) solutions (1.3 mL) of
propylene oxide (3 mmol), two in aqueous hydrogen
peroxide (37% and 16% w/v, respectively) and the third
one in water alone, were kept under stirring for 1 h at
30 °C. In the first experiment, GC analysis revealed the
formation of 67% of 1,2-propanediol (in addition to 20%
of the 2(1)-hydroperoxy-1(2)-propanol mixture), while
47% of the above diol was formed (along with 4.5% of
the corresponding regioisomeric â-hydroperoxy alcohols)
when hydrogen peroxide of lower strength was used, the
remaining part being substantially unconverted propy-
lene oxide. Finally, in the last experiment the diol
formed was only 34%.
Reverting now to the aforementioned two-phase hy-
drolysis experiments in the presence of hydrogen peroxide
and catalyst Ia (Table 2, runs 1-4), it appears reasonable
that the effects on the acid hydrolysis of epoxyalkanes
2a ,b related to the different concentration of the hydro-
gen peroxide used are much more pronounced than in a
monophasic system (apart from the efficiency of the
reaction which, of course, is lower).
In fact, the difference in acidity existing betweeen the
two types of experiments (16% and 40% w/v H2O2) with
regard to the aqueous phase of the Ia /H2O2 two-phase
system (see Experimental Section) could be greater at
the interface, where the epoxyalkane hydrolysis presum-
ably occurs, as a consequence of the fact that the “more
lipophilic” hydrogen peroxide is expected to pass into the
organic phase preferably to water, in proportion to its
concentration. On the other hand, depending on its
concentration, hydrogen peroxide could also exert some
“homogenizing action” on the phases, which favors acid-
activation of the epoxide and its interaction with the
aqueous medium. The polar features of the products
themselves arising from the oxirane opening could con-
tribute to the purpose.
The trend of runs 1-4 of Table 2 is confirmed by the
data of separate experiments, reported in the same table
(runs 5-8), which correlate the aptitude to hydrolysis of
epoxyalkanes 2a ,b with the concentration (16% or 40%
w/v) of hydrogen peroxide used when the former are
caused to react with the sole aqueous phase of the Ia /
H2O2 biphasic system (procedure B) rather than with the
entire system (procedure A).
1
2
3
4
5
6
7
8
9
2a
2a
2b
2b
2a
16
40
16
40
16
40
16
40
40
40
A
A
A
A
B
B
B
B
13.5 (8.5)
35 (44)
77
11
4.2 (6.7)
22 (47)
88.5
18
6.3 (0.8)
15 (7)
92
77
2a
2b
2b
2bd
2be
0.8 (tr)
99
94
7.5
79
4.2 (1.5)
34.5 (42)
3.8 (14.6)
10
A
a
Unless otherwise stated, all experiments were performed on
3 mmol of epoxyalkane by reacting the epoxide under stirring with
the Ia /H2O2 (16% or 40% w/v) biphasic system (procedure A) or
with the sole aqueous phase of the latter (procedure B) at the
indicated temperature for 30 min, as described in detail in
b
Experimental Section. GC determination (against an internal
standard). c In procedure A, the indicated value also includes the
R-ketol formed by subsequent oxidation of the 1,2-diol under the
reaction conditions. The value in parentheses refers to % conver-
sion of the epoxide into the â-hydroperoxy alcohol. The remaining
part of the converted epoxide essentially consisted of condensation
products. In procedure A, small amounts of the related aldehyde
d
and carboxylic acid of one less carbon were also formed. 1,2-
Epoxydecane (3 mmol) was caused to react under stirring with
40% w/v aq H2O2 (1.27 mL) in which 10 mg (ca. 0.04 mmol) of
tungstic acid was previously dissolved. e The neutral complex
methyltrioctylammonium µ-oxo-bis[oxodiperoxotungstate(VI)]60a
was used as catalyst in place of Ia . To equalize the amount of
tungsten present in complex Ia , 0.06 mmol of the former complex
(equivalent to 0.12 mmol of tungsten) was used.
depending on the concentration of the hydrogen peroxide
used at first sight may seem surprising in view of the
fact that the primary source of the acidity promoting the
hydrolytic opening of the oxirane ring is the same: the
metal catalyst. However, this puzzling feature is expli-
cable if one keeps in mind that the aqueous phase in the
two biphasic reaction systems is different in character.
It has long been known56 that the pH meter readings
in aqueous solutions of hydrogen peroxide are lower than
the “real pH” (in water) and the difference, revealed by
the glass-calomel electrode system as a shift in the E0
value, increases with the increasing concentration of
hydrogen peroxide. This behavior has been related56 to
the difference in free energy of solvation of the proton in
water and in nearly anhydrous hydrogen peroxide (about
9 kcal/mol), which indicates that the proton is more
firmly bound to water. Accordingly, a water molecule,
by interaction with neighboring hydrogen peroxide mol-
ecules, has its basicity reduced, and the extent of this
reduction changes with the percentage of hydrogen
peroxide in the aqueous medium.57
From the comparison of the data of runs 1-4 and 5-8
of Table 2 it turns out, however, that in the latter case
the hydrolysis rate of 2a ,b is much lower. In particular,
the acidity of the sole aqueous phase of the Ia /H2O2
biphasic system appears to be virtually inadequate to
promote hydrolysis of more lipophilic long-chain epoxy-
alkanes such as 1,2-epoxydecane (2b) under two-phase
conditions, even when more concentrated hydrogen per-
oxide is used (Table 2, run 8). This leads us to argue
that the results of runs 1-4 of Table 2 are attained
mainly due to the catalytic contribution of acidic species
(arising from the hydrolytic degradation of catalyst Ia )
which are likely held more effectively in the organic
phase by the “onium” group under the oxidizing reaction
conditions.58
Thus it appears that the experiments using hydrogen
peroxide in acidic aqueous solution but differing in the
concentration of this reagent differ also in the acidity of
the reaction medium, which is invariably higher than
when the medium is exclusively aqueous, despite the fact
the “real pH” values are the same. Therefore, if a
(55) (a) Spa¨th, E.; Pailer, M.; Schmid, M. Chem. Ber. 1941, 74, 1552.
(b) Swern, D. Organic Peroxides; Wiley-Interscience: New York, 1970;
Vol. 1, Chapter 1, pp 24-26.
(56) Kolczynski, J . R.; Roth, E. M.; Shanley, E. S. J . Am. Chem. Soc.
1957, 79, 531.
(57) Mitchell, A. G.; Wynne-J ones, W. F. K. Discuss. Faraday Soc.
1953, 15, 161.
That hydrolysis of 2a ,b occurs mostly by Bro¨nsted
rather than Lewis acid catalysis is suggested also by the

7200 J . Org. Chem., Vol. 63, No. 21, 1998
Antonelli et al.
following observations. First, when (paralleling proce-
dure B of Table 2) 1,2-epoxydecane (2b) was put into
contact under stirring with an aqueous solution of
hydrogen peroxide (40% w/v) containing an amount of
tungstic acid ca. 5 times greater than that present (along
with phosphoric acid) in the aqueous phase of the Ia /
H2O2 biphasic system (for details, see Experimental
Section under the heading Acidity of the Aqueous Phase...),
the reaction trend was comparable to that observed with
the same epoxide adopting procedure A (Table 2, cf. runs
3 and 9).59 Second, when a tungsten peroxocomplex able
to ensure nearly neutral reaction conditions, [(C8H17)3-
NCH3]2[(O2)2W(O)O(O)W(O2)2 (methyltrioctylammonium
µ-oxo-bis[oxodiperoxotungstate(VI)]),60a was employed
with the epoxide 2b as a catalyst in place of Ia under
the same conditions, a much less satisfactory result was
obtained (Table 2, cf. runs 3 and 10).
Needless to say, the use of more concentrated hydrogen
peroxide for the reasons above stated also favors the
perhydrolysis of 2a ,b in addition to the hydrolysis. As
emerges from the data of Table 2 (cf. runs 1-4 and 5-8),
the presence of peroxocomplex Ia as such (procedure A)
leads to an increased proportion of the perhydrolysis
product compared to that of hydrolysis, other conditions
being equal. A similar effect is also observed employing
the neutral tungsten peroxocomplex mentioned above
(run 10) or tungstic acid (in substantial amount) dissolved
in hydrogen peroxide (run 9) as a catalyst. It may well
be that the presence of a tungsten peroxidic species
(preformed or generated in situ) makes hydrogen perox-
ide much more accessible than water to interaction with
the epoxyalkane. Ascertaining the origin of this effect
requires further study, however.
Thus, according to the results of the experiments
reported above, the choice we made initially of utilizing,
in the present olefin-to-acid oxidation, hydrogen peroxide
more concentrated than that usually used by us (16% w/v)
in the alkene epoxidation13 proves itself, a posteriori, a
prerequisite for the success of the reaction in the case of
medium- and long-chain alkenes. With these substrates,
the perhydrolytic and hydrolytic pathways are both
favored by the above circumstance, given the fact that
the epoxides involved are highly lipophilic and little prone
to undergo ring opening by hydrogen peroxide or water
(initiating the two pathways) under the heterogeneous
reaction conditions.
Conversely, with arylalkenes and C5-C7 cycloalkenes,
the use of hydrogen peroxide of lower strength would not
affect the reaction course appreciably, owing to the
hydrolytic/perhydrolytic lability of the respective inter-
mediate epoxides even under two-phase conditions.
B. Th e Solven t a s Hin d r a n ce to th e Oxir a n e-
Op en in g Step in a Tw o-P h a se Rea ction System . The
fact that for medium- and long-chain alkenes, under the
heterogeneous conditions adopted, ring opening of the
intermediate epoxyalkanes by hydrogen peroxide or
water represents the critical point of the reaction explains
why the use of the solvent leads, in these cases, to a
progressive, marked worsening of the oxidation on pass-
ing from a C8 to a C16 alkene.
Under these conditions, the interaction between the
epoxyalkane and the nucleophile, already rather difficult
in itself as we have seen, is made even more difficult, in
right ratio to the increasing lipophilic character of the
epoxyalkane involved, with obvious negative effects on
the course of the whole oxidation. As confirmed by
control experiments carried out with added solvent, the
slowing of the reaction is accompanied by a larger
nonproductive consumption of hydrogen peroxide per
mole of carboxylic acid formed, which ultimately results
in a decrease in efficiency and selectivity of the oxidation,
depending on the substrate and solvent used.
C. Bip h a sic (w it h ou t Ad d ed Solven t ) ver su s
Mon op h a sic P r oced u r e. For what has been said with
regard to the difficulty of the oxirane-opening step in a
two-phase system for the substrates mentioned above,
one should expect the adoption of a monophasic procedure
(t-BuOH as the solvent) to be particularly advantageous
in these cases.
In fact, on carrying out the oxidation of medium- and
long-chain alkenes with Ia /H2O2 under homogeneous
conditions (20 mL of t-BuOH per 10 mmol of alkene, 0.01
mmol of catalyst Ia , and 44-55 mmol of 40% w/v H2O2),
the formed intermediate epoxyalkanes were converted
into the corresponding â-hydroperoxy alcohols/1,2-diols
much faster than under our “special” two-phase condi-
tions (lack of the solvent). However, the advantages
offered by the monophasic procedure in this stage of the
reaction were found to be by far frustrated by the
slowdown of the other stages, due to both the dilution of
the reactants and the presence of water. It is known1c
that in monophasic systems water seriously retards the
hydrogen peroxide epoxidation of alkenes, particularly
the less reactive terminal ones. This can account for the
modest results (33-45% yield after 24 h) reported,7,8 for
example, for the tungsten-promoted oxidation of 1-octene
to octanoic acid with aqueous hydrogen peroxide under
homogeneous conditions.
It follows that the special two-phase procedure here
described for the oxidation of alkenes to carboxylic acids
has to be regarded as a compromise solution ensuring in
general the best overall execution of all stages of the
reaction.
V. Th e Role of Ca ta lyst Ia in th e Rea ction . From
this investigation it is apparent that complex Ia plays a
manifold role in the reaction as a catalyst. First, it exerts
(in conjunction with hydrogen peroxide) an oxidizing
action toward alkenes,11-13 transferring its active oxygen
to the double bond of these substrates. Which of the
structurally different peroxo groups (terminal or bridg-
ing)12 of Ia are involved in the oxygen transfer has been
estimated so far by theoretical studies on model com-
plexes.61 These studies suggest that the bridging peroxo
group is primary involved in the olefin epoxidation, with
the “most electrophilic” oxygen atom adjacent to the triply
shared oxygen presumably being transferred to the
alkene. Likewise, complex Ia acts as oxidant of the 1,2-
diol, â-hydroperoxy alcohol, and aldehyde17 intermediates
to give the corresponding R-ketols, R-hydroperoxy ke-
tones, and final carboxylic acids, respectively.
(58) In the absence of hydrogen peroxide, no appreciable contribution
to the hydrolysis of epoxyalkanes 2a ,b by such species was observed.
(59) In contrast, poor results were obtained using 70% HClO4, 96%
H2SO4, or 85% H3PO4 as acid catalysts in place of H2WO4 (the first
one in the same molar amount, the other two in a molar amount 5-fold
greater), under identical conditions. This points to the importance of
the lipophilic character of the used acid catalyst for a satisfactory
oxirane opening under two-phase reaction conditions.
(60) (a) Prepared according to the procedure described in the
literature for the analogous phosphonium complex.60b (b) Prandi, J .;
Kagan, H. B,; Mimoun, H. Tetrahedron Lett. 1986, 27, 2617.
(61) Fantucci, P.; Lolli, S.; Venturello, C. J . Catal. 1997, 169, 228.

Oxidative Cleavage of Olefins to Carboxylic Acids
J . Org. Chem., Vol. 63, No. 21, 1998 7201
The role of Ia as a catalyst is not limited to oxidation,
however. Being subject to hydrolysis, it can also act as
an acid, favoring the “normal” opening of the formed
epoxides by hydrogen peroxide or water as well as the
“anomalous” rearrangements40 of medium-sized ring cy-
cloalkene oxides (leading, for example, to formation of
diketones or ketodiacids), which characterize the oxida-
tion of the parent cycloalkenes. Acting as an acid, Ia also
favors the oxidative cleavage of the intermediate R-ketols
by hydrogen peroxide. Finally, according to the behavior
of transition-metal compounds which form one-electron
redox systems,43 Ia can initiate the homolytic decomposi-
tion of hydroperoxides such as the â-hydroperoxy alcohol
and R-hydroperoxy ketone intermediates.
hydrogen peroxide with difficulty under the heteroge-
neous reaction conditions.
In connection with this observation, a rationale has
been provided for the negative effect of the solvent
observed in the oxidation of the above substrates to acids.
Finally, a comparison of the effectiveness of the present
biphasic procedure (without added solvent) and that of
an analogous monophasic procedure has shown the
superiority of the former.
Exp er im en ta l Section
Infrared spectra were recorded with a FT-IR spectrometer.
1H and 13C NMR spectra were obtained at 300 and 75.5 MHz,
respectively, using (CH3)4Si as an internal standard. Mass
spectra were obtained at 70 eV on a triple quadrupole mass
spectrometer coupled to a gas chromatograph. The columns
used were the same as for the GC analyses (see below). Unless
otherwise stated, CI mass spectra were recorded using isobu-
tane as the reactant gas. GC analyses were performed on a
flame ionization gas chromatograph (with column-temperature
programming) using three different bonded-phase, fused silica
capillary columns: (a) SPB-5 (30 m × 0.32 mm i.d., 0.25-µm
film, Supelco), (b) TAP (5 m × 0.25 mm i.d., 0.10-µm film,
Chrompack, for detection of â,â′-dihydroxyalkyl peroxides and
ethers as well as of â-hydroperoxy alcohol-aldehyde adducts),
and (c) methyl phenyl cyanopropyl silicone (30 m × 0.53 mm
i.d., 3.0-µm film, Quadrex, for acetic acid determination).
Thin-layer chromatography (TLC) was performed on Merck
precoated silica gel 60 F-254 plates, and spots were detected
by exposure to iodine fumes or by spraying with a bromophenol
blue or potassium permanganate solution, according to the
product. Column chromatography was carried out using
Merck Kieselgel 60 (70-230 mesh). HPLC analyses were
performed using an instrument equipped with a UV detector
(at 260 nm). A 10-µL volume of a CH3CN solution of the
sample was injected on an Li Chrospher 100RP-18 end-capped
column (Merck, 4 × 250 mm, 5 µm) using a mobile phase of
CH3CN/water with a linear gradient (% CH3CN from 2 to 100
in 25 min) and a flow rate of 0.8 mL/min. LC/TSP-MS
analyses were performed on the same liquid chromatograph
interfaced to a triple quadrupole mass spectrometer equipped
with a thermospray interface, proceeeding as above. Postcol-
umn addition of 0.6 mL/min of 0.1 M ammonium acetate in
water was performed with a Waters 510 pump.
Products were identified by combustion analysis and spec-
tral data (when isolated) or by gas chromatography-mass
spectroscopy (GC-MS) (if necessary, after methylation with
CH2N2 or silylation with N,O-bis(trimethylsilyl)acetamide
(BSA)/pyridine or -trifluoroacetamide (BSTFA)/pyridine (1:1))
in comparison with authentic samples and/or the literature
values. Except for 2-hydroperoxy-2-phenylethanol 5If,26 all of
the isolated â-hydroperoxy alcohols (as either a single com-
ponent or a regioisomeric mixture) have not been described
before. They were usually stored in the freezer (-25 °C) for
some weeks without appreciable decomposition. Also R-hy-
droperoxy ketones 6IIb (detected in the reaction mixture) and
6d (isolated) have so far not been reported.
Silylation (BSTFA) of compounds 5a -f and 6d was per-
formed at room temperature. For compounds 5c,d , it usually
required about 24 h to be complete. In the case of 5If and 6d ,
if silylation was performed on the isolated product, pyridine
was omitted and diethyl ether was used as the solvent. When
obtained from the isolated product, the silyl derivative of
6-hydroperoxy-5-decanone (6d ) was relatively stable (90%
unchanged after 24 h). However, if formed in situ in the
reaction mixture (in the presence of the catalyst), it gradually
transformed through the formation of 5,6-decanedione62a into
the enol trimethylsilyl (TMS) ether derivative of the latter (E/Z
mixture; configuration not assigned) according to the known62b
mode of reaction of the silylating agents with compounds
containing an enolizable ketone group. 5-Trimethylsiloxy-4-
decen-6-one, C4H9COC(OSiMe3)dCHC3H7: GC-MS m/z (rel
Con clu sion
An efficient procedure has been described for the
oxidation of alkenes to carboxylic acids with 40% w/v
hydrogen peroxide catalyzed by tungsten peroxocomplex
Ia under two-phase conditions in the absence of organic
solvents. The investigation of the reaction has estab-
lished that the oxidation proceeds via epoxidation of
alkene 1 followed by acid hydrolysis or perhydrolysis of
the formed epoxide 2 to give a 1,2-diol 3 and a â-hydro-
peroxy alcohol 5, respectively. Metal-catalyzed hydrogen
peroxide oxidation of 3 to R-ketol 4 or of 5 to R-oxo
hydroperoxide 6 followed, in the former case (hydrolytic
pathway), by oxidative cleavage (acid-catalyzed) of 4 and,
in the latter (perhydrolytic pathway), by thermal/metal-
induced decomposition of 6, leading to the final carboxylic
acid(s) 8. In the course of the perhydrolytic process, as
an alternative to the oxidation of 5 to 8 via 6 (way a′),
compounds 5 can decompose to carbonyl derivatives 7 or
to the corresponding diols (and R-ketols, if the hydrop-
eroxy group is secondary), which also are oxidized to acids
8 (way a′′). The relative contribution of ways a′ and a′′
(denoted as the oxidation and decomposition ways,
respectively) largely depends on the oxidizability of the
hydroxy group (primary or secondary) present in 5 and
also on the reaction medium, with hydrogen peroxide
exerting a stabilizing action on the â-hydroperoxy alcohol
and significantly retarding its decomposition by the metal
catalyst.
The perhydrolytic reaction pathway has been found to
play a primary role in the oxidation of medium- and long-
chain alkenes to acids, while it intervenes to a rather
limited extent in the oxidation of arylalkenes and C5-
C7 cycloalkenes, for which the hydrolytic reaction path-
way is predominant. Evidence for the occurrence of the
perhydrolytic pathway and related ways a′ and a′′ has
been acquired by the isolation or detection of the relevant
intermediates (â-hydroperoxy alcohols, R-hydroperoxy
ketones, aldehydes, etc.) involved.
The study of the reaction has evidenced different roles
played by catalyst Ia at various stages of the process.
Likewise, it has also shown that hydrogen peroxide, in
addition to acting in the reaction as an oxidant (in
combination with catalyst Ia ) and a nucleophile, exerts
a third important function as a solvent. In this role,
depending on its concentration, hydrogen peroxide influ-
ences the medium acidity to a different extent and
indirectly conditions the efficiency of the hydrolytic
reaction pathway (and also of the perhydrolytic one),
markedly in the case of medium- and long-chain alkenes,
whose epoxides undergo acid ring opening by water or

7202 J . Org. Chem., Vol. 63, No. 21, 1998
Antonelli et al.
intensity) (EI) 242 (5, M+), 227 (77), 183 (34), 157 (9, [M -
C4H9CO]+), 143 (10), 75 (45), 73 (100). The above transforma-
tion was greatly accelerated by the presence of pyridine.
The evaluation of the aldehyde(s) present in the reaction
mixture by GC analysis required prior reduction of the
hydroperoxide compounds with triphenylphosphine.
rt (except for entries 10 and 15, where the temperature was
kept around 70 °C), the mixture was worked up as follows.
En tr ies 1-4, 7, a n d 11-14. Dichloromethane (Et2O, in
entry 2) (20 mL) was added, and the two-phase mixture was
stirred for a few minutes, poured into a separatory funnel (the
organic layer only, in entry 4), and extracted with 10% Na2-
CO3 (4 × 15 mL). After being washed with Et2O (10 mL), the
basic extracts were acidified with dilute HCl (pH ≈ 1) under
stirring, treated with Na2SO3 to decompose the unreacted
hydrogen peroxide (the pH was adjusted to its original value
with some HCl, if necessary), salted out (NaCl) in entries 7
and 12-14, and re-extracted with n-pentane (Et2O in entries
3 and 12-14; CH2Cl2 in entry 11) [4 × 30 mL (6 × 50 mL in
entry 11)]. The combined organic extracts, washed with water
(3 × 7 mL) in entry 4 to remove acetic acid thoroughly, were
dried over Na2SO4, filtered, and (except in entries 12-14)
evaporated to give the desired acid. The product so obtained
was either purified by distillation under reduced pressure
(entries 1 and 2) or, after dissolution in Et2O (20 mL), passed
through a short column (2.5-cm diameter) of silica gel (30 g)
and then eluted with more Et2O (500 mL) (entry 11). In entry
12, the combined organic extracts were mixed with 12 g of
silica gel and the resulting slurry was further dried on a rotary
evaporator to yield a powder, which was poured onto the top
of a chromatographic column (4.5-cm diameter) filled with
silica gel (200 g). CHCl3/acetic acid (9:1) was then passed
through the column, and the fraction Rf 0.64 was collected.
After removal of the solvent under water pump vacuum,
n-pentane (50 mL) was added to the isolated solid product
repeatedly, and the solvent was evaporated after each addition
to remove acetic acid thoroughly. Finally, the product was
dried under reduced pressure. In entries 13 and 14, the
combined organic extracts were analyzed directly by GC after
silylation (BSTFA) at 60 °C using n-eicosane and n-tetra-
cosane, respectively, as the internal standard.
En tr ies 5 a n d 8. Diethyl ether (80 mL) was added under
stirring (in entry 8, the solid formed was dissolved). After the
unreacted hydrogen peroxide was decomposed as above, the
organic layer was separated and the aqueous one was salted
out (NaCl) and extracted with Et2O (3 × 80 mL). The ether
extracts were combined with the original organic layer, dried
(Na2SO4), and analyzed by GC after silylation (BSA) at 60 °C
using n-dodecane as the internal standard.
En tr y 6. 1,2-Dichloroethane (DCE, 80 mL) was added
under stirring to dissolve the solid formed. The resulting two-
phase mixture was poured into a separatory funnel, more DCE
(120 mL) was added, and the organic layer was separated. Into
this phase a 10% Na2CO3 solution (60 mL) was dropped quickly
with stirring to precipitate the sodium salt, which was
separated, made thoroughly solvent-free, and dissolved in hot
water (600 mL). The aqueous solution was stirred, acidified
with dilute HCl, cooled, and extracted with n-pentane (4 × 80
mL), and the extract was then worked up as in entries 1 and
2.
Ma ter ia ls. Catalyst complex methyltrioctylammonium tet-
rakis(oxodiperoxotungsto)phosphate (Ia ) was prepared accord-
ing to the reported procedure.13 All olefins used in this work
were commercial products (Fluka, Aldrich, or Ega-Chemie) and
were purified by distillation before use. Hydrogen peroxide
(40% w/v aqueous solution, Fluka) was used as such or after
appropriate dilution. Noncommercial epoxides tested in this
work were made by the standard method from the correspond-
ing olefins and m-CPBA. Specimens for the comparison of
noncommercial vic-diols63 and R-ketols64 involved in this work
as well as of δ-acetoxypentanoic acid,65 4-oxooctanedioic acid,66
4-hydroxyoctanedioic acid γ-lactone,66 and 1,4-cyclooctanedi-
one67 were prepared according to literature procedures. Speci-
mens of 1-hydroperoxy-2-octanol (5IIa ) and -decanol (5IIb) free
of the respective regioisomer I were prepared from the corre-
sponding epoxides and alkaline hydrogen peroxide adapting
the procedure described in the literature26 for analogous
1-hydroperoxy-2-hydroxy derivatives. The epoxide (2a or 2b,
30 mmol), dissolved in a solution (230 mL) of methanol/water
(11:1) containing 7 g of 85% KOH, was allowed to react under
stirring for 7 h at 30 °C with 40% w/v aqueous hydrogen
peroxide (80 mL). At the end, the solution was poured into
water (1200 mL), salted out with (NH4)2SO4 (30 g), and
extracted with Et2O (4 × 100 mL). The combined ether
extracts were dried (Na2SO4), filtered, and evaporated, and
the residue was chromatographed over silica gel (CH2Cl2/
EtOAc 7:3). From the product thus isolated (62-63% yield),
containing ca. 5% of the undesired regioisomer, a pure sample
of 5IIa (as a viscous oil) or 5IIb (as a white solid, mp 33-34
°C, from Et2O/n-hexane at -25 °C) could be obtained after
further elutions (acetone/n-hexane 3:7).
Oxid a t ive Clea va ge of Olefin s t o Ca r b oxylic Acid s.
Gen er a l P r oced u r e. A three-necked round-bottom flask
equipped with a Teflon-coated magnetic bar stirrer, thermom-
eter, reflux condenser, and dropping funnel was charged with
methyltrioctylammonium tetrakis(oxodiperoxotungsto)phosphate
(Ia )68 (0.5 or 0.6 mmol, see Table 1) and the appropriate olefin
(50 mmol). The vigorously stirred mixture was warmed to 60-
70 °C (30-40 °C in entry 9), and then a 40% w/v aqueous
hydrogen peroxide solution [150, 200, or 250 mmol depending
on the substrate,19 with (except in entry 15) an additional 10%
to 37.5% molar excess (see Table 1)] was dropped into the
reaction mixture cautiously, at such a rate as to keep the
temperature below 80 °C. After the initial exothermicity had
subsided, the hydrogen peroxide addition could proceed quickly
and was usually over within 15-20 min. The resulting two-
phase mixture was then heated to 85 °C for the appropriate
time (see Table 1). At the end, after being allowed to cool to
En tr y 9. Dichloromethane (20 mL) was added under
stirring, and the aqueous layer was separated, washed with
CH2Cl2 (2 × 5 mL), treated with concentrated HCl (ca. 1 mL)
and then Na2SO3 (1 g) to decompose the unreacted hydrogen
peroxide, salted out with (NH4)2SO4 (17 g), and extracted with
Et2O (4 × 50 mL). The combined and dried (Na2SO4) ether
extracts were concentrated to 50 mL, passed through a short
column of silica gel as in entry 11, and then eluted with more
Et2O (250 mL). On evaporation of the solvent, the desired acid
was obtained as a white solid.
(62) (a) Nonsilylated secondary R-hydroperoxy ketones are known
to easily form R-diketones by dehydration, see: Bayley, E. J .; Barton,
D. H. R.; Elks, J .; Templeton, J . F. J . Chem. Soc. 1962, 1578 and
references therein. (b) Poole, C. F. In Handbook of Derivatives for
Chromatography; Blau, K., King, G. S., Eds.; Heyden & Son Ltd.:
London, 1978; Chapter 4, pp 160-165.
(63) (a) Swern, D.; Billen, G. N.; Findley, T. W.; Scanlan, J . T. J .
Am. Chem. Soc. 1945, 67, 1786. (b) Fieser, M.; Fieser, L. F. Reagents
for Organic Synthesis; Wiley: New York, 1967; Vol. 1, p 796.
(64) (a) Tsuji, T. Tetrahedron Lett. 1966, 22, 2413. (b) Srinivasan,
N. S.; Lee, D. G. Synthesis 1979, 520. (c) Sakata, Y.; Ishii, Y. J . Org.
Chem. 1991, 56, 6233.
(65) Andreev, V. M.; Polyakova, S. G.; Bazhulina, V. I.; Khrustova,
Z. S.; Smirnova, V. V. J . Org. Chem. USSR (Engl. Transl.) 1981, 17,
79.
(66) Holmquist, H. E.; Marsh, F. D.; Sauer, J . C.; Engelhardt, V. A.
J . Am. Chem. Soc. 1959, 81, 3681.
En tr ies 10 a n d 15. 1,2-Dichloroethane (1 × 20 mL, 2 × 5
mL) was added, and after each addition, the two-phase mixture
was stirred for a few minutes and then the lower organic layer
was siphoned off.69 The aqueous layer was stored in
a
refrigerator (5 °C) overnight to afford a crystalline precipitate,
which was filtered, washed with CH2Cl2 (3 × 5 mL), and dried.
(67) Cope, A. C.; Fenton, S. W.; Spencer, C. F. J . Am. Chem. Soc.
1952, 74, 5884.
(68) Due to its syrupy nature, complex Ia was preferably employed
as a dichloromethane solution of known concentration, with the solvent
subsequently removed.
(69) In entry 10, small amounts (ca. 0.2 g) of product separated from
the siphoned organic layer by standing. They were collected, washed
with CH2Cl2, and combined with the crystals precipitated from the
aqueous layer.

Oxidative Cleavage of Olefins to Carboxylic Acids
J . Org. Chem., Vol. 63, No. 21, 1998 7203
In a method similar to that for the aqueous layer in entry 9,
the mother liquor was extracted with Et2O (4 × 30 mL). On
evaporation of the dried (Na2SO4) ether extracts, a solid (pasty,
in entry 15) residue was obtained, which was washed with
CH2Cl2 (3 × 3 mL) to give a second crop (0.3-0.4 g) of the
desired acid as a white powder. In entry 15, the major and
minor crop of the product so obtained were combined and
dissolved in boiling CH3CN (20 mL), from which the pure acid
recrystallyzed on standing at -25 °C.
ing the actual reaction conditions), a somewhat modified
oxidation procedure (cf. also the text) was used to obtain the
desired compounds to a more satisfactory extent. It implied
(a) with epoxides trans-2c,d the use of 70% w/w H2O2 (in a
2:1 molar ratio to the epoxide), directly charged in the reactor,
at 60 °C (reaction time, 75 min; 97-99% conversion of the
epoxide; 55-58% conversion of H2O2) and (b) with epoxides
2e,f the dropwise addition of 40% w/v H2O2 (in a 5:1 molar
ratio to the epoxide) over 5-6 min into the epoxide/catalyst
Ia /water mixture (cooled with an ice/water bath), which was
then kept under stirring for 30 min at room temperature (96-
97% conversion of the epoxide; 5-6% conversion of H2O2).
To isolate 5a -f, the above-mentioned organic extracts were
concentrated under reduced pressure and the residue was
chromatographed on SiO2 (50 g, ethyl acetate/CH2Cl2 as an
eluent). The fraction of appropriate Rf was collected and
further purified on SiO2, eluting with acetone (Et2O)/n-hexane
(for details, see the individual compounds). Careful rotoevapo-
ration (30-40 °C) of the solvent, at first under slightly reduced
pressure and then at 1-2 mmHg (to remove the last traces of
solvents and moisture), gave the desired product, 95-97% pure
by GC. In the case of compounds 5a ,b, prior to the second
elution the crude product obtained was dissolved into Et2O
(15 mL) and stirred with a solution of Na2CO3 (80-120 mg)
in a minimum of water at rt for 15-20 min to remove residual
contamination by organic acids as evidenced by GC. Yields;
chromatography eluents; and Rf, spectral, physical, and ana-
lytical data of the isolated products 5a -f are as follows.
2(1)-Hyd r op er oxy-1(2)-octa n ols (5a ) were isolated as a
colorless oil in 35% yield (ethyl acetate/CH2Cl2 3:7 and then
acetone/n-hexane 3:7 as eluents; Rf 0.47 and 0.38). The two
regioisomers 5Ia and 5IIa (ratio ca. 2.7:1)76 could not be
separated. Their NMR data were assigned from the mixture.
The signals for 5IIa were confirmed on the pure compound
obtained by an independent synthesis (see above).
6-Oxoh ep ta n oic Acid : low-melting solid, mp 32-33 °C
1
(Et2O/n-pentane at -25 °C) [lit.70 mp 33-34 °C]; H and 13C
NMR spectra7 and the mass spectrum (as methyl ester)71 were
in accord with published data.
5-Oxod eca n ed ioic Acid : mp 115-116 °C (DCE) [lit.72 mp
115-116 °C]; GC-MS (as dimethyl ester) m/z (rel intensity)
(EI) 244 (1, M+), 213 (3), 181 (18), 143 (20, [CO(CH2)4-
COOCH3]+), 129 (28, [CO(CH2)3COOCH3]+), 111 (48), 101 (40),
59 (79), 55 (100), 41 (44); 1H and 13C NMR spectra were in
accord with published data.73
In ter r u p ted Oxid a tion of Ep oxid es (or th eir P a r en t
Alk en es) w ith Ia /H2O2 (40% w /v). F or m a tion of â-Hy-
d r op er oxy Alcoh ols. The reaction is illustrated by the
following examples referred to the examined model epoxides
2a -f. Into a stirred mixture of catalyst Ia (0.226 g, 0.1 mmol)
and 1,2-epoxyoctane (2a ) (97% pure; 1.321 g, 10 mmol),
warmed to 60-70 °C, was dropped 40% w/v aqueous hydrogen
peroxide (4.25 mL, 50 mmol)74 over a period of 5-6 min, with
care taken to control the initial exothermicity. The resulting
two-phase mixture was then heated at 85 °C for 30 min (almost
quantitative conversion of the epoxide, ca. 20% conversion of
H2O2). After the mixture cooled, Et2O (15 mL) was added with
stirring, the organic layer was separated, and the aqueous
layer was salted out with (NH4)2SO4 (2(3 g) and then extracted
with Et2O (2 × 15 mL). The combined, dried (Na2SO4), and
filtered organic layers were analyzed, after silylation (BSTFA),
by GC using n-tetradecane as an internal standard.
The same procedure was followed with 1,2-epoxydecane (2b),
trans-4,5- and -5,6-epoxydecane (2c and 2d ), and cyclohexene
and styrene oxides (2e and 2f), except that with the last two
compounds water (1 mL) was added initially to the epoxide/
catalyst Ia mixture in order to achieve better thermal control
during hydrogen peroxide addition (cf. also footnote d of Table
1) and the reaction time was shortened to 10 min, while it
was prolonged to 1 h with epoxides trans-2c,d . When the
reaction was stopped, epoxide conversion was 92-93% for 2b,
less than 30% for trans-2c,d , and almost quantitative for 2e,f,
and H2O2 conversion was 17-18% for 2b and 8-9% for 2c-f.
â-Hydroperoxy alcohol intermediates 5a -f were formed
along with the corresponding vic-diols/R-ketols and small or
moderate amounts of the final acids and related aldehydes (for
the pertinent data, cf. text). Compounds 5a -c were obtained
as a regioisomeric mixture. Where the two regioisomers
(silylated) could be revealed by GC as distinct peaks (5a ,b),
regioisomer I was found to elute first.
5Ia : 1H NMR (DMSO-d6) δ 0.86 (t, 3 H), 1.05-1.75 (m, 10
H), 3.46 (ddd after OH proton decoupling, J ) 11.4, 5.0, 4.7
Hz, 2 H, CH2OH), 3.58-3.82 (m, 1 H, CHOOH), 4.48 (t, J )
5.7 Hz, 1 H, CH2OH), 11.21 (s, 1 H, CHOOH) ppm; 13C NMR
(CDCl3) δ 14.08, 22.65, 25.68, 28.72, 29.37, 31.77, 62.95, 85.86
ppm; GC-MS (as bis(TMS) derivative) m/z (rel intensity) (CI)
307 (100, MH+), (EI) M+ absent, 201 (7), 187 (28), 147 (100,
[Me2Si)O-SiMe3]+), 103 (62, [CH2OSiMe3]+), 89 (50), 73 (57,
[SiMe3]+).
5IIa : 1H NMR (DMSO-d6) δ 0.86 (t, 3 H), 1.05-1.75 (m,
10 H), 3.58-3.82 (m, 3 H, CH2OOH + CHOH), 4.55 (d, J )
5.0 Hz, 1 H, CHOH), 11.58 (s, 1 H, CH2OOH) ppm; 13C NMR
(CDCl3) δ 14.08, 22.65, 25.46, 29.37, 31.79, 32.96, 69.60, 80.79
ppm; GC-MS (as bis(TMS) derivative) m/z (rel intensity) (CI)
307 (100, MH+), (EI) M+ absent, 261 (6), 187 (100, [C6H13
CHOSiMe3]+), 147 (83), 73 (34).
-
5a : IR (neat) νmax 3321, 1075 (sh), 1052, 1039, 816 cm-1
.
2(1)-Hyd r op er oxy-1(2)-d eca n ols (5b) were isolated as a
colorless oil in 45% yield (ethyl acetate/CH2Cl2 3:7 and then
acetone/n-hexane 3:7 as eluents; Rf 0.53 and 0.48). The two
regioisomers 5Ib and 5IIb (ratio ca. 2:1)76 could not be
separated. Their NMR data were assigned from the mixture.
The signals for 5IIb were confirmed on the pure compound
obtained by an independent synthesis (see above).
Interrupted-oxidation experiments with the parent alkenes
1a -f (in place of epoxides 2a -f) were carried out under
identical conditions. When the reaction was stopped, alkene
conversion ranged from 90% with 1a ,f to 96% with 1b and to
98-99% with 1c-e, whereas the unreacted (on converted
olefin) intermediate epoxide was around 1% for 2a , 48% for
2b, 78-80% for 2c,d , and absent for 2e,f.
Isolation of â-hydroperoxy alcohols 5a -f (formed from
epoxides) was accomplished on the organic extracts resulting
from workup75 of the reaction as described below. However,
in the case of compounds 5c-f, for which an adequate
accumulation of product (to be isolated) was not possible with
the interrupted-oxidation procedure reported above (parallel-
5Ib: 1H NMR (DMSO-d6) as for 5Ia except the signal at
δ 1.05-1.75 ppm corresponds to 14 H; 13C NMR (CDCl3) δ
14.11, 22.72, 25.74, 28.73, 29.35, 29.56, 29.75, 31.94, 62.92,
85.83 ppm; GC-MS (as bis(TMS) derivative) m/z (rel intensity)
(CI) 335 (100, MH+), (EI) M+ absent, 229 (3), 215 (22), 147
(100), 103 (40), 89 (27), 73 (32).
5IIb : 1H NMR (DMSO-d6) as for 5IIa except the signal
at δ 1.05-1.75 ppm corresponds to 14 H; 13C NMR (CDCl3) δ
(70) Acheson, R. M. J . Chem. Soc. 1956, 4232.
(75) While preferred as a solvent for analytical purposes, Et2O was
replaced with CH2Cl2 to achieve the isolation of compounds 5a -f. This
change made it possible to minimize the extraction of hydrogen
peroxide into the organic phase, which was found to complicate the
separation. Furthermore, except for 5e, salting out of the aqueous layer
with (NH4)2SO4 prior to extraction with CH2Cl2 was omitted.
(76) Determined by GC (after silylation) and NMR.
(71) Thoma, H.; Spiteller, G. Liebigs Ann. Chem. 1983, 1237.
(72) Leonard, N. J .; Goode, W. E. J . Am. Chem. Soc. 1950, 72, 5404.
(73) Osowska-Pacewicka, K.; Alper, H. J . Org. Chem. 1988, 53, 808.
(74) The high H2O2:substrate molar ratio (5:1) used in these experi-
ments was chosen to simulate the reaction conditions (volume ratio of
the organic versus aqueous phase) as close as possible.

7204 J . Org. Chem., Vol. 63, No. 21, 1998
Antonelli et al.
14.11, 22.72, 25.52, 29.35, 29.56, 29.75, 31.94, 32.97, 69.55,
OSiMe3]+), 179 (24, [ArCHOSiMe3]+), 147 (100), 133 (16), 119
(4), 103 (29), 73 (81). Anal. Calcd for C8H10O3: C, 62.32; H,
6.54; (O2)2-, 20.76. Found: C, 61.95; H, 6.50; (O2)2-, 20.60.
Rea ction of â-Hyd r op er oxy Alcoh ols (s-OH) w ith Ia /
H2O2 a t Low Con ver sion of th e Su bstr a te. F or m a tion
of In ter m ed ia te r-Hyd r op er oxy Keton es. The procedure
is illustrated by the following examples referred to the
examined model compounds 5IIb and 5d . A stirred mixture
of catalyst Ia (0.01 mmol) and 1-hydroperoxy-2-decanol (5IIb)
or erythro-6-hydroperoxy-5-decanol (5d ) (1 mmol) was reacted
at 85 °C with 40% w/v hydrogen peroxide (5 mmol).74 The
progress of the reaction was monitored by GC or HPLC (see
text, section IIIAa). After 20 min, the maximum yield of the
corresponding R-hydroperoxy ketone was observed (78% for
6IIb and 82% for 6d , at 17% and 40% conversion of 5IIb and
5d , respectively). The amounts of this and the other (minor)
intermediates formed (aldehyde, vic-diol, and R-ketol, see text)
as well as of the final carboxylic acid and unconverted
â-hydroperoxy alcohol were quantitated by GC as follows. The
cooled two-phase reaction mixture was salted out with (NH4)2-
SO4 (1 g) and shaken with Et2O (3 × 2 mL). To the combined
ether extracts was added n-tetradecane (for 5IIb) or n-
dodecane (for 5d ) as the internal standard. The solution was
then diluted to a 10-mL volume with more Et2O and dried (Na2-
SO4). On an aliquot (1 mL) of this solution, after treatment
with BSTFA-pyridine, the products vic-diol, R-ketol, and
carboxylic acid were determined as their silyl derivatives.
Another aliquot was cooled to 4 °C, and an excess of triph-
enylphosphine was added, whereupon the evaluation of the
aldehyde was accomplished. The same aliquot was then
treated with BSTFA-pyridine, and the R-hydroperoxy ketone
and unconverted â-hydroperoxy alcohol were determined as
ketol and diol (silylated), respectively. In the case of 6d , the
resulting R-ketol was analyzed in part as the bis(silyl) deriva-
tive (enol-TMS-ether-TMS-ether).62
80.78 ppm; GC-MS (as bis(TMS) derivative) m/z (rel intensity)
(CI) 335 (100, MH+), (EI) M+ absent, 289 (6), 215 (100, [C8H17
CHOSiMe3]+), 191 (27), 147 (82), 73 (27).
-
(4RS,5SR)-5(4)-Hyd r op er oxy-4(5)-d eca n ols (5c) were
isolated as a colorless oil in 63% yield (ethyl acetate/CH2Cl2
1:4 and then acetone/n-hexane 3:7 as eluents; Rf 0.57 and 0.53).
The two regioisomers 5Ic and 5IIc (ratio 1:1)77 could not be
separated. 1H NMR (DMSO-d6) δ 0.70-1.02 (m, 6 H), 1.10-
1.65 (m, 12 H), 3.42-3.56 (m, 1 H, CHOOH), 3.56-3.76 (m, 1
H, CHOH), 4.313 and 4.319 (d, J ) 5.8 Hz, CHOH) (1 H),
11.115 and 11.120 (s, CHOOH) (1 H) ppm; 13C NMR (CDCl3)
δ 14.04, 14.08, 14.10, 19.47, 19.58, 22.56, 22.62, 25.97, 26.09,
26.50, 28.58, 31.72, 31.92, 33.84, 71.68, 71.96, 88.05, 88.39
ppm; GC-MS (as bis(TMS) derivative) 5Ic, m/z (rel intensity)
(CI) 335 (100, MH+) (EI) M+ absent, 147 (85), 145 (95, [C3H7-
CHOSiMe3]+), 73 (100), 5IIc, m/z (CI) 335 (100, MH+) (EI) M+
absent, 173 (78, [C5H11CHOSiMe3]+), 147 (82), 73 (100). The
configuration of 5c was established as follows. A sample of
5c (96% pure; 0.198 g, 1 mmol) in Et2O (10 mL) was stirred
for 15 min at rt with a solution of (C6H5)3P (0.341 g, 1.3 mmol)
in Et2O (15 mL). Elution on an SiO2 column (15 g, Et2O as
an eluent, Rf 0.78) gave a 90% yield of erythro-4,5-decanediol,
mp 115-117 °C (MeCN) [lit.33b mp 118-118.5 °C].
er yth r o-6-Hyd r op er oxy-5-d eca n ol (5d ) was isolated as
a colorless oil78 in 61% yield (ethyl acetate/CH2Cl2 1:4 and then
acetone/n-hexane 3:7 as eluents; Rf 0.59 and 0.57): 1H NMR
(DMSO-d6) δ 0.87 (two overlapping t, 6 H), 1.10-1.80 (m, 12
H), 3.48 (five-line m, 1 H), 3.55-3.70 (m, simplifies after OH
proton decoupling, 1 H), 4.32 (d, J ) 5.9 Hz, 1 H), 11.11 (s, 1
H) ppm; 13C NMR (CDCl3) δ 13.93, 14.00, 22.75, 26.18, 28.46,
28.55, 31.41, 71.91, 88.33 ppm; GC-MS (as bis(TMS) deriva-
tive) m/z (rel intensity) (CI) 335 (100, MH+) (EI) M+ absent,
159 (100, [C4H9CHOSiMe3]+), 147 (73), 73 (28). Reduction of
5d with (C6H5)3P and workup as described above for 5c (Et2O
as eluent, Rf 0.80) gave meso-5,6-decanediol, mp 134-136 °C
(MeCN) [lit.33b mp 136-137.5 °C].
1-Hyd r op er oxy-2-d eca n on e (6IIb), obtained from (5IIb)
as indicated above, was identified in the reaction mixture by
its GC-mass spectrum: m/z (rel intensity) (CI) 189 (75, MH+),
171 (100, [MH - H2O]+) (EI) M+ absent, 143 (18), 142 (6), 141
(5), 125 (7), 83 (42), 75 (11, [COCH2OOH]+), 69 (100), 47 (19,
[CH2OOH]+), 45 (30). The identity of 6IIb was confirmed by
LC/TSP-MS analysis which showed the presence of a [M +
NH4]+ adduct ion at m/z 206 in the positive-ion mode and a
[M + AcO]- adduct ion at m/z 247 in the negative-ion mode.
After treatment of the organic solution containing 6IIb with
(C6H5)3P, the peak corresponding to this compound disap-
peared in GC or HPLC and a new peak was observed whose
GC-mass spectrum was identical to that of an authentic
sample of 1-hydroxy-2-decanone. Alternatively, addition of
BSTFA-pyridine to the above organic solution led to the
formation of a product identified as the 1-trimethylsiloxy-1-
N-(trimethylsilyl)trifluoroiminoacetoxy-2-decanone adduct C8H17-
COCH(OSiMe3)OC(dNSiMe3)CF3 (see text) by its GC-mass
spectrum: m/z (rel intensity) (CI) 428 (60, MH+), 356 (30), 338
tr a n s-2-Hyd r op er oxycycloh exa n ol (5e) was isolated as
a white solid in 21% yield (ethyl acetate/CH2Cl2 1:1 and then
Et2O/n-hexane 4:1 as eluents; Rf 0.53 and 0.46): mp 34-36
1
°C (Et2O/n-pentane at -25 °C); H NMR (DMSO-d6) δ 0.90-
1.45 (m, 4 H), 1.45-1.68 (m, 2 H), 1.68-1.85 (m, 1 H), 1.85-
2.10 (m, 1 H), 3.32-3.48 (m, simplifies after OH proton
decoupling, 1 H), 3.48-3.60 (m, 1 H), 4.66 (d, J ) 4.2 Hz, 1
H), 11.25 (s, 1 H) ppm; 13C NMR (CDCl3) δ 23.87, 24.12, 28.49,
32.87, 72.80, 88.50 ppm; GC-MS (as bis(TMS) derivative] m/z
(rel intensity) (CI) 277 (100, MH+) (EI) M+ absent, 187 (19,
[CyOOSiMe3]+), 171 (17, [CyOSiMe3]+), 147 (82), 97 (25), 75
(48), 73 (100). Anal. Calcd for C6H12O3: C, 54.53; H, 9.15;
(O2)2-, 24.22. Found: C, 54.78; H, 9.30; (O2)2-, 23.96. Reduc-
tion of 5e with (C6H5)3P and workup as described above for
5c (Et2O as an eluent, Rf 0.28) gave trans-cyclohexanediol, mp
104-105 °C (C6H6) [lit.79 mp 104.5-105.5 °C].
(100) (EI) M+ absent, 412 (3), 303 (8), 286 (12, [M - C8H17
-
2-Hyd r op er oxy-2-p h en yleth a n ol (5If) was isolated as a
white solid in 22% yield (ethyl acetate/CH2Cl2 1:1 and then
Et2O/n-hexane 4:1 as eluents; Rf 0.54 and 0.48): mp 70-71
CO]+), 168 (20, [CF3C(NSiMe3]+), 147 (10), 102 (40), 73 (100).
In the case of the R-hydroperoxy ketone 6d , identification
was performed on the compound isolated from a scaled-up
experiment operating as follows.
1
°C (Et2O/n-pentane at -25 °C) [lit.26 mp 71-73 °C]; H NMR
(DMSO-d6) δ 3.56 (ddd after OH proton decoupling, J ) 11.9,
7.0, 4.5 Hz, 2 H), 4.83 (dd, J ) 7.0, 4.5 Hz, 1 H), 4.91 (t, J )
5.9 Hz, 1 H), 7.27-7.34 (m, 5 H, ArH), 11.61 (s, 1 H) ppm, the
1H NMR data of 5If previously described26 were obtained in
CDCl3 and not in DMSO-d6 as here reported; 13C NMR
(CDCl3) δ 64.63, 88.43, 127.07, 128.59, 128.63, 137.00 ppm;
GC-MS (as bis(TMS) derivative) m/z (rel intensity) (EI) M+
absent,80 209 (1), 195 (3, [ArCHOOSiMe3]+), 193 (14, [ArCHCH2-
6-Hyd r op er oxy-5-d eca n on e (6d ). A stirred mixture of
erythro-6-hydroperoxy-5-decanol (5d ) (96% pure, 0.99 g, 5
mmol), catalyst Ia (0.113 g, 0.05 mmol), and 40% w/v hydrogen
peroxide (2.15 mL, 25 mmol)74 was heated at 85 °C for 30 min
(ca. 13% conversion of H2O2). After the mixture cooled, CH2-
Cl2 (8 mL) was added with stirring, the organic layer was
separated, and the aqueous layer was extracted with CH2Cl2
(2 × 15 mL). The combined, dried (Na2SO4), and filtered
organic layers were concentrated under reduced pressure, the
residue was chromatographed on SiO2 (Et2O/n-hexane 1:1),
and the fractions Rf 0.75 and Rf 0.42 were collected. From
the latter fraction, 0.615 g of unreacted 5d (90% pure) was
(77) Determined by NMR.
(78) 1,1-Dihydroperoxypentane was usually revealed as a contami-
nant by the 1H NMR spectrum. Its formation was confirmed by an
independent preparation from pentanal (7d ) and hydrogen peroxide:
1H NMR (DMSO-d6) (selected values) δ 4.98 (t, J ) 6.0 Hz, 1 H,
CH(OOH)2, 11.578 (s, 2 H, OOH) ppm; GC-MS (as bis(TMS) deriva-
tive) m/z (rel intensity) (CI-NH3) 298 (94, [M + NH4]+), 208 (13), 152
(100) (EI) M+ absent, 175 (14, [M - OOSiMe3]+), 147 (6), 119 (100,
[CH2OOSiMe3]+), 91 (67, [HOOSiMe2]+), 75 (60), 61 (38).
(79) Rigby, W. J . Chem. Soc. 1950, 1911.
(80) The parent peak could be observed on the mass spectrum of
the compound as recorded by direct inlet (DIP-EI): MS m/z (rel
intensity) 154 (3, M+), 138 (1), 136 (2.5), 123 (52), 121 (65), 106 (63),
105 (100), 77 (58).

Oxidative Cleavage of Olefins to Carboxylic Acids
J . Org. Chem., Vol. 63, No. 21, 1998 7205
recovered (58%). The former fraction was eluted once more
with acetone/n-hexane 1:9 (Rf 0.24). Careful rotoevaporation
of the solvent, at first under slightly reduced pressure and then
at 1-2 mmHg (to remove the last traces of solvents and
moisture), afforded 0.307 g of 6-hydroperoxy-5-decanone (6d )
as a colorless oil, 94% pure by GC81 (75% yield based on
presence (along with the unreacted starting compounds) of two
major peaks, roughly estimated at 34% on the whole on the
basis of charged 5d , in a 1.6-1.8:1 ratio (attributable to
stereoisomers) was observed, whose GC-mass spectra were
consistent with the structure of the title adduct: GC-MS (as
bis(TMS) derivative) m/z (rel intensity) (CI-NH3) 438 [25, [M
+ NH4]+), (CI-i-C4H10) MH+ absent, 245 (75, [Me3SiOCH-
(C4H9)CH(C4H9)O]+), 175 (100, [C4H9CH(OSiMe3)O]+), 159 (53,
[C4H9CHOSiMe3]+) (EI) M+ absent, 229 (2.5, [Me3SiOCH-
(C4H9)CH(C4H9)]+), 191 (1.5, [C4H9CH(OSiMe3)OO]+), 159
(100), 103 (50), 75 (70), 73 (63). In a duplicate experiment,
the stirred mixture of 5d and 7d was heated under helium
(85 °C, 90 min). Pentanoic acid (8d ) and 6-hydroxy-5-decanone
(4d ) were formed (molar ratio 8d :4d ) 2.8:1),84 and evolution
of H2 was observed. It was checked that, operating under
these conditions but in the presence of the sole aldehyde 7d ,
the acid 8d was not formed.
As a proof of the chemical structure of 9 (R2, R3, R4 ) C4H9),
in another experiment 65% HClO4 (10 µL) was added to the
reaction mixture containing the adduct and cyclization in situ
of the latter to the corresponding 1,2,4-trioxane derivative was
effected (35 °C, 24 h).85 The formed product was dissolved in
Et2O, and the resultant solution was stirred (30 min) with solid
NaHCO3 and some droplets of water (to neutralize the acidity
present), filtered, and eluted on an SiO2 column (acetone/n-
hexane 1:9) to afford 0.203 g of 3,5,6-tributyl-1,2,4-trioxane
(ca. 94:6 mixture of stereoisomers) as a colorless oil, 96% pure
by GC (70% yield): 1H NMR (CDCl3) δ 0.80-1.05 (m, 9 H),
1.16-1.47 (m, 12 H), 1.47-1.66 (m, 6 H), 3.64 (dt, J ) 10.6,
3.1 Hz, 1 H), 3.80-4.05 (m, 1 H), 5.195 and 5.34 (t, J ) 5.4
Hz) (1 H) ppm, the signal at δ 5.34 ppm integrates to 1:15.3
the area of the signal at δ 5.195 ppm; 13C NMR (DMSO-d6) δ
13.66, 13.76, 13.89, 21.83, 21.89, 21.95, 24.06, 25.23, 26.87,
27.05, 29.93, 31.40, 76.20, 80.66, 104.05 ppm; GC-MS m/z (rel
intensity) (CI) 259 (100, MH+), 173 (19), 157 (37), 139 (24).
Hyd r oxym eth yl 2-Hyd r oxyd ecyl P er oxid e (Ad d u ct 9;
R2, R4 ) H, R3 ) n -C8H17). 1-Hydroperoxy-2-decanol (5IIb)
(97% pure, 0.168 g, 1 mmol) was stirred for 30 min at 25 °C
with 40% w/v formaldehyde (75 µL, 1 mmol) under an
atmosphere of helium. At the end, Et2O was added. The
resultant solution was dried (Na2SO4), filtered, and analyzed
by GC after silylation (BSTFA) using n-tetradecane as an
internal standard. It showed the presence (along with the
starting â-hydroperoxy alcohol) of a major peak, roughly
estimated at 39% on charged 5IIb, whose GC-mass spectrum
was consistent with the structure of the title adduct: GC-
MS (as bis(TMS) derivative) m/z (rel intensity) (CI) 365 (9,
MH+), 247 (86), 245 (56, [C8H17CH(OSiMe3)CH2O]+), 229 (100,
[C8H17CH(OSiMe3)CH2]+), 119 (70, [OCH2OSiMe3]+) (EI) M+
absent, 229 (9), 215 (8), 147 (16), 103 (100, [CH2OSiMe3]+), 75
(76), 73 (61).
consumed 5d ): IR (neat) νmax 3379, 1712 cm-1 1H NMR
;
(DMSO-d6) δ 0.80-0.92 (m, 6 H), 1.10-1.40 (m, 4 H), 1.40-
1.70 (m, 6 H), 2.61 (ddt, J ) 18.1, 7.3 Hz, 2 H), 4.12 (dd, J )
7.4, 6.3 Hz, 1 H), 12.01 (s, 1 H) ppm; 13C NMR (CDCl3) δ 13.76,
13.86, 22.37, 22.46, 25.32, 27.56, 29.18, 37.77, 89.66, 212.26
ppm; GC-MS m/z (rel intensity) (CI) 189 (100, MH+) (EI) M+
absent, 170 (9, [M - H2O]+), 103 (1.5), 99 (1.4), 85 (100, [C4H9-
CO]+), 57 (85), 44 (45), 41 (65). As additional chemical
structure proof, reduction of 6d with (C6H5)3P (and relative
workup) as described above for 5c (acetone/n-hexane 1:9 as
an eluent; Rf 0.40) afforded the expected 6-hydroxy-5-decanone
(4d ) as an oil in 80% yield. IR, 1H NMR, and mass spectra
were consistent with published data.64b
â-Hyd r op er oxy Alcoh ol Decom p osition in th e P r es-
en ce of Wa ter a n d Ca ta lyst Ia (Bla n k Exp er im en ts). In
a typical experiment, a stirred mixture of 1-hydroperoxy-2-
decanol (5IIb) (97% pure, 0.168 g, 1 mmol) and catalyst Ia
(22.6 mg, 0.01 mmol) was heated for 45 min at 85 °C together
with water (425 µL) under an atmosphere of helium. At the
end, the reaction mixture was worked up and analyzed by GC
as described in the preceding experimentation. The formation
of 38% of nonanal (7b), 30.5% of 1,2-decanediol (3b), and 13%
of nonanoic acid (8b) was observed, at 34% conversion of 5IIb.
Likewise, the presence (roughly estimated at 7% of converted
5IIb) of hydroxymethyl 2-hydroxydecyl peroxide (adduct 9; R2,
R4 ) H, R3 ) n-C8H17; see the following heading) and (in a
small amount) of 1-hydroxynonyl 2-hydroxydecyl peroxide
82
(adduct 9; R2 ) H, R3, R4 ) n-C8H17
)
was ascertained.
In an analogous experiment with erythro-6-hydroperoxy-5-
decanol (5d ) (96% pure, 0.198 g, 1 mmol) in place of 5IIb, after
1 h (at 63% conversion of 5d ) GC analysis revealed the
formation of 124% of pentanal (7d ) (62% referred to decom-
posed 5d ), 8% of 5,6-decanediol (3d ), 14% of 6-hydroxy-5-
decanone (4d ), 22% of pentanoic acid (8d ) (11% referred to
decomposed 5d ), and approximately 2% of 1-hydroxypentyl
1-butyl-2-hydroxyhexyl peroxide (adduct 9; R2, R3, R4 ) n-C4H9;
see the following paragraph).83a
â-Hyd r oxyp er oxyh em ia ceta ls fr om â-Hyd r op er oxy Al-
coh ols a n d Ald eh yd es. 1-Hyd r oxyp en tyl 1-Bu tyl-2-h y-
d r oxyh exyl P er oxid e (Ad d u ct 9; R2, R3, R4 ) n -C4H9). A
mixture of erythro-6-hydroperoxy-5-decanol (5d ) (96% pure,
0.215 g, 1.08 mmol) and freshly distilled pentanal (7d ) (98%
pure, 0.095 g, 1.08 mmol) was stirred under an atmosphere of
helium for 10 min at 25 °C. At the end, Et2O was added and
the resultant solution was analyzed by GC after silylation
(BSTFA), using n-dodecane as an internal standard. The
If a 10-fold excess of 40% w/v formaldehyde (750 µL) was
used (30 °C, 2 h), 5(II)b could be completely converted, but
this essentially led to sizable amounts of the addition product
of the title adduct to formaldehyde.
(81) Purity was determined after converting 6d into the correspond-
ing R-ketol.
(82) Identified by its GC-mass spectrum: (as bis(TMS) derivative)
m/z (rel intensity) (CI) 477 (5, MH+), 245 (85, [C8H17CH(OSi-
Me3)CH2O]+), 231 (100, [C8H17CH(OSiMe3)O]+).
(83) (a) The formation of the acid 8d to a larger extent (22%) as
compared to the R-ketol 4d (14%), which in addition to resulting from
the decomposition of 5d is also coproduced together with 8d in the
thermal decomposition of the formed adduct 9 (R2, R3, R4 ) n-C4H9)
according to eq 2, leads us to hypothesize a catalytic activity for 4d .
The latter might react with 5d to give a new adduct which, on
decomposition, would afford one molecule of C5 aldehyde and one of
the corresponding acid, forming the R-ketol again with the evolution
of H2. Ketones are known to react with hydroperoxides to give
peroxyhemiketals and peroxyketals.54a,83b A similar behavior might also
be exhibited by 2-hydroxydecanal 4Ib (not detected), which is expected
to be coproduced in the decomposition of hydroxymethyl and 1-hy-
Acid ity of th e Aqu eou s P h a se of th e Ia /H2O2 Bip h a sic
System . In parallel to the reaction conditions, complex Ia
(0.12 mmol), dissolved in 1,2-dichloroethane (3 mL), was kept
in contact under stirring for 15 min at rt with 16% and 40%
w/v aqueous hydrogen peroxide (5.1 mL; 24 and 60 mmol,
respectively). The resulting aqueous phase was separated and
filtered (syringe filter, 25 mm; membrane, cellulose acetate;
pore size, 0.45 µm). A pHapp of 1.9 and 1.18 (0.1 M solution in
KCl), as measured by glass electrode, was observed at rt in
the two cases.86 (If the aqueous hydrogen peroxide was
replaced by an equal volume of water alone, the resulting pH
droxynonyl 2-hydroxydecyl peroxide adducts (9; R2, R4 ) H, R3
)
(84) The value of the ratio decidedly in favor of 8d parallels the trend
observed in the blank decomposition of 5d (see Experimental Section)
and discussed in ref 83a.
(85) For the formation of 1,2,4-trioxanes by the reaction of aldehydes
with â-hydroperoxy alcohols in the presence of acids or dehydrating
agents, see: (a) Swern, D. Organic Peroxides; Wiley-Interscience: New
York, 1970; Vol. 3, Chapter 2, pp 108-110. (b) Singh, C. Tetrahedron
Lett. 1990, 31, 69. (c) Refs 28 and 29.
n-C8H17 and R2 ) H, R3, R4 ) n-C8H17, respectively; cf. eq 2) formed in
the blank decomposition of 5IIb (see Experimental Section). In this
case, the reaction of 4Ib with 5IIb followed by decomposition of the
resultant adduct would give nonanal and formic acid, with the
formation of the R-hydroxy aldehyde again and the evolution of H2.
(b) Dickey, F. H.; Rust, F.; Vaughan, W. E. J . Am. Chem. Soc. 1949,
71, 1432.

7206 J . Org. Chem., Vol. 63, No. 21, 1998
Antonelli et al.
of the aqueous phase was 2.4 under the same conditions.) Ion
chromatography analysis of the aqueous phase (after treat-
ment with NaOH) revealed the presence of a quantity of
phosphate and tungstate ions amounting to 65-7587 and
1500-1600 mg/L, respectively. This solution was used in the
following experimentation (procedure B).
Ia -Ca ta lyzed Hyd r olysis of Ep oxya lk a n es u n d er Tw o-
P h a se Con d ition s (w ith ou t Ad d ed Solven t) in th e P r es-
en ce of Hyd r ogen P er oxid e of Differ en t Con cen tr a tion .
Pertinent experiments were performed on 1,2-epoxyoctane (2a )
and -decane (2b) following procedures A and B.
P r oced u r e A. The reaction was carried out by keeping
under stirring the two-phase mixture made up of the epoxy-
alkane (3 mmol), catalyst Ia (0.03 mmol), and 16% or 40% w/v
aqueous hydrogen peroxide (1.27 mL; 6 and 15 mmol, respec-
tively) at 70 °C (2a ) or 85 °C (2b) for 30 min.
P r oced u r e B. An aliquot (1.27 mL) of the acidic aqueous
phase of the Ia /H2O2 (16% or 40% w/v) biphasic system,
obtained as described in the preceding heading, was allowed
to react under stirring with the epoxyalkane (3 mmol) as in
procedure A.
u n d er Tw o-P h a se Con d ition s w ith Ad d ed Solven t (Com -
p a r a tive Exp er im en ts). In a typical experiment, 1-decene
(98% pure, 2.86 g, 20 mmol) was reacted for 6 h at 85 °C under
stirring with the catalyst Ia (0.542 g, 0.24 mmol)/40% w/v H2O2
(9.35 mL, 110 mmol) oxidizing system in the presence of
toluene (7.35 mL). At the end, after the usual workup the
phases were separated. The amount of unconverted hydrogen
peroxide was determined by iodometric titration of the aqueous
layer. The organic layer was analyzed after silylation (BSTFA)
by GC using n-tetradecane as an internal standard. Nonanoic
acid was formed in 55% yield at 84% conversion of hydrogen
peroxide (with 4-5% of the intermediate epoxide being unre-
acted), as compared to an isolated yield of 80% obtained in
the absence of the solvent (Table 1, entry 2). When tetrachlo-
roethane was used as a solvent under the same conditions,
the formed C9 acid dropped to 34% at 73% conversion of
hydrogen peroxide (with 33% of the epoxide being unreacted).
Analogously, upon oxidation with Ia (0.2 mmol)/40% w/v
H2O2 (88 mmol) in the presence of toluene (7.35 mL), trans-
4-decene (20 mmol) after 8 h gave 24% of butanoic and
hexanoic acid at 74% conversion of hydrogen peroxide, with
51% of the intermediate epoxide being unreacted (GC analysis,
n-dodecane as an internal standard), as compared to an 80%
and 77% yield of the same acids (after 6 h) obtained in the
absence of the solvent (Table 1, entry 5). If tetrachloroethane
was used as a solvent under the same conditions, the yield of
the C4 and C6 acids fell to 16% at 84% conversion of hydrogen
peroxide (with 70% of the epoxide being unreacted).
In both procedures, at the end of the reaction Et2O was
added under stirring and the aqueous phase was salted out.
The organic phase was separated, dried (Na2SO4), filtered, and
analyzed after silylation (BSTFA) by GC using n-tetradecane
as an internal standard. The data pertaining to these experi-
ments are tabulated in Table 2.88
Hyd r ogen P er oxid e Oxid a tion of Med iu m - a n d Lon g-
Ch a in Alk en es to Ca r boxylic Acid s Ca ta lyzed by Ia
(86) The pHapp of the 16% and 40% w/v hydrogen peroxide employed
in these experiments was 3.3 and 2.2 (free acid as H3PO4, 20.6 and 48
mg/L, respectively).
Ack n ow led gm en t. We thank Mr. G. C. Bacchilega
for the NMR spectra determination and Mr. G. F.
Guglielmetti for LC/TPS-MS analyses.
(87) This value does not include the quota due to the phosphoric
acid already contained in hydrogen peroxide itself (cf. ref 86).
(88) Through the use of procedure B (85 °C, 30 min) but with water
alone in place of aqueous hydrogen peroxide, from cyclohexene and
styrene oxides (2e and 2f) the corresponding 1,2-diols were easily
obtained in 85-86% yields (GC analysis). In contrast, from epoxyal-
kanes 2a ,b only small or negligible amounts of the respective 1,2-diols
were formed under the same conditions.
J O980481T
(90) Most of the glutaric and adipic acid formed (about 5% and 1%,
respectively) is found in the mother liquor from acid extraction.
(91) The GC-mass spectrum (as dimethyl ester) was consistent with
published data: Heller, J .; Yogev, A.; Dreiding, A. S. Helv. Chim. Acta
1972, 55, 1003.
(89) The GC-mass spectrum (as methyl ester) was consistent with
published data.71