Selective oxidation of cyclooctane to cyclootanone with molecular oxygen in the presence of compressed carbon dioxide

Article abstract of DOI:10.1039/b111212k

The oxidation of cyclooctane (1) to cyclooctanone (3) with molecular oxygen and acetaldehyde (2) as a co-reductant occurs efficiently in the presence of compressed CO₂. Up to 20% yields of 3 are obtained under optimised multiphase conditions.

Full text of DOI:10.1039/b111212k

Selective oxidation of cyclooctane to cyclootanone with molecular oxygen
in the presence of compressed carbon dioxide
Nils Theyssen and Walter Leitner*
Max-Planck-Institut für Kohlenforschung, Kaiser-Wilhelm-Platz 1 45470. Mülheim an der Ruhr, Germany.
E-mail: leitner@mpi-muelheim.mpg.de
Received (in Cambridge, UK) 7th December 2001, Accepted 11th January 2002
First published as an Advance Article on the web 1st February 2002
The oxidation of cyclooctane (1) to cyclooctanone (3) with
molecular oxygen and acetaldehyde (2) as a co-reductant
occurs efficiently in the presence of compressed CO2. Up to
20% yields of 3 are obtained under optimised multiphase
conditions.
high pressure IR-spectroscopy. † Fig. 1 shows the carbonyl
region of a typical experiment in the initial stage.
At t0, 2 (n˜CNO = 1735 cm21) starts to diffuse into the reactor
where it is converted to peracetic acid (n˜CNO = 1765 cm21) and
acetic acid 4 (n˜CNO = 1720 cm21) during the course of the
reaction. The CNO stretching frequency of 3 is obscured by
these strong bands. Reactions were allowed to proceed until a
constant intensity of the acetic acid band was reached. With this
new system very high conversions and selectivities of 3 could
be obtained in the absence of a catalyst with a maximum yield
of 3 up to 20% (Table 1).
The selective oxidation of hydrocarbons with molecular oxygen
is one of the great challenges for synthetic chemistry from an
economic as well as an ecological point of view.1 Compressed
carbon dioxide is a very attractive reaction medium for
oxidation reactions, because it allows the combination of high
concentrations of substrates and oxygen with the additional
safety of a totally inert environment. There are, however,
surprisingly few examples for selective oxidation reactions in
carbon dioxide media.2
Recently we reported the steel-promoted epoxidation of
olefins under Mukaiyama-type reaction conditions in super-
critical carbon dioxide.3 In the absence of an extra catalyst good
to excellent yields were observed for substrates with internal
double bonds and for long chain terminal olefins. Here, we
describe the application of a similar methodology to alkane
oxidation (Scheme 1).
The yield of ketone 3 obtained in the presence of carbon
dioxide compares favourably to the best values under conven-
tional conditions with related oxidation systems (11%).5,6
Conversion and especially selectivity are also significantly
higher than those reported for the oxidation of cyclohexane in
scCO2. Only 3% of cyclohexanone was achieved with an iron
porphyrin catalyst at 70 °C using 0.25 equivalents of acet-
aldehyde7 or by autoxidation in the absence of aldehyde at 160
°C.8
The use of aldehydes as co-reductants is a very well
established methodology in oxidation chemistry. Acetaldehyde
was used in the present study as the cheapest suitable co-
reductant. Cyclooctane was chosen as a model substrate
because all reactive positions are identical and the main
oxidation product, cyclooctanone, was proven to be inert
towards Baeyer–Villiger-oxidation under these reaction condi-
tions.4 The reactions were monitored by online time-resolved
Fig. 1 Time resolved ATR-IR spectra of cyclooctane oxidation in
compressed carbon dioxide showing the region of carbonyl stretching
frequencies (conditions as in entry 4).
Scheme 1 Idealised reaction scheme for the formation of 3.
Table 1 Oxidation of cyclooctane 1 with molecular oxygen in the presence of acetaldehyde 2a
Cyclooctane 1
Cyclooctanone 3
Cyclooctanol
Selectivityb (%)
Cyclooctane-1,4-dione
Selectivityb (%)
Entry
r (CO2)/g ml21 T/°C Time/h Conversionb (%)
Selectivityb (%)
Yieldb (%)
1
2
3
4
5
6
7c
0.62
0.58
0.18
0.18
0.18
0.18
0.18
42
52
42
52
70
90
70
36
24
46
26
10
9.3
10.4
23.8
27.6
30.9
33.7
37.3
61.5
64.5
66.0
66.9
62.1
57.6
59.6
5.7
6.7
2.1
2.9
7.1
8.5
6.2
5.9
4.1
23.5
18.2
8.0
15.7
18.4
19.2
19.5
22.3
7.8
10.9
11.9
14.2
1.4
10
a N (1+O2+2) = 1+2.1+2.0; for standard reaction conditions see note †. b Determined by GC analysis. Selectivities were calculated from the mole fractions
of the products. Total recoveries varied between 85 and 100%. Other typical by-products were 9-oxabicyclo(3.3.1)nonan-1-ol (isolated by preparative GC
and identified by 13C-NMRspectroscopy),9 cyclooctyl hydroperoxide and adipic acid. c N (1+O2+2) = 1+3.1+4.0.
410
CHEM. COMMUN., 2002, 410–411
This journal is © The Royal Society of Chemistry 2002

to 20% yield of cyclooctanone were obtained under optimised
multiphase conditions.
Notes and references
† Safety warning: The use of compressed gases and especially O2 in the
presence of organic substrates requires appropriate safety precautions and
must only be carried out using suitable equipment. In a typical experiment
the stainless steel reactor (V = 205 ml, austenitic steel No. 1.4571, teflon O-
seals) equipped with thick walled glass windows and an IR-ATR element
(Attenuated Total Reflection, ReactIR, Mettler Toledo) was filled with
cyclooctane (50 mmol), carbon dioxide and molecular oxygen in that order.
The amount of CO2 was weighed into the reactor to achieve the desired
density. Defined amounts of oxygen were added by a pressure difference of
at least 30 bars (resolution 1 bar) from a storage vessel with known volume.
The required pressure difference was calculated from the virial equation
including the second virial coefficient and the temperature of the storage
vessel. Acetaldehyde (100 mmol) was added last by diffusion (n˙ ~ 1 mmol
min21) from a completely filled dosing unit into the reactor at the desired
reaction temperature. When the absorbance at 1720 cm21 (including
carbonyl stretching frequencies of acetic acid and cyclooctyl compounds)
achieved constant values the reaction was stopped by cooling to rt followed
by slow depressurisation ( ~ 5 h) from the bottom of the reactor through a
cold trap of 250 °C. The two-phase mixture remaining in the cold trap was
homogenized by adding diethyl ether and analysed by GC-MS using n-
decane as internal standard. The FID-signals of cycloctanone (f = 1.230
0.009) and cyclooctanol (f = 1.246 0.015) were corrected according to
their exact response factors. The correction factors for 9-oxabicyclo-
(3.3.1)nonan-1-ol (f = 1.32), cyclooctane-1,4-dione (f = 1.40) and all other
found products (f = 1.40) were estimated. The contents of peroxides was
< 4% in all experiments according to iodometric analysis;14 the exact values
showed no correlation with conversion or reaction temperature.
Fig. 2 3-D-Simplex-optimisation in the high density regime. * Reaction
became heterogeneous during the reaction. # Experiment is not part of the
statistical algorithm.
Notably, other gases were considerably less efficient as inert
diluting agents as compared to CO2. Using largely identical
molar ratios of inert gas, oxygen and reagents, the yields of 3
were 10.7% for N2, 7.1% for SF6 and 4.3% for Ar under the
conditions of entry 3. Furthermore, our results indicate that the
density of carbon dioxide is instrumental to control the
reactivity and selectivity of the catalyst-free oxidation.
Fig. 2 shows a 3-D-Simplex-Optimisation10 of the relevant
reaction parameters temperature, oxygen-to-substrate ratio and
CO2 density. The 3-D-Simplex-optimisation shows a trend to
improved yields with lower CO2-densities, lower oxygen to
substrate ratio and higher reaction temperatures.
Further optimisation over a wider range of reaction condi-
tions indicated a particularly pronounced effect of the carbon
dioxide density (Table 1). Reducing the carbon dioxide density
from around 0.60 g cm23 to 0.18 g cm23 results in a two-phase
reaction system in all cases from the beginning.11 This emerged
as quite effective since the yield of 3 increased by a factor of 2.7
under otherwise identical reaction conditions (cf. Table 1,
entries 1/3 and 2/4). The conversion of 2 was complete,
however, in all cases. This indicates that the higher yield of 3
can be mainly attributed to a more efficient oxygen transfer
relative to the autoxidation of 2. This pronounced density effect
can be related to a more efficient interception of free radicals in
the liquid phase.8
The presence of a free radical pathway is strongly supported
by two observations. First of all the exclusive formation of
cyclooctane-1,4-dione as side product with two keto functions
is best ascribed to the intramolecular attack of a cyclooctyl
peroxo radical intermediate on the kinetically favoured 4-posi-
tion. Secondly the observed 1-adamantol to 2-adamantone ratio
of 4.5+1 for adamantane as substrate clearly indicates the
operation of a radical rather than a metaloxo pathway6
(conditions similar to those of entry 4).
The radical chain is believed to be initiated by steel-promoted
formation of the peroxo acyl radical.3 H-Abstraction from 1
followed by addition of molecular oxygen leads to the
formation of the cyclooctyl peroxo radical. This species can
either react with another cyclooctyl peroxo radical to give an
equimolar amount of cyclooctanol and 3 or react with a peroxo
acyl radical to give 3 and peracetic acid. We favour the latter
pathway on basis of the high ketone to alcohol ratio12 and the
observed influence of the reaction parameters.13
1 (a) A. E. Shilov and G. B. Shul’pin, Activation and Catalytic Reactions
of Saturated Hydrocarbons in the Presence of Metal Complexes,
Kluwer Academic Publishers, 2000; (b) J. M. Thomas, R. Raja, G.
Sankar and R. G. Bell, Nature, 1999, 398, 227; (c) A. K. Suresh and T.
Sridhar, Ind. Eng. Chem. Res., 2000, 39, 3589; (d) I. W. C. E. Arends
and R. A. Sheldon, Applied Catalysis A: General, 2001, 212, 175–187;
(e) U. Schuchardt, D. Cardoso, R. Sercheli, R. Pereira, R. S. da Cruz, M.
C. Guerreiro, D. Mandelli, E. V. Spinacé and E. L. Pires, Applied
Catalysis A: General, 2001, 211, 1.
2 (a) Chemical Synthesis Using Supercritical Fluids, eds. P. G. Jessop and
W. Leitner, Wiley-VCH, 1999; (b) D. A. Morgenstern, R. M.
LeLacheur, D. K. Morita, S. L. Borkowsky, S. Feng, G. H. Brown, L.
Luan, M. F. Gross, M. J. Burk and W. Tumas, in Green Chemistry, ACS
Symp. Ser. 626, eds. P. T. Anastas and T. C. Williamson, American
Chemical Society, Washington DC, 1996, 132; (c) A. Martin and B.
Kerler, Chem. Ing. Tech., 2000, 72, 382.
3 W. Leitner and F. Loeker, Chem. Eur. J., 2000, 6, 2011.
4 C. Palazzi, C. Bolm, G. Franciò and W. Leitner, manuscript in
preparation.
5 S.-I. Murahashi, Y. Oda and T. Naota, J. Am. Chem. Soc., 1992, 114,
7913.
6 A. M. Khenkin, A. Rosenberger and R. Neumann, J. Cat., 1999, 182,
82.
7 X.-W. Wu, Y. Oshima and S. Koda, Chem. Lett., 1997, 1045.
8 P. Srinivas and M. Mukhopadhyay, Ind. Eng. Chem. Res., 1997, 36,
2066.
9 A. J. Bloodworth, J. A. Khan and M. E. Loveitt, J. Chem. Soc., Perkin.
Trans. 1, 1981, 621.
10 M. Otto, Chemometrie—Statistik und Computereinsatz in der Analytik,
Wiley-VCH, 1997, 127.
11 Although the exact explosion limits of the system under scrutiny are not
known, the large amount of carbon dioxide can be expected to allow for
safe operation even in the low density regime. See for example B. Lewis
and G. von Elbe, Combustion, Flames and Explosions of Gases,
Academic Press, New York, 1951.
12 Oxidation of cyclooctanol to 3 by intermediate high concentrations of
peracetic acid can not be ruled out completely.
13 V. I. Timokhin, M. T. Lisovska and A. P. Pokutsa, Kinet. Catal. (Engl.
Transl.), 2000, 41, 159.
14 R. Alcantara, L. Canoira, P. G. Joao, J.-M. Santos and I. Vázquez,
Applied Catalysis A: General, 2000, 203, 259.
In summary, compressed carbon dioxide provides an inert
reaction medium for efficient and selective oxidation of
cyclooctane using O2–aldehyde mixtures. It is more effective
than other inert diluting gases under identical conditions and up
CHEM. COMMUN., 2002, 410–411
411