Increasing the ketone selectivity of the cobalt-catalyzed radical chain oxidation of cyclohexane

Article abstract of DOI:10.1002/1521-3765(20020816)8:16<3724::AID-CHEM3724>3.0.CO;2-W

A variety of heterogeneous catalysts for the radical chain oxidation of cyclohexane has been prepared by immobilization of the well-defined cobalt acetate oligomers [py₃Co₃(μ₃-O)(OH)(O₂ CCH₃)₅](PF₆) (1) and [py₄ Co₂(OH)₂(O₂CCH₃)₃] (PF₆) (2) on carboxy-modified mesoporous silica supports A-D by carboxylate exchange. The catalytic oxidation of cyclohexane with tert-butyl hydroperoxide (TBHP) in the presence of these homogeneous and immobilized cobalt acetate complexes afforded the corresponding alcohol and ketone in high yield. The immobilization of 1 and 2 results in a significant increase of catalytic activity. TBHP acts as a radical initiator and as source of molecular oxygen, which is also involved in the overall oxidation process. The rate of cyclohexane conversion is limited by the diffusion of molecular oxygen, and steady-state concentrations of cyclohexanone (K, ketone) and cyclohexanol (A, alcohol) are established; these determine the maximum K:A ratio.

Full text of DOI:10.1002/1521-3765(20020816)8:16<3724::AID-CHEM3724>3.0.CO;2-W

FULL PAPER
Increasing the Ketone Selectivity of the Cobalt-Catalyzed Radical Chain
Oxidation of Cyclohexane
Mathias Nowotny,* Lone N. Pedersen, Ulf Hanefeld, and Thomas Maschmeyer*[a]
Dedicated to Professor Sir John Meurig Thomas on the occasion of his 70th birthday
Abstract: A variety of heterogeneous
catalysts for the radical chain oxidation
of cyclohexane has been prepared by
immobilization of the well-defined co-
balt acetate oligomers [py3Co3(m3-
O)(OH)(O2CCH3)5](PF6) (1) and [py4
Co2(OH)2(O2CCH3)3](PF6) (2) on car-
boxy-modified mesoporous silica sup-
ports A D by carboxylate exchange.
The catalytic oxidation of cyclohexane
with tert-butyl hydroperoxide (TBHP)
in the presence of these homogeneous
and immobilized cobalt acetate com-
plexes afforded the corresponding alco-
hol and ketone in high yield. The
immobilization of 1 and 2 results in a
significant increase of catalytic activity.
TBHP acts as a radical initiator and as
source of molecular oxygen, which is
also involved in the overall oxidation
process. The rate of cyclohexane con-
version is limited by the diffusion of
molecular oxygen, and steady-state con-
centrations of cyclohexanone (K, ke-
tone) and cyclohexanol (A, alcohol) are
established; these determine the max-
imum K:A ratio.
Keywords: carboxylate ligands
cobalt ¥ heterogeneous catalysis ¥
immobilization oxidation
mesoporous materials
¥
¥
¥
It is equally well-established that the autoxidation of
Introduction
cyclohexane proceeds by free-radical pathways (Scheme 1).
However, the high degree of complexity of these reactions is
very challenging, and (predictive) chemical models of the
overall oxidation process are evolving only slowly.[6]
The aerobic oxidation of cyclohexane to a mixture of cyclo-
hexanol and cyclohexanone (the so-called K A oil), under
mild conditions is of great industrial significance, since these
two products are intermediates in the manufacture of nylon-6
and nylon-6,6.[1] The worldwide production of cyclohexanol
and cyclohexanone exceeds 106 tonnesperyear. Industrially,
cyclohexane is oxidized at 398 438 K and 8 15 bar air to
give an initial mixture of alcohol, ketone, and the intermedi-
ate cyclohexylhydroperoxide (CHHP), which is decomposed
either directly or in a separate step to yield additional ketone
and alcohol.[2] Cobalt carboxylate complexes are the most
widely employed homogeneous catalysts to achieve these
transformations. However, the application of these homoge-
neous catalysts is restricted by their generally short lifetime.
The role of cobalt acetate as catalyst in the catalytic process
has been studied intensely,[3] and it is now well-established
that ™cobalt(iii) acetate∫ consists of various combinations of
monomers, dimers, trimers, and mixed-valence trimers.[4]
However, the well-defined oligomers [py3Co3(m3-O)(OH)-
(O2CCH3)5](PF6) (1) and [py4Co2(OH)2(O2CCH3)3](PF6) (2)
have been isolated in high yields from cobalt acetate solutions
in the presence of pyridine (py).[5]
Scheme 1. Radical chain oxidation of cyclohexane.
[a] M. Nowotny, Prof. T. Maschmeyer, L. N. Pedersen, U. Hanefeld
Laboratory of Applied Organic Chemistry and Catalysis
Delft University of Technology, Julianalaan 136
2628 BL Delft (The Netherlands)
Conversions in the industrial process have to be limited to a
maximum of below 10%, to avoid overoxidation to other,
unwanted products, since the desired products, cyclohexanol
Fax : (‡31)15-278 4289
E-mail: th.maschmeyer@tnw.tudelft.nl
3724
¹ 2002 WILEY-VCHVerlag GmbH& Co. KGaA, Weinheim
0947-6539/02/0816-3724 $ 20.00+.50/0
Chem. Eur. J. 2002, 8, No. 16

3724 3731
and cyclohexanone, are more
readily oxidized than cyclohex-
ane itself. Another drawback of
the industrial process is the fact
that the cyclohexanol produced
has to be subsequently convert-
ed into the desired product
cyclohexanone, which repre-
sents an additional step. Thus,
an increase of the ketone:alco-
hol ratio (K:A), which is gen-
erally close to 1 in the industrial
production, would mean a sig-
nificant improvement of the
overall process.
In a preceding study, it was
demonstrated that the immobi-
lization of the well-defined tri-
nuclear cobalt acetate 1 inside
the channels of the hexagonal
mesoporous support MCM-41,
which had been functionalized
with carboxylic acid surface
grafts, results in a durable and
Scheme 2. Immobilization of oligonuclear cobalt acetate complexes on carboxy-functionalized MCM-41.
selective catalyst 1A for the
oxidation of cyclohexane (Scheme 2).[7] Moreover, this im-
mobilized catalyst exhibited a high selectivity for cyclohex-
anone in a catalytic system which employed tert-butyl hydro-
peroxide (TBHP) as a sacrificial oxidant and no solvents other
than the substrate cyclohexane. The use of 1 had been
prompted by the observation that this cobalt acetate oligomer
had shown promising selectivities in the catalytic oxidation of
adamantane.[8] However, the catalytic behavior of the homo-
geneous complex 1 and of other cobalt acetate oligomers in
the cyclohexane/TBHP system has not been reported so far.
Therefore, we present herein the results of our study on the
performance of the homogeneous cobalt acetate oligomers 1
and 2 in comparison to their respective immobilized ana-
logues on a variety of carboxylate-modified silica supports.
In the preceding study,[7] the functionalization of the
support surface was effected by derivatizing previously
complex. Moreover, we wished to study the influence of the
nature of the immobilized cobalt acetate complex and to gain
a deeper insight into the solvent-free cyclohexane/TBHP
system in an attempt to optimize the achievable K:A ratio.
Results and Discussion
Our first approach towards the synthesis of a carboxy-
functionalized mesoporous support comprised the acid-cata-
lyzed hydrolysis of 3-cyanopropyl-grafted siliceous MCM-41
(3), which had been obtained from a reaction of the mesopore
with 3-(trichlorosilyl)butyronitrile (Scheme 3). The grafted
compound 3 was converted into the carboxylate derivative B
by treatment with aqueous sulfuric acid at 1508C in an
analogous manner to a reported procedure for functionalizing
grafted
3-bromopropylsilane
tethers with glycine. However,
the low efficiency of the reac-
tion between the amine part of
the glycine and the bromoalkyl
group limited the degree of
glycine derivatization and,
hence, the loading of the modi-
fied support with the catalyst
complex. The target of the
present study was to overcome
this limitation by developing
new, more efficient pathways
of derivatizing MCM-41 with
carboxylic acid grafts and,
thereby, to achieve a higher
loading of the support with the
oligonuclear cobalt acetate
Scheme 3. Preparation of novel carboxy-functionalized MCM-41 derivatives.
¹ 2002 WILEY-VCHVerlag GmbH& Co. KGaA, Weinheim 0947-6539/02/0816-3725 $ 20.00+.50/0
Chem. Eur. J. 2002, 8, No. 16
3725

M. Nowotny, T. Maschmeyer et al.
FULL PAPER
a cyanoalkyl-modified silica surface with carboxylic acid
groups.[9] However, the harsh conditions of the hydrolysis step
resulted in a partial collapse of the mesoporous support
during the synthesis of B. This mesopore collapse manifested
itself in a reduction of the BET surface area Atot from 975 to
690 m2 gÀ1, and in a reduction of the average pore diameter
Dav from 24 to 22 ä, as determined before and after
hydrolysis, respectively. An alternative, a more gentle ap-
proach for functionalizing the mesopore surface with carbox-
ylic acid grafts is represented by the reaction of MCM-41 with
the linker 4, which afforded the modified support D. Linker 4,
which itself is accessible through the ring-opening aminolysis
of glutaric anhydride by 3-aminopropyltrimethoxysilane,[10]
cannot be stored indefinitely because it undergoes slow
polymerization as a result of carboxylic acid attack at the
trialkoxysilane terminus. Therefore, it has to be freshly
prepared before the immobilization step. An analogous
reaction of glutaric anhydride with 3-(methylamino)propyl-
trimethoxysilane afforded the N-methyl derivative 5, which in
turn was grafted onto MCM-41 to give support C. The
kinetically controlled protection of the silanol groups on the
external surface of MCM-41 with an equivalent quantity of
dichlorodimethylsilane[7] guaranteed the location of the sub-
sequently introduced carboxylic acid grafts inside the meso-
porous channels in all of the novel modified supports.
Moreover, any residual silanol groups on the surface of the
carboxy-derivatized mesoporous supports B and C were
protected by reaction with excess dimethoxydimethylsilane.
However, in case of derivative D, the residual silanol groups
and the amide proton were left unprotected for an assessment
of the influence of these weakly acidic sites on the course of
the catalytic reaction.
For the immobilization of the cobalt acetate oligomers 1
and 2 on the carboxylate-modified supports, we utilized the
tendency of the bridging acetate ligands in these complexes to
undergo exchange reactions with other carboxylic acids (see
Scheme 2).[5b] However, an irreversible anchoring of 1 and 2
by a partial ligand exchange with surface-tethered carboxy
groups requires the removal of the released acetic acid from
the equilibrium. The immobilization was, thus, carried out by
stirring the respective cobalt acetate complex with an excess
of the carboxy-functionalized support in boiling CHCl3, while
the released acetic acid was trapped in the gas phase by
passing the reflux through a Soxhlet thimble filled with
anhydrous Na2CO3. Any residual, physically adsorbed cobalt
acetate was subsequently removed by Soxhlet extraction of
the as-prepared material with CH2Cl2. The application of this
procedure on all possible permutations of cobalt acetate
complexes and carboxy-functionalized supports afforded the
series 1B D of immobilized trinuclear complex 1 and the
series 2B D of immobilized dinuclear complex 2 (Scheme 2).
Under identical reaction conditions, the extent of cobalt
acetate incorporation, expressed by the Co content deter-
mined by X-ray fluorescence (XRF) analysis, spanned the
range between 0.237 and 0.925 wt% for 2C and 1B, respec-
tively. Hence, the loadings achieved by this immobilization
procedure generally exceed the Co content of 0.189 wt%
determined for a sample of 1A that was prepared in
accordance with the previously reported procedure.
The homogeneous cobalt acetate oligomers 1 and 2, and all
immobilized catalysts 1A 1D and 2B 2D were subjected to
catalytic cyclohexane oxidation tests in the presence of TBHP
under conditions similar to those of the original report. In
agreement with the reported behavior for the catalytic system
involving 1A, cyclohexanone, after an initial phase of
predominant production of cyclohexanol, represents the
principle product in all systems studied. A comparison of
the catalytic runs with either the homogeneous complexes 1, 2,
or the immobilized catalyst 1B (performed in the presence of
equimolar quantities of Co), reveals some significant differ-
ences in the catalytic performance (Figure 1). Both homoge-
Figure 1. Comparison of the course of the catalytic oxidation of cyclo-
hexane with TBHP in presence of cobalt acetate oligomers 1 and 2 and
immobilized catalyst 1B, respectively. A quantity of 2.6 mmolLÀ1 of Co was
present in all reactions.
neous cobalt acetate catalysts exhibit inferior activities under
the conditions applied, as indicated by the lower rate of
ketone and alcohol formation relative to the immobilized
counterpart 1B. Furthermore, this reduced ability of the
homogeneous catalysts to catalyze the decomposition of
peroxide intermediates leads to relatively higher concentra-
tion of CHHP, which represents the main product in the
reaction mixture of both 1 and 2 at runtimes under 2200 and
1200 min, respectively. The higher activity of the immobilized
catalyst 1B may be explained by the high dispersion of
complex 1 on the support surface, while the low solubility of 1
and 2 in the reaction medium might prevent a complete
dissolution of the homogeneous complexes under the reaction
conditions employed.
An accurate monitoring of the TBHP concentration
revealed that all catalytic systems underwent an initial phase
of rapid TBHP decomposition, accompanied by vigorous
evolution of molecular oxygen, which in some cases resulted
in the complete consumption of TBHP and termination of the
catalytic cyclohexane conversion. This initial TBHP decom-
position activity decreased rapidly within the first hour.
Addition of neat TBHP after that time did result in further
hydroperoxide decomposition, but not at the initial rate. A
3726
¹ 2002 WILEY-VCHVerlag GmbH& Co. KGaA, Weinheim
0947-6539/02/0816-3726 $ 20.00+.50/0
Chem. Eur. J. 2002, 8, No. 16

Radical Chain Oxidation of Cyclohexane
3724 3731
simple concentration dependence of the TBHP decomposi-
tion rate must therefore be excluded, while a change in the
activity pattern resulting from a structural change of the
catalyst complex is a more likely explanation. Hints for such
structural changes were found in an in situ EXAFS study on
1A in the active state.[7] A modification of the experimental
procedure by direct addition of half of the oxidant at the start,
followed by the gradual addition of the second half during the
first hour of the catalytic experiment left a sufficient concen-
tration of TBHP at the end of the initial phase of rapid
decomposition to guarantee further cyclohexane conversion.
The decomposition of large amounts of TBHP under
formation of molecular oxygen does not represent an
unproductive loss of the oxidant. Volumetric monitoring of
the released gas revealed that the amount of dioxygen reaches
a maximum shortly after the end of the initial TBHP
decomposition phase after ꢀ150 min. Subsequently, it is
consumed again in the course of the catalytic cyclohexane
oxidation. The formation, as well as consumption, of dioxygen
is indicative of a catalytic system that proceeds via free-radical
chains, which originate from a Haber Weiss hydroperoxide
decomposition cycle[11] (Scheme 1, Eqs. (1) and (2)) and
partially terminate in Russell recombination steps[12]
(Scheme 1, Eq. (7)). This underlying mechanism also explains
the formation of the detected monooxygenated products
CHHP, cyclohexanol, cyclohexanone, and of the coupling
products tert-butylcyclohexylperether (TBCP) and tert-butyl
peroxide (TBP), while the detected adipic acid (AA) repre-
sents the principal product of overoxidation.[2, 3] The typical
variation of the concentrations of these compounds during the
course of the catalytic reaction, exemplified by a reaction that
employed catalyst 1B, is represented in Figure 2.
confirmed in a separate experiment by the addition of excess
cyclohexanol to a catalytic run involving catalyst 1B. Con-
sequently, the system responded by a relaxation of the
elevated alcohol concentration towards the original steady-
state value (Figure 3a), thereby indicating the consumption of
Figure 3. a) Response of the catalytic system with catalyst 1B to the
addition of an excess of cyclohexanol at t ˆ 300 min. b) Response of the
same system to the addition of an excess of cyclohexanone at t ˆ 100 min.
Dashed lines represent the behavior of the undisturbed system.
cyclohexanol in subsequent oxidation steps. The CHHP and
TBP concentrations adopt similar steady-states in all catalytic
systems within the first 500 min of the catalytic reaction. The
cyclohexanone concentration increases steadily in all systems
during the course of the reaction, thereby resulting in a
continuous rise of the K:A ratio with increasing runtime, until
the cyclohexanone concentration approaches a steady-state
level near 500 mmolLÀ1. Yet, this steady-state of cyclohex-
anone concentration is never fully established under the
applied standard conditions since the near-complete con-
sumption of the oxidants TBHP and molecular oxygen at
extended reaction times results in a general decrease in the
rate of reaction.
Figure 2. Course of the oxidation of cyclohexane with TBHP in the
presence of immobilized catalyst 1B. The amount of released molecular
oxygen is given with respect to the liquid volume of the reaction mixture to
facilitate comparison.
The initially high rate of cyclohexanol formation decreases
in all systems investigated until the cyclohexanol concentra-
tion reaches a steady-state level of 160 Æ 5 mmolLÀ1. The
steady-state nature of this limiting concentration has been
An artificial increase in the cyclohexanone concentration
well above its steady-state level by addition of a surplus of
Chem. Eur. J. 2002, 8, No. 16
¹ 2002 WILEY-VCHVerlag GmbH& Co. KGaA, Weinheim
0947-6539/02/0816-3727 $ 20.00+.50/0
3727

M. Nowotny, T. Maschmeyer et al.
FULL PAPER
cyclohexanone at the beginning of the catalytic reaction
resulted in a decrease from this elevated level because of the
increased efficiency of substrate-dependent overoxidation
processes (Figure 3b). However, the rate of this decrease
towards the steady-state level is significantly lower than the
respective rate for cyclohexanol. This difference in rates of
oxidation is one of the causes of the increasing K:A ratio in
the catalytic system studied. An additional contribution to the
predominance of cyclohexanone over cyclohexanol in the
TBHP-promoted oxidation of cyclohexane derives from the
.
lack of an a-proton in tBuOO , which will favor the decay
under formation of cyclohexanone in a mixed Russell
.
termination between this radical and c-C6H11OO (Scheme 4).
Figure 4. Time-dependence of the concentrations of cyclohexanone (solid
lines) and cyclohexanol (dashed lines) in the course of the catalytic
oxidation of cyclohexane in presence of different amounts of immobilized
catalyst 1B.
overoxidation processes. The gradually increasing concentra-
tion of carboxylic acids in the reaction mixture may also
explain the observed leaching of the cobalt acetate complexes
from the supports at extended reaction times (Table 1), since
these acids are able to replace the carboxylate tethers of the
immobilized complexes, thereby reversing the original im-
mobilization procedure.
The initial rate of the catalytic cyclohexane conversion that
used the immobilized catalysts 1A 1D and 2B 2D, ex-
pressed in terms of the respective cyclohexane conversion at a
runtime of 100 min[13] and under consideration of blank
effects, exhibits a linear dependence with respect to cobalt
loading (Figure 5). An influence of the nature of the various
immobilized cobalt acetate complexes, supports, or tethers on
the course of the catalytic reaction was not noticed, even at
this early stage. Moreover, even the early correlation with the
Co concentration will be obscured when the catalytic system
evolves towards its equilibrium state. Therefore, any compar-
ison of the catalytic performance after the early stage of the
catalysis in terms of turnover numbers is meaningless.
.
Scheme 4. Mixed Russell termination step between tBuOO and
c-C6H11OO .
.
The independence of the steady-state concentrations of
cyclohexanol and cyclohexanone on the catalyst concentra-
tion was also revealed in a series of experiments that used the
catalyst 1B at various concentrations (Figure 4). The exis-
tence of steady-state concentrations of both cyclohexanol and
cyclohexanone limits the theoretically achievable K:A ratio in
the system under consideration. Moreover, this limiting K:A
value can be approached only at the price of a decreased
selectivity for monooxygenated products, since allowing the
system to proceed to a conversion of >15% results in a
considerable degree of overoxidation under predominant
formation of adipic acid (Table 1). Trace amounts of 6-hy-
droxycaproic, glutaric, succinic, and valeric acids were also
detected in the reaction mixture as a result of further
Table 1. Oxidation of cyclohexane with TBHP catalyzed by cobaltic acetate oligomers.[a]
t ˆ 100 min
t ˆ 4000 min
[CHHP] [TBCP] [AA] K/A
[c]
Cat
no.
[Co]
[K]
[A]
[CHHP] [TBCP] Conv.[b]
[K] [A]
[mm] [mm]
Smono
Conv.[b] Leach.[d]
[mm] [mm] [mm] [mm]
[mm]
[mol%]
[mm]
[mm] [mm]
[%]
[%]
[%]
[e]
0
2.59
2.6
0.32
1.58
1.15
1.25
0.60
1.06
0.82
12
24
10
18
38
14
52
72
70
74
44
47
49
54
51
65
1
5
1
1.51
2.65
1.81
2.47
6.25
5.86
5.59
3.82
5.20
4.22
49
316
301
474
497
468
445
509
532
545
36
111
104
136
128
144
121
142
150
140
159
137
171
108
80
39
85
64
60
4
14
15
29
23
42
33
35
34
39
3
62
53
111
108
101
113
128
119
121
1.36
2.85
2.89
3.49
3.89
3.24
3.69
3.58
3.56
3.89
98.81
90.31
91.77
87.06
87.08
87.28
85.82
85.42
86.70
86.54
4.55
1
2
12.21
12.32
15.13
15.42
14.36
14.41
15.82
16.15
16.48
1A
1B
1C
1D
2B
2C
2D
24
37
3
0.08
0.98
2.07
0.90
3.06
2.87
2.60
147
141
124
57
107
62
131
121
123
92
116
94
17
15
13
9
14
9
54
[a] General conditions: 6 mL of a mixture of cyclohexane (5.58 molLÀ1), TBHP (3.14 molLÀ1) and chlorobenzene (0.44 mol LÀl), amount of catalyst as
indicated in text, 343 K. [b] Conversion of cyclohexane.[13] [c] Selectivity for monooxygenated products after 4000 min, calculated as 100([mono])/([mono] ‡
[AA]), where [mono] ˆ [K] ‡ [A] ‡ [CHHP] ‡ [TBCP]. [d] amount of Co in solution after 4000 min, referred to total amount of Co. [e] Blank reaction in
absence of immobilized catalyst. K ˆ cyclohexanone, A ˆ cyclohexanol, CHHP ˆ cyclohexyl hydroperoxide, TBCP ˆ tert-butylcyclohexylperether, AA ˆ
adipic acid.
3728
¹ 2002 WILEY-VCHVerlag GmbH& Co. KGaA, Weinheim
0947-6539/02/0816-3728 $ 20.00+.50/0
Chem. Eur. J. 2002, 8, No. 16

Radical Chain Oxidation of Cyclohexane
3724 3731
escape of radical chains from their origin (at the catalyst
inside the mesopores) into the bulk solution. An analysis of
the already intricate catalytic system is further complicated by
the fact that molecular oxygen (which originates from an
initial phase of rapid TBHP decomposition in presence of the
catalyst) is present as a second terminal oxidant in large
quantities. After the initial stage, the catalytic system is
governed by mass-transfer limitation in molecular oxygen,
which obscures any differences that arise from nature or
concentration of the immobilized cobalt acetate catalysts. The
achievable K:A ratio in this catalytic system is limited by the
appearance of catalyst-independent steady-state concentra-
tions of both cyclohexanol and cyclohexanone.
Figure 5. Correlation of the cyclohexane conversion at t ˆ 100 min with
the Co loading of the immobilized catalyst employed, 1A 1D and 2B
2D, respectively.
Experimental Section
The observation of TBHP- and catalyst-independent
steady-state concentrations for cyclohexanol and cyclohexa-
none indicates a mass-transfer limitation in molecular oxygen
of the radical chain oxidation. Hence, radical chains that were
initiated by the immobilized cobalt acetate complexes inside
the mesopores of 1A 1D and 2B 2D must escape from the
porous space into the bulk solution during the course of the
reaction. Consequently, cyclohexane conversion is increas-
ingly governed by freely evolving radical chains, and any
effect of confinement control by the mesoporous host
becomes negligible. The observed mass-transfer limitation
underlines the significance of the molecular oxygen released
from the TBHP decomposition for the product formation
within the radical chain (Scheme 1, Eq. (5)). The influence of
molecular oxygen on the course of the radical chain oxidation
of cyclohexane by TBHP has also been observed in a different
catalytic system.[14]
In contrast, the role of the second potential oxidant TBHP
in the system under consideration is that of a radical initiator
(Scheme 1, Eqs. (1) and (2)) rather than that of a direct
oxygen-transferring agent towards the substrate (apart from
being the source of the released molecular oxygen), because
its contribution to product formation is not rate-determining.
However, TBHP plays an important role in the subsequent
oxidation of cyclohexanol to cyclohexanone, and hence in the
increase of the K:A ratio.[15]
General remarks: All reactions involving alkoxysilanes or chlorosilanes
were carried out under nitrogen with standard Schlenk techniques.
Siliceous MCM-41,[16] complexes 1 and 2,[5] and the immobilized catalyst
1A[7] were prepared according to published procedures. 3-(Methylamino)-
propyltrimethoxysilane was acquired from Fluka, while all other chemicals
were purchased from Aldrich. All stoichiometric calculations involving
support surfaces are based on an average surface silanol density of 7.8 Â
[17]
10À6 molSiOHmÀ2
.
1Hand 13C liquid-state NMR spectra were recorded
on a Varian Unity Inova spectrometer and referenced against internal
TMS. 13C MAS NMR spectra were acquired on a Varian VXR400S
spectrometer. XRD data were obtained from a Philips PW1840 diffrac-
tometer operating with CuKa radiation. XRF analyses were carried out on a
Philips PW1480 instrument. IR spectra were recorded on a Perkin Elmer
Spectrum 1000FT-IR spectrometer. GC analyses were carried out on an
Agilent6890 gas chromatograph equipped with a split inlet (2008C, split
ratio 10.0), a Chrompack Sil5CB capillary column (50 m  0.53 mm;
constant flow of N2 4.0 mLminÀ1; temperature program: isothermal at
458C for 33 min, followed by heating to 3008C with a rate of 208CminÀ1
,
final time 15 min at 3008C) and a FID detector (set to 3258C); retention
times (min): cyclohexane 8.83, TBHP 10.97, TBP 15.23, methyl valeriate
23.44, chlorobenzene 26.58, cyclohexanone 33.59, cyclohexanol 34.18,
dimethyl succinate 39.02, CHHP 39.48, TBCP 40.21, dimethyl glutarate
40.47, dimethyl adipate 41.49. Cobalt concentrations in the filtrates of the
catalytic experiments were determined by AAS analysis on a Perkin El-
mer4100ZL spectrometer. BET surface areas Atot, BJHpore volumes Vtot
,
and average pore diameters Dav were determined by nitrogen adsorption
analysis at 77 K on a Quantachrome Autosorp-6B instrument.
Preparation of carboxylic acid-grafted supports
[MCM-41]-(CH2)3CN (3): MCM-41 (25 g, Atot ˆ 1025 m2 gÀ1, 200 mmol-
SiOH) was dehydrated for 3 h at 2008C in vacuo and subsequently
suspended in dry hexane (70 mL). Dichlorodimethylsilane (0.69 g,
5.36 mmol) was added dropwise at 208C to protect the external surface.
The mixture was stirred for 12 h, then 3-cyanopropyltrichlorosilane (2.43 g,
12 mmol, 0.2 equiv) was added dropwise, and stirring was continued for
additional 12 h. All volatiles were removed in vacuo and the remaining
solid was suspended in toluene (100 mL). Dimethoxydimethylsilane (8.4 g,
70 mmol) was added to protect the residual silanol groups, and the resulting
mixture was heated under reflux for 20 h. The product was collected by
filtration, washed with acetone and ether and dried in vacuo at 808C for
Conclusion
The catalytic cyclohexane oxidation that employs cobalt
acetate complexes as catalysts and TBHP as the oxidant is a
typical radical chain-oxidation process. The moderate cata-
lytic activity of the oligomeric catalyst complexes 1 and 2 is
significantly increased by their immobilization on carboxy-
modified mesoporous silica supports. In the early phase of the
catalytic reaction involving the immobilized catalysts 1A
1D and 2B 2D, the principal parameter determining the
system reactivity is the respective Co loading. A substantial
effect of confinement on the selectivity of the radical chain
oxidation is not observed. This is consistent with a rapid
12 h. Yield 25.5 g; XRD: 2Vˆ 2.488; BET: Atot ˆ 973 m2 gÀ1
,
Vtot
ˆ
0.873 cm3 gÀ1, Dav ˆ 24 ä; IR (KBr): nÄCN ˆ 2259 cmÀ1
;
13C CP-MAS NMR
(100 MHz): d ˆ 0.0 (SiCH3), 10.7 (CH2), 19.1 (2  CH2), 48.8 (residual
OCH3), 111.3 ppm (CN).
[MCM-41]-(CH2)3COOH (B): Compound 3 (5 g) was suspended in a
mixture of H2O and H2SO4 (40 mL, 1:1 (v/v)) and heated to 1508C for 1 h.
The solid was collected by filtration and washed with water until the filtrate
gave a neutral reaction. Subsequent washing with methanol and ether
followed by drying in vacuo at 808C overnight afforded 4.94 g of B. XRD:
2Vˆ 2.548; BET: Atot ˆ 690 m2 gÀ1
,
Vtot ˆ 0.612 cm3 gÀ1
,
Dav ˆ 21 ä; IR
Chem. Eur. J. 2002, 8, No. 16
¹ 2002 WILEY-VCHVerlag GmbH& Co. KGaA, Weinheim
0947-6539/02/0816-3729 $ 20.00+.50/0
3729

M. Nowotny, T. Maschmeyer et al.
FULL PAPER
(KBr): nÄCO ˆ 1733 cmÀ1
;
13C CP-MAS NMR (100 MHz): d ˆ 0.0 (SiCH3),
2D: From 2 (0.2 g, 0.274 mmol) and D (2.0 g). Yield 1.983 g. XRF: [Co]
0.484 wt%.
11.66 (CH2), 18.10 (CH2), 35.71 (CH2), 177.32 ppm (CO).
Catalytic investigations: In all catalytic experiments, a stock solution of
TBHP in cyclohexane was used which had been prepared by extraction of
commercial TBHP (Aldrich, 70% in water) into an equal volume of
cyclohexane. Phase separation was promoted by saturation of the aqueous
layer with solid NaCl. The organic layer was dried over MgSO4, filtered and
stored at 48C.
(H3CO)3Si(CH2)3NHCO(CH2)3COOH (4): Glutaric anhydride (0.384 g,
3.37 mmol) in dry THF (20 mL) was added dropwise to a solution of
3-aminopropyltrimethoxysilane (0.556 g, 3.12 mmol) in THF (30 mL).
After the mixture had been stirred for 12 h at 208C, the solvent was
removed in vacuo. Compound
4 was isolated as a colorless oil in
quantitative yield. 1HNMR (300 MHz, CDCl 3, TMS): d ˆ 0.65 (m, 2H;
SiCH2), 1.63 (m, 2H; CH2), 1.96 (m, 2H; CH2), 2.28 (m, 2H; CH2), 2.41 (m,
2H; CH2), 3.24 (m, 2H; CH2), 3.57 (s, 9H; OCH3), 6.25‡6.44 (1H; NH),
11.21 ppm (s, 1H; COOH); 13C NMR (75 MHz, CDCl3, TMS): d ˆ 6.5,
Homogeneous catalytic experiments were carried out with finely powdered
1 (4.6 mg, 5.19 mmol) or 2 (6.1 mg, 7.8 mmol). A 60-mg quantity of dry
catalyst was used in all catalytic experiments that used immobilized
cobaltic acetate oligomers.
ˆ
20.9, 22.6, 33.2, 35.4, 42.0 (CH2), 50.6 (OCH3), 172.9 (C( O)N), 176.9 ppm
(COOH); IR (CH2Cl2): nÄ ˆ 1713 (COOH), 1670 (amide I), 1522
The respective catalyst was added to 3 mL of a mixture consisting of
cyclohexane (5.58 molLÀ1), TBHP (3.14 molLÀ1), and chlorobenzene
(0.44 mol LÀl) as internal standard in a 10-mL reactor fitted with a reflux
condenser and a magnetic stirring bar. The mixture was rapidly heated to
708C by immersion in a thermostated bath. An additional 3 mL of the
reagent mixture was added in 0.5 mL portions at intervals of 10 min. The
course of the reaction was monitored by GC analysis of liquid samples,
beginning after the evolution of molecular oxygen had mostly ceased at
100 min.[13] The production and consumption of molecular oxygen was
monitored volumetrically with an attached gas burette. The concentration
of carboxylic acid side products was determined by GC analysis from
separate samples after conversion into the respective methyl esters.[18] The
extent of Co leaching was determined by AAS analysis of the filtered
reaction mixture after the end of the catalytic experiment.
(amide II) cmÀ1
.
(H3CO)3Si(CH2)3N(CH3)CO(CH2)3COOH (5): Compound 5 was prepared
in quantitative yield in an analogous manner to the synthesis of 4 from
3-(methylamino)propyltrimethoxysilane (0.929 g, 4.8 mmol) and glutaric
anhydride (0.547 g, 4.8 mmol). 1HNMR (300 MHz, CDCl 3, TMS): d ˆ 0.59
(m, 2H; SiCH2), 1.65 (m, 2H; CH2), 1.95 (m, 2H; CH2), 2.38 2.52 (m, 4H;
2CH2), 2.92‡3.00 (2s, 3H; NCH3), 3.31 (m, 2H; NCH2), 3.57 (s, 9H; OCH3
), 11.58 ppm (s, 1H; COOH); 13C NMR (75 MHz, CDCl3, TMS): d ˆ 6.3,
20.2, 21.7, 31.9‡32.5, 33.4, 51.3‡52.2 (CH2), 34.7‡35.4 (NCH3), 50.6
ˆ
(OCH3), 172.6‡172.7 (C( O)N), 177.0 ppm (COOH); IR (CH2Cl2): nÄ ˆ
1715 (COOH), 1651 (amide) cmÀ1
.
[MCM-41]-(CH2)3NHCO(CH2)3COOH (D): After dehydration for 3 h at
2008C in vacuo, MCM-41 (5 g, Atot ˆ 1025 m2 gÀ1, 40 mmolSiOH) was
suspended in hexane (50 mL), and the external surface was protected by
addition of dichlorodimethylsilane (0.138 g, 1.07 mmol). After the mixture
had been left to stir for 12 h at 208C, all volatiles were removed in vacuo. A
solution of freshly prepared 4 (0.863 g, 2.94 mmol) in THF (30 mL) was
added, and the resulting suspension was stirred at 208C for 44 h. The title
compound was isolated by filtration and washed with CH2Cl2. Drying in
vacuo at 808C overnight afforded 5.293 g of D; 112 mg of nonadsorbed 4
was recovered from the filtrates. 13C CP-MAS NMR (100 MHz): d ˆ À0.3
Acknowledgements
We thank Dr. James K. Beattie, Dr. Anthony F. Masters, and John A.
Klepetko from the University of Sydney for a generous gift of complex 1.
Helpful discussions with Prof. Sir John Meurig Thomas and Dr. Robert
Raja are also gratefully acknowledged. M.N. thanks DSM Research for a
research fellowship at the TU Delft. U.H. acknowledges a fellowship of The
Royal Netherlands Academy of Arts and Sciences (KNAW), and L.N.P.
thanks the TU Delft and IAESTE Netherlands for funding a two-month
studentship.
ˆ
(SiCH3), 18 49 (br, CH2), 172.0 (C( O)N), 174.0 ppm (COOH); IR
(KBr): nÄ ˆ 1725 (COOH), 1650 (amide I), 1558 (amide II) cmÀ1
.
[MCM-41]-(CH2)3N(CH3)CO(CH2)3COOH (C): MCM-41 (15 g, Atot
ˆ
1025 m2 gÀ1, 120 mmolSiOH) was dehydrated for 3 h at 2008C in vacuo,
and the external surface was protected by addition of dichlorodimethylsi-
lane (0.414 g, 3.22 mmol) in hexane (100 mL). After the slurry had been
stirred for 12 h at 208C, all volatiles were removed in vacuo. The residue
was suspended in THF (100 mL), and 5 (0.934 g, 3 mmol, freshly prepared
in THF (20 mL)) was added under stirring at 208C. Stirring was continued
for 12 h, and the mixture was subsequently heated under reflux for 6 h.
Dimethoxydimethylsilane (6.2 mL, 45 mmol) was added dropwise, and
heating under reflux was continued for additional 12 h. Filtration, washing
with CH2Cl2, and drying in vacuo overnight at 808C afforded 15.33 g of C.
13C CP-MAS NMR (100 MHz): d ˆ À3.5 (SiCH3), 20 55 (br, CH3 ‡ CH2),
[1] a) W. B. Fisher, J. F. VanPeppen in Kirk Othmer Encyclopedia of
Chemical Technology, 4th ed., Vol. 7 (Ed.: M. Howe-Grant), Wiley,
New York, 1996, pp. 871 859; b) M. T. Musser in Ullmann×s Ency-
clopedia of Industrial Organic Chemicals, Vol. 3, VCH-Wiley, Wein-
heim, 1999, pp. 1807 1821.
[2] R. A. Sheldon, J. K. Kochi, Metal Catalyzed Oxidations of Organic
Compounds, Academic Press, New York, 1981.
[3] a) N. M. Emanuel, Z. K. Maizus, I. P. Skibida, Angew. Chem. 1969, 81,
91 101; Angew. Chem. Int. Ed. Engl. 1969, 8, 97 107; b) E. P. Talsi,
V. D. Chinakov, V. P. Balenko, V. N. Sidelnikov, K. I. Zamarev,
Molecular Catalysis, Wiley, New York, 1980; c) G. W. Parshall, S. D.
Ittel, Homogeneous Catalysis, 2nd ed., Wiley, New York, 1992,
pp. 237 268.
[4] a) A. R. Blake, J. R. Chipperfield, S. Lau, D. E. Webster, J. Chem. Soc.
Dalton Trans. 1990, 3719 3724; b) C. F. Hendriks, H. C. A. van Beek,
P. M. Heertjes, Ind. Eng. Chem. Prod. Res. Dev. 1979, 18, 43 46; c) W.
Partenheimer, R. K. Gipe in Catalytic Selective Oxidation (Eds.: S. T.
Oyama, J. T. Hightower) ACS Symp. Ser. 1993, 523, pp. 81 88; d) J. K.
Beattie, T. W. Hambley, J. A. Klepetko, A. F. Masters, P. Turner,
Polyhedron 1998, 17, 1343 1354.
[5] a) S. Uemura, A. Spencer, G. Wilkinson, J. Chem. Soc. Dalton Trans.
1973, 2565 2571; b) C. E. Sumner, Inorg. Chem. 1988, 27, 1320 1327.
[6] a) M. Spielman, AIChE Journal 1964, 10, 496 501; b) J. Alagy, P.
Trombouze, H. van Landeghem, Ind. Eng. Chem. Proc. Des. Dev.
1974, 13, 317 323; c) C. A. Tolman, J. D. Druliner, P. J. Krisic, M. J.
Nappa, W. C. Seidel, I. D. Williams, S. D. Ittel, J. Mol. Catal. 1988, 48,
129 148; d) C. A. Tolman, J. D. Druliner, M. J. Nappa, N. Herron in
Activation and Functionalization of Alkanes (Ed.: C. L. Hill), Wiley,
New York, 1989, pp. 303 360; e) J. D. Druliner, L. D. Geller, E.
Wasserman, J. Phys. Chem. 1991, 95, 1519 1521; f) R. Pohorecki, J.
ˆ
173.0 (C( O)N), 177.1 ppm (COOH); IR (KBr) : nÄ ˆ 1731 (COOH), 1636
(amide) cmÀ1
.
General procedure for the immobilization of cobalt acetate clusters on
carboxylic acid grafted supports: The amounts of cobalt(iii) acetate
oligomer and modified support that are stated below were suspended in
CHCl3 (30 mL), and the resulting slurry was heated under reflux for 24 h in
a flask which was equipped with a Soxhlet extractor that contained a
thimble filled with anhydrous Na2CO3 (0.5 g) that was used to trap any
acetic acid released. The product was collected by filtration, extracted with
CH2Cl2 in a Soxhlet device for 20 h, and finally dried in vacuo at 808C for
24 h.
1B: From 1 (0.2 g, 0.226 mmol) and B (2.0 g). Yield 2.128 g. XRF: [Co]
0.925 wt%.
1C: From 1 (0.2 g, 0.226 mmol) and C (2.0 g). Yield 2.064 g. XRF: [Co]
0.678 wt%.
1D: From 1 (0.1 g, 0.113 mmol) and D (2.0 g). Yield 2.074 g. XRF: [Co]
0.735 wt%.
2B: From 2 (0.1 g, 0.137 mmol) and B (2.0 g). Yield 2.012 g. XRF: [Co]
0.356 wt%.
2C: From 2 (0.1 g, 0.137 mmol) and C (2.0 g). Yield 1.993 g. XRF: [Co]
0.623 wt%.
3730
¹ 2002 WILEY-VCHVerlag GmbH& Co. KGaA, Weinheim
0947-6539/02/0816-3730 $ 20.00+.50/0
Chem. Eur. J. 2002, 8, No. 16

Radical Chain Oxidation of Cyclohexane
3724 3731
¬
¬
Ba¸dyga, W. Moniuk, W. Podgorska, A. Zdrojkowski, P. Wierzchow-
[14] P. A. MacFaul, K. U. Ingold, D. D. M. Wayner, L. Que, Jr., J. Am.
Chem. Soc. 1997, 119, 101594 10598.
ski, Chem. Eng. Sci. 2001, 56, 1285 1291.
[7] T. Maschmeyer, J. M. Thomas, A. F. Masters, J. K. Beattie, G. Sankar,
R. D. Oldroyd, J. A. Klepetko, Angew. Chem. 1997, 109, 1713 1716;
Angew. Chem. Int. Ed. Engl. 1997, 36, 1639 1642
[8] A. F. Masters, J. K. Beattie, personal communication.
[9] a) J. H. Clark, D. J. Macquarrie, Chem. Commun. 1998, 853 854;
b) P. M. Price, J. H. Clark, D. J. Macquarrie, J. Chem. Soc. Dalton
Trans. 2000, 101 110.
[10] U. Nagel, E. Kinzel, J. Chem. Soc. Chem. Commun. 1986, 1098 1099.
[11] a) F. Haber, J. Weiss, Naturwissenschaften, 1932, 20, 948 950; b) R. A.
Sheldon, J. K. Kochi, Adv. Catal. 1976, 25, 272 413.
[12] G. A. Russell, J. Am. Chem. Soc. 1957, 79, 3871 3877.
[13] Inevitable evaporation losses of cyclohexane–which was carried
away by the evolved gasses–prevented a complete mass balance in
the GC analysis. The cited conversions were calculated by referring
the sum of the determined concentrations of all cyclohexane-derived
products to the initial cyclohexane concentration in the respective
catalytic experiment.
[15] Neat cyclohexanol (3 mL) was converted almost stoichiometrically
into cyclohexanone in the presence of TBHP (70% in water, 3 mL)
and catalyst 1B (60 mg) at 708C in an independent catalytic experi-
ment.
[16] J. S. Beck, J. C. Vartuli, W. J. Roth, M. E. Leonowicz, C. T. Kresge,
K. D. Schmitt, C. T-W. Chu, D. H. Olson, E. W. Sheppard, S. B.
McCullen, J. B. Higgins, J. L. Schlenker, J. Am. Chem. Soc. 1992, 114,
10834 10843.
[17] a) L. T. Zhuralev, Colloids Surf. 1993, 74, 71 90; b) P. Basu, D.
Panayotov, J. T. Yates, Jr., J. Am. Chem. Soc. 1988, 110, 2074 2081.
[18] a) R. Raja, P. Ratnasamy, Catal. Lett. 1997, 48, 1 10; b) L. D.
Metcalfe, A. A. Schmitz, Anal. Chem. 1961, 33, 363 364.
Received: December 27, 2001
Revised: April 8, 2002 [F3761]
Chem. Eur. J. 2002, 8, No. 16
¹ 2002 WILEY-VCHVerlag GmbH& Co. KGaA, Weinheim
0947-6539/02/0816-3731 $ 20.00+.50/0
3731