Transformations of cyclohexene oxide over silica-supported Cu, Pd, and Rh catalysts in H2/D2 atmosphere

Article abstract of DOI:10.1006/jcat.1998.2284

Transformations of cyclohexene oxide were studied in a recirculation reactor and a flow system over silica-supported transition and early transition metal catalysts. Hydrogen pressure dependence of the transformations was measured over Cu/SiO₂ at 403 K in the 0-98.7 kPa range. Deuterium distribution in the reactant and products were determined at 20 kPa D₂ pressure. The reactions were studied in the flow reactor under D₂ stream over Cu/SiO₂ (373 K, 403 K, 443 K, 523 K), Pd/SiO₂ (403 K), and Rh/SiO₂ (403 K) catalysts too. Deuterium distributions were determined at the commonly used 403 K. Over each catalyst, in both reactor types three main transformation pathways were found. Single C-O scission occurred via isomerization and hydrogenation. Double C-O cleavage, that is, deoxygenation was also found to be important. Product distribution varied from catalyst to catalyst and it was influenced by the conditions, as well as the reactor types. Deuterium distribution revealed significant differences in the elementary surface processes over the catalysts. Detailed mechanistic interpretation of the transformations is offered.

Full text of DOI:10.1006/jcat.1998.2284

Journal of Catalysis 181, 28–36 (1999)
Article ID jcat.1998.2284, available online at http://www.idealibrary.com on
Transformations of Cyclohexene Oxide over Silica-Supported
Cu, Pd, and Rh Catalysts in H2/D2 Atmosphere
,1
Andra´s Fa´si and Istva´n Pa´linko´†
Chemical Research Center, Institute of Chemistry, Hungarian Academy of Sciences, Pusztaszeri u´t 59-67, Budapest, H-1025 Hungary;
†Department of Organic Chemistry, Jo´zsef Attila University, Do´m te´r 8, Szeged, H-6720 Hungary
Received May 27, 1998; revised September 9, 1998; accepted September 9, 1998
was tried with various epoxides (4–11), to the best of our
knowledge, the cyclohexene oxide-H2 (or D2) system
has never been the subject of detailed studies as far
as the identification of transformation types and the mech-
anism of reactions are concerned. Closest to it were the deu-
terogenation of (i) 1-methyl-1,2-epoxycyclobutane over
Ni (4) and Pd (5), (ii) 1-methyl-1,2-epoxycyclopentane over
Pt, Pd, and Ni (7), and (iii) the conformationally rigid
cis- and trans-tert-butylcyclohexene epoxide and their
1-methyl or 4,5-dimethyl substituted derivatives over Pd/C
in the liquid phase (8). For these molecules deuterium
distribution provided some information about the mode
of adsorption; however, mechanistic details were only
discussed for 1-methyl-1,2-epoxycyclobutane where the
cyclobutane ring showed peculiar behaviour (4, 5). The
stereochemistry of deuterium addition could be studied
with the conformationally rigid cyclohexene epoxides and
was found to be predominantly anti. It was considered as
the proof of the rollover of the O-(mono)adsorbed species
on the catalyst before the addition of the second hydrogen
(or deuterium). Otherwise the reactions in various reactors
were not compared and the mechanisms of other reactions
than single C-O scission were not investigated. Albeit
having only one metal, the Pd, common in this work and
the ones previously cited, considerations discussed there
may be useful here, except that direct evidences on the
stereochemistry of hydrogen (deuterium) addition cannot
be obtained with our model.
In this contribution we give experimental data on the
single as well as the double C-O scission reactions of cyclo-
hexene oxide with or without added hydrogen or deuterium
over silica-supported Cu, Pd, and Rh catalysts. Since a re-
action temperature applicable in the (closed) static recir-
culation reactor and the flow system for every catalyst was
found, we are able to compare phenomenological features,
such as the activities and selectivities of the catalysts. More-
over, deuterium distribution data obtained at this temper-
ature allows even deeper comparisons to be made; surface
events can be and are analysed in detail.
Transformations of cyclohexene oxide were studied in a recircu-
lation reactor and a flow system over silica-supported transition and
early transition metal catalysts. Hydrogen pressure dependence of
the transformations was measured over Cu/SiO2 at 403 K in the
0–98.7 kPa range. Deuterium distribution in the reactant and prod-
ucts were determined at 20 kPa D2 pressure. The reactions were
studied in the flow reactor under D2 stream over Cu/SiO2 (373
K, 403 K, 443 K, 523 K), Pd/SiO2 (403 K), and Rh/SiO2 (403 K)
catalysts too. Deuterium distributions were determined at the com-
monly used 403 K. Over each catalyst, in both reactor types three
main transformation pathways were found. Single C-O scission oc-
curred via isomerization and hydrogenation. Double C-O cleavage,
that is, deoxygenation was also found to be important. Product dis-
tribution varied from catalyst to catalyst and it was influenced by
the conditions, as well as the reactor types. Deuterium distribution
revealed significant differences in the elementary surface processes
over the catalysts. Detailed mechanistic interpretation of the trans-
c
formations is offered.
1999 Academic Press
Key Words: cyclohexene oxide; Cu/SiO2; Pd/SiO2; Rh/SiO2; hy-
drogenative ring opening; H-D exchange; mechanism of ring open-
ing.
INTRODUCTION
It is well known that due to the highly strained oxirane
ring epoxides may undergo easy ring opening. The ring
may be opened by all sorts of nucleophiles, making these
compounds versatile intermediates in organic synthesis
(1). The reaction can be performed stoichiometrically as
well as catalytically. For the latter method, beside applying
a nucleophile with Lewis acid as activator (nucleophilic
catalysis) or solid acids (electrophilic catalysis) (2), the ring
may be opened using metals or supported metals in the pre-
sence of hydrogen (3). When hydrogen is replaced by
deuterium, monitoring deuterium distribution in the reac-
tant and the product(s) allows deeper insight into surface
events composing the reactions. Although similar approach
1 To whom all correspondence should be addressed.
28
0021-9517/99 $30.00
c
Copyright
1999 by Academic Press
All rights of reproduction in any form reserved.

METAL CATALYSIS OF CYCLOHEXENE OXIDE
29
EXPERIMENTAL
temperature 323 K) with helium as the carrier gas (flow
rate 1.3 cm3/min).
Reactants and Catalysts
The reactants (0.8 kPa of cyclohexene oxide and various
amounts of H2 or D2) were premixed in the circulation part
of the system before the reaction. Cyclohexene oxide was
subjected to several freeze–evacuate–thaw cycles before
preparing the reaction mixture. The reaction temperature
in the hydrogen pressure-dependence measurements was
403 K. The hydrogen pressure range was 0–98.7 kPa.
Temperature dependence was studied at 20 kPa hydrogen
pressure in the 373 K–523 K temperature interval. Initial
rates were determined from the product yield versus time
(t) functions obtained at certain hydrogen pressures. These
rates were converted to turnover frequency data (TOF:
molecule exp.atom 1s 1) on the basis of exposed metal
atoms. Measurements were reproducible within 10% . They
were always performed over a fresh sample of catalyst
(20 mg), which was activated in the reactor under 26.6 kPa
hydrogen pressure at 573 K for 1 h. In the circulation
reactor only Cu/SiO2 was used.
Cyclohexene oxide was a Fluka product with the purity
of 99.9+% by the GC-MS method.
Hydrogen was produced by a Matheson 8326 hydro-
gen generator equipped with a Pd membrane. Deuterium
was obtained through the electrolysis of D2O (purity was
99.8% ) in a general Electric 15EHG generator. The gas was
deoxygenated with a MODEL 1000 Oxygen Trap (CRS,
USA) before use.
The Cu/SiO2 catalyst was prepared by the ion-exchange
method. Support wasStrem large pore (120-230mesh) silica
gel. It was immersed in a tetramminecopper(II) solution at
pH 11. After 24 h the blue product was washed thoroughly
with double distilled water and dried at 393 K (24 h). After
calcination (773 K, 3 h in air) the catalyst was reduced in a
hydrogen stream (673 K, 2 h) and kept in a vacuum desic-
cator until use. Metal loading was 6.36% . The percentage
of surface copper atoms was 14.6 by N2O titration at 363 K.
After reduction 7% of the copper atoms remained ionic,
possibly as Cu+ ions (12).
Deuterogenation was studied at 403 K on mixture with
0.8 kPa of cyclohexene oxide and 20 kPa of D2.
The Rh/SiO2 catalyst was prepared by impregnating
BDH Cab-O-Sil support with an acidic aqueous RhCl3 so-
lution. The catalyst was dried at 403 K for 16 h, washed
until neutral with double distilled water, and reduced un-
der hydrogen flow at 483 K for 3 h and then at 673 K for 4 h.
Finally, it was cooled to room temperature under nitrogen
flow and was stored in air in a vacuum desiccator. Rhodium
loading was 3.2% . Percentage of exposed rhodium atoms
was 27.0 by static hydrogen chemisorption at room temper-
ature.
The Pd/SiO2 catalyst was prepared similarly, but an aque-
ous PdCl2 solution was applied. Reduction and storing con-
ditions were the same as well. Palladium loading was 3.0% .
The percentage of exposed atoms was 10.0 by static hydro-
gen chemisorption at 340 K.
Flow Reactor
Reactions were carried out over 20 mg of Cu/SiO2 or
10 mg of Pd/SiO2 or 10 mg of Rh/SiO2 sample under D2
flow (15 cm3/min). The partial pressure of cyclohexene ox-
ide was 1.3 kPa. Four reaction temperatures (373 K, 403 K,
443 K, and 523 K) were used over the Cu/SiO2. Measure-
ments were performed at 403 K over the other two catalysts.
A fresh sample of catalyst was used in each measurement.
Activation conditions were as follows, Cu/SiO2: 20 cm3/min
D2 flow for 1 h at 573 K, Pd/SiO2 and Rh/SiO2: 20 cm3/min
D2 flow for 1 h at 423 K. Diffusion resistance was checked
by varying the quantity of the catalysts. The chosen amount
allowed investigations in the kinetic regime. Product accu-
mulation was followed by the GC-MS technique described
in the previous paragraph, exactly like in measurements in
the recirculation reactor.
Apparatus, Methodology, and Analysis
Deuterium Distribution
Static Recirculation Reactor
Deuterium distribution was calculated on the basis and
vicinity of the parent peaks. In the calculations the natural
abundance for 13C and D, the C-H and C-D fragmentation
reactions were taken into account.
Reactions were carried out in a conventional closed cir-
culation reactor. The volume of the reactor was 69 cm3, and
the total volume of the system was 169.4 cm3. The reactor
was heated with an air thermostat. The volume of the sam-
pling capillary was 0.05 cm3 and the total sampling volume
was 0.15 cm3. A Hewlett-Packard (HP) Model 5890 gas
chromatograph equipped with a quadrupole mass-selective
RESULTS
Cyclohexene oxide reacts over the catalysts with and
detector was attached to the system. Data analysis was without added hydrogen. Two reaction types are observed:
performed on an HP 5970 Chemstation. A 25-m-long single and double C-O cleavage. The main products are car-
Carbowax 20M capillary column (film thickness: 0.2 mm) bonyl compounds, alcohols, and cyclic hydrocarbons, de-
was used for separating the reactants and the products. pending on the catalyst, reactor type as well as reaction
The column was operated in the isothermal mode (oven conditions (Scheme 1).

´
´
´
30
FASI AND PALINKO
SCHEME 1. Products from the metal-catalyzed transformations of cyclohexene oxide.
Transformations of Cyclohexene Oxide over Cu/SiO2
in the Recirculation Reactor
The hydrogenative ring opening route increased its share
on increasing hydrogen pressure, just like deoxygenation.
This latter process, resulting from double C-O scission,
was a major transformation pathway from low hydrogen
pressures and it became predominant over 53.3 kPa.
Despite replacing hydrogen by deuterium the reactant
was free of deuterium, making it clear that the epoxide ring
is destroyed on touching the catalyst surface; that is, there is
no cyclohexene oxide desorption. There was no deuterium
in the cyclohexene product either after the 5-min sampling.
Later, deuterium appeared in cyclohexene as the result of
secondary H-D exchange of the cyclic olefin.
Without added hydrogen the reaction started with a long
induction period at 403 K. After 20 min cyclohexene (2)
and 1,3-cyclohexadiene (3) emerged. The reaction is deoxy-
genation (double C-O scission), followed by further hydro-
gen losses. When there was added hydrogen the induction
period disappeared. Deoxygenation and dehydrogenation
remained major transformation routes but further reaction
pathways also opened. The hydrogenative ring opening of
the epoxide ring gave cyclohexanol (5) and its isomeriza-
tion resulted in cyclohexanone (6) and 2-cyclohexane-1-ol
(7). Due to the limited hydrogenation ability of copper,
the main product was cyclohexene even with added hydro-
gen. Under these conditions the adsorbed olefin precursor
was not hydrogenated but dehydrogenated over the surface
and desorbed as 1,3-cyclohexadiene (3). On elevation of
the temperature further dehydrogenation may occur over
the surface of the catalyst, and in line with this benzene
(4) formation could be detected. Temperature dependence
measurements allowed the calculation of apparent activa-
tion energies for products appearing at each temperature.
They are given in Table 1.
The isomerization pathway of single C-O cleavage pro-
duced cyclohexanone-d1 and cyclohexanone-d2 in nearly
2 : 1 initial ratio. The products of the hydrogenative route
were cyclohexanol-d2 and cyclohexanol-d3 in 3 : 1 initial
TABLE 1
Apparent Activation Energies for the Overall
Transformation of Cyclohexene Oxide and the
Formation of Major Products
1
Compounds
Ea/kJmol
39.6
Turnover frequency data of initial product formation
rates at 403 K are listed in Table 2.
Reactant
1
At this temperature benzene was not found, even at
the lowest hydrogen pressure applied, and above 33.3 kPa
there was no 1,3-cyclohexadiene formation either. The oxi-
rane ring opened via isomerization in the entire hydrogen
pressure range, but it was suppressed at high hydrogen
pressures. The formation of the allyl alcohol analogue 2-
cyclohexene-1-ol ceased above 53.3 kPa hydrogen pressure.
Products
2
3
5
6
38.4
38.8
30.2
36.5
Note. 373 K–523 K temperature range, 0.8 kPa
cyclohexene oxide, 20 kPa H2.

METAL CATALYSIS OF CYCLOHEXENE OXIDE
31
Table 4. However, observations made at the other two tem-
peratures (373 K and 443 K) are also described in the fol-
lowing.
TABLE 2
4
Initial Product Formation Rates (Turnover Frequencies/10
Molecule exp.atom 1s 1) in the Function of H2 Pressure over Cu/
SiO2 Catalyst at 403 K in the Recirculation Reactor (0.8 kPa Cy-
clohexene Oxide, 20 mg Catalyst)
The catalyst stabilized after about half an hour, that is,
steady-state activity was attained at 373 K and 403 K. The
steady-state conversion levels were in the range of 5–7%
and 42–49% , respectively. At 443 K, although nearly full
conversion was achieved initially, the catalysts gradually
deactivated. At still higher temperature (523 K) a similar
phenomenon was observed. Products differed, depending
on temperature. At 373 K cyclohexene, cyclohexanol, cyclo-
hexanone, and 2-cyclohexene-1-ol were found. At 403 K
another product, 2-cyclohexene-1-one (8), could be de-
tected in small, but not negligible quantities. Although ini-
tially, cyclohexene was the most significant product, its
formation rate decreased over the steady-state catalyst at
both temperatures. The significance of 2-cyclohexene-1-ol
changed in the opposite way. It was either not observed over
the initial catalyst (373 K) or its quantity was low (403 K).
Later on, its formation rate increased and surpassed that
of the cyclohexene. The initial cyclohexanol formation was
significantly faster than that of cyclohexanone both at 373 K
and 403 K. However, once again, the tendency of formation
rate change of these two molecules was the opposite over
the steady-state catalyst. At the 120-min samplingtheywere
about the same at 373 K, while at 403 K cyclohexanone was
formed faster than cyclohexanol.
At higher temperatures the tendency of changes in the
formation rates of cyclohexene, cyclohexanone, cyclohexa-
nol was the same. However, products not seen at lower
temperatures appeared from 443 K: they are 1,3-cyclohexa-
diene and cyclopentane carbaldehyde (9). Cyclopentane
carbaldehyde was formed with a low rate at 443 K but it
became an important product at 523 K (210-min sampling).
Further dehydrogenation of the adsorbed cyclohexene
precursor and desorption as 1,3-cyclohexadiene started at
443 K and went to benzene at 523 K.
H2 pressure/kPa
Products 1.3
6.7 13.3 20.0 33.3 46.7 53.3 69.3 98.7
2
3
5
6
7
6.0 10.9 15.5 19.5 24.0 25.7 26.5 28.7 30.3
1.2
0
1.5
0.5
1.7
3.2
1.2
4.3
0.5
6.1
0
7.5
0
7.9
0
0
9.0 10.1
9.3 16.5 20.9 25.6 26.8 24.0 20.6 18.0 15.5
2.8 3.2 3.3 3.2 2.8
0
0
0
0
ratio. In both cases these ratios approached 1 : 1 on increas-
ing the residence time. Data are listed in Table 3.
Transformations of Cyclohexene Oxide over Cu/SiO2
Catalyst in the Flow Reactor
Similarly to the recirculation reactor, cyclohexene oxide
transformed in the flow system without added hydrogen
(under He stream). The products at 403 K were cyclohex-
anol, cyclohexanone, and 2-cyclohexene-1-ol. Contrary to
observations made in the recirculation reactor, neither
cyclohexene nor 1,3-cyclohexadiene (product of deoxy-
genation and further dehydrogenation) were formed. The
conversion was low, 1.2–1.5% , over the steady-state sur-
face. Isomerization and the hydrogenative ways of single
C-O scission were equally important transformation routes.
When the temperature was raised to 523 K the conversion
increased to 47% . Products of deoxygenation and further
dehydrogenation appeared but the isomerization pathway
of C-O bond scission was overwhelming (riso/rdeox = 6.7).
It is to be noted that benzene was not formed, even at this
temperature. No sign of the hydrogenative ring opening
was found either.
Transformations of cyclohexene oxide were studied un-
der deuterium flow as well, at various temperatures. Data
corresponding to 403 K (to allow comparison with product
distribution in the recirculation reactor) and 523 K (the
highest temperature applied) are shown as examples in
Deuterium distribution in the reactant and the products
was determined in the reaction at 403 K (Table 5).
There was no deuterium in cyclohexene oxide and it
was negligible in cyclohexene. Cyclohexanone-d1 was over-
whelming over the initial catalyst and there was not any
TABLE 3
Deuterium Distribution in the Ring-Opening Products of Cyclohexene Oxide over Cu/SiO2 Catalyst at 403 K in the Recirculation Reactor
Deuterium distribution/%
Cyclohexene (2)
Cyclohexanol (5)
Cyclohexanone (6)
t/min
Conv/%
d0
100
98.0
93.2
d1
d2
d2
d3
d4
d5
d0
d1
d2
d3
d4
5
20
35
8.3
42.0
66.0
0
2.0
5.1
0
0
1.7
66.2
46.4
31.9
22.1
33.3
40.3
11.7
20.3
24.9
0
0
2.9
11.2
3.4
1.7
56.5
43.3
31.6
31.0
40.6
41.1
1.3
12.7
22.4
0
0
3.2
Note. 20 mg catalyst, 0.8 kPa cyclohexene oxide, 20.0 kPa D2.

´ ´ ´
FASI AND PALINKO
32
Transformations of Cyclohexene Oxide over Rh/SiO2
TABLE 4
and Pd/SiO2 Catalysts in the Flow Reactor
Product Distribution in the Flow Reactor over Cu/SiO2
at 443 K and 523 K
The reactions and the parallel analysis of deuterium dis-
tribution in the reactants and the products were performed
at 403 K allowing certain comparisons to be made with ob-
servations over Cu/SiO2 under the same conditions.
As it was expected cyclohexene oxide reacted signifi-
cantly faster over Pd/SiO2 and Rh/SiO2 than over Cu/SiO2
(Table 6).
Conversions were close to 100% or even reached this
value over both transition metals. Three products were
only detected. They were cyclohexane (10), cyclohexanone,
and cyclohexanol. Cyclohexane could be derived from de-
oxygenation (double C-O scission), cyclohexanone was the
product of the isomerization route and cyclohexanol was
that of the hydrogenative pathway of single C-O cleavage.
The hydrogenative route was found to be the overwhelm-
ing transformation pathway on both metals. The isomeriza-
tion route was more significant over Pd/SiO2, while deoxy-
genation was the more important transformation pathway
over Rh/SiO2. Deoxygenation was faster than isomeriza-
tion over the initial Pd/SiO2. However, after the initial stage
the rate of cyclohexane formation decreased, finally, it was
even ceased.
Data concerning deuterium distribution in the reactant
and the product molecules are summarised in Table 7.
Deuterium in cyclohexene oxide was not found over
any of the catalyst, however, it could be observed in cy-
clohexane in every possible position (cyclohexene-d1 is not
one of them) with about the same abundance. Cyclohexa-
none-d1 is more abundant than cyclohexanone-d2 over
the initial catalysts and remains so over Rh/SiO2, albeit its
relative abundance decreases on continuous reactant flow.
Cyclohexanone-d3 is unimportant over Rh/SiO2, while
cyclohexanone-dy (y = 1–7) appears over Pd/SiO2 catalyst
with approximately equal probability at the 65-min sam-
pling. The relative importance of cyclohexanone-d4 was
found to be the highest among these species. Although
Product distribution/mol%
Products 5 min 20 min 35 min 50 min 65 min 120 min 210 min
403 K
2
5
6
7
8
32.2
21.1
0.5
0.3
0.3
20.0
35.6
0.6
0.6
0.6
10.3
37.0
0.8
1.0
0.7
9.6
29.3
1.5
1.7
0.5
9.5
26.5
2.7
2.6
0.5
8.2
10.6
16.8
8.0
5.8
8.2
18.9
11.1
1.3
0.4
Conv./% 54.4
57.4
49.8
42.6
41.8
44.4
45.3
523 K
2
3
4
5
6
8
9
31.7
0
24.3
0
22.7
0
20.2
0
19.9
0
17.7
0
11.0
5.0
0.5
56.4
4.7
0
1.1
61.2
5.7
0
1.2
61.8
5.7
0
1.2
61.7
6.8
0
1.0
62.0
6.5
0
1.1
59.5
9.4
0
0
27.0
36.0
0.4
0
0
92.3
0
91.4
0
89.9
0
89.4
0.4
88.1
6.0
Conv./% 93.3
85.4
Note. 20 mg catalyst, 1.3 kPa cyclohexene oxide, 15 cm3/min D2 flow.
cyclohexanone molecule without deuterium. The steady-
state catalyst exhibited peculiar behaviour. Increasing
amounts of cyclohexane-d0 were detected. The proportion
of cyclohexane-d1 dropped to close to half of the value
on the initial catalyst, but hardly changed any further.
The percentage of cyclohexanone-d2 increased to about
40% from the initial 11.5% , but dropped again as results
of later samplings attest. There was no cyclohexanol-d1
over the initial surface, while cyclohexanol-d2, d3, and d4
were detected in similar amounts. Over the steady-state
surface, however, cyclohexanol-d1 was found in increasing
proportions, just as cyclohexanol-d2, while the percentage
of cyclohexanol-d3 and d4 decreased on further samplings.
Note, that cyclohexanol-d5, d6, and d7 became less signifi-
cant in this order.
TABLE 5
Deuterium Distribution in the Ring-Opening Products of Cyclohexene Oxide over Cu/SiO2 Catalyst at 403 K in the Flow Reactor
Deuterium distribution/%
Cyclohexene (2)
Cyclohexanol (5)
Cyclohexanone (6)
t/min
d0
d1
d2
d1
d2
d3
d4
d5
d6
d7
d0
d1
d2
d3
5
20
35
50
65
96.5
96.9
98.0
98.8
98.2
99.2
99.7
3.2
3.0
2.0
1.2
1.8
0.8
0.3
0.3
0.1
0
0
0
0
0
0
25.3
26.8
27.5
32.2
32.9
51.6
62.4
33.3
34.9
33.8
32.4
32.7
25.5
16.4
25.4
25.1
25.7
23.8
22.9
12.5
5.3
11.6
9.8
9.7
8.9
8.3
3.8
0.7
4.3
3.2
3.1
2.6
2.7
0
0.1
0.2
0.2
0.1
0.1
0
0
0
88.5
100
11.5
0
0
0
0
6.2
8.7
2.0
0.2
10.6
9.7
15.3
35.9
47.9
49.3
47.7
42.6
45.8
43.0
40.1
36.4
33.4
16.3
8.9
0
0.4
6.6
15.2
120
210
0
0
0
0
Note. 20 mg catalyst, 1.3 kPa cyclohexene oxide, 15 cm3/min D2 flow.

METAL CATALYSIS OF CYCLOHEXENE OXIDE
33
occurs either via deoxygenation or the cleavage of one C-O
bond either through isomerization or hydrogenation. De-
tails are different, however.
TABLE 6
Product Distribution in the Flow Reactor over Pd/SiO2
and Rh/SiO2 at 403 K
As it may be expected, significant differences arise on
changing the catalysts. Perhaps, it is more surprising that
change in the reactor type also influences reaction charac-
teristics.
Product distribution/mol%
Products
5 min
20 min
35 min
50 min
65 min
Pd/SiO2
10
At the beginning, let us note that added hydrogen/
deuterium is needed for the reactions to occur over Rh/SiO2
and Pd/SiO2. However, there are reactions over Cu/SiO2
when there is no added hydrogen/deuterium. The reactions
are slow and they are different in the static and the flow
reactors. In the first one, an induction period indicates that
surface transformation is needed for the exclusive reaction,
the deoxygenation to occur. In the latter C-O bond scis-
sion via isomerization route and even hydrogenation took
place. For this reaction hydrogen is possibly obtained from
hydrogen-rich weakly adsorbed carbonaceous residues re-
sembling the reactant. Deoxygenation products were not
found in the flow reactor, nevertheless, 2-cyclohexene-1-ol
mayverywellbe the intermediate not desorbingin the static
system on the way to cyclohexene.
Now, let us compare the main features of the three main
reaction pathways in the presence of deuterium at the com-
monly used 403 K in the flow system. The activities of the
catalysts largely differ. The transition metal catalysts are
very much more active than Cu/SiO2. Their hydrogenation
ability also differs significantly. The deoxygenated product
is an olefin over Cu/SiO2, while it desorbs hydrogenated
from Rh/SiO2 and Pd/SiO2. The deoxygenation pathway
is significant over Cu/SiO2 and Rh/SiO2, but it is negligible
over Pd/SiO2. The hydrogenative C-O bond scission is the
3.4
88.6
2.7
1.0
96.1
2.2
0.5
97.3
2.2
0.3
97.7
2.0
0.1
98.0
1.9
5
6
Conv./%
94.7
99.3
100
100
100
Rh/SiO2
10
5
6
26.4
71.6
0.6
19.3
77.5
0.5
14.4
83.5
0.7
10.7
88.7
0.4
10.3
89.1
0.5
Conv./%
98.6
97.3
98.6
99.8
99.9
Note. 10 mg catalyst, 1.3 kPa cyclohexene oxide, 15 cm3/min D2 flow.
multiply exchanged cyclohexanol molecules were ob-
served over both catalysts, cyclohexanol-d2 is absolutely
overwhelming over Pd/SiO2. Its abundance is the highest
over Rh/SiO2, nevertheless, significant amount of cyclo-
hexanol-d3 was also formed over this metal.
DISCUSSION
Cyclohexene oxide reacts over silica-supported Cu, Pd as
well as Rh catalysts under relatively mild conditions. Over-
all, the picture is similar over each catalyst. In the presence
of H2/D2 three main reaction types take place. Each reac-
tion leads to the disappearance of the oxirane ring. This
TABLE 7
Deuterium Distribution in Cyclohexanol and Cyclohexanone Derived from the Hydrogenative Ring Opening of Cyclohexene Oxidea
over Pd/SiO2 and Rh/SiO2 at 403 K in the Flow Reactor
Deuterium distribution/%
Cyclohexanol (5)
Cyclohexanone (6)
t/min
d2
d3
d4
d5
d6
d7
d8
d9
d10
d11
d1
d2
d3
d4
d5
d6
d7
Pd/SiO2
5
20
35
50
65
88.4
88.4
88.4
88.4
88.4
8.0
8.0
8.0
7.8
8.0
2.1
2.1
2.2
2.3
2.2
1.4
1.4
1.4
1.5
1.4
0.1
0.1
0.1
0.1
0.1
0
0
0
0.1
0.1
0
0
0
0
0
0
0
0
0
0
0
0
37.1
18.5
6.5
3.4
2.8
18.3
10.8
12.7
7.6
13.7
18.9
18.8
17.6
19.6
14.2
23.1
24.3
25.5
23.3
11.9
14.6
19.2
22.0
20.4
4.8
12.0
11.6
15.9
14.8
0
0
0
0
0
0
0
0
0
2.1
6.9
8.0
7.8
11.3
Rh/SiO2
5
20
35
50
65
47.3
45.5
47.8
54.6
60.0
17.5
19.1
20.6
20.4
19.3
10.9
11.8
11.8
10.4
8.8
8.3
8.6
8.1
6.5
5.6
5.9
6.0
5.3
4.0
3.2
4.1
3.9
2.9
2.0
1.6
2.7
2.5
1.8
1.1
0.9
1.9
1.6
1.1
0.7
0.5
1.1
0.9
0.6
0.3
0.1
0.3
0.1
0
0
0
86.6
100
60.6
69.2
48.4
13.4
0
27.5
24.2
38.5
0
0
11.9
6.6
13.1
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
Note. 10 mg catalyst, 1.3 kPa cyclohexene oxide, 15 cm3/min D2 flow.
a For deuterium distribution in cyclohexane, see text.

´ ´ ´
FASI AND PALINKO
34
major transformation pathway over Rh/SiO2 and Pd/SiO2 2-epoxycyclobutane (5) and 1-methyl-1,2-epoxycyclopen-
throughout the measurementsand it ismore important than tane (7). Isomerization of both compounds occurred over
the isomerization route in the initial stages of the run over Pd and the hydrogenative route of C-O scission was not de-
Cu/SiO2 as well. On this catalyst the relationship between tected. It is to be remarked that during this transformation
the two waysofsingle C-O bond scission turnsto the reverse the bridge C-C bond was cleaved first. The oxo compound
later. Deoxygenation transforms the surface of the Cu/SiO2 from this molecule in the deuterium stream was mostly
catalyst dramatically. Further deoxygenation and hydro- cyclopentanone-d2. Nearly 60% of the oxo compound from
genative C-O scission are diminished gradually over the 1-methyl-1,2-epoxycyclopentane was deuterium-free. The
working surface and the isomerization ability of the catalyst most abundant of those containing deuterium was 2-meth-
increases dramatically. Deoxygenation promotes the for- ylcyclopentanone-d2 (11.2% ).
mation of surface oxides; that is, ionic centres become abun-
As another preliminary, let us give a brief and necessarily
dant. Therefore, isomerization is favoured to hydrogena- simplified picture about the possible geometrical arrange-
tion over Cu/SiO2. It is to be noted that the ionic surface ments of our reactant, the cyclohexene oxide. For making
becomes active in rearrangement reactions at high enough, considerations visual molecules are placed on a flat sur-
but still moderate (443 K), temperatures. This reaction, face (first simplification). The geometry of the molecule is
leading to cyclopentane carbaldehyde, is a typical ionic ring optimized as an isolated entity without added hydrogen/
contraction. The noble metals are much less prone to sur- deuterium (second simplification), with the AM1 semiem-
face oxidation. Here hydrogenation remains overwhelm- pirical method (13) (third simplification) in minimum con-
ing. Since the activity towards direct deoxygenation over formation but without concentrating on finding the global
Cu/SiO2 decreases there is more chance for the formation minimum (fourth simplification). The possible arrange-
of the intermediate on the deoxygenation pathway, con- ments are depicted in Fig. 1.
sequently, 2-cyclohexene-1-ol and at higher temperature
2-cyclohexene-1-one formation rates gradually increase.
Adsorption form A represents edgewise attachment, B
depicts flat adsorption, and C is adsorption on the tip of the
A comparison of the transformation properties over epoxide ring. All those arrangements are discarded where
Cu/SiO2 in the flow system and the static recirculation re- the oxygen atom is not in close vicinity of the surface, since it
actor reveals that increased residence time and moderate is known that (irreversible) adsorption starts at the oxygen
hydrogen availability facilitates further dehydrogenation atom. In form A the hydrogen atom on the pillar carbon and
during deoxygenation. In the static, recirculation system cy- another on the neighbouring carbon are in easy exchange
clohexadiene is also formed, while cyclohexene is the sole positions. The two hydrogens on the next carbon are not
product not having oxygen in the flow reactor. Hydrogen too far from the surface, therefore, they might participate in
availability is crucial in determining cyclohexanone forma- H-D exchange. Adsorption form B may allow the exchange
tion rates relative to the hydrogenative way of single C-O of hydrogens on the C-C bridge. Seemingly, form C should
scission. Cyclohexanone formation, that is, the isomeriza- be discarded at once, since none of the C-O bonds are in
tion route of ring opening, is faster in the recirculation reac- contact with the surface. Nevertheless, we decided to keep
tor than the hydrogenative route of C-O cleavage. It is the this form because of a mechanism described by Teplyakov
reverse in the flow system until deoxygenation transforms and Bent recently (14). It is a type of the Eley–Rideal
the catalyst to an oxidic surface with abundant ionic centres. mechanism, where the H2/D2 molecule is loosely attached
Then, isomerization becomes faster than hydrogenation.
to the surface at one end of the molecule and the hydro-
Before analyzing H-D exchange results let us summa- gen/deuterium on top attacks. Thus, a ring opening of this
rize those pieces of information from earlier works which type might be a viable form of transformation route and
concern the Pd-catalyzed deuterogenation of 1-methyl-1, can be an alternative explanation for the anti addition in
FIG. 1. Adsorption geometries of cyclohexene oxide over a flat surface: A, edgewise; B, flat; C on tip.

METAL CATALYSIS OF CYCLOHEXENE OXIDE
35
conformationallyrigid cyclohexene epoxidesto the rollover
TABLE 8
mechanism (8). The rollover of such bulky adsorbed
molecules is difficult to envisage anyway. In form C two
hydrogen atoms of the C-C bond in the cyclohexane ring
opposite to the oxirane ring get close to the surface. Be-
fore going into detail let us nail down once again that the
lack of exchanged cyclohexene oxide shows that the oxirane
ring is destroyed on landing on the metal surface. Further-
more, let us distinguish two main adsorption types, depend-
ing on what bond(s) is (are) broken in the first step of sur-
face transformations. The cleavage of the O-C bond leads
the associatively adsorbed species (associative adsorption),
while the concomitant scission of the O-C bond and one or
more C-H bonds results in dissociatively adsorbed moieties
(dissociative adsorption). Cyclohexanone can only form
through dissociatively adsorbed species, whatever adsorp-
tion geometries are assumed. Form A would facilitate the
cleavage of the O-C bond on the surface, providing the pos-
sibility for cyclohexanol formation only. (Cyclohexanone
formation would need the dissociation of a hydrogen from
the carbon atom far from the surface.) If there is associative
adsorption, two deuterium atoms will be in the cyclohex-
anol. If it is dissociative, then the number will be three, but
it also can be four (and it is more probable) as the neigbour-
ing carbon has a hydrogen close to the surface, which may
take part in the primary adsorption process. In a secondary
exchange two more hydrogens may take place. When form
B is assumed dissociative adsorption seems to be the most
probable. Beside the O-C bond two hydrogens (on the C-C
bridge) have the possibility of taking part in the primary
process. Here, cyclohexanone would contain two and cy-
clohexanol four deuterium atoms. The products may form
via -allyl adsorbed species too. This means dissociative ad-
sorption (thus, cyclohexanone may also form), which would
lead to one new deuterium in cyclohexanone or three new
deuterium atoms on alcohol formation. The other hydrogen
on the C-C bridge then can take part in secondary H-D ex-
change. Perhaps, associative adsorption may take place for
B as well. Then d2 alcohol may form (and, of course, there is
no cyclohexanone). Although form C was not discarded at
once, but it is taken as not very possible since a rational ex-
planation for O-C bond scission on adsorption is not found.
If it happens, it is an associative adsorption giving two new
deuterium atoms in cyclohexanol. However, on adsorption
H-D exchange may occur parallel to the ring opening. Then
two hydrogens on the C-C bond opposite to the bridge
C-C bond may take part in the process. The overall result
will then be four new deuterium atoms in the molecule. The
above described possibilities are summarized in Table 8.
A comparison of the experimental findings and consid-
erations summarized above and in Table 8 results in the
following:
Deuterium Distribution to Be Expected from Primary H-D
Exchange of Adsorbed Species Based on the Model in Fig. 1
Geometry of
adsorption
Mode of
adsorption
Cyclohexanone
Cyclohexanol
A (edgewise)
Associative
Dissociative
-Allyl
Associative
Dissociative
-Allyl
Associative
Dissociative
-Allyl
—
—
—
—
d2
d1
—
—
—
d2
d3 and/or d4
—
B (flat)
Perhaps, d2
d4
d3
d4
—
—
C (on tip)
system; i.e., the adsorption is dissociative. In the flow system
this is the main adsorption type over the fresh Pd/SiO2 and
clearly overwhelming over the fresh Cu/SiO2 and Rh/SiO2
catalysts as well. When the catalysts are oxidized in the
course of transformation, other adsorption type, dissocia-
tive through flat-lying species, becomes competitive in both
reactor types over Cu/SiO2 and over Rh/SiO2 in the flow
system. After the initial stages Pd/SiO2 displays close to
even distribution in the d1-d7 region. An explanation has
not been found yet and it is clearly beyond the capabilities
of our simplified model.
(2) Cyclohexanol-d2 is predominant in the recirculation
reactor over the initialCu/SiO2 and over Rh/SiO2 in the flow
system throughout the run. It is overwhelming over Pd/SiO2
at each sampling. Thismolecule indicatesthe hydrogenative
ring opening via edgewise associatively adsorbed species.
Ring opening through either edgewise or flat-lying, but cer-
tainly dissociatively adsorbed species becomes competitive
over the used/transformed Cu/SiO2 surface.
(3) Cyclohexene coming from double C-O scission is
practically not exchanged, indicating that the cyclohexane
ring is not in contact with the surface. Over the hydrogenat-
ing catalyst cyclohexane with every possible deuterium po-
sition shows that H-D exchange is faster than hydrogena-
tion followed by desorption.
The above described simplified model performed satis-
factorily in rationalization of adsorption leading to single
C-O scission. Characteristic adsorption types are depicted
in Fig. 2.
CONCLUSIONS
It has been demonstrated through the metal-catalyzed
transformations of cyclohexene oxide that if a common
temperature can be found, analysis and comparison of pro-
duct distribution and deuterium distribution in the prod-
(1) Cyclohexanone isformed mainlyvia -allyl adsorbed ucts in different reactor types, over various metals can lead
species over the initial Cu/SiO2 catalyst in the recirculation to a deeper understanding of surface transformations. It

´
´
´
36
FASI AND PALINKO
2. Barto´k, M., in “The Chemistry of Functional Groups. Supplement E2:
The Chemistry of Hydroxyl, Ether and Peroxide Groups” (S. Patai,
Ed.), Chap. 15, p. 843. Wiley, Chichester/Brisbane/New York/Toronto/
Singapore, 1993.
3. Barto´k, M., and Notheisz, F., in “Stereochemistry of Heterogeneous
Metal Catalysis,” Chap. IX, p. 435. Wiley, Chichester/Brisbane/New
York/Toronto/Singapore, 1985.
4. Duchet, J. C., and Cornet, D., Bull. Soc. Chim. Fr., 1135 (1975).
5. Duchet, J. C., and Cornet, D., Bull. Soc. Chim. Fr., 1141 (1975).
6. Se´ne´chal, G., and Cornet, D., Bull. Soc. Chim. Fr., 773 (1971).
7. Se´ne´chal, G., and Cornet, D., Bull. Soc. Chim. Fr., 783 (1971).
8. Accrombessi, G. C., Geneste, P., Olive´, J.-L., and Pavia, A. A., J. Org.
Chem., 45, 4139 (1980).
FIG. 2. Adsorbed intermediates on the way to cyclohexanone and/or
cyclohexanol: (a) flat-lying, -allyl; (b) flat-lying, dissociative; (c) edge-
wise, associative; and (d) edgewise, dissociative.
9. Fa´si, A., Notheisz, F., and Barto´k, M., J. Catal. 167, 242 (1997).
10. Fa´si, A., Notheisz, F., and Barto´k, M., J. Catal. 169, 114 (1997).
11. Fa´si, A., Notheisz, F., and Barto´k, M., J. Catal. 175, 40 (1998).
has also been shown that the possibilities are enhanced if
models of realistic background can be used.
´
12. Pa´linko´, I., Molna´r, A., Kiss, J. T., and Barto´k, M., J. Catal. 121, 396
REFERENCES
(1990).
13. Hyperchem 4.5, AM1 routine, Hypercube Inc., 1994.
1. Denmark, S. E., Barsanti, P. A., Wong, K.-T., and Stavenger, R. A., 14. Teplyakov, A. V., and Bent, B. E., J. Chem. Soc., Faraday Trans. 91,
J. Catal. 63, 2428 (1998) and appropriate references therein.
3645 (1995).