The effect of dehydration in the oxidation of ethylbenzene to acetophenone with supported catalysts

Article abstract of DOI:10.1021/op0000632

The influence of the product water formed in the catalytic oxidation of ethylbenzene to acetophenone using air in a batch reactor was studied. The water adsorbed in the hydrophilic catalyst supports, that is, alumina or silica, has a significant negative effect on the oxidation processes. The application of a dehydration unit in the catalytic oxidation of ethylbenzene with air in a batch reactor was studied using sulphuric acid, molecular sieves, and silica gel as dehydrants. Processes were developed to use the dehydration unit to dehydrate the condensed product, from a Dean Stark trap, and the reaction mixture. Experimental results confirmed that an improvement in reaction performance was achieved using the dehydration unit to decrease the water concentration in the reaction mixture below its saturation solubility. The effect of a continuous increase in the temperature of the oxidation process, compared to isothermal operation, is shown to improve the oxidation performance.

Full text of DOI:10.1021/op0000632

Organic Process Research & Development 2001, 5, 204−210
Articles
The Effect of Dehydration in the Oxidation of Ethylbenzene to Acetophenone
with Supported Catalysts
Simon T. Man,†,§ Colin Ramshaw, Keith Scott,* James Clark, D. J. Macquarrie, and Roshan Jachuck
†
,†
‡
‡
†
Department of Chemical and Process Engineering, UniVersity of Newcastle upon Tyne, Newcastle upon Tyne, NE1 7RU,
UK, and Department of Chemistry, UniVersity of York, Heslington, York YO1 5DD, UK
Abstract:
due to the adsorption of water vapour on both the support
and the metal ion sites, with the impact on the support being
more significant than that on the catalyst sites.4
In this paper, the role of water in the oxidation of organic
substrates with air using alumina- and silica-supported
catalysts is discussed. The application of a dehydration unit
for the catalytic oxidation of ethylbenzene to acetophenone
with three dehydrants: sulphuric acid, molecular sieves, and
silica gel, is described.
The influence of the product water formed in the catalytic
oxidation of ethylbenzene to acetophenone using air in a batch
reactor was studied. The water adsorbed in the hydrophilic
catalyst supports, that is, alumina or silica, has a significant
negative effect on the oxidation processes. The application of a
dehydration unit in the catalytic oxidation of ethylbenzene with
air in a batch reactor was studied using sulphuric acid,
molecular sieves, and silica gel as dehydrants. Processes were
developed to use the dehydration unit to dehydrate the
condensed product, from a Dean Stark trap, and the reaction
mixture. Experimental results confirmed that an improvement
in reaction performance was achieved using the dehydration
unit to decrease the water concentration in the reaction mixture
below its saturation solubility. The effect of a continuous
increase in the temperature of the oxidation process, compared
to isothermal operation, is shown to improve the oxidation
performance.
2. Microeffects of Water in Catalysts
The schematic representation of Cr ions bound on a
chemically modified support surface is shown in Figure 1.
6
+
3+
The active Cr (dichromate) or Cr (chromium) cations
used as catalysts are chemically immobilised on both external
and internal surfaces of the supports, on which the catalytic
reaction takes place. In the catalytic oxidation of ethylben-
zene
EPAD or CHRISS
ethylbenzene + air
8
1
. Introduction
Many catalysts used in organic oxidation are based on
acetophenone + benzoic acid + water (1)
metal ions immobilised on porous supports such as silica
gel and alumina. The use of a heterogeneous catalyst, based
on alumina-supported dichromate [EPAD (EPAD is the
commercial name of the catalyst manufactured by Contract
Catalysts, Knowsley Business Park, Merseyside)], for the
oxidation of ethylbenzene to acetophenone with air has been
both ethylbenzene and oxygen must transfer from bulk liquid
to the active Cr cation sites on both external and internal
surfaces of solid supports, and acetophenone, benzoic acid,
and water must transfer from the active Cr cation sites to
bulk liquid.
On alumina the amount of chemisorbed water is propor-
tional to the surface area, corresponding to binding of one
2
1
recently reported. A more recent communication reported
a heterogeneous catalyst, based on silica gel-supported
chromium (CHRISS), for the oxidation of the alkyl aromatics
H O to two O atoms in the surface. Additional water is
7
adsorbed on the monolayer physically. Study of dielectric
2
with air. These studies showed that the oxidation was
behaviour of water adsorbed on γ-alumina suggested that in
hindered by long induction periods and catalyst deactivation,
which limited overall reaction rates. It has been reported that
water acts as an inhibitor for catalytic organic oxidation,
a multilayer system the first layer is strongly bound, while
layers above the first freely orient in the applied field. Bound
8
3
-6
water continues to be lost from γ-alumina even after heating
9
to 1000 °C. Aluminas used as catalyst supports are prepared
*
Corresponding author. E-mail address: k.scott@ncl.ac.uk.
University of Newcastle upon Tyne.
University of York.
†
by heating hydrated oxide to various temperatures so that
‡
§
Tong Wen.
(4) Zhang, M.; Zhou, B.; Chuang, K. T. Appl. Catal., B 1997, 13, 123.
(5) Papaefthimiou, P., etc. Appl. Catal., B 1998, 15, 74.
(6) Sunada, F.; Heller, A. EnViron. Sci. Technol. 1998, 32, 282.
(7) DeBoer, J. H.; Fortuin, J. M. H.; Lippens, B. C.; Meijs, W. H. J. Catal.
1963, 1.
(
1) Chisem, I. C.; Martin, K.; Shieh, M. T.; Chisem, J.; Clark, J. H.; Jachuck,
R.; Macquarrie, D. J.; Rafelt, J.; Ramshaw, C.; Scott, K.Org. Process Res.
DeV. 1997, 1, 365.
(
2) Chisem, I. C.; Rafelt, J.; Shieh, M. T.; Chisem, J.; Clark, J. H.; Jachuck,
R.; Macquarrie, D. J.; Ramshaw, C.; Scott, K. Chem. Commun. 1949 1998.
3) VandeBeld, L.; et al. Chem. Eng. Sci. 1994, 49, 4361.
(8) Baldwin, M. J.; Morrow, J. C. J. Chem. Phys. 1962, 36, 1591.
(9) Peri, J. B.; Hannan, R. B. J. Phys. Chem. 1960, 64, 1526.
(
2
04
•
Vol. 5, No. 3, 2001 / Organic Process Research & Development
10.1021/op0000632 CCC: $20.00 © 2001 American Chemical Society and The Royal Society of Chemistry
Published on Web 02/24/2001

In the first few hours of the ethylbenzene oxidation at
ambient pressure and 130 °C, water is removed from the
surface of the catalyst particles while the water in the core
of the catalyst particles diffuses to the surface due to the
water concentration gradient. This dehydration process
removes the liquid water in the mesopores and the water
released during the dehydration of alumina, enabling greater
access of ethylbenzene to the Cr cation sites in the porous
structure. The oxidation of ethylbenzene can hardly start until
the catalyst is partially dehydrated, an “induction” period is
1
Figure 1. Effects of water covering active Cr site.
usually observed.
the surfaces may be partially or wholly dehydrated. The
activity of the alumina depends critically on the pretreatment,
Assuming that the mass transfer rate of water, from
catalyst sites to bulk liquid, is lower than the rate of
production of water by reaction, water accumulation can lead
to a thin “water film” covering, at least partially, the active
Cr cation sites (Figure 1), which has significant negative
consequences. As ethylbenzene is virtually insoluble and
benzoic acid and acetophenone are only slightly soluble in
1
0
subsequent exposure to moist air, and other factors.
Silica gel, SiO , is a form of amorphous silica that is
obtained by the hydrolysis of alkoxides such as Si(OEt) ; it
often contains ∼4% water. Si NMR suggests the presence
(OH) in
2
4
of Si(OSi ≡)
4
, Si(OSi ≡)
3
(OH), and Si(OSi ≡)
2
2
11
5+
3+
silica gel. Previous studies suggest that one water molecule
is adsorbed for every two silanol groups on the gel surface.
Loss of water from silica gel may begin at 50 °C, reaching
water, a thin “water film” covering the active Cr or Cr
sites will, at least, increase the mass transfer resistance for
both reactants and products, or in the worst case, block the
mass transfer completely and hence inhibit the oxidation
process.
12
a maximum rate of desorption at about 140 °C. Due to the
hydrophilicity of alumina and silica, water can wet the
surface of the supports and be drawn into the mesopores.
As the support is mesoporous (e.g., pore diameter 20 Å for
EPAD, or 100 Å for CHRISS), there will be a considerable
hydrostatic pressure due to the capillary rise.13 This means
that transfer of water from the mesopores is more difficult
than that of ethylbenzene, which is the opposite the process
requires.
3
. Experimental Section
.1. Catalysts, Chemicals, and Sample Analysis. Cata-
lysts were made according to the procedures previously
3
1
2
described (EPAD and CHRISS ). The dichromate/alumina
catalyst EPAD has a dichromate loading of .∼0.075 mmol
-
1
g
(determined by atomic absorption spectroscopy), an
For the three-phase catalytic oxidation of ethylbenzene
average pore size of 20 Å, a particle size of 2-30 µm, and
(see eq 1) studied, product water has to be transported from
2
-1
a surface area of 86.9 m g . The chromium/silica catalyst
the active sites of internal (and external) catalyst surface to
the bulk liquid by diffusion. The driving force is the different
water concentration between the active sites and the bulk
liquid. In current industrial processes, to achieve required
product specification, the water of reaction is usually
removed by condensing the water and organic vapours in
the reactor headspace.14 With a Dean-Stark trap, or equip-
ment operating on the same principle, condensed water is
separated and removed while condensed organic (that has a
water concentration equal to its equilibrium solubility) flows
back to the reactor directly.
In a liquid aromatic reaction system with a Dean-Stark
trap, the dissolved water concentration in the reaction mixture
is quite low (e.g., at 18 °C and ambient pressure, the
equilibrium solubility of water in ethylbenzene is ∼300 ppm).
However, additional removal of water, as proposed in this
work, should benefit the reaction performance in terms of
conversion, yield, and reaction rate.
-
1
CHRISS has a chromium loading of .∼0.10 mmol g
(
determined by atomic absorption spectroscopy), an average
pore size of 100 Å and a particle size of 30-140 µm.
The chemicals used were GPR grade and obtained from
Aldrich, England. Chemical analysis was by gas chroma-
tography using a Unicam 610 system, with a capillary
column and a FID detector.
3
.2. Reactors. Two reactors were used in this study, a
small glass reactor fabricated in-house and a computer-
controlled autoclave.
The small reactor reaction system consisted of a 500 cm3
vessel fitted with baffles, a stirrer, an air sparger, a reflux
condenser, a decanter,and a heater (Figure 2). A thermo-
couple in the reactor was linked to a hot plate to control the
temperature. A chiller was used to supply cooling water to
the condenser at 0-1 °C to minimise loss of aromatic
compounds from the reactor. Air was supplied to the reactor
from a cylinder (industrial grade, obtained from BOC).
Initial studies of the oxidation of ethylbenzene considered
the following ranges of parameters. The catalyst amount was
changed from 0.125 to 5.0 g, air flow rate was changed from
Due to the hydrophilicity of alumina and silica supports,
the catalysts can adsorb moisture from air during storage.
(
10) Cotton, F. A.; Wilkinson, G. AdVanced Inorganic Chemistry, 5th ed.; John
Wiley & Sons Inc.: New York, 1988; pp 211-212.
3
-1
3
-1
2
00 cm min to 600 cm min , temperature was changed
(11) King, R. B. Encyclopaedia of Inorganic Chemistry; John Wiley & Sons
Inc.: Exeter, England, 1994; p 3768.
from 115 °C to 130 °C, and agitator speed was changed from
3
(12) Darlow, B. B.; Ross, R. A. Nature 1963, 198, 988.
500 to 2000 rpm, using 300 cm of ethylbenzene and
(
13) Glasstone, S.; Lewis, D. Elements of Physical Chemistry, 2nd ed.; D. Van
Nostrand Company, Inc.: reprinted in Hong Kong, 1980; pp 140-143.
14) Nardin, D. Chemspec Eur. 95 BACS Symp. 1995, 594.
operating at atmospheric pressure. The preliminary “opti-
mised” experimental conditions were
(
Vol. 5, No. 3, 2001 / Organic Process Research & Development
•
205

dehydrants were tested: 98% sulphuric acid, 1.7-2.4 mm
and 2.4-4.5 mm bead 3 Å molecular sieves (theoretical
water adsorption capacity 21%), and 2-5 mm granulated
self-indicating silica gel (water adsorption capacity 28.5%).15
Two dehydration strategies were tested: (1) dehydration
of the condensed organic phase from the Dean-Stark trap
and (2) direct dehydration of the whole reaction mixture.
Figure 2 shows a schematic of the reaction system with the
dehydration unit, used in conjunction with a Dean-Stark
trap. Using a Dean-Stark trap (or equipment operating on
the same principle) alone, condensed water is separated and
removed, while the condensed organic phase, that has a
water concentration equal to its equilibrium solubility, flows
back to the reactor directly. Thus, the water concentration
in the reaction mixture is approximately equal to its
equilibrium solubility at the reaction temperature.
However, using the dehydration unit after the Dean-Stark
trap (Figure 2), the condensed organic phase is further
dehydrated before it flows back to the reactor. Thus, the
water concentration in the reaction mixture is less than its
equilibrium solubility at the reaction temperature.
The dehydration unit consisted of a glass column (internal
Figure 2. Dehydration processes.
diameter 48 mm, height 350 mm) filled with 500 cm3
dehydrant: either molecular sieves (360 g) or silica gel (395
g). The molecular sieves were thermally activated at 250
°C and the silica gel thermally activated at 150 °C before
amount of catalyst in reactor ) 1.0 g
3
-1
air flow rate
temperature
agitator speed
) 400 cm min
) 130 °C
) 1500 rpm
3
testing. Before each oxidation test, 250 cm of ethylbenzene
was added to the glass column to fill the porous space. The
working temperature of the dehydration unit was ambient,
at approximately 20 °C.
All of the oxidations performed with the small glass
reactor were carried out with the above experimental
conditions unless otherwise indicated. Each test was carried
out with fresh ethylbenzene, and the reactor was cleaned prior
to each test.
With the sulphuric acid as the dehydrant, the glass column
was filled with glass balls to decrease the volume of the
sulphuric acid and improve mass transfer. With this dehydra-
tion unit, the condensed aromatic phase entered the bottom
of the column and dispersed into small drops in the sulphuric
acid. The sulphuric acid and aromatics were separated into
two layers due to their density difference, with sulphuric acid
at the bottom. The dehydrated ethylbenzene top layer then
flowed back to the reactor.
For the catalytic oxidation of ethylbenzene at 130 °C and
ambient pressure, the condensed organic phase consisted
mainly of ethylbenzene due to its lower boiling point, relative
to those ofthe products (136 °C for ethylbenzene, 202 °C
for acetophenone and 249 °C for benzoic acid).
The computer-monitored autoclave reactor consisted of
3
a 1000 cm glass autoclave (AE P/N 5010-2019 Glass
Reactor, Autoclave Engineers Group, Erie, Pennsylvania,
USA) with baffles, stirrer, a sparger, a reflux condenser, and
a decanter. The autoclave was heated by circulated hot oil.
A chiller was used to supply cooling water to the condenser
at 2-3 °C to minimise organic chemical loss from the
reactor. The air was supplied to the reactor from a cylinder.
A computer with a software package (wizcon) was used to
monitor the stirrer speed, reactor temperature, and pressure.
The oxidation tests, duration 8 h, with the autoclave
3
reactor were carried out with 600 cm of reaction mixture
To dehydrate the reaction mixture, a stream is continu-
ously pumped from the reactor, through a filter, to retain
the solid catalyst particles in the reactor, to the dehydration
unit, and back to the reactor. In the absence of catalyst and
an air supply to the dehydration unit, and with a working
temperature in the dehydration unit of approximately 20 °C,
it is assumed that no ethylbenzene oxidation occurs in the
dehydration unit (i.e., all oxidation products were produced
in the reactor). In this case, the dehydration unit consisted
and 2.0 g of CHRISS catalyst and an air flow rate of 800
3
-1
cm min . The temperature was varied between 130 and
20 °C; the agitator speed was varied from 600 to 300 rpm,
1
and the pressure was varied from 2.0 to 3.0 bar. Each test
used fresh feed and catalyst, and the reactor was cleaned
prior to each test.
3
.3. Dehydrants, Dehydration Processes, and the
Dehydration Units. For the proposed dehydration processes
the dehydration medium should meet the following require-
ments: (1) not react chemically with the catalysts, organic
reactants, and products, (2) be easily separated from the
organic phase, (3) provide the maximum practical extent of
water removal, and (4) be economically feasible (e.g.
molecular sieves are cheap and can be regenerated). Three
3
of a glass column filled with 500 cm of dehydrant, which
could be used for several 24-h runs until the dehydrant was
3
saturated with water. For the first 24-h run, 320 cm of
(
15) Sigma-Aldrich: Mineral adsorbents, filter agents and drying agents;
Aldrich Technical Information Bulletin Number AL-143.
206
•
Vol. 5, No. 3, 2001 / Organic Process Research & Development

Table 1. Aerial oxidation of ethylbenzene with dehydration of condensed organic phase (24 h)
conversion to
acetophenone
(%)
conversion to
benzoic acid
(%)
total
conversion
(%)
increase of
total conversion
(%)
catalyst
(CHRISS)
dehydration
unit
test
1
2
3
4
5
without
with
with
with
with
without
without
with
with
with
43.1
48.7
53.4
54.2
53.2
2.6
6.7
6.8
7.8
7.4
45.7
55.4
60.1
62.0
60.5
-
9.7
a
14.4
16.3
14.8
b
c
a
3
b
3
Test 3 used 500 cm (360 g) 2.4-4.5 mm bead 3 Å molecular sieves (water adsorption capacity 21%) as dehydrant. Test 4 used 500 cm (360 g) 1.7-2.4 mm
bead 3 Å molecular sieves (water adsorption capacity 21%) as dehydrant. c Test 5 used 500 cm (395 g) 2-5 mm granulated self-indicating silica gel (water adsorption
capacity 28.5%) as dehydrant.
3
ethylbenzene was added to the glass column, and about 250
cm of liquid was removed at the end (i.e. 70 cm of liquid
improvement of the total conversion was due to the dehydra-
tion process.
3
3
was adsorbed by the dehydrants). For subsequent 24-h runs,
In tests 3 and 4 there is only a marginal difference in
product yield with two different sizes of dehydrant, with the
smaller-size molecular sieves giving slightly higher conver-
sion. This may be due to faster water adsorption obtained
with larger solid/liquid surface area of the smaller-size
dehydrant.
The use of 2-5 mm of granulated self-indicating silica
gel as dehydrant produced results similar to those obtained
with molecular sieves (test 5 in Table 1). The conversion to
acetophenone and benzoic acid were 53.2 and 7.4% respec-
tively.
3
250 cm of ethylbenzene was added to the glass column at
the start, and approximately the same volume of liquid was
removed.
In the following discussion the conversion is calculated
as (P
weight and total liquid weight in the reactor respectively,
is weight of the free flow liquid in the dehydration unit,
and P is the product weight in the free flow liquid. It should
1
+ P
2
)/(W
1
2 1 1
+ W ), where P and W are the product
W
2
2
be pointed out that there are products adsorbed in the
dehydrants, which are not included in the calculation as direct
analysis of the amount was not possible.
4.2. Dehydration and the Induction Period. Previous
1
work on the catalytic oxidation of ethylbenzene with the
4
. Results and Discussions
.1. Dehydration of the Condensed Organic Phase.
Sulphuric acid (H SO ) (98%) was used in preliminary test
dichromate/alumina catalyst EPAD has shown that there is
a long induction period, of some 2 to 3 h, before significant
reaction proceeds. This induction period is believed to be
associated with initial hydration of catalyst during storage,
that is, the hydrophilic alumina support adsorbs moisture
from air during storage. Due to the porous structure and
hydrophilicity of alumina, the unfavourable capillary action
makes it difficult for the adsorbed water to escape and for
the reactant to reach the Cr cation sites inside the meso-
pores, during the initial stages of reaction.
4
2
4
because it does not chemically react with ethylbenzene at
lower temperatures, it can be easily separated from the
aromatics due to the density difference, and it is an effective
dehydrant. In this test, only a small quantity of acetophenone
was produced in the first hour, with no further subsequent
production detected. In addition, white polymer gel floc-
culation and flashing metal crystals were observed in the
reactor. The low yields of acetophenone are believed to be
due to dissolution of sulphuric acid into the ethylbenzene
reaction mixture which quickly destroyed the catalyst at 130
2
Previous work has also shown that a shorter induction
period and a higher conversion rate were obtained with the
chromium/silica catalyst CHRISS than with the alumina-
based catalyst. One reason for this may be due to the larger
pore size (100 Å of the silica supports compared with 20 Å
of the alumina supports), which leads to a lower capillary
rise and easier removal of water from the meso-pores.
To eliminate the induction period the catalyst was dried
at 110 °C overnight before use in the oxidation process.
Figure 3 gives a typical example of the oxidation of
ethylbenzene with the predried chromium/silica catalyst using
the dehydration unit (silica gel, dehydration of condensed
organic phase). As can be seen there was a steady rise in
the acetophenone concentration with time, and no induction
period was observed.
°
C.
Molecular SieVes and Silica Gel as Dehydrants. Table 1
summarises one set of typical results of the oxidation of
ethylbenzene under different conditions of operation using
CHRISS catalyst. Each test was carried out for a period of
2
4 h, with fresh reactants used in each case. Test 1 was
operated without catalyst and the dehydration unit. Test 2
was operated with catalyst but without the dehydration unit.
Tests 3 and 4 were operated with catalyst and the dehydration
unit; 2.4-4.5 mm bead 3 Å molecular sieves were used for
test 3, and 1.7-2.4 mm bead 3 Å molecular sieves were
used for test 4.
With the catalyst, the conversion to acetophenone and
benzoic acid was increased from 45.7 to 55.4% (w/w). With
the catalyst and dehydration unit, the conversion to acetophe-
none and benzoic acid was further increased to 62.0% (w/
w). Table 1 shows that about 32% (test 3) to 40% (test 4)
4.3. Dehydration of the Reaction Mixture. Table 2
shows the results obtained for the direct dehydration of the
ethylbenzene reaction mixture with 3 Å molecular sieves and
2-5 mm granulated self-indicating silica gel, using the small
glass reactor.
Vol. 5, No. 3, 2001 / Organic Process Research & Development
•
207

of the background oxidation, which will decrease the rate
of the background oxidation, and hence decrease the apparent
conversion rate.
Oxidation carried out without dehydration using alumina-
and silica-supported catalysts give some support to the second
effect. Figure 4 shows that for the oxidation of ethylbenzene
with a Cr/alumina catalyst (EPAD), when the quantity of
catalyst was increased from 1.0 to 5.0 g, the induction period
increased, indicating a greater amount of catalyst support
may inhibit the oxidation. Also, using an unmodified alumina
support in place of the Cr/alumina catalyst, a conversion of
1
3.4% was obtained after 24-h of operation, demonstrating
that the alumina support (Al ) inhibits the inherent
2
O
3
background oxidation. In previous research, alumina has been
used as a stationary phase in column chromatography to
16
remove organic peroxides effectively. Thus, circulating the
reaction mixture through the dehydration unit could decrease
the intermediate organic radicals in the substrate, thus
decreasing the level of the background oxidation.
Figure 5 shows that, for the oxidation of ethylbenzene
with the Cr/silica catalyst (CHRISS), when the quantity of
catalyst was increased, from 1.0 to 5.0 g, the induction period
increased, which suggests that silica also adsorbs the
intermediate organic radicals, but not as significantly as
alumina (Figure 4). This behaviour gives some support to
the superior performance achieved with silica gel as dehy-
drant (Table 2) compared with the use of molecular sieves.
Figure 3. Physical dehydration.
The test with silica gel used the same dehydration column
for four 24-h runs. Each run used the same amount of fresh
ethylbenzene and catalyst. The conversion rates increased
with the use of dehydrant in the column and particularly after
the second run. This effect is due to adsorption of oxidation
products by the silica gel during run 1. The amounts of
product adsorbed during subsequent runs were thus reduced.
Overall, by using the silica gel as dehydrant the conver-
sion of ethylbenzene was increased by approximately 6%,
compared with no dehydrant.
Table 2 also shows the results obtained with molecular
sieves as dehydrant. The average conversion of ethylbenzene
of two runs was 37.5%, compared with 41.3% with no
dehydrant. The result differed from that obtained when the
molecular sieves were used for dehydration of condensed
organic phase, where positive improvement of the conversion
rate was observed. The difference in behaviour is believed,
in part, to be due to the fact that in dehydrating the condensed
organic phase, only the reactant ethylbenzene goes through
the dehydration unit, whereas in dehydrating the reaction
mixture, the whole reaction liquid goes through the dehydra-
tion unit. In addition, compared with silica gel, the difference
in performance using molecular sieves in both dehydration
systems may be due to the outer surface of the molecular
sieves or the binder material chemically reacting with, or
physically adsorbing the intermediate radicals of the back-
ground oxidation.
4.5. Effect of Circulation Rate of the Reaction Mixture.
One advantage of direct dehydration of the reaction mixture
is that the flow rate of the organic phase can be controlled.
Two sets of tests were carried out, with identical experimental
conditions, to assess the effect of the organic phase circula-
tion rates. In each set, the same silica gel dehydration column
was used for three 24-h runs, and each run used the same
amount of fresh ethylbenzene and catalyst. Table 3 shows
the results of the tests where a higher circulation rate gave
a slightly lower average conversion rate, although there is
an indication that, at least in the final test, a higher flow
rate may be beneficial.
Overall there is only a small influence of flow rate on
the conversion of ethylbenzene, with an indication of an
improvement in performance, at higher flow rate, due to a
possible improvement in mass transfer of the dehydration
process.
Increasing the circulation rate, during dehydration of the
reaction mixture, could improve the mass transfer condition
in the dehydration unit, which could, on one hand, increase
the dehydration rate, and hence the apparent conversion rate,
and on the other hand, increase adsorption of the intermediate
organic radicals, which could decrease the rate of the
background oxidation.
4.4. Effect of the Background Oxidation of Ethylben-
zene. For the aerial oxidation of ethylbenzene, without
catalyst and the dehydration unit, a conversion of 45.7% was
obtained after 24 h of operation (test 1 in Table 1). This
background oxidation might be from small quantities of
1
organic peroxides, present in ethylbenzene as impurities.
4
.6. Reaction with a Continuous Temperature In-
With catalyst, the apparent total conversion is therefore the
sum obtained via two different reaction mechanisms: back-
ground oxidation and catalytic oxidation.
Dehydration of the reaction mixture may lead to two
opposite effects: adsorption of water from the mixture, which
assists catalyst performance, and the dehydrants may chemi-
cally react with, or physically adsorb the intermediate radicals
crease. For the oxidation processes with both EPAD and
CHRISS, the reaction rates were quite low, typically, at 130
°
2
C, approximately 50% of ethylbenzene was converted in
4 h. Thus, it might be possible to exploit the difference in
(16) Eggersgluss, W. Organische Peroxyde; Monographieran zu Angewandte
Chemie, No. 61; Verlag Chemie: Weinheim, 1951.
208
•
Vol. 5, No. 3, 2001 / Organic Process Research & Development

Table 2. Aerial oxidation of ethylbenzene with direct dehydration of the whole reaction mixture (24 h run)
conversion to acetophenone
(%)
conversion to benzoic acid
(%)
conversion of acetophone and benzoic acid
(%)
dehydrant
molecular sievesb
no dehydrant
testa
run 1
run 2
run 1
run 1
run 2
run 3
33.5
33.9
37.5
38.4
42.1
42.2
3.6
4.0
3.7
4.6
4.7
4.9
37.1
37.9
41.3
43.0
46.7
47.1
c
silica gel
a
b
3
Fresh CHRISS catalyst and ethylbenzene were used for each 24-h run. The same 500 cm (360 g) 1.7-2.4 mm bead 3 Å molecular sieves (water adsorption
capacity 21%) was used as dehydrant for two 24-h runs. c The same 500 cm (395 g) 2-5 mm granulated self-indicating silica gel (water adsorption capacity 28.5%)
was used as dehydrant for four 24-h runs.
3
Table 3. Effect of the circulation rate on the oxidation of
ethylbenzene with dehydration of the reaction mixture
flow rate
2.6 mL/min
7.5 mL/min
run 1
run 2
run 3
average
43.0%
46.7%
46.6%
45.4%
36.3%
45.5%
49.3%
43.7%
Figure 4. Oxidation of ethylbenzene with Cr/alumina catalyst.
Figure 6. Boiling point of reaction mixture.
continuously increase the reaction temperature during the
oxidation. Ideally, the reaction temperature would be con-
trolled some 5-6 °C below the boiling points during the
oxidation process (dashed line in Figure 6).
The experimental procedure adopted was to use temper-
ature step increases (0-3 h at 130 °C, 3-6 h at 135 °C,
6
-9 h at 140 °C, and 9-24 h at 145 °C) for the oxidation
of ethylbenzene with Cr/Silica catalyst, but without the
dehydration unit. Figure 7 shows the experimental results,
where the 24 h conversion obtained was 68.5%, which
compared favourably with the conversion of 55.4% with a
constant temperature at 130 °C.
Figure 5. Oxidation of ethylbenzene with Cr/silica catalyst.
The experiment with temperature steps was repeated using
the dehydration unit (silica gel). The 24 h conversion
obtained was 68.2% which whencompared to the conversion
of 68.5% without the dehydration unit, showed that dehydra-
tion did not improve the reaction performance when the
reaction was carried out at increased temperature. One
explanation for this is that, at higher reaction temperature
the boiling points of ethylbenzene, acetophenone, and
benzoic acid (at atmospheric pressure 136, 202, and 249 °C,
respectively) and operate at higher temperatures. Thus, as
the boiling point of the reaction mixture increases with
conversion of ethylbenzene, for example, from 136 to 168
°
C at a conversion of 68.5% (Figure 6), it is possible to
Vol. 5, No. 3, 2001 / Organic Process Research & Development
•
209

dehydrants for the process dehydrating the condensed organic
phase, and (3) silica gel is a suitable dehydrant for the process
of dehydrating the reaction mixture. Overall, selection of a
suitable dehydrant is the key in the dehydration process.
By operating the reactor with an increase in reaction
temperature, better performance is achieved than under
isothermal conditions (at atmospheric pressure). With silica-
supported catalysts, the dehydration process cannot improve
the reaction performance under conditions of increased
reaction temperature.
The advantages of dehydrating the condensed organic
phase, as opposed to dehydrating the reaction mixture, are:
(
1) no filter is needed as the solid catalyst particles remain
in the reactor, and
2) no pump is needed as, with a suitable hydraulic
(
arrangement, the condensed organic phase will flow back
to the reactor due to gravity.
The main shortcoming in dehydrating the condensed
organic phase is that the flow rate of the organic phase is
controlled by the rate of condensation.
Figure 7. Oxidation of ethylbenzene with increasing temper-
ature steps. (2) Benzoic acid, (9) acetophenone, ([) acetophe-
none and benzoic acid.
(above 140 °C), the mass transfer rate of water from Cr cation
sites is higher than the rate of production of water by reaction,
and therefore water mass transfer is no longer a control
barrier of the process. Previous research has reported that
the loss of water from silica gel may begin at 50 °C, reaching
a maximum rate of desorption at about 140 °C. 12
Acknowledgment
The Engineering and Physical Sciences Research Council
(EPSRC) provided financial support. This research was
undertaken with the support of the Chemical Engineering/
Chemistry collaboration between Newcastle and York Uni-
versities. J.H.C. thanks the RAEng-EPSRC for a Clean
Technology Fellowship. Dr. Keith Martin at Contract
Chemicals is thanked for helpful discussions.
5. Conclusions
For the catalytic oxidation of ethylbenzene to acetophe-
none, investigation of the influence of dehydration has shown
that: (1) 98% sulphuric acid is not a suitable dehydrant for
the proposed dehydration processes due to destruction of the
catalyst, (2) the molecular sieves and silica gel are suitable
Received for review June 8, 2000.
OP0000632
210
•
Vol. 5, No. 3, 2001 / Organic Process Research & Development