Continuous Heterogeneously Catalyzed Oxidation of Benzyl Alcohol in a Ceramic Membrane Packed-Bed Reactor

Article abstract of DOI:10.1021/acs.oprd.5b00220

A ceramic membrane reactor was investigated for the continuous catalytic oxidation of benzyl alcohol with oxygen. The reactor had a concentric configuration. An inner tube created an annulus for the catalyst packed-bed (0.9 wt % Au-Pd/TiO₂, par

Full text of DOI:10.1021/acs.oprd.5b00220

Article
pubs.acs.org/OPRD
Continuous Heterogeneously Catalyzed Oxidation of Benzyl Alcohol
in a Ceramic Membrane Packed-Bed Reactor
Achilleas Constantinou,†,¶ Gaowei Wu,† Albert Corredera,† Peter Ellis,‡ Donald Bethell,§
Graham J. Hutchings,∥ Simon Kuhn,⊥ and Asterios Gavriilidis*,†
†Department of Chemical Engineering, University College London, Torrington Place, London, WC1E 7JE, United Kingdom
¶Division of Chemical and Petroleum Engineering, School of Engineering, London South Bank University, London, SE1 0AA, United
Kingdom
‡Johnson Matthey, Blounts Court Road, Reading, RG4 9NH, United Kingdom
§Department of Chemistry, University of Liverpool, Crown Street, Liverpool L69 7ZD, United Kingdom
∥School of Chemistry, Cardiff University, Main Building, Park Place, Cardiff, CF10 3AT, United Kingdom
⊥Department of Chemical Engineering, KU Leuven, W. de Croylaan 46, 3001 Leuven, Belgium
S
* Supporting Information
ABSTRACT: A ceramic membrane reactor was investigated for the continuous catalytic oxidation of benzyl alcohol with
oxygen. The reactor had a concentric configuration. An inner tube created an annulus for the catalyst packed-bed (0.9 wt % Au−
Pd/TiO2, particle size 90−125 μm) through which the liquid phase (benzyl alcohol, neat or dissolved in o-xylene) flowed. This
was followed by the tubular ceramic membrane, which consisted of layers of alumina and a zirconia top layer with a nominal
average pore size of 50 nm. The role of the membrane was to provide an interface for gas and liquid to come in contact. Pure
oxygen was fed to the opposite side of the membrane in the outer shell of the reactor. Temperature affected conversion but not
selectivity, possibly because of insufficient supply of oxygen. However, increasing catalyst contact time or decreasing benzyl
alcohol concentration improved selectivity and conversion, indicating that a key parameter was the balance between oxygen
supply by the membrane vs oxygen demand by the reaction. By adjusting the operating parameters, reaction performance
improved. Selectivity to benzaldehyde 88% and conversion of benzyl alcohol 75% were obtained at 3.2 bara of gas pressure,
24444 gcat·s/galcohol catalyst contact time, 0.5 M benzyl alcohol concentration, and temperature of 120 °C. This performance was
comparable to simulated trickle bed operation, where oxygen and substrate were premixed before entering the catalyst packed
bed. The membrane reactor offers safer operation, since flammable oxygen/organic mixtures formed in the trickle bed are
avoided.
which can affect product selectivity. Various studies have been
performed for multiphase reactions with catalytic membranes
(i.e., where the catalyst is located in the membrane itself).
Vospernik et al.11 used ceramic membranes, impregnated with
Pt for liquid phase oxidation of aqueous formic acid solutions.
The ceramic membranes consisted of multiple layers of Al2O3/
TiO2 and a thin layer of ZrO2, located on the inner surface of
membrane, having a thickness of ∼6 μm and a nominal average
pore size of 20 nm. Experiments and modeling showed that
only a small fraction of the deposited Pt takes part in the
reaction. Furthermore, the productivity of the reactor was
affected by the dissolved oxygen in the reaction area and the
molar ratio of reactants. Aran et al.12 examined the performance
of ceramic membrane reactors varying characteristic length,
catalyst support thickness, wetting properties, and operational
1. INTRODUCTION
Aldehydes, produced by catalytic oxidation of alcohols, are
valuable precursors for the production of pharmaceutical
compounds, dyes and fragrances.1,2 Traditionally high value
chemical products (fine chemicals, pharmaceuticals) are
manufactured in batch units, generating enormous waste and
the direct contact of the phases raises safety concerns for
hazardous reactions such as oxidations.3 Continuous processing
can facilitate safe manipulation of potentially hazardous
reagents by minimizing local inventory of hazardous materials
and allowing efficient heat transfer in highly exothermic
processes.
Membrane contactors allow two phases to come into direct
contact with each other to achieve efficient interfacial mass
transfer, without the need to disperse one phase into the other.
The concept of using membranes covers many industrial
processes such as extraction, stripping, and absorption.4−6 They
can also be employed for catalytic reactions, since the
membrane can either be catalytic itself or can be used to
contain a catalyst particle packed-bed.7,8 Ceramic membranes
have great potential for gas−liquid−solid reactions in
continuous flow due to their chemical resistance and gas
permeability.9,10 They allow controlled dosing of gas reactants,
−
parameters for the hydrogenation of nitrite ions (NO2 ) in
water. The membrane consisted of α-Al2O3 and γ-Al2O3 as a
support for a Pd catalyst. The reaction rate per active catalyst
surface area decreased by increasing the γ-Al2O3 thickness
indicating internal mass transfer limitations, whereas decreasing
Received: July 5, 2015
© XXXX American Chemical Society
A
DOI: 10.1021/acs.oprd.5b00220
Org. Process Res. Dev. XXXX, XXX, XXX−XXX

Organic Process Research & Development
Article
2PhCH2OH → PhCHO + PhCH3 + H2O
the inner tube diameter and using slug flow in the reactor
channel improved the external mass transfer. Pashkova et al.13
investigated the synthesis of hydrogen peroxide directly from
hydrogen and oxygen in a gas−liquid ceramic membrane
reactor. Different membrane materials were used such as Al2O3,
TiO2, and carbon-coated Al2O3 with a Pd catalyst deposited
into the finest porous layer on the inner side of the membranes.
It was found that the diffusive transport of the reactants to the
catalytically active zone, located on the inner walls of the
membrane channel was crucial. Furthermore, important process
parameters such as solvent type, system pressure, and flow
regime were evaluated showing that ceramic membrane
reactors might be suited for H2O2 production. Iojoiu et al.14
studied wet air oxidation of industrial effluents in a catalytic
membrane reactor. Two different ceramic membranes were
used; the first one consisted of pure titania support with two
intermediate layers of ceria-doped-zirconia-covered titania, and
the second membrane consisted of α-alumina/titania as support
with an intermediate layer of zirconia. The catalyst (Pt) was
deposited on the inside fine layer of the membrane. The
catalytic membrane reactor and the operation conditions were
successfully scaled-up from lab scale to a pilot unit. The
approach of impregnating the catalyst on the inner side of the
ceramic membranes is beneficial in terms of oxygen mass
transfer and gas−liquid contacting, although it requires defined
gas-permeability properties of the ceramic membranes, which
might require pore modification.15 A major drawback of the
impregnated membranes is that in case of deactivation the
catalyst is not exchangeable. An alternative approach is to use
the ceramic membrane to contain a catalyst packed-bed. The
advantages of this concept are the safety (since gas and liquid
can be kept separated), and the flexibility of exchanging the
catalyst in case of deactivation. Packed-bed membrane reactors
have been used for gas phase catalytic reactions of hydro-
carbons16,17 but not for three-phase catalytic reactions.
Aerobic oxidation of benzyl alcohol using molecular oxygen
is one of the most common catalytic reaction systems studied
using a variety of metal based catalysts.18−21 Hutchings’ group
discovered that a bimetallic catalyst of Au and Pd led to a 25-
fold enhancement in turnover frequencies compared to
monometallic supported Au and Pd22 for oxidations of alcohols
(including benzyl alcohol). A few studies investigated benzyl
alcohol oxidation under flow conditions using gold-based
catalysts. Cao et al.23 studied the catalytic oxidation of benzyl
alcohol using Au−Pd/TiO2 catalyst in a silicon-glass micro-
packed-bed reactor. It was observed that the conversion of
benzyl alcohol was comparable with the conversion found in a
conventional stirred batch reactor. Wang et al.18 developed a
gold-immobilized microchannel flow reactor through cross-
linking of copolymer for oxidations of different alcohols. No
leaching of gold was observed and the gold-immobilized
capillary column could be used up to 4 days without loss of
activity. Kaizuka et al.24 used Au−Pt and Au−Pd bimetallic
nanoclusters for catalytic aerobic oxidation of alcohols in
multiphase flow systems. It was found that the flow systems
were superior to the batch systems in terms of both yield and
selectivity. Two main pathways are suggested for the solvent-
free synthesis of benzaldehyde from benzyl alcohol on Au−Pd/
TiO2 catalysts,23,25,26 which are the oxidation reaction (eq 1)
and the disproportionation reaction (eq 2).
(2)
Aerobic catalytic oxidations are avoided in industry due to
safety concerns. Packed-bed membrane reactors for such
applications have rarely been studied. In our previous work27
we investigated the catalytic oxidation of benzyl alcohol using
Au−Pd/TiO2 catalyst in a Teflon AF-2400 tube-in-tube
configuration. This design allowed continuous penetration of
oxygen through the gas-permeable tube during the reaction,
and as a result conversion was significantly improved compared
to a reactor operating with an oxygen presaturated feed. In this
work, we study a packed-bed porous ceramic membrane reactor
for a similar catalytic system. This configuration allows
continuous addition of the oxidant along the length of the
reactor safely, since the gas does not come in direct contact
with the organic mixture in the packed-bed area. The gas/liquid
interface is stabilized at the nanoporous top membrane layer,
and from there oxygen diffuses into the packed-bed liquid
phase. Ceramic membrane reactors offer easier scalability, as
multichannel ceramic membranes are commercially avail-
able.13,14
2. EXPERIMENTAL SECTION
2.1. Catalyst Preparation. Au−Pd/TiO2 catalyst was
prepared by impregnation, similar to the procedure described
in.28 Titania (Evonik P25, 49.5 g) was suspended in 200 mL
demineralized water by stirring. Tetrachloroauric acid solution
(0.061g of HAuCl4 solution (41.22 wt % Au), 0.025g Au) and
palladium nitrate solution (3.16g of Pd(NO3)2 solution
(15.05% Pd), 0.475g Pd) were added to the stirred suspension.
The suspension was spray dried at 220 °C (Buchi B-290) to
give a pale yellow powder which was calcined in static air at 400
°C for 1 h. The product was analyzed by ICP-AES and found to
contain 0.85 wt % Pd and 0.05 wt % Au, with metal particle size
1−2 nm as observed by TEM. The catalyst powder was
pelletized and then crushed to obtain the desired particle size
fraction (90−125 μm) by sieving.
2.2. Packed-Bed Membrane Reactor Setup. The reactor
consisted of several concentric sections (an inner tube, a
ceramic membrane tube, and an outer stainless steel tube
housing the whole assembly) (see Figure 1). The inner tube
(outer diameter O.D. 4 mm, inner diameter I.D. 2.96 mm) and
ceramic membrane created an annulus for the catalyst packed-
bed, through which the liquid phase flowed. The ceramic
membrane tube of length 250 mm, O.D. 10 mm and I.D. 7 mm
(Pall, Europe) consisted of layers of alumina (support layer
thickness 1500 μm, sublayer thickness 35 μm) and a zirconia
top layer with nominal average pore size 50 nm and thickness 8
μm. Oxygen was fed to the opposite side of the membrane. The
housing was made out of stainless steel (Orion Alloys, UK)
with an I.D. of 12 mm. Liquid inlet and outlet tubes were
welded on the two stainless steels caps at an angle (see Figure
1b) at the two ends of the reactor. For the sealing of the
membrane against the two stainless steel caps, two O-rings
made of ethylene propylene diene monomer (EPDM) were
used. In addition, the two ends of the membranes were glazed
using enamel (Johnson Matthey, UK). Two stainless steel nuts
sealed and secured the inner tube on the two stainless steel
caps. For retaining the catalyst, a stainless steel frit was inserted
at one end of the reactor. Packing of the catalyst was achieved
by suction using a vacuum pump. The amount of catalyst used
in all experiments was 440 mg. The catalyst bed length was
approximately 2.2 cm (total length of reactor 25 cm). Glass
2PhCH2OH + O2 → 2PhCHO + 2H2O
(1)
B
DOI: 10.1021/acs.oprd.5b00220
Org. Process Res. Dev. XXXX, XXX, XXX−XXX

Organic Process Research & Development
Article
USA) for performing experiments at 3.2 bara gas pressure. To
avoid breakthrough of one phase into the other the reactor was
operated with a pressure difference between the gas and the
liquid phase PG − PL ≈ 0.1 bar. The outlet of the gas phase
passed through a glass separator, where was diluted with pure
N2 to reduce the concentration of O2 in order to avoid
formation of flammable mixtures. The liquid outlet passed first
through a custom-build bubble detector (based on detecting
changes in the transmitted light intensity in the presence of a
gas phase) and then to a liquid collector. The liquid collector
was also purged with N2 to degas the solution from O2 and
reduce risk. The collected sample was quantitatively analyzed
by a gas chromatograph (Agilent 7820A, USA) fitted with a
HP-INNOWax capillary column and a flame ionization
detector.
The heating of the reactor was achieved by using three
flexible heaters (10 W/in2 each) (Omega, UK) covering the
outside surface of the reactor. The heaters were insulated with
calcium−magnesium silicate thermal wool (RS, UK) to reduce
heat losses. Two thermocouples were placed at the inlet/outlet
of the liquid phase, one in the packed-bed and one in the gas
outlet to monitor the temperatures. The inclination of the
inlet/outlet of the reactor allowed the insertion of the
thermocouple before packing the reactor with catalyst. The
temperature was measured in three different locations “a” to “c”
inside the reactor (see Figure 3) by pulling up the
thermocouple inside the area packed with glass beads. For
the investigated temperature range no temperature gradient was
observed, confirming that the reaction fluid reached the
required temperature. The catalyst location toward the end of
the reactor ensured that the reaction zone was isothermal. To
further, check reactor isothermality, a reaction experiment was
performed where the catalyst was located between “b” and “c”
in Figure 3. The temperature difference between the inlet and
the outlet of the catalyst packed-bed was just 1 °C, which was
within experimental accuracy. After the temperature tests the
thermocouple was located at the location “c” shown in Figure 3.
Measures to ensure safe operation of the setup are presented in
the Supporting Information.
Experiments were started by setting the liquid flow rate, the
reactor temperature, gas flow rate and pressure difference
ΔPG‑L= PG − PL to desired values. After the system stabilized,
collection of samples started. The gas flow rate was fixed at 30
mL/min, while the liquid flow rate was varied between 0.02 and
0.08 mL/min, and the reaction temperature was varied between
80 and 120 °C.
Figure 1. Schematics of membrane reactor: (a) concentric sections of
the reactor (not to scale) with an SEM picture of the ceramic
membrane, (b) components of the reactor assembly.
beads (100−200 μm, ∼0.1 g) were first sucked into the reactor
creating a bed ∼0.5 cm long, followed by a layer of catalyst,
after which the reactor was filled up to the top with glass beads.
A schematic of the experimental setup used for the catalytic
oxidation experiments is shown in Figure 2. A HPLC pump
(Vapourtec R4, UK) was used to direct the liquid benzyl
alcohol (99.0%, Sigma-Aldrich, US) in the catalyst packed-bed
area, while the gas (pure O2) was controlled by a mass flow
controller (Brooks 4800, Netherlands) and flowed on the
opposite side of the membrane. The flow of liquid and gas
reactants was cocurrent. Four pressure sensors (Omega, UK)
were placed at the inlets/outlets of the gas/liquid phases to
monitor gas/liquid pressures (all pressures reported are
absolute). The differential pressure between the two phases
was controlled by metering valves (Upchurch, USA) at the
outlet of the liquid and gas phase after the separator. Metering
valves were replaced with back pressure regulators (Zaiput,
Benzyl alcohol conversion (X) and selectivity (S) of each
product were calculated based on the following equations:
Calcohol,in − Calcohol,out
X =
× 100%
Calcohol,in
(3)
where Calcohol,in and Calcohol,out were the concentration of benzyl
alcohol at the inlet and outlet, respectively,
Ciνi
Calcohol,inXtotal
Si =
× 100%
(4)
where νi was the number of moles of benzyl alcohol consumed
for the production of 1 mol of product i. Catalyst contact time
was used to characterize the reaction time of benzyl alcohol and
was defined as
C
DOI: 10.1021/acs.oprd.5b00220
Org. Process Res. Dev. XXXX, XXX, XXX−XXX

Organic Process Research & Development
Article
Figure 2. Schematic diagram of the experimental setup.
3. RESULTS AND DISCUSSION
3.1. Effect of Temperature. Figure 4 shows the effect of
temperature on benzyl alcohol conversion and selectivities to
Figure 4. Effect of temperature on benzyl alcohol conversion and
selectivities to toluene and benzaldehyde. Reaction conditions: gas
pressure 1.1 bara, liquid flow rate 0.04 mL/min, gas flow rate 30 mL/
min, and inlet pure BnOH.
Figure 3. Picture of the ceramic membrane packed-bed reactor. Lines
indicate the locations where temperature was measured with the
thermocouple inside the reactor.
mass of catalyst (gcat)
benzaldehyde and toluene when the temperature was changed
from 60 to 120 °C. Benzyl alcohol conversion increased up to
19.5% by increasing the temperature to 120 °C. Despite the
temperature change, selectivity to benzaldehyde changed only
slightly and varied between 63 and 64%. The same trend was
observed for toluene selectivity (variation between 33 and
35%). This is in contrast to the results of Cao et al.,23 where the
catalytic oxidation of benzyl alcohol using Au−Pd/TiO2
(particle size 50−63 μm, based on sol immobilization
preparation method) in a silicon-glass micropacked-bed reactor
resulted in increased conversion of benzyl alcohol and
CCT =
mass flow rate of benzyl alcohol (galcohol/s)
(5)
For each experiment, three samples were collected, and the
results were averaged. The reproducibility of the experiments
and catalyst stability were checked by a standard run (100 °C,
catalyst space time 635 galcohol·s/gcat, liquid flow rate 0.04 mL/
min, gas flow rate 30 mL/min, gas pressure 1.1 bara), and the
relative difference was less than 5%.
D
DOI: 10.1021/acs.oprd.5b00220
Org. Process Res. Dev. XXXX, XXX, XXX−XXX

Organic Process Research & Development
Article
a
Table 1. Comparison of Benzyl Alcohol Conversion and Selectivities between Membrane and Trickle-Bed Operation
mode of
operation
Sbenzaldehyde
(%)
Stoluene
(%)
Sbenzene
(%)
Sdibenzyl ether
Sbenzoic acid
(%)
Sbenzyl benzoate
X (%)
(%)
(%)
oxygen consumption rate (mol/gcat·min)
membrane
trickle bed
11
29
64
77
33.6
17
0.2
0.9
2
0.3
1.3
0.94
1.9
1.5 × 10−5
7.6 × 10−5
0.85
a
Reaction conditions: reactant pure benzyl alcohol, liquid flow rate 0.04 mL/min, temperature 100 °C, catalyst contact time 635 gcat·s/galcohol
.
Membrane operation: gas flow rate 30 mL/min, gas pressure 1.1 bara. Trickle bed operation: gas flow rate 20 mL/min, gas pressure 1 bara.
decreased selectivity to benzaldehyde (from 85 to 61%) when
the temperature was increased from 80 to 120 °C. In the
micropacked-bed reactor the gas and liquid phases were mixed
at the inlet (gas to liquid flow rate ratio = 100), whereas in this
work a gas/liquid interface was stabilized at the nanoporous top
membrane layer, where the liquid was saturated with oxygen
and from there oxygen diffused in the packed-bed liquid phase.
No bubbles were created in the liquid that flowed inside the
packed-bed. The different trend with respect to benzaldehyde
selectivity with temperature might be related to insufficient
supply of oxygen in the catalyst packed-bed area, but also to the
Figure 5. Effect of catalyst contact time on benzyl alcohol conversion
and selectivities to toluene and benzaldehyde. Reaction conditions: gas
pressure 1.1 bara, liquid flow rate 0.02−0.08 mL/min, gas flow rate 30
mL/min, inlet pure BnOH, temperature 100 °C.
different catalyst preparation method.
To further explore the importance of oxygen on conversion
and selectivities, an experiment was performed in which the gas
and liquid phases were mixed before entering the reactor and
the gas inlet/outlet of the reactor were closed, thus simulating a
trickle bed operation. The gas flow rate was 20 mL/min with
liquid flow rate 0.04 mL/min resulting in a gas to liquid flow
rate ratio of 500. Table 1 compares the results of the membrane
and trickle bed operation. It is observed that for identical
temperature, catalyst amount, and catalyst contact time, the
conversion of benzyl alcohol (X) and selectivity to
benzaldehyde (Sbenzaldehyde) are higher and selectivity to toluene
(Stoluene) is lower for the trickle bed operation. For the latter,
more oxygen is present in the catalyst packed-bed, and as a
result the oxidation reaction (see eq 1) is enhanced. During
membrane operation insufficient supply of oxygen to the
catalyst packed-bed area results in decreased selectivity to
benzaldehyde and benzyl alcohol conversion. Table 1 also lists
selectivities to other products and for the membrane condition
their values were always lower than 3%. These findings seem to
corroborate that the lack of sufficient oxygen is the reason for
the low and constant selectivity to benzaldehyde with
temperature under membrane operation. Under the operating
conditions above, a trickle bed reactor is beneficial for
increasing both conversion and selectivity.
oxygen by the oxidation reaction (see eq 1). As a result,
selectivity to benzaldehyde increased with catalyst contact time.
The same observation for the selectivities to benzaldehyde and
toluene was reported in our previous work27 where the
oxidation of benzyl alcohol was performed in a tube-in-tube
configuration.
3.3. Effect of Benzyl Alcohol Dilution. To increase the
oxygen supply in relation to oxygen demand by the reaction
and thus improve conversion of benzyl alcohol and
benzaldehyde selectivity, benzyl alcohol was diluted 5−20
times using o-xylene as a solvent. Figure 6 shows the results
3.2. Effect of Catalyst Contact Time. The effect of
catalyst contact time was investigated by varying the liquid flow
rate from 0.02 to 0.08 mL/min and keeping the same amount
of catalyst (440 mg) under membrane operation. By increasing
the space time from 317 to 1269 gcat·s/galcohol the benzyl alcohol
conversion increased up to 19.2%, while the selectivity to
benzaldehyde increased from 59% to 69% and the selectivity to
toluene decreased from 39% to 28% (Figure 5). Cao et al.23
studied benzyl alcohol oxidation in a silicon-glass micropacked-
bed and a batch reactor. Oxygen remained in excess in these
two reactors, which may explain the reason why the selectivities
to benzaldehyde and toluene remained almost the same at
different catalyst contact time. In our membrane operation the
differences in selectivities might be due to oxygen deficiency in
the catalyst packed-bed area. By decreasing the liquid flow rate
and keeping the catalyst bed length constant the catalyst
contact time was increased and hence, oxygen had more time to
permeate and react. This allowed increased consumption of
Figure 6. Effect of dilution of benzyl alcohol on benzyl alcohol
conversion and selectivities to toluene and benzaldehyde. Reaction
conditions: gas pressure 1.1 bara, liquid flow rate 0.08 mL/min, gas
flow rate 30 mL/min, temperature 100 °C.
when the concentration of benzyl alcohol was varied from 0.5
M to 9.6 M. Diluting benzyl alcohol 20 times resulted to
conversion increase to 57%. Correspondingly selectivity to
benzaldehyde increased to 72%, while selectivity to toluene
dropped to 27%. The reason for this behavior is that the oxygen
supply remained the same (as in previous experiments), but the
benzyl alcohol concentration was decreased. As a result, oxygen
demand was less, and the oxygen supply through the membrane
resulted in a higher dissolved oxygen concentration in the
catalyst packed-bed area.
E
DOI: 10.1021/acs.oprd.5b00220
Org. Process Res. Dev. XXXX, XXX, XXX−XXX

Organic Process Research & Development
Article
a
Table 2. Operating Conditions for Optimizing Reaction Performance
mode of operation
X (%)
Sbenzaldehyde (%)
Stoluene (%)
oxygen consumption rate (mol/gcat·min)
membrane
trickle bed
75
80
88
85
9
6
6.7 × 10−6
7.2 × 10−6
a
Reaction conditions: benzyl alcohol concentration 0.5 M, liquid flow rate 0.02 mL/min, temperature 120 °C, gas pressure 3.1−3.2 bara, catalyst
contact time 24440 gcat·s/galcohol. Membrane operation: gas flow rate 30 mL/min. Trickle bed operation: gas flow rate 20 mL/min.
3.4. Optimized Conditions for Higher Benzaldehyde
Selectivity. In a next step, operating conditions were adjusted
with the aim to further increase the selectivity to benzaldehyde
and the conversion of benzyl alcohol. Benzyl alcohol was
diluted with o-xylene (BnOH 0.5 M), gas pressure was
increased to 3.2 bara (upper limit for the setup), and catalyst
contact time was increased to 24444 (gcat·s/galcohol) and
temperature to 120 °C. Dilution of benzyl alcohol resulted in
lower oxygen demand by the reaction, increasing the gas
pressure increased dissolved oxygen concentration, higher
catalyst contact time gave oxygen more time to permeate and
react, and higher temperature resulted to a higher reaction rate.
Table 2 shows that under membrane operation the conversion
of benzyl alcohol increased to 75% (previous maximum 57%,
see Figure 6) and most importantly the selectivity to
benzaldehyde increased to 88% (previous maximum 72%, see
Figure 6). This is a significant improvement in selectivity and
conversion compared to previous results (sections 3.1−3.3).
The selectivity improvement is attributed to higher oxygen
concentration not only in the bulk liquid in the packed-bed but
also inside the catalytic particles. The possibility of solvent
influencing the selectivity cannot be excluded. As a comparison,
experiments were performed under the same reaction
conditions in trickle bed operation, which resulted in
conversion of benzyl alcohol of 80% and selectivity to
benzaldehyde of 85% (Table 2). This conversion and selectivity
were similar to the values obtained under membrane operation,
which indicates sufficient oxygen supply for both membrane
and trickle bed operation.
catalyzed aerobic oxidation of benzyl alcohol. This reactor
configuration allows continuous addition of the oxidant along
the length of the reactor, and since the gas does not come in
direct contact with the organic mixture in the packed-bed area,
it enables safer, continuous oxidation of benzyl alcohol. In
addition, it allows expedient catalyst replacement in case of
deactivation which is not possible for wall-coated catalytic
membrane reactors. However, the rate of oxidant supply may
not be sufficient for fast reactions. Experiments showed that
benzyl alcohol conversion increased with increasing temper-
ature, catalyst contact time and dilution with o-xylene. The
selectivity to benzaldehyde increased with catalyst contact time,
since oxygen had more time to penetrate and react in the
catalyst packed-bed area. The same trend was observed by
diluting benzyl alcohol with o-xylene, since less oxygen was
required by the reaction. Temperature changes had no
significant effect on the selectivity to benzaldehyde possibly
due to deficiency of oxygen in the catalyst packed-bed area
under membrane operation. Selectivity and conversion were
improved by increasing gas pressure, diluting benzyl alcohol,
increasing catalyst contact time and increasing temperature.
This work demonstrates that, if oxygen supply and demand are
not balanced, operating and design parameters can be altered to
overcome the deficiency of oxygen in the catalyst bed area.
ASSOCIATED CONTENT
* Supporting Information
■
S
The Supporting Information is available free of charge on the
ACS Publications website at DOI: 10.1021/acs.oprd.5b00220.
The catalytic system studied in this work is relatively fast. For
undiluted reactant and atmospheric oxygen pressure, average
oxygen consumption rate is 7.6 × 10−5 mol/gcat min (see Table
1) for trickle bed operation which provides the best
performance. For the same conditions, the membrane reactor
shows reaction rates of 1.5 × 10−5 mol/gcat min, indicating that
oxygen supply is not sufficient. However, under the conditions
for optimized selectivity (see Table 2) the oxygen reaction rates
are closer: 7.2 × 10−6 mol/gcat min for trickle bed operation and
6.7 × 10−6 mol/gcat min for membrane operation. Hence, for
benzyl alcohol oxidation on the catalyst and the membrane
reactor used in this work, reaction rates need to be reduced for
improving selectivity so that an acceptable balance between
oxygen supply and demand is achieved. If this is acceptable, it
would also depend on consideration of the safety advantages
that the membrane reactor offers and the downstream
separation costs. For slower catalytic systems, with oxygen
reaction rates below 7 × 10−6 mol/gcat min, the membrane
reactor would be satisfactory, without having to slow down the
reaction. For such systems, filling the reactor with catalyst or
recirculating the liquid reactant can be utilized to increase
reactant conversion.
A detailed description of safety precautions which
provided several layers of safety in the experimental
setup (PDF)
AUTHOR INFORMATION
Corresponding Author
*E-mail: a.gavriilidis@ucl.ac.uk.
■
Notes
The authors declare no competing financial interest.
ACKNOWLEDGMENTS
■
Financial support by EPSRC (grant EP/L003279/1) is
gratefully acknowledged. We thank Simon Barrass for help
with the experimental setup, John Langdon for the fabrication
of the reactor, and Winson Kuo for catalyst characterisation.
REFERENCES
■
2004.
(1) Backval, J. E. Modern Oxidation Methods; Wiley-VCH: Weinheim,
̈
(2) Hudlicky, M. Oxidation in Organic Chemistry; American Chemical
Society: Washington, DC, 1990.
(3) Mannel, D. S.; Stahl, S. S.; Root, T. W. Org. Process Res. Dev.
4. CONCLUSIONS
2014, 18, 1503−1508.
(4) Cai, Z.-X.; Fang, Q.; Chen, H.-W.; Fang, Z.-L. Anal. Chim. Acta
2006, 556, 151−156.
A catalytic packed-bed ceramic membrane reactor was
developed and used for the continuous heterogeneously
F
DOI: 10.1021/acs.oprd.5b00220
Org. Process Res. Dev. XXXX, XXX, XXX−XXX

Organic Process Research & Development
Article
(5) Constantinou, A.; Ghiotto, F.; Lam, K. F.; Gavriilidis, A. Analyst
2014, 139, 266−272.
(6) Constantinou, A.; Barrass, S.; Gavriilidis, A. Ind. Eng. Chem. Res.
2014, 53, 9236−9242.
(7) Coronas, J.; Santamaría, J. Catal. Today 1999, 51, 377−389.
(8) Julbe, A.; Farrusseng, D.; Guizard, C. J. Membr. Sci. 2001, 181, 3−
20.
(9) Vospernik, M.; Pintar, A.; Berci
Eng. Res. Des. 2004, 82, 659−666.
̌ ̌
c, G.; Batista, J.; Levec, J. Chem.
(10) Aran, H. C.; Chinthaginjala, J. K.; Groote, R.; Roelofs, T.;
Lefferts, L.; Wessling, M.; Lammertink, R. G. H. Chem. Eng. J. 2011,
169, 239−246.
(11) Vospernik, M.; Pintar, A.; Levec, J. Chem. Eng. Process. 2006, 45,
404−414.
(12) Aran, H. C.; Klooster, H.; Jani, J. M.; Wessling, M.; Lefferts, L.;
Lammertink, R. G. H. Chem. Eng. J. 2012, 207−208, 814−821.
(13) Pashkova, A.; Dittmeyer, R.; Kaltenborn, N.; Richter, H. Chem.
Eng. J. 2010, 165, 924−933.
(14) Iojoiu, E. E.; Landrivon, E.; Raeder, H.; Torp, E. G.; Miachon,
S.; Dalmon, J.-A. Catal. Today 2006, 118, 246−252.
(15) Murru, M.; Prabowo, A.; Gavriilidis, A. J. Membr. Sci. 2005, 248,
27−36.
(16) Klose, F.; Wolff, T.; Thomas, S.; Seidel-Morgenstern, A. Appl.
Catal., A 2004, 257, 193−199.
(17) Oyama, S. T.; Zhang, X.; Lu, J.; Gu, Y.; Fujitani, T. J. Catal.
2008, 257, 1−4.
(18) Wang, N.; Matsumoto, T.; Ueno, M.; Miyamura, H.; Kobayashi,
S. Angew. Chem. 2009, 121, 4838−4840.
(19) Zotova, N.; Hellgardt, K.; Kelsall, G. H.; Jessiman, A. S.; Hii, K.
K. Green Chem. 2010, 12, 2157−2163.
(20) Kesavan, L.; Tiruvalam, R.; Rahim, M. H. A.; bin Saiman, M. I.;
Enache, D. I.; Jenkins, R. L.; Dimitratos, N.; Lopez-Sanchez, J. A.;
Taylor, S. H.; Knight, D. W.; Kiely, C. J.; Hutchings, G. J. Science 2011,
331, 195−199.
(21) Obermayer, D.; Balu, A. M.; Romero, A. A.; Goessler, W.;
Luque, R.; Kappe, C. O. Green Chem. 2013, 15, 1530−1537.
(22) Enache, D. I.; Edwards, J. K.; Landon, P.; Solsona-Espriu, B.;
Carley, A. F.; Herzing, A. A.; Watanabe, M.; Kiely, C. J.; Knight, D. W.;
Hutchings, G. J. Science 2006, 311, 362−365.
(23) Cao, E.; Sankar, M.; Firth, S.; Lam, K. F.; Bethell, D.; Knight, D.
K.; Hutchings, G. J.; McMillan, P. F.; Gavriilidis, A. Chem. Eng. J. 2011,
167, 734−743.
(24) Kaizuka, K.; Lee, K.-Y.; Miyamura, H.; Kobayashi, S. J. Flow
Chem. 2012, 2, 1−4.
(25) Sankar, M.; Nowicka, E.; Tiruvalam, R.; He, Q.; Taylor, S. H.;
Kiely, C. J.; Bethell, D.; Knight, D. W.; Hutchings, G. J. Chem. - Eur. J.
2011, 17, 6524−6532.
(26) Meenakshisundaram, S.; Nowicka, E.; Miedziak, P. J.; Brett, G.
L.; Jenkins, R. L.; Dimitratos, N.; Taylor, S. H.; Knight, D. W.; Bethell,
D.; Hutchings, G. J. Faraday Discuss. 2010, 145, 341−356.
(27) Wu, G.; Constantinou, A.; Cao, E.; Kuhn, S.; Morad, M.;
Sankar, M.; Bethell, D.; Hutchings, G. J.; Gavriilidis, A. Ind. Eng. Chem.
Res. 2015, 54, 4183−4189.
(28) Dimitratos, N.; Lopez-Sanchez, J. A.; Anthonykutty, J. M.; Brett,
G.; Carley, A. F.; Tiruvalam, R. C.; Herzing, A. A.; Kiely, C. J.; Knight,
D. W.; Hutchings, G. J. Phys. Chem. Chem. Phys. 2009, 11, 4952−4961.
G
DOI: 10.1021/acs.oprd.5b00220
Org. Process Res. Dev. XXXX, XXX, XXX−XXX