Novel synthesis of thick wall coatings of titania supported Bi poisoned Pd catalysts and application in selective hydrogenation of acetylene alcohols in capillary microreactors

Article abstract of DOI:10.1039/c4lc01066c

Catalysis in microreactors allows reactions to be performed in a very small volume, reducing the environmental problems and greatly intensifying the processes through easy pressure control and the elimination of heat- and mass-transfer limitations. In this study, we report a novel method for the controlled synthesis of micrometre-thick mesoporous TiO₂ catalytic coatings on the walls of long channels (>1 m) of capillary microreactors in a single deposition step. The method uses elevated temperature and introduces a convenient control parameter of the deposition rate (displacement speed controlled by a stepper motor), which allows deposition from concentrated and viscous sols without channel clogging. A capillary microreactor wall-coated with titania supported Bi-poisoned Pd catalyst was obtained using the method and used for the semihydrogenation of 2-methyl-3-butyn-2-ol providing 93 ± 1.5% alkene yield for 100 h without deactivation. Although the coating method was applied only for TiO₂ deposition, it is nonetheless suitable for the deposition of volatile sols.

Full text of DOI:10.1039/c4lc01066c

Lab on a Chip
View Article Online
View Journal | View Issue
PAPER
Novel synthesis of thick wall coatings of titania
supported Bi poisoned Pd catalysts and
application in selective hydrogenation of
acetylene alcohols in capillary microreactors†
Cite this: Lab Chip, 2015, 15, 1952
a
a
b
Nikolay Cherkasov,‡ Alex O. Ibhadon* and Evgeny V. Rebrov
Catalysis in microreactors allows reactions to be performed in a very small volume, reducing the
environmental problems and greatly intensifying the processes through easy pressure control and the
elimination of heat- and mass-transfer limitations. In this study, we report a novel method for the con-
2
trolled synthesis of micrometre-thick mesoporous TiO catalytic coatings on the walls of long channels
(>1 m) of capillary microreactors in a single deposition step. The method uses elevated temperature and
introduces a convenient control parameter of the deposition rate (displacement speed controlled by a
stepper motor), which allows deposition from concentrated and viscous sols without channel clogging. A
capillary microreactor wall-coated with titania supported Bi-poisoned Pd catalyst was obtained using the
method and used for the semihydrogenation of 2-methyl-3-butyn-2-ol providing 93 ± 1.5% alkene yield
Received 10th September 2014,
Accepted 18th February 2015
DOI: 10.1039/c4lc01066c
for 100 h without deactivation. Although the coating method was applied only for TiO
2
deposition, it is
www.rsc.org/loc
nonetheless suitable for the deposition of volatile sols.
the particles of supported catalysts, is possible in micro-
reactors. However, great care should be taken to ensure that
the catalyst particles do not block the channels and that
Introduction
Mesoporous metal oxide coatings have many applications in
photocatalysis, gas chromatography, electronics and many
1
5–18
mass and heat transfer limitations do not apply.
A slurry
1
–7
other areas.
In particular, they play an important role as
of catalyst nanoparticles can be introduced into a flow reac-
tor to overcome the problems of clogging and mass-transfer,
but this approach creates a problem of catalyst separation
catalysts or catalyst supports for microfluidic applications
where a reaction is performed in a microreactor with the
volume smaller than a millilitre. Microreactors have many
advantages over traditional large-scale chemical reactors: the
small dimensions of microreactors facilitate mass- and heat-
transfer, small reaction volume allows the handling of hazard-
ous or expensive chemicals safely and high-pressure reactions
to be carried out with inexpensive equipment, hence signifi-
cant process intensification is possible with microreactor
1
9
from the product at the end of the reaction. Hence, the
most convenient way of performing catalytic reactions in flow
systems is to coat the reactor walls with a catalyst. The coat-
ing can be done by grafting the reactor walls with meta-
llorganic molecules, but this approach requires very expen-
20
sive chemicals and complex synthesis procedures. A more
common approach is the deposition of metallic catalysts
supported on oxide materials. There are well-developed
methods of introduction of catalytically active metal nano-
particles into oxide matrixes, but subsequent deposition of
the matrix uniformly on the reactor walls is still a very chal-
lenging task, especially for concentrated sols and long
8–14
technology.
In heterogeneous catalytic reactions however, the advan-
tages of microreactors are offset by the problem of catalyst
introduction into the reactor. A convenient laboratory prac-
tice to use packed-bed reactors, i.e. fill the microreactor with
2
1–23
reactors.
a
Two main methods are used to obtain a uniform coating
on a substrate: spin- and dip-coating. In the spin-coating
technique, the substrate is first covered with an excess of pre-
cursor solution that is later removed by centrifugal forces
Catalysis and Reactor Engineering Research Group, Department of Chemistry
and School of Biological, Biomedical and Environmental Sciences, University of
Hull, Cottingham Road, Hull, HU6 7RX, UK. E-mail: a.o.ibhadon@hull.ac.uk
b
School of Engineering, University of Warwick, Coventry, CV4 7AL, UK
†
Electronic supplementary information (ESI) available. See DOI: 10.1039/
2
4
leaving a thin uniform coating after solvent evaporation.
c4lc01066c
Thus, spin-coating is suitable only for flat or slightly curved
surfaces and cannot be used to create coatings inside
‡
Present address: School of Engineering, University of Warwick, Coventry, CV4
AL, UK.
7
1952 | Lab Chip, 2015, 15, 1952–1960
This journal is © The Royal Society of Chemistry 2015

View Article Online
Lab on a Chip
Paper
capillary channels. The dip-coating method on the other
hand, is also traditionally applied to flat substrates that are
submerged and withdrawn from a solution containing a pre-
cannot be achieved by simply changing the solvent. There-
fore, the solvent removal from the reactor should be
performed under unsteady conditions.
2
4
cursor of metal oxide in a volatile solvent. Depending on
the withdrawing speed, three deposition regimes can be
observed: the capillary, the draining or a combination of both
Static and dynamic coating methods were developed for
chromatographic applications and addressed the problem of
solvent evaporation. With dynamic coating, a plug of the pre-
cursor solution is pushed through the reactor leaving some
solution on the column walls. The solution is later dried by
passing excess gas through the capillary. Static coating
implies filling the reactor with a precursor solution, closing
one end of the capillary and applying vacuum to the other
2
5
regimes. The capillary regime is observed for very slow
−
1
withdrawing speeds (<0.1 mm s ) and is associated with the
capillary rise in the formed solid film. The draining regime,
on the contrary, dominates at high withdrawing speeds
−
1
(
>1 mm s ) due to quick dragging of the viscous fluid. For a
3
4,35
given system, the conditions that separate these regimes and
the resulting coating thickness is determined by a combina-
tion of liquid viscosity, surface tension and most importantly,
end to facilitate evaporation.
The static coating method
offers a simple way of achieving uniform coating synthesis
with a thickness that can be adjusted by changing the precur-
2
5
35–37
the liquid evaporation rate. By careful optimisation of the
dip-coating conditions, very fine control of the coating can be
sor solution concentration.
Other experimental parame-
ters such as the evaporation temperature and the pressure
should be optimised to slow the evaporation rate required to
obtain uniform coatings. However, if thick coatings are to be
obtained, the concentration of the precursor solution should
be increased. This leads to increased viscosity, greatly limits
the range of adjustable parameters, decreases evaporation
rate, and significantly raises the probability of plug formation
2
6
achieved as demonstrated by Krins et al., who obtained a
crack-free TiO coating with a thickness of up to 0.5 μm in a
2
single deposition stage.
A method similar to dip coating adapted for tubular sub-
strates, is called the gas displacement method. A horizontal
tube to be coated is first filled with the precursor solution,
then gas displaces the liquid leaving some liquid along the
3
6,37
inside the capillaries.
Hence, the preparation of long
2
7
walls. This method was applied to obtain a 5–20 μm thick
microreactors with thick catalytic coatings is still a very chal-
lenging task.
2
8
coating inside shorter (0.3 m) reactors or a thin coating
2
3,29
(90 nm) inside 10 m long reactors.
Another derivative of
In this work, a novel sol–gel-based method is reported for
the synthesis of metal oxide coatings suitable for long capil-
lary microreactors. The method is similar to the static coating
method, but uses elevated temperature and introduces an
additional external parameter, displacement speed, that con-
trols the deposition rate. In order to demonstrate the applica-
tion of the method in the synthesis of catalytic coatings, the
liquid phase semihydrogenation of 2-methyl-3-buyne-2-ol
(MBY) to 2-methyl-3-buten-2-ol (MBE) was selected because of
the precise reaction control that is needed to stop the reac-
tion at the alkene stage and the importance of this reaction
in the synthesis of many fine chemicals such as vitamins
the dip-coating method, the fill-and-dry method, can be used
to obtain 10–20 μm thick catalytic coatings inside short
3
0–32
reactors.
However, the essential component of these methods, sol-
vent evaporation, requires years to remove solvent from a
longer (5 m) reactor according to estimations performed
3
3
using Stefan's equation, Fig. 1. As expected, evaporation
time decreases at high vapour pressure, i.e. close to the boil-
ing point of the solvent, however, the shortest evaporation
time observed in the presented parameter range, 1.3 years, is
still unacceptably high and requires very tight temperature
control. Clearly, a 1000-fold increase in the evaporation rate,
which is needed to perform the coating within several hours,
3
8
(A, E), and fragrances (Linalool). Traditionally, semihydro-
genation reactions are performed in batch reactors with lead-
containing Lindlar catalysts, where lead blocks some active
sites and affects the adsorption energy of intermediate
3
9,40
species, improving the alkene yield.
However, the high
toxicity of lead raises environmental concerns and as a result,
4
1
Bi was used as a doping agent, a notably less toxic metal.
In addition, the synthesis of Bi-poisoned Pd catalyst is
4
0,42
simple, meaning low catalyst costs.
Experimental section
2
TiO sol preparation
A titanium dioxide sol was prepared using an established
4
3
method by hydrolysis of titanium alkoxide solution. Hot
distilled water (360 K, 8.1 g) was quickly added to a solution
of 20.2 g titaniumIJIV) butoxide (Sigma-Aldrich, 97 wt.%) in
Fig. 1 Effect of temperature and pressure on the evaporation time of
6
.60 g of methanol (Fisher Scientific, 99.9 wt.%) under vigor-
the 5 m long reactor filled with ethanol, estimated using Stefan's
equation.
3
3
ous stirring to obtain TiO2 particles (hydrolysis ratio
This journal is © The Royal Society of Chemistry 2015
Lab Chip, 2015, 15, 1952–1960 | 1953

View Article Online
Paper
Lab on a Chip
[
H O]/[Ti] = 15). After 15 min of stirring, 0.78 g of concen-
coated using the same method to ensure the reproducibility
of the method.
2
trated HNO3 (Fisher Scientific, 65 wt.%) was added to dis-
perse any agglomerated TiO and the dispersion was refluxed
for 4 h. Pluronic F127 surfactant (Sigma-Aldrich), 1.0 g, was
added and the solution was left stirring for 2 h at room tem-
perature. The dispersion obtained was left in an ultrasonic
bath for 20 min and diluted with methanol to obtain titania
2
To obtain enough material for characterisation, a glass
tube with an internal diameter of 1.5 mm was also coated
−
1
with the 80 g L titania sol using the method described
above. The coating obtained was mechanically removed and
used for further characterisation.
−
1
sols with concentrations ranging from 5 to 80 g L .
2
Synthesis of the Pd–Bi/TiO coatings
2
Synthesis of TiO coatings
The selective poisoning of Pd active sites with Bi was
performed by adapting the method of Anderson and co-
The coating apparatus consists of a stepper motor, controlled
by an Arduino microcontroller, which pushes the capillary
vertically at a constant velocity into the oven, Fig. 2. The cap-
illary is guided into the oven by a 3 mm coiled copper tubing
insulated with ceramic wool for the first 5 cm that allows
gradual temperature increase along this length. The total
length of the copper guide is longer than that of the capillary
to ensure that the capillary is held at high temperature to
prevent condensation of the solvent within the capillary.
Fused silica capillaries with internal diameter of 250 μm
were used unless otherwise stated. Prior to the coating depo-
sition, the capillaries were activated by passing the following
4
0,42
workers.
Unsupported Bi-poisoned palladium nano-
particles were synthesised by reducing 45 mg of palladiumIJII)
acetate (Sigma-Aldrich, 98 wt.%) in the presence of 250 mg of
polyvinylpyrrolidone (Sigma-Aldrich) which were dissolved in
15 mL of methanol. A solution of 20 mg sodium borohydride
(Sigma-Aldrich, 98 wt.%) in 5 mL of methanol was quickly
added and left stirring for 1 h. Any excess of sodium borohy-
dride was neutralised by adding 0.2 ml of acetic acid (Fisher
Chemical, 99 wt.%). Afterwards, 1.7 mL of an aqueous 0.02 M
BiIJNO ) solution in 2% acetic acid was added to the disper-
3
3
sion obtained with stirring for 3 h at 320 K in hydrogen atmo-
sphere (1 bar). The solvent was evaporated using a rotary
evaporator until the total solution volume reached 7 mL.
The Pd–Bi nanoparticle dispersion was mixed with the
−
1
solutions at 20 μL min for 15 min: water, a 1 M NaOH solu-
tion, water, a 2 M Na Cr O in 5 M H SO solution and water.
2
2
7
2
4
−
1
The capillary (0.5–1 m long) was then slowly (7 μL min )
filled with the titania sol using a syringe pump and the free
end of the capillary was closed with a shut-off valve. The first
−
1
2
80 g L TiO sol in 1 : 1 volume ratio and left in an ultrasonic
bath for 20 min. The dispersion obtained was used to fill a
2.5 m fused silica capillary (0.53 mm internal diameter) and
prepare a coating as described above using a displacement
5
cm of the capillary was left without the dispersion to mini-
mise clogging during the initial stages of the coating. The
filled capillary was displaced into the oven at a constant lin-
ear velocity (displacement speed) of 0.1–10 mm s by the
stepper motor (step length 20 μm).
−
1
velocity of 0.2 mm s and the oven temperature of 450 K.
−
1
Semihydrogenation
The capillaries obtained left at 473 K in a vacuum oven for
1
0 h to stabilise the coating, flushed with ethanol at 300 μL
A schematic view of the hydrogenation apparatus is shown in
Fig. 3. It consists of two mass-flow controllers (Aalborg) for
hydrogen and nitrogen (CK gas, 99.999 vol.%), a continuous-
flow syringe pump (SyrDos2) connected with a T-joint to the
−
1
min for 20 min to remove any loose fragments of the coat-
ing and to extract the surfactant. The capillaries were then
dried at 390 K in air. At least, three identical capillaries were
capillary microreactor. The capillary was placed in
a
temperature-controlled water bath and connected to a sample
collector. The residence time in the reactor was estimated by
injection of a pulse of octane 10 μL (Alfa Aesar, 99 vol.%) into
the T-joint and analysing its concentration at the reactor out-
−
1
let. At the liquid flow rate of 40 μL min , the average resi-
dence time was determined by pulsing methylene blue to be
1
tion.
20 s, which agrees with the Lockhart–Martinelli correla-
4
4,45
The semihydrogenation of 2-methyl-3-butyn-2-ol
(
MBY, Sigma-Aldrich, 98 wt.%) was performed using a 1200
−
3
mol m MBY solution in hexane (VWR, 98 wt.%). The reac-
tion conditions were optimised in the temperature range
from 310 to 335 K, hydrogen flow rate of 10 mL min (STP)
and a solution flow rate of 30–50 μL min . During the opti-
misation, the experimental conditions were set, the system
was allowed to reach a steady state within 30–60 min, then
−
1
−
1
3
–5 samples were collected (100 μL) within 3 h of operation.
Fig. 2 Schematic view of the coating apparatus.
The samples were diluted in a 1 : 10 ratio with hexane and
1954 | Lab Chip, 2015, 15, 1952–1960
This journal is © The Royal Society of Chemistry 2015

View Article Online
Lab on a Chip
Paper
Results and discussion
Titania coated capillary reactors
Metal oxide sols in high concentrations present many practi-
cal problems for the coating process. Firstly, these sols are
very viscous non-Newtonian fluids that easily clog the capil-
2
8
laries on deposition. Secondly, the boiling points of metha-
nol and butanol of 338 and 390 K, respectively, require a tem-
perature above 395 K for complete solvent removal. However,
controlled and slow solvent evaporation at high temperature
is very difficult due to vigorous solvent boiling. The method
reported in this work addresses these problems by using an
elevated temperature and pressure as opposed to convention-
3
4–37
ally used vacuum-based techniques.
The temperature of
the copper guide tubing at the entrance to the oven (Fig. 2)
increases linearly from about 350 K (air-cooled side) to the
oven temperature. Hence, when the capillary is moving into
the oven, its temperature gradually increases leading to the
preferential evaporation of the most volatile components in
the solution followed by full evaporation.
Fig. 3 Scheme of the hydrogenation apparatus.
The method offers several adjustable experimental param-
eters such as the titania sol concentration, the temperature
of the oven and the displacement speed. Aiming for the
highest coating thickness, a titania sol with a concentration
analysed using a Varian 430 gas chromatograph fitted with
an autosampler and a 30 m Stabiwax capillary column.
−
1
of 80 g L was used in the parametric study. First, the capil-
laries were coated at oven temperatures of 390, 420 and
Characterization of TiO coatings
−1
2
450 K and a displacement speed of 0.1 mm s . At the tem-
Scanning electron microscopy analysis. Scanning electron
microscopy (SEM) study was performed using a TM-1000
Hitachi Tabletop Microscope. The capillary was cut into
peratures of 390 and 420 K, the capillaries were clogged by a
solid deposit, while at 450 K the deposition process was
reproducible and the coating thickness was uniform along
the length. The absence of clogging at 450 K might be due to
(i) the decreased viscosity at the higher temperature, and
importantly, (ii) higher vapour pressure. Because methanol
vapour pressure is 14 and 27 bar at 420 and 450 K, respec-
tively, the removal of a plug formed inside the capillary is
more likely at a higher temperature due to pressure build up.
Fig. 4a shows a typical SEM image of the cross-section and
the pore size distribution (PSD) of a 250 μm capillary coated
with a 4 μm-thick layer of titania obtained in a single deposi-
tion stage. A mean pore size of 5.7 nm was observed. The
4
mm sections evenly distributed along the capillary length.
These sections were vertically glued on a sample holder. For
every capillary, up to 30 individual cross-sections were
studied measuring between 10 and 20 positions along the
angular direction for analysis of thickness distribution.
Pressure drop analysis. The hydrodynamic diameter of the
coated capillaries was calculated from the Hagen–Poiseuille
equation. In these experiments, ethylene glycol was fed at a
−
1
desired flow rate in the range of 50 to 300 μL min and the
pressure drop was measured with a pressure transducer
(Alicat PC100). The viscosity of the ethylene glycol at room
2
coating consisted of two phases of TiO : anatase and brookite
temperature was determined using an untreated capillary
with a diameter of 250 ± 0.7 μm. The thickness of the coating
was calculated as the difference between the initial diameter
and the hydraulic diameter of the coated capillary.
in a 30/70 ratio according to the PXRD data (Fig. 4b). The
crystallite size of the sol was relatively small (5.1 ± 1.8 nm
estimated using Williamson–Hall plot) due to high hydrolysis
4
3
ratio used for sol preparation.
BET and PXRD measurements. The surface area and pore
size distribution of the titania coatings were measured from
In the deposition method, the displacement speed serves
as a convenient external control parameter of the evaporation
−
1
N
2
adsorption–desorption isotherms (Energas, 99.999 vol.%)
rate. At a displacement speed above 5 mm s , the coating
thickness was not constant in the angular direction (Fig. 5a).
using a TriStar 3000 micrometrics analyser using standard
multipoint BET analysis and BJH pore distribution methods.
Prior to nitrogen adsorption, the material was dried in nitro-
gen flow at 410 K for 3 h. PXRD patterns were recorded using
an Empyrean powder X-ray diffractometer equipped with a
monochromatic Kα-Cu X-ray source in the 2θ range 20–80°,
step size 0.039° and step time 180 s. The patterns obtained
were analysed using PANalytical Highscore Plus software.
−
1
However, at a displacement speed below 0.2 mm s , the
coating thickness was uniform with random deviations from
the mean value (Fig. 5b). At an intermediate displacement
speed, the thickness distribution of the coatings along the
capillary circumference was intermediate between these two
cases. The non-uniform coating along the angular direction
can be explained considering that at high displacement
This journal is © The Royal Society of Chemistry 2015
Lab Chip, 2015, 15, 1952–1960 | 1955

View Article Online
Paper
Lab on a Chip
Fig. 4 (a) PSD of the TiO
2
coatings, (insert) SEM image of the cross-section of the titania coated 250 μm capillary, and (b) the PXRD pattern of
the coating.
−
1
−1
Fig. 5 Cross-sectional SEM images of the capillaries coated at (a) 0.2 mm s , (b) 5 mm s and the coating thickness as a function of angular
position.
speed, there was not enough time for the solvent to evaporate
in the vertical part of the guide tubing. Hence, solvent evapo-
ration took place at the bend or in horizontal section (Fig. 2),
which led to a spreading of the sol at the capillary bottom by
gravitational forces. At low displacement speed, the evapora-
tion rate was slow enough so the solvent evaporation took
place in the vertical section forming a uniform coating.
The effect of sol concentration on the coating properties
are in good agreement (considering standard deviation of the
experimental data indicated as confidence intervals in the
plot), which indicates that a uniform coating was obtained
with the absence of plugs or necks, where the pressure drop
was expected to be the highest.
For capillaries with smaller internal diameters, solid plugs
2
8
are likely to be formed due to capillary forces, but the coat-
ing method reported can be applied to coat capillaries with
larger diameters. For a given TiO2 precursor concentration,
the coating thickness is proportional to the internal diameter
of the capillary (Fig. 7) allowing one to obtain coating thick-
ness above 10 μm in the tubes about 1 mm in diameter in a
single deposition step.
−
1
was studied at a displacement speed of 0.2 mm s and a sol
−
1
concentration ranging from 5 to 80 g L . In these cases, the
distribution of the coating thickness in the angular direction
was uniform as in Fig. 5b, while the axial thickness changed
with concentration. Fig. 6a shows the axial thickness distri-
bution determined by studying cross-sectional SEM images.
As expected, the coating thickness increased at higher sol
concentrations, because more titania was introduced into the
capillaries. Fig. 6b shows a comparison of the average coating
thickness determined using the SEM images and hydrody-
namic measurements. These data on the coating thickness
Semihydrogenation in the capillary reactor
In order to test the coating method for catalytic applications,
MBY semihydrogenation on the Bi-poisoned Pd catalyst was
performed. To prepare the catalyst,
a
dispersion of
1956 | Lab Chip, 2015, 15, 1952–1960
This journal is © The Royal Society of Chemistry 2015

View Article Online
Lab on a Chip
Paper
Fig. 6 (a) Distribution of the TiO
2
coating thickness obtained in the 250 μm fused silica capillaries using different precursor sol concentrations.
−
1
Displacement speed: 0.2 mm s . Oven temperature: 450 K. (b) The comparison of the average coating thickness determined by (●) SEM and (♦)
hydrodynamic study.
−
1
unsupported Pd–Bi nanoparticles was mixed with a 80 g L
parameter optimisation, the solution flow rate was varied in
−
1
TiO sol in a 1 : 1 volume ratio and the mixture was deposited
the interval of 30–50 μL min at a constant hydrogen gas
2
−
1
onto the inner walls of a 2.5 m silica capillary with an inter-
nal diameter of 530 μm. After thorough surfactant extraction
with ethanol followed by drying, a coating thickness of 3.6 ±
.1 μm was obtained with a nominal Pd loading of 3.3 wt.%,
and a total catalyst loading of 5 mg m . During the
flow rate of 10 mL min (STP), which corresponded to super-
−
1
−1
ficial flow velocities of 2–4 mm s for liquid and 0.75 m s
for hydrogen. Reynolds numbers were below 5 showing lami-
nar liquid and gas flows. Image analysis showed that a slug-
1
−
1
46,47
annular flow regime was realised.
Fig. 8 shows the performance of the capillary microreactor
−
3
in the hydrogenation of a 1200 mol m MBY solution in hex-
ane at reaction temperatures of 310 and 335 K. As the flow
rate increased, the MBY conversion decreased due to shorter
residence time, while the selectivity towards MBE increased
reaching 98% at MBY conversion up to 80%. Importantly,
even at high MBY conversions, the MBE selectivity was high,
allowing for a maximum MBE yield of 95.0 and 94.3% at 335
and 310 K, respectively.
The very high selectivity observed on the Pd–Bi catalyst
can be explained by a combination of thermodynamic effects
and active site poisoning. The former means the displace-
ment of alkene species from the catalyst surface due to much
higher heat of adsorption of alkynes which results in about
48–51
90–95% alkene selectivity over monometallic Pd catalysts.
The latter, poisoning of Pd catalysts with Bi, further in-
creased selectivity by blocking the step and edge sites of the
Fig. 7 Effect of the capillary inner diameter on the coating thickness
−1
(
2
SEM data) using a 80 g L TiO sol.
−
3
Fig. 8 Summary of the performance of the capillary microreactor with Pd–Bi/TiO
2
in MBY semihydrogenation (1200 mol m MBY in hexane,
−1
1
0 mL min (STP) H
2
flow, (a) 310 K and (b) 335 K). Selectivity is calculated towards MBE.
This journal is © The Royal Society of Chemistry 2015
Lab Chip, 2015, 15, 1952–1960 | 1957

View Article Online
Paper
Lab on a Chip
5
6
Pd nanoparticles, as those sites are responsible for over
hydrogenation of MBY molecules.
A high long term stability is required for industrial appli-
cations of the microreactor. The catalyst stability was studied
0.81 bar, which corresponds to the hydrogen partial pres-
sure of 0.97 and 0.49 bar, respectively. As the hydrogenation
4
0,48,52
5
5
is a first order reaction in hydrogen, the reaction rate
was expected to decrease by a factor of 2.3 with a tempera-
ture increase from 310 to 335 K due to dilution of hydrogen
with hexane (ESI†). As a result, the increase in intrinsic
reaction rate due to higher temperature is balanced by its
decrease due to lower hydrogen pressure. No mass transfer
limitations were observed. The supporting calculations are
provided in ESI.†
−
1
at 335 K using an MBY solution flow rate of 42.5 μL min
−
1
and a hydrogen flow of 10 mL min for 100 hours. Fig. 9
shows that the MBE yield was 93 ± 1.5% with neither deacti-
vation nor decrease in selectivity observed. However, there
were some fluctuations in the product yield, also observed
during the parametric study (Fig. 8). These are likely related
to minor fluctuations in the flow rates, the thermostat tem-
perature (about ±0.5 K), and the pressure drop in the reactor
Conclusions
5
3
due to hydrodynamic reasons.
Interestingly, it can be seen from Fig. 8 that the reaction
temperature had a marginal effect on MBY conversion. In
particular, the maximum MBE yield was observed at almost
A novel method for the controlled deposition of micrometre-
thick mesoporous titania coatings inside long capillary
microreactors is proposed. The two main advantages of the
method over the conventional static coating procedure are
associated with the use of elevated temperature and introduc-
tion of the capillary at a constant displacement speed into
the oven. Elevated deposition temperature prevents capillary
clogging even with concentrated solutions, while the dis-
placement speed controls the evaporation rate. This allows
very slow solvent removal leading to the formation of uni-
form coatings. The coating thickness can be precisely con-
trolled and it is proportional to the sol concentration and the
internal diameter of the capillary, while textural properties
−
1
the same liquid flow rates of 40.0 and 42.5 μL min at 310
and 335 K, respectively, showing very close apparent reaction
rates (Fig. 8). However, an increase in the hydrogenation
reaction rate by a factor of 2.5 would be expected at 335 K as
5
4,55
compared with 310 K,
which should lead to much higher
conversion at a given liquid flow rate. This fact can be
explained by a substantial decrease in the hydrogen partial
pressure due to solvent evaporation and its diffusion to the
gas phase (Fig. 9). The hexane partial pressure at 310 and
3
35 K, estimated by the Antonie equation, was 0.33 and
4
3
can be adjusted with a surfactant. Although only titania
coatings were investigated in this work, the method can be
extended to the deposition of any thermostable materials,
particularly metal oxide sols.
The method was applied to coat the inner walls of a capil-
lary microreactor (i.d. 0.53 mm, 2.5 m long) with a titania
supported Bi-poisoned Pd catalyst. The microreactor demon-
strated exceptionally high selectivity in semihydrogenation of
MBY with the alkene yield up to 95% and with no observed
catalyst deactivation for a time on stream of 100 h. The
alkyne hydrogenation performance was significantly higher
than for the previously reported capillary microreactors
(
Table 1).
The capillary microreactor produced in this work offers
opportunities for further intensive and extensive perfor-
mance improvement using high hydrogen pressure, selective
microwave catalyst heating, increasing the reactor length and
Fig. 9 Long-term stability of the capillary microreactor with Pd–Bi/
−
3
TiO
2
in MBY semihydrogenation (1200 mol m MBY in hexane flow
−
1
−1
4
2.5 μL min , 10 mL min (STP) H flow, 335 K).
2
Table 1 Comparison of the microreactor performance in semihydrogenation
a
b
c
Max. alkene yield, %
R
app
,
R
vol
,
cat
R ,
−1
react−3 −1
−1
cat−1
References
Reactor
Catalyst
Pd/TiO
PdZn/TiO
PdZn/TiO
Alkyne
μmol s
μmol m
s
μmol s kg
2
2
3
9
d
Rebrov et al.
Rebrov et al.
Capillary
Capillary
2
PA
80.3
89.5
97.0
86.4
93.7
93.0 ± 1.5
65
23
19
1000
166
850
148
46
38
5680
950
1540
53
38
32
34
6
2
MBY
2
+ pyridine MBY
5
7
Liguori and Barbaro
Monolith Pd@MonoBor
Pd@MonoBor
Capillary
MBY
d
PA
Current work
PdBi/TiO
2
MBY
64
a
b
c
Reactor alkyne hydrogenation performance. Performance normalized for the reaction volume. Performance normalized per kg of the
d
catalyst. Phenylacetylene.
1958 | Lab Chip, 2015, 15, 1952–1960
This journal is © The Royal Society of Chemistry 2015

View Article Online
Lab on a Chip
Paper
the coating thickness or by numbering-up. Capillary
microreactors open a promising and cost-efficient way for
the replacement of traditional, energy- and labour-
inefficient batch reactor systems in the synthesis of high-
value pharmaceutical and fine chemistry products.
21 R. J. White, R. Luque, V. L. Budarin, J. H. Clark and D. J.
Macquarrie, Chem. Soc. Rev., 2009, 38, 481–494.
22 J. A. Schwarz, C. Contescu and A. Contescu, Chem. Rev.,
1995, 95, 477–510.
23 E. V. Rebrov, A. Berenguer-Murcia, H. E. Skelton, B. F. G.
Johnson, A. E. H. Wheatley and J. C. Schouten, Lab Chip,
2
009, 9, 503–506.
Acknowledgements
2
2
4 L. E. Scriven, Mater. Res. Soc. Symp. Proc., 1988, 121, 717–729.
5 M. Faustini, B. Louis, P. A. Albouy, M. Kuemmel and D.
Grosso, J. Phys. Chem. C, 2010, 114, 7637–7645.
Authors are grateful to Dr. C. Willies, Dr. R. Lewis, Dr. M.
Hird, Professor P. Fletcher and Professor B. Binks for the
access to their equipment, to the EU for an FP7 Grant for the
MAPSYN Project (MAPSYN.eu no. CP-IP 309376) and Royal
Society for an International Joint Project Grant 2010/R.
2
6 N. Krins, M. Faustini, B. Louis and D. Grosso, Chem. Mater.,
2
010, 22, 6218–6220.
2
2
7 G. I. Taylor, J. Fluid Mech., 1961, 10, 161–165.
8 J. Bravo, A. Karim, T. Conant, G. P. Lopez and A. Datye,
Chem. Eng. J., 2004, 101, 113–121.
9 E. V. Rebrov, E. A. Klinger, A. Berenguer-Murcia, E. M.
Sulman and J. C. Schouten, Org. Process Res. Dev., 2009, 13,
991–998.
30 S.-M. Hwang, O. J. Kwon and J. J. Kim, Appl. Catal., A,
2007, 316, 83–89.
31 J. F. Ng, Y. Nie, G. K. Chuah and S. Jaenicke, J. Catal.,
2010, 269, 302–308.
32 M. Ueno, T. Suzuki, T. Naito, H. Oyamada and S. Kobayashi,
Chem. Commun., 2008, 1647–1649.
33 J. Mitrovic, Chem. Eng. Sci., 2012, 75, 279–281.
34 G. Lambertus, A. Elstro, K. Sensenig, J. Potkay, M. Agah, S.
Scheuering, K. Wise, F. Dorman and R. Sacks, Anal. Chem.,
2004, 76, 2629–2637.
35 S. Reidy, G. Lambertus, J. Reece and R. Sacks, Anal. Chem.,
2006, 78, 2623–2630.
36 K. J. Grob and K. Grob, Chromatographia, 1977, 10, 250–255.
37 K. Grob and G. Grob, J. High Resolut. Chromatogr., 1983, 6,
133–139.
References
2
1
2
L. Liu and X. Chen, Chem. Rev., 2014, 114, 9890–9918.
Y. Shiraishi, D. Tsukamoto, Y. Sugano, A. Shiro, S. Ichikawa,
S. Tanaka and T. Hirai, ACS Catal., 2012, 2, 1984–1992.
H. Chen, C. E. Nanayakkara and V. H. Grassian, Chem. Rev.,
3
4
5
2
012, 112, 5919–5948.
P. Pfeifer, T. Zscherpe, K. Haas-Santo and R. Dittmeyer,
Appl. Catal., A, 2011, 391, 289–296.
T. Fröschl, U. Hörmann, P. Kubiak, G. Kučerová, M.
Pfanzelt, C. K. Weiss, R. J. Behm, N. Hüsing, U. Kaiser, K.
Landfester and M. Wohlfahrt-Mehrens, Chem. Soc. Rev.,
2
012, 41, 5313–5360.
6
7
8
9
Z. Li, B. Gao, G. Z. Chen, R. Mokaya, S. Sotiropoulos and G.
Li Puma, Appl. Catal., B, 2011, 110, 50–57.
A. M. Ramirez, K. Demeestere, N. De Belie, T. Mäntylä and
E. Levänen, Build. Environ., 2010, 45, 832–838.
K. Jähnisch, V. Hessel, H. Löwe and M. Baerns, Angew.
Chem., Int. Ed., 2004, 43, 406–446.
M. Irfan, T. N. Glasnov and C. O. Kappe, ChemSusChem,
38 W. Bonrath, J. Medlock, J. Schutz, B. Wüstenberg and T.
Netscher, Hydrogenation in the Vitamins and Fine
Chemicals Industry – An Overview, in Hydrogenation, ed. I.
Karame, InTech, Rijeka, 2012, pp. 69–90.
39 B. Bridier, N. López and J. Pérez-Ramírez, Dalton Trans.,
2010, 39, 8412–8419.
2
011, 4, 300–316.
1
1
0 V. Hessel, G. Cravotto, P. Fitzpatrick, B. S. Patil, J. Lang and
W. Bonrath, Chem. Eng. Process., 2013, 71, 19–30.
1 D. Reay, C. Ramshaw and A. Harvey, PROCESS
INTENSIFICATION: Engineering for Efficiency, Sustainability
and Flexibility, Butterworth-Heinemann, 2011.
40 J. A. Anderson, J. Mellor and R. K. P. K. Wells, J. Catal.,
2009, 261, 208–216.
1
2 G. Chen, J. Yue and Q. Yuan, Chin. J. Chem. Eng., 2008, 16,
6
63–669.
41 B. Bradley, M. Singleton and A. L. Wan Po, J. Clin. Pharm.
Ther., 1989, 14, 423–441.
42 J. Sá, J. Montero, E. Duncan and J. A. Anderson, Appl. Catal.,
B, 2007, 73, 98–105.
43 R. Bleta, P. Alphonse and L. Lorenzato, J. Phys. Chem. C,
2010, 114, 2039–2048.
1
3 V. Hessel, D. Kralisch, N. Kockmann, T. Noël and Q. Wang,
ChemSusChem, 2013, 6, 746–789.
4 J. R. Burns and C. Ramshaw, Lab Chip, 2001, 1, 10–15.
5 D. van Herk, P. Castaño, M. Makkee, J. A. Moulijn and M. T.
Kreutzer, Appl. Catal., A, 2009, 365, 199–206.
1
1
1
6 M. W. Losey, M. A. Schmidt and K. F. Jensen, Chem.
Innovation, 2001, 40, 2555–2562.
44 R. W. Lockhart and R. C. Martinelli, Chem. Eng. Prog.,
1949, 45, 39–48.
1
1
7 K. F. Jensen, Chem. Eng. Sci., 2001, 56, 293–303.
8 B. P. Mason, K. E. Price, J. L. Steinbacher, A. R. Bogdan and
D. T. McQuade, Chem. Rev., 2007, 107, 2300–2318.
9 A.-K. Liedtke, F. Bornette, R. Philippe and C. de Bellefon,
Chem. Eng. J., 2013, 227, 174–181.
45 D. Chisholm, Int. J. Heat Mass Transfer, 1967, 10, 1767–1778.
46 N. Shao, A. Gavriilidis and P. Angeli, Chem. Eng. Sci.,
2009, 64, 2749–2761.
47 J. Yue, G. Chen, Q. Yuan, L. Luo and Y. Gonthier, Chem. Eng.
Sci., 2007, 62, 2096–2108.
1
2
0 J. Kobayashi, Y. Mori, K. Okamoto, R. Akiyama, M. Ueno, T.
Kitamori and S. Kobayashi, Science, 2004, 304, 1305–1308.
48 M. Crespo-Quesada, A. Yarulin, M. Jin, Y. Xia and L. Kiwi-
Minsker, J. Am. Chem. Soc., 2011, 133, 12787–12794.
This journal is © The Royal Society of Chemistry 2015
Lab Chip, 2015, 15, 1952–1960 | 1959

View Article Online
Paper
Lab on a Chip
4
5
5
5
9 A. N. R. Bos and K. R. Westerterp, Chem. Eng. Process.,
53 V. van Steijn, M. T. Kreutzer and C. R. Kleijn, Chem. Eng. J.,
2008, 135, S159–S165.
1
993, 32, 1–7.
0 K. Reuter and M. Scheffler, Phys. Rev. B: Condens. Matter
Mater. Phys., 2006, 73, 1–17.
1 A. Borodziński and G. C. Bond, Catal. Rev.: Sci. Eng.,
54 M. Crespo-Quesada, M. Grasemann, N. Semagina, A. Renken
and L. Kiwi-Minsker, Catal. Today, 2009, 147, 247–254.
55 N. Semagina, E. Joannet, S. Parra, E. Sulman, A. Renken and
L. Kiwi-Minsker, Appl. Catal., A, 2005, 280, 141–147.
56 G. F. Carruth and R. Kobayashi, J. Chem. Eng. Data, 1973, 18,
115–126.
2
006, 48, 91–144.
2 J. A. Bennett, G. A. Attard, K. Deplanche, M. Casadesus, S. E.
Huxter, L. E. Macaskie and J. Wood, ACS Catal., 2012, 2,
5
04–511.
57 F. Liguori and P. Barbaro, J. Catal., 2014, 311, 212–220.
1960 | Lab Chip, 2015, 15, 1952–1960
This journal is © The Royal Society of Chemistry 2015