8
N. Thanh-Binh et al. / Applied Catalysis A: General 520 (2016) 7–12
are prone to overheating and require special reactor design to
enhance local heat transfer. In 1972, Batist et al. reported sev-
eral ways for synthesis of Bi-Mo-O based on Bi(NO3)3· 5H2O,
Recently, the bismuth molybdates were obtained by several meth-
ods such as combining complexation and spray drying [25,26],
surfactant assisted hydrothermal treatment [27], hydrothermal
treatment [28], reflux and solid state [29], and hard-templating
[30]. However, the specific surface area of the bismuth molyb-
2–4 m2/g). Therefore, preparing bismuth molybdate with high sur-
face area requires attention. On the other hand, the different phases
of bismuth molybdate also show different effects on catalytic per-
formance [4,31]. Bismuth molybdates have three phases including
˛ (Bi2Mo3O12), ˇ (Bi2Mo2O9), and ꢀ (Bi2MoO6). For ammoxida-
tion of propylene, the ˛ phase shows good NH3 activation and ˛-H
abstraction. The ꢀ phase is good for the reoxidation process. Finally,
the ˇ phase combines ˛ and ꢀ structures, so this phase is the most
effective for ammoxidation of propylene. No report has however
described the effect of the phase on the ammoxidation of acrolein
Metal oxides supported on mesostructured silica are widely
applied in the catalysis field because of their large specific sur-
face area, high pore volume, large pore size, and high stability
[32,33]. The bismuth and molybdate precursors precipitate from
their solutions upon mixing, so that a suitable method should be
considered. Several reports indicated that in a strong acid, or base,
or in glycerol, the mixture remains homogeneous [24,34,35]. How-
ever, strong acids or bases will change silica surface properties. In
addition, glycerol has high viscosity, making it difficult for the pre-
cursors to diffuse into the pores of the mesoporous silica. Yen et al.
synthesized mesostructured metal oxides using a hard-templating
technique based on a dual-solvent and solid–liquid impregnation
method [36]. A non-polar solvent (heptane in our work) was used
for pre-wetting, so as to reduce surface tension (surface tensions of
solid–gas are larger than surface tensions of solid–liquid) [36,37].
Therefore, metal salts can be easily transported inside the pores of
the mesoporous silica. In addition, using this method in absence of
water will reduce precipitation of precursors.
compare supports, catalyst C3 was synthesized using commercial
porasil silica (Milipore Corporation, 34 Maple Street, Milford, MA
01757) as support and same elemental composition as catalyst n3
active phase.
In preparing the non-supported catalyst a similar impregnated
silica was prepared and the KIT-6 silica support was then removed
by digestion in 2M NaOH solution at room temperature (three times
over one day). A total amount of 4.5 g of the mixed precursors was
impregnated on 1.25 g of KIT-6. The resulting solids were filtered
and calcined at 500 ◦C for 3 h before silica removal. The final bis-
muth molybdate mixed oxide was then washed for several times
in water and aqueous ethanol and dried at 100 ◦C.
2.2. Characterization
N2 physisorption isotherms at 77 K were measured using a
Quantachrome Nova 2000 series instrument. The samples were
preliminarily degassed in vacuumn at 150 ◦C for 6 h. Specific sur-
face area of the catalysts was calculated using the linear part of the
BET plot (0.05–0.2 in relative pressure). The pore size distribution is
calculated from the adsorption branch following the NLDFT (non-
local density functional theory) method for cylinder shape. The pore
volume is taken as the adsorbed nitrogen volume at 0.99 relative
pressure. Wide-angle X-ray diffraction (XRD) analysis was per-
formed with a Siemens 80 Model D5000 diffractometer using CuK˛
radiation (ꢁ = 0.15496 nm). In addition, the catalysts phases were
monitored by Raman spectroscopy LABRAM HR800 (Horiba Jobin
Yvon, Villeneuve d’Ascq, France) coupled with an Olympus BX30
fixed stage microscope using Ar+ laser (514.5 nm) as an excitation
light source (Coherent, INNOCA 70C Series Ion Laser, Santa Clara,
CA). The chemical composition of the samples was established by
Inductively Coupled Plasma-Optical Emission Spectroscopy (ICP-
OES) using a Perkin Elmer Optima 4300DV spectrometer.
2.3. Catalysts test
50 mg of the catalysts was loaded in a fixed-bed quartz reac-
tor (inner diameter and length 8 mm and 420 mm, respectively) at
(GC, HP 5890). The low catalyst mass allows avoiding hot spots.
The reactants and products were analyzed using a TCD detector
and Haysep P and molecular sieve 13X columns. The reactor sys-
tem is shown in Fig. 1. Gas feeds including acrolein, air, ammonia,
and nitrogen enter the quartz reactor through two separate inlets
in order to avoid polymerization at the contact between ammonia
and acrolein. One feed comprises air, ammonia (0.47 cc/min), and
nitrogen diluent. Acrolein is vaporized from a 95% aqueous solu-
tact time) on catalytic reaction, reactant molar ratio (AC/NH3/O2)
was fixed whereas the diluent nitrogen flow rate was varied. All
inlet gas compositions were selected outside of the flammability
region (see Figs. S1 and S2 in supporting information). On the other
hand, studying the effect of reactant molar ratio on catalytic per-
formance, the flow rates of air and diluent nitrogen were changed
to keep total flow rate constant (detail information in Table S1). In
order to avoid condensation of polyacrolein and polyacrylonitrile,
the system lines are heated (red-lines) to 180 ◦C and several three
way valves (V1, V3, V4, V5) were used. After each catalytic test,
methanol is flown through part of the system for cleaning. In order
to protect the GC sampling loop, the reactor exhaust is send to the
vent using valves V3 and V4 between samplings.
Hence, the effect of surface area and the phases of the bismuth
molybdate catalysts could be studied for ammoxidation of acrolein
to acrylonitrile by using Yen et al. one-step impregnation catalyst
preparation method. The catalysts including supported on meso-
porous 3D KIT-6 and non-supported bismuth molybdate will be
investigated in this work.
2. Experimental
2.1. Synthesis of bismuth molybdate catalysts
large pore size (8.1 nm), and 3D mesopore connectivity structure,
synthesized, at 100 ◦C aging temperature, according to a previous
report [38] was used as support. The catalysts designed as differ-
ent mixtures between Bi2O3 and MoO3 were synthesized based
on the method developed by Yen et al. [36]. Specifically, the mix-
tures Bi2O3.nMoO3 with n = 1, 2, and 3 supported on KIT-6 were
designated as n1, n2, and n3, respectively. First, 1.25 g evacuated
KIT-6 was pre-wetted by n-heptane, and then pre-mixed with
0.94 g of the precursors including calculated weights of bismuth
nitrate Bi(NO3)3· 5H2O and ammonium molybdate tetrahydrate
(NH4)6Mo7O24· 4H2O. The mixture was transferred to a round bot-
tom flask and heated at 85 ◦C overnight. The solids were filtered and
dried at 50 ◦C, and then calcined at 550 ◦C for 3 h in air. In order to
A carbon balance was calculated based on all detected prod-
ucts including carbon dioxide, acrolein, acetonitrile (ACE), and
acrylonitrile (carbon monoxide was never detected). The reported