G Model
CATTOD-9881; No. of Pages9
ARTICLE IN PRESS
2
B. Yan et al. / Catalysis Today xxx (2015) xxx–xxx
RbNa and CsNa catalysts with respectively widely varied Rb+-
and Cs+-exchange degrees would be deserving as the appropriate
ion-exchange windows of Rb+ and Cs+ ions for the LTA reaction
remained unexplored in the earlier report.
We report herein the surface acid–base property of these
RbxNa1−x and CsxNa1−x catalysts and their performance in cat-
alyzing the LTA reaction. The data obtained in this study are
discussed by comparison with those obtained previously on the
KxNa1−x samples, which confirm that suitably balanced acidity
and basicity at the catalyst surface is the key to the high efficiency
quartz reactor (50 cm × 9 mm (i.d.)). A constant weight of catalyst
(500 mg, 20–40 mesh) was sandwiched in the middle of the reactor
with quartz wools. About 2 ml of quartz sands (ca. 2 cm by height)
were placed above the catalyst bed to ensure complete evapora-
tion of the liquid feed. Prior to injection of the reaction feed, an
aqueous solution containing 10 mol% (35.7 wt%) LA, the catalyst
was pretreated at the reaction temperature for 1 h in flowing N2
(25 ml/min). A flow of N2 (15.5 ml/min) was used as the reaction
carrier gas, which combined with the vaporized liquid feed to make
a feeding gas mixture containing 7.4 kPa LA and 66.9 kPa H2O (N2
of the LTA reaction. Besides, the present RbxNa1−x and CsxNa1−x

balance). The reaction weight hourly space velocity by LA (WHSVLA)
catalysts are found superior to the previous KxNa1−x catalysts in
was 2.1 h−1. Each run of the reaction was conducted for 10 h, dur-
trap and collected hourly for off-line analysis by gas chromato-
graph (GC) and ion chromatograph (IC). The materials balance of
the reaction was always higher than 95%, except in the very first
hour. As reported earlier [9], the GC analysis would usually not be
[9]. Further details of the GC and IC analyses including their cross
calibrations can be found in Refs. [9,30].
catalyzing the LTA reaction in terms of AA selectivity and yield.
2. Experimental
Raw powders of as-synthesized Beta zeolite (SiO2/Al2O3 = ca. 40,
Nankai University Catalyst Co., Ltd.) were calcined in ambient air
at 420 ◦C for 1 h, then at 540 ◦C for 5 h to completely remove any
organic residue. The calcined sample (15 g) was then converted
to its sodium form, i.e., Na, by subjecting for four times to ion-
exchange in 0.5 M NaNO3 aqueous solution (300 ml) at 80 ◦C for
1 h. The solid product powders were filtered, dried overnight at
110 ◦C and calcined at 500 ◦C for 3 h, and were then used to prepare
the Rb+- and Cs+-exchanged samples by ion-exchange with aque-
ous solutions of RbNO3 and CsNO3, respectively. Variation of the
exchange degrees for Rb+ or Cs+ ions was done by adjusting the
concentration of aqueous RbNO3 or CsNO3 (0.005–0.5 M), and rep-
etition times (1–4 times) of the ion-exhange reaction. Typically, ca.
2 g of Na was dispersed and stirred in the RbNO3 or CsNO3 solu-
tion (40 ml) at 80 ◦C for 1 h. The ion exchange reaction was repeated
for 1–4 times to achieve the desirable exchange degrees. At the ter-
mination of the final ion exchange, the sample was filtered, dried
are coded as MxNa1−x (M = Rb+ or Cs+), in which x refers to the frac-
tional exchange degree of M (x = M/[M+Na] (molar)). In particular,
an additional calcination step (500 ◦C, 3 h) should be added after
the second ion-exchange reaction with 0.5 M RbNO3 or CsNO3 for
preparing the Rb1.00 and Cs1.00 samples. Table 1 presents the
different MxNa1−x samples investigated in this study.
The LA conversion, product selectivity and AA yield were calcu-
lated according to the following equations [9,30]:
moles of LA consumed
moles of LA in the feed
LA conversion (%) =
× 100,
Product selectivity (mol% or C%)
moles of carbon atoms in the product defined
moles of carbon atoms in LA consumed
=
× 100,
moles of AA produced
moles of LA in the feed
AA yield (mol% or C%) =
× 100.
3. Results and discussion
3.1. Composition and physicochemical properties
The contents of alkali ions and SiO2/Al2O3 ratio in the MxNa1−x

The two series of samples (RbxNa1−x and CsxNa1−x) prepared
from the parent Na by ion-exchange with the aqueous solutions
of RbNO3 and CsNO3, respectively, are listed in Table 1, together
with their preparation descriptors (including the concentration of
RbNO3 or CsNO3 and repetition times of the ion-exchange), frac-
tional exchange degree (x) of Rb+ or Cs+, SiO2/Al2O3 ratio, and
samples were determined by X-ray fluorescence (XRF) analysis
on a Shimadzu XRF-1800 fluorescence spectrometer. The BET sur-
face areas, pore volumes and average pore diameters were derived
from adsorption–desorption isotherms at −196 ◦C measured on a
Micromeritics ASAP 2010C instrument after outgassing at 200 ◦C
for 5 h. Surface acid–base properties of the samples were probed
by temperature-programmed desorption of NH3 (NH3-TPD) and
CO2 (CO2-TPD) on a Catalyst Analyzer (BEL JAPAN INC.) equipped
with a well-calibrated quadrupole mass spectrometer (Inprocess
Instruments, GAM 200) as the detector [9]. The sample was placed
in a quartz microreactor and pretreated in a flow of 20% O2/Ar
(40 ml/min) at 500 ◦C for 1 h. After cooling the reactor to 100 ◦C,
the reactor was switched to a flow of 2% NH3/Ar or 2% CO2/Ar
(40 ml/min) for NH3 or CO2 adsorption (1 h). Then, the sample was
purged with flowing Ar (40 ml/min) to remove weakly adsorbed
NH3 or CO2 for 1–1.5 h (until the background signals of NH3 and CO2
became stabilized). The reactor temperature was then increased to
600 ◦C at a rate of 10 ◦C/min and the desorption profiles of NH3 and
CO2 were monitored by recording the signals at m/z = 15 and 44,
respectively. The calibration of the mass spectrometer and quan-
tification of the desorbed NH3 and CO2 were documented earlier
[9].
textural properties. The exchange degree of Cs+ in the CsxNa1−x


samples always appeared lower than that of Rb+ in RbxNa1−x
when the ion-exchange reactions were conducted under the sim-
ilar conditions. Compared with the preparation of the KxNa1−x

samples of comparable x in our previous study [30], more severe
conditions for the ion-exchange reactions had to be taken for pro-
ducing the current RbxNa1−x and CsxNa1−x samples. For instance,
the K1.00 sample was obtained by conducting at 80 ◦C the ion-
exchanging reaction of Na in an aqueous solution of 0.5 M KBr
for four times, while the Rb1.00 and Cs1.00 samples could not be
obtained simply by repeating four times the ion-exchange reaction
temperature (500 ◦C) calcination step was added. This is because
that some Na+ ions in the parent Na are located inside the cages
+
˚
with small openings and are not accessible to the larger Rb (1.48 A)
+
+
˚
˚
and Cs (1.69 A) ions but smaller K ions (1.33 A) [31], and the calci-
nation of the partially exchanged samples can cause redistribution
by thermal migration of the alkali ions within the zeolite, making
those Na+ accessible to and replaced by the coexisting Rb+ or Cs+
The gas-phase dehydration of LA was carried out under atmo-
spheric pressure at 360 ◦C in a vertical down-flow fixed-bed tubular
Please cite this article in press as: B. Yan, et al., Sustainable production of acrylic acid: Rb+- and Cs+-exchanged Beta zeolite catalysts for