2
Y. Wang et al. / Journal of Molecular Catalysis A: Chemical 385 (2014) 1–6
100
content of reaction mixture, the addition of auxiliary containing
silicon could effectively inhibit the loss of zeolite in basic circum-
stance and significantly expand the duration of stable performance
of zeolite. Jiang et al. [19] investigated the reaction of cyclohex-
of silicon 50 ∼ 60 g/g from catalyst framework by dissolution. The
continual addition of the auxiliary silicon could effectively inhibit
the dissolve run-off of framework silicon, and expand the lifetime
of catalyst without negative effects on the performance of catalytic
reaction. Choudhary et al. [20] proposed that the poisoning of TS-
1 by ammonia decreased the conversion of styrene epoxidation
reaction.
80
60
40
20
0
Acetone Conv.(%) a
H O Eff.(%) b
2
2
Oxime Sle.(%)
Deactivation time (h) c
For the reaction of acetone ammoximation on TS-1 zeolite,
however, current studies are mainly focusing on the conversion,
selectivity and productivity, while the deactivation of TS-1 catalyst
still remain to be investigated.
Acetone d H O& t-BuOH e
t-BuOH
H O
2
2
Fig. 1. Influence of solvent on the acetone ammoximation. aFor the liquid sampled
after 0.5 h continual reaction; bmolar ratio of acetoxime to H2O2; cthe continual
operation time until acetone conversion was less than 60%; dextra amount of acetone
as solvent; eH2O: t-BuOH = 1:1 (wt).
In this work, the deactivation of HTS catalyst in the reaction of
acetone ammoximation with H2O2 was investigated. Based on the
characterization of fresh and deactivated catalysts by various meth-
ods, and the identification of impurities from both reaction liquid
and deactivated catalyst surface by GC–MS and NMR, the mecha-
nism of HTS deactivation and by-products formation was discussed
in detail.
GC–MS. Some of by-products were identified by Brucker AVANCE
400 NMR using DMSO-d6 as solvent and TMS as internal standard.
2.4. Characterization techniques
2. Experimental
The UV–vis spectra were recorded on a Perkin-Elmer Lambda
950 spectrometer, BaSO4 being used as a reference. Powder X-
ray diffraction patterns were recorded at RT on a Bruker-AXS
D8 Advance diffractometer using nickel-filtered Cu K␣ radiation.
Micromeritic ASAP 2020M system was used to measure nitrogen
adsorption–desorption isotherms of samples.
2.1. Material
Industrial HTS catalyst containing 3.0% ∼ 3.5% Ti was provided
by SINOPEC. Reagents t-BuOH (99%), CH2Cl2 (99.5%), CHCl3 (99%),
HNO3 (65%) and TMPDO (97%) were analytical grade. Ammonia
solution (NH3, 25%), hydrogen peroxide solution (H2O2, 28%) and
acetone (99.5%) were industrial grade.
3.1. Acetone ammoximation
2.2. Operation procedure
Data in Fig. 1 showed the results of acetone ammoximation in
various solvents. One can see that solvent property had signifi-
cant effects on acetone conversion, H2O2 efficiency and catalyst
deactivation, even though all solvents gave high selectivity of prod-
uct oxime (more than 97%). With using t-BuOH as solvent, both
acetone conversion and H2O2 efficiency were high, but the time
causing catalysts deactivation was less than 15 h. With using H2O
as solvent, the time deactivating catalyst became longer although
acetone conversion and H2O2 efficiency decreased a little, indi-
cating the positive effect of water on the maintenance of catalyst
reactivity. In the case of t-BuOH–H2O mixture as solvent, the time
causing deactivation was longer than that of single t-BuOH sol-
vent but shorter than single water solvent. According to solvent
polarity, it was concluded that an increase of solvent polarity could
extend the lifetime of catalyst but decrease the initial conversion
of acetone and the efficiency of H2O2. In addition, with the exist-
ence of excessive acetone as solvent, the H2O2 efficiency was much
lower, the time deactivating catalyst became shorter, and also a
little decrease of oxime selectivity was observed which could be
caused by self-condensation of excessive acetone [21].
After comprehensive consideration to the solvent effects on
reaction process, water was finally chosen as a preferable sol-
vent. Fig. 2 showed the change of acetone conversion in water
solvent with continual operation time. As seen, the conversion
was slowly reduced during 30 h, and then decreased quickly
with further increase of time, suggesting that the loss of cata-
lyst reactivity before 30 h might be caused losing of active sites
whereas the rapid decrease of reactivity after 30 h could be caused
by blocking pores of catalyst. It was interesting to find that
2,3-dimethyl-2,3-dinitrobutane (DMNB) was generated in large
Continuous operation was conducted in a 5 L slurry bed reactor
connecting with a ceramic membrane separator. At the begin-
ning, 300 g HTS zeolite, 1000 g acetone and 2500 g solvents with
the molar ratios of NH3:acetone = 1.17:1 and H2O2:acetone = 1.1:1
were filled into the reactor. During the operation, the pre-
mixed solution with the molar ratios of H2O2:acetone = 1.1:1,
solvent:acetone = 2.5:1 and NH3:acetone = 1.08:1 was continually
sent to the reactor at rate of 25 ∼ 32 mL/min. The operation con-
ditions were 343 ∼ 348 K, 0.1∼0.25 MPa and 70 min resident time.
Once the reactor system reached steady-state in a half hour, prod-
ucts were sampled at different operation times for GC analysis.
Reaction solution was removed by ceramic membrane separation
when reaction was finished. The used catalyst was boiled in deion-
ized water at 348 K for 1 h, and then collected for characterization.
The boiled liquid was extracted by CH2Cl2 for determination of
by-products by GC–MS.
Batch operation was carried out to verify the influence of by-
products on reaction. 35 g HTS catalyst, 105 g deionized water,
1.81 mol acetone and 1.45 mol ammonia were added into a 500 mL
flask in a hot water bath at 343 K. Then, 1.32 mol of ammonia and
1.54 mol of H2O2 were dropped into the flask with stirring in 1 h.
After another half hour reaction, the catalyst was separated by cen-
trifugation and the liquid solution was collected for GC analysis.
2.3. Samples analysis
The samples of reactants and products were analyzed by Agilent
6890N GC. By-products were identified by Agilent 6890N/5937N