22
A. Yamaguchi et al. / Catalysis Communications 69 (2015) 20–24
dinitrogen, carbon monoxide, and methane) and a PorapakTypeQ col-
umn (for carbon dioxide, methane, ethane, and ethylene) with a ther-
mal conductivity detector and using a SP1700 column (for methane,
propane, propylene, n-butane, i-butane, n-pentane, and n-hexane) and
a Gaskuropack54 column (for methane, n-heptane, benzene, toluene,
and xylene) with a flame ionization detector. Conversion of n-hexane
and product yield based on carbon were defined as given below.
ꢀ
ꢁ
mol of unreacted n−hexane in products
mol of reactant n−hexane
ꢀ 100
Conversion ð%Þ ¼ 1−
ð1Þ
ð2Þ
Yield based on carbon ðC %Þ
ꢀ
ꢁ
mol of carbon atom in product
¼
ꢀ 100
mol of carbon atom in reactant n−hexane
Fig. 2. TG-DTA of P-ZSM-5 catalysts after the 2nd run (24 h of time on stream) of n-hexane
Catalytic steam cracking of n-hexane was carried out for 24 h of time
on stream (1st run). Then, the P-ZSM-5 catalyst was calcined under
O2(20%)/He(80%) flow at 893 K for 5 h to remove the coke deposition
and the calcined catalyst was used for the n-hexane cracking again for
24 h of time on stream (2nd run).
cracking at 923 K. The reaction and calcination conditions are the same as Fig. 1.
To understand the effect of S/nC6 ratio on the coke deposition, we
carried out TG-DTA measurement for the P-ZSM-5 catalysts after the
2nd run of n-hexane cracking with different S/nC6 ratios (Fig. 2). The
exothermic peak around 880 K and the weight loss was observed,
which were attributed to the combustion of coke deposited on the P-
ZSM-5 catalysts during the 2nd run of n-hexane cracking for 24 h [13].
The amount of coke deposition increased with decreasing the S/nC6
ratio (Fig. 3), which was consistent with the results that the catalyst
deactivated by the coke deposition at the low S/nC6 ratios. In the case
of thermal cracking of naphtha, the coking decreases with increasing
the steam ratio to naphtha from simulation results [25] and experimen-
tal results [26]. On the other hand, in the case of catalytic cracking, the
effect of steam ratio to naphtha on the amount of coke has not been in-
vestigated. The effect of steam treatment on sweeping of the deposited
coke was investigated on H-ZSM-5 [27]. The amount of deposited coke
was reduced by the steam treatment at 823 K and the temperature for
the combustion of the coke which remained on H-ZSM-5 was raised
after the steam treatment, indicating that the coke with low H/C ratio
remained [27]. In this study, the amount of the deposited coke with
the high S/nC6 ratio was smaller and the coke combustion started
from a little higher temperature (Fig. 2), which was consistent with
the sweeping effect of steam [27]. The mechanism of coke formation
on H-ZSM-5 catalysts during the catalytic cracking of naphtha has
been reported. At first, light olefins such as propylene are formed from
naphtha and the sequential reactions of the light olefins produce BTX
2.3. Characterization
Thermogravimetric and differential thermal analysis (TG-DTA) was
carried out for the catalysts after the reaction of 2nd run on a TG-
DTA2100SA thermal analyzer (Bruker AXS, Japan) at a heating rate of
10 K min−1 in flowing dry air.
3. Results and discussion
Fig. 1 shows the n-hexane conversion during the catalytic cracking
with different S/nC6 ratios at 923 K over the P-ZSM-5 catalyst. The con-
version of n-hexane with 1.0 of S/nC6 ratio over the P-ZSM-5 catalyst
decreased from 90 to 60% for 24 h of time on stream (Fig. 1(a)), consis-
tent with the previous paper [21]. The P-ZSM-5 catalyst was calcined
under O2(20%)/He(80%) flow at 893 K for 5 h to remove the coke depo-
sition after the 1st run of 24 h and the calcined catalyst was used for the
n-hexane cracking again. Most of deposited coke could be burned by the
catalyst calcination at 893 K, which was higher than the reported tem-
perature (ca. 773–823 K) of the calcination for the regeneration of the
coke-deposited ZSM-5 catalysts [22–24]. The initial conversion of the
2nd run with 1.0 of S/nC6 ratio was 84%, which was higher than the con-
version measured at the last of the 1st run (60%) and was lower than the
initial conversion of the 1st run (90%), indicating that the P-ZSM-5
catalyst deactivated because of both the coke deposition and the
dealumination of ZSM-5. We previously reported that the deactivation
caused by the coke deposition could be recovered by the calcination at
893 K and that the irreversible deactivation was caused by the
dealumination [13]. In the case of 0.5 of S/nC6 ratio (Fig. 1(b)), the con-
version of n-hexane over the P-ZSM-5 catalyst decreased from 98 to 89%
for 24 h of time on stream (Fig. 1(b)), where the slope of deactivation
was smaller than that in the case of 1.0 of S/nC6 ratio. The initial conver-
sion of the 2nd run with 0.5 of S/nC6 ratio was 91%, which was a little
higher than the conversion measured at the last of the 1st run (89%)
and was lower than the initial conversion of the 1st run (98%), indicat-
ing that the deactivation was also caused by both the coke deposition
and the dealumination of ZSM-5. On the other hand, the initial conver-
sions of both the 1st and the 2nd runs were 100% in the cases of 0.2, 0.1,
and 0 of S/nC6 ratios (Fig. 1 (c, d, and e)), indicating that the deactiva-
tion was reversible under these conditions. The P-ZSM-5 catalyst was
not seriously deactivated by the dealumination irreversibly and mainly
deactivated reversibly by the coke deposition. Interestingly, among the
conditions for 0.2, 0.1, and 0 of S/nC6 ratios, the final conversions with
less steam showed lower value.
Fig. 3. Amount of coke deposition over the P-ZSM-5 catalysts after the 2nd run (24 h of
time on stream) of n-hexane cracking at 923 K. The reaction and calcination conditions
are the same as Fig. 1.