Inter-conversion of light olefins on ZSM-5 in catalytic naphtha cracking condition

Article abstract of DOI:10.1016/j.cattod.2013.09.060

The inter-conversion of light olefins over four types of HZSM-5 based catalysts under cracking conditions was investigated systematically and various methods including XRD, Ar adsorption-desorption, NH₃-TPD, ²⁷Al and ³¹P MAS-NMR were used to characterize the effects of P modification and steaming on ZSM-5. Regardless the types of catalyst, the same behaviors of light olefin inter-conversion were observed only depending on conversion of light olefins. Also, the conversion and selectivity were not influenced by the presence of hydrogen, suggesting that light paraffins were mainly produced from hydrogen transfer during cracking rather than hydrogenation of light olefins. It can be suggested that the inter-conversion of light olefins occurs through oligomerization of light olefins and then re-cracking of the oligomerized products. To guarantee high light olefin yield in catalytic naphtha cracking, it is strongly required to suppress oligomerization of light olefins during catalytic cracking.

Full text of DOI:10.1016/j.cattod.2013.09.060

Catalysis Today 226 (2014) 52–66
Contents lists available at ScienceDirect
Catalysis Today
journal homepage: www.elsevier.com/locate/cattod
Inter-conversion of light olefins on ZSM-5 in catalytic naphtha
cracking condition
Dan Liua,b, Won Choon Choia, Na Young Kanga, You Jin Leea, Hun Soo Parkc,
Chae-Ho Shind, Yong-Ki Parka,∗
a Green Chemistry Division, Korea Research Institute of Chemical Technology, Daejeon 305-343, Republic of Korea
b School of Environmental and Chemical Engineering, Tianjin Polytechnic University, Tianjin 300387, PR China
c Heesung Catalyst Corp., Chung-ku, Seoul, Republic of Korea
d Department of Chemical Engineering, Chungbuk National University, Chungbuk 361-763, Republic of Korea
a r t i c l e i n f o
a b s t r a c t
Article history:
Received 17 June 2013
Received in revised form 26 August 2013
Accepted 28 September 2013
Available online 29 October 2013
The inter-conversion of light olefins over four types of HZSM-5 based catalysts under cracking conditions
was investigated systematically and various methods including XRD, Ar adsorption–desorption, NH3-
TPD, 27Al and 31P MAS–NMR were used to characterize the effects of P modification and steaming on
ZSM-5. Regardless the types of catalyst, the same behaviors of light olefin inter-conversion were observed
only depending on conversion of light olefins. Also, the conversion and selectivity were not influenced by
the presence of hydrogen, suggesting that light paraffins were mainly produced from hydrogen transfer
during cracking rather than hydrogenation of light olefins. It can be suggested that the inter-conversion
of light olefins occurs through oligomerization of light olefins and then re-cracking of the oligomerized
products. To guarantee high light olefin yield in catalytic naphtha cracking, it is strongly required to
suppress oligomerization of light olefins during catalytic cracking.
Keywords:
Inter-conversion
Ethene
Propene
HZSM-5
© 2013 Elsevier B.V. All rights reserved.
Cracking
Oligomerization
1. Introduction
hydrothermal stability and low product selectivity. Because the
ZSM-5 has intrinsically high activities for the side reactions such
Together with BTX (benzene, toluene, and xylene), light olefins
such as ethene and propene are important as chemical feed stocks
in chemical industry. More than half of light olefins are still pro-
duced by thermal cracking of naphtha and a lot of energy is used
for their production due to high cracking temperature. Also, due
to higher demand for propene than ethene, it is required to pro-
duce more propene in naphtha cracking and catalytic cracking is
considered as one of good alternative to replace conventional ther-
mal cracking [1–4]. Recently, catalytic naphtha cracking process,
applying ZSM-5 based-catalyst and having the name of ACOTM was
developed by SK Innovation co., KBR co., and KRICT in commer-
cial scale. This catalytic cracking process revealed improved light
olefin yield with less energy consumption than conventional ther-
mal cracking process due to the use of acid cracking catalyst and
reduced cracking temperature [1]. To guarantee high light olefin
yield, especially propene yield, in catalytic naphtha cracking there
is no option in selecting catalyst except for ZSM-5. However, the
ZSM-5 is still suffered from commercialization due to the lack of
as hydrogen transfer, oligomerization and aromatization together
with cracking activity, the cracking reaction should be carried out in
diluted steam condition at high temperature to prevent side reac-
tions. Therefore high hydrothermal stability of ZSM-5 is required.
As an effective modifier to improve hydrothermal stability of ZSM-
5, various types of phosphate have been suggested and widely used
for the preparation of commercial cracking catalyst [5].
According to our recent study, the light olefins such as ethene
and propene produced in catalytic naphtha cracking over ZSM-5
are unstable and considerable amount of them is inter-converted
each other through dimerization and re-cracking process. That is,
the inter-conversion rate of ethene and propene is relatively high
and comparable to that of cracking, which resulted in the transfor-
mation of produced light olefins to other light olefins and saturated
hydrocarbons such as methane, ethane and propane. Therefore, it
was investigated more systematically the inter-conversion behav-
ior of light olefins on ZSM-5 in catalytic naphtha cracking condition
and tried to estimate yield pattern of catalytic naphtha cracking
based on the behavior of light olefin inter-conversion.
Already, it is known that oligomerization and isomerization
of light olefins occur very fast together with cracking over acidic
zeolite catalyst, which leads to the formation olefinic products
∗
Corresponding author. Tel.: +82 42 860 7672; fax: +82 42 860 7590.
E-mail address: ykpark@krict.re.kr (Y.-K. Park).
0920-5861/$ – see front matter © 2013 Elsevier B.V. All rights reserved.
http://dx.doi.org/10.1016/j.cattod.2013.09.060

D. Liu et al. / Catalysis Today 226 (2014) 52–66
53
with different carbon numbers. Also, irreversible cyclization and
hydrogen transfer reactions are responsible for the formation of
cycloolefins, alkyl aromatics and paraffins. However, most of the
previous studies are focused on the transformation of light olefins
for the production of aromatics or heavy hydrocarbon fuel in
the low temperature range of 200–450 ◦C [6–16]. Therefore, it
is not easy to estimate the behavior of light olefins at the reac-
tion temperature higher than 650 ◦C. According to Garwood et al.,
the shape-selective oligomerization of C2–C10 olefins occurs more
favorably at the temperature lower than 400 ◦C over ZSM-5, which
leads to the formation of aromatics [6]. Quann et al. also reported
that distillate-range olefins with high quality fuel properties are
produced through oligomerization from C3 to C6 olefins over
ZSM-5 at 200–300 ◦C and relatively high pressure of 30–100 bar
and lower molecular weight olefins are produced at higher tem-
perature and lower pressure due to high cracking rates of olefins
[8]. As the temperature increased, the consumption of ethene and
propene decreased, whereas that of butene increased [9]. The activ-
ity of oligomerization of light olefins is also strongly influenced by
the acidity of zeolite and reaction temperature. The light olefins
such as ethene, propene and i-butene are transformed more eas-
ily on more acidic zeolite [10]. According to Berg et al., the ethene
oligomerization occurs only on strong Brönsted acid sites at low
temperature of 27 ◦C but weak acid sites become active at high
temperature. The formed oligomers begin to crack from 127 ◦C and
the rate of cracking becomes faster than that of oligomerization at
T > 227 ◦C [11].
(iv) St-P/HZSM-5: the P/HZSM-5 was steamed at the same condi-
tion as that of St-HZSM-5.
2.2. Characterization of catalyst
2.2.1. XRD analysis
The XRD analysis was carried out on Bruker D5000 oper-
ated at 50 kV and 30 mA, with Cu K␣ monochromatized radiation
(ꢀ = 0.15468 nm).
2.2.2. Pore analysis
The specific surface area, pore volume were measured by Ar
adsorption–desorption isotherms (ASAP 2010, Micromeritics Ins.
Corp.). Prior to the measurement, all samples were degassed at
250 ◦C until a stable vacuum of ca. 10−3 Torr was reached.
2.2.3. Pyridine adsorption
FTIR spectra were taken at room temperature in the range of
400–4000 cm−1 using an AVATAR 360 Nicolet spectrophotome-
ter in a vacuum cell. Prior to pyridine adsorption, the samples
were degassed under vacuum (10−3 Pa) at 400 ◦C for 24 h. Then
the samples were exposed with pyridine for 30 min and degassed
for 1 h under vacuum at increasing temperatures (150, 250 and
350 ◦C), followed by IR measurement. The relative ratio of Brön-
sted and Lewis acidity was derived from the area of the IR bands at
1550 cm−1 and 1450 cm−1, respectively.
According to our study, the behavior of light olefins on ZSM-5 at
high temperature is much different from that at low temperature
and it is important to interpret the behavior of catalytic naphtha
cracking with consideration of inter-conversion of light olefins.
That is, in catalytic naphtha cracking, the produced light olefins
are unstable and considerable amount of them are transformed
irreversibly to other side products through oligomerization and re-
cracking processes. To suppress the side reactions and maximize
light olefins yield in catalytic naphtha cracking, understanding the
behavior of produced light olefins is quite importance.
However, there has been no detailed study for the behavior of
light olefins in catalytic naphtha cracking condition over ZSM-5
based catalysts. Therefore, the reactivity of ethene and propene
was investigated thoroughly depending on the types of catalyst in
catalytic naphtha cracking condition and it was tried to suggest a
generalized mechanism for light olefins inter-conversion.
2.2.4. Solid MAS NMR
The 27Al MAS NMR spectra were obtained on a Bruker AVANCE
500 spectrometer at a 27Al frequency of 130.325 MHz in 4 mm
rotors at a spinning rate 12.0 kHz. The spectra were obtained with
an acquisition of ca. 3000 pulse transient, which were reported with
a ꢁ/4 rad pulse length of 5.00 ␮s and a recycle delay of 1.0 s. The
27Al chemical shifts were referenced to Al(H2O)6 solution. The
3+
31P MAS NMR spectra at a spinning rate of 11.0 kHz were obtained
on the same spectrometer at a 31P frequency of 202.450 MHz with
an acquisition of about 100 pulse transients, which were repeated
with a ꢁ/2 rad pulse length of 5.0 ␮s and a recycle delay of 30 s. The
31P chemical shifts were referenced to H3PO4 solution.
2.3. Reactions
The reactivity of light olefins was evaluated in quartz micro-
reactor. 0.05 g catalyst was loaded into the micro-reactor and
reactions were carried out at 675 ◦C in the WHSV range of
4–128 h−1. As a reactant, three types of ethene, propene and
both of them were used and hydrogen was also introduced
together in the case of seeing the effect of hydrogenation. Because
the product volume ratio of C2H4/C3H6/H2 in catalytic naph-
tha cracking is 20/15/5, the reactivity of light olefins were
evaluated in the following reactant volume compositions: (i)
C2H4/He = 20/80, (ii) C2H4/H2/He = 20/5/75, (iii) C3H6/He = 15/85,
(iv) C3H6/H2/He = 15/5/80, (v) C2H4/C3H6/He = 20/15/65, and (vi)
C2H4/C3H6/H2/He = 20/15/5/60, respectively. The products were
analyzed with GC (6890, Agilent) equipped with NP-1 column and
FID detector.
2. Experimental
2.1. Preparation of catalyst
For activity test of light olefins, four types of ZSM-5s were pre-
pared in the following preparation conditions.
(i) HZSM-5: parent HZSM-5 (Si/Al = 14.5) was obtained from Albe-
marle.
(ii) St-HZSM-5: the parent HZSM-5 was steamed at 770 ◦C in
100 vol% steam flow for 24 h. The steam treatment was carried
out in a fixed bed reactor after loading 4 g of catalyst while flow-
ing evaporated water at WHSV of 10 h−1. The temperature was
programmed from 70 to 770 ◦C for 70 min, maintained at 770 ◦C
for 24 h, and then cooled down quickly to room temperature
after steaming.
3. Results and discussion
3.1. Catalyst characterization
(iii) P/HZSM-5: the parent HZSM-5 was modified with phosphoric
acid by impregnation method. The amount of P corresponding
to the Al/P atomic ratio = 1 was loaded on HZSM-5, where Al
means framework Al of ZSM-5. After impregnation of P, it was
dried at 100 ◦C for 24 h and then calcined at 650 ◦C for 6 h.
3.1.1. XRD analysis
All of P-free and P-modified HZSM-5s revealed typical XRD
patterns of MFI structure with no other extra peaks regardless
of steaming at 770 ◦C for 24 h. However, in the case of HZSM-5,

54
D. Liu et al. / Catalysis Today 226 (2014) 52–66
total pore volume (Vtotal). However, after P modification (P/HZSM-
5), SBET, micropore area (Smicro), Sext, Vmicro and Vtotal decreased,
suggesting the formation of polymeric phosphate species located
in inside of zeolite pore channels or over the external surface of
zeolite and blocks the channel. If the P/HZSM-5 was steamed, the
SBET, Smicro, Vmicro and Vtotal increased compared with P/HZSM-5
due to the removal of free phosphorus species and the hydroly-
sis of polymeric phosphate species during steaming. According to
our previous study, hydrolysis and re-polymerization of polymeric
phosphate species occurs simultaneously by steam and the phos-
phate species are reassembled in the P modified catalyst [16], which
finally leads to the increase of surface area and pore volume.
(d) St-P/HZSM-5
(c) P/HZSM-5
(b) St-HZSM-5
3.1.3. Pyridine adsorption
(a) HZSM-5
The acid property of samples was monitored by FT-IR after
pyridine adsorption and then outgassing at different tempera-
tures (Table 1). The increase of outgassing temperature leads to
the removal of weakly coordinated pyridine and only pyridine
adsorbed on strong acid sites exists. The spectra outgassed at 150,
250 and 350 ◦C reveals information of total, medium + strong, and
strong acid sites of the samples, respectively. Parent HZSM-5 shows
the strongest acid strength, and 96% of total Brönsted acid sites and
1/3 of total Lewis acid sites appear as strong ones. Meanwhile, total
Lewis acid sites of HZSM-5 are more than total Brönsted acid sites,
which agrees well with the previous result that a large fraction of
acid sites would be transformed to Lewis sites by calcination at
650 ◦C [18]. If HZSM-5 was steamed at 770 ◦C for 24 h, nearly all
the Brönsted acid sites disappeared due to heavy dealumination
by severe hydrothermal treatment, and half of its total Lewis acid
sites was lost and only small amount remain as strong sites. The
impregnation of phosphate (P/HZSM-5) also brought an obvious
decrease of both strong Brönsted and Lewis acid sites. Especially,
the strong Brönsted acid sites decreased preferentially due to the
neutralization by phosphate species, and the medium and weak
Brönsted acid sites increased caused by the increased P OH group.
The steaming of P/HZSM-5 brought further decrease of Brönsted
acid sites, different from steamed HZSM-5 (St-HZSM-5), about 47%
of total Brönsted acid sites still remained even after steaming. This
means that some improvement in hydrothermal stability could be
found by phosphorus modification, which is caused by preventing
dealumination from framework [16]. As we know that the crack-
ing capability is directly related to the Brönsted acid sites, their
relative quantities were observed to decrease as follows: HZSM-
5 > P/HZSM-5 > St-P/HZSM-5 > St-HZSM-5.
5
10
15
20
25
30
35
40
45
50
2 Theta (degree)
Fig. 1. XRD Patterns of (a) HZSM-5, (b) St-HZSM-5, (c) P/HZSM-5 and (d) St-P/HZSM-
5.
slightly decrease of intensity at 2ꢂ of 24◦ was observed and rela-
tive crystallinity was decreased to 96% due to the dealumination
of framework Al by steaming. Different from HZSM-5, P-modified
HZSM-5 showed no change in XRD peak intensity and relative crys-
tallinity even after steaming (Fig. 1). These results indicate that
the main framework of HZSM-5 is scarcely damaged by steaming
and P modification. Also, no XRD peaks corresponding to the extra
framework alumina or AlPO4 were observed over the steamed St-
P/HZSM-5. However, according to Ramesh et al., it would not be
ruled out the formation of ultrafine particles of extra framework
˚
alumina and AlPO4 lower than 40 A in particle size on the surface
of zeolite crystals [17].
3.1.2. Ar adsorption–desorption
Textural properties of four different types of ZSM-5s are pre-
sented in Table 1. All the samples except St-HZSM-5 revealed the
adsorption–desorption isotherms of type I (Fig. 2). In the case of
St-HZSM-5, an adsorption–desorption hysteresis loop can be seen,
which can be attributed to the formation of mesopores due to
heavy dealumination of framework Al of HZSM-5 by steaming at
770 ◦C for 24 h. Therefore, the steaming treatment of parent HZSM-
5 resulted in a decrease of both micropore volume (Vmicro) and
surface area (SBET) but an increase of external surface area (Sext) and
3.1.4. Solid-state MAS NMR
Various types of Al and P species were observed in NMR
depending on treatment conditions (Fig. 3). Before steaming, most
of Al exists as tetrahedral-coordinated Al at 53 ppm, Altetra in
zeolite framework but the framework Al comes out from the
framework of zeolite by steaming and the intensity of Altetra at
53 ppm decreased almost to zero regardless of phosphate mod-
ification (Fig. 3(A)). However, the degree of dealumination by
steaming is more severe in phosphate-free HZSM-5 (St-HZSM-5)
than in phosphate-modified one (St-P/HZSM-5). If the phosphate-
free HZSM-5 is steamed (St-HZSM-5), most of the tetrahedral
Altetra species at 53 ppm transformed to other various Al species
and Al MNR spectra of broad and low intensity were observed.
However, different from St-HZSM-5, new Al peak with high
intensity at 40 ppm assigned to distorted framework Al with
either tetra- or penta-coordination was observed in the steamed
phosphate-modified HZSM-5 (St-P/HZSM-5) [19]. These distorted
Al species are formed through the interaction of framework Al
with introduced phosphate and they are thought to prevent dea-
lumination of framework Al and enhance hydrothermal stability
of ZSM-5. This NMR result is well coincided with that of pyridine
200
160
120
80
(a) HZSM-5
(b) St-HZSM-5
40
(c) P/HZSM-5
(d) St-P/HZSM-5
0
0.0
0.2
0.4
0.6
0.8
1.0
Relative pressure (p/po)
Fig. 2. Adsorption–desorption isotherms of Argon: (a) HZSM-5, (b) St-HZSM-5, (c)
P/HZSM-5 and (d) St-P/HZSM-5.

D. Liu et al. / Catalysis Today 226 (2014) 52–66
55
Table 1
Textural and acid properties of HZSM-5 and phosphate modified HZSM-5.
Catalysts
SBET(m2/g)
Micropore area
(m2/g)
External surface
areas (m2/g)
Pore volume (cm3 g−1
)
Lewis acid site
Brönsted acid site
Micro
Meso
Total
150 ◦
C
250 ◦
C
350 ◦
C
150 ◦
C
250 ◦
C
350 ◦
C
HZSM-5
447
408
271
357
432
379
268
357
15.3
29.1
3.36
0.05
0.15
0.13
0.10
0.13
–
0.05
–
0.17
0.18
0.12
0.13
109
58
70
51
13
11
18
37
5
9
100
0
78
47
97
0
44
32
96
0
31
19
St-HZSM-5
P/HZSM-5
St-P/HZSM-5
–
81
3
adsorption that the relative amount of strong Brönsted acid sites
of four catalyst is HZSM-5 > P/HZSM-5 > St-P/HZSM-5 > St-HZSM-
5. Also, over the all phosphate-modified HZSM-5s (P/HZSM-5
and St-P/HZSM-5), new Al species at −11 ppm assigned to extra-
framework AlPO4 appeared with the disappearance of octahedral
Al species at 0 ppm assigned to extra-framework Al2O3. Therefore,
this extra-framework AlPO4 is thought to be formed through the
interaction of non-framework octahedral Al2O3 with introduced
phosphate.
This means that P species become more condensed after steaming
[7,20–22].
From the results of characterizations, the followings could be
suggested. Even the HZSM-5 is unmodified with phosphate, it
maintains MFI structure after steaming but loses all of its acidity
and mesopore is formed due to the dealumination of framework
Al. If the HZSM-5 is modified with phosphate, some of its acidity
is maintained even after steaming and hydrothermal stability is
enhanced due to prevention of dealumination.
Over the phosphate-modified HZSM-5s, many kinds of P species
were observed in 31P MAS NMR spectra (Fig. 3(B)). After steaming,
low field peaks at 0.9, −6, −16 and −25 ppm decreased dramati-
cally or even disappeared and higher field peaks in the region of
−32 to −46 ppm assigned to polyphosphates became dominant.
3.2. The behavior of light olefins inter-conversion in catalytic
naphtha cracking condition
Typically, catalytic cracking of naphtha is carried out in the
reaction conditions of 650–700 ◦C and WHSV of 16–32 h−1 over
phosphate-modified HZSM-5 and, as mentioned in Section 2.3, light
olefins of about 20 vol% ethene and 15 vol% propene are produced
together with 5 vol% H2. Therefore, the behavior of light olefins
inter-conversion was monitored in this product composition and
reaction condition with and without catalyst (Table 2).
40
A
-11
53
The light olefins such as ethene and propene produced in cat-
alytic naphtha cracking are quite reactive and about 22.0% and
55.0% of ethene and propene are transformed selectively to other
types of light olefins even at high WHSV of 32 h−1 in the presence of
catalyst through inter-conversion. That is, the selectivity of propene
is 62.9 mol% from ethene and 67.1 mol% of ethene selectivity was
obtained when feeding propene. However, if the reaction is carried
out thermally without catalyst, the light olefins are quite stable
and the conversion of ethene and propene were only 2.3% and 5.8%
could be obtained.
(d) St-P/HZSM-5
36
(c) P/HZSM-5
(b) St-HZSM-5
3.3
-1.5
Also, quite different product distributions are obtained in the
absence/presence of catalyst. In the presence of catalyst, other
type of light olefin was produced preferentially with only small
fraction of saturated hydrocarbons such as methane, ethane, and
propane; i.e. propene was produced mainly when ethane was
fed and ethene was the main product of propene reaction. How-
ever, in thermal cracking condition without catalyst, hydrogenation
reaction occurred dominantly and saturated hydrocarbons were
produced preferentially together with considerable amount of
higher hydrocarbons (C4–C6 fraction). The different reaction path-
ways in thermal and catalytic conditions could be seen clearly from
the product distributions depending on the type of reactants.
In thermal cracking condition, if ethene was used as a reactant,
ethane, propene and C4s would be produced selectively without
any methane and propane, but if propene as the feed, methane,
ethene, propane and C4 and small fraction of C6 were produced
without any ethane. This means that saturated hydrocarbons are
formed directly from olefin through hydrogenation reaction under
thermal cracking conditions. Another thing to be considered in the
product distribution is that higher hydrocarbons such as C4 and
C6 can be formed also from ethene and propene through dimeriza-
tion reaction even though its rate is not so high in thermal reaction
condition; C4 and C6 are formed preferentially by the dimeriza-
tion of ethene and propene, respectively. However, because the
formed C6 are more reactive than C4 , more complicated product
(a) HZSM-5
-20
100
80
60
40
20
0
-40
27
Al
δ
(ppm)
-32
-39
B
-46
-6
-25
-16
(d) St-P/HZSM-5
(c) P/HZSM-5
0.9
40
20
0
-20
-40
-60
-80
31
P δ (ppm)
Fig. 3. (A) 27Al MAS NMR spectra and (B) 31P MAS NMR spectra of (a) HZSM-5, (b)
St-HZSM-5, (c) P/HZSM-5 and (d) St-P/HZSM-5.

56
D. Liu et al. / Catalysis Today 226 (2014) 52–66
Table 2
The degree of inter-conversion of olefins in thermal and catalytic cracking conditions.a
Reaction condition
Reactant
Conv. (%)
Product selectivity (mole %)
CH4
C2H4
C2H6
C3H6
C3H8
C4
C5
C6
BTX
c
C2H4
C3H6
C2H4
C3H6
2.3
5.8
22.0
55.0
0
–
25.9
–
53.0
0
3.4
1.5
12.3
–
62.9
–
0
34.6
40.1
11.8
10.1
0
0
1.2
0.6
0
0
Thermal cracking
d
11.0
10.3
8.5
18.6
1.5
5.0
3.7
0.3
0.2
0.7
8.5
6.9
c
Catalytic crackingb
d
67.1
a
Reactions were carried out at 675 ◦C and WHSV of 32 h−1
HZSM-5 as the catalyst.
The feed is C2H4/He/H2 = 20/75/5 (vol. ratio).
The feed is C3H6/He/H2 = 15/80/5 (vol. ratio).
.
b
c
d
distribution is obtained in the reaction of propene due to the re-
cracking of formed C6 . That is, the formed C6 by dimerization of
propene are re-cracked, which leads to only small fraction of C6 in
propene reaction. It was found that the produced ethene and C4 is
not equal in moles in the reaction of C3 although the dimerization
of propene to C6 and then re-cracking to C2 and C4 are dominant
reaction. It is thought that this is due to the complicated side reac-
tions including dimerization, hydrogen transfer, cyclization and
aromatization. So, equal mole of C2 and C4 can only be formed
without these side reactions. In the real condition, the dimerization
of C2 –C4 occurs quickly together with main reaction, however,
C4 is difficult to crack directly to C2 under monomolecular crack-
ing mechanism, which probably lead to higher selectivity to C4
100
80
60
40
20
0
than that of C2
.
The different pattern of methane formation in the reaction of
ethene and propene provides good information about the intrinsic
stability of light olefins. In the absence of catalyst under the cracking
conditions, ethene and its dimer C4 is stable enough not to crack
further to methane; but propene can dimerized into C6 , which is
most likely to recrack into methane (11.0 mole%) and other light
hydrocarbons. This means that methane is formed not by cracking
of ethene and its dimer, but by further cracking of C6 , the dimer
of propene.
Different from thermal reaction, more complicated product dis-
tribution was obtained in the catalytic reaction (Table 2). The
inter-conversion rate of light olefins in the presence of catalyst
is very high and even comparable to that of naphtha cracking.
That is, considerable amount of ethene is inter-converted selec-
tively to propene (62.9 mol%) and propene to ethene (67.1 mole%)
in catalytic naphtha cracking condition through dimerization → re-
cracking process. Therefore, to estimate product distribution more
correctly in catalytic naphtha cracking, it is required to understand
the behavior of light olefins inter-conversion in catalytic naph-
tha cracking condition. Therefore the behavior of light olefins was
investigated more thoroughly depending on the reaction condition
and the type of catalysts.
400
450
500
550
600
650
700
Temperature (oC)
Fig. 4. Conversion of (ꢀ) C2H4, ( ) C3H6, ( ) n-C4H10 in their corresponding reaction
and conversion of ( ) C2H4 and ( ) C3H6 in the reaction of C2H4 and C3H6 mixture
depending on temperature over H-ZSM-5 (T: 675 ◦C, WHSV: 32 h−1, reactant com-
position (vol%): C2H4/H2/He = 20/5/75, C3H6/H2/He = 15/5/80, n-C4H10/He = 10/90,
C2H4/C3H6/H2/He = 20/15/5/60).
40
30
20
10
0
40
30
20
10
0
C2H4+ H2
C2H4
C3H6+ H2
C3H6
3.3. The stability of light olefins
To estimate the degree of light olefins inter-conversion in
catalytic naphtha cracking condition, the inter-conversion rates of
ethene and propene were compared with that of n-C4H10 crack-
ing over HZSM-5 catalyst while changing reaction temperature
(Fig. 4). The inter-conversion of light olefins and n-C4H10 cracking
showed opposite dependence on the reaction temperature. As
temperature increased, the rate of cracking increased and the
higher conversion of n-C4H10 was obtained (Fig. 4c). However,
the inter-conversion rate of light olefins deceased and the lower
conversion was obtained. Only 5% of n-C4H10 was converted at
500 ◦C and increased to 80% as the temperature increased to 700 ◦C
(Fig. 4(c)). Different from n-C4H10, however, conversion of light
olefins decreases with temperature (Fig. 4(a and b)). The reason for
0
32
64
96
128
WHSV(h-1)
Fig. 5. The effect of hydrogen on the conversion of ethene and propene; (•) C2H4with
H2, (ꢀ) C2H4 without H2, ( ) C3H6with H2, and ( ) C3H6without H2 (cata-
lyst: St-P/HZSM-5, T: 675 ◦C, reactant composition (vol%): C2H4/H2/He = 20/5/75,
C2H4/He = 20/80, C3H6/H2/He = 15/5/80, C3H6/He = 15/85).

D. Liu et al. / Catalysis Today 226 (2014) 52–66
57
(a) C2H4+ H2
(b) C2H4
100
80
60
40
20
0
BTX
C s
6
C s
5
C s
4
C H
3
8
C H
3
6
C H
2
6
C H
2
4
CH
4
4
8
16
32
4
8
16
32
(c) C3H6+ H2
(d) C3H6
100
80
60
40
20
0
BTX
C s
6
C s
5
C s
4
C H
3
8
C H
3
6
C H
2
6
C H
2
4
CH
4
4
16
32
128
4
16
32
128
WHSV (h-1)
WHSV (h-1)
Fig. 6. The effect of hydrogen on selectivity in ethene and propene reaction; (a) C2H4 with H2, (b) C2H4 without H2, (c) C3H6 with H2, and (d) C3H6 without H2 (catalyst = St-
P/HZSM-5, T = 675 ◦C, reactant composition (vol%): C2H4/H2/He = 20/5/75, C2H4/H2/He = 20/80, C3H6/H2/He = 15/5/80, C3H6/He = 15/85).
the decrease of conversion of light olefins at high temperature can
be explained by the increased rate of cracking compared with that
of polymerization. That is, at low temperature the rate of polymer-
ization is higher than that of cracking, so that the conversion of
light olefin is high. However, relatively low conversion of olefins
is obtained at high temperature due to the increased cracking
activity. Also, quite different reactivity was observed depending
on the type of light olefins. The conversion of propene decreased
gradually from 70% to 55% as the temperature increased from
400 ◦C to 700 ◦C (Fig. 4(b)). However, the conversion of ethene
decreased more rapidly from 49% to 23% as the temperature
increased (Fig. 4(a)). This means that the conversion of propene is
higher than that of ethene in catalytic naphtha cracking condition.
For more precise comparison of inter-conversion rate of ethene
and propene, the reaction was carried out using a mixture of 20 vol%
ethene and 15 vol% propene (Fig. 4(d and e)). The ethene and
propene revealed big difference in reactivity. At lower temperature
below 500 ◦C, almost the same conversions of ethene and propene
were obtained but higher conversion of propene was obtained than
ethene at high temperature. The conversion of ethene and propene
was around 50% and 50% at 500 ◦C, respectively but it decreased to
3% and 43% at 700 ◦C. That is, ethene is more stable than propene
at high temperature. This means that the ethene exists as it is once
it is formed in catalytic naphtha cracking but the formed propene
transforms more easily to other products through inter-conversion
reaction. Therefore, it is required to maintain short contact time in

58
D. Liu et al. / Catalysis Today 226 (2014) 52–66
60
40
20
catalytic naphtha cracking to guarantee high light olefin yield by
preventing especially secondary reaction of produced propene.
As can be seen in the patterns of n-C4H10 cracking and light
olefins inter-conversion, because the activities of cracking and
inter-conversion are trade-off each other, it is required to carry
out catalytic naphtha cracking at relatively high reaction temper-
ature to guarantee high light olefin yield by increasing cracking
activity and prohibiting inter-conversion of light olefins. However,
if the cracking temperature is too high, a lot of undesired saturated
hydrocarbons such as methane, ethane and propane will be formed
irreversibly due to the secondary re-cracking of high hydrocar-
bons and direct hydrogenation of light olefins by thermal reaction.
Therefore, the reaction temperature should be optimized in cat-
alytic naphtha cracking to guarantee high light olefin yield and to
minimize saturated hydrocarbon formation.
HZSM-5
(a)
St-HZSM-5
P/HZSM-5
St-P/HZSM-5
0
4
8
16
32
64
128
256
60
HZSM-5
(b)
St-HZSM-5
P/HZSM-5
40
20
0
St-P/HZSM-5
3.4. The effect of hydrogen in inter-conversion of light olefins
Because the economy of process is closely related to the light
olefin yield in catalytic naphtha cracking, it is required to produce
lightolefins such as etheneandpropeneselectively andtominimize
the formation of invaluable saturated light hydrocarbons such as
methane, ethane and propane. Therefore, it is quite meaningful to
figure out where these saturated light hydrocarbons are originated.
Normally, considerable amount of hydrogen is produced (about
5 vol% in product) during catalytic naphtha cracking. There-
fore, the effect of hydrogen on the formation of saturated light
hydrocarbons was monitored. The inter-conversion reaction of
ethene and propene was carried out without hydrogen and with
5 vol% extra hydrogen in the reactant over steamed P/HZSM-5
(Figs. 5 and 6).
The conversions of light olefins and product distribution were
not influenced much by hydrogen, which agrees well with that of 2-
methylpentane cracking reported by Zhao et al. that the presence
of molecular hydrogen would not affect cracking capability [23].
Thissuggeststhatthesaturatehydrocarbons producedweremainly
caused by hydrogen transfer reaction, not by direct hydrogenation
of olefin. Only slight difference exist in the product distribution
of ethene conversion in the absence/presence of hydrogen, that is,
ethane increase slightly in the presence of hydrogen; while almost
same product distribution was obtained in the reaction of propene
even in the presence of hydrogen. This means that small amount
of ethane may be caused by the hydrogenation of ethene (reaction
(1)).
4
8
16
32
64
128 256 512 1024
HZSM-5
(c)
60
40
20
0
60
40
20
0
St-HZSM-5
P/HZSM-5
St-P/HZSM-5
: C2H4
: C3H6
4
8
16
32
64
128 256 512 1024
WHSV(h-1
)
Fig. 7. Conversion of (a) ethene, (b) propene and (c) ethene and propene mixture
depending on the types of catalysts; (ꢀ) HZSM-5, ( ) St-HZSM-5, ( ) P/HZSM-5 and
(
) St-P/HZSM-5 (T = 675 ◦C, reactant composition (vol%): C2H4/H2/He = 20/5/75,
C3H6/H2/He = 15/5/80, C2H4/C3H6/H2/He = 20/15/5/60).
and (4), because the cracking of C6 to two C3 or C2 and C4 pro-
ceeds more easily than C4 , only small fraction of C6 is observed in
the product. Increasing the contact time finally leads to the inter-
conversion of ethene to propene selectively and 62% of propene is
C2 + H2 → C2H6
(1)
Another interesting thing found in the product distribution is
that relatively large amount of oligomers are formed in the reac-
tion of light olefins; C4s are formed preferentially in the reaction
of ethene and some fraction of C6s are formed preferentially in the
reaction of propene (Fig. 6).
obtained at WHSV of 4 h−1
.
C6 → 2C3H6
(3)
(4)
In the reaction of ethene, the fraction of C4s reached to about
80% when the contact time decreased (WHSV of 32 h−1). With the
increase of contact time, conversion of ethene increased, while
selectivity to C4s decreased and propene selectivity increased. If
the propene is formed by hydrocracking of dimerized ethene, i.e.
C4 , through the following reaction (2), equal mole of CH4 should
be formed. However, there is no CH4 in the product when WHSV is
C6 → C2H4 + C4
As mentioned before, because propene is more active than
ethene, more dimer, C6 , is formed at the same contact time. How-
ever, only small amount of C6 was observed in the product even at
short contact time due to the re-cracking to the shorter hydrocar-
bons through the reactions of (3) and (4). Therefore, the same as in
the reaction of ethene, the propene is selectively inter-converted
to ethene at long contact time through series reaction steps of
oligomerization → re-cracking and 66% of ethene was obtained
higher than 16 h−1
.
C4 + H2 → CH4 + C3H6
(2)
Therefore, it is thought that the propene, C3 is formed by re-
cracking of the trimer (C6 ) or even higher oligomers of ethene,
which also indicating that butene is more difficult to follow
monomolecular cracking mechanism. As shown in reactions (3)
selectively at WHSV of 4 h−1
.
From the above results, the following conclusion can be
obtained that propane and ethane are mainly produced by hydro-
gen transfer during cracking of oligomerized hydrocarbons or

D. Liu et al. / Catalysis Today 226 (2014) 52–66
59
30
20
10
80
(a)
(c)
(e)
(b)
HZSM-5
60
St-HZSM-5
P/HZSM-5
St-P/HZSM-5
40
20
0
0
0
0
0
0
10
10
10
20
20
20
20
30
30
30
30
40
40
40
40
0
10
20
30
40
80
5
4
3
2
1
(d)
(f)
60
40
20
0
0
5
0
10
20
30
40
100
80
60
40
20
4
3
2
1
0
0
0
10
20
30
40
2.5
30
(g)
(h)
2.0
1.5
1.0
0.5
0.0
20
10
0
10
0
10
20
30
40
C2H4 Conversion (%)
C2H4 Conversion (%)
Fig. 8. Selectivity of (a) CH4, (b) C2H6, (c) C3H6, (d) C3H8, (e) C4s, (f) C5s, (g) C6s and (h) BTX depending on the conversion in the reaction of ethene over different types of
catalysts; (ꢀ) HZSM-5, ( ) St-HZSM-5, ( ) P/HZSM-5 and ( ) St-P/HZSM-5 (T = 675 ◦C, reactant composition (vol%): C2H4/H2/He = 20/5/75).
naphtha, although small part of ethane is produced from ethene
through direct hydrogenation. And another important thing to be
considered is that the inter-conversion of ethene and propene is
the dominant reaction in catalytic naphtha cracking and should
be considered seriously in interpretation of catalytic naphtha
cracking. With these backgrounds, the following experiments
were carried out to figure out how much the inter-conversion of
light olefin is influenced by the types of catalyst.

60
D. Liu et al. / Catalysis Today 226 (2014) 52–66
3.5. The inter-conversion depending on the types of catalyst
The higher inter-conversion of light olefins means the higher
secondary reaction in catalytic naphtha cracking. Because the
cracking activity is strongly influenced by the types of catalyst
in catalytic naphtha cracking it was investigated whether the
inter-conversion of light olefins is also influenced by the types
of catalyst. The inter-conversion was carried out over four dif-
ferent types of catalyst: HZSM-5, P/HZSM-5, St-P/HZSM-5, and
St-HZSM-5 (Figs. 7 and 8). The inter-conversion activity was
strongly influenced by the types of catalyst as expected. The cat-
alyst having higher naphtha cracking activity revealed the higher
inter-conversion activity for all the reactants of ethene, propene
and mixture of ethene + propene over the full range of contact time.
Different from activity, however, the selectivity was not influ-
enced by the types of catalyst and almost same product distribution
was obtained at the same conversion of ethene, propene and mix-
ture of ethene + propene regardless the types of catalyst.
3.5.1. The inter-conversion activity depending on the types of
catalyst
Fig. 9. Product distribution of C4s depending on the conversion in the
reaction of ethene over HZSM-5 (T = 675 ◦C, reactant composition (vol%):
C2H4/H2/He = 20/5/75).
The catalyst having higher naphtha cracking activity also
revealed the higher inter-conversion activity for all the reac-
tants of ethene, propene and mixture of ethene + propene;
HZSM-5 > P/HZSM-5 > St-P/HZSM-5 > St-HZSM-5 (Fig. 7). Also, the
inter-conversion activity is well correlated with acidity of cata-
lyst as that of naphtha cracking (Table 1). The HZSM-5 having
higher acidity revealed higher inter-conversion activity than that
of P/HZSM-5. The acidity of St-HZSM-5 decreased dramatically and
almost all of Brönsted acid sites disappeared due to the dealumina-
tion of framework aluminum and revealed lowest inter-conversion
activity. However, the hydrothermal stability of P/HZSM-5 catalyst
is improved; and higher inter-conversion activity of St-P/HZSM-5
can be obtained than St-HZSM-5 due to good hydrothermal stabil-
ity after P modification. To sum up, inter-conversion activity of four
types of catalyst reveal can be ordered depending on the acidity:
HZSM-5 > P/HZSM-5 > St-P/HZSM-5 > St-HZSM-5.
Another thing to be considered is that the intrinsic inter-
conversion activity is strongly influenced by the types of olefins.
As Garwood et al. suggested in the oligomerization reaction [6],
the higher conversion was obtained in the oligomerization of
propene than ethene at the same reaction condition. This means
that propene is easier to oligomerize than ethene. However, if
the ethene and propene exist together in the reactant, the inter-
conversion rate will be lower than that of single reactant itself.
Different from 55% propene and 22% ethene conversions from the
single reactant of propene and ethene, 44% propene and 12% ethene
conversions were obtained from the mixture of propene and ethene
over HZSM-5 at WHSV of 32 h−1 (Fig. 7(c)). That is, the conver-
sion of ethene and propene are inhibited to a certain degree by the
inter-conversion of propene and ethene.
At very low conversion of ethene, C4s and ethane were domi-
nantly formed and then decreased rapidly with increasing contact
time. This means that C4s are initially formed by dimerization of
ethene (reaction (5)) and ethane by direct hydrogenation of ethene
with hydrogen (reaction (1)), respectively.
2C2 → C4
(5)
As the conversion increased, propene, C5s, and C6s were begun
to be formed as secondary products and revealed a maximum
at certain conversions. The formation of higher olefinic hydro-
carbons from ethene can be explained by following reactions;
the C6s are formed by further oligomerization of initially formed
C4s with ethene (reaction (6)) and then the formed C6s are re-
cracked to propene (reaction (7)) and the C5s by dimerization of
formed propene with ethene (reaction (8)). However, the frac-
tion of C5s and C6s is quite low (<1%) compared with propene
(>40%) because the re-cracking rate of C5s, and C6s is much higher
than that of oligomerization. That is, through the reaction of
oligomerization → re-cracking, considerable amount of ethene is
inter-converted to propene because all of the H-ZSM-5 catalysts
have relatively high oligomerization activity comparable to that of
cracking.
C4 + C2 → C6
C6 → 2C3
(6)
(7)
(8)
C3 + C2 → C5
Different from these olefinic species, the fraction of methane,
propane, and BTX increased linearly with initial induction time as
the conversion increased. This means these saturated hydrocarbons
and aromatic species are tertiary products and produced irre-
versibly through hydrogen-transfer or aromatization of the formed
olefinic species (reactions (2), (9), (10)).
3.5.2. The inter-conversion selectivity of ethene depending on the
types of catalyst
As mentioned in previous section, the inter-conversion activity
is strongly influenced by the types of catalyst. That is, the con-
version of light olefins is function of catalyst activity and contact
time of reactant with catalyst. The higher inter-conversion of light
olefins is obtained at long contact time or over the fresh catalysts.
However, different from activity, product selectivity was not
influenced by the types of catalyst. If the product distribution is
plotted based on the conversion, quite expectable general trends
in product distribution can be obtained regardless the types of cat-
alyst and contact time (Figs. 8 and 9). That is, quite interestingly,
the same product distribution is obtained at the same conversion
regardless of the types of catalyst and contact time (Fig. 8).
C6 + 3C3 → C6H6 + 3C3
C6 → C6H6 + 3H2
(9)
(10)
Because many kinds of isomer are produced during the
inter-conversion reaction, the distribution of produced C4s was
monitored depending on the conversion (Fig. 9). Olefinic C4
s
such as 1-butene, trans-butene, cis-butene, and iso-butene were
preferentially formed with only small fraction of saturated C4s.
The concentration of C4 isomers has following order: iso-butene

D. Liu et al. / Catalysis Today 226 (2014) 52–66
61
ethene (17%) and propene (21%). This means that the isomerization
rate is higher than that of oligomerization and cracking.
1-butene can be oligomerized to olefinic hydrocarbons higher
than C8 but they are not detected in the product because their
cracking rate is much higher than that of formation. During this
oligomerization → re-cracking process, ethene and propene are
selectively produced the same as the reaction of ethylene.
Also, as a secondary product, BTX were produced through dehy-
drogenation of olefinic oligomer to alkadiene and then cyclization
as the contact time was increased. That is, the increase of contact
time brought linear increase of BTX and H2 together with decrease
of C4 isomer. This means that the contact time in catalytic naphtha
cracking should be optimized to guarantee high light olefin yield
by minimizing the formation of BTX because they are irreversible
product.
100
90
80
70
60
50
(a)
In the reaction of 1-butene, the light olefins such as ethene and
propene were mainly produced by further cracking of dimer of
C4 s, more ethene than propene was produced as the contact time
increased due to higher stability of ethene than propene at high
temperature.
32
64
128
256
WHSV (h-1)
100
80
60
40
20
0
The distribution of hydrocarbon products observed from ethene
conversion was the same and there is no obvious difference in their
selectivity over the four different catalysts. Propene, as the dom-
inant product, can be regarded as a secondary product produced
by re-cracking of the ethene oligomers. Therefore, its initial selec-
tivity was quite low at short contact time and then reached to a
maximum with increase of contact time due to further reaction
of produced propene. Butene, the dimer of ethene, showed higher
selectivity and maximum at short contact time and decreased
gradually with increasing contact time. Methane mainly comes
from the thermal cracking of ethene oligomers, increased lin-
early with contact time, which indicating that the concurrence
of catalytic cracking and radical cracking under catalytic naph-
tha cracking condition. C5s and C6s olefins were produced both
as secondary products, so their selectivity are low at short con-
tact time, then increased to maximum and finally decreased due to
re-cracking.
BTX
C6s
C5s
(b)
C4s
C3H8
C3H6
C2H6
C2H4
CH4
H2
32
64
128
256
The formation of saturates is due to hydrogen transfer reac-
tions catalyzed by strong Brönsted acid sites. Ethane selectivity
decreased with increasing contact time due to the decreasing
concentration of ethene. Propane appears only when propene
selectivity reaches the maximum at higher conversion of
ethene, suggesting that propane is produced from propene and
increased with contact time. BTX, mainly formed irreversibly
by hydrogen transfer reaction, increased linearly with contact
time.
WHSV (h-1)
Fig. 10. (a) Conversion and (b) product selectivity in the reaction of 1-butene over
HZSM-5 (T = 675 ◦C, reactant composition (vol%): 1-C4H8/H2/He = 10/5/85).
(33%) > 1-butene and trans-butene (21%) > cis-butene (17%) > cyclo-
butane (5%) ꢁ 1,3-butadiene (trace amount). Also, the product
distribution at short contact time is same as that at long con-
tact time. That is, the product distribution is almost same without
change regardless of contact time. This suggests that the isomeriza-
tion of C4 occurs very quickly and reach to an equilibrium prior to
further cracking to more light hydrocarbons. This result agrees well
with the previous results that the rate of isomerization of C5–C8
alkenes is faster than that of their cracking over H-ZSM-5 [25].
Compared with olefinic C4s, the faction of saturated C4s includ-
ing butane and i-butane is very low (generally less than 1%), so it can
be suggested that the hydrogenation or hydrogen transfer reaction
of C4 occurs very slowly and can be neglected in catalytic naphtha
cracking.
For further understanding of inter-conversion behavior of
ethene and propene, the reaction of 1-butene was carried out at
the same reaction condition of ethene over HZSM-5 (Fig. 10). 1-
butene revealed higher reactivity than ethene and propene and the
higher conversion was obtained; 1-butene ꢁ propene > ethene.
In product selectivity, also quite predictable result was obtained
as that of ethene reaction. At short contact time of 256 h−1 cor-
responding to the low 1-butene conversion, isomerized C4s (45%)
were obtained preferentially together with cracked product of
3.5.3. The inter-conversion selectivity of propene depending on
the types of catalyst
The selectivity depending on propene conversion also shows
the same trends over different catalysts (Fig. 11), indicating that
the reaction mechanism of inter-conversion also does not change
depending on the acidity of ZSM-5.
In propene reaction, ethene is formed predominantly and its
fraction increases with contact time. Because ethene stability is rel-
atively high compared with other hydrocarbons, once it is formed,
it is maintained as it is without further conversion to other hydro-
carbons and about 70% selectivity is obtained even at long contact
time. Propane selectivity from propene was high at short con-
tact time but it decreased rapidly due to the decrease of propene
concentration with increasing contact time. However, the other sat-
urates e.g. methane and ethane, increased with the conversion of
propylene. C4s come from the cracking of propene oligomers or
the dimerization of produced ethene, increased to maximum and
then decreased with increasing contact time due to re-cracking.

62
D. Liu et al. / Catalysis Today 226 (2014) 52–66
30
20
10
80
(b)
(a)
(c)
HZSM-5
St-HZSM-5
P/HZSM-5
60
40
20
St-P/HZSM-5
0
0
0
0
0
0
15
15
15
15
30
30
30
30
45
45
45
45
60
60
60
60
0
0
0
0
15
15
15
15
30
30
30
30
45
45
45
45
60
60
60
60
10
30
(d)
(f)
8
6
4
2
20
10
0
0
60
10
(e)
(g)
8
6
4
2
45
30
15
0
0
10
10
(h)
8
6
4
2
0
8
6
4
2
0
C3H6 Conversion (%)
C3H6 Conversion (%)
Fig. 11. Selectivity of (a) CH4, (b) C2H4, (c) C2H6, (d) C3H8, (e) C4s, (f) C5s, (g) C6s and (h) BTX depending on the conversion in the reaction of propene over different types of
catalysts; (ꢀ) HZSM-5, ( ) St-HZSM-5, ( ) P/HZSM-5 and ( ) St-P/HZSM-5 (T = 675 ◦C, reactant composition (vol%): C3H6/H2/He = 15/5/80).
The inter-conversion of propene occurs dominantly through the
following pathway (reaction (11)).
Same as in ethene reaction, the dimerized product of C6 is
formed preferentially at the initial stage of reaction and then
decreased with increasing contact time due to the further sec-
ondary reaction of re-cracking and aromatization (Fig. 11(g)). The
2C3 → C6 → C2 + C4
(11)

D. Liu et al. / Catalysis Today 226 (2014) 52–66
63
80
60
40
20
20
HZSM-5
(a)
(b)
16
12
8
St-HZSM-5
P/HZSM-5
St-P/HZSM-5
4
0
0
0
0
0
0
15
15
15
30
30
30
30
45
45
45
60
60
60
60
0
0
0
15
15
15
30
30
30
45
45
45
60
60
60
40
80
(d)
(f)
(c)
(e)
30
20
10
60
40
20
0
60
20
15
10
5
45
30
15
0
0
80
C3H6 Conversion (mol%)
(g)
60
40
20
0
10
20
40
50
C3H6 Conversion (mol%)
Fig. 12. Selectivity of (a) CH4, (b) C2H4, (c) C3H8, (d) C4s, (e) C5s, (f) C6s and (g) BTX depending on the conversion in the reaction of ethane and propene mixture over different
types of catalysts; (ꢀ) HZSM-5, ( ) St-HZSM-5, ( ) P/HZSM-5and ( ) St-P/HZSM-5 (T = 675 ◦C, reactant composition (vol%): C2H4/C3H6/H2/He = 20/15/5/60).

64
D. Liu et al. / Catalysis Today 226 (2014) 52–66
C2
C4
+
C3
=
=
(C )
+
C6
4
+
isomers
+
+
+
=
C4
=
C7
+
=
=
C8
C5
+
C5
Aromatics
(a) Catalytic inter-conversion network of ethene
C3
C4
C2
=
=
+
C6
C4
+
+
isomers
+
=
=
+
C3
C3
+
=
=
C3
C5
=
=
C7
C5
+
+
C5
=
C9
Aromatics
(b) Catalytic inter-conversion network of propene
C3
C2
+
+
=
C5= isomers
C8
=
2C4
C4
+
=
C4= isomers
+
+
C
+
3
=
C7
=
=
C6
+
C2
+
C6= isomers
Aromatics
(c) Catalytic inter-conversion network of ethene & propene mixtures
Fig. 13. Catalytic inter-conversion networks of light olefins at high temperature over acidic zeolite.

D. Liu et al. / Catalysis Today 226 (2014) 52–66
65
formed C6 is quite reactive and re-cracked to most stable ethene.
Once ethene is formed other dimerized products such as C4 and
C5 begin to be formed (Fig. 11(e and f)). However, the level of
dimerized fraction is not so high because they are re-cracked to
ethene again as the contact time is increased.
Different from the dimerized products, the selectivity of BTX
increased linearly with contact time same as in the ethene reaction
(Fig. 11(h)). The two routes for the formation of BTX can be sug-
gested such as hydrogen transfer reaction by strong Brönsted acid
site and dehydrogenation over stronger Lewis acid site. Therefore,
the highest yield of BTX is obtained over parent HZSM-5 with the
highest strong acid sites.
Except for the main reaction of oligomerization or copoly-
merization of light olefins and re-cracking of oligomers and
copolymers, the other reactions such as hydrogen transfer, iso-
merization, dehydrogenation, cyclization and aromatization also
occur simultaneously to form a complex reaction network. So
the following complicated catalytic reaction networks of inter-
conversion of light olefins could be suggested over HZSM-5 based
catalyst in naphtha cracking condition (Fig. 13).
(1) The conversion of ethene
The inter-conversion of ethene over acid zeolite occurs
through the following procedures. Ethene is dimerized firstly
to C4 , which is difficult to undergo monomolecular cracking
and then the formed C4 is further polymerized to C6 . In naph-
tha cracking condition, because the cracking rate of C6 is faster
than the formation rate, C3 is formed dominantly as an inter-
conversion product.
From the result of propene reaction, it can be seen also clearly
that the propene is inter-converted preferentially to ethene though
the series reaction of oligomerization and re-cracking.
3.5.4. The inter-conversion selectivity of ethene and propene
mixture depending on the types of catalyst
(2) The conversion of propene
For better understanding of light olefin inter-conversion, reac-
tion was carried out with a mixture of ethene and propene over
four different catalysts with different WHSV of stream (Fig. 12).
An interesting result was obtained with the mixture different from
single reactant. Because the ethene and propene exist together at
the initial stage of reaction it was expected to observe all types
of dimerized products such as such as C4 , C5 and C6 even at
short contact time. However, only C5 and C6 were observed at
short contact time without C4 (Fig. 12(d–f)). This means that the
dimerization rate of propene is higher than that of ethane. Because
propene is reactive enough to dimerize even with ethene as well as
propene itself, the dimerized C5 and C6 are formed preferentially
from the beginning of reaction but C4 begins to be formed after
some of propylene is used for dimerization and have a maximum
at a certain degree of conversion. This result agrees well with pre-
vious result that the copolymerization of ethene with propene is
faster than the dimerization of ethene itself over SAPO-34 zeolite
[24].
The concentration of methane increased linearly with contact
time and reached to 40% at the conversion of 60%. This level of
methane formation is about two times higher than that in the
reaction of single reactant because of increased concentration of
light olefins. According to the previous study, methane is usually
regarded as the product of radical cracking, which may be associ-
ated with the acidity from extra-framework aluminum (IV) [26],
so the St-HZSM-5, most severely dealuminated steamed HZSM-
5, induces more favorably the radical cracking of dimerized light
olefins and revealed the highest methane formation among the four
types of catalyst (Fig. 12(a), red filled circle).
The inter-conversion of propene over acid zeolite occurs
through the following procedures. Propene is dimerized firstly
to C6 and then the formed C6 is re-cracked to 2C3 or C2 and
C4 . The formed C3 and C4 is further dimerization and then re-
cracked to C2 and C3 at longer contact time (see Fig. 11). The
main product in inter-conversion of propene is ethene because
the formed propene is neglected as the reactant.
(3) The conversion of ethene and propene mixture
When the mixture of ethene and propene is used as the
reactant, the reaction of ethene and propene will undergo
oligomerization and copolymerization simultaneously to form
C4 , C5 and C6 . The formed C4 and C5 will be further dimer-
ized to higher olefins and then they are re-cracked to small
olefins while C6 undergoes monomolecular cracking. More
methane is formed due to the increased concentration of light
olefins.
In catalytic naphtha cracking, quite similar light olefin
distributions are obtained regardless of catalyst types and reac-
tion conditions, which means that inter-conversion of light
olefins through oligomerization → re-cracking occurs simulta-
neously together with naphtha cracking. Therefore, it is difficult to
control the selectivity of ethene and propene in catalytic naphtha
cracking.
4. Conclusions
The inter-conversion of light olefins was carried out over four
different types of ZSM-5 based catalysts in catalytic naphtha crack-
ing condition and the following conclusions were obtained.
The presence of hydrogen did not affect activity and selectiv-
ity in the inter-conversion of ethene and propene, suggesting that
light paraffin would formed mainly by hydrogen transfer reaction
in catalytic naphtha cracking.
The inter-conversion of ethene and propene is dominant side
reaction in catalytic naphtha cracking. That is, the dimerization
rate of light olefins is comparable to that of cracking and the
inter-conversion occurs through the procedure of oligomeriza-
tion → re-cracking. The product distribution in inter-conversion of
light olefins was not influenced by the types of catalyst and reaction
conditions and almost the same product distribution was obtained
at the same conversion, which suggests that the inter-conversion
of light olefin follows the same reaction mechanisms.
All the other paraffins and BTX product would increase with
the contact time, and the St-P/HZSM-5 always showed the lowest
selectivity for paraffins and BTX. That is, the St-P/HZSM-5 revealed
lower by-products formation at the same feed conversion due to the
decreased hydrogen transfer reaction, which suggesting that the
modification of HZSM-5 with P will reduce the secondary reaction
of light olefins catalytic naphtha cracking.
3.6. Catalytic reaction network in the inter-conversion of light
olefins
As discussed in previous sections, the inter-conversion of ethene
and propene is not negligible in catalytic naphtha cracking con-
dition and the following four main reaction steps should be
considered in the inter-conversion of light olefins: (1) oligomeriza-
tion of light olefins, (2) re-cracking and isomerization of oligomers,
(3) copolymerization of mixture of light olefins, and (4) re-cracking
and isomerization of copolymer.
Finally, to guarantee high light olefin yield in catalytic naphtha
cracking the catalyst and reaction conditions should be optimized
to minimize inter-conversion of light olefins.

66
D. Liu et al. / Catalysis Today 226 (2014) 52–66
[11] J.P. Van Den Berg, J.P. Wolthuizen, J.H.C. Van Hooff, J. Catal. 80 (1983) 139.
References
[12] T.J. Gricus Kofke, R.J. Gokte, J. Catal. 115 (1989) 233.
[13] A. Ghosh, R.A. Kydd, J. Catal. 100 (1986) 185.
[14] J.P. Van Den Berg, J.P. Wolthuizen, A.D.H. Clague, G.R. Hays, R. Hurs, J.H.C. Van
Hooff, J. Catal. 80 (1983) 130.
[15] E.G. Derouane, J.P. Gilson, J.B. Nagy, J. Mol. Catal. 10 (1981) 331.
[16] L.G. Galya, M.L. Occelli, J.T. Hsu, D.C. Young, J. Mol. Catal. 32 (1985) 391.
[17] K. Ramesh, C. Jie, Y.-F. Han, A. Borgna, Ind. Eng. Chem. Res. 49 (2010) 4080.
[18] C.T. O’Connor, M. Kojima, Catal. Today 6 (1990) 329.
[19] T. Blasco, A. Corma, J. Martinez-Triguer, J. Catal. 237 (2006) 267.
[20] P.M. Bautista, J.M. Campelo, A. Garcia, D. Luna, J.M. Marinas, A.A. Romero, Appl.
Catal. A 96 (1993) 175.
[1] Y.K. Park, C.W. Lee, N.Y. Kang, W.C. Choi, S. Choi, S.H. Oh, D.S. Park, Catal. Surv.
Asia 14 (2010) 75.
[2] Y. Yoshimura, N. Kijima, T. Hayakawa, K. Murata, K. Suzuki, F. Mizukami, K.
Matano, T. Konishi, T. Oikawa, M. Saito, T. Shiojima, K. Shiozawa, K. Wakui, G.
Sawada, K. Sato, S. Matsuo, N. Yamaoka, Catal. Surv. Jpn. 4 (2000) 157.
[3] S.Y. Han, C.W. Lee, J.R. Kim, N.S. Han, W.C. Choi, C.H. Shin, Y.K. Park, Stud. Surf.
Sci. Catal. 153 (2004) 157.
[4] Y. Wei, Z. Liu, G. Wang, Y. Qi, L. Xu, P. Xie, Y. He, Stud. Surf. Sci. Catal. 158 (2005)
1223.
[5] D. Liu, W.C. Choi, C.W. Lee, N.Y. Kang, Y.J. Lee, C.H. Shin, Y.K. Park, Catal. Today
164 (2011) 154.
[6] W.E. Garwood, in: D. Galen, Stucky, G. Francis, Dwyer (Eds.), Intrazeolite Chem-
istry, American Chemical Society, Washington, DC, 1983, p. 383.
[7] S.A. Tabak, F.J. Krambeck, W.E. Garwood, Am. Inst. Chem. Eng. 23 (1986) 1526.
[8] R.J. Quam, L.A. Green, S.A. Tabak, F.J. Krambeck, Ind. Eng. Chem. Res. 27 (1988)
565.
[21] W.W. Kaeding, S.A. Butter, J. Catal. 61 (1980) 155.
[22] A.R. Grimmer, U. Haubenreisser, Chem. Phys. Lett. 99 (1983) 487.
[23] Y.X. Zhao, B.W. Wojciechowski, J. Catal. 144 (1993) 377.
[24] H.Q. Zhou, Y. Wang, F. Wei, D.Z. Wang, Z.W. Wang, Appl. Catal. A 348 (2008)
135.
[25] J.S. Buchanan, J.G. Santiesteban, W.O. Haag, J. Catal. 158 (1996) 279.
[26] A. Corma, V. Fornés, A. Martínez, A.V. Orchillés, in: W.H. Flank, T.E. Whyte
(Eds.), Perspectives in Molecular Sieve Science, American Chemical Society,
Washington, DC, 1988, p. 542.
[9] P. Borges, R. Ramos Pinto, M.A.N.D.A. Lemos, F. Lemos, J.C. Védrine, E.G. Der-
ouane, F. Ramôa Ribeiro, Appl. Catal. A 324 (2007) 20.
[10] I.P. Dzikh, J.M. Lopes, F. Lemos, F. Ramôa Ribeiro, Appl. Catal. A 177 (1999) 245.