10 Journal of the Japan Petroleum Institute, 61, (1), 10-19 (2018)

[Review Paper]

Selective Formation of Light Olefins by Catalytic Cracking of

Naphtha Components over ZSM-5 Zeolite Catalysts

Kohei KUBO, Hajime IIDA, Seitaro NAMBA, and Akira IGARASHI＊

Dept. of Applied Chemistry, Faculty of Engineering, Kogakuin University, 2665-1 Nakano-machi, Hachioji, Tokyo 192-0015, JAPAN

(Received July 4, 2017)

Catalytic cracking of C6-8 hydrocarbons as model naphtha components was evaluated over H-ZSM-5 catalysts
with various Si/Al ratios at 723-923 K to selectively form light olefins. The highest ethylene＋propylene yield
was 59.7 C-% with a propylene/ethylene ratio of ca. 0.72 at 99.6 % conversion in the cracking of n-heptane over
H-ZSM-5 (Si/Al＝31) at 923 K. The selectivities and activation energies indicated that monomolecular cracking
was predominant at temperatures higher than 923 K. Next, cracking of various C6-8 hydrocarbons on H-ZSM-5
at high temperatures in the presence of another hydrocarbon was also investigated. Coexistent 1-hexene did not
affect the cracking rate of n-heptane, cyclohexane, or methylcyclohexane, also indicating that monomolecular
cracking was predominant at temperatures as high as 923 K. In contrast, coexistent cyclohexane and methyl-
cyclohexane considerably affected the cracking rate of n-heptane because of their slow diffusion in the pores of
H-ZSM-5. Moreover, catalysts with activity equal or higher than that of H-ZSM-5 and steaming stability higher
than that of H-ZSM-5 were examined. The 194 %Cu-ZSM-5 and 95 %Ag-ZSM-5 zeolites exhibited high steam-
ing stabilities and higher cracking activities than H-ZSM-5 and P/H-ZSM-5. In particular, 194 %Cu-ZSM-5
exhibited an ultra-high steaming stability under conditions such as 1023 K and 10 h presumably because Cu2＋ and
Cu＋ cations in the 194 %Cu-ZSM-5 are not reduced to Cu0 at a high temperature of 1023 K in the presence of O2.

Keywords
Catalytic cracking, Naphtha, Light olefin, ZSM-5, Steaming stability

1. Introduction compounds of naphtha have used n-paraffins because

naphtha is a mixture of hydrocarbons and the main
Light olefins, such as ethylene, propylene, and components are C5-8 n-paraffins. Zeolite is a typical
butene, are important raw materials for the industrial catalyst used for naphtha cracking3)∼23). Specifically,
manufacture of various polymers and petrochemicals. ZSM-5 zeolite is one of the most efficient catalysts for
Ethylene is the major product of the naphtha thermal the naphtha catalytic cracking process at high tempera-
cracking process, whereas propylene is a byproduct of tures (873-923 K)13)∼23). The reaction mechanism of n-
this process 1). The petrochemical industry is currently paraffin cracking has been extensively investigated3)∼24),
facing a major shortage of propylene as a result of the but has not been completely clarified. Understanding
increasing demand for propylene derivatives (e.g., poly- of the activity and selectivity of catalysts for naphtha
propylene). However, the naphtha thermal cracking cracking requires not only kinetic studies but also con-
process cannot meet the demand for propylene since the sideration of the reaction mechanism.
selectivity for specific olefins is difficult to control (pro- Coke formation is considerable during naphtha
pylene/ethylene weight ratio is 0.5-0.6). Moreover, cracking over zeolite catalyst, resulting in coverage of
this process is usually operated at high temperatures the catalytic surfaces and blocking of the micropores,
(＞ 1073 K), and so requires a large amount of energy 2). resulting in deactivation of the catalyst. Therefore, re-
Therefore, the naphtha catalytic cracking process is cur- generation of H-ZSM-5 by calcination in air is neces-
rently under investigation as an alternative to naphtha sary. However, a small amount of steam is generated
thermal cracking, because of its superior ability to obtain during the regeneration, because coke contains a small
a propylene/ethylene weight ratio higher than 0.6 at amount of hydrogen25). Permanent deactivation of the
lower reaction temperature. Most research into model catalyst can occur related to dealumination from the
zeolitic framework in a hydrothermal atmosphere.
DOI: doi.org/10.1627/jpi.61.10
Therefore, a much higher stability against steaming
＊
To whom correspondence should be addressed. conditions than that of H-ZSM-5 is required for im-
＊
E-mail: [email protected] proved catalysts for naphtha cracking.

J. Jpn. Petrol. Inst., Vol. 61, No. 1, 2018

 11

This article reviews the reaction mechanism of cata- Table 1 kc and Ea for the Catalytic Cracking of C6-8 Hydrocarbons
lytic cracking of C6-8 hydrocarbons as a model reac- kc [mol gcat.–1 h–1]
tion of naphtha catalytic cracking over H-ZSM-5 Reactant Ea [kJ mol–1]
723 K 823 K 923 K
zeolites with various Si/Al ratios at 723-923 K, and dis-
cusses the stability against steaming conditions of n-Hexane 0.0726 0.570 2.35 97
n-Heptane 0.177 0.719 2.60 74
H-ZSM-5 and modified ZSM-5.
n-Octane 0.275 1.06 3.28 69
3-Methylpentane 0.0453 0.409 2.07 106
2. Cracking of Hydrocarbons over H-ZSM-5 2-Methylhexane ‐ ‐ 3.72 ‐
Zeolite at High Temperature Methylcyclopentane 0.368 1.98 6.29 79
Cyclohexane 0.0959 0.704 2.78 94
Methylcyclohexane 0.132 0.881 2.37 81
H-ZSM-5 catalysts (Si/Al＝31, 34, 51, 106, and 200)
were synthesized by a hydrothermal procedure followed
by conventional ion exchange in 1 M (1 mol L–1) HCl
solution26). Hydrocarbon cracking was performed at ( S i / A l＝25) a t 753-813 K w a s 149 k J m o l –1 6).
723-923 K15),22). The contribution of thermal cracking Narbeshuber et al. found the activation energy for the
was not negligible at 773 K or higher temperatures. monomolecular cracking of n-hexane over H-ZSM-5
Therefore, the rate constant of catalytic cracking was (Si/Al＝35) at 723-823 K was 105 kJ mol–1 and that the
subtracted from the contribution of thermal crack- activation energy decreased longer hydrocarbon chain
ing15),17),21)∼23). length7). Nakasaka et al. reported that the activation
Haag et al. proposed that cracking of n-hexane and energy for cracking of n-heptane over H-ZSM-5 (Si/Al
of 3-methylpentane proceeds through two simultaneous ＝320) at 823-923 K was 128 kJ mol–1, and that at 748-
pathways, bimolecular cracking and monomolecular 823 K was 93 kJ mol–1 20). Therefore, monomolecular
cracking3),4). Bimolecular cracking is the classical cracking absolutely predominated in the cracking of n-
chain process involving hydride transfer between a hexane at 923 K. The reactivity of n-paraffins in-
paraffin molecule and an adsorbed carbenium ion. creased with higher carbon number.
Formation of a new carbenium ion is followed by skele- Table 1 shows the rate constants and apparent acti-
tal isomerization prior to β-scission to avoid the forma- vation energies for the catalytic cracking of C6-8
tion of primary carbenium ions. Monomolecular hydrocarbons22). The rate constants increased with
cracking involves direct protonation of a paraffin to higher carbon number at 923 K as observed at low tem-
form a high energy transition state that may resemble a peratures7). However, the difference in the rate con-
penta-coordinated carbonium ion, which is subsequently stants due to carbon number at 923 K was small
cracked or dehydrogenated. Bimolecular cracking compared to that at 723 K, probably because mono-
predominates at low temperature and high olefin con- molecular cracking was predominant at high tempera-
centration, i.e. high conversion. Monomolecular tures whereas both monomolecular and bimolecular
cracking is favored by the opposite conditions. cracking proceeded simultaneously at low temperatures.
bimolecular cracking: The rate constants for the catalytic cracking of
RH + R 1+ Æ R + + R 1H 2-methylhexane and methylcyclopentane containing a
tertiary carbon were higher than those of n-heptane and
R + = Olefin + R 1+ cyclohexane without the tertiary carbon. The transition
where RH＝original paraffin, R1＋＝smaller carbenium state in the monomolecular cracking of 2-methyhexane
ion. and methylcyclohexane is not so bulky compared with
monomolecular cracking: that in the bimolecular cracking of 2-methylhexane and
C m H 2m+2 + H + Æ C m H 2m+3 + methylcyclopentane, so no restricted transition state
selectivity occurs in contrast to the cracking of 3-methyl-
Æ C n H 2n+2 + C m-n H 2( m-n )+1+
pentane on H-ZSM-5 at 811 K5). On the other hand,
C m-n H 2( m-n )+1+ Æ C x H 2x + C y H 2 y + H + the rate constant for catalytic cracking of 3-methyl-
[n＝0 ∼ (m−1), x＋y＝m−n] pentane with a tertiary carbon was lower than that of n-
Monomolecular cracking only forms hydrogen, hexane at low temperatures, because bimolecular crack-
methane and ethane as primary products, excluding the ing of 3-methylpentane via a bulky transition state is
effects of thermal cracking. more restricted than that of n-hexane in narrow reaction
Haag et al. reported an activation energy of 126 kJ spaces5). However, the effect on the rate constant at
mol–1 for the monomolecular cracking of n-hexane over 923 K was much smaller than that at 723 K, because
H-ZSM-5 (Si/Al＝35) at 723-813 K and 153 kJ mol–1 monomolecular cracking was predominant at high tem-
for the bimolecular cracking of n-hexane at 513-573 K4). peratures. The rate constants for the catalytic cracking
Babitz et al. reported that the activation energy for the of cyclohexane and methylcyclohexane were lower than
monomolecular cracking of n-hexane over H-ZSM-5 that of methylcyclopentane, probably because of their

J. Jpn. Petrol. Inst., Vol. 61, No. 1, 2018

 12

Table 2 Selectivity (C-%) at ca. 38 % Conversion for the Cracking of C6-8 Hydrocarbons at 923 K

Reactant
Product
n-Hexane n-Heptane n-Octane 2-Methylhexane Methylcyclopentane Cyclohexane Methylcyclohexane
Ethylene 16.1 18.2 18.0 13.4 9.7 15.8 11.4
Propylene 37.1 31.9 33.1 40.3 61.5 44.3 33.3
Butenes 16.2 18.6 17.6 22.0 12.5 10.0 19.1
Benzene 0.5 0.5 0.4 0.2 1.7 8.3 7.3
Toluene 0.4 0.4 0.4 0.2 1.3 5.9 12.3
Xylene 0.1 0.1 0.1 0.0 0.6 2.5 3.1
Methane 3.6 2.5 2.3 5.2 0.9 0.9 2.8
Ethane 9.5 6.6 4.9 4.0 0.2 0.5 0.5
Propane 10.1 8.6 6.6 4.0 0.6 1.9 1.1
Butanes 4.1 7.4 7.0 6.6 0.3 0.4 0.4
C5+ 2.4 5.2 9.6 3.9 10.6 9.3 8.7

slow diffusion rates.

Table 2 shows the selectivity (C-%) at a conversion
of ca. 38 % for the cracking of C6-8 hydrocarbons at
923 K22). The selectivities for light olefins showed little
difference between the C6-8 paraffins. The selectivity
for propylene was very high for methylcyclopentane,
because methylcyclopentane is stoichiometrically
decomposed into two molecules of propylene. The
selectivity for aromatics [benzene, toluene, and xylene
(BTX)] was very high for cycloparaffins, especially for
cyclohexane and methylcyclohexane. Therefore, the
direct dehydrogenation reaction to the corresponding
BTX, such as cyclohexane to benzene and methylcyclo-
hexane to toluene, occurs in addition to the cracking
reaction11),12),19). In fact, the selectivity for benzene
was high for cyclohexane, and that for toluene was high
for methylcyclohexane.
The relationships between selectivities expressed as
C-% at a conversion of ca. 50 % and reaction tempera-
tures for n-heptane cracking over H-ZSM-5 (Si/Al＝
31) are shown in Fig. 115). Selectivities for ethylene
and propylene increased at higher reaction temperature,
whereas the propylene/ethylene ratio decreased from 3
at 723 K to 1.4 at 923 K. The selectivities for propane
and butanes decreased with higher reaction tempera-
ture. The selectivities for methane and ethane
increased, whereas those for butenes and BTX did not
change noticeably with higher reaction temperatures.
The change in selectivity with reaction temperature
may be mainly caused by greater dominance of mono-
molecular cracking over bimolecular cracking at higher
reaction temperature. The selectivities for methane, Fig. 1●Relationships between Selectivities at a Conversion of ca.
ethane, and light olefins increase, and the selectivities 50 % and Reaction Temperatures for n-Heptane Cracking
for propane and butanes decrease at a higher ratio. over H-ZSM-5 (Si/Al＝31) (TOS＝15 min)
The relationships between the selectivity expressed
as C-% and conversion for n-heptane cracking over
H-ZSM-5 (Si/Al＝31) at 923 K are shown in Fig. 215). Hydride transfer from olefins to carbenium inter-
The selectivities for propylene and butenes decreased mediate species occurs on acid catalysts at high conver-
with higher conversion. Only a small amount of BTX sion levels to form paraffins and hydrogen-deficient
formed at low conversion, whereas the selectivity for species. Such hydrogen-deficient species, e.g. allyl-
BTX increased with higher conversion. carbenium ion, would be further transformed into

J. Jpn. Petrol. Inst., Vol. 61, No. 1, 2018

 13

Fig. 3●Relationships between Sum of Selectivities for Hydrogen,

Methane, or Ethane and Reaction Temperatures for n-Hep-
tane Cracking over H-ZSM-5(51) (TOS＝15 min) (n-heptane
conversion＝29-55 %)

products and little is formed or consumed by secondary

reactions up to 55 % conversion. Consequently, that
the ratio of monomolecular cracking to bimolecular
cracking may not decrease at higher n-heptane conver-
sion, i.e. olefin pressure, at the low initial pressure of n-
heptane.
The relationship between the sum of selectivities for
hydrogen, methane, and ethane, and reaction tempera-
ture for heptane cracking over H-ZSM-5 (Si/Al＝51)
(n-heptane conversion＝29-55 %) is shown in Fig. 3.
Hydrogen＋methane＋ethane selectivity (mol%)
Fig. 2●Selectivities for n-Heptane Cracking over H-ZSM-5 (Si/Al
increased with higher reaction temperature, and the
＝31) (reaction temp.＝923 K, TOS＝15 min)
s e l e c t iv i t y r e a c h e d 65 % a t 873 K a n d 923 K .
Therefore, monomolecular cracking predominantly pro-
BTX9). The results shown in Fig. 2 indicate that BTX ceeded at 873 K and higher temperatures.
is secondary reaction product and forms mainly from The relationship between the sum of selectivities for
propylene and butenes. Therefore, the formation of hydrogen, methane, and ethane, and n-heptane pressure
BTX reduces the olefin yields. On the other hand, for n-heptane cracking is shown in Fig. 4. The sum of
little ethylene is consumed by the secondary reaction, selectivities for hydrogen, methane, and ethane decreased
because of the low reactivity of ethylene. with higher heptane pressure at 773 K. The partial
The ethylene＋propylene selectivity increased, pressure of the produced olefin increases with higher
whereas the propylene/ethylene ratio decreased with heptane pressure so the contribution of bimolecular
higher reaction temperature. Therefore, high tempera- cracking is increased by greater carbenium ion concen-
tures are required to obtain high ethylene＋propylene tration on the catalyst surface. On the other hand, the
yield. The highest ethylene＋propylene yield obtained sum of selectivities for hydrogen, methane, and ethane
in this study was 59.7 C-% (propylene/ethylene ratio of remained almost unchanged even at high heptane pres-
ca. 0.72) at 99.6 % conversion over H-ZSM-5 (Si/Al＝ sure. Presumably monomolecular cracking pre-
31) at 923 K. This propylene/ethylene ratio is 1.4 dominantly proceeded even at high heptane pressure.
times higher that obtained from conventional thermal From these results, we conclude that high tempera-
cracking (ca. 0.5)2). tures such as 873 K or 923 K, at which monomolecular
The yields of hydrogen, methane and ethane are cracking predominates, are desirable to increase the
almost proportional to the conversion up to 55 % con- olefin yield.
version15). Therefore, the selectivities for methane and
ethane remain constant up to 55 % conversion. Most
of the hydrogen, methane and ethane forms the primary

J. Jpn. Petrol. Inst., Vol. 61, No. 1, 2018

 14

3. Coexistence of other Hydrocarbons in Cracking single and binary reactants at 923 K22). The rate con-
on H-ZSM-5 Catalyst at High Temperatures stants (kc) for the cracking of n-heptane, cyclohexane,
and methylcyclohexane were almost the same in the
In order to understand the cracking chemistry of presence and absence of 1-hexene. Therefore, coexist-
naphtha, both cracking of a single hydrocarbon and ing 1-hexene does not affect the cracking rate of n-hep-
cracking of binary mixtures of hydrocarbons are impor- tane, cyclohexane, or methylcyclohexane. These find-
tant to study, because naphtha is a mixture of hydro- ings indicate that monomolecular cracking is
carbons. Moreover, understanding the effect of co- predominant at 923 K. Cracking of n-hexane acceler-
existence of other hydrocarbons in cracking on H-ZSM-5 ates at 643 K with the coexistence of an olefin4),
is important to design the catalytic process for the because the cracking proceeds by bimolecular cracking,
cracking of naphtha. which predominates at such a low temperature. The
First-order plots for the catalytic cracking of n- reaction rate of a low reactivity reactant will accelerate
heptane and 1-hexene, and binary n-heptane/1-hexene in the presence of high reactivity reactant in bimolecular
mixtures at 923 K were obtained22). Proportional rela- cracking.
tionships were observed in the catalytic cracking of n- If additivity in product yields can be established in
heptane coexisting with 1-hexene, and for the catalytic binary cracking, the selectivity for each product in binary
cracking of cyclohexane and methylcyclohexane co- cracking can be calculated from the cracking yield of
existing with 1-hexene. These results indicate that the each product. Cracking of n-heptane and 1-hexene,
catalytic cracking of n-heptane, cyclohexane, and methyl- and binary n-heptane/1-hexene was carried out under
cyclohexane coexisting with 1-hexene obeys first-order constant reaction conditions; constant temperature,
kinetics. On the other hand, no proportional relation- W/Ftotal, and initial pressure of each reactant. The cal-
ship was observed for the catalytic cracking of culated selectivity agrees precisely with the selectivity
1-hexene, because the conversion of 1-hexene was too obtained experimentally22). Therefore, cracking of n-
high (＞ 93 %) to determine the rate constant. Table 3 heptane and 1-hexene proceeded almost independently.
shows the rate constants for the catalytic cracking of First-order plots for the catalytic cracking of n-
heptane, methylcyclohexane, and binary n-heptane/
m e t h y l cy c l o h ex a n e a t 923 K w e r e o b t a i n e d 22).
Proportional relationships were observed in the catalytic
cracking of n-heptane and methylcyclohexane and in
the cracking of the binary mixture, and in the catalytic
cracking of cyclohexane and n-heptane and in the
cracking of the binary mixture. These results indicate
that the catalytic cracking of n-heptane, cyclohexane,
and methylcyclohexane in a binary mixture obeys first-
order kinetics. Table 4 shows the rate constants for
catalytic cracking of single and binary reactants at
923 K22). The rate constants (kc) for the cracking of
cyclohexane and methylcyclohexane are almost the
same in the presence and absence of n-heptane.
Therefore, n-heptane does not affect the cracking rate
of cyclohexane and methylcyclohexane. On the other
hand, the rate constant (kc) for the cracking of n-heptane
Fig. 4●Relationships between Sum of Selectivities for Hydrogen,
Methane, or Ethane and n-Heptane Pressure for n-Heptane
was lower in the presence than in the absence of cyclo-
Cracking over H-ZSM-5(51) (TOS＝15 min) (n-heptane hexane or methylcyclohexane. Therefore, cyclo-
conversion＝40 %) hexane and methylcyclohexane affect the cracking rate

Table 3 kc for the Catalytic Cracking of Single and Binary Reactants at 923 K

kc [mol gcat.–1 h–1]

Combination Component
Single Binary
n-Heptane/1-hexene n-Heptane 2.60 2.61
1-Hexene - -
Cyclohexane/1-hexene Cyclohexane 2.78 2.85
1-Hexene - -
Methylcyclohexane/1-hexene Methylcyclohexane 2.37 2.46
1-Hexene - -

J. Jpn. Petrol. Inst., Vol. 61, No. 1, 2018

 15

of n-heptane. The reaction obeys first-order kinetics, butenes. These observations indicate that the acid site
indicating that coverage of the reactant on the active density little affects the selectivity at a relatively high
sites is very low. Therefore, retardation of n-heptane conversion of 70 %.
cracking by coexistent cyclohexane and methylcyclo- On the other hand, CMR ratios, which Post et al. de-
hexane is not caused by their adsorption on the active fined as a parameter of the ratio of monomolecular
sites, but by their slow diffusion in the pores of H-ZSM-5, cracking to bimolecular cracking 24), increased signifi-
as observed in n-octane cracking on H-ZSM-5 at cantly with higher Si/Al ratio, suggesting that the ratio
673 K10). The apparent activation energies of cracking of monomolecular cracking to bimolecular cracking
of cyclohexane and methylcyclohexane were small, may increase with higher Si/Al ratio. This consider-
probably because the rate determining step is diffusion ation agrees with Post et al.24). However, the CMR
in the pores (Table 1). may not be a suitable parameter in our reaction condi-
Catalytic cracking of n-heptane, cyclohexane, and tions, because ethylene is produced by a secondary re-
methylcyclohexane, and binary n-heptane/cyclohexane action at high reaction temperatures such as 923 K.
and n-heptane/methylcyclohexane mixtures at 923 K at Moreover, the observation that the acid site density little
relatively low conversion (13-21 %) were performed. affects the selectivity suggests that the ratio of mono-
If two reactants crack individually in the cracking of a molecular cracking to bimolecular cracking may not
binary reactant, additivity in the selectivity will be greatly change with the Si/Al ratio, because the ratio
established. Thus, the selectivity for the cracking of must affect the selectivity. Brønsted acid strength is
binary reactants was calculated from the yields for the constant or slightly weakened with higher Al con-
cracking of n-heptane and cyclohexane or methylcyclo- tents27). Therefore, we consider that there is a distri-
hexane. The calculated selectivity agreed well with bution of acidity and the acid strength increases with
the selectivity obtained experimentally22). Therefore, higher acid site density of H-ZSM-5 based on the acti-
the cracking of n-heptane and cyclohexane or methyl- vation energy.
cyclohexane proceeds independently.

4. Effect of Si/Al Ratio of H-ZSM-5

The effect of reaction temperature and acid site den-

sity on n-heptane cracking activity can be inferred from
the activation energy as a function of Al content of
H-ZSM-5 catalysts. The activation energies obtained
from the Arrhenius plots over various Si/Al ratios are
shown in Fig. 515). The activation energy for the cata-
lytic cracking of n-heptane tended to decrease with
higher acid site density. Post et al. reported a similar
phenomenon for n-hexane cracking over H-ZSM-5 with
various Si/Al ratios, and concluded that the bimolecular
cracking route is favored by the presence of two neigh-
boring acid sites24).
The selectivities (not C-% but mol%) at 923 K as a
function of Si/Al ratio at conversion of ca. 70 % are
shown in Table 515). Since the selectivity is a func-
tion of the conversion, meaningful comparison of selec-
tivities can only be made at similar conversions, which
was achieved by adjusting the catalyst weight (W). All
H-ZSM-5 zeolites exhibited almost the same selectivi- Fig. 5●Effect of Acid Site Density of H-ZSM-5 on the Apparent
ties, olefin/paraffin ratios (O/P), and ratio of i-butene/n- Activation Energy of n-Heptane Cracking

Table 4 kc for the Catalytic Cracking of Single and Binary Reactants at 923 K

kc [mol gcat.–1 h–1]

Combination Component
Single Binary
n-Heptane/cyclohexane n-Heptane 2.60 2.35
Cyclohexane 2.78 2.77
n-Heptane/methylcyclohexane n-Heptane 2.60 2.10
Methylcyclohexane 2.37 2.40

J. Jpn. Petrol. Inst., Vol. 61, No. 1, 2018

 16

Table 5●Effect of Acid Site Density of H-ZSM-5 on Selectivities (mol%) for Cracking of n-Hep-
tane at ca. 70 % Conversion (reaction temp.＝ 923 K, TOS＝15 min)

Si/Ala) 31 34 51 106 200

C2H4 32.5 33.0 34.0 32.3 31.9
C3H6 28.7 26.3 29.9 29.3 28.3
n-C4H8 6.0 5.2 5.8 6.0 6.2
i-C4H8 3.2 2.8 3.2 3.3 3.4
BTX 1.5 2.5 1.0 1.0 0.7
CH4 6.4 8.4 5.9 6.4 7.6
C2H6 8.2 9.2 7.1 8.4 9.0
C3H8 8.2 7.9 7.8 7.7 7.1
n-C4H10 3.5 3.1 3.5 3.6 3.6
i-C4H10 0.4 0.4 0.3 0.3 0.2
C5+ 1.5 1.4 1.5 1.7 2.1
CMRb) 118 131 158 174 202
O/P c) 3.47 3.28 3.90 3.54 3.50
i-C4H8/n-C4H8 0.54 0.55 0.55 0.55 0.55
a) Determined by ICP measurement.
b) (C1＋∑C2)/i-C4 (as described24)).
c) CnH2n/CnH2n＋2 (n＝2-4).

5. Steaming Stability of H-ZSM-5, Ag-ZSM-5 and temperature peak was assigned to ammonia weakly
Cu-ZSM-5 held or physically adsorbed on the zeolite30),31). In
contrast, the desorption peak of ammonia at 660 K was
Phosphorus impregnation is known to be effective for scarcely observed in the case of 95 %Ag-ZSM-5 and
improving the steaming stability of H-ZSM-528). 194 %Cu-ZSM-5, but desorption peaks were observed
However, phosphorus-impregnated H-ZSM-5 (P/ at 756 K for 95 %Ag-ZSM-5 and at 847 K for the
H-ZSM-5) exhibits lower activity for catalytic cracking 194 %Cu-ZSM-5. We will discuss the nature of these
than H-ZSM-5, because the Brønsted acidity of P/ peaks later.
H-ZSM-5 is lower than that of H-ZSM-5. Sano et al. The Fourier transform infrared (FTIR) spectra of pyr-
reported that protons catalyze the dealumination of idine adsorbed on H-ZSM-5, 95 %Ag-ZSM-5, and
H-ZSM-529). Therefore, a catalyst with lower Brønsted 194 %Cu-ZSM-5 were obtained23). The band at
acidity may exhibit higher steaming stability, but lower 1545 cm–1 was assigned to pyridinium ions formed by
activity for catalytic cracking. Catalyst regeneration is proton transfer from Brønsted acid sites, and the band
carried out in the presence of O2; whereas cracking over at 1455 cm –1 was identical to pyridine adsorbed on
H-ZSM-5 at high temperatures takes place in a reduc- Lewis acid sites. Clearly oxidized 95 %Ag-ZSM-5
tive atmosphere, due to the formation of a considerable and oxidized 194 %Cu-ZSM-5 had few Brønsted acid
amount of H2. Therefore, catalysts with low Brønsted sites and considerable amounts of Lewis acid sites.
acidity in an oxidative atmosphere and high Brønsted The NH3-TPD desorption peaks at 756 K for 95 %Ag-
acidity in a reductive atmosphere are likely to have both ZSM-5 and at 847 K for the 194 %Cu-ZSM-5 may be
high steaming stability and high catalytic activity. We attributed to these Lewis acid sites. On the other hand,
considered that Ag-ZSM-5 and Cu-ZSM-5 with redox the amounts of Brønsted acid sites on reduced 95 %Ag-
properties were candidate catalysts for the naphtha cat- ZSM-5 and reduced 194 %Cu-ZSM-5 are similar to
alytic cracking process at high temperatures. those on H-ZSM-5. These findings indicate that these
Exchange rates of Ag＋ and Cu2＋ were determined to catalysts have low Brønsted acidity in an oxidative
be 95 % and 194 % by inductively coupled plasma- atmosphere and high Brønsted acidity in a reductive
atomic emission spectroscopy (ICP-AES) measure- atmosphere.
ments, respectively. Such high exchange rate of The first-order rate constants for the cracking of
194 % for Cu2＋ indicates that [Cu(OH)]＋ cations occu- cumene over the catalysts are shown in Table 623).
py almost all of the cation exchange sites17). The activities of reduced 194 %Cu-ZSM-5 and reduced
The temperature-programmed desorption of ammonia 95 %Ag-ZSM-5 were much higher than those of oxi-
(NH3-TPD) spectra for H-ZSM-5, 95 %Ag-HZSM-5, dized 194 %Cu-ZSM-5 and oxidized 95 %Ag-ZSM-5.
and 194 %Cu-ZSM-5 were obtained23). H-ZSM-5 These findings indicate that the Brønsted acid sites were
exhibited two desorption peaks at low (477 K) and high generated by reduction of copper cations and silver cat-
(660 K) temperatures. The high temperature peak was ions. Moreover, the activities of reduced 194 %Cu-
assigned to the desorption peak of ammonia, which was ZSM-5 and reduced 95 %Ag-ZSM-5 were higher than
adsorbed on the Brønsted acid sites, whereas the low that of H-ZSM-5. Catalytic activity for cracking or

J. Jpn. Petrol. Inst., Vol. 61, No. 1, 2018

 17

isomerization of hydrocarbons in the presence of H2 is Brønsted acidity.

enhanced by the addition of a metal capable of H2 dis- In order to examine the reduction temperatures of
sociation to acidic zeolites32),33). Hattori et al. pro- Cu2＋ in Cu-ZSM-5 and Ag＋ in Ag-ZSM-5, we carried
posed that such enhancement is due to the generation of out H2-temperature-programmed reduction (H2-TPR)
molecular hydrogen-originated protonic acid sites33). measurements. The H2-TPR spectra for 194 %Cu-
The higher activity of Cu-ZSM-5 and Ag-ZSM-5 than ZSM-5 and 95 %Ag-ZSM-5 are shown in Fig. 623).
that of H-ZSM-5 may be caused by the reduction of The low temperature peak of 194 %Cu-ZSM-5 was
Cu2＋ and Ag＋, resulting in the formation of Brønsted ascribed to Cu2＋ reduction to Cu＋ and the high temper-
acid sites, Cu metals, and Ag metals with H2 dissocia- ature peak to Cu＋ reduction to Cu0 34). Therefore, Cu2＋
tion ability at 923 K. On the other hand, the activities cations in Cu-ZSM-5 were reduced to copper metal at
of oxidized Cu-ZSM-5, oxidized Ag-ZSM-5, and P/ around 723 K. The low temperature peak of 95 %Ag-
H-ZSM-5 were lower than that of H-ZSM-5, because ZSM-5 was ascribed to the silver cationic clusters (Agn＋)
the modifications decreased the Brønsted acidity of the formed by reaction between Ag0 and Ag＋, and the high
H-ZSM-5. Moreover, the activities of the oxidized temperature peak to Agn＋ reduction to silver metal35)∼37).
Cu-ZSM-5 and oxidized Ag-ZSM-5 decreased with Therefore, some silver cations in ZSM-5 were reduced
higher exchange rate. These findings are consistent to silver metal at low temperatures at around 473 K,
with the NH3-TPD spectra. The order of cumene and all silver cations in Ag-ZSM-5 were reduced to sil-
cracking activities was H-ZSM-5 ＞ P/H-ZSM-5 ≫ oxi- ver metal at around 653 K. These findings indicate
dized 95 %Ag-ZSM-5 ＞ oxidized 194 %Cu-ZSM-5. that the Ag＋ cations in Ag-ZSM-5 are readily reduced
These findings show that Cu-ZSM-5 and Ag-ZSM-5 to metal, compared with the Cu2＋ cations in Cu-ZSM-5.
have low Brønsted acidities in the presence of O2 and The amounts of H2 consumption for each peak in the
high Brønsted acidities in a reductive atmosphere com- H2-TPR spectra were calculated from the peak areas
pared with H-ZSM-5. and theoretical values were calculated from the cation
In order to clarify the steaming stabilities, we carried exchange rate23). H2 consumptions at low and high
out cracking of n-heptane over catalysts after steaming temperatures were almost the same and their sum was
under more severe conditions (1023 K, 10.0 h). The nearly equal to the theoretical value for 194 %Cu-
steaming stabilities are shown in Table 723). The
order of retained percentage activity, which is a func-
tion of the steaming stability, was 194 %Cu-ZSM-5 ＞
95 %Ag-ZSM-5 ≫ P/H-ZSM-5 ＞ H-ZSM-5, which
was the reverse order of the cumene cracking activities
in the presence of O2. Cumene cracking activity is a
function of the Brønsted acidity. Therefore, the steam-
ing stability of a catalyst is strongly related to its

Table 6 First-order Rate Constants for Cracking of Cumene

Catalysts First-order rate constanta) [mol gcat.–1 h–1]

H-ZSM-5 1.16
194 %Cu-ZSM-5(O2) 0.11
194 %Cu-ZSM-5(H2) 1.44
95 %Ag-ZSM-5(O2) 0.16
95 %Ag-ZSM-5(H2) 1.89
P/H-ZSM-5 0.79
a) Data at 15 min on stream. Fig. 6 H2-TPR Spectra for 194 %Cu-ZSM-5 and 95 %Ag-ZSM-5

Table 7 Steaming Stabilities of Catalysts (steaming conditions: 1023 K, 10.0 h, H2O/N2 or H2O/O2 ＝ 1/1.83)

kca) [mol gcat.–1 h–1]

Catalysts Percent activity retainedb) [%]
Fresh Steamed
H-ZSM-5 1.78 0.30 [N2] 17
P/H-ZSM-5 1.51 0.59 [N2] 39
95 %Ag-ZSM-5 2.07 1.22 [O2] 59
194 %Cu-ZSM-5 2.27 2.36 [O2] 104
a) Data at 15 min on stream.
b) (Percent activity retained)＝(kc for steamed catalyst)/(kc for fresh catalyst)×100 %.

J. Jpn. Petrol. Inst., Vol. 61, No. 1, 2018

 18

ZSM-5. Therefore, all copper cations in Cu-ZSM-5 increase with higher acid site density of H-ZSM-5.
were Cu2＋ in the form of [Cu(OH)]＋ after the pretreat- 194 %Cu-ZSM-5 exhibits an ultra-high steaming stabil-
ment followed by oxidation, Cu2＋ was reduced to Cu＋ ity and higher cracking activity than H-ZSM-5, P/
at around 540 K, and finally Cu＋ was reduced to Cu0 H-ZSM-5 and Ag-ZSM-5, as Cu2＋ and Cu＋ cations in
by H2 at around 723 K as shown below: the 194 %Cu-ZSM-5 are not reduced to Cu0 at a high
temperature of 1023 K in the presence of O2.
ZO - [ Cu ( OH )]+ + 1 2H 2 Æ ZO - Cu + + H 2 O （1）
Acknowledgment
ZO - Cu + + 1 2H 2 Æ ZO - H + + Cu 0 （2）
We thank Professor T. Komatsu at Tokyo Institute of
where ZO – denotes the zeolite framework. If re- Technology for catalyst preparation and ICP measure-
oxidation of Cu0 is ideal, Cu0 is oxidized to [Cu(OH)]＋ ments. And we thank Professor J. N. Kondo and Mr.
by O2 as follows: H. Mochizuki of Tokyo Institute of Technology for their
technical assistance with the pyridine-IR measurements
ZO - H + + Cu 0 + 1 2O 2 Æ ZO - [ Cu ( OH )]+ (3)
and helpful discussion.
H2 consumption for reduction of silver cations at low
temperature was lower than that at high temperature References
and their sum was lower than the theoretical value for
95 %Ag-ZSM-5. This indicates that some silver cat- 1) Plotkin, J. S., Catal. Today, 106, 10 (2005).
2) Wan, J., Wei, Y., Liu, Z., Li, B., Qi, Y., Catal. Lett., 124, 150
ions were thermally reduced by the oxidation treatment (2008).
at 923 K. The H2 consumption at high temperature 3) Haag, W. O., Dessau, R. M., “Proceedings of the Eighth
was approximately half of the theoretical value. International Congress on Catalysis,” vol. II, Verlag Chemie,
Therefore, Ag2＋ was formed at around 473 K, and fi- Weinheim, Berlin (1984), p. 305.
nally Ag2＋ was reduced to silver metal by H2 at around 4) Haag, W. O., Dessau, R. M., Lago, R. M., Stud. Surf. Sci.
Catal., 60, 255 (1991).
653 K as shown below: 5) Haag, W. O., Lago, R. M., Weisz, P. B., Faraday Discuss.
ZO - Ag + + 1 2H 2 Æ ZO - H + + Ag 0 (4) Chem. Soc., 72, 317 (1981).
6) Babitz, S. M., Williams, B. A., Miller, J. T., Snurr, R. Q., Haag,
- + - +
Ag + ZO Ag Æ ZO Ag 2
0
(5) W. O., Kung, H. H., Appl. Catal. A: General, 179, 71 (1999).
7) Narbeshuber, T. F., Vinek, H., Lercher, J. A., J. Catal., 157,
ZO - Ag 2 + + 1 2H 2 Æ ZO - H + + Ag 2 0 (6) 388 (1995).
8) Kotrel, S., Rosynek, M. P., Lunsford, J. H., J. Phys. Chem. B,
0 0
If reoxidation of Ag2 is ideal, Ag2 is oxidized to 103, 818 (1999).
2Ag＋ by O2 as follows: 9) Komatsu, T., Ishihara, H., Fukui, Y., Yashima, T., Appl. Catal. A:
General, 214, 103 (2001).
ZO - H + + 1 2 Ag 2 0 Æ ZO - Ag + + 1 2H 2 (7) 10) Namba, S., Sato, K., Fujita, K., Kim, J. H., Yashima, T., Stud.
2＋ ＋ Surf. Sci. Catal., 28, 661 (1986).
We conclude that Cu and Cu cations in 194 %Cu- 11) Cerqueira, H. S., Mihindou-Koumba, P. C., Magnoux, P.,
ZSM-5 are not reduced to Cu0 at a high temperature of Guisnet, M., Ind. Eng. Chem. Res., 40, 1032 (2001).
1023 K in the presence of O2 and, therefore, 194 %Cu- 12) Slagtern, A., Dahl, I. M., Jens, K. J., Myrstad, T., Appl. Catal. A,
ZSM-5 exhibits ultra-high steaming stability in the 375, 213 (2010).
presence of O2. The Brønsted acidity of 95 %Ag- 13) Yoshimura, Y., Kijima, N., Hayakawa, T., Murata, K., Suzuki,
K., Mizukami, F., Matano, K., Konishi, T., Oikawa, T., Saito,
ZSM-5 is extremely low, almost equal to that of M., Shiojima, T., Shiozawa, K., Wakui, K., Sawada, G., Sato,
194 %Cu-ZSM-5. However, thermal reduction of Ag＋ K., Matsuo, S., Yamaoka, N., Catal. Surveys Jpn., 4, 157
occurs during steaming even in the presence of O2 at a (2000).
high temperature of 1023 K and, therefore, the steam- 14) Mochizuki, H., Yokoi, T., Imai, H., Watanabe, R., Namba, S.,
ing stability of 95 %Ag-ZSM-5 in the presence of O2 is Kondo, J. N., Tatsumi, T., Micropor. Mesopor. Mater., 145, 165
(2011).
much lower than that of 194 %Cu-ZSM-5. 15) Kubo, K., Iida, H., Namba, S., Igarashi, A., Micropor. Mesopor.
Mater., 149, 126 (2012).
6. Conclusion 16) Konno, H., Okamura, T., Kawahara, T., Nakasaka, Y., Tago, T.,
Masuda, T., Chem. Eng. J., 207-208, 490 (2012).
This study reviewed the reaction mechanism of the 17) Kubo, K., Iida, H., Namba, S., Igarashi, A., Catal. Commun.,
29, 162 (2012).
catalytic cracking of C6-8 hydrocarbons as naphtha 18) Mochizuki, H., Yokoi, T., Imai, H., Namba, S., Kondo, J. N.,
components over H-ZSM-5 catalysts at high tempera- Tatsumi, T., Appl. Catal. A: General, 449, 188 (2012).
ture to selectively form light olefins, and methods for 19) Konno, H., Tago, T., Nakasaka, Y., Ohnaka, R., Nishimura, J. I.,
enhancing the steaming stability of H-ZSM-5. High Masuda, T., Micropor. Mesopor. Mater., 175, 25 (2013).
temperature such as 873 to 923 K at which monomolec- 20) Nakasaka, Y., Okamura, T., Konno, H., Tago, T., Masuda, T.,
Micropor. Mesopor. Mater., 182, 244 (2013).
ular cracking progressed predominantly is desirable to 21) Kubo, K., Iida, H., Namba, S., Igarashi, A., Micropor. Mesopor.
obtain high ethylene and propylene yields. There may Mater., 188, 23 (2014).
be a distribution of acidity and the acid strength may

J. Jpn. Petrol. Inst., Vol. 61, No. 1, 2018

 19

22) Kubo, K., Takahashi, T., Iida, H., Namba, S., Igarashi, A., 30) Niwa, M., Katada, K., Catal. Surveys Jpn., 1, 215 (1997).
Appl. Catal. A: General, 482, 370 (2014). 31) Suzuki, K., Aoyagi, Y., Katada, N., Choi, M., Ryoo, R., Niwa,
23) Kubo, K., Iida, H., Namba, S., Igarashi, A., Appl. Catal. A: M., Catal. Today, 132, 38 (2008).
General, 489, 272 (2015). 32) Fujimoto, K., Maeda, K., Aimoto, K., Appl. Catal. A: General,
24) Wielers, A. F. H., Vaarkamp, M., Post, M. F. M., J. Catal., 127, 91, 81 (1992).
51 (1991). 33) Hattori, H., Top. Catal., 53, 432 (2010).
25) Nakasaka, Y., Tago, T., Konno, H., Okabe, A., Masuda, T., 34) Sultana, A., Nanba, T., Haneda, M., Sasaki, M., Hamada, H.,
Chem. Eng. J., 207-208, 368 (2012). Appl. Catal. B: Environmental, 101, 61 (2010).
26) Yashima, T., Sakaguchi, Y., Namba, S., Stud. Surf. Sci. Catal., 7, 35) Shibata, J., Shimizu, K., Takada, Y., Shichi, A., Yoshida, H.,
739 (1981). Satokawa, S., Satsuma, A., Hattori, T., J. Catal., 227, 367
27) Haag, W. O., Stud. Surf. Sci. Catal., 84, 1375 (1994). (2004).
28) Blasco, T., Corma, A., Martínez-Triguero, J., J. Catal., 237, 36) Ausavasukhi, A., Suwannaran, S., Limtrakul, J., Sooknoi, T.,
267 (2006). Appl. Catal. A: General, 345, 89 (2008).
29) Sano, T., Ikeya, H., Kasuno, T., Wang, Z. B., Kawakami, Y., 37) Bartolomeu, R., Henriques, C., Da Costa, P., Ribeiro, F., Catal.
Soga, K., Zeolites, 19, 80 (1997). Today, 176, 81 (2011).

要   旨

ZSM-5 系ゼオライト触媒上でのナフサ成分の接触分解による低級オレフィンの選択的合成

久保 光平，飯田  肇，難波 征太郎，五十嵐 哲

工学院大学工学部応用化学科，192-0015 東京都八王子市中野町2665-1

ナフサから低級オレフィンを選択的に得るために，723 ∼ た。一方，シクロヘキサンやメチルシクロヘキサンの共存は，
923 K で種々の Si/Al 比の H-ZSM-5 系ゼオライト触媒上で，ナ n-ヘ プ タ ン の 分 解 速 度 に か な り の 影 響 を 与 え る こ と か ら，
フサ成分である C6 ∼ 8 炭化水素の接触分解を行った。最初に， H-ZSM-5 細孔内でそれらの拡散速度が遅いことが示唆された。
923 K，H-ZSM-5(Si/Al＝31) 上で n-ヘプタンの接触分解を行っ さ ら に，H-ZSM-5 と 同 等 も し く は そ れ 以 上 の 触 媒 活 性 と
たところ，転化率 99.6 % で 59.7 C-% の高いエチレン＋プロピ H-ZSM-5 よりも高い耐スチーム性を有する触媒の開発につい
レン収率（プロピレンス／エチレン比＝約 0.72）が得られた。 て 検 討 し た と こ ろ，194 %Cu-ZSM-5 や 95 %Ag-ZSM-5 が
そして，生成物分布と活性化エネルギーから，923 K 以上の高 H-ZSM-5 や P/H-ZSM-5 に比べて耐スチーム性が高いことが明
温では単分子反応機構が支配的であることを明らかにした。次 らかとなった。特に，194 %Cu-ZSM-5 は，1023 K，10 h のスチー
に，2 成分系炭化水素の接触分解を行った。その結果，1-ヘキ ム 条 件 で 極 め て 高 い 耐 ス チ ー ム 性 を 示 し た。 こ の 理 由 は，
センの共存は，n-ヘプタン，シクロヘキサン，またはメチルシ 194 %Cu-ZSM-5 中の Cu2 ＋および Cu＋が，1023 K，酸素共存下
クロヘキサンの分解速度に影響を与えないことから，923 K 以 において Cu0 に還元されないためである。
上の高温では単分子反応機構が支配的であることが支持され

J. Jpn. Petrol. Inst., Vol. 61, No. 1, 2018