Fluoride-treated H-ZSM-5 as a highly selective and stable catalyst for the production of propylene from methyl halides

Article abstract of DOI:10.1016/j.jcat.2012.08.014

Among several typical zeolites, H-ZSM-5 was found to be a promising catalyst for the conversion of methyl halides (CH₃Cl and CH ₃Br) into propylene. The increase in Si/Al ratio or Na⁺ exchange degree in ZSM-5 increased the selectivity of propylene but decreased the conversion of methyl halides. The treatment of H-ZSM-5 with ammonium fluoride followed by calcination significantly improved its catalytic performance. With a proper concentration of fluoride, not only the propylene selectivity but also the catalyst stability could be enhanced significantly. We have demonstrated that the acidity and the pore structure are two crucial factors determining the catalytic behaviors. The weaker acidity of the fluoride-treated H-ZSM-5 suppressed the hydrogen-transfer and aromatization reactions, leading to higher selectivity to light olefins. Larger micropores with sizes of 0.73-0.78 nm, which were generated after the fluoride treatment, changed the distribution of methylbenzenes in the "hydrocarbon pool" over catalyst and contributed to higher propylene selectivity.

Full text of DOI:10.1016/j.jcat.2012.08.014

Journal of Catalysis 295 (2012) 232–241
Contents lists available at SciVerse ScienceDirect
Journal of Catalysis
journal homepage: www.elsevier.com/locate/jcat
Fluoride-treated H-ZSM-5 as a highly selective and stable catalyst for the
production of propylene from methyl halides
⇑
⇑
Ting Xu, Qinghong Zhang , Hang Song, Ye Wang
State Key Laboratory of Physical Chemistry of Solid Surfaces, National Engineering Laboratory for Green Chemical Productions of Alcohols, Ethers and Esters,
College of Chemistry and Chemical Engineering, Xiamen University, Xiamen 361005, PR China
a r t i c l e i n f o
a b s t r a c t
Article history:
Among several typical zeolites, H-ZSM-5 was found to be a promising catalyst for the conversion of
methyl halides (CH3Cl and CH3Br) into propylene. The increase in Si/Al ratio or Na+ exchange degree in
ZSM-5 increased the selectivity of propylene but decreased the conversion of methyl halides. The treat-
ment of H-ZSM-5 with ammonium fluoride followed by calcination significantly improved its catalytic
performance. With a proper concentration of fluoride, not only the propylene selectivity but also the
catalyst stability could be enhanced significantly. We have demonstrated that the acidity and the pore
structure are two crucial factors determining the catalytic behaviors. The weaker acidity of the fluo-
ride-treated H-ZSM-5 suppressed the hydrogen-transfer and aromatization reactions, leading to higher
selectivity to light olefins. Larger micropores with sizes of 0.73–0.78 nm, which were generated after
the fluoride treatment, changed the distribution of methylbenzenes in the ‘‘hydrocarbon pool’’ over cat-
alyst and contributed to higher propylene selectivity.
Received 20 June 2012
Revised 14 August 2012
Accepted 18 August 2012
Available online 23 September 2012
Keywords:
Propylene
Methyl halides
H-ZSM-5
Fluoride treatment
Acidity
Pore structure
Ó 2012 Elsevier Inc. All rights reserved.
1. Introduction
to propylene would provide an important alternative to the con-
ventional naphtha-based propylene route. Propylene may be pro-
Propylene is one of the most important bulk chemicals for the
production of polypropylene, acrylonitrile, and propylene oxide.
Currently, propylene is mainly produced as a co-product of ethyl-
ene via steam cracking of naphtha. The demand for propylene is
growing quickly in recent years because of the increased needs
for propylene derivatives such as polypropylene. On the other
hand, the depletion of petroleum resources requires a decreased
reliance on petroleum. Therefore, the development of non-petro-
leum routes for propylene production has attracted much attention
in recent years.
Several reaction routes including the dehydrogenation or oxida-
tive dehydrogenation of propane [1,2], the metathesis of ethylene
and butenes [1,3], the conversion of either ethylene [4–7] or bu-
tenes [8–10], and the conversion of methanol [11–13] have been
proposed as alternatives for the production of propylene. The costs
of the raw materials such as propane, ethylene, and butenes largely
influence the economical feasibility of these routes [1].
duced from methane via synthesis gas (H2 + CO) through a three-
step route, that is, (i) the high-temperature steam reforming of
methane to produce syngas, (ii) methanol synthesis from syngas,
and (iii) the conversion of methanol to propylene. However, the
steam reforming of methane is an energy-intensive and high-cost
process [14]. A large number of studies have been devoted to the
direct conversion of methane to high-valued chemicals such as
methanol, formaldehyde, ethylene, and benzene [14–20]. How-
ever, yields of target products are still low for most of these direct
conversion reactions and none of them meet the requirements for
commercialization.
Methyl halides (CH3Cl or CH3Br) may also work as platform
molecules for the transformation of CH4 to chemicals. Olah et al.
reported that CH4 could be converted to CH3Cl or CH3Br by Cl2 or
Br2 with high selectivities over supported superacids or noble met-
als and that the subsequent hydrolysis of CH3Cl or CH3Br over
Al2O3-supported metal oxides or metal hydroxides such as ZnO/
Al(OH)3/ -Al2O3 could produce methanol and dimethyl ether
c-
Methane is the main constituent of the vast natural gas, coal-
bed gas, shale gas, landfill gas, biogas, and methane hydrates.
The transformation of the abundant and cheap methane resources
c
[21]. Zhou et al. showed that methyl or ethyl bromide, which
was formed by the reaction of CH4 or C2H6 with Br2, could be con-
verted to C2H4 or oxygenates through the reaction with metal oxi-
des, and metal bromides were formed at the same time [22]. For
these systems, the generated HCl, HBr, or metal bromides must
be oxidized to Cl2 or Br2 to complete the catalytic cycle. Thus,
the Deacon process has to be incorporated.
⇑
Corresponding authors. Fax: +86 592 2183047 (Y. Wang).
E-mail addresses: zhangqh@xmu.edu.cn (Q. Zhang), wangye@xmu.edu.cn
(Y. Wang).
0021-9517/$ - see front matter Ó 2012 Elsevier Inc. All rights reserved.
http://dx.doi.org/10.1016/j.jcat.2012.08.014

T. Xu et al. / Journal of Catalysis 295 (2012) 232–241
233
Very recently, we reported a two-step route for the conversion
of CH4 to C3H6 via CH3Cl or CH3Br [23]. The reactions in the two
steps are shown as follows.
dryness, and the obtained solid product was calcined in air at
873 K for 6 h. The fluoride-treated H-ZSM-5 sample was denoted
as H-ZSM-5-F-x M, where x was the concentration of NH4F used
for treating H-ZSM-5.
Step 1 : CH4 þ HClðHBrÞ þ 1=2O2 ! CH3Cl ðCH3BrÞ þ H2O
Step 2 : CH3ClðCH3BrÞ ! 1=3C3H6 þ HCl ðHBrÞ
Net : CH4 þ 1=2O2 ! 1=3C3H6 þ H2O
ð1Þ
ð2Þ
ð3Þ
2.2. Catalyst characterization
The catalysts were characterized by various techniques. Induc-
tively coupled plasma mass spectrometry (ICP-MS) and X-ray fluo-
rescence (XRF) spectroscopy were used to detect the composition
of catalysts. The ICP-MS analysis was performed on an Agilent ICP-
MS 4500 instrument, while the XRF measurements were carried
out on a Panalytical Axois Petro XRF instrument with rhodium target
(50 kV, 50 mA). Powder X-ray diffraction (XRD) and scanning elec-
tron microscopy (SEM) were employed to measure the crystalline
structure and the morphology of catalysts. XRD patterns were col-
The HCl or HBr generated in the second step could be directly used
in the first step, and there is no need of the expensive Deacon pro-
cess for the re-oxidation of HCl or HBr to Cl2 or Br2. The net reaction
of this two-step route is the oxidative dehydrogenation of CH4 to
C3H6, that is, Eq. (3).
The development of efficient catalysts for both steps is crucial
for this novel two-step route. Several research groups have con-
tributed to develop efficient catalysts such as Ru/SiO2 [24], Rh/
SiO2 [25], FePO4/SiO2 [26], LaOCl [27–29], and CeO2 [23] for the
first-step reaction, that is, the oxidative chlorination or bromina-
tion of CH4 to CH3Cl or CH3Br by HCl or HBr in the presence of
O2. The present work focuses on studying the second step, that
is, the catalytic transformation of CH3Cl or CH3Br to C3H6.
There are some publications on the conversion of methyl ha-
lides to hydrocarbons including light olefins [30–39]. Zeolite H-
SAPO-34 has mainly been employed as a catalyst for the formation
of light olefins in these studies [33–39]. H-SAPO-34 is known to be
an efficient catalyst for the conversion of methanol to olefins
(MTO) [11–13,40]. It can be expected that H-SAPO-34 may also
be promising for the conversion of CH3Cl or CH3Br to light olefins
because of the chemical similarities between methyl halides and
CH3OH. However, over H-SAPO-34, the fraction of C3H6 in light ole-
fins was not high, and in most cases, the selectivity of C3H6 (20–
45%) was lower than that of C2H4 [33–39]. Moreover, H-SAPO-34
underwent quick deactivation in the conversion of methyl halides.
On the other hand, the studies using H-ZSM-5 for the conversion
of methyl halides to propylene are scarce. H-ZSM-5 is known as an
efficient catalyst for the methanol-to-gasoline (MTG) process, and it
can also be modified for the methanol-to-propylene (MTP) process
[13,41]. Olsbye et al. once showed that the main products for the con-
version of CH3Cl over H-ZSM-5 were C2–C4 paraffins and C5+ aromat-
ics [38]. Proper modifications may be required for obtaining better
selectivities of C2–C4 olefins. In a recent communication, we have
demonstrated that the treatment of H-ZSM-5 by ammonium fluoride
followed by calcination is a very efficient means to improve the selec-
tivity of C3H6 [23]. This paper reports our detailed studies on the ef-
fects of fluoride treatment on the structure, acidity, and catalytic
behaviors of H-ZSM-5 for the conversions of CH3Cl and CH3Br. The
key factors determining the catalytic behaviors, the roles of fluoride
treatment, and the possible reaction mechanism are also discussed.
lected on a Panalytical X’pert Pro diffractometer using Cu Ka radia-
tion (40 kV, 30 mA). SEM measurements were performed on a
Hitachi S-4800 operated at 15 kV. Argon physisorption, which was
used to gain information about the pore structure of catalysts, was
performed on an ASAP2010M Micromeritics apparatus. The pore size
distribution was evaluated by the Horváth–Kawazoe (HK) method.
NH3 temperature-programmed desorption (NH3-TPD) was per-
formed on a Micromeritics AutoChem 2920 II instrument for mea-
suring the acidity. Typically, the sample loaded in a quartz reactor
was first pretreated with high-purity He at 673 K for 1 h. After the
sample was cooled to 373 or 573 K, NH3 adsorption was performed
by switching the He flow to a NH3–He (10 vol.% NH3) gas mixture
and then keeping at 373 or 573 K for 1 h. Then, the gas phase or
the weakly adsorbed NH3 was purged by high-purity He at the same
temperature. NH3-TPD was performed in the He flow by raising the
temperature to 1073 K at a rate of 10 K minꢀ1, and the desorbed
NH3 molecules were detected by ThermoStar GSD 301 T2 mass spec-
trometer with the signal of m/e = 16. A similar procedure was
adopted for measuring the adsorption amounts of p-xylene and o-xy-
lene. Temperature-programmed oxidation (TPO) measurements
were also performed on the Micromeritics AutoChem 2920 II instru-
ment. The catalyst after catalytic reactions was first pretreated in
pure He gas flow at 673 K for 0.5 h. Then, TPO was performed by
switching the He flow to an O2–He (5 vol.% O2) flow after the temper-
ature was decreased to 373 K. The temperature was raised to 1073 K
at a rate of 10 K minꢀ1, and the formed CO2 and CO were detected by
the mass spectrometer with signals of m/e = 44 and m/e = 28.
Fourier-transform infrared (FT-IR) and 27Al magic-angle spin-
ning nuclear magnetic resonance (27Al MAS NMR) were applied
to study the acidic property and the coordination environment of
Al. FT-IR measurements were performed with a Nicolet Avatar
instrument equipped with an MCT detector with a resolution of
4 cmꢀ1. For the measurements of pyridine-adsorbed FT-IR, the
sample was pressed into a self-supported wafer and placed in an
in situ FT-IR cell. After pretreatment under vacuum at 673 K for
30 min, the sample was cooled down to 473 K. Then, pyridine
was adsorbed at 473 K onto the sample for a sufficient time. FT-
IR spectra were recorded after gaseous or weakly adsorbed pyri-
dine molecules were removed by evacuation at 473 K. 27Al MAS
NMR experiments were conducted on a Varian Infinity plus
AS400 spectrometer (Varian Inc.) operated with frequency at
2. Experimental
2.1. Catalyst preparation
Zeolites including H-ZSM-5 with different Si/Al ratios were pur-
chased from Nankai University Catalyst Co. ZSM-5 (Si/Al = 100)
samples with different Na+ exchange degrees were prepared by
the ion exchange technique using H-ZSM-5 and an aqueous solu-
tion of 0.03 M NaNO3 at 343 K.
The H-ZSM-5 with a Si/Al ratio of 100 was subjected to fluoride
treatment by the following procedure. H-ZSM-5 (typically 2.0 g)
was first immersed into an aqueous solution (typically 40 mL) of
NH4F. The concentration of NH4F in the aqueous solution was var-
ied in a range of 0–0.104 mol dmꢀ3 (M). After being stirred for 10 h
at room temperature, the suspension was evaporated at 343 K to
104.26 MHz, pulse width at 0.5 ls, radiofrequency field strength
at 50 G, pulse delay at 0.5 s, spinning rate at 7 kHz, and with
85,000 scans. The samples used for 27Al MAS NMR measurements
did not undergo hydration preliminarily.
2.3. Catalytic reaction
The catalytic conversion of CH3Cl or CH3Br was carried out on a
fixed-bed tubular quartz reactor operated at atmospheric pressure.

234
T. Xu et al. / Journal of Catalysis 295 (2012) 232–241
Table 1
The catalyst loaded in the reactor was pretreated in high-purity N2
(purity >99.999%) gas flow at 673 K for 1 h before reaction. Then,
the reaction was started by introducing CH3Cl or CH3Br (purity
>99.99%) diluted with high-purity N2. All the products were ana-
lyzed by on-line gas chromatography with TCD detector using
Porapak T, RT PLOTQ and Porapak Q columns. The performance
typically after 2 h of reactions was used for discussion unless
otherwise mentioned. Over our catalysts, the carbon-containing
products included CH4, C2–C4 paraffins (including C2H6, C3H8, and
C4H10), C2–C4 olefins (including C2H4, C3H6, and C4H8), and C5+
hydrocarbons. The selectivity of a given product was calculated
from the amount (mol) of the product divided by the total amount
(mol) of the CH3Cl or CH3Br converted on a carbon basis. The car-
bon balance without considering carbon deposition was typically
better than 90%.
Catalytic performances of typical zeolites for the conversions of methyl halidesa.
Catalystb
Conv.
(%)
Selectivityc (%)
C3H6
yield (%)
C02—C04
CH4 C2H4 C3H6 C4H8
C5+
(A) CH3Cl conversion
H-SAPO-34 50
H-SAPO-11 <1
H-ZSM-5
(100)
H-MOR
(12)
H-beta (50) 26
H-MCM-22 16
(30)
1.1
–
46
–
32
–
0.5
–
19
–
0.9 16
–
Trace
89
1.4
18
16
5.6
37
22
14
8.7
1.9
23
26
25
7.8
16
2.3
0.4
1.0
20
10
31
37
20
28
16
12
12
12
8.3
5.9
H-Y (4.8)
24
0.5
12
17
19
20
30
4.1
(B) CH3Br conversion
H-SAPO-34 43
H-SAPO-11 Trace
1.1
–
54
–
35
–
5.1
–
4.7
–
0.1 15
–
Trace
H-ZSM-5
(100)
99
1.0
12
21
8.0
34
24
21
3. Results and discussion
H-MOR
(12)
H-beta (50) 23
H-MCM-22 12
(30)
5.9
30
39
11
0
2.0
18
0.7
3.1. Catalytic performances of typical H-form zeolites for the
conversion of methyl halides
9.5
8.1
8.4
15
32
49
9.1
19
13
7.9
28
7.4
1.0 5.7
14 2.3
P(CH3Cl) = 3.3 kPa,
Catalytic performances of several typical H-form zeolites for the
conversions of CH3Cl and CH3Br are shown in Table 1. H-SAPO-34
and H-ZSM-5 displayed significantly better performances, provid-
ing C3H6 yields of P14% for both reactions. H-MCM-22 and H-beta
exhibited higher selectivities to light olefins, but their activities
were lower. H-SAPO-11, which could catalyze the MTO reaction
with a good efficiency [42], was almost inactive for the conversions
of methyl halides under our reaction conditions. H-MOR and HY
showed lower yields to C3H6 due to the lower selectivities and/or
lower activities.
H-Y (4.8)
19
24
25
12
1.4
22
a
Reaction
conditions:
(A)
T = 673 K,
W = 0.30 g,
F(total) = 15.5 mL minꢀ1
, t = 2 h; (B) T = 673 K, W = 0.15 g, P(CH3Br) = 9.2 kPa,
F(total) = 11 mL minꢀ1, t = 2 h.
b
The number in parenthesis denotes the Si/Al ratio.
C20—C04 and C5+ denote C2–C4 paraffins and C5+ hydrocarbons.
c
We compared H-ZSM-5 and H-SAPO-34 in more detail for the
conversions of CH3Cl and CH3Br. H-ZSM-5 provided higher conver-
sions of both CH3Cl and CH3Br but lower selectivities to C3H6. The
yields of C3H6 over the two catalysts were similar to each other for
the conversion of CH3Cl (14% and 16%), while for the conversion of
CH3Br, H-ZSM-5 exhibited a higher C3H6 yield (21% vs. 15%). We
further investigated the stability of the two catalysts in the conver-
sion of CH3Cl. As displayed in Fig. 1, H-ZSM-5 could keep the C3H6
yield of 14% for ꢁ10 h and then deactivated gradually. On the other
hand, H-SAPO-34 deactivated more quickly, and C3H6 yield became
<5% at a time on stream of 7 h. In other words, H-ZSM-5 was more
stable than H-SAPO-34. As described above, most of the present
studies on the conversion of methyl halides to light olefins are
focusing on H-SAPO-34 [33–39]. Only a few scattered studies have
been devoted to ZSM-5-catalyzed conversions of methyl halide.
Therefore, we have chosen H-ZSM-5 for further studies. Clearly,
the increase in the selectivity of C3H6 is one of the most important
tasks for H-ZSM-5-based catalysts.
Fig. 1. Changes of catalytic performances with time on stream in the conversion of
CH3Cl over H-SAPO-34 and H-ZSM-5. Reaction conditions: T = 673 K, W = 0.30 g,
P(CH3Cl) = 3.3 kPa, F(total) = 15.5 mL minꢀ1
.
respectively. The selectivities to C2H4 and C4H8 also showed a trend
of rising with increasing Si/Al ratio up to 260, while those to C2–C4
paraffins (C0—C0) and C5+ hydrocarbons (hydrocarbons with car-
3.2. Effects of Si/Al ratio and Na+ exchange degree on catalytic
performances of ZSM-5 for the conversion of methyl halides
2
4
bon numbers P5) decreased significantly. The highest yield of
C3H6 was obtained at a Si/Al ratio of 260 for each reaction, and
the values were 22% and 29% for the conversions of CH3Cl and CH3-
Br, respectively.
The Si/Al ratio significantly affected the selectivity of H-ZSM-5
in the MTP reaction [41]. Lurgi employed highly siliceous H-ZSM-
5 for the MTP process [13]. Here, we examined the effect of Si/Al
ratio on catalytic performances of H-ZSM-5 for the conversion of
methyl halides. The results are shown in Table 2. Similar trends
were observed for both reactants. The increase in Si/Al ratio or
the decrease in Al content gradually decreased the conversion of
CH3Cl or CH3Br. Almost no conversion or the conversion was very
low over the purely siliceous zeolite. The product selectivity also
changed with Si/Al ratio. With increasing Si/Al ratio, the selectivity
of C3H6 increased steadily for both reactions, reaching 44% and 52%
at a Si/Al ratio of 260 for the conversions of CH3Cl and CH3Br,
Similar to the Si/Al ratio, the Na+ exchange degree is another
important parameter influencing the acidity of ZSM-5. We have
examined the effect of Na+ exchange degree on catalytic behaviors
of ZSM-5 with a Si/Al ratio of 100 for the conversion of CH3Cl. The
results in Table 3 show that, with increasing Na+ exchange degree,
the conversion of CH3Cl decreases, while the selectivity of C3H6 in-
creases. At the same time, the selectivity of C4H8 also increased and
those of C2–C4 paraffines and C5+ hydrocarbons decreased. For this
series of catalysts, the highest yield of C3H6 (26%) was obtained
over the ZSM-5 with a Na+ exchange degree of 61%. These observa-

T. Xu et al. / Journal of Catalysis 295 (2012) 232–241
235
Table 2
Table 4
Catalytic performances of H-ZSM-5 with different Si/Al ratios for the conversions of
Effect of the concentration of NH4F used for treating H-ZSM-5 on the catalytic
methyl halidesa.
performances for the conversions of methyl halidesa.
Si/Al
ratio
Conversion
(%)
Selectivityb (%)
C3H6
yield (%)
NH4F
conc.
(M)
Conversion
(%)
Selectivityb (%)
C3H6
yield
(%)
C02—C04
C02—C04
CH4 C2H4 C3H6 C4H8
C5+
CH4 C2H4 C3H6 C4H8
C5+
(A) CH3Cl conversion
(A) CH3Cl conversion
50
92
89
81
70
50
26
0
1.0
1.4
1.7
1.0
2.2
4.8
–
15
18
19
20
20
21
–
13
16
21
30
44
51
–
5.0
5.6
6.2
13
20
20
–
41
37
34
20
7.1
1.8
–
25
22
18
16
6.1 22
1.3 13
12
15
17
21
0
89
83
89
76
68
55
29
1.4
1.5
1.0
1.2
1.4
1.3
1.5
18
20
12
4.8
3.5
3.0
2.7
16
27
46
64
67
67
60
5.6
9.5
20
23
23
26
29
37
26
22
16
12
6.0 49
4.5 45
1.9 37
6.5 17
14
22
41
100
140
200
260
300
1
0.026
0.039
0.052
0.065
0.078
0.104
8.1
1.0
0.6
0.5
0.3
–
–
(B) CH3Br conversion
(B) CH3Br conversion
50
93
99
98
89
56
18
4.0
2.8
1.2
1.8
1.0
3.0
17
11
12
12
15
15
14
3.8
17
21
23
33
52
54
41
5.2
8.5
8.2
15
21
13
28
41
34
36
25
7.8
0.6
0
23
23
19
11
1.2 29
1.4 9.8
16
21
22
29
0
92
98
98
94
87
36
0
1.2
0.9
1.0
1.3
1.0
3.5
–
13
11
7.0
4.2
3.0
1.8
–
30
38
46
56
59
63
–
13
20
25
23
25
30
–
23
11
5.0
1.5
0.1
0
19
18
16
14
12
27
38
45
53
51
100
140
200
260
300
1
0.026
0.039
0.052
0.065
0.078
0.104
1.7 23
–
27
0
1.1
–
a
Reaction
conditions:
(A)
T = 673 K,
W = 0.30 g,
P(CH3Cl) = 3.3 kPa,
a
Reaction
conditions:
(A)
T = 673 K,
W = 0.30 g,
P(CH3Cl) = 3.3 kPa,
F(total) = 15.5 mL minꢀ1
, t = 2 h; (B) T = 673 K, W = 0.15 g, P(CH3Br) = 9.2 kPa,
F(total) = 15.5 mL minꢀ1
, t = 2 h; (B) T = 673 K, W = 0.10 g, P(CH3Br) = 9.2 kPa,
F(total) = 11 mL minꢀ1, t = 2 h.
F(total) = 11 mL minꢀ1, t = 2 h.
C20—C04 and C5+ denote C2–C4 paraffins and C5+ hydrocarbons.
b
b
C20—C04 and C5+ denote C2–C4 paraffins and C5+ hydrocarbons.
Table 3
Catalytic performances of ZSM-5 catalysts with different Na+ exchange degrees for the
conversion of CH3Cla.
Na+ exchange Conv.
Selectivityc (%)
C3H6
yield
(%)
degreeb (%)
(%)
C02—C04
CH4 C2H4 C3H6 C4H8
C5+
0
28
61
89
71
56
34
1.4
1.6
1.7
1.6
18
17
15
13
16
24
46
57
5.6
9.9
15
37
29
12
2.8
22
18
10
14
17
26
100
22
3.6 19
a
Reaction
conditions:
T = 673 K,
W = 0.3 g,
P(CH3Cl) = 3.3 kPa,
F(total) = 15.5 mL minꢀ1, t = 2 h.
b
Na+ content was measured by ICP.
C20—C04 and C5+ denote C2–C4 paraffins and C5+ hydrocarbons.
c
tions further confirm that the conversion of CH3Cl decreases with
decreasing acidity, but a proper acidity is important for obtaining
high selectivity and high yield of C3H6.
3.3. Effect of the treatment of H-ZSM-5 by fluoride on the conversion of
methyl halides
Fig. 2. Changes of catalytic behaviors with time on stream in the conversions of
CH3Cl (A) and CH3Br (B) over H-ZSM-5-F-0.052 M catalyst. Reaction conditions: (A)
T = 673 K, W = 0.30 g, P(CH3Cl) = 3.3 kPa, F(total) = 15.5 mL minꢀ1
; (B) T = 673 K,
W = 0.10 g, P(CH3Br) = 9.2 kPa, F(total) = 11 mL minꢀ1
.
We found that the treatment of H-ZSM-5 in NH4F aqueous solu-
tion followed by calcination was a very effective means to improve
the selectivity of C3H6. Table 4 shows the effect of the concentra-
tion of NH4F used for treating H-ZSM-5 on catalytic performances
for the conversions of CH3Cl and CH3Br. The increase in the concen-
tration of NH4F significantly increased the selectivity of C3H6 in
both reactions. When the concentration of NH4F reached
0.052 M, the selectivities of C3H6 reached 64% and 56% for the con-
versions of CH3Cl and CH3Br, respectively. These values were about
four times and twice higher than those over H-ZSM-5 without fluo-
ride treatment. The increase in the concentration of NH4F also in-
creased the selectivity of C4H8. At the same time, the selectivities
to C2–C4 paraffins, C2H4, and C5+ hydrocarbons decreased. The in-
crease in the concentration of NH4F in a range of 0–0.052 M only
slightly varied the conversion of CH3Cl or CH3Br, but a further in-
crease in the concentration of NH4F would cause a significant de-
crease in the conversions of both reactants. The catalyst treated
by fluoride with a concentration of 0.052 M (H-ZSM-5-F-0.052 M)
afforded the best yields of C3H6 in both reactions; C3H6 yields of
49% and 53% were attained with CH3Cl and CH3Br selectivities of
64% and 56% for the conversions of CH3Cl and CH3Br, respectively.
To the best of our knowledge, these excellent C3H6 selectivities and
yields have never been reported for the conversions of haloge-
nomethanes before our work.
The deactivation of catalyst was observed in H-SAPO-34-cata-
lyzed conversions of methyl halides [33–39]. We found that H-
ZSM-5 also deactivated after 10 h of reaction during the conversion
of CH3Cl, although it was more stable than H-SAPO-34 (Fig. 1).
After 22 h of reaction, CH3Cl conversion and C3H6 yield decreased
from 88% to 40% and from 16% to 8.3%, respectively, over H-ZSM-
5 with a Si/Al of 100. We demonstrated that the treatment of

236
T. Xu et al. / Journal of Catalysis 295 (2012) 232–241
H-ZSM-5 by fluoride could significantly improve the catalyst stabil-
ity. Fig. 2 displays that the selectivity of C3H6 kept almost unchanged
with prolonging the time on stream over the H-ZSM-5-F-0.052 M
catalyst. After 50 h of reaction, CH3Cl conversion and C3H6 yield de-
creased slightly from 75% to 66% and from 49% to 41%, respectively
(Fig. 2A). For the conversion of CH3Br, CH3Br conversion and C3H6
yield decreased from 94% to 88% and from 52% to 50%, respectively
(Fig. 2B). Thus, the fluoride-treated H-ZSM-5 was highly stable for
the transformations of both CH3Cl and CH3Br into C3H6.
To gain information about the effect of the fluoride treatment
on reaction pathways, we have compared the catalytic behaviors
of H-ZSM-5 and the fluoride-treated H-ZSM-5 (H-ZSM-5-F-
0.052 M) for the conversion of CH3Cl at different contact times.
The conversion of CH3Cl increased almost linearly with the contact
time over both catalysts (Fig. 3A). The fluoride-treated H-ZSM-5
showed a slightly lower CH3Cl conversion than H-ZSM-5 at the
same contact time. At shorter contact times, that is, the initial reac-
tion stage, C3H6 was formed as a main product together with C2H4
and C4H8 over H-ZSM-5 (Fig. 3B–D). The increase in contact time
decreased the selectivities to C3H6, C2H4, and C4H8 and increased
those to C2–C4 paraffins (Fig. 3E) and C5+ hydrocarbons (Fig. 3F)
over H-ZSM-5. These results suggest that, over H-ZSM-5, the light
olefins, that is, C2H4, C3H6 and C4H8, are the primary products,
whereas the C2–C4 paraffins and C5+ hydrocarbons (including aro-
matics) are the secondary products formed from light olefins prob-
ably via hydrogen-transfer and/or aromatization reactions.
the initial reaction stage (short contact times) showed that the
selectivities to C3H6 and C4H8 over the fluoride-treated H-ZSM-5
were higher than those over H-ZSM-5 (Fig. 3B and D), whereas
the selectivity to C2H4 became lower after fluoride treatment
(Fig. 3C). Thus, the fluoride treatment also changed the distribution
of light olefins at the initial reaction stage, increasing the fractions
of C3H6 and C4H8 and decreasing that of C2H4.
We have investigated the carbon deposition on H-ZSM-5 and
the fluoride-treated H-ZSM-5 catalysts by the TPO technique. The
TPO profiles for H-ZSM-5 and H-ZSM-5-F-0.052 M after the conver-
sion of CH3Cl for 32 h at 673 K are shown in Fig. 4. For H-ZSM-5,
two larger CO2 (m/e = 44) formation peaks were observed at
ꢁ703 K and ꢁ890 K. The formation of CO (m/e = 28) occurred al-
most simultaneously with that of CO2. After the treatment by fluo-
ride, the main peaks for CO and CO2 formations were observed at
ꢁ800 K, and the intensity became significantly lower. It was re-
ported that the temperatures for COx formations in the TPO profile
may be related to the nature of the carbon species with different H/
C ratios and their locations over the catalyst [43,44]. Quantitative
calculations revealed that the amount of carbon species deposited
over H-ZSM-5 (22.5 mmol gꢀ1) was ꢁ5 times larger than that over
H-ZSM-5-F-0.052 M (4.1 mmol gꢀ1). The lower carbon deposition
rate may result in the higher stability of the fluoride-treated H-
ZSM-5 catalyst (Fig. 2).
3.4. Characterizations of fluoride-treated H-ZSM-5
On the other hand, over the fluoride-treated H-ZSM-5, the de-
crease in the selectivity to C3H6 as well as C2H4 and C4H8 became
insignificant with increasing contact time (Fig. 3B–D). The selectiv-
ities to C2–C4 paraffins and C5+ hydrocarbons became remarkably
lower over the fluoride-treated H-ZSM-5 even at longer contact
times (Fig. 3E and F). This clearly indicates that the fluoride treat-
ment of H-ZSM-5 suppresses the hydrogen-transfer and the aro-
matization reactions. Further comparisons of the two catalysts at
The contents of Al and F in the fluoride-treated H-ZSM-5 sam-
ples were measured by ICP-MS and XRF techniques. Both tech-
niques showed that the content of Al increased with increasing
concentration of NH4F used for treating H-ZSM-5, although there
were small differences between the values obtained from the
two techniques (Table 5). This suggests that the fluoride treatment
mainly causes desilication in the zeolite. It is quite unexpected that
Fig. 3. Effect of contact time on catalytic behaviors for the conversion of CH3Cl over H-ZSM-5 and H-ZSM-5-F-0.052 M catalysts. (A) CH3Cl conversion, (B) C3H6 selectivity, (C)
C2H4 selectivity, (D) C4H8 selectivity, (E) C2–C4 paraffin selectivity, (F) C5+ selectivity. Reaction conditions: T = 673 K, W = 0.30 g, P(CH3Cl) = 3.3 kPa, F(total) = 15.5 mL minꢀ1
.

T. Xu et al. / Journal of Catalysis 295 (2012) 232–241
237
Fig. 5. XRD patterns for the H-ZSM-5 samples treated by fluoride with different
concentrations.
Fig. 4. TPO profiles for H-ZSM-5 and H-ZSM-5-F-0.052 M catalysts after the
conversion of CH3Cl for 32 h. Reaction conditions: T = 673 K, W = 0.30 g, P(CH3-
Cl) = 3.3 kPa, F(total) = 15.5 mL minꢀ1
.
presence of larger micropores became distinct in the samples trea-
ted by fluoride with concentration of P0.052 M.
Table 5
The contents of Al and F in the fluoride-treated H-ZSM-5 samples.
To confirm the bimodal pore structure of the fluoride-treated
samples, we measured the adsorption amounts of p-xylene and
o-xylene, which possess kinetic diameters of 0.58 and 0.68 nm,
respectively [45]. It is known that the adsorption amount of the
former molecule in H-ZSM-5 is much larger than that of the latter
molecule because of the shape selectivity of H-ZSM-5 [46]. We
have measured the adsorption amounts of p-xylene and o-xylene
at 323 K in H-ZSM-5 and the fluoride-treated H-ZSM-5 (H-ZSM-
5-F-0.052 M). Fig. 8 confirms that, in the case of H-ZSM-5, the
adsorption amount of o-xylene is remarkably lower than that of
p-xylene. The treatment of H-ZSM-5 by fluoride significantly in-
creased the adsorption amount of the larger o-xylene molecule.
This further supports the result obtained from argon adsorption
study that larger micropores are generated after the fluoride treat-
ment. Considering the possible location of fluoride anions in the
channels of H-ZSM-5 [47], we speculate that the fluoride anion
might react with the nearby framework Si atom, and the removal
of a Si atom may create a distorted 14-membered ring in the struc-
ture of H-ZSM-5. This may explain the generation of the larger
micropores with sizes around 0.73–0.78 nm.
Fig. 9 shows the NH3-TPD profiles for the H-ZSM-5 samples
with and without fluoride treatment. For H-ZSM-5, two NH3
desorption peaks were observed at 473 and 683 K (Fig. 9A). These
two peaks are typically attributed to the NH3 molecules from the
strong Brønsted acid sites and the weakly held (probably hydro-
gen-bonded) NH3 molecules or NH3 molecules from the weak Le-
wis acid sites [48]. The intensities of these peaks became lower
for the fluoride-treated H-ZSM-5 samples (Fig. 9A). It is well known
that only NH3 molecules adsorbed on strong acid sites can remain
during NH3 adsorption at 573 K followed by He purge [8,48]. Thus,
we have also investigated the strong acid sites by NH3-TPD with
NH3 adsorption at 573 K followed by purging with He at the same
temperature. From the result shown in Fig. 9B, it becomes clear
that the amount of the strong acid sites decreases gradually with
increasing concentration of NH4F used for treating H-ZSM-5. At
an NH4F concentration of 0.078 M, a small amount of acid sites still
remained, but no acid sites could be observed when the concentra-
tion of NH4F was raised to 0.104 M.
Sample
Al content (wt.%)
F content (wt.%)
XRF
ICP
XRF
H-ZSM-5
0.45
0.66
0.82
0.99
1.18
0.44
0.56
0.73
0.89
0.99
0
0
0
0
0
H-ZSM-5-F-0.026 M
H-ZSM-5-F-0.052 M
H-ZSM-5-F-0.078 M
H-ZSM-5-F-0.104 M
no fluorine can be detected in any of the fluoride-treated H-ZSM-5
samples. This was also confirmed by XPS measurements. From
these results, we speculate that the fluoride anions introduced into
H-ZSM-5 may react with Si atoms in the framework of zeolite dur-
ing the calcination stage, being removed as gaseous SiF4.
Fig. 5 shows the XRD patterns of fluoride-treated H-ZSM-5 cat-
alysts. The diffraction lines of fluoride-treated H-ZSM-5 samples
could all be assigned to typical crystalline H-ZSM-5, suggesting
that the crystalline structure of H-ZSM-5 was sustained after the
fluoride treatment. It is unexpected that the intensity of the dif-
fraction lines becomes stronger with increasing concentration of
NH4F used for treating H-ZSM-5. This implies that the crystallinity
of H-ZSM-5 is increased after the fluoride treatment.
SEM measurements were performed to gain information about
the morphology of the fluoride-treated H-ZSM-5. H-ZSM-5 is com-
posed of bigger crystallites with some amorphous aggregates or
smaller particles (Fig. 6a). After the fluoride treatment, the amount
of the amorphous aggregates or smaller particles decreased
(Fig. 6b–f). This observation indicates that the amorphous aggre-
gates or smaller particles may be easier to react with the fluoride
anions and thus may be facile to be removed. The removal of the
amorphous aggregates could cause the increase in the intensity
of XRD peaks ascribed to H-ZSM-5 crystalline structure after the
fluoride treatment (Fig. 5).
Fig. 7 shows the argon adsorption–desorption isotherms and
the corresponding pore-diameter distributions for the fluoride-
treated H-ZSM-5 samples. All these samples exhibited type I iso-
therms, which are typical for microporous zeolites. The pore-diam-
eter distribution derived from the isotherms by the HK method
showed that H-ZSM-5 without fluoride treatment possessed
micropores with sizes around ꢁ0.52 nm, which are typical for H-
ZSM-5. In addition to the micropores with sizes around 0.51–
0.55 nm, new micropores with sizes around 0.73–0.78 nm were
observed in the fluoride-treated H-ZSM-5 samples (Fig. 7B). The
The change in acidity after the fluoride treatment was also de-
tected from FT-IR studies for the hydroxyl bands. In the region of
3800–3500 cmꢀ1, H-ZSM-5 exhibited IR bands at ꢁ3740, 3700
and 3610 cmꢀ1 (Fig. 10), and the three bands could be ascribed
to terminal silanol groups, internal (defective) silanol groups (also
known as silanol nests) and structural Si(OH)Al groups, respec-
tively [49]. The fluoride treatment significantly decreased the

238
T. Xu et al. / Journal of Catalysis 295 (2012) 232–241
Fig. 6. SEM micrographs for the H-ZSM-5 samples treated by fluoride with different concentrations. (a) H-ZSM-5, (b) H-ZSM-5-F-0.013 M, (c) H-ZSM-5-F-0.026 M, (d) H-ZSM-
5-F-0.052 M, (e) H-ZSM-5-F-0.078 M, and (f) H-ZSM-5-F-0.104 M.
Fig. 7. (A) Argon adsorption–desorption isotherms for the H-ZSM-5 samples
Fig. 9. NH3-TPD profiles for the H-ZSM-5 samples treated by fluoride with different
treated by fluoride with different concentrations and (B) corresponding pore-
concentrations. (A) NH3 adsorption at 373 K and (B) NH3 adsorption at 573 K.
diameter distributions obtained from argon adsorption–desorption isotherms by
the HK method.
Fig. 8. Adsorption amounts of p-xylene and o-xylene over H-ZSM-5 and H-ZSM-5-
F-0.052 M samples.
intensity of the band at 3610 cmꢀ1 belonging to the Brønsted acid
sites. When the concentration of fluoride exceeded 0.078 M, this
band disappeared, in agreement with the NH3-TPD result (Fig. 9).
The band at ꢁ3740 cmꢀ1 belonging to terminal silanol groups did
not undergo significant changes, whereas that at ꢁ3700 cmꢀ1 as-
cribed to silanol nests became weaker after the fluoride treatment.
This observation implies that the fluoride treatment may preferen-
tially remove the defective Si sites, and this is in agreement with
Fig. 10. IR spectra in the OH region for the H-ZSM-5 samples treated by fluoride
with different concentrations.
the XRD and SEM results (Figs. 5 and 6). It is reported that the desi-
lication of ZSM-5 by NaOH treatment also started from the defec-
tive silanol sites [50]. IR bands at 3664 and 3683 cmꢀ1 were also

T. Xu et al. / Journal of Catalysis 295 (2012) 232–241
239
and six-coordinated Al species in H-ZSM-5. After the treatment
of H-ZSM-5 by fluoride, a broad signal at ꢁ0 ppm became distinct
although the intensity of this signal was not high. This suggested
the appearance of the octahedrally coordinated Al, that is, extra-
framework Al. However, no signal at ꢁ30 ppm was observed, indi-
cating the absence of the five-coordinated Al species in these sam-
ples. At the same time, the signal at 55.5 ppm shifted to ꢁ52 ppm
and became broader (Fig. 12). The deconvolution of the broader
signal provided three components with chemical shifts at 55, 52,
and 47 ppm. Thus, we speculate that there might exist several dif-
ferent types of Al species in the fluoride-treated H-ZSM-5 samples.
The intensity of the component at 55 ppm assigned to the tetrahe-
dral Al species in the framework decreased with increasing con-
centration of fluoride, indicating the distortion of tetrahedral Al
species in framework. The components at 52 and 47 ppm might
be related to perturbed tetrahedral Al species [53,54]. Combining
with the results of acidic properties (Figs. 9–11), we speculate that
these Al species do not contribute to the acidity.
Fig. 11. IR spectra of adsorbed pyridine for the H-ZSM-5 samples treated by
fluoride with different concentrations.
observed for the samples treated by fluoride with medium concen-
trations (0.026 and 0.052 M). These bands could be attributed to
the hydroxyl groups associated with extra-framework Al [49], indi-
cating the generation of extra-framework Al due to the fluoride
treatment.
We have performed FT-IR studies of adsorbed pyridine for the
fluoride-treated H-ZSM-5 samples to gain further information
about the acid sites. Three IR bands at 1542, 1487, and
1450 cmꢀ1 were observed for H-ZSM-5 (Fig. 11). The bands at
1542 and 1450 cmꢀ1 are attributable to Brønsted acid and Lewis
acid sites, respectively, while the band at 1487 cmꢀ1 could stem
from both Brønsted acid and Lewis acid sites [51]. The fluoride
treatment decreased the intensities of the IR bands belonging to
both Brønsted acid and Lewis acid sites. For the sample treated
by fluoride with a concentration of 0.078 M, only small amounts
of both types of acid sites still remained.
3.5. Possible reaction mechanism and the roles of fluoride treatment
A large number of studies have been devoted to uncovering the
reaction mechanism for the conversion methanol to olefins, and
the ‘‘hydrocarbon pool’’ mechanism has been widely accepted in
recent years [11–13,55–60]. It is proposed that methylbenzenes
with different numbers of methyl substituents and the correspond-
ing carbenium ions work as active intermediates in the ‘‘hydrocar-
bon pool’’ in the working catalyst for the formation of light olefins
[55–60]. A similar ‘‘hydrocarbon pool’’ mechanism has also been
proposed for the conversion of CH3Cl to light olefins over H-
SAPO-34 [35,38].
We have analyzed the hydrocarbon compounds trapped in the
used catalysts. After the reaction for 2 h, the catalyst was dissolved
in HF solution and the organic compounds were extracted by CH2-
Cl2. The extracted organic compounds were then analyzed by GC–
MS. We observed various methylbenzenes over both H-ZSM-5 and
fluoride-treated H-ZSM-5 catalysts. Moreover, the distributions of
methylbenzenes were quite different over the two catalysts
(Fig. 13). Mono-, di-, and tri-methylbenzenes were the main organ-
ic compounds in the ‘‘hydrocarbon pool’’ in H-ZSM-5, whereas tet-
ra-, penta-, and hexa-methylbenzenes became predominant in the
fluoride-treated H-ZSM-5 (H-ZSM-5-F-0.052 M). Our studies have
indicated the generation of larger micropores with sizes of 0.73–
0.78 nm after the fluoride treatment (Figs. 7 and 8). We speculate
that this may have caused the change in the distribution of meth-
ylbenzene intermediates over the fluoride-treated H-ZSM-5
catalyst.
27Al MAS NMR is an effective technique to provide information
on the coordination state of Al species, and generally, the Al species
in four-, five-, and six-coordination states exhibit signals at 55–60,
ꢁ30, and ꢁ0 ppm, respectively [52]. To gain insights into the loca-
tion of Al in our catalysts, we have performed 27Al MAS NMR stud-
ies. H-ZSM-5 without fluoride treatment exhibited a symmetrical
and narrow signal at 55.5 ppm (Fig. 12), which was attributable
to the tetrahedral Al in the framework of H-ZSM-5. No signals at
30 and 0 ppm were observed, indicating the absence of the five-
Fig. 14 shows the reaction mechanism that we have proposed
for the conversion of methyl halides (denoted as CH3X, including
CH3Cl and CH3Br) over our catalysts. The analysis of the hydrocar-
Fig. 12. 27Al MAS NMR spectra for the H-ZSM-5 samples treated by fluoride with
Fig. 13. Distribution of methylbenzenes in the trapped hydrocarbons over H-ZSM-5
different concentrations.
and H-ZSM-5-F-0.052 M samples.

240
T. Xu et al. / Journal of Catalysis 295 (2012) 232–241
bon intermediates trapped in the used catalysts suggested the for-
mation of methylbenzenes in the ‘‘hydrocarbon pool’’ (Fig. 13). Like
the MTO process [11–13,55–60], light olefins are assumed to be
formed also from methylbenzenes in our case. Our results demon-
strated that light olefins, that is, C2H4, C3H6 and C4H8, were all the
primary products and should be formed in parallel (Fig. 3). The de-
tailed mechanism for the formation of light olefins from methyl-
benzenes might resemble that proposed for the MTO process
[12,13]. The main byproducts, that is, C2–C4 paraffins and C5+
hydrocarbons, were produced from the light olefins probably via
hydrogen-transfer and aromatization reactions.
It has been reported that the hydrogen-transfer and aromatiza-
tion reactions generally require the participation of strong acid
sites [61]. Our results clearly showed that the fluoride treatment
decreased the acidity of H-ZSM-5 (Figs. 9–11). This was believed
to inhibit the secondary conversions of light olefins over the fluo-
ride-treated H-ZSM-5 catalysts (Fig. 3). However, over the catalyst
with very low acidity or without acidity, for example, the H-ZSM-
5-F-0.078 M and H-ZSM-5-F-0.104 M catalysts, the conversion of
CH3Cl or CH3Br became significantly low. Thus, we can conclude
that an appropriate acidity is required for obtaining high selectivity
of light olefins at a reasonably high activity. This conclusion is also
supported by the catalytic results obtained for the H-ZSM-5 cata-
lysts with different Si/Al ratios and the ZSM-5 catalysts with differ-
ent Na+ exchange degrees. Similarly, there exists an optimum Si/Al
ratio or Na+ exchange degree, that is, a proper acidity, for obtaining
a high C3H6 yield (Tables 2 and 3). However, the comparison of Ta-
bles 2–4 demonstrates that the fluoride treatment is a more effi-
cient strategy to increase the selectivity of light olefins while
sustaining the conversion of methyl halides.
strated that for H-beta with larger sizes of micropores
(0.76 ꢂ 0.64 nm), penta- and hexa-methylbenzenes were mainly
involved in the conversion of methanol, leading to C3H6 and higher
olefins, whereas for H-ZSM-5, the lower methylbenzenes were the
main intermediates, forming C2H4 as the main primary product
[58,59]. Although H-beta provided higher selectivity of C3H6 than
H-ZSM-5, its activity for CH3OH conversion was lower [58,59].
For the conversions of methyl halides, H-beta was also less active
than H-ZSM-5 although it was more selective (Table 1). We clari-
fied that the fluoride treatment caused the generation of larger
micropores with sizes of 0.73–0.78 nm in H-ZSM-5, resulting in
the change of fraction of methylbenzenes in ‘‘hydrocarbon pool’’.
This increased the selectivity of C3H6 while keeping the high activ-
ity of H-ZSM-5.
4. Conclusions
H-ZSM-5 and H-SAPO-34 were found to exhibit higher C3H6
yields for the conversion of methyl halide (CH3Cl or CH3Br) than
several other zeolites including H-mordenite, H-beta, H-MCM-22,
HY, and H-SAPO-11. H-ZSM-5 was more stable than H-SAPO-34.
The catalytic performance of ZSM-5 depended on the Si/Al ratio
and the Na+ exchange degree; a higher Si/Al ratio or Na+ exchange
degree favored the selectivity of C3H6 but decreased the conversion
of methyl halide. We demonstrated that the treatment of H-ZSM-5
by NH4F followed by calcination afforded a highly selective and
stable catalyst for the production of C3H6 from both CH3Cl and CH3-
Br. The concentration of NH4F significantly affected the catalytic
performance of the fluoride-treated H-ZSM-5. At a NH4F concen-
tration of 0.052 M, the selectivities of C3H6 reached 64% and 56%
at CH3Cl and CH3Br conversions of 76% and 94%, respectively.
Our characterizations suggest that the strong Brønsted acidity of
H-ZSM-5 has been weakened and new micropores with sizes of
0.73–0.78 nm have been generated after fluoride treatment. It is
proposed that the decrease in the acidity contributes to increasing
the selectivity of light olefins by suppressing the hydrogen-transfer
and aromatization reactions, by which the primary light olefins are
transformed to C2–C4 alkanes and C5+ hydrocarbons. The presence
of larger micropores increased the fraction of higher methylbenz-
enes, that is, tetra-, penta-, and hexa-methylbenzenes, in the
‘‘hydrocarbon pool’’ and increased the selectivities of C3H6 and
C4H8 at the expense of that of C2H4.
We have demonstrated that the fluoride treatment can change
the distribution of light olefins at the initial reaction stage
(Fig. 3). We speculate that this stems from the change in the frac-
tions of methylbenzenes in the ‘‘hydrocarbon pool’’ after the fluo-
ride treatment (Fig. 13). It has been reported that mono-, di-, and
tri-methylbenzenes are mainly responsible for the formation of
C2H4, whereas tetra-, penta-, and hexa-methylbenzenes favor the
formation of C3H6 [12,13,59]. Thus, it is reasonable to consider that
the increase in the fraction of tetra-, penta-, and hexa-methylbenz-
enes after the fluoride treatment contributes to the increases in the
selectivities to C3H6 and C4H8 and the decrease in that to C2H4 (the
initial stage of Fig. 3). Actually, Bjørgen and co-workers demon-
Acknowledgments
This work was supported by the National Basic Research Pro-
gram of China (2010CB732303), the Natural Science Foundation
of China (20923004, 21033006 and 21161130522), the Program
for Changjiang Scholars and Innovative Research Team in Univer-
sity (No. IRT1036), and the Key Scientific Project of Fujian Province
(2009HZ0002-1).
References
[1] A.H. Tullo, Chem. Eng. News 81 (50) (2003) 15–16.
[2] Y.M. Liu, W.L. Feng, T.C. Li, H.Y. He, W.L. Dai, W. Huang, Y. Cao, K.N. Fan, J. Catal.
239 (2006) 125–136.
[3] J.C. Mol, J. Mol. Catal. A 213 (2004) 39–45.
[4] H. Oikawa, Y. Shibata, K. Inazu, Y. Iwase, K. Murai, S. Hyodo, G. Kobayashi, T.
Baba, Appl. Catal. A: Gen. 312 (2006) 181–185.
[5] M. Iwamoto, Y. Kosugi, J. Phys. Chem. C 111 (2007) 13–15.
[6] M. Taoufik, E. Le Roux, J. Thivolle-Cazat, J.M. Basset, Angew. Chem., Int. Ed. 46
(2007) 7202–7205.
[7] B. Lin, Q. Zhang, Y. Wang, Ind. Eng. Chem. Res. 48 (2009) 10788–10795.
[8] N. Xue, X. Chen, L. Nie, X. Guo, W. Ding, Y. Chen, M. Gu, Z. Xie, J. Catal. 248
(2007) 20–28.
[9] G. Zhao, J. Teng, Z. Xie, W. Jin, W. Yang, Q. Chen, Y. Tang, J. Catal. 248 (2007) 29–
37.
Fig. 14. Proposed reaction mechanism for the conversion of methyl halide (CH3Cl or
CH3Br) over the H-ZSM-5-based catalysts.

T. Xu et al. / Journal of Catalysis 295 (2012) 232–241
241
[10] E. Mazoyer, K.C. Szeto, S. Norsic, A. Garron, J.M. Basset, C.P. Nicholas, M.
Taoufik, ACS Catal. 1 (2011) 1643–1646.
[38] U. Olsbye, O.V. Saure, N.B. Muddada, S. Bordiga, C. Lamberti, M.H. Nilsen, K.P.
Lillerud, S. Svelle, Catal. Today 171 (2011) 211–220.
[11] M. Stöcker, Microporous Mesoporous Mater. 29 (1999) 3–48.
[12] J.F. Haw, W. Song, D.W. Marcus, J.B. Nicholas, Acc. Chem. Res. 36 (2003) 317–
326.
[39] A. Zhang, S. Sun, Z.J.A. Komon, N. Osterwalder, S. Gadewar, P. Stoimenov, D.J.
Auerbach, G.D. Stucky, E.W. McFarland, Phys. Chem. Chem. Phys. 13 (2011)
2550–2555.
[13] U. Olsbye, S. Svelle, M. Bjørgen, P. Beato, T.V.W. Janssens, F. Joensen, S. Bordiga,
K.P. Lillerud, Angew. Chem., Int. Ed. 51 (2012) 5810–5831.
[14] A. Holmen, Catal. Today 142 (2009) 2–8.
[40] Z. Liu, J. Liang, Curr. Opin. Solid State Mater. Sci. 4 (1999) 80–84.
[41] C. Mei, P. Wen, Z. Liu, H. Liu, Y. Wang, W. Yang, Z. Xie, W. Hua, Z. Gao, J. Catal.
258 (2008) 243–249.
[15] K. Otsuka, Y. Wang, Appl. Catal. A: Gen. 222 (2001) 145–161.
[16] K. Tabata, Y. Teng, T. Takemoto, E. Suzuki, M.A. Bañares, M.A. Peña, J.L.G. Fierro,
Catal. Rev. 44 (2002) 1058.
[42] W. Dai, X. Wang, G. Wu, N. Guan, M. Hunger, L. Li, ACS Catal. 1 (2011) 292–299.
[43] J. Ortega, A. Gayubo, A. Aguayo, P. Benito, J. Bilbao, Ind. Eng. Chem. Res. 36
(1997) 60–66.
[17] A. Arena, A. Parmaliana, Acc. Chem. Res. 36 (2003) 867–875.
[18] Y. Xu, X. Bao, L. Lin, J. Catal. 216 (2003) 386–395.
[19] Q. Zhang, W. Deng, Y. Wang, Chem. Commun. 47 (2011) 9272–9275.
[20] H. Schwarz, Angew. Chem., Int. Ed. 50 (2011) 10096–10115.
[21] G.A. Olah, B. Gupta, M. Farina, J.D. Felberg, W.M. Ip, A. Husain, R. Karpeles, K.
Lammertsma, A.K. Melhotra, N.J. Trivedi, J. Am. Chem. Soc. 107 (1985) 7097–
7105.
[22] X.P. Zhou, A. Yilmaz, G.A. Yilmaz, I.M. Lorkovic, L.E. Laverman, M. Weiss, J.H.
Sherman, E.W. McFarland, G.D. Stucky, P.C. Ford, Chem. Commun. (2003)
2294–2295.
[44] C. Minh, R. Jones, I. Craven, T. Brown, Energy Fuels 11 (1997) 463–469.
[45] C.D. Baertsch, H.H. Funke, J.L. Falconer, R.D. Noble, J. Phys. Chem. 100 (1996)
7676–7679.
[46] J.H. Kim, A. Ishida, M. Okajima, M. Niwa, J. Catal. 161 (1996) 387–392.
ˇ ˇ
[47] E. Aubert, F. Porcher, M. Souhassou, V. Pettrícek, C. Lecomte, J. Phys. Chem. B
106 (2002) 1110–1117.
[48] N. Katada, H. Igi, J.H. Kim, M. Niwa, J. Phys. Chem. B 101 (1997) 5969–5977.
[49] P. Wu, T. Komatsu, T. Yashima, J. Phys. Chem. 99 (1995) 10923–10931.
[50] B. Gil, Ł. Mokrzycki, B. Sulikowski, Z. Olejniczal, S. Walas, Catal. Today 152
(2010) 24–32.
ˇ
[23] J. He, T. Xu, Z. Wang, Q. Zhang, W. Deng, Y. Wang, Angew. Chem., Int. Ed. 51
(2012) 2438–2442.
[51] J.B. Koo, N. Jiang, S. Saravanamurugan, M. Bejblová, Z. Musilová, J. Cejka, S.E.
Park, J. Catal. 276 (2010) 327–334.
[24] K.X. Wang, H.F. Xu, W.S. Li, X.P. Zhou, J. Mol. Catal. A: Chem. 225 (2005) 65–69.
[25] Z. Liu, L. Huang, W.S. Li, F. Yang, C.T. Au, X.P. Zhou, J. Mol. Catal. A: Chem. 273
(2007) 14–20.
[52] J. Klinowski, Chem. Rev. 91 (1991) 1459–1479.
[53] J.A. van Bokhoven, A.L. Roest, D.C. Koningsberger, J. Phys. Chem. B 104 (2000)
6743–6754.
[26] R. Lin, Y. Ding, L. Gong, W. Dong, J. Wang, T. Zhang, J. Catal. 272 (2010) 65–73.
[27] E. Peringer, S.G. Podkolzin, M.E. Jones, R. Olindo, J.A. Lercher, Top. Catal. 38
(2006) 211–220.
[28] S.G. Podkolzin, E.E. Stangland, M.E. Jones, E. Peringer, J.A. Lercher, J. Am. Chem.
Soc. 129 (2007) 2569–2576.
[54] P. Sazama, B. Witchterlova, J. Dedecek, Z. Tvaruzkova, Z. Musilova, L. Palumbo,
S. Sklenak, O. Gonsiorova, Microporous Mesoporous Mater. 143 (2011) 87–96.
[55] B. Arstad, S. Kolboe, J. Am. Chem. Soc. 123 (2001) 8137–8138.
[56] S. Svelle, F. Joensen, J. Nerlov, U. Olsbye, K.P. Lillerud, S. Kolboe, M. Bjørgen, J.
Am. Chem. Soc. 128 (2006) 14770–14771.
[29] E. Peringer, M. Salzinger, M. Hutt, A.L. Lemonidou, J.A. Lercher, Top. Catal. 52
(2009) 1220–1231.
[57] D.M. McCann, D.L. Lesthaeghe, P.W. Kletnieks, D.R. Guenther, M.J. Hayman,
V.V. Speybroeck, M. Waroquier, J.F. Haw, Angew. Chem., Int. Ed. 47 (2008)
5179–5182.
[58] S. Svelle, U. Olsbye, F. Joensen, M. Bjørgen, J. Phys. Chem. C 111 (2007) 17981–
17984.
[59] M. Bjørgen, F. Joensen, K.P. Lillerud, U. Olsbye, S. Svelle, Catal. Today 142
(2009) 90–97.
[60] J. Li, Y. Wei, J. Chen, P. Tian, X. Su, S. Xu, Y. Qi, Q. Wang, Y. Zhou, Y. He, Z. Liu, J.
Am. Chem. Soc. 134 (2012) 836–839.
[30] D.K. Murray, J.W. Chang, J.F. Haw, J. Am. Chem. Soc. 115 (1993) 4732–4741.
[31] C.E. Taylor, Stud. Surf. Sci. Catal. 130D (2000) 3633–3638.
[32] D. Jaumain, B.L. Su, J. Mol. Catal. A: Chem. 197 (2003) 263–273.
[33] Y. Wei, D. Zhang, L. Xu, Z. Liu, B.L. Su, Catal. Today 106 (2005) 84–89.
[34] Y. Wei, D. Zhang, Z. Liu, B.L. Su, J. Catal. 238 (2006) 46–57.
[35] Y. Wei, D. Zhang, F. Chang, Q. Xia, B.L. Su, Z. Liu, Chem. Commun. (2009) 5999–
6001.
[36] S. Svelle, S. Aravinthan, M. Bjørgen, K.P. Lillerud, S. Kolboe, I.M. Dahl, U. Olsbye,
J. Catal. 241 (2006) 243–254.
[61] M.A. Sanchez-Castillo, R.J. Madon, J.A. Dumesic, J. Phys. Chem. B 109 (2005)
2164–2175.
[37] M.H. Nilsen, S. Svelle, S. Aravinthan, U. Olsbye, Appl. Catal. A: Gen. 367 (2009)
23–31.