The influence of acidities of boron- and aluminium-containing MFI zeolites on co-reaction of methanol and ethene

Article abstract of DOI:10.1039/c1cp20338j

A series of boron- and aluminium-containing MFI zeolites were synthesized and various characterization techniques, such as NMR (²⁷Al, ²⁹Si and ¹¹B), were employed to study the acidities of zeolites. Moreover, in situ IR was applied to investigate the interaction of methanol and ethene with the acid sites, and those catalytic materials were used for co-reaction of methanol and ethene to produce propene. The production of propene was related to the Al content of the zeolites with Si/Al ratios of higher than 90. It is implied that the presence of boron during the synthesis directed the aluminium to occupy certain tetrahedral sites in the zeolite framework, thus preventing the formation of ethene oligomers, and resulting in increased propene selectivity. the Owner Societies 2011.

Full text of DOI:10.1039/c1cp20338j

View Article Online / Journal Homepage / Table of Contents for this issue
Dynamic Article Links
PCCP
Cite this: Phys. Chem. Chem. Phys., 2011, 13, 14598–14605
www.rsc.org/pccp
PAPER
The influence of acidities of boron- and aluminium-containing MFI
zeolites on co-reaction of methanol and ethene
a
a
a
b
Qingjun Zhu, Junko N. Kondo, Toshiyuki Yokoi, Tooru Setoyama,
b
b
c
Masashi Yamaguchi, Takahiko Takewaki, Kazunari Domen and
a
Takashi Tatsumi*
Received 9th February 2011, Accepted 31st May 2011
DOI: 10.1039/c1cp20338j
A series of boron- and aluminium-containing MFI zeolites were synthesized and various
2
7
29
11
characterization techniques, such as NMR ( Al, Si and B), were employed to study
the acidities of zeolites. Moreover, in situ IR was applied to investigate the interaction of
methanol and ethene with the acid sites, and those catalytic materials were used for co-reaction
of methanol and ethene to produce propene. The production of propene was related to the
Al content of the zeolites with Si/Al ratios of higher than 90. It is implied that the presence
of boron during the synthesis directed the aluminium to occupy certain tetrahedral sites in
the zeolite framework, thus preventing the formation of ethene oligomers, and resulting in
increased propene selectivity.
1
. Introduction
The role of methylation reaction, i.e., methyl group substi-
tution in the product formations in the MTO process, is an
interesting topic since methanol is one of the most important
methylation agents in heterogeneous catalysis. The importance
of the methylation of hydrocarbons with methanol in the MTO
The methanol to olefins (MTO) process offers a practical
reaction pathway for the future production of ethene and
propene which are building blocks for the polymer industry.
The topology and acidity of the catalytic materials play crucial
roles in the transformation of methanol and the formation of
15
process was addressed by Dessau and Lapierre. Because alkenes
are apparently the primary products from the transformation of
methanol, the sequential methylation of the produced alkene
16
1
,2
desired products. The experimental results obtained in the
last decade, mainly by the isotope exchange, indicate that the
hydrocarbon-pool mechanism plays decisive roles in the olefins
are believed to be important reactions in the MTO process.
The experimental evidence, however, was scarce even though
the MTO process has been developing for more than thirty
years, possibly caused by the difficulty in distinguishing the
addition of methanol to ethene from the interconversion of
alkenes over acidic materials. The pioneering experimental works
3
–10
production.
However, it has been claimed recently that
propene is also formed through alkene methylation to a
11
considerable extent over HZSM-5 zeolites. This is in agreement
with the theoretical calculation results that the methylation of
smaller branched alkenes rather than bulkier methylbenzenes
is more favorable on HZSM-5 due to the geometric constraints
on methylation of alkenes with methanol in the MTO were
17,18
carried out by Svelle et al.
Isotopically labeled reagents
1
2
imposed by the MFI topology. In other words, the reaction
mechanisms of methanol to olefins over silicoalumino-
phosphate SAPO-34 and aluminosilicate HZSM-5, the most
important two catalysts for the MTO process, could be
different significantly; that is, propene formation from methylation
of ethene with methanol may be pronounced on HZSM-5
whereas such methylation only plays a minor role over
were employed at extremely high feed rates on HZSM-5, and
the obtained isotopomer products have unequivocally confirmed
that the methylation of the alkenes with methanol plays
important roles in the methanol to olefins process.
The methylation reaction is catalyzed by the acid sites of
HZSM-5 in the MTO process. The study of the interaction
between methanol and the acidic catalysts is extensively reported,
and recent research has focused on the function of the
1
3,14
SAPO-34.
1
9,20
methoxy group.
Theoretical simulation of the methylation
a
Chemical Resources Laboratory, Tokyo Institute of Technology,
259 Nagatsuta-cho, Midori-ku, Yokohama 226-8503, Japan.
of alkene has predicted that the end-on adsorbed methanol on
the acid site, with a methanol perpendicular to the zeolite surface,
4
E-mail: ttatsumi@cat.res.titech.ac.jp; Fax: +81-45-924-5282;
Tel: +81-45-9245238
Mitsubishi Chemical Group, Science and Technology Research
Center, 1000 Kamoshida-cho, Aoba-ku, Yokohama 227-8502, Japan
Department of Chemical System Engineering, The University of
Tokyo, 7-3-1 Hongo, Bunkyo-ku, Tokyo 113-8656, Japan
2
1
reacts with the activated hydrocarbons on the siliceous part.
b
c
This suggests that the methoxy group, the methyl group
directly attached to the oxygen atoms in the zeolite, does not
participate in the methylation reaction. Nevertheless, the other
1
4598 Phys. Chem. Chem. Phys., 2011, 13, 14598–14605
This journal is c the Owner Societies 2011

View Article Online
important aspect of the alkene substrate activation in the
methylation reaction is not illustrated in detail despite the fact
that alkene is readily adsorbed on the acid sites, and the
experimental knowledge of interaction of the substrates with
the catalysts under the reaction conditions is limited.
In this work, a series of boron and aluminium containing
MFI zeolites was prepared and a variety of characterization
techniques such as NMR and IR were employed to characterize
intrinsic properties of the catalysts. We investigated the inter-
action of methanol and ethene with the MFI zeolite acidities
under the reaction conditions. The catalysts are tested for
co-reaction of methanol and ethene (CME) to produce propene.
The main objective is to study the correlation of the catalytic
performance with zeolite acidities, particularly in terms of the
interaction between the adsorbed species and the nature of
acid sites.
with a p/2 pulse width of 2 ms and a pulse delay of 5 s.
Tetramethylsilane was used as the reference sample for Si
2
9
2
9
NMR. Si MAS NMR spectra were obtained with a p/2 pulse
width of 7 ms and a pulse delay of 60 s. B(OH) was used as the
3
1
1
11
reference sample for B NMR. B MAS NMR spectra were
obtained with a p/2 pulse width of 2 ms and a pulse delay of 5 s.
The boron resonance frequency employed was 128.34 MHz
and the sample spinning rate was 7.0 kHz.
3
Ammonia temperature programmed desorption (NH -TPD)
profiles were recorded on a Multitrack TPD equipment
(Japan BEL). Typically, 25 mg of the catalyst was pretreated
À1
at 773 K for 1 h in He (50 ml min ), and then cooled to an
adsorption temperature of 373 K. Prior to the adsorption of
NH , the sample was evacuated. The sample was exposed to
3
approximately 2500 Pa NH3 for 10 min. Subsequently, the
sample was evacuated to remove weakly adsorbed NH for
3
À1
3
0 minutes. Finally, He (50 ml min ) was passed through
the reactor that was heated from 373 to 923 K at a ramping
À1
2
. Experimental
rate of 5 K min . A mass spectrometer was used to monitor
2
.1 Catalyst preparation
desorbed NH (m/z = 16).
3
The surface acidity and the interaction of methanol or
The MFI zeolites were prepared by hydrothermal synthesis in
ethene molecules with the acidic sites of zeolites were studied
autoclaves. Cab-O-Sil M5, Al(NO) Á9H O and H BO were
3
2
3
3
by in situ infrared (IR) spectroscopy at elevated temperatures.
2
employed as silicon, aluminium and boron sources, respectively.
Tetrapropylammonium bromide (TPABr) was used as the
template. The prepared mother gel was autoclaved at 443 K
for 7 days under tumbling conditions. The obtained crystals
were filtered, washed and dried at 373 K overnight. The template
was removed by calcination at 823 K for 10 h. The sodium
2,23
The setup was elaborated in our previous publication.
The
zeolite sample of approximately 10 mg was pressed into a self-
supporting wafer and placed in an in situ IR cell that was
heatable. All the IR spectra were recorded on an FT/IR 7300
À1
spectrometer with an MCT detector at a resolution of 4 cm
and 64 scans. Prior to the measurements, the sample was
heated to 773 K in a simulated air atmosphere for 1 h to
remove adsorbed water. After the sample was cooled down to
type of MFI zeolites was treated with aqueous NH
+
4
NO
3
to
+
ion-exchange Na for NH
4
, followed by calcination to
produce proton type catalysts. X-Ray diffraction (XRD)
patterns confirmed the crystal structure of the zeolites. BET
measurements, despite being not strictly valid for the quantitative
analysis of the microporous materials, showed the micro-
porosity of these MFI zeolites. Their chemical compositions
are listed in Table 1. The catalysts are denoted by H[B,Al]MFI
5
73 K, the wafer was exposed to the methanol vapor diluted
with N for 30 min. The ethene adsorption was carried out at
73 K. Finally, the sample was flushed with N until neither
2
4
2
methanol nor ethene in the gas phase was observed by IR
spectroscopy.
(x, y), where x and y stand for the Si/B ratio and Si/Al ratio in
the zeolites, respectively. A sample synthesized in the absence of
boron is designated as HZSM-5 (92) with the Si/Al ratio of 92.
The average particle size of synthesized zeolite was ca. 1 mm.
2.3 Catalytic reaction
Co-reaction of methanol and ethene was carried out in a
continuous flow reactor at atmospheric pressure. Methanol
conversion was defined as the consumed methanol divided by
the methanol fed to the reactor, and its dehydrated product
dimethyl ether was considered to be an unconverted reactant.
The product selectivity was calculated on a carbon basis; that
is, the amount of carbon atoms in each obtained product was
divided by the converted carbon atoms from ethene and
methanol together. We estimated that carbon mass balance
was close to 95%. Methanol (Kanto, 499.8%) was injected
into an evaporator by using an automotive fuel injector
2
.2 Catalyst characterization
2
7
29
11
Solid-state Al, Si and B magic angle spinning nuclear
magnetic resonance (MAS NMR) spectra were obtained on a
JEOL ECA-400 spectrometer. The Al resonance frequency
employed was 104.23 MHz and the sample spinning rate was
2
7
7
.0 kHz. Al chemical shifts were referenced to a saturated
2
7
3 3
Al(NO ) solution. Al MAS NMR spectra were obtained
Table 1 The compositions of boron and aluminium containing MFI
zeolites and HZSM-5 (92)
(
Harvard). Methanol was vaporized in an evaporator and
fed to the catalyst bed with N (499.99%) as a carrier gas.
2
a
a
Catalyst
Si/B
Si/Al
The ethene (499.99%) flow rate was regulated by a mass flow
controller, and the flow rate was calibrated by a bubble meter.
The tubing of the reaction system was heated to avoid condensa-
tion of heavy product molecules to guarantee their detection.
Blank runs in an empty reactor (no catalyst) resulted in no
observable conversion of methanol and ethene at elevated
temperatures up to 773 K, demonstrating that no gas-phase
H[B,Al]MFI (36, 2100)
H[B,Al]MFI (41, 331)
H[B,Al]MFI (47, 162)
H[B,Al]MFI (58, 100)
HZSM-5 (92)
36
41
47
58
0
42100
331
162
100
92
a
ICP analysis.
This journal is c the Owner Societies 2011
Phys. Chem. Chem. Phys., 2011, 13, 14598–14605 14599

View Article Online
reaction occurred. Three well-calibrated online gas chromato-
graphs (Shimadzu) were used as the analyzing equipments
with thermal conductivity detectors (TCD). Thermon_3000 on
Chromosorb and Gaskuropack 54 columns were used to
detect the aliphatic products of C –C and aromatic products
with tetrahedrally coordinated Al of the H[B,Al]MFI zeolites
2
7
are centered at about 53 ppm. Han et al. studied the 2D Al
3Q MAS NMR of ZSM-5 and concluded that 12 distinct
tetrahedral (T) sites in ZSM-5, when occupied by Al, have
2
5
different chemical shifts ranging from 57 to 51 ppm. The
tetrahedrally coordinated Al peaks should consist of the
contributions of different T-sites. The T-sites are divided into
L (57–54 ppm) and R (53–51 ppm) groups, depending on the
2
6
of C –C , respectively. Gases were fractionated by a silica gel
6 10
column. Typically, 0.1 g of the catalyst was placed at the
center of a quartz reactor, 4 mm in inner diameter, in a vertical
2
5
27
furnace. The mixture containing gas phase methanol and
À1
T–O–T angles. Our observation of the Al NMR signal at
56 ppm for HZSM-5 (92) corresponds to the literature on the
point that the majority of Al atoms are located in the T-sites in
ethene in N
2
was fed at a total flow rate of 10 ml min
.
The overall gas hourly space velocity (GHSV) of feed was
À1
25
kept constant at approximately 3000 h ; the apparent density
À1
the L-groups for HZSM-5 with high Si to Al ratio. However,
of the employed materials was approximately 0.5 g ml
.
if both boron and aluminium are present in the gel, the
majority of Al species seem to be incorporated into the
R-groups of T-sites, leading to the appearance of the peak
Catalysts were calcined prior to the reaction at 823 K in a
simulated air atmosphere for 2 h. The temperature of the
catalyst bed was then decreased from 823 to 523 K and the
mixture of the reactants was introduced into the reactor.
Reaction products were analyzed after 30 min reaction time
during the steady state. The reaction temperature was elevated
from 523 to 723 K in a stepwise manner at intervals of 50 K,
and the ramping process took 30 min at each step. At each
temperature, the reaction products were analyzed after the
reaction was stabilized. The catalytic performance presented in
this report was obtained at 723 K, 4.5 h after the reactants
were introduced into the reactor.
2
7
at ca. 53 ppm, as evidenced by Al NMR spectra of the
H[B,Al]MFI zeolites. In other words, the presence of boron
affects the incorporation of Al into the framework. Because
¨
Bronsted acidity is induced by the framework Al in zeolites, it
is reasonable to conclude that the acidities of these
H[B,Al]MFI zeolites differ from their HZSM-5 counterpart.
1
1
Fig. 2 shows the B MAS NMR spectra of boron-containing
zeolites. Note that the samples are exposed to the air, and thus
can be regarded as hydrated samples. Detailed discussion of
the local structures of framework boron in borosilicates in the
2
6
hydrated state was elaborated by Hwang et al. The broad
band (8 to À2 ppm) is assigned to the second-order quadrupole
3
. Results and discussion
broadening of trigonally coordinated B(O–Si) in the zeolites
3
1
framework, as confirmed by B 2D MQMAS NMR spectra.
1
26
3
.1 NMR
The presence of this type of boron in the zeolites results in
silanol groups (Si–OH) if boron is not located in the external
surface of the zeolites. The signal centered at ca. À3 ppm
Information concerning the incorporation and nature of Al in
2
7
the zeolites has been obtained by means of Al NMR. The
2
7
Al MAS NMR spectra of the calcined samples are shown in
represents the tetrahedrally coordinated boron, i.e., B(O–Si)
4
Fig. 1. H[B,Al]MFI (36, 2100) shows no signal, indicating that
its low amount of Al is beyond the limitation of detection by
in the zeolite framework, and the asymmetric line shape
implies that several tetrahedrally coordinated boron sites are
27
2
7
Al NMR. The octahedrally coordinated Al with a character-
co-existent in the sample. The presence of tetrahedrally
istic peak at 0 ppm, usually assigned to extra-framework Al,
are observed clearly on H[B,Al]MFI (58, 100) by a slight and
broad band although such signals are hardly detected on other
studied samples, indicating that the majority of the Al are
coordinated boron in the zeolite lattice requires cations to
counterbalance the negative charge of the zeolite framework.
Since the sodium amounts in our zeolites are negligible, we suggest
+
that the proton, or a large cluster like H (H O) , compensates
2 n
2
4
located in the zeolite frameworks. Note that the chemical
shift is 56 ppm for HZSM-5 (92), whereas the peak associated
the negative charges caused by the framework boron, leading
to boron-induced Bronsted acid sites in H[B,Al]MFI zeolites.
However, it was reported that such acidity might disappear
¨
27
Fig. 1
Al MAS NMR spectra for H[B,Al]MFI zeolites and
11
HZSM-5 (92).
Fig. 2
B MAS NMR spectra for H[B,Al]MFI zeolites.
1
4600 Phys. Chem. Chem. Phys., 2011, 13, 14598–14605
This journal is c the Owner Societies 2011

View Article Online
29
3
Fig. 4 NH TPD for H[B,Al]MFI zeolites and HZSM-5 (92).
Fig. 3
Si MAS NMR spectra for H[B,Al]MFI zeolites and
HZSM-5 (92).
on all samples except H[B,Al]MFI (36, 2100), indicating that
there is negligible strong acidity on H[B,Al]MFI (36, 2100).
The temperature for ammonia desorption maximum for strongly
acidic sites differs slightly. The shapes for the strongly acidic
sites are similar, and the desorption amount increases with the
Al content. However, we would refrain from estimating the
exact ammonia uptake based on the intensity for those
under dehydration conditions (e.g., 723 K as our reaction
2
8
temperature). Moreover, nonframework trigonal boron with
a characteristic band at 17–10 ppm is not observed. Hence, we
propose that the majority of the boron species in the zeolite are
incorporated in the zeolite framework after synthesis, either in
the trigonal manner or tetrahedral manner.
2
9
materials of low acid amounts. The NH -TPD results confirm
3
Si MAS NMR was used to study the Si environment in the
the presence of strongly acidic sites in the studied zeolites
synthesized zeolites, and the spectra are displayed in Fig. 3.
The spectra of the zeolites are dominated by the resonance at
around À110 ppm, which is assigned to the tetrahedrally
coordinated Si in the zeolite frameworks. The small and
narrow band at À105 ppm on HZSM-5 (92) is assigned to
except H[B,Al]MFI (36, 2100).
3
.3 IR
The infrared spectra of the dehydrated zeolites in the range of
À1
2
9
the Si atoms associated with the Bro
¨
nsted acid sites, i.e., the
4000 to 1250 cm are presented in Fig. 5. All the samples
Si(1Al) unit. The band might be distinguishable for
H[B,Al]MFI (58, 100), but not visible for the other studied
zeolites, possibly due to their low Al contents. There is about
exhibit typical proton-type zeolite patterns with overtones of
À1
frameworks centered at ca. 1880 cm . The boron-containing
À1
samples show broad bands at 1385 cm that are attributed to
3
the asymmetric stretching vibration of the B–O bond in BO3;
1
2
ppm chemical shift difference between HZSM-5 (92) and the
H[B,Al]MFI zeolites. In addition, the H[B,Al]MFI zeolites
show a broader line width than HZSM-5 (92), which might
indicate a variety of crystallographic T sites. The main peak at
À111 ppm and the shoulder at around À115 ppm could be
ascribed to different inequivalent T-sites of the zeolite
such a band is absent on HZSM-5 (92). The boron-containing
À1
zeolites show broad bands from 3750 to 3350 cm , which
are due to the hydrogen bond between the Si–OH group
3
2
and the vicinal B–OH groups. The various band shapes of
the hydroxyl group stretching vibrations suggest that boron
and aluminium incorporation into the zeolite frameworks is
dependent on the compositions of the synthesis gel. Despite its
2
9
frameworks.
3
.2 NH -TPD
3
Temperature-programmed desorption of ammonia is applied
to obtain the information of the nature and distribution of the
acid sites on the zeolites. The ammonia desorption profiles are
shown in Fig. 4. Only the desorption strengths of overall acid
sites can be distinguished while the correlation between acid sites
¨
structures (Bronsted and Lewis acid sites) and the assignments
of ammonia associated with them are debated in the literature.
All the samples exhibit ammonia desorption at relatively low
temperatures (B420 K) which is identified as physically
adsorbed or weakly held ammonia. This band is not related
with the weak acid sites and should have no significant function
3
0
in the catalysis. The NH
usually assigned to strongly acidic sites, is directly associated
with the Bronsted and Lewis acid sites which are important for
3
desorption at high temperatures,
¨
Fig. 5 As-observed IR spectra of H[B,Al]MFI zeolites and HZSM-5
(92) at 573 K after the samples are treated in vacuum at 773 K for
30 min for the removal of water.
the catalysis. The high temperature ammonia desorption,
characterized by the broad band around 610 K, is observed
This journal is c the Owner Societies 2011
Phys. Chem. Chem. Phys., 2011, 13, 14598–14605 14601

View Article Online
Fig. 6 FT-IR differential spectra of the zeolites exposed to methanol
at 573 K and then purged by N
Fig. 7 FT-IR differential spectra of the zeolites exposed to ethene at
2
.
4
2
73 K and then purged by N .
À1
relatively low Al content, two bands at 3738 and 3605 cm
can be observed on HZSM-5 (92).
catalyst, the adsorption of ethene is clearly visible in the form
of the products therefrom. This is consistent with the observation
34
Fig. 6 displays difference IR spectra measured after exposure
of catalysts to methanol followed by nitrogen purge at 573 K.
The difference spectra are obtained by subtracting the zeolite
spectra taken at 573 K before adsorption. The negative bands
indicate the consumption of corresponding functional groups
on zeolite catalysts whereas the positive bands point to the
generated adsorbed species. Obviously, both boron and
made by Stepanov et al. Evidence for residual species on the
surface of the HZSM-5 (92) catalyst is clearly observed in the
corresponding spectrum, where adsorption bands appear at
À1
3
450, 2952, 2933, 2860, 1468 and 1380 cm , accompanied by
À1
the consumption of the hydroxyl group band at 3605 cm
.
À1
Intensive bands at 2933 and 2860 cm are assigned to the
asymmetric and symmetric stretching vibrations of the C–H
bond in ethene oligomers, corroborated by their deformation
aluminium in the framework of the zeolites interact with CH OH
3
at 573 K, as confirmed by the negative bands in the hydroxyl
À1
À1
vibrations at 1468 and 1380 cm . The broad band centered at
À1
stretching region (from 3750–3500 cm ) and the positive
À1
3
450 cm , due to the hydrogen bond formation, indicates the
bands at around 2956 and 2855 cm , which are associated
interaction of physically adsorbed hydrocarbons with the hydroxyl
35
group. Accordingly, it is concluded that ethene oligomers are
with the asymmetric and symmetric stretching vibrations of the
33
C–H bond of the adsorbed methoxy groups. In contrast to the
well-defined peaks observed for HZSM-5 (92), the H[B,Al]MFI
zeolites show broad bands, suggesting that the hydrogen bond
participates in the interaction with methanol. The asymmetric
features of these bands for boron containing samples indicate
¨
readily formed by the interaction of ethene with the Bronsted
acid sites of HZSM-5 (92). Note that no bands characteristic
of ethene oligomers are found on H[B,Al]MFI (58, 100)
although its Al content is similar to that of HZSM-5 (92). It is
apparent that boron plays a more elaborate role in this catalyst
system than what has been generally assumed; namely, it has
been supposed that the boron and the aluminium are independent
of each other in the zeolite catalysis, and that the Al incorporation
and B incorporation during the synthesis do not interfere. The
different behavior of these H[B,Al]MFI zeolites from HZSM-5
toward ethene adsorption can only originate from the substantial
differences in the structure between the zeolites, for example, the
different acidities caused by Al atoms at different tetrahedral
sites in the framework.
that they might include the contributions from SiOCH Al,
3
3
3
SiOCH B, SiOCH3 and hydrogen bonded methanol, etc.
3
Moreover, the bending modes (positive bands) of the generated
methoxy groups were swamped by the negative bands caused
–
1
by the disappearance of the bands at 1385 cm from trigonally
À1
coordinated borons, leading to complex features below 1500 cm
for the boron-containing samples. Therefore, the mode of
interaction between methanol and the trigonally coordinated
borons remains unclear. Furthermore, no band attributed
to the acidic hydroxyl groups can be found on H[B,Al]MFI
(
36, 2100) although there appeared slight bands due to the
3
.4 Reaction results
adsorbed methanol.
The adsorption behavior of ethene and related species on
the zeolitic materials is investigated by the use of IR spectroscopy
to gain a better understanding of the events on the catalyst
surface, and the difference IR spectra taken after exposure of
catalysts to ethene followed by nitrogen purge at 473 K are
presented in Fig. 7. No band is observed for the H[B,Al]MFI
zeolites on which ethene was adsorbed, indicating a complete
removal of physically adsorbed ethene and/or the products
therefrom, if any, on the catalyst surface. On the other hand,
when the same experiment is performed on the HZSM-5 (92)
It is well known that both methanol and ethene can be trans-
formed into other hydrocarbons on acidic zeolites. Methanol
can be converted into olefins by the MTO process, and ethene
transformation, herein denoted by ET, to other alkenes and
3
6
alkanes was reported on acid catalysts such as SAPO-34. To
study the inter-reaction between highly reactive methanol and
ethene, it is necessary to distinguish the contributions from
different reactions, particularly for the transformation of
ethene to propene; namely, we need to know if propene is
formed from ethene alone or from the methylation of ethene
1
4602 Phys. Chem. Chem. Phys., 2011, 13, 14598–14605
This journal is c the Owner Societies 2011

View Article Online
a
Table 2 Different reactions on H[B,Al]MFI (41, 331)
b
b
ET
b
Reaction
MTO
CME
Feed composition
Reactants concentration (%)
Conversion (%)
Selectivity to propene (%)
Propene production (mmol h
CH
6
100
64
3
OH C
6
2
H
4
CH
6
3
OH C
6
2 4
H
o2 100
65 58
o0.1 3.7
14
À1 À1
g
) 3.2
a
Reaction conditions: 723 K; reactant concentration in volume;
À1
CH
3
OH and C
2
H
4
diluted in N
2
; overall GHSV of 3000 h ; the
catalysts were exposed to the feed for 4.5 h at elevated temperature
b
prior to the measurements. MTO: methanol to olefins; ET: ethene
transformation; CME: co-reaction of methanol and ethene; the balance
2
N concentration (94%) in the feeds in MTO and ET was higher than
that (88%) in CME.
Fig. 8 The dependence of propene formation on the Al content
of H[B,Al]MFI zeolites and HZSM-5 (92). Reaction conditions:
723 K; 6% CH OH and 6% C H diluted in N ; overall GHSV of
with methanol. Table 2 lists a variety of reactions carried out
on a representative catalyst of H[B,Al]MFI (41, 331) with
different feed components at 723 K. Note that the catalyst had
been exposed to the reactants for 4.5 h at elevated temperatures
3
2
4
2
À1
3000 h
.
relationship between the aluminium contents and the propene
formation can be found, we suggest that the propene formation
(
from 523 to 723 K). When only methanol was introduced
into the reactor, it was transformed completely with a high
selectivity toward propene under the reaction conditions. This
is in agreement with the literature concerning the MTO on
¨
is directly related with the number of Bronsted acid sites asso-
ciated with the framework Al. Moreover, another sample of
HZSM-5 (92), containing no boron, produced a considerable
amount of propene, confirming the proposal that the propene
1
6
MFI zeolites with a low Al content. In contrast, when only
ethene was employed as the reactant, the conversion of ethene
was very low (o2%) and the propene yield was negligible. The
low conversion of ethene obtained at such high temperature is
explained by an extensive coke deposition on the catalyst. In
fact, we observed that ethene was highly reactive on the catalyst
only at the initial stage of the ET process. Nevertheless, when
both methanol and ethene were present in the feed (CME
process), a significant increase in the ethene transformation (14%)
was observed, whereas the conversion of methanol remained
high. Therefore, ethene is more effectively transformed into
the desired product propene in the presence of methanol. In
other words, the co-existence of methanol and ethene resulted
in at least two parallel processes of MTO and CME together
on the MFI zeolite. This is also in agreement with the previous
reports that propene can be formed by the methylation of
ethene with methanol, as evidenced by both experimental
formation is associated with the Bro
¨
nsted acidity related with
framework Al in the zeolites.
The incorporation of different trivalent cations into the
zeolite framework is an efficient method for adjusting the
zeolite acidities. The capability of boron-induced acidity for
the acid-catalyzed reactions remains elusive; there are many
3
7
27
11
conflicting reports. The Al and B NMR spectra (Fig. 1
and 2) of our H[B,Al]MFI zeolites clearly indicate that the
majority of aluminium and boron are incorporated into the
zeolite framework during the synthesis. Accordingly, strong
¨
and relatively weak Bronsted acidities, induced by the frame-
work aluminium and boron, respectively, are co-existent on
H[B,Al]MFI zeolites. Nevertheless, our observation that the
H[B,Al]MFI (36, 2100) catalyst showed negligible catalytic
performance in the CME process, together with the fact that
the propene formation was related to the Al content, point to
1
7,18,21
results (isotope exchange) and theoretical calculation.
the necessity of the Al-induced Bro
¨
whereas the boron-induced acidity, either of Bro
nsted acidity for the reaction,
nsted acids or
However, we have no intention to distinguish between the
contributions of CME and MTO for the propene production
at present because CME can be regarded as a sub-reaction of
the MTO process.
¨
of Lewis acids, is inactive in this reaction. This is also in
agreement with the report that H-[B]-ZSM-5 showed negligible
3
8
activity in MTO, and implicates that ethene cannot be activated
by boron-induced acidity. As we reported in our previous
communication, the H-[B]-MFI catalyst was an efficient catalyst
in the co-reaction of methanol with ethene for the production
of propene, and the catalysts showed remarkable stability
for the reaction at elevated temperatures. We noted in the
communication that Al impurity was significantly present in
Since it is proposed that the methylation of ethene with
methanol is catalyzed by the Bro
¨
nsted acid sites of the
zeolites, it is particularly interesting to study the relation
between the amount of Bronsted acidity with the propene
2
1
¨
production. A series of boron- and aluminium-containing MFI
zeolites with relatively low Al contents, where the majority of
Al atoms were incorporated in the zeolite framework as
3
9
the catalyst and might play catalytic roles. The reaction
results upon H[B,Al]MFI (36, 2100) in this report clearly
2
7
evidenced by the Al NMR (Fig. 1), was employed for the
CME process. The influence of Bronsted acidity of the zeolites
¨
¨
confirmed the conclusion that Al-induced Bronsted acidities
upon the propene formation under the same reaction conditions
is exhibited in Fig. 8. H[B,Al]MFI (36, 2100), with negligible
play the decisive roles. However, it is noted that incorporation
of both boron and aluminium into the MFI zeolite during the
zeolite synthesis might enhance the catalytic performance of
the CME process, particularly in terms of the catalyst life
compared to HZSM-5 that contains only heteroatom of Al.
¨
Bronsted acidity due to its extremely low Al content, showed
little propene formation, but an increase in the Al content led
to enhanced propene formation. Although no strictly linear
This journal is c the Owner Societies 2011
Phys. Chem. Chem. Phys., 2011, 13, 14598–14605 14603

View Article Online
Although theoretical calculation predicts that the methoxy
and aromatics are transformed from the primary olefinic
products, we suggest that the hydrogen-transfer reactions are
catalyzed by the FSAS, whereas the methylation of methanol
with alkene could be catalyzed by the FWAS. Since coke consists
of carbonaceous deposits that are deficient in hydrogen, the
formation of coke must be related with the hydrogen-transfer
reactions. The absence of FSAS in H[B,Al]MFI zeolites, together
with the fact of their relatively low Al contents (Si/Al 490),
might inhibit the coke formation, thus extending the catalyst life.
The knowledge of zeolite acidities is particularly important
for the HZSM-5 zeolite because it is a suitable catalyst for
21
group does not participate in the ethene methylation reaction,
1
it is concluded by the C NMR that the methoxy group
3
4
0
is responsible for aniline methylation on H-Y zeolite. Alter-
natively, direct methylation of methanol may also contribute
to the enhanced production of propene. Our methanol adsorption
experiments indicate that methoxy groups can be formed on
all the hydroxyl groups in the zeolites and that the nature of
these methoxy groups could be different (Fig. 6). In contrast,
the chemisorbed ethene is not a prerequisite for the methylation
of ethene with methanol, because H[B,Al]MFI zeolites, on
which no chemisorption of ethene occurs, show pronounced
activities in the CME process to produce propene (Fig. 7 and 8).
The substrate ethene, if it is activated, is presumed to be
physisorbed on the weak Lewis acid sites or siliceous part of
the zeolite, and such physisorbed ethene could play crucial roles
in the reaction. However, the difference in the adsorption behavior
of ethene between H[B,Al]MFI zeolites and HZSM-5(92) indicates
that there are some intrinsic structural differences between the
catalysts depending on the presence or absence of boron during
the synthesis procedure.
4
5
both MTO and methanol to gasoline (MTG) processes. The
reaction mechanisms between these two processes are not
clear, although stronger acids are generally supposed to be
required for MTG. In our opinion, various crystallographic Al
sites in HZSM-5, due to different geometric features, play
different catalytic roles in the transformation of hydrocarbons.
For HZSM-5 with high aluminium contents, which is used for
MTG, the Al atoms occupy all the T sites although the spatial
Al distribution among the T sites might be different; hence both
FSAS and FWAS are present. After alkenes are produced by
the hydrocarbon-pool mechanism, further transformations
occur on both framework strong acid sites, leading to the
formation of aromatics and alkanes, and framework weak acid
sites, resulting in further sequential methylation of alkenes.
For HZSM-5 with low aluminium contents which has the
potential for the further improvement of MTO, the majority of
the crystallographic aluminium might occupy some certain
T sites that are associated with the framework weak acid sites.
In this case, the hydrogen-transfer reactions are effectively
inhibited, and the olefins formed through hydrocarbon-pool
intermediates become dominant. Therefore, the amount and
nature of the catalyst acidities, together with the shape selectivity
imposed by the topologies of the catalytic materials, decide the
components and the distributions of the products in the
methanol to hydrocarbon transformations.
The MFI topology has 12 T sites in the asymmetric unit and
9
6 T sites in the unit cell for the orthorhombic (Pnma) space
4
1
group. Thus far, the local structure of the catalytically active
Bronsted acid sites in ZSM-5 zeolites has rarely been studied
due to the experimental limitations, and much insight was
¨
42–44
obtained by the theoretical modeling.
Different simulations
result in divergent judgments about the natures of the acidities
on different T sites, but there is a unanimous conclusion that
the acid strengths of various crystallographic Al sites are
different; i.e., the slight different bond angles and distances of
the hydroxyl groups to their corresponding Al result in a variety
2
7
of acid strengths. Moreover, Al MQ MAS NMR spectra clearly
indicate that these crystallographically inequivalent Al sites can be
2
5
divided into two groups. Here, we tentatively assign hydroxyl
2
7
groups associated with the framework Al of Al NMR (Fig. 1)
chemical shift from 57–54 ppm to the framework strong acid sites
2
7
(
FSAS), and framework Al of Al NMR chemical shift between
4. Conclusions
53–51 ppm to framework weak acid sites (FWAS).
Boron- and aluminium- containing zeolites are found to have
different acidities compared with their boron-free MFI counter-
part (HZSM-5). It is proposed that boron substitution for Si,
during the synthesis, prevents the Al atoms from occupying the
T sites that are associated with framework strong acid sites,
and facilitates the allocation of Al to the framework positions
that are related with framework weak acid sites. The different
natures of the acidities owing to different crystallographical Al
sites can be differentiated by the reactivity of adsorbed ethene
to form ethene oligomers at elevated temperatures. Probe
reactions of co-reaction of methanol and ethene are tested
on the catalysts. It is found that the propene production is
related with the Al content in the zeolite, while the acidities
induced by the boron incorporation into the zeolites do not
catalyze this reaction.
The assignments above are used to account for the difference
in the ethene adsorption behavior between the H[B,Al]MFI
and HZSM-5 (92) zeolites (Fig. 7). The criterion between the
FSAS and FWAS is whether ethene can be protonated on the
acid sites to form ethene oligomers. FSAS is observed on
27
HZSM-5 (92) with the Al NMR band at 56 ppm (Fig. 1), being
corroborated by the notable feature in ethene oligomerization
when the catalyst is exposed to ethene at 473 K (Fig. 7). In
contrast, during the H[B,Al]MFI zeolite synthesis, the boron
might occupy, or prevent aluminium from siting in, the T sites
which are associated with the FSAS, resulting in the location of
Al atoms in the T positions related with FWAS. Therefore, the
majority of framework Al atoms of H[B,Al]MFI zeolites appear
2
7
in the Al NMR spectra at 53 ppm (Fig. 1), and their relatively
weak Bronsted acidity is evidenced by the incapability to activate
ethene to form oligomers (Fig. 7). NH is a stronger base than
ethene so it is suggested that NH can interact with both FSAS
and FWAS.
¨
3
3
References
1
2
¨
M. Stocker, Microporous Mesoporous Mater., 1999, 29, 3–48.
C. D. Chang, Catal. Rev., 1983, 25, 1–118.
Because hydrogen-transfer reactions are accompanied by the
formation of ethene oligomers on the zeolite surface and alkanes
3 I. M. Dahl and S. Kolboe, J. Catal., 1994, 149, 458–464.
1
4604 Phys. Chem. Chem. Phys., 2011, 13, 14598–14605
This journal is c the Owner Societies 2011

View Article Online
4
5
J. F. Haw, J. B. Nicolas, W. Song, F. Deng, Z. Wang, T. Xu and
C. S. Heneghan, J. Am. Chem. Soc., 2000, 122, 4763–4775.
W. Song, J. F. Haw, J. B. Nicholas and C. S. Heneghan, J. Am.
Chem. Soc., 2000, 122, 10726–10727.
25 O. H. Han, C.-S. Kim and S. B. Hong, Angew. Chem., Int. Ed.,
2002, 41, 469–472.
26 S.-J. Hwang, C.-Y. Chen and S. I. Zones, J. Phys. Chem. B, 2004,
108, 18535–18546.
6
7
B. Arstad and S. Kolboe, J. Am. Chem. Soc., 2001, 123, 8137–8138.
A. Sassi, M. A. Wildman, H. J. Ahn, P. Prasad, J. B. Nicolas and
J. F. Haw, J. Phys. Chem. B, 2002, 106, 2294–2303.
M. Bjørgen, U. Olsbye, D. Petersen and S. Kolboe, J. Catal., 2004,
2
27 L. Chen, M. Zhang, Y. Yue, C. Ye and F. Deng, Microporous
Mesoporous Mater., 2004, 76, 151–156.
28 C. Fild, D. F. Shantz, R. F. Lobo and H. Koller, Phys. Chem.
Chem. Phys., 2000, 2, 3091–3098.
29 D. Ma, F. Deng, R. Fu, X. Han and X. Bao, J. Phys. Chem. B,
2001, 105, 1770–1779.
30 M. Niwa and N. Katada, Catal. Surv. Jpn., 1997, 1, 215–226.
31 A. Cichocki, J. Datka, A. Olech and Z. Piwowarska, J. Chem. Soc.,
Faraday Trans., 1990, 86, 753–756.
8
9
21, 1–10.
B. Arstad, J. B. Nicholas and J. F. Haw, J. Am. Chem. Soc., 2004,
26, 2991–3001.
0 W. Wang, Y. Jiang and M. Hunger, Catal. Today, 2006, 113,
02–114.
1
1
1
1
1
1 S. Svelle, F. Joensen, J. Nerlov, U. Olsbye, K. Lillerud, S. Kolboe
and M. Bjørgen, J. Am. Chem. Soc., 2006, 128, 14770–14771.
2 D. Lesthaeghe, B. De Sterck, V. van Speybroeck, G. B. Marin and
M. Waroquier, Angew. Chem., Int. Ed., 2007, 46, 1311–1314.
3 I. M. Dahl and S. Kolboe, Catal. Lett., 1993, 20, 329–336.
4 F. Bleken, M. Bjørgen, L. Palumbo, S. Bordiga, S. Svelle,
K. Lillerud and U. Olsbye, Top. Catal., 2009, 52, 218–228.
5 R. M. Dessau and R. B. Lapierre, J. Catal., 1982, 78, 136–141.
6 M. M. Wu and W. W. Kaeding, J. Catal., 1984, 88, 478–489.
7 S. Svelle, P. O. Rønning and S. Kolboe, J. Catal., 2004, 224,
32 J. Datka and Z. Piwowarska, J. Chem. Soc., Faraday Trans. 1,
1989, 85, 837–841.
33 S. M. Campbell, X.-Z. Jiang and R. F. Howe, Microporous
Mesoporous Mater., 1999, 29, 91–108.
34 A. G. Stepanov, M. V. Luzgin, V. N. Romannikov,
V. N. Sidelnikov and E. A. Paukshtis, J. Catal., 1998, 178, 466–477.
35 E. Yoda, J. N. Kondo and K. Domen, J. Phys. Chem. B, 2005, 109,
1464–1472.
36 H. Oikawa, Y. Shibata, K. Inazu, Y. Iwase, K. Murai, S. Hyodo,
G. Kobayashi and T. Baba, Appl. Catal., A, 2006, 312, 181–185.
37 R. Millini, G. Perego and G. Bellussi, Top. Catal., 1999, 9, 13–34.
38 E. Unneberg and S. Kolboe, Appl. Catal., A, 1995, 124, 345–354.
39 Q. Zhu, J. N. Kondo, M. Yamaguchi, T. Setoyama, K. Domen
and T. Tatsumi, Chem. Commun., 2008, 5164–5166.
40 I. I. Ivanova, E. B. Pomakhina, A. I. Rebrov, M. Hunger,
Y. G. Kolyagin and J. Weitkamp, J. Catal., 2001, 203, 375–381.
41 D. H. Olson, G. T. Kokotailo and S. L. Lawton, J. Phys. Chem.,
1981, 85, 2238–2243.
1
1
1
1
1
1
15–123.
8 S. Svelle, P. O. Rønning, U. Olsbye and S. Kolboe, J. Catal., 2005,
34, 385–400.
9 J. F. Haw, D. M. Marcus and P. W. Kletnieks, J. Catal., 2006, 244,
30–133.
0 Y. Jiang, W. Wang, V. R. R. Marthala, J. Huang, B. Sulikowski
and M. Hunger, J. Catal., 2006, 244, 134–136.
1 S. Svelle, B. Arstad, S. Kolboe and O. Swang, J. Phys. Chem. B,
1
1
2
2
2
2
1
2003, 107, 9281–9289.
42 A. Redondo and P. J. Hay, J. Phys. Chem., 1993, 97, 11754–11761.
2 Q. Zhu, J. N. Kondo, T. Tatsumi, S. Inagaki, R. Ohnuma,
Y. Kubota, Y. Shimodaira, H. Kobayashi and K. Domen,
J. Phys. Chem. C, 2007, 111, 5409–5415.
43 R. Crau-Crespo, A. G. Peralta, A. R. Ruiz-Salvador, A. Go
and R. Lopez-Cordero, Phys. Chem. Chem. Phys., 2000, 2,
5716–5722.
44 S. Sklenak, J. De
´
mez
´
23 Q. Zhu, J. N. Kondo, S. Inagaki and T. Tatsumi, Top. Catal., 2009,
decek, C. Li, B. Wichterlova, V. Gabova,
ˇ ˇ ´ ´ ´
5
2, 1272–1280.
M. Sierka and J. Sauer, Phys. Chem. Chem. Phys., 2009, 11,
1237–1247.
45 S. M. Campbell, D. M. Bibby, J. M. Coddington, R. F. Howe and
R. H. Menhold, J. Catal., 1996, 161, 338–349.
2
4 Q. Zhu, B. L. Mojet, R. A. J. Janssen, E. J. M. Hensen,
J. van Grondelle, P. C. M. M. Magusin and R. A. van Santen,
Catal. Lett., 2002, 81, 205–212.
This journal is c the Owner Societies 2011
Phys. Chem. Chem. Phys., 2011, 13, 14598–14605 14605