Catalysis Letters 62 (1999) 139–145 139

Conversion of methanol to hydrocarbons on zeolite HZSM-5

investigated by in situ MAS NMR spectroscopy under flow
conditions and on-line gas chromatography
Michael Seiler, Udo Schenk and Michael Hunger ∗
Institute of Chemical Technology I, University of Stuttgart, D-70550 Stuttgart, Germany
E-mail: michael hunger@po.uni-stuttgart.de

Received 30 June 1999; accepted 13 August 1999

In situ MAS NMR spectroscopy under flow conditions and on-line gas chromatography have been applied to study the onset of the
conversion of methanol on zeolite HZSM-5 at temperatures between 373 and 573 K. In the steady states of methanol conversion at
T > 523 K, by on-line gas chromatography mainly the formation of ethene and propene was observed. Simultaneously recorded in
situ 13 C MAS NMR spectra show signals at 12–25 ppm and at ca. 125–131 ppm indicating the presence of adsorbed C4 –C6 olefins.
The observation of these adsorbates on a working catalyst supports the “hydrocarbon pool” mechanism previously proposed for the
methanol-to-hydrocarbon conversion on acidic zeolites. Methanol conversion at 473 and 573 K and subsequent purging of the catalyst
with dry nitrogen at 293 K led to a 13 C MAS NMR signal at 59 ppm due to methoxy groups. No hints to the presence of ethoxy, propoxy
or butoxy groups and the formation of alkyl oxonium ions were found by in situ 13 C MAS NMR spectroscopy under flow conditions.
Keywords: heterogeneous catalysis, methanol-to-hydrocarbon conversion, MTG, zeolite HZSM-5, Brønsted acid sites, in situ MAS
NMR spectroscopy, flow conditions

1. Introduction Based on GC-MS experiments with 13 C- and 12 C-

labelled reactants, Dahl and Kolboe (see [12,13]) explained
Since the first reports on methanol-to-gasoline (MTG) the formation of olefins by a “hydrocarbon pool” mecha-
conversion on zeolite HZSM-5 in 1976, this process is nism which represents adsorbates such as
one of the mostly investigated reactions in heterogeneous
catalysis [1–3]. Although a number of experimental and
theoretical works have established various mechanisms for
the conversion of methanol on zeolites (see [4–8], and ref-
erences therein), the chemical reactions proceeding at the
Brønsted acid sites of these catalysts are poorly understood.
Generally, three reaction steps are distinguished: (i) the de- According to this mechanism, “big” carbonaceous species
hydration of methanol to dimethyl ether, (ii) the conversion exist inside the zeolite pores and add and split off reactants
of a mixture of methanol and dimethyl ether to olefins and and products, respectively [12,13]. The “hydrocarbon pool”
(iii) a bond chain polymerization of olefins and isomeriza- represents adsorbates with many characteristics of ordinary
tion [8]. Most of the recently published theoretical works coke and would perhaps be better represented by (CHx )n
investigate the mechanisms for the first C–C bond forma- with 0 < x < 2 [11].
tion. As a reactive intermediate in steps (i) and (ii) methoxy Since 1989, with the first in situ MAS NMR study
groups bound to oxygen atoms in Si–O–Al bridges of the of methanol conversion on zeolite HZSM-5 carried out
zeolite framework are proposed. by Anderson and Klinowski [14–17], this spectroscopic
An often discussed way for the formation of the first method has been developed to a useful tool for the inves-
C–C bond is the oxonium ylide mechanism [9,10]. Ac- tigation of heterogeneously catalyzed reactions. In the last
cording to this mechanism, trimethyl oxonium ions formed decade a growing number of groups performed in situ MAS
by a reaction of dimethyl ether molecules are deprotonated NMR spectroscopic studies of the conversion of methanol
by a basic site and yield dimethyl oxonium methyl ylide. on various aluminosilicate- and silicoaluminophosphate-
After an intramolecular Stevens rearrangement [9] or an type zeolites under batch conditions [18–25]. Applying
intermolecular methylation [10], ethylmethyl oxonium or a pulse-quench reactor, Goguen et al. [26] made methanol
conversion experiments on zeolite HZSM-5 under semi-
dimethylethyl oxonium ions are formed and split off the
flow conditions. After giving a reactant pulse on the cat-
reaction product ethene (see also [11]).
alyst filled into an external reactor tube, the reaction was
∗ To whom correspondence should be addressed. rapidly stopped by quenching with liquid nitrogen. Sub-

 J.C. Baltzer AG, Science Publishers

140 M. Seiler et al. / MTG on zeolite HZSM-5 studied by in situ 13 C MAS NMR spectroscopy

sequently, the catalyst loaded with reactants and reaction iodide, C∗ H3 I (Aldrich, 27.718-5, the asterisk indicates a
13
products was transferred into the precooled MAS NMR C-enrichment of the carbon atom) was adsorbed using a
probe. Recently, Isbester et al. [27] introduced a MAS vacuum line and, subsequently, heated at 473 K for 1 h.
NMR probe with an isolated reactant flow. In a first ap- Before the in situ MAS NMR experiments under flow
plication, this probe was used to observe the formation of conditions, the calcined zeolite was filled into the MAS
dimethyl ether due to conversion of methanol on zeolite NMR rotor reactor under dry nitrogen in a glove box and
HZSM-5 without further gas chromatographic investiga- pressed to a cylindrical catalyst bed (see [29]). After trans-
tions of hydrocarbon formation. By our group, a technique ferring the rotor into the MAS NMR probe, a second dehy-
for the continuous injection of reactant molecules into the dration of the catalyst was performed at 673 K for 1 h un-
spinning rotor of commercial variable temperature MAS der flowing nitrogen (30 ml/min). During the in situ MAS
NMR probes was introduced [28]. In a very recent work, NMR experiments under continuous-flow conditions, nitro-
this injection technique was coupled with on-line gas chro- gen loaded with methanol (me), C∗ H3 OH (Campro Sci-
matography allowing simultaneous in situ MAS NMR stud- entific, 83-00005-9), or a mixture of C∗ H3 OH and ethene
ies of adsorbate complexes formed under flow conditions (C=2 ), C2 H4 (Fluka, 398815/1), were injected into a 7 mm
and gas chromatographical analysis of the reaction prod- MAS NMR rotor applying the equipment described else-
ucts [29]. In the present study, a high temperature MAS where [29]. Via an exhaust tube on the top of the MAS
NMR probe was equipped with such an injection system NMR rotor reactor, the probe was connected with the sam-
allowing investigations of heterogeneously catalyzed reac- pling loop of a gas chromatograph (see [29]). The in situ
tions under flow conditions at temperatures up to 873 K. MAS NMR flow probe is based on a modified DSI-740
This equipment was applied to study the formation of hy- 7 mm STD MAS NB NMR probe delivered by Doty Sci-
drocarbons by conversion of methanol on zeolite HZSM-5. entific Instruments, Columbia, Texas, allowing measure-
As done in previous works [16,17,24,25], especially the on- ments at temperatures up to 873 K with sample spinning
set of this reaction at temperatures up to 573 K was studied. rates up to 3.5 kHz. By 207 Pb MAS NMR spectroscopy
of Pb(NO3 )2 , the temperature inside the MAS NMR rotor
reactor and the temperature gradient over the whole sam-
2. Experimental ple volume of the modified in situ MAS NMR flow probe
were investigated. According to Ferguson and Haw [31],
Zeolite NaZSM-5 with the chemical composition the 207 Pb NMR shift of Pb(NO3 )2 changes by 0.775 ppm/K
Na4.2 Al4.2 Si91.8 O192 was synthesized as described else- with the temperature. The 207 Pb MAS NMR spectra of
where [30]. The ammonium form was prepared by four- Pb(NO3 )2 shown in figure 1 were recorded using the mod-
fold ion exchange at 353 K in a 0.4 M aqueous solution of ified in situ MAS NMR flow probe during injection of ni-
NH4 NO3 . After reaching an ion exchange degree of 98%, trogen (15 ml/min) into the MAS NMR rotor spinning with
the zeolite powder was washed in demineralized water and
dried at room temperature. The material was characterized
by AES-ICP, XRD, 27 Al and 29 Si MAS NMR spectroscopy.
No signal of extra-framework aluminum atoms was found
in the 27 Al MAS NMR spectra of the hydrated samples.
The MAS NMR investigations were performed on a
Bruker MSL 400 spectrometer at resonance frequencies of
400.1 MHz for 1 H, 100.6 MHz for 13 C nuclei and 83.5 MHz
for 207 Pb and with sample spinning rates of 2.4 kHz for 1 H
and 1.8 kHz for 13 C and 207 Pb MAS NMR spectroscopy.
For bearing and driving the MAS NMR turbine, the same
dry nitrogen gas (produced by a nitrogen evaporator) was
used as for the carrier gas in the catalytic flow experiments
(vide infra). The 13 C MAS NMR spectra were obtained
after direct excitation and proton decoupling. For each 1 H
MAS NMR spectrum, 100 free induction decays and for the
13
C and 207 Pb MAS NMR spectra from 100 to 2000 free
induction decays were accumulated with a repetition time
of 5 s. The 1 H and 13 C MAS NMR spectra were referred
to TMS.
Prior to the NMR investigations, the ammonium form of Figure 1. 207 Pb MAS NMR spectra of Pb(NO3 )2 recorded using the
modified in situ MAS NMR flow probe (see text) during injection of
zeolite was heated in vacuum with a rate of 20 K/h up to nitrogen (15 ml/min) into the MAS NMR rotor spinning with a rate of
the final temperature of 723 K leading to zeolite HZSM-5. 1.8 kHz. The temperature gradients given in the figure were calculated
There, it was calcined at a pressure below 10−2 Pa for by the residual linewidth of the 207 Pb MAS NMR signals obtained at the
12 h. On one sample of calcined zeolite HZSM-5, methyl corresponding temperatures and the factor 0.775 ppm/K [28].
 M. Seiler et al. / MTG on zeolite HZSM-5 studied by in situ 13 C MAS NMR spectroscopy 141

1
a rate of 1.8 kHz. At the maximum reaction temperature of H MAS NMR signals of a dehydrated and intact zeolite
573 K applied in the present work a temperature gradient HZSM-5: a weak shoulder at 1.8 ppm due to silanol groups
of ∆T = 1.5 K was calculated via the residual linewidth of (SiOH), a dominating large signal at 4.2 ppm caused by
the 207 Pb MAS NMR signal determined at this temperature. acidic bridging OH groups (SiOHAl) and a weak signal at
The chemical shift was referred to 0 ppm for Pb(NO3 )2 at 6.8 ppm which indicates a small amount of residual ammo-
293 K. nium ions [32].
The reaction products were analyzed using a gas chro- The catalytic investigations were started at a temperature
matograph HP 5890 (Hewlett–Packard) equipped with a of 373 K (figure 3(a)). At this temperature no formation of
Coating Poraplot Q capillary column (Chrompack Plot dimethyl ether (dme) and hydrocarbons was found by GC
fused silica, length 50 m, inner diameter 0.32 mm). The (data on the left-hand side). The simultaneously recorded
exhaust flow was sampled and analyzed in steps of 30 min. 13
C MAS NMR spectrum consists only of a single line at
The reactant flow was adjusted by the carrier gas (dry 50 ppm due to methanol [33]. The integral intensity of
nitrogen) flow rate and the temperature-dependent partial the methanol peak obtained by GC was used as an inten-
pressure of methanol in the saturator. In the in situ MAS sity standard for the further experiments. After raising the
NMR flow experiments, 100 mg of dry zeolite HZSM-5 reaction temperature to 423 K, a yield of dimethyl ether
and a modified residence time of W/Fme = 25 g h/mol (dme) of 10.7% was determined by GC and the 13 C MAS
were used. NMR spectrum shows an additional signal at 61 ppm due to
dme [33] (figure 3(b)). At a reaction temperature of 473 K,
a significant increase of the signal at 61 ppm as well as
3. Results and discussion of the yield of dme occurred (figure 3(c)). The intensities
3.1. Conversion of methanol on zeolite HZSM-5 at of the 13 C MAS NMR signals of methanol and dme de-
temperatures of 373–573 K viate from the amounts of reactant molecules determined
by GC. This is due to differences in the residence times
The first step of an in situ MAS NMR experiment under of these weakly adsorbed molecules inside the MAS NMR
flow conditions was the dehydration of the zeolite filled rotor reactor. A residence time shorter than the spin–lattice
into the MAS NMR rotor reactor. To prove the correct relaxation time T1 causes a decrease or disappearance of
preparation of the zeolite, a 1 H MAS NMR spectrum was the MAS NMR signals of the corresponding molecules and
taken at 293 K after treating the catalyst at 673 K under surface compounds.
in situ conditions. Because of the low sample spinning Starting at the reaction temperature of 523 K, the forma-
rate (νrot = 2.4 kHz), the central line of the 1 H MAS tion of hydrocarbons was observed (figure 3 (d)–(f)). The
NMR spectrum shown in figure 2(a) is strongly broad- yield of dimethyl ether (dme) as well as those of ethene
ened. Inspite of this broadening, it shows the characteristic (C= = =
2 ), propene (C3 ), propane (C3 ), butenes (C4 ) and bu-
tanes (C4 ) increased (GC data on the left-hand side). In
the 13 C MAS NMR spectra, the formation of hydrocar-
bons is accompanied with the appearance of signals at 12–
25 ppm and ca. 125–131 ppm. Interestingly, no signal of
ethene was found at 123 ppm [33] which indicates a short
residence time of this molecule inside the MAS NMR ro-
tor reactor. The increase of the hydrocarbon yields at the
reaction temperatures of 548 and 573 K is accompanied
by an increase of the 13 C MAS NMR signals at 15 and
18 ppm and of broad signals at ca. 125–131 ppm (figure 3
(e) and (f)). Although the latter signals lie in the aromatic
region, neither benzene (128.5 ppm [33]) nor toluene (21.4,
125.7, 128.5, 129.3 and 137.8 ppm [33]), p-xylene (20.9,
129.1 and 134.5 ppm [33]) or naphthalene (125.6, 127.7
and 133.3 ppm [33]) can be assigned to these signals. All
these molecules would cause additional signals out of the
range of resonances observed in figure 3 (e) and (f) and/or
they do not allow an explanation of the signals in the re-
gion between 12 and 25 ppm. Another situation occurs,
considering trans-butene-2 (16.8 and 125.4 ppm [33]), cis-
Figure 2. 1 H MAS NMR spectrum of zeolite HZSM-5 recorded after butene-2 (11.4 and 124.2 ppm [33]), n-pentene-2 (13.6,
dehydration in the in situ MAS NMR probe at 673 K immediately before
starting the methanol flow (a). Spectrum (b) was obtained at 293 K after
17.3, 25.8, 123.8 and 133.2 ppm [33]), n-hexene-3 (14.1,
conversion of methanol at 573 K and purging the catalyst with dry nitrogen 25.9 and 131.2 ppm [33]) and n-hexadiene-2,4 (17.5, 126.2
gas (30 ml/min) at this temperature. Both spectra were recorded with a and 132.5 ppm [33]). A mixture of these C4 –C6 olefins
sample spinning rate of 2.4 kHz. Asterisks denote spinning sidebands. would lead to signals at 12–25 ppm as well as to broad res-
142 M. Seiler et al. / MTG on zeolite HZSM-5 studied by in situ 13 C MAS NMR spectroscopy

Figure 3. 13 C MAS NMR spectra of the conversion of C∗ H3 OH on zeolite HZSM-5 recorded under flow conditions with a modified residence time
of W/Fme = 25 g h/mol and at temperatures between 373 and 573 K. On the left-hand side the yields (in %) of dimethyl ether (dme), ethene (C=
2 ),
propene (C=
3 ), propane (C3 ), butenes (C= ) and butanes (C ) determined by simultaneously performed on-line GC analysis are given.
4 4

onances at ca. 125–131 ppm. Hence, the 13 C MAS NMR 3.2. Formation of methoxy groups
spectra recorded at 523–573 K under flow conditions hint
to the presence of a “hydrocarbon pool”, which indicates In most of the reaction mechanisms discussed for the
a formation of hydrocarbons according to the mechanism MTG process, methoxy groups bound to oxygen atoms of
proposed by Dahl and Kolboe [12,13]. A formation of hy- Si–O–Al bridges in the zeolite lattice play a important role
drocarbons according to the “rake mechanism” proposed by as methylating intermediates (see, e.g., [8]). In a previ-
Cormerais et al. [34] cannot be supported. In this case, the ous work, Bosacek [40] prepared methoxy groups on ze-
“hydrocarbon pool” would consist of ethoxy (68 ppm [35]), olite HZSM-5 by loading methyl iodide and observed a
propoxy (87 ppm [36]) and butoxy groups (77 ppm [37]) narrow 13 C MAS NMR signal at the isotropic chemical
which were not observed in the present work. Furthermore, shift of 59.4 ppm. Applying two-dimensional 13 C magic-
no indication for the formation of trimethyl oxonium ions angle-turning NMR spectroscopy (MAT NMR) on zeolite
(80 ppm [38,39]) was found. HZSM-5 after loading with methanol, heating at 473 K for
 M. Seiler et al. / MTG on zeolite HZSM-5 studied by in situ 13 C MAS NMR spectroscopy 143

from the value obtained by Philippou et al. [41] applying

2D 13 C MAT NMR spectroscopy (∆σ = −49 ppm). There-
fore, the zeolite HZSM-5 used in the present study was
pretreated as described by Bosacek [40], i.e., the zeolite
calcined at 673 K was loaded with methyl iodide (C∗ H3 I)
and, subsequently, heated at 473 K for 1 h. The 13 C MAS
NMR signal of the methoxy groups on the zeolite HZSM-5
prepared in this way (figure 4(c)) agrees very well with
the low-field patterns of the spectra in figure 4 (a) and (b).
This supports the assignment of the signal at 59 ppm in
figure 4 (a) and (b) to methoxy groups formed at frame-
work oxygen atoms. Hence, the conversion of methanol
on zeolite HZSM-5 under flow conditions is accompanied
by the formation of methoxy groups which are absent in
the spectra recorded at 573 K (figure 3(f)) because of their
short lifetime at high reaction temperatures. It is important
to note that also in the spectra recorded at room temperature
after conversion of methanol at 473 and 573 K, no signals
due to ethoxy, propoxy or butoxy groups (77–87 ppm [35–
37]) and trimethyl oxonium ions (80 ppm [38,39]) were
observed.

3.3. Conversion of a methanol/ethene feed on zeolite

HZSM-5

In a number of works [7,12,13], the role of ethene

in the MTG process was investigated. Ethene has been
Figure 4. 13 C MAS NMR spectra recorded at 293 K after conversion of discussed as a product of the first C–C bond formation.
C∗ H3 OH on zeolite HZSM-5 at 473 (a) and 573 K (b) with W/Fme = Dahl and Kolboe [13] studied the conversion of a feed of
13
25 g h/mol and purging the catalyst with dry nitrogen gas (30 ml/min) at C-methanol and 12 C-ethene on SAPO-34 by GC-MS and
293 K. The spectrum in (c) was obtained after loading the calcined zeolite
found that a large part of ethene leave the catalyst unreacted
HZSM-5 with C∗ H3 I and heating at 473 K. All spectra were recorded with
a sample spinning rate of 1.8 kHz. Asterisks denote spinning sidebands. while methanol was almost completely converted to hydro-
The decomposition and simulation of the spectra were performed with the carbons. According to this finding, the authors suggested
Bruker software WINFIT. that higher hydrocarbons are not formed by a successive
methylation of ethene.
15 min and subsequent evacuation, Philippou et al. [41] In the present work, the co-reaction of 13 C-methanol
found a signal at 58.4 ppm which the authors assigned to and ethene (natural abundance of 13 C-isotopes) on zeolite
methoxy groups formed on this zeolite. Considering the HZSM-5 was studied applying in situ 13 C MAS NMR spec-
13
C MAS NMR spectra shown in figure 3 (b)–(d), a signal troscopy under flow conditions. These experiments were
of methoxy groups at ca. 59 ppm should be covered by the performed at reaction temperatures of 548 and 573 K with
signal of dimethyl ether at 61 ppm. At higher reaction tem- a molar feed ratio of ṅme /ṅC=
2
= 2 : 1 and a modified resi-
peratures, the lifetime of the methoxy species could be too dence time of methanol of W/Fme = 25 g h/mol (figure 5
short for detection by 13 C MAS NMR spectroscopy. There- (a) and (b)). Excepting two weak signals at 47 and 61 ppm,
fore, after the conversion of methanol at a certain reaction the spectrum shown in figure 5(a) agrees well with that ob-
temperature, the methanol flow was stopped and the catalyst tained at 548 K for conversion of pure methanol on zeolite
was cooled down to 293 K and purged with dry nitrogen gas HZSM-5 (compare figure 3(e)). The product distribution
(30 ml/min). Subsequently, 13 C MAS NMR spectra were determined by GC deviates only slightly from that found for
recorded of the purged samples. In figure 4 (a) and (b), the conversion of pure methanol. In contrast to this find-
the 13 C MAS NMR spectra obtained after conversion of ing, the spectrum recorded at 573 K (figure 5(b)) shows
methanol at 473 and 573 K, respectively, and purging are differences in comparison with that obtained for conver-
shown. The low-field range of both spectra consists of an sion of pure methanol at the same temperature (compare
MAS NMR sideband pattern with a central line at 59 ppm. figure 3(f)). The high-field region is dominated by the nar-
The simulation of these sideband patterns yielded a chem- row signal at 18 ppm and the broad signals in the low-
ical shift anisotropy of ∆σ = −23 ppm and an asymmetry field region are increased. These effects can be explained
parameter, η, of 0.2–0.4. While the isotropic chemical shift by a higher content of n-hexadiene-2,4 (17.5, 126.2 and
of 59 ppm agrees well with that given in literature [40,41], 132.5 ppm [33]) in the “hydrocarbon pool” formed in the
the chemical shift anisotropy of ∆σ = −23 ppm differs zeolite pores. This variation in the composition of the “hy-
144 M. Seiler et al. / MTG on zeolite HZSM-5 studied by in situ 13 C MAS NMR spectroscopy

Figure 5. 13 C MAS NMR spectra of the conversion of a C∗ H3 OH/ethene feed on zeolite HZSM-5 recorded under flow conditions with a modified
residence time of W/Fme = 25 g h/mol, a molar methanol/ethene feed ratio of ṅme /ṅC= 2
= 2 : 1 and at temperatures of 548 (a) and 573 K (b). The
spectrum shown in (c) was recorded after catalysis according to (b) and subsequent purging of the catalyst with dry nitrogen gas (30 ml/min) at 573 K.
On the left-hand side the yields (in %) of dimethyl ether (dme), propene (C= =
3 ), propane (C3 ), butenes (C4 ), butanes (C4 ) and higher hydrocarbons
(C5+ ) determined by simultaneously performed on-line GC analysis are given.

drocarbon pool” seems to be caused by the co-injection tion temperatures, the residence time of these reactants on
of ethene, but has no significant influence on the product the catalysts surface becomes short in comparison to their
distribution (GC data on the left-hand side). Purging the residence time in the MAS NMR rotor reactor and their
catalyst at 573 K with dry nitrogen after conversion of the 13
C MAS NMR signals disappear. A similar behaviour
methanol/ethene feed leads to a continuous decrease of the was found for methoxy groups which are discussed in the
13
C MAS NMR signals (figure 5(c)). After purging the literature as methylating intermediates. At reaction temper-
catalyst at 573 K for 2 h, no 13 C MAS NMR signals could atures of T > 423 K their detection by 13 C MAS NMR
be observed, i.e., all hydrocarbons were removed from the spectroscopy is difficult. However, a signal observed at
catalyst. In agreement with this finding, the 1 H MAS NMR 59 ppm in the 13 C MAS NMR spectra recorded at 293 K,
spectrum recorded at 293 K after the purging consists of a after methanol conversion at reaction temperatures of 473
strong sideband pattern due to bridging OH groups without or 573 K and purging with nitrogen, indicates that these
a hint to residual adsorbates (figure 2(b)). Interestingly, dur- species are present on the working catalyst. Starting at
ing the first 15 min of the catalyst purging at 573 K, ethene 523 K, hydrocarbons were determined by on-line GC, ac-
and propene were determined by GC as main desorptives. companied with the occurrence of 13 C MAS NMR signals
This shows that the higher olefins contributing to the previ- at 12–25 ppm and ca. 125–131 ppm. Considering the res-
ously existing “hydrocarbon pool” are split off into smaller
onance positions of aromatic and olefinic molecules, the
compounds before the desorption from the surface of the
above-mentioned signals are a hint to the formation of a
zeolite HZSM-5 occurs.
pool of hydrocarbons probably consisting of a mixture of
C4 –C6 olefins. According to the “hydrocarbon pool” mech-
4. Conclusions anism dicussed in the literature, these carbonaceous species
exist inside the zeolite pores and add and split off reactants
In situ 13 C MAS NMR spectroscopy under flow con- and product molecules. Co-injection of ethene led to a
ditions and on-line gas chromatography were successfully shift of the chemical equilibrium of the olefins contributing
applied to study the conversion of methanol on zeolite the “hydrocarbon pool” to higher olefins, probably to hexa-
HZSM-5. At reaction temperatures of T 6 523 K methanol diene. However, no significant variation of the product
and dimethyl ether could be observed in the 13 C MAS NMR distribution was determined by simultaneously performed
spectra at 50 and 61 ppm, respectively. At higher reac- on-line GC analysis. Purging of the zeolite HZSM-5 after
 M. Seiler et al. / MTG on zeolite HZSM-5 studied by in situ 13 C MAS NMR spectroscopy 145

conversion of methanol at 573 K caused a disappearance [17] M.W. Anderson, M.L. Occelli and J. Klinowski, J. Phys. Chem. 96
of the 13 C MAS NMR signals at 12–25 ppm and ca. 125– (1992) 388.
[18] E.J. Munson, N.D. Lazo, M.E. Moellenhoff and J.F. Haw, J. Am.
131 ppm and a desorption of light olefins.
Chem. Soc. 113 (1991) 2783.
[19] E.J. Munson, A.A. Kheir, N.D. Lazo and J.F. Haw, J. Phys. Chem.
96 (1992) 7740.
Acknowledgement [20] H. Ernst, D. Freude and T. Mildner, Chem. Phys. Lett. 229 (1994)
291.
The authors gratefully acknowledge the Deutsche For- [21] H. He, L. Zhang and J. Klinowski, J. Phys. Chem. 99 (1995) 6980.
schungsgemeinschaft, Max-Buchner-Forschungsstiftung [22] M.D. Alba, A.A. Romero, M.L. Occelli and J. Klinowski, J. Chem.
Soc. Faraday Trans. 93 (1997) 1221.
and Fonds der Chemischen Industrie for financial support
[23] M.D. Alba, A.A. Romero, M.L. Occelli and J. Klinowski, J. Phys.
and Jens Weitkamp (University Stuttgart, Germany) for Chem. B 101 (1997) 5166.
helpful discussions. [24] F. Salehirad and M.W. Anderson, J. Catal. 164 (1996) 301; 177
(1998) 189.
[25] F. Salehirad and M.W. Anderson, J. Chem. Soc. Faraday Trans. 94
References (1998) 1911; 94 (1998) 2857.
[26] P.W. Goguen, T. Xu, D.H. Barich, T.W. Skloss, W. Song, Z. Wang,
[1] S.L. Meisel, J.P. McCullough, C.H. Lechthaler and P.B. Weisz, J.B. Nicholas and J.F. Haw, J. Am. Chem. Soc. 120 (1998) 2650.
CHEMTECH 6 (1976) 86. [27] P.K. Isbester, A. Zalusky, D.H. Lewis, M.C. Douskey, M.J. Pomije,
[2] C.D. Chang and A.J. Silvestri, J. Catal. 47 (1977) 249. K.R. Mann and E.J. Munson, Catal. Today 49 (1999) 363.
[3] C.D. Chang, Catal. Rev. Sci. Eng. 25 (1983) 1; 26 (1984) 323. [28] M. Hunger and T. Horvath, J. Chem. Soc. Chem. Commun. (1995)
[4] G.J. Hutchings, D.F. Lee and M. Lynch, Appl. Calal. A 106 (1993) 1423.
115. [29] M. Hunger, M. Seiler and T. Horvath, Catal. Lett. 57 (1999) 199.
[5] P.E. Sinclair and C.R.A. Catlow, J. Chem. Soc. Faraday Trans. 92 [30] S. Ernst and J. Weitkamp, Chem. Ing. Tech. 93 (1991) 748.
(1996) 2099. [31] D.B. Ferguson and J.F. Haw, Anal. Chem. 67 (1993) 3342.
[6] R. Shah, J.D. Gale and M.C. Payne, J. Phys. Chem. B 101 (1997) [32] M. Hunger, Catal. Rev. Sci. Eng. 39 (1997) 345.
4787. [33] H.-O. Kalinowski, S. Berger and S. Braun, 13C-NMR-Spektroskopie
[7] S.R. Blaszkowski and R.A. van Santen, J. Am. Chem. Soc. 119 (Georg Thieme, Stuttgart, 1984) pp. 97, 115, 156, 164.
(1997) 5020. [34] F.X. Cormerais, G. Pérot, F. Chevalier and M. Guisnet, J. Chem.
[8] N. Tajima, T. Tsuneda, F. Toyama and K. Hirao, J. Am. Chem. Soc. Res. (S) (1980) 362.
120 (1998) 8222. [35] D.K. Murray, J.-W. Chang and J.F. Haw, J. Am. Chem. Soc. 115
[9] J.P. van den Berg, J.P. Wolthuizen and J.H.C. van Hooff, in: Proc. (1993) 4732.
5th International Zeolite Conference, ed. L.V. Rees (Heyden, Lon- [36] J.F. Haw, B.R. Richardson, J.S. Oshiro, N.D. Lazo and J.A. Speed,
don, 1980) p. 649. J. Am. Chem. Soc. 111 (1989) 2052.
[10] G.A. Olah, H. Doggweiler, J.D. Felberg, S. Frohlich, M.J. Grdina, R. [37] M.T. Aronson, R.J. Gorte, W.E. Farneth and D. White, J. Am. Chem.
Karpeles, T. Keumi, S. Inaba, W.M. Ip, K. Lammertsma, G. Salem Soc. 111 (1989) 840.
and T.C. Tabor, J. Am. Chem. Soc. 106 (1984) 2143. [38] G.A. Olah, H. Doggweiler, J.D. Felberg and S. Frohlich, J. Org.
[11] M. Stöcker, Micropor. Mesopor. Mater. 29 (1999) 3. Chem. 50 (1980) 4847.
[12] S. Kolboe, Acta Chem. Scand. A 40 (1986) 711. [39] E.J. Munson, A.A. Kheir and J.F. Haw, J. Phys. Chem. 97 (1993)
[13] I.M. Dahl and S. Kolboe, J. Catal. 149 (1994) 458; 161 (1996) 304. 7321.
[14] M.W. Anderson and J. Klinowski, Nature 339 (1989) 200. [40] V. Bosacek, J. Phys. Chem. 97 (1993) 10732.
[15] M.W. Anderson and J. Klinowski, J. Am. Chem. Soc. 112 (1990) [41] A. Philippou, F. Salehirad, D.P. Luigi and M.W. Anderson, in: Proc.
10. 12th International Zeolite Conference, eds. M.M.J. Treacy, B.K.
[16] M.W. Anderson, B. Sulikowski, P.J. Barrie and J. Klinowski, J. Phys. Marcus, M.E. Bisher and J.B. Higgins (Materials Research Society,
Chem. 94 (1990) 2730. Warrendale, PA, 1999) p. 2465.