Kinetic studies of the dehydration of methanol over aluminosilicate and gallosilicate offretites

Article abstract of DOI:10.1039/a607290i

The kinetics of the dehydration of CH₃OH by the hydrogen forms of three offretites with different contents of framework Al and Ga have been monitored by ¹³C magic-angle-spinning (MAS) NMR spectroscopy. No reaction takes place at room temperature. At 150°C methanol is dehydrated to dimethyl ether. As expected, the catalytic activity of offretite for this reaction is lower than that of zeolite ZSM-5, and it decreases with increasing framework gallium content. The rates of the dehydration reaction are 0.045, 0.863 and 1.136 a.u.^-1 s^-1 for Ga-, Al,Ga- and Al-offretites, respectively.

Full text of DOI:10.1039/a607290i

View Article Online / Journal Homepage / Table of Contents for this issue
Kinetic studies of the dehydration of methanol over aluminosilicate
and gallosilicate o†retites
Mar•a D. Alba,a¤ Antonio A. Romero,a” Mario L. Occellib and Jacek Klinowskia*°
a Department of Chemistry, University of Cambridge, L ensÐeld Road, Cambridge UK CB2 1EW
b Zeolite and Clay Program, Georgia T ech Research Institute, Georgia Institute of T echnology,
Atlanta, GA 30332, USA
The kinetics of the dehydration of CH OH by the hydrogen forms of three o†retites with di†erent contents of framework Al and
3
Ga have been monitored by 13C magic-angle-spinning (MAS) NMR spectroscopy. No reaction takes place at room temperature.
At 150 ¡C methanol is dehydrated to dimethyl ether. As expected, the catalytic activity of o†retite for this reaction is lower than
that of zeolite ZSM-5, and it decreases with increasing framework gallium content. The rates of the dehydration reaction are
0.045, 0.863 and 1.136 a.u.~1 s~1 for Ga-, Al,Ga- and Al-o†retites, respectively.
Microporous molecular sieves, such as zeolites and AlPO -
based materials, are BrÔnsted acids and can selectively accom-
proton, the material exhibits BrÔnsted acidity. Proton forms
of zeolites are prepared by Ðrst replacing the charge-balancing
cations in the as-prepared material (typically sodium) by the
ammonium cation which is then decomposed thermally with
the evolution of ammonia gas. The BrÔnsted site is thus a
“bridgingÏ hydroxy group between SiIV and the substituent
element, with a dative bond involving a pair of unshared elec-
trons of oxygen and an unoccupied orbital of the substituent
element. Gallium-substituted zeolites are known to have lower
BrÔnsted acidity than their aluminosilicate counterparts.
Based on the measured IR l(OH) frequencies, Chu and
Chang26 have shown that the BrÔnsted acidity of the bridging
hydroxyl groups increases in the sequence Si(OH) @
Ga(OH)Si \ Al(OH)Si. On the other hand, o†retite is a less
powerful catalyst than, for example, zeolites Y and ZSM-5.
Nonetheless, it is of considerable interest to catalysis in view
of its very di†erent selectivity in a variety of organic reactions.
We are fully aware that the conditions inside a sealed glass
microreactor are not the same as those in the Ñow system
which is actually used. Furthermore, given the experimental
procedure, the precise pressure inside the microreactor is not
known accurately. Alternative procedures, such as magic-
angle spinning using 13C-enriched spinning gas, are totally
impracticable in view of the enormous cost involved. With all
these reservations, the technique used in this work allows us
to monitor in situ the evolution of the various products and
intermediates in the MTG process, providing information
which is unavailable to other techniques, such as gas chroma-
tography (in view of the di†erential adsorption of MeOH and
DME and their di†erent di†usional properties).
4
modate a variety of small molecules, which makes them
powerful shape-selective catalysts.1 Among the multitude of
chemical reactions catalysed by these materials, one of the
most important industrially is the methanol-to-gasoline
(MTG) process in which methanol (MeOH) is Ðrst dehydrated
to dimethyl ether (DME), which is then converted to gasoline,
a mixture of C ÈC hydrocarbons.2,3 There has been much
5
10
interest in the nature of the intermediate species involved, but
the mechanism of the various consecutive and parallel reac-
tions is still a matter of heated debate. IR spectroscopy,4 gas
chromatography,3,5
mass
spectrometry
and
NMR
spectroscopy6h10 show that MeOH is initially adsorbed at the
framework BrÔnsted acid sites and then dehydrated to DME.
The mixture of MeOH and DME is then converted to
alkenes, aliphatics and aromatics up to C
.
10
The chemical status of the species in the intracrystalline
space in the course of catalytic reactions on molecular sieves
can be studied conveniently using 13C magic-angle-spinning
(MAS) NMR spectroscopy.4,11h17 Reaction rates are of con-
siderable practical importance given the need to be able to
predict how quickly a reaction mixture will reach equilibrium
and to optimise the rate by the appropriate choice of condi-
tions, such as the pressure, temperature and the nature of the
catalyst. We have applied 13C MAS NMR spectroscopy in
situ11,12 to study the kinetics of the Ðrst step of this reaction:
the conversion of MeOH into DME over an o†retite with iso-
morphous substitution of gallium and/or aluminium for
silicon.
The cages of o†retite are connected through unrestricted
twelve-membered ring apertures in such a way as to form
parallel microporous channels ca. 6.3 Ó in diameter which
span the entire length of the crystal in the c direction.18h20 As
a result of this arrangement, o†retite adsorbs organic mol-
ecules of up to 6.0 A in diameter.
Isomorphous substitution of silicon by a trivalent element
such as Ga,21,22 Al23,24 or B25 in tectosilicate frameworks
creates a net negative framework charge which, for electrical
neutrality, must be balanced by a cation. When the cation is a
Experimental
Preparation of samples
Molecular sieves with the o†retite structure were synthesized
from hydrogels of composition 1.0 M O : 12 SiO : 1.5
2
2
3
2
Na O : 2.2 K O : 4.0 (TMA) O : 250 H O (M \ Al, Ga;
2
2
2
TMA \ tetramethylammonium from tetramethylammonium
chloride). The required Al/Ga ratio of the gel was obtained by
adding a clear aqueous solution of sodium gallate. Crys-
tallization (without stirring) was performed in PTFE-lined
Bergho† autoclaves heated at 95 ¡C for 48 h. The product was
washed with distilled water, dried at 100 ¡C in an air oven,
activated by removing the organic templete by calcining at
550 ¡C for 10 h in air and stored under ambient conditions.
¤ On leave from: Departamento de Qu•mica Inorganica, Instituto
Ciencia de los Materiales, Universidad de Sevilla, C.S.I.C., P.O. Box
874, 41012 Sevilla, Spain.
” On leave from: Departamento de Qu•mica Organica, Universidad
de Cordoba, San Alverto Magno s/n, 14004 Cordoba, Spain.
° Email: jk18=cam.ac.uk
J. Chem. Soc., Faraday T rans., 1997, 93(6), 1221È1224
1221

View Article Online
Powder X-ray di†raction (XRD) indicates that all samples are
highly crystalline. Ammonium-exchanged forms of o†retite
were prepared by repeated treatment of the crystals with a 2 M
aqueous solution of NH NO at 60 ¡C. Finally, the hydrogen
4
3
forms were made by heating the ammonium-exchanged forms
at 500 ¡C for 4 h in Ñowing air.18 XRD, surface area measure-
ments and 29Si, 27Al and 71Ga MAS NMR spectroscopies
(results not shown) indicate a decrease of sample crystallinity
as a result of the calcination procedure. Some six-coordinate
(extra-framework) Al and Ga are generated in the process and
the removal of extra-framework Ga is easier than the removal
of Al. Microcalorimetry with ammonia as a probe molecule
indicates that the insertion of gallium into the aluminosilicate
framework results in a heterogeneity of the acid sites:
BrÔnsted acidity is decreased and Lewis acidity increased. The
bulk compositions of the three samples of o†retite examined
in this work are given in Table 1.
For MAS NMR measurements, a small quantity (ca. 200
mg) of each catalyst was placed in a Pyrex capsule designed to
Ðt inside a zirconia MAS rotor27 and the capsule was con-
nected to a vacuum line. The catalyst was activated by heating
at 400 ¡C for 6 h under vacuum, charged with 50 Torr of 30
wt.% 13C-enriched CH OH at 20 ¡C and allowed to equili-
3
brate for 1 h before being isolated from the bulk of the
gaseous MeOH. The capsule was sealed o† under liquid nitro-
gen to prevent the onset of catalytic processes. The capsule
was then heated to the desired reaction temperature, for
various lengths of time, quenched and placed in the NMR
rotor to record 13C NMR spectra at ambient temperature.
Fig. 1 13C MAS NMR spectra (ppm from TMS) (1H decoupling
only) of (a) Ga-o†retite, (b) Al,Ga-o†retite and (c) Al-o†retite with 50
Torr of adsorbed CH OH and no thermal treatment
3
NMR spectra
After heating the sample of Ga-o†retite at 150 ¡C for 40 min
the spectrum consists of two lines, at 50.8 and 60.5 ppm [Fig.
2(b)], corresponding to MeOH and DME, respectively. The
DME/MeOH concentration ratio is 0.17 and increases when
thermal treatment of the sample continues at the same tem-
All spectra were recorded on a Chemagnetics CMX-400
spectrometer with zirconia rotors 7.5 mm in diameter, spun at
the magic angle at rates of up to 2 kHz in nitrogen gas. 13C
spectra were recorded using high-power decoupling without
cross-polarization in order to make the spectra quantitatively
reliable. 13C MAS spectra without proton decoupling and
cross-polarization (CP) were also measured. 1HÈ13C CP MAS
NMR spectra were measured only in order to establish
whether protons are attached to a given carbon atom or to
discriminate between mobile and immobile species. High-
power proton decoupling experiments were carried out with
45¡ 13C pulses with a repetition time of 10 s and 300 scans.
The longest 13C T relaxation time is ca. 1.5 s, so that a 10 s
1
recycle time is sufficiently long to provide quantitatively reli-
able spectra. The decoupler was gated o† between scans in
order to prevent nuclear Overhauser enhancement. Experi-
ments without high-power decoupling used the same acquisi-
tion parameters. Cross-polarization spectra were acquired
with a 7.3 ls proton 90¡ pulse, a 4 ms contact time and a 4 s
repetition time. Chemical shifts are referred to external tetra-
methylsilane (TMS). Relative concentrations of MeOH and
DME were determined by integration of the two well resolved
NMR resonances.
Results
The 13C MAS NMR spectra of MeOH adsorbed on the three
catalysts at room temperature (Fig. 1) and not thermally
treated each consist of a single resonance at 50.8 ppm from
the methyl groups in adsorbed MeOH.11,12
Fig. 2 13C MAS NMR spectra (ppm from TMS) (1H decoupling
only) of Ga-o†retite heated at 150 ¡C for (a) 10, (b) 40, (c) 100 min, (d)
3, (e) 4, (f) 5, (g) 7, (h) 10, (i) 12 and (j) 18 h
Table 1 Oxide composition (mole ratios) of samples of o†retite after ion exchange and calcination
Al O Ga O SiO Na O
sample
K O
2
3
2
3
2
2
2
Ga-o†retite
Al,Ga-o†retite
Al-o†retite
È
0.85
1.00
1.00
0.15
È
7.65
7.71
7.79
0.030
0.001
0.033
È
0.09
0.17
1222
J. Chem. Soc., Faraday T rans., 1997, V ol. 93

View Article Online
in agreement with the mechanism proposed by Anderson and
Klinowski12 who suggested that the Ðrst step of the reaction
involves protonation of MeOH
CH OH ] HZ ] (CH OH )` ] Z~
3
3
2
where Z stands for the zeolitic framework and the methoxon-
ium ion is a transition structure.31 The reaction then con-
tinues:
(CH OH )` ] CH OH ] CH OCH ] H O`
3
2
3
3
3
3
Discussion
As has been shown previously,11,12 at 150 ¡C only DME is
produced from MeOH when zeolite ZSM-5 is the catalyst.
However, while with ZSM-5 the equilibrium between MeOH
and DME is approached after heating the sample at 150 ¡C
for 5 min, a much longer period is required when the reaction
is carried out over o†retite. The di†erence can be explained in
terms of acid strength and shape selectivity as follows.
Zeolite ZSM-5, with a ten-membered ring pore opening, has
an e†ective pore diameter of ca. 5.5 Ó, whereas the o†retite
frameworks cages are connected through twelve-membered
oxygen rings linked by eight-membered ring apertures leaving
parallel channels of ca. 6.3 Ó in diameter. Shah et al.32 demon-
strated that when methanol is situated in the more open cage
regions of zeolites, it appears to be unprotonated and less
chemically active. It follows that medium-sized pores should
be the most active for the MTG reaction because this is where
methanol appears to be protonated. Di†erences in the
BrÔnsted acidity of ZSM-5 and o†retite have been assessed by
measuring the 1H NMR chemical shifts of the MeOH
hydroxy groups. Anderson et al.19 found a lower acidity of
o†retite compared to that of ZSM-5.
The percentage of DME formed during the MeOH dehy-
dration, calculated from the intensity (peak area) of the 13C
MAS NMR resonances vs. the time of reaction t, is plotted in
Fig. 4(a). As far as we are aware, this is the Ðrst time that the
kinetics of MeOH dehydration have been monitored in situ. It
is clear that the kinetics over the three o†retite catalysts of
di†erent composition are quite di†erent.
Fig. 3 13C MAS NMR spectra (ppm from TMS) of Ga-o†retite with
50 Torr of adsorbed CH OH: (a) heated at 150 ¡C for 100 min and
3
recorded with proton decoupling only; (b) MAS without proton
decoupling; and (c) 1HÈ13C CP MAS NMR spectrum
perature. Equilibrium is reached after heating for 18 h [Fig.
2(j)], when ca. 45% of MeOH has been converted.
With Al,Ga-o†retite, dehydration proceeds faster. The
DME resonance appears after only 10 min treatment at
150 ¡C (not shown), and the conversion of MeOH to DME
reaches ca. 80% after 5 h. With Al-o†retite the lines corre-
sponding to DME and MeOH are of comparable intensity
after heating at 150 ¡C for just 20 min, and the reaction is
complete within 100 min.
The 13C MAS NMR spectra with proton decoupling of
Al,Ga-o†retite [Fig. 3(a)] and Al-o†retite (not shown) treated
at 150 ¡C for 40 min reveal a new line at ca. 57 ppm which we
assign to the DME vapour. We conÐrmed this assignment as
follows. First, we note that the species in question gives rise to
a single resonance with a small upÐeld shift in comparison
with that from adsorbed DME. Second, the 13C MAS spec-
trum without proton decoupling [Fig. 3(b)] shows that the 57
Chang and Silvestri3 found that, when used as the feed gas,
DME is partly converted to MeOH, even in the absence of
water. However, below 200 ¡C the conversion of DME to
MeOH via this route is insigniÐcant,10 and the reaction can be
written simply as:
ppm line is split into a quadruplet withJ
B 140 Hz, charac-
k
ChH
teristic of DME.28 The fact that a spectrum without decoup-
ling can be obtained at all indicates that the species in
question is highly mobile. Third, in the spectrum with cross-
polarization [Fig. 3(c)] the intensity of the line is considerably
lower than the intensity of the line from adsorbed DME,
which almost disappears.
We have also found that the MeOH peak moves downÐeld
and its linewidth increases as the concentration of MeOH
decreases. We attribute the increased linewidth to a restriction
in motion of the adsorbed species, which increases the dipoleÈ
dipole and dipoleÈquadrupole interactions, as well as the
chemical shift anisotropy.29 The downÐeld shift of the MeOH
line is caused by the exchange between adsorbed methoxon-
ium ions and adsorbed methanol29,30 according to the scheme
2CH OH ÈÈ CH OCH ] H O
3
3
3
2
The reaction rate is proportional to the concentration of
methanol raised to a power equal to the order of the reaction
with respect to that species33
1 d[CH OH] d[CH OCH ]
3
3
3
l \ [
\
\ k[CH OH]2
3
2
dt
dt
where the coefficient k is the rate constant of the reaction, and
is independent of the concentration but dependent on the tem-
perature.
The concentration of MeOH has been calculated by
assuming that the areas of the individual deconvoluted NMR
lines are directly proportional to the populations of the
respective chemical species. Fig. 4(b) shows the plot of
1/[MeOH] vs. the duration of the reaction from which the
order of the reaction can be obtained. As expected for a
second-order reaction, the plot is a straight line in agreement
with the integrated form of the second-order rate law:
1
1
\
] kt
[MeOH]
[MeOH]
t
0
J. Chem. Soc., Faraday T rans., 1997, V ol. 93
1223

View Article Online
Fig. 4 (a) Percentage of MeOH dehydrated at 150 ¡C over the three samples of o†retite. (b) Plot of 1/[MeOH] vs. reaction time.
t
Table 2 Kinetics of the conversion of MeOH over the three o†retites at 150 ¡C
Al
Ga
Al ] Ga ] Si
Al ] Ga ] Si
k/a.u.~1 h~1
Ga-o†retite
Al,Ga-o†retite
Al-o†retite
È
0.175
0.204
0.207
0.031
È
0.045
0.863
1.136
a.u. \ atomic unit.
13 M. W. Anderson, B. Sulikowski, P. J. Barrie and J. Klinowski, J.
Phys. Chem., 1990, 94, 2730.
where [MeOH] and [MeOH] are the concentration of
0
t
CH OH at t \ 0 and at time t.
3
14 M. W. Anderson and J. Klinowski, J. Chem. Soc., Chem.
Commun., 1990, 918.
The rate constant can be calculated from the slope of the
plots. Table 2 gives k for the conversion of MeOH over the
three catalysts and the regression coefficients of the corre-
sponding linear equation to which the experimental points
have been Ðtted. The rate constant towards DME increases
with increasing aluminium content: 0.045, 0.863 and 1.136
a.u.~1 s~1 for Ga-, Al,Ga- and Al-o†retites, respectively. We
conclude that the rates of formation of DME under di†erent
experimental conditions can be determined directly from the
NMR spectra and rationalized in terms of acidity and shape
selectivity of the individual catalysts.
15 M. W. Anderson and J. Klinowski, Chem. Phys. L ett., 1990, 172,
275.
16 J. Klinowski and M. W. Anderson, Magn. Reson. Chem., 1990, 28,
68.
17 M. W. Anderson, P. J. Barrie and J. Klinowski, J. Phys. Chem.,
1991, 95, 235.
18 M. L. Occelli, R. A. Innes, S. S. Pollack and J. V. Sanders, Zeo-
lites, 1987, 7, 265.
19 M. W. Anderson, M. L. Occelli and J. Klinowski, J. Phys. Chem.,
1992, 96, 388.
20 C. L. Cavalcante, M. Eic, D. M. Ruthven and M. L. Occelli, Zeo-
lites, 1995, 15, 293.
21 C.-F. Cheng, and J. Klinowski, J. Chem. Soc., Faraday T rans.,
1996, 92, 289.
We are grateful to the European Community for grants to
M.D.A. and A.A.R. and to Dr. H. He for experimental
assistance.
22 C.-F. Cheng, M. D. Alba and J. Klinowski, Chem. Phys. L ett.,
1996, 250, 328.
23 Z. Luan, C.-F. Cheng, W. Zhou and J. Klinowski, J. Phys. Chem.,
1995, 99, 1018.
References
24 R. Mokaya, W. Jones, Z. Luan, M. D. Alba and J. Klinowski,
Catal. L ett., 1996, 37, 113.
1
2
P. B. Weisz and V. J. Frilette, J. Phys. Chem., 1960, 64, 382.
S. L. Meisel, J. P. McCullogh, C. H. Lechlhaler and P. B. Weisz,
ChemT ech, 1976, 6, 86.
C. D. Chang and A. J. Silvestri, J. Catal., 1977, 47, 249.
G. Mirth, J. A. Lercher, M. W. Anderson and J. Klinowski, J.
Chem. Soc., Faraday T rans., 1990, 86, 3039.
25 S. X. Liu, H. Y. He, Z. H. Luan and J. Klinowski, J. Chem. Soc.,
Faraday T rans., 1996, 92, 2011.
3
4
26 C. T. W. Chu and C. D. Chang, J. Phys. Chem., 1985, 89, 1569.
27 T. A. Carpenter, J. Klinowski, D. T. B. Tennakoon, C. J. Smith
and D. C. Edwards, J. Magn. Reson., 1986, 68, 561.
28 E. Pretsh, W. Simon, J. Seibl and T. Clerc, T able of Spectral Data
for Structure Determination of Organic Compounds, Springer-
Verlag, Berlin, 1989, C-220.
5
6
R. M. Dessau and R. B. La Pierre, J. Catal., 1982, 78, 136.
J. F. Haw, B. R. Richardson, I. S. Oshiro, N. D. Lazo and J. A.
Speed, J. Am. Chem. Soc., 1989, 111, 2052.
7
8
9
J. B. Nagy, J. P. Gilson and E. G. Derouane, J. Mol. Catal., 1979,
5, 393.
29 A. ThursÐeld and M. W. Anderson, J. Phys. Chem., 1996, 100,
6698.
E. G. Derouane, P. DejaÐve and J. B. Nagy, J. Mol. Catal., 1977,
3, 453.
30 G. A. Olah and A. M. White, J. Am. Chem. Soc., 1969, 91, 5801.
31 F. Haase and J. Sauer, J. Am. Chem. Soc., 1995, 117, 3780.
32 R. Shah, M. C. Payne, M. H. Lee and J. D. Gale, Science, 1996,
271, 1395.
33 P. W. Atkins, Physical Chemistry. Oxford University Press, 1994,
p. 861.
E. G. Derouane, J. P. Gilson and J. B. Nagy, Zeolites, 1982, 2, 42.
10 C. E. Bronnimann and G. E. Maciel, J. Am. Chem. Soc., 1986,
108, 7154.
11 M. W. Anderson and J. Klinowski, Nature (L ondon), 1989, 339,
200.
12 M. W. Anderson and J. Klinowski, J. Am. Chem. Soc., 1990, 112,
10.
Paper 6/07290I; Received 25th October, 1996
1224
J. Chem. Soc., Faraday T rans., 1997, V ol. 93