Comparative vapor phase synthesis of ETBE from ethanol and isobutene over different acid zeolites

Article abstract of DOI:10.1006/jcat.2001.3280

Vapor phase synthesis of ETBE from ethanol and isobutene was studied over US-Y, Beta, and ZSM-5 zeolites with different Si/Al ratios, using Amberlyst-15 as a reference catalyst. The sequence of activity was Beta zeolite > US-Y> Mordenite > Omega ≥ ZSM-5. At maximum isobutene to ETBE conversion, the Beta zeolites yielded more ETBE than the commercial samples and acid resin. With external surface area of > 200 sq m/g, Beta zeolites were the most active among the zeolites. Amberlyst-15 and Beta zeolites were 100% selective below 55°C. Extra-framework Al species showed negative effect on the reaction, and their removal by a mild acid leaching was beneficial.

Full text of DOI:10.1006/jcat.2001.3280

Journal of Catalysis 202, 68–77 (2001)
doi:10.1006/jcat.2001.3280, available online at http://www.idealibrary.com on
Comparative Vapor Phase Synthesis of ETBE from Ethanol
and Isobutene over Different Acid Zeolites
F. Collignon and G. Poncelet2
1
Universit e´ Catholique de Louvain, Unit e´ de Catalyse et Chimie des Mat e´ riaux Divis e´ s, Place Croix du Sud 2/17, B-1348 Louvain-la-Neuve, Belgium
Received November 13, 2000; revised May 14, 2001; accepted May 20, 2001
and well-performing catalysts. Because ether synthesis is
Vapor phase synthesis of ETBE from ethanol and isobutene exothermic, temperature needs to be carefully controlled to
has been investigated over US-Y, Beta, and ZSM-5 zeolites with avoid the release of temperature-sensitive sulfonic groups
different Si/Al ratios, using Amberlyst-15 as a reference catalyst.
The zeolites were characterized by X-ray diffraction and nitrogen
adsorption–desorption isotherms at 77 K. Total acidity of the Beta
zeolites was determined by ammonia-TPD measurements whereas
IR spectroscopy was used to investigate the OH stretching region
formation in liquid phase conditions have been investigated
and evaluate the Brønsted and Lewis acid contents by pyridine ad-
sorption. Beta zeolites, with higher external surface area, were more
active than the other zeolites. The highest yields of ETBE were ob-
tained for Beta samples with high Brønsted acid content and low
from the matrix, which causes a loss of activity and corro-
sion problems (5). Since the process is operated with al-
cohol/isobutene molar ratios is excess of stoichiometry, re-
cycling is necessary. The kinetics and mechanism of ETBE
by several authors. Fran c¸ oisse and Thyrion (6) observed
that ethanol was preferentially adsorbed inside the poros-
ity of Amberlyst-15 due to its polarity, and they proposed
Eley–Rideal and Langmuir–Hinshelwood reaction mech-
SiOH/AlOHSi ratio. �c 2001 Academic Press
Key Words: ETBE; vapor phase synthesis; zeolites; H-Beta; US-Y; anisms depending on ethanol concentration. Heat capac-
H-ZSM-5; external surface area; Brønsted acidity; silanol groups.
ity and reaction enthalpy measurements obtained by Sola
et al. (7) with reaction calorimetry techniques confirmed
the E–R model established from kinetic studies (8). Cunill
et al. (9) observed that the production rates of ETBE and
MTBE were strongly lowered by small amounts of water,
particularly at the beginning of the reaction. This inhibitor
effect disappeared with progressive formation of tert-butyl
alcohol.
The synthesis of ETBE over nonresin catalysts is much
less documented relative to MTBE. Acid zeolites are sta-
ble in the temperature domain where the ethers are formed.
They could be attractive catalysts provided that their effi-
ciency at the low temperatures imposed by the thermody-
namics of the reaction (10) is high enough. Results of Tau
and Davis (11) showed that the isobutene to ETBE con-
version over H-ZSM-5 was relatively low compared with
Amberlyst-15. The protective effect of ethanol clusters on
the acid sites of H-mordenite would be responsible for the
inhibiting effect of ethanol; it would also prevent isobutene
oligomerization and ensure the zeolite stability (12). A re-
action pathway involving the attack of the C == C bond of
INTRODUCTION
Methyl tertiary-butyl ether (MTBE) and ethyl tertiary-
butyl ether (ETBE) are octane enhancers of unleaded gaso-
lines. Both ethers easily mix with gasoline (10–15 vol.% ),
have excellent antiknocking behavior, and reduce emis-
sion of carbon monoxide and unburned hydrocarbons (1).
MTBE production largely overwhelms that of ETBE be-
cause of the greater availability and lower production cost
of methanol relative to ethanol (2). ETBE has the advan-
tage over MTBE of a lower vapor pressure, and ethanol
is a renewable energy source (bio-ethanol). Recently, a
novel ether, the 3-methoxy-3-methylpentane, obtained by
the etherification of 2-ethyl-1-hexene with methanol, has
been proposed as a possible candidate (3).
The synthesis of these oxygenates is commercially oper-
ated in liquid phase over sulfonic acid resins, at tempera-
�
tures in the 50–70 C range, pressures between 10 and 15 bar,
and alcohol/isobutene molar ratios in excess of stoichiom-
etry (4). Sulfonic acid resins are relatively inexpensive
isobutene by a single hydrogen-bonded ethanol molecule
was proposed. Vapor phase synthesis of ETBE over gal-
lophosphate cloverite gave a maximum yield of ETBE of
1
Present address: Katholieke Universiteit Leuven, Centrum voor Op-
�
2
0% at 102 C and a selectivity of 83–84% (13, 14).
pervlaktechemie en Katalyse, Kasteelpark Arenberg 23, B-3001 Leuven,
Belgium.
A previous study on the vapor phase synthesis of MTBE
over different zeolites showed that H-Beta was as ac-
tive as Amberlyst-15 and superior to US-Y, H-mordenites,
2
To whom correspondence should be addressed. E-mail: poncelet
@
cata.ucl.ac.be.
68
0021-9517/01 $35.00
Copyright �c 2001 by Academic Press
All rights of reproduction in any form reserved.

VAPOR PHASE SYNTHESIS OF ETBE
69
H-ZSM-5, and other zeolites (15), a result confirmed both
TABLE 1
Si/Al Ratio, Relative Crystallinity (%), External Surface Area
in vapor (16–19) and in liquid phase conditions (20).
This study reports on comparative screening experiments
of the vapor phase synthesis of ETBE over different zeo-
2
� 1 � 1
3
(
g
Sext, m g ), Micropore Volume (V� , cm g ), Al Content (mmol
� 1
� 1
), NH3-TPD (mmol g ), and Enthalpy of Ammonia Desorption
�
1
lites and Amberlyst-15. Emphasis has been given to H-Beta (1H, kJ mol
zeolites because of their higher catalytic performances. In
particular, their activation mode and the role of the acid
and silanol contents have been addressed.
)
Rel.
Si/Al Cryst.
Sext
V�
Altot.a NH3-TPD 1Hb
Zeolite Beta
Bsc
B75
B25
B25
B25 0.02
24.6
36.0
12.2 100
13.4
17.4
19.1
22.0
23.6
31.4
9.2
97
96
182
261
282
241
n.d.
255
252
n.d.
263
188
179
278
263
167
225
233
n.d.
n.d.
0.18
0.18
0.19
0.16
n.d.
0.18
0.16
n.d.
0.18
0.19
0.13
0.21
0.18
0.16
0.13
0.19
n.d.
n.d.
0.65
0.45
1.26
1.16
0.89
0.87
0.74
0.73
0.56
1.63
1.63
1.07
1.07
0.88
0.88
1.30
1.30
1.11
0.45
0.41
0.89
0.46
0.55
0.52
0.54
0.50
0.42
1.12
1.43
0.95
1.02
0.81
0.77
1.05
1.02
1.24
123
123
121
119
119
119
121
120
120
122
117
123
122
121
120
119
118
118
EXPERIMENTAL
Catalysts
The different zeolites considered in this study are listed
�
�
�
�
�
98
92
92
89
96
89
88
83
c
c
in Table 1. The US-Y zeolites (series CBV 500 to CBV
B25 0.03
�
7
80), Beta zeolites (B75, B25, B25 , with the asterisk indi- B25 0.05
cating a different batch, and B25t, with t standing for a sam- B25 0.06
�
B25 0.09
ple still containing its template tetraethyl ammonium), and
lB18t c
ZSM-5 (CBV3020 to CBV8020, and CBV2802, a silicalite-
lB18t cec
9.2
14.6 100
type) were commercial samples kindly supplied by PQ Ze-
lB28t c
olite. Sample B25 is the same batch used in previous studies lB28t cec
15, 20). Bsc was a Beta zeolite provided by S u¨ d-Chemie. lB36t c
14.6
17.9
17.9
11.8
11.8
14.0
82
93
93
93
78
92
(
lB36t cec
B25t c
B25t cec
B25t 0.1 c
Mordenite was a large-port variety (HSZ640 from Tosho
Corp., Japan) that was steam-dealuminated at 600 C and
acid leached to remove extra framework Al species (21).
This sample was among the most active mordenites in a
previous study (15). LZ-6 (Omega zeolite) was a sample
�
US-Y
CBV500
2.6 100
2.8 91
5.8 121
30.0
37.0
85
96
116
119
136
0.26
0.22
0.25
0.26
0.26
4.63
4.38
2.47
0.54
0.44
1.51
0.83
0.69
0.27
0.17
116
115
118
127
127
from Union Carbide. The series of US-Y, Beta, and H- CBV600
CBV712
CBV760
CBV780
ZSM-5 were used as received, except for three samples in
the ammonium form (CBV500, CBV 712, and CBV8020),
which were previously decationated under the conditions
40
89
ZSM-5
CBV3020
CBV5020
indicated below.
16.5
25.0
38.3
137.5 100
117.4 93
4.5 100
79
85
81
63
35
56
0.16
0.18
0.17
n.d.
0.95
0.64
0.42
0.12
0.93
0.62
0.47
n.d.
128
130
129
n.d.
�
Samples of B25 were dealuminated using a method sim-
ilar to that of Fajula et al. (22). Zeolite (1 g) was added to CBV8020
00 ml of HNO3 solution previously heated at 80 C, and the CBV2802
�
2
n.d.
mixture was stirred for 3 h at this temperature. Acid concen- Mordenite
�
M-S600H6
87
0.19
n.d.
0.14
3.03
0.13
n.d.
n.d.
n.d.
trations were in the range 0.02–0.09 M (samples B25 0.02
�
�
to B25 0.09 in Table 1). The solids were washed and dried Omega
LZ-6
n.d.
at 100 C for 1 h.
Beta zeolites with Si/Al ratios of 9.2, 14.6, and 17.9 (sam-
ples lB-t for lab-synthesized Beta in Table 1) were synthe-
sized in the presence of tetraethyl ammonium hydroxide
a
From ICPS;
Calculated according to (26, 27).
n.d.: not determined.
b
c
(
(
TEAOH) as a template, following a reported procedure
23). Briefly, TEAOH (35 wt% aqueous solution, Aldrich) Samples were either calcined to eliminate the template
was mixed in deionised water with sodium aluminate (lBt c), or calcined, subsequently ammonium-exchanged,
37 wt% Na2O, 56 wt% Al2O3, Carlo Erba) in a teflon and calcined again to obtain the protonic forms (lBt cec).
cylinder and stirred for 10 min. Silica (FK300DS, Degussa) Elimination of the template from B25t and the lab-
(
was progressively added to the solution. The SiO2/TEAOH synthesized Beta zeolites (lB-t) was achieved in leached-
and H2O/SiO2 molar ratios were 6 and 6.1, respectively. bed conditions by temperature-programmed calcination
�
The Si/Al molar ratio of the “dry” mixture was changed by between room temperature and 120 C, with a plateau for
�
adjusting the amount of sodium aluminate. The powder was 1 h, and from 120 to 540 C in flowing nitrogen (250 ml
mixed at 60–70 rpm for 30–40 min until a thick slurry was min ) with a ramp of 5 C min . At 540 C, nitrogen was
obtained. Crystallization was carried out in a nonstirred switched with oxygen, and the treatment was maintained
�
1
�
� 1
�
Parr pressure vessel (using Teflon containers) heated for for an additional 2.5 h (samples marked “t c” in tables).
�
4
8 h at 170 C. The solid was washed by filtration with hot
A few Beta zeolites (cec in Table 1) were calcined to re-
�
+
demineralized water (25 ml per g of solid) and dried at 60 C. move the template and then converted to the NH -form
4

7
0
COLLIGNON AND PONCELET
by adding 1 g zeolite to 200 ml of a 2 M ammonium acetate Catalytic Measurements
�
solution heated at 80 C and stirred overnight. After fil-
The reaction was carried out in vapor phase in a fixed
�
tration, washing, and drying at 60 C for 1 h, the sam-
ples were calcined to obtain the H -form. Calcination was
done under flowing air using the same ramp as above, ex-
cept that the final temperature (500 C) was maintained for
0–15 h. The ammonia evolved during the whole treat-
ment was collected and quantified as for the NH3-TPD
measurements.
Amberlyst-15 (a sulfonic acid resin from Aldrich), with
surface area of 50 m g and cation exchange capacity of
bed U-shaped glass microreactor operated at atmospheric
pressure, with ethanol/isobutene (EtOH/IB) molar ratio of
+
�
1
1
and WHSV of 2 h . Ethanol vapor was generated by
�
�
4
� 1
flowing a stream of helium (7.9 � 10 mol min ) through
1
�
a glass saturator thermostated at 28 C. The flows were mon-
itored with mass-flow controllers. The reaction was stu-
�
died in the temperature range 36–100 C at a heating rate of
� � 1
0
.3 C min , with a thermocouple placed near the catalyst
2
� 1
bed allowing continuous control of temperature. Tubings,
valves, and connections in front of the reactor were heated
�
1
4
.9 mequiv. g , was used as a reference catalyst. It was
previously washed following a procedure reported in the
literature (24).
�
at 60 C, and those between the reactor outlet and the gas
�
chromatograph were heated to 120 C. Prior to the catalytic
evaluation, activation of the zeolites (typically 200 mg) was
Characterization
�
done in situ at 400 C for 3 h in flowing helium. The acid
�
The relative crystallinity of the dealuminated samples resin was treated in situ at 90 C for 3 h in flowing helium.
was evaluated from the X-ray diffraction patterns recorded Online gas phase analysis was done in a gas chromatograph
with a Siemens D-5000 apparatus (with monochromatized (5890 from Hewlett-Packard) equipped with TC detector
Cu-K� radiation) relating, for the Beta zeolites, the area of and 30-m capillary column (cross-linked methyl silicone
�
the [302] reflection (at 22.43 2�) to that of sample B25. For from Hewlett-Packard). Sampling of the gaseous effluent
the ZSM-5 and the US-Y zeolites, the sum of the areas was monitored every 16.67 min by means of an automated
of the main reflections was reported to that of samples six-way sampling valve. The conversions, yields, and selec-
CBV2802 and CBV500, respectively. The bulk Si/Al ratios tivities were calculated from the number of moles of reacted
were established from the Al contents (Altot in Table 1) ob- and nonreacted isobutene.
tained byinductivelycoupled plasma spectroscopy. The tex-
tural properties were obtained from nitrogen adsorption–
desorption isotherms recorded at 77 K with an ASAP 2000
sorptometer (Micromeritics), using the t-plot method for
the determination of the external surface areas (Sext) and
RESULTS
Characterization
The bulk Si/Al ratios, relative crystallinities, and textu-
micropore volumes (V� ) (for t values in the range 0.35– ral characteristics (Sext and V� ) of the different zeolites are
.5 nm) (25). compiled in Table 1. Except for CBV760, the relative crys-
Temperature-programmed desorption of ammonia tallinities were satisfactory with, for all the Beta zeolites,
NH3-TPD) was carried out adopting the experimental values above 80% .
conditions of Niwa et al. (26) and Katada et al. (27). The As expected, the US-Y zeolites had greater micropore
0
(
release of ammonia in the gaseous effluent was followed volumes than the H-Beta zeolites and H-mordenite, with
with a thermal conductivity detector, with the ammonia the lowest values for the H-ZSM-5 zeolites. Dealuminated
being continuously collected at the outlet of the TCD in US-Y zeolites exhibited increased external surface area.
a 0.2 M solution of H3BO3. Back titration with 0.005 M All the Beta zeolites had higher external surface areas than
H2SO4 provided the total amount of ammonia evolved the other zeolites, owing to their very small crystallite size
from the zeolite.
(20 nm) (28).
FTIR spectroscopy was used to examine the OH stretch-
The acid contents obtained from the ammonia-TPD mea-
�
1
ing region of the Beta zeolites before and after adsorption surements, the total Al content (Altot, in mmol g ), and
2
of pyridine. Self-supporting wafers (1.54 cm ) obtained by the desorption enthalpies providing an estimation of the av-
�
compression were outgassed in the IR cell at 400 C for 2 h erage strength of the acid sites calculated according to Refs.
�
5
(
10 torr). After cooling at room temperature and record- (26, 27) are given in Table 1. ZMS-5 and the most dealu-
ing the spectrum, pyridine was admitted in the cell and minated US-Y samples exhibited slightly higher desorption
equilibrated for 15 min. The cell was then outgassed un- enthalpies.
der vacuum at increasing temperatures, with intermediate
Several commercial samples and the lab-synthesized
spectra being recorded after cooling at ambience. The areas Beta zeolites had NH3-TPD values close to the total Al con-
of the different bands were normalized to the weight of the tent, suggesting that each Al generated an acid site. B25 had
�
wafers (10–30 mg). It was verified that the integrated area of more accessible acid sites (NH3-TPD) than B25 , in spite
�
1
the Si–O band at 1870 cm was proportional to the weight of a same origin and similar Si/Al ratio (probably diffe-
�
of the wafer.
rent syntheses). Leaching B25 with 0.02 to 0.09 M HNO3

VAPOR PHASE SYNTHESIS OF ETBE
71
TABLE 2
�
1
Normalized Surface Area of Different OH Bands (u.a. g ), and Relative Amounts of Brønsted and
�
1
�
Lewis Acidities (in mmol � g ) after Pyridine Adsorption and Outgasing at 150 C
SiOH
AlOHSi
Pyridine-
Brønsted
1545 cm
Pyridine-
Lewis
1454 cm
VHF
3780 cm
3738–3745
AlOH
� 3660 cm
3607–3615
�
1
� 1
� 1
� 1
� 1a
� 1a
�
cm
cm
Bsc
B75
B25
B25
n.b.b
n.b.
2
13
28
31
33
8
1132
1466
1045
1552
1181
1807
1830
1133c
n.b.
n.b.
6
47
2
144
141
199
46
57
97
0.29
0.18
0.34
0.29
0.39
0.25
0.33
0.33
0.52
0.63
0.28
0.21
0.58
0.22
0.07
0.28
0.28
0.30
0.17
0.19
0.19
0.23
0.27
0.15
0.08
0.37
�
�
B25 0.03
�
B25 0.05
1
�
B25 0.06
n.b.
n.b.
0.2
n.b.
n.b.
n.b.
n.b.
112
79
�
B25 0.09
c
lB18t c
7
813
607
801
440
267
690
lB18t cec
lB28t cec
lB36t cec
B25t c
n.b.
n.b.
n.b.
7
953c
1046c
c
756
1350c
a
Calculated with εB = 1.67 cm �mol� 1 and εL = 2.22 cm �mol� 1 (42).
b
c
n.b.: no band.
� 1
Majority of SiOH groups in external position (� � 3743 cm ).
increased the Si/Al ratio from 12.5 to 31.4. For these sam-
The ZSM-5 zeolites had similar Altot and NH3-TPD val-
�
1
ples, the normalized area of the band near 3610 cm (bridg- ues, thus ruling out significant amounts of EFAL species.
ing AlOHSi groups) increased with increasing acid concen- They also had acid sites with higher average strength com-
trations (except for the sample treated with 0.09 M HNO3) pared with the other zeolites, as inferred from the higher
(
Table 2), indicating that extra-framework Al (EFAL) desorption enthalpies.
�
species neutralizing acid sites or blocking their access (28) Typical IR spectra of commercial (B25 ) and lab-
were removed by a mild acid leaching. Sample B75 had very synthesized Beta zeolites (lB36t cec) are shown in Fig. 1.
few EFAL as inferred from the NH3-TPD /A ltot molar ratio Five bands appeared in the OH stretching region. The
�
1
close to 1 (Table 1) and the low content of Lewis acid sites two intense bands at 3745 and 3738 cm are character-
Table 2). istic of silanol groups at external and internal (hydroxyl
B25t c (supplied with its template and calcined) exhib- nest) positions, respectively (30–36). The high relative in-
(
ited a greater amount of acid sites compared with B25 and tensity of these bands in Beta zeolite has been accounted
�
B25 (both with similar Al contents). Interestingly, B25t for by the small crystallite size and the existence of struc-
0
.1-c (treated with 0.1 M HNO3 and calcined to remove the tural defects (coexistence of three unit cells) (37, 38). The
template) desorbed the highest amount of ammonia among
the Beta zeolites. The samples with Si/Al ratio higher than
1
1 (B25t, lB28t, and lB36t) calcined without (-c) and with an
intermediate ammonium exchange (-cec) had similar acid
contents, thus indicating that an intermediate ammonium
exchange is not indispensable.
The US-Y zeolites exhibited large differences between
the Altot and the NH3-TPD values, particularly CBV500,
CBV600, and CBV712. 27Al MAS-NMR spectra of those
zeolites showed signals at 0 and 30 ppm typical of EFAL
species (29). CBV760 and CBV780 had very few EFAL
species. The number of acid sites measured by TPD agreed
27
well with the tetrahedral Al determined by Al MAS-
�
1
NMR (1.64, 0.88, 0.75, 0.22, and 0.13 mmol g , for CBV500,
CBV600, CBV712, CBV760, and CBV780, respectively)
(
15, 29). This correspondence supports the validity of the
procedure adopted for the quantification of the accessible
acid sites.
�
FIG. 1. IR spectra in the O–H stretching region of B25 (commercial)
and lB36t cec (lab-synthesized) Beta zeolites.

7
2
COLLIGNON AND PONCELET
integrated areas of the different OH bands normalized to
the sample weights are given in Table 2. Compared with
the commercial samples, the lab-synthesized zeolites (sam-
ples lBt) had fewer SiOH groups, with a higher proportion
at external than internal position, and also more bridging
�
1
AlOHSi groups (band at 3610 cm ), consistently with the
higher NH3-TPD values (Table 1). The bands at 3660 and
�
1
3
780 cm
were mainly observed over the commercial
samples. According to Jia et al. (33), the intensity of
these bands increases when hot spots or self-steaming
conditions are favored during the calcination (deep-bed
�
conditions and/or temperature higher than 500 C). The
band at 3780 cm
�
1
has recently been attributed to a
hydroxyl group on a tri-coordinated Al partially con-
nected to the framework (39). These OH groups have a
weaker acid character than those absorbing at 3610 and
� � 1
1
3
660cm (31, 40). The band at 3660cm hasbeen assigned
to extra-framework AlOH groups (32, 34) or OH groups
bonded to tetrahedral Al partially disconnected from the
framework (41).
The acidity of the Beta zeolites was also investigated
by IR spectroscopy of adsorbed pyridine. After adsorp-
tion of pyridine and subsequent outgassing, the bands at
�
1
3
780 (VHF), 3660, and 3610 cm disappeared, thus veri-
fying their acidic character. The relative quantities of the
Brønsted and Lewis acid sites estimated from the normal-
�
1
ized areas of the bands at 1545 (Py-B) and 1446 cm (Py-L)
(
FIG. 2. (a) Relation between total acidity determined by NH3-TPD
�
samples outgassed at 150 C) corrected by the respective
extinction coefficients of Emeis (42) are given in Table 2. (in a.u. g ). (b) Relation between Lewis acidity determined by pyridine
The relation between the integrated areas of the Py-B band
� 1
(
mmol g ) and normalized band area of the bridging AlOHSi groups
�
1
� � 1
1
(
normalized area of the band at 1456 cm , a.u. g ) and the difference
�
1
�
1
� 1
between total Al and acid content determined by NH3-TPD (mmol g ).
at 1545 cm versus the bridging OH at 3610 cm , both
normalized to the sample weight, is shown in Fig. 2a. A
similar relationship was observed between the NH3-TPD
values and the integrated area of that same band (figure
not shown here). The fact that the relationship did not pass
through the origin is probably related to the presence of
reversible Al, typical of Beta zeolites (30), and/or to
different AlOH Brønsted acid sites (bands at 3780 and
�
to ETBE conversion, and the yields of C8 at 100 C ob-
tained over the different series of zeolites, are compiled
in Table 3. CBV760 and CBV712 were more active than
the other US-Y zeolites, with a maximum ETBE yield of
�
1
0–11% at about 80 C. CBV760 also was the most active
one in the vapor phase synthesis of MTBE (15). None of the
�
US-Y zeolites was active at 45 C. The ZSM-5 zeolites were
�
1
3
660 cm ). A relation was also observed when plotting
all less efficient than the US-Y ones, with yields at max-
�
1
the normalized areas of the Py-L band at 1445 cm against
the difference between the total Al content (Altot) and the
�
imum conversion barely exceeding 5% (at around 90 C),
namely, half the value obtained over the best US-Y zeolites,
and with only a limited effect of the Si/Al ratio (between
�
NH3-TPD values (Fig. 2b). Except for the EFAl-rich B25
and the nonexchanged lB18t c samples, the Lewis acid con-
tent increased with the EFAL content.
1
6.5 and 38.3). ZSM-5 with Si/Al = 137.5 was markedly
less active. The yields of ETBE at maximum IB conver-
sion obtained over a dealuminated large-port mordenite
and Omega (LZ-6) were 8 and 5% , thus in the range of
Catalysis
The yield of ETBE versus reaction temperature curves values found for the US-Y and ZSM-5, respectively.
all exhibited a similar shape. With increasing temperature, The yields versus temperature curves obtained over the
the IB to ETBE conversion increased, reached a maximum, commercial Beta zeolites and the reference Amberlyst-15
�
and decreased beyond this maximum, with the most active are compared in Fig. 3. B25 , B75, and Bsc, with nearly
catalysts operating the synthesis at lower temperatures. The identical acid content (Table 1), exhibited similar catalytic
decreasing part of the curves superimpose, as imposed by performances and were less active than the resin, while B25,
the thermodynamic equilibrium of the reaction at each tem- with more acid sites, was a little more active than the resin
�
�
�
perature. The yields of ETBE at 45 C and at maximum IB (yield of 27% at 54 C vs 25% at 60 C). Greater amounts

VAPOR PHASE SYNTHESIS OF ETBE
73
TABLE 3
mances, except for the sample treated with 0.09 M nitric
acid, which also contained less acid sites. A treatment with
0.02–0.05 M solutions (Si/Al ratio in the range 17.4–22) gave
the best results, but the yields of ETBE at maximum con-
version remained below that of the resin. This table also
shows that the lab-synthesized samples that were either cal-
cined (samples lBt c) or calcined, ammonium-exchanged,
and decationed (samples lBt cec) were all more active than
the other H-Beta zeolites and Amberlyst-15, with yields of
ETBE at maximum conversion above 28% . None of these
samples, however, equalized the performance of B25t 0.1 c,
the most efficient sample (Fig. 3).
�
Yield of ETBE at 45 C, at Maximum IB to ETBE Conversion
Ymax, mol %), Temperature at Maximum Conversion (Tmax, C),
and Yield of Di-isobutenes at 100 C (YC8, mol %)
�
(
Y45
5
Ymax
25
Tmax
60
YC8
7
Amb.-15
H-Beta
Bsc
B75
B25
B25*
B25* 0.02
B25* 0.03
B25* 0.05
B25* 0.06
B25* 0.09
lB18t c
lB18t cec
lB28t c
lB28t cec
lB36t c
lB36t cec
B25t c
4
3
15
3
5
7
4
6
3
23
22
30
28
15
23
21
23
32
20
19
27
20
22
23
22
21
20
28
29
31
31
28
30
32
33
34
65
65
55
67
64
60
60
66
66
53
55
48
51
55
55
51
46
50
8
11
14
3
11
12
12
9
13
15
14
21
15
13
14
19
19
18
For all the Beta zeolites, the selectivity to ETBE was
�
100% in the temperature range 36–55 C (up to the maxi-
mum yield of ETBE). Di-isobutenes started to appear at
temperatures beyond that of maximum IB to ETBE con-
�
version and, at 100 C, substantial amounts were produced
over the most active samples. The other common side reac-
tion products in the synthesis of ETBE (43), namely, tert-
butyl alcohol, diethyl ether, and 2,4,4-trimethyl-2-ethoxy-
pentane, were also identified in the reaction products, but
in amounts less than 1% .
B25t cec
B25t 0.1 c
US-Y
DISCUSSION
CBV500
CBV600
CBV712
CBV760
CBV780
0
0
0
0
0
8
9
10
11
9
86
80
80
80
86
0
1
0
1
0
The activity of the different zeolites followed the seque-
nce: H-Beta > US-Y � H-mordenite > ZSM-5 = Omega. It
has been shown in the synthesis of MTBE that the catalytic
performance could not be related to a single characteristic
feature of the zeolite, but merely to the resultant of several
factors acting together, with some having a prevailing influ-
ence (15). Since the reaction mainly occurs at the periphery
of the zeolite crystallites, it is expected that the Beta zeo-
lites, with higher external surface area (Sext), will be more
active than those with lower Sext. Another feature distin-
guishing Beta zeolites from the other zeolites is the higher
hydroxyl contents of the former ones, which intervene as
adsorption sites for the alcohol molecules (16–19, 31, 33,
H-ZSM-5
CBV3020
CBV5020
CBV8020
CBV2802
0
0
0
0
5
5
5
3
92
95
95
1
0
0
0
105
H-Mordenite
M-S600H6
0
0
8
5
80
90
2
0
H-Omega
LZ-6
of ETBE were produced over B25t c, B25t cec, and B25t 34). These aspects will be briefly discussed hereafter.
.1-c with, at maximum IB to ETBE conversion, yields be- As shown in Fig. 4, the yield at maximum IB to ETBE
0
tween 32 and 35% (25% for the resin), namely, a relative conversion isproportionalto the externalsurface area (Sext)
increase of between 28 and 40% . These three samples ex- of the zeolite. It verifies similar relationships obtained in
�
hibited high conversions below 40 C. It has to be remarked previous studies (15, 20) where it was inferred that the re-
that the comparison with the resin may be questionable action was strongly diffusion limited. The external surface
since this catalyst is normally used in liquid phase condi- area was also an important parameter in the alkylation of
tions where the polar medium causes a swelling of its mi- isobutane with 1-butene (44) and in that of phenol with
croreticular structure, which maysubstantiallyinfluence the tert-butyl alcohol (45). However, the sole external surface
catalytic performance. However, B25 was slightly more ac- area cannot account for the differences of activity observed
tive at low temperatures than Amberlyst-15 (15) in vapor among the H-Beta zeolites. Indeed, although the series of
phase synthesis of MTBE; it was significantly more efficient H-Beta zeolites had similar Sext, the lab-synthesized sam-
than the resin in liquid phase conditions (20).
ples (lBt) and those derived from B25t (with template) were
�
�
The results obtained over samples of B25 leached with substantially more active than B25, B25 (and the acid-
�
nitric acid and calcined at 400 C are given in Table 3. Dea- leached forms), B75, and Bsc (Fig. 4 and Table 3), therefore
lumination had a beneficial effect on the catalytic perfor- pointing to the intervention of other factors.

7
4
COLLIGNON AND PONCELET
FIG. 3. Variation of the yield of ETBE vs reaction temperature over commercial Beta zeolites and Amberlyst-15.
Partial dealumination (increasing Si /A l ratio) enhanced
It has been observed in the vapor phase synthesis of
the ETBE yield by increasing the accessibility of the acid MTBE that the presence of EFAl species had a negative
sites (e.g., US-Y zeolites) and/or removing some EFAL effect on the reaction (15). This effect was clearly estab-
�
species (e.g., the acid-leached forms of B25 ). For the Beta lished for the series of US-Y zeolites for which a linear
�
�
zeolites, the yield at 45 C (and at maximum conversion) relationship exists between the reaction rate at 60 C and
increased with the total amount of acid sites determined the proportion of EFAL species quantified by 27Al MAS
by NH3-TPD, as shown in Fig. 5. The different slopes are NMR spectroscopy (signals at 0 and 30 ppm) (29). CBV760,
interpreted as an indication that at low temperature (and with the lowest EFAL content, was the most active one.
conversion), the reaction occurs at the external surface and In the present study, this zeolite is also more active among
partly in the micropores of the zeolite. At maximum IB to the series of US-Y zeolites. The higher yields of ETBE ob-
�
ETBE conversion (higher temperatures), the less marked tained over B25 relative to B25 can be accounted for by
acidity effect suggests that the reaction is more diffusion the same reason. These two zeolites have, indeed, similar
limited and merely takes place at the external surface.
Si /A l ratios but different amounts of acid sites, as found by
FIG. 4. Yield at maximum IB to ETBE conversion vs external surface area of the different zeolites.

VAPOR PHASE SYNTHESIS OF ETBE
75
�
FIG. 5. Beta zeolites: yield of ETBE at 45 C and maximum conversion vs acid content measured by NH3-TPD.
NH3-TPD (Table 1) as well as by the normalized area of investigated (Fig. 1 and Table 2). Figure 6 shows the re-
�
1
�
the AlOHSi band at 3610 cm (Table 2), both pointing to lationship between the yields of ETBE at 45 C and the
�
27
the presence of EFAL species in B25 . Quantitative Al amount of acid sites. Similar yields were reached for sam-
MAS-NMR analysis of sample B25 showed the presence ples having high contents of AlOHSi sites (lBt). These
�
IV
VI
of 58% Al (signal at 54 ppm), 25% Al (0 ppm), the samples were also characterized by the presence of fewer
difference (17% ) obtained by subtracting the Al content silanol groups at the outer surface of the crystallites (band
2
7
� 1
established by quantitative Al MAS-NMR from that de- at 3743–45 cm ). Hunger and co-workers (16–19) pro-
termined by ICPS representing the “invisible” Al. Removal posed that the SiOH groups act as a reservoir of reactant,
of these EFAL species by mild acid leaching had a beneficial forming clusters of alcohol molecules near the active sites.
effect on the catalytic performances.
Plotting the ETBE yields against the SiOH/AlOHSi nor-
The zeolites lBt and B25t exhibited higher acid con- malized band area ratio of different Beta zeolites (Fig. 7)
tents and bridging AlOHSi than the other Beta zeolites shows that those with large ratios are less active and,
�
� 1
FIG. 6. Beta zeolites: yield of ETBE at 45 C and maximum conversion vs normalized band area of bridging AlOHSi groups (a.u. g ).

7
6
COLLIGNON AND PONCELET
�
FIG. 7. Beta zeolites: yield of ETBE at 45 C and maximum conversion vs SiOH/AlOHSi normalized band area ratio.
conversely, the more active ones have smaller ratios, with
the most active catalysts apparently describing an opti-
mal SiOH/AlOHSi value. This suggests that ETBE syn-
thesis over Beta zeolites is favored by a high amount of
bridging AlOHSi (active sites) and a moderate amount
of SiOH groups at the external surface, preventing the
formation of too large and/or numerous ethanol clus-
ters inhibiting the access of isobutene to the AlOHSi
acid sites.
CONCLUSION
Vapor phase synthesis of ETBE performed over differ-
ent acid zeolites has shown the following sequence of ac-
tivity: Beta zeolite > US-Y > Mordenite > Omega � ZSM-
2
� 1
,
5
. Beta zeolites, with externalsurface area above 200m g
are much more active than the other zeolites. At maximum
isobutene to ETBE conversion, the lab-synthesized Beta
zeolites produce more ETBE than the commercial samples
and acid resin. Both Beta zeolites and Amberlyst-15 are
No significant effect on the yield of ETBE was no-
ticed whether calcination of Beta zeolites to remove the
template was followed by an ammonium-exchange or not
�
1
00% selective below 55 C. The reaction occurs on bridg-
ing AlOHSi acid sites of the zeolites, with higher yields of
ETBE for those with high content of acid sites (AlOHSi)
and low SiOH/AlOHSi ratios, where the number of ethanol
clusters around the silanol groups still allows an easy access
of isobutene to the AlOHSi active sites. Extra-framework
Al species have a detrimental influence on the reaction,
and their removal by a mild acid leaching is beneficial. Cal-
cination of as-synthesized Beta zeolites to remove the tem-
plate without and with subsequent ammonium exchange
and thermal decomposition provide or a treatment with di-
luted acid prior to the elimination of the template provide
very efficient catalysts.
(
compare lBt cec and lBt c, and B25t cec and B25t c, Ta-
ble 3). It has been reported that for an Al/(Al + Si) molar
ratio of 0.085 (Si/Al = 10.8), all the template molecules
are associated with framework Al, with the Na cations
being located at structural defects (SiONa) (46). How-
ever, it has been shown that the ammonium ions resulting
from the decomposition of the template (tetraethyl am-
monium) occupy negative SiOAlO3Si3 centers, with the
protons generated by decationation neutralizing the neg-
atively charged framework sites (47). These observations
may explain the catalytic results. For these samples, the
similar amounts of acid sites determined by NH3-TPD
ACKNOWLEDGMENTS
(
Table 1), on the one hand, and the corresponding yields
of ETBE (Table 3), on the other hand, are therefore
consistent.
F. Collignon gratefully acknowledges F. R. I. A. (Fonds pour la For-
mation a` la Recherche dans l’Industrie et l’Agriculture), Belgium, for a
Obviously, the activity was improved when Beta zeolite doctoral grant. The authors are indebted to P. Grobet, (Centrum voor
2
7
stillcontainingitstemplate (B25t) wastreated with 0.1M ni- Oppervlaktechemie en Katalyse, K. U. Leuven), for the Al MAS-NMR
measurements.
tric acid prior to the thermal decomposition of the template
(
B25t 0.1 c). This particular sample produced 32% ETBE
REFERENCES
�
at 45 C, a substantial gain compared with the other Beta
samples and the acid resin, thus pointing to the beneficial
effect using this activation method.
1
. H a¨ lsig, C. P., Gregory, P., and Peacock, T., Spec. Publ.—R. Soc. Chem.
97, 311 (1991).

VAPOR PHASE SYNTHESIS OF ETBE
77
2
3
4
. Peaff, G., Chem. Eng. News 8, Sept. 26 (1994).
25. de Boer, J. H., Linsen, B. G., and Osinga, T. J., J. Catal. 4, 643 (1964).
26. Niwa, M., Katada, N., Sawa, M., and Murikami, Y., J. Phys. Chem. 99,
8812 (1995).
27. Katada, N., Iijima, S., Iigi, H., and Niwa, M., in “Studies in Surface
Science and Catalysis,” Vol. 105, p. 1227. Elsevier Sci., Amsterdam,
1997.
. Karinen, R. S., and Krause, A. O. I., Catal. Lett. 67, 73 (2000).
. Leprince, P., and Travers, P., in “Le raffinage du p e´ trole. 3 Proc e´ d e´ s
de transformation,” p. 299. Publ. IFP, Technip, Paris, 1998.
. Vicencio, G. T., Ramos, A. O., Villanueva, J. R., and Lopez, G. P., Rev.
Inst. Mexicano Petrol. 24, 63 (1992).
5
6
7
. Fran c¸ oisse, O., and Thyrion, F. C., Chem. Eng. Process. 30, 141 (1991).
. Sola, L., Pericas, M. A., Cunill, F., and Tejero, J., Ind. Eng. Chem. Res.
28. Coutanceau, C., Da Silva, J. M., Alvarez, M. F., Ribero, F. R., and
Guisnet, M., J. Chim. Phys. 94, 765 (1997).
3
4, 3718 (1995).
29. Remy, M. J., Stanica, D., Poncelet, G., Feijen, E. J. P., Grobet, P. J.,
Martens, J. A., and Jacobs, P. A., J. Phys. Chem. 100, 12440 (1996).
30. Bourgeat-Lami, E., Massiani, P., Di Renzo, F., Fajula, F., and Des
Couri e` res, T., Catal. Lett. 5, 265 (1990).
31. Kiricsi, I., Flego, C., Pazzuconi, G., Parker, W. O., Jr., Millini, R.,
Perego, C., and Bellussi, G., J. Phys. Chem. 98, 4627 (1994).
32. Beck, L. W., and Haw, J. F., J. Phys. Chem. 99, 1076 (1996).
33. Jia, C., Massiani, P., and Barthomeuf, D., J. Chem. Soc., Faraday Trans.
89, 3659 (1993).
8
. Fit e´ , C., Iborra, M., Tejero, J., Izquierdo, J. F., and Cunill, F., Ind. Eng.
Chem. Res. 33, 581 (1994).
. Cunill, F., Vila, M., Izquierdo, J. F., Iborra, M., and Tejero, J., Ind. Eng.
Chem. Res. 32, 564 (1993).
0. Izquierdo, J. F., Cunill, F., Vila, M., Iborra, M., and Tejero, J., Ind. Eng.
Chem. Res. 33, 2830 (1994).
1. Tau, L.-M., and Davis, H., Appl. Catal. 53, 263 (1989).
2. Larsen, G., Lotero, E., Marquez, M., and Silva, H., J. Catal. 157, 645
9
1
1
1
(
1995).
34. Maache, M., Janin, A., Lavalley, J. C., Joly, J.-F., and Benazzi, E.,
Zeolites 13, 419 (1993).
35. Guisnet, M., Ayrault, P., Coutanceau, C., Fernanda Alvarez, M., and
Datka, J., J. Chem. Soc. Faraday Trans. 93, 1661 (1997).
36. Paz e´ , C., Bordiga, S., Lamberti, C., Salvalaggio, M., Zecchina, A., and
Bellussi, G., J. Phys. Chem. 101, 4740 (1997).
1
3. Fricke, R., Richter, M., Zubowa, H. L., and Schreier, E., in “Stud-
ies in Surface Science and Catalysis,” Vol. 105, p. 655. Elsevier Sci.,
Amsterdam, 1997.
4. Richter, M., Zubowa, H. L., Heckelt, R., and Fricke, R., Microporous
Mater. 7, 119 (1997).
1
1
1
5. Collignon, F., Mariani, M., Moreno, S., Remy, M., and Poncelet, G.,
J. Catal. 166, 53 (1997).
6. Hunger, M., Horvath, T., and Weitkamp, J., in Proc. DGMK Conf. “C4
Chemistry-Manufacture and Uses of C4 Hydrocabons,” p. 65. German
Soc. Petrol. Coal Sci. Techn., 1997.
37. Newsam, J., Treacy, M., Koetsier, W., and de Gruyter, C., Proc. R. Soc.
London, Ser. A 420, 375 (1988).
38. Treacy, M., and Newsam, J., Nature 332, 249 (1988).
39. Vimont, A., Thibault-Starzyk, F., and Lavalley, J. C., J. Phys. Chem. B
104, 286 (2000).
1
1
7. Hunger, M., and Horvath, T., Catal. Lett. 49, 95 (1997).
8. Hunger, M., Horvath., T., and Weitkamp, J., Microporous. Mater. 22,
40. Su, B.-L., and Norberg, V., Zeolites 19, 65 (1997).
41. Loeffler, R. E., Lohse, U., Peuker, Ch., Oelhmann, G., Kustov, L. M.,
Zholobenko, V. L., and Kazansky, V. B., Zeolites 10, 266 (1990).
42. Emeis, A., J. Catal. 141, 347 (1993).
43. Richter, M., Fisher, H., Bartoszek, M., Zubowa, H. L., and Fricke, R.,
Microporous Mater. 8, 69 (1997).
44. Loenders, R., Martens, J. A., and Jacobs, P. A., J. Catal. 176, 545 (1998).
45. Zhang, K., Huang, C., Zhang, H., Liu, S., Xu, D., and Li, H., Appl.
Catal., A 166, 89 (1998).
46. Vaudry, F., Di Renzo, F., Fajula, F., and Schulz, P., in “Studies in Surface
Science and Catalysis,” Vol. 84, p. 163. Elsevier Sci., Amsterdam, 1994.
3
57 (1998).
9. Horvath, T., Seiler, M., and Hunger, M., Appl. Catal., A. 193, 227
2000).
1
2
(
0. Collignon, F., Loenders, R., Martens, J. A., Jacobs, P. A., and Poncelet,
G., J. Catal. 182, 302 (1999).
1. Moreno, S., and Poncelet, G., Microporous Mater. 12, 197 (1997).
2. Fajula, F., Bourgeat-Lami, E., Zivkov, C., Des Couri e` res, T., and
Anglerot, D., Fr. Patent. 2, 069, 618 (1992).
2
2
2
3. Borade, R. B., and Clearfield, A., Microporous Mater. 5, 289 (1996).
2
4. Volosch, M., Ladisch, M. R., and Tsao, G. T., React. Polym. 4, 9 47. Bougeat-Lami, E., Di Renzo, F., and Fajula, F., J. Phys. Chem. 96, 3087
1986). (1992).
(