N-Butanol to iso-butene in one-step over zeolite catalysts

Article abstract of DOI:10.1016/j.apcata.2011.05.037

The conversion of n-butanol to iso-butene in one-step has been studied over zeolite catalysts (Theta-1, ZSM-2, Ferrierite, ZSM-5, SAPO-11 and Y). The zeolites were characterized by XRF, XRD, BET, and NH₃-TPD. Coke formation on the spent catalysts was measured by TGA. The unidirectional channel zeolites Theta-1 and ZSM-23 gave high yields of iso-butene and showed high stability. In contrast, ferrierite, although initially active, deactivated due to the presence of water from dehydration giving progressively lower yields of iso-butene. ZSM-5 showed high activity but low selectivity to iso-butene; SAPO-11 was structurally unstable under the reaction conditions. Ferrierite deactivation was studied in more detail by NH₃-TPD, ²⁷Al MAS NMR and by a comparison of 1-butene isomerisation with and without the presence of added steam. A similar comparison was made for ZSM-23 and Theta-1 which further emphasised their stability and showed that for these catalysts under the reaction conditions used, the product steam from n-butanol dehydration promoted iso-butene yield by acting as a diluent.

Full text of DOI:10.1016/j.apcata.2011.05.037

Applied Catalysis A: General 403 (2011) 1–11
Contents lists available at ScienceDirect
Applied Catalysis A: General
journal homepage: www.elsevier.com/locate/apcata
n-Butanol to iso-butene in one-step over zeolite catalysts
∗
Dazhi Zhang, Sami A.I. Barri, David Chadwick
Department of Chemical Engineering, Imperial College London, South Kensington Campus, London SW7 2AZ, UK
a r t i c l e i n f o
a b s t r a c t
Article history:
The conversion of n-butanol to iso-butene in one-step has been studied over zeolite catalysts (Theta-1,
ZSM-2, Ferrierite, ZSM-5, SAPO-11 and Y). The zeolites were characterized by XRF, XRD, BET, and NH3-
TPD. Coke formation on the spent catalysts was measured by TGA. The unidirectional channel zeolites
Theta-1 and ZSM-23 gave high yields of iso-butene and showed high stability. In contrast, ferrierite,
although initially active, deactivated due to the presence of water from dehydration giving progressively
lower yields of iso-butene. ZSM-5 showed high activity but low selectivity to iso-butene; SAPO-11 was
structurally unstable under the reaction conditions. Ferrierite deactivation was studied in more detail by
NH3-TPD, Al MAS NMR and by a comparison of 1-butene isomerisation with and without the presence
of added steam. A similar comparison was made for ZSM-23 and Theta-1 which further emphasised their
stability and showed that for these catalysts under the reaction conditions used, the product steam from
n-butanol dehydration promoted iso-butene yield by acting as a diluent.
Received 29 December 2010
Received in revised form 27 May 2011
Accepted 28 May 2011
Available online 6 June 2011
Keywords:
n-Butanol
iso-Butene
Dehydration
Isomerisation
Zeolite
2
7
©
2011 Elsevier B.V. All rights reserved.
1
. Introduction
has been widely studied [12–16] and zeolites have been shown to
be more effective catalysts than other solid acids such as resins and
modified alumina. In particular, zeolites with medium pore (10-
ring pore structure) unidirectional channels have high activity and
selectivity as well as performance stability. Such zeolites included
ferrierite, SAPO-11, Theta-1/ZSM-22, and ZSM-23. Ferrierite, in par-
ticular, was found to be the most active and selective for 1-butene
isomerisation [1,17–19]. The high performance of these zeolites is
attributed to a combination of their high acidity and their ability
to restrict side reactions such as dimerization and aromatization.
The reaction mechanism has been debated extensively in the lit-
erature and both monomolecular and bimolecular pathways have
been proposed [20–24]. The monomolecular mechanism is more
selective and it predominates in aged catalyst as coke deposition
can also suppress the bimolecular reaction. Coke formation has
also been proposed to be part of the catalytic cycle as a pseudo-
monomolecular reaction mechanism [16].
The one-step conversion of n-butanol to iso-butene combines
dehydration and skeletal isomerisation, but also introduces water
into the reaction fluid. In our preliminary work [5], we demon-
strated that zeolites with 10-ring pores and unidirectional channels
gave high activity and selectivity for iso-butene production. We also
found that ferrierite was unstable at the reaction conditions and
speculated that it was a consequence of the product water from the
dehydration step. In this paper, we compare the performances of
zeolite catalysts, including SAPO-11, in the conversion of n-butanol
to iso-butene and investigate in more detail the contribution of
framework composition and framework structure to the catalyst
activity, product selectivity and catalyst stability. We report on the
extent of coke formation, and further elucidate the influence of the
iso-Butene is an important chemical intermediate for the pro-
duction of polyisobutene, alkylates and gasoline additives such as
diisobutene and ethyl tertiary butyl ether (ETBE) [1–4]. The source
of isobutene has traditionally been through dehydrogenation,
steam cracking and fluid catalytic cracking. There are increasing
environmental pressures to develop production routes to chemical
intermediates based on non-fossil fuels such as biomass. Recently,
we have reported n-butanol dehydration and isomerisation to iso-
butene in one-step over zeolite catalysts with high productivity
[
5]. This is a promising route to iso-butene from bio-derived n-
butanol (Scheme 1) since n-butanol can be obtained as a feedstock
by fermentation of biomass potentially on large scale [6,7].
n-Butanol dehydration is a facile reaction and takes place over
many acidic catalysts at low temperature[8–10]. However, only
linear butenes are produced at such conditions. Dehydration of
various butanols over sulphated ␥-alumina at higher reaction tem-
perature was studied by Macho et al. [11]. In their study, iso-butanol
and tert-butanol produced a much higher iso-butene yield than
linear butanols. A very low space velocity was required to pro-
duce around 31% iso-butene selectivity from n-butanol giving low
iso-butene productivity.
Linear butenes skeletal isomerisation requires stronger acidity
and higher temperatures than butanol dehydration. This reaction
∗ Corresponding author. Tel.: +44 020 7594 5579; fax: +44 020 7594 5637.
E-mail address: d.chadwick@imperial.ac.uk (D. Chadwick).
0
926-860X/$ – see front matter © 2011 Elsevier B.V. All rights reserved.
doi:10.1016/j.apcata.2011.05.037

2
D. Zhang et al. / Applied Catalysis A: General 403 (2011) 1–11
Craking
Petroleum source
-
H O
Skeletal isomerisation
2
Fermentation
-H O
2
Biomass
OH
gasification,
Skeletal isomerisation
alcohol synthesis
Scheme 1. Reaction route to produce iso-butanol from biomass.
product water from dehydration by use of solid-state NMR and a
comparison between the reaction of n-butanol to iso-butene and
that of 1-butene to iso-butene with and without added water.
40 ml/min helium for 30 min. After the temperature was decreased
◦
to 110 C, and ammonia was absorbed to saturation over the cata-
◦
◦
lyst. Then ammonia was desorbed from 110 to 660 C at 10 C/min
in 40 ml/min helium. The desorbing ammonia was identified by
m/z = 16.
2
. Experimental
2.1. Catalysts
2.5. Catalyst evaluation
ZSM-23, Theta-1 and SAPO-11 were synthesised according to
published methods [25–27]. The dried as-synthesised zeolites were
calcined at 550 C to remove the organic templates in a flow of
n-Butanol dehydration was carried out at steady conditions in a
◦
fixed bed reactor. 0.50 g of the catalyst granules was diluted with
inert silicon carbide granules of approximately the same size to
make up a volume of 7.0 ml. The diluted catalyst was loaded into a
tubular reactor (500 mm in length and 11 mm internal diameter),
sandwiched between shallow beds of inert granules. SiC was used
as diluent because of its inertness and heat transfer properties, and
the particle size, dilution ratio and pre/post packing were used to
maintain plug flow and isothermality. Before reaction, the catalyst
air, raising the temperature from room temperature at the rate
◦
of 1 C/min. The calcined zeolites were ion exchanged with con-
centrated ammonium nitrate solution several times to ensure no
metal cations were present. The ammonium exchanged samples
◦
were then calcined at 550 C to convert to the H-forms. Ferrierite
(
CP 914 and CP 914C), ZSM-5 (CBV 5524G), and Y (CBV 760) were
supplied by Zeolyst International. These zeolites were supplied in
◦
the ammonium or proton form and were calcined at 550 C before
◦
was pre-activated at 550 C for 6 h in argon, and then lowered to
use. The zeolite powders were pressed into tablets (without use
of lubricant or binder), crushed and sieved to granules in the size
range of 0.5–0.8 mm for the catalytic tests.
◦
the set reaction temperature between 300 to 500 C. n-Butanol was
fed into the catalyst bed via an evaporator at 160 C, through which
◦
the gas diluent (argon) was passed. The total flow rate was in the
range of 110–215 ml/min and the molar ratio of n-butanol to Ar
was 1:1, except for a few experiments at 2:1 and 1:2. GHSV values
are given with reference to the volume of the undiluted catalyst.
The mixture was passed through the catalyst and the liquid prod-
uct was collected in a condenser kept at room temperature. The
flow of the gaseous product was measured and the liquid prod-
uct was weighed. The hydrocarbon and aqueous products were
analysed by GC with FID detector. The gaseous, hydrocarbon liquid
and aqueous liquid products were analysed on Plot Al O /KCl, BP
2.2. Catalyst characterisation
The SiO /Al O ratio of each of the zeolites was determined
2
2
3
directly on the powder using Bruker S4 explorer X-ray Fluores-
cence spectrometer. X-ray powder diffraction was carried out
using PANalytical X’Pert PRO equipped with PW3015 X’Celerator
detector and Cu K␣ radiation (ꢀ = 1.5406 A˚ ). Solid-state 27Al MAS
NMR spectra were obtained on a Bruker 600 MHz NMR spec-
trometer. Surface areas (BET) and pore volumes were measured
on a Micromeritrics TriStar instrument. The pore volumes were
calculated according to the procedure of Remy and Poncelet
2
3
and HP-FFAP capillary columns respectively. The study of 1-butene
isomerisation was carried out using the same reactor system.
Water or n-butanol was introduced into the 1-butene via the
evaporator operated at the appropriate temperature. Two evap-
orators were used for the experiments with cofeeding of water and
n-butanol.
[
28] utilising the Dubinin and Radushkevich equation. TGA was
carried out using a TA Instruments Q500 Thermogravimetric
Analyzer.
The overall conversion comprised n-butanol dehydration fol-
lowed by linear butenes isomerisation. Thus, the reaction was
evaluated by dehydration and isomerisation parameters. Because
of the rapid equilibrium of linear butenes, the three linear butenes
(1-butene, t-butene-2 and c-butene-2) were considered as reactant
to the isomerisation reaction. The n-butanol conversion (Xn-butanol),
2
.3. Temperature programmed surface reaction (TPSR)
TPSR was carried out using a quartz reactor equipped with mass
spectra detector. Before reaction, the catalyst (0.02 g) was pre-
activated at 550 C for 1 h in argon. n-Butanol was introduced by
◦
a flow of argon at 65 ml/min through a n-butanol evaporator at
product yields (Y ), linear butenes conversion (Xn-butene), isomerisa-
tion selectivity (Siso) and isomerisation activity (Aiso) were defined
as:
i
◦
◦
◦
5
0 C. The catalyst sample was heated to 400 C at 10 C/min and
◦
held at 400 C for 0.5 h. The m/z ratios used to identify species were:
m/z = 31 for butanol, m/z = 41 for butenes, m/z = 18 for water.
ꢀ
ꢁ
Detected n-butanol in products
n-Butanol in reactant
Xn-butanol(%) = 1 −
× 100
(1)
(2)
2
.4. Temperature program desorption of ammonia (NH -TPD)
3
Ammonia temperature program desorption was carried out
Detected compoundi
All the detected compounds
using a mass spectrometer detector. 0.1 g catalyst granules were
loaded in a stainless steel reactor and pre-treated at 660 C in
Yi(wt.%) =
× 100
◦

D. Zhang et al. / Applied Catalysis A: General 403 (2011) 1–11
3
Table 1
SiO2/Al2O3 ratios and textural properties of zeolites.
2
3
Catalyst
SiO2/Al2O3
SBET (m /g)
Pore volume (cm /g)
Ferrierite (CP 914C)
Ferrierite (CP 914)
ZSM-5 (CBV 5524G)
Y (CBV 760)
Theta-1
ZSM-23
20
55
50
60
63
68
-
367
323
408
814
208
204
199
0.148
0.130
0.175
0.352
0.079
0.071
0.106
SAPO-11
perature (LT) and the other at high temperature (HT). The HT was
due to ammonia adsorbed on the Bronsted acid sites on the sur-
face of the frameworks. The LT corresponded to ammonia hydrogen
bonded to those adsorbed and contained within the micropores of
the channels [32]. As all the TPD-NH experiments were carried out
at the same conditions, the temperature corresponding to the HT
3
Fig. 1. The XRD patterns of the synthesised zeolite.
All the detected dehydration compounds-linear butenes
X
n−butene(%) =
× 100
× 100
(3)
(4)
All the detected dehydration compounds
Amount of iso-butene
All the detected hydrocarbon compounds − linear butenes
Siso(wt.%) =
Amount of iso-butene
Amount of linear butenes and iso-butene
peaks was characteristic of the strength of the acid site. As seen, the
strength of the acid sites from TPD-NH3 had the following trend:
ferrierite > ZSM-23 > Theta-1 = ZSM-5 > Y ꢀ SAPO-11. However, the
trend should be treated with caution as the channel structure can
impose diffusion constrains on the ammonia desorbed.
Aiso
=
× 100
(5)
In these equations, the water product is ignored and is considered
as a diluent. The detected dehydration compounds are non-
oxygenated hydrocarbons. The concentrations given are grouped in
terms of the number of the carbon atoms in the product. In evaluat-
ing yield by Eq. (2) the n-butanol feed is regarded as butene + water
and only hydrocarbons are considered since conversion was always
3.2. Catalytic performance
1
00% and the sum of oxygenated products was less than 0.01%.
The two reaction steps (dehydration and isomerisation,
Schemes 1 and 2) occur at different temperatures. Dehydration is
a more facile reaction and can be catalysed at a lower tempera-
ture and with a lower acid strength catalyst than the case for the
skeletal isomerisation of butenes as noted above. TPSR provided a
good technique to rapidly screen the temperature values to iden-
tify a range where the two reactions occurred to acceptable levels.
Catalyst testing at steady-state enabled comparison of the overall
performance of each zeolite at varying conditions.
Under these conditions yield becomes equivalent to selectivity.
3
. Results and discussion
The zeolite catalysts gave varying performances for n-butanol
to iso-butene reaction depending on their composition and struc-
tural properties. Theta-1(ZSM22) and ZSM-23 are known to have
related structures. Both zeolites have medium pore unidirectional
channels and share the same secondary building unit. Although the
structure of ferrierite is based on the same secondary building unit,
it contains 8-member ring channels perpendicular to and intersect-
ing with the unidirectional medium pore channels. SAPO-11, on the
other hand, has an entirely different structure and composition (sil-
icoaluminophosphate), but shares the presence of unidirectional
non-intersecting medium pore channels as those present in Theta-
A typical TPSR profile (of ferrierite) is shown in Fig. 3. The
◦
dehydration reaction started at 200 C and the level of conversion
1
3
and ZSM-23. Zeolite Y has an aluminosilicate framework with
-dimensional large pore (12-member rings) channels.
3.1. Catalyst characterisation
The X-ray powder diffraction patterns of the as-synthesised zeo-
lites are shown in Fig. 1. The patterns were analysed and found to
contain no crystalline impurities. The as-supplied zeolites were also
checked by XRD and were entirely consistent with the known struc-
tures. The surface areas of the samples are given in Table 1 and were
as expected from a pure phase of crystalline structure. The surface
areas of the zeolites are consistent with the literature values and the
supplier data sheets. The micropore volumes are consistent with
the structure of the zeolites and their framework densities [29–31].
The SiO /Al O ratios of the zeolites are also given in Table 1. The
2
2
3
level of acidity present corresponded to the aluminium content of
the zeolite. The TPD-NH3 profiles of the H-forms of the zeolites are
shown in Fig. 2. All the zeolites exhibited 2 peaks, one at low tem-
Fig. 2. NH3-TPD profiles of Y, ZSM-5, SAPO-11, Theta-1, ZSM-23, ferrierite (55)
zeolites.

4
D. Zhang et al. / Applied Catalysis A: General 403 (2011) 1–11
◦
the thermodynamic value (0.47) at 400 C. Further increase in the
◦
temperature to 500 C led to a marginal decrease of X
and
n-butene
Aiso, most probably due to thermodynamic equilibrium limitations
33]. A suitable reaction temperature for the one-step dehydra-
[
tion and isomerisation of n-butanol over zeolites lies in the range
◦
of 400–500 C. The highest iso-butene productivity is achieved at
◦
4
00 C.
3.3. Effect of zeolite type
◦
The performances of the catalysts with time-on-stream at 400 C
are shown in Fig. 4. The SiO /Al O ratios were within the narrow
2
2
3
range of 50–70, except for the ferrierite. Ferrierite with a SiO /Al O
2
2
3
ratio of 20 performed substantially better than a sample that had a
SiO /Al O ratio of 55, Table 3. Therefore, the performances of the
2
2
3
other zeolites were compared with the ferrierite (20) rather than
with the ferrierite with the higher SiO /Al O ratio. The composi-
2
2
3
tion of the framework provided the acidity required for the reaction
and the pore structure of the zeolite had an effect on the overall per-
formance especially the selectivity. Theta-1 and ZSM-23, as shown
in Fig. 4, gave very similar catalytic performances and achieved the
highest iso-butene yields amongst all the catalysts over the time-
on-stream period. This is due directly to their similar unidirectional
channel structure and composition. The iso-butene yield for both
zeolites actually increased with time-on-stream, reaching values
of 34 wt.% for Theta-1 and 31 wt.% for ZSM-23. The reason for this
is most probably due to coke laydown narrowing the pore struc-
ture and increasing the selectivity to iso-butene. Although ferrierite
initially gave comparable results to Theta-1 and ZSM-23, its per-
formance dropped steadily with time-on-stream, Fig. 4. Ferrierite,
like Theta-1 and ZSM-23, is known for its good, stable performance
Fig. 3. n-Butanol TPSR profile over ferrierite.
◦
quickly reached a maximum at 300 C. The signals for n-butanol,
butenes and water reached a plateau at a similar temperature indi-
cating the complete dehydration of n-butanol. A low level of carbon
chain growth and cracking reaction was detected.
The steady-state reaction of n-butanol over Theta-1 zeolite as a
function of temperature is given in Table 2. Complete dehydration
◦
was achieved at >300 C as found in the TPRS study. Additionally,
the linear butenes conversion X
increased with temperature
n-butene
◦
and reached a maximum at 400–450 C. The isomerisation reaction
selectivity Siso was greater than 60 wt.% at all the studied temper-
atures. The isomerisation activity as measured by Aiso was close to
Fig. 4. n-Butanol conversion over zeolites: (a) iso-butene yield, (b) isomerisation activity, (c) n-butene conversion, (d) isomerisation selectivity. Conditions: GHSV n-
−1
◦
butanol = 5200 h ; n-butanol/Ar = 1 mol; T = 400 C.

D. Zhang et al. / Applied Catalysis A: General 403 (2011) 1–11
5
Table 2
Products distribution of n-butanol conversion over Theta-1 at different temperatures.
a
◦
Yi (wt.%)b
i-C4=
−1 −1
Temperature ( C)
Xn-butanol (%)
Xn-butene (%)
Siso (wt.%)
Aiso
Productivity (kg kg cal
h
)
C4=−1
t-C4=−2
c-C4=−2
C5+
300
350
400
450
500
100
100
100
100
100
4.7
25.6
34.0
33.9
33.5
17.6
13.6
11.0
12.1
14.3
45
29.6
19.6
14.4
13.8
14.6
1.7
7.8
38.4
55.0
55.5
51.7
60.6
66.7
61.9
61.2
64.8
0.05
0.29
0.43
0.43
0.41
0.98
5.34
7.09
7.07
6.99
28.5
19.6
18.6
19.4
7.0
11.5
11.9
10.7
a
−1
Reaction conditions: n-butanol:Ar (mol) = 1:1, GHSV(n-butanol) = 5200 h
.
b
i-C4= iso-butene; C4=−1 1-butene; t-C4=−2 trans-butene-2; c-C4=−2 cis-butene-2; C5+ hydrocarbons of C5 and higher.
Table 3
n-Butanol conversion over ferrierite with different SiO2/Al2O3 ratio.a
Sample
Xn-butanol (%)
Yi (wt.%)b
i-C4=
Xn-butene (%)
Siso (wt.%)
Aiso
n-C4=
C3+C4
C3=+C5=
C6+
FER(20)
FER(55)
100
100
33.8
23
48.7
57.5
2.2
0.9
8.6
13.5
6
4.7
51.3
42.5
65.9
54.2
0.41
0.29
a
◦
−1
Reaction conditions: T = 400 C; n-butanol:Ar (mol) = 1:1, GHSV(n-butanol) = 5200 h .
b
i-C4= iso-butene; n-C4= 1-butene + 2-butenes; C3 propane; C4 butanes; C3=+C5= propene + pentenes; C6+ hydrocarbons of C6 and higher.
for the isomerisation of n-butenes to iso-butene [1,17–19]. Fur-
thermore the TPD-NH3 results, Fig. 2, indicated high acidity of the
surface. Therefore, it was surprising to see its performance decline
with time-on-stream. This decline was suspected to be due to the
presence of water, which was a product of the n-butanol dehydra-
tion step. As can be seen in Fig. 4d, the isomerisation selectivity
of the ferrierite (Siso) increased with time-on-stream in a similar
manner to Theta-1 and ZSM-23, and therefore the observed drop
in the yield of iso-butene was due to a reduction in the conversion
seen in the same figure that steaming Theta-1 at the same condi-
tion did not affect the catalyst acidity – either the acid strength or
the acid site concentration. It is evident from these results that the
presence of product water affected the stability and performance
of ferrierite but did not affect the Theta-1 or ZSM-23 catalysts over
the duration of the experiments. The permanent loss in activity of
ferrierite was confirmed by a study of n-butanol conversion over
samples regenerated by removal of the coke with air, Fig. 7.
SAPO-11 catalysed the dehydration of n-butanol, but not the
butenes to iso-butene reaction to the same level as achieved with
Theta-1 or ZSM-23. This is in contrast to the reported activity of
SAPO-11 for butenes isomerisation [34]. The reason for this dif-
ference in performance was the instability of the structure in the
presence of the product water. This was confirmed by XRD of the
spent SAPO-11 catalyst.
Both ZSM-5 and Y zeolites gave inferior performances to those
of Theta-1 and ZSM-23, as shown in Fig. 4. ZSM-5 demonstrated
high activity but low selectivity due to the production of higher
molecular weight and gaseous hydrocarbon products. The chan-
nel intersections present in the structure allowed dimerization,
oligomerisation and aromatisation to take place, thus lowering the
selectivity of the catalyst, Fig. 4. Zeolite Y performance, on the other
hand, gave high conversion of the n-butanol to n-butenes but low
isomerisation activity. However, this zeolite is known for its high
cracking activity and the most likely cause of its low isomerisation
activity is coke laydown.
(
Xn-butene) and therefore in the catalyst isomerisation activity. This
2
7
was investigated by Al solid state NMR and TPD-NH of steamed
3
and un-steamed ferrierite (20) samples. 27Al solid state NMR of the
ferrierite after steam treatment (at the same condition and with
similar water concentration as the catalytic reaction) showed an
increase in the levels of octahedral and asymmetrical aluminium
levels (0 ppm and 44 ppm) compared with an un-steamed sample,
as shown in Fig. 5 [32]. Furthermore, the TPD-NH3 of the steamed
sample compared with the un-steamed sample showed a drop in
the acidity with steaming, as shown in Fig. 6. In contrast, it can be
Characterisation of coke on the spent catalysts was carried out
by TGA in air and the differential loss of weight as a function of tem-
perature is shown in Fig. 8. The colours of the spent catalysts were
black for Y, ferrierite and ZSM-5; green for Theta-1 and grey for
ZSM-23. After the TGA experiment all the catalysts came out white.
◦
The peak of weight loss for Theta-1 and ZSM-23 at around 300 C
can be assigned to aliphatic carbon deposits, which do not affect
the catalytic activity appreciably except perhaps in improving the
◦
selectivity via subtle pore narrowing. The peak at around 425 C is
assigned to aliphatic higher molecular weight carbon deposits. The
◦
peak loss at around 525 C present for zeolite Y, which is assigned to
polynuclear aromatic coke, constituted around 12–15% weight loss
and accounts for the deactivation of the zeolite and its low olefin
isomerisation activity as discussed above. The peak is also present
in the TGA of spent ZSM-5 but to a lesser extent (4–5%) and in this
case it is consistent with the known aromatisation activity of this
zeolite. For the other zeolites, this peak was present only at a low
Fig. 5. Solid State 27Al NMR of: (a) fresh ferrierite (20), (b) steam treated ferrierite
(
20).

6
D. Zhang et al. / Applied Catalysis A: General 403 (2011) 1–11
Fig. 6. NH3-TPD profiles of: (a) ferrierite (20), (b) Theta-1.
Fig. 7. n-Butanol conversion over regenerated ferrierite: (a) iso-butene yield, (b) isomerisation activity, (c) n-butene conversion, (d) isomerisation selectivity. Conditions:
−1
◦
GHSV n-butanol = 5200 h ; n-butanol/Ar (mol) = 1; T = 400 C.

D. Zhang et al. / Applied Catalysis A: General 403 (2011) 1–11
7
level of weight loss, perhaps due to coke outside the channels, and
therefore did not contribute significantly to catalyst deactivation.
3.4. Comparison of 1-butene and n-butanol conversion
It is important to demonstrate that ferrierite has stable isomeri-
sation performance in the absence of water, and that differences
in the unidirectional channel zeolite catalyst performances with n-
butanol are due directly to the product water from dehydration.
Therefore, the effect of water on catalyst stability was investigated
further by a comparative study of 1-butene and n-butanol conver-
sion in the absence and presence of water. The results are shown
in Figs. 9–11 for ferrierite, ZSM-23 and Theta-1, respectively.
The effect of water on the stability of the ferrierite catalyst is
demonstrated in Fig. 9. In the absence of water, the catalyst was sta-
ble with time-on-stream for the isomerisation of linear butenes as
shown in the figure and in agreement with several earlier literature
reports [35,36]. In the same figure the performance is compared
to the isomerisation and iso-butene yield in the presence of water,
either by co-feeding it with the butene, or co-feeding water with the
n-butanol. As seen, when the water concentration is increased the
performance declined with time-on-stream more progressively.
This is a clear effect of the water presence and when taken together
27
with the results of NH -TPD and Al solid state NMR reported
3
above, demonstrated that the framework acidity was not stable
in a product environment containing steam. This contrasts with
the performance of Theta-1 and ZSM-23 in Figs. 10 and 11, respec-
tively, where it is clear that the presence of steam in the case of
these two catalysts acted as a diluent enhancing the performance in
the early stage of the time-on-stream. The isobutene yield from n-
butanol or 1-butene in Figs. 10 and 11 is seen to rise in the presence
of added steam (steam was added to the 1-butene after 100 min
time-on-stream).
The best catalyst performance for the conversion of n-butanol
to iso-butene in terms of yield has been found to be from catalysts
with aluminosilicate framework and 10-member ring unidirec-
tional channels. This finding is in agreement with that found for
the isomerisation of n-butenes to iso-butene [1,17–19] with the
exception of the ferrierite catalyst, which as shown, proved to be
unstable with time-on-stream.
3.5. Reaction mechanism effect on performance
The selectivity of the catalyst was highly dependent on the struc-
ture of the zeolite and its shape selectivity. For example, although
ZSM-5 demonstrated high acidity (high Aiso) and was a highly
active catalyst for the conversion of n-butanol, the isomerisation
selectivity (Siso) was relatively low because of the pore structure,
as discussed above. The unidirectional channels of both Theta-1
and ZSM-23 promote this selectivity and hence high yields of iso-
butene are obtained. The absence of intersections promoted the
mono-molecular reaction and suppressed the bi-molecular reac-
tion mechanism. The selectivity was additionally influenced by the
reaction conditions such as dilution, space velocity and tempera-
ture. The reaction conditions influenced the selectivity also via the
promotion of the mono-molecular mechanism.
Within the performances of the zeolites with unidirectional
medium pore channels, subtle differences in the selectivities are
observed. Fig. 12 shows the selectivities of iso-butene and higher
hydrocarbons as a function of linear butene conversion (Xn-butene
obtained for reaction of 1-butene or n-butanol over Theta-1, ZSM-
3 and ferrierite. As seen, the selectivity is higher at low Xn-butene
The n-butanol conversion (Fig. 12B) is more selective at lower
Xn-butene than is the case for the 1-butene conversion (Fig. 12A),
because of the extra dilution effect of the product water. Further
reaction of iso-butene proceeds via a bulky interaction and the
)
2
.
Fig. 8. Thermal gravimetric analyses of spent catalysts.

8
D. Zhang et al. / Applied Catalysis A: General 403 (2011) 1–11
Fig. 9. The effect of water with different reactants over ferrierite. Conditions: GHSV n-butanol or 1-butene = 5200 h−1; n-butanol or 1-butene/Ar(mol) = 1; H2O/Ar (mol) = 1;
◦
T = 400 C.
product is also large in volume. The marginally higher selectivity
of Theta-1 relative to ZSM-23 is probably due to the elliptically
shaped channels of Theta-1 as compared to the pear-shaped chan-
nels of ZSM-23 [37]. Fig. 13 shows the propene yield from reaction
of n-butanol (A) or 1-butene (B) as a function of the zeolite acidity as
measured by the isomerisation activity, Aiso. The low acidity of the
ferrierite in the presence of water is again clearly demonstrated,
showing low cracking resulting in low propene yield. Theta-1
shows a higher acidity function (Aiso) but lower cracking activity
than ZSM-23, most probably due to the subtle difference in the
channel shape as discussed above.
chain growth and cracking. Dehydration of butanol leaves an
adsorbed carbenium ion on the surface of the catalyst that is active
and ready to react further. The carbenium ion is initially a primary
one and can convert quickly into a secondary ion via a hydride shift,
which may donate a proton to the conjugate base on the surface of
the catalyst and thus form a 1-butene or a 2-butene. The carbenium
ion can also undergo a methyl shift and transforms into a tertiary
butyl primary carbenium ion, or after a hydride shift convert into a
stable tertiary ion. This is summarised in Scheme 2.
The most desirable mechanism is the mono-molecular pathway.
However, there are a number of publications [22] reporting skele-
tal isomerisation of butene which advocate a bi-molecular reaction
pathway via the formation of the dimers, or even oligomers, of
The product selectivity over the zeolites in general was governed
by carbenium ion chemistry involving skeletal rearrangement,
Fig. 10. The effect of water with different reactants over ZSM-23. Conditions: GHSV n-butanol or 1-butene = 5200 h−1; n-butanol or 1-butene/Ar(mol) = 1; H2O/Ar (mol) = 1;
◦
T = 400 C. For 1-butene + H2O, steam was added to the 1-butene feed after 100 min time-on-stream.

D. Zhang et al. / Applied Catalysis A: General 403 (2011) 1–11
9
Fig. 11. The effect of water with different reactants over Theta-1. Conditions: GHSV n-butanol or 1-butene = 5200 h−1; n-butanol or 1-butene/Ar(mol) = 1; H2O/Ar(mol) = 1;
◦
T = 400 C. For 1-butene + H2O, steam was added to the 1-butene feed after 100 min time-on-stream.
Table 4
a
Products distribution of n-butanol conversion over Theta-1 with different diluted ratio.
n-Butanol:Ar (mol)
Xn-butanol (%)
Yi (wt.%)b
i-C4=
Xn-butene (%)
Siso (wt.%)
Aiso
n-C4=
C3+C4
C3=+C5=
C6+
1:2
1:1
2:1
100
100
100
34.4
31.7
30.1
39.4
36.9
36.9
3.3
3.4
3.4
14
13.8
14
8.4
13.7
15.1
60.6
63.1
63.1
56.8
50.2
47.6
0.47
0.46
0.45
a
◦
−1
Reaction conditions: T = 400 C; GHSV(n-butanol) = 5200 h
Symbols as in Table 3.
.
b
butenes. The main reason for this assertion is based on considera-
tion of the instability of the primary carbenium ion resulting from
the methyl shift via the formation of a cyclopropyl carbenium ion.
The bi-molecular mechanism is more likely to occur when high
concentration of iso-butene is present. It is also the main source
of the side products, especially propene and pentenes, which are
products of the ␤-scission of the octyl carbenium ion. In Fig. 12 we
have plotted the selectivities of iso-butene and the C6+ (mostly C8=)
versus the conversion of the linear butenes. As seen the selectiv-
ity approaches 90–100% if the conversion is extrapolated to a very
low level. This suggests that the iso-butene was a primary prod-
uct whereas those of the octenes are secondary products, forming
only when the concentration of iso-butene increases. Dilution and
low pressure suppresses the formation of the octenes and pro-
Fig. 12. iso-Butene and C6+ selectivity vs. linear butene conversion (Xn-butene) with different reactant: n-butene (A), n-butanol (B). Selectivities were evaluated using Eq. (4),
−
1
◦
modified in the case of C6+. Conditions: GHSV n-butanol or 1-butene = 5200 h ; n-butanol or 1-butene/Ar(mol) = 1; T = 400 C.

1
0
D. Zhang et al. / Applied Catalysis A: General 403 (2011) 1–11
Fig. 13. Propene yield vs. iso-butene/total butenes (Aiso) with different reactant: n-butanol (A), butene + water (B). Conditions: GHSV n-butanol or 1-butene = 5200 h−1;
◦
n-butanol or 1-butene/Ar(mol) = 1; H2O/Ar (mol) = 1; T = 400 C.
motes the yield of iso-butene. Table 4 confirms this and shows
an increase in the selectivity to iso-butene production with dilu-
tion at the expense of C6+ products. The main effect of dilution is,
therefore, to reduce the dimerization of the butenes. These results
highlight the importance of the mono-molecular reaction mecha-
nism to the selectivity of iso-butene production. The dimerization
reaction is not only a source of propene and pentenes by-products,
it also consumes iso-butene as branched olefins are more likely to
form dimers than linear olefins. This trend is consistent with the
reaction network shown in Scheme 2.
conversion, and a mixture of mono- and bi-molecular mechanisms
at high conversion.
Finally, to address the issue of the primary carbenium ion stabil-
ity, one could look at the reaction conditions and the driving force
of the reaction. The distribution of the butenes was always close to
equilibrium for all the sufficiently acidic catalysts. Therefore, one
could conclude that the driving force was thermodynamic rather
than kinetic. However, the formed primary carbenium ion needs
to be stable enough kinetically to rearrange to a tertiary one, or to
eliminate the proton and produce 1-butene. For the one-step con-
version of the butanol to iso-butene, one could also propose an E2
elimination mechanism, forming an alkoxide intermediate on the
surface of the catalyst. Such an intermediate has been proposed
previously [38–40] in support of the mono-molecular mechanism.
A low level of coke formation restricts the space available for the
bi-molecular mechanism and as a result the selectivity increases as
seen in the time-on-stream trends in Figs. 4 and 7. Thus, the exper-
imental evidence points to the mono-molecular mechanism at low
H
+
+
-
H O
OH
+
OH
2
+ H
H
Zeo-
+
Zeo-
Zeo-
+
+
Zeo-
(1)
Zeo-
-
H+
-H+
+
+
+
(2)
+
+
+
-
H+
+
+
Pentenes
(3)
(4)
+
+
+
+
Aromatics
+
alkanes
+
higher olefins
Scheme 2. Reaction network of n-butanol conversion over zeolite catalysts.

D. Zhang et al. / Applied Catalysis A: General 403 (2011) 1–11
11
4
. Conclusions
[7] N. Qureshi, T.C. Ezeji, Biofuels Bioprod. Biorefining 2 (2008) 319–330.
[
8] O. Macias, J. Largo, C. Pesquera, C. Blanco, F. Gonzalez, Appl. Catal. A-Gen. 314
(
2006) 23–31.
The one-step dehydration and skeletal isomerisation of n-
[
9] M.A. Makarova, E.A. Paukshtis, J.M. Thomas, C. Williams, K.I. Zamaraev, J. Catal.
149 (1994) 36–51.
butanol to iso-butene over zeolite catalysts has been demonstrated
◦
[10] R.M. West, D.J. Braden, J.A. Dumesic, J. Catal. 262 (2009) 134–143.
11] V. Macho, M. Kralik, E. Jurecekova, J. Hudec, L. Jurecek, Appl. Catal. A-Gen. 214
2001) 251–257.
in the range of 400–500 C. The results open the possibility of a
[
process to iso-butene based on biomass derived n-butanol.
The yield of iso-butene was determined by the zeolite frame-
work structure, acidity, hydrothermal stability, and reactant partial
pressure. Aluminosilicate zeolites, Theta-1 and ZSM-23, with 10-
ring pores and unidirectional channels showed high activity for
conversion of n-butanol, high selectivity to iso-butene, and high
stability. In contrast, ZSM-5 with 10-ring pore size but intersecting
channels showed very low iso-butene selectivity. The observed low
activity for linear butene conversion over SAPO-11 and Y zeolite
is attributed to structural instability and coke formation, respec-
tively. The deactivation of ferrierite was shown to be irreversible
and due to dealumination and reduction of acid sites under the
steam environment created by dehydration of n-butanol. In con-
trast, steam acted as a diluent for Theta-1 and ZSM-23 enhancing
the selectivity and yield of iso-butene. The relationship of iso-
butene selectivity with linear butene conversion indicated that the
iso-butene was a primary product. Overall, the product distribu-
tions, effects of steam and reactant dilution are consistent with
the mono-molecular isomerisation mechanism at low conversion,
whereas both mono- and bi-molecular mechanisms operate at high
conversions.
(
[12] F.J. Maldonado-Hodar, A.F. Perez-Cadenas, J.L.G. Fierro, C. Moreno-Castilla, J.
Phys. Chem. B 107 (2003) 5003–5007.
[
13] C.L. Oyoung, R.J. Pellet, D.G. Casey, J.R. Ugolini, R.A. Sawicki, J. Catal. 151 (1995)
67–469.
4
[14] Y.W. Suh, J.W. Lee, H.K. Rhee, Appl. Catal. A-Gen. 274 (2004) 159–165.
[15] Y. Nakano, T. Iizuka, H. Hattori, K. Tanabe, J. Catal. 57 (1979) 1–10.
16] S. van Donk, J.H. Bitter, K.P. de Jong, Appl. Catal. A-Gen. 212 (2001) 97–116.
17] B. de Menorval, P. Ayrault, N.S. Gnep, M. Guisnet, Appl. Catal. A-Gen. 304 (2006)
[
[
1
–13.
[18] L. Domokos, L. Lefferts, K. Seshan, J.A. Lercher, J. Mol. Catal. A-Chem. 162 (2000)
47–157.
19] G. Onyestyak, J. Valyon, G. Pal-Borbely, L.V.C. Rees, Appl. Surf. Sci. 196 (2002)
01–407.
[20] M. Kangas, N. Kumar, E. Harlin, T. Salmi, D.Y. Murzin, Ind. Eng. Chem. Res. 47
2008) 5402–5412.
21] F. Bauer, W.H. Chen, E. Biiz, A. Freyer, V. Sauerland, S.B. Liu, J. Catal. 251 (2007)
58–270.
[22] B. de Menorval, P. Ayrault, N.S. Gnep, M. Guisnet, J. Catal. 230 (2005) 38–51.
1
[
4
(
[
2
[
[
[
23] R. Byggningsbacka, N. Kumar, L.E. Lindfors, J. Catal. 178 (1998) 611–620.
24] M. Boronat, P. Viruela, A. Corma, J. Phys. Chem. A 102 (1998) 982–989.
25] S.A.I. Barri, P. Howard, C.D. Telford, EU Pat., 0,057,049, 1982.
[26] S.A.I. Barri, GB Pat. 2,190,910, 1987.
[27] C.M. Lopez, V. Escobar, M.E. Arcos, L. De Nobrega, F. Yanez, L.V. Garcia, Catal.
Today 133 (2008) 120–128.
[
28] M.J. Remy, G. Poncelet, J. Phys. Chem. 99 (1995) 773–779.
[29] S.A.I. Barri, G.W. Smith, D. White, D. Young, Nature 312 (1984) 533–534.
[
[
30] P.A. Wright, J.M. Thomas, R.G. Millward, S. Ramdas, S.A.I. Barri, J. Chem. Soc.
Chem. Commun. 16 (1985) 1117–1119.
31] D. Breck, Zeolite molecular sieves: structure, in: Chemistry and Use, John Wiley,
Acknowledgement
1973.
We thank Dr. Rob Laws for the solid state NMR.
References
[32] N. Katada, H. Igi, J. Kim, M. Niwa, J. Phys. Chem. B 101 (1997) 5969–5977.
[
33] D.R. Stull, E.F. Westrum Jr., G.C. Sinke, The Chemical Thermodynamics of
Organic Compounds, Wiley, New York, 1989.
[34] V. Nieminen, N. Kumar, T. Heikkila, E. Laine, J. Villegas, T. Salmi, D.Y. Murzin,
Appl. Catal. A-Gen. 259 (2004) 227–234.
[
[
1] J. Houzvicka, V. Ponec, Catal. Rev. Sci. Eng. 39 (1997) 319–344.
2] J. Aittamaa, J. Jakkula, P. Lindqvist M. Koskinen, J. Linnekoski, O. Krause, M.
Sourander, J. Ignatius, A. Pyhalahti, US Pat., 6,613,108, 2003.
3] E. D. Kezel, E.P. 2105428 (A1), 2009.
[35] A.C. Butler, C.P. Nicolaides, Catal. Today 18 (1993) 443–471.
[36] W.Q. Xu, Y.G. Yin, S.L. Suib, J.C. Edwards, C.L. Oyoung, J. Phys. Chem. 99 (1995)
9443–9451.
[37] J.F. Denayer, A.R. Ocakoglu, W. Huybrechts, J.A. Martens, J.W. Thybaut, G.B.
Marin, G.V. Baron, Chem. Commun. (2003) 1880–1881.
[
[
4] European Fuel Oxygenates Association (EFOA), ETBE technical product bulletin,
2
006. Available from http://www.efoa.org/docs.
5] D. Zhang, R. Al-Hajri, S.A.I. Barri, D. Chadwick, Chem. Commun. 46 (2010)
088–4090.
6] Y. Ni, Z.H. Sun, Appl. Microbiol. Biotechnol. 83 (2009) 415–423.
[38] M. Boronat, P. Viruela, A. Corma, Phys. Chem. Chem. Phys. 3 (2001) 3235–3239.
[39] P. Meriaudeau, V.A. Tuan, N.H. Le, G. Szabo, J. Catal. 169 (1997) 397–399.
[40] V. Nieminen, M. Sierka, D.Y. Murzin, J. Sauer, J. Catal. 231 (2005) 393–404.
[
[
4