Conversion of 1-hexanol to di-n-hexyl ether on acidic catalysts

Article abstract of DOI:10.1016/j.apcata.2009.11.024

Conversion, selectivity and yield of 1-hexanol liquid phase dehydration to di-n-hexyl ether (DNHE) were determined at 150-190 °C on three acidic catalysts, the thermally stable resin Amberlyst 70, the perfluoroalkanesulfonic Nafion NR50 and the zeolite H-BEA-25, in a batch reactor. The highest conversion and yield were achieved on Amberlyst 70 at 190 °C, but the most selective catalyst was Nafion NR50. Good results were obtained at 190 °C on the zeolite. Apparent activation energies for the three catalysts were in the range 108-140 kJ/mol. Unlike H-BEA-25, the reaction of DNHE synthesis on Amberlyst 70 and NR50 was a bit more active but less selective than the analogous 1-pentanol dehydration to di-n-pentyl ether (DNPE).

Full text of DOI:10.1016/j.apcata.2009.11.024

Applied Catalysis A: General 374 (2010) 41–47
Contents lists available at ScienceDirect
Applied Catalysis A: General
journal homepage: www.elsevier.com/locate/apcata
Conversion of 1-hexanol to di-n-hexyl ether on acidic catalysts
Eduardo Medina, Roger Bringu e´ , Javier Tejero *, Montserrat Iborra, Carles Fit e´
Department of Chemical Engineering, University of Barcelona, Mart ı´ i Franqu e` s, 1, E-08028 Barcelona, Spain
A R T I C L E I N F O
A B S T R A C T
Article history:
Conversion, selectivity and yield of 1-hexanol liquid phase dehydration to di-n-hexyl ether (DNHE) were
determined at 150–190 8C on three acidic catalysts, the thermally stable resin Amberlyst 70, the
perfluoroalkanesulfonic Nafion NR50 and the zeolite H-BEA-25, in a batch reactor. The highest
conversion and yield were achieved on Amberlyst 70 at 190 8C, but the most selective catalyst was
Nafion NR50. Good results were obtained at 190 8C on the zeolite. Apparent activation energies for the
three catalysts were in the range 108–140 kJ/mol. Unlike H-BEA-25, the reaction of DNHE synthesis on
Amberlyst 70 and NR50 was a bit more active but less selective than the analogous 1-pentanol
dehydration to di-n-pentyl ether (DNPE).
Received 29 July 2009
Received in revised form 4 November 2009
Accepted 17 November 2009
Available online 22 November 2009
Keywords:
Di-n-hexyl ether
1
-Hexanol
ß 2009 Elsevier B.V. All rights reserved.
Amberlyst 70
Nafion NR50
H-BEA-25
Dehydration reaction
1
. Introduction
Oil industry is interested in increasing the cetane number of low
5
quality diesel streams, e.g. linear C olefins. The production of high
After sweeting diesel streams in order to decrease its sulfur
content, formulation of future diesel fuels will be characterized by
a higher cetane number and a lower density and aromatic content
cetane ethers from linear olefins could be a promising way, since it
offers a solution for species with a very high ozone formation
potential, such as pentenes, which should be removed from
gasolines. With this aim, 1-hexanol may be synthesized by
[
1]. Due to the huge oil demand, heavy crude amounts processed in
refineries are increasing and total quantity of such compounds
increases. Refineries have made an effort introducing hydrocrack-
ing units in order to decrease sulfur content and obtain high added
value products from heavy streams. In addition, legislation is in
continuous evolution, and a hypothetical change in cetane number
specifications would help to decrease particle matter emissions as
5
hydroformylation of 1-pentene, a feedstock of C olefin streams.
The industrial synthesis consists of selective hydroformylation and
hydrogenation of 1-pentene in the presence of Rh and Co
phosphines [6]. Afterwards, the molecular dehydration reaction
of the alcohol gives DNHE. The alcohol dehydration reaction needs
an acidic catalyst to proceed. Ion exchange resins are today the best
option in terms of selectivity and activity to achieve good linear
ether yields [7–9].
x
well as CO, NO , unburned hydrocarbons and smoke [2,3].
In order to fulfill new European specifications, a possible option
of reformulating diesel blends may be the introduction of
oxygenates in commercial blends. Previous studies concluded
that linear ethers with at least 9 carbon atoms showed high cetane
numbers and desirable cold flow properties [4].
Di-n-hexyl ether (DNHE) has been selected to carry out the
present study since itbehavesitselfas a lightdieselfuel. Boilingpoint
of DNHE is about 220 8C, close to light end of commercial boiling
point of diesel fuels (200–380 8C). Density and viscosity of DNHE are
actually lower than that of standard diesel fuels. Furthermore, since
DNHE cetane number is as high as 118, its introduction in
commercial blends would contribute to upgrade diesel properties
particularly in the light end of the boiling point curve [5].
Literature on dehydration of 1-hexanol to ether is rather scarce.
6 2
Dehydration in the gas phase of some C alcohols on a Nafion/SiO
composite and Y zeolites was studied at 200 8C ꢀ T ꢀ 300 8C [10]. It
was found that the formation of alkenes was the main reaction (2-
hexene was the predominant isomer) and DNHE synthesis was a
secondary reaction. Also in the gas phase, the dehydration of C
12 linear 1-alcohols over -alumina was studied in a fixed bed
flow reactor at 250–350 8C, 0–4 MPa, and WHSV of 1–4 h . The
5
–
C
h
ꢁ1
ꢁ1
highest ether yield was 54% at 300 8C, 1 MPa and WSHV of 1 h
[11]. The dehydration of some alcohols (including 1-hexanol) in
the liquid phase at 200 8C over montmorillonite showed that
primary alcohols underwent preferential intermolecular dehydra-
tion to ethers while secondary alcohols favored the formation of
alkenes via intramolecular dehydration [12]. Finally, in the liquid
phase synthesis of DNHE over Nafion-H yields of 95% in overnight
experiments at 145–150 8C were found [13].
*
Corresponding author. Tel.: +34 93 402 1308; fax: +34 93 402 1291.
E-mail address: jtejero@ub.edu (J. Tejero).
0926-860X/$ – see front matter ß 2009 Elsevier B.V. All rights reserved.
doi:10.1016/j.apcata.2009.11.024

4
2
E. Medina et al. / Applied Catalysis A: General 374 (2010) 41–47
Nomenclature
d
pore
ap
pore diameter (nm)
E
apparent activation energy (kJ/mol)
number DNHE moles
n
DNHE
ꢁ
1
ꢁ1
)
r
r
DNHE
0
DNHE
reaction rate of DNHE synthesis (mol h kg
initial reaction rate of DNHE synthesis
ꢁ
1
ꢁ1
)
(
mol h kg
S
S
t
selectivity (%)
2
ꢁ1
)
g
surface area (m g
time (h)
T
temperature (K)
3
ꢁ1
)
Fig. 1. Distribution of bead size of Amberlyst 70 in air, 1-hexanol, DNHE and
V
g
pore volume (cm g
water.
V
sp
specific volume of the swollen polymer phase
3
ꢁ1
(cm g
)
alcohol supplied by Fluka (ꢂ98%). Di-n-hexyl ether was obtained in
our lab and purified to ꢂ98%. 1-Hexene (ꢂ99%) from Aldrich, trans-
W
weight of dry catalyst (g)
conversion of 1-hexanol (%)
DNHE yield
X
HeOH
2
-hexene (ꢂ98%), cis-2-hexene (ꢂ95%), trans-3-hexene (ꢂ97%),
Y
DNHE
cis-3-hexene (ꢂ95%), and 2-methyl-1-pentanol (ꢂ99%) were
supplied by Fluka and used for analysis purposes.
Greek letters
Three acidic catalysts were tested. Amberlyst 70 (A70), a
chlorinated low crosslinked polystyrene-divinylbenzene (PS-DVB)
copolymer from Rohm and Haas France, with an acidic capacity of
3
r
u
s
skeletal density (g/cm )
porosity
+
3
3.01 meq H /g, and skeletal density of 1.52 g/cm , mean bead
diameter of 570 m, and a maximum operating temperature of
m
On the other hand, acidic ion exchange resins have shown to be
200 8C [17]. Some structural properties are given in Table 1.
Amberlyst 70 is a macroporous resin flexible enough to accom-
modate to aqueous media because of its low crosslinking degree.
This fact is stated in the determination of bead size made in water,
DNHE, 1-hexanol and air by a laser technique with a Microtrack
SRA analyzer. As Fig. 1 shows, beads swell clearly in water, but
hardly swell in alcohol and ether. The same fact is observed by
Inverse Steric Exclusion Chromatography (ISEC) structural mea-
surements in water (Table 1) [18].
highly selective to the ether formation from 1-alkanol, avoiding
secondary reactions, e.g. dehydration to olefin. Recently, the
thermally stable ion exchange resin Amberlyst 70 based on a
styrene-divinylbenzene copolymer has been commercialized. It
has shown to be an effective catalyst in 1-pentanol dehydration to
di-n-pentyl ether (DNPE) [14,15]. Actually, compared to Nafion
NR50, Amberlyst 70 was more active than Nafion catalyst, but only
a bit less selective.
We have studied recently the chemical equilibrium of liquid
phase dehydration of 1-hexanol to di-n-hexyl ether on Amberlyst
Nafion NR50 (NR50) is a perfluoroalkane sulfonic resin well
described in the literature [14,15,19], with an acidic capacity of
0.81 meq H /g, and skeletal density of 2.042 g/cm , mean bead
+
3
7
0 [16]. The present work is devoted to study the catalyst activity
of Amberlyst 70 in the 1-hexanol conversion to DNHE. In addition,
experiments on Nafion NR50 and H-BEA-25 zeolite have been also
performed in order to compare the behavior of the three catalysts.
Finally, the syntheses of DNHE and DNPE on these solids, in the
same experimental set-up, are compared.
diameter of 2350 mm and a maximum operating temperature of
220 8C.
Finally, the zeolite H-BEA-25 (BEA25) with a SiO /Al O ratio of
2
2 3
+
25.1, an acidic capacity of 1.2 meq H /g, and skeletal density of
3
2.237 g/cm and a mean bead diameter of 8.1
mm. An exhaustive
report of H-BEA-25 properties as well as the pore distribution
curve is available elsewhere [15,20].
2
. Experimental
.1. Materials
-Hexanol (99.5% pure, <0.3% 2-methyl-1-pentanol, 0.1%
2
2.2. Experimental set-up
1
Experiments were carried out in a 100 mL stainless steel auto-
clave which operated in batch mode, stirred by means of a
water) was used after purification in a distillation column of the
Table 1
Structural properties of Amberlyst 70.
a
Dry state
b
2
c
g
3
d
e
S
g
(m /g)
V
(cm /g)
d
pore (nm)
u (%)
3
0
0.153
19,3
18.8
Swollen in water (ISEC method)
2
3
3
d
e
S
g
(m /g)
V
g
(cm /g)
V
sp (cm /g)
d
pore (nm)
u
(%)
1
76
0.355
1.189
1.19
57.4
a
b
c
Dried by successive percolation with methanol, toluene and isooctane.
BET method.
Determined by adsorption–desorption of N
2
at 77 K.
d
e
Assuming pore cylindrical model.
In dry state
u
= 100 V
g
/(V
g
+ (1/r
s
)). Swollen in water,
u
= 100(V
g
+ Vspꢁ(1/
s g
r ))/(V + Vsp).

E. Medina et al. / Applied Catalysis A: General 374 (2010) 41–47
43
magnetic drive turbine. Temperature was controlled to within
obtain the variation in concentration over time of all com-
pounds. The assayed temperatures were selected in the range
150–190 8C.
ꢃ1 8C by an electric furnace. To carry out the reaction in the liquid
2
phase pressure was set at 2.1 MPa using N as inert gas. 0.2 ml
samples were taken from the reactor through a liquid sampling valve
directly connected to a GLC apparatus.
In each experiment, 1-hexanol conversion (XHeOH), selectivity to
DNHE (SDNHE), and similarly to alkenes (Salkenes) and to branched
ethers (Sethers), and yield of DNHE with respect 1-hexanol (YDNHE
)
2.3. Analysis
were calculated by the expressions:
The composition of liquid mixtures was analyzed by using a
split-mode operation in a HP6890A GLC apparatus equipped with
TCD, which allowed quantifying also water as a reaction product. A
mole of 1-hexanol reacted
X
HeOH
¼
(1)
(2)
initial mole of 1-hexanol
5
0 m ꢄ 0.2 mm ꢄ 0.5
mm methyl silicone capillary column was
temperature programmed with a 6 min initial hold at 45 8C,
followed by a 30 8C/min ramp up to 180 8C and holding for 10 min.
Helium was used as a carrier gas at a total flow rate of 30 ml/min. It
mole of 1-hexanol reacted to form DNHE
mole of 1-hexanol reacted
S
DNHE
¼
¼
allowed to separate 1-hexanol, DNHE, water and by-products: C
olefins (1-hexene, 2-hexene and 3-hexene) and branched ethers
2,2-oxybis hexane and 1,2-oxybis hexane). Identification of by-
6
mole of 1-hexanol reacted to form DNHE
initial mole of 1-hexanol
Y
DNHE
(
products was carried out with an additional GLC apparatus
equipped with a MS detector, where samples were injected
manually.
¼ XHeOH
S
DNPE
(3)
Reaction rates of DNHE formation at any time were calculated
from the function of variation of nDNHE (number of DNHE moles
2.4. Methodology
produced) versus time:
ꢀ
ꢁ ꢂ
ꢃ
1
dnDNHE
mol DNHE
kg h
Amberlyst 70 and NR50 were dried at 110 8C in an oven,
firstly at atmospheric pressure during 15 h, and then 2 h
under vacuum. H-BEA-25 was activated by calcination at
r
DNHE
¼
(4)
W
dt
t
Finally, the turnover frequency (TOF) can be computed by
dividing the reaction rate by the acidic capacity:
5
00 8C for 3 h under air stream in a muffle furnace and kept
under vacuum overnight. Dried catalyst (1 g) and 70 ml of 1-
hexanol were charged into the reactor and when pressure
achieved 2.1 MPa leaking problems were checked. Then, the
reactor was heated until the desired reaction temperature was
reached. This moment was considered as the zero time of
experiment. For 6 h, liquid samples were analyzed hourly to
ꢂ
ꢃ
r
DNHE
þ
mol DNHE
TOF ¼
(5)
þ
eq H =kg eq H
h
3. Results and discussion
.1. Description of an experiment
3
Fig. 2 shows a typical plot of DNHE and by-products mole
evolution on Amberlyst 70 over an experiment conducted at
90 8C. After a heating period of about 20 min, the alcohol
1
conversion was about 3.5%. From this time, the reaction
proceeds smoothly, DNHE being the main product. By-products
appeared as soon as the reaction begins, and their amount
increased continuously through the experiment. Detected by-
products were C
trans-3-hexene),
6
olefins (1-hexene, trans- and cis-2-hexene, cis/
alcohols (2- and 3-hexanol) and C
12
C
6
branched ethers (1,2-oxybis hexane and 2,2-oxybis hexane). It
is to be noted that the amount of water in the liquid phase was
lower than that of DNHE despite that reactions of dehydration to
DNHE and olefins release water. This can be explained by the
preferential adsorption of water on the resin, as seen in Fig. 1,
6
and by the formation of by-products such as C alcohols other
than 1-hexanol. As for by-products, olefins are the main ones. As
Fig. 2 shows, 2-hexene is the most favored olefin, the quantities
of 3-hexene and 1-hexene being much smaller. Branched ethers
and, especially, 2- and 3-hexanols appear in a much lesser
quantity.
Fig. 3 shows the evolution of 1-hexanol conversion and
selectivity with time over the experiment. XHeOH increased with
time as expected, with a slow rise in the experiment ending by the
influence of the reverse reaction. SDNHE decreased with time,
whereas selectivity to olefins and branched ethers increased but
very slowly. Selectivity to alcohols other than 1-hexanol was very
low, and it is not shown in Fig. 3.
Fig. 2. Variation of composition of the reaction medium with time at 190 8C on 1 g of
Amberlyst 70 and 500 rpm (70 ml 1-hexanol). Up: 1-pentanol, DNHE, and water;
down: by-products.
Based on the evolution of the liquid phase composition, the
following reaction scheme could be proposed:

4
4
E. Medina et al. / Applied Catalysis A: General 374 (2010) 41–47
(
1) Dehydration of 1-hexanol to DNHE is the main reaction.
(
(
2) Dehydration to 1-hexene is the main side-reaction,
50–80 8C with methanol [22]. Consequently, reaction between
1
1
-hexene and 1-hexanol could contribute to the formation of
,2 oxybis hexane since: (1) working temperature is substan-
tially higher in our study, and (2) hexenes concentration in the
liquid phase is much higher than that of 2-hexanol. It is to be
considered that, in the same way that 2-pentanol dehydration
on acidic resins yields preferably olefins to the corresponding
ether [23], it is likely that 2-hexanol gives place to hexenes as
liquid phase composition suggests.
3) 1-Hexene isomerizes to 2-hexene (cis and trans), and 3-hexene
(
cis and trans).
(
(
4) Alkenes may react with water giving place to C
and 3-hexanol).
5) The reaction between the appropriate pair of alcohols yields
branched ethers: 2-hexanol originates 2,2-oxybis hexane; 2-
hexanol and 1-hexanol react to give 1,2-oxybis hexane.
6
alcohols (2-
3
.2. Synthesis of DNHE: effect of temperature
Preliminary experiments were performed to check whether
measured reaction rates were free of mass transfer effects over
Amberlyst 70 at 190 8C [24], the higher temperature of the range
(
6) Finally, the reaction between the corresponding pair of C
6
explored. In this way, the effect of the amount of catalyst was
checked by changing the mass of dried resin from 0.5 to 5 g. It was
observed that the initial reaction rate did not change when the
olefin and alcohol could also give branched ethers, i.e. 1-hexene
reacting with 1-hexanol to give 1,2-oxybis hexane.
Literature on etherification of
shows that they hardly react on acidic resins as compared to
tertiary olefins. Still, literature also shows that on specially
a
-olefins with shorter alcohols
catalyst load was lower than 4 g. Next, the influence of the external
mass transfer was tested in a series of experiments with stirring
speeds of 50–800 rpm. It was found that over 200 rpm initial
reaction rates were the same within the limits of the experimental
error. Finally, to evaluate the effect of the catalyst size on the
reaction rate, experiments were performed over sieved batches of
Amberlyst 70. No effect of diffusion was observed with batches
favorable conditions
a
-olefins can be etherified, i.e. de Klerk
cut containing 86.6% of 1-
[
21] reported that treating a C
6
hexene at 69 8C with a methanol/tert-olefin initial molar ratio
of 0.3 a 1-hexene conversion of 8% was observed. sec-Butyl
methyl ether formation as by-product of MTBE synthesis has
between 450 and 670
mm. Thus, further experiments were
been also reported when a C
4
cut containing isobutene (40–
performed by using 1 g of dried catalyst with commercial
distribution of particle sizes, and stirring speed of 500 rpm. It
was assumed that obtained results on NR50 and H-BEA-25 were
free of mass transfer influence as it was shown in the study of 1-
pentanol dehydration to di-n-pentyl ether conducted in the same
set-up [15].
5
5 wt %), 1-butene (20–30), 2-butene (8–18) was etherified at
To test the effect of temperature, a series of experiments was
performed on Amberlyst 70 and NR50 at the temperature range of
1
50–190 8C, and on H-BEA-25 at 160–190 8C. Table 2 shows 1-
hexanol conversions, selectivities to DNHE and by-products,
yields to DNHE after 6 h of reaction, as well as initial reaction rates
of ether synthesis. As expected, 1-hexanol conversion increased
with temperature for all tested catalysts, Amberlyst 70 being the
most active due to its higher acidic capacity. Despite acid capacity
of NR50 is lower than that of H-BEA-25, NR50 was more active
because of its higher acid strength [15]. Although the high
conversion achieved after 6 h at 190 8C with Amberlyst 70 (ꢅ71%),
reaction system was still far from chemical equilibrium, as seen in
preliminary experiments with 5 g of resin where 1-hexanol
Fig. 3. Hexanol conversion (XHexanol) and selectivity to DNHE (SDNHE), branched
6
ethers (Sethers) and C olefins (Salkenes) at 190 8C (1 g Amberlyst 70, 500 rpm, 70 ml 1-
hexanol).

E. Medina et al. / Applied Catalysis A: General 374 (2010) 41–47
45
Table 2
Conversion, selectivity, ether yield after 6 h of reaction and initial reaction rate of dehydration of 1-hexanol to DNHE on Amberlyst 70, NR50 and H-BEA-25 (Wcat = 1 g, 70 mL 1-
hexanol, 500 rpm).
0
T (8C)
X
HeOH (%)
SDNHE (%)
Y
DNHE (%)
Salkenes (%)
S
ethers (%)
r
DNHE (mol/h kg)
A-70 NR50 BEA25 A-70 NR50 BEA25 A-70 NR50 BEA25 A-70 NR50 BEA25 A-70 NR50 BEA25 A-70
NR50 BEA25
4.7
150
160
170
180
190
16.5
29.7
47.6
63.7
70.9
10.3
18.4
35.6
53.9
66.1
97.7
96.2
94.3
91.1
86.9
97.9
98.0
97.9
96.9
93.4
16.1
28.6
44.9
58.0
61.7
10.0
18.0
34.8
52.2
61.7
1.5
2.4
3.5
5.9
9.4
1.5
1.2
1.2
1.9
4.7
0.8
1.4
2.1
3.0
3.6
0.6
0.8
0.9
1.2
2.0
10.7
18.7
6.7
15.1
31.2
53.8
92.2
93.0
90.0
88.8
6.1
14.0
28.0
47.8
2.8
0.3
4.8
6.8
5.0
6.7
5.2
4.4
9.4
1.9 ꢃ 0.1
5.3 ꢃ 0.3
15 ꢃ 3
32.3 19.2
73.3 36.1
151
89.7
26 ꢃ 1
conversions above 90% were found. Moreover, from liquid phase
equilibrium constants of synthesis of DNHE [16] an equilibrium
conversion of 1-hexanol of about 93% has been estimated at
reaction rate was observed for Amberlyst 70 and in lower extent
for NR50 on increasing XHeOH. This decrease can be ascribed
primarily to the effect of reverse reaction, since the formation
DNHE and water decreases the driving force of the reaction. The
presence of increasing amounts of water as reaction proceeds can
also play a role: water is preferably adsorbed on the resin and
1
90 8C.
On the other hand, SDNHE decreased on increasing temperature
due to the formation of C alkenes, which was the main side-
6
reaction; although some branched ethers were also detected. It is
likely that branched ethers were formed by reaction between
swells the catalyst (Fig. 1), favoring diffusion of 1-hexanol and
ꢁ
accessibility to HSO
3
groups. Nevertheless, some inhibiting effect
olefins and alcohols, since the detected amounts of C
6
alcohols
because of preferential adsorption of water on sulfonic groups
diminishing the actual number of active sites [25], and even by a
other than 1-hexanol were very low. Moreover, Salkenes and Sethers
increased on decreasing SDNHE. The decrease in SDNHE on rising the
temperature was a bit higher on Amberlyst 70 than on NR50 and H-
BEA-25, in particular at 190 8C. NR50 was the most selective
catalyst in the whole temperature range, followed by Amberlyst 70
at 150–180 8C, and H-BEA-25 at 190 8C.
Despite the decrease on the SDNHE, DNHE yields increased with
temperature, reaching a value of 61.7% after 6 h on Amberlyst 70
and NR50 at 190 8C. In the entire range of temperature, the highest
DNHE yields were obtained on Amberlyst 70, but at 190 8C DNHE
yields on NR50 and Amberlyst 70 are the same. At present working
conditions, the upgrade in ether yield observed on increasing the
temperature is due to higher 1-hexanol conversions which balance
the slight decrease in selectivity. Thus NR50 could be a good option
for industrial operation at 190 8C, whereas Amberlyst 70 is the best
option when the whole temperature range is considered.
reaction mechanism change from a concerted one (the true
ꢁ
catalytic species is the HSO
3
group) to an ionic one (the true
+
catalytic species is the solvated H ) could take place [26]. NR50
shows a similar behavior. Instead, reaction rates decrease slightly
on H-BEA-25 with XHeOH, maintaining the activity level during the
whole experiment, probably because this catalyst does not swell
on adsorbing water, in such a way that the observed rate decrease
is only due to the effect of reverse reaction.
Fig. 5 shows the variation of turnover frequency (TOF) with
X
HeOH at 190 8C. NR50 showed the highest TOF values. This
behavior was expected since fluorine atoms confer NR50 superacid
) of 11
characteristics, with a Hammett acidity function (ꢁH
0
(Table 3). Table 3 also shows the Hammett acidity functions of
Amberlyst 35 and H-BEA-25. The acid strength of Amberlyst 70 is
assumed to be the same as Amberlyst 35, as shown by ammonia
adsorption flow calorimetry tests carried out by Siril et al. [27]. At
low XHeOH values, TOFs of Amberlyst 70 are higher than those of H-
As Table 2 shows, initial reaction rates of DNHE synthesis
strongly increase with temperature and, again, the highest value
was achieved at 190 8C with Amberlyst 70, followed by NR50. Fig. 4
plots the reaction rate versus 1-hexanol conversion for the three
catalysts at 190 8C. Amberlyst 70 was the more active at XHeOH
BEA-25, which correlates with its slightly higher values of (ꢁH
0
).
But as reaction pass by TOFs of NR50 and Amberlyst 70 decrease
dramatically whereas those of H-BEA-25 only decreases slightly.
TOF decrease of NR50 and Amberlyst 70 can be explained by the
influence of reverse reaction and the inhibitor effect of water.
Moreover, the fact that acid strength of both catalysts decreases
when the content of water of the liquid phase increases (Table 3)
could additionally contribute to TOF decrease. On the other hand,
ꢀ
40%. At higher conversions, rates on Amberlyst 70 and NR50 were
the same within the limits of the experimental error. At low XHeOH
values, reaction rate on Amberlyst 70 is the highest since it has
the highest acid capacity, followed by NR50 since it has by far the
highest acid strength, and finally H-BEA-25. A sharp decrease in the
Fig. 4. Reaction rate versus 1-hexanol conversion for the assayed catalysts at 190 8C
Fig. 5. TOF versus 1-hexanol conversion for the assayed catalysts at 190 8C (1 g of
(1 g of dried catalyst, 500 rpm, 70 ml 1-hexanol).
dried catalyst, 500 rpm, 70 ml 1-hexanol).

4
6
E. Medina et al. / Applied Catalysis A: General 374 (2010) 41–47
Table 3
bit higher in the hexanol/DNHE system. Instead, on H-BEA-25
Hammett acidity function (ꢁH
0
) in aqueous and non-aqueous conditions [14,15].
higher ether yields are observed in the pentanol/DNPE system. In
this case, both alcohol conversions and selectivities to ether were
slightly higher in DNPE synthesis. Accordingly, selectivities to
olefins and branched ethers are slightly lower in the dehydration
reaction of 1-pentanol to DNPE.
Catalyst
Aqueous
Non-aqueous
a
A35
2.65
8.4
5.6
11
NR50
BEA 25
4.4 ꢀ ꢁH ꢀ 5.7
0
a
Initial reaction rates show the same trend as conversions so that
they were higher in the 1-hexanol/DNHE system on Amberlyst 70
and NR50 in the whole temperature range. On the contrary, higher
initial reaction rates were observed in the 1-pentanol/DNHE
system on H-BEA-25. It seems that swelling of both polymeric
resins due to water released allows accommodate both reaction
intermediates reasonably well. The fact that the longest chain
alcohol, 1-hexanol, is more acid than 1-pentanol could also help to
explain the higher dehydration reaction for DNHE synthesis on
resins [28,29]. As for H-BEA-25, observed activity data could result
from the fact that 1-hexanol is slightly bulkier than 1-pentanol.
In Table 4 the activation energies for both reaction systems are
shown. It should be noted that Eap,DNHE values were estimated
Amberlyst 35 and Amberlyst 70 have similar acid strength [27].
directly from initial reaction rates, whereas
Eap,DNPE were
computed from a kinetic model in which adsorption–desorption
processes were taken into account. Notwithstanding that, activa-
tion energies for both systems are similar, particularly, for
Amberlyst 70 and NR50. However, the highest value for H-BEA-
Fig. 6. Arrhenius plot of initial reaction rates for the assayed catalysts.
25 in DNHE synthesis suggests that this reaction is a bit more
sensitive to temperature on the zeolite than DNPE synthesis.
Table 4
Apparent activation energies (kJ/mol) for DNHE (this work) and DNPE [15]
synthesis reactions.
4. Conclusions
Catalyst
E
ap,DNHE
E
ap,DNPE
The thermally stable catalysts Amberlyst 70, Nafion NR50 and
zeolite H-BEA-25 have proven to be good catalysts to obtain di-n-
hexyl ether. The best results were obtained with Amberlyst 70, but
the most selective catalyst was Nafion NR50. In the reaction
conditions, the best ether yields were obtained at the highest
temperature because the gain in alcohol conversions outweighs
the decrease in selectivity to ether. Apparent activation energies
for DNHE synthesis were in the range 108–148 kJ/mol, showing a
moderate temperature dependence of reaction rate. Compared
with the related synthesis of DNPE it is seen that DNHE yields on
Amberlyst 70 and NR50 are slightly better than DNPE ones, despite
that selectivity to DNHE is a bit lesser than that to DNPE. On the
contrary, on H-BEA-25 slightly better yields in DNPE are found.
A70
108 ꢃ 7
118 ꢃ 6
148 ꢃ 11
115 ꢃ 5
109 ꢃ 3
121 ꢃ 2
NR50
BEA 25
the slight decrease of TOF for H-BEA-25 suggests that only the
reverse reaction has a significant role. Finally, the fact that at low
HeOH values TOF of H-BEA-25 was nearly a 40% of that of
Amberlyst 70, despite that in non-aqueous media their acid
strengths are similar, agrees with the assumption that only
mesopore surface (about a 40% of total surface area) is accessible to
X
1
-pentanol and larger alcohols [15].
Arrhenius plots of initial reaction rates are straight lines for the
three catalysts (Fig. 6). Thus, neither NR50 nor H-BEA-25 show
diffusion concerns in the temperature range studied. From
Arrhenius plots apparent activation energies for DNHE synthesis,
Acknowledgments
Financial support was provided by the State Education,
Universities, Research & Development Office of Spain (project
CTQ2004-01729/PPQ). The authors thank Rohm and Haas France
for providing A-70 thermally ion exchange resin.
E
ap, were estimated (Table 4). Similar values were obtained for
Amberlyst 70 and NR50, while a slightly higher value was
computed for H-BEA-25, showing that the reaction is a bit more
sensitive to temperature on the zeolite.
3.3. Comparison of DNPE and DNHE formation reactions
References
[
[
1] A. Douaud, Hydrocarb. Process. 74 (2) (1995) 55–58.
2] F. Giavazzi, D. Terna, D. Patrini, F. Ancillotti, G.C. Pecci, R. Trer e` , M. Benelli, IX
International Symposium on Alcohol Fuels, vol. 1, 1991, 327–335.
It is interesting to compare the behavior of the tested catalysts
in the related synthesis of DNPE and DNHE, since both ethers have
been proposed as options to reformulate diesel fuels. Experiments
concerning the dehydration of 1-pentanol to DNPE and water were
performed in the same set-up, using the same amount and type of
catalysts and under the same conditions of operation free of mass
transfer influence [15].
The dehydration of 1-hexanol to DNHE on Amberlyst 70 and
NR50 is slightly faster than the dehydration of 1-pentanol to DNPE.
On both catalysts, 1-hexanol conversions are higher than those of
[3] J. Van Heerden, J.J. Botha, P.N.J. Roets, XII International Symposium on Alcohol
Fuels, vol. 1, Tsinghua Univ. Press, Beijing, 1998, pp. 188–199.
[4] G.C. Pecci, M.G. Clerici, M.G. Giavazzi, F. Ancillotti, M. Marchionna, R. Patrini, X
International Symposium on Alcohol Fuels, vol. 1, 1991, 321–335.
[5] O.A. Olah, US Patent 5.520.710 (1996).
[6] J. Hagen, Industrial Catalysis: A Practical Approach, 1st ed., Wiley-VCH, 1999.
[
[
[
7] R. Patrini, M. Marchionna, UK Patent GB 2.323.844 A (1998).
8] O.N. Karpov, R.M. Bistroya, L.G. Fedoysuk, Zh. Prikl. Khim. 40 (1967) 219–223.
9] E. Swistak, P. Mastagli, Comptes Rend. 239 (1954) 709–711.
[10] C. Park, M.A. Keane, J. Mol. Catal. A: Chem. 166 (2001) 303–322.
[
[
11] R.J.J. Nel, A. de Klerk, Ind. Eng. Chem. Res. 48 (2009) 5230–5238.
12] J.A. Ballantine, M. Davies, I. Patel, J.H. Purnell, M. Rayanakorn, K.J. Williams, J. Mol.
Catal. 26 (1984) 37–56.
1
-pentanol, so that despite that selectivity to DNHE is slightly
lower that selectivity to DNPE measured DNHE yields are slightly
better. Moreover, selectivity to alkenes and branched ethers was a
[13] G.A. Olah, T. Shamma, G.K. Surya Prakash, Catal. Lett. 46 (1997) 1–4.

E. Medina et al. / Applied Catalysis A: General 374 (2010) 41–47
47
[
[
[
14] J. Tejero, F. Cunill, M. Iborra, J.F. Izquierdo, C. Fit e´ , J. Mol. Catal. A: Chem. 182–183
2002) 541–554.
15] R. Bringu e´ , J. Tejero, M. Iborra, J.F. Izquierdo, C. Fit e´ , F. Cunill, J. Catal. 244 (2006)
3–42.
16] R. Bringu e´ , J. Tejero, M. Iborra, C. Fit e´ , J.F. Izquierdo, F. Cunill, J. Chem. Eng. Data 53
2008) 2854–2860.
17] J.R. Collin, D. Ramprasad, European Patent EP 1 479 665 (2004).
18] K. Jerabek, ACS Symp. Ser. 635 (1996) 211–220.
19] Q. Deng, Y. Hu, R.B. Moore, C.L. McCormick, K.A. Mauritz, Chem. Mater. 9 (1997)
[21] A. de Klerk, Ind. Eng. Chem. Res. 43 (2004) 6349–6354.
[22] M. Vila, F. Cunill, J.F. Izquierdo, J. Gonz a´ lez, A. Hern a´ ndez, Appl. Catal. A: Gen. 117
(1994) L99–L108.
(
3
[23] R. Bringu e´ , Ph.D. Thesis, University of Barcelona, 2006.
[24] E. Medina, MS Chemical Engineering Thesis, University of Barcelona, 2007.
[25] R. Bringu e´ , J. Tejero, M. Iborra, J.F. Izquierdo, C. Fit e´ , F. Cunill, Top. Catal. 45 (2007)
181–186.
[26] B.C. Gates, W. Rodr ı´ guez, J. Catal. 31 (1973) 27.
[27] P.F. Siril, H.E. Cross, D.R. Brown, J. Mol. Catal. A: Chem. 279 (2008) 63–68.
[28] F. Ancillotti, M. Massi Mauri, E. Pescarollo, J. Catal. 46 (1977) 49–57.
[29] F. Ancillotti, M. Massi mauri, E. Pescarollo, L. Romagnoni, J. Mol. Catal. 4 (1978)
37–48.
(
[
[
[
3
6–44.
20] M. Iborra, J. Tejero, C. Fit e´ , J.F. Izquierdo, F. Cunill, J. Catal. 231 (2005)
7–91.
[
7