Synthesis of ethyl hexyl ether over acidic ion-exchange resins for cleaner diesel fuel

Article abstract of DOI:10.1039/c4cy01548g

The synthesis of ethyl hexyl ether as a suitable diesel additive was investigated using 1-hexanol and diethyl carbonate as reactants and acidic ion-exchange resins as catalysts. Liquid-phase experiments were performed in a batch reactor at the temperature range of 403-463 K and 2.5 MPa. The formation of ethyl hexyl ether proceeded from two routes: thermal decomposition of ethyl hexyl carbonate and intermolecular dehydration of 1-hexanol with ethanol. Both pathways require a previous transesterification reaction between diethyl carbonate and 1-hexanol. It was revealed that this reaction is favoured in polymer zones of 0.4 nm nm^-3 polymer density (equivalent to 2.6 nm diameter pores in inorganic materials). Acidic ion-exchange resins containing a significant volume fraction of this polymer density are Dowex 50W×2 and Amberlyst 70. By using this kind of catalyst, reaction rate and selectivity are significantly increased. Finally, it was observed that working at low temperature would favour the selectivity to ethyl hexyl carbonate and hinder the undesired formation of alkenes. This journal is

Full text of DOI:10.1039/c4cy01548g

Catalysis
Science &
Technology
View Article Online
View Journal | View Issue
PAPER
Synthesis of ethyl hexyl ether over acidic
ion-exchange resins for cleaner diesel fuel
Cite this: Catal. Sci. Technol., 2015,
5, 2238
*
J. Guilera, E. Ramírez, C. Fité, J. Tejero and F. Cunill
The synthesis of ethyl hexyl ether as a suitable diesel additive was investigated using 1-hexanol and diethyl
carbonate as reactants and acidic ion-exchange resins as catalysts. Liquid-phase experiments were
performed in a batch reactor at the temperature range of 403–463 K and 2.5 MPa. The formation of ethyl
hexyl ether proceeded from two routes: thermal decomposition of ethyl hexyl carbonate and inter-
molecular dehydration of 1-hexanol with ethanol. Both pathways require a previous transesterification
reaction between diethyl carbonate and 1-hexanol. It was revealed that this reaction is favoured in polymer
zones of 0.4 nm nm−3 polymer density (equivalent to 2.6 nm diameter pores in inorganic materials). Acidic
ion-exchange resins containing a significant volume fraction of this polymer density are Dowex 50W×2
and Amberlyst 70. By using this kind of catalyst, reaction rate and selectivity are significantly increased.
Finally, it was observed that working at low temperature would favour the selectivity to ethyl hexyl carbonate
and hinder the undesired formation of alkenes.
Received 25th November 2014,
Accepted 16th January 2015
DOI: 10.1039/c4cy01548g
www.rsc.org/catalysis
group can be renewable bioethanol.13–17 Ethyl ethers can be
synthesized directly from the dehydration reaction of two
1. Introduction
Complete combustion of diesel fuel only produces carbon
dioxide and water, but actual combustion in diesel engines is
incomplete and forms toxic air contaminants. In particular,
diesel vehicles are the largest worldwide contributors to
particulate matter pollution in urban areas, where most
people live. Compared to similar-sized gasoline engines,
diesel combustion generates about 100-fold more particles for
travelled distance. These particles form deposits in lungs with
what an increased risk of premature mortality ensues.1–4 Far
from being undervalued, Yim and Barret estimated that road
pollution in the UK is twice as deadly as traffic accidents.5
A feasible technical solution able to reduce the amount of
particulates in diesel exhaust, with little or no penalty on
nitrogen oxide emissions, is the use of diesel blends
containing oxygenated compounds, commercially known as
oxygenated diesel.6–9 Traditional oxygenates used in gasoline
formulation are unsuitable for incorporation into diesel fuel
blends due to their low cetane numbers, leading to dimin-
ished fuel performance. In contrast to the branched ethers
currently used in gasoline blends (e.g. methyl and ethyl
tert-butyl ether), linear ethers are preferred in the case of
diesel fuel for cleaner fuel burning purposes.10–12
ethanol molecules, or else, by the reaction between ethanol
with a longer alcohol. From an industrial perspective, both
options have their pros and cons, and actually they are per-
fectly compatible. The first option forms diethyl ether, which
is the most economical ethyl ether compound, but as a result
of its low flash point, it changes considerably the initial
diesel properties. Thus, the addition of large quantities of
diethyl ether to diesel should be avoided. A second option,
more expensive due to the lower availability of longer
alcohols, forms asymmetrical linear ethers with excellent
properties as diesel fuels. As a result, it is not necessary to
restrict their addition from a technical standpoint but it
should be optimized from an economical one. Ethyl hexyl
ether is an ethyl ether of the latter group that provides a good
compromise between blending cetane number and cold flow
properties.18,19
The literature dealing with ethyl hexyl ether synthesis is
scarce. However, on the analogy of similar ether syntheses,
there are several potential organic routes able to produce
industrial amounts of ethyl hexyl ether, all consisting in the
ethylation of 1-hexanol. Three cost-effective ethylating agents
showed interesting performance in analogous reactions: etha-
nol, diethyl carbonate and diethyl ether. Among these three
environmentally benign ethylating agents, diethyl carbonate
is the most reactive one and gives rise to successful conver-
sion at lower temperature.20–22
Among linear ethers, some attention has been focused on
ethers containing ethyl groups as the origin of the ethyl
The chance of decreasing the reaction temperature in the
formation of ethyl hexyl ether is especially relevant to hinder
Chemical Engineering Department, Faculty of Chemistry, University of Barcelona,
c/Martí i Franquès 1, 08028 Barcelona, Spain. E-mail: jtejero@ub.edu
2238 | Catal. Sci. Technol., 2015, 5, 2238–2250
This journal is © The Royal Society of Chemistry 2015

View Article Online
Catalysis Science & Technology
Paper
alkene formation, in the present case, C6 alkenes.23 The
formation of alkenes, by means of intramolecular dehydration
of an alcohol (reactant) or via cracking of ether (product),
has higher dependency on the temperature than the desired
production of ethers. In this sense, at temperatures above
523 K the undesired formation of alkenes becomes the
dominant process.20,24,25 Therefore, reactor temperature
should be kept below 523 K. At this relatively low temperature,
acidic ion-exchange resins are interesting candidates to be
used as catalysts because they exhibit higher concentration of
acid sites (≈5 mmol H+ g−1) compared to most other solid
acids. Thus, even though the strength of the resin acid sites
tends to be lower than those found in zeolites and similar solid
acids, they are typically more active at lower temperatures.26–29
In line with this, the production of the ethers used in gasoline
blends is nowadays carried out on acidic ion-exchange resins
on the industrial scale.
Catalytic performance of different types of acidic ion-
exchange resins can be extremely different because the
exchange capacity of acidic resins is chiefly conditioned by
their molecular accessibility, namely, by their ability to be
crossed by reactants and products moving to and from the
active sites. On these grounds, it appears quite obvious that
any application of acidic ion-exchange resins ought to be pre-
ceded by a careful examination of the resin morphology.30–34
Morphology of polymeric resins cannot be well character-
ized by conventional porosimetry methods such as mercury
intrusion or nitrogen adsorption–desorption since they
require completely dry samples. Using such data to interpret
the different performances observed (e.g. in reactions carried
out in solvents) requires the assumption that the resin
morphology is not significantly changed when it is wet with
solvent. This assumption is clearly not valid using hydro-
philic polymeric catalysts in liquid-phase polar reaction
environments.31
Therefore, other characterization techniques are needed
in order to study the actual morphology of acidic ion-
exchange resins in the reaction media. To the best of our
knowledge, the only procedure employed to assess the
morphology of ion-exchange resins in a swollen state has
been the inverse steric exclusion chromatography (ISEC)
technique.
Previous studies revealed that most catalytic performances
can be qualitatively predicted by the morphological descrip-
tion in the swollen state given by ISEC.18,27,34,35 The aim of
this work is to study the synthesis of ethyl hexyl ether from
1-hexanol and diethyl carbonate over acidic ion-exchange
resins (Fig. 1). In this sense, catalytic performances are evalu-
ated and relationships between ISEC data of catalysts and
their catalytic behaviour are obtained. The effect of the
temperature on the reaction system is also studied.
2. Experimental
2.1 Materials
1-Hexanol (≥98%, Acros) and diethyl carbonate (≥99%, Acros)
were used as reactants. Distilled water, ethanol (≥99%,
Panreac), 1-hexene (≥99%, Aldrich), di-n-hexyl ether (≥97%,
Fluka) and ethyl hexyl ether (≥99%, produced and purified by
rectification in our lab) were used for analysis purposes. N2
was used to pressurize the reactor (≥99.995%, Abelló Linde)
and He as chromatographic carrier gas (≥99.998%, Linde).
Nine commercial acidic ion-exchange resins based on a
styrene-co-divinylbenzene polymer, supplied by Rohm and
Haas France (Amberlyst 15, 16, 39, 46 and 70), Aldrich
(Dowex 50W×8, 50W×4 and 50W×2) and Purolite (Purolite
CT482), were used as catalysts. As Table 1 shows, tested
resins cover a wide range of structural properties and
acid capacities, with the goal of evaluating their influence on
the catalyst activity. Over-sulfonated resins were not used
since they hardly increased the catalytic performance in similar
reactions,18,34 while the over-sulfonation procedure supposes
an increase in manufacturing cost.
2.2 Characterization of ion-exchange resins
Morphology of ion-exchangers in the swollen state was
analysed by inverse steric exclusion chromatography (ISEC).
An ISEC apparatus consisted of a HPLC pump (Waters 510),
a sampling valve, a stainless steel column (4.27 cm3) and a
refractometric detector (Shodex RI-100). The detector signal
was connected to a computer and the sampling data were
synchronized with the mobile phase flow rate using a drop
counter.
Catalysts were crushed, sieved in the swollen state
(0.125 ≤ dp ≤ 0.250 mm) and placed overnight in the mobile
phase (a 0.2 N solution of Na2SO4 for aqueous phase
measurements). Then, the swollen catalyst was packed in the
column by flowing the mobile phase for 30 min (≈5 mL min−1),
and connected to the apparatus. During chromatographic
measurements, the standard solutes of known molecular
sizes (deuterium oxide, sugars and dextrans) were injected
independently (20 μL). Elution volumes were determined on
the basis of the first statistical moments of the chromato-
graphic peaks. For each solute the measurement was repeated
three times for determination of the experimental error. At
the end of the measurements, the catalyst was washed with
distilled water, quantitatively extruded from the column and
dried overnight (T = 383 K), and finally it was weighted. Acid
capacity was measured by titration against standard base.
2.3 Experimental set-up
Experiments were conducted in a 100 mL stainless steel
autoclave which operates in batch mode. A magnetic driven
turbine was the mixing system and an electric furnace was
the heating one ( 1 K). One of the reactor outlets was
connected directly to a liquid sampling valve, which injected
0.1 μL of pressurized liquid into a gas–liquid chromatograph
Fig. 1 Target reaction of this work.
This journal is © The Royal Society of Chemistry 2015
Catal. Sci. Technol., 2015, 5, 2238–2250 | 2239

View Article Online
Paper
Catalysis Science & Technology
Table 1 Specific characteristics of used ion-exchange resins27,34
Structure
Acid capacitya (mmol H+ g−1
)
DVB (%)
Highb [4.80–4.95]
Mediumc [2.65–4.25]
Lowd [0.87]
Macroreticular
Gel-type
20
12
8
8
4
Amberlyst 15
Amberlyst 16
Amberlyst 39
Dowex 50W×8
Dowex 50W×4
Dowex 50W×2
Amberlyst 46
Purolite CT482
Amberlyst 70
2
a
b
Titration against standard base. Conventionally sulfonated, with a stoichiometric amount of sulfonic groups: a sulfonic group by styrene
ring. c Conventionally sulfonated chlorinated resins. d Surface sulfonated: sulfonation restricted to external gel-phase layers.
(HP6890A, Hewlett Packard) equipped with a thermal con-
ductivity detector. A 50 m × 0.2 mm × 0.5 μm methyl silicone
capillary column was used to determine the composition of
liquid mixtures. The column was temperature programmed
to start at 323 K with a 10 K min−1 ramp to reach 573 K and
held for 6 min. The He flow rate was set at 30 mL min−1. All
the species were identified using a second chromatograph
equipped with a mass spectrometer (GC/MS 5973, Agilent).
corresponds to the standard deviation between replicated
experiments.
In each experiment, conversion of reactant j (Xj) and selec-
tivity (Skj ) and yield (Ykj ) of product k with respect to reactant j
were calculated by:
moles of j reacted
X j 
100 %


(1)
(2)
(3)
initial moles of j
moles of j reacted to form k
moles of j reacted
Sjk 
100 %



2.4 Procedure
Catalysts were dried at room temperature for 24 h prior to
mechanical sieving (Dowex catalysts, dp ≤ 0.40 mm; the other
catalysts, 0.40 ≤ dp ≤ 0.63 mm). After sieving, resin samples
were dried at 383 K, firstly at atmospheric pressure for 3 h
and, subsequently, under vacuum overnight. This drying pro-
cedure assures a water content of resins of less than 3 wt.%.34
The reactor was filled with 70 mL of a mixture of
1-hexanol and diethyl carbonate (molar ratio RHeOH/DEC = 2).
Then, the liquid was heated up to the working temperature
and stirred at 500 rpm. The pressure was kept at 2.5 MPa
with N2 ensuring that the reaction medium was in the liquid
phase. In these conditions, the gas phase was composed of
N2 and carbon dioxide, the only non-condensable product. The
change in volume of the liquid phase by the CO2 exit to the gas
phase was partially neutralised by the change in density of the
liquid phase so that no change in the volume of the liquid
phase was observed. When the mixture reached the working
temperature (approximately after 15 minutes), the dry catalyst
was injected into the reactor from an external cylinder by a N2
pulse. Experiments were carried out in the temperature range
of 403–463 K. The amount of the used catalyst was changed at
each temperature within the range of 0.2–3.7 g in order to
obtain similar conversion levels during the whole series of
experiments performed over each single catalyst.
moles of j reacted to form k
initial moles of j
Yjk

100 %

Initial reaction rates of formation of compound j (r0j ) were
found from the slope of curves of the number of moles of
compound j (nj) versus time, where Wcat is the mass of the
dry catalyst:
dn
dt








j 

1
mol
rj0 
(4)
W
cat 
h kgcat
t0 
Under the selected reaction conditions (temperature,
pressure, stirring speed, particle size and catalyst mass), it is
assumed that the obtained reaction rates were not influenced
by mass transfer effects in this reaction set-up.23,27,36
3. Results
3.1 Morphology of ion-exchange resins
Ion exchange resins swell in liquid polar media, particularly in
the aqueous phase, and non-permanent pores between poly-
mer chains (spaces) appear. A description of resin morphology
in the swollen state can be obtained by analysis of ISEC data.
The porous structure is modelled as a set of discrete volume
fractions each composed of pores having simple geometry and
uniform sizes. A portion of open spaces in macromolecular
resins can be characterized by the cylindrical pore model
(“true pores”). The true pore diameter ranged in the 10–118 nm
zone for Amberlyst 15 and 16, and in the 8–41 nm zone for
Amberlyst 39 and 70, and Purolite CT482; most of such spaces
Catalyst injection was taken at time zero. At that moment,
no conversion of 1-hexanol was observed but the conversion
level of diethyl carbonate, which suffers thermal decomposi-
tion, was always less than 1%. The composition of the liquid
mixture was analysed hourly until the end of the experiment
(6 h) to obtain the mole variation over time for all species. All
the catalytic tests were, at least, replicated once to assure data
reproducibility. The experimental error shown in this work
2240 | Catal. Sci. Technol., 2015, 5, 2238–2250
This journal is © The Royal Society of Chemistry 2015

View Article Online
Catalysis Science & Technology
Paper
lie in the 8–10 nm zone. Swollen gel-type resins do not show
spaces in the range of mesopores.
both reactants followed the same pattern (see Fig. 2A). Thus,
it can be inferred that the reactivity of diethyl carbonate and
1-hexanol on acidic ion-exchange resins is quite similar. The
number of 1-hexanol moles present in the liquid phase was
always higher than that of diethyl carbonate since the alcohol
was initially in excess.
From the very beginning, reactants were consumed
leading to the formation of several compounds. Ethyl hexyl
carbonate, ethanol, ethyl hexyl ether, di-n-hexyl ether and
diethyl ether were always formed in significant amounts;
water and di-hexyl carbonate were formed in much lower
quantities (not plotted for the sake of clarity) and C6 alkenes
were only detected on some ion-exchange resins.
The reaction products can be divided into two groups
depending on their evolution pattern. On the one hand, ethyl
hexyl carbonate and ethanol showed the typical behaviour of
reaction intermediate compounds in a batch-type experiment
(see Fig. 2B). Their number of moles highly increased during
the first hours of reaction, and subsequently, they showed a
plateau in the case of ethanol and a slight maximum in the
case of ethyl hexyl carbonate. This behaviour suggests that
they were firstly produced in one reaction and then consumed
in the second one, revealing that these compounds take part
in some reactions in series. On the other hand, the number of
moles of the three ethers formed, ethyl hexyl ether, di-n-hexyl
ether and diethyl ether, keeps increasing until the end of the
experiment (see Fig. 2C). As a result, conversion of reactants
and selectivity and yield of ethyl hexyl ether (see Fig. 3A–C,
respectively), as well as yield of other ethers, steadily increase
during the whole experiment. Therefore, it is clear that for
industrial purposes, longer contact times are necessary.
The observed distribution of products suggests the reaction
network shown in Fig. 4. As seen, several plausible reactions are
involved forming a series–parallel scheme. According to the pro-
posed scheme, the synthesis of ethyl hexyl ether proceeds from
two simultaneous reaction pathways. The first one forms ethyl
hexyl ether and carbon dioxide, involving transesterification
reaction of diethyl carbonate to ethyl hexyl carbonate and
ethanol (reaction (1)) and thermal decomposition of the ethyl
hexyl carbonate (reaction (2)). The second one forms ethyl
hexyl ether and water, involving reaction (1) and intermolecular
dehydration of ethanol with 1-hexanol (reaction (3)).
The cylindrical pore model is unable to describe the
spaces between polymer chains formed in the micropore
range. A good view of the three-dimensional polymer network
of the swollen gel-phase is given by the Ogston geometrical
model, in which micropores are described by spaces between
randomly oriented rigid rods.37 From ISEC data, the Ogston
model makes possible to estimate the specific volume of swol-
len polymer (free space plus that occupied by the skeleton),
Vsp, and also to distinguish gel zones of different densities or
polymer chain concentrations. Gel-phase porosity is described
as a set of zones of different chain densities. According to the
Ogston model, the pore size of each gel-phase zone is repre-
sented as the total length of polymer chains per unit of volume
of swollen polymer (nm nm−3 or nm−2).35,38–42
In particular, the specific volume of swollen polymer (Vsp)
has been modelled in 5 zones of polymer chain concentrations
of 0.1, 0.2, 0.4, 0.8 and 1.5 nm nm−3; the sum of estimated
volumes of the swollen gel fractions used is the total Vsp value.
The spaces between chains in such zones are equivalent to
pore diameters of 9.8, 4.3, 2.6, 1.5 and 1 nm, respectively.43,44
Table 2 shows gel-phase morphology data for tested catalysts
in the swollen state. As seen, Vsp decreases on macroreticular
and gel-type resins and, accordingly, the predominant gel
phase is denser, on increasing DVB%. The densest resins in
the swollen state are Amberlyst 15 and 16 and Dowex 50W×8
which show chain concentrations of 1.5 nm−2, and less dense
ones are Dowex 50W×2 with polymer density in the range of
0.2–0.4 nm−2 and Amberlyst 70 (0.4 nm−2). Dowex 50W×4
(0.4–0.8 nm−2), Amberlyst 39 (0.4–1.5) and Purolite CT482
(0.8–1.5 nm−2) show intermediate density values.
Table 2 also shows the acid capacity and the maximum
operation temperature recommended by the manufacturers
of tested ion-exchange resins.
3.2 Reaction network
Fig. 2 shows the evolution of the number of moles of each
compound over time in an experiment conducted on Dowex
50W×2, which can be taken as a representative of the whole
series of experiments. As shown, the consumption rate of
Table 2 Morphology of the polymer gel-phase swollen in water, acid capacity and maximum operation temperature of catalysts
a
Acid capacityb
Catalyst
Vsp (cm3 g−1
)
Polymer fraction density [nm nm−3
]
0.1
0.2
0.4
0.8
1.5
Total
(mmol H+ g−1
)
Tmax (K)
c
Amberlyst 15
Amberlyst 16
Amberlyst 39
Amberlyst 46
Amberlyst 70
Purolite CT482
Dowex 50W×2
Dowex 50W×4
Dowex 50W×8
0.622
0.913
0.595
0.164
0.622
1.136
1.643
0.190
1.149
1.081
2.677
1.920
1.404
4.81
4.80
4.95
0.87
2.65
4.25
4.83
4.95
4.83
393
393
403
423
463
463
423
423
423
0.208
0.218
0.010
0.071
0.016
0.243
0.016
0.588
0.007
0.729
0.034
1.072
0.093
1.949
0.639
0.249
0.740
1.325
1.281
0.046
a
ISEC method. b Titration against standard base. c Manufacturer data.
This journal is © The Royal Society of Chemistry 2015
Catal. Sci. Technol., 2015, 5, 2238–2250 | 2241

View Article Online
Paper
Catalysis Science & Technology
Fig. 2 Evolution of the mole number of species over time, A (♦ 1-hexanol;
● diethyl carbonate), B (● ethanol; ♦ ethyl hexyl carbonate), C (○ ethyl
hexyl ether; ◊ di-n-hexyl ether; Δ diethyl ether). Experimental conditions:
T = 423 K, P = 2.5 MPa, Wcat = 1 g, RHeOH/DEC = 2, catalyst Dowex 50W×2.
Fig. 3 Evolution of reactant conversion (A) and selectivity (B) and yield
of ethyl hexyl ether (C) over time. Compound j: diethyl carbonate
(round circles), 1-hexanol (diamonds). Experimental conditions T = 423 K,
P = 2.5 MPa, Wcat = 1 g, RHeOH/DEC = 2, catalyst Dowex 50W×2.
Altogether, they form two moles of ethyl hexyl ether, one
of carbon dioxide and one of water by consuming two moles
of 1-hexanol and one of diethyl carbonate (see target reaction,
Fig. 1). Reaction (1) takes place in both pathways and thereby
2242 | Catal. Sci. Technol., 2015, 5, 2238–2250
This journal is © The Royal Society of Chemistry 2015

View Article Online
Catalysis Science & Technology
Paper
diethyl carbonate and 3–16% with respect to 1-hexanol.
Nevertheless, the significant amount of intermediate products
still present at the end of the run (Fig. 3B) and the continuous
increase in the yield with time (Fig. 3C) indicate that higher
amounts of ethyl hexyl ether can be easily achieved by increas-
ing contact time and/or the reaction temperature.
Among tested catalysts, the most active in terms of
reactant conversion and yield of ethyl hexyl ether was Dowex
50W×2 (Xj = 63–66% and YEj HE = 15–19%). Even though this
low crosslinked gel-type resin was initially designed for fine
chemical and pharmaceutical column separation purposes,
lately it has shown satisfactory catalytic performance in
several types of reactions, e.g. bio-oil upgrading, glycerol acet-
ylation, dehydration of alcohols and synthesis of glycol and
bisphenol A.45–48 Besides, it has been reported as a suitable
support for embedding Ru nanoparticles as bifunctional
catalysts for γ-valerolactone formation.49 The high capacity of
Dowex 50W×2 to swell the polymer backbone in the polar
medium due to a low divinylbenzene content was found to
be crucial for accessing the inner sulfonic acid sites.
Fig. 4 Series–parallel reaction scheme between 1-hexanol and diethyl
carbonate over acidic ion-exchange resins.
it is a mandatory step for the further formation of ethyl
hexyl ether.
The routes of diethyl ether and di-n-hexyl ether formation
are analogous to those of ethyl hexyl ether (reactions (4) and
(5) and (6)–(8), respectively). Moreover, small amounts of C6
alkenes were formed from the intramolecular dehydration of
1-hexanol on some resins (reaction (9)). For simplicity, the
nine reactions can be grouped into four different kinds of
reactions:
Amberlyst 46 was the less active catalyst for obtaining
ethyl hexyl ether (Xj = 12–24% and YEj HE = 2–3%). A special
property of Amberlyst 46 is the small amount of sulfonic
groups within the gel phase because the sulfonation is
restricted to only the first few layers of styrene rings. Consid-
ering the low catalytic activity shown by Amberlyst 46, it can
be assured that the presence of sulfonic groups located
within the gel phase clearly promotes the formation of ethyl
hexyl ether. However, as several reactions are involved in this
catalytic process, special attention has to be paid to the selec-
tivity within the gel phase.
As in most series–parallel reaction systems, the selectivity
was clearly influenced by the conversion level, in such a way
that the selectivity to ethyl hexyl ether increased with reactant
conversion (see Fig. 3B). For comparison purposes, Table 4
shows selectivity data at the same level of diethyl carbonate
conversion (XDEC = 10%), estimated from the mole–time
curves. That conversion level was reached with all catalysts in
less than 2 hours of reaction, but with Amberlyst 46 more
than 5 hours were necessary to achieve it. At this initial stage
of reaction, it has been assumed that all the products were
formed directly from reactants. Thus, diethyl carbonate should
only form ethyl hexyl carbonate by the transesterification
with 1-hexanol (reaction (1)) or it can be thermally decomposed
to diethyl ether (reaction (4)); and analogously 1-hexanol
should only form ethylhexyl carbonate (reaction (1)), react
with another 1-hexanol molecule producing di-n-hexyl ether
(reaction (6)) or form C6 alkenes by dehydration (reaction (9)).
With this approach, the reaction network at the initial reaction
time was simplified from nine series–parallel reactions to four
parallel ones.
(i) Transesterification of carbonates (reactions (1) and (7)),
(ii) Thermal decomposition of carbonates (reactions (2),
(4) and (8))
(iii) Alcohol intermolecular dehydration (reactions (3), (5)
and (6))
(iv) Alcohol intramolecular dehydration (reaction (9)).
3.3 Catalyst screening
Data obtained for all tested catalysts are summarized in
Table 3. After 6 h of reaction, conversion of diethyl carbonate
ranged 12–66% and that of 1-hexanol ranged 24–63%. As a
rule, a similar level of conversion was achieved from both
reactants, 1-hexanol being slightly more reactive than diethyl
carbonate. These results highly differ from those obtained in
a
previous study when 1-octanol was used instead of
1-hexanol for obtaining ethyl octyl ether.34 In that case,
diethyl carbonate was much more reactive than the alcohol
and the formation of diethyl ether was more favoured than
ethyl octyl ether synthesis. In the present reaction, the higher
reactivity of 1-hexanol contributes positively to the formation
of the asymmetrical ether.
As mentioned above, yields obtained were rather low for
industrial purposes; they ranged 2–19% with respect to
Table 3 Conversion of reactants and yield of ethyl hexyl ether (T = 423 K,
P = 2.5 MPa, Wcat = 1 g, RHeOH/DEC = 2, t = 6 h)
EHE
DEC
EHE
XDEC (%)
XHeOH (%)
Y
(%)
Y
(%)
HeOH
Amberlyst 15
Amberlyst 16
Amberlyst 39
Amberlyst 46
Amberlyst 70
Purolite CT482
Dowex 50W×2
Dowex 50W×4
Dowex 50W×8
32.2 0.5
41.8 0.2
54.8 1.5
11.7 0.9
44.8 0.8
53.9 1.8
65.9 0.1
61.1 1.6
48.6 0.5
44.1 0.4
52.2 0.5
58.4 1.5
24.0 1.2
56.6 0.2
57.8 0.2
63.4 0.2
61.2 1.2
56.7 0.2
10.1 0.1
13.8 0.1
17.2 0.5
2.1 0.5
12.6 0.7
17.3 0.6
19.4 0.1
19.1 1.5
15.7 0.1
10.2 0.1
13.1 0.1
14.6 0.1
2.7 0.6
10.7 0.5
15.1 0.8
15.3 0.1
15.5 0.8
13.9 0.1
In general, diethyl carbonate reacted readily with
1-hexanol and interesting selectivities to ethyl hexyl carbon-
EHC
DEC
ate were achieved (84% < S
< 98%), which indicates that
the loss of ethyl groups to diethyl ether formation was
DEE
DEC
certainly hindered for some catalysts (2% < S
< 16%).
This journal is © The Royal Society of Chemistry 2015
Catal. Sci. Technol., 2015, 5, 2238–2250 | 2243

View Article Online
Paper
Catalysis Science & Technology
Table 4 Estimated selectivity at the conversion level of XDEC = 10% (T = 423 K, P = 2.5 MPa, Wcat = 1 g, RHeOH/DEC = 2)
DEE
DEC
EHC
HeOH
DHE
Time (min)
XHeOH (%)
SEDHECC (%)
S
(%)
S
(%)
S (%)
HeOH
SaHlekOenHes (%)
Amberlyst 15
Amberlyst 16
Amberlyst 39
Amberlyst 46
Amberlyst 70
Purolite CT482
Dowex 50W×2
Dowex 50W×4
Dowex 50W×8
123
89
61
311
70
64
37
47
71
24.5 0.1
25.5 1.0
22.1 1.0
21.2 0.2
22.6 0.4
22.6 0.4
19.8 0.1
20.9 0.6
24.3 0.3
84.4 2.4
89.2 0.8
95.2 1.6
97.6 1.6
96.2 2.4
91.6 1.4
96.3 0.7
95.0 1.2
93.1 2.1
15.6 2.4
10.8 0.8
4.8 1.6
2.4 1.6
3.8 2.4
8.4 1.4
3.7 0.7
5.0 1.2
6.9 2.1
43.8 1.1
49.1 1.3
59.3 2.0
68.9 1.6
64.6 1.2
54.6 1.4
72.4 0.5
67.0 1.3
52.0 0.7
39.4 1.2
43.8 1.4
39.0 0.7
30.0 0.7
35.7 0.5
42.1 0.6
27.6 0.4
33.2 1.4
44.6 1.3
16.9 1.1
7.1 1.7
2.2 2.0
1.1 0.7
0.6 0.6
3.3 1.8
0.0 0.0
0.1 0.1
3.4 1.2
On the contrary, the ability of 1-hexanol to react with diethyl
to be reusable without activity loss after three 50 h on
stream cycles, and desulfonation after 24 h at 463 K in water
stream was negligible.28,51
EHC
carbonate is significantly lower (44% < S
< 72%),
HeOH
mainly because such reaction competes with di-n-hexyl ether
formation by the dehydration reaction of two hexanol mole-
In relation to catalytic activity of ion-exchange resins the
presence of water is outstanding. As discussed in the literature,
water is essential to swell the resin making easier the access to
sulfonic groups. The size of spaces formed on swelling deter-
mines selectivity to moderately bulky compounds, such as lin-
ear ethers of 8 or more carbon atoms. As a result, as seen in
previous discussion, activity and selectivity to ethyl hexyl ether
is higher for the more swollen resins. Kinetically, water inhibits
the reaction rate since it competes with the alcohol for active
sites. The inhibitory effect of water on the formation of ethyl
hexyl ether will be analysed in further work by performing
a kinetic study. The catalytic activity decay observed in liquid-
phase dehydration reaction of alcohol to ether over ion-
exchange resins is ascribed to the strong competitive adsorp-
tion on acid sites, blocking the adsorption of alcohol, and to
hydrolysis of the polymer backbone releasing sulphuric acid,
causing desulfonation in this way. Fortunately, loss of activity
originating from water adsorption is fully recovered as soon
as water is removed from the reaction medium.36,51
Finally, the least selective resin Amberlyst 15, despite its
maximum operating temperature being as low as 393 K,
might be reused several times at 423 K without activity loss
as it was shown in the dehydration of 1-butanol to di-n-butyl
ether.54 However, in this case morphology changes gain
surface area but some deactivation is detected. The
maintaining of catalytic activity for a few runs would be the
result of those opposite changes in the resin.
DHE
cules (28% < S
< 45%). Finally, it is observed that the
HeOH
formation of C6 alkenes via intramolecular dehydration of
1-hexanol is hardly significant for some particular catalysts
alkenes
HeOH
(0% < S
< 17%). In summary, the dehydration of two
molecules of 1-hexanol to give di-n-hexyl ether and water is
the main competitive reaction of the desired reaction
between diethyl carbonate and 1-hexanol, a necessary step for
the further formation of ethyl hexyl ether.
From all the data collected, it is seen that Dowex 50W×2
is the most active catalyst, and also the most selective to
form ethyl hexyl carbonate. Interestingly, Dowex 50W×2
(fully sulfonated) showed selectivities very close to those of
Amberlyst 46 (only surface sulfonated), while it needed 8-fold
less time to achieve a comparable level of conversion (37 min
and 311 min). Therefore, among the tested acid ion-exchange
resins, Dowex 50W×2 seems to be the most suitable catalyst
for the synthesis of ethyl hexyl ether, whereas Amberlyst 15,
perhaps the most industrially used acidic ion-exchange resin
as a catalyst, is definitely not a suitable one for ethyl hexyl
ether production due to the poor selectivity showed.
On the selection of a suitable catalyst for ethyl hexyl ether
synthesis, it is necessary that it retains its activity for a long
time. On that score, it has been reported that gel-type resins
maintain their activity and selectivity in aqueous media for
more that 80 h on stream at 393–423 K.50,51 In particular, in
the dehydration reaction between ethanol and 1-octanol to
ethyl octyl ether in a fixed bed reactor, reaction rates on
Dowex 50W×2 only dropped by 10% after 70 h at 423 K.51
Interestingly, the catalyst recovered its activity after drying in
a N2 flux and, in addition, its acid capacity was unchanged.
Taking into account that the maximum operating temper-
ature of Dowex 50W×2 is 423 K (Table 2), if temperatures
higher than 423 K were needed for operation, Amberlyst 70 is
a good option. It is a resin with upgraded thermal stability
(Tmax = 463 K), and selective to ethyl hexyl ether with respect
to diethyl carbonate as Dowex 50W×2 is. It is to be noted that
Amberlyst 70 is as selective as gel-type resins with 2–4 DVB%
in other liquid-phase dehydration reactions of alcohol to
linear ether.27,34,52,53 Moreover, in the reaction of synthesis
of ethyl octyl ether in a fixed bed reactor, Amberlyst 70 was found
3.4 Relationship between swollen-state morphology
and catalytic activity
From a chemical point of view, all acidic ion-exchange resins
used in this work have a simple and uniform composition: sul-
fonic groups bound to aromatic rings. Thus, the acid strength
of sulfonic groups, the actual catalytic active species, was simi-
lar in most resins and only ≈5% higher in the particular case
of chlorinated ones.28 In addition, the number of sulfonic
groups was quite similar in most resins (4.8–4.95 mmol H+ g−1).
Despite the chemical similarity, tested catalysts showed differ-
ent catalytic performances which typically are a consequence of
differences in their swollen morphology.31
2244 | Catal. Sci. Technol., 2015, 5, 2238–2250
This journal is © The Royal Society of Chemistry 2015

View Article Online
Catalysis Science & Technology
Paper
In the present system, where the mixture was composed of
alcohols, carbonates, ethers and water, the reaction proceeds
essentially in a polar environment.21 As a consequence, the
actual working-state morphology of resins could be quite similar
to that deduced from aqueous ISEC measurements discussed in
section 3.1 (see Table 2). A characteristic property of acidic ion-
exchange resins obtained by means of ISEC is Vsp; the more
swollen resins having the higher Vsp values. Interestingly, recent
studies showed that Vsp values allowed prediction of the catalyst
performance in polar environments with high accuracy.18,34
Likewise, it is assumed that based on resin Vsp values their cata-
lytic behaviour in this reaction system can be predicted.
Fig. 5 shows the dependence of initial consumption rates
of reactants on Vsp. It is clearly seen that consumption rates
of reactants are enhanced in catalysts having high Vsp values.
This fact could be explained because a high expanded poly-
mer skeleton favours the accommodation of reaction inter-
mediates and allows the access of reactants to a larger num-
ber of acid centres. In this way, the reaction rate can be
increased up to 4-fold by using resins able to swell its poly-
mer skeleton in polar environments. The differences between
the consumption rate profiles of both reactants and the Vsp
are essentially due to by-product formation.
1
dn



r0  rD0EE



DEE 
(6)
(7)
(8)
4
W
W
W
dt
t=0
cat 
1
dn


t=0
r0  rD0HE
DHE 
6

dt
cat 
1
dn


t=0
r0  r0
alkenes 
9
alkenes

dt
cat 
Fig. 6 displays the influence of the Vsp values on product
formation. Initial reaction rates of ethyl hexyl carbonate for-
mation showed a strong dependence on the catalyst Vsp
values, which might indicate that it is the most sterically
demanding reaction. An entirely opposite behaviour was
observed in the formation of alkenes as they were only
detected over resins with low Vsp values. Therefore, the intra-
molecular dehydration to olefin occurred preferentially on
resins with a less expanded polymer. An intermediate behav-
iour was observed in the formation of di-n-hexyl ether and
diethyl ether, where the reaction rate dependence on Vsp was
not relevant. This behaviour needs further discussion.
According to the reaction scheme (Fig. 4), the reactions
taking place at the beginning of the reaction (t = 0) were
essentially reactions (1), (4), (6) and (9); namely, the forma-
tion of ethyl hexyl carbonate, diethyl ether, di-n-hexyl ether
and C6 alkenes, respectively. Thus, it is assumed that the only
relevant reaction that took place at t = 0, not coming directly
from diethyl carbonate and/or 1-hexanol, was the thermal
decomposition of ethyl hexyl carbonate to ethyl hexyl ether
(reaction (2)), while the extent of reactions (3), (5), (7) and (8)
was negligible. Accordingly, the reaction rates of product for-
mation were calculated as follows:
As a consequence of the observed dependence of initial
reaction rates on the Vsp, it might be assumed that each sin-
gle reaction is favoured on different zones of the polymer,
depending on the steric hindrances for reactants, products
and intermediates of reaction. In this sense, the intrinsic
reaction rate achieved by a catalyst can be expressed as the
sum of contributions of individual polymer domains.35,42 The
individual contribution of each of the five polymer domains
of a characteristic chain density was computed as follows:
1.5
1.5






Vsp,n
Vsp
rj0 
r0

TOF  H

0
i,n

 

(9)








i,n

n0.15
n0.15




1
dn



r0  rE0HC

EHC 
(5)
1
W
cat 
dt
t=0
Fig. 6 Influence of the specific volume of swollen polymer (Vsp) on
initial reaction rates of formation of compound (♦ ethyl hexyl
carbonate; Δ diethyl ether; ◊ di-n-hexyl ether; x alkenes). Experimental
conditions: T = 423 K, P = 2.5 MPa, RHeOH/DEC = 2.
Fig. 5 Influence of the specific volume of swollen polymer (Vsp) on
the consumption rate of reactants (● diethyl carbonate; ♦ 1-hexanol).
Experimental conditions: T = 423 K, P = 2.5 MPa, RHeOH/DEC = 2.
j
This journal is © The Royal Society of Chemistry 2015
Catal. Sci. Technol., 2015, 5, 2238–2250 | 2245

View Article Online
Paper
Catalysis Science & Technology
For a given catalyst r0i is the initial reaction rate of reaction
containing higher volume zones of the dense swollen gel phase
should be avoided as a catalyst in order to favour the formation
of ethyl hexyl carbonate. These results are in agreement
with the lower selectivity to ethyl hexyl ether of the catalysts
containing the densest zones.
As for ethyl hexyl carbonate, di-n-hexyl ether and alkene
(all formed from 1-hexanol) formation reaction, TOFi,n values
are displayed in Fig. 8. As it is seen, the polymer zone of
0.4 nm nm−3 density favours equally two competing reactions:
the formation of ethyl hexyl carbonate and ethanol, and that
of di-n-hexyl ether and water. Unfortunately, the selectivity to
products formed from 1-hexanol cannot be optimized with
this kind of catalyst as both reactions took place preferentially
in the same polymer zone. However, the intermolecular
dehydration of 1-hexanol to C6 alkenes only takes place within
the densest zone of the polymer. Thus, catalysts containing
1.5 nm nm−3 polymer zones should be avoided.
As shown, eqn (9) provides a useful phenomenological
assessment to study the catalytic behaviour at the molecular
level and to select a suitable catalyst depending on the desired
reaction. The transesterification reaction of 1-hexanol and
diethyl carbonate and the 1-hexanol intermolecular dehydration
showed similar steric limitations and they occurred preferen-
tially in medium expanded polymer zones (0.4 nm nm−3 zone).
In inorganic materials, this polymer chain density is equivalent
to pores of 2.6 nm.43,44 In contrast, the thermal decomposition
of diethyl carbonate is favoured in slightly denser ones
(0.8 nm nm−3 zone, equivalent to 1.5 nm (ref. 43 and 44)).
Finally, the formation of C6 alkenes only takes place in zones
with severe steric limitations (1.5 nm nm−3 zone, equivalent to
1.0 nm (ref. 43 and 44)).
i, r0i,n the contribution to the reaction rate induced by polymer
volume fraction n, TOF0i,n, initial specific activity of polymer
fraction n for reaction i defined as reaction rate per acid site,
[H+] the acid capacity, Vsp,n the specific volume of swollen
polymer occupied by fraction n and Vsp the total specific
volume of swollen polymer. As only Dowex 50W×2 shows a
relevant swollen specific volume with 0.2 nm nm−3 polymer
density (see Table 2), the polymer chain concentrations of 0.1
and 0.2 nm nm−3 were grouped as a 0.15 nm nm−3 fraction.
Then, a set of equations (ri, one for each catalyst) and four
unknown parameters (TOF0i,n, initial specific activity of each
polymer fraction) can be obtained for each reaction. In order
to obtain TOF0i,n, values, the squared difference between exper-
imental and computed values of ri was minimized.
Distribution of the specific activity values for each polymer
chain density, TOF0i,n, is depicted in Fig. 7 for species coming
from diethyl carbonate. Relevant differences were observed
depending on the reaction considered. The transesterification
reaction of 1-hexanol with diethyl carbonate to form ethyl
hexyl ether and ethanol, which is by far the most favoured
reaction, showed a maximum of specific activity in the
0.4 nm nm−3 zone. A further increase in the polymer chain
density would hinder the formation of the necessary reaction
intermediate due to steric limitation, and as result, polymer
zones denser than 0.4 nm nm−3 led to a decrease in specific
activity. The most expanded zone (0.15 nm nm−3) also
showed lower specific activity since the polymer volume
fraction with the lowest chain density fractions contains
excessively dispersed acid groups. Tested resins contain one
sulfonic group per aromatic ring of the polymer backbone,
thus, the number of acid groups present in the more
expandable zones was much lower.
3.5 Influence of temperature on catalyst performance
On the contrary, the highest-performance polymer fraction
for the thermal decomposition of diethyl carbonate to diethyl
ether is shifted to a denser zone (0.8 nm nm−3) showing that
it is a less sterically demanding reaction. Therefore, resins
The previous section revealed that the morphology of
ion-exchangers is decisive to favour the transesterification
reaction of diethyl carbonate with 1-hexanol and thereby
Fig. 7 TOF0i,n distribution for each polymer chain density zone for
Fig. 8 TOF0i,n distribution for each characteristic polymer chain density
on reaction i (♦), ethyl hexyl carbonate (○), di-n-hexyl ether and (x)
alkenes. Experimental conditions: T = 423 K, P = 2.5 MPa, RHeOH/DEC = 2.
reaction
i (♦), ethyl hexyl carbonate and (Δ), and diethyl ether.
Experimental conditions: T = 423 K, P = 2.5 MPa, RHeOH/DEC = 2.
2246 | Catal. Sci. Technol., 2015, 5, 2238–2250
This journal is © The Royal Society of Chemistry 2015

View Article Online
Catalysis Science & Technology
Paper
EHC
enhance the further yield of ethyl hexyl ether. On the other
hand, using the most suitable catalyst, a careful selection
of the reaction temperature can play a key role in the process
optimization. The effect of temperature was checked over the
chlorinated resins Purolite CT482 and Amberlyst 70 which are
thermally stable up to 463 K (see Table 2). Experiments were
performed in the temperature range of 403–463 K with initial
1-hexanol to diethyl carbonate molar ratio RHeOH/DEC of 2. The
amount of the catalyst was varied from 0.2 to 3.7 g to achieve
a similar conversion level in the whole temperature range.
Moreover, the influence of the catalyst load on the reaction rate
is negligible at the selected catalyst mass range.23,27,55
As seen in Table 5, despite changing the catalyst mass
upon increasing the working temperature, conversion levels
in the range of 42–50% over Amberlyst 70 and 43–63% over
Purolite CT482 for XDEC, and 55–65% over Amberlyst 70 and
57–63% over Purolite CT482 for XHeOH, make it difficult to
clearly evaluate the influence of temperature. As a whole, data
show a high dependence on temperature since about the same
conversion level is obtained upon increasing temperature by
20 K and decreasing the catalyst mass to a third. It is to be
noted that the yield of ethyl hexyl ether with respect to both
diethyl carbonate and 1-hexanol also increases with tempera-
ture, which could be because the reaction rate of the decom-
position of ethyl hexyl carbonate to ethyl hexyl ether highly
increases with temperature.
with respect to diethyl carbonate (91% < S
< 98%) or
DEC
with respect to 1-hexanol (44%< SEHC
HeOH < 77%). Again, it is to
be noted that at this low conversion levels a small quantity of
ethyl hexyl ether is also formed.
Apparent activation energies of the four competitive reac-
tions at low conversion levels were obtained from Arrhenius
plots of initial reaction rates. Fig. 9 shows the Arrhenius plot
for Purolite CT482. As seen the plot displays very good straight
lines which indicates that the rate limiting step does not
change over the temperature range explored. The Arrhenius
plot for initial reaction rates on Amberlyst 70 shows a similar
trend. Apparent activation energies found from the slope of
Arrhenius plots are shown in Table 7.
Apparent activation energies for linear ether formation
are in total agreement with those previously reported for
diethyl ether and di-n-hexyl ether formation on Amberlyst
70 by means of intermolecular dehydration of ethanol
To see the effect of temperature on selectivity, Table 6
shows XHeOH and selectivity over Amberlyst 70 at each tem-
perature and XDEC = 10%. As seen, 1-hexanol conversions
increase slightly which reveals that side reactions of forma-
tion of di-n-hexyl ether and olefins accelerate with tempera-
ture more than the reaction of transesterification to ethyl
hexyl carbonate. As a whole, selectivity to diethyl ether from
diethyl carbonate and selectivity to di-n-hexyl ether and olefins
from 1-hexanol increase with temperature. As a consequence,
the ability to obtain ethyl hexyl ether decreases with tempera-
ture as shown by the selectivity to ethyl hexyl carbonate
Fig. 9 Arrhenius plot of initial reaction rates of reaction i on CT482.
(♦ ethyl hexyl carbonate; Δ diethyl ether formation; ○ di-n-hexyl ether
formation; x alkene formation). Experimental conditions: P = 2.5 MPa,
RHeOH/DEC = 2.
Table 5 Conversion of reactants and yield of ethyl hexyl ether (P = 2.5 MPa, RHeOH/DEC = 2, t = 6 h)
EHE
DEC
EHE
HeOH
Catalyst
T (K)
Wcat (g)
XDEC (%)
XHeOH (%)
Y
(%)
Y
(%)
Amberlyst 70
403
423
443
473
403
423
443
473
3.678 0.001
1.044 0.001
0.351 0.001
0.192 0.001
3.645 0.001
1.061 0.001
0.353 0.001
0.193 0.001
49.7 1.9
44.8 0.8
42.3 0.9
46.1 0.8
43.8 1.4
53.9 1.8
51.0 1.9
62.7 1.3
55.9 0.3
56.6 0.2
59.2 0.2
65.0 0.4
60.2 0.2
57.8 0.2
61.0 0.2
62.7 0.2
11.0 0.6
12.6 0.7
11.3 0.6
19.4 1.1
13.5 0.5
17.3 0.6
18.8 0.5
26.6 0.9
9.1 0.4
10.7 0.5
12.5 0.6
16.8 0.4
7.5 0.5
15.1 0.8
16.6 0.7
21.5 1.2
Purolite CT482
Table 6 Estimated selectivity at the conversion level of XDEC = 10% over Amberlyst 70 (P = 2.5 MPa, RHeOH/DEC = 2)
DEE
DEC
EHC
HeOH
DHE
T (K)
Wcat (g)
XHeOH (%)
SEDHECC (%)
S
(%)
S
(%)
S (%)
HeOH
SaHlekOenHes (%)
403
423
443
463
3.678 0.001
1.044 0.001
0.351 0.001
0.192 0.001
20.4 3.5
22.6 0.4
24.1 0.1
24.4 2.0
97.6 0.8
96.2 2.4
91.7 5.1
90.9 2.9
2.4 1.2
3.8 2.4
8.3 0.7
9.6 3.7
76.3 0.5
64.6 1.2
52.4 0.4
44.4 0.1
23.1 0.2
35.7 0.5
45.5 0.4
50.1 0.6
0.2 0.2
0.6 0.6
2.6 0.1
5.7 0.4
This journal is © The Royal Society of Chemistry 2015
Catal. Sci. Technol., 2015, 5, 2238–2250 | 2247

View Article Online
Paper
Catalysis Science & Technology
Table
7 Apparent activation energies of thermally stable acid ion-
savings brought about by working at low temperature is a very
important benefit to take into account.
exchange resins at RHeOH/DEC = 2
Ea (kJ mol−1
)
A possible reactor design might use two unit reactors
connected in series. According to estimated activation energy
values, the selectivity to ethyl hexyl carbonate is clearly favoured
at low temperatures. On the contrary, ethyl hexyl ether, diethyl
ether and di-n-hexyl ether syntheses are favoured at intermedi-
ate temperature, whereas alkene formation is favoured at higher
ones. The first reactor should operate at low temperature
increasing in this way the selectivity to ethyl hexyl carbonate
and hindering the production of by-products (di-n-hexyl ether,
diethyl ether and olefins). The second unit would work at
higher temperature than the first one to increase the reaction
rate of thermal decomposition of ethyl hexyl carbonate into
ethyl hexyl ether. Nevertheless, further research is necessary
to implement such two-reactor units.
Catalyst
EHC
EHE
DEE
DHE
Alkenes
143 18
Amberlyst 70
Purolite CT482
68
78
3
1
98
100
3
6
103 13
98
102
101
4
4
6
136
4
(Ea,DEE = 100 5 kJ mol−1 36
and 1-hexanol (Ea,DHE = 108
)
7 kJ mol−1),52 respectively. On the other hand, apparent acti-
vation energy for ethyl hexyl ether synthesis is similar to that
one found for ethyl octyl ether over Amberlyst 70 (Ea = 105
4 kJ mol−1).36 Finally, apparent energies for linear ether forma-
tion over Amberlyst 70 are also similar to those of dehydration
of 1-pentanol to di-n-pentyl ether (118.7 0.2 kJ mol−1)56 and
that of 1-octanol to di-n-octyl ether from (114 7 kJ mol−1).57
Since in all quoted studies it was experimentally checked
that resistance to external mass transfer and diffusion was
negligible, it can be assumed that reported data in the present
work are free of mass transfer effects.
4. Conclusions
Ethyl hexyl ether can be successfully formed using diethyl
carbonate and 1-hexanol over acidic ion-exchange resins by
means of two reaction pathways: thermal decomposition of
ethyl hexyl carbonate and intermolecular dehydration of ethanol
with 1-hexanol. Nevertheless, both routes require a previous
transesterification step of diethyl carbonate with 1-hexanol for
ethyl hexyl reaction to proceed by the mentioned pathways. This
reaction is favoured in gel-phase zones of 0.4 nm nm−3 chain
polymer density (equivalent to 2.6 nm diameter pores).
Unfortunately, the main by-product, di-n-hexyl ether, is also
preferentially formed in polymer zones of the same density.
Therefore, it would be significantly formed together with
ethyl hexyl ether. However, both ethers are excellent diesel
additives and they can be introduced into the diesel pool
jointly without further separation.
On the contrary, the thermal decomposition of diethyl
carbonate is favoured in slightly denser polymer zones
(0.8 nm nm−3) and the formation of C6 alkenes only takes place
in zones with severe steric limitations (1.5 nm nm−3 zone).
Both products are less desired as diesel additives. Thus,
the synthesis of ethyl hexyl ether should be carried out on
polymers with medium to low dense polymer fractions.
To the best of our knowledge, ion-exchange resins with a
swollen gel phase with only a zone of low polymer density are
not currently available. Therefore, acid ion-exchangers with a
high swollen volume fraction of 0.4 nm nm−3 polymer density
are selected for industrial production, e.g. Dowex 50W×2 or
Amberlyst 70.
As seen in Table 7, the lowest temperature dependence
was observed in the desired transesterification of diethyl car-
bonate to ethyl hexyl carbonate (Ea,EHC = 68–78 kJ mol−1)
whereas the highest dependence was found in the side reac-
tion of intramolecular dehydration to C6 alkenes (Ea,alkenes
136–143 kJ mol−1). Ether formation reactions showed interme-
diate temperature dependencies (Ea,EHE ≈ Ea,DEE ≈ Ea,DHE
=
≈
98–103 kJ mol−1). These apparent activation energy values are
in line with the literature which shows that in the dehydration
of C5–C12 linear alcohols, olefin formation is favoured at
high temperatures but that of linear ether is favoured at low
temperature.12 Accordingly, activation energy for alcohol
dehydration to olefin is higher than that of linear ether for-
mation. On the other side, transesterification reactions are
much less sensitive to temperature as it is also shown in the
transesterification of ethyl acetate with methanol on several
ion-exchange resins (48–52 kJ mol−1).32
On a technical standpoint, given that the formation of the
different types of compounds (carbonates, ethers and alkenes)
showed different activation energy values, an important task
in order to implement an industrial process for obtaining
ethyl hexyl ether by reaction between diethyl carbonate and
1-hexanol is the selection of a working temperature range.
This is essential to produce readily and selectively ethyl hexyl
ether, as data in Table 6 shows, since an increase in tempera-
ture enhances the reaction rate but decreases selectivity to
ethyl hexyl ether. This behaviour agrees with the fact that acti-
vation energy values for ether formation are intermediate
between those of transesterification to ethyl hexyl carbonate
and dehydration of 1-hexanol to olefin. Since in industry, selec-
tivity has priority over conversion, it follows that it is preferable
working at low temperature. A relevant drawback of conducting
the reaction at low temperature would be the requirement of
a higher amount of the catalyst and/or a longer reaction time
(or reactor volume, in continuous processes). However, energy
Finally, a decrease in the reaction temperature would favour
the reaction between 1-hexanol and diethyl ether to ethyl hexyl
ether that showed lower temperature dependence (EaEHC
=
67–80 kJ mol−1), while all the non-desired reactions showed
higher temperature dependence (Ea = 91–136 kJ mol−1). This
fact would be useful for designing the reactor unit of a poten-
tial industrial process to enhance the selectivity to ethyl hexyl
carbonate and further, that of ethyl hexyl carbonate decompo-
sition to ethyl hexyl ether.
2248 | Catal. Sci. Technol., 2015, 5, 2238–2250
This journal is © The Royal Society of Chemistry 2015

View Article Online
Catalysis Science & Technology
Paper
2 R. O. McClellan, Annu. Rev. Pharmacol. Toxicol., 1987, 27,
279–300.
3 J. Kagawa, Toxicology, 2002, 181–182, 349–353.
4 M. Riedl and D. Diaz-Sanchez, J. Allergy Clin. Immunol.,
2005, 115, 221–228.
Nomenclature
[H+]
DEC
DEE
acid capacity (mmol H+ gcat
diethyl carbonate
)
−1
diethyl ether
DHE di-n-hexyl ether
5 S. H. L. Yim and S. R. H. Barrett, Environ. Sci. Technol.,
2012, 46, 4291–4296.
dp
Ea
DVB
EHC
EHE
particle diameter (mm)
apparent activation energy (kJ mol−1
divinylbenzene
)
6 F. Liotta and D. Montalvo, SAE Tech. Pap. Ser., 1993, 932734.
7 C. Y. Choi and R. D. Reitz, Fuel, 1999, 78, 1303–1317.
8 C. C. Barrios, C. Martín, A. Domínguez-Sáez, P. Álvarez,
M. Pujadas and J. Casanova, Fuel, 2014, 132, 93–100.
9 E-diesel (oxygenated diesel), http://www.oxydiesel.com, 2014.
10 B. G. Harvey and H. A. Meylemans, J. Chem. Technol.
Biotechnol., 2011, 86, 2–9.
ethyl hexyl carbonate
ethyl hexyl ether
HeOH 1-hexanol
ISEC inverse steric exclusion chromatography
i
reaction
j
k
reactant
product
11 B. Martin, P. Aakko, D. Beckman and N. Del Giacomo, SAE
Tech. Pap. Ser., 1997, 972966.
n
nj
P
R
r0i,n
gel-resin zone characterized by a given chain density
number of moles of compound j (mol)
pressure (MPa)
12 R. J. J. Nel and A. de Klerk, Ind. Eng. Chem. Res., 2009, 48,
5230–5238.
13 D. C. Rakopoulos, C. D. Rakopoulos, E. G. Giakoumis and
A. M. Dimaratos, Energy, 2012, 43, 214–224.
14 B. Bailey, J. Eberhardt, S. Goguen and J. Erwin, SAE Tech.
Pap. Ser., 1997, 972978.
15 İ. Sezer, Int. J. Therm. Sci., 2011, 50, 1594–1603.
16 K. Górski and M. Przedlacki, Energy Fuels, 2014, 28, 2608–2616.
17 M. Karabektas, G. Ergen and M. Hosoz, Fuel, 2014, 115,
855–860.
molar ratio (mol mol−1)
contribution to the initial reaction rate of reaction
−1
i induced by polymer fraction n (mol kgcat h−1)
rj0
rate of formation of compound j (mol kgcat h−1)
−1
ri0
initial reaction rate of reaction i (mol kgcat h−1)
−1
Sjk
t
selectivity to k with respect to reactant j (mol %)
time (h)
T
Tmax
temperature (K)
maximum operating temperature (K)
18 R. Bringué, E. Ramírez, M. Iborra, J. Tejero and F. Cunill,
J. Catal., 2013, 304, 7–21.
19 G. Olah, US Pat., 5,520,710, 1996.
20 A. J. Parrott, R. A. Bourne, P. N. Gooden, H. S. Bevinakatti,
M. Poliakoff and D. J. Irvine, Org. Process Res. Dev., 2010, 14,
1420–1426.
21 J. Guilera, R. Bringué, E. Ramírez, M. Iborra and J. Tejero,
Ind. Eng. Chem. Res., 2012, 51, 16525–16530.
22 P. Tundo, S. Memoli, D. Herault and K. Hill, Green Chem.,
2004, 6, 609–612.
TOF0i,n the initial specific activity of the polymer fraction
n for reaction i (mol (meq H+)−1 h−1)
Vsp
Vsp,n
specific volume of swollen polymer (cm3 g−1)
specific volume of swollen polymer occupied by
fraction n (cm3 g−1
)
Wcat
Xj
mass of the dry catalyst (g)
conversion of compound j (mol.%)
YEj HE yield of ethyl hexyl ether with respect to reactant
j (mol.%)
23 R. Bringué, E. Ramírez, M. Iborra, J. Tejero and F. Cunill,
Chem. Eng. J., 2014, 246, 71–78.
[H+]
acid capacity (mmol H+ g−1)
24 X. Gao, A. Krzywicki, D. S. R. Johnston and A. Kalivoda,
US Pat., 7,576,250, 2009.
25 R. Bringué, J. Tejero, M. Iborra, C. Fité, J. F. Izquierdo and
F. Cunill, J. Chem. Eng. Data, 2008, 53, 2854–2860.
26 M. A. Harmer and Q. Sun, Appl. Catal., A, 2001, 221, 45–62.
27 R. Bringué, M. Iborra, J. Tejero, J. F. Izquierdo, F. Cunill,
C. Fité and V. Cruz, J. Catal., 2006, 244, 33–42.
28 P. F. Siril, H. E. Cross and D. R. Brown, J. Mol. Catal. A:
Chem., 2008, 279, 63–68.
Acknowledgements
Financial support was provided by the Science and Education
Ministry of Spain (project: CTQ2010-16047). The authors
thank Rohm and Haas France and Purolite for providing
Amberlyst and CT ion-exchange resins, respectively. Finally, the
authors would like to express their gratefulness to Dr. Karel
Jerabek of the Institute of Chemical Process Fundamentals
(Prague, Czech Republic) for discussing the ISEC analyses.
29 S. Koujout and D. R. Brown, Thermochim. Acta, 2005, 434,
158–164.
30 J.-A. Dalmon and G. A. Martin, J. Chem. Soc., Faraday Trans.,
1979, 75, 1011–1015.
31 K. Jerabek, Kem. Ind., 2013, 62, 171–176.
32 E. Van de Steene, J. De Clercq and J. W. Thybaut, Chem. Eng. J.,
2014, 242, 170–179.
33 Y. Bing, C. Hailin, Y. Hua, L. Xuesong, J. Mingming and
D. Wang, J. Nanosci. Nanotechnol., 2014, 14, 1790–1798.
References
1 S. Salvi, A. Blomerg, B. Rudell, F. Kelly, T. Sandstrom,
S. T. Holgate and A. Frew, Am. J. Respir. Crit. Care Med.,
1999, 159, 702–709.
This journal is © The Royal Society of Chemistry 2015
Catal. Sci. Technol., 2015, 5, 2238–2250 | 2249

View Article Online
Paper
Catalysis Science & Technology
34 J. Guilera, R. Bringué, E. Ramírez, M. Iborra and J. Tejero,
Appl. Catal., A, 2012, 413–414, 21–29.
47 L. Rios, D. Echeverri and F. Cardeño, Ind. Crops Prod.,
2013, 43, 183–187.
35 J. H. Badia, C. Fité, R. Bringué, M. Iborra and F. Cunill, 16th
Nord. Symp. Catal., 2014.
36 J. Guilera, R. Bringué, E. Ramírez, C. Fité and J. Tejero,
AIChE J., 2014, 60, 2918–2928.
37 A. G. Ogston, Trans. Faraday Soc., 1958, 54, 1754.
38 K. Jerábek and K. Setínek, J. Polym. Sci., Part A: Polym.
Chem., 1990, 28, 1387–1395.
39 Y. Yao and A. M. Lenhoff, J. Chromatogr. A, 2004, 1037, 273–282.
40 A. A. D'Archivi, L. Galantini, A. Panatta, E. Tettamanti and
B. Corain, J. Phys. Chem. B, 1998, 102, 6774–6779.
41 I. Bacskay, A. Sepsey and A. Felinger, J. Chromatogr. A,
2014, 1339, 110–117.
42 K. Jerabek, ACS Symp. Ser., 1996, 635, 211–224.
43 K. Jerabek and L. Holub, in 13th International IUPAC
Conference on Polymers and Organic Chemistry, Montreal, 2009.
44 S. Sterchele, P. Centomo, M. Zecca, L. Hanková and
K. Jeřábek, Microporous Mesoporous Mater., 2014, 185, 26–29.
45 Z. Zhang, Q. Wang, P. Tripathi and C. U. Pittman Jr, Green
Chem., 2011, 13, 940–949.
48 W. Schutyser, S.-F. Koelewijn, M. Dusselier, S. de Vyver,
J. Thomas, F. Yu, M. J. Carbone, M. Smet, P. Van Puyvelde,
W. Dehaen and B. F. Sels, Green Chem., 2014, 16, 1999–2007.
49 P. Barbaro and C. M. Marrodan, Green Chem., 2014, 16,
3434–3438.
50 L. Petrus, E. J. Stamhuis and G. E. H. Josteen, Ind. Eng.
Chem., Proc. Res. Develop., 1981, 20, 366–371.
51 J. Guilera, E. Ramírez, C. Fité, M. Iborra and J. Tejero, Appl.
Catal., A, 2013, 467, 301–309.
52 E. Medina, R. Bringué, J. Tejero, M. Iborra and C. Fité, Appl.
Catal., A, 2010, 374, 41–47.
53 C. Casas, R. Bringué, E. Ramírez, M. Iborra and J. Tejero,
Appl. Catal., A, 2011, 396, 129–139.
54 M. A. Pérez, R. Bringué, M. Iborra, J. Tejero and F. Cunill,
Appl. Catal., A, 2014, 482, 38–48.
55 M. Granollers, J. F. Izquierdo, J. Tejero, M. Iborra, C. Fité,
R. Bringué and F. Cunill, Ind. Eng. Chem. Res., 2010, 49,
3561–3570.
56 R. Bringué, E. Ramírez, C. Fité, M. Iborra and J. Tejero, Ind.
Eng. Chem. Res., 2011, 50, 7911–7919.
57 C. Casas, Ph. D. Thesis, University of Barcelona, 2013.
46 I. Dosuna-Rodríguez and E. M. Gaigneaux, Catal. Today,
2012, 195, 14–21.
2250 | Catal. Sci. Technol., 2015, 5, 2238–2250
This journal is © The Royal Society of Chemistry 2015