The key parameters for selecting suitable ILs are the
availability of VLE data and selectivity. Research studies on the extractive
distillation of ethanol and water have involved a pool of various ILs,
particularly imidazolium-based ones 8 28. However, neither their
comparison of performances in vapour-liquid laboratory experiments, process
simulations, or pilot scale experiments produces a conclusive choice of IL as the
best entrainer. Different papers are contradictory of one another in the
selection of suitable IL, as they are judged using different solvent criteria. Using
the available comparisons and results from literature, a fair choice of IL is
made based on the overall analysis, with the following justifications.
The first step in the selection is to ascertain the
availability of VLE data of the ILs for ternary system involving ethanol and
water. The respective VLE data of ILs that is available in literature is
compiled by Figueroa et al. 29.
Among all the possible choices of ILs for ethanol dehydration, the most frequently
discussed cations for are Emim+ and Bmim+, whereas the
anions are BF4-, Cl-, OAc-.
The second step for IL screening is to compare the
selectivities at infinite dilution between the ILs, which is reported by
several papers 8 27 17.
ILs with higher selectivity will lead to a greater difference in volatility
between the components in the feed, therefore increasing the separation. With a
greater separation, the operation would require a lower entrainer mass and
number of column stages, which reduces the operating and capital costs 14.
As inferred by Ge et al. 17, the selectivity
depends on both the cation and anion of the IL. The larger the cation size, the
lower the difference in relative volatility. By comparing the relative
volatility between ethanol and water using Emim+ and Bmim+,
Ge et al. found that the enhancement effect of Emim+ was larger 17. This was further
supported by a more recent study by Pereiro et al. 11, which investigated
three cations, namely (in ascending order of chain length) Emim+,
Bmim+ and Hmim+, with Cl- as cation. The
same result was found and attached in Figure
where the relative volatility decreased with size of cation at mole fractions of
IL above 0.03. Furthermore, when the molar fraction of IL increased, there was
a greater enhancement in relative volatility.
Figure 3: Relative
volatility of chlorinated ILs with different cations at different IL molar
Figure 4: Relative
volatility of ILs 11.
As for anions, they have a greater impact on the relative
volatility than cations 17
The relative volatility between ethanol and water was plotted against the
different ILs studied by Pereiro et al. 11
EmimCl had the highest selectivity among the ILs listed in the figure, followed
by EmimOAc. This was also supported by other VLE study, by Ge et al. 17. Both papers reported
results of the relative volatility enhancement effect by the ILs to be in this
descending order: Cl-, OAc-, BF4, paired
with cations Emim+ or Bmim+. However, the viscosities
of chlorinated ILs are high, and would have the mass transfer resistance,
leading to the poor mass and heat transfer performance in separation units.
This would also incur a high operating cost as more energy would be spent to overcome
these resistances in agitation and other operations 8. In addition to
that, chlorinated ILs are also corrosive 28.
Therefore, they are not chosen.
The IL with the second highest selectivity is EmimOAc,
and is claimed to be a promising entrainer due to its lower viscosity compared
to EmimCl 17.
However, several other papers have disputed its use because of its thermal
instability 8 12. It is has a low
flash point of 164°C and is therefore infeasible to be
recycled under high temperature conditions. Meindersma et al. 27 explained that the
strong attachment of the acetate IL to water would cause the separation to be
difficult during solvent recovery. The flash drum would need to operate at high
vacuum conditions (0.1 mPa in Meindersma et al.’s study) in order for the
recovery to be feasible 27.
This would incur a high energy demand, which leads to costly operation.
Arlt et al. was the pioneer for the research of ILs in ethanol
dehydration. Their patent for BASF was published in 2004, concluding that it was
feasible to break the azeotrope using EmimBF4 as the entrainer. Since
then, EmimBF4 has become a popular choice to be used in multiple
studies for the process design of the ethanol dehydration and comparison with
other solvents 8 13.
This is because EmimBF4 has favourable characteristics over
other ILs such as a high thermal stability (up to 450°C), high selectivity and is commercially available 24. Hence, EmimBF4
is selected as the IL to be studied in this dissertation.
This section of the literature review focuses on two
aspects, the feasibility of EmimBF4 in ethanol dehydration and
the process economics. The IL is evaluated to determine the achieved purity
under set operating parameters, and its feasibility in meeting the baseline
purity in industry. As for the economics, the differences in the capital costs
and the utilities cost are analysed, with comparison to conventional solvents.
EmimBF4 was used in the research by Seiler
et al. to carry out process simulation and optimisation of the extractive
distillation of ethanol and water, with comparison to 1,2-ethanediol 13. The extractive
column was operated at Patm whereas the solvent recovery unit
consisted of a flash vessel operating at 10 kPa, and was connected to a
stripping column operating at Patm. The results showed that the
overall heat consumption for IL was about 25% lower than 1,2-ethandiol at the
same purity of ethanol (99.8 mol%). The capital cost was not taken into account
in the economic analysis. Nevertheless, this 13-year-old study has presented
the feasible design of the IL in the extractive distillation process and its
energy advantages. It provided a good base for future work such as exploring
more aspects of economic analysis and more optimisation parameters.
A more recent study was done by Meindersma et al. in 2016 27, which compared the
energy consumption by different types of configuration, where the solvent
recovery unit consisted of either flash drum only or flash drum and stripping
column. The configuration with only flash drum required about 4% lower in energy
than flash drum+stripping column. As the comparison was done under the same set
of operating conditions, this configuration with flash drum appears to be well
validated to be the superior choice. The process configuration was modified and
in accordance with other literature, such as Zhu et al. 8.
Zhu et al. focused on the process design and optimisation
of the ethanol dehydration using different solvents, including EmimBF4.
The general process design consists of two units, the extractive distillation
column and the flash drum for solvent recovery. The IL is added to the
extractive distillation column along with the ethanol and water feed. The top
product obtained would be the anhydrous ethanol whereas the bottom product would
contain the mixture of IL and water. The bottom product is subsequently fed
into the solvent recovery unit. As ILs have negligible vapour pressure and do
not vaporise, a conventional distillation column cannot be used to recover IL. Instead,
it is recovered using a flash vessel instead under vacuum conditions (P