This thesis has shown that the energy requirements of ethanol purification can be mitigated
by blending a partially purified ethanol mixture directly into petrol as opposed to fully
purifying the ethanol first.
In Chapter 2 this concept was verified using rigorous simulation based on phase equilibrium
measurements taken from literature [25]. A process flowsheet was developed to illustrate this
concept, and based on simulations a simple process using two-stage counter-current liquid-
liquid extraction proved to be sufficient to produce a fuel mixture of desirable ethanol
content, while recovering 98% of ethanol feed in the fuel phase.
This process represents a significant cost-saving when compared to conventional processes,
since it foregoes the energy required for final purification. However, the process flowsheet
from Chapter 2 does not necessarily constitute a fully optimized process. Further, the
optimality of any particular version of the process is contingent on the circumstances in
121 In light of this, Chapter 3 investigated the process in greater detail, examining the effects of
various operating parameters and looking more deeply at the requirements of two different
contexts for implementation: the South African market, where 2% ethanol blends have been
mandated for use by new legislation and the United States market, where 10% ethanol blends
are already in widespread use.
In that chapter, different versions of the bioethanol blending process were examined
specifically for their suitability in those two contexts.
It must be very clearly noted, however, that the ethanol pre-blending concept creates a large
optimization space for the design of bioethanol production processes, and this thesis has
explored only a small portion of that optimization space. I am confident that there are a
number of possible refinements to achieve better performance using this concept and to
integrate it into other creative approaches to bioethanol separation.
One example of such a possibility is a setup resembling that of heteroazeotropic distillation
where the phase split and distillation both occur within the same unit. Such a setup could
potentially achieve a high recovery of ethanol into a fuel mixture, without prior partial
purification as is necessary for the process presented in this thesis. Another exotic possibility
is a reactive distillation unit with phase-split occurring within the column.
These and numerous other possibilities are yet to be investigated, and there may be other
significant advantages to them. For instance, a process integrating the liquid-liquid phase split
into another form of separation could potentially benefit from the phase split in terms of
energy consumption, while still producing an under-saturated fuel mixture and thereby
providing higher stability and avoiding the complications that result from dealing with a
122 According to the South African Petroleum Industry Association [38], in 2009 national petrol
useage was 11.3 billion liters, with steady increases in that figure expected. This means that
the 2% ethanol content mandated by legislation corresponds to at least 226 million liters of
ethanol to be blended per year. With the process presented in chapter 2 reducing energy
useage per liter by between 0.916MJ and 2.13MJ that places the potential energy saving
nationwide at between 2.07x108 MJ and 4.81 x108 MJ per annum, approximately equivalent
to the household electricity consumption of between ten and twenty thousand average
households.
While this is a significant sum to a nation suffering an energy crisis, it is trivial in comparison
to the potential implications in the United States and elsewhere in the world. This thesis has
addressed the South African context, as that is where this university is situated and therefore
is the market with the most immediacy and relevance. I have also considered the context of
the United States not only because they are the single largest ethanol producer but also for
reasons of convenience. Their biofuels legislation has established highly uniform ethanol
content, which creates a clear target when synthesizing a new process such as that presented
in Chapter 2. Brazil, for instance, permits wide-ranging ethanol content in fuel. Without a
clear target for ethanol content, process viability is more difficult to assess, and optimization
becomes more complex because of the additional degree of freedom. Having clear-cut targets
for ethanol content allows for a straightforward demonstration of the viability and value of
the proposed process.
That is not to say, however, that the concepts and flowsheets presented in this thesis are any
less applicable elsewhere in the world. Flexibility in terms of ethanol content could in fact
make our specific process more attractive. Any specified ethanol content more or less dictates
the blending ratio that must be used in this process, constraining the design within narrow
123 designer the flexibility needed to best implement the concept of using phase equilibrium to
assist separation.
Furthermore, the development of new processes using this concept need not adhere to any
particular specification of ethanol content. Since the particulars of any such process will tend
to be highly dependent on the required ethanol content, this flexibility offers a large design
space for creative designs.
While our specific process design offers immediate benefits, the true potential of the core
concept is yet to be explored. It is optimistic to believe that the implementation of this simple
concept will change the face of the bioethanol industry, but the findings of this thesis suggest
that it is possible.