Distillation Improvement Opportunities Part 4: Hybrid Schemes and Analysis

Izak Nieuwoudt reviews the hybrid schemes that can help process engineers reduce the energy use of their distillation processes

DISTILLATION is the most prevalent technique for separating organic chemicals, but the energy consumption is significant – especially for mixtures with low relative volatility. There are schemes that chemical engineers can employ to reduce this energy consumption and in turn reduce the carbon emissions of the plants they design and operate.

So, what are the separation hurdles we are looking to overcome? The relative volatility of a system can be low due to vapour pressures being similar and/or because of attractive/repulsive forces between molecules. Separating these components with conventional distillation schemes may require the use of a large number of stages and a high reflux ratio, which means high energy consumption per unit of product. The attractive/repulsive forces may lead to the formation of azeotropes, which makes it impossible to produce pure components with simple distillation schemes. As we will see below, there are several enhanced and hybrid distillation schemes that we can use to address these separation challenges.

Extractive distillation

In extractive distillation the aim is to select a high boiling solvent, or solvent mixture, that preferentially associates with one or more of the components without forming azeotropes. This association will render these components less volatile and facilitate their separation from the other components. This association reduces the volatility of the components so that they exit the tower at the bottom with the low volatility solvent. The components that do not associate with the solvent exit the tower as the distillate product. The components that associate with the solvent can be distilled off from the high boiling solvent. Due to the low volatility of the solvent, and the absence of azeotropes, it is easily separated from the product and is recycled without having to evaporate the solvent, which significantly benefits the energy consumption. Extractive distillation not only applies to systems with low relative volatility, but also to systems where components in the feed form azeotropes. In the case where feed components form azeotropes, the association between the solvent and some of the components can change the system so that the azeotrope does not form. The energy consumption of an extractive distillation system can be significantly lower than that of conventional distillation systems. When setting up an extractive distillation system it is important to consider the following:

• The solvent should be chemically and thermally stable since it will be subjected to a very large number of cycles
• You should plan to clean up a slip stream of the solvent. This includes the removal of high boiling or non-volatile components
• Choose a solvent that maximises the relative volatility between the key components
• Ensure that the boiling point of the solvent is high enough that it can be kept from going overhead
• Ensure the solvent will not form azeotropes with any of the components
• Consider the melting point of the solvent
• Select a solvent that does not pose significant environmental, health and safety risks
• It should be possible to regenerate the solvent to a degree that does not cause the recycle of the component that you planned to remove to the extractive distillation tower
• Computer-aided molecular design (CAMD) methods are available^{1, 3, 5, 7} that can help you select a solvent or solvent mixture. In some instances, solvent blends can significantly outperform pure solvents
• The relative volatility can typically be increased by increasing the solvent concentration in the column (higher solvent flowrate). You must optimise the solvent to feed ratio, equipment sizing, and regeneration energy demand by using process simulation software
• Reflux in the extractive distillation tower is primarily used to prevent the solvent from exiting with the overhead product. Increased reflux can have a negative effect since it dilutes the solvent in the tower and reduces the relative volatility. This effect is system dependent, and you must use process simulation to determine the optimum reflux ratio
• You should also use process simulation to determine the optimum location of the feed point, which is also system dependent
• Industrial experience has shown that the tray efficiency or packing height equivalent to theoretical plate (HETP) in extractive distillation towers are poorer than in conventional systems. The vapour to liquid ratio and the effect of the solvent on mass transfer coefficients could be responsible for this effect. It is best to determine the tray efficiency or packing HETP in pilot plant studies
• Group contribution methods and using the pair-wise-additivity assumption with binary vapour-liquid equilibrium (VLE) data typically gives an unsatisfactory description of the equilibria in extractive distillation systems and should not be used for finals designs. The preferred approach is to fit activity coefficient models to multi-component equilibrium data that includes the solvent and the components that need to be separated
• Energy integration schemes, including dividing wall arrangements, can be used to further reduce the energy consumption of extractive distillation units²

A simple energy-integrated extractive distillation flowsheet is shown in Figure 1.

Case study one

The separation of ethanol and 2-propanol is challenging. At atmospheric pressure the relative volatility of the system is approximately 1.18.

Van Dyk³ showed that the addition of a high-boiling non-polar solvent can increase the relative volatility to 1.36. Compared to a conventional distillation unit with 100 stages requiring 100 units of energy, an extractive distillation unit only requires 87 stages and 53 units of energy. This technology was successfully commercialised by Sasol⁴.

Case study two

With a relative volatility of 1.08 at atmospheric pressure, the separation of n-propanol and 2-butanol with conventional distillation is very capital and energy intensive. For a particular separation, 200 stages and 100 units of energy are required in a conventional distillation arrangement. The addition of a high-boiling non-polar solvent increased the relative volatility to 1.31 and the separation required 80 stages and 41 units of energy. This technology was successfully commercialised by Sasol⁴. It is also important to note that the VLE of the multi-component system could not be predicted by combining sets of VLE parameters regressed on binary data. This is most likely the result of the liquid activity coefficient parameters being cross correlated, resulting in multiple combinations of parameters giving the same fit to the data. This was resolved by measuring multi-component VLE data and simultaneously regressing the multi-component and binary VLE data while ensuring that the regression does not get stuck in a local optimum. This approach should be followed for all solvent-driven separations since the molecular interactions are too complex to expect pair-wise additivity of binary systems to accurately describe the multi-component interactions.

Case study three

The system acetone-methanol forms a homogenous azeotrope containing approximately 80 mole% acetone at atmospheric pressure. This means that conventional distillation cannot be used to separate these components. Water, being a highly polar solvent, interacts with methanol more strongly than acetone¹. This association, and the boiling point of water, “drags” the methanol to the bottom of the tower and acetone is recovered as the distillate. This association is strong enough to “break” the acetone-methanol azeotrope. This is shown in Figure 2.

In this case extractive distillation turns a separation that is not feasible with conventional distillation into a separation that is not only feasible, but where the relative volatility is increased, and the energy consumption is very attractive. The water does not have to be boiled, so that the high latent heat of water does not affect the energy consumption. Energy integration schemes, including dividing wall technology, can also be used to minimise the energy consumption of this extractive distillation separation².

Azeotropic distillation

In some chemical systems the interaction between components is such that low boiling or high boiling homogenous azeotropes form. In homogenous azeotropic systems, conventional distillation will yield a low boiling azeotropic mixture as distillate, or the high boiling azeotrope as the bottoms product. A solvent can be introduced into the system that will form an azeotrope that is rich in one of the components and where the boiling point of this solvent azeotrope is significantly different from that of the homogenous azeotrope formed by the components in the feed. This allows a distillation column to separate the solvent azeotrope from the rest of the feed. It is important that the solvent azeotrope, when condensed and cooled down, separates into two liquid phases with very different compositions. This allows us to recycle the solvent to the first column and concentrate the product in a second column. A classic example of azeotropic distillation is the separation of ethanol and water by using cyclohexane as a solvent. A simplified arrangement is shown in Figure 3. The boiling point of the water-cyclohexane azeotrope is lower than that of the ethanol-water azeotrope. The overheads of the first column is thus approximately the water-cyclohexane azeotrope and the bottoms is the purified ethanol. When this azeotrope is condensed and cooled down it splits into two liquid phases. The light phase is rich in cyclohexane and is returned to the first tower as reflux to continuously remove water from the feed. The heavier phase is rich in water and is sent to a second column where the water-cyclohexane azeotrope is removed overheads and pure water is the bottoms product.