Distillation Fundamentals
Distillation is still the most widely used method to separate chemical species in the process industries.
It’s therefore probably worthwhile revisiting the fundamentals of this process, for the benefit of the current new generation of chemical and process engineers, who otherwise would rely exclusively on computer simulation tools without having gained the insights of the previous generation, who perhaps did not have the ‘benefit’ of all these tools early in their careers. These insights give a better understanding of how to configure a distillation process; what features can improve energy efficiency and when to use them, and how to get a computer solution more quickly – avoiding the blind ‘sledge-hammer to crack a nut’ approach.
Its a fairly big subject and will start with the basic concepts of phase equilibrium and build on this foundation.
The distillation operation that may come to mind – for some of us – is production of whisky or other spirits which comprise up to about 95% ethyl alcohol or ethanol (C2H5OH) from a weak mixture of alcohol and water – which has been produced by fermentation of a carbohydrate (malt or other grain) with the help of yeast. (This weak mixture is called ‘beer’ – even in the whisky industry, but is not drunk as the beer we know and love? (no hops).
The ethanol content of the ‘beer’ boils or vapourises more easily than does its water content, so the vapour leaving the ‘still’ is enriched in ethanol. It may be condensed to form a ‘spirit’. This is called a single stage separation.
It relies upon the differences in ‘volatility’ (relative tendency to escape from the liquid into the vapour phase) between the components in a mixture. By extending this simple one step process into a series of steps, each one taking either the vapour from the basic single stage separation and cooling it to partially condense it, or the liquid and partially boiling it, we can achieve a further extent of separation, and so on for several steps or ‘stages’. In practice and to make it easier, we normally only cool and condense vapour from the very top or 1st stage and only supply heat to boil the liquid at the bottom stage. The condensed liquid from the top stage passes (physically downward by gravity) to mix with fluids on the next-to-top stage and so on. Similarly the vapour boiled up in the bottom stage passes up to the penultimate stage where it mixes with the liquid there – causes some its light (alcohol) to vapourise and leaves to the next stage above and so on.
The stages or trays are horizontal and mounted in a vertical cylindrical shell. The trays may be anything from a few cm apart to a metre or more depending on the application.
Each of the stages is typically a flat tray, with many holes or passages to allow the vapour to bubble through and contact with the tray liquid, and ‘downcomers’ for the contacted liquid to leave the tray and descend to the next tray down. Liquid typically flows across the tray from an incoming downcomer onto the tray and across to the outlet downcomer. Therefore the vapour and liquid are normally therefore said to be in cross-flow – similar to the cross-flow heat transfer in an air cooled heat exchanger.
This multi stage ‘cascade of steps’ is basically what is typically found in a distillation process, where the ‘lightest’ or most volatile component(s) (in this case ethanol) becomes concentrated at the topmost stage and the ‘heavier’ or least volatile (in this case water) reaches a highest concentration at the bottom.
In any particular stage, when the liquid from the stage above and the vapour arriving from the stage below are mixed together and come to perfect equilibrium, we call that an ideal equilibrium stage.
In practice, it is not quite achieved, as some liquid may bypass and leave the stage without ‘seeing’ any vapour, or some liquid drops may carry upward with the vapour to the next tray. To approach the ideal efficiency of a perfect equilibrium stage the tray hydraulic designers have a few clever tricks and rules.
Reverting to the example of alcohol-water distillation, for the process to make sense, fresh feed (‘beer’) from the fermenation step is continuously fed to the column, usually somewhere in the middle and alcohol rich liquid is drawn off the top condenser and water concentrate drawn off the bottom.
This is called a continuous process. However, in the old days whisky was produced via a ‘batch’ process, where the distillation ‘still’ was filled with beer, brought to the boil, and the alcohol-rich vapour was steadily drawn off and condensed to produce the ‘moonshine’? - until the charge to the still was sufficiently depleted and mostly water was left.
People of a certain generation may recall the Joan Baez folk song ‘Copper Kettle’.
Copper was traditionally used in the still as it reacted with and removed the traces of suphur present in the grain and which would otherwise produce an unpleasant flavour.
A word of warning, the preceding fermentation process produces various other alcohols and other components such as aldehydes and esters. Some of these guarantee a hangover and others such as methyl alcohol or methanol are deadly, so at least one additional preceding column is needed to remove these components prior to the main column, so that its ‘distillate’ (about 95% ethyl alcohol) can be relatively safely used after dilution to produce whisky.
The process is in outline what happens in any typical distillation separation . Alcohol-water is called a binary (2-component) system and requires one column, but the same approach may be applied to a multi-component system. In general at least one additional column is needed for each component beyond two to be recovered in a fairly pure state.
For further study by the curious, the alcohol-water binary forms what is called a minimum boiling azeotrope, at an alcohol content of 95.6%. What this means is that beyond 95.6% alcohol, the alcohol no longer behaves as the most volatile component, but this duty flips over to water. So without doing something clever, there is a road-block preventing us from distilling to more than about 90-95% maximum ethanol content.