Jul-2001

Working with an extractive distillation process

A cost comparison between progressive extractive distillation of aromatics versus a more conventional technology has shown there can be considerable savings in operating costs during the production of high purity benzene

Gerd Emmrich, Helmut Gehrke & Uwe Ranke
Krupp Uhde GmbH

Article Summary

In the mid-60s, Krupp Uhde developed the Morphylane process (extractive distillation process) originally for the recovery of high-purity benzene from hydro-refined coke oven benzole, the reason being that the existing liquid-liquid methods of extraction were unable to process coke oven benzole due to its high aromatics content. The liquid-liquid extraction techniques exploit the different solubilities of aromatics and non-aromatics in a polar solvent (Figure 1).

Two liquid phases are formed: the aromatics, which tend towards the solvent phase, usually the heavy phase, due to their higher polarity and thus higher solubility, and the light phase, which consists mainly of non-aromatics.

The extraction process essentially consists of four process units:
- Extractor, where the feedstock is brought into contact with the solvent
- Extractive stripper, where the non-aromatics which have also been dissolved are stripped off
- Solvent stripper, where the aromatics are separated from the solvent
- Raffinate washer for the recovery of the solvent residue which has been dissolved in the raffinate.

It is obviously a very long-winded process. However, if the aromatics content of the feed material is too high, formation of two phases does not take place as the aromatics act as a solubility agent for non-aromatics. As a result, extractive separation of the aromatics and non-aromatics is not possible because the entire feed material dissolves in the solvent.

Polar solvents, as commonly used for this purpose, also demonstrate another exceptional physical property in addition to the solubility differences of aromatics and non-aromatics. Their polarity can influence hydrocarbon vapour pressures. They lower vapour pressures of all hydrocarbons contained in the solution, albeit to different degrees. What does this mean, exactly?

The molecular structure of N-formylmorpholine is shown in Figure 2, the solvent used in Krupp Uhde’s proprietary Morphylex and Morphylane processes. The different molecular groups on both sides of the molecule have an electrical effect, making the molecule act like a small dipole.

The electrical effect acts on the double bond of the hydrocarbons. A loose bond is formed impeding the mobility of the molecules concerned. The more double bonds a molecule has, the greater its movement is impeded and the more difficult for it to be converted from a liquid to a vapour state in the presence of such solvents.

How significant are these vapour pressure changes?
Example: the first temperature column in Table 1 shows the boiling points of the pure components such as benzene (80°C), methyl-cyclohexane (MCH) (101°C) and n-heptane (98°C). If these components are now mixed with NFM to a ratio of 15 mol% hydrocarbons:85 mol% NFM, the resulting mixtures have boiling points of 135°C for benzene, approximately 110°C for MCH and 103°C for n-heptane. The reduction of the vapour pressure is dramatic in the case of benzene, while in the case of n-heptane, it is only minimal.

The most important point, however, is that in the pure component mixture benzene is the low boiler, while in the mixture containing the solvent, it is the high boiler. This is the principle of extractive distillation where, at first, a close-boiling distillation cut is produced; this is then mixed with a polar solvent to increase boiling differences between components, thus increasing the relative volatility of the non-aromatics. Azeotropes are also destroyed. The principle arrangement of extractive distillation is very simple (Figure 3).

The extractive distillation column shown in Figure 3 consists of three sections: Stripping section, rectifying section and solvent recovery section.

The feed material, in this case the benzene fraction, is added between the stripping and the rectifying section; the solvent NFM is introduced above the rectifying section. The vapourised benzene is removed in the enrichment zone by the influence of NFM. Some light non-aromatics are also dissolved and stripped off in the stripping section.

Solvent vapours also flow to the top of the column together with non-aromatics, according to their partial pressure. These solvent traces are washed back into the column in the solvent recovery section by non-aromatics reflux. The solvent-benzene mixture which accumulates in the bottom of the ED column is transported to the stripper column, which operates under a slight vacuum. Here the benzene is separated from the solvent.

The NFM in the benzene vapours is washed back by the reflux to less than 1ppm. The condensed benzene meets all product specifications. The solvent, which contains a minimal benzene content, is fed back to the extractive distillation column from the bottom of the stripping column after intensive heat exchange.

In principle, this is a conventional two-column distillation system in which the low-boiling component is taken overhead in the first column, the medium boiling component leaves as overhead in the second column and the high-boiling component accumulates in the bottom of the second column. The only difference is that the high-boiling component is recycled back to the first column as extraction solvent. Separation of a three-component mixture (ABC) into individual components A, B and C, by distillation in direct sequence is shown in Figure 4a. The three components can also be separated using a main column and a side column. Depending on the boiling conditions of component B, either a side stripper or a side rectifier can be used. These are both thermally coupled columns as the two columns are connected with each other at the reboiler or condenser side.

These side columns can also be incorporated in one column by dividing the column into sections. If the two aforementioned methods are combined, a fully thermally coupled column is obtained. If the side column is then incorporated in the main column, the result is the divided-wall column (Figure 4b).