Extractive Distillation

Introduction Extractive distillation is a partial vaporization process in the presence of a miscible, high-boiling, non-volatile massseparation agent, normally called the solvent, which is added to an azeotropic or nonazeotropic feed mixture to alter the volatilities of the key components without the formation of any additional azeotropes. Extractive distillation is used throughout the petrochemical- and chemical-processing industries for the separation of close-boiling, pinched, or azeotropic systems for which simple single-feed distillation is either too expensive or impossible. It can also be used to obtain products which are residue-curve saddles, a task not generally possible with single-feed distillation.

Fig. 13-71 illustrates the classical implementation of an extractive distillation process for the separation of a binary system. The configuration consists of a double-feed extractive column and a solvent-recovery column. The components A and B may have a low relative volatility or form a minimum-boiling azeotrope. The solvent is introduced into the extractive column at a high concentration a few stages below the condenser, but above the primary-feed stage. Since the solvent is chosen to be nonvolatile it remains at a relatively high concentration in the liquid phase throughout the sections of the column below the solvent-feed stage.

One of the components, A (not necessarily the most volatile species of the original mixture), is withdrawn as an essentially pure distillate stream. Because the solvent is nonvolatile, at most a few stages above the solvent-feed stage are sufficient to rectify the solvent from the distillate. The bottoms product, consisting of B and the solvent, is sent to the recovery column. The distillate from the recovery column is pure B, and the solvent-bottoms product is recycled back to the extractive column.

Extractive distillation works by the exploitation of the selective solvent-induced enhancements or moderations of the liquid-phase nonidealities of the components to be separated. The solvent selectively alters the activity coefficients of the components being separated. To do this, a high concentration of solvent is necessary. Several features are essential:

1. The solvent must be chosen to affect the liquid-phase behavior of the key components differently, otherwise no enhancement in separability will occur.

Alcohol Distillation
FIG. 13-71 Typical extracting distillation sequence. Component A is less associated with the solvent.

2. The solvent must be higher boiling than the key components of the separation and must be relatively nonvolatile in the extractive column, in order to remain largely in the liquid phase.

3. The solvent should not form additional azeotropes with the components in the mixture to be separated.

4. The extractive column must be a double-feed column, with the solvent feed above the primary feed; the column must have an extractive section.

As a consequence of these restrictions, separations of binary systems by extractive distillation correspond to only two possible three-component distillation region diagrams, depending on whether the binary system is pinched or close boiling (DRD 001), or forms a minimum-boiling azeotrope (DRD 003). The addition of high-boiling solvents can also facilitate the breaking of maximum-boiling azeotropes (DRD 014), for example splitting the nitric acid-water azeotrope with sulfuric acid. However, as explained in the section on azeotropic distillation, this type of separation might better be characterized as exploitation of extreme boundary curvature rather than extractive distillation, as the important liquid-phase activity-coefficient modification occurs in the bottom of the column. Although many references show sulfuric acid being introduced high in the column, two separate feeds are in fact not required.

Examples of industrial uses of extractive distillation grouped by distillation region diagram type are given in Table 13-19. Achievable compositions in dual-feed extractive distillation columns are very different from the bow-tie regions for single-feed columns. Either pure component (the higher-boiling of which is a saddle) for close-boiling systems, and either pure component (both of which are saddles) for minimum-boiling azeotropic systems can be obtained as distillate.

Extractive distillation is generally only applicable to systems in which the components to be separated contain one or more different functional groups. Extractive distillation is usually uneconomical for separating stereoisomers, homologs, or homology or structural isomers containing the same functional groups, unless the differences in structure also contribute to significantly different polarity, dipole moment, or hydrophobic character. One such counter-example is the separation of ethanol from isopropanol, where the addition of methyl benzoate raises the relative volatility from 1.09 to 1.27 [Berg et al., Chem. Eng. Comm., 66,1 (1988)].

Solvent Effects in Extractive Distillation In the distillation of ideal or nonazeotropic mixtures, the component with the lowest pure-component boiling point is always recovered primarily in the distillate, while the highest boiler is recovered primarily in the bottoms. The situation is not as straightforward for an extractive-distillation operation. With some solvents, the component with the lower pure-component boiling point will be recovered in the distillate as in ordinary distillation. For another solvent, the expected order is reversed, and the component with the higher pure-component boiling point will be recovered in the distillate. The possibility that the expected relative volatility may be reversed by the addition of solvent is entirely a function of the way the solvent interacts with and modifies the activity coefficients and, thus, the volatility of the components in the mixture.

In normal applications of extractive distillation (i.e., pinched, close-boiling, or azeotropic systems), the relative volatilities between the light and heavy key components will be unity or close to unity. Assuming an ideal vapor phase and subcritical components, the relative volatility between the light and heavy keys of the desired separation can be written as the product of the ratios of the pure-component vapor pressures and activity-coefficient ratios whether the solvent is present or not:

where L and H denote the lower-boiling and higher-boiling pure component, respectively.

The addition of the solvent has an indirect effect on the vapor-pressure ratio. Because the solvent is high boiling and is generally added at a relatively high mole ratio to the primary-feed mixture, the temperature of an extractive-distillation process tends to increase over that of a simple distillation of the original binary mixture (unless the system pressure is lowered). The result is a corresponding increase in the vapor pressure of both key components. However, the rise in operating temperature generally does not result in a significant modification of the relative volatility, because the ratio of vapor pressures often remains approximately constant, unless the slopes of the vapor-pressure curves differ significantly. The ratio of the vapor pressures typically remains greater than unity, following the "natural" volatility ofthe system.

Since activity coefficients have a strong dependence on composition, the effect of the solvent on the activity coefficients is generally more pronounced. However, the magnitude and direction of change is highly dependent on the solvent concentration, as well as the liquidphase interactions between the key components and the solvent. The solvent acts to lessen the nonidealities of the key component whose liquid-phase behavior is similar to the solvent, while enhancing the nonideal behavior of the dissimilar key.

The solvent and the key component that show most similar liquidphase behavior tend to exhibit little molecular interactions. These components form an ideal or nearly ideal liquid solution. The activity coefficient of this key approaches unity, or may even show negative deviations from Raoult's law if solvating or complexing interactions occur. On the other hand, the dissimilar key and the solvent demonstrate unfavorable molecular interactions, and the activity coefficient of this key increases. The positive deviations from Raoult's law are further enhanced by the diluting effect of the high-solvent concentration, and the value of the activity coefficient of this key may approach the infinite dilution value, often a very large number.

The natural relative volatility of the system is enhanced when the activity coefficient of the lower-boiling pure component is increased by the solvent addition (yL/yH increases and PLat/PHat > 1). In this case, the lower-boiling pure component will be recovered in the distillate as expected. In order for the higher-boiling pure component to be recovered in the distillate, the addition of the solvent must decrease the ratio YL/yH such that the product of the yL/yH and PLat/PHat (i.e., aLH) in the presence of the solvent is less than unity. Generally, the latter is more difficult and requires higher solvent-to-feed ratios. It is normally better to select a solvent that forces the lower-boiling component overhead.

The effect of solvent concentration on the activity coefficients of the key components is shown in Fig. 13-72 for the system methanol-acetone with either water or methylisopropylketone (MIPK) as solvent. For an initial-feed mixture of 50 mol % methanol and 50 mol % acetone (no solvent present), the ratio of activity coefficients of methanol and acetone is close to unity. With water as the solvent, the activity coefficient of the similar key (methanol) rises slightly as the solvent concentration increases, while the coefficient of acetone approaches the relatively large infinite-dilution value. With methyliso-propylketone as the solvent, acetone is the similar key and its activity coefficient drops toward unity as the solvent concentration increases, while the activity coefficient of the methanol increases.

TABLE 13-19 Examples of Extractive Distillation, Salt Extractive Distillation




Benzene-cyclohexane Ethyl acetate-ethanol


Acetone-methanol Isoprene-pentane Pyridine-water Methyl acetate-methanol

Minimum-boiling azeotrope

Minimum-boiling azeotrope Minimum-boiling azeotrope

Minimum-boiling azeotrope

Minimum-boiling azeotrope Minimum-boiling azeotrope Minimum-boiling azeotrope Minimum-boiling azeotrope

Ethylene glycol, acetate salts for salt process Aniline

Higher esters or alcohols, aromatics Propylene glycol

Water, aniline, ethylene glycol Furfural, DMF, acetonitrile Bisphenol

Ethylene glycol monomethyl ether

Alternative to azeotropic distillation, pressure swing distillation

Process similar for other alcohol-ester systems

Alternative to pressure swing distillation

Element of recovery system for alternative to production of methyl acetate by reactive distillation; alternative to azeotropic, pressure, swing distillation

C4 alkenes/C4 alkanes/ C4 dienes

C5 alkenes/C5 alkanes/ C5 dienes Heptane isomers-

cyclohexane Heptane isomers-toluene

Close-boiling and minimum-boiling azeotropes Close-boiling and minimum-boiling azeotropes Close-boiling

Close-boiling and minimum-boiling azeotropes

Furfural, DMF, acetonitrile, n-methylpyrolidone Furfural, DMF, acetonitrile, n-methylpyrolidone Aniline, phenol

Aniline, phenol

Making Your Own Wine

Making Your Own Wine

At one time or another you must have sent away for something. A

Get My Free Ebook

Post a comment