Calculated azeotropic distillation

The methods used here to give the phase equilibria are reviewed, and the Azeotropic Distillation Program ADP/ADPLLE is described. Application of the program to calculate an azeotropic distillation problem is shown and discussed, and a sample computer output is given and is briefly discussed. Finally, calculated azeotropic distillation results are compared for dehydrating aqueous ethanol for the three entrainers, n-pentane, benzene, and diethyl ether. 

Copolymerizations were performed at 70 C using an ampoule technique similar to that used for MMA. Monomers were purified by distillation. Most of the runs had an initial weight fraction styrene of 0.767 and 1.45 mole % AIBN initiator. Also utilized is one run using 0.235 wt. fraction styrene (0.350 mole % AIBN) and one at 0.557 (1.45 mole % AIBN). Gruber and Knell (10) used both the former compositions. The latter one is the calculated azeotropic composition using their values of the reactivity ratios. 

Levy SG, Van Dongen DB and Doherty MF (1985) Design and Synthesis of Homogeneous Azeotropic Distillation 2. Minimum Reflux Calculations for Non-ideal and Azeotropic Columns, Ind Eng Chem Fund, 24 463. 

Polyurethane networks were prepared from polyoxypropylene (POP) triols(Union Carbide Niax Polyols) after removal of water by azeotropic distillation with benzene. For Niax LHT 240, the number-average molecular weight determined by VPO was 710 and the number-average functionality fn, calculated from Mjj and the content of OH groupSj determined by using excess phenyl isocyanate and titration of unreacted phenyl isocyanate with dibutylamine, was 2.78 the content of residual water was 0.02 wt.-%. For the Niax LG-56, 1 =2630, fn=2.78, and the content of H2O was 0.02wt.-%. The triols were reacted with recrystallized 4,4"-diphenylmethane diisocyanate in the presence of 0.002 wt.-% dibutyltin dilaurate under exclusion of moisture at 80 C for 7 days. The molar ratio r0H = [OH]/ [NCO] varied between 1.0 and 1.8. For dry samples, the stress-strain dependences were measured at 60 C in nitrogen atmosphere. The relaxation was sufficiently fast and no extrapolation to infinite time was necessary. 

Norman, W. S. Trans. Inst. Chem. Eng. 23 (1945) 66. The dehydration of ethanol by azeotropic distillation. Ibid. 89. Design calculations for azeotropic dehydration columns. 

The theoretical and mathematical treatment of azeotropic data has been covered by several authors, including Benedict and coworkers (1), and Scheibel and Friedland (50). Licht and Denzler (37) have discussed the thermodynamic conditions necessary and sufficient for azeotropism. Colburn (8) has reviewed the calculations associated with azeotropic and extractive distillation, and Hodgson (24) stresses the relationship between azeotropic distillation, extractive distillation, and liquid-liquid extraction. 

Figure 13.29. Composition profiles and flowsketches of two azeotropic distillation processes (adapted by King, 1980). (a) Separation of ethanol and water with benzene as entrainer. Data of the composition profiles in the first column were calculated by Robinson and Gilliland, (1950) the flowsketch is after Zdonik and Woodfield (in Chemical Engineers Handbook, McGraw-Hill, New York, 1950, p. 652). (b) Separation of n-heptane and toluene with methylethylketone entrainer which is introduced in this case at two points in the column (data calculated by Smith, 1963).

Design. When the vapor-liquid equilibria are known, in the form of UNIQUAC parameters for instance, the calculation of azeotropic distillation may be accomplished with any of the standard multicomponent distillation procedures. The Naphthali-... 

Sandholm algorithm (Fig. 13.20) and the 0-method of Holland (1981) are satisfactory. Another tray-by-tray algorithm is illustrated for azeotropic distillation by Black, Golding, and Ditsler [Adv. Chem. Ser. 115, 64 (1972)]. A procedure coupling the tower, decanter, and stripper of Figure 13.29(a) is due to Prokopakis and Seider [AIChE 3. 29, 49 (1983)]. Two sets of composition profiles obtained by tray-by-tray calculations appear in Figures 13.29(a) and (b). 

In conclusion, recent developments in solvent selection, phase nonideality description, and tray-to-tray calculation schemes have greatly facilitated the design of extractive and azeotropic distillation schemes, and use of salts give new methods for extractive distillation separations. Finally, the work of Gerster (30), Black and Ditsler (29), and Black et al. (25) compare these two schemes. 

Extractive distillation has been extensively used for nearly three decades in laboratory, pilot plant, and commercial plant operations. Calculation or prediction of phase equilibria for such separations has often been discussed (I, 2, 3). Some have discussed the selection of solvents for extractive distillation (4, 5). Others have discussed its recent application to particular separations (6, 7, 8). A comparison of extractive distillation, as a separation method, with azeotropic distillation and with liquid-liquid extraction has recently been discussed briefly by Gerster (9). 

The use of digital computers to carry out complete calculations in the design of separation processes has been the goal of many. To do this effectively, suitable methods for phase equilibria and tray-to-tray distillation calculations are required. Results calculated by the application of such methods to dehydrate aqueous ethanol mixtures using ethylene glycol as the extractive distillation solvent is discussed below. A brief review of the methods used for phase equilibria and enthalpies is followed by a discussion of the results from distillation calculations. These are compared for extractive distillation with corresponding results obtained by azeotropic distillation with n-pentane. 

The results of calculations for the two separation methods are summarized in Table VIII. Fewer trays are required in the azeotropic distillation column than the extractive distillation column. The heat loads are also smaller. The quality of the ethanol product is also slightly better for the azeotropic distillation method. Including the stripper for processing the aqueous phase, the total heat load for reboilers for the azeotropic distillation method is less than half that for the extractive distillation method. The total condenser load is roughly two-thirds that for the extractive distillation method. 

Although the azeotropic distillation scheme, using n-pentane, operates at a higher pressure, comparative calculations indicate this to be better than the extractive distillation scheme using diethylene glycol to dehydrate aqueous ethanol. 

Azeotropic Distillation Results from Automatic Computer Calculations... 

To calculate phase equilibria suitable for most azeotropic distillation problems, the methods should be applicable to three-phase equilibria. Vapor-liquid and liquid-liquid equilibria are usually required. A suitable method for this purpose has already been discussed (5). It is applied here to calculate completely all phase equilibria involved in the usual azeotropic distillation process. 

This ternary azeotropic distillation program uses a special system of utility subroutines with programmed initialization. Eight main controls, KNTRL, are used with various options on each. Four parameter options are built into the program, but the values are changed by the user by using PRMTR cards. Twenty-one DATA cards allow the user to give the pertinent conditions and specifications for the separation to be calculated. 

The second way in which the program is operated is to calculate rigorously with the subprogram ADPLLE. These calculations are reliable for either two- or three-phase equilibria. The azeotropic distillation columns and the liquid-liquid equilibria for the condensed overhead in the accumulator is calculated with this program. 

A sample calculation is given showing the use of the program to calculate the azeotropic distillation of aqueous ethanol mixtures using the entrainer n-pentane (KNTRL 2, Option 7). 

The computer program for azeotropic distillation ADP/ADPLLE makes possible not only a comparison of entrainers for a separation but also gives results of a quality required for actual design calculations. 

For azeotropic distillation especially the systems are non-ideal which makes calculating vapor-liquid equilibrium properties more difficult than, for example, in distillation of mixtures of simple hydrocarbons. Work predicting the vapor-liquid equilibrium properties of ternary mixtures of... 

---