Relative volatility correlation example

Equilibrium data correlations can be extremely complex, especially when related to non-ideal multicomponent mixtures, and in order to handle such real life complex simulations, a commercial dynamic simulator with access to a physical property data-base often becomes essential. The approach in this text, is based, however, on the basic concepts of ideal behaviour, as expressed by Henry s law for gas absorption, the use of constant relative volatility values for distillation and constant distribution coeficients for solvent extraction. These have the advantage that they normally enable an explicit method of solution and avoid the more cumbersome iterative types of procedure, which would otherwise be required. Simulation examples in which more complex forms of equilibria are employed are STEAM and BUBBLE. 

Measurements of binary vapor-liquid equilibria can be expressed in terms of activity coefficients, and then correlated by the Wilson or other suitable equation. Data on all possible pairs of components can be combined to represent the vapor-liquid behavior of the complete mixture. For exploratory purposes, several rapid experimental techniques are applicable. For example, differential ebulliometry can obtain data for several systems in one laboratory day, from which infinite dilution activity coefficients can be calculated and then used to evaluate the parameters of correlating equations. Chromatography also is a well-developed rapid technique for vapor-liquid equilibrium measurement of extractive distillation systems. The low-boiling solvent is deposited on an inert carrier to serve as the adsorbent. The mathematics is known from which the relative volatility of a pair of substances can be calculated from the effluent trace of the elutriated stream. Some of the literature of these two techniques is cited by Walas (1985, pp. 216-217). 

The predictive techniques are rather accurate. However, significant errors have been observed in few cases (4, 13, 27, 40). No direct comparison between the three predictive methods is available. The authors of the parachor method (27) claim that their method yields equal or better results than the PDD method for the cases considered in their study it is believed (42), however, that the latter is more reliable and it is recommended. The Weimer-Prausnitz method probably gives less accuracy than the PDD method, but it is more general. For example, Hanson and Van Winkle (40) report that their data on the hexane-hexene pair were not successfully correlated by the WP method. The Helpinstill-Van Winkle modification is recommended over the WP method. Recently, Null and Palmer (43) have presented a modification of the WP method which provides better accuracy but it is less general. The PDD method should be used cautiously when extrapolation with respect to temperature is used (27). When the GLC method is used, reliable results are expected. Evaluation of infinite dilution relative volatilities is recommended (36). 

Electronic nose technology and analysis of volatiles has long been apphed in the food industry to control the quahty of food products and to determine shelf hves. For example, sensor arrays based on different Sn02 gas sensors can be used to distingiush milk products of different rancidity levels [41]. Standard microbial test prediction of shelf hfe of milk products has a low level of reliability due to relatively poor correlation between microbial counts and actual shelf hfe. Several alternative methods have therefore been developed. One method is based on dynamic headspace capillary gas chromatography analyses of volatiles in mUk followed by MDA analyses. [42]. Principals of this method were later used for development of a faster and simpler test, where the extraction was performed by the SPME technique, the extracts... 

Use the Gilliland correlation to estimate the number of ideal stages required for the separation of Example 6.4. Assume that, for the system benzene-toluene at atmospheric pressure, the relative volatility is constant at a = 2.5. Use the results of Problems 6.15 and 6.16. Use the Kirkbride equation to estimate the optimum feed-stage location. 

Solution. The Lockhart-Leggett version of the O Connell correlation given in Fig. 13.5 can be used to estimate E . Because absorber oil is used, it is preferable to compute relative volatility and viscosity at average liquid conditions. This is most conveniently done for this example by averaging between the top and bottom stages. From the solution based on the Soave-Redlich-Kwong correlation, with ethane and propane as light key and heavy key, respectively... 

If the top temperature is too cold and the bottom tenperature is too hot to allow sandwich conponents to exit at the rate they enter the column, they become trapped in the center of the column and accumulate there fKister. 20041. This accumulation can be quite large for trace conponents in the feed and can cause column flooding and development of a second liquid phase. The problem can be identified from the simulation if the engineer knows all the trace conponents that occur in the feed, accurate vapor-liquid equilibrium (VLE) correlations are available, and the simulator allows two liquid phases and one vapor phase. Unfortunately, the VLE may be very nonideal and trace conponents may not accumulate where we think they will. For example, when ethanol and water are distilled, there often are traces of heavier alcohols present. Alcohols with four or more carbons (butanol and heavier) are only partially miscible in water. They are easily stripped from a water phase (relative volatility 1), but when there is litde water present they are less volatile than ethanol. Thus, they collect somewhere in the middle of the column where they may form a second liquid phase in which the heavy alcohols have low volatility. The usual solution to this problem is to install a side withdrawal line, separate the intermediate component from the other components, and return the other components to the column. These heterogeneous systems are discussed in more detail in Chapter 8. 

Such expectations typically involve attempts to infer what spectromet-ric data do not imply. Few conclusions concerning sample purity, for example, follow from a routine mass spectrum. The intensity of a molecular ion peak in the spectrum observed is not necessarily correlated to the concentrations of components in the sample investigated. A volatile but stable minor component can produce an intense molecular ion peak. The most abundant component, however, if it is comparatively involatile or insensitive to the ioiiization method, may yield a relatively weak peak. Alternatively, the molecular ion formed by the major component may fragment into smaller ions faster than it can be detected. In these circumstances, it outrages logic to conclude that a compound is absent from a sample because a mass spectrum failed to show the corresponding molecular ion. Such negative evidence deserves no trust. 

When the calibration curves were compared, several compounds at the low end of the calibrated concentration range were affected by components already present in the diesel/oil extract. For example, low-levels of some PNAs and phthalates, present naturally in these refined petroleum products, were detected in the unspiked diesel/oil extract. Also, some of the phenols in this dirty matrix were reactive in the injection liner indeed, the matrix itself can passivate the liner for some target compounds. Passivation in this sense means that the liner surface becomes coated with non-volatile components, forming a barrier between the analyte and the bare, more reactive glass surface. While this issue is not related to ion trap mass spectrometry per se, it will be present in any analytical GC/MS system. As illustrated in the example below, calibration curve linearity (as represented by relative percent standard deviation, or RSDs, of the relative response factor at each calibration concentration level) and correlation coefficients for most compounds in the pure solvent were identical statistically to those prepared in the 3000 ppm diesel/oil matrix spikes, as are shown in Figures 15.36 and 15.37. 

---