Repository logo

Masters Degrees (Chemical Engineering)

Permanent URI for this collectionhttps://hdl.handle.net/10413/6658

Browse

Browsing Masters Degrees (Chemical Engineering) by Author "Avoseh, Funmilola Elizabeth."

Filter results by typing the first few letters
Now showing 1 - 1 of 1
  • Thumbnail Image
    Item
    Binary vapour-liquid equilibrium for systems of industrial importance.
    (2015) Avoseh, Funmilola Elizabeth.; Ramjugernath, Deresh.; Narasigadu, Caleb.
    Most industrial chemical engineering separation processes such as distillation, extraction, absorption and adsorption rely absolutely on accurate phase equilibrium data for effective design, optimization and simulation. Carbonyls and alcohols are known to be of important use in the petrochemical industries. Ketones alongside with alcohols and carboxylic acids are found both in the product stream and waste stream of the Fischer-Tropsch process. 4-methyl-2-pentanone forms parts of these by-products and it is used in a number of industrial applications. It is generally used as solvent, as chemical intermediate in the production of paints, rubber products, chemicals, resins and drugs to mention a few, due to its low solubility in water; it is used for liquid-liquid extraction. This work focuses on measurement of new vapour-liquid equilibrium (VLE) data for binary mixtures of : 1-Propanol (1) + 4-methyl-2-pentanone (2) (at 338.15 K, 353.15 K, and 368.15 K), 2-propanol (1) + 4-methyl-2-pentanone (2) (at 323.15 K, 338.15 K, and 353.15 K) and 2-pentanone (1) + 2-methylpropan-1-ol (2) (at 343.15 K, 358.15 K, and 363.15 K). A modified (Bhownath, 2008) low pressure dynamic VLE glass recirculating still originally designed by Raal (Raal & Mühlbauer, 1998) was used for the measurements. This work also presents the infinite dilution activity coefficients and the excess thermodynamic properties (i.e. molar excess Gibbs energy GE, heat of mixing HE, and excess entropy SE). These properties were derived from the measured isothermal VLE data. A highly non-ideal system comprised of cyclohexane + ethanol was chosen as a test system and was used to verify the reproducibility and repeatability of the apparatus. The test system had been measured in our laboratory (Joseph, 2001) and the data were found to agree excellently with those of Morachevsky and Zharov (1963) and were reported to be thermodynamically consistent according to Gmehling and Onken (1977). The results for the test system measured in this work were in excellent agreement with literature. Thus, there was confidence in the new data measured since the apparatus and the operating procedures used for the test system were able to give accurate results. The vapour pressures measured in this study were also in good agreement with literature. The temperature, pressure and composition measuring devices were well calibrated and the uncertainty acquired for each is included. The uncertainty in the pressure measurement was estimated to be ± 0.02 kPa and controlled within 0.01 kPa. The uncertainty in the temperature measurement was estimated to be ± 0.06 K (Type B uncertainty, NIST) and was controlled within 0.04 K during manual operation. The uncertainty in the composition measurement was estimated as ± 0.002. The 1-propanol (1) + methyl isobutyl ketone (2) system was found to exhibit a minimum boiling azeotrope at 353.15 K. The gamma-phi (γ-Φ) or combined method was used for the regression of the measured VLE data. Three activity coefficient models were investigated to account for the liquid phase deviation of the mixture from ideality: NRTL (Renon and Prausnitz, 1968), Wilson (1964) and the UNIQUAC (Abrams and Prausnitz, 1975) models. Two equation of state models were used to account for the vapour phase non- ideality: the virial EoS with the Hayden O’ Connell (Hayden & Connell, 1973) correlation for the calculation of the second virial coefficient, and the Nothnagel (Nothnagel, Abrams, & Prausnitz, 1973) formulation. The maximum likelihood regression technique was used to determine the regressed parameters of the activity coefficient models. These models were found to fit the measured data well. The measured VLE data passed the point test of Van Ness(et al., 1973) and the direct test (Van Ness, 1995).