Development of novel apparatus for vapour-liquid equilibrium measurements at moderate pressures.
In this work, a novel experimental apparatus has been designed, constructed and commissioned for the measurement of VLE at pressures up to 750 kPa and temperatures up to 600 K. The project undertaken represents a complete re-working of the design of Harris (2004), which was plagued by irregularities in the equipment operation and in the acquisition of experimental data. As in the work of Harris (2004), the design of the apparatus presented here is based upon the highly successful glass VLE still design of Raal (Raal and Muhlbauer, 1998). The novel apparatus is principally constructed from machined 316 stainless steel and features sight glasses in strategic positions to allow for an observation of the fluid flow characteristics in specific sections of the apparatus. The key criteria that encompassed the design of the equipment were expediency, operational efficiency and versatility in the acquisition of reliable VLE data. An initial test of the performance of the equipment was achieved through the measurement of pure-component vapour pressures of selected hydrocarbons (n-alkanes and a cycloalkane) and alkanols. The test system for vapour-liquid equilibrium (P-T-x-y) measurements with the novel apparatus was that of cyclohexane + ethanol at a pressure of 40 kPa. Good agreement between the literature and the experimental data was observed. Isobars for the cyclohexane + ethanol system at 69.8 kPa, 97.7 kPa and ISO kPa were also measured. The latter constitutes new data that have been measured for this system. Novel vapour-liquid equilibrium data were also obtained for the systems of I-propanol + 2-butanol, I-propanol + n-dodecane and 2-butanol + ndodecane at temperatures of 373.15 K, 393.15 K and 423.15 K. For the very high relative volatility alkanol + n-dodecane systems, uncertainties in the measurement of the vapour phase (y) resulted in only P-T-x experimental data being presented here, where the vapour phase composition was computed with the Wilson equation. The theoretical treatment of the experimental VLE data was achieved through a combination of the gamma-phi and the phi-phi approaches in the fitting of the VLE data to various thermodynamic models. In the gamma-phi method, a variety of activity coefficient models (Wilson, T-K Wilson, NRTL, UNIQUAC and modified UNIQUAC) together with the truncated virial equation of state were employed to find the best fit for the data. In the phi-phi method, the isothermal data sets were treated with the Peng-Robinson-Stryjek-Vera equation of state with the original Huron-Vidal (HV) and the modified Huron-Vidal mixing rules (MHVI and MHV2) in the correlative procedure. Thermodynamic consistency testing was also performed with the Direct Test of Van Ness (1995) to assess the quality of the experimental P-T-x-y VLE data sets measured in this study.