Use of 1-ethyl-3-methylimidazolium ethyl sulfate for liquid-liquid equilibria for ternary mixtures

Abstract

This thesis forms part of the Durban University of Technology Thermodynamics Research Unit’s project which is aimed at developing a method for determination of the liquid-liquid equilibria (LLE) data for the azeotrope {methanol + water} with an ionic-liquid (IL) using DSA5000M to assess the efficiency of the ionic liquid to be used in liquid-liquid extractions for the recovery and recycling of methanol from petroleum refinery. The objective of this study was to determine the liquid-liquid equilibria data of the azeotrope {methanol + water} using 1-ethyl-3-methylimidazolium ethyl sulfate ionic liquid with the intention to recycle methanol from the Fischer-Tropsch (FT) process by- products in petroleum industries and to utilize it in gasoline additives in a new methanol to gasoline (MTG) petroleum process. LLE studies of systems containing alcohols and water are important due to the increasing demands of oxygenated compounds to produce lead free gasoline. Light alkanols such as methanol and ethanol are reported to be suitable compounds in order to produce lead free gasoline, but the use of methanol in gasoline blends can cause phase separation problems in: 1. dry conditions, these are due to its partial solubility in saturated hydrocarbons.

2. the presence of water from ambient humidity or in storage tanks, this depend on unfavourable distribution factor between aqueous and the hydrocarbon phase.

To determine the possibility of separating methanol from water using ionic liquid, the liquid-liquid equilibria data was determined at room temperature, T = 298.15 K and atmospheric pressure to investigate whether it separate from water and/or a non-phase separation if it is used as an additive. The experimental data generated was compared to that of the literature for the system

{methanol (1) +toluene (2) + dodecane (3)} and showed good agreement with the literature data with only maximum deviation of ± 0.0015 in the mole fraction using density calculations and ± 0.0092 in the mole fraction when using refractive index calculations The selectivities and distribution coefficients for this system were also calculated and the maximum deviation between the two methods (nD and ρ) was ± 1.33 in selectivities and found to be ±0.001 for distribution coefficients. The maximum deviation in distribution coefficients from literature when using nD calculations for system 1 was ±0.04 and ±0.01 for ρ. For the selectivity values the deviation from that of literature of nD when compared was found to be ± 1.28 and 0.29 for ρ respectively. The selectivity values from the density calculations were found to be in the range 2.82 –

7.66 for this system with the distribution coefficient values reported in the range 0.17 – 0.23.

In the second system (system 2) the generated experimental data was also compared to that of the literature for the system {water (1) + methanol (2) + cyclohexane (3)} and in good agreement with literature values with only maximum deviation of ± 0.0091 in the weight fraction based on density calculations. The selectivities and distribution coefficients were also calculated and the maximum deviation between the literature and the experimental data was computed to be at ± 0.0003 for selectivity and ±0.09 in distribution coefficient.

The selectivity values were found to be in a range 0.00 - 0.04 for this system and were constant throughout the phases but significantly less than one; with the distribution coefficient values in the range 0.00 – 0.008. For 1-ethyl-3-methylimidazolium ethyl sulfate system (Ionic liquid system) the selectivity values were not constant throughout the two-phase region and the values were found to be in the range 0.63 -0.99 still below one which indicates that the ionic liquid used in this study could not be considered as a potential solvent for the separation of the investigated azeotrope. The distribution coefficients for this system were determined and found to be in the range 0.23 – 0.74. The certainty and reliability of experimentally measured tie-line data was ascertained by applying Othmer-Tobias (OT) correlations and the Non-Random, Two Liquid (NRTL) parameters. The OT correlations for system 1 was linear and indicated the certainty of the five tie-lines prepared for this system. In system 2 the OT correlation was not linear and indicated extensively high errors as well as high systematic multiplicative and additive errors in calculations of mole fractions. For the IL system the OT correlation was linear throughout the whole tie-line range and indicated the adequate precision, which denotes that the investigation was carried out with minimal random and systematic errors and indicated the efficiency of the DSA 5000 M to generate the liquid-liquid equilibria data.

All the ternary systems were well correlated and in good agreement with the estimated NRTL data.

It was only system 1{methanol (1) + toluene (2) + dodecane (3)} that gave a high maximum deviation ( %RSMD) of 1.288 when using the RI measurements with the minimum error margin of 0.6320, this account as to why RI measurements were not applied in other systems (system 2 and ionic liquid system). Similarly for the same system; system 1{methanol (1) + toluene (2) + dodecane (3)} when using the density measurements; the NRTL model gave a maximum deviation of 0.5620 and minimum error margin of 0.2590. The NRTL obtained for system 2 {water (1) + methanol (2) + cyclohexane (3)} gave the maximum deviation of 0.5752 and minimum error margin of 0.0127. The NRTL for the ionic liquid ternary system {[EMIM][EtSO4](1) + methanol (2) + water (3)}showed a good agreement between the experimental data and the NRTL model tie- line data with the %RSMD of 1.0201 on the upper limit and 0.1620 as a lower deviation.