Energy efficient trace removal by extractive distillation

Separation processes contribute for about 40–70 % to the total energy requirements of the chemical process industry. Especially when trace removal is required to manufacture high purity products, traditional separation technologies become extremely expensive and are not providing satisfying solutions. In this work, trace removal is defined as the separation of impurities or (by) products at a level of 100–1000 ppm. The main reason for trace removal separation is the need to meet the purity demanded by customers further processing and/or legislation. Distillation is currently used to conduct 90–95 % of all separations in the chemical process industry. The drawbacks of distillation are the poor energy efficiency and selectivity. These drawbacks are above all applicable to cases where impurities have to be removed with molecular structures closely resembling the main component (and hence, having similar boiling points) and are present at ppm level in a highly concentrated, organic stream. Two representative industrial examples are the separation of ethylbenzene traces (~100 ppm) from styrene and dichloroacetic acid (DCA) traces (~1000 ppm) from monochloroacetic acid (MCA). Both are examples of close-boiling mixtures (= low relative volatility). A well-established technology to separate close-boiling mixtures is extractive distillation. In extractive distillation, a solvent is added to a distillation column to alter the relative volatility of the system to be separated, and, thereby, to reduce the capital and operational expenditures (OPEX and CAPEX). Nowadays, extractive distillation is already applied to produce high purity products in benzene, toluene, xylene processes with purities up to 99.995 wt%. Therefore, the main objective of the research described in this thesis was to investigate whether extractive distillation could be a promising technology to obtain high purity products for the two representative industrial examples. For both cases, the trace removal and bulk separation were both performed by extractive distillation processes to be able to replace the current separation processes, and thus also achieving process intensification. Sulfolane, which is a commonly applied organic solvent, and a wide range of ionic liquids (ILs) were investigated for the separation of ethylbenzene from styrene to obtain ethylbenzene impurity levels lower than 10 ppm in styrene. Complexing agents (extractants) were studied to separate DCA from MCA to obtain high purity MCA (wDCA <50 ppm). First, a solvent/extractant screening study was performed for both industrial cases followed by the measurement of binary and ternary liquid–liquid and vapor–liquid equilibrium data for the selected solvents and the regression of these data by thermodynamic models. Subsequently, the parameters obtained from the regressions were used in equilibrium stage process models to setup a conceptual process design for the extractive distillation processes including solvent regeneration. The conceptual process models that were developed using Aspen Plus®, were applied for the calculation of the energy requirements, leading to the operational expenditures (OPEX). The capital expenditures (CAPEX) were estimated using the Aspen Process Economic Analyzer®, and finally the total annual costs (TAC) were estimated from the OPEX and CAPEX. The current distillation unit was used as the benchmark process for the ethylbenzene/styrene system. The extractive distillation process for the MCA/DCA system could not be compared to the current separation process, because no process data is available about this process. Therefore, the added production costs per tonne of MCA product were calculated and compared to the current high purity MCA market price. Ethylbenzene/styrene A proper solvent should have a high selectivity as well as a high solvent capacity for the mixture to be separated by extractive distillation. Preferably, the solvent is even fully miscible with the ethylbenzene/styrene system, which was observed for the benchmark solvent sulfolane, whereas from the IL screening study it was found that ILs form multiple liquid phases with the ethylbenzene/styrene system. However, several ILs outperformed the benchmark solvent sulfolane with selectivities up to 2.5, whereas sulfolane displayed a solvent selectivity of about 1.6. Furthermore, it was observed from the IL screening study that a clear trade-off exists between the solvent capacity and selectivity. The ILs 3-methyl-N-butylpyridinium tetracyanoborate ([3-mebupy][B(CN)4]), 4-methyl-N-butylpyridinium tetrafluoroborate ([4-mebupy][BF4]), and 1-ethyl-3-methylimidazolium thiocyanate ([EMIM][SCN]) were selected as promising candidates to study in more detail. The solvent capacities for ethylbenzene and styrene increase in the order of [EMIM][SCN] > [4-mebupy][BF4] > [3-mebupy][B(CN)4]. The selectivity for the ethylbenzene/styrene system decreases in the same order. Binary and ternary phase equilibrium data were obtained for the systems ethylbenzene + styrene + selected ILs, and ethylbenzene + styrene + sulfolane. The NRTL model was able to correlate the phase equilibrium data adequately, and the parameters that were obtained by regression of the experimental data were used in the following chapters in the process simulations. A very important factor in extractive distillation processes is the recovery of the solvent. If that cannot be done adequately and cost effectively, the process cannot be competitive. Therefore, a screening of several IL recovery technologies was performed, for which the IL [4-mebupy][BF4] was taken as a model IL. It was found that this IL should be purified to at least 99.6 wt% in the regeneration section to maintain the distillate and bottom purity requirements for the primary separation of ethylbenzene and styrene in the extractive distillation column. From the TAC, the overall conclusion can be drawn that evaporation using very low pressures (P <10 mbar) and ethylbenzene stripping, both in combination with an evaporator operating at mild conditions (T = 130 oC, Tcondenser = 20 oC), are the most promising technologies to recover styrene monomer from ILs, although the differences in TAC compared to stripping with nitrogen for example are rather small. The conceptual design study of the total extractive distillation process showed that the IL [4-mebupy][BF4] outperforms the other two selected ILs [3-mebupy][B(CN)4] and [EMIM][SCN] with up to 11.5 % lower energy requirements. The OPEX of the [4-mebupy][BF4] process are 43.2 % lower than the current distillation process and 5 % lower than the extractive distillation process using sulfolane. However, the CAPEX were about 23 % higher for the [4-mebupy][BF4] process compared to the sulfolane process. Finally, from the TAC, the conclusion can be drawn that all extractive distillation processes clearly outperform the current distillation process to obtain high purity styrene with up to 40 % lower TAC, but that ILs do not perform better than sulfolane. MCA/DCA An adequate extractant should enhance the relative volatility of the MCA/DCA system to a value above 1.8 to obtain an energy efficient extractive distillation process, and the formed complex between the extractant and the solute should be reversible to regenerate the solvent. Moreover, the extractant should be thermally/chemically stable in the strongly acidic environment to avoid irreversible extractant loss and considerable by-product formation, and the extractant should have a sufficiently high boiling point to facilitate easy regeneration. The extractant screening study showed that many extractants outperform the benchmark extractant sulfolane with relative volatilities of the MCA/DCA mixture up to 5.9, whereas sulfolane displays a relative volatility of 1.6. There is, however, a clear trade-off between the effect of the extractant on the relative volatility of the MCA/DCA system and the regeneration ability of the extractant. Extractants with high effect on the relative volatility of the MCA/DCA system appeared to be difficult to regenerate in the regeneration column. Moreover, only the tested ethers, phosphine oxides, ketones, and the benchmark extractant sulfolane were found to be stable in the strongly acidic MCA/DCA environment. Finally, the extractant diethylene glycol dipentyl ether (DGDP) was selected, which improves the relative volatility of the MCA/DCA system up to 2.1–2.4 at a DCA/extractant mole ratio of 1. Moreover, DGDP could easily be regenerated using distillation and it was found to be thermally and chemically stable. After succeeding in finding a promising extractant, binary and ternary VLE data were determined for the system MCA + DCA + DGDP at 5, 7.5, and 10 kPa to be able to simulate the extractive distillation process. The non-ideal behavior in the liquid and vapor phase were well correlated with the NRTL and Hayden-O’Connell (HOC) model, respectively. The conceptual design study showed that the TAC for the extractive distillation process are 1.6 M€/year for a 70.000 metric tonnes annually MCA plant. The addition to the MCA production costs are 23 €/tonne MCA product that is rather low compared to the current (high purity) MCA price of 1500 €/tonne. Therefore, the conclusion can be drawn that extractive distillation using DGDP as extractant is a promising technology to obtain high purity MCA. Overall conclusion and outlook From both industry cases, the conclusion can be drawn that extractive distillation is a very promising concept to obtain high purity products from close-boiling mixtures by replacing the currently applied separation technology. For the ethylbenzene/styrene separation, sulfolane slightly outperformed the ILs, and since sulfolane is already applied on industrial scale it is easier to implement this solvent on short term. However, for implementation on industrial scale, still some further research needs to be performed. The polymerization rate of styrene in sulfolane requires attention, because some solvents can enhance polymerization rates, which affects the required inhibitor flow. Moreover, it is crucial to find a separation method to effectively separate heavies, like styrene polymers, from ILs. Adding an antisolvent like methanol was tried and seems promising. Finally, pilot plant experiments needs to be performed, because mass transfer efficiency influences the required packing height and the hydrodynamic capacity of a packing depends strongly on the system to be separated. On the long term, the energy efficiency of the ethylbenzene/styrene separation by extractive distillation using ILs can be improved by screening for an IL, which has a high solvent capacity as well as high selectivity for the ethylbenzene/styrene system. Furthermore, finding more possibilities for solvent heat recovery in the styrene production process can also improve the energy efficiency significantly. For the MCA/DCA separation, the possible savings in OPEX and CAPEX by the application of the extractive distillation process in the current production process should be investigated. Furthermore, the thermal/chemical degradation of the extractant DGDP in the strongly acidic environment needs to be studied in more detail, because it is essential to determine the required extractant make-up flow. Furthermore, it is very important to determine which degradation products are formed and where these products will end up. Next to that, the effect of impurities in the extractant recycle requires further investigation as well. Also for this case, pilot plant experiments should be carried out to determine the mass transfer efficiency and the hydrodynamic capacity