STUDIES ON EXTRACTION OF AROMATICS FROM KEROSENE FRACTION

Abstract

Dearomatisation of petroleum refinery fractions like naphtha, kerosene, gas oil and heavier fractions is practiced in refineries to improve their properties and to meet the specifications with respect to aromatics for various end products. Bureau of Indian Standards (BIS) specifications for aviation turbine fuel (ATF) (IS 1571: 1992) put a limit on the concentration of aromatics (22 % by volume or -24 % by weight) and require a smoke point of minimum 20 mm whereas BIS specifications (IS 1459: 1974) for superior kerosene (SK) require aromatic content to be low enough to give a non-smoky flame height of at least 18 mm. In India, typical kerosene fraction (140-270°C) obtained from Assam mix crudes is rich in aromatics and contains about 38 wt.% aromatics with smoke point of 15 mm and thus does not meet the above specifications. Hence dearomatisation of this fraction is necessary to produce SK and ATF. Hydrotreatment and solvent extraction are the conventional processes used for dearomatisation of kerosene fractions. Solvent extraction is an attractive route for dearomatisation of high aromatics kerosene due to requirement of low capital and operating cost compared to hydrotreatment and availability of more valuable aromatic extract as a by product. Comparison of the old Edleanu process using liquid sulphur dioxide as the solvent with a more environment friendly sulpholane extraction technology developed jointly by the Indian Institute of Petroleum, Dehradun (IIP), Engineers India Limited (EIL) Delhi and Hindustan Petroleum Corporation Limited (HPCL), Mumbai exhibited several advantages e.g. compact plant design, environment friendly solvent, increased raffinate yield, etc. This process requires high utilities and puts a constraint on the final boiling point of kerosene fraction to 240°C instead of 270°C. Up-gradation of this technology is possible by use of re-extraction route for aromatics recovery from the extract phase, rather than distillation for the recovery of kerosene hydrocarbons from NMP and/or sulpholane. In the re-extraction scheme, the extract phase obtained after aromatics extraction is re-extracted generally with a Low Boiling Paraffinic Stream (LBPS), to recover aromatic hydrocarbons from the extract phase. The resulting aromatic lean solvent could then be recycled back into the upstream extractor and LBPS could be recovered by simple fractionation for re-use. This thesis reports the detailed studies on dearomatisation of full range straight-run kerosene fraction (140-270°C) and model hydrocarbons representing this kerosene as feedstocks with sulpholane and NMP +5 wt.% water as solvents under varying operating conditions followed by the solvent recovery step using the re-extraction concept, using hexane fraction as re-extraction solvent, with varying operating parameters i.e. water content of NMP, solvent to feed ratio, extraction temperature, etc. Batch and continuous extraction and re-extraction runs were carried out for the dearomatisation of full range kerosene (140-270°C cut) by using sulpholane and NMP+5 wt.% water as extraction solvent. The detailed characterization e.g. ASTM distillation data, hydrocarbon class type analysis, smoke point, flash point, freezing point, total sulphur, etc. of the straight-run kerosene feedstock was carried out. It was found that the BIS specifications for aromatics content and smoke point are not met. Batch LLE data for extraction step were generated for typical model hydrocarbons such as propyl benzene and decane, representing aromatics and saturates in the kerosene feedstock, with solvents in a 200-ml capacity jacketed mixer-settler. For re-extraction step, LLE data were generated for the extract phase represented by two model compounds, namely, propyl benzene (representing alkyl benzenes) and methyl naphthalene (representing diaromatics) as the solutes, sulpholane/NMP+5% water as the diluent and n-heptane/nhexane as the solvents. Continuous extraction runs were carried out in a packed glass column of 34 mm internal diameter and packed with 6 mm ceramic Intallox Saddles to 1.5 m f v height. All extraction and re-extraction runs were carried out at 40°C except the extraction runs with sulpholane which were carried out at 80°C. The S/F ratio (by volume) of 2 and 4 for extraction runs and 1 and 2 for re-extraction runs were used. The results obtained in continuous extraction runs showed that S/F ratios of 2 and 4 for both sulpholane and NMP+5 wt% water are capable of producing raffinate with lower aromatics content and higher smoke points than stipulated for ATF/SK by BIS. The sulpholane extract phase distillation in Oldershaw column revealed that even with stringent operating conditions i.e. high vacuum, high reboiler temperature and use of stripping steam, the lean solvent (solvent obtained at the bottom ofthe column) contained about 3 wt.% residual aromatics, which impaired the selectivity and extraction performance. Application of re-extraction route as an alternative to distillation for recovery of hydrocarbons from kerosene extract phase was therefore studied. For this purpose, corresponding extract phases obtained during continuous extraction runs were re-extracted with hexane under different operating conditions in batch and continuous extraction units. The re-extraction studies showed that the kerosene hydrocarbon content of the extract phase could be substantially reduced in the re-extraction step under ambient operating conditions. The results obtained on continuous column indicated that for dearomatisation of kerosene, solvent as dispersed phase gives better performance than when used as the continuous phase. This was observed in the case of both the solvents. The LLE data for model hydrocarbon systems were correlated by NRTL and UNIQUAC thermodynamic models and the corresponding binary interaction parameters were estimated by minimisation of objective functions using maximum likelihood principle. The results showed a very good match between experimental and correlated values for both these models. UNIFAC group contribution approach was used for the prediction of LLE data for kerosene feedstock. For this purpose, the kerosene feedstock was represented by six IV model hydrocarbons representing hydrocarbon class types such as paraffins, condensed and non-condensed cycloparaffins, indanes and tetralins, alkyl benzenes and diaromatics. Selection of these components was based on mean average boiling point of the kerosene fraction. Interaction parameters for NMP were estimated from the LLE data reported in the literature of the model hydrocarbons with NMP containing different amounts of water. Substantial improvements in results were obtained after incorporation of these revised parameters. UNIFAC group interaction parameters were further used for simulating the results of countercurrent extraction and re-extraction runs using equilibrium stage approach. A good agreement between the predicted and experimental values confirmed the application of UNIFAC method with the parameter sets generated above for simulations. By using these parameters and sum rate method, continuous extraction and re-extraction runs were simulated successfully. The process flow sheet for the dearomatisation of full range kerosene by using re-extraction route for solvent recovery was proposed and simulations were carried out with sulpholane and NMP+5 wt.% water for S/F ratio of 1 and 2 by weight using ASPEN PLUS (Release 9.2) flowsheet simulator. It is concluded that both sulpholane and NMP+5 wt.% water can be used for the dearomatisation of kerosene using re-extraction approach for recovery of hydrocarbons from the extract phase. The products meet the specifications for SK/ATF. This approach will require approximately 40% less utilities over the earlier distillation approach, besides other advantages like production of better quality Aromex, feasibility of using high capacity solvent like NMP and reduction in solvent degradation due to solvent recovery at lower temperature.