Solubility of p-xylene in methane and ethane and implications for freeze-out at LNG conditions

Abstract Even at trace concentrations the presence of heavy hydrocarbons such as BTEX compounds (benzene, toluene, ethylbenzene, and xylenes) in liquefied natural gas (LNG) production streams poses a significant risk of blockage and eventually plant shutdown. However, although p- xylene has the highest melting temperature of all the BTEX compounds, no data are available for its solubility in liquid methane or ethane. In this work, sapphire equilibrium cells housed in either an air bath or a cryogenic thermostat fitted with periscopes were used to measure melting (liquidus) temperatures for synthetically prepared methane +  p- xylene and ethane +  p- xylene mixtures at temperatures from 200 K and pressures up to 22.5 MPa. Peltier-driven copper posts, which could be sub-cooled relative to the bulk liquid mixture, controlled the location of the solid formed within each cell. For the methane +  p- xylene system both VLE and SLVE data sets were measured and used separately to tune the binary interaction parameter (BIP) within the Peng Robinson equation of state. The VLE-tuned BIP systematically under-predicted the measured melting temperatures for this binary by up to 5.2 K, while the SLVE-tuned BIP could represent the melting temperature data with an r.m.s. deviation of 0.4 K. For ethane +  p- xylene, a BIP estimated using a group contribution method systematically over-predicted the measured melting temperatures by as much as 14.6 K. Tuning the ethane +  p- xylene BIP to the measured SLE data more than halved the model’s r.m.s. deviation to 3.1 K. The use of BIPs tuned to VLE data rather than SLVE data has significant implications for freeze-out risk assessments in LNG production. For example, at operating conditions typical of an LNG plant’s main cryogenic heat exchanger, the solubility of p -xylene in liquid methane predicted using a BIP tuned to VLE data is 20 times larger than the solubility predicted using a BIP tuned to SLVE data.