(202e) Thermodynamics Modelling of CO2 + n-Alkane + Alcohol Systems with the RK-Aspen Thermodynamic Model

Detergent range alcohols (octanol and larger) are used in the production surfactants. These alcohols can be produced, amongst others, from an alkane feedstock. Typically, the alkane feedstock has a range of molecular masses and therefore so to do the alcohols. However, the conversion of alkanes to alcohols is not complete, resulting in the requirement of a post-production separation process.

Evaluation of the physical properties of the alkanes and alcohols show that cross-over boiling and melting points exits [1]. Additionally, association is present between n-alkanes and 1‑alcohols leading to low boiling point azeotropic behavior in many subsystems where the boiling points are similar [2–5]. Advanced separation methods are thus required to separate the alkanes and alcohols. While heterogeneous azeotropic distillation has been shown to be technically viable, high temperatures are required, leading to significant thermal degradation.

Supercritical CO₂ fractionation has been shown to be a suitable alternative [6–8]. Pilot plant scale studies have shown that it is possible to fractionate an alkane + alcohol mixture in this molecular mass range. However, these pilot plant scale experiments are time consuming and costly to implement. Further, they have shown that there appears to be a temperature dependent pinch in the separation system. Thus to facilitate the design of such systems without the costly and time consuming pilot plant experiments, process simulation is required.

In order to model the fractionation of alkanes + alcohol mixtures in a process simulator like Aspen Plus^® a reliable thermodynamic model is required. To date a number of studies have considered various thermodynamics implemented in Aspen Plus® and the RK-Aspen model has been shown to be the most suitable model outperforming more advanced models such as PC-SAFT and CPA [9–11].

The RK-Aspen model [12] is a modification of the SRK model [13] where an additional polar parameter is added to the alpha function. This polar parameter allows for improved prediction of the pure component properties while still retaining the cubic nature of the EOS. The equation of state is extended to mixtures using quadratic mixing rules and two temperature dependent BIPs.

To date a number of studies have measured high pressure bubble and dew point data for CO₂ + n-alkane + 1-alcohol systems [9–11,14]. Essentially, in these studies the phase transition pressure is measured at a set composition and temperature. This data are relatively easy and cheap to measure but comes with the disadvantage of not providing information on the compositions of the co-existing phases. However, a previous study [10] has shown that for these types of systems the RK-Aspen model can be correlated using phase transition data to a similar accuracy than when using phase equilibrium data where the compositions of the co-exiting boundaries are known.

A number of studies have correlated the RK-Aspen model to various CO₂ + n-alkane + alcohol systems [9–11,14]. Each of these studies have used different data and methods to correlate the polar parameter and the BIPs, resulting in slightly different values and predictions. Further, limited implementation of the temperature dependence of the model parameters was considered although results indicate the need therefore. These individually determined BIPs also do not allow for interpolation and extrapolation of the polar parameters and BIPs.

The aim of this work is to systematically apply the RK-Aspen model to the CO₂ + n-alkane + 1‑alcohol system for n-alkanes between n-decane and n-eicosane and for 1-alcohols between 1‑octanol and 1-dodecane. The study was conducted in Aspen Plus^® V14 to closely resemble the process followed during the process simulation of a supercritical fluid fractionation study. Where suitable, the Aspen Simulation Workbook (ASW) in MSExcel was used to aid calculations.

The polar parameters were correlated against vapor pressure data estimated from the extended Antoine equation. The parameters were obtained from the NIST ThermoData Engine in Aspen Plus^® V14.0. in order to apply a more universal approach to the model, the final polar parameters were determined as a function of molecular mass for each of the homologous series.

For each solvent + solute pair the literature was scoured for phase behavior data. At each temperature the data were compared and if required phase transition data were converted to phase equilibrium data. To ensure a sufficient pressure range a minimum pressure range of 2 MPa was required for the data and at least 10 co-existing phases were required. In many cases very limited data are available near the critical point of CO₂. Here most of the data are bubble and dewpoint data with little or no measurements in the CO₂ rich side due to the asymmetry of the phase behavior. The inclusion of temperature dependence in the BIPs was also investigated. Linear temperature dependence was included where significant improvement in the correlation was observed with the inclusion thereof.

A large quantity of the reliable ternary data needed for the correlation of the solute + solute BIPs is high-pressure bubble and dewpoint data. The Aspen Plus^® DRS is not able to regress parameters against this type of data. The flash drum function in Aspen Plus^® is however able to estimate the bubble or dew point pressure of a stream at a set temperature and composition. Therefore, using the ASW a fine grid optimization was implemented to determine the BIP values where the %AAD was found to be a minimum. Previously determined BIPs were used as starting values. Where the inclusion of temperature dependence in the BIPs significantly improves the correlation of the data, linearly temperature dependent BIPs were used, else these were omitted.

This work has conducted a systematic study on the modelling of CO₂ + n-alkane + 1-alcohol system using the RK-Aspen model. A more universal approach was proposed with the intention of providing more generalized BIPs rather than parameters correlated to specific systems. While the approach was implemented for a single series of data, this approach can be extended to other CO₂ + solute + solute systems in the future.

References

[1] B.E. Poling, J.M. Prausnitz, J.P. O’Connell, The Properties of Gases and Liquids, 5th Edition, McGraw-Hill, New York, 2001.

[2] M. Ferreira, C.E. Schwarz, Low-Pressure VLE Measurements and Thermodynamic Modeling, with PSRK and NRTL, of Binary 1-Alcohol + n-Alkane Systems, J. Chem. Eng. Data 63 (2018) 4614–4625. https://doi.org/10.1021/acs.jced.8b00680.

[3] S.H. Du Plessis, C. Latsky-Galloway, C.E. Schwarz, Experimental Measurement and PSRK and NRTL Modeling of Binary 1-Alcohol (C_m) + n-Alkane (C_m+3) Vapor–Liquid Equilibrium Data, J. Chem. Eng. Data 67 (2022) 1915–1931. https://doi.org/10.1021/acs.jced.1c00915.

[4] S.H. Du Plessis, C. Latsky-Galloway, C.E. Schwarz, Low-Pressure Phase Equilibrium Data for Binary Systems of 1-Alcohols (C_nOH) and n-Alkanes (C_n+2), J. Chem. Eng. Data 68 (2023) 2898–2912. https://doi.org/10.1021/acs.jced.3c00026.

[5] S.H. du Plessis, S.A.M. Smith, C. Latsky-Galloway, C.E. Schwarz, Investigation of the Low-Pressure Phase Behavior and SAFT Modeling of 1-Alcohol and n-Alkane Binary Systems, J. Chem. Eng. Data 69 (2024) 623–638. https://doi.org/10.1021/acs.jced.3c00346.

[6] G.J.K. Bonthuys, C.E. Schwarz, A.J. Burger, J.H. Knoetze, Separation of alkanes and alcohols with supercritical fluids. Part I: Phase equilibria and viability study, J. Supercrit. Fluids 57 (2011) 101–111. https://doi.org/10.1016/j.supflu.2011.02.012.

[7] C.E. Schwarz, G.J.K. Bonthuys, R.F. van Schalkwyk, D.L. Laubscher, A.J. Burger, J.H. Knoetze, Separation of alkanes and alcohols with supercritical fluids. Part II. Influence of process parameters and size of operating range, J. Supercrit. Fluids 58 (2011) 352–359. https://doi.org/10.1016/j.supflu.2011.07.005.

[8] M. Zamudio, C.E. Schwarz, J.H. Knoetze, Methodology for process modelling of supercritical fluid fractionation processes illustrated for the separation of alkane/alcohol isomer mixtures using CO₂, J. Supercrit. Fluids 104 (2015) 272–280. https://doi.org/10.1016/j.supflu.2015.07.004.

[9] F.C. v. N. Fourie, High pressure phase behaviour of detergent range alcohols and alkanes, PhD Thesis in Chemical Engineering, Stellenbosch University, 2018.

[10] M. Ferreira, Phase Equilibria & Thermodynamic Modelling of the Ternary System CO₂ + 1-Decanol + n-Tetradecane, PhD Thesis in Chemical Engineering, Stellenbosch University, 2018.

[11] C. Latsky, High pressure phase equilibria of the system CO₂ + n-dodecane + 3,7 dimethyl-1-octanol + 1-decanol, PhD Thesis in Chemical Engineering, Stellenbosch University, 2019.

[12] P.M. Mathias, A versatile phase equilibrium equation of state, Ind. Eng. Chem. Proc. Dev. 22 (1983) 385–391. https://doi.org/10.1021/i200022a008.

[13] G. Soave, Equilibrium constants from a modified Redlich-Kwong equation of state, Chem. Eng. Sci. 27 (1972) 1197–1203. https://doi.org/10.1016/0009-2509(72)80096-4.

[14] M. Zamudio, The Separation of Detergent Range Alkanes and Alcohol Isomers with Supercritical Carbon Dioxide, PhD Thesis in Chemical Engineering, Stellenbosch University, 2014.