High pressure phase behaviour of the CO2 + 1-decanol + n-dodecane system
Introduction
Detergent range alcohols (C8–C20 isomers) are used as raw materials for surfactants and lubricants [1]. The production of detergent range alcohols by the Oxo-process results in a product stream that contains large amounts of unreacted alkanes, necessitating further product purification. The alcohols and alkanes in the product stream typically have similar boiling points. The separation of close-boiling components is a concern in industry, since conventional distillation methods become uneconomical and alternative processing methods are required [2], [3]. Amongst others, supercritical fluid fractionation (SFF) is considered for this separation problem, a process in which the unique properties of supercritical fluids are exploited to bring about the separation of compounds [3], [4], [5]. Accurate models are required for the design and optimisation of industrial processes. However, good experimental data is required to improve the predictive capabilities of these models and to verify their applicability [5], [6], [7].
As a first approximation to the modelling of the separation of detergent range alkanes and alcohols Zamudio et al. [4], [6] considered the separation of 1-decanol and n-dodecane using supercritical CO2. The solubility of two binary systems, namely CO2 + 1-decanol and CO2 + n-dodecane, was considered in their investigation. However, SFF processes typically deal with streams that contain multiple components and it is necessary to determine to what extent the interaction between solutes influences the behaviour of the system [8]. For this purpose Zamudio et al. [6] investigated a ternary system, CO2 + (77.8 wt% 1-decanol + 22.2 wt% n-dodecane) to observe the interaction between a close boiling primary alcohol and linear alkane.
Closer examination of the two binary and one ternary systems points towards interesting phase behaviour. It is expected that the solubility pressures for the ternary system, containing 77.8 wt% 1-decanol, will be closer to that of the CO2 + 1-decanol system, should the interaction between the two solutes be negligible. However, the solubility pressures of the ternary system are approximately halfway between that of the CO2 + 1-decanol and CO2 + n-dodecane systems, indicating that the interaction between 1-decanol and n-dodecane in the presence of supercritical CO2 is appreciable.
Previous modelling attempts of CO2 + 1-alkanol, CO2 + n-alkane, 1-alkanol + n-alkane, and CO2 + 1-alkanol + n-alkane systems have been conducted with different variations of the cubic equations of state (EOS): Soave–Redlich–Kwong (SRK) [9], [10], [11], SRK with Huron–Vidal mixing rule [12] Peng–Robinson (PR) [9], [13], [14], PR with the Stryjek–Vera modification [10], [11], PPR78 [15], and Patel–Teja [10]. Some investigations also considered statistical mechanics-based EOSs, such as CSPHC [10], APACT [11], and SAFT [16].
Zamudio et al. [6] modelled phase equilibrium data within the Aspen Plus® process simulator with the specific objective of improving the process modelling of supercritical fluid processes using commercial simulators. The authors found that within Aspen Plus®, and amongst the EOSs investigated in their study, the RK-Aspen EOS showed the best predictive capability. For the binary systems solute–solvent interaction parameters were determined and implemented to predict the solubility of the ternary system. However, the interaction between the two solutes was not accounted for and led to significant errors in the prediction of the solubility of the ternary system. Zamudio et al. [6] concluded that the inclusion of solute–solute interaction parameters might lead to predicted values closer to those measured experimentally.
The current work therefore aims to further characterise the interaction between 1-decanol and n-dodecane in supercritical CO2 by (i) conducting additional solubility measurements of (1-decanol + n-dodecane)-mixtures in supercritical CO2, and (ii) analysing and interpreting the experimental data to evaluate the effect that n-dodecane has over 1-decanol. Furthermore, (iii) the correlative capability of the RK-Aspen EOS is improved by determining suitable values for the binary solute–solute interaction parameters.
Section snippets
Experimental set-up
Phase transition measurements are conducted in two previously constructed static synthetic view cells with variable volume that operate on exactly the same principle. The main difference between the two cells is the size: the maximum internal volume of the small cell is 45 cm3 [16], while that of the large cell is 80 cm3 [17]. Details regarding the design of the two cells can be found in Schwarz and Nieuwoudt [16] and Fourie et al. [17], respectively. Typically the large cell was used for low
Measured data
Three mixtures of (1-decanol + n-dodecane) were prepared such that the 1-decanol mass fraction in each mixture was 0.10, 0.20, and 0.60, ±0.0001. These mixtures are subsequently referred to as Mixtures A, B, and C, respectively.
Due to fluctuations in the ambient and heating medium temperatures measurements cannot always be taken at the target temperature. To enable the comparison between different systems interpolation of data is required. Interpolation has been used successfully in previous
Conclusions
High pressure solubility measurements were conducted for three mixtures of CO2 + (1-decanol + n-dodecane) at temperatures ranging from 308 K to 348 K and at solute mass fractions between 0.017 and 0.621. The measurements show that the presence of n-dodecane improves the solubility of 1-decanol in supercritical CO2, and that the solubility pressure of (1-decanol + n-dodecane)-mixtures is not a linear function of the mixture composition. Furthermore, the interaction between 1-decanol and n-dodecane is of
References (23)
- et al.
Phase equilibria of branched isomers of C10-alcohols and C10-alkanes in supercritical carbon dioxide
J. Supercrit. Fluids
(2011) - et al.
Experimental measurement and modelling with Aspen Plus® of the phase behaviour of supercritical CO2 + (n-dodecane + 1-decanol + 3,7-dimethyl-1-octanol)
J. Supercrit. Fluids
(2013) High-pressure investigations of fluid mixtures—review and recent results
J. Supercrit. Fluids
(1998)- et al.
Measurement of phase equilibria of supercritical carbon dioxide and paraffins
J. Supercrit. Fluids
(2002) - et al.
High pressure VLE in alkanol + alkane mixtures. Experimental results for n-butane + ethanol, +1-propanol, +1-butanol systems and calculations with three EOS methods
Fluid Phase Equilib.
(1995) - et al.
Investigation of phase equilibria in the ternary system carbon dioxide + 1-heptanol + n-pentadecane
Fluid Phase Equilib.
(2007) - et al.
Modeling the solubilities of high molecular weight n-alkanes in supercritical carbon dioxide
Fluid Phase Equilib.
(2004) - et al.
Predicting the phase equilibria of CO2 + hydrocarbon systems with the PPR78 model (PR EOS and kij calculated through a group contribution method)
J. Supercrit. Fluids
(2008) - et al.
Phase equilibrium of propane and alkanes: Part I. Experimental procedures, dotriacontane equilibrium and EOS modelling
J. Supercrit. Fluids
(2003) - et al.
Phase equilibria of alcohols in supercritical fluids: Part I. The effect of the position of the hydroxyl group for linear C8 alcohols in supercritical carbon dioxide
J. Supercrit. Fluids
(2008)
Equilibrium constants from a modified Redlich–Kwong equation of state
Chem. Eng. Sci.
Cited by (17)
High pressure phase behaviour of ternary systems containing CO<inf>2</inf> + fatty acid methyl esters
2024, Fluid Phase EquilibriaHigh-pressure fluid-phase equilibria: Experimental methods, developments and systems investigated (2013–2016)
2024, Fluid Phase EquilibriaHigh pressure phase behaviour of binary systems containing supercritical solvents and uneven acids
2021, Journal of Supercritical FluidsHigh pressure bubble- and dew-point data for systems containing CO<inf>2</inf> with 1-decanol and n-hexadecane
2020, Fluid Phase EquilibriaCitation Excerpt :Two sets of interaction parameters were regressed, the one set accounts for the interactions between the solvent (CO2) and the solutes (n-hexadecane and 1-decanol) and the other set accounts for the interactions between the solutes. The solute-solute interaction parameters were optimized by minimizing the %AAD at constant temperature [6]. This manual regression procedure is highly time consuming and therefore not all possible combinations of ka,ij0 and kb,ij0 could be investigated.
CO<inf>2</inf> + n-dodecane + 1-decanol: High pressure experimental phase equilibria data and thermodynamic modelling
2019, Journal of Supercritical FluidsCitation Excerpt :CO2 + nC12 is the more soluble of the constituent binaries and a mixture thereof with a similar bulk CO2 mass fraction of 0.733 transitions at higher pressures of 75.0, 113.2 and 151.2 bar at 35, 55 and 75 °C [42]. Zamudio et al. [42] and Smith and Schwarz [43] performed synthetic measurements of different CO2 + nC12 + 1-alcohol systems and identified enhanced solubility in the low-nC12 region and co-solvency in the high-nC12 region. For the CO2 + nC12 + C10OH system, this behaviour led to the formation of detached two-phase regions [43].
High pressure phase behaviour for the CO<inf>2</inf> + n-dodecane + 3,7-dimethyl-1-octanol system
2019, Journal of Supercritical FluidsCitation Excerpt :The p-xAred diagram presented in Fig. 3, was constructed by combining isothermal p-xAred sections of the p-xs-xAred diagrams at a constant total solute mass fraction (xs) of 0.5 [17] and the identified trends hold true for the entire total solute mass fraction (xs) range. Since the experiments for the different mixtures were not conducted at the exact same solute mass fractions, polynomial pressure-composition curves were fitted to the isothermal data to allow interpolation at specific solute mass fractions [3]. Analysis of Fig. 3 not only illustrates the non-linear decrease in phase transition pressure, but it also reveals a pressure minimum in the isothermal p-xAred curves.