Density correlation of MEA and water mixtures

https://doi.org/10.1155/2020/7051368

“This section discusses the empirical correlations developed for different types of MEA solutions. It also highlights the theoretical background of those correlations especially the excess volume of MEA and water mixtures. Table 8 summarizes the various published correlations for the density of pure, aqueous, and CO₂-loaded aqueous MEA mixtures. The density of pure liquids at different temperatures was fitted into a second-order polynomial as shown in equations (6), and coefficients were found through a regression [3, 6, 9, 12, 49]. Table 9 lists the parameters found for the polynomial correlation. Valtz et al. [18] used the correlation presented in Reid et al. [50] as given in equation (7) to predict the density of pure MEA at different temperatures. The parameters are given in Table 10.”

”

In binary mixtures, excess molar volume  as given in equations (8) and (9) arises due to the different shape and size of the component molecules, physical interactions, and specific or chemical interactions among the component molecules [51–53]. Mathematically, it is defined as the difference of molar volumes between real and ideal mixtures. The theory of Prigogine–Flory–Patterson [54, 55] discusses  as a summation of interactional contribution, a free volume contribution, and a pressure contribution [56].

Redlich and Kister [57] illustrate an algebraic representation to adopt the excess thermodynamic properties of nonelectrolyte solutions. Therefore, the excess molar volume is presented in a power series with temperature-dependent parameters. This approach has been adopted to correlate excess molar volumes of the MEA and water binary mixture. The effect of temperature on excess volume is figured by introducing a second-order polynomial for parameters in the Redlich–Kister type correlation as shown in equations (10) and (11).

Amundsen et al. [2] and Lee and Lin [14] calculated the coefficients  for different temperatures while Hsu and Li [49] presented  for the entire temperature range of (303.15–353.15) K. A similar work was performed by Han et al. [4] in which the temperature dependence was correlated as a linear relation with respect to temperature. Hartono et al. [1] and Yang et al. [6] also developed a simplified Redlich–Kister type algebraic representation to fit the measured data as given in equations (12) and (13), respectively. The influence of pressure on the density of aqueous MEA was studied by Sobrino et al. [25]. The measured densities from 0.1 MPa up to 120 MPa under different temperatures (293.15–393.15) K and MEA compositions (10–40 mass%) were fitted to a modified Tammann–Tait equation as given in equation (14). Cheng et al. [47] developed a correlation as illustrated in equation (15) based on densities of pure liquids and mass fraction of MEA in the mixture. The correlation is capable of representing densities at different temperatures.

The construction of a proper correlation to fit the density of CO₂-loaded aqueous MEA solutions is challenging as the CO₂ dissolve and react with MEA forming various ions including carbamate, bicarbonate, and protonated MEA. The solution becomes an electrolyte and molecular interactions are more dominant than a MEA and water mixture without CO₂. Various attempts have been taken to build an effective correlation that can be easily used in process design and simulations. Licht and Weiland [48] proposed a correlation to predict the density of CO₂-loaded aqueous amines including MEA as described in equation (16). Weiland et al. [26] proposed a new correlation as from equations (17) to (20) for several amines, and it is extensively used in various studies related to MEA. The correlation shown from equations (21) to (23) was developed by Hartono et al. [1] for CO₂-loaded mixtures. The correlation requires the density of unloaded mixtures to represent the density data of CO₂-loaded mixtures. Literature can be found related to the verification and parameter estimation of Weiland’s correlation for various MEA concentrations and temperatures. Weiland’s correlation was used to fit measured density under different MEA concentrations (10–40 mass%) and CO₂ loading 0.05–0.25 mol CO₂/mol MEA at 298.15 K. Amundsen et al. [2] extended the temperature range of density measurement from 298.15 K to 353.15 K and used the same parameter values as given by Weiland et al. [26] to validate the correlation. The maximum deviation between the measurement and the correlation obtained by Amundsen et al. [2] is 1.6% at 353.15 K. Jayarathna et al. [31] extended the measured MEA concentration up to 70 mass% of aqueous MEA and CO₂ loading 0.1–0.5 mol CO₂/mol MEA in the temperature range of 303.15–333.15 K. The parameters of Weiland’s correlation were estimated within that range and accuracy of the data fit was reported as 2.03 kg·m⁻³ of AAD. Han et al. [4] also used Weiland’s correlation for the density prediction in an extended temperature range up to 413.15 K of the CO₂-loaded solutions. It introduced a nonlinear temperature dependence for the correlation parameters and gained a deviation between measured and correlated as 3.8 kg·m⁻³ of AAD. The main difference between Hartono’s correlation and Weiland’s correlation is that Hartono’s correlation needs the density of unloaded density to calculate the density of loaded solutions.

A study was performed to investigate the accuracies of correlations proposed for aqueous MEA and CO₂-loaded aqueous MEA mixtures. The calculated AARD and AMD for different density correlations of aqueous MEA are listed in Table 11. Hartono’s correlation for density of aqueous MEA used density data from Maham et al. [11] while Han et al.’s correlation used data from their own experiments [4]. The highest AARD of 0.16% was observed for Han’s correlation for the density data published by Amundsen et al. [2] while a maximum deviation of 4.07 kg·m⁻³ at  and T = 293.15 K for the presented data by Ma et al. [27]. For Hartono’s correlation, a maximum AARD of 0.05% and a maximum deviation of 1.79 kg·m⁻³ at  and T = 293.15 K were found for measured viscosities given by Ma et al. [27].

”

“Table 12 lists the calculated AARD and AMD of correlations proposed for density of CO₂-loaded aqueous MEA mixtures. Therefore, Hartono’s correlation and Weiland’s correlation, which was modified by Han et al. [4] for CO₂-loaded aqueous MEA, were studied with different literature for density data of 30 mass% CO₂-loaded mixtures. Correlations were able to represent literature data with less than 1% AARD. Weiland’s correlation showed a higher deviation for data presented by Amundsen et al. [2] and Zhang et al. [29] compared to Hartono et al. [1] and Han et al. [4]. Hartono’s correlation showed a good agreement with data given by Zhang et al. [29]. The maximum deviation is beyond the expanded combined uncertainties reported in data sources, and calculated AARD shows that the agreement between correlated and experimental densities is satisfactory.”