Elsevier

Chemosphere

Volume 144, February 2016, Pages 2247-2256
Chemosphere

Measurement of Setschenow constants for six hydrophobic compounds in simulated brines and use in predictive modeling for oil and gas systems

https://doi.org/10.1016/j.chemosphere.2015.10.115Get rights and content

Highlights

  • Setschenow constants were measured using solid phase microextraction.

  • Literature Setschenow constants obtained at low salt concentrations can be extrapolated.

  • Setschenow constants were additive in moderate to high ionic strengths.

  • The Setschenow Equation was log-linear up to 2 -- 5 M NaCl and 1.5 -- 2 M CaCl2.

  • Two new models were developed for prediction of CaCl2 and KCl Setschenow constants.

Abstract

Treatment and reuse of brines, produced from energy extraction activities, requires aqueous solubility data for organic compounds in saline solutions. The presence of salts decreases the aqueous solubility of organic compounds (i.e. salting-out effect) and can be modeled using the Setschenow Equation, the validity of which has not been assessed in high salt concentrations. In this study, we used solid-phase microextraction to determine Setschenow constants for selected organic compounds in aqueous solutions up to 2–5 M NaCl, 1.5–2 M CaCl2, and in Na–Ca binary electrolyte solutions to assess additivity of the constants. These compounds exhibited log-linear behavior up to these high NaCl concentrations. Log-linear decreases in solubility with increasing salt concentration were observed up to 1.5–2 M CaCl2 for all compounds, and added to a sparse database of CaCl2 Setschenow constants. Setschenow constants were additive in binary electrolyte mixtures. New models to predict CaCl2 and KCl Setschenow constants from NaCl Setschenow constants were developed, which successfully predicted the solubility of the compounds measured in this study. Overall, data show that the Setschenow Equation is valid for a wide range of salinity conditions typically found in energy-related technologies.

Introduction

Many energy-related technologies, such as carbon capture, utilization, and storage (CCUS); enhanced oil recovery (EOR); and unconventional oil and gas extraction; and desalination technologies (e.g. reverse osmosis) produce highly saline waters containing dissolved organic compounds (e.g. petroleum and synthetic hydrocarbons). The solubility of organic compounds in water usually decreases with increasing salinity. Therefore, assessing the risks of mobilization of organics from brine and/or CO2 leakage from geologic storage formations (e.g. leakage from Sleipner gas field in the North Sea (Monastersky, 2013)) or migration of produced waters from unconventional oil and gas reservoirs into freshwater aquifers require reliable ways of predicting the solubility of a number of different apolar and semi-polar organic compounds in brines. The composition of the brines mentioned above is typically dominated by dissolved alkaline and alkali metal chlorides (e.g. NaCl, CaCl2) which sometimes can reach total dissolved solids concentration of 400,000 mg/L (DOE/NETL-2003/119, 2003, Shaffer et al., 2013). The effect of dissolved salts on the aqueous solubility of organic compounds, called the “salting-out” effect, is typically modeled by the empirical Setschenow Equation (Burant et al., 2013, Sechenov, 1889) (Eqn. (1)). The model is empirical, with Setschenow constants (Ksi,k Eqn. (1)) determined experimentally for each organic compound as well as each salt of interest.APTARANORMALlog(γwsaltγwDI)=APTARANORMALlog(SWDISwsalt)=APTARANORMALlog(K1saltK1DI)=Ksi,k[salt].

In this equation, γwDI is the activity coefficient of an organic compound in deionized water (DI water), γwsalt is the activity coefficient of the organic compound in water at the salt concentration of interest, SWDI is the aqueous solubility of the organic compound in DI water, Swsalt is the aqueous solubility of the organic compound at the salt concentration of interest, K1salt is a partitioning coefficient of an organic compound between water and another phase (air, solid, for example) at the salt concentration of interest, K1DI is the partitioning coefficient of an organic compound between water and another phase at the salt concentration of interest, Ksi,k is the Setschenow constant of the organic compound (i) for the specific salt (k) (M−1), and [salt] is the concentration of salt (M). The presence of salt causes an increase in the activity coefficient of the organic compound in the saline solution relative to the DI water. This corresponds to a decrease in aqueous solubility and an increase in partitioning from water. In mixed salt solutions the Setschenow constant is determined based on the mole fractions of the salts present:KsmixedkKsi,salt,k·xk.where Ksmixed is the Setschenow constant of the mixed electrolyte system, Ksi,salt,k is the Setschenow constant of organic compound (i) in a single electrolyte solution (k), and xk is the mole fraction of each type of salt solution comprising the mixture. The Ksmixed can then be applied to Eqn. (1), replacing Ksi,k, to predict the aqueous activity coefficient, aqueous solubility, or partition coefficient of a compound in the mixed electrolyte solution.

The ability of the Setschenow Equation to predict the solubility for many organic compounds has been confirmed only for aqueous solutions with salt concentrations up to 1 M NaCl (Xie et al., 1997). Data for the solubility of organic compounds for high salinity NaCl and CaCl2 solutions is scarce. In addition, there are only 19 reported CaCl2 Setschenow constants, despite the fact that CaCl2 is a major component of oil and gas reservoir brines (Kharaka and Hanor, 2003). The additive nature of the salting-out effect (Eqn. (2)) has only been demonstrated for benzene and naphthalene at moderate ionic strengths (Gordon and Thorne, 1967a, Gordon and Thorne, 1967b, McDevit and Long, 1952).

Reported aqueous solubility data are available for benzene, toluene, ethylbenzene, and the xylene isomers (BTEX) up to 5 M NaCl (Keeley et al., 1991, Keeley et al., 1988). However, these compounds are relatively small and monopolar, and more data on representative compounds from different classes of organic compounds are needed to assess the validity of the Setschenow Equation over a broader range of salt concentrations, and to determine if Setschenow constants determined at low salt concentration can be extended to brines (≥1 M). The Setschenow Equation may not be valid for larger and monopolar/apolar or polar organic compounds, i.e. organic compounds with more than one aromatic ring or organic compounds capable of hydrogen bonding interactions (Eisen and Joffe, 1966, Janado et al., 1983, Jochmann et al., 2006, Lee, 1997, Meranda and Furter, 1974). For example, Janado et al., 1983 found that the Setschenow Equation did not predict the salting-out effect for naphthalene and biphenyl in aqueous NaSCN and KSCN solutions, which exhibited both salting-out and then salting-in effects, while benzene only exhibited salting-out behavior. Although deviations from the Setschenow Equation are not common for solutions of moderate salt concentration (i.e. up to 1 M), Whitehouse, 1985 observed deviations from the Setschenow Equation for 1,2-benzanthracene, a PAH, in concentrations of NaCl up to seawater (Whitehouse, 1985). Since there are ∼193 reported NaCl Setschenow constants (Endo et al., 2012, Ni and Yalkowsky, 2003), the extension of the previously measured Setschenow constants up to high salt concentrations would be ideal, since additional experimental data for derivation of Setschenow constants would not need to be collected. However due to lack of reliable solubility measurements at high salinity it is uncertain if the Setschenow Equation can be used to accurately predict the aqueous solubility of a variety of organic compounds at high salt concentrations expected in brines (Endo et al., 2012).

The additivity of the Setschenow constants was confirmed for benzene and naphthalene, over a range of different salt compositions with different anions and cations, and for a several hydrophobic organic compounds up to seawater salinity (Eganhouse and Calder, 1976, Rossi and Thomas, 1981, Sutton and Calder, 1975). The additivity of Setschenow constants for mixed electrolytes (Eqn. 2) has not been confirmed with extensive data collection, and never for high salinity fluids (i.e. with higher than seawater salinity).

Avoiding experimental determination of new Setschenow constants is desirable, especially for Ca2+ as data are especially sparse. Modeling approaches to predict Setschenow constants have had moderate success using correlation with molar volume (Endo et al., 2012, Jonker and Muijs, 2010, Long and McDevit, 1952, McDevit and Long, 1952, Xie et al., 1997) and octanol–water partitioning coefficients (Endo et al., 2012, Ni and Yalkowsky, 2003). Compounds with larger molar volumes tend to have higher Setschenow constants. In addition, apolar and monopolar compounds with high octanol–water partition coefficient (Kow) tend to have larger Setschenow constants than polar compounds. The first attempt to capture these trends in a model to predict NaCl Setschenow constants was obtained by regressing log Kow with measured NaCl Setschenow constants (Eqn. (3)) (Ni and Yalkowsky, 2003). This study showed a fairly good correlation (n = 101, R2 = 0.772, reported standard error = 0.041) between the predicted and experimental NaCl Setschenow constants. Log Kow is a good qualitative proxy for Setschenow constants, because both follow trends in size and polarity. However, Endo et al., 2012 could not reproduce this simple fit using different compounds. Rather, Endo et al., 2012 developed a poly-parameter linear free energy relationship (pp-LFER), which incorporates Abraham solvation parameters to account for size and intermolecular interactions (Eqn. (4)) to predict NaCl Setschenow constants (n = 43, R2 = 0.83, reported standard deviation = 0.031).Ksi,NaCl=0.0400.25em0exAPTARANORMALlog0.25em0exKow+0.114Ksi,NaCl=0.1120.020R20.042π20.047α20.060β2+0.171V2

The coefficients in Eqn. (4) are the Abraham solvation parameters. The R2 is the index of refraction of the organic compound, π2 is the organic compounds polarizability, α2 is the hydrogen bonding acidity of the organic compound, β2 is the hydrogen bonding basicity of the organic compound, and V2 is the molar volume of the compound.

Wang et al. (2014) developed a similar pp-LFER using the Abraham solvation parameters, however for the prediction of (NH4)2SO4 Setschenow constants. However, no models have been developed to predict Setschenow constants for CaCl2. However these will be needed to predict solubility of organic compounds in saline waters associated with CCUS and EOR.

The objectives of this work were (1) to determine whether reported Setschenow constants, measured at low salt concentration, are applicable to high salinity solutions (from 2 M- 5 M NaCl and in the range of 1.5 – 2 M CaCl2); (2) to determine if Setschenow constants are additive for selected compounds in mixed electrolyte brines; and (3) to develop and test two respective LFER models for predicting CaCl2 and KCl Setschenow constants from NaCl Setschenow constants. The effect of salt on the aqueous solubility of the PAHs, naphthalene, fluorene, and phenanthrene, and of several heterocyclic sulfur compounds, thiophene, benzothiophene, and dibenzothiophene were studied for both NaCl and CaCl2 solutions. Data from this study and literature were used to train models that predict Setschenow constants in CaCl2 and KCl solutions from experimental values derived with NaCl. In addition, data from the literature was used to train the KCl model. The last objective enables accurate prediction of largely unavailable CaCl2 and KCl Setschenow constants from the more often reported NaCl Setschenow constants determined in lieu of experimentation.

Section snippets

Materials and methods

The estimation of Setschenow constants was completed using solid phase microextraction (SPME), followed with analysis of the adsorbed organic compounds using gas-chromatography coupled with a flame ionization detector (GC-FID). SPME, has been used in previous studies to measure Setschenow constants and has proven to produce consistent and accurate results (Endo et al., 2012, Jonker and Muijs, 2010). The Setschenow constants were estimated by linear regression of Eqn. (1), where differences in

NaCl Setschenow constants

The Setschenow constants for selected compounds were determined using the SPME method (Fig. 1 and Table 1). Previously reported NaCl Setschenow constants for the PAHs are found in Table 1, while thiophenes have no reported NaCl Setschenow constants. The standard error of those estimates for Ksi,NaCl were low (0.008–0.043 M−1), and values are comparable to previously reported measurements of Setschenow constants determined at lower salt concentration, for the PAHs (Jochmann et al., 2006, Jonker

Conclusions

The salting-out effect for hydrophobic compounds in both this study and the literature in NaCl and CaCl2 solutions exhibited log-linear behavior up to 2–5 M NaCl and 1.5–2 M CaCl2. This has been shown for BTEX compounds, (Keeley et al., 1991, Keeley et al., 1988) as well as the PAHs and sulfur heterocyclics in this study. Fluorene and dibenzothiophene displayed log-linear salting-out behavior up to 2 M NaCl, phenanthrene displayed log-linear salting-out behavior up to 3 M NaCl, benzothiophene

Acknowledgments

This technical effort was performed under the auspice of the US DOE National Energy Technology Laboratory, under the RES contract DE-FE0004000. We thank the Jared and Maureen Cohon Graduate Fellowship in Civil and Environmental Engineering and the Bradford and Diane Smith Fellowship in Engineering for support. Many thanks to Unnati Rao for help with the experiments and Clinton Noack for Matlab help and thorough reviews.

References (44)

  • S. Portier et al.

    Modelling CO2 solubility in pure water and NaCl-type waters from 0 to 300 °C and from 1 to 300 bar

    Chem. Geol.

    (2005)
  • S. Poulson et al.

    The solubility of toluene in aqueous salt solutions

    Talanta

    (1999)
  • B.G. Whitehouse

    Observation of abnormal solubility behavior of aromatic hydrocarbons in seawater

    Mar. Chem.

    (1985)
  • B.G. Whitehouse

    The effects of temperature and salinity on the aqueous solubility of polynuclear aromatic hydrocarbons

    Mar. Chem.

    (1984)
  • W.-H. Xie et al.

    A review of the effect of salts on the solubility of organic compounds in seawater

    Mar. Environ. Res.

    (1997)
  • J.T. Andersson et al.

    A method for measuring 1-octanol−water partition coefficients

    Anal. Chem.

    (1999)
  • M. Aquan-Yuen et al.

    Solubility of hexane, phenanthrene, chlorobenzene and p-dichlorobenzene in aqueous electrolyte solutions

    J. Chem. Eng. Data

    (1979)
  • K.L. Benko et al.

    Produced water in the Western United States: geographical distribution, occurrence, and composition

    Environ. Eng. Sci.

    (2008)
  • A. Burant et al.

    Partitioning behavior of organic contaminants in carbon storage environments: a critical review

    Environ. Sci. Technol.

    (2013)
  • M.G. Cacace et al.

    The hofmeister series: salt and solvent effects on interfacial phenomena

    Q. Rev. Biophys.

    (1997)
  • DOE/NETL-2003/119

    U.S. Brine Wells Database

    (2003)
  • E.O. Eisen et al.

    Salt effects in liquid–liquid equilibria

    J. Chem. Eng. Data

    (1966)
  • Cited by (15)

    • Enhanced adsorption performance for selected pharmaceutical compounds by sonicated Ti<inf>3</inf>C<inf>2</inf>T<inf>X</inf> MXene

      2021, Chemical Engineering Journal
      Citation Excerpt :

      This is due to a relatively strong competition between Ca2+ as a divalent cation and AMT+ for adsorption sites on S28-MXene [38]. In addition, another plausible reason could be due to a salting-out effect, in which the hydrophobic interaction between organic compounds can induce a higher ionic strength in the solution [41]. Due to the salting-out effect, reduced solubility of AMT can cause a larger molecular size and/or precipitation of some AMT, resulting in hardly any adsorption.

    • Production and application of O<inf>2</inf> enriched air produced by fresh and salt water desorption in chemical plants

      2018, Journal of Environmental Management
      Citation Excerpt :

      Salinity affects also desorption of both gases. For these reason, we studied the influence of salt concentration in water and we applied the Setschenow equation to determine the effect of salts on air solubility (Burant et al., 2016), by defining a simplified thermodynamic approach suitable for the interpretation of these equilibria. This first evaluation concerning the use of sea water is encouraging.

    View all citing articles on Scopus
    View full text