Evaluating the reliability of equilibrium dissolution assumption from residual gasoline in contact with water saturated sands

https://doi.org/10.1016/j.jconhyd.2016.12.003Get rights and content

Highlights

  • Quantification of gasoline dissolution at the interface with water saturated sands

  • Local equilibrium assumption verified at the scale of investigation for BTEX and TMBs

  • Characterization of the impact NAPL saturation and medium heterogeneities on dissolution kinetics

  • Equilibrium assumption more sensitive to heterogeneities for the most soluble BTEX compounds

Abstract

Understanding dissolution dynamics of hazardous compounds from complex gasoline mixtures is a key to long-term predictions of groundwater risks. The aim of this study was to investigate if the local equilibrium assumption for BTEX and TMBs (trimethylbenzenes) dissolution was valid under variable saturation in two dimensional flow conditions and evaluate the impact of local heterogeneities when equilibrium is verified at the scale of investigation. An initial residual gasoline saturation was established over the upper two-thirds of a water saturated sand pack. A constant horizontal pore velocity was maintained and water samples were recovered across 38 sampling ports over 141 days. Inside the residual NAPL zone, BTEX and TMBs dissolution curves were in agreement with the TMVOC model based on the local equilibrium assumption. Results compared to previous numerical studies suggest the presence of small scale dissolution fingering created perpendicular to the horizontal dissolution front, mainly triggered by heterogeneities in the medium structure and the local NAPL residual saturation. In the transition zone, TMVOC was able to represent a range of behaviours exhibited by the data, confirming equilibrium or near-equilibrium dissolution at the scale of investigation. The model locally showed discrepancies with the most soluble compounds, i.e. benzene and toluene, due to local heterogeneities exhibiting that at lower scale flow bypassing and channelling may have occurred. In these conditions mass transfer rates were still high enough to fall under the equilibrium assumption in TMVOC at the scale of investigation. Comparisons with other models involving upscaled mass transfer rates demonstrated that such approximations with TMVOC could lead to overestimate BTEX dissolution rates and underestimate the total remediation time.

Introduction

Non-aqueous phase liquids (NAPL) such as petroleum hydrocarbon fuels present long-term risks of contamination in groundwater. After accidental release or leakage in natural environments, hydrocarbon contaminants infiltrate through soils to reach aquifers where LNAPL will typically volatilise and spread on the water table and release soluble components into groundwater via dissolution. Degradation of fuel components in groundwater environments is common (e.g., Patterson et al., 1993, Franzmann et al., 1999) but despite this plumes emanating from fuel releases can be extensive and prolonged (e.g., Davis et al., 1999, Sookhak Lari and Safavi, 2008) due to slow and persistent dissolution into groundwater.

Petroleum hydrocarbon dissolution is often described in terms of mass transfer at the NAPL/water interface assuming local equilibrium (Seagren et al., 1999, Sulaymon and Gzar, 2011) or rate limited mass transfer (Miller et al., 1990, Imhoff et al., 1994, Vasudevan et al., 2014). Both approaches to dissolution have been investigated in one dimensional column experiments for single components (Powers et al., 1992, Geller and Hunt, 1993) or synthetic multi-component mixtures (Borden and Kao, 1992, Garg and Rixey, 1999) from homogeneous NAPL pools or residual saturations (Lee and Chrysikopoulos, 2006). In porous media, rate-limited dissolution is a function of a mass transfer coefficient and the specific interfacial area between the water and the NAPL phase (Cho et al., 2005, Mobile et al., 2012). Due to difficulties inherent to accurate estimation of the interfacial area, lumped mass transfer coefficients have been defined through empirical mass transfer correlations to calibrate model predictions as a function of water pore velocities and NAPL saturations (Miller et al., 1990, Chrysikopoulos and Kim, 2000, Nambi and Powers, 2003). However, these unidirectional flow experiments are not fully representative of realistic field case situations in which channelling and flow bypassing may occur due to the morphology of sources as well as permeability and NAPL saturation heterogeneities which may have a non-negligible influence on mass transfer (Imhoff et al., 1996, Chen and Jawitz, 2009, Farthing et al., 2012, Kokkinaki et al., 2013). Oostrom et al. (2006) reported the most relevant two and three dimensional, intermediate scale experiments, focusing on aqueous dissolution processes (Eberhardt and Grathwohl, 2002) and those carried out to enhance remediation procedures (Jawitz et al., 1998, Palomino and Grubb, 2004). Most of the experiments investigated the dissolution of single (Saba et al., 2001) and multi-component DNAPL pools (Dela Barre et al., 2002, Roy et al., 2004).

Overall, few multi-dimensional, multi-component studies have been undertaken for gasoline dissolution. Voudrias and Yeh (1994) conducted the first two dimensional LNAPL pool dissolution experiment with toluene overlying water saturated sands. Results demonstrated that the vertical concentration gradient in the water phase was sharp and total dissolution would take decades in such configuration. From a similar design performed in three dimensions with benzene, Sulaymon and Gzar (2011) confirmed that transverse dispersion increases rate-limited mass transfer of benzene compared to column experiments in which the flow was usually forced in one direction through NAPL resident in the column. Other works with single components were also performed with residual entrapped LNAPL sources (Saba and Illangasekare, 2000) to quantify mass transfer rate correlations when horizontal fingering occurred for various lengths of entrapped zones of p-xylene. Two dimensional sand packs were also used to investigate dissolution kinetics due to the effects of permeability changes between an entrapped source zone and the surrounding material, geometry and perimeter of the source, as well as the grain size and the initial residual saturation (Nambi and Powers, 2000). Eberhardt and Grathwohl (2002) developed long-term experiment where dissolution was monitored for a year in three dimensional layered sands to estimate the reliability of Raoult's law to predict equilibrium aqueous concentrations, the influence of the flow velocity as well as the impact of the transverse dispersion on BTEX (benzene, toluene, ethylbenzene, xylenes) and PAHs (polycyclic aromatic hydrocarbons) dissolution from a coal tar pool.

No two-dimensional experiments of preferential dissolution of compounds from gasoline or complex gasoline-type mixtures have been reported. Such studies would need to consider complexities associated with variable saturations and interfacial water-NAPL surface areas, and preferential dissolution of the most soluble compounds, such as BTEX and TMBs (Lekmine et al., 2014). This would favour the increase in molar fractions of other aromatics and aliphatics during dissolution, changing the interactions between compounds over the course of dissolution (Vasudevan et al., 2016, Sookhak Lari et al., 2016a, Sookhak Lari et al., 2016b). Together these features, coupled to gasoline initial composition, saturation, and permeability heterogeneities have a key role in local equilibrium and potential rate-limited dissolution (Grant and Gerhard, 2007).

The objective of this work is to evaluate the role of local heterogeneities on the reliability of the apparent equilibrium dissolution assumption verified at the scale investigation during dissolution from a fresh residual gasoline fuel in a two dimensional flow field with variable NAPL saturation. We developed a two dimensional laboratory experiment to quantify the dissolution sequence of BTEX and trimethylbenzene isomers (TMBs) from a gasoline at residual saturation focusing on two particular areas of interest: 1) in a homogeneous residual gasoline (NAPL) saturated zone and 2) in the transition zone from residual gasoline to full water saturated conditions. A multi-phase and multi-component model was built with TMVOC to evaluate the underlying processes, equilibrium assumption and heterogeneities governing the dissolution dynamics, and to determine temporal changes in BTEX molar fractions within portions of the residual NAPL contaminated and transition zones.

Section snippets

Experimental design

Tests were carried out in a two dimensional stainless steel tank of dimension 71 cm long × 30 cm high × 5.5 cm thick (Fig. 1). Five injectors equally spaced at 5 cm intervals were set on both edges and connected to vertical water wells (2 × 30 × 5.5 cm) separated from the sand pack in the tank by stainless steel screen meshes. A transparent glass window was set on the front wall to visually monitor packing, fluid emplacement and dissolution procedures. On the opposite wall 38 water sampling ports were

Numerical modelling

TMVOC (Pruess and Battistelli, 2002) is a Darcy-scale finite-difference simulator for three-phase flow of water, gas and a multi-component mixture of volatile organic chemicals in heterogeneous porous media. A two dimensional numerical model was built with TMVOC to characterize the equilibrium dissolution sequence across all zones of the sand pack. The feasibility of the governing equations applied in the numerical model for simulation of field-scale interfacial mass transfer has recently been

Results and discussion

Fig. 3 presents the distribution of benzene aqueous concentrations for 6 sampling steps during the dissolution experiment, compared to the modelled sections obtained with TMVOC. The two dimensional contours of the other compounds showed the same level of consistency. Here we focused our investigation on two key zones; (i) the residual NAPL zone vertically extended from top to Z = 16 cm where the dissolution behaviour is expected to be similar to one dimensional flow through residual gasoline in a

Summary and conclusions

This paper aimed to experimentally evaluate the implications of local pore volume heterogeneities while the equilibrium dissolution assumption is verified at a higher scale of investigation. An experiment was performed to simulate the dissolution of a fresh gasoline (unleaded 91) in water saturated conditions at the water table. The objective was to determine dissolution characteristics of gasoline and test the veracity of the equilibrium assumption for BTEX and TMBs dissolution from a

Acknowledgements

This research was partly supported by the Australian Government Department of Industry, Innovation and Science under the Australia-India Strategic Research Fund (Project Number: ST050041). The authors would like to give special thanks to Yasuko J. Geste for her support in undertaking the analytical analysis, Dave Walton and Stuart Dahl for their important input in the experimental design, Bradley Patterson and the anonymous reviewers for helpful comments and suggestions.

References (56)

  • I.M. Nambi et al.

    NAPL dissolution in heterogeneous systems: an experimental investigation in a simple heterogeneous system

    J. Contam. Hydrol.

    (2000)
  • A.M. Palomino et al.

    Recovery of dodecane, octane and toluene spills in sandpacks using ethanol

    J. Hazard. Mater.

    (2004)
  • B.M. Patterson et al.

    Biodegradation and retardation of PCE and BTEX compounds in aquifer material from Western Australia using large-scale columns

    J. Contam. Hydrol.

    (1993)
  • J.W. Roy et al.

    Laboratory evidence of natural remobilization of multicomponent DNAPL pools due to dissolution

    J. Contam. Hydrol.

    (2004)
  • T. Saba et al.

    Investigation of surfactant-enhanced dissolution of entrapped nonaqueous phase liquid chemicals in a two-dimensional groundwater flow field

    J. Contam. Hydrol.

    (2001)
  • E.A. Seagren et al.

    A critical evaluation of the local-equilibrium assumption in modelling NAPL-pool dissolution

    J. Contam. Hydrol.

    (1999)
  • M.A. Seyedabbasi et al.

    The influence of wettability on NAPL dissolution fingering

    Adv. Water Resour.

    (2008)
  • K. Sookhak Lari et al.

    Incorporating hysteresis in a multi-phase multi-component NAPL modelling framework; a multi-component LNAPL gasoline example

    Adv. Water Resour.

    (2016)
  • A.H. Sulaymon et al.

    Experimental investigation and numerical modeling of light nonaqueous phase liquid dissolution and transport in a saturated zone of the soil

    J. Hazard. Mater.

    (2011)
  • M. Vasudevan et al.

    Effect of compositional heterogeneity on dissolution of non-ideal LNAPL mixtures

    J. Contam. Hydrol.

    (2016)
  • R.C. Borden et al.

    Evaluation of groundwater extraction for remediation of petroleum contaminated aquifers

    Water Environmental Research

    (1992)
  • X. Chen et al.

    Convergence of DNAPL source strength functions with site age

    Environ. Sci. Technol.

    (2009)
  • J. Cho et al.

    Measured mass transfer coefficients in porous media using specific interfacial area

    Environ. Sci. Technol.

    (2005)
  • J.A. Christ et al.

    Estimating mass discharge from dense nonaqueous phase liquid source zones using upscaled mass transfer coefficients: an evaluation using multiphase numerical simulations

    Water Resour. Res.

    (2006)
  • C.V. Chrysikopoulos et al.

    Local mass transfer correlation for nonaqueous phase liquid pool dissolution in saturated porous media

    Transp. Porous Media

    (2000)
  • B.K. Dela Barre et al.

    Measuring and modelling the dissolution of nonideally shaped dense nonaqueous phase liquid pools in saturated porous media

    Water Resour. Res.

    (2002)
  • J.D. Drummond et al.

    Effects of solute breakthrough curve tail truncation on residence time estimates: a synthesis of solute tracer injection studies

    J. Geophys. Res.

    (2012)
  • M.W. Farthing et al.

    Influence of porous media heterogeneity on nonaqueous phase liquid dissolution fingering and upscaled mass transfer

    Water Resour. Res.

    (2012)
  • Cited by (26)

    • Tracking NSZD mass removal rates over decades: Site-wide and local scale assessment of mass removal at a legacy petroleum site

      2022, Journal of Contaminant Hydrology
      Citation Excerpt :

      The underlying processes have been well summarised in two reasonably recent reviews (Garg et al., 2017; Sookhak Lari et al., 2019) and expanded upon by Smith et al. (2022). The two reviews brought together literature and understanding spanning over 30 years; combining features of LNAPL movement (Sookhak Lari et al., 2016; Tomlinson et al., 2017; Lenhard et al., 2017), LNAPL partitioning to groundwater and soil gas phases (Lang et al., 2009; Lekmine et al., 2017; Davis et al., 2021), LNAPL compositional changes and fingerprinting during mass loss (Johnston et al., 2007; DeVaull et al., 2020; Rayner et al., 2020), groundwater natural attenuation (Wiedemeier et al., 1999); microbiology and biodegradation processes (Franzmann et al., 1996; Franzmann et al., 1999; Bruckberger et al., 2021), and biodegradation of LNAPL mass and vapours in the vadose zone (Amos et al., 2005; Davis et al., 2005). Linking a number of these features together was largely birthed at two sites that have undergone detailed investigation; the Bemidji Minnesota crude oil release site (Chaplin et al., 2002; Amos et al., 2005; Molins et al., 2010) and the Guadalupe Oil Field site in California (Lundegard and Johnson, 2006; Johnson et al., 2006).

    • Towards a digital twin for characterising natural source zone depletion: A feasibility study based on the Bemidji site

      2022, Water Research
      Citation Excerpt :

      Some of the assumptions resulting in Darcy scale modeling of multi-phase systems have been extensively studied. For instance, it is believed that the equilibrium partitioning assumption holds (Lekmine et al., 2017). However, subsurface modeling at a Darcy scale still has some limitations with respect to the constitutive relationships.

    View all citing articles on Scopus
    View full text