Remediation of NAPL below the water table by steam-induced heat conduction

https://doi.org/10.1016/j.jconhyd.2003.11.001Get rights and content

Abstract

Previous experimental studies have shown that NAPL will be removed when it is contacted by steam. However, in full-scale operations, steam may not contact the NAPL directly and this is the situation addressed in this study. A two-dimensional intermediate scale sand box experiment was performed where an organic contaminant was emplaced below the water table at the interface between a coarse and a fine sand layer. Steam was injected above the water table and after an initial heating period the contaminant was recovered at the outlet. The experiment was successfully modeled using the numerical code T2VOC and the dominant removal mechanism was identified to be heat conduction induced boiling of the separate phase contaminant. Subsequent numerical modeling showed that this mechanism was insensitive to the porous medium properties and that it could be evaluated by considering only one-dimensional heat conduction.

Introduction

Steam injection is a promising technology for remediating subsurface hydrocarbon contamination. Full-scale clean-up operations have suggested that the technique ensures rapid and satisfactory clean-up of even very complicated contaminations Newmark et al., 1994, Newmark et al., 1998. Several one-dimensional laboratory experiments have shown that volatile NAPL will rapidly vaporize when it contacts with steam Hunt et al., 1988, Hadim et al., 1993, Betz et al., 1997. Two-dimensional laboratory studies have shown that heterogeneous porous media can also be remediated by steam injection Itamura, 1996, She and Sleep, 1999. In these experiments the steam zone eventually occupied the entire sand pack and the emplaced NAPL was completely removed. Thus, the investigations suggest that the soil volume contacted by steam will be cleaned. However, if NAPL is present in a low permeable heterogeneity or in a saturated zone, steam may not contact directly. This will significantly reduce the removal rate, but remediation may still be faster than with other technologies. The focus of this study is to quantify the removal when steam does not contact the contaminant directly. Specifically, we address the situation where a steady-state steam zone overlies a saturated zone contaminated with NAPL. This is a very common situation in a full-scale application of steam injection. When injecting steam below the water table, a steam zone will tend to move upwards due to buoyancy and may override a part of the contamination, e.g. when a DNAPL is pooled on top of a low permeable layer. Likewise, Schmidt et al. (2002) showed how injection of pure steam in an unsaturated zone containing NAPL could lead to downward migration resulting in a lens of LNAPL residing between the steam zone and the water table. In these cases, there will be no steam flow through the contaminated soil and the only heating mechanism will be conduction from the overlying steam zone. At a certain temperature, below the normal boiling point of water, the two immiscible liquids will boil because the sum of their individual vapor pressures will equal the surrounding pressure (Atkins, 1994). The produced gas will, due to buoyancy, be transported from the saturated zone into the steam zone where it can be extracted. In clean-up operations limited by conduction, a performance assessment should not be based on the steam zone development, which would provide a too optimistic estimate of the clean-up time.

The objective of this study is to examine the removal mechanisms when steam is not in direct contact with NAPL. A two-dimensional intermediate scale sand box experiment was conducted where separate phase trichloroethylene (TCE) was emplaced below the water table at the interface between a coarse and a fine sand layer. Steam was injected above the water table and the remediation process was monitored by measurements of temperature inside the box and recovery of contaminant at the outlet. The experimental data were interpreted with the numerical code T2VOC that subsequently was used for analyzing the involved processes in details.

Section snippets

Experimental setup

Fig. 1 illustrates the experimental setup. The sand box had the interior dimensions 122×8.5×58 cm and was constructed of stainless steel with a front glass cover allowing for inspection of the sand packing. During the experiment, the sand box was insulated to minimize heat loss. The sand box was equipped with 98 Pt-100 temperature sensors connected to a data acquisition system that continuously logged the temperature. At the location of the bottom row of temperature sensors, pressure

Results

The injection of steam created a zone with steam temperature in the upper part of the coarse sand layer with convection as the dominant heat transfer process. A vertical steam front, where the temperature rapidly changed from steam temperature to the temperature of the surroundings, moved with an almost constant velocity through the sand box. Fig. 2 compares the measured and the simulated temperature after 1, 2 and 3 h of steam injection. The experimental and simulated temperatures compare well

Discussion

Fig. 5a shows a contour plot of the simulated gas saturation after 12 h, which is in the period of the TCE-recovery. Above the contaminated layer, an unsaturated zone has formed from gas created by boiling of separate phase TCE and water. The gas moves upwards due to buoyancy and in the steam zone it is transported towards the extraction well. This is shown in Fig. 5b where a plume of TCE is present in the gas phase from the contaminated area to the extraction well.

Fig. 6 shows the temporal

Perspectives

The above findings show how TCE is rapidly removed in an experiment due to a boiling mechanism. The numerical model can be used to generalize this into something more applicable at field scale by varying the significant parameters within the interval to be expected in the field. The parameters significant to the process are identified as (a) porous medium properties, (b) the dimensions and (c) the type of contaminant. We assume that a 1-D model is sufficient to capture the dominant mechanisms

Conclusions

Steam injection was performed in a two-dimensional sand box to remediate TCE present as a separate phase below the water table. Steam was injected above the water table and after an initial heating period the separate phase TCE was recovered at the outlet. The experiment was successfully modeled with the numerical code T2VOC using independently measured parameters and no calibration. By means of the numerical model, the dominant removal mechanism was identified as boiling induced by heat

Acknowledgements

The experiment was performed at the VEGAS facility at the University of Stuttgart, Germany. We express our great gratitude for being offered access to the experimental facilities and to the staff members for helpful assistance during the experiments.

References (26)

  • Itamura, M.T., 1996. Removal of Chlorinated Hydrocarbons from Homogeneous and Heterogeneous Porous Media Using Steam....
  • F.P Incropera et al.

    Introduction to Heat Transfer

    (1996)
  • H.H Liu et al.

    Reconciliation between measured and theoretical temperature effects on soil water retention curves

    Soil Science Society of America Journal

    (1993)
  • Cited by (17)

    • Effects of geologic heterogeneity on migration of gaseous CO<inf>2</inf> using laboratory and modeling investigations

      2015, International Journal of Greenhouse Gas Control
      Citation Excerpt :

      The Brooks and Corey (1964) model is chosen as the functional relationship for describing the capillary pressure–saturation relationship instead of the Van Genuchten (1980) model as the test sand (Table 1) has a very high lambda value (narrow pore size distribution) compared to field soils. Furthermore, we use the minor modification introduced by Gudbjerg et al. (2004) to handle entry pressures when using the Brooks and Corey (1964) formulation. The 1D column is represented by elements with dimensions of 1.2 cm × 5 cm × 5 cm, resulting in 100 grid elements in the vertical direction.

    • Numerical analysis of contaminant removal from fractured rock during boiling

      2012, Journal of Contaminant Hydrology
      Citation Excerpt :

      The T2VOC and TMVOC codes are members of the TOUGH2 family of codes developed for simulation of multiphase fluid and heat flow with single and multiple volatile organic chemicals, respectively (Falta et al., 1995; Pruess and Battistelli, 2002). They have been used to simulate several laboratory and field experiments (Falta et al., 1992b; Gudbjerg et al., 2004; Ochs et al., 2003; Tse et al., 2006). Falta et al. (1992b) simulated a steam displacement experiment on a water-saturated sand column contaminated with TCE NAPL.

    • Henry's law constants of chlorinated solvents at elevated temperatures

      2012, Chemosphere
      Citation Excerpt :

      As boiling occurs, volatile compounds partition to and are removed by the moving vapor phase. This process effectively removes dissolved contaminants from soils and fractured rocks (Gauglitz et al., 1994; Heron et al., 1998b, 2009; Ochs et al., 2003; Gudbjerg et al., 2004; Hodges et al., 2004; Beyke and Fleming, 2005; Chen et al., 2010). Knowing the Henry’s law constants of CVOCs that are frequently encountered at contaminated sites at higher temperatures is important for understanding the process of contaminant removal and for predicting the performance of thermal remediation.

    • Combining steam injection with hydraulic fracturing for the in situ remediation of the unsaturated zone of a fractured soil polluted by jet fuel

      2011, Journal of Environmental Management
      Citation Excerpt :

      In general, when steam is injected in a porous medium, three zones are developed: the steam zone, the variable temperature zone and the ambient temperature zone (Davis, 1998; Class and Helmig, 2002; Class et al., 2002). The main NAPL recovery mechanisms with steam injection are (Davis, 1998): (i) the physical displacement by the hot water formed from the condensation of steam, and by the steam itself; (ii) the decrease of the interfacial tension and capillary pressure at elevated temperatures that reduces the residual NAPL saturation; (iii)the decrease of NAPL viscosity, which increases its mobility and the efficiency of its displacement by condensed water (She and Sleep, 1998; Gudbjerg et al., 2004a,b); (iv) the volatilization and steam distillation (Tzovolou et al., 2010); (v) the transport of contaminant vapor in the steam phase (steam stripping) to a condensation front where accumulated light hydrocarbons assist in the reduction of interfacial tension, capillary pressure, and viscosity with similar effect as noted in mechanisms (ii) and (iii). Enhanced desorption of contaminants from the solid surface at elevated temperatures may also occur (Braass et al., 2003).

    View all citing articles on Scopus
    View full text