Surfactants can solubilise super-hydrophobic organic contaminants (SHOCs), thereby facilitating their transport through soil in the aqueous phase. This may result in unexpected groundwater contamination and associated offsite transfer, posing risks for environmental and human exposure to toxic compounds such as polychlorinated dibenzo-p-dioxins and dibenzofurans (PCDD/Fs). Current knowledge on surfactant facilitated transport (SFT) stems mainly from laboratory studies focused on surfactant enhanced aquifer remediation of moderately hydrophobic organic compounds in saturated soil. The present study investigated the transport potential and congener-specific distribution of PCDD/Fs in agricultural and industrial field soils where surfactants were released at relatively high levels or high frequencies, and derived modelling approaches to predict SFT in the vadose zone environment.
To assess PCDD/F transport potential under high surfactant loads, deep soil cores were collected at a pesticide factory where contaminated wastewater containing over 1% surfactants was stored in bunded ponds. PCDD/Fs were present in soils beneath the ponds to depths of 5.7 metres, with congener profiles similar to those in overlying wastewater. The maximum ΣPCDD/F concentrations were observed at 2-2.5 metres depth (68-130 ng/g), far below their expected mobility range based on physico-chemical properties. Simulations of water flow and estimated surfactant loads in the wastewater suggested SFT as the dominant transport mechanism resulting in rapid vertical migration of PCDD/Fs (2.4 metres in <4 months). Subsequent lateral transport of higher-chlorinated PCDD/Fs was indicated by the presence of these compounds throughout deep soils collected in the direction of groundwater flow. Congener-specific analysis further indicated that PCDD/F mobility under SFT was reversed compared to that predicted from physico-chemical properties, with the least water soluble congener migrating to the greatest extent.
To investigate PCDD/F mobility under longer-term, high frequency surfactant release, additional cores were obtained from a sugarcane field and a suspected chemical/pesticide waste disposal site. Whilst upper soil layers contained the highest PCDD/F concentrations (25-96 ng/g dw) which generally decreased with depth, concentrations remained elevated at ppb levels over several metres. Identical isomer profiles in the surface soils to those at the pesticide factory site indicated pesticide impurities as the most likely source of PCDD/Fs in these soils. Congener-specific analysis revealed a progressive shift towards PCDD/Fs fully chlorinated in the 1,4,6,9-positions, consistent with reports of preferred lateral dechlorination. A model developed to simulate changes in PCDD congener profile distributions, with parameters fitted to predict observed field results, provided a novel means of estimating previously unknown field parameters that can be used as inputs for mechanistic simulations. The outcomes suggested the release of a continuous, but decreasing OCDD-dominated source over ~60 years, dechlorination half-lives of 8-40 years for tetra- to octa-CDDs, respectively, and facilitated migration rates in the order of metres per decade.
Field data from the pesticide factory site (where surfactant release was quantified) were used to evaluate modelling approaches to predict, for the first time, SFT of SHOCs in unsaturated field soils. An assessment of SFT processes suggested that the key mechanisms influencing SHOC migration in the vadose zone are advective transport and the sorption behaviour of surfactants and SHOCs under dynamic flow. In some instances, mixture effects and degradation of SHOCs and surfactants, as well as SHOC transport with surfactant monomers, may be important, and often unrecognised, field processes. Few models were identified that could simulate these processes under vadose zone conditions, and it was concluded that existing modelling approaches have limited application to predict SFT under long term, frequent surfactant release, such as at agricultural sites. Existing modelling approaches could, however, be adapted to simulate SFT at high volume surfactant release sites, such as at the pesticide factory, and methods to represent transient flow phases using an existing steady state SFT model were derived. Based on a detailed review, SMART (Streamtube Model for Advective and Reactive Transport) was selected as the most appropriate steady state model for this study. Additional SMART functionality requirements to simulate vadose zone processes and current data gaps for empirical solute transport parameters were identified; these were addressed where possible by defining alternative modelling approaches and predictive models.
The SMART model, parameterised to conditions at the study site, could predict SFT of PCDDs in the vadose zone to a depth of several metres, with most movement of PCDDs occurring during the initial ponded-infiltration phase. Simulated migration depths were, however, lower than those observed in the field. Sensitivity analyses indicated that the migration potential of SHOCs in field soils is highly dependent on the input parameters for SHOC and surfactant partitioning to soil; in particular, model predictions may be improved by incorporating non-equilibrium sorption of these compounds into vadose zone modelling approaches. The different migration potentials of TCDD and OCDD observed in the field, i.e. the mobility reversal, could not be simulated when simple equilibrium PCDD sorption processes were assumed, parameterised using predictive models (based on linear property-property relationships with KOW). Hypotheses to account for the reversal of mobility were proposed; these related to differential sorption affinities and kinetics between PCDD congeners for stationary and mobile subsurface phases, and competitive solubilisation of PCDDs to micelles. These results highlight the need for empirical data for surfactant and SHOC sorption kinetics and competitive partitioning processes under transient flow conditions, and incorporation of these, where necessary, into vadose zone modelling approaches.
This study demonstrated that relatively rapid and extensive migration of otherwise immobile contaminants can occur at field sites with a history of surfactant release. The ability of surfactants to fundamentally change chemical behaviour, so that the most immobile compounds have the greatest potential to be readily transported with water, suggests that SFT in the vadose zone may need to be considered in areas where surfactants and SHOCs are released at high levels and/or frequencies, to facilitate protection of groundwater via appropriate chemical management.