The effective life of PRB depends on a complex interplay of degradation of the reactive media, either by exhaustion or coating, changes to hydraulic characteristics due to clogging, biofouling, channel formation or the production of gases. The coating of reactive surfaces with precipitates is a limiting factor on the long-term performance of some PRB (Kamolpornwijit et al. 2003). For example, the ability of granular zero valent iron to reductively degrade trichloroethene is compromised in the presence of permanganate, which induces insoluble precipitates and oxide coatings (Okwi et al. 2005). Clogging of pores by precipitates, with the reduction of hydraulic conductivity in barrier media, was noted by Mackenzie et al. (1999), Kamolpornwijit et al. (2003) and Simon and Biermann (2005), but not by Okwi et al. (2005). Gavaskar (1999) reported no significant degradation of performance following 5 years' operation of a zero valent iron barrier in Canada. Wilkin et al. (2002) and Lai et al. (2006) noted a loss of porosity but not of hydraulic performance, indicating that the onset of clogging is specific to the size of the granular materials and the chemistry of the aquifer being treated, and as a consequence feasibility tests of the long-term performance of barrier systems should be conducted using site groundwater supplemented by numerical modelling (Blowes et al. 2000; Jeen et al. 2007). The types of minerals precipitating can show zonation within the barrier; Li et al. (2006) found that carbonates dominated on the upstream side of a zero valent iron barrier, and that ferrous hydroxide dominated on the downstream side. These spatial patterns will control the pattern of clogging and ultimately the nature and location of barrier failure.
The accumulation of gas in barrier media as a consequence of microbial activity or carbonate dissolution can reduce hydraulic conductivity over time (Oberdorfer and Peterson 1985; Soares et al. 1991; Schipper et al. 2004; Williams et al. 2007), although there is a possibility of hydraulic recovery as bubbles migrate (Fryar and Schwartz 1998). Barriers that support naturally occurring microbial activity, or in the case of petroleum hydrocarbons, are supplied with nutrients in order to encourage microbial growth and activity, can be prone to biofouling or bioclogging. The growth of biomass can lead to changes in the hydraulic conductivity of the media, causing reduced residence time in the contaminated aquifer as well as parts of the barrier (Scherer et al. 2000; Thullner et al. 2004; Seki et al. 2006). Because bioclogging might only occur in parts of the barrier, the average hydraulic performance may not be affected, and as a consequence the reduction in treatment efficiency may not be detected (Seki et al. 2006). The period to the onset of biofouling will depend on the particular site water quality, particularly in terms of dissolved nutrients, contaminants and type of barrier media. A zero valent iron barrier exhibited no sign of biofouling following 6 months operation with a trichlororethene plume (Vogan et al. 1999), although few long-term studies exist to determine over what time frame biofoul-ing is likely (Kalin 2004). Roehl et al. (2005b) contains a range of case studies of the long-term performance of PRB.
Permeable reactive barriers are a relatively young technology, and long-term studies of their behaviour are few. As a consequence, modelling can offer valuable insights into barrier performance over longer times. Model simulations of barriers with different hydraulic conductivity and thickness indicate that flow will be greatest (and thus residence time the least) around the edges of the gate of a barrier, and flow will be least in the centre of the gate (Benner et al. 2001; Painter 2004). Increasing barrier thickness exacerbates this effect by enhancing flow convergence. Increasing the hydraulic conductivity of the barrier increases con-vegent flow and water flux, but does not change the edge flow enhancement (Benner et al. 2001). Heterogeneities in aquifer hydraulic conductivity are transmitted through homogeneous barrier material, but this effect is moderated with increasing barrier thickness. Further model simulations indicate that a few localized high hydraulic conductivity layers are more effective at introducing heterogeneous flow in the barrier than smaller, better-distributed high hydraulic conductivity layers (Benner et al. 2001). These simulations have implications for the performance of barriers operated in areas of freezing ground, particularly where ice lenses develop within the barrier or the aquifer upstream or downstream. Spatial heterogeneity in hydraulic conductivity within both the barrier and aquifer control the pattern of water flow (Gupta and Fox 1999; Benner et al. 2001). The hydraulics of the aquifer control the flux of water and contaminants within thinner barriers, whereas thicker barriers are more sensitive to inhomo-geneities within the barrier itself. In either case, though, contaminant breakthrough will most likely occur at localized zones of high-velocity flow, which includes the edges of homogeneous barriers. Elongating the funnel in the down-flow direction (termed "velocity equalization walls"; Christodoulatos et al. 1996) reduces this enhanced edge flow. An alternative modification to the velocity equalization walls is non-uniform barrier thickness in the down-flow direction, which places thicker barrier media in areas of enhanced throughflow, in order to achieve the desired residence time throughout the barrier (Painter 2004). A further design enhancement is the downwards extension of the funnel wall, which reduces vertical capture of groundwater flow (Painter 2004).
Was this article helpful?