Granular iron reacts with organic compounds dissolved in water through a surface redox reaction that corrodes the metal. In the early 1990s, it was suggested by Dr. R.W. Gillham at the University of Waterloo that this material could be placed in the ground to intercept plumes of contaminated groundwater and remove the pollutants by chemical transformations. This idea led to the concept of permeable reactive barriers (PRBs) for groundwater remediation. Barriers are now constructed with a variety of materials to handle numerous contamination problems. However, granular iron remains one of the most versatile and effective barrier materials available. To the right is an image of a sample of granular iron subjected to a solution of several hundred micromolar of 4-chloronitrobenzene. The black color is characteristic of magnetite that develops on the surface. The orange color is indicative of iron (III) oxides and hydroxides (rust). In our research we examine the kinetics of the surface reactions of nitroaromatic compounds and chlorinated aliphatic compounds (solvents) and evaluate their dependencies on such factors as reactant concentration, product concentration, competing substance concentration and ionic composition of the water.

The research to date has demonstrated the saturation effects of high concentrations of reactants, the reaction enhancing effects of low concentrations of sulphate and chloride (relative to perchlorate) on the metal, and the inhibiting effects of nitrate and carbonate (at 8 millimolar each in separate tests). We have also gathered evidence of multiple kinds of reactive "sites" on the solid surface, and associated sorption behavior with these sites.

Our results have implications for reactive barrier perfomance and design, and some of the methods we have developed could be advantageous in treatability studies.

It has been well documented that granular iron will clog if exposed to water with high dissolved oxygen content, or high carbonate content. It has also been shown that the precipitates causing the clogging reduce the reactivity of the medium. The image to the right shows a granular iron column clogging with rust following exposure to aerated water. However, virtually all of the previous laboratory studies were conducted under controlled conditions that were hydraulically unrepresentative of a field barrier. There is some indication from field studies that iron-based reactive barriers have long lives inspite of the shortcomings evident in the lab studies. Some of this disparity may be due to large differences in the flushing rates of the columns compared to the field, but some may also be attributable to differences in the flow systems. We propose to examine the relationship between reactivity and flow in more detail with particular attention on duplicating the conditions that would exist in actual reactive barriers as closely as possible, while maintaining the control of a laboratory column experiment for the imaging of pore-scale changes.

Permeable reactive barriers (PRB) have been in use for over a decade now and much has been learned about their strengths and weaknesses. Most iron-based PRBs have enjoyed considerable success once they were installed, and the site owners now benefit from low operation and maintenance costs for their remediation systems. The relatively few PRBs that have run into difficulties have generally suffered from problems at the site characterization stage rather than poor performance of the PRBs themselves. This work set out to address the factors that might affect PRB performance in cases where site characterization was adequately done. In particular, it has been shown that groundwater dominated by different anionic species imposed either an enhanced reactivity or a depressed reactivity on commercial granular iron, relative to solutions of perchlorate (a surface inactive substance) at identical ionic strengths. However, the literature does not address the effects of anion mixtures on iron reactivity, so this was investigated here.

Ian Bowen, an undergraduate chemist shown presenting his work at a meeting of the AEG/AIH in 2007, performed a series of experiments systematically examining solutions with different proportions of carbonate and sulfate, maintaining a constant overall ionic strength. His work showed that while carbonate alone causes a decrease in iron reactivity, and sulfate alone causes an enhancement, under the conditions of these experiments, mixtures of carbonate and sulfate can lead to a super-enhancement in reactivity. This work may explain discrepencies in the literature where some studies have reported carbonate induced enhancements to reactivity and others have reported the reverse.

The community of Baden, Ontario, once drew its drinking water from a well field located a short distance out of town. In the mid 1990s, the concentration of nitrate in the well water rose above the the recommended drinking water limit, leading the town to seek new and more costly sources of water. Since above ground treatment of nitrate in water is also very expensive, this project was aimed at treating the groundwater in situ by stimulating denitrification in the aquifer. The image to the right is that of a section of core collected between 120 and 122 ft depth (36 to 37 m), corresponding to a preferred horizon for flow and nitrate transport in the Baden aquifer. This and other similar zones were targeted for treatment by injecting acetate (a source of carbon) in weekly pulses across the ambient flow field, and allowing the acetate to be consumed as it migrated toward the municipal well between injections. The injection method was an adaptation of previous research published with Barker in 1994 and 1996 in the journals Ground Water and Water Resources Research (see Publications).

The aquifer was characterized by coring, aquifer testing and geochemical sampling. Hydraulic responses at multilevel points were found to be excellent predictors of solute transport pathways. Subsequent acetate injections were found to stimulate the the desired denitrification, indicating that this methodology may be useful for extending the useful life of this and other water supplies affected by non-point source nitrate. The first paper from this work, a comparison of electron acceptors for stimulating denitrification, was published in the Journal of Contaminant Hydrology (JCH) in 2000 (Devlin, Eedy and Butler). The second paper, which deals with assessing heterogeneity at the the study site(Gierczak, Devlin and Rudolph) was published by JCH in 2006. The third paper dealt with the results of a pilot denitrification system, and was published by JCH in 2007(Gierczak, Devlin and Rudolph).

The first priority in assessing contaminant transport at a contaminated site is to determine the direction and speed the pollutants are moving. This, of course, depends on the direction and speed the groundwater is moving. In this project, a novel probe was designed (PVP), modelled and field tested for the direct measurements of point scale groundwater velocities. Pictured at the right is Mike McGlashan, a former graduate student at the University of Kansas, standing beside a multilevel PVP. Five of these devices were installed across a petroleum plume in the Borden aquifer and monitored for changing velocities in response to geochemical and microbiological changes in a hydrocarbon plume undergoing aerobic bioremediation. The transect was also monitored using borehole ground penetrating radar, and the two data sets correlated very well, as shown in the images to the right. In this case, zones of enhanced microbial activity corresponded to zones of low groundwater velocities and low radar wave velocities (cool colors). Work continues to evaluate the underlying causes of changing PVP and GPR signals over time in the aquifer, but the results have twice been duplicated in the laboratory, in projects undertaken by graduate student Peter Schillig and undergraduate Elisheva Patterson. Faculty collaborators on this project include Dr. G. Tsoflias, geophysicist, and Dr. J. Roberts, geomicrobiologist, both from the Dept. of Geology at the University of Kansas.

The PVP instrument was the subject of a 2004 Ph.D. thesis at the University of Waterloo, Ontario, Canada, by Walid Labaky under the supervision of J.F. Devlin and R.W. Gillham. That work is in press with Environmental Science and Technology.

The floodplain of the Kansas River has been studied for its geology, soils, quaternary geology and hydrogeology. Recently, a small part of the floodplain (the GEMS site) has been used in the development of slug test methods for detailed characterization of aquifer heterogeneity. However, most of the regional hydrogeological work was completed several decades ago when the primary research issues related to determining the magnitude of of the resource.

In this project, we propose to update our knowledge of the floodplain hydrogeology using state-of-the-art investigative tools. In addition, we propose to use current hydrogeological knowledge to ask questions that were not considered in the previous studies, including questions related to contaminant transport, geologic heterogeneity and groundwater - surface water interactions. Pictured above is a view of the northern half of the Kansas River Floodplain north of Lawrence, Kansas. The water in the foreground is a flooded meander scar.

GPR response to microbial growth in porous media. Undergraduate Elisheva Patterson, and graduate student Peter Schillig, shown at right, were awarded the top two places for their research presentations in the G-Hawker Symposium, held at the University of Kansas in October, 2007. The presentations dealt with complimentary studies in the field (Schillig) and laboratory (Patterson) demonstrating a relationship between ground penetrating radar (GPR) wave velocities and microbial activity in biostimulated sand. This research has application in the bioremediation industry and in the monitoring of natural attenuation of pollutant plumes.

Effect of microbial activity on flow. With the rise of permeable reactive barriers for groundwater remediation, questions need to be asked concerning how long they last and what controls their performance. It is frequently noted that rates of pollutant transformation in laboratory tests do not match rates observed in the field. In this work it is hypothesized that residence time in the reactive zone of an aquifer undergoing bioremediation is affected by the microbial activity there. Peter Schillig performed his Masters research on this topic for which he was awarded best student paper at the AGU Joint Assembly, Acapulco, Mexico, in May, 2007. Peter was also awarded the prestigious Self Fellowship to continue his research at the doctoral level at KU. At right Peter poses beside the grill of a Dept. SUV containing feathers from a turkey vulture he bagged in flight.

Sorption vs. reaction on granular iron. Observed reaction rates of organics with granular iron depends on both the quantity sorbed and the surface reaction rate. Similar rates might be observed for the two distinct cases in which either sorption is large but surface reactions are low, or sorption is low but surface reaction rates are high. Unfortunately, uniquely calculating the sorption and reaction parameters necessary to model such scenarios is difficult for reacting species. The literature reports only one extraction based method to accomplish this separation of processes. However, a new method, involving combined column and batch tests, was tested by Melissa Marietta with promising results (undergraduate thesis work) for three nitroaromatic compounds. The work won her the AEG student paper award, 2005. At right Melissa presents her work to theAEG membership after winning the award.

In situ sequential treatment. Over the past few years, field trials were carried out to test the viability of sequencing in situ technologies for groundwater remediation. The first such test was conducted at the Alameda Naval Air Station (pictured at right), where granular iron and a biosparge unit were coupled to remove a mixture of chlorinated solvents and petroleum hydrocarbons from the groundwater The test facility performed very well, removing over 99% of the cis-dichloroethene that was initially present in the water (> 200 mg/L in some locations) and all detectable hydrocarbons (max. BTEX = 10 mg/L). This work was published in 2000 in the Journal of Contaminant Hydrology with Morkin, Barker and Butler. See Publications for the complete citation.

Sequenced in situ treatment was also tested in a controlled field experiment at the CFB Borden site. In that test, two biodegradation systems were constructed in series to treat a plume containing carbon tetrachloride, chloroform, tetrachloroethene and toluene. The first treatment was anaerobic biodegradation stimulated by benzoate and nutrient additions made from a nutrient injection wall (NIW). The second treatment was aerobic biodegradation from a residual oxygen barrier (ROB). The two treatment units were constructed in a 24 m long sheet pile alleyway. Pictured at right is the enclosure built to protect the experiment. Removal rates exceeding 99% were experienced for all primary compounds except toluene. Also, low but detectable amounts of cis-dichloroethene were close to breakthrough a the end of the experiment. Both these problems could be alleviated with diminished benzoate additions at the NIW, thus reducing oxygen demand at the ROB. This work was coauthored with Katic and Barker, and was published in the Journal of Contaminant Hydrology in 2004.