The feasibility of installing biocidal beads into a catch basin filter insert for simultaneous stormwater disinfection and contaminant removal was tested. The catch basin filter insert (DrainPac®) was tested for its sediment, oil and grease, and coliform bacteria removal efficiency in conjunction with bench-scale testing of biocidal polymer beads. DrainPac® catch basin filters are composed of a metal frame, polypropylene filter fabric, and a high-density polymer support basket, and are installed below storm drains. A 12 x 41 in. DrainPac® filter (United Stormwater, Inc.) insert was set in a flume that simulated a large-scale catchment basin. Pond water was gravity fed to the flume at flow rates up to 200 gpm. The pond water contained fine sediments at concentrations ranging from 30-50 mg/L. The biocidal beads were tested in a small laboratory column for potential application to stormwater treatment. The head loss through the clean filter insert varied from 0.5 cm at 20 gpm to 9.1 cm at 200 gpm. Head loss of 21.5 cm occurred after 625 g of solids were added to the filter at 200 gpm at which point water began bypassing the filter fabric and flowing through the mesh screen. The highest flow rate that could be filtered through the loaded filter was 80 gpm. The DrainPac® filter removed total suspended solids with efficiencies ranging from 83% to 91% at flow rates of 20 to 200 gpm, with higher removal efficiencies at lower flow rates. The oil and grease removal efficiency of the DrainPac® filter ranged from 40% to 82%. The DrainPac® filter exhibited no removal of coliform bacteria under these test conditions. Biocidal brominated polystyrene beads, developed by Dr. S. D. Worley at Auburn University, were tested in a 1-cm diameter laboratory column apparatus for leaching of bromine compounds, which is important for determining if the filters will meet water quality regulations of the receiving waters. Removal efficiencies of coliform bacteria were tested in a companion study by Cal Poly graduate student Alex Bowerman. Pond water was passed through a 1-cm thick bed of beads in an up-flow direction. Leachate was analyzed for bromine, bromide, and bromoform. Leaching from both 0.3-mm and 0.8-mm biocidal beads was tested in the column apparatus. Samples collected for bromoform analysis were quenched with sodium thiosulfate, and the time before samples were quenched had a drastic effect on bromoform concentrations. Samples quenched after the collection of the total sample were assumed to be the most realistic for stormwater conditions since stormwater entering catch basins isn’t immediately discharged into its receiving waters. Effluent from a 1-cm bed of 0.3-mm beads at 0.56 mL/sec contained the following average concentrations: 0.47 mg/L bromine, 2.45 mg/L bromoform, and 0.53 mg/L bromide. The same conditions for 0.8 mm beads resulted in effluent containing the following average concentrations: 0.27 mg/L bromine, 0.06 mg/L bromoform, and 0.50 mg/L bromide. The much lower concentrations of bromine measured in the 0.8-mm bead leachate was likely due to the greater surface area of the smaller beads. The greater leaching of bromoform from the 0.3 mm beads may be partly attributed to the greater surface area of the smaller beads, but also the manufacturer of the beads reported that different production methods were used for the 0.3 mm beads. Higher bromine compound concentrations were found at lower flow rates, due to the longer contact times with the beads. Deionized (DI) water that was used to initially wet the biocidal beads before passing pond water through the column was also analyzed for bromoform. The DI flush water from the 0.3 mm beads contained 1.25 mg/L bromoform at a flow rate of 0.28 mL/sec. Bromoform concentrations in the DI flush water from the 0.8 mm beads at the same flow rate were below detection. Bromoform is expected to be formed as bromine is reduced to bromide by the presence of organics. Since there are no organics present in DI water to form such high concentrations of bromoform as those found from the 0.3-mm bead DI flush water, it appears that bromoform could have leached from the surface of the 0.3 mm beads. The longevity of the bromine treatment of the biocidal beads was tested on one set of 0.3-mm beads by simulating five use cycles, and also by testing another set of beads after dry storage. Each use cycle was simulated by pumping pond water through a 1-cm bed of 0.3 mm beads at 0.56 mL/sec for one hour and then connecting the column to an air pump to dry for 23 hours. After five simulated uses, leachate from the beads showed only a slight reduction in concentrations of bromine and bromide (21% and 4% less than fresh beads, respectively), while the concentration of bromoform was nearly 100 times less. The drastic decrease in bromoform concentration suggests that after five simulated uses, much of the bromoform was exhausted or conditions for the formation of bromoform were no longer present. For the dry storage test, 0.3 mm beads were wetted with 1-L DI water and then stored dry for 162 days. Then pond water was pumped through the column at 0.56 mL/sec and the leachate was analyzed. After dry storage for 162 days, the leachate showed no reduction in bromine concentrations compared to fresh beads, a 97% reduction in bromoform, and a 30% increase in bromide concentrations. This significant reduction in bromoform could be due to the volatilization of bromoform off the surface of the 0.3 mm beads during dry storage. The 0.3 mm beads are no longer being manufactured, and leachate from the 0.8 mm beads contained bromoform at concentrations below the potable drinking water maximum contaminant level of 80 µg/L. Under all tested conditions, bromine, bromide, and bromoform are present in the leachate from the biocidal beads, and thus their applicability for stormwater disinfection depends on the longevity of the bromine compounds in receiving waters, and on the regulations governing these compounds.

Urban stormwater is today recognized as a significant source of pollution that has contributed to the deterioration of water quality in lakes and streams. Of the pollutants, metals are commonly occurring in stormwater and can cause major damage when released into the receiving waters. A promising treatment method for removing these metals before it reaches the receiving waters is to filter the water in various filter technologies such as catch basin inserts. In this study, the potential of five materials to remove dissolved copper (Cu) and zinc (Zn) from stormwater was investigated. The study was initiated with batch equilibrium tests to assess the sorption capacity of the materials at different metal concentrations relevant for stormwater. This was done in both single and binary batches to investigate whether or not any competitive sorption occurred between the metals. Langmuir, Freundlich and Sips isotherm models were adapted to the obtained data. Through geochemical modelling, it was also possible to investigate whether precipitation of metals was likely. Finally, kinetic studies were conducted to investigate whether the reaction rate of Cu and Zn was relevant in the context of catch basin inserts. The results showed that the biochar had the highest sorption capacity followed by peat, bark, milkweed, and polypropylene. It could also be noted that Cu competed with Zn binding sites at high concentrations and that the sorption capacity of biochar, peat and bark, at concentrations between 50-100 μg/l Cu and 50-500 μg/l Zn, was sufficient to meet the limits set for stormwater emissions. For these sorbents, the reaction rate was also significant. Kinetics tests showed that at an initial concentration of 1,000 μg/l, 83% of the metals were sorbed after 5 minutes and after 10 minutes this number was 93%. Bark and peat also showed low effluent pH and leaching of dissolved organic carbon (DOC). The Langmuir and pseudo-second order equation could be well adapted to the data while geochemical modelling showed that precipitation of metals was unlikely. This suggests that chemical adsorption may be the mechanism that largely accounted for the removal of Cu and Zn. The results of this study can hardly be used to estimate the field performance of stormwater filters, but can be used as a basis for comparing and selecting sorbents for subsequent column tests. / I denna studie undersöktes potentialen för fem material att avskilja löst koppar (Cu) och zink (Zn) från dagvatten. Studien inleddes med skaktest för att testa sorptionsförmågan hos materialen vid olika metallkoncentrationer relevant för dagvatten. Detta gjordes i både singel och binära test för att undersöka om någon konkurrerande sorption skedde mellan metallerna. Langmuirs, Freundlichs och Sips isotermmodeller anpassades även till resultatet. Genom geokemisk modellering var det även möjlig att undersöka om utfällning av metaller var sannolik. Till sist gjordes kinetiska studier för att undersöka om reaktionshastigheten för sorption av Cu och Zn till filtermaterialen var relevanta i ett brunnsfiltersammanhang. Resultaten visar att biokol hade den största sorptionsförmågan följt av torv, bark, sidenört och polypropylen. Det kunde även konstateras att koppar konkurrerade om sorptionsplatser för zink vid högre koncentrationer och att sorptionskapaciteten för biokol, torv och bark, vid koncentrationer mellan 50-100 µg/l Cu och 50-500 µg/l Zn, är tillräckligt hög för att uppfylla gränsvärdena för utsläpp av dagvatten. För dessa sorbenter var reaktionshastigheten betydande. Kinetiktesten visade att vid en initial koncentration på 1,000 µg/L var 83% av metallerna sorberade efter 5 minuter och efter 10 minuter var denna siffra uppe i 93%. Bark and torv uppvisade även ett lågt uppmätt pH och lakade löst organisk kol (DOC). Langmuir och pseudo-andra ordningens ekvation kunde anpassas väl till datat samtidigt som den geokemiska modelleringen visade att utfällning av metaller var osannolikt. Detta antyder att kemisk adsorption kan vara den mekanism som till störst del stod för avskiljningen av Cu och Zn. Studien kan inte fastställa något om filtermaterialens fältprestanda, men kan användas som underlag för att jämföra och att välja sorbenter för efterföljande kolumntester. / Formas project 2016-75