Investigation of the adsorption and biodegradation capacity of filter media towards airport de-icing chemicals

At airports in Norway, propylene glycol (PG) and potassium formate (PF) are used as de-icing chemicals to provide safe air traffic. Expected climate changes will lead to temperatures more frequently around zero degrees, when de-icing chemicals consumption is at its highest. In addition, air traffic is expected to increase. De-icing chemicals degradation has a finite capacity, mainly dependent on nutrients, oxygen and temperature. In this thesis, some filter media have been tested to evaluate adsorption and biodegradation capacity of de-icing chemicals. Filtralite media is today used for drinking- and wastewater treatment. Filtralite media is a promising media in a bioretention system, due to low head loss and high storage capacity. Batch experiments and column experiment have been conducted to investigate adsorption and biological degradation of the de-icing chemicals propylene glycol and potassium formate. Batch experiments were performed to investigate adsorption onto granular activated carbon (GAC), Filtralite NC 0-2 and Filtralite NC 0.8-1.6. Adsorption was shown onto GAC. Linear isotherm was found as best-fit for both chemicals (PG, R2 = 0.99 and PF, R2 = 0.94). PG and PF showed similar power indexes with Freundlich isotherm, close to 1 (1/n of 0.883 and 0.871 respectively), which indicate that adsorption onto GAC was not very high for any of the de-icing chemicals. Second order kinetics fitted best for both chemicals. No adsorption onto the Filtralite media was observed. Columns were filled with Filtralite NC 0.8-1.6 (Column 1) and Filtralite NC 1.5-2.5 (Column 2). Synthetic stormwater was prepared with concentration of 20 mg/l PG and 13.4 mg/l PF. Degradation in a 1 m3 feed tank was investigated prior to the column experiment, to find that 50% of the de-icing chemicals were degraded after 5 days. This resulted in changing the synthetic stormwater every day to prevent the risk of degradation. Degradation in the columns was shown to be significant in the upper ~15 cm for both columns, while degradation in the lower parts of the columns were not significant. Initial removal in the upper ~15 cm was 20% for column 1 and 15% for column 2. Increased phosphorous and nitrogen in the synthetic stormwater showed improved degradation (C:N:P=24:7:1), with the average of 50% removal through both the filters in the columns. Results from Iwasaki modelling showed increased filtration coefficients after nutrient addition. Clogging occurred in the top 15 cm of the filter. From these results it was found a lifetime of 2 years for the filters tested in the column experiment, which is relatively low, probably due to high sediment loading. In reality the sediment load might actually be less than in the column experiment, however sediment loading was necessary due to the need of bacteria for the biofilter form the sediments. A top layer of GAC in the biofilter was considered, however a biofilter with a GAC layer on top reduces the lifetime of the filter due to clogging, and is therefore not recommended due to low adsorption capacity. Filtralite media showed interesting results and recommended to be further investigated