Phytoremediation of a highly creosote-contaminated soil by means of Salix viminalis

Sammanfattning: The aims of this study were to investigate whether Salix viminalis could grow in a highly creosote contaminated soil and, if so, whether the introduction of the plant enhanced the dissipation of polycyclic aromatic hydrocarbons (PAHs). The purpose was also to examine whether inoculation of two different bacterial strains further enhanced the decrease of the contaminant. The experiment was conducted in a greenhouse during ten weeks. The soil was collected at a former wood impregnation site belonging to the Swedish railway administration, in Krylbo, Sweden. The experiment comprised two different levels of PAH contaminated soil, the creosote soil used as such or diluted with an arable soil (1:1). The main PAHs studied in this experiment were phenanthrene, anthracene, fluoranthene, pyrene and benzo[a]pyrene. The soil samples were extracted with toluene and analysed by GC-MS. Microbial analyses were made to measure the number of extracted and cultivable bacteria (CFU/g soil) and dominating bacteria were tested for gram, fluorescence and oxidase reactions. The plants were infected by some bacterial disease which probably was an inherent infection of the cuttings and the number of "healthy" plants was reduced by 2/3 in both the creosote soil and the diluted soil. The initial PAH concentrations were very high, particularly for fluoranthene and pyrene. Some PAH degradation occurred in pots without plants, but the dissipation of all studied PAHs was enhanced in the presence of Salix viminalis and also the reduction of more recalcitrant compounds such as benzo[a]pyrene was impressive. In the presence of plants the PAH compounds were reduced in the creosote soil by 67 %, 79%, 77% and 43 % for anthracene, fluoranthene, pyrene and benzo[a]pyrene respectively compared to the initial values and by 61 %, 66%, 79% and 30% in the mixed soil. For all studied PAHs the concentrations in the rhizosphere soil were found to be greater than in the bulk soil and this could probably be explained mostly by the movement of compounds from the surrounding soil into the rhizosphere, but partly also by an increased solubility. The results indicate that the introduction of the inocula was not succesful, but due to the inadequate number of pots with "healthy" plants no conclusions could be drawn. The microbial analyses showed a significantly higher number of cultivable bacteria per gram of soil in treatments with plants both compared to the initial soil and the treatments without plants.