Andre, CM; (2006) Adsorption treatment for the removal of humic substances from drinking water supply, using granular activated carbon and iron-containing adsorbents. Doctoral thesis , UCL (University College London).

Preview

Text Andre.C.M.Thesis.Redacted.pdf Download (24MB) | Preview

Abstract

This thesis investigates the removal by adsorption of potentially harmful natural organic matter, humic substances (HS), from the water supply using three different media granular activated carbon (GAC), iron-coated alumina (AAFS) and ferric oxihydroxide (akaganeite P-FeOOH). The adsorbents were characterised using Fourier transform infrared, scanning electron microscopy, pH of zero point of charge and surface area analysis. While GAC is microporous, the two iron-containing adsorbents are mesoporous with low microporosity. Further work is required to investigate media attrition. The HS solutions were fractionated into three ranges of molecular weights (from 0 to 50 kDa) and were characterised by UV absorbance at 254 nm, DOC and the SUVA (ratio UV254/DOC). Results from the kinetics study indicated that different mechanisms control the adsorption rate at defined stages throughout the process. Five kinetic adsorption models were assessed to describe the adsorbent-adsorbate systems behaviour pseudo-first and second order models, the Elovich equation, intraparticle diffusion and external diffusion models. Adsorption in batch systems showed that low molecular weight HS preferentially adsorbed onto GAC whereas heavy molecular weight HS would adsorb better onto AAFS and p-FeOOH. Adsorption capacities were estimated using the Freundlich isotherm. Large residual DOCs were required before adsorption would commence, caused by a non-adsorbable fraction with very low molecular weights. This also led to a minimum residual DOC concentration in the column effluent during continuous flow experiments. Furthermore, significant HS removal was only observed in the columns once a concentration gradient had built up. The Thomas equation accurately predicted the adsorption and the bed depth service time model indicated that contact time is a more important criterion than bed depth. Finally, simulations of adsorption in the columns were carried out using the computer model GACMan. These were only successful over a narrow range of molecular weights.

Download activity - last month

Export as JSONXMLCSV

Download activity - last 12 months

Export as JSONXMLCSV

Downloads by country - last 12 months

Export as CSVXMLJSON

Archive Staff Only

View Item