Methods for Scaling and Comparing Adsorption Datasets
Type of DegreeThesis
MetadataShow full item record
The presence of reactive groundwater contaminants are a great concern to many government agencies and private entities. The fate and transport of these contaminants is determined by their interactions with subsurface solids and hence a thorough understanding of these processes is necessary for the remediation of contaminated sites. There is an abundance of adsorption literature; however, these studies vary in their goals and methods. Some studies are general in nature, dealing with the general chemistry of adsorption in a well-controlled laboratory setting while others examine a specific problem at a contaminated site, using the soils and conditions at the site. This inherent difference, along with inadequate reporting of experimental conditions, makes it difficult to take information from one study and compare it confidently to others. Proper comparisons are an important part of studying these systems. Furthermore, developing predictive reactive-transport models requires the ability to scale the reactions to various systems. As models capable of predicting the fate and transport of groundwater contaminants are developed, the need to compare adsorption data in a reliable manner is important. In this thesis, various approaches for comparing adsorption data were examined to determine the best method for designing and analyzing the results of adsorption experiments. The interactions of arsenate and goethite were examined as a representative system although results can be extended to other systems. A study of model-generated data was first executed to examine the basic theoretical principles of dataset comparison. These principles were then tested on a suite of experimental data from our laboratory and from the literature. The results indicated that the most important factor in the comparison of sorption datasets is the total adsorbate-adsorbent ratio (e.g. mol As/mol Fe). If this value is too low, the system will be limited only by the amount of adsorbate present and accurate comparisons between datasets cannot be made. Additionally, it was found that normalization techniques can have a marked effect on comparisons, especially between systems with differing adsorbent types (i.e. natural and synthetic goethite). It is nearly always better to use the specific surface area of the adsorbent as a scaling parameter rather than the reactive mass of the adsorbent (in this case, Fe content). The incorporation of these techniques will improve the methods for scaling and comparing adsorption data in future studies.