Abstract

Several recent studies that have proposed MUSIC models for gibbsite surfaces have purported to achieve a very good fit with potentiometric titration data. However, in order to achieve such results, several significant parameters, such as the number of surface sites, site densities, and pKa values were sometimes re-introduced in the model as fitted parameters, and physically unrealistic modeling assumptions were sometimes used. In addition, recent evidence supports the idea that some of the gibbsite potentiometric titration data from these studies may be unreliable. In order to re-interpret the potentiometric titration data, we used several recently published methods. In order to detect possible problems with estimates of gibbsite basal and edge surface area, we synthesized two gibbsite samples with different aspect ratios and characterized their surface areas using BET, AFM, and computerized image analysis routines. We also estimated pKa values for acid/base reactions at gibbsite surfaces by applying a new bond-valence method to gibbsite (001)-type and (100)-type surface structures based on ab initio calculations. The resulting pKa estimates are not to be taken as precise values due to difficulties and assumptions associated with calculating reasonable ab initio surface structures. Instead, we believe they represent a more reasonable range than has been previously estimated. Using these estimates, we propose an improved MUSIC model for gibbsite, which seems to predict surface adsorption, not perfectly, but within a reasonable range for a number of titration data sets without re-introducing any of our estimated parameters as adjustable parameters. Discrepancies that exist between model predictions and various potentiometric titration data sets are likely due to error associated with potentiometric titrations and pKa predictions.

Degree

College and Department

Physical and Mathematical Sciences; Geological Sciences

Rights

http://lib.byu.edu/about/copyright/