A homologous series of n-alcohols was used as model contaminants to investigate the effect of hydrophobicity on the hydrogen peroxide concentration necessary in Fenton-like treatment for near-complete (>99%) destruction of compounds sorbed to soil. These probe compounds were selected because they exhibit equal reactivities with hydroxyl radicals, but have varied hydrophobicities. The standard Fenton reaction was first used to confirm equal hydroxyl radical reactivity for the n-alcohols. Central composite rotatable design experiments were then used to determine the conditions in an iron (III)-hydrogen peroxide system that resulted in 99% degradation of each of the probe compounds when sorbed to soil. The hydrogen peroxide concentrations required for 99% destruction of the sorbed compounds increased with probe compound hydrophobicity. Furthermore, hydrogen peroxide concentration requirements were directly proportional to the log octanol-water partition coefficients (log K-OW) of each probe compound. This quantitative relationship may not be directly applicable to other organic contaminants, but a strong correlation between log K-OW and hydrogen peroxide requirements for other contaminants will likely be found. These results confirm that hydrogen peroxide requirements for soil treatment increase as a function of contaminant hydrophobicity and provide a basis for the development of an algorithm for hydrogen peroxide requirements when modified Fenton's reagent is used for in situ chemical oxidation (ISCO). (C) 2003 Elsevier B.V. All rights reserved.