Stochastic track-structure simulation methods and the determination of product yields in the radiolysis of water containing various solutes

An investigation was made into the use of Monte Carlo based models to simulate the radiolysis of liquid water solutions. Using computer codes developed at Oak Ridge National Laboratory, two areas of inquiry were explored: (1) a study of the basic premise used in the ORNL and other models that bimolecular reactions occur as a result of the proximity of the species alone, the reactants having undergone diffusion by random walk, and (2) an investigation into the agreement of the predicted model results with some new data compilations, which are considerably broader than those previously available. The study of the simulation approach to reaction modeling was made in a spherical "reaction chamber" in which a randomly distributed number of species was allowed to diffuse and react. The model generated values of the reaction rate coefficients at long times which were very near to the entered values, and the calculated values of the coefficients exhibited recognized time-dependent behavior. The calculated values were found to be dependent on the initial density of species within the chamber, the probability of reaction per encounter, and to a lesser extent, to the size of the chamber. In addition, the possibility that species may have "jumped through" each other during a diffusive step was found to be important for a few of the reactions. The investigation of the model agreement with extant data compilations employed four newly developed versions of the ORNL codes which treat the early chemistry within a radiation track. Testing of these codes against several new experimental data compilations revealed several areas in which some of the basic model assumptions needed adjustment. Several improvements to the computer model were made, including adjustments to fractional disappearance constants governing the apportionment of species during the prechemical stage, the addition of explicit dependence on the pH of the solution, and minor adjustments to the reaction rate coefficients for two reactions. With these changes, the model- generated results are in good agreement with a broad range of experimental data from a number of different aqueous systems containing different solutes. This study has resulted in a better understanding of these models to simulate track-structure events in the radiolysis of aqueous systems. The use of the reaction chamber to study in detail the behavior of the Monte Carlo codes within the basic assumptions inherent in the physical and chemical systems modeled, the development of computer codes to model four new chemical systems, and the adjustments to the prechemistry portions of the ORNL model represent new contributions to this ongoing area of inquiry.