Modeling of radiation-induced segregation in austenitic Fe-Cr-Ni alloys.

Radiation-induced segregation (RIS) was studied in Fe-Cr-Ni alloys irradiated with protons to better understand the mechanisms causing changes in grain boundary chemistry and to improve the ability to predict RIS in austenitic Fe-Cr-Ni alloys. Ni-18Cr, Ni-18Cr-9Fe, Ni-18Cr-0.08P, and Fe-20Cr-9Fe were irradiated with 3.2MeV protons at temperatures from 200$\sp\circ$C to 500$\sp\circ$C and to doses from 0.1 to 3 dpa. Grain boundary chemistry was measured using both Auger electron spectroscopy (AES) and scanning transmission electron microscopy with energy dispersive x-ray spectroscopy (STEM/EDS). The significant driving mechanism far segregation in Fe-Cr-Ni alloys is shown to be the inverse Kirkendall (IK) mechanism, specifically the coupling between alloying elements and the vacancy flux. The inclusion of interstitial binding effects to RIS models results in poor agreement between model predictions and segregation measurements, severely overpredicting the measured Ni enrichment and Fe depletion. Grain boundary segregation is unique for each bulk alloy composition in that the amount and the rate of segregation differs for alloys irradiated under the same conditions. Kinetic parameters must be known for each alloy to accurately predict segregation, but the kinetic parameters in Fe-Cr-Ni alloys at low temperature are not well studied. Additionally, short range ordering interactions are important in determining the segregation in all Fe-Cr-Ni alloys. Ordering enthalpies must be included in RIS models to correctly describe the segregation process. Therefore, to develop a predictive RIS model, a method for calculating diffusivities from the bulk composition that includes ordering enthalpies was developed. The Perks (IK) model has been modified to account for composition dependent segregation kinetics by calculating the migration energy using pair interaction potentials, ordering enthalpies, and the local concentration. Based on segregation measurements from seven different alloys, irradiated to five different temperatures and six different doses, this model is in better agreement with the segregation measurements than the original Perks model. Using a single set of input parameters, the modified IK model predicts the measured segregation of Fe, Cr, and Ni with an error less than about 5 at%.Ph.D.Applied SciencesMaterials scienceNuclear engineeringUniversity of Michigan, Horace H. Rackham School of Graduate Studieshttp://deepblue.lib.umich.edu/bitstream/2027.42/130191/2/9721937.pd