On Agreement of Experimental Data and Calculated Results in Grain Boundary Segregation

Abstract

There are two sources of quantitative data on grain boundary segregation: careful experimental results and calculated data. These values can be compared in various ways. Here we show a comparison of average concentrations of silicon, vanadium, and tin at the grain boundaries of bcc iron determined in three ways: (i) on the basis of calculations of the segregation energy for individual sites; (ii) experimentally; and (iii) using a phenomenological prediction for selected systems characterized by satisfactory solid solubility of the segregant in bulk. We found very good agreement between the results of these approaches for all three of the segregants. The results clearly show the indispensable role of so-called ‘anti-segregation sites’ in the determination of average grain boundary concentration as well as the importance of segregation entropy and consequently, of entropy-dominated grain boundary segregation.

Keywords

• DFT calculations
• experimental data
• energy
• entropy
• grain boundaries
• machine learning