On the effect of slip transfer at grain boundaries on the strength of FCC polycrystals

TL;DR

This study uses computational modeling to analyze how slip transfer at grain boundaries influences the flow strength and damage nucleation in FCC polycrystals, showing that slip transfer reduces strength and affects damage sites.

Contribution

It introduces a physically-based crystal plasticity model that incorporates slip transfer criteria at grain boundaries, improving predictions of strength and damage in FCC polycrystals.

Findings

Slip transfer reduces flow stress compared to opaque boundaries.

Opaque boundaries have higher dislocation densities and stresses, favoring damage.

Including slip transfer improves Hall-Petch predictions for small grains.

Abstract

The effect of slip transfer on the flow strength of various FCC polycrystals was analyzed by means of computational homogenization of a representative volume element of the microstructure. The crystal behavior was governed by a physically-based crystal plasticity model in the framework of finite strains where slip transfer at grain boundaries was allowed between slip systems suitably oriented according to geometrical criteria. Conversely, slip transfer was blocked if the conditions for slip transfer were not fulfilled, leading to the formation of dislocation pile-ups. All the model parameters for each material were identified from either dislocation dynamics simulations or experimental data from the literature. Slip transfer led to a reduction in the flow stress of the polycrystals (as compared with the simulations with opaque grain boundaries) which was dependent on the fraction of translucent and transparent grain boundaries in the microstructure. Moreover, dislocation densities and Von Mises stresses were much higher around opaque grain boundaries, which become suitable places for damage nucleation. Finally, predictions of the Hall-Petch effect in Al, Ni, Cu and Ag polycrystals including slip transfer were in better agreement with the literature results, as compared with predictions assuming that all grain boundaries are opaque, particularly for small grain sizes ($<$ 20 $\mu$m).