Observations of grain boundary phase transformations in an elemental metal

While the theory of grain boundary (GB) structure has a long history(1) and the proposition that grain boundaries can undergo phase transformations was made already 50 years ago(2,3), the search for these GB transitions has been in vain. The underlying assumption was that multiple stable and metastable states exist for different grain boundary orientations(4,5,6). The terminology complexion was recently proposed to distinguish between those interfacial states that differ in any equilibrium thermodynamic property(7). Different types of complexions and transitions between them have been characterized mostly in binary or multicomponent systems (8,9,10,11,12,13,14,15,16,17,18,19). Novel simulation schemes have provided a new perspective on the phase behavior of interfaces and showed that grain boundary transitions can occur in a multitude of material systems(20,21,22,23,24). However, their direct experimental observation and transformation kinetics in an elemental metal remained elusive. Here, we show atomic scale grain boundary phase coexistence and transformations at symmetric and asymmetric [11-1] tilt grain boundaries in elemental copper (Cu). We found the coexistence of two different grain boundary structures at Σ19b grain boundaries by atomic resolution imaging. Evolutionary grain boundary structure search and clustering analysis(21,25,26) finds the same structures and predicts the properties of these GB phases. Furthermore, finite temperature molecular dynamics simulations explore their coexistence and transformation kinetics. Our results demonstrate how grain boundary phases can be kinetically trapped enabling atomic scale room temperature observations. Our work paves the way for atomic scale in situ studies of grain boundary phase transformations in metallic grain boundaries. In the past, only indirect measurements have indicated the existence of interfacial transitions(9,15,27,28,29). Now, their atomic scale role on the influence of abnormal grain growth, non-Arrhenius type diffusion or liquid metal embrittlement can be explored.