Title: Assessment of deformation mechanisms in Mg alloys by means of advanced characterization techniques and machine learning
Speaker: Prof. Javier LLorca
Date/Time: 2023.07.11 09:00-10:00
Location: Yiucheng Lecture Hall (500), Xu Zuyao Building
Inviter: Prof. Xiaoqin Zeng
Biography
Prof. Javier LLorca is scientific director and founder of the IMDEA Materials Institute, where he leads the research group on Bio/Chemo/Mechanics of Materials, and professor and head of the research group on Advanced Structural Materials and Nanomaterials at the Polytechnic University of Madrid.
A Fulbright scholar, Prof. LLorca is Fellow of the European Mechanics Society and of the Materials Research Society, member of the Academia Europaea and has held visiting positions at Brown University, Indian Institute of Science, China Central South University and Shanghai Jiao Tong University, where he is currently Visiting Chair Professor. He has received -among many others- the Research Award from the Spanish Royal Academy of Sciences, the Distinguished Scientist Award of the Structural Materials Division of TMS, the Research Award from the Polytechnic University of Madrid and the Career Award from the Spanish Society of Materials (SOCIEMAT). His current research interests – within the framework of Integrated Computational Materials Engineering – are aimed at the design of advanced materials for engineering applications in transport, health care (implants) as well as energy (catalysis), so new materials can be designed, tested and optimized in silico before they are actually manufactured in the laboratory.
Abstract
Plastic deformation of Mg alloys takes place through different mechanisms, namely basal, prismatic and pyramidal slip as well as extension and compression twinning. The large differences in the critical resolved shear stress between basal and non-basal slip and the fact that twinning is a polar mechanism lead to complex stress states at the local level, even in samples subjected to uniaxial loading. As a result, understanding the effect of the microstructure (grain size, texture, grain boundaries, etc,) on the progression of plastic deformation -and eventually fracture- in Mg alloys is still a pending issue.
In this investigation, advanced characterization tools -based on electron backscattered diffraction (EBSD) and high-resolution digital image correlation (HR-DIC)- are used in combination with in situ mechanical deformation within the scanning electron microscope to ascertain the dominant deformation mechanisms in pure Mg and Mg alloys with different microstructural features. In particular, the effect of grain boundary orientation on basal-to-basal slip transfer was revealed while it was found that the twins connected at a common grain boundary do not necessarily result from twin transfer events. Instead, the majority of twin pairs at the grain boundary form through the co-nucleation of two twins in neighboring grains. To gain further knowledge, twin nucleation was ascertained in > 3000 grains, including 28 microstructural parameters for each grain. This information was used to train supervised Machine learning classification models to ascertain the link between twinning and microstructural features. In particular, Bayesian network models not only confirm that grains with large sizes and large twinning Schmid factors have a high probability of twinning but also reveal that many-body relationships, such as differences in stiffness and size between a given grain and its neighbors, are crucial for twin nucleation in grains which are unlikely to twin (e.g., small grain size, low or even negative twinning Schmid factor).