Automatic search versus chemical rules in materials structure study

Maosheng Miao
California State University, Northridge (CSU Northridge)

The increase of the computer power in the past decades not only allow us to calculate
larger systems with higher accuracy in materials studies, but also provide the opportunity
to explore large configuration spaces such as structures and compositions. Automatic
structure searches have been very successful in predicting structures of bulk materials. It
seems out of question whether the automatic search is advantageous over traditional
structure design based on chemical knowledge and intuition. Here we give several
examples of computational materials, both related to structure search and design. In the
first example, we show the development of an efficient method that can automatically
explore the surface structures by virtue of structure swarm intelligence. While applying
the method on the "simple" diamond (100) surface, we discovered a hitherto unexpected
surface reconstruction featuring self-assembly of carbon nanotubes (CNTs) arrays. The
intriguing covalent bonding between the neighboring tubes creates a unique feature of
carrier kinetics--one dimensionality of hole states whereas two dimensionality of
electron states, which may lead to novel design of superior electronics. In the second
example, we propose and demonstrate a large family of two-dimensional semiconductors (2DSC), all adopting the same structure and consisting of only main group elements. We demonstrate the attainability of these materials, and show that they cover a large range of lattice constants, band gaps and band edge states, therefore are good candidate materials for heterojunctions. 2DSCs are currently the focus of many studies. Comparing with the 3D semiconductors, the choice of the 2D materials is very limited. The new 2DSCs may pave a way toward fabrication of 2DSC devices at the same thriving level as 3D semiconductors. Our examples show that automatic structure search and chemical knowledge and rules can both be important in materials discovery and can benefit from each other.

Back to Workshop I: Optimization and Optimal Control for Complex Energy and Property Landscapes