General Description

The objects of our group's research are functional electronic and optoelectronic materials and related systems (material complexes), which are in most cases solid-state matter that can be described in terms of three- or two-dimensional periodicity in space. Today, the performance (efficiency, stability, etc.) of these materials and relevant devices is closely related to their properties on nanoscale, i.e., crystal structure, atomic structure, and electronic structure. Quantum mechanical first-principles (ab initio) modelling is fundamental to explore these properties. The success of Density-functional-theory (DFT) provides a most efficient method to handel these problems, as it can properly balance the computational accuracy and expense. In particular, DFT methods are the only practical ones for "complex" materials systems, the size of whose model systems easily becomes close to the upper limit for periodic DFT calculations and is thus unfeasible for other though more accurate approaches.


The research profile of our group can be summarized according to different classifications as follows:

Application-oriented: Halide perovskites for optoelectronics; oxide perovskite for ferro- and piezoelectrics; other functional-materials systems
System-oriented: Bulk materials; alloys; defective systems; surfaces and interfaces
Method-oriented: Conventional DFT modelling; data-driven / machine-learning computational materials science










There are two major characters in today's functional materials study that motivate us to go beyond conventional DFT calculations on the simple bulk materials. On the one hand, doping and alloying are commonly applied to capture the merits of different compositions, such as [Pb(Mg1/3Nb2/3]O3]x[PbTiO3]1-x (xPMN-PT) for ferroelectrics and K-doped CsxFA1-xPb(IyBr1-y)3 for photovoltaics. Addressing the many possible alloying configurations and to have a full picture over the whole alloying range is beyond the power of conventional DFT, because in principles we need to study an infinite number of super large model systems. We are managing to employ the novel data-driven and machine-learning computational materials science methods which are supposed to be very powerful for solving these problems.


On the other hand, we pay much attention to models systems beyond simple bulk materials, e.g., (quasi-)two-dimensional materials, surfaces, and interfaces. Interface zwischen the "active" material and another "functional" material plays an important role to the device performance. To correctly describe the interfacial atomic and electronic structure and simulate interfacial processes (such as charge transfer and ion exchange), we first need to have correct surface models, for which DFT-based atomistic thermodynamics is required.