To overcome performance limits of 2D poly-graphene as a transparent conductor,  we have invented a concept of 2D conductor (e.g. poly-graphene) doped with metal nanonet (e.g. silver nanowire). The resistance of the hybrid networks is lower than the individual films. Graphene grains allow electrons to bypass the highly resistive NW (nanowire)–NW junctions; likewise the NWs bridge the grain boundaries. Together, these elements mutually erase the percolating bottlenecks, which is an example of percolation engineering. This pioneering concept was experimentally verified and it turned out that the hybrid structures have a multi-fuctional use (i.e., electrical, thermal, chemical, mechanical) beyond original use as a transparent conducting electrode. In the future, we pursue to identify novel 2D layered materials with new functionality: for example, a flexible solid-state power generator for wearable electronics. 

There is a growing consensus that physics-based model needs to be coupled with machine learning (ML) model relying on data or vice versa in order to fully exploit their combined strengths to address scientific or engineering problems that cannot be solve separately. For semiconductor design, we have proposed several innovative methodologies bridging physical simulation and AI to overcome conventional challenges of physical simulation: simulation runtime, coverage, and so on. The use cases of the hybrid modeling and our research focus are as follows: i) accelearting traditional TCAD simulation of semicondor process and devices, ii) scaling-briding simulation of mechanical properties of semiconductor devices, iii) identifying the unknowns with physics-based priors when physical equations are either fully or partially unknown,  iv) to accelerate atomistic MD simulation for new material design, and so on. 

We have studied quasi-ballistic transport and limits of contact resistance in nanoscale transistor by a characterization, a sophisticated transport modeling, and physics-based compact model. Works at Samsung contribute to securing competiveness of logic device technology. Key achievements are as follows: i) simulation-based design on high performance logic transistor for FinFET, ii) exploration of novel materials and architecture for future logic technologies, iii) atomistic simulation of carrier transport modeling in SiGe and III-V channel, iv) self-heating modeling of advanced nanotransistors, and  v)  simulation-based analysis of local layout effect. We are interested in computational modeling of new materials and structures to explore new electronic device possibilites. 

We have contributed to formal extension of the Landauer approach to theromelectric problems involving electrons and phonons.  The Landauer approach presents a simple, clear picture of electron, phonon transport at the nanoscale and also provides a bridge to sophisticated, a quantum simulation using the non-equilibrium Green's function (NEGF) method. Most of concepts are discussed in co-authored book with Prof. Mark Lundstrom, 'Near-equilibrium Transport: Fundamentals and Applications'.

Another key accomplishment is  to examine a variety of new concepts for thermoelectrics such as 2D graphene superlattice, molecular devices,  and materials with distorted bandstructures and to identify the most promising ideas. We will focus on exploring energy transport and conversion in nano-engineered devices and systems to enhance the energy efficiency, performance, and reliability.

For neuromorphic application, phase-change and ovonic-threshold switching materials have regained an attention because a continuous transition between resistance levels of phase-change memory (PCM) devices in an analog manner can be used to program them to mimic the behavior of a synapse. We have studied the scaling limits of PCMs through a solid understanding of electrical/heat conduction and reliability. We are interested in a physcis-based compact model that is universally applicable to a wide range of neuromophic devices, and in a possibility of phase-change materials as new energy conversion device that is compatible with conventional silicon-based devices. This could open up new application such as Si-compatible solid-state cooling.