top of page

RESEARCH INTERESTS

Beyond Von Neumann computing architectures

Neuromorphic computing devices

NM3.jpg

Neuromorphic computing that emulates the learning and cognitive abilities of human neural networks can define a new era of the futuristic computational architecture due to its extremely efficient energy saving and parallel data processing capability. In biological environment, neurons and synapses, which utilize a distributed, parallel, and event-driven operation, are the heart of neural circuitry to complete the learning and memorization of the human brain. In this sense, the construction of highly effective artificial neurons and synapses is the prerequisite to embed the biological functions in the digital architecture. Our goal is to deploy many tools available to introduce the biological nervous behavior into the digital architecture.

Multi-valued logic beyond 0 and 1

The advent of solid-state transistors revolutionized the electronics industry by opening the era of binary devices, where information is represented using two logic states: ON and OFF. However, this binary framework, while efficient in some respects, also definitely imposes limitations, particularly in terms of information density per interconnect. Multi-valued logic (MVL) offers a promising alternative by allowing for more than two logic levels, thereby potentially increasing computational efficiency and reducing power consumption. This approach has garnered increased attention, especially given the constraints posed by the power density limits of conventional complementary metal–oxide–semiconductor (CMOS) technology. Researchers have been exploring various MVL unit devices based on both conventional and emerging materials. These efforts aim to develop devices capable of accommodating multiple logic states while addressing the challenges associated with power consumption and computational efficiency. The goal of Lee group includes the implementation of the various inspiration to the eye-opening technologies to realize the new generation multi-valued logic circuits.

Logic-in-memory computing

Current computing systems are built on the Von Neumann architecture, where the CPU is tasked with control and arithmetic/logic operations, and memory serves only to store and retrieve data. However, as technology scales down, the performance gap between these two separate units grows, creating a critical bottleneck known as the Von Neumann bottleneck. To address this, Processing-in-Memory (PIM) has been introduced—a paradigm that enables computation to occur directly within memory, eliminating the need for data transfers between the CPU and memory over the system bus, thus significantly boosting efficiency and performance. The Lee group is dedicated to pioneering advancements in the field of next-generation computing, aiming to break through the limitations of traditional computing architectures. By integrating logic operations directly into memory, we focus on developing new device architectures that combine logic and memory at the material and circuit levels, reducing data transfer bottlenecks in traditional computing systems.

Fundamentals of plasmonic hot carrier and its applications

HC.png

Hot carriers are in a highly excited state and out of thermal equilibrium while confined to a metal surface during photon absorption and exothermal chemical reactions. Since carriers with high-kinetic energies of 1−3 eV away from the Fermi energy of the metal relax to phonon scattering within a ten of few femtoseconds, and disappear via energy dissipation into the surrounding materials, it has been difficult to deal with the fast-disappearing charges from the metal surface. However, recent research has proposed that these hot carriers could be applied to effective energy converters in optoelectronic and photocatalytic applications. The focus in our Lab is on investigating the fundamental physics of plasmonic hot carriers as well as realizing futuristic energy harvesting, environment science, and opto-electronic applications.

Developments of compound semiconducting materials and devices

CS.png

The Lee group also takes a profound interest in exploring compound semiconducting materials and their next generation applications/ architectures, such as micro light-emitting diodes (LEDs), ultraviolet (UV) LED, power device, and monolithic 3D heterogeneous integration. Our strategies cover facile and efficient methods to fabricate compound semiconducting materials, such nanocrystal or nanovoid formation. We are also developing a distinctive deposition technique to grow compound semiconducting materials using plasma-enhanced atomic layer deposition (PEALD). 

bottom of page