All-in-one smart sensors overcome physical limitations by leveraging embedded artificial intelligence algorithms to detect and analyze light, color, heat, and other phenomena that cannot be directly measured. It has been believed that all spectroscopic analyses rely on light dispersion and require bulky, heavy optical modules or mechanical components. However, miniaturized computational spectrometers without light dispersion can be powered by combining electrically tunable spectral responses based on tunable interlayer transport, band alignments, and exciton dynamics of van der Waals heterostructures with reconstruction algorithms (Figure A). This enables us to analyze thousands of visible and invisible colors, identify gas or biomarker compositions, and detect polarization or thermal radiation on a chip, not in a lab (Figure B).

We explore optical interconnects, topological gate dielectrics, topological semimetal contacts, and single-channel logic gates for next-generation logic, memory, and AI systems, aiming to overcome the fundamental limitations of Moore's Law and to maximize sensing, storage, processing, and communication capabilities while minimizing energy dissipation and managing heat. Emerging devices designed to break Moore's Law, improve performance, suppress power consumption, reduce footprint, minimize process costs, and diversify applications. For example, we implemented anti-ambipolar phototransistors switchable by incident light wavelength (Figure A) by integrating a confined quantum well with a charge trapping layer (Figure B). This can build multivalued logic gates within a single device (Figure C), maximizing information density.

We investigate one-dimensional or zigzag superlattices formed in two-dimensional Dirac fermionic graphene, in which current can be treated as light (Figure A). If electrons are treated as photons, focusing, collimation, diffraction, and interference phenomena considered in traditional optics can be implemented in electronic or quantum devices. Our goal is to control the trajectory of ballistic carriers in graphene by using quantum point contacts, Veselago lens, Klein collimator, and local artificial atom substitution (Figure B), and to observe interference in the double slit or Aharonov-Bohm ring structures. In particular, electron interferometers are typically known to require magnetic fields for tuning, but we may be able to tune interference solely with electric fields, which could pave the way for electron-interference-based quantum transistors.

Various van der Waals quantum materials are promising platforms for emerging quantum devices, such as Josephson junctions/diodes and transmon/gatemon qubits, as well as for exotic quantum phenomena, including Mott transitions and proximity/Stark/spin-galvanic effects. We fabricate van der Waals heterostructures with a wide variety of quantum materials, like superconductors, topological/Mott insulators/semimetals, and ferromagnets, engineer their interfaces or induce phase transition via plasma treatment (Figure A) or laser illumination (Figure B), and characterize them in a cryogenic, high-magnetic-field superconducting magnet measurement system. We aim to implement coherent compact superconducting transmon qubits for practical quantum computing, which requires advances in Josephson-junction materials and processes.