arXiv subjects

Seungwon Jung

Publications and source records attributed to Seungwon Jung.

7 recordsLinked to original sources

arXiv2026

AIMS: an AI experimentalist turns uncertainty into quantum matter discovery

Most AI agents act only after scientists have defined the task. Discovery is harder under practical uncertainties: the probe may not be where it is expected, the signal may occupy only a small region of a disordered sample, and the evidence may not distinguish among competing explanations. Here we show that an AI agent can decide what evidence an uncertain experiment needs next, and act on it. Beyond automation, AIMS, an uncertainty-aware experimentalist for cryogenic microwave impedance microscopy, quantifies uncertainty where it originates, in perception, sampling, and interpretation, and converts each into its own corrective action rather than a single confidence score. Given only an open objective, AIMS relocated a probe lost during cooldown while flagging its own unreliable estimates, mapped twist angle disorder to locate the strongest correlated states in twisted bilayer MoSe$_2$, and uncovered a paradox: the half-filled stripe that classical theory predicts should melt first survived longest. Distinguishing an incomplete model from a wrong mechanism, AIMS commissioned a beyond-mean-field calculation and an independent structural measurement as the decisive tests, revising its interpretation as each arrived: quantum motion reverses the classical hierarchy, stabilizing the half-filled stripe while destabilizing its neighbors. These uncertainty-to-action loops are generic to scanning probe experiments, and AIMS turns uncertainty from an obstacle into a driver of discovery.

cond-mat.str-el↗

arXiv2025

Local thermodynamic DOS measurement and twist-angle mapping in graphene-hBN superlattices

Moir\'e patterns arising from twisted van der Waals stacks fundamentally reshape their electronic properties, enabling band-structure engineering that has driven rapidly growing interest in this field. In studying electronic properties, however, structural disorder present in real devices often leads to twist-angle inhomogeneity and obscures angle-dependent electronic effects when measured with bulk-averaged measurements. Probes that can access local thermodynamic response of the electronic systems with high sensitivity would be highly valuable. Here, we adopt Kelvin probe force microscopy (KPFM) to locally investigate graphene-hBN superlattices. By additionally modulating the chemical potential of the system, we obtain the inverse compressibility with high signal-to-noise ratio, enabling extraction of the local thermodynamic DOS. From this information, we determine the local twist angle along the device and find that twist-angle deviations are strongly correlated with bubble-induced strain features. Furthermore, by simultaneously tracking the offsets in the contact potential difference and in the net charge, we identify which interface within the heterostructure hosts the trapped bubbles. This capability to identify local electro-chemical environments provides a practical tool for strain-based studies and future device designs utilizing nanoscale engineering in moir\'e systems.

cond-mat.mes-hall↗

arXiv2025

Cryogen-free variable-temperature Kelvin probe force microscopy for probing local chemical potential in a graphene heterostructure

We report the development of a variable-temperature Kelvin probe force microscopy (KPFM) system capable of stable and highly sensitive operation over a wide temperature range based on a GM-cooler-based cryogen-free cryostat. The system incorporates a custom-designed phase-locked loop and automatic gain control, along with passive vibration isolation, enabling precise measurements of local chemical potential even under cryogenic conditions. We demonstrate the performance of this setup by measuring hBN encapsulated monolayer graphene (MLG), revealing spatially resolved electronic inhomogeneities and charge puddles. Our measurements clearly capture temperature-dependent variations in the chemical potential near the charge neutrality point (CNP), consistent with the linear band dispersion of MLG and interaction-driven renormalization of Fermi velocity. This work highlights the robust sensitivity and stability of our system, making it a versatile local probe of quantum phases in van der Waals heterostructures.

physics.app-ph↗

arXiv2024

Stop using Landau gauge for Tight-binding Models

To analyze the electronic band structure of a two-dimensional (2D) crystal under a commensurate perpendicular magnetic field, tight-binding (TB) Hamiltonians are typically constructed using a magnetic unit cell (MUC), which is composed of several unit cells (UC) to satisfy flux quantization. However, when the vector potential is constrained to the Landau gauge, an additional constraint is imposed on the hopping trajectories, further enlarging the TB Hamiltonian and preventing incommensurate atomic rearrangements. In this paper, we demonstrate that this constraint persists, albeit in a weaker form, for any linear vector potential ($\mathbf{A}(\mathbf{r})$ linear in $\mathbf{r}$). This restriction can only be fully lifted by using a nonlinear vector potential. With a general nonlinear vector potential, a TB Hamiltonian can be constructed that matches the minimal size dictated by flux quantization, even when incommensurate atomic rearrangements occur within the MUC, such as moiré reconstructions. For example, as the twist angle $θ$ of twisted bilayer graphene (TBG) approaches zero, the size of the TB Hamiltonian scales as $1/θ^4$ when using linear vector potentials (including the Landau gauge). In contrast, with a nonlinear vector potential, the size scales more favorably, as $1/θ^2$, making small-angle TBG models more tractable with TB.

cond-mat.mes-hall↗

arXiv2023

You Truly Understand What I Need: Intellectual and Friendly Dialogue Agents grounding Knowledge and Persona

To build a conversational agent that interacts fluently with humans, previous studies blend knowledge or personal profile into the pre-trained language model. However, the model that considers knowledge and persona at the same time is still limited, leading to hallucination and a passive way of using personas. We propose an effective dialogue agent that grounds external knowledge and persona simultaneously. The agent selects the proper knowledge and persona to use for generating the answers with our candidate scoring implemented with a poly-encoder. Then, our model generates the utterance with lesser hallucination and more engagingness utilizing retrieval augmented generation with knowledge-persona enhanced query. We conduct experiments on the persona-knowledge chat and achieve state-of-the-art performance in grounding and generation tasks on the automatic metrics. Moreover, we validate the answers from the models regarding hallucination and engagingness through human evaluation and qualitative results. We show our retriever's effectiveness in extracting relevant documents compared to the other previous retrievers, along with the comparison of multiple candidate scoring methods. Code is available at https://github.com/dlawjddn803/INFO

cs.CL↗

arXiv2010

Reversible Metal-Semiconductor Transition of ssDNA-Decorated Single-Walled Carbon Nanotubes

A field effect transistor (FET) measurement of a SWNT shows a transition from a metallic one to a p-type semiconductor after helical wrapping of DNA. Water is found to be critical to activate this metal-semiconductor transition in the SWNT-ssDNA hybrid. Raman spectroscopy confirms the same change in electrical behavior. According to our ab initio calculations, a band gap can open up in a metallic SWNT with wrapped ssDNA in the presence of water molecules due to charge transfer.

cond-mat.mes-hall↗

arXiv2010

Dissociation of ssDNA - Single-Walled Carbon Nanotube Hybrids by Watson-Crick Base Pairing

The unwrapping event of ssDNA from the SWNT during the Watson-Crick base paring is investigated through electrical and optical methods, and binding energy calculations. While the ssDNA-metallic SWNT hybrid shows the p-type semiconducting property, the hybridization product recovered metallic properties. The gel electrophoresis directly verifies the result of wrapping and unwrapping events which was also reflected to the Raman shifts. Our molecular dynamics simulations and binding energy calculations provide atomistic description for the pathway to this phenomenon. This nano-physical phenomenon will open up a new approach for nano-bio sensing of specific sequences with the advantages of efficient particle-based recognition, no labeling, and direct electrical detection which can be easily realized into a microfluidic chip format.

cond-mat.mes-hall↗