arXiv subjects

Chulin Wang

Publications and source records attributed to Chulin Wang.

6 recordsLinked to original sources

arXiv2022

Electron subband degeneracy heat pump for cryogenic cooling

An unconventional method of continuous solid-state cryogenic cooling utilizing the electron subband degeneracy of semiconductor heterostructures is proposed in this Letter. An electrostatic heat pump is modeled, which employs subband "expansion" and "compression" to reach sub-dilution refrigeration temperatures with the fundamental limit set by electron-phonon interaction. Using an ultra-wide GaAs quantum well as an example, the cooling power per unit volume is estimated to reach $4.5\ \rm mW/cm^3$ with a hot-side temperature of $300\ \rm mK$, suitable for applications such as quantum computers or infrared detectors.

cond-mat.mes-hall↗

arXiv2021

S3RP: Self-Supervised Super-Resolution and Prediction for Advection-Diffusion Process

We present a super-resolution model for an advection-diffusion process with limited information. While most of the super-resolution models assume high-resolution (HR) ground-truth data in the training, in many cases such HR dataset is not readily accessible. Here, we show that a Recurrent Convolutional Network trained with physics-based regularizations is able to reconstruct the HR information without having the HR ground-truth data. Moreover, considering the ill-posed nature of a super-resolution problem, we employ the Recurrent Wasserstein Autoencoder to model the uncertainty.

cs.LG↗

arXiv2021

Sensitivity Analysis for Optimizing Electrical Impedance Tomography Protocols

Electrical impedance tomography (EIT) is a noninvasive imaging method whereby electrical measurements on the boundary of a conductive medium (the data) are taken according to a prescribed protocol set and inverted to map the internal conductivity (the model). This paper introduces a sensitivity analysis method and corresponding inversion and protocol optimization that generalizes the criteria for tomographic inversion to minimize the model-space dimensionality and maximize data importance. Sensitivity vectors, defined as rows of the Jacobian matrix in the linearized forward problem, are used to map targeted conductivity features from model-space to data-space, and a volumetric outer-product of these vectors in model-space called the sensitivity parallelotope volume provides a figure-of-merit for data protocol optimization. Orthonormal basis functions that accurately constrain the model-space to features of interest can be defined from a priori information. By increasing the contact number to expand the number of possible measurements Dmax, and by reducing the model-space to a minimal number M0 of basis functions that describe only the features of interest, the M0 << Dmax sensitivity vectors of greatest length and maximal orthogonality that span this model-space can be identified. The reduction in model-space dimensionality accelerates the inversion by several orders of magnitude, and the enhanced sensitivity can tolerate noise levels up to 1,000 times larger than standard protocols.

physics.med-ph↗

arXiv2021

Observation of current-induced switching in non-collinear antiferromagnetic IrMn$_3$ by differential voltage measurements

There is accelerating interest in developing memory devices using antiferromagnetic (AFM) materials, motivated by the possibility for electrically controlling AFM order via spin-orbit torques, and its read-out via magnetoresistive effects. Recent studies have shown, however, that high current densities create non-magnetic contributions to resistive switching signals in AFM/heavy metal (AFM/HM) bilayers, complicating their interpretation. Here we introduce an experimental protocol to unambiguously distinguish current-induced magnetic and nonmagnetic switching signals in AFM/HM structures, and demonstrate it in IrMn$_3$/Pt devices. A six-terminal double-cross device is constructed, with an IrMn$_3$ pillar placed on one cross. The differential voltage is measured between the two crosses with and without IrMn$_3$ after each switching attempt. For a wide range of current densities, reversible switching is observed only when write currents pass through the cross with the IrMn$_3$ pillar, eliminating any possibility of non-magnetic switching artifacts. Micromagnetic simulations support our findings, indicating a complex domain-mediated switching process.

cond-mat.mes-hall↗

arXiv2020

Physics-Informed Neural Network Super Resolution for Advection-Diffusion Models

Physics-informed neural networks (NN) are an emerging technique to improve spatial resolution and enforce physical consistency of data from physics models or satellite observations. A super-resolution (SR) technique is explored to reconstruct high-resolution images ($4\times$) from lower resolution images in an advection-diffusion model of atmospheric pollution plumes. SR performance is generally increased when the advection-diffusion equation constrains the NN in addition to conventional pixel-based constraints. The ability of SR techniques to also reconstruct missing data is investigated by randomly removing image pixels from the simulations and allowing the system to learn the content of missing data. Improvements in S/N of $11\%$ are demonstrated when physics equations are included in SR with $40\%$ pixel loss. Physics-informed NNs accurately reconstruct corrupted images and generate better results compared to the standard SR approaches.

cs.CV↗

arXiv2020

Investigating two super-resolution methods for downscaling precipitation: ESRGAN and CAR

In an effort to provide optimal inputs to downstream modeling systems (e.g., a hydrodynamics model that simulates the water circulation of a lake), we hereby strive to enhance resolution of precipitation fields from a weather model by up to 9x. We test two super-resolution models: the enhanced super-resolution generative adversarial networks (ESRGAN) proposed in 2017, and the content adaptive resampler (CAR) proposed in 2020. Both models outperform simple bicubic interpolation, with the ESRGAN exceeding expectations for accuracy. We make several proposals for extending the work to ensure it can be a useful tool for quantifying the impact of climate change on local ecosystems while removing reliance on energy-intensive, high-resolution weather model simulations.

physics.ao-ph↗