arXiv subjects

Eric Ngo

Publications and source records attributed to Eric Ngo.

2 recordsLinked to original sources

arXiv2026

Unlocking extreme doping and strain in epitaxial monocrystalline silicon

Hyperdoping, overcoming the solubility limit of dopants in a crystalline semiconductor, is a fertile method for the enhancement of the electrical, structural and optical devices' performances and for the exploration of exotic phases such as superconductivity. We demonstrate an unprecedented control on the dopant concentration and lattice deformation via nanosecond laser doping in epitaxial boron doped silicon, achieving record carrier concentrations (8 at.%) and lattice deformations (3 %). Probing the microscopical hyperdoping limitations, we show that the relevant mechanisms are caught by a simple combinatorial model, which quantitatively explains both the experimental carrier concentration and lattice deformation evolution. First principle calculations complete and support such simple model. Indeed, at the high doping levels now attainable, the maximum carrier concentration is inherently limited by the probability of two or three substitutional dopants occupying neighboring lattice sites, forming partially inactive complexes that we detail. This description is valid in the case of perfect layers with no crystallographic defects and a fully substitutional dopant occupation, highlighting the quality of the epitaxial layers realized.

cond-mat.mtrl-sci↗

arXiv2025

Feature Engineering vs. Deep Learning for Automated Coin Grading: A Comparative Study on Saint-Gaudens Double Eagles

We challenge the common belief that deep learning always trumps older techniques, using the example of grading Saint-Gaudens Double Eagle gold coins automatically. In our work, we put a feature-based Artificial Neural Network built around 192 custom features pulled from Sobel edge detection and HSV color analysis up against a hybrid Convolutional Neural Network that blends in EfficientNetV2, plus a straightforward Support Vector Machine as the control. Testing 1,785 coins graded by experts, the ANN nailed 86% exact matches and hit 98% when allowing a 3-grade leeway. On the flip side, CNN and SVM mostly just guessed the most common grade, scraping by with 31% and 30% exact hits. Sure, the CNN looked good on broader tolerance metrics, but that is because of some averaging trick in regression that hides how it totally flops at picking out specific grades. All told, when you are stuck with under 2,000 examples and lopsided classes, baking in real coin-expert knowledge through feature design beats out those inscrutable, all-in-one deep learning setups. This rings true for other niche quality checks where data's thin and know-how matters more than raw compute.

cs.LG↗