A Foundation Model for Non-Destructive Defect Identification from Vibrational Spectra

2025-05-31

AI Summary

Quantitative, nondestructive characterization of multiple coexisting point defects in solids remains a long-standing challenge. This work introduces DefectNet, a foundational deep learning model that enables end-to-end inversion of both chemical identities and concentrations (0.2–25 at.%) of substitutional point defects directly from phonon density-of-states (DOS) spectra. DefectNet integrates a vibration-spectrum-aware attention mechanism with a transfer-learning architecture and is trained on over 16,000 simulated phonon DOS spectra. It exhibits strong generalization across 56 elemental crystals. The model achieves high accuracy on experimental data from SiGe alloys and MgB₂ superconductors and supports fine-tuning for real-world measurements. This work establishes vibrational spectroscopy—specifically phonon DOS—as a new paradigm for quantitative defect characterization, delivering a scalable, noninvasive analytical tool for materials defect engineering.

Technology Category

• Machine Learning: Calibration & Uncertainty Quantification
• Knowledge Representation and Reasoning: Diagnosis and Abductive Reasoning
• Computer Vision: Other Foundations of Computer Vision

Application Category

• Graph Algorithms and Modeling for the Web: Graph embeddings and representation learning for Web-related graphs
• Systems and Infrastructure for Web, Mobile and WoT: Web performance, measurement, and characterization
• Web Mining and Content Analysis: Large pretrained models with web data

Abstract

Defects are ubiquitous in solids and strongly influence materials' mechanical and functional properties. However, non-destructive characterization and quantification of defects, especially when multiple types coexist, remain a long-standing challenge. Here we introduce DefectNet, a foundation machine learning model that predicts the chemical identity and concentration of substitutional point defects with multiple coexisting elements directly from vibrational spectra, specifically phonon density-of-states (PDoS). Trained on over 16,000 simulated spectra from 2,000 semiconductors, DefectNet employs a tailored attention mechanism to identify up to six distinct defect elements at concentrations ranging from 0.2% to 25%. The model generalizes well to unseen crystals across 56 elements and can be fine-tuned on experimental data. Validation using inelastic scattering measurements of SiGe alloys and MgB$_2$ superconductor demonstrates its accuracy and transferability. Our work establishes vibrational spectroscopy as a viable, non-destructive probe for point defect quantification in bulk materials, and highlights the promise of foundation models in data-driven defect engineering.

Problem

• Non-destructive identification of multiple coexisting defects in solids
• Quantifying defect types and concentrations from vibrational spectra
• Developing a foundation model for defect analysis in diverse materials

Innovation

• Uses DefectNet for defect identification
• Analyzes vibrational spectra non-destructively
• Trains on 16,000 simulated semiconductor spectra