Scholarlink

<p>Infrared (IR) spectroscopy is a fundamental tool for molecular identification and characterization, yet comprehensive IR spectral databases remain limited, particularly for small organic molecules with well-defined theoretical baselines. Here we introduce SQuIRL, the Spectral Quantum Chemistry and Infrared Resonance Library, a collection of computed IR spectra for 133,885 organic molecules. Each entry provides vibrational frequencies and intensities with near-benchmark accuracy, thereby extending the QM9 dataset by incorporating vibrational fingerprints alongside its structural and electronic descriptors. The resulting dataset enables data-driven spectrum prediction, machine-learning model training, and automated molecular identification. SQuIRL establishes a new foundation for high-fidelity, quantum-accurate infrared spectroscopy in computational chemistry. Distributed in structured HDF5 format with an accessible API, it integrates seamlessly into AI-based spectroscopy workflows and molecular discovery pipelines, offering a widely accessible resource for data-driven approaches in vibrational spectroscopy.</p>

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

Spectral Quantum Chemistry and Infrared Resonance Library for Data-Driven Molecular Spectroscopy

  • 期刊论文

  • 出版模式：

    开放获取

  • 发表日期：

    2026年4月18日

查看全文
  • Anirudh Krishnadas,
  • Jatin Kansal,
  • Nicholas E. Charron,
  • Anita Ragyanszki

摘要

Infrared (IR) spectroscopy is a fundamental tool for molecular identification and characterization, yet comprehensive IR spectral databases remain limited, particularly for small organic molecules with well-defined theoretical baselines. Here we introduce SQuIRL, the Spectral Quantum Chemistry and Infrared Resonance Library, a collection of computed IR spectra for 133,885 organic molecules. Each entry provides vibrational frequencies and intensities with near-benchmark accuracy, thereby extending the QM9 dataset by incorporating vibrational fingerprints alongside its structural and electronic descriptors. The resulting dataset enables data-driven spectrum prediction, machine-learning model training, and automated molecular identification. SQuIRL establishes a new foundation for high-fidelity, quantum-accurate infrared spectroscopy in computational chemistry. Distributed in structured HDF5 format with an accessible API, it integrates seamlessly into AI-based spectroscopy workflows and molecular discovery pipelines, offering a widely accessible resource for data-driven approaches in vibrational spectroscopy.