Spectral Library Search

Identifying an unknown compound from its spectrum is one of the most common tasks in analytical spectroscopy. Traditional library searching requires expensive software with proprietary reference databases — tools that can cost thousands of dollars and lock your workflow to a single platform. For many researchers, teaching labs, and QA teams, that barrier makes compound identification slower and more costly than it needs to be.

SpectralBench provides a free, browser-based spectral library search that removes that friction entirely. Upload an unknown spectrum, search against the bundled reference library or your personal collection, and review HQI-ranked matches with visual overlays — all without installing software or uploading your data to a server. All matching runs client-side, so your data never leaves your machine. Once you have candidate matches, use the Compare Spectra tool for a detailed overlay, or the Peak Identifier to correlate absorption bands with known functional groups.

How It Works

Upload a spectrum in any supported format — JCAMP-DX, SPC, Bruker OPUS, CSV, or TXT — and SpectralBench preprocesses it and compares it against every entry in the reference library using HQI similarity scoring. Results are ranked by match quality, with the highest scoring candidates listed first alongside their HQI scores.

Click any result to overlay the reference spectrum against your query spectrum for a direct visual comparison. This makes it easy to spot whether the agreement is strong across the full spectral range or confined to a few key bands. You can adjust search parameters to focus on specific spectral regions — for example, restricting the comparison to the fingerprint region for more discriminating organic compound identification.

The search runs simultaneously across both the bundled library and any personal references you have saved, so you get a unified ranked list from a single query. No need to run separate searches against different databases.

Understanding Spectral Matching

The HQI (Hit Quality Index) algorithm computes the correlation between the query spectrum and each reference entry after normalization. A score of 100 indicates a perfect match; scores above 90 are generally considered strong candidates for positive identification, while scores in the 70-90 range warrant closer inspection.

The quality of library search results depends heavily on how well the spectra are preprocessed before comparison. Baseline correction removes fluorescence background, drift, and instrument artifacts that would otherwise reduce correlation scores across the board. Normalization puts the query and reference spectra on the same intensity scale, preventing mismatches caused by concentration differences or detector gain settings.

The fingerprint region (600–1500 cm⁻¹) is often the most discriminating for organic compound identification in FTIR spectra, as subtle differences in molecular structure produce distinct absorption patterns in this range. Restricting the HQI calculation to this region can improve discrimination between closely related compounds. For Raman spectra, the low-frequency region below 500 cm⁻¹ can be equally diagnostic. SpectralBench's preprocessing tools can prepare your spectra — including baseline correction and normalization — before you run a library search.

Building a Personal Library

Save verified spectra as reference entries in your personal library directly from the workbench. Organize references by category — solvents, polymers, analytical standards, raw materials — to keep your collection manageable as it grows. When you run a library search, SpectralBench queries your personal collection alongside the bundled database and presents a single unified ranked list.

Personal libraries are stored in your browser's local storage, so they are private, persistent across sessions, and never transmitted to any server. You can export your library as a file for backup or to share a curated set of references with colleagues.