EPR Spectroscopy Analysis Platform

ChemFitX provides an online EPR spectroscopy analysis workflow for researchers working with electron paramagnetic resonance data from radicals, paramagnetic defects, transition-metal systems and other materials containing unpaired electrons. The platform combines data import, signal preprocessing, resonance analysis, visualization and export in one browser-based workspace.

Upload an EPR spectrum, select the correct magnetic-field unit and microwave frequency, choose derivative or absorption mode, and then inspect the processed spectrum together with calculated resonance parameters. The current platform can report signal amplitude, resonance field, linewidth, g-value, integrated intensity, estimated signal-to-noise ratio, apparent hyperfine spacing and, when a suitable reference standard is supplied, a reference-based estimate of spin concentration.

The analysis is performed locally in your web browser by this application. Your spectroscopy file is read in the browser and is not uploaded to ChemFitX servers during the analysis.

What the EPR Analysis Platform Can Do

Import EPR datasets from XLSX, XLS, CSV, TSV and TXT files.
Choose the magnetic-field and signal columns from uploaded tabular data.
Work with first-derivative CW-EPR spectra or absorption spectra.
Use magnetic-field units in gauss (G), millitesla (mT) or tesla (T).
Enter the experimental microwave frequency in GHz for g-value calculation.
Apply linear, constant, second-order polynomial, third-order polynomial or manual constant baseline correction.
Apply Savitzky–Golay, moving-average or Gaussian smoothing.
Analyze resonance field, linewidth, amplitude, integrated intensity and estimated SNR.
Inspect an apparent hyperfine-spacing estimate and a table of detected resonance features.
Optionally estimate spins per gram using a reference standard and sample/standard masses.
Zoom, pan, inspect cursor values, undo or redo processing steps, and reset the current analysis.
Export processed numerical data to CSV or Excel.

Supported Data and Required Inputs

File formats

The platform accepts XLSX, XLS, CSV, TSV and TXT files. The selected data should contain at least two numerical columns: one for magnetic field and one for the measured EPR signal. After import, choose the appropriate X and Y columns if the automatic selection does not match your file.

Magnetic-field units

Select the unit used in the uploaded spectrum: G, mT or T. The platform converts the field internally to gauss for resonance calculations and then displays the results in the selected unit. Incorrect field units will directly produce incorrect resonance fields, linewidths, hyperfine spacing and g-values.

Microwave frequency

Enter the actual microwave frequency used to acquire the EPR spectrum. The default value is 9.500 GHz, which is representative of X-band measurements, but it should be replaced with the experimental value whenever the measured frequency is known. A precise g-value requires both the correct resonance magnetic field and the correct microwave frequency.

Derivative or absorption spectrum

Choose “First derivative” for conventional field-modulated CW-EPR data and “Absorption spectrum” when the uploaded signal is already an absorption-like spectrum. The analysis logic and linewidth definition change according to this selection.

EPR Preprocessing

Baseline correction

Baseline correction removes slowly varying offsets or slopes that can distort amplitudes, integrals and feature detection. ChemFitX provides five baseline options: linear end-region baseline, constant end-region mean, polynomial order 2, polynomial order 3 and a user-defined manual constant. For automatic end-region methods, the selected end-region fraction controls how much of the beginning and end of the spectrum is used to estimate the baseline.

The end regions should represent baseline as well as possible. If a broad EPR signal occupies much of the scan range, an end-region baseline may remove real spectral intensity. In that case, inspect the corrected trace carefully and use a baseline model that is justified by the experiment.

Smoothing

Noise reduction can be applied using Savitzky–Golay, moving-average or Gaussian smoothing. The Savitzky–Golay implementation performs local polynomial least-squares smoothing and allows selection of the window size and polynomial order. Smoothing can improve visual inspection and peak detection, but an overly large window can broaden or suppress genuine spectral structure. Use the smallest amount of smoothing that achieves the analytical purpose.

How EPR Parameters Are Calculated

Resonance field and g-value

For a field-swept EPR experiment, the g-value is related to microwave frequency and resonance field by the resonance condition g = hν/(μB Bres), where h is Planck’s constant, ν is the microwave frequency, μB is the Bohr magneton and Bres is the resonance magnetic field. In derivative mode, ChemFitX estimates the principal resonance field from the dominant zero crossing. In absorption mode, it uses the field position of the dominant absorption maximum.

The calculated g-value is only as accurate as the microwave-frequency and field values supplied to the platform. Field calibration, frequency precision and the definition of the resonance position therefore matter when high-accuracy g-values are required.

Linewidth

For first-derivative spectra, the platform reports the peak-to-peak linewidth ΔHpp as the magnetic-field separation between the dominant positive and negative derivative extrema. For absorption spectra, it reports the full width at half maximum (FWHM). These two linewidth definitions are not interchangeable and should be identified clearly when comparing results.

Signal amplitude

In derivative mode, signal amplitude is reported as the peak-to-peak difference between the dominant positive and negative extrema. In absorption mode, the reported amplitude is the difference between the maximum and minimum signal levels of the processed trace.

Integrated intensity

For derivative CW-EPR data, the platform first numerically integrates the derivative spectrum to obtain an absorption-like trace and then performs a second numerical integration to obtain a double-integrated intensity. For absorption-mode data, it performs a single numerical integration of the absorption signal. Integrated intensity can be useful for relative quantification, but meaningful comparison requires consistent acquisition, processing and calibration conditions.

Estimated signal-to-noise ratio

The platform reports an estimated SNR in decibels. Noise is estimated from signal values near both ends of the processed spectrum, while the signal term is based on the measured amplitude. This is a convenient diagnostic rather than a universal metrological SNR definition. If the EPR signal extends into the spectral edges, the noise estimate may be biased.

Apparent hyperfine spacing

When multiple resonance features are detected, ChemFitX calculates an apparent spacing from the separation of detected features. This can help with quick inspection of simple, well-resolved spectra, but it is not a spin-Hamiltonian fit and does not determine a complete hyperfine tensor. Complex, overlapping, anisotropic or strongly coupled EPR spectra should be interpreted using dedicated EPR simulation and fitting software.

Reference-Based Spin Quantification

The optional spin-quantification panel is designed for relative, reference-based estimation of spins per gram. Absolute spin concentration cannot be obtained from one arbitrary, uncalibrated spectrum without appropriate calibration information. For this ChemFitX workflow, provide a reference standard measured under comparable conditions.

The implemented calculation uses the sample integrated intensity relative to the standard integrated intensity, corrects for the sample and standard masses, and applies an optional response-correction factor:

Csample = Cstandard × (Isample / Istandard) × (mstandard / msample) × response correction factor

Use a response correction factor of 1 only when the relevant acquisition and response conditions are sufficiently matched. Differences in microwave power, receiver gain, modulation amplitude, resonator response, temperature, sample geometry or other instrument conditions can affect EPR intensity and may require validated correction or a more rigorous quantitative protocol.

Detected Resonance Features

After analysis, the platform lists detected resonance features in a table together with field position, calculated g-value, feature type and signal value. In derivative mode, detected features are based on relevant zero crossings. In absorption mode, the platform detects absorption maxima subject to the selected prominence and minimum-separation settings.

Feature detection is intended as an inspection aid. A detected peak or zero crossing should not automatically be assigned to a specific paramagnetic species without considering the complete spectrum, experimental conditions and appropriate literature or simulation.

How to Use the ChemFitX EPR Analysis Tool

Upload an EPR data file in XLSX, XLS, CSV, TSV or TXT format.
Confirm the magnetic-field/X column and signal/Y column.
Select whether the data are a first-derivative CW-EPR spectrum or an absorption spectrum.
Select the correct magnetic-field unit and enter the experimental microwave frequency.
Inspect the raw trace before preprocessing.
Apply baseline correction if the spectrum contains a meaningful offset, slope or curvature.
Apply smoothing only when needed, using a window small enough to preserve real spectral features.
Set the peak-prominence and minimum-separation controls when feature detection needs adjustment.
Click “Analyze Spectrum” and inspect the numerical results, spectrum and detected-feature table together.
If spin quantification is required, enter the reference standard information and sample/standard masses after the spectrum has been analyzed.
Export processed data and results to CSV or Excel when you are satisfied with the analysis.

Interpreting the Main EPR Results

g-value

The g-value describes the magnetic resonance condition of the electron spin system. Values close to the free-electron region are common for many organic radicals, whereas transition-metal and anisotropic systems can show substantially different or direction-dependent g-values. A single scalar g-value from a one-dimensional spectrum should therefore be interpreted in the context of the sample and experiment.

Linewidth

Linewidth can reflect unresolved hyperfine structure, relaxation, dipolar interactions, exchange, magnetic-field inhomogeneity and other broadening processes. ChemFitX measures linewidth from the processed trace; it does not decompose the observed width into specific physical broadening mechanisms.

Integrated intensity

Integrated EPR intensity can track the amount of paramagnetic material when acquisition and response conditions are appropriately controlled. For derivative spectra, the double integral is the relevant intensity measure in this platform. Quantitative comparison requires consistent processing and calibration.

Hyperfine spacing

Regularly spaced resonance features can indicate hyperfine coupling in simple spectra. The platform’s reported value is an apparent spacing estimate only. It should not be treated as a complete hyperfine analysis for anisotropic, multi-component or strongly overlapping spectra.

Why the Processing Choices Matter

EPR preprocessing changes the numerical trace used for downstream calculations. Baseline correction can strongly affect integration, while smoothing can alter extrema, zero crossings, linewidths and small features. For reproducible analysis, keep a record of the selected spectrum type, field unit, microwave frequency, baseline method, smoothing method, window size and feature-detection settings together with the exported data.

A visually smooth spectrum is not automatically a more accurate spectrum. Processing should be selected according to the measurement quality and analytical objective, and important conclusions should be checked against the unprocessed data.

Data Privacy and Browser-Based Processing

ChemFitX reads and analyzes the EPR file locally in the user’s browser for this application. The uploaded spectroscopy file is not sent to ChemFitX servers by the analysis code. This design allows researchers to process local datasets without uploading the scientific file to a remote analysis server.

As with any browser-based workflow, users should still follow their institution’s own data-governance, confidentiality and device-security requirements.

Scientific Scope and Limitations

This platform is a data-processing and parameter-extraction tool; it is not a general spin-Hamiltonian simulation engine.
The automatic resonance and hyperfine-feature routines are most suitable for spectra with sufficiently resolved features.
Complex anisotropic spectra, transition-metal spectra, overlapping radicals, powder patterns and multi-component systems can require dedicated EPR simulation or global fitting.
The reported g-value depends on the correctness and calibration of the magnetic field and microwave frequency.
Baseline and smoothing settings can change linewidths, amplitudes and integrated intensities.
The SNR value is an estimate based on edge-region noise and may be unreliable when signal is present at the spectrum edges.
Reference-based spin quantification assumes that the standard information is valid and that instrumental response differences are controlled or corrected.
The platform does not identify a radical, defect or paramagnetic species solely from a g-value, linewidth or apparent hyperfine spacing.

Frequently Asked Questions

What is EPR spectroscopy?

Electron paramagnetic resonance (EPR), also called electron spin resonance (ESR), detects systems containing unpaired electrons by observing microwave-induced transitions between electron-spin energy levels in a magnetic field.

Can I analyze a normal first-derivative CW-EPR spectrum?

Yes. Select the first-derivative spectrum type. The platform can estimate the dominant resonance field from a zero crossing, report peak-to-peak linewidth, calculate a g-value from the entered microwave frequency and perform numerical double integration.

Can I analyze an absorption EPR spectrum?

Yes. Select the absorption spectrum option. The platform then uses the dominant absorption maximum for the resonance field, reports FWHM linewidth and integrates the absorption signal directly.

What magnetic-field units are supported?

The platform supports gauss (G), millitesla (mT) and tesla (T). Select the unit that matches your uploaded field column.

Why do I need to enter the microwave frequency?

The g-value is calculated from the resonance magnetic field and microwave frequency. Using an approximate frequency when the actual experimental frequency is known reduces the accuracy of the calculated g-value.

Does the tool automatically identify radicals?

No. The platform extracts numerical features but does not assign a chemical or paramagnetic species automatically. Reliable identification can require reference spectra, experimental context and dedicated simulation or fitting.

Is the hyperfine spacing a fitted hyperfine coupling constant?

Not necessarily. ChemFitX reports an apparent spacing between detected resonance features as an inspection aid. It is not a full spin-Hamiltonian fit or tensor determination.

Can ChemFitX calculate spin concentration?

The optional module can estimate spins per gram using a reference standard, integrated intensities, sample and standard masses, and an optional response-correction factor. The result depends on the validity of the standard and comparability of the measurement conditions.

Are my EPR files uploaded to ChemFitX?

For this platform, the supplied code reads and analyzes the spectroscopy file locally in the browser and does not upload the file to ChemFitX servers.

Can I export the processed spectrum?

Yes. Processed numerical data can be exported to CSV, and an Excel workbook can include a summary, processed data and detected features.

Scientific References

Bruker, “EPR 101.” Overview of EPR resonance and field-swept measurements. Source
Bruker, WIN-EPR SimFonia User’s Manual. g-factor determination from microwave frequency and magnetic field. Source
EasySpin documentation. Definitions of peak-to-peak and FWHM linewidth conventions used in EPR simulation. Source
Bruker, ELEXSYS-II E500 / Xepr data-processing materials. Baseline correction, double integration and quantitative EPR workflows. Source