📚 EPR AI Analysis Assistant - User Guide

Supports various EPR data formats, including standard text formats and instrument raw data.

• Supported Formats: .txt, .csv, .asc, .dat, .epr, .eprp, .spc, .xlsx
• Bruker Data: Supports uploading .DSC and .DTA files simultaneously (please select both files).
• Microwave Frequency: If not included in the file, you will be prompted to enter it manually (common X-band is ~9.8 GHz).

1.1 Text-like Formats

• .txt / .csv / .asc：
  • Recommended to have two columns of numbers: first column for magnetic field (G or mT), second for intensity.
  • Columns can be separated by spaces, Tabs, or commas. Ensure consistency within the file.
  • First row can be headers (e.g. Field,Intensity) or data; lines starting with # are ignored as comments.
  • Example (space separated):
    # B(G) Intensity
3300 0.0023
3301 0.0028
• .xlsx：
  • Uses the first sheet;
  • First two columns are Field and Intensity (e.g., Col A = Field / Col B = Intensity);
  • Do not merge cells or insert complex headers in the data region. Extra text above the data is fine.

1.2 Instrument Raw Formats

• .dat / .spc： Directly upload raw spectral files exported from instruments. No formatting needed.
• .epr / .eprp： Software-specific formats containing parameters and data. Upload directly.
• .DSC + .DTA： For Bruker EPR data, select both files to upload together. The system parses them automatically.

The system provides three analysis modes for different experimental scenarios:

New Feature: After loading data, you don't have to wait for AI analysis. Click the 🎯 Manual Peak button next to the delete button to enter manual mode instantly.

1. Click 🎯 Manual Peak to enter peak selection mode (button changes to ✓).
2. Click on peak positions in the spectrum area; the system will mark them with red dots.
3. It is recommended to mark at least 2 main peaks (or all obvious peaks for complex spectra).
4. After marking, click 🚀 AI Analysis at the top. The algorithm will force the fit to pass through your marked peaks.
5. Click the manual peak button again to exit the mode; click "Clear Peaks" to reset marks.

After analysis, the interface displays comprehensive results:

• Fitted Spectrum: Red dashed line is the fit, black solid line is the experiment. Zoom and pan to inspect details.
• Parameter List: The left table shows the Top 5 matches. Click different rows to view fits for different candidates.
• Multi-component Details: In multi-component mode, click a result to see parameters (g-value, hyperfine coupling A, linewidth) for each component below.

Click "Generate Detailed AI Report" to let the large language model create a professional analysis report.

• Experimental Conditions: Click "Experiment Settings" before generating to add spin trap and solvent info for a more accurate report.
• Smart Q&A: After report generation, you can continue asking questions about it in the chat box below.

Generate publication-ready figures and captions with one click.

• Format: Generates Nature-style figures by default (high resolution). PNG download supported.
• Caption: Automatically generates English/Chinese figure captions containing experimental conditions and fit parameters. Text download supported.
• WYSIWYG: Generated based on the selected fit result. Click "Download Details" to get full Figure Description Markdown.

Example: Analyzing DMPO-OH: