Physiscore

An integrated, physics-constrained analysis suite for thin-film photovoltaic device and material characterization.

Physiscore unifies nine measurement techniques behind one interface and one set of conventions, so that every parameter reported for a device is derived from the raw measured curve rather than from an instrument summary block — and so that quantities measurable by more than one technique (e.g. the optical band gap) can be cross-checked within a single workflow.

Technique	Derived quantities
J–V (light)	Voc, Jsc, FF, PCE, Vmpp, Jmpp, Pmax, Rs, Rsh — all from the curve
J–V (dark)	ideality factor n, saturation current J₀, dark Rsh
EQE	AM1.5G-integrated Jsc (ASTM G173-03, 2002 points), Eg, Urbach energy
Abs	Tauc-plot band gap with automatic linear-region selection
XRD	ALS background, simultaneous pseudo-Voigt fitting of overlapping reflections, d-spacing, Scherrer size, Williamson–Hall
PL	ALS baseline, windowed Gaussian/Voigt fitting, FWHM, quenching efficiency
Raman / FTIR	multi-layer loading, peak detection, peak-shift comparison
SEM	watershed grain segmentation, size distribution statistics
XPS	singlet / spin-orbit-doublet Voigt fitting, atomic concentrations

Why another analysis tool

Device parameters are routinely copied from the summary block a source-meter writes into its export file. Those numbers depend on the instrument's own fitting settings, are not reproducible across labs, and silently hide problems in the underlying sweep. Physiscore recomputes every quantity from the measured points, keeps the derivation explicit, and exports both its own values and the instrument's for cross-validation.

Installation

Requires Python 3.10 or newer.

git clone https://github.com/jongsu-king/physiscore.git
cd physiscore
python -m venv .venv && source .venv/bin/activate   # Windows: .venv\Scripts\activate
pip install -r requirements.txt

Quick start

Launch the unified interface:

python Physiscore_Launcher.py

Or use the analysis cores directly as a library — they are GUI-free and importable:

import numpy as np
from scripts.JV.jv_core import compute_jv_parameters

v = np.linspace(-0.2, 1.1, 261)
j = ...                      # measured current density [mA/cm^2]
p = compute_jv_parameters(v, j, pin_mw_cm2=100.0)
print(p["Voc"], p["Jsc"], p["FF"], p["PCE"], p["Rs"], p["Rsh"])

from scripts.Dark_JV.dark_jv_core import analyze_dark_diode
d = analyze_dark_diode(v_dark, j_dark_abs)   # -> n, J0_mA_cm2, Rsh_dark_Ohm_cm2, fit_R2

from scripts.XRD.xrd_core import detect_peaks_and_fit
peaks_df, y_fit = detect_peaks_and_fit(two_theta, intensity,
                                       instrumental_fwhm_deg=0.05, scherrer_k=0.9)

Configuration

Copy the templates and edit them locally — the real files are git-ignored because they hold machine-specific paths and an optional API key:

cp physisfera_config.example.json physisfera_config.json

cell_area_cm2 (default 0.04) sets the device area used for current-density conversion; it is re-read for every analysis, so a correction applies to the next folder you process. The AI-report feature is optional: with no API key configured, every physics computation and export path behaves identically.

Example data

Measured example files (a perovskite XRD pattern, XPS region scans for a glass reference and a PEI-coated sample) live in a local rawdata/ directory that is not published here — those measurements are unpublished. The code does not require them: the ASTM G173-03 reference spectrum used for J–V and EQE work is bundled in the repository, and the test suite skips its measured-file checks when rawdata/ is absent.

Tests

pip install pytest
pytest -v

The suite encodes analytical regression checks against closed-form synthetic data with known ground truth (single-diode J–V curves, overlapping pseudo-Voigt reflections, Gaussian PL peaks) and against measured reference files.

Citing

See CITATION.cff. A software paper is in preparation; please cite the archived release DOI in the meantime.

License

MIT — see LICENSE.

Copyright © 2026 Hybrid Materials and Optoelectronics Laboratory (H-MOL), Incheon National University.