Graph_PE: An Advanced Standalone Application for Preprocessing, Multi-Component Peak Fitting, and Quantitative Defect/Strain Analysis of PECVD Graphene Raman Spectra (Version 2.0 Enterprise Edition)
Abstract
Graph_PE (PECVD Graphene Raman Analyzer) is an advanced, publication-grade scientific desktop application engineered for the automated preprocessing, multi-component peak deconvolution, and quantitative structural evaluation of Raman spectra from Plasma-Enhanced Chemical Vapor Deposition (PECVD) graphene, functionalized carbon nanostructures, and 2D materials. Developed by dr. Algirdas Lazauskas at Kaunas University of Technology (KTU), the software addresses the complex spectroscopy of disordered and nanostructured carbon, where pronounced defect activation, severe G and D′ band overlap, turbostratic 2D band broadening, and amorphous carbon contributions hinder conventional single-peak fitting routines. Distributed as a standalone, zero-dependency compiled Nuitka executable for Windows, Graph_PE integrates a comprehensive set of automated analytical capabilities: Universal Spectroscopic File Ingestion: Features an internal binary parser for Galactic and Thermo SPC files (supporting both legacy LabCalc and modern formats, 16-bit and 32-bit fixed-point encodings, IEEE floating-point representations, and multi-subfile spectra) without external binary dependencies; native Renishaw WiRE WDF binary reading; universal delimiter auto-detection for text, CSV, DAT, XY, ASC, DPT, and RSI files; automatic laser wavelength sniffing from metadata; and automated Raman axis reconstruction heuristics for corrupted data. Robust Preprocessing Pipeline: Includes cosmic-ray despiking via modified Z-score and median absolute deviation filtering; Savitzky–Golay and Gaussian smoothing; and penalized least-squares baseline removal featuring adaptive iteratively reweighted penalized least squares (airPLS) with exponential overflow protection, asymmetric least squares (AsLS), anchor-guided polynomial fitting, rolling-ball morphological opening, and convex-hull rubberband baseline subtraction. Multi-Stage Peak Detection & Band Classification: Employs a four-stage adaptive cascade algorithm combining multi-prominence continuous sweeps, region-specific first-order and second-order scans, window-maxima fallbacks, and a specialized D′ shoulder rescue routine to classify all major graphene features (D*, D, D′′, G, D′, D+D′′, 2D, D+D′, and 2D′). Nonlinear Peak Fitting & Lineshape Library: Implements six distinct lineshape profiles—Lorentzian, Gaussian, Faddeeva-based Voigt, Pseudo-Voigt, Pearson VII, and asymmetric Breit–Wigner–Fano (BWF) for metallic or heavily doped graphene—coupled with five optimization algorithms (Levenberg–Marquardt, Trust-Region Reflective, Nelder–Mead, Differential Evolution, and Dual Annealing), physical parameter bounding, center-locking controls, and trapezoidal numerical area integration. Quantitative Defect, Strain, and Doping Characterization: Automatically computes crystallite size, point-defect density, and inter-defect distance according to Cançado and Lucchese relations; identifies defect nature (sp³-type, vacancy-type, or grain boundary edges) via the Eckmann D-to-D′ intensity ratio; estimates layer thickness and turbostratic disorder from 2D-to-G ratios and 2D line widths; and performs Lee vector decomposition to optically separate biaxial mechanical strain from charge-carrier doping. Regime Intelligence & Error Propagation: Features a smart Tuinstra–Koenig regime classifier that alerts users when transitioning into amorphous Stage 2 disorder where standard crystallite scaling relations lose physical validity; provides a composite 0-to-100 defect severity score; and executes rigorous Delta-method error propagation from the fit covariance matrix to assign definitive standard errors to all derived physical metrics. Interactive Graphical Interface & Batch Workflows: Built on a high-performance, 60-frames-per-second PySide6 and pyqtgraph canvas with synchronized spectral and residual plots, draggable marker tags, interactive peak insertion, undo/redo session tracking, and a One-Click Auto PECVD Fit workflow. High-Throughput Mapping, Diagnostics & Reporting: Includes multi-threaded background workers for unattended directory batch processing; 2D Raman hyperspectral spatial mapping with heatmap generation; multi-spectrum growth series overlays; parameter correlation and covariance diagnostics; and export options encompassing publication-ready vector (SVG, PDF, EPS) and high-DPI raster graphics, full-traceability raw-to-fit CSV tables, complete JSON session files, and standardized three-page technical PDF summary reports.