ThinOptix

ThinOptix is a Python workflow for modelling and optimising low-emissivity (Low-E) coated glass stacks. The project evaluates spectral transmission and reflectance for a multilayer stack using a transfer-matrix approach, then computes visible performance metrics such as photopic transmittance and CIE Lab colour coordinates.

The core stack model is:

air | L3 | L2 (Ag) | L1 | glass | air

The project is used to:

• compute T / F / G spectra for a coated-glass design
• search for thickness combinations that meet transmission and colour constraints
• compare different silver layer thicknesses and material datasets
• generate plots and summary CSV files for candidate designs

---

Project purpose

This repository is built around optical coating design rather than a general-purpose app. It is intended for research and engineering exploration of low-E coating stacks where the design goal is often:

• maximise visible transmission (Tvis)
• keep visible reflectance on both sides under limits
• maintain colour coordinates within acceptable ranges
• evaluate multiple Ag thicknesses and material candidates

The code implements a thin-film optical model and uses numerical search over thickness combinations to identify feasible designs.

---

Main files

Root scripts

• lowe_design.py — primary exhaustive grid search for low-E coating designs.

  • Reads material files and experiment files.
  • Evaluates every L1/L3 thickness combination for selected Ag thicknesses.
  • Applies design constraints and saves passing designs.
  • Generates design-space and 3D visualisation plots.

• optical_coating_tool.py — general optical coating utility and optimizer.

  • Computes T/F/G spectra for a single stack.
  • Converts spectra to CIE Lab values.
  • Implements thickness optimization for L1/L3 with Ag fixed at 15 nm.
  • Includes optional IGU / NFRC-style performance calculations.

Scripts folder

• scripts/thickness_sweep.py — sweeps one layer thickness at a time to study sensitivity.
• scripts/calculate_coated_glass_TFG_V2_percent.py — older or alternative coated-glass calculation script.

Colour analysis utilities

• color_analysis/colorLAB_TFG_csv_v4_table_output.py — processes TFG/Lab CSV outputs to build table-based summaries and plots.

---

Repository structure

ThinOptix/
├── README.md
├── lowe_design.py
├── optical_coating_tool.py
├── data/
│   ├── experiments/
│   └── materials/
├── output/
│   ├── csv/
│   └── png/
│       ├── 3d_surfaces/
│       ├── color_patches/
│       └── design_space/
├── scripts/
│   ├── calculate_coated_glass_TFG_V2_percent.py
│   ├── thickness_sweep.py
│   └── ...
├── color_analysis/
│   ├── colorLAB_TFG_csv_v4_table_output.py
│   ├── *.csv
│   └── ...
└── ...

---

Data format

Material files

Material CSVs live in data/materials/ and should contain at least:

• wavelength column (or NK)
• n column
• k column

Examples include files such as:

• glass-industr-2018-mat.csv
• L1-Si3N4-mat.csv
• L2-Ag-mat.csv
• L3-top-ZrO2-mat.csv

The loader code accepts wavelength values in either nm or microns and converts if needed.

Experiment files

Experiment CSVs in data/experiments/ represent optical data for the Ag layer or similar material variants used in the grid-search workflow. These files are loaded by lowe_design.py and used as the L2 layer in the search.

---

Setup

This project uses Python 3 and the following packages:

pip install numpy pandas scipy matplotlib colour

A typical virtual environment setup is:

python3 -m venv .venv
source .venv/bin/activate
pip install -U pip
pip install numpy pandas scipy matplotlib colour

---

Typical workflows

1) Run the full Low-E design search

From the project root:

python lowe_design.py

This script:

• loads material files
• evaluates a 1 nm thickness grid for L1 and L3
• evaluates Ag thicknesses at 10, 12.5 and 15 nm
• checks the optical constraints
• writes passing designs to output/csv/
• produces design-space and colour patch plots in output/png/

Expected output files include patterns such as:

• output/csv/lowe_Ag10p0nm_passing.csv
• output/csv/best_designs_all_experiments.csv
• output/png/design_space/*.png
• output/png/color_patches/*.png

2) Run the interactive optical coating tool

python optical_coating_tool.py

This script is more interactive and can:

• evaluate one specific stack (d1, d2, d3)
• compute spectral T/F/G data and CIE Lab values
• optimise d1 and d3 for a fixed Ag layer thickness
• optionally run an IGU / NFRC-style performance estimate

3) Sweep a single layer thickness

python scripts/thickness_sweep.py

This is useful when you want to see how changing one layer affects T/F/G colour and Tvis while holding the other layers fixed.

4) Compute a coated glass spectrum directly

python scripts/calculate_coated_glass_TFG_V2_percent.py

This older script asks for file names and thicknesses interactively and writes a spectral CSV containing percent transmission and reflection values.

---

What the model calculates

The core calculations model a stack normal-incidence transfer matrix with a thick glass substrate treated incoherently to avoid unrealistic fringe effects from the substrate thickness.

For each wavelength:

• T = transmitted fraction through the coated glass
• F = film-side reflectance
• G = glass-side reflectance

Those spectra are then integrated against CIE photopic weighting to estimate visible performance metrics such as:

• Tvis
• Y, L*, a*, b* in CIE Lab space

Design checks are applied using the thresholds defined in lowe_design.py, including visible reflectance limits and colour-coordinate constraints.

---

Output files

CSV outputs

The project writes output CSV files in output/csv/, including:

• passing designs from each experiment and Ag thickness
• best design summary across experiments
• spectral T/F/G results
• Lab colour summaries

Plot outputs

The project writes PNG files under output/png/:

• 3d_surfaces/ for 3D Tvis and colour surfaces
• color_patches/ for candidate design colour patches
• design_space/ for 2D thickness-search maps

---

Notes and practical guidance

• The project is highly parameterised, so the main knobs are material CSVs, the Ag thickness list, and the search ranges defined near the top of lowe_design.py.
• For experiments and material comparisons, the workflow is easiest to use when all input data are placed under data/materials/ and data/experiments/ with consistent wavelength and optical property formatting.
• If you are exploring a new coating, start with optical_coating_tool.py for a single-stack run, then switch to lowe_design.py for the exhaustive design sweep.

---

Summary

ThinOptix is a research-oriented optical design toolkit for low-E coatings. It combines thin-film transfer-matrix physics, photometric colour calculations, and optimization/search routines to identify coating stacks that satisfy visible-light transmission and reflectance requirements.