Grating Characterization Toolkit

This repository contains Python tools developed during my 2026 research internship for the spectral characterization of optical gratings.

It includes:

• Reflection spectrum analysis
• Transmission spectrum analysis
• Transfer Matrix Method (TMM) simulation for grating modeling

The project is still evolving and may expand with additional analysis tools.

---

Table of Contents

1. Required Input Files
2. Recommended Repository Structure
3. File Naming Convention
4. reflection_analysis.py
5. transmission_analysis.py
6. tmm.py
7. Requirements
8. Notes

---

1. Required Input Files

All analysis scripts expect .csv files exported directly from an Optical Spectrum Analyzer (OSA).

The parser is designed for the standard OSA format containing a [TRACE DATA] section with traces labeled A–G.

---

2. Recommended Repository Structure

Grating-research/
├── reflection_analysis.py
├── transmission_analysis.py
├── tmm.py
├── data/
│ ├── reflection/
│ ├── transmission/
│ └── Example/
│ ├── comparison/
│ ├── cross/
│ └── full/
└── results/
├── reflection/
├── transmission/
└── tmm/

All scripts use relative paths, so the toolkit works on any machine without modification.

---

3. File Naming Convention

Measurement files follow this pattern:

SAMPLENAME_DD-MM-YY.csv

Examples:

BOX2_IN_OUT_10-06-26.csv
BOX1_0.050_10-06-26.csv

Why this convention?

• The sample name appears first for quick identification.
• The date at the end keeps files chronologically sorted.
• Scripts automatically extract the date for labeling plots and outputs.

If you use a different naming scheme, simply update the file name variables (FILE, FILE_1, etc.).
The scripts do not depend on the date format.

About the “box” terminology

In the original experiment, gratings were grouped into physical containers (“boxes”), and each OSA file contained multiple traces (A–G) corresponding to gratings in that box.

This is only a naming convention.
You can replace it with:

• Sample1
• Setup_A
• Test01
• anything that fits your workflow

The scripts only rely on the file name you provide.

---

4. reflection_analysis.py

Analyzes reflection spectra from OSA measurements.

Modes

Mode	Description
single	Analyze one file using a source and a reflected trace
compare	Compare two files side-by-side
cross	Overlay selected traces across multiple files
box	Compare all traces in one file against a reference trace
full	Plot the full spectrum of a file (all traces)
single-box	Compute reflection for each trace in a file against a shared source trace

Set the mode at the top of the script:

MODE = "single"   # or "compare", "cross", "box", "full", "single-box"

---

Mode Details

single

Computes the reflection spectrum from one source trace and one reflected trace in a single file.

Configure:

FILE            = DATA_DIR / "SAMPLE1_0.050_10-06-26.csv"
TRACE_SOURCE    = "A"
TRACE_REFLECTED = "E"

COMPARE_TRACES  = True    # set False to skip
TRACE_COMPARE_1 = "D"
TRACE_COMPARE_2 = "E"

Outputs:

• Full spectrum plot (all traces)
• Source vs reflected power plot + reflection curve
• (Optional) Side-by-side trace comparison

---

compare

Compares reflection spectra from two separate files.

Configure:

FILE_1            = DATA_DIR / "SAMPLE1_08-06-26.csv"
TRACE_SOURCE_1    = "A"
TRACE_REFLECTED_1 = "D"

FILE_2            = DATA_DIR / "SAMPLE2_08-06-26.csv"
TRACE_SOURCE_2    = "A"
TRACE_REFLECTED_2 = "E"

Outputs:

• Full spectrum plot for each file
• 3×2 side-by-side comparison figure (source, reflected, and overlay for each file)
• CSV for each file

Typical use cases:

• Same grating measured at different OSA resolutions
• Measurements taken at different times
• Forward vs backward coupling directions
• Gratings from different sample groups

---

cross

Overlays up to 4 individual traces drawn from any combination of files. Useful for directly comparing specific gratings across measurements.

Configure:

FILE_1     = "SAMPLE1_IN_OUT_10-06-26"
FILE_2     = "SAMPLE1_OUT_IN_10-06-26"
FIBER_TYPE = "GratingName"

CROSS_TRACES = [
    (DATA_DIR / f"{FILE_1}.csv", "B", "GratingName_0.05 - in->out"),
    (DATA_DIR / f"{FILE_2}.csv", "G", "GratingName_0.05 - out->in"),
    (DATA_DIR / f"{FILE_1}.csv", "C", "GratingName_0.5 - in->out"),
    (DATA_DIR / f"{FILE_2}.csv", "F", "GratingName_0.5 - out->in"),
]

Each entry in CROSS_TRACES is a tuple: (file_path, trace_letter, label).

Labels containing "in->out" or "out->in" are automatically grouped into separate overlay figures.

Outputs:

• in->out overlay plot
• out->in overlay plot
• Isolated subplots for each trace (split across two figures if more than 2 traces)

---

box

Computes the reflection of every listed trace against a single reference trace, all from the same file. Produces an individual plot and CSV per trace.

Configure:

BOX_FILE            = DATA_DIR / "SAMPLE1_OUT_IN_10-06-26.csv"
BOX_REFERENCE_TRACE = "A"

BOX_TRACES = [
    ("B", "Grating_1-0.050"),
    ("C", "Grating_1-0.500"),
    ("D", "Grating_2-0.050"),
    ("E", "Grating_2-0.500"),
    ("F", "Grating_3-0.050"),
    ("G", "Grating_3-0.500"),
]

Each entry is (trace_letter, label).

Outputs (per trace):

• Reference vs trace power plot + reflection curve
• CSV with reflection and return loss values
• Console summary (min / max / mean)

---

full

Plots every trace in a file on a single figure. Useful for a quick visual overview of a measurement.

Configure:

FULL_FILE    = DATA_DIR / "SAMPLE1_OUT_IN_10-06-26.csv"
FULL_EXCLUDE = []    # e.g. ["F", "G"] to hide specific traces

Output:

• Single figure with all traces overlaid

---

single-box

Similar to box, but tailored for a single measurement session: computes the reflection of each listed trace against a shared source trace, generating one plot per grating.

Configure:

SINGLE_BOX_FILE   = DATA_DIR / "SAMPLE1_OUT_IN_10-06-26.csv"
SINGLE_BOX_SOURCE = "A"

SINGLE_BOX_FIBERS = [
    ("B", "Grating_1_0.050"),
    ("C", "Grating_1_0.500"),
    ("D", "Grating_2_0.050"),
    ("E", "Grating_2_0.500"),
    ("F", "Grating_3_0.050"),
    ("G", "Grating_3_0.500"),
]

Each entry is (trace_letter, grating_name).

Outputs (per grating):

• Source vs grating power (top panel)
• Reflection curve in dB (bottom panel)

---

Output Locations

Results are saved under:

results/reflection/<mode>/<resolution_tag>/<file_tag>/

For example:

results/reflection/box/RES_TEST/SAMPLE1_OUT_IN_10-06-26/
├── Grating_1-0.050/
│   ├── Grating_1-0.050_reflection.csv
│   └── Grating_1-0.050_reflection.png
└── Grating_1-0.500/
    ├── Grating_1-0.500_reflection.csv
    └── Grating_1-0.500_reflection.png

---

Computed Values

For all modes that compute reflection, the output CSV contains:

Column	Description
wavelength_nm	Wavelength axis (nm)
source_dBm	Source power (dBm)
reflected_dBm	Reflected power interpolated to source grid (dBm)
reflection_dB	Reflected − Source (dB)
return_loss_dB	−(Reflected − Source) (dB)

---

5. transmission_analysis.py

Analyzes transmission spectra and computes insertion loss.

Outputs

• Full spectrum plot
• Source vs transmitted power comparison
• Transmission and insertion loss curves
• CSV file with computed values
• Console summary (min / max / mean transmission)

Example output location:

results/transmission/Example/
├── transmission_SAMPLE1_DATE.csv
└── transmission_SAMPLE1_DATE.png

---

6. tmm.py

Simulates theoretical grating spectra using the Transfer Matrix Method.

Features

• Reflection spectrum (linear and dB)
• Transmission spectrum (dB)
• Wavelength shift vs applied strain
• CSV export of simulated spectra
• CSV export of peak wavelength vs strain
• Sensitivity calculation

Results are saved in:

results/tmm/

---

7. Requirements

• Python 3.10+
• numpy
• pandas
• matplotlib

Install dependencies:

pip install -r requirements.txt

---

8. Notes

• All scripts use relative paths for portability across machines.
• Input files must be OSA-exported .csv files with a [TRACE DATA] section.
• The file parser caches loaded files to avoid re-reading the same file multiple times within a session.
• Naming conventions are flexible — adjust file name variables and labels to match your own workflow.
• The project is actively evolving; additional modes and analysis tools may be added.