Optical Forces Dataset (optical_forces_dataset)

Overview

This DeepTrackAI repository contains simulated optical force data for a microsphere trapped in an optical tweezer, following the work of Bronte Ciriza et al., ACS Photonics, 2022. The simulations were generated using the Optical Tweezers in Geometrical Optics (OTGO) toolbox by Callegari et al., JOSA B, 2015.

Summary

The dataset contains:

• Theoretical optical forces along the z-axis (exact analytical solution).
• Geometrical optics (GO) approximation results with 100 rays.

Experimental parameters

• Laser power (P): 5 mW
• Medium refractive index (nₘ): 1.33 (water)
• Particle refractive index (nₚ): 1.50 (glass)
• Particle radius (R): 1 μm
• Focal length (f): 0.1 mm
• Numerical aperture (NA): 1.3
• Beam waist (w₀): 0.1 mm

---

File Description

• fz_vs_z_theory.txt
  Two columns:

  1. z-position (μm)
  2. Theoretical z-component of the optical force Fz (pN)
     These values serve as the ground truth for comparison with GO results and ML predictions.

• xyz_go_100rays.npy
  NumPy array (4D) containing the particle positions (x, y, z) where optical forces were calculated.

• fxyz_go_100rays.npy
  NumPy array (4D) containing the corresponding force components (Fx, Fy, Fz) at each position in piconewtons.

• sphere_100rays/
  Contains approximately 10⁵ optical force data points simulated via OTGO.

  • Data is split into 101 files: force_grid_3D=1.txt, force_grid_3D=2.txt, …
  • Each file contains rows of eight numbers:

    R  np  x  y  z  fx  fy  fz
    

    where:
    • R: particle radius in meters
    • np: particle refractive index
    • x, y, z: particle position in meters
    • fx, fy, fz: optical force components in newtons

---

Original Source

• Title: Faster and More Accurate Geometrical-Optics Optical Force Calculation Using Neural Networks

• Authors: David Bronte Ciriza, et al.

• Journal: ACS Photonics, 10, 234–241 (2022)

• DOI: 10.1021/acsphotonics.2c01565

• Toolbox: Computational Toolbox for Optical Tweezers in Geometrical Optics (OTGO)

• Authors: Agnese Callegari, et al.

• Journal: Journal of the Optical Society of America B, 32, B11–B19 (2015)

• DOI: 10.1364/JOSAB.32.000B11

If you use this dataset in your research, please follow the licensing requirements and properly attribute the original authors.

---

Dataset Structure

/optical_forces_dataset  
├── fz_vs_z_theory.txt        # Theoretical Fz vs. z-position (z [μm], Fz [pN])
├── xyz_go_100rays.npy        # (4D array) Particle positions [x, y, z]
├── fxyz_go_100rays.npy       # (4D array) Force components [Fx, Fy, Fz] in pN
└── sphere_100rays/           # ~10^5 GO data points for NN training
    ├── force_grid_3D=1.txt
    ├── force_grid_3D=2.txt
    └── ...

---

How to Access the Data

Clone the Repository

git clone https://github.com/DeepTrackAI/optical_forces_dataset.git
cd optical_forces_dataset

---

Usage Example

import numpy as np

# Load theoretical Fz data
z_theory, fz_theory = np.loadtxt("fz_vs_z_theory.txt", unpack=True)

# Load GO simulation data
xyz_go = np.load("xyz_go_100rays.npy")     # positions
fxyz_go = np.load("fxyz_go_100rays.npy")   # forces

# Example: read one of the sphere_100rays files
R, np_val, x, y, z, fx, fy, fz = np.loadtxt("sphere_100rays/force_grid_3D=1.txt", unpack=True)

print("Positions shape:", xyz_go.shape)
print("Forces shape:", fxyz_go.shape)
print("First few Fz theoretical values:", fz_theory[:5])

---

Attribution

Cite the dataset original paper:

Bronte Ciriza D, et al. Faster and More Accurate Geometrical-Optics Optical Force Calculation Using Neural Networks. ACS Photonics, 10: 234–241 (2022). https://doi.org/10.1021/acsphotonics.2c01565

@article{bronte2022faster,
  title={Faster and More Accurate Geometrical-Optics Optical Force Calculation Using Neural Networks},
  author={Bronte Ciriza, David and others},
  journal={ACS Photonics},
  volume={10},
  pages={234--241},
  year={2022},
  publisher={ACS},
  doi={10.1021/acsphotonics.2c01565}
}

Cite the toolbox original paper:

Callegari A, et al. Computational Toolbox for Optical Tweezers in Geometrical Optics. Journal of the Optical Society of America B, 32: B11–B19 (2015). https://doi.org/10.1364/JOSAB.32.000B11

@article{callegari2015comp,
  title={Computational Toolbox for Optical Tweezers in Geometrical Optics},
  author={Callegari, Agnese and others},
  journal={JOSA B},
  volume={32},
  pages={B11--B19},
  year={2015},
  publisher={Optica},
  doi={10.1364/JOSAB.32.000B11}
}

---

License

This repository is released under the MIT License.
You are free to use, modify, and distribute this work, provided that you include the original license notice and attribution.