THzTools is an open-source Python package for data analysis in terahertz time-domain spectroscopy (THz-TDS). As the name suggests, THz-TDS involves measurements of terahertz-frequency electromagnetic waveforms that are acquired as a function of time, which users typically represent as a function of frequency for analysis. THzTools makes it easier for researchers to use statistically optimal methods for doing this analysis, as described in L. Mohtashemi et al., Opt. Express 29, 4912 (2021).

This is beta software that is currently under development.

Information	Links
Project
Build
Documentation	https://dodge-research-group.github.io/thztools/
Cite	L. Mohtashemi et al., Opt. Express 29, 4912 (2021). (DOI)

The original MATLAB code is available at Zenodo.

You can install THzTools with pip:

pip install thztools

THzTools is also available through conda-forge:

conda install -c conda-forge thztools

See the Getting started tutorial for additional information.

In the conventional approach to THz-TDS analysis, one transforms the time-domain measurements into the frequency domain for further analysis. This approach has well-known problems, however, which can be resolved by using a maximum-likelihood estimation procedure in the time domain. To support this mode of analysis, the THzTools package provides functionality and documentation that are unavailable in existing THz-TDS analysis software. It provides functions to simulate THz-TDS measurements (eg, timebase and wave), apply a frequency response function to a THz-TDS waveform (apply_frf), characterize the noise of a THz-TDS system (noisefit), and fit a parameterized frequency response function to a pair of input and output waveforms (fit).

With THzTools, you can:

• Simulate a terahertz waveform
• Simulate realistic time-domain noise for a waveform
• Apply a frequency response function (FRF) to a time-domain waveform
• Rescale and shift a waveform
• Estimate time-domain system noise parameters from a set of measured waveforms
• Fit a parameterized FRF to a pair of time-domain waveforms

For example, the following code creates an ideal waveform and applies a frequency response function to it.

import numpy as np
import thztools as thz

# Set the waveform parameters
n = 256  # Number of samples
dt = 0.05  # Sampling time [ps]
a = 0.5 # Scale factor
eta = 1.0 # Delay [ps]

# Simulate the waveform
t = thz.timebase(n, dt=dt)
mu = thz.wave(n, dt=dt)

# Define a frequency response function
def frfun(omega, _a, _eta):
    return _a * np.exp(-1j * omega * _eta)

# Apply the frequency response function to the waveform
psi = thz.apply_frf(frfun, mu, dt=dt, args=(a, eta))

If you use THzTools, please consider citing the paper and/or the software package.

@article{Mohtashemi:21,
author = {Laleh Mohtashemi and Paul Westlund and Derek G. Sahota and 
Graham B. Lea and Ian Bushfield and Payam Mousavi and J. Steven Dodge},
journal = {Opt. Express},
number = {4},
pages = {4912--4926},
publisher = {Optica Publishing Group},
title = {Maximum-likelihood parameter estimation in terahertz time-domain 
spectroscopy},
volume = {29},
month = {Feb},
year = {2021},
url = {https://opg.optica.org/oe/abstract.cfm?URI=oe-29-4-4912},
doi = {10.1364/OE.417724},
}

@software{Posada_Loaiza_THzTools_2024,
author = {Posada Loaiza, Jonathan and Higuera-Quintero, Santiago and Noori, 
Alireza and Mohtashemi, Laleh and Hall, R. P. and Yimam, Naod Ayalew and 
Dodge, J. Steven},
license = {MIT},
month = oct,
title = {{THzTools}},
url = {https://github.com/dodge-research-group/thztools},
version = {0.5.3},
year = {2024}
}