Corrective Re-gridding Techniques for Non-Uniform Sampling in Time Domain Terahertz Spectroscopy

TL;DR

This paper evaluates corrective re-gridding algorithms, specifically cubic spline and Shannon methods, for improving signal quality in time domain terahertz spectroscopy affected by non-uniform sampling.

Contribution

It introduces and experimentally validates re-gridding techniques to mitigate sampling noise, enhancing signal-to-noise ratio in THz spectroscopy.

Findings

Re-gridding algorithms improve SNR from 100 GHz to 2 THz.

Experimental results confirm effectiveness of cubic spline and Shannon re-gridding.

Re-gridding becomes more critical as THz technology advances.

Abstract

Time domain terahertz spectroscopy typically uses mechanical delay stages that inherently suffer from non-uniform sampling positions. We review, simulate, and experimentally test the ability of corrective cubic spline and Shannon re-gridding algorithms to mitigate the inherent sampling position noise. We present simulations and experimental results that show re-gridding algorithms can increase the signal to noise ratio within the frequency range of 100 GHz to 2 THz. We also predict that re-gridding corrections will become increasingly important to both spectroscopy and imaging as THz technology continues to improve and higher frequencies become experimentally accessible.