General theory for super-sensitive dual-wavelength phase metrology: error-free unwrapping and signal-to-noise ratio

TL;DR

This paper develops a comprehensive mathematical framework for dual-wavelength phase metrology, enabling error-free phase unwrapping and SNR analysis, thereby enhancing the accuracy and sensitivity of optical measurements.

Contribution

It introduces the first detailed mathematical analysis and formulas for SNR and error-free unwrapping in dual-wavelength phase metrology.

Findings

Provides formulas for SNR of phase-difference and phase-sum

Demonstrates error-free unwrapping of noisy phase-sum from phase-difference

Achieves super-sensitive height measurements in profilometry

Abstract

From 1971 to 2012 dual-wavelength optical-metrology used only the demodulated low-sensitivity phase-difference of two close-sensitive fringes. Dual-wavelength phase-metrology that additionally uses the phase-sum was first reported by Di et al. in 2013 [28]; this was an important step to increase the phase-accuracy in optical metrology. This and its derived papers however do not offer mathematical analysis for signal-to-noise ratio (SNR) for the phase-difference and phase-sum. Neither provide the mathematical analysis for unwrapping the phase-sum without errors. Here a general theory for super-sensitive two-wavelength phase-metrology is given. In particular mathematical analysis and formulas for SNR and error-free phase-unwrapping for two-wavelength metrology is provided. We start by phase-demodulating two close-sensitivity fringes by phase-shifting algorithms (PSAs). We then calculate their phase-difference and their phase-sum; the phase-difference is assumed non-wrapped. However the phase-sum is highly wrapped, super-sensitive and has much higher SNR. Spatial phase unwrapping for a highly discontinuous phase-sum is precluded. However as we show, it is possible to unwrap the noisy phase-sum from the noisier phase-difference without errors. We apply this super-sensitive phase-metrology theory to profilometry allowing us to obtain super-sensitive height measurements. To the best of our knowledge the mathematical analysis and formulas herein presented for the SNR and error-free unwrapping have not been reported before.