Electro-optic frequency comb Doppler thermometry

TL;DR

This work introduces a Doppler thermometer based on direct optical frequency comb spectroscopy of rubidium vapor using a chirped electro-optic frequency comb (EOFC). By ensuring the comb repetition rate exceeds the excited-state decay rate, transit-induced optical pumping distortion is suppressed, enabling higher optical power and improved signal-to-noise without introducing systematic shifts. Optical Bloch equation simulations and experiments comparing direct EOFC spectroscopy to conventional stepped-scan methods show that EOFC Doppler thermometry can achieve fast averaging with reduced bias, pointing to a compact, fast primary thermometer. The technique holds promise for industrial applications such as pharmaceutical manufacturing and nuclear waste monitoring where small size and high measurement rate are critical.

Abstract

We demonstrate a Doppler thermometer based on direct optical frequency comb spectroscopy of an $^{85}$Rb vapor with a chirped electro-optic frequency comb (EOFC). The direct EOFC Doppler thermometer is accurate to within its approximately 1 K statistical uncertainty. We experimentally compare direct EOFC spectroscopy with conventional Doppler spectroscopy using a single-frequency, step-scanned laser probe. Our results show that direct EOFC spectroscopy mitigates transit-induced optical pumping distortion of the atomic lineshape, which is the dominant systematic temperature shift in alkali atom Doppler thermometry. Optical Bloch equation simulations of conventional and direct EOFC Doppler spectroscopy confirm that EOFC spectroscopy can use higher optical power to reduce statistical noise without optical pumping distortion. Our results indicate that EOFC Doppler thermometry is a promising approach to realizing a primary thermometer with size and measurement rate sufficient for applications including pharmaceutical manufacturing and nuclear waste monitoring.

Figures (4)

• Figure 1: Schematic of the apparatus and measurement scheme: (a) Electro-optic frequency comb (EOFC) generation system and scan probe laser (shown in (a.i)), (b) temperature measurement and EOFC recombination setup, (c) state diagram for the rubidium D$_2$ line, (d) EOFC spectrum with the relevant rubidium hyperfine transitions. Abbreviations -- ADC: analog-to-digital converter, AM: amplitude modulation, AOM: acousto-optic modulator, DAC: digital-to-analog converter, DMM: digital multimeter, EOM: electro-optic modulator, GT: Glan-Thompson polarizer, LO: local oscillator, PD: photodiode, PID: proportional-integral-differential controller, VOA: variable optical attenuator, WP: Wollaston prism. The Glan-Thompson polarizer and Wollaston prism are depicted as beamsplitting cubes.
• Figure 2: Difference between distorted and Lorentzian lineshapes for (a) stepped scan spectroscopy and (b) direct EOFC spectroscopy. The effect of transit-induced distortion is shown for $s=10^{-4}$ (blue, dotted), $s=10^{-3}$ (orange, dashed), $s=10^{-2}$ (teal, dash-dotted), $s=10^{-1}$ (red, dash-dash-dotted), $s=1$ (gray, solid), and $s=10$ (magenta, dash-dot-dotted).
• Figure 3: Example stepped scan and direct EOFC Doppler spectra. (a) Stepped scan transmission spectrum at $s\approx 0.1$ (blue points) with Voigt model fit (orange line). (b) Residuals for the fit and spectrum in (a). (c) Direct EOFC transmission spectrum at $s\approx 0.4$ (blue points) with Voigt model fit (orange line). (d) Residuals for the fit and spectrum in (c). The residuals in (b) and (d) are plotted on different vertical scales. The feature in (b) near $(\omega-\omega_{3,4})/2\pi\approx -1200~MHz$ arises due to a small relative abundance of $^{87}$Rb in the vapor cell (see text).
• Figure 4: Fitted Doppler temperature as a function of probe saturation parameter. Blue circles and red squares show the temperature determined from stepped scan spectra and direct EOFC spectra, respectively. The horizontal dashed line denotes the temperature of the vapor cell reported by the two PRTs. Error bars represent the statistical standard uncertainty. Many error bars are smaller than the data points. The uncertainty in the vapor cell temperature is less than the width of the horizontal dashed line.