Optical remote sensing of water temperature using Raman spectroscopy

Abstract

A detailed investigation into the use of Raman spectroscopy for determining water temperature is presented. The temperature dependence of unpolarized Raman spectra is evaluated numerically, and methods based on linear regression are used to determine the accuracy with which temperature can be obtained from Raman spectra. These methods were also used to inform the design and predict the performance of a two-channel Raman spectrometer, which can predict the temperature of mains supply water to an accuracy of ± 0.5 °C.

Remote Sensing of Natural Waters Using a Multichannel, Lidar-Compatible Raman Spectrometer and Blue Excitation

“…The authors of Artlett and Pask (2015) evaluated the trade-offs between spectral channels and sensitivities by performing simulations with unpolarized Raman signals acquired from ultrapure (Reverse-Osmosis) water samples. The mean-scaled markers sensitivities calculated from two spectral channels of 250 cm −1 width exhibited values around 0.68%/ • C; and sensitivities for channels widths around 150 cm −1 were estimated to be around 1.03%/ • C. Considering that the spectral channels used in our work had widths of 234 cm −1 and 137 cm −1 at the FWHM, the sensitivities found for both depolarisation(A) and two-color(||) markers were reasonably in agreement with the values proposed in Artlett and Pask (2015).…”

“…In Milli-Q water, the measured sensitivities for the various markers reflect this dependence, plus the positions and widths of the spectral channels. According to simulations performed by Artlett and Pask (2015) for ultrapure (Reverse-Osmosis) water, an optimum trade-off between Raman signals strength and RMSTEs would be obtained for acquisition channels with spectral widths of around 200 cm −1 . Optimum spectral positions for such channels were explored using simulations in Artlett and Pask (2017), with the "low shift" channel central position at 3,200 cm −1 and the "high shift" channel central position at 3,600 cm −1 .…”

“…When trying to conduct the same analysis for temperature predictions in natural waters, we found substantially lower accuracies, which we attributed to the overlapping of the Raman peak for 532 nm excitation and fluorescence signals (de Lima Ribeiro et al, 2019a). The commercial dispersive Raman spectrometer (RS) used in Artlett and Pask (2015) and de Lima Ribeiro et al (2019a) did not fulfill LIDARcompatibility requirements and, in order to transition from commercial equipment toward LIDAR-compatible technologies, we designed and assembled a LIDAR-compatible multichannel RS integrated to a 532 nm pulsed excitation laser (de Lima Ribeiro et al, 2019b). The equipment allowed for simultaneous Raman signal collection in four spectral channels, and twocolor and depolarisation markers were estimated for ultrapure (Milli-Q) and natural water sample, achieving best accuracies of ±0.3 • C in both cases.…”

“…These markers can be calculated from Raman signals having the same polarization state and are known as "two-color" ratios, or from the number of photons having perpendicular/parallel polarization, referred to as "depolarisation" ratios. These ratios form the basis for numerous studies undertaken from the 1970s until the present time, aimed at using Raman spectroscopy to remotely determine water temperature (Chang and Young, 1972;Leonard et al, 1979;Leonard and Caputo, 1983;Artlett and Pask, 2015). When used in combination with LIDAR methods, there exists great potential to obtain depth-resolved measurements of subsurface water temperature.…”

“…In (Artlett and Pask, 2015) accuracies of ±0.1 • C were reported for water temperature measurements performed in the laboratory using a commercial dispersive Raman spectrometer (Enwave-EZRaman I), incorporating a 532 nm, continuous wave, excitation laser. That work utilized unpolarized Raman spectra, two-color markers, and Reverse-Osmosis laboratory water.…”