A comprehensive investigation into the impact of spectral baseline on temperature prediction in natural marine water samples by Raman spectroscopy is presented. The origin of baseline signals is investigated using principal component analysis and phytoplankton cultures in laboratory experiments. Results indicate that fluorescence from photosynthetic pigments and dissolved organic matter may overlap with the Raman peak for 532 nm excitation and compromise the accuracy of temperature predictions. Two methods of spectral baseline correction in natural waters are evaluated: a traditional tilted baseline correction and a new correction by temperature marker values, with accuracies as high as ± 0.2°C being achieved in both cases.

The design and operation of a custom-built LIDAR-compatible, four-channel Raman spectrometer integrated to a 532 nm pulsed laser is presented. The multichannel design allowed for simultaneous collection of Raman photons at two spectral regions identified as highly sensitive to changes in water temperature. For each of these spectral bands, the signals having polarization parallel to (∥) and perpendicular to (⟂), the excitation polarization were collected. Four independent temperature markers were calculated from the Raman signals: two-colour(∥), two-colour(⟂), depolarization(A) and depolarization(B). A total of sixteen datasets were analysed for one ultrapure (Milli-Q) and three samples of natural water. Temperature accuracies of ±0.4 °C–±0.8 °C were achieved using the two-colour(∥) marker. When multiple linear regression models were constructed (linear combination) utilizing all simultaneously acquired temperature markers, improved accuracies of ±0.3 °C–±0.7 °C were achieved.

No abstract

The design and operation of a custom-built LIDAR-compatible, four-channel Raman spectrometer integrated to a 473 nm pulsed laser is presented. The multichannel design allowed for simultaneous collection of Raman photons at spectral regions identified as highly sensitive to changes in water temperature. Four independent temperature markers were calculated for ultrapure (Milli-Q) and natural water samples [two-color(||), two-color(⊥), depolarisation(A), and depolarisation(B)]. Temperature accuracies of up to ±0.5 • C were achieved for both water types when predicted by two-color(||) markers. Multiple linear regression models were constructed considering all simultaneously acquired temperature markers, resulting in improved accuracies of up to ±0.2 • C. The potential benefits of blue laser excitation in relation to avoiding overlap between the Raman signal and fluorescence by chlorophyll-a are discussed, along with the higher Raman returns anticipated compared to the more-conventional green laser excitation.

Binary sorting between ABS and PS polymers is a challenge for the recycling industry, particularly when black pigments are present. We propose the sequential application of a hyperspectral sensor in the short-wave infrared (HSI-SWIR) and a Raman sensor unit (532 nm excitation). HSI-SWIR created maps which allowed for initial spectral and spatial assessment of the material stream and Raman point measurements enabled specific identification of ABS (white and black) and PS. The operationalisation of this sensor network requires advanced solutions for fast data acquisition, processing and classification.