TY - GEN

T1 - Single-path supercontinuum laser spectroscopy combining MIR-ATR and NIR transmission

AU - Sonoiki, Doyinsola S.

AU - Kwarkye, Kyei

AU - Sørensen, Klavs M.

AU - Engelsen, Søren B.

AU - Bang, Ole

AU - Petersen, Christian R.

N1 - Publisher Copyright: © 2024 SPIE.

PY - 2024

Y1 - 2024

N2 - The strong absorption of water in the mid-infrared (MIR) causes difficulties in performing combined near-infrared (NIR) and MIR spectroscopy on aqueous samples using a single instrument. Combining spectra from different instruments can result in unwanted spectral variations, which can influence the prediction models and mitigate the advantages of the combination approaches. In this work, absorption spectra were collected in the NIR (1300nm-2500nm) and MIR (2500nm-3800nm) region by combining a single high-brightness broadband supercontinuum (SC) laser spanning from ∼1-4.0 μm with attenuated total reflectance (ATR) and a transmission cuvette in a single-path configuration to provide a uniform spectral response across the NIR and MIR regions. The measured NIR- and MIR- spectra were assessed based on their ability to predict varying concentrations of ethanol, sucrose, and L-proline in aqueous solutions. The NIR-based partial least square regression (PLSR) model gave higher prediction accuracy for sucrose (R2 = 0.95) as compared to both ethanol and L-proline (R2 = 0.75 and R2 = 0.57 respectively). On the other hand, the MIR-based model enhances the prediction accuracy of ethanol (R2 = 1.00) and L-proline (R2 = 0.62) while demonstrating no significant change in prediction accuracy for sucrose (R2 = 0.96). The prediction models based on the combined NIR-MIR spectra performed similar but slightly worse than the MIR-only models for ethanol and L-proline (R2 = 0.97 and R2 = 0.54 respectively), while for sucrose, it was slightly improved (R2 = 0.99).

AB - The strong absorption of water in the mid-infrared (MIR) causes difficulties in performing combined near-infrared (NIR) and MIR spectroscopy on aqueous samples using a single instrument. Combining spectra from different instruments can result in unwanted spectral variations, which can influence the prediction models and mitigate the advantages of the combination approaches. In this work, absorption spectra were collected in the NIR (1300nm-2500nm) and MIR (2500nm-3800nm) region by combining a single high-brightness broadband supercontinuum (SC) laser spanning from ∼1-4.0 μm with attenuated total reflectance (ATR) and a transmission cuvette in a single-path configuration to provide a uniform spectral response across the NIR and MIR regions. The measured NIR- and MIR- spectra were assessed based on their ability to predict varying concentrations of ethanol, sucrose, and L-proline in aqueous solutions. The NIR-based partial least square regression (PLSR) model gave higher prediction accuracy for sucrose (R2 = 0.95) as compared to both ethanol and L-proline (R2 = 0.75 and R2 = 0.57 respectively). On the other hand, the MIR-based model enhances the prediction accuracy of ethanol (R2 = 1.00) and L-proline (R2 = 0.62) while demonstrating no significant change in prediction accuracy for sucrose (R2 = 0.96). The prediction models based on the combined NIR-MIR spectra performed similar but slightly worse than the MIR-only models for ethanol and L-proline (R2 = 0.97 and R2 = 0.54 respectively), while for sucrose, it was slightly improved (R2 = 0.99).

KW - aqueous solution

KW - ATR

KW - MIR

KW - NIR

KW - PLSR

KW - single-path

KW - spectroscopy

KW - Supercontinuum

U2 - 10.1117/12.3000302

DO - 10.1117/12.3000302

M3 - Article in proceedings

AN - SCOPUS:85191516983

T3 - Proceedings of SPIE - The International Society for Optical Engineering

BT - Terahertz, RF, Millimeter, and Submillimeter-Wave Technology and Applications XVII

A2 - Sadwick, Laurence P.

A2 - Yang, Tianxin

PB - SPIE

T2 - Terahertz, RF, Millimeter, and Submillimeter-Wave Technology and Applications XVII 2024

Y2 - 29 January 2024 through 1 February 2024

ER -