Monitoring the temperature of dilute aqueous solutions using near-infrared water absorption.

An alternative spectroscopic approach for monitoring the temperature of aqueous solutions is presented. The method is based upon the temperature-induced spectral changes undergone by the second overtone (around 960 nm) of the near-infrared (NIR) water absorption band. Single and multilinear regression analysis are tested in order to evaluate the predictive ability of temperature. A linear dependence is found when measurements are performed at a single wavelength, but a lower prediction error is obtained when multilinear models are applied. No matrix effects produced by moderately concentrated common dissolved ions are found in a broad range of pH. A signal-to-noise ratio allows a precision of 0.5 degrees C for temperature monitoring. A prediction error of 0.77 degrees C (single linear regression) and 0.25 degrees C (multilinear approach) are achieved in a range from 15 to 90 degrees C. Advantages in terms of instrumentation and data analysis required are discussed.