Deep deoxidation of water in a miniaturized annular rotating device: Experimental investigation and machine learning modeling

Abstract

Conventional gas–liquid devices exhibit limitations, including inadequate dispersion efficiency and significant backmixing, leading to low conversion and large equipment volume. Thus, a miniaturized annular rotating device (m-ARD) was proposed based on microscale effects and rotating flow field to achieve efficient gas–liquid mass transfer and high conversion. Nitrogen-oxygenated water was selected as the experimental system to evaluate the performance of the m-ARD. The gas–liquid flow characteristics were investigated. The analysis focused on examining the impact of different factors on deoxygenation efficiency. The deoxygenation performance was optimized and compared with different devices. The study found that the m-ARD enables a continuous counter-current gas–liquid flow mode, with superior mixing performance and a narrow residence time distribution. The deoxygenated water with an oxygen concentration as low as 0.14 ppm can be produced, with a volumetric mass transfer coefficient (kLa) of 0.26 s−1, a theoretical stage value of 4.22, and a handling capacity of 30 mL/min under optimized experimental conditions. In addition, kLa was predicted using an artificial neural network model and showed good agreement with the simulated values. This work provides a promising reactor for chemical reactions that require high conversions.