Non-invasive monitoring of CO2 concentration in aqueous diethanolamine (DEA), methyldiethanolamine (MDEA) and their blends in high CO2 loading region using Raman spectroscopy and partial least square regression (PLSR)

Chemical absorption using amines is a suitable method to separate CO2 from CO2 rich natural gas stream. An instantaneous monitoring of CO2 concentration in amine solvent is essential for an efficient chemical absorption process. A spectroscopic technique such as Raman spectroscopy along with multivariate modeling is considered as a robust and fast analytical method. It has been applied to monitor CO2 concentration in a chemical absorption process. However, these studies are limited to low CO2 loadings (<0.5 molCO2/molamine) and cannot be extrapolated to high CO2 loading conditions. The evaluation of Raman method at high CO2 loading is essential for the application at high pressure gas streams. In the present study, Raman spectroscopy is non-invasively applied to monitor CO2 concentration in aqueous amines (DEA, MDEA, and their blends) over a wide range of CO2 loadings (0.04�1.3 molCO2/molamine). The partial least square regression (PLSR) calibration models are developed and validated accordingly. The prediction accuracy is reported using determination coefficient (R2) and root mean square error (RMSE). The average validation R2 V and RMSEV for all the studied systems are calculated as 0.94 and 0.064 molCO2/molamine respectively. These values show that Raman spectroscopy with PLSR is a promising technique to monitor CO2 concentration for a wide range of CO2 loading. The improvement in CO2 monitoring is expected to enhance the process efficiency of natural gas processing plants. © 2017 Elsevier Ltd