Análise computacional da absorção química simultânea de Sulfeto de Hidrogênio (H2S) e Dióxido de Carbono (CO2) em solução aquosa de Hidróxido de Sódio (NaOH)

Abstract

Hydrogen Sulfide (H2S) and Carbon Dioxide (CO2) are chemical species naturally generated during the decomposition process of organic matter, being present in the composition of natural gases from different sources. In general, these gases have a higher amount of Carbon Dioxide (CO2), varying between 15% and 40% of the gas composition, than Hydrogen Sulfide (H2S), which varies between 10 to 3000 ppm, depending on the source of that gas. Due to the high toxicity of Hydrogen Sulfide (H2S), operational work demands a high level of attention regarding safety criteria. Thus, the use of a calculation method to simulate operating conditions before starting the experimental work is a very valuable tool. This work presents a computational tool developed using the python programming language to simulate the chemical absorption of both Hydrogen Sulfide (H2S) and Carbon Dioxide (CO2), initially using the conditions of Bontozoglou & Karabelas (1993), in an absorption column with 27.2 cm in length and 0.34 cm in diameter, using a gas flow of 2.94 m³/h, to parameterize the system and later compare with the authors' experimental data. The gas stream is synthetic gas, composed of Hydrogen Sulfide (H2S) with a concentration of 900 ppm, Carbon Dioxide (CO2) with a concentration of 0.4 mol/m³, and an inert gas that complements the composition of the gas. A sodium hydroxide solution with a concentration of 500 mol/m³ was used as the solvent. A model was obtained that obtained a removal of Hydrogen Sulfide (H2S) of up to 89.7% with an average percentage error between the value calculated by the model and the experimental data of 8.22%. For Carbon Dioxide (CO2), the removal value obtained was 5.01%, which, despite being low, was in agreement with the experimental value, with an average error of 5.94%. The tool was tested with experimental data from Turpin et al. (2008), performing the simulation under the same conditions employed by the authors, obtaining an average error of 8.64%.