Abstract

The NOXSO flue gas cleanup process uses a dry regenerable sorbent capable of adsorbing 95+% of SO$\sb2$ and 90+% of NO$\rm\sb{x}$ from boiler flue gas. The process consists of two main steps, namely an adsorption-step where sulfur and nitrogen oxides are adsorbed, and a regeneration-step where the spent sorbent can be regenerated using various reducing gases, such as CH$\sb4,$ CO, H$\sb2,$ H$\sb2$S, and reforming gas. During sorbent regeneration, first NO$\rm\sb{x}$ comes out and is sent back to the boiler where the recycle NO$\rm\sb{x}$-rich gas stream decreases the NO$\rm\sb{x}$ production by upsetting the equilibrium of the reaction that produces NO$\rm\sb{x}$. Then a reducing gas is brought in contact with the sorbent at about 620$\sp\circ$C where several chemical reactions take place. The product of these reactions are mainly H$\sb2$S and SO$\sb2$ which can be used in the Claus plant for the production of elementary sulfur. After regeneration, the sorbent is transferred to a cooler before sending it back to the adsorber. The adsorption-step has been thoroughly investigated and modeled by NOXSO Corporation, whereas the regeneration-step had not previously been completely understood.

The main objective of this research is to study and model the regeneration-step of the NOXSO process using various reducing gases. Screening experiments showed that CH$\sb4,$ H$\sb2,$ and Reforming gas were equally effective for the sorbent regeneration at 620$\sp\circ$C and CH$\sb4$ was then selected as a potential candidate for sorbent regeneration. The effects of various operating variables, including bed height, temperature, sulfur loading, residence time, and inlet methane concentration on the sorbent regeneration efficiency and time were studied. The kinetics constants and activation energies of the important reactions with methane were obtained from the experimental data. These values were employed in the development of a comprehensive mathematical model for the sorbent regeneration process.

The experimental results and model predictions suggested that temperature, sorbent sulfur loading, and inlet CH$\sb4$ concentration are the most important operating variables that affect sorbent regeneration. The temperature effect was significant within the operating range used, since the process was found to be kinetically controlled. The CH$\sb4$ concentration effect, on the other hand, was only significant below 40 mol%. The effect of sulfur loading on the sulfur removal during the induction period was well pronounced whereas its effect on the sulfur removal after the induction period was not significant.

The model predicted that it would require more than twice the time to increase the sorbent regeneration efficiency from 80% to 100%, which could be attributed to the increased carbon deposit (coking) on the sorbent during regeneration. Considering the presence of carbon deposits and any sulfur remaining in the form of Na$\sb2$S on the sorbent after regeneration using CH$\sb4,$ steam treatment of the sorbent becomes a crucial step in the overall NOXSO process, since it removes both carbon deposit and Na$\sb2$S.