Mercury Removal from Cement Plants by Sorbent Injection upstream of a Pulse Jet Fabric Filter

There are growing concerns over mercury emissions due to their toxicity, volatility, persistence, and bioaccumulation in the environment. Mercury emissions from cement plants are being regulated by environmental agencies in most ountries. Among the available technologies for mercury removal from flue gas, sorbent injection upstream of a polishing fabric filter is considered as the most promising and suitable technology for cement plant application. Cement plants are quite different from power plants and waste incinerators regarding the flue gas composition, temperature, gas and solid residence time, and inherent material circulation. Thus knowledge obtained from mercury removal in power plants and incinerators might not be applied to cement plants directly and fundamental investigation under well controlled cement kiln condition is imperative.

Tests in simulated cement kiln flue gas show that the red brass converter developed for waste incinerator application does not work properly for either elemental or total mercury measurement. Sodium sulfite converter is developed and
optimized for oxidized mercury reduction and total mercury measurement. The response time of the sulfite converter is short, which makes it appropriate for dynamic measurement of mercury adsorption and oxidation by sorbents.

Screening tests of sorbents for mercury removal from cement plants have been conducted in the fixed-bed reactor system using simulated cement kiln flue gas with elemental mercury and mercury chloride sources. The tested sorbents include commercial activated carbons, commercial non-carbon sorbents, and cement materials. With elemental mercury present in the flue gas, no mercury adsorption or oxidation by non-carbon based sorbents and cement materials is observed. Generally larger amount of adsorbed mercury is obtained with sorbents that have larger mercury oxidation capacity. While all the non-carbon based sorbents and cement materials show some adsorption of mercury chloride. Among the tested sorbents the Darco Hg activated shows the best performance of adsorption of both elemental and oxidized mercury and is recommended as the reference sorbent for fundamental investigation.

Parametric studies of mercury adsorption by activated carbon have been conducted in the fixed-bed reactor regarding the effects of adsorption temperature, flue gas rate, mercury level, carbon particle size, carbon load, and flue gas composition. The mercury adsorption isotherm follows Henry’s law for the applied mercury inlet levels in this project. Henry’s constant and heat of adsorption are derived for model input. The mercury adsorption capacity does not change with O₂, CO, and NO levels in the flue gas, but decreases when CO₂, H₂O, SO₂, and NO₂ concentrations increase. Slight promoting effects of HCl on mercury adsorption are observed with HCl in the flue gas up to 20 ppmv. Larger mercury adsorption capacity is obtained when HCl is removed from the gas. Similar adsorption behaviors of mercury chloride and elemental mercury by Darco Hg activated carbon are observed using simulated cement kiln flue gas, due to the effective catalytic oxidation of elemental mercury by the activated carbon.

Mathematical models are developed to simulate mercury adsorption by a single carbon particle, fixed carbon bed, in the duct and fabric filter. The developed fixed bed model can reasonably simulate the mercury breakthrough curve of the fixed carbon bed. Comparison with fabric filter model simulations and experimental data from slipstream tests at a cement plant shows that the developed two-stage model is a valuable tool and can reasonably predict the mercury removal from cement plants by carbon injection upstream of a fabric filter.