![[logo] US EPA](https://webarchive.library.unt.edu/eot2008/20081107212232im_/http://www.epa.gov/epafiles/images/logo_epaseal.gif){kind=logo}
Final Report: Control of Mercury Emissions from Coal-Fired Power Plants
EPA Grant Number: R828170Title: Control of Mercury Emissions from Coal-Fired Power Plants
Investigators: Helble, Joseph J. , Qiu, Joseph , Sarofim, Adel F. , Sterling, R.
Institution: University of Connecticut , University of Utah
EPA Project Officer: Shapiro, Paul
Project Period: July 1, 2000 through June 30, 2002 (Extended to September 30, 2003)
Project Amount: $224,642
RFA: Exploratory Research - Environmental Engineering (1999)
Research Category: Engineering and Environmental Chemistry
Description:
Objective:The overall objective of the project was to explore and potentially develop an alternative, low-cost method for capturing mercury in coal-fired power plants. Specific program objectives included: (1) determination of the fundamental gas-phase oxidation pathways for mercury in simulated coal flue gas under a broad range of conditions; and (2) assessment of the feasibility of increasing the oxidation of mercury through manipulation of the oxidizing radical pool in coal combustion gases.
Summary/Accomplishments (Outputs/Outcomes):Research in this project therefore focused on four key tasks in support of the program objectives. These were to: (1) conduct baseline kinetic modeling of Hg oxidation using mechanisms described in the technical literature to identify conditions that could be used to increase Hg oxidation; (2) conduct baseline bench scale experiments in a unique flat flame-fired flow reactor to measure the effects of constituents such as HCl, Cl2, SO2, NO, and O2 on the extent of mercury oxidation; (3) conduct a detailed examination of the rate constants associated with the Hg oxidation mechanism and improve existing kinetic models to predict the effects of gas injection on Hg oxidation; and (4) conduct bench scale experiments at the University of Connecticut and pilot scale combustion experiments at the University of Utah to assess the feasibility of gas injection as a means of promoting Hg oxidation.
Initial work conducted under Task 1 demonstrated that increases in Cl radical concentration could be effected through the injection of combustion products, but might be sensitive to changes in the temperature of the flue gases at the point of injection. Injecting the combustion products of a methane/oxygen adiabatic reactor at different temperatures in the post combustion zone enhances the chlorine radical concentration by diminishing amounts as the injection temperature is decreased. This is related to a decrease in OH with decreases in the injection temperature, and shows the importance of recombination reactions in decreasing the superequilibrium free radical concentration.
Subsequent work under Task 2 focused on a bench scale study of the effects of gaseous constituents such as HC1, Cl2, SO2, NO, and O2 on mercury oxidation. The investigation of the homogeneous oxidation of mercury in simulated flue gases was carried out using Cl2 concentrations of 0, 150, 250, or 500 ppm, HCl concentrations of 0, 100, 150, 200, 250, or 300 ppm, SO2 concentrations of 0, 100, or 300 ppm, and (added) NO concentrations of 0, 100, or 300 ppm. In all cases, homogeneous oxidation was investigated using a flat-flame fired system consisting of a mercury vapor generation unit, a flue gas generation unit, a quartz reactor, a flue gas sampling system, a gas injection system and data acquisition systems. Use of a flame provided the radical pool believed to be important for initiating mercury oxidation. Mercury was injected into the post-flame zone and the mercury containing gases then were exhausted into an insulated quartz tube reactor for study of the relevant mercury chemistry. The quartz reactor was an 80 cm long 12.7 cm diameter tube mounted at 5-10 degrees to the horizontal to prevent the accumulation of products of condensation. As noted above, heat tapes and insulation were wrapped around the circumference of the reactor to provide additional heating and to reduce heat losses to the environment. Flue gases were sampled using a quartz probe that was connected to the reactor through one of four sampling ports. Flue gases were cooled and passed to a Semtech mercury (Hg) analyzer for determination of elemental mercury content.
Results of the bench scale experiments confirmed that concentrations of Cl2 and HCl played an important role in determining the extent of Hg oxidation, that SO2 inhibited Hg oxidation by Cl2, and that NOx played a lesser role, but the magnitude of any effect was affected by concentrations of NO and the oxidizing (chlorine-containing) species. For example, the addition of NO (beyond any thermal NO) had a slight inhibitory effect on Hg oxidation by HCl at HCl concentrations of 300 ppm. In contrast, SO2 has a large inhibitory effect on oxidation by Cl2. In experiments with HCl and SO2, very little effect was observed. The scavenging of O radicals by SO2(g) is a plausible interpretation for the indirect reduction of the chlorine-containing radicals generated via reaction. The overall effect is a reduction in mercury oxidation represented by the eight step mechanism where these chlorine-containing radicals are needed for initiating Hgo(g) conversion. In the presence of 100 ppm HCl at a flame equivalence ratio of 1, 6% of the Hg present was oxidized to HgCl2. In the presence of 100 ppm SO2, oxidation remained unchanged at 6%. Increasing the SO2 concentration to 400 ppm resulted in a statistically insignificant change in Hg oxidation to 7%. Experimental results indicated that increasing oxygen levels also contribute to an increase in Hg oxidation. Results from a series of experiments at two different flame stoichiometries in the presence of varying levels of HCl show little difference at low HCl concentrations, but differences become apparent as HCl concentrations increase above 200 ppm.
Results were interpreted through the use of elementary reaction-based kinetic modeling. A detailed kinetic model involving several modified Hg reaction rate constants calculated using quantum chemistry methods and involving NOx/SOx chemistry and S-Cl-O reactions was therefore developed. This resulted in a comprehensive kinetic model consisting of 168 elementary reactions that describes mercury oxidation under post-coal combustion conditions. This model was used to identify important pathways and used to interpret results of the experiments conducted for this project.
Calculations showed that the reaction HCl + O = OH + Cl is an important pathway for releasing Cl into the system. The rate of this reaction is strongly temperature dependent. Cl radical is the most abundant Cl-containing radical from 1373 K to 600 K. In the early post-flame region, calculated Cl mole fractions are much higher than other Cl-containing species, with the exception of HCl. As the flue gases cool, HOCl, ClO and Cl2 form through reactions of Cl with OH, O, and a third-body recombination reaction. A portion of the Cl radicals was also predicted to be converted back to HCl. At low temperature, Hg oxidation was seen to be limited by the lack of Cl radicals and Cl2 formed by recombination reactions. This suggested that moderate temperatures were necessary for the formation of the radical species that can oxidize Hg in systems (such as coal combustion systems) where HCl is the dominant chlorine species.
SO2 showed a strong inhibitory effect on mercury oxidation by Cl2. Mercury oxidation was reduced as the concentration of SO2 increased in the absence of added NO, consistent with experimental results. Sensitivity analysis indicated that SO2 consumes OH radicals and O radicals via the reactions SO2 + OH = SO3 + H, SO2 + OH + M = HOSO2 + M and SO2 + O + M = SO3 + M leading to a reduction in Cl radicals.
The final task under this project examined the addition of a small amount of hydrocarbons to high temperature post-combustion gases as a potential method for increasing concentrations of OH, Cl, and Cl2. This suggested the possibility of promoting Hg oxidation by utilizing hydrocarbon injection. To identify the preferred conditions for hydrocarbon injection, methane oxidation chemistry was added to the kinetic model of this study. It was found after a series of calculations that the mole fractions of Cl radicals would rise upon addition of one of several hydrocarbon additives; for example, the injection of 1000 ppm CH4 increased Cl radicals to values almost five times the initial level without injection. Calculations were also carried out to examine the effectiveness of combustion product injection (as opposed to straight hydrocarbon injection) on mercury oxidation. In this case, a small amount of CH4 was burned to form combustion products including CxHy radicals under various equivalence ratios and then injected into the main combustion flue gas. Here the injection of combustion products fired under close to stoichiometric conditions (0.9–1.2) was seen to be more efficient, and injection of combustion products at 973 K was found to have the potential to enhance mercury oxidation.
Experiments were conducted with a retrofitted mixing chamber to study hydrocarbon-air injection. In these experiments, small quantities of CH4(g) and oxidizer in various proportions were added to the simulated flue gases containing mercury and chlorine species at specific flue gas temperatures and the level of mercury oxidation was then determined. The CH4(g) was added counter current to or co-current to the flow of the flue gases. A subsequent comparison of the amount of Hgo(g) converted to Hg2+(g) by Cl2(g) with and without CH4(g) injection provided an indication of the effect of the radicals and/or gases added on the mercury speciation. The data reported for the CH4(g) injection experiments indicated that Hgo(g) oxidation was 80% with no CH4(g) added and 26% when 1250 ppmv of CH4(g) was added in a CH4(g)-O2(g)-N2(g) gas mixture to flue gases. For experiments where CH4(g) was replaced with an equivalent amount of N2(g) in the injected gas stream, the amount of Hgo(g) oxidized was 81% with no gas injection and 79% with gas injection. These results clearly indicated that mercury oxidation was significantly affected by the presence of methane, but not in the desired direction.
To further understand the effect of CH4(g) addition to the flue gas mixture, experiments were conducted with the gases injected counter current to and co-current to the flow of the flue gases. The concentrations of initial Hgo(g), Cl2(g), and CH4(g) added to the combustion gases were the same as the previous experiments and a similar approach was used to analyze the results obtained. From the level of mercury oxidation obtained from these two sets of experiments, the direction of the injected gases containing CH4(g) influenced the amount of Hgo(g) converted to Hg2+(g) in the presence of 250 ppmv Cl2(g). It is speculated that the difference in impact of the CH4(g) injection for the counter current flow conditions versus co-current flow is most probably a manifestation of the consumption of the different radicals needed to promote mercury oxidation. Specifically, it is suggested that radicals such as O(g) and OH(g) that play a secondary role in the oxidation reaction mechanism, are consumed when CH4(g) is injected counter current to the flue gas flow.
Based on modeling, it was postulated that changing the composition of the combustion reactant gases injected increased the concentration of the radicals responsible for mercury speciation. To verify this, a total of ten experiments were conducted in sets of two where only the equivalence ratio of the downstream injected gas mixture was manipulated. From the first set of experiments, with no CH4(g) injected 74.8% of the initial Hgo(g) present in the flue gas sample was oxidized by the chlorine species whereas the subsequent injection of a CH4(g)-O2(g)-N2(g) gas mixture with φin = 0.8 reduced the amount of Hgo(g) oxidized to 62%. With the same Hgo(g) and Cl2(g) concentrations in flue gas, changing the injected CH4(g)-O2(g)-N2(g) gas mixture equivalence ratio to φ = 1.0 led to 69% conversion of the Hgo(g) species present to the oxidized form. A comparison of the latter result with the data collected from the first gas injection experiment showed that the inhibitory effect of the gases injected decreased with an increasing gas injection equivalence ratio.
While these results were not able to show an enhancement of mercury oxidation through hydrocarbon addition, they did show that the homogeneous chemistry was quite sensitive to the temperature and stoichiometric ratio of the injected mixture. Calculations suggest that under higher temperature conditions, there may be some enhancement of mercury oxidation. Calculations conducted by the University of Utah also suggest that rapid quench through water addition would promote oxidation. It may therefore be possible to enhance Hg oxidation through a combination of quench and hydrocarbon addition. Attempts to demonstrate this experimentally were not successful.
Several important conclusions regarding mercury in combustion systems can be drawn from this study. These include: a clear indication that fuel-lean conditions (v. stoichiometric combustion) promote Hg oxidation by either HCl or Cl2; a determination that in the presence of Cl2, SO2 can suppress mercury oxidation, whereas in the presence of HCl(g), SO2(g) concentrations of 100 and 400 ppm have little to no effect on mercury oxidation; a finding that homogeneous chemistry alone appears insufficient to account for levels of oxidized Hg observed in coal combustion systems, confirming that surface reactions play a role in the oxidation of mercury in many full scale coal combustion systems; and finally, a determination that injection temperatures above 1000°C followed by rapid quench, perhaps through subsequent water spray injection, would increase the Cl radical pool sufficiently to produce the desired increase in oxidized mercury concentrations.
Journal Articles:No journal articles submitted with this report: View all 6 publications for this project
Supplemental Keywords:toxics, metals, heavy metals, engineering, environmental chemistry,
,
Toxics, Air, Scientific Discipline, Waste, RFA, Engineering, Chemistry, & Physics, HAPS, Incineration/Combustion, air toxics, particulate matter, Environmental Chemistry, 33/50, Environmental Monitoring, tropospheric ozone, heavy metals, coal combustion, anthropogenic stresses, hydrocarbons, ambient emissions, combustion byproducts, ambient air, benzene, chemical composition, hydrocarbon, benzene emissions, combustion, mercury & mercury compounds, particulates, methane, flue gas emissions, air pollution, mercury, stratospheric ozone, ion chromatography, mercury speciation, anthropogenic stress, Benzene (including benzene from gasoline), Mercury Compounds, coal fired power plants
Progress and Final Reports:
2000 Progress Report
2001 Progress Report
Original Abstract