By Jim Jarvis, URS, and Dr. Frank Meserole, Codan Development LLC
The United States utility industry must significantly reduce mercury emissions. Depending on location, individual power plants may have to reduce existing mercury emissions by up to 90 percent of the mercury introduced with the coal. To respond to this challenge, the industry has been engaged in research to quantify mercury emissions as a function of fuel type, determine the effectiveness of existing or planned emissions control equipment and develop and optimize innovative control processes.
In many cases, utilities are counting on their flue gas desulfurization (FGD) scrubbers and selective catalytic reduction (SCR) systems to control mercury as well as pollutants such as SO2 and NOX. As more stringent emission limits are phased in, utilities will likely need supplemental measures such as activated carbon injection (ACI) to further reduce emissions. However, SO3 could foul this strategy and potentially jeopardize the ability of some coal-fired power plants to meet future mercury emission limits.
A growing consensus exists among emissions control engineers and researchers that the presence of SO3 in the flue gas from coal-fired boilers reduces the capacity of fly ash and/or activated carbon to adsorb gaseous mercury species. The impact of SO3, which is actually a combination of SO3 and H2SO4 vapor, has been investigated on a fundamental level using approaches such as fixed-bed reactors and activated carbons. Pilot testing has provided mercury removal efficiency data on actual flue gas over a wide range of SO3 concentrations.
If SO3 threatens the feasibility of low capital-cost approaches involving activated carbon, utilities will be forced to explore other more costly options. One example is the use of mercury catalytic oxidation to boost mercury removal efficiencies in wet scrubbers. However, catalytic oxidation of mercury using SCR catalysts may have unintended consequences. These may include the formation of additional flue gas SO3, which can increase air heater plugging and downstream corrosion and may lead to visible emissions (a “blue” sulfuric acid mist plume).
A better alternative may be to remove the SO3 so it doesn’t interfere with the removal of mercury via the fly ash and/or activated carbon. This approach has the further advantage of avoiding the SO3-related operational problems mentioned earlier. Further, removing SO3 can allow a reduction in the air heater outlet temperature, which can improve heat rate and further increase the capacity of sorbents to capture mercury.
The Mercury Challenge
Mercury can exist in the gas phase in either the elemental state (Hg0) or in the oxidized state (Hg2+). Most of the mercury in coal appears to be present as sulfide compounds, usually associated with the pyritic fraction, which are destroyed during the combustion process. At temperatures above about 1,300 F, the thermodynamically stable form is elemental mercury; the preferred state exiting the boiler will tend to be vaporous, elemental mercury.
The rate and ultimate degree of mercury oxidation after the boiler will depend on several factors, including fuel type and composition, which affects the flue gas components such as HCl and fly ash that are present as the flue gas cools. Research has shown that for low-sulfur, low-chloride coals, the majority of the mercury tends to be in the elemental state. In high-sulfur, high-chloride situations, a significant fraction of the mercury exists in the oxidized state. Furthermore, it has been found that SCR catalysts and other catalysts can, to varying degrees, convert elemental mercury to the oxidized state.
Oxidized mercury can be removed effectively in wet FGD systems (in contrast, elemental mercury is not soluble and simply passes through the scrubber). Consequently, many utilities are relying on the combination of SCR and FGD to control mercury emissions. This strategy relies on the SCR reactor to remove NOX and to oxidize the remaining elemental mercury, and on the FGD scrubber to remove both the SO2 and the oxidized mercury.
The SCR/FGD strategy may not be as widely applicable as it might seem. For example, many smaller boilers are not equipped with both SCR and FGD equipment. In addition, mercury removed by the FGD may be re-emitted from the scrubber. Finally, mercury in the scrubber waste and byproduct materials is an area of increasing concern.
For example, the distribution of mercury between scrubber solutions and precipitated solids has been found to vary from process to process. In some cases, most of the mercury is associated with the solids. In other cases, as much as half of the mercury remained dissolved in the scrubber solutions.
As a consequence of the limitations of the SCR/FGD approach, other technologies are being evaluated and developed. Activated carbon injection has been found effective in removing both the elemental and oxidized mercury species. This technology removes mercury upstream of the scrubber (decreasing the mercury contained in the gypsum byproduct) and has lower capital costs than other alternatives. Although ACI has emerged as the currently preferred technology for the low-sulfur, low-chloride coal situations, it may also see applications on boilers firing high sulfur, high chloride coals
SO3 Fouls Compliance Strategy
SO3 is an important consideration for compliance strategies based on both ACI and SCR/FGD. For SCR/FGD, there is a conflict in the design goals for the SCR catalyst. Conceptually, it is important to recognize that oxidation of SO2 to SO3 is also promoted by SCR catalysts and designing a catalyst for both low oxidation of SO2 and high oxidation of mercury could be difficult, especially if extremely high mercury oxidation rates are required to meet emission limitations. One option is to use SO3 controls to help optimize the SCR/FGD approach.
SO3 plays an equally important role in the application of ACI technology. Research clearly shows that SO3 interferes with the ability of the carbon to adsorb mercury. In simple terms, SO3 and mercury compete for the active adsorption sites on the carbon particle (as well as on the carbon in the fly ash itself).
Several years ago, URS and Codan Development LLC observed that mercury retention on fly ash is strongly affected by the presence of SO3. These results were developed based on the analysis of fly ash samples obtained from ESP hoppers for five different units where a high-efficiency SO3 control process (SBS Injection) was in operation. The SBS Injection technology involves injecting a sodium-based solution such as soda ash into the duct either upstream or downstream of the air heater. The sodium-based particulate reacts to remove the SO3, and is then removed in the ESP with the fly ash.
By turning the SBS technology on and off and by varying the reagent feed ratio, correlations were obtained at all of these units for the amount of mercury retained on the ash as a function of the SO3 concentration in the flue gas. As illustrated in Figure 1, the amount of mercury retained on the native fly ash increased substantially as the amount of SO3 in the flue gas was significantly reduced. When very low levels of SO3 were achieved (less than 3 ppmv), the amount of mercury retained on the fly ash increased dramatically (by a factor of 10 at one plant).
Although gas-phase mercury measurements were not made during the fly ash sampling, the dramatic increase in the mercury concentrations in these ash samples suggests significant reductions in gas phase mercury concentrations. These results provided the impetus to conduct additional testing to explore the relationships between SO3 concentration and mercury adsorption. Two recent projects speak directly to the relationship between SO3 and mercury:
- A series of sorption tests was performed using packed-bed canisters on actual flue gas from low-sulfur Powder River Basin (PRB) coal. In some tests, the flue gas contained SO3 that was added using the plant’s ash conditioning system.
- A systematic evaluation of the impact of SO3 on ash and ACI mercury adsorption was conducted at the Mercury Research Center at Gulf Power’s Plant Crist. The testing was performed over a wide range of SO3 concentrations and carbon injection rates and included an evaluation of the impacts of SO3 removal.
In the first project, sorption tests were performed on a PRB-fired boiler using a slip stream taken from upstream of an ESP. The testing provided a direct comparison of the effect of flue gas with the low SO3 levels typical of PRB coals versus the same flue gas with higher SO3 levels (provided by the plant’s ash conditioning system). Results showed that the carbons’ mercury adsorption capacities decreased by a factor of four or more as a result of the higher SO3 concentration. In actual ACI installation, the mercury removal efficiency will often vary in proportion to the carbon adsorption capacity. Thus, a reduction in adsorption capacity due to SO3 would cause a significant reduction in mercury removal efficiency.
In the second project, a systematic evaluation of the impact of SO3 on ACI performance was conducted at Southern Co.’s Mercury Research Center (MRC). The testing made use of the MRC’s capability to manipulate key process variables such as SO3 concentration, activated carbon injection rate and air heater outlet temperature over wide ranges. The testing demonstrated the effect of SO3 on mercury capture and the beneficial effect of SO3 removal on the mercury removal capability of both the carbon in the native fly ash and the activated carbon. When high-efficiency SO3 removal was used (SBS injection in these tests), the mercury removal capability of the native ash alone was greater than with activated carbon and no SO3 controls.
Figure 2 summarizes key MRC test results. The data include 28 tests where the initial SO3 flue gas concentration was elevated (to between 35 and 50 ppm). ACI rates ranged between 0 and 10 lb/Macf, and mercury removal efficiencies shown in the figure represent the removals measured between the ACI location (just upstream of the ESP) and the ESP outlet. With no SO3 control, the mercury removal efficiency remained below 35 percent, even at the relatively high 10 lb/Macf carbon injection rate.
On the other hand, by removing the SO3, mercury removal efficiencies were more than 50 percent with no activated carbon injection at all (mercury removal in this case resulted from adsorption of the mercury onto the carbon in the fly ash).
Mercury removal efficiencies rose to above 70 percent with SO3 removal and as little as 2 lb/Macf of activated carbon. For a large coal-fired power plant, a mercury control strategy involving SO3 removal could cut net carbon costs by several million dollars a year or even eliminate the need for ACI entirely.
Solution? SO3 Controls
How should SO3 removal be implemented to obtain the greatest benefit? Are there other effects on plant operations?
One question to address is how much SO3 must be removed to obtain high ACI efficiencies. Additional data collected during the MRC testing are presented in Figure 3. Two conclusions are apparent from the figure.
First, high mercury removal efficiency was obtained only when the ESP outlet SO3 concentration was below about 1 ppm. This result suggests that low ESP outlet SO3 concentrations are required to maximize sorption and the mercury removal efficiency.
Second, although SO3 concentrations on the order of 1 ppm were noted in tests without SO3 control, the mercury removal efficiency did not significantly improve relative to tests that included SO3 control. This suggests SO3 should be removed upstream of the air heater or at least upstream of the carbon injection location.
Removing SO3 upstream of the air heater offers an additional benefit. Low air heater-inlet SO3 concentrations allow a reduction in the air heater outlet temperature and a corresponding increase in efficiency. Flue gas temperatures at the air heater outlet must be kept above about 300 F for units firing medium- to high-sulfur coals. At lower temperatures, excessive sulfuric acid condensation results in air heater fouling and corrosion.
High-efficiency SO3 removal allows air heater outlet temperatures to be reduced to as low as 250 F. Besides improving unit heat rate, the adsorption capacity of activated carbon increases at lower temperatures. The effect is illustrated in Figure 4, which shows the temperature dependence of oxidized mercury adsorption on activated carbon. Thus, high-efficiency SO3 removal, employed upstream of the air heater, allows further reductions in mercury removal together with a reduction in operating costs.
One concern is the effect of removing SO3 on ESP particulate emissions. Sorbents such as hydrated lime can remove SO3; however, by removing the SO3, fly ash resistivity increases and an increase in ESP outlet emissions or opacity can occur. Sodium sorbents do not have an adverse effect on resistivity. Operating experience from nearly 10,000 MW of installed commercial SBS Injection system indicates that the impact on ESP emissions and opacity is typically neutral or even slightly beneficial.
Authors: Jim Jarvis is a Project Manager with URS responsible for emission control process research and development. Dr. Frank Meserole is a principal and founder of Codan Development and runs Meserole Consulting.