Two pollutants for the price of one
When devising your strategy for complying with the new Clean Air Interstate Rule and Clean Air Mercury Rule, you'd be wise to consider the potential synergism of available pollution control technologies. Taking the time to carefully develop a co-benefits mercury strategy may reduce your overall cost of compliance and exposure to performance and financial risk.
There are ways to reduce the production of SO2, NOx, and particulates at their source by fine-tuning the combustion process within a boiler. But for most power plants, the preferred strategy for reducing levels of these pollutants to new compliance levels (see "The new rules") is to install an air pollution control device (APCD) specific to the pollutant downstream of the boiler. What you may not know is that many APCDs designed to capture SO2, NOx, and particulates also can remove mercury from boiler flue gases, to varying degrees. This approach to mercury control is popularly known as a "co-benefits" strategy.
The amount of mercury removed by an APCD depends on its type and the variety of coal being burned. Industry experience (Figure 1) indicates that at plants burning bituminous coals, a fabric filter (FF), the combination of an FF and a spray dryer-absorber (SDA), or a flue-gas desulfurization (FGD) system (scrubber)—can reduce mercury emissions by 60% to 90%. By contrast, cold-side electrostatic precipitators (CESPs) remove only 40% of mercury from flue gases, and hot-side ESPs (HESPs) have virtually no effect on the pollutant.
1. All over the lot. The amount of mercury removed by an air pollution control device is primarily a function of the type of coal being burned. Sources: EPA ICR and DOE-sponsored demonstrations
At plants burning subbituminous coals or lignite, APCDs are much less capable of removing mercury. For the former fuel, only FFs (60%) and FGD systems (15%) have exhibited much effectiveness. In the case of lignite-burning plants, only scrubbers (20%) have shown any ability to capture mercury, and the few demonstrations of FFs have produced inconclusive results.
Based in part on the data in Figure 1, Salt Lake City–based Reaction Engineering International—a consulting firm specializing in combustion and environmental engineering—has identified two combinations of APCDs that appear to hold the most promise for removing mercury as a co-benefit:
• A wet scrubber plus a selective catalytic reduction (SCR) system.
• A low-NOx combustion system plus a particulate control device (an FF or ESP).
Scrubber + SCR
The mercury-removal effectiveness of scrubbers varies considerably from boiler to boiler, as a function of the species of mercury entering the scrubber. Unlike other pollutants, mercury can be found in several forms in boiler flue gas. Mercury leaves the boiler in gaseous, elemental form (sometimes written as Hg0) but can react with acid gases in the flue gas (particularly chlorine compounds) to form gaseous oxidized species (collectively called Hg2+). When gaseous mercury adsorbs on flyash, the result is particulate-bound mercury (Hgp).
Wet scrubbers can remove up to 90% of oxidized mercury, but they do not remove elemental mercury because it is not highly water-soluble. SDAs remove both oxidized and elemental mercury from the product of combustion of bituminous coals, but they only remove oxidized mercury from the flue gases of plants burning low-rank coals.
Most of the difference in mercury speciation among the different ranks of coal can be attributed to two variables:
• The level of unburned carbon in the flyash.
• The coal's chlorine content.
Bituminous coal flyash often has more unburned carbon than flyash from low-rank coals, particularly on boilers with low-NOx combustion systems. Unburned carbon can adsorb mercury (Hg0 and Hg2+) to form particulate-bound mercury. Unburned carbon can also act as a catalyst of oxidation and convert elemental mercury to oxidized mercury if the level of chlorine compounds in the flue gas is sufficient.
As Figure 2 shows, the chlorine content of bituminous coals varies from several hundred to several thousand ppm by weight. Conversely, the chlorine content of low-rank coals has a much lower and narrower range: from 5 to 100 ppm. In both cases, chlorine compounds in the flue gas decrease the amount of gaseous elemental mercury present at the inlet to APCDs and increase the amount of oxidized mercury there.
2. Chlorine content is key. A coal's chlorine content influences how much mercury can be removed from its combustion gases by a scrubber or the combination of a spray dryer-absorber and a fabric filter. Sources: EPA ICR and DOE-sponsored demonstrations
Figure 2 also shows that scrubbers are not as effective at removing mercury as the combination of an SDA and FF, especially at higher levels of coal chlorine content. But if the amount of oxidized mercury could be increased at the inlet to the scrubber, it could remove more mercury. SCR systems have demonstrated their ability to convert elemental mercury to oxidized mercury, with the coal's chlorine content affecting the level of conversion (Figure 3). Other factors that influence mercury oxidation are the space velocity and catalyst formulation, the temperature of the catalyst, and the sulfur content of the fuel.
3. Mercury oxidized. Elemental mercury is oxidized across an SCR system.
Source: Full-scale measurements compiled by the author
As Figure 1 shows, a wet FGD system can reduce the mercury emissions of a bituminous coal-burning power plant by about 60%. But the addition of an SCR system raises the removal potential to 80% (or more) because the system oxidizes elemental mercury and thereby increases the efficiency of the scrubber.
For plants burning a subbituminous coal or lignite, the combination of an SCR system and a scrubber has not been shown to increase mercury removal across the scrubber to the same extent as expected with bituminous coal. A number of demonstration projects now under way aim to increase the amount of oxidized mercury in the flue gas by adding halogen-containing compounds to the fuel or the boiler. Low-temperature catalysts specific to mercury oxidation are also being tested in the field. If successful, these new technologies could improve the ability of scrubbers to reduce mercury emissions from coal-fired boilers.
Low-NOx system + FF/SDA
As mentioned earlier, the combustion process can be modified to minimize NOx production, either through the use of low-NOx burners, overfire air, or reburning. However, these modifications may increase the amount of unburned carbon in the flyash. The type of boiler and the type of fuel will also have a bearing on this parameter.
The amount of unburned carbon in the flyash and the temperature of the particulate control device play important roles in mercury removal across fabric filters and ESPs. Figure 4 illustrates several examples of the removal of mercury across full-scale ESPs as a function of the loss on ignition (LOI) in the ash for various plants burning bituminous coals. Although LOI is often viewed as undesirable in flyash, the unburned carbon in the ash can provide positive benefits for mercury control by particulate control devices.
4. Reduce your LOI. Mercury removal across electrostatic precipitators from full-scale boilers burning bituminous coals is a function of loss on ignition (LOI) in flyash.
Source: Full-scale measurements compiled by the author
Recently, GE Energy & Environmental and Southern Research Institute have each demonstrated increases in mercury oxidation and removal in pilot-scale combustion systems when the combustion system was manipulated to increase unburned carbon.
Making good predictions
Under the CAMR, mercury emissions will be monitored and traded, giving them an economic value. Accordingly, if a utility chooses to pursue a "co-benefit" compliance strategy during Phase I of the CAMR, reliable prediction of mercury emissions will become important.
There is ample evidence that scrubbers and SCR, fabric filters, and cold-side ESPs can remove mercury. But the ability to predict how much mercury will be in oxidized form at the inlet to a scrubber, or how much will be converted to particulate-bound mercury at the inlet to a particulate control device, will be essential to maximizing the success and cost-effectiveness of the compliance strategy.
Predictions of mercury speciation and removal can be made by several methods, including empirical correlations, kinetic-based process models, and more-comprehensive computational fluid dynamics (CFD) models.
At Reaction Engineering International, a CFD model of a 500-MW boiler was used to quantify the impacts of variations of burner air and fuel flows on NOx, unburned carbon in flyash, and mercury speciation in flue gases. One of the project's key findings was that unburned carbon was not evenly distributed in the flue gas leaving the upper furnace (Figure 5). The CFD model was then integrated with one for mercury oxidation in the flue gas. This combined model indicated that, starting in the economizer and proceeding through the air preheater and into the particulate control device, the chlorine in the flue gas and the unburned carbon in the flyash both contribute to the oxidization of elemental mercury.
5. Shot from a canon. The trajectories of selected coal particles in a 500-MW boiler, color-coded by the level of burnout (blue = 0% burnout; red = 100% burnout).
Courtesy: Reaction Engineering International
Together, the mercury kinetic model and the boiler CFD model were able to represent the effects of changes in boiler operating conditions on unburned carbon and mercury speciation. This is just one example of how the use of modeling tools can help power plants predict the effectiveness of mercury removal by air pollution control devices.