Although existing APCD can capture some mercury, innovative control technologies will be needed to comply with the CAMR Phase II mercury emissions cap. To date, activated carbon injection (ACI) has shown the most promise as a near-term mercury control technology, although continuous long-term operation is required to determine the effect on plant operations. In a typical configuration, Powdered activated carbon (PAC) is injected downstream of the power plant's air heater and upstream of the particulate control device—either an ESP or a fabric filter. The PAC adsorbs the mercury from the combustion flue gas and is subsequently captured along with the fly ash in the particulate control device. A variation of this concept is the TOXECON™ process, where a separate baghouse is installed after the primary particulate collector (especially when it is a hot-side ESP) and an air heater and PAC is injected prior to the TOXECON unit (i.e., TOXECON I™). This concept allows for separate treatment or disposal of fly ash collected in the primary particulate control device. A variation of the process is the injection of PAC into a downstream ESP collection field to eliminate the requirement of a retrofit fabric filter and allow for potential sorbent recycling (i.e., TOXECON II™ configuration).
Overview of Powdered Activated Carbon Injection for Mercury Control
The performance of PAC in capturing mercury is influenced by the flue gas characteristics, which is determined by factors such as coal type, APCD configuration, and additions to the flue gas, including SO3 for flue gas conditioning [87]. Research has shown that HCl and sulfur species (i.e., SO2 and SO3) in the flue gas significantly impact the adsorption capacity of fly ash and activated carbon for mercury. Specifically, the following results have been found [86]:
- •
-
HCl and H2SO4 accumulate on the surface on the carbon.
- •
-
HCl increases the mercury removal effectiveness of activated carbon and fly ash for mercury, particularly as the flue gas concentration increases from 1 to 10 ppm. The relative enhancement in mercury removal performance is not as great above 10 ppm HCl. Other strong Brønsted acids, such as the hydrogen halides HCl, HBr, or HI should have a similar effect. Halogens such as Cl2 and Br2 should also be effective at enhancing mercury removal effectiveness, but this may be the result of the halogens reacting directly with mercury rather than the halides thereby promoting the effectiveness of the activated carbon.
- •
-
SO2 and SO3 reduce the equilibrium capacity of activated carbon and fly ash for mercury. Activated carbon catalyzes SO2 to H2SO4 in the flue gas. Because the concentration of SO2 is much higher than mercury in the flue gas, the overall adsorption capacity of mercury is likely dependent on the SO2 and SO3 concentrations in the gas, since these form H2SO4 on the surface of the carbon.
Figure 9.26 shows some results from conventional PAC injection tests performed through numerous U.S. DOE test programs [86]. Conventional PAC injection was the focus of initial field testing (in 2001–2002) and serves as the benchmark for all field PAC injection tests. This work showed that a maximum of approximately 65 percent mercury capture could be achieved when firing a subbituminous coal in a power plant using an electrostatic precipitator. Also, when using conventional PAC costs of 50¢/lb, it was found that the cost to remove 70 percent mercury from a bituminous coal-fired plant with an ESP would cost about $70,000/lb of mercury removed, which is higher than the targets of $30,000 to $45,000/lb of mercury removed (in 1999 dollars) [88].
Figure 9.26. Mercury removal as a function of activated carbon injection concentration.
Source: From Feeley (2006).
The conventional powdered activated carbon testing was followed by work with chemically treated PACs, which were developed for low-rank coal applications due the low mercury capture results in the initial testing. Some results from the chemically treated PAC testing are shown in Figure 9.27 [86]. With PAC injection at 1 lb/MMacf (million actual cubic feet of flue gas), mercury removal ranges from 70 to 95 percent for low-rank coals. For 90 percent mercury removal, costs are estimated at about $6,000/lb mercury removed for a subbituminous coal/SDA-FF combination to more than $20,000/lb mercury removed for a subbituminous coal/ESP combination assuming chemically treated PAC costs of $0.75 to $1.00/lb [86].
Figure 9.27. Mercury removal as a function of chemically treated activated carbon injection concentration.
Source: From Feeley (2006).Results from testing using the TOXECON technology are given in Table 9.13 (fabric filter configuration) and Figure 9.28 (ESP configuration) [87]. In the fabric filter configuration, mercury removals of 70 to 90 percent have been achieved using low-rank coals, while removal efficiencies of 50 to 90 percent are expected when using low-sulfur bituminous coals [87]. From Figure 9.28, 50 to 80 percent mercury reduction was achieved when injecting chemically treated PAC at 4–5 lb/MMacf into the next-to-last ESP field (i.e., F5) and last ESP field (i.e., F7) [86].
Table 9.13. Demonstrated Mercury Removal at 2 or 5 lb/MMacf
| Empty Cell | ESP (5 lb/MMacf) | ESP with SO3a (5 lb/MMacf) | TOXECON (2 lb/MMacf) | SDA+FF (2 lb/MMacf) |
|---|---|---|---|---|
| Low S, very low Cl (PRB or North Dakota lignite) | 78–95% with brominated PAC | 40–91% with brominated PAC | 70–90% | 90–95% with brominated PAC |
| Low S, >50 ppm Cl (some Texas lignites) | 78–95% | 40–91% | 70–90% | 90–95% with brominated PAC |
| Low S, bituminous coal | 55–75% | 40–91% | NDb | ND |
| Low S, bituminous coal with SCR | 15–70% | NAc | ND | ND |
| Low S, bituminous coal | <15% | NA | NA | NA |
- a
- SO3 from SO3 injection
- b
- Not available
- c
- Not applicable or configuration unlikely
Source: Modified from Sjostrom et al. (2007).
Figure 9.28. Mercury removal as a function of activated carbon injection concentration using the TOXECON II configuration at the Independence Station.
Source: From Feeley (2006).The improved mercury capture efficiency of the advanced chemically treated sorbent injection systems has given U.S. coal-fired power plant operators the confidence to begin deploying the technology. As of April 2008, nearly 90 full-scale ACI systems have been ordered by U.S. coal-fired power generators [89]. These contracts include both new and retrofit installations. The ACI systems have the potential to remove more than 90% of the mercury in many applications.
Balance-of-Plant Issues
Long-term tests are necessary to evaluate the effect of PAC injection on plant performance, and projects addressing this have recently started. To date, through short-term field tests with plants using ESPs, no impact of PAC injection upstream of the air preheater has been observed. In general, no balance-of-plant issues, including increases in particulate emissions and changes in ESP operation, have been identified as a result of PAC injection in the ductwork upstream of the APCDs. One exception was an increase in the arc rate at low boiler load conditions as compared to baseline arcing during PAC injection [87]. Particulate measurements collected at the outlet of the ESP when injecting PAC in the TOXECON II configuration indicated that the PAC injection did not impact particulate emissions [87]. However, stack opacity and ESP outlet particulate spikes associated with PAC injection following fourth field raps were observed.
Two significant balance-of-plant issues were discovered during TOXEON I testing at Presque Isle [87]. The first was that PAC can self-ignite in the hoppers of the baghouse when it is allowed to accumulate and when exposed to external heating from hopper heaters. The second was that the ash/PAC mixture becomes stickier than ash alone and harder to remove from the hoppers and transport with the ash removal system when it is heated. The first issue is being addressed by frequently evacuating the hoppers and controlling the maximum temperature of the heating elements to less than 300°F. The second issue can be addressed through equipment design.
No balance-of-plant problems have been noted at sites using spray dryer absorbers during PAC injection. This includes opacity or changes in spray dryer absorber or fabric filter operation.
English 
German 
Italian 
Japanese 
Korean 
Portuguese 
Russian 
Spanish 
Turkish 
Vietnamese