Bench-Scale Studies of In-Duct Mercury Capture Using Cupric

Mar 12, 2009 - in ducts upstream of particulate matter control devices. ... mercury capture achieved in the duct portion with a residence time of 0.54...
0 downloads 0 Views 597KB Size
Download PDF
Environ. Sci. Technol. 2009, 43, 2957–2962

Bench-Scale Studies of In-Duct Mercury Capture Using Cupric Chloride-Impregnated Carbons SANG-SUP LEE,† JOO-YOUP LEE,‡ AND T I M C . K E E N E R * ,† Department of Civil and Environmental Engineering, University of Cincinnati, Cincinnati, Ohio 45221, and Department of Chemical and Materials Engineering, University of Cincinnati, Cincinnati, Ohio 45221

Received July 14, 2008. Revised manuscript received December 15, 2008. Accepted January 30, 2009.

A brominated activated carbon (Darco Hg-LH) and cupric chloride-impregnated activated carbon (CuCl2-ACs) sorbent have been tested in a bench-scale entrained-flow reactor system which was developed for simulating in-flight mercury capture in ducts upstream of particulate matter control devices. The bench-scale experimental system has been operated with the conditions of a residence time of 0.75 s and a gas temperature of 140 °C to simulate typical conditions in the duct of coal-fired exhaust gas. In addition, sorbent deposition on walls which can occur in a laboratory-scale system more than in a full-scale system was significantly reduced so that additional mercury capture by the deposited sorbent was minimized. In the entrained-flow system, CuCl2-ACs demonstrated similar performance in Hg adsorption and better performance in Hg0 oxidation than Darco Hg-LH. In addition, the carbon content of those sorbents was found to determine their Hg adsorption capability in the entrained-flow system. The bench-scale entrained-flow system was able to demonstrate the important Hg adsorption and oxidation characteristics of the tested sorbents.

Introduction Sorbent injection is one of the most promising control technologies for reducing mercury emissions from coal-fired power plants (1-3). Since the use of economical and efficient sorbents is most important in sorbent injection, various sorbents have been studied and developed. Among these sorbents, activated carbons have been extensively studied and shown their capability for mercury capture (4-6). However, raw activated carbon has low capability in mercury removal from flue gases with low hydrogen chloride (HCl) concentration so that a relatively large amount of raw activated carbon is required for mercury emissions control from subbituminous or lignite coal-burning units producing a low concentration of HCl in a flue gas (7-9). For subbituminous or lignite-burning units, the use of chemically treated activated carbons is an effective option to enhance mercury removal efficiency and lower the amount of carbon injection. For example, Ghorishi et al. have reported a significant increase in Hg0 removal by impregnating HCl onto * Corresponding author e-mail: [email protected]; tel: 1-513556-3676; fax: 1-513-556-2599. † Department of Civil and Environmental Engineering. ‡ Department of Chemical and Materials Engineering. 10.1021/es801943t CCC: $40.75

Published on Web 03/12/2009

 2009 American Chemical Society

activated carbon (10). Especially, brominated activated carbon has been tested in several full-scale systems and has demonstrated good performance such as more than 90% mercury removal efficiency at an injection level of less than 80 mg/m3 (5 lb/MMacf) (1, 11). EPA’s Information Collection Request (ICR) data show that mercury removal efficiency varies with a particulate control system (3, 12). Full-scale test results also show significantly higher mercury removal in a fabric filter (FF) than an electrostatic precipitator (ESP) with a similar level of sorbent injection (1, 11). However, while 83% of the U.S. coal-fired boiler units operating with particulate control devices are equipped with ESPs, only 14% are equipped with FFs (8). In addition, full-scale test results (13) show that the mercury capture achieved in the duct portion with a residence time of 0.54 s reaches approximately 90% of the total mercury capture achieved in both duct and an ESP with a residence time of 14.04 s. These results stress the importance of using an entrained-flow system to simulate in-duct mercury capture upstream of an ESP rather than inside an ESP for mercury sorbent tests. Tests reported here therefore were designed to measure the in-duct portion of mercury capture in order to more accurately determine the effect of kinetics on the reaction in the duct stream. In this study, a bench-scale entrained-flow system was constructed and operated for simulating in-duct mercury capture using a mixture of air and elemental mercury vapor. In addition, cupric chloride-impregnated carbons, which demonstrated excellent performance in our previous studies (7, 14), were tested and their results were compared with those of a representative brominated activated carbon. Although a few test results were obtained in air with sulfur dioxide (SO2), most entrained-flow tests were conducted in air to obtain the Hg0 oxidation capability of cupric chloride and Hg adsorption capability under the minimum influence of other flue gas components. Therefore, the performance results obtained from the bench-scale entrained-flow system may not be directly used for the projection of full-scale performance.

Experimental Section Sorbent Preparations. Three concentrations of cupric chloride were impregnated onto Darco-Hg which is raw activated carbon (Norit Americas Inc., Marshall, TX). Portions of 0.24, 0.5, and 1.125 g of copper chloride dihydrates (CuCl2 · 2H2O, Aldrich Chemical. Co., Milwaukee, WI) were added to 100 mL of isopropyl alcohol (99.8%, Pharmco Products Inc., Brookfield, CT), respectively, under vigorous stirring until its complete dissolution. Darco-Hg activated carbon (4.5 g) was then added in each cupric chloride solution, and the mixtures were kept at room temperature for about 3 h with continuous stirring. Each solution was then filtered, and each resulting solid was dried at 100 °C for about 2 h to create cupric chloride-impregnated activated carbon sorbents. To find the concentrations of cupric chloride in the sorbents, the filtered solutions were analyzed for the residual concentrations of copper and chloride in the solutions using atomic absorption spectrophotometry (AAnalyst 300, PerkinElmer, Waltham, MA) and ion chromatography (DX-600, Dionex, Sunnyvale, CA), respectively. Then, the concentrations of cupric chloride impregnated onto each sorbent were determined from the difference before and after the impregnation process. The results showed that 3.8%, 6.5%, and 10% (w) cupric chlorides (as CuCl2) were impregnated onto each Darco-Hg activated carbon, respectively. These sorbents VOL. 43, NO. 8, 2009 / ENVIRONMENTAL SCIENCE & TECHNOLOGY

9

2957

FIGURE 1. Schematic of an entrained-flow reactor system simulating in-duct mercury capture.

TABLE 1. Surface Areas and Mean Particle Sizes of the Sorbents Tested in This Study surface area (m2/g) sorbent

description

from Norit’s datasheet

from measurements

mean particle diameter (µm)

Darco-Hg 3.8% CuCl2-AC 6.5% CuCl2-AC 10% CuCl2-AC Darco Hg-LH

lignite coal-based activated carbon 3.8% (w) CuCl2-impregnated activated carbon 6.5% (w) CuCl2-impregnated activated carbon 10% (w) CuCl2-impregnated activated carbon brominated, lignite coal-based activated carbon

600

450 ((30) 400 ((50) 420 ((30) 400 ((70) 350 ((60)

23 23 23 23 22

were found to have similar surface areas and mean particle diameters as those of Darco-Hg from the analysis using a single point BET technique (Monosorb Surface Area Analyzer, model MS-12, Quantachrome Instruments, Boynton Beach, FL) and laser particle counting system (PC-2000, Spectrex Corporation, Redwood city, CA) as shown in Table 1. In addition, brominated activated carbon known commercially as Darco Hg-LH (Norit Americas Inc., Marshall, TX) served as a benchmark sorbent in this study. Experimental Conditions. An entrained-flow reactor was designed and constructed to simulate mercury capture in the ductwork. The 8 m long 2.5 cm diameter reactor was made of borosilicate glass which is inert to mercury vapor. Sections were connected using glass spherical joints and spring clamps (35/25 size, Ace Glass, Inc. Vineland, NJ). Heating tapes (Fisher Scientific, Inc. Pittsburgh, PA) with separate controllers (Omega Engineering, Inc., Stamford, CT) were used for every 2-m section of the reactor, and type K thermocouples were placed in leak tight fittings at 2-m intervals for monitoring the duct temperature. The tips of the thermocouples were placed so as to not interfere in the gas flow. Fiberglass pipe insulation (McMaster-Carr, Aurora, OH) was used to insulate the entire length of the reactor. The inlet air flow was preheated in a furnace at a temperature of 700 °C (Lindberg, Watertown, WI) before entering the reactor. Sorbent particles were injected through a venturi-type Bernoulli feeder. The feeder operates by allowing particles to be entrained off of a rotating variable speed circular top which has a precise groove cut that contains the particles. An aspirator (shown in Figure 1) is mounted above and close to the groove and picks up the particles by means of suction and then they are conveyed to the reactor where they are injected. The sorbent injection rate is controlled by the speed of the rotating top. Operating conditions consisted of a total air flow rate of 139.5 L/min at a reactor temperature of 140 °C and have a flow velocity of 6.7 m/sec approaching that (∼15 m/sec) of full-scale ductwork. With this air flow rate, a residence time 2958

9

ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 43, NO. 8, 2009

550

of 0.75 s was also obtained, and this residence time is within the range of that in the full-scale ductwork upstream of an ESP. To minimize additional mercury capture by sorbents deposited on the reactor wall, Hg0-laden gas was injected after the first 3 m of the reactor where most of sorbent depositions were found. In addition, the reactor was continuously vibrated by using vibrating pens (McMaster-Carr, Aurora, OH) located at 2-m intervals as shown in Figure 1. This technique was found to be extremely successful in minimizing the amount of sorbent that deposited on the walls. These experimental conditions are also summarized in Figure 1. Experimental Procedure and Data Analysis. During the sorbent injection, a portion of the effluent gas was sampled at a flow rate of 1.5 L/min to measure the outlet mercury concentrations. The outlet mercury speciation was conducted by using the Ontario Hydro Method (15). The sampled gas passed through a 1 M KCl impinger solution and a 4% (w/v) KMnO4/10% (v/v) H2SO4 impinger solution to capture oxidized mercury and elemental mercury in the sampled gas, respectively. Oxidized mercury (Hg2+out) and elemental mercury (Hg0out) concentrations in the effluent gas were determined by analyzing those solutions using a cold vapor atomic absorption spectrophotometer (CVAAS, model 400A, Buck Scientific Inc., East Norwalk, CT). At the beginning of each test, the reactor was heated to an operating temperature of 140 °C by injecting a sorbentfree, filtered, dehumidified, and purified air flow supplied by an air compressor (Balston compressed air filters, model A912A, Parker Hannifin Corp., Cleveland, OH). An Hg0-laden air flow generated from elemental mercury permeation tubes (VICI Metronics, Inc., Poulsbo, WA) in a permeation tube oven (Dynacalibrator model 500, VICI Metronics, Inc., Poulsbo, WA) was also injected. Blank Hg0 concentrations were measured at several locations such as the outlet of the mercury permeation oven, 1 m after Hg0 injection, and the outlet of the reactor system in the absence of sorbent, and

similar blank Hg0 concentrations were obtained from all measurements. Throughout the blank tests, it was confirmed that neither Hg0 oxidation nor removal took place inside the reactor without sorbent injection, and injected Hg0 was quickly mixed with air flow. The average inlet Hg0 concentration was determined to be 6.5 ppbv from the blank measurements, and its variation was confirmed within 0.5 ppbv. Each sorbent (0.025-0.3 g) was premixed with 2 g of montmorillonite K 10 clay (Aldrich Chemical Co. Milwaukee, WI) and then injected through a venturi-type Bernoulli feeder at a constant feeding rate for 10 min. The sorbent particles were separated from the flue gas stream at the end of the reactor by a cyclone made of borosilicate material. After each test run during 10 min, sorbent particles collected in a cyclone and deposited on the reactor wall were obtained by washing those with acetone (99.5%, Reagent grade ACS/USP, Pharmco Products Inc., Brookfield, CT). Then, a sorbent mass balance was taken to determine an actual sorbent injection rate into the reaction zone. All the collected spent sorbent particles were analyzed for the amount of mercury adsorption (Hgads) by following the digestion procedure described in the Ontario Hydro Method.

Results and Discussion Performance of the Entrained-Flow System. The flow conditions in the full-scale ductwork such as a residence time of less than 1 s and a gas temperature of around 140 °C can be obtained in the entrained-flow system by providing it with flow and heating as explained in the previous experimental section. However, compared to the full-scale ductwork, the entrained-flow system has a significantly higher ratio of inside surface area to the volume of the system (S/V ratio ) 2/R; where R is a radius of the ductwork). Assuming that a Reynolds number of 500,000 and a flue gas velocity of 15.24 m/sec (50 ft/sec) are applied to the full-scale ductwork, the full-scale system has an S/V ratio of 4.37 (m2/m3) while the entrained-flow system has an S/V ratio as high as 160 (m2/m3). This significantly higher S/V ratio of the entrainedflow system can cause a significant amount of sorbent deposition on reactor walls so that mercury vapor can be additionally captured by the deposited sorbent. To minimize the sorbent deposition in our tests, the entrained-flow system was provided with the highest flow rate eligible to the current system configurations. In addition, Hg0 vapor was injected after the first 3 m of the reactor which was found to have relatively higher sorbent deposition. Vibrating pens were also applied to prevent sorbent deposition, and montmorillonite clay was injected together with each sorbent because coinjection of montmorillonite clay was found to decrease the amount of sorbent deposition and make the sorbent particles more easily detached from the reactor walls by vibration. In addition, montmorillonite clay was found to be inert toward Hg0 vapor from our blank tests. After each test, individual sorbent and mercury mass balances were taken with respect to fractions escaping from the reactor, collected in the cyclone, and deposited on the reactor wall. As a result, approximately 10% of the injected sorbent was found to be deposited on the reactor wall, and only 10-15% of the total mercury adsorption was achieved by the deposited sorbent for all of the tests conducted in this study. Therefore, additional mercury capture by deposited sorbent was also minimized in the entrained-flow system. From the entrained-flow tests, Hg adsorption efficiency was determined from the amount of mercury adsorbed (Hgads) onto each sorbent. The amount of effluent oxidized mercury (Hg2+out) was measured in the outlet of the reactor system, and effluent oxidized mercury percentage was determined. In addition, the sum of these amounts was defined as Hg0-

FIGURE 2. Mercury speciation results of replicate tests with 6.5% CuCl2-AC at 20 mg/m3 based on inlet Hg0. removal total because elemental mercury was removed in the form of oxidized mercury, which can be captured by fly ash and wet scrubbers in full-scale systems. Three replicate tests were conducted with 6.5% CuCl2-AC at 20 mg/m3 to examine reproducibility of entrained-flow test results. Figure 2 shows mercury speciation results obtained based on the inlet Hg0. As shown in the figure, mass balance closures of three replicate tests are in the range of the inlet Hg0 concentration (6.5 ( 0.5 ppbv), and Hgads, Hg2+out, and Hgads + Hg2+out generally increase in proportion to the mercury mass balance closure. These results indicate that the variations in Hgads, Hg2+out, and Hgads + Hg2+out are ascribed to the variation in the inlet Hg0 concentration rather than the errors from the measurements of Hgads, Hg2+out, and Hg0out. Therefore, the efficiency values reported here were determined based on the sum of the amount of mercury adsorbed (Hgads), the amount of effluent oxidized mercury (Hg2+out), and the amount of effluent elemental mercury (Hg0out). In addition, the mercury mass balance closures of all tests were in a reasonably acceptable range (87-110%). Performance of Mercury Sorbents. Cupric chlorideimpregnated sorbents showed remarkable performance in Hg0 removal in the previous fixed-bed tests (7) conducted under typical simulated flue gases of subbituminous/lignite coals. In this study, raw activated carbons were chemically promoted with three different concentrations of cupric chloride and tested in the entrained-flow system. One (6.5% CuCl2-AC) of the cupric chloride-impregnated carbons was compared for its test results with raw (Darco Hg) and brominated (Darco Hg-LH) activated carbons in Figures 3, 4, and 5. As shown in Figure 3, raw activated carbon did not demonstrate good performance in Hg adsorption efficiency under air flow in the entrained-flow system. The Darco HgLH showed Hg adsorption performance comparable to that of 6.5% CuCl2-AC. On the basis of these results and surface areas of both sorbents shown in Table 1, the relationship between the Hg adsorption efficiency and the total surface area of the sorbent injected during each test is presented in Figure 4. The relative standard deviation of 7% in Hg adsorption efficiency determined based on the sum of Hgads, Hg2+out, and Hg0out from the previous replicate tests shown VOL. 43, NO. 8, 2009 / ENVIRONMENTAL SCIENCE & TECHNOLOGY

9

2959

FIGURE 3. Hg adsorption efficiencies of Darco Hg-LH, 6.5% CuCl2-AC, and Darco Hg.

FIGURE 4. Relationship between Hg adsorption efficiency and total surface area of injected sorbent. in Figure 2 was applied to all data points for error bars shown in the figure. The Darco Hg-LH showed unnoticeable Hg adsorption performance compared with 6.5% CuCl2-AC with respect to surface area. The Hg adsorption efficiencies of both sorbents increased with an increase in their total surface areas available in the entrained-flow system, suggesting that the Hg adsorption is limited by the readsorption of the resultant oxidized mercury generated from the reaction between Hg(0) and CuCl2. In Figure 5, the effluent oxidized mercury results obtained with Darco Hg-LH and 6.5% CuCl2AC showed that 6.5% CuCl2-AC had slightly higher effluent oxidized mercury. Therefore, except for slightly higher Hg0 oxidation, the 6.5% CuCl2-AC sorbent demonstrated very similar performance in mercury control to Darco Hg-LH. In addition, significant amounts of Hg0 were found to be oxidized by the sorbents, but the resultant oxidized mercury was not adsorbed onto both sorbents. These results indicate again that Hg removal capability of both sorbents is limited by the readsorption of the resultant oxidized mercury. 2960

9

ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 43, NO. 8, 2009

FIGURE 5. Percentages of effluent oxidized mercury determined from entrained-flow tests with Darco Hg-LH and 6.5% CuCl2-AC.

FIGURE 6. Hg adsorption efficiencies of cupric chlorideimpregnated activated carbons with different cupric chloride loadings obtained from digestion of their spent sorbents. Hg adsorption efficiencies of CuCl2-AC sorbents obtained from the digestion analysis of their spent sorbents after each test are also shown in Figure 6. Similar to the test results of Darco Hg-LH and 6.5% CuCl2-AC, Hg adsorption efficiency increased almost in proportion to the sorbent injection rate for all CuCl2-AC sorbents, and very slightly increased with an increase in CuCl2 loading in each sorbent. This suggests that mercury adsorption onto CuCl2-AC sorbents is limited by external mass transfer resistance due to the low mercury concentration rather than limited by the CuCl2 loading in the sorbent. Meanwhile, effluent oxidized mercury was slightly dependent on CuCl2 loading, but was not strongly dependent on sorbent injection rate as shown in Figure 7. This indicates that only the raw carbon surfaces may be responsible for the readsorption of the resultant oxidized mercury and Hg(0) oxidation may be independent of surface area above a certain critical value. Figure 8 illustrates Hg0removal totals (the sum of Hg adsorption and effluent

FIGURE 7. Percentages of effluent oxidized mercury determined from entrained-flow tests with cupric chloride-impregnated activated carbons.

FIGURE 9. Entrained-flow test results of (a) 6.5% CuCl2-AC and (b) Darco Hg-LH at an injection rate of approximately 30 mg/m3 in air without and with 1500 ppmv SO2.

FIGURE 8. Hg0-removal totals of brominated (Darco Hg-LH) and cupric chloride impregnated (CuCl2-ACs) activated carbons. oxidized mercury) obtained with brominated and cupric chloride-impregnated sorbents in the entrained-flow system. It clearly shows that the Hg0-removal total increases with an increase in CuCl2 loading in sorbent due to the Hg0 oxidation capability of CuCl2. In addition, the entrained-flow results of Darco Hg-LH show a similar trend as full-scale test results obtained from the literature (1). Additional entrained-flow tests were conducted in air with 1500 ppmv SO2 to find the effects of other flue gas components on the performance of Darco Hg-LH and CuCl2AC sorbents. These tests also examine possible biases by the presence of Cl2 in the KCl solution used for measuring effluent oxidized mercury as reported by Cauch, et al. (16). Figure 9 compares entrained-flow results obtained in air to those in air only with SO2 for Darco Hg-LH and 6.5% CuCl2-AC. The efficiency values were determined from two replicate tests

for each sorbent. As shown in the figure, only very slight performance decrease by SO2 injection was found for both sorbents, although SO2 is known as the most applicable inhibitor of Hg0 oxidation and adsorption for carbon-based sorbents (8, 17). Therefore, while flue gas components may have relatively minimal effects on the performance of Darco Hg-LH and CuCl2-AC sorbents, significant amounts of effluent oxidized mercury found from the tests of both sorbents suggest further investigation into the capability of fly ash to adsorb the resultant oxidized mercury using the entrainedflow system. In addition, similar amounts of effluent oxidized mercury found from both tests without and with SO2 confirm that biases by the presence of Cl2 in the KCl solution are negligible in these entrained-flow tests. As shown in the figure, 6.5% CuCl2-AC still shows slightly higher effluent oxidized mercury and Hg0-removal total than Darco Hg-LH in the tests with 1500 ppmv SO2. Considering that bromine loading in Darco Hg-LH is approximately 15% (18), CuCl2-AC sorbents have better Hg0 oxidation capability than Darco Hg-LH. In addition, CuCl2-AC sorbents will benefit from the utilization of a waste stream from the printed circuit board (PCB) industry, and would thus be environmentally beneficial to both of the utility and electronic industries. This study evaluated Hg0 oxidation and adsorption capabilities of Darco Hg-LH and CuCl2-AC sorbents in an entrained-flow system. As a result, CuCl2-AC sorbents demonstrated better performance in Hg0 oxidation than Darco Hg-LH while similar Hg adsorption performance was shown between these sorbents. In addition, both sorbents showed minimal effects of SO2 on their Hg0 oxidation and adsorption capabilities. Further studies using a fixed-bed system and X-ray absorption fine-structure spectroscopy are expected to be able to determine the reaction mechanism between CuCl2 and Hg0 and to identify the species responsible for Hg0 oxidation of CuCl2-AC sorbents. VOL. 43, NO. 8, 2009 / ENVIRONMENTAL SCIENCE & TECHNOLOGY

9

2961

Literature Cited (1) Jones, A. P.; Hoffmann, J. W.; Smith, D. N.; Feeley, T. J.; Murphy, J. T. DOE/NETL’s phase II mercury control technology field testing program: preliminary economic analysis of activated carbon injection. Environ. Sci. Technol. 2007, 41, 1365–1371. (2) Sjostrom, S.; Ebner, T.; Ley, T.; Slye, R.; Richardson, C.; Machalek, T.; Richardson, M.; Chang, R. Assessing sorbents for mercury control in coal-combustion flue gas. J. Air Waste Manage. Assoc. 2002, 52, 902–911. (3) Staudt, J. E.; Jozewicz, W. Performance and cost of mercury and multipollutant emission control technology applications on electric utility boilers, EPA-600/R-03/110; U.S. Environmental Protection Agency, National Risk Management Research Laboratory: Research Triangle Park, NC, 2003. (4) Hsi, H. C.; Chen, S.; Rostam-Abadi, M.; Rood, M. J.; Richardson, C. F.; Carey, T. R.; Chang, R. Preparation and evaluation of coal-derived activated carbons for removal of mercury vapor from simulated coal combustion flue gases. Energy Fuels 1998, 12, 1061–1070. (5) Miller, S. J.; Dunham, G. E.; Olson, E. S.; Brown, T. D. Flue gas effects on a carbon-based mercury sorbent. Fuel Process. Technol. 2000, 65-66, 343–363. (6) Serre, S. D.; Gullett, B. K.; Ghorishi, B. Entrained-flow adsorption of mercury using activated carbon. J. Air Waste Mange. Assoc. 2001, 51, 733–741. (7) Lee, J.-Y.; Ju, Y.; Lee, S.-S.; Keener, T. C.; Varma, R. S. Novel mercury oxidant and sorbent for mercury emissions control from coal-fired power plants. Water Air Soil Pollut.: Focus 2008, 8, 333–341. (8) Kilgroe, J. D.; Sedman, C. B.; Srivastava, R. K.; Ryan, J. V.; Lee, C. W.; Thorneloe, S. A. Control of mercury emissions from coalfired electric utility boilers: interim report; EPA-600/R-01/109; U.S. Environmental Protection Agency, National Risk Management Research Laboratory: Research Triangle Park, NC, 2001. (9) Pavlish, J. H.; Holmes, M. J.; Benson, S. A.; Crocker, C. R.; Galbreath, K. C. Application of sorbents for mercury control for utilities burning lignite coal. Fuel Process. Technol. 2004, 85, 563–576.

2962

9

ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 43, NO. 8, 2009

(10) Ghorishi, S. B.; Keeney, R. M.; Serre, S. D.; Gullett, B. K.; Jozewicz, W. S. Development of a Cl-impregnated activated carbon for entrained-flow capture of elemental mercury. Environ. Sci. Technol. 2002, 36, 4454–4459. (11) Feeley, T. J.; Jones, A. P. An update on DOE/NETL’s mercury control technology field testing program; http://www.netl.doe. gov/technologies/coalpower/ewr/mercury/pubs/netl%20Hg% 20program%20white%20paper%20FINAL%20Jan2008.pdf. (12) Levin, L. Atmospheric mercury research update; 1005500; EPRI: Palo Alto, CA, 2004; http://www.epa.gov/mercury/pdfs/OAR2002-0056-2589.pdf. (13) Durham, M.; Bustard, J.; Starns, T.; Sjostrom, S.; Lindsey, C.; Martin, C.; Schlager, R. Chang, R.; Renninger, S.; Monroe, L.; Berry, M.; Johnson, D. Full-scale results of mercury control by injecting activated carbon upstream of ESPs and fabric filters. Presented at PowerGen 2003 Las Vegas, NV, December 9-11, 2003; http://www.adaes.com/documents/Pubno.03010PowerGen2003.pdf. (14) Varma, R. S.; Ju, Y.; Sikdar, S.; Lee, J.-Y.; Keener, T. C. Compositions and Methods for Removing Mercury from MercuryContaining Fluids; U.S. Patent 0140940, 2007. (15) ASTM Method D6784-02. Standard test method for elemental, oxidized, particle-bound and total mercury in flue gas generated from coal-fired stationary sources (Ontario Hydro Method); ASTM International: West Conshohocken, PA, 2006. (16) Cauch, B.; Silcox, G. D.; Lighty, J. S.; Wendt, J. O.; Fry, A.; Senior, C. L. Confounding effects of aqueous-phase impinger chemistry on apparent oxidation of mercury in flue gases. Environ. Sci. Technol. 2008, 42, 2594–2599. (17) Olson, E. S.; Laumb, J. D.; Benson, S. A.; Dunham, G. E.; Sharma, R. K.; Mibeck, B. A.; Miller, S. J.; Holmes, M. J.; Pavlish, J. H. Chemical mechanisms in mercury emission control technologies. J. Phys. IV France 2003, 107, 979–982. (18) Norit MSDS No. 109. Material Safety Data Sheet of Darco HgLH; Norit Americas Inc.: Marshall, TX, 2007; http://www. norit-americas.com/pdf/MSDS109_REV03.pdf.

ES801943T