and NO on Homogeneous Hg Oxidation by Cl - American

Foster Wheeler North America Corporation, 12 Peach Tree Hill Road, LiVingston, New ... nitric oxide (NO) does not have a significant impact on Hg oxid...
0 downloads 0 Views 167KB Size
Download PDF
1068

Energy & Fuels 2006, 20, 1068-1075

Effects of H2O, SO2, and NO on Homogeneous Hg Oxidation by Cl2 Hans Agarwal and Harvey G. Stenger* Department of Chemical Engineering and Energy Research Center, Lehigh UniVersity, 117 ATLSS DriVe, Bethlehem, PennsylVania 18015

Song Wu and Zhen Fan Foster Wheeler North America Corporation, 12 Peach Tree Hill Road, LiVingston, New Jersey 07039 ReceiVed NoVember 22, 2005. ReVised Manuscript ReceiVed February 13, 2006

Several researchers have determined that water (H2O) and sulfur dioxide (SO2) in a flue gas stream have an impact on the amount of elemental mercury (Hg0) that is homogeneously oxidized by a chlorine-containing species. Generally, it is concluded that H2O inhibits Hg oxidation by chlorine (Cl2). However, doubt remains as to whether SO2 promotes or inhibits Hg oxidation. Further, most published results seem to indicate that nitric oxide (NO) does not have a significant impact on Hg oxidation. This paper will present data taken in a laboratory-scale apparatus designed to test these observations. In this work, Cl2 is intentionally added to a synthetic flue gas stream containing known amounts of elemental mercury. This gas blend is similar to a flue gas obtained by burning Powder River Basin coal in a pulverized coal fired power plant and is subject to a time-temperature profile similar to a power plant. The results obtained show that H2O, SO2, and NO all have an inhibitory effect on the homogeneous oxidation of Hg by Cl2. Further, the presence of H2O increases the inhibitory effect of SO2 and NO. Two new reactions are proposed to explain these results, in which SO2 and NO react with Cl2. The consequences of these reactions are a reduction in the oxidative interactions that take place between Hg and Cl2, thus decreasing the amount of Hg oxidation that occurs.

Introduction In 1997, the United States Environmental Protection Agency (USEPA) submitted a mercury study report to the U.S. Congress. They projected that of the 158 tons of mercury being released into the environment, 48 tons was derived from coal-fired combustion sources.1,2 On March 15, 2005, the USEPA issued a federal rule to permanently cap and reduce mercury (Hg) emissions from coal-fired power plants. Upon complete implementation, Hg emissions will be reduced by 21% by 2010, and by 70% by 2018. Because of these impending reductions, it is important to find efficient means to remove Hg from power plant flue gases. Mercury in the flue gas is most commonly classified in three forms: elemental mercury (Hg0), oxidized mercury (Hg2+), and particulate bound mercury (HgP). Particulate bound mercury is usually trapped by ash collection devices within power plants, such as electrostatic precipitators (ESP), or baghouses. Elemental mercury is relatively inert and difficult to capture because of its nonreactivity. It is also volatile at high temperatures and insoluble in water. Conversely, oxidized mercury is very water soluble and has an affinity for adsorbing onto particulate matter such as fly ash or on metal surfaces in the duct. As a result of these physical and chemical properties of Hg0 and Hg2+, the removal of mercury is enhanced when elemental Hg is converted to its oxidized form. * Corresponding author. Telephone: (610) 758-4791. E-mail: HGS0@ LEHIGH.edu. (1) Mercury Study Report to Congress, Volume I: ExecutiVe Summary; Report No. EPA-452/R-97-003; U S. Environmental Protection Agency, Office of Air Quality Planning and Standards and Office of Research and Development, U.S. Government Printing Office: Washington, DC, 1997. (2) Brown, T. D.; Smith, D. N.; Hargis, R. A., Jr; O’dowd, W. J. J. Air Waste Manage. Assoc. 1999, June, 1-97.

Temperature and flue gas components, such as SO2, Cl2 or HCl, and H2O have strong influences on mercury speciation.3,4 Thermodynamic equilibrium calculations performed by Frandsen et al. predict that all gas-phase mercury in a coal combustion flue gas should exist as mercuric chloride (HgCl2) at temperatures below 450 °C.5 However, actual mercury speciation data collected by EPA’s Information Collection Request (ICR) database shows that, at even lower temperatures, elemental mercury still exists in the flue gas.6 Further studies have also shown that the distribution of mercury species is strongly dependent on the type of coal being burnt, such as lignite, subbituminous, or bituminous types of coal. Different coal types have different chlorine contents, but also vary in the amount of sulfur and nitric oxides produced.7 Laudal et al. performed several bench-scale experiments to determine the effects of various flue gas components on Hg oxidation.8 Experimental work was performed by blending individual flue gas components at a maximum temperature of 175 °C. They did not use a continuous emission monitor (CEM) to measure mercury concentrations, but used the Ontario Hydro method instead. They found that 10 parts per million by volume (ppmv) Cl2 in a gas stream consisting of N2, O2, CO2, and H2O, (3) Senior, C. L.; Sarofim, A. F.; Zeng, T.; Helble, J. J.; Mamani-Paco, R. Fuel Process. Technol. 2000, 63 (2-3), 197-213. (4) Galbreath, K. C.; Zygarlicke, C. J. EnViron. Sci. Technol. 1996, 30, 2421-2426. (5) Frandsen, F.; Dam-Johansen, K.; Rasmussen, P. Prog. Energy Combust. Sci. 1994, 20, 115-138. (6) Cao, Y.; Duan, Y.; Kellie, S.; Li, L.; Xu, W.; Riley, J. T.; Pan, W.P.; Chu, P.; Metha, A. K.; Carty, R. Energy Fuels 2005, 19, 842-854. (7) Kellie, S.; Cao, Y.; Duan, Y.; Li, L.; Chu, P.; Mehta, A.; Carty, R.; Riley, J. T.; Pan, W.-P. Energy Fuels 2005, 19, 800-806. (8) Laudal, D. L.; Brown, T. D.; Nott, B. R. Fuel Process. Technol. 2000, 65-66, 157-165.

10.1021/ef050388p CCC: $33.50 © 2006 American Chemical Society Published on Web 04/25/2006

Effects of H2O, SO2, and NO on Hg Oxidation by Cl2

with elemental Hg, resulted in 84.8% Hg oxidation. They also found that certain flue gas components, such as 50 ppmv HCl, 1500 ppmv SO2, or 630 ppmv NOx, independently did not cause any Hg oxidation. However, similar to Kilgroe, they found that 1500 ppmv SO2 did have an inhibitory effect on Hg oxidation, which when added with 10 ppmv Cl2 and 50 ppmv HCl resulted in less than 2% Hg oxidation. Their results also suggest that 630 ppmv NOx may inhibit Hg oxidation, which when added with 10 ppmv Cl2 and 50 ppmv HCl resulted in approximately 78.7% Hg oxidation. Kilgroe et al. showed a negative influence of water and sulfur on mercury oxidation.9 Experiments were performed at temperatures above 755 °C by blending gases to form a flue gas stream. They found that increasing the HCl content of the flue gas resulted in an increased amount of mercury oxidation, where 50 ppmv HCl resulted in approximately 7% Hg oxidation, while 200 ppmv HCl resulted in approximately 27% Hg oxidation. The addition of 500 ppmv SO2 in the system resulted in a decreased amount of mercury oxidation; 50 ppmv HCl (with SO2) resulted in approximately 3% Hg oxidation, while 200 ppmv HCl (with SO2) resulted in approximately 15% Hg oxidation. Addition of H2O further decreased the amount of mercury oxidation, where 50 ppmv HCl (with SO2 and H2O) resulted in approximately 1% Hg oxidation, while 200 ppmv HCl (with SO2 and H2O) resulted in approximately 10% Hg oxidation. Norton et al. reported experimental results similar to those of Laudal et al., where the individual impact of NO and NO2 was studied in greater detail.10 Experiments were performed by blending gases at temperatures between 120 and 180 °C. They suggested that NO2, HCl, and SO2 seemed to promote Hg oxidation while NO inhibited Hg oxidation. However, the experiments were performed in the presence of fly ash, where the reaction mechanism becomes more complicated and difficult to interpret. Studies conducted by the Mercury Emissions Monitoring Lab at Western Kentucky University (WKU) have shown a relationship between coal chlorine concentration and mercury oxidation.7 Experimental work was performed in a 100-MW utility boiler, and temperatures ranged from above 750 °C in the boiler to 365 °C at the air preheater (APH) inlet and 155 °C at the electrostatic precipitator (ESP) outlet.6 Higher chlorine concentration within the flue gas stream resulted in higher mercury oxidation, where Hg2+ was the dominant species formed. In one coal sample, the reported chlorine and Hg concentrations were 1006 and 0.12 parts per million by weight (ppmw), respectively. Upon combustion of the coal, approximately 79% of the initial amount of Hg was detected to be in oxidized form. In a second coal sample, the reported chlorine and Hg concentrations were 1449 and 0.10 ppmw, respectively. Upon combustion of the coal, approximately 89% of the initial Hg in the coal was detected to be in oxidized form. These Hg measurements were taken at the exit of the ESP, where temperatures are at 160 °C. Because of its affinity for adsorbing onto particles, high quantities of Hg2+ were found adsorbed on the fly ash in their system. They also concluded that sulfur has a tendency to encourage the oxidation of Hg, but that it also inhibited the adsorption of Hg2+ onto the fly ash. Further experimental results suggest that both HCl and SO2 may directly (9) Kilgroe, J. D.; Sedman, C. B.; Srivastava, R. K.; Ryan, J. V.; Lee, C. W.; Thorneloe, S. A. Interim Report No. EPA-600/R-01-109; U.S. Environmental Protection Agency, U.S. Government Printing Office: Washington, DC, December 2001. (10) Norton, G. A.; Yang, H.; Brown, R. C.; Laudal, D.; Dunham, G. E.; Erjavec, J. Fuel 2002, 82, 107-116.

Energy & Fuels, Vol. 20, No. 3, 2006 1069

Figure 1. Schematic of testing facility to perform Hg oxidation tests. Base gas compositions and temperature profile are shown in Tables 1 and 2.

affect the mercury oxidation mechanism, which may be a cause for the high amounts of Hg2+ detected in their work. Also, coal combustion results in fly ash formation, which results in a more complex problem of heterogeneous Hg oxidation. In summary, it was found that H2O inhibits the oxidation of Hg by a chlorine species. The question remains as to whether Cl2 or HCl is the primary chlorine species responsible for Hg oxidation. Opinions also differ as to whether SO2 encourages or inhibits Hg oxidation. NO seems to inhibit Hg oxidation, but the published work was done in the presence of particulate matter and not in a homogeneous environment. This paper will present data obtained from a laboratory-scale apparatus, where the effects of H2O, SO2, and NO on Hg oxidation are studied in systematic detail. Most of the experimental work done by other researchers was performed at temperatures greater than 700 °C or below 200 °C. Additionally, the individual components of the flue gas were studied in the absence of other major flue gas components such as H2O and CO2. This work focuses on temperatures between 550 and 180 °C and attempts to provide a comprehensive view of the cumulative effects of each of the major flue gas components on Hg oxidation. Testing Facility Figure 1 is a schematic of the testing facility that was used to perform the Hg oxidation tests. The system was designed and built such that two different types of mercury experimentation could be performed: homogeneous Hg oxidation in the gas phase and heterogeneous Hg oxidation with sorbent adsorption work. This paper focuses on homogeneous Hg oxidation. The various components of the flue gas are metered, blended, and heated to the desired temperature. Table 1 shows the base case concentrations of the various components of the flue gas used in this work. The specifications closely follow a flue gas stream derived from the combustion of Powder River Basin (PRB) coal. The following subsections describe the components of the testing facility in greater detail.

1070 Energy & Fuels, Vol. 20, No. 3, 2006

Agarwal et al.

Table 1. Summary of Components and Composition component

composition

nitrogen (N2) oxygen (O2) carbon dioxide (CO2) water (H2O) sulfur dioxide (SO2) nitric oxide (NO) carbon monoxide (CO) chlorine (HCl or Cl2) elemental mercury (Hg)

70% 3.5% 13.5% 13% 370 ppmv 170 ppmv 300 ppmv 100 ppmv 10 µg/m3

Table 2. Values of Temperature versus Residence Time Profiles, and Respective Locations on the Experimental Apparatus location

temp (°C)

cumul residence time (s)

exit R1 exit R2 exit HX1 baghouse inlet

538 389 177 150

0.75 2.4 2.8 3.7

Air Preheater (APH). The APH was designed to preheat the bulk gas blend, which consists of nitrogen (N2), carbon dioxide (CO2), and air. The APH is a 20 in. long, 1 in. diameter stainless steel tube heated externally with two 425 W ceramic heating elements. The tube is filled with stainless steel packing to enhance mixing and heat transfer. The flow rate of the bulk gases through the APH used in this work was between 70 and 86 standard liters per minute (slpm). The exit gas temperature was approximately 315 °C. Steam Preheater (SPH). The SPH is made with the same size tube and heater as the APH, and is used to vaporize liquid water (H2O) to superheated steam. In this work, a liquid water flow rate of 10 milliliters per minute (mlpm) was pumped into the bottom of the SPH. The exiting steam temperature was approximately 300 °C. Final Preheater (R1). The preheated bulk gases and steam are mixed with the trace gases and elemental mercury, before the gas blend enters the R1. The trace gases include sulfur dioxide (SO2), nitric oxide (NO), and carbon monoxide (CO). The R1 is 44 in. long and made of a 1.5 in. schedule 40 SS310 pipe. It is filled with stainless steel material to enhance heat transfer and mixing. The pipe is heated externally with four 1295 W ceramic heating elements. In this work, the total flow rate of the gas blend entering the R1 was approximately 87 slpm. The gas temperature at the exit of the R1 was controlled at 540 °C for most experiments; however, the heater is capable of heating the gas to 900 °C when needed. Reactor (R2). R2 is an inconel shell and tube heat exchanger. The tube has a length of 38 in. and an outer diameter of 4.5 in. The shell is 34 in. long and has an outer diameter of 6.625 in. The reacting gases flow through the tube, and cooling gas flows in the shell. The residence time of the gas in R2 is approximately 2 s. In most experiments in this work, the gas temperature at the exit of R2 was controlled at 389 °C. The oxidant used in this work is chlorine gas (Cl2). A cylinder of 1.0% Cl2 in N2 was used as the chlorine source. This gas was metered and injected into the heated gas stream between R1 and R2. The same temperature profile, as shown in Table 2, was used in all three experiments. The concentration of each flue gas component used in this experimental work is shown in Table 1, except for the concentration of Cl2 which changes with each experiment. Heat Exchanger (HX1). HX1 is also an inconel shell and tube heat exchanger. The pipe section has a length of 26.75 in. with an outer diameter of 1.9 in. The shell section has a length of 22 in. and an outer diameter of 3.5 in. The gas temperature

at the exit of HX1 for most experiments was controlled at 177 °C. The gas was sampled for Hg at the exit of HX1 through a 1/4 in. port. The temperature of the sample line was fixed at approximately 150 °C. Mercury Generation System (CAVKIT). Elemental mercury present in a reservoir within the CAVKIT (PS Analytical) vaporizes upon heating. A small flow of nitrogen, called “reservoir flow”, passes over the reservoir and carries the vaporized elemental Hg. This stream is then diluted by a larger nitrogen stream before flowing to the main apparatus. For example, a temperature of 45 °C, a reservoir flow of 18 mlpm, and a dilution flow of 5 slpm, when mixed with the other gases in the apparatus, will result in an approximate elemental Hg concentration of 10 µg/m3. Sorbent Injection Section. The gas exits the bottom of HX1, makes a 180° turn, and enters the sorbent injection system with upward flow. In this section, sorbent can be injected into the flue gas stream to study its ability to capture mercury by adsorption. The section is an 89 in. tall, 3 in. diameter pipe, heated externally with a heating coil. The size of the pipe provides a 1-2 s residence time for capture. The sorbent/flue gas mixture then enters a baghouse unit where the sorbent is trapped and can be removed for Hg content analysis. The flue gas stream exiting the baghouse is cooled in a condenser where most of the water is condensed and drained. The dry flue gas passes through a carbon trap where any acid gases are removed and finally flows through a gas totalizer where the volumetric flow rate of the total dry gas is measured. An important feature of the apparatus is its ability to follow a typical power plant time-temperature profile. Figure 2 shows the close comparison of a power plant temperature profile with the profile obtained in this apparatus when operated at 87 slpm. Table 2 lists these values of temperature versus residence time as obtained in this apparatus. Semicontinuous Emission Monitoring (SCEM) System. The SCEM device used to measure elemental and total Hg is the PS Analytical Sir Galahad 10.525. It uses a gold trap, which collects mercury from a known volume of gas at low temperature and then desorbs it by heating. The desorbed Hg is analyzed by an atomic fluorescent detector. The SCEM system has a gas conditioning system, which splits the sample stream into two streams. One stream passes through an impinger containing a solution of sodium hydroxide (NaOH) and potassium chloride (KCl). The NaOH removes any traces of acid, while the KCl removes any oxidized Hg species. As a result, only the elemental Hg species present in the gas stream leaves this scrubber and goes to the analyzer. The second stream passes through a second impinger, which contains a solution of NaOH and tin(II) chloride (SnCl2). The SnCl2 reduces the oxidized Hg in the gas stream to elemental Hg. The amount of Hg measured from this stream is the total Hg present in the system. The difference in Hg concentration value between the two impinger units is the amount of oxidized Hg present in the stream. A stream selector system, which is controlled by the PSA software, switches the gas feed to the analyzer between the total Hg stream and the elemental Hg stream. The SCEM device is regularly calibrated. A known volume of air with a known concentration of elemental, vapor-phase Hg at a measured ambient temperature is injected into the gold trap via a pressure-lock syringe. The syringe samples an air space, via a septum, above liquid mercury in a closed container. Because the concentration in the air space is fixed by temperature, the amount of elemental Hg injected into the SCEM can be varied by varying the sample volume. The analyzer records

Effects of H2O, SO2, and NO on Hg Oxidation by Cl2

Energy & Fuels, Vol. 20, No. 3, 2006 1071

Figure 2. Comparing temperature versus residence time profiles of a typical power plant and the experimental apparatus. The dashed line represents the profile of a power plant; the solid line represents the profile of the apparatus in this experimental work.

the peak height and area as detected by the fluorescent detector, for each sample. The PSA software plots the peak height or area (as specified by the user) versus the amount of elemental Hg injected. Typically, four mercury amounts were used for the calibration of these experiments: 0, 1, 2, and 5 nanograms (ng). When unknown gases are analyzed, typically a sample volume of 250 mL is used, which if it contained 5 ng would correspond to a concentration of 20 µg/m3. Results There are two major goals for the experiments in this work. First, the experiments are conducted at temperatures different from those of most other researchers. For example, Kilgroe et al. performed their analysis of the effects of SO2 and H2O on Hg oxidation at temperatures greater than 755 °C. According to the model by Frandsen, most of the Hg will exist at an elemental state at these temperatures. Conversely, Laudal et al. performed experiments at temperatures below 200 °C, and according to Frandsen, most of the Hg will exist in its oxidized form. The temperature profile chosen for this work was between 550 and 200 °C. The second goal of these experiments is to measure the effects of H2O, SO2, and NO on the oxidation of Hg by Cl2. Other researchers agree that H2O has an inhibitory effect on Hg oxidation, but questions remain concerning the effects of SO2 and NO. This paper will attempt to answer some of these questions. Experiment 1. The first experiment performed studied the effects of the gas components on the reaction of Hg with Cl2. The system was heated with gas flowing until the temperature profile (in Table 2) was attained. Initially the total gas flow through the system was 87 slpm and consisted of 96.5% N2, 3.5% O2, and 10 µg/m3 elemental Hg. Upon reaching the desired temperature, Cl2 was injected between the R1 and R2 at a rate that would give a 2 ppmv concentration. The SCEM monitored the change in Hg concentration. Once steady Hg values were achieved, the Cl2 flow was discontinued and the system was allowed to reequilibrate. Figure 3 shows the results of this experiment, where elemental Hg concentration (in µg/m3) is

plotted against the duration of the entire experiment. In the first stage, labeled 1a, 71.7% of the initial amount of Hg was oxidized. Next, CO2 was added and the gas flowing through the system was adjusted accordingly to give 83% N2, 3.5% O2, 13.5% CO2, and 10 µg/m3 Hg0. At steady temperatures, 2 ppmv Cl2 was again injected and changes in Hg were measured. This process was repeated for SO2, NO, CO, and H2O, respectively. Table 3 shows the composition of the gas used in each stage of this experiment, with the percent of Hg oxidized for each stage. Figure 3 also shows that the apparatus has no effect on Hg oxidation by itself. The theoretical concentration of Hg injected into the apparatus was approximately 12 µg/m3. As seen in Figure 3, stages 1a, 1b, 1d, and 1e show that the concentration of Hg before the addition of Cl2 was approximately 12 µg/m3. In the cases of 1c and 1f, the concentration of mercury is slightly different from 12 µg/m3 due to drift in flows, but within experimental error. It is clear from the results in Table 3 that Cl2 oxidizes Hg effectively. Stages 1a and 1b show that over 70% of Hg was oxidized. Upon the addition of SO2 (stage 1c), the percent of Hg oxidized dropped from over 70% to 52.3%. This agrees with the results obtained by Kilgroe and Laudal, that SO2 has an inhibitory effect on Hg oxidation. The addition of NO into the gas stream (stage 1d) further reduced the percent of Hg oxidized, from 52.3% to 45.7%. Norton et al. showed similar trends in their work. The addition of CO into the gas stream did not have an impact on the observed Hg oxidation. In stage 1f, the addition of water completely removes the oxidation effect of Cl2. Also, the addition of H2O results in a huge spike of mercury emitted from the apparatus. It is speculated that some elemental and oxidized Hg accumulates on the inner walls of the apparatus over the duration of the experiment, and the presence of H2O in the gas stream caused an immediate desorption of this elemental Hg. This effect does not last long, since within a few minutes the concentration of elemental Hg returns to approximately 10 µg/m3. Following the spike, the addition of 2 ppm Cl2 showed no measurable oxidation effect. This indicates that 2 ppmv Cl2 is

1072 Energy & Fuels, Vol. 20, No. 3, 2006

Agarwal et al.

Figure 3. Plot of elemental Hg concentration (µg/m3) with respect to time for experiment 1. Each stage of the experiment is labeled along with the respective gas composition and the period of addition and removal of Cl2. Table 3. Composition of Gas Used in Each Stage of Experiment 1, and Elemental Hg Oxidation Results components in flue gas stream stage

elem Hg (µg/m 3)

N2 (%)

O2 (%)

CO2 (%)

1a 1b 1c 1d 1e 1f

11.6 13.4 14.9 12.7 11.7 10.0

96.5 83.5 83.5 83.5 83.5 70.0

3.5 3.5 3.5 3.5 3.5 3.5

13.5 13.5 13.5 13.5 13.5

SO2 (ppmv)

370 370 370 370

Table 4. Composition of Gas Used in Each Stage of Experiment 2, and Elemental Hg Oxidation Results components in flue gas stream stage

elem Hg (µg/m 3)

N2 (%)

O2 (%)

H2O (%)

Cl2 (ppmv)

elem Hg oxidn (%)

2a 2b 2c 2d

11.6 12.7 11.6 11.4

83.5 83.5 83.5 83.5

3.5 3.5 3.5 3.5

13 13 13 13

2.0 12.4 30.9 52.5

0 83 88 92

not able to overcome the oxidation canceling effects of H2O. Experiments 2 and 3 were then designed and run to determine if increasing the Cl2 concentration could overcome the effects of water. Experiment 2. In this experiment, the gas flow through the system was again 87 slpm and the gas blend consisted of 83% N2, 3.5% O2, and 13.5% H2O. No CO2, SO2, NO, or CO was added to the gas. The concentration of Cl2 was increased from 2 to 52 ppmv, and the amount of Hg oxidized was measured. The same temperature profile was used as shown in Table 2. Table 4 shows the composition of the gas used in each stage of this experiment, along with the percent of Hg oxidized for each stage. In the first stage of experiment 2 (2a), the same result was obtained as in experiment 1f, where 2 ppmv Cl2 had no effect on Hg oxidation. When 12 ppmv Cl2 was added (stage 2b), 82.7% of elemental Hg was oxidized. Further increases of Cl2 to 31 and 52 ppmv resulted in smaller increases in Hg oxidation of 88.3% and 92.4%, respectively. Figure 4 summarizes the results of this experiment.

NO (ppmv)

170 170 170

CO (ppmv)

300 300

H2O (%)

Cl2 (ppmv)

elem Hg oxidn (%)

13

2.0 2.0 2.0 2.0 2.0 2.0

72 77 52 46 47 0

Experiment 3. This experiment is the same as experiment 1, except that 13% water was added and 27.5 ppmv Cl2 was injected into the heated gas stream, rather than 2 ppmv in experiment 1. This amount of Cl2 was chosen in the expectation that the oxidation rates would be strong, allowing the effects of SO2 and NO to be observed. Table 5 shows the composition of the gas used in each stage of this experiment along with the percent of Hg oxidized for each stage. The results of stages 3a and 3b (effect of CO2) are similar to those of stages 1a and 1b in experiment 1 in that the amount of Hg oxidized is high, but the presence of CO2 had little effect. However, in stage 3c, when SO2 was added, the percent Hg oxidized dropped to 56.7%. When NO was added into the gas (stage 3d), the observed Hg oxidation decreased from 56.7% to 45.6%. Stage 3e showed that the addition of CO did not have an impact on Hg oxidation. Figure 5 summarizes the results of this experiment and compares them to the results of experiment 1. Discussion It is known that chlorine compounds can oxidize elemental Hg effectively. However, it has not been shown whether Cl2 or HCl is the primary species responsible for this oxidation. For instance, Laudal et al. injected Cl2 into their experimental work and concluded that Cl2, and not HCl, was responsible for Hg oxidation. Conversely, Kilgroe injected HCl in their system and found that, at temperatures above 750 °C, Hg oxidation occurred very effectively. One way to explain this discrepancy is by

Effects of H2O, SO2, and NO on Hg Oxidation by Cl2

Energy & Fuels, Vol. 20, No. 3, 2006 1073

Figure 4. Plot of percent Hg oxidation versus concentration of chlorine (ppmv). See Table 4 for the gas composition used in each stage of this experiment.

Figure 5. Plot of percent Hg oxidation versus stage of experiments 1 and 3. See Tables 3 and 5 for the gas composition used in each stage of these two experiments. Table 5. Composition of Gas Used in Each Stage of Experiment 3, and Elemental Hg Oxidation Results components in flue gas stream stage

elem Hg (µg/m 3)

N2 (%)

O2 (%)

CO2 (%)

3a 3b 3c 3d 3e

10.5 10.6 10.3 10.6 10.8

83.5 70.0 70.0 70.0 70.0

3.5 3.5 3.5 3.5 3.5

13.5 13.5 13.5 13.5

SO2 (ppmv)

370 370 370

assuming the reaction of Cl2 and water to form HCl and O2 is reversible.

Cl2(g) + H2O(g) T 2HCl(g) + (1/2)O2(g)

(1)

NO (ppmv)

170 170

CO (ppmv)

H2O (%)

Cl2 (ppmv)

elem Hg oxidn (%)

300

13 13 13 13 13

27.5 27.5 27.5 27.5 27.5

90 90 57 46 45

Thermodynamic equilibrium calculations for this reaction at the conditions used in experiment 3 show that, at temperatures below 475 °C and for typical O2 and H2O concentrations, Cl2 dominates HCl and, at temperatures above 475 °C, HCl

1074 Energy & Fuels, Vol. 20, No. 3, 2006

Agarwal et al.

Figure 6. Thermodynamic plot of chlorine species concentration versus temperature. The concentration of O2 and H2O used in the calculations were 3.5% and 13.5%, respectively. Initial concentration of Cl2 used was 27.5 ppmv.

dominates over Cl2. Figure 6 shows a plot of equilibrium Cl2 and HCl concentrations with respect to temperature. The constants for the reaction were calculated using the enthalpy and Gibbs free energy of the reaction. The concentrations of water and O2 were constant at 13.5% and 3.5%, respectively. In this work, Cl2 was injected at 538 °C. According to Figure 6, some of the Cl2 injected would have converted to HCl, but at temperatures of 538 °C and below, Cl2 is the dominant species present in the gas stream. Figure 6 also helps explain the results obtained by Kilgroe and Laudal. Kilgroe performed experimental work at temperatures above 700 °C, and according to Figure 6, HCl is the dominant chlorine species present which would most likely oxidize elemental Hg. Conversely, Laudal performed experimental work at temperatures below 200 °C, and according to Figure 6, Cl2 would be the dominant chlorine species present and would most likely oxidize elemental Hg. As a result of electronic structure, Cl2 is a much stronger oxidizing agent than HCl. The ability of the chlorine atom to accept a valence electron from elemental mercury is significantly lower for HCl since H has already contributed an electron to the outer shell of Cl. This makes the chlorine atom in HCl less reactive than Cl2, which has a covalent structure that permits its outer shell to be available to those valence electrons available from Hg. Figure 4 shows that a critical concentration of Cl2 exists between 2 and 20 ppmv, where it overcomes water inhibition and rapidly oxidizes mercury. Beyond 50 ppmv, mercury oxidation reaches above 95%, and it is assumed that at excessively high Cl2 concentrations the percent of Hg oxidation will eventually reach 100%. In experiments 1 and 3 (with and without water), the addition of SO2 resulted in decreasing Hg oxidation. Comparing stages 1b and 1c in Table 3, when 370 ppmv SO2 was added, the observed Hg oxidation dropped from 76.9% to 52.3%. When 370 ppmv SO2 was added in experiment 3, comparing stages 3b and 3c in Table 5, Hg oxidation dropped from 90.0% to 56.7%. This indicates that H2O causes SO2 to have a stronger inhibition on Hg oxidation.

The inhibitory effect of SO2 is possibly due to the effect of SO2 reacting irreversibly with Cl2. The concentrations of SO2 in experiments 1 and 3 are in greater excess of Cl2, suggesting that the following reaction may be occurring:

SO2(g) + Cl2(g) f SO2Cl2(g)

(2)

The reaction of SO2 and Cl2 to form SO2Cl2 is slightly favorable at temperatures near the end of a typical power plant. For example, at 100 °C the equilibrium constant for the reaction SO2 + Cl2 T SO2Cl2 is 102 atm-1. A rough equilibrium extent calculation using the RGIBBS reactor model in Aspen Plus shows that, for a partial pressure of SO2 equal to 0.0002 atm (200 ppm), a chlorine concentration of 2 ppm, and a temperature of 100 °C, up to 5% of the added chlorine could be scavenged by SO2. This does not cause enough of an effect to account for the data presented in this work. The exact analysis of the equilibrium extent of these reactions is planned as part of future work. This rough estimate is only to indicate that the reaction of Cl2 with SO2 is feasible and is potentially a reason for decreased mercury oxidation when SO2 is added. Similar to SO2, the addition of NO also resulted in decreasing Hg oxidation. Comparing stages 1c and 1d in Table 3, when 170 ppmv NO was added, the observed Hg oxidation dropped from 52.3% to 45.7%. In experiment 3, when 170 ppmv NO was added, comparing stages 3c and 3d in Table 5, the observed Hg oxidation dropped from 56.7% to 45.6%. Similar to SO2, NO inhibits Hg oxidation and the presence of H2O causes NO to exhibit a stronger inhibitory effect on Hg oxidation. The effect is possibly due to a reaction between NO and Cl2 which would be

2NO(g) + Cl2(g) f 2NOCl(g)

(3)

The reaction of NO and Cl2 to form NOCl is thermodynamically favorable at temperatures near the end of a typical power plant. For example, at 200 °C the equilibrium constant for the reaction NO + 0.5Cl2 T NOCl is 92 atm-0.5. A rough

Effects of H2O, SO2, and NO on Hg Oxidation by Cl2

calculation using Aspen Plus (similar to the previous reaction) shows that, at a partial pressure of NO equal to 0.0004 atm (400 ppm) and a concentration of Cl2 of 2 ppm, up to 75% of the added chlorine could be scavenged by NO at 200 °C. The exact analysis of the equilibrium of these reactions is planned as part of future work. This rough estimate is only to indicate that the reaction of Cl2 with NO is feasible and is potentially a reason for decreased mercury oxidation when NO is added. The presence of CO2 did not have a significant effect on Hg oxidation. Comparing stages 1a and 1b, the percent of Hg oxidation increased by 5%; comparing stages 3a and 3b, the percent of Hg oxidation increased by 0.5%. However, neither value is significant enough to make a conclusion about the effects of CO2. Finally, the presence of CO also did not have a significant effect on the percent of Hg oxidation. Comparing stages 1d and 1e, the percent of Hg oxidation increased by 1.3%; comparing stages 3d and 3e, the percent of Hg oxidation reduced by 1.1%. This helps draw the conclusion that CO, in the flue gas stream at these temperatures, does not have an impact on Hg oxidation. Summary and Conclusions Three experiments were performed to determine the effects of various components of coal combustion flue gas on Hg oxidation. To study Hg oxidation by chlorine, various concentrations of Cl2 were added. It was found that H2O, SO2, and NO exhibit inhibitory effects on Hg oxidation. Additionally, two reaction pathways are suggested that may cause these components to demonstrate these characteristics. The following summarize the results in this paper: (a) A testing facility was built to simulate a temperature versus residence time profile similar to that of a power plant at locations beyond the economizer.

Energy & Fuels, Vol. 20, No. 3, 2006 1075

(b) Chlorine gas (Cl2) oxidizes Hg effectively. In all these experiments, Cl2 was intentionally added into a flue gas stream and the percent of Hg oxidation was observed. (c) In a gas stream consisting of only N2, O2, and CO2, the observed Hg oxidation increased with increasing Cl2 concentration and the presence of H2O. CO2 did not have appreciable effects on Hg oxidation. (d) The addition of SO2 results in a decrease in Hg oxidation. This may be caused by the reaction of SO2 with Cl2 at these temperatures. (e) The addition of NO resulted in a further decrease in Hg oxidation. This may be caused by the reaction of NO with Cl2 at these temperatures. (f) The presence of H2O seems to cause SO2 and NO to exhibit a stronger inhibitory effect on Hg oxidation, and this is clarified by experiment 3. (g) The presence of CO in the flue gas had no effect on Hg oxidation. (h) At a fixed gas composition consisting of 83% N2, 3.5% O2, and 13.5% H2O, an increasing amount of Cl2 resulted in a higher percent of Hg oxidation. It is assumed that at high enough Cl2 concentrations the percent of Hg oxidation will eventually reach 100%. Acknowledgment. The authors extend their sincere gratitude to Ms. Sandhya Eswaran, Mr. William Betchel and Mr. Matthew Chalmers, who helped design and build the apparatus used in this experimental work. Funding for this experimental work was provided by Foster Wheeler North America Corp. Much appreciation is extended to the Energy Research Center (ERC) at Lehigh University for allowing the authors to use their facility to build the apparatus and perform all experimental work. EF050388P