ROLE OF IIC! IN ADSORPTION OF ELEMENTAL MERCURY VAPOR BY CALCIUM-BASED SORBENTS Behrooz Ghorishi ARCADIS Geraghly & Miller, Inc., P.O. Box 13109 Research Triangle Park, NC 27709 Brian K. Gullett U.S. Environmental Protection Agency National Risk Management Research Laboratory (MD-65) Research Triangle Park, NC 27711 Wojciech Jozewicz and Wojciech Kozlowski ARCADIS Gcraghty & Miller, Inc., P.O. Box 13109 Research Triangle Park, NC 27709 Abstract Elemental mercury (Hg°) capture has been mostly investigated using high surface area, expensive activated carbons. Previous investigations showed that the presence of hydrogen chloride (HC1) in the flue gas had a pronounced positive effect on the adsorption of IIg° by a lignite-coal-based activated carbon with high calcium (Ca) content. The objective of this study was to identify active sites and surface functional groups that may contribute to the adsorption of Hg° by less expensive Ca-based sorbents. The study investigated the formation of chlorine (CI) sites in Ca- based sorbents as well as their role and reactivity in the adsorption of Hg°. Initially four Ca-based sorbents were exposed to HC1 and subsequently to Hg°: calcium oxide (CaO), calcium hydroxide [Ca(OH)2], calcium carbonate (CaCOs), and calcium sulfate dihydrate (CaS04*2H20; gypsum). HCl-exposed CaS0,t-2H20 exhibited a superior IIg° sorption capability. Through comparison to anhydrous calcium sulfate (CaS04) and calcium sulfate hemihydrate (CaSOvV&feO; plaster), crystalline water molecules on the surface of CaS04*2H20 were confirmed to contribute indirectly to Hg° uptake. These surface molecules may have adsorbed HC1 through hydrogen bond formation between an oxygen atom of a crystalline water molecule and a hydrogen atom of an HC1 molecule. Two adjacent, physically adsorbed HC1 molecules could then trap an Hg° molecule through formation of a mercuric chloride (IIgCl2)-like molecule. Thus, two adsorbed CI atoms that are spaced at a distance equal to the length of an HgCl2 molecule could be regarded as an IIg° active site. Significant correlations between CI and Hg° uptake by CaS04*2H20 were noted. ------- Introduction Mercury present in hazardous/municipal waste and in coal is readily volatilized during incineration and combustion processes1. The capability of the existing flue gas desulfurization (FGD) systems to remove mercury from a given flue gas was thought to be affected by the form of mercury present. Oxidized mercury compounds such as mercuric chloride (HgCl2) can be removed from the flue gas by conventional sulfur dioxide (S02) scrubbers2. However, a major part of elemental mercury (Hg°), the most volatile form among trace metal compounds, can pass through particulate matter (PM) control devices3, if uncontrolled. Injection of dry sorbents upstream of a PM control device is one potential method for controlling mercury emissions from combustion processes. The importance of the mercury species form for the efficiency of mercury capture by dry sorbents lies in the fact that the oxidized forms of mercury, such as HgCh, seem to be more easily controlled than Hg° via solid adsorption4-6. Past research and field experience have identified two different classes of sorbents to be effective in mercury removal: activated carbons and alkaline sorbents such as calcium-based (Ca-based) compounds. Activated carbons have been found to be efficient Hg° and HgCl2 sorbents4,7"11. Alkaline Ca-based sorbents (in addition to activated carbons) were effective HgCl2 sorbents5,12*14 in addition to effectively controlling emissions of acid gases, such as S02 and hydrogen chloride (IIC1). The dual role of Ca-based sorbents in controlling emissions of acid gases and IIgCl2 may be considered as a cost-effective, multi-pollutant control strategy. This strategy could be further enhanced by focusing on improving the adsorption of Hg° by Ca-based sorbents in order to increase total mercury capture from flue gas. Dry sorbents that can react with Hg° species shift them from the vapor phase to the particulate (sorbent) phase, thus facilitating removal. Two cr itical steps need to be accomplished for effective Hg° adsorption by injected dry sorbents The first step is film mass transfer by diffusion of Hg° species from the bulk flue gas to the vicinity of the particle surface within the available residence time (sorbent particles assumed to be adequately dispersed). The second step involves the reaction of Hg° molecules with sorbent active sites at the vicinity of the surface. A fundamental analysis15 concluded that only a veiy small surface area of active sites was theoretically required to trap Hg°. The implication was that the number of reactive surface sites is much more important than the amount of surface area. The importance of the reactive surface sites underscores the need for comprehensive research on characterizing the role of active sites and surface functional groups on the adsorption of different mercury species Previous investigation13 illustrated that calcium hydroxide [Ca(OH)2] and Ca(OH)2-fly ash sorbents had insignificant removal capacity for Hg° in the absence of S02 in the simulated flue gas. However, in the presence of S02, IIg° capture was enhanced for selected Ca(OII)2-fly ash sorbents. This enhancement was likely due to the effects of a combination of parameters such as generation of sulfur-activated sites, surface area, and internal pore structure. More studies are needed to conclusively determine the effects of these parameters. Another study14 indicated that Ca(OH)2 had no affinity for Hg° in nitrogen (N2) but captured HgCl2 effectively. Sorbent alkalinity was hypothesized to be the underlying reason for the capture 2 ------- of acidic HgCI2. The presence of acid gases (S02 and HC1) drastically decreased the uptake of HgCl2 by Ca(OII)2. It was hypothesized that the inhibition effect was due to competition between these acid gases and HgCh for the same available alkaline sites. Conversely, the same study14 showed that the presence of acid gases (especially HC1) had a positive effect on the capture of Hg° by a high-Ca lignite-coal-based activated carbon. This enhancement effect was then attributed to the high concentration of Ca (acid gas sorbent) in carbon. It was hypothesized that the attraction of HC1 to Ca/carbon sites in the activated carbon created active chlorine (CI) sites. These sites captured tig0 through formation of Cl-Hg bonds in the activated carbon. This hypothesis was further confirmed by another study16. In summary, our previous study14 was significant for two observations. These observations were: an abundance of HgCl2 sites in Ca-based sorbents and the enhancement effect of HC1 on Hg° uptake by the activated carbon. They suggest that generation of suitable CI sites in Ca-based sorbents may promote adsorption ofHg0 as an HgCl2-like molecule. Therefore, the focus of this study was to elucidate the following: • potential for facilitation of IIg° uptake (IIg° captured as HgCl2) by Ca-based sorbents following their exposure to a CI source (HC1), • comparison of rates and capacities of Hg° uptake by fresh and HCl-exposed Ca-based sorbents, • correlation between the amount of CI uptake and subsequent Hg° uptake by Ca-based sorbents, and • the nature of active sites in HCl-exposed, Ca-based sorbents that are instrumental in Hg° capture. Experimental Apparatus Detailed descriptions of the experimental setup, a fixed bed reactor system (Figure 1), have been presented previously13, l4. Briefly, pure Hg° liquid in a permeation tube was the source of Hg° vapor. HC1 [500 ppm in nitrogen (N2)] was delivered using certified gas cylinders. The concentration of Hg° in the gas stream was controlled by adjusting the temperature of the permeation tube's water bath. Hg° vapor generated was delivered to the fixed bed by an N2 stream at a constant total system flow rate of 300 cmVmin (at standard temperature and pressure, STP). The sorbent to be studied (0.5 g) was placed in the reactor between two quartz wool plugs and maintained at the desired bed temperature by a temperature controller. A furnace kept at 850° C was placed downstream of the reactor to convert any oxidized mercury (Hg44, as in HgCl2) to Hg°. The presence of the furnace enabled detection of oxidized, non-adsorbed Hg° as gas-phase Hg° by the on-line ultraviolet (UV) Hg° analyzer. Thus, continuous Hg° capture data by the fixed bed of sorbent could be acquired. 3 ------- Methodology In each test, the fixed bed was first exposed to 500 ppm HC1 in N2 for different exposure times. The bed was then subsequently exposed to 32 ppb Hg° in N2 for 60 minutes or until 100 percent breakthrough (saturation) was achieved (whichever came first). During the Hg° contact period, the exit concentration of mercury was continuously monitored. The instantaneous removal of Hg° at any time (Qt) was obtained as given by equation (1): Qt ~ [(Hg° in"*Hg°out)/ Hg° in] X 100 (1) Hg° uptake was determined by integrating and evaluating the area under the removal curve. Hg° uptake was defined as a cumulative removal of Hg° up to time t and was expressed as a weight ratio of mercury species to sorbent (|ig Hg°/g sorbent). Similarly, CI uptake was defined as a cumulative removal of CI up to time t and was expressed as a weight ratio of CI species to sorbent (ng Cl/g sorbent). Selected experiments conducted during this test program were run in duplicate and indicated a range of ±5 percent about the mean in the experimental results. Analytical Sorbent samples to be analyzed for CI uptake were removed from the fixed-bed reactor after the completion of the HC1 exposure period. Each sample was transferred to a beaker containing 20 ml of deionized water and a small amount of 2N sulfuric acid. The sample was stirred for 15 min and was then filtered into a 100-ml volumetric flask to remove the quartz wool residue. An aliquot was taken from the original 100-ml dilution for determination of CI by ion chromatography. Chlorination products of sorbent reaction with HC1 in the fixed-bed reactor were evaluated in a Perkin Elmer DSC7 differential scanning calorimeter (DSC). A detailed description of the DSC analysis can be found elsewhere17. Particle size distributions were determined in a Micromeritics Sedigraph using a gravity sedimentation method (Sedisperse A-l 1 dispersant). Porosity and surface area were measured in a Micromeritics ASAP 2600 using N2 adsorption and desorption with a Brunauer-Emmett-Teller (BET) method. The amount of crystalline water in certain sorbents was determined in a Perkin Elmer TGA7 thermogravinietric analyzer. A detailed description of TGA analysis is outlined elsewhere17. Results Hg° Uptake by HCI-Exposed Sorbents The effect ofHCl on the capture ofHg0 by Ca-bascd sorbents was studied by carrying out sequential exposures to HC1 and Hg°. An initial exposure to HCI (for 24 hours) was followed by an exposure to Hg° at a bed temperature of 60 °C. Initially, four Ca-based sorbents (all reagent grade, Fisher Scientific) were chosen for this study: Ca(OH)2, calcium oxide (CaO), calcium carbonate (CaCC^), and calcium sulfate dihydrate (CaS04'2H20, gypsum). The Hg° uptake (fig Hg°/g sorbent) by the HCl-exposed sorbents is shown in Figure 2. It should be noted that no 14g° 4 ------- removal was obtained with fresh (not exposed to HCl) sorbents (detection limit of about 0.05 pg Hg°/g sorbent). Table 1 summarizes IIg° uptakes by sorbents previously exposed to HCl. Table 1 Hg° Uptake by the HCl-exposed Ca-based Sorbents Sorbent Type CI Uptake, ng/g Ilg0 Uptake, jig/g Ca(OH)2 60000 0.108 CaO 16000 0.828±0.046 CaC03 2400 0.587±0.065 CaS04*2H20 3600^200 2.27±0.056 The CI uptake did not correlate with the Hg° uptake. The sorbent with the highest CI uptake, Ca(OH)2, exhibited the lowest Tig0 removal. Exposure to HCl clearly induced Hg° removal capabilities in certain sorbents. HCl-exposed CaS04*2H20 exhibited superior IIg° removal capability compared to other HCl-exposed sorbents. One obvious difference between CaS04-2H20 and other sorbents considered here is that the former contains crystalline water within its structure. Effect of Crystalline Water The important role of crystalline water in adsorption of HCl and Hg° on the surface of CaS04*2H20 was indicated by results presented above. This role was further confirmed by alternate tests using anhydrous CaS04 and calcium sulfate hemihydrate (CaS04 !/2H20, plaster) at identical solid/gas-phase contacting conditions. Results are shown in Figure 3. HCl-exposed CaS04 was not active in capturing Hg°, and HCl-exposed CaS04*1/2H20 was less active than CaS04*2H20. This figure shows the importance of crystalline water and its quantity on the adsorption of HCl and subsequent creation of a suitable active site for capturing Hg°. To further investigate the effect of crystalline water, two more crystalline water containing sorbents were tested in addition to CaS04*2H20. They were: sodium phosphate tribasic (Na3P04-12H20) and sodium borate (Na^CVlOH^O). Table 2 summarizes the CI uptake (in pg/g sorbent) versus Hg° uptake (in pg/g sorbent, after 40 minutes of exposure) together with the physical properties of these sorbents. The higher amount of crystalline water in Na2B407- 10II2O and Na3P04* 121120 led to higher uptakes of CI. No correlation existed between the IIg° and CI uptake and the physical properties of the sorbents such as surface area, pore volume, average pore diameter, and median particle size. 5 ------- f Table 2 CI and Hg° Uptake by Sorbcnts Containing Ciystalline Water Sorbcnt Type CaS04-2H20 Na2B4O7*10H2O Na3P0412H20 Total CI Fed, g/g 0.68 0.68 0.68 CI Uptake, ng/g 3600±200 24000 38000 H2O/CI, mole ratio 115 39 29 Hg° Uptake, jig/g 2.27±0.056 0.197 0.697 Cl/Hg°, mole ratio 8637 343147 153515 Surface Area, m2/g 6.6 26.3 2.8 Pore Volume, cm3/g 0.033 0.135 0.013 Average Pore Diameter, A 198 205 182 Median Particle Size, \im 29 47 >300 Saturation Capacity of HCI-exposed CaS04m2H20 A series of tests was performed to evaluate the saturation capacity (breakthrough points) of CI by CaS04*2H20. Different batches of CaS04-2H20 were exposed to 500 ppm HC1 in N2 for different exposure times, and resulting CI and subsequent IIg° uptakes were determined. The results are summarized in Table 3. Table 3 Effect of Exposure Time on CI and Hg° Uptake by CaS04-2II2Q Exposure Time, h 0 1 4 8 12 24 48 Total CI Fed, g/g 0 0.028 0.114 0.220 0.340 0.680 1.36 CI Uptake, |ig/g <80 1502 1640 2600 3200 3600 3600 H2O/CI, mole ratio >5160 274 252 159 129 115 115 Hg° Uptake, \ig/g <0.05 0.123 0.613 0.974 1.80 2.27 2.21 Cl/Hg°, mole ratio - 67742 12432 12333 8181 7286 7846 As can be seen in Table 3, the CI saturation capacity of CaS04*2H20 batches was obtained at 24 hour exposure time. Hg° uptake, interestingly, reached its saturation point at the same exposure time as IIC1. Total CI fed at the breakthrough point was 0.68 g Cl/g CaS04*2H20; whereas, the total CI uptake at this point was 0.0036 g Cl/g CaS04*2H20. This indicated that only a small fraction of the gas-phase HC1 that passed through the fixed bed was actually captured. Temperature Effect on the Adsorption ofHCI by CaS04'2H20 The effect of bed temperature on the adsorption of IIC1 by CaS04*2Il20 and subsequent capture of IIg° by the HCI-exposed CaS04*2H20 was studied next. In these studies, CaS04*2II20 was exposed to a high concentration HC1 in N2 stream (10000 ppm) for a short period of 1.2 hours to expedite the previous lengthy (24-hour) experiments. Consequently, these results may not be directly comparable to the previous ones outlined above since they were obtained at a different HC1 concentration. The capture of HC1 by CaS04*2H2Q was studied at three temperatures: 60, 6 ------- 100, and 120 °C. The Hg° capture by the HCl-exposed CaS04*2H20 was studied at 60 °C. In a separate study, CaS04*2H20 was exposed to HCI at 60 °C, and the resulting HCl-exposed sorbent was exposed to Hg° (32 ppb in N2) at three temperatures: 60, 100, and 120 °C. Table 4 summarizes the effect of temperature on CI and Hg° uptake by HCI-exposed CaSO^IkO. Increasing the bed temperature from 60 to 100 °C during the HCI exposure period increased the CI uptake and subsequently resulted in a higher Hg° uptake. Table 4 Effect of Bed Temperature on CI and Hg° Uptake by CaS04*2H:0 IICI /Ilg0 Exposure Temperatures, °C/ °C Hg° Uptake, ng/g CI Uptake, ng/g 60/60 0.992±0.123 1380±300 100/60 1.934 4200 120/60 1.736 2640 60/100 .0-1 1380±300 60/120 0.1 1380i 300 Discussion Mechanism of Hg° Uptake by HCI-exposed CaS04-2H20 Two observations were made following experiments investigating the uptake of Hg° by HCl- exposed sorbents as previously summarized in Table 1: ~ CI uptake by Ca-based sorbents promoted their capability to remove Kg0, and • there was no correlation between CI and Hg° uptake The second observation may indicate that the reactivity (and not the quantity) of captured CI is the determining factor in the subsequent adsorption of Hg°. To investigate this hypothesis, DSC analysis of the HCl-exposed sorbents was performed to determine the nature of captured CI by these sorbents. When scanned by DSC, only Ca(OH)2 (and to a very limited extent CaO) revealed an endotherm. An endotherm indicates occurrence of a chemical reaction (as opposed to simple adsorption). An earlier study14 identified this endotherm as a solid solution of CaCl2 and CaClOH. This high uptake of CI by Ca(OII)2 (DSC endotherm) can be, therefore, explained by the formation of this solid solution through a chemical reaction between gas-phase HCI and Ca(OH)2. This chlorinated compound was subsequently shown to be inactive in capturing IIg°; thus, no correlation between the CI and Hg° uptake could be demonstrated. A similar DSC analysis of HCl-exposed CaS04-2H20 and CaC03 indicated the absence of this endotherm, suggesting that the uptake of CI by these two sorbents may have occurred through a physical adsorption mechanism. However, IlCl-exposed CaS04*2H20 subsequently revealed hiah uptake Ofllg0. The high uptake of Hg° by the HCl-exposed CaS04*2H20 may be explained as follows. Crystalline water on the surface of the CaS04*2II20 lattice may be regarded as an attractive site 7 ------- for the uptake of CI. Here, the nature of surface crystalline water (as differentiated from surface moisture) needs to be discussed. Crystalline water was in solid form and was attached to the crystalline structure of the bulk of CaS04-2H20. The oxygen atom (O) of a solid crystalline water molecule located at the surface, as depicted in Figure 4, may have attracted an IIC1 molecule through formation of hydrogen bonds between 0 and the hydrogen atom (H) of HC1. This bond is a subset of bonds formed by an electron donor-acceptor (EDA) interaction, also known as Lewis base-acid interactions. Lewis base-acid interactions have been studied extensively. For example, the case of adsorption of HC1 in the liquid phase by activated carbon was studied18. The study explored adsorption of HC1 by basic sites on activated carbon. One of the hypotheses presented18, related this adsorption to surface functional groups containing O and an EDA interaction. The assertion was made that the H atoms of HC1 molecules could interact with electron pair donors, O atoms, on the surface of activated carbon (a Lewis acid-base interaction). This electrostatic interaction would lead to formation of hydrogen bonds. Similarly, in the case considered here, the O atoms of crystalline water molecules on the surface of CaS04*2H20 interacted with HCI molecules, forming hydrogen bonds. Hydrogen bonds, formed by attraction of H and O atoms, are only one-tenth to one- thirtieth as strong as covalcnt chemical bonds (3 to 250 kcal/mole). Therefore, they could be classified, in this case, as physical rather than chemical adsorption mechanisms. As shown in Figure 4, two neighboring crystalline water molecules can attract and adsorb two IIC1 molecules and, if these two crystalline water molecules (or the adsorbed IIC1 molecules) are located at an optimum distance, they subsequently could attract and adsorb an Hg° molecule possibly through attractive van der Waals forces to form an HgClj-like molecule. This was in agreement with our previous investigation14 and results by others16. These results showed that HgCI2 was easier to capture than Hg° and that mercury species captured in the presence of IIC1 were most probably in the form of an HgCl2-type molecule. CI and IIg° uptake analysis of the exposed CaS04*2H20 showed that for every 10,000 CI atoms one Iig° molecule was captured (refer to Table 1). This illustrated the fact that an optimum spacing between crystalline water molecules or CI atoms on the surface of CaS04*2H20 was required to trap and capture molecules of Hg(l as HgCl2. Apparently, only one out of 5000 such sites (two CI atoms constitute one site) had this optimum spacing. Optimum distance may be the length of an IlgCh molecule from one chlorine atom to the other. Ilg-Cl bond distance in IlgCh was reported16 to be approximately 2.25 A. Bond lengths and angles of selected chemical compounds are also listed in the CRC handbook19. Such data are available for sulfur dichloride (SCI2). S-Cl bond length is I.99±0.03 A and Cl-S-Cl angle is 101±4 . Assuming that SCI2 and HgCl2 have similar bond angles, the distance between the two chlorine atoms can be calculated as 3.47 A. This distance (3.47 A) may be the optimum distance between two O atoms that, through formation of hydrogen bonds with two H atoms of two HCI molecules, can attract and trap an IIg° molecule. It was further hypothesized that, through their O atoms, surface crystalline water molecules were the active sites for the capture of HCI and subsequent physical adsorption of Hg°. It was of interest to evaluate the surface coverage of the crystalline water (its projected area) and compare that to the total surface area of CaS0.i*2H20. Assuming that crystalline water at the surface 8 ------- » resembled an SCI2 molecule, the distance between the two H atoms (see Figure 4) could be calculated as 3.07 A. The projected area of one crystalline water molecule would thus be 7.4* 10" 20 m2/molecule. As indicated in Tables 1 and 3, maximum (saturation) uptake of CI by CaS04-2H20 (at 60 °C) was achieved after 24 hours of exposure (3600 jig Cl/g CaS04-2H20). This corresponded to l.OlxlO-4 mole Cl/g CaSC>4*2H20. If, through hydrogen bond formation, each surface crystalline water adsorbed one molecule of HC1 at the saturation point, then the number of surface crystalline water molecules would be: (1.01xl0"4)x(6.02xl02 ) = 6.1*1019 molecules H20(s)/g. The total projected area occupied by these molecules would thus be (6.1xl019)x(7.4*10"2°) = 4.5 m2/g. The total surface area of CaS04*2H20 was measured to be 6.6 m2/g (see Table 3). The close agreement between these two numbers was another indication that a bond (most probably a hydrogen bond) between gas-phase HC1 and surface crystalline water was formed. As indicated above, the amount and spacing of crystalline water molecules on the surface of a sorbent may have controlled the amount of HC1 uptake and subsequent capture of Hg°. Of the two tested compounds with crystalline water content higher than CaS(V2H20: \Ta3P04-12H20 demonstrated greater IIC1 uptake than Na2B407* 10II2O. This observation may confirm the "hydrogen bonding" theory outlined above. However, the IIg° uptake results (see Table 2) indicate that not all of the captured CI was instrumental in capturing Hg°; CaS04*2H20 with the lowest CI uptake exhibited the highest Hg° uptake. This observation may support the theory that an optimum spacing between CI atoms is required to trap Hg° as HgCl2 on the surface; not all captured CI can provide this optimum spacing. Hydrogen Bond Formation Since the fixed-bed sorption process was not an external mass transfer controlled process14, the CI capture must have been controlled by an adsorption kinetic mechanism. This mechanism, specifically, involved the formation of hydrogen bonds between the H atom of an HC1 molecule and the O atom of a hydrated water molecule as depicted by equation (2): HCl(g) + II20(s) - C1II OII2(s) (2) The rate of this hydrogen bond formation can be expressed as the difference between the adsorption (formation) rate and the desorption (dissociation) rate. The adsorption rate should be proportional to the HCI concentration in the gas phase and to the difference between the number of sites (crystalline water molecules) available for adsorption and the number of occupied sites. The desorption rate should simply be proportional to the number of occupied active sites; hence the rate of the overall process is expressed by equation (3): r - k)(qmax-q)c - k?q (3) Where r is the overall rate of sorption, ki and k2 are rate coefficients of the adsorption and desorption reactions (respectively), qmax is the asymptotic adsorbate concentration (fully occupied 9 ------- surface), q is CI concentration in the solid phase (adsorbate), and c is HC1 concentration in the gas phase. Both rate coefficients are Arrhenius functions of the bed temperature. It may be hypothesized that in the 60-100 °C range (data in Table 4), ki in equation (2) is greater than k2. Thus, increasing the temperature increased the adsorption rate of HC1 through a higher rate of hydrogen bond formation. Further increase in temperature (from 100 to 120 °C) caused a decrease in CI and subsequent Hg° uptake. This behavior may be interpreted two different ways. If it is assumed that in the 100-120 °C range desorption rate (k2) was larger than the adsorption rate (ki), then an increase in temperature resulted in higher desorption of CI and subsequent lower capture of Hg°. More likely, however, the loss of HC1 uptake capability by CaS(V2H20 was the result of the compound's losing its crystalline water. According to literature, CaS04*2H20 loses its crystalline water at 140 °C. However, TGA analysis performed on CaS04-2H20, used in this study, indicated that the loss of crystalline water occurred as early as 80 °C and approached completion at about 140 °C. The partial loss of crystalline water in the 100-120 °C range may be the underlying reason for the loss of the HC1 uptake capability of CaSC>4*2H20 at these elevated temperatures. Increasing the temperature to 100 or 120° C during the Hg° exposure period completely suppressed the Hg° capture capability of the HCl-exposed CaSCV2H20 As illustrated in Figure 4, the Hg° attraction to the CI sites may be of a weak physical adsorption nature (van der Waals bonds). Increasing system temperature significantly decreased the stability of these bonds. Therefore, no He0 capture could be observed at higher temperatures. This indicates that Hg° was not strongly attached to the active sites and could be easily desorbed through thermal treatment. Conclusions A series of experiments was performed to simulate adsorption of IIg° by Ca-based sorbents in the presence of HQ. The following conclusions are based on the results of those experiments: 1) For the family of four Ca-based sorbents tested here, the amount of Hg° uptake did not correlate with the amount of the preceding HC1 capture because of the different nature of HC1 interactions with those Ca-based sorbents. Chlorinated compounds, produced as a result of chemical reaction between IIC1 and CaO or Ca(OII)2, did not promote subsequent IIg° capture. 2) HQ physical adsorption, through hydrogen bond formation, in Ca-based compounds with crystalline water (CaS04*2H20 and CaSCV'/^O) was shown to be responsible for promoting Hg° uptake. It was hypothesized that a pair of physically adsorbed HQ molecules can attract and adsorb an Hg° molecule possibly through attractive van der Waals forces to form an HgCl2-like molecule. The amount of HgC) uptake was correlated to the amount of crystalline water for Ca- based sorbents. 3) Optimum spacing between crystalline water molecules, or between adsorbed CI atoms, was required to trap Hg° as an HgCl2-typc molecule. 10 ------- 4) As indicated by the effect of temperature studies on the uptake of HC1 by CaSO^P^O, and depending on the range of temperatures, the rate of hydrogen bond formation and/or dissociation was the rate limiting step for CI and subsequent Hg° uptake by CaS04*2H20. 5) Increasing the temperature during the Hg° exposure period completely suppressed the Hg° capture capability of the HCl-exposed CaS04*2II20. As mentioned, the Hg° attraction to the CI sites may have been of a weak physical adsorption nature (van der Waals bonds). Increasing the system temperature significantly decreased the stability of these bonds. This indicated that Hg° was not strongly attached to the active sites and could be easily desorbed through thermal treatment. 11 ------- References 1. D. H. Klein, A. W. Andren, J. A. Carter, J. F. Emery, C. Feldman, W. Fulkerson, W. S. Lyon, J. C. Ogle, Y. Talmi, R. I. Van Hook, and N. Bolton. "Pathways of 37 Trace Elements Through Coal-Fired Power Plant." Environmental Science & Technology. Vol. 9, p. 973 (1975). 2. C.B. Sedman. "Controlling Emissions from Fuel and Waste Combustion." Chemical Engineering. Vol. 106, No.l, p. 82 (1999). 3. K.E. Redinger, A.P. Evans, R.T. Bailey, and P.S. Nolan. "Mercury Emissions Control in FGD Systems" Presented at the First EPRT-DOE-EPA Combined Utility Air Pollutant Control Symposium (The MEGA Symposium), Washington, DC (August 1997). 4. S. V. Krishnan, B. K. Gullett, and W. Jozewicz. "Sorption of Elemental Mercury by Activated Carbons" Environmental Science & Technology. Vol. 28, No. 8, p. 1506 (1994). 5. S. V. Krishnan, H Bakhteyar, and C. B. Sedman. "Mercury Sorption Mechanisms and Control by Calcium-Based Sorbents." In the Proceedings of the 89th Air & Waste Management Association Annual Meeting, Nashville, TN (1996). 6. A. Lancia, D. Musmarra, F. Pepe, and G. Volpicelli. "Control of Emissions of Metallic Mercury from MSW Incineration by Adsorption on Sorbalit™." In the Proceedings of the 1997 International Conference on Incineration and Thermal Treatment Technologies, Oakland, CA(May 1997). 7. R. D. Vidic and J. B. McLaughlin. "Uptake of Elemental Mercury Vapors by Activated Carbons." Journal of Air & Waste Management Association. Vol 46, p. 241 (1996). 8. D. Karatza, A. Lancia, D. Musmara, F. Pepe, and G. Volpicelli. "Removal of Mercuric Chloride from Flue Gas by Sulfur Impregnated Activated Carbon." Hazardous Waste & Hazardous Material. Vol. 13, p. 95 (1996). 9. G. E. Dunham, S. J. Miller, R Chang, and P. Bergman "Mercury Capture by an Activated Carbon in a Fixed-Bed Bench-Scale System." Environmental Progress. Vol. 17, No. 3, p. 203 (1998). 10. T. R. Carey, O. W. Hargrove, C. F. Richardson, R. Chang, and F. B. Meserole. "Performance of Activated Carbon for Mercury Control in Utility Flue Gas Using Sorbent Injection." In the Proceedings of the First EPRI-DOE-EPA Combined Utility Air Pollutant Control Symposium (The MEGA Symposium), Washington, DC (August 1997). 12 ------- 11. T. R Carey, O. W. Hargrove, C. F. Richardson, R. Chang, and F. B. Meserole. "Factors Affecting Mercury Control in Utility Flue Gas Using Activated Carbon." Journal of Air & Waste Management Association. Vol. 48, p. 1166 (1998). 12. A. Lancia, D. Musmarra, F. Pepe, and G. Volpicelli. "Adsorption of Mercuric Chloride Vapors from Incinerator Flue Gases on Calcium Hydroxide Particles." Combustion Science and Technology. Vol. 93, p. 277 (1993). 13. S.B. Ghorishi and C.B. Sedman. "Low Concentration Mercury Sorption Mechanisms and Control by Calcium-Based Sorbents: Application in Coal-Fired Processes." Journal of Air & Waste Management Association. Vol. 48, p. 1191 (1998). 14. S.B. Ghorishi and B.K. Gullett. "Sorption of Mercury Species by Activated Carbons and Calcium-Based Sorbents: Effect of Temperature, Mercury Concentration and Acid Gases." Waste Management & Research. Vol.16, No. 6, p. 582 (1998). 15. M Rostam-Abadi, S. G. Chen, II-C. Ilsi, M. Rood, R. Chang, T. Carey, B. Hargrove, C. Richardson, W. Rosenhoover, and F. Meserole. "Novel Vapor Phase Mercury Sorbents." In the Proceedings of the First EPR1-DOE-EPA Combined Utility Air Pollutant Control Symposium (The MEGA Symposium), Washington, DC (August 1997). 16 F.E. Iluggins, G.P. Huffman, G.E. Dunham, and C.L. Senior. "XAFS Examination of Mercury Sorption on Three Activated Carbons." Energy & Fuels. Vol 13, No. 1. p 114 (1999). 17. W. Jozewicz and B.K. Gullett. "Reaction Mechanisms of Dry Ca-Based Sorbents with Gaseous HC1." Industrial and Engineering Chemistry Research. Vol. 34, p. 607 (1995). 18. C.A. Leon Y Leon, J.M. Solar, V. Calemma, and L.R Radovic. "Evidence for the Protonation of Basal Plane Sites on Carbon." Carbon. Vol. 30, "No. 5, p 797 (1992) 19. CRC Handbook of Chemistry and Physics, 62nd edition, CRC Press, Boca Raton; Florida; Editors: R.C. Weast and M.J. Aslle, (1981-1982) 13 ------- M ercury Source Carbon Trap 1 Manifold On-Off Valves I Wate r Batht 3-Way Valve Reactor By-Pass e—- o 3-Way Valves HCI/N Carbon Trap Data Acquisition Mercury ^ Analyzer | Rotameter Figure 1 Schematic of the Fixed-bed Sorption Reactor System 14 ------- Hg° uptake (jig/g sorbent) Figure 2 IIg° Uptake by ITCl-exposed, Ca-based Sorbents CaO Ca(OH)2 CaS04.2H20 * CaC03 ~ 15 ------- Hg° uptake (^g/g sorbent) 2.5 0.5 0 10 20 30 40 50 60 70 time (min) Figure 3 Effect of Crystalline Water (0, V2, and 2) on Hg° Uptake by the HCl-exposed Forms of CaS04 16 ------- H3° Hg? van der Waals bond • ^ \ 2 25^ cjji -347A - (j:i H H i | I Hydrogen bond A H H CaSCX Figure 4 Mechanism of IIC1 and IIg° Capture by CaS04'2H20 17 ------- VCMDT n j -i o TPrMWirAI OPPOOTHATA NRMRJy RT • 418 (Plct fore completing) 1. REPORT NO. 2. 600/A-99/060 3. RECIPIENT'S ACCESSION NO. 4. TITLE AND SUBTITLE Role of HC1 in Adsorption of Elemental Mercury Vapor by Calcium-based Sorbents 5. REPORT OATE 6. PERFORMING ORGANIZATION CODE 7.author(s) a Ghorishi (ARCADIS), B. K. Gullett (EPA), and W. Josewicz and W.Kozlowski (ARCADIS) 8. PERFORMING ORGANIZATION REPORT NO. 9. PERFORMING OROANIZATION NAME ANO ADDRESS ARCADIS Geraghty and Miller, Inc. P. C. Box 13ICS Research Triangle Park, North Carolina 2770S 10. PROGRAM ELEMENT NO. 11. CONTRACT/GRANT NO. 68-C-99-201, WAs 5-026 and 0-014 12. SPONSORING AGENCY NAME ANO ADDRESS EPA, Office of Research and Development Air Pollution Prevention and Control Division Research Triangle Park, NC 27711 13. TYPE OF REPORT AND PERIOD COVERED Published paper; 9/98-5/99 14. SPONSORING AGENCY CODE EPA/600/13 15.supplementarynotesappcd project officer is Brian K. Gullett, Mail Drop 65, 919/541- 1534. For presentation at EPRI/DOE/EPA Combined Utility Air Pollutant Control Symposium, 8/16-20/99, Atlanta, GA. i6. abstract 'j-he paper gives results of a study to identify active sites and surface func- tional groups that may contribute to the adsorption of elemental mercury (Hg°) by relatively inexpensive calcium (Ca)-based sorbents. (NCTE: Hg° capture has been mostly investigated using high-surface-area, expensive activated carbons. Previous investigations showed that hydrogen chloride (HC1) in the flue gas had a pronounced positive effect on the adsorption of lJg° by a lignite-coal-based activated carbon with high Ca con;ent.) The study investigated the formation of chlorine (Cl) sites in Ca- based sorbents as well as their role and reactivity in the adsorption of Hg°. HC1- exposed calcium sulfate dihydrate (gypsum) exhibited a superior IIg° sorption capa- bility. Crystalline water molecules on the surface of the gypsum were confirmed to contribute indirectly to Hg° uptake. These surface molecules may have adsorbed 1JC1 through hydrogen bond formation between an oxygen atom of a crystalline water molecule and a hydrogen atom of an HC1 molecule. Two adjacent, physically adsor- bed HC1 molecules could then trap an IIg° molecule through formation of a mercuric- chloride-like molecule. Significant correlations between Cl and Hg° uptake by gyp- sum were noted. 17. KEY WORDS AND OOCUMENT ANALYSIS a. DESCRIPTORS b.IDENTIFIERS/OPEN ENDED TERMS c. COSATI Field/Group Air Pollution Gypsum Mercury (Metal) V apors Sorption Calcium Hydrogen Chloride Air Pollution Control Stationary Sources Elemental Mercury 13 B 08G 07B 07D 18. DISTRIBUTION STATEMENT Release to Public 19. SECURITY CLASS (This Report) Unclassified 21. NO. OF PAGES 20. SECURITY CLASS (This pexe) Unclassified 22. PRICE EPA Form 2220-1 (9-73) ------- |