Catalytic Oxidation and Stabilized Adsorption of Elemental Mercury

Institute of Coal Chemistry, Chinese Academy of Sciences, Taiyuan 030001, People's Republic of China. ‡ Graduate School of Chinese Academy of Scienc...
1 downloads 0 Views 961KB Size
Article pubs.acs.org/EF

Catalytic Oxidation and Stabilized Adsorption of Elemental Mercury from Coal-Derived Fuel Gas He Zhang,†,‡ Jiantao Zhao,*,† Yitian Fang,† Jiejie Huang,† and Yang Wang† †

Institute of Coal Chemistry, Chinese Academy of Sciences, Taiyuan 030001, People’s Republic of China Graduate School of Chinese Academy of Sciences, Beijing 100039, People’s Republic of China



ABSTRACT: The aim of this study is to develop an efficient process for the stabilized removal of elemental mercury (Hg0) from coal-derived fuel gas by the aid of the catalytic oxidation of Hg0 over activated carbon (AC). The experiments were performed in the temperature range of 120−240 °C with simulated coal-derived gases containing varying concentrations of CO, H2, CO2, H2O, and H2S. O2 was introduced to promote the oxidation reaction between H2S and Hg0 over the AC sorbent to form mercury sulfide (HgS). The results indicated that the Hg0 removal capacity of AC could be dramatically improved in the presence of both H2S and O2. The presence of O2 was indispensable for the efficient and stable removal of Hg0 from H2S-containing fuel gas. The high temperature and high content of reducing gases, such as CO and H2, may inhibit the oxidation reaction and decrease the Hg0 removal efficiency, whereas the high content of H2O can promote Hg0 removal. On the basis of thermodynamic analysis, as well as the temperature-programmed decomposition (TPD) and X-ray photoelectron spectroscopy (XPS) characterization of the sorbents, it is suggested that the partial oxidation of H2S with O2 to active sulfur may contribute to the stabilized removal of Hg0 by the reaction of active sulfur with Hg0 to form HgS.

1. INTRODUCTION Gasification is an important strategy for the effective use of finite coal reserves.1 However, various mercury compounds are generated by the production of coal-derived fuel gas. The direct emission of mercury compounds would cause serious environmental problems because these compounds are powerful neurotoxins and can bioaccumulate in the food chain. Therefore, efficient gas-cleaning systems must be developed to remove the mercury generated from coal gasification.2,3 Because the content of mercury compounds in coal-derived fuel gas is higher than that in coal combustion flue gas,4,5 it is more practical and challenging to remove mercury from coal-derived fuel gas. Among the mercury compounds produced from coal gasification, elemental mercury (Hg0) is the predominant species formed in such a reducing environment.6 Additionally, Hg0 is difficult to capture because of its low melting point, high equilibrium vapor pressure, and low solubility in water.4−6 Hence, the removal of Hg0 is quite demanding and has received significant attention. Until now, very little attention has been paid to the removal of Hg0 from coal-derived fuel gas. Portzer et al.7 have demonstrated that chemically reactive solid sorbents remove Hg0 from coal-derived synthesis gas at elevated temperatures (204−315 °C), but the specific name and composition of the sorbents were not mentioned in the paper. Granite et al.1 found that sorbents supported with noble metals, such as iridium, platinum, palladium, and ruthenium effectively remove Hg0 from simulated fuel gas at temperatures of 204−371 °C. Palladium sorbents appear to be the most promising for the high-temperature removal of mercury from fuel gases. However, the high production cost of such sorbents may seriously limit their industrial utility. To solve this problem, a more economical and efficient sorbent is necessary. As a relatively cheap sorbent, activated carbons (ACs) are widely © 2012 American Chemical Society

used in the removal of mercury from coal combustion flue gas.8−10 However, virgin ACs show poor removal of Hg0. It has been reported that AC impregnated with sulfur can remove Hg0 very effectively.11−13 Unfortunately, impregnated ACs are more expensive and less stable, and their regeneration is slower than that of virgin ACs. These drawbacks may inhibit the large-scale use of impregnated ACs. Thus, if AC is to be used to remove Hg0 from coal-derived fuel gas, it is of vital importance to develop an efficient method that can overcome the abovementioned shortcomings. The high adsorption ability of sulfurimpregnated AC results from the active sulfur species on the surface of the AC.13−15 Note that H2S can be oxidized by oxygen radicals to elemental sulfur on the virgin carbon surface.15 These observations suggest that Hg0 in coal-derived fuel gas may be removed efficiently by AC in the presence of H2S because H2S is usually present in the unrefined gas derived from coal gasification. If the coexisting pollutants can react to form HgS, the Hg0 removal efficiency would be greatly improved. Furthermore, the used sorbent waste needs no further treatment because HgS is stable and nontoxic to the environment. These features would reduce the cost, improve the adsorption efficiency of AC and, importantly, consume a part of the H2S contaminant load present in the fuel gas. To verify the above assumptions, a thermodynamic analysis of the reaction between H2S and Hg0 was performed. For integrated gasification combined cycles (IGCC), the thermodynamic efficiency of the cycle can be improved if the thermal energy of the syngas can be preserved at a higher cleaning temperature. It has been reported that sorbent-based desulfurization can be accomplished reliably at 260−315 °C.7 Received: September 25, 2011 Revised: February 7, 2012 Published: February 7, 2012 1629

dx.doi.org/10.1021/ef201453d | Energy Fuels 2012, 26, 1629−1637

Energy & Fuels

Article

Table 1. Thermodynamic Parameters of the Reaction among Mercury, Hydrogen Sulfide, Sulfur, and Oxygen H2S(g) + Hg(g) → HgS(s) + H2(g) temperature (°C) 120 160 200 240

ΔG (kJ/mol) 59.10 68.44 77.78 87.12

ΔG (kJ/mol)

K 1.410 5.595 2.596 1.357

× × × ×

H2S(g) + 1/2O2(g) + Hg(g) → HgS(s) + H2O(g)

S(s) + Hg(g) → HgS(s)

−8

10 10−9 10−9 10−9

−71.48 −66.84 −62.31 −57.75

ΔG (kJ/mol)

K 3.130 1.146 7.549 7.544

× × × ×

9

10 108 106 105

−258.2 −250.0 −241.9 −233.8

K 1.995 1.380 5.012 6.310

× × × ×

1034 1030 1026 1023

Figure 1. Schematic diagram of the experimental setup.

that the conversion of Hg0 to HgS could be effective and nearly complete in the presence of H2S and O2. In this study, the performance of AC for Hg0 removal from simulated coal-derived fuel gas containing H2S was evaluated in a fixed-bed reactor. To trigger the reaction, a small amount of O2 that could not react with H2 or CO at relatively low temperatures was introduced into the reducing atmosphere. Specifically, the impact of the reaction temperature and coal gas composition (CO, H2, H2O, and CO2) was systemically investigated. The temperature-programmed decomposition (TPD) technique was applied to understand the contribution of surface oxygen groups on AC in mercury removal. The exhausted AC was characterized using X-ray photoelectron spectroscopy (XPS) to gain information regarding the speciation and binding of mercury and sulfur to AC.

However, the temperature in most industrial gasification processes is approximately 200 °C after the water quench cleaning, such as in the Texaco and Shell gasification processes. Therefore, in the present work, we are targeting this temperature range (∼200 °C), which is close to the industrial condition, for mercury removal. As shown in Table 1,16,17 under standard conditions (constant pressure and temperature), the change in the Gibbs free energy (G) of the reaction between H2S and Hg0 is positive. Thus, the H2S−Hg system is a thermodynamically stable system at temperatures of 120−240 °C, and H2S cannot react spontaneously with Hg0 to form HgS. In contrast, the change in G of the reaction between S and Hg0 is negative, and in the temperature range of 120−240 °C, the S−Hg system is thermodynamically unstable. This implies that the reaction between S and Hg0 is spontaneous. Furthermore, the values of the equilibrium constant, K, are much greater than 1, indicating that, at equilibrium, the reaction would proceed mostly to the right. As proposed by Hedding and Rao,18 H2S can be oxidized by oxygen radicals to elemental sulfur on virgin carbon surfaces. Inspired by these findings, we hypothesized that the efficient conversion of H2S to elemental sulfur and its further reaction with Hg0 to form HgS may be achieved in the presence of AC and O2. To test this assumption, a thermodynamic analysis of the reactions among H2S, O2, and Hg0 in the 120−240 °C temperature range was performed. As listed in Table 1, the H2S−O2−Hg system is a thermodynamically unstable system, and the ΔG of the H2S−O2−Hg system is much more negative than that of the S−Hg system. Moreover, the values of K are very large. This result indicates that the reaction among H2S, O2, and Hg0 is spontaneous and

2. EXPERIMENTAL SECTION 2.1. AC Sample. The sample was a coal-based AC prepared from Jiexiu (Shanxi province, China) coal. The carbonization was conducted at 850 °C for 6 h under a 10.5 L/min N2 flow. The activation was performed at 950 °C with a CO2 flow rate of 360 cm3/min. During heating to the activation temperature and cooling to room temperature, an argon atmosphere was used. The Brunauer− Emmett−Teller (BET) surface area and total pore volume of the AC sample, as measured by nitrogen adsorption using a Micromeritics (Gemini) BET analyzer (SpectRx, Inc., Norcross, GA), were approximately (ca.) 800 m2/g and 0.312 cm3/g, respectively. 2.2. Apparatus and Procedure. Hg0 removal by AC was conducted in the laboratory-scale fixed-bed reactor system shown in Figure 1. The reactor was a quartz glass reactor with a length of 65 cm and a diameter of 2.2 cm, which was positioned vertically in a clam shell furnace. The furnace temperature was controlled by a K-type thermocouple within the furnace, whereas the reaction temperature 1630

dx.doi.org/10.1021/ef201453d | Energy Fuels 2012, 26, 1629−1637

Energy & Fuels

Article

was monitored by another K-type thermocouple axially centered in the reactor tube. The Hg0 removal experiments were conducted at atmospheric pressure with a gas hourly space velocity (GHSV) of 3000 h−1. A total of 2 g of a 30−60-mesh AC sample was loaded into the isothermal section of the reactor. The system was gradually heated under N2 gas to the desired temperature (120−240 °C). The simulated fuel gas feed used was composed of 57.9 μg/m3 (6.47 ppb) Hg0, 0.3% H2S (3000 ppm), 25% H2, 5% CO2, 4% H2O, and 60% CO, and the balance was N2 (v/v) (unless otherwise noted). The flow rates of the feed gases were controlled by mass flowmeters. A stable feed of Hg0 gas was realized using a mercury permeation device (VICI Metronics, Inc., Santa Clara, CA). The permeation device was sealed in a glass permeation tube holder with inlet and outlet ports. The permeation tube, which was certified by the manufacturer to release 80 ng of mercury/min at 50 °C, was located at the bottom of a glass U-tube that was immersed in a temperature-controlled water bath. A flow of ultrahigh-purity nitrogen carrier gas passed over the permeation tube and was maintained at all times by a mass-flow controller. The stream, laden with Hg0, was transferred throughout the system using Teflon tubing, which was heated to prevent condensation of Hg0. The water vapor in the simulated fuel gas was produced by heating the water, which was injected into the stream by a syringe pump through the inlet tubing (stainless steel) at the desired flow rate. O2 was introduced into the simulated fuel gas stream at the rate necessary to achieve an O2/ H2S molar ratio of 0.5. The Hg0 concentrations at the inlet and outlet of the reactor were analyzed using a QM201G portable mercury analyzer (Suzhou Greencalm, Inc., China) with a detection limit of 0.001 μg/m3. The QM201G portable mercury analyzer consists of the Hg analyzer and a standard mercury gas box. The sample gas was first dehumidified by silica gel and then injected into the mercury collector, in which mercury was collected by a gold membrane. After collecting for 3 min, according to the instructions of the manufacturer, the gold membrane was heated to the desired high temperature to liberate mercury. The mercury vapor was carried into the absorption cell by a flow of ultrahigh-purity nitrogen and was then measured using a cold-vapor atomic fluorescence spectroscopy (CVAFS) analyzer. Because acidic gases may erode the gold film surface and reduce the efficiency of amalgamation, after every 10 experiments, the standard controlled amount of element mercury sample gas was analyzed. When the error was greater than 3%, the gold membrane was washed using a 10% (v/ v) HNO3 solution for 10 min. The Hg mass balance closure was checked for each run. The amount of Hg adsorbed by the AC at the end of each run was analyzed in accordance with United States Environmental Protection Agency (U.S. EPA) Method 7473 using a direct mercury analyzer (Milestone DMA-80, Italy). The mass balance of Hg for each run was within the range of 100 ± 10%, which is acceptable.19 The outlet mercury speciation was also measured using the Ontario hydro method. The results showed that, after adsorption, the mercury remaining in the simulated gas was dominantly in elemental form. The Hg0 removal efficiency (η) was defined as shown below η=

C in − Cout × 100% C in

The XPS analysis of used AC samples was performed in an ESCALAB250 spectrometer (Thermo Electron Co., East Grinsted, U.K.) with an Al Kα radiation source. All of the binding energies were referenced to C 1s at 284.6 eV. Both survey scan spectra and highresolution scan spectra of Hg 4f and S 2p were collected from used AC samples. The XPS peak 4.1 software package was used to fit the XPS spectra peaks of Hg 4f and S 2p. Specifically, the spectra of Hg 4f and S 2p were fitted using a nonlinear least-squares method with a Lorentzian−Guassian function.

3. RESULTS AND DISCUSSION 3.1. Mercury Removal Efficiency in the Presence of H2S and O2. The effect of H2S and H2S−O2 on the Hg0 removal from simulated coal-derived gas by AC was examined at 160 °C. As shown in Figure 2, the initial Hg0 removal

Figure 2. Hg0 removal efficiency under different atmospheres at 160 °C. Feed gas: Hg0 (6.47 ppb), H2S (0 or 3000 ppm), O2 (0 or 1500 ppm), H2 (25%), CO2 (5%), H2O (4%), CO (60%), and N2 (balance gas) at 300 cm3/min.

efficiency by AC alone was 68%, which sharply decreased to 28% after 180 min of adsorption. In contrast, the initial Hg0 removal efficiencies increased to 78 and 89% when H2S and H2S−O2 were added and remained at ca. 68 and 78%, respectively, throughout the adsorption period. This finding indicated that the Hg0 adsorption ability of AC was significantly enhanced in the presence of H2S, particularly in the presence of both H2S and O2. Lopez-Antón et al.20 studied the mercury adsorption efficiency of a virgin AC and the same AC impregnated with sulfur. The removal efficiency of mercury by the former was approximately 18−30%, and the mercury removal efficiency by the latter reached 70−92%. The performance of AC under the atmosphere containing H2S and O2 in this study is close to that of sulfur-impregnated AC. In the H2S-containing fuel gas system, there was a consistent dip in the data at approximately 10 min, after which the removal efficiency increased. This might be due to the development of some new active sites for mercury adsorption, which are created during the adsorption process. To further identify the effect of the atmosphere on the Hg0 adsorption capacity, the adsorption time was prolonged to 100 h and the saturated amount of removed Hg0 was calculated. As shown in Figure 3, 1 g of virgin AC can only remove 50.4 μg of Hg in the absence of H2S and O2. However, the amount of captured Hg0 drastically increased to 210.3 μg/g in the presence of H2S, and this amount further increased to 390.8

(1)

where Cin and Cout represent the Hg0 concentrations at the inlet and outlet of the reactor, respectively. 2.3. Sorbent Characterization. The TPD of AC before and after mercury adsorption under different atmospheres (AC, AC−Hg, AC− H2S−Hg, and AC−H2S−O2−Hg) was studied with a chemisorption analyzer (Autochem II2920, Micromeritics, Inc., Norcross, GA), consisting of a quartz U-shaped tubular microreactor placed inside an electrical furnace. The flow rate of the helium carrier gas (50 mL/min) and the heating rate of the furnace (10 °C/min) were controlled with appropriate units. The amounts of CO and CO2 desorbed from the AC samples (0.1 g) were monitored with a quadrupole gas analyzer mass spectrometer (Omnistar, Pfeiffer Vacuum, Inc., Germany). 1631

dx.doi.org/10.1021/ef201453d | Energy Fuels 2012, 26, 1629−1637

Energy & Fuels

Article

Active sites for mercury adsorption are acidic oxygencontaining functional groups on AC surfaces and carbon lattice inside AC.21,22 Surface oxygen-functional groups on carbon materials decompose into CO and CO2 upon heating to different temperatures.23 Thus, the TPD technique was applied to understand the contribution of surface oxygen groups on AC in mercury removal. Figures 5 and 6 show the CO−TPD and

Figure 3. Comparison of the amounts of Hg0 captured by AC over 100 h under different atmospheres at 160 °C. Feed gas: Hg0 (6.47 ppb), H2S (0 or 3000 ppm), O2 (0 or 1500 ppm), H2 (25%), CO2 (5%), H2O (4%), CO (60%), and N2 (balance gas) at 300 cm3/min.

μg/g when O2 was added. This result showed that the AC sorbent exposed to H2S and H2S−O2 exhibited higher adsorption capacity. It is obvious that the presence of H2S and O2 increases the Hg0 removal efficiency of AC. To elucidate the adsorptionpromoting effects of H2S and O2, the transient method was used in this work. As shown in Figure 4, in the absence of H2S

Figure 5. TPD spectra (CO evolution) of fresh AC and used AC samples.

Figure 6. TPD spectra (CO2 evolution) of fresh AC and used AC samples.

0

Figure 4. Effect of the presence of H2S and O2 on Hg removal. Feed gas: Hg0 (6.47 ppb), H2S (0 or 3000 ppm), O2 (0 or 1500 ppm), and N2 (balance gas) at 300 cm3/min.

CO2−TPD profiles of AC before and after mercury adsorption under different atmospheres. For fresh AC, two kinds of CO desorption peaks appeared: a shoulder peak at low temperatures (from 400 to 700 °C), which may result from phenols, ethers, and carboxylic anhydrides,23−26 and a main peak at higher temperature (from 700 to 1000 °C), which may result from carbonyls and quinones.25,27 Furthermore, the CO2 peak at 530 °C may be attributed to carboxylic anhydrides and lactones on the AC surface.23,25,27 After mercury adsorption with and without H2S (AC−H2S−Hg and AC−Hg), the CO desorption peak at 400−700 °C disappeared. Meanwhile, the CO2 peak at 530 °C also disappeared. After mercury adsorption with H2S and O2 (AC−H2S−O2−Hg), the intensity of the CO desorption peak at 700−1000 °C decreased dramatically.

and O2 in the feed gas, the Hg0 removal efficiency sharply decreased from 67 to 44% after adsorption for 60 min. However, when H2S was introduced, the Hg0 removal efficiency drastically increased from 44 to 63% and then dropped to 52.5% after adsorption for 60 min. When O2 was present in the feed gas, the efficiency climbed from 52.5 to 70% and no obvious drop was observed even after adsorption for 120 min. Furthermore, after O2 was subsequently removed from the feed gas stream, the efficiency declined gradually from 69 to 51% within 60 min. The efficiency further decreased to 47.5% once H2S feeding was discontinued and decreased from 47.5 to 28% after 60 min. 1632

dx.doi.org/10.1021/ef201453d | Energy Fuels 2012, 26, 1629−1637

Energy & Fuels

Article

When these observations are combined, it can be inferred that mercury is preferentially adsorbed on the more active sites, i.e., phenols, ethers, lactones, and carboxylic anhydrides, whereas less adsorption occurs on the inactive sites, such as carbonyls and quinines. In the presence of H2S and O2, these inactive sites become active for mercury adsorption. As a result, more adsorption sites are available for mercury adsorption in the presence of H2S and O2. This could be one reason for the improvement in Hg0 removal efficiency by AC in the presence of H2S and O2. Another reason for this improvement may be the reaction between Hg0 and H2S. It has been reported that H2S adsorbed on the AC surface can react with O2 to form adsorbed S species (S*) on the surface of the AC. With the addition of O2, the reaction between Hg0 and H2S intensified. Therefore, we assume that the introduction of O2 facilitated the formation of S* from H2S on the AC surface, which may act as new active sites capable of reacting with Hg0 to form HgS. As a result, the number of available adsorption sites for Hg0 increased, leading to an improvement in the Hg0 removal efficiency. The following reactions are suggested for the capture of Hg0 on the AC surface in the presence of H2S and O2: H2S + 1/2O2 ⇔ S* + H2O

(R1)

S* + Hg ⇒ HgS

(R2)

where S* is active sulfur that is adsorbed on the surface of the AC. Thus, in the presence of H2S and O2, the following reaction could be proposed as the overall reaction for mercury conversion: H2S + 1/2O2 + Hg ⇒ HgS + H2O

(R3)

Figure 7. XPS spectra of Hg 4f for AC samples after Hg0 adsorption (a, AC−Hg; b, AC−H2S−Hg; c, AC−H2S−O2−Hg).

3.2. Mercury Adsorption Analysis. To verify the above hypothesis, the XPS spectra of Hg 4f and S 2p on the AC surface after Hg0 adsorption were compared, as shown in Figures 7 and 8. The Hg 4f spectrum (spectrum a of Figure 7) showed only one peak, and its binding energy was located at 102.5 eV. This ruled out the possibility of metallic Hg because the binding energy for Hg 4f7/2 associated with Hg0 was located at 99.85 eV. This peak may result from Hg0 that was adsorbed on the oxygen functional groups of the AC.28 In contrast, the Hg 4f spectra (spectra b and c of Figure 7) were more complex and showed two additional doublet peaks. The binding energies of the Hg 4f7/2 peaks were situated at 100.6 and 103.8 eV, which can be assigned to sulfide mercury and sulfuric mercury, respetively.28,29 This result indicated the formation of HgII surface complexes. The above result indicated a multiplicity of Hg adsorption modalities, which could be classified as “sulfide” [Hg(su1)], “sulfuric” [Hg(su2)], and “oxyhydroxide” [Hg(ox)]. Furthermore, the S 2p spectra are shown in Figure 8. The S 2p spectra of AC−H2S−Hg and AC−H2S−O2−Hg included two peaks. One, with a binding energy value of 161.9 eV, was characteristic of SII, while the other, with a binding energy value of 163.9 eV, was characteristic of elemental S. This result indicated that surface S species were in the form of SII and elemental S. The existence of SII was consistent with the values for HgS. In the S 2p spectra, the peak that is characteristic of SVI was not observed, indicating that no sulfate compounds were formed on the surface of the AC. Future studies with advanced instruments, such as X-ray absorption near edge structure (XANES), should be conducted to clarify whether

Figure 8. XPS spectra of S 2p for AC samples after Hg0 adsorption (b, AC−H2S−Hg; c, AC−H2S−O2−Hg).

and how sulfuric mercury forms on AC surfaces. For comparison, an XPS analysis of S 2p for the AC after exposure to fuel gas containing H2S (no O2 or Hg0) and the AC after exposure to fuel gas containing H2S and O2 (no Hg0) is shown in Figure 9. In both cases, only elemental sulfur was observed, as indicated by the peak at 163.9 eV. The absence of SII species indicated that H2S was oxidized to elemental sulfur on the surface of AC with or without O2. 1633

dx.doi.org/10.1021/ef201453d | Energy Fuels 2012, 26, 1629−1637

Energy & Fuels

Article

Figure 10. Effect of the temperature on Hg0 removal. Feed gas: Hg0 (6.47 ppb), H2S (3000 ppm), O2 (1500 ppm), H2 (25%), CO2 (5%), H2O (4%), CO (60%), and N2 (balance gas) at 300 cm3/min.

Figure 9. XPS spectra of S 2p for AC samples after exposure to the fuel gas (a, AC−H2S; b, AC−H2S−O2).

Figure 10. For this reason, a high Hg0 adsorption efficiency is not favored at high temperatures. 3.4. Effect of H2. H2 is a major component of coal gasification gas. The effect of the H2 content on Hg0 removal performance was investigated. Figure 11 shows the impact of

When the XPS and TPD results are combined, it can be inferred that, in the absence of H2S and O2, Hg0 could only adsorb on the oxygen functional groups on the AC surface, resulting in a moderate Hg0 removal efficiency. With increasing adsorption times, the number of available active sites would decrease on the AC surface, leading to a decrease in the Hg0 removal efficiency of AC. In contrast, when added in the reaction atmosphere, H2S could be adsorbed on the AC surface, as indicated by the presence of S2− on the AC surface (XPS result) and the decrease in surface oxygen groups on the AC (TPD result), and this adsorbed H2S could act as a source of additional active sites for removing Hg0. As a result, an improvement in the Hg0 removal efficiency of AC was observed in the presence of H2S. When both H2S and O2 were added in the reaction atmosphere, part of the adsorbed H2S on the AC surface could react with O2 to S* on the surface of the AC, which may be more active than elemental S in reactions with Hg0. The reaction between Hg0 and H2S was intensified by the presence of O2. Similar to the adsorbed H2S, the S* groups could act as active sites and react with the adsorbed Hg0 to form HgS. This mechanism would result in a further increase in the available active sites for Hg0 removal and would contribute to a higher and more stable Hg0 removal efficiency by AC. 3.3. Effect of the Temperature. To investigate the temperature effect on the performance of AC in the presence of both H2S and O2, AC adsorption studies were conducted for a 180 min duration at temperatures of 120, 160, 200, and 240 °C. The mercury concentration was 57.9 μg/m3 in the baseline simulated fuel gas with a flow rate of 300 cm3/min. The results are shown in Figure 10. It was observed that the Hg0 removal activity of AC decreased with an increasing adsorption temperature. Similar results have been reported by Liu et al.4 for S-impregnated AC. It is known that the reactions of H2S with O2 (reaction R1) and Hg0 with S* (reaction R2) are exothermic. Furthermore, the overall reaction (reaction R3) of H2S, O2, and Hg0 is also highly exothermic. Additionally, the thermodynamic calculation results shown in Table 1 indicate that the values of the equilibrium constant, K, show a sharp decrease with an increasing temperature. Because the K value at 120 °C is several orders of magnitude higher than the K values at 160, 200, and 240 °C, a larger amount of HgS can be formed at 120 °C, which explains the difference in AC performance shown in

Figure 11. Effect of H2 on Hg0 removal. Feed gas: Hg0 (6.47 ppb), H2S (3000 ppm), O2 (1500 ppm), H2 (20, 25, 30, or 35%), CO2 (5%), H2O (4%), CO (60%), and N2 (balance gas) at 300 cm3/min.

the H2 content in the simulated fuel gas on mercury uptake by AC at 160 °C. When the concentration of H2 increased from 20 to 35%, the mercury removal efficiency of AC decreased from an average of 80 to 66%. Walker et al.30 reported that H2 is preferentially retained by AC. Because of their extremely small size, hydrogen molecules can easily reach different size carbon pores to form strong hydrogen−carbon complexes. As a result, the available surface area in the mesoporous region decreases, and the reaction between mercury and active sulfur retained in those pores becomes more difficult. Moreover, H2 is readily adsorbed on the surface of AC and is protonated to H+. The formation of H+ is unfavorable with respect to eq R1, because the H+ ions formed from H2 compete with those H+ ions that arise from H2S to react with O2. With increasing concentrations of H+ ions from H2, the balance of reaction R1 shifts toward the left. Consequently, the mercury removal efficiency decreases along with the sulfuration rate of H2S. For all of these reasons, 1634

dx.doi.org/10.1021/ef201453d | Energy Fuels 2012, 26, 1629−1637

Energy & Fuels

Article

it is reasonable to assume that increasing the concentration of H2 will reduce the efficiency of Hg0 removal. 3.5. Effect of CO2. Figure 12 compares mercury removal by AC in the presence of different CO2 concentrations in the

Figure 13. Effect of H2O on Hg0 removal. Feed gas: Hg0 (6.47 ppb), H2S (3000 ppm), O2 (1500 ppm), H2 (25%), CO2 (5%), H2O (0, 4, 8, or 12%), CO (60%), and N2 (balance gas) at 300 cm3/min.

3.7. Effect of CO. Because CO is a major component of coal gasification gas, the effect of the CO content on the Hg0 removal performance is of great importance. As shown in Figure 14, increasing the CO concentration decreased the Hg0

Figure 12. Effect of CO2 on Hg0 removal. Feed gas: Hg0 (6.47 ppb), H2S (3000 ppm), O2 (1500 ppm), H2 (25%), CO2 (5, 10, 15, or 20%), H2O (4%), CO (60%), and N2 (balance gas) at 300 cm3/min.

simulated fuel gas at 160 °C. As shown in this figure, CO2 affects the performance of AC. The mercury removal efficiency decreased from an average of 77 to 56% as the concentration of CO2 increased from 5 to 20% in the simulated fuel gas. The above result indicated that mercury adsorption was suppressed by CO2, because of the decrease in the active site numbers that resulted from CO2 filling a part of the microporous structure of AC and the competitive adsorption of CO2 and Hg0 on AC. Because physisorption is limited at high adsorption temperatures, such as 160 °C, the occupation of similar chemisorption sites between Hg0 and CO2 is a possible explanation for the observed competitive adsorption. These active sites are represented mainly as lactones and carbonyls, which were located at the edges of the AC graphitic layers.31−33 The decrease in Hg0 adsorbed on the surface of the AC is unfavorable to reaction R2; thus, the presence of CO2 decreased the performance of AC. 3.6. Effect of H2O. Figure 13 shows the impact of the moisture content in the simulated fuel gas on mercury removal by AC. The temperature of the adsorption bed was 160 °C. The mercury removal efficiency of AC increased from an average of 70 to 82.5% as the concentration of H2O increased from 0 to 12%. Li et al.21 showed that interactions between H2O and carbon−oxygen complexes might create certain active sites or surface conditions favoring Hg0 adsorption. Thus, the presence of H2O is favorable to mercury removal by AC. This result suggests that the presence of H2O is favorable to Hg0 removal from a dynamic point of view, while it is unfavorable to the Hg0 removal from a thermodynamic point of view (as shown in Table 1). For the H2S system in the presence of O2, the presence of H2O greatly affected the Hg0 removal rate. The enhanced effect of H2O could be explained by the hypothesis proposed by Klein et al.34 for the catalytic oxidation of H2S on AC. The authors claimed that another possible explanation for the role of water is the formation of a film of water on the carbon surface, where H2S and O2 are dissolved and the oxidation of H2S (reaction R1) proceeds through some intermediate radicals.

Figure 14. Effect of CO on Hg0 removal. Feed gas: Hg0 (6.47 ppb), H2S (3000 ppm), O2 (1500 ppm), H2 (25%), CO2 (5%), H2O (4%), CO (40, 50, 60, or 70%), and N2 (balance gas) at 300 cm3/min.

removal efficiency at 160 °C. With the increase of the CO concentration from 40 to 50, 60, and 70% in the simulated fuel gas, the Hg0 removal efficiency decreased from an average of 82 to 80, 78, and 72%. CO is a reductive gas and does not favor the oxidation of H2S to active sulfur, which is necessary for the oxidation of Hg0. Furthermore, with an increasing temperature in the presence of CO, some surface active sulfur species may react with CO to form COS during the removal of Hg0 with H2S.22 The formation of COS is unfavorable to reaction R2, which may explain why the presence of CO lowered the performance of AC. It appears that the mercury removal unit should be located immediately prior to the wet desulfurization unit of a gas cleanup system. Generally, COS produced during the gasification of coal is converted to H2S catalytically before entering the wet gas desulfurization unit, which is not effective for the absorption of COS. Therefore, if COS is produced during the removal of Hg0 with H2S in the presence of CO, this 1635

dx.doi.org/10.1021/ef201453d | Energy Fuels 2012, 26, 1629−1637

Energy & Fuels

Article

(7) Portzer, J. W.; Albritton, J. R.; Allen, C. C.; Gupta, R. P. Development of novel sorbents for mercury control at elevated temperature in coal-derived syngas: Results of initial screening of candidate materials. Fuel Process. Technol. 2004, 85, 621−630. (8) Skodras, G.; Diamantopoulou, I.; Pantoleontos, G.; Sakellaropoulos, G. P. Kinetic studies of elemental mercury adsorption in activated carbon fixed bed reactor. J. Hazard. Mater. 2008, 158 (1), 1−13. (9) Yan, R.; Liang, D. T.; Tay, J. H. Control of mercury vapor emissions from combustion flue gas. Environ. Sci. Pollut. Res. 2003, 10 (6), 399−407. (10) Lee, S. H.; Park, Y. O. Gas-phase mercury removal by carbonbased sorbents. Fuel Process. Technol. 2003, 84 (1−3), 197−206. (11) Hsi, H. C.; Rood, M. J.; Rostam-Abadi, M.; Chen, S. G.; Chang, R. S. Effects of sulfur impregnation temperature on the properties and mercury adsorption capacities of activated carbon fibers (ACFs). Environ. Sci. Technol. 2001, 35 (13), 2785−2791. (12) Korpiel, J. A.; Vidic, R. D. Effect of sulfur impregnation method on activated carbon uptake of gas-phase mercury. Environ. Sci. Technol. 1997, 31 (8), 2319−2325. (13) Liu, W.; Vidic, R. D.; Brown, T. D. Impact of flue gas conditions on mercury uptake by sulfur-impregnated activated carbon. Environ. Sci. Technol. 2000, 34 (1), 154−159. (14) Kwon, S.; Vidic, R. D. Evaluation of two sulfur impregnation methods on activated carbon and bentonite for the production of elemental mercury sorbents. Environ. Eng. Sci. 2000, 17 (6), 303−313. (15) Feng, W. G.; Borguet, E.; Vidic, R. D. Sulfurization of a carbon surface for vapor phase mercury removalI: Effect of temperature and sulfurization protocol. Carbon 2006, 44 (14), 2990−2997. (16) Smith, J. M.; Ness, H. C. V. Introduction to Chemical Engineering Thermodynamics, 4th ed.; McGraw-Hill: New York, 1987; p 105. (17) Chase, M. W. Jr.; Davles, C. A.; Downey, J. R. Jr.; Frurip, D. J.; McDonald, R. A.; Syverud, A. N. J. JANAF thermochemical tables, third edition. Phys. Chem. Ref. Data 1985, 1318. (18) Hedding, K.; Rao, B. R. VDI-Bericht 253 S, 37/42; VDI-Verlag: Düsseldorf, Germany, 1976. (19) Wang, S. X.; Zhang, L.; Li, G. H.; Wu, Y.; Hao, J. M.; Pirrone, N.; Sprovieri, F.; Ancora, M. P. Mercury emission and speciation of coal-fired power plants in China. Atmos. Chem. Phys. 2010, 10, 1183− 1192. (20) Lopez-Antón, M. A.; Tanscón, J. M.; Martinez-Tarazona, M. R. Retention of mercury in activated carbons in coal combustion and gasification flue gases. Fuel Process. Technol. 2002, 77−78, 353−358. (21) Li, Y. H.; Lee, C. W.; Gullett, B. K. The effect of activated carbon surface moisture on low temperature mercury adsorption. Carbon 2002, 40 (1), 65−72. (22) Wu, S. J.; Oya, N.; Ozaki, M.; Kawakami, J.; Uddin, M. A.; Sasaoka, E. Development of iron oxide sorbents for Hg0 removal from coal derived fuel gas: Sulfidation characteristics of iron oxide sorbents and activity for COS formation during Hg0 removal. Fuel 2007, 86 (17−18), 2857−2863. (23) Figueiredo, J. L.; Pereira, M. F. R.; Freitas, M. M. A.; Ó rfão, J. J. M. Modification of the surface chemistry of activated carbons. Carbon 1999, 37, 1379−1389. (24) Zhuang, Q. L.; Kyotani, T.; Tomita, A. The change of TPD pattern of O2-gasified carbon upon air exposure. Carbon 1994, 32, 539−540. (25) Zielke, U.; Hüttinger, K. J.; Hoffman, W. P. Surface-oxidized carbon fibers: I. Surface structure and chemistry. Carbon 1996, 34, 983−998. (26) Marchon, B.; Carrazza, J.; Heinemann, H.; Somorjai, G. A. TPD and XPS studies of O2, CO2, and H2O adsorption on clean polycrystalline graphite. Carbon 1988, 26, 507−514. (27) Zhuang, Q. L.; Kyotani, T.; Tomita, A. DRIFT and TK/TPD analyses of surface oxygen complexes formed during carbon gasification. Energy Fuels 1994, 8, 714−718. (28) Ehrhardt, J. J.; Behra, P.; Bonnissel-Gissinger, P.; Alnot, M. XPS study of the sorption of Hg(II) onto pyrite FeS2. Surf. Interface Anal. 2000, 30 (1), 269−272.

process cannot be put into practical use unless a COS converter is installed after the Hg0 removal unit. Thus, further research should be conducted with a view toward understanding and reducing the formation of COS.

4. CONCLUSION For simulated coal-derived fuel gas, the Hg0 removal performance of AC was significantly improved in the presence of H2S, as indicated by the drastic increase in the Hg0 adsorption capacity (from 50.4 to 210.3 μg/g). The capacity further increased to 390.8 μg/g in the presence of O2, implying that O2 promoted Hg0 removal from H2S-containing fuel gases. The TPD results indicate that the inactive mercury adsorption sites on the AC become active in the presence of H2S and O2. Furthermore, the XPS results show that H2S could react with O2 to form active sulfur on the surface of AC, which could react with Hg0 to form the stable and nontoxic compound HgS. Active sulfur acted as a source of new active sites for mercury adsorption. These mechanisms are responsible for the promotional effect of H2S and O2 on the Hg0 removal efficiency of AC. In the H2S + O2 coal-derived fuel gas system, the Hg0 removal was favored at lower temperatures (160−240 °C). Because of the competitive adsorption of CO2 and Hg0 on similar adsorption sites, the presence of CO2 suppressed Hg0 removal. CO and H2 decreased the Hg0 adsorption efficiency of AC because they inhibited the formation of active S species from H2S and covered the mesoporous region by forming hydrogen−carbon complexes. In contrast, the presence of H2O favored Hg0 removal through the formation of new active sites.



AUTHOR INFORMATION

Corresponding Author

*Telephone: 86-351-2021137. Fax: 86-351-2021137. E-mail: [email protected]. Notes

The authors declare no competing financial interest.



ACKNOWLEDGMENTS The authors gratefully acknowledge the financial support of the National High Technology Research and Development Planning of China (HTRDP; 2007AA05Z325) and the National Natural Science Foundation of China (20706055).



REFERENCES

(1) Granite, E. J.; Myers, C. R.; King, W. P.; Stanko, D. C.; Pennline, H. W. Sorbents for mercury capture from fuel gas with application to gasification systems. Ind. Eng. Chem. Res. 2006, 45, 4844−4848. (2) Uddin, M. A.; Ozaki, M.; Sasaoka, E.; Wu, S. J. Temperatureprogrammed decomposition desorption of mercury species over activated carbon sorbents for mercury removal from coal-derived fuel gas. Energy Fuels 2009, 23, 4710−4716. (3) Eswaran, S.; Stenger, H. G.; Fan, Z. Gas-phase mercury adsorption rate studies. Energy Fuels 2007, 21 (2), 852−857. (4) Liu, W.; Vidic, R. D.; Brown, T. D. Impact of flue gas conditions on mercury uptake by sulfur-impregnated activated carbon. Environ. Sci. Technol. 2000, 34 (1), 154−159. (5) Li, Y. H.; Lee, C. W.; Gullett, B. K. Importance of activated carbon’s oxygen surface functional groups on elemental mercury adsorption. Fuel 2003, 82 (4), 451−457. (6) Lu, D. Y.; Granatstein, D. L.; Rose, D. J. Study of mercury speciation from simulated coal gasification. Ind. Eng. Chem. Res. 2004, 43, 5400−5404. 1636

dx.doi.org/10.1021/ef201453d | Energy Fuels 2012, 26, 1629−1637

Energy & Fuels

Article

(29) Huston, N. D.; Attwood, B. C.; Scheckel, K. G. XAS and XPS characterization of mercury binding on brominated activated carbon. Environ. Sci. Technol. 2007, 41 (5), 1747−1752. (30) Chemistry and Physics of Carbon; Walker, P. L., Jr., Ed.; Marcel Dekker: New York, 1970; Vol. 6, p 264. (31) Diamantopoulou, I.; Skodras, G.; Sakellaropoulos, G. P. Sorption of mercury by activated carbon in the presence of flue gas component. Fuel Process. Technol. 2010, 91 (2), 158−163. (32) Montoya, A.; Mondragón, F.; Truong, T. N. CO2 adsorption on carbonaceous surfaces: A combined experimental and theoretical study. Carbon 2003, 41 (1), 29−39. (33) Gauden, P. A.; Wiśniewski, M. CO2 sorption on substituted carbon materials: Computational chemistry studies. Appl. Surf. Sci. 2007, 253 (13), 5276−5731. (34) Klein, J.; Henning, K. D. Catalytic oxidation of hydrogen sulphide on activated carbons. Fuel 1984, 63 (8), 1064−1067.

1637

dx.doi.org/10.1021/ef201453d | Energy Fuels 2012, 26, 1629−1637