Mercury adsorption characteristics of carbon sorbent with low surface area

ABSTRACT

Several studies have been conducted to decrease the cost of sorbents used for the control of mercury emissions. Thus far, several sorbents with low surface areas have been reported to exhibit promising mercury removal capacities. However, based on the results reported, it is difficult to understand the mechanisms of adsorption and oxidization of elemental mercury on sorbents with low surface areas compared to those with higher surface areas. Three types of materials with different surface areas were evaluated herein for use as carbon sorbents for the adsorption of elemental mercury: (1) coal, (2) sewage sludge, and (3) unburned carbon. The respective raw sorbents and FeCl₃-impreganted congeners were evaluated. Each sorbent was tested in a fixed-bed reactor system under two simulated flue gas conditions (1) without and (2) with 20 ppm hydrogen chloride (HCl). The injection of HCl increased the mercury adsorption efficiency of all tested sorbents by decreasing the emission of elemental mercury. Doping the sorbent with FeCl₃ increased the mercury adsorption efficiency during the earlier test period under both simulated flue gas conditions (without and with HCl). FeCl₃-impregnated activated carbon and FeCl₃-impregnated unburned carbon emitted large amounts of oxidized mercury during the later test periods.

Implications: We tested three types of sorbents to investigate the mercury adsorption characteristics of sorbents with low surface area. The mercury adsorption test was conducted by varying the raw material of the sorbent, chemical impregnation of the sorbent and the simulated flue gas composition. We found that HCl in simulated flue gas increased the mercury adsorption efficiency of both the raw and FeCl₃-impregnated sorbents by decreasing the emission of elemental mercury.

Introduction

Mercury has been reported to exist in three forms in combustion flue gases: elemental mercury (Hg⁰), oxidized mercury (Hg²⁺), and particulate mercury (Hg^P) (Galbreath and Zygarlicke 2000; Granite, Pennline, and Senior 2015; Lee and Wilcox 2017; UNEP 2013; Wilcox et al. 2012). Among the three forms of mercury, elemental mercury is difficult to remove using existing air pollution control devices. Injection of activated carbon has been suggested as an effective and simple method of controlling the adsorption of elemental mercury from combustion flue gases (Fuente-Cuesta et al. 2015; Pavlish et al. 2003; Wilcox et al. 2011; Yang et al. 2019). Many studies have been conducted to decrease the cost of sorbents (González-García 2018; Xu et al. 2018; Yu et al. 2016). Thus, many inexpensive materials have been tested and suggested as raw sorbent materials to control mercury emissions, such as coconut husk (Johari et al. 2016), waste wood (Yu et al. 2016), fruit peel (Bhakta et al. 2017), fly ash (Yang et al. 2019), sewage sludge (Park and Lee 2018), and gasification char (Fuente-Cuesta et al. 2012). Materials with a high content of bromine or chlorine have been reported to have high Hg⁰ adsorption capacities (Fuente-Cuesta et al. 2012). In addition, chemically promoted sorbents demonstrate good performance in Hg⁰ adsorption (Hu et al. 2006; Liu, Xu, and Liu 2020; Qi et al. 2015; Wang et al. 2018, 2015).

Bamboo charcoal with a surface area of 91 m²/g exhibited very low efficiency for adsorbing elemental mercury, whereas FeCl₃-impregnated bamboo charcoal afforded almost 100% elemental mercury adsorption efficiency under 6% oxygen (O₂), 2800 mg/m³ sulfur dioxide (SO₂), 130 mg/m³ nitrogen monoxide (NO), 12% carbon dioxide (CO₂), and nitrogen (N₂) (Tan, Niu, and Chen 2015). Fly ash afforded approximately 40% elemental mercury adsorption efficiency, whereas cupric chloride (CuCl₂)- and FeCl₃-impregnated fly ash afforded elemental mercury adsorption efficiencies of 95% and 75%, respectively, under N₂ gas environment (Xu et al. 2013). In addition, fly ash promoted with manganese acetate (Mn(CH₃COOH)₃ · 4H₂O) and iron(III) nitrate (Fe(NO₃)₃) afforded 98% elemental mercury removal efficiency, whereas raw fly ash showed 20% elemental mercury removal efficiency under air environment (Xing, Xu, and Zhong 2012). It has been reported that the elemental mercury removal efficiency of a sorbent can be enhanced by the oxidation of elemental mercury via iron and Cl present in the sorbent (Han et al. 2016; Hou et al. 2013; Liu, Xu, and Liu 2020; Shu, Lu, and He 2013; Tan et al. 2012).

The elemental mercury adsorption efficiencies of the sorbents are also significantly affected by the flue gas composition. Tire pyrolysis char showed 90% elemental mercury adsorption efficiency when exposed to simulated flue gas consisting of 490 ppm NO, 17 ppm nitrogen dioxide (NO₂), 3.5 μg/m³ Hg⁰, and N₂ balance, whereas it afforded 47% efficiency under N₂ environment. When 6% O₂, 12% CO₂, and 1800 ppm SO₂ were added to the simulated flue gas, the efficiency decreased to 80%. When 200 ppm of hydrogen chloride (HCl) was added to the simulated flue gas, the efficiency increased to approximately 99% (Seneviratne et al. 2007). It has been reported that SO₂ inhibits elemental mercury adsorption because SO₂ competes with elemental mercury for the active site of the sorbent (Xu et al. 2013; Zhou et al. 2018) on the other hand, NO increased the elemental mercury adsorption efficiency of the sorbent (Yang et al. 2014; Zhou et al. 2018). In particular, HCl gas plays an important role in increasing the elemental mercury adsorption efficiency of sorbents (Diamantopoulou, Skodras, and Sakellaropoulos 2010; Min et al. 2017; Olson et al. 2004; Tong et al. 2015; Yang et al. 2007). In a previous study, we also showed that sewage sludge activated carbon sorbents with a low surface area of 4–64 m²/g afforded high elemental mercury adsorption capability with the injection of HCl gas (Park and Lee 2018). Many researchers have reported that sorbents with low surface areas have promising capability for the control of elemental mercury. Specifically, these sorbents demonstrated high elemental mercury adsorption efficiency in the presence of HCl gas. However, most previous studies investigated only elemental mercury emissions after adsorption onto sorbents. Based on the results reported, it is difficult to understand the effect of low surface area sorbents on elemental mercury oxidation and adsorption compared to other sorbents with higher surface areas.

The mercury adsorption characteristics of sorbents with low surface areas are investigated in this study by varying the raw material, chemical promotion of the sorbent, and simulated flue gas composition. The emission of both elemental and oxidized mercury after the adsorption of elemental mercury on the carbon sorbents is investigated. Three types of materials are used to make the carbon sorbents for the adsorption of elemental mercury: (1) coal, (2) sewage sludge, and (3) unburned carbon. The sorbent made from each material is tested as a raw sorbent or as the FeCl₃-impregnated congener. Each sorbent was also tested under two simulated flue gas conditions, i.e., with and without HCl gas.

Experimental

Sorbent preparation

Sewage sludge and coal were used as raw materials to prepare the activated carbon sorbents. Using a rotary kiln reactor, activated carbon was manufactured through the carbonization and activation processes. Carbonization of the raw material was performed under N₂ gas environment. The carbonized material was activated by injecting water vapor. A detailed description of the manufacturing method can be found in our previous publication (Min, Ahmad, and Lee 2017; Park and Lee 2018). Sewage sludge was selected as the raw material for generating activated carbon with a low surface area. The sewage sludge was carbonized at 400°C for 1 h and then activated at 850°C for 1 h by injecting water vapor. The Brunauer–Emmett–Teller (BET) surface area of the activated carbon made from sewage sludge (sludge AC) was as low as 40.5 m²/g, as shown in Table 1.

Table 1. Physical properties of prepared samples

	BET surface area (m^2/g)	Mesopore volume (cm^3/g)	Micropore volume (cm^3/g)
Coal AC	219.4	0.067	0.074
FeCl3-coal AC	86.3	0.074	0.019
Sludge AC	40.5	0.049	0.008
FeCl3-sludge AC	34.0	0.066	0.004
Unburned C	6.5	0.020	N.D.
FeCl3-unburned C	6.1	0.016	N.D.

Coal was selected to produce activated carbon with a high surface area. Coal is a general raw material for activated carbon used for the control of mercury emissions. In this study, coal was carbonized at 700°C for 30 min and then activated at 900°C for 30 min. As shown in Table 1, the BET surface area of the activated carbon made from coal (coal AC), 219.4 m²/g, was significantly higher than that of the sludge AC.

Unburned carbon (unburned C) from fly ash was selected as the non-activated sorbent with a low surface area. Unburned carbon has been reported to show mercury adsorption ability (Xu et al. 2013; Yu et al. 2016). Table 2 shows the ultimate analysis data for sludge AC, coal AC, and unburned C, respectively, acquired using an elemental analyzer (Vario MICRO cube, Elementar, Germany).

Table 2. Results of ultimate analysis

	C (%)	H (%)	N (%)	S (%)
Coal AC	72.32	1.03	0.75	0.90
Sludge AC	8.30	0.30	0.39	0.39
Unburned C	13.03	0.32	0.31	0.52

Each sorbent was impregnated with FeCl₃ (≥97.0%, Junsei, Japan). FeCl₃ (0.5 g) was first dissolved in 100 mL of isopropyl alcohol (≥99.9%, Kanto Chemical, Japan). Each sorbent (4.5 g) was then added to the FeCl₃ solution and stirred for 2 h. The solution was filtered, and the filtered solid was dried at 100°C. The Fe content of the filtrate was analyzed using ion chromatography (Dionex Integrion + Dionex Aquion, Thermo Scientific, USA). The concentration of FeCl₃ on the sorbent was 7.3, 4.5, and 1.0% for coal AC, sludge AC, and unburned C, respectively. This shows that the concentration of FeCl₃ impregnated on the sorbent is largely dependent on the surface area of the sorbent. The BET surface area, micropore volume, and mesopore volume of each sorbent were determined using an analyzer (BELSORP-mini, BEL Japan, Inc., Japan). The results are shown in Table 1.

Mercury adsorption test

Six sorbents were tested for the adsorption of elemental mercury using a fixed-bed reactor system. The adsorption tests were performed by injecting 1 L/min simulated flue gas with two compositions: 12% CO₂, 5% O₂, 7% H₂O, 65–75 μg/m³ Hg⁰, 500 ppm SO₂, and 200 ppm NO (1) without and (2) with 20 ppm HCl. The sorbent (30 mg) was pre-mixed with 6 g silica granules (Model S150-3, Fisher Scientific, USA), placed in a fixed-bed reactor at 140°C, and tested for 280 min. A detailed description of the experimental system is provided in our previous publication (Park and Lee 2018). The Ontario Hydro Method was used to determine the concentrations of oxidized mercury and elemental mercury, respectively. The outlet gas after flowing through the fixed-bed reactor first entered a 1 N potassium chloride (KCl) impinger train to absorb oxidized mercury and then entered a 4% (w/w) potassium permanganate (KMnO₄) and 10% sulfuric acid (H₂SO₄) impinger train to absorb elemental mercury. The concentrations of the outlet elemental mercury ($H{g}_{out}^0$) and outlet oxidized mercury ($H{g}_{out}^{2+}$) were then determined. The impinger trains were changed to new impinger trains at test times of 10, 20, 30, 40, 160, and 280 min. During the first 40 min of the test, the impinger trains were changed every 10 min because the outlet mercury concentrations varied faster during the earlier time of the test. Subsequently, the impinger trains were changed every 120 min to investigate the mercury emission over a longer period of time. The concentration of inlet elemental mercury ($H{g}_{in}^0$) was determined by analyzing the simulated flue gas flowing through the bypass line before and after each test. At each time interval, the oxidized mercury emission percentage ($H{g}_{out}^{2+}\left(\mathrm{\%}\right)$) and elemental mercury emission percentage ($H{g}_{out}^0\left(\mathrm{\%}\right)$) were determined from the following equations, respectively:

(1) $H{g}_{out}^{2+}\left(\mathrm{\%}\right)=\frac{H{g}_{out}^{2+}}{H{g}_{in}^0}\times 100$(1)

(2) $H{g}_{out}^0\left(\mathrm{\%}\right)=\frac{H{g}_{out}^0}{H{g}_{in}^0}\times 100$(2)

The mercury adsorption efficiency was also determined from the following equation:

(3) $Mercuryadsorptionefficiency\left(\mathrm{\%}\right)=100-H{g}_{out}^{2+}\left(\mathrm{\%}\right)-H{g}_{out}^0\left(\mathrm{\%}\right)$(3)

All tests were performed in duplicate, and the average and standard error of the results were calculated.

Results and discussion

Analysis of coal AC

Coal AC, which has the highest surface area among the sorbents evaluated in this study, was first tested for the adsorption of elemental mercury. Figures 1 and Figures 2 show the emission of oxidized mercury and elemental mercury from coal AC and FeCl₃-coal AC, respectively. Because these sorbents demonstrated high mercury adsorption efficiencies, the y axes of Figures 1 and Figures 2 span 0 to 20%.

Figure 1. Emission of oxidized and elemental mercury from (a) coal AC and (b) FeCl₃-coal AC under simulated flue gas without HCl

Figure 2. Emission of oxidized and elemental mercury from (a) coal AC and (b) FeCl₃-coal AC under simulated flue gas with HCl

When exposed to simulated flue gas without HCl, the emission of elemental mercury from coal AC was in the range of 2.7‒3.8%, accompanied by minimal oxidized mercury emission, for the overall test period, as shown in Figure 1a. The emission of elemental mercury from FeCl₃-coal AC decreased significantly to less than 1% for the test time of 0–40 min, as shown in Figure 1b. However, the emission of oxidized mercury from FeCl₃-coal AC was higher than that from coal AC during the test time of 40–280 min. This shows that elemental mercury is oxidized by FeCl₃, but some of the resultant oxidized mercury is not adsorbed onto the sorbent. This suggests that the raw carbon surface on FeCl₃-coal AC may be responsible for the re-adsorption of the resultant oxidized mercury generated from the reaction between FeCl₃ and elemental mercury (Lee, Lee, and Keener 2009).

Figure 2a,b show the emission of oxidized mercury and elemental mercury from the coal AC and FeCl₃-coal AC, respectively, using simulated flue gas with HCl. Comparison of the results in Figures 1a and Figures 2a demonstrates that coal AC exhibited higher mercury adsorption efficiency when HCl was contained in the simulated flue gas. Comparison of the data in Figures 1b and Figures 2b illustrates that FeCl₃-coal AC also showed higher mercury adsorption efficiency with the injection of HCl. In particular, FeCl₃-coal AC showed significantly lower elemental mercury emissions in the later test period when HCl was contained in the simulated flue gas. This indicates that the presence of HCl gas is effective for the adsorption of elemental mercury onto the raw carbon surface of the FeCl₃-coal AC.

Analysis of sludge AC

Sludge AC and FeCl₃-sludge AC were also tested under simulated flue gas without and with HCl, respectively. Sludge AC has been reported to have Hg⁰ oxidation capacity because of its large ash content, including metal oxides (Park and Lee 2018). However, sludge AC emitted a large amount of oxidized mercury under simulated flue gas without HCl, as shown in Figure 3a. This may be ascribed to the low BET surface area of the sludge AC. On the other hand, FeCl₃-sludge AC showed high mercury adsorption efficiencies of more than 95% within the first 30 min, as shown in Figure 3b. The tests for FeCl₃-coal AC showed that doping sludge AC with FeCl₃ was also effective in increasing its mercury adsorption efficiency during the first tens of minutes. In addition, similar to the results for coal-AC and FeCl₃-coal AC, FeCl₃-sludge AC showed higher oxidized mercury emissions than sludge AC after 30 min. Although the oxidation capacity of sludge AC was enhanced by doping with FeCl₃, a large amount of oxidized mercury was not adsorbed onto the FeCl₃-sludge AC. Because of the low raw carbon surface area of FeCl₃-sludge AC, re-adsorption of the resultant oxidized mercury was limited after 30 min.

Figure 3. Emission of oxidized and elemental mercury from (a) sludge AC and (b) FeCl₃-sludge AC under simulated flue gas without HCl

With the injection of HCl, the mercury adsorption efficiency of sludge AC increased significantly within the first 20 min, as shown in Figure 4a. Comparison of the results after 30 min in Figures 3a and Figures 4a shows that the elemental mercury emissions decreased and the oxidized mercury emissions increased with the injection of HCl. The decreased elemental mercury emissions may be ascribed to an increase in the adsorption of elemental mercury onto the raw carbon surface of sludge AC in the presence of HCl. The increased oxidized mercury emissions may be ascribed to the limited raw carbon surface area of sludge AC at the later time of the test. Figure 4b shows the oxidized mercury and elemental mercury emission data for the FeCl₃-sludge AC. Similar to the results for sludge AC, FeCl₃-sludge AC also showed high mercury adsorption efficiency during the earlier time of the test. Comparing the results after 40 min in Figures 3b and Figures 4b shows that FeCl₃-sludge AC exhibited significantly lower elemental mercury emission when HCl was contained in the simulated flue gas, as also observed for FeCl₃-coal AC. This again indicates that HCl gas is effective in increasing the adsorption of elemental mercury onto the raw carbon surface of FeCl₃-sludge AC.

Figure 4. Emission of oxidized and elemental mercury from (a) sludge AC and (b) FeCl₃-sludge AC under simulated flue gas with HCl

Analysis of unburned carbon

Unburned C and FeCl₃-unburned C were tested under simulated flue gas conditions with and without HCl. Figure 5a,b show the mercury emission as a function of the test time for unburned C and FeCl₃-unburned C, respectively, under simulated flue gas without HCl. Unlike coal AC and sludge AC, unburned C did not demonstrate mercury adsorption capacity. This may be ascribed to the very low surface area and non-detectable micropore volume of unburned C. FeCl₃-unburned C showed higher mercury adsorption efficiency than unburned C during the earlier time of the test. However, the minimal increase in the mercury adsorption efficiency of FeCl₃-unburned C may be ascribed to the low concentration of FeCl₃ impregnated on the unburned C.

Figure 5. Emission of oxidized and elemental mercury from (a) unburned C and (b) FeCl₃-unburned C under simulated flue gas without HCl

Figure 6a,b show the emission of mercury as a function of time for the unburned C and FeCl₃-unburned C, respectively, under simulated flue gas with HCl. The injection of HCl increased the mercury adsorption efficiency of both unburned C and FeCl₃-unburned C by decreasing the emission of elemental mercury, as also observed for coal AC and sludge AC. In addition, the mercury adsorption efficiency of FeCl₃-unburned C was higher than that of unburned C, especially during the earlier time of the test, as observed for FeCl₃-coal AC and FeCl₃-sludge AC. However, the emission of oxidized mercury was higher for FeCl₃-unburned C than for unburned C. This indicates that the re-adsorption of the resultant oxidized mercury onto the raw carbon surface of FeCl₃-unburned C is limited owing to its very low surface area. Although the unburned C sorbents demonstrated significantly lower mercury adsorption efficiency than coal AC and sludge AC, their Hg⁰ oxidation and adsorption properties were similar to that of the coal AC and sludge AC sorbents depending on the chemical promotion and simulated flue gas composition.

Figure 6. Emission of oxidized and elemental mercury from (a) unburned C and (b) FeCl₃-unburned C under simulated flue gas with HCl

Conclusion

Coal AC, sludge AC, unburned C, and their FeCl₃-impregnated counterparts were tested as sorbents for elemental mercury under two simulated flue gas conditions (with and without HCl gas, respectively). For all tested sorbents, HCl gas was effective in increasing the mercury adsorption efficiency of both the raw and FeCl₃-impregnated sorbents by decreasing the emission of elemental mercury. The FeCl₃-impregnated sorbents demonstrated higher mercury adsorption efficiency than the raw sorbents, especially during the earlier test period. The mercury adsorption efficiency of coal AC with a surface area of 219.4 m²/g exceeded 95% for the duration of the test when exposed to simulated flue gas with HCl. However, sludge AC with a surface area of 40.5 m²/g showed high mercury adsorption efficiency under simulated flue gas conditions with HCl during the first 20 min of the test. The mercury adsorption efficiency of FeCl₃-sludge AC with a surface area of 34 m²/g exceeded 95% under both simulated flue gas conditions (without and with HCl) during the first 30 min of the test. Considering the particle residence times in full-scale particulate control devices, the sludge AC sorbent is promising for the control of mercury emissions. In addition, this can be a promising alternative method to ocean dumping for the disposal of sewage sludge. However, pilot- or full-scale tests are still recommended to confirm if the sludge AC sorbents are technically and economically feasible to be used for the control of mercury emissions. FeCl₃-sludge AC and FeCl₃-unburned C emitted large amounts of oxidized mercury during the later test periods. This indicates that FeCl₃ impregnated on the sorbent effectively oxidizes elemental mercury, but the raw carbon surface of the sorbent may be responsible for the re-adsorption of the resultant oxidized mercury. The unburned C sorbents demonstrated significantly lower mercury adsorption efficiencies than coal AC and sludge AC because of the low surface areas of the former: 6.5 m²/g for unburned C and 6.1 m²/g for FeCl₃-unburned C. However, the Hg⁰ oxidation and adsorption behavior of these sorbents was similar to that of the coal AC and sludge AC sorbents depending on the chemical promotion and simulated flue gas composition.

Disclosure statement

No potential conflict of interest was reported by the author(s).

Additional information

Funding

This research was supported by the Basic Science Research Program through the National Research Foundation of Korea (NRF) funded by the Ministry of Education [NRF-2018R1D1A1A09083083].

Notes on contributors

Jeongmin Park

Jeongmin Park is a graduate student in the Department of Environmental Engineering at Chungbuk National University, Korea.

Sang-Sup Lee

Sang-Sup Lee is a professor in the Department of Environmental Engineering at Chungbuk National University, Korea.