activated carbon sorbents for mid-temperature capture of mercury from coal-derived fuel gas

JO U R N A L OF EN V I RO N M EN T A L S CI E NC ES 2 6 ( 20 1 4 ) 1 4 9 7–1 5 0 4

Available online at www.sciencedirect.com

ScienceDirect www.elsevier.com/locate/jes

Pd/activated carbon sorbents for mid-temperature capture of mercury from coal-derived fuel gas Dekui Li1 , Jieru Han1 , Lina Han2,⁎, Jiancheng Wang1 , Liping Chang1,⁎ 1. Key Laboratory of Coal Science and Technology, Ministry of Education and Shanxi Province, Taiyuan University of Technology, Taiyuan 030024, China 2. College of Materials Science and Engineering, Taiyuan University of Technology, Taiyuan 030024, China

AR TIC LE I N FO

ABS TR ACT

Article history:

Higher concentrations of Hg can be emitted from coal pyrolysis or gasification than from

Received 6 August 2013

coal combustion, especially elemental Hg. Highly efficient Hg removal technology from

Revised 20 November 2013

coal-derived fuel gas is thus of great importance. Based on the very excellent Hg removal

Accepted 28 January 2014

ability of Pd and the high adsorption abilities of activated carbon (AC) for H2S and Hg, a

Available online 12 June 2014

series of Pd/AC sorbents was prepared by using pore volume impregnation, and their

Keywords:

laboratory-scale fixed-bed reactor. The effects of loading amount, reaction temperature and

Coal-derived fuel gas

reaction atmosphere on Hg removal from coal-derived fuel gas were studied. The sorbents

performance in capturing Hg and H2S from coal-derived fuel gas was investigated using a

Hg removal

were characterized by N2 adsorption, X-ray diffraction (XRD) and X-ray photoelectron

Pd/AC sorbents

spectroscopy (XPS). The results indicated that the efficiency of Hg removal increased with

H2S removal

the increasing of Pd loading amount, but the effective utilization rate of the active component Pd decreased significantly at the same time. High temperature had a negative influence on the Hg removal. The efficiency of Hg removal in the N2–H2S–H2–CO–Hg atmosphere (simulated coal gas) was higher than that in N2–H2S–Hg and N2–Hg atmospheres, which showed that H2 and CO, with their reducing capacity, could benefit promote the removal of Hg. The XPS results suggested that there were two different ways of capturing Hg over sorbents in N2–H2S–Hg and N2–Hg atmospheres. © 2014 The Research Center for Eco-Environmental Sciences, Chinese Academy of Sciences. Published by Elsevier B.V.

Introduction During the combustion and gasification of coal, mercury can be mostly emitted into the vapor phase. Mercury can travel a long distance and is easily accumulated in the organisms. Once Hg is converted into methylmercury by bacteria in the sediments, it enters the food chain and becomes hazardous to wildlife and humans (Otero-Rey et al., 2003). It has been reported that more amounts and higher concentrations of elemental Hg can be emitted from coal pyrolysis or gasification than that from coal combustion (Yan et al., 2004). There are three forms of mercury that may be present in the coal-derived fuel or flue gas, namely, elemental mercury (Hg0), oxidized mercury (Hg2+, Hg+) and

particle-bound-mercury (HgP) (Presto and Granite, 2006). Oxidized mercury can be removed by wet flue gas desulfurization (WFGD) facilities, and particle-bound-mercury can be captured along with fly ash by the typical air pollution control devices (APCD) such as electrostatic precipitators and bag houses (Yan et al., 2005, 2011). However, elemental mercury (Hg0 ) vapor is very difficult to capture with typical APCD because Hg is highly volatile and insoluble in water (Ozaki et al., 2008). Coal gasification is the leading technology in clean coal conversion and will be widely used in the next ten or twenty years (Pavlish et al., 2010). Therefore, highly efficient Hg control technology, especially for coal gasification, is of great importance.

⁎ Corresponding authors. E-mails: [email protected] (Lina Han); [email protected] (Liping Chang).

http://dx.doi.org/10.1016/j.jes.2014.05.016 1001-0742/© 2014 The Research Center for Eco-Environmental Sciences, Chinese Academy of Sciences. Published by Elsevier B.V.

1498

J O U R NA L OF EN V I RO N M EN T A L S CI EN C ES 2 6 ( 20 1 4 ) 1 4 97–1 5 0 4

Recently, some researchers found that several noble metals, such as Ag, Au, Pd and Pt, etc. showed high Hg0 removal efficiency in the mid-temperature range of 150–350°C (Baltrus et al., 2008; Granite et al., 2006). Among these metal-based sorbents, Pd showed the highest removal ability of Hg0 according to modeling and experimental results (Granite et al., 2006). Moreover, Pd supported on Al2O3 can capture Hg at mid-temperatures (Han et al., 2012). As we know, the content of H2S, another harmful gas generated during the coal gasification process (Liu et al., 1998), is several orders of magnitude higher than that of mercury. It has been reported that H2S has a negative effect on the capture of Hg over Pd/Al2O3 sorbents (Han et al., 2012). Activated carbon (AC) can be used as a catalyst support due to the abundant pore structure and large BET surface area. Furthermore, it also possesses high adsorption abilities (Shen et al., 2010; Bandosz, 2002; Murakami et al., 2010). It has been reported that AC, as well as AC impregnated with sulfur, chlorine and iodine, and active metal oxides, was effective for Hg removal (Baltrus et al., 2011; Feng et al., 2006; Kamakoti et al., 2005; Poulston et al., 2007; Hu et al., 2010; Sun et al., 2013). In this work, Pd and AC were selected as the active metal and support, respectively. Pd/AC sorbents were prepared by pore volume impregnation and the removal of Hg and H2S from simulated coal gas was investigated using a laboratory-scale fixed-bed reactor. N2 adsorption, X-ray diffraction (XRD), transmission electron microscopy (TEM) and X-ray photoelectron spectroscopy (XPS) characteristic technologies were used to analyze the physical and chemical properties of the fresh and used sorbents, and a possible approach to capture Hg over Pd/AC sorbents was discussed.

1. Experimental

500°C for 5 hr. The 0.5Pd/AC, 1Pd/AC and 3Pd/AC sorbents were prepared by loading 0.5 wt.%, 1.0 wt.%, 3.0 wt.% Pd on the AC, respectively.

1.2. Adsorption test The adsorption tests were carried out using a fixed-bed quartz flow reactor at atmospheric pressure (Fig. 1). The test apparatus includes five sections: gas system, Hg generating device, fixed bed reactor and online mercury analyzer (Lumex-Marketing JSC, Russia) as the detection system. The Hg generator consists of a mercury permeation tube (Valco Instruments Company Inc., Houston, TX, USA) placed in a constant water bath system at 40°C. Hg vapor was brought into the reactor using ultra high purity N2 as a carrier. The flow rate of each gas stream was accurately controlled by the mass flow controller. Sorbents (0.50 g) were placed in a quartz tube reactor (5.0 mm inner diameter and 64.2 cm length). The gas flow rate was 470 mL/min and was controlled by a mass flow controller where the space velocity was 4.5 × 104/hr. The simulated coal gas included 10% H2 (when used), 20% CO (when used), 300 ppm H2S and N2 as a balance gas. The initial concentration of mercury was 40 ± 3 μg/m3. The contents of Hg vapor and H2S in the inlet and outlet gases of the reactor could be monitored using the mercury analyzer and gas chromatograph with a flame photometry detector, respectively. The removal efficiency (η) of Hg or H2S and Hg content (Q) was used to evaluate the performance of the prepared sorbents capturing Hg or H2S from simulated coal gas. Q and η were calculated using the following formulas

1.1. Sample preparation A commercial coal-based AC was selected as the carrier, in the form of pellets with diameters of 0.25–0.38 mm. A series of sorbents were prepared by pore volume impregnation using palladium nitrate as the precursor. After the impregnation, the sorbents were air-dried at room temperature for 10 hr, and then dried at 100°C for 12 hr, and lastly, calcined in N2 at

η ¼ ðn0 −n1 Þ=n0  100%

ð1Þ

Q ¼ m1 =m0

ð2Þ 3

3

where, n0 (μg/m or ppm) and n1 (μg/m or ppm), are the concentrations of Hg or H2S in the feed and effluent gases, respectively, and m1 (mg) and m0 (g) are the quantities of Hg

Furnace Catalyst

N2

H2S LUMEX/RA915M Mercury analyzer Heating tape CO

Computer

GC950 H2 Activated carbon

N2

Absorber

Mass flowmeter Water bath

Mercury permeation tube

Fig. 1 – Schematic diagram of adsorption apparatus for removing Hg and H2S.

1499

JO U R N A L OF EN V I RO N M EN T A L S CI E NC ES 2 6 ( 20 1 4 ) 1 4 9 7–1 5 0 4

captured by sorbents and the quality of fresh sorbents, respectively.

1.3. Sample characterization The BET surface area, pore size and pore volume of sorbents were measured by BET analysis (TriStar 3000, Micromeritics Instrument Corporation, China) using N2 adsorption at −196°C. The samples were initially purged at 150°C for 3 hr before adsorption. XRD was employed to investigate the crystal structures of the sorbents. The instrument was a Rigaku D/max2500 diffractometer, fitted with a nickel-filtered Cu Kα radiation source operating at a voltage of 40 kV and 100 mA. The scan rate was 8°/min at the range from 5 to 85°. XPS surface analysis was conducted to determine the surface concentration and binding energy of samples, using an ESCALAB 250 spectrometer (VG Scientific Ltd., UK) equipped with an Al Kα source (hγ = 1486.6 eV, 150 W). Energy correction was performed using the C1s peak at 284.6 eV. No smoothing routine of data was applied to analyze the results. The crystalline phases of samples were measured by high resolution TEM at room temperature with a JEM-2010 microscope operating at 200 kV.

2. Results and discussion 2.1. Effect of Pd loading amount and dispersion on Hg removal Fig. 2 shows the removal efficiency of Hg and H2S over the Pd/ AC sorbents with different Pd loading amounts. At 250°C, the initial efficiency of Hg removal over 0.5Pd/AC, 1Pd/AC and 3Pd/AC sorbents was about 89%, 93% and 94%, respectively. The efficiency of Hg removal over the 3Pd/AC sorbent remained at around 95% over 150 min. That of 0.5Pd/AC and 1Pd/AC sorbents showed a decrease, reducing to 20% and 65% after 150 min, respectively. It is obvious that the efficiency of

Hg removal increased with the increasing of Pd loading amount on the sorbent. The 3Pd/AC sorbent was selected and the mercury balance was evaluated. The total amount of mercury at the inlet was 3.38 μg during the reaction time in N2–Hg–H2S atmosphere at 250°C (40 μg/m3 Hg, 300 ppm H2S in N2, and SV = 4.5 × 104/hr). The mercury adsorbed and catalytic oxidation by sorbents (calculated by Hg content of sorbents) was 2.42 μg. The amount of mercury at the output was 0.17 μg. Because the mercury analyzer can only detect Hg0, the remaining 0.79 μg mercury should be in the form of Hg2+. Fortunately, Hg2+ can be efficiently removed by WFGD facilities, as it is soluble in most aqueous solutions (Ozaki et al., 2008). The Hg content (Q) of the sorbent and the mercury amount adsorbed per mole of Pd (Hg/Pd) are shown in Table 1. For example, the Hg content of 0.5Pd/AC was 3.85 mg Hg/1000 g sorbent at 250°C, with the ratio of Hg/Pd at 0.49 mmol/mol. Although the Hg content of 3Pd/AC was the highest among the three sorbents, the Hg/Pd ratio of the 3Pd/AC sorbent was much lower than the 1Pd/AC and 0.5Pd/AC sorbents. This indicated that the dispersion of Pd on the surface of 0.5Pd/AC and 1Pd/AC was better and their effective utilization rate of Pd was also higher than that of 3Pd/AC. It is suggested that both the dispersion and the loading amount of Pd contribute to the efficiency of Hg removal over the Pd/AC sorbent. There was good dispersion of Pd over the 0.5Pd/AC sorbent, but the active sites were limited compared to the 1Pd/AC and 3Pd/AC sorbents. On the other hand, the Hg/Pd ratio can partly reflect the utilization efficiency of Pd over the Pd/AC sorbents. Fig. 3 shows the TEM images of the Pd/AC sorbents with different Pd loading amounts. It can be found that the Pd loading amount affected the dispersion of Pd and it became slightly worse as the Pd loading amount increased. Therefore, the 1Pd/AC sorbent was selected for the next experiments, considering the effective utilization rate of Pd, the loading amount and the efficiency of Hg removal. The influence of the Pd loading amount on the H2S removal efficiency at 250°C is also presented in Fig. 2. It was

100

H2S removal efficiency (%)

Hg removal efficiency (%)

100

80

60

40

0.5Pd/AC 1Pd/AC/AC 3Pd/AC/AC

20

0

20

40

60

80

60

40 0.5% Pd/AC 1% Pd/AC 3% Pd/AC

20

80

100

Time (min)

120

140

160

0

30

60

90

120

150

180

210

240

Time (min)

Fig. 2 – Efficiency of Hg (left) and H2S (right) removal over 0.5Pd/AC, 1Pd/AC and 3Pd/AC sorbents in N2–Hg–H2S atmosphere at 250°C. Reaction conditions: 40 μg/m3 Hg, 300 ppm H2S in N2, and SV = 4.5 × 104/hr.

1500

J O U R NA L OF EN V I RO N M EN T A L S CI EN C ES 2 6 ( 20 1 4 ) 1 4 97–1 5 0 4

Table 1 – Hg content of sorbents with different Pd loading amounts. Samples

Temperature (°C)

Hg content (Q) (mg Hg/1000 g sorbent)

Hg/Pd (mmol/mol)

0.5Pd/AC 1Pd/AC 3Pd/AC

250 250 250

3.85 4.29 4.84

0.49 0.23 0.09

AC: activated carbon

found that all of the desulfurization efficiency over the three sorbents can be maintained at above 85% for 4 hr at 250°C. The 3Pd/AC sorbent did not show very good desulfurization activity, which suggests that the Pd loading amount has no obvious influence on the H2S removal efficiency.

2.2. Effect of the temperature on the removal of Hg and H2S over sorbents The effect of the temperature on the removal efficiency of Hg and H2S over 1Pd/AC was evaluated at the temperatures of 250, 300 and 350°C, and the results are shown in Fig. 4. At 250°C, the Hg removal efficiency over the 1Pd/AC sorbents could reach 92% and maintain a relatively high efficiency of 68% within 150 min. When the reaction temperature rose to 350°C, the Hg removal efficiency sharply dropped to 14% and was only 1% after 130 min. The results indicated that high temperatures had a negative effect on the efficiency for Hg removal. Pd (mainly Pd0 from Section 2.4) loaded on AC could form amalgams with mercury (Granite and Pennline, 2002; Granite et al., 2000). The decrease of the solubility of Hg in Pd with the increasing temperature and the de-amalgamation at high temperatures were possible reasons for the decrease of Hg removal efficiency from 250 to 350°C (Poulston et al., 2007). The H2S adsorption curves of the 1Pd/AC sorbent at different temperatures are also presented in Fig. 4. It was found that the change of H2S removal efficiency with temperature was different from that of Hg, increasing slightly from 250 to 350°C.

2.3. Effect of H2 and CO atmosphere on the removal of Hg over sorbent The curves of the Hg removal efficiency over the 1Pd/AC sorbent in different atmospheres at 250°C are shown in Fig. 5. It was found that the efficiency for Hg removal over the 1Pd/ AC sorbent in N2–Hg atmosphere at 250°C could be maintained above 80% for at least 300 min. However, in N2–H2S–Hg

a

50 nm

b

50 nm

atmosphere, the efficiency of Hg removal reached over 90% but only for the first 50 min, then decreased to 68%. Moreover, in N2–H2S–H2–CO–Hg atmosphere, the efficiency of Hg removal over the 1Pd/AC sorbent could be maintained above 92% for at least 360 min. These results indicated that Hg removal efficiency over 1Pd/AC can be clearly improved when CO, H2 and H2S exist simultaneously. H2S is known as a poison for Pd (Castro et al., 2002) in N2–H2S–Hg atmosphere. The reason may be that the PdO (in Section 2.4) on the surface of 1Pd/AC sorbent can be significantly converted to PdS, which is difficult to reduce to element Pd (Pd0 is the main active component of Hg removal). PdO was more easily reduced to elemental Pd in N2–H2S–H2–CO–Hg atmosphere. The efficiency of Hg capture can be promoted in the reducing gas (Yu and Shaw, 1998), due to the formation of the Pd–Hg alloy. This result showed that the Pd/AC sorbents have a high potential as a candidate for capturing Hg in the presence of H2S from coal-derived fuel gas, which contains CO, H2, H2S and other gases.

2.4. Characteristics of physical and chemical properties of sorbents The physical properties of the AC support and the Pd/AC sorbents including the BET surface area, the pore volume and the pore diameters are summarized in Table 2. Generally, the surface characteristics of all sorbents were similar, which means physical characteristics are not the main factor in Hg removal. Fig. 6 shows the XRD patterns of 0.5Pd/AC, 1Pd/AC and 3Pd/ AC sorbents before and after the reaction. Representative diffraction peaks of Pd (2θ = 40.1, 46.2, 67.8, 82.0°) were found from the sorbents (Lineberry et al., 2009). These results showed that the active component of the sorbents was mainly elemental Pd, except for the 0.5Pd/AC sorbent due to the relatively low loading amount and the high dispersion of Pd over the AC support. There were no peaks for palladium oxide observed for any of the sorbents. This may be the result of the

c

50 nm

Fig. 3 – TEM of the Pd/AC sorbents with different Pd loading amount. (a) 0.5Pd/AC; (b) 1Pd/AC; (c) 3Pd/AC sorbents.

1501

JO U R N A L OF EN V I RO N M EN T A L S CI E NC ES 2 6 ( 20 1 4 ) 1 4 9 7–1 5 0 4

1Pd/AC 250°C

1Pd/AC 300°C

1Pd/AC 350°C

H2S removal efficiency (%)

Hg removal efficiency (%)

100 80

60

40

20

0

20

40

60

80

100

120

140

160

80

60

40

20

0

30

60

Time (min)

90

120

150

Time (min)

Fig. 4 – Efficiency of Hg and H2S removal over 1Pd/AC sorbent in N2–Hg–H2S atmosphere at different temperatures. Reaction conditions: 40 μg/m3 Hg, 300 ppm H2S in N2, and SV = 4.5 × 104/hr.

palladium oxide being amorphous, or having low crystallinity (Baltrus et al., 2011). This result differed from that of the Pd/ Al2O3 sorbents in our previous work, whose active component was mainly PdO (Han et al., 2012). The intensity of Pd characteristic peaks of used sorbents became notably smaller compared with the fresh sorbents. In particular, the Pd diffraction peak at 2θ = 46.2° of used 3Pd/AC sorbent disappeared and the intensities of peaks at 2θ = 40.1, 67.8, and 82.0° were clearly smaller than that of the fresh 3Pd/AC sorbent. Similar results can be found from the XRD patterns of the 1Pd/ AC sorbent. It is possible that the consumption of Pd in the reaction resulted in the decrease of the Pd peak intensities. The Pd diffraction peaks at 2θ = 40.1 and 67.8° of the used 1Pd/ AC sorbent were also weakened. Thus, it is suggested that elemental Pd should be the main active component of the Pd/ AC sorbents for the capture of Hg and H2S.

100

Hg removal efficiency (%)

90 80 70 60 50 40

XPS analysis of the fresh and used 1Pd/AC sorbent in N2–Hg and N2–Hg–H2S atmospheres was performed in order to obtain more information about the mechanism of the Hg capture over the Pd/AC sorbents. The Pd 3d XPS patterns are given in Fig. 7a. It can be seen that the binding energies (BEs) 335.3 eV of Pd 3d5/2 and 340.4 eV of Pd 3d3/2 of the fresh 1Pd/AC sorbent correspond to Pd0, which is in agreement with the results of the XRD patterns of the fresh 1Pd/AC sorbent. After capturing Hg in N2–Hg–H2S atmosphere, the BEs 336.9 and 342.1 eV of Pd 3d3/2 of the used 1Pd/AC sorbent were assigned to Pd2+ species. There were also Pd0 characteristic peaks, although the intensity was notably lower compared with Pd2+. After capturing Hg in N2–Hg atmosphere, there was more Pd0 than Pd2+ species on the surface of the sorbents. This indicated that there are different mechanisms of the Hg removal over the Pd/ AC sorbent in the two kinds of atmospheres. The Hg4f XPS patterns are shown in Fig. 7b. In N2–Hg–H2S atmosphere, the BE 101.6 eV of Hg4f7/2 of the 1Pd/AC sorbent was assigned to Hg0, and 103.4 eV of Hg4f 7/2 was assigned to Hg2+. In N2–Hg atmosphere, it can be found that Hg0 and Hg2+ also occurred. Moreover, Hg0 was the main component on the surface of the used sorbents after the capturing of Hg in N2–Hg atmosphere, while Hg2+ was the main component on the surface of the used sorbents in N2–Hg–H2S atmosphere. These results indicated that there are obviously different reaction

30

N2+Hg N2+H2S+Hg N2+H2S+H2+CO+Hg

20 10

Table 2 – Physical properties of the AC support and the Pd/ AC sorbents.

0 0

50

100

150

200

250

300

Time (min) Fig. 5 – Efficiency of Hg removal over 1Pd/AC sorbent in different atmospheres at 250°C. Reaction conditions: 40 μg/m3 Hg, 300 ppm H2S, 10% H2, 20% CO in N2, and SV = 4.5 × 104/hr.

Samples

BET surface area (m2/g)

Pore volume (cm3/g)

Average pore diameter (nm)

AC 0.5Pd/AC 1Pd/AC 3Pd/AC

915.56 879.43 926.48 899.50

0.51 0.49 0.52 0.51

2.39 2.29 2.44 2.47

350

AC: activated carbon

1502

J O U R NA L OF EN V I RO N M EN T A L S CI EN C ES 2 6 ( 20 1 4 ) 1 4 97–1 5 0 4

C

Pd

Pd

Relative intensity (a.u.)

Pd

c1 c b1 b a1 a 20

30

40

50

60

70

80

90

2θ (°) Fig. 6 – XRD spectra of sorbents before and after the reaction. Line a, b, c: fresh 0.5 Pd/AC, 1Pd/AC and 3 Pd/AC sorbents; line a1, b1, c1: the sorbents after reaction in N2–Hg–H2S atmosphere at 250°C.

336.9

342.1 Pd 3d

340.4

335.3

in N2-Hg-H2S

101.9

b

103.4

Hg 4f

in N2-Hg-H2S

Intensity (a.u.)

Intensity (a.u.)

a

mechanisms of Hg removal with and without H2S in the reaction atmosphere. The C1s XPS patterns are shown in Fig. 7c. The C1s BEs at 284.8, 286.3, 287.5, 288.87 eV of the fresh 1Pd/AC sorbent corresponded to C\C, C\O, C_O, COO, respectively (Xu et al., 2013). It is noted that the peak intensities of these oxygencontaining groups become weak after the capture of Hg in N2– Hg–H2S atmosphere, in particular, the peaks of C_O and COO almost disappear. It is possible that these oxygen-containing groups take part in the reaction to oxidize Pd0 into Pd2+ (PdO). The S2p XPS pattern is shown in Fig. 7d. The BEs 163.36 eV and 162.02 eV of S2p of the used 1Pd/AC sorbent corresponded to the formation of HgS and PdS in N2–Hg–H2S atmosphere (Baltrus et al., 2008; Zylberajch-Antoine et al., 1991), respectively. Therefore, it can be deduced that there are two ways of capturing Hg over the Pd/AC sorbents in N2–Hg–H2S atmosphere. On the one hand, a part of Hg can react with Pd0 on the surface of sorbents to produce Hg–Pd amalgam (Poulston, et al., 2007). On the other hand, the oxygen-containing groups of the AC can oxidize Pd0 to Pd2+ (PdO), and the PdO may react with H2S and Hg to form PdS and Hg2+ (HgS). It can also be deduced from the above XPS analysis that H2S competes with

in N2-Hg

in N2-Hg fresh 330

332

334

336

338

340

342

344

98

100

102

Binding energy (eV) 284.8

c

104

106

Binding energy (eV)

d

C 1s

S 2p

286.3

Intensity (a.u.)

Intensity (a.u.)

163.36 in N2-Hg-H2S

287.5 fresh

282

284

286

288

Binding energy (eV)

162.02

in N2-Hg-H2S

288.87

290

167

166

165

164

163

162

161

Binding energy (eV)

Fig. 7 – XPS spectra of the fresh and used 1Pd/AC sorbents over the spectral regions of Pd3d (a), Hg4f (b), C1s (c) and S2p (d).

JO U R N A L OF EN V I RO N M EN T A L S CI E NC ES 2 6 ( 20 1 4 ) 1 4 9 7–1 5 0 4

Pd

C=O COO

Pd Pd

Hg

S H 2 PdS PdO

Hg

Hg2+

Pd-Hg

Fig. 8 – Diagram of the reaction approach of capturing Hg over Pd/AC sorbent in N2–Hg–H2S atmosphere.

Hg on Pd active sites of sorbents in N2–Hg–H2S atmosphere. H2S can react with PdO to generate PdS, which is difficult to be reduced to elemental Pd in N2–Hg–H2S atmosphere. Therefore, H2S has a negative effect on the removal of Hg. A diagram is given to clearly display the reaction approach of capturing Hg over the Pd/AC sorbents in N2–Hg–H2S atmosphere (Fig. 8).

3. Conclusions The performance of Hg and H2S capture over Pd/AC sorbents and the effects of the Pd loading amount, reaction temperature and reaction atmosphere were studied. The Pd/AC sorbents showed an excellent adsorption activity for the Hg and H2S removal from coal-derived fuel gas. The efficiency of Hg removal increased with the increasing of Pd loading amount on the sorbents, but the effective utilization of Pd gradually decreased at the same time. High temperature was not benefit of the Hg removal, but it was conducive to desulfurization. The addition of CO and H2 in atmospheres can promote the removal efficiency of Hg removal, and prevent H2S poisoning of the sorbents due to the formation of PdS. Pd0 was the main active component of the Pd/AC sorbents for the removal of Hg. It can be deduced from the XPS results that there are two ways of capturing Hg over the sorbents in N2–Hg–H2S atmosphere. One way is a part of Hg reacting with Pd0 on the surface of the sorbents to produce Hg–Pd amalgam. Another way is via Pd0 oxidized by the containing-oxygen groups on the AC to form Pd2+ (PdO), which reacts with H2S and Hg to form PdS and Hg2+ (HgS).

Acknowledgments This work was supported by the National Natural Science Foundation of China (Nos. 21006067, 21276170), the Shanxi Province Natural Science Foundation (Nos. 2010021008-1, 201101008-4) and the National High-Tech Research and Development Program (863) of China (No. 2013AA065404E).

REFERENCES

Baltrus, J.P., Granite, E.J., Stanko, D.C., Pennline, H.W., 2008. Surface characterization of Pd/Al2O3 sorbents for mercury capture from fuel gas. Main Group Chem. 7 (3), 217–225. Baltrus, J.P., Granite, E.J., Rupp, E.C., Stanko, D.C., Howard, B., Pennline, H.W., 2011. Effect of palladium dispersion on the

1503

capture of toxic components from fuel gas by palladium–alumina sorbents. Fuel 90 (5), 1992–1998. Bandosz, T.J., 2002. On the adsorption/oxidation of hydrogen sulfide on activated carbons at ambient temperatures. J. Colloid Interface Sci. 246 (1), 1–20. Castro, F., Meyer, G., Zampieri, G., 2002. Effects of sulfur poisoning on hydrogen desorption from palladium. J. Alloys Compd. 330–332, 612–616. Feng, W.B., Borguet, E., Vidic, R.D., 2006. Sulfurization of a carbon surface for vapor phase mercury removal-II: sulfur forms and mercury uptake. Carbon 44 (14), 2998–3004. Granite, E.J., Pennline, H.W., 2002. Photochemical removal of mercury from flue gas. Ind. Eng. Chem. Res. 41 (22), 5470–5476. Granite, E.J., Pennline, H.W., Hargis, R.A., 2000. Novel sorbents for mercury removal from flue gas. Ind. Eng. Chem. Res. 39 (4), 1020–1029. Granite, E.J., Myers, C.R., King, W.P., Stanko, D.C., Pennline, H.W., 2006. Sorbents for mercury capture from fuel gas with application to gasification systems. Ind. Eng. Chem. Res. 45 (13), 4844–4848. Han, L.N., Lü, X.Y., Wang, J.C., Chang, L.P., 2012. Palladium–iron bimetal sorbents for simultaneous capture of hydrogen sulfide and mercury from simulated syngas. Energy Fuels 26 (3), 1638–1644. Hu, C., Zhou, J., Luo, Z., Cen, K., 2010. Oxidative adsorption of elemental mercury by activated carbon in simulated coal-fired flue gas. Energy Fuels 25 (1), 154–158. Kamakoti, P., Morreale, B.D., Ciocco, M.V., Howard, B.H., Killmeyer, R.P., Cugini, A.V., et al., 2005. Prediction of hydrogen flux through sulfur-tolerant binary alloy membranes. Science 307 (5709), 569–573. Lineberry, Q.J., Cao, Y., Lin, Y., Ghose, S., Connell, J.W., Pan, W.P., 2009. Mercury capture from flue gas using palladium nanoparticle-decorated substrates as injected sorbent. Energy Fuels 23 (3), 1512–1517. Liu, W., Vidić, R.D., Brown, T.D., 1998. Optimization of sulfur impregnation protocol for fixed-bed application of activated carbon-based sorbents for gas-phase mercury removal. Environ. Sci. Technol. 32 (4), 531–538. Murakami, A., Uddin, M.A., Ochiai, R., Sasaoka, E., Wu, S.J., 2010. Study of the mercury sorption mechanism on activated carbon in coal combustion flue gas by the temperature-programmed decomposition desorption technique. Energy Fuels 24 (8), 4241–4249. Otero-Rey, J.R., López-Vilariño, J.M., Moreda-Piñeiro, J., AlonsoRodríguez, E., Muniategui-Lorenzo, S., López-Mahía, P., et al., 2003. As, Hg, and Se flue gas sampling in a coal-fired power plant and their fate during coal combustion. Environ. Sci. Technol. 37 (22), 5262–5267. Ozaki, M., Uddin, M.A., Sasaoka, E., Wu, S.J., 2008. Temperature programmed decomposition desorption of the mercury species over spent iron-based sorbents for mercury removal from coal derived fuel gas. Fuel 87 (17–18), 3610–3615. Pavlish, J.H., Hamre, L.L., Zhuang, Y., 2010. Mercury control technologies for coal combustion and gasification systems. Fuel 89 (4), 838–847. Poulston, S., Granite, E.J., Pennline, H.W., Myers, C.R., Stanko, D.P., Hamilton, H., et al., 2007. Metal sorbents for high temperature mercury capture from fuel gas. Fuel 86 (14), 2201–2203. Presto, A.A., Granite, E.J., 2006. Survey of catalysts for oxidation of mercury in flue gas. Environ. Sci. Technol. 40 (18), 5601–5609. Shen, Z.M., Ma, J., Mei, Z.J., Zhang, J.D., 2010. Metal chlorides loaded on activated carbon to capture elemental mercury. J. Environ. Sci. 22 (11), 1814–1819. Sun, C.G., Snape, C.E., Liu, H., 2013. Development of low-cost functional adsorbents for control of mercury (Hg) emissions from coal combustion. Energy Fuels 27 (7), 3875–3882.

1504

J O U R NA L OF EN V I RO N M EN T A L S CI EN C ES 2 6 ( 20 1 4 ) 1 4 97–1 5 0 4

Xu, W.Q., Wang, H.R., Zhu, T.Y., Kuang, J.Y., Jing, P.F., 2013. Mercury removal from coal combustion flue gas by modified fly ash. J. Environ. Sci. 25 (2), 393–398. Yan, R., Liang, D.T., Tsen, L., Wong, Y.P., Lee, Y.K., 2004. Bench-scale experimental evaluation of carbon performance on mercury vapour adsorption. Fuel 83 (17), 2401–2409. Yan, N.Q., Liu, S.H., Chang, S.G., 2005. Method for the study of gaseous oxidants for the oxidation of mercury gas. Ind. Eng. Chem. Res. 44 (15), 5567–5574.

Yan, N.Q., Chen, W.M., Chen, J., Qu, Z., Guo, Y.F., Yang, S.J., et al., 2011. Significance of RuO2 modified SCR catalyst for elemental mercury oxidation in coal-fired flue gas. Environ. Sci. Technol. 45 (13), 5725–5730. Yu, T.C., Shaw, H., 1998. The effect of sulfur poisoning on methane oxidation over palladium supported on γ-alumina catalysts. Appl. Catal., B 18 (1–2), 105–114. Zylberajch-Antoine, C., Barraud, A., Roulet, H., Dufour, G., 1991. XPS characterization of inserted mercury sulfide single layers in a Langmuir-Blodgett matrix. Appl. Surf. Sci. 52 (4), 323–327.