Article pubs.acs.org/EF
Impact of Sulfur Loading on Brominated Biomass Ash on Mercury Capture Teresa M. Bisson,† Zong Qian Ong,† Aimee MacLennan,‡ Yongfeng Hu,*,‡ and Zhenghe Xu*,†,§ †
Department of Chemical and Materials Engineering, University of Alberta, Edmonton, Alberta, Canada Canadian Light Source Inc., University of Saskatchewan, 44 Innovation Boulevard, Saskatoon, Saskatchewan, Canada § Institute of Nuclear and New Energy Technology, Tsinghua University, Beijing, P.R. China 1000084 ‡
S Supporting Information *
ABSTRACT: Brominated carbon injection is a commonly applied technology for capture of mercury in flue gases emitted from coal-fired power plants. However, a previous study has shown leaching potential of mercury (Hg) and bromine (Br) from brominated carbon sorbents, with recommendations to reduce Br concentration in the sorbents. To reduce the Br content required on the sorbent, a strategy of coimpregnation of sulfur and Br was employed in this study. Sulfur (S) loading was found to increase the mercury uptake by the sorbent, with the maximum uptake occurring at a S:Sorbent mass ratio of 1:20. Co-loading of a low concentration of bromine (0.7−0.9%) to the sulfur-loaded sample further improved mercury capture by the sorbent and appeared to have a synergistic effect on Hg removal. Leaching tests confirmed that combining both Br and S onto the sorbent reduced the leaching potential of Br and Hg from the spent sorbent material. In addition, analysis of Hg X-ray absorption spectra (XAS) revealed a mixed Hg−Br, Hg−S, and Hg−C binding environment on the sorbent coimpregnated with bromine and sulfur.
■
materials.31,32 AC sorbents were prepared by activating soybean straw, rice straw, and corn stalks with zinc chloride.31 The optimum temperature and zinc chloride loading were found to be dependent on the type of biomass used, leading to 95% Hg0 capture under optimal conditions.31 Olive seed biomass, pine wood, oak wood, and waste tires were used as carbon sources by Skodras et al.32 The Hg0 removal mechanism by these materials seemed to depend on oxygen-containing surface functional groups. While these waste materials were considered to be low cost, the KOH required for activation should be considered if comparing the cost effectiveness to other AC sorbents.32 Recently, a novel carbon-based sorbent (Br-Ash) has been developed by brominating a carbon-rich waste byproduct of biomass combustion (wood ash) in order to reduce sorbent costs.33 In contrast to the other biomass AC sorbents, no chemical activation was required as the ash was directly impregnated with Br using a novel chemical−mechanical bromination method. The mercury binding of the Br-Ash sorbent was studied by X-ray absorption near edge spectroscopy (XANES) and extended X-ray absorption fine structure (EXAFS) in comparison to a commercial brominated AC.3 The mechanism of mercury binding to Br-Ash was proposed to be surface enhanced oxidation by the Br, followed by binding of the oxidized mercury on the carbon surface. In the case of the commercial brominated activated carbon, the oxidized mercury was found to be coordinated to the sulfide species present on the surface of the commercial sorbent. This finding suggested the possibility of a synergetic effect by combining both sulfur and bromine onto the carbon surface for mercury removal.
INTRODUCTION Mercury is one of the toxic air pollutants emitted from coalfired power plants. Over the past several years, much work has been done to develop more efficient and cost-effective methods to remove mercury from coal-fired power plant flue gases. A very promising technology, which has been proven on the lab scale, pilot scale, and full scale, involves the injection of powdered activated carbon (AC).1 The powdered AC is often impregnated with additional reactive chemical species, such as halogens or sulfur. These species improve the mercury capture efficiency of AC, particularly from flue gases of power plants burning coals with low chlorine content. To synthesize more effective sorbents, several studies on understanding the mercury removal mechanism by various AC sorbents have been performed.1−17 These studies showed that several factors influence the adsorption of Hg0 onto AC, including contact time, flue gas temperature and composition,18 unburned carbon content in the fly ash,19 and carbon type (which influences amount of mercury active sites).19 For sulfur impregnated sorbents, previous studies indicated that mercury uptake is influenced by sulfur content20−23 and speciation,22,24−26 impregnation temperature,21,22 and sorbent specific surface area23−25,27,28 and pore characteristics.20,24,25 Most mechanistic studies showed that Hg is present on the surface of the AC in its oxidized form.2−9,14,16,17 Mechanisms of mercury removal by chemically impregnated sorbents have therefore been generally proposed to include some form of oxidation of the Hg0, followed by binding to available surface species on the AC.3−6 However, Huggins et al.6 also suggested that a different Hg0 capture mechanism could be present for each type of AC sorbent. Several studies aimed to reduce the cost of AC injection by using less expensive source materials such as biochar,29 agricultural residue,24 petroleum coke,30 and various biomass © 2015 American Chemical Society
Received: June 1, 2015 Revised: October 22, 2015 Published: October 22, 2015 8110
DOI: 10.1021/acs.energyfuels.5b01213 Energy Fuels 2015, 29, 8110−8117
Article
Energy & Fuels Table 1. XRF Analysis of Raw Biomass Ash34a
Adding similar sulfur species to the Br-Ash sorbent was hypothesized to provide another site for mercury binding after oxidation of elemental mercury by the Br. Furthermore, the addition of sulfur could reduce the amount of bromine required for sorbent synthesis while maintaining adequate mercury removal capacity. Reducing bromine content on brominated sorbent is important, as our previous tests showed a significant leaching of bromine from brominated carbon sorbents, in particular at low liquid to solid mass ratios.34 The high concentration of bromine in the leachate was also found to increase mercury leaching from the spent sorbent material. Similar studies were reported on mercury leaching from fly ash after coal combustion,8,35−37 activated carbons,35,38 and brominated activated carbons.39 These studies concluded that the mercury was stable (except under highly acidic conditions), but the tests did not include analysis for bromine leaching and were not conducted at a low liquid to solid mass ratio condition. Decreasing the bromine content of the sorbents is therefore desirable in order to reduce the potential environmental impact of spent sorbent disposal. On the basis of the previous studies, combining both Br and S species on the biomass ash sorbent could reduce the amount of Br required. A possible additional benefit of sulfur impregnation is its low cost. On the basis of publicly available commodity market pricing information, the cost of elemental sulfur could be as much as 14× lower than bromine.40 However, it is uncertain if this cost advantage would offset the cost of impregnating two species. Many previous studies investigated the effect of sulfur addition to activated carbon on mercury capture. The main methods of sulfur addition include adsorption of H2S,20,41−45 vapor impregnation by heating the carbon sorbent with elemental sulfur (S0),21,22,27,43,44,46−51 or liquid impregnation using other sulfur species with subsequent thermal treatment to remove the moisture.23,25,27,28 In addition, experiments involving the injection of sulfur chlorides (SCl2 and S2Cl2)52 and sulfur monobromide (S2Br2)53 into flue gases for Hg0 removal have also been investigated. Speciation of sulfur on Simpregnated sorbents has been found to be a key parameter for mercury removal,26,44,46 with sulfide and elemental sulfur being effective species.3,44 For this reason, the physical impregnation of S0 onto the biomass ash was chosen for this study at the known maximum sulfur loading temperature of 600 °C.54 At this temperature, approximately 16% of the sulfur is in the form of the more reactive S2 allotrope.22 To our best knowledge, no previous studies have investigated coimpregnating both Br and S onto carbon-based sorbents synthesized from waste materials to maintain high mercury capture while reducing Br leaching and lowering the cost. This paper is to fill this gap by combining both S and Br on a high carbon content biomass ash (waste material), testing the sorbent effectiveness for mercury removal, investigating the role of these species on mercury capture and determining the leachability of both mercury and bromine from the new sorbent.
■
element
raw ash (wt %)
Al Si P S K Ca Ti V Cr Mn Fe Cu Zn Br Sr Zr
1.61 8.38 1.76 2.94 9.99 55.88 1.32 0.09 0.10 3.34 13.22 0.10 0.85 0.08 0.22 0.15
a
Note: light elements such as carbon, hydrogen, nitrogen, and oxygen are not detected in XRF analysis. As a result, the XRF composition reported in this table represents a relative concentration among the detected elements.
(S0) (99% purity from Sigma-Aldrich) was weighed into a ceramic crucible followed by Raw Ash, which was placed on the top of the sulfur. S-Ash (sulfur loaded Raw Ash) was prepared using a 20:1 Raw Ash to Sulfur mass ratio (i.e., 2 g Raw Ash with 100 mg S0). In order to determine the optimal sulfur concentration on the wood ash for mercury removal, two other samples at a 100:1 Raw Ash to Sulfur mass ratio (5 g Raw Ash and 50 mg S0) and 4:1 mass ratio (750 mg Raw Ash and 187.6 mg Sulfur) were also prepared and tested. The crucible containing ash and sulfur was placed in a custom made holder and inserted into a tube furnace (see schematic in Figure S1 in Supporting Information). Nitrogen was used to purge the tube furnace throughout the experiment at a flow rate of 500 mL/min. After 15 min, the air was purged out of the system, and the tube furnace was heated from room temperature to 100°C, followed by holding the sample at this temperature for 1 h in order to remove any moisture from the sample. The furnace was then heated at a rate of 10°C/min to 200 °C (just past the melting point of sulfur) and held at this temperature for 1 h to allow liquid sulfur to mix with the Raw Ash. At this time, the furnace was further heated at a rate of 5 °C/min to 600 °C and held at this temperature for 2 h before cooling the sample to room temperature. The sulfur-loaded sample was removed and mixed with mortar and pestle for 5 min to obtain a homogeneous powder. In this study, lower bromine loading was used to reduce the possible environmental issues with Br release from spent sorbent in landfill. Similar bromination procedures described by Bisson et al.33 were used. In summary, samples of low bromine content, 2D Br-Ash and 2D SAsh sorbents (where 2D represents two drops of Br) were prepared by placing 9.0 g of Raw Ash (for 2D Br-Ash) or S-Ash (for 2D S-Ash) on top of 63 g of 6 mm glass beads contained in a 250 mL gastight FEP jar. After addition of two drops (approximately 50 mg) of liquid Br2 on top of the ash, the jars were quickly capped and sealed. The jars were then packed tightly into a 10 L carboy and rolled for 30 min on a set of mechanical rollers. After completing the chemical−mechanical bromination, the mixture was poured onto a tray inside a fume hood and rested for 30 min. The mixture was then transferred into a vacuum oven at 200 °C for 30 min to remove any loosely bound bromine. Two sets of each brominated sorbent sample were prepared to ensure the repeatability of synthesis. Sorbent Characterization. Sorbents were characterized by elemental analysis, BET and pore size distribution measurements, particle size distribution determination, and scanning electron microscopy (SEM) and transmission electron microscopy (TEM) imaging as described in Supporting Information.
EXPERIMENTAL PROCEDURES
Synthesis of S-Ash and Br-Ash. Biomass ash was obtained from a power station burning waste wood chips. The as-received ash was ball milled using the same procedure reported by Bisson et al.33 The ground ash is called “Raw Ash” with composition given in Tables 1 and 2. Particle size distribution of the ash is shown in Table S1 in Supporting Information. The appropriate amount of elemental sulfur 8111
DOI: 10.1021/acs.energyfuels.5b01213 Energy Fuels 2015, 29, 8110−8117
Article
Energy & Fuels Table 2. Elemental Analysis Results and BET Data sample Raw Ash Low S-Ash S-Ash High S-Ash 2D Br-S-Ash 2D Br-Ash
C (wt %) 37.9 44.8 34.1 29.0 31.7 35.6
± ± ± ± ± ±
4.9 7.5 3.1 1.9 1.0 0.3
H (wt %) 0.5 0.4 0.3 0.2 0.4 0.4
± ± ± ± ± ±
0.1 0.1 0.1 0.1 0.1 0.1
N (wt %) 0.1 0.1 0.1 0.1 0.1 0.1
± ± ± ± ± ±
Mercury Pulse Injection Tests. Mercury pulse injection tests are a convenient way to determine if a sorbent is capable of removing mercury from gases. The test is designed to simulate the short contact time between the powdered sorbents and mercury in flue gases during powdered sorbent injection. The pulse injection test is similar in setup to a fixed bed test, but differs in the duration of mercury exposure, reporting only the initial capture of mercury and not sorbent capacity. A detailed description of the pulse injection test can be found elsewhere.33 Briefly, 40 mg of sample was loaded into a borosilicate utube, which was heated to the desired temperature using a GC oven. A mercury standard consisting of a pure Hg0 ball in equilibrium with air was kept in a tightly sealed glass vessel with a syringe port and was kept at room temperature. An accurately measured sample of 200 μL of Hg0 in air was taken and injected into 40 mL/min argon flow, upstream of the u-tube. The sorbent adsorbed the Hg0, and any portion that remained in the gas stream was preconcentrated on a gold trap downstream of the sorbent. After 5 min of collection, the gold trap was quickly heated to release the mercury which was detected by a Tekran 2500 CVAFS (cold vapor atomic fluorescence spectrophotometer). The mercury that was not adsorbed or captured by the sample is divided by the amount of mercury injected, which is defined as “Hg0 Breakthrough” in this paper. Each sample was repeated three times to ensure repeatability, with the average and standard deviation being reported. Hg Loading for XAS Analysis. To investigate the binding mechanism of mercury on the sorbent by XAS, a sufficient amount of Hg0 was loaded onto the sorbent samples in air at 100 °C. Although typical application temperatures for AC injection are in the range of 121−221 °C,18 our previous study over the temperature range of 100 and 200 °C showed no detectable difference in Hg binding for the BrAsh samples.3 For this reason, the mercury adsorption temperature for determining mercury binding environment in this study was set at 100 °C. A schematic for mercury loading experiments is shown in Figure S2 in Supporting Information. 75 mg of sample was weighed into a borosilicate u-tube containing quartz wool to keep the sorbent in place. The tube was then placed into a GC oven (Figure S2) and heated to 100 °C. After the tube was heated for 5 min, 100 mL/min of air was introduced, flowing over the top of a ball of mercury (Figure S2) and contacting the sorbent. The setup was left for 24 h to achieve high mercury loading on the sample. After 24 h the air flow was stopped, and the sample cooled down to room temperature. The sample was then removed from the u-tube and mixed for 5 min with mortar and pestle to ensure sample homogeneity. The mercury content of the sample was measured using a Milestone DMA-80 mercury analyzer. The mercury content in the sample was found to be too high for direct measurement by the DMA-80. To reduce mercury concentration for analysis, the sample was diluted by mixing 10 mg of the mercury loaded sorbent with 600 mg of raw activated carbon for 15 min using a mortar and pestle. The mercury content of the raw activated carbon was also determined using the DMA-80 in order to subtract from the diluted mixture the contribution of mercury from the raw carbon. To ensure the validity of the method, a similar dilution procedure was applied to a standard sample (NIST 2451) containing a high mercury loading (certified value of 688 mg/kg ± 28 mg/kg). The measured value of the standard was always within 10% of the reported value, within the known accuracy of the DMA-80 instrument. To further ensure the reliability of the data, two mixtures were made for each sample (and repeat sample), and mercury concentration was
0.1 0.1 0.1 0.1 0.1 0.1
S (wt %)
Br (wt %)
BET (m2/g)
± ± ± ± ± ±
n/a n/a n/a n/a 0.9 ± 0.1 0.7 ± 0.1
210 185 130 147 119 178
0.4 1.7 4.1 3.8 3.8 0.4
0.1 0.2 0.5 0.6 0.2 0.1
measured three times for each mixture. The results reported are the average and standard deviation of these measurements. Leaching Tests. Leaching tests were conducted to compare the amount of Br and Hg leached from the sorbents with the results reported in our previous study.34 Mercury was loaded onto the sorbents by spreading the sorbents in the bottom of a sealed glass vessel, and inserting a mercury source to saturate the air in the vessel for 1 week. The mercury content in the sample was measured by a DMA-80 mercury analyzer (Milestone, Inc.). The first set of leaching tests was based on slight modification of the standard method (called toxicity characteristic leaching procedure or TCLP) recommended by the EPA to determine if a waste is hazardous. This method involved adding 20 mL of 5.7 mL/L glacial acetic acid aqueous solution to 1 g mercury-loaded sample in a 250 mL container tightly sealed with the cap. Additional tests were conducted at a low liquid to solid ratio of 2. In this case, 2 mL of ultrapure water was added to 1 g of mercuryloaded sorbent in the 250 mL container. The mixtures were placed on a shaker at room temperature for 18 ± 2 h and then filtered. The filtrate was acidified and diluted to 100 mL for subsequent analysis. The total Hg content of the samples was determined by a PSA Millennium Merlin mercury analyzer, as well as the Milestone DMA80 mercury analyzer. Bromide content was determined at the Biogeochemical Analytical Service Laboratory at the University of Alberta using ion chromatography (Dionex 600 with 9 mM Na2CO3 Eluent and Dionex IonPac AS9-HC column). XAS Analysis. Samples with and without Hg0 loading were analyzed by XAS at Canadian Light Source (CLS). S K-edge XANES measurements were performed at the SXRMB beamline. Both surface sensitive total electron yield (TEY) and bulk sensitive fluorescence yield (FY) were recorded. It should be noted that fluorescence measurement for Hg L-edge is necessary as the Hg concentration is generally low. Much longer analysis time is needed for other analytical techniques to have sufficient sensitivity for the detection. Some surface oxidation (in the form of sulfate peak) could be observed in the TEY spectra; thus FY data of all testing samples were used for data fitting. Selected sulfur model compounds, including elemental sulfur, sulfides, calcium sulfide, sulfoxide, thiophene, and sulfate, were also analyzed as reference. TEY data of model compounds were used for spectral fitting, as the FY spectrum of concentrated sample is known to suffer the self-absorption problem. Hg LIII-edge XANES and EXAFS were acquired using the HXMA beamline at the CLS. A 32 element Ge detector was used for fluorescence measurements, similar to the one used in our previous study.3 IFEFFIT55,56 software was used for all XAS data analysis. Further details regarding testing procedure and data analysis can be found in Supporting Information.
■
RESULTS AND DISCUSSION Elemental Analysis of Sorbents. Elemental analysis results for all the samples are shown in Table 2. The sulfur concentration for Low S-Ash is 1.7%, while for the S-Ash sample it is 4.1%, which is similar to the optimum sulfur loading found by others.20 It is interesting to note that the High S-Ash had twice the amount of elemental sulfur during impregnation compared to S-Ash, but resulted in similar final sulfur content. This result indicates a diminishing return of adding more elemental sulfur during the sample preparation. Table 2 also shows a low Br concentration on the two brominated sorbents 8112
DOI: 10.1021/acs.energyfuels.5b01213 Energy Fuels 2015, 29, 8110−8117
Article
Energy & Fuels
200 °C (and below), the breakthrough of mercury is very low ( Low S-Ash > High S-Ash > S-Ash. Pore size distribution over the mesopore region given in Figure S3 of Supporting Information follows a similar trend as the results of surface area measurements over the particle size range of 2−5 nm. The lower BET surface area of sulfur-loaded samples appears to arise from the blockage of the pores20,42 and/or sintering,57 which would make the pores inaccessible. The slightly higher specific surface area of High S-Ash than that of the S-Ash is possibly due to the higher sulfur concentration during impregnation of High S-Ash. The high sulfur concentration may have partially damaged the physical structure of the ash, creating a larger surface area even for similar sulfur content. Yao et al.25 suggested that sulfur vapor could act as an activating gas that etches the outer carbon surface, causing an increase in pore volume and specific surface area. Hg Capture Results. It is important to note that due to the short residence time available in some particulate control devices, sorbent reactivity has been thought to be more important than capacity in these cases.58 Therefore, it is critical to consider the results of the pulse injection tests which have a short contact time between the mercury and sorbent material. When considering these results, it is important to recognize that real flue gases contain several components such as SO2, NO, HCl, and H2S, which can impact mercury uptake and reaction chemistry. The current study contains an idealized system which does not include all constituents anticipated to be present in flue gases. The results from the Hg0 pulse injection tests for the sulfur-impregnated samples are shown in Figure 1.
Figure 1. Mercury breakthrough of raw34 and sulfur-loaded ash samples.
An effective sorbent for Hg0 capture by powdered activated carbon sorbents should have a low mercury breakthrough (i.e., high Hg0 capture). The untreated Raw Ash captured mercury at low temperatures, but performed poorly above 100 °C, exhibiting the worst mercury capture among all the samples tested. Above 200 °C, all the sorbents in Figure 1 show high Hg0 breakthrough, indicating that they are not effective sorbents for Hg0 capture at these temperatures. However, at 8113
DOI: 10.1021/acs.energyfuels.5b01213 Energy Fuels 2015, 29, 8110−8117
Article
Energy & Fuels
Figure 2. TEM images and EDS mapping of (a) S-Ash and (b) High S-Ash.
Table 4. Br Leaching Test Results using TCLP Leaching Method
mercury inlet concentration, which is different from the testing conditions of the current study.) It can also be seen in Figure S5 that the samples loaded with both bromine and sulfur (2D Br-S-Ash) outperformed the brominated (2D Br-Ash) or sulfurimpregnated (S-Ash) samples, even though 2D Br-S-Ash has a lower specific surface area than the other two sorbents (Table 2). The improved mercury uptake demonstrates the benefit of combining both Br and S on the sorbent. Leaching Tests. Leaching tests were conducted for the SAsh, 2D Br-Ash, and 2D Br-S-Ash sorbents, with the results for the mercury and bromine leaching tests given in Tables 3 and 4, Table 3. Hg Leaching Test Results using TCLP Leaching Method sample
average (mg/L) 0.01 0.18 0.02 0.01
± ± ± ±
0.00 0.13 0.01 0.00
average (%) 0.05 0.7 0.1 0.05
± ± ± ±
average (mg/L)
average (%)
145 ± 10.0 209 ± 15.6 2630 ± 37
42.1 ± 2.8 46.9 ± 3.6 70.9 ± 3.6
the other samples the 2D Br-Ash sample has a higher mercury concentration in the leachate, indicating that for low Br concentrations, addition of sulfur onto the ash appears to stabilize the mercury. It should be noted that the mercury content of the sorbents in the current study is much greater than the anticipated Hg concentration of sorbent/fly ash mixtures in industrial operations. The higher concentrations in this study represent a worst case scenario, and the leaching of Hg from the sorbent/fly ash mixture in an operating plant would be expected to be much lower than the value shown in Table 3. The Br leaching test results (Table 4) show that the bromine concentration in the leachate (as bromide) is much lower for samples 2D Br-Ash and 2D Br-S-Ash than for the BrAsh. The lower leachate bromine concentration is likely due to a much lower bromine concentration on the 2D Br-Ash and 2D Br-S-Ash samples. Compared to Br-Ash, the 2D Br-S-Ash and 2D Br-Ash also show a lower Br % leached, indicating that the bromine present at low concentrations is more stable. As the TCLP test has a high liquid:solid ratio (20), a test was also conducted at a low liquid:solid ratio (2) for the 2D Br-S-Ash with the results shown in Figure S8 in Supporting Information. The mercury concentrations in the leachate of 2D Br-S-Ash was higher than that in the TCLP test, but was still below the 0.2 mg/L limit, indicating that the lower amount of liquid does not severely impact Hg leaching. The amount of bromine and mercury leached was also significantly lower than that from the Br-Ash (shown in Figure S8). Overall, the amount of Br and Hg leached from the 2D Br-S-Ash was lower than that from the BrAsh sorbent, even under the most extreme condition found in our previous study to cause Hg and Br to leach from Br-Ash. In addition, the mercury concentration in the leachate of 2D Br-SAsh was lower than the 0.2 mg/L limit set by the EPA.59 Sulfur K-Edge XANES. The speciation of sulfur is known to impact removal of Hg0 by activated carbon.3,20,22,44,46 Sulfur Kedge spectra were obtained for all the samples tested in this study to better understand Hg0 capture mechanisms and improve future sorbent design. The sulfur K-edge spectra of the
Figure 3. Mercury breakthrough of ash samples containing sulfur and low Br.
Br-S-Ash 2D Br-Ash 2D Br-S-Ash Br-Ash34
sample 2D Br-Ash 2D Br-S-Ash Br-Ash34
0.05 0.7 0.1 0.01
respectively. In addition to the new samples, the results for the Br-Ash34 are included for comparison. It is important to note that even though much of the Br leached from Br-Ash, 1.9 wt % still remained, which is theoretically sufficient to capture the mercury on the sorbent.34 All samples in Table 3 show a mercury content below the 0.2 mg/L target. Compared with 8114
DOI: 10.1021/acs.energyfuels.5b01213 Energy Fuels 2015, 29, 8110−8117
Article
Energy & Fuels
Figure 4. Sulfur K-edge results for varying sulfur contents. (a) FY spectra; (b) Results of S K-edge peak area analysis.
Table 5. Summary of Mercury XANES and EXAFS Results sample
a
IPD (eV)
k-range (Å)
2D Br-S-Ash + Hg
7.494
3−13
S-Ash + Hg
7.672
4−12
2D Br-Ash + Hg
7.434
4−12
Path c
Hg−S Hg−Brb Hg−Ca Hg−Sc Hg−Ca Hg−Brb Hg−Ca
CN 1 1 1 2 2 2 2
R (Å) 2.320 2.503 1.962 2.335 2.084 2.500 2.065
± ± ± ± ± ± ±
0.027 0.015 0.090 0.015 0.138 0.008 0.039
σ2 (Å2)
ΔE (eV)
R-factor
± ± ± ± ± ± ±
−5.40 ± 3.24
0.003
5.77 ± 3.57
0.012
−3.02 ± 1.77
0.007
0.006 0.005 0.030 0.002 0.020 0.006 0.032
0.004 0.002 0.014 0.002 0.017 0.001 0.009
Crystal structure from HgC2.61 bCrystal structure from HgBr2.62 cCrystal structure from Hg(mpgH)2.63
sulfur-impregnated sorbents are shown in Figure 4a. There are three dominant sulfur peaks, showing the differences between the S present in the raw ash and in the sulfur-impregnated samples. The sulfur in the raw ash is predominantly in the oxidized (sulfate) form (E0 ≈ 2481.6 eV), while the sulfurloaded samples (Low S-Ash, S-Ash, High S-Ash and 2D Br-SAsh) contain the sulfate peak as well as two other peaks around E0 ≈ 2472 and 2474 eV. The peak at 2472 eV is characteristic of sulfur in the reduced form (e.g., elemental sulfur, S0, and sulfide, S2−). The peak around 2478 eV indicates the presence of sulfur as an intermediate sulfur species (e.g., sulfoxide and thiophene) and is also similar to peaks previously seen for CaS, which has pronounced features around 2474 and 2478 eV.60 After sulfur addition, the amount (intensity) of S0 and S2− increased while the relative intensity of sulfate decreased. This observation confirms successful loading of S0 and S2−, which are known to be effective in Hg0 capture.3,44 The amount of intermediate sulfur species likely increased partially due to the formation of CaS for biomass ash of relatively high calcium content, but it was not known if these species contribute to Hg0 capture. To better understand the changes in sulfur speciation after sulfur loading, the relative amount of sulfur species were calculated based on peak area and were found to be quite different among the Low S-Ash, S-Ash, and High S-Ash samples as shown in Figure 4b. Considering these results with those of mercury capture shown in Figures 1 and S5, the following observations about the speciation of sulfur and the mercury uptake can be made. The sulfur in the raw ash was predominantly in the form of sulfate and exhibited a low mercury uptake, indicating that sulfate does not appear to participate in mercury removal by the functionalized sorbents. From Low S-Ash to S-Ash, the amount of oxidized S decreases, while the intermediate and reduced sulfur species increases, leading to an increased mercury capture as shown in Figures 1
and S5. Although the High S-Ash sample also contains a higher content of intermediate and reduced sulfur species and lower content of oxidized S, the mercury uptake by the High S-Ash is lower than that by the S-Ash. Overall higher concentrations of intermediate and reduced sulfur species are shown to improve mercury capture, which is in agreement with previously published results.3,26,44,46 However, the pore blockage, lower formation of mesopores, and/or bonding of S to Ca (as described earlier) appears to limit further increase in mercury capture with increasing addition of reduced S species for the sorbents synthesized in the current study under the idealized mercury loading conditions. Sulfur K-edge spectra were also obtained for the S-Ash sample and 2D Br-S-Ash sample with and without Hg0, and are discussed in Supporting Information. Hg LIII-edge XANES. Hg LIII-edge spectra in Figure 5 for Hg-S-Ash, Hg-2D-Br-S-Ash, and Hg-2D-Br-Ash samples show very different binding environments of Hg among the three samples. The sample loaded with both Br and S exhibits a binding environment that is a combination of the other two samples (i.e., the spectrum lies in between the other two). The inflection point difference (IPD) was calculated from the derivative spectra (inset of Figure 5), and the results are given in Table 5. IPD values can provide insight into the type of bonding environment around the Hg atoms. The higher the IPD, the higher the ionic nature of the Hg bond.6 The IPD values for the samples were in the range of 7.4−7.7 eV, which is more ionic than the Br-Ash previously tested (IPD = 5.6).3 Hg EXAFS. Although IPD value can indicate the type of bonding environment around the Hg atoms, it is an approximate measure. To better understand the Hg0 speciation, IPD should be used in combination with EXAFS analysis.3 Hg LIII-edge EXAFS were obtained to determine more precisely the bonding environment of the Hg species on the samples and are shown in Figure S7 in Supporting Information and in Table 8115
DOI: 10.1021/acs.energyfuels.5b01213 Energy Fuels 2015, 29, 8110−8117
Article
Energy & Fuels
Figure 5. Mercury LIII-edge XANES spectra. Inset: First derivative of the normalized XANES spectra. Figure 6. Illustration of proposed Hg0 capture mechanism by 2D Br-SAsh. Fused rings represent carbon in wood ash.
5 along with the fits to the spectra. In the case of S-Ash, the EXAFS data show the association of mercury with sulfur and carbon on the ash, which is similar to the commercial activated carbon tested in our previous work.3 The IPD value of 7.5 eV is similar to the value reported by others for HgS.4,6 A likely mechanism of mercury capture could involve oxidation and subsequent binding to sulfur or carbon species on the S-Ash, similar to the mechanism previously proposed.3,25 In the case of 2D Br-Ash, the Hg LIII-edge EXAFS results show an association of mercury with Br and carbon. It is interesting to note that even though the Br concentration is very low (
