Effects of CO and CO2 on the Removal of Elemental Mercury over Carbonaceous Surfaces.

Coal gasification is a popular method for the optimization of coal utilization and the reduction of environmental pollutant emissions. However, the reductive atmosphere of its products is disadvantageous for removing elemental mercury (Hg0). Activated cokes (AC) was employed in this work for mercury capture in a reducing atmosphere. The high-temperature heating decreases the mercury-removal capability of carbon sorbents because the carbonaceous surface is becoming oxygen-depleted and micropore-decreased after the heating treatment. The mechanism of mercury adsorption in pure nitrogen follows the Mars-Maessen mechanism over the carbon sorbents. To identify the effects of carbon monoxide (CO) and carbon dioxide (CO2) on Hg0 removal, the Hg0-adsorption and thermal desorption experiments were carried in a fixed-bed reaction system. CO inhibits both the chemisorption and physisorption of Hg0. CO2 competes for the active sites, lactone groups and hydroxyl groups, and occupies the micropores, which is beneficial to adsorb Hg0 physically. When CO and CO2 coexisted, the removal efficiencies show steadier than those in monocomponent gas (only CO or CO2). CO2 can resist the negative effect of CO on Hg0 removal, to some extent, because CO2 can inhibit the oxidation and disproportionation of CO. This experimental study provides practical guidance for the development of mercury-removal technology with carbon materials in the coal gasification plant.

Introduction

World Health Organization (WHO) has picked out mercury as one of chemicals of the major public health concern for its high toxicity.[1] Mercury is the only heavy metal participating in the atmospheric cycle, which can transport for a long distance in the form of gaseous Hg0. In ecosystems, such as soils, sediments, and aquatic systems, elemental mercury can partly convert into methylmercury (MeHg), which can readily bioaccumulate in the bodies of aquatic animals.[2,3] When people consume a contaminated fish and seafood, MeHg can get into the human body and damage the central nervous system or even cause death in a condition of severe exposure.[4,5] Anthropogenic sources emitted 2220 tons of mercury to the atmosphere in 2015,[6] and coal-burning was responsible for nearly 21%.[6] Global coal consumption made up 27% in primary energy in 2019, which is the lowest level in 16 years, but coal remains as the dominant energy source in some emerging countries with huge economic development demands, particularly in China, Indonesia, and Vietnam.[7] Thus, more attention should be given to reduce mercury emissions during the process of coal utilization.

In recent years, gasification has become a core technology for the high-efficient and clean utilization of coal resources. Coal, or mixed with biomass,[8,9] reacts with gasification agents at about 700 °C, and then it is converted into synthesis gas (syngas) consisting of carbon monoxide, carbon dioxide, and hydrogen. These gaseous products are used for combusting for advanced power generation,[10,11] producing fine chemicals,[12] or synthesizing fuel via the Fischer–Tropsch process.[13] The predominant species of mercury, releasing from coal, among the gaseous products is Hg0 due to the high temperatures and reducing atmospheres.[14−16] Given that Hg0 has high volatility and low solubility in water, capturing elemental mercury with heterogeneous catalytic adsorbents,[17−19] especially modified carbon materials, is one of the favorable methods. CO2 needs to be separated from syngas and condensed for storage to reduce the greenhouse effect, but trace mercury in CO2 stream can corrode aluminum equipment[20,21] when it flows through purification and compression units. Also, mercury poisons metal catalysts in synthetic reactions.[20] It is necessary to remove mercury after the pyrolysis process considering the endangerment of mercury.

Porous carbon materials are widely used in the removal of pollutants from the exhaust gases because they have flourishing pore structures and prodigious surface areas as well as plentiful functional groups. Hydrocarbon functional groups can activate neighboring carbon atoms and promote the capacity of carbonaceous surfaces for Hg0 removal.[22] Meanwhile, it has been reported that the C=O groups in carbonyl groups (C=O) and ester groups (C(O)–O–C) could oxidize Hg0 and turn into C–O,[23,24] while phenol and carboxyl groups could adsorb mercury physically.[25]

The performance of two gases over a carbonaceous adsorbent surface has been evaluated by experiments and DFT simulations. The micropore diffusivity of CO in activated carbon is faster than CO2’s at 303 K,[26] but the order of the stabilities is just the opposite.[27] In addition, when CO concentration is high, it competes with Hg0 for adsorption sites, and CO2 does not disturb the mercury adsorption.[27] Shen et al.[18] supposed CO has a negligible effect on a CuCl2-AC sorbent, whereas other researchers have suggested CO takes a passive role because it can generate deposited carbon over the adsorbent surface[28] and block the pore structures.[18] The CO2 concentration in the syngas varies from 7 to 40 vol %.[29−31] When Diamantopoulou et al.[32] added 12 vol % CO2 into a nitrogen atmosphere, the mercury-removal efficiency dropped drastically by 85%. The result shows CO2 is also harmful to mercury adsorption. However, there is still a lack of detailed experimental study on the situation that CO and CO2 coexist.

In this work, activated cokes were pretreated under different timings at 1000 °C for further study. The removal of Hg0 was investigated in a laboratory-scale fixed-bed reactor, and the adsorption products were identified by the temperature-programmed desorption (TPD). Moreover, the characterization methods of the fresh and used sorbents were as follows: scanning electron microscope (SEM), Brunauer–Emmett–Teller (BET), X-ray photoelectron spectroscopy (XPS), and Fourier transform infrared spectroscopy (FTIR). The influence of CO and CO2 on activated cokes was discussed.

Results and Discussion

Characterization of Activated Cokes

Figure shows the microphotographs of activated cokes with different heating times under 1000 °C. The surface of the raw activated coke is rich in porosity, but the high temperature flatted the other surface.

Figure 1

SEM images for (a, b) AC, (c, d) AC2, and (e, f) AC10.

The parameters of the pore structure of ACs are shown in Table , the branches of isotherms are shown in Figure a, and the pore distributions are presented in Figure b. AC had the largest specific surface area (470.5m2/g); the microporous areas of two calcined ACs dropped by more than 50%, compared with the raw AC. The curves in Figure a have hysteresis loops, the characteristic feature of mesoporous matters,[33] which means they belong to Type IV isotherms. Figure b exposes that the pore diameters of three kinds of activated cokes were mainly disturbed in the range of 0–2.0 nm (micropores) and 2–20 nm (mesopores). An excessively pretreating temperature (1000 °C) caused the microporous walls to collapse, but the mesopores were developed during heating duration. The longer the heating process, the larger the mesoporous volume. It may explain the relative extension in the total volume of AC10 compared with AC2.

Table 1

Textural Properties of AC, AC2, and AC10

	micropore (<2 nm)		mesopore (2–50 nm)		total pore
sample	Vmic (cm3/g)	Smic (m2/g)	Vmes (cm3/g)	Smes (m2/g)	Vtotal (cm3/g)	Stotal (m2/g)	average pore diameter (nm)
AC	0.14	334.0	0.20	136.0	0.37	470.5 ± 1.5	5.9
AC2	0.05	110.2	0.18	127.1	0.27	263.5 ± 0.8	6.0
AC10	0.07	160.4	0.22	152.1	0.32	332.9 ± 0.6	5.9

Figure 2

Nitrogen-adsorption–desorption isotherms (a) and distribution of pore size (b) of AC, AC2, and AC10.

The kinetic diameter of mercury is 0.30 nm.[34] The pores, whose diameters are in the range of 0.5–0.9 nm, are suitable for adsorbing mercury atoms in a physical way, according to the adsorption theory. The micropore volume of AC2 in that range (Figure b) is close to that of AC10, which implies that its mercury physisorption capacity is similar.

FTIR (Figure ) was carried out to explore the functional groups over the carbonaceous surfaces in the study. The broad peak at 3600–3200 cm–1[11,35] is from the stretching vibration of self-associated O–H or pyrrolic N–H; a sharp peak at 3642 cm–1[36] in AC2 and AC10 is related to a free or unassociated hydrogen bond of −OH among molecules. The peaks at 2960, 2922, and 2854 cm–1 belong to stretching vibration of C–H in −CH3, −CH2–, and −CH–, respectively.[35] These peaks of AC, AC2, and AC10 (Figure ) indicate that the intensity of aliphatic hydrocarbons was of positive correlation with the heating time. The population of aliphatic hydrocarbons enlarged when purged in the CO atmosphere. It demonstrates that CO could directly react with oxygen-containing functional groups increasing the content of saturated carbon at 80 °C. The physical adsorption of CO2 forms the peaks at 2300–2400 cm–1.[11] The vibration peaks at 1775–1345 cm–1[11,35] are identified as the oxygen-containing functional groups (carbonyl and lactone) vibration and aromatic C=C stretching vibration. The distribution of the peaks of purged AC2s among this zone differs from the fresh AC2, which emphasizes that CO and CO2 would react with the oxy-groups that are involved in the oxidation of mercury. The peaks at 877 and 798 cm–1 represent the substituted benzene ring with isolated hydrogen and two neighboring hydrogen or angular condensation ring systems. The types and numbers of oxygen-containing functional groups greatly changed during the heating treatment. Therefore, the influence of CO and CO2 should not be ignored when studying the mercury-removal capacity of activated cokes.

Figure 3

FT-IR spectra of different activated coke samples (AC2-CO: AC2 purged by CO for 3 h at 80 °C, AC2-CO2: AC2 purged by CO2 for 3 h at 80 °C, AC2-CO/CO2: AC2 purged by CO and CO2 for 3 h at 80 °C, used AC2: AC2 used in the adsorption experiment in N2 for 2 h at 80 °C).

XPS analysis was taken to characterize the surface functional groups and their proportions on the surface of the samples with different treatments. The multipeaks of C1s and O1s were simulated by Avantage software, and the atomic concentrations of two elements were calculated at the same time. The relevant data are shown in Figure and Table .

Figure 4

C1s and O1s XPS spectra of (a)–(c) AC2, (c, d) AC2-CO, (e, f) AC2-CO2, (g, h) AC2-CO/CO2, and (i, j) used AC2.

Table 2

Results of XPS Spectra of C1s and O1s for Different Active Cokes

sample			relative intensity (%)
state	structure	position (eV)	AC	AC2	AC10	AC2-CO	AC2-CO2	AC2-CO/CO2	used AC2
C1s	C–C	284.8	43.2	46.9	45.7	54.2	49.9	51.0	58.4
	C–O	285.5	31.8	38.8	39.2	37.4	35.8	38.1	34.7
	C=O	288.0	5.2	0.8	2.8	0.2	1.1	0.9	0.3
	C(O)O	289.1	2.5	7.7	5.1	6.4	6.8	4.9	6.0
	π–π*	291.2	7.1	5.8	7.4	1.8	6.4	5.2	0.6
O1s	C=O	532.1	81.1	77.1	80.6	62.2	86.7	94.0	79.9
	C–O	533.1	3.4	11.9	7.5	27.1	5.1	0.2	19.8
	–OH	534.2	3.9	2.8	1.3	0.6	0.6	0.6	0.4
	free oxygen	535.6	11.6	8.2	10.6	10.1	7.5	5.3	0

The C1s peak could be recognized as five main types:[37−39] C-graphite (∼284.8 eV), the C in C–O bonding (∼285.5 eV), the carbonyl carbon (∼288.0 eV), the carboxylate C (C(O)O, ∼289.1 eV), and π–π* (∼291.2 eV). The relative intensity of the carbonyl carbon drops from 5.2% (AC) to 0.8% (AC2), and the relative intensity of the carboxylate carbon lifts from 2.5% (AC) to 7.7% (AC2). As carbonyl and lactone groups can oxidize mercury in the temperature range of 30–210 °C,[40] the loss of C=O would negatively impact the capturing process of Hg0 but the increment of C(O)O would do the opposite. It coincides with the adsorption experiment results that the Hg0-removal efficiencies of AC and AC2 were close (53.46 and 49.31%, respectively). AC2 lost the major physical adsorption capacity, but the increase of C(O)O covered the loss. Also, the C=O and C(O)O proportions of AC2-CO were 0.2 and 6.4%, these fractions in AC2-CO2 were 1.1 and 6.8%, and their percentages of AC2-CO/CO2 were 0.9 and 4.9%. These discrepancies show that CO could decay the C=O but CO2 enhance it, and both gases would reduce the contents of C(O)O. When AC2 processed with CO and CO2 together, the effects of both sides on C=O could be counteracted, but the adverse impact on C(O)O would be worsened.

The O elemental mainly exists in four forms as follows:[38] C=O (∼532.1 eV), C–O (∼533.1 eV), −OH (∼534.2 eV), and free oxygen (∼535.6 eV). After the purging treatment and adsorption test, the O/C atomic ratio of the samples has fallen from 0.10 to 0.06–0.07 (purging treatment) and 0.05 (adsorption) because the oxygen was consumed in these processes. The −OH portions of AC, AC2-CO, AC2-CO2, AC2-CO/CO2, and the used AC2 are 2.8, 0.6, 0.6, 0.6, and 0.4%, respectively. The data indicates that the hydroxyl and carboxyl oxygen would oxide Hg0, and CO and CO2 would compete for the active sites with Hg0.

Mercury-Removal Performance

The mercury-removal efficiencies of different types of active cokes under the nitrogen atmosphere are illustrated in Figure . The Hg0-removal efficiencies at 80 °C of AC, AC2, and AC10 are 53.46, 49.31, and 10.03%, respectively. The higher temperature restrained the performance of AC2 in nitrogen.

Figure 5

Mercury average-removal efficiencies of AC and ACs within 2 h in nitrogen.

The terrible performance of AC10 could be interpreted by the poor physical adsorption and low oxygen contents. As mentioned above (Table ), the low microporous volume of AC10 determined its terrible physical adsorption.[41,42] The drop of oxygen species is due to the prolonged heating duration. The oxygen–carbon (O/C) ratio is 0.06 over the surface of AC10, which is approaching to the used AC2’s. The active oxygen comes from free oxygen and oxygen-containing functional groups (carbonyls, lactones, etc.),[40] and the amount of active oxygen is relative to the intensity of chemisorption. Thus, AC and AC2 show a better performance than AC10.

In a pure nitrogen atmosphere, the chemical adsorption of Hg0 would follow the Mars–Maessen mechanism.[43,44] The specific reaction formula is as follows:

As AC2 and AC10 are in a similar distribution of the micropores, which play an important role in physisorption, it could be deduced that chemical adsorption of AC2 is the dominant one. Within the temperature range of 80–120 °C, the efficiency decreases notably as the temperature increases. The results are consistent with the reported literature data.[40] It is thought that the higher temperature could supply sufficient energy for the Hg0 oxidation, but the results are the exact opposite. The higher temperature does no favor to the physisorption, which is exothermic, so mercury atoms have less possibility to stay on the surface. The first reaction R1R1 was restrained at the higher temperature, and, of course, the second reaction R2R2 could hardly take place when Hg0 could hardly get close to the active oxygen.

Since AC2 has enough oxygen-containing functional groups and not enough micropore structure, it is a suitable adsorbent for exploring the effects of CO and CO2 on the carbonaceous surface. All the adsorbents used in the subsequent experiments are AC2 without special notification.

Mercury-Removal Activities in CO Atmosphere

CO is one of the main components of coal gasification products. Studying the effect of CO on the mercury-removal performance can provide the suggestions for the application of adsorbents in a reducing atmosphere. The average Hg0-adsorption efficiency of AC2 is dotted in Figure , in various concentrations of CO within 1 h at 80 °C.

Figure 6

Hg0-removal efficiency of various CO concentrations within 1 h at 80 °C.

When CO concentration is 20 vol %, the Hg0-removal efficiency is 29.66%. As the concentration of CO rises to 30 and 40 vol %, the removal efficiency is 26.08 and 28.16%, respectively. No question that CO inhibited Hg0 removal, and the main reasons are as follows:

The CO purging treatment may decline the quantity of oxygen-containing functional groups and free oxygen on the surface of AC based on Table . CO could compete with Hg0 for the active oxygen (carbonyls and lactones, especially), impeding the mercury (R4R4).

The following experiment was carried out to prove the deduction. An amount of 0.1 g of AC2 was placed in a quartz tube on the fixed-bed reactor. Then, the AC2 adsorbent was purged with CO at a flow rate of 400 mL/min for an hour at 80 °C. The treated AC2, named as AC2-CO1, was used in adsorption experiments in nitrogen at 80 °C.

The adsorption curves of AC2-CO1 and fresh AC10 are pictured in Figure . The adsorption curves converge together with time. The average adsorption efficiency of AC2-CO1 is only 8.28%. It meets the conclusion that the oxidation of Hg0 may slow down due to CO.

Figure 7

Mercury-removal efficiency curves of sorbents within 1 h in an N2 atmosphere.

The inverse reaction of Boudouard reaction (R5R5)[45] could take place. CO may transform into CO2 and graphite, and the latter would deposit on the pore structures.

When the reaction temperature rose to 120 °C, black deposits were observed on the inner wall of the quartz tube. When the quartz tube with a black substance was heated in air at a high temperature (1000 °C) for a moment, the black substance disappeared without special odors. Black deposits did not appear during the adsorption experiment in an N2 atmosphere at 120 °C. The black substance should be graphite.

CO, the strong reducing agent, would facilitate the formation of low valence states mercury (R6R6). This part is discussed further in Thermal Desorption Experiments.

Mercury-Removal Activities in a CO2 Atmosphere

Trace mercury can corrode the aluminum equipment in the carbon capture process.[20,21] It is essential to remove mercury from the gas before gathering CO2. Figure shows that Hg0-removal efficiencies in CO2 stay at about 30%. CO2 also has a negative impact on the adsorption of element mercury.

Figure 8

Hg0-removal efficiency of various CO2 concentrations within 1 h at 80 °C.

The previous studies believed that CO2 would not react with elemental mercury on the carbon-based surface,[27] but others[46] put forward a standpoint that functional groups such as lactone groups, carboxyl groups, and hydroxyl groups could enhance the adsorption of CO2. Hydrogen bonds are formed between the functional groups and the CO2 adsorbed on them.[47] Therefore, Hg0 has a competition with CO2 for physical adsorption around the functional groups.

The diameter of the micropores is very tiny, of which the van der Waals potentials generated by the pore walls overlap. Its van der Waals force is stronger than that of the mesopores. While the kinetic diameters of the Hg atom and CO2 molecule are 2.97[34] and 3.30 Å, respectively,[48] the adsorption behaviors of Hg and CO2 in pore structures are similar and are affected by pore sizes. At a pressure of 1 bar, the adsorption effect is optimal when the pore size is less than or equal to 0.8 nm.[49,50] CO2 molecules occupy the mercury physisorption position in the micropores, causing a loss in adsorption efficiency. Compared with the N2 experiment of AC2, the Hg0-adsorption efficiencies in the CO2 experiments decrease by about 20%.

Interaction between CO and CO2

Figure shows the tendency of mercury efficiency of various samples over time. The Hg0 average-removal efficiencies of 20% CO/5% CO2, 20% CO/10% CO2, and 20% CO/20% CO2 are 14.97, 22.68, and 29.83% within 1 h at 80 °C. With the proportion of CO2 increases, the adsorption efficiency enhances accordingly. When the experiment atmosphere contained only CO or only CO2, their adsorption efficiencies are higher than those of the CO and CO2 coexisting experiments at first. Soon the efficiencies of single-gas experiments (CO or CO2) fall rapidly from 74.62% (CO) and 56.30% (CO2) to below 30% at 20 min, whereas the efficiencies of dual-gas experiments (CO coexisting with CO2) remain parallel with the x-axis.

Figure 9

Hg0-removal efficiency in different atmospheres within 1 h at 80 °C.

As mentioned above, CO2 would cut off the path to physisorption and chemisorption of mercury. The higher concentration of CO2 in the dicomponent experiments, the more Hg0-adsorption sites occupied at the first 5 min. It made the 20% CO/20% CO2 group to have the lowest efficiency at first.

As one of the products in the inverse Boudouard reaction (R5R5) and the reduction reaction of HgO, CO2 could limit the disproportionation reaction of CO and the higher CO2 concentration could protect the divalent mercury from reducing. CO2 only occupied the active sites physically; as the reactions proceeding, the active oxygen would be released and react with Hg0.

Considering the high amounts of CO and CO2 in gasification products, it is reasonable to purify the gas before separating the syngas because CO2 could promote the adsorbent performance when CO exists.

Thermal Desorption Experiments

Different mercury-containing compounds can be converted to elemental mercury at different temperatures and flow out with the gas,[24,51] so the species and portions of mercury were analyzed by thermal desorption experiments.

After adsorbing and purging process, the adsorbents were heated in N2 at a flow rate of 0.25 L/min, and the temperature was programmed at the rate of 3 °C/min from 80 to 500 °C. The different purging conditions and distribution of various mercury components are given in Table . The temperature-programmed desorption curves are displayed in Figure . AC2 purged with N2 has two obvious desorption peaks, which are at 195 and 280 °C.

Table 3

Amount of Mercury Desorption of Activated Cokes at Various Conditions

							products
group	sorbents	purging conditions	total adsorption (μg)	purged amount (μg)	total desorption (μg)	recovery rate (%)	physisorption	chemi sorption
(a)	AC2-CO1	N2, 5 h	0.34	0.10	0.26	105.9%	78.6%	21.4%
(b)	AC2	CO, 5 h	2.02	0.80	1.21	99.5%	80.9%	19.1%
(c)	AC2	CO, 1 h; N2, 4 h	2.80	1.24	1.76	107.1%	47.2%	52.8%
(d)	AC2	N2, 5h	3.22	0.61	2.18	86.6%	38.7%	61.3%

Figure 10

Mercury-desorption curves of the AC2s with different treatments. ((a)AC2-CO1 purged by N2 for 5 h at 80 °C, (b) AC2 purged by CO for 5 h at 80 °C, (c) AC2 purged by CO for 1 h and N2 for 4 h at 80 °C, and (d) AC2 purged by N2 for 5 h at 80 °C).

The mercury-adsorption capacity of AC10 can be approximately considered as physical adsorption. The physical adsorption amount of AC2 can be thought as same as AC10’s due to their similar microporous structures. The total adsorption amount of 0.1 g AC10 for 2 h in N2 is 0.82–1.56 μg. Supposing that the physical and chemical adsorption contents did not change after purging in nitrogen, the mercury content of the peak at 195 °C in group d is 1.24 μg, which is in the range of the total adsorption amount of AC10. Therefore, the desorption peak at 195 °C is attributed to physical adsorption. In Hong’s study,[24] the desorption peak of physical adsorption was at 160 °C. The difference is related to the purging time of the desorption experiment and the flow rate of the desorption carrier gas.

The decomposition temperature range of HgO is between 200 and 380 °C,[51] and the decomposition temperature range of HgS is between 180 and 350 °C.[51] But neither the active cokes nor the experimental atmosphere contained the sulfur, so the desorption peak at 280 °C is attributed to HgO.

After adsorbing in nitrogen, the used activated cokes were separately purged with a mercury-free CO gas for 1 h (Figure c) and 5 h (Figure b). The proportion of desorption products corresponding to the 280 °C peak decreases from 61.33 to 52.8% (CO, 1 h) and 19.1% (CO, 5 h). Furthermore, the ratio of the purged amount to the total adsorption of group d is only 18.9%, which is much lower than the ratio of groups b and c (39.6%, 44.3%). These results prove that CO could directly reduce high-valence mercury as R8R8.

The adsorption curve of AC2-CO1 (Figure ) points out that CO purging treatment can cause the depletion of the adsorption performance of fresh AC. The adsorption amount of group a (0.34 μg) is only one-tenth of that of group d (3.22 μg). Not only AC2-CO1 had a poor chemical adsorption capacity but also its physical adsorption performance was weakened, which supports the view that CO disproportionation reaction can happen.

Conclusions

With the adsorption and temperature-programmed desorption, the mechanisms of CO and CO2 on the Hg0-removal performance of calcined activated cokes have been analyzed and verified. The conclusions are as follows:

The prolonged heating time can destroy micropores and the C=O but generate C(O)O and the saturated carbon. The chemical adsorption of Hg0, with oxygen functional groups, follows the Mars–Maessen mechanism in N2, and the reactions are vulnerable to temperature changes. The mercury-removal efficiencies of AC2 dived when the temperature rose to 120 °C.

CO could react with functional groups, especially C=O. It could reduce HgO directly and decompose into graphite blocking the microporous structures.

CO2 impedes the Hg0 oxidation through occupying the active sites and micropores physically.

When adding CO and CO2 into the same system, CO2 would slow down the negative effect of CO on mercury removal. The tendency of the removal efficiencies in dicomponent groups was stabilized after the first 10 min.

The desorption peak at 195 °C is attributed to physical adsorption, and the one at 280 °C belongs to chemical adsorption.

Due to the performance of AC2 in the dicomponent experiments, it would be better to remove mercury from the gasification products before syngas processing and usage. Further investigations are needed in developing a detailed mechanism of multiple reactants at the atomic level.

Materials and Methods

Preparation of Activated Cokes

The activated coke was purchased from the Henan Ultra-clean Water Treatment Technology Co., Ltd. (China). The activated coke was crushed and sifted through 60–80 mesh, washed with distilled water, and then dried in an oven at 110 °C for 24 h. The dried coke was named as AC. Afterward, two groups of AC were calcined to remove oxygen-containing functional groups under the N2 atmosphere for 2 and 10 h, respectively; and the resulting products were designated as AC2 and AC10.

Mercury Adsorption and TPD Experiments

All the adsorption experiments were performed on a fixed-bed catalytic reactor, as shown in Figure . A mercury vapor generating device (Valco Instruments Company Inc., the U.S.) was used to provide a constant initial concentration (50 μg/m3) of elemental mercury vapor carried by high-purity N2. The mixture contained CO (20, 30, or 40 vol %) and CO2 (5, 10, or 20 vol %), and the balance gas was N2. All the gases were precisely controlled by mass flowmeters (Beijing Sevenstar Electronics Co., Ltd., China). The total gas flow rate was 1.2 L/min in adsorption experiments; in other words, the gas hourly space velocity (GHSV) was approximately 120,000 h–1. The concentrations of mercury were accurately measured by a continuous mercury emissions monitor, Tekran 3310Xi (Tekran Instruments Corporation, the U.S.), whose sampling interval is 150 s. A 0.1 g sample was used in each test when the adsorption temperature was kept at 80 °C. Also, the reaction time was 120 min in pure nitrogen and 60 min in the mixture. The purging process was conducted in situ. CO, CO2, or N2 were used as purging gases at the total rate of 400 mL/min at 80 °C, and the time of this step lasts as required.

Figure 11

Diagram of the fixed-bed reaction system.

The TPD experiments were started instantly after the adsorption or purging process in the same reactor. The steps were as follows. First, the used adsorbent was purged with nitrogen gas (1 L/min) until the mercury concentration was less than 2 μg/m3 at the outlet. Second, while nitrogen (250 mL/min) flew through the gas lines, the desorption temperature increased from 80 to 500 °C at the rate of 3 °C/min. In the desorption experiment, the outlet mercury concentration was monitored in real-time by RA 915+ (Lumex Instrument, Russia), whose sampling interval was 1 s.

Tekran 3310Xi can monitor Hg0 and Hg2+ at the same time. During the adsorption experiments, the amount of Hg2+ was too low to consider. The mercury-removal efficiency (η) of the adsorbent is defined as:

In the equation, cin and cout are the Hg0 concentration at the inlet and outlet of the reactor (μg/m3), respectively.

The mercury average-removal efficiency is defined as the average value of η in a certain period.

In the thermal desorption experiment, the definitions of the mercury-adsorption amount (mads), purging amount (mpurge), and desorption amount (mdes) are as follows:

In this formula, Δt is the desorption time (s); v is the total gas flow rate (m3/s). The influence of thermodynamic factors on gas volume was neglected in this work.

Vent gas went through the off-gas treatment system before discharged into the air.

Characterization of Activated Cokes

The morphology of ACs was photographed by a field emission scanning electron micrograph (German Carl Zeiss, Sigma 500).

Nitrogen-adsorption isotherms were measured at −196.3 °C on physisorption analyzers (US Micromeritics Instruments Corporation, ASAP 2020) and so were the specific surface area and pore parameters of the sorbents. The pretreatment was degassing samples at 250 °C for 3 h. The total pore surface area (Stotal) was derived by the Brunauer–Emmett–Teller[29] method fitting for values of relative pressure (p/p0) varying from 0.05 to 0.35, while the total pore volume (Vtotal) used single-point total pore volume evaluation at p/p0 = 0.99. The average pore diameter and volume (Vmec) and area (Smec) of mesopore were calculated by the Barrett–Joyner–Halenda (BJH) algorithm with the data of the desorption branch of the isotherm. The volume (Vmic) and area (Smic) of the micropore were deducted from the t-plot analysis. Meanwhile, the DFT model was applied for the porosity distribution.

The elemental valence states were obtained by an X-ray photoelectron spectrometer (Thermo Fisher Scientific, ESCALAB 250Xi) with an Al K-Alpha X-ray source. The C1s binding energy value (284.8 eV) helped to calibrate the results.

To investigate the functional groups, samples were blended with potassium bromide (KBr) powders, then pressurized into thin slices to 15 MPa for 10 min by a tablet machine, and detected by Fourier transform infrared spectroscopy (Thermo Fisher Scientific, Nicolet iS10) at last. The resolution was set to be 4.0 cm–1, and the detection range was from 500 to 4000 cm–1.