Influences of Copper ( II ) Chloride Impregnation on Activated Carbon for Low-Concentration Elemental Mercury Adsorption from Simulated Coal Combustion Flue Gas

In this study, the Hg adsorption equilibrium and kinetics of a coconut-shell-based activated carbon impregnated with CuCl2 were examined with respect to their resulting physical and chemical properties. Integrating the results from N2 adsorption isotherm at 77 K, scanning electron microscopy, elemental analysis, X-ray photoelectron spectroscopy, and Hg adsorption experiments under N2 and simulated coal-combustion flue gases conditions, it was found that HCl pretreatment could enhance Hg adsorption of crude activated carbon; the Hg adsorption capacities of crude and HCl-pretreated activated carbon under N2 condition were 95.8 and 225.4 μg g, respectively. Additionally, CuCl2 impregnation further increased the adsorption capacity of crude. The Hg adsorption capacity of crude activated carbon with 8% CuCl2 impregnation was 631.1 μg g. However, the equilibrium Hg adsorption capacity decreased when Cu loading exceeded 8 wt%, suggesting that adequate forms of surface Cu, O and Cl interacting with flue gas components and Hg, as well as the presence of pores with specific size ranges allowing rapid transport of the Hg molecules into the interior of the activated carbon and as energy sinker govern the overall chemisorption process. Pseudo-second-order kinetic model could best describe the adsorption behaviors of tested samples under both test conditions, indicating that Hg adsorbed on the activated carbon surface could be explained by bimolecular reaction mechanisms.


INTRODUCTION
Mercury (Hg) and its compounds emitted from anthropogenic sources, e.g., coal-fired power plants, industrial boilers, waste incinerators, sinters, and cement plants, have tempted substantial attention due to their high toxicity, bioaccumulability, and global transport behaviors in atmosphere (Kumari et al., 2015;Chen et al., 2016;Marusczak et al., 2016;Wang et al., 2016).Coal-fired power plants were reported as the largest single source in most countries in Hg emissions (Pacyna et al., 2010).Hg is present in the coal-combustion flue gases in three major forms, namely, particle-bound (Hg p ), oxidized (Hg 2+ ), and elemental (Hg 0 ) forms (Hsi et al., 2010;Wilcox et al., 2012).Hg p and Hg 2+ can be readily captured by traditional air pollution control devices, such as electrostatic precipitators and wet flue gas desulfurization.In contrast, Hg 0 is highly volatile, insoluble in water, and therefore more difficult to be removed.Furthermore, low-concentration Hg 0 (i.e., at 1-10 ppb v level) present in coal combustion flue gas streams leads to extreme challenge to control due to masstransfer limitation.Consequently, novel approaches for low-concentration Hg 0 removal from coal-combustion flue gases have lured marked attention in recent years.
Numerous studies pertaining to develop effective technologies on low-concentration Hg 0 control have been conducted (Li et al., 2015).Using porous materials, especially activated carbons as adsorbents, have been shown as profitable Hg 0 emission control approaches (Lin et al., 2015).The adsorptive efficiency of activated carbon largely depends on the surface and porous characteristics of activated carbon, including surface area, pore volume and size distribution, and surface functionality.The adsorptive environment, namely, flue gas condition, also plays a key role influencing the Hg 0 adsorption effectiveness of activated carbon.For the past decade, sulfur impregnation has been widely reported to considerably enhance the equilibrium Hg 0 adsorption capacity of activated carbon (Hsi et al., 2001(Hsi et al., , 2002;;Vitolo et al., 2002;Feng et al., 2006;Ho et al., 2008;Hsi et al., 2011Hsi et al., , 2012Hsi et al., , 2013Hsi et al., , 2014)).Besides sulfur, activated carbon impregnated with metal salts has demonstrated the competitive adsorption performance to those treated with sulfur.Copper salts, such as CuCl 2 and Cu(NO 3 ) 2 , have been extensively investigated as impregnation agents since they can not only increase the adsorption capacity of lowconcentration Hg 0 but also enhance the oxidation of Hg 0 into water-soluble Hg 2+ for activated carbon (Williams et al., 1997;Vidic et al., 2001;Kim et al., 2005;Nguyen-Thanh et al., 2005;Lee et al., 2008;Lee et al., 2009a, b, c;Zheng et al., 2012;Li et al., 2013;Du et al., 2014;Liu et al., 2015;Yang et al., 2016a, b, c, d). Lee et al. (2009a) suggested that activated carbon containing CuCl 2 may possess different sites available for Hg 0 oxidation and Hg adsorption, and the resulted oxidized Hg generated from the reactions between Hg 0 and CuCl 2 may be re-adsorbed at other available sites of the CuCl 2 -treated activated carbon.Li et al. (2013) indicated that CuCl 2 was the active species responsible for Hg 0 oxidation.In addition, activated carbons with greater chloride contents were found to have larger dynamic adsorption capacities than those with smaller chloride contents (Yang et al., 2016a).The degree of conversion of Hg 0 to Hg 2+ species was also observed to be directly related to the amount of chloride on the activated carbon (Vidic et al., 2001).Furthermore, Yang et al. (2016b) fabricated CuCl 2 loaded magnetospheres catalyst; they suggested that there were two different Hg adsorption sites, namely, Cl adsorption and Cu adsorption sites.Cu(NO 3 ) 2 may show competitive Hg 0 adsorption enhancement as CuCl 2 .However, activated carbon impregnated with Cu(NO 3 ) 2 needed subsequently calcination to increase oxidizing capability of activated carbons.Therefore, CuCl 2 was considered a more suitable impregnation agents for enhancing Hg 0 adsorption and oxidation.
Surface oxygenated groups could not only increase Hg 0 adsorption but also enhance the distribution of metal oxides/halides on the surface of activated carbon.Li et al. (2003) indicated that oxygenated functionality on the carbon surface, such as lactone and carbonyl groups, could be the active sites for Hg 0 capture.Yang et al. (2016d) also reported that the C=O group could be an effective electron acceptor, assisting the electron transfer for Hg 0 oxidation.Tseng et al. (2006) further reported that carbonyl groups were generated in activated carbon after HCl treatment; the oxygenated groups could be beneficial for subsequent distribution of CuO on the activated carbon surface.Therefore, to improve the extent of active sites for Hg 0 capture and oxidation, it may be feasible to pretreat activated carbon with acids to produce a variety of surface oxygenated groups acting as bridging sites, followed by metal precursor impregnation.
In the present study, we impregnated a high-quality, coconut-shell-based activated carbon with various amounts of CuCl 2 to produce effective adsorbents for removing low-concentration Hg 0 (i.e., in ppb level) from gas streams.Additionally, some activated carbons were pretreated with HCl solution to increase the extent of oxygenated groups on the activated carbon surface followed by subsequent CuCl 2 impregnation.The influences of CuCl 2 impregnation with and without HCl pretreatment on the physical and chemical properties and Hg 0 adsorption equilibrium/kinetics of resulting samples were then examined and better understood.Results obtained from this study is crucial from practical viewpoint because Taiwan Environmental Protection Administration has announced regulations in October 2013 to limit Hg emissions from coal-fired steam and cogeneration boilers; the Hg emissions should be lowered than 2 and 5 µg Nm -3 for new and existing facilities, respectively (Taiwan Environmental Protection Administration, 2014), which addresses the concerns and demands of successful control strategies for Hg emissions from coal-fired power plants.

Preparation of HCl Pretreated and CuCl 2 -impregnated Activated Carbon
A high-quality, coconut-shell-based activated carbon with a total surface area of approximately 1113 m 2 g -1 and > 90% microporosity was commercially obtained.The received activated carbon was initially immersed in hot deionized water for 2 h and then washed with cold deionized water several times to remove impurity.After oven-dried at 105°C for 24 h, the cooled sample was ground and passed through a 200-mesh sieve to obtain the homogeneous aliquot sample.The sample was surface-modified with two approaches: (1) HCl pretreatment followed by CuCl 2 impregnation, or (2) direct CuCl 2 impregnation.For HCl pretreatment, activated carbon sample of 10 g was mechanically stirred in a flask containing 50 mL HCl (37 v/v% from J.T. Baker) for 48 h.The treated activated carbon was separated from acid solution by percolating using a vacuum pump, washed with d.i.water, and then oven-dried at 105°C for 24 h.
The crude and HCl-pretreated activated carbon samples were subsequently impregnated with 2−16 wt% CuCl 2 (as Cu, 99.3% purity from J.T. Baker).The samples were immersed in CuCl 2 solution at 60−70°C for approximately 6 h that the water was completely vaporized.The resulting samples were then oven-dried at 105°C for 24 h to obtain the final products.The samples are designated as HCAC or CAC for samples with and without HCl pretreatment, respectively, and x wt% indicating the CuCl 2 impregnation amount.

Physical and Chemical Characterizations of Activated Carbon
The surface morphology was observed using a scanning electron microscope (Hitachi, model S-4700).Brunauer-Emmett-Teller (BET) specific surface area (S BET ), total pore volume (V total ) micropore (pore width < 2 nm) surface area (S micro ), micropore volume (V micro ), and pore-size distribution (PSD) were analyzed using a Quantachrome NOVA 2000e analyzer based on the N 2 adsorption isotherms obtained at 77 K. S BET was determined using the BET equation according to the ASTM D4820-96a method.S micro and V micro were calculated from t-plot evaluation based on the Jura-Harkins equation: t = [13.99/0.0340-log(p/p 0 )] 0.5 (Lippens et al., 1965).The range of relative pressures chosen for determining S micro and V micro was based on the values of thickness t between 0.45 and 0.8 nm.Micropore size distribution was simulated based on the quenched solid density functional theory (QSDFT).The mesopore size distribution was determined by the Barrett-Joyner-Halenda (BJH) method (Gregg et al., 1982).The chemical composition, including the mass concentration of C and H of samples, was determined by using an elemental analyzer (Thermo Flash EA 1112).Cu mass percent in the sample was verified by acid digestion followed by analysis with flame atomic absorption spectroscopy (FAAS; GBC AA932).The Cl content was determined with an energy dispersive spectrometer (EDS; JEOL JSM-7000F).The O content was estimated based on 100% -(C + H + Cu + Cl)%.The surface functional groups of the samples were further examined by using X-ray photoelectron spectroscopy (XPS; Physical Electronics, model ESCA PHI 1600).

Simulated Coal Combustion Flue Gas Hg 0 Adsorption Test
Detailed descriptions pertaining to experimental apparatus and procedures for Hg 0 adsorption tests have been described elsewhere (Hsi et al., 2011(Hsi et al., , 2012(Hsi et al., , 2013;;Chiu et al., 2014;Hsi et al., 2014); nevertheless, we present them here again for clarity.The adsorption tests were carried out under two conditions: (1) Hg 0 /N 2 condition and (2) simulated coalcombustion flue gas condition containing 6 vol% O 2 , 14 vol% CO 2 , 10 vol% H 2 O, 50 ppm v HCl, 200 ppm v SO 2 , 200 ppm v NO, and balance N 2 .The simulated coalcombustion flue gas composition was selected to reflect the typical condition of Taiwan's coal-fired power plants, in which low-sulfur bituminous and sub-bituminous coal blends are generally fired.The gas flow rate was controlled at 1.2 L min -1 with a Hg 0 concentration at 82 g Nm -3 (i.e., 10 ppb v ).Hg 0 was generated with a certificated Hg 0 permeation tube (VICI Metronics) at 70 ± 0.1°C to ensure a constant Hg 0 diffusion rate.The Hg 0 -containing gas homogeneously mixed with N 2 or simulated coal-combustion flue gas was passed through a temperature-controlled fixedbed column (0.5-in i.d.) containing a 10 mg sample mixed with 3 g quartz sand.The column length of the sample/sand mixture was approximately 5 cm and the duration for gas stream to pass the mixture was around 0.3 s.The temperature of fixed-bed column and heated Teflon tubes was controlled at 150°C and 110−130°C respectively to avoid moisture condensation.The effluent gas from the fixed-bed column flowed through the heated lines to the first impinger containing 20% SnCl 2(aq) that reduces any oxidized Hg compounds to Hg 0 followed by the second impinger containing 12% Na 2 CO 3(aq) to remove the acidic components.The gas then flowed through a Nefion tube to remove H 2 O and protect the downstream detector.The gas finally flowed through a gold amalgamation system (Brooks Rand model AC-01) where the Hg 0 in the gas was adsorbed.Hg 0 concentrated on the gold was subsequently desorbed at temperature > 400°C, and was sent as a concentrated Hg 0 stream to a cold-vapor atomic fluorescence spectrophotometer (CVAFS; Brooks Rand Lab Model III) for analysis.The test was stopped when 100% breakthrough reached, or conducted for up to 975 min.Hg 0 adsorption capacities, presented as g-Hg 0 g -1 adsorbent at given time, were determined by summing the mass of Hg 0 removed from the gas stream on the basis of the obtained breakthrough curves and then dividing by the mass of the adsorbent in the adsorption bed: where m i is the mass of adsorbed Hg, m adsorbent is the total mass of adsorbent, t' is the adsorption time, C i,in is the inlet Hg concentration, C i,out is the outlet Hg concentration at time t, Q g is the gas flow rate, and t is the time interval during the breakthrough test.

Properties of Adsorbents Prior to and After CuCl 2 Impregnation
The surface morphology of activated carbon before and after CuCl 2 impregnation is shown in Fig. 1.Dramatic changes on the surface shape and roughness of the samples after HCl pretreatment and CuCl 2 impregnation were not observed, implying that the distribution of CuCl 2 on activated carbon surface was fairly uniform.
The BET surface area, total pore volume, micropore surface area, and micropore volume of samples are shown in Table 1.Generally, CuCl 2 impregnation decreased the surface area and pore volume of all resulting activated carbons.The S micro of crude CAC was 1113 ± 117 m 2 g -1 and S BET was 1168 ± 77.0 m 2 g -1 .The calculated S micro /S BET , namely 95.2 ± 3.7%, suggests that crude CAC contains a marked amount of micropores.Additionally, pronounced difference in the S micro , S BET , and S micro /S BET between crude CAC and HCAC samples was not observed, indicating that the surface area and pore structure of activated carbon are less affected by acid treatment, or in other words, by the introduction of additional oxygenated groups.In contrast, pore blockage influencing surface area and pore volume by CuCl 2 impregnation was found; the extent of influence greatly depended on the amount of impregnated CuCl 2 .For instance, after 16 wt% CuCl 2 impregnation, the S micro dropped to 799.3 ± 35.6 m 2 g -1 and the S BET was about 901.3 ± 24.5 m 2 g -1 .The V micro also reduced from 0.58 ± 0.02 cm 3 g -1 (crude CAC) to 0.42 ± 0.01 cm 3 g -1 (CAC 16%) and the V total decreased from 0.65 ± 0.12 cm 3 g -1 (crude CAC) to 0.52 ± 0.02 cm 3 g -1 (CAC 16%).HCAC 2%, 8% and 16% samples also showed a similar trend of physical property changes with CAC series samples after CuCl 2 impregnation.It is important to note that, the S micro /S BET had no significant variation after CuCl 2 impregnation, suggesting that CuCl 2 uniformly blocked not only micropores but also macropores and mesopores of activated carbon.
The physical property changes due to surface modification could also be observed by the PSD simulation results, which are illustrated in Fig. 2  test samples can be classified as microporous activated carbons.Both crude CAC and HCAC samples had a unimodal micropore size distribution with a peak at around 0.6−0.8nm (Fig. 2(a)); peak shift after HCl pretreatment was not found.In contrast, CuCl 2 impregnation caused significant changes in PSDs for both CAC and HCAC samples.Notably, the change in peak volume, not in peak position was markedly influenced by the impregnation content.This observation supports that the decrease in micropore volume is mainly due to pore blockage, not pore shrinkage.Fig. 2(b) further shows that CuCl 2 impregnation had no significant effects on altering the mesopore structure of CAC and HCAC, which occupied < 10% of the total porosity.

. Based on the results in Table 1, all
Table 2 lists the results of element analyses for the activated carbon prior to and after HCl pretreatment and CuCl 2 impregnation with various concentrations.C and O were the main elements in CAC and HCAC samples.HCl pretreatment enhanced the balanced O content of crude activated carbon from 4.39 to 8.02 wt%, as well as increasing the Cl content to 0.34 wt%.CuCl 2 impregnation substantially increased the total Cu amount based on acid digestion/FAAS analysis, which increased with elevating impregnated CuCl 2 content.Cl content also increased with increasing CuCl 2 impregnation amount based on EDS analysis.However, the increase in Cl extent was much smaller than the increase in Cu content, which may be due to vaporization of Cl compounds during impregnation and oven-drying process.In contrast, the C contents of samples were shown to correspondingly decrease with increasing the Cu contents.For example, the C and Cu amounts of CAC 2% were 93.0 and 1.38 wt%, respectively, but CAC 16% had C content of 76.6 wt% and Cu content of 13.0 wt%.A similar tendency was also observed in CuCl 2 -impregnated HCACs.These experimental results support our previous observation that the impregnated agents could be doped into the porous adsorbents by means of surface coverage or pore filling, or by both mechanisms at the same time (Hsi et al., 2002(Hsi et al., , 2011(Hsi et al., , 2012(Hsi et al., , 2013(Hsi et al., , 2014)).
were further deconvoluted and their corresponding percentages of total C 1s peak area were then determined.The relative potions of C 1s peak area for crude CAC were 73.6% for graphite, 11.0% for hydroxyl, 5.55% for carbonyl, 9.32% for carboxyl, and 0.58% for π-electron resonance, respectively.After CuCl 2 impregnation, the graphitic content of activated carbon samples decreased.For example, the relative potions of C 1s for CAC 16% were 66.6% (graphite), 9.41% (hydroxyl), 19.3% (carbonyl), 4.18% (carboxyl), and 0.61% (π-electron resonance), respectively.These deconvolution results suggest that carbonyl content could increase with elevating CuCl 2 impregnation content.It is important to note that lactone and carbonyl functionality groups have been suggested to promote Hg adsorption through chemisorption based on the bonding energies calculated by density functional theory (Padak et al., 2006).Furthermore, Skodras et al. (2007) reported that oxygen functional groups can facilitate the electron-transfer process and Hg 0 oxidation on the carbon surface and act as potential adsorption centers of Hg 0 .Fig. 3 also illustrates that Cu 2p3/2 XPS spectrum obtained from the CAC 16% sample could be deconvoluted into two peaks within 932.0-934.2eV, including Cu + (31.6% of total Cu peak area with 932.0-932.8eV) and Cu 2+ (68.4% of total Cu peak area within 932.9-934.2eV).Cl 2p peak at 198.7 eV also indicated the presence of metal chloride after CuCl 2 impregnation.

Hg 0 Removal Performance under N 2 and Simulation Flue Gas Conditions
The Hg 0 adsorption capacities of the crude, HCl-pretreated, and CuCl 2 -impregnated samples are illustrated in Fig. 4. The crude CAC and HCAC adsorbents, in general, demonstrated smaller adsorption capacities than the CuCl 2 -impregnated samples under N 2 condition (Figs. 4(a) and (b)).For instance, the equilibrium Hg 0 adsorption capacity of crude CAC was 95.8 µg g -1 , but that for CuCl 2 -impregnated samples, such as CAC 2% and CAC 16%, was 417.2 and 435.8 µg g -1 , respectively (Fig. 4(a)).In addition, the Hg 0 adsorption capacity of CAC 8% achieved 631.1 µg g -1 , which was the greatest adsorption performance compared to other crude and CuCl 2 -impregnated CACs.These adsorption results may stem from two reasons: first, the S BET of CAC 8% is greater than those of CAC 2% and CAC 16% samples and consequently CAC 8% has more surface area containing active sites for Hg 0 adsorption.Second, although the S BET of crude CAC is greater than that of CAC 8%, crude CAC lacks Hg 0 adsorption sites that were provided by the impregnated CuCl 2 .Fig. 4(b) illustrates the Hg 0 adsorption results for HCAC series samples under N 2 condition.The Hg 0 adsorption capacity of HCAC 8% was 479.8 µg g -1 , which was larger than 458.4 µg g -1 (HCAC 16%), 332.6 µg g -1 (HCAC 2%), and 225.4 µg g -1 (crude HCAC).Notably, the crude HCAC sample (225.4 µg g -1 ) had a greater Hg 0 adsorption capacity than crude CAC (95.8 µg g -1 ), indicating that oxygenated groups and remained Cl introduced by HCl pretreatment could enhance Hg 0 adsorption of activated carbon without CuCl 2 impregnation.After CuCl 2 impregnation, HCAC series samples did not show a significant increase in Hg 0 capture.Notably, S BET and Cu amount of HCAC 8% were smaller than those of CAC 8% sample, which could be one of the reasons that HCAC 8% had a lower Hg 0 adsorption capacity than CAC 8%.(2015) have reported that different Cu loading on the sorbent surface caused various Cu coordination; isolated Cu 2+ at low Cu loading and associated Cu 2+ at high Cu loading ions existed in chlorine-free and chlorine-enriched situations, respectively.The active adsorption sites for Hg 0 were typically created by the chlorine-enriched coordination, not chlorine-free.Therefore, in our study, we expected that when CuCl 2 impregnation amount was increased, the chlorine-enriched coordination on the impregnated CAC dominated and enhanced the Hg 0 adsorption.Nevertheless, it is important to address that a large amount of CuCl 2 in activated carbon does not guarantee great Hg 0 adsorption performance.For instance, CAC 16% and HCAC 16% had a Cu content of approximately 13 wt% (Table 2) but the Hg 0 adsorption capacities are smaller than those of CAC 8% and HCAC 8%.The microporous structure is shown to be retained after CuCl 2 impregnation, indicating that the extent of pore surface area/volume is not the limiting factor in Hg 0 capture.In contrast, the decrease in microporosity, namely, the decrease in certain "key" micropores with specific pore sizes, may play a more important role in adsorption.In addition, a high CuCl 2 impregnation amount may also result in hindering effects on inherent, effective surface functional groups (such as inherent and introduced oxygenated groups), which could be another cause resulting in the smaller Hg 0 adsorption.
crude and CuCl 2 -impregnated CAC and HCAC samples under simulated coal-combustion flue gas condition.Similar to those observed under N 2 condition, the crude activated carbon without impregnated with CuCl 2 had smaller adsorption capacities (245.6 µg g -1 for CAC and 491.9 µg g -1 for HCAC, respectively) than CuCl 2 -impregnated samples.However, the effects of CuCl 2 impregnation for activated carbons under these two test conditions are dissimilar.Under the flue gas condition, CAC 2% and HCAC 2% samples had the largest Hg 0 adsorption capacities of 759.5 and 606.3 µg g -1 respectively compared to the other samples.Additionally, the Hg 0 adsorption capacities of samples under the flue gas condition were in general greater than those under N 2 condition, suggesting the enhancing effects of flue gas components, such as SO 2 , HCl, and O 2 , on Hg 0 adsorption (Hsi et al., 2012).These observations again support that the total surface area and total CuCl 2 amount of adsorbents could markedly influence Hg 0 adsorption but not the only determining factors on both Hg 0 adsorption equilibrium and kinetics, which was implied by the slope of the breakthrough curves and adsorption duration.Specific forms of Cu and oxygenated groups playing as catalytic and adsorption sites and pores with specific size ranges allowing rapid transport of the Hg molecules into the interior of the activated carbon and as energy sinker govern the adsorption process (Hsi et al., 2002(Hsi et al., , 2011(Hsi et al., , 2013)).

Kinetic Analysis of Hg 0 Adsorption on Crude and CuCl 2impregnated Activated Carbons
To further understand the mechanisms of Hg 0 adsorption on crude and CuCl 2 -impregnated adsorbents and the conceivable rate limiting steps, kinetic models were employed to the adsorption results from Hg 0 adsorption breakthrough tests.Because various functionalities including oxygen, chloride, and copper groups on the carbon surface may cause various types of adsorbent-adsorbate interactions, a lumped and simplified kinetic analysis is a practical approach from a system design viewpoint (Juang et al., 2000;Yang et al., 2005;Hsi et al., 2011Hsi et al., , 2012)).The pseudo-first and secondorder kinetic models assuming that adsorption is resolved by pseudo chemical reaction processes were chosen and the adsorption rates can be determined respectively by the following first-order and second-order reaction rate equations: (2) where q t (µg g -1 ) is the Hg 0 adsorption capacity at time t (min), q e (µg g -1 ) is the equilibrium Hg 0 adsorption capacity, and k 1 (min -1 ) and k 2 (g µg -1 min -1 ) are the pseudo first and second-order rate constants, respectively.These equations represent initial value problems and have analytical solution when combined with the conditions q t = 0 at t = 0 and q t = q t at t = t.The solutions for Eqs. ( 2) and ( 3) respectively are expressed as: ln(q e -q t ) = ln(q e ) -k 1 t (4) t/qt = 1/(k 2 q e 2 ) + t/q e (5) For the pseudo-first-order kinetic model, a linear driving force similarity is achieved when the driving force is depicted as a concentration dissimilarity (Ho et al., 1998).Some researchers have thus used the pseudo-first-order simulation to describe reaction, adsorption, and unsteady state diffusion.In contrast, when the adsorption seems to follow pseudo-second-order kinetic model and the rate is mostly concluded by chemisorption, the model on the basis of Langmuir-type second-order mass action rate expression could better address the adsorption characteristics (Skodras et al., 2008).Moreover, the chemisorption occurring on a strong heterogeneous surface can also be modelled by the Elovich equation, which is given by Given that q t = q t at t = t and qt = 0 at t = 0, the integrated form of Eq. ( 7) is where t 0 = 1/αβ.If t >> t 0 , Eq. ( 7) can be simplified as The modeling kinetic constants, R 2 values, and sum of squared error (SSE) using pseudo first-order, pseudo secondorder, and Elovich equations for simulation are shown in Table 3 and Fig. 6.R 2 value of pseudo first-order and pseudo second-order were 0.726−0.935and 0.849-0.997,able 3. Kinetic constants, R 2 and sum of squared values for pseudo-first-order, pseudo-second-order, and Elovich simulations.

Sample
Pseudo-first-order simulation Pseudo-second-order simulation Elovich simulation  respectively.The R 2 value of Elovich equation was within 0.728−0.962.Therefore, pseudo-second-order equation was demonstrated to best describe the Hg 0 adsorption kinetics over the entire fractional approach to equilibrium for all tested samples, which was consistent with those shown in (Skodras et al., 2008) and implied the nature of a chemisorption process.These simulation results suggest that Hg 0 adsorption on CuCl 2 -impregnated adsorbents appear to occur in bimolecular form, namely, two active sites are occupied to capture one molecule of Hg 0 .The active sites could be supplied by CuCl 2 impregnation, including both Cu and Cl functionality, or provided by other surface groups including inherent and additional oxygenated groups resulted from HCl pretreatment.

CONCLUSION
Effective Hg adsorbents via CuCl 2 impregnation on activated carbons to enhance equilibrium Hg 0 adsorption were successfully prepared.Overall, CuCl 2 impregnation altered the physical and chemical properties of activated carbons, increasing the total Cu and Cl content and lowering the surface area and pore volume of all prepared samples.SEM images suggested that the surface morphology of crude and treated activated carbons was similar.Elemental and XPS analyses not only verified the increase in O content after HCl pretreatment and the presence of Cu in activated carbons after CuCl 2 impregnation, but also showed that Cu was present in both Cu 2+ and Cu + forms.
CAC and HCAC samples after 2 wt% and 8 wt% CuCl 2 impregnation possessed the largest Hg 0 adsorption capacity under N 2 and simulated coal combustion flue gas conditions, respectively.Results from this study support our earlier finding that a large amount of CuCl 2 in activated carbon does not necessarily guarantee great Hg 0 adsorption performance.Furthermore, total pore surface area/volume is not the limiting factor in Hg 0 capture.The decrease in certain key microporosity with specific pore sizes may play a more critical role in Hg 0 adsorption.Specific forms of Cu and oxygenated groups and the presence of Cl playing as catalytic and adsorption sites, adequate interaction of surface functionality, flue gas components and Hg 0 , and pores with specific size ranges allowing rapid transport of the Hg molecules into the interior of the activated carbon and as energy sinker govern the chemical adsorption process.

Fig. 6 .
Fig. 6.Comparison of pseudo-first-order and pseudo-second-order kinetic simulations for (a) N 2 condition and (b) simulated flue gas condition.

Table 1 .
Total/micropore area and volume and microporosity of activated carbon samples prior to and after HCl pretreatment and CuCl 2 impregnation.

Table 2 .
Chemical composition of activated carbon samples prior to and after HCl pretreatment and CuCl 2 impregnation.
The Hg 0 adsorption results obtained under N 2 condition for CuCl 2 -impregnated activated carbons indicate the