Mercury Migration Behavior from Flue Gas to Fly Ashes in a Commercial Coal-Fired CFB Power Plant

: Mercury (Hg) emissions from coal-ﬁred power plants are of increasing concern around the world. In this study, ﬁeld tests were carried out to understand the Hg emission characteristics and its migration behaviors in a commercial CFB boiler unit with the electricity generation capacity of 25 MW. This boiler is equipped with one electrostatic precipitator (ESP) and two fabric ﬁlters (FFs) in series for removing particulates from the ﬂue gas. The EPA 30B method was used for simultaneous ﬂue gas Hg sampling at the inlet of the ESP and the outlet of the second FF. The Hg mass balance in the range of 104.07% to 112.87% was obtained throughout the CFB unit by measuring the Hg contents in the feed fuel, the ﬂy ash and the bottom ash, as well as in the ﬂue gas at the outlet of the particulate control device (PCD) system. More than 99% of Hg contained in the feed fuel was captured by the ﬂy ash, whilst less than 1% of Hg was remained in the bottom ash or the ﬂue gas after passing the PCD system. The gaseous Hg obviously migrated from the ﬂue gas to the ﬂy ash in the air pre-heater, where the ﬂue gas temperature decreased from 250 ◦ C at the inlet to 120 ◦ C at the outlet. Other gaseous Hg migrated from the ﬂue gas to the ﬂy ash in the PCD system, as the Hg concentrations in the ﬂue gas ranged from 3.14 to 4.14 µ g / m 3 at the inlet of the ESP and ranged from 0.30 to 0.36 µ g / m 3 at the outlet of the second FF. The average Hg contents in the ﬂy ash samples collected from the ESP, the ﬁrst FF and the second FF were 912.3, 1313.6 and 1464.9 ng / g, respectively, while the mean particle diameters of these ﬂy ash samples tend to decrease along the ﬂow pass in the PCD system. Compared to large ﬂy ash particles, smaller ﬂy ash particles exhibit higher Hg capture performance due to their high unburned carbon (UBC) content and large speciﬁc surface area. The migration of gaseous Hg from the ﬂue gas to the ﬂy ash downstream of the CFB boiler unit was easier than that downstream of the PC boiler unit due to high UBC content and speciﬁc surface area. the Hg the flue at the of the ESP and the of the following: (1) the Hg mass balance and distribution throughout the CFB unit, (2) the migration of the gaseous Hg from the flue gas to the fly ash in the APH, (3) the migration of the gaseous Hg across the PCDs system, (4) the Hg adsorption mechanism by the fly ash collected from ESP/FFs and (5) comparison of the Hg behavior between PC boiler unit and CFB boiler unit. The objective of this paper is to further understand the characteristics of Hg emissions from the commercial CFB boiler unit and to clarify the gaseous Hg migration from the flue gas to the fly ash throughout the and the provide a guidance for the control of the Hg emissions from CFB boiler


Introduction
Mercury (Hg) is a serious toxic heavy metal for human health due to its volatility, persistence and bioaccumulation [1]. Coal combustion for electricity and heat generation is the dominant source of anthropogenic mercury emissions [2][3][4]. It is estimated that coal-fired power plants account for 26% of the global anthropogenic Hg emissions to the atmosphere [5]. Many countries have taken relevant measures and formulated relevant regulations [6]. For example, the United States proposed the Mercury and Air Toxics Standards (MATS) rule to be enforced by April 15, 2015 [7]; China issued the latest coal-fired power plants air pollutant emission standard (GB13223-2011) in 2011 [8], which limited Hg concentration to less than 0.03 mg/Nm 3 . As a result, Hg emissions from coal-fired power plants should be considered to be a high-priority regulatory environment concern.
Hg compounds existing naturally in coal are released as gaseous elemental Hg (Hg 0 ) during the high-temperature combustion process. Hg 0 can easily enter the atmosphere because it is a kind of thermodynamically stable form with high vapor pressure and low water solubility [9]. As the flue gas is cooled down, Hg 0 is partially oxidized via the gas-phase reactions involving the oxygen and halogen species forming gaseous oxidized (Hg 2+ (g)) [10]. In addition to the oxidization of Hg 0 , Hg behavior in coal-fired boilers also includes the migration of gaseous Hg (Hg 0 (g) + Hg 2+ (g)) from the flue gas to the fly ashes that forms particulate-bound mercury (Hg P ) by the physisorption and the chemisorption. If the gaseous Hg adsorbed on the fly ashes increases, the gaseous Hg concentration in the flue gas will decrease correspondingly. As mentioned by Zhao et al. [6] in a critical review, the migration behavior of Hg in boiler units needs to be understood for reducing the mercury emissions from coal-fired power plants.
As a key component of the back-end in coal-fired boiler units, an air pre-heater (APH) is generally used to reduce the exhaust heat loss and improve the thermal efficiency of boilers by recovering the residual heat from the flue gas [11]. Several researchers [12][13][14][15] have investigated the effect of APH on the oxidization and the adsorption of Hg 0 (g) in the flue gas. Schofield et al. [12] found that the untreated steel or platinum surfaces could significantly convert Hg 0 into Hg 2+ without the need of any catalysts. The amalgamation can occur in the temperature range of 230 ± 30 • C, which is just located in APH of coal-fired boiler units [12]. Later, Schofield [13] pointed out that the oxidation of Hg 0 (g) to Hg 2+ (g) can be enhanced by inserting metal surfaces into the flue gas under an appropriate temperature range in a coal-fired pilot unit [13]. Romero et al. [14] investigated the impact of the operating conditions on the Hg emissions from a pulverized coal (PC) boiler unit. The authors found that the oxidation and the adsorption of gaseous Hg were enhanced in APH due to a high flue-gas quenching rate. Senior [15] suggested that the oxidation of Hg 0 (g) to Hg 2+ (g) in APH was mainly affected by the chlorine radical concentration in the flue gas and the UBC content of the fly ash. However, this research only investigated the change of the gaseous Hg concentrations at the inlet and outlet of APH, which is insufficient to fully understand the Hg conversion mechanism within APH.
Particulate control devices (PCDs), such as electrostatic precipitators (ESP) and fabric filters (FF), have been extensively applied to control the particulates emissions from coal-fired power plants. If Hg exists in the particulate phase at the inlet of ESP/FF, these devices can remove it efficiently [16][17][18]. According to Information Collection Request (ICR) database of United States Environmental Protection Agency [19], the average Hg removal efficiencies of ESP and FF are 27% and 58%, respectively. Furthermore, ESP/FF not only capture Hg P , but also enhance the migration of Hg 0 (g) and Hg 2+ (g) to the fly ash due to a good gas-solids contact. Wang et al. [20] investigated the Hg emission and speciation of six PC boiler units in China. The test results suggested that the gaseous Hg concentrations in the flue gas at the inlet of PCDs were higher than those in the flue gas at the outlet of PCDs. Several investigations have identified that FF showed better performance of Hg removal than ESP due to an excellent gas-solids contact across the dust cake on filters [21,22]. In addition, ESP and FF have been adopted in series in some coal-fired power plants, to control the particulate emissions. Zhao et al. [23] measured the mercury concentrations in the flue gas at the inlet and outlet of the PCD system (ESP+FF) and showed that the total Hg P capture efficiency could reach up to 99.95-99.97%. While the total Hg capture efficiency could be obtained easily, the individual contribution of Hg capture in either ESP or FF has not yet been distinguished.
It should be mentioned that the majority of investigations were based on PC boiler units [6], whilst only a few considered circulating fluidized bed (CFB) boiler units. In recent twenty years, CFB boilers Energies 2020, 13, 1040 3 of 15 have been rapidly employed in coal-fired power plants due to their excellent fuel flexibility, low NOx emission, high sulfur capture efficiency and so on [24]. Wichlinski et al. [25] reported that very high Hg content was determined in the fly ash collected from a CFB boiler unit. Wang et al. [26] obtained nearly 100% gaseous Hg removal in a CFB boiler unit, compared to only 27.56% or 33.59% Hg removed in two PC boiler units. Similarly, Zhang et al. [27] pointed out that gaseous Hg emissions from CFB boiler units was much lower than that from PC units. Our previous study found that the Hg content in the fly ash collected from the 350 MW CFB unit was much higher than that in the fly ash collected from the 330 MW PC boiler unit due to high UBC content and specific surface area [28]. This research mainly focused on the characteristic of Hg emissions from CFB boiler units and suggested that a stronger adsorption force between the gaseous Hg and the fly ash may occur in CFB boilers. Therefore, the migration behavior of Hg in CFB boiler units and the impact of APH and PCDs for the reduction of Hg emission needs to be further investigated.
In this study, full-scale field tests on Hg emissions were carried out in a commercial CFB boiler unit with the electricity generation capacity of 25 MW. This boiler unit is equipped with a tubular APH, an ESP and two FFs in series. To estimate the migration of Hg in this CFB boiler unit, the flue gas samples, the fly ash samples, the bottom ash samples and the feed fuel samples were collected. The EPA 30 B method, the internationally recognized standard method, was used to sample the gaseous Hg in the flue gas at the inlet of the ESP and the outlet of the second FF. The main contents included are the following: (1) the Hg mass balance and distribution throughout the CFB unit, (2) the migration of the gaseous Hg from the flue gas to the fly ash in the APH, (3) the migration of the gaseous Hg across the PCDs system, (4) the Hg adsorption mechanism by the fly ash collected from ESP/FFs and (5) comparison of the Hg behavior between PC boiler unit and CFB boiler unit. The objective of this paper is to further understand the characteristics of Hg emissions from the commercial CFB boiler unit and to clarify the gaseous Hg migration from the flue gas to the fly ash throughout the APH and the PCDs, which could provide a guidance for the control of the Hg emissions from CFB boiler units.

CFB Boiler Unit and Measurement Method
Full-scale field tests were performed in a commercial coal-fired CFB unit co-firing gangue and coal with the electricity generation capacity of 25 MW. A tubular APH was adopted to preheat the combustion air, and the PCD system consisted of one ESP and two FFs in series (ESP + FF1 + FF2) and was adopted to remove particulates from the flue gas. Figure 1 shows the CFB boiler unit together with two gas-sampling locations and seven solid-sampling locations during the field tests. Two flue-gas-sampling (G in ESP and G out FF ) points were located at the inlet of the ESP and the outlet of the second FF. The solid samples included the feed fuel collected from the coal conveying belt of the feeder, the bottom ash from the sampling port of the residue extraction mechanism, five fly ash samples from the inlet of the APH, the outlet of the APH, the discharging valve of the ESP, the discharging valve of the first FF and the discharging valve of the second FF, respectively, which were denoted as C, BA, FA 1 APH , FA 2 APH , FA ESP , FA FF1 and FA FF2 , respectively, in this study. The EPA 30B method was employed to obtain the gaseous Hg in the flue gas at the flue gas sampling points, in which the sorbent traps filled with the activated carbon were adopted to collect the total gaseous Hg. The sorbent traps were filled with two parts of activated carbon sorbents, which were used for sampling and verifying the Hg breakthrough of the first part, respectively. The Hg concentrations in the flue gas was defined as the Hg contents in the used activated carbon divided by the volume of the sampled flue gas. The sampling was considered to be successful, and the data was recognized as reliable, as the following two conditions were satisfied: (1) The Hg concentration in the second part activated carbon adsorbent was 10% less than that in the first part; and (2) the testing results of the parallel samplings were with less than 10% deviation. The self-designed fly ash sampling device was used for obtaining the fly ash at the inlet and the outlet of the APH, which consists of one Energies 2020, 13, 1040 4 of 15 vacuum pump and one cyclone separator connected in series. The fly ash particles were separated from the flue gas by the centrifugal force when the gas passed through the cyclone separator. During each test, the feed fuel, the bottom ash, the fly ash from the ESP, the fly ash from the first FF and the fly ash from the second FF were collected at 0.5 h intervals. The Hg content in the feed fuel samples, the fly ash samples, the bottom ash samples and the activated carbon used for measuring Hg in the flue gas were detected by LUMEX RA915 + analyzer (LUMEX Co., Ltd., Russia). Each solid sample was measured 10 times, to ensure the accuracy of the test data, and the average value was used for further analyses.

CFB Boiler Unit and Measurement Method
Full-scale field tests were performed in a commercial coal-fired CFB unit co-firing gangue and coal with the electricity generation capacity of 25 MW. A tubular APH was adopted to preheat the combustion air, and the PCD system consisted of one ESP and two FFs in series (ESP + FF1 + FF2) and was adopted to remove particulates from the flue gas. Figure 1 shows the CFB boiler unit together with two gas-sampling locations and seven solid-sampling locations during the field tests. Two fluegas-sampling (G in ESP and G out FF ) points were located at the inlet of the ESP and the outlet of the second FF. The solid samples included the feed fuel collected from the coal conveying belt of the feeder, the bottom ash from the sampling port of the residue extraction mechanism, five fly ash samples from the inlet of the APH, the outlet of the APH, the discharging valve of the ESP, the discharging valve of the first FF and the discharging valve of the second FF, respectively, which were denoted as C, BA, FA 1 APH, FA 2 APH, FAESP, FAFF1 and FAFF2, respectively, in this study.

Characterization of Fly Ash Samples
The particle size distribution and the specific surface area of the fly ash samples were detected by Mastersize 2000-layer analyzer (Malvern Co., Ltd, England) and ASAP2020M apparatus (Micromeritics Co., Ltd, USA), respectively. The UBC content in the fly ash sample was measured by the thermal gravimetric analysis (TGA) technique [29] (TA 4000 analyzer, Mettler Co., Ltd, Switzerland). In each TGA test, a 5 ± 0.5 mg sample was weighed and loaded into the crucible. N 2 with the flow rate of 80 mL/min was first introduced for approximately 30 min to purge the lines for stabilizing the apparatus. Then, the sample was heated from the room temperature to 750 • C at the fixed heating rate of 20 • C/min in the N 2 atmosphere to drive off volatile compounds. As the temperature reached 750 • C, it was held constant for about 30 min, to make sure no further weight loss occurred. Finally, N 2 was replaced by air, and the flow rate of air was also maintained at 80 mL/min, which oxidized UBC in the fly ash to carbon dioxide. The experiment was terminated when the weight loss no longer occurred.

Hg Content in Feed Fuel
Four full-scale field tests were conducted during two days, with the same boiler load and the feed fuel in this study. Test 1 and Test 2 were done in the first day, and Test 3 and Test 4 were done in the second day. The proximate and ultimate analysis of the feed fuel samples are listed in Table 1. The Hg contents in the feed fuel samples collected at each test are shown in Figure 2. It can be found that the Hg content ranges from 219.6 to 395.4 ng/g for the Test 1, 195.4 to 389.7 ng/g for the Test 2, 177.9 to 361.6 ng/g for the Test 3 and 189.5 to 381.6 ng/g for the Test 4, respectively. Correspondingly, their average values range from 254.2 to 291.2 ng/g, with the standard deviation of 50.79-72.63 ng/g. To minimize the error from both sampling and analysis procedures of field tests, the Hg content data fulfilling the following condition was used: Energies 2020, 13, 1040 where C i is the used Hg content data; C Hg and δ stand for the average Hg content and the standard deviation of ten measured Hg content values, respectively. For all the measured results shown in Figure 2, there are only two outliers, which demonstrates that almost all testing data are acceptable. These two outliers are removed based on the classical statistics method, and the calibrated average values are shown in the inset of Figure 2. As a result, the average Hg contents in four feed fuel samples are 270.4, 290.7, 241.8 and 245.4 ng/g, respectively. The average value of remaining data Outlier Outlier

Hg Mass Balance and Distribution in the CFB Unit
The Hg mass balance is a key indicator for verifying the accuracy of the field test data that is defined as the ratio of the input Hg amount to the output Hg amount. The Hg contents in the feed fuel samples are considered as the input Hg amount and also as the datum (100%) for the Hg mass balance calculation, while the summation of the Hg contents in the fly ash, the bottom ash and the stack gas stream is considered as the output Hg amount of the CFB boiler. Generally, the Hg mass balance is acceptable to be in the range of 70% and 130%, considering some uncertainty, such as fluctuation in the feed fuel, slight change in the boiler load, error in the Hg sampling and analysis procedures [30]. The Hg mass balances for the four tests are shown in Figure 3. It can be seen that the Hg mass balances across the CFB boiler unit are 112.87%, 104.63%, 105.18% and 104.07%, respectively. This confirms that the data of field tests in this study are reliable.

Hg Mass Balance and Distribution in the CFB Unit
The Hg mass balance is a key indicator for verifying the accuracy of the field test data that is defined as the ratio of the input Hg amount to the output Hg amount. The Hg contents in the feed fuel samples are considered as the input Hg amount and also as the datum (100%) for the Hg mass balance calculation, while the summation of the Hg contents in the fly ash, the bottom ash and the stack gas stream is considered as the output Hg amount of the CFB boiler. Generally, the Hg mass balance is acceptable to be in the range of 70% and 130%, considering some uncertainty, such as fluctuation in the feed fuel, slight change in the boiler load, error in the Hg sampling and analysis procedures [30]. The Hg mass balances for the four tests are shown in Figure 3. It can be seen that the Hg mass balances across the CFB boiler unit are 112.87%, 104.63%, 105.18% and 104.07%, respectively. This confirms that the data of field tests in this study are reliable.  The mass distribution of Hg in the combustion by-products, such as the fly ash, the bottom ash and the flue gas, is of importance to understand the migration of Hg in the CFB boiler unit. Table 2 provides the mass distribution of Hg in the combustion by-products, based upon the amount Hg output in the power unit. It is found that the Hg contents in the fly ashes collected by the ESP/FF account for 99.29%-99.45% of the total Hg output amount, whilst only 0.07% to 0.08% of Hg remains in the bottom ash, and 0.48% to 0.64% of Hg is in the flue gas leaving from the second FF. This fact suggests that almost all of gaseous Hg migrate from the flue gas to the fly ash in this CFB boiler unit. Hall et al. [31] pointed out that Hg adsorption on the fly ash cannot occur at temperatures above 350 °C. It is clear that the APH, the ESP and the FFs play key roles in the migration of Hg from the flue gas onto the fly ash in this CFB boiler unit.

Gaseous Hg migration across APH
APH is a kind of gas-gas heat exchanger equipped in coal-fired boiler units, and its role is to preheat the combustion air and reduce the exhaust heat loss by contacting the combustion air with the flue gas. In order to investigate gaseous Hg migration from the flue gas to the fly ash through the APH, the fly ash samples were collected by a self-designed particulate sampling device at its inlet and outlet. The UBC content and the specific surface area of these two fly ash samples are shown in Table 3. It can be seen that the UBC content and the specific surface area of these two fly ash samples are very close, so the fly ash samples collected by the self-designed particulate sampling device are suitable for exploring the migration behavior of gaseous Hg from the flue gas to the fly ash in the APH. The mass distribution of Hg in the combustion by-products, such as the fly ash, the bottom ash and the flue gas, is of importance to understand the migration of Hg in the CFB boiler unit. Table 2 provides the mass distribution of Hg in the combustion by-products, based upon the amount Hg output in the power unit. It is found that the Hg contents in the fly ashes collected by the ESP/FF account for 99.29-99.45% of the total Hg output amount, whilst only 0.07% to 0.08% of Hg remains in the bottom ash, and 0.48% to 0.64% of Hg is in the flue gas leaving from the second FF. This fact suggests that almost all of gaseous Hg migrate from the flue gas to the fly ash in this CFB boiler unit. Hall et al. [31] pointed out that Hg adsorption on the fly ash cannot occur at temperatures above 350 • C. It is clear that the APH, the ESP and the FFs play key roles in the migration of Hg from the flue gas onto the fly ash in this CFB boiler unit.

Gaseous Hg migration across APH
APH is a kind of gas-gas heat exchanger equipped in coal-fired boiler units, and its role is to preheat the combustion air and reduce the exhaust heat loss by contacting the combustion air with the flue gas. In order to investigate gaseous Hg migration from the flue gas to the fly ash through the APH, the fly ash samples were collected by a self-designed particulate sampling device at its inlet and outlet. The UBC content and the specific surface area of these two fly ash samples are shown in Table 3. It can be seen that the UBC content and the specific surface area of these two fly ash samples are very close, so the fly ash samples collected by the self-designed particulate sampling device are suitable for exploring the migration behavior of gaseous Hg from the flue gas to the fly ash in the APH.  Figure 4 presents the Hg contents in these two fly ash samples. The Hg content in the fly ash sample collected from the inlet of the APH is in the range from 978.9 to 1169.0 ng/g, while that in the fly ash sample collected from the outlet of the APH is in the range from 1842.0 to 2210.0 ng/g. As shown in the inset of Figure 4, the average values of these two samples are 1030.9 and 2018.6 ng/g, indicating that the gaseous Hg migrates from the flue gas to the fly ash in the APH.   Figure 4 presents the Hg contents in these two fly ash samples. The Hg content in the fly ash sample collected from the inlet of the APH is in the range from 978.9 to 1169.0 ng/g, while that in the fly ash sample collected from the outlet of the APH is in the range from 1842.0 to 2210.0 ng/g. As shown in the inset of Figure 4, the average values of these two samples are 1030.9 and 2018.6 ng/g, indicating that the gaseous Hg migrates from the flue gas to the fly ash in the APH.

Test1
Hg content (ng/g)   Figure 5 shows that the temperatures of the flue gas at the APH inlet and outlet were 260 ± 10 °C and 120 ± 10 °C, respectively. The chlorine radical concentration is closely related to the flue gas cooling rates and plays an important role in the oxidization of Hg 0 in the flue gas [32]. Compared to the gaseous Hg 0 , the gaseous Hg 2+ is easily adsorbed on the fly ash [33]. In addition, the adsorption of gaseous Hg on the fly ash is progressively increases with decreasing the flue gas temperature below 400 °C [20,34]. Our previous work also investigated the effect of the temperature on the adsorption of Hg 0 on the fly ash collected from another CFB boiler unit [35]. The results suggested that the adsorption of Hg onto the fly ash samples increased with the decrease of the temperature, and the adsorption capacity of the fly ash samples was the highest at 150 °C. Accordingly, the gaseous Hg obviously migrates from the flue gas to the fly ash through the APH, mainly due to the decrease of the temperature.  Figure 5 shows that the temperatures of the flue gas at the APH inlet and outlet were 260 ± 10 • C and 120 ± 10 • C, respectively. The chlorine radical concentration is closely related to the flue gas cooling rates and plays an important role in the oxidization of Hg 0 in the flue gas [32]. Compared to the gaseous Hg 0 , the gaseous Hg 2+ is easily adsorbed on the fly ash [33]. In addition, the adsorption of gaseous Hg on the fly ash is progressively increases with decreasing the flue gas temperature below 400 • C [20,34]. Our previous work also investigated the effect of the temperature on the adsorption of Hg 0 on the fly ash collected from another CFB boiler unit [35]. The results suggested that the adsorption of Hg onto the fly ash samples increased with the decrease of the temperature, and the adsorption capacity of the fly ash samples was the highest at 150 • C. Accordingly, the gaseous Hg obviously migrates from the flue gas to the fly ash through the APH, mainly due to the decrease of the temperature.

Gaseous Hg Migration Across PCD
ESP and FF are two kinds of commercial PCDs to remove particulates from the flue gas in coalfired boiler units, which also can affect the migration of the gaseous Hg and capture Hg P in the flue gas [15][16][17]. To investigate the effect of the whole PCDs system on the migration of the gaseous Hg, the Hg concentrations in the flue gas at the inlet of the ESP and at the outlet of the second FF were measured. The data in Figure 6 show that the Hg concentrations in the flue gas range from 3.14 to 4.14 μg/m 3 at the ESP inlet and from 0.30 to 0.36 μg/m 3 at the outlet of the second FF, respectively. These results confirm that the gaseous Hg noticeably migrate from the flue gas to the fly ash in the whole PCDs system.

The Effect of Fly Ash Properties on Hg Migration
The migration behavior of Hg in coal-fired boiler units is not only affected by APH and PCDs, but also affected by the properties of the fly ashes. Figure 7 outlines the migration route of the gaseous Hg from the flue gas to the fly ash. It is found that the UBC content and the specific surface area play key roles in the adsorption of gaseous Hg on the fly ash [6,[36][37][38]. In addition, the particle size of the

Gaseous Hg Migration Across PCD
ESP and FF are two kinds of commercial PCDs to remove particulates from the flue gas in coal-fired boiler units, which also can affect the migration of the gaseous Hg and capture Hg P in the flue gas [15][16][17]. To investigate the effect of the whole PCDs system on the migration of the gaseous Hg, the Hg concentrations in the flue gas at the inlet of the ESP and at the outlet of the second FF were measured. The data in Figure 6 show that the Hg concentrations in the flue gas range from 3.14 to 4.14 µg/m 3 at the ESP inlet and from 0.30 to 0.36 µg/m 3 at the outlet of the second FF, respectively. These results confirm that the gaseous Hg noticeably migrate from the flue gas to the fly ash in the whole PCDs system.

Gaseous Hg Migration Across PCD
ESP and FF are two kinds of commercial PCDs to remove particulates from the flue gas in coalfired boiler units, which also can affect the migration of the gaseous Hg and capture Hg P in the flue gas [15][16][17]. To investigate the effect of the whole PCDs system on the migration of the gaseous Hg, the Hg concentrations in the flue gas at the inlet of the ESP and at the outlet of the second FF were measured. The data in Figure 6 show that the Hg concentrations in the flue gas range from 3.14 to 4.14 μg/m 3 at the ESP inlet and from 0.30 to 0.36 μg/m 3 at the outlet of the second FF, respectively. These results confirm that the gaseous Hg noticeably migrate from the flue gas to the fly ash in the whole PCDs system.

The Effect of Fly Ash Properties on Hg Migration
The migration behavior of Hg in coal-fired boiler units is not only affected by APH and PCDs, but also affected by the properties of the fly ashes. Figure 7 outlines the migration route of the gaseous Hg from the flue gas to the fly ash. It is found that the UBC content and the specific surface area play key roles in the adsorption of gaseous Hg on the fly ash [6,[36][37][38]. In addition, the particle size of the

The Effect of Fly Ash Properties on Hg Migration
The migration behavior of Hg in coal-fired boiler units is not only affected by APH and PCDs, but also affected by the properties of the fly ashes. Figure 7 outlines the migration route of the gaseous Hg from the flue gas to the fly ash. It is found that the UBC content and the specific surface area play Energies 2020, 13, 1040 9 of 15 key roles in the adsorption of gaseous Hg on the fly ash [6,[36][37][38]. In addition, the particle size of the fly ash can also affect the adsorption of the gaseous Hg because the UBC content and the specific surface area are closely related to the particle size [39].
Energies 2020, 13, 1040 9 of 15 key roles in the adsorption of gaseous Hg on the fly ash [6,[36][37][38]. In addition, the particle size of the fly ash can also affect the adsorption of the gaseous Hg because the UBC content and the specific surface area are closely related to the particle size [39]. It is well-known that the properties of the fly ash collected from CFB boiler units are quite different from those collected from PC boiler units. Our previous work [27] confirmed that the UBC content and the specific surface area of the CFB fly ash were higher than those of the PC fly ash. Chen et al. [40] found lots of microspheres in the PC fly ashes, but most of the particulates were porous and with irregular shape in the CFB fly ashes. In order to understand the migration of the gaseous Hg from the flue gas to the fly ash in the CFB boiler unit, the Hg contents in the fly ash samples collected from the ESP/FFs were measured. As shown in Figure 8, the Hg contents in the fly ash samples vary from 769.2 to 1045.0 ng/g collected from the ESP, 1217.2 to 1440.9 ng/g collected from the first FF and 1329.0 to 1605.0 ng/g collected from the second FF. The average Hg contents in the fly ash samples collected from the ESP, the first FF and the second FF were 912.3, 1313.6 and 1464.9 ng/g, respectively. These results show that the gaseous Hg adsorption ability is different for these fly ashes. As mentioned above, the particle size distributions, the UBC contents and the specific surface areas play important roles in the adsorption of the gaseous Hg on the fly ash. The particle size distributions, the mean particle diameter, the UBC contents and the specific surface areas of these fly ash samples are shown in Figures 9-12, respectively. Figure 9 demonstrates that the proportion of coarser particles in these fly ash samples collected from the ESP is higher than that from the FFs. Figure 10 shows that their mean particle diameters tended to decline along the flow pass in the PCD system. Referring to the data in Figure 8, it is clear that the gaseous Hg adsorption capacity of the small fly ash particles is higher than that of the coarser ones, which is quite different from the fly ash collected from the PC boiler unit. Figure 11 shows that the UBC contents in the fly ash samples fluctuate from 6.9% to 7.4% for the ESP, 8.6% to 9.2% for the first FF and 10.9% to 13.1% for the second FF, respectively. It is illustrated that the UBC content increases with decreasing the particle size of the fly ash, which is similar to the results from 17 CFB boiler units [41]. As shown in Figure 12, the specific surface areas of these fly ash samples are in the range of 4.43 to 4.85 m 2 /g collected from the ESP, 5.55 to 6.06 m 2 /g collected from the first FF and 7.35 to 8.29 m 2 /g collected from the second FF, indicating that the specific surface area also increases with decreasing the particle size of the fly ash. The above results suggest that the gaseous Hg easily migrates onto the small-sized fly ash particles in the CFB boiler unit due to a high UBC content and a large specific surface area. It is well-known that the properties of the fly ash collected from CFB boiler units are quite different from those collected from PC boiler units. Our previous work [27] confirmed that the UBC content and the specific surface area of the CFB fly ash were higher than those of the PC fly ash. Chen et al. [40] found lots of microspheres in the PC fly ashes, but most of the particulates were porous and with irregular shape in the CFB fly ashes. In order to understand the migration of the gaseous Hg from the flue gas to the fly ash in the CFB boiler unit, the Hg contents in the fly ash samples collected from the ESP/FFs were measured. As shown in Figure 8, the Hg contents in the fly ash samples vary from 769.2 to 1045.0 ng/g collected from the ESP, 1217.2 to 1440.9 ng/g collected from the first FF and 1329.0 to 1605.0 ng/g collected from the second FF. The average Hg contents in the fly ash samples collected from the ESP, the first FF and the second FF were 912.3, 1313.6 and 1464.9 ng/g, respectively. These results show that the gaseous Hg adsorption ability is different for these fly ashes. As mentioned above, the particle size distributions, the UBC contents and the specific surface areas play important roles in the adsorption of the gaseous Hg on the fly ash. The particle size distributions, the mean particle diameter, the UBC contents and the specific surface areas of these fly ash samples are shown in Figures 9-12, respectively. Figure 9 demonstrates that the proportion of coarser particles in these fly ash samples collected from the ESP is higher than that from the FFs. Figure 10 shows that their mean particle diameters tended to decline along the flow pass in the PCD system. Referring to the data in Figure 8, it is clear that the gaseous Hg adsorption capacity of the small fly ash particles is higher than that of the coarser ones, which is quite different from the fly ash collected from the PC boiler unit. Figure 11 shows that the UBC contents in the fly ash samples fluctuate from 6.9% to 7.4% for the ESP, 8.6% to 9.2% for the first FF and 10.9% to 13.1% for the second FF, respectively. It is illustrated that the UBC content increases with decreasing the particle size of the fly ash, which is similar to the results from 17 CFB boiler units [41]. As shown in Figure 12, the specific surface areas of these fly ash samples are in the range of 4.43 to 4.85 m 2 /g collected from the ESP, 5.55 to 6.06 m 2 /g collected from the first FF and 7.35 to 8.29 m 2 /g collected from the second FF, indicating that the specific surface area also increases with decreasing the particle size of the fly ash. The above results suggest that the gaseous Hg easily migrates onto the small-sized fly ash particles in the CFB boiler unit due to a high UBC content and a large specific surface area.        Test 1  Test 2  Test 3  Test 4 Fly ash from ESP Fly ash from FF1 Fly ash from FF2      Test 1  Test 2  Test 3  Test 4 Fly ash from ESP Fly ash from FF1 Fly ash from FF2      Test 1  Test 2  Test 3  Test 4 Fly ash from ESP Fly ash from FF1 Fly ash from FF2   Test 1  Test 2  Test 3  Test 4 Fly ash from ESP Fly ash from FF1 Fly ash from FF2

Comparison of Hg Behavior between PC Boiler Unit and CFB Boiler Unit
CFB boilers and PC boilers are the two major types of boilers adopted in coal-fired power plants. Wang et al. [25] found that nearly 100% gaseous Hg was removed in a CFB boiler unit, compared to only 27.56% or 33.59% Hg removed in two PC boiler units. This result indicates that Hg behavior between CFB boiler units and PC boiler unit is different. As mentioned in Section 3.3, APH plays a key role in the migration of gaseous Hg due to the change of flue gas temperature. Our previous study investigated the effect of the temperature (50-300 °C) on the Hg adsorption ability of fly ashes collected from a PC boiler unit and a CFB boiler unit [35]. The results indicated that Hg adsorption by these two fly ashes increased with a decrease of the temperature and that Hg adsorption on the CFB fly ashes was much higher than that of the PC fly ashes at the same temperature. It suggests that the gaseous Hg migration across APH in the CFB boiler unit is stronger than that across APH in the PC boiler unit. Wang et al. [19] investigated the Hg emission and speciation of six PC boiler units in China. They pointed out that the average mercury removal efficiency of ESP was 24%. However, the results in our study show that more than 99% of Hg contained in the feed fuel was captured by the fly ash collected from the PCD system in the CFB unit. It is well-known that Hg adsorption ability is closely related to the properties of fly ashes. Due to the difference of combustion situation, the fly ashes collected from CFB boiler units are quite different from those collected from PC boiler units. Based on the above results, it is clear that the gaseous Hg easily migrated from the flue gas to the fly ash at downstream of the CFB boiler unit due to high UBC content and specific surface area.

Conclusions
The Hg emission and its migration behavior in the CFB boiler unit equipped with a tubular APH, an ESP and two FFs in series have been investigated, and the following results are obtained: (1) The Hg mass balance across the CFB boiler unit for four tests are between 100.38% and 112.88%. More than 99% of Hg contained in the feed fuel is captured by the fly ash, whilst less than 1% of Hg remained in the bottom ash or the flue gas at the outlet of the PCD system.
(2) The migration of the gaseous Hg from the flue gas to the fly ash is enhanced in the APH, where the flue gas temperature decreases from 250 to 120 °C, and the average Hg content in the fly ash samples increases from 1030.9 to 2018.6 ng/g.
(3) Continuous migration of Hg from the flue gas to the fly ash is observed across the PCD system. The gaseous Hg could further migrate to the fly ash in the subsequent FFs due to a long gassolid contact time.

Comparison of Hg Behavior between PC Boiler Unit and CFB Boiler Unit
CFB boilers and PC boilers are the two major types of boilers adopted in coal-fired power plants. Wang et al. [25] found that nearly 100% gaseous Hg was removed in a CFB boiler unit, compared to only 27.56% or 33.59% Hg removed in two PC boiler units. This result indicates that Hg behavior between CFB boiler units and PC boiler unit is different. As mentioned in Section 3.3, APH plays a key role in the migration of gaseous Hg due to the change of flue gas temperature. Our previous study investigated the effect of the temperature (50-300 • C) on the Hg adsorption ability of fly ashes collected from a PC boiler unit and a CFB boiler unit [35]. The results indicated that Hg adsorption by these two fly ashes increased with a decrease of the temperature and that Hg adsorption on the CFB fly ashes was much higher than that of the PC fly ashes at the same temperature. It suggests that the gaseous Hg migration across APH in the CFB boiler unit is stronger than that across APH in the PC boiler unit. Wang et al. [19] investigated the Hg emission and speciation of six PC boiler units in China. They pointed out that the average mercury removal efficiency of ESP was 24%. However, the results in our study show that more than 99% of Hg contained in the feed fuel was captured by the fly ash collected from the PCD system in the CFB unit. It is well-known that Hg adsorption ability is closely related to the properties of fly ashes. Due to the difference of combustion situation, the fly ashes collected from CFB boiler units are quite different from those collected from PC boiler units. Based on the above results, it is clear that the gaseous Hg easily migrated from the flue gas to the fly ash at downstream of the CFB boiler unit due to high UBC content and specific surface area.

Conclusions
The Hg emission and its migration behavior in the CFB boiler unit equipped with a tubular APH, an ESP and two FFs in series have been investigated, and the following results are obtained: (1) The Hg mass balance across the CFB boiler unit for four tests are between 100.38% and 112.88%.
More than 99% of Hg contained in the feed fuel is captured by the fly ash, whilst less than 1% of Hg remained in the bottom ash or the flue gas at the outlet of the PCD system. (2) The migration of the gaseous Hg from the flue gas to the fly ash is enhanced in the APH, where the flue gas temperature decreases from 250 to 120 • C, and the average Hg content in the fly ash samples increases from 1030.9 to 2018.6 ng/g.