Nitrogen Oxides, Sulfur Trioxide, and Mercury Emissions during Oxy

Nov 17, 2014 - Therefore, to control harmful Hg0, a mercury removal system may need to be ..... using a micro-gas chromatograph (micro-GC) (Agilent GC...
1 downloads 0 Views 1MB Size
Download PDF
Article pubs.acs.org/est

Nitrogen Oxides, Sulfur Trioxide, and Mercury Emissions during Oxyfuel Fluidized Bed Combustion of Victorian Brown Coal Bithi Roy, Luguang Chen, and Sankar Bhattacharya* Department of Chemical Engineering, Monash University, P.O. Box 36, Clayton Campus, Melbourne, Victoria 3800, Australia ABSTRACT: This study investigates, for the first time, the NOx, N2O, SO3, and Hg emissions from combustion of a Victorian brown coal in a 10 kWth fluidized bed unit under oxy-fuel combustion conditions. Compared to air combustion, lower NOx emissions and higher N2O formation were observed in the oxy-fuel atmosphere. These NOx reduction and N2O formations were further enhanced with steam in the combustion environment. The NOx concentration level in the flue gas was within the permissible limit in coal-fired power plants in Victoria. Therefore, an additional NOx removal system will not be required using this coal. In contrast, both SO3 and gaseous mercury concentrations were considerably higher under oxy-fuel combustion compared to that in the air combustion. Around 83% of total gaseous mercury released was Hg0, with the rest emitted as Hg2+. Therefore, to control harmful Hg0, a mercury removal system may need to be considered to avoid corrosion in the boiler and CO2 separation units during the oxy-fuel fluidized-bed combustion using this coal.

■

INTRODUCTION Oxy-fuel combustion is a developing technology for easier capture of the CO2 generated from coal-fired power plants.1 This technology, together with the well established fluidized bed combustion technology, offers several advantages, such as fuel flexibility, reduced NOx formation due to the low concentration of nitrogen in the oxy-fuel combustion environment, and potential for in-bed SO2 capture via the addition of sorbents.2,3 A combination of these two technologiesoxyfuel-based fluidized bed (Oxy-FB) combustionis an emerging technology for power generation using all grades of coals, including low-rank coals. Moreover, Oxy-FB combustion also offers several technical and economic advantages over oxy-fuel pulverized fuel (Oxy-PF) combustion.1 Some of the major advantages of Oxy-FB combustion over Oxy-PF combustion are the lower furnace temperature and lower leakage of air, which in turn lowers NOx emissions. However, the emission level of N2O, one of the major greenhouse gases, is also an important consideration because of its contribution to the destruction of the ozone layer in the stratosphere.4 In coal combustion, nitrogen is first released during devolatilization in the form of either HCN or NH3, which are then further oxidized to NOx and N2O during char combustion. The main reaction paths during devolatilization and char combustion of the coal−nitrogen to NO and N2O were provided by Winter.4 The main reactions involved in the NOx chemistry in a fluidized bed reactor were summarized by Lupianez et al.5 A number of experimental studies have been carried out to investigate the NOx emissions characteristics in Oxy-FB combustion.2,3,5−9 On the other hand, only a few experimental © XXXX American Chemical Society

data are available on the N2O emissions during Oxy-FB combustion.7 Moreover, the limited number of prior studies considered mainly bituminous coals. The data on N2O and NOx emissions from brown coals during oxy-fuel fluidized bed combustion do not exist. Moreover, in oxy-fuel combustion it is also important to know the extent of SO3 formation in flue gas as SO3 can corrode the air-heater and economizer surfaces.10,11 The formation of SO3 depends on several factors, such as O2 concentration, sulfur content of coal, temperature, residence time, fly ash composition, presence of catalysts (e.g., Fe2O3), and flue gas components (e.g., NOx and CO).10,12 Normally at high temperature, most of the organic-S is converted to SO2. During flue gas cooling a portion of this SO2 is converted to SO3, primarily via the oxidation of SO2 according to reaction 1, and secondarily via the formation of HOSO2 according to reactions 2 and 3 for temperatures below 1150 K.10,12 SO2 + O( +M) → SO3( +M)

(1)

SO2 + OH( +M) → HOSO2 (+ M)

(2)

HOSO2 + O2 → SO3 + HO2

(3)

In presence of steam in flue gas, this SO3 may be subsequently converted to gaseous sulfuric acid (H2SO4), according to Received: September 23, 2014 Revised: November 8, 2014 Accepted: November 17, 2014

A

dx.doi.org/10.1021/es504667r | Environ. Sci. Technol. XXXX, XXX, XXX−XXX

Environmental Science & Technology

Article

reaction 4 which can cause corrosion if it is allowed to condense at temperatures between 400 and 200 °C.12 SO3 + H 2O → H 2SO4

Victoria has an abundant resource of brown coal with an estimated reserve of 430 billion tonne which is expected to last for over 500 years at the current consumption rate.25 With the use of three Victorian brown coals, Loy Yang, Morwell, and Yallourn, in our previous study, mercury speciation under oxyfuel combustion was predicted by thermodynamic equilibrium calculation.23 With the use of one Victorian brown coal, Morwell, in two drop-tube furnaces, another study was performed to investigate the sulfur emissions at different operating temperatures and different residence times under oxy-fuel combustion.17 However, the fundamental data for SO3 and Hg emissions during oxy-fuel fluidized bed combustion using Victorian brown coal is nonexistent. Moreover, the N2O and NOx emissions data for Oxy-FB combustion using this huge reserve of coal is nonexistent. In this present work, experiments were conducted in a 10 kWth fluidized bed combustor under air and oxy-fuel combustion conditions using one Victorian brown coal, Loy Yang. This paper presents, for the first time, for Oxy-FB combustion of a Victorian brown coal, the emissions characteristics of NO, NO2, N2O, SO3, and Hg. The effects of oxygen and steam concentrations in the feed gas, and bed temperature on the emissions characteristics are discussed. In addition, the results from oxy-fuel combustion are also compared to those from air combustion.

(4)

It is known that the measurement of SO3 concentration in flue gas is difficult owing to the highly reactive nature of SO3. Several techniques have been used for the measurement of SO3, such as controlled condensation method, isopropyl alcohol drop method, isopropyl alcohol absorption bottle method, salt method, Pentol SO3 monitor (using Severn Science Analyzer), Fourier transform infrared (FTIR) spectroscopy, and mass spectrometer.13 Another method was developed by Ibanez et al.,14 who used calcium oxalate as a reagent which reacts only with SO3 (not with SO2) under certain conditions to produce a more stable and conveniently measured molecule. In the oxyfuel combustion atmosphere due to the low concentration of air-borne nitrogen and production of water vapor, higher levels of SO2 (and consequently SO3) was produced compared to that in air combustion.10,15,16 Therefore, in oxy-fuel combustion it is very important to know the level of SO3 formation. Hence, several researchers studied the fate and mechanism of SO3 formation under oxy-fuel combustion.10−13,15−18 However, studies on the extent of SO3 formation using brown coal or lignite in oxy-fuel fluidized bed combustion are limited. Furthermore, the fate of mercury (Hg) is also an important consideration on the operation of oxy-fuel combustors using a variety of coals as the presence of this element in the gas beyond the permissible limits is harmful to health and the environment, with additional implications for CO2 transport and storage. However, different forms of mercury determine the ultimate health effects and environmental impacts. It is wellknown that there are mainly three forms of mercury from coalfired boilers: gaseous elemental mercury (Hg0), gaseous oxidized mercury (Hg2+), and particulate bound mercury (HgP).19 At high temperature combustion zone, most of the coal-Hg is converted to Hg0. As the combustion flue gas cools at the downstream of the furnace, this Hg0 can be oxidized by the flue gas components, (such as HCl, SOx and NOx) to form Hg2+. This Hg2+ can react with the particulate in the flue gas to form HgP, which is adsorbed onto fly ash.19,20 Among these forms of mercury, gaseous elemental mercury (Hg0) is the most toxic form, and can stay longer in the environment and disperse long distances from the source of emission. In contrast, gaseous oxidized mercury (Hg2+) is easier to capture due to its high solubility in water. Therefore, it is essential to know the extent of different forms of gaseous mercury, especially the harmful Hg0, generated during combustion using different types of coals. A number of studies have been performed to investigate the mercury speciation during oxy-fuel combustion.11,20−24 To avoid corrosion in an Al alloy containing CO2 compression unit,24 Hg needs to be removed from the flue gas at the upstream section of the CO2 compressor. Owing to the recirculation of the flue gas containing Hg, a higher concentration of Hg was observed at the outlet of the burner during oxy-fuel combustion compared to that in air combustion.11 In the exhaust gas from a 90 kWth bubbling fluidized bed oxy-combustion pilot plant, most (82%) of the Hg was found to be present as Hg0 in comparison to Hg2+, owing to the high volatility of Hg0.22 However, using a selective catalytic reactor with TRAC catalyst, almost all the Hg0 was found to be converted into Hg2+.11

■

MATERIALS AND METHODS Feedstock. One air-dried Victorian brown coal, Loy Yang, was selected as the solid fuel. The dried coal sample was first crushed and sieved, and then the particles with a size range of 1−3 mm were used in the experiments. The composition of this coal is given in Table 1. Table 1. Composition of Coal Used in the Experiments ultimate analysis (wt % dry basis) carbon hydrogen nitrogen sulfur oxygen ash

65.00 4.60 0.72 0.50 25.48 3.70

chlorine

0.11

minerals and inorganic (wt % ash basis) SiO2 Al2O3 Fe2O3 TiO2 K2O MgO Na2O CaO SO3

56.90 20.64 4.63 1.51 1.31 3.63 4.73 1.61 5.04

trace element (mg/kg dry coal basis) Hg

0.06

As bed material, silica sand of 350−400 μm was used in the reactor. In addition, char (77.97% C, 2.42% H, 1.29% N, 0.11% S, 15.77% O, and 2.44% ash) of 1−3 mm was used during startup. Experimental Installation. The experiments were conducted in a ∼10 kWth fluidized bed unit which consists of a combustor, electric furnace, coal feeder, gas supply unit, gas preheating system, cyclone separators, particulate filter, cooler, water condensation unit, and gas analyzer. The inner diameter of the main reactor zone and freeboard zone of the reactor are 0.1 and 0.15 m, respectively, and the overall height of the reactor is 4 m. Prior to entering the reactor, all reacting gases, air/typical oxy-fuel gas mix in the reactor (O2, CO2 and/or steam), were mixed and preheated in a gas heater. To maintain the riser temperature during combustion, an electrically heated B

dx.doi.org/10.1021/es504667r | Environ. Sci. Technol. XXXX, XXX, XXX−XXX

Environmental Science & Technology

Article

(10% H2SO4−4% KMnO4). Gaseous oxidized mercury (Hg2+) was captured in KCl solution containing impingers, whereas gaseous elemental mercury (Hg0) was captured in HNO3− H2O2 and H2SO4−KMnO4 solutions containing impingers. After the experiment, the impinger solutions were well shaken before several 1 mL solutions were taken for analysis. These impinger solutions were analyzed for gaseous mercury concentration using a HACH DR-5000 UV-vis spectrophotometer with 0.03 μg/L Hg sensitivity. This UV spectrophotometer was calibrated beforehand with four different concentrations of HgCl2 standards. It is noted that during the experiment the temperature at the flue gas sampling point was around 200 ± 20 °C.

furnace was encapsulated with the bottom section of the reactor. The solids were fed to the combustor by a screw feeder (coupled with a variable-speed motor) located just below the furnace but above the distributor. Flue gases first passed through a primary and then a secondary cyclone to separate the solids before exhausting the clean flue gas to the atmosphere. A portion of the flue gases was sampled continuously through a particulate filter, cooler, and water condensation system, and then analyzed. NO and NO2 were measured using a gas analyzer (MX6 iBrid), while N2O was measured using a gas chromatograph (Agilent GC 7890A). General process data (such as total gas flow, gas composition, and temperature) were continuously recorded during the experiment. The steady state period was 1−3 h during which the measurements were taken. The experimental conditions are summarized in Table 2. The typical O2 concentration in oxy-

■

RESULTS AND DISCUSSION Results are presented first for evaluation of the effects of oxygen, carbon dioxide, and steam concentrations in the

Table 2. Operating Conditions Used in the Experiments combustion atmosphere (% volume) air 15% 21% 30% 15% 15% 15%

O2 O2 O2 O2 O2 O2

air 15% 21% 30% 15%

O2 O2 O2 O2

gas velocity (m/s)

Measurements of Nitrogen 0.82 + 85% CO2 0.76 + 79% CO2 0.81 + 70% CO2 0.81 + 85% CO2 0.74 + 85% CO2 0.79 + 73% CO2 + 12% H2O 0.75 Measurements of Sulfur Trioxide and 0.82 + 85% CO2 0.76 + 79% CO2 0.81 + 70% CO2 0.81 + 73% CO2 + 12% H2O 0.75

bed temperature (°C)

coal feed rate (g/h)

Oxides 840 ± 10 450−670 850 ± 10 670−750 845 ± 5 670−710 845 ± 10 600−630 820 ± 10 490−670 880 ± 10 800−830 850 ± 10 565−600 Gaseous Mercury 850 ± 10 510−515 850 ± 10 570−670 850 ± 10 670−735 850 ± 10 600−670 850 ± 10 600−640

Figure 1. Nitrogen oxides emissions at different combustion environments.

fuel combustion is 15−30% by volume,26 therefore this range of O2 concentration was chosen during these experiments. Moreover, the typical H2O concentration in the oxidant of oxy-fuel combustion is 4−15%.27 From the steady-state process model on Oxy-FB combustion using Aspen Plus simulation, 10−12% H2O was observed in the case of Victorian brown coal. Therefore, 12% H2O was chosen in these set of experiments. Determination of SO3. The majority of the generated gases, such as CO, CO2, SO2, and O2, were monitored continuously using a micro-gas chromatograph (micro-GC) (Agilent GC 490). SO3 was determined by the method developed by Ibanez et al., 14 and its operation procedure was detailed elsewhere.17 In brief, calcium oxalate mixed with glass beads was used in the SO3 trap, for which the temperature was controlled in between 300 °C and 375 °C. At this temperature range, calcium oxalate reacts only with SO3 to produce additional CO and CO2 , according to reaction 5. SO 3 concentration in the sampling gas was measured indirectly by measuring the CO using the micro-GC. SO3 + CaC2O4 → CaSO4 + CO2 + CO

Figure 2. Effect of bed temperature on nitrogen oxides emissions.

(5)

Determination of Hg. The sequential selective extraction method of Hg(g) measurement by Mitsui et al.11 was used in this study. A portion of the flue gas was extracted through a series of impingers in ice−water baths. The fractions obtained with corresponding sequential adsorbents were coded as follow: I1 and I2 (KCl, 1M), I3 (5% HNO3−10% H2O2), I4 and I5

Figure 3. SO3 emission at different combustion environments. C

dx.doi.org/10.1021/es504667r | Environ. Sci. Technol. XXXX, XXX, XXX−XXX

Environmental Science & Technology

Article

Figure 4. Hg emission at different combustion environments.

combustor, and combustor temperature on emissions of nitrogen species. It is followed by the SO3 and Hg emissions under different combustion conditions. All the emissions are expressed on a dry flue gas basis. In a dry oxy-fuel experiment, the reacting gas composition consisted of 21% v/v O2 and 79% v/v CO2, and in wet oxy-fuel experiment, the reacting gas consisted of 12% v/v H2O balanced by O2 and CO2 (as shown in Table 2). NOx emissions. The nitrogen oxides emissions using Loy Yang coal under different combustion atmospheres are compared in Figure 1. All these experiments were conducted at the same bed temperature of 850 °C. It can be seen that compared to that in air combustion, NOx emissions are lower (by 4%) in dry oxy-fuel combustion, in common with several literature sources.2,28,29 This lower amount of NOx formation in Oxy-FB combustion results from having zero nitrogen in the oxy-fuel combustion atmosphere.5 On the other hand, N2O formation is higher (by 24%) during oxy-fuel combustion compared to that in air combustion, also in line with the observation by other researchers.28 Yoshiie et al.28 also reported that by the recirculation of the flue gas, this N2O concentration further increased owing to the accumulation of additional N2O resulting from recirculation. The experimental results also indicate that the addition of steam (keeping oxygen concentration in oxidant same as 15% v/v) in the combustion environment promotes NOx reduction (by 5.5%) and N2O formation (by 10%). These findings are consistent with the observations by several researchers.6,7 Steam is known to influence the HCN formation by its reaction with char-N and simultaneously lowers the NOx formation by char-N and O2.7 However, HCN is also known to convert into N2O, according to reactions 6 and 7,30 and as a result N2O formation increases with steam. A maximum 200 ppmv NOx and 20 ppmv N2O were observed using 12% v/v steam in the oxy-fuel combustion atmosphere. HCN + O → NCO + H

(6)

NCO + NO → N2O + CO

(7)

Figure 5. Hg2+/Hgtotal ratio as functions of (a) NO, (b) NO2, and (c) N2O.

With an increase in oxygen concentration, temperature tends to increase, thereby promoting NOx conversion. In addition as the oxygen concentration increases, the gas velocity in the riser becomes lower and the residence time of the fuel particles in the combustor becomes longer, which promotes the fuel-N to NOx conversion3 They also concluded that excess oxygen is the main variable controlling the NOx formation in both air and oxy-fuel combustion. In our work, we observed the emission of N2O to increase slightly with the oxygen concentration, as shown in Figure 1. The N2O emissions increased by 12% from 59 mg/MJ to 66 mg/MJ, when the oxygen concentration was increased from 15% (v/v) to 30% (v/v) in the combustion atmosphere. This increase in N2O emissions with the increase in inlet oxygen

These experiments also showed clearly that the NOx emissions increased with the O2 concentrations in feed gas for both nitric oxide and nitrogen dioxide. The NOx emissions significantly increased by 74% from 350 mg/MJ to 610 mg/MJ, when the oxygen concentration was increased from 15% (v/v) to 30% (v/v) in the combustion atmosphere. This finding is consistent with the observations on oxy-fuel fluidized bed combustion by other researchers3,5 working with different coals. D

dx.doi.org/10.1021/es504667r | Environ. Sci. Technol. XXXX, XXX, XXX−XXX

Environmental Science & Technology

Article

In our study using Victorian brown coal in oxy-fuel fluidized bed combustion, a maximum 228 ppm of NO was observed, which is much lower than the level from oxy-fuel pulverized fuel combustion of other lignites. Using Canadian lignite having a coal-N level of 1% in a 0.3 MW test facility, Tan et al.29 observed 555 ppm of NO in an oxy-fuel pulverized fuel combustion, which was considerably higher than our study having a coal-N level of 0.72%. A maximum 0.68 g/m3 @ 4% O2 concentration of NOx was observed in the flue gas, which was within the allowable NOx emission limit of 0.78 g/m3 in coal-fired power plants in Victoria, Australia.33 Therefore, an additional NOx removal system (such as selective catalytic reduction or low-NOx burner) will not be required during oxyfuel fluidized bed combustion using the Loy Yang coal. SO3 Emissions. From Figure 3, it can be observed that compared to air combustion, SO3 emissions were lower in dry oxy-fuel combustion with similar oxygen concentrations in the combustion atmosphere. In terms of concentrations, however, the SO3 concentrations were higher (by 6.7%) during oxy-fuel combustion, in common with several literature sources.10,11,16 Fleig et al.10 concluded that because of the absence of airborne N2 in oxy-fuel combustion, SO2 concentration was higher, which in turn increased the SO3 concentration. On the other hand, steam promotes the SO3 formation by 7%. This increase in SO3 emission in the presence of steam was explained by Fleig et al.10,15 The presence of steam increases the amount of OH-radicals, according to reaction 13. Therefore, it promotes the formation of HOSO2, according to reaction 2, and then the formation of SO3, according to reaction 3.

Figure 6. Hg2+/Hgtotal ratio as a function of SO3.

concentration is due to the higher conversion of volatiles (such as HCN) to N2O, according to reactions 6 and 7.31 The effect of bed temperature on nitrogen oxides emissions is shown in Figure 2. All these experiments were conducted at the same oxygen concentration (15% v/v and balance CO2) in the feed gas. In fluidized bed combustion, the NOx emissions are known to increase with bed temperature.3,5,31 However, a reverse trend was observed for the temperature ranges used in this study. Several researchers32 also did not find any clear trend of NOx measurement with temperature during fluidized bed combustion attributing this to the fact that the NOx formation is mainly influenced by the oxygen partial pressure surrounding the burning particle and type of coal rather than the bed temperature. N2O emissions were found to decrease with the increase in bed temperature, supporting the observations in previous literature,31,32 which reported that the nitrous oxide emission is mainly influenced by operating temperature. The N2O emissions decreased by 26% from 82 mg/MJ to 60 mg/MJ, when the bed temperature was increased from 820 to 880 °C. The lower N2O emission at higher temperature is due to the decomposition of N2O, according to reactions 8 to 12.32 The increase in bed temperature increases the rates of these reactions and increases the concentration of H and OH radicals, important intermediates in the N2O decomposition. N2O + H → N2 + OH

(8)

N2O + OH → N2 + HO2

(9)

N2O + M → N2 + O + M

(10)

N2O + O → N2 + O2

(11)

N2O + O → 2NO

(12)

H 2O + O → OH + OH

(13)

These experiments showed clearly that sulfur trioxide emissions significantly increased with the oxygen concentrations in the feed gas. The SO3 emissions increased by 61% from 9 mg/MJ to 14.5 mg/MJ, when the oxygen concentration was increased from 15% (v/v) to 30% (v/v) in the combustion atmosphere. This finding is consistent with the observations during oxy-fuel combustion by many researchers.8−10,15 Duan et al.9 reported that with the increase in oxygen concentration in their experiments, the combustor temperature increased which promoted sulfur release in the gas phase. However, it should be noted that an increase of O2 concentration does not always mean an increase of the combustor temperature. Furthermore, it is noted that the measured SO3 concentration levels in the flue gas were close to the SO3 levels observed by other researchers18 with similar S level of 0.5% in our coal (Table 1). In our study, around 2.5 ppm of SO3 was observed at 30% oxygen concentration in the oxy-fuel combustion environment; this is slightly lower than the SO3 concentration level of 4 ppm presented in the study by

Table 3. Hg Mass Balance under Different Combustion Conditions combustion atmosphere (% volume) air 15% 21% 30% 15%

O2 O2 O2 O2

+ + + +

85% 79% 70% 73%

CO2 CO2 CO2 CO2 + 12% H2O

input (mg/h)

output (mg/h) 0

coal

Hg

0.0281 0.0308 0.0363 0.0329 0.0320

0.0136 0.0162 0.0219 0.0177 0.0212

Hg

E

2+

0.0061 0.0061 0.0083 0.0075 0.0043

balance Hg

P

0.0002 0.0007 0.0006 0.0004 0.0008

total

(%)

0.0199 0.0231 0.0308 0.0256 0.0263

70.89 74.78 84.88 78.01 82.25

dx.doi.org/10.1021/es504667r | Environ. Sci. Technol. XXXX, XXX, XXX−XXX

Environmental Science & Technology

Article

Fujimori and Yamada,18 presumably due to the higher oxidant level of 35% in their study. Hg Emissions. As shown in Figure 4, the experimental results showed that compared to the air combustion, Hg0 emissions were higher (by 28%) during oxy-fuel combustion with same oxygen (21% v/v) concentration in feed gas, in common with the literature.20 In our experiment, the total gaseous mercury concentration level of 20 μg/m3N was observed in air combustion, whereas 26 μg/m3N was observed in oxy-fuel combustion. About 69% of gaseous mercury was observed to emit as Hg0 in air combustion, while 73% of gaseous mercury emitted as Hg0 in oxy-fuel combustion. This is due to the lower NOx formation for having zero nitrogen in the oxy-fuel combustion atmosphere, since NOx is known to promote Hg0 oxidation.34 The addition of steam (even with same 15% v/v oxygen concentration in oxidant) in the combustion environment was found to promote (by 32.6%) the Hg0 emissions. This finding is consistent with the observations by Niksa et al.,35 who reported that H2O strongly inhibited the mercury oxidation. The higher amount of Hg0 in the presence of H2O is due to the production of HgO at higher temperature (>450 °C) initially, according to reaction 14 between the dominant mercury species at low temperature (