Fate of Hazardous Air Pollutants in Oxygen-Fired Coal Combustion

Energy & Environmental Research Center, University of North Dakota, 15 North 23rd Street, Stop 9018, Grand Forks, North Dakota 58202-9018, United Stat...
2 downloads 0 Views 4MB Size
Download PDF
Article pubs.acs.org/est

Fate of Hazardous Air Pollutants in Oxygen-Fired Coal Combustion with Different Flue Gas Recycling Ye Zhuang*,† and John H. Pavlish† †

Energy & Environmental Research Center, University of North Dakota, 15 North 23rd Street, Stop 9018, Grand Forks, North Dakota 58202-9018, United States S Supporting Information *

ABSTRACT: Experiments were performed to characterize transformation and speciation of hazardous air pollutants (HAPs), including SO2/SO3, NOx, HCl, particulate matter, mercury, and other trace elements in oxygen-firing bituminous coal with recirculation flue gas (RFG) from 1) an electrostatic precipitator outlet or 2) a wet scrubber outlet. The experimental results showed that oxycombustion with RFG generated a flue gas with less volume and containing HAPs at higher levels, while the actual emissions of HAPs per unit of energy produced were much less than that of air-blown combustion. NOx reduction was achieved in oxycombustion because of the elimination of nitrogen and the destruction of NO in the RFG. The elevated SO2/SO3 in flue gas improved sulfur self-retention. SO3 vapor could reach its dew point in the flue gas with high moisture, which limits the amount of SO3 vapor in flue gas and possibly induces material corrosion. Most nonvolatile trace elements were less enriched in fly ash in oxycombustion than air-firing because of lower oxycombustion temperatures occurring in the present study. Meanwhile, Hg and Se were found to be enriched on submicrometer fly ash at higher levels in oxy-firing than in air-blown combustion.

■

INTRODUCTION Oxyfuel combustion with recirculation flue gas (RFG) has been considered as one promising solution to carbon capture and sequestration,1 and extensive studies were conducted with much focus on combustion performance, i.e., ignition,2 flame temperature and propagation,3,4 and heat transfer.5 As for air pollutant, soot formation was suppressed in oxycombustion flame as a result of less fuel and more oxygen containing species such as OH in the region of high temperature.6 The formation of thermal NOx in oxycombustion is expected to be minimum because of the absence of N2 in the feed gas. Laboratory study of oxy-firing without RFG has shown that the conversion of fuel N to NO increased with elevated combustion temperature.5,7 Meanwhile, experimental data of oxycombustion with RFG indicated that the recycled NO was reduced to N2 through homo- and heterogeneous reactions with CxHyOz and its derivatives, whereas the extent of reduction is affected by combustion conditions.8−10 Since NOx formation highly depends on combustion conditions, the resulting NOx concentration in oxy-fuel combustion flue gas could be either higher or lower than air-combustion flue gas.9 Oxycombustion usually not only raises the level of CO2 but SO2 concentration as well in the combustion zone, which could change sulfur partitioning in flue gas. Croiset and Thambimuthu reported averaged 1600 ppm SO2 for oxycombustion compared to around 600 ppm SO2 in air-firing,11 and the © 2012 American Chemical Society

concentrated SO2 and O2 in the combustion zone also promote formation of SO3 through the following reactions SO2 + O → SO3

(1)

SO2 + OH → HOSO2

(2)

HOSO2 + O2 → SO3 + HO2

(3)

The increased CO concentration could enhance the formation of OH-radical that subsequently facilitates SO3 formation through Reaction 2.12 Sulfur retention by fly ash may be enhanced during oxycombustion through either direct sulfation or indirect calcination/sulfation or in combination, depending on the temperature, CO2 partial pressure, and SO2 partial pressure.13 However, the effect of CO2 concentration on calcium sulfation is not well understood.14 Transformation of trace elements (TEs) during coal combustion depends upon combustion temperature, gas composition, the forms of occurrence of the TEs in the coal, and the interactions among different elements both during combuston and postcombustion.15 TEs were found to be enriched in a fine fraction of fly ash through vaporization and subsequent condensation.16,17 The significantly changed Received: Revised: Accepted: Published: 4657

January 12, 2012 March 22, 2012 March 22, 2012 March 22, 2012 dx.doi.org/10.1021/es300143q | Environ. Sci. Technol. 2012, 46, 4657−4665

Environmental Science & Technology

Article

For comparison purposes, a baseline air-blown coal combustion test was also performed at the same fuel firing rate as that of oxycombustion test.

in-furnace gas compositions, such as elevated O2, CO2, and SO2 in oxycombustion, will affect the coal combustion process including devolatilization, ignition, char reaction, and pollutant formation, e.g., toxic TEs. Modeling results showed that the high CO2 concentration in bulk gas changed the CO to CO2 ratio within char, which could affect the vaporization and condensation of refractory oxides.18 Experimental data by Suriyawong et al.19 and Sheng et al.20 concluded the geometric mean diameter of the formed fine ash shifted to smaller size under a higher CO2/O2 environment in comparison to airfiring conditions, while the CO2/O2 combustion did not change the formation mechanism of submicrometer particles.21 However, only limited data of partitioning of trace elements during oxycombustion are available. A comprehensive understanding of the fate of various hazardous air pollutants (HAPs) during oxycombustion under different scenarios of RFG will facilitate the implementation of full-scale oxycombustion technology while meeting regulations on air pollutant emissions.

■

RESULTS AND DISCUSSION Coal Analysis. A bituminous coal was combusted during the test and analyzed for proximate/ultimate, major inorganic elements, and TEs of interest. The analysis results are listed in Table 1. Oxycombustion Performance. Table 2 lists combustion parameters for air-blown and oxygen-blown with RFG from the ESP outlet and the wet scrubber outlet, respectively. The coal feed rate was maintained at the same 24.5 kg/h, equivalent to 173.5 kWth, for all test conditions. The actual amount of O2 into the combustor was jointly determined by the feed pure O2 stream and residual oxygen in the RFG with the goal to attain 3%-5% excess O2 at the combustor exit. The equivalence ratio happened to be 0.93 for oxycombustion vs 0.83 for air-firing due to the system fluctuation, resulting in lower combustion temperature in oxycombustion than that of air-firing. Another reason for the lower combustion temperature can be ascribed to the higher specific heat of CO2 in oxycombustion. Average O2 concentration in the gas stream into the combustor during oxycombustion was 26.2% and 28.3% for RFG at the ESP outlet and the scrubber outlet, respectively, indicating that the amount of recycled flue gas feeding into the combustion was less than the corresponding nitrogen used in air-blown combustion. The resulting total volume of flue gas generated during oxycombustion was expected to be ∼75%− 80% of the total flue gas generated in the air-blown combustion test, which has been verified by the measured flue gas flow rate at the furnace exit as shown in Table 2. Higher LOI, as listed in Table 3, were observed for ash generated in oxycombustion than that of air-blown combustion, indicating insufficient burnout during oxycombustion, which may be ascribed to the lower O2 diffusivity and less efficient O2 mixing in a CO2-enriched gas stream. Subsequently, CO concentrations in flue gas were 78.1−56.6 ppm during oxycombustion, higher than 17.7 ppm in air-blown mode. A similar high CO spike was also reported by others in oxycombustion condition22 and partially contributed to the enhanced CO formation through NO/char reaction under an elevated CO2 environment.23 Achieving high levels of CO2 in the postcombustion flue gas is the main focus of the oxy-blown combustion, while the concentration of CO2 is affected by air infiltration of the combustion system. The highest CO2 attained at the furnace exit during oxycombustion was ∼70% and further decreased to 65%−68% at the stack, which indicates that the combustor has an approximate 8.5% air leakage through the entire combustion system. Further oxygen measurement along the system indicated that ∼6% air infiltration was from the combustor, while the rest of the air leakage occurred between the convection duct and stack. As the result of the low volumetric flue gas generated, the moisture content and dust loading during oxycombustion were higher than that in air-blown combustion. Emission Characteristics of HAPs. NOx/SOx Emissions. Table 3 lists the NOx and SO2 distributions and the actual stack emissions under the three testing conditions. Even with the elimination of atmospheric N2, 732−758 mg/dNm3 of NOx were measured in oxycombustion flue gas, consistently higher than that in air-blown flue gas, mainly due to the accumulating

■

EXPERIMENTAL FACILITY AND METHOD In order to characterize the transformation of various HAPs during oxygen−coal combustion and understand the impacts of different RFG scenarios on these HAPs, experiments were conducted with a pulverized coal (pc)-fired combustor followed by an electrostatic precipitator (ESP) for particulate matter (PM) control and a wet scrubber for SO2 capture. As shown in Figure 1, during the oxygen-blown test, part of the

Figure 1. Experimental schematic diagram (M29 stands for U.S. Environmental Protection Agency [EPA] Method 29; FGD stands for flue gas desulfurization).

postcombustion flue gas was recycled and mixed with oxygen stream to maintain flame temperature and stability in coal combustion. Two different scenarios of RFG were evaluated: flue gas was recycled either from 1) the ESP outlet or 2) the wet scrubber outlet, separately, to investigate the corresponding effect on the transformation of HAPs. Flue gas was drawn from the furnace exit, the ESP outlet, and the scrubber outlet; filtered; and analyzed for O2, CO, CO2, NOx, and SO2. Flue gas samples were collected at the ESP inlet, the ESP outlet, and the scrubber outlet for analyses of SO3, H2O, PM, HCl, mercury, and other TEs. 4658

dx.doi.org/10.1021/es300143q | Environ. Sci. Technol. 2012, 46, 4657−4665

Environmental Science & Technology

Article

Table 1. Analysis Results of Test Coal and Ash parameter

value

proximate analysis, wt% moisture volatile matter fixed carbon ash ultimate analysis, wt% hydrogen carbon nitrogen sulfur oxygen ash heating value, kJ/kg Major Inorganic Elements in the Ashed Coal and Fly Ash

6.7 33.52 47.51 12.27 5.32 62.85 1.34 3.63 14.59 12.27 25,475

element, wt%

ashed coal

fly ash in air combustion

Si Al Fe Ti P Ca Mg Na K

45.86 23.02 21.95 0.95 0.19 2.84 1.18 0.27 2.98

44.16 18.89 26.14 1.18 0.14 3.94 0.90 0.44 3.43

Sb

As

Be

Cd

Cr

Co

Mn

Hg

Ni

Se

Pb

Cl

0.26

3.17

0.95

0.29

14.9

5.69

33.0

0.08

22.2

1.26

7.62

137

fly ash in oxycombustion with RFG at ESP outlet

43.60 18.36 27.04 1.16 0.14 3.98 0.86 0.27 3.39 Trace Elements of Tested Bituminous Coal, μg/g, dry

fuel feed rate, kg/h thermal power, kW feed O2, vol.%, dry feed N2, vol.%, dry feed CO2, vol.%, dry feed others combustion temperature, K equivalence ratio flue gas flow rate at furnace exit, Nm3/ min flue gas O2 at furnace exit, vol.%, dry flue gas CO2 at furnace exit, vol.%, dry flue gas CO at furnace exit, ppmv, dry flue gas H2O at furnace exit, vol.% PM at furnace exit, g/Nm3 LOIa in ash, % a

air-blown

oxygen-blown with RFG at ESP outlet

oxygen-blown with RFG at wet scrubber outlet

24.5 173.5 21 79 0 − 1412 ± 4

24.5 173.5 26.2 5.3 68.5 SO2, NOx 1348 ± 6

24.5 173.5 28.3 7.2 64.5 SO2, NOx 1387 ± 6

0.83 3.34 ± 0.03

0.93 2.55 ± 0.11

0.93 2.35 ± 0.11

3.7 ± 0.3

4.9 ± 0.4

5.1 ± 0.5

14.7 ± 2.5

71.4 ± 1.2

72.8 ± 2.7

17.7 ± 1.8

78.1 ± 26.4

56.6 ± 12.6

8.5 ± 0.2

18 ± 1.4

15 ± 0.8

8.46

13.9

13.3

0.74

2.24

1.38

43.29 18.14 27.08 1.16 0.15 3.97 0.84 0.28 3.36

effect from RFG and the reduced gas volume generated during the oxycombustion. However, the actual NOx stack emissions were in the range of 17.2 to 30.1 mg/MJ during oxyfuel combustion with two different RFG scenarios, and ∼90% NOx reduction was achieved compared to the 167.7 mg/MJ NOx emission from air-blown combustion. Similar up to 75% NOx reduction during oxycombsution was reported by others,5,9 while the main reason was ascribed to enhanced NO destruction rate as NO was recirculated to the combustion even though fuel NOx was slightly favored in oxycombustion. The absence of atmospheric N2 enabled the reversed Zeldovich mechanism to destroy NO. The relatively low oxycombustion temperature may also contribute to the low NO formation.8−10 The high-sulfur coal generated ∼9407 mg/dNm3 SO2 in conventional air-blown combustion. The packed-tower wet scrubber reduced SO2 concentration in the stack exhaust to 482.7 mg/dNm3, equivalent to 146.3 mg/MJ. As the combustor switched to oxy-blown mode with RFG at the ESP outlet, the SO2 concentration at the furnace exit was >18,186 mg/dNm3, while the emission of SO2 from the wet scrubber was kept as low as 7.7 mg/dNm3, equivalent to 9.4 × 10−2 mg/MJ, due to the high liquid−gas ratio of the wet scrubber since only ∼12% of the flue gas generated flowed through the wet scrubber. The pH value of the wet scrubber slurry was ∼5.6 in oxycombustion compared to 5.3 obtained in air-blown combustion, which would result in elevated residual limestone concentration because of a highly sensitive limestone dissolution rate above 5.5 in the presence of high levels of CO2,24 thereby higher

Table 2. Combustion Performance under Air- and OxygenBlown Conditions combustion parameter

fly ash in oxycombustion with RFG at wet scrubber outlet

Loss of ignition. 4659

dx.doi.org/10.1021/es300143q | Environ. Sci. Technol. 2012, 46, 4657−4665

Environmental Science & Technology

Article

Table 3. Comparison of NOx, SO2, and HCl Concentrations in Air- and Oxy-Firing Conditions NOx oxy-wet scrubber RFG

air-blown

559 ± 65 539 ± 66

758 ± 124 857 ± 91

732 ± 122 779 ± 80

573 ± 70

719 ± 80

167.7 ± 21.5

17.2 ± 2.1

air-blown ESP in, mg/dNm3 ESP out, mg/ dNm3 scrubber out, mg/ dNm3 stack emission, mg/MJ

SO2

oxy-ESP RFG

HCl

oxy-ESP RFG

oxy-wet scrubber RFG

airblown

oxy-ESP RFG

oxy-wet scrubber RFG

9407 ± 63 8879 ± 81

18186 ± 980 12979 ± 670

11294 ± 1050 10906 ± 931

− 10.5

16.7 17.9

17.6 18.8

692 ± 134

482.7 ± 15

7.7 ± 1.3

146.6 ± 85

−

−

−

30.1 ± 5.6

146.3 ± 4.3

9.4 × 10−2 ± 0.01

6.4 ± 3.8

−

−

−

case of oxy-firing with RFG at the ESP outlet, the calculated equilibrium SO3 vapor was 75.5 mg/dNm3 in comparison with the measured 52.2 mg/dNm3 SO3 in actual flue gas. Considering the inherent variability of the flue gas, flue gas may approach a condition where SO3 dew point condensation could occur, which limits the amount of SO3 vapor in flue gas. Therefore, acid gas corrosion on plant equipment needs to be addressed for oxycombustion. The ratio of SO 3 −SO x was 0.3% during air-blown combustion and increased to 0.4%−0.46% when switched to oxycombustion, which is much lower than the 5%−6% reported by Tan et al.,5 while the reason can be ascribed to the 35% O2 presence in their experiment compared to the 26%−28% in the present study. Sulfur retention efficiency, ηSR, is defined by the percent of the total sulfur in coal retained by ash after combustion

degrees of desulfurization. For the test where the RFG from the scrubber outlet, not much SO2 enrichment was observed in the flue gas as that in the oxycombustion with RFG at the ESP outlet since most of the SO2 had already been captured with the wet scrubber prior to recycling. The average SO2 concentration in flue gas was 10,906−∼11,294 mg/dNm3 prior to the wet scrubber and decreased to ∼146 mg/dNm3 at the scrubber outlet, equal to 6.4 mg/MJ. Again, the improved performance of the wet scrubber was because the total gas volume across the wet scrubber was 2.35 N m3/min, less than the 3.34 N m3/min during air-blown combustion. SO3 condensation sampling was conducted at the ESP outlet for the three test conditions. The SO3 data are plotted in Figure 2.

S ·A ηSR = ash × 100% Scoal

(4)

where Sash and Scoal are the weight contents of sulfur in the ash and parent coal, respectively, and A is the ash content of the coal. The calculated sulfur retention is plotted as a function of SO2 concentration in flue gas shown in Figure 3. Due to low Ca/S

Figure 2. SO3 concentrations at the ESP outlet during air- and oxygenblown combustion.

It appears that SO3 concentrations during oxycombustion tests were nearly doubled compared to air-blown combustion, and there was no significant difference on SO3 concentrations between the two scenarios of RFG. The enhanced formation of SO3 was mainly because of the elevated SO2 concentration in flue gas, while higher levels of H2O during oxycombustion may also facilitate SO3 formation.25 Also included in Figure 2 are the equilibrium SO3 vapor concentrations at the ESP outlet, calculated based on the empirical correlation as a function of temperature and moisture content of the flue gas.26 While maintaining a constant flue gas temperature of 425 K, due to the RFG, the moisture content of the flue gas at the ESP outlet raised from 8.5% in air-firing mode to 18.6% and 14.1% in oxycombustion with RFG at the ESP outlet and scrubber outlet, respectively, causing the corresponding equilibrium SO3 vapor concentration to decrease as shown in Figure 2. For the

Figure 3. Comparison of sulfur retention between air- and oxy-firing conditions.

molar ratio of 0.06 of the testing coal, the sulfur self-retention efficiency during air-blown combustion is 4.7%, which agrees with the data in literature.27 The sulfur retention efficiency increased as the coal combustion switched to oxy-firing mode. Since the molar ratio of Ca/S in the parent coal is unchanged during the tests, the observed increase in sulfur retention with ash during oxycombustion can be ascribed to the increased SO2/ SO3 concentrations in flue gas. The elevated concentrations of 4660

dx.doi.org/10.1021/es300143q | Environ. Sci. Technol. 2012, 46, 4657−4665

Environmental Science & Technology

Article

the SO2/SO3 in the combustion zone could accelerate the rates of reaction between SO2/SO3 and alkali/alkaline ash. Moreover, the higher SO2 concentration could stabilize the formed CaSO4 and prevent its dissociation that could take place at temperatures above 1123 K. Hence the self-desulfurization efficiency of oxy− coal combustion is higher compared to air-firing. HCl Emission. HCl sampling using EPA Method 26 was conducted at the inlet and outlet of the ESP during the tests, and the results in Table 3 show HCl enrichment during oxycombustion. Meanwhile, analysis on resulting fly ash showed no detectable chloride on the collected ash from both air-blown and oxy-blown combustions. Assuming all HCl in the recycled flue gas was removed either by the flue gas direct cooling system or by the wet scrubber for oxycombustion with RFG, the reason for the enriched HCl in flue gas during oxycombustion was most likely because of the concentrated effect of the reduced gas volume during oxycombustion. However, additional HCl data in flue gas at the scrubber outlet will be helpful to fully understand the fate of HCl in the combustion system. Mercury and Other TEs Emission. One EPA Method 29 sample was collected for each test condition at the inlet and outlet of the ESP and the scrubber outlet, respectively, to establish mercury partitioning between total gas-phase mercury vs particulate-bound mercury. Speciated sorbent trap samples were also collected at the ESP outlet and scrubber outlet under the same three test conditions, to verify Method 29 results and

to provide gas-phase mercury speciation data. A summary of the mercury data is plotted in Figure 4. The flue gas from air-blown combustion contained a total mercury value of 11.4 μg/dNm3 with 97% gas-phase mercury, which matched the theoretical expectation based on mercury content in coal and proximate/ultimate data. There was virtually no mercury capture across the ESP since most of the mercury was in the gas phase. Both EPA Method 29 and the sorbent trap data indicated that mercury reduction, in the range of 40% to 50%, occurred through the wet scrubber because ∼52% gaseous mercury was in the soluble oxidized form, indicated by the sorbent trap data. Mercury speciation data established during the conventional air-blown combustion are consistent with mercury emission data published elsewhere.28 As for oxy-blown combustion with RFG, mercury concentrations in the flue gas at the ESP inlet were concentrated to ∼20 μg/dNm3, almost double compared to total mercury data in the air-blown combustion test, which can be ascribed to the accumulative effect of flue gas recycling and the reduced gas volume during oxycombustion. Under oxycombustion conditions, 3.0−4.4 μg/dNm3 particulate associate mercury was present in the flue gas at the ESP inlet and was being removed by the ESP, resulting in ∼14.5% mercury reduction across the ESP. Moreover, higher percentages of oxidized mercury vapor were observed in the ESP outlet flue gas under oxycombustion compared to that in air-firing mode: 81.7% and 63.7% of the mercury vapor escaping from the ESP was in oxidized form for oxycombustion with RFG at the ESP outlet and wet scrubber outlet, respectively. The enhanced Hg(g)-to-Hg(p) conversion and Hg0-to-Hg2+ oxidation in oxycombustion were most likely a result of the continuous interactions of mercury with enriched HCl, unburned carbon (higher LOI during oxycombustion in Table 3), and other reagents in the flue gas.20 As a result, the overall mercury capture across the oxycombustion system was 80.9%, much higher than the 40.4% mercury removal by the airblown combustion system. The partitioning of the total eleven toxic metals in flue gas from air- and oxy-firing condition is summarized in Table 4. Similar to Hg, the other 10 TEs were also enriched in the oxycombustion flue gas in comparison to the corresponding concentrations in air-firing flue gas. Considering the fact of the same coal firing rate under three testing conditions, the measured elevated concentrations of the TEs are mainly the result of higher dust loadings in flue gas because of the

Figure 4. Mercury species partitioning during air- and oxygen-blown combustion.

Table 4. Comparison of Toxic Metals in the Fly Ash between Air- and Oxy-Firing Conditions air-firing, μg/dNm3 at 3% O2 Sb As Be Cd Cr Co Pb Mn Hg Ni Se a

oxy-firing with RFG at ESP out, μg/dNm3 at 3% O2

oxy-firing with RFG at scrubber out, μg/dNm3 at 3% O2

vapor

particulate

vapor

particulate

(EFoxy)/(EFair)

vapor

particulate

(EFoxy)/(EFair)

NDa ND ND ND 1.40 ND 0.75 15.2 11.0 5.46 114.08

36.6 365.8 57.3 125.4 1289.1 326.6 407.6 2177.6 0.31 1498.2 106.2

ND ND ND ND 23.42 ND ND 30.16 16.99 25.9 179.17

61.1 627.6 97.9 226.2 3203.8 561.5 693.6 4128.7 3.04 3550.6 194.8

0.91 0.93 0.93 0.98 1.35 0.93 0.92 1.03 5.3 1.29 1.10

ND ND ND ND 10.9 ND 0.94 2.92 15.2 10.7 112.2

59.1 675.8 60.6 224.7 2415.9 591.3 675.8 4341.9 4.40 2804.5 195.9

0.87 0.99 0.57 0.96 1.01 0.97 0.89 1.07 7.52 1.01 1.1

Nondetectable. 4661

dx.doi.org/10.1021/es300143q | Environ. Sci. Technol. 2012, 46, 4657−4665

Environmental Science & Technology

Article

Se concentration increased to 373.97 μg/dNm3, with 48/54 of vapor vs particulate. Se concentration further changed to 308.1 μg/dNm3, and partitioning between vapor and particulate also changed to 36/64 as the RFG switched to the scrubber outlet. The experimental data show that oxycombustion favors retention of volatile elements, i.e., Hg and Se, with fly ash. The reason for the change of Se behavior under the two oxy-firing tests was that part of the soluble Se vapor was scavenged by the scrubber prior to recirculating into combustion, which did not take place for the test when RFG from the ESP outlet.

reduced gas volume in oxycombustion. Meanwhile, any residual TEs contained in the recycling would also contribute to the enrichment. The majorities of the toxic elements, including Sb, As, Be, Cd, Cr, Co, Pb, Mn, and Ni, were associated with fly ash under air-firing and oxy-firing conditions. In addition to Hg, Se is another volatile element that has shown a fair amount of vapor species in coal combustion flue gas. Under air-firing conditions, the concentration of Se in flue gas was 220.28 μg/dNm3, with a 52/48 split between vapor and particulate phases. When switching to oxycombustion and RFG from the ESP outlet, the

Figure 5. continued 4662

dx.doi.org/10.1021/es300143q | Environ. Sci. Technol. 2012, 46, 4657−4665

Environmental Science & Technology

Article

Figure 5. Comparison of TE size distributions under air combustion and oxycombustion.

scenarios. The main reason can be ascribed to the relatively lower temperatures of oxycombustion as compared to the air firing temperature (as shown in Table 2). Because of the low combustion temperature, these elements might not be mobilized and associated with refractory elements within bottom ash.29 For those relatively volatile elements of Cr, Ni, Se, and, possibly, Mn, there was slightly higher TE enrichment with fly ash in oxycombustion than that of air-firing. The reason is that 1) the oxycombustion temperature was high enough to attain at least the same levels of vaporization of these elements as that in aircombustion and 2) the elevated reducing condition induced by the O2−CO2 mixture in combustion might favor vaporization of these elements.21 Hg, the most volatile element, clearly showed higher retention by ash in oxy-firing, which was

In order to eliminate the variability in different ash, the normalized enrichment factor for the TEs is calculated as follows EF =

(Ci /CAl )flyash (Ci /CAl )coalash

(5)

where Ci is the concentration of element i in fly ash and coal ash, respectively. CAl is the concentration of alumina in fly ash and coal ash, respectively. The enrichment factors of TEs in oxycombustion relative to that in air-firing are also included in Table 4, indicating that elements of Sb, As, Be, Cd, Co, and Pb, were slightly depleted within fly ash from oxycombustion compared to that of air-firing, and there is no significant difference on fly ash enrichment between the two RFG 4663

dx.doi.org/10.1021/es300143q | Environ. Sci. Technol. 2012, 46, 4657−4665

Environmental Science & Technology mainly caused by higher HCl in flue gas and higher LOI in ash that facilitate mercury adsorption on fly ash. TEs enriched in fine fly ash cause environmental and human health risks. Therefore, size-segregated fly ashes were collected using the multicyclone to establish TE partitioning as a function of fly ash size as shown in Figure 5. The eleven elements can be divided into three groups according to their size distributions within fly ash between air combustion and oxycombustion conditions. Group I includes Sb, As, Be, Cd, and Pb, which showed inversely proportional to fly ash diameter under air-firing. When switched to oxycombustion, Group I elements consistently showed lower concentrations on fly ash than with air-firing because the lower oxycombustion temperature caused less amount of Group I elements mobilized during combustion but rather associated with refractory elements in bottom ash. Meanwhile, Group I elements showed a local peak at fly ash diameter of 1−2 μm, implying that the fine fragmentation mechanism proposed by Linak30 and Seames 17 may be favored under oxycombustion conditions. Note that the location of peak concentration shifted to a larger particle size as the RFG was changed from the ESP outlet to the wet scrubber outlet, while the reason was not clear at this stage. Group II elements, including Co, Mn, Cr, and Ni, showed the highest concentrations at particle sizes of 0.6−1.5 μm under air-blown combustion. For oxycombustion with RFG from the ESP outlet, Group II elements maintained the same pattern of size distribution but with slightly lower concentrations than that in air-firing, indicating some degree of suppression under a high O2/CO2 combustion zone. However, as the RFG took place at the scrubber outlet, a bimodal distribution was observed for each of four elements. Part of the reason might be ascribed to the fact that some of these elements were found leaching from the stainless steel pipe into the scrubber slurry, while the residual slurry carryover was recirculated into the combustion and reformed into submicrometer particles. Volatile Se and Hg were categorized as Group III, and they both showed enrichment in submicrometer fly ash with some degree of depletion in supermicrometer particles under oxycombustion scenarios compared to that in air-firing. The observation was in agreement with Method 29 results and was mainly because 1) the oxycombustion temperature was high enough to attain at least the same levels of vaporization of these elements as that in air combustion and 2) the elevated reducing condition induced by the O2−CO2 mixture in combustion might favor vaporization of these elements.

■

ACKNOWLEDGMENTS

■

REFERENCES

This paper was prepared with the support of the Center for Air Toxic Metals under Award No. DE-FC26-08NT43291 of the U.S. Department of Energy.

(1) Buhre, B. J. P.; Elliott, L. K.; Sheng, C. D.; Gupta, R. P.; Wall, T. F. Oxy-Fuel Combustion Technology for Coal-Fired Power Generation. Prog. Energy Combust. Sci. 2005, 31 (4), 283−307. (2) Molina, A. S. C. Ignition and Devolatilization of Pulverized Bituminous Coal Particles During Oxygen/Carbon Dioxide Coal Combustion. Proc. Combust. Inst. 2007, 31, 1905−1912. (3) Chen, R.; Axelbaum, R. L. Scalar Dissipation Rate at Extinction and the Effects of Oxygen-Enriched Combustion. Combust. Flame 2005, 142 (1/2), 62−71. (4) Yamamoto, A. S. T.; Okazaki, K. Mechanism of Ignition Delay in O2/CO2 Pulverized Coal Combustion. In Proceedings of the Annual International Pittsburgh Coal Conference; Pittsburgh, PA, 2004. (5) Tan, Y; Croiset, E.; Douglas, M. A.; Thambimuthu, K. V. Combustion Characteristics of Coal in a Mixture of Oxygen and Recycled Flue Gas. Fuel 2006, 85, 507−512. (6) Du, J.; Axelbaum, R. L. The Effect of Flame Structure on SootParticle Inception in Diffusion Flames. Combust. Flame 1995, 100 (3), 367−375. (7) Liu, H.; Zailani, R.; Gibbs, B. M. Comparisons of Pulverized Coal Combustion in Air and in Mixtures of O2/CO2. Fuel 2005, 84, 833. (8) Okazaki, K.; Ando, T. NOx Reduction Mechanism in Coal Combustion with Recycled CO2. Energy 1997, 23 (2/3), 207−215. (9) Andersson, K.; Normann, F.; Johnsson, F.; Leckner, B. NO Emission During Oxy-Fuel Combustion of Lignite. Ind. Eng. Chem. Res. 2008, 47, 1835−1845. (10) Hu, Y. Q.; Kobayashi, N.; Hasatani, M. The Reduction of Recycled-NOx in Coal Combustion with O2/Recycled Flue Gas under Low Recycling Ratio. Fuel 2001, 80 (13). (11) Croiset, E.; Thambimuthu, K. V. NOx and SO2 Emissions from O2/CO2 Recycle Coal Combustion. Fuel 2001, 80 (14). (12) Glarborg, P.; Bentzen, L. Chemical Effects of High CO2 Concentration in Oxy-Fuel Combustion of Methane. Energy Fuels 2008, 22, 291. (13) Duan, L. D.; Zhou, W.; Li, H.; Chen, X.; Zhao, C. Sulfur Fate During Bituminous Coal Combustion in an Oxy-Fired Circulating Fluidized Bed Combustor. Korean J. Chem. Eng. 2011, 28 (9), 1952− 1955. (14) Garcia-Labiano, F.; Rufas, A.; de Diego, L. F.; de las ObrasLoscertales, M.; Gayan, P.; Abad, A.; Adanez, J. Calcium-Based Sorbents Behaviours During Sulphation at Oxy-Fuel Fluidised Bed Combustion Conditions. Fuel 2011, 90, 3100−3108. (15) Senior, C.; Helble, J. J.; Sarofim, A. F. Emissions of Mercury, Trace Elements, and Fine Particles from Stationary Combustion Sources. Fuel Process. Technol. 2000, 65−66, 263−288. (16) Helble, J. J.; Mojtahedi, W.; Lyyränen, J.; Jokiniemi, J.; Kauppinen, E. Trace Element Partitioning During Coal Gasification. Fuel 1996, 75 (8), 931−939. (17) Seames, W. S. An Initial Study of the Fine Fragmentation Fly Ash Particle Mode Generated During Pulverized Coal Combustion. Fuel Process. Technol. 2003, 81, 109−125. (18) Krishnamoorthy, G.; Veranth, J. M. Computational Modeling of CO/CO2 Ratio Inside Single Char Particles During Pulverized Coal Combustion. Energy Fuels 2003, 17, 1367−1371. (19) Suriyawong, A.; Gamble, M.; Lee, M.-H.; Axelbaum, R.; Biswas, P. Submicrometer Particle Formation and Mercury Speciation Under Oxygen-Carbon Dioxide Coal Combustion. Energy Fuels 2006, 20, 2357−2363. (20) Sheng, C.; Li, Y. Experimental Study of Ash Formation During Pulverized Coal Combustion in O2/CO2 Mixtures. Fuel 2008, 87, 1297−1305.

ASSOCIATED CONTENT

* Supporting Information S

Additional details. This material is available free of charge via the Internet at http://pubs.acs.org.

■

■

Article

AUTHOR INFORMATION

Corresponding Author

*Phone: (701) 777-5236. Fax (701) 777-5181. E-mail: [email protected]. Notes

This report was prepared as an account of work sponsored by the U.S. Department of Energy. The views and opinions of authors expressed herein do not necessarily state or reflect those of the United States Government or any agency thereof. The authors declare no competing financial interest. 4664

dx.doi.org/10.1021/es300143q | Environ. Sci. Technol. 2012, 46, 4657−4665

Environmental Science & Technology

Article

(21) Sheng, C.; Li, Y.; Liu, X.; Yao, H.; Xu, M. Ash Particle Formation During O2/CO2 Combustion of Pulverized Coal. Fuel Process. Technol. 2007, 88, 1021−1028. (22) Hu, Y. N.; Naito, S.; Kobayashi, N.; Hasatani, M. CO2, NOx and SO2 Emissions from the Combustion of Coal with High Oxygen Concentration Gases. Fuel 2000, 79, 15. (23) Li, Y. H.; Radovic, L. R.; Lu, G. Q.; Rudolph, V. A New Kinetic Model for the NO−Carbon Reaction. Chem. Eng. Sci. 1999, 54, 4125− 4136. (24) Chan, P.; Rochelle, G. T. Limestone Dissolution: Effect of pH, CO2 and Buffers Modeled by Mass Transfer. ACS Symp. Ser. 1982, 188, 75−97. (25) Fleig, D.; Normann, F.; Andersson, K.; Johnsson, F.; Leckner, B. The Fate of Sulphur During Oxy-Fuel Combustion of Lignite. Energy Procedia 2009, 1, 383−390. (26) Banchero, J. T.; Verhoff, F. H. Evaluation and Interpretation of the Vapor Pressure Data for Sulphuric Acid Aqueous Solutions with Application to Flue Gas Dewpoints. Inst. Fuel 1975, 76−80. (27) Sheng, C.; Xu, M.; Zhang, J.; Xu, Y. Comparison of Sulphur Retention by Coal Ash in Different Types of Combustors. Fuel Process. Technol. 2000, 64, 1−11. (28) Zhuang, Y. P.; Martin, C.; Pavlish, J.; Botha, F. Cobenefit of SO3 Reduction on Mercury Capture with Activated Carbon in Coal Flue Gas. Fuel 2011, 90, 2998−3006. (29) Seames, W. S.; Wendt, J. O. L. Partitioning of Arsenic, Selenium, and Cadmium During the Combustion of Pittsburgh and Illinois #6 Coals in a Self-Sustained Combustor. Fuel Process. Technol. 2000, 63 (2−3), 179−196. (30) Linak, W. P.; Miller, C. A.; Wendt, J. O. L. Comparison of Particle Size Distributions and Elemental Partitioning from the Combustion of Pulverized Coal and Residual Fuel Oil. J. Air Waste Manage. Assoc. 2000, 50, 1532−1544.

4665

dx.doi.org/10.1021/es300143q | Environ. Sci. Technol. 2012, 46, 4657−4665