Article pubs.acs.org/est
Ash Particulate Formation from Pulverized Coal under Oxy-Fuel Combustion Conditions Yunlu Jia and JoAnn S. Lighty* Department of Chemical Engineering, University of Utah, 50 South Central Campus Drive, Salt Lake City, Utah 84112, United States S Supporting Information *
ABSTRACT: Aerosol particulates are generated by coal combustion. The amount and properties of aerosol particulates, specifically size distribution and composition, can be affected by combustion conditions. Understanding the formation of these particles is important for predicting emissions and understanding potential deposition. Oxy-fuel combustion conditions utilize an oxygen-enriched gas environment with CO2. The high concentration of CO2 is a result of recycle flue gas which is used to maintain temperature. A hypothesis is that high CO2 concentration reduces the vaporization of refractory oxides from combustion. A hightemperature drop-tube furnace was used under different oxygen concentrations and CO2 versus N2 to study the effects of furnace temperature, coal type, and gas phase conditions on particulate formation. A scanning mobility particle sizer (SMPS) and aerodynamic particle sizer (APS) were utilized for particle size distributions ranging from 14.3 nm to 20 μm. In addition, particles were collected on a Berner low pressure impactor (BLPI) for elemental analysis using scanning electron microscopy and energy dispersive spectroscopy. Three particle size modes were seen: ultrafine (below 0.1 μm), fine (0.1 to 1.0 μm), and coarse (above 1 μm). Ultrafine mass concentrations were directly related to estimated particle temperature, increasing with increasing temperature. For high silicon and calcium coals, Utah Skyline and PRB, there was a secondary effect due to CO2 and the hypothesized reaction. Illinois #6, a high sulfur coal, had the highest amount of ultrafine mass and most of the sulfur was concentrated in the ultrafine and fine modes. Fine and coarse mode mass concentrations did not show a temperature or CO2 relationship. (The table of contents graphic and abstract graphic are adapted from ref 27.) governed by: MOn(s) + CO(g) ↔ CO2(g) + MOn−1(g). The vapor pressure of the vaporizing oxides or metal (MOn−1) is determined by the equilibrium of the reaction between the refractory metal oxides MOn and CO inside the particle at high temperatures. Therefore, increasing CO2 concentration could effectively reduce the vaporization of oxides, and reduce the yield of ultrafine and fine particulate. Wall et al.9 investigated oxy-fuel combustion in a pilot-scale furnace with an average of 27% oxygen. They found no significant differences in chemical composition and size distribution of bulk fly ash between O2/CO2 and O2/N2 combustion. They did find differences in slagging and fouling in some coal tests. Ash and deposit formation from oxy-fuel combustion, with 27% and 32% oxygen, in a pilot-scale furnace were studied by Yu et al.10 They found no significant impacts on bulk ash particle size distributions (PSDs) and composition in O2/CO2 combustion. Apparent impacts of oxy-fuel combustion were found for the composition of deposits,
1. INTRODUCTION Coal is the least expensive and most widely used solid fuel to produce electricity. The U.S. Department of Energy and the Energy Information Administration1 estimates that the world’s coal consumption will increase by 56% from 2007 to 2035. In the U.S., 90% of the coal is used in electricity generation. Emissions from coal combustion have been of significant concern because of their damage to both the environment and human and animal health. In oxy-fuel combustion, oxygen is used versus air in the combustion zone, with recycled flue gas to keep the system cooler. The result is a flue gas with mostly water and CO2. Ultrafine particulate is a consequence of the minerals in the coal volatilizing and recondensing or nucleating to form submicrometer particles. Raask2 and Bryers3 have investigated particulates, focusing on organic mineral matter in the ash, and they found that formation was highly dependent on the local temperature and gas environment. The formation of submicrometer-size aerosols during coal combustion is known to be due to mineral vaporization under local combustion conditions and subsequent particle formation. Quann and Sarofim,4 supported by the studies of Neville et al.,5 Senior and Flagan,6 Kaupppinen and Pakkanen,7 and Haynes et al.,8 hypothesized that the vaporization of refractory oxides was © 2012 American Chemical Society
Received: Revised: Accepted: Published: 5214
November 29, 2011 March 22, 2012 April 2, 2012 April 2, 2012 dx.doi.org/10.1021/es204196s | Environ. Sci. Technol. 2012, 46, 5214−5221
Environmental Science & Technology
Article
Figure 1. High temperature drop tube furnace.
especially for Ca, Fe, and S oxides. Suriyawong et al.11 studied a subbituminous coal with a drop tube furnace over a range of oxygen conditions, 21% to 50%, and observed that the mean size and the total amount of submicrometer ash particles formed in O2/CO2 combustion was smaller than that formed in O2/N2 combustion for same O2 concentration. Similar results were found in Sheng and co-worker’s drop tube studies at 20 and 40% oxygen.12,13 They concluded that the combustion temperature was the dominant factor, versus CO2 effect on ash formation. Krishnamoorthy and Veranth14 modeled the CO/ CO2 ratio inside a single char particle when it combusted. Their results suggested that changing CO2 concentration in the bulk gas significantly changed the CO/CO2 ratio which could affect the vaporization ratio of refractory oxides and the formation of fine ash particulates. The focus of this study was to investigate the effects of furnace temperature, coal type, and gas phase conditions, namely, CO2 versus N2 on ash particulate formation and refractory oxides composition under controlled conditions. The study utilized particle sizing equipment allowing a wide range of particle sizes, from submicrometer to micrometer. In addition, composition was determined for the different size fractions.
A water-cooled collection probe was inserted along the axis through the bottom of the furnace. The probe position was adjustable to allow for variable residence times. All gas and particulate combustion products were withdrawn through a collection probe for further sampling tests. Pure N2 was added and permeated through a porous tube to quench chemical reactions inside the collection probe. The temperature of the flue gas after quenching was approximately 350 K. The outlet stream was further diluted by particle-free air while passing through a dilution tunnel for sampling with the particle sizing instruments. A stream (0.3 L/min) was introduced into a TSI scanning mobility particle sizer (SMPS, model 3080) and an aerodynamic particle sizer (APS, model 3321) to determine the PSDs. The SMPS has a theoretical measurement range of 14.3−673.2 nm, and the APS ranges from 532 nm to 20 μm. A Berner low pressure impactor (BLPI) was utilized for ash size segregation and collection, as discussed in Hillamo and Kauppien16 and Linak et al.17 The BLPI was connected to the outlet of the collection probe with N2 gas dilution to maintain the total gas flow rate at a constant value of 25.4 L/min (standard). The BLPI uses low pressure and high jet velocities for particle size segregation, collecting particles on 11 stages with particle sizes (50% aerodynamic cutoff diameterin micrometers). Sample was collected on greased aluminum foils for gravimetric measurements. To collect sufficient mass of ultrafine particles for analysis, the BLPI was operated in two steps. First, a cyclone was attached to the outlet of the collection probe to remove particles larger than 1 μm. Operation with the cyclone lasted about 180 min to ensure enough ultrafine particles were collected on the bottom stages, without larger particles overloading the top stages (stages 11− 7). The cyclone was then removed and larger particles were collected on the upper stages. Data from the BLPI and SMPS were compared and found to be the same except for the
2. EXPERIMENTAL METHODS AND MATERIALS A schematic diagram of the high-temperature, drop-tube furnace (HDTF)15 is shown in Figure 1. Synthetic flue gas was used to simulate oxy-fuel conditions and compare these results with air. The furnace was heated by a pair of graphite electric heating elements. An alumina muffle tube (50 mm inside diameter, 40 cm long) served as the central reaction zone with an alumina honeycomb on the top to act as a preheater and flow straightener. The main gas flow rate was maintained at 0.2 L/s and coal feed rate was set at 1.5 g/h. The coal feed rate is low enough that the drop tube operates with little change in gas environment (high excess oxygen conditions). 5215
dx.doi.org/10.1021/es204196s | Environ. Sci. Technol. 2012, 46, 5214−5221
Environmental Science & Technology
Article
ultrafine particle range (less than 0.1 μm) where the small amount of sample is difficult to weigh. Particulates were also collected on polycarbonate membranes for elemental analysis. Particles were deposited on each BLPI impactor stage for 15−20 min. Collected samples were examined by a FEI Nova Nano scanning electron microscope (SEM) and an energy dispersive X-ray spectrometer (EDS) as discussed by Vassilev and Vassileva.18 Six major refractory oxides and sulfur were considered in the elemental analysis (reported as oxides): Na, Mg, Al, Si, Ca, and Fe. Two bituminous coals, Utah Skyline and Illinois #6, and one subbituminous coal, Powder River Basin Black Thunder (PRB) are reported on in this paper. The properties of the coals is summarized in Table 1. All coal samples were dried at 105 °C
3. SINGLE-PARTICLE MODEL Particle temperature changes as a function of oxygen, with higher temperatures at higher concentrations (Suriyawong et al.11). To elucidate temperature differences over different combustion conditions, a simplified, single-particle model was developed for coal char to estimate particle temperatures. As in the studies of Badzioch et al.19 and Carslaw and Jaeger,20 a spherical, homogeneous particle was considered, reactions in the boundary layer were ignored, and no temperature gradient within the particle was used. The only reaction considered at the char particle surface was the direct oxidation of carbon to form carbon monoxide, C + 1/2O2 →CO, following the approach of Field21 and Gavalas,22 and supported by the findings of Tognotti et al.23 Gas-phase oxidation of CO to CO2 was assumed to occur in the free stream and, therefore, did not enter into the energy and material balance equations. The following heat balance expression was used which gives the rate of change of particle temperature with time: dTp 4 ) − hA p(Tp − Tg) + ΔHγ mpCp = −εσA p(Tp4 − T W dt
Table 1. Coal Analysis Dataa Utah Skyline
PRB
Illinois #6
3.2 8.8 38.6 49.4
23.7 4.9 33.4 38.0
9.7 8.0 36.8 45.6
Ultimate Analysis (%), Dry, Ash-Free carbon 77.4 hydrogen 5.9 nitrogen 1.3 sulfur 0.5 oxygen (by difference) 14.9
56.5 6.8 0.8 0.2 35.7
70.3 6.1 1.2 4.3 18.1
Ash Composition (%) Al as Al2O3 Ca as CaO Fe as Fe2O3 Mg as MgO Mn as MnO P as P2O5 K as K2O Si as SiO2 Na as Na2O S as SO3 Ti as TiO2
14.78 22.19 5.2 5.17 0.01 1.07 0.35 30.46 1.94 8.83 1.3
17.66 1.87 14.57 0.98 0.02 0.11 2.26 49.28 1.51 2.22 0.85
Proximate Analysis LOD (%) ash (%) volatile matter (%) fixed C (%)
14.52 6.11 5.09 1.39 0.02 0.59 0.57 60.89 1.41 2.33 0.88
(1)
In eq 1 the wall temperature and gas temperature were considered equal and gas temperatures were measured. The overall reaction rate, γ, kg/s considered oxygen diffusion from the bulk gas to the surface and the surface reaction rate, where, γ = ko·PO2·A p
(2) 24
Properties were from studies of Murphy and Shaddix and the oxidation reaction rates of the chars were assumed to be the same for all the coals. A diameter of 65 μm was used. The properties are shown in the Supporting Information.
4. RESULTS AND DISCUSSION Temperatures were estimated from the above-mentioned model. Constant properties were considered comparisons are made for each coal at the combustion conditions in question separately. Table 2 indicates the estimated temperatures for the Table 2. Char Temperatures for a Particle Diameter of 65 μm (K) Tg = 1410 K
a
Note: Loss on drying (LOD) was determined in air at 105°C for one hour and is reported on an as received sample weight basis. Note: Ash analysis results are reported on an ashed sample weight basis.
O2(%) O2/N2 O2/CO2
21 1747 1694
31.5 1940 1843
Tg = 1520 K 21 1860 1807
31.5 2044 1949
particles at the indicated gas temperatures. It is estimated that temperatures are within ±40 K. Comparing O2/N2 and O2/CO2 combustion conditions, the particle temperature in O2/N2 was estimated to be higher than that in O2/CO2 at the same gas temperature and oxygen concentration conditions. The most likely reason is that the mass diffusivity of O2 in CO2 is lower than that of O2 in N2 which inhibits the oxygen at the particle surface. These results are consistent with data in the literature25,26 and the temperatures are within the same range. Typical PSDs for the higher gas temperature condition, 1520 K, are shown in Figure 2 for the three coals. As seen in this figure, there were three modes of particles: ultrafine, less than 0.1 μm; fine, 0.1 to 1.0 μm; and coarse, greater than 1 μm. Illinois #6 had the most defined segregation between modes.
and sieved to a size range of 54−72 μm to ensure a stable feeding rate. The residence time of particles passing through the drop-tube reaction zone was about 0.9 s which allowed for burnout of the particle; no soot was detected in the exhaust. An O2/CO2 mixture was used as the oxidant in the HTDT experiments to simulate O2/recycled flue gas combustion, and these results were compared with O 2/N 2 combustion conditions. Two oxygen concentrations, 21% and 31.5%, and furnace wall temperatures were studied. The gas temperature was measured and found to be relatively constant along the length of the reactor and under the different gas compositions. The higher gas temperature was approximately 1520 K, whereas the lower was 1410 K. 5216
dx.doi.org/10.1021/es204196s | Environ. Sci. Technol. 2012, 46, 5214−5221
Environmental Science & Technology
Article
Figure 2. PSDs of Utah Skyline (top), PRB (middle) and Illinois #6 (bottom) at higher furnace temperature with a gas temperature of 1520 K.
Figure 3. Mass concentration of ultrafine particles for Utah Skyline, PRB and Illinois #6 coals.
The mean size of each particle mode was found and the SMPS and APS PSD data were integrated to obtain total mass for ultrafine, fine, and coarse modes. Results were normalized to 1000 mg of ash. As illustrated in Figure 2, the diameter and mass for the coarse modes (particles greater than 1 μm) did not change for the three coals. This result is expected since particles in this size range are usually the result of particle attrition. Changes in the ultrafine and fine modes were found for all three coals under the different conditions. In these ranges, particles are formed from volatilization and subsequent condensation and/or nucleation and coagulation.27 These steps are more dependent upon the temperature and gas phase environment. The mass concentrations of ultrafine particle modes, under 100 nm, are shown in Figure 3 for all three coals. The mass concentrations were obtained by integrating the SMPS data over the ranges previously discussed. This number was normalized by the total mass, which was obtained by summing all the fractions. The data are plotted as temperature which was obtained from the model; the error bars show both temperature (horizontal) and concentration (vertical) errors. Utah Skyline
(Figure 3a) showed a significant increase in ultrafine particle mass concentration with temperature. The data for PRB (Figure 3b) are more scattered but also show a slight increase with temperature. Illinois #6 (Figure 3c) had the highest ultrafine mass concentration of the three coals (notice the change in scale). This coal had a high concentration of sulfur (see Table 1) and the ultrafine mass was relatively consistent over all conditions, with the exception of one point. This is likely due to the vaporization of the sulfur which would be independent of temperature and gas environment at these conditions. Both Utah Skyline and PRB coals show a second order effect of CO2 at the higher temperatures where, for the same temperature, the mass of ultrafine was lower for the CO2 condition versus the N2 (comparing dark and light bars). This result is due to the high content of silica in the Utah Skyline and calcium and silica in the PRB. These compounds would be expected to have the hypothesized relationship with CO2 where reaction is driven toward the solid oxide not a reduced vaporized oxide. 5217
dx.doi.org/10.1021/es204196s | Environ. Sci. Technol. 2012, 46, 5214−5221
Environmental Science & Technology
Article
Figure 4. Mean diameter ultrafine for Utah Skyline as a function of temperature.
Figure 5. Mass concentration of fine mode particles.
Figure 6. Elemental mass fraction size distribution of Utah Skyline at 1520K gas condition. The lines show the three modes: ultrafine, fine, and coarse.
5218
dx.doi.org/10.1021/es204196s | Environ. Sci. Technol. 2012, 46, 5214−5221
Environmental Science & Technology
Article
Figure 7. Elemental mass fraction size distribution of PRB at 1520 K gas condition. The lines show the three modes: ultrafine, fine, and coarse.
Figure 8. Elemental mass fraction size distribution of Illinois #6 at 1520K gas condition. The lines show the three modes: ultrafine, fine, and coarse.
higher across the stages and silica is in higher concentrations for this condition. The fine mode also shows enrichment in silica for the highest temperature case. PRB coal has a high content of Ca and Si, and Ca is the predominant ash compound, as Ca is more easily vaporized than Si (Figure 7). Again, the highest amount of calcium was seen at the conditions of the lower, right graph. There is also slightly more silica in this case. The coarse fraction concentrations are also relatively constant. The Illinois #6 data are shown in Figure 8. Illinois #6 coal has a high content of sulfur, Si, and Fe. Large concentrations of sulfur were found in the ultrafine and fine particles, with almost no sulfur in the coarse mode. Si and Fe showed more dependence on O2 concentration versus the combustion environment (N2/CO2) for all stage samples. These results are consistent with the fact that the sulfur was easily vaporized at the temperatures in question.
Figure 4 shows the mean sizes for the Utah Skyline ultrafine particles; as seen in this figure, the diameters are constant across the temperatures and conditions. The mass concentration of the fine mode did not vary significantly except at the highest temperature for the Utah Skyline coal (Figure 5). The fine mode mass concentrations of Illinois #6, not shown here, were much higher than the other two coals in this mode. Figure 6 shows the results from the BLPI and EDS analysis for the Utah Skyline coal with the three modes, ultrafine (
