Formation of Fine Particulate Matter in a Domestic Pellet-Fired Boiler

Article pubs.acs.org/EF

Formation of Fine Particulate Matter in a Domestic Pellet-Fired Boiler U. Fernandes and M. Costa* Mechanical Engineering Department, Instituto Superior Técnico, Technical University of Lisbon, Avenida Rovisco Pais, 1049-001 Lisbon, Portugal ABSTRACT: This study concentrates on the formation of fine particulate matter (PM) in a domestic pellet-fired boiler and includes a detailed characterization of the combustion process inside the combustion chamber of the boiler. Measurements of local mean gas temperatures, major gas species concentrations, and PM concentrations and size distributions are reported for a representative boiler operating condition. In addition, a number of selected PM samples were also morphologically and chemically characterized. The results revealed that (i) most particles present aerodynamic diameters below 1 μm, regardless of the measurement location, and the mass size distributions peak in the size range of 50−130 nm; (ii) early in the flame, the PM size distributions present a bimodal size distribution, but this attribute vanishes as the final stages of the combustion process are approached; (iii) soot particles are rapidly formed early in the combustion process, but these carbon-rich particles also oxidize rapidly in the near burner region; (iv) the aerodynamic diameter corresponding to the maximum amount of particles increases slightly from the fuel bed until ≈4/5 of the combustion chamber height; (v) for a given measurement location and particle sizes below 1 μm, the PM chemical composition varies marginally among size ranges, indicating that the existing chemical compounds tend to form different size aggregates of similar composition; (vi) the element mass concentration varies according to the measurement location; early in the combustion process, the PM is dominated by carbon, with smaller amounts of O and Si, while in the final stages of the combustion process, the PM composition is dominated by inorganic material; and (vii) large fractions of the alkali metals are present as chlorides in the fine PM, and small fractions of the alkalis are present as sulfates.

1. INTRODUCTION In a recent work,1 we have quantified and characterized morphologically the particulate matter (PM) emissions from a domestic pine pellet-fired boiler during steady-state operation for different boiler thermal inputs. In common with other studies, e.g., the study by Tissari et al.,2 the results revealed that large fractions of the PM emitted had dimensions smaller than 2.5 μm, including ultrafine and sub-micrometer- and micrometer-sized particles. The ultrafine particles were mostly composed of inorganic compounds (O, K, Cl, Na, Ca, and Mg), and the submicrometer- and micrometer-sized particles were made of carbon and inorganic compounds (O, K, Ca, Mg, and P). These fine particles have a large impact on human health, with their capture using conventional gas cleaning devices being very difficult. Under these circumstances, it is important to find ways to reduce their formation during the combustion process, which requires a good understanding of the formation mechanisms of fine PM during biomass combustion. Related previous studies on domestic pellet-fired boilers concentrated primarily on PM emissions,2−7 with little attention devoted to the particle formation process from the flame zone to the post-flame region. The few exceptions available in the literature include the interesting studies by Wiinikka et al.8,9 These authors investigated the various stages of high-temperature aerosol formation in an 8 kW updraft fired wood pellet reactor that has been custom-designed for systematic investigation of particle emissions. From spatially resolved PM measurements along the certerline of the combustor, they concluded that the particle size distribution below 2.5 μm was bimodal, with distinct fine and coarse modes. Early in the flame, both modes were dominated by particles from incomplete combustion © 2013 American Chemical Society

(i.e., soot particles), whose concentrations decrease rapidly in the hot oxygen-rich region because of carbon oxidation. After the hot flame, the fine mode concentration and particle diameter increase gradually because of condensation on preexisting particles in the gas of alkali sulfates, alkali chlorides, and Zn species formed from constituents vaporized in the fuel bed, while the concentration of the coarse mode, dominated by carbon, refractory oxides, and alkali metals, decreased very little. In the context of the PM formation, important knowledge has been derived from studies carried out in small-scale reactors, e.g., in the studies by Sippula et al.10 and Knudsen et al.,11 pilotscale circulating fluidized-bed reactors, e.g., in the study by Lind et al.,12 large-scale circulating fluidized-bed boilers, e.g., in the study by Valmari et al.,13 and utility boilers, e.g., in the study by Christensen et al.,14 which must be mentioned here. For example, a number of studies11−13 showed that the chlorine present in the fuel play an important role in the release of ash-forming elements to the gas phase, which leads to an increase in the fine particle formation. Also, the fuel sulfur may affect the particle formation in various ways, as discussed by Sippula et al.10 Moreover, there is evidence that fuel potassium released to the gas phase is significant at combustion temperatures above 700 °C, but the quantity released is largely determined by the ash composition.11 The subject is complex and certainly needs further investigation. The aim of this study is to enhance the current Received: August 31, 2012 Revised: January 3, 2013 Published: January 3, 2013 1081

dx.doi.org/10.1021/ef301428m | Energy Fuels 2013, 27, 1081−1092

Energy & Fuels

Article

Figure 1. Schematic of the experimental setup, showing the boiler and the measuring techniques.

Figure 2. Schematic of the combustion chamber, showing the burner and the measurement traverses (ports 1−5).

understanding of the fine PM formation mechanisms during biomass combustion through a detailed characterization of the combustion process inside a domestic pine pellet-fired boiler. The characterization includes detailed measurements of local mean gas temperatures, O2, CO2, unburnt hydrocarbon (HC), CO, and NOx concentrations, and PM concentrations and size distributions for a representative boiler operating condition. In addition, a number of selected PM samples were also morphologically and chemically characterized.

2. MATERIALS AND METHODS The present tests have been performed in a domestic wood pelletfired boiler with a maximum thermal capacity of 22 kW, with forced draft.1 Figure 1 shows a schematic of the experimental setup. The pellets are manually loaded into a hopper with a capacity of 45 kg and are fed to the burner through a screw feeder that works by impulses. The feeding rate of the pellets is regulated by the boiler load, and the pellet consumption rate is measured with the aid of a loss-in-weight technique, for which the boiler is mounted on a weighbridge. 1082


Energy & Fuels

Article

Table 1. Calibrated Cutoff Diameters, d50, of the 13-Stage Dekati Low-Pressure Cascade Impactor stage

1

2

3

4

5

6

7

8

9

10

11

12

13

d50 (μm)

0.0282

0.0559

0.0927

0.153

0.259

0.379

0.608

0.940

1.59

2.37

3.96

6.63

9.84

Figure 2 shows a schematic of the rectangular combustion chamber and burner. Note that the front side of the rectangular combustion chamber incorporates a door with a large ceramic glass window, which allows for viewing of the flame. The combustion of the pellets takes place within a hemispherical basket (brazier) with a diameter of 120 mm. The basket is top-fed with pellets by the screw. Ignition is accomplished with the aid of an electrical resistance placed close to the basket, and the primary air is supplied by a dedicated fan to the basket through several small orifices located across the basket bottom. A short cleaning period of the basket is programmed to occur once every 11.5 min. During the cleaning process, the fuel supply decreases and the air supply increases for a few minutes to remove the ashes accumulated at the bottom of the basket (bottom ashes). The resulting hot gases from the combustion exchange heat with water circulating in a heat exchanger located at the top of the combustion chamber. The heat transferred to the water in the boiler is dissipated through a plate heat exchanger with the aid of an external water circuit. For the present study, the combustion chamber was modified to allow for the introduction of probes inside the rectangular combustion chamber. The probes could be inserted horizontally inside the combustion chamber through five ports located along the middle of the side wall, so that measurements were taken in a horizontal plane cutting through the burner axis. Figure 2 shows schematically the location of the five measurement traverses (ports 1−5) and the x positions where in-flame measurements of temperature, gas species concentration and PM were performed. Estimated residence times from the burner to ports 1−5 are 28, 56, 88, 115, and 147 ms, respectively, with the overall residence time in the combustion chamber being 197 ms. Local mean temperatures were measured using uncoated 76 μm diameter fine wire platinum/platinum:13% rhodium (type R) thermocouples (Figure 1). The hot junction was installed and supported on 300 μm wires of the same material located in a twinbore alumina sheath with an external diameter of 5 mm and placed inside a stainless-steel tube. Because radiation losses represent the major source of uncertainty in the mean temperature measurements, an attempt was made to quantify them on the basis of a theoretical expression developed by De.15 The expression requires temperature measurements obtained at the same point with three thermocouples of the same material but of different diameter. In this study, platinum/ platinum:13% rhodium thermocouples with wires of diameters of 25 and 300 μm were also used to measure the temperature. The calculations led to uncertainties of 8% in the regions of highest temperature and lower elsewhere. An additional source of uncertainty in the temperature measurements relates to the deposition of PM on the thermocouple wires. This results in an increase in radiation losses, owing to the increase in surface area. The error because of PM deposition on the thermocouple wires was minimized by examining the condition of the thermocouple frequently during measurements and, if necessary, by replacing the thermocouple. The sampling of the gases for the measurement of local mean O2, CO2, HC, CO, and NOx concentrations was achieved using a stainlesssteel water-cooled probe (Figure 1), which has been designed to minimize the major sources of uncertainty in the concentration measurements inside the combustor, namely, those associated with the quenching of the chemical reactions and the aerodynamic disturbances of the flow. The probe was composed of a central 1.3 mm inner diameter tube through which quenched samples were evacuated. This central tube was surrounded by two concentric tubes for probe cooling. The gas sample was drawn through the probe and part of the system by an oil-free diaphragm pump. A condenser removed the main particulate burden and condensate. A filter and a drier removed any residual particles and moisture, so that a constant supply of clean dry combustion gases was delivered to the analyzers through a manifold to

give species concentration on a dry basis. The analytical instrumentation included a magnetic pressure analyzer for O2 measurements, a non-dispersive infrared gas analyzer for CO2 and CO measurements, a flame ionization detector for HC measurements, and a chemiluminescent analyzer for NOx measurements. Quenching of the chemical reactions was rapidly achieved upon the samples being drawn into the central tube of the probe because of the high water cooling rate in its surrounding annulus; our best estimate indicated quenching rates of about 107 K/s. No attempt was made to quantify the probe flow disturbances. On average, the repeatability of the gas species concentration data was within 10% of the mean value. Jiménez and Ballester16 presented a very interesting comparative study of different methods for the sampling of high-temperature combustion aerosols, which included three different probes based on aerodynamic quenching, nitrogen dilution, and thermophoresis. In the present study, PM sampling was accomplished with the aid of a rapid nitrogen dilution sampling quartz probe similar to those used by Wiinikka et al.8,9 and Strand et al.17 Figure 1 shows a schematic of the PM sampling probe and associated sampling system used in this study. At the probe tip, the combustion gases were diluted with nitrogen using a dilution ratio of 25, whose calculation was made with the aid of NOx concentration measurements at the probe outlet. Upon leaving the probe, a part of the flow (18 NL/min) was sent to a Tecora total filter holder, where the total mass concentration of PM was determined by weighting the quartz microfiber filter used in the filter holder. The remaining part of the flow (10 NL/min) was sent to a 13-stages Dekati low-pressure cascade impactor (DLPI, Dekati, Ltd.) that size-classified the PM according to their aerodynamic diameter in

Table 2. Characteristics of the Pine Pellets parameter

value

Proximate Analysis (wt %, As Received) volatile matter 80.5 fixed carbon 10.9 moisture 7.3 ash 1.3 Ultimate Analysis (wt %, Dry and Ash Free) carbon 46.0 hydrogen 6.2 nitrogen 0.5 sulfur 10 cm, near the side walls of the combustion chamber, the relatively high concentrations of O2 and the very low HC and CO concentrations reveal the absence of chemical reactions in these zones and the relatively high excess air level associated with the operation of this type of boiler. The fall in the temperature noted near the side walls at port 1 is due to the upstream convection of colder gases from downstream. At port 2 (z = 7.1 cm), the main reaction zone is also located at x ≈ 5 cm, with the lower HC and CO concentrations, as compared to those measured at port 1, revealing the progress of the combustion process. It is interesting to note that the NOx concentrations reach maximum values at ports 1 and 2, where the maximum mean temperature does not exceeds 1200 °C. This indicates that the fuel NO mechanism is the main contributor to the NOx emissions from the present domestic boiler. 1085


Energy & Fuels

Article

Figure 6. Typical SEM images: (a) port 1 at x = 0 cm, (b) port 1 at x = 5 cm, (c) port 2 at x = 0 cm, (d) port 2 at x = 7.5 cm, (e) port 3 at x = 0 cm, (f) port 3 at x = 10 cm, (g) port 4 at x = 0 cm, and (h) port 5 at x = 0 cm.

Figure 5 shows the PM size distribution in eight measurement locations (five along the burner axis and three near the visible flame boundary) for the boiler operating condition studied here. Note that the x axis in Figure 5 represents the aerodynamic diameter, which means that the particle diameter is not corrected for different material densities. The figure reveals that most of the particles have aerodynamic diameters below 1 μm, regardless of the measurement location, and that the mass size distributions peak in the size range of 50− 130 nm. It is interesting to note that PM size distributions at port 1 and near the visible flame boundary at ports 2 and 3

present a bimodal size distribution, as also observed by Wiinikka et al.8 It is also evident that, with the increasing distance from the burner, along its axis, the bimodal nature of the size distribution vanishes presumably because of the oxidation of the soot present in the ultrafine particle mode. Figure 5 also reveals that, along the burner axis, the total PM mass concentration decreases significantly from port 1 to port 2 (cf. Figure 4). In biomass combustion, fine PM is produced from incomplete combustion (i.e., soot) and from vaporization and condensation of easily volatile ash elements. The high 1086


Energy & Fuels

Article

Figure 7. PM chemical composition in six substrates from port 1 at x = 0 cm.

usually heavily sintered. With the information available, we are unable to offer an explanation for this observation, but the three-dimensional nature of the flow may play a role in the decrease of the aerodynamic diameter, corresponding to the maximum amount of particles from port 4 to port 5. Away from the burner axis, in port 2 at x = 7.5 cm and in port 3 at x = 10 cm, near the visible flame boundary, Figure 5 shows that the total PM mass concentrations are higher than those measured in port 2 at x = 0 cm and in port 3 at x = 0 cm, respectively. This is presumably because the soot formation rates in port 2 at x = 7.5 cm and in port 3 at x = 10 cm are higher than those at the burner axis, owing to the lower oxygen concentrations and higher temperatures that exist in these outer x positions. Of course, the soot destruction rates may also contribute to the differences observed. 3.3. PM Morphology and Chemical Composition. Figure 6 shows typical SEM images of PM sampling from the eight measurement locations considered in Figure 5. This figure intends to show the general appearance of the PM collected in the impactor substrates rather than providing information on single particle morphology because they were in the gas phase inside the combustion chamber. Panels a−d and f of Figure 6 show micrographs of PM collected from locations early in the

amounts of PM encountered at the axis of port 1 reveal that soot particles are rapidly formed early in the combustion process. Moreover, it is seen that these carbon-rich particles also tend to oxidize rapidly between ports 1 and 2 because of the relatively high temperatures and oxygen availability in this region. Beyond port 2, along the burner axis, the decrease in the total PM mass concentration is significantly smaller for two reasons: first, because there is progressively less soot formation and, second, because the inorganic vapors released in the fuel bed tend to condensate as the temperature decreases. This condensation process can be either a homogeneous nucleation process, where fine particles are formed directly from the gas, or a heterogeneous condensation process, where condensation takes place on existing particles. In addition, still along the burner axis, Figure 5 discloses that the aerodynamic diameter corresponding to the maximum amount of particles increases slightly from port 1 to port 4: specifically, from 0.0927 μm for port 1 to 0.259 μm for port 4. This is because, besides heterogeneous condensation, coagulation/ agglomeration of PM occurs under such high particle concentrations. From port 4 to port 5, however, a decrease of the aerodynamic diameter (to 0.0559 μm) is observed, corresponding to the maximum amount of particles, which cannot be attributed to PM fragmentation because the agglomerates found in biomass combustion processes are 1087


Energy & Fuels

Article

Figure 8. PM chemical composition in five substrates from port 3 at x = 0 cm.

flame, where a large number of soot aggregates are present. In all of these micrographs, the presence of inorganic compounds is hardly discernible. Given the relatively high temperatures encountered in this region (cf. Figure 3), inorganic species should exist in the vapor phase in these sampling locations. During the sampling process, the inorganic vapors should condense to the particulate phase in the quartz probe as the sample cools; e.g., see the study by Jiménez and Ballester.16 The reduced presence of inorganic compounds observed in panels a−d and f of Figure 6 suggests that losses of inorganic vapors occur in the sampling probe. This unavoidable artifact16 has to be carefully taken into consideration when analyzing data obtained using the present method. Figure 6e shows a micrograph of PM collected from port 3 at x = 0 cm, where it is possible to observe PM (soot) covered by inorganic compounds that have condensed over them. In light of the argumentation above, it can be concluded with reasonable justification that the condensation of the inorganic vapors over the soot aggregates has occurred, to a large extent, inside the combustion chamber. Finally, panels g−h of Figure 6 show micrographs of PM collected from ports 4 and 5 at x = 0 cm. Here, some crystalline compounds are clearly identified. These compounds are expected to exist inside the combustion chamber at ports 4 and 5 as a result of the relatively lower

temperatures in this region (cf. Figure 3), which have permitted the condensation of inorganic vapors. Figures 7−10 show the PM chemical composition in a number of substrates for the ports and x positions considered in Figure 6. For a given port and x position, the results reveal that the PM chemical composition varies marginally among substrates; note that the particle sizes analyzed here are below 1 μm (see Table 1), which may explain this observation. This indicates that the existing chemical compounds tend to form different size aggregates of similar composition. Because of this, substrate number 3 was chosen for the remaining analyses; this substrate was the one where more PM mass was collected. Figure 11 shows the PM chemical composition in substrate number 3 in the eight measurement locations showed in Figure 5, and Figure 12 shows the evolution of the PM chemical composition in substrate number 3 with the temperature along the burner axis (ports 1−5). Figures 11 and 12 reveal that the concentrations of K and Cl are lower at higher temperatures, near the fuel bed, than at lower temperatures, near the exit of the combustor chamber. Because the elements K and Cl are released from the fuel bed, its concentration cannot increase from port 1 to port 5. The explanation for these trends is the 1088


Energy & Fuels

Article


present at port 5, it is clear that some soot aggregates were able to escape from the combustion zone, contributing to the fine PM emissions, as observed in our previous work. 1 Figure 11 also reveals that, at the visible flame boundary, the presence of carbon in the PM is generally higher than that at the burner axis, probably because of the enhanced soot formation in this region, as discussed earlier. The concentrations of the inorganic elements in this outer region are, however, similar to those at the burner axis. To gain information on the formation of alkali metal species, Table 3 shows the molar ratios (Na + K)/Cl and (Na + K)/ (Cl + 2S) in the eight measurement locations shown in Figure 11. According to Lind et al.,12 these ratios are 1 if all of the alkali metals are bound as chlorides or as sulfates and chlorides, respectively, and if all of the Cl and S are bound with alkali metals. Along the burner centerline, the (Na + K)/Cl molar ratio increases from 0.89 in port 1 at x = 0 cm to almost 2 in port 5 at x = 0 cm. This indicates that almost all of the alkali metals were present as chlorides in the fine PM. Moreover, in port 3 at x = 0 cm, port 4 at x = 0 cm, and port 5 at x = 0 cm, the alkali metal/chlorine ratios were 1.18, 1.45, and 1.98, respectively, which indicates that, in addition to chlorides, alkali metals were present as other compounds. Furthermore, from port 3 at x = 0 cm to port 5 at x = 0 cm, the (Na + K)/(Cl + 2S) molar ratios are very

losses of K and Cl vapors in the sampling probe, as discussed above. Figures 11 and 12 show that the elements mass concentration varies according to the measurement location and, thus, temperature. At the burner axis of ports 1 and 2, the PM is dominated by atomic carbon atoms (≈70%) from incomplete combustion, with smaller amounts of O (≈15%) and the nonvolatile refractory metal Si (≈15%), whose origin was probably from the fragmentation of char particles ejected from the fuel bed. Note that, in this near fuel bed region, elements such as K and Cl have to exist in the gas phase, but the experimental method used was unable to detect them, as pointed out earlier. With an increasing distance from the burner, the carbon oxidizes and the inorganic elements K, Na, and Cl condensate, which is particularly evident from port 3 to port 5, after the main reaction zone. In this far fuel bed region, the temperature is lower than in ports 1 and 2, and thus, a significant fraction of these inorganic elements is already in the particulate phase, which enables their successful sampling with the present dilution probe. At the burner axis of port 5, the PM composition is dominated by inorganic material; specifically, the PM is composed of C (≈15%), O (≈29%), Si (≈8%), Na (≈3%), S (≈6%), Cl (≈13%), K (≈22%), and Zn (≈1%). Despite the relatively low concentration of carbon 1089


Energy & Fuels

Article


of the soot particles within the higher temperature regions and/or higher temperatures in the combustion zone. The latter strategy may, however, enhance the vaporization of the easily volatile ash elements, which may originate higher ashbased PM emissions. Consequently, at the present time, an important challenge regarding the fine PM formation and emission in domestic applications is to design burners/ boilers capable of satisfying both the requirements of increasing the soot oxidation rates and decreasing the vaporization rates of the inorganic materials present in the biomass fuels.

Table 3. Molar Ratios (Na + K)/Cl and (Na + K)/(Cl + 2S) as a Function of the Sampling Location sampling location port port port port port port port port

1, 2, 3, 4, 5, 1, 2, 3,

x x x x x x x x

= = = = = = = =

0 cm 0 cm 0 cm 0 cm 0 cm 5 cm 7.5 cm 10 cm

temperature (°C)

(Na + K)/Cl

(Na + K)/(Cl + 2S)

1178 1053 918 859 733 1220 1007 921

0.89 1.03 1.18 1.45 1.98 0.72 1.78 1.05

0.81 0.90 0.93 0.95 0.98 0.69 1.21 0.85

4. CONCLUSION Measurements of local mean gas temperatures, major gas species concentrations, and PM concentrations and size distributions are reported for a representative domestic pelletfired operating condition. In addition, a number of selected PM samples were also morphologically and chemically characterized. The results revealed that (i) most particles present aerodynamic diameters below 1 μm, regardless of the measurement location, and the mass size distributions peak in the size range of 50−130 nm; (ii) early in the flame, the PM size distributions present a bimodal size distribution, but this attribute vanishes as the final stages of the combustion process are approached presumably because of the oxidation of the soot

close to 1, which indicates that the fraction of alkali metals that was not bound in the fine PM as chlorides was present as sulfates. Away from the burner axis, in port 1 at x = 5 cm and port 3 at x = 10 cm, near the visible flame boundary, the molar ratios are similar to those at the burner centerline, but in port 2 at x = 7.5 cm, both molar ratios are higher than 1, indicating that the alkali metals were bound as other compounds in addition to chlorides and sulfates. As a final remark, it should be pointed out that the presence of soot particles in the flue gas can be minimized through combustion modifications that lead to longer residence times 1090


Energy & Fuels

Article

Figure 11. PM chemical composition in various measurement locations.

molar ratios (Na + K)/Cl and (Na + K)/(Cl + 2S) suggest that large fractions of the alkali metals are present as chlorides in the fine PM and that small fractions of the alkalis are present as sulfates.

■

AUTHOR INFORMATION

Corresponding Author

*Telephone: +351218417186. E-mail: [email protected]. Notes

The authors declare no competing financial interest. Figure 12. Evolution of the PM chemical concentration with the temperature.

■

ACKNOWLEDGMENTS

■

REFERENCES

The authors express their gratitude to the COST Action CM0901 (European Union). The authors also thank undergraduate student José Madeira, who assisted them in conducting the work presented here. U. Fernandes is pleased to acknowledge the Fundaçaõ para a Ciência e a Tecnologia (FCT) for the provision of Scholarship SFRH/BD/82438/ 2011.

present in the ultrafine particle mode; (iii) soot particles are rapidly formed early in the combustion process, but these carbon-rich particles also oxidize rapidly because of the relatively high temperatures and oxygen availability in the near burner region; (iv) the aerodynamic diameter corresponding to the maximum amount of particles increases slightly from the fuel bed until ≈4/5 of the combustion chamber height, because the occurrence of heterogeneous condensation and PM coagulation/ agglomeration; (v) for a given measurement location and particle sizes below 1 μm, the PM chemical composition varies slightly among size ranges, indicating that the existing chemical compounds tend to form different size aggregates of similar composition; (vi) the elements mass concentration varies according with the measurement location; early in the combustion process, the PM is dominated by carbon from incomplete combustion, with smaller amounts of O and Si, while in the final stages of the combustion process, the PM composition is dominated by inorganic material; and (vii) the

(1) Fernandes, U.; Costa, M. Particle emissions from a domestic pellets-fired boiler. Fuel Process. Technol. 2012, 103, 51−56. (2) Tissari, J.; Lyyränen, J.; Hytönen, K.; Sippula, O.; Tapper, U.; Frey, A.; Saarnio, K.; Pennanen, A. S.; Hillamo, R.; Salonen, R. O.; Hirvonen, M. R.; Jokiniemi, J. Fine particle and gaseous emissions from normal and smouldering wood combustion in a conventional masonry heater. Atmos. Environ. 2008, 42, 7862−7873. (3) Johansson, L. S.; Leckner, B.; Gustavsson, L.; Cooper, D.; Tullin, C.; Potter, A. Emission characteristics of modern and old-type residential boilers fired with wood logs and wood pellets. Atmos. Environ. 2004, 38, 4183−4195. 1091


Energy & Fuels

Article

(4) Wiinikka, H.; Gebart, R. Critical parameters for particle emissions in small-scale fixed-bed combustion of wood pellets. Energy Fuels 2004, 18, 897−907. (5) Wiinikka, H.; Gebart, R. The influence of air distribution rate on particle emissions in fixed bed combustion of biomass. Combust. Sci. Technol. 2005, 177, 1747−1766. (6) Bignal, K. L.; Langridge, S.; Zhou, J. L. Release of polycyclic aromatic hydrocarbons, carbon monoxide and particulate matter from biomass combustion in a wood-fired boiler under varying boiler conditions. Atmos. Environ. 2008, 42, 8863−8871. (7) Atkins, A.; Bignal, K.; Zhou, J.; Cazier, F. Profiles of polycyclic aromatic hydrocarbons and polychlorinated biphenyls from the combustion of biomass pellets. Chemosphere 2010, 78, 1385−1392. (8) Wiinikka, H.; Gebart, R.; Boman, C.; Boström, D.; Nordin, A.; Ö hman, M. High-temperature aerosol formation in wood pellets flames: Spatially resolved measurements. Combust. Flame 2006, 147, 278−293. (9) Wiinikka, H.; Gebart, R.; Boman, C.; Boström, D.; Ö hman, M. Influence of fuel ash composition on high temperature aerosol formation in fixed bed combustion of woody biomass pellets. Fuel 2007, 86, 181−193. (10) Sippula, O.; Lind, T.; Jokiniemi, J. Effects of chlorine and sulfur on particle formation in wood combustion performed in a laboratory scale reactor. Fuel 2008, 87, 2425−2436. (11) Knudsen, J. N.; Jensen, P. A.; Dam-Johansen, K. Transformation and release to the gas phase of Cl, K, and S during combustion of annual biomass. Energy Fuels 2004, 18, 1385−1399. (12) Lind, T.; Kauppinen, E. I.; Hokkinen, J.; Jokiniemi, J. K.; Orjala, M.; Aurela, M.; Hillamo, R. Effect of chlorine and sulphur on fine particle formation in pilot-scale CFBC of biomass. Energy Fuels 2006, 20, 61−68. (13) Valmari, T.; Kauppinen, E. I.; Kurkela, J.; Jokiniemi, J. K.; Sfiris, G.; Revitzer, H. Fly ash formation and deposition during fluidized bed combustion of willow. J. Aerosol Sci. 1998, 29, 445−459. (14) Christensen, K. A.; Stenholm, M.; Livbjerg, H. The formation of submicron aerosol particles, HCl and SO2 in straw-fired boilers. J. Aerosol Sci. 1998, 29, 421−444. (15) De, D. S. Measurement of flame temperature with a multielement thermocouple. J. Inst. Energy 1981, 54, 113−116. (16) Jiménez, S.; Ballester, J. A Comparative study of different methods for the sampling of high temperature combustion aerosols. Aerosol Sci. Technol. 2005, 39, 811−821. (17) Strand, M.; Bohgard, M.; Swietlicki, E.; Gharibi, A.; Sanati, M. Laboratory and field test of a sampling method for characterization of combustion aerosols at high temperatures. Aerosol Sci. Technol. 2004, 38, 757−765.

1092


Formation of Fine Particulate Matter in a Domestic Pellet-Fired Boiler

Recommend Documents