Elemental and Organic Carbon in Flue Gas Particles of Various
Transcription
Elemental and Organic Carbon in Flue Gas Particles of Various
Elemental and Organic Carbon in Flue Gas Particles of Various Wood Combustion Systems Project management: Christian Gaegauf, dipl. Ing. ETH/SIA gaegauf@oekozentrum.ch Project team: Martin Schmid, dipl. Ing. FH schmid@oekozentrum.ch Pierre Güntert, dipl. Ing. FH guentert@oekozentrum.ch Langenbruck, December 2005 ¨Ökozentrum Langenbruck, CH-4438 Langenbruck www.oekozentrum.ch Elemental and Organic Carbon in Flue Gas Particles of Various Wood Combustion Systems 2 Elemental and Organic Carbon in Flue Gas Particles of Various Wood Combustion Systems Content Zusammenfassung 5 Messmethoden 5 Resultate 6 Diskussion 6 Schlussfolgerungen 7 1 Introduction 8 2 Materials and Methods 8 2.1 Flue gas sampling 2.2 Particle analysis 10 2.3 Carbon analysis 10 2.4 Wood combustion systems 11 3 Results 8 11 3.1 Coulometer 11 3.2 Particle size distribution 12 3.3 Emission factors 13 4 Discussion 14 4.1 EC- and OC-content in total particle emissions 14 4.2 Particle emissions in the different particle sizes 16 4.3 EC- and OC-emissions in the different particle classes 16 4.4 Products of incomplete combustion 18 4.5 Comparison of particle emissions 19 5 Conclusions 20 Acknowledgements 21 References 21 3 Elemental and Organic Carbon in Flue Gas Particles of Various Wood Combustion Systems English text is based on the paper presented at the 8th International Conference on Energy for a Clean Environment, Lisbon 2005 Authors C. K. Gaegauf, M. R. Schmid and P. Güntert Centre of Appropriate Technology and Social Ecology, Langenbruck, Switzerland Abstract The airborne particulate matter (PM) in the environment is of ever increasing concern to authorities and the public. The major fractions of particles in wood combustion processes are in the size less than 1 micron, typically in the range of 30 to 300 nm. Of specific interest is the content of the elemental carbon (EC) and organic carbon (OC) in the particles since these substances are known for its particular potential as carcinogens. Various wood combustion systems have been analysed (wood chip boiler, pellet boiler, wood log boiler, wood stove and open fire). The sampling of the particles was done by mean of a multi-stage particle sizing sampler cascade impactor. The impactor classifies the particles collected according to their size. The 7 stages classify the particles between 0.4 and 9 microns aerodynamic diameter. The analytical method for determining the content of EC and OC in the particles is based on coulometry. The coulometer measures the conductivity of CO2 released by oxidation of EC in the samples at 650°C. The OC content is determined by pyrolysis of the particle samples in helium atmosphere. Keywords wood combustion, smoke emission, particulates, elemental carbon, organic carbon 4 Elemental and Organic Carbon in Flue Gas Particles of Various Wood Combustion Systems Zusammenfassung Holzfeuerungen sind unter den stationären Anlagen eine relevante Emissionsquelle für Staub- und Feinpartikelemissionen. Von speziellem Augenmerk sind die Feinpartikel unter 10 µm (PM10), da sie über den menschlichen Atmungstrakt in den Körper penetrieren können. Wegen der Karzinogenität des organischen (Organic Carbon, OC) und elementaren (Elemental Carbon, EC) Kohlenstoffes ist der OC- und EC-Anteil in den Staubemissionen für die Beurteilung der Umweltbelastung von Holzfeuerungen von Bedeutung. In der vorliegenden Untersuchung wurden die Emissionsfaktoren sowohl des Gesamtstaubes (Total Particle Emissions, TPE) wie der Feinpartikelfraktionen verschiedener Holzfeuerungen ermittelt. Zusätzlich wurden die Emissionsfaktoren der OC- und EC-Anteile im Gesamtstaub und den Feinpartikelfraktionen bestimmt. Für die zu untersuchenden Holzfeuerungen wurde je ein typischer Vertreter der Kategorien Holzöfen, Stückholz-, Pellet- und Schnitzelkessel ausgewählt und im Felde unter Praxisbedingungen gemessen. Die Staub- und Partikelemissionen eines offenen Feuers lieferten Vergleichsdaten zu den untersuchten Feuerungssystemen mit geschlossenen Brennräumen. Messmethoden Gesamtstaub Die Gesamtstaubmessung wurde auf gravimetrischer Basis mit Planfiltern durchgeführt. Die verwendeten Planfilter basierten auf Quarzfasergewebe, damit sie für die EC/OC-Analytik tauglich waren. Feinpartikel Die Feinpartikel-Klassifizierungen wurden mit einem Kaskaden-Impaktor (Firma Andersen) durchgeführt. Die Abstufung der Partikelklassen waren wie folgt gegeben: > 9 µm, 5.8 - 9 µm, 4,7 - 5.8 µm, 3.3 - 4.7 µm, 2.1 3.3 µm, 1.1 - 2.1 µm, 0.7 - 1.1 µm, 0.4 - 0.7 µm und < 0.4 µm. EC/OC-Analytik Die Bestimmung des EC/OC-Anteils in den Partikelemissionen erfolgte nach dem coulometrischen Standardverfahren (VDI 2465, Chemisch-analytische Bestimmung des elementaren Kohlenstoffes nach Extraktion und Thermodesorption des organischen Kohlenstoffes). Das SUVA-Analyselabor in Luzern führte die Analysen durch. Es wurden sowohl die Gesamtstaubproben wie die Feinpartikelfraktionen mit dem Standardverfahren auf den EC/OCGehalt hin untersucht. Bei der Schnitzelfeuerung wurde die EC/OC-Analytik bei sämtlichen Partikelklassen durchgeführt, bei den übrigen Anlagen nur in den drei Klassen >9 µm, 2.1 - 3.3 µm und <0.4 µm. Die Gesamtmenge des organischen Kohlenstoffs (Total Carbon, TC) bestimmt sich aus der Summe von EC und OC. Abgase In Ergänzung zu den Staub- und Partikelmessungen wurde Kohlenmonoxid (CO) als gasförmiger Schadstoff im Abgas erfasst. Vormessung Die EC/OC-Emissionsuntersuchungen der Feuerungen erfolgt in einem zweistufigen Vorgehen. In Vormessungen wurden die vorgesehenen Messverfahren an einer Schnitzelfeuerung exemplarisch durchgeführt. Anhand der Labor-Analysemethoden und der Datenauswertungen wurde das Messprozedere für die Hauptmesskampagne festgelegt. Für die coulometrischen Untersuchungen analysierte das SUVA-Labor den EC-Gehalt bei Analysetemperaturen von 650°C bzw. 800°C. Von den Filterproben wurde eine Filterhälfte mit Phosphorsäure behandelt, um die Carbonate aus dem Filterstaub herauszulösen. Tabelle 2 zeigt, dass der EC-Anteil der unbehandelten Probe Faktoren höher ist als der EC-Anteil der behandelten. Dies ist darauf zurückzuführen, dass sich die Carbonate über 650°C verflüchtigen. Weiter kann festgestellt werden, dass mit Phosphorsäure behandelte Filterproben bei der EC5 Elemental and Organic Carbon in Flue Gas Particles of Various Wood Combustion Systems Analyse auf dem Temperaturniveau 650°C und 800°C ähnliche EC-Werte aufweisen. Daraus lässt sich schliessen, dass die EC-Analyse bei einer Coulometertemperatur von 650°C verlässliche Werte liefert. Damit ist bei dem Analysetemperaturniveau von 650°C die Vorbehandlung der Filterproben mit Phosphorsäure nicht erforderlich. Resultate Vormessung an Schnitzelfeuerung Die Vormessung an der Schnitzelfeuerung über alle Impaktorstufen zeigte, dass die TC-Konzentration in den Partikelgrössen-Klassen von >9 µm und <0.4 µm am höchsten ist (Fig. 6). Der EC-Anteil am Gesamtkohlenstoff (TC) nimmt zu kleineren Partikelgrössen hin ab, der OC-Anteil zu. Der prozentuale Anteil bezogen auf den Gesamtstaub liegt für EC bei 1%, für OC bei 3.1%. Der TC-Anteil als Summe aus EC und OC erreicht im Gesamtstaub 4.1%. In den einzelnen Partikelgrössen-Klassen liegt der Anteil für EC zwischen 1 - 5%, für OC zwischen 1.8 - 3.4%. Der TC-Anteil liegt bei allen Klassen unter 7%. Die Schnitzelfeuerung wies eine weitgehend vollständige Verbrennung auf, was sich aus den geringen CO-Werten im Abgas zwischen 43 und 64 mg/m3n (@13% O2) schliessen lässt. Hauptmesskampagne Die EC/OC-Analyse wurde in der Hauptmesskampagne in den drei Impaktorstufen >9 µm, 2.1 - 3.3 µm und <0.4 µm durchgeführt. Diese drei Klassen erwiesen sich auf Grund der Vormessungen als sinnvoll: >9 µm und <0.4 µm hatten den höchsten Staub und TC-Anteil, die Impaktorstufe 2.1 - 3.3 µm liefert einen Anhalt der Werte in den Zwischenklassen. Die Zusammenstellung der Emissionsfaktoren findet sich in Tabelle 3. In Tabelle 4 sind die Emissionsfaktoren für den Gesamtstaub in den Partikelgrössenklassen PM10, PM2.5 und PM1 abgeschätzt. Die prozentuale Aufteilung der Grössenklassen erfolgte mit dem Impaktor. Die Absolutwerte lieferte die Gesamtstaubmessung. In Tabelle 5 sind Bereiche von Emissionsfaktoren aus diversen Feldmesskampagnen des Ökozentrums Langenbruck aufgeführt. Neben den auf die Brennstoffmasse bezogenen Emissionsfaktoren finden sich auch die auf dem unteren Heizwert basierenden Emissionswerte. Als Anhaltspunkt für die gemessenen Emissionsfaktoren enthält die Tabelle auch die Staubgehalt-Anforderungen des Schweizer Qualitätssiegels für Holzfeuerungen sowie die EURO-3- und EURO-4-Partikel-Grenzwerte für Dieselfahrzeuge differenziert nach Personen/Lieferwagen und Lastwagen. Diskussion EC/OC-Gehalt im Gesamtstaub Den geringsten Auswurf an Gesamtstaub (TPE) weist der Stückholzkessel auf. Der höchste hat das offene Feuer gefolgt vom Cheminéeofen (Fig. 7). Die Anteile EC/OC in den Gesamtstaubpartikeln variieren. Feuerungen mit schlechtem Ausbrand zeigen einen erhöhten EC-Anteil. Die Schnitzelfeuerung hat einen deutlich höheren Anteil an OC gemessen am TC, das allerdings auf einem insgesamt sehr tiefen Niveau an TC (Fig. 9). Dies ist mit der hohen Glutbetttemperatur in der Schnitzelfeuerung zu erklären, die die Verflüchtigung der mineralischen Komponenten in der Asche erhöht (z.B. Carbonate) was zu salzartigen Partikeln führt. Partikelauswurf und EC/OC-Gehalt in den Partikelklassen Den höchsten Ausstoss an Partikeln haben alle Feuerungen in dem Bereich der Ultrafeinpartikeln <0.4 µm (Fig. 11). Wie bei dem Gesamtstaub haben das offene Feuer und der Cheminéeofen den höchsten Staubauswurf in den einzelnen Partikelklassen. Stückholz- und Pelletkessel liegen ähnlich. Der Kohlenstoffanteil (TC) ist bei offenem Feuer und Cheminéeofen deutlich am höchsten. Alle Kessel haben nur einen geringen Anteil an TC (Fig. 12). Die EC-Emissionsfaktoren liegen in allen gemessenen Partikelklassen über denen von OC (Fig. 13, Fig.14). 6 Elemental and Organic Carbon in Flue Gas Particles of Various Wood Combustion Systems Gesamtstaubauswurf und Partikelgrössenklassen Die PM10-Partikel enthalten zwischen 83% und 98% der gesamten Staubmasse (Fig. 10). In der Grössenklasse PM2.5 ist dieser Anteil noch 73% bis 91%, bei PM1 61% bis 80% bezogen auf den gesamten Staubauswurf. Das offene Feuer weist in den Grössenklassen PM2.5 und PM1 mit 73% und 61% den geringsten Anteil auf. Dies lässt sich auf die unvollständige Verbrennung mit gröberen Verbrennungspartikeln zurückführen. Einfluss der Verbrennung auf TC, EC und OC Die Untersuchung am Stückholzkessel erlaubte die Unterscheidung von Betriebszuständen in zwei Testläufen. Ein Testlauf beinhaltete auch die Anfeuerphase des Kessels mit erhöhtem Anteil an Unverbranntem (Products of incomplete combustion, PIC). Betrachtet man die TC bzw. EC/OC-Emissionen stellt man fest, dass sich sowohl der Gesamtstaub wie der TC bei guter Verbrennung reduzieren. In Abbildung 15 (Fig. 15) zeigt sich, dass die OCEmissionen in etwa bei beiden Betriebszuständen (Anfeuern bzw. steady-state-Betrieb) gleich sind, die ECEmissionen aber deutlich bei besserer Verbrennung (CO low) zurückgehen. Schlussfolgerungen • Bei allen untersuchten Feuerungen (Schnitzel-, Pellet- und Stückholzkessel sowie Cheminéeofen und offenes Feuer) weisen Gesamtstaub (TPE) und Gesamtkohlenstoff (TC) die höchsten Emissionsfaktoren in der Partikelklasse <0.4 µm auf. • Zwischen 61% (offenes Feuer) und 80% (Stückholzkessel) des Gesamtstaubes werden mit der Partikelgrössenklasse PM1 emittiert. Die PM2.5-Partikel enthalten bereits 73% (offenes Feuer) bis 91% (Holzofen) des Gesamtstaubes, PM10 83% bis 98%. • Der Massenanteil des Gesamtkohlenstoffs (TC) im Gesamtstaub variiert zwischen 9% und 84%. • Der Anteil des elementaren Kohlenstoffes (EC) am Gesamtkohlenstoff (TC) liegt im Bereich von 20% - 70%, der des organischen Kohlenstoffes (OC) im Bereich von 27% - 79%. • Es gibt keinen eindeutigen Zusammenhang für den Split von EC/OC im Gesamtkohlenstoff. Ein Einflussfaktor kann die Verbrennungsgüte sein: bei erhöhtem Anteil an unverbrannten Produkten (PIC) im Abgas ist der ECAnteil am Gesamtkohlenstoff deutlich höher. • Das offene Feuer und der Cheminéeofen weisen für alle Werte (TPE, TC bzw. EC und OC) die höchsten Emissionsfaktoren auf. Die Gesamtstaubemissionen des offenen Feuers liegen um das 7-fache höher als die des Stückholzkessels, der die tiefsten TPE-Wert aufwies. • Der Stückholzkessel erzielte unter den Praxisbedingungen im Feld die Staubemissionsanforderungen des Schweizer Qualitätssiegels für Holzfeuerungen. Es gilt zu betonen, dass diese Anforderungen für Messungen unter Prüflaborbedingungen mit klaren Prüfprozederes und standardisiertem Brennstoff gelten. • Die Emissionsfaktoren für Gesamtstaub von Stückholz- und Pelletkessel erreichen unter Praxisbedingungen ähnliche Werte wie Dieselfahrzeuge mit EURO-4-Staubemissionsgrenzwerten. Auf Grund verschiedener Untersuchungen an Holzfeuerungen mit Partikelabscheider (Elektro- und Gewebefilter) kann gesagt werden, dass sich mit solchen Filtersystemen TPE-Emissionsfaktoren im Reingas von 1 - 6 mg/MJ erzielen lassen. • Die Untersuchungen haben einen klaren Praxischarakter mit entsprechenden Zufälligkeiten für die erhobenen Daten. Die ausgewiesenen Emissionsfaktoren sind statistisch nicht gesichert. Von Interesse wäre eine wissenschaftliche Untersuchung, um die Prozessparameter in der Holzverbrennung zu eruieren, die die TC- bzw. EC/OC-Bildung begünstigen bzw. senken. 7 Elemental and Organic Carbon in Flue Gas Particles of Various Wood Combustion Systems 1 Introduction Life cycle assessments of wood combustion processes show many advantages over fossil fuels. There is nevertheless a remarkable disadvantage looking at the wood combustion emission, particularly at particulate emissions contributing to the total suspended particles in the ambient air. Airborne particulate matter (PM) in the environment is of ever increasing concern to authorities and the public since it is known to be carcinogens and respiratory irritants. The major fractions of the particles in the wood combustion process are in the size less than 1 micron, typically in the range of 30 to 300 nm [1]. These nanoparticle fractions behave like gaseous effluents and penetrate deeply into the human respiratory tract [2]. The Swiss Federal Environment Protection Agency (BUWAL) wanted to know more about the carbon fraction in the particles of wood fire exhausts. Of specific interest is the content of the elemental carbon (EC) and organic carbon (OC) in the particles of different size since these substances are known for its particular potential as carcinogens. The wood combustion systems examined are typical wood combustion systems such as wood chip boiler, pellet boiler, wood log boiler, wood stove and an open fire. All but the open fire were tested in the field. Thus the results reflect practice condition. The open fire has been tested on a test rig in the laboratory. 2 Materials and Methods 2.1 Flue gas sampling 2.1.1 In stack measurement Three sampling lines are applied in the measuring section of the flue duct to determine particles, carbon dioxide (CO2) and carbon monoxide (CO). The set up (Figure 1) allows the simultaneous detection of the effluents during the test run. Flue gas TI Dry gas meter Filter Dryers Pumps TI TPE sampling train Dry gas meter Impactor Cooler O2 O2 Impactor sampling train CO CO2 Gaseous effluents sampling train Wood boiler/stove Figure 1. Set up for in stack emission measurement. Three sampling lines allow simultaneous analysis of total particle emissions (TPE), particle classes in an impactor and gaseous effluents. The arrangement of the dust probes for the flat filter and the impactor is shown in Figure 2. 8 Elemental and Organic Carbon in Flue Gas Particles of Various Wood Combustion Systems Figure 2. Sampling probes for dust collection in stack. The flat filter is shown with filter and back up filter and the electric heater. The third probe collects the gaseous effluents in the flue gases. The flue gas analyses detects the emission concentration. For the conversion of the emission concentration to emission factors the specific gas volume per unit dry fuel is required. The specific flue gas volume was calculated by carbon balance method which derives from CO2 and CO analysis in the flue gas [3]. 2.1.2 Dilution tunnel To analyse the particle emission of the open fire the method of constant volume sampling in a dilution tunnel was applied [4]. Flue gas and ambient air are mixed in the dilution tunnel and kept at constant flow conditions. The sampling train for particles is installed in the measuring section of the dilution tunnel (Figure 3). The particle sampling starts with the ignition of the fire and ends, when fuel is burnt. A weighing scale indicates the end of a burn cycle when all fuel is fired. The emission factors can be determined by the particle mass sampled over the entire burn cycle and the fuel mass burnt. Dilution tunnel TI Filter TPE sampling train FI Dry gas meter Exhaust Dryers Pumps TI Dry gas meter Blower Impactor Impactor sampling train Open fire Scale Figure 3. Test rig with dilution tunnel to analyse the total particle emissions (TPE) and particle classes in an impactor of the open fire. 9 Elemental and Organic Carbon in Flue Gas Particles of Various Wood Combustion Systems 2.2 Particle analysis The sampling of the particle classes was done by means of a particle sizing sampler (Figure 4). The sampler used is a multi-stage, multi-orifice cascade impactor [5]. The impactor classifies the particles collected according to the aerodynamic dimensions which are a valuable measure to assess the particles lung penetrability [2]. The 7 stages classify the particles between 0.4 and 9 microns aerodynamic diameter. Figure 4. Set up of impactor and handling of filter paper while loading the impactor stages. The paper is thermally conditioned for gravimetric particle analyses after loading. Parallel to the impactor total particle emission (TPE) were measured in the flue gases. The TPE-filter combines a filter and back up filter with a filter efficiency of 99.75%. The measurement uncertainty for the range used (20 - 300 mg/mn3) is 2%. All stage filters of the impactor were weighted for total particle emission to compare with the figures of the TPE-filters. The impactor particle mass values were adapted to the total TPE-emission so they summed up to equal mass of TPE-emissions. That implies that the impactor was solely used to define the fraction of the particle mass in the impactor stages. The combination of the particle distribution fraction in the impactor and the TPE lead to the mass distribution in the different particle classes. 2.3 Carbon analysis Carbon in combustion particles occurs in the organic fraction as elemental carbon (EC) and as organic carbon (OC). Different techniques are used to distinguish between EC and OC. The distinction between EC and OC is usually made by coulometric or thermo gravimetric analysis [6]. The coulometric analysis gives the content of EC for the non-volatile fraction and OC for the volatile fraction of organic carbon. Coulometry is the VDI reference method for determining EC and OC (VDI, 1996). A filter sample is heated in an oxygen flow to a temperature where all carbon is converted to CO2. The coulometer measures the conductivity of CO2 released by oxidation of EC in the samples at 650°C. The OC content is determined by pyrolyses in helium atmosphere. The total carbon (TC) is the sum of EC and OC analysed. 10 Elemental and Organic Carbon in Flue Gas Particles of Various Wood Combustion Systems The EC and OC were analysed in the TPE-filters as well as in the impactor stage filters. The EC/OC-values of the TPE-filter gives an answer of organic carbon from coarse (>10 µm) to ultra fine particles (< 1 µm) whereas the EC and OC values of the impactor filters relates EC and OC to the various particle classes. According to the SUVA-laboratories the accuracy of the coulometry is 2 - 7% depending on the filter load (Eurachem-Guide). 2.4 Wood combustion systems All wood combustion systems were tested in the field but the open fire, which was analysed in the laboratory of the institute. The heat output of the boilers and appliances were arbitrary according to the heat demand of the heating system. The emission figures thus have real live character. The following wood combustion systems were investigated for particles, EC, OC and TC resp. Table 1 Wood combustion systems analysed in the field tests. Type Fuel Heat output Test site Wood chip boiler wood chips 70 kW institute building Pellet boiler pellets 10-32 kW apartment house Wood log boiler wood logs 35 kW apartment house Stove wood logs 4 kW living room Open fire wood logs – laboratory test rig 3 Results 3.1 Coulometer Some pre-tests were performed to establish the coulometric analysis for the EC and OC content of the particles from wood combustion. A crucial parameter is the temperature level of the coulometer when detecting EC. To verify the optimal coulometer temperature two parts of the same and the one filter sample were measured at temperatures of 800°C and 650°C resp. (Table 2). Table 2 Verification of the temperature level of elemental carbon (EC) analytics in the coulometer. Phosphoric acid treatment of the samples leaching out carbonates which evaporate at temperature over 650°C. Samples No. Leaching Coulometer temperature EC mass [°C] [µg] Ia untreated 800°C 2164 Ib treated 800°C 346 IIa treated 650°C 51 IIb treated 800°C 63 One part of the sample was treated with phosphoric acid to leach the carbonates. Since carbonates evaporate over 650°C this can lead to misinterpretation of EC content since part of the particle mass escapes with the carbonates. The results show, that there is a considerable difference between leached and unleached samples analysed by coulometer at 800°C. The difference of the two samples in EC content is due to evaporation of car11 Elemental and Organic Carbon in Flue Gas Particles of Various Wood Combustion Systems bonates. If the sample is leached either 650°C or 800°C coulometer temperature show similar content of EC. After the pre-test it has been decided to run the coulometer at a temperature of 650°C for all samples. 3.2 Particle size distribution The multi-stage particle sizing sampler cascade impactor classifies the particles in 7 stages between 0.4 and 9 microns aerodynamic diameter. In a pre-test campaign the distribution of the particles were analysed in all 7 classes (Figure 5). The relevant classes in terms of particle emissions are those in the range of >9 µm and <1.1 µm. TPE-emissions in different particle classes TPE concentration [mg/m3n @ 13% O2] 80 60 40 20 0 .0 >9 .0 -9 5.8 .8 .7 -4 -5 4.7 3.3 .3 -3 2.1 .1 -2 1.1 .1 -1 0.7 .7 -0 0.4 0.4 0- Particle size [m] Figure 5. Total particle emission (TPE) of wood chip boiler measured in the various impactor stages. There is a similar distribution as TPE for the TC emission (Figure 6). The major part of the said emission are in the range of particles <1.1 µm. EC-OC-TC-emissions in differen particle clases 2.50 TPE concentration [mg/m3n @ 13% O2] OC EC TC 2.00 1.50 1.00 0.50 0.00 .0 >9 .0 -9 5.8 .8 -5 4.7 .7 -4 3.3 .3 -3 2.1 .1 -2 1.1 .1 -1 0.7 .7 -0 0.4 0.4 0- Particle size [m] Figure 6. Distribution of EC, OC and TC emission in the different particle classes. It was decided to analyse the impactor stages of >9 µm, 2.1-3.3 µm and <0.4 µm in the main measurement campaign. 12 Elemental and Organic Carbon in Flue Gas Particles of Various Wood Combustion Systems 3.3 Emission factors The emission factors are calculated from standardised particle concentration at 13% oxygen measured in the flue gas. The emission factors are based on the dry fuel mass. Since the wood combustion systems were tested in the field the emission figures have real live character. They give some indication of the EC/OC characteristic even though the data are not statistically proven. Each wood combustion systems was tested over two runs. The data in Table 3 are mean values of the two test runs. The figures are given as TPE, TC, EC and OC derived of TPEfilter measurement as well as impactor measurement in the particle classes of > 9 µm, 2.1 - 3.3 µm and < 0.4 µm. Table 3 Emission factors for total particle, total carbon, elemental carbon and organic carbon of the wood combustion systems analysed based on fuel dry. TPE: total particle emissions, PE: particle emission, TC: total carbon, EC: elemental carbon, OC: organic carbon. The measurement uncertainty for TPE is 2%, for EC/OC 4.5%. Emission factors Wood combustion system Wood chip boiler Pellet boiler Wood log boiler Wood stove Open fire Particle class PE TC EC OC [µm] [g/kg] [g/kg] [g/kg] [g/kg] TPE 1.450 0.129 0.027 0.101 % 100% 8.9% 1.9% 7.0% >9 0.138 0.010 0.007 0.003 2.1-3.3 0.042 0.002 0.0004 0.001 < 0.4 0.704 0.027 0.003 0.024 TPE 0.825 0.593 0.401 0.192 % 100% 71.9% 48.6% 23.3% >9 0.028 0.009 0.007 0.003 2.1-3.3 0.048 0.024 0.020 0.004 < 0.4 0.387 0.157 0.145 0.012 TPE 0.621 0.463 0.197 0.265 % 100% 74.5% 31.8% 42.7% >9 0.029 0.010 0.005 0.004 2.1-3.3 0.024 0.009 0.005 0.004 < 0.4 0.336 0.096 0.067 0.029 TPE 2.201 2.162 1.501 0.661 % 100% 83.9% 62.3% 21.6% >9 0.066 0.037 0.032 0.005 2.1-3.3 0.096 0.067 0.061 0.006 < 0.4 1.126 0.770 0.737 0.033 TPE 4.520 2.724 1.998 0.727 % 100% 60.3% 44.2% 16.1% >9 0.365 0.154 0.094 0.059 2.1-3.3 0.293 0.133 0.080 0.052 < 0.4 1.476 0.857 0.574 0.283 The airborne particle are classified according to their size. The commonly used classes are PM10 (< 10 µm), PM2.5 (< 2.5 µm) and PM1 (< 1 µm). With the gravimetric data of the impactor stage filters it is possible to estimate the emission factors of the particle classes PM10, PM2.5 and PM1. The share of the impactor 13 Elemental and Organic Carbon in Flue Gas Particles of Various Wood Combustion Systems stages were summed up according to the PM-class. PM1 was calculated from the impactor stages <0.4 µm, 0.40.7 µm and 0.7-1.1 µm. For the PM2.5 the stages 1.1-2.1 µm and 2.1-3.3 µm were added. The stages 3.3-4.7 µm, 4.7-5.8 µm and 5.8-9 µm were additionally used to get PM10 figures. Table 4 PM-emission factors of the wood combustion systems analysed based on fuel dry. TPE: total particle emissions; PM10: particles with an aerodynamic diameter <10 µm, PM2.5: aerodynamic diameter <2.5 µm, PM1: aerodynamic diameter <1 µm. Emission factors Wood combustion system Wood chip boiler Pellet boiler Wood log boiler Wood stove Open fire Particle class TPE PM10 PM2.5 PM1 [g/kg] 1.45 1.29 1.135 0.997 % 100% 83.3% 78.2% 68.8% [g/kg] 0.825 0.805 0.774 0.638 % 100% 97.6% 90.2% 77.3% [g/kg] 0.659 0.63 0.582 0.526 % 100% 95.6% 88.2% 79.8% [g/kg] 2.2 2.12 2 1.75 % 100% 96.3% 90.8% 79.6% [g/kg] 4.52 4.0 3.3 2.7 % 100% 89.4% 72.7% 60.8% 4 Discussion 4.1 EC- and OC-content in total particle emissions The TPE-emission factors show lowest values for wood log boiler followed by pellet and wood chip boiler (Figure 7). The open fire and the wood stove have the highest emission factors. The correlation of EC and OC in the TPE is varying: the domination of elemental carbon for the open fire and the stove indicates a combustion with products of incomplete combustion (PIC). These test runs show also high emission factors for carbon monoxide (CO). The middle field of the EC- and OC-emission are for wood log and pellet boiler. They burned with fairly complete combustion emitting little EC and OC. The emission characteristic of wood chip boiler indicates a complete combustion at a high combustion temperature. The salty components of the minerals in the ash evaporate due to elevated charcoal bed temperature as carbonates. The particles are aerosols from these condensed inorganic carbon components. The low content of EC and OC demonstrates the complete combustion in the wood chip boiler. 14 Elemental and Organic Carbon in Flue Gas Particles of Various Wood Combustion Systems Figure 7. Emission factors of total particle emission (TPE). The elemental (EC) and organic carbon (OC) was analysed from TPE-filters. The content of TC as a fraction of TPE is in the range of 60% to 80% (Figure 8). Just the wood chip boiler is an exemption with a content of less than 9%. The EC fraction is between 30% and 60% of the TPE, OC 10% to 40%. Fraction of EC/OC/TC based on TPE 100% Percentage [%] 80% 60% 40% 20% 0% Wood chip boiler Pellet boiler Wood log boiler EC Figure 8. OC Stove Open fire TC Fraction of elemental carbon (EC), organic carbon (OC) and total carbon (TC) based on total particle emissions (TPE). The share of EC and OC in the TC is indicated in Figure 9. The content of EC in the TC spreads in the range of 21% - 74%, the one of OC 26% - 78%. 15 Elemental and Organic Carbon in Flue Gas Particles of Various Wood Combustion Systems Fraction of EC/OC based on TC 100% Percentage [%] 80% 60% 40% 20% 0% Wood chip boiler Pellet boiler Wood log boiler EC Figure 9. 4.2 Stove Open fire OC The share of elemental (EC) and organic carbon (OC) in the total carbon (TC) for total particle emission (TPE). Particle emissions in the different particle sizes PM10 contains 83% (wood chip boiler) up to 98% (pellet boiler) of the total particle mass (Figure 10). PM2.5 holds 73% (open fire) to 91% (stove) of TPE. PM1 still includes 61% (open fire) to 80% of total particle mass. This indicates that the mayor part of dust particles are of ultra fine size in the flue gas of wood combustion processes. Figure 10. The particle emission in the particle sizes PM10, PM2.5 and PM1 as fraction of total particle emissions (TPI) The open fire with more products of incomplete combustion shows a higher content of coarser particles. The particle emission of PM1 are just 61%, PM2.5 just 73% of the total particle emission. 4.3 EC- and OC-emissions in the different particle classes In all tests the impactor stages >9 µm, 2.3-2.5 µm and <0.4 µm were analysed for EC, OC and TPE. Figure 11 summarises the emission factors of the classes for total particles. 16 Elemental and Organic Carbon in Flue Gas Particles of Various Wood Combustion Systems Emission factor [g/kg fuel dry] TPE-emission in different particle classes 1.6 1.4 1.2 1.0 0.8 0.6 Wood chip boiler Pellet boiler Wood log boiler Stove Open fire 0.4 0.2 0.0 PC >9 Figure 11. PC 2.3-2.5 PC 0-0.4 Total particle emission (TPE) in different particle classes. The open fire has highest TPE-emission in all classes. All wood combustion systems have most TPEemission in the ultra fine fraction of particle classes <0.4 µm. Highest TC-emission occurs like the TPE-emission in the particle class <0.4 µm (Figure 12). Emission factor [g/kg fuel dry] TC-emission in different particle classes 1.0 0.8 0.6 Wood chip boiler Pellet boiler Wood log boiler Stove Open fire 0.4 0.2 0.0 PC >9 Figure 12. PC 2.3-2.5 PC 0-0.4 Total carbon emission (TC) in different particle classes. When EC- and OC-emission are compared over the various classes EC dominates the emission (Figure 13). 17 Elemental and Organic Carbon in Flue Gas Particles of Various Wood Combustion Systems EC-emission in different particle classes Emission factor [g/kg fuel dry] 1.0 0.8 0.6 Wood chip boiler Pellet boiler Wood log boiler Stove Open fire 0.4 0.2 0.0 PC >9 Figure 13. PC 2.3-2.5 PC 0-0.4 Elemental carbon emission (EC) in different particle classes. OC is emitted at a lower level than EC in all classes (Figure 14). Emission factor [g/kg fuel dry] OC-emission in different particle classes 1.0 0.8 0.6 0.4 Wood chip boiler Pellet boiler Wood log boiler Stove Open fire 0.2 0.0 PC >9 Figure 14. 4.4 PC 2.3-2.5 PC 0-0.4 Organic carbon emission (OC) in different particle classes. Products of incomplete combustion The test runs with the wood log boiler comprised two phases: one test was with a start phase including the ignition of the wood logs. The later was a phase under normal operating conditions. The start phase had periods with incomplete combustion. Thus the CO emission factor was 117 g/kg fuel dry (10'400 mg/m3n @ 13% O2). The value for CO under normal operating conditions was 35 g/kg (3'050 mg/m3n @ 13% O2). Total carbon (TC) is clearly reduced (Figure 15). Where OC remains almost identical EC is low for the combustion with little product of incomplete combustion (PIC). 18 Elemental and Organic Carbon in Flue Gas Particles of Various Wood Combustion Systems Emission factor [g/kg fuel dry] Products of incomplete combustion 1 0.8 0.6 0.4 0.2 0 Test run 1 CO high TPE Figure 15. 4.5 Test run 2 CO low EC OC TC Elevated products of incomplete combustion (PIC) in a test run is indicated by high carbon monoxide (CO) and low PIC with low CO resp. Comparison of particle emissions In Table 5 the total particle emissions are compared with other particle sources. Since the emissions measured in the combustion system has some arbitrary character, there are ranges given. All data derive from field measurements done by our institute. The European standards for Diesel vehicles are listed as emission factors based on net heating value of fuel. This allows the comparison of wood emission with Diesel engines as an other source of fine particles. Table 5. Emissions of various particle sources. Diesel vehicles emission factors are based on European emission standards (EURO-3 and Euro-4). There are also the requirements of the Swiss quality label for wood heat boilers and appliances. TPE-Emission factor based on fuel net heating value TPE-Emissions Swiss quality label requirement TPE-Emission [mg/MJ] 4) [mg/mn3] @ 13% O2 [mg/mn3] @ 13% O2 Wood chip boilers 42 - 85 62 - 125 90 Pellet boilers 19 - 48 28 - 71 60 Wood log boilers 16 - 39 24 - 57 60 Wood log boilers with flue gas filter 1 - 5.8 1.5 - 8.5 Wood stoves 24 -129 35 - 190 100 Open fire 265 390 - Light duty vehicles (EURO 3) 1) 22.7 - - Heavy duty vehicles (lorries, EURO-3) 2) 11.4 - - Light duty vehicles (EURO-4) 1) 11.4 - - Heavy duty vehicles (lorries, EURO-4) 2) 2.3 - - Source 1) Diesel fuel consumption: 6 liter/100 km 2) Diesel fuel consumption: 205 g/kWh 3) specific flue gas volume: 11.6 mn3/kg @13%O2 4) net heating value of wood: 17 MJ/kg 19 Elemental and Organic Carbon in Flue Gas Particles of Various Wood Combustion Systems In Table 5 the total particle emissions are compared with the requirement of the Swiss quality label for wood heat boilers and appliances. The Swiss label requirements give some indications what the threshold values of clean air requirements for areas suffering from air pollution could be. Wood log and pellet boilers have the potential to achieve similar emission values for TPE as Diesel vehicles according to EURO-3 standard. Wood chip boilers with dust filter systems such as electrostatic precipitators or fabric filters [7] have TPE-emission lower than EURO-3 and -4 from Diesel vehicles. 5 Conclusions The distribution of total particle emission (TPE) over the particle classes show highest emission factors in the particle class <0.4 µm. This is also true for total carbon (TC) emission as the sum of elemental (EC) and organic carbon (OC). The content of TC in the TPE varies between 9% up to 84%. The share of EC and OC fraction in the TC is in the range of 20% to 70% for EC and 27% to 79% for OC. There is no correlation of the EC and OC share in the TC: there are values measured for higher EC content as OC and visa versa. The crucial parameters seem to be the grade of complete combustion and charcoal bed temperature for EC and OC formation. Little products of incomplete combustion also result in low EC-emission factors. PM10 contains 83% (wood chip boiler) up to 98% (pellet boiler) of the total particle mass. PM2.5 holds 73% (open fire) to 91% (stove) of TPE. PM1 still includes 61% (open fire) to 80% of total particle mass. This indicates that the mayor part of dust particles are of ultra fine size in the flue gas of wood combustion processes. The open fire and the wood stove have the highest emission factors for TPE, EC and OC. The average TPEemission factor of the open fire is 7 fold higher than lowest emission of wood log boiler. The wood log boiler achieved in the field the requirements of the Swiss quality label. It must be stressed that the requirements are tailored to emission of type tests under test lab condition. Wood log and pellet boilers can achieve TPE-emission values in the range of EURO-3 Diesel vehicles. The TPE-emission figures of the wood chip boilers and wood log appliances are higher. The particle emissions of wood combustion can be significantly reduced by dust filter systems such as electrostatic precipitators or fabric filters. 20 Elemental and Organic Carbon in Flue Gas Particles of Various Wood Combustion Systems Acknowledgements The analytical work of the EC and OC coulometry is greatly acknowledged to the labs of the Swiss National Accident Insurance Fund (SUVA), Analytical Section. The Swiss Federal Environment Protection Agency (BUWAL) granted the research work. References [1] Hueglin C., Gaegauf C., Kuenzel S., Burtscher H.:Characterization of Wood Combustion Particles. Morphology, Mobility and Photoelectric Activity. Environmental Science and Technology, 1997. [2] Wells, W.F.: Airborne Contagion and Air Hygiene, Harvard University Press, Cambridge, Massachusetts, 1955. [3] P.E. Denyer, D. Wilkins: Further investigations into the use of a flue loss method for measurement of the thermal efficiency of wood-burning roomheaters, CRE Group Ltd., Gloucestershire GB, F&AT Report no. 70, 1994. [4] C. K. Gaegauf, Y.Macquat: Comparison of draft ISO- and CEN-standard to determine efficiency and emission of solid fuel burning appliances, 1st World Conference and Exhibition on Biomass for Energy and Industry, Sevilla, 2000. [5] Andersen 1 ACFM Particle sizing sampler. [6] Physical characterization of particulate emissions from diesel engines: a review H. Burtscher, Journal of Aerosol Science, Elsevier, in press. [7] C. K. Gaegauf, M.P. Schmid, Andres Jenni: Wirksamkeit neuer Abscheidetechniken in Holzfeuerungsabgasen bezüglich Feinpartikeln, Final Report of Swiss Environment Protection Agency (BUWAL), 2003. 21