QUT Digital Repository: http://eprints.qut.edu.au/

This is the authors’ version of the following conference paper:

Marsh, Jack, Morawska, Lidia, Ristovski, Zoran, & Wardoyo, Arinto (2005) Particle Number and Mass Emission Factors from Combustion of Wood Samples Collected from Forest. In: Towards a New Agenda: Proceedings of the 17th International Clean Air & Environment Conference 2005, 3 May ‐ 6 May 2005, Australai, Tasmania, Hobart.

© Copyright 2005 Please consult the authors.

PARTICLE NUMBER AND MASS EMISSION FACTORS FROM COMBUSTION OF QUEENSLAND FOREST VEGETATION SAMPLES Arinto Y.P. Wardoyo, Lidia Morawska, Zoran D. Ristovski, and Jack Marsh International Laboratory for Air Quality and Health Queensland University of Technology, GPO Box 2434, , Queensland 4001, Australia

Abstract Knowledge of particle emission characteristics associated with forest fires and in general, biomass burning, is becoming increasingly important due to the impact of these emissions on human health. Of particular importance is developing a better understanding of the size distribution of particles generated from forest combustion under different environmental conditions, as well as provision of emission factors for different particle size ranges. This study was aimed at quantifying particle emission factors from four types of wood found in South East Queensland forests: Spotted Gum (Corymbia citriodora), Red Gum (Eucalypt tereticornis), Blood Gum (Eucalypt intermedia), and Iron bark (Eucalypt decorticans); under controlled laboratory conditions. The experimental set up included a modified commercial stove connected to a dilution system designed for the conditions of the study. Measurements of particle number size distribution and concentration resulting from the burning of woods with a relatively homogenous moisture content (in the range of 15 to 26 %) and for different rates of burning were performed using a TSI Scanning Mobility Particle Sizer (SMPS) in the size range from 10 to 600 nm and a TSI Dust Trak for PM2.5. The results of the study in terms of the relationship between particle number size distribution and different condition of burning for different species show that particle number emission factors and PM2.5 mass emission factors depend on the type of wood and the burning rate; fast burning or slow burning. The average particle number emission factors for fast burning conditions are in the range of 3.3 x 1015 to 5.7 x 1015 particles/kg, and for PM2.5 are in the range of 139 to 217 mg/kg. Keywords: Biomass burning, ultrafine particles, particle number emission.

1. Introduction biomass (Hueglin et al., 1997). Forest fires as a The burning of biomass emits significant amounts type of biomass burning contribute significantly to of particles that impact on human health and are the particle burden of the atmosphere. linked to morbidity and mortality (Dockery et al., The state of Queensland has a problem with 1993; Samet et al., 2000a; Samet et al., 2000b). forest fires. The state experienced its worst forest Characterization of particle emissions has been fires in 1991, burning 37,000 hectares (Hamwood, conducted in previous studies for several types of 1992). The state still suffers from forest fires from biomass burning, such as: wood burning (Hays et year to year. In the last several months there were al., 2002; Hedberg et al., 2002), grassland burning a number of such fires in Queensland, for example: (Dennis et al., 2002), agriculture product forest fires at San Fernando and Canugra in July (Anderson, 2002; Reddy and Venkataraman, 2002), 2004; at Wallaby Hill Mudgeeraba, Gold Coast, and biofuel briquette (Mukherji et al., 2002). Other Minden, Gilston/Tallai Range, and Tamborine in factors that influence the characteristics of the August 2004; at Tamborine, Lowry/Hinze Dam, and emissions have also been studied, these include: Nerang in October 2004. Characterizing the particle moisture content related to completeness of emissions from forest fires is an important issue in burning (Leese et al., 1989; Todd, 1991), and rate providing better information related to their impacts of burning associated with the supply of oxygen for on human health. This paper presents an initial completeness of burning and the amount of burned study that is aimed at characterizing the particle emissions from burning vegetation samples typically growing in South East Queensland. probe of 1 cm in diameter. The sample flow was Additional emphasis is given to developing a better introduced to an ejector dilutor (Dekati) where it understanding of the size distribution and emission was diluted 10 times with heated, compressed, factors of ultrafine particles generated under particle-free air to obtain a dry, diluted sample and different environmental conditions. The to prevent further coagulation. The concentration at quantification of the emissions is conducted under the inlet of the ejector dilutor was very high and controlled conditions in the laboratory. caused frequent blockage of the dilutor inlet. This was especially observed during slow burning. The 2. Experiment Methods dilutor had to be cleaned before each new experiment. The sample flow was next mixed with a 2.1. Experimental Setup constant flow of ambient air filtered by a HEPA filter in a dilution tunnel to further reduce the concentration below 106 per cm3. The flow rate and temperature of the samples in the dilution tunnel Heater were measured using an air velocity meter. The HEPA filter Diluter 10X 1 tunnel temperature was observed to be between Thermocouple 0 cm

Flue gas 28°C and 30°C. Thermocouple analyser Dust The dilution ratio was calculated by measurement SMPS Track Hepa Filter 4 of the concentrations of CO2 in the flue, the dilution 0cm Pump Dilution tunnel tunnel, and the ambient air that was used for

5 5cm dilution. The concentration of CO2 in the flue was Q trak Air velocity meter continuously measured using a flue gas analyser. At the same time the concentration of CO2 in the

7 dilution tunnel and ambient air were recorded using Air flow 4.5 cm a TSI Q Trak. The dilution ratio was found to vary Valve Air velocity Sample flow meter from 150 to 200 for fast burning and from 100 to Blower 300 for slow burning and depended on the type of Figure 1. Experimental set up. wood burned.

The experimental setup is schematically shown in 2.4. Particle Measurement System Figure 1. The system consists of a stove, dilution The number concentration and size distribution of and sampling system, and a particle measurement the particles produced during the burning process system. were measured using a Scanning Mobility Particle Sizer (SMPS). The SMPS consisted of a 3071 TSI 2.2. Burning System electrostatic classifier and a 3010 condensation Stoves have been used previously as burning particle counter. The SMPS was operated in a systems to characterize biomass burning emissions window of 10 – 600 nm. The sheath air flow was (Todd, 1991; McDonald et al., 2000; Hedberg et al., selected as 4 lpm and the sample flow was 0.4 lpm. 2002). A modified commercial stove with The scanning time and retrace time were set at 120 dimensions of 66 x 74.5 x 55 cm3 was used as a s and 60 s respectively. The continuous burning system in this study. The part of the stove measurement of the total particle number originally used to adjust air flow was replaced by a concentration was determined using a ventilation system that enabled the introduction of a condensation particle counter (CPC) TSI 3022, which is able to measure particle concentrations of controlled amount of air into the stove. In order to 6 3 obtain a homogeneous rate of air flow, the outlet of more than 10 particle/cm . The sampling interval the ventilation system was connected to a was 20 s. Approximation of mass concentration of rectangular hood. The hood was connected to a fine particles (PM2.5) was measured by a TSI dust blower with a maximum capacity of 14 L/s through Trak using a PM2.5 inlet. The sampling time was 10 a pipe 30 mm in diameter. The connection was s. adjusted by a valve to control the flow rate of air. The purpose of the system was to simulate the 2.5. Sample Material and Preparation burning rates that would occur in forest fires. The samples consisted of different species of wood collected from the trees growing in a forest in Mount 2.3. Dilution and Sampling System Samson, located about 40 km from Brisbane City. The sampling system was a modified version of a Four species of Eucalypt were selected as samples system previously used for vehicle emission according to their prevalence in the forest. These measurements (Morawska, 1998; Ristovski, 1998). were: Spotted gum (Eucalypt citriodora), Red gum The samples were taken from the flue through a (Eucalypt.tereticornis), Blood wood (Eucalypt intermedia) and Iron bark (Eucalypt decorticans). All four species are known as hard woods. The sample. Pieces of wood were used as kindling for hardest of these is the Red gum, followed in order burning the following samples. The second and of decreasing hardness by Blood gum, Iron bark subsequent samples were logged in the stove just and then Spotted gum. The logs of wood were before burning of the previous sample finished. The placed in an open area of the laboratory for several processes of burning; such as igniting, flaming, and months. They were cut into pieces of 15 - 25 cm in smouldering (Simoneit, 2002); were observed length with diameter of 5 – 12 cm. The logs were through the window glass of the stove and recorded treated to obtain moisture contents that were as every 3 minutes to see the relationship between the homogeneous as possible throughout the samples processes and the particle size distribution, and that were also in the range of optimum concentration and emission factors. Temperature moisture for burning of 20 – 30 % (Core et al., measurements were conducted by attaching a K- 1982; Core et al., 1984). Moisture content was type thermocouple where the probe was inserted measured by a moisture meter. The moisture through a hole that was positioned 50 cm in the contents of the samples varied for different wood flue. The thermocouple was connected to a data types, from 18– 26 % for Spotted gum, 15-26 % for logger through an analogue to digital converter. The Red gum, 14-24 % for Blood gum, and 17-25 % for temperature was recorded every 30 seconds during Iron bark. The wood samples were weighted to the burning process. about 2 kg and their moisture was measured before they were burned in the stove. The unburned wood 3. Results and Discussion and ash were weighed after the burning finished. 3.1. Particle Size Distribution 2.6 Burning Conditions a) The wood samples were burned in the stove under 4.50E+08 two different conditions of burning called ‘fast Spotted Gum Igniting 4.00E+08 Flaming burning’ and ‘slow burning’. During fast burning the Smouldering stove was connected to a blower that introduced 3.50E+08 fresh air with a rate of 840 lpm, whilst the valve was 3.00E+08 fully open to let air into the stove with maximum 2.50E+08 velocity. Under slow burning conditions, the stove 2.00E+08 (particles/cm3) was not connected to the blower. In this condition 1.50E+08

the air supply through the ventilation system during Particle Number Concentration 1.00E+08 the burning process was not forced. 5.00E+07 The air velocity in the base of the stove for the 0.00E+00 fast burning condition was measured at several 1 10 100 1000 Diameter (nm) points using an air velocity meter while the door of b) the stove was closed. The distribution of the air 1.00E+08 velocity is presented in Figure 2. Igniting Spotted Gum Flaming 9.00E+07 Smouldering 8.00E+07

7.00E+07

30 6.00E+07 1.5

5.00E+07 25

4.00E+07 (particles/cm3) 20 3.00E+07

Particle Number Concentration Concentration Number Particle 2.00E+07 15 1.7 1.8 2.0

1.00E+07

Length (cm) 10 0.00E+00 1 10 100 1000 5 Diameter (nm)

0 0 1020304050 Figure 3 Typical particle size distributions for a) Width (cm) fast and b) slow burning of Spotted gum.

Figure 2 Velocity distribution of air in the base of Figure 3 shows typical size distributions for a) fast the stove for fast burning (in m/s). burning and b) slow burning of Spotted Gum during ignition, flaming and smouldering processes. The The burning of the wood samples was replicated nature of the particle size distributions is mono continuously 5 times for fast burning and 3 times for modal with most of the values for particle number slow burning. To start the burning, newspaper and centred at 30 to 40 nm. The diameter of the small pieces of the same wood types were used as particles is higher during the igniting process, a starter and kindling for the burning of the first associated with the beginning of the burning, and is during fast burning. Red gum, Blood gum, and Iron about 50 nm for fast burning and 120 nm for slow bark, having relatively the same hardness, burning. statistically produce particles in the same concentrations during fast burning. On the other hand, Spotted gum emits particles at the highest 3.2. Particle Number Concentrations and concentration of (3.3 ± 0.9)×107 particles/cm3 Emission Factors during slow burning. The three other samples emit particles in similar concentrations of around 4×106 1.20E+09 3 Red Gum fast burning particles/cm .

1.00E+09 160 Fast burning Slow burning 140

8.00E+08

120

6.00E+08 100 (particles/cm3) 80 4.00E+08

Particle Number Concentration Concentration Number Particle 60

2.00E+08

40 PM2.5 Mass Concentration (mg/m3) 0.00E+00 20 0 2000 4000 6000 8000 10000 12000 14000 Time (seconds) 0 Figure 4 Total number concentration for fast Spotted Gum Red Gum Blood Gum Iron Bark Type of Wood burning of Red gum. Figure 6 Average PM2.5 concentration for the Figure 4 shows an example of the change in different types of wood burned. particle number concentration for four repeated runs during fast burning of Red gum. The average PM2.5 average concentrations per burned particle concentration is calculated by integrating sample during fast and slow burning are shown in the total concentration during one and dividing Figure 6 for the different wood types. The PM2.5 by the burning time. A similar calculation was concentrations do not show any statistically conducted for the PM2.5 measured by the dust Trak. significant differences between different wood types 3 for fast burning. These values are: 4.6 ± 1.2 mg/m 3 1.800E+08 for Spotted gum, 4.2 ± 1.2 mg/m for Red gum, 5.0 Fast burning 3 3 Slow burning 1.600E+08 ± 1.5 mg/m for Blood gum, and 3.1 ± 0.6 mg/m for

1.400E+08 Iron bark. However during slow burning, Spotted gum emits particles at a higher concentration 1.200E+08 (113.3 ± 37.73 mg/m3) than any other wood types 1.000E+08 (22.2 ± 12.2 mg/m3 for Red gum, 6.8 ± 4.6 mg/m3 8.000E+07 for Blood gum, and 12.2 ± 5.8 mg/m3 for Iron bark). 6.000E+07 7.0E+15

4.000E+07 6.0E+15

Particle (particles/cm3) Concentration Number 2.000E+07

5.0E+15 0.000E+00 Spotted Gum Red Gum Blood Gum Iron Bark Type of Wood 4.0E+15 Figure 5 Average particle number concentration 3.0E+15 for the different types of wood burned.

2.0E+15 Figure 5 shows the average concentration of 1.0E+15 particles produced by the burning of one of each of (particles/kg) Factor Emission Number Particle

0.0E+00 the samples of wood for both fast and slow burning Spotted Gum Red Gum Blood Gum Iron Bark conditions. The standard deviation is derived from 4 Type of woods repeated runs. As can be seen from Figure 5, fast Figure 7 Total particle number emission factors burning produces higher average particle per kg of burned wood. concentrations than slow burning. The measured average concentrations were (8.6 ± 1.9)×107 Figure 7 presents the particle number emission particles/cm3 for Spotted gum, (1.1 ± 0.2)×108 factors per 1 kg of burned wood for fast burning of particles/cm3 for Red gum, (1.2 ± 0.2)×108 the different wood types. The particle number particles/cm3 for Blood gum, and (1.4 ± 0.2)×108 emission factor of Spotted gum was (3.2 ± 15 particles/cm3 for Iron bark. Spotted gum, which is 0.4)×10 particles/kg, which is the smallest the softest wood, emits the lowest particle number emission factor compared with the other samples. concentration compared to the other three samples Red gum, Blood gum, and Iron bark emit particles with emission factors of (5.7 ± 0.7)×1015 References particles/kg, (5.1 ± 0.6)×1015 particles/kg, and (5.7 Anderson, N. D., R; (2002), Airborne reduced ± 0.4)×1015 particles/kg respectively: the three nitrogen: ammonia emissions from agriculture samples produce particles in statistically the same and other resources, Environment International, amounts. 1011, 1 - 10. 350 Core, J. E., Cooper, J. A., DeCaesar, R. T. and Houck, J. E. (1982), Residential wood 300 combustion study.

250 Core, J. E., Cooper, J. A. and Neulicht, R. M. (1984), Current and projected impacts of

200 residential wood combustion on Pacific Northwest air quality, Journal Air Pollution 150 Control Association, 34, 138 - 143. Dennis, A., Fraser, M., Anderson, S. and Allen, D. 100 (2002), Air pollutant emissions associated with PM2.5 emisssion factor (mg/kg) factor PM2.5 emisssion forest, grassland, and agricultural burning in 50 Texas, Atmospheric Environment, 36, 3779- 3792. 0 Spotted Gum Red Gum Blood Gum Iron Bark Dockery, D. W., Pope, A., Xu, X., Spengler, J., D, Type of Wood Ware, J. H., Fay, M. E., Ferris, B. G. and Figure 8 PM2.5 emission factors per kg of burned Speizer, F. E. (1993), Mortality risk of air wood. pollution: a prospective cohort study, New Journal of Medicine, 329, 1753-1759. In terms of PM2.5, as shown in Figure 8, Fine, P. M., Cass, G. R. and Simoneit, B. R. T. statistical analysis shows that fast burning produces (2002), Chemical characterization of fine particle PM2.5 in the same amounts across the selected emissions from the fire place combustion of samples with emission factors of 168 ± 52 mg/kg woods grown in the Southern United States, for Spotted gum, 217 ± 69 mg/kg for Red gum, 208 Environ. Sci. Technol, 36, 1442-1451. ± 77 mg/kg for Blood gum, and 139 ± 23 mg/kg for Fine, P. M., Cass, G. R. and Simonet, B. T. (2001), Iron bark. The study of an open burning of mixed Chemical characteristization of fine particle hardwood forest foliage in the US showed that the emissions from fireplace combustion of woods PM2.5 emission factor was 10.8 ± 3.9 g/kg (Hays et grown in the Notheastern United States, al., 2002). Fine et al (2002) studied PM2.5 emissions Environ. Sci. Technol, 35, 225-2675. from fireplace combustion of wood grown in the Hamwood, K. R. (1992), Large forest plantation fire Northeastern and Southern United States (Fine et in Queensland, IFFN, 7, 2-3. al., 2001; Fine et al., 2002). They measured that Hays, M. D., Geron, C. D., Linna, K. J., Smith, N. D. the emission factors were in the range of 2.7 to 6.8 and Schauer, J. J. (2002), Speciation of gas- g/kg for hard wood. The differences between our phase and fine particle emissions from burning results and previously reported ones could be due of foliar fuels, Environment Science and to the different wood types considered as well as Technology, 36, 2281-2294. the different burning conditions applied. Hedberg, E., Kristensson, A., Ohlsson, M., Johansson, C., Johansson, P.-A., Swietlicki, E., 4. Summary Vesely, V., Wideqvist, U. and Westerholm, R. This paper presents a preliminary study aimed (2002), Chemical and physical characterization at quantifying particle number and mass emission of emissions from birch wood combustion in a factors from the combustion of trees growing in wood stove, Atmospheric Environment, 36, South East Queensland forests. Emissions from 4823-4837. four types of woods were quantified in terms of Hueglin, C. H., Gaegauf, C. H., Kunzel, S. and particle number and PM2.5 mass emissions for Burtscher, H. (1997), Characterization of wood different condition of burnings, specifically fast combustion particles: morphology, mobility, and burning and slow burning. The results show that the photoelectric activity, Environmental Science softest wood emits particles with the smallest and Technology, 31, 3439-3447. number concentration and the smallest emission Leese, K. E., Harkins, S. M. and MCCrillis, R. C. factors. On the other hand, it produces the highest (1989), Effects of burn rate, wood species, concentration of PM2.5. The emission factors of moisture content and weight of wood loaded on PM2.5 under fast burning conditions are the same woodstove emissions, EPA 600/2-89/025; U.S. for all four types of wood. Environmental Protection Agency, Washington, DC. McDonald, J. D., Zielinska, B., Fujita, E. M., Sagebiel, J. C., Chow, J. C. and Watson, J. G. (2000), Fine particle and gaseous emission rates from residential wood combustion, Environmental Science and Technology, 34, 2080-2091. Morawska, L., Bofinger, N.D., Kosic, L., and Nwankwoala, A. (1998), Submicron and supermicron particles from diesel vehicles emissions, Environment Science and Technology, 32, 2033 - 2042. Mukherji, S., Swain, A. K. and Venkataraman, C. (2002), Comparative mutagenicity assessment of aerosols in emissions from biofuel combustion, Atmospheric Environment, 36, 5627-5635. Reddy, M. S. and Venkataraman, C. (2002), Inventory of aerosol and sulphur dioxide emissions from India. Part II--biomass combustion, Atmospheric Environment, 36, 699- 712. Ristovski, Z., Morawska, L., Bofinger, N.D. (1998), Submicrometer and supermicrometer particles from spark ignition vehicle emissions, Environment Science and Technology, 345, 3845 - 3852. Samet, J. M., Dominici, F., Zeger, S. L., Schwartz, J. and Dockery, D. W. (2000a) Health Effects Institute Research Research Report 94, Part I. Samet, J. M., Zeger, S. L., Dominici, F., Curreiro, F. C., Coursac, I., Dockery, D. W., Schwartz, J. and Zanobetti, A. (2000b) Helath Effects Institute Research Report 94, Part II. Simoneit, B. R. T. (2002), Biomass burning -- a review of organic tracers for smoke from incomplete combustion, Applied Geochemistry, 17, 129-162. Todd, J. J. (1991), Emission and performance of wood heaters when burning softwood, Report 3, Centre for environment Studies, University of Tasmania,hobart.