Characterization and Kinetics Study of Off-Gas Emissions from Stored Wood Pellets

Transcription

1 Ann. Occup. Hyg., pp. 1 9 Ó The Author Published by Oxford University Press on behalf of the British Occupational Hygiene Society doi: /annhyg/men053 Characterization and Kinetics Study of Off-Gas Emissions from Stored Wood Pellets XINGYA KUANG 1,3, TUMULURU JAYA SHANKAR 1, XIAOTAO T. BI 1, SHAHAB SOKHANSANJ 1,4, C. JIM LIM 1 * and STAFFAN MELIN 2 1 Department of Chemical and Biological Engineering, University of British Columbia, 2360 East Mall, Vancouver, British Columbia V6T 1Z3, Canada; 2 Delta Research Corporation, 501 Centennial Parkway, Delta, British Columbia V4L 2L5, Canada; 3 Department of Occupational Medicine, Yangpu District Central Hospital, Shanghai , China; 4 Environmental Sciences Division, Oak Ridge National Laboratory, Oak Ridge, TN 37831, USA Received 10 March 2008; in final form 19 July 2008 The full potential health impact from the emissions of biomass fuels, including wood pellets, during storage and transportation has not been documented in the open literature. The purpose of this study is to provide data on the concentration of CO 2, CO and CH 4 from wood pellets stored in sealed vessels and to develop a kinetic model for predicting the transient emission rate factors at different storage temperatures. Five 45-l metal containers (305 mm diameter by 610 mm long) equipped with heating and temperature control devices were used to study the temperature effect on the off-gas emissions from wood pellets. Concurrently, ten 2-l aluminum canisters (100 mm diameter by 250 mm long) were used to study the off-gas emissions from different types of biomass materials. Concentrations of CO 2, CO and CH 4 were measured by a gas chromatograph as a function of storage time and storage temperature. The results showed that the concentrations of CO, CO 2 and CH 4 in the sealed space of the reactor increased over time, fast at the beginning but leveling off after a few days. A first-order reaction kinetics fitted the data well. The maximum concentration and the time it takes for the buildup of gas concentrations can be predicted using kinetic equations Keywords: biomass; decomposition kinetics; emission factors; off-gassing emission; storage; temperature effect; wood pellets INTRODUCTION A gradual shift from fossil fuels to biomass fuel for electricity and heat generation is under way everywhere especially in Europe. Wood pellets, as a densified renewable fuel, are widely used for heat and electricity production as well as cofiring with coal and natural gas. Of roughly tonnes of wood pellets consumed during 2007 in Europe, tonnes are transported from Canada mainly from British Columbia (BC). One of the issues related to the safe handling of wood pellets is excessive concentration of CO, CO 2,CH 4 and depletion of O 2 in storage tanks and shipping vessels. It is well known that all biomass gradually decomposes over time, both chemically and biologically, *Author to whom correspondence should be addressed. Tel: þ ; fax: þ ; cjlim@chml.ubc.ca slowly releasing toxic and oxygen-depleting gases such as CO, CO 2 and CH 4 (Reuss and Pratt, 2001; Johansson et al., 2004; Arshadi and Gref, 2005). The off-gases from wood pellets have not been considered an issue until recently when incidents of injuries and even fatalities occurred among people who worked around large storage tanks and vessels. Limited studies (Svedberg et al., 2004) have reported the composition of the off-gas emissions from stored wood pellets. CO, CO 2,CH 4 and non-methane organic compounds are commonly identified in the off-gases from biomass (Johansson et al., 2004). It was postulated that CO is formed due to auto-oxidative degradation of fats and fatty acids (Svedberg et al., 2004; Arshadi and Gref, 2005). Pellets made from aged sawdust might generate less volatile organic compounds (VOCs) than pellets made from fresh sawdust due to the oxidation process. The level of off-gassing was also found to be related to wood species, pellets made from spruce sawdust emitting less 1of9

2 2 of 9 X. Kuang et al. VOCs than pellets made from pine (Arshadi and Gref, 2005). High levels of hexanal and CO were identified in the off-gases released from stored wood pellets (Svedberg et al., 2004; Ernstgard et al., 2006). Toxicological literature survey showed that the available scientific information on hexanal is insufficient to determine its potential risks to human health. Most recently, Svedberg et al. (2008) reported a fatal accident onboard of a vessel in Port of Helsingborg, Sweden, when unloading wood pellets from BC, Canada. One seaman was killed, a stevedore was seriously injured and several rescue workers were slightly injured due to the lack of O 2 and the emission of CO when they entered an unventilated stairway next to the cargo hold. Gaseous concentrations from wood pellets in the closed cargo hatches of five ocean vessels from Canada to Sweden during ocean transportation were monitored. The concentration ranges of CO, CO 2 and CH 4 were reported as , and p.p.m., respectively, which were well above the safety threshold values. The oxygen concentration was found to vary from 0.8 to 16.9%. The objective of this study is to characterize emissions of CO 2, CO and CH 4 from wood pellets in an enclosed storage space and to develop a kinetic model for predicting the evolution of emission rate factors at different storage temperatures. MATERIALS AND METHODS Equipment and experimental materials Ten 2-l aluminum containers (100 mm diameter and 250 mm long) were used to study the gas emissions when loaded with switchgrass, wood pellets made in BC, Canada, wood pellets made in Europe and torrefied (partially heat treated and carbonized) woodchips also made in Europe. The BC wood pellets (from BC, Canada) were made from fresh mill residues such as sawdust and planer shavings from pine trees harvested 2 years after they died as a result of attack by the mountain pinebeetles. As received, they were 6 mm diameter and mm long, with the size distributions by screening: 1% mm, 5% mm and 94%.6.3 mm. We also tested emissions from wood pellet fines, which consisted of 30.8% 0 1 mm, 28.8% 1 2 mm, 40.4% mm and 4%.3.15 mm. The European wood pellets were made from Pinus sylvestris (Scots pine) harvested in Europe. The pellets came directly from the manufacturing plant by air and were less than a week old when tested. As received, they were 6 mm diameter and mm long, with the size distributions by screening: 0% mm, 12.5% mm and 87.5%.6.3 mm. European torrefied woodchips were also made from P. sylvestris harvested in Europe, by thermal treatment in the absence of oxygen at temperatures from 200 to 300 C. The chips were 6 months old when we started. The torrefied woodchips as received are mm 2 and 5 mm thick. The weight of biomass used in each test was 350 g for switchgrass, 1000 g for BC wood pellets, 800 g for fines of BC wood pellets, 1200 g for the European wood pellets and 1000 g for the European torrefied woodchips. For each test, two identical canisters were loaded with the same amount of biomass samples, sealed and then placed in a series of ovens with their temperature controller set at 20 and 40 C. Five milliliter gas samples were drawn from the two duplicate containers by a syringe daily. The composition of the sampled gas was analyzed by the gas chromatographic (GC) methods (hydrogen flame ionization detector and thermal conductivity detector) using a GC-14A (Shimadzu Corporation, Japan) to quantify the concentration of CO, CO 2,CH 4 and O 2. Duplicate samples were also measured by GC. The GC was calibrated regularly with standard calibration gases, e.g. standard CO, CO 2 and CH 4 gases. Argon and compressed air were used as the reference and carrier gases. A fused silica capillary column of inner diameter of 0.1 mm and length 50 m was used. Five 45-l steel containers (305 mm diameter and 610 mm long) were also used in the experiment, with the inner and outer walls of the reactors all coated by rust-proof Teflon, to prevent interaction between the wood pellets and the metal wall and to prevent the vessel rusting. Only BC wood pellets were tested, and each container was filled with 25 kg of wood pellet during testing. The containers were sealed airtight. The containers were heated to central temperatures of 20, 30, 40, 50 and 55 C using electrical heating tapes wrapped around on the outside of the reactors. The moisture content of the pellets at the start of trials was %. Ten millilitre gas was drawn from the reactor daily for the measurement of the gas compositions, including CO, CO 2 and CH 4. The tests were typically run until the gas concentration did not increase any further, which usually happen over a period of 30 to 60 days. RESULTS Characteristics of off-gas emissions Figures 1 3 show typical results on CO 2, CO and CH 4 emissions, respectively, obtained from the 45-l containers with BC wood pellets. The gas concentrations increased over time, fast at the beginning and approaching a plateau after a few days. As the storage temperature increased from 20 to 55 C, the emission rate increased significantly. However, there was very little change of the container pressure for tests at room temperature, and the maximum increase in container pressure at 50 C ranged from 6.9 to 8.1

3 Characterization and kinetics study of off-gas emissions 3 of 9 Fig. 1. CO 2 concentrations in the 45-l containers as a function of storage time at different storage temperatures using BC wood pellets. Lines are the best fitted lines using equation (1). Fig. 2. CO concentrations in the 45-l containers as a function of storage time at different storage temperatures using BC wood pellets. Lines are the best fitted lines using equation (1). kpa over a complete test run. This insignificant change in the container pressure over the test period may be a result of the counter balance between the oxygen depletion and the CO 2 /CO formation. One mole CO 2 consumes one mole of oxygen, and one mole CO consumes half a mole of oxygen; 79% of air in the container is nitrogen which does not participate in the reaction. In comparing the three species of gas, CO 2 concentration in general is the highest, while CH 4 is the lowest. Figures 4 and 5 show the concentration ratios of CO/CO 2 and CH 4 /CO 2 as a function of storage time and storage temperature. Figure 4 shows a decrease in CO/CO 2 ratio with the increase in storage temperature. The result may imply a shift in equilibrium conditions for CO and CO 2 concentrations. According to the following reversible exothermic reaction: 2CO þ O 2 5 2CO 2, a high temperature will favor a high CO/CO 2 ratio, but a high pressure will favor a low CO/CO 2 ratio. The counter effects of temperature and pressure may lead to either the increase or decrease of the CO/CO 2 ratio. Current experimental results seem to suggest that the pressure has more significant effect on the reversible reaction process, although further investigation is still needed. As the temperature rises, both CH 4 and CO 2 emissions increase according to Figs 1 and 3. At the same time, the CH 4 /CO 2 ratio also increases, as shown in

4 4 of 9 X. Kuang et al. Fig. 3. CH 4 concentration in the 45-l containers as a function of storage time at different storage temperatures using BC wood pellets. Lines are the best fitted lines using equation (1). Fig. 4. CO/CO 2 molar concentration ratios in the 45-l containers as a function of storage time at different storage temperatures using BC wood pellets. Fig. 5. This indicates that CH 4 generation is favored over CO 2 at a high temperature. In a conventional biomass composting system, CH 4 generation is usually associated with the anaerobic decomposition of biomass, while CO 2 likely is generated from the thermal oxidation or aerobic degradation. In a system containing CO 2,CH 4 and O 2, the following reversible exothermic reaction may take place: CH 4 þ O 2 5 CO 2 þ H 2 O. By assuming that the equilibrium is established in the above reaction among O 2,CO 2 and CH 4, high temperature will favor a high CH 4 / CO 2 ratio for such a reversible exothermic reaction, which seems to be consistent with the results shown in Fig. 5. Comparison of gas emissions from various woody materials The effect of wood properties on the off-gassing emissions of wood pellets was studied using five materials: switchgrass, BC wood pellets, fines of BC wood pellets, European wood pellets and European torrefied woodchips. The peak concentrations of the gas emissions measured over a storage period of days for different species at 20 and 40 C are listed in Table 1. Table 1 shows a large increase in emission concentrations of CO 2,CH 4 and CO as the storage temperature increased from 20 to 40 C. At 20 C, the torrefied woodchip released more CO 2 than the other

5 Characterization and kinetics study of off-gas emissions 5 of 9 Fig. 5. CH 4 /CO 2 molar concentration ratios in the 45-l containers as a function of storage time at different storage temperatures using BC wood pellets. Table 1. Plateau (peak) concentrations of CO 2, CO and CH 4 for five different woody materials stored in 2-l containers at two temperatures Material Storage CO 2 (p.p.m.) CH 4 (p.p.m.) CO (p.p.m.) time (days) 20 C 40 C 20 C 40 C 20 C 40 C Switchgrass Fines of BC wood pellets BC wood pellets European wood pellets European torrefied woodchips materials. Switchgrass and European wood pellets produced low CO, CO 2 and CH 4 emissions. Although switchgrass emits less CO 2 at room temperature than any of the other materials, it emits more CO 2 at 40 C. Similarly, the European wood pellets are very sensitive to the temperature increases. Torrefied woodchips, on the other hand, are less sensitive to temperature increases, emitting the highest amount of CO 2 at 20 C of all materials, but the least at 40 C. All the pelletized materials emit more gases, perhaps as a result of the heat treatment during the manufacturing process, which opens the porous structure of wood and thereby increases the surface area of the material. This hypothesis is supported by the slightly higher emissions from the fines of BC wood pellets, which have a larger total surface area than the original wood pellets. It should be noted that the emissions from all the tested materials are related to the age of the biomass. The longer it has been aged, the less gas is expected to be emitted. However, the gas emission at elevated temperatures (e.g. 40 C) is likely less affected by the aging of the materials. Switchgrass and torrefied woodchips generally emit less gas at 40 C, which may be attributed to the less porous structure of the material and less surface area. The removal of VOCs during the torrefaction process is a likely explanation why emission of gases from torrefied wood is less sensitive to the temperature. Development of simple off-gassing kinetics Figures 1 3 show that emissions from all species follow an exponential function, a characteristic of the first-order chemical reactions. Most oxidation reactions can be approximated as first-order reactions at low hydrocarbon concentrations (Wark et al., 1999). The approximation is that the woody material only decomposes into CO, CO 2 and CH 4 under approximately constant pressures established by present study. The concentrations of CO, CO 2 and CH 4 can be derived from the first-order reaction kinetic equation, with the final expression in form of the following equation: C i t 5 Ci;N 1 exp ki t ; ð1þ where C i,} represents the maximum concentration at the plateau, k i is the kinetic rate constant of the firstorder kinetic equation and t is the storage time in days. Under a constant temperature (T) and pressure

6 6 of 9 X. Kuang et al. (P) of the reaction, the concentration can be converted to an emission factor, f i (in gram gas per kilogram materials) by: f i g kg 5 P Ci V M wt RTMp ; ð2þ where R is the gas constant, M wt is the gas molecular weight (gram per mole), M p is the total mass of material in the container (kilogram), V is the total gas volume in the reactor (cubic meter), which equals to the container volume less the pellets volume and P is the pressure of the vessel and is approximated as a constant based on the observation that there is minimal change in P at low temperatures and,8% at 50 C, the highest tested temperature. The volume of pellets can be calculated based on the weight of pellets divided by the pellet density, while the pellet density can be calculated as the average value of the measured weight of single pellets divided by the measured volume of single pellets. Equation (1) is fitted to the experimental data shown in Figs 1 3. The fitted parameter of C i, } is then converted to an emission factor using equation (2). Fitted values of f i,} and k i are also given in Table 2, together with a half-life time (s 1/2 ), i.e. the time for the emission concentration to reach half of its final peak (maximum) or plateau value, which characterizes how fast the off-gassing will occur. It is seen from Table 2 that both the peak emission factor (f } ) and the reaction rate constant (k) increased with the increase in storage temperature for all three gases (CO 2, CO and CH 4 ). The number of days to reach the half plateau emissions decreased as the storage temperature increased. The implication is that the reaction is faster at higher temperatures and, as a result, a higher concentration of gases will buildup in a shorter time period at higher storage temperatures. For first-order reactions, the reaction rate constant k follows the Arrhenius relationship: k 5 A 0 exp E : ð3þ RT The activation energy E, and pre-factor A 0, can be obtained from the plot of the reaction rate constant k versus the reactor temperature in the semilog form lnðkþ 5 lnða 0 Þ E=RT: The following three kinetic rate constant expressions are thus obtained by least square linear regression of each line: k CO2 5 2: exp ; ð4þ RT k CO 5 1: exp ; ð5þ RT k CH4 5 3: exp : ð6þ RT The activation energies for CO 2, CO and CH 4 are 59.88, and kj/mole, respectively, and of the same order for the first-order oxidation reactions of most hydrocarbons (Wark et al., 1999). The estimated activation energy values show that CO 2 is emitted with the least energy of formation (application of energy) followed by CH 4 and CO in this order. Figure 6 shows the plateau (peak) emission factors f i, } for CO, CO 2 and CH 4 as a function of temperature. A linear regression of the curves gives three linear equations for the plateau (peak) emission factors as a function of temperature (T in C): f CO2 ;N 5 0:0022T 0:0184; ð7þ f CO;N 5 0:0001T þ 0:0109; f CH4 ;N 5 0:00004T 0:0005: ð8þ ð9þ By combination of equations (1) (9), the following three equations are obtained for the prediction of CO, CO 2 and CH 4 emissions in gram per kilogram, respectively, from BC wood pellets used in this experimental study in sealed vessels for a temperature range of C and at the atmospheric pressure: f CO2 ðtþ 5 ð0:0022t 0:0184Þ exp 2: exp t ; 8:31T ð10þ Table 2. Plateau (peak) gas concentrations, emission factor (f } ), reaction rate constant (k) and time (s 1/2 ) to reach 50% plateau (peak) concentrations for gases emitted from BC wood pellets stored in the 45-l containers at different temperatures T ( C) CO 2 CO CH 4 C (p.p.m.) f } (g kg 1 ) k (day 1 ) s 1/2 (day) C (p.p.m.) f } (g kg 1 ) k (day 1 ) s 1/2 (day) C (p.p.m.) f } (g kg 1 ) k (day 1 ) s 1/2 (day)

7 Characterization and kinetics study of off-gas emissions 7 of 9 Fig. 6. The plateau (peak) emission factors for CO 2, CO and CH 4 from BC wood pellets in 45-l sealed containers. Lines are linearly fitted lines. f CO ðtþ 5 ð0:0001t þ 0:0109Þ exp 1: exp t ; 8:31T ð11þ f CH4 ðtþ 5 ð0:00004t 0:0005Þ exp 3: exp t : 8:31T ð12þ DISCUSSION Temperature impact It is observed from the current study that temperature is one of the critical factors that affect the gas emission (off-gassing) from stored wood pellets. From the kinetic study, it is found that for a volume of biomass (in this case wood pellets) enclosed in a sealed space, the temperature affects both the plateau (peak) emission factor and the emission rate, with higher peak emission factor and faster emission rate as the temperature increases. As biomass can

8 8 of 9 X. Kuang et al. decompose both chemically and biologically, it is anticipated that the emission rate will increase with temperature if the chemical degradation is in dominance. At the same time, the biological process may peak at certain temperature and decrease at higher temperatures when bacteria and fungi perish. The results from the current study suggest that chemical process could be the dominant mechanism for off-gassing of CO 2, CO and CH 4, although biological process may also contribute to the emissions for moist biomass such as woodchips. Further studies of the thermal and biological decomposition of wood pellets under controlled conditions are needed to verify to what extent the biological process contributes to the decomposition of woody materials. Concentration buildup in confined storage spaces and the potential health impact Gas emissions from woody material in a confined storage space will change the composition of the contained gases, leading to the buildup of toxic compounds (e.g. CO). For a given storage space, if the temperature is relatively stable, the peak concentration and the time it takes for such a concentration to buildup can be predicted using the currently developed kinetic equations for given storage temperature, the weight of wood pellets and the storage space volume. Air contains 380 p.p.m. CO 2, a normal nonreactive and non-toxic by-product from combustion and decomposition of biodegradable materials as well as human and animal metabolism. At low concentrations, CO 2 is categorized as a simple asphyxiant when allowed to replace air in a given space. A CO 2 concentration p.p.m. in the air starts to cause headaches and dizziness, increases the heart rate and blood pressure and induce coma. The threshold limit value time weighted average (TLV TWA) for 8 h exposure is set at 5000 p.p.m. in most jurisdictions. CO acts toxically by preventing the uptake of oxygen by the hemoglobin of the red blood cells and thereby preventing the red blood cells from carrying the required oxygen to the brain and other parts of the body. At a CO level of 300 p.p.m., people start to experience headaches, fatigue and discomfort. The TLV TWA for 8 h exposure is between 25 and 50 p.p.m. in most jurisdictions. The peak concentrations of CO 2,COandCH 4 emitted from BC wood pellets in 45-l containers for C are listed in Table 2. These values, which are comparable with those measured in closed cargo ship hatches reported by Svedberg et al. (2008), are well above their respective TLV TWA values. In addition to CO and CO 2,oneepidemiological investigation on acute effects caused by exposure to hexanal vapors from wood pellets involving 12 volunteers has been reported (Ernstgard et al., 2006). Future study incorporating the real temperature and ventilation data from specific storage and shipping vessels and the emission factor data from the current study should provide useful safety guidelines. Effect of oxygen level on off-gassing kinetics Oxygen content in fresh air is 21% by volume. In a sealed container, oxygen is consumed by the chemical and likely biological decomposition of wood pellets (Svedberg et al., 2008), leading to the depletion of oxygen and the generation of CO 2 and CO in the container. Figure 7 shows strong correlation between the buildup of CO and CO 2 concentration and decrease in O 2 concentration in the 45-l containers. It demonstrates that O 2 was depleted steadily when the emission of CO and CO 2 increased. In other Fig. 7. The relationships between oxygen and CO þ CO 2 concentrations for BC wood pellets in the 45-l containers. Line is linearly fitted line.

9 Characterization and kinetics study of off-gas emissions 9 of 9 words, the quantity of O 2 in a sealed container affects the generation of the off-gases. The trend of emissions under different free air space to pellet volume ratios (Vg/Vp) is currently under investigation in our laboratory to elucidate the relationship between the depletion of oxygen and the generation of offgas emissions. A better understanding of this relationship will allow useful prediction of requirements for ventilation and risk assessment for unhealthy working conditions. CONCLUSIONS 1. Storage temperature is one of the critical factors that affect the off-gassing from stored wood pellets. Higher peak emission factors and faster emission rates are associated with higher temperatures. 2. A first-order kinetic equation has been developed for predicting CO, CO 2 and CH 4 emission rates and their peak concentrations from the wood pellets in a storage vessel at different temperatures. 3. The concentrations of CO 2,COandCH 4 at room temperature or higher temperatures can be close to or well exceed the safety threshold values of the accepted TLV TWA standards in most jurisdictions. 4. Chemical decomposition of wood pellets is the dominant mechanism for off-gassing CO, CO 2 and CH 4 within the temperature ranges (20 55 C) studied. However, the relationship between the oxygen depletion and off-gas emissions needs to be further studied. FUNDING Natural Sciences and Engineering Research Council of Canada (NSERC-CRDPJ ); Wood Pellet Association of Canada (Grant11R42500). REFERENCES Arshadi M, Gref R. (2005) Emission of volatile organic compounds from softwood pellets during storage. For Prod J; 55: Ernstgard L, Iregren A, Sjogren B et al. (2006) Acute effects of exposure to hexanal vapors in humans. J Occup Environ Med; 48: Johansson LS, Leckner B, Gustavsson L et al. (2004) Emission characteristics of modern and old-type residential boilers fired with wood logs and wood pellets. Atmos Environ; 38: Reuss R, Pratt S. (2001) Accumulation of carbon monoxide and carbon dioxide in stored canola. J Stored Prod Res; 37: Svedberg URA, Hogberg HE, Hogberg J et al. (2004) Emission of hexanal and carbon monoxide from storage of wood pellets, a potential occupational and domestic health hazard. Ann Occup Hyg; 48: Svedberg U, Samuelsson J, Melin S. (2008) Hazardous off-gassing of carbon monoxide and oxygen depletion during ocean transportation of wood pellets. Ann Occup Hyg; 52: Wark K, Warner CF, Davis WT. (1999) Air pollution: its origin and control. 3rd edn. Berkeley, CA: Addison-Wesley. pp