MOISTURE SORPTION AND GAS EMISSIONS DURING THE STORAGE OF HIGH MOISTURE WOODY BIOMASS by XIAO HE B. Eng., Nanjing University of Science and Technology, 2007 M.A.Sc., Nanjing University of Science and Technology, 2009 A THESIS SUBMITTED IN PARTIAL FULFILMENT OF THE REQUIREMENTS FOR THE DEGREE OF DOCTOR OF PHILOSOPHY in THE FACULTY OF GRADUATE STUDIES (CHEMICAL AND BIOLOGICAL ENGINEERING) THE UNIVERSITY OF BRITISH COLUMBIA (VANCOUVER) April 2013 © Xiao He, 2013 ii ABSTRACT Moisture sorption and gas emissions are major processes associated with biomass storage. Depending on the storage conditions, these processes alter the structure and composition of biomass. The objectives of this research are (1) to develop moisture relations for woody biomass exposed to drying and wetting environments; (2) to quantify gas emissions from biomass stored under aerobic and anaerobic conditions; and (3) to develop dry matter loss equations for the stored biomass. Moisture adsorption and desorption (drying) experiments were carried out on Aspen branches in a controlled temperature and humidity chamber. Frequent wetting-drying cycles were simulated by spraying water on the biomass. A lump model for simulating moisture adsorption-desorption was developed and calibrated with experimental results. The model was applied to the Aspen bales stored for one year in the field under natural conditions. The predicted moisture contents using the lump moisture transfer model were found to be in reasonably good agreement with the moisture contents measured in the stored bales. In another set of experiments, gas emissions from stored Western Red Cedar (WRC) and Douglas fir (DF) were analyzed. The emissions of CO2, CO, H2 and CH4, and the depletion of O2 were measured. The highest total CO2 emissions from WRC stored in the non-aerobic and aerobic reactors were 2.8 g/kg DM and 6.6 g/kg DM, respectively. Higher gas emissions were measured from stored DF materials than from WRC. Common volatile organic compounds (VOCs) measured using GC-MS were methanol, aldehydes, terpene, acid, acetone, hexane, ketone, benzene, ethers and esters from WRC and DF. The total VOC concentrations were found to have a positive correlation with temperature. The results of microbial analysis were compatible with gas emission results. Positive correlations between percent dry matter losses and gas emissions were found for both aerobic and non-aerobic storage conditions. The summation of gas emissions from aerobic reactors is greater than accumulated gas emissions from non-aerobic reactors over the same storage period. It was found that DF is more readily degradable than WRC. Greens (leaves and twigs) degrade faster than wood chips. iii PREFACE Part of Chapter 6 has been published: He, X., Lau, A.K., Sokhansanj, S., Lim, C.J., Bi, X.T. and Melin, S. 2012. Dry matter losses in combination with gaseous emissions during the storage of forest residues. Fuel, 95, 662-664. The experiment was designed and conducted by Xiao He under supervision of Dr. Lau and Dr. Sokhansanj. Data analysis and discussion was performed by Xiao He, with the guidance from Dr. Lau and Dr. Sokhansanj. The manuscript was written by Xiao He, with insightful feedback from all the other co-authors. Part of Chapter 3 was submitted for publication and currently under revision: He, X., Lau, A.K., Sokhansanj, S., Lim, C.J. and Bi, X.T. 2012. Modeling the drying and wetting processes of Aspen (Populus tremuloides). Bioenergy Research. The cycles of drying and wetting experiment was conducted by Xiao He. The model was developed and calibrated by Xiao He, with helpful advice from Dr. Lau and Dr. Sokhansanj. Xiao He was responsible for data collection, analysis and manuscript preparation. Dr. Lau and Dr. Sokhansanj provided feedback and guidance throughout the process. Part of Chapter 2 was submitted for publication and under review: He, X., Lau, A.K., Sokhansanj, S., Lim, C.J., Bi, X., Melin, S. and Keddy, T. 2012. Moisture sorption isotherms and drying characteristics of Aspen. Biomass & Bioenergy. All the experiment and data analysis, as well as manuscript preparation were conducted by Xiao He, with guidance from Dr. Lau and Dr. Sokhansanj. The co-authors provided valuable feedback on the written work. Part of Chapter 5 was submitted for publication and under review: He, X., Lau, A.K., Sokhansanj, S., Lim, C.J., Bi, X. and Melin, S. 2013. Gas emissions from stored Western Red Cedar chips. International Journal of Environmental Science and Technology. The reactors as well as the lab work were designed by Xiao He, with helpful advice from Dr. Lau and Dr. Sokhansanj. Xiao He was responsible for data collection and analysis. The written work was a collaborative effort by Xiao He, Dr. Lau, Dr. Sokhansanj, Dr. Lim and Dr. Bi. Part of Chapter 4 titled Application of a model to simulate the wetting and drying of Aspen bales in the field and part of Chapter 6 titled Comparison of gas emissions from different biomass materials during storage will be submitted for publication. iv TABLE OF CONTENTS ABSTRACT ............................................................................................................................. ii PREFACE ................................................................................................................................ iii TABLE OF CONTENTS ......................................................................................................... iv LIST OF TABLES ................................................................................................................. viii LIST OF FIGURES ................................................................................................................. ix LIST OF ABBREVIATIONS ................................................................................................ xiii LIST OF SYMBOLS ............................................................................................................. xiv ACKNOWLEDGEMENTS ................................................................................................... xvi DEDICATION ...................................................................................................................... xvii Chapter 1. Introduction ............................................................................................................. 1 1.1 Background and problem statement ........................................................................... 1 1.2 Objectives ................................................................................................................... 3 1.3 Organization of the thesis ........................................................................................... 5 1.4 Literature review ........................................................................................................ 6 1.4.1 Biomass ............................................................................................................... 6 1.4.2 Biomass storage .................................................................................................. 7 1.4.3 Moisture sorption characteristics ...................................................................... 13 1.4.4 Gas emissions from stored biomass .................................................................. 22 1.5 Concluding remarks ................................................................................................. 29 Chapter 2. Moisture sorption characteristics of biomass during storage ................................ 30 2.1 Introduction .............................................................................................................. 30 2.2 Materials and methods ............................................................................................. 32 2.3 Results and discussion .............................................................................................. 33 v 2.3.1 Moisture sorption characteristics ...................................................................... 33 2.3.2 Moisture sorption isotherms ............................................................................. 35 2.3.3 Drying rate of Aspen ......................................................................................... 39 2.4 Conclusion ................................................................................................................ 43 Chapter 3. Modelling the drying and wetting processes of Aspen (Populus tremuloides) ..... 44 3.1 Introduction .............................................................................................................. 44 3.2 Model description ..................................................................................................... 45 3.3 Materials and methods ............................................................................................. 48 3.3.1 Materials ........................................................................................................... 48 3.3.2 Experiment ........................................................................................................ 48 3.4 Results and discussion .............................................................................................. 49 3.4.1 Internal moisture content .................................................................................. 49 3.4.2 External moisture content ................................................................................. 50 3.5 Conclusion ................................................................................................................ 55 Chapter 4. Model application for moisture variation of Aspen bales ..................................... 56 4.1 Introduction .............................................................................................................. 56 4.2 Model description ..................................................................................................... 57 4.2.1 Moisture content ............................................................................................... 57 4.2.2 Equilibrium moisture content ........................................................................... 58 4.2.3 Evaporation rate ................................................................................................ 59 4.2.4 Parameter estimation ......................................................................................... 60 4.3 Materials and methods ............................................................................................. 61 4.3.1 Model structure and simulation procedure ....................................................... 61 4.3.2 Field Test .......................................................................................................... 63 vi 4.4 Results and Discussion ............................................................................................. 64 4.4.1 Model application ............................................................................................. 64 4.4.2 Prediction of the moisture content of a bale with transparent cover ................. 71 4.5 Conclusion ................................................................................................................ 73 Chapter 5. Gas emissions from stored Western Red Cedar chips ........................................... 74 5.1 Introduction .............................................................................................................. 74 5.2 Materials and methods ............................................................................................. 76 5.2.1 Materials ........................................................................................................... 76 5.2.2 Experimental setup ........................................................................................... 77 5.2.3 Gas emission measurement ............................................................................... 78 5.2.4 Microbial analysis ............................................................................................. 79 5.2.5 Data analysis ..................................................................................................... 79 5.3 Results and discussion .............................................................................................. 80 5.3.1 Gas emissions ................................................................................................... 80 5.3.2 Gas emissions from sterilized woodchips ......................................................... 87 5.3.3 Characteristics of VOCs ................................................................................... 90 5.3.4 Gas emissions from wood chips with different initial moisture content ........... 91 5.4 Conclusion ................................................................................................................ 95 Chapter 6. Gas emissions from stored Douglas fir (Pseudotsuga menziesii) ......................... 97 6.1 Introduction .............................................................................................................. 97 6.2 Materials and methods ............................................................................................. 99 6.2.1 Materials ........................................................................................................... 99 6.2.2 Experimental setup ......................................................................................... 100 6.2.3 Gas emission measurements ........................................................................... 101 vii 6.2.4 Microbial analysis ........................................................................................... 102 6.2.5 Data analysis ................................................................................................... 102 6.3 Results and discussion ............................................................................................ 102 6.3.1 Characteristics of gas emissions from Douglas fir chips (Test Series #1) ...... 102 6.3.2 Characteristics of gas emissions from Douglas fir greens (Test Series #2) .... 108 6.3.3 Characteristics of gas emissions from mixed Douglas fir wood chips and greens (Test Series #3) .............................................................................................................. 112 6.3.4 Comparing the results from the three types of DF materials .......................... 116 6.3.5 Comparison of gas emissions between stored Douglas fir and Western Red Cedar wood chips .......................................................................................................... 130 6.3.6 Comparison of gas emissions with wood pellets ............................................ 131 6.3.7 Comparison of gas emissions with other woody materials ............................. 132 6.4 Conclusion .............................................................................................................. 133 Chapter 7. Conclusions and Recommendations ................................................................... 136 7.1 Conclusions ............................................................................................................ 136 7.2 Recommendations for future research.................................................................... 142 References ............................................................................................................................. 144 Appendix A. Summary of different drying methods ............................................................ 156 Appendix B. Derivation of Equation (5.1) ........................................................................... 157 Appendix C. Summary of bales temperatures ...................................................................... 158 Appendix D. GC and GC/MS spectra ................................................................................... 162 viii LIST OF TABLES Table 1.1. The Threshold Limit Value of carbon dioxide, carbon monoxide and methane ... 23 Table 2.1. Estimated coefficients and error parameters of four moisture sorption isotherm models fitted to experimental data .......................................................................................... 37 Table 2.2. Parameters of Page’s equation (k and n) under different temperatures ................. 42 Table 6.1a. Total bacterial counts of the Douglas fir samples (cfu/g sample) ...................... 118 Table 6.1b. Mold counts of samples (cfu/g sample) ............................................................. 119 Table 6.2. Dry matter losses from each test (%) ................................................................... 120 Table 6.3. Estimated constants a and b for DF wood chips in Eq (6.7) ................................ 126 Table 6.4. A comparison of VOCs identified by qualitative GC/MS analysis ..................... 130 Table 6.5. Comparison of gas emissions from wood pellets and chips under non-aerobic condition after 30 days storage ............................................................................................. 132 Table A1. Summary of different drying methods ................................................................. 156 Table C1. Summary of monthly average temperature data for the four bales ...................... 158 ix LIST OF FIGURES Figure 1.1. Supply chain of biomass feedstock from harvest to biorefinery ............................ 7 Figure 2.1. Aspen pieces used in the experimental study ....................................................... 32 Figure 2.2. Schematic diagram of the controlled environment chamber ................................ 33 Figure 2.3. Moisture desorption curves for Aspen at different air temperatures and relative humidity .................................................................................................................................. 34 Figure 2.4. Moisture adsorption curves for Aspen: (a) under 90% relative humidity and three temperatures; (b) under 35oC temperature and three levels of relative humidity ................... 35 Figure 2.5. Adsorption and desorption curves as predicted by the Modified Oswin equation38 Figure 2.6. Drying rate versus moisture content at 40oC. Mc is the critical moisture constant, a transition from constant rate drying to falling rate drying ................................................... 40 Figure 3.1. Change in total moisture content with time when biomass is wetted and dried repeatedly at 20oC. This graph emulates frequent wetting of biomass due to rain ................. 51 Figure 3.2. Change in total moisture content with time when biomass is wetted and dried repeatedly at 30oC ................................................................................................................... 52 Figure 3.3. Drying rate versus moisture content at 20oC ........................................................ 53 Figure 3.4. Variation of the three coefficients with the number of simulated precipitation (wetting) events at 20oC and 30oC (a1: ∆; a2: ▪; a3: ●) .......................................................... 54 Figure 4.1. Flow chart of the model to calculate the moisture content of Aspen materials at time t ....................................................................................................................................... 62 Figure 4.2. Stored Aspen bales at UBC (uncovered: bales #1 and #2; covered: bales #3 and #4) ........................................................................................................................................... 64 Figure 4.3. Daily mean air temperature for Vancouver, British Columbia from June 1, 2011 to May 31, 2012 ...................................................................................................................... 67 Figure 4.4. Daily air relative humidity for Vancouver, British Columbia from June 1, 2011 to May 31, 2012 .......................................................................................................................... 68 x Figure 4.5. Daily precipitation for Vancouver, British Columbia from June 1, 2011 to May 31, 2012 .................................................................................................................................. 68 Figure 4.6. Daily net radiation from June 2011 to May 2012 ................................................. 69 Figure 4.7. Calculated daily evaporation rate for the uncovered bales ................................... 69 Figure 4.8. Calculated daily evaporation rate for the covered bales ....................................... 70 Figure 4.9. Predicted and measured moisture contents of the uncovered Aspen bales (bale #1 and #2) in the field (June, 2011 to May 2012) ........................................................................ 70 Figure 4.10. Predicted and measured moisture contents of the covered Aspen bales (bale #3 and #4) in the field (June, 2011 to May 2012) ........................................................................ 71 Figure 4.11. Predicted moisture contents of Aspen bale placed under a transparent cover .... 72 Figure 5.1. Western Red Cedar chips as received from the recycling yard. The size of chips varied from 30-80 mm in length on average ........................................................................... 77 Figure 5.2. Left reactors for aerobic tests were ventilated every 24 hours during the experiment. Right reactors for non-aerobic tests remain sealed for the duration of the experiment .............................................................................................................................. 78 Figure 5.3. Gas emission profiles from stored WRC chips under different temperatures (non- aerobic) (□: 5oC; ●: 20oC; ∆: 35oC; ▼: 45oC; *:50oC) ....................................................... 82 Figure 5.4. Effect of temperature on CO2 emission factors during the testing period under non-aerobic conditions ............................................................................................................ 83 Figure 5.5. Cumulative gas emissions from stored WRC chips under aerobic conditions (□: 5oC; ●: 20oC; ∆: 35oC; ▼: 45oC; *:50oC) ............................................................................. 85 Figure 5.6. Effect of temperature on CO2 emissions during the storage period under aerobic conditions ................................................................................................................................ 86 Figure 5.7. Gas emission from treated WRC chips under different temperatures (non-aerobic) (□: 5oC; ●: 20oC; ∆: 35oC; ▼: 45oC; *:50oC) ..................................................................... 89 Figure 5.8. Cumulative concentration of TVOC from the aerobic reactors at various temperatures ............................................................................................................................ 91 xi Figure 5.9. Gas emission profiles from stored WRC chips with initial moisture content of 35% wet basis under different temperatures (non-aerobic) (□: 5oC; ●: 20oC; ∆: 35oC; ▼: 45oC; *:50oC) .......................................................................................................................... 93 Figure 5.10. Cumulative gas emissions from stored WRC chips with initial moisture content of 35% wet basis under aerobic conditions (□: 5oC; ●: 20oC; ∆: 35oC; ▼: 45oC; *:50oC) 94 Figure 6.1. Douglas fir wood chips with the size range of 5-30 mm .................................... 100 Figure 6.2. Three series of tests with different materials (a: DF wood chips; b: DF greens; c: mixed DF chips and greens); and two kinds of reactors (left: non-aerobic reactor; right: aerobic reactor) ..................................................................................................................... 101 Figure 6.3. Gas emission profiles from stored DF chips at different temperatures under non- aerobic conditions (+: 5oC; ●: 20oC; □: 35oC; ▲: 50oC) .................................................. 104 Figure 6.4. Gas emission profiles from stored DF chips at different temperatures under aerobic conditions (+: 5oC; ●: 20oC; □: 35oC; ▲: 50oC) .................................................. 105 Figure 6.5. Cumulative concentration of total VOCs (TVOC) at different temperatures from aerobic reactors - wood chips ............................................................................................... 106 Figure 6.6. Gas emissions profiles of stored DF greens at different temperatures under non- aerobic conditions (+: 5oC; ●: 20oC; □: 35oC; ▲: 50oC) .................................................. 108 Figure 6.7. Gas emission profiles from DF greens at different temperatures under aerobic conditions (+: 5oC; ●: 20oC; □: 35oC; ▲: 50oC) ............................................................... 110 Figure 6.8. Cumulative concentration of TVOCs from aerobic reactors for DF greens ....... 111 Figure 6.9. Gas emissions from stored DF mixed chips and greens at different temperatures under non-aerobic conditions (+: 5oC; ●: 20oC; □: 35oC; ▲: 50oC) ................................. 113 Figure 6.10. Gas emissions from DF mixed chips and greens at different temperatures under aerobic conditions (+: 5oC; ●: 20oC; □: 35oC; ▲: 50oC) .................................................. 115 Figure 6.11. Cumulative concentration of TVOC at various temperatures from aerobic reactors for mixed materials ................................................................................................. 116 xii Figure 6.12. Correlation between dry matter losses and total CO2 emission of WRC and DF materials at all temperatures under non-aerobic conditions ................................................. 121 Figure 6.13. Correlation between temperature and dry matter losses from DF materials under non-aerobic conditions .......................................................................................................... 121 Figure 6.14. Correlation between dry matter losses and total CO2 emission from WRC and DF materials at all temperatures under aerobic conditions ................................................... 125 Figure 6.15. Dry matter losses as a function of temperature for DF materials under aerobic conditions .............................................................................................................................. 125 Figure 6.16. The number of days before DF wood chip loses 1% dry matter for temperatures from 0 to 50oC ...................................................................................................................... 126 Figure 6.17. The appearance of the materials in three series tests (DF chips, greens and mixed chips and greens) before and after the storage ........................................................... 129 Figure C1. The locations of thermocouples in each bale ...................................................... 161 Figure D1. GC spectra of the gas samples from the reactors with experimental materials .. 163 Figure D2. GC/MS spectra of the gas samples from the reactors with experimental material .............................................................................................................................................. 164 Figure D3. Identification of compounds by using mass-charge ratio ................................... 170 xiii LIST OF ABBREVIATIONS CO2 carbon dioxide CO carbon monoxide CH4 methane d.b. dry basis DM dry mass EMC equilibrium moisture content ERH equilibrium relative humidity GC gas chromatography GC-MS gas chromatography- mass spectrometry SEE standard error of estimation TBCs total bacteria counts TVOC total volatile organic compound VOC volatile organic compound w.b. wet basis xiv LIST OF SYMBOLS a1, a2, a3 coefficients, mm-1 ∆ slope of the saturation vapor pressure curve, kPa/oC γ psychrometric coefficient, kPa/oC δe vapor pressure deficit, kPa λ latent heat of vaporization, MJ/kg A, B, C coefficients Ci volumetric concentration of a particular gas, m3 gas species per m3 total gas Cn0 initial concentration of nitrogen, % Cnt concentration of nitrogen at time t, % Ep potential evaporation rate, mm/h f emission factor, g/kg DM k drying rate constant ka adsorption rate constant m total mass of materials in the container, kg M moisture content, decimal, dry basis M0 initial moisture content, decimal, dry basis Me equilibrium moisture content, decimal, dry basis Mi internal moisture content, decimal, dry basis Ms external moisture content, decimal, dry basis xv Mwt gas molecular weight, g/mol n constant P precipitation rate, mm/h Ps absolute pressure of gas in the container, Pa rh relative humidity R universal gas constant, J/(mol.K) Rn net radiation, MJ/(m2.d) t time, h T temperature, oC u wind speed, m/s V gas volume in the reactor (m3) xvi ACKNOWLEDGEMENTS I would like to express my greatest gratitude to my supervisors Dr. Anthony K. Lau and Dr. Shahab Sokhansanj. This thesis would not be able to be completed without their patient guidance, inspiration and constant support. I greatly appreciate their valuable advice, constructive criticism, thoughtful suggestions, rewarding discussions, and whenever available help throughout the course of this thesis research. I am also thankful to my thesis committee members Dr. Jim Lim and Dr. Stavros Avramidis for their encouragement and insightful comments to improve my thesis work. In particular, I am very grateful to Dr. Jim Lim for his valuable advice and constructive suggestions on experimental design and instrumentation. Special thanks are also expressed to Dr. Xiaotao Bi for his critical suggestions on my experiments and interpretation of data. I would like to give thanks to Dr. Karen Bartlett for her assistance with equipment to analyze gas emissions in this research. I wish to thank Mr. Staffan Melin for providing his valuable insight in the current critical challenges of the biomass and bioenergy from industry. I also appreciate the help from Ehsan Oveisi and Bahman Ghiasi for the set-up of Aspen bales and data collection in Gas Gun site on campus. I am thankful to Dr. Zhiwei Chen for spending his time helping me with the laboratory work. I would also like to thank the BBRG colleagues, CHBE workshop and administrative personnel for their technical and administrative supports. Finally, I wish to extend my thanks to my parents for their endless love, understanding and continuous support throughout my graduate study. xvii DEDICATION Dedicated to my parents for their endless love, support and encouragement 1 Chapter 1. Introduction 1.1 Background and problem statement Due to the growing demand on energy and awareness on sustainability, people are actively seeking alternative fuel sources and researching for clean energy. Renewable resources such as biomass are abundant and generally have positive environmental impacts, hence biomass appears to be an attractive feedstock (Karunanithy et al., 2013). Biomass can be converted into three main types of products, electrical/thermal energy, transportation fuel and chemical feedstock. Today, biomass utilization technologies such as combustion, pyrolysis, gasification, liquefaction and biochemical technology have undergone a lot of development (Baxter, 2005; Caputo et al., 2005; Di Blasi, 2008; Solantausta et al., 1992). In Canada, large amounts of lignocellulosic biomass in the form of forest and agricultural residues constitute renewable resources for conversion into biofuels. Woody biomass is the by-product of forest management, which include tree branches, tops of trunks, stumps, branches, needles, leaves and other woody parts left on the forest floor or landing after logging and thinning operations have taken place. Woody biomass is utilized to produce bioenergy and a full range of biobased products including lumber, composites, paper and pulp, furniture, housing components, round wood, ethanol and other liquids, chemicals, and energy feedstocks. Based on a study conducted by the Biomass Energy Centre (Biomass Energy Center, 2008-2011) in the UK, the price and energy output are compared between wood pellets, wood chips, natural gas, heating oil, bulk LPG, and electricity. Amongst all, wood pellet provides the highest energy output (4800 kWh/ton) at the second lowest price (4.2 pence/kWh). Wood chips (at 30% moisture content, wet basis) have the second highest energy output of 3500 kWh/ton with the lowest price at 2.9 pence/kWh. In contrast, coal has the heating value of 30MJ/kg with the price of 84.5 Euro/ton, when comparing to wood pellets of 18 MJ/kg. 2 Wood chips can be used as solid biofuel for direct combustion or making pellets. Traditionally, the raw materials for making wood pellets are sawdust, planer shavings and wood chips (Lehtikangas, 2001), and sometimes from woody residues left in the forest after logging. Wood pellets are renewable, clean, efficient and cost stable (Pellet Fuels Institute, 2011), and thus are widely used in North America and Europe (Wood Pellet Association of Canada, 2012). Forest materials have moisture contents as high as 50%. In order to produce wood pellets or other bioproducts, the reduction of the moisture content to 10% or somewhat lower is desirable. Biomass with high moisture content has a relatively low calorific value; thus moisture causes deterioration in pellet quality and increased cost of pelletizing. Also, moisture reduces the maximum combustion temperature and combustion efficiency (Maciejewska et al., 2006). Besides, high moisture content promotes microbial activities and raises the probability of self-heating, thereby leading to decomposition of dry matter during storage. For long-time storage, the moisture content of biomass has to be low, usually under 20% (wet basis) in order to improve the calorific value of biomass and keep dry matter losses moderate. Drying is usually applied before storage or other operations to reduce the moisture content of biomass. Conventional methods of drying include natural drying and artificial drying (over 60oC). Some reduction in moisture content may be achieved by natural drying, but the process takes a long time and weather is a major constraint. For artificial drying at high temperatures, biomass can achieve very low moisture content in a short time, but it requires substantial energy input. Hence artificial drying increases the operating costs and has impact on long-term environmental sustainability. The immediate use of forest residues after harvest is often unfeasible. The logging operations produce large quantities of biomass that are often left behind in the field for a period of time before they are transported and processed to make wood pellets or other forms of solid fuel products, depending on the production cycles. During the holding period, the properties of biomass residues can change due to physical, chemical and microbial processes. Some problems have been noticed during the storage: moisture sorption (environment dependent) and gas emissions (accompanied by dry matter losses). Self heating is another problem. 3 Biomass adsorbs or desorbs moisture continually from its immediate surrounding environment during storage in the field. When exposed to an environment with certain temperature and relative humidity, the materials will gain or lose moisture until reaching the equilibrium moisture content (EMC) (Zomorodian et al., 2010). The difference between the instantaneous and equilibrium moisture contents of the biomass represents the potential for moisture desorption (that is, drying) or adsorption. An increase in moisture content due to adsorption will lead to a series of problems to the materials including deterioration in quality, lower calorific value, higher transportation and processing costs, etc. It would be beneficial to estimate the time-dependent moisture sorption of the biomass when left in the field, and knowledge in this aspect of biomass storage is rather limited. The main emissions from woody biomass storage are carbon dioxide (CO2) and VOCs (volatile organic compounds); carbon monoxide (CO) and methane (CH4) emissions have also been reported. In terms of safety and occupational health, the threshold limit value- time weighted average (TLV-TWA) for 8h exposure to CO2, CO and CH4 are set at 5000 ppm, 25 ppm and 1000 ppm, respectively (Kuang et al., 2008). VOCs emitted from the process are potential air pollutants that have malodorous and even hazardous properties. According to Brosseau and Heitz (1994), VOCs in combination with nitrogen oxides (NOx) in the presence of sunlight are precursors to ground-level ozone production, and ozone has been identified as a key chemical substance affecting the environment via smog formation, as well as human health in terms of general toxicity and carcinogenicity (Brosseau & Heitz, 1994). With prolonged exposure to these substances, eye and throat irritation, damage to liver and central nervous system may occur. Some VOCs are major malodorous compounds (Eitzer, 1995; Komilis et al., 2004) and the odours are sometimes due to the synergistic action of these VOCs. Odours can induce indirect health effects such as nausea, vomits and reactions of hypersensitivity. 1.2 Objectives Storage is one of the key components in the biomass supply chain. During storage, the physical and chemical properties of the biomass change in terms of moisture content and gas emissions. This may lead to a deterioration in the biomass quality or affect the subsequent 4 process as well as the surrounding environment. This research is focused on the storage aspect of high-moisture woody biomass. The overall goal is to better understand the mechanism of moisture sorptions and gas emissions during storage, and to gain a deeper insight into various factors such as environmental conditions surrounding storage system. This will assist in the better management of fresh biomass. The specific objectives of the thesis are: 1) To investigate the sorption characteristics of woody biomass and develop a mathematical model to describe the relation between equilibrium moisture content and equilibrium relative humidity. This work can provide information on the stage of moisture in the material; it can also provide the relevant data and guidelines for expediting drying, and eventually terminating the drying process in a timely manner. This would contribute towards energy saving, as well as the planning and logistics of regional biomass storage. 2) To describe the wetting and drying of lignocellulosic biomass under natural conditions; and to develop a model to simulate the time-dependent moisture content of the materials during storage and apply the model to an experimental storage of biomass bales. This model can be used to simulate the moisture desorption and adsorption processes and thus predict the time-dependent moisture contents after being stored for a period of time under natural conditions, which can help to devise a better way to store and manage these high-moisture materials. 3) To study the composition of gas emissions from different types of woody biomass under different storage conditions and to quantify the emitted gases. The outcomes can help to determine the extent of emission control, and assess the impacts on human health and environment. 4) To measure the dry matter changes of woody materials during storage and investigate the relationship between gas emissions and dry matter losses. 5 The results shall demonstrate the relationship between dry matter losses and gas emission. Hence, gas emission control could also be based on minimizing percent dry matter losses during biomass storage. 1.3 Organization of the thesis To achieve the above-mentioned objectives, several series of experiments were designed and conducted in this study. The results are presented in seven chapters. Chapter 1 introduces the problems and knowledge gaps relevant to biomass storage, and thus defines the objectives of the thesis research. A comprehensive literature review is also presented in this chapter. Chapter 2 describes the sorption characteristics of woody biomass along with details of the desorption (drying) and adsorption processes. The moisture sorption isotherm is derived from the experimental data in order to understand the relation between equilibrium moisture content and environmental conditions for specific species of woody biomass. Chapter 3 depicts the cycles of wetting and drying of stored woody biomass under natural weather conditions. Changes in moisture content during storage are studied. With the lab-scale experimental results, the model is developed and calibrated to predict the time- dependent moisture contents of the biomass. Subsequently, in chapter 4, this model is applied to estimate the moisture content of biomass bales stored in the field during a one-year period under natural conditions. Chapters 5 and 6 present the results of gas emissions from the storage of different types of woody biomass. Different storage conditions are simulated with respect to temperature, initial moisture content and oxygen availability. Dry matter changes of the materials after storage, as well as the relationship with emitted gases are also presented in this chapter. Finally, Chapter 7 summarizes the main results and contributions of this thesis. Also, recommendations for future work are suggested. 6 1.4 Literature review 1.4.1 Biomass The potential threat posed by global warming problems due to high emission levels of greenhouse gases from fossil fuels combustion has become a major stimulus for renewable energy sources in general. Biomass and bioenergy as an alternative has received a lot of attention during the past decade. For different varieties of biomass, the contents of C, H, O are similar while the contents of N and ash-forming elements exhibit significantly different (Obernberger et al., 1997). The diversity of biomass has been identified, including woody biomass; agricultural biomass; aquatic biomass; animal and human wastes; contaminated biomass; industrial biomass wastes and biomass mixtures (Vassilev et al., 2010). Woody material is a major source of biomass. The source of woody biomass are mainly from the wood industry (paper mills, sawmills, and others) and urban wood wastes (tree trimmings, land clearance, construction and others). Forestry residues are also a significant source of wood residues (Easterly & Burnham, 1996). The main components of wood are cellulose, hemicelluloses and lignin. Cellulose is a polysaccharide with large molecular weight (usually around 100000). Hemicelluloses have much lower molecular weight, which are normally less than 30000. Hemicellulose consists of various monosaccharides. Lignin is primarily formed of phenylpropane (McKendry, 2002). Cellulose and hemicelluloses are the structural components in wood. Lignin connects the wood cells together. Cellulose constitutes 40-50% and hemicellulose 25-35% of wood. The content of lignin varies between softwood and hardwood, which is 25-32% and 18-25% respectively. Extractives, which make up less than 5% of wood, are not structural components. They include terpenes, fats, waxes, and phenols, most of which can evaporate easily during the heat treatment. Elemental compositions analysis shows the biomass contains around 50% carbon, 44% oxygen, 6% hydrogen and 0.1% nitrogen regardless of species (Barton, 1984). The woody biomass usually contains less ash, N and S and more C and Ca in comparison with other biomass. Woody biomass also has higher moisture content than the other biomass. Generally, woody biomass has 80% volatile matter and 20% fixed carbon (moisture free and ash free basis). Volatile matter is the portion that can be heated to 7 gas including water vapor. The fixed carbon content is the remained dry mass after combustion, excluding volatile matter and ash (Maciejewska et al., 2006). The volatile matter yield of biomass commonly includes light hydrocarbons, CO, CO2, H2, moisture and tars (Vassilev et al., 2010). When correctly managed, biomass is a sustainable low-carbon fuel versus fossil fuels. Biomass utilization has the following benefits: 1) Biomass is renewable and widely available source of energy; 2) It has a significant reduction in greenhouse gas and hence net carbon emissions; thus, it improves human health through better air quality; 3) Biomass has low contents of ash, C, S, N, and trace elements; 4) The cost-effective biomass energy plants can make a substantial, positive impact on regional economies; 5) The use of biomass fuel encourages an economic incentive to better manage forestry which improves biodiversity and reduces wildfire risk (American Renewables, 2012; Biomass Energy Center, 2008-2011; Vassilev et al., 2010). 1.4.2 Biomass storage Figure 1.1. Supply chain of biomass feedstock from harvest to biorefinery Figure 1.1 shows a typical supply chain of woody biomass feedstock from harvest point to biorefinery. A large proportion of the biomass residues are left in the forest after logging operations. These materials are then collected and gathered together. The immediate use of these biomass residues is often infeasible for various reasons. Therefore, prior to transportation the woody biomass is usually stored in the field for an extended period. Then, the materials are transported to the station for preprocessing, such as selection, comminution, 8 drying and so on. Afterwards, the materials are kept in the depot before transporting to biorefineries. To ensure continuous availability for biofuel production during the growing season and winter months, as well as heat production during the winter season, it is inevitable to store the high moisture biomass feedstock for a period of time up to one year through the supply chain. The prolonged storage of these biomass prior to operation could cause problems. One of the most difficult problems in biomass harvesting is how to store the material and how to avoid the mass losses due to degradation (Krupińska et al., 2007). Biomass decomposes both chemically and biologically over time. During the period of fast decomposition, there exist risks of emissions, energy losses, and fires (Wihersaari, 2005a). Microorganisms, especially fungi motivate the heat generation in the stored biomass. This heat will favor the growth of microbes and induce higher temperatures in the pile. In extreme cases, this can result in self-ignition and potentially fire (Jirjis, 1995). Some allergenic microbes are released to the air, which is associated with health risk. Furthermore, the loss of dry matter and carbohydrates has a negative economic impact (Wiselogel et al., 1996). Research studies conducted in the past have identified the following problems with the storage of biomass: 1) Moisture content. Depending on the form of storage, biomass may be exposed to elements of weather conditions in the field. Forest residues normally have moisture content of 45-55% and even up to 70% (wet basis). The most important fuel quality factor is moisture content, since it affects the calorific value, storage properties and transportation costs. It is taken into account in the pricing of the fuel (Pettersson & Nordfjell, 2007). Problems with the storage of biomass at high moisture content could become critical, including deterioration in quality, dry matter losses, fire risk, or even generating microbes that are harmful to human health (Rentizelas et al., 2009). Feedstocks with high moisture require more energy for processing and drying; this will affect the efficiency of the combustion process and emissions (Maciejewska et al., 2006). Another example is the increase in char yield during pyrolysis, resulting in a higher cost of thermochemical conversion. Dry biomass is more stable and safer to store, and easier 9 to preprocess, which are important for the successful operation of biorefineries (Singh, 2004). Thus, prediction of time-dependent moisture content is important and helpful towards better management of the fresh materials in the field. 2) Gas emission. Gas emission is one of the major problems during the handling process due to its effect on the surrounding environment and human health, as well as its relation with dry matter losses which affect the heating value of the materials. Thus, the investigation and quantification of gas emissions from woody biomass under different storage conditions become important. This knowledge can also assist in the better management of fresh materials. 3) Dry matter losses. Dry matter is the mass of material without moisture content. It is the source of energy when combustion. Thus, it is important to minimize the dry matter losses during storage. Previous research has dealt with dry matter changes under different storage conditions. Factors that exert influence on dry matter losses include temperature, relative humidity, precipitation, and storage period. Moreover, characteristics of biomass such as the type of feedstock, initial moisture content and particle size impose significant effects (Afzal et al., 2010; Casal et al., 2010; Jirjis, 2005; Nurmi, 1999). Two piles of chips were set up by Gjoelsjoe (1995) to study the effect of size on dry matter losses for the period May to December. Total dry matter losses of 8.7% dry weight were slightly higher in a pile of large chips (collected on 19mm screen), as compared to 7.5% in a pile of small chips (retained on 6mm screen). The moisture contents of both piles were reduced from around 40% (wet basis) to 30% after storage (Gjoelsjoe, 1995). In order to avoid re-wetting of feedstocks exposed to wet weather, Jirjis and Lehtikangas (1993) conducted research on the storage of residues covered with impregnated paper. The materials were piled into a windrow after felling and covered immediately. Results showed the fuel quality from covered windrows was improved as compared to uncovered windrows. The moisture content in covered windrows decreased by around 10% (wet basis). The dry matter losses from covered residues were below 1% per month (Jirjis & Lehtikangas, 1993). 10 Wihersaari (2005) reported that the material losses during storage were the highest in the beginning, right after the temperature has rapidly risen. The losses were estimated to be 3.6 wt% per week (measured during the second week of storage) and 0.4-0.7 wt% per week thereafter. His results also showed that the dry matter losses over a six-month period were twice as high for fresh forest residues versus dried forest residues (15.5% and 6.6% respectively). In general, the higher the initial moisture content of the stored feedstock, the higher the dry matter losses. Furthermore, smaller chip size led to less ventilation of the pile and eventually higher dry matter losses, whereas forced ventilation decreased the dry matter losses significantly (Wihersaari, 2005a). He suggested that it would be advisable to avoid storing biomass feedstock with high moisture content; hence, drying should be applied before storage. Biomass undergoes a change in moisture content, calorific value and dry matter content due to degradation processes and climate conditions during storage. Pettersson and Nordfjell (2007) studied fuel quality of logging residues (LR), with respect to moisture content and dry matter losses before and after large-scale storage and handling of compacted young trees. The experimental treatments involved the storage of uncompacted and compacted LR, with and without cover. Results indicated that the moisture content of LR declined to 18.0-20.7% for the covered parts of the windrows and to 18.8-24.9% for the uncovered parts during 9 and 12 months storage. They found dry matter losses of 8.7-11.7% and 14.4-17.4% after 9 months and 12 months of storage, respectively. Windrow with loose LR re-moistened to 40.8% by snowfall resulted in a 6% lower net calorific value as received, compared to the cylindrical bales with higher bulk density (Pettersson & Nordfjell, 2007). The effect of storage on fuel quality was evaluated with respect to moisture content, calorific value, particle size distribution, and ash content by Jirjis (2005). Results showed that temperature development was rapid and prominent in the chipped willow particularly in the 6 m high pile. By comparison, temperature rise was very slow in the chunk wood piles and became notably higher than ambient temperature after 2 months of storage in the 6 m high pile. In general, salix chips had a relatively lower moisture content and energy value than chunk wood by the end of storage (Jirjis, 2005). 11 Casal et al. (2010) observed the change in moisture contents of pine woodchips stored under the climatic conditions typical of Northern Spain for 12 months. While the moisture content of the pile increased with storage time, the woodchips underwent a slight deterioration especially during the first three months, but remained practically unaltered for longer periods of storage. The most important effect observed was a significant decrease in the heating value due to a marked increase in moisture content (Casal et al., 2010). Use of a breathable tarp to reduce moisture content of wood chips and dry matter losses was suggested by Afzal et al. (2010). Their study involved three forms of woody biomass - wood chips pile, bundle, and loose slash during a one-year storage period. Wood chips made from fresh birch wood stems were piled on forest floor, either covered with a breathable tarp to prevent precipitation, or uncovered. The rate of moisture content increment was lower in bundle form than in an uncovered wood chips pile. Loss of calorific value and dry matter loss were higher in wood chip piles as compared to the bundles at the end of the storage period. The maximum dry matter losses were observed in the uncovered wood chips pile (Afzal et al., 2010). Nurmi (1999) investigated the effect of storage on the fuel wood properties of Norway spruce logging residues (Nurmi, 1999). The stored materials include uncomminuted residue piles on the clear-cut, uncomminuted residues in large windrows at road-side landing and comminuted residue piles at a terminal. After one year storage, the moisture content of residues at the clear-cut and the landing had decreased from 56% to 28.5 and 42.2%, respectively. The moisture content of comminuted materials had risen to 65.3% after 9 months. The needle content of the residues decreased from 27.7% to 6.9% and 18.9% on the clear-cut and the landing, respectively after one year. The content of carbon in the comminuted materials had little change from 50.0% to 51.2%. The effect of size on wood chips storage was studied by Heding et al. (1993). Results reveal that heat generation in the piles with large particles (large chunk and firewood) was small. By comparison, temperature in the piles of smaller particles (chips and fine chunk) increased higher. And the heat generation was less dependent on surrounding environment. 12 Lower final moisture content was observed in the piles with larger particles as the moisture content decreased by 9.5% (Heding et al., 1993). Eriksson & Gustavsson (2010) compared a bundle system with a chip system for collection and transport of biomass residues in Sweden (Eriksson & Gustavsson, 2010). The particle size was observed to have a negative relation with dry matter losses and moisture contents. They found that the bundle system had less dry matter losses and 5-10% lower costs than the chip system. Bundles made from both green and brown residues had a reduction of moisture content by about 6%. Total dry matter losses in the bundles after 5.5 months were 5.8% for the originally green residues and less than 1% for the originally brown residues. However, after 8 months the dry matter losses increased to 11.5% and 8.7%, respectively. The moisture content of chip pile has a large influence on dry matter losses. Tests carried out in Sweden showed that the initial moisture content was proportional to the losses. Specifically, the initial moisture content of 42%, 51% and 58% in chips was associated with monthly dry matter losses of 1.1, 2.2 and 2.6 wt%, and total dry matter losses during the 6- month storage period of 6.6, 13.2 and 15.6 wt%, respectively (Thörnqvist, 1983; Thörnqvist, 1984). In another 9-month study by Thörnqvist (1984), the initial moisture content of 32 wt% and under 20 wt% in chips led to monthly losses of 1.03 and 0.23-0.35 wt%. According to Nurmi (1999), the temperature in a pile of chipped forest residue either fresh or naturally dried usually rose rapidly after an initial period of approximately one week, which is a definitive sign that the materials began to decompose leading to losses in materials and energy value (Nurmi, 1999). The moisture content of wood residues varies widely depending on the materials and storage circumstances. For instance, forest residues had a moisture content of 45-55 wt% upon delivering (Wihersaari, 2005b); after 6-month storage as “compost heaps”, the forest residues were naturally dried to 40 wt% moisture content. Jirjis (2003) studied cylindrical bales of green forest residues having an initial moisture content of 31-38%. After ten months, the moisture content decreased to 21% for the outdoor, covered bale and indoor bale, while the outdoor, uncovered stack showed only slight decreases. The percent dry matter losses due to biological activity were highest in 13 the outdoor, uncovered stack with an average value of 18.5%, as compared to a value of 14% in the other two stacks (Jirjis, 2003). He concluded that storage of newly harvested logging residues in bales for ten months could produce a fuel with acceptable quality; however, intensive microbial activity leading to high substance losses and reduction in total energy content can be problematic. 1.4.3 Moisture sorption characteristics Wood contains water in two forms: bound water and free water. Bound water exists in the cell wall and is hydrogen bonded to the free hydroxyl groups in cellulose, hemicelluloses, and lignin. Free (capillary) water is the bulk of water contained in the cell lumens and voids of the wood. The amount of free water is limited by the void space of wood. It is held by capillary forces, not bound by hydrogen bond. Energy is required to overcome the capillary forces. The change of free water will not cause changes in physical properties (swelling, shrinkage) because the cell wall is saturated by bound water. Biomass containing abundant free water is treated as high moisture content material, normally higher than 30%. The fiber saturation point (FSP) expresses the value of moisture content at which all of the free water is removed. The cell cavities are empty while the cell walls are still completely saturated (Krupińska et al., 2007). When the moisture content of biomass is lower than FSP, the microbial activities will be inhibited (Gislerud, 1990). For most woody biomass, an average value of FSP is 0.3 (dry basis) (Siau, 1984). For willow, the fiber saturation point achieves a value of 0.25 (Gigler et al., 2000a). 1.4.3.1 Drying Biomass is expected to be stored under ambient atmospheric conditions or in an enclosure due to its bulky nature. The moisture content of feedstock as well as storage type can lead to deterioration of the materials. Moisture content exerts a strong influence on biomass harvest, preprocessing, transportation, storage, conversion and the resultant products. The moisture content for safe storage depends upon the type of feedstocks. For long-term storage, the moisture content should be below 17.5% (dry basis) for most of the feedstocks so there is no 14 proliferation in microorganisms development (Arabhosseini et al., 2010; Karunanithy et al., 2013; Kudra & Strumillo, 1998). Low moisture content enables long term storage with low microbial activity, and thus low dry matter losses, and reduces health risks. It also increases conversion efficiency into electricity and reduces gas emissions. In the biochemical or thermochemical conversion of biomass into biofuels, moisture content is also of significant interest. High moisture content can affect the applications of biomass for thermo-chemical conversion processes including combustion. Drying of biomass is recommended, even required in order to easily handle with, store the materials safely, reduce the cost of transportation, increase boiler efficiency, increase the efficiency of thermal applications, reduce gas emissions and achieve desired quality of product (Gigler et al., 2000a). Drying is a process involving heat and mass transfer. The drying process may change the quality of product. Drying occurs by heating the materials in terms of convection, conduction and radiation (Mujumdar & Devahastin, 2000). Drying is necessary whether the biomass is used in densified form (such as pellets and briquettes) or non-densified form. An increase in temperature leads to the activation of water molecules. This will cause the water molecules to become less stable and to break away from the water binding sites of the material, thus decreasing the moisture content (Chowdhury et al., 2006). Drying of biomass occurs in three phases for high-moisture materials. During the initial phase, the rate of drying increases due to increase in temperature with some free moisture being removed. In the second phase, free moisture is evaporated from the saturated surface; the drying rate is high and essentially constant. This phase is also recognized as constant-rate stage. The third phase corresponds to the falling-rate drying period; the area of the saturated surface gradually decreases as the moisture movement within the solid can no longer supply enough moisture to wet the surface. The importance of biomass drying has motivated a number of works on drying techniques. The drying methods can be briefly divided into two categories: the commonly used procedures of air drying and kiln drying, and the specialized techniques using chemicals, solvents, vacuum retorts, solar energy dehumidifiers, high frequency generators, and so on 15 (Bousquet, 2000). The description of different drying methods together with the advantages and disadvantages are listed in Table A1. The merits of advanced drying methods are fast drying rate, high drying efficiency and good control. However, most of them have expensive operating costs and capital investment. A wide variety of woody biomass and the forest residues generated by logging operations have lower calorific values and complex composition compared to densified biomass. The high energy cost makes it not feasible to set up an advanced drying facility in or close to the field, and this encourages the search for more economically attractive techniques (Moreno et al., 2004). Conventional methods of drying include natural drying, low-temperature drying and high-temperature drying. High-temperature drying requires substantial energy input and specialized mechanical equipment. Low temperature drying can be used for forest residues because of its low loss of the volatile compounds (Stahl et al., 2004). Low temperature drying is easy to operate, but it may require a long time to reach the desired final moisture content. Some reductions in moisture content may be achieved by natural drying, but the process generally takes a long time as weather is a major constraint. The air drying or low temperature drying of forest residues in site are possible, which can improve the energy density of the residues for subsequent thermal application (Phanphanich & Mani, 2009). Thus, optimized operation of natural drying and low-temperature drying is needed to achieve better drying efficiency and energy saving. The study of drying can help to better understand the drying mechanisms, predict drying rates, and possibly modify the factors that affect drying. The factors that affect drying are the surrounding environment (air temperature and humidity, wind, precipitation, solar radiation and soil moisture); the properties of materials (species, maturity and yield) and the treatment methods (chemical and mechanical). The drying rate was found to be positively related to solar radiation and vapor pressure deficit, whereas it was negatively correlated to the thickness and initial moisture content of the materials (Savoie & Mailhot, 1986). A mathematical model is an appropriate tool to help understand the drying process of biomass. The drying process can be predicted by appropriate models with product mass, 16 drying characteristics and drying air conditions as the parameters. Numerous models have been studied to describe the drying process of different biological materials (Jayas et al., 1991). With a simulation model, the drying time and cost of drying can be calculated with the information of moisture removal rate of the drier, the initial and final moisture contents and the fuel consumed. Moisture diffusion models have been developed by researchers for the theoretical analysis of the physics of drying. Complex partial differential equations that encompass mass balance, heat balance, heat transfer and drying rate are solved with numerical techniques. These models have been primarily developed to study high temperature drying, though some have subsequently been applied to low temperature drying (Sharp, 1982). They have shown more accurate than the thin-layer models, but they require much computer time. Using simple models (thin-layer drying models), drying rates can be readily estimated. Some of these models have been found to be useful and sufficient by research scientists who studied the drying of agricultural crops (Akpinar et al., 2003; Bruce, 1985; Ertekin & Yaldiz, 2004; Sacilik & Elicin, 2006). Besides, there were previous studies that involve biomass as feedstocks; for instance, corn, alfalfa, flax fiber, and willow chips and stems. Thin layer drying of pine residues with bark, needles, leaves, and chips was studied by Phanphanich and Mani (2009) using three different drying models (Lewis, Page, and Henderson and Pabis equations) (Phanphanich & Mani, 2009). These equations have also been successfully used in agricultural materials and other biomass feedstocks (Yang et al., 2007). The Lewis model assumes that the moisture within the material can move to the surface without resistance. The drying rate is proportional to the moisture difference between the material being dried and the equilibrium moisture content (Lewis, 1921). Page’s model is an empirical model and is the modification of Lewis model. A parameter was added to time to improve the original model (Gigler et al., 2000b). It has been utilized to predict the thin layer drying of grain and rough rice, white bean, barley (Vijayaraj et al., 2007), pistachio, peanuts (Yang et al., 2007) and rapeseeds (Panchariya et al., 2002). Page’s model has also been modified to better fit the drying results of different biomass such as soybean (Overhults et al., 1973). The Henderson and Pabis model was developed on the basis of diffusion 17 process. During the falling rate stage of drying, the drying rate is related to effective diffusivity. This model has been applied to numbers of agricultural materials. The Lewis model is the special case of the Henderson and Pabis model when the intercept of the line (moisture content vs time) is zero. 1.4.3.2 Adsorption In general, woody biomass is hygroscopic in nature. It can desorb or adsorb the moisture from the surrounding atmosphere (Singh, 2004). The hygroscopic phenomenon in wood is a chemical interaction between the water molecules and the structure of cellulose (Merakeb et al., 2009). When the moisture content of materials is relatively low, or relative humidity of the air is high, the materials adsorb moisture from the air in order to attain the equilibrium between the materials and environment. Adsorption is an opposite process to drying. Various physical properties as mass, dimensions and density, as well as its mechanical properties are affected by the moisture content (Droin et al., 1988). Absorption of water has a negative effect on wood quality in terms of facilitating fungal attack (Baronas et al., 2001). To predict and describe the moisture content, few models including theoretical, empirical and semi-empirical approaches have been developed to simulate the moisture adsorption of biomass. The theoretical models include the diffusion equations based on Fick’s second law. These models involve numerous functions and complex computations, which can not be carried out in some cases (Khazaei, 2008). At this point, empirical models are preferred since they are easy to operate and interpret. The empirical models that have been popularly used to simulate water adsorption processes of biomass include Peleg model, Exponential model and Weibull models (Gowen et al., 2007). Peleg model, which has been applied to adsorption processes of different kinds of foods, is a two parameter and non- exponential equation (García-Pascual et al., 2006). The Weibull distribution model estimates the adsorption process with two parameters. The Weibull model showed good fit for the description of water adsorption of a variety of dried foods, and adequately described adsorption processes controlled by different mechanisms (Marabi et al., 2003). 18 1.4.3.3 Equilibrium moisture content Feedstocks are subject to different temperatures and relative humidity (RH) during harvesting, preprocessing, transportation, and storage in a wide variety of climates. Knowledge of the relationship between the air temperature, relative humidity and the moisture content of biomass can help to correctly dry and store the materials in order to preserve the quality of feedstocks (Nilsson et al., 2005). The moisture content of a material in equilibrium with the environment with respect to temperature and relative humidity is termed the equilibrium moisture content (EMC) (Baker, 1997). The EMC of a product is the final result of moisture exchange between the product and the air surrounding the sample (Arabhosseini et al., 2010). In this condition, the water in the material is in balance with the moisture in the surrounding environment (Silakul & Jindal, 2002). To optimize the drying and storage processes, an accurate knowledge of EMC is essential. EMC would allow specifying the progress of the drying process and determining whether the feedstock will gain or lose moisture under known storage conditions (Zomorodian & Tavakoli, 2007). The traditional method of removing moisture is to use dryers. Proper design of dryers requires information about the EMC of the biomass (Singh, 2004). The equilibrium moisture content of biomass is related to several parameters. It depends on the temperature and relative humidity of the surrounding environment (Karunanithy et al., 2013), the species, variety (Brooker et al., 1992), degree of maturity (Hartley & Avramidis, 1994), porosity and microstructure (Choudhury et al., 2011), specific surface area (Arslan, 2006), amount of extractives and strength of feedstock/wood (Moreno et al., 2004), and type of processing or treatment the feedstock was subjected to (Acharjee et al., 2011). 1.4.3.4 Moisture sorption isotherm Plots of the equilibrium moisture content per unit dry mass versus relative humidity at constant temperature is called adsorption or desorption isotherm, depending on wetting or drying sample preparations, respectively (Arslan, 2006). Desorption isotherm is the 19 equilibrium process from wet materials and reach equilibrium by losing moisture. It determines the lowest attainable moisture content of biomass at a particular drying temperature and relative humidity. Adsorption isotherm is the equilibrium process from dry materials and reach equilibrium by adding moisture (Arabhosseini et al., 2010; Baker, 1997). Therefore, sorption isotherm represents thermodynamic equilibrium between the biomass material and environment, and it is necessary for determining the optimum moisture conditions for storage stability, drying, and other processes involving humid air and woody materials (Krupińska et al., 2007). Drying processes require a lot of energy consumption. Thus, with the knowledge of sorption properties of the biomass, the drying time can be shortened as well as energy saving. The sorption isotherms give information about the equilibrium of the material and help to understand the stability of the material after drying. This determination is crucial to determine the dryer’s thermal efficiency, drying rates, product heating and material quality (Iguaz & Virseda, 2007). The knowledge of moisture sorption isotherms is also valuable in solving engineering problems such as equipment design, drying and storage processes (Arogba, 2001; Mohamed et al., 2005b). More than 200 isotherm equations have been developed theoretically, semi- theoretically or empirically to model the relationship between EMC or ERH and temperature of different biological materials (Van den Berg & Bruin, 1981). Different types of models include monolayer models and multilayer models. These models involve semi-empirical models, empirical models and theoretical models (Krupińska et al., 2007). They have been used to predict the equilibrium moisture content of the feedstocks starting from harvesting, drying and preprocessing through transportation, storage, and processing. Much work have been done on the moisture sorption isotherms of different biomaterials (Boquet et al., 1978a; Cassells et al., 2003; Soysal & Öztekin, 1999; Zomorodian et al., 2010; Zomorodian & Tavakoli, 2007). The Henderson equation was built upon the Gibbs' thermodynamic adsorption model; while the Oswin equation expanded the sigmoid shaped curves mathematically (Sun, 1999). Subsequently, the Modified Chung-Pfost Equation was developed. It was assumed that there is relationship between the energy change and moisture content during sorption 20 process. The Modified Henderson Equation and Modified Chung-Pfost Equation are found to be the most appropriate models for starchy grains including rough rice and fibrous materials and barley (Basunia & Abe, 2005; Basunia & Abe, 2001). The EMC/ERH sorption isotherms represented by the Strohman-Yoerger equation work well for rice (Sun, 1999), whereas Henderson's model was a better predictor of the biscuit isotherm (Arogba, 2001). The Modified Halsey equation is recommended for high oil and protein products. The Modified Oswin equation is a good model for popcorn, corncobs, red beans, soybean, whole pods of peanut and some varieties of corn and wheat (Chen & Morey, 1989). The GAB model is adequate to describe the experimental data for amaranth (Pagano & Mascheroni, 2005). For switchgrass and prairie cord grass, the experiment results fit Modified Halsey as the best model followed by the Modified Oswin equation (Karunanithy et al., 2013). Sun and Woods (1994) conducted a study to fit more than 1000 data points of wheat to a number of EMC/ERH equations in order to compare the different equations. Result shows the Modified Chung-Pfost equation is the most appropriate equation for wheat (Sun & Woods, 1994). Furthermore, the Modified Chung-Pfost equation, Modified Oswin equation and Modified Halsey equation were exhibited to fit best for the sorption of wheat, shelled corn and rapeseed, respectively (Sun, 1998; Sun & Byrne, 1998). The modified Henderson equation showed the worst fitting for these materials. Nevertheless, Lahsasni et al. (2002) conducted an experimental study on modeling of sorption isotherms of prickly pear peel (Opuntia ficusindica), and found the BET and Henderson models as the best fit models to their EMC experimental data (Lahsasni et al., 2002). Four equations (Modified Henderson, Modified Chung-Pfost, Modified-Halsey and Modified-Oswin equations) were chosen by the American Society of Agricultural and Biological Engineers for use in their Standards. All of the equations have three coefficients and can be interpreted easily as a function of temperature and relative humidity (ASABE, 2006). The GAB equation was also recommended in ASABE; however, it does not include the effect of sorption temperature. Each of the four models have successfully predicting the EMC for biological materials under certain relative humidity and temperature (Boquet et al., 1978b). 21 Although several mathematical models exist to describe moisture sorption isotherms of wood materials, none of these equations has been found to be suitable to describe the EMC/ERH relations for various types of woody material accurately in a large range of relative humidities and temperatures. Since wood has a sophisticated structure and chemical composition (cellulose, hemicellulose, lignin) and it differs among the species (softwoods, hardwoods) (Krupińska et al., 2007), it is necessary to find the most suitable EMC/ERH equation for each specific feedstock (Chen & Morey, 1989). The sorption isotherms are characterized by three zones. In the first region, water is tightly bound to the cell and unavailable for reaction. Monolayer of water exists in the cell. In the second region, more water presents and it is bound loosely. In the last region, multilayer of water appears and is held in capillaries (Mujumdar & Devahastin, 2000). The shape of sorption isotherms for hygroscopic materials are indentified to be sigmoid (Merakeb et al., 2009). Depending on adsorption and desorption processes, the amount of water at any relative humidity may be different, a phenomenon known as hysteresis (Lahsasni et al., 2003). Under the same environment condition, the direction of sorption can result in different moisture contents of biomass. The desorption curve lies above the adsorption curve. That is, the moisture content from the desorption process is higher than that from adsorption under same condition. In most cases, the hysteresis effect decreased and the sorption isotherms shift downwards with increasing temperature (Krupińska et al., 2007). Researchers have explored some explanations for the hysteresis phenomenon in hygroscopic materials. One of them is related to the change of the active polar sites where the water molecules are bonded to. Originally when the material is wet, the polar sites are filled with water. When the material is dried, the water molecules and the sites are held more closely to each other, which will reduce the holding capacity of water to the material during the subsequent adsorption process. Thus, the moisture content of material is higher from desorption curve as compare to that from the adsorption curve (Zomorodian & Tavakoli, 2007). 22 1.4.4 Gas emissions from stored biomass Organic materials are subject to decomposition over time during storage, principally due to either biological (anaerobic or aerobic) or auto-oxidative process. The extractives in the woody biomass degrade more readily, and these compounds evaporate from the wood during storage. The main emissions from these materials are carbon dioxide (CO2), carbon monoxide (CO), methane (CH4) and volatile organic compounds (VOCs). Carbon monoxide is a leading cause of chemical poisoning in both the workplace and at home. CO2 and CH4 are greenhouse gas (GHG) emissions. Total GHG emission in 2010 was 6,821.8 million metric tons CO2 eq, which was estimated as 84.6% from CO2, 7.9% from CH4, 5.5% from N2O and 2% from hydrochlorofluorocarbons, chlorofluorocarbons and sulfur hexafluorides (EPA, 2012). 1.4.4.1 Emission of carbon-based gases When organic materials are stored in confined space, the gas emissions from decomposition may accumulate and eventually reach toxic levels. The Threshold Limit Value (TLV) of a chemical substance sets the concentration level in the environment, where a worker can repeatedly expose to it safely. TLV is a reserved term of the American Conference of Governmental Industrial Hygienists (ACGIH). It is commonly used in the field of occupational health and toxicology. The TLV of carbon dioxide, carbon monoxide and methane are listed in Table 1.1 (ACGIH, 2004; Cairelli et al., 1994; EnviroMed Detection Services; Ontario Ministry of Labour, 2012; The National Institute for Occupational Safety and Health (NIOSH), 2007). The Time Weighted Average (TWA) is TLV based on an 8-hr workday and a 40-hr workweek. For example, the 8-hr TWA for CO is 25 ppm, meaning that an average of 25 ppm is considered to be the safe TLV for an 8-hr workday. Short Term Exposure Limit (STEL) is TLV based on a 15-min average. Concentrations that are immediately dangerous to life and health are quite high for these three compounds. 23 Table 1.1. The Threshold Limit Value of carbon dioxide, carbon monoxide and methane Chemical Substance TWA (8-hr average) STEL (15-min average) Immediately Dangerous to Life and Health Carbon dioxide 5,000 ppm 30,000 pm 40,000 ppm Carbon monoxide 25 ppm 100 ppm 1,200 ppm Methane 1000 ppm - - Boddy (1983) studied the effects of temperature and moisture content on the gas emission from wood under aerobic conditions. Results showed that increases in temperature and moisture content of the materials led to an increase in CO evolution (Boddy, 1983). When higher temperatures occurred with high moisture contents, CO evolution leveled off or decreased, which was attributed to a decline of O2. The decomposition of wood by microorganisms releases CO2, H2O and heat leaving chemically altered wood and the tissues of the decomposer organism. Respiration rate also increased linearly with increasing moisture content, although at very high moisture content the CO2 evolution curve shifted downwards. Gas emissions from stored biomass as affected by various factors have been studied in the past few years. Wihersaari (2005) found GHG emissions almost three times higher in case of the fresh versus dried forest residues. The potential amount of CH4 emission seemed to be of larger concern than the N2O emission. Besides, he suggested storage heaps should not be mixed or moved during the storage period, as this would probably make the decomposition process more intensive, which causes increase in emission rates. Eriksson and Gustavsson (2010) studied gas emission from bundled forest residues. They pointed out that the bundle system had higher primary energy use and CO2 emissions, but the lower dry-matter losses in the bundle system chain give CO2 emissions per delivered MWh almost as low as for the chip system. The Finnish bundle system with its more effective compressing and forwarding emitted less CO2 emissions than the current Swedish 24 bundle system, but with the theoretical improvements considered here the Swedish bundle system will emit less than the current system (Eriksson & Gustavsson, 2010). Wood pellets with high energy density are used as a high quality feedstock for manufacturing liquid fuels and chemicals. A major existing or potential issue is the self- heating of these pellets either at the terminal during prolonged storage or during ocean transport (Feist et al., 1973). The long-term storage of wood pellets in enclosed spaces might lead to accumulation of compounds that are either toxic or causing asphyxiation (CO2, CO, CH4, VOCs) due to microbial activities and/or chemical reactions. The resulting depletion of oxygen along with emission of gases can endanger the life and health of workers. Incidents of injuries and even fatalities have occurred among workers in recent years (Svedberg et al., 2008). Kuang et al. (2008) monitored CO, CO2 and CH4 off-gases (gas emissions) and oxygen depletion from stored wood pellets in ocean vessels. They postulated that emissions are likely due to biodegradation of lipids and fatty acids, and auto-oxidative reactions involving other organic constituents naturally present in wood (Kuang et al., 2008). In a lab- scale study by (Kuang et al., 2009), they found the higher peak emission factors for CO2, CO, and CH4 were always associated with higher temperature and increased humidity in the headspace of the reactors. Other researchers have suggested that the oxidation of unsaturated fatty acids in wood can be a reason for gas emissions from wood pellets and it is significantly influenced by storage temperature (Shankar et al., 2008; Svedberg et al., 2004). In this regard, Svedberg also identified high levels of hexanal and carbon monoxide emissions caused by the degradation of wood. The generation of CO2 along with oxygen depletion are mainly caused by microbiological activity pertinent to wood pellets, while the CO generation is attributed to the chemical oxidative processes. During the transportation of logs and wood chips in confined space, complete depletion of O2 was observed only after 37 hours. The CO2 concentrations ranged from 0.5 to 15%, while CO concentrations were from 2 to 174 ppm (Svedberg et al., 2009). The depletion of oxygen in the cargo was suggested to be the result of microbiological activities (CO2 formation) and chemical oxidation of wood. High concentration of CO2 indicated the intensive microbial activities in logs and wood chips. This showed a big difference with their previous findings with stored wood pellets. For wood 25 pellets, the auto-oxidative degradation processes was dominant (Svedberg et al., 2008). Only 70% of consumed O2 was observed to convert into CO2 during the sea transportation, which was suggested to be due to the different solubilities of CO2 and O2. Some of the CO2 is maintained in water within the wood as carbonic acid (H2CO3). Temperature affects the equilibrium between CO2, H2CO3 and the solubility of CO2. He et al also found that CO2 decreased in the form of carbonic acid in the materials during the storage of the Douglas fir branches (He et al., 2012). 1.4.4.2 VOCs emission VOCs are made up of a wide range of organic compounds with vapour pressure greater than 0.01 kPa at 20oC (VOC-directive EU, 1999). They are also characterized by their low water solubilities. Their great mobility make them capable to be inhaled by people working or living in places with high concentrations (Das et al., 2004; Domingo & Nadal, 2009). The major VOCs emitted from wood pellets are aldehydes, some of which are known to cause irritation to the respiratory system (Hagstrm, 2008). Furthermore, Kuang et al. (2009) found higher peak emission factors were always associated with higher temperature and relative humidity in the headspace of reactors. Arshadi et al. (2009) suggested a high temperature used during the drying of sawdust would subsequently lead to higher emissions of aldehydes and ketones from the manufactured pellets (Arshadi et al., 2009). It was also reported that high levels of hexanal and pentanal together with minor quantities of other aldehydes were detected in softwood pellets storage (Arshadi & Gref, 2005). Some organic acids such as acetic acid are likely to be emitted from the breakdown of wood hemicellulose (Johansson & Rasmuson, 1998). According to Stahl et al. (2004), the release of VOCs was rapid early in the drying process, with a small second emission peak at 10% moisture content (Stahl et al., 2004). Increased drying temperature increases the total amount of VOCs. (Leinonen & tutkimuskeskus, 2004) mentioned the considerable amount of VOCs emitted from woody biomass are mainly terpene compounds. In comparison with conventional drying method for wood, Beakler et al. (Beakler et al., 2005) measured Total Organic Compounds released from drying of hardwood and found that the type of wood affects the level of released TOCs. The results indicated that mixed red oak and white oak 26 lumber with initial moisture content of 21% released the highest amount of TOCs. The common VOC produced during wood drying is potentially carcinogenic other than its short- term health effects such as eye and throat irritation (Granstrom & Mansson, 2008). Monoterpenes, mainly α-pinene were found from logs and wood chips (Svedberg et al., 2009). Terpenes often treat as attractants for wood-destroying insects. During the long- term, they are released to the environment and their quantities in wood decreases. In a study performed by (Kačík et al., 2012), the terpenes in “recent fir wood” were found to be about 60 times higher than the old wood from the 17th century (ratio of 186:3 mg/kg). Thermal wood treatment accelerated the release of terpenes. (Rupar & Sanati, 2005) investigated wood chip piles in a terminal storage located in southern Sweden, stretching from June through January. The release of terpenes from the bark/wood chips pile was found to be high in the middle of the storage period and low in the beginning and end of the storage period. Air emission increased when the temperature directly above the pile increased. More terpenes were released when wood chips were mixed with bark, especially when the amount of precipitation increased. (Hoell & Piezconka, 1978) and (Piispanen & Saranpaa, 2002) found that the polyunsaturated acid, linoleic acid comprise most of the free fatty acids and triglycerides in wood. The oxidation of linoleic acids and its esters produces hexanal as the main VOC component, which was found by (Back & Allen, 2000). These reactions may be either enzyme-induced or happen through an auto-oxidation process (Frankel et al., 1989; Noordermeer et al., 2001; Schieberle & Grosch, 1981). Hagstrm (2008) found that pellets, under certain conditions, emitted high levels of VOCs. The major VOCs emitted from wood pellets are aldehydes, some of which are known to cause irritation to the respiratory system (Hagstrm, 2008). 1.4.4.3 Microbial activities The use of woody biomass as an alternative energy source has led to an increase in the number of wood piles in open environments. The presence of these piles may have an adverse impact on the ambient air quality because of microorganisms on the wood. Wood 27 materials have been found to support the growth of a wide range of microorganisms, particularly fungi. With indoor storage of wood, these microbes can become airborne and be inhaled by workers. Many of the fungi, actinomycetes and other bacteria have been reported to be potentially pathogenic or toxigenic. Inhalation of these microorganisms at high concentrations can lead to serious allergic reactions or pulmonary diseases (Hellenbrand & Reade, 1992). Under certain conditions, especially in the presence of sufficient moisture, wood is a good substrate for microbial growth. Wood chips with large surface area also have more extensive microbial growth. Large wood chip piles often heat up spontaneously due to microbial activity which is responsible for dry matters losses. Heat is generated by the decomposition of wood by microorganisms. Once the temperature reaches a certain point, chemical reaction occurs. This reaction produces more heat and raises the acidity of the pile. The role of rapidly developing organisms dominated the early stages of deterioration. At higher temperature, chemical reactions are mostly active and dominated (Fuller, 1985). Wood deterioration is due to three types of microorganisms in the wood cells, including decay fungi, staining fungi and bacteria. Decay fungi, involving white rot, brown rot and soft rot, can metabolize the wood cell wall constituents (both cellulose and lignin). They affect the strength of wood and even cause the complete destruction of the wood. Staining fungi and molds inhabit in sapwood and obtain nutrients from the xylem parenchyma and discolor the wood, while bacteria can consume the parenchyma cells of wood. All of the microorganisms metabolize the wood substances to gain the nutrients (Scheffer, 1966). The principal factors that influence infection of the chips are temperature, moisture, oxygen conditions in the piles and storage time. Chip size and tree species appeared to have little influence on fungal growth (Bjoerklund, 1983). The survival of most fungi is temperature dependent. They can be killed when exposed to 65oC for several hours or 60oC for a longer time. The moisture content is also known to be one important factor to fungi. Minimum and optimum moisture are about 18 and 28-45% (w.b.), respectively, whereas maximum moisture ranges from 60-75%. Wood-attacking fungi are mostly aerobic; hence an 28 adequate amount of oxygen is essential for fungal growth. For some particular fungal species, it was found that the optimum water activity for rapid growth was the lowest at temperatures close to the maximum (Ayerst, 1969). The microbial activity of wood chips, logs and wood pellets was assessed. It was found to be high in fresh wood chips and bark. By comparison, both dry and wet wood pellet showed none microbiological activity. The number of microbial counts were further supported the findings with high value in wood chip and bark samples and none in wood pellets (Svedberg et al., 2009). The reasons are supposed to be the difference in moisture contents. Fresh logs and wood chips have moisture content around 50% (w.b.) while wood pellets only have 8% (w.b.) after drying process. Microbiological activity is temperature dependent; high temperature during drying process kills the microorganisms. For pellets, both high temperature during the pelletizing process and the low moisture content are unfavorable for microbiological activity. Madsen et al. (2004) reported the similar results on straw and wood chips. They found high concentrations of bacterial in dusts from straw and wood chips between 8 x104 and 3.1 x106 cfu/mg dust, and very low bacterial counts in dusts from briquettes and wood pellets between 20 and 60 cfu/mg dust. In an earlier study, Feist et al. (1973) found bacterial populations in wood chip piles as high as 5 × l08 cfu/g dry wood, and they contributed significantly to the self-heating of the pile (Feist et al., 1973). Different species of wood have been stored to study the microbial activity on mass losses. The losses in rough pine stored in summer (April to October) ranged from 2-4% (2 months), 5-8% (4 months), and 7-10% (6 months). During winter storage (October to April), reductions due to decay were about one-third of those in the summer. Overall, for a full year of storage, loss of materials was 11-15%. For aspen and balsam fir, the density losses were 25-30% during 4-year storage which are higher than those for jack pine and spruce. The weight reductions in rough jack pine were approximately 5% after 1 year and 9% after 2 years (Lindgren & Eslyn, 1961). Greaves observed biodeterioration in the form of mass losses in tropical wood substance as a result of microbiological activity. The wood substance loss amounted to l.5% 29 per month. A number of microflora species are found during storage. Temperature was found to increase in chip piles, which influenced the trend of microorganisms (Greaves, 1975). 1.5 Concluding remarks Based on literature review, it may be concluded that little previous work has been done on moisture adsorption and desorption characteristics and gas emissions from fresh logging residues and wood chips during storage. Although the sorption characteristics of wood in terms of logs and sawdust have been studied by several researchers, there is little information on the sorption behaviour of forest residues. Moreover, there is a lack of sorption data for fresh biomass, especially for wood from energetic plantations. Conducting experiments, the development of mathematical models that represent these processes, and application of the calibrated models are considered an appropriate approach to address these knowledge gaps. 30 Chapter 2. Moisture sorption characteristics of biomass during storage 2.1 Introduction Due to the non-renewable nature of fossil fuels and their negative effect on environment, there is an increasing demand for alternative fuels including biomass, such as those derived from forest and agricultural residues. These materials may be used as solid biofuels for combined heat and power generation, or processed in biorefineries to produce liquid biofuels primarily for transportation. Usually, the immediate use of lignocellulosic biomass after harvest is infeasible. The logging operations produce large quantities of high moisture residues that are left behind in the forest. Traditionally the material is either left to rot or burned intentionally to reduce the risk of wild fires. New efforts are under way to salvage the logging residue by chipping it and using it either directly in nearby boilers or pelletizing it for long distance transport and ease of handling. Aspen (Populus tremuloides) that grows in northern climates can be a readily available woody biomass source for chipping and making fuel pellets and animal bedding. Its moisture content varies from 80 to 115% on dry basis (or, 44 to 54% on wet basis) depending on the season (Jensen & Davis, 1953). High moisture content increases the cost of transport and pelletizing, which will be reflected in the pricing of fuel (Pettersson & Nordfjell, 2007). Besides, moisture reduces the maximum combustion temperature and combustion efficiency (Maciejewska et al., 2006). Problems with the storage of biomass at high moisture content could become critical, including the deterioration in quality, dry matter losses, fire risk, or even generating microbes that are harmful to human health (Rentizelas et al., 2009). Therefore, drying is usually applied before storage or other operations to reduce moisture content of biomass to a safe and manageable level. Biomass equilibrates with the surrounding environment’s temperature and relative humidity to eventually reach the equilibrium moisture content (EMC) (Zomorodian et al., 2010). The difference between the instantaneous and equilibrium moisture contents of the biomass represents the potential for moisture desorption (drying) or adsorption. In order to estimate drying rates, it is necessary to investigate the EMC at a range of equilibrium relative humidity and temperature prevalent to conditions in which biomass is dried or stored. Plots of EMC per unit dry mass versus relative humidity at a constant temperature are referred to 31 as the moisture sorption isotherms (Arslan, 2006). An isotherm may be used as a guide to terminate the drying process before the moisture content reaches a specified value in order to save energy. It would also help to estimate the moisture content of biomass after being stored for a period of time under certain conditions. Different types of biomass have different EMC’s under similar environmental conditions because of their physical and chemical characteristics. A large number of theoretical or empirical equations have been developed to model the relationship between EMC and equilibrium relative humidity and temperature for different materials (Van den Berg & Bruin, 1981). Extensive research has been published on agricultural products but not much on forestry residues (Basunia & Abe, 2005; Basunia & Abe, 2001; Lahsasni et al., 2003; Mohamed et al., 2005b; Sun, 1999). Forintek (Forintek, 2004) discussed a commercial drying technique for logs and lumber from Aspen and birch to minimize shrinking and cracking. Drying occurs in several phases for high-moisture materials. During the initial phase, the rate of drying increases due to increase in temperature with some free moisture being removed. In the second phase, free moisture persists on the surface and moisture is evaporated from the saturated surface; the drying rate is high and essentially constant. The third phase corresponds to the falling-rate drying period; the area of the saturated surface gradually decreases as the moisture movement within the solid can no longer supply enough moisture to wet the surface. These three phases might not be distinguishable in some biological materials. Estimates of the drying rate for biomass would provide useful information for the drying industry. There are several methods for modeling the drying process. As for lumber boards, the moisture diffusion shows a good work to model the drying process. This approach can estimate the time-dependent moisture gradients. This model also considers the effects on drying process and moisture gradients, including temperature, relative humidity and air velocity (Cai, 2005; Simpson, 1993). The objectives of this research are to develop a mathematical model to describe the relation between equilibrium moisture content and equilibrium relative humidity of Aspen (Populus tremuloides), and to investigate its drying characteristics. 32 2.2 Materials and methods The biomass used as the materials in this study was obtained from a natural regenerating Trembling Aspen (Populus tremuloides) stand in central Alberta (~40 km north of Plamondon, Alberta, Canada; 54o49’N and 112o19’W). The original stand was harvested in 2004 for the production of hardwood pulp. During April 2011, the immature stems in the stand with an average height of 4.2 m were harvested using a “Bio-Baler” system. The Aspen samples consisted of small-size stems, having bark content around 25%. Samples were put in a cold storage at 4oC before the tests. Pieces of Aspen with their bark intact were then cut to uniform length of 200-250 mm, with diameter varying from 5-10 mm. A picture of materials is shown in Figure 2.1. Figure 2.1. Aspen pieces used in the experimental study A controlled environment chamber (Temperature & Humidity Cabinet, Model LHU- 113, ESPEC Corp., Japan) was used in the experiments (Figure 2.2). Three series of tests were conducted to study the sorption characteristics of Aspen. Materials were equilibrated for 24 h before each test. Test series #1 was for desorption or drying process, whereby temperature ranged from 20 to 70oC. The next two series of tests were for adsorption process. In test series #2, Aspen was placed in the humid chamber at relative humidity of 90% and temperature settings of 25, 35 and 45oC. The last series of experiment (#3) involved testing the materials under relative humidity settings of 60, 70 and 80%, with a constant temperature 33 of 35oC. For all tests, the change in weight of the materials with time was recorded using a digital balance. The initial and final moisture contents of the materials were measured. Moisture content of the sample was determined in triplicate in a forced-air convection oven at 103°C for 24 h to obtain the bone dry biomass according to ASABE Standards S358.2 (ASABE, 2010a). Figure 2.2. Schematic diagram of the controlled environment chamber 2.3 Results and discussion 2.3.1 Moisture sorption characteristics The effects of air temperature and relative humidity on the rate of moisture sorption by Aspen are shown in Figures 2.3-2.4. The drying curves at different temperatures ranging from 20-70oC and corresponding relative humidity are presented in Figure 2.3. Evidently, a considerable time period is required to achieve complete drying and to reach equilibrium moisture content, EMC. With higher ambient temperature and lower relative humidity, the moisture desorption rate increased, as demonstrated by a lower EMC and shorter time to reach the EMC. The EMC of Aspen decreased from 12.1% (that is 0.121 in decimal) to 2.5% dry basis (or, 10.8% to 2.4% wet basis) as the temperature increased from 20 to 70oC. Figure 2.4(a) depicts the effect of temperature on moisture sorption process and the EMC of Aspen, 34 when relative humidity was kept constant at 90%. Moisture adsorption is seen to increase with decreasing temperature. As temperature dropped from 45 to 25oC, Aspen adsorbed additional 3.9% (d.b.) moisture. However, the sorption rate increased with increasing temperature, as the time to reach EMC shortened. This is due to the high activity of water molecules inside the wood, which move faster to the surface at high temperatures. Figure 2.4(b) illustrates the impact of relative humidity on moisture sorption and EMC at a constant temperature of 35oC. EMC was observed to increase with relative humidity of ambient air. The EMC was around 11.7% (d.b.) at 60% relative humidity, and it increased to 17% at 80% relative humidity. This large increase in moisture content with a mild increase in relative humidity shows capillary condensation as described by Yang et al. (Yang et al., 1997). The curve at the higher relative humidity of 80% has a steeper slope at the initial stage, indicating a high adsorption rate. Figure 2.3. Moisture desorption curves for Aspen at different air temperatures and relative humidity 35 Figure 2.4. Moisture adsorption curves for Aspen: (a) under 90% relative humidity and three temperatures; (b) under 35oC temperature and three levels of relative humidity 2.3.2 Moisture sorption isotherms The desorption isotherm is relevant to the drying process and storage of Aspen, while the adsorption isotherm represents the rewetting process during storage. The sorption isotherms can be applied to predict the EMC of materials under certain environmental conditions, which is very useful to industrial processes and systems. A large number of equations have been used to describe EMC with respect to temperature and equilibrium relative humidity (ERH) on different biological materials. According to ASABE Standard D245.6 (ASABE, 2010b), the Modified Henderson, Modified Chung-Pfost, Modified Halsey and Modified Oswin equations are recommended to represent the EMC-ERH relationship for plant-based agricultural products. These four sorption isotherms are empirical equations. These equations were adopted in this study to analyze the EMC and ERH data, as described below. Modified Henderson equation (Henderson, 1952): 1 exp[ ( ) ]BERH A T C EMC= − − + ⋅ (2.1) 36 Modified Chung-Pfost equation (Pfost et al., 1976): exp[ exp( )]AERH B EMC T C = − − ⋅ + (2.2) Modified Halsey equation (Iglesias & Chirife, 1976): exp( )exp[ ]C A B TERH EMC + ⋅ = − (2.3) Modified Oswin equation (Oswin, 1946): 1 1 CA B TERH EMC − + ⋅  = +      (2.4) where ERH is the relative humidity in decimal, EMC is the equilibrium moisture content in decimal (d.b.), T is temperature in oC, and A, B and C are coefficients. The coefficients of the equations were estimated using the non-linear regression module in MATLAB. The performance of each model was evaluated by the error parameter, mean relative deviation (MRD) defined as: 1 1 n P i P EMC EMC MRD n EMC = − =  (2.5) where EMCp is the predicted value of EMC, n is the number of data points, and df is the degree of freedom for the model. Table 2.1 lists the model coefficients and the MRD based on curve fitting to the experimental data. The Modified Chung-Pfost and Modified Oswin equations fitted reasonably well for both adsorption and desorption. Since the Modified Oswin equation has a MRD lower than the Modified Chung-Pfost equation, it was selected as the more appropriate model to describe the moisture sorption relationship for Aspen. 37 Table 2.1. Estimated coefficients and error parameters of four moisture sorption isotherm models fitted to experimental data Model Modified Henderson Modified Chung-Pfost Modified Halsey Modified Oswin Adsorption A 0.1211 72.2 -4.775 0.1211 B 1.161 12.31 -0.01758 -0.00074 C 67.56 3.892 2.18 2.41 MRD 0.000107 0.000246 -4.5E-05 2.13E-05 Desorption A 0.3879 476.4 -0.08227 0.1248 B 2.007 22.33 0.05167 -1.1E-05 C 99.96 25.42 -0.615 2.059 MRD -0.0429 0.0246 -0.06159 -0.0208 The adsorption and desorption curves at 35oC as predicted by the Modified Oswin equation display a sigmoidal shape, as shown in Figure 2.5. There are three stages for the adsorption of water and the state of water molecules within the pore spaces of particles. In the first stage, the water is mostly bound water, and it is adsorbed as monolayer molecules. More water molecules fill the void space. In the last stage, water molecules fill up the pores. Multilayer of water molecules is present within the pores. The desorption curve rides above the adsorption curve, and the desorption and adsorption curves meet at a point where relative humidity equals zero. This demonstrates the hysteresis effect due to adsorption and desorption, and it is comparable to findings pertinent to biological materials (such as rice, barley, grain, leaves and other agricultural and forestry products) by other researchers (Brooker et al., 1974). The average sorption curve lies between the adsorption and desorption curves which can be used for general applications. 38 One theory used to explain hysteresis postulates that during drying process, the water- binding sites are pulled close together with shrinkage, resulting in a reduction of these sites and thus a smaller capacity for attracting water molecules during subsequent adsorption. Therefore, during wetting process the polar sites onto which water is adsorbed are not entirely occupied by moisture (Mohamed et al., 2005a). Moreover, it is known that adsorption and desorption are accompanied by swelling and shrinkage, respectively. Mechanical stresses can cause different equilibrium states from stress-free conditions, acting either in accordance with swelling or to partially prevent it. Figure 2.5. Adsorption and desorption curves as predicted by the Modified Oswin equation Adsorption process is accompanied with a release of heat as the water vapor from the environment is adsorbed into the material; whereas heat is taken up as water in the wood convert to water vapor during desorption process. Some factors have influences on the sorption hysteresis, including temperature and the material properties (Yang et al., 1997). As the temperature increases, both adsorption and desorption isotherms shift downward. The hysteresis effect will be less and less after several cycles of adsorption and desorption tests. 39 These results may be used in a simulation model (Integrated Biomass Supply Analysis and Logistics) (Sokhansanj et al., 2006) for the design of feedstock supply systems for biofuel production. This model can predict the natural drying of forest biomass, when subject to outdoor weather conditions. The moisture adsorption and desorption equations that have been validated are required to calculate the time-dependent moisture contents. 2.3.3 Drying rate of Aspen Equilibrium moisture content (EMC) is a significant factor in biomass drying; it can provide a guideline for expediting drying and for terminating the drying process in a timely manner to save energy. Knowledge of EMC is also necessary for planning the logistics of regional biomass storage. For example, ambient temperature and relative humidity of a region may prove to be associated with high equilibrium moisture and thus the product may not dry in time when left in the field. During the drying process, it is important to know the biomass drying rate, which will exert considerable influence on the process. Aside from ambient temperature and relative humidity, the rate of drying of materials is determined by the velocity of air that flows past its surface, and the heat supply (Lewis, 1921). Figure 2.6 depicts the relation between drying rate and moisture content of Aspen at 40oC. Again, it is generally compatible with the theoretical trend of the three drying stages. The turning point, Mc, is the critical moisture content, whereby drying switches from constant-rate period (dominated by free water) to falling-rate period (dominated by bound water). It depends on several factors which are characteristics of the materials being dried. The constant-rate line and the falling-rate line intersect at the critical moisture content. The constant-rate line was drawn by fitting a horizontal line to the data points with moisture content greater than the upper limit of the fibre saturation point FSP (30% d.b.). As for the falling-rate curve, it was fitted using the data points with moisture content smaller than the lower limit of FSP (25% d.b.). After the critical moisture content is attained, the drying process continues at a decreasing rate, until it reaches the equilibrium moisture content. Similar curves were derived from experimental treatments that involve other temperatures. 40 The drying rate curves agree with the theory of drying. At the onset, moisture is removed from the wood surface; during this stage, sensible heat is transferred from drying air to the moist wood, which enables subsequent processes to take place. The rate of evaporation increases during this period. The next stage is the constant-rate drying period, wherein moisture keeps being removed from the saturated surface. During this period, the rate of evaporation is the highest, and in theory it would change very little as the moisture content is reduced. As illustrated in Figure 2.6, the actual drying rate is seen to fluctuate within this period; it may be attributed to the phase transition of water from the wood surface to the surrounding air. The falling-rate period coincides with the last stage of drying. During this period, bound water migrates from inside to the surface of wood since free water has already completely evaporated. Bound water is more difficult to evaporate due to the hydrogen bonds in wood, which becomes the limiting factor for the drying rate. Drying rate is affected by the wood structure and the moisture gradients within the wood. The reduction in drying rate with time was observed for all experimental treatments. Temperature has a significant effect on the drying rate of Aspen. As temperature was increased from 20 to 70oC, the drying period decreased from 200 hours to 30 hours when EMC was established. Figure 2.6. Drying rate versus moisture content at 40oC. Mc is the critical moisture constant, a transition from constant rate drying to falling rate drying 41 Several mathematical models have been proposed to simulate water movement during drying. Page’s model was specifically developed to determine the drying characteristics of agricultural crops (ASABE, 2006; Phanphanich & Mani, 2009) as shown in Eq. (2.6), ( ) 0 exp ne e M M kt M M − = − − (2.6) where M is the instantaneous moisture content (decimal, d.b.), M0 is the initial moisture content (decimal, d.b.), Me is the equilibrium moisture content (decimal, d.b.), k is the drying rate constant, n is constant, and t is time (h). For materials stored in the field, the drying rate constant is related to solar radiation, temperature, wind speed and biomass density. The rate of moisture movement from the interior of the materials is proportional to the difference between the instantaneous moisture content and the equilibrium moisture content, in units of concentration. The experimental data were analyzed for estimating Page’s model parameters n and k at different temperatures and relative humidity using the non-linear regression module in MATLAB, and the results are listed in Table 2.2. It can be seen that n and k varied from (0.94 to 1.26), and (0.016 to 0.081 h-1), respectively, while temperature ranges from 20-70oC. By comparison, n and k values varied from (1.10 to 1.37), and (0.28 to 0.62 h-1), respectively, with temperature ranging from 40-80oC in a study of the drying characteristics of pine forest residues by Phanphanich and Mani (Phanphanich & Mani, 2009). We also compare the values of these constants with those reported for agricultural materials such as grass (3.37 h- 1), wheat (1.02h-1) and lentils (0.2 h-1) (ASABE, 2006). The much faster drying rates exhibited by the pine forest residues versus aspen stems could be attributed to the much smaller sizes of particles including ground leaves and needles (Phanphanich & Mani, 2009). Since the experimental n values do not have a high variability, we can define a uniform exponent n as the average of all n values at different temperatures. Thus, Eq. (2.6) may be expressed as: ( )1.109 0 expe e M M kt M M − = − − (2.7) 42 Temperature is a significant factor that affects the drying rate constant. Evidently, the drying rate constant of aspen increases with an increase in the drying temperature. This relationship was shown to resemble the Arrhenius equation which is generally applied in the analysis of chemical reaction rates: 0 exp b Ek k RT   = −   (2.8) where k0 is a constant calculated from the plot at the intercept (1/T = 0); Eb is the activation energy in J/mol, which was estimated to be 34.9 kJ/mol. Table 2.2. Parameters of Page’s equation (k and n) under different temperatures Temperature (oC) k (h-1) n R2 20 0.0160 0.943 0.9996 30 0.0355 1.069 0.9995 40 0.0286 1.260 0.9997 50 0.0390 1.202 0.9993 60 0.0785 1.085 0.9997 70 0.0807 1.093 0.9996 43 2.4 Conclusion The effects of temperature and relative humidity on moisture sorption of Aspen were studied. Results showed that low temperature and high relative humidity led to higher equilibrium moisture content for both adsorption and desorption processes at the end of the drying process. Higher relative humidity promotes the adsorption process; it led to a higher EMC under the same temperature. Under the same relative humidity, higher temperature resulted in greater sorption rates and a lower EMC. Four sorption isotherm models were fit to the experimental data in order to predict the drying process of Aspen during storage. The Modified Oswin equation was found to provide the best fit for both desorption and adsorption processes pertinent to Aspen. The adsorption and desorption curves displayed a sigmoidal shape similar to the characteristics of other biological materials, and the curves exhibited hysteresis effect between adsorption and desorption. The drying rates of Aspen obtained in this study generally agree with the theory of drying for wood. Furthermore, the drying rate of Aspen was analyzed by applying Page’s model. Results indicated that the trend of drying rate constant as a function of temperature followed the Arrhenius equation, and Page’s model is appropriate for predicting the drying characteristics of Aspen. 44 Chapter 3. Modelling the drying and wetting processes of Aspen (Populus tremuloides) 3.1 Introduction Lignocellulosic biomass may be used as solid or liquid biofuels. Usually, the immediate use of these materials after harvest is infeasible. The logging operations produce large quantities of high moisture biomass that are left behind in the field. These materials will undergo a series of operations before they can be used in the form of chips or densified forms such as pellets, the latter primarily for long distance transport and ease of handling. There is an abundant supply of softwood forest biomass such as spruce, pine and fir in British Columbia. However, in some parts of North America, hardwood forest biomass is common. Aspen, a hardwood species that grows in northern climates, can be a readily available woody biomass source for chipping and making fuel pellets. Biomass may be exposed to elements of weather conditions during storage in the field for up to one year prior to its removal. Depending on the form of storage, the biomass can adsorb or desorb moisture continually from its immediate environment. This could in turn lead to deterioration in the quality such as calorific value of the materials (Pettersson & Nordfjell, 2007). Besides, high moisture content increases the cost of transport and pelletizing, which will be reflected in the pricing of fuel (Rentizelas et al., 2009). Temperature, relative humidity and precipitation are the major factors that would affect moisture content. The drying process of materials is affected by several variables and driven by the differences of water vapour pressure between the materials and surrounding air (Savoie & Mailhot, 1986); while precipitation (rain or snow) and condensed water are adsorbed by the materials resulting in an increase in moisture content. Eventually, the material equilibrates with the surrounding environment’s temperature and relative humidity to reach the equilibrium moisture content (Zomorodian et al., 2010). Investigation of moisture sorption of woody biomass under different storage and natural drying conditions have been reported in the published literature (Afzal et al., 2010; Casal et al., 2010; Eriksson & Gustavsson, 2010; Gislerud, 1990; Jirjis, 2005). Lots of 45 previous studies have been done on the sorption isotherms of biomass and the effects of temperature, for instance, those reported by (Basunia & Abe, 2005; Lahsasni et al., 2003; Mohamed et al., 2005b; Sun, 1999). However, few studies have applied modeling to simulate the drying and wetting processes due to precipitation (Johansson & Salin, 2011; Nilsson, 1999; Nilsson & Karlsson, 2005; Satin, 2011; Tonn et al., 2011). In this regard, Stewart and Lievers (Stewart & Lievers, 1978) developed a model from experimental data for the drying and rewetting processes of wheat straw in the field. Subsequently, a modified model was applied and validated for the field drying of wheat straw by Nilsson (Nilsson, 1999) and cut flax by Nilsson and Karlsson (Nilsson & Karlsson, 2005). Models can be used to predict the time-dependent moisture sorption of woody biomass during natural drying, and hence help to seek a better way to store and manage these high-moisture materials. The objective of this study is to adopt a model from literature, and calibrate the model for its future application to simulate the wetting and drying of Aspen materials under natural drying conditions. 3.2 Model description The drying process involves moisture evaporation from the surface of wood and moisture diffusion from the interior of wood to the surface. It occurs in three phases. During the initial phase, sensible heat is transferred to the materials and the contained moisture; the rate of drying increases due to increase in temperature with some free external moisture being removed. In the second phase, free moisture on the saturated surfaces is removed by evaporation; the rate of drying is high and essentially constant. The third phase is characterized by falling-rate drying; the area of the saturated surface gradually decreased as the moisture movement within the solid can no longer supply enough moisture to wet the surface. Migration of bound water from the inner to the outer surface takes place since free water is already completely evaporated. Bound water is more difficult to evaporate due to the hydrogen bonds in wood, which leads to the reduction in the drying rate. The falling-rate period expresses the movement of bound water, which can be used to estimate the internal moisture sorption. 46 Biomass adsorbs or desorbs moisture continually from its immediate surrounding environment during storage in the field. Equations are developed to represent moisture relations for biomass. Biomass that exchanges moisture with its surrounding environment is divided into two parts: an external part that exchanges moisture with the surroundings and an internal part within which moisture transfer occurs through diffusion. The average moisture content for the bulk biomass is estimated as the sum of these two types of moistures, thus i s mass of internal water mass of external waterM M M total dry mass total dry mass = + = + (3.1) where M is the overall moisture content, Mi is the internal moisture content, and Ms is the external moisture content, all of which are expressed with reference to the total mass of dry matter (that is, dry mass basis). Internal moisture is the bound water of biomass originating from the uptake of water by the roots. External moisture is primarily the water that originates from precipitation and dew; it also includes water that has moved to the surface. The above formulation was first suggested by Stewart and Lievers (Stewart & Lievers, 1978) for the drying and wetting processes of wheat straw in the field. Subsequently, a modified model was applied and validated for the field drying of wheat straw by Nilsson (Nilsson, 1999) and cut flax by Nilsson and Karlsson (Nilsson & Karlsson, 2005). The change in the internal moisture content depends on the difference between the instantaneous moisture content M and the equilibrium moisture content of biomass Me. Several mathematical models have been proposed to simulate the movement of bound water during drying. These models integrate the factors of temperature, relative humidity and airflow, which can estimate the time-dependent moisture gradients (Cai, 2005; Simpson, 1993). In this study, the semi-empirical drying equation developed by Lewis (Lewis, 1921) was used to describe the change in internal moisture content. Lewis model is a simplified version of Page’s model. It is assumed that the moisture content in the biomass is uniform, and the water can be dried without resistance. Thus, the time dependency of the internal moisture content is represented by the following first-order drying equation: ( )1i p i edM a E M Mdt = − − (3.2) 47 where a1 is a coefficient (mm-1), Ep is the pan or potential evaporation rate (that is, evaporation rate of water from a free surface or open water) (mm/hr) and Me is equilibrium moisture content (decimal, dry basis). In Lewis’ model, a1Ep is presented as k, which is defined as the drying rate constant. Hence, the drying rate constant k is a combination of Ep, which is related to the surrounding environment and a1, which is a property of the material. The dimensions of the materials, and specifically, the surface area-to-volume ratio, will affect the value of a1. The equilibrium moisture content Me is also a physical characteristics of the material, under specified ambient temperature and relative humidity conditions. For the external moisture Ms, the following mass balance that relates the rate of moisture change to precipitation and evaporation rate may apply (Nilsson, 1999): (3.3) where P is the precipitation rate (mm/hr), a2 and a3 are constants (mm-1). The term a2P accounts for the adsorbed precipitation by the materials, whereas the term a3Ep represents the actual evaporation (as a fraction of potential or maximum evaporation). A large number of theoretical or empirical equations have been developed to model the relationship between equilibrium moisture content and relative humidity at a certain temperature for different materials. Extensive research findings have been published for agricultural products, but not much are available for forestry residues (Basunia & Abe, 2005; Lahsasni et al., 2003; Mohamed et al., 2005b; Sun, 1999). Based on our previous research work in Chapter 2, the modified-Oswin equation was found to provide the best fit for the sorption processes pertinent to Aspen (Populus tremuloides). Therefore, we applied this equation to calculate the equilibrium moisture content Me, ( )1/1/ 1 CeM A B T rh= + ⋅ − (3.4) where rh is the relative humidity in decimal, Me is the equilibrium moisture content in decimal (dry basis), T is ambient temperature in oC, and A, B and C are coefficients. p32 s EaPa dt dM −= 48 3.3 Materials and methods 3.3.1 Materials The Canadian Wood Fibre Centre and the Alberta-Pacific Forest Industries set up an immature regenerating trembling aspen (Populus tremuloides) stands to enhance the available volume of biomass at its disposal. The site is 24.8 ha in size, and is located in central Alberta (~ 40 km north of Plamondon, Alberta, Canada; 55o2’N and 111o57’W). After seven years of natural regeneration, the stand now consists of trembling aspen, balsam poplar and white birch stems. All the biomass used as raw materials in this study was obtained from the natural regenerating trembling aspen stand. Harvesting was done in April 2011 using a system that harvests and bundles the biomass in one process. The Aspen materials were then delivered to the University of British Columbia, Vancouver. For the experiment, the Aspen samples consisted of small-size stems, having a bark content around 25%. Pieces of Aspen with their bark intact were then cut to uniform length of 200-250 mm, with diameter varying from 5-10 mm (Figure 2.1). 3.3.2 Experiment In this study, experiments were conducted under controlled environment in the lab to obtain the coefficients of the model. The coefficients a1, a2 and a3 in Eqs (3.2) and (3.3) were determined by applying regression analysis (curve-fitting) to the data collected. A controlled environment chamber (Temperature & Humidity Cabinet, Model LHU-113, ESPEC Corp., Japan) was used for the experiments. The materials were stored in a cold room at 4oC, and equilibrated for 24 h before each test. Two series of tests were conducted to study the sorption characteristics of Aspen. Test series #1 was for desorption or drying process, with temperature ranging from 20 to 70oC. These tests were operated without simulated precipitation, which is relevant to the coefficient a1 for the internal moisture (Eq 2). Test series #2 was conducted to study the effect of simulated precipitation on the change of moisture with time under two temperatures, 20oC and 30oC, with the primary aim to determine the coefficients a2 and a3 in Eq 3. In the 49 beginning of Test series #2, no water was sprayed on the material, and the material would be expected to dry to a sufficiently low point close to the EMC via evaporation only; this would enable the coefficient a3 to be determined. Subsequently, water was added to simulate different amounts of rainfall; this was implemented to study the simultaneous effects of evaporation and precipitation on external moisture content and thus estimate the coefficient a2. The drying curves derived from Test series #2 can also be used to determine the coefficient a1. Natural precipitation was simulated by using a sprayer to deliver various amounts of tap water to the materials, according to the average precipitation in Vancouver, BC, Canada (Environment Canada). The amounts of 25 ml, 50 ml, 75 ml and 100 ml delivered correspond to 0.4 mm, 0.8 mm, 1.2 mm and 1.6 mm per hour, respectively. Samples with mass of 330 g and moisture content around 40% (d.b.) were placed on a 310 x 200 mm screen in single (thin) layer. The mesh with the materials was then placed in a plastic box to allow for any water that percolates through the materials to be collected as leachate. Replicate runs were performed for each test. For all tests, the change in mass of the materials with time was recorded using a digital balance. The initial and final moisture contents of the materials were measured in triplicate in a forced-air convection oven at 103°C for 24 h according to ASABE Standards S358.2 (ASABE, 2010a). The evaporation rate of water under different temperatures was simultaneously measured from a pan during the experiment. The coefficients of the equations were estimated using the non-linear regression module in MATLAB (The MathWorks Inc., MA, USA, 2011). 3.4 Results and discussion Parameter estimation was performed for the coefficients of the simulation model, which are pertinent to internal moisture content and external moisture content. 3.4.1 Internal moisture content The change in internal moisture content within the materials is due to diffusion from inside to the outer surface. Drying tests without precipitation were set up to obtain the coefficient a1 50 in the internal moisture equation (Eq 3.2). The drying curves of Aspen at different temperatures ranging from 20-70oC and the corresponding relative humidity are shown in Figure 2.1. A relatively long drying period was required to enable the development of a complete drying curve, including the EMC (equilibrium moisture content) point. At higher temperatures, the moisture desorption rate increased, thus the EMC was lower and it took a shorter time to reach the EMC. The change in the internal moisture content (bound water) corresponds to the falling-rate period of drying. Bound water is more difficult to evaporate due to the hydrogen bonds in wood, which becomes the limiting factor for the drying rate. The reduction in drying rate with time was observed; evidently, temperature had a significant effect on the drying rate of Aspen. A correlation between measured evaporation rate and temperature (over the range 20-70oC) was developed, and it may be expressed as 5 29 10 0.0192 0.282pE T T − = − × ⋅ + − (3.5) where Ep is evaporation rate (mm/h) and T is temperature (oC). Thus, in Eq (3.2), Ep varies with temperature and a1 is a constant. The drying rate constant (k =a1Ep) was determined by linear regression (MATLAB); thus the coefficient a1 was calculated as k/Ep and determined from the drying curves to be 0.208 mm-1 on average. 3.4.2 External moisture content When there is no precipitation, the equation that represents external moisture content becomes: 3 s p dM a E dt = − (3.6) The drying rate for external moisture without precipitation is constant; this corresponds to the constant-rate period of the drying process. During this period, the rate of evaporation is highest and constant. Moisture (free water) keeps being removed from the saturated surface. 51 Figures 3.1 and 3.2 depict the changes in total moisture content, M, when subject to the effects of precipitation and evaporation at 20oC and 30oC, respectively. The first drying curve in both graphs represents the drying of the materials from its original moisture content in the absence of precipitation (no water sprayed on the materials). After the samples were dried to near EMC, water was added to simulate different amounts of rainfall. The coefficient a3 was then determined during the constant-rate period according to Eq (3.6). For the drying curves after each simulated rainfall or wetting event, the moisture content had a sharper decrease at the beginning; it was then gradually reduced until the EMC was attained. This may be attributed to the evaporation of the free water on the surface of the materials (originating from precipitation), which preceded the evaporation of free water within the cell capillary and then the bound water in the cell wall. Figure 3.1. Change in total moisture content with time when biomass is wetted and dried repeatedly at 20oC. This graph emulates frequent wetting of biomass due to rain 52 Figure 3.2. Change in total moisture content with time when biomass is wetted and dried repeatedly at 30oC Figure 3.3 describes the relation between the drying rate and the instantaneous moisture content at 20oC temperature. The data were derived from all of the 10 cycles in Figure 3.1. The critical moisture content is defined by the intersection of the two straight lines (y1 = C; y2 = b1x + b2), at which drying switches from the constant-rate period (dominated by free water) to the falling-rate period (dominated by bound water). It depends on the physical and chemical characteristics of the material being dried. After the critical moisture content is attained, the drying process continues at a decreasing rate, until it reaches the equilibrium moisture content. In Figure 3.3, a3 can be estimated from the line (y1 = C), which equals C/Ep. A similar relationship was observed for the experiment conducted at 30oC temperature. 53 Figure 3.3. Drying rate versus moisture content at 20oC When there is precipitation, no evaporation was assumed to occur. The equation that represents external moisture content would then take the following form, 2 sdM a P dt = (3.7) where P is the precipitation rate. The amount of water added had an impact on the subsequent initial moisture contents. All of the increased moisture contributed to the change in external moisture content. It is evident from Figures 3.1 and 3.2 that the initial moisture content increased substantially when the amount of sprayed water on the wood was raised from 0.4 mm to 1.2 mm per hour. However, further increase in the amount of simulated rainfall from 1.2 mm to 1.6 mm per hour did not induce much greater change in the initial moisture content, as water percolated through the thin-layer materials as leachate. Since the leachate collected at the bottom of the box was not in direct contact with the materials, it is assumed to have negligible contribution to moisture adsorption by biomass. The amount of water held on the surface of the materials (620 cm2 surface area) was estimated to be 90 ml; simulated rainfall over this adsorption capacity would generate leachate. The changes in external moisture contents were used to calculate the coefficient a2 based on Eq (3.7). 54 In order to estimate the coefficients a2 and a3 by MATLAB, a restriction was placed on the external moisture content ( 0sM ≥ ). The coefficient a1 may also be estimated from the falling-rate period of the drying curves in Figure 3.1 and Figure 3.2, whereby the measured Ep which follows the empirical relation (Eq 3.5) were used in the calculations. The ranges of estimated values for the three coefficients are illustrated in Figure 3.4. They are all seen to fluctuate by no more than 25%, 16% and 13% for a1, a2 and a3 around their respective means; no obvious trends were observed after several cycles of re-wetting. The coefficient a1 varies from 0.15 to 0.24 mm-1, while a2 and a3 ranges from 0.11-0.15 mm-1 and 0.21-0.26 mm-1, respectively. According to the data presented in Figure 3.4, the average values of a2 and a3 are 0.129 mm-1 and 0.239 mm-1, respectively. The overall average value of a1 is 0.206 mm-1 based on the two series of tests, with an error parameter (standard error of estimation SEE) of 0.026, whereas the SEE for a2 and a3 are 0.01 and 0.024. The small SEE values demonstrate that the model is well-calibrated with these coefficients. Figure 3.4. Variation of the three coefficients with the number of simulated precipitation (wetting) events at 20oC and 30oC (a1: ∆; a2: ▪; a3: ●) 55 In Nilsson’s study on wheat straw, the three coefficients a1, a2 and a3 were estimated to be 1.2 mm-1, 0.23 mm-1 and 0.18 mm-1, respectively. The large difference in the coefficient a1 between our study and Nilsson’s implies it is more difficult to evaporate the internal moisture of Aspen than wheat straw under the same environmental conditions. That straw could be readily dried to an equilibrium moisture content as low as 0% (Stewart & Lievers, 1978) verified its higher evaporation capability versus Aspen. The higher value of a2 for wheat straw suggests that it is more sensitive to precipitation; as a result wheat straw could gain more moisture than Aspen when exposed to the same amount of precipitation. It shall be noted that the coefficients (a1, a2 and a3) derived from the experiment are relevant to the Aspen stem materials with diameter 5-10 mm and length 200-250 mm. If materials with different dimensions are used, the surface area-to-volume ratio will change and the values of these coefficients will be different. The model is being applied to describe the moisture relations for covered and uncovered Aspen materials during long-term storage under natural weather conditions. 3.5 Conclusion In this chapter, a mathematical model was adopted to simulate the wetting and drying of lignocellulosic biomass. The moisture of Aspen is divided into two parts, internal moisture content (bound water) and external moisture content (free water), which exchanges with the surrounding environment by evaporation and precipitation. The time-dependent internal moisture content was represented by the Lewis equation. The change of external moisture content is modeled on the basis of evaporation versus precipitation. Aspen (Populus tremuloides) stems were used in the experiment for calibrating the model. The three coefficients of the model were estimated to be 0.206 mm-1, 0.129 mm-1 and 0.239 mm-1, with small standard errors of estimation. 56 Chapter 4. Model application for moisture variation of Aspen bales 4.1 Introduction Aspen is a readily available hardwood biomass source for chipping and making standard grade or premium grade fuel pellets depending on its ash content. After harvest, large quantities of high moisture biomass are piled or baled, and then left in the field for storage. These materials will be exposed to various weather conditions during storage for up to one year prior to transportation, and they will adsorb or desorb moisture continually from the surrounding environment. The variation of moisture content in Aspen exerts an influence on its quality such as calorific value upon utilization (Pettersson & Nordfjell, 2007). The moisture of the materials will vary when subject to natural weather conditions during storage in the field. In the event of precipitation, the moisture content of the materials will increase, whereas the materials will lose moisture during dry season with high temperature and low relative humidity. When the moisture in the material is in balance with the surrounding atmosphere, it reaches the equilibrium moisture content (Silakul & Jindal, 2002). Since the raw materials are usually dried to a certain degree before processing and utilization, it would be beneficial to estimate the moisture content of the materials before drying. Being able to predict the moisture content can help to better understand and manage the storage process in the field in order to minimize the losses in calorific value. This information can also be utilized to design a cost-effective natural drying process, thus minimizing the energy consumption. The model for simulating the moisture sorption during storage has been described in Chapter 3, section 2. The time-dependent moisture content of Aspen is divided into internal and external moisture contents, which represent the bound water and free water associated with the wood, respectively. The materials exchange moisture with the surrounding environment through evaporation and precipitation. The internal moisture equation originates from the Lewis’ equation, while the change in external moisture content is related to the 57 differences between evaporation and precipitation. Several research studies have been done on the drying and wetting processes of wheat straw and cut flax in the field (Nilsson, 1999; Nilsson & Karlsson, 2005; Stewart & Lievers, 1978). However, previous researches have not studied the processes of drying and wetting Aspen (Populus tremuloides) upon exposure to the natural weather conditions when these materials are left in the field after harvest, and before delivery to the plant. The objectives of this Chapter are to apply a lumped model for simulating the moisture variations of the Aspen materials, with available weather data as inputs, and to verify the model with data collected from field tests. 4.2 Model description The model is comprised of a series of equations that describe various processes involved in drying and wetting. Some of the equations presented in this section have been shown in Chapters 2 and 3. 4.2.1 Moisture content As described in Chapter 3, section 2, the moisture content of biomass is estimated as the sum of internal and external moistures, i s mass of internal water mass of external waterM M M total dry mass total dry mass = + = + (3.1) where M is the overall moisture content, Mi is the internal moisture content, and Ms is the external moisture content, all of which are expressed with reference to the total mass of dry matter (that is, dry mass basis). The internal moisture content and external moisture content are represented by the following equations (Lewis, 1921; Nilsson, 1999), ( )1i p i edM a E M Mdt = − − (3.2) 58 (3.3) where a1 is a coefficient (mm-1), Ep is the potential evaporation rate (that is, evaporation rate of water from a free surface or open water) (mm/d) and Me is equilibrium moisture content (dry basis, decimal). P is precipitation rate (mm/d), a2 and a3 are constants (mm-1). During simulation, Ms is restricted to be ≥ 0. When the moisture content of the Aspen matertials are low and the surrouding environment is humid, they adsorb water from the air. According to Chapter 2 section 3.1, The adsorption process of the mateirals is expressed by a exponential model ( )i a e i dM k M M dt = − (4.1) where ka is the adsorption rate constant, which is related to temperature and relative humidity. Its value was determined by fitting a regression equation to the experimental data as follows : 0.87 0.017 0.92 0.009122ak T rh T rh= − + + − ⋅ ⋅ (4.2) where T is ambient temperature (oC), rh is relative humidity (%). It is assumed that the lumped model describes the moisture content of Aspen bales during prolonged storage (months). Moisture is assumed the same everywhere within the bale. Hence, the moisture content estimated by this model represents the average value within the bales. 4.2.2 Equilibrium moisture content Based on Chapter 2, section 3, the Modified-Oswin equation was found to provide the best fit for the sorption processes pertinent to Aspen (Populus tremuloides). Here, this equation was applied to calculate the equilibrium moisture content Me, ( )1/1/ 1 CeM A B T rh= + ⋅ − (3.4) p32 s EaPa dt dM −= 59 where rh is the relative humidity in decimal, Me is the equilibrium moisture content in decimal (dry basis), T is ambient temperature in oC, and A, B and C are coefficients. 4.2.3 Evaporation rate In Eq (3.2) and (3.3), the evaporation rate is estimated by the modified Penman equation as proposed by Shuttleworth (Penman, 1948; Shuttleworth & Evaporation, 1993): ( ) ( ) 6.43 1 0.536n e p R u E γ δ λ γ Δ ⋅ + ⋅ + ⋅ = Δ + (4.3) where Ep is evaporation rate (mm/d), Rn is net radiation (MJ/m2.d), u is wind speed (m/s), δe is vapor pressure deficit (kPa), ∆ is the slope of the saturation vapor pressure curve (kPa/oC), γ is the psychrometric coefficient (kPa/oC), and λ is the latent heat of vaporization (MJ/kg). Net radiation comprises net short-wave (solar) radiation and net long-wave radiation. Net solar radiation is that portion of the incident solar radiation captured by the biomass taking into account losses due to reflection. The albedo of the Aspen materials was assumed to be 0.18; this value is similar for deciduous trees (Barry & Chorley, 2009). Psychrometric relations (ASABE., 2010) were used to calculate the saturated vapor pressure, as well as the latent heat of vaporization. Some of the parameters in Eq (4.3) are calculated by the procedure outlined by Allen et al (1998). ∆ and γ were calculated by: ( )2 17.272504exp 237.3 237.3 T T T   + Δ = + (4.4) 2.501 0.002361Tλ = − (4.5) 0.00163 atmPγ λ= (4.6) where T is air temperature (oC), and Patm is atmospheric pressure (kPa). Vapor pressure deficit is defined by the following expression, 60 e s ae eδ = − (4.7) where, es is saturation vapor pressure (kPa) and ea is actual vapor pressure of ambient air (kPa) at the given temperatures. Net radiation is the balance of the incoming net shortwave radiation (Rns) and the outgoing net longwave radiation (Rnl). Net radiation is generally positive during daytime and negative during nighttime. It provides the energy to drive the evaporation process. n ns nlR R R= − (4.8) As a component of net radiation, the net shortwave radiation is expressed as: ( )1ns sR Rα= − (4.9) where α is albedo or reflection coefficient [dimensionless], Rs is the incoming solar radiation (MJ/m2.d). And the net longwave radiation is given by: ( )4 4max, min, 0.34 0.14 1.35 0.352K K snl a so T T RR e R σ  +   = − −      (4.10) where σ is the Stefan-Boltzmann constant [4.903×10-9 MJ K-4 m-2 day-1], Tmax,K is the maximum temperature during the day [K], Tmin,K is the minimum temperature during the day [K] and Rso is the clear sky radiation (MJ/m2.d). Rs/Rso is relative shortwave radiation, which has a value less than 1.0. 4.2.4 Parameter estimation Lab-scale experiments were conducted, as described in Chapter 3, section 4 to determine the coefficients a1, a2, a3 in Eq (3.2) and (3.3) of the model. This was done for model calibration purposes. The Aspen materials were subject to periods of artificial precipitation and evaporation under constant temperature in a controlled environment chamber. Time- 61 dependent moisture data along with pan evaporation rates were collected and used to estimate the parameters of the equations. The coefficients a1, a2, a3 were estimated to be 0.206 mm-1, 0.129 mm-1 and 0.239 mm-1, respectively, with small standard errors of estimation. 4.3 Materials and methods 4.3.1 Model structure and simulation procedure A flow chart of the model for estimating the time-dependent moisture content of the materials is shown in Figure 4.1. The time step (Δt) in the model is 1 day. When the running cycles reach the total simulation time n, the procedure ends. The model is implemented and executed using MATLAB (MathWorks Inc., MA, USA 2011). The moisture content of bales is measured at the beginning of the storage period and that provides the input as the initial value in the simulation. After starting the program (t = 0), the first step is to calculate the equilibrium moisture content (EMC) using temperature and relative humidity. The EMC is then compared to the instantaneous total moisture content to decide the sequent sorption process. If EMC is larger than bale moisture, the simulation would proceed to the adsorption equation; otherwise, desorption process would take over. When it is the desorption process, the EMC is further compared to internal moisture content. If internal moisture content is larger than EMC, both internal and external moisture contents would be recalculated by Eqs (3.2) and (3.3). Otherwise, the external moisture content would be calculated by Eq (3.3) while the internal moisture content stays the same. A restriction is applied such that the external moisture content could not be less than 0. The total moisture content is the sum of internal and external moisture contents. After the new total moisture content is obtained at time t+Δt, the program will proceed to the next cycle until time t = n. 62 Figure 4.1. Flow chart of the model to calculate the moisture content of Aspen materials at time t 63 4.3.2 Field Test The Canadian Wood Fibre Centre and the Alberta-Pacific Forest Industries set up an immature regenerating trembling aspen stands to enhance the available volume of biomass at its disposal. The site is 24.8 ha in size, and is located in central Alberta (~ 40 km north of Plamondon, Alberta, Canada; 55o 2’ N and 111o 57’ W). After seven years of natural regeneration, the stand now consists of trembling aspen, balsam poplar and white birch stems. All the biomass used as raw materials in this study was obtained from the natural regenerating trembling aspen stand. Harvesting was done in April 2011 using a system that harvests and bundles the biomass in one process. Four Aspen bales received from a natural regenerating trembling aspen stand in central Alberta were used in the field experiment (Figure 4.2). The bale is cylindrical in shape, with dimensions 1.24 x 1.24 m (length and diameter), giving a volume of 1.4 m3 and bulk density of 190 kg/m3. Data was collected for one year from June 2011 to May 2012. Two bales (#1 and #2) were stored in an open field as replicates; they were uncovered and exposed to outdoor weather conditions, at the Vancouver campus of the University of British Columbia (UBC). Bale #1 was oriented East-West, while bale #2 was placed in North-South orientation. The other two bales (#3 and #4) were placed under cover (3.6 m x 3.6 m tent), as replicates, and were protected from direct solar radiation and precipitation. The side walls of the tent are made of mesh which can prevent passage of insects and rain drops but help the air circulation. Bale #3 and #4 were also placed as East-West orientation and North-South orientation, respectively. Each bale was placed on a pallet in order to prevent the adsorption of soil moisture and become muddy. Once a month, approximately 100 g of sample was collected from a number of spots in each bale. The samples were mixed and reduced to smaller pieces for moisture content determination. The moisture content thus obtained is an average value, and it was used to verify the model. Temperature of the biomass was measured using several thermocouples inserted in various places within the bale, and automatically logged (Appendix C). A weather station was located in the Department of Earth and Ocean Sciences at UBC, 200 m from the test site. 64 The following weather data were recorded daily: temperature, precipitation, wind speed, solar radiation and relative humidity. The data thus collected from the field test were used to verify the model. Subsequently, model application was extended to predict the moisture content of a bale with transparent cover, with an aim to assess the effect of protection from precipitation while allowing solar radiation to penetrate. Hence, solar radiation would affect the evaporation rate of the materials. Nowadays, transparent water-proof covers are available for sheltering piles of logging residues from rain, snow and ice in the field. The sides of the pile can still remain open in order to allow moisture to evaporate. Figure 4.2. Stored Aspen bales at UBC (uncovered: bales #1 and #2; covered: bales #3 and #4) 4.4 Results and Discussion 4.4.1 Model application Field experimental data and the data collected from the weather station were used for model validation purposes. The weather data (temperature, relative humidity and precipitation) 65 during the storage period (June 2011-May 2012) are shown in Figures 4.3-4.5. The climate may be described as warm and dry in summer time (June-September 2011) and then (April- May 2012), as compared to the cool and wet climate conditions (November 2011-March 2012). Temperature was high around 20oC in late August and September and occasionally reaching 24oC. Temperatures averaged around 17oC in the summer, and gradually decreased to average 5oC during late fall to early spring. On some days, the temperature dropped to below zero. Relative humidity was seen to fluctuate around 70% during summer and around 85% in the winter. At times, the relative humidity reached almost 95% amidst frequent precipitation during November-March. During this period, the average daily precipitation was 12 mm. Considerable precipitation over 15 mm was observed for a few days. In some days, the daily precipitation reached as high as 40 mm. Precipitation was also quite frequent in April; however, the daily amount was smaller than those in the wet season. By comparison, it is obvious that there was little precipitation during the period June-September; in particular, precipitation was observed only once in August, and the climate in May was dry relative to April. The calculated values of net radiation (using Eqs 4.8-4.10) are displayed in Figure 4.6. It is the radiation captured by the biomass, which is partially related to evaporation. The value of net radiation is seen to range from -2 to 18 MJ/m2.d. The net radiation was as low as zero and sometimes negative in the November-March period, as compared to a high average value of 10 MJ/m2.d during the April-September period. This is compatible with the sunny summer and rainy winter. The daily evaporation rates for both the covered and uncovered bales were calculated based on Eq (4.3) and results are presented in Figures 4.7 and 4.8, respectively. The computed evaporation rates of 2-5 mm/d associated with the uncovered bales were high from June to September when solar radiation was revealed as the dominant driving force (Figure 4.7). Evaporation rates had a pronounced decrease from late fall through winter to early spring (0.2-2 mm/d) when vapor pressure deficit and wind speed became the main driving forces. Again, the reverse trend was observed as of March. By comparison, from Figure 4.8, the computed evaporation rates of the covered bales were significantly lower, ranging from 0-2 mm/d for the whole year; the evaporation rates were higher in summer and spring and 66 lower in winter, which exhibits a similar trend as the uncovered bales. As the evaporation rates of the covered bales were unaffected by solar radiation, again, vapor pressure deficit and wind constituted the two major driving forces. With the daily weather data as inputs, the simulation then proceeded to predict the moisture contents of the baled Aspen materials during storage by applying Eqs (3.1) to (3.3) and (4.1), and implementing a time-step of one day. Results are shown in Figures 4.9 and 4.10, along with the moisture contents that were measured once a month. For initial conditions, Mi and Ms values were assumed to be 90% and 10% of the initial total moisture content, M, respectively. This assumption was based on the fiber saturation point of woody biomass (Perré, 2007) and the measured initial moisture content of the Aspen bales. Samples were taken from different locations within the bales. For the covered bales (#3 and #4), results indicated small variations, for instance, 16-19% (wet basis) or 19-23% (dry basis) during the wet season. However, for the uncovered bales (#1 and #2), larger variations were observed; for instance, the moisture content ranged from 29-42% (wet basis) or 41-70% (dry basis). For the uncovered bales, the measured moisture contents of materials were low, around 10% dry basis in the beginning. It increased gradually to around 20% dry basis in August. The moisture content had a sharp increase to 90% dry basis in November due to the frequent precipitations. Subsequently, the moisture content fluctuated around 70% dry basis from January to April. The measured moisture contents of the uncovered bales #1 (E-W orientation) and #2 (N-S orientation) showed similar values when moisture content was lower than 40% dry basis. During wet and cold season, the moisture content increased and the difference in moisture content between two bales increased. However, the trends of moisture content for both bales were similar. There is no obvious correlation observed between bale orientation and moisture content. The predicted high moisture contents of the Aspen materials across the winter months were compatible with the intensive rain and low evaporation rate during that period (Figure 4.9). It can be seen that the trend of measured moisture contents follows the trend of calculated moisture contents reasonably well, despite some larger deviations of the measured versus predicted values. Overall, the percent 67 differences between the predicted and actual values for bales #1 and #2 are 18.6% and 30.4%, respectively. As for the covered bales, the predicted and measured moisture contents are illustrated in Figure 4.10. The measured moisture contents had a smaller range compared to the moisture contents of the bales without cover. From June to September, the moisture contents were low, around 15% dry basis. The moisture contents increased to 25% dry basis in the winter time, and dropped back to 17% dry basis on average in spring time. The trends of the moisture content for both bales (with different orientations) were similar during the one-year storage period. In general, there was no big difference between the moisture contents of the two bales, except for July 2011 and February 2012. The predicted moisture content increased as of October, and then slightly decreased as of March. This pattern is in line with the high relative humidity and low evaporation rates during the period. Overall, the predicted moisture content differed from the measured values by 14.2% and 12.7%, for bales #3 and #4, respectively. However, like the case of the uncovered bales, the trends of these two sets of values (predicted versus measured) are also compatible. Figure 4.3. Daily mean air temperature for Vancouver, British Columbia from June 1, 2011 to May 31, 2012 68 Figure 4.4. Daily air relative humidity for Vancouver, British Columbia from June 1, 2011 to May 31, 2012 Figure 4.5. Daily precipitation for Vancouver, British Columbia from June 1, 2011 to May 31, 2012 69 Figure 4.6. Daily net radiation from June 2011 to May 2012 Figure 4.7. Calculated daily evaporation rate for the uncovered bales 70 Figure 4.8. Calculated daily evaporation rate for the covered bales Figure 4.9. Predicted and measured moisture contents of the uncovered Aspen bales (bale #1 and #2) in the field (June, 2011 to May 2012) 71 Figure 4.10. Predicted and measured moisture contents of the covered Aspen bales (bale #3 and #4) in the field (June, 2011 to May 2012) 4.4.2 Prediction of the moisture content of a bale with transparent cover The model was extended to predict moisture contents of a bale with transparent cover. With this type of cover, the materials can be protected from precipitation but not solar radiation. The logging residues left in the field typically have moisture content as high as 50% wet basis (equivalent to 100% dry basis). Hence, the simulation started with an initial moisture content of 50% wet basis. The weather data records during the 2011-2012 period were again used as inputs to the simulation model. The predicted results are shown in Figure 4.11. Moisture content of the covered bales keeps decreasing from June to September as the weather is warm and dry together with adequate sunshine in this period. Accordingly, the evaporation rate is relatively high during these four months, leading to a continuous reduction of moisture to around 13% dry basis in the materials. As the climate becomes cool and wet from November to March, the predicted moisture content is seen to increase somewhat and fluctuate around 25% dry basis. This increase in moisture content may be attributed to the high relative humidity of 85-90% in the surrounding atmosphere, which 72 leads to moisture adsorption by the materials. However, there is no large increase in moisture content as the bales are protected from precipitation. When it progresses to April and May with warmer and drier climate conditions, the moisture in the materials starts to evaporate again. The predicted moisture contents in this case demonstrate that the natural drying process could be effectively improved by sheltering the bales under a transparent cover. Although the final moisture content has not yet reached a sufficiently low value which makes the material suitable for pelletizing if it needs be, the natural drying process can already help to reduce energy consumption. Thus, the lumped model adopted in this study may be used to estimate the moisture variations in the biomass when subject to various weather conditions, at least as a first approximation. Figure 4.11. Predicted moisture contents of Aspen bale placed under a transparent cover 73 4.5 Conclusion A lumped model was applied to predict the moisture relations for Aspen materials. With the inputs of available weather data which include temperature, relative humidity, wind speed, solar radiation and precipitation, the model was able to estimate the daily evaporation and hence the moisture content of Aspen under both covered and uncovered conditions during the one-year storage period (June 2011-May 2012). For the uncovered bales, the predicted moisture contents fluctuated around 20% (dry basis) in the summer time. The high moisture contents of the materials estimated during the November-March period were in line with the low temperature and high precipitation conditions. By comparison, for the covered bales, higher moisture contents as predicted for the period extending from October to March result from the low evaporation rates which corresponded to high relative humidity conditions. On average, the predicted moisture contents were 24% and 14% different from the actual values, for the uncovered bales and the covered bales, respectively. Nevertheless, the measured moisture contents exhibited the same trend as the predicted moisture contents for both situations. The model was extended to predict the moisture content of a bale with transparent cover, such that the materials are protected from precipitation but not solar radiation. Results show that the natural drying process could be speeded up and effectively improved by sheltering the biomass. In conclusion, the lumped model presented in this thesis research may be used as a first approximation, and applied to estimate the moisture content of Aspen or similar biomass during relatively long-term field storage with a reasonable degree of accuracy. 74 Chapter 5. Gas emissions from stored Western Red Cedar chips 5.1 Introduction Due to the non-renewable of nature of fossil fuels and their negative impacts on the environment, the utilization of alternative fuel to generate heat and power has become an important element of sustainability in today’s world. Biomass has been increasingly used due to its notable lower GHG and criteria air pollutant emissions as compared to fossil fuels, and no regional restrictions as compared to other renewable energy sources such as wind and solar (Zhang et al., 2009). The development of bioenergy and biomaterials from sustainable biomass will lead to a new scene of industry (Ragauskas et al., 2006). In Canada, many clean energy projects have been put in place. A large amount of lignocellulosic biomass are converted to biofuels and integrated into the energy generation systems. Gas emissions (CO2, CO, CH4 and volatile organic compounds VOCs) from woody biomass storage systems have been studied in recent years. Leinonen and Tutkimuskeskus reported the emission of considerable VOCs that are mainly terpene compounds from woody biomass (Leinonen & tutkimuskeskus, 2004). The major VOCs emitted from stored wood pellets were aldehydes, some of which are known to cause irritation to the respiratory system (Hagstrm, 2008). Svedberg identified that the storage of wood pellets led to the emission of high levels of hexanal as a result of the general degradation processes of wood (Svedberg et al., 2004). VOCs are also generated during wood drying. For instance, Granstroem and Mansson (Granstrom & Mansson, 2008) suggested that formaldehyde, a common VOC produced during wood drying, has short-term health effects such as eye and throat irritation, and factors such as temperature and moisture content would affect the emissions from the materials. Arshadi et al. found that high temperatures used for drying sawdust led to higher emissions of aldehydes and ketones from pellets (Arshadi et al., 2009), whereas Stahl et al. observed increases in the total amount of VOCs, including the faster release of terpenes with increased drying temperature (Stahl et al., 2004). In terms of CO2, CO, and CH4 emissions, higher emission factors were associated with higher temperatures, whereas increased relative humidity in the enclosed container 75 increased the rate of gas emission and a corresponding depletion of oxygen (Kuang et al., 2009; Shankar et al., 2008). Wihersaari (Wihersaari, 2005a) concluded that the CO2 emissions from fresh forest residues were almost three times higher than the dried materials, and suggested that mixing the heaps during the storage period would probably cause increased emissions rates. Rupar and Sanati investigated wood chip piles in an existing terminal storage and revealed an increase in air emission when the temperature directly above the pile increased; also, more terpenes were released with a greater amount of precipitation (Rupar & Sanati, 2005). In a lab-scale study of gas emission from stored fresh Douglas fir residues by He et al., results showed that higher temperature led to higher gas concentrations and greater dry matter losses (He et al., 2012). Emissions of CO, CO2 and CH4 are likely due to biodegradation and auto-oxidative reactions of organic constituents naturally present in wood (Kuang et al., 2008). CO2 can be generated from thermal oxidation, aerobic biodegradation or anaerobic biodegradation. Svedberg et al. and Arshadi and Gref postulated that CO is formed from the auto-oxidative degradation of lipids and fatty acids present in wood, and storage temperature is one of the critical factors (Svedberg et al., 2004; Svedberg et al., 2008) (Arshadi & Gref, 2005). Hellebrand and Schade suggested that CO generation is independent of microbial activity in the feedstock, but is promoted by increased temperatures and available oxygen (Hellebrand & Schade, 2008). Moreover, CO produced from plant litter was most likely caused by thermochemical oxidation rather than a biological process (He et al., 2012). Some organic acids such as acetic acid are likely to be emitted from the breakdown of wood hemicellulose (Johansson & Rasmuson, 1998). The extractives in wood have also been realized as one source of VOCs; aromatic and ether extractives are released when temperature increases. Under low temperature, the breakdown of polysaccharides and functional groups of hemicellulose and lignin result in the emission of methanol, light aldehydes, formic acid and acetic acid as the major substances (Koppmann et al., 2005). Recently, the University of British Columbia (UBC) has installed a gasification plant as part of a District Energy System on campus at Vancouver, BC, Canada. Western Red Cedar (WRC) chips are being considered as a possible fuel for this plant. Other biomass feedstocks such as spruce-pine-fir residues and urban tree trimmings and wood waste are 76 also used. WRC is an abundant softwood species in British Columbia. The residue from WRC harvest can be used in the form of chips or pellets for combustion, or converted to liquid biofuels (Liu et al., 2010; Nakamura et al., 2010; Zhang et al., 2011). Problems pertinent to the storage of high moisture WRC include gas emissions, odours, dry matter losses and fire risk (Rentizelas et al., 2009). Western Red Cedar is known to have a higher extractives (around 10%) than other wood species (for instance, Douglas Fir 5.9%) (Gonzalez, 1997). Cellulose content is slightly lower but the lignin content is higher than other species. Hemicellulose compounds are in the same range (13-14%). Potential emissions from storage and drying of cedar chips have generated concerns about the storage and utilization of WRC than other wood species. The objective of this chapter is to quantify gas emissions from Western Red Cedar under different storage conditions. Findings from this study can provide the data required for assessing the potential impact of emissions from stored Western Red Cedar on the environment and human health. This can assist in the better handling and management of fresh biomass prior to their utilization in different energy production processes. 5.2 Materials and methods 5.2.1 Materials Western Red Cedars (WRC) are harvested in British Columbia, primarily on Vancouver Island and mid and south coast of British Columbia, Canada. Subsequently, the logs are transported in water booms and tug-boated to sawmills in the Lower Mainland Vancouver area. The residual chips are created from waste wood which is cut off to make the appropriately sized products for sale as boards, siding and fencing. They are taken from a chipper, which chips and screens the chips for fines and oversized pieces, leaving a uniform chip for pulp or for bioenergy applications. Fresh WRC chips as shown in Figure 5.1 were obtained from a recycling yard in Langley, BC. The chips were stored in a cold room at 4oC at the UBC laboratory. The chips had an average size of 10-20 mm, as measured by the Gilson Testing Screens TS-1 and TS-2 77 (Gilson Company, Worthington, Ohio). Moisture content of the samples was determined in triplicate in a forced-air convection oven at 103°C for 24 h according to ASABE Standards S358.2 (ASABE, 2010a). Initial moisture content of the chips was around 50% (wet mass basis). A number of glass containers (2L) were fitted with valves and sampling ports and assembled. Each glass container (reactor) was loaded with 330 g wood chips. The wood chips with initial moisture content of 35% (wet basis) were also used to study the gas emissions. Figure 5.1. Western Red Cedar chips as received from the recycling yard. The size of chips varied from 30-80 mm in length on average 5.2.2 Experimental setup Two series of tests were conducted in order to simulate the storage environment, one under aerobic and the other under non-aerobic conditions. The materials in a pile experience different environmental conditions depending on the availability of oxygen. Biomass in the outer layers of pile close to the surface may be assumed to be more aerobic due to adequate oxygen supply from the air. However, at a certain depth from the surface, it may be subject to non-aerobic conditions after a period of time because of oxygen deficiency. At the extreme, the inner core of the pile may even experience anaerobic condition due to a complete lack of oxygen. Ten reactors were divided into two equal groups. As shown in Figure 5.2, non- aerobic reactors were sealed at all times to study gas emission under airtight (non-aerobic) conditions. Air was pumped into the aerobic reactors after daily gas sampling event in order 78 to replenish and maintain a high oxygen level (aerobic condition) in the reactors. After the reactors were loaded with the materials, they were sealed and placed either in a cooler at 5oC, or in ovens with temperature maintained at 20oC, 35oC, 45oC and 50oC. The range of temperature adopted for the test represents cool to hot climate conditions in different geographic locations, and involving seasonal variations. As in summer time, some part within the pile can heat up to as high as 50oC with the help of solar radiation and microbial activities. Two replicates were performed for each test. In all cases, the experiment was run for approximately two months. The same experimental procedure was also conducted on the gas emissions from wood chips with initial moisture content of 35% (wet basis). Figure 5.2. Left reactors for aerobic tests were ventilated every 24 hours during the experiment. Right reactors for non-aerobic tests remain sealed for the duration of the experiment 5.2.3 Gas emission measurement The concentrations of CO2, CO and CH4 along with O2 were analyzed by gas chromatography (Model SRI 8610C, Mandel, USA). The GC was calibrated regularly with the corresponding standard gases. A GC/MS analyzer (Model 5975B/6890N, Agilent Technologies, USA) was used for qualitative analysis of the VOCs. During each sampling event, gas sample was drawn from each reactor to measure CO2, CO, CH4 and O2 concentrations. Gas sampling occurred daily for the aerobic reactors, while gas samples from 79 the non-aerobic reactors were measured more frequently in the earlier period of the test. 250µL gas sample was drawn from each reactor to analyze the VOC composition. The examples of GC, GC/MS spectra are shown in Appendix D. For the aerobic reactors, the total concentration of VOCs (TVOC) was measured by a portable VOC monitor (Model PGM, RAE Systems, San Jose, CA) on a daily basis. For the reactors under non-aerobic conditions, the same procedure cannot be applied since this would cause a larger amount of gases to be released from the reactor, thus affecting the accuracy of gas analysis for the remaining test (storage) period. Hence, TVOC for non- aerobic containers was measured at the end of the test period. 5.2.4 Microbial analysis At the end of all tests, chip samples were taken from each reactor and sent to a microbiology laboratory in Vancouver, BC for microbial analysis. The methodology used for total bacterial counts followed the MFHPB-18 Standard (Canada Health, 2001); and enumeration of yeasts and molds in the samples was performed using MFHPB-22 Standard (Canada Health, 2004). 5.2.5 Data analysis In this study, the emissions of CO2 and CO are expressed as emission factors, in units of [gram gas species per kilogram dry matter DM). Emission factor is a cumulative parameter. For the non-aerobic (airtight) reactors, the gas concentrations are expected to accumulate with time. For the aerobic reactors, the day-to-day gas concentrations will not accumulate, and emission factor would be calculated based on the sum of the daily gas concentrations. The measured gas concentrations were converted from volumetric percentage to emission factors by using the N2 balance method (Kuang et al., 2009), assuming that nitrogen would not be consumed during the test period. At constant temperature and pressure, the emission factor f in [g/kg DM] is related to volumetric gas concentration Ci as follows (Appendix B): 80 ( ) 0s i wt n i nt P CV M C f RTmC = (5.1) where i is the gas species, Ps is absolute pressure of gas in the container (Pa), Ci is volumetric concentration of a particular gas (m3 of gas species per m3 of gas going through the GC), V is gas volume in the reactor (m3), Mwt is gas molecular weight (g/mol), Cn0 is initial concentration of nitrogen (%), Cnt is the concentration of nitrogen at time t (%), R is universal gas constant (8.31 J/mol.K), T is temperature (K), and m is the total mass of materials in the container (kg). The gas volume is the numeric difference between the volume of container and volume of chips. Change in pressure during the whole test was assumed to be minimal in general. 5.3 Results and discussion 5.3.1 Gas emissions 5.3.1.1 Non-aerobic conditions Results from the non-aerobic reactors in terms of emission factors are shown in Figure 5.3. For each temperature, the CO2 emission factor (fCO2) increased gradually due to the accumulation of CO2 with time. It is evident that the emission factors were the highest at 20oC followed by 35oC, but the emission factors were substantially reduced at temperatures of 45 and 50oC. The CO2 emission factor was also lower at 5oC. The trends of the CO2 profiles were exactly opposite to the O2 profiles for all temperatures. After about three weeks, fCO2 from the 20oC reactor reached a plateau of 2.8 g/kg DM as oxygen content was depleted to nearly 0%. This value corresponds to a CO2 concentration of 16% (on volumetric basis). For the 35oC reactor, fCO2 increased to an asymptotic value of 2.7 g/kg DM after 45 days storage. By comparison, fCO2 increased slowly at 5, 45 and 50oC, reaching only 1.1, 0.6 and 0.4 g/kg DM (or, CO2 concentration 2-7%), respectively after 60 days’ storage, implying that CO2 generation could be dominated by biological mechanism rather than chemical mechanism. 81 Based on the CO2 emission data (Figure 5.3a), the effect of temperature on CO2 emission factors was analyzed and plotted in Figure 5.4. The quadratic polynomial ( 2f aT bT c= + + ) was found to fit well to the data recorded at storage times beyond day 5. The coefficients (a, b, c) have different values for each curve at different times in Figure 5.4. Again, it shows that fCO2 is the highest at 20oC, supporting the postulation that the generation of CO2 was dominated by biological mechanism rather than chemical mechanism. CO was only observed to emit from the 35, 45 and 50oC reactors (Figure 5.3b). For each temperature, the CO emission factor fCO increased gradually with time, though it is two orders of magnitude less than the CO2 emission factor. Furthermore, fCO increased significantly with temperature, being 0.033 g/kg DM at 50oC vs. 0.002 g/kg DM at 35oC at the end of storage, and this trend is opposite to that observed for CO2. In general, CO generation is due to the chemical oxidation of materials and is promoted by increased temperature and the availability of oxygen. Our observations support this mechanism. Biomass can be decomposed both chemically and biologically. When chemical oxidation is dominant, the emitted gases will generally increase with increasing temperature according to the Arrhenius kinetics relationship. Temperature is among the most important environmental factors that control biological processes, whereby responses to temperature can be categorized in terms of three cardinal temperatures, namely the base or minimum temperature, the optimum temperature, and the maximum temperature. In the case of microbial ecology, at temperature lower or higher than the optimum, bacteria and fungi could become dormant or even killed (Agrios, 2004). 82 Figure 5.3. Gas emission profiles from stored WRC chips under different temperatures (non- aerobic) (□: 5oC; ●: 20oC; ∆: 35oC; ▼: 45oC; *:50oC) 83 Figure 5.4. Effect of temperature on CO2 emission factors during the testing period under non-aerobic conditions 5.3.1.2 Aerobic conditions The gas emissions from the aerobic reactors are shown in Figures 5.5a and 5.5b. At each temperature, CO2 and CO emissions increased slowly with time while oxygen levels remained close to 20% during the entire test period (Figure 5.5c). Daily pumping of air into the aerobic reactor was effective in keeping the O2 content relatively high and constant. The CO2 emission profiles exhibited trends opposite to the O2 profiles for all temperatures. In terms of temperature effect, aerobic reactors displayed similar trends to non-aerobic reactors. The peak CO2 emission factor was again observed at 20oC which indicated biological process may contribute primarily to the emission. Depending on the temperature, the emission factors increased with time at different rates. The total CO2 emissions ranged from 1-7.5 g/kg DM depending on temperature. CO was only observed to emit from the 35, 45 and 50oC reactors 84 (Figure 5.5b). The CO emission factor exhibits the same trend as the CO2 emission factor, and it had a positive relationship with temperature. By comparison of Figure 5.5 with Figure 5.3, the total CO2 produced during the entire storage period was higher for the aerobic reactors. Over 55 days, the highest total CO2 emission from the non-aerobic reactors and the aerobic reactors was 2.8 g/kg DM and 6.6 g/kg DM, respectively, for the 20oC temperature. The CO emissions from both conditions were similar. The difference between the CO2 emissions from non-aerobic and aerobic conditions is essentially due to oxygen content. As oxygen was depleted to zero at 20oC in the non-aerobic reactors, CO2 emission reached the plateau and did not increase further. In contrast, CO2 was produced continuously in the aerobic reactors with the oxygen concentration almost close to ambient level. The effect of temperature on CO2 emission factors was plotted in Figure 5.6. The curves at different storage times may be represented by a polynomial equation ( 4 3 2f aT bT cT dT e= + + + + ). Like the case of non-aerobic conditions, it shows that the highest CO2 emission again occurred at 20oC during the whole storage period. This supports the argument that the biological mechanism could be dominant during the process. 85 Figure 5.5. Cumulative gas emissions from stored WRC chips under aerobic conditions (□: 5oC; ●: 20oC; ∆: 35oC; ▼: 45oC; *:50oC) 86 Figure 5.6. Effect of temperature on CO2 emissions during the storage period under aerobic conditions The CO2 and CO emissions could have safety and health implications, depending on the duration of exposure. The threshold limit value-time weighted average (TLV-TWA) values for 8h exposure to CO2 and CO are set at 0.5% and 0.0025%, respectively. The TLV- STEL (Short-Term Exposure Limit) values for 15 min exposure to CO2 and CO are 3% and 0.01%, respectively. The corresponding concentrations that are deemed “Immediately Dangerous to Life and Health” for CO2 and CO are 4% and 0.12% (ACGIH, 2004). Hence, the CO2 emissions observed from the non-aerobic reactors in this study exceeded these threshold values by a rather large margin. Another comparison was made with respect to stored wood pellets (3.7-10% moisture content, wet basis) under oxygen-depleting environment. (Kuang et al., 2009) reported the CO2 emission factor to increase from 0.025 to 0.23 g/kg DM as temperature was raised from 10 to 45oC after 30 days storage, while the corresponding CO emission factor increased from 0.001 to 0.058 g/kg DM. They concluded that chemical auto-oxidation process could be the 0 2 4 6 8 0 20 40 60 Em is si on fa ct or (g /k g D M ) Temperature (oC) day 1 day 5 day 10 day 20 day 40 day 60 87 dominant mechanism for the gas emissions from wood pellets. The highest fCO2 from Western Red Cedar chips in this study was almost 10 times greater than those derived from wood pellets, whereas fCO was similar to those from pellets. When the initial O2 content was high in the reactors, it is possible for the fresh wood chips to release a greater amount of CO2 than the wood pellets due to biological process in addition to chemical oxidation. At the later stage of storage when anaerobic condition would prevail at O2 content close to zero, the biological process was inhibited so that no more CO2 was generated; thus the emission factor of CO2 reached the peak value. Moisture content is a key factor that governs biological reactions. Kuang et al.(Kuang et al., 2009) postulated that biological process may contribute to the emissions for moist biomass such as wood chips; our findings from this study confirm their suggestion. The large differences in fCO2 may be attributed to the much higher moisture content of the fresh wood chips (50% wet mass basis) with live microbes, whereas most of the microbes should have been killed during pelletization and the pre-pelletization drying process at high temperature. 5.3.2 Gas emissions from sterilized woodchips In order to verify the dominance of CO2 emission by the biological process, another series of experiment was conducted. The WRC woodchips were sterilized in a bench-top steam sterilizer (Sterilemax Series 1277 Table Top Model, Thermo Scientific, USA) at 135oC for 15 min to eliminate the microbes. Steam condition was adjusted to maintain the moisture content of the chips at the same level. After cooling down, 330 g of steam-sterilized woodchips was loaded into each reactor. The experiment was conducted at temperatures of 5oC, 20oC, 35oC, 45oC and 50oC. Results from running the test under non-aerobic condition are shown in Figure 5.7. After 60 days storage, the CO2 emission factors for the treated (sterilized) woodchips were lower than those from the untreated woodchips with temperature ranging from 5-50oC (Figure 5.3). It shows the generated CO2 due to chemical mechanism was around 4 times less when comparing to that due to biological mechanism. Evidently, the emissions of CO2 and CO increased significantly at higher temperatures. It indicated that biological mechanism 88 was no longer dominant for CO2 emission as the microbes would have been killed by the sterilization process. Thus, according to the findings, biological process played a leading role in generating CO2 emissions from the stored WRC wood chips. Microbial analysis results revealed that the total bacteria counts (TBC) for the wood chip samples obtained from the aerobic reactors at the end of the tests were 3000, 10000, 4500 and 30 cfu/g sample whereas fungi counts ranged from 5-60 cfu/g sample at 5, 20, 35 and 50oC, respectively. The highest microbial activity was in line with the highest CO2 emission at 20oC. By comparison, TBC were substantially lower for the wood chip samples obtained from the non-aerobic reactors at the end of the tests, with a maximum value of 140 cfu/g also found at 20oC. It shall be noted that the samples were sent for microbial analysis at the end of the tests. The microbial activity could have been higher with shorter storage time, leading to higher bacterial counts. It is known that Western Red Cedar wood is resistant against decay as extractives such as thujaplicins are present. Percent dry matter losses of the materials were determined to be 0.06-0.26% for the non-aerobic reactors, as compared to 0.15-0.57% for the aerobic reactors under different temperatures. Such small degree of dry matter losses over the storage period compared to those reported for other materials in the literature seems to reaffirm the decay-resistance characteristics of WRC. Therefore, the CO2 emissions observed in this study could be primarily due to microbial respiration rather than decomposition of the biomass. 89 Figure 5.7. Gas emission from treated WRC chips under different temperatures (non-aerobic) (□: 5oC; ●: 20oC; ∆: 35oC; ▼: 45oC; *:50oC) 90 5.3.3 Characteristics of VOCs The components of VOCs and their total concentration were analyzed in this study. A range of chemical compounds were identified from the qualitative GC/MS analysis. The molecules are separated at different retention times by GC, and they are further identified using mass- charge ratio by MS. Some examples of GC/MS spectra are shown in Appendix D. Measurements indicated that at all temperatures tested, the VOCs emitted were aromatic compounds, including benzene and its derivatives. Methanol, aldehydes, terpene, acid, acetone, hexane, ketone, ethers and esters were found in the gas emissions from the WRC woodchips, similar to other wood products (Arshadi et al., 2009; Hagstrm, 2008; He et al., 2012; Leinonen & tutkimuskeskus, 2004; Svedberg et al., 2004). Furan was also detected in the emitted gases; it may be attributed to the uptake of chloride ions in the WRC wood chips during transport in the salty waters of Pacific Ocean. Indole was found to emit from the reactors under high temperatures of 45oC and 50oC. These VOCs could be inhaled by people working near the woodchip storage area, or nearby residents, and causing irritation to the respiratory system. The odorous VOCs such as indole, terpenes, aldehydes and ketones can also induce odour nuisance problems. Figure 5.8 illustrates the cumulative concentrations of total VOCs (TVOCs) with time from the aerobic reactors. It can be seen that temperature is positively correlated with TVOC concentration. Temperature can induce the release of VOCs; the results are similar to those reported by other researches (Arshadi et al., 2009; Stahl et al., 2004). At the end of 60 day storage, the TVOC concentrations were 20, 110, 360, 660 and 800 ppm for temperature ranging from 5 to 50oC. By comparison, the concentrations of TVOC emitted from the non- aerobic reactors were somewhat lower, at 11, 89, 237, 429 and 560 ppm, for the same range of temperature. Taking the total CO2 emissions together with the total TVOC emissions over the entire storage period, a positive correlation between the percent dry matter losses and gas emission was obtained. 91 Figure 5.8. Cumulative concentration of TVOC from the aerobic reactors at various temperatures 5.3.4 Gas emissions from wood chips with different initial moisture content The gas emissions from stored WRC chips with lower initial moisture content of 35% (wet basis) were also studied. For the non-aerobic condition, the gas emission trends (Figure 5.9) are similar to those for initial moisture content of 50% wet basis (Figure 5.3). The peak fCO2 from the 20oC reactor was around 1.9 g/kg DM; this is 30% lower than fCO2 for WRC chips with 50% moisture content wet basis). Similar observations were made at other temperatures. CO was observed to emit from the reactors under 35, 45 and 50oC, with emission factors varying from 0.001 to 0.02 g/kg DM at the end of the storage. These values are also lower than those associated with WRC chips having higher initial moisture content. Materials with higher moisture content could generate higher emissions of CO2, either due to biological process (Wihersaari, 2005a) or chemical reaction process (Kuang et al., 2009). In terms of biological aspect, microbial activities have an important effect on the process. As previously mentioned, moisture content is known to be an important factor 92 affecting the gas emission from stored biomass (Kuang et al., 2009; Wihersaari, 2005a). It is one of the key factors that govern biological reactions. Moisture content was found to significantly affect the size and activity of the microbial community because of its control on the respiration (Bruce, 1985). The respiration rate of the microbes increased linearly with increasing moisture content (Boddy, 1983). A positive correlation was found between the percentage humidity, the microbial growth rate constant and the evolution of CO2. And dryness inhibits the microbial activity (Ertekin & Yaldiz, 2004). It was reported the naturally dried forest residue (moisture 40 wt%) is evaluated to have emitted totally 58 kg CO2 eq when comparing to 144 kg CO2 eq from the fresh forest residue (moisture 60 wt%) after a 6- month storage (Wihersaari, 2005a). It was also reported that moisture is an important factor affecting the chemical reaction rate (Kuang et al., 2009). High moisture content is favorable for the decomposition process. It allows rapid degradation circumstances. So, materials with higher moisture content could generate higher emissions of CO2 and CO. 93 Figure 5.9. Gas emission profiles from stored WRC chips with initial moisture content of 35% wet basis under different temperatures (non-aerobic) (□: 5oC; ●: 20oC; ∆: 35oC; ▼: 45oC; *:50oC) 94 The cumulative gas emissions from the aerobic reactors were presented in Figure 5.10. CO2 emission was the highest at 20oC and lowest at 50oC. Again, fCO2 was lower for all temperatures when compared to emissions from the WRC chips having higher moisture content of 50% (w.b.). The lower moisture WRC chips also emitted less CO. Comparison with the non-aerobic reactors shows that total CO2 emissions from the aerobic reactors were greater, but the total CO emissions were lower. Figure 5.10. Cumulative gas emissions from stored WRC chips with initial moisture content of 35% wet basis under aerobic conditions (□: 5oC; ●: 20oC; ∆: 35oC; ▼: 45oC; *:50oC) 95 5.4 Conclusion Gas emissions from stored Western Red Cedar chips were studied. Under non-aerobic conditions, the emissions of CO2 and CO increased gradually with time for all temperatures. At higher temperatures, higher CO emissions were measured, which may be attributed to chemical oxidation mechanism. In contrast, lower emissions of CO2 along with higher O2 concentrations were observed at higher temperatures. At 20oC, CO2 emission was the highest, followed by 35oC. After three weeks, the emission factor of CO2 reached a plateau of 2.8 g/kg DM (or, 16% concentration) as oxygen content was depleted to 0%, indicating that the storage environment has turned anaerobic. By comparison, oxygen content was lowered to 10-15% for the other temperatures. Results further showed that CO2 and CO emissions from the aerobic reactors exhibit similar trends as the non-aerobic reactors with respect to the effect of temperature. The cumulative emissions of both CO2 and CO at each temperature increased slowly with time, with higher oxygen levels of close to 20% being maintained during the entire test period. Biological reaction could be dominant during the storage of wet biomass, leading to the highest level of microbial activity at 20oC rather than at higher temperatures. Wood chips were then sterilized to delineate the mechanism that dominates gas emissions. Results indicated that the biological mechanism was no longer the dominant mechanism for CO2 emission as the microbes would have been killed by the sterilization process. Thus, biological process played a leading role in generating CO2 emissions from the stored WRC wood chips. Microbial analysis results are compatible with the CO2 emission results. Qualitative GC/MS analysis results indicated that the major VOCs emitted include benzene and its derivatives, methanol, terpene, aldehydes, acids, alkane, indole, furan, acetone, ethers and esters. The total concentration of VOCs was found to be higher at higher temperatures for both aerobic and non-aerobic storage conditions. Under non-aerobic conditions, the concentrations of TVOC at the end of 60 days storage were 11, 89, 237, 429 and 560 ppm with the temperature ranging from 5 to 50oC. The TVOC concentrations from the non-aerobic reactors were lower than the TVOC concentrations from the aerobic reactors 96 over the entire storage period. Taking the total CO2 emissions together with the total TVOC emissions over the entire storage period, a positive correlation between the percent dry matter losses and gas emission was found. The gas emissions from the materials with lower initial moisture content of 35% (wet basis) exhibited the same trends as those from higher initial moisture content of 50% (wet basis) under both aerobic and non-aerobic conditions. However, the amounts of gases were found to be lower to different extents, indicating that moisture content is another important factor that affect the gas emissions from stored biomass, as it affects microbial activity as well as chemical reaction rate. Overall, the gas emission results from this study reaffirm the importance of ensuring safe and environmentally friendly storage conditions for wet woody biomass such as the WRC woodchips, and controlling the emission of odorous VOCs. 97 Chapter 6. Gas emissions from stored Douglas fir (Pseudotsuga menziesii) 6.1 Introduction In Canada, large amounts of woody biomass constitute renewable resources for conversion into solid or liquid biofuels. Woody biomass are partially generated by the wood products industry, including paper mills, sawmills, and furniture manufacturing; also it could be the by-product of forest management including mostly of tree branches, tops of trunks, needles, leaves, and other woody parts left on the forest floor (Easterly & Burnham, 1996); urban wood wastes is another source, including tree trimmings, land clearance and other wood products. Douglas fir as a popular softwood species makes up some 25% of the tree species in coastal British Columbia. Douglas fir residues mainly include wood chips, bark, needles and leaves. Traditionally, wood chips are used in pulp mills. Today, these woody residues can also be the source of fuel for bioenergy production (Phanphanich & Mani, 2009). In the pulp industry, wood chips are normally made from debarked round wood. However, for heating purposes, whole trees or logging residues including bark and needles are used (Hellenbrand & Reade, 1992). After logging and other process operations, the forest residues are usually piled in the forest or on site. Biomass feedstocks need to be stored for a period of time to ensure continuous availability for biofuel production during the growing season and winter months. These biomass have high moisture content over 40% (wet basis), which induce problems during their storage, including gas emissions and dry matter losses due to degradation (Krupińska et al., 2007). It is well known that all biomass gradually decomposes over time, both chemically and biologically. Losses of dry matter are due to the decomposition of forest residues. Microbial activity in the stored biomass, especially fungal growth, is the major cause of the initial heat development which in turn encourages further growth leading to high temperatures in the pile. In extreme cases, this would cause self- ignition and potentially fire (Jirjis, 1995). Materials with green parts are relatively rich in nutrients, and are even more favourable for microbial growth (Thörnqvist, 1984). During the 98 period of fast decomposition, there exist the risks of emissions, energy losses, and fires (Wihersaari, 2005a). Previous studies have been conducted on dry matter changes for a variety of woody biomass (Afzal et al., 2010; Eriksson & Gustavsson, 2010; Pettersson & Nordfjell, 2007; Thornqvist, 1985; Wihersaari, 2005a). A number of factors such as the physical characteristics of biomass feedstocks, weather conditions, and the method and duration of storage can influence dry matter losses during storage. Dry matter produces energy in thermal processes such as combustion; hence dry matter losses would reduce the calorific value and the potential revenue. It is important to minimize such losses during biomass storage. The common gas emissions from biomass are identified to be CO2, CO, CH4 and non- methane volatile organic compounds (VOCs) (Johansson et al., 2004). The generation of CO2 can be from thermal oxidation, aerobic biodegradation or anaerobic biodegradation. Aerobic degradation relies on aerobic microorganisms, which decompose the organic matter producing heat, CO2 and H2O (Tonn et al., 2011). For instance, the breakdown of carbohydrates with oxygen consumption may be represented by the reaction C6H12O6 + 6O2 → 6CO2 + 6H2O. (Svedberg et al., 2004) and (Arshadi & Gref, 2005) postulated that CO is formed from the auto-oxidative degradation of lipids and fatty acids present in wood, and storage temperature is one of the critical factors. (Hellebrand & Schade, 2008) suggested that CO generation is independent of microbial activity in the feedstock, but is promoted by increased temperatures and available oxygen. Moreover, CO produced from plant litter was most likely caused by thermochemical oxidation rather than a biological process. CH4 generation is usually due to anaerobic decomposition of biomass due to the action of microorganisms. The methanogens can use CO2 and H2, or acetic acid to produce CH4. Some VOCs can be emitted from the breakdown of wood hemicellulose (Johansson & Rasmuson, 1998). Wood extractives have also been realized as one source of VOCs. Several factors such as temperature, moisture content and freshness of the materials were found to have effects on the gas emissions from stored biomass. Kuang et al. (2009) concluded from their lab-scale study that peak emissions of these gases were associated with 99 higher temperature and relative humidity in the headspace of the reactors with stored wood pellets. Wihersaari (2005a) found that the greenhouse gas emissions were almost three times higher for fresh versus dried forest residues. Pellets made from aged sawdust were suggested to generate less VOCs than pellets made from fresh sawdust due to the oxidation process (Kuang et al., 2008). Wood chip piles were investigated in an existing terminal storage and it was revealed that air emission increased when the temperature directly above the pile increased (Rupar & Sanati, 2005). When the biomass are stored in confined spaces, these emissions due to decomposition will accumulate to toxic levels, while oxygen will be depleted simultaneously (Svedberg et al., 2009). There are few published studies on the storage of fresh woody biomass in regard to gas emissions. The objectives of this chapter were to investigate the gas emissions from the stored fresh Douglas fir residues under different temperatures, and to measure the dry matter changes during the storage period. Gas emissions were reported in the literature for stored wood pellets with low moisture contents, suggesting that microbes might not be totally eliminated by pelletizing. It is desirable to extend the study to forest residues which have much higher moisture contents that would promote greater microbial activities. 6.2 Materials and methods 6.2.1 Materials The materials used in this study were Douglas fir (Pseudotsuga menziesii) wood chips and greens. Douglas fir (DF) chips as shown in Figure 6.1 were obtained from Fibreco Export Inc., North Vancouver, BC. They originate from suppliers in British Columbia, and are produced from sawmill operations’ residual chips and through whole log chipping. The wood chips contain no more than 1% bark. Fresh DF greens were obtained from trimmings and cut branches of DF trees in the Vancouver area. All materials were stored in a cold room at 4oC to minimize any degradation before the experiment. The size of chips as measured by Gilson Testing Screen (TS-1 & TS-2, Serial No. 2920, Gilson Company, Worthington, Ohio) had a range of 5-30 mm. Moisture 100 content of the samples was determined in triplicate in a forced-air convection oven at 103°C for 24 h to obtain the bone dry biomass according to ASABE Standards S358.2 (ASABE, 2010a). Initial moisture contents of the DF chips and greens were 48% (wet basis) on average, and 54% (wet basis) on average, respectively. A number of glass containers (2L) were fitted with valves and sampling ports and assembled as reactors for the experiment. Each bottle was loaded with 170 g dry mass materials. Figure 6.1. Douglas fir wood chips with the size range of 5-30 mm 6.2.2 Experimental setup The green parts of the residues are expected to be more readily degraded as compared to the woody parts due to the presence of more nutrients. Thus, the emitted gases would be different as a result of the mixing ratio of greens and wood. In order to delineate the differences in gas emission from the greens and the chips, three series of tests were conducted in this study. For each test series, eight reactors were divided into two groups to simulate the storage situation, one group under aerobic and the other group under non- aerobic conditions. The actual storage environment would be in-between these two conditions in terms of oxygen content. In Test Series #1, reactors were only filled with DF chips (Figure 6.2a). Test Series #2 involved only DF greens (Figure 6.2b). For Test Series #3, chips and greens were mixed in the ratio of 1:1 on dry mass basis (Figure 6.2c). In each test, eight reactors were divided into two equal groups. Group 1 reactors are “non-aerobic reactors”, which were sealed at all times to study gas emission under airtight (non-aerobic) conditions. After daily gas sampling 101 event, air was pumped into an aerobic reactor via a metal tube inserted into the bottom of the reactor in order to replenish and maintain high oxygen level (aerobic condition) in the reactor. After the reactors were loaded with the materials, they were sealed and placed either in a cooler at 5oC, or in ovens with temperature maintained at 20, 35 and 50oC. The range of temperature adopted for the test represents cool to hot climate conditions in different geographic locations, and involving seasonal variations. Two replicates were performed for each test. In all cases, the experiment was run for two months. Figure 6.2. Three series of tests with different materials (a: DF wood chips; b: DF greens; c: mixed DF chips and greens); and two kinds of reactors (left: non-aerobic reactor; right: aerobic reactor) 6.2.3 Gas emission measurements Details about the instrumentation for measuring gas emissions and sampling have been described in Chapter 5, section 2.3. The concentrations of carbon-based gas emissions (CO2, CO, CH4) along with O2 and VOCs were analyzed by GC, GC/MS and VOC monitor. Gas samples were taken from the Group 2 reactors daily, while gas samples from the Group 1 reactors were measured more frequently in the beginning of the test than the later period. 102 6.2.4 Microbial analysis Samples were taken from each reactor and sent to a microbiology laboratory in Vancouver, BC for microbial analysis at the end of all tests. The methodology used for total bacterial counts and mold counts has been described in Chapter 5, section 2.4. 6.2.5 Data analysis In this study, the emissions of CO2, CO and CH4 are expressed as emission factor, in units of (gram gas species per kilogram dry matter DM). The conversion of measured gas concentrations into emission factors have been described in Chapter 5 section 2.5. 6.3 Results and discussion Results are presented in this section for the three series of tests which involved DF chips, greens, and mixed chips and greens, respectively. Gas emission characteristics, microbial analysis and dry matter losses will be discussed. 6.3.1 Characteristics of gas emissions from Douglas fir chips (Test Series #1) Results pertinent to the non-aerobic conditions are presented first, followed by the results pertinent to the aerobic conditions. Gas emissions include CO2, CO, CH4, H2 and VOCs. 6.3.1.1 Non-aerobic conditions Figure 6.3 shows the results of gas emissions obtained from the reactors with DF wood chips under non-aerobic conditions. The emission factors of CO2, CO and CH4 gases increased with time under all temperatures due to accumulation in the airtight reactors. The trends of the O2 profiles were exactly opposite to the CO2 profiles for all temperatures. At 20, 35 and 50oC, when the initial O2 content was high in the reactors, CO2 was generated rather rapidly and the emission factor fCO2 reached a plateau of 2.8-3.0 g/kg DM after 10 days. This corresponds to a CO2 concentration of 16.5%, and there are no statistically significant differences between the three temperatures. At the later stage, when oxygen content was 103 depleted to zero, no more CO2 was generated, thus cumulative CO2 emission remained constant. By comparison, under 5oC, fCO2 went up gradually and it attained 2 g/kg DM (or, 11% CO2 concentration) after 50 days storage. That the CO2 emission factor increased with temperature suggested that the emission was likely dominated by chemical oxidation mechanism during storage. In general, CO generation is due to the chemical oxidation of materials and is promoted by increased temperatures and the availability of oxygen. In this test, CO was only observed in the reactors at 35 and 50oC, but not the lower temperatures of 5 and 20oC. At the end of storage, fCO was 2×10-4 and 5.5×10-4 g/kg DM, respectively. These values are equivalent to 0.0019% and 0.005% concentration, and they are four orders of magnitude less than fCO2. Likewise, the emissions of CH4 from the reactors were very low, which is five orders of magnitude lower than fCO2. In biogas production, one microbial metabolic pathway is for methane to be produced from hydrogen and carbon dioxide during the biodegradation process. Since the measured H2 content was low compared to CO2 content (Figure 6.3), the storage environment is deemed unfavorable for the methanogens. As for VOCs, a range of chemical compounds were found by GC/MS at all temperatures tested. The VOCs were identified to be alkanes (hexane), alkenes, terpenes (α- pinene), aldehydes, acids, ketones (acetone), benzene and its derivatives, ethers, esters, and nitrogen compounds. The cumulative concentrations of total VOCs (TVOCs) from the non- aerobic reactors at the end of two-month storage were 12.6, 256, 677 and over 1000 ppm (the detection limit) for temperatures 5-50oC. As temperature increased, there was a significant increase in the TVOC concentrations. 104 Figure 6.3. Gas emission profiles from stored DF chips at different temperatures under non- aerobic conditions (+: 5oC; ●: 20oC; □: 35oC; ▲: 50oC) 105 6.3.1.2 Aerobic conditions As a result of daily pumping of air into the aerobic reactors, the oxygen content in the reactors was kept relatively high during the entire test period (Figure 6.4a). CO2 emission factors (fCO2) over the same period are shown in Figure 6.4b; it can be seen that the daily increment in CO2 emission decreased with time for all temperatures. This could be attributed to the reduction in moisture content of the DF chip samples with time. More emission was observed under higher temperatures, and the fCO2 values varied from 18 g/kg DM at 50oC to 3 g/kg DM at 5oC. This phenomenon is similar to the non-aerobic reactors; thus chemical oxidation mechanism could contribute primarily to the emission. Figure 6.4. Gas emission profiles from stored DF chips at different temperatures under aerobic conditions (+: 5oC; ●: 20oC; □: 35oC; ▲: 50oC) 106 CO was only observed to emit from the 50oC reactor, with emission factor equal to 6×10-4 g/kg DM after 50 days. This is five orders of magnitude lower than the CO2 emission factor. It shall be noted that at the beginning of the tests, both CO2 and CO emission factors were similar in magnitude for the non-aerobic and the aerobic reactors. But thereafter, their differences became increasingly larger. After 50 days, fCO2 was 3 g/kg DM and 18 g/kg DM respectively for the non-aerobic and aerobic reactors at 50oC. No CH4 emission was detected from the aerobic reactors. The components of VOCs emitted from the reactors under aerobic conditions were similar to those detected from non-aerobic reactors. The results from GC-MS analysis showed the following: branched and straight chain hydrocarbons (alkanes, alkenes, alkynes), terpenes, aldehydes, acids, ketones, benzene and its derivatives, ethers, esters, and other compounds. Figure 6.5. Cumulative concentration of total VOCs (TVOC) at different temperatures from aerobic reactors - wood chips Figure 6.5 illustrates the profiles of cumulative TVOC concentrations from the aerobic reactors for DF wood chips. On day 1, the TVOC concentration of 500 ppm was 107 significantly higher at 50oC than other temperatures, for instance, 0.5 ppm at 5oC. Like the carbon-based emissions, the daily increments of TVOC concentration also decreased with time to different extents depending on temperature during the storage period. Other researchers have also reported that temperature could induce the release of VOCs; however, heat treatment of the materials may lower the subsequent VOC emissions (Arshadi et al., 2009; Hyttinen et al., 2010; Stahl et al., 2004). Aside from temperature, moisture content is another factor that affects the emissions. Since VOCs originate from the biological and chemical breakdowns of wood and extractives, their generation rates would slow down as the moisture content of materials decreased with time. After 50 days storage, the cumulative TVOC concentration was almost 10,000 ppm at 50oC, while the corresponding values at 5, 20 and 35oC were substantially lower, ranging from less than 10 to 1000 ppm. 108 6.3.2 Characteristics of gas emissions from Douglas fir greens (Test Series #2) 6.3.2.1 Non-aerobic conditions Figure 6.6. Gas emissions profiles of stored DF greens at different temperatures under non- aerobic conditions (+: 5oC; ●: 20oC; □: 35oC; ▲: 50oC) The CO2 emission factors increased slowly at 35 and 50oC, while the generation of CO2 at 5 and 20oC was fast it the beginning and then it slowed down at the later period of storage. After two-months storage, at 35 and 50oC, fCO2 increased from 8.5 and 5 g/kg DM to 10.8 and 7 g/kg DM (which correspond to CO2 concentrations of 43% and 32%), respectively. In contrast, fCO2 was the highest at 20oC in the first two weeks of storage, approaching a plateau of 17.5 g/kg DM. Thereafter, it was overtaken by fCO2 at 5oC, when the peak value of 25 g/kg DM was attained after two months. This phenomenon of higher emission factor under lower 109 temperature implies that biological mechanism was dominant in this test. It shall be noted that CO emissions were 4-5 orders of magnitude lower than CO2 emission (Figure 6.6). Oxygen content depleted quickly within 2 days from the beginning of the test, and no oxygen was detected at all temperatures during the test. As there was little to no oxygen in the reactors but CO2 emission kept increasing at 5oC, the emission of CO2 could be due to the anaerobic biodegradation of the materials. Furthermore, CH4 was practically not produced (6 orders of magnitude lower than CO2 as seen in Figure 6.6). GC-MS analysis results show that the VOC components from stored DF greens under non-aerobic conditions are mainly hydrocarbons, aromatic hydrocarbons, terpenes, ketones, aldehydes, methanol, acids and esters. For instance, α-pinene and acetone were found at all temperatures, whereas camphene and β-pinene were detected at 20 and 35oC. Hydrocarbons such as hexane and heptane were present. Pyridine, ammonia, indole and phenyl-furoxan were detected at 50oC, while nitrogen and sulfur compounds were found at 35 and 50oC temperatures. Higher temperature could have induced the breakdown of organic matter and generate various VOCs. The release of odorous VOCs was observed upon opening the 50oC reactor after the test. The total VOCs concentrations from DF greens under non-aerobic conditions are quite high. After two-month storage, the concentrations from the reactors at 20oC, 35oC and 50oC all exceeded 1000 ppm, which is significantly higher than the TVOC concentration of 210 ppm for the 5oC reactor. These results suggested that the DF greens are more readily degradable materials. 6.3.2.2 Aerobic conditions Figure 6.7 shows that daily consumption of oxygen under 20oC was considerable, as the oxygen concentration was depleted to almost zero at all times whereas oxygen content was relatively high around 18% at 50oC during the test. CO2 and CO were measured from the DF greens under aerobic conditions during the whole storage period. Unlike the DF chips, the daily increment of CO2 emission factor was somewhat constant for all temperatures. After 60 days storage, fCO2 was found to be the highest at 20oC (190 g/kg DM), followed by 35oC 110 (130 g/kg DM), whereas lower fCO2 values were observed at 5 and 50oC, being 70 and 25 g/kg DM, respectively. These results may again be attributed to the dominance by biological rather than chemical oxidation mechanism. Figure 6.7. Gas emission profiles from DF greens at different temperatures under aerobic conditions (+: 5oC; ●: 20oC; □: 35oC; ▲: 50oC) 111 CO emission was 2-4 orders of magnitude lower than CO2 emission (Figure 6.7c); maximum fCO values were 0.28 g/kg DM and 0.01 g/kg DM at 50 and 20oC, respectively. As oxygen was consumed during the biological degradation process, the environment in the reactor was not favorable for CO generation. Again, CH4 was not detected from the aerobic reactors. The VOCs produced from DF greens under aerobic conditions were mainly terpenes, ketones, hydrocarbons, benzene and its derivatives. Acids, esters and alcohols were also found. Indole and some sulfur compounds were detected under the higher temperatures of 35 and 50oC. The emission from the 5oC reactor was much cleaner than that from the 50oC reactor, as fewer compounds were found. As a result of microbial process, terpene is the most popular component from stored DF greens. Othere researchers have found some of the above-mentioned VOCs including terpenes in the drying process of grass (Johansson & Salin, 2011) and composting of garden wastes (grass and leaves) (Eitzer, 1995; Komilis et al., 2004; Wilkins & Larsen, 1996). Figure 6.8. Cumulative concentration of TVOCs from aerobic reactors for DF greens 112 As seen in Figure 6.8, temperature is positively correlated with TVOC concentration no matter what mechanism might be dominant during the test. The initial concentration was less than 20 ppm at 5oC but as high as 670 ppm at 50oC. The decrease in daily increments of TVOC concentrations could be a result of the reduction in moisture content over the test period as well as limitation of degradable substrates. Cumulative TVOC concentration was 16,000 ppm for the 50oC reactor and 8,000 ppm for the 20oC reactor, respectively after two- month storage. 6.3.3 Characteristics of gas emissions from mixed Douglas fir wood chips and greens (Test Series #3) 6.3.3.1 Non-aerobic conditions Figure 6.9 illustrates the gas emissions from the non-aerobic reactors. At the early period of test, CO2 emission from the 20oC reactor was the highest. However, the emission factor pertinent to the 5oC reactor took over and became the highest after about 3 weeks. The values of fCO2 under all temperatures lied between the values for the chips and the greens, as expected. The peak CO2 emission factors, fCO2 were (14, 12, 8.3 and 5 g/kg DM), which correspond to CO2 concentration of (48, 42, 34 and 24%) for temperatures ranging from 5- 50oC. Again, the lower CO2 emission at higher temperature indicated that anaerobic biodegradation of the substrate could be the dominant process with no oxygen present in the reactors. In fact, oxygen was only detected at 5oC in the first few days, and it dropped quickly to zero after 8 days. There was no oxygen detected from the beginning of the test for the other three temperatures. CO and CH4 emissions were low to negligible. 113 Figure 6.9. Gas emissions from stored DF mixed chips and greens at different temperatures under non-aerobic conditions (+: 5oC; ●: 20oC; □: 35oC; ▲: 50oC) The main VOCs include branched, straight and cyclic chain hydrocarbons, terpenes, acids, esters, aldehydes, alcohol and benzene. Specifically, compounds such as α-pinene, ocimene, butane, methanol, methoxyacetic acid, heptyl ester, cyclopentane, cycloheptatriene, and cyclohexene were found. Indene and hydrazide (formic acid hydrazide) were also emitted from the reactors. Furanone was detected from the reactor under 50oC. At the end of the test, the TVOC concentrations from the reactors with mixed DF wood chips and greens were measured. The results showed that the concentrations at 20, 35 and 50oC were beyond 1000 ppm (the detection limit) due to the green parts in the substrate. Much lower TVOC concentration of 120 ppm was detected at 5oC, reaffirming that 114 temperature can induce the release of VOCs leading to high emission under higher temperature. 6.3.3.2 Aerobic conditions Figure 6.10a indicates that, at the beginning, the oxygen concentrations at all temperatures except 5oC were low as a result of fast degradation rates. Subsequently, oxygen concentrations increased, though at different rates during the test with respect to the various temperatures, ranging from 10-18% at the end of the test. CO2 and CO emission results were plotted in Figure 6.10 for the reactors with mixed DF chips and greens under aerobic conditions. The CO2 emission factor was found to be 30 g/kg DM at 50oC versus 150 g/kg DM at 20oC (Figure 6.10b). By comparison, the CO emission factor reached a value of 0.08 g/kg DM after 60 days of storage at 50oC (Figure 6.10c). Components of VOC emission from these mixed materials under aerobic conditions include alkanes, alkenes, alkynes, terpenes, benzene and its derivatives, acids, esters, ketones and so on. These are similar to the non-aerobic reactors. However, specific compounds were also found in aerobic reactors such as pentene, benzonitrile, butene, heptadiene, toluene, ketoprofen, benzoic acid and ethyl ester. 115 Figure 6.10. Gas emissions from DF mixed chips and greens at different temperatures under aerobic conditions (+: 5oC; ●: 20oC; □: 35oC; ▲: 50oC) 116 As shown in Figure 6.11, the daily increment of TVOC concentration diminished with time. This trend was similar to the other materials tested under aerobic conditions. Temperature was again in accord with the TVOC concentrations. The initial TVOC was measured to be 550 ppm at 50oC, while it was only 10 ppm at 5oC. In terms of cumulative TVOCs, the values were 11,000 ppm and 1,500 ppm at 50 and 20oC, respectively. Figure 6.11. Cumulative concentration of TVOC at various temperatures from aerobic reactors for mixed materials 6.3.4 Comparing the results from the three types of DF materials 6.3.4.1 Non-aerobic conditions Overall, the profiles of gas emissions from DF chips exhibited differences from the DF greens, but gas emissions from mixed chips and greens have similar trends as those from the greens. 117 The trends of CO2 profiles were different for wood chips versus greens. Greens are more readily degradable materials than wood chips; hence it is reasonable that CO2 emission factor for the greens was much higher than those for the wood chips by 8 folds. The emission factors for the mixed materials are in between, as expected, though the mixed materials and the chips differ by several times. Thus, during storage, emission of CO2 from the reactor increased significantly in the presence of greens which played a critical role. Chemical oxidation was suggested to be the dominant mechanism for CO2 emission from stored DF chips, while biological degradation could be the main mechanism in stored DF greens and mixed materials. As oxygen content was depleted to zero, the gas emission rate became very slow and CO2 increased little for wood chips. In contrast, the greens and mixed materials went through the anaerobic degradation process without oxygen, and produced CO2 emissions rather than CH4. Due to their greater nitrogen content, greens have lower C:N ratio (around 45) as compared to wood chips (around 250), which favors the living of microbes. Very low levels of CO emission were observed for all three types of materials. In general, CO is produced by the chemical oxidation of organic matters such as lipids and fatty acids, and it is promoted by storage temperature and availability of oxygen. Apparently, these factors were not significant in this experiment. Very low levels of CH4 were also observed for all three types of materials. The main VOC components detected from the three types of materials are similar, including hydrocarbons, terpenes, ketones, aldehydes, benzene and its derivatives, acids and esters. Some of the VOCs were also reported by other researchers who worked with yard wastes, for instance, (Sinicio et al., 1995). In this study, odorous compounds such as pyridine, ammonia, sulfur compounds, indole and furanone were associated with the storage of DF greens and mixed materials under higher temperatures. As green materials contain greater nitrogen content for microorganisms to assimilate, more VOCs were generated during biodegradation. Thus, the total concentrations of VOCs (TVOC) were much higher for the greens and mixed materials as compared to wood chips. 118 Microbial analysis Results in terms of total microbial counts (bacteria and molds) for all three types of materials are listed in Tables 6.1a and 6.1b. These results suggested that temperature is a key factor affecting the activity of microbes. For all materials, the counts for the 50oC reactor were consistently much lower than those pertinent to the other temperatures, whereas a higher level of microbial activities was observed at 20oC. Microbes living on biomass have a maximum temperature; the rate of microbial growth may become progressively slow and sometimes cease at temperatures close to the maximum (Ayerst, 1969). Apparently, the microbes could not withstand a temperature as high as 50oC. In addition, the very low counts of microbes for the wood chips amidst a high emission of CO2 at 50oC is suggesting that CO2 generation could be attributed to chemical oxidation mechanism. Microbial respiration might also contribute to CO2 emissions from the chips at 5, 20 and 35oC. The total bacterial counts (TBC) for the DF wood chips were higher than the greens and the mixed materials have intermediate TBCs. This might be due to the inhibition of microbial activity by the high concentration of CO2 in the reactors with the greens while there was little to no oxygen. For the mixed materials, the trend of microbial counts is consistent with the emission of CO2. Table 6.1a. Total bacterial counts of the Douglas fir samples (cfu/g sample) Test #1 (DF wood chips) Test #2 (DF greens) Test #3 (DF mixtures) Temperature, oC Non- aerobic Aerobic Non- aerobic Aerobic Non- aerobic Aerobic 5 20000 25000 110 80 13000 120000 20 23000 30000 40 130000 8500 200000 35 18000 12000 20 130000 160 180000 50 50 3800 10 10 50 15000 119 Table 6.1b. Mold counts of samples (cfu/g sample) Test #1 (DF wood chips) Test #2 (DF greens) Test #3 (DF mixtures) Temperature, oC Non- aerobic Aerobic Non- aerobic Aerobic Non- aerobic Aerobic 20 < 5 2600 <5 6×106 120 1×105 35 < 5 nm <5 4×106 < 5 20 Dry matter losses For Douglas fir chips, the reduction in dry matter after the storage test was presented in Table 6.2. Dry matter losses were highest at 50oC, which corresponds to the gas emissions results in Figure 6.3. Higher gas emissions at higher temperatures implied that chemical oxidation mechanism was dominant. Dry matter losses of 0.24% as measured for 5oC was much lower than those for the reactors at other temperatures. As the total bacteria counts for 5, 20 and 35oC are relatively high compared to that at 50oC, biological process may also contribute to the gas emissions. For the DF greens, temperature had a negative relationship with respect to the dry matter losses. The trend of dry matter losses is compatible with the trend of emission factors. The highest loss was 3.2% under 5oC, while under 50oC the dry matter loss was 2.5%. Hence, the dry matter losses from the DF greens were 3-10 times greater than those from the wood chips, depending on the temperature. Results of dry matter losses from mixed materials under non-aerobic conditions (Table 6.2) and results of gas emissions are in good agreement. Highest dry matter losses were found to be 2.2% at 5oC. Considerable losses of dry matter along with no oxygen detected indicated that the generation of gases was due to the biological degradation rather than chemical oxidation and microbial respiration. Again, these relatively high losses might be largely due to the greens in the samples. This corresponds with the gas emission results, and reaffirms that green materials are much easier to degrade than wood chips. With data for WRC chips from Chapter 5 included, the correlation between dry matter losses and total CO2 emission under non-aerobic conditions is then shown in Figure 120 6.12. The dry matter losses have a positive correlation with CO2 emissions. This reaffirms that gas emission is an important factor leading to dry matter losses. The effect of temperature on the dry matter losses for DF under non-aerobic condition is depicted in Figure 6.13. The correlations between dry matter losses and temperature are different for different DF samples. With an increase in temperature, dry matter losses from DF chips increased. But dry matter losses have a negative correlation with temperature for DF greens. DF mixed materials had similar behaviour as the greens. The degradation of biomass is a major factor causing the dry matter losses. For greens and mixed materials, biological mechanism is dominant during the storage process. Higher temperature might inhibit the biological activity and slow down the degradation process, thus exhibiting slightly less dry matter losses. The relationships between temperature and dry matter losses for different DF samples are expressed by DF chips 0.014 0.18DMloss T= + (6.1) DF greens 0.016 3.35DMloss T= − + (6.2) DF mixed materials 0.0044 2.21DMloss T= − + (6.3) Table 6.2. Dry matter losses from each test (%) Test #1 (DF wood chips) Test #2 (DF greens) Test #3 (DF mixtures) Temperature, oC Non- aerobic Aerobic Non- aerobic Aerobic Non- aerobic Aerobic 5 0.24 0.97 3.20 8.03 2.17 5.26 20 0.47 1.92 3.14 15.6 2.15 11.1 35 0.68 2.09 2.76 14.7 2.06 8.71 50 0.87 2.36 2.52 8.95 1.98 6.79 121 Figure 6.12. Correlation between dry matter losses and total CO2 emission of WRC and DF materials at all temperatures under non-aerobic conditions Figure 6.13. Correlation between temperature and dry matter losses from DF materials under non-aerobic conditions 122 6.3.4.2 Aerobic conditions Obviously, with mixed materials, the trends of gas emission under all temperatures were similar to the trends of the greens. In fact, the measured gases and their general trends are the same for all three types of materials. Again, CO2 emissions were the highest for the greens, followed by the mixed materials. The components of VOCs are similar for all three materials, and the trends of total VOC emissions are the same despite different concentrations. Microbial analysis The microbial counts from aerobic reactors with DF wood chips under all the temperatures are also shown in Tables 6.1a and 6.1b. TBCs for reactors at 5, 20 and 35oC (12,000-30,000 cfu/g sample were relatively high when compared to that at 50oC (3,800 cfu/g sample). These results seem to support the argument that chemical oxidation was dominant for CO2 production during the storage of DF wood chips. For the greens, the results of microbial analysis can also be found in Table 6.1. TBCs were high (around 1.3×105 cfu/g sample) and close to each other at 20 and 35oC. In contrast, the values obtained at 5 and 50oC were very low, being 80 and 10 cfu/g sample respectively, implying that 20-35oC may be the most favorable temperatures for microbes living on the greens. The mold counts detected from the green samples were dramatically higher at 6×106 and 4×106 cfu/g of sample at 20 and 35oC respectively, as compared to the wood chips, indicating that the abundant nutrients in the greens favor the growth of mold. Furthermore, when compared to the extremely low mold counts (<5 cfu/g of sample) for samples stored under non-aerobic conditions, it means molds tend to thrive under aerobic condition rather than in an environment with limited oxygen. This has been proven by the other researches (McGinnis, 2007; Reed et al., 2007; Shi et al., 2012; Zhang et al., 2010). Overall, the microbial analysis results are in accord with gas emissions and they support the argument of dominance by biological mechanism. The results of microbial counts for mixed materials also support the argument that CO2 emission was dominated by biological degradation mechanism. As the highest CO2 emission was observed at 20oC, the microbial counts (bacteria plus molds) were also highest 123 at 2×105 cfu/g sample. By comparison, the relatively low microbial counts of 1.5×104 cfu/g sample are compatible with the lowest emission of CO2 at 50oC. Dry matter losses For the wood chips under aerobic conditions, the dry matter losses increased from 0.97% to 2.36% as temperature ranged from 5oC to 50oC (Table 6.2). The greater dry matter losses versus the non-aerobic reactors may be due to the daily air pumping which may remove some dry matter from the chips. The dry matter losses from the reactors with DF greens under aerobic conditions were substantially greater compared to those under non-aerobic conditions as a result of the daily ventilation to replenish oxygen in the reactor. The highest dry matter loss was 15.6% under 20oC after two-month storage, which is in parallel with the gas emission results in Figure 6.7. Whereas, dry matter losses under 35 and 50oC were 14.7% and 8.95% respectively, and the lowest loss was 8% at 5oC. Considerable losses of dry matter indicated the generation of gases was due to biological degradation rather than microbial respiration. Although MacGregor et al. suggested that the maximum composting activity may be achieved under thermophilic conditions of 50-60°C (MacGregor et al., 1981), Rao et al. and Vikman et al. reported faster biodegradation rate of organic matter, and increased rate of O2 uptake and rate of mineralization of carbon to CO2 within the mesophilic temperature range of 35-43oC (Rao et al., 1996; Vikman et al., 2002). Ananda et al. did a comprehensive review of green waste composting under mesophilic and thermophilic conditions (Anand et al., 2012). They concluded that aerobic mesophilic condition is most favorable for the growth and development of cellulolytic microbes and increasing their cellulose activity required for lignocellulosic degradation. It is also the best in terms of the quality of finished compost. Dry matter losses from the greens under aerobic conditions are significantly higher than the wood chips. The losses from mixed materials were lower than greens, but they are still 3-5 times greater than the wood chips. The greens with more nutrients than the wood chips are more readily degradable. Results from composting of leaves with 3 mm particle size showed that around 30% dry matter was lost after 6 weeks in the process (Satin, 2011). 124 For the mixed materials, dry matter losses in the reactors at temperatures ranging from 5 to 50oC amounted to 5.3%, 11.1%, 8.7% and 6.8%, respectively. Again, with data for WRC chips from Chapter 5 included, the results show that dry matter losses have a positive correlation with gas emissions, as demonstrated in Figure 6.14. The dry matter losses as a function of temperature for DF materials under aerobic condition are shown in Figure 6.15. The dry matter losses exhibited exponential relationship with temperature for DF chips; while the curves for greens and mixed materials displayed similar polynomial shape. The equations used to describe the relation between temperature and dry matter losses for chips, greens and mixed materials are DF chips ( )1.96 exp 0.069 2.36DMloss T= − ⋅ − + (6.4) DF greens 20.0148 0.829 4.409DMloss T T= − + + (6.5) DF mixed materials 20.00973 0.557 3.005DMloss T T= − + + (6.6) 125 Figure 6.14. Correlation between dry matter losses and total CO2 emission from WRC and DF materials at all temperatures under aerobic conditions Figure 6.15. Dry matter losses as a function of temperature for DF materials under aerobic conditions 126 As an extension of the research results, data from tests using DF wood chips under both non-aerobic and aerobic conditions were compiled to estimate the number of days, d, to reach 1% loss in dry matter, using the IBSAL approach (Sokhansanj et al., 2003): ( )ln d a bT= + (6.7) where d is the number of storage days; T is the temperature (oC); a and b are constants. Table 6.3 lists the estimated a and b values along with the coefficient of determination R2. Table 6.3. Estimated constants a and b for DF wood chips in Eq (6.7) a b R2 Non-aerobic 5.36 -0.028 0.95 Aerobic 3.90 -0.018 0.79 Eqn (6.7) is graphically represented by Figure 6.16 temperatures ranging from 0 to 50oC. Under non-aerobic conditions, the wood chips hardly degraded and they would take a very long time to lose even just 1% dry matter. Figure 6.16. The number of days before DF wood chip loses 1% dry matter for temperatures from 0 to 50oC 127 6.3.4.3 Comparison of changes in visual appearance of materials The changes in visual appearance of the tested materials during storage are presented in Figure 6.17. It shows that wood chips were not perceptible; there was no color change observed before and after storage. For non-aerobic conditions, green materials became brown only after one day of storage under 35 and 50oC. Materials started to turn brown after five days at 20oC; whereas at the end of the storage, there were still some greens detected from the reactor at 5oC. In contrast, the materials changed to dark brown from day 33 at 50oC, while the others stayed as brown. The greens experienced similar color changes under both aerobic and non-aerobic conditions. However, more greens remained at 5oC under aerobic condition. Molds were clearly visible in the reactors under 20 and 35oC. Mixed materials appear to behave as wood chips and greens in combination. The green parts went through similar color changes as those from Test Series #2 (for greens only), while the wood chips did not have any visible changes as in Test Series #1 (for chips only). 128 129 Figure 6.17. The appearance of the materials in three series tests (DF chips, greens and mixed chips and greens) before and after the storage 130 6.3.5 Comparison of gas emissions between stored Douglas fir and Western Red Cedar wood chips 6.3.5.1 Non-aerobic conditions The results from stored Douglas fir chips are compared to Western Red Cedar chips. The dominant mechanisms of gas generation are considered to be different between these two materials. The CO2 emission was positively related to temperature (5, 20, 35 and 50oC) for DF chips, whereas the highest CO2 emission was measured at 20oC for WRC chips. However, the highest CO2 emission factors were both around 3 g/kg DM. CO emissions were much higher from the stored WRC chips than the DF chips, which might be attributed to the different oxygen contents in the reactors. CH4 emissions were negligible for both types of wood chips. The major VOCs from these two types of chips are similar, consisting of hydrocarbon, methanol, aldehydes, terpene, acid, ketone, esters and aromatic compounds. A comparison of VOCs from qualitative GC/MS analysis for these two types of woody chips is listed in Table 6.4. The total concentrations of VOCs from the WRC chips are approximately 3 times lower than those from the DF chips especially under higher temperatures. The total bacterial counts from WRC chips were substantially lower; for instance TBC was 140 cfu/g sample as compared to 2.3×104 cfu/g sample for the DF chips under 20oC. Table 6.4. A comparison of VOCs identified by qualitative GC/MS analysis WRC chips DF chips alkanes, alkenes, alkynes, aldehydes, acid, benzene and its derivatives, methanol, terpene, ketone, ethers and esters; furan, indole alkanes (hexane), alkenes, aldehydes, acids, benzene and its derivatives, terpenes (α- pinene), ketones (acetone), ethers, esters; nitrogen compounds The dry matter losses from stored DF chips ranged from 0.24% to 0.87%, as compared to 0.058% to 0.26% for WRC chips, suggesting that DF chips are somewhat more 131 degradable than WRC chips. It reaffirms the decay-resistance characteristics of WRC, which is also reflected in the low microbial counts mentioned above. 6.3.5.2 Aerobic conditions Gas emissions exhibit the same trends for both the WRC and DF chips. At 20oC, the highest CO2 emission factor fCO2 for WRC chips was 6.6 g/kg DM, as compared to 10.1 g/kg DM for DF chips after 50 days storage. At 50oC, the highest fCO2 was 17.7 g/kg DM for the DF chips. With the same initial moisture content, DF chips are more susceptible and degradable than WRC chips in the presence of oxygen. The initial total concentrations of VOCs from DF chips are almost 10 times higher than WRC chips, especially under high temperatures. The trends were similar for both types of materials. At the end of the test, TVOCs from DF are much higher than WRC chips. The dry matter losses from DF chips were about 4 times greater than those derived from WRC chips under aerobic conditions. It indicates the DF chips can be degraded more readily with the existence of oxygen. 6.3.6 Comparison of gas emissions with wood pellets Gas emissions from stored wood pellets (3.7-10% moisture content, wet basis) under oxygen-depleting environment (Kuang et al., 2009) were compared with Douglas fir chips. The peak emission factor of CO2 from wood pellets increased from 0.025 g/kg DM to 0.23 g/kg DM as temperature was raised from 10 to 45oC after 30 days storage, while the corresponding peak CO emission factor changed from 0.001 to 0.058 g/kg DM. The chemical auto-oxidation process was suggested to be the dominant mechanism for the gas emissions from wood pellets. Table 6.5 lists gas emission values for non-aerobic storage of wood chips and pellets after 30 days storage. The gas emissions from regular white pellets with moisture contents of 4, 9, 15, 35 and 50% (wb) were measured at 25oC, 40oC, and 60oC during a 9-week storage period (Yazdanpanah F., 2012). Results showed the CO2 emission factors, which ranged from 132 0.007-0.415 g/kg DM, had a positive correlation with temperature as well as moisture content. CO emission was also found to be positively related to temperature. The same mechanism was found in stored DF wood chips. As temperature rises, CO2 emissions increase. The highest fCO2 from DF chips under 50oC was almost 10 times larger than those derived from wood pellets with moisture contents ranging from 4% to 50% (wb), whereas fCO was lower than those from wood pellets. The large differences in CO2 emissions may be attributed to the higher moisture content of the fresh chips (50% wet mass basis) with live microbes, whereas most of the microbes should have been killed during pelletization and the pre-pelletization drying process at high temperatures. Moisture content is a key factor affecting the reaction rate. When the initial O2 content was high in the reactors, it is possible for the fresh wood chips to release a greater amount of CO2 than the wood pellets. By comparison, alkanes, hexanal, acetone and benzene were detected from wood pellets storage (Arshadi & Gref, 2005; Svedberg et al., 2004). Table 6.5. Comparison of gas emissions from wood pellets and chips under non-aerobic condition after 30 days storage Wood pellets (5.1% w.b.) (g/kg DM) (Kuang et al., 2009) WRC chips (50% w.b.) (g/kg DM) DF chips (50% w.b.) (g/kg DM) Temperature (oC) CO2 CO Temperature (oC) CO2 CO Temperature (oC) CO2 CO 10 0.028 0.001 5 0.63 - 5 1.21 - 23 0.037 0.0074 20 2.8 - 20 2.78 - 35 0.15 0.036 35 2.12 0.0015 35 2.95 0.00012 45 0.24 0.058 50 0.25 0.018 50 3.02 0.00047 6.3.7 Comparison of gas emissions with other woody materials The levels of oxygen and toxic gases from logs and wood chips stored in confined spaces during sea transportation were studied (Svedberg et al., 2009). The materials had been on 133 board for 37-66 hours. The concentration of oxygen in the cargo was 10% in average; however, it was depleted to 0% in some places. The average concentrations of CO2 and CO were 7.5% and 46 ppm, respectively. The major emission of the hydrocarbons from stored materials was monoterpenes (α-pinene). It was also observed that oxygen level was higher during the cold season. A high level of microorganisms present in wood chips versus wood pellets was found by others. (Madsen et al., 2004) reported that high bacterial activities were found in dusts from straw and wood chips (8×104 - 3.1×106 cfu/mg dust), while lower bacterial counts were measured in dusts from briquettes and wood pellets (between 20 and 60 cfu/mg dust). 6.4 Conclusion Gas emissions were studied using three test series involving stored Douglas fir chips, greens and mixed materials (chips plus greens). Experiments were conducted using lab-scale reactors for a range of temperatures under both aerobic and non-aerobic conditions, which would correspond to the environment within a pile depending on the availability of oxygen. Results from the tests showed that CO and CH4 emissions were 2-5 orders of magnitude less than CO2 emission for Douglas fir residues. The relationship between CO2 emissions and temperature differs with the stored materials and depends on the mechanism - chemical oxidation versus biological (biodegradation or microbial respiration). Under non-aerobic conditions, the CO2 emissions increased with time for all temperatures. At higher temperatures, higher CO2 emissions were measured from stored chips which suggested that chemical reaction was in dominance. In contrast, lower emissions of CO2 at higher temperatures were observed from the greens and the mixed materials. The CO2 emissions from stored DF greens were found to be 8 times higher than those produced from wood chips. The significantly higher emissions of CO2 together with the eventual lack of oxygen from stored DF greens indicated that the storage environment has turned anaerobic and the materials went through the anaerobic degradation. It reaffirms that the green materials are easier to be degraded than wood chips due to the higher content of nutrients. 134 The amount of gas emissions from mixed materials is generally in between the values obtained from wood chips and greens. Results also showed that CO2 emissions from the aerobic reactors exhibit similar trends as the non-aerobic reactors with respect to the effect of temperature. However, the total gas emissions are higher from the aerobic reactors over the same storage period. The major VOCs components detected from the three types of materials by GC/MS are similar, including hydrocarbons, terpenes, ketone, aldehyde, acetone, methanol, benzene and its derivatives, acid and esters. However, some unique VOCs from each material are odorous or have potential impact on human health. These include pyridine, ammonia, sulfur compounds, indole and furanone found from stored greens and mixed materials under higher temperatures. The total concentration of VOCs (TVOC) was found to have a positive correlation with temperature for all materials under both aerobic and non-aerobic storage conditions. The TVOCs are much higher from greens and mixed materials than wood chips, which may also be attributed to the readily degradation of green materials. Under aerobic conditions, the cumulative TVOC concentrations were greater versus the non-aerobic reactors over the storage period. Microbial analysis results in terms of total bacterial counts and mold counts are compatible with the CO2 emission results. Different extents of dry matter losses from the three materials were found. Reactors with greens were measured to have the largest dry mass losses, which are generally 4 times more than those for wood chips. The percent dry matter losses from the aerobic reactors are greater than those from the non-aerobic reactors. When the total CO2 emissions together with the total TVOC emissions over the entire storage period were taken into account, a positive correlation was obtained between the percent dry matter losses and gas emission under both aerobic and non-aerobic conditions. In conclusion, Douglas fir materials may be degraded to a certain extent during the storage as a result of the chemical and/or biological processes. Green materials can be much more readily degraded than wood chips; thus the presence of greens exerts significant effect 135 on gas emissions. Comparison of Douglas fir with Western Red Cedar and other woody materials shows that Douglas fir is a more readily degradable material than Western Red Cedar. Moreover, much more gas emissions are produced from Douglas fir chips than wood pellets as a result of higher moisture content. Overall, the gas emission results from this study can help to understand the production and evolution of gases from stored wet biomass. It gives the concept regarding what types of gases would be produced and how they behave. It provides a background to better manage and handle the storage of high-moisture biomass. This work reaffirms the importance of work safety and eco-friendly storage conditions for wet woody biomass and controlling the emission of odorous VOCs. 136 Chapter 7. Conclusions and Recommendations 7.1 Conclusions The overall goal of this thesis was to study the biomass moisture content and gas emissions during storage. The physical and chemical properties of the biomass may change due to moisture variation and gas emissions. The specific objectives were: 1) to investigate the drying characteristics of woody biomass; 2) to describe the wetting and drying processes of woody biomass and adopt a model to simulate the time-dependent moisture content of the materials during storage; 3) to apply the model to a biomass pile stored in the field under natural weather conditions; 4) to study the gas emissions from different kinds of woody biomass under different storage conditions and to quantify the emitted gases. In chapter 1, recent published research on biomass storage was reviewed. The problems associated with storage, which include variation of moisture content, gas emissions and dry matter losses were identified. A comprehensive literature review about moisture sorption characteristics of biomass was also included in this chapter, along with the equilibrium moisture content and moisture sorption isotherms. Literature review also covered gas emissions from different stored biomass. The effects of temperature, moisture content, storage time and microbial activities were discussed. Most studies have been carried out on wood pellets and woody materials with low moisture content. Very few works were conducted on the fresh woody biomass with high moisture content, especially the gas emissions from different high moisture materials. The sorption characteristics of Aspen materials with the details of drying and adsorption processes were presented in Chapter 2. The effects of temperature and relative humidity on the sorption characteristics and the drying rate of woody biomass were studied. The moisture sorption isotherm was obtained based on the experimental data. Experiments using Trembling Aspen (Populus tremuloides) as materials were conducted in a controlled environment chamber. Results showed that low temperature and high relative humidity of ambient air led to higher equilibrium moisture content (EMC) for both desorption and adsorption processes. At higher temperature, the EMC was reached over a shorter drying 137 time indicating a higher sorption rate; relative humidity was also positively correlated with the adsorption rate. The Modified Oswin model that relates equilibrium relative humidity to temperature and EMC was found to provide the best fit to the experimental data for both desorption and adsorption processes. The adsorption and desorption curves displayed a sigmoidal shape and the curves exhibited hysteresis effect between adsorption and desorption. The trend of drying rate constant versus temperature followed the Arrhenius equation, and Page’s model was appropriate for predicting the drying characteristics of Aspen. In Chapter 3, a mathematical model was adopted. The model was further developed and calibrated for simulating the wetting and drying processes of Aspen. The changes of moisture content during storage were studied. The moisture of the materials was divided into internal and external moisture contents, which represent the bound water and free water associated with the wood, respectively. The biomass exchanges moisture with the surrounding environment by evaporation and precipitation. The internal moisture equation originated from the Lewis’ equation, while the change in external moisture content was related to the differences between evaporation and precipitation. The Aspen materials were subject to periods of artificial precipitation and constant temperature evaporation in the laboratory. Time-dependent moisture data along with pan evaporation rates were used to estimate the parameters of the equations. The three coefficients of the model were estimated to be 0.206 mm-1, 0.129 mm-1 and 0.239 mm-1, with small standard errors of estimation. The model was then applied to the Aspen bales stored in the field during one-year storage under natural conditions in Chapter 4. With the inputs of available weather data which included temperature, relative humidity, wind speed, solar radiation and precipitation, the model was used to estimate the daily evaporation and hence the moisture content of Aspen under both covered and uncovered conditions during the one-year storage period (June 2011-May 2012). The uncovered bales were exposed to the natural weather conditions while the covered bales were protected from precipitation and solar radiation. For the uncovered bales, the predicted moisture contents fluctuated around 20% (dry basis) in the summer time. The high moisture contents of materials estimated during the November- March period were in line with the low temperature and high precipitation conditions. The predicted moisture content ranged from 15% to 80% (dry basis). By comparison, for the 138 covered bales, higher moisture contents as predicted for the period extending from October to March resulted from the low evaporation rates which corresponded to high relative humidity conditions. The predicted moisture contents were approximately 14% different from the actual values. Nevertheless, the measured moisture contents exhibited the same trend as the predicted moisture contents for both situations. The predicted moisture contents and the profiles were in reasonably good agreement with the measured in-field results. This indicated that the lumped model presented in chapter 3 may be used as a first approximation, and applied to estimate the moisture content of Aspen or similar biomass during relatively long-term field storage with a reasonable degree of accuracy. Chapters 5 and 6 presented the results of gas emissions from the storage of two important types of woody biomass in British Columbia under different storage conditions with respect to temperature, initial moisture content and oxygen availability. Experiments were conducted using lab-scale reactors for a range of temperatures under both aerobic and non-aerobic conditions, which would correspond to the environment within a pile of woodchips depending on the availability of oxygen. In chapter 5, the gas emissions from stored Western Red Cedar chips were studied. Results from WRC chips with initial moisture content of 50% wet basis in non-aerobic reactors showed that the highest CO2 emission factor of 2.8 g/kg DM (equivalent to 16% in concentration) was observed at 20oC along with the lowest O2 concentration of zero while temperature ranged from 5 to 50oC, suggesting biological reaction could be the dominant mechanism for CO2 generation during the storage period. Gas emissions from sterilized woodchips confirmed this phenomenon. Although the CO emission factor was much lower at 0.03 g/kg DM, it increased with increasing temperatures due to chemical oxidation. CO2 and CO emissions from the aerobic reactors exhibited similar trends as the non-aerobic reactors with respect to the effect of temperature. The cumulative emissions of both CO2 and CO at each temperature increased slowly with time, with higher oxygen levels of close to 20% being maintained during the entire test period. Over a storage period of 55 days, the total CO2 emissions from the non-aerobic and aerobic reactors were 2.8 g/kg DM and 6.6 g/kg DM, respectively for the 20oC temperature. Similarly, the total CO2 emission was greater for 139 the aerobic reactors at other temperatures. Microbial analysis results in terms of total bacteria counts supported the argument of dominance by biological mechanism for CO2 emission. Results from the qualitative GC/MS analysis indicated that the volatile organic compounds (VOCs) emitted from the stored WRC woodchips included benzene and its derivatives, methanol, terpene, aldehydes, acid, alkane, indole, furan, acetone, ethers and esters. Some of these VOCs may be associated with the characteristics pungent smell of WRC which can cause odor nuisance to the neighboring community. The total VOC concentration was found to have a positive correlation with temperature. Percent dry matter losses were less than 0.3% for the non-aerobic reactors as compared to 0.15-0.57% for the aerobic reactors under different temperatures; the differences between aerobic and non- aerobic reactors were compatible with the estimated total gas emission over the storage period. The accumulated concentrations of TVOC emitted from the non-aerobic reactors were lower than the summation of the daily TVOC concentrations from the aerobic reactors over the storage period. Taking the total CO2 emissions together with the total TVOC emissions over the entire storage period, a positive correlation between the percent dry matter losses and gas emission was found. The plots of gas emission data from the materials with lower initial moisture content of 35% wet basis exhibited the same trends as those from higher moisture content under both aerobic and non-aerobic conditions. However, the amounts of gases were found to be lower to different extents, indicating the moisture content is another important factor affecting the gas emission from stored biomass. The moisture content was found to have a positive correlation with the microbial activities, the chemical reaction rate and the evolution of gas emissions. In chapter 6, the gas emissions were studied from three tests involving stored Douglas fir chips, greens and mixed chips and greens (with ratio of 1 chip:1 green in dry mass). The initial moisture contents of these materials were all around 50% wet basis. Results showed that the test with green materials more likely experienced biologically dominant degradation rather than chemical oxidation. Under non-aerobic conditions, higher emissions of CO2, CO and CH4 were measured at higher temperatures from stored chips which suggested the 140 chemical mechanism was in dominance. By comparison, the emissions of CO2 were lower at higher temperatures from DF greens and mixed chips and greens. The emissions of CO2 from stored DF greens were as much as 8 times higher than DF wood chips. The amount of gas emissions from mixed greens and chips was generally between wood chips and greens. The oxygen was not detected from stored DF greens and mixed materials under non-aerobic conditions. The outstanding high emissions of CO2 along with no oxygen from stored DF greens indicated that the biomass had experienced the anaerobic degradation during storage period. This performed that the green materials are much easier to be degraded than wood chips as a result of the higher content of nutrients. Further results from the aerobic reactors showed that the emissions of CO2 and CO exhibited similar trends as the non-aerobic reactors with respect to the effect of temperature. However, the total gas emissions from the aerobic reactors were calculated to be higher than those from non-aerobic reactors after the entire storage period. Microbial analysis results in terms of total bacterial counts and mold counts were compatible with the CO2 emission results. The major VOCs components detected from DF chips, greens and mixed materials by GC/MS were similar, which included hydrocarbons, terpenes, ketone, aldehyde, acetone, methanol, benzene and its derivatives, acid and esters. In addition, some odorous compounds such as pyridine, ammonia, sulfur compounds, indole and furanone were found from stored DF greens and mixed materials under high temperatures. The total concentration of VOCs was found to be positively related to temperature from all three materials under both aerobic and non-aerobic storage conditions. The TVOCs were much higher from greens and mixed materials comparing to wood chips. This might be attributed to the easily degradation of green materials. Results showed that the summations of the daily TVOC concentrations from the aerobic reactors were higher than the accumulated concentrations of TVOC emitted from the non-aerobic reactors over the storage period. The dry matter losses during storage were measured from three types of materials. DF greens had the largest dry mass losses, which were almost 4 times higher than those from wood chips. The dry matter losses from aerobic reactors were larger than those from non-aerobic reactors. Taking account of the total CO2 emissions along with the total TVOC emissions over the entire storage period, a positive correlation between the percent dry matter losses and gas emission was found. 141 Results indicated that Douglas fir materials would be degraded during the storage process as a result of the chemical and biological processes. Greens were more readily degradable materials comparing to wood chips. When comparing to Western Red Cedar chips, higher gas emissions and larger dry matter losses were measured from DF chips. It showed Douglas fir is a more easily degradable material than Western Red Cedar. Small degree of dry matter losses from WRC chips over storage period also reaffirmed the decay- resistance properties of WRC. Overall, this study was focused on the storage and natural drying aspects of the supply chain of biomass intended for bioenergy production. A model was adopted for predicting the moisture content of biomass stored in the field, and lab-scale tests were conducted under controlled environment conditions to calibrate the model, using artificial and drying wetting cycles. The model was verified using one-year in-field data. This model would be useful as a first approximation tool for predicting moisture content of biomass during storage. Another series of lab-scale tests were conducted under controlled environment conditions (aerobic and non-aerobic) to quantify the gas emissions from high- moisture biomass. Results showed there was a good correlation between dry matter losses and gas emissions. The measurement techniques used in this study are applicable for estimating gas emissions. A correlation between CO2 emission and dry matter losses was established. The number of days required for the biomass to reach 1% dry matter loss was also determined. This can help to devise ways to prevent excessive losses of dry matter during storage, which have economic implications. Information may be derived from the results regarding work safety associated with the storage of high-moisture biomass in confined space. 142 7.2 Recommendations for future research Several studies are recommended in the future. 1) This thesis did not consider self-heating of biomass during storage, but this is an area that can be explored in future. 2) It would be helpful to take account of the leachate from precipitation tests of biomass in lab-scale experiment. It is useful to collect and test the contents of the leachate. 3) A lumped model was used to describe the moisture variation in a pile. The average moisture content of the pile was estimated. In order to predict moisture stratification in the pile, the phenomenon of moisture diffusion shall be investigated. In order to get more accurate results, extra parameters should be considered to optimize the model in the future study. 4) The moisture content of the bale under the transparent cover during one-year storage in the field has been estimated in this thesis. It would be worthwhile to conduct a field test with covered bales during one-year storage to validate the predicted moisture content. 5) The aerobic and non-aerobic tests in the lab provided a concept of gas evolution within a pile of biomass. The measurement of gas emissions from a large-scale pile is recommended. 6) Moisture content was found to be an important factor affecting gas emissions. Initial moisture contents of 35% and 50% wet basis in Western Red Cedar chips have been studied. Various initial moisture contents can be investigated to study the specific relations between moisture content and gas emissions. 7) Green materials have a major effect on gas emissions from stored Douglas fir materials. The greens-to-chips ratio of 1 (dry weight) has been used as the basis of the study. It would be worthwhile to consider different ratios. The study on the relation between gas emissions and percent of greens is recommended. 143 8) The VOCs from stored materials have only been qualitatively studied due to limitation of the laboratory instrumentation. It would be worthwhile to quantify those VOC components. It would give a better understanding how the temperature affects the concentration of different VOC components. 9) Dry matter losses were found not to be totally accounted for by gas emissions. Some dry mass might be lost as it was dissolved in the leachate. It is recommended to measure the dry matter in the leachate from the stored materials. 144 References ACGIH. 2004. TLVs and BEIs: Based on the Documentation of the Threshold Limit Values for Chemical Substances and Physical Agents & Biological Exposure Indices. American Conference of Governmental Industrial Hygienists. Acharjee, T.C., Coronella, C.J., Vasquez, V.R. 2011. Effect of thermal pretreatment on equilibrium moisture content of lignocellulosic biomass. Bioresource technology, 102(7), 4849-4854. Afzal, M., Bedane, A., Sokhansanj, S., Mahmood, W. 2010. Storage of comminuted and uncomminuted forest biomass and its effect on fuel quality. BioResources, 5(1), 55- 69. Agrios, G.N. 2004. Plant pathology. 5th ed. Academic Press, New York. Akpinar, E.K., Bicer, Y., Yildiz, C. 2003. Thin layer drying of red pepper. Journal of food engineering, 59(1), 99-104. American Renewables. 2012. Benefits of Biomass Energy, http://www.amrenewables.com/biomass-energy/biomass-energy-benefits.php. Anand, D., Veerakumar, V., Gabhane, J., William, S.P.M.P., Bhange, V., Vaidya, A., Patil, M., Bhattacharyya, J., Wate, S. 2012. Why and How Aerobic Mesophilic Composting is Effective? A Comprehensive Study on Aerobic and Anaerobic Composting of Green Waste under Mesophilic and Thermophilic Conditions. International Journal of Recent Trends in Science And Technology, 5(1), 9-15. Arabhosseini, A., Huisman, W., Müller, J. 2010. Modeling of the equilibrium moisture content (EMC) of Miscanthus (Miscanthus×giganteus). Biomass and Bioenergy, 34(4), 411-416. Arogba, S.S. 2001. Effect of temperature on the moisture sorption isotherm of a biscuit containing processed mango (Mangifera indica) kernel flour. Journal of food engineering, 48(2), 121-125. Arshadi, M., Geladi, P., Gref, R., Fjallstrom, P. 2009. Emission of Volatile Aldehydes and Ketones from Wood Pellets under Controlled Conditions. Annals of Occupational Hygiene, 53(8), 797-805. Arshadi, M., Gref, R. 2005. Emission of volatile organic compounds from softwood pellets during storage. Forest products journal, 55(12), 132-135. Arslan, N. 2006. The fitting of various models to water sorption isotherms of tea stored in a chamber under controlled temperature and humidity. Journal of Stored Products Research, 42(2), 112-135. ASABE. 2010a. Moisture measurement - forage (S358.2), American Society of Agricultural Engineers. St. Joseph, MI, USA. ASABE. 2010b. Moisture relationship of plant-based agricultural products (D245.6), American Society of Agricultural Engineers. St. Joseph, MI, USA. 145 ASABE. 2006. Thin-layer drying of agricultural crops (S448.1), American Society of Agricultural Engineers. St. Joseph, MI, USA. ASABE. 2010. Psychrometric data (D271.2), American Society of Agricultural Engineers. St. Joseph, MI, USA. Ayerst, G. 1969. The effects of moisture and temperature on growth and spore germination in some fungi. Journal of Stored Products Research, 5(2), 127-141. Back, E., Allen, L. 2000. Pitch control, wood resin and deresination. Tappi Press Atlanta, GA. Baker, C.G.J. 1997. Industrial drying of foods. Springer. Baronas, R., Ivanauskas, F., Juodeikienė, I., Kajalavicius, A. 2001. Modelling of moisture movement in wood during outdoor storage. Nonlinear Analysis: Modelling and Control, 6(2), 3-14. Barry, R.G., Chorley, R.J. 2009. Atmosphere, weather and climate. Taylor & Francis. Barton, G. 1984. Definition of biomass samples involving wood, Park and foliage. Biomass, 4(4), 311-314. Basunia, M., Abe, T. 2005. Adsorption isotherms of barley at low and high temperatures. Journal of food engineering, 66(1), 129-136. Basunia, M., Abe, T. 2001. Moisture desorption isotherms of medium-grain rough rice. Journal of Stored Products Research, 37(3), 205-219. Baxter, L. 2005. Biomass-coal co-combustion: opportunity for affordable renewable energy. Fuel, 84(10), 1295-1302. Beakler, B., Blankenhorn, P., Stover, L., Ray, C. 2005. Total organic compounds released from dehumidification drying of air-dried hardwood lumber. Forest Products Journal and Index, 55(2), 57-61. Biomass Energy Center. 2008-2011. http://www.biomassenergycentre.org.uk/portal/page?_pageid=75,59188&_dad=portal. Bjoerklund, L. 1983. Storage of whole-tree chips of different species and in different fractions. Rapport-Sveriges Lantbruksuniversitet, Institutionen foer Virkeslaera, 143, 53. Boddy, L. 1983. Carbon dioxide release from decomposing wood: effect of water content and temperature. Soil Biology and Biochemistry, 15(5), 501-510. Boquet, R., Chirife, J., Iglesias, H. 1978a. Equations for fitting water sorption isotherms of foods. International Journal of Food Science & Technology, 13(4), 319-327. Boquet, R., Chirife, J., Iglesias, H.A. 1978b. Equations for fitting water sorption isotherms of foods. International Journal of Food Science & Technology, 13(4), 319-327. Bousquet, D. 2000. Lumber Drying: An Overview of Current Processes. University of Vermont Extension. Brooker, D.B., Bakker-Arkema, F.W., Hall, C.W. 1992. Drying and storage of grains and oilseeds. Springer. 146 Brooker, D.B., Bakker-Arkema, F.W., Hall, C.W. 1974. Drying cereal grains. AVI Pub. Co. CT. Brosseau, J., Heitz, M. 1994. Trace gas compound emissions from municipal landfill sanitary sites. Atmospheric environment, 28(2), 285-293. Bruce, D. 1985. Exposed-layer barley drying: three models fitted to new data up to 150 C. Journal of Agricultural Engineering Research, 32(4), 337-348. Cai, L. 2005. Determination of diffusion coefficients for sub-alpine fir. Wood Science and Technology, 39(2), 153-162. Cairelli, S., Ludwig, H., Whalen, J. 1994. Documentation for immediately dangerous to life or health concentrations (IDLHS), Springfield, VA: NTIS. Canada Health. 2004. Enumeration of Yeasts and Moulds in Foods (MFHPB-22), Vol. Volume 2, Health Products and Food Branch. Ottawa, Ontario. Canada Health. 2001. HPB methods of microbiological analysis of foods (MFHPB-18), Vol. Volume 2, Health Products and Food Branch. Ottawa, Ontario. Caputo, A., Palumbo, M., Pelagagge, P., Scacchia, F. 2005. Economics of biomass energy utilization in combustion and gasification plants: effects of logistic variables. Biomass and Bioenergy, 28(1), 35-51. Casal, M., Gil, M., Pevida, C., Rubiera, F., Pis, J. 2010. Influence of storage time on the quality and combustion behaviour of pine woodchips. Energy, 35(7), 3066-3071. Cassells, J., Caddick, L., Green, J., Reuss, R. 2003. Isotherms for Australian canola varieties. Proceedings of the Australian Postharvest Technical Conference, 25-27 June, Canberra. pp. 59-63. Chen, C.C., Morey, R.V. 1989. Comparison of four EMC/ERH equations. American Society of Agricultural Engineers, 32(3), 983-990. Choudhury, D., Sahu, J.K., Sharma, G. 2011. Moisture sorption isotherms, heat of sorption and properties of sorbed water of raw bamboo (Dendrocalamus longispathus) shoots. Industrial Crops and Products, 33(1), 211-216. Chowdhury, M., Huda, M., Hossain, M., Hassan, M. 2006. Moisture sorption isotherms for mungbean (Vigna radiata). Journal of food engineering, 74(4), 462-467. Das, D., Gaur, V., Verma, N. 2004. Removal of volatile organic compound by activated carbon fiber. Carbon, 42(14), 2949-2962. Di Blasi, C. 2008. Modeling chemical and physical processes of wood and biomass pyrolysis. Progress in Energy and Combustion Science, 34(1), 47-90. Domingo, J., Nadal, M. 2009. Domestic waste composting facilities: A review of human health risks. Environment international, 35(2), 382-389. Droin, A., Taverdet, J., Vergnaud, J. 1988. Modeling the kinetics of moisture adsorption by wood. Wood Science and Technology, 22(1), 11-20. Easterly, J.L., Burnham, M. 1996. Overview of biomass and waste fuel resources for power production. Biomass and Bioenergy, 10(2), 79-92. 147 Eitzer, B. 1995. Emissions of volatile organic chemicals from municipal solid waste composting facilities. Environmental Science & Technology, 29(4), 896-902. EnviroMed Detection Services. Threshold Limit Values from ACGIH 2008 TLVs and BEIs, http://www.enviromed.ca/qa_tox_01.htm. Mount Pearl. Environment Canada. http://climate.weatheroffice.gc.ca/climate_normals/index_e.html. EPA. 2012. Inventory of US greenhouse gas emissions and sinks: 1990-2010. United States Environmental Protection Agency 430-R-12-004. Eriksson, L., Gustavsson, L. 2010. Comparative analysis of wood chips and bundles-Costs, carbon dioxide emissions, dry-matter losses and allergic reactions. Biomass and Bioenergy, 34(1), 82-90. Ertekin, C., Yaldiz, O. 2004. Drying of eggplant and selection of a suitable thin layer drying model. Journal of food engineering, 63(3), 349-359. Feist, W., Springer, E., Hajny, G. 1973. Spontaneous heating in piled wood chips- contribution of bacteria. Tappi, 56(4), 148-151. Forintek. 2004. Drying Aspen and Birch - Northern Hardwoods Need A Special Approach. Forintek Canada publication, TP-03-04W. Frankel, E., Hu, M., Tappel, A. 1989. Rapid headspace gas chromatography of hexanal as a measure of lipid peroxidation in biological samples. Lipids, 24(11), 976-981. Fuller, W.S. 1985. Chip pile storage - a review of practices to avoid deterioration and economic losses. Tappi Journal, 68(8), 48-52. García-Pascual, P., Sanjuán, N., Melis, R., Mulet, A. 2006. Morchella esculenta (morel) rehydration process modelling. Journal of food engineering, 72(4), 346-353. Gigler, J., Van Loon, W., Seres, I., Meerdink, G., Coumans, W. 2000a. Drying Characteristics of Willow Chips and Stems. Journal of Agricultural Engineering Research, 77(4), 391-400. Gigler, J., Van Loon, W., Vissers, M., Bot, G. 2000b. Forced convective drying of willow chips. Biomass and Bioenergy, 19(4), 259-270. Gislerud, O. 1990. Drying and storing of comminuted wood fuels. Biomass, 22(1-4), 229- 244. Gjoelsjoe, S. 1995. Storage of comminuted birch piles in Norway. Uppsatser och Resultat- Sveriges Lantbruksuniversitet, Institutionen foer Skogsteknik (Sweden), 278, 76-86. Gonzalez, J.S. 1997. Growth, Properties and Uses of Western Redcedar (Thuja Plicata Donn Ex D. Don.). Forintek Canada Corporation. No. SP-37R. Gowen, A., Abu-Ghannam, N., Frias, J., Oliveira, J. 2007. Influence of pre-blanching on the water absorption kinetics of soybeans. Journal of food engineering, 78(3), 965-971. Granstrom, K., Mansson, B. 2008. Volatile organic compounds emitted from hardwood drying as a function of processing parameters. International Journal, 5(2), 141-148. Greaves, H. 1975. Microbiological aspects of wood chip storage in tropical environments. Australian journal of biological sciences, 28(3), 315-322. 148 Hagstrm, K. 2008. Occupational exposure during production of wood pellets in Sweden. in: Department of Natural Sciences, Vol. Doctoral thesis, Örebro University, pp. 75. Hartley, I., Avramidis, S. 1994. Water clustering phenomenon in two softwoods during adsorption and desorption processes. Journal of the Institute of Wood Science, 13(4), 467-474. He, X., Lau, A.K., Sokhansanj, S., Jim Lim, C., Bi, X.T., Melin, S. 2012. Dry matter losses in combination with gaseous emissions during the storage of forest residues. Fuel, 95, 662-664. Heding, N., Kofman, P., Morsing, M. 1993. Large scale study on storage of woodfuel comminuted in different sizes: chips, chunkwood and firewood. Skovbrugsserien (Denmark), 7, 48. Hellebrand, H., Schade, G. 2008. Carbon monoxide from composting due to thermal oxidation of biomass. Journal of environmental quality, 37(2), 592-598. Hellenbrand, K., Reade, A. 1992. Microorganisms associated with fuel wood chips and their impact on indoor air quality: a review. International Biodeterioration & Biodegradation, 29(1), 19-43. Henderson, S. 1952. A basic concept of equilibrium moisture. Agricultural Engineering, 33(1), 29-32. Hoell, W., Piezconka, K. 1978. Lipids in sap and heartwood of Picea abies (L.) Karst. Z Pflantzenphysiol, 87, 191-8. Hyttinen, M., Masalin-Weijo, M., Kalliokoski, P., Pasanen, P. 2010. Comparison of VOC emissions between air-dried and heat-treated Norway spruce (Picea abies), Scots pine (Pinus sylvesteris) and European aspen (Populus tremula) wood. Atmospheric Environment, 44(38), 5028-5033. Iglesias, H., Chirife, J. 1976. Prediction of the effect of temperature on water sorption isotherms of food material. International Journal of Food Science & Technology, 11(2), 109-116. Iguaz, A., Virseda, P. 2007. Moisture desorption isotherms of rough rice at high temperatures. Journal of food engineering, 79(3), 794-802. Jayas, D.S., Cenkowski, S., Pabis, S., Muir, W.E. 1991. Review of thin-layer drying and wetting equations. Drying technology, 9(3), 551-588. Jensen, R.A., Davis, J.R. 1953. Seasonal moisture variations in aspen. Minnesota Forestry Notes(19). Jirjis, R. 2005. Effects of particle size and pile height on storage and fuel quality of comminuted Salix viminalis. Biomass and Bioenergy, 28(2), 193-201. Jirjis, R. 1995. Storage and drying of wood fuel. Biomass and Bioenergy, 9(1-5), 181-190. Jirjis, R. 2003. Storage of forest residues in bales. Department of Bioenergy, Swedish University of Agricultural Sciences, Uppsala, 23. 149 Jirjis, R., Lehtikangas, P. 1993. Fuel Quality and Dry Matter Loss during Storage of Logging Residues in a Windrow. Sveriges Lantbruksuniversitet, Ultunabiblioteket, (Sweden) SLU. Johansson, A., Rasmuson, A. 1998. The release of monoterpenes during convective drying of wood chips. Drying Technology, 16(7), 1395-1428. Johansson, J., Salin, J.G. 2011. Application of percolation modelling on end-grain water absorption in aspen (Populus tremula L.). Wood Material Science & Engineering, 6(3), 112-118. Johansson, L.S., Leckner, B., Gustavsson, L., Cooper, D., Tullin, C., Potter, A. 2004. Emission characteristics of modern and old-type residential boilers fired with wood logs and wood pellets. Atmospheric Environment, 38(25), 4183-4195. Kačík, F., Veľková, V., Šmíra, P., Nasswettrová, A., Kačíková, D., Reinprecht, L. 2012. Release of Terpenes from Fir Wood during Its Long-Term Use and in Thermal Treatment. Molecules, 17(8), 9990-9999. Karunanithy, C., Muthukumarappan, K., Donepudi, A. 2013. Moisture sorption characteristics of switchgrass and prairie cord grass. Fuel, 103, 171-178. Khazaei, J. 2008. Characteristics of Mechanical Strength and Water Absorption in Almond and Its Kernel. Cercetări Agronomice În Moldova, 133, 37-51. Komilis, D., Ham, R., Park, J. 2004. Emission of volatile organic compounds during composting of municipal solid wastes. Water Research, 38(7), 1707-1714. Koppmann, R., Von Czapiewski, K., Reid, J. 2005. A review of biomass burning emissions, part I: gaseous emissions of carbon monoxide, methane, volatile organic compounds, and nitrogen containing compounds. Atmospheric Chemistry and Physics Discussions, 5(5), 10455-10516. Krupińska, B., Strømmen, I., Pakowski, Z., Eikevik, T. 2007. Modeling of sorption isotherms of various kinds of wood at different temperature conditions. Drying technology, 25(9), 1463-1470. Kuang, X., Shankar, T., Bi, X., Lim, C., Sokhansanj, S., Melin, S. 2009. Rate and Peak Concentrations of Off-Gas Emissions in Stored Wood Pellets--Sensitivities to Temperature, Relative Humidity, and Headspace Volume. Annals of Occupational Hygiene, 53(8), 789-796. Kuang, X., Shankar, T., Bi, X., Sokhansanj, S., Jim Lim, C., Melin, S. 2008. Characterization and kinetics study of off-gas emissions from stored wood pellets. Annals of Occupational Hygiene, 52(8), 675-683. Kudra, T., Strumillo, C. 1998. Thermal processing of bio-materials. Gordon & Breach Science Publishers, The Netherlands. Lahsasni, S., Kouhila, M., Mahrouz, M., Fliyou, M. 2003. Moisture adsorption-desorption isotherms of prickly pear cladode (Opuntia ficus indica) at different temperatures. Energy conversion and management, 44(6), 923-936. 150 Lahsasni, S., Kouhila, M., Mahrouz, M., Kechaou, N. 2002. Experimental study and modelling of adsorption and desorption isotherms of prickly pear peel (Opuntia ficus indica). Journal of food engineering, 55(3), 201-207. Lehtikangas, P. 2001. Quality properties of pelletised sawdust, logging residues and bark. Biomass and Bioenergy, 20(5), 351-360. Leinonen, A., tutkimuskeskus, V.t. 2004. Harvesting technology of forest residues for fuel in the USA and Finland. VTT TIEDOTTEITA. Lewis, W. 1921. The Rate of Drying of Solid Materials. Industrial & Engineering Chemistry, 13(5), 427-432. Lindgren, R.M., Eslyn, W.E. 1961. Biological deterioration of pulpwood and pulp chips during storage. Tappi, 44(6), 419-429. Liu, Z.M., Daniels, C., Morris, P.I. 2010. New Method to Isolate Gamma-Thujaplicin from Western Red Cedar (Thuja plicata Donn.). Journal of Wood Chemistry and Technology, 30(3), 299-314. MacGregor, S., Miller, F., Psarianos, K., Finstein, M. 1981. Composting process control based on interaction between microbial heat output and temperature. Applied and Environmental Microbiology, 41(6), 1321-1330. Maciejewska, A., Veringa, H., Sanders, J., Peteves, S. 2006. Co-firing of biomass with coal: constraints and role of biomass pre-treatment. Petten, The Netherlands: Institute for Energy, 113, 100. Madsen, A., Mårtensson, L., Schneider, T., Larsson, L. 2004. Microbial dustiness and particle release of different biofuels. Annals of Occupational Hygiene, 48(4), 327-338. Marabi, A., Livings, S., Jacobson, M., Saguy, I. 2003. Normalized Weibull distribution for modeling rehydration of food particulates. European Food Research and Technology, 217(4), 311-318. McGinnis, M.R. 2007. Indoor mould development and dispersal. Medical Mycology, 45(1), 1-9. McKendry, P. 2002. Energy production from biomass (part 1): overview of biomass. Bioresource technology, 83(1), 37-46. Merakeb, S., Dubois, F., Petit, C. 2009. Modeling of the sorption hysteresis for wood. Wood Science and Technology, 43(7), 575-589. Mohamed, A., Kouhila, M., Jamali, A., Lahsasni, S., Mahrouz, M. 2005a. Moisture sorption isotherms and heat of sorption of bitter orange leaves (Citrus aurantium). Journal of food engineering, 67(4), 491-498. Mohamed, L.A., Kouhila, M., Lahsasni, S., Jamali, A., Idlimam, A., Rhazi, M., Aghfir, M., Mahrouz, M. 2005b. Equilibrium moisture content and heat of sorption of Gelidium sesquipedale. Journal of Stored Products Research, 41(2), 199-209. Moreno, R., Antolı́n, G., Reyes, A., Alvarez, P. 2004. Drying Characteristics of Forest Biomass Particles of Pinus radiata. Biosystems engineering, 88(1), 105-115. 151 Mujumdar, A.S., Devahastin, S. 2000. Fundamental principles of drying. Exergex, Brossard, Canada. Nakamura, A., Miyafuji, H., Saka, S. 2010. Liquefaction behavior of Western red cedar and Japanese beech in the ionic liquid 1-ethyl-3-methylimidazolium chloride. Holzforschung, 64(3), 289-294. Nilsson, D. 1999. SHAM--a simulation model for designing straw fuel delivery systems. Part 1: model description. Biomass and Bioenergy, 16(1), 25-38. Nilsson, D., Karlsson, S. 2005. A model for the field drying and wetting processes of cut flax straw. Biosystems engineering, 92(1), 25-35. Nilsson, D., Svennerstedt, B., Wretfors, C. 2005. Adsorption equilibrium moisture contents of flax straw, hemp stalks and reed canary grass. Biosystems engineering, 91(1), 35- 43. Noordermeer, M., Veldink, G., Vliegenthart, J. 2001. Fatty acid hydroperoxide lyase: a plant cytochrome P450 enzyme involved in wound healing and pest resistance. Chembiochem, 2(7-8), 494-504. Nurmi, J. 1999. The storage of logging residue for fuel. Biomass and Bioenergy, 17(1), 41- 47. Obernberger, I., Biedermann, F., Widmann, W., Riedl, R. 1997. Concentrations of inorganic elements in biomass fuels and recovery in the different ash fractions. Biomass and Bioenergy, 12(3), 211-224. Ontario Ministry of Labour. 2012. Control of Exposure to Biological or Chenlical Agents. Reg. 833, R.R.O.1990, Ministry of Labour. Toronto, Ontario, Canada. Oswin, C. 1946. The kinetics of package life. III. The isotherm. Journal of the Society of Chemical Industry, 65(12), 419-421. Overhults, D., White, G., Hamilton, H., Ross, I. 1973. Drying soybeans with heated air. Transactions of the ASAE, 16(2), 195-200. Pagano, A., Mascheroni, R. 2005. Sorption isotherms for amaranth grains. Journal of food engineering, 67(4), 441-450. Panchariya, P., Popovic, D., Sharma, A. 2002. Thin-layer modelling of black tea drying process. Journal of food engineering, 52(4), 349-357. Penman, H.L. 1948. Natural evaporation from open water, bare soil and grass. Proceedings of the Royal Society of London. Series A. Mathematical and Physical Sciences, 193(1032), 120-145. Perré, P. 2007. Fundamentals of wood drying. Nancy, France : A.R.BO.LOR. Pettersson, M., Nordfjell, T. 2007. Fuel quality changes during seasonal storage of compacted logging residues and young trees. Biomass and Bioenergy, 31(11-12), 782-792. Pfost, H., Maurer, S., Chung, D., Milliken, G.A. 1976. Summarizing and reporting equilibrium moisture data for grains. St. Joseph: ASAE. 152 Phanphanich, M., Mani, S. 2009. Drying characteristics of pine forest residues. BioResources, 5(1), 108-121. Piispanen, R., Saranpaa, P. 2002. Neutral lipids and phospholipids in Scots pine (Pinus sylvestris) sapwood and heartwood. Tree Physiology, 22(9), 661. Ragauskas, A.J., Williams, C.K., Davison, B.H., Britovsek, G., Cairney, J., Eckert, C.A., Frederick Jr, W.J., Hallett, J.P., Leak, D.J., Liotta, C.L. 2006. The path forward for biofuels and biomaterials. Science, 311(5760), 484-489. Rao, N., Grethlein, H., Reddy, C. 1996. Effect of temperature on composting of atrazine- amended lignocellulosic substrates. Compost science and utilization, 4, 83-88. Reed, C., Doyungan, S., Ioerger, B., Getchell, A. 2007. Response of storage molds to different initial moisture contents of maize (corn) stored at 25° C, and effect on respiration rate and nutrient composition. Journal of Stored Products Research, 43(4), 443-458. Rentizelas, A., Tolis, A., Tatsiopoulos, I. 2009. Logistics issues of biomass: The storage problem and the multi-biomass supply chain. Renewable and Sustainable Energy Reviews, 13(4), 887-894. Rupar, K., Sanati, M. 2005. The release of terpenes during storage of biomass. Biomass and Bioenergy, 28(1), 29-34. Sacilik, K., Elicin, A.K. 2006. The thin layer drying characteristics of organic apple slices. Journal of food engineering, 73(3), 281-289. Satin, J.G. 2011. Fibre level modelling of free water behaviour during wood drying and wetting. Maderas. Ciencia y tecnología, 13(2), 153-162. Savoie, P., Mailhot, A. 1986. Influence of eight factors on the drying rate of timothy hay. Canadian Agricultural Engineering, 28(2), 145-148. Scheffer, T.C. 1966. Natural resistance of wood to microbial deterioration. Annual Review of Phytopathology, 4(1), 147-168. Schieberle, P., Grosch, W. 1981. Model experiments about the formation of volatile carbonyl compounds. Journal of the American Oil Chemists' Society, 58(5), 602-607. Shankar, T., Xingya, K., Sokhansanj, S., Lim, C., Bi, X., Melin, S. 2008. Effect of storage temperature on off-gassing and physical properties of wood pellets. ASABE Paper, 84248. Sharp, J. 1982. A review of low temperature drying simulation models. Journal of Agricultural Engineering Research, 27(3), 169-190. Shi, J., Sharma-Shivappa, R.R., Chinn, M.S. 2012. Interactions between Fungal Growth, Substrate Utilization and Enzyme Production during Shallow Stationary Cultivation of Phanerochaete chrysosporium on Cotton Stalks. Enzyme and Microbial Technology, 51(1), 1-8. Shuttleworth, W., Evaporation, D. 1993. Maidment, Editor, Handbook of Hydrology, McGraw-Hill, New York. Siau, J.F. 1984. Transport processes in wood. Springer-Verlag, New York. 153 Silakul, T., Jindal, V. 2002. Equilibrium moisture content isotherms of mungbean. International Journal of food properties, 5(1), 25-35. Simpson, W. 1993. Determination and use of moisture diffusion coefficient to characterize drying of northern red oak (Quercus rubra). Wood Science and Technology, 27(6), 409-420. Singh, R. 2004. Equilibrium moisture content of biomass briquettes. Biomass and Bioenergy, 26(3), 251-253. Sinicio, R., Muir, W., Jayas, D., Cenkowski, S. 1995. Thin-layer drying and wetting of wheat. Postharvest Biology and Technology, 5(3), 261-275. Sokhansanj, S., Khoshtaghaza, H., Schoenau, G., Arinze, E., Tabil, L. 2003. Heat and moisture transfer and quality changes in containerized alfalfa cubes during transport. Transactions of the ASAE, 46(2), 423-432. Sokhansanj, S., Kumar, A., Turhollow, A.F. 2006. Development and implementation of integrated biomass supply analysis and logistics model (IBSAL). Biomass and Bioenergy, 30(10), 838-847. Solantausta, Y., Beckman, D., Bridgwater, A., Diebold, J., Elliott, D. 1992. Assessment of liquefaction and pyrolysis systems. Biomass and Bioenergy, 2(1-6), 279-297. Soysal, Y., Öztekin, S. 1999. Equilibrium moisture content equations for some medicinal and aromatic plants. Journal of Agricultural Engineering Research, 74(3), 317-324. Stahl, M., Granstrom, K., Berghel, J., Renstrom, R. 2004. Industrial processes for biomass drying and their effects on the quality properties of wood pellets. Biomass and Bioenergy, 27(6), 621-628. Stewart, D., Lievers, K. 1978. A simulation model for the drying and rewetting processes of wheat. Canadian Agricultural Engineering, 20, 53-59. Sun, D.W. 1999. Comparison and selection of EMC/ERH isotherm equations for rice. Journal of Stored Products Research, 35(3), 249-264. Sun, D.W. 1998. Selection of EMC/ERH isotherm equations for shelled corn based on fitting to available data. Drying technology, 16(3-5), 779-797. Sun, D.W., Byrne, C. 1998. Selection of EMC/ERH isotherm equations for rapeseed. Journal of Agricultural Engineering Research, 69(4), 307-315. Sun, D.W., Woods, J. 1994. The selection of sorption isotherm equations for wheat based on the fitting of available data. Journal of Stored Products Research, 30(1), 27-43. Svedberg, U., Hogberg, H., Hogberg, J., Galle, B. 2004. Emission of hexanal and carbon monoxide from storage of wood pellets, a potential occupational and domestic health hazard. Annals of Occupational Hygiene, 48(4), 339-349. Svedberg, U., Petrini, C., Johanson, G. 2009. Oxygen depletion and formation of toxic gases following sea transportation of logs and wood chips. Annals of Occupational Hygiene, 53(8), 779-787. 154 Svedberg, U., Samuelsson, J., Melin, S. 2008. Hazardous off-gassing of carbon monoxide and oxygen depletion during ocean transportation of wood pellets. Annals of Occupational Hygiene, 52(4), 259-266. The National Institute for Occupational Safety and Health (NIOSH). 2007. Documentation for immediately dangerous to life or health concentrations (IDHL), Springfield, VA: NTIS. Thornqvist, T. 1985. Drying and storage of forest residues for energy production. Biomass, 7(2), 125-134. Thörnqvist, T. 1983. Fuel chips change during one year of storage. Rapport, Institutionen för Virkeslära, Sveriges Lantbruksuniversitet, 148, 69. Thörnqvist, T. 1984. Logging residues as a feedstock for energy production-drying, storing, handling and grading. in: Rapport, Institutionen för Virkeslära, Sveriges Lantbruksuniversitet, Vol. Doctoral, pp. 115. Tonn, B., Dengler, V., Thumm, U., Piepho, H.P., Claupein, W. 2011. Influence of leaching on the chemical composition of grassland biomass for combustion. Grass and Forage Science, 66(4), 464-473. Van den Berg, C., Bruin, S. 1981. Water activity and its estimation in food systems: Theoretical aspects. L.B.Rockland and G.F.Stewart eds., Academic Press, New York. Vassilev, S.V., Baxter, D., Andersen, L.K., Vassileva, C.G. 2010. An overview of the chemical composition of biomass. Fuel, 89(5), 913-933. Vijayaraj, B., Saravanan, R., Renganarayanan, S. 2007. Studies on thin layer drying of bagasse. International journal of energy research, 31(4), 422-437. Vikman, M., Karjomaa, S., Kapanen, A., Wallenius, K., Itävaara, M. 2002. The influence of lignin content and temperature on the biodegradation of lignocellulose in composting conditions. Applied microbiology and biotechnology, 59(4), 591-598. VOC-directive EU. 1999. Council Directive 1999/13/EC of 11 March 1999 on the limitation of emissions of volatile organic compounds due to the use of organic solvents in certain activities and installations. Official Journal of the European Communities: Brussels. Wihersaari, M. 2005a. Evaluation of greenhouse gas emission risks from storage of wood residue. Biomass and Bioenergy, 28(5), 444-453. Wihersaari, M. 2005b. Greenhouse gas emissions from final harvest fuel chip production in Finland. Biomass and Bioenergy, 28(5), 435-443. Wilkins, K., Larsen, K. 1996. Volatile organic compounds from garden waste. Chemosphere, 32(10), 2049-2055. Wiselogel, A., Agblevor, F., Johnson, D., Deutch, S., Fennell, J., Sanderson, M. 1996. Compositional changes during storage of large round switchgrass bales. Bioresource technology, 56(1), 103-109. Yang, C., Fon, D., Lin, T. 2007. Simulation and validation of thin layer models for peanut drying. Drying technology, 25(9), 1515-1526. 155 Yang, W., Sokhansanj, S., Cenkowski, S., Tang, J., Wu, Y. 1997. A general model for sorption hysteresis in food materials. Journal of food engineering, 33(3-4), 421-444. Yazdanpanah F., I.L., S. Sokhansanj, C.J. Lim, X. Bi, S. Melin. 2012. Concentration of CO2, CO, CH4, N2, O2, and H2 for 4%, 9%, and 15% moisture content wood pellets stored at temperature of 25oC, 40oC, and 60oC for 62 days. Unpublished report, Biomass and Bioenergy Research Group, The University of British Columbia. Zhang, D., Spadaro, D., Garibaldi, A., Gullino, M.L. 2010. Efficacy of the antagonist Aureobasidium pullulans PL5 against postharvest pathogens of peach, apple and plum and its modes of action. Biological Control, 54(3), 172-180. Zhang, L., Ninomiya, Y., Wang, Q., Yamashita, T. 2011. Influence of woody biomass (cedar chip) addition on the emissions of PM10 from pulverised coal combustion. Fuel, 90(1), 77-86. Zhang, Y., McKechnie, J., Cormier, D., Lyng, R., Mabee, W., Ogino, A., MacLean, H.L. 2009. Life cycle emissions and cost of producing electricity from coal, natural gas, and wood pellets in Ontario, Canada. Environmental science & technology, 44(1), 538-544. Zomorodian, A., Kavoosi, Z., Momenzadeh, L. 2010. Determination of EMC isotherms and appropriate mathematical models for canola. Food and Bioproducts Processing, 89(4), 407-413. Zomorodian, A., Tavakoli, R. 2007. The adsorption-desorption hysteresis effect on pistachio nuts. Journal of Agricultural Science and Technology, 9(4), 259-265. 156 Appendix A. Summary of different drying methods Table A1. Summary of different drying methods Method Description Advantages Disadvantages Conventional methods Natural drying (air drying) Materials are exposed to the natural weather environment (with/without cover) No energy input Low drying efficiency, depend on weather and take a long time (up to 6-12 months) Forced air drying Materials are exposed to the outside environment under cover with circulation by fans Little energy input Low drying efficiency Low temperature drying Temperature from 25- 55oC with circulation by fans Energy efficient, maintain the quality of materials Long drying time Conventional kiln drying In an insulated structure with temperatures up to 80ºC, with humidity control and circulation by fans Good control Relatively high cost High temperature drying Temperature from 100oC, up to 150oC High throughput, good control Bugs and insects are killed High cost, heat waste, strict requirement on insulation Specialized methods Dehumidification drying Moisture condenses and water is drained from the system Efficient utilisation of heat (recycled); good control Slow drying rate, low temperature (60oC), well insulation, chemical condensation Solar drying Materials are stored in a construction with solar collection Low cost, faster drying rate than air drying, materials protected from outside, free and abundant energy Depend on weather Vacuum drying Materials are dried in a reduced pressure environment, air is drier than the object being dried Less energy needed, work faster, little damage on materials High investment cost, products with an increased hygroscopic and thermo plasticity Radio frequency drying It heats at the molecular level from within the material and the middle as well as the surface Heat materials only, fast drying, no waste of heat, lower drying temperature, save space Expensive, risk of failure, technology needed 157 Appendix B. Derivation of Equation (5.1) Mass of materials in the reactor: m Void space in the container: V At t = 0, N2 concentration is Cn0 At time t, the N2 concentration is Cnt, the gas concentration is Ci, and the gas volume is Vt. Since N2 is an inert gas, which is not consumed or generated, the mass of N2 remains the same over the entire test. VCn0 = VtCnt Vt = VCn0/Cnt Based on Ideal Gas Law, PsV = nRT, for each gas species produced, PsCiVt = (mi/Mwt)RT PsCi (VCn0/Cnt) = (fi m/Mwt)RT Hence, 0s i wt n i nt PCVM Cf mRTC = 158 Appendix C. Summary of bales temperatures Table C1. Summary of monthly average temperature data for the four bales June July August September 1 2 3 4 1 2 3 4 1 2 3 4 1 2 3 4 1 13.5 13.5 13.5 13.5 14.8 14.8 14.8 14.8 18.0 18.0 18.0 18.0 16.2 16.2 16.2 16.2 2 11.9 11.9 11.9 11.9 17.6 17.6 17.6 17.6 20.1 20.1 20.1 20.1 16.1 16.1 16.1 16.1 3 13.8 13.8 13.8 13.8 17.2 17.2 17.2 17.2 19.5 19.5 19.5 19.5 18.1 18.1 18.1 18.1 4 14.0 14.0 14.0 14.0 17.0 17.0 17.0 17.0 20.1 20.1 20.1 20.1 19.9 19.9 19.9 19.9 5 15.3 15.3 15.3 15.3 18.6 18.6 18.6 18.6 17.7 17.7 17.7 17.7 20.7 20.7 20.7 20.7 6 16.5 16.5 16.5 16.5 18.7 18.7 18.7 18.7 17.9 17.9 17.9 17.9 20.9 20.9 20.9 20.9 7 14.4 14.4 14.4 14.4 16.3 16.3 16.3 16.3 18.0 18.0 18.0 18.0 20.5 20.5 20.5 20.5 8 14.5 14.5 14.5 14.5 15.6 15.6 15.6 15.6 18.1 18.1 18.1 18.1 21.7 21.7 21.7 21.7 9 15.5 15.5 15.5 15.5 15.3 15.3 15.3 15.3 18.1 18.1 18.1 18.1 19.6 19.6 19.6 19.6 10 15.3 15.3 15.3 15.3 17.5 17.5 17.5 17.5 17.0 17.0 17.0 17.0 19.4 19.4 19.4 19.4 11 14.6 14.6 14.6 14.6 17.9 17.9 17.9 17.9 17.8 17.8 17.8 17.8 20.2 20.2 20.2 20.2 12 15.4 15.4 15.4 15.4 17.1 17.1 17.1 17.1 19.1 19.1 19.1 19.1 19.3 19.3 19.3 19.3 13 16.2 16.2 16.2 16.2 16.3 16.3 16.3 16.3 18.0 18.0 18.0 18.0 17.2 17.2 17.2 17.2 14 14.3 14.3 14.3 14.3 15.5 15.5 15.5 15.5 18.1 18.1 18.1 18.1 17.2 17.2 17.2 17.2 15 14.1 14.1 14.1 14.1 17.3 17.3 17.3 17.3 17.4 17.4 17.4 17.4 15.9 15.9 15.9 15.9 16 13.9 13.9 13.9 13.9 16.2 16.2 16.2 16.2 17.8 17.8 17.8 17.8 14.1 14.1 14.1 14.1 17 16.0 16.0 16.0 16.0 16.8 16.8 16.8 16.8 17.7 17.7 17.7 17.7 13.9 13.9 13.9 13.9 18 14.8 14.8 14.8 14.8 17.8 17.8 17.8 17.8 17.3 17.3 17.3 17.3 15.0 15.0 15.0 15.0 19 14.8 14.8 14.8 14.8 18.3 18.3 18.3 18.3 18.4 18.4 18.4 18.4 15.7 15.7 15.7 15.7 20 16.7 16.7 16.7 16.7 17.3 17.3 17.3 17.3 20.6 20.6 20.6 20.6 16.5 16.5 16.5 16.5 21 17.7 17.7 17.7 17.7 16.2 16.2 16.2 16.2 24.1 24.1 24.1 24.1 17.7 17.7 17.7 17.7 22 16.7 16.7 16.7 16.7 16.2 16.2 16.2 16.2 18.6 18.6 18.6 18.6 17.2 17.2 17.2 17.2 23 15.1 15.1 15.1 15.1 16.8 16.8 16.8 16.8 19.2 19.2 19.2 19.2 21.3 21.3 21.3 21.3 24 14.0 14.0 14.0 14.0 19.7 19.7 19.7 19.7 19.3 19.3 19.3 19.3 20.6 20.6 20.6 20.6 25 14.8 14.8 14.8 14.8 18.5 18.5 18.5 18.5 20.1 20.1 20.1 20.1 15.3 15.3 15.3 15.3 26 14.7 14.7 14.7 14.7 16.5 16.5 16.5 16.5 20.5 20.5 20.5 20.5 12.6 12.6 12.6 12.6 27 17.7 17.7 17.7 17.7 17.6 17.6 17.6 17.6 20.7 20.7 20.7 20.7 14.5 14.5 14.5 14.5 28 18.7 18.7 18.7 18.7 18.1 18.1 18.1 18.1 19.6 19.6 19.6 19.6 11.8 11.8 11.8 11.8 29 17.4 17.4 17.4 17.4 18.5 18.5 18.5 18.5 17.5 17.5 17.5 17.5 14.5 14.5 14.5 14.5 30 16.0 16.0 16.0 16.0 20.5 20.5 20.5 20.5 17.2 17.2 17.2 17.2 14.5 14.5 14.5 14.5 31 18.1 18.1 18.1 18.1 16.8 16.8 16.8 16.8 159 October November December January 1 2 3 4 1 2 3 4 1 2 3 4 1 2 3 4 1 21.1 21.2 21.2 26.4 7.1 6.5 6.4 5.9 4.0 3.8 3.0 3.5 11.8 11.6 5.1 6.9 2 21.1 20.9 20.9 26.0 6.0 5.5 5.6 5.3 4.4 4.5 3.2 3.6 11.3 11.1 5.0 6.1 3 21.7 21.7 21.7 26.9 7.0 6.3 6.7 6.9 4.2 4.0 3.3 5.5 14.6 13.4 7.6 9.2 4 20.7 20.4 20.4 25.7 5.7 5.5 4.9 5.4 12.7 13.2 7.6 7.1 13.1 12.2 5.8 7.5 5 15.1 14.6 14.6 20.3 5.9 5.4 5.4 5.9 12.1 12.1 7.5 6.5 16.9 15.8 10.2 11.6 6 12.4 12.1 12.1 12.0 5.7 4.3 3.7 4.1 11.9 11.5 6.8 6.6 16.8 15.7 10.2 11.5 7 13.2 13.0 13.0 12.9 5.9 3.9 4.2 4.4 12.5 12.3 7.1 6.7 17.0 16.3 11.2 11.4 8 12.2 11.8 11.8 11.5 7.2 5.7 6.0 6.0 11.7 11.8 6.8 6.0 17.3 16.9 11.1 12.1 9 12.7 12.5 12.5 12.3 8.9 7.9 8.2 8.3 10.6 10.4 6.3 5.8 15.7 15.4 10.8 10.1 10 11.3 11.1 11.1 10.8 9.6 9.1 9.8 9.9 11.3 10.8 6.5 6.2 16.8 16.1 11.5 10.8 11 12.1 11.7 11.7 11.6 7.0 6.1 6.9 7.1 11.4 11.2 6.3 5.9 15.7 15.4 10.8 10.1 12 11.5 11.3 11.3 11.0 4.4 3.4 3.8 4.0 10.6 10.4 6.0 5.4 16.8 16.1 11.5 10.8 13 10.5 10.1 10.1 10.1 5.1 4.5 4.5 5.0 11.0 10.6 6.2 5.8 13.2 12.6 6.7 7.6 14 10.9 10.2 10.2 10.3 5.5 4.8 5.1 4.9 11.7 11.2 6.1 6.9 12.9 12.5 6.6 7.4 15 10.0 9.4 9.4 9.4 4.9 4.2 4.3 3.8 13.2 12.6 7.3 8.3 15.1 14.1 8.3 9.5 16 9.6 8.9 8.9 9.0 5.3 3.2 3.2 3.4 14.5 13.9 8.6 9.8 14.6 13.5 7.5 8.8 17 10.5 10.0 10.0 10.1 3.1 2.3 3.5 2.7 15.8 15.1 10.0 9.9 15.6 15.1 9.9 10.0 18 11.8 11.1 11.1 11.0 2.3 0.8 1.7 0.8 16.0 15.7 10.4 9.8 16.6 15.7 10.1 11.1 19 12.4 11.7 11.7 11.4 2.3 1.1 1.9 0.9 15.4 15.2 10.0 10.0 13.2 13.0 8.8 7.5 20 12.3 12.3 12.3 12.0 0.8 -0.3 0.3 - 0.4 14.6 14.7 9.1 8.6 13.7 14.1 9.0 8.2 21 11.2 10.6 10.6 10.7 4.4 2.8 2.7 2.0 12.3 12.4 7.4 6.7 11.3 10.9 6.7 5.5 22 10.9 10.6 10.6 10.7 7.4 5.7 6.9 6.2 10.8 11.0 6.2 6.7 11.5 11.2 6.7 5.7 23 10.5 10.3 10.3 10.5 5.8 4.7 5.8 5.4 13.3 12.2 7.1 8.7 10.6 10.2 5.6 5.2 24 9.6 9.0 9.0 9.2 4.3 3.0 3.7 3.4 15.5 14.5 9.3 10.1 11.2 10.8 5.5 5.9 25 8.9 8.2 8.2 8.5 4.9 3.4 4.3 4.0 15.6 15.4 9.9 9.2 10.9 10.3 5.1 5.5 26 7.1 6.7 6.7 6.6 6.7 4.8 5.5 5.2 13.5 13.5 7.9 8.4 11.8 10.3 5.1 5.6 27 7.2 6.6 6.6 6.8 8.5 6.9 8.4 7.8 14.2 13.3 8.1 10.5 10.7 10.3 4.6 5.4 28 7.5 6.8 6.8 6.8 5.5 4.5 5.5 4.7 17.9 16.7 11.9 11.0 10.9 10.4 4.9 5.4 29 8.5 7.6 7.6 7.5 4.5 3.5 4.4 3.5 15.3 15.1 10.0 9.2 10.7 10.3 4.7 5.2 30 9.6 8.8 8.8 8.6 4.2 2.8 4.0 3.0 14.4 13.8 8.8 7.7 10.7 10.3 4.6 5.2 31 9.4 9.0 9.0 8.5 12.2 12.1 6.8 6.5 10.7 10.3 4.7 5.2 160 February March April May 1 2 3 4 1 2 3 4 1 2 3 4 1 2 3 4 1 14.1 12.9 8.1 7.4 13.4 12.3 7.2 8.5 15.7 14.7 10.0 11.2 19.9 18.4 13.5 14.0 2 14.0 13.4 9.7 7.1 13.9 12.7 8.1 8.6 17.1 16.1 11.0 12.0 18.4 17.8 12.2 13.2 3 15.0 14.6 10.2 6.2 15.3 14.5 9.6 10.0 16.9 15.9 10.8 11.4 18.5 17.4 12.3 13.2 4 15.9 15.0 10.9 6.4 16.4 15.6 11.2 11.3 16.3 15.4 10.7 11.1 19.5 18.2 13.4 14.2 5 13.9 12.9 7.7 7.7 15.4 13.7 9.3 9.7 17.0 15.9 11.0 11.5 20.1 18.6 13.6 14.3 6 14.0 13.2 8.5 7.9 13.2 12.0 7.2 7.6 17.3 16.0 11.3 12.0 20.2 19.4 13.4 14.9 7 15.1 14.6 9.3 9.6 12.8 11.2 5.9 7.2 17.9 16.6 12.3 12.9 22.1 21.5 15.7 16.7 8 16.4 15.8 9.9 10.7 14.1 12.6 7.2 8.6 18.9 17.9 13.7 14.3 23.1 21.9 16.5 17.8 9 17.1 16.5 10.8 11.7 15.4 14.6 9.9 10.6 20.7 19.7 14.6 15.6 20.0 18.1 13.3 14.3 10 16.1 15.5 9.7 10.6 15.7 14.7 10.4 10.5 21.3 20.0 14.9 15.6 18.9 17.7 12.7 14.0 11 18.1 17.4 12.3 12.4 15.2 13.7 9.0 9.4 20.6 19.5 14.6 15.0 19.7 19.5 13.8 14.5 12 16.8 16.1 11.3 10.8 13.4 12.8 8.0 8.5 20.1 18.9 14.2 14.8 21.7 21.8 15.7 16.7 13 16.0 15.4 10.9 10.2 13.5 12.3 7.5 8.1 20.1 19.0 14.6 14.9 24.0 23.5 17.9 18.8 14 15.0 14.5 10.2 9.5 12.6 11.6 6.9 7.2 20.9 19.8 14.5 15.0 25.4 24.2 19.2 19.4 15 14.5 14.0 8.9 8.6 15.9 15.8 10.9 11.7 20.4 19.3 13.7 14.9 25.0 23.7 18.9 19.0 16 14.4 13.8 8.1 9.0 16.0 14.4 9.8 10.6 19.7 18.5 12.5 13.3 23.5 22.0 17.3 17.8 17 14.5 13.7 8.3 9.3 15.5 14.3 9.3 10.6 17.7 16.2 12.4 13.4 21.8 20.1 15.3 16.2 18 14.2 13.2 8.0 9.2 14.8 13.4 8.5 9.5 18.9 18.3 12.4 14.0 20.8 19.6 14.7 15.3 19 13.6 12.2 6.9 8.3 13.4 12.1 7.5 7.8 17.9 17.0 12.4 13.6 21.6 21.1 15.5 16.0 20 13.6 12.2 6.7 8.2 13.5 12.6 7.6 8.9 19.4 18.3 13.3 14.1 21.9 21.7 15.9 16.7 21 14.4 13.7 8.2 9.3 14.7 13.8 12.4 9.4 19.4 18.6 14.2 14.6 22.4 22.2 16.9 18.3 22 15.7 14.8 10.1 10.3 14.2 13.0 8.4 8.4 20.8 19.9 16.0 16.6 22.2 21.7 16.5 17.0 23 15.2 13.6 8.9 9.7 15.5 14.9 9.7 10.8 23.1 22.2 16.6 17.0 21.1 20.8 15.4 16.0 24 13.6 12.6 6.9 8.5 16.8 15.8 11.2 11.2 21.6 20.6 15.8 16.3 21.0 20.3 14.7 16.1 25 12.5 11.4 5.8 7.2 17.9 17.2 12.2 12.2 21.0 20.3 15.6 16.1 23.6 23.0 17.7 18.4 26 11.7 9.6 4.5 5.9 17.9 16.7 11.9 12.4 21.2 19.9 14.6 15.0 25.4 24.6 18.8 19.6 27 10.9 9.1 3.7 5.6 17.3 16.5 11.6 12.2 19.7 18.3 13.3 14.2 24.0 22.4 17.5 18.5 28 10.9 10.2 4.3 6.4 17.6 16.6 11.8 12.8 19.3 18.3 13.9 15.1 24.0 22.8 18.1 18.4 29 11.4 10.7 4.8 6.8 16.1 15.4 10.5 11.2 20.9 20.1 14.3 15.2 22.6 21.2 16.2 16.8 30 15.8 15.0 10.2 10.4 19.8 18.7 13.8 14.5 21.7 21.1 15.6 16.5 31 14.6 13.6 9.2 8.8 22.9 22.3 17.1 18.1 161 Figure C1. The locations of thermocouples in each bale 162 Appendix D. GC and GC/MS spectra 163 Figure D1. GC spectra of the gas samples from the reactors with experimental materials 164 Figure D2. GC/MS spectra of the gas samples from the reactors with experimental materials 165 (1) 2-Cyclohexen-1-one, 2-methyl-5-(1-methylethenyl)- (2) N-(m-Nitrophenyl)benzamide 166 (3) Bicyclo[2,2,1]heptane, 7,7-dimethyl-2-methylene- (4) Pentanal, 3 methyl- 167 (5) Ocimene (6) 1R-, alpha-Pinene 168 (7) Camphene (8) 2(3H)-Furanone, dihydro-3-methyl- 169 (9) 1H-Indole, 2-methyl-3-phenyl- (10) Isopropyl Alcohol 170 (11) 2-Phenazinecarboxylic acid Figure D3. Identification of compounds by using mass-charge ratio