Open Collections

UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

Preventive control of ammonia and odor emissions during the active phase of poultry manure composting Zhang, Wenxiu 2008

Your browser doesn't seem to have a PDF viewer, please download the PDF to view this item.

Item Metadata

Download

Media
24-ubc_2009_spring_zhang_wenxiu.pdf [ 1.78MB ]
Metadata
JSON: 24-1.0058524.json
JSON-LD: 24-1.0058524-ld.json
RDF/XML (Pretty): 24-1.0058524-rdf.xml
RDF/JSON: 24-1.0058524-rdf.json
Turtle: 24-1.0058524-turtle.txt
N-Triples: 24-1.0058524-rdf-ntriples.txt
Original Record: 24-1.0058524-source.json
Full Text
24-1.0058524-fulltext.txt
Citation
24-1.0058524.ris

Full Text

PREVENTIVE CONTROL OF AMMONIA AND ODOR EMISSIONS DURING THE ACTIVE PHASE OF POULTRY MANURE COMPOSTING  by WENXIU ZHANG B.Sc., The Inner Mongolia Agricultural University, China, 1984 M.Sc., China Institute of Water Resources and Hydro-power Research, 1987 M.A.Sc., The University of British Columbia, 2002  A THESIS SUBMITTED IN PARTIAL FULFILLMENT 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)  September 2008 © Wenxiu Zhang 2008  ABSTRACT  Traditional measures used in the composting industry for ammonia and odor emissions control are those involving collection and treatment such as thermal oxidation, adsorption, wet scrubbing and biofiltration. However, these methods do not address the source of the odor generation problem. The primary objective of this thesis research was to develop preventive means to minimize ammonia and odor emissions, and maximize nitrogen conservation to increase the agronomic value of compost product.  Laboratory-scale  experiments were performed to examine the effectiveness of various technologies to minimize these emissions during the active phase of composting. These techniques included precipitating ammonium into struvite in composting matrix before it release to outside environment; the use of chemical and biological additives in the form of yeast, zeolite and alum; and the manipulation of key operational parameters during the composting process. The fact that struvite crystals were formed in manure composting media, as verified by both XRD and SEM-EDS analyses, represents novel findings from this study. This technique was able to reduce ammonia emission by 40-84%, while nitrogen content in the finished compost was increased by 37-105%. The application of yeast and zeolite with dosages of 5-10% enhanced the thermal performance of composting and the degree of degradation, and ammonia emission was reduced by up to 50%. Alum was found to be the most effective additive for both ammonia and odor emission control; ammonia emission decreased by 4590% depending on the dosage, and odor emission assessed via an upgraded dynamic dilution olfactometer was reduced by 44% with dosages above 2.5%. This study reaffirmed that aeration is the most influential factor to odor emission, and an optimal air flowrate for odor control would be 0.6 L/min.kg dry matter with an intermittent aeration system. Quantitative relationships between odor emission and the key operational parameters were determined, which would enable “best management practices” to be devised and implemented for composting. An empirical odor predictive model was developed to provide a simple and direct means for simulation of composting odor emission. The effects of operating conditions were incorporated into the model with the multiplicative algorithm and linearization approximation approach. The model was validated with experimental observations. ii  TABLE OF CONTENTS Abstract ............................................................................................................................... ii Table of Contents ............................................................................................................... iii List of Tables .................................................................................................................... vii List of Figures .................................................................................................................... ix Acknowledgements ........................................................................................................... xii Statement of Co-Authorship ............................................................................................ xiii Chapter 1 Introduction .........................................................................................................1 1.1 General Introduction ................................................................................................1 1.1.1 Problem Statement ..........................................................................................1 1.1.2 Research Needs ...............................................................................................3 1.1.3 Objectives .......................................................................................................6 1.1.4 Thesis Organization ........................................................................................8 1.2 Background Research ............................................................................................10 1.2.1 Characteristics of Odors ................................................................................10 1.2.2 Odor Measurement........................................................................................15 1.2.3 Odor Emissions from Composting................................................................18 1.2.4 Principles of Odor Control ............................................................................23 1.2.5 Ammonia Emission and Control...................................................................25 1.3 References ..............................................................................................................34 Chapter 2 Lab-Scale Bioreactor System Design and Olfactometer Development ............44 2.1 Introduction ............................................................................................................44 2.2 Bioreactor System Design and Evaluation ............................................................45 2.2.1 System Configuration ...................................................................................45 2.2.2 Process Control and Data Acquisition ..........................................................48 2.2.3 Overall System Performance ........................................................................48 2.3 Dynamic Dilution Olfactometer ............................................................................53 2.3.1 Design ...........................................................................................................53 2.3.2 Calibration.....................................................................................................55  iii  2.3.3 Operation.......................................................................................................60 2.3.4 Data Evaluation .............................................................................................60 2.3.5 Overall Odor Laboratory Performance Quality ............................................63 2.4 Conclusions ............................................................................................................65 2.5 References ..............................................................................................................65 Chapter 3 Reducing Ammonia and Odor Emissions Via Struvite Formation ...................71 3.1 Introduction ............................................................................................................71 3.2 Materials and Methods ...........................................................................................73 3.2.1 Substrate Characterization and Recipe Formulation.....................................73 3.2.2 Experimental Design and Setup ....................................................................74 3.2.3 Experimental Procedure ................................................................................76 3.2.4 Analytical Methods .......................................................................................77 3.2.5 Statistical Analysis ........................................................................................79 3.3 Results and Discussion ..........................................................................................79 3.3.1 Struvite Formation ........................................................................................79 3.3.2 Ammonia Emission.......................................................................................85 3.3.3 Nitrogen Retention ........................................................................................89 3.3.4 Changes in Water Soluble Orthophosphate ..................................................92 3.3.5 Impacts on Composting Process ...................................................................94 3.4 Conclusions ..........................................................................................................103 3.5 References ............................................................................................................105 Chapter 4 Effects of Additives in the Form of Yeasts, Zeolite and Alum on Composting Process and Odor Emissions ..........................................................................109 4.1 Introduction ..........................................................................................................109 4.2 Background Research ..........................................................................................110 4.3 Materials and Methods .........................................................................................113 4.3.1 Experimental Set-up and Treatments ..........................................................113 4.3.2 Sampling and Analytical Measurement ......................................................115 4.3.3 Ammonia and Odor Emission Measurements ............................................116 4.3.4 Statistical Analysis ......................................................................................117  iv  4.4 Results and Discussion ........................................................................................117 4.4.1 Composting Process and Parameters ..........................................................117 4.4.2 Ammonia Emission.....................................................................................126 4.4.3 Odor Emissions ...........................................................................................130 4.5 Conclusions ..........................................................................................................139 4.6 References ............................................................................................................141 Chapter 5 Effects of Operational Parameters on Odor Emissions ...................................146 5.1 Introduction ..........................................................................................................146 5.2 Materials and Methods .........................................................................................147 5.3 Results and Discussion ........................................................................................150 5.3.1 Effect of Aeration Flowrate ........................................................................150 5.3.2 Effect of Initial Moisture Content ...............................................................156 5.3.3 Effect of Temperature Setpoint ...................................................................158 5.3.4 Effect of Biodegradable Volatile Solid .......................................................161 5.4 Conclusions ..........................................................................................................167 5.5 References ............................................................................................................168 Chapter 6 Predictive Odor Model ....................................................................................170 6.1 Introduction ..........................................................................................................170 6.2 Model Development.............................................................................................172 6.2.1 General Description ....................................................................................172 6.2.2 Reference Odor Predictive Model ..............................................................177 6.2.3 Generalized Odor Predictive Model ...........................................................180 6.2.4 Determination of Effect Factors..................................................................183 6.3 Model Validation .................................................................................................185 6.4 Sensitivity Analysis .............................................................................................189 6.5 Conclusions ..........................................................................................................196 6.6 References ............................................................................................................198 Chapter 7 Conclusion and Recommendations .................................................................201 7.1 Summary and Conclusions ..................................................................................201  v  7.2 Contributions to Research and Practical Applications in Composting  .........204  7.3 Recommendations for Future Research ...............................................................206 7.4 References ............................................................................................................208  Appendix I. Results of Olfactometer Calibration ............................................................209 Appendix II. Temperature Profiles of the Tests at Different Operation Conditions .......213 Appendix III. Summary of Statistical Analysis ...............................................................218  vi  LIST OF TABLES Table 1.1 Major odorous compounds from composting process .......................................14 Table 1.2 Properties of struvite ..........................................................................................30 Table 1.3 Struvite solubility products reported in literature ..............................................30 Table 2.1 Temperature-time data from laboratory scale bioreactor system ......................52 Table 2.2 Properties of the force choice dynamic dilution olfactometer ...........................59 Table 2.3 Results of olfactometer airflow calibration .......................................................60 Table 2.4 Result code driving from panel observation ......................................................63 Table 2.5 Calculation of precision for odor lab performance ............................................64 Table 3.1 Characterization of substrates ............................................................................70 Table 3.2 Experimental design for reducing ammonia emission via struvite formation ...75 Table 3.3 Summary of the temperature parameters and total ammonia emission for replicates of treatment with Mg/P molar ratio of 1.25 ..................................87 Table 3.4 Inorganic ammonia nitrogen retained in compost after curing stage.................92 Table 3.5 Changes in water-soluble orthophosphate in compost during the active phase 97 Table 3.6 Thermal performances of different treatments during the active phase of composting ..........................................................................................97 Table 3.7 Changes in volatile solid and degree of degradation of different treatments ....88 Table 4.1 Experimental treatments for composting additive test ....................................114 Table 4.2 Initial physical and chemical properties of the composting mixture for additive tests ...............................................................................................114 Table 4.3 Thermal performances of additive treatments during active phase of composting...................................................................................................124 Table 4.4 Changes in pH and nitrogen contents of additive treatments ..........................125 Table 4.5 Changes in volatile solids and degree of degradation of additive treatments ..125 Table 4.6 Summary of the temperature parameters and odor emission for replicates of yeast and zeolite treatments ....................................................131 Table 4.7 Summary of the temperature parameters and odor concentrations for replicates of alum treatment .......................................................................132 Table 4.8 Summary of the additives tested for reducing ammonia and odor emissions .138  vii  Table 5.1 Experimental design for effects of operation conditions on odor emission ....149 Table 5.2 Initial physical and chemical properties of the composting mixture for Operational parameter tests..............................................................................150 Table 5.3 Odor emission rates of treatments with different aeration flowrate ................152 Table 5.4 Odor emission rates of treatments at different initial moisture levels .............154 Table 5.5 Odor emission rates of treatments at different temperature setpoint ...............157 Table 5.6 Odor emission rates of treatments with different level of biodegradable volatile solid .....................................................................................................160 Table 5.7 Statistical analysis on experimental data from the control treatment ..............162 Table 6.1 Results of parameter estimation for reference odor model ..............................179 Table 6.2 Case scenarios for model validation ................................................................186 Table 6.3 Relative sensitivity for odor emission with respect to aeration, moisture content and biodegradable volatile solids .........................................190 Table 6.4 Parameter values used in the sensitivity analysis ............................................191 Table 6.5 Summary of sensitivity analysis for peak odor emission and cumulative odor emission to deviations in model parameters ..........................192  viii  LIST OF FIGURES  Figure 1.1 Odor measurement methods .............................................................................16 Figure 1.2 Substrate breakdown process ...........................................................................19 Figure 1.3 Relative concentration of NH3 and NH4 in solution .......................................26 Figure 1.4 Magnesium speciation versus pH at 20 oC .......................................................33 Figure 2.1 Setup of composting unit: Bioreactor with periphery equipment ....................47 Figure 2.2 Temperature feedback control flowchart..........................................................49 Figure 2.3 Drawing language program for composting process control ...........................50 Figure 2.4 Generic composting process temperature profile .............................................51 Figure 2.5 Schematic of the forced choice dynamic dilution olfactometer .......................56 Figure 2.6 Layout of air chambers and airflow lines .........................................................57 Figure 2.7 Olfactometer calibration setup .........................................................................58 Figure 3.1 Change in pH over time from experimental set 1.............................................80 Figure 3.2 Change in pH over time from experimental set 2.............................................80 Figure 3.3 X-ray diffraction spectrum of the precipitates obtained from the final compost ..............................................................................................83 Figure 3.4 Scanning electron micrograph (SEM) of the precipitates obtained from compost after active phase of composting ...............................................84 Figure 3.5 Energy dispersive X-ray spectroscopy (EDS) analysis of the precipitates obtained from compost after active phase of composting ...............................84 Figure 3.6 Cumulative ammonia emissions of different treatments as affected by Mg and P salt addition (Mg:P ratio of 1:1 with turning) ..................................88 Figure 3.7 Cumulative ammonia emissions of different treatments during the active phase of composting (Mg:P ratio of 1.25:1 with turning) ............88 Figure 3.8 Variation of water soluble ammonium nitrogen in compost over time as affected by Mg and P salt addition (Mg:P molar ratio of 1:1 with turning) ...90 Figure 3.9 Variation of water soluble ammonium nitrogen in compost over time as affected by Mg and P salt addition (Mg:P ratio of 1.25:1 with turning) ........90 Figure 3.10 Changes in water-soluble orthophosphate with time for treatments with Mg:P molar ratio of 1:1 and turning ..............................................................96  ix  Figure 3.11 Changes in water-soluble orthophosphate with time for treatments with Mg:P molar ratio of 1:1.25 and turning .........................................................96 Figure 3.12 Temperature profiles of treatments with Mg/P molar ratio of 1:1 .................99 Figure 3.13 Temperature profiles of treatments with Mg/P molar ratio of 1.25:1 (Replicate 1) .................................................................................................100 Figure 3.14 Temperature profiles of treatments with Mg/P molar ratio of 1.25:1 (Replicate 2) .................................................................................................101 Figure 3.15 Changes in EC with time for experimental Set 1 .........................................102 Figure 3.16 Changes in EC with time for experimental Set 2 .........................................102 Figure 4.1 Temperature profiles of the treatment with yeast combined with zeolite ......121 Figure 4.2 Temperature profiles of the treatments with yeast .........................................122 Figure 4.3 Temperature profiles of the treatments with alum .........................................123 Figure 4.4 Ammonia emission rate over time as a function of the application rate of yeast combined with zeolite ......................................................................127 Figure 4.5 Ammonia emission rate over time as a function of the application rate of yeast ..........................................................................................................127 Figure 4.6 Ammonia emission rate over time as a function of the application rate of alum ..........................................................................................................128 Figure 4.7 Cumulative ammonia emission over time for treatment with yeast and zeolite ............................................................................................................128 Figure 4.8 Cumulative ammonia emission over time for treatment with yeast ...............129 Figure 4.9 Cumulative ammonia emission over time for treatment with alum ...............129 Figure 4.10 Variation of odor concentration with time for treatment with yeast combined with zeolite ...................................................................................133 Figure 4.11 Variation of odor concentration with time for treatment with yeast ............133 Figure 4.12 Variation of odor concentration with time for treatment with alum ............134 Figure 4.13 Cumulative odor emissions from yeast and zeolite treated compost ...........136 Figure 4.14 Cumulative odor emissions from yeast treated compost ..............................136 Figure 4.15 Cumulative odor emissions from alum treated compost ..............................137 Figure 5.1 Cumulative odor emissions over time as affected by aeration flowrate .........155 Figure 5.2 Cumulative odor emissions over time as affected by initial moisture............158  x  Figure 5.3 Cumulative odor emissions over time as affected by temperature setpoint ...161 Figure 5.4 Cumulative odor emissions over time as affected by biodegradable volatile solid .................................................................................................163 Figure 6.1 Schematic of a typical microbial growth curve ..............................................175 Figure 6.2 Schematic of generalized effect function with linear approximation approach .........................................................................................................182 Figure 6.3 Measured and model predicted odor emission rate as a function of time for case 1 ........................................................................................................187 Figure 6.4 Measured and model predicted odor emission rate as a function of time for case 2 ........................................................................................................188 Figure 6.5 Measured and model predicted odor emission rate as a function of time for case 3 ........................................................................................................188 Figure 6.6 Sensitivity of cumulative odor emission to changes in parameter G0 ............193 Figure 6.7 Sensitivity of cumulative odor emission to changes in parameter G1 ............193 Figure 6.8 Sensitivity of cumulative odor emission to changes in parameter B1 ............194 Figure 6.9 Sensitivity of cumulative odor emission to changes in parameter M1............194 Figure 6.10 Sensitivity of cumulative odor emission to changes in parameter B2 ..........195 Figure 6.11 Sensitivity of cumulative odor emission to changes in parameter M2..........195  xi  ACKNOWLEDGEMENTS I would like to express my greatest gratitude and deepest appreciation to my academic supervisor Dr. Anthony K. Lau for making the entire program possible. I greatly appreciate his valuable advice, constructive criticism, thoughtful suggestions, rewarding discussions, and whenever available help throughout the course of this thesis research. I am very thankful to Drs. Victor Lo, Tony Bi and Madjid Mohseni for taking time to serve as my committee members and for their valuable advice and constructive suggestions to my research proposal. In particular, I am very grateful to Dr. Victor Lo for allowing me to work in his analytical laboratory to carry out the majority of physical and chemical analyses. Special thanks are due to Dr. Ping H. Liao for his ideas, criticism, and assistance with laboratory analyses to my research work. I leaned a great deal from him. I also would also like to thank Drs. Elisabetta Pani and M. Raudsepp for allowing me to work in their Electro Microbeam/X-ray Diffraction Laboratories and spending many hours teaching me to do the XRD and SEM-EDS analyses and helping to interpret the X-ray diffractgram and EDS results. I am also thankful to Winnie Chan, Olivia Lo and Karen Lau for their enthusiasm for my research work and spending their summer helping me the laboratory work. Finally, I wish to extend my love and thanks to my wife and my son for their understanding and support. I would not be able to endure such a long journey without their encouragement.  Funding by the Natural Sciences and Engineering Research Council of Canada towards this study is gratefully acknowledged.  xii  Statement of Co-Authorship  Five chapters of the thesis have been published, submitted or will be submitted for publication in refereed journals. The co-authors include Lau, A.K. and G., Lemus, and Z.P., Wen. My contributions include:  Identified and designed the research program; Conducted all laboratory work and developed the predictive model; Performed experimental data analyses; The principal author of the manuscripts.       xiii     CHAPTER 1. INTRODUCTION  1.1 General Introduction  1.1.1  Problem Statement  Animal wastes are one of the most significant and growing worldwide problems today. Globally, the estimated volume of animal manure is in the order of 13 billion tonnes per annum (Weiss, 2007). Canadian livestock production generates an estimated 0.36 million tonnes of manure daily. This translated to over 132 million tonnes of manure per year (Statistics Canada, 2001) or 1% of the world’s total, while Canada’s share of population is about 0.5%. The poultry industry sector is the largest producer of surplus nitrogen in the Lower Fraser Valley of British Columbia, Canada. At present, there is close to 300,000 tonnes of poultry manure produced annually (BCMAL 2006). The large amount of poultry manure together with other organic waste materials such as pulp and paper sludge, municipal biosolids, yard waste, food waste, wood waste and crop residues have imposed a significant challenge to environmental protection and sustainable development for the region.  The direct application of manure on cropland near production areas at the present level is not considered sustainable for several reasons. The land-base in many regions of the world for the utilization of these nutrients is finite at present and declining, and land use changes are further reducing the land available for manure spreading. The transportation of manure over long distances to crop producers can contribute climate change due to the use of fossil fuels. While fuel costs are escalating indefinitely, the delivery of large volume of manure to croplands can also become not economically feasible, aside from environmental concerns. In British Columbia, Canada, as the government is developing and implementing the Environmental Farm Planning (EFP) programs and nutrient management plans, the amount of poultry manure nutrients application on croplands can be restricted and decreased by up to 50% in the coming future. Some vegetable farmers are now shifting away from fresh manure due to heightened food safety concerns.  1  As the poultry industry grows rapidly, the competitiveness of the industry could be jeopardized by inadequate and costly manure management practices. Composting has proven to be one of the most efficient and effective ways to manage manure wastes. Through converting raw manure to an organic fertilizer, which has controlled release characteristics, less manure with its uncontrolled and often polluting nature will be applied.  Present policy of the various levels of governments in Canada in organic waste management has evolved towards bioconversion technologies for the multiple purposes of resource recovery and protection of soil, water and air quality (Industry Canada, 2005; Lay et al., 1999). While composting is being perceived as an essential element of the holistic approach in solving the environmental challenges via converting organic wastes into a value-added product, from general perspective of waste management, we are only composting a small percentage of the organic residuals today.  The growth of the composting industry is dependent on both the economics of composting and air emissions that concern health and the environment. Composting, although be considered to be eco-friendly, is associated with “secondary pollutant emissions”. Odor emission is perhaps the biggest and most difficult problem the composting industry has (Ohio EPA, 1999). Even though good management guidelines exist, odor emissions can be unavoidable due to practical issues such as inadequate odor control measures, non-ideal feedstocks, facility downtime and mismanagement (Gage, 2003). Some malodorous compounds are naturally present in wastes and released from composting to the environment regardless of whether or not they become anaerobic. Failure to adequately address odor problems has led to numerous complaints by neighbors and directly caused the closures of a number of large-scale composting facilities (Haug, 1993; Richard, 1996; Ohio EPA, 1999; Gage, 2000).  Ammonia is the most common odor found at composting facilities. Ammonia is toxic, reactive, and corrosive with sharp odor. Ammonia can be smelt when it is in the air at a level higher than 50 ppm. The US Occupation Safety and Health Administration (OSHA) has set a  2  limit of 50 ppm over an eight-hour work day, 40 hour work week for ammonia vapor in ambient air. Exposure at levels of 50-100 ppm ammonia in air can give rise to eye, throat and nose irritation (Busca et al., 2003). According to the research in literature, nitrogen loss through gas emissions during composting ranged from 16% to as high as 77% of the initial nitrogen, with 40% as an average (Martins and Dewes, 1992; Raviv, et al., 2002; Hansen et al., 1989; Kithome et al. 1999; Liang, 2000; Tiquia and Tam, 2000; DeLaune et al., 2004). Most of the nitrogen is lost as ammonia gas (Day et al., 1999; Barrington et al., 2003; Raviv, et al., 2002). Poultry manure composting generally has a high potential for ammonia emission because of high nitrogen concentration in the poultry litter and low C/N ratio (DeLaune et al., 2004). The substantial loss of nitrogen through ammonia emission not only reduce the agronomic value of compost product, but also lead to onsite odor and potential health problems as well as environmental concerns. Excessive amounts of nitrogen deposition in hydrosphere by the route of ammonia emission could be potential environmental problems by causing nitrogen enrichment, acidification of soils and surface waters, and the pollution of ground and surface waters (Ongley, 1996).  In the Lower Mainland/Fraser Valley of British Columbia, which has high population and livestock densities, odor emission control is now considered to be a high priority issue for air quality because of its immediate impact on the population (GVRD, 2005). Indeed, some European countries such as Netherlands, Germany, and Denmark have established stringent regulations with regard to allowable odor emissions.  1.1.2  Research Needs  Traditional measures used by the majority of operators for odor and ammonia emission control are those involving collection and treatment such as thermal oxidation, absorption, wet scrubbing and chemical oxidation (Haug, 1993; Mills, 1996). In recent years, the use of biofiltration is widespread (Goldstein, 2006; Sironi and Botta, 2001). However, these solutions to the odor challenge require additional equipment and space, which could be costly, and do not address the sources of problems. For instance, odor concentrations (dilutions-to-threshold, or ou) were found to be as high as 50,000 ou in municipal solid  3  wastes in some European composting facilities (Haug, 2004), thus it can be very expensive to achieve tolerable working level with odor treatment. Prevention, a strategy that reduces odor emission potential from its sources of generation, could be more cost-effective.  Research efforts have been made to develop preventive means to alleviate composting odors in the past decades. Examples include adopting forced aeration to destruct anaerobic odorous compounds (Elwell et al, 2002a, 2002b); adjusting pH with lime to prevent the formation of hydrogen sulfide (BioCycle, 1999), and using organic materials with large surface area of particles such as peat moss (Mathur et al, 1990), wood ash (Rosenfeld et al, 2003), coal ash (Das, 2000) to adsorb odorant molecules. Though these measures have shown some positive effects on reducing odorous compounds such as hydrogen sulfide, volatile fatty acids (VFA) or volatile organic compounds (VOC), it is obvious that reducing a few specific odorants cannot guarantee a reduction in total odor emitting from composting, since a number of odorus compounds can be generated from composting if optimum operating conditions cannot be maintained from time to time. Besides, no consistent correlation between these parameters has been identified (Powers, 2004c; McCrory and Hobbs, 2001; CIWMB, 2007). Few researchers have directed further experimentation efforts towards the effectiveness of additives on composting odor emissions. Current literature suggests that further research is needed in several areas that are potentially significant to composting odor control: •  ·  The technology of struvite crystallization is relatively new in area of composting  research. While reducing ammonia emission via struvite formation in high moisture food waste composting appears possible, and has a considerable potential because of the increased fertilizer value of compost (Jeong and Kim, 2001), there is no report in the literature on struvite formation in other composting medium. Knowledge gaps exist in various aspects, such as the feasibility of struvite precipitation at the optimal moisture condition (about 55%) without compromising the performance of the composting process; the possibility of forming mineral compounds other than struvite in compost, which may adversely affect struvite formation; the effect of magnesium to phosphate (Mg:P) molar ratio, which could be an engineering tool to facilitate struvite formation; the dynamic changes of water-soluble phosphate in composting process, which could  4  help to understand the kinetics of struvite precipitation; and nitrogen conservation at the maturation stage to understand the overall efficiency of struvite formation in compost. More in-depth research study is needed to fill these knowledge gaps. •  ·  A number of chemical and biological additives have been investigated in animal  production industry to control ammonia and odor emissions. Their general performance to odor control is inconsistent and has been a subject of debate, especially for bio-based additives (McCrory and Hobbs, 2001). Some manure additives such as zeolite and alum have been shown to be effective in reducing ammonia emission from composting (Kithome, 1999; Witter and Kirchmann, 1989a; DeLaune et al., 2004; Lefcourt and Meiseinger, 2001; Venglovsky, 2005; Hu et al., 2006). However, ammonia is of major concern only on-site, as it could be readily diluted in the atmosphere and does not pose an odor problem off-site. Researchers have not directed further experimentation efforts towards the effectiveness of additives in reducing pervasive odors other than ammonia, and in terms of odor detectability as perceived by human. Even though others have studied the use of additives in treating manure in pits or lagoons in which anaerobic bacteria dominate and have shown positive effects for reducing offensive odors, further research is required to understand their effectiveness in odor control during composting whereby aerobic microorganisms prevail. The combination of yeast and zeolite as one additive for composting odor control has not been investigated. •  ·  It is generally recognized that operating conditions are critical to composting odor  generation, emission and control (Haug 1993; Binner el al., 2002; Gage, 2003; CIWMB, 2007; Goldstein, 2007). While the majority of past studies concentrated on chemically identifying individual or groups of malodorous compounds (via gas chromatography or detection tubes) in relation to operational conditions (Michel and Reddy, 1998; Edwell et al., 2001; Fraser and Lau, 2000; CIWMB, 2007), the information and knowledge thus generated do not represent the experience of odor sensation as perceived by human because the perception of a mixture of odorants is very different from how each chemical would be perceived individually (Power, 2004; Mills, 1995). So far, published information about the quantitative relationships between human perceived odor concentration and key operational parameters such as the type of substrate, aeration rate and moisture content are very limited. Binner et al. (2002) tested the composting  5  conditions for preventing odors, but only the odor profile relative to oxygen level was reported. Other olfactory odor studies with respect to composting operation conditions were sporadic and rarely provided odor profiles (Bruce, 1998; Nobel et al., 2001). The lack of systematic experimental studies on odor emissions in terms of olfactometry provides little information for developing odor predictive model, which is useful for regulation purposes when used along with odor dispersion modeling, since at present olfactometry remains to be the industry and regulatory standard. In addition, some studies suggested that aeration increases odor emission rates (Walker, 1993; Fraser and Lau, 2000; Schlegelmilch et al., 2005); while others reported that aeration might oxidize intermediate odorous compounds generated from anaerobic niches due to insufficient oxygen, thus reducing odor concentration (Haug, 1993; Richard, 1996; Michel and Reddy, 1998; CIWMB, 2007). Technically, both claims are correct because very high airflow rates could actually increase odor emission rate though odor concentration could be somewhat reduced. Therefore, it is desirable to determine the optimal aeration strategy in terms of minimizing odor emission rate, which constitutes a key parameter in odor dispersion model. •  ·  Predicting odor generation and movement off-site can yield valuable insight into  composting site selection, neighborhood impacts, and the required level of odor control technology (Wu, 2000). Several studies attempted to use dispersion modeling with measured source odor data to predict the distributions of odor emanated from composting facilities (Wu, 2000; McGinley, 2005; William and Servo, 2005). To date, no odor source generation models have been developed for the prediction of odor emissions from composting in terms of odor concentration (strength) as perceived by human. It is desirable to develop such model that could be linked to a dispersion model to predict the odor concentration in the neighborhood of composting facilities. This is particularly useful for new facilities; even for existing composting facilities, sampling and measuring odor sources are time consuming and costly.  6  1.1.3  Objectives  In light of the problems that the composting industry is facing with and the research needs identified above, the overall goal of this study is defined to be developing preventive means to minimize the ammonia and odor emissions during the active phase of composting, and maximize nitrogen retention in finished compost. The preventive techniques to be chosen were on the bases that have the potential of producing value-added compost as high-grade organic fertilizer and soil amendment derived from compost. The specific objectives are: •  To study the feasibility of precipitating ammonium into struvite in poultry manure composting media to reduce ammonia emission and enhance nitrogen conservation, thus increasing the market value of compost product.  This study differs from the previous trials of struvite precipitation in compost in several aspects. Firstly, it examines a different type of organic waste. Jeong and Kim (2001) and Jeong and Hwang (2004) tested with food waste that contains little phosphate and magnesium, hence struvite formation in compost relied largely on chemical addition. This study used poultry manure that is rich in nitrogen and phosphorus, and has greater potential for ammonia emission, thus it is of more practical and economic significance. Secondly, previous studies only verified struvite presence with scanning electronic microscopy (SEM), while this study examines the possible presence of all potentially formed minerals with X-ray diffraction (XRD) in addition to SEM to understand the mechanism of chemical reaction in the composting process. Thirdly, previous studies were only limited to ammonia emission and nitrogen conservation due to struvite formation during the active phase of composting. This study expands the investigation to the effects of other parameters, including the Mg:P molar ratio, changes in ortho-phosphate over time; stability of nitrogen retention in compost after the active phase; and compost salinity as affected by salts addition. Studies of these parameters are of theoretical and practical importance in struvite formation and its effects on compost products. •  To examine the effectiveness of chemical and biological additives in the form of yeast, zeolite and alum in reducing ammonia and odor emissions.  7  The hypotheses to be tested are that zeolite would have positive effects on the reduction of ammonia because of its highly adsorptive capability; and that the enzymes in yeast can proliferate the microorganisms which may help to accelerate the degradation of slower degradable fractions of organic matter in the feedstock, and may have deodorization capability at the same time. Therefore, the combination of yeast and zeolite may offer a potential solution for enhancing composting efficacy while reducing ammonia and odor emissions. For alum, when it dissolves, ammonium can react with sulphate to form ammonium sulphate, thus reducing ammonia emission while enhancing the fertilizer value of finished compost. Moreover, there are no previous reports on the effectiveness of alum on odor control for aerobic composting. •  To determine the effects of composting operating conditions on odor emissions during the active phase of composting as a first line of odor defense.  The results of quantitative correlation between odor emission and key operational parameters will provide the basis for developing a predictive odor model. •  To develop a predictive odor model that can simulate odor emission for various operation conditions of aerobic composting.  Such a model would enable composting facilities to modify their design, determine the extent of odor control, and manage operations to reduce the odors. Moreover, this model may combine with odor dispersion modeling, and it may potentially offer a solution to the off-site odor problems by providing objective means for sitting new composting plants, establishing operating strategies, determining the extent of odor control, demonstrating compliance with regulations, and assessing the impacts on surrounding communities.  1.1.4  Thesis Organization  To address the objectives outlined above, a series of experiments were designed and conducted in this thesis study. The results are presented in eight chapters. Chapter 1 is a general introduction to identify the problems and research needs, and to define the objectives of the study. A further section in this chapter presents a detailed literature review associated  8  with this research, with emphasis on the state of art in composting odor theories and practices. Chapter 2 depicts the bioreactor system with the details of system control and data acquisition, and demonstrates the performance of the bioreactor system. An upgrade olfactometer in compliance with the international odor quantification standard (EU Standard) and the calibration and accuracy of the instrument are also presented in this chapter.  Chapters 3 through 5 present the detailed experimental results of this thesis study. A throughout experimental study on reducing ammonia emissions via struvite formation is presented in chapter 3. Chapter 4 details the study results on minimizing the odor and ammonia emissions with biological and chemical additives in the forms of yeast, zeolite and alum. In Chapter 5, a summary of the experimental results on odor emissions with respect to key operating parameters is presented. The results from this chapter also serve as the basis for development of odor predictive model. An empirical odor predictive model was developed and presented in Chapter 6. Finally, Chapter 7 summarizes the main findings and contributions from this thesis study. Recommendations for further research are also presented in this chapter.  9  1.2  Background Research  1.2.1 Characteristics of Odors Odors and Their Potential Environmental and Health Effects Odors are defined as sensations resulting from the reception of a stimulus by the olfactory sensory system, which consists of two separate subsystems: the olfactory epithelium and the trigeminal nerve (WEF & ASCE, 1995). Substances that stimulate the human olfactory sensory system are known as odorants. Most odors discharged to atmosphere from industrial and agricultural sources consist of complex mixtures of many odorants namely odorous chemical compounds. The detection of odors raises health concerns simply because citizens become aware of exposure to chemicals. Odor is considered as an important environmental pollution issue because of its immediate impact on the population. In general, odors are ranked among the major generators of public complaints to regulatory agencies throughout North America. The Greater Vancouver Regional District (GVRD) received in excess of 2000 complaints concerning air pollution from public every year, and of these, approximately 65% are related to odors (GVRD, 2005). Attention to odor as an environmental nuisance has been growing as a result of increasing awareness of people’s need for a clean environment. Traditionally, odor has been regarded primarily as a nuisance issue, but the health effects of odors are now receiving rigorous scientific study (Schiffman and Williams, 2005). Some negative effects of odors on human health and welfare have been reported. Odors may affect well-being by eliciting unpleasant sensations, by triggering possibly harmful reflexes and other physiologic reactions, and by modifying olfactory function (NRC, 1979). Offensive odors are capable of causing nausea, vomiting and headache; inducing shallow breathing and coughing; upsetting sleep and stomach; irritating eyes, nose, and throat; curbing appetite, impairing nutrition and curtailing water intake; and interfering with the enjoyment of food, home, and external environment (Schiffman et al., 2000; Nimmermark, 2004; Dalton, 2003; GVRD, 1993; NRC, 1979). The exposure to some odorous compounds may also lead to 10  decrease in heart rate, constriction of blood vessels of the skin and muscles, release of epinephrine, and even alterations of the size and condition of cells in olfactory bulbs of the brain (NRC, 1979). Many odorous gases are also volatile organic compounds (VOCs), which, along with oxides of nitrogen (NOx), are known precursors for the formation of ground level ozone, a major component of urban smog (GVRD, 2005). Large amount of ammonia is evolved from composting process and volatilizes into environment, which is now being viewed as an air quality problem.  Sensory Properties of Odors Odors are generally characterized using sensory properties. These include detectability, intensity, quality, and hedonic tone. The detectability (or threshold) of an odor refers to the theoretical minimum concentration of odorant that arouses an olfactory response or sensation in a specified percentage of the test population. Threshold concentration values are not fixed physiological facts or physical constants but are statistically representing the best estimate value from a group of individuals (CEN, 2003). The odor threshold concentration is determined by diluting the odor to the point where 50% of the odor panel or test population can correctly detect the odor. The ability to detect odor by panelist or test population is influenced by physiological factors and criteria used in producing the response. The sample presentation parameters such as flowrate of gaseous, odorous sample also introduce variations. However, a concentration range exists below which the odor of a substance will not be detectable under practice circumstances, and above which individuals with a normal sense of smell would readily detect the presence of the odor (ASTM, 2004). Odor intensity is defined as the perceived strength or magnitude of the odor sensation. It measures how strong the odor smells. Odor intensity is evaluated by comparing the intensity of a standard odorant (e.g. n-butanol) at various concentrations. Quantitatively, intensity increases as a function of odor concentration. The relationship of between intensity and  11  concentration may be described as theoretically derived logarithmic function according to Weber and Fechner method (CEN, 2003): S = k w ⋅ log I / I o Where: S  is the perceived intensity of sensation (theoretically determined),  I  is the physical intensity (odor concentration),  Io  is the threshold concentration,  kw  is the Weber-Fechner or Weber ratio.  Or as a power function in accordance with Stevens Psychophysical Law (Stevens, 1960):  S = K ⋅In Where: S  is the perceived intensity of sensation (empirically determined),  I  is the physical intensity (odor concentration),  n  is the Stevens’ exponent,  K  is a constant.  Odor quality is a measure of the character of the odor, i.e. what the odor smells like. It is usually expressed in descriptors. Odor quality has not been routinely measured in odor studies conducted at composting and wastewater facilities (Haug, 2004). Hedonic tone is a category judgment of the relative pleasantness or unpleasantness of the odor. The degree of pleasantness or unpleasantness is determined by each panelist’s experience and emotional association. Hedonic tone is also not routinely measured in composting studies. This is because even pleasant odors become a source of annoyance and complaint when they occur as odorous air pollution.  12  Odorous Compounds from Composting Odors are mixtures of numerous gases, vapors and dusts. More than 300 odorous compounds have been identified in odors from animal manure facilities (Schiffman et al., 2001; Sweeten et al, 1994; O’Neil et al. 1992). The main groups of odorous compounds that are of major concerns from composting odor appear to be sulfur compounds, nitrogen compounds, and volatile fatty acids (Haug, 1993; Day et al., 1999; Goldstein, 2002). A list of these odorous compounds and their detection limits is provided in Table 1.1.  Sulfur Compounds Total reduced sulfur such as hydrogen sulphide, methyl mercaptan, dimethyl sulphide and dimethyl disulphide are identified as major malodorous compounds from composting. During composting, organic sulphurous compounds are first converted to sulphides as an intermediate production. Under anaerobic conditions, little sulfur can be oxidized. Instead reduction of the sulfur takes place, leading to the production of volatile organic sulphides. Reduced sulfur gases smell horribly. Two compounds detectable at very low concentrations are hydrogen sulfide (0.7 µg/m3) and methane thiol (0.04 µg/m3) (Goldstein 2002).  Nitrogen Compounds Ammonia is produced from either aerobic or anaerobic decomposition of proteins and amino acids. It is the most common odor problem at composting facilities. Approximately 99% of nitrogen emission is ammonia during biosolids composting. Ammonia has a sharp and irritating odor. Breathing levels of 50-100 ppm ammonia in air can give rise to eye, throat and nose irritation (Busca et al., 2003). However ammonia has relative high threshold odor concentration, and can be diluted rapidly to below detection. Amines, which are produced during anaerobic decomposition of proteins and amino acids, are also nitrogen-based compounds with bad smelling, and can also have high toxicity (Busca et al., 2003). Trimeththyl amine has a pungent and fishy odor. It can be detected at very low  13  Table 1.1 Major odorous compounds from composting process* Class of Compounds Smell Detection limit (µg/m3) Sulfur compounds Dimethyl disulfide  Rotten cabbage  0.1  Dimethyl sulfide  Rotten cabbage  2.5  Carbon disulfide  Rotten pumpkin  24  Hydrogen sulfide  Rotten egg  0.7  Pungent sulfur  0.04  Medicinal  27  Fishy  0.11  Sour (vinegar)  1019  Propionic acid  Rancid  28  Butyric acid  Putrid  0.3  Methane thiol  Nitrogen compounds Ammonia gas Trimethyl amine  Volatile fatty acids Acetic acid  * Modified from Goldstein (2002)  14  Structural representation  concentration (0.11 µg/m3). Trimeththyl amine is also difficult to biodegrade because microbes cannot easily break it apart (Goldstein, 2002).  Volatile Fatty Acids Volatile fatty acids are one of the most significant odorous compounds from manures. VFAs smell like vinegar and are by-products of anaerobic decomposition. VFAs decompose rapidly when air passes through them (Cooper et al., 1978). Rapid, early composting removes most of the VFAs through bacterial consumption, which could substantially reduce the potential for later release (Elwell et al., 2001; Elwell et al., 2003). Other odorous compounds found in composting facilities include ketones, aldehydes, alcohols, terpenes, guaiacol, and so on. These odorous compounds could be contained in feedstock or formed as intermediates during composting. Relatively they may not be very significant if optimum composting conditions are maintained (Goldstein, 2002).  1.2.2 Odor Measurements Odor measurement is a complicated and difficult task due to the complex composition of odors, variable sources, effects of environmental factors, and varying human perceptions of offensive smells (Chapin et al., 1998). Odors are generally measured using analytical or sensory methods. Analytical Methods determine the components and their concentrations, whereas sensory methods measure human responses to the odors. Figure 1.1 is a categorization of the analytical and sensory methods.  Analytical Methods Analyzing emission of specific odorants is a means of measuring odor that does not rely on human sensory perception but instead aims at identifying surrogate compounds that are major contributors to the perceived odor. From the standpoint of odor control, it might be useful to identify the surrogate compounds, so that these few compounds can be targeted with control strategies (Powers, 2004c). Attempts to correlate specific odorous compounds with  15  Odor Sample  Analytical Testing  GC Analysis  GC/MC Analysis  Sensory Testing  Monitoring Individual Compounds  Odor Intensity (Strength)  Odor Threshold (Detection)  Odor Quality (Descriptor)  Hedonic Tone (Annoyance)  Figure 1.1 Odor measurement methods. GC = gas chromatography; MS = mass spectrometry (Modified from WEF and ASAE, 1995)  16  olfactometric odor concentration have been reported in a few studies (Defoer et al., 2002, Noble et al. 2001, Sato et al., 2001, Elwell et al. 2001, Elwell et al., 2003). Examples of the surrogate compounds include total reduced sulfur compounds, ammonia, and volatile organic compounds. Volatile organic acids concentrations, volatile fatty acids, and phenolics have also shown some sort of correlation to odor sensory measurements (Zahn, 1997, Zahn et al., 2001). The analysis of chemical odorants can be accomplished using a variety of analytical methods ranging from wet chemical methods, colorimetric detector tubes to GC/MC techniques. Wet chemical method involves several absorption techniques. This method is frequently used for measuring hydrogen sulfite and ammonia that are captured by passing measured volume of exhaust gas through trapping solutions in series. From the trappings, individual compounds can be measured with standard methods. Wet chemical method is also capable of measuring odorous compounds such as methyl mercaptan, dimethyl sulfite, dimethyl disulfite, aldehydes, and amines (WEF and ASCE, 1995). It is suitable for continuous monitoring. Colorimetric detector tubes can provide useful information by identifying specific gaseous compounds that emit odors instantly. However, this method usually has a large error range and can only be used for discrete sampling purpose. Odor emissions often consist of a complex mixture of numerous odorous compounds. In order to identify and quantify the constituents of odor, gas chromatography coupled with mass spectrometery (GC/MS) technique is most frequently used, where chemical compounds are separated based on their volatility. However, the interpretation of results in this method is complicated because odors that are equal in concentration may not be equal in offensiveness or intensity. Furthermore, two odors of equal concentration may not be perceived as having different intensities (Powers, 2004b). Electronic nose analysis with a sensor array is an emerging new tool for odor evaluation. The electronic noses have been developed in attempt to mimic the human response to odor, and can be calibrated to recognize a specific odor and the relationship between the response of an  17  electronic nose and odor concentration. While this technology is potentially useful in providing an immediate, objective, non-sensory method, it appears currently to provide little practical use in composting industry (Zhang et al., 2002, Powers, 2004b).  Sensory Methods Perception of a mixture of odorants is usually very different from how each chemical would be perceived individually. As a result, odor sensory methods, instead of instrumental methods, remain the most valid methods for odor evaluation at the current time (Sweeten et al., 1994; WEF & ASCE, 1995; Chapin et al., 1998; Powers, 2004b). Several sensory methods involving the human nose have been developed and used worldwide. The main techniques used are absorption media, olfactometry and sentometry. The most popular method is olfactometry. The major advantage of olfactometry is the direct correlation of the odor to human’s sensitive sense of smell. Moreover, olfactometry analyzes the complete gas mixture so that contribution of each compound in the odor is included in the analysis. The major challenges of sensory methods are variation in performance of panelists and odor fatigue. Proper procedural protocol is very important. The most recent advance is the development of “European Standard” (CEN, 2003). It is a comprehensive document providing new guidelines on a number of areas: the scope and field of application, performance quality requirements, quality requirements for dilution apparatus (olfactometer and pre-dilution equipment), materials and gases, screening of panel members, presentation of odorants to assessors and objective data analysis. The use of this standard has improved the reliability of odor concentration analysis and reduced the degree of subjectiveness of the human perception of odors.  1.2.3 Odor Emissions from Composting Mechanism of Odor Generation and Emissions Odor generation from composting is a complex process and still poorly understood. In composting, there are always both aerobic and anaerobic niches. It is generally believed that  18  offensive odors result from anaerobic biochemical metabolism process, which produces many intermediate compounds that are odorous. Figure 1.2 illustrates the anaerobic decomposition process of complex organic substrate such as manure. The main constituents of typical composting materials, such as manures, include carbohydrate, lipid and proteins. Decomposition of carbohydrate produces glucose first. Under anaerobic conditions, glucose is then broken down producing intermediate compounds: alcohols, aldehydes, ketones and organic acids. These intermediate compounds are odorous before further transforming into carbon dioxide, methane and water.  Complex substrate  Carbohydrate  Lipid  Protein  Alcohols, aldehydes, ketones, organic acids  Fatty acids, alcohols, acetate, organic acids  Petones, petides, amino aids, organic acids, sulfides, mercaptans, phenols  Figure 1.2 Substrate breakdown process (adopted from Powers, 2004a). Lipids or fats are easily degradable (Haug, 1993; Lefebvre et al., 1998, Lemus, 2003). Bacteria use fats as energy source, hydrolyzing them first to the corresponding long-chain fatty acid and alcohols. These acids, along with those produced in the deamination of amino acids, undergo further breakdown in which acetic acid is produced (Powers, 2004c). Volatile fatty acids such as acetic, propionic, iosbutyric, butyric, and isovaleric acids are believed to  19  be the most significant odorous compounds in manures (Elwell et al, 2001; Wiles et al., 2001; Kuroda et al., 1996). Anaerobic degradation of protein is regarded as the main source of malodors from composting. The breakdown of protein proceeds to ever-simpler proteses, peptones, amino acid and further to ammonia and volatile organic acids. Various sulfides and mercaptans could be formed as a result of catabolism (Powers, 2004c). Reduced sulfur gases have horrible smells and can be detected at very low concentration. Offensive odors generally result from anaerobic metabolism. However, odors may also be generated by aerobic decomposition although aerobic intermediates may sometimes be less obnoxious. Some malodorous compounds naturally present in plants and animal wastes may also objectionable regardless of whether or not they become anaerobic (CIWMB, 2007). Odor emissions from composting process involve quite complicated mechanisms, and are generally believe to be a combined result of microbial activity and contribution of environmental factors. In composting process, odorous compounds with low molecular weight (typically 17-152 g/mole) and high volatility are generated as intermediate or end products by either aerobic or anaerobic bacteria (Yuwono et al., 2004). The low molecular weight compounds generally have high potential for emission to atmosphere (WEF & ASCE, 1995). However, the volatility of molecules is not solely determined by their molecular weight. The strength of the interactions between the molecules also plays an important role, with non-polar molecules being more volatile than polar ones (Yuwono et al., 2004). Anaerobic bacteria generate malodors. On the other hand, odors generated by aerobic bacteria are not offensive but rather musty or earthy (White et al., 1971). Composting odors are primarily dependent on the type and age/state of the feedstocks. For instance, manure odors are caused by chemicals, primarily volatile fatty acids (VFAs), indolics, phenolics, and sulfur compounds. Swine manure contained much higher concentrations of the longer chain (most offensive) VFAs than with cattle manure (Elwell et al., 2001). Sewage sludge with wood based amendments, odor concentrations from 100 to  20  over 1000 ou are typical, while municipal solid waste composting facilities have reported odor concentration over 20000 ou (Haug, 1993). High odor emissions normally occur during the high active phase of composting (Bidlingmaier, 1993; Benedict et al., 1988; Day et al., 1999). As the process progresses, some odorous compounds such as amines and VAFs decrease (Goldstein, 2002). Although odors continue to release throughout the high rate composting process, the total odor intensity or threshold is significantly low at the end of active phase.  Factors Affecting Odor Emissions Odor emission from composting can be affected by many factors such as substrate characteristics (C/N ratio, bioavailability, moisture, particle size, and pH), environmental conditions (temperature, moisture, oxygen concentration), and operational parameters (aeration, temperature), but key factors that can be easily controlled to promote aerobic conditions and reduce odor emission are aeration, moisture content, temperature, C/N ratio and particle size distribution (Cornell Composting, 1996; Day et al., 1999; Ohio EPA, 1999). Aeration is probably the most influential factor to odor emissions. The basic principle for odor reduction is that aeration provides dissolve oxygen to aerobic bacteria such that they can actively decompose the odorous compounds, hence achieving odor reduction (Zhu, 2000). Aeration also helps to maintain temperature and moisture at the appropriate levels, which in turn positively influence odor emissions. Sufficient air (oxygen) promotes aerobic conditions and reduces potential of anaerobic odors that are more offensive than aerobic odors. On the other hand, aeration also serves as a driving force to transport odorous compounds to outside environment, thus increase odors (Fraser and Lau, 2000; San Diego State University, 2007). Aeration depends on particle sizes, moisture and density of the compost mix as well as aeration rate in the case of forced air used. Small particles, excess moisture and overly dense material will impede aeration and increase the potential for odor generation (Epstein and Wu, 2000; Ohio EPA, 1999).  21  Moisture content is a very critical factor as it affects many aspects including free air space, microbial activity, heat and mass transports, and nitrogen dynamics. Oxygen diffusion is 6000 to 10,000 times greater in air than in water. Excess moisture fills the pore space, and thus creates larger water filled zones between particles, which slow oxygen diffusion and result in anaerobic clumps. With most composting materials, as moisture content increases beyond 60%, the pores will rapidly fill and anaerobic conditions will result (Cornell Composting, 1996). Temperature is also an important factor to odor emissions. The concentration of a particular odorous chemical in air emission is dependent upon its vapor pressure, which is a dependant of composting temperature. It is reported that the concentrations of some of the familiar odorous compounds such as pinene and ethyl butyrate increase approximately 10-fold when temperature of the compost pile increase from 20 to 65 oC (Day et al., 1999). Therefore, high temperature increases the concentrations of chemicals in the air, and as a result, may increase odor emissions. On the other hand, for forced aeration composting system, composting temperature is generally regulated by aeration. More air has to be applied to lower the high temperature to setpoint, which means that bigger volume of exhaust odorous air will be generated. The overall effect of high composting temperature on odor emissions is not clear (Epstein and Wu, 2000). C/N ratio determines nutrient balance. A lower ratio (too much nitrogen) can result in higher ammonia emission, which is foul smelling. Excess nitrogen can also cause accelerated microbial growth, which will rapidly use up the available oxygen, resulting in anaerobic conditions (Ohio EPA, 1999). A higher C/N ratio, however, may result in insufficient nitrogen for optimum growth of microorganisms and slow down the composting process. Inadequate particle size distribution can lead to anaerobic conditions. Larger particle sizes and loosely packed material can make a compost pile highly porous, which increases the airflow and reduces the accumulation of moisture (Ohio EPA, 1999). Small particle sizes reduce the number of large pores and increase the likelihood that oxygen will need to diffuse  22  a long way through small pores. Small particles can also interact with high moisture levels to reduce oxygen transport and generate anaerobic odors (Richard, 1996).  1.2.4 Principles of Odor Control A wide variety of measures for composting odor control have been investigated and some of them are used in practice. The main principles behind these measures are inhibition of anaerobic condition, biological or chemical transformation, and treatments. Odor treatment includes a wide array of means such as chemical stripping, thermal destruction, and biofiltration. A full description of the odor treatment measures for composting is beyond the scope of this thesis study. Haug (1993) and Schlegelmilch et al. (2005) gave excellent review on this subject. The following reviews the use of preventive means in control of composting odor emissions.  Inhibition of Anaerobic Condition Since malodor generation is mainly the result of biological anaerobic decomposition, inhibiting anaerobic condition could minimize the opportunity for odorant generation, thus reduce the odor emission potential. Oxygen availability is the key to prevent anaerobic conditions. It was found that anaerobic pockets start forming, which then leads to formation of odorous compounds VFAs, ketones, aldehydes and potential methane, when oxygen content drops below 17 percent (Goldstein, 2002). Forced aeration is probably the most viable choice to prevent anaerobic odors technically as well as economically. The basic principle of aeration for odor control is that aeration provides dissolve oxygen to aerobic bacteria such that they can actively decompose the odorous compounds, hence achieving odor reduction (Zhu, 2000). Some studies suggested that increased aeration generally results in decreased concentration but an increase in total emission of odorous compounds (Walker, 1993; Fraser and Lau, 2000). In contrast, a study conducted by Elwell et al. (2001) gave a mixed result. Their data  23  showed that a 76% reduction in airflow resulted in a substantial reductions in acetic, propionic and butyric acids release, but isobutyric, isovaleric and valeric acid emissions increased. While it has been recognized that aeration strongly affects odor concentration, its net effects on odor production is still not clear.  Biological or Chemical Transformation Minimizing the opportunities of odorous volatile compound formation and volatilization can reduce odor emissions. This could be achieved by using chemical or biological transformation. Biocatalysts purport to degrade odorous compounds via biologically generated enzymes (Richard, 1996). A catalyst facilitates a reaction without itself being permanently changed by the reaction, and thus each enzyme can act on many molecules of an odorous compound before it is eventually degraded (Richard, 1996). The role of yeast in environmental biotechnology has received increasing attention in recent years (Grommen and Verstraete, 2002). Enzymes that contain β-glucan and β-glucosidase (cellulase) are present in certain yeasts, which can readily degrade cellulose, a major component of plant and animal wastes. In a study of grass composting with olive mill wastewater added, a high correlation was observed between phenol degradation and βglucosidase activity (Grego et al., 2002). Yeast might be working at the biochemical level, shifting microbial population. Tequia et al. (2002) found that the population of fungi and actinomycetes were most positively correlated with the activities of the α–galactosidase and β-glucosidase enzymes. Kim et al. (2002) isolated and identified yeasts (Candida spp.) from soil and compost sources, which were effective to reduce odorous VFAs from pig feces. Wood ash incorporation in composting has been shown to be effective in composting odor control. Rosenfeld et al. (2002) investigated the effectiveness of high carbon wood ash to control odor emissions from yard waste composting. They found that significant reduction in dilution-to-threshold values, but ammonia emission were increase due to the strong alkaline pH of wood ash. A similar research conducted at a biosolids composting facility in Maryland also found that the use of wood ash could reduce odor emission (Composting research  24  update, 1999). Some other adsorption materials such as zeolite and some chemical amendments like lime and salts have also been tested in reducing ammonia emission (Kithome, 1999; Witter and Kirchmann, 1989a; DeLaune et al., 2004; Lefcourt and Meiseinger, 2001; Venglovsky, 2005; Hu et al., 2006). Zeolite has been shown positive effectiveness in reducing ammonia odor, but its effectiveness on offensive odor is less conclusive (DeLaune et al., 2004; Venglovsky, 2005). Alum was proved to be effective to control of ammonia and odor emissions from animal manure storage, no reports were found in the literature on its effectiveness to control of offensive odors from composting (McCrory and Hobbs, 2001).  1.2.5 Ammonia Emission and Control Ammonia is the most common odor found at composting facilities. Unlike other malodors, ammonia odor has a low dilution to threshold ratio. Gas ammonia can disperse easily because it is lighter than air. Therefore ammonia is more regarded as an onsite odor rather than an off site problem (Richard, 1996). For these reasons, ammonia odor is treated separately from other malodorous odors, and thus has its own set of management options. This review concentrates on the preventive measures of ammonia emission, as they are the focus of this thesis study.  Ammonia Volatilization Ammonia can be formed either aerobically or anaerobically. In typical compost feedstock, besides predominant carbohydrates, nitrogen is also presented in the form of proteins. Under aerobic conditions the proteins are broken down as follows (Day et al., 1999): Proteins ⇒ peptides ⇒ amino acids ⇒ NH4 species ⇒ bacterial proplasm + NH3 In simplistic chemical terms, the overall reaction can be written as follows: CH3CHNH2COOH + 3O2 ⇒ 3CO2 + NH3 +2H2O  25  Consequently, ammonia is a direct product of oxidative composting of proteins. The magnitude of ammonia volatilization is strongly affected by pH. Figure 1.3 shows the general equilibrium relationship in solution as a function of pH. NH3 (gaseous ammonia) and NH4+ (aqueous ammonium ion) are in equilibrium at a pH of about 9. Higher pH forces more NH4+ into the gas form and has a much higher potential to be volatilized. Although the volatilization of ammonia during the course of composting cannot be avoided, its release to atmosphere could be minimized by manipulating air supply, carbon to nitrogen ratio, pH, as well as using preventive means such as adsorption, chemical transformation and precipitation techniques.  Figure 1.3 Relative concentrations of NH4+ and NH3 in solution (adopted from Richard, 1996)  Preventive Control of Ammonia Emission by Adsorption and Chemical Transformation Various amendments have been studied to adsorb NH3 and NH4+ or chemically transform them before releasing. Witter and Kirchmann (1989a) used peat, zeolite and basalt to reduce ammonia volatilization via adsorption. All these adsorbents were proved to be effective in adsorbing ammonia in gases, but less effective when mixed with manure. Liao et al. (1997) 26  also reported that peat moss is effective in reducing ammonia emission from a full-scale fish waste composting system. Kithome et al. (1999) investigated reducing NH3 losses from poultry manure composting with various additives including natural zeolites, clay, coir (mesocarp of coconut fruit), chloride and sulfate salts of calcium and magnesium, and alum. They found that a layer of 38 percent zeolite placed on the surface of composting material reduced ammonia emission by 44%, whereas 33 percent coir amendment placed on the surface of composting material reduced ammonia emission by 49%. The advantages of using either zeolite or peat are that they are nonhazardous and can act as good soil conditioners. Alum and phosphoric acid have both been shown effective in reducing NH3 volatilization from composting. Alum produces H+ when dissolves, which reacts with NH3 to form NH4+. NH4+ can then react with the sulfate to form ammonium sulfate. Phosphoric acid can directly react with NH3 to form ammonium phosphates. DeLaune et al. (2004) reported that using alum and phosphoric acid could reduce NH3 volatilization from composting poultry litter by as much as 76% and 54%, respectively. Witter and Kirchmann (1989b) stated that adding calcium and magnesium salts to control the pH in compost could reduce ammonia emission. The results from Kithome et al. (1999) indicated that mixing 20% CaCl2 with composted poultry manure reduced ammonia emission by 90%, whereas 20% alum reduced ammonia emission by 26%. But CaSO4 and MgSO4 were found ineffective in reducing ammonia emission.  Preventive Control of Ammonia Emission Via Struvite Precipitation  Characteristics of Struvite Struvite is an inorganic crystalline mineral consisting of magnesium, ammonia and phosphate in equal molar concentrations (MgNH4PO4·6H2O). Table 1.2 summarizes the properties of struvite (Barthelmy, 2002; Wu and Bishop, 2004). Pure struvite has white color, but it can also be yellowish or brownish white when other elements are incorporated. Struvite crystal  27  can be generally identified by X-ray diffraction (XRD). Since struvite crystal has a distinctive orthorhombic structure, it can also be identified by scanning electron microscope. A number of studies have been conducted to measure the struvite solubility. Some solubility data for struvite were summarized by Doyle and Parsons (2002) and are listed in Table 1.3. Because the solubility of struvite in water is extremely low, this special property makes it particularly interesting in using as a slow release fertilizer. Bridger (1968) concluded that struvite is a highly efficient source of magnesium, nitrogen and phosphorus for plants. Furthermore, its release rate to plants can be controlled by use of different granule sizes, and it can be applied in massive doses without burning. According to the author, the release of nutrients from struvite in soil is determined by bacterial nitrification rather than solubility. Several authors have shown that struvite is a good fertiliser for variety of grasses, ornamental plants, field crops, citrus as well as planting forestry seedlings and transplants Their studies indicated that of struvite proved better than conventional NPK fertilizer, fused magnesium phosphate and triple super phosphate fertilizer (Bridger et al., 1962; Leiser; Grambell et al., 1964; Greaves et al., 1999).  Factors Controlling Struvite Precipitation Struvite (MgNH4PO4.6H2O) is formed as Mg2+, NH +4 , and PO 34− react in 1:1:1 molar ratio in aqueous phase. The simplified reaction equation is as follows: Mg2+ + NH +4 + PO 34− + H2O ⇔ MgNH4PO4·6(H2O)  (2.1)  It is generally believed that struvite precipitation consists of two stages: nucleation and crystal growth. Nucleation occur when constitute ions aggregate together to form crystal embryos. Crystal nuclei grow to crystals until equilibrium is reached (Ohlinger et al., 1999). Nucleation is primarily affected by the degree of supersaturation, and therefore is a reactioncontrolled process. The crystal growth rate, however, is transport controlled (Nelson et al.,  28  2002). Struvite will be favored to precipitate when the product of Mg2+, NH +4 and PO 34− activities (IAP) exceeds the equilibrium ion-activity product of 7.08 × 10-14. The precipitation of struvite is primarily controlled by pH, degree of supersaturation, temperature and presence of other ions in the solution such as calcium (Doyle and Parsons, 2002). Figure 1.4 shows the main magnesium and phosphate species as a function of their total concentration plotted against the pH (Mijangos et al., 2004). As seen in figure 1.4, struvite precipitation may not occur when the total concentration of their constitutes is below 0.015 mole/L. At pH<6, newberite (MgHPO4·3H2O) probably is the main species precipitated. Struvite precipitation may proceed over a range of pH from 7 to 11. Abbona et al. (1982) found that struvite solubility decrease rapidly at a pH of 7.5. Research on phosphate removal from wastewater stream via struvite precipitation indicates that pH value of 7.5 could result in definite reduction in phosphate. A pH value of 8.0 is desirable (Momberg and Oellermann, 1992). As seen in Figure 1.4, Mg3(PO4)2 may also be formed in alkaline conditions especially when pH value is higher than 11. Since struvite formation is highly pH dependent, chemical and mechanical measures have been used to raise pH. NaOH has been suggested as the most effective chemical mean for the pH adjustment, but dosing base will significantly increase the cost. Mechanical means offer the advantages of being relatively cheap and nonhazadours. Increasing turbulence leads to CO2 stripping, an increase in pH, and hence an increase in struvite precipitation. Burns et al. (2001) found that mechanical stirring during agitation of swine manure holding pond prior to land application can effectively raise pH by one point over two hour period. Aeration, which strips CO2 and thus alters the carbonate chemistry, can also increase pH (Doyle and Parsons, 2002).  29  Table 1.2 Properties of struvite Parameters  Description  Chemical Formula  Mg (NH4)PO4·6(H2O)  Molecular Weight  245.41 gm  Composition  Magnesium 9.90 % Mg 16.42 % MgO Phosphorus 12.62 % P  28.92 % P2O5  Hydrogen  1.64 % H  Hydrogen  4.93 % H  Nitrogen  5.71 % N 10.61 % (NH4)2O  Oxygen  65.20 % O  44.05 % H2O  Color  White, Yellowish white, Brownish white  Density  1.7  Diaphaniety  Transparent to Translucent  Habit  Euhedral Crystals - Occurs as well-formed crystals showing good external form  Solubility  Very insoluble in water and alcohol, soluble in acid  Table 1.3 Struvite solubility products reported in literature (Doyle and Parsons, 2002) Solubility product Ksp  pKsp  References  2.51×10-13  12.6  Bube K.  7.08×10-14  13.15  Taylor et al., 1963  3.98×10-10  9.4  Borgerding, 1972  4.36×10-13  12.6  Stumm and Morgan, 1972  7.58×10-14  13.12  Burns and Finlayson, 1982  1.0×10-13  13.0  Mamais et al., 1994  2.51×10-13  12.6  Loewenthal et al., 1994  4.36×10-13  12.36  Buchananet al., 1994  1.15×10-13  12.94  Aage et al., 1997  5.50×10-14  13.26  Ohlinger and Young, 1998  30  It has been demonstrated that increasing the ion concentration stoichiometrically could increase crystal yield of struvite. This is consistent with chemical reaction. An excess of ammonium concentration relative to magnesium and phosphate appears to increase the purity of struvite precipitates in terms of struvite composition (Stratful et al., 2001). On the other hand, a few studies show that increasing Mg/P molar ratio could enhance struvite precipitation (Stratful et al., 2001; Nelson et al., 2002). The effect of temperature on struvite formation in aqueous media shows decrease with increased temperature. However, the decrease in struvite yield appears not very significant when temperature is under 60oC, although a temperature of 25oC or less is optimal (Mijangos et al., 2004). Struvite formation is also affected by the interaction of the calcium and magnesium. It is reported that magnesium ions kinetically inhibit the nucleation and subsequent growth of Caphosphate because Mg-phosphate precipitate faster than Ca-phosphates (Amjad et al., 1984; Salimi et al., 1985; Abbona, 1990). At the molar ratio of Mg/Ca > 0.6, struvite precipitation is dominant (Mosvoto et al., 2000). It has been concluded that the precipitation of calcium phosphate apatite occurs at pH above 9.5, whereas effective struvite precipitation occur at pH value of 8 and above (Doyle and Parsons, 2002).  Struvite Formation in Compost Published information about struvite precipitation in compost media is limited to only two reports. Jeong and Kim (2001) first confirmed that struvite was formed in food waste compost by adding magnesium and phosphate salts to feedstock. They found that ammonia volatilization was greatly reduced as a result of struvite formation, whereas the ammonical-N content in compost significantly increased. A subsequent study by Jeong and Hwang (2005) further investigated the optimal dosages of Mg and P salts for struvite precipitation in food waste compost, and concluded that optimal doses of Mg and P salts are about 20% of initial N in compost mixture. The major advantage of this technology is that the produced compost  31  product has high agronomic values because struvite is a valuable slow-release fertilizer (Bridger et al., 1962; Greaves et al., 1999). To date, the vast majority of struvite precipitation research has been conducted on supersaturated liquid medium from wastewater treatment plants. Little research has been done in wet porous media such as composting. As mentioned earlier, one disadvantage of forming struvite in food waste composting is that food waste contains little phosphate and magnesium that are necessary for forming struvite, thus struvite formation relied largely on chemical addition. Therefore, it is desirable to explore the possibility of forming struvite in other compostable materials that are rich in phosphate and magnesium. Animal manure and in particular poultry manure appears to fit the requirements. Manure presents a bigger problem than food waste in terms of ammonia emission from composting because of its higher nitrogen content and low C/N ratio (DeLaune et al., 2004). When compared to food waste and some other organic wastes, poultry manure contains larger amounts of phosphate and magnesium, which are helpful for struvite formation. Food waste usually has low pH value; in contrast, poultry manure generally has higher pH value, which is also favorable for struvite crystallization. Moreover, the feasibility of struvite precipitation at the optimal moisture condition (i.e. about 55%), the effect of magnesium to phosphate molar ratio on struvite formation, the dynamic changes of water-soluble phosphate in composting process, and the nitrogen conservation at the maturation stage of composting have not studied yet. All these knowledge is of theoretical and practical importance for struvite crystallization during composting.  32  Figure 1.4 Magnesium speciation versus pH at 20 oC. Total concentration of magnesium, CT: (a) CT = 0.015 M, (b) CT = 0.100M, and (c) CT = 0.15 M. (Adopted from Mijangos et al., 2004).  33  1.3 References Aage H. K, Anderson BL, Blom A, Jensen I. 1997. The solubility of struvite. J Radioanal Nucl Chem:223:213–5. Abbona, F., 1990, Crytallization of calcium and magnesium phosphates from solutions of low concentration. J. Cryst. Growth, 104:661-671. Amjad, Z., P.G. Koutsoukos and G.H. Nancollas, 1984. The crystallization of hydroxyapatite and fuorapatite in the presence of magnesium ions, J. Colloid Sci. Interface Sci. 101 (1):250-256. Anonymous, 1999. Reducing odor and VOC emissions. BioCycle. 40(3):68-72. ASTM International. 2004. Standard Practice for Determination of Odor and Taste Threshold By a Forced-Choice Ascending Concentration Series Method of Limits. E679-04, Philadelphia, PA. Barthelmy,  D.,  2002,  Mineralogy  Database.  Website  available  at:  http://webmineral.com/data/Struvite.shtml Beck-Friis, B., S. Smasrs, H. Jonsson, and H. Kirchmann. 2001. Gaseous emissions of carbon dioxide, ammonia and nitrous oxide from organic household waste in a compost reactor under different temperature regimes. J. Agric. Eng Res.78 (4):423 –430. Benedict, A.H., E. Epstein, and J. Alpert. 1988. Composting Municipal Sludge. A Technology Evaluation. Pollution Technology Review No. 152. Noyes Data Corporation. Park Ridge, N J. Bidlingmaier, W. 1993. Odour Emissions from Composting Plants. Compost Science and Utilization. 1(4):64-68. Borgerding J. Phosphate deposits in digestion systems, 1972. J Water Pollut Control Fed. 44:813–9. Bridger G.L., Murrell L. Salustky and R. W. Starostka, 1962. SCOPE Newsletter No.43 page 3. Website available at: http://www.ceep-phosphates.org/scope/articles/scope43/scope43.pdf Bridger, G. 1968. “New Fertiliser Materials 1968” - Centre International des Bube K. Uber magnesiumammoniumphosphat. Z Anal Chem 1910; 49:525–96.  34  Buchanan JR, Mote CR, Robinson RB, 1994. Struvite control by chemical treatment. Trans Am Soc Agric Eng. 37:1301–8. Burns JR, and B Finlayson, 982. Solubility of magnesium ammonium phosphate hexahydrate at various temperatures. J. Urol;128:426–8. Burns, R.T., Moody, L.B., 2002. Phosphorous recovery from animal manures using optimized struvite precipitation. Proceedings of coagulants and Flocculants: Global Market and Technical Opportunities Treatment Chemicals, Chicago, Illinois. May 22-24, 2002. Busca, G. and C. Pistarino. 2003. Abatement of ammonia and amines from waste gases: a summary. Journal of Loss Prevention in the Process Industry. 16:157-163. California Integrated Waste Management Board (CIWMB). 2007. Comprehensive composting odor response project. Report prepared by San Diego State University. Chapin, A., C. Boulind, A. Moore. 1998. Controlling odor and gaseous emission problems from industrial Swine facilities. Yale Environmental Protection Clinic Handbook Committee for European Normalization (CEN). 2003. EN13725: Air Quality – Determination of Odour Concentration by Dynamic Olfactometry. Brussels, Belgium. Cooper, P., and I.S. Cornforth. 1978. Volatile fatty acids in stored animal slurries. J. Sci. Food Agric. 29(1):83-87. Dalton, P. 2003. How people sense, perceive and react to odors. BioCycle 43(11):26-29. Day, M., K. Shaw and M. Krymien. 1999. Composting Odours: What Can Chemistry Tell Us? Institute for Chemical Process and Environmental Technology, National Research Council. Defoer, N., I.de Bo, H. van Langenhove, J. Dewulf and T. van Elst. 2002.Gas chromatography-mass spectrometry as a tool for estimating odour concentrations of biofilter effluents at aerobic composting and rendering plants. J. Chromatography A, 15 pp. DeLaune, P.B., Moor P.A., Daniel, T.C., Lemunyon, J.L., 2004. Effect of chemical and microbial amendments on ammonia volatilization from composting poultry litter. J. Environ. Qual. 33:728-734. Doyle, J.D., Parsons, S.A., 2002. Struvite formation, control and recovery. Water Research 36:3925-3940.  35  Elwell, D.L., H.M. Keener, M.C. Wiles, D.C. Borger and L.B.Willett. 2001. Odorous emission and odor control in composting swine manure/sawdust mixes using continuous and intermittent aeration. Trans. ASAE , 44:1307-1316. Elwell, D.L., D.C. Borger, D.V. Blaho, J.K. Fahrini, H.M. Keener, and L.B. Willett. 2002a. Changes in concentrations of malodorous compounds from fresh and aged dairy manure during controlled aeration composting. Proceedings, 2002 International Symposium, Composting and Compost Utilization. Track 3, 11 pgs. The JG Press, Emmaus, PA 18049, USA. Elwell, D.L., J.H. Hong, and H.M. Keener, 2002b. Composting hog manure/sawdust mixtures using intermittent and continuous aeration: Ammonia emissions. Compost Science and Utilization 10(2): 142-149. Elwell, D.L., D.C. Borger, H.M. Keener, and L.B. Willett. 2003. Reduction of volatile odorous chemicals in composting of dairy manure. Presented at the 2003 ASAE Annual International Meeting, July 27-30, Las Vegas, Nevada, USA. Paper No. 034049. Engrais Chimiques - Noyes Development Corporation - “Magnesium Fertilisers - Section 3 – Magnesium ammonium phosphate and related compounds”. pp 236-284. Epstein, E. and N. Wu. 2000. Planning, design, and operation factors that affect odor control at composting facilities. Website available at: http://www.p2pays.org/ref/11/10493.pdf Fraser, B.S. & A.K. Lau. 2000. Effect of process control strategies on composting rate and odor emission. Compost Science & Utilization. Vol. 8. No 4. 274-292. Gage, J. 2000. Operating by Progressive Odor Management Plan. BioCycle. 41(6):52-55. Gage, J., 2003. Checklist for odor management at compost facilities. BioCycle, 44(5):42-47. Goldstein, N. 2002. Getting to know the odor compounds. BioCycle, 43(7):42-44. Goldstein, J. and N. Goldstein. 2005. Controlling Odors at Composting Facilities. Pages 22 in BioCycle. Vol. 46. Goldstein, N. 2006. Rapid Treatment OPtions. BioCycle, 47(6):35-37. Greaves J., P. Hobbes, D. Chadwick, P. Haygarth. 1999. Prospects for the recovery of phosphorus from animal manures: a review. Journal of Environmental Technology. 20:697-708. Grego, S., D. Corradini, K. Liburdi, M.C. Moscatelli, R. Bernini and S. Marinari. 2002. Evaluation of aerobic composting processes of agricultural wastes by reversed phase high  36  performance liquid chromatography. Proc. Int’l Sym. composting & compost utilize. May 6-8, Columbus, OH. pp.101-109. Greater Vancouver Regional District (GVRD), 1993. GVRD Air Quality Management Plan for Greater Vancouver - Odor Management. Greater Vancouver Regional District (GVRD), 2005. Air Quality Management Plan for Greater Vancouver. Available at: http://public.metrovancouver.org/about/publications/Publications/AQMPSeptember2005. pdf. Grommen, R. and W. Verstraete, 2002. Environmental biotechnology: the ongoing quest. Journal of Biotechnology, Vol. 98: 113-123. Hansen, R.C., H.M. Keener, and H.A.J. Hoitink. 1989. Poultry manure composting-An exploratory study. Trans. ASAS 36:2151-2157. Haug, R.T. 1993. The Practical Handbook of Compost Engineering. Lewis Publishers, Boca Raton, FL. Hu. H.Y. G.M. Zeng, D.L.Huang, H.Y. Yu, X.Y. Jiang F. Dai and G.H. Huang. 2006. Use of potassium and dihydrogen phosphate and sawdust as adsorbents of ammonical nitrogen in aerobic composting process. J. Hazardous Materials. Industry Canada, 2005, Innovation roadmap on bio-based feedstocks, fuels & industrial products,  BioProducts  Canada  (2005)  ISBN  0-662-36411-2www.bio-  productscanada.org/pdf/en_roadmap_book.pdf. Jeong, Y. K. and J.S. Kim. 2001.A new method for conservation of nitrogen in aerobic composting processes. Bioresource Technol.79:129-133. Jeong, Y. K. and S.J. Hwang, 2005. Optimum doses of Mg and P salts for precipitating ammonia into struvite crystals in aerobic composting. Bioresour. Technol. 96:1-6. Kim, T.I., J.D. Han, Y.H. Yoo, T.Y. Thak, K.N. Kim, C.H.K. Kim and M.K.Kim. 2002. Studies on the deodorization of pig feces by yeasts. Proc. International Symposium on composting and compost utilization. May 6-8, Columbus, OH. pp.1274-1283. Kithome, M., J. W. Paul and A.A.Bomke.1999. Reducing nitrogen losses during simulated composting of poultry manure using adsorbents or chemical amendments, J. Environ.Qual.28 (1):194-201.  37  Kuroda, K., T. Osada, M. Yonaga, A. Kane Matu, T. Nitta, S. Mouri and T. Kojima. 1996. Emissions of malodorous compounds and greenhouse gases from composting swine faeces. Bioresource Technology 56: 265-271. Lay, J.J., Y.J. Lee and T. Noike.1999. Feasibility of biological hydrogen production from organic fraction of municipal solid waste. Water Res.33: 2579-2586. Liang, Y., 2000. Nitrogen Retention in the High Rate Stage of Composting. Ph.D. Thesis. Bioresource and Food Engineering, Department of Agricultural, Food and Nutritional Sciences. University of Alberta. Edmonton, AB. Lefcourt, A.M. and J.J. Meiseinger, 2001. Effect of adding alum or zeolite to dairy slurry on ammonia volatilization and chemical composition. 84:1814-1821. Lefebvre, X., E. Paul, M. Mauret, P. Baptiste, and B. Capdeville. 1998. Kinetic Characterization of Saponified Domestic Lipid Residues Aerobic Biodegradation. Water Research. 32(10):3031-3038. Leiser A.T. : Magnesium Ammonium Phosphate for highway Plantings. A progress Report. 42nd Annual Meeting. W.R. Grace and Co. Lemus, Gladis R. 2003. Biodegradation and environmental impact of lipid-rich wastes under aerobic composting conditions. Ph.D. Dissertation, Department of Chemical and Biological Engineering, University of British Columbia. Liao, P.H., L. Jones, A.k. Lau, S. Walkemeyer, B. Egan, and N. Holbek. 1997. Composting of fish wastes in a full-scale in vessel system. Bioresource Technology 59:163–168. Loewenthal R.E., Kornmuller U.R.C., van Heerden E.P., 1994. Modelling struvite precipitation in anaerobic treatment systems. Water Sci Technol. 30:107–16. Mamais D, Pitt PA, Yao Wen Cheung, Loicano J, Jenkins D, 1994. Determination of ferric chloride dose to control struvite precipitation in anaerobic digesters. Water Environ Res. 66:912–8. Martins, O., and T. Dewes, 1992. Loss of nitrogenous compounds during composting of animal wastes. Bioresource Technology 42: 103-111. McCrory, D.F. and P.J. Hobbs, 2001. Additives to reduce ammonia and odor emissions from livestock wastes: A review. J. Environ. Qual. 30:345-355.  38  Mijangos F., Kamel, M., Lesmes, G., Muraaviev, D.N., 2004. Synthesis of struvite by ion exchange isothermal supersaturation technique. Reactive & Functional Polymers 60:151161. Mills, B., 1996. Review of methods of odour control. Filtration and Seperation, 2:147-152. Momberg, G.A. and R.A. Oellermann, 1992. The removal of phosphate by hydroxyapatite and struvite crystallization in South Africa, Water Sci. Tech, 26:987-996. Musvoto EV, Wentzel MC, Ekama GA, 2000. Integrated chemical–physical process modelling II. Development of a kinetic based model for weak acid/base systems. Water Res. 34:1868–80. Nelson N.O., R.L. Mikkelsen, D.L. Hesterberg, 2003. Struvite precipitation in anaerobic swine lagoon liquid: effect of pH and Mg:P ratio and determination of rate constant. Bioresource Technology 89:229–236. Nicholson, J. P., 2001. ANIMAL MANURE: Challenge and Opportunity. Will U.S. Leadership be Responsive? Water Environment Federation Animal Residuals Management Conference Chicago, Illinois. June 6-8, 2001. Nimmermark S. Odour influence on well-being and health with specific focus on animal production emissions. Ann Agric environ Med. 11: 163-173. Noble, R., P.J. Hobbs, A. Dobrovin-Pennington, T.H. Misselbrook, and A. Mead. 2001.Olfactory response to mushroom composting emissions as a function of chemical concentration. J. Environ. Qual. 30:760 –767. NRC. 1979. Odors from stationary and mobile sources. National Academic of Sciences., Washington, D.C. Ohio EPA, 1999. Question and answers for composting facilities Web page available at: http://www.epa.state.oh.us/dsiwm/document/guidance Ohlinger K.N., Young T.M., Schroeder E.D., 1998. Predicting struvite formation in digestion. Water Res. 32: 3607–14. Ohlinger K.N., Young T.M., Schroeder E.D.,1999. Kinetics effects on preferential struvite accumulation in wastewater. J Environ Eng.125:730–7. Ongley, E.D., 1996. Control of Water Pollution from Agriculture. FAO paper 55.  39  O’Neill, D. H. and V. R. Phillips, 1992. A review of the odor control of odour nuisance from livestock buildings: Part 3, Properties of the odorous substances which have been identified in livestock wastes or in the air around them. J. Agric. Eng. Res. 53(1): 23-50. Ongley, E.D. 1996. Control of Water Pollution from Agriculture. FAO paper 55. Powers, W.J., 2004a. The Science of Smell Part 2: Odor chemistry. Website available at http://www.extension.iasate.edu/airquaity. Powers, W.J., 2004b. The Science of Smell Part 3: Odor detection and measurement. Website available at http://www.extension.iasate.edu/airquaity. Powers, W.J., 2004c. The Science of Smell Part 4: Principles of odor Control. Website available at http://www.extension.iasate.edu/airquaity. Raviv, M., S. Medina, A. Krasnocsky, and H. Ziadna, 2002. Conserving nitrogen during composting. BioCycle. 42(9):48-51. Richard, T.L., 1996. "The Science and Engineering of Composting." Web page available at http://compost.css.cornell.edu/science.html. Rosenfeld, P., M. Grey and M. Suffet, 2002. Controlling odors using high carbon wood ash. BioCycle. 42(3):43-45. Rynk, R. (ed). 1992. On farm composting handbook. NRAES-54, Cooperative Extension Service, Northeast Regional Agricultural Engineering Service, Ithaca, NY, USA. Salimi, M.H., Heughebaert, J.C., Nancollas, G.H.,1985. Crystal growth of calcium phosphates in the presence of magnesium ions. Langmuir 1,119–122. Sato, H., T. Hirose, T. Kimura, Y. Moriyama and Y. Nakashima. 2001.Analysis of malodorous volatile substances of human waste: feces and urine. J. Health Sci.47(5):483490. Schiffman, S.S., J.M. Walker, P. Dalton, T.S. Lorig, J.H. Raymer, and C.M. Williams, 2000. Potential health effects of odor from animal operations, wastewater treatment, and recycling of byproducts. J. Agromed. 7:7–81. Schiffman, S.S., J.L. Bennett, J.H. Raymer, 2001. Quantification of odors and odorants from swine operations in North Carolina. Agricultural and forest Meteorology. 108:213–240. Schiffman, S.S., and C.M. Williams. 2005. Science of odor as a potential health issue. J. Envirn. Qual. 34:129-138.  40  Schlegelmilch, M., J. Streese and R. Stegmann, 2005. Odour Management and Treatment Technologies – an overview, Waste Management 25 (2005) (9), pp. 928–939. Sironi, S. and D. Botta, 2001. Biofilter efficiency in odor abatement at composting plants. Compost Science & Utilization, 9(2):149–155. Smet, E, H. Van Langenhove, I. De Bo, 1998. The emission of volatile compounds during the aerobic and the combined anaerobic/aerobic composting of biowaste. Atmospheric Environment 33:1295-1303. Statistics Canada, 2001. A Geographical Profile of Manure Production in Canada. Available at: http://www.statcan.ca/english/freepub/ Stratful I., M.D. Scrimshaw and J.N. Lester, 2001. Conditions influencing the precipitation of magnesium ammonia phosphate. Wat. Rec. Vol. 35, No.17: 4191-4199. Stevens, S.S. 1960. The Psychophysics of Sensory Function. American Scientist, 48: 226253. Stumm, W., and J.J. Morgan, 1970. Aquatic chemistry. Wiley-Interscience, New York, NY. Sweeten, J. M. and M. R. Akabani, 1994. Approaches to Odor Measurement for Dairy and Cattle Feedlots: A State of the Art Review. In: Final Report: Preliminary research Concerning the Character, Sources and Intensity of Odors from Dairy Operations in Earth County, Texas; Texas Natural Resources Conservation Commission Office of Air Quality. Taylor A.W., Frazier A.W., Gurney E.L., 1963. Solubility products of magnesium ammonium and magnesium potassium phosphates. Trans Faraday Soc.59:1580–4. Timmenga and Associates Inc., 2003. Evaluation of Option for Fraser Valley Poultry Manure. Report prepared for: Broiler Hatching Egg Producers’ Association, BC Chicken Growers Association; BC Turkey Association, and Fraser Valley Egg Producers’ Association. Tiquia, S.M. and N.F.Y. Tam, 2000. Fate of nitrogen during composting of chicken litter. Environmental Pollution 110:535-541. Tiquia, S.M.,  J.H.C. Wan, and N.F.Y Tam. 2002. Microbial population dynamics and  enzyme activities during composting. Compost Science & Utilization, 10(2): 150–161. Venglovsky, J., N. Sasakova M. Vargova, Z. Pacajova, I. Placha, M. Petrovsky, D. Harichova, 1996. Evolution of temperature and chemical parameters during composting  41  of the pig slurry solid fraction amended with natural zeolite. Bioresource Technology, 96 (2005) 181–189 Walker, J.M. 1993. Control of Composting Odors. Science and Engineering of Composting Design, Environmental, Microbiological and Utilization Aspects. In Hoitink, H.A. J. and Keener, H.M. eds. Proceedings of the International Composting Research Symposium. 1992. WEF & ASCE, 1995. Odor control in wastewater treatment plants. Alexandria, VA. Water Environment Federation White, R. K., E. P. Taiganides and G. D. Cole, 1971. Chromatographic Identification of Malodors from Dairy animal Waste. 110-113. In: Livestock Waste Management and Pollution Abatement. Proc. International Symposium on Livestock Wastes. Wiles, M.C., D.L. Elwell, H.M. Keener, J.C. Amburgey, D.C. Borger and Willett, 2001. Volatile fatty acid emission during composting of swine manure as a measure of odor potential. Compost Science & Utilization, 9(1): 27–37. Weiss, K., 2007. The rise of slime. New Internationalist 2007, number 371 (Jan-Feb):6-8. Witter, E., Kirchmann, H., 1989a. Peat, zeolite, and basalt as absorbent of ammonical nitrogen during manure decomposition. Plant Soil 115:43-52. Witter, E., Kirchmann, H., 1989b. Effects of addition of calcium and magnesium salts on ammonia volatilization during manure decomposition. Plant Soil 115:53-58. Wu, Q.Z. and P.L. Bishop, 2004. Enhancing struvite crystallization from anaerobic supernatant. J. Environ. Eng. Sci. 3:21-29. Yuwono, A.S. and P.S. Lammers, 2004. Odor Pollution in the Environment and the Detection Instrumentation. Agricultural Engineering International: the CIGR Journal of Science Research and Development. Vol. VI. P1-33. Zahn, J.A., J.L. Hatfield, Y.S. Do, A.A. DiSpirito, D.A. Laird, and R.L. Pfeiffer. 1997. Characterization of volatile organic emissions and wastes from a swine production facility. J. Environ. Qual. 26:1687-1696. Zahn, J.A., DoSpirito, A.A., Do Y. S., Brooks, B.E., Cooper, E.E., and Hatfield, J.L., 2001. Correlation of human olfactory responses to airborne concentrations of malodorous volatile organic compounds emitted from swine effluent. Journal of Environmental Quality. 30:624-634.  42  Zhang, Q., J.J.R. Feddes, I.K. Edeogu and X.J. Zhou. 2002.Correlation between odour intensity assessed by human assessors and odour concentration measured with olfactometers. Can. Biosystems Eng, 44: 6.27 –6.32. Zhu, J. 2000. A review of microbiology in swine manure odor control. Agriculture, Ecosystem and Environment 78:93-106.  43  CHAPTER 2 LABORATORY SCALE BIOREACTOR SYSTEM DESIGN AND OLFACTOMETER DEVELOPMENT 2.1  Introduction  Composting emissions of carbon dioxide, ammonia, nitrous compounds and odors depend on microbial activity, and are closely associates with temperature, oxygen, pH levels and other operating conditions, but this dependency is sparsely recorded and not well understood (Smars et al., 2001, Komilis and Ham, 2000). To reduce the air emissions and determine the optimal course of composting, systematic experiments are needed to improve the knowledge in this field. The right experimental setup that can yield reproducible results is critical to enable such systematic simulation studies. The application of laboratory-scale composting reactor systems has been reviewed by a number of authors (Finstein et al., 1983, Hogan el al., 1989, Smars et al., 2001, Mason, and Milke, 2005a and 2005b). A wide range of laboratory-scale composting reactor systems including different scales, equipment and operation modes, are used worldwide. However, the results from different experimental systems are hardly comparable (Korner et al., 2003), partially because the boundary conditions of bioreactors vary widely from one to another. This is especially true for self-heating reactors at laboratory scale. Since laboratory-scale self-heating reactors have relatively small size, they may, therefore, suffer high heat losses across boundaries even with substantial quantities of insulation. To provide a suitable thermodynamic environment, the use of highly insulated bioreactor is of over-riding importance in laboratory-scale composting experiment.  The quantification of odor emissions from composting has been a significant challenge for odor research. Considerable study efforts have been made to determine odorous gas compounds through GC/MS or other chemical measurements and then attempted to relate the individual odorant to the total odor strength (Elwell et al., 2001, 2002, 2003, 2004, Fraser and Lau, 2000, Kuroda, 1996, Noble and Hobbs, 2001, Noble, et al., 2002, Rosenfels et al. 2002, Smet et al., 1999). Electronic noses have aroused a lot of interest lately (Nicolas et al. 2000, Krzymien et al., 1997). Unfortunately, neither instrumental method (GC/MS) nor A version of this chapter will be submitted for publication. Zhang, W., Lau, A.K. and Lemus, G. Upgrading of a Dynamic Dilution Olfactometer for Odor Concentration Measurements.  44  electronic noses are highly correlated with olfactometry (Gralapp, et al., 2001). Accordingly, olfactometry remains the most valid method for odor evaluation at present. Since the perception of odor by humans is still subjective, standard protocols have been developed for their use over the years. The current European Standard (CEM, 2003), which is regarded as international standard for odor quantification, has improved the reliability of odor concentration analysis significantly.  The aim for bioreactor system design and olfactometer development, described in this chapter, was to develop efficient and reliable tools for the subsequent systematic simulation experiments. Specific objectives are to:  x  Design a composting reactor system that can effectively simulate controlled operations and has high reproducibility and continuous emission monitoring capability;  x  Develop a dynamic olfactometer in compliance with the latest international standard for odor evaluation.  2.2  2.2.1  Bioreactor System Design and Evaluation  System Configuration  Thermal condition is of critical importance as it affects biological activities, moisture and vapor transfer, and oxygen status. Ordinary laboratory-scale bioreactors typically suffer significant losses of heat through the walls, even with substantial insulation present (Mason, I.G. and M.W. Milke, 2005). Dewar flask, also referred to as super-insulation vessel, is reported to have remarkable reproducibility, and hence being suggested as a standardized means in compost self-heating tests (Brinton, et al., 1992, 1993, 1994). Therefore, Dewar flasks were adopted as bioreactors for the laboratory-scale composting system.  The design of the bioreactor system in this study was on based on previous works conducted by Fraser (1997) and Lemus (2003). The system was modified to be airtight, and had the capability for continuous air emission monitoring and flexibility for automatic control. 45  Figure 2.1 shows a schematic of the laboratory-scale composting system. A set of four Dewar reactors was used in the bioreactor system. The Dewar reactor employed is a 6-L doublewalled stainless steel flask with vacuum in between (Cole Parmer Instruments Company, Verson Hills, IL). A supported metallic mesh was raised about 500 mm from the bottom of the reactor, so that sufficient space is allowed for the aerator and air distribution. An aquarium air diffuser (porous stone) is used to ensure that the air is distributed uniformly. The composting vessel is positioned inside an insulated (adiabatic) box in order to simulate typical in-vessel composting process, or the core part of a compost pile.  All reactors are closed in an airtight manner to allow balancing of all mass flows and prevent the dilution of exhaust air by ambient air before sampling. For additional thermal insulation, each reactor was wrapped in a custom made Reflectix sleeve (heating tank insulation, Reflectix Inc., Markville, IN) and placed inside a 0.8×0.8×1m box (1 m high) constructed with R-5 Styrofoam board; the space between the board and the reactor was filled with fiberglass insulation material.  Compressed air supplied by a series of air pumps was used for forced aeration. The aeration was controlled automatically and continuously measured using air flow meters. The baseline aeration rate was set to 0.72 L/min.kg dry mass (DS) as suggested in literature (Rynk, 1992). Exhausted air was led into an empty flask for cooling, then went through a sampling port for sampling, and finally passed through the second flask with acid scrubbing solution to remove ammonia for measurement.  The automation system was designed to provide process control up to eight reactors at a time. The laboratory-scale unit was equipped with online temperature and airflow rate monitoring system, and emission traps. The system allows for the simulation of composting process with various operational controls.  46  Relay  2  1. aeration pump 2. control valve 3. air flowmeter  4. leachate 5. bioreactor 6. thermocouple  7. condense trap 8. oxygen 9. ammonia trap  Figure 2.1 Setup of composting unit: Bioreactor with periphery equipment 47  11. control boards 12. computer  2.2.2  Process Control and Data Acquisition  The process control strategy used is the industry standard (the Rutger's temperature-feedback method). The temperature setpoint chosen was 65qC, which aims at reducing aeration and maximizing microbial activity, while meeting the regulatory requirements for pathogen destruction. Aeration was intermittent with 33% duty cycle below the temperature setpoint, and continuous above the setpoint. Temperature was continuously measured with retrievable thermocouple probes.  A PC-based process control system was used to monitor and login data. The computer used was fitted with a multiplexer board and a wiring terminal (Advantech, California). The system runs on Labtech Control V.9.02 software. Interfaces to external measuring and control devices are provided through the data acquisition and control boards. Other external devices include thermocouples for temperature monitoring, and relay boxes and solenoid valves for air supply control. A schematic flowchart of the process control is shown in Figure 2.2. The automated process control and data acquisition was programmed using drawing language, a built-in feature in Labtech. The program was written to control up to eight reactors simultaneously. Figure 2.3 shows the program developed for the data acquisition and process control.  2.2.3  Overall System Performance  Composting experiment is costly and time-consuming because composting is a very slow biodegradation process. A bioreactor system capable of reproducing thermal regime is of critical importance to experimental studies. To evaluate the overall performance of the bioreactor system used in this study, data collected from a series of experiments were used to statistically analyze the errors and reproducibility of the bioreactor system. A quantitative assessment method based on temperature distribution pattern, which was modified from the method proposed by Mason and Milke (2005), was employed. The evaluation parameters used include:  48  Start  Input DC, DI  No On duty ? Yes LF off  LF On  Measure T  No T > SP ? Yes HF on  Figure 2.2 Temperature feedback control flowchart DC: Duty cycle DI: Duty interval LF: Low flowrate HF: High flowrate SP: Temperature setpoint  49  HF off  Figure 2.3 Drawing language program for composting process control  50  (1) the time of mesophilic phase, thermophilic phase and pathogen destruction requirement, which are denoted as Tmp, Ttp and Tpd, respectively; (2) the areas bounded by temperature curve with temperature baselines of lower disinfection limit (Apd), with temperature baselines of lower thermophilic phases (Atp), and the upper and lower limits of mesophilic phase (Amp) as shown in Figure 2.4. 80 Lower disinfection limit  70  Temperature ( oC )  60  Lower thermophilic limit (Upper mosophilic)  50 40 30 20  Lower mosophilic limit  10  Time  Figure 2.4 Generic composting process temperature profile (Modified from Rynk, 1992)  The baseline temperatures of different composting stages are as follows (Brock et al, 1984, Tchobannoglous et al, 1993): Mesophiles:  25-45oC  Thermophilies:  > 45oC  Pathogen destruction: > 55oC  The defined areas shown in Figure 2.4 were calculated by summation of the production of time step multiplied with corresponding temperature. Table 2.1 shows the evaluation parameters and statistical analysis results. These data are from eight trial-runs with identical composting recipe but at different time periods. The recipe used was 65% of chicken manure, 25% of sawdust and 10% of hog fuel on wet mass basis. The selection of amendments was largely in accordance with those locally available materials and commonly used in 51  commercial facilities. To formulate the recipe, preliminary trials were also conducted with broiler litter amended by different percentage of tree leaves, sawdust and hog fuel. The recipe that gave the best degradation and thermal performance was selected and used as a standard recipe throughout this thesis study. All trials were operated under similar conditions in which moisture was adjusted to 55% that is optimal for composting, and the aeration rate was set to the industrial standard of 0.72 L/min·kg DS (Jeris and Regan; 1973; Rynk, 1992; Haug, 1993; Day and Shaw, 2001; Agnew and Leonard, 2003). The temperature data were from the same time span of 216 hours.  For all runs, the period of pathogen destruction (as indicated by the temperature profile) ranged from 3.6 to 4.9 days, with an average of 4.2 days and standard deviation of 0.47 days, which exceeded the regulatory requirement of three consecutive days (BCMOE, 2002). The average mesophilic phase and thermophilic phase lasted for 3.0 and 5.8 days with standard deviation of 0.35 and 0.38 days, respectively. In terms of the area parameters, Amp was found to be 146±9.13 oC-days, while Atp and Apd were 91.5±7.91 oC-days and 40.9±5.28 oC-days, respectively. These results demonstrate that the bioreactor system used has high reproducibility in terms of thermal performance.  Table 2.1 Temperature-time data from laboratory scale bioreactor system Test Amp  Area (oC-days) Atp  Apd  Tmp  Time (days) Ttp  Tpd  Run 1  132.0  93.0  45.5  3.5  5.2  4.3  Run 2  155.4  104.6  49.2  2.8  6.2  4.9  Run 3  147.5  95.1  42.3  2.7  5.9  4.5  Run 4  154.0  97.7  44.2  2.9  6.1  4.6  Run 5  137.5  81.4  35.9  3.7  5.2  3.8  Run 6  143.5  85.6  36.0  3.0  5.8  3.6  Run 7  158.5  82.7  34.6  2.9  6.1  3.7  Run 8  144.7  91.9  39.2  3.0  5.9  4.5  Average  146.6  91.5  40.9  3.0  5.8  4.2  STDEV  9.13  7.91  5.28  0.35  0.38  0.47  52  2.3  Dynamic Dilution Olfactometer  In the search for effective means for composting odor control, the odor must be quantified first. Olfactometry, which is directly related to receptors using the human sense of smell, is widely regarded as principal measurement technique for odor qualification (Berglund et al. 1987, Callan 1993, Chen et al. 1999, Gostelow, 2001, Feddes et al. 2001, Krzymien and Day 1997, Powers, 2004). However, the uncertainty of olfactometry method has been a major concern to researchers, environmental and regulatory stakeholders (Jiang, 2001). In many cases, the results from olfactometry method were comparable within studies, but were not compatible to other studies due to different olfactometry standards being used (Jiang, 2005).  The development of European (EU) Standard EN13725 “Air quality-determination of odor concentration by dynamic olfactometry” (CEN, 2003), which also supports and exceeds the most recent ASTM Standard of Practice E679-04 “Determination of Odor and Taste Threshold by a Forced-Choice Ascending Concentration Series Method of Limits” (ASTM, 2004), has improved the reliability of odor measurement (SRF Consulting Group, Inc., 2004). The new EU standard has become the international standard for odor testing (McGinley, 2005).  The introduction of EU Standard EN13725 presents both opportunities and challenges for olfactometry laboratories. To implement this standard, a reliable dynamic olfactometer is required to meet the stringent instrumental and panelist performance as well as procedure criteria. The objective of this research, as presented in this section, was to design a dynamic olfactometer that complies with the new international standard such that the quality of odor quantification in the subsequent systematic experimental studies can be assured.  2.3.1  Design  The Environmental Laboratory of the Chemical and Biological Engineering Department has a prototype of olfactometer previously assembled in accordance with the 1991 ASTM E-679  53  Standard (Bruce, 1998). This prototype olfactometer formed the base for the upgraded olfactometer design. The new design focused on the following improvements: x  Re-design of airflow lines to minimize the length of tubes and residence time in order to prevent odorant contamination and deliver airflow to ports uniformly;  x  Improving the control and measuring devices to increase the accuracy; and  x  Upgrading the capacity of olfactometer to meet the requirements of the EU standard.  Figure 2.5 shows the schematic of the triangular forced-choice dynamic dilution olfactometer. The olfactometer consists of two stations. Each station has three sniffing ports. The sniffing cylinders are made of glass, with 50 mm in diameter and 400 mm in length to avoid any turbulent airflow when reaching panelist’s nose. The design air flowrate at the sniffing port is 20 L/min, but it may be adjusted up to 30 L/min if desired. The resultant face velocity is 0.2 m/s for the designed airflow rate. Air flowrate can be visually monitored with flowmeters installed on the operator panel.  Two air chambers are positioned in the center of the olfactometer. One is to delivery odorfree air to the panelist sniffing ports. The other is used for mixing odorous air. The special layout of air chambers and airflow lines minimizes the length of tubes that convey odorous air, and hence reduce odorants contamination. Also the tube length from air chamber to each sniffing port are similar, thus airflow can be distributed uniformly to panelists (Figure 2.6).  Pressured odor-free air is introduced through an air regulator, and then passes though two activated carbon filters to further remove any odorants before entering the air chambers. Two multi-head peristaltic pumps are used for sampling odor and delivering the odor sample to the air mixing chamber. The smaller pump has flowrate capacity ranging from 0.3 to 38 ml/min. Once the requirements for odor sampling exceed the capacity of the smaller pump, the larger pump will be activated, which has a flowrate capacity of 15-700 ml/min.  An automated system was used for panelists' input for recording and processing. Each station of the olfactometer has an input panel that consists of six electrical switches arranged in two sets. One set is for decision input. The panelist is forced to choose which of the sniffing ports 54  has odor. The other set has switches for level of certainty, with which the panelist needs to indicate the decision based on a pure “guessing”, a possible “inkling”, or a positive “certainty”. The electrical switches are connected to a computer with input/output boards (Advantech, CA). The computer processes the panelists’ input and randomly assigns the odor port for each station for the next level of dilution. A customized software program developed with Turbo-C was used for processing the input signals (Bruce, 1998).  All olfactometer parts that come into direct contact with odorous air are made of glass or stainless steel. The olfactometer is capable of producing dilutions ranging from 25 up to 218. Table 2.2 compares the properties of the olfactometer with ASTM and European (EU) standards. The forced choice dynamic dilution olfactometer complies with the EU Standard and the current ASTM Standard.  2.3.2  Calibration  To obtain consistent and accurate odor measurements, the sample flows were calibrated at all delivery settings using a set of primary volume devices. Figure 2.6 shows the setup of the olfactometer calibration. A set of soap bubble flow meters was used in the calibration. For each pump selection and their delivery setting, flows were measured in triplicate and averaged to ensure stability.  55  SNIFFING PORTS  SAMPLING PUMP CONTROLLER  PANEL STATION 2  SNIFFING PORTS  PANEL STATION 1  ODOR AIR CHAMBER  FRESH AIR CHAMBER  MULTI-HEAD SAMPLING PUMP 1  P-1 P-2  P-5  P-3  COMPUTER  ROTAMETERS  P-4  P-6 RELIEF VALVE  MULTI-HEAD SAMPLING PUMP 2 ACTIVE CARBON FILTER  FRESH AIR ODOR SAMPLE TEDLAR BAG  Figure 3.5 Schematic of forced choice dynamic dilution olfactometer  Figure 2.6 Layout of air chambers and airflow lines of the olfactometer  The calibration involved 522 single measurements comprising 87 measurements for each of the six pump selections. Table 2.3 presents the calibration results. The odor sampling PumpA has four selections P-1, P-2, P-3 and P-4. The standard deviations of airflow rate ranged from 0.001 to 0.391. The odor sampling Pump-B has two selections P-5 and P-6. Their airflow rates have standard deviations ranged from 0.21 to 9.45. The results indicate that odor sampling Pump-A is more accurate than Pump-B. In other words, the olfactometer is more accurate in the higher dilution range of 210 to 218 (1024 to 262144) when using small Pump-A. The accuracy of the instrument becomes somewhat sensitive in the lower dilution range of 25 to 29 (32 to 512), when using large Pump-B. The airflow calibration curves for the triangular forced choice dynamic olfactometer are provided in Appendix I.  57  Figure 2.7 Olfactometer calibration setup  58  Table 2.2 Properties of the force choice dynamic dilution olfactometer Parameters  ASTM Standard (2004)  EU Standard (2003)  Dynamic Dilution Olfactometer  Yes  Yes  Yes  Mode  Forced or Nonforced choice  Forced-choice or Yes/No  Forced choice  Station  Not required  Not required  2  Sniffing port diameter  5-10 cm  Not specified  5 cm  Number of sniffing ports  2 or 3  Minimum of 2  3  Panel size  4-16  4 to 8  4  Presentation order  Ascending  Ascending or random  Ascending  Face velocity in port  0.01-0.1 m/s  0.2 - 0.5 m/s  0.2 m/s  Presentation flow  >3 L/min  20 L/min  20 L/min  Maximum (lower) dilution ratio  <10  Less than 27  25  Minimum (upper) dilution ratio  • 5000  At least 214  218  Dilution ratio range  Not specified  At least 213  214  Dilution factor increase  ”2  1.4 - 2.0  2.0  Evaluation time  15-30 seconds  15 seconds  15 seconds  Inter-stimulus interval /Purging time  Not specified  • 30 seconds  30 seconds  Dynamic dilution  59  Table 2.3 Results of olfactometer airflow calibration (n=3) P-1 P-2 P-3 Delivery Settings Avg SD Avg SD Avg SD  P-4  P-5  P-6  Avg  SD  Avg  SD  Avg  SD          15.40  0.137  20.50  0.164  29.00  0.072  1.12  0.005  33.90  0.651             0.909 0.969 0.625 2.501 1.176 1.279  0.8 1 1.2 1.4 1.6 1.8 2                                      0.31 0.41 0.48 0.61 0.67 0.76  0.007 0.006 0.005 0.006 0.008 0.001  0.56 0.66 0.77 0.89 1.04 1.15  0.006 0.006 0.008 0.012 0.011 0.013  1.50 1.88 2.37 2.80 3.54 3.77  0.014 0.021 0.025 0.028 0.063 0.046  1.93 2.45 3.35 3.84 4.51 5.12 5.85  0.038 0.015 0.041 0.018 0.060 0.041 0.062  39.92 44.53 49.57 54.77 58.02 67.08 75.67  0.767 0.417 0.562 0.214 2.699 0.262 1.081  94.83 104.73 115.92 126.11 142.60 159.04  2.2 2.4  0.84 0.95  0.009 0.006  1.31 1.48  0.006 0.040  4.15 4.85  0.077 0.008  6.58 7.39  0.085 0.072  80.69 85.88  2.117 1.676      186.44  3.013  2.6  1.05  0.007  1.63  0.033  5.48  0.032  8.26  0.140  89.78  1.673  2.8 3  1.15 1.29  0.024 0.011  1.77 1.98  0.018 0.015  5.91 6.39  0.121 0.026  9.08 9.84  0.193 0.043  98.26 111.81  0.385 2.756        233.06  5.233  3.33  1.44  0.024  2.25  0.012  7.09  0.089  11.31  0.096  122.01  3.435  3.67 4  1.62 1.78  0.033 0.007  2.47 2.79  0.009 0.018  8.00 8.80  0.077 0.162  12.64 14.25  0.214 0.089  134.95 147.70  3.870 2.184        313.63  2.895  4.33              2.23  0.026  385.66  1.059  5.5 6          2.75  0.047  460.90  6.837  6.5 7 8 9          3.31 3.85 4.24  0.046 0.023 0.022  5.08 5.77 6.60  0.096 0.130 0.078  10      7.15  0.178  23.71  0 0.2 0.4 0.6  4.67 5  3.01  0.022  9.77  0.125  15.17  0.274  159.40  2.047  3.24 3.53  0.037 0.019  10.51 11.25  0.166 0.185  16.42 17.85  0.072 0.134  170.97 181.45  0.839 2.252  3.86 4.23  0.045 0.005  12.52 13.69  0.236 0.180  19.54 21.71  0.076 0.175  197.16 213.71      0.391 0.056 0.133 0.317        0.352 0.150 0.280  23.62 25.70 30.45 34.33    16.46 19.33 21.33  257.02 283.80 317.54  3.171 2.018 8.947  538.43 611.35 671.14  4.715 5.341 9.452  0.355  38.36  0.149  349.09  3.572  702.49  5.606  60  3.618 6.416  2.3.3  Operation  The European Standard EN13725 was strictly followed in odor test. Panelists were selected through a screening process using the n-butanol method to ensure a “normal” sensitivity. The olfactometer operator was responsible for training the panelist. But the detail procedures of odor testing were not disclosed to the panelists in order to reduce the subjective errors.  Total odor concentration was measured within 30 hours of sampling. Olfactometer operator initiates the computer software, sets the dilution ratio, observes each panelist’s input on computer screen, and informs the panelists once they complete their assessment session. Each test sample was presented to the panelist in ascending order of sample concentrations, with a series of dilution ratios increasing by two-fold each time, so as to avoid olfactory fatigue (odor habituation and loss of sensitivity) according to ASTM Standard (1991). The panelists are allowed to sniff, compare, but then forced to choose which of the ports is the one with odor within 15 seconds. The interval between dilution series is 30 seconds to purge the olfactometer and avoid adaptation of the assessors to the odor.  2.3.4  Data Evaluation  Olfactometry analysis gives the total odor concentration of the sample, expressed in odor units (OU, or previously called dilution-to-threshold, D/T]. The current European standard defines an odor unit, OUE, as the amount of odorants that, when evaporated into 1 m3 of neutral gas (air or nitrogen) at standard conditions, elicits a physiological response from an odor panel equivalent to that elicited by one European Reference (CEN, 2003) odor mass (with 40-80 ppb n-butanol) under the same conditions. This standardized procedure effectively attempts to express odor concentrations in terms of “n-butanol mass equivalents”, and by definition, 1 OUE and 1 OUE/m3 have the same meaning.  61  For each set of panelist responses, individual threshold estimate ZITE is estimated as the geometric mean of the dilution factors where panelist responses change from “false” to consistently “true”, when sorted on ascending order (Equation 3.1). Table 2.4 shows the result code driven from the combination of choice and level of certainty for the forced choice presentation mode.  Z ITE  Z i 1 ˜ Z i  (3.1)  where Zi is the dilution factor; and i is the point where panelist responses change from “false” to consistently “true”. Retrospective screening was then carried out according to the parameter ¨Z as described in the European Standard to exclude panel members that show deviant responses. The parameter ¨Z is calculated with the following equations:  Z ITE ­ °° 'Z Z ITE ® Z °'Z  ITE °¯ Z ITE n  Z ITE  Z ITE t Z ITE  (3.2)  Z ITE  Z ITE  (– Z ITE ( i ) )  1 n  (3.3)  i 1  where: ǻZ is retrospective screening parameter; Z ITE is dilution factor at panel threshold; n is number of panelist members. The parameter ǻZ must comply with the following criteria: -5 < ǻZ < 5  (3.4)  62  The screening procedure was repeated until all panel members comply with the above requirements. The odor concentration of a sample was determined by the numerical value of geometric dilution factor at panel threshold ( Z ITE ) after retrospective screening process. A spreadsheet was developed to implement the odor evaluation procedure as described above. Table 2.4 Result code driven from panel observation* Choice  Certainty  Response  Result code  Incorrect  Guess  False  0  Correct  Guess  False  0  Incorrect  Inkling  False  0  Correct  Inkling  False  0  Incorrect  Certain  False  0  Correct  Certain  True  1  * Adopted from EU Standard EN13725 (CEN, 2003). 2.3.5  Overall Odor Laboratory Performance Quality  To test compliance with the overall quality criteria for accuracy of the odor concentration measurements within our odor laboratory, a series of odor measurements was carried out using the n-butanol in odor-free air. The n-butanol odor used was a certified reference material. Twelve (12) odor tests were conducted under repeatability conditions at different times over the course of one month. The measurements were carried out with four panelists in accordance with the normal procedures used in our olfactometry laboratory. The testing results are presented in Table 2.5.  The repeatability standard deviation for the laboratory, sr, is calculated with: n  ¦(y sr  i   y w )2  i 1  (n  1)  0.1475  (3.5)  where n is the number of test results; yw is the average test results; and yi is the single test result.  63  The repeatability limit, r, is then calculated by:  r  t ˜ 2 ˜ sr  2.201 u 1.4142 u 0.1475  0.459  (3.6)  When r = 0.459 < 0.477 (the criterion for the repeatability limit), the olfactometry or odor lab shall meet the EU Standard for the precision requirement, expressed as “repeatability”. The repeatability limit in non-logarithmic terms is: 10r = 2.88 < 3.0  This implies that the ratio between two single measurements, performed on the same testing material in this laboratory under repeatability conditions, will not be larger than 2.88 in 95% of the cases.  Table 2.5 Calculation of precision for odor lab performance Test No.  Test Results (OUE/m3)  log 10 (yi)  y1  11585  4.06  y2  11585  4.06  y3  11585  4.06  y4  11585  4.06  y5  11585  4.06  y6  11585  4.06  y7  8192  3.91  y8  5793  3.76  y9  11585  4.06  y10  11585  4.06  y11  4871  3.69  y12  5793  3.76  64  yw  sr  r  3.97  0.1475  0.459  2.4  Conclusions  Composting odor emissions are closely associates with operating conditions, but this dependency is still not well understood. Systematic simulation studies are needed to improve the knowledge of optimal course of composting in reducing odor emissions. To enable such a systematic study, a composting bioreactor system that can effectively simulate controlled operations and has high reproducibility and continuous emission monitoring capability was developed. The overall performance of the bioreactor system was evaluated with data collected from a series of lab experiments during different time. The results demonstrate that the bioreactor system statistically has high reproducibility in terms of thermal performance. The bioreactor system was proven an efficient and reliable tool for subsequent laboratory experiments of this study.  Because electronic nose technology is not yet applicable to complex odors tests, olfactometry and odor panel tests remain the international standard for odor measurements. A forced-choice dynamic dilution olfactometer previously assembled in accordance with the ASTM Standard was successfully upgraded to be in compliance with the European/international standard EN13725 in regard to performance quality requirements, quality requirements for the dilution apparatus (olfactometer), odor panel procedure and data analysis. Aside from being used as a tool in odor research in this study, the upgraded olfactometer has since been used by the Greater Vancouver Regional District (Metro Vancouver) as well as other industrial operations.  65  2.5  References  Agnew, J.M. and J.J. Leonard, 2003. The Physical Properties of Compost. Compost Science and Utilization, Vol. 11 No. 3, 238-264. ASTM. American Society for Testing and Materials. 1991. Standard Practice for the Determination of Odor and Taste Thresholds by a Forced-Choice Ascending Concentration Series Method of Limits. E679-79. Philadelphia, PA. ASTM International. 2004. Standard Practice for Determination of Odor and Taste Threshold By a Forced-Choice Ascending Concentration Series Method of Limits. E679-04, Philadelphia, PA. Bari, Q., Koenig, A., Guihe, T., 2000. Kinetic analysis of forced aeration composting-I. Reaction rates and temperature. Waste Management and Research 18 (4), 303–312. Brock, T.D., D.W. Smith, and M.T. Madigan. 1984. Biology of Microorganisms. Prentice-Hall Inc., Englewood Cliffs, New Jersey, p. 240. Bruce, M.P. 1998. Odour Production and Oxygen Consumption under Controlled Composting Conditions. Master’s Thesis. Chemical and Bio-Resource Engineering. University of British Columbia. Vancouver, BC. Committee for European Normalization (CEN). 2003. EN13725: Air Quality – Determination of Odour Concentration by Dynamic Olfactometry. Brussels, Belgium. Day, M., Shaw, K., Cooney, D., Watts, J., Harrigan, B., 1997. Degradable polymers: the role of the degradation environment. Journal of Environmental Polymer Degradation 5 (3), 137–151. Elwell, D.L., H.M. Keener, M.C. Wiles, D.C. Borger and L.B. Willett. 2001. Odorous emission and odor control in composting swine manure/sawdust mixes using continuous and intermittent aeration. Trans. ASAE , 44:1307-1316. Elwell, D.L., D.C. Borger, D.V. Blaho, J.K. Fahrini, H.M. Keener, and L.B. Willett. 2002. Changes in concentrations of malodorous compounds from fresh and aged dairy manure during controlled aeration composting. Proceedings, 2002 International Symposium, Composting and Compost Utilization. Track 3, 11 pgs. The JG Press, Emmaus, PA 18049, USA.  66  Elwell, D.L., D.C. Borger, H.M. Keener, and L.B. Willett. 2003. Reduction of volatile odorous chemicals in composting of dairy manure. Presented at the 2003 ASAE Annual International Meeting, July 27-30, Las Vegas, Nevada, USA. Paper No. 034049. Elwell, D.L., D.C. Borger, D.V. Blaho, J.K. Fahrini, H.M. Keener, and L.B. Willett. 2004. Changes in concentrations of malodorous compounds during controlled aeration composting. Compost Science & Utilization, 12(2):102-107. Feddes, J.J.R., G. Qu, C.A. Oullette, and J.J. Leonard. 2001. Development of an Eight-Panelist Single Port, Forced-Choice, Dynamic Dilution Olfactometer. Canadian Biosystems Engineering. 43:6.1-6.5. Fraser, B.S. & A.K. Lau. 2000. Effect of process control strategies on composting rate and odor emission. Compost Science & Utilization. 8 (4), 274–292. Finstein, M.S., Miller, F.C., Strom, P.F., MacGregor, S.T., Psarianos, K.M., 1983. Composting ecosystem management for waste treatment. Biotechnology (June), 347–353. Gralapp, A. K., W. J. Powers, D. S. Bundy. Comparison of olfactometry, gas chromatography, and electronic nose technology for measurement of indoor air from swine facilities. 2001. Trans. ASAE , 44 (5):1283-1290. Haug, R.T. 1993. The Practical Handbook of Compost Engineering. Lewis Publishers, Boca Raton, FL. Hogan, J.A., Miller, F.C., Finstein, M.S., 1989. Physical modeling of the composting ecosystem. Applied and Environmental Microbiology. 55 (5): 1082–1092. Jeris, J.S. and R.W. Regan, 1973. Controlling environmental parameters for optimum composting. II. Moisture, free air space and recycle. Compost Science, 14:8-15. Jiang, J., 2001. What are the areas for further research? Water Science & Technology, Vol. 44 No 9:7-25. Jiang, J., 2000. Development of the Next Generation Dynamic Olfactometer – Dynascent. Vol. 44, No 9:7-25. Komilis, D., Ham, R., 2000. A Laboratory Method to investigate Gaseous Emissions and Solid Decomposition During Composting of Municipal Solid Wastes. Compost Science & Utilisation 8 (3), 254–265.  67  Korner, I., J. Braukmeier, J. Herrenklage, K. Leikam, M. Ritzkowski, M. Schlegelmilch, R. Stegmann. 2003. Investigation and optimization of composting processes— test systems and practical examples. Waste Management. 23:17-26. Krzymien, M.E., and M. Day.  1997.  Odours and Volatile Organics Emissions from a  Commercial Composting Operation. In Proceedings of the Air and Waste Management Association 90th Annual Meeting & Exhibition. June 8-13. Toronto, ON. Krzymien, M., Day, M., Shaw, K., Mohmad, R., Sheehan, S., 1999. The role of feed composition on the composting process. II. Effect on the release of volatile organic compounds and odours. Journal of Environmental Science & Health Part A – Toxic/Hazardous Substances & Environmental Engineering 34 (6), 1369–1396. Kuroda, K., T. Osada, M. Yonga, A. Kanematu, T. Nitta, S. Mouri, and T. Kojima. 1996. Emissions of Malodorous Compounds and Greenhouse Gases from Composting Swine Feces. Bioresource Technology. 56:265-271. Lemus, Gladis R. 2003. Biodegradation and environmental impact of lipid-rich wastes under aerobic composting conditions. Ph.D. Dissertation, Department of Chemical and Biological Engineering, University of British Columbia. Liao, P.H., May, A.C., Chieng, S.T., 1995. Monitoring process efficiency of a full-scale in-vessel system for composting fisheries wastes. Bioresource Technology 54 (2), 159–163. Mason, I.G. and M.W. Milke, 2005. Physical modelling of the composting environment: A review. Part 1: Reactor systems. Waste Management. 25:481-500. Mason, I.G. and M.W. Milke, 2005. Physical modeling of the composting environment: A review. Part 2: Simulation performance. Waste Management. 25:501-509. McGinley, M.A. and C.M. McGinley. 2005. Measuring Composting Odors for Decision Making. U.S. Composting Council 2005 Annual Conference. 24-26 January 2005, San Antonio, TX. International Water Association. Singapore Conference. June 19-21. Salt Lake City, UT. Michel Jr., F.C. and C.A. Reddy. Effect of oxygen level on yard trimmings composting rate, odor production, and compost quality in bench-scale reactors. Compost Science & Utilization. Vol. 6. No.4, 6-14. Pereira-Neto, J.T., Stentiford, E.I., Mara, D.D., 1987. Low cost controlled composting of refuse and sewage sludge. Water Science and Technology 19 (Rio), 839–845.  68  Powers, W.J. “Development of Procedures for Odor Evaluation Techniques” 1st IWA International Conference on Odour and VOC’s: Measurement Regulation and Control Techniques. Sydney, Australia, March 25-19, 2001. Rosenfels, P. M. Grey and M. Suffet. 2002. Controlling odors using high carbon wood ash. Biocycle, 42-45. Rynk, R. (ed). 1992. On farm composting handbook. NRAES-54, Cooperative Extension Service, Northeast Regional Agricultural Engineering Service, Ithaca, NY, USA. Schiffman, Susan S, Jeanette L. Bennett, and James H. Raymer. “Quantification of odors and odorants from swine operations in North Carolina.” Agricultural and Forest Meteorology 108(2001): 213-240. Schloss, P.D., Chaves, B., Walker, L.P., 2000. The use of the analysis of variance to assess the influence of mixing during composting. Process Biochemistry 35 (7), 675–684. Seki, H., 2000. Stochastic modeling of composting processes with batch operation by the Fokker–Planck equation. Transactions of ASAE 43 (11), 169–179. SRF Consulting Group, Inc., 2004. A Review of National and International Odor Policy, Odor Measurement Technology and Public Administration. Prepared for Minnesota Pollution Control Agency (MPCA). Smet, E., H. Van Langenhove, I. De Bo. 1999. The emission of volatile compounds during the aerobic and the combined anaerobic/aerobic composting of biowaste. Atmospheric Environment 33: 1295-1303. Smars S., B. Beck-Friis, H. Jonsson, H. Kirchmann. 2001. An Advanced Experimental Composting Reactor for Systematic Simulation Studies. J. Agric. Engng Res. 78 (4), 415422. Tchobanoglous, G., H. Theisen, and S. Vigil. 1993. Biological and chemical conversion technologies, p. 671–716. In: G. Tchobanoglous (ed.). Integrated Solid Waste Management: Engineering Principles and Management Issues, 2nd edition. McGraw-Hill Higher Education, New York. VanderGheynst, J., Gossett, J., Walker, L., 1997. High-solids aerobic decomposition: pilot-scale reactor development and experimentation. Process Biochemistry 32 (5), 361–375.  69  Zahn, J.A., DoSpirito, A.A., Do Y. S., Brooks, B.E., Cooper, E.E., and Hatfield, J.L., 2001. Correlation of human olfactory responses to airborne concentrations of malodorous volatile organic compounds emitted from swine effluent. Journal of Environmental Quality. 30:624634.  70  CHAPTER 3 REDUCING AMMONIA EMISSION VIA STRUVITE FORMATION  3.1  Introduction  Poultry manure composting generally has a high potential for ammonia emission because of high N concentration in the poultry litter and low C/N ratio (DeLaune et al., 2004). Hansen et al. (1989) reported that nitrogen loss from poultry manure composting was up to 33% of the initial nitrogen. Kirchmann and Witter (1989) composted poultry manure with straw and found that 44% of initial nitrogen in the composting mixture lost through ammonia emission. A study by Kithome et al. (1999) shown that the ammonia loss was from 47 to as high as 62% of initial nitrogen during 25 days of poultry manure composting. The substantial loss of nitrogen through ammonia emission can not only reduce the agronomic value of compost product, but also lead to onsite odor and potential health problems.  The traditional means to reduce ammonia emission at composting facilities are those involving collection and treatment via absorption, biofiltration or wet scrubbing. However, these solutions require additional equipment and space, which could be costly, and do not address the sources of the problems. Moreover, the scrubbing solution containing ammonia must be utilized or disposed of in some way. Another strategy is to use preventive measures via adsorbing or precipitating ammonia into other compounds before its release from composting matrix. Some adsorption materials such as peat, zeolite and basalt have been investigated (Kithome et al. 1999; Liao et al., 1995; Witter and Kirchmann, 1989). However, Witter and Kirchmann (1989) concluded that these adsorbents were effective in adsorbing ammonia in gases, but less effective when mixed with manure.  Jeong and Kim (2001) reported a new approach to reduce ammonia emission by adding magnesium and phosphate salts in food waste composting mixture to precipitate ammonia into struvite. Their results demonstrated significant increase in free ammonium nitrogen content in compost and substantial decrease in ammonia loss as ammonium was precipitated into struvite before releasing into outside environment. The major advantage of this method A version of this chapter has been published. Zhang, W. And Lau, A.K. (2007) Reducing Ammonia Emission from Poultry Manure Composting via Struvite Formation. J. Chem. Technol. Biotechnol. 82: 598-602.  71  is that the produced compost product has higher agronomic values because struvite is a valuable slow-release fertilizer. Several studies have shown that comparing to conventional NPK fertilizers, struvite is a better fertilizer for planting, forestry seedlings and transplants, and can increase the growth of grasses, fruit and high value crops such as ornamentals, strawberry (Scope newsletter, 2001; Wrigley et al., 1992).  Thus far, the great majority of struvite precipitation studies in waste management have been concentrated on domestic and swine wastewater for phosphorus recovery in which struvite are formed in supersaturated liquid medium (Wrigley et al, 1992; Ohlinger et al., 1999; Wu and Bishop, 2004; Nelson et al. 2003; Burns and Moody, 2002). While struvite formation in high moisture food waste composting also appears possible (Jeong and Kim, 2001), there is no report on struvite formation in other composting media. One of the limitations of forming struvite in food waste is that food waste has little phosphate and magnesium that are necessary for struvite formation. The sources of P and Mg have to rely on chemical addition. Therefore, it is desirable to explore the feasibility of forming struvite in poultry manure composting, because poultry manure is relatively rich in phosphate. In addition, poultry manure composting generally has excessive ammonium and higher pH, which are favorable for struvite crystallization. Furthermore, research on struvite formation in composting media is still in a very early stage. Knowledge is lacking about the possibility of forming other mineral compounds such as calcium struvite that may adversely affect struvite formation, the effects of Mg/P molar ratio to struvite formation, the nitrogen conservation at the maturation stage due to struvite formation, and the potential effects of salts addition on compost quality in terms of salinity. These knowledge gaps will need to be addressed and warrant more research. The objectives of this series of experimental study were:  (1) To investigate the feasibility of struvite formation in poultry manure composting at optimal moisture condition; (2) To examine the effectiveness of struvite formation in minimizing ammonia emission and maximizing the nitrogen retention in compost, hence increasing the market value of compost; and (3) To evaluate the effects of magnesium and phosphate salt addition on composting process.  72  3.2  3.2.1  Materials and Methods  Substrate Characterization and Recipe Formulation  To compare different techniques and measures for ammonia and odor control across a series of systematic composting experiments, a single composting recipe was used throughout this thesis study. Broiler litter was used as primary composting material. Sawdust and hog fuel were added as amendment and bulking agent to adjust the C/N ratio and improve the porosity of feedstock. Physical and chemical characterization of each ingredient was measured with standard method (APHA, 1995) in the analytical laboratories of Chemical & Biological Engineering Department and Earth & Ocean Science Department, the University of British Columbia. The results of analyses are presented in Table 3.1. Table 3.1. Characterization of substrates Ingredients  Boiler liter  Sawdust  Hog fuel  Moisture (% wb )  30.0  10.7  40.5  Volatile solids (%)  82.8  94.3  ND[b]  Carbon (%)  39.1  41.6  52.9  Nitrogen (%)  4.38  1.42  0.38  NH 4 -N (g/kg)  5.58  1.46  ND  CaO (%)  3.55  0.9  0.51  Fe2O3 (%)  0.13  0.04  0.04  K2O (%)  1.43  0.04  0.01  MgO (%)  0.55  0.17  0.07  Mn3O4 (%)  0.06  0.01  0.01  Na2O (%)  0.36  0.15  0.02  P2O5 (%)  3.98  0.42  0.04  SiO2 (%)  0.85  0.48  0.21  TiO2 (%)  0.01  0.00  0.00  Al2O3 (%)  3.53  0.81  0.15  [a]  Each composition value is the average of triplicate samples [a] wb = wet basis; [b] ND = not determined 73  Getting the right combination of ingredients is the key for efficient composting. To achieve optimal course of composting, a typical recipe was developed based on desirable carbon to nitrogen ratio (C/N), moisture content and bulk density. Initially, the poultry manure had the following characteristics: 2.4 g/kg soluble phosphate, 1.1 g/kg soluble magnesium and 5.7 g/kg ammonium-nitrogen (dry weight basis). The recipe consists of 65% chicken manure, 25% sawdust and 10% hog fuel on dry mass basis. Tap water was added to attain initial moisture content of 55% for the mixture, which is generally believed to be optimal for composting.  3.2.2  Experimental Design and Setup  Struvite, or magnesium ammonium phosphate hexahydrate (MgNH4PO4·6H2O), is a white crystaline mineral. Struvite could be formed as Mg2+, NH 4 , and PO 34 reacts in 1:1:1 molar ratios and when the product of Mg2+, NH 4 , and PO 34 activities (IAP) exceeds the equilibrium ion-activity product IAPeq (Nelson et al., 2003): Mg2+ + NH 4 + PO 34 + H2O œ MgNH4PO4 · 6(H2O)  (4.1)  IAPeq = (Mg2+)( NH 4 ) ( PO 34 ) = 7.08 u 10-14  (4.2)  Since the ionic movement and activities of struvite precipitation occur in aqueous phase, the experiment was designed on the basis of ion strength in solution in the porous composting medium. The experimental design, as depicted in Table 3.2, included two sets. Set 1 was designed based on the theoretical reaction of struvite formation, in which the magnesium to phosphate (Mg/P) molar ratio in compost solution was 1.0. Set 2 increased Mg/P molar ratio to 1.25 to examine if it is advantageous to enhancing struvite precipitation and reducing the phosphate residue in compost. Experimental Set 2 was run in replicates (n = 2).  74  To form struvite, ammonium, magnesium and phosphate ions must be available for solubility reaction. For poultry manure composting, ammonium is generally in excess relative to magnesium and phosphate, as it is produced in situ by degradation of nitrogenous materials. Since magnesium is usually limited in manure wastes, magnesium salt (MgCl2) was added to feedstock to raise the magnesium ionic strength in compost solution to 0.2-0.4 mol/L, which is sufficient to induce struvite precipitation under adequate pH conditions (Stratful et al., 2001; Mijangos et al., 2004). Phosphate salt (KH2PO4) was also added in supplement to the soluble phosphate already present in poultry manure, as a further aid for struvite formation.  The experimental setup consists of four 6-liter Dewar reactors. Details of the composting system are described in Chapter 3. The temperature setpoint used was 65qC. The baseline aeration rate was set to 0.72 L/min.kg dry matter (DS) with 33% duty cycle below the setpoint and continuous above the setpoint. The compost was agitated manually at 1 to 4 days interval during the experiment. The experiments were allowed to complete a high-rate, active phase of composting, which is defined with respect to the composting mix temperature when it has dropped back to ambient level. Each run lasted for 10 to13 days. Table 3.2 Experimental design for reducing ammonia emission via struvite formation Test Series Treatment  Set 1  Set 2  [a]  Dosage (g/kg DS[a])  Mg/P ratio  Concentration (mol/L)  Mg salt  P salt  (mb[b])  Mg2+  PO 34  S1  47.8  62.6  1  0.4  0.4  S2  28.6  36.4  1  0.3  0.3  S3  18.8  22.4  1  0.2  0.2  S4 (control)  0  0  N/A  N/A  N/A  S5  59.8  62.6  1.25  0.5  0.4  S6  35.8  36.4  1.25  0.38  0.3  S7  23.5  22.4  1.25  0.25  0.2  S8 (control)  0  0  N/A  N/A  N/A  DS = dry solids or dry matter; [b] mb = molar basis; N/A: not applicable  75  3.2.3  Experimental Procedure  The uniformity of waste and environmental conditions is important in controlling the variability of substrate characteristics from run to run. In this study both substrate preparation and sampling approach are designed to address the challenges of heterogeneity and environmental variation. The broiler litter used, which comprised mainly of manure and wood shavings, was collected from a commercial chicken farm in Abbotsford, British Columbia. Broiler litter had relatively low moisture content of 30%; moreover, it was stored in a cold room whereby temperature was controlled below 4°C to prevent degradation before the composting tests. It was allowed to reach ambient temperature after retrieving from storage the night before composting mix preparation. The differences in feedstock characteristics from run to run were minimal.  In order to improve the homogeneity of the substrate, broiler litter was screened with 10 mm sieve to remove any large lumps. Substrates were mixed manually for at least 10 minutes to ensure uniformity of the mixture. Where applicable, additives tested were applied in solid form during mixing. Tap water was then added with a sprinkler and mixed thoroughly for another 10 minutes. The feedstock was allowed to sit in room temperature for 30 minutes to achieve desired moisture uniformity. Triplicate feedstock samples were taken and grouped into three sub-samples for determination of moisture content, volatile solids, pH, EC, ammonium nitrogen, ortho-phosphate, total nitrogen and total carbon in compost. The feedstock was loaded without any compaction. The weight of each reactor was registered to determine the required aeration rate and mass changes during the active phase of composting.  During the composting process, the ambient temperature was fairly constant at about 24°C. Agitation was performed manually at one to four days interval. About 50g samples were taken immediately after agitation to determine moisture content, volatile solid, pH, EC and ammonium nitrogen content. Reactors, aeration flow lines, thermocouples and other equipment were frequently checked for the signs of leakage, wear, or other potential failures throughout the course of experiments. Flowmeters were monitored regularly to ensure all connections are airtight and designated air is supplied. The reactor caps and connections  76  between reactor and the ammonia trap are of particular concern because any leaks would result in underestimate of ammonia emission.  Ammonia was captured continuously by passing the exhaust gas through a condensate trap and acid trap solution in series. The acid trap solution contains 4.8% sulphuric acid solution that matches the sulphuric acid concentration of TKN non-digestion working standards of auto analyzer method. Samples were collected every 24 hours. This method is suitable for continuous capture of ammonia, and is preferred over colorimetric tubes, which can only be used for discrete sampling purposes.  The experiments were allowed to complete a high-rate, active phase of composting, which is defined with respect to the composting mix temperature when it has dropped back to ambient level. Each run lasted for 10 to13 days. At the completion of each run, reactors were weighed and triplicate grab samples were taken to determine the final physical and chemical properties of compost. Compost for selected runs was then transferred to another reactor for curing. During the curing phase, periodical turning was applied instead of aeration. The compost moisture was kept in a range of 30% to 60% by adding tap water during turning. At the end of curing phase, the nitrogen content of the finished compost was analyzed.  3.2.4  Analytical Methods  Physical and Chemical Properties of Compost Grab samples in triplicates (about 30 g each) were used for the determination of moisture content and volatile solids content. Both moisture content and volatile solids were measured gravimetrically. Moisture content was determined by oven-drying the compost samples at 110ºC for 24 hours. The oven-dried samples were subsequently transferred to a furnace to determine the volatile solids content by igniting compost mixture at 550ºC for 2 hours (APHA, 1995).  77  Water-soluble NH4-N and ortho-p were extracted with DI water by adding 100 ml DI water to 20 g fresh compost sample, continuously stirring the suspension for 5 minutes, and shaking for another 30 minutes before measurement. The waste suspension was let stand for about 15 minutes to allow most of the suspended waste to settle out for pH and EC measurements. The pH and EC values were then measured using digital pH and EC meters, respectively. The sample solution was then centrifuged for 30 minutes and filtered twice using 0.45μm glass fiber filter before analysis. Ammonium and ortho-phosphate concentrations were measured colorimetrically following the standard method (APHA, 1995) using a flow injection analyzer (FIA) (Lachat Instrument, Model QuickChem 8000). Total nitrogen and carbon contents of the compost mixture were determined by using a Carlo Erba NA-1500 CN Analyzer at the Earth and Ocean Sciences Laboratory at UBC. Carbon-tonitrogen ratio (C/N) was then calculated with the measured total carbon and total nitrogen contents. Total P and Mg contents were measured via ignition at 950ºC for 2 hours and analyzed with an atomic adsorption spectrometer.  Ammonia Emission There are a large number of samples from each run. The use of automatic method is essential to make the experiment more efficient. Since the acid solution is not suitable for the standard method for ammonia or TKN measurement with QuickChem autoanalyzer, a modified TKN method were employed, which includes omitted the digestion procedure by modifying both standards and carrier. The ammonia emission rate was then be calculated as following:  Er  CV TM  (4.3)  Where: Er = ammonia emission rate (g/hr.kg), C = the ammonium concentration in acid trap solution (mg/L), V = volume of acid trap solution (L), T = sampling interval (hour), M = dry mass of initial compost mixture (kg).  78  Struvite Confirmation At the end of each run, about half of compost from each reactor was air-dried, and large particles were screened out from the compost for precipitate separation. Precipitates were obtained manually by gravity settling, decanting and washing the compost. The precipitates obtained were examined by X-ray diffraction (XRD) in parallel with scanning electron microscopy (SEM) equipped with energy dispersive X-ray spectroscopy (EDS) to identify the mineral compounds formed in compost and help to elucidate a possible reaction mechanism. All measurements were performed in the X-ray Diffraction Lab of the Earth and Ocean Science Department, University of British Columbia. 3.2.5  Statistical Analysis  Statistical analyses were performed using the Analysis ToolPak that Microsoft Excel provides. The single factor analysis of variance (ANOVA) was used and the results were reported in terms of p-values that represent the degree of statistical significance for the thermal performance and cumulative ammonia emission of the treatments. Differences in results between treatments are significant at P < 0.05, unless stated.  3.3  Results and Discussions  3.3.1. Struvite Formation Struvite precipitation is highly pH dependent because the activities of both PO 34 and NH 4 are pH dependent (Doyle and parsons, 2002; Nelson et al., 2003). Figures 3.1 and 3.2 show the variation of pH over time from experimental Sets 1 and 2, respectively. The changes in pH of control treatment exhibited a typical pattern found in literature. The initial pH was slightly alkaline with a value of 8.2. Following a slight increase in the first few hours of composting, the pH declined on the second day due to the production of organic acids. Thereafter, the pH began to increase as the acids produced were converted to carbon dioxide.  79  10  9  pH  8  7  6  0.4 mol/L PO4  0.3 mol/L PO4  0.2 mol/L PO4  Control  5 0  24  48  72  96  120 144 168 192 216 240 264 288 312 Time (hours)  Figure 3.1 Change in pH and standard variations of readings from Set 1 (Mg/P = 1:1)  10  9  pH  8  7  6  0.4 mol/L PO4  0.3 mol/L PO4  0.2 mol/L PO4  Control  5 0  24  48  72  96  120 144 168 192 216 240 264 288 312 Time (hour)  Figure 3.2 Change in pH and standard variations of readings from Set 2 (Mg/P = 1.25:1)  80  However, with the addition of phosphate and magnesium salts, which are acidic, the initial pH values dropped below 6.6. As composting proceeded, pH increased gradually due to aeration stripping carbon dioxide from the compost matrix, thus altering the carbonate chemistry of the solution in compost. The pH values of all treatments increased to above 7 within 3.5 days. It is generally believed that struvite precipitation occurs when the pH is within a range of 7 to 11 (Mijangos et al., 2004; Doyle and parsons, 2002; Burns and Moody, 2002). According to Stumm and Morgan (1970), the fraction of total ortho-phosphate present as the anion increased 250-fold as pH increased from 7 to 9, whereas the total ammoniacal N present as NH 4 decreased from 99% to 64% in the same range of pH. As pH further increase above 9, the solubility of struvite begins to increase since the NH3 will predominate and NH 4 ion strength will be greatly reduced (Doyle and Parsons, 2002). Therefore, a pH range of 7-9 is favorable for struvite precipitation because of the substantial increase in ortho-phosphate. The fact that pH values of the treatments were increased to 7 within 3.5 days and ended up with 8.5 to 9.0, which were within the desirable range for struvite precipitation, indicates that favorable conditions for struvite precipitation were created. The pH variation trends and final values of treatments from Sets 1 and 2 were similar, whilst he final pH values of treatments were virtually same, meaning that increasing Mg/P ratio does not significantly affect the final pH value of compost.  To verify the presence of struvite in compost, precipitates were collected from the finished compost. The precipitates were then used for the analysis of crystalline mineral phases by the means of X-ray diffractometry (XRD). Figure 3.3 shows the results of XRD analysis for the precipitates separated from the compost. The identification of the minerals was based on the location and intensity of the peaks matching with those collected in database for crystal structures. The vertical lines on the graph represent where the peaks of the mineral analyzed should occur. The results indicate that both the positions of the peaks on x-axis and the intensity of the peaks on y-axis identified struvite as main phase present in the precipitates. Some quartz has also been identified in the precipitates likely because poultry manure initially contains quartz. The quartz found in the compost might assist the struvite crystal growth as it has been used as seed material in struvite crystallization in wastewater. All other  81  elements in the precipitates were minor and could not be established through the XRD analysis. Struvite crystals were not observed from the control treatment.  This fact has also been corroborated by SEM/EDS microanalysis performed on the precipitates obtained. Struvite crystals have a distinctive shape and can be identified by their orthorhombic structure. The typical shapes of struvite could be needle-like, rod-like or pyramidal depending its forming environment. As shown in Figure 3.4, the crystals display a typical orthorhombic pyramidal shape that is close to the SEM view of the Sigma pure struvite reported by Wu and Bishop (2004), indicating that the struvite crystals formed in poultry manure composting media are relatively pure struvite. Jeong and Kim (2001) also found the orthorhombic pyramidal shape of struvite from food waste composting media.  In fact, the size of struvite crystals discovered in the poultry manure composting media was as large as 0.5mm in length and can be visualized without magnification. Furthermore, EDS microanalysis revealed that magnesium and phosphate were definitely presented in the crystals, and the peaks and position matched very well with struvite, as shown in Figure 3.5, reaffirming that the precipitates obtained from compost are indeed struvite. It was noticed that a small potassium-peak also appeared on the EDS graph. This is likely from the residual of K as it was added with KH2PO4 as supplemental source of phosphate. It is also possible to form K-struvite (KMgPO4·6H2O) and HAP (Ca5(PO4)3·O) in the composting process. However, neither K-struvite nor HAP could be established by XRD, indicating that their quantities were negligible.  The results demonstrate that struvite precipitation does not necessarily limit to supersaturated liquid media or high moisture media such as manure slurry. It also could be formed in porous composting media at moisture content around 55%, which is optimal for composting. This finding is of practical importance, because it means that the goal of struvite formation in compost media could be achieved without compromising optimal moisture condition, thereby the negative effects of odor emissions due to high moisture can be avoided.  82  800 0  700 0  600 0  Lin (Cps)  500 0  400 0  300 0  200 0  100 0  0 4  10  20  30  40  50  60  2-Theta - S cale Wzhang-1 - F ile: Wzhang-1.raw - Ty pe: Loc ked Coupled - Start: 3.119 ° - End: 70. 097 ° - Step: 0.040 ° - St ep time: 1. s - Temp.: 25 ° C (Room ) - Time Started: 7 s - 2-Thet a: 3 00-015-0762 (* ) - Struvit e, sy n - NH4MgPO4·6H2O - Y: 21. 43 % - d x by: 1. - WL: 1.5406 - Ort horhom bic - a 6.94500 - b 11.20800 - c 6.13550 - alpha 90.000 - beta 90.000 - g 00-046-1045 (* ) - Quartz , s yn - SiO2 - Y: 79.17 % - d x by: 1. - WL: 1.5406 - Hex agonal - a 4.91344 - b 4.91344 - c 5. 40524 - alpha 90.000 - beta 90.000 - gam ma 120.000 - P  Figure 3.3 X-ray diffraction spectrum of the precipitates obtained from the final compost  83  PP  A  PP  B  Figure 3.4 Scanning electron micrograph (SEM) of the precipitates obtained from compost after active phase of composting. (A) SEM view of air-dried crystals; (B) Magnified view of air-dried struvite.  Figure 3.5 Energy dispersive X-ray spectroscopy (EDS) analysis of the precipitates obtained from compost after active phase of composting.  84  3.3.2. Ammonia Emission To examine the effectiveness in reducing ammonia emission via struvite formation, ammonia released through exhaustion was captured and measured. Figures 3.6 and 3.7 show the cumulative ammonia emissions from all treatments. The cumulative ammonia emissions in Figure 3.7 are the average values of two replicates. Lines were computer generated trend lines. The behavior of ammonia emission was dramatically changed due to Mg and P salt addition. Ammonia from the control treatment released from the onset of composting, and continued to release to some extent through the end of active phase, whereas ammonia emissions from treatments with magnesium and phosphate salts had a time lag of 2 to 5 days, depending on the dosage of Mg and P slats. The lag of ammonia release of treatments could be explained by two reasons. Firstly, the pH values of the treatments dropped to below 6.6 due to Mg and P salt addition. When pH is below 7, ammonium almost exclusively exists in aqueous phase, and ammonia volatilization is very limited (Sawyer and McCarty, 1978; Liang, 2000). Secondly, ammonia volatilization is greatly influenced by temperature. Because of the acidic conditions within the composting matrix, the temperatures of the treatments picked up slowly and were consistent with the lag phase of ammonia release. In comparison with the control treatment, the peak ammonia emission rates were reduced by 43 to 74%, depending on the dosages of Mg and P salts.  In all cases, the cumulative ammonia emission decreased proportionally with increased magnesium and phosphate concentration in compost solution. The total N loss through ammonia emission was 4.16g per kg of initial dry mass from control treatment, whereas treatments at phosphate and magnesium solution concentration level of 0.2, 0.3 and 0.4 mol/L lost only 2.48, 1.64 and 0.68 g N through exhaust gas. Total ammonia emission was reduced by 40% to 84% due to Mg and P salt addition, suggesting that some ammonium in compost solution was precipitated in the form of MgNH4PO4 in the compost matrix before being transformed to NH3 and volatilized.  85  As shown in Figure 3.7, increasing Mg/P molar ratio from 1.00 and 1.25 resulted in lower ammonia emission rates and less total ammonia emission. Table 3.3 summarizes the results of total ammonia emission for replicates of treatment with Mg/P molar ratio of 1.25. The details of statistical analysis can be found in Appendix III. The difference in total ammonia emission between treatments is significant (p < 0.005). The further reduction in ammonia emission may attribute to struvite formation, because increasing Mg/P molar ratio has little effect on pH as discussed previously. Therefore, higher Mg/P molar ratio may benefit to struvite formation. The results are in agreement with those reported from struvite precipitation in wastewater. A number of researchers have shown that increased Mg/P molar ratio resulted in lower phosphate residual and subsequently greater amount of struvite precipitation (Stratful et al., 2001, Nelson et al., 2003). The substantial ammonia reduction is of practical significance because high ammonia concentration on-site is a major health concern for workers at composting facilities.  86  Table 3.3 Summary of the temperature parameters and total ammonia emission for replicates of treatment with Mg/P molar ratio of 1.25 Treatment 1 Treatment 2 Treatment 3 Control Replicate 1 Replicate 2 Replicate 1 Replicate 2 Replicate 1 Replicate 2 Replicate 1 Replicate 2 Thermal Performance Peak temperature (oC)  65  63  67  60  68  65  73  70  Time to reach peak temperature (hour) 122  100  124  98  50  72  19  40  Duration of temperature > 55 oC (hour) 97  123  103  129  118  149  106  82  296  1205  855  1508  1505  2744  2981  Ammonia Emission  Cumulative ammonia emission at the end of active phase (mg/kg)  810  Treatment 1: 0.4 mol/L P salt; Treatment 2: 0.3 mol/L P salt; Treatment 3: 0.2 mol/L P salt.  87  Cumulative Ammonia Emission (mg NH3-N/kg)  4,500 4,000  0.4 mol/L P Salt 0.3 mol/L P Salt  3,500  0.2 mol/L P Salt 3,000  Control  2,500 2,000 1,500 1,000 500 0 0  24  48  72  96  120 144 168 192 216 240 264 288 312 Time (hours)  Figure 3.6 Cumulative ammonia emissions of different treatments as affected by Mg and P salt addition (Mg:P = 1:1) 3,500  Ammonia Emission (mg NH3-N/kg)  0.4 mol/L P salt 0.3 mol/L P salt  3,000  0.2 mol/L P salt Control  2,500 2,000 1,500 1,000 500 0 0  24  48  72  96  120  144  168  192  216  240  264  288  Time (hours)  Figure 3.7 Cumulative ammonia emissions of different treatments as affected by Mg and P salt addition (Mg:P = 1.25:1) (all curves are the means of two replicates)  88  3.3.3. Nitrogen Conservation One of the primary objectives of this study is to maximize nitrogen conservation to increase the fertilizer value of compost. To evaluate nitrogen conservation of the treatments, watersoluble ammonium nitrogen retained in compost after the active phase was measured with the procedure described in section 4.2 and results are shown in Figures 3.8 and 3.9. The water-soluble NH 4 -N contents generally increased as test progressed regardless of treatments. The increased trend was expected. On one hand, the ammonification during composting produces ammonium nitrogen through transforming some of the organic N to inorganic NH3 or NH 4 . On the other hand, the mass of compost decreases as a result of decomposition of organic materials. Both factors would result in increase in ammonium nitrogen content in compost. At the end of active phase, the water-soluble ammonium nitrogen content of control treatment was 5.4 g/kg DS from Set 1. With the addition of Mg and P salts, however, the ammonium nitrogen contents were 7.2, 8.2 and 8.3 g/kg DS for Mg and P concentration levels of 0.2, 0.3 and 0.4 mol/L, respectively, which were 32.3% to 52.3% higher than that of control treatment. The results are in a general agreement with those by Jeong and Kim (2001), who reported that the addition of Mg and P salts in food waste composting increased ammonium nitrogen content up to 14.3 g/kg DS. The high ammonium nitrogen observed implies that ammonium was excessive relative to magnesium and phosphate concentration requirement for struvite formation. The high free ammonium concentration is actually favorable for struvite precipitation. Stratful et al. (2001) reported that excess ammonia is highly beneficial to the precipitation of struvite.  Since water-soluble ammonium nitrogen retained in compost during active phase may not be totally stable and would continue to release during the curing phase. Curing experiment was carried out to further examine the nitrogen conservation in finished compost. After eight months of curing, finished compost samples were taken to measure the nitrogen content in soluble and insoluble phases. These samples are considered to be reasonably stable, as the compost had gone through such a long period of maturation.  89  9.0 8.0  NH4-N Content (g/kg)  7.0 6.0 5.0 4.0  0.4 mol/L P Salt 0.3 mol/L P Salt  3.0  0.2 mol/L P Salt  2.0  Control 1.0 0.0 0  24  48  72  96  120 144 168 192 216 240 264 288 312 Time (hours)  Figure 3.8 Variation of water soluble ammonium nitrogen in compost over time as affected by Mg and P salt addition (Mg:P =1:1). 10.0 9.0  NH4-N Content (g/kg)  8.0 7.0 6.0 5.0 4.0  0.4 mol/L P Salt  3.0  0.3 mol/L P Salt 0.2 mol/L P Salt  2.0  Control 1.0 0.0 0  24  48  72  96  120 144 168 192 216 240 264 288 312 Time (hour)  Figure 3.9 Variation of water soluble ammonium nitrogen in compost over time as affected by Mg and P salt addition (Mg:P = 1.25:1).  90  Inorganic ammonium nitrogen in compost is commonly measured by water extraction because ammonia in ordinary compost exists as either ammonia gas or in an ionized state (Klleher et al., 2002). In this study, some ammonium nitrogen was precipitated into struvite, which is normally insoluble in water. Therefore, the nitrogen conservation in the compost should also include the insoluble form of ammonium nitrogen (i.e. struvite). Since struvite can be dissolved in low pH solution, in this study the total ammonium nitrogen in compost was measured by extraction with low pH solution. The ammonium nitrogen in solid phase was then calculated from the differences between total ammonium nitrogen extracted with low pH solution and the water-soluble ammonium nitrogen extracted with DI water. Without digestion, the chances for transforming organic nitrogen to inorganic form should be minimal during a 30-minute extraction period. Therefore, the nitrogen from the struvite can be estimated with the ammonium nitrogen in solid phase, because other than struvite, insoluble compounds containing ammonium would not be stable enough to exist for a period of eight months.  Sulphuric acid (4.8%) was used to dissolve struvite bound nitrogen and extract the inorganic ammonium nitrogen in both soluble and insoluble phases. Extraction was performed by adding 100 ml DI water to 20 g of fresh sample, shaking for 30 minutes, centrifuging for 20 minutes, and filtering twice with 0.45μm glass fiber filters before analysis, which is similar to the DI water extraction procedure. The NH4+-N was colorimetrically determined following the standard methods (APHA, 1995) using an automated ion analyzer (Lachat QuickChem, Zellweger Analytic, Inc.). The nitrogen content in soluble phase was measured with DI water extraction as described previously.  Table 3.4 presents the ammonium-nitrogen contents in the finished compost measured with samples from experimental Set 1. The results show that the ammonium nitrogen from insoluble part that represents ammonium nitrogen from struvite, were 4.63 and 3.47 g/kg for treatment with highest and lowest Mg and P dosages, respectively, which were about 45% and 65% of the total inorganic NH 4 -N found in the compost, implying that increasing dosage resulted in a slightly lower efficiency of struvite precipitation. The results also suggest that the efficiency of precipitating ammonia into struvite might not be high (< 65%). This can be  91  attributed to the interference of the organics during the struvite formation in compost solution. Schuilling and Andrade (1999) found that the struvite crystallization process in wastewater is interfered when total suspended solids concentration is above 1000mg/L.  It is particularly interesting to note that the ammonium nitrogen of treatments with highest Mg and P dosages was more than doubled in total compared with the control treatment. Even for the treatment with lowest dosage, a 37% increase in ammonium nitrogen was achieved. The ammonium nitrogen concentrations in solid phase were also increased by 49.4% and 99.3% for the treatments with highest and lowest dosages, respectively. Although the struvite precipitation efficiency in composting was lower than that in wastewater, these are still favorable results as the primary goal of this study is to reduce ammonia emission and enhance the nitrogen conservation in compost rather than producing pure struvite. Table 3.4 Inorganic ammonium nitrogen retained in compost after curing stage[a] S1 S3 S4 Treatment (0.4 mol/L)[b] (0.2 mol/L) (Control) Total NH4-N (g/kg DS)  9.95  6.65  4.85  Soluble NH4-N (g/kg DS)  5.32  3.18  2.52  Insoluble NH4-N[c] (g/kg DS)  4.63  3.47  2.32  Total NH4-N increase (%)  105.4%  37.2%  -  Insoluble NH4-N increase (%)  99.3%  49.4%  -  [a]  The curing stage lasted for 8 months; Molar concentration of Mg2+ and PO 34 in compost solution; [c] The NH4-N from insoluble struvite precipitates. [b]  3.3.4. Changes in Water Soluble Orthophosphate Figures 3.10 shows the changes in orthophosphate (PO4-P) concentrations in compost extracted with DI water. Without Mg and P salt addition, water-soluble PO4-P concentration in compost was low because the majority of inorganic phosphate is particulate-bound. Consequently, changes in water-soluble PO4-P concentration during composting were  92  insignificant. With Mg and P salt addition, however, the initial water-soluble PO4-P concentrations were considerably high due to the water-soluble phosphate salt added. As test progressed, the water-soluble PO4-P concentration gradually decreased and leveled off after five days of composting. At the end of active phase, water-soluble PO4-P concentrations were decreased by 60.3 to 82.4%. The significant decrease may attribute to the struvite precipitation, given the fact that struvite presence in the compost has been confirmed by XRD in parallel with SEM/EDS analysis, and significant amount of ammonium nitrogen in solid phase was found in the finished compost. Although it is also possible to form calcium struvite (HAP, Ca5(PO4)3OH) during the composting process, which could also contribute to the reduction of water-soluble phosphate, the kinetics of this process is extremely slow (Musvoto et al, 2000). Furthermore, magnesium concentration was significantly higher than that of calcium in compost solution due to addition of Mg salt. Magnesium ion can kinetically inhibits the precipitation of HAP (Abbona, 1990; Amjad et al., 1984; Doyle and Parsons, 2002), especially when Mg/Ca molar ratio is higher than 0.6.  It was noticed that the residue of water-soluble phosphate of treatments observed in the final compost was high, suggesting that available Mg ion might be a limited factor for struvite precipitation. Therefore, further experiment was conducted by increasing the Mg/P molar ratio to 1.25 to reduce the soluble phosphate residue in compost. As shown in Figure 3.11, water-soluble PO4-P concentrations in final compost were decreased by 72.1 to 85% compared to their initial concentrations. The phosphate residues were significantly reduced. The results are similar to the findings from struvite precipitation in water and wastewater. Stratful et al. (2001) chemically formed struvite in water and demonstrated that a lower phosphate residue can be achieved by increasing the Mg/P molar ratio. Nelson et al. (2003) also reported that increasing the Mg/P molar ratio resulted in decreased water-soluble PO4-P concentration in swine lagoon liquid.  The removal efficiency of phosphate from compost solution was evaluated with the changes in total mass of water-soluble PO4-P in initial and final compost. As shown in Table 3.5, the phosphate removal efficiency of treatments with Mg/P molar ratio of 1:1 was in the range of 66.5-85.3%. Increasing Mg/P molar ratio to 1.25:1 improved phosphate precipitation  93  efficiency to 76.4-87.5%. Higher Mg and P salt dosages resulted in slightly lower phosphate removal efficiency, and thus a little higher residue of soluble phosphate in compost. Therefore, a dosage of phosphate salt lower than 22 g/kg on dry mass basis (i.e. PO4-P molar concentration in compost solution < 0.3 mol/L) is recommended for struvite formation in poultry manure composting.  3.3.5. Impacts on the Composting Process Jeong and Kim (2001) suggested that adding Mg and P salts into substrate could suppress the microbial activity responsible for the composting reaction, and thus might have potential negative effects on the composting process. Adding salts may also cause salinity problem and affect the quality of the finished compost. This section addresses these issues via examining some important indicators of composting process and soluble salts in compost, including temperature, degree of degradation and electrical conductivity.  Figures 3.12, 13 and 14 show the temperature profiles of all treatments. The thermal performance parameters are tabulated in Table 3.6. The addition of P and Mg salts slowed the composting process, reduced the peak temperature and prolonged the thermophilic phase (i.e. temperature above 45oC). The control treatment heated up faster and reached peak temperature within 17 hours. With P and Mg salt addition, it took 50 to 127 hours to reach the peaks depending on the dosages used. The more the P and Mg salts added, the slower the temperature picked up. The peak temperatures of the control treatment were 72-73oC, whereas the peak temperatures of treatments with highest salt dosage were never beyond 65oC. For Set 2 tests with two replicates, ANOVA analysis indicated that differences in the peak temperatures for the various treatments are quite insignificant (p = 0.15). The thermophilic phase of the control lasted for about 150 hours. However, the thermophilic phase of the other treatments was maintained up to 218 hours, which is almost 3 days longer than that of the control treatment. The lowered peak temperature and extended thermophilic phase may in fact be beneficial to the microorganisms responsible for composting, as most species of microorganisms may not survive at temperatures above 65°C and microbial activity is maximized in the 35 to 45°C temperature range (Day and Shaw, 2001; Richard,  94  1996). All treatments met the pathogen destruction standard that requires minimum temperature of 55oC for at least three consecutive days. Therefore, under the dosages tested there were no adverse effects on the composting process in terms of temperature condition.  Table 3.7 shows the volatile solids and degree of degradation of all treatment. The degree of degradation was calculated on the basis of total mass changes of organic material during the composting process. At the end of active phase, the degree of degradation of all treatments was in a narrow range of 21.6 to 24.6% with standard deviations ranged from 0.05 to 1.16. Statistically, there are no significant differences in the degree of degradation between the treatments. The results further demonstrate that the application of magnesium and phosphate does not adversely affect the composting process as long as the dosages are adequately controlled.  The soluble salts in compost were measured in terms of electrical conductivity (EC) during the composting process and presented in Figure 3.15 and 3.16. As expected, the addition of Mg and P salts had pronounced effects on EC. Regardless of treatments, changes in EC were insignificant during the active phase. However, EC values increased proportionally to the application rate of P and Mg salts. In the conclusion of tests, EC values of the treatments were between 0.29 and 0.61 mmhos/cm. It is generally recognized that compost EC level of 0.35-0.64 mmhos/cm is a desirable range for most of the plants grown on compost (Alexander, 1994). Therefore, the addition of P and Mg salts would not cause salinity problems to the compost product as long as the dosages are control below the high level tested in this study.  95  12 0.4 mol/L P Salt  10 Water soluble P (g/kg)  0.3 mol/L P Salt 0.2 mol/L P Salt  8  Control 6  4  2  0 0  24  48  72  96  120 144  168 192 216 240 264 288 312  Time (Hour)  Figure 3.10 Changes in water-soluble orthophosphate with time (Mg:P = 1:1) 14 1412 0.4 mol/L P Salt Water soluble P (g/kg) Water soluble P (g/kg)  12 10  0.3 mol/L P Salt 0.4 mol/L 0.2 mol/L P Salt 0.3 mol/L Control 0.2 mol/L  10 8 8  Control  6  64 4  2  20 0  24  48  72  96  0 0  24  48  72  96  120 144  168 192 216  Time (hours) 120 144 168 192 216  240 264 240 264  288 312 288 312  Time (hours)  Figure 3.11 Changes in water-soluble orthophosphate with time (Mg:P = 1.25:1).  96  Table 3.5 Changes in water-soluble orthophosphate in compost during the active phase Test Series  P Salt Conc. (mole/L)  Mg/P ratio (mb[a])  Initial PO4-P g/kg DS  Final PO4-P g/kg DS  Precipitation Efficiency  Set 1  0.4  1  10.65  4.22  66.5%  0.3  1  9.66  2.62  77.6%  0.2  1  5.95  1.05  85.3%  Control  N/A  0.40  0.62  í  0.4  1.25  11.04  3.29  74.3%  0.3  1.25  8.16  1.63  83.2%  0.2  1.25  4.77  0.67  88.3%  Control  N/A  0.36  0.53  í  Set 2  Values are the means of two replicates for Set 2 Test [a] mb = molar basis  Table 3.6 Thermal performances of treatments with Mg and P salts Test Series  P Salt Conc. (mole/L)  Mg/P ratio (mb)  Tp (oC)  tp (hour)  t55 (hour)  Set 1  0.4  1  65  127  124  0.3  1  66  124  134  0.2  1  67  54  138  Control  N/A  72  10  102  0.4  1.25  64  121  110  0.3  1.25  63  120  116  0.2  1.25  67  55  134  Control  N/A  71  19  94  Set 2  Values are the means of two replicates for Set 2 Test mb = molar basis; Tp = peak temperature; tp = time elapsed to reach to peak temperature; t55 = cumulative hours with temperature above 55oC.  97  Table 3.7 Changes in volatile solids and degree of degradation of different treatments  Test Series  Treatment  Mg/P ratio  Volatile solids (%)  mb[a] Set 1  Set 2  Degree of degradation %  Increase  0.4  1  81.0  23.8  0.2  0.3  1  82.4  24.4  0.8  0.2  1  84.1  22.0  -1.6  Control  N/A  87.1  23.6  0.0  0.4  1.25  81.1  24.0  1.9  0.3  1.25  82.5  22.3  0.1  0.2  1.25  84.4  22.0  -0.1  Control  N/A  87.5  22.1  0.0  Values are average of triplicate samples for Set 2 Test [a] mb: molar based  98  80 75 70 65 60 55 Temperature (°C)  50 45 40  Mixing point Mixing points  35 30 25 20 15  0.4 mol/L  10  0.3 mol/L  0.2 mol/L  Control  Ambient  5 0 0  24  48  72  96  120  144  168  192  216  240  Time (hours)  Figure 3.12 Temperature profiles of treatments with Mg/P molar ratio of 1:1 99  264  288  312  80 75 70 65 60  Temperature (°C)  55 50 45 40  Mixing point  35 30 25 20 15 0.4 mol/L  10  0.3 mol/L  0.2 mol/L  Control  Ambient  5 0 0  24  48  72  96  120  144  168  192  216  240  Time (hours)  Figure 3.13 Temperature profiles of treatments with Mg/P molar ratio of 1.25:1  100  264  288  75 70 65 60 55  Temperature (°C)  50 45 40 35 30 25 20 15 0.4 mol/L  10  0.3 mol/L  0.2 mol/L  Control  Ambient  5 0 0  24  48  72  96  120  144  168  192  216  240  264  Time (hours)  Figure 3.14 Temperature profiles of treatments with Mg/P molar ratio of 1.25:1 (replicates)  101  288  312  0.7  Soluble Salts (mmhos/cm)  0.6 0.5 0.4 0.3 0.2 0.1  0.4 mol/L  0.3 mol/L  0.2 mol/L  Control  0.0 0  24  48  72  96  120 144 168 192 216 240 264 288 312 Time (hr)  Figure 3.15 Changes in EC with time from Experimental Set 1 ((Mg:P = 1:1)  0.8  Soluble Salts (mmhos/cm)  0.7 0.6 0.5 0.4 0.3 0.2 0.1  0.4 mol/L  0.3 mol/L  0.2 mol/L  Control  0.0 0  24  48  72  96  120 144 168 192 216 240 264 288 312 Time (hours)  Figure 3.16 Changes in EC with time for Experimental Set 2( Mg:P = 1:25)  102  3.4  Conclusions  Reducing ammonia emission via struvite formation has the advantage of binding the two significant fertilizer components, nitrogen (N) and phosphorus (P) nutrients in manure to form struvite, which is a slow release inorganic fertilizer, albeit derived from organic resource - the poultry manure. So far, there is no report on struvite formation in composting media other than high moisture food waste. A series of laboratory-scale experiments were conducted to examine the feasibility of struvite formation in poultry manure composting medium, and determine the effectiveness of reducing ammonia emission via struvite formation. The main conclusions from this series of experiment are as follows:  (1) This study demonstrated that struvite precipitation in poultry manure composting media is technically feasible. The struvite crystals formed in compost were confirmed by both X-ray diffraction (XRD) and scanning electron microscopy coupled with energy dispersive X-ray spectroscopy (SEM-EDS) analyses. This is the first time to discover struvite crystals formed in poultry manure composting media. This discovery could be practically beneficial to the composting industry in the future, as struvite has proven to be good slow-released fertilizer. (2) It is possible to form struvite at a moisture level of 55%, implying that struvite could be precipitated in composting media without compromising the optimal moisture for composting. (3) The addition of Mg and P salts in poultry manure composting mixture could reduce ammonia emission by 40-84% depending on the salt dosages. The substantial reduction in ammonia emission could be attributed to struvite formation during composting, given the fact that struvite crystals were observed in compost. (4) After curing phase, ammonium nitrogen in finished compost increased by 37.2% to as high as 105.4 % compared with control treatment. Nitrogen retention was greatly enhanced as a result of struvite formation in compost. (5) Adding Mg and P salts to feedstock could affect the composting process. However, under the dosage levels tested, no harmful or detrimental effects were observed. As a matter of fact, it extended thermophilic phase, which may be beneficial for pathogen destruction.  103  (6) The phosphate removal efficiency of treatments with Mg/P molar ratio of 1.0 was in a range of 66.5-85.3%. Increasing Mg/P molar ratio to 1.25 improved phosphate removal efficiency to 76.4-87.5%. A dosage of phosphate salt lower than 22 g/kg on dry mass basis is recommended. (7) More research is needed to understand the kinetics of struvite precipitation in composting process. To implement this technology in practice, organic wastes or other cost-effective materials rich in magnesium and phosphate need to be investigated so that external source of chemicals can be avoided or minimized.  104  3.5  References  Abbona, F., 1990, Crytallization of calcium and magnesium phosphates from solutions of low concentration. J. Cryst. Growth, 104:661-671. Alexander, R.A., 1994. Standards and guidelines for compost use. Biocycle 35:37-41. Amjad, Z., P.G. Koutsoukos and G.H. Nancollas, 1984, The crystallization of hydroxyapatite and fuorapatite in the presence of magnesium ions, J. Colloid Sci. Interface Sci. 101 (1):250-256. APHA, AWWA, WEF, 1995. Standard methods for the examination of water and wastewater. 19th ed. Publication Office APHA, Washington, DC. Burns, R.T., Moody, L.B., 2002. Phosphorous recovery from animal manures using optimized struvite precipitation. Proceedings of coagulants and Flocculants: Global Market and Technical Opportunities Treatment Chemicals, Chicago, Illinois. May 22-24, 2002. Busca, G. and C. Pistarino. 2003. Abatement of ammonia and amines from waste gases: a summary. Journal of Loss Prevention in the Process Industry. 16:157-163. Dao, T.H., L.J. Sikora, A. Hamasaki, and R.L.Chaney. 2001. Manure phosphorous extractability as affected by Aluminum- and Iron Bu Product and Aerobic Composting. J. Environ. Qual. 30:1693-1698. DeLaune, P.B., Moor P.A., Daniel, T.C., Lemunyon, J.L., 2004. Effect of chemical and microbial amendments on ammonia volatilization from composting poultry litter. J. Environ. Qual. 33:728-734. Doyle, J.D., Parsons, S.A., 2002. Struvite formation, control and recovery. Water Research 36:3925-3940. Folton, G.K., L.E. Carr, C.E. Prigge and J.C. Bouwkamp. 2004. Nitrogen and Phosphorous Dynamics in Cocomposted Yard Trimmings and Broiler Litter. Compost Science & utilization. Vol. 12. No.4 349-355. Hansen, R.C., Keener H.M., and Hoitink H.A.J., 1989. Poultry manure composting-An exploratory study. Trans. ASAE 36:2151-2157.  105  Haug, R.T. 1993. The Practical Handbook of Compost Engineering. Lewis Publishers, Boca Raton, FL. Jeong, Y. K., J.S. Kim 2001. A new method for conservation of nitrogen in aerobic composting processes. Bioresource Technology 79:129-133. Jeong, Y. K., J.S. Hwang, 2004. Optimum doses of Mg and P salts for precipitating ammonia into struvite crystals in aerobic composting. Bioresource Technology 79:129-133. Kelleher, B.P., J.J. Leahy, A.m. Henihan. T.F. O’Dwyer. D. Sutton, M.J. Leahy, 2002. Advances in poultry litter disposal technology- a review. Bioresource Technology 83:2736. Kerem Gungor and K.G. and Karthikeyan, 2007. Phosphorous forms and extractability in dairy manure: A case study for Wisconsin on-farm anaerobic digesters. Bioresource Technology. Online available at WWW.sciencedirect.com. Kithome, M., J. W. Paul and A.A.Bomke.1999. Reducing nitrogen losses during simulated composting of poultry manure using adsorbents or chemical amendments, J. Environ. Qual. 28 (1):194-201. Kirchmann, H. and E. Witter, 1989, Ammonia volatilization during aerobic and anaerobic manure decomposition. Plan Soil 115:35-41. Lachat Instruments, 1995. QuickChem 8000 continuum series automated ion analyzer, Lachat, Inc., Milwaukee. Liao, P.H., A.C. May, and S.T. Chieng. 1995. Monitoring process efficiency of a full-scale in-vessel system for composting fisheries wastes. Bioresource Technology 54:159–163. Liao, P.H., A.T. Vizcarra, A. Chen, and K.V. Lo. 1996. Composting of salmon farm mortalities with passive aeration. Compost Science & Utilization 2(4):58–66. Liang, Y. 2000. Nitrogen retention in the high rate stage of composting. Ph.D. Dissertation, Department of Agricultural, Food and Nutritional Science, University of Alberta. Mijangos F., Kamel, M., Lesmes, G., Muraaviev, D.N., 2004. Synthesis of struvite by ion exchange isothermal supersaturation technique. Reactive & Functional Polymers 60:151161. Miller, F.C., B.J. Macauley, and E.R. Harper. 1991. Investigation of various gases, pH and redox potential in mushroom composting phase I stacks. Australian Journal of Experimental Agriculture 31:415–425.  106  Musvoto EV, Wentzel MC, Ekama GA. Integrated chemical–physical process modelling II. Development of a kinetic based model for weak acid/base systems. Water Res. 2000, 34:1868–80. Nelson N.O., R.L. Mikkelsen, D.L. Hesterberg, 2003. Struvite precipitation in anaerobic swine lagoon liquid: effect of pH and Mg:P ratio and determination of rate constant. Bioresource Technology 89:229–236. Ohlinger, KN, Young TM, Schroeder ED,1999. Kinetics effects on preferential struvite accumulation in wastewater. J Environ Eng.125:730–7. Ongley, E.D., 1996. Control of Water Pollution from Agriculture. FAO paper 55. Richard, T.L., 1996. "The Science and Engineering of Composting." Web page available at http://compost.css.cornell.edu/science.html. Rynk, R. (ed). 1992. On farm composting handbook. NRAES-54, Cooperative Extension Service, Northeast Regional Agricultural Engineering Service, Ithaca, NY, USA. Sawyer, C.N. and P.L. McCarty, 1978. Chemistry for Environmental Engineering, 3rd Edition. McGraw-Hill Book Company, NY. Schuiling, R.D., Andrade, A., 1999. Recovery of struvite from calf manure. Environ. Technol. 20,765–768. Scope Newsletter, 2001. Fertilizer value of struvite. Scope Newsletter 43:3-4. Strarful I., M.D. Scrimshaw and J.N. Lester, 2001. Conditions influencing the precipitation of magnesium ammonia phosphate. Wat. Rec. Vol. 35, No.17: 4191-4199. Stumm, W., and J.J. Morgan, 1970. Aquatic chemistry. Wiley-Interscience, New York, NY. Szanto, G.L., H.V.M. Hamelers, W.H. Rulkens, A.H.M. Veeken, 2006. NH3, NO2 and CH4 emissions during passively aerated composting of straw-rich pig manure. Bioresource Technology. Online available at WWW.sciencedirect.com. Timmenga and Associates Inc., 2003. Evaluation of Option for Fraser Valley Poultry Manure. Report prepared for: Broiler Hatching Egg Producers’ Association, BC Chicken Growers Association; BC Turkey Association, and Fraser Valley Egg Producers’ Association. Toor G.S., J.D. Peak and J.T. Sims, 2005. Phosphorous speciation in broiler litter and turkey manure produced from modified diets. J. Envirn. Qual. 34:687-697.  107  Van Ruiten Adviesbureau, 1998. Phosphate recovery from animal manure: the possibilities in the Netherlands. Report for CEEP. Website available at: http://www.nhm.ac.uk/researchcuration/projects/phosphate-recovery/VanRuiten.pdf Witter, E., Kirchmann, H., 1989. Peat, zeolite, and basalt as absorbent of ammonical nitrogen during manure decomposition. Plant Soil 115:43-52. Wrigley, T.J., Webb, K.M., Venkitachalm, H., 1992. A laboratory study of struvite precipitation after anaerobic digestion of piggery wastes. Bioresource Technology 41:117-121. Wu, Q.Z. and P.L. Bishop, 2004. Enhancing struvite crystallization from anaerobic supernatant. J. Environ. Eng. Sci. 3:21-29. Zhang, Q., J.J.R. Feddes, I.K. Edeogu and X.J. Zhou. 2002.Correlation between odour intensity assessed by human assessors and odour concentration measured with olfactometers. Can. Biosystems Eng, 44: 6.27 –6.32.  108  CHAPTER 4 EFFECTS OF ADDITIVES IN THE FORM OF YEASTS, ZEOLITE AND ALUM ON THE COMPOSTING PROCESS AND ODOR EMISSIONS  4.1  Introduction  Animal manure management is one of the environmental issues that have been presented a significant challenge to environmental protection and sustainable development today. The current practice of land application of manure wastes, for instance, in the Lower Fraser Valley of British Columbia, Canada, which has both high population and livestock densities, is considered unsustainable because of environmental concerns about water and air pollution. Composting is being perceived as an essential element of the holistic approach in solving these environmental challenges via converting organic wastes into useful resources such as organic fertilizer. However, several factors are hindering the further growth and development of the composting industry; these include slow degradation rates, undesirable secondary emissions (odor and ammonia), and economics. Even with proper management practices, ammonia and odor generation are inevitable results of the decomposition of organic matters during the composting process; hence, odor pollution is considered a major problem for existing composting facilities (CIWMB, 2007; Chou and Büyüksönmez, 2006; Haug, 2004; Gage, 2003). Traditionally, odor has been regarded primarily as a nuisance, but the effects of odors on health and ambient air quality are now receiving more rigorous scientific study (Schiffman and Williams, 2005).  Conventional means for odor control involve collection and treatment using absorption, adsorption, biofiltration or thermal oxidation. These solutions require additional equipment and/or space, which could be costly and do not address the source of the problem. Another strategy is pollution prevention via reducing the release of odorous compounds and ammonia from the composting process. The use of manure additives, in the form of biological or chemical products, is not a new concept. Nevertheless, products continue to evolve in recent years. Bioadditives that are more eco-friendly than chemical additives have the potential for achieving this goal with continued research efforts. A version of this chapter has been accepted for publication. Zhang, W. and Lau, A.K. (2008) Effects of Bioadditive in the Form of Yeasts and Zeolite on the Composting Process and Odor Emission. International Journal of Environment and Waste Management. 109  4.2  Background Research  The substantial loss of ammonia during composting not only leads to odor and potential health problems on-site, but also reduces the fertilizer and market value of compost due to nitrogen nutrient loss. Natural zeolites are known to be effective in reducing ammonia emanating from animal houses and livestock slurry because of their high adsorptive capabilities (Nakaue et al., 1981 and Miner et al., 1997). Alum is one of the most effective manure amendments for ammonia emission control. Alum additions to poultry litter (without composting) have been shown to reduce ammonia emission by 99% in lab studies, resulting in much higher total nitrogen in alum-treated litter than normal litter (Moore et al., 2000). Kithome et al. (1999) investigated reducing ammonia losses during composting of poultry manure with various amendments including zeolite, clay, coir, alum [or aluminum sulfate Al2(SO4)3.14H2O], and chloride and sulfate salts of calcium and magnesium. They found that a layer of 38% zeolite clinoptilolite placed on the surface of the material reduced ammonia losses by 44%, whereas the 20% alum treatment reduced ammonia losses by 28%. The advantages of using zeolite or peat are that they are non-hazardous and good soil conditioners. Addition of alum and/or sulfuric acid was also found to decrease ammonia losses by Ekinci et al. (2000) in their study of co-composting broiler litter with short paper fiber. DeLaune et al. (2004) reported that alum and phosphoric acid could reduce ammonia volatilization from large-scale poultry litter composting by as much as 76% and 54% respectively. Like zeolite, alum is relatively inexpensive and widely available. Koenig et al. (2005) suggested that low moisture and temperature could reduce ammonia release. However, managing temperature and moisture for this end would adversely impact microbial activity and other desired benefits of composting. They evaluated various chemical amendments on ammonia produced during layer manure composting in 1-litre vessels. With the exception of zeolite and cellulose, all amendments reduced ammonia capture; an alkaline pH (> 8) was thought to have enhanced ammonia volatilization and masked the adsorbent properties of the zeolite clinoptilolite.  In the past decade and until more recently, a number of commercial products and organic materials have been suggested to farmers for addition to manure to obtain beneficial effects  110  (Barrington et al. 2002; Heber et al., 2001). Claims of such beneficial effects include reduction in odor and ammonia losses, stimulation of bacterial activity, improvement in solids handling, increased manure decomposition, increase in composting rates and so forth. However, very little information about the test results has been reported in scientific publications, especially for the effects on odor emission during manure composting.  Starbuck and Wesley (1998) were among those who insisted at the time that commercial additives such as microorganisms, mineral nutrients, vitamins, enzymes or readily available forms of carbon would not be very effective when compared to good composting pile management. On the other hand, studies using biologically based additives such as selected microbial strains, bacterial-enzymatic preparations and plant extracts for the reduction and control of odor emissions from animal houses and stored manure slurry have continued (McCrory and Hobbs, 2001; Richard and Trautmann 1995). Illmer et al. (2007) studied the influence of a microbial compost starter kit, which consisted of an organic matrix and two seeded microorganisms, on the maturity and stability of the end-product from the lawn clippings composting process. Geotrichum klebahnii was isolated during their preliminary experiments with fresh, degrading material, whereas Trichoderma viridea was known as a potent cellulase-producer. Based on the results of ammonium-nitrogen, nitrate-nitrogen, dry matter, organic matter and pH measurements, the starter kit was considered to be an effective tool for composting preparation on a household scale.  Research studies have indicated the use of yeasts in the treatment of wastes to be technically feasible and may have economic significance (Arnold et al, 2000; Cristiani-Urbina et al. 2000). Some of the yeast strains can facilitate the growth of beneficial microbes by providing them with the energy and growth factors. Enzymes in yeast can proliferate the microorganisms that may have deodorization capability. Using a lab-scale composter for composting of food waste at 50oC, Choi and Park (1998) observed an early increase in the growth of yeast followed by the growth of thermophilic bacteria, albeit with a subsequent rapid decline in yeast population with the onset of thermophilic temperatures. Yeast might be working at the biochemical level, shifting microbial population. For instance, Tiquia et al. (2002) studied microbial population dynamics and enzyme activities during composting, and  111  found that the population of fungi and actinomycetes were most positively correlated with the activities of the Į–galactosidase and ȕ-glucosidase enzymes, which can readily degrade cellulose, a major component of plant and animal wastes. Kim et al. (2002) isolated and identified the yeasts Candida rugosa and Candida maris from soil and compost source. Both of the yeasts were found to be effective in reducing ammonium-nitrogen and biochemical oxygen demand of pig feces; the former was also highly effective in the deodorization of odorous volatile fatty acids (propionic acid, butyric acid and iso-valeric acid).  Yulipriyanto et al. (2002) used three different additives (lingo-cellulosic waste, microbial additive and Yucca juice) to assess effects of the additives on the nitrification-denitrification activities during composting of chicken manure piles. Their results indicated that the microbial additive had modified the composting conditions and microbial environment, thus conserving more nitrogen; lingo-cellulose waste had increased the carbon-to-nitrogen (C:N) ratio and hence less ammonia volatilization; and Yucca juice was effective in regulating the ammonia emission via promoting denitrification activities. Körner et al (2003) observed from lab-scale experiments that odor reducing additives might limit microbial degradation under aerobic conditions; if a higher concentration of the additive was used, oxygen consumption rate would be lowered and the period of microbial inhibition could be lengthened. Nevertheless, they also found microbial adaptation to the additive at a later part of the trial, resulting in effective biodegradation.  In summary, zeolite would have positive effects on the reduction of ammonia because of its highly adsorptive capability, and yeast may help to accelerate the degradation of slower degradable fractions of lignocellulosic ingredients in feedstock, and may have deodorization capability at the same time because the enzymes can proliferate the microorganisms involved in aerobic degradation. Therefore, the combination of yeast and zeolite may offer a potential solution for enhancing composting efficacy and reducing ammonia and odor emissions at the same time. Although the effectiveness of alum to ammonia and odor control in manure storage has been proven, little information is available on its effectiveness on composting odor control.  112  The objectives of this chapter are to determine the effectiveness of additives in the form of alum, zeolite and/or yeasts in reducing odor and ammonia emissions during the active phase of poultry manure composting and to assess how the use of these additives affects the composting process. The application rates or dosages that give rise to their effectiveness on ammonia and odor emission reductions were the major parameters being tested.  4.3  4.3.1  Materials and Methods  Experimental Set-up and Treatments  The substrate used in the laboratory-scale composting experiments was poultry manure in the form of broiler litter, which comprised mainly of manure and wood shavings. The composting recipe consists of 65% chicken manure, 25% sawdust and 10% hog fuel on dry weight basis. The broiler litter was collected from a commercial chicken farm in Abbotsford, British Columbia. Broiler litter had relatively low moisture content of 30%; moreover, it was stored in a cold room whereby temperature was controlled below 4°C to prevent degradation before the composting tests. It was allowed to reach ambient temperature after retrieving from storage the night before composting mix preparation. The characterization of the substrates and the initial physical and chemical properties of the composting mixture were measured with standard method (APHA, 1995) in the analytical laboratories of Chemical & Biological Engineering Department and the Earth & Ocean Science Department, the University of British Columbia. The results of analyzes are presented in Table 4.1 and Table 4.2. The initial moisture content of feedstock was adjusted to about 55% by adding tap water.  113  Table 4.1 Characterization of substrates Ingredients  Broiler litter  Sawdust  Hog fuel  moisture content (% wb) 31.0 10.7 VS content (% db) 82.8 94.3 carbon content (% db) 39.1 41.6 nitrogen content (% db) 4.38 1.42 C:N ratio 9 29 Each composition value is the average of triplicate samples VS: volatile solids; wb: wet basis; db: dry basis; ND: not determined.  40.5 ND 52.9 0.38 139  Table 4.2 Initial physical and chemical properties of the composting mixture Test Series Moisture Carbon Nitrogen VS content content content content (% db) (% wb) (% db) (% db) Set I Set II Set III  55.9 57.2 55.8  43.0 35.5 39.5  3.20 3.15 3.25  90.1 90.3 89.3  pH  8.47 7.87 4.10 - 6.09  Three sets of experiments were implemented for testing the application of the bioadditive during the active phase of poultry manure composting. Set I tests involved a powder-form product made up of 5% (w/w) yeasts, 85% zeolite and 10% wheat bran (developed and patented by CK Life Sciences International Inc., Hong Kong). These are non-genetically modified but specialized yeast strains that include primarily strains of Baker’s yeast Saccharomyces cerevisiae. The zeolite used is clinoptilolite, a natural zeolite ground to powder with particle sizes in the range of 0.063-0.5 mm, and having a bulk density of 700 kg/m3. The zeolite product contains 80% clinoptilolite and 20% clay. In Set II tests only the yeasts were used, which are the same species used in Set I. A third series of experiments (Set III) was performed for testing the application of alum. In each set of tests, three application rates or dosage levels were tested in addition to the control treatment without any bioadditive or alum. For Set I tests, the dosage levels were 1%, 5% and 10% (w/w) of the bioadditive “yeast and zeolite”, and they were run in replicates. For Set II tests, the dosage levels were 0.5%, 1% and 2% (w/w) of the bioadditive “yeast”. Application rates of alum were 1%, 2.5% and 5% (w/w) for Set III tests, also conducted in replicates.  114  Composting bioreactors were double-walled stainless steel Dewar flasks with vacuum in between, which gives thermos-like characteristics (Cole Parmer Instruments Company, Vernon Hills, IL, USA), with a working volume of 6 L. The composting vessel was positioned inside an insulated (adiabatic) box in order to simulate typical in-vessel composting process, or the core part of a compost pile. The laboratory-scale unit was equipped with continuous online temperature (copper-constantan thermocouples) and airflow rate monitoring system. The process control strategy used was the Rutger's temperature feedback method. Aeration was intermittent with 33% duty cycle below the setpoint, and continuous above the set point. The baseline of aeration rate was set to 0.36 L/min per kg dry mass or dry solids (DS), which represents a less ideal oxygen supply condition when compared to 0.72 L/min per kg DS. The temperature setpoint was 65°C to ensure that the regulatory requirements for pathogen destruction could be met. Temperature data were collected for the first 7-10 days during the active phase of composting, which is deemed to be finished when the temperature of the composting mixture has dropped back to ambient level.  4.3.2  Sampling and Analytical Measurement  At the beginning and end of composting, compost samples were taken to determine the moisture, carbon, nitrogen and volatile solids contents, as well as pH and free ammonia in the compost. Total composting mass was measured gravimetrically before and after composting. Moisture content was measured by gravimetric analysis and oven drying (at 101ºC) for 24 hours. The amount of volatile solids was measured by gravimetric analysis and ash content (ignition at 550ºC for 2 hours). Carbon and nitrogen contents were determined by using a Carlo Erba NA-1500 CN Analyzer at the Earth and Ocean Sciences Laboratory at UBC. Water-soluble NH4+-N in compost and feedstock was extracted DI water by adding 100ml DI water to 20 g of fresh sample, continuously stirring the suspension for 5 minutes, and shaking for another 30 minutes. The waste suspension was allowed to stand for about 15 minutes so that most of the suspended particles would settle out for pH and EC measurements using digital meters. The sample solution was centrifuged for 20 minutes and filtered twice using  115  0.45μm glass fiber filters before analysis. NH4+-N was determined colorimetrically using an automated ion analyzer (Lachat QuickChem, Zellweger Analytic, Inc.).  4.3.3  Ammonia and Odor Emission Measurements  Ammonia emitted from each reactor was captured by passing the exhaust gas through a condensate trap and an acid solution trap in series. The acid trap solution contains 4.8% sulphuric acid, which was replaced with fresh solution on each sampling day. Ammonium concentration in the acid solution was determined colorimetrically with the TKN method (but omitting the digestion procedure) using an automated ion analyzer (Lachat QuickChem, Zellweger Analytic, Inc.).  Electronic nose technology has been studied for odor measurements since the 1990s (Nicolas et al. 2000, Krzymien and Day 1997). Methods to apply electronic nose devices in a practical way still need further research and development, as electronic noses are not able to analyze complex odors such as those emanating from composting, landfill and wastewater treatment plants. For odor measurements, therefore, the olfactometry method remains a widely used and valid method for determining the total odor concentration, which may be interpreted as a composite indicator of odor strength (Zhang et al. 2002).  A dynamic dilution olfactometer, as described in Chapter 3, was used in odor measurement. The apparatus was designed, assembled, and calibrated in accordance with major provisions in the European Standard (CEN, 2003), while retaining features pertinent to the updated ASTM E679-04 Standard (ASTM, 2004). Odorous samples from the exhaust air were manually collected using Tedlar bags with a volume of 5L and fitted with one plastic valve (Safety Instruments Inc., Edmonton, AB). Tedlar bags were directly connected to odor sampling ports of the reactors to fill exhaust gas during aeration. In each sampling event, approximately 2-3 L of exhaust gas sample was collected in each bag for analysis. Odor sampling took place on day 0 (prior to the start of experiment), and then on days 1, 2, 4, 6 and at the end of each run. Odor measurements were carried out within 30 hours of sampling. Odor panelists were selected via panel screening procedure recommended by the EU  116  standard using n-butanol vapor at 40-80 ppbv to ensure a “normal” sense of smell. Each test sample was presented to an odor panel in ascending order of sample concentrations, so as to avoid olfactory fatigue. A series of dilution ratios with a factor of 2 was used in order to avoid olfactory fatigue (odor habituation and loss of sensitivity). Four panelists were used in the odor tests, which met the requirement of international standard. Each odor sample was presented twice to the panelists. The data set of each odor sample test includes eight individual scores from the four panelists.  4.3.4  Statistical Analysis  Statistical analyses were performed using the Analysis ToolPak that Microsoft Excel provides. The single factor analysis of variance (ANOVA) was used and the results were reported in terms of p-values that represent the degree of statistical significance for the thermal performance, peak odor concentration and cumulative odor emission of the treatments. Differences in results between treatments are significant at P < 0.05, unless stated.  4.4  4.4.1  Results and Discussion  Composting Process and Parameters  Temperature is a factor affecting the composting process, as well as a good indicator of its thermal performance. It shows how well the composting system is working and how far along the decomposition has progressed. As shown in Table 4.3 the duration of temperature exceeding 55oC was between 79-185 hours for all treatments, thus fulfilling the pathogen destruction criteria of at least 55oC for no less than three consecutive days for composting of manure or other organic matter (BCMOE, 2002)  Figures 4.1 and 4.2 present the temperature profiles of the compost treated with the bioadditives. The general temperature patterns of the treatments with yeasts alone, as well as yeast mixed with zeolite were similar to that of the control treatments. After the start-up with 117  aeration, the temperature of all treatments increased rapidly and reached thermophilic conditions (45oC or above) within 12 hours, and attaining peak temperatures approximately after 24 hours. As the readily available microbial food supply was consumed, temperatures began to decline and fell to ambient after 10 days of composting. In both sets of tests, the treatments with higher dosage of the bioadditives reached thermophilic temperatures sooner than the control treatment and the thermophilic phase was maintained for relatively a longer time period by more than three days, which could be advantageous for the decomposition of organic matter. These observations of thermal performance agree with those by Choi and Park (1998), who found that seeding with yeast stimulated the growth and increased the population of thermophilic bacteria in composting food waste.  As illustrated in Figure 4.3, the incorporation of alum in the composting feedstock altered the temperature evolution in the composting process and led to slow start, lower peak temperature, and longer thermophilic stage. The initial pH values were reduced to a range of 6.1 to 4.1 for alum application rates of 5%, 2.5 % and 1%, respectively (Table 4.2). The low pH might have shifted the microbial population to species that can adapt to the acidic environment, thus resulting in slow heat up and a gradual increase in temperature. When compared with the control treatment, the peak temperatures of the alum treatments were reduced by 2-8oC, while the thermophilic phase (above 45oC) was extended by 23-96 hours, depending on the application rate. The lower peak temperature and longer thermophilic phase may actually benefit the decomposition of organic matter. According to Stentiford (1996), temperatures between 45oC and 55oC may maximize biodegradation rate, while temperatures between 25oC and 45oC may maximize microbial diversity.  The addition of yeast and zeolite or yeast alone had little effect on the pH of compost. For the composting feedstock subjected to yeast and zeolite treatment, the initial pH (8.5) was alkaline. At the end of the experiment, the pH declined slightly with higher dosages, although the differences are insignificant (Table 4.4). This decreasing trend in pH is probably associated with the zeolite added. Venglovsky et al. (2005) also observed lower pH values in zeolite-amended compost versus the control for most of the thermophilic phase. With the  118  addition of alum at different dosages, the final pH values were within a narrower range of 8.5-8.7.  The final moisture contents of the treatments ranged from 52 to 63%, indicating that moisture was adequately maintained during the composting process. In comparison with the control, the treatment with 10% yeast combined with zeolite had higher total nitrogen content and smaller nitrogen losses, which are consistent with its lower ammonia emissions. For the treatments with yeasts applied alone, the greater losses in the mass of biodegradable matters contributed to the higher nitrogen contents with increasing yeast application rates.  The application of yeast with zeolite had marked effects on the free ammonia-nitrogen in compost, whereas the addition of yeast alone appeared to have minimal effect. In Set I tests, the concentration of water-soluble ammonia nitrogen decreased with higher application rate; this can be attributed to the zeolite, which has high adsorption capability for ammonia, thus binding the free ammonia and reducing the water extractable ammonia in compost.  Volatile solids (VS) content is a useful parameter for evaluating the extent of degradation during composting. As shown in Table 4.4, the VS content decreased with increasing dosages of the yeast and zeolite treatments. This observed trend could be attributed to two factors. Firstly, zeolite is a mineral. Unlike organic materials in which a portion of them will be lost as metabolic water and carbon dioxide, the inorganic minerals generally remain unchanged in the composting process. Hence, the addition of minerals increases the ash content and lowers the VS content. Secondly, because of the ability of the enzymes in yeasts that can proliferate the microbial population in the early stage of composting and enhance thermophilic conditions as previously discussed, the bioadditive would have positive effects on the decomposition of organic matter, thus lowering the volatile component.  Based on the change in mass of the volatile solids, the application of yeast combined with zeolite was found to be effective on biodegradation. At a dosage of 1%, the degree of degradation was virtually the same as the control treatment indicating the dosage was too low to have an influential effect on the decomposition process of organic wastes. However, as the  119  dosages increased to 5% and 10%, the degree of degradation was raised to 25.2 and 27% respectively, which is significantly higher than that of the control treatment (19%). In Set II tests, the final VS contents were virtually unaffected by the application of yeasts alone, and the degree of degradation was somewhat higher for the 1% yeast treatment (at 24%) versus the control (at 20%). For alum treatments, application rate of 5% resulted in a degree of degradation 3.6% higher than that of control; however, dosages below 5% showed neither positive nor negative effects on the degradation of the organic materials.  120  80  70 10% yeast and zeolite 5% yeast and zeolite 60  Temperature (°C)  1% yeast and zeolite control  50  Amb 40  30  20  10  0 0  24  48  72  96  120  144  168  192  216  Time (hours)  Figure 4.1 Temperature profiles of the treatment with yeast combined with zeolite  121  240  80  70  60  Temperature (°C)  50  40  30  20  2% Yeast  10  1% Yeast  0.5% Yeast  Control  Amb  0 0  24  48  72  96  120  144  168  Time (hours)  Figure 4.2 Temperature profiles of the treatments with yeast only  122  192  216  80  70  60  Temperature (°C)  50  40  30  20  5% alum  10  2.5% alum  1% alum  control  Amb  0 0  24  48  72  96  120  144  168  192  216  240  264  Time (hours)  Figure 4.3 Temperature profiles of the treatments with alum  123  288  312  336  Table 4.3 Thermal performances of additive treatments during active phase of composting Test Series Set I  Set II  Set III  Tp  tp  t45  t55  (oC)  (hr)  (hr)  (hr)  10% Yeast and Zeolite  71  23  141  108  5% Yeast and Zeolite  73  19  138  104  1% Yeast and Zeolite  72  22  130  108  Control  70  17  117  79  2% Yeast  71  23  168  127  1% Yeast  71  27  166  114  0.5% Yeast  72  18  134  108  Control  71  25  146  88  5% Alum  65  128  236  185  2.5% Alum  67  73  204  162  1% Alum  70  51  163  126  Treatment  Control 71 34 140 87 Values are the means of two replicates for Set I test and Set III test series db = dry basis Tp = peak temperature; tp = time elapsed to reach to peak temperature; t55 = cumulative hour when temperature above 55 oC; t45 = cumulative hour when temperature above 45 oC (i.e. thermophilic phase); A45 = areas bounded by temperature curve with temperature baselines of thermophilic phases.  124  Table 4.4 Changes in pH and nitrogen contents of additive treatments Test Series  Treatment  pH  Total N (%)  NH4-N (g/kg DS)  Set I  10% Yeast and Zeolite  8.78  3.32  3.32  5% Yeast and Zeolite  8.76  4.40  4.40  1% Yeast and Zeolite  8.74  4.42  4.42  Control  8.64  4.45  4.45  2% Yeast  8.79  5.18  5.18  1% Yeast  8.79  7.81  7.81  0.5% Yeast  8.74  6.54  6.54  Control  8.82  6.19  6.19  5% Alum  8.45  12.93  12.93  2.5% Alum  8.41  8.58  8.58  1% Alum  8.69  7.20  7.20  Control  8.85  5.90  5.90  Set II  Set III  Table 4.5 Changes in volatile solids and degree of degradation of additive treatments Test Series  Treatment  Volatile solid (%)  Degree of degradation (%)  Set I  10% Yeast and Zeolite  80.1  27.0  5% Yeast and Zeolite  83.9  25.2  1% Yeast and Zeolite  87.2  18.6  Control  88.1  18.9  2% Yeast  87.7  21.0  1% Yeast  87.5  23.8  0.5% Yeast  87.6  22.7  Control  87.7  20.3  5% Alum  84.5  24.2  2.5% Alum  86.3  20.5  1% Alum  87.7  21.0  Control  88.3  20.6  Set II  Set III  125  4.4.2  Ammonia Emission  Figures 4.4, 4.5 and 4.6 illustrate the variation of ammonia emission as a function of bioadditive and alum application rates during the active phase of composting. In all cases, the peak ammonia emission rates occurred on the second day of composting regardless of the dosages used, whereas the ammonia emission rates were much lower after nine days. For Set I tests (yeast and zeolite), the treatment with 1% dosage was probably too low to have any effects on the reduction in ammonia emission when compared with the control. This is in line with the results showing the degree of degradation as discussed previously. However, the effect of treatment at 5% dosage level is more obvious, while the effect at 10% dosage level is significant, as cumulative ammonia emission was seen to decrease by 50% from 2.69 g/kg DS to 1.35 g/kg DS (Figure 4.7). The addition of yeasts alone (Set II tests) with 2% dosage brought about a 14% reduction in cumulative ammonia emission (Figure 4.8), which could be resulted from the incorporation of some available nitrogen into the microbial biomass; however, similar effect was not evident with a lower dosage of 0.5%.  The application of alum greatly reduced the peak ammonia emission rate and resulted in a lag phase in peak ammonia emission (Figure 4.6). These effects can be explained by the fact that Alum produces H+ when it dissolves, which can react with NH3 to form ammonium. Ammonium can then react with the sulfate to form ammonium sulfate, which is a fertilizer. Also, the low initial pH due to the alum addition means less ammonia volatilization, as ammonia volatilization is highly dependent on pH. Over the active phase of composting, incorporating 1% alum into the feedstock reduced total ammonia emission by 45%, while a 5% dosage cut the total ammonia emission substantially by up to 90% (Figure 4.9). These results are in agreement with those reported in the literature í 76% reduction from DeLaune et al. (2004) and up to 99% reduction from Moore et al. (2000).  126  Ammonia Emission Rate (mg/h kg DS)  25  10% yeast and zeolite  20  5% yeast and zeolite 1% yeast and zeolite 15  control  10  5  0 0  24  48  72  96  120  144  168  192  216  240  Time (hour)  Figure 4.4 Ammonia emission rate over time as a function of application rate of yeast and zeolite 25 Ammonia Emission Rate (mg/h kg DS)  2% Yeast 1% Yeast 20 0.5% Yeast Control 15  10  5  0 0  24  48  72  96  120  144  168  192  216  240  Time (hr)  Figure 4.5 Ammonia emission rate over time as a function of yeast application rate  127  30  Emission Rate (mg/h kg DS)  25 5% alum 20  2.5% alum 1% alum  15  Conntrol  10  5  0 0  24  48  72  96  120  144  168  192  216  240  Time (hour)  Figure 4.6 Ammonia emission rate over time as a function of alum application rate  Cumulative Ammonia Emission (mg NH4-N/kg)  3,000 10% yeast and zeolite 2,500  5% yeast and zeolite 1% yeast and zeolite  2,000  control  1,500  1,000  500  0 0  24  48  72  96  120  144  168  192  216  240  Time (hours)  Figure 4.7 Cumulative ammonia emissions over time for treatments with yeast and zeolite (values are means of two replicates) 128  Cumulative Ammonia Emission (mg NH3-N/kg)  3,500 2% yeast 3,000  1% yeast 0.5% yeast  2,500  control 2,000 1,500 1,000 500 0 0  24  48  72  96  120  144  168  192  216  240  Time (hours)  Cumulative Ammonia Emission (mg NH3-N/kg  Figure 4.8 Cumulative ammonia emissions over time for treatments with yeast 3,500 5% alum 3,000  2.5% alum 1% alum  2,500  control 2,000 1,500 1,000 500 0 0  24  48  72  96  120  144  168  192  216  240  Time (hours)  Figure 4.9 Cumulative ammonia emissions over time for treatments with alum (values are means of two replicates)  129  264  4.4.3  Odor Emissions  Odor emissions were characterized via measured odor concentrations and odor emission rates. Figures 4.10 and 4.11 illustrate the variation of odor concentrations over time for Sets I and II tests. Regardless of the bioadditives used, odor concentrations increased to peak values within 24 hours after the start-up of composting, which were coincident with the high heat generation period. Thereafter, the odor concentrations gradually declined towards the end of active phase composting. For Set I tests, the higher dosage (10% w/w) results in lower odor concentrations. The percent odor reduction, in comparison with the control treatment, varied from 18% on day 2 to 40% on day 4, whereas the lower dosage treatment (1% w/w) led to 11% less odor level on day 2, but no further reduction in odor was observed thereafter.) This suggests that yeast combined with zeolite may have somewhat positive effects on reducing offensiveness of composting odors. Tables 4.6 and 4.7 present the results of replicates for peak odor concentration and cumulative odor emission. The ANOVA analysis indicated no significant differences between treatments for peak odor concentration (p > 0.25, as shown in Appendix III), though the differences were slightly more significant (p = 0.10) in terms of cumulative odor emission at the end of active phase of composting.  The differences among the odor concentration profiles of alum treatments were evident (Figure 4.12). The application of Alum resulted in a lag phase in peak odor occurrence, which is coincident with the peak temperature (Figure 4.3). Compared with the control, the odor concentrations from alum-treated compost with higher dosages were significantly lower (p = 0.05, see Appendix III). The lower odor concentrations observed would suggest less offensive odor generation. According to White et al. (1971), the source of odor is primarily due to microbial activities. Higher temperature indicates more microbial activities and greater potential for the formation and release of intermediate odorous compounds. While reduced sulfur gases including mercaptans and hydrogen sulfide are characteristics of anaerobic conditions during manure decomposition, it has been established that odorous gases such as dimethyl sulfide (DMS) and dimethyl disulfide (DMDS) could be produced even under aerobic composting conditions (Kuroda et al., 1996).  130  Table 4.6 Summary of the temperature parameters and odor emission for replicates of yeast and zeolite treatments Treatment 1 Treatment 2 Treatment 3 Control Replicate 1 Replicate 2 Replicate 1 Replicate 2 Replicate 1 Replicate 2 Replicate 1 Replicate 2 Thermal Performances Peak temperature (oC)  71  69  73  71  72  71  70  72  Time to reach peak temperature (hour) 23  27  19  31  22  24  17  20  Duration of temperature > 55 oC (hour) 108  74  104  105  108  96  79  77  Odor Emission Peak odor concentration (OU/m3)  7003  11585  7900  8192  7626  5793  8545  5793  Cumulative odor emission at t = 96 h (OU/kg)  6950  8972  7806  8815  8100  8941  8346  8275  Cumulative odor emission at end of active phase (OU/kg)  10084  13269  12447  14476  13431  11823  11672  10506  Treatment 1: 10% yeast and zeolite; Treatment 2: 5% yeast and zeolite; Treatment 3: 1% Yeast and zeolite.  131  Table 4.7 Summary of the temperature parameters and odor concentrations for replicates of alum treatment Treatment 1 Treatment 2 Treatment 3 Control Replicate 1 Replicate 2 Replicate 1 Replicate 2 Replicate 1 Replicate 2 Replicate 1 Replicate 2 Thermal Performances Peak temperature (oC)  65  63  67  64  70  68  71  70  Time to reach peak temperature (hour)  128  97  73  66  51  47  34  43  Duration of temperature > 55 oC (hour)  185  162  162  139  126  102  87  110  Peak odor concentration (OU/m3)  4820  4096  4344  4096  5792  4360  6992  8192  Cumulative odor emission at t = 96 h (OU/kg)  1635  1251  1755  1480  3822  2796  7524  5262  Cumulative odor emission at end of active phase (OU/kg)  5921  3867  5249  5642  10187  9680  11620  9427  Odor Emission  Treatment 1: 5% alum; Treatment 2: 2.5% alum; Treatment 3: 1% alum.  132  Odor concentration (OU/m 3)  10000 control 1% yeast and zeolite 5% yeast and zeolite 10% yeast and zeolite  7500  5000  2500  0 0  24  48  72  96  120  144  168  192  216  240  Time (hour)  Figure 4.10 Variation of odor concentration with the application rates of yeast and zeolite and standard deviations  Odor concentration (OU/m 3)  10000  control 0.5% yeast 1% yeast 2% yeast  7500  5000  2500  0 0  24  48  72  96  120  144  168  192  216  240  Time (hour)  Figure 4.11 Variation of odor concentration with the application rates of yeast and standard deviations  133  10000  Odor concentration (OU/m 3)  Control 1% alum  7500  2.5% alum 5% alum 5000  2500  0 0  24  48  72  96  120  144  168  192  216  240  Time (hour)  Figure 4.12 Variation of odor concentration with alum application rate and standard deviations  It would be useful to make a comparison between the strength of composting odor from this study and those reported in literature. Odors from municipal solid waste composting with windrows (5000-25000 OU/m3) and rotating drum (25000-50000 OU/m3) have been reported by Giggey et al. (1995) and Haug (1993). Noble et al (2001) measured odor concentrations ranging from 666 to 20139 OU/m3 for aerated materials, and from 538-395533 OU/m3 for unaerated materials, with odor samples collected from mushroom composting facility. Defoer et al. (2002) used a forced choice dynamic dilution olfactometer to analyze the effluents from composting plants with biofilters installed, which ranged from 390±60 to 13050±7628 OU/m3. Biasioli et al. (2004) monitored odor emissions from composting plants, and showed that odor concentration at the outlets of biofilters ranged from 780 to 5700 OU/m3. Even if the odor removal efficiencies of the biofilters were assumed to be only 50%, the concentrations of odors emitting from the composting plants could have been doubled. In fact, Sironi and Botta (2001) have shown that the composting odors at the biofilter inlets varied from 1630-32000 OU/m3. The reported values of odor concentrations are in the same order of magnitude as the findings from the lab-scale composting tests in this study.  134  Moreover, since the odor threshold value of DMS is quite low (Fraser and Lau, 2000), odor concentrations with values about 5000 OU/m3 are deemed reasonable for ideal condition. Odor concentration assessed via olfactometry represents the strength of the odor. It may not completely reflect the magnitude of odor emission, as odor emission is a dependent of airflow rate. Specifically, odor emission rate (OER) is a function of odor concentration and exhaust airflow rate, multiplied together; it is an essential parameter for dispersion modeling to predict odor concentration (level) downwind at the receptors (neighboring community).  In general, the highest odor emission rate occurred at a time coinciding with the peak temperature. This is a predictable situation because the Rutger’s strategy of temperature control induces greater aeration for cooling the compost, which directly raises the OER. Regardless of treatments, the cumulative or total odor emissions, as depicted in Figures 4.13 and 14, increased at a faster rate in the first four days, which accounted for 72-82% of the total odor emissions for both Sets I and II tests. The 10% yeast and zeolite application rate caused a decrease in odor emission by 15%, although a dosage of 1% had essentially no effect on the odor emission (Figure 4.13). While zeolite has a high adsorptive capability for ammonia, it is not known to be effective for removing organic odorants. The reduced odor emission might be explained by the higher degradation rate attained with yeast and zeolite added, whereby some odorous compounds could have been oxidized. In contrast, the degrees of degradation were similar for the various treatments when yeasts alone were applied. Hence, yeast alone appeared ineffective in reducing total odor emission (Figure 4.14).  Results show that alum addition could reduce odor emissions. As shown in Figure 4.15, a dosage of 1% had little effect on cumulative odor emission. However, the statistical analyses indicate that increasing dosage to 2.5% or higher did significantly reduce total odor emission (p < 0.05, as shown in Appendix III). The mechanism could be the change of the microenvironment of composting, with shifting of the microbial community to species that either generates fewer odors or has the capability to degrade odorous compounds. This is an area that needs further investigation. The effective alum dosage for odor reduction (by about 50%) was found to be 2.5%. It appears that dosage higher 2.5% would not reduce odor further.  135  Cumulative odor emission (OU/kg)  15000  12500  10000  7500 Control 5000  1% yeast and zeolite 5% yeast and zeolite  2500 10% yeast and 0 0  24  48  72  96  120  144  168  192  216  240  Time (hr)  Figure 4.13 Cumulative odor emissions from yeast and zeolite treated compost (values are means of two replicates)  Cumulative odor emission (OU/kg)  15000  12000  9000  6000 control 0.5% yeast 1% yeast 2% yeast  3000  0 0  24  48  72  96  120  144  168  192  216  Time (hr)  Figure 4.14 Cumulative odor emissions from yeast treated compost  136  240  15000 control 1% alum  Cumulative odor emission (OU/kg)  12500  2.5% alum 5% alum  10000  7500  5000  2500  0 0  24  48  72  96  120  144  168  192  216  240  Time (hours)  Figure 4.15 Cumulative odor emissions from alum treated compost (values are means of two replicates) Table 4.8 compares the additives tested in this study. All additives have advantages and disadvantages. The yeasts tested were effective in improving biodegradation of organic materials, but its effectiveness on ammonia and odor control appeared limited. Zeolite mixed with yeast showed positive effects on ammonia and odor reduction, but a high dosage (10% w/w on dry mass basis) is required, which means higher cost. Compared with yeast and zeolite, alum is the most effective additive for ammonia and odor control. One concern over alum application is the possibility of adverse effect on compost quality. As a matter of fact, alum is routinely used by the poultry industry. Over 700 million chickens were grown with alum in the US in 2006 alone (Moore and Edwards, 2007). The maximum dosage used in this study is well below the range of commercial broiler house practicing. A number of studies have shown that the addition of alum to poultry litter reduced phosphate and heavy metal, and resulted in greater crop yield, and had no negative effects on aluminum availability in  137  soil, runoff and uptake by plants (Moore et al., 1995, 1996; Line, 2002; Moore and Edwards, 2005).  Table 4.8 Comparison of composting additives tested in reducing ammonia and odor emissions Additive Effectiveness  Yeast and Zeolite x x  x  Advantages  x x x x  Disadvantages  x  Yeast  Reducing NH3 emission by up to 50%; Somewhat effective in reducing offensiveness of odors (peak odor concentration); Less effective in reducing odor in total (less than 15%).  x  Improving biodegradation; End product is good fertilizer; Natural product; Potential market.  x  Low efficiency in odor control.  x  x  x x  x  Costs  x x  High capital cost; Lower operation cost  x x  Alum  A little effects on ammonia emission; Not effective in total odor reduction.  x  Effective in reducing both NH3 emission (up to 90%) and odor emissions (up to 50%).  Improving biodegradation; Natural product; Potential market.  x  High nitrogen conservation; Good fertilizer.  The yeast strains tested appeared not effective in odor control; Difficult to handle due to small dosage required. High capital cost; Higher operation cost  x  Potential compost quality problem if over-dosed  x  Medium capital cost; Medium operation cost.  x  x  From the economic perspective, alum is a cost-effective additive in composting odor control; it has high market demand, thus it can be obtained with a reasonable price Zeolite is relatively cheap as its cost is primarily associated with transportation cost depending on its local availability. The costs of yeasts or yeast-based products are high at present. However, the economic viability of a product depends on the availability of markets for the product.  138  Biological based additives can become popular treatment approach for farmers; potentially there is a commercial market and hence lower price could result.  4.5  Conclusions  The study demonstrates that within the limits of application rates tested, the addition of bioadditives in the form of yeast alone or yeast in combination with zeolite had no adverse effects on the composting process. In both sets of tests, the treatments with higher dosage of the bioadditives reached thermophilic temperatures sooner than the control treatment and the high temperatures stayed for a longer time period by more than three days, which could be advantageous for the decomposition of organic matter.  All treatments had negligible effects on the pH of compost after the active phase of composting. The application of yeast and zeolite was found to be effective on biodegradation. At a higher dosage of 10%, the degree of degradation was significantly greater than that of the control treatment. When the yeasts were applied alone, the degree of degradation was somewhat greater for the 2% dosage level versus the control. Enzymes present in the yeasts may stimulate the growth of thermophilic microorganisms, which is beneficial to degradation; this phenomenon should be investigated in future studies.  In terms of reducing ammonia emission, alum appeared to be the most effective additive, cutting it by up to 90% at 5% (w/w) dosage. The combination of yeast and zeolite was also effective in reducing ammonia release by 50%, but an application rate of 5% (w/w) or higher on a dry mass basis is required. This is in line with the enhanced degree of degradation. The effectiveness of applying yeasts alone on ammonia emission reduction was relatively limited because its mechanism is possibly due to the incorporation of some available nitrogen into the microbial biomass.  The assessment of odor emission via olfactometry generates mixed results in response to the addition of yeast and zeolite, and to yeasts alone. Up to day 4 lower odor concentrations were observed with the higher dosage treatment (10% w/w) of yeast and zeolite, resulting in a  139  decrease in total odor emission by 15%. The reduced odor emission may be explained by the higher degradation rate attained with yeast and zeolite added, whereby some odorous compounds could have been oxidized. Because of the effectiveness on reducing ammonia and odor emissions, the application of the bioadditive could result in a smaller size of odor treatment system such as biofilter downstream of composting.  Alum addition exerted a strong influence on both ammonia and odor emissions. The incorporation of alum into feedstock reduced ammonia emission by 45% to as high as 90% depending on the dosage used, at the same time, odor emissions assessed via olfactometry could be reduced by 44% with dosages above 2.5% on dry mass basis. The study also demonstrates that nitrogen conservation can be greatly enhanced by alum addition. As a result of alum addition, water extractable ammonia in compost increased from 22% to as high as doubled compared with control. To be both effective and economic, a dosage of 2.5% is recommended.  By comparison, alum is most effective in reducing both ammonia and odor emissions, whereas the combination of yeast and zeolite has a great potential in enhancing degradation and reducing ammonia odor. The study demonstrates that within the limits of dosages tested in this study, the addition of yeast, yeast mixed zeolite, and alum have no adverse effects on the composting process.  140  4.6  References  Arnold, J.L., Knapp, J.S., Johnson, C.L. 2000. The use of yeast to reduce the polluting potential of silage effluent, Water Research, Vol. 34, No. 15, pp.3699-3708. ASTM. American Society for Testing Materials. 2004. Standard Practice for Determination of Odor and Taste Thresholds by a Forced-Choice Ascending Concentration Series Method of Limits. Annual Book of ASTM Standards. Designation E 679-04. Philadelphia. Barrington, S., D. Choiniere, M. Trigui, and W. Knight. 2002. Effect of carbon source on compost nitrogen and carbon losses, Bioresource Technol Vol. 83, pp.189 –194. B.C. Ministry of Environment. 2002. Organic Matter Recycling Regulation (OMRR). B.C. Reg. 18/2002. Biasioli, F., F. Gasperi, G. Odorizzi, E. Aprea, D. Mott, F. Marini, G. Autiero, G. Rotondo and T.D. Mark. 2004. PTR-MS monitoring of odor emissions from composting plants, International Journal of Mass Spectrometry Vol. 239, pp.103-109. California Integrated Waste Management Board (CIWMB). 2007. Comprehensive composting odor response project. Report prepared by San Diego State University. 171 pp. CEN. 2003. Air quality determination of odor concentration by dynamic olfactometry. prEN13725, CEN/TC 264/N 134. Committé Européen de Normalization, Brussels, Belgium. Choi, M.H. and Y.H. Park. .1998. The influence of yeast on thermophilic composting of food waste. Letters in Applied Microbiology, Vol. 26, pp.175-178. Chou, C.H. and F. Büyüksönmez. 2006. Biogenic emissions from green waste and comparison to the emissions resulting from composting: Part I: Ammonia, Compost Science & Utilization, Vol. 14, No. 1, pp. l6-22. Cristiani-Urbina, E. A.R. Netzahuatl-Munoz, F.J. Manriquez-Rojas, C. Juarez-Ramirez, N. Ruiz-Ordaz, Galindez-Mayer. 2000. Batch and fed-batch cultures for the treatment of whey with mixed yeast cultures, Process Biochemistry Vol. 35, No. 7, pp.649-657.  141  Defoer, N., I. de Bo, H. van Langenhove, J. Dewulf and T. van Elst. 2002. Gas chromatography-mass spectrometry as a tool for estimating odour concentrations of biofilter  effluents  at  aerobic  composting  and  rendering  plants,  Journal  of  Chromatography A, Vol. 970, pp. 259-273. DeLaune, P.B., Moor P.A., Daniel, T.C., Lemunyon, J.L. 2004. Effect of chemical and microbial amendments on ammonia volatilization from composting poultry litter, Journal of Environmental Quality, Vol. 33, pp.728-734. Ekinci, K., H.M. Keener and D.L. Elwell. 2000. Composting short paper fiber with broiler litter and additives, Part I: Effects of initial pH and carbon/nitrogen ratio on ammonia emission, Compost Science and Utilization, Vol. 8, No. 2, pp.160-172. Fraser, B.S. & A.K. Lau. 2000. Effect of process control strategies on composting rate and odor emission. Compost Science & Utilization, Vol. 8. No. 4, pp.274-292. Gage, J. 2003. Checklist for odor management at compost facilities, BioCycle, Vol. 44, No. 5, pp. 42-47. Giggey, M.D., J.R. Pinnette, and C.A. Dwinal. 1995. Odor control factors in compost site selection, BioCycle, Vol. 36, pp.74-79. Haug, R.T. (2004) Elements of Odor Management. In: Odor Management at Composting Facilities, J. G. Press, Emmaus, PA. Heber, A.J., J. Ni, A.L. Sutton, J.A. Patterson, K.J. Fakhoury, D.T. Kelly and P. Shao. 2001. Laboratory testing of commercial manure additives for swine odor control, Agricultural Air Quality Laboratory, Purdue University, West Lafayette, IN. Illmer, P., A.O. Wagner, J. Mair, C. Malin and S. Faxbmacher. 2007. Chemical and biochemical parameters during composting of lawn clippings with special regard to the efficiency of a compost starter kit, Compost Science and Utilization, Vol. 15, No. 1, pp.40-46. Kim, T.I., J.D. Han, Y.H. Yoo, T.Y. Thak, K.N. Kim, C.H.K. Kim and M.K.Kim. 2002. Studies on the deodorization of pig feces by yeasts. Proc. International Symposium on composting and compost utilization. May 6-8, Columbus, OH. pp.1274-1283. Kithome, M., J. W. Paul and A.A. Bomke. 1999. Reducing nitrogen losses during simulated composting of poultry manure using adsorbents or chemical amendments, Journal of Environmental Quality, Vol. 28, No. 1, pp.194-201.  142  Koenig, R.T., M.D. Palmer, F.D. Miner Jr., B.E. Miller, J.D. Harrison. 2005. Chemical amendments and process controls to reduce ammonia volatilization during in-house composting, Compost Science and Utilization Vol. 13, No. 2, pp.141-149. Körner, I., J. Braukmeier, J. Herrenklage, K. Leikam, M. Ritzkowski, M. Schlegelmilch and R. Stegmann. 2003. Investigation and optimization of composting processes – test systems and practical examples, Waste Management Vol. 23, No. 1, pp.17-26. Krzymien, M.E., and M. Day. 1997. Odours and Volatile Organics Emissions from a Commercial Composting Operation. In: Proceedings of the Air and Waste Management Association 90th Annual Meeting & Exhibition. June 8-13. Toronto, ON. Kuroda, K., T. Osada, M. Yonga, A. Kanematu, T. Nitta, S. Mouri, and T. Kojima. 1996. Emissions of malodorous compounds and greenhouse gases from composting swine feces, Bioresource Technology. Vol. 56, pp.265-271. Line, J.E. 2002. Campylobacier and Salmonella populations associated with chickens raised on acidified litter. Poult. Sci. 81:1473-1477. McCrory, D.F. and P.J. Hobbs. 2001. Additives to reduce ammonia and odor emissions from livestock wastes: A review, Journal of Environmental Quality, Vol. 30, pp.345-355. Miner, J.R., S.N. Raja, and W. McGregor. 1997. Finely ground zeolite as an odour control additive immediately prior to sprinkler application of liquid dairy manure. p. 717–720. In: Proc. of the Int. Symp. on Ammonia and Odour Emissions from Animal Production, Vinkeloord, the Netherlands. 6–10 Oct. 1997. NVTL, Rosmalen, the Netherlands. Moore, PA., Jr., T.C. Daniel, D.R. Edwards, and D.M. Miller. 1995. Effect of chemical amendments on ammonia volatilization from poultry litter. J. Environ. Qual. 24:293-300. Moore, PA., Jr., T.C. Daniel, D.R. Edwards, and D.M. Miller. 1996. Evaluation of chemical amendments to inhibit ammonia volatilization from poultry litter. Poult. Sci. 75:315-320. Moore, P.A., Jr., T.C. Daniel and D.R. Edwards. 2000. Reducing phosphorus runoff and inhibiting ammonia loss from poultry manure with aluminum sulfate. Journal of Environmental Quality, 29 (1): 37-49. Moore, PA., Jr., and D.R. Edwards. 2005. Long-term effects of poultry litter, alum-treated litter, and ammonium nitrate on aluminum availability in soils. J. Environ. Qual. 34:2104-2111.  143  Moore, PA., Jr., and D.R. Edwards. 2007. Long-term effects of poultry litter, alum-treated litter, and ammonium nitrate on phosphoroud availability in soils. J. Environ. Qual. 36:163-174. Nakaue, H.S., J.K. Koelliker, and M.L. Pierson. 1981. Studies with clinoptilolite in poultry: II. Effect of feeding broilers and the direct application of clinoptilolite zeolite on clean and reused broiler litter on broiler performance and house environment, Poultry Science, Vol.60, pp.1221–1228. Nicolas, J., A.C. Romain, V. Wiertz, J. Maternova, and P.H. Andre. 2000. Using the classification model of an electronic nose to assign unknown malodors to environmental sources and to monitor them continuously, Sensors and Actuators B. Vol. 69, pp.366-371. Noble, R., P.J. Hobbs, A. Dobrovin-Pennington, T.H. Misselbrook, and A. Mead. 2001. Olfactory response to mushroom composting emissions as a function of chemical concentration, Journal of Environmental Quality Vol. 30, pp.760 –767. Richard, T and N. Trautmann. (1995) The Science and Engineering of Composting. Available at: http://compost.css.cornell.edu/science.html. Schiffman, S.S., and C.M. Wiliams. 2005. Science of odor as a potential health issue, Journal of Environmental Quality Vol. 34, pp.129-138. Sironi, S., L. Capelli, P. Centola, R. Del Rosso and M. II Grande. 2007. Continuous monitoring of odours from a composting plant using electronic noses, Waste Management, Vol. 27, No. 3, pp. 389-397. Sironi, S. and D. Botta. 2001. Biofilter efficiency in odor abatement at composting plants, Compost Science and Utilization, Vol. 9, No. 2, pp.149-155. Starbuck, C. and T. Wesley. 1998. Compost additives do not add much, Horticulture Vol. 95, No. 2, p. 18. Statistics Canada, 2001. A geographical profile of manure production in Canada. http://www.statcan.ca/english/freepub/ Tiquia, S.M., J.H.C. Wan, and N.F.Y Tam. 2002. Microbial population dynamics and enzyme activities during composting, Compost Science & Utilization, Vol. 10, No. 2, pp.150–161. Venglovsky, J., N. Sasakova M. Vargova, Z. Pacajova, I. Placha, M. Petrovsky, D. Harichova. 1996. Evolution of temperature and chemical parameters during composting  144  of the pig slurry solid fraction amended with natural zeolite, Bioresource Technology, Vol. 96, pp.181–189. White, R. K., E. P. Taiganides and G. D. Cole, 1971. Chromatographic Identification of Malodors from Dairy animal Waste. 110-113. In: Livestock Waste Management and Pollution Abatement. Proc. International Symposium on Livestock Wastes. Yulipriyanto, H., P. R. Philippe-Morand, G. Tricot and C. Aubert. 2002. Effect of additives on the nitrification-denitrification activities during composting of chicken manure. In: Microbiology of Composting. (eds. Insam, H. et al.), Springer, New York. Zhang, Q., J.J.R. Feddes, I.K. Edeogu and X.J. Zhou. 2002. Correlation between odour intensity assessed by human assessors and odour concentration measured with olfactometers, Canadian Biosystems Engineering, Vol. 44, pp.6.27–6.32.  145  CHAPTER 5 EFFECTS OF OPERATIONAL PARAMETERS ON ODOR EMISSIONS  5.1  Introduction  Odor has been a persistent problem for composting industry. Significant efforts have been made in seeking effective solution to minimize the odor problem. While mitigation and control measures are important to odor control, better understanding and proper handling operation conditions are inarguably the first line of defense for composting odor control. Odor emissions from composting are affected by a number of variables such as biodegradability of feedstock, nutrient balance, oxygen (aeration), moisture, temperature, free air space (air-filled porosity), and pH. Where particular materials are composted, aeration, moisture, temperature and biodegradability become more important because they are not only influential to odor generation and release, but also controllable in composting operation.  It has been recognized that aeration, moisture, temperature, and biodegradable volatile solid have pronounced effects on composting odor emissions. Walker (1993) reported that increased aeration resulted in a decrease in concentration of odorous compounds, but an increase in total mass emission; conversely, decreased aeration resulted in increased concentration but decreased total emission rate of odorous compounds. Elwell et al. (2001; 2004) investigated the effects of aeration on ammonia and volatile fatty acid emissions. They demonstrated that a linear relationship existed between ammonia emission and the total airflows with a trend toward higher ammonia emissions with greater airflows. Their results also show that airflow reduction resulted in decreases in odorous VFA (acetic, propionic, butyric acid) emissions, though higher chain VFA (isobutyric, isovaleric, and valeric acid) emissions were increased. Similarly, Fraser and Lau (2000) found that mass emission rate of odorous compounds methyl mercaptan and dimethyl sulfide typically increased with higher aeration rate. Day et al. (1999) stated that the concentrations of some odorous compounds such as pinene and ethyl butyrate could be 10-fold higher when composting pile temperature increased from 20 to 65 oC. While it is possible to use GC to measure the concentration of A version of this chapter has been accepted for publication. Zhang, W., A.K. Lau, and Z.P. Wen (2008) Preventive Control of Odor Emission through Manipulation of Operational Parameters during the Active Phase of Composting. J. Environ. Sci. Health Part B - Pesticides, Food Contaminants and Agric. Wastes.  146  individual components in odor emissions, this chemical knowledge is of dubious value in assessing odor problems (Mills, 1995), because there are no reliable chemical indicators for odors caused by complex biological materials such as manure. In addition, odorous compounds are interactive, not additive, in their effect. That is, the combination of several odorous compounds may create a unique odor and not several odors perceived independently. Most composting studies with experimental treatments that involve operational parameters did not focus on odor emission, but rather on process dynamics, compost quality and ammonia emission. Hence, there is a need to quantify the relationship between composting odor emissions and operational parameters. Olfactometry remains to be the industry and regulatory standard, and provides the best possible method currently available for odor evaluation. The general objectives of this series of experiments are twofold:  (1) Examine the effects of key operational parameters on odor emissions and determine the operation conditions that are optimal for composting in terms of reduced odor emission; (2) Quantitatively determine correlations between peak odor emission rate and the operational parameters. The odor emission data generated from this series of experiment will also serve as the basis for developing predictive odor model for composting.  Specifically, aeration rate, moisture content, temperature setpoint, and biodegradable volatile solids are the primary parameters of interest. The dynamic changes of odor emissions during active phase of composting were quantified through olfactometric analysis. The peak odor emission rate data were then correlated with operational parameters. These correlation equations will be applied to the odor predictive base model to simulate various operation conditions. More details will be discussed in Chapter 7.  5.2  Materials and Methods  In order to evaluate the effects of operational parameters on odor emissions, the dynamic changes in odor emissions during the active phase of composting were examined by varying some key operational parameters under controlled conditions. Four variables, including  147  aeration rate, initial moisture content, temperature setpoint, and biodegradable volatile solids (BVS) that are deemed important to odor generation and release were chosen in this series of experimental study. These factors were assumed to be independent of each other. The experimental design involves each factor at four different levels. The levels of airflow rate tested correspond to the low, average and high aeration rates used in commercial composting facilities. For temperature setpoint, the lowest setpoint of 55ºC is the temperature required for pathogen destruction (BCMOE, 2002). The highest setpoint of 70ºC corresponds to the limit of temperature above which many forms of microbes may not be able to survive. The levels of initial moisture content and biodegradable volatile solids are within the general range of low and high limits found in industrial composting practice. A partial factorial experimental design was adopted by varying one variable while holding others constant, such that a total of 16 treatments were tested (Table 5.1). Tests concerning two important parameters, aeration and BVS, were performed with replicates.  The experiments were conducted with a lab-scale bioreactor system. The experimental setup and its process monitoring and control system are described in detail in Chapter 3. A recipe common to all tests comprised of 65% of chicken manure, 25% of sawdust and 10% hog fuel. The feedstocks were characterized with standard methods (APHA, 1995) in the analytical laboratories of the Chemical & Biological Engineering Department and the Earth & Ocean Science Department, the University of British Columbia. Table 5.2 shows the initial physical and chemical properties of the feedstocks.  The experimental procedure, analytical measurements and olfactometry tests are similar to those reported in Chapters 4 and 5, and briefly described here. Aeration was intermittent with 33% duty cycle below the setpoint, and continuous above the set point. Air flowrate and temperature were continuously monitored with mass flowmeters and thermocouples. Biodegradable volatile solids content (BVS) of the feedstock was determined indirectly via measuring the volatile solids content (VS), and applying a known degradability coefficient of each ingredient, which was 75.6% for chicken manure and 42% for sawdust (Haug, 1993).  148  Table 5.1 Experimental design for the effects of operational conditions on odor emission Treatment  Aeration rate (L/min·kg DS)  P1 P2 P3 P4 P5 P6 P7 P8 P9 P10 P11 P12 P13 P14 P15 P16  0.36 0.54 0.72 1.08 0.72 0.72 0.72 0.72 0.72 0.72 0.72 0.72 0.72 0.72 0.72 0.72  Moisture content (% wb) 55 55 55 55 45 55 65 75 55 55 55 55 55 55 55 55  Temperature setpoint (oC) 65 65 65 65 65 65 65 65 55 60 65 70 65 65 65 65  Biodegradable volatile solids (% DS) 55 55 55 55 55 55 55 55 55 55 55 55 45 50 55 65  DS: dry solids; wb: wet basis. Odor samples were collected with Tedlar bags on days 0, 1, 2, 4, 6, and at the end of each run. Odor measurements were carried out within 30 hours of sampling with a triangular forced-choice dynamic dilution olfactometer that complies with EU standard. Details of the instrument have been presented in Chapter 3. Four panelists were used in the odor tests, which met the requirement of international standard. Each odor sample was presented twice to the panelists. The data set of each odor sample test includes eight individual scores from the four panelists.  Statistical analyses were performed using the Analysis ToolPak that Microsoft Excel provides. The single factor analysis of variance (ANOVA) was used and the results were reported in terms of p-values that represent the degree of statistical significance for the 149  thermal performance, peak odor concentration and cumulative odor emission of the treatments. Differences in results between treatments are significant at P < 0.05, unless stated.  The standard deviations of cumulative odor emissions were calculated with following method to show the statistical dispersion (how widely spread the values in a data set are):  If the mean or expected value of a parameter X is the sum of its component values in a series, i.e. E(X) = μo + μ1 + μ2 +……+ μN Then the variance will be given by: V(X) = Vo2 + V12 + V22 +……+ VN2 Hence, standard deviation: SD  V( X )  Table 5.2 Initial physical and chemical properties of the composting mixture Treatment Moisture  Carbon  Nitrogen  VS[b]  Free NH3  (% db[a])  (% db)  (% db)  (% db)  (NH3 g/kg)  P1-P4  55.8  43.4  3.19  90.3  2.58  7.70  0.22  P5  43.1  43.9  3.30  90.7  2.76  7.24  0.24  P6  53.9  43.9  3.30  90.5  2.76  7.24  0.24  P7  63.6  43.9  3.30  90.3  2.76  7.24  0.24  P8  73.1  43.9  3.30  90.7  2.76  7.24  0.24  P9-12  55.3  36.8  2.97  90.4  4.18  7.65  0.28  P13  62.9  42.8  1.71  94.7  2.64  7.61  0.10  P14  56.0  41.6  2.04  93.1  2.54  7.69  0.14  P15  54.7  39.4  2.97  90.5  3.72  7.70  0.21  P16 56.9 37.8 4.50 86.6 4.33 7.68 Each composition value is the average of triplicate samples [a] db: dry basis; [b] VS: volatile solids; [c] EC: electrical conductivity.  0.24  150  pH  EC[c] (dS/m)  5.3  5.3.1  Results and Discussions  Effects of Aeration Rate  Odor concentration quantitatively reflects the odorants in unit air eliciting a physiological response from odor panelists (CEN, 2003). It represents how strong the odor is. Consequently, odor concentration is an important parameter for on-site composting odor evaluation with regard to the extent of anaerobic conditions that might be present in composting process despite the use of aeration. The changes of odor concentrations over time under different aeration regime are presented in Table 5.3.  During the active phase, the odor concentrations were significant only in the first four days of composting and peaked on day 2 regardless of treatment. The highest odor concentration was 11585 OU/m3 generated by the treatment with airflow rate of 0.36 L/min·kg DS, which was probably too low to supply sufficient oxygen and led to anaerobic pockets in the composting matrix. The highest aeration rate of 1.08 L/min·kg DS was assumed to provide more oxygen supply to reduce odor concentration; however, such high aeration rate could also free up and strip the more odorous compounds from the composting matrix before they were further decomposed in place (CIWMS, 2007). It is interesting to note that the treatments receiving medium aeration rates of 0.54 and 0.72 L/min·kg DS are associated with peak odor concentration of 2896 OU/m3, which was much lower than that of the treatment with 0.36 L/min·kg DS aeration rate. The results reaffirm that aeration rate has profound influence on odor generation. Table 5.4 summarizes the odor concentrations of the replicates of treatments with different aeration rates. The results of ANOVA analysis suggested that the various treatments had significantly different peak odor concentrations (p < 0.005, as shown in Appendix III).  When evaluating the extent of off-site odor impact on surrounding communities, odor emission rate (OER) becomes more meaningful, because odor concentration at receptor may be estimated from OER using odor dispersion modeling. The odor emission rate is calculated with the product of measured odor concentration and exhaust airflow rate. As shown in Table  151  5.3, the general trends of odor emission rate are similar to those of odor concentration. There are significant differences in peak OER among treatments. The highest peak OER observed is from treatment with aeration rate of 1.08 L/min·kg because it had high airflow rate and produced greater odor concentration. The peak OER of the treatment with aeration rate of 0.36 L/min·kg was also high due to the fact that offensive odor was generated (11585 OU/m3). Treatments with airflow rates of 0.54 and 0.72 L/min·kg resulted in significant lower peak OER. After four days of composting, however, the differences in OER between treatments were insignificant.  Table 5.3 Odor emission rates of treatments with different aeration regimes Treatment  Parameters OC (OU/m3)  Elapsed time (hr) 48 96  0  216  304  11585  1024  362  0.36 AFR (L/min) L/min.kg DS OER (OU/hr)  0.24  0.36  0.25  0.12  4.45  250.24  18.46  4.15  OC (OU/m3)  304  2896  609  609  0.54 AFR (L/min) L/min.kg DS OER (OU/hr)  0.24  0.54  0.32  0.18  4.45  93.84  21.37  7.48  OC (OU/m3)  304  2896  362  431  0.72 AFR (L/min) L/min.kg DS OER (OU/hr)  0.24  0.72  0.38  0.24  4.45  156.40  8.09  8.68  OC (OU/m3)  304  5793  1448  256  1.08 AFR (L/min) 0.24 1.08 0.36 L/min.kg DS OER (OU/hr) 4.45 375.36 33.18 Values are the means of two replicates involved in the aeration test series OC: Odor concentration; AFR: Airflow rate; OER: Odor emission rate.  152  0.36 5.86  Figure 5.1 shows the cumulative mass of odor emissions. Since odor measurement was discrete, the cumulative odor was calculated by the product of odor emission rates (OER) and their corresponding time intervals. The value of odor emission rate calculated is a geometric average of two consecutive measurements in the same time span. As illustrated in Figure 5.1, the highest airflow rate yielded greatest odor emissions in total, suggesting a mechanism that odor release from compost matrix may primarily be a mechanical conveyance by air movement. This is in agreement with the conclusion by Elwell et al. (2004) who stated that aeration might remove odorous compounds from compost mass before they have the opportunity to be further decomposed in place. On the other hand, total odor emission of treatment with the lowest airflow rate was also high, apparently due to anaerobic odor generation as indicated by the high odor concentration observed from this treatment.  Practically, maximum odor emission rate (OER) is of more importance as it can be used for odor dispersion modeling for planning and design of composting plant. In this study, the correlation between maximum OER and environmental factors will serve as the basis for the development of odor predictive model in Chapter 7. To obviate influences of the data obtained from different runs, the observed peak OER data were normalized by dividing their values with the peak OER at airflow rate of 0.72 L/min·kg, which is then termed “relative OER”. The relative OER data were then used to correlate with aeration rate. The relationship was found to fit well with a second order polynomial equation (R2 = 0.94), indicating that the relative OER was highly correlated to airflow rate. The correlation equation thus obtained is:  fm  8.98Q 2  11.64Q  4.53  (6.1)  where fm is the relative peak odor emission rate (dimensionless), and Q is the aeration rate for forced aeration system (L/min·kg dry mass).  Treatments with both the lowest and highest aeration rates resulted in higher peak OER, implying that an optimal aeration rate could exist in terms of minimizing odor emissions. Through taking the derivative of Equation 6.1, the optimal aeration rate for minimizing odor emission from poultry manure composting was determined to be 0.63 L/min·kg dry mass.  153  Table 5.4 Summary of the temperature parameters and odor emission of the replicates of treatments with different aeration rates. Treatment 1 Treatment 2 Treatment 3 Treatment 4 Replicate 1 Replicate 2 Replicate 1 Replicate 2 Replicate 1 Replicate 2 Replicate 1 Replicate 2 Thermal Performances Peak temperature (oC)  73  69  75  73  73  72  72  73  Time to reach peak temperature (hour)  64  51  40  32  23  25  21  22  Duration of temperature > 55 oC (hour)  132  111  159  137  103  119  85  117  Peak odor concentration (OU/m3)  11585  11585  2896  2896  2896  5793  5793  5793  Cumulative odor emission at t = 96 h (OU/kg)  5919  6589  3999  3524  2784  7525  6336  13048  Cumulative odor emission at end of active phase (OU/kg)  6166  7363  4906  4188  3812  8631  8920  14627  Odor Emission  Treatment 1: 0.36 L/min.kg DS; Treatment 2: 0.54 L/min.kg DS; Treatment 3: 0.72 L/min.kg DS; Treatment 4: 1.08 L/min.kg DS.  154  Overall, airflow rate of 0.72 L/min·kg dry mass performed better in reducing odor emission, whereas airflow rate of 0.54 L/min·kg dry mass appears to be a better choice in terms of power consumption without compromising odor emission. In terms of degradation, both aeration rates are acceptable. Therefore, aeration rate between 0.54 and 0.72 L/min·kg dry mass is recommended.  Cumulative odor emission (OUe/kg)  12000  0.36 L/min.kg  10000  0.54 L/min.kg 0.72 L/min.kg 1.08 L/min.kg  8000  6000  4000  2000  0 0  24  48  72  96  120  144  168  192  216  Time (hr)  Figure 5.1 Cumulative odor emissions over time as affected by aeration rate  155  5.3.2  Effects of Initial Moisture Content  Moisture is known to exert profound effect on composting odor generation (Haug, 1993; Miller, 1993; Cornell Composting, 1996). Initial moisture content may be readily adjusted to optimize moisture condition during feedstock preparation (Richard et al., 2002). In practice, however, it may not be possible to start composting at optimum moisture level due to difficulties to obtain adequate amendments in time and the requirement to accurately measure the moisture content of each ingredient, which is usually well mixed before loading into the composter. From this perspective, quantitatively determining the correlation between odor emissions and initial moisture contents is of practical importance for odor control.  Table 5.5 shows the changes in odor concentration during the course of high rate composting with initial moisture contents. The cumulative odor emissions from different treatments are shown in Figure 5.2. The highest odor concentration was 5793 OU/m3 recorded with the highest initial moisture content of 75%, because high moisture occupies more void space, thereby reducing free air space and increasing the possibility of forming anaerobic pockets. Treatment with moisture level of 45% had lowest odor concentration. However, moisture at such a low level resulted in less degradation probably because microbial growth was somewhat limited by water availability. Similar to the odor concentration observed, the highest odor emission rate was generated by the treatment with initial moisture of 75%, followed by the 65% moisture treatment, as shown Table 5.5.  The results reveal that cumulative odor emissions in the first four days of composting account for approximately 80-90% of odor emissions in total during the active phase of composting, suggesting that for composting system with forced aeration, the first few days constitute a critical period for odor control. As shown in Figure 5.2, the total odor emission of the treatment with initial moisture of 75% was very high; nevertheless, there were no significant differences among the treatments with initial moisture below 65%.  156  Table 5.5 Odor emission rates of treatments with different initial moisture contents Treatment  45%  55%  65%  75%  Elapsed time (hr) 46 94  166  255  2896  2896  2048  512  512  0.24  0.72  0.55  0.24  0.24  OER (OU/hr)  3.69  125.12  95.58  29.49  0.24 7.37  7.37  OC (OU/m3)  256  4096  2896  2048  362  256  AFR (L/min)  0.24  0.26  0.72  0.24  0.24  OER (OU/hr)  3.69  63.90  125.12  29.49  0.24 5.21  3.69  OC (OU/m3)  256  2896  4096  724  724  181  AFR (L/min)  0.24  0.72  0.67  0.24  0.24  OER (OU/hr)  3.69  125.12  164.66  10.43  0.24 10.43  2.61  OC (OU/m3)  256  5793  4096  1448  1448  362  AFR (L/min)  0.24  0.72  0.72  0.24  0.24  OER (OU/hr)  3.69  250.24  176.95  20.85  0.24 20.85  Parameters  0  22  OC (OU/m3)  256  AFR (L/min)  5.21  OC: Odor concentration; AFR: Airflow rate; OER: Odor emission rate  The values of peak OER were normalized at a moisture content of 55% and correlated to moisture. There was a very close correlation between the relative peak OER and initial moisture content (R2 §1). The best-fitted equation is:  fm  17.1T 2  16.52T  4.944  (6.2)  where T is initial moisture content on wet basis (%).  The correlation indicates that a moisture level of 55%, which is generally recognized as optimal for composting, also generates less odor emission. When initial moisture is above 55%, odor emissions increase with moisture as a second-order function.  157  It is generally believed that moisture above 60% on wet basis would generate anaerobic conditions (Day and Shawn; 2001; CIWMB, 2007). The results of this study suggest that for poultry manure composting with forced aeration, initial moisture content could be up to 65% without generating significant odors. It should be noted that the results obtained from this series of test were based on industrial standard aeration rate of 0.72 L/min per kg DS, and they may not be applicable to composting systems with aeration below the industrial standard.  14000  Cumulative odor emission (OUe/kg)  12000  10000  8000  6000 45% 55%  4000  65% 75% 2000  0 0  24  48  72  96  120  144  168  192  216  240  264  Time (hr)  Figure 5.2 Cumulative odor emissions over time as function of initial moisture content  5.3.3  Effects of Temperature Setpoint  There are some contradictory reports in the literature regarding the effects of temperature on odor emissions. Wilber and Murray (1990) cited by Epstein (1997) demonstrated that odor emissions decreased with increased temperature at a biosolids composting facility. In contrast, Day et al. (1999) stated that the concentrations of some odorous compounds in exit gas increased when composting pile temperature increased from 20 to 65oC. Miller (1993) 158  also reported that up to a limit of 60oC, decreasing temperature reduced odors. A study conducted by Bruce (1998) showed that odor emission rates were insensitive to temperature setpoint.  The results of this study partially support the conclusions of both Miller (1993) and Epstein (1997) under certain circumstances. Table 5.6 shows the odor concentration and OER observed at different temperature setpoint. The peak odor concentration recorded at temperature setpoint of 55oC was 1448 OU/m3, whereas peak odor concentration recorded at temperature setpoint of 60 oC was 5793 OU/m3, four times higher than at 55oC. Hence, up to a limit of 60oC, odor concentration increased with higher temperature setpoint. This is in good agreement with the results reported by Miller (1993). On the other hand, temperature setpoint above 60oC led to a decrease in odor concentrations, which partially supports Epstein’s conclusions. Peak odor emission rate was similar to the odor concentration observed. Although peak odor emission rate of treatment with temperature setpoint of 55oC was significantly lower, it may not be viable for composting because the pathogen destruction requirement could not be met at temperature setpoint of 55oC. It was found that the peak odor emission rate did not correlate very well with temperature setpoint. As illustrated in Figure 5.3, increasing temperature setpoint from 55 to 60oC doubled total odor emission, whereas, increasing temperature setpoint from 60 to 65oC reduced odor emission by 14%. Further increase of the temperature setpoint to 70oC reduced odor emission by 46%. This phenomenon could be explained by the fact that low temperature setpoint requires additional aeration for cooling; hence, it tends to reduce the probability for anaerobic odor generation. On the other hand, if temperature goes too high beyond 65oC, the microbial organisms responsible for odor generation may stop growing or start to die. This is particularly evident for temperature setpoint at 70oC. As seen in Figure 5.3, at a temperature setpoint of 70oC, the odor concentration dropped sharply following the pattern of its temperature profile (Appendix II). This study reveals that the temperature setpoint of 60oC appeared to be a turning point for odor emission. Below this point, odor emission increased with higher temperature setpoint, whereas above this point, odor emission decreased with  159  increasing temperature setpoint. Indeed there are reports in literature, which regarded temperature at 60 oC as a critical point for microorganism growth (Kaiser, 1996).  Table 5.6 Odor emission rates of treatments at different temperature setpoints  0  23  Elapsed time (hr) 47  OC (OU/m )  256  1448  1448  724  128  AFR (L/min)  0.24  0.72  0.54  0.24  0.24  OER (OU/hr)  3.69  62.6  46.9  10.4  1.84  OC (OU/m3)  256  2896  5793  724  256  AFR (L/min)  0.24  0.72  0.41  0.24  0.24  OER (OU/hr)  3.69  125.1  142.5  10.4  3.69  OC (OU/m3)  256  2896  2896  1024  128  AFR (L/min)  0.24  0.72  0.52  0.24  0.24  OER (OU/hr)  3.69  125.1  90.4  14.8  1.84  OC (OU/m3)  256  2896  1024  724  256  AFR (L/min)  0.24  0.72  0.41  0.24  0.24  OER (OU/hr)  3.69  125.1  25.2  10.4  3.69  Temperature Parameters setpoint 3  55oC  o  60 C  o  65 C  70oC  95  216  OC: Odor concentration; AFR: Average airflow rate; OER: Odor emission rate.  160  8000  Cumulative odor emission (OUe/kg)  7000  6000  5000  4000  3000  2000 55 C  60 C  65 C  70 C  96  120  144  168  1000  0 0  24  48  72  192  216  Time (hr)  Figure 5.3 Cumulative odor emissions over time as affected by temperature setpoint  5.3.4  Effects of Biodegradable Volatile Solids Content  Odor emission mainly depends on the biodegradable fraction of feedstock rather than the inert fraction. Indeed many researchers have recognized that the decomposition of organic wastes may be better understood by being distinguished as readily degradable, slow degradable and non-biodegradable fractions (Haug, 1993; Kaiser, 1996; Richard, 1997). In this study, the biodegradable volatile solids (BVS), which represents biodegradable fraction of volatile solids, was investigated in relation to odor emission. As shown in Table 5.7, the peak odor concentrations of the treatments ranged from 2896 OU/m3 to 6889 OU/m3, with percent standard error of less than 35%. Higher odor concentrations observed in the composting process were consistent with higher BVS. This is especially true for the treatment with the highest BVS of 65%, which had high odor concentration even at the end of active phase of composting.  161  Table 5.7 Odor emission rates of treatments with different biodegradable volatile solids BVS  Parameters 3  Elapsed time (hr) 24 48 96  0  240  45%  OC (OU/m ) AFR (L/min) OER (OU/hr)  65 0.24 1.47  2896 0.24 41.71  1722 0.24 34.70  215 0.24 3.55  215 0.24 3.55  50%  OC (OU/m3) AFR (L/min) OER (OU/hr)  90 0.24 1.62  3444 0.72 151.03  2435 0.35 52.6  1024 0.24 18.43  304 0.24 4.45  55%  OC (OU/m3) AFR (L/min) OER (OU/hr)  181 0.24 2.61  2048 0.72 93.84  4871 0.72 213.59  861 0.24 12.59  554 0.24 8.27  OC (OU/m3) 362 6889 5793 1281 AFR (L/min) 0.24 0.72 0.72 0.32 65% OER (OU/hr) 5.21 338.71 250.24 22.71 Values are the means of two replicates involved in the BVS test series OC: Odor concentration; AFR: Airflow rate; OER: Odor emission rate  1108 0.24 23.91  Figure 5.4 shows the cumulative odor emissions of treatments with different levels of BVS. The total odor emissions were 2140, 4110, 7423, and 13766 OU for treatments with 45, 50, 55, and 65% BVS, respectively. The results demonstrated that total odor emission is highly dependent on BVS. This could be explained by a few factors. Firstly, with high BVS, aerobic intermediate odorous compounds could accumulate quickly due to fast decomposition and become released into the environment with forced air stream. Secondly, high BVS feedstock decomposes rapidly such that oxygen demand may outpace its supply and increase the potential to produce anaerobic odors. However, this does not mean that anaerobic condition would necessarily prevail due to greater oxygen consumption, as the odor concentrations observed were not too high compared to levels associated with industrial composting activities. Thirdly, the high BVS feedstock might cause a reduction in porosity due to smaller  162  particles that result from fast decomposition; this would negatively affect aeration, and contribute to odor generation.  The peak odor emission rates were normalized by BVS at a reference condition (i.e. 50% BVS) and correlated to BVS itself. The best fit was found to be a second order polynomial equation with R2 = 0.984. The regression equation is:  fm  53.721BVS 2  64.51BVS  17.82  (6.3)  18000  Cumulative odor emission (OUe/kg)  16000 45% 50% 55% 65%  14000 12000 10000 8000 6000 4000 2000 0 0  24  48  72  96  120  144  168  192  216  240  264  Time (hr)  Figure 5.4 Cumulative odor emissions over time as affected by biodegradable volatile solids Table 5.8 summarizes the temperature parameters and odor emission of the replicates of treatments with different biodegradable volatile solids. Statistical analysis is presented in Appendix III. According to the ANOVA analysis performed on the various performance indicators, significant differences were found among the BVS treatments in terms of thermal performance (peak temperature; time to reach peak temperature and duration of T > 55oC) with p < 0.05, and for cumulative odor emission at the end of the experiment with p < 0.005.  163  To examine the experimental repeatability of the tests involving the various key operational parameters, a statistical analysis was performed against the cumulative odor emissions of all control treatments used in these tests, which are supposed to produce similar results because they were tested under the same composting conditions (i.e. aeration rate of 0.72 L/min.kg dry mass, initial moisture content of 55%, temperature setpoint of 65oC and biodegradable volatile solids content of 55%) The data analyzed include cumulative odor emissions ȈOE, in [OU/kg], during the first 4 days (96 hrs) and at the end of the active phase of composting (240 hrs). Table 5.9 shows the standard deviations and the coefficients of variation, CV. In view of the nature of odor measurements, the experiments are considered to possess good reproducibility.  164  Table 5.8 Summary of the temperature parameters and odor emission of the replicates of treatments with different biodegradable volatile solids Treatment 1 Treatment 2 Treatment 3 Treatment 4 Replicate 1 Replicate 2 Replicate 1 Replicate 2 Replicate 1 Replicate 2 Replicate 1 Replicate 2 Thermal Performances Peak temperature (oC)  72  69  70  73  70  72  63  66  Time to reach peak temperature (hour) 40  44  25  24  25  24  26  32  Duration of temperature > 55 oC (hour) 125  160  100  109  77  79  39  54  Odor Emission Peak odor concentration (OU/m3)  5793  11585  5793  4096  4096  2896  4096  2896  Cumulative odor emission at t = 96 h (OU/kg)  9247  13484  5761  6456  3197  4618  1780  1105  Cumulative odor emission at end of active phase (OU/kg)  13493  15211  7263  7924  3948  6751  2045  1855  Treatment 1: 65%; Treatment 2: 55%; Treatment 3: 50%; Treatment 4: 45%.  165  Table 5.9 Statistical analysis of experimental data from the control treatments Ȉ OE96  Ȉ OE240  OU/kg  OU/kg  2784 6213  P6 P11  Treatment  P3  P15  Replicates*  R1 R2  R1 R2  Log(Ȉ OE96)  Log(Ȉ OE240)  3547 7323  3.445 3.793  3.550 3.865  5399  6682  3.732  3.825  4781  5412  3.680  3.733  5953 6456  7423 7647  3.775 3.810  3.871 3.883  Mean Std Dev  5264 6339 3.721 1356 1589 0.136 CV 25.8% 25.1% 3.7% * R1, R2: replicates within the test series involving aeration and BVS  166  3.802 0.129 3.4%  5.4  Conclusions  It is possible to devise and implement the best management practices using preventive odor control measures, knowing the odor emissions and the associated operating conditions. This study quantitatively examined the relationship between odor emission and key operational parameters by independently changing aeration rate, initial moisture content, temperature setpoint and biodegradable volatile solids content in the test series. The main conclusions are as follows:  (1)  Aeration rate had strong influence on odor emissions. Low aeration rate could lead to anaerobic pockets within the compost matrix, thus resulting in higher odor concentrations. However, high aeration rate might also result in higher odor emission rate as well as total odor emission. An optimal aeration rate was approximately 0.6 L/min.kg dry mass for intermittent aeration and a duty cycle of 33%.  (2)  Odor emissions were greatly affected by initial moisture content of feedstock. Both peak odor concentration and emission rate generally increased with initial moisture. However, odor emission appeared significant only at moisture levels higher than 65%. Similar results were also observed for the ammonia emission. An initial moisture level below 45% is not recommended due to concern with lower degree of degradation.  (3)  Biodegradable volatile solids content (BVS) of feedstock had pronounced effect on odor emissions. Peak odor concentration and emission rate increased dramatically as BVS increased up to 55%. Further increase in BVS resulted in little change in odor emission rate. However, total odor emission increased exponentially with BVS.  (4)  Temperature setpoint at 60oC appeared to be a turning point for odor emission. Below this point, odor emissions increased with increasing temperature setpoint. Conversely, above this point, odor emissions decreased with increasing temperature setpoint.  (5)  Peak odor emission rate was closely correlated with aeration rate, initial moisture content and BVS content. The correlation between peak odor emission rate and these operational parameters could be useful for odor prediction.  167  5.5  References  B.C. Ministry of Environment. 2002. Organic Matter Recycling Regulation (OMRR). B.C. Reg. 18/2002. Bruce, M.P. 1998. Odour Production and Oxygen Consumption under Controlled Composting Conditions. Master’s Thesis. Chemical and Bio-Resource Engineering. University of British Columbia. Vancouver, BC. California Integrated Waste management Board (CIWMB), 2007. Comprehensive composting odor response project. Report prepared by San Diego State University. Composting Council of Canada. 1998. Production and Use of Compost Regulation. Web page available at http://www.env.gov.bc.ca/epd/cpr/regs/pauocreg.html. Day, M., K. Shaw and M. Krymien. 1999. Composting Odours: What Can Chemistry Tell Us? Institute for Chemical Process and Environmental Technology, National Research Council. Day M. and K. Shaw, 2001. Biological, Chemical, and Physical Processes of Composting. Chapter in: Compost Utilization in Horticultural Cropping Systems. P.J. Stoffella and B.A. Kahn (eds.) CRC Press. 2001. Elwell, D.L., H.M. Keener, M.C. Wiles, D.C. Borger and L.B. Willett. 2001. Odorous emission and odor control in composting swine manure/sawdust mixes using continuous and intermittent aeration. Trans. ASAE , 44:1307-1316. Epstien, E., 1997. The Science of Composting. Technomic Publishing Company, Penn. Fraser, B.S. & A.K. Lau. 2000. Effect of process control strategies on composting rate and odor emission. Compost Science & Utilization. Vol. 8. No 4. 274-292. Haug, R.T. 1993. The Practical Handbook of Compost Engineering. Lewis Publishers, Boca Raton, FL. Kirchmann, H. and E. Witter, 1989, Ammonia volatilization during aerobic and anaerobic manure decomposition. Plan Soil 115:35-41. Kaiser, J., 1996. Modeling composting as a microbial ecosystem: a simulation approach. Ecological Modeling. 91: 25-37 Miller, F.C., 1993. Minimizing Odor Generation. In Science and Engineering of Composting. Ohio State University.  168  Mills, B., 1996. Review of methods of odor control. Filtration and Separation, 2:147-152. Richard, T.L., 1996. The Science and Engineering of Composting. Web page available at http://compost.css.cornell.edu/science.html. Richard, T.L., 1997. The kinetics of solid-state aerobic biodegradation. Ph.D. Dissertation. Cornell University. Richard, T.L., B. Hamelers, A. Veeken, and T. Silvia, 2002. Moisture relationships in composting processes. Compost Science & Utilization. Vol. 10. No 4. 286-302. Sironi, S., L. Capelli, P. Centola, R. Del Rosso and M. II Grande. 2007. Continuous monitoring of odours from a composting plant using electronic noses, Waste Management, Vol. 27, No. 3, pp. 389-397. Tiquia, S.M. and N.F.Y. Tam, 2000. Fate of nitrogen during composting of chicken manure. Environmental Pollution. 110:535-541 Walker, J.M., 1993. Control of composting odors. In Hoitink, H.I.J. and H.M. Keener (Eds.): Science and Engineering of Composting. The Ohio State University, USA. Wilber, C., and C. Murray, 1990. Odor Source Evaluation. BioCycle 31(3):68-72.  169  CHAPTER 6 ODOR PREDICTIVE MODEL  6.1  Introduction  Odor emissions from composting facilities and their off-site impacts have been major problems for the development of composting industry. With urban encroachment spreading rapidly, the problems have further escalated and forced the composting industry to upgrade the technologies and employ best management practice to minimize the composting odor impact to resolve the conflicts between composting facilities and the neighboring communities without closing or impairing composting operation (San Diego State University, 2007). The combination of odor emission prediction and dispersion modeling may offer a potential solutions to these problems by providing objective means for siting new plants, establishing operating strategies, determining the extent of odor control, demonstrating compliance with regulatory and community odor restrictions, and assessing the impacts on surrounding communities.  Research efforts have been made on composting odor modeling and primarily concentrated on dispersion modeling. Heinemann and Wahanik (1998) linked an odor dispersion model to an odor source generation model to predict odor distribution emanated from mushroom composting facilities. In their odor source generation model, odor emission rate was in fact calculated based on geometric dimension of the composting pile and the emission of dimethyl sulfide reported by Derikx et al. (1990). The limitation of such models is that specific odorant cannot represent the total odor strength (previously known as dilution-tothreshold D/T and now standardized as odor units OU) from composting as no single odorous compound has been identified as a proper predictor of odor sensation. Wu (2000) successfully incorporated the odor strength data (D/T) to the Industrial Source Complex Short Term (ISCST3) air dispersion model recommended by the United States Environmental Protection Agency (USEPA) to analyze the potential impact of composting odor on surrounding communities. This effort demonstrated the usefulness of composting odor modeling, but the source odor parameters used in the model was also a simple A version of this chapter will be submitted for publication. Zhang, W. and Lau, A.K. Development and Validation of an Odor Predictive Model  170  calculation based on the source sizes and odor concentration observed from the facilities studied. Similar approach was also used by Williams and Servo (2005) in odor modeling at a biosolids composting facility. All these models neither take into account the effects of composting operating conditions, nor consider the changes in odor emissions over time. . It is no doubt that predicting the odor emission presents a significant challenge in composting odor modeling. Despite the fact that a number of models for off-site composting odor modeling have published, a source odor predictive model in terms of olfactometry has still not been proposed. This could be partly explained by the fact that composting is characterized by a high degree of intricacy related to a number of physical and biochemical factors in heterogeneous matrix of gas, liquid and solid phases, which fluctuates considerably over time. This is further complicated by the fact that many odors, including those from composting, involve hundreds of odorous compounds. The perception of a mixture of odorants is very different from how each odorant would be perceived independently because odorous compounds do not act in an additive fashion when mixed. The complex nature of odor makes it difficult to predict. For these reasons, deterministic modeling is very difficult for such a complex system. However, odors from organic waste decomposition have been known to be the products of microbial activities (Zhu, 2000; Haug, 1993; Cornell Composting, 1996). It is thus desirable to develop a empirical model with fewer parameters to predict the odor emissions based on the concept of microbiology such that source odor predictive model can be combined with dispersion model for better overall odor modeling.  Various models have been proposed to describe the dynamics of microorganisms at different physical and chemical conditions. One of the most well-known models used in predictive modeling in microbiology is the Gompertz function that originally formulated for actuarial science for fitting human mortality data based on an exponential relationship between specific growth rate and population density (Okpokwasili and Nweke, 2005). The Gompertz model has been modified and applied successfully to bacterial growth prediction (Gibson et al., 1987; Zwietering et al., 1990; McMeekin et al., 1993; Banani et al., 2007). The modified Gompetz model has also been used to predict methane from a landfill bioreactor (Lay, 2000) as well as hydrogen generation and gas production by activities of anaerobic bacteria during  171  fermentation (Beuvink and Kogut, 1993; Lay et al., 1998). Chang et al. (2005) applied the modified Gompertz model to aerobic composting and demonstrated that carbon dioxide evolution associated with thermophilic bacteria during composting process was well fitted with a modified Gompertz equation. Therefore, the Gompertz model may also be suitable for the description of odor evaluation from composting system where anaerobic and aerobic conditions essentially coexist, given the fact that sigmoidal pattern of odor emission from composting has been observed  In this study, an odor predictive model was developed by analogy with microbial activities. The well-known Gompertz model commonly used in predictive microbiology was employed. The proposed odor predictive model constitutes a basic model and adjustment functions to reflect the influences of environmental factors. The structure of the model is easily adaptable for future improvement. Three sets of experimental data were used as a simple verification of the proposed model. A detailed sensitivity analysis of the model parameters was performed to assess the relative importance of individual parameters on the overall model performance. The model can be easily linked to dispersion model for odor modeling at composting facilities.  6.2  6.2.1  Model Development  General Description  Since microbial activities are regarded to be responsible for odor generation, it is reasonable to assume that evolution of odors in compost can be related to microbial activities or microbial population size. However, generalizations of odor emission to microbial population size are difficult without a predefined condition because anaerobic bacteria are known to generate much higher odors than aerobic bacteria do. Therefore, to relate odor to microbial activity, an assumption was made. That is on ideal composting conditions (a reference condition discussed later), odor emission is proportional to microbial activities (population size). Under this assumption, the odor emission can be simulated by analogy with the  172  dynamic changes of microbial population during composting process. In mathematical term, this can be represented by:  G( t )  P ˜ log N ( t )  (7.1)  Where: G = odor emission rate (OUE/h), t = time (hour),  P = proportionality factor representing the odor units generated per capita microorganisms in terms of log counts, N = the population size of microorganisms. A number of studies have demonstrated that the Gompertz model and its modified forms could be regarded as the best model to describe the data of microbial population growth in terms of the goodness-of-fit to the data and its easy use (Zwietering et al., 1990; McMeekin et al., 1993; Gibson et al., 1987; Banani et al., 2007). The modified Gompertz equation by Gibson et al. (1987), which is commonly applied to the description of cell density versus time in bacterial growth curves in terms of exponential growth rates and lag phase duration, is as follows:  log N ( t )  A  De  e   B1 ( t  M1 )  (7.2)  Where: N and t have the same meaning as in Eqn. 7.1, A = value of the lower asymptote {i.e., Log N(-’)}, D = difference in value between the lower and upper asymptote of growth curve {i.e., Log N(-’) - Log N(-’)}, M1 = time at which exponential growth rate is maximal, B1 = a constant related to the slope of the growth curve at point M.  173  It has been reported that microbial survival (death) curves could also be adequately modeled by the Gompertz equation (Bhaduri et al., 1991; Linton et al., 1995, 1996; Xiong et al., 1999). When the Gompertz equation is applied for sigmoidal survival curves, it can be expressed by (Bhaduri et al., 1991):  log N ( t )  H  Ce  e   B2 ( t  M 2 )  (7.3)  Where: N and t have the same meaning as in Eqn. 7.1, H = upper asymptotic value (e.g. maximum population density), C = the difference in value of the upper and lower asymptote of decay curve, M2 = the time at which the absolute death rate is maximal, B2 = a constant related to the slope of the death curve at time M2. Microbial growth process can generally be distinguished into four typical phases: the log, exponential, stationary and death phases as shown in Figure 6.1 (McMeekin et al., 1993; Hajmeer and Cliver, 2002). The size of microbial population varies as a consequence of birth and death. If birth rate of the population exceeds the death rate, the population will increase, conversely, when the death rate exceeds birth rate, population will decrease (Jones and Walker, 1993). Following this assumption, if we use equation 7.2 and 7.3 to represent the growth and death phases, respectively, the combination of these two equations could describe the complete picture of dynamic change process of microbial population. The combined Gompertz equation can be expressed as:  log N ( t )  A  De  e   B1 ( t  M1 )   Ce  e   B2 ( t  M 2 )  (7.4)  Substituting Equation 7.4 into 7.1 and let G0  P ˜ A , G1  P ˜ D and G2  modified Gompertz equation for predicting composting odor is given by:  G( t ) G0  G1e  e   B1 ( t  M 1 )   G2 e  e   B2 ( t  M 2 )  (7.5)  174  P ˜ C , the  Where G0 = value of the lower asymptote of growth curve, G1 = the difference in value of the upper and lower asymptote of growth phase, which can be interpreted as the difference in value between peak and initial odor emission, G2 = the difference in value of the upper and lower asymptote of death phase, which can be interpreted as the difference in value between peak odor and residue odor at the end of active composting, All others have the same meaning as described in previous equations.  Figure 6.1 Schematic of a typical microbial growth curve (adopted from Hajmeer and Cliver, 2002).  The modified Gompertz model (Equation 7.5) is biological relevant, although it does not implicitly include biological parameters. The equation essentially includes two parts. The first part, G0  G1e  e   B1 ( t  M 1 )  , could be regarded as odor production as a result of microbial  growth. The duration of lag phase, which is generally defined as the x-axis intercept of the tangent through inflection point, can be derived from this growth curve (Gibson et al., 1987):  175  t0  M1   1 B1  (7.6)  When microbial growth reaches stationary phase, this part can be approximated by:  G0  G1e  e   B1 ( t  M 1 )  | G0  G1  Gm  (7.7)  Where Gm is the maximum odor emission. The parameter G0 could be interpreted as emission resulting from initial odors inherited from feedstock.  The second part of the equation 7.5 could be seen as the decay of odor emission that is directly related to the magnitude of microbial population death. Subsequently, once reaching the stationary phase, equation 7.5 then becomes:  G  Gm  G2 e  e   B2 ( t  M 2 )  (7.8)  This equation is consistent with the survival curve model by Bhaduri et al. (1991, equation 7.3). Similar to the growth curve, we defined the x-axis intercept of the tangent through inflection point of the decay curve as tail time of odor emission, at where odor emission tends to be steady over time and become significantly low. Based on this definition, the slope of the tangent can be derived by finding the first order derivative of the function (equation 7.8) with respect to t:  dG dt  G 2 e  e   B2 ( t  M 2 )  ˜ [ e  B2 ( t  m2 ) ] ˜ (  B2 )  (7.9)  The slope of tangent can then be obtained by calculating the first order derivative at the inflection point, where t = M2:  176  k  § dG · ¸ ¨ © dt ¹ t   M2  B2 G 2 e  (7.10)  where k is the slope of tangent.  From equation 7.8, at deflection point where t = M2, G = Gm-G2/e. The description of the tangent line through the inflection point M2 is:  G    B2 G 2 G BG t  Gm  2  2 2 M 2 e e e  (7.11)  Thus, the tail odor time tc can be obtained by having G = 0 in equation 7.11:  tc  M2   G e 1  m B2 B2 G 2  (7.12)  Since the tail odor emission Gc is generally small compared with the maximum odor emission Gm, G2  tc  M2   6.2.2  Gm  Gc | Gm . Therefore, equation 7.12 can be rewritten by:  e 1 B2  (7.13)  Reference Odor Predictive Model  An empirical predictive model is essentially a numerical representation of experimental results by means of simple function, like the Gompertz function used here. This is generally true regardless of disciplines. In order to determine the parameters in a predictive model, the response curves must be measured. Clearly, the parameters in the proposed odor predictive model (equation 7.5) are environmentally dependent. Because there are so many environmental variables for a composting system, it is neither practical nor possible to test all cases for each type of composting feedstock and various environmental conditions. To obviate the need to define unique parameters for every case, a two-step model building 177  approach used in crop science for the determination of crop evapotranspiration (ET) was employed. The approach was first introduced in crop science by Penman (1948) to investigate crop ET. Because there are numerous crops and various climate conditions all over the world, the idea is first to determine the ET (Penman equation) for a specific plant alfalfa under a pre-defined reference conditions which are ideal for plant growth. The ET of all other crops is then related to the reference ET by means of correction with crop coefficients. Following this approach (Penman, 1948), the concept of reference composting conditions is introduced in this study. Odor emissions from composting with various feedstocks and environmental conditions can then be related to the odor emission from the reference conditions by means of correction coefficients (effective functions). The reference composting conditions, which are generally considered optimal, may be defined as follows:  (1)  Feedstock: broiler litter amended by sawdust and hogfuel  (2)  Carbon-to-nitrogen ratio: 25  (3)  Moisture content: 55% on wet basis  (4)  Biodegradable volatile solids content: 55%  (5)  Aeration: forced aeration with flowrate of 0.72 L/min.kg DS at 33% duty cycle  Laboratory experiments were carried out under the reference conditions defined above. It was found that the initial and tail odor emission rates are low and close each other under the reference composting conditions. Therefore, the proposed Gompetaz model (Equation 7.5) was further modified by replacing parameter G2 with G1 so that the model has fewer parameters. The modified form was then fitted to the experimental data obtained from four runs. Nonlinear regression was performed using the DataFit software developed by the Oakdale Engineering. The results of parameter estimation from the nonlinear regression analysis are presented in Table 6.1.  Thus, the modified Gompetaz model, referred as reference odor predictive model thereafter, can be expressed by:  Gr  4.2  124.9( e  e  0.128( t 10.8 )   e e  0.08( t  76.6 )  )  178  (7.14)  Where Gr is reference odor emission rate (OUE/h). Using equation 7.6 and 7.13 yields that the lag time t0 = 3 hours, whilst the tail odor time tc = 98 hours, which are quite consistent with the observations throughout this study. As discussed in previous chapters, without significantly altering the composting environment (e.g. extremely low pH or high moisture), temperature usually picked up within 2 to 5 hours after aeration startup, which is an indication of microbial population growth. On the other hand, odor emissions were no longer significant after four days of intensive composting.  Table 6.1 Results of the parameter estimation for reference odor model 95% Confidence Intervals  t-ratio  Prob(t)  25.8  0.409  0.687  -0.070  0.326  1.361  0.190  0.080  -0.811  0.971  0.189  0.853  M1  10.8  -7.1  28.8  1.273  0.219  M2  76.6  -133.0  286.2  0.768  0.453  G1  124.9  82.6  167.1  6.206  0.000  Parameter  Value  G0  4.2  -17.4  B1  0.128  B2  Lower Limit Upper Limit  R2  0.82  The test statistic t-ratio is the ratio of the estimated parameter value to the estimated parameter standard deviation. The larger the ratio is, the more significant the parameter is in the regression model. The p-value (Prob(t)) is used to test the null hypothesis for each parameter. The smaller the value of Prob(t), the less likely the parameter is actually zero. From statistical results presented in Table 6.1, the parameter G1 had the largest t-ratio and smallest Prob(t) values, meaning that it is the most influential parameter in the regression model. Accordingly, the model reflects the fact that the type of feedstock composted is dominating factor to odor emission, as G1 is primarily governed by the nature of feedstock. It was noticed that the parameters representing exponential growth had greater t-ratio and wider variation range for the 95% confidence interval, suggesting that microbial growth rather than death may play a greater role in odor emissions.  179  The nonlinear regression with R2 = 0.82 is reasonably satisfactory, given the fact that odor concentration measurements can vary considerably. In fact, the European standard for dynamic olfactometry suggests that statistical analysis should consider using the logarithm of the odor concentration values (CEN, 2003) because of such nature of odor concentration measurements. It is no doubt that logarithmic transformation would be able to yield better statistics for the odor concentration data. However, a preliminary comparison of nonlinear regression performed with and without logarithmic transformation indicated that fitting the odor data without logarithmic transformation would make the parameters in the predictive odor model more meaningful, that is, reflecting the physical reality. Whether logarithmic transformation of odor concentration data should be applied towards developing the odor predictive model shall be a subject of further investigation.  6.2.3  Generalized Odor Predictive Model  To develop an odor predictive model under non-reference composting conditions, a generalized model with multiplicative structure, which is widely used in composting rate modeling for environmental factors (Schulze, 1961; Jeris and Regan, 1973; Fingers, 1976; Whang and Menaghan, 1980; Cathcart et al., 1986; Nakassaki et al., 1987; Haug, 1993; Stombaugh and Noke, 1996; Richard, 1997), was employed. We define the effect function as the ratio of the odor emission under non-standard composting conditions to that of the reference composting conditions. To generalize the odor model with multiplicative function, an assumption was made that effects of environmental factors at individual levels are assumed to be much greater than their interaction effects. In other words, the interaction effects of environmental factors are small and negligible as compared with the effects of environmental factors at their individual levels. Under this assumption, the generalized multiplicative model that includes multiple correction factors can be expressed by:  n  G( t ) Gr ( t ) ˜ – f i ( xi )  (7.15)  i 1  180  f i ( xi )  Gi ( t ) Gr ( t )  (7.16)  Where n = the number of environmental factors, xi = the effect factors (e.g. aeration, moisture, etc.), ƒi = the effect functions of the environmental factors. When operating under reference conditions, all effect functions equal to 1.  The effect functions apparently vary over time because of the changing characteristics of the microbial growth over the active phase of composting. Since odor measurement is very costly and time consuming, it may not be worthy to experimentally determine the effect function for each of the environmental factors. Considering that the peak odor emission rate is the main concern in odor dispersion modeling, a linear approximation is good enough to estimate the effect of environmental factors. The linear effect function can be easily determined by the data observed at key boundary conditions.  To linearize the effect function, the boundary conditions must be defined first. Similar to the microbial grow phases, the typical phases of odor emission during active phase of composting are characterized by four stages following the work in predictive microbiology (McMeekin et al., 1993):  (1) The lag phase in which microbial cells do not grow, instead store energy and acclimate to the new environment, and microorganisms do not generate new odors in this stage; (2) The exponential phase in which microorganisms proliferate and generate odors rapidly; (3) The decay phase where microorganism death exceeds the birth, and both microbial population and odor generation gradually decline; (4) The stabilizing phase where odor emissions vary little and stabilize at tail odor.  Figure 6.2 illustrates odor emission stages defined above during the active phase of composting. Because odor emissions during lag and tail phase vary little, both can be treated  181  as constant. Therefore, the linearized effect function that can be determined by the boundary conditions in accordance with the four phases is as follows:  f i ( xi )  0  t d to  ­ f 0 ( xi ) ° ° ° f ( x )  f m ( xi )  f 0 ( x i ) ( t  t ) 0 ° 0 i tm  t0 ° ® ° f m ( xi )  f c ( xi ) ( t  tm ) ° f m ( xi )  tm  tc ° ° °[ f ( x )  f ( x )] / 2 e i ¯ c i  t0  t d tm  (7.17) tm  t d tc tc  t d te  Where : t0 = the lag time (hour), tm = the time of peak odor emission occurs, tc = the time of tail odor starts, te = the end of active phase of composting, ƒ0, ƒc, ƒm and ƒe are values of effect functions at corresponding time t0, tc, tm and te. 6 fm  5  Effect function  4 3 2  f0  fc  fe  1 Lag Exponential  Decay  Tail  0 to  tm  Time (hour)  tc  te  Figure 6.2 Schematic of generalized effect function with linear approximation approach  182  6.2.4  Determination of Effect Factors  A number of factors such as feedstock, aeration, free air space, moisture content, temperature and pH are known to affect odor emissions. Idealistically, the model should include all the environmental factors. However, an important assumption underlying the multiplicative model is that the effect factors are independent. Apparently, this is not always the case. As a matter of fact, they are mostly interactive (Richard et al., 2002; San Diego State University, 2007). Therefore, it may not be justified to building a sophisticated model that involves too many factors because of the potential implications and overlaps of the environmental factors. The strategy used in the model building is to choose these more influential, controllable and independent.  The base odor model was developed based on a pre-defined standard feedstock composition. Where the composting feedstock encountered in practice differs from the standard composition, an adjustment is required. As discussed in Chapter 6, odor emissions are primarily influenced by the biodegradable fraction rather than the inert fraction of volatile solid. Thus, the effects of feedstock on odor emission can be related to the biodegradable volatile solid (BVS). The initial, peak and tail odor emissions all are influenced by the BVS content of feedstock. Through a series laboratory experiments as described in Chapter 6, the initial, peak and tail odor emission rates in relation to BVS are determined through regression and given by following equations:  f0  31.614 BVS 2  23.241BVS  4.113  fm  0.01  e  e  fe  2 u 10 5 e19.272˜BVS  (7.18)  43.287 ( BVS 0.466 )  (7.19)  (7.20)  Among the environmental factors, aeration and moisture are considered to be the most influential factors as they directly affect the oxygen level in composting matrix, and hence  183  the potential generation of offensive anaerobic odors. The experimental results presented in Chapter 5 indicate that the initial and tail odor emissions vary little with moisture level and airflow rate, but the peak odor emissions are strongly affected by both of them. So the effects of aeration and initial moisture on initial and tail odor emission are considered to be minimal (i.e. f0 =1). The peak odor emission rates in relation to aeration flowrate and moisture content can be best described by following second-degree polynomial regression equations:  fm  30.864Q 2  41.667Q  15  (7.21)  fm  17.1T 2  16.52T  4.944  (7.22)  where Q is aeration flowrate (L/min.kg DS); and ș is moisture content (% wb). For other environmental factors, free air space (FAS) is considered to be very important to odor generation. However, with forced aeration composting system, its influence on odor emissions is primarily through affecting the effectiveness of aeration. The large variability of FAS in terms of measurement as well as its interaction with aeration flowrate and moisture level also makes it difficult to be incorporated into the model.  It is well known that temperature also has profound effect on odor emissions because higher temperature increases the volatility of odorous compounds. However, rise in temperature is a natural process of decomposition. In practice, temperature is usually controlled though aeration via the temperature setpoint. With lower temperature setpoint, volatile compounds could become less volatile; but it is also accompanied by greater amount of aeration due to cooling requirements, which may carry more volatile compounds from the composting mass to ambient air. So the net effect of lower temperature setpoint on odor emission depends on specific circumstances. In fact, this study and others (Bruce, 1998) found that temperature set point might not be as significant as suggested by some reports in literature. Therefore, the temperature setpoint parameter was not included in the model.  184  Because of the implications among some environmental factors and the difficulties in generalizing the relationships between all environmental factors and odor emissions, only three factors - feedstock, aeration, and moisture were selected and incorporated into the odor predictive model. This does not necessarily mean that other environmental factors should be excluded from the odor predictive model; rather it shall also be a subject of future investigation.  The mathematical equations that describe the odor predictive model are all explicit functions. Therefore, model simulations were readily performed with a spreadsheet developed for the odor predictive model.  6.3  Model Validation  Validation is an important part of the model development. At present, little published olfactometric odor data is available for poultry manure composting with forced aeration. There is also no composting odor predictive model to compare. In order to assess the validity and capability of the predictive model, three sets of olfactometric odor data collected from laboratory experiment under various nonstandard environmental conditions, which is completely unrelated to the model development, were used to compare with the model prediction for providing a simple verification. Table 6.2 describes the case scenarios used in the model validation. The experimental setup and analytical methods were the same as those described in previous chapters.  Figures 6.3 through 6.5 show the measured olfactometric odor data and odor emission profiles predicted by odor predictive model for cases described above. In case #1, composting was operated at low airflow rate, while other environmental conditions were held at standard conditions. As shown in Figure 6.3, the model over predicted odor emission rate at low aeration flow rate. The largest difference between model and experimental observations was at the time of peak odor occurrence, where the model over predicted peak odor emission rate by 33.2%. Odor olfactometric measurement usually varies by a factor of two (2), meaning that the any individual deviation will be at least 100%. Consequently,  185  larger variations in odor concentration are expected at low aeration because odor concentrations are high at low airflow due to anaerobic odor presence. Given the nature of odor measurement, the prediction errors are actually in the acceptable range.  Table 6.2 Scenarios for model validation Parameters  Unit  Case #1  Case #2  Case #3  L/min kg DS  0.36  0.72  0.72  Moisture Content (MC)  %  55  65  61  Biological Volatile Solid (BVS)  %  55  55  46  Low AFR  High MC  Low BVS and high MC  Aeration Flow Rate (AFR)  Composting conditions  Case #2 is to examine the reliability of the odor predictive model under higher moisture condition, while other environmental factors was held at standard levels. The comparison between simulated and experimental results demonstrates very good agreement between the odor emission profile predicted by the odor model and odor emissions observed in laboratory experiment (Figure 6.4).  Case #3 is to test the validity of the odor predictive model under the influences of multiple factors. The composting experiment was operated at a lower level of biodegradable volatile solid combined with a higher moisture level. As illustrated in Figure 6.5, the model predicted the peak odor emission rate fairly well, but the time of peak odor occurrence was not coincident with each other. This discrepancy was likely caused by the low biological volatile solid, which might limit microbial growth due to quick depletion of readily degradable materials that would normally be sufficient for microbial consumption during the first a few days of composting. Thus, it is possible that the peak odor occurrence advanced as a result of limitation of the BVS availability.  186  This prediction discrepancy is also an indication that the model does not have the capability to predict the peak odor occurrence especially when the content of biodegradable volatile solid is low. As a matter of fact, the model did not include the function to simulate the time of peak odor occurrence, in part because for poultry manure composting with forced aeration, peak odor usually occurs in a narrow range of 24 to 48 hours after startup. Nevertheless, in spite of the inconsistency between prediction and observation in peak odor occurrence, the overall performances of the odor predictive model were encouraging because our major concern is the magnitude of peak odor emission instead of the time of peak odor occurrence.  Odor Emission Rate (OUe/hr)  600  500  Measured Predicted  400  300  200  100  0 0  24  48  72  96  120  144  168  192  216  240  264  288  Time (hour)  Figure 6.3 Measured and model predicted odor emission rate as a function of time for case 1.  187  180  Odor Emission Rate (OUe/hr)  160 Measured  140  Predicted  120 100 80 60 40 20 0 0  24  48  72  96  120  144  168  192  216  240  264  288  Time (hour)  Figure 6.4 Measured and model predicted odor emission rate as a function of time for case 2.  45  Odor Emission Rate (OUe/hr)  40 35  Measured  30  Predicted  25 20 15 10 5 0 0  24  48  72  96  120  144  168  192  216  240  264  288  Time (hour)  Figure 6.5 Measured and model predicted odor emission rate as a function of time for case 3. 188  6.4  Sensitivity Analysis  To assess the relative sensitivity of the model to the key composting operational parameters, a sensitivity analysis was performed to determine the relative changes in odor emission with respect to changes in aeration, moisture content and biodegradable volatile solid. This kind of sensitivity analysis is also known as “parametric study”. The relative sensitivity is defined as the percent change in total odor emission for each percent change in the input factor for the particular range being considered. Following the method reported by Zerihum et al. (1996) and Liang (2002), the relative sensitivity may be assessed using the following equation:  Sr  'G / Gmean 'P / Pmean  (7.23)  Where Sr = relative sensitivity (%); ¨G = changes in total odor emission (OU); Gmean = mean value of total odor emission (OU); ¨P = changes in input factor over the rang e being considered; Pmean = mean value of the input factor. The relative sensitivity for each input factor was calculated for specified “practical range” range within the factor, while keeping the other factors not being tested constant at their mean values. The mean values of the input factors used were airflow or aeration rate = 0.72 L/min.kg DS, moisture content = 55%, and BVS = 55%. Table 6.3 presents the calculated values of the relative sensitivity. The results show that the predictive odor model is most sensitive to biodegradable volatile solids. This is reasonable because biodegradable volatile solids content reflects the nature of feedstock. It has been recognized that feedstock is probably the most important determinant of odor generation (CIWMB, 2007). The results also demonstrate that the model is more sensitive to aeration rate than to moisture content. Therefore, where particular materials are composted, more attention should be paid to aeration so as to minimize odor generation.  189  Table 6.3 Relative sensitivity for odor emission with respect to aeration, moisture content and biodegradable volatile solids Aeration rate (L/min kg DS) AFR  TOE  0.36  16553  0.48  10612  0.60  MC  TOE  35  5777  3.93  40  5266  6738  2.57  45  0.72  4930  1.20  0.84  5188  0.96  TOE  45  784  0.61  50  3252  4.30  4954  0.22  55  4930  2.92  50  4842  0.17  60  6578  2.87  0.17  55  4930  0.57  65  8518  3.38  7513  1.54  60  5217  0.96  1.08  11904  2.91  65  5704  1.35  1.08  11904  2.91  70  6392  1.74  75  7278  0.61  55  5596  0.65  55  4812  4.42  9062  2.05  Sr  Biodegradable VS (%) BVS  0.72  Sr  Moisture content (%)  Sr  * Italic numbers are average values; AFR = air flow rare; TOE = total odor emission. In addition, a detailed sensitivity analysis was conducted to investigate the relative importance of individual parameters of the model on the overall model performance, and to determine where error propagation is likely to occur. This kind of sensitivity analysis of the model is essentially against the base or default values of the parameters of the model, whereas the environmental factors are treated as variables in the “effect function” and incorporated in the odor model. The model parameters investigated include the initial odor emission rate (G0), the difference between maximum and initial odor emission rate (G1), relative growth rate (B1), the time of maximal growth rate occurs (M1), relative decay rate (B2), and the time of maximal decay rate occurs (M2). Each of the parameters examined was varied with a deviation of r50% from its base value. The model was run with individual changes of the examined parameter, while other parameters were set at their base values. Table 6.4 shows the parameter values used in the sensitivity analysis.  The results of sensitivity analysis are illustrated in Figures 6.6 through 6.11 and also summarized in Table 6.5. The model appeared insensitive to the changes in parameter G0 (Figure 6.6). Increasing base G0 by 50% resulted in about 6% increase in total odor emission, 190  but virtually no changes in peak odor emission rate (Table 6.5), whilst decreasing base G0 by 50% led to 3% and 6% decreases in peak odor emission rate and total odor emission, respectively. The lack of sensitivity of the parameter G0 could be explained by the small magnitude of initial odor emission rate from poultry manure composting because broiler litter does not have a strong smell. However, the parameter G0 essentially represents the inherent odors. The model could become more sensitive to G0 for substrates such as swine manure and biosolids that have stronger smell.  The model is most sensitive to the deviation of parameter G1. As shown in Figure 6.7 and Table 6.5, variations of both peak odor emission rate and total odor emission were almost proportional to the changes in G1. This is expected because G1is primarily governed by the type and degradability of feedstock that determine what particular intermediate and potentially odorous compounds develop as well as the amount of odors could be produced. Accordingly, significant errors could be introduced from parameter G1. Special attention must be paid in its determination.  Table 6.4 Parameter values used in the sensitivity analysis Parameter  Unit  -50%  Base  +50%  G0  OU/hr  2.1  6.3  G1  OU/hr  62.45  4.2 124.9  187.35  B1  Dimensionless  0.064  0.128  0.192  M1  Hour  5.4  10.8  16.2  B2  Dimensionless  0.04  0.080  0.12  M2  Hour  38.3  76.6  114.9  As illustrated in Figure 6.8 and Table 6.5, the sensitivity of the model predictions increased with decreased values in parameter B1. Increasing B1 by 50% showed little effect on both peak odor emission rate and total odor emission. Conversely, reducing B1 by 50% resulted in 6.4% and 6.7% decreases in peak odor emission rate and total odor emission, respectively. The overall sensitivity of this parameter is low.  191  The sensitivity of the model to changes in M1 is demonstrated in Figure 6.9 and Table 6.5. The overall effects due to changes in M1 are limited. A deviation of 50% in M1 virtually had no influences on peak odor emissions, but resulted in either 7% decrease or 6.6% increase in total odor emission. The parameter M1 primarily determines the length of lag phase. For well aerated composting, it may not have a significant impact on peak odor emissions.  Table 6.5 Summary of sensitivity analysis for peak odor emission and cumulative odor emission to deviations in model parameters* Parameter  -50%  +50%  Peak OER  Cumulative OE  Peak OER  Cumulative OE  G0  -3.1%  -5.8%  0.2%  5.8%  G1  -49.1%  -44.2%  46.2%  44.2%  B1  -6.4%  -6.7%  -1.1%  2.0%  M1  -1.3%  6.6%  -1.8%  -7.0%  B2  -4.4%  9.5%  -0.8%  -3.2%  1.1%  50.2%  M2 -19.7% -49.4% * Changes in percentage are relative to base condition.  Similar to the parameter B1, the sensitivity of the model predictions also increases with decreased values of B2 (Figure 6.10). The Model had little response when increasing B2 by 50%. But, decreasing B1 by 50% resulted in 5.4% and 9.5% decreases in peak odor emission rate and total odor emission (Table 6.5), respectively.  As shown in Figure 6.11, the parameter M2 also plays a significant role in the interpretation of odor emission rate profile. A 50% increase of M2 resulted in about 50% increase in total odor emission. Conversely, decreasing M2 by 50% led to 49% decrease in total odor emission. Because M2 represents how long the microorganisms survive, longer period of high microbial population density could lead to greater cumulative odor emissions, but may not generate higher peak in odor emission rate (Table 6.5).  192  16000  Cumulative Odor Emission (OUe)  14000 12000 10000 8000 6000 4000  50% Increase Base  2000  50% Decrease 0 0  24  48  72  96  120  144  168  192  216  240  264  288  Time (hour)  Figure 6.6 Sensitivity of cumulative odor emission to changes in parameter G0  Cumulative Odor Emission (OUe)  12000  10000  8000  6000  4000  50% Increase Base  2000  50% Decrease  0 0  24  48  72  96  120  144  168  192  216  240  264  288  Time (hour)  Figure 6.7 Sensitivity of cumulative odor emission to changes in parameter G1 193  Cumulative Odor Emission (OUe)  12000  10000  8000  6000 50% Increase Base  4000  50% Decrease 2000  0 0  24  48  72  96  120  144  168  192  216  240  264  288  Time (hour)  Figure 6.8 Sensitivity of cumulative odor emission to changes in parameter B1  Cumulative Odor Emission (OUe)  12000  10000  8000  6000 50% Increase  4000  Base 50% Decrease 2000  0 0  24  48  72  96  120  144  168  192  216  240  264  288  Time (hour)  Figure 6.9 Sensitivity of cumulative odor emission to changes in parameter M11  194  Cumulative Odor Emission (OUe)  12000  10000  8000  6000 50% Increase 4000  Base 50% Decrease  2000  0 0  24  48  72  96  120  144  168  192  216  240  264  288  Time (hour)  Figure 6.10 Sensitivity of cumulative odor emission to changes in parameter B2 16000 50% Increase Cumulative Odor Emission (OUe)  14000  Base 50% Decrease  12000 10000 8000 6000 4000 2000 0 0  24  48  72  96  120  144  168  192  216  240  264  288  Time (hour)  Figure 6.11 Sensitivity of cumulative odor emission to changes in parameter M2  195  In summary, the results of sensitivity analysis generally are in agreement with the statistic results of the regression model presented in section 7.3. The model is very sensitive to parameters G1 and M2, implying that they are significant to the model. Thus, special attention must to be paid to theses two parameters in base model development. On the other hand, they could also be the major sources of errors from the model development. Despite the model appeared not very sensitive to G0 and M1, both of them play an important role in characterizing the odor emission profile. Moreover, the model could become sensitive to G0 for wastes having high odor potential. The low sensitivity of both B1 and B2 suggests that there is a wider range of values that would fit well with the data used for model development, and consequently further investigation is needed. 6.5  Conclusions  Predicting odor emission from composting process presents a considerable challenge due to complexity of the microbial involvement, the variability of odor quantification, and the intricacy of multiple environmental factor effects. A practical strategy to provide realistic odor prediction would be a compromise between empirical and theoretical approach. This thesis study proposed an algorithm to develop the odor predictive model by analogy with microbial activity dynamics during the composting process. Similar to the predictive microbiology, the conventional Gompertz model was employed as the base in the novel model. In the innovative approach, the effects of various environmental factors were incorporated into the predictive model by multiplicative algorithm and linearization approximation. The odor predictive model and corresponding parameters introduced in this work constitute contributions to reducing odor emission at existing composting facilities and improving planning and design for new composting plants.  The preliminary validation of the proposed odor predictive model demonstrated that the model was able to predict the overall behavior of odor emission process, and the prediction was in general agreement with lab experimental observations. Although some prediction errors could be as high as 33%, the errors were in the acceptable range given the considerable variation nature of odor measurement.  196  A sensitivity analysis revealed that the model is most sensitive to the parameter G2 that are dependent on the type of substrate and the time of maximal decay rate M2 that is relative to environmental conditions. The model was less sensitive to so-called relative growth and death rate B1 and B2, suggesting that the two parameters are a subject of further investigation. This model provides a simple and direct mean for estimating odor emissions during composting. The parameters in the model were well defined and easy to obtain. It can serve as a useful tool in predicting odor emission without extensive and expensive odor measurements. Although the model developed in this study was substrate specific (i.e. poultry manure), it could also have general applicability to other substrates since the effects of substrate on odor emissions were incorporated into the model with four parameters initial, maximum, and final correction coefficients (i.e. f0, fm, fe) as well as biodegradability. Finally the limitation of the odor predictive model shall be noted. The multiplicative model assumes that the interaction effects of environmental factors are small and ignorable comparing to the effects of environmental factors at individual levels. This assumption constitutes a limitation of the model, despite the model validation results indicate that the model can sufficiently represent the essential aspects of the system studied. The effects of pH values and free air space were not built into the current model. The changes in peak odor emission occurrence due to different environmental conditions were also not considered. Fortunately they appear insignificant to predicting the peak odor emission rate that is generally a major concern in odor dispersion modeling. Yet, there were some large discrepancies between the model prediction and experimental observations, which indicates the difficulties in odor prediction, and further improvement is needed for the proposed model. To make the odor predictive model more practical, research efforts are required in the future to link the predictive model with an odor dispersion model and combine with geographical information system (GIS), as the impacts of odor off-site are spatially dependent. Further research should also be directed to the calibration and validation of the model and extension of the model capability to different type of organic waste substrates that can generate odor emissions to different extents.  197  6.6  References  Banani, R. C., R. Chakraborty, U.R. Chaudhuri, 2007. Validity of modified Gompertz and Logistic models in predicting cell growth of Pediococcus acidilactici H during the production of bacteriocin pediocin AcH. Journal of Food Engineering 80: 1171–1175. Beuvink, J.M.W. and J. Kogut, 1993. Modeling Gas Production Kinetics of Grass Silages Incubated with Buffered Ruminal Fluid. J. Anim. Sci. 71:1041-1046. Bhaduri, S., Smith, P. W., Palumbo, S. A., Turner- Jones, C. O., Smith, J. L., Marmer, B. S., Buchanan, R. L., Zaika, L. L. and Williams, A. C., 1991. Thermal destruction of L. monocytogenes in liver sausage slurry. Food Microbiol. 8: 75-78. Bruce, M.P. 1998. Odor Production and Oxygen Consumption under Controlled Composting Conditions. Master’s Thesis. Chemical and Bio-Resource Engineering. University of British Columbia. Vancouver, BC. Campbell, A.G., R.L. Folk and R.R. Tripepi, 1997. Wood ash as an amendment in municipal sewage sludge and yard waste composting processes. Compost Science and Utilization, 5(1):62-73. Chang, J.I., J.J. Tsai and K.H. Wu, Mathematical model for carbon dioxide evolution from the thermophilic composting of synthetic food wastes made of dog food. Waste Management 25: 1037–1045. Committee for European Normalization (CEN). 2003. EN13725: Air Quality – Determination of Odour Concentration by Dynamic Olfactometry. Brussels, Belgium. Cornell Composting, 1996. "The Science and Engineering of Composting." Web page available at: http://compost.css.cornell.edu/science.html. Derikx, P.J.L., H.J.M Op Den camp, C. van der Dift, L.J.L.D. van Griensven, and G.D. Vogels. 1990. Odorous sulfur compounds emitted during production used as a substrate in mushroom cultivation. Appl. Envirn. Microbiol. 56(1): 176-180. Gibson, A.M., N. Bratchell, T.A. Roberts, 1987. The effects of sodium chloride and temperature on the rate and extent of growth of Clostridium botulinum type A in pasteurized pork slurry. J. Appl. Bacteriol. 62: 479-490.  198  Hajmeer M.N., and Cliver D.O. 2002. Microbiology of Food Preservation and Sanitation. In Foodborne Diseases. 2nd Ed. Cliver, D.O. and Riemann, H. (Eds.) Academic Press, New York, NY. Chapter 22. pp.330–352. Hamelers, H.V.M., 2004. Modeling composting kinetics: A review of approaches. Reviews in Environmental Science & Bio/Technology 3: 331–342. Haug, R.T. 1993. The Practical Handbook of Compost Engineering. Lewis Publishers, Boca Raton, FL. Heinemann, P. and D. Wahanik, 1998. Modeling the generation and dispersion of odors from composting mushroom facilities. Transactions Of the ASAE, vol. 41(2): 437-446. Jones, J.E. and S.J. Walker, 1993. Advances in modeling microbial growth. Journal of Industrial Microbiology, 12: 200-205 Lay, J.J., 2000. Biohydrogen generation by mesophilic anaerobic fermentation of microcrystalline cellulose. Biotechnol. Bioeng. 74 (4), 280–287. Lay, J.J., Li, Y.Y., Noike, T., 1998. A mathematical Model for methane production from a landfill bioreactor. J. Environ. Engrg., ASCE 124, 730–736. Liang, Z. S. 2002. Modeling ammonia emission from swine anaerobic lagoons. Transactions of the ASAE, Vol. 45(3): 787–798. Linton, R. H., Carter, W. H., Pierson, M. D. and Hackney, C. R., 1995. Use of a modified Gompertz equation to model nonlinear survival curves for Listeria monocytogenes Scott A. J. Food Prot. 58: 946-954. Linton, R. H., Carter,W. H., Pierson,M. D., Hackney, C. R. and Eifert, J. D. (1996) Use of a modified Gompertz equation to predict the effects of temperature, pH, and NaCl on the inactivation of Listeria monocytogenes Scott A heated in infant formula. J. Food Prot. 59: 16-23. Lystad, H., J.P. Hammer, R. Sørhein, O. Molland, O. Bergersen, and P. Smits, 2002. Using lime to reduce odors from biowaste composting. BioCycle, 43(9) 38-41. McMeekin, T.A.., J.N. Olley, T. Ross, and D.A. Ratkowsky, 1993. Predictive Microbiology: Theory and Application. RSP, Taunton, England. Montgomery County Regional Composting Facility, 1999. Reducing odor and VOC emissions. BioCycle, Vol. 40(3): 68-72.  199  Okpokwasili, G.C. and C.O. Nweke, 2005. Microbial growth and substrate utilization kinetics. African Journal of biotechnology Vol. 5(4): 305-317. Penman, H.L., 1948. Natural evaporation from open water, bare soil and grass. Proceedings of the Royal Society, Series A. 193:120-145. Richard, T.L., B. Hamelers, A. Veeken, and T. Silvia, 2002. Moisture relationships in composting processes. Compost Science & Utilization. Vol. 10. No 4. 286-302. Rosenfeld, P., M. Grey and M. Suffet, 2002. Controlling odors using high carbon wood ash. BioCycle. 42(3):43-45. San Diego State University, 2007. Comprehensive composting odor response project. Report to the California Integrated Waste management Board. Sawyer, C.N. and P.L. McCarty, 1978. Chemistry for Environmental Engineering. McGrawHill, Inc., Boston, MA. Williams, T.O. and S. Servo, 2005. Odor modeling at a biosolids composting facility. BioCycle, Vol. 46(3): 25-28. Wu, N., 2000, Odor modeling as first line of defense. BioCycle, Vol. 41(5): 64-68. Xiong, R.G., G. Xie, A.S. Edmondson, R.H. Linton and M. A. Sheard, 1999. Comparison of the Baranyi model with the modified Gompertz equation for modelling thermal inactivation of Listeria monocytogenes Scott A. Food Microbiology, 1999, 16: 269-279. Zerihun, D., J. Feyen, and J. M. Reddy. 1996. Sensitivity analysis of furrow–irrigation performance parameters. J. Irrigation and Drainage Eng. 122(1): 49–57 Zhu, J., 2000. A review of microbiology in swine manure odor control. Agriculture, Ecosystem and Environment 78:93-106. Zwietering, M.H., Jongenburger, I., Rimbouts, F.M., van t Riet, K., 1990. Modeling of the bacteria growth curve. Appl. Environ. Microbiol. 56 (6), 1857–1881.  200  CHAPTER 7 SUMMARY AND CONCLUSIONS  Traditional measures used by the majority of composting facility operators for odor and ammonia emission control are those involving collection and treatment such as thermal oxidation, absorption, wet scrubbing, chemical oxidation and biofiltration (Haug, 1993; Mills, 1996; Goldstein, 2005; Sironi and Botta, 2001). However, these solutions to the odor challenge require additional equipment and space, which could be costly, and do not address the sources of problems. This thesis systematically studied preventive means for composting odor control. Laboratory experiments were performed to examine the effectiveness of various technologies or techniques to minimize ammonia and odor emissions. Ammonia was treated separately from other composting odors because of its unique characteristics. These techniques include the addition of magnesium and phosphate salts for struvite precipitation; the use of chemical and biological additives in the form of yeast, zeolite and alum to reduce ammonia and odor emissions; and the manipulation of key operational parameters during the composting process. A predictive odor model was then developed, which can potentially be used along with odor dispersion modeling to become a more comprehensive mathematical model for simulating the impact of odor emissions on the neighboring communities. In this concluding chapter, the key findings and general conclusions from the thesis research will be summarized. It will also discuss the limitations of present research efforts, and identify the areas for future research.  7.1 Summary and Conclusions 1.  In searching for effective means in minimizing odor emissions from composting, the problem must be quantified first. In this study, a forced-choice dynamic dilution olfactometer previously assembled in accordance with the ASTM E-679-91 Standard (ASTM, 1991) formed the basis for the upgrading design of the olfactometer. The upgraded olfactometer would be in compliance with the European/international standard EN13725 (CEN, 2003), in regard to performance quality requirements, quality requirements for the dilution apparatus (olfactometer), odor panel procedure and data  201  analysis. The upgraded olfactometer was also used to evaluate odor removal efficiency of the biofilters in Lulu Island Wastewater Treatment Plant and Annacis Island Wastewater Treatment Plant in the Greater Vancouver Regional District as well as other industrial operations. The instrument has been proven to be a useful tool in odor research and industry use.  Results of odor measurements in terms of odor concentrations as obtained from the various lab-scale composting tests in this study were in the same order of magnitude (103–104 OU) as those reported in the literature for composting systems with forced aeration.  1.  An innovative technology to precipitate ammonia into struvite before its release from the composting matrix was investigated. This is the first time that struvite crystals were discovered in manure composting medium, that is, other than food waste composting. It constitutes a major contribution to the composting science and industry because it proved that struvite, as a valuable slow-release fertilizer, could be formed in a porous composting medium, and is not necessarily limited to a supersaturated liquid medium. This technique may be used to maximize nitrogen retention in finished compost, thereby increasing its agronomic value. Furthermore, the study demonstrated that it is possible to form struvite in the composting matrix at a moisture level of 55%, suggesting that struvite could precipitate without compromising the optimal moisture for composting.  The results showed that the addition of magnesium and phosphate salts in poultry manure composting mixture could substantially reduce ammonia emission by 40-84%, while ammonium nitrogen in finished compost increased by 37%-105% compared with the control treatment, depending on the dosages used. These phenomena can be attributed to struvite formation, given the fact that struvite presence in the compost was verified by both XRD and SEM-EDS. There were no adverse effects on the composting process, with respect to substrate degradation and salinity level of compost product, within the range of application rates tested,  202  2.  The evaluation of chemical and biological additives for odor emissions reduction showed that the dosage of yeast and zeolite at 10% (dry mass basis) could increase the degree of degradation by 6-8%, whereas ammonia emission could be reduced by up to 50%. Their effectiveness in reducing total odor emission by 15% was not very high. The application of yeast alone could also enhance biodegradation, but its effectiveness on ammonia emission reduction was finite. The study demonstrated that yeast has the potential in boosting the composting processes; for instance, yeast could be used as an effective activator in case of composting failure.  Alum was found to be the most effective, among the additives tested, in reducing both ammonia and odor emissions. The incorporation of alum into feedstock reduced ammonia emission by up to 90% depending on the dosage used, thus greatly enhanced nitrogen conservation. At the same time, odor emission assessed via olfactometry was reduced by 44% with a dosage of 2.5% (dry mass basis), which may be considered costeffective.  3.  Better understanding of the effects of key operational parameters or environmental factors on odor emission is of critical importance for minimizing composting odors. This study reaffirmed that aeration is the most influential factor to odor emissions from composting. The study revealed that airflow rates that were too high or too low could result in higher total odor emissions. An optimal flowrate for odor control would be approximately 0.6 L/min.kg dry matter with an intermittent aeration system and a duty cycle of 33%. Temperature setpoint at 60oC appeared to be a turning point for odor emission. Below this point, odor emissions increased with increasing temperature setpoint; conversely, odor emissions decreased with increasing temperature setpoint above this point  With regard to the composting material properties, odor emissions were greatly affected by the initial moisture content of feedstock. Both peak odor concentration and emission rate generally increased with higher initial moisture content. Odor emission appeared significant only at moisture levels higher than 65%. Similar results were observed for  203  ammonia emission. An initial moisture level below 45% is not recommended due to concern with the resulting lower degree of degradation. The biodegradable volatile solids (BVS) content of feedstock had profound influence on odor emissions. Peak odor concentration and emission rate increased dramatically as BVS increased to 55%; further increase resulted in negligible changes.  4.  An odor predictive model was developed via making analogy with microbial activity dynamics during the composting process. The correlations established between peak odor emission rate and the operational parameters - aeration rate, initial moisture content and BVS content were incorporated into the predictive model by the multiplicative algorithm and linearization approximation approach. The odor predictive model would fill the gaps in composting odor modeling, as currently there is no source model available to predict composting odor emission.  The preliminary validation of the proposed model demonstrated that it was able to predict the overall trends of odor emission, and the prediction was in general agreement with experimental observations. The odor predictive model can provide a simple and direct means for simulating odor emissions during the active phase of composting; the parameters in the model were well-defined and may be readily obtained.  7.2 Contributions to Research and Practical Applications in Composting This thesis systematically studied preventive means for composting odor control. Laboratory experiments were performed to examine the effectiveness of various technologies or techniques to minimize ammonia and odor emissions.  1.  Because electronic nose technology is not yet applicable to complex odors, olfactometry and odor panel tests remain the international standard for odor measurements. A forcedchoice dynamic dilution olfactometer previously assembled in accordance with the ASTM Standard was successfully upgraded to be in compliance with the European/international standard EN13725 in regard to performance quality requirements  204  for the dilution apparatus (olfactometer), odor panel procedure and data analysis. Aside from being used as a tool in odor research, the upgraded olfactometer has since been used by the Greater Vancouver Regional District (Metro Vancouver) as well as other industrial operations. To the knowledge of the author, there are no other olfactometers or odor analysis laboratories in the region while there is demand for odor panel tests from time to time. Further fine-tuning and development of this instrument can make it highly compatible with commercially available olfactometers.  2.  Struvite crystals were discovered during the manure composting process, and the presence of struvite in the compost was verified by both XRD and SEM-EDS. It constitutes a major contribution to composting science and technology because the study proved that struvite, as a valuable slow-release fertilizer, could be formed in a wet porous composting medium, and is not necessarily limited to a supersaturated liquid medium. Struvite precipitation during food waste composting has once been reported by other researchers; however, their study did not apply the XRD method for more precise confirmation of the struvite crystals. The use of poultry manure in the thesis research was also based on the fact that it has higher concentrations of magnesium, ammoniumnitrogen and phosphate versus food waste; hence, it had greater potential for struvite formation. Even though the small size of the struvite crystals makes it impractical to separate them from finished compost, its presence in the compost would enhance the agronomic/fertilizer value via two desirable characteristics – nitrogen retention and slow-release.  3.  The study demonstrated that yeast could boost the composting process; potentially yeast could be used as an effective activator in case of process failure for the composting operation. In contrast to yeast and zeolite, alum was found to be the most effective in reducing both ammonia and odor emissions, and this has not been previously investigated.  4.  Previous studies as reported in the literature have shown that aeration rate might not be effective in reducing odor emissions. Findings from this study, however, revealed that  205  airflow rates that were too high or too low could result in higher odor emission rates and total odor emissions, and an optimal air flowrate for odor control was determined. This optimal aeration rate is smaller than the current industry standard aeration rate, which is based on oxygen requirements for aerobic degradation during the active phase of composting. Potentially, using a lower aeration rate could save some energy cost for a composting operation.  5.  An odor predictive model was developed via making analogy with microbial activity dynamics during the composting process. It would fill the gaps in composting odor modeling, as currently there is no source generation model available to predict composting odor emission. Although the empirical model is specific for the type of feedstock used in the study, the coefficients of the model could be obtained using a similar procedure developed in this thesis if other organic wastes such as food waste, fish waste and biosolids are used as the feedstock. The predictive odor model can potentially be used along with odor dispersion modeling and become a more comprehensive mathematical model for simulating the impact of odor emissions on the neighbouring communities. Dispersion modeling technique is generally used by regulatory agencies to establish the air pollutant emission standards for industrial sources, with respect to the stipulated ambient air quality standard.  7.3 Recommendations for Future Research 1. This study demonstrated that adding magnesium and phosphate salts to feedstock to form struvite could substantially reduce ammonia emission from poultry manure composting. However, this would increase the cost of composting. Further experiments are needed to replace magnesium and phosphate salts with waste materials or other low cost materials (preferably organic wastes) to reduce the cost and be more compatible with the principle of recycling and hence sustainability. In addition, more research is needed to understand the kinetics of struvite precipitation in the composting matrix.  206  2. More detailed compost quality tests shall be conducted in future investigations of the various technologies or techniques involved in this thesis research. Moreover, for the techniques that were found to be effective in preventive odor control to different extents, it is recommended that crop growth trials/field studies be conducted with the finished compost product, in order to assess their overall benefits.  3. Quantitatively understanding the effects of operational parameters on odor emissions can not only gain important insights into odor generation and release, but also provide the basis for odor predictive modeling. While results from the odor predictive model developed in this study appeared to be in good agreement with odor observations from the lab-scale composting tests, it is important to note that the predictive odor model has not been validated through independent investigation. Additional experiments across the entire range of typical composting practices, and using different types of organic wastes (substrates), are needed to provide more valuable insights into the mechanism of odor production and enhance the applicability and accuracy of the model. Future work should also incorporate the important environmental factors pH and free air space into the odor predictive model.  4. Finally, to implement the odor predictive model in practice, research efforts are required in the future to link the predictive model with an odor dispersion model, and more significantly to combine with geographical information system (GIS) in composting odor modeling, as off-site odor impacts from composting facilities are spatially dependent.  207  7.4 References ASTM. American Society for Testing Materials. 1991. Standard Practice for Determination of Odor Thresholds By a Forced-Choice Ascending Concentration Series Method of Limits. Annual Book of ASTM Standards. Designation E 679-79. Committee for European Normalization (CEN), 2003. EN13725: Air Quality–Determination of Odour Concentration by Dynamic Olfactometry. Brussels, Belgium. Goldstein, J. and N. Goldstein. 2005. Controlling Odors at Composting Facilities. Pages 22 in BioCycle. Vol. 46. Haug, R.T. 1993. The Practical Handbook of Compost Engineering. Lewis Publishers, Boca Raton, FL. Mills, B., 1996. Review of methods of odour control. Filtration and Seperation, 2:147-152. Sironi, S. and D. Botta, 2001. Biofilter efficiency in odor abatement at composting plants. Compost Science & Utilization, 9(2):149–155.  208  APPENDIX I  RESULTS OF OLFACTOMETER CALIBRATION  The following figures summarize the calibration results of the triangular forced choice dynamic olfactometer using bubble flowmeters. Detailed calibration procedures are presented in Chapter 3.  209  Figure 3.8 Dilution calibration curves (P1 & P2) 140000  120000  P-1  Dilution Settings  100000  P-2  80000  60000  40000  20000  0 0  1  2  3  4  5 Delivery settings  210  6  7  8  9  10  Figure 3.9 Dilution calibration curves (P3 & P4) 40000  35000  P-4 30000  Dilution ratio  25000  20000  P-3 15000  10000  5000  0 0  1  2  3  4  5 Delivery settings  211  6  7  8  9  10  Figure 3.10 Dilution calibration curves (P5 & P6) 3000  2500  P-5  Dilution ratio  2000  1500  1000  P-6  500  0 0  1  2  3  4  5 Delivery settings  212  6  7  8  9  10  APPENDIX II  TEMPERATURE PROFILES OF THE TESTS WITH DIFFERENT OPERTION CONDITIONS  The following figures present the temperature profiles of all tests under various operation conditions. The parameters tested include aeration rate, initial moisture content, temperature set point, and biodegradable volatile solids. Detailed discussions are presented in Chapter 6.  213  80  0.36 L/min kg DS  70  0.54 L/min kg DS 0.72 L/min kg DS  60  Temperature (°C)  1.08 L/min kg DS Amb  50  40  30  20  10  0 0  24  48  72  96  120  144  168  192  216  240  Time (hours)  Temperature profiles of treatments with different aeration rates  214  264  288  Temperature profiles of treatments with different initial moisture contents 80  70  45% 55% 65%  60  75% Amb  Temperature (°C)  50  40  30  20  10  0 0  24  48  72  96  120  144  168  Time (hrs)  215  192  216  240  264  288  312  80  70 55 °C 60  60 °C 65 °C  Temperature (oC)  50  70 °C Amb  40  30  20  10  0 0  24  48  72  96  120  144  168  Time (hours)  Temperature profiles of treatments with different temperature set points  216  192  216  80 65 BVS 55 BVS  70  50 BVS 45 BVS 60  Amb  Temperature (°C)  50  40  30  20  10  0 0  24  48  72  96  120  144  168  Time (hours)  Temperature profiles of treatments with different biodegradable volatile solids  217  192  216  APPENDIX III  SUMARRY OF STATISTICAL ANALYSIS  The following summarize the results of the experimental results and statistical analysis for thermal parameters, ammonia emission and odor emission from all test series.  218  Struvite Test - ANOVA Analysis with EXCEL I.  Anova: Single Factor SUMMARY Groups TMT1 TMT2 TMT3 CONTROL  II.  III.  Peak Temperature  Count  Sum 128 127 133 143  2 2 2 2  ANOVA Source of Variation Between Groups Within Groups  SS 80.375 35.5  df 3 4  Total  115.875  7  Anova: Single Factor  MS F 26.791667 3.018779 8.875  P-value 0.15683  F crit 6.591392  P-value 0.002385  F crit 6.591392  Ammonia  SUMMARY Groups TMT1 TMT2 TMT3 CONTROL  Count 2 2 2 2  ANOVA Source of Variation Between Groups Within Groups  SS df 5950694.1 3 221455.44 4  Total  6172149.6 7  Anova: Single Factor SUMMARY Groups TMT1 TMT2 TMT3 CONTROL  Average Variance 64 2 63.5 24.5 66.5 4.5 71.5 4.5  Sum 1105.443 2059.629 3012.726 5725.786  Average 552.72153 1029.8147 1506.3632 2862.8928  Variance 131969.6 61397.38 4.818444 28083.61  MS F 1983564.7 35.82779 55363.86  Time of Peak (Tp)  Count 2 2 2 2  Sum 186 187 147 57  219 203  Average 93 93.5 73.5 28.5  Variance 1682 1860.5 1104.5 264.5  Struvite Test - ANOVA Analysis with EXCEL  IV  ANOVA Source of Variation Between Groups Within Groups  SS 5595.375 4911.5  Total  10506.875 7  df 3 4  F 1.518986  P-value 0.338873  F crit 6.591392  T55oC  Anova: Single Factor SUMMARY Groups TMT1 TMT2 TMT3 CONTROL  MS 1865.125 1227.875  Count 2 2 2 2  ANOVA Source of Variation Between Groups Within Groups  SS 1597.375 1444.5  Total  3041.875  Sum 220 232 267 188  df  MS 3 532.45833 4 361.125 7  220 204  Average Variance 110 338 116 338 133.5 480.5 94 288  F 1.474443  P-value 0.348402  F crit 6.591392  Yeast and Zeolite Test - ANOVA Analysis with EXCEL I.  Anova: Single Factor SUMMARY Groups TMT1 TMT2 TMT3 CONTROL  II.  III.  Peak Temperature  Count  Sum Average Variance 140 70 2 144 72 2 143 71.5 0.5 142 71 2  2 2 2 2  ANOVA Source of Variation Between Groups Within Groups  SS 4.375 6.5  df 3 4  Total  10.875  7  MS F P-value F crit 1.458333 0.897436 0.516064 6.591382 1.625  Time of Peak (Tp)  Anova: Single Factor SUMMARY Groups TMT1 TMT2 TMT3 CONTROL  Count 2 2 2 2  Sum 50 50 46 37  Average 25 25 23 18.5  ANOVA Source of Variation Between Groups Within Groups  SS 56.375 86.5  df 3 4  MS F P-value F crit 18.79167 0.868979 0.527061 6.591382 21.625  Total  142.875  7  T55oC  Anova: Single Factor SUMMARY Groups TMT1 TMT2 TMT3 CONTROL  Variance 8 72 2 4.5  Count 2 2 2 2  Sum 182 209 204 156  221 205  Average 91 104.5 102 78  Variance 578 0.5 72 2  Yeast and Zeolite Test - ANOVA Analysis with EXCEL  IV  ANOVA Source of Variation Between Groups Within Groups  SS 878.375 652.5  Total  1530.875 7  Anova: Single Factor SUMMARY Groups TMT1 TMT2 TMT3 CONTROL  V  df 3 4  Odor Concentration  Count  Sum Average Variance 2 17377.86 8688.928 16777216 2 13984.62 6992.309 2878515 2 9889 4944.5 1439905 2 11585.24 5792.619 0  ANOVA Source of Variation Between Groups Within Groups  SS 15819976 21095636  Total  36915611  Anova: Single Factor SUMMARY Groups TMT1 TMT2 TMT3 CONTROL  MS F P-value F crit 292.7917 1.794891 0.287509 6.591382 163.125  df 3 4  MS F P-value F crit 5273325 0.999889 0.478987 6.591382 5273909  7  Cumulative Odor Emission  Count 2 2 2 2  ANOVA Source of Variation Between Groups Within Groups  SS 39813559 12285866  Total  52099425  Sum 28059.65 24053.25 19318.89 16386.46  df  MS F P-value F crit 3 13271186 4.320798 0.09567 6.591382 4 3071467 7  222 206  Average Variance 14029.82 7324259 12026.63 4202066 9659.445 752630 8193.23 6911.285  Alum Test - ANOVA Analysis with EXCEL I.  SUMMARY Groups TMT1 TMT2 TMT3 CONTROL  II.  Count  Sum Average Variance 128 64 2 131 65.5 4.5 138 69 2 141 70.5 0.5  2 2 2 2  ANOVA Source of Variation Between Groups Within Groups  SS 54.5 9  df 3 4  Total  63.5  7  Anova: Single Factor SUMMARY Groups TMT1 TMT2 TMT3 CONTROL  III.  Peak Temperature  Anova: Single Factor  Time of Peak (Tp)  Count  Sum Average Variance 225 112.5 480.5 139 69.5 24.5 98 49 8 77 38.5 40.5  2 2 2 2  ANOVA Source of Variation Between Groups Within Groups  SS df 6424.375 3 553.5 4  Total  6977.875 7  MS F P-value F crit 2141.458 15.47576 0.011481 6.591382 138.375  T55oC  Anova: Single Factor SUMMARY Groups TMT1 TMT2 TMT3 CONTROL  MS F P-value F crit 18.16667 8.074074 0.035835 6.591382 2.25  Count 2 2 2 2  ANOVA Source of Variation Between Groups Within Groups  SS df 6985.375 3 1081.5 4  Total  8066.875 7  Sum Average Variance 347 173.5 264.5 301 150.5 264.5 228 114 288 197 98.5 264.5  MS F P-value F crit 2328.458 8.611959 0.032155 6.591382 270.375  223 207  Alum Test - ANOVA Analysis with EXCEL  IV  Anova: Single Factor SUMMARY Groups TMT1 TMT2 TMT3 CONTROL  V  Odor Concentration  Count 2 2 2 2  ANOVA Source of Variation Between Groups Within Groups  SS 13271262 2824827  Total  16096089  Anova: Single Factor SUMMARY Groups Groups TMT1 TMT2 TMT3 CONTROL  Sum 9640.155 8440.464 10152.31 15184.31  df  Average Variance 4820.077 1048576 4220.232 30867.2 5076.155 1025755 7592.155 719628.8  MS F P-value F crit 3 4423754 6.264106 0.054251 6.591382 4 706206.8 7  Cumulative Odor Emission  Count Count 2 2 2 2  ANOVA Source of Variation Between Groups Within Groups  SS 49817172 9578355  Total  59395527  Sum Sum 11842.5 10890.82 19867.49 22446.92  df  Variance Variance 5471854 77296.69 128772.3 3900432  MS F P-value F crit 3 16605724 6.934687 0.046055 6.591382 4 2394589 7  224 208  Average Average 5921.251 5445.408 9933.744 11223.46  Aeration Test - ANOVA Analysis with EXCEL I.  SUMMARY Groups TMT1 TMT2 TMT3 TMT4  II.  Count 2 2 2 2  Sum Average 141.8097 70.90485 147.7568 73.8784 144.9872 72.4936 144.6554 72.3277  ANOVA Source of Variation Between Groups Within Groups  SS df 8.870246 3 10.19134 4  Total  19.06158 7  Anova: Single Factor SUMMARY Groups TMT1 TMT2 TMT3 TMT4  III.  Peak Temperature  Anova: Single Factor  MS F P-value F crit 2.956749 1.160495 0.427975 6.591382 2.547834  Time of Peak (Tp)  Count  Sum Average Variance 115 57.5 84.5 72 36 32 48 24 2 43 21.5 0.5  2 2 2 2  ANOVA Source of Variation Between Groups Within Groups  SS df 2455.375 3 1102.5 4  Total  3557.875 7  MS F P-value F crit 818.4583 2.969463 0.160205 6.591382 275.625  T55oC  Anova: Single Factor SUMMARY Groups TMT1 TMT2 TMT3 TMT4  Variance 7.256907 1.543173 0.487282 0.903975  Count 2 2 2 2  ANOVA Source of Variation Between Groups Within Groups  SS df 2455.375 3 1102.5 4  Total  3557.875 7  Sum Average Variance 243 121.5 220.5 296 148 242 222 111 128 202 101 512  MS F P-value F crit 818.4583 2.969463 0.160205 6.591382 275.625  225 209  Aeration Test - ANOVA Analysis with EXCEL  IV  Anova: Single Factor SUMMARY Groups TMT1 TMT2 TMT3 TMT4  V  Odor Concentration  Count 2 2 2 2  ANOVA Source of Variation Between Groups Within Groups  SS 87031808 4194304  Total  91226112  Anova: Single Factor SUMMARY Groups TMT1 TMT2 TMT3 TMT4  Sum 23170.48 5792.619 8688.928 11585.24  df  Average Variance 11585.24 0 2896.309 0 4344.464 4194304 5792.619 0  MS F P-value F crit 3 29010603 27.66667 0.003902 6.591382 4 1048576 7  Cumulative Odor Emission  Count 2 2 2 2  ANOVA Source of Variation Between Groups Within Groups  SS 47104151 37290600  Total  84394751  Sum 13929.54 9693.054 12042.81 22546.71  df  Variance 317765.8 868616 13616190 22488029  MS F 3 15701384 1.684219 4 9322650 7  226 210  Average 6964.771 4846.527 6021.403 11273.36  P-value F crit 0.30665 6.591382  BVS Test - ANOVA Analysis with EXCEL I.  SUMMARY Groups TMT1 TMT2 TMT3 TMT4  II.  Count 2 2 2 2  Sum Average 140.8004 70.4002 142.8099 71.40495 141.5666 70.7833 129.1055 64.55275  ANOVA Source of Variation Between Groups Within Groups  SS df 60.7539 3 15.11567 4  Total  75.86958 7  Anova: Single Factor SUMMARY Groups TMT1 TMT2 TMT3 TMT4  III.  Peak Temperature  Anova: Single Factor  MS F P-value F crit 20.2513 5.359021 0.069282 6.591382 3.778918  Time of Peak (Tp)  Count  Sum 2 2 2 2  84 49 49 58  ANOVA Source of Variation Between Groups Within Groups  SS 411 27  df 3 4  Total  438  7  Average Variance 42 8 24.5 0.5 24.5 0.5 29 18  MS 137 6.75  F 20.2963  P-value F crit 0.006977 6.591382  T55oC  Anova: Single Factor SUMMARY Groups TMT1 TMT2 TMT3 TMT4  Variance 5.11872 3.947769 1.227118 4.822065  Count 2 2 2 2  ANOVA Source of Variation Between Groups Within Groups  SS df 9939.375 3 767.5 4  Total  10706.88 7  Sum Average Variance 285 142.5 612.5 209 104.5 40.5 156 78 2 93 46.5 112.5  MS F 3313.125 17.2671 191.875  227 220 211  P-value F crit 0.009401 6.591382  BVS Test - ANOVA Analysis with EXCEL  IV  Anova: Single Factor SUMMARY Groups TMT1 TMT2 TMT3 TMT4  V  Odor Concentration  Count 2 2 2 2  ANOVA Source of Variation Between Groups Within Groups  SS 36071455 19654355  Total  55725810  Anova: Single Factor SUMMARY Groups TMT1 TMT2 TMT3 TMT4  ANOVA Source of Variation Between Groups Within Groups Total  Sum 17377.62 9888.619 6992.309 6992.309  df  Average 8688.809 4944.309 3496.155 3496.155  Variance 16775840 1439258 719628.8 719628.8  MS F P-value F crit 3 12023818 2.447054 0.203691 6.591382 4 4913589 7  Cumulative Odor Emission  Count 2 2 2 2  SS 1.64E+08 5641035  Sum 28704.08 15186.36 10698.67 3900.797  df  1.7E+08  MS F P-value F crit 3 54826688 38.87704 0.002039 6.591382 4 1410259 7  228 219 212  Average Variance 14352.04 1474791 7593.182 218470.6 5349.335 3929701 1950.399 18071.86  

Cite

Citation Scheme:

        

Citations by CSL (citeproc-js)

Usage Statistics

Share

Embed

Customize your widget with the following options, then copy and paste the code below into the HTML of your page to embed this item in your website.
                        
                            <div id="ubcOpenCollectionsWidgetDisplay">
                            <script id="ubcOpenCollectionsWidget"
                            src="{[{embed.src}]}"
                            data-item="{[{embed.item}]}"
                            data-collection="{[{embed.collection}]}"
                            data-metadata="{[{embed.showMetadata}]}"
                            data-width="{[{embed.width}]}"
                            async >
                            </script>
                            </div>
                        
                    
IIIF logo Our image viewer uses the IIIF 2.0 standard. To load this item in other compatible viewers, use this url:
http://iiif.library.ubc.ca/presentation/dsp.24.1-0058524/manifest

Comment

Related Items