~_, _ _ ‘_ -.__ ‘M- .3 \ THESIS . . r???“ ." "'29" ,fio.’ f .....-;......=. .. . ....te ?#’:W“T“‘I 1 f i “sink“? **-~b3 ‘\_______~__ _ _ / This is to certify that the dissertation entitled Coupled High Solids Fermentation and Anaerobic Filtration of Cellulosic Residues presented by Yew-Ming Lin has been accepted towards fulfillment of the requirements for Doctor of Philosophy degree in Sanitary Engineering ///fl£m M ajor professor Date November 10, 1983 MSU is an Affirmative Action/Equal Opportunity Institution 0-12771 MSU RETURNING MATERIALS: Place in book drop to remove this checkout from LIBRARIES “- your record. FINES will be charged if book is returned after the date stamped below. a"? .3 ‘3' l" '-~?"s.‘-':-‘ . “Hg-r 'a COUPLED HIGH SOLIDS FERMENTATION AND ANAEROBIC FILTRATION OF CELLULOSIC RESIDUES By Yow-Ming Lin A DISSERTATION Submitted to Michigan State University in partial fulfillment of the requirements for the degree of DOCTOR OF PHILOSOPHY Department of Civil and Sanitary Engineering 1983 ABSTRACT COUPLED HIGH SELIDS FERMENTATION AND ANAEROBIC FILTRATON OF CELLULOSIC RESIDUES By Yow-Ming Lin A coupled high solids fermentation and attached-growth anaerobic filtration process to produce methane from.cellulosic residues was developed and successfully operated for 18 months using wheat straw as the substrate. The process was conducted in eight 600 ml packed reac- tors and two anaerobic filters connected in series allowing semi-continuous feeding of straw at a solids concentration of 34%. A mobil liquid phase was circulated at a constant rate to carry COD from the packed reactors to the anaerobic filters where 85% of the total methane was generated. The major fUnctions of the packed reactors were the hydrolysis of the solid substrate and the production of organic acids. The volatile fatty acid COD, composed mostly by acetic, propionic, and butyric acids, was produced at a slower rate than soluble COD in the packed reactors. The initial soluble COD, 651 of the COD produced, was con- tributed by leaching while the subsequent slow substrate degradation was attributed to microbial hydrolysis. Specific methane production rates as high as 2.1 liter CH“ per day per liter reactor volune and a volatile fatty acid COD removal effi- ciency of 98% were obtained fran the anaerobic filters at loadings of 297 to 594 lb soluble GOD/day/lo3 ft3 with a hydraulic retention time Vow—Ming Lin of 34 hours. Methane contributed 73% to 79% of the total gas produc- tion. Total methane production per unit weight of substrate input was 1014.3 m1 CH4 per gram of straw input with 110 days substrate solids retention time and 76.3 ml CHu/g straw with 18 days solids retention time. At a 40 days solids retention time overall degradation of the un-pretreated straw was 30% with 43% and A1% degradation of cellulose and hemi-cellulose respectively. At an 18 days soids retention time degradations were 20% overall, 26% for cellulose and 31% for hemi-cellulose. During most of the study, a liquid reservoir served as an equali- zation basin preventing shock loading to the anaerobic filters. Results of a direct input study, without the liquid reservoir, suggest- ed that the methanogenic bacteria in the anaerobic filters could sustain methane production during transient loading, although the sud- denly increased substrate could not be completely utilized on the first pass. A mathematical model of solid substrate degradation in the packed reactor was developed. The curves computed from the model agreed closely with the experimental data. Biological hydrolysis of the um-pretreated wheat straw was found to be the rate limiting step in the system. DEDICATED TO MY BELOVED MOTHER AND TO THE MEMORY OF MY FATHER ii ACKNOWLEDGEMENTS To all the individuals who have helped and encouraged me during the course of this study, I offer my sincere thanks. Foremost appreciation is extended to my wife, Chu-Fen, for her understanding, patience and assistance during the last several years. Grateful thanks are given to my family members for their long time support and continuous encouragement. Special appreciation is extended to Dr. John A. Eastman, my major professor, for his assistance and valuable comments during the preparation of this dissertation. Thanks also to the members of my doctoral committee: Dr. Mackenize L. Davis, Dr. Harold L. Sadoff, and Dr. David A. Cornwall for their advice and suggestions. Thanks are also due to the Division of Engineering Research and the Department of Civil and Sanitary Engineering for their financial support which made this research possible. TABLE OF CONTENTS LIST 0? TABLE 0 O O O O O O O O O O O O O O O O O O 0 LIST OF FIGU m D O O O C O O O O O O O O O O O O O O 0 CHAPTER 1. 2. INTROWCI‘ION O O O C O O I O O O O O O O O O O O THEORETICAL BACKGROUND AND LITERATURE REVIEW. . 2.1 2.2 Chemical and Microbiological Background. . . . . .101 .102 .1 3 NNN 2.1." 2.1.5 2.1.6 Anaerobic Fermentation Process Stability . . . . . Metabolic Stages of Anaerobic Fermentation. . . Methanogenic Habitats . . . . . . . . . . . . . Production of Fatty Acids in Anaerobic Methane Femntation. O O O O O O O O O I O O O O O O 0 2.1.3.1 Physical and Chemical Properties of Fatty Acids. . . . . . . . . . . . . Methanogens and Methanogenesis. . . . . . . . Biochemical Pathway of Methane Formation. . . . . . . . . . . . . .5 Unique Properties 0f Methanogens . Role of Hydrogen in Anaerobic Fermentation. 2.1.5.1 Effects Of Hydrogen Concentration To The Fermentative Products . . . Role Of Nitrate And Sulfate In Anaerobic Methane Fermentation. . . . . . . . . . . . The Effects of Volatile Fatty Acids And pH. Ammonia Toxicity. . . . . . . . . . . . . . salt TOXicity O O O O O O O O O O O O O O 0 Heavy Metal Toxicity. . . . . . . . . . . . iv A 1 Physiology of Methane Bacteria . . . A 2 Methane Formation From Hydrogen And Carbon Dioxide . . . . . . . . . . . .A.3 Methane Production From Acetate. . . A A a Page viii 33 37 40 1:1 112 "3 an CHAPTER 2.3 Anaerobic Methane Fermentation Process Control . . 2.4. 2.5 Alkalinity, pH and Fatty Acid Effect of Tamperature . . . . Absence Of Toxic Material . . Effect Of Retention Time. . . . Effect Of Mixing. . . . . . . . Nutrient Requirenents . . . . . NNEUNNN wwwwww O ONU'Ib'UJN—A Anaerobic Fermentation Process Models. Conventional Anaerobic Digester Anaerobic Contact Reactor . . . Batch Load Model. . . . . . . . Plug Flow With Recycle. . . . . Anaerobic Expended Bed. . . . . High Solids Anaerobic Fermentation. Anaerobic Filtration Process. . . . Cellulosic Substrate . . . . . . . . . . . 2.5.1 2.5.2 Hydrolysis Of Cellulose Material. . 2.5.3 Properties Of Wheat Straw . . . . . EXPERIMENTAL METHODS AND MATERIALS. . . . . . . 3.1 Description Of The Experimental Apparatus. Concentration 3.2 Experimental Procedures. . . . . . . . . . . 3.3 3.2.1 3.2.2 System Start Up And First Stage 3.2.3 Second Stage Experiment . . . . 3.2.4 Third Stage Experiment. . . . . Methods Of Sample Analysis . . . . . . pH Measurement. . . . . . . . . Chemical Oxygen Demand. . . . . Individual Volatile Fatty Acids Gas Composition . . . . . . . . 303-501 Cellulose, Hanicellulose And Ligni n Detenmination Of Cell wall Substrate Preparation . . . . . . . . . Experiment. Physical And Chemical Properties Of Cellulose Constituent (Neutral Detergent Fiber). . . . . . . Determination Of Acid Detergent Fiber. Determination Of Acid Detergent Lignin Reagents For Fiber Analysis. . . . . . Page 46 A6 49 50 51 55 56 59 59 62 62 63 63 6A 65 68 68 70 72 73 73 8O 80 81 84 85 86 86 87 9O 92 99 99 101 102 10a CHAPTER 4. EXPERIMENTAL RESULTS AND DISCUSSION . . . . . . . . . 4.1 Performance of the High Solids Packed Reactors . 4.1.1 Total Soluble COD Production. . . . . . . 4.1.2 Volatile Fatty Acids Production . . . . . 4.1.3 Gas Production From The Packed Reactor. . 4.2 Role of the Liquid Equalization Reservoir. . . . 4.3 The Performance of the Anaerobic Filter. . . . . 4.3.1 Gas Production From the Anaerobic Filters 4.3.1.1 Stage I Gas Production . . . . . 4.3.1.2 Stage II Gas Production. . . . . M3 1. 3 Specific Methane Production. . . 4.3.2 COD Rm oval Efficiency. . . . . . . . . . . . . 4.3.2.1 Total Soluble COD Removal. . . . . . . 4. 3. 2.2 Volatile Fatty Acid COD Removal. . . . 4. 3. 2. 3 Summary of The COD Removal Efficiency. 4.3.3 Individual Volatile Fatty Acids in Anaerobic Filters 0 O O O O O O O O O O O O O O O O O 0 4.3.4 Estimation of Particulate COD Degradation in AnaerobicFilters.............. 4.4 Extent of Wheat Straw Degradation. . . . . . . . . . 4.4.1 Percent Substration Degradation Based on weight Loss 0 O O O O O O O O O O O I O O O 0 4.4.2 Percent Substrate Degradation Based on Fiber malYSj-s. O O O O O O O O O O O O O O O O O 0 4.4.3 Percent of Substrate Degradation Based on Mass Balance Calculation. . . . . . . . . . . 4.5 Summary of The Performance of Anaerobic Filtrs And The Extent of Substrate Degradation. . . . . . . . . 4.6 Production Rate of Nonbiodegradable COD. . . . . . . 4.7 Response of Anaerobic Filters to Transient Substrate Loading. 0 O O O O O O C O O O O O O O O O O O O O C MATHEMATICAL MODEL OF SOLID SUBSTRATE DEGRADATION IN PAcm REAflOR. O O O O I I O O O I O O O O O O O O O O O 5 O 1 “Odel Dev e1 oment O O O O O O O O O O O O O O O O O O O 5.2 Limitations and Discussion of The Application of MODEL vi Page 106 106 107 117 120 124 128 130 130 132 136 138 138 141 145 148 151 152 153 155 158 161 164 165 169 169 181 CHAPTER Page 6. ENGINEERINGAPPLICATION.................. 184 6.1 Configuration and Liquid Flow Pattern of The Packed Reactors. . . . . . . . . . . . . . . . . . . . 184 6.2 Comparsion of The Three Types of Reactor Systems . . . 185 2 1 The Need For A Liquid Reservoir . . . . . . . . 185 2 2 Extent of Solid Substrate Hydrolysis in The PackedReactors................186 .2.3 The Extent of Leaching. . . . . . . . . . . . . 187 2 4 Operation Simplicity. . . . . . . . . . . . . . 187 6.3 Ultimate Usage of The Treated Solid Substrate. . . . . 188 6.4 Operation of Packed Reactor at Short Solid Retenti on Tim 0 O O O O O O O O O O O O O O O C O O O 1 88 6.5 Conclusion of The Engineering Application. . . . . . . 189 7. CONCLUSIONS . . . . . . . . . . . . . . . . . . . . . . . . 190 8. SUGGESTIONS FOR FUTURE RESEARCH . . . . . . . . . . . . . . 195 APPENDICES A. List of Symbols . . . . . . . . . . . . . . . . . . . . . . 196 B. Mathematical Calculation for the Mass of COD produced by LeaChing O O O O O O O O O O O O O O O O O I O O O O O O 198 C. COD Convertion Factors. . . . . . . . . . . . . . . . . . . 200 D. Data for Figures. . . . . . . . . . . . . . . . . . . . .‘. 203 E. FORTRANProgramListings.................. 211 REFERENCES............................229 TABLE 1-1 2-1 2-2 2-3 2-4 2—5 2-6 2-7 2—8 2-9 2-10 2-1 1 2-12 2-13 2-14 2—15 2-16 3-1 3-2 LIST OF TABLES U. S. Food Crop Residue Generation . . . . . . . . . . . . Metabolic Function of HZ-producing Acetogenic Bacteria . . Physiological Characteristics of Four Groups of Bacteria . Physical and Chemical Properties of Volatile Fatty Acids . Methanogenic Species in Pure Culture . . . . . . . . . . . Transformations of Methanosarcina. . . . . . . . . . . . . Chemical Reactions For The Syntrophic Association of BaCteI'ia and HZ-UtIlIZIng ”Ethanogen o e e o e o e o o o 0 Effects of Hydrogen Partial Pressure on Free Energy Change Fermentation Of Cellulose By BL flayefacicus And RiflaiefacicusPlusmrminantim...n...nn The Oxidation-reduction Potentials Of Some Redox Pairs . . Concentration For Salt Toxicity (mg/1) . . . . . . . . . . Toxic Concentrations of Some Heavy Metals in Anaerobic Digesters. . . . . . . . . . . . . . . . . . . . Minimum Solids Retention Time For Anaerobic Methane Fementati on O I O O O O O O O O O O O l O O O O O I O O 0 Approximate Elementary Composition of Microbial Cells. . . General Physiological Functions Of The Principal Elements. Summary of Previous Anaerobic Filter Studies . . . . . . . Usage 0F Wheat Straw . . . . . . . . . . . . . . . . . . . Response Of Thermal Conductivity Detector To CH“ And C02 . H-BromideSolution.................... viii page 2 9 11 18 20 26 31 34 36 38 44 45 53 57 58 66 72 96 106 TABLE 4-1 u-3 4-4 u-s 4-6 4-8 u-9 4-10 14-11 11-12 11-13 4.1:: Individual Volatile Fatty Acid Concentration in the PaCRed ReaCtor. O O O O I O O I O O O 0 O O O O O O 0 Comparison of Methane Production in Packed Reactorsand in The Entire System (Stage I). . . . . . . . . . . . Composition of Gas Produced in Stage II Anaerobic Filters Influent and Effluent COD Concentrations for The Anaerobic Filters (Stage II). . . . . . . . . . . . . COD Removal Efficiency in the Anaerobic Filters(Stage Volatile Fatty Acid COD Removal Efficiency in Stage I AnaerObic Filters 0 0 O O O O O O O O O O O O O O O 0 Individual Volatile Fatty Acid Concentrations in Filter No.1 Effluent. (Stage II). . . . . . . . . . . Individual Volatile Fatty Acid Concentrations in Filter No.2 Effluent. (Stage II). . . . . . . . . . . Straw Weight Loss After Fermentation (Stage I). . . . Straw Weight Loss After Fermentation. . . . . . . . . II) Cellulose, Hemi-cellulose, Lignin Contents of Wheat Straw. Adjustment of Straw Fiber Contents (Stage I) . . . . . . . Adjustment of Straw Fibrous Contents (Stage II). . . . . . Sumnary of The Performance of Anaerobic Filters and The Extent of Substrate Degradation. . . . . . . . . . . . . . Comparision of The Three Types of Reactor Systems. . . . . ix page 117 12a 1311 145 146 146 148 149 154 154 156 157 158 163 189 FIGURE 2-1 2—1 2-3 2—4 2-5 2-6 2-7 2-8 2-9 2-10 2-11 2-12 2-13 3-2 3-3 3-4 LIST OF FIGURES Three stage scheme of anaerobic methane fermentation. . . Type B habitat. Rumen, caecum, instestine . . . . . . . . TypeChabitat.Thermalspring.............. Pathways in the rumen fermentation of the major insoluble carbohydrates present in plants . . . . . . . . . . . . . Some major end products of the microbial fermentations of sugarsfrompyruvicacid. . . . . . . . . . . . . . . . . Barker's schene of methane formation. . . . . . . . . . . Modification of Barker's schene for 002 reduction to ”wane I I I I I I I I I I I I I I I I I I I I I I I I I Effect of hydrogen partial pressure on the free energy change for the oxidation of ethanol, propionate, and butyrateI I I I I I I I I I I I I I I I I I I I I I I I I Fermentation interactions between W Wu: and Hz—utilizing methanogens. . . . . . . . . Electron tower for 02/820, N03/Nog, SOZ/HZS, and COZ/CI-Iuredoxpairs.0.0000000000000000 Schenatic diagrams of anaerobic fermentation processes. . Conformation formula of cellulose . . . . . . . . . . . . Cellulose molecule with interaction hydrogen bonds. . . . Schenatic diagram of experimental systen. . . . . . . . . Schematic diagram of packed reactor and gas collection systh I I I I I I I I I I I I I I I I I I I I I I I I I Experimental systen for high solids fermentation and anaerobic filtration of wheat straw . . . . . . . . . . . ReactorsystenatStateI................ Page 8 12 13 15 17 29 32 35 39 60 69 69 74 75 79 83 FIGURE Page 3—5 Reactor system at Stage II and Stage III. . . . . . . . . 83 3—6 Calibration curve for colorimetric COD test . . . . . . . 89 3-7 Volatile fatty acids standard curves. . . . . . . . . . . 93 3—8 Chromatograms of fatty acids standard . . . . . . . . . . 94 3-9 Chromatograns of volatile fatty acids . . . . . . . . . . 95 3-10 Chromatograms of gases from No.1 and No.2 anaerobic filters I I I I I I I I I I I I I I I I I I I I I I I I I 97 3-11 Thermal conductivity detector response to methane and carbon diOXideI I I I I I I I I I I I I I I I I I I I I I 98 4-1 Total soluble COD and volatile acid COD of effluent from paCRedreactOE‘SQOOOOooooooooooooooeo 108 4-2 Total soluble COD of effluent liquid from packed reactors 110 4-3 Average total soluble COD of effluent from packed reactorsaftermak0000.000.000.000...111 4-4 Straw leaching test in a packed reactor . . . . . . . . . 113 4-5 Straw leaching test in a batch reactor. . . . . . . . . . 114 4-6 Total soluble COD production from packed reactor. . . . . 116 4-7 Individual volatile acid concentration in one packed reactor I I I I I I I I I I I I I I I I I I I I I I I I I 1 18 4-8 Gas composition in Stage I packed reactors. . . . . . . . 121 4-9 Cumulative gas production from a single packed reactor. . 122 4-10 Individual volatile fatty acid concentration in liquid reserVOir I I I I I I I I I I I I I I I I I I I I I I I I 126 4-11 Effluent total soluble COD and volatile fatty acid DOD of liquid reservoir . . . . . . . . . . . . . . . . . . . 127 4-12 Methane gas composition of Filters No.1 and No.2, and the pH change of anaerobic filters, liquid reservoir and the oldest packed reactor in Stage I. . . . . . . . . . . . . 129 4-13 Total gas production fren Filter No.1 and Filter No.2 . . 133 4-14 Cunulative methane and carbon dioxide production fran anaerObic filters I I I I I I I I I I I I I I I I I I I I 135 xi mum-z 11-15 4-16 11-17 4-18 11-19 4-20 Page Total soluble COD of effluent from the liquid reservoir and the anaerobic filters . . . . . . . . . . . . . . . . 139 Effluent volatile fatty COD of the liquid reservoir and anaerobic filters. (Stage II) . . . . . . . . . . . . . . 142 Volatile fatty acid COD of effluent from liquid reservoir and anaerobic filters. (Stage I). . . . . . . . . . . . . 144 Individual volatile fatty acid concentration in Filter No.1 and Filter No.2 . . . . . . . . . . . . . . . 150 Total soluble COD, before (Fig. A) amd after (Fig. B) direct substrate input into anaerobic filters . . . . . . 166 Inidividual volatile fatty acid concentration of anaerobic filter effluent . . . . . . . . . . . . . . . . 168 Schematic diagram of a packed reactor divided into five sementSI I I I I I I I I I I I I I I I I I I I I I I I I 170 Experimental effluent soluble COD concentration from the newest packed reactor . . . . . . . . . . . . . . . . . . 178 Effluent soluble DOD concentration of packed reactors obtained from MODEL and from experimental measurements. . 180 Estimated soluble COD concentrations of packed reactors by MODEL at various segment numbers . . . . . . . . . . . 182 Schematic diagram of parallel connected reactor system. . . 184 xii CHAPTER ONE INTRODUCTION Anaerobic digestion (anaerobic methane fermentation) is a common and successfully used process in wastewater treatment. Methane and carbon dioxide are the final end products of this process resulting from microbial decomposition of organic matter in the absence of molec- ular oxygen. Although methane gas is a useful er1 that some treatment plants recover and use for heating the digesters or driving pumps, anaerobic digestion has been mainly used for sludge and wastewater sta- bilization; the methane gas has usually been treated as a by-product. However, in recent years, the increasing interest in alternative energy sources has altered the traditional role of anaerobic digestion. In addition to waste stabilization, anaerobic digestion has been consi- dered as an energy production process and organic waste as the alternative energy source. Among organic wastes, cellulosic agricultural residues are most abundant. It is estimated that, considering only food crop residues, over 290 million tons are produced every year in the United State (Ben- son, 1977). Table 1-1 shows the quantity of some agricultural residues produced in the United States. Most of these cellulosic residues exist in dry form or at very high solids concentration. In order to utilize these natural products to produce methane by anaerobic fermentation, a new process needs to be developed to overcome difficulties of mixing and punping slurries with high solids content in the reactor and asso- ciated piping as well as the dificulties of substrate input to and 2 removal from the reactor. The current anaerobic fermentation processes require relatively low solids concentration for operation. Sawyer (1960) suggested that the optimum solids concentration in the conven- tional digester should not exceed six to eight percent. If conventional digesters were to be used for the fermentation of high solids cellulosic material, a large quantity of water would have to be added to reduce the solids concentration resulting in a greater reactor volume and larger residual sludge volume. This would make the process uneconomical. Table 1-1. U. S. Food Crop Residue Generation Crop 1O6 Tons/Year Corn 156 Wheat ug Soy Bean 40 Sorghun Grain 24 Oats 12 Barley 11 Rice Straw 4.4 Peanut 1.6 Rye 0.9 * After Jewell (1980). 3 This research investigates a new process that consists of fixed-film anaerobic filters and packed bed reactors which contain a high solids concentration of stationary phase and a mobil liquid phase. The special configuration of the packed bed reactors enable this pro- cess to have semi-continuous input of high solids cellulosic substrate. The anaerobic filters convert the fatty acids produced in the packed reactors to methane and carbon dioxide after being transported by the mobil liquid phase. Because the required reactor volume is inversely proportion to solids concentration, increasing the substrate concentra- tion would allow a reduction in the total system volume and system cost, making this process economically competitive. Detailed presenta- tion of this process and the experimental system used to conduct the process will be given in the following chapters. The objectives of this research were to: 1. Design an anaerobic fermentation systen which would allow semi-continuous feeding of high solids substrate without caus- ing retardation of methane production. 2. Evaluate the performance of the high solids packed reactor and to investigate the behavior of cellulosic substrate degrada- tion, total soluble COD production and volatile fatty acid production. 3. Evaluate the performance of the anaerobic filters including methane production and COD renoval efficiency. 4. Evaluate the operational parameters, include liquid flow rate, hydraulic retention time, substrate input interval, and the percent of substrate solids concentration. Determine the degradation rate of a cellulosic substrate without pretreatment. Develop a mathematical model for the liquid soluble COD pro- duction by solid substrate in the packed reactor. CHAPTER TWO THEORETICAL BACKGROUND AND LITERATURE REVIEW Methane fermentation is a complex biological process in which a mixed culture of microorganisms decomposes organic matter to gaseous end products, methane and carbon dioxide, in the absence of exogenous electron acceptors other than carbon dioxide (McInerney et al., 1981a). One distinct characteristic of this process is that only a small por- tion of the chemical energy from the decomposition of organic substrate is used for bacterial cell growth and about 90% of the energy is retained in the methane produced. The advantages of low biological growth and the production of an energy rich gas have made anaerobic digestion a favorite treatment method for sludges and strong organic wastes for a long time. The mechanism of anaerobic fermentation was not clearly understood until the 1950's, although it had been successfully operated for many years. After Barker (1956) and Buswell et a1. (1952), reported their studies of methane fermentation, extensive research was conducted by many investigators. Those studies have provided a better understanding of the complicated anaerobic fermentation process. This chapter will review the literature regarding important chemical, microbiological, engineering process control prameters, and other related concepts of methane fermentation as well as recent developments of this process. 2-1 ChanieaLanLUiembielcgicaLBackgmund Anaerobic fermentation is carried out mainly by a diverse group of bacteria. The microbial population in the anaerobic fermentation ecosystem is composed of obligate and facultative anaerobic bacteria. Obligate anaerobes (aerophobic anaerobes) can only survive in the strict anaerobic condition, while the facultative anaerobes (aeroto— lerent anaerobes) can also use molecular oxygen during metabolism. According to current knowledge, the bacterial population involved in anaerobic fermentation can be classified into four different groups, namely, (1) fermentative bacteria, (2) obiligate hydrogen producing acetogenic bacteria, (3) methanogenic bacteria, and (4) homoacetogenic bacteria. The following paragraphs will discuss the microbiological functions of these bacteria and chemical reactions that exist in the system during microbial decomposition of organic substances. Traditionally, methane fermentation has been considered to have only two metabolic stages, an acid forming stage and a methane forming stage. In the first stage, a complex of fermentative acid forming bac- teria degrade high molecular weight organic compounds such as polysaccharides and protein to volatile fatty acids, hydrogen, carbon dioxide, ammonia and sulfide. The second, or methane forming stage, involves a complex group of strict anaerobic methane bacteria. These methane bacteria convert the products from the first stage to methane and carbon dioxide. Recently, a three stage scheme, first proposed by Bryant et al. (1967), has become widely accepted (McCarty, 1981) and further expanded by microbiologists (Bryant, 1976, 1979; Kasper and WUhrmann, 1978; Boone and Bryant, 1980; McInerney et al., 1981a, 1981b). The estab- lishment of the three stage scheme (Figure 2-1) was based on the finding that fatty acids other than formate and acetate are degraded by syntrophic association of hydrogen-producing acetogenic bacteria and hydrogen—utilizing methanogens, and not by methanogens alone. The first stage of the three-stage scheme is the same as in the two stage scheme model; the fermentative bacteria hydrolize polysac- charides to smaller organic sugars and degrade these products to fatty acids, alcohols, hydrogen, and carbon dioxide. The second stage involves hydrogen-producing acetogenic bacteria which are involved in (1) -oxidation of fatty acids of even numbered carbon to acetate and hydrogen and odd-nunbered fatty acids to acetate, propionate, hydrogen; (2) oxidation of alcohols such as ethanol to acetate and hydrogen; and (3) decarboxylation of propionate to acetate, hydrogen and carbon diox- ide (McInerney et al., 1981b). The chemical reactions for the conversion of longer-chain fatty acids to acetic acid by acetogenic bacteria are shown in Equations 2-1 to 2-4 in the Table 2-1. Evidence of the second metabolic stage in anaerobic methane fermentation was obtained by the successful isolation of two fatty acid oxidizing acetogenic bacteria, Syntrgpngmgnas.uglfeii and Syntrophobact r wolinii, via coculture with hydrogen-utilizing bac- teria. The microbiological characteristics of these two bacteria have also been studied (McInerney et al., 1981b; Boone and Bryant, 1980). Polysaccharides (I) Hydrolysis and Fermentation (I) (II ) Organic Acids (II). 1 1 H2, C02, Formate (IV) Acetate (III) (III) CH4, C02 Figure 2-1 Three stage scheme of anaerobic methane fermentation. Involves four groups of microorganisms, (I) fermentative bacteria; (11) obiligate H -production acetogenic bacteria; (III) methane production bacte ia; (IV) homoacetogenic bacteria. Table 2-1 Metabolic Function Of H2-producing Acetogenic Bacteria A. B - oxidation of longer-chain fatty acids a. Even-nunbered carbon to acetate and hydrogen CH3CH2CH2COO' + 2 H20 .—= 2 CH3COO" + 2 H2 + H" butyrate AG:D -.- + 11.5 Keel/reaction b. Odd-nunbered carbon to acetate, propionate, and hydrogen valerate + 2 H2 + H+ AG; : + 11.5 Keel/reaction B. Decarboxylation of propionate to acetate, C02, H2 propionate AG; = + 18.2 Kcal/reaction C. Oxidation of alcohols to acetate and hydrogen ethonal AG:3 -.- + 2.3 Keel/reaction (2-1) (2-2) (2-3) (2-4) In the third, or terminal, stage of methane fermentation, methano- genic bacteria split acetate to methane and carbon dioxide, and use hydrogen to reduce carbon dioxide to methane. An additional metabolic group, homoacetogenic bacteria, which is capable of oxidizing hydrogen anaerobicly with the reduction of carbon dioxide to acetate, was discovered in an anaerobic fermentation ecosys- ten (Zeikus, 1979; Wolfe and Higgins, 1979). So far, Acetobactenim noodii is the only physiologically well characterized 10 hydrogen-consuning homoacetogenic bacteriun (Balch et a1. , 1979). Table 2-2 taken from Zeikus (1979) shows the physiological charac— teristics of four groups of bacteria that are often isolated frcm anaerobic sludge digesters and have been discussed in the previous par- agraphs. In a mixed culture anaerobic fermentation ecosystem, the effective metabolism of one group of bacteria is closely related to interaction with other groups of bacteria. Therefore, their metabolism may not be separated into distinct steps for metabolic optimization (McIncrney and Byrant, 1981a; Zeikus, 1979). 2.1.2 MemancgeniLflabitats Several different types of methanogenic habitats can be feund in the ecosystem, and they may be classified into three types (Welfe and Higgins, 1979). Type A habitat shown in Figure 2-1, includes aquatic sediments, anaerobic sludge digesters and marsh; it is a complete anaerobic fermentation system which involves all four groups of bacter- ia. Animal tracts such as rumen and caecum are Type B habitats (Figure 2-2). In a Type B habitat, only fermentative bacteria (Group I) and hydrogen-utilizing bacteria (Group III) are involved in the ecosystem. Fatty acids are the major end products, and longer-chain acids are not converted to acetic acid but are absorbed into the blood- stream where they serve as the major energy source for the ruminant. Another special methanogenic habitat, Type C (Figure 2-3), only involves methanogens that utilize acetate, hydrogen and carbon dioxide present in the system to produce methane. Some thermal springs in the 11 Noe .ezu em A Iooumxu - N m N e NH oH m I” INN maempoen weseeee ou . :0 .IO Io . OU\ I eceemoeeeeez eeeeeeeeeeewez N e N N mwcmuoen seaweaoewowseoeeeew ou . :0 e-N 00\ I ewemmoeeesez eeeeeeeeeeeeeesz .e.eee< em Noe\NI . ewemuoen wweeoe .uwamo< m owuoem .mmouuacl owcmmoumomoso: eseeewoeoeeeu< .uanee< ..NI. - .eeeeam .owpmo< ewempuen omcmmoumoe .Pocmeum .Nou\N: - wuw>=e>a mcwoacoca-N: .54~:emeo m. .omuoen .owumo< N mmowno_Pmu ewcmuuen eswwsu°eeoew .Foeeesm .NOU\N: A eme_=__eu e>anmseeeLel eeeeeeweeee muoauoea Aczv wave caecumnam :owpeucmsgml mow—3:00 caponeumo azocw owponmomz smwcemco mwcooomm Co mazoeu czol mo momumwgmuomcmzu —mowmopowmxea NaN mpame 12 Polysaccharides (I) Acetate Lactate Formate Propionate Succinate CO2 Butyrate H2 Valerate (III) 1 Absorbed into Intermediates CH4 + C02_] bloodstream in propionate * synthesis Fl'Uure 2-2 Type 8 Habitat. Rumen, caecum, intestine. 13 H2, coz l—- (111) Acetate (111) (III) (mesophilic) (thermophilic) (III) CH4 + C02 CH4 CH4 Figure 2-3 Type C Habitat. Thermal Spring. 14 Yellow Stone National Park and Lake Kivu (in Africa) (Wolfe and Hig- gins, 1979) have been found belonging to this type of habitat. Thermal springs contain hydrogen, carbon dioxide, sulfide, and mineral salts. Some methanogens have been isolated from several thermal springs in Yellow Stone Park. In Lake Kivu, methanogens reduce carbon dioxide by using volcanic hydrogen to produce methane (Deuser et al., 1973). Among all the methanogenic habitats mentioned above, the rumen is the most studied anaerobic fermentation ecosystem (Hungate, 1975; Hobson, 1971, 1974, 1982; Prins and Clarke, 1979). 2.1.3 .ELQduQIiQn_Qfl_EaLC1_AQidS_in_AnaacQDiQ_M§Lhan§_E£cm§nLaLion Organic substrates that are subject to anaerobic degradation are mostly carbohydrate, protein, and lipids. Agricultural residues con- tain mainly polysaccharides, such as cellulose, hemi-cellulose, and pectin. As described in the Section 2.2.1, these materials are first hydrolized by extracellular enzymes secreted by fermentative bacteria into lower molecular weight compounds, such as monosaccharides (glu- cose, fructose, xylose), and oligosaccharides (sucrose, cellobiose, short-chain fructosans). These smaller organic compounds can be tran- sported through the fermentative bacterial cell wall and further degraded to fatty acids and other organic acids, alcohols, hydrogen, and carbon dioxide. Figure 2-4 shows the pathway of hydrolysis and fermentation of cellulose, hemi-cellulose, and pectin to monosaccharides and to pyru- vate via the Embden-Meyerhof-Parnas pathway. Pyruvate is the key intermediate product in the first metabolic stage of anaerobic fermen- 15 Hemicellulose l ‘ 7 ** 1M1 Maltose Cellobiose Pectic Acid Xylobiose l J l l Glucose Glucose Galacturonic -—————-—Xylose ( Acid 1 Glucose 1 P Xylose P l -Glucose 6 P- Pentose PhOSphate Pathway l Fructose 6 P Fructose 6 P l Embden Meyerhof Parnas Pathway [Pyruvate J F1'Slure 2-4 Pathways in the rumen fermentation of the major insoluble carbohydrates present in plants. ** Pectic substances consist of D-galacturonic acid and its methyl ester, D-galactose and L-arabionse. The details of its structure are not completely known, but the major part consists of (1-4)-linked c-D-galacturonic acid residues (Wood, 1970). 16 tation, because most of the important organic acids produced in this stage are obtained from pyruvate. Figure 2-5 gives the identified pathways of microbial fermentation of organic acids from pyruvic acid in pure culture. From Figure 2-5, it can be seen that a wide variety of acids and alcohols can be produced from pyruvic acid depending on the particular type of microorganism involved. In mixed cultures, such as Type A ecosystem, the major fermentation products in the first meta- bolic stage are acetic acid, propionic acid, butyric acid, hydrogen, and carbon dioxide. Certain fermentation products of monosaccharides may involve reaction of the Pentose Phosphate Pathway and Entner-Doudoroff Pathway (Lehninger, 1975). In a normal fermentation system, acetic acid is the predominant acid while in a stressed system, propionic and butyric acids may have higher concentrations (Hobson et al., 1981). 2.1.3.1 iEhisical_and_Cbemieal_Ereperties_of_Eattx_Acids Saturated fatty acids are single lipids with the general formula, IV 0 CH3 - (CH2)n - coon n (2-5) The terminal carboxyl group of the fatty acid is very hydrophilic and the hydrogen carbon chain, constructed from two identical carbon monomers, is almost insoluble in water. The hydrophilic-hydrophobic character gives fatty acids a polar carboxyl head and a non-polar hydrocarbon tail. Table 2-3 gives the physical and chenical charac- teristics of some major fatty acids commonly found in the complete 17 Polysaccharides 1 Sugars ’ -co +2H ‘ Lactate H Pyruvate HAcetaldehyde ——-|Tithanol:l ’///<fi;o +co2 Acrylate H2 + CD2 oL-Acetolactic Acid , l-coz Oxalacetate [Formate ] Acetyl-S-CoA Acetoin +2H 5 +4H 1+2” Malate [H2 + COZJ l2,3,-Butanediol] +2H -H20 ATP Fumarate AcetateJ Acetoacety::::CoA\\\\\\E::anolJ [Succin:te_j [AEET6:::j//‘fa/////H Butyryl- -S- CoA l.{Propionate J [Iso-propanol] [Butyric Acid] [Butanol ] Figure 2-5 Some major end products of the microbial fermentations of sugars from pyruvic acid. 2H represents two hydrogen atoms being donated in a reductive step. Reduced form products have more hydrogens an electrons per carbon atom. (after Lynch, 1979) l8 mwnga use acumweogu mo xooaocm: ”moczom topaz PE oo~ Lon seem eow.~ e.o NNm.o o.~o~ ea.eaa Icoeeimzuvmze eae< slogans e eom.~ speemapm mmm.o N.NON ea.eaa IOOUNANIQVIUNAMIUV ease eaeeaeu-0mi e hmo.N ee.m mma.o o.ewa mH.NoH Iooumimzevmxu ease ocea2e> m emo.~ N.e comm.o o.eea mH.~oH Iooumxezumimzev ece< sweepe>-omi m eam.a Ne.m Nemm.o mm.mea Na.mw Iooomzumzemze eaua aceseem e mam.a o.o~ Hwem.o N.mmH Ha.ew Ioouzemimzev nice ewesesm-omi e Nam.a Axe omam.o mm.oea mo.eg :ooumzemzu ease uieocaoea m eeo.a has Nmeo.a a.eaa mo.oe :ooemze ecu< scarce N eeem.o has oNN.H N.ooH mo.ee zoos: ease scarce a N . sea>eee e .neaoa castes e\aoe a ean_um eccauaam m=a_iom .2.z eczee0l maez to .02 meae< Nasal epise_o> co mm_sieeaoea emanates nee _eecmsea m-~ a_eec 19 anaerobic methane fermentation systen. 2.1.4 We Methane bacteria are a unique group of microorganisms involved in the terminal stage of anaerobic fermentation to produce methane. This section will review some of the current knowledge of this special group of bacteria and the mechanism of methane formation. 2.1.4.1 WWW Although methanogens are a morphologically diverse group of bac- teria, varying from short, lancet—shaped cocci to long, filamentous rods, they share some common physiological properties that are not found in any other group of bacteria. They require strict anaerobic conditions and a very low redox potential (-330mv) for growth (Zehnder, 1978). Studies of methane bacterial cells, both gram-negative and gram-positive types have failed to find muramic acid and peptidoglycan which are present in all other bacterial cells. Fox et a1. (1977) found that both transfer RNA and ribosomal RNA oligonucleotide sequences of methanogens are different from typical bacteria. They also indicated that methanogens are one of the most ancient groups of organisms (Fox et al., 1977; Balch et al., 1979). Woese (1978, 1981) declared that methanogens are neither prokaryotes nor eukaryotes but are the largest group of archebacteria. Balch et a1. (1979) presented a new taxonomic scheme for the methanogens based on the relationship of oligonucleotide sequences of the 16 S ribobomal RNA. Table 2-4 gives the new taxonomic scheme as well as some characteristics of methanogens 20 a2_eaeee> .z :oou: .NOU\N: we «gem - m_wpoe maouoo mpepo> mewEemoozpm come» sow: :oouz .No0\N: moccansm camooea . opauoe mzcuou wmpmwcce> mauoouocmnumz Eappommpc opacwm now: 1000: .NOU\N: camczeoczmmn + mcoc anpcecwe:e .2 we usem Fwequm e:_FmamFm Nou\N: :wmeaeouamma + one: mpmcwm mmcog neocm mapwnawcoaee Eappmmepe o: ”mace ueocm co Ioou: .NOU\N= :wwc320ezmma + one: voooo mumaeemuomucep Ezwpcecw52c couoenmmeaocecuoz NOU\N: camc35oe=wma + one: .l .2 we maem e:o_;aoeaop:eoEcm;u Noo\N: camezsoeamma + mcoc .l .2 me mEem FenceALD mangmrvc oneseoc .Nou\N: cwmczeoezmmg + one: ”no; um>c=u op uemwepm E=UVUwELom sawemuueaocezumz weeeemasm Fae: Pies cocauwwm secpas_oz smopoeaeoz meaeaam mezupau mesa cw mmwumam owcmmocegumz euN open» 21 mpepmoe cease Panama mewcecooem mgmxoea Iomxu .Nou\N: -zroaocmom; + mcoc cw mzccoo cm—zmmecw wgmxemn ecpocemocmnpcz :pmmgm pmccmuxm sow: moccEmpwe ”Eapchwgm zoom: .NOU\N: muwcznam :quoca . mpwuoe ”moo; cm>e=u .mcop wmuomczz Ezppwenmocesomz Ioouz .Nou\N: muwcaazm campoca - mpwuoe mzccoo empzmceew wemmcmwems :oou: .Noo\N: muwcanzm :mmuoca . mpwpoe mppmo vacuuou cepzmmccw Fumwcmu Ezwcmmocmspmz :oouz .NOU\N: muwzcnsm cwmuoca . mpwpoe to; cm>e=o .ocogm mpmnos Eamnocuweocmcuwz acetamaem Fae: Fame caecumww sciences: smopoeaeez meceeam umzcwucoo eiN epoch 22 that have been isolated in pure culture. From Table 2-4, it can be seen that methane bacteria are diverse in morphology. Some species are motile while others are not, and it is interesting to note that nonmotile species are gramepositive and motile species are gran-negative. All species share the common metabolic capacity to produce methane from hydrogen and carbon dioxide. Several species can utilize fermate but only one is able to use acetate as a substrate. Most methanogens are most active in the temperature range fren 33°C to 45°C. Nethenobacterimjhermoautctmphiem is the only known thermophilic methane bacteriun with an optimun tenperature of 65°C to 70°C (Zehnder, 1979). Methane bacteria are very sensitive to pH changes, growing best in the pH range from 6.5 to 7.7 (Smith and Hun- gate, 1958), with the optimum being 7.05 to 7.20 (Harmeer and Borchardt, 1969). It has also been found that important hydrogen-producing and hydrogen consuming anaerobes do not grow at pH values below 6.0 (weimer and Zeikus, 1977). Although some non-methanogenic anaerobes can grow at lower pH values (Cohen et al., 1979; Eastman, 1981), even as low as 2.0 (Canale-Parola, 1970), no methanogens can grow well at pH values less than 6.0 or above 8.0. Zeikus (1979) indicated that the inhibition of hydrogen oxidizing methanogens by high proton concentration may be related to thermodynam- ic regulation; at lower pH conditions, proton reduction to hydrogen become the thermodynamically favored process rather than the normal oxidation of hydrogen to proton. All methanogenic bacteria contain several types of special coen- zymes (Bryant, 1979), such as coenzyme M, coenzyme 420, and coenzyme 23 factor B. Coenzyme M is a methyl carrier and participates in methano— genesis from. methanol and acetate (Smith and Mah, 1978; Taylor and Welfe, 1974). Coenzyme 420 is involves in the electron transfer and serves as an electron carrier similar to the function of ferredoxin (Tzeng et al., 1975). Coenzyme factor 420 is a low molecular weight, heat stable cofactor and is believed to be involved in the enzymatic formation of methane from methyl coenzyme M (Gunsalus and Wolfe, 1976). Two other cofactors; F430 and F342 were also discovered but their fUnc- tions are still not known (Gunsalus and Wolf, 1978). 2.1.4.2 WWW As indicated in Table 2-4, the only known methanogenic substrates are H2/C02, fermate, methylamine, and acetate. In spite of different morphologies, all known pure cultures of methane bacteria can use hydrogen as the electron source to reduce carbon dioxide, to methane according to Equation 2-6. un2+ncog+n+==cau+3azo (2-6) 250; = -32.7 Kcal/reaction Equation 2-6 shows that eight electrons, derived from four moles of hydrogen, were used to reduce one mole of carbon dioxide. The carbon dioxide utilized by methane bacteria is partly reduced to methane and partly metabolized and fixed for cell material. This character is dif- ferent from other autctrophs that just use carbon dioxide as the single carbon source for growth. It is also noted that the standard free energy change of Equation 2-6 is very negative, indicating that methane 24 bacteria have a very strong affinity for hydrogen gas. The special ability of methanogens to consume hydrogen is the major factor main- taining very low hydrogen concentration in anaerobic environment. Hungate (1970) reported that the Km for the utilization of hydrogen in -6 the runen was only 10 M. 2.1.4.3 WW It has been reported that about 70 - 73 i of methane produced is from the decarboxylation of acetate (Jeris and McCarty, 1965; Smith and Mah, 1966). A few species of bacteria have been reported to utilize acetate as an energy source and to produce methane (Barker, 1936; Bryant, 1974). However, up to the present time, only one species of acetate utilizing methanogen, Methanosarcina.barkezi, has been isolated in pure culture (see Table 2-4). Stadtman and Barker (1949) performed a series of experiments by using 1uC-methyl or 1i'C-carboxyl labeled acetate and formate and reported that methane was derived from the methyl group of acetate and carbon dioxide was derived from the carboxyl group of acetate as Equa- tion 2-7 shows: 0 - o - ”CH3 000 + H20 ..——..—. 'CHL, + H CO3 (2-7) AG:D = —7.4 Kcal/mole The hydrogen atoms on the methane were obtained fren the methyl group of acetate plus the fourth hydrogen atom contributed by water (Pine and Barker, 1956). These observations cleared up the earlier controversy 25 of methane formation from acetate, i.e., whether acetate is completely oxidized to hydrogen and carbon dioxide and then the carbon dioxide is reduced with hydrogen to methane versus the direct conversion of the methyl group of acetate to methane (Smith and Mah, 1980). In addition to acetate, Methanesarcina barkeri can also utilize CH3OH, CH3NH2 and H2/C02 as the substrate to produce methane. The chemical reactions as well as the standard free energy changes for these reactions are shown in Table 2-5. Weimer and Zeikus (1979) reported that M; ,barkeri grow about four times faster on H2/C02 or methanol than on acetate. Equation 2-7 shows that the standard free energy change is a small negative value (-7.4 Kcal/mole) that is insufficient to produce one ATP, since values for the free energy change of ATP hydrolysis have been estimated to range fran -8.5 Keel/mole to -12.5 Kcal/mole (Decker et al., 1970). Normal efficiency of energy transfer in bacteria are 30% to 70% (McCarty, 1975; Decker et al., 1970), and, therefore, mini- mun energy required to produce one ATP would be 11.1 Kcal/mol. As mentioned earlier, the growth rate of Methanosarcina on acetate is very slow (doubling time greater than 24 hours). and its growth yield is only 1.6 - 3.0 mg dry weight/m. mol CH4 (Smith and Mah, 1978). This slow growth rate may be related to the small energy yield. However the actual mechanisms of methane production from acetate and its energy production mechanisms are not fully understood today. 26 Table 2-5 Transformations of Methanosarcina Equations AG; Kcal/reaction A.mmmd 1. u cn3on === 3 can + H00; + 8* + H20 - 75.2 2. u cn3ou + cn3coo- ==: 8 can + 2 H00; + 8* - 82.6 3. ca3on + H2 === can + H20 - 26.9 B. Methylamine 1. u CH3NH§ + 3 H20 === 3 can + H00; + 4 NHK + 8* — 53.8 2. 2(CH3)2NH§ + 3 H20 === 3 can + ace; + 2 NHfi + 8* - 52.5 3. 4(CH3)3NH"' + 9 1120 == 9 CH4 + 3 1100; + 11 NH; + 3 1r -159.8 C. Carbon Dioxide 1. 4 H2 + 8* + 8003 === CH4 + 3 H20 - 32.4 D. Acetate 1. CH3COO' «1» H20 = CH“ + HCO§ - 7.4 * Source: Smith and Mah, 1980 2.1.4.4 .Bicchemieal_Eatbwax_cf_Eethane_Eonmatien Barker (1956) first introduced a schene to explain the possible pathways of carbon in methane formation from various sources as shown in Figure 2-6. Barker's schene suggests that an unidentified one—carbon carrier, R, is bound to various substrates which are reduced to methane with the regeneration of carrier. McBride and Wolfe (1971) discovered coenzyme M (HS-CHZ-CHZ-SO3H, 2-mercaptoethan sulfonic acid, 27 C02 + RH -* RCO0H 2 H H 0 R-CHO 2 H R-CH 0H 2 H H20 H20 3 7\ : R-CHB 7— 12113011 + RH Hco‘ HO 3 2H/\ 2 CH4 + RH Figure 2-6 Barker's scheme of methane formation. 28 abbreviated as HS-CoM) which was later shown to be one of the unknown carriers in Barker's scheme. It is now widely accepted that HS-CoM is a necessary methyl carrier for methyl transfer in methanogesesis. Barker's scheme was modified by Wolfe (1979) to a cyclic pathway (Figure 2-7) by the findings that the methyl-reductase reaction was coupled to the activation and reduction of carbon dioxide, and that an intermediate, involved in the primary step of C02 activation, is gen— erated from the terminal reaction in Barker's scheme, H n+2 ATP 2’ g ’ ~ cm, + H-S—CoM (2-8) CHB-S-COM methylreductase 2.1.4.5 WW Reviewing the material given above, some of the unique properties found in the methane bacteria can be summarized as follows: 1. Methane is the common metabolic product for all methanogens and only for methanogens. 2. Methanogens can use only a narrow range of substrates; hydrogen/carbon dioxide, fermate, methanol, methylamine, and Acetate. 3. Methanogens require strict anaerobic conditions and an extreme low redox potential (-330 mV) for growth. 4. Methanogens contain unique coenzymes and cofactors: CoM, F420, F430, F342. 5. The 16 S rRNA sequences are unique. 6. Cell walls contain no D-anino acids or muramic acid. 7. Cytochromes (electron transferring proteins containing an 29 HS - CoM CH4 [ 7 ] 02 ATP 2H Mg+2 CH3 -S- COM XCOOH (unknown) 2H H20 HOCH2 -5- COM XCHO (unknown) 2H Figure 2-7 Modification of Barker's scheme for C0 reduction to methane to emphasize a cycle where the unknown activated- intermediate produced by the methylreductase is involved in 002 activation. (after Wolfe, 1979) 30 iron-porphyrin group) and quinones (electron carrying coen- zymes) are absent. 8. Unique carbon dioxide fixation reactions are involved in cell systhesis. 245 MW Hydrogen is an intermediate product in the anaerobic fermentation of organic matter. However, it is rarely detectable in a normal diges- ter, because it is consumed by the hydrogen utilizing methanogens as rapidly as it is formed. It has already been stated that hydrogen utilizing methane bacter- ia have a great affinity for hydrogen (Equation 2-6). From Table 2-1, the standard free energy change for the reactions in Equations 2—1 to 2—N all show positive values which indicate that the degradation of butyrate, valerate, propionate, and ethanol to acetate are thermodynam- ically unfavorable, unless the partial pressure of hydrogen is maintained at a very low level. One way to achieve low hydrogen con- centrations is the consumption of hydrogen by hydrogen utilizing methanogens. As has mentioned before, H2-producing acetogenic bacteria are syntrophicly associated with hydrogen utilizing methane bacteria and can not survive if separated from hydrogen utilizing bacteria. By combining Equation 2—6 and Equations 2-1 to Z-H, as shown in Table 2—6, the result of the syntrophic association is obvious, the free energy changes for Equations 2-9 to 2-12 all having negative values. Therefore the degradation of the above substrates becomes energetically favorable. 31 Table 2—6 Chemical Reactions For The Syntrophic Association of Bacteria and Hz-utilizing Methanogen (2-6) + (2-1) 2 CH3CH20H + H00; === 2 CH3COO’ + CH, + H2O + H+ (2—9) AG; = - 27.8 Kcal/reaction (2—6) + (2-2) 2 CH3CH2CHZCH2COO' + H00; + H20 === 2 CH3CH2000' + 2 cu3c00' (2-10) + CH" + ("1+ Axe; = - 9.u Kcal/reaction (2—6) + (2-3) u CH3CH2COO' + 3 320 === u CH3COO' + 3 can + H+ (2-11) 15G; = - 2H.5 Kcal/reaction (2-6) + (2-4) 2 cn3cuzcnzcoor + H003 + H20 === u CH3COO' + CH, + H+ (2-12) £560 = - 9.“ Kcal/reaction The effect of hydrogen partial pressure on the free energy change for the degradation of fatty acids can best be described by Figure 2-8 (Zeikus, 1979, McInery and Bryant, 1980). This shows that hydrogen partial pressure has to be lowered to 2x10’3 atm fOr the degradation of butyric acid and 9x10"5 atm for the degradation of propionate. Methane formation from hydrogen and carbon dioxide is energetically favorable at hydrogen partial pressures greater than 2x10"6 atm. When a reactor is stressed, such as shortened retention time or transient organic 32 Aowmfi .ucmxcm gmuwmv .mumczuzn new .mamcowaoga .Focmcpm do cowumnwxo mg» Low mmcmgu zmcmcm mmgw mg» no Amxav mcammmgg meugmn :mmoguxs we uommeu mum mesmwu 5: 00 mm A: Ia “and can 87 ow- on: o o: ow _ _ . . _ _ _ _ _ 1 u 1 _ _ . a u 4 _ a n .. _ n r1 _ 1 " TI “ mumcoonLa 1 r. ¢1m1/// " I . Pecans” " rl. _ l1 _ mumcxpzm _ _ _ h _ _ _ c . 1E (wie) sz 601 33 loading, hydrogen produced from fermentation of organic substrate can not be effectively consumed by methanogens causing the hydrogen concen- tration to increase which results in a supression of fatty acid degradation. Figure 2—8 shows that a slight increase of hydrogen par- tial pressure above 10’5 will cause propionate degradation to become unfavorable, while further increases of H2 partial pressure will cause other acids to accumulate in the system. 2.1.5.1 WWW .Bnoducts Hydrogen concentration in the anaerobic fermentation system also plays an important role in regulating the quantity and types of organic products formed by the fermentative bacteria by .interspegies .hygrggen tnansfsr (Wolin, 1974). Production of molecular hydrogen by fermentative bacteria is through the reoxidizing of the reduced NAD+ (NADH, diphosphopyridine nucleotide) generated in the glycolysis pathway as Equation 2—13 shows: 4. NADH + H ==== NAD+ + H2 (2-13) AG' 0 = + ”.3 Kcal/reaction The above equation is thermodynamically unfavorable at hydrogen partial pressure above 10'3 to 10'” atm as indicated by Wolin (1974). The effects of hydrogen partial pressure on the free energy change for Equation 2-13 are shown in Table 2-7. Several investigators (Kaspar and Wuhrmann, 1978; WOlin 1974) have reported that when hydrogen is effectively consumed by methanogens, the oxidized fermentation products 34 such as acetate and carbon dioxide will increase and the reduced (elec- tron sink) products such as propionate, ethanol will decrease, and an increase of hydrogen production which in turn is used by hydrogen-utilizing methanogens. This phenomenon can be seen from the experimental results performed by WOlin (1974) as shown in Figure 2-9 and Table 2-8. Table 2-7 Effects of Hydrogen Partial Pressure on Free Energy Change H2 (atm) Ac; 10° + n.33 10'1 + 2.97 10"2 + 1.61 10'3 + 0.25 1o"l - 1.11 10-5 - 2.u7 10"6 - 3.83 35 RUMINOCOCCUS RUMINOCOCCUS FLAVEFACICUS - FLAVEFACICUS ONLY CELLULOSE ] + * ; METHANOBACTERIUM : RUMINANTIUM l + J K. + + H + NADH : NADH + H = NAD ++§> 1 1 1 FORMATE === { PYRUVATEJ === FOijTE + C0 + H ' FF!- VEG- 2 -2 i : 2 2: ' L... “'F"""' NADHl 1 ACETATE C02 . ACETATE l u 1 ‘ l 3 I [ UCCINATE ] ' CH4 ' 1 Figure 2-9 Fermentation Interactions Between Bumjnggggcus flayefacius And Hz-utilizing Methanogens. 36 Table 2—8 Fermentation Of Cellulose By.R..£1a1e£agiQu§ AndRiflavefacicusPlusmminantim moles/100 moles Cellulose Products R. flavefacicus R. flavefacicus + M. ruminatium Acetate 7H 1H5 Formate 35 3 Succinate 94 25 Hydrogen 33 0 Carbon Dioxide 37 79 Methane O 63 * Source: Wolin (197“) W W is an important celluloytic species found in the runen. Figure 2-9 shows that when 3., W grows alone on cellulose, the main products are succinate and acetate with small amounts of carbon dioxide and hydrogen; but no methane is found. When a coculture of L flavefacicus and flethanobagtgnim minantim is grown, the main products are acetate, carbon dioxide and methane. In the coculture environment, hydrogen concentration is maintained at a low level, shifting electron flow from the production of succinate to the regeneration of NAD+ and hydrogen. Pyruvate metabolism is shifted from succinate to more acetate formation. Therefore, if the hydrogen concentration is high in the system the reduced fermention products and hydrogen will accunulate and substrate utilization may be inhibited (Mah et al., 1977). The results of the interaction between hydrogen utilizing methano- gens and nonmethanogens in anaerobic fermentation may be summarized as follows: (1) increase substrate utilization; (2) different proportions 37 of reduced end products; (3) more ATP systhesized by the nonmethanogens; (4) increased growth of both organisms (WOlin, 1974). 2.1.6 Role Of Nitrate egg Sglflete In Ageerobic Meghane Eementation If sulfate and nitrate are present in the system, methane fermen- tation will be inhibited because nitrate and sulfate have higher electron affinity than carbon dioxide and will compete seriously with carbon dioxide for electrons. Table 2-9 gives the redox potential of some redox pairs. Figure 2-10 illustrates the relationship of fOur electron acceptors, 02, N03, 80;, and 002, according to the order of magnitude of their redox potential. From Table 2-9, it can be seen that hydrogen has the greatest ten- dency to donate electrons and oxygen has the greatest tendency of accepting electrons. In natural ecosystem, nitrate is first reduced, followed by the reduction of sulfate and finally the formation of methane (FiEUre 2-10). Therefore, methane can only be formed in the absence of nitrate and sulfate. If nitrate exists in the system, methane is produced only after all the nitrate is reduced to nitrogen. 38 Table 2-9 The Oxidation-reduction Potentials 0f Some Redox Pairs Redox Pair Redox Potential , 8; (Volt) 2 H+/H2 - 0.u1 NAD+/NADH - 0.32 COZ/Acetate - 0.29 cog/ca, - 0.2a 50fi/H23 - 0.22 Fumarate/Succinate + 0.03 hog/N0 + 0.36 N0§/N0§ + 0.u3 Fe+3/Fe+2 + 0.77 1/2 ob/Hzo + 0.82 * Source: Brock, 1979 Most of the nitrate reducing bacteria are facultative anaerobes, they can transfer electrons to oxygen or to nitrate when oxygen is absent. Since sulfate reducing bacteria are obligate anaerobes, they can use hydrogen as the major electron donor; 0 H2 + so: ==== H28 + 2 H20 + 2 on' (2-10) AG'O = -36.ll Kcal/reaction u H2 + H00; + H+ ==== CH, + 3 H20 (2-15) AG'O = —32.11 Kcal/reaction It appears that sulfate-reducing bacteria can successfully compete with 39 REOOX POTENTIAL ELECTRON ACCEPTOR PRODUCT < 02 AEROBIC RESPIRATION +0.82 =: H20 N03 DENITRIFICATION +0.93( _ S0,,= ANAEROBIC RESPIRATION -0.22( C02 METHANOOENESIS -0.2M( 3:: CH4 A fl Figure 2-10 Electron tower for OZ/HZO’ NO3/NOE. SOZ/HZS. and C02/CH4 redox pairs. 40 hydrogen utilizing methanogens for hydrogen. If sulfate-reducing anaerobes are present in the system, electron flow is diverted from methane formation to H28 production. 2.2 Wm In view of the previous discussion, a well operated anaerobic fer- mentor must be low in hydrogen concentration, have near neutral pH, and balanced production and utilization of volatile fatty acids. In other words, the stability of an anaerobic fenmentation system may be dis- torted by inproper pH, high hydrogen partial pressure, and high fatty acid concentration. These three parameters are actually closely relat- ed, variation of one factor causing other parameters to be affected. For instance, when fatty acids begin to accumulate in the system, the pH value will drop and inhibit the activity of hydrogen utilizing methanogens. Therefore the hydrogen concentration will increase, which in turn will supress the degradation of volatile fatty acids, resulting in a further pH decrease. In addition to these three factors, process instability may be also caused by sudden changes of environmental and operational conditions, such as a sudden change in temperature , organ- ic loading, and hydraulic loading. Several organic and inorganic compounds, such as ammonia and heavy metals, also play a significant role in process instability. Further discussion of some of these fac- tors will be given in the following sections. 41 2.2.1 W Accunulation of fatty acids and reduction of pH in a reactor are two common signs of a failing anaerobic fermentation system. Inhibition resulting from high concentrations of fatty acids has been studied by several investigators. Two major conclusions may be drawn from their studies which conflict with each other. One group of researchers (McCarty and McKinney, 1961a; Cassell and Sawyer, 1959; Sawyer et al., 1954; Kaplonsky, 1951) believed that methane bacteria were inhibited because of the drop of pH value caused by high fatty acid concentration in the system, and that this inhibition may be removed by the addition of buffering chemicals to raise the pH value. Another group (Buswell, 1939; Schulze and Raju, 1958; Mneller et al., 1959) argued that fatty acids themselves were directly toxic to methane bacteria at concentrations above 2000 mg/l regardless of the pH main- tained, and the toxic condition can be released only by diluting the reactor substrate or reducing the substrate loading rate. Buswell and Mogan (1962) further reported that propionic acid would inhibit the methane bacteria. However, studies by McCarty et al. (1964) found another controversial result that propionic acid had little effect on methane bacteria but did inhibit the acid forming bacteria. Andrews (1969) tried to solve the conflicting ideas about fatty acid toxicity and reported that the toxicities were caused by the non-ionized portion of volatile acids. Thus toxicity is directly related to both the pH value and the acid concentration because the relative concentrations of ionized and un-ionized fatty acids are affected by hydrogen ion concen- tration, for instance, 42 CH3COOH = CH3COO' + 11+ (2—16) When pH decreases, equilibuiun shifts to the left hand side and causes the un-ionized acid concentration to increase. Krocker (1979) also reported the same results that toxicity would increase when the pH dropped and un-ionized volatile acid concentration increased. Because non-methanogenic bacteria can grow in a low pH environ- ment, an unbalanced reactor with a lower pH value would favor rapid growth of non-methane bacteria and faster production of fatty acids. This will result in a fUrther pH drop, increasing inhibition of methanogenic activity and causing the accumulation of hydrogen and fatty acids. The result of this adverse cyclic interaction between pH, acid concentration, and hydrogen concentration is the total failure of fermentation system. 2.2.2 Amonialoxicitx Ammonia may be present in the anaerobic fermentation system in the form of ammonium ion (NHE) or free ammonia (NH3). Concentration of these two forms of ammonia are affected by the hydrogen ion concentra- tion in the system; low pH favors the formation of ammonium ion (NHfi) and high pH favors free ammonia (NH3) production (Equation 2-17). + + NH” ==== NH3 + H (2-17) 0 -5 The dissociation constant for ammonia at 35 C is 1.849 x 10 , or pKa : ”.733. 43 Amonia serves as the nitrogen source for the microbial growth in fermentation systems. However, it can also be a toxic agent if excess concentration is present in the system. McCarty (1964) reported that ammonia nitrogen concentrations of 150 to 300 mg/l are inhibitory to the system at pH values greater than 7.4 to 7.6, and, if the concentra- tion exceeds 3000 mg/l, anmoniun ion itself becomes very toxic regardless of the pH. However, Krocker (1979) reported that process inhibition by ammonia was the result of excessive concentration of free ammonia rather than anmoniun ion. 2.2.3 W A nunber of earth-metal salts such as sodiun, potassiun, calciun, and magnesiun may be associated with the substrate and may be intro- duced into the system. The presence of these substances may inhibit process opeeration if high concenetrations is present. McCarty and McKinney (1961b) performed a series of experiments and found the pro- cess instability due to metal salts was associated with the metal cations rather than volatile acid anions. They also reported various cation concentrations that would cause inhibition, as shown in the Table 2—10. 44 Table 2-10 Concentration For Salt Toxicity (mg/l) Cation Moderately Inhibitory Strongly Inhibitory Sodium 3500 - 5500 8000 Potassium 2500 - 4500 12000 Calcium 2500 - 4500 8000 Magnesium 1000 - 1500 3000 * Source: McCarty, 1964 In the laboratory, the existance of these salts is mostly contri- buted by the agents used for pH control. Therefore, concentration of these substances are usually fairly low and do not cause inhibition effects unless large amounts of chemicals are added. 2.2.4 SW Several heavy metals such as copper, nickel, zinc, and chromium are frequently toxic to microbial activity in many biological processes. The maximun allowable concentrations of these heavy metals vary as shown in Table 2-11 which summarizes the results reported by previous investigators. 45 Table 2-11 Toxic Concentrations of Some Heavy Metals in Anaerobic Digesters Metal Toxic Concen. (mg/l) Reference Copper 150 - 250 Rudgel, 1941 500 Rudgel, 1946 1000 Barnes & Braidech, 1942 Nickel 200 Barnes & Braidech, 1942 1000 Wischmeyer & Chapman, 1947 Zinc 1000 Rudolphs & Zeller, 1932 350 McDermott et al., 1963 Chromium“ 2000 Barnes 5. Braidech, 1942 200 Pagano et al., 1950 Source: Kugelman and Chin, 1970 * At normal pH levels, chromium normally reduces to the trivalent form which is very insoluble and consequently is not as toxic as the hexa- valent chromium. Heavy metal toxicity may be released by precipitation of the metals by adding sulfides such as sodiun sulfide into the reactor. The solubility product of heavy metal sulfides range fran 3.7 x 10"19 for FeS to 8.5 x 10.“5 for CuS (McCarty et al., 1964; Lawrence and McCarty, 1965). At pH values higher than 7.6, concentrations of zinc greater than 1000 mg/l, can be precipitated out as zinc carbonate (Mosey et a1, 1971, 1975). 46 Heavy metals do not exist or are only present in trace amounts in cellulosic agricultural residues. Therefore, they are not considered as potential toxicants in this research. 2.3 W In order to maintain stable process operation and to obtain optimum. efficiency, it is important to understand the controlling parameters. Some biological and chemical factors that directly and indirectly influence process stability have been discussed in the pre- vious sections. Application of these concepts and other factors required for effective operation will be disscussed in this section. Important operating variables include, 1. pH 2. Alkalinity 3. Volatile fatty acid concentration 4. Tamperature 5. Absence of toxic material 6. Nutrient availability 7. Retention time 8. Degree of mixing 47 2.3.1 Won A complete anaerobic fermentation system (type A habitat) termi- nates with the formation of methane. Because methanogens are more sensitive to pH changes than other groups of bacteria (see Section 2.1.4.1), the optimum pH range for methanogens (6.5 - 7.7) automatical- ly becomes the optimum pH range for the entire system. The pH value of a digester is a function of three parameters: alkalinity, volatile fatty acid concentration, and the fraction of carbon dioxide in the reactor's gas phase. Alkalinity is the measurement of carbonate and bicarbonate concentration in the reactor and it acts to buffer against pH fluctuation due to changing acid concentrations. Under normal con- ditions, pH in the reactor is maintained in the proper range by the destruction of fatty acids and formation of bicarbonate buffering. The main buffering substance in most anaerobic digesters is NHuHCO3. A suitable ammonia nitrogen concentration, 50 - 200 mg/l (McCarty, 1964), can provide both the nutritional requirement for microbial growth and the necessary bicarbonate buffering. Ammonium ion ... (NHu) does not provide bicarbonate buffering directly but only through: 4. .- In the anaerobic fermentation system, total alkalinity is composed of both bicarbonate alkalinity and fatty acid alkalinity and has the rela- tionship expressed in Equation 2-18, TA : BA + (0.85 x O.833)(TFA) (2-18) 48 where TA = total alkalinity, mg/l as C3003 BA = bicarbonate alkalinity, mg/l as Ca003 TEA = total fatty acid concentration, mg/l as acetic acid Acetic acid is converted to the equivalent alkalinity as CaCO3 by a factor of 0.833. The factor of 0.85 in Equation 2-18 is an adjust fac- tor because 85% of the volatile acid alkalinity is measured by titration of total alkalinity to pH 4 (McCarty, 1964). To ensure a sufficient buffering capacity, a bicarbonate alkalinity in the range of 1000 - 5000 mg/l at pH range of 6.6 to 7.6 must be maintained. According to the relationship in Equation 2-18, bicarbonate alkalinity will be decreased due to the increased concentration of total volatile' fatty acid. One control parameter often used by anaerobic digester operators is the total volatile fatty acid concentration (mg/l as acet- ic acid) to total alkalinity (mg/l as CaCO3) ratio. If the value of this ratio drops lower than 0.8 the reactor becomes unbalanced. Low pH reactor may be restored by reducing the substrate feeding rate or adjusting pH by the addition of chemical reagents such as bicarbonate, phosphate, lime, sodash etc.. Among those buffering reagents, lime is the most popular chemical being used by many wastewa- ter treatment plants for pH control. However, lime is good only fOr completely mixed reactors where the pH has dropped below 6.5. Also, the amount of lime dosage must be carefully controlled; lime should be added only to raise the pH to about 6.7 (McCarty, 1964). Over dose of lime will cause excessive consuming of 002 and resulting in high pH (about 8.0). 49 Equation 2-19 shows that lime initially reacts with C02 to form calcium bicarbonate. When the bicarbonate alkalinity reaches some point between 500 and 1000 mg/l, and the pH is about 6.7, additional lime will result in the formation of insoluble calcium carbonate (Equation 2-20) without increasing the pH or alkalinity until the C02 in the gas phase is depleted. Sodium bicarbonate is also a good pH control agent because it can provide 5000 - 6000 mg/l of alkalinity without causing toxic effects. 2.3.2 W The chemical composition of a microbial cell, the activities of cellular enzymes, and bacterial nutrition are all influenced by the temperature at which a bacteriun is grown. Therefore, the growth rate of microorganisms is a function of temperature. Conceptually, tempera- ture ranges for the optimal growth of microorganisms can be divided into three temperature regions: a thermophilic zone (above 45 °C) , a mesophilic zone (20 - 45 oC), and a psychrophilic zone (below 20 °C). The effect of temperature on anaerobic fermentation has been intensive- ly studied by nany investigators (Golueke, 1958; Malina, 1962; Farrel et al., 1967; Speece et al., 1970; Maly and Fadrus, 1971; Pfeffer, 1974; van Velsen et al., 1979). The recommended temperature range for efficient anaerobic sludge digestion is between 30°C and 35°C for the mesophilic digesters (Malina, 1964). 50 It is a common understanding that a mesophilic organism operating optimally at 30°C should not be expected to function well at an elevat- ed temperature of 60°C. Therefore, a system normally operating at 30°C could be upset if the temperature is raised to 45°C. Buswell (1952) reported that in a sudden change of temperature of as little as one or two degrees (centigrade), inhibited methane formation and volatile fatty acids accunulate. A temperature change also affects the 002 concentration in both liquid and gas phases. The solubility of carbon dioxide decreases with increasing temperature. Thus, the 002 concentration will decrease in the aqueous phase and increase in the gas phase at higher temperatures. The carbonate equilibrium constants are also affected by temperature change; pKa1, for H2003 = H00; + H+, decreases fran 6.52 at 5°C to 6.30 at 60°C, and pKaz, for H003 == C0; + H+, decreases from 10.56 at 5°C to 10.14 at 65°C (Snoeyink A Jenkins, 1980). Therefore bicarbonate concentration will be decreasing with increasing of temperature. 2.3.3 AbsencLQfloxiLMatenial If toxic materials are presence in the reactor, two signals of inhibition may be exhibited: (1) a decrease in methane gas production; (2) a decrease in volatile fatty acid concentration. In a mixed cul- ture ecosystem with mixed substrates, it is difficult to obtain a definite concentration at which a component becomes toxic. The magni- tude of a toxic effect may be relieved or enhanced by complex interactions, known as antagonism (a reduction of the toxic effect of one substance by the presence of another) and synergism (an increase of 51 the toxic effect of one substance by the present of another). Microbial cultures may also become acclimated to the toxic substances. For instance, McCarty (1964) indicated that an anaerobic digester is inhibited by un-ionized anmonia nitrogen at a concentration greater than 150 mg/l as NH3-N. However, Krocker et a1. (1975) performed a successful anaerobic digestion experiment with swine manure at un-ionized amnonia concentrations of 500 mg/l as NH3’N. The degree of acclimation may also explain the variability in toxic concentrations reported by various invistigators (Table 2-11). Parkin et al. (1983) studied the response of methane fermentation to several toxicants (including ammonia-nitrogen, copper, nickel, chlo— roform, formaldehyde, hydrazine) and reported that the system could recover after extended periods of zero gas production, provided the microbial solids retentation time is long enough. Therefore, those processes with high solids retention time and short hydraulic retention time, such as anaerobic filters and anaerobic biological rotating disks, should have the highest potential for recovery from toxic inhi- bition. 2.3.4 Wine Hydraulic retention time and microbial solids retention time are the two most important control parameters for process design and opera- tion. Hydraulic retention time (HRT) is defined as the ratio of effective reactor volume to the flow rate of substrate stream passing through the reactor and can be expressed as: 52 HRT - V (221) ’ 0 where V .. effective reactor volune, (L3) 0 liquid substrate flow rate, (L3/T) Solid retention time (SRT) is defined as the total active microbial mass in the system divided by the total quantity of active microbial mass that is withdrawn fran the system per unit of time and can be expressed as: xt 00 = ———— (AX /AT) where 00 : microbial solids retention time, (time) (2-22) or sludge age, or mean cell residence time Xt = total active microbial mass in system, (mass) (AX /A'r) = total quantity of active bianass withdrawn per time, (mass/time) . Biological solids retention time is nunerically equal to the hydraulic retention time for a steady state, completely mixed reactor without recycle. Adequately long solids retention time is crucial for effec- tive operation of anaerobic fermentation processes; low solids retention time will cause washout of the microbial mass fran the reac- tor resulting in system failure. A sunmary of minimun solids retention times for anaerobic digestion of various substrates are shown in Table 2.12. 53 Table 2-12 Minimun Solids Retention Time For Anaerobic Methane Fermentation o m Temperature, C Substrate 0c (day) Reference 15 Municipal sludge 6Oa O'Rourke, 1968 20 Acetic acid 7.8 O'Rourke, 1968 Stearic & palimitic 7.2 O'Rourke, 1968 acid Mixed acids 7.2 O'Rourke, 1968 Municipal sludge 103 O'Rourke, 1968 25 Acetic acid 4.2 Lawrence & McCarty, 1969 Propionic acid 2.8 Lawrence & McCarty, 1969 Stearic & palimitic 5.9 O'Rourke, 1968 acids Mixed acids 5.9 O'Rourke, 1968 Municipal sludge 7.5a O'Rourke, 1968 30 Acetic acid 4.2 Lawrence & McCarty, 1969 35 Acetic acid 3.1 Lawrence & McCarty, 1969 Propionic acid 3.2 Lawrence & McCarty, 1969 Butyric acid 2.7 Lawrence & McCarty, 1969 Stearic & palimitic 4.0 O'Rourke, 1968 acid Mixed acids 4.0 O'Rourke, 1968 Municipal sludge O'Rourke, 1968 Municipal sludge 2.6a Torpey, 1955 Source: Lawrence and McCarty, 1970 m 0c = limiting minimum solids retention time, determined by calculation from experimental data except as noted. m a = 0c determined by washout. 54 The limiting,minimum solids retention time is defined as the value an of 0c which occurs when influent substrate concentration is much greater than the half velocity coefficient, K (Lawrence & McCarty, s 1970). If the following assumptions hold: (1) A constant proportion of the organisms are viable; (2) The primary substrate serves as the essential limiting nutrient; (3) Microbial growth can be expressed by Monod's model, the solids retention time for a steady state, completely mixed, single reactor without recycle can be expressed as: KS/S + 1 Ge = (2-23) YK - b( KS/S + l) where KS = half velocity coefficient, equal to the substrate concentra- tion when dF/dt = 1/2 (K), in which dF/dt = rate of microbial substrate utilization per unit volune, K = maximum rate of substrate utilization per unit weight of microorganism; S : substrate concentration; b = microorganism decay coefficient, time"; Y = growth yield coefficient, mass of organism formed per mass of substrate utilized. When the influent substrate concentration, S, is much greater than the half velocity coefficient, Ks, then Equation 2—23 can be simplified as, m 1 0 .-.—__ 2-224 0 YK-b ( ) The limiting minimun solids retention time listed in Table 2-12 were calculated using Equation 2-24. In general, solids retention times of 10 to 30 days at 35°C are employed by many anaerobic sludge digesters: these retention times are 3 to 10 times greater than the limiting values. McCarty (1970) sug- 55 gested that a safety factor of about 3 to 10 should be applied to the minimum solids retention for operation and design of anaerobic diges- ters. Another control strategy occasionally used is the volumetric organic loading rate. It is defined as the rate per unit volume at which organic substrate is fed into the reactor, and can be expressed as: (organic concentration) x Q V Organic Loading Rate = organic substrate concentration : HRT (2-25) Therefore organic loading rate is related both to the hydraulic reten- tion time (HRT) and the percentage of organic contents in the influent substrate. The values of loading rate can be changed by changing sub- strate BOD or HRT at a given substrate concentration. At lower organic loading rate and when substrate concentration does not change, higher percent of organic substrate will be degraded but less CH“ will be pro- duced per volune of reactor as compared to that at the higher organic loading rate. 2.3.5 W For many conventional anaerobic fermentation processes, mixing is an important operational parameter to achieve satisfactory treatment efficiency. Sufficient mixing of the reactor can provide the fallowing benefits: (1) uniform distribution of substrate, microorganisms, and temperature; (2) substrate is kept in continuous contact with the 56 microorganism; (3) biological intermediates and end-products are uni- formly distributed; and (4) prevention of a scan blanket. Finney and Evans (1975) hypothesized that methane production is influenced by the phase transfer rate and suggested that vigorous agi- tation, low pressure (vacuum), and high temperature would increase the rate of phase transfer resulting in higher methane production rates. However, Coppinger et al. (1979) reported no decrease in gas produc- tion when mixing was discontinued. They indicated that the gas bubbling and thermal convection currents provided sufficient mixing for the reactor. Hashimoto (1982) reported that although a continuous mixed fermentor produced significantly higher methane than the fermen- tors mixed only two hours per day, the methane production rate from the continuously mixed fermentor was only slightly higher than the rate produced from another fermentor with intermittent mixing. Therefore, he concluded that there is little potential for increasing fermentation rates by excess increased mixing, and that phase transfer controling mechanisms have minimal effect on the CH4 production rate. 2.3.6 Nutrientmmbts Microorganisms require a variety of substances for synthesis of cell material and for generation of energy. Since microorganisms are extremely diverse in their physiological properties, nutrient require- ments for each species of bacteria are not identical. The chemical composition of cell material gives the basic idea of the major material that are required for cell growth. The approximate elementary composi- tion of a microbial cell is given in Table 2-13. 57 Table 2-13 Approximate Elementary Composition of Microbial Cells Element Percent of Dry Weight Carbon 50 Oxygen 2O Nitrogen 14 Hydrogen 8 Phosphorus 3 Sulfur 1 Sodiun 1 Calcium 0.5 Magnesiun 0.5 Chlorine 0.5 Iron 0.2 All others 0.3 * Adapted from Stanier et al., 1976 The major components of a microbial cell are hydrogen, oxygen, carbon, nitrogen, phosphorus, and sulfur, these elements accounting for about 95% of the total cellular dry weight. The function of these material as well as other nutrients are summarized in Table 2-14. 58 Table 2-14 General Physiological Functions Of The Principal Elements Element Physiological Functions Hydrogen Constituent of cellular water, organic cell materials. Oxygen Constituent of cellular water, organic cell materials; electron acceptor in aerobes. Carbon Component of organic cell material. Nitrogen Component of proteins, nucleic acids, coenzymes. SulfUr Component of proteins and coenzymes. Phosphorus Constituent of nucleic acids, phOSpholipids, and coenzymes. Potassium Principal inorganic cation in cells, cofactor for some enzymes. Magnesiun Cofactor for enzymatic reactions; functions in binding enzymes to substrate; component of chlorophylls. Calciun Cofactor of some enzymes. Iron Constituent of cytochromes and heme or nonheme proteins; cofactor for some enzymes. Cobalt Component of Vitamin B12 Copper, Inorganic constituents of special enzymes. Zinc Source: Stanier et al., 1976 Most of the wastewaters treated by anaerobic fermentation processes contain sufficient nutrients for microbial growth although certain types of substrates such as cellulosic residues may be deficient in some nutrients such as nitrogen, phosphorous, sulfide, and iron. These deficient materials may be supplied as inorganic salts. Ammonium car- 59 bonate or hydroxide and anhydrous ammonia have been found as suitable nitrogen sources for methane fermentation. Other inorganic salts such as (NHu)2HP0u, NaHC03 may also be used to meet the nutrients require- ments. Micronutrients such as manganese, cobalt, copper, molybdenum, and zinc are required in very small quantities. They are usually present in adequate amounts in tap water or as contaminants of the major inorganic constituents in the growth media or influent substrate. 2.4. MW: Anaerobic fermentation processes that have been developed so far may be sunmarized into seven different configurations (Figure 2-11): (1) conventional - completely mixed without solids recycle, (2) anaero- bic contact - completely mixed with solids recycle, (3) batch-load, (4) plug flow with solids recycle, (5) anaerobic expanded bed, (6) high solids or dry anaerobic fermentation, (7) anaerobic filtration. Brief descriptions of each process will be given in the following paragraphs. 2.4.1 WW Most of the early anaerobic digesters were designed without main- taining a high microbial population in the system and belong to the flow-through type reactor without sludge recycle. The term "conventional" was used to describe this type of anaerobic digestor. Conventional digester has been used mostly in municipal sewage treat- ment plants for sludge stabilization. Sludge and microorganisms are uniformly distributed in the entire reactor and the digested effluent 60 (1) Conventional-Completely mixed, no solids recycle \ l Influent \AD I Effluent (2) Anaerobic Contact-Completely mixed with solids recycle CH4 + C02 Influent Mixed Liquor ‘ Effluent 1 ” \ l is Solids Recycle (3) Batch-Load CH4 + €02 ___ Effluent X} Influent \\\\~//,J CH4 + 002 \ l I ,,__.__ Effluent b Figure 2-11 Schematic Diagrams of Anaerobic Fermentation Processes. 61 (4) Plug Flow - with solids recycle CH4 + C02 l Influent ------- 0-1 DO-ooo-Ilo-d .-.----- -1 p-----—-— .-.- o ----l —----.-d h------ (5) Anaerobic Expanded Bed CH 4 1 L__.. Effluent Solids Recycle + CO2 Recycle Liquid Recycle Influent (7) Anaerobic Filter Filter Media Influent Effluent (6) Dry Fermentation CH4 + C02 Substrate F igure 2-11 Continued .____.._ Effluent 62 is withdrawn at the same rate as the influent sludge to maintain a con- stant reactor volume. This process assumes that the microbial concentration in the reactor and in the effluent are equal. 2.4.2 Water Anaerobic contact process, or anaerobic activated sludge process, is a modified form of the conventional anaerobic digester. This pro- cess employs a settler to separate the microbial solids from the effluent liquid and mix the recycled sludge with raw waste to maintain a high microbial concentration in the main reactor for more efficient and rapid treatment. This process assunes that all substrate stabili- zation occurs in the main reactor and the total biological mass in the system is equal to the biological mass in the main reactor. This pro- cess has been applied to treat medium strength industrial wastes. 2.4.3 W The batch load model is composed of two identical completely mixed anaerobic digesters. When the first one is fermenting the waste as a batch reactor, the second one is receiving raw sludge. When the second reactor is full of substrate, and begins its digestion period, the first reactor is emptied to receive incoming sludge. This process has been found to have a higher treatment efficiency than that obtained from conventional reactors. 63 2.4.4 W12 The plug flow model applies a longitudinal reactor without longi- tudinal mixing during the passage of substrate through the reactor. Similar to the contact process, this model has a settling tank to recy- cle active anaerobic sludge to mix with raw waste. The biodegradable organics decrease along the tank while the microbial concentration increases. The plug flow anaerobic reactor has been applied for animal waste treatment both in pilot scale and full scale (Hutchinson, 1972; Bell, 1973; Jewell, 1976). Jewell (1976) indicated that with the absence of internal mixing and at lower operating temperatures (22.5 0C), a successful plug flow anaerobic digester could be obtained which would be economically competitive with the conventional process. 2.4.5 WM Anaerobic expended bed, or anaerobic fluidized bed, are newly developed fixed-film anaerobic processes. These closely related processes consist of a column packed with an inert material, often send, which will expand as the waste flows upward through the column. The particles serve as a support surface for microorganisms to attach, and provides a large surface area to mass ratio. These processes pro- vide a high degree of contact between the substrate and biomass and therefore result in high treatment efficiency. 64 2.4.6 Wm Previous literature has provided little information about high solids fermentation. Early studies by Buswell et a1. (1936), Keefer (1947), and Schulze (1958) showed that anaerobic digestion could be performed at solids concentration of up to 20% , although, higher con- centrations of volatile fatty acid would build up in the reactor. Wong-Chong (1975) first studied the "dry anaerobic digestion" of animal wastes (dairy and poultry) at solids concentration greater than 20% in both batch and batch feed reactors. He concluded that anaerobic diges- tion of waste with high solids concentrations is feasible and provided economies in reduced reactor volume, digested sludge handling, and avoiding treatment of digester supernatant. However, he also pointed out the potential inhibition by ammonia due to high concentrations of nitrogenous waste. In his study, WOng-Chung did not resolve the sub- strate input and output problems for the semi-continous or continous flow reactors. He also did not solve the possible inhibition problem due to high volatile acid concentration built up in the reactor. Another "dry anaerobic fermentation" study done by WUjicik and Jewell (1979) dealt with a substrate of mixed dairy cow manure and agricultural residues in batch reactor; the study concluded that at a solids content of 40%, the methane bacteria were inhibited. And at 55% total solids concentration, volatile fatty acids reached a maximum. At higher solids content, acid production declined indicating that the fermentative bacteria were inhibited too. Again, their study of high solids anaerobic fermentation did not provide information about sub- strate input and output methods, and no information was given about the 65 possible resolution of inhibition due to high volatile acid concentra- tion. 2.4.7 AW Anaerobic filtration is a fairly new process. It was first introduced by Coulter et al. in 1957, but their study did not arose much attention at that time. The real development of this process was started by Young and McCarty in 1967 and subsquently by several other investigators working with a variety of wastes (Table 2-15). Up to the present time, anaerobic filtration has not been coupled with high sol- ids anaerobic fermentation of cellulosic residues. Anaerobic filtration has several advantages over suspended growth processes. ‘With the suspended growth processes mentioned above, either long hydraulic retention times or solids seperation and solids recycle are required to provide an adequate solids retention time (SRT) fOr efficient treatment. In anaerobic filtration, the substrate is passed upward through a colunn reactor which is filled with a median for microorganisms to attach and grow in a manner similar to the anaerobic expended bed. Because the biological solids affix to the surface of the filter medium or become trapped within the interstices void spaces and are not washed out in the effluent stream, a long solids retention time can be obtained without solids seperation. This process is only suitable for soluble or colloidal wastes. 66 Table 2—15 Summary of Previous Anaerobic Filter Studies 6000-27000 mg/l COD 000 Loading" Detention 000 Removal Type of Waste 1b””4031”; Time (Hr) Eg§iciency References Volatile Acid Young 8 and Protein 26.5 - 212 4.5-72 56 - 98 McCarty, Carbohydrate 1968 375-12000 mg/l COD Food Processing 30 - 86 Pulmmer 8 (carbohydrate) 100 - 640 13 - 83 (55-86 Malina, 8500 mg/l C00 soluble) 1968 Acetic Acid 370 12 30 - 80 Clark & 6400 mg/l COD Speece, 1970 Potato —t Pailthorp Processing Waste 33 - 145 13 - 59 41 - 79 et al, 1971 3000 mg/l C00 Wheat Starch Richter et al Waste 237 22 65 (76 % 1971 5930-13100 mg/l soluble) Taylor, 1972 COD Lab Scale . Organicvalcohols :gv;$us aldelydes, acids 35 - 130 17 - 46 64 - 76 1972 ’ amine, glycol, phenol 2000 mg/l COD (20000 mg/l COD system failure) Pilot scale 40 - 145 72 10 - 13 Petrochemical, 2000-8000 mg/l ACOD Brewery Press Liquor 50 - 400 15 - 330 30 - 97 Foree et al, 1972 * Based on total (empty bed) volume ** Based on initial unseeded void volume 67 Table 2-15 Continued * Type of Waste COD Loading Detention C00 Removal 3 3 Time (hr) Efficiency References lb/day-10 ft ** (%) "Metrecal" 427 18 70 - 95 El-Shafie & 11000 mg/l COD Bloodgood, 1973 Dilute Waste Sulfide Liquor 125 - 375 89 - 95 27 - 58 Wilson 8 1300-5300 mg/l BOD Timpany, 8005 -Removal 1973 Pharmaceutical Waste (95 % 14 - 220 12 - 48 94 — 98 Dennis & methanol) Jennett, 1250-16000 mg/l 1974 COD Grain Alcohol Dahab & Stillage 31 - 124 36 - 72 74 - 84 Young, 1981 3000 mg/l COD Pharmaceutical Sachs, Waste 34.9-104.6 36 18 - 80 Jennett & 2000-6000 mg/l (25 - 94 Myrton, 1982 C00 800) * Based on total (empty bed) Volume ** Based on initial Unseeded void volume 68 2.5 Who Cellulose is an important consitituent of all plant tissue, close- ly associate with lignin and non-cellulosic polysaccharides such as hemicellulose and pectic substance. To study the microbial degradation of cellulosic material there must first be a clear understanding of the physical and chemical nature of the material involved. This section will discuss the physical and chemical properties of the cellulosic substrate. 2.5.1 WWW Cellulose is a water insoluble, high molecular weight polymer com- posed of D-glucose residues jointed by B -1,4-glucoside bonds. Figure 2-12 illustrates the conformation formula (chair form) of cellulose, showing that the hydroxyl groups are located in the equatorial position and the H-atans are located in the axial position of the (LAM-linked anhydroglucose units. Every other glucose unit is rotated 180o around the main axis. The nunber of repeating cellobiose units ranges fran 500 to 5,000 giving molecular weights from 100,000 to 1,000,000. The chemical properties of cellulose are mainly determined by the glycosid- ic linkage and three hydroxyl groups on the glucose unit. Major chemical and enzymatic reactions occur at these locations (Fan et al., 1980). It has also been reported (Baney, 1968) that the hydroxyl group in the 3-position is bound by intramolecular hydrogen bonding to the ring oxygen atom of the next unit as shown in Figure 2-13. Therefore, cel- lulose molecules are linked longitudinally to form fibrils, which 69 Cellobiose Unit Figure 2-12 Conformation Formula of Cellulose 0H CHZOH Figure 2-13 Cellulose Molecule With Intrachain Hydrogen Bonds 7O laterally linked by hydrogen bonding of adjacent fibrous chains. The length of one anhydroglucose unit is about 0.515 nm (5.15 A), and the total length of the native cellulose molecule (10,000 units) is about 511m- The regular cellulose molecule exhibits a crystalline X-ray dif- fraction pattern. This periodic structure of diffraction repeats itself every 10.3 A in the direction of the fibre axis (WOod, 1970). X-ray diffraction patterns also show that cellulosic fibers contain amorphous areas. The degree of crystallinity varies with the type of cellulosic material. In general, the proportion of crystalline materi- al ranges from 50 to 90%. 2.5.2 MW Unlike starch, which is a poly-cx-1,4-D-glucosan, cellulose does not act as a carbonhydrate reservoir which can be readily broken down to glucose whenever necessary. Only a few specialized species of aero- bic bacteria and fungi can utilize cellulose. However, nany species of facultative and strict anaerobes can secrete extracellular enzymes, called cellulases, to hydrolyze cellulose. The biochemical transforma- tion of cellulose into smaller soluble carbohydrates is known as "cellulolysis". The degradability of cellulose by microorganisms varies greatly with the nature of the cellulose. The physical and chemical features that may affect biodegradation of cellulose include the following: 1. The degree of crystallinity: cellulose with a higher degree of crystallinity has stronger resistance to enzymatic hydro- 71 lysis. 2. The unit cell dimension of the cellulose: a smaller unit cell is more easily hydrolyzed. 3. Degree of polymerization of the cellulose: a higher degree of polymerization is expected to have a slower rate of hydro- lysis. 4. The nature of the substances with which the cellulose is associated: lignin, hemicellulose, mineral constituents, and trace amount of N and P may associate with cellulose. The higher percentage of lignin associated with the cellulose, the more difficult the enzymatic hydrolysis. Lignin, one of the most chemically and biologically resistant materials, is a polymer of aromatic compounds. Lignin supplys strength and rigidity to the plant tissues and acts as a physical barrier, min- imizing water permeation across the cell walls, and providing protection against infection. The degradability of cellulosic material is inversely correlated to the amount of lignin present in the sub- strate. Hemicelluloses are amorphous, non-cellulose, heterogeneous mixtures of linear or branched polymers that may contain D-xylose, D-mannose, D-glucose, D-galactose, and D-glucuronic acid. They can be isolated from the original or delignified tissue by extraction with dilute alkali and acid. The content of hemicellulose in plant tissue ranges from 6 to 40% depending on the type of plant. Homicelluloses are readily degradable by bacteria. 72 2.5.3 Wm Wheat straw was selected as the substrate for this research, because it is a major agricultural residue and it is available in large quantity on the M.S.U. campus and also because it is easy to handle. Wheat straw has been mostly used for animal bedding and has also been used for animal feed, construction material, and paper pulp. Other uses of wheat straw are listed in Table 2-16. Table 2-16 Usage oF Wheat Straw Methods Products Direct Uses Fuel, fertilizer, soil conditioner, feed, packing materials, bedding for animals. Mechanical Conversion Pulp and paper. Chemical Conversion Sugar, alcohol, cellulose derivatives, phenolic compounds, lignin, etc. Biological Conversion Sugar, alcohol, enzymes, fermented feed, methane. Source: Han, 1979 As with many other plant materials, wheat straw is mainly composed of cellulose, hemi-cellulose, and lignin. In addition to three major components, wheat straw also contains trace amounts of protein, calci- un, potassiun, magnesium, phosphorus, and sulfur. CHAPTER THREE EXPERIMENTAL LETHODS AND MATERIALS A specially designed reactor system was used to conduct the exper- iments of this research. The experimental procedures were designed to verify the feasibility of the proposed process for semi-continuous treatment of cellulosic residues. This chapter will present the exper- imental system, the experimental procedures, as well as the analytical methods that have been used in this research. 3.1 WOW The reactor system, as shown in Figure 3-1, consists of two fixed-film anaerobic filters, ten high solids packed reactors, a liquid equalization reservoir and a gas collection system. The reactors were connected in series and which was allowed to have a semi-continuous substrate feeding. The packed reactors were designed to have a sta- tionary solid phase and a mobil liquid phase. The liquid phase in the packed reactors was punped by a constant flow rate pump and was recir- culated through two anaerobic filters that served as the methane generators. Figure 3-2 shows the details of the packed reactor and the gas collection system. The packed reactor was made of 1/8 in. thick acrylic cylinder with the inside diameter of 7.3 cm. (2 7/3 in.) and the total height of 14.4 cm. (5.7 in.), and which was divided into three sections; a substrate holding section having length of 10.4 cm. in the middle, and two liquid sections with 2 cm. in length for 73 74 I 4'(12L9 am) #_J 12 1 10 9 8 7 1 {‘1 f-\ 1C:/ \I;/ 71 9 ‘ —r— ’ i . ’ I .. i D: r. i "132 \‘ I” : * y I ,J 1’ I y I o __J_ r- /- / w M GIG—6949 5’ \/ SLOTTED ANGLE FRAME 3 4 5 5 ___L_____ - 1 J1 FL —T_ //N Ag;2;B?INE afi GAS EOLLECTION C N N3 SOLUTION 1 z . CYLINDER RESERVOIR l ‘k~.nL "J HWJ .lTJ IIIJ 5" 8. GAS! 5,.e I :1 ’7 r ,3 5 - :1 - a L _ CO (\I A' / I CO I :1 CYLINDER SUPPORT RODS Lk\ 1 *2 \ HIGH SOLIDS PACKED REACTOR LIQUID "“-J*\‘ ~ F1 RESERVOIR LIQUID L NE ‘ STIRRER METERING PUMP 57 L. ITEZAH I (30.5) __JL_ 79’79C/)’/’/’ zC/9CZ/VC/ Figure 3-1 Schematic Diagram of Experimental System 75 Brine Solution Gas Collection / Reservoir /Cvlinder l‘l Fl). FFL Fl é , Gas Line E E. + E -¥~ Stop 3 E- Valve 6‘ E Check Valve;§’ _ I If x W K 7 7 1’, Brine Network / E1385. Line G S LI 6 ,/ ‘ L'lQU'ld a n Effluent No. 13.5 Stopper __ I : (3%— ' g 7 7/ :3 5 .0 Packed <_ '; Reactor , es .9 ../ Reactor SupportIng *1 //// —' Plate /// \ 1L / §3L_.fi I Liquid Influent Line 7" I 7.3 I L 12 7 cm _J No. 7 Rubber F77 ° I Stopper Liquid will ing Pot Figure 3-2 Schematic Diagram of Packed Reactor And Gas Collection System. 76 each section located on the top and the bottom of the reactor. Sixty grams of air dried, chopped wheat straw was placed in the center sub- strate section for each experimental run. The total volume of one packed reactor was 600 ml and the volume for the center substrate sec- tion was 435 ml. The volume occupied by the dry straw was 150 ml; thus, based on the volune to volune ratio, the solids concentration in the packed reactor was: 150/435 = 34.5 1 . The anaerobic filters have the same size and dimensions as the packed reactors, and were constructed by using the same material. Wheat straw was used as the filter media and provided surface area and interstitial void spaces for microorganisms to attach and growth. Both types of reactor have a 5 inches by 5 inches, 0.375 inches thick acrylic plate base which was designed to support the reactor and for the convenience of installing and removing the reactor. The reac- tors were sealed using a No. 13 1/2rubber stopper on the top and a No. 7 rubber stopper at the bottom. Tygon tubing (0.375 in. I.D.) was used as the effluent line and also served as the gas vent line. An inverted Y-connector was installed in this gas-liquid line about 10 inches above the top of the reactor to separate the gas and liquid. The effluent liquid line (0.375 in. I.D.) was then connected from the side branch of the Yeconnector to the influent of the next reactor. The gas line (3’16 in I.D.) was connected from the straight branch of the connector to the gas collection cylinder. The gas volume produced was measured by displacement of an acid brine solution. The brine solution consisted of 10 1 NaCl and 2 1 “2504 which has a very low solubility for gases. Another, smaller Y-connector was installed in the effluent line of each reactor so that liquid samples could be with- 77 drawn easily without disconnecting the tubing between two reactors. A slotted angle frame with dimension of four feet long, 1.5 feet wide, and 6.0 feet high was built to support all experimental reactors. Individual reactors can be easily installed by slipping the base plate into guide rails mounted on the supporting plate. As shown in Figure 3-2, the reactors designated as No. 1 and No. 2 in the top view are the anaerobic filters. The other reactors sur- rounding the supporting frame, designated from No. 3 to No. 10, are the high solids, packed reactors. The effluent line of each reactor was connected by a plastic, quick release connector to the influent of the next reactor, so that all reactors were connected in series. The liquid phase was pumped at a constant rate from the liquid reservoir to the anaerobic filters from which it flowed by gravity through the packed reactors and back to the liquid reservoir. The liquid flow in each reactor was upward. The liquid flow pattern in the reactor system was always from the reactor with the oldest substrate through a series of packed reactors to the one with the newest sub- strate, and than into the liquid reservoir. Effluent from the liquid reservoir was pumped into Filters No. 1 and No. 2 and the flow circula- tion pattern was completed by introducing the effluent of Filter No. 2 to the oldest reactor. The gas collection system was composed of gas cylinders, a brine solution network, and a brine reservoir (Figure 3-1). Ten gas collec- tion cylinders for 10 packed reactors were made from 1/3 inches thick, 2.0 in. LB. acrylic pipe, each Cylinder having an effective volune of 800 ml. A two liter plastic graduated cylinder served for gas collec- tion from Filter No. 1 and a one liter plastic graduated cylinder was 78 used for Filter No. 2. All gas cylinders and the brine reservoir were connected together by tygon tubing (5’16 in.) so that brine displaced from the cylinders could be stored in the reservoir. A protection dev- ice, consisting of a check valve (one way flow into the reservoir) and a by-pass tube with a stop valve, was installed at the effluent port of the brine reservoir for the purpose of preventing loss of brine due to possible accidental loosening of the tygon tubing. All gas cylinders were mounted to aluniniun supporting rods by moveable clamps. When gas production was measured, the individual gas cylinders were moved so that level of brine inside the cylinder was at the same height as that in the reservoir. Therefore, atmospheric pressure could be maintained every time the gas volume was measured. A one liter, glass aspirator bottle was used as the liquid equali- zation reservoir. This reservoir was connected between the newest substrate packed reactor and Filter No. 1. Besides acting as an equal- ization basin, this bottle held enough volume of liquid for daily sampling. The liquid reservoir sat on a magnetic stirrer to keep the liquid well mixed. A constant voltage transformer and a variable voltage regulator were used to control the positive displacement piston punp (Model PR-GZO by Fluid Metering, Inc., Oystering Bay, N.Y.) fer a desired con- stant flow rate. A photograph showing the entire experimental apparatus is given in Figure 3-3. 79 Figure 3-3 Experimental System For High Solids Anaerobic Fermentation And Anaerobic Filtration of Wheat Straw. 80 3.2 W The entire experimental apparatus was kept in a walk-in constant temperature room. The temperature was always maintained at 36 t 1°C during the entire 18 months of experimental work. The experimental program was divided into three stages. The first stage was the time for conceptual studies, system debugging and modification, and the improvement of analytical techniques, as well as the study of the per- formance of packed reactors and anaerobic filters. The second stage was the extended study of the performance of packed reactors and anaerobic filters; the results obtained from this period were used for mathematical model development. The third stage was a study of the. performance of the anaerobic filter under transient loading. Experimental methods performed in this research will be described in this section. 3.2.1 MW Baled wheat straw was obtained from the straw storage room belong- ing to the MSU Department of Animal Science. Straw was first chopped with a communiting machine (Model D, by ‘W. J. Fitzpatric Co.) into pieces about 0.5 inches long and dried at room temperature. The mois- ture content of the chopped straw was measured by oven drying at 103° C. The average moisture content was found to be 5.67% (S. D. = 0.11). 81 3.2.2 W The experiment was first started with only one reactor (later designated as the No. 1 anaerobic filter), packed with 60 grams of wheat straw, and the liquid reservoir in the system. After connecting the influent line of the reactor, 450 ml of distilled water were added into the reactor, and the top rubber stopper with the effluent line was installed. No nitrogen or other inert gas was applied to try to purge oxygen. The liquid reservoir was filled with 1,000 ml of distilled water. About 50 ml of active anaerobic digester sludge obtained from the Mason Wastewater Treatment Plant, Mason, MI., was injected into the reactor. Liquid was circulated through Reactor 1 and the liquid reser- voir. Three days later, the second reactor (anaerobic filter No. 2) was added into the system by using the same procedures. Tubing was reconnected to let the effluent fran Reactor 1 flow into the bottom of Reactor 2, and effluent from Reactor 2 went into the liquid reservoir. Another 50 ml of digester sludge was also seeded into Reactor 2. Every three days thereafter, one more reactor was added into the system until Reactor 12 was installed. Besides daily gas production and pH measurement, no other sample was taken from the system during this period. Several chemical reagents including NaHC03, NaOH, and NHuHCO3 were used to adjust pH value to around 6.8 in the liquid reservoir. After 36 days, Reactors 1 and 2 were moved to the middle part of the supporting frame and stayed in that position serving as the anaerobic filters thereafter. Also, Reactors 11, and 12 were disconnected from the system and acted as long term batch reactors. The position of reactors at that stage is shown 82 in Figure 3-4. After Reactors 1 and 2 started to serve as the anaerobic filters, about 100 ml of active digester sludge was injected into these two filters, and the pH of the liquid reservoir was monitored. On the 36th day, a new reactor with 60 grams fresh straw was installed to replace Reactor 3, since, at that time Reactor 3 was the oldest one in the sys- tem. The effluent line of Reactor 3 was connected to the liquid reservoir and the effluent line of No.10 was reconnected to the influent line of Reactor 3. Also the effluent of Filter No. 2 was reconnected to the influent line of Reactor 4. In this way, as men- tioned earlier, liquid flow was always fran the oldest reactor to the newest reactor. Three days later, Reactor 4 was replaced by a new reactor in the same manner. The substrate input interval of three days was maintained until Reactor 10 (the oldest packed reactor in the system at that time) was replaced by a new substrate reactor. During this period, samples were taken for volatile fatty acids, total soluble COD and pH. However, samples were only taken once a day and the sampling times were not con- sistent. It was latter found that sampling frequency and sampling time were both crucial fOr this research. Every time a new reactor was added, about 50 to 100 ml distilled water was added into the liquid reservoir to make up the amount of liquid lost from the system as the result of daily sampling. After two cycles were completed for a three—day substrate input interval, the substrate input interval was changed to five days until the end of the first stage experiments. 83 \l G 0 0 . Figure 3—4 Reactor System at Stage I Experiment 0 0 7 0 0 6 Figure 3'5 Reactor System at Stage II and Stage III Experiments. 84 The pump flow rate was calibrated from time to time. Unfortunately, however, it was found, after using the first pump for three months, that the flow rate was not constant due to the leaking valves. The actual flow rate in the system was lower because the dis- charge pressure was higher than when the punp was removed fran the system for calibration. A new punp head was then ordered but about two months were lost during the pump failure period. The new pump head provided a fairly constant flow rate and was able to maintain a high discharge pressure. This pump was used continuously until the end of the experimental program. Liquid samples were first filtered through 0.45 m Millipore filters for soluble (DD measurenent and for volatile fatty acids. A microscale colorimetric COD method (HACH CO.) was applied for COD meas- urement (see further discussion in section 3.4.2). 3.2.3 Wimp: Six packed reactors and two anaerobic filters were involved in this stage's experiment (Figure 3-5). New substrate was input at the same time at intervals of three days. The substrate input and output, and influent and effluent tubing connection methods were the same as that used in the first stage experiment. The temperature was also kept at 36 t 1 oC. Samples were collected two to eight times in a day from the effluent of the packed reactor and anaerobic filters for 00D and volatile fatty acid analysis. Gas production and gas composition were measured daily. The fermented straw was dried in the 103°C oven and weighed to compare with the initial weight (60 grams). The dried straw 85 was then stored in a refrigerator for cellulose, bani-cellulose, and lignin analysis at a later date. A constant voltage transformer and a variable voltage regulator were installed to control the punp flow rate. The punp flow rate was checked daily and it was maintained at a fairly constant flow rate (26 ml/hr). Data were obtained during this stage and a mathematical model for substrate degradation in the packed reactors was developed. The liquid reservoir always stayed in the sys- tem serving as the equalization basin for the anaerobic filters and also retaining enough liquid volune in the system for daily sampling Chemical buffering reagents such as NHuHCD3 and NaOH were occasionally used when pH adjustment was necessary. 3.2.4 Wheat The liquid reservoir was essential for this research because a certain amount of liquid volune had to be retained in the systen for daily sampling. However, in a full scale reactor system, it would be benefical to reduce the overall cost by reducing the system's total volune. This could be partly achieved by excluding the liquid reser- voir from the system. Therefore, it is important to examine the performance of anaerobic filters under transient loading, i.e. under the condition when the liquid reservoir is not in the system. A 20 ml glass U-tube was installed between the metering punp and the effluent line of the newest packed reactor to replace the liquid reservoir. The reason for installing the glass U-tube was that an amount of liquid had to be maintained on the suction side of the punp to prevent a possible "dry punp". Thus, the effluent liquid from the 86 newest packed reactor with sharply increased COD concentration could be directly introduced into the filters. Because the liquid reservoir was not in the system, the amount of liquid volune in the system was not enough for a long term experiment and the third stage experiment only lasted for two weeks. 3.3 MW Methods for sample analysis including pH, COD, volatile fatty acids, gas composition, cellulose, hemicellulose, and lignin will be described in the following sections. 3.3.1 Mammal; A Fisher Accunet Model 325 expanded scale research pH meter with Corning semi-micro combination electrode was used for pH measurement. Before the pH as a sample was measured, the electrode was first cali- brated against a standard buffer solution with pH = 6.98 at 35 0C (pH meter and buffer solution were placed in the constant temperature room). The pH value was then read to 0.01 unit inmediately after the sample was withdrawn from the system. When the electrode was not in use, it was submerged in a pH = 4.02 buffer solution. The standard buffer solution was changed every week. 87 3.3.2 Wm The colorimetric method with micro digestion procedures (Hach Co., Loveland, Colordo) was used for measuranent of total soluble 000. This method needs only 2.0 ml sample for digestion. The 000 vials along with COD reagent, were first purchased from Hach. Later, the frequent- ly used OOD reagent was made by the author according to the recipe provided by Hach as shown bellow: H2804 I<20r20z A330,, HgSOu 2.5 ml 0.0245 g 0.03 g 0.03 g The chemical quantities listed above are for one OOD vial only. A batch of reagent for 200 vials was made at one time. The procedures for COD reagent preparation are as follow: 1. Disolve 6.0 grams of silver sulfate in 500 ml of concentrated sulfuric acid. 2. Weight exactly 4.90 grams of anhydrous potassium dichromate and mix into H2804 solution prepared in step (1) until com- pletely dissolved. 3. Transfer about 0.03 grams of mercuric sulfate into each cleaned COD vial. 4. Use repipetor to pipet exactly 2.50 ml of solution prepared from step (2) into each OOD vial prepared from step (3). 5. Store COD vials with reagent in the refrigerator. A Bausch & Lomb SP—20 spectrophotometer was first used for the 88 measurement of percent transmittance by directly inserting the digested COD vial (made by HACH) into SP-20. However, it was found that COD vials did not have uniform circumference and that the inside diameter for different vials was not identical. Therefore, COD measured during the first stage period were not precise and were not useful, these data were not used for analysis but were only used as a reference. The COD measurement technique were improved by using precision test tubes (by Bausch& LOInb CO.) and replacing the SP-20 by a SP—7O spectrophotaneter. Good calibration curves of COD vs percent transmittance were obtained, as shown in Figure 3-6, by using the improved method. This new method was then used for stage two and stage three experiments. Methods of sample preparation for OOD measurement and the pro- cedures for micro-digestion, and spectrophotometric measurement are described in the following: 1. Filter 2.0 ml of fresh liquid sample through 0.45 p m filter paper. 2. Accurately pipet 1.0 ml of filtrate into 10 m1 volumetric flask and add distilled water to the mark. 3. Gently shake the volunetric flask to completely mix the dilut- ed filtrate. 4. Turn on the COD reactor (Hach, Model 16500) to preheat to 150 degree C. Remove the cap from a OOD vial. Holding the vial at a 450 angle, carefully pipet exactly 2.0 ml of diluted fil- trate sample into the vial. 5. Replace the cap and tighten cautiously. Hold at the cap and invert the vial several times to»mix the sample with the COD 89 coca 5.“: :8 8252538 mo“. MES 22:25.2”. 91m use: AF\mEV Qzxo 4 Qz< Goo madman—ow .._<._.o._. Huc $58; QHN mma mmH Amaozv mz_p sea oma mm m: :N o “1-1.-. -....- q .. _ I So 32 3.523 1.1 new @533 p38. “NH 1 r ' eouoocm IIIIloIIIIIi II n .02 i . .-..-Lm..i..i.-- 4 Louoemm m .02 qfiuuq—d-u4J-uqfid—_—dd——uu——-_--u~—---—--—u—--fi—I—-—-q———-—:du--——quuu—duqd-—-—:——dd-—--qd—qd-u—--q o8 <.._> m .02 O D l O I I Louuoom w .02 O JllllllllllflllllllllllllllllilliLLILllljllalIJHIIALLIJJ (1/3 ) Ovaaq N39AXO Tvolwanj 109 approximate the soluble COD concentration for the points after the peak concentration obtained from four packed reactors. A FORTRAN program CUVFIT was written to handle the matrix manipulation and curve fitting procedures. This computer program was executed both on the CDC CYBER/750 mainframe computer and a DEC PDP—11/23 minicomputer. Program CUVFIT can be found in Appendix E-1. Figure 4-2 shows the experimental data points from four packed reactors as well as the approximation curve calculated from the 7-degree, least-squares polynomial equation generated from CUVFIT as shown below: u 2 2 y = 1.u3o1 x 10 - 3.7852 x 10 x + 7.272u x -2 3 -4 u -6 5 - 7.5897 x 10 X + “.5712 x 10 X - 1.5929 x 10 X 9 7 - 6 -12 + 2.9847 x 10 X - 2.3287 x 10 X in which y soluble COD, (mg/l). X = time, (hr). If the experimental data are plotted on a log-log scale, as shown in Figure 4-3, it can be seen that COD concentration in the packed reactor is decreasing logarithmically with respect to time after the peak occurs. The characteristic of the sharply increasing COD concentration during the first 12 hours resulted from rapid leaching of soluble and colloidal substances associated with the wheat straw. The subsequent slow COD production was contributed by microbial degradation of the straw. This leaching from the straw was verified by pumping distilled 110 CNN .mmo»o 1759205 We Miami : axe: .855”. 535 7: Se: Eéemmsm zoEv m2: cm mm om ma ma 3H NH oa w m s N L1I4m. u 1;. T“ I d _ _ . o 67:33 0 o .. l . . .. 8N cwcowaoea . . . . - . a L 8: .0 . o . .. . 1 com _L Axum uesmu< I J coca l I OQNH l m—onE‘nm l GOSH w eouoemm emgpo l m eouommm O 1 coma T L coma (3.11339 59 l/fiw) mIlVHlNEDNOj 013V E'lIlV'IOA 119 (as acetic acid) at about 12 hours after new substrate was introduced while propionic acid and acetic acids peaked about 18 hours later at 660 mg/l and 1800 mg/l as acetic acid, respectively. About 60 hours after substrate input, the butyric acid concentration had decreased to less than 50 mg/l (as acetic acid) and then gradually declined to as low as 6 mg/l. This suggests that the butyric acid produced at the earlier time has been washed out or been converted partly to acetic and propionic acids. Also the concentration of iso—butyric acid (Appendix D-u-2) at 6.5 hours was higher than the concentrations of propionic and butyric acid. By 20 hours the iso-butyric acid concentration had dropped to less than 20 mg/l. This suggests that the straw surface . contains small amounts of substances, possibly a certain kind of amino acid, fOr instance, valine, which is readily converted to iso-butyric acid (Barker, 1961). The total of the volatile fatty acids, expressed as COD, is presented in Appendix D-4-2 and plotted in Figure N-l. Comparing the VFA COD curve in Figure fl-l with the total soluble COD curve shows that VFA were produced at a slower rate than total soluble COD. The VFA COD reaches a peak concentration approximately 2N hours after the peak for total soluble COD. These characteristics are further evidence that the initial soluble COD production in the packed reactor was not contribut- ed by microbial activity but rather by leaching. 120 14.1.3 W The main function of the packed reactors was the production of soluble COD from solid substrate, not the production of methane gas. However, a anall quantity of methane was produced daily from each packed reactor. Figure 4-8 shows the composition of methane and carbon dioxide of several packed reactors, and Figure 4-9 shows the cumulative gas production from a single packed reactor in Stage I. As shown in Figure H-9, more carbon dioxide than methane was produced from the packed reactor. Figure fl-8 shows that the methane content gradually increased from a low level, about 15%, at the beginning when the packed reactor was installed. By Day 8, the methane content had increased to a value higher than the percentage of carbon dioxide where it remained until Day 31. At Day 31 the packed reactor had moved to the position as the second oldest reactor in the system, and at Day 36, it was the oldest reactor in the system. Examining Figures fl—B and 4-9 reveals that the methane content is decreasing and the total gas production rate is increasing after Day 31. These phenomena can be explained by two observations. First the pH value in the influent liquid, which is the effluent of Filter No.2, was higher than the pH inside the last two packed reactors; the drop of pH resulted in more carbon dioxide escap- ing from the liquid phase. Secondly, a larger bacterial population may have developed in the last two packed reactors than in the other reac- tors due to the carrying over of microorganisms from Filter No.2; higher microbial population in the two older reactor resulted in the higher gas production rate. The change of pH versus time of a single packed reactor from the first day of installation until the last day 121 meopuemm cmxoea H mmepm cw cowuwmoasoo mew w-e acumen lease mace emmaapm we ow mm mm em mm om mm mm um um cm 3 3 3 NH 3 m o e N o a _ AT _ _ _ a. T. .4 l _ _ _ _ _ Tc _ _ a I O] o o 1. Cl 1. n1 0 0 ,9 O OxouuzAT.sT.:orn. . r10 M. o. .\\. o. o. nor/VTwr. “w I o.Av o..u o 0 IL 0 o 10:41:96.6! ‘ o I a“. o 15. o o no 0 M xrho 0 I” it 0 ll 0 0 J o o o o .19 1| “IQ .l IIOJ fil NS 1.6:... w 7. emu 110' 01 oH om om ow om a an ow co (%) 0011;50dw03 sea Hd 122 .Louommm umxoma opacwm a See» cowpozuoea mew m>wpe_:E=u a-e wesmwm Azmov weep ummam_m cc ow mm Nm wN «N oN ma NH m e o l__lfia__ll o oofi CON cam ooe com com ooN com com e p _ _ _ _ _ _ Te _ oooH (lw) uoiionpoud $99 aALielnan 123 when it was removed was also plotted in Figure 4-8. The pH of a new packed reactor was as low as 5.5 and then gradually increased to about 6.8 on the last day. The initial low pH value was due to the high con- centration of volatile fatty acids produced from carbohydrate fermentation. Hydrogen was also detected in first day after a new packed reactor was installed. This finding agreed with the theoretical concepts dis- cussed in Section 2.1.1 that hydrogen would be produced during the first stage (substrate hydrolysis and organic acids fonmation) of anaerobic fermentation. Methane produced in this period mostly came from the reduction of carbon dioxide by using hydrogen (see section 2.1.“.2). The daily production of methane and carbon dioxide was calculated based on the normalized percentages for CH“ and C02 only. Since the nitrogen gas was assumed to result from the air originally present in the reactor. The total amount of methane producted from the packed reactors was about 15 T of the total methane produced from the entire system. Table 4-2 shows the volume of methane produced from anaerobic filters and packed reactors in two long term periods in Stage I. 124 Table 4-2 Comparison of Methane Production in Packed Reactors and in The Entire System (Stage I) Methaneimductimumll Eemntage Dates Days Packed Filter Total Packed Filter 1982 Reactors 1+2 System. Reactors 1+2 (1) (2) (3) (4) (5) (6) (7) 1/21 - 3/10 48 10676 58900 69576 15.3 84.7 3/29 - 5/15 47 8269 46684 54917 15.1 84.9 The above values were calaulated based on the data obtained from the extented experimental period. Therefore it is reasonable to assune that the packed reactors can produce about 15 1 of the total methane that is produced from the entire system. This information will be used for data analysis in the latter section. However, the information given above is not suitable for use to estimate the daily methane pro- duction from a specific packed reactor; for instance, if the daily methane production from the filters is 1000 ml it is not always true that the packed reactors will produce 150 ml of methane in the same day. 4.2 Won The average volume of liquid stored in the liquid reservoir during Stage II was about 1.0 liter for a detention time of about 1.6 days. Every 3 to 6 days, a measured amount of distilled water (50 — 100 ml) was added to the reservoir to maintain approximately the same liquid volune. As mentioned before, pH control chemicals such as NH4HPO3, 125 and NaOH were also added to the liquid reservoir when pH adjustment was necessary. Liquid samples taken from the reservoir were analyzed for total soluble 00D and volatile fatty acid concentrations. The change of individual volatile acid concentrations in the reservoir with respect to time are presented in Figure 4-10, and the total soluble COD and total VFA 00D versus time are plotted in Figure 4-11. By comparing Figure 4-7 with Figure 4-11, it was found that the variation of VFA COD inside the liquid reservoir was not as significant as in the packed reactors. Although Figure 4-1 and Figure 4-11 are not plotted with the same scale and therefore are not directly comparable, it still can be seen that the total soluble COD has been equalized by . the reservoir to a large extent. The difference between the highest and the lowest COD concentration, was about 1300 mg/l or about 19 1 of the average soluble COD concentration. The maximum difference of total soluble COD concentration for the influent and the effluent of a packed reactor was greater than 5000 mg/l as can be seen from Figure 4-1. It is apparent that the liquid reservoir played an important role as an equalization basin that reduced the variability of COD before it was brought into the anaerobic filters, preventing possible damage to the anaerobic filters due to shock loading. The hydraulic retention time in the liquid reservoir during Stage II was approximately 1.67 days which is not long enough for acid util- izing methane bacteria to grow. Although methane gas was produced in the liquid reservoir, the amount of CH” produced, 450 ml in 40 days period, was not significant when compared with the gas produced from the anaerobic filters. Another important function of the liquid reser- voir was that it stored enough liquid volune for the necessary 126 .mHo>mmmmm OHDGHJ z— onkHh 4HQZH oH!cmm:me mpPDm . .1 Dow ulzolaoma n .. o_»wu< o .. r. IJOONH T a L 83 h 2 l L .. m. z e. z m. z _ HT _ _ _ _ _ _ _ _ h h _ _ Doom (D?130v vv 3/6w) NOIIVHINBDNOD 127 wH .Lwo>emmme emacHH we now scum Scum» mpecaPo> use Doc m_a=_0m .acos cassette He-e messed NH mH mH «H mH NH HH memo OH 1.0 U) c A~w\M\oHv m . _ _ Doc alga Hesse mpasapo> O ecu manpom Hmuop O 0 O O 0 O O O O o oooH oooN ooom oooe ooom coco oooN Doom ooom (l/5w) 003 128 sampling. The arrow marks in Figure 4-11 indicate the times when a new packed reactor was installed in the systen. The drop in COD concentra- tion when a new reactor was added is due to dilution from the make up water added at the same time. Valeric and iso—valeric acids were found in the liquid reservoir in concentrations ranging from.15 to 30 mg/l as acetic. The presence of these acids which are rarely found in the packed reactors suggests that further degradation of organic substrate was taking place in the liquid reservoir. The change of pH in the liquid reservoir was also plotted in Figure 4-12. It should be noted that buffering chemicals were used occasionally to adjust pH in the liquid reservoir. 4.3 111W As mentioned in Chapter One, the anaerobic filters used in this study were up-flow reactors in to which straw was packed and served as the filter medium. Because the majority of the microorganisms were attached on the surface of the medium or trapped within the intersti- cial void spaces, a long biological solids retention time was able to be maintained. Due to the special substrate input method, the packed reactors received substrate intermittently and produced effluent with highly variable COD concenetration. Although the variability was reduced by the liquid reservoir, the effluent from the liquid reservoir exhibited a cyclic COD concentration fluctuation, as can be seen from Figure 4-11. Therefore, the influent soluble 00D concentration of the anaero— bic filters was never at a constant state but was changing periodically 129 .H momum :H eopommg emxoma ummeHo we» can .LHo>cmmme cHaaHH .meoHHHH oHaoemmce Ho oucmgo In .N. oz ecu H. oz mgwuHHu Ho :oHpHmogEoo mam mcmgawz NH-¢ mezmHa H>emmma eHzeHH map soeH HamsHHHm Ho goo «Ha=_0m HAHSH mH-H mezmHi open \0 LO c Hmm\m\oHV m emuHHa N.oz o emuHHm H.oz 0 $3.83”. 33.3.. o m oz v.02 £062 T'VS) 6898 (% _ o oooH oooN ooom oooe ooom oooo coon ooow ooom (I/fiw) 003 annIOS 19101 140 the average effluent COD concentration for Filters No.1 and No.2 were 3729 mg/l and 3047 mg/l, respectively. It is noted that the average effluent COD concentration of Filter No.1 was computed from. the data after 10/11/82 when this filter had stabilized. Therefore, the average COD removal efficiency for Filter No.1 can be calculated as follows: 6636 - 3729 100 : 4 .8 4- 6636 x 3 % ( 1) E1 (1) = If the two filters are considered as a unit, then the total COD ranoval efficiency become: E (z) — 6636 ' 3047 100 - 54 1% (4 2) t ’ 6636 x ’ ° ’ The average COD loading of Filter No.1 was: 6636 (mg/l) x 26 (ml/hr) x (l/ml) x 24 (hr/day) x 1000 435 (ml) COD/da = 9.52 _ g y (4-3) liter reactor volume or _3 9.52 x 2.205 x 10 59“ lb COD/day (4 4) 0.03531 ’ 103rt3 ' The total average COD loading for two filters together was: 6636 x 0.026 x 24 g COD/da = 4.76 , y (4-5) 430 x 2 liter or _3 4.76 x 2.205 x 10 lb COD/day (4-6) = 2 0.03531 97 103 ft3 141 -3 * 1 gran = 2.205 x 10 lb. 1 liter: 0.03531 ft3 In view of the results obtained from the above calculations, the overall removal efficiency of 54.1 1 does not seem significantly large. The reason for the high effluent COD is that non-biodegradable GOD is leached from the straw and accunulates due to recirculation, stabiliz- ing at a value determined by the rate at which liquid is removed due to sampling and replacement of packed reactors. As shown in the next sec- tion, degradation of volatile fatty acids was nearly complete indicating that the filters were not overload. It should be emphasized that the anaerobic filters were designed to produce methane from the organic substances contained in the liquid phase and not necessarily to produce a high quality effluent since the effluent is not being dis- charged from the system. 4.3.2.2 WW Data obtained both from Stage II and Stage I will be presented for VFA COD removal efficiency. 4.3.2.2.1 W The change of VFA COD concentration for the influent and effluent of the anaerobic filters is presented in Figure 4-16 (data in Appendix D-4-4). It can be seen that the curve of effluent VFA 00D for Filter No.2 still shows a declining trend, which indicates that the microbio- logical population in the Filter No.2 was still developing during that 142 HHH mampmv .mewuHHH oHnoemecm use eHo>emmwe qucHH ecu Ho coo uHom HuueH mHHueHo> pcozHHHm mH-e meamHa memo b'86) 5 .. mMHDHa N .02 o mMHHHI H .02 .II m_o>mmmmm QHDGHD o O / IO zIII oooH OOON Doom ooov (I/Pw) 003 PIDV aIIIPIOA 143 period. In other words, the COD removal efficiency of Filter No.1 was increasing and eventually would reach a level that would make the Filter No.2 unnecessary. As shown in Figure 4-16, the average influent VFA COD concentra- tion of Filter No.1 was 3044 mg/l and the average concentration of effluent from Filter No.2 was 48 mg/l. Using the same method for the calculation of total soluble COD removal efficiency, described earlier, the VFA COD removal efficiency for the two filters was 98.4 1. For the purpose of comparing the performance of the two filters, the last four days of Filter No.1 were used to compute the average of 253 mg/l. Therefore the VFA COD removal efficiency for Filter No.1 was 91.7 1.. The VFA COD loading for Filter No.1 was 4.37 g COD/day/liter or 273 lb COD/day/lO3 ft3. And the total VFA COD loading for two filters together was 2.19 g COD/day/liter or 136 lb COD/day/lO3 ft3. It is noted that the VFA COD produced in the filters was not included in the above calculation. 4.3.2.2.2 W As has been pointed out in Chapter 3, the original pump used in the early period of Stage I failed and it was replaced by a new pump. The experimental results to be presented in this section were obtained from the 35 days of "stable" data from 4/7/82 to 5/10/82 (Day 229 to Day 263) after the new punp was used. From Figure 4-17 and Appendix D-4-5 it was found that the average VFA COD concentration of the influent liquid for Filter No.1 was 5356 mg/l and the effluent VFA COD concentrations were 682 mg/l and 125 mg/l 144 HH manomv.mCmHHHl uHaoemacq e=< CHo>tmmmm eHaoHH seen HamsHHHm Ho coo eHu< Hosea SHHHaHo> HH-H emstHi HHSDV mEHH mm em Nm om mN 0N eN NN ON NH H: 3 NH 3 w o e N ..“mfiluul'”: .loll - . oI‘l II III. I o o o 1‘. . |I|III"III'III'IIII. Ilnl|ll Inil|fil|nl lllll ..I. IL I. g .v ) Z ) 9 6 E 8 I. I9 I. .l. I. .V II I. w CC w .0 ..D J % / atm / ( l ( l T ..mHHH... oHnomeé N.oz o L 1 L3:... oHaocmSE H.oz o J T cHoimmmm 3:55 o J _ H _ _ F H H H H _ _ H P r _ _ H (I/6w EOIX) 000 PIDV K1193 3111910A 145 for Filters No.1 and No.2 respectively. Therefore, the average VFA COD ranoval efficiency for Filter No.1 was 87.3 5, and the total removal efficiency fbr two filters was 97.7 1. Furthermore, the COD loading, in terms of VFA COD, fOr Filter No.1 in this stage was 7.94 g VFA COD/liter/day or 496 1b COD/day/lO3 ft3, and the total VFA COD loading for two filters was 3.97 g VFA COD/liter/day or 278 lb COD/day/103 ft3. 4.3.2.3 Wu The above calculations for the COD removal efficiency may be sum- marized as in the tables shown below. Table 4—4 shows the influent and effluent COD concentrations of the anaerobic filters; Table 4-5 summar- izes the COD removal efficiency for Stage II; and Table 4-6 shows the COD removal efficiency for the Stage I. Table 4-4 Influent and Effluent COD Concentrations for The Anaerobic Filters (Stage II) Influent Effluent COD Removed Filter No.1 Filter No.2 (1) Total Soluble COD, mg/l 6636 3047 3589 (2) VFA COD 3044 48 2996 Difference (1) & (2) 3592 2999 593 146 Table 4-5 COD Removal Efficiency in the Anaerobic Filters (Stage II) COD Loading COD Removal (lb COD/Day/lOOO ft3) Efficiency (1) Soluble COD VFA COD Soluble COD VFA OOD Filter No.1 594 , 273 , 43.8 91.7 (9.52) (4.37) NO.1+ No.2 297 , 136 , 54.1 98.4 Filter (4.76) (2.19) * unit in g COD/day/liter Table 4-6 Volatile Fatty Acid COD Removal Efficiency in Stage I Anaerobic Filters COD L d' 1b COD/day R 81 Eff' i (1) ca ing 103 ft3 emov lc ency Filter No.1 496 ‘ 37.3 (7.94) No.1 + No. 2 278 97.7 filters * (3.97) * unit = g COD/day/liter As mentioned earlier, the influent liquid to the anaerobic filters contained non-biodegradable and biodegradable substances which included particulate COD and soluble COD, and the soluble COD included VFA COD and non-VFA COD. Examination of Tables 4-4 and 4-5 shows that 3589 mg/l of soluble COD was removed by the anaerobic filters while only 147 2996 mg/l of VFA COD was removed. This indicated that 593 mg/l of soluble COD, about 16.5 1 of the degradable soluble COD, was fermented to volatile acids and further converted to methane and carbon dioxide inside the filters. In other words, about 16.5 1 of non-VFA COD in the influent stream was converted to VFA COD in the anaerobic filters, and this portion of the COD was not included in the calculation of VFA COD removal efficiency. If this 593 mg/l is added to the influent VFA COD then the removal efficiency becomes: (593 + 3044) - 48 593 + 3044 X 100 : 98.7 % (ll-7) for both filters in Stage II. It is concluded that at a total soluble COD loading of 4.63 g COD/liter per day (289.12 lb. COD/day/103 ft3), for two anaerobic filters, 98.7 1 of volatile fatty acids were convert— ed to CH”, C02 and microbial cell solids. The fact that a volatile acid COD removal efficiency of greater than 90 1 was obtained suggests that efficient methane production may be accomplished at even higher COD loading rates once sufficient micro— bial mass is established in the anaerobic filters. Also by comparing the removal efficiency and the operating paraneters of this study, such as type of substrate, COD loading, hydraulic retention time, with the results obtained from previous studies (Table 2-15), it is found that the anaerobic filters conducted in this study performed better than those of many previous studies. It also concluded from Table 4-5 that 81 1 of the soluble COD remoyal was accomplished by Filter No.1. These finding agreed with the results reported by several other investigators (Young and McCarty, 148 1967; Lovan & Force, 1971) that the major portion of the COD is removed in the lower section of an anaerobic filter. 4.3.3 W A rough picture of the fate of volatile fatty acid degradation in the anaerobic filters can be observed by determining the individual acid concentrations of the effluent stream. Tables 4-7 and 4-8 give the concentrations of individual acids that were measured after 10/6/1982. The data in these two tables are plotted in Figure 4-18. Table 4-7 Individual Volatile Fatty Acid Concentrations in Filter No.1 Effluent. (Stage II) Volatile Acid Concentration (mg/l as HAc) Date Time HAO HP iHB HB iHV HV 10/6 1000 433 367 10 57 17 18 10/7 1000 399 550 40 126 _, ._ 10/8 1000 131 230 9 93 -- -- 10/9 1400 189 280 17 2 22 14 10/10 1030 150 270 13 -- l6 -- 10/11 1000 133 185 8 -- 12 4 10/12 0400 145 189 9 -- l4 6 10/14 2200 167 122 6 -- ll 4 10/15 1000 106 85 1 1 5 2 10/16 1200 76 84 2 -- 4 2 10/18 1300 77 51 2 -- -- -- ' Stage II, Year 1982. 149 Table 4-8 Individual Volatile Fatty Acid Concentrations in Filter No.2 Effluent. (Stage II) Volatile Acid Concentration (mg/l as HAc) Date Time HAO HP iHB HB iHV HV 10/5 1600 103 1.5 -- -- -_ -- 10/6 1000 58 0.6 -- -- __ __ 10/7 1000 55 2.5 -- -- -- _- 10/8 1000 32 -- -- _- -- -- 10/9 1400 27 -- -- H- -_ __ 10/10 1030 22 -- -- H- -_ _- 10/11 1000 30 1.0 _- -- -- -_ 10/12 0400 35 -- -- -_ -- _- 10/14 2200 6 -- __ -- __ -_ 10/15 1000 8 -- -- -- -_ -- 10/16 1200 5 —- -- H- _- _- 10/17 1200 43 28 -_ -- -_ -_ 10/18 1300 47 -- -- -_ -- _- * Stage II, Year 1982. AS has mentioned in Section 4.3.1.2, the original Filter No.1 was replaced by a spare filter. The new filter was acclimated for two weeks before VFA COD measurements were resumed. However, the new Filter did not achieve stable operation until the final days of Stage II. This can be observed from the continuously decreasing concentra- tion of volatile fatty acids in the effluent of Filter No.1. The VFA COD escaping from Flter No.1 was removed by Filter No.2. Therefore, the over-all VFA COD removal capability of the anaerobic filters was not affected. As shown in Figure 4-18, the effluent acetic acid concentration in Filter No.1 decreased sharply from above 400 mg/l to about 150 mg/l and then decreased at a slower rate. Figure 4-18 also shows that butyric acid was not effectively removed before 10/8 and the propionic acid 150 H: 833 Nd... ES: 92 H.oz HE: 5 2225258 2% EH: 5539 .5222 OT: e58: mHaa H~m\m\eHv ON O T. L T. l OOH T l I . . . I 98 HI L OOm T 25% ESE N62 0 1 Sufism c T 22288... n 1 OOO I 252 ES: H.oz o J 2 F _ H F _ _ H r H r _ _ 8m (311932 59 [/fiw) NOIlVHlNBDNOj 151 concentration was higher than that of acetic acid before 10/13 in Filter No.1. These phenomena suggest that a longer time is required for HZ-producing acetogenic bacteria to grow to a sufficient microbial population to completely degrade the propionic and butyric acids to acetic acid than for degradation of acetic acid. L103.“ 11.: .l 9 .2 - 1 ‘ .. .‘1 -.°- .1 911-." .... Biodegradable organic substances that contribute to methane pro- duction include particulate COD and soluble COD. Although liquid samples were not analyzed for particulate COD, the portion of particu- late COD that was degraded in the anaerobic filters may be estimated by theoretical calculations. As previously described in Section 4.3.1.2, the average daily methane production from both anaerobic filters in Stage II was 1052 ml/day. The theoretical maximun methane production per unit COD destroyed is 0.35 l/g COD at standard conditions, equivalent to 0.396 1 CHu/g COD at 36°C (see Appendix C calculation). If assumed that 98 1 of the COD destroyed was converted to methane with the remaining COD used for cell growth, then the amount of the total biodegradable COD destroyed by the anaerobic filters was: 1 1 0.98 x 0.628 1.052 (lxday) x (8/1) x (day/1) 0.396 = 4.32 (gm/1) = 4320 (mg/l) (4-8) From Table 4-7, the total soluble COD removed by the anaerobic filters was 3589 mg/l. Therefore, the theoretical particulate COD removed by 152 the filters was: 4,320 - 3,589 = 731 (mg/l) (4-9) The above computational results are only an estimation and they are only valid providing the following conditions are valid: (1) the average liquid flow rate was equal to 0.628 llday (S.D. = 0.074 l/day), (2) no methane produced diffused out or leaked out from the gas collec- tion system, (3) the percent of methane composition measurements were accurate, and (4) 98 1 of the total COD destroyed were converted to methane. Of the factors which affect this calculation, the consistency of the pump flow probably has the grestest impact. The standard devia- tion of the average punp flow rate was 0.074 l/day which was 11.8 1 of the average flow rate. Therefore, the variation of flow rate could substantially affect the computational result. 4.4 Waugh The substrate used in the packed reactor was un-pretreated, chopped wheat straw. The extent to which this substrate was degraded has been studied and will be presented in terms of three parameters: (1) the substrate weight loss after fementation, (2) the percent of cellulose and hemi-cellulose degradation, and (3) by calculation from methane production. 153 4.4-1 BMW After a packed reactor was ranoved from the system at the end of its full retention period, the fermented straw inside the reactor was carefully transfered to a 185 mm dia., 765 ml evaporation dish. The fermented straw was placed in a 103°C oven for two days to dry. It was then cooled and weighed. The weighed sample was ground and stored for analysis of cellulose, hemi-cellulose, and lignin at a later date. The moisture content of the fresh straw was 5.671 (S.D. = 0.11). The straw weight loss after fermentation was obtained by subtracting the fermented weight from the dry fresh straw weight. Table 4-9 shows the weight loss of fermented straw from 3/6/82 to 5/28/82 (Day 197 to Day 280) in Stage I. Table 4-10 shows the percent of straw weight loss in Stage II. 154 Table 4-9 Straw Weight Loss After Fermentation. (Stage I) Date Days Reactor ‘Weight Percent weight Loss (g) Loss (1) ‘ (1) (2) (3) (4) (5) 3/06 - 4/18 43 5 14.79 26.13 3/11 - 4/28 43 6 16.02 28.30 3/18 - 4/28 41 7 17.08 30.18 3/23 — 5/03 41 8 15.27 26.98 3/28 - 5/08 41 9 15.65 27.65 4/03 - 5/15 42 10 15.15 26.77 4/08 - 5/20 42 3 16.19 28.60 4/14 - 5/24 40 4 14.12 24.95 4/18 - 5/28 40 5 15.98 28.23 Average 15.58 27.53 S.D. 0.87 1.53 Fresh Substrate Weight : 60.0 x 0.9433 = 56.6 grams (5) = (4)/56.6 x 100 Year : 1982 Table 4-10 Straw Weight Loss After Fermentation. (Stage II) Date Days Reactor Weight Percent Weight Loss (g) Loss (1) (1) (2) (3) (4) (5) 9/28 - 10/16 18 5 10.44 18.45 10/01 - 10/22 21 6 12.51 21.10 Average = 20.28 * Fresh Substrate Weight = 60.0 x 0.9433 = 56.6 grams (5) = (4)/56.6 x 100 155 The weight loss is apparently a function of solids residence time. The weight loss in Stage I with a solids retention time of 40 to 43 days was 1.36 times the weight loss in Stage II when the solids reten- tion time was only 18 to 21 days. From Section 4.3.1.3, the methane production per gram straw input was 104.3 ml/g in Stage I and 76.3 ml/g in Stage II for a ratio of 1.37, essentially the same as the weight loss ratio. 4.4.2 WOW The Goering and van Soest (1970) method, as described in Chapter 3, was used to analyze the cellulose, hemi-cellulose, and lignin con- tents for fresh straw and fermented straw samples. Table 4-11 gives the results of the fiber analysis for Stage I and Stage II. The fresh straw samples were analyzed at the same time as the fermented samples. 156 Table 4-11 Cellulose, Hemi-cellulose, Lignin Contents of Wheat Straw Date Days Cellulose Hemi- Lignin Cellulose (1) (1) (1) Fresh Straw 43.64 30.25 6.70 43.64 29.75 6.9 44.46 31.63 5.69 43.22 33.03 6.78 Average 43.74 31.17 6.52 S.D. 0.52 1.48 0.56 Stage I 3/06 - 4/18 43 43.30 29.47 10.90 41.48 31.61 10.75 3/11 - 4/23 41 41.40 30.01 11.44 40.20 31.50 11.14 3/18 - 4/28 41 41.47 28.85 11.06 39.51 30.84 11.35 3/23 — 5/03 42 43.28 30.36 10.74 41.75 31.79 11.01 4/08 - 5/20 41 43.41 31.20 10.92 41.79 32.70 10.04 Average 41.75 30.83 10.94 S.D. 1.30 1.17 0.39 Stage II 10/16 - 11/5 20 46.11 30.55 9.35 45.83 30.82 9.20 Average 45.97 30.69 9.28 Because lignin is almost nonbiodegradable, it is reasonable to assume that the lignin content in the straw was not changed during the fermen- tation period. Therefore, based on the 6.52 1 lignin content for the fresh straw, the percent of fiber contents in the fermented samples may be adjusted as in Table 4-12. 157 Table 4-12 Adjustment of Straw Fiber Contents (Stage I) Cellulose Hemi- (1) Lignin Ash Other (1) Cellulose (1) (1) (1) (1) Fresh Straw 43.74 31.17 6.52 4.90 13.67 Fermented (2) Before Adjust 41.75 30.83 10.94 8.22 8.26 Fermented (3) After Adjust 24.88 18.37 6.52 4.90 4.93 (4) Difference (1) - (3) 18.86 12.80 0.0 0.0 8.75 Assume 501 of the materials under "other" item.are degradable, then the total material degraded in Stage I was 36.041 (18.86 + 12.80 + 8.75 x 0.5) of the initial dry weight. The percentage of cellulose and hemi-cellulose degraded were 43.11 (18.86/43.74) and 41.11 (12.8/31.17) respectively. Using the same approach, fiber contents for the second stage sam- ple can be adjusted as shown in table 4-13. 158 Table 4-13 Adjustment of Straw Fibrous Contents (Stage II) Cellulose Hemi- (1) Lignin Ash Other (1) Cellulose (1) (1) (1) (1) Fresh Straw 43.74 31.17 6.52 4.90 13.67 Fermented (2) Before Adjust 45.97 30.69 9.28 6.97 7.09 Fermented (3) After Adjust 32.29 21.56 6.52 4.90 4.98 (4) Difference (1) - (3) 11.45 9.61 0.0 0.0 8.69 The amount of material degraded in Stage II is therefore equal to 25.401 (11.45 + 9.61 + 8.69 x 0.5) of the initial dry weight. The per- centage of cellulose and hemi-cellulose degraded in this stage were 26.21 and 30.81 respectively. Stage I to Stage II is then equal to : 4.4.3 .Bencent_o£JSAxflu3u2LIknu3nEuainljknsuiJnlltuaLEEAance .Calculation Considering the packed reactors, liquid reservoir, filters as one unit, The ratio of percent degradation in 6.04 - 3 /25.40 - 1.42. and anaerobic and using the simplified sketch shown below, a mass balance can be used to relate substrate degradation and methane production. 159 10 9 8 7 6 5 4 3 ll? 1 2 '5 The total amount of biodegradable COD produced from straw degrada- tion in the system is further converted by microorganisms to fermentative product COD and microbial cell solids COD. The fermenta- tive COD includes organic acid COD in the liquid phase and methane COD in the gas phase. Although the type of COD is changed, the over-all COD should be not reduced. Based on this concept, the extent of sub- strate degradation can be calculated according to the total methane produced from the system. The experimental results presented in Section 4.3 have shown that no organic acids accunulated in the systen. Therefore, most of the organic acids should have been converted to methane and carbon dioxide. It has been pointed out in the previous sections that the majority, about 98 1, of the degraded substrate COD is converted to methane while the rest is utilized by microorganisms for cell growth. Under "stable" operating conditions, it may be assuned that the total organic acid COD in the system remains unchanged. Then, based on mass balance concepts, the product COD must be equal to the summation of methane COD and 160 microbial cellular COD as the expression below shows: [Product COD) .-. [ CH” COD] + [Cellular COD) (4-10) In Stage I, the total amount of straw input into the system in the period from 3/29/82 to 5/8/82 (Day 220 to Day 260) was 453 grams (60 x 8 x 94.331) and the total methane produced from the system was 50,038 ml (see Section 4.3.1.3). Thus, the equivalent methane COD can be cal- culated as: 2.525 (8 COD/ 1) x 50.038 (1) = 126.35 gm COD (4-11) Assune y percent of the straw is lost due to degradation and leaching. Since 1.0 gram of cellulose can produce 1.186 gram of COD (Appendix C), and the fresh straw contains 74.91 1 of cellulose and hemi-cellulose, the equivalent product COD is: 453(y) x 1.186 x 0.7491 = 402.46(y) gm COD (4-12) The amount of VFA COD wasted from the system in this period resulting from liquid sampling was 2.76 g COD. Assune 501 of the total soluble COD was VFA COD. And assume 200 ml of liquid was wasted every time the oldest packed reactor was removed and that the total COD concentration of the oldest reactor was 3000 mg/l. Then, the total amount of COD wasted was: 2.76/0.5 + 0.2 x 8 x 3.0 = 10.32 g COD Since 98 1 of the product COD was converted to CH4, the percent of 161 straw degraded can be obtained from the relationship: 126.35 = (402.46(y) - 10.32) x 0.98 y = 34.6 1 (4-13) In Stage II, thelamount of methane produced in the period between 10/4/82 and 11/10/82 was 41,197 ml and the total straw introduced into the system during this period was 509.4 gram (60 x 9 x 94.31). The amount of VFA COD wasted from the system was 7.82 g COD (1.21/0.5 + 0.2 x 9 x 3.0). Again, assuning that K percent of the straw was degraded, and applying the same approach as above, the percent of substrate- degradation (K) can then be calculated from the following expression: (509.4(K) X 1.186 X 0.7491 - 7.82) X 0.98 = 2.525 X “1.197 K : 25.2 % (4-14) The ratio of percent substrate degradation in Stage I to Stage II based 34.6 on this method is equal to: = 1.37, which is the same as the 25.2 ratio of methane production per gram substrate input in Stage I to Stage II. 4.5 WWW W The performance of the anaerobic filters in Stage I and Stage II as well as the extent of substrate degradation in these two stages are sunmarized in Table 4-14. As shown in Table 4-14, the extent of solid 162 substrate degradation obtained by mass balance calculation from methane production was 34.61 in Stage I and 25.21 in Stage II. These two values agreed favorably with the values obtained from fiber analysis. The values obtained from the weight loss measurement were both smaller than the data obtained from the other two methods. A possible reason for the difference may be due to liquid phase solids that adhering to the straw surface and then being included when the straw was dried for weighing. The ratio of the percent of substrate degradation in Stage I to Stage II range from 1.34 to 1.37, that are very close to the number of the ratio of methane production per unit weight of straw input. 163 Table 4-14 Sunmary of The Performance of Anaerobic Filters and The Extent of Substrate Degradation Stage I Stage II Ratio (1) CH4 Production (ml/day) Filter 1 761 (S.D.:158) 925 (S.D.:116) Filter 1+2 1035 (S.D.:230) 1052 (S.D.:149) (2) Specific CH4 Production (a) l/day/l reactor Filter 1 1.75 2.12 filter 1+2 1.19 1.20 (b) ml GHQ/g substrate input Filter 1+2 104.3 76.3 1.37 (3) Extent of Substrate Degradation (1) (a) Weight Loss 27.5 20.3 1.36 (b) Fiber Analysis 36.0 25.4 1.34 (c) Mass Balance 34.6 25.2 1.37 3 (4) COD Loading, lb COD/day/1000 ft ISEQD .YEAQQD .ISQQD .MEAQQD Filter 1 -- 496 594 273 Filter 1+2 -- 278 297 136 (5) COD Removal Efficiency (1) Filter 1 -- 87.3 43.8 91.7 Filter 1+2 -- 97.7 54.1 98.4 TSCOD Total Soluble COD VFACOD = Volatile Fatty Acid COD 164 4.6 W If assume that the nonbiodegradable COD is constant throughout the system and that the soluble COD in Filter 2 is nonbiodegradable, the rate of nonbiodegradable COD (NDCOD) production in the system can be calculated according to the following expression: Rate of NDCOD Average Vol. of New Water Added = X NDCOD Production Day In stage II, 243 liquid samples were taken from the system from 10/1/82 to 11/9/82 (39 days) and the average volume of one sample was 4.5 ml. During this period, nine packed reactors were removed and replaced with the fresh straw. Assume 200 ml of liquid was wasted each time the oldest reactor was removed. Then, the total liquid wasted was: 4.5 (ml/sample) x 243 (samples) + 200 (ml) x 9 = 2894 ml. (4-15) and the mass of NDCOD wasted was: 3000 (mg/l) x 2.894 (I) = 8682 mg COD. Therefore, the average nonbiodegradable COD production rate in Stage II was: 8682 (mg COD) / 39 (days) = 222.6 (mg COD/day) (4-16) The amount of total methane produced from the whole system from 10/4/82 to 11/10/82 (37 days) was 41,197 ml. As has descrbed in Sec- tion 4.4.3, 98% of the degraded COD was converted to methane and the 165 amount of COD wasted from the system was 7.82 grams. The total COD produced in this period was: 2.525 (g con/1 can) x n.1197 (1) = 113 g con (4-17) 0.98 Because the rate of total 00D production was: 113 g/37 (days) = 3063 mg COD/day. Thus, the ratio of the nonbiodegradable COD production rate to the total COD production rate is: 222.6/(222.6 + 3063) : 6.8%. 4.7 i' 0.00 ‘ 0 ‘9-.‘ be ‘ o to ‘1 --0 - ‘ out: In Stages I and II when the liquid euqalization reservoir was con- nected in the system, it has been shown that the anaerobic filters were capable of accommodating the mild fluctuations of influent substrate concentration and temporary periods of inoperation. In order to deter— mine whether the liquid reservoir was essential for the system or whether it may be excluded from the system to minimize the overall cost, the third stage experiment was conducted without the liquid equalization reservoir in the system. Therefore, the anaerobic filters were receiving the highly variable COD concentration directly from the newest packed reactor. This section will present the response of the anaerobic filters to this kind of transient substrate loading. The filter response to loading changes as indicated by total solu- ble COD is illustrated in Figure 4-19 which includes two figures; Figure B represents the change of soluble COD versus time after the liquid reservoir was removed, and Figure A shows the effluent COD con- 166 .mmmpsll osmomm in which: Vj : volume in segment j, (cm3) Aj : cross sectional area in segment 3, (cm2) Q : liquid flow rate, (ml/hr) C- = soluble COD concentration in segment 3, (mg/l) t = time, (hr.) Dij = dispersion coefficient, into segment j, (cmz/hr) D0:] = dispersion coefficient, out from segment j, (cmZ/hr) X = length of one segment, (cm) Pnj = rate of soluble COD production from solid substrate in segment 3, (mg/l/hr), n =1, 2 R. = rate of COD utilization in segment 3, (mg/I/hr) 3013- 3003- The disperSIOn differentiaton terms, AjDij(— 3x ) and AJDOJ(T ), in Equation 5-1 can be simplified by applying the finite difference technique: 172 Segment Segment Segment 3-1 .1 3+1 £5 . £3 . . l XJ_1 _j XJ _J £5XJ+1 _J let: 3 C03 _ Cj+1 ’ CI _ CJ+1 ' CI (5 2) - 1 - _ 3cm“ _ CI ' Cj-I _ CI " 03-1 (5 3) x ' T . . ' " a /2 ( IIJ + xJ_,) L Substitute Equation 5-2 and Equation 5-3 into Equation 5-1, to obtain: (Cj - Cj-1) - AjD0j L + [Pnj - Rj] (5'14) 173 Rearrange the terms in Equation 5-4: acj _ C. I ( Q AjDiJ’ AJDOJ ) at ' J v. L ' L J 1 AJD C 1 AJDlJ + C3 1 —V-( Q + ) + 3+1 T X L J J + [an - R3] (5-5) Now define; A D-. A-D . 1 J 13 J OJ B - = — (-Q - - ) (5-6 ‘3 VJ- L L ) 1 A3903 82, =—v. ( Q + > <5-7) J A.D-. 1 1 B . _-_..__ x ._J_._J_ (5-8) 3J v. L J Substitute Equations 5-6, 5-7, 5-8 into Equation 5-5, to obtain: 3C3 :2 R ( > Y: B13- C3 + sz Cj-1 + 333- CJ '1' nj - 3 5‘9 174 The COD production term, Pnd in Equatin 5-9 has two forms: (1) From the "zero" time when the packed reactor was connected into the system to time "hour 6.0", the COD production term is: (2) After hour 6, the COD production term is: P23 : Sj K23: 13 (5-11) Where fj of straw when water is just added to the reactor due to instantaneous is the soluble COD concentration per unit time per unit weight leaching. While K13, sz, and are 00D production kinetic constants. Equations 5-10 and 5-11 were obtained according to the characteristic curves of the experimental COD measurements shown in Figure 4-1. Therefore, Equation 5-9 actually contains two series of system dif- ferential equations. Since: AJ=A1=A2:°”:A Vj=V1=V2='° :Vt/sz Sj=S1=52=...:St/N=S where: A = cross-sectional area Vt total volune of the straw holding section St total straw weight. 175 And assune: Rj = R1 = R2 = ... = R fj = f1 = f2 = --- = r Dij = D11 = D12 = ... = D1 Doj = Do1 = D02 = --° = DO K1J - x1, = K12 : ... _ K1 K2J - K21 = K22 - -- = K2. 1 ”1 ADo B1 : T(_Q _ L .. L ) (5-12) 1 ADo 1 ADi B3 = T( L ) (5-14) c. 3 J n = 1, 2 in which: A =sr+sx1mgw 64m P2 = 5 K2 t" (5-17) 176 As can be seen from Figure 4-1, the COD production rate was not a constant value. Therefore, Equation 5-15 cannot be simplified by just making = 0. Instead, it has to be solved for the COD concentra- tion for any given time t. Examining Equation 5-15, it is found that the equation contains three unknown variables. Hence, two boundary conditions have to be provided in order to satisfy the equation. If a packed reactor is divided into five segments as shown in Fig- ure 5-1, the first segment, shown with a dashed line and designated Segment 0, is a dummy segment which provides the lower boundary condi- tion for Segment 1; and the upper boundary condition of Segment 5, is given by making C5 = C5 +1. In other words, assune that the influent COD concentration is the concentration in the dummy segment and the effluent COD concentration is the same as the COD concentration inside Segment 5. Thus, there are six system partial differential equations describing the six sements. These equations are readily solved numer- ically by using a computer. The Runge—Kutta method was applied to solve the equations and a FORTRAN program MODEL was written to handle the calculations and to create the data files for computer plotting. The program MODEL is included in Appendix E-2. The three known variables, B1, B2, and B3, can be computed from Equations 5-12 to 5-14 given the basic hydraulic and reactor dimension data. The COD utilization term, R, can be estimated from the amount of methane produced by the packed reactor. The average methane produced from a packed reactor in one solids retention time was about 1.98 T of the total methane produced from two anaerobic filters in the same per- iod of time. From Figure 4-14, the total methane produced from the two filters from.10/10/82 to 10/27/82 was about 17,000 ml. Therefore, the 177 daily COD consunption in a packed reactor can be calculated by: 2.525 I8 C0D/1 ) x 17.0 (1) x 0.0198 18 (day) = 0.0’472 (gm COD/dBY) : 1.967 (mg COD/hr) (5-18) The volume of the straw holding section in a packed reactor is equal to 435 ml. Therefore, COD utilization per unit volume in the packed reac- tor is: 1.976 mg/l = ———- = “052 0.435 hr R 4. 2 R. :——- :—5—— (5'19) 3 N N where N : No. of segments in a packed reactor. Before Equation 5-15 can be solved, four constants: f, K1, K2, (1, for Equations 5-16 and 5-17 have to be provided. The value of f can be obtained from Figure 5-2 which is the semi-log plot of the meas- ured effluent soluble 00D concentration versus time for a packed reactor before the COD concentration reached the peak point. If we let F = St f, the interception point of the straight line and the Y-axis in Figure 5-2 gives the value of F which is approximately equal to 1,000. Because St = 60 grans, then f = F/60 =16.7. The exponentional con- stant, ar, in Equation 5—17 can be obtained from the slope of the straight line in Figure 4-3. The value of ¢z is approximately - 0.23. The other two constants, K1 and K2, were determined by trial and error. The best values of K1 and K2 as well as the final values of f and A- mmmumv Couommm cmxuma ammzmz seem . cowumgucmocou coo mpazpom ucwzpewm Pavemewcmaxm NIm meaowl Agnosv wave mH oH 178 m N o m e m m g oooA _ _ _ a _ l _ A I 88 .L soon I 88 l 88 I 88 J 82 I 88 II ooom J oooofi oooHH (I/fiw) 003 aIRflIos 19101 179 1: are those that result in the best fit of the experimental data by Equation 5—15. The dispersion coefficients, DO and Di, were assumed identical and their values were selected from the literature (Motta, 2 1976) as 0.00482 CM/hr. The program MODEL was executed several times on a DEC PDP-11/23 minicomputer until the best fit curve was obtained and the best values of K1 and K2 were determined. Program MODEL was designed to be interactive so that the variables of Equation 5-15 could be changed conveniently from the terminal screen without exiting and restarting the program. Good results were obtained from the model, since the soluble COD _ concentration calculated from MODEL agreed with the experimental data very well. Figure 5-3 demonstrates the computational results from MODEL compared with the real experimental COD data to evaluate how well the results computed from the model agreed with the measured data. The final parameters for Equations 5-16 and 5—17 were found to be: f = 16.7 (mg/l/hr'gm) c1 = - 0.238 K1 : 9.86 (mg/l/hr‘gm) K2 : 7.98 (mg/l/hr'gm) R = 4.52 (mg/l/hr) Other input data for I‘DDEL can be found fran Appendix E—2. 180 .mucmsmcsmmme Foucmspgmaxw sage new Ammo: cog» cmcpmuno mgouoomg coxom; co copumgucmocoo coo mpazpom scmzpweu MIm mesopm .888 .88“ @301 ZH mZHH .88 a .8 8m .88 .8v .8 ___L________e__T________._______;L I .888 a I .888m .I .888m I .888 8. I .888m I .8888 . I dean fll .8888 I .888m I883 II .882 I.888Nd 'I/EIN NI 003 181 5.2 W Most mathematical models, including the one presented in this chapter, are established under certain assumptions. Therefore, their applications are subjected to some limitations attributed by their fun- damental conditions. The assumptions described in the preceding section limit the application of this model as well. Among those con- ditions, the two factors most affecting the computional results when MODEL is used are: (1) the liquid flow characteristics, in that the flow rate was assumed strictly steady, was identical everywhere inside the straw holding area, and no short circuiting existed in the reactor; and (2) the straw substrate input conditions, such as substrate solids retention time and substrate solids concentration. The two COD produc- tion kinetic paraneters, K1 and K2, were determined, and are only valid for a substrate input interval of three days, a straw weight of 60 grams, and a packed reactor divided into five segments. In order to test the suitablity of the values of K1 and K2 in MODEL when the packed reactor is divided into a different number (N) of segments than five, the same values of K1 and K2 as obtained from N = 5 were used to run mDEI.. The nunber N was varied fran 2 to 6. Figure 5-4 shows the theoretical soluble COD concentrations for five different conditions, the lowest curve (N = 2) represents the effluent soluble COD concentration when the reactor was divided into two segments, and the highest curve, six segments. It can be seen that using the same values of K1 and K2 resulting in similar curves for different segment nunbers. Therefore, it can be concluded that the two COD production parameters obtained from this study were suitable for various segment 182 .mgm353: ucmsmmm maopem> um Ammo: >5 meouumme umxomq Co mcopumeacmocoo oou mpazpom umumspgmm e m mezmam mDOI ZH mzHH 88a .8va .88a .88 .8 fi__..____E 32;; ___..E_._____ E: .. Elk-4 .P.pb-P--b-b:. (1‘41 ‘I/CJN NI 003 183 nembers (N > 1) in MODEL. The model presented in this chapter is still a prototype. Once the relationships between different hydraulic retention times and sub- strate solids retention times with the values of K1, K2, and a: are developed in the future, MODEL will be able to predict the effluent COD concentration of a high solids packed reactor at any hydraulic reten- tion time and solids retention time without running wet chemical measurements, and can be applied to associate with other automatic pro- cess control devices, such as pH control. CHAPTER SIX ENGINEERING APPLICATION The experimental results of the proposed process have shown the technical feasibility of producing methane from un-pretreated wheat straw. This chapter will discuss the application of this process to a full scale system. 6.1 '0 1-2.! t!!.°-'.;'i. '. 'I' 1‘ .- ‘-. I‘v. 0‘ For full scale operation, there are three possible basic flow con- figurations for the packed reactors: (1) counter-current series; reactors connect in series with the liquid flow opposite to the solid substrate movement as studied in this project, (2) co-current series; reactors connect in series but with the liquid flow in the same direc- tion as the solid substrate movement, and (3) parallel; packed reactors connected in parallel as shown in Figure 6-1. O O O O O O O Figure 6-1 Schematic diagram Of para11e1 Connected reactor system 184 185 6-2 WW Among the three reactor systems metioned above, the last two types have not been experimentally studied. The likely nature of their per- formance will be disscused and compared based on the knowledge obtained from the studies of the first reactor configuration. Chmparsion of the operation for the three reactor systems will be made in terms of the following parameters: (1) the need for a liquid reservoir, (2) the extent of solid substrate hydrolysis, (3) the extent of leaching, and (4) operational simplicity. 6.2.1 W The first configuration of counter-current, series connected reac- tors always has the liquid phase passing through the newest packed reactor last so that the high 00D and low pH liquid can be removed shortly after it is produced. As previously mentioned, a liquid reser- voir can be used to equalize the highly variable 00D before it is introduced into the anaerobic filter. The second configuration, on the contrary, has the newest packed reactor located at the begining of the liquid flow stream. Thus the high COD liquid flows fran the newest reactor through a series of packed reactors that will provide addition- al liquid volume to dilute the high 00D fluid due to dispersion. Therefore, the liquid reservoir can more easily be eliminated from the co—current reactor system. The third system, packed reactors connected in parallel, may require a liquid reservoir downstream of the packed reactors to equal- ize the peak of high COD liquid when a new packed reactor is added to 186 the systen and to simplify the operation procedure. 6.2.2 - ‘ - ‘ ‘ - . - .- s -- Q . A. . .. - !\ 0 0 0 '0 1. \ . l '_ .1 . The extent of solid substrate hydrolysis is affected by several factors such as pH, microbial population, and reaction time. The pH in a single packed reactor for the counter-current system increased gradu- ally from 5.5 when new to 6.8 when removed as shown in Figure 4-7. In the co-current system, the pH values in all the packed reactors are expected to be lower than 6.0 because the newest reactors are located at the upstream end. The high 00D and VFA's produced in the newest packed reactor would be augmented as the liquid flows through the other reactors. Therefore. the pH would remainn low, probably not higher than 6.0. In the parallel system, the initial pH value in a new packed reac- tor would fall between 5.5 and 6.0; then it would quickly rise to a higher value because the influent to each reactor would always be high pH liquid from the anaerobic filter. Since the best pH range fOr sub- strate hydrolysis is between 6.2 and 7.4, the parallel reactor system would have the most favorable pH range for substrate hydrolysis. Effluent from the anaerobic filter is likely to contain active bacteria which can be partly retained in the packed reactors. The par- allel connected reactor system has every packed reactor directly connected to the anaerobic filter. Therefore, packed reactors in this type of system would contain more bacteria and hence provide a higher degree of hydrolysis than the two types of series connected reactor systems. 187 If the nunber and the size of packed reactors were the same for the three types of reactor systems, the total solids retentation time in each system would be the same for all three cases. As a summary of the above discussion, packed reactors connected in parallel should provide the highest degree of substrate hydrolysis fol- lowed by the counter-current series systen. 6.2.3 Thu-3W Although the initial production rate of soluble 00D by leaching from a new packed reactor may be affected by the liquid flow rate through the reactor. the total amount of COD produced by leaching should not be affected by the flow conditions or the initial 00D con— centration in the influent liquid. Therefore, the extent of leaching from the solid substrate should be independent of reactor configura- tion. 6.2.4 WW As has stated in Chapter 3, when a new reactor is installed into the series reactor system, the connecting piping between the liquid reservoir, the newest reactor. the oldest reactor, the second oldest reactor, and the anaerobic filter (see Figure 3-5) have to be discon- nected and then reconnected in the new positions such that the newest reactor is always located last in the flow stream and the effluent from the anaerobic filter can always flows into the oldest reactor. Unlike the laboratory system in which flexible tubing can be used and only two tubes (one influent and one effluent) are needed, the full scale system 188 has to use rigid plastic or metal pipes and four pipes (two influent and twOIeffluent pipes) with associated valving are required for each packed reactor to connect to the liquid reservoir, the anaerobic filters and the adjacent packed reactors. However, fOr the parallel reactor configuration, the position of the newest packed reactor in the system is not important and only two pipes (one influent and one effluent pipe) are needed. A packed reactor needs only to be isolated with two valves while being,emptied and refilled; it does not need to be relocated in the series flow pattern. Therefore, the piping, valv- ing, and operational procedures are simplier when a parallel configuration is used. 6.3 W In lieu of disposal, the treated substrate may be used for pulp and paper making in which case it can be transported directly to the pulp and paper making facilities as a slurry without needing drying. The treated straw can be also used for animal bedding after the straw is dried. If the treated straw is not to be further utilized, it may be disposed by landfill or by spreading on farm land. 6-4 W The experimental results have shown that 65% of the soluble 00D produced from a new packed reactor was contributed by rapid leaching of straw. Therefore, an alternative process would be to operate the packed reactors at a short solids retention time, to produce a high COD liquid simply by leaching without temperature control. The effluent 189 from the packed reactors can then be introduced to the temperature con- trolled anaerobic filter for conversion toImethane gas. 6.5 W Based on the above disscusion, the proposed process can be applied to a full scale system, using any of three types of reactor configura- tions. A sunmary of the comparsion between three reactor systems is shown in Table 6-1: Table 6-1 Comparision of The Three Types of Reactor Systems Counter-current Co—current Parallel Series Series Liquid Needed Not Needed Needed Reservoir Extent of Intermediate Lowest Highest Hydrolysis Extent of Same Same Same Leaching Operation More More Simpler Simplicity Complicated Complicated According to the information shown above, the parallel system is most suitable for a full scale system.due to its simpler operation and greater extent of substrate hydrolysis. CHAPTER SEVEN CONCLUSIONS The following conclusons about the coupled high solids anaerobic fermentation and anaerobic filtration of cellulosic residues can be drawn from the results of this research. 1. The proposed process of coupled high solids fermentation and anaerobic filtration has been successfully operated for an extended period of time. The use of packed reactors has suc- cessfully overcome the difficulties of handling substrate at the very high concentration of 34.4%. Recirculation of the liquid phase provided the buffer capacity to prevent pH inhi- bition of methanogenesis. The major function for the packed reactors was the leaching and hydrolysis of the solid substrate and the production of organic acids. In addition, about 15% of the total methane production occured in the packed reactors, largely it is presumed, from utilization of the hydrogen produced during acid formation. Initially, soluble COD production from a packed reactor is contributed mostly by rapid leaching of the straw. The effluent total soluble COD from a packed reactor reaches a peak concentration in about 12 to 14 hours after fresh sub- strate is introduced. After the peak occures, the soluble COD concentration decreased logarithmically with respect to time. 190 u. 191 After the initial leaching, microbial activity was responsible for organic acid formation and further slow substrate degrada- tion. The major fatty acids found in the packed reactor were acetic, propionic, and butryic acids. Acetic acid had the highest concentration, accounting for 55% of the total VFA COD, fOllowed by propionic (31%) and butyric acids (14%). Other volatile fatty acids persisted in small concentrations of less then 20 mg/l. The volatile fatty acid COD was produced at a slower rate than total soluble COD in the packed reactors, reaching a peak con- centration approximately 24 hours for total soluble COD. This is additional evidence that initial soluble COD production was contributed by leaching rather than microbial activity. The liquid reservoir served an important role as an equaliza- tion basin that reduced the variability of COD, preventing possible damage to the anaerobic filters due to shock loading. The hydraulic retention time in the liquid reservoir during Stage II was approximately 1.6 days which is not long enough for acid utilizing methanogens to grow as evidenced by negli- gible gas production in the liquid reservoir. Another important function of the liquid reservoir was to stored enough liquid volume for long term sampling. The anaerobic filters were the major methane generators, pro- ducing 85% of the methane from the entire system. The highest specific methane production rate for one anaerobic filter was 2.12 liter CHu/day/liter reactor volume during stable opera- tion. 8. 10. 11. 192 Total methane production per unit weight of substrate input was 104.3 ml methane/g substrate added in Stage I and 76.3 ml CHu/B substrate input was obtained in Stage II. It is clear that the system. produce more methane per gram of substrate input in Stage I than in Stage II. This was caused by the longer straw retention time of 40 days in Stage I compared with 18 days in Stage II. Therefore, a longer solids reten- tion time would provide higher methane production per unit weight of solid substrate input due to increased straw degra- dation. The VFA COD removal efficiency for Filter No.1 + No.2 in Stage II was 98.4% at a loading rate of 136 lb VFA COD/day per 103 ft3 (2.19 g COD/day/liter) with a hydraulic retention time of 34 hours. The total soluble COD removal efficiency in Stage II for both reactors was 54.1% at a loading rate of 297 lb COD/day per 103 ft3 (4.76 g COD/day/liter), with the hydraulic retention time of 34 hours. Since the volatile fatty acids produced in the packed reactors were almost completely converted to methane and carbon dioxide in the anaerobic filters, no volatile acids were found to accumulate in the liquid phase. Therefore, the rate limiting step for the entire reactor system was the hydrolysis of solid substrate rather than the methanogenic step. The extrenely high VFA COD removal efficiency in both experi- ments suggests that efficient methane production may be accomplished at yet higher COD loading rates or at shorter hydraulic retention times. 12. 13. 193 More than 80% of the total soluble COD removed and over 88% of the volatile fatty acid COD removed was accomplished in Filter No.1. These data suggest that Filter No.1 was nearly large enough by itself. Filter No.2 essentially acted as a backup reactor which removed the remaining COD escaping from Filter No.1. With the recirculation of the liquid phase, the residu- al COD could be removed on the next pass. The extent of straw substrate degradation was 34.6% in Stage I and 25.2% in Stage II obtained by mass balance calculation from methane production. These two values quite favorably agreed with the results obtained by fiber analysis, 36.0% in Stage I and 25.4% in Stage II. The extent of degradation obtained by weight loss measurements were 27.5% in Stage I and 20.3% in Stage II, lower than the data obtained fran the other two methods. This discrepancy was most likely attributed by dissolved solids in liquid phase retained on the fermented straw when it was dried for weighing. The percent of cellu- lose and hemi-cellulose degrdation were 43.1% and 41.1% respectively in Stage I, and 26.2% and 30.8% respectively in Stage II. 14. When the liquid reservoir was excluded from the system so that the anaerobic filters received a transient substrate loading, the effluent soluble COD from Filter No.1 rose sharply. This result indicated that the existing microbial population was not able to completely utilize the sudden increase in sub- strate concentration. However, the effluent pH did not drop to a value lower than 6.5, suggesting that the methanogenic 15. 16. 194 bacteria could sustain methane production during transient loading. A mathenatical model was developed based on the mass balance concept expressing all paraneters in terms of equivalent COD. The model includes terms for inflow, diffusion, OOD production and COD utilization in a series of reactor segments. The results obtained from. the mathematical model agreed very favorably with the experimental data. This model can be used to predict the soluble COD concentration inside the packed reactor as well as the effluent soluble COD concentration, provided the proper constants are given. The proposed process can be applied to a full scale system using any of the three types of reactor configurations: (1) counter-current series, (2) co-current series, (3) parallel. The parallel system is most suitable for a full scale system due to its simpler operation and greater extent of substrate hydrolysis. CHAPTER EIGHT SUGGESTIONS FOR FUTURE RESEARCH As a result of this investigation, several ideas are suggested as possible topics for future research: 1. Investigation of the behavior of different substrate solids concentrations, different solids retention times, and dif- ferent liquid flow rates in the packed reactors and anaerobic filters in order to determine the relationships between COD production and hydraulic retention time, the extent of sub- strate degradation and solids retention time. Also, more kinetic parameters can be determined to refine and extend the mathematical model. Improvement of anaerobic filter design to determine the optimun organic loading. Operation of packed reactors at lower (or ambient) tempera- tures and at short hydraulic retention times to study the feasibility of methane production from leachate of residues. Further investigation of the effects of transient substrate loading on the performance of the anaerobic filters over an extended period. Investigation of suitable pretreatment methods for lignocellu- losic substrate to enhance the extent of substrate degradation. Study of COD production from different types of cellulosic residues for comparsion with wheat straw. 195 APPENDICES APPENDIX A LIST OF SYMBOLS Reactor cross-sectional area in segment j Parameter in mathematical model Parameter in mathematical model Parameter in mathematical model Microorganism decay coefficient Soluble OOD concentration in segment j of packed reactor Dispersion coefficient, into segment j of packed reactor Dispersion coefficient, out from segment j of packed reactor Rate of substrate utilizaton Constant for mathematical model Hydraulic retention time Maximum rate of substrate utilization Constant for mathematical model Constant for mathematical model Constant for mathematical model Longitudinal segment length of packed reactor Number of segments in one packed reactor Nonbiodegradable COD OOD production term in mathematical model Liquid flow rate Rate of COD utilization in packed reactor Solids retention time Substrate concentration Substrate weight in segment j of a packed reactor 196 197 Total substrate weight in one packed reactor Time Packed reactor volume in segment j Total packed reactor volume Volatile Fatty Acids Longitudinal length of one segment in packed reactor Active microbial mass in system Growth yield coefficient Constant for COD production term in mathematical model Limiting minimun solids retention time Solids retention time Appendix B Mathematical Calculation for the Mass of COD Produced by Leaching (A) (B) Wat Area under OOD~Time curve in Figure 4-4 (by using Simpson's Rule) m hr. = 99,273 3 liter Flow Rate = 0.0365 liter/hour Mass of COD Produced = 99,237 x 0.0365 = 3623 mg COD Weight of straw in the reactor : 60 grams. mg OOD Mass COD per gram straw = 3623/6O = 60.4 —— gm straw Hydraulic retention time = 12.5 hours Elapsed time of flow through leaching test = 48 hours T : 48/12.5 = 3.84 W The maximun soluble COD concentration in Figure 4-5 : 2,880 mg/l Liquid volume in the batch reactor : 1,500 m1 Mass of COD produced = 2,880 (mg/l) x 1.50 (liter) 4,320 mg COD Weight of straw = 50 grams 4320/50 35.4._EEL£EEL_ gm straw Mass of COD produced per gram of straw 198 199 (C) W Flow Rate = 0.628 liter/day Hydraulic Retention Time = 17.2 hours Elapsed time = 17.2 x 3.84 (from A) = 66 hours Area between solid curve and dashed curve in Figure 4-2 hr. (by Simpson's Rule) in 66 hours = 213,463 mg iter Mass of COD produced per gram of straw : 0.628 1 mg COD 213,463 x _ = .1____ 24 x 60 93 gm straw APPENDIX C COD CONVERSION FACTORS COD equivalent of the volatile fatty acids, CH”, and H2 can be calculated as the methods described bellow. C.1 .AQeLiQ_AQid Molecular Weight: 60.05 CH3COOH + 2 02 ————> 2 CO2 + 2 H20 (C-1) COD = 2 x 16 x 2 = 1.066 g/g HAc 60.05 C.2 PLODiODiLAcid Molecular Weight: 74.0801 CH3CH2COOH + 7/2 02 -——> 3 CO2 + 3 H20 (0-2) COD - 16 x 2 x 3'5 - 1 512 g/ HP ’ 74.0801 ' ' g C.3 .BDL¥£19_AQid Molecular Weight: 88.1072 16 x 2 x 5 COD = = 1.816 HB 88.1072 g/g 0.4 W Molecular Weight: 102.1343 200 0.5 C.6 At standard conditions one mole of CH“ occupied 22.4 liters, 201 CH3(CH2)BCOQ'I + 13/2 02 ——-> 5 002 + 5 H20 - 16 x 13 2 037 HO ' 102.1343 ’ ' g/g EEDQEi£_AQid Molecular Weight: 116.1613 CH3 6 C02 + 6 H20 COD - 16 x 8 x 2 - 2 204 g/ HC ' 116.1613 ' ° 8 Methane Molecular Weight: 16 16 x 2 x 2 COD = 16 = 4.0 g/g CH” (C-4) (C-5) (C-6) one gram of CH“ = 22'4/16 = 1.40 liters, therefore, one liter of CH4 is 0 equal to I'D/1.4 = 2.86 gram OOD destroyed. At 36 C, one gram of can is equal to 1.585 liters and one liter of CH“ is equal to 2.52 grams COD destroyed. 0.? Human Molecular Weight: 1.008 16 x 2 COD = 1.008 x 4 = 7.94 8/8 H2 (C-7) weeefl-ewl 202 At standard temperature and pressure, 22.4 1 H .-.____=11.11 lite s g 2 1.008 x 2 r Therefore, 1 liter of H2 is equal to 7.94/11.11 : 0.7146 g COD des- troyed. 0.8 W Molecular Weight: 162.0 n (C6 H12 06)!) + On 02 —-9 6n “)2 + 5n H2 0 (C-B) _ 32 x 6n 162.0n COD = 1.185 g/g cellulose Appendix D Data for Figures film 203 Appendix D-3-1 Integrated Area Counts for Volatile Fatty Acids Standard Solution, Data for Figure 3-7. Concentration Area Counts Average S.D. (mg/l) (1) (2) (3) Acetic Acid M.W. : 60.05 1.066 g COD/g HAc 6972.5 907367 870703 865420 881163 22846 3486.2 443550 434930 -- 439240 1743.1 212289 206525 205320 208045 3724 871.6 99825 102064 98313 100067 1887 435.8 49066 47190 48822 48359 1020 217.9 23454 24633 -- 24044 108.9 11792 12031 12031 12020 223 Propionic Acid M.W. = 74.08 1.512 g COD/g HP 4983.8 1072597 1073717 1098282 1081532 14517 2491.9 544382 546451 -- 545417 1246.0 266081 258349 261657 262029 3879 623.0 129712 129688 124916 128105 2762 311.5 64345 61214 63273 62944 1591 155.8 30984 30370 -- 30677 77.9 14699 14602 15236 14846 342 iso—Butyric Acid M.W. = 88.11 1.816 g COD/g iHB 997.1 277484 274885 282389 278253 3811 498.6 140705 140460 -- 140533 249.3 68949 66837 68635 68140 1140 124.6 34427 33963 32842 33744 815 62.3 17081 15946 16604 16544 570 31.2 8292 8045 -- 8169 15.6 3834 3872 4080 3929 132 Butyric Acid M.W. = 88.11 1.816 g COD/g HB 2988.1 791948 809559 818484 806664 13503 1494.1 400403 407570 - 403987 747.0 195524 190216 191265 192335 2811 373.5 97241 97577 89815 94878 4388 186.5 48026 46197 47721 47315 980 93.2 23059 22717 - 22888 46.6 11047 10898 11241 11062 172 204 Appendix D-3-1 Continued Area Counts Average S.D. (mg/l) (1) (2) (3) Concentration iso—Valeric Acid M.W. = 102.13 2.037 g COD/g iHV 997.8 306012 316335 318756 313701 6768 498.9 155961 158818 -- 157390 249.5 77220 75168 -- 75642 1063 124.7 37474 37425 36225 37041 707 62.4 18432 17617 18239 18096 426 31.2 8878 8558 8338 8591 272 15.6 3887 4167 4301 4118 211 Valeric Acid M.W. =102.13 2.037 g COD/g HV 1990.7 607809 633102 640895 627269 17297 995.3 309392 316318 -- 312855 497.7 152659 149024 147831 149838 2515 248.8 72704 73410 70460 72191 1540 124.4 34794 44956 35407 34719 728 62.2 16811 16535 -- 16673 31.1 7757 7754 8185 7399 243 Caporic Acid M.W. : 116.16 2.204 g COD/g HC 994.6 335074 346071 351875 344340 8533 497.7 167969 171050 —- 169510 248.7 81356 78389 77139 78961 2166 124.3 36110 37744 36481 36778 867 62.2 16682 17561 18291 17515 806 31.1 8656 8412 -- 8534 15.6 3581 3602 4000 3728 236 * S.D. = Standard Deviation 205 Appendix D-4-1 Effluent Total Soluble COD From the Packed reactors (Stage II) Total Soluble COD (mg/l) Time* Reactor Reactor Reactor Reactor Reactor Average S.D. (hr.) No. 4 No. 6 No. 7 No. 8 No. 9 (mg/l) 2.0 -- -- 880 1070 902 951 104 4.0 -- 6612 -- -- 6796 6704 6.0 -- 9565 9860 -- 8338 9254 807 6.5 -- -- -- 10695 -- 10695 8.0 9373 9123 9915 -- 9860 9568 384 9.0 -- 10476 -- -- -- 10476 10.0 11315 -- 9591 11662 9395 10491 1164 11.0 -- 10700 -- -- -- 10700 12.0 -- -- -- 12707 9252 10980 19.0 —- 8220 -- -- -- 8220 20.0 -- -- -- 9252 -— 9252 21.0 8798 -- -- -- -- 8798 24.0 -- -- 7762 7628 -— 7695 26.0 -- -- 7650 -- -- 7650 27.0 -- 7024 7474 -- -- 7249 28.0 7548 -- -- 7540 -- 7544 34.0 -- -— -- 7430 -- 7430 36.0 -- -- -- -- 7650 7650 40.0 -- -- -- -- 7303 7303 43.0 -- 6333 -- -- -- 6333 44.0 7240 -- 6382 -- -- 6811 48.0 -- -- -- 6775 -- 6775 49.0 -- 6352 -- -- 6953 6653 50.0 -- -- 6472 -- -- 6472 54.5 -- -- -- 6412 -- 6412 56.0 6892 -- 6412 -- -- 6652 58.0 -- -- -- 6572 -- 6572 68.0 -- 6042 6119 -- -- 6081 68.5 -- -- -— 6432 -- 6432 72.0 -— -- -- -- 6496 6496 75.5 -- 5814 -- -— -- 5814 79.0 -- 5965 - -- -- 5965 80.0 -- -- 5852 —- -- 5852 91.0 -- 5628 -- -- -— 5628 92.0 -- -- 5908 6216 -- 6062 95.0 -- -- -- -- 5891 5891 96.0 —- —- 5683 -- -- 5683 98.0 5536 -- -- —- -- 5536 206 Appendix D-4-1 Continued Total Soluble COD (mg/l) S.D. = Standard Deviation Time No. 4 No. 6 No. 7 No. 8 No. 9 Average S.D. (hr.) Reactor Reactor Reactor Reactor Reactor (mg/1) 100.0 -- -- 6157 -- -- 6157 102.0 -- -- -- -- 5928 5928 106.0 -- -- -- 6099 5991 6045 115.0 -- 5558 -- -- -- 5558 118.0 -- -- —- -- 5836 5836 121.0 -- 5628 -- -- -- 5628 124.0 -- -- -- -- 5836 5836 126.5 -- -- 5758 -- -- 5758 130.0 5628 -- 5814 -- 5781 5741 99 139.0 -- 5390 -- -- -- 5390 142.0 -- -- -- 5600 -- 5600 144.0 -- -- -- -- 5405 5405 149.5 -- -- 5536 -- -- 5536 161.0 -- -- 5445 -- -- 5445 168.0 -- -— -- -- 5426 5426 175.0 -- -- -- 5370 -- 5370 177.0 -- 5053 -- -— -- 5053 186.0 -- -- -- 5481 -- 5481 187.0 -- -- 5354 -- -- 5354 191.0 -- 5247 -- -- -- 5247 198.0 -- -- -- 5076 -- 5076 204.0 -- -- -- 5370 5053 5212 223.0 -— 5070 5127 -- -- 5099 224.0 -- -- —- 5059 -- 5059 235.0 -- 4948 -- -- -- 4948 249.0 —- 4890 —- -- -- 4890 273.0 -- 4707 -- 4573 -- 4640 281.0 -- -— 4510 -- -- 4510 285.0 -- 4414 -- -- -- 4414 321.0 -- 4238 -— —— —- 4238 337.0 -- 4275 -- -- -- 4275 357.0 -- 4210 -- -- -- 4210 381.0 —— 4082 —- -- -- 4082 * Time zero was installation of reactor in the system 207 Appendix D-4-2 Individual Volatile Fatty Acid Concentration In The Packed Reactors. (Stage II) Elapsed Volatile Acids Concentration Total VFA COD Time(hr) (mg/l as Acetic Acid) (mg/l) HAc HP iHB HB iHV Packed Reactor No.8 6.5 816 44 137 84 -- 1849 12.0 1169 223 47 343 -- 2748 20.0 1485 528 4 304 -- 3450 28.0 1743 658 8 237 -- 3935 48.0 1715 572 7 98 -- 3566 58.0 1668 531 8 51 -- 2998 68.5 1361 415 13 58 16 2505 130.0 1097 340 12 27 14 1978 142.0 1017 326 8 22 12 1833 165.0 931 307 5 15 10 1670 188.0 835 275 4 1O -- 1485 214.0 773 260 3 15 -- 1396 238.0 715 252 2 15 -- 1301 263.0 748 260 3 6 -- 1341 Packed Reactor No.9 12.0 1259 374 35 92 -- 2422 22.0 1521 563 12 202 -- 3354 40.0 1510 504 17 103 20 2991 60.0 1488 449 17 68 20 2756 72.0 1558 468 16 107 20 2998 82.0 1320 394 15 48 25 2420 95.0 1183 362 10 31 14 2112 106.0 1109 257 9 12 13 1782 118.0 1091 337 10 20 11 1932 144.0 926 298 5 23 10 1671 168.0 901 306 5 22 9 1658 193.0 800 275 3 20 7 1673 * samples taken from reactor effluent Note: HAc = Acetic Acid HP = Propionic Acid iHB = ios-Butyric Acid HB = Butyric Acid iHV = iso-Valeric Acid VFA = Volatile Fatty Acid 208 Appendix D-4-3 Individual Volatile Fatty Acid Concentrations In the Liquid Equilization Reservoir (Stage II) Date Volatile Acid Concentration (mg/l as HAc) VFA COD 1982 Time HAc HP iHB HB iHV HV (mg/l) 10/5 1600 1347 454 10 182 14 18 2905 10/6 1000 1534 545 10 175 15 18 3263 10/7 1000 1312 486 7 98 11 12 2667 10/8 1000 1476 481 8 204 14 19 3153 10/9 1400 1568 520 10 211 14 21 3349 10/10 1030 1501 482 8 127 13 14 2952 10/11 1000 1546 463 11 191 14 33 3212 10/12 0400 1509 471 6 181 17 32 3157 10/12 2400 1576 487 11 138 19 24 3137 10/13 1200 1416 418 16 57 18 -- 2548 10/15 1000 1403 409 14 175 16 20 2887 10/16 1200 1498 440 15 246 18 -- 3235 10/17 1200 1575 430 16 223 18 23 3262 10/18 1300 1369 403 14 196 16 19 2891 Average 1474 464 11 172 16 21 3044 S.D. 89 41 3 51 2 6 240 * S.D. : Standard Deviation 209 Appendix D-4-4 Effluent COD Concentration From The Liquid Reservoir and Anaerobic Filters. (Stage II) Date Time Elapsed L.R. Filter No.1 Filter No.2 Time TSCOD VFACOD TSCOD VFACOD TSCOD ‘VFAOOD (1) (2) (3) (4) (5) (6) (7) (8) (9) 10/3 2400 0 -- -— -- -- -- -_ 10/4 2200 22 6353 -- -- -- 2728 -_ 10/5 1000 34 7024 -- -- -- 3014 -- 10/5 1600 40 6775 2905 -- -- 2899 113 10/6 1000 58 6940 3263 5229 1445 3014 67 10/6 2200 70 6940 -- —— -- -- -- 10/7 1000 82 6294 2667 4983 -- 3161 70 10/7 1400 86 6003 -- -- -— -- .. 10/8 1000 104 6612 3153 4034 907 2899 39 10/8 1800 112 6472 -- -- -- -- -- 10/9 1400 132 7300 3349 4540 901 3044 32 10/9 1600 134 6633 -- -- -- -- -- 10/9 2100 139 6432 -- -- -- .. .. 10/10 1100 153 6796 2952 4194 776 3146 27 10/10 1200 155 6157 -- -- -- -- .. 10/10 2200 165 6392 -- -— -- -- _. 10/11 1000 177 6372 3212 3611 567 2942 38 10/11 2400 191 6892 -- -— -_ -- -- 10/12 0400 195 6796 3157 3844 602 3015 42 10/12 1300 204 7172 -- -- -- -- -- 10/12 2400 215 7003 3137 -- -— -- _. 10/13 1200 227 6612 2548 -— -_ -- -- 10/13 1800 233 5965 -- -- —— -- .. 10/14 1100 250 6392 -- 3719 -- 3043 -- 10/14 1800 257 6796 -- -- - -- -- 10/14 2200 261 -- -- 3750 472 3161 12 10/15 1000 273 6633 2887 3907 305 2986 20 10/15 2200 285 7172 -- -— -_ -- -_ 10/16 1200 299 6592 3235 3766 264 3161 13 10/16 1600 303 5909 -— -- -- -- -- 10/17 1200 323 7172 3263 3642 247 3384 104 10/18 1300 348 6492 2891 3596 195 3161 50 Average 6636 3044 3729a 253b 3047 48 Standard Deviation 372 240 111 46 150 32 a Calculated from data after hour 177 6 Calculated from data after Hour 273 * 00D Unit = mg/l; Time Unit: Hour 210 Appendix D-4-5 Effluent Volatile Fatty Acid COD Concentration of Anaerobic Filters and Liquid Reservoir. (Stage I) Date Liquid Reservoir Filter No.1 Filter No.2 4/07 5039 458 32 4/08 4762 319 28 4/09 - - 56 4/10 5146 199 252 4/11 4692 1048 138 4/12 5908 -- - 4/13 5333 1394 207 4/14 5042 - - 4/15 5272 - 54 4/17 6038 -- -— 4/18 5532 376 249 4/19 5422 - -- 4/22 6148 - -- 4/23 5613 - -- 4/24 5450 - -- 4/26 5832 477 135 4/19 5119 - 342 5/02 5582 1144 68 5/04 4529 842 35 5/06 5309 199 23 Average 5356 682 125 S. D. 434 453 106 * S. D. = Standard Deviation Data for Figure 4-17 APPENDIX E FORTRAN PROGRAM LISTINGS Appendix E-l PROGRAM CUVFIT 211 PROGRAM CUVFIT LEAST-SQUARES POLYNOMIAL CURVE FITTING FOR EXPERIMENTAL DATA. FOR HIGH SOLIDS ANAEROBIC FERMENTATION AND ANAEROBIC FILTRATION 0F CELLULOSIC MATERIAL BY YOU-MING LIN, OCTOBER 1982, AT MICHIGAN STATE UNIVERSITY *H:******************m*******comnTs ““**************************** A I STARTING MATRIX 0R INVERSE MATRIX DETER = DETERMINENT EPS = THE MINIMUN PIVOT MAGNITUDE PERMITTED ITER - o, READ NEw DATA SET FOR ANOTHER RUN. ITER - 1, READ N,M,EPS,XSTART AND USE OLD x, AND Y VALUES ITER - 2, STOP PROGRAM EXECUTION. IPLOT - 0, NO CURVE PLOTING DATA PILE IS RANTED. IPLOT . 1, DATA PILE(S) WILL BE CREATED. XCUV - 1, TO PLOT ARITHMETIC SCALE CURVE. KCUV - 2, To PLOT LOG. SCALE CURVE. INDIC - NEGATIVE VALUE, DO N*N MATRIX INVERSE ONLY INDIC - 0, N*N+1 MATRIX INVERSE AND SOLVE POLYNOMIAL EQU. COEF. M . DEGREE OP POLYNOMIAL. N - NUMBER OF DATA POINTS. X = VALUES FOR X AXIS. Y - VALUES FOR Y AXIS. PNAME a FILE NAME XSTART' STARTING X VALUE ********************************************************************** OOOOOOOOOOOOOOOOOOOOOOOCOCO PROGRAM CUVFIT IMPLICIT REAL*8(A-H,O-Z) DIMENSION A(41,41),E(41),X(41),Y(41),C(41,41),DX(41) *,IR0w(41),JCOL(41),JORD(41),T(41),CNY(360),TIME(360) LOGICAL*1 FNAME(10) ITER=O ROUNT-l 15 PRINT*,'INPUT N,M,INDIC,EPS,XSTART' READ(5,*)N,M,INDIC,EPS,XSTART IF(N .EQ. 0)G0 TO 99 wRITE(7,2OO)N,M,INDIC,EPS,ITER,XSTART 200 FORMAT(1H1,9X,'N s',14/ln ,9X,'M =',14/IR ,9X *,'INDIC -',14/IE ,9X,'EPS =',E10.2/IE ,9X,'ITER -' *,14/IE ,9X,'XSTART -',F8.3) IP(ITER .NE. 0)Go To 25 PRINT*,'INPUT X(I),I-1,N' READ<5,*)(X(I),I-1,N) PRINT*,'INPUT Y(I),I-1,N' READ(5,*)(Y(I),I-1,N) wRITE(7,210) 210 FORMAT(1H ,///,9X,'GIVEN DATA'IIH ,//,12X,' TIME (HR)',14X,'C O D *(MG/L)') WRITE(7,220)(X(I),Y(I),I-l,N) 220 PORMAT IMPLICIT REAL*8(A-E,O-z) DIMENSION TIME(360).CNY(360) DO 105 IG-ISTP,LAST,INCRM IF(RCUV .EQ. 1)GO TO 300 TIME