LIBRARY Michigan State University PLACE IN RETURN BOX to remove this chockom from your roeord. TO AVOID FINES row monorbdoroddoduo. DATE DUE DATE DUE DATE DUE _E[:! j _W usu chnNflnnlfivo Mien/Equal Opportunity mum * Li fl RESPONSE OF ANAEROBIC SYSTEMS TO LONG-TERM PERIODIC SUBSTRATE PERTURBATIONS By Jian Xing A DISSERTATION Submitted to Michigan State University in partial fulfillment of the requirements for the degree of DOCF OR OF PHILOSOPHY Department of Civil and Environmental Engineering 1996 ABSTRACT RESPONSE OF ANAEROBIC SYSTEMS TO LONG-TERM PERIODIC SUBSTRATE PERTURBATIONS By Jian Xing Three 1.5 L working volume anaerobic chemostats ("daughter" reactors), fed with glucose as sole carbon and energy source, were used to investigate the response of anaerobic systems to long-term periodic substrate perturbations. The daughter reactors were originally operated under constant fed conditions (inlet glucose concentration of 8 g/L and hydraulic retention time of 10 days). After reaching steady-state, the daughter reactors were subjected to periodic applied organic loading patterns in which the influent glucose concentrations were periodically changed in amplitude from 16 g/L to 0 g/L (mineral media only) with time on total cycle times of 2, 6 and 10 days. The average organic loading rates for all systems equal to the steady-state organic loading rate of 0.8 g glucose/L-d. A 10 day hydraulic retention time was maintained for all systems. The anaerobic microbial communities were sensitive to the periodic substrate perturbations. Rapid accumulation of glucose fermentation intermediates occurred immediately after the initiation of the substrate perturbations. The response of the reactor microbial communities to the substrate perturbations can be broken down into four distinct stages: (1) rapid VFA and COD accumulation; (2) establishment of a transient ”pseudo steady-state" at reduced COD removal efficiency; (3) rapid VFA degradation; and (4) re-establishment of steady-state with high COD removal efficiency. The anaerobic systems were able to adapt to the periodic substrate perturbations through changes in microbial communities. This was verified by comparison of DNA extracts and FAME profiles from the steady-state control reactor and the daughter reactors, changes in substrate degradation rates, and microsc0pic examinations. The perturbation interval appeared to significantly affected the structure and function of the perturbed anaerobic communities. Stable microbial communities was only be re- established in two of three systems. The periodic substrate addition mode appears to provide a strong selection force on the populations of the anaerobic chemostat communities. The operational and kinetic performance of the re-established daughter reactors, especially the system operated with 1/1 day substrate perturbation, was superior to the performance of the stable control reactor. This suggests a more robust microbial community structure can be selected and established using this type of substrate perturbation resulting in enhanced treatment efficiency and operational stability. Copyright by JIAN XING 1996 Dedicated to my father Jinzhong Xing, my mother Jia He, and my wife Yi Yu, my daughter Fei and Jing whom I love. ACKNOWLEDGMENT I would like to express gratitude to my major professor and research advisor, Dr. Robert F. Hickey, for his instruction, guidance and encouragement during my Ph. D study. Special thanks are extended to my other research advisor, Dr. Craig S. Criddle for his instruction, guidance and encouragement through out this research work. I wish to express my sincere thanks to Dr. James Tiedje and Dr. Thomas C. Voice for their advice, assistance and valuable comments on this research. Thanks are extended to Dr. Bruce Dale for his time and effort spent on reviewing this dissertation, and for his valuable comments and suggestions. Special thanks are extended to Dr. Poul Loconto, Dr. Wei-min Wu, Dr. Jizhong Zhou, Dr. Dave Odelson, Y. L. Pan, Hung Nguyen, Daniel Wigner, Helen L. Garshow and Mathuram Stanley for their invaluable help and assistance during this work. I also thank my fellow students for _ their friendship, support and stimulating discussions during my years at Michigan State University. This study was funded by a grant from the National Science Foundation Center for Microbial Ecology at Michigan State University under Microbial Ecology Grant NSF BIR 9120006. vi TABLE OF CONTENTS LIST OF TABLES ........................................................................................ xi LIST OF FIGURES ....................................................................................... xiv KEY TO ABBREVIATIONS ....................................................................... xix CHAPTER 1 INTRODUCTION ......................................................................................... 1 CHAPTER 2 BACKGROUND ........................................................................................... 2.1 Anaerobic degradation process .................................................. 2.2 Methanogenesis .......................................................................... 2.3 Acetogenesis ............................................................................... 2.4 Anaerobic fermentation of glucose ............................................. 2.5 Ecological point of view of anaerobic degradation .................... r—r—OOUIWU) mu— CHAPTER 3 CULTIVATION OF ANAEROBIC CULTURE IN LAB-SCALE ANAEROBIC REACTOR FED WITH GLUCOSE AS SOLE CARBON AND ENERGY SOURCE ....................................................................................................... 20 3.1 Introduction ................................................................................ 20 3.2 Materials and Methods ............................................................... 21 3.2.1 Laboratory- scale anaerobic reactor .............................. 21 3. 2. 2 Substrate ....................................................................... 23 3. 2. 3 Nutrients ....................................................................... 23 3. 2. 4 Analytical methods ....................................................... 25 3.2.5 MPN enumeration method ........................................... 25 3. 2. 6 Inoculum material ......................................................... 26 3 3 Results ........................................................................................ 27 3.3.1 Mixing characteristics of the mother reactor ................ 27 3. 3. 2 Reactor start-up ............................................................ 27 3. 3. 3 Reactor operational performance under steady-state conditions .............................................. 33 3. 3. 4 Biomass characteristics ................................................... 36 3.3.5 Mass balance calculation for the baseline steady- -state operation ................................................ 39 3.3.6 MPN enumeration results ............................................. 40 3.3.7 Microscopic examination ............................................. 41 3.4 Discussion .................................................................................. 42 3.5 Summary .................................................................................... 43 vii CHAPTER 4 RESPONSE OF ANAEROBIC COMMUNITY TO A LONG-TERM ONE DAY FEAST-ONE DAY FAMINE PERIODIC SUBSTRATE PERTURBATION... 44 4.1 Introduction ................................................................................ 44 4.2 Materials and methods ................................................................ 46 4.2.1 Anaerobic reactor ........................................................ 46 4. 2. 2 Substrate and nutrients ................................................ 48 4. 2. 3 Analytical methods ...................................................... 50 4. 2. 4 Measurement of maximum substrateconversion rate" 50 4. 2. 5 Measurement of decay coefficients of biomass and substrate utilization capacities ............................ 52 4. 2. 6 Reactor inoculation ...................................................... 53 4. 2. 7 Experimental procedure .............................................. - 53 4. 3 Results ........................................................................................ 56 4.3.1 Effects of the periodic substrate perturbation .......................... 56 4. 3. 2 Gas production during perturbation period ................. 64 4. 3. 3 Maximum substrate conversion rate ............................ 67 4. 3. 4 Decay characteristics ................................................... 68 4. 3. 5 Protein measurement results ........................................ 75 4. 3. 6 Enumeration results ..................................................... 79 4. 3. 7 Microscopic observation ............................................. 84 4.4 Discussion .................................................................................. 86 4.5 Summary .................................................................................... 92 CHAPTER 5 RESPONSE OF ANAEROBIC COMMUNITY TO A LONG-TERM THREE DAY FEAST-THREE DAY FAMINE PERIODIC SUBSTRATE PERTURBATION ................................................................. 93 5.1 Introduction ................................................................................ 93 5.2 Materials and methods ................................................................ 94 5.2.1 Anaerobic reactor ........................................................ 94 5.2.2 Substrate and nutrients ................................................ 94 5. 2. 3 Analytical methods ...................................................... 96 5.2.4 Free energy calculations .............................................. 96 5.2.5 Reactor inoculation ...................................................... 97 5. 2. 6 Experimental procedure .............................................. 98 5.3 Results ........................................................................................ 100 5.3.1 Effects of periodic substrate perturbation ................... 100 5. 3. 2 Changes 1n gas production rate during substrate perturbation period ..................................... 109 5.3.3 Changes in un-dissociated VFA concentrations during the perturbation period ................................... 112 5.3.4 Free energy calculations for anaerobic oxidation of glucose metabolic intermediates ........................... 116 5.3.5 Electron equivalents mass balance .............................. 121 5.3.6 Microscopic observations ............................................ 124 5.4 Discussion .................................................................................. 126 5.5 Summary .................................................................................... 132 viii CHAPTER 6 RESPONSE OF ANAEROBIC COMMUNITY TO A LONG-TERM FIVE DAY FEAST-FIVE DAY FAMINE PERIODIC SUBSTRATE PERTURBATION ......................................................................................... 6.1 Introduction ................................................................................ 6.2 Materials and methods ................................................................ 6.2.1 Anaerobic reactor ......................................................... 6.2.2 Substrate and nutrients ................................................. 6.2.3 Analytical methods ....................................................... 6.2.4 Reactor inoculation ....................................................... 6.2.5 Experimental design ..................................................... 6.3 Results ........................................................................................ 6.3.1 Effects of the periodic substrate perturbation ............... 6.3.2 Observation of protozoa density change ...................... 6.3.3 Microscopic observations ............................................. 6.3.4 Results of genetic analysis ........................................... 6.4 Discussion .................................................................................. 6.5 Summary .................................................................................... CHAPTER 7 OPERATIONAL CHARACTERISTICS OF THE RE-ESTABLISHED MICROBIAL POPULATIONS UNDER STEADY-STATE AND SHOCK LOADING CONDITIONS ............................................................. 7.1 Introduction ................................................................................ 7.2 Materials and methods ................................................................ 7.2.1 Anaerobic reactors ........................................................ 7.2.2 Substrate and nutrient solutions ................................... 7.2.3 Analytical methods ....................................................... 7.2.4 Experimental procedures .............................................. 7.3 Results ........................................................................................ 7.3.1 Operational performance of the daughter reactors ....... 7.3.2 Response of anaerobic populations to glucose and temperature shocks ................................ 7.4 Discussion .................................................................................. 7.5 Summary .................................................................................... CHAPTER 8 CHARACTERIZATION OF MICROBIAL COMMUNITY STRUCTURE CHANGE IN RESPONSE TO LONG-TERM PERIODIC SUBSTRATE FEED BY FATTY ACID METHYLESTERS (FAME) ANALYSIS .......... 8.1 Introduction ................................................................................ 8.2 Materials and methods ................................................................ 8.2.1 Anaerobic reactors ........................................................ 8.2.2 Sample collection and storage ...................................... 8.2.3 FAME extraction .......................................................... 8.2.4 FAME analysis ............................................................. 8.2.5 Numerical analysis methods ......................................... 8.2.6 Experimental design ..................................................... 8.2.7 Nonmenclature for fatty acids configuration ................ 8.3 Results ........................................................................................ 8.3.1 Determination of minimum sample size for FAME analysis .......................................................... 8.3.2 Characterization of microbial community structure in anaerobic reactor populations by FAME analysis ..... ix 133 133 134 134 134 136 136 138 138 150 154 156 158 162 189 189 191 191 191 191 191 192 193 194 194 196 8.3.3 Feasibility of monitoring microbial community structure changes by FAME analysis ........................ 202 8.3.4 Effects of starvation on FAME profiles ........................ 218 8.3.5 Effects of shock loading on FAME profiles ................. 222 8.3.6 FAME analysis results for mother and 1/1 day daughter reactor MPN samples ........................... 226 8.4 Discussion .................................................................................. 230 8.5 Summary .................................................................................... 235 CHAPTER 9 CONCLUSIONS AND ENGINEERING SIGNIFICANCE ........................ 236 CHAPTER 10 SUGGESTIONS FOR FUTURE RESEARCH ............................................ 240 APPENDIX A. General analytical methods ...................................................................... 241 B. MPN enumeration method ........................................................................ 244 C. Epifluorescence microscopy direct counting methods ............................. 248 D. Protein measurement procedure ............................................................... 252 E. Effects of pH and VFA concentrations on hydrogen-utilizing methanogens maximum hydrogen utilization capacity ............................. 254 F. Development of free energy equation ....................................................... 258 G. FAME extraction procedure .................................................................... 260 H. Nomenclature of cell membrane fatty acids ............................................. 262 LIST OF REFERENCES .............................................................................. 267 TABLE 2-1. Reactions involved in conversion of glucose to methane ................ 2-2. Metabolic pathways of glucose fermentation under low and high hydrogen partial pressures .................................................. 3-1. Nutrient supply for the glucose feed anaerobic reactor (mg/ 10 g Glucose) ................................................................. 3-2. Mineral solution for the glucose feed anaerobic reactor (mg/10 g Glucose) ................................................................. 3-3. Characteristics of the inoculum taken from Jackson Wastewater Treatment Plant ............................................................. 3-4. Batch operation results of the mother reactor ................................... 3-5. Operation parameters used in the mother reactor step-up loading increase period .......................................................... 3-6. Operational characteristics of the mother reactor under steady-state conditions ............................................................ 3-7. Volumetric and specific loading rates and gas production rates during the steady-state operation ....................................................... 3-8. Characteristics of the anaerobic biomass in the mother reactor during the steady-state operation ........................................... 3-9. CODN SS ratio measured during the experimental period ............... 3-10. Percentage contents of carbon, nitrogen and hydrogen in volatile fraction of biological sludge observed in various different studies... 3-11. COD mass balance calculation results for the baseline data acquisition period of the mother reactor ............................................ 3-12. MPN values of anaerobic microorganisms in the mother reactor evaluated using 5 tube dilution .............................................. 4- 1 Concentration of the nutrients in the combined influent LIST OF TABLES fed to the anaerobic reactor ............................................................... xi l3 17 24 24 26 29 32 35 36 37 38 39 41 49 4-2. Substrate concentrations applied to the mother and the daughter reactors during the maximum substrate conversion rate measurements ........................................................... 52 4-3. Summary of reactor operational results during the different stages of the perturbation period ........................................ 63 4.4. Specific maximum substrate utilization rates observed for the mother and the re-established daughter reactor .................... 67 4-5. Decay parameters of the mother and the daughter reactor cultures. 68 4-6. Decay rates of the glucose degrader and hydrogen-utilizing methanogens in the mother and the daughter reactor cultures .......... 74 4-7. Summary of protein and VSS measurement results .......................... 79 4-8. DTAF direct counting results for the mother and the daughter reactor cultures ................................................................... 80 4-9. Acridine orange direct counting results for the mother and the daughter reactor cultures ....................................................... 81 4- 10. Summary of acridine orange and DTAF direct counting results for the mother and the 1/1 day daughter reactor cultures .................. 82 4-11. MPN values of anaerobic microorganisms in the re-established daughter reactor culture .............................................. 83 5-1. Free energy associated with the anaerobic oxidation of glucose fermentation intermediates ................................................... 97 5-2. Summary of reactor operational results during the different stages of the perturbation experiment- ................................ 110 5-3. Summary of non-ionized Ac and non-ionized VFA concentrations (as acetic acid) observed during the 3/3 day and 1/1 day substrate perturbations ............................................ 114 5-4. Free energy values (KJ/mol) observed for the anaerobic oxidation of glucose metabolic intermediates during substrate perturbation.... 119 5-5. Electron equivalents (eq) of the anaerobic reactor substrate, intermediates and biomass ................................................ 122 5-6. Electron equivalents distribution (mmol) in anaerobic reactor effluent during the three day on/ three day off long-term periodic substrate perturbation ......................................... 122 6- 1. Apparent propionate degradation rate changes during the first four substrate perturbation cycles (40 days) .............................. 141 xii 7-3. 7-4. 7-5. 8-1. 8-2. 8-4. 8-5. Summary of reactor operational performance during the baseline steady—state and new established steady-state .................... COD and VFA accumulation rate in the beginning of the substrate perturbation .................................................................. Experimental conditions used in the glucose and temperature shocks ............................................................................ Summary of operational performance of the mother and the daughter reactors ......................................................................... Specific maximum substrate utilization rates measured for the mother and the daughter reactor cultures .................................... Baseline operation results for reactor-Ill and m-l ............................ Peak levels of glucose fermentation intermediates and other operational parameters in reactor 1/1 and m-l during glucose shock loading tests .................................................... Sample volumes and equivalent biomass (as VSS) of the mother reactor culture used for minimum sampling size determination ....... Effect of sample volume on FAME analysis results ......................... FAME profiles for the mother and the daughter reactor microbial populations ............................................................ Abundance (%) of the saturated, branched, unsaturated and cyclopropyl fatty acids in the different reactor populations .............. Similarity matrix (So, %) calculated based on FAME data listed in Table 8-3 .............................................................................. VFA concentrations examined in the Hz-C02 degradation experiment. .................................................................... xiii 146 158 166 173 173 177 183 195 195 197 198 198 249 FIGURE 2-1. Substrate flow for anaerobic digestion of complex organic materials... 2-2. Graphical representation of hydrogen-dependent thermodynamic favorability of acetogenic oxidations and inorganic respirations associated with the anaerobic degradation of waste organics ........... 2-3. General outline of possible biochemical conversion occurring on anaerobic degradation of glucose ................................. 2-4. Population dynamic control model for anaerobic acidogenesis in anaerobic digestion .................................................. 2-5. Acidogenic phase of glucose fermentation under low and high hydrogen partial pressures to form acetic acid, propionic acid, H2 gas and C02 ........................................................ 3-1. Schematic diagram of laboratory-scale anaerobic reactor ................. 3-2. Tracer test results of the mother reactor ............................................ 3-3. Batch operation results for COD reduction and gas production ....... 3-4. Batch operation results for biomass production and pH change ....... 3-5. Reactor performance during stepwise loading increase period ......... 4 1. Experimental set up ........................................................................... 4-2. Substrate feeding pattern used for perturbation experiment ............. 4-3. VFA concentration changes during the perturbation period ............. 4-4. COD variation during the perturbation period .................................. 4-5. pH change during the perturbation period ......................................... 4-6. Average gas production in the two-day cycle during the perturbation period ............................................................................ 4-7 Daily gas production during the substrate perturbation period ......... LIST OF FIGURES xiv 10 12 14 16 22 28 30 31 34 47 55 57 58 60 62 65 4-8. Deterioration of gas production on days when glucose was not supplied during stage 1 ........................................................ 66 49. Biomass decay of the mother and the daughter reactor cultures under starvation condition ................................................... 69 4-10. Estimation of biomass decay parameters for the mother and the daughter reactor cultures ....................................................... 70 4-11. Estimation of decay parameters for hydrogen-utilization activity in the mother and the daughter reactor cultures ................................ 72 4- 12. Estimation of decay parameters for glucose degradation activity in the mother and the daughter reactor cultures ................................ 73 4-13. Protein and biomass concentration changes in the mother and the daughter reactor cultures ....................................................... 76 4-14. Protein and biomass concentration changes in the daughter reactor during two complete perturbation cycles .............................. 77 4-15. Protein/V SS ratio changes in the mother and the daughter reactor cultures .................................................................................. 78 4—16. Microscopic observation of the mother (a) and daughter (b) reactor cultures ............................................................................. 85 5-1. Schematic of the experimental set up ................................................ 95 5-2. Substrate feeding pattern used for perturbation experiment ............. 99 5-3. VFA concentration changes during the perturbation period ............. 101 5-4. COD variation during the perturbation period .................................. 102 5-5. Hydrogen and ethanol concentration changes during the first stage of the perturbation period ................................................. 104 5-6. pH change during the perturbation period ......................................... 106 5-7. Gas production during the perturbation period ................................. 107 58 Daily gas production during the perturbation period ........................ 111 5-9. Summary of non-ionized volatile fatty acids, pH and anaerobic reactor performance .......................................................... 113 5-10. Changes in non-ionized VFA concentrations during the l/ 1 day and 3/3 day substrate perturbation tests ................................ 115 5-11. Free energy available for the anaerobic oxidation of hydrogen and carbon dioxide, and acetate during the substrate perturbation period ................................................................................................. 117 XV 5-12. 5-13. 5-14. 5-15. 6-7. 6—8. 6-9. 6— 10. 6-1 1. 6—12. 6-13. 7-1. 7-2. Free energy available for the anaerobic oxidation of butyrate and propionate during the substrate perturbation period ................... Free energy available for the anaerobic oxidation of ethanol during the substrate perturbation period ............................................ Electron mass balance during the perturbation period ...................... Microscopic observation of the mother (a) and the daughter (b, c) reactor cultures .......................................................... Schematic of the experimental set up ................................................ Substrate feeding pattern for the 5/5 perturbation experiment .......... COD and VFA concentration changes during the first stage of the 5/5 perturbation experiment .......................................... Hydrogen concentration changes during the first stage of the 5/5 perturbation period ...................................................................... pH and non-ionized VFA concentration changes during the first stage of the perturbation period .......................................... Free energy available for the anaerobic oxidation of propionate and butyrate during the substrate perturbation period ....................... COD and VFA concentration changes during the entire substrate perturbation period ............................................................. Biomass concentration change during the substrate perturbation period ............................................................................ Microscopic observation of the protozoa in the 5/5 reactor culture... Protozoa density changes during observation period ........................ Changes in protozoa density and total biomass concentration under starvation condition .......................................... Microscopic observations of the mother (a), the 5/5 (b), III (C) and 3/3 (d) daughter reactor cultures ..................................... DNA analysis results for the mother and the daughter reactor cultures. ................................................................................ Recovery processes of the 1/1, 3/3 and 5/5 day daughter reactor populations ............................................................. COD concentration changes in the re-established daughter reactors. VSS concentration changes in the re-established daughter reactors ............................................................................... xvi 118 120 123 125 135 137 139 142 144 145 147 149 151 152 153 155 157 168 169 170 7-7. 7-8. 7-9. 7-10. 7-11. 7-12. 8-1. 8-2. 8-4. 8-5. 8?. Hydrogen concentration pressure changes in the re-established daughter reactors ........................................................ Gas production changes in the re-established daughter reactors ....... Effect of perturbation interval on maximum glucose and pr0pionate utilization rates ......................................................... Effect of perturbation interval on maximum acetate, butyrate and hydrogen utilization rates ............................................. Changes in (a) COD and (b) glucose concentrations during the glucose shock loading experiment. .................................. Changes in concentrations of glucose fermentation inter mediates for the a) m-l and b) 1/1 reactors during the glucose chock loading expe1iment. ................................................................ Changes in a) hydrogen and b) biomass concentrations during the glucose shock loading experiment. .................................. Changes in pH observed during the glucose shock loading experiment. ........................................................................... Changes in (a) temperature and (b) COD concentrations during the temperature shock experiments ........................................ Principal component analysis of the FAME profiles presented in Table 8-3 ....................................................................... Clustering of the anaerobic reactor populations using the l-Pearson correlation coefficient, median linkage method with FAME profiles presented in Table 8-3 ...................................... Changes in major cell membrane fatty acids abundance during perturbation period in 3/3 day reactor populations ........................... Changes in grouped cell membrane fatty acids abundance during perturbation period in 3/3 day reactor populations ................ Changes in similarity coefficient So during perturbation period in 3/3 day reactor populations ................................................ Principal-component analyses of the FAME profiles obtained from 3/3 day reactor populations ........................................ Clustering of the anaerobic reactor populations using the l-Pearson correlation coefficient, single linkage method with FAME profiles obtained from 3/3 day reactor populations ........................................ xvii ‘ 171 172 175 176 178 180 181 182 184 200 201 203 205 206 208 209 8-8. 8- 10. 8-11. 8-12. 8-18. 8-19. 8-20. 8-21. E-2. Correlation between the reactor effluent COD and the similarity coefficient So, the abundance of the saturated fatty acids and the branched fatty acids ............................................. Changes in major cell membrane fatty acids abundance during the 5/5 day substrate perturbation period ............................... Changes in grouped cell membrane fatty acids abundance during the 5/5 day substrate perturbation period ............................... Changes in similarity coefficient So during the 5/5 day substrate perturbation period..-. .......................................................... Principal-component analysis of FAME profiles obtained from 5/5 day substrate perturbation experiment. ............................... Clustering of the FAME profiles obtained from 5/5 day experiment using the l-Pearson correlation coefficient, single linkage method. Changes in grouped fatty acids abundance during the starvation period ................................................................................ Changes in similarity coefficient So during the starvation period for the mother and the 1/1 day daughter reactor cultures .................. Principal component analysis results for the FAME profiles obtained during the starvation test ....................................... Changes in major cell membrane fatty acids abundance during shock loading experiment ...................................................... Changes in grouped cell membrane fatty acids abundance during the glucose shock loading test ............................. Similarity coefficient changes during the glucose shock loading test .............................................................................. FAME profiles obtained from mother and 1/1 daughter reactor glucose MPN enrichment cultures ........................................ DNA analysis results for mother and 1/1 daughter reactor glucose MPN enrichment cultures ........................................ pH effect on hydrogen conversion rates ............................................ VFA effect on hydrOgen conversion rates at pH 7.4. (a) acetate; (b) propionate; (c) butyrate .................................................. xviii 210 212 213 215 216 217 219 220 221 223 224 225 227 228 256 257 ATP BOD COD DNA TCD VFA VSS KEY TO ABBREVIATIONS Adenosine diphosphate Adenosine triphosphate Biological oxygen demand Chemical oxygen demand Deoxyribonucleic acid Dissolved oxygen ' Fatty acid methyl esters Flame ionization detector Gas chromatography Hydraulic retention time Most probable number Oxidized form of nicotinamide adenine dinucleotide Reduced form of nicotinamide adenine dinucleotide Optical density Suspended solids Standard pressure and temperature Thermal conductivity detector Volatile fatty acid Volatile suspended solids xix Chapter 1 INTRODUCTION Anaerobic processes have received increased application in recent years, both for the treatment of industrial wastewaters and for the conversion of biomass to methane. Despite the many advantages possessed by the anaerobic processes (McCarty, 1982), due to the inherent slow growth rate of the methanogenic and acetogenic bacteria and the complex interactions between different groups of bacteria, anaerobic systems can be sensitive to environmental changes. In engineered systems, such as anaerobic digesters, a sudden increase in the organic or hydraulic applied loading rate can result in serious operational problems including reactor failure. Long recovery periods are often required. Many studies have examined the response of anaerobic systems to environmental perturbations, especially changes in organic loading rate (Cohen etal., 1981; Mosey et al., 1989; Hickey and Switzenbaum, 1991b, 1991c; Gupta, 1994). These studies are of three different types: (1) pulse feed experiments in which a substrate pulse is injected suddenly into an anaerobic reactor; (2) step feed experiments in which the reactor organic or hydraulic loading rate is suddenly increased to a higher level for a period of time then returned to the steady-state conditions; and (3) periodic feed experiments in which the influent substrate concentration is periodically changed. Essentially all of the perturbation experiments described to date were performed over a relatively short time period (from a few days to a few weeks). Little is known about the effects of long-term, continuous periodic substrate perturbations on anaerobic community structure and function. This is despite the fact that long-term periodic substrate loadings are common in nature and in engineered systems. 2 This project was designed to investigate the response of anaerobic chemostats to long- terrn, continuous periodic substrate perturbations. The major objectives were to (1) demonstrate the influence of long—term, periodic substrate perturbations on anaerobic system operational, kinetic, thermodynamic and metabolic performance and (2) investigate whether long-term adaptive changes in microbial community might occur in such systems. Chapter 2 BACKGROUND 2- l. Anaerobic degradation process Anaerobic degradation of complex organic materials can be considered as a step-wise processes which proceeds as both successive stages and parallel reactions (McCarty, 1964; Kaspar and Wuhrmann, 197 8; Zehnder, 1978; Gujer and Zehnder, 1983; Pavlostathis, 1991). The whole sequence of anaerobic degradation of organic materials can be: regarded as a microbial energy web with a product produced by one bacterium being used by another. The conversion of complex organic materials to methane takes place in four stages and involves at least three groups of microorganisms (Figure 2-1). First, Complex polymeric materials such as polysaccharides, proteins, and lipids are hydrolyzed into small, soluble products (amino acids, sugars and long-chain fatty acids) which can be aSSirnilated by the bacterial cells. Second, the assimilated compounds are fermented or anaerobically oxidized to short—chain fatty acids plus hydrogen gas and carbon dioxide. In the third step, the short-chain fatty acids, mainly butyrate and propionate, are converted to a‘Cetate, hydrogen gas and carbon dioxide. Finally, methanogenisis occurs from carbon dioxide reduction by hydrogen and from decarboxylation of acetate. The three major groups of bacteria which mediate the anaerobic degradation of organic materials are as follows: 1. Acidogenic bacteria, which include hydrolytic and fermentative bacteria (Ia and 1b); 2. acetogenic bacteria (H); 3. methanogenic bacteria, which include acetophilic and hydrogenophilic methanogens (111a and IIIb). In addition to the three major groups of bacteria mentioned above, several other groups of organisms . Complex organic Stages 9mm [ materials 1 ' la Hydrolysis Amino acids Sugars Fatty acids Acidogens‘ lb Acidogenesis Propionate Butyrate Acetogens * Acetogenesis Acetate J Illa lllb Methanogenesis Methanogens Methane gig 2-1. Substrate flow for anaerobic digestion of complex organic materials (adapted from alns Sam-Soon, (1987) and Zinder (1984)) 5 also involved in anaerobic degradation including sulfate reducing bacteria (Zeikus, 1979; Zehnder et al., 1981), homoacetogens (Zeikus, 1979; Widdel, 1988) and protozoa (Williams and Harfoot, 1976; Fenchel, 1987). Due to the limited amount of available energy and the normal environmental conditions, these organisms usually only contribute a small portion of the COD removal from anaerobic systems. Although the anaerobic microorganisms and the reactions that they mediate can be theoretically separated into several groups/stages, overall methanogenesis is strongly dependent on close interaction between the bacterial groups in each stage. 2-2. Methanogenesis Methanogenic bacteria, which mediate the terminal step in anaerobic degradation, play the most important role in anaerobic degradation. Methanogenic bacteria are a morphologically diverse group of bacteria unified by their ability to produce methane. They possess some unique features, which distinguish them as a special group of bacteria, including the absence of peptidoglycan in the cell wall; the presence of mainly ether-linked isoprenoids rather than ester-linked phospholipids in the membranes; and the presence of unusual coenzymes such as coenzyme M, factor F420, Factor F430, methanopterin, and l“nethanofuran (Jain et a1. 1991). Methanogens utilize a very limited number of simple Carbon compounds as carbon and energy sources for methanogenesis, i.e. acetate, H2 plus C02, forrnate, methanol and CO. Acetate-utilizing methanogens normally control the pH value of the fermentation by l‘emoval of acetate and formation of carbon dioxide. They are responsible for the majority of COD removal and organic compound stabilization. McCarty (1964) reported that for Waste solids digestion about 70% of the COD was converted into methane via acetate decarboxylation. Based on theoretical calculations, Mosey (1981) estimated that 67% of CH4 produced from glucose was generated by acetate decarboxylation. These organisms are quite sensitive to pH changes and grow well only within a narrow pH range of ca. 6.5 to 7.4 (McCarty, 1964; van den Berg et al., 1976; Yong and Okos, 1987). It has been 6 found that at a pH value about 7, the methanogens can tolerate an acetate concentration as high as ca. 6,000-7 .000 mg/L without loss their normal acetate metabolic capacity. Yang and Okos (1987) published a study on pure culture of methanogens utilizing acetate at pH 7 and 35 0C, in which the optimum acetate concentrations for the methanogenic bacteria Methanosarsina barkeri and Methanosarsina mazei were 7,000 and 6,800 mg/L, respectively. Van den Berg et al. (1976) reported similar results for an acetate enrichment culture growth at pH 7 and 35 0C. In this study it was observed that acetate conversion rates were not significantly affected by acetate concentration up to 6,000 m g/L, but when the reactor pH dropped to below 6.5, methanogenic activity decreased rapidly. In a pilot study, Duarte and Anderson (1982) found that a pH level of 6.4 was critical for a continuously fed acetate catabolizing digester (35 0C, 4.5 day HRT). When reactor pH was decreased from 6.5 to 6.3, methane production decreased by 65%, even though the digester had previously operated successfully at the same acetate concentration (ca. 2500 rug/1). This inhibition of acetate metabolism was mainly attributed to the increased unionized acetic acid concentration at the lower pH (Andreson et a1, 1982; Attal et al., 1 98 8). The acetate-utilizin g methanogens function at relatively high saturation levels (the ratio of steady-state substrate utilization rate V0 to the maximum substrate utilization rate Vmax) Compared to other anaerobic bacteria. When digesting sludge from a full-scale municipal digester (33 0C, 40 day HRT) was tested for its acetate degradation rate by spiking acetate in a lab-scale digester, the rate increased from 18.3 mg acetate/L-h at steady-state to 27.8 mg acetate/L-h at enzyme saturation conditions (Kaspar and Wuhrmann, 1978). The eStimated half-saturation constant varied from 14 mg/L to 31 m g/L acetate. These results indicate, that in the full-scale digester, the acetate-degrading system was saturating ca. 50%. A similar result, ca. 45% saturation level, was reported by Valcke and Verstraete ( l 983) in their study of sludge fermentation. The relatively high saturation level indicates that the acetate-utilizing methanogens have limited additional capacity to handle sudden increase in substrate loading rate. 7 The free energy available for growth of acetate-utilizing methanogenes is relatively low. Mosey (1981) estimated that only ca. 0.5 moles of ATP can be obtained for acetate- utilizing methanogens to break down one mole acetate. The acetate-utilizing methanogens are among the slowest growing bacteria in the anaerobic microbial community with an experimentally observed growth yield of 0.01 to 0.05 g/g-acetate, and the kinetics of their growth often dominates the overall rate of anaerobic reactions (Harper and Pohland, 1985; Lin et a., 1986). Due to the inherent characteristics mentioned above, the acetate-utilizing methanogenes are sensitive to environmental changes, and long periods are likely to be required for this group of bacteria to adjust to new environmental conditions. The hydrogen-utilizing methanogens contribute to ca. 20% to 30% of methane production depending on original substrate composition (Pavlostathis and Giraldo- Gomez, 1991). In spite of the quantitative methane production, interspecies hydrogen transfer and utilization is far more important function of these methanogens since it regulates the rate of H2-producing reactions by controlling the partial pressure of hydrogen. In anaerobic systems no other compounds can so quickly change system thermodynamic conditions like molecular hydrogen (or formate). Under inhibitory or slll'ge loading conditions, the hydrogen partial pressure can increase by several orders of magnitude within a few hours (Harper and Pohland, 1985; Smith, 1987; Mosey and Fernandez, 1989; Bae and McCarty, 1993). This can seriously inhibit or stop acetogenic 1‘eactions, and significantly change fermentation intermediate distribution in anaerobic digesters. In well-operating systems, the hydrogen partial pressure is maintained within a relatively low range of ca. 30 to 200 ppm (gas phase) by hydrogen-utilizing methanogens, which is necessary for many anaerobic reactions to occur (Mosey, 1983; Lovely, 1985; Cord-Ruwisch, 1988; Zinder 1990; Bea and McCarty, 1993). The hydrogen-utilizing methanogens are relatively fast growing species, with estimated cloubling time of 4 to 12 hours (Gujer and Zehnder, 1983; Archer and Powell 1985; Mosey and Femandes, 1989; Pavlostathis et al., 1990). They have been found to be 8 resistant to elevated concentrations of VFA. Hobson and Shaw (197 8) reported that at a pH of 7.1, an acetate or butyrate concentration up to 10,000 mg/L (as acetic acid) did not inhibit M. formicicum, the principal hydrogen-utilizing bacterium in a piggery-waste digester. Experimental results also indicate that the Hz-utilizing methanogens are also quite resistant to changes in pH values. A optimum pH range of 5.4 to 7.2 was reported for hydrogen-utilizing methanogens in digester sludge by Attal et al. (1988). Kasper and Wuhrrnann (1978) reported that under steady state condition the saturation level of the hydrogen-utilizing methanogens in a mature digester was only about 1%. Similar results have also been reported by Shea et a1. (1968). There is, therefore, a vast extra utilization capacity available for increased H2 oxidation during surge loading conditions. These characteristics endow the hydrogen-utilizing methanogens with the ability to quickly and effectively respond to changes in environmental conditions. Prolonged hydrogen accumulation is rarely observed under organic-hydraulic perturbation experiments (Smith and McCarty, 1989; Mosey and Femandes, 1989; Hickey and Switzenbaum, 1991c). 2-3. Acetogenesis The anaerobic oxidation of short chain fatty acids and ethanol is an important stage in anaerobic degradation, because no methanogens have been found to acquire the ability to directly utilize these fermentation intermediates (Zeikus, 1977). Nineteen species of sYtrophic acetogens have been identified so far, including five species capable of converting butyrate and the others capable of converting propionate and higher VFAs (Li er. al., 1994). Enrichment culture studies indicate that acetogens growth relatively slowly even under optimum conditions, with estimated minimum doubling time of 1.5-5.0 days (Lawrence and McCarty, 1969; Gujer and Zehnder, 1983; Heyes and Hall, 1983; Lin et al., 1 986). The reactions mediated by acetogens are thermodynamically unfavorable under Standard conditions. In order for these reactions to proceed, it is necessary for the reaction products (acetate and Hz) to be present at sufficient low concentrations to yield a negative 9 value for the actual available free energy change (AGO') under physiological concentrations (Thauer et al., 1977). Therrnodynamically related product inhibitions have been illustrated by experimental observations (Heyes and Hall, 1981; Kaspar and Wuhrmann, 1989; Hickey and Switzenbaum, 1991c; Smith and McCarty, 1989). Ahring and Westerman (1987) reported that a hydrogen partial pressure in the gas phase up to 2 x 10‘2 atrn could completely block butyrate degradation. Dwyer et a1. (1988) found that removal of acetate from the butyrate rich media increased the butyrate degradation rate. Mosey (1981) reported a reactor failure which was directly related to thermodynamic inhibition of hydrogen to propionate degradation. As shown in Figure 2-2, thermodynamic calculations associated with acetogenic and methanogenic reactions indicate that in order for anaerobic degradation to proceed for all the important anaerobic fermentation intermediates the hydrogen partial pressure in the digester headspase (assuming equilibrium with the liquid phase) must be maintained in a narrow range of ca. 104 to 10'6 atm. Actually, in well operated anaerobic systems the hydrogen partial pressure has been found to be within a range of ca. 4 x 105 to 2 x 10 '4 atrn (Kaspar and Wuhrmann, 1978; Heyes and Hall, 1983; Mosey, 1983b; Smith, 1987). This is sufficient low so that butyrate and propionate OXidation is energetically favored. The number of acetogens existing in an anaerobic process has been reported dependent upon the feeding schedule of a given reactor and its associated stability history (Harper and Pohland, 1985). Following a shock loading or other process upset, the proportion of acetogens will increase to accommodate accumulated short chain fatty acids. In contrast, under steady-state conditions, the number of electrons channeled through DrOpionic and butyric acids can be greatly decreased (Palns, 1987; Mosey 1981), and the Ilumber of acetogens will decrease accordingly. Considering the inherent slow growth rate of acetogenic bacteria, this dynamic situation is essential for understanding the effects of environmental perturbations on the overall stability and efficiency of anaerobic systems, eSpecially during transient accumulation and recovery stages (Fongastitkul et a1. 1994). 10 Gibb's Free Energy Change (A66) Per Reaction, KJ -80 -100 +120 if T l a 0 .4 9 .. Hydrogen Partial Pressure, (atm) Fig 2- 2. Graphical representation of hydrogen-dependent thermodynamic favorability of aCetogenic oxidations and morganic respirations associated with the anaerobic degradation of waste organics. (1) Propionic acid oxidation to acetic acid. (2) Butyric acid oxidation to acetic acid. (3) Ethanol to acetic acid. (4) lactic acid to acetic acid. (5) Acetogenic l‘EESpiration of bicarbonate (C02). (6) Methanogenic respiration of bicarbonate (C02). (7) Respiration of sulfate to sulfide. (8) Respiration of sulfite to sulfude. (9) Methanogenic Cleavage of acetic acid. (”10) Sulfate reducing bacteria mediated cleavage of acetic acid. Acetic acid, 25 mM; propionic, butyric, lactic acids, and ethanol, 10 mM; bicarbonate, 20 mM, methane, 0.7 atm. After Harper and Pohland (1985). 1 1 With further research and exploration, it can be regarded as one of the key considerations to stabilizing and improving anaerobic treatment. 24. Anaerobic fermentation of glucose In this study a simple carbohydrate, glucose, was used as the sole energy and carbon source. Studies on methane fermentation of glucose using 14C tracers have shown that glucose fermentation primarily involves the Embden-Meyethof-Pamas pathway (Jeris and McCarty, 1962). After glycolysis, pyruvate is formed which can be fermented by a number of pathways (Kisaalita et al., 1989; Linden, 1989). The most important metabolic routes of anaerobic decomposition of glucose can be assembled in a general outline, showing the three degradative stages (Figure 2-3). The major intermediates produced from glucose fermentation are acetate, butyrate, propionate, ethanol and gaseous products of H2 (and C02). These fermentation intermediates are eventually converted to methane and carbon dioxide via subsequent acetogenic and methanogenic stages. The most important reactions, which may occur during glucose anaerobic degradation, are depicted in Table 2- 1 ('Ihauer et al., 1977). For acidogenic bacteria, the reaction which converts glucose into acetate, is preferred one. It provides the acid-forming bacteria with the largest energy yield for growth and provides methanogens with their prime substrates. In spite of this, significant changes in distribution of glucose fermentation intermediates are often observed, especially under stressed conditions (Zoetemeyer et al., 1982a, 1982b; Cohen et al., 1983; Harper and Pohland, 1986; Bae and McCarty, 1993). In a chemostat glucose acidogenesis experiment, Cohen et a1. (1983) recognized the existence of two glucose fermentation routes which occurred complementary to each other. The so called "butyric acid type" fermentation is characterized by production of acetate, butyrate, carbon dioxide and hydrogen as the main fermentation products; while the "propionic acid type" fermentation is characterized by the formation of acetate, propionate, and carbon dioxide, with much lower hydrogen 12 HYDROLYSIS __ WLYSACEHARIDES TCD ACIDOGENESIS GLUCOSE J -‘-——C02 .3 f H ® @ Hr H 6 ,____.., H @ PvagvATE LACTATE -¥- ROPlONAT H. H 7525 4 -FORMATE H @ ® j H 1 { ® ACETYLICOA _L_., ETHANOL [ BUTYRATE ' © 5 l i J I L C02 | ACETATE ORGAN C ELECTRON SINK PRODUCTS ACETOGENESIS ACETOGENIC HYDROGENATION r re We ‘\ PROPIONATE . H2 CO, ACETATE BUTYRATE . LACTATE ETHANOL L L k 653 L J #1 ACETOGENIC DEHYDROGENATION (pH,< 10" ATM) METHANOGENESIS 2 CO ACETATE %1 2 l I cow-Reiucnow ACETATE DECE%BOXYLATION CH4 * H20 CH. 0 C02 Figure 2:3. General outline of possible biochemical conversion occurring on anaerobic degradatron of glucose. After T. Cohen (1982). 13 Table 2-1. Reactions involved in conversion of glucose to methane AG°' (KJ) 1. Acidogenic reactions (1) C6H1206 + 4H20 = 2CH3COO‘ + 4H+ + 2HCO3' + 4H2 -206.0 (2) C6H1205 + 2H20 = CH3CH2CH2COO' + 3H+ + 2H2 + 2HCO3' -255.0 (3) C6H1206 = 4/3CH3CH2COO'+ 2/3CH3COO'+ 8/3H++ 2/3HCO3' -220.0 (4) C6H1205 + 2H20 = 2CH3CH20H + 2H+ + 2HC03‘ -226.0 2. Acetogenic reactions (5) CH3CH20H + H20 = CH3COO' + H+ + 2H2 9.6 (6) CH3CH2COO‘ + 2H20 = CH3COO' + 3H2 + C02 76.1 (7) CH3CH2CH2COO' + 2H20 = 2CH3COO' + 2H2 + H+ 48.1 3. Methanogenic reactions (9) CH3COO' + H‘l' = CH4 + C02 — 35.8 (8) 4H2 + C02 = CH4 + 2H20 -130.7 concentrations. The two different types of glucose fermentation were assumed to be carried out by different fermentation species. Based on observation of VFA accumulation patterns during overload and recovery periods, McCarty and Mosey (1991) proposed a population dynamics model for anaerobic acidogenesis of carbohydrates. In this model the anaerobic fermentation of carbohydrates is proposed as competition between "propionic" and "butyric" bacteria (Figure 2-4). According to this model, under normal operating conditions (low concentration of hydrogen and modest organic loading rate) butyrate forming bacteria will directly produce acetate plus C02 and H2, so that the maximum free energy can be 14 233 .682 28 ban 0 .5“ .553? 03885 E emocowoemoa 03885 he BEE 35:8 ems—SEC cowaueomiv-m “95$.“ Acozflcno .2505 NI + N00 + magma/x Amos; In 30.; 9.38m NI + N00 + 39:33 2.53 2033588 35:33 :95 35 mum—Eon. . 3.303 3334. 3.8305 A 8820 39530.5 15 obtained. At low pH, the butyric forming bacteria will produce butyrate plus C02 and H2 to reduce the production of protons. As proposed by this model, the propionic bacteria become predominant only under surge loading conditions (high substrate concentration). The main fermentation products would be propionate, acetate and formate. Although this model generally agreed with experimental observations, as the authors noted, this model, which was deduced from digester operational performance, remains to be rigorously tested. Based on thermodynamic calculations, Duarte and Anderson (1982) proposed that under normal conditions (pH 7 and temperature of 35 0C), the butyrate type fermentation is thermodynamically more favorable than the propionate type fermentation. Palns Sam- soon et al. (1987) proposed two alternative glucose acidogenesis metabolic pathways based on biochemical and thermodynamic considerations (Figure 2-5). According to Palns Sam—soon, depending on digester hydrogen partial pressure the acidogenesis of glucose could go through either of the two alternative metabolic pathways as shown in Table 2-2. Pathway I is the favorable metabolic pathway employed by acidogens. Four moles of ATP can be generated from one mole glucose metabolized. However, this pathway is available for acidogens only under low hydrogen partial pressure (< 10‘4 atm.). Under conditions of stress, a shift in the metabolic pathway to pathway 11 could occur, resulting in a shift towards more propionate production. A model using a regulator function based on concentrations of hydrogen gas in digester headspace was developed by Mosey (1983) to simulate accumulation and decay of the major volatile fatty acids during hydraulic overload and recovery in a glucose fed digester. In this model, the hydrogen partial pressure is assumed to be directly linked to the NADH regeneration and, therefore, regulates the acidogenic reactions at several points in the glycolytic pathway. By providing a framework to thermodynamically relate hydrogen accumulation to the rate of acidogenic and acetogenic reactions, this model is a good start in helping to uncover how these systems operate. But as mentioned by the 16 63.38:: 353:5 .m .< 0:52.80 :08" .o2-:oen::m .m2m .m:o::_>o.5n< Anna: 2: :0 58-8% 23m :82 .80 2: Ba N: e8: cacao:— e8: 38: E8 9 85:83 3:3 Ewe—ob— .35 2:: Be. :25: 5:85.58 8003”“ .«o 829 253292 .m-~ Paw—n— osmmma :3th N... :9: 3v 8:32: 3:3 N: 26.. AS ON: + Bo: 2:28... Eon 0302 N: 20: uzmu< N A_._. 0.5 mM CH4) was checked by gas chromatography to establish whether or not the tubes were positive for methanogens and acetogens. Hydrolytic and fermentative bacterial populations were enumerated in the tubes containing 1111 C01 11} 31 26 the basal medium with 2 g/L of glucose. The tubes were incubated at 37 0C for two weeks. Tubes were scored as positive on the basis of increase in optical density (> 0.2) at 650 nm and production of volatile fatty acids (> 5 mM total VFA), when compared with non- inoculated control tubes. The MPN values and 95% confidence limits were calculated by using a MPN index computation table in Standard Methods (APHA, 1989). A more detailed description of MPN enumeration is provided in Appendix B. 3.2.6 Inoculum material The inoculum for the mother reactor was anaerobic digester sludge taken from the Jackson Wastewater Treatment Plant (Jackson, Michigan). The characteristics of the inoculum are shown in table 3-3. Table 3-3. Characteristics of the inoculum taken from Jackson Wastewater Treatment Plant Parameter Value Color black pH 7.4 SS (g/L) 17.3 VSS (g/L) 9.8 VSS/SS (%) 56.6 C (% of ash free dry weight) 51.5 H (% of ash free dry weight) 8.2 N (% of ash free dry weight) 5.9 27 3.3 RESULTS 3.3.] Mixing characteristics of the mother reactor The mother reactor was designed to operate as a completely-mixed stirred tank reactor (CSTR) without recycle. In order to illustrate the mixing characteristics of the reactor, a methyl blue dye tracer test was performed. In this test 1.5 ml concentrated methyl blue dye was injected into the reactor as an impulse and samples were then continuously collected from reactor effluent port at a pre determined time schedule. The collected samples were immediately transferred into 1.5 ml Standard polystyrene curettes (Life Science Products, Inc., Denver, Colorado) and the adsorption of the samples was obtained using a Perkin Elmer Lambda 6 UVNIS spectrophotometer (Perkin-Elmer Co., Norwalk, CT) at a wavelength of 425 run against D. I. water. The dye concentrations were obtained from a dye adsorption—concentration calibration curve prepared prior to the experiment. The effect of methyl blue dye adsorbed to the inner wall of the reactor was negligible. The data obtained from the tracer test along with the theoretical curve of an ideal CSTR (Levenspiel, 1972) is shown in Figure 3-2. The good fit of the experimental data with theoretical suggests that the mixing behavior of the mother reactor is close to an ideal CSTR. 3.3.2 Reactor start-up The digester sludge added to the ”mother" reactor was first passed through a 2 mm screen to remove coarse debris, and then diluted with tap water to obtain a VSS concentration of 1 g/L. The reactor liquid phase was continuously purged with nitrogen gas during inoculation to prevent the anaerobic microorganisms from being exposed to oxygen. After inoculation, a fast start-up operation was unsuccessfully attempted. During this unsuccessful start-up operation, the reactor was operated at an inlet glucose concentration of 8 g/L and hydraulic retention time of 15 days. Failure was caused by a rapid glucose to CilCo (%) 28 100 0 Measured data 80 . —— Ideal CSTR 6O ' 40 r 20 . O I t I 0 1 2 s Frgure 3-2. Results Obtained from the mother reactor tracer test 29 methane and carbon dioxide. The average daily COD reduction rates observed after the three consecutive batch feeding operations indicated increasing methanogenic activity. During batch operation the reactor pH ranged from 6.9 to 7.2. accumulation of VFA accompanied by a sharp decrease in reactor pH and gas production. Under this condition the reactor soon became a so called "stuck" reactor (McCarty, 1982) with poor COD removal efficiency (< 20%). Attempts to effect a quick recovery of the reactor failed. Subsequently a combined batch and stepwise increase in the organic loading rate was employed for start-up. It took ca. 150 days to complete the start-up process in this reactor including ca. 45 days of batch operation. Initially, 5 grams glucose was added into the reactor re-inoculated with active anaerobic microorganisms. When continuous gas production and COD reduction were observed, three batch feeding operations were subsequently performed. In the first run, 20 grams glucose was injected into the reactor. For the following two operations 35 grams of glucose was injected into the reactor. Proper amounts of buffer and nutrient solutions were also added Variations of COD, gas production, VSS and pH in the reactor during batch operation are illustrated in Figure 3-3 and 3-4, respectively. The experimental data obtained from batch operations are presented in Table 34. Close to stoichiometric amounts of gas Table 3-4. Batch operation results of the mother reactor Parameters Run 1 Run 2 Run 3 Amount of glucose addition (g) 20 35 35 Operation time (days) 15 16 13 Initial COD (mg/L) 1490 2470 2640 Final COD (mg/L) 160 220 310 Total gas production (L)1 14.2 25.9 25.3 Final VSS (mg/L) 880 890 910 Final pH 7.12 7.04 7.10 COD reduction rate (mg/L-d) 90 140 180 1: Gas production under standard condition (0 0C and 1 atrn) 30 1000 500‘ O I I r I 0 10 20 30 . . 40 50 Time (day) 40 Gas Production (L) Time (day) Figure 3-3. Batch operation results for COD reduction and gas production 31 1500 1200 ‘ I3 900 h E g, 600 J > 300 " O I I I I 0 10 20 30 40 50 Time (days) 7.4 J 7.2 '1 :1: 7.0 " a. 6.8 ‘ 66 1 I I I 0 10 20 30 40 50 Time (days) Figure 34. Batch operation results for biomass production and pH change 32 production were observed during batch operation. This indicates complete conversion of glucose to methane and carbon dioxide. The average daily COD reduction rates observed after the three consecutive batch feeding operations indicated increasing methanogenic activity. During batch operation the reactor pH ranged from 6.9 to 7.2. Following the third batch feed, the reactor was switched to continuous feed mode. An initial inlet glucose concentration of 4 g/L and hydraulic retention time of 20 days was used. Under this condition the reactor had a volumetric COD loading rate of 0.21 g/L-d, which was close to the average daily COD utilization rate of 0.18 g/L-d obtained in the last run of the batch operation. After that, an increase in either the glucose feed concentration or reactor flow rate (Table 3-5) was employed when the system appeared able to accommodate increased loading rates. About 110 days were required to reach the desired operating conditions of 8 g/L glucose feed and 10 day HRT. By the end of the start-up period, a new community had been established in the reactor. Theoretically less than 0.1% of the original inoculum remained after over 7 HRTs. After that, analyses of the contents reflected only biological activity resulting from utilization of the glucose fed to the reactor. Table 3-5. Operation parameters used in the mother reactor step-up loading increase period No. HRT Glucose feed Loading rate (dayS) (g/L) (C0D g/L-d) 1 20 4 0.21 2 20 6 0.32 3 15 6 0.42 4 15 8 0.56 5 10 8 0.85 33 Operational data including influent and effluent COD concentrations, COD removal efficiency and volumetric COD loading rate, are illustrated in Figure 3-5. The pH in the reactor was maintained between 6.8 to 7.1, with the alkalinity ranging from 2,300 to 4,400 mg/L (as CaCO3). During the stepwise loading increase period, no glucose was detected in the reactor liquid phase (< 0.5 mg/L). 3.3.3 Reactor operational performance under steady-state conditions After start-up, the reactor was continuously operated at an inlet glucose concentration of 8 g/L, temperature of 35 0C and hydraulic retention time of 10 days. Once steady-state conditions were established in the reactor, as indicated by consistent effluent COD and VSS concentrations and gas production, the steady-state baseline operation parameters of the mother reactor were intensively collected and analyzed over a period of ca. 50 days. The information collected during this period included influent and effluent COD, pH, SS and VSS, gas production and gas composition, and VFA concentrations. No significant variations were noted for any of the parameters measured during the baseline information collection period. The baseline experimental data are summarized in Table 3-6. The calculated volumetric and specific loading rates and gas production rates are presented in Table 3-7. The yield of bacteria was 0.156 (g VSS formed/ g COD consumed) or 0.165 (g VSS formed/g glucose consumed). Glucose was not detected in the effluent (< 0.5 mg/l) throughout the experimental period. 34 10000 1 Influent . 0. . . fl ' 80m '10.. .1 ee e 'e*' am ° ° " é a e ' le're’ O IDDD U 2000 Effluent O 1. . I 're'ietflllleenlle' I" 1111111111 w 50 70 90 110 130 150 170 Time (days) a 1.0 ' r g... e O... .eet’e ‘. ".'O . O «"1110 lm a O. . 0 . O 8 0.8 . COD removal _ 80 U 3 «1 l- 8 0.6 ‘ ’60 a h d an I: 1§ 0.4‘ “'40 T: i COD loading rate '5 0.2 ‘ '20 0 E 3 g 00 I I f I I I I O 50 70 90 110 130 150 170 Time (days) Figure 3-5. Reactor performance during stepwise loading increase period COD removal efficiency (%). 35 Table 3-6. Operational characteristics of the mother reactor under steady-state conditions Parameter Value No of observations CODin (m g/L) 8400 $2501 30 CODc (mg/L) 320 $40 30 COD removal (%) 96.2 pH 7.03 $0.05 45 SS (g/L) 1.34 $0.11 33 VSS (g/L) 1.26 $0.11 33 VSS/SS (%) 94.0 Gas production (ml/L-d) 4302 $60 45 CH4 (%) 51 $2 15 H2 @pm) 37 $7.8 15 Alkalinity (mg/L as CaCO3) 3800 $80 30 Acetate (mg/L) 120 $50 20 Propionate (mg/L) 12 $5 20 Butyrate (mg/L) 4 $3 20 1. Standard deviation 2. Value presented at 273 K, 1 atm. 36 Table 3-7. Volumetric and specific organic loading rates and gas production rates during steady-state operation Parameter Value Volumetric Loading rate (g COD/L-d) 0.85 Specific Loading rate (g COD/g VSS-d) 0.68 Volumetric gas production (ml/L-d) 438 Specific gas production (ml/g VSS-d) 348 Volumetric methane production (ml/L-d) 223 Specific methane production (ml/g VSS-d) 178 Gas production (ml/g-COD) 522 Gas production (ml/g-glucose) 554 Methane production (ml/g-COD) 265 Methane production (mllg-glucose) 282 3.3.4 Biomass characteristics Results of eight independent elemental analysis of the biological solids in the mother reactor are presented in Table 3-8. Based on the average content of the elemental carbon, hydrogen, nitrogen and oxygen in the ash-free volatile fraction of the anaerobic sludge, an empirical chemical formula of the anaerobic biomass in the mother reactor was: CH1.700.4N0.2 (3-1) Based on this empirical formula, the oxidation reaction of the biomass in the mother reactor can be written as: CH1.700.4N0_2 + 1.075 02 = C02 + 0.2 NH3 + 0.55 H20 (3-2) 22.9 + 34.4 = 44 + 3.4 + 9.9 Based on equation (3-2), to completely oxidize one mole biomass, which has a theoretical molecular weight of 22.9 g, 1.075 moles of oxygen is required, which gives a CODN SS ratio of 1.5 (g/g). This theoretically calculated CODN SS ratio was confirmed by experimentally measured COD/V SS ratio. As shown in Table 3-9, an averaged 37 COD/V SS ratio of 1.515 $ 0.177 (g/g) was obtained from 18 groups of independently measured data. The empirical formula of anaerobic solids found in this study compared closely with empirical formulations of anaerobic and aerobic biological solids reported in the literature. The elemental content of carbon, nitrogen and hydrogen in the ash-free (volatile) fraction of various anaerobic and aerobic biomass reported in the literature are compared to those obtained in this study in Table 3-10. Table 3-8. Characteristics of the anaerobic biomass in the mother reactor during steady-state operation Elemental content in ash-free dry weight (%) Carbon Hydrogen Nitrogen Oxygenl Ash (%) 1 52.29 7.83 9.59 30.29 5.1 2 51.55 7.80 9.34 31.31 7.7 3 53.40 7.44 10.30 28.86 6.6 4 51.85 7.44 10.20 30.51 5.4 5 53.69 7.70 10.60 28.01 6.2 6 54.56 7.67 10.92 26.85 5.5 7 54.41 7.68 10.53 27.38 8.1 8 54.02 7.69 10.45 27.84 5.4 Ave. 53.22 7.66 10.24 28.88 6.3 Std2 1.10 0.14 0.49 1.53 1.1 1. Estimated 2. Standard deviation 38 Table 3-9. COD/V SS ratio measured for the mother reactor No CODT CODc CODVSS VSS CODN SS 1 1854 439 1415 840 1.685 2 1756 366 1390 880 1.580 3 1805 341 1464 980 1.493 4 1902 341 1561 840 1.858 5 1805 292 1513 1000 1.513 6 1902 341 1561 900 1.734 7 1854 293 1561 1100 1.419 8 2000 293 1707 1060 1.610 9 1658 546 1112 940 1.183 10 1756 530 1226 1020 1.202 11 2000 585 1415 987 1.433 12 2000 415 1585 920 1.722 13 1902 488 1414 1020 1.386 14 1951 463 1488 1080 1.378 15 . 2000 536 1464 910 1.609 16 2000 463 1564 1100 1.422 17 1951 536 1415 920 1.538 Ave. 1888 428 1462 970 1.516 Std. $101 $97 $135 $83 $0.176 39 Table 3-10. Percentage contents of carbon, nitrogen and hydrogen in volatile fraction of biological sludges observed in various different studies C N H Formulae Substrate Reference Content in VSS (%) 51.2 1 1.7 8.2 CH1.900,4N0,2 glucose 1 50.4 12.8 7.3 CH1,700,4N0,2 glycerol 2 47.4 10.7 7. 1 CH1.gOo,6No.2 glucose 3 48.6 10.8 7.0 CH1.700,5N0_2 Starch 3 45.2 1 1.1 6.4 CH1.700.6N0.2 Glycine 3 45.4 10.6 7.6 CH2,000,5N0,2 Acetate 3 53.2 10.2 7.7 CH1.700.4N0_2 Glucose This study 1: Cohen (1982) 2: Herbert (1975) 3: Speece (1964) 3.3.5 Mass balance calculation for the baseline steady-state period Removal efficiency and loading rates shown in Figure 3-5 and Table 3-6 were based upon the COD concentration in the feed and on the average COD of the solids-free effluent. As a check on the reliability of the data, a mass balance was made based on the COD input to the system and the COD output from the system. The feed glucose was the only source of COD input, while the effluent and methane were the two sources of COD output. By determining the effluent COD which included the solids-free dissolved COD and the COD equivalents of effluent solids and methane, a COD balance was made for the entire baseline operation period and is recorded in Table 3-11. Based on this calculation, an overall COD recovery of ca. 102% was achieved. The mass balance calculation results show that of the total COD removed from the glucose fed mother reactor, ca. 77 percent was converted to methane and 23 percent was converted to biomass. This result was in good agreement with theoretically predicted 40 Table 3-11. COD mass balance calculation results for the baseline data acquisition period of the mother reactor Total COD inputl 12.570 g/d Effluent soluble COD 0.479 g/d COD equivalent of biomass2 2.835 g/d COD equivalent of methane 9.557 g/d Total COD output 12.871 g/d COD recovery (%) 102 1: Calculated based on measured average influent COD concentration. 2: Calculated based on the theoretical CODN SS ratio of 1.5/1 obtained in this study. glucose conversion ratio of 75/25 (methane/biomass based on COD) in anaerobic methane fermentation ( Mosey, 1981). 3.3.6 MPN enumeration results When the anaerobic reactor reached its steady-state condition at a HRT of 10 days and a glucose feed concentration of 8 g/L, most probable number (MPN) enumeration was performed to determine some measure of the microbial community structure. The results of MPN enumeration for different microbial trophic groups in the glucose fed anaerobic reactor culture are presented in Table 3-12. Based on the MPN values, the major microorganisms in the glucose fed anaerobic reactor were fermentative bacteria. Ferrnentative bacteria contributed ca. 69% of the total culturable microbial populations present. H2- and acetate-utilizing methanogens contributed about 10% of the population. Sytrophic acetogens contributed ca. 21% of population measured via the MPN test. Based on the MPN results, the number of total anaerobic bacteria was estimated to be ca. 1.3 x 1011 cells/g VSS. 41 Table 3-12. MPN values of anaerobic microorganisms in the mother reactor evaluated using 5 tube dilution (x108/m1) 95% confidence limits (x103/ml) Substrate MPN (x103/m1) lower upper Glucose 1.10 0.40 3.00 Acetate 0.13 0.05 0.39 Butyrate 0.17 0.07 0.48 Propionate 0.17 0.07 0.48 Hydrogen 0.028 0.012 0.07 3.3.7 Microscopic examination The digester sludge used for inoculation was black in color with large amount of microbial flocks and organic and inorganic particles. A vast number of different mophotypes were present. Biomass in the steady-state mother reactor was white in color and consisted of well-separated dispersed cells. Differences in the microbial composition of these cultures were easily observed using phase contrast microscopy. The predominant morphologies in the steady-state mother reactor culture were long filamentous rods and short rods. 42 3.4 DISCUSSIONS It has been noted that in complex microbial habitats, such as anaerobic reactors, the composition of the microbial community depends on the type and amounts of substrates supplied (T oerien, 1966; Hattingh, 1967). When the composition of the substrate supplied is changed, a change in the composition of the microbial population will occur (Chynoweth and Mah, 1977). Due to the significant differences in growth rates of the various anaerobic microorganisms, adaptation will not be rapid. In the present study, the failure of a "fast start-up" attempt indicates that the inoculated community, which was acclimated to convert biomass to methane, did not quickly adapt to the new substrate. A significant imbalance among the different trophic groups occurred. In view of this, a combined batch and stepwise increased loading strategy was employed. As proposed by Barford (1989), this start-up method is essentially an empirical ”trial and error" procedure. This start-up strategy performed more smoothly and effectively. The start-up procedure of the mother reactor was completed within ca. 150 days, including ca. 50 days of batch operation. This start-up time is reasonable compared with similar anaerobic reactor start-up operations (Bae and McCarty, 1993). It has been reported that, depending on inoculation materials and target compounds, two to six months can be required for start-up of anaerobic reactors (Barford, 1989; Hickey et al, 1991a). If the inocula for reactor start-up is not adapted to the target compounds, as in the present study, longer times can be expected (Hattingh, 1967). Under steady-state conditions, a specific loading rate of 0.67 g COD/g VSS-d and a COD removal efficiency of 96% was achieved in the mother reactor. These values are relatively high when compared with values obtained from other lab-scale anaerobic reactors operated under similar conditions (Speece and McCarty, 1964; Cohen et al., 1980). The operational performance data obtained from the mother reactor during the intensive data collection period are therefore, valid reference points for comparison to the following periodic substrate perturbations. an $111111 lit-ii; in VSS . tin: 43 3.6. SUMMARY The combined batch-stepwise start-up procedure employed in this study is a workable approach to start-up glucose-fed anaerobic reactors. By this method a stable microbial community was established within ca. 150 days. Under steady-state conditions (temperature of 35 0C, HRT of 10 days and organic loading rate of 0.85 g COD/L—d) the following operational performance data were obtained in the mother reactor: COD removal of 96%, Gas Production of 430 mllL-d, VSS of 1.26 g/L and methane content of 50% in the biogas. An empirical biomass elemental formula of CH1,700.4N0.2 was obtained from this study. Compared with other lab-scale anaerobic reactors, the mother reactor was operated under optimum conditions with relatively high specific loading rate and COD removal efficiency. Reactor performance data obtained in this study are used as references for the following long-term periodic substrate perturbation studies. Chapter 4 RESPONSE OF ANAEROBIC COMMUNITY TO A LONG-TERM ONE DAY FEAST-ONE DAY FAMINE PERIODIC SUBSTRATE PERTURBATION 4.1 INTRODUCTION Methane fermentation results in the conversion of organic materials to methane and carbon dioxide in the absence of molecular oxygen. Conversion of carbohydrates, fats and proteins to methane requires the combined activity of fermentative, acetogenic, and methanogenic bacteria (Novaes, 1986; Pavlostathis, 1991). Due to the inherent slow growth rate of methanogenic and acetogenic bacteria, anaerobic systems can be sensitive to environmental changes (Lawrence, 1967). In engineered systems, such as anaerobic digesters, a sudden increase in the organic or hydraulic applied loading rate can result in serious operational problems including reactor failure-long recovery periods are often required (McCarty, 1964; McCarty and Mosey, 1991; Fongastitkul et a1, 1994). As a result, anaerobic systems, and particularly suspended growth systems, are sometimes characterized as difficult to operate (Ross and Smollen, 1981). Many studies have examined the response of anaerobic systems to environmental perturbations, especially changes in organic loading (Cohen et a1, 1981; Pavlostathis and Giraldo-Gomez, 1981; Hickey and Switzenbaum, 1991b, 1991c; Gupta et al, 1994; Fox and Suidan, 1996). These studies are of three different types: (1) pulse feed experiments in which a substrate pulse is injected suddenly into an anaerobic reactor; (2) step feed experiments in which the reactor organic or hydraulic loading rate is suddenly increased to a higher level for a period of time then returned to the steady-state condition; and (3) periodic feed experiments in which the influent substrate concentration is periodically 44 45 changed. Fongastitkul et a1. (1994) conducted a combined hydraulic-organic overloading test in a two phase UASB reactor to investigate system feasibility, maximum loading capacity and system failure and recovery. Using a lab scale anaerobic reactor Hickey and Switzenbaum (1991c) examined the acetate accumulation and its subsequent effect on the build-up of other volatile fatty acids during step feed organic and hydraulic perturbation experiments. Following a series of pulse feed perturbations, Smith and McCarty (1988) studied energetic and reaction-rate interactions between hydrogenic and hydrogentrophic bacteria. The response of anaerobic populations to a periodically fed skimmed-milk solution was studied by Mosey and Fenader ( 1981) using a lab-scale digester. These perturbation experiments have provided useful information on the response of anaerobic populations to sudden environmental changes, the upset and recovery patterns, and kinetic and thermodynamic changes. Essentially all of the perturbation experiments described to date were performed over a relatively short time period (from a few days to a few weeks). Little is known about the effects of long—term, continuous periodic substrate perturbations on anaerobic community structure and function. This is despite the fact that long-term periodic substrate loadings are common in nature and in engineered systems. Examples include aquifer environments subject to tidal influence, the rumen, high rate anaerobic processes treating effluents from food processing, agriculture, and other industries. In order to determine whether long-term adaptive changes might occur in anaerobic systems, a long term (> 200 days) periodic substrate perturbation was applied to an anaerobic chemostat. In this experiment, identical steady-state communities were established in a "mother" and a "daughter” reactor. After reaching steady-state, the daughter reactor was subjected to a periodic organic loading pattern in which the influent glucose concentration was alternately varied from 16 g/L to 0 g/L on a two day cycle. The average organic loading rate for the perturbation cycle was equal to the steady-state glucose loading rate and the dilution rate was maintained unchanged during the perturbation period. The response of the daughter reactor microbial community to the 46 periodic substrate perturbation, in terms of operational performance, substrate kinetics, genetic and overall microbial composition changes, were investigated following the initiation of the substrate perturbation. 4.2 MATERIALS AND METHODS 4.2.1 Anaerobic reactor A schematic representation of the experimental apparatus used for the perturbation experiment is shown in Figure 4-1. The cylindrical all glass reactor (Constructed by the Glass Shop, Chemistry Department, Michigan State University), was approximately 15 cm in diameter and 12 cm in height. The reactor had a working volume of 1.5 liters and a headspace of 0.5 liters. Three, 2-cm diameter threaded all glass outlets and two 0.5-cm diameter glass outlets located in top of the reactor were used to connect with liquid and gas sampling valves and gas production meter. The substrate and nutrient solutions were injected into the reactor through a rubber septum. A bell shaped, internal mixing chamber was installed to enhance mixing and to improve liquid and gas phase mass transfer. Liquid continuously entered the internal mixing chamber through four 0.5 cm diameter holes evenly located near the reactor liquid surface and was "pumped" out of the mixing chamber through four 0.5 cm diameter holes located near the bottom of the chamber. A spinning stir bar, located inside the mixing chamber, acted as an impeller and created a high speed down-flow liquid stream that continuously entrained gas from the reactor headspace mixing it with liquid inside the chamber. The magnetic stir bar was driven by a Corning Model PC-310 magnetic stir plate (Corning Inc., Corning, New York) at a speed of 500 run per minute. Continuous feed was introduced into the reactor with a timer controlled pump system. Two Watson-Marlow Model lOlU/R pumps (W atson-Marlow, Inc., Wilmington, MA) were used to alternately supply glucose plus mineral media and mineral media alone (no glucose) at a flow rate 150 ml/d to the daughter reactor. Power to the Watson-Marlow 101 47 $008 .8052 0.8020 .t ..... ...... .............. ................ ........... ......... ......... ......... ............ ....... ............ :05: 6.5200 F 9.5. to: 000: @5383 :: .00 3:08.093: TV 0.3:... .8820 9.38 3808. .8000: :8:: 805:2 @5383 0.00 8028.0 1.. .308 000 48 U/R pumps was channeled through a electronic ChronTrol XT timer (ChronTrol Co., San Diego, CA). This programmable timer was used to control on/off operation of the two feed pumps to obtain the desired substrate feed schedule. The nutrient solution was directly injected into the reactor from a 60 ml plastic B-D syringes (Becton Dickinson & CO., Rutherford, NJ) mounted on a Harvard Model 22 syringe infusion pump (Harvard Apparatus, South Natick, Mass). Nutrients were supplied separately to minimize growth in the glucose feed reservoirs. Effluent exited the reactor through an inverted U-shaped air-tight water seal effluent tube to prevent the reactor content from being exposed to oxygen. Liquid samples were taken from the reactor liquid sampling port at approximately the mid-point of the reactor liquid phase. Samples were centrifuged using a Hennle compact centrifuge (Model 2203, National Labnet Co., Woodbridge, H1) at 3,500 g for 10 minutes, and either analde immediately or frozen at -30 0C and stored for analysis. Gas production was quantified using a digital liquid displacement based gas meter and discharged. Gas samples were drawn from the gas sampling port located at the top of the anaerobic reactor. The entire apparatus was located in a constant temperature room (Nor-Lake Scientific, Hudson, WI) set at a temperature of 35 $ 0.5 0C. 4.2.2 Substrate and nutrients Glucose was supplied as the sole carbon and energy source. Analytical reagent grade granular glucose (Mallinckrodt Specific Chemical Co., Paris, KE) was used to prepare the 8 g/L and 16 g/L glucose solutions. Sodium bicarbonate was added to the feed to provide buffer for pH control. A buffer solution, prepared with reagent grade powder sodium bicarbonate (J, T. Baker Inc., Philipsburg, NJ), was mixed with the above glucose solutions to obtain 7 g/L and 14 g/L NaHCOg respectively. Both the glucose solution and the sodium bicarbonate buffer solution were prepared with deionized water. The mineral nutrient solution was selected based on the compounds and their concentrations used by several researchers (Cohn et a1. 1982; Zoetemeyer et al. 1982a; 49 Zehnder et a1. 1987). The concentration of the various components in the substrate feed solution is presented in Table 4-1. In order to dissolve the salts in the stock nutrient solution, the pH was reduced by the addition of 30 ml concentrated HCl per 1 liter of stock nutrient solution. The prepared stock nutrient solution was stored at 4 0C. This stock solution was diluted to prepare the nutrient feed solution for the anaerobic reactors. TABLE 4-1. Concentration of the nutrients in the combined influent fed to the anaerobic reactor Chemicals Concentration (mg/l) NH4C1 1070 KH2PO4 275 NaCl 120 MgC12-6H20 49 NazSO4 43 FeSO4-7H20 5.53 MnC12-4H20 2.02 CaC12-2H20 0.59 ZnC12 0.67 NiC12-6H20 0.65 Cu C12-2H20 0.16 C0C12-6H20 0.48 H3BO3 0.063 NazMoO4-2H20 0.0045 50 4.2.3 Analytical methods Reactor performance was monitored daily for effluent pH, total gas production and gas composition. Analyses for COD, VFAs and VSS were performed on a daily to weekly basis, depending on reactor performance. Analytical methods used for determination of COD, pH, VSS, VFA, gas composition are described in Appendix I. Epifluorescence microscopy direct count methods, which were used for estimating living cells in the anaerobic reactor cultures, are described in Appendix C. Protein was determined using a commercial Coomassie protein assay reagent (No. 23200, Pierce, Rockford, IL 61105). The sampling and analysis procedures for protein measurement are described in Appendix D. Microscopic observations were performed using an Olympus BH-2 phase-contract microscope (Scientific Supply Company, Schiller Park, IL 60176). 4.2.4 Measurement of maximum substrate conversion rate When the perturbed daughter reactor again reached steady-state, the maximum substrate conversion rates by cultures in the mother and the daughter reactors were measured based on substrate consumption. Maximum substrate conversion rates (Vmax) were estimated from the glucose and glucose metabolic intermediates degradation curves following pulse substrate injection. Assuming the substrate degradation can be described by Michaelis-Menten kinetics, the rate of substrate utilization becomes: -dS/dt = v'max ms + KS) (44) Where: V'max = maximum volumetric rate of substrate conversion, mg/l-d S = concentration of growth-limiting substrate, mg/l Ks = substrate saturation constant, mg/l 51 After addition of high substrate concentration, S > Ks, the substrate utilization equation can be written as: -ds/dt == V'max (4-2) Therefore as indicated by equation (4-2), at high substrate concentrations, substrate conversion rate is independent of substrate concentration, following zero-order kinetics. That is true if there is no substrate inhibition and the increase of biomass due to growth is small compared with the total biomass concentration. V'max can be estimated by using the integrated form of equation (4-2): 8 = So - V'max t (4-3) Where: S = substrate concentration at time t, mg/l So = original substrate concentration, mg/l t = reaction time, day Specific maximum conversion rates were then determined by: Vmax = V'max/x (4'4) Where: X = reactor biomass concentration, mg/l In this experiment, the maximum substrate conversion rates for different substrates, except hydrogen, were determined directly in the mother and the daughter reactors by pulse injecting concentrated stock solutions of different substrates to obtain an elevated substrate concentration in the systems. Maximum hydrogen conversion rates were measured indirectly in pH controlled batch experiments with 158 ml serum bottles. The substrate concentrations used for the maximum substrate conversion rate measurements are listed in Table 4-2. 52 Table 4-2. Substrate concentrations applied to the mother and the daughter reactors during the maximum substrate conversion rate measurements Substrate Concentration (mg/L) Glucose 200 Butyrate 300 Propionate 300 Acetate 600 Hydrogen 20 (psig) 4.2.5 Measurement of decay coefficients of biomass and substrate utilization capacities Under starvation conditions the change in biomass concentration is proportional to the existing biomass concentration (McCarty, 1966): dX/dt = - Kd X (4-5) Where: X = biomass concentration, g/l Kd = decay coefficient, day‘1 Kd can be estimated using the integrated form of equation (4-5): lnX=lnX0-Kdt (4-6) Where: i X = biomass concentration at time t, g/l X0 = original biomass concentration, g/l t = decay time, day 53 Biomass decay coefficients were measured in batch experiments using 158 m1 serum bottles using cultures taken from the anaerobic reactors. For mixed cultures, like anaerobic sludge, it is difficult to directly measure the actual decay rate of the individual trophic groups of bacteria. Assuming the maximum substrate conversion rates possessed by the anaerobic cultures are proportional to the active population for that substrate, the decrease in the maximum substrate conversion rates under starvation conditions will directly reflect the decay rate of the associated trophic group. To use this concept, the decay coefficients of the different trophic groups can also be estimated by using equation (4-6). In this case the only difference is that the terms of X and x0 in equation (445) are replaced by vmax and vmax, 0 respectively (Wu etal., 1995). 4.2.6 Reactor inoculation The inoculum (1,500 ml) for the daughter reactor was obtained from a 15 L laboratory-scale mother reactor operated under steady state conditions with a 10 day HRT and a constant 8 g/L glucose feed, as described in Chapter 3. The mother reactor was operated under steady-state conditions for more than two hundred days before the daughter reactor was started. The mother reactor also served as a stable control and as reference for changes in activity levels and community structure in the daughter reactor. The inoculated "daughter" reactor was initially operated under the same conditions as the mother reactor (10 day HRT with continuous input of 8 g/l glucose) for 50 days. Once steady-state was reached and sufficient "base-line" information obtained, a square wave perturbation in feed concentration was applied to the daughter reactor. 4.2.7 Experimental procedure _ In this study, a long term periodic substrate perturbation was applied to the anaerobic "daughter" reactor. The response of the system to the perturbation was investigated. During the perturbation, the reactor influent glucose concentration was alternately changed in amplitude from 16 g/L to 0 g/L (mineral media only) on a 2 day square wave cycle (see 54 Figure 4—2), i.e. the reactor was fed 16 g/l glucose solution for one day followed by glucose-free mineral media at the same flow rate on the next day. A 10 day hydraulic retention time (dilution rate of 0.1 d'l) was maintained throughout the experiment. The perturbation loading pattern was chosen so that the average influent glucose concentration during the perturbation period was equal to the steady-state concentration of 8 g/L applied to the control mother reactor. The daughter reactor was continuously operated for a period of 250 days including ca. 50 days under steady-state baseline conditions and ca. 200 days under perturbation conditions. A.I\MIV ICII.II!IIIII0':IIIII.I 5.3:.II.'II.«I ' Glucose concentration (g/l) 55 Perturbation Conc. (g/l) 20 - "'""'" Steady-state Cone. (g/I) 16 " r" '—"r — *——r —' ‘r—- 12 " 3.---u_--—---—----—---a-—------—n—-----u--- .1 O u 0 2 4 6 8 10 12 Time (Day) Figure 4—2. Substrate feeding pattern used for perturbation experiment 56 4.3 RESULTS 4.3.1 Effects of the periodic substrate perturbation Based on the experimental results, the response of the daughter reactor community to the periodic substrate perturbation can be broken down into four distinct stages: (1) rapid VFA and COD accumulation; (2) establishment of a metastable steady-state at reduced COD removal efficiency; (3) rapid VFA degradation; and (4) re-establishment of steady- state with high COD removal efficiency. Stage 1 can be further divided into two sub- stages: (1a) rapid acetate and propionate accumulation, and (1b) rapid butyrate accumulation while acetate and propionate levels stabilized. Each of the above stages is discussed in greater detail in the following sections. Stage 1a (day 0 to day 20) As shown in Figure 4—3, the steady-state microbial community was highly sensitive to the variation in organic feed concentration. A rapid accumulation of VFAs occurred nearly immediately after the start of the substrate perturbation. The acetate-utilizing methanogens were immediately impacted by the substrate perturbation as evidenced by a sharp increase in the effluent acetate concentration and a drop in methane production. Effluent acetate concentration increased almost linearly from 120 mg/l to ca. 2,200 mg/l, accumulating at a rate of 105 mg/l-d. Effluent COD concentration continuously increased from 370 mg/L to ca. 3,800 mg/L, accumulating at a rate of 170 mg/l-d (Figure 4-4). Thus, acetate contributed ca. 65% of the total increase in COD observed during this period. Subsequently, acetate concentration stabilized. A linear accumulation of propionate was also observed during the initial 20 days of the substrate perturbation. Effluent propionate concentration increased from less than 40 m g/l to a peak concentration of ca. 700 mg/l at an average daily accumulating rate of ca. 33 mg/l-d. Thus propionate contributed about 30% of the total accumulated COD. Propionate formation from glucose can be written as (Thauer et al., 1977): C5H1205 = 4/3CH3CH2COO‘ + 2/3CH3COO’ +2/3HC03' +8/3H+ (4-7) 57 a, a te.: C #30763. ...°a_'-m 0302‘- 0530 4 0.4 ‘ Mother 0.0 ' I ‘ 1 j i ‘ 0 10 20 30 Time (day) 2.0 y = 1.6108 * 10"(-7.6644e-2x) R"2 = 0.992 y = 1.0195 "' 10"(-7.1233e-3x) R"2 = 0.888 1.6 3 1.2 ' B ‘8 0.8 - > 0.4 ' Daughter 0.0 ' I T l l 0 10 20 30 Time (day) Figure 4-10. Estimation of biomass decay parameters for the mother and the daughter reactor cultures. 71 As presented in Table 4-5, the daughter reactor culture had significantly higher first stage decay rates. The second stage decay rate of the daughter reactor culture was essentially the same as the decay rate observed in the mother reactor culture. The biomass decay rates observed in the second stage (Kdz) are very close to the literature reported values of 0.020 to 0.025 d'1 for glucose fed anaerobic sludge (Stewart, 1958; Agardy et al., 1963). b. Decay characteristics of hydrogen-utilizing methanogens and acidogens during a starvation period Cultures taken from the mother and the daughter reactors were transferred into 158 ml anaerobic serum bottles and incubated at 35 0C without feed. Maximum substrate utilization rates were measured at different incubation times using concentrated substrates (glucose or H2-C02 gas) over a period of ca. 70 days. The decay rates of the hydrogen-utilizing methanogens in the mother and the daughter reactor cultures are presented in Figure 4-11. From this figure it can be seen that the decay curves of the hydrogen-utilizing methanogens could be divided into two stages. During the first stage, the maximum hydrogen utilization capacity decreased rapidly, while in the second stage a much slower decay rate was observed. The decay rates observed in the daughter reactor culture, both in the first and the second stages, were significantly lower than the values observed in the mother reactor culture. The decay progress of the acidogens in the mother and the daughter reactor cultures is shown in Figure 4-12. From this figure it can be seen that the decay process of the acidogens in the mother and the daughter reactor cultures also can be divided into two stages: in the first stage the maximum glucose degradation capacity decreased rapidly with decay rates of 0.24 d'1 and 0.17 d'1 for the mother and the daughter reactor cultures, respectively; during the second stage the decay rate was relatively slow with calculated rates of 0.019 d'1 and 0.023 d'1 for the mother and the daughter reactor cultures, respectively. 72 1.00 y = 1.0010 "' 10"(-5.2563e-2x) R"2 = 1.000 y = 0.49093 * 10“(-2.9459e-3x) R"2 = 0.952 O E > Time (day) 1-00 y = 1.0173 * 10"(-8.2980e-3x) 1w = 0.982 y = 0.63393 * 10"(—l.5582e—3x) R"2 = 0.931 0.80 " o 0.60 i Z . > 0.40 ‘ 1 0.20 ‘ Daughter 0.“) ' I f l ' I ' I ' I ' I ' I 0 10 20 30 40 50 60 70 Time (day) Figure 4-11. Estimation of decay parameters for hydrogen-utilization activity in the mother and the daughter reactor cultures. 73 1.00 y = 1.0113 * 10"(-0.10309x) R"2 = 0.942 y = 0.11570 * lO“(-8.3508e-3x) R"2 = 0.970 0.80 ‘ O 0.60 ' Z . > 0.40 ‘ J 0.20 " # Mother 000 , r . . m 0 40 50 60 70 Time (day) 1.00 y = 0.92413 * 10"(-6.37l7e-2x) R"2 = 0.849 y = 0.14854 "' 10"(-9.8964e-3x) R"2 = 0.902 0.80 c 0.60 ‘ Z > 0.40 " 0.20 ‘ Daughter 0.00 ' ' 0 10 20 30 40 50 60 70 Time (day) Figure 4—12. Estimation of decay parameters for glucose degradation activity in the mother and the daughter reactor cultures. 74 Table 4-6. Decay rates of the glucose degrader and hydrogen-utilizing methanogens in the mother and the daughter reactor cultures Mother reactor Daughter reactor Gimme K41 (d']) 0.240 0.170 K42 (d'l) 0.019 0.023 Hydrogen Kdr (d-l) 0.120 0.019 Kdg (d-l) 0.007 0.004 The decay rates, estimated based on the reduction of the maximum substrate utilization rates, of the acidogens and hydrogen-utilizing methanogens in the mother and the daughter reactor cultures are summarized in Table 4-6. 75 4.3.5 Protein measurement results. Changes in protein concentrations were investigated when the daughter reactor again established steady-state. Samples were collected on a daily basis for two weeks from the mother and the daughter reactors. More intensive sample collection (every 4 to 6 hours) was employed for the daughter reactor culture over two complete perturbation cycles (4 days). During this period, changes in VSS and COD were also monitored. Cyclic changes in protein and VSS concentrations were observed in the periodically fed daughter reactor (Figure 4- 13). Although both protein and VSS concentrations responded positively to the periodic substrate feed, much larger fluctuations in VSS concentrations were observed (Figure 4-14). This indicates that the change in protein concentration was not proportional to changes in VSS concentration. A large variation in the ratio of protein/V SS occurred during the substrate perturbation period (Figure 4-153). The daily change in the ratio of protein/V SS varied inversely with reactor VSS changes (Figure 4-15b). The highest ratio of protein/V SS (and lowest VSS) was always observed at the end of the no—substrate period, while the lowest ratio of protein/V SS (and the highest VSS) was observed at the end of the substrate feed period. The protein and VSS measurement results of the mother reactor culture are presented in Figures 4-13 and 4—15, respectively. Based on the experimental data, the average protein concentration of the mother reactor culture was 462 mg/L, while the average VSS concentration was 1,183 mg/L. This gives an average proteinN SS ratio of 0.391 in the mother reactor culture. The protein and VSS measurement results obtained during this experiment are summarized in Table 4-7. The average protein/V SS ratio measured in the mother reactor culture was higher than the average value measured in the daughter reactor culture. That was mainly due to the high VSS concentration observed in the daughter reactor culture, because the average protein concentration was essentially the same in the mother and in Protein (mg/L) VSS (mg/L) 76 l4 600 - a 500 r 400 ' 300 ‘ 200 r 100 .. -—D— Daughter . + Mother 0 ' I ' l ' II ' I v I f 1 r 0 2 4 6 8 10 12 Time (day) 2000 b 1600 ' 1200 ‘ 800 - 400 d —D— Daughter 1 —0'— Mother 0 f I ' l V I v I v I u l 0 2 4 6 8 10 12 Time (day) 14 Figure 4-13. Protein and biomass concentration changes in the mother and the daughter reactor cultures. 77 2000 1600 Q 1200‘ a” r v —0— Protein E 800- —e— vss 400 4W 0 ‘ ' ' ‘ 0 1 2 3 4 Time (day) Figure 4-14. Protein and biomass concentration changes in the daughter reactor during two complete perturbation cycles. 78 —0— Mother —0— Daughter 0.6 0.5 - m c m g 0.4 - E e 2 0.3 . a. u d O -1 3 0.2 E II 0.1 4 . . 0.0 l ' I ' 1 ' U 6 8 10 12 14 Time (day) m E ,. :1 '5’. e E =- O.30r 1 a... 800 U) a ‘ gr ,3 0.20" ' a ‘ .- °‘ 010. —D— ProteinNSS 400 ' . —O— VSS 0.00 . ' . ‘ 0 o 2 3 4 Time (day) Figure 4-15. Protein/V SS ratio changes (a) in the mother and the daughter reactor cultures during the experimental period, and (b) in the daughter reactor during two complete perturbation cycles. 79 Table 4-7. Summary of protein and VSS measurement results Protein vss PNSS COD (g/L) (mg/L) (mg/L) Mother 462126 11831 87 039110.024 369123 Daughter Ave. 439139 13571158 032410.027 340188 Daughter(l6g/L)1 43120 15231 79 031110.012 416132 Daughter(0g/L)2 45116 11911 64 0.33910013 264149 1: Data obtained in the end of 16g/L glucose feed period. 2: Data obtained in the end of 0 g/L glucose feed period. the daughter reactor cultures. The values of the protein content in total biomass obtained in this experiment are in good agreement with values reported in the literature. Hattingh et a1. (1967) reported that a protein/V SS ratio of 0.362 was obtained for anaerobic sludge at the end of the sludge adaptation period. 4.3.6 Enumeration results Epifluorescence microscopy direct counting methods were employed to estimate total living cells in the daughter reactor and the mother reactor cultures, while the MPN enumeration method was used to estimate the distribution of each major trophic group of anaerobic bacteria in the daughter reactor and the mother reactor. a. Epifluorescence microscopy direct count results Epifluorescence microscopy using acridine orange or DTAF as fluorescent dyes is used worldwide to determine the total number of microorganisms in environmental samples (Schirndt et al., 1982; Bratbak, 1985; Bitten, et al., 1993). Fluorescent dyes can be used to improve the visualization of individual microorganisms by binding to specific cell components, such as protein and DNA, and fluorescing under UV light. The living cells 80 stained by acridine orange (DNA stain) fluoresced orange under UV light, while the living cells stained by DTAF (protein stain) fluoresced green under UV light. Samples taken from the mother and the daughter reactors were diluted to 1 x 104, and analyzed for total bacterial number by using acridine orange and DTAF direct counting methods (Appendix C). One of the DTAF direct counting results for the mother and the daughter reactor cultures is presented in Table 48. Based on the experimental results, the total number of living cells in the re-established daughter reactor was twice the number found in the mother reactor culture. Table 4-8. DTAF direct counting results for the mother and the daughter reactor cultures Mother reactor Daughter reactor (x 108/m1) (x 108/m1) 2.74 4.48 2.17 5.74 2.38 5.36 2.07 4.90 2.66 5.15 2.47 4.41 2.42 $0.24 4.96 i053 One of the acridine orange direct counting results for the mother and daughter reactor cultures are shown in Table 4-9. Higher values for the total number of living cells in both of the mother and the daughter reactor cultures were estimated by the acridine orange direct counting method. However the ratio of the total number of living cells in the daughter and the mother reactor cultures was close (ca. 2:1) to the value estimated by DTAF direct counting method. 81 Table 4-9. Acridine orange direct counting results for the mother and the daughter reactor cultures Mother reactor Daughter reactor (x 103/ml) (x 103/ml) 3.10 5.31 3.23 6.36 3.04 6.64 2.71 5.27 2.95 4.65 2.83 5.84 3.44 5.76 3.04 $0.23 5.69 21:0.63 The total number of living cells in the mother and the daughter reactor cultures was estimated by two independent tests using DTAF and acridine orange direct counting methods. The total number of living cells estimated in this experiment are summarized in Table 4-10. As illustrated by the experimental results, the culture in the daughter reactor contained ca. two times more living cells than the culture in the mother reactor. The enumeration results suggest that the microorganisms in the daughter reactor culture were relatively small in size since the average VSS concentration in the daughter reactor was just slightly higher than the VSS concentration in the mother reactor (< 1.2 times). This was confirmed by direct microscope observation. 82 Table 4-10. Summary of DTAF and acridine orange direct counting results for the mother and the daughter reactor cultures Mother reactor Daughter reactor DIM (x 1081611) (x 108/m1) DTAF 2.42 110.24 4.96 1053 2.05 2.44 $0.27 5.05 $0.34 2.07 A0 3.04 10.23 5.69 21037 1.87 4.01 21:0.34 6.54 i053 1.63 Average 2.98 5.56 1.87 Blank runs, using DI. water, were performed with each experiment. The total bacterial number estimated from the blank was 1,000 times lower than the number obtained from reactor culture samples. b. MPN enumeration results In order to estimate the bacterial distributions of the major trohpic groups, a MPN enumeration was performed for the daughter reactor culture according to the procedure provided in Appendix I. The MPN enumeration results are presented in Table 4-11. Based on the MPN values, the major anaerobic microorganisms in the periodically perturbed daughter reactor were fermentative bacteria. They contributed to 78% of the total culturable microbial populations presented in the anaerobic reactor. H2- and acetate- utilizing methanogens contributed about 10.7% of the population. Sytrophic acetogens contributed ca. 11% of population. The proportion of methanogens present in the total culturable anaerobic bacteria population were in the range reported previously, 1 to 10%, in MPN enumeration tests (Anderson et al., 1994; Zhang and Noike, 1994) 83 Table 4-11. MPN values of anaerobic bacteria in the re-established daughter reactor culture estimated using 5 tube dilution (x108lml) 95% confidence limits (x103/ml) Substrate MPN (x103/ml) lower upper Glucose 1.70 0.70 4.80 Acetate 0.17 0.07 0.48 Butyrate 0.17 0.08 0.41 Propionate 0.07 0.03 0.21 Hydrogen 0.06 0.03 0.18 A comparison of the MPN enumeration results from the mother and the daughter reactors shows some differences between these two cultures. The total number of culturable anaerobic bacteria in the daughter reactor was significantly higher (ca. 1.4 times) than the number in the mother reactor. Differences could also been observed in the distribution of the major trophic groups in the total culturable microbial populations. MPN enumeration results were in good agreement with results obtained from epifluorescence microscopy direct counting experiments. Both enumeration results indicate a higher number of anaerobic bacteria in the daughter reactor culture. Based on the total number of anaerobic bacteria estimated by epifluorescence microscopy direct counting methods, the MPN enumeration method cultured ca. 38% to 52% of viable anaerobic bacteria present in the anaerobic cultures. 84 4.3.7 Microscopic observations Differences in the microbial composition of the daughter reactor and the mother reactor cultures were easily observed using phase-contrast microscopy. Cultures from the daughter reactor were white in color with mostly dispersed cells. The microbial community in the daughter reactor appeared more homogeneous with less diversity. The predominant morphologies were short rods and cocci, with some long spirillum type bacteria (Figure 4-16a, scale bar in the photo represents 2 run). In contrast, the microbial community in the mother reactor appeared to be more diverse, being comprised of more different morphotypes (Figure 4-16b). The predominant morphologies were long filamentous rods and short rods. 85 Figure 4— 16. Microscopic observation of the mother (a) and daughter (b) reactor cultures. 86 4.4 DISCUSSION For complete conversion of carbohydrates, such as glucose, to methane, five groups of bacteria: fermentative bacteria, propionate- and butyrate-utilizing aceto gens and hydrogen- and acetate-utilizing methanogens are required (Mosey, 1983; Novaes, 1986; Pavlostathis and Giraldo-Gomez, 1991). These bacteria must work syntrophically as they are linked physiologically, kinetically and thermodynamically. Sudden environmental changes can cause changes in individual groups which eventually affect the whole microbial community (McCarty, 1964; Fongastitkul, 1994). Changes in the concentrations of intermediate volatile fatty acids during the perturbation period indicate that all the major groups of the anaerobic microbial community were impacted by the application of a substrate perturbation to the anaerobic chemostat. A change in the fermentation pattern was accompanied by a shift in the predominant microbial populations. Typically, acetate-utilizing methanogens are responsible for the majority of COD removal and organic compound stabilization. (McCarty, 1964; Mosey 1981). The acetate- utilizing methanogens function at relatively high saturation levels (the ratio of steady-state substrate utilization rate Vt to the maximum substrate utilization rate Vmax) compared to other anaerobic bacteria. Kaspar and Wuhrrnann (1978) observed that the saturation level for acetate decarboxylation was about 43%. Similar results were reported by Valcke and Verstraete (1983) in their study of sludge fermentation. In the present study, a saturation level of 45 % was estimated under the baseline steady-state (constant feed) conditions. This indicates that the acetate-utilizing methanogens have limited additional capacity to handle sudden substrate increases. The acetate-utilizing methanogens are also among the slowest growing bacteria in the anaerobic microbial community with an experimentally observed growth yield of ca. 0.025 to 0.04 g VSS/g acetate (Lawrence and McCarty, 1967; Ahring and Westerrnann, 1987; Aivasidis et al., 1988). Consequently, these organisms are sensitive to environmental changes, and long periods are likely to be 87 required for this group of bacteria to adjust to new environmental conditions. This expectation was confirmed by the results. Gas production on "famine days" (when glucose was not present in the feed) was a useful indicator of acetate-catabolizing methanogenic activity. After initiating the substrate perturbation, gas production on such days declined rapidly (Figure 4—7). This indicates a significant reduction in methanogenic acetate-catabolizing activity. Gas production on famine days decreased to less than 5 ml/d (detection limit) for about 30 days. Subsequently, limited gas production resumed on famine days, indicating a shift in the predominant acetate-utilizing population and/or an adaptation of the original acetate- catabolizing methanogenic bacteria. It took about 90 days for the complete recovery of acetate-catabolizing methanogenic activity. The optimum pH for methane fermentation is between pH 6.5 to 7.4 (McCarty, 1964; Zehnder et al., 1981; Aivasidis et al., 1988; Attal, et a1. 1988). At a pH value of 7.0, the methanogens could tolerate an acetate concentration as high as 6,000 to 7,000 mg/L without loss of normal acetate metabolic capacity, but they were quite sensitive to pH change (Duarte and Anderson, 1982; van den Berg et al., 1976; Yang and Okos, 1987). Yang and Okos (1987) published a study on a pure culture of methanogens utilizing acetate at pH 7 and 35 0C. In their study, the optimum acetate concentrations for the methanogenic bacteria Methanosarsina barkeri and Methanosarsina mazei were 7,000 and 6,800 m g/L, respectively. van den Berg et a1 (1976) reported similar results for an acetate enrichment culture operated at pH 7 and 35 0C, in which acetate conversion rates were not significantly affected by acetate concentration up to 6,000 mg/L, but when the rextor pH dropped to below 6.5, methanogenic activity decreased rapidly. In a pilot study, Duarte and Anderson (1982) found that a pH level of 6.4 was critical for a continuously fed acetate catabolizing digester operated at a 4.5 day hydraulic retention time and 35 9C. When reactor pH decreased from 6.5 to 6.3, methane production decreased by 65%, even though the digester had previously operated successfully at the same acetate concentration 88 (ca. 2,500 mg/L). This pH inhibition of acetate metabolism was attributed to either the pH drop itself or to the increased unionized acetic acid concentration at the lower pH. McCarty (1964b) emphasized that a decrease in pH in anaerobic reactors should be considered as a sign of a imbalance between the different anaerobic trophic groups rather than the cause of the imbalance. This conclusion was also conformed with our experimental results. In the beginning of the perturbation experiment, due to the sudden change of the substrate feeding pattern, VFAs, especially acetate, were formed more rapidly than they were removed. The accumulated VFAs destroyed the reactor buffering capacity resulting in a decrease in reactor pH (Figure 4-5). The lower pH then, in turn, discouraged the growth of methanogens and, therefore, seriously prolonged the recovery process of the acetate-utilizing methanegons. This may help to explain why such a protracted period was required for the anaerobic community to adapt to the periodic substrate feeding conditions. Although glucose accumulation was not observed throughout the whole experimental period, the fermentative bacteria were also affected by the periodic substrate perturbation. The most evident sign of this was the dramatic change in glucose fermentation products. For glucose, two complementary fermentation routes can occur (Cohen et al, 1979; Zoetemeyer et al, 1982a; McCarty and Mosey, 1991; Fongastitkul et a1, 1994). Butyrate type fermentation is characterized by production of acetate, butyrate, carbon dioxide and hydrogen as the main fermentation products. Propionate type fermentation is characterized by the formation of acetate, propionate, and carbon dioxide, with much lower hydrogen concentrations . Based on thermodynamic calculations, Duarte and Anderson (1982) proposed that under normal conditions (pH 7 and temperature 35 0C), the butyrate type fermentation is thermodynamically favored over the propionate type fermentation. Normally, the butyrate fermentation route is predominant in anaerobic reactors (McCarty and Mosey, 1991). This is indirectly supported by observations that anaerobic systems usually have limited 89 capacity for propionate metabolism but can metabolize butyrate at relatively high rates (Bae and McCarty, 1994). Thus, in anaerobic systems, butyrate is only present in trace amounts. This phenomenon was also observed in this experiment for the steady-state mother reactor and the daughter reactor prior to' the start of the substrate perturbation (Table 44). Compared with the maximum butyrate degradation rate of 250 mg/g VSS-d, the maximum propionate degradation rate of 4.8 mg/ g VSS-d possessed by the culture in the mother reactor is negligible. During the perturbation period, glucose fermentation quickly shifted from the butyrate fermentation to a mixed butyrate-propionate fermentation. Propionate accumulation occurred immediately after initiating the perturbation (Figure 4-3). Since the propionate degradation ability of the microbial community was extremely limited, the observed propionate accumulation rate could be considered to approximate the actual pr0pionate production rate from glucose. Mass balance calculations based on COD, indicated that during Stage 1a of the substrate perturbation, propionate fermentation accounted for as much as 37% of substrate flow. Prior to the initiation of the perturbation, propionate accumulation was observed to be less than 3% of substrate flow. Subsequently, accelerated butyrate accumulation (Stage 1b) and a mixed butyrate- propionate type fermentation (Stage 2) were observed. During this period propionate accumulated continuously but at a reduced rate. The substrate flow through propionate fermentation decreased from ca. 37% to 12%. The molar ratio of accumulated butyrate to propionate was 1.8:1. This mixed butyrate-propionate type fermentation continued for about 70 days during the metastable steady-state. As the final transition period began, the butyrate type fermentation again became predominant. The lithotropic Hz-utilizing methanogens were the organisms least affected in the anaerobic community. H2 turnover rate appeared was affected over only a short period of time (the first 20 days). Mosey and Fernandez (1989) reported that the H2-utilizing methanogens are relatively fast growing species, with an estimated doubling time of 6 90 hours. Kasper et al (1978) reported that under steady-state condition the saturation level of the H2 utilizing methanogens was only about 1%. There is, therefore, a vast extra utilization capacity available for increased hydrogen oxidation during the substrate perturbation. The hydrogen-utilizing methanogens have also been found to be quite resistant to environmental changes. Hobson and Shaw (1978) found that at a pH of 7.1, an acetate or butyrate concentration up to 10,000 mg/l and propionate concentration up to 1,000 mglL (as acetic acid) did not inhibit M. formicicum, the principal hydrogen-utilizing bacterium in a piggery-waste digester. Results of the present study indicate that the H2 utilizing-methanogens in the mother reactor were resistant to changes in VFA concentration and pH (See Appendix V). The maximum H2 utilization rate changed by only ca. 15% over a pH range of 5.5 to 7.5. These findings are in good agreement with the work of Attal et al. (1988) who reported an Optimum pH range of 5.4 to 7.4 for hydrogen-utilizing methanogens. These characteristics apparently endow the Hz-utilizing methanogens with the ability to quickly and effectively respond to changes in substrate levels. Results of protein analysis show that the average protein concentration in the mother and the daughter reactor cultures were essentially the same, but a much larger fluctuation in VSS concentration was observed in the daughter reactor. It appears that microorganisms in the daughter reactor acquired the ability to produce "stress products" during the feast period and then consume the stored products during the following famine period. This apparent storage ability acquired by the daughter reactor community allowed likely it to better tolerate the periodic substrate feed condition. The substrate storage capacity was most likely possessed by the acidogens, because: (1) this trophic group constituted the majority of the anaerobic population (79% based on MPN enumeration), and (2) the acidogens are the only major trophic group which could quickly respond to the substrate concentration change. This interpretation was supported by the decay test results. Based on results in Table 4-5 and 4-6, the decay of biomass in the daughter 91 reactor culture was faster than the mother reactor culture (Kdl). However, the maximum glucose utilization activiity of the daughter reactor culture decreased more slowly than that of the mother reactor culture (K61)- A reasonable explain is that the stored substrate was consumed during the first stage of the starvation period of the daughter reactor culture resulting in a ”faster" decay based on volatile solids concentration; while more acidogens remained alive due to the consumption of the stored substrate resulting in ”slower" decay based on maximum substrate utilization rate. Substrate storage by acidogens has also being observed by the other researchers in glucose fed anaerobic reactors (Bae and McCarty, 1994). The results in the perturbation experiment indicate that the structure of the anaerobic community gradually changed, enabling more effective utilization of substrate fed in a cyclic pattern. The remarkable morphological differences between the cultures in the mother and the daughter reactors suggest that the ability of the daughter reactor to adapt to the periodic substrate perturbation was most likely the result of a shift in the predominant species. Kinetic and enumeration results indicated that through a long-term selection the originally dominant microorganisms were replaced by some new/or adapted species which were smaller in size (in average) and were capably of higher substrate utilization rates and reduced decay rate. DNA and FAME (see Chapter 7 and 8) analyses support this interpretation. 92 4.5 SUMMARY This study establishes that the initial steady-state anaerobic microbial community was sensitive to fluctuations in the pattern of organic loading rate of the carbon source. The rapid accumulation of VFAs reflects an apparent decrease in the number or activity of acetate-utilizing methanogens and acetogens during the initial 110 days. Fermentative bacteria were also impacted by the substrate perturbation. A shift in the products of glucose fermentation occurred during the first and the second stages of the perturbation. The pH decrease associated with VFA accumulation appeared to be, in-part, responsible for inhibiting acetate-utilizin g methanogenic activity and prolonged the system recovery process. The anaerobic system did gradually adapt to the perturbation conditions through a long term change in community structure as evidenced by changes in the kinetics of substrate utilization and decay, microbial composition and morphological changes. Chapter 5 RESPONSE OF ANAEROBIC COMMUNITY TO THREE DAY F EAST-THREE DAY FAMINE LONG-TERM PERIODIC SUBSTRATE PERTURBATION 5.1 INTRODUCTION Conversion of carbohydrates, fats and proteins to methane requires the combined activity of fermentative, acetogenic, and methanogenic bacteria (McCarty, 1981; Novaes, 1986; Pavlostathiand Giraldo-Gomez, 1991). Many studies have examined the response of anaerobic systems to environmental perturbations, especially changes in organic loading (Cohen et al., 1981; Smith and McCarty, 1989; Hickey and Switzenbaum, 1991b, c; Gupta, et al., 1994). Essentially all of the perturbation experiments described to date were performed over a relatively short time period (from a few days to a few weeks). Little is known about the effects of long-term, continuous periodic substrate perturbations on anaerobic community structure and function. In order to determine whether long-term adaptive changes might occur in such systems, a series of long-term periodic substrate perturbations with different cyclic frequencies were applied to anaerobic chemostats. In Chapter 4, the response of an anaerobic chemostat subjected to an one day feast-one day famine cyclic feeding pattern was examined. This chapter extends the conditions evaluated to a long-term (>400 days) three day feast-three day famine feeding pattem. The response of the anaerobic chemostat population to the 3/3 day perturbation, especially the thermodynamic changes during the perturbation period, were investigated. 93 3116 1‘0” We 1‘11 94 5.2 MATERIALS AND METHODS 5.2.1 Anaerobic reactor A schematic representation of the experimental apparatus used for the perturbation experiment is shown in Figure 5-1. The experimental system consisted of a 1,500 ml working volume all glass reactor, timer controlled pumps for substrate and mineral media feed, and provisions for effluent liquid and gas effluent collection. Two Watson-Marlow 101 UIR pumps (W atson-Marlow, Inc., Wilmington, MA) were used to alternately supply glucose plus mineral media and mineral media alone (no glucose) at a flow rate ca. 150 ml/d. Power to the Watson-Marlow 101 UIR pumps was channeled through a electronic ChronTrol XT timer (ChronTrol Co., San Diego, CA). This programmable timer was used to control on/off operation of the two substrate feed pumps to obtain the desired substrate feed schedule. The contents of the anaerobic reactor were kept completely mixed with a magnetic stir bar driven by a Corning magnetic stir plate (Model PC-310, Corning Inc., Corning, New York). The entire apparatus was located in a constant temperature chamber (Nor-Lake Scientific, Hudson, WI) at 35 :t 0.5 0C. 5.2.2 Substrate and nutrients Glucose was supplied as the sole carbon and energy source. Analytical reagent grade granular glucose (Mallinckrodt Specific Chemical Co., Paris, KB) was used to prepare the 16 g/L glucose solutions. A buffer solution, prepared with reagent grade powder sodium bicarbonate (J, T. Baker Inc., Philipsburg, NJ), was mixed with the glucose solutions to obtain 7 g/L NaHCO3 in the feed solution. The mineral nutrient solution was selected based on the compounds and their concentrations used by several researchers worked on anaerobic digestion (Cohn et a1. 1982; Zoetemeyer et al. 1982a; Zehnder et a1. 1987). The concentration of the various components in the nutrient solution was described in detail previously (Chapter 4). Both the feed substrate solution and the feed nutrient solution were prepared with deionized water. 95 58E .2252 -.. ...... ..... ...... d: “8 358598 65 we ouannow ._ -n 05mm .3826 P 955 ton econ. @5383 BE: .obcou aEaa Emtuaz @5388 «mo Swarm 33E mac 10 96 5.2.3 Analytical methods Reactor performance was monitored daily for effluent pH, total gas production and composition, and gas-phase hydrogen partial pressure. COD, VFA, ethanol and VSS analysis were performed on a daily to weekly basis, depending on reactor performance. Total gas production was measured with a digital gas meter based on liquid displacement. Procedures and equipment used for performing the above analyses are provided in Appendix A. 5.2.4 Free energy calculations Free energy changes were calculated using a modified equation (Appendix F) which allows the effect of reactor pH to be counted for compared to standard physiological conditions (T = 35 0C and pH = 7). AG' = AGO. + m RT ln ([H+]/10‘7) + RT 2 Vi ln Ai (5-1) where: AG' = free energy change under operation conditions, KJ/mol AGO' = the increment of free energy under standard physiological conditions (25 0C, 1 atrn and pH 7) m = net number of protons in the reaction (m is negative when more protons are consumed than formed) R = gas constant, 8.3143 J/K-mol T = temperature, K [H'l’] = proton concentration in the reactor solution, mol Vi = the stoichiometric coefficient for component Ai in a biological reaction ( i = l,2,3,... ) and is negative for reactants and positive for products A1 = the physiological concentration of the component i in the reaction, mol. Under experimental conditions, the reactor temperature was maintained at 35 0C (308 K). therefore equation (5-1) can be rewritten as: 40' = AGO' + 5.9 [log ([H+]/10'7)m + 2 Vi 108 Ai] (5'2) 97 TABLE 5-1. Free energy associated with the anaerobic oxidation of glucose fermentation intermediates. Equation AGO' (KJlmol) (1) CH3CH2COO‘ + 2 H20 = CH3COO' + 3 H2 + C02 +761 (2) CH3CH2CH2COO' + 2 H20 = 2 CH3COO‘ + 2 H2 + H+ +48.1 (3) CH3CH20H + H20 = CH3COO' + H+ + 2 H2 +9.6 (4) CH3COO' + H+ 2 CH4 1 C02 - 35.8 (5) 4H2 + C02 = CH4 + 2H20 - 135.8 Equations used to calculate standard free energies for anaerobic oxidation of glucose metabolic intermediates at pH 7 are presented in Table 5-1. Free energy values were calculated according to Thauer et a1. (1977). 5.2.5 Reactor inoculation The inoculum (1,500 ml) for the daughter reactor was obtained from a 15 L laboratory-scale "mother" reactor ope:rated under steady state conditions with a 10 day HRT and a constant 8 g/L glucose feed. The mother reactor was inoculated with digester sludge taken from Jackson Wastewater Treatment Plant (Jackson, Michigan). The inoculated "daughter" reactor was initially operated under the same conditions as the mother reactor (10 day HRT with continuous input of 8 g/L glucose) for 50 days. Once 98 steady-state was reached and sufficient baseline information obtained, a square wave substrate perturbation was applied to the daughter reactor. 5.2.6 Experimental design In this study, a long term periodic substrate perturbation was applied to an anaerobic "daughter" reactor. The response of the system to the perturbation was then investigated. During the perturbation, the reactor influent glucose concentration was alternately changed in amplitude from 16 g/L to 0 g/L (mineral media only) on a 6 day square wave cycle (Figure 5—2), i.e. the reactor was fed 16 g/L glucose solution for three days followed by glucose-free mineral media at the same flow rate for the next three days. A 10 day hydraulic retention time (dilution rate of 0.1 d'l) was maintained throughout the experiment. The perturbation loading pattern was chosen so that the average influent glucose concentration during the perturbation period was equal to the steady-state concentration of 8 g/L applied to the control "mother" reactor. The daughter reactor was continuously operated for a period of 450 days including ca. 50 days under steady-state baseline conditions and ca. 400 days under perturbation conditions. Glucose concentration (glL) 99 20. Perturbation conc. '"'"' Steady-state Cone. 12‘ 8 q---- I---I|-------- pn-ununnuunnn nun-dr---.--- 4.. O 1 1 i v 1 v 0 3 6 9121518 2124 27 Time (day) Figure 5-2. Substrate feeding pattern used for perturbation experiment. 100 5.3 RESULTS 5.3.1 Effects of the periodic substrate perturbation. The response of the daughter reactor community to the periodic substrate perturbation can be broken down into four distinct stages (Figure 5-3): 1. rapid accumulation of metabolic intermediates of glucose fermentation (acetate, propionate, hydrogen and ethanol) and COD accumulation; 2. establishment of a metastable "steady-state" at reduced COD removal efficiency (Figure 5-4); 3. rapid VFA degradation; and (4) re- establishment of steady-state with high COD removal efficiency. Stage 1 can be further divided into two distinct sub-stages: stage 1a. rapid acetate and propionate accumulation, and stage 1b. rapid ethanol and hydrogen accumulation while acetate and propionate concentrations stabilized. Stage 2 also can be divided into two sub-stages: stage 2a. a cyclic pattern of acetate and butyrate concentration changes and stage 2b. all the major glucose degradation intermediates become more stabilized. Each of the above stages is discussed in greater detail in the following sections. Stage 1a (day 0 to day 18) As shown in Figures 5-3 and 5-4, the initial microbial community was highly sensitive to variation in organic feed concentration. A rapid accumulation of glucose metabolic intermediates including acetate, propionate, as well as ethanol and hydrogen, occurred nearly immediately after initiation of the substrate perturbation. Effluent COD concentration quickly increased from 370 m g/L to ca. 6,000 mg/L during the first 18 days (three complete perturbation cycles), accumulating at an average rate of 310 mg/L-d (Figure 5-4). The acetate-utilizing methanogens were immediately impacted by the substrate perturbation as evidenced by a sharp increase in the effluent acetate concentrations. Effluent acetate concentration increased from 150 mg/L to ca. 3,000 m g/L during the first 18 days, accumulating at an average rate of 158 m g/L-d. Acetate contributed ca. 50% of 101 2935 lol 920390 '9' o§mo< Ilul 60th cornea—Eon 05 wars 5 823.5588 E mowcasu .m-m 8:3". 53 sec. 8% com ..-.A,..r..- .11 a J «’43.???- . ’oneeesoemqoewlck'.’ . — e-Ic . _ a. . - a. . _ 6 _ _ . - . _ u _ — _ _ n. _ . . _ _ e _ _ . _ - _ _ _ u . . .... . . . _ . _ . e _ . u . _ . u . . _ . - - . - _ - . _ - . _ - e . m . «L F 89. (1,801) uoyenuaouoo vg A 102 ¢ n o 0 ”(can .u' 0 ”D lP------——-h Cl (115"!) (10:) ---------------------------- a DO o o—.0 I n — E.- 0 O (40 0‘0 D'\_ —.. 5r) in u 4‘? tool be. ago out “DC (0 O O ---P -.u‘_‘_-_’ - ----------------I I 400 300 200 100 Time (day) Figure 5-4. Changes in COD concentrations during the perturbation period. 103 the total increase in COD observed during this period, with acetate concentration fluctuating around 3,000 mg/L. A rapid accumulation of propionate was also observed during the initial 18 days of the substrate perturbation. Effluent propionate concentration increased from less than 30 mg/L to a level of ca. 1,000 mg/L, at an average daily accumulation rate of 53 mg/L-d. Propionate contributed about 26% of the total COD accumulated during this period. Ethanol and gas phase hydrogen concentrations increased appreciably during the glucose feed period. Although the accumulated ethanol and hydrogen degraded during the following fasting periods (mineral media only), peak concentrations of accumulated ethanol and hydrogen during the feed periods gradually increased. Butyrate did not accumulate during this period. Effluent butyrate concentration remained at ca. 30 mg/L. Stage 1b (day 19 to day 48) During this period, the acetate and propionate concentrations stabilized while an accelerated accumulation of hydrogen and ethanol was observed. (Figure 5-5a). The accumulation of hydrogen suggests a significant decrease in hydrogen-utilizing methanogenic activity. As shown in Figure 5-5a, hydrogen accumulated immediately after initiation of the substrate perturbation. A maximum hydrogen partial pressure of ca. 6.6 x 10‘2 atrn (6.6% headspace gas composition) was observed at day 45. Subsequently, hydrogen concentrations decreased rapidly. Ethanol accumulation was coincident with the periodic hydrogen accumulation (Figure 5-5b). The peak values of accumulated ethanol quickly increased to ca. 800 mg/l within four complete perturbation cycles (24 days). Thus, ethanol contributed as much as 23% of the accumulated COD during this period. In contrast to the accumulation pattern of the VFAs, the ethanol that accumulated during the three day glucose feed period was quickly metabolized during the following three day fasting period. Periodic ethanol accumulation n~~hn 111}: iv" V l vav (133133 50‘10’ [sear wit-1.1. AIIIIIIIU fifi '31" .11 1 v A~\fiialhv havlhluahIVH 104 H2 (ppm) Time (day) 1000 800‘ Ethanol (mg/l) Time (day) Figure 5-5. Changes in hydrogen and ethanol concentrations during the first stage of the perturbation period. 105 continued, until a complete recovery of hydrogen-utilizing methanogenic activity was observed. No significant ethanol accumulation was observed thereafter. During the first stage of the substrate perturbation, pH decreased from a steady-state value of 7.0 to ca. 6.2 with large fluctuations (Figure 5-6). Daily gas production rate (average value for the 6—day cycle) decreased from about 430 ml/L-d to a minimum level of ca. 135 ml/L—d at the end of this period (Figure 5-7). Most of the gas was produced during the glucose feeding period. The average gas production during the famine period decreased from ca. 80 mllL-d to ca. 10 ml/L-d within about 50 days. Effluent butyrate and iso-butyrate concentrations remained low (less than 50 mg/L and 90 m g/L, respectively) during the first stage of the substrate perturbation. No VFAs, other than acetate and propionate, accumulated during this period. No glucose (< 0.5 mg/l) was detected in the reactor liquid phase during the glucose feed period. Stage 2. (day 49 to 289) When the average effluent COD concentration reached ca. 5,000 mg/L, a metastable "steady-state" was established. During this period, the peak hydrogen concentrations observed during the glucose feed period soon returned to near their prior perturbation levels of less than 50 ppm, and remained stable at this level thereafter. COD removal efficiency averaged about 41%, compared to ca. 95% during the baseline period. Average daily gas production decreased to 54 % of the baseline value, and CH4 content in the reactor headspace was 26% compared to 50% prior to the perturbation. Although effluent COD remained relatively stable during the metastable steady-state, significant changes in the composition of the VFAs were observed. 211. (day 49 to 189) For the period of from day 49 to 189, a dramatic accumulation of butyrate was initially observed. Between days 49 and 52, butyrate concentration quickly increased pH 106 7.4 Time (day) Figure 5-6. Changes in pH during the perturbation period. 1 ' r 1 : 2 : 3 : 4 | I 7.2- i g . : r - ' i l". 7.0 8 : : (03:3 5;.” A . : 1 or 68 I ' ' ' .8 r : “i : 1 : I a° I - r ' ° ' 6.6 D " | : . : . 0.. 1 | a , . ”0.4.1 ° ' i 6'4 ”ll. ‘0' o n l r q . or. 10.": '.° to on o o u o o I : . r l 1 it my .1. : ° ‘1 Li'loo‘ » i 11.11 6.0- o r n on 000000 a a : : a : 0 n U ' ' t I i 5.8 " I ' 1 ' ' 0 100 200 300 AIUI\I\-I-V Dal.“ V 107 3 300 200 100 'l"|l' Gas—5 95 Time (day) Figure 5-7. Gas production during the perturbation period. 108 from less than 50 mg/L to ca. 1,100 mg/L, and in the following 20 days, butyrate concentration further increased to a peak level of ca. 2,200 mg/L, accounting for as much as 65% of the accumulated COD during this period. Acetate concentration varied concurrently with butyrate concentration in a cyclic pattern, when butyrate levels increased acetate levels decreased and vice-versa. In each cycle, the acetate and butyrate concentrations varied concurrently with amplitudes from ca. 1,700 mg/L to ca. 3,400 mg/L and from ca. 2,000 mg/L to ca. 500 mg/L, respectively. The cyclic pattern of acetate and butyrate concentration changes had a frequency of approximating 48 days for three complete cycles. Propionate concentration decreased steadily from ca. 1,200 m g/L to a minimum level of ca. 300 mg/l. It then increased slightly to an average value of ca. 400 mg/L, accounting for about 12% of the accumulated COD during this period. 2b. (day 190 to day 289) During this period, the 50 day butyrate-acetate concentrations cyclic variation pattern was broken. Butyrate concentration steadily decreased within several mini cycles in two apparent steps. First the butyrate concentration steadily decreased from ca. 1,200 mg/L to a minimum level of ca 150 mg/L. Acetate concentration concurrently increased to levels of over 3,200 mg/L. This peak acetate concentration apparently temporarily blocked butyrate degradation, as a result butyrate concentration quickly increased until a peak level of ca. 1,100 mg/L was reached. In the following ca. 60 days, the butyrate concentration again continuously decreased and stabilized at a level of ca. 300 mg/L. The acetate concentration fluctuated around ca. 2,800 mg/L. Stage 3 (day 290 to day 340) A rapid decrease in VFA concentrations began on day 290. As shown in Figure 3a, during this period a sharp decrease in acetate concentration was observed. Within about Mm 150 n Pm COD fun um C9310 new aver heft; M W 32. 109 30 days, the acetate concentration decreased from ca 3,200 mg/L to less than 200 mg/L. Meanwhile, butyrate and iso-butyrate concentrations quickly decreased from ca. 450 and 250 mgll to their pre-perturbation concentrations of less than 30 mg/L each, respectively. Propionate concentration decreased steadily from 600 mg/L to ca. 300 mg/L, and effluent COD decreased from ca. 5,500 mg/L to less than 800 rug/L. The pH quickly increased from 6.2 to about 7.1, and the gas production rate also reflected a drastic change in VFA utilization. Average gas production rate peaked at 560 mllL—d on the same day what acetate concentration began to fall (Figure 5-7). Stage 4 (day 340 to day 400) With the exception of propionate, all VFAs stabilized at new steady-state levels within ca. 30 days subsequent to the start of rapid acetate degradation. After another 20 days, a new steady-state was re-established in the anaerobic reactor. During this period the average COD concentration, pH and biogas production rate were similar to those observed before the substrate perturbation. A comparison of the reactor performance during the pre- perturbation steady-state condition, the metastable steady-state condition, and the new steady-state condition established during the substrate perturbation, is presented in Table 5-2. 5.3.2 Changes in gas production rate during substrate perturbation period Gas production directly represented the effect of the periodic substrate perturbation on the anaerobic microbial community. During the first stage of the substrate perturbation, a period marked by rapid VFA accumulation, the average daily gas production (average value for the 6-day cycle) concurrently decreased from about 430 ml/L-d to a minimum level of ca. 130 ml/L-d. The average daily gas production increased slightly and fluctuated around a level of 230 ml/L-d (Figure 5-7). As observed in previous one day feast-one day famine perturbation, most of the gas was produced on days when glucose was provided Table 5-2. Summary of reactor operation results during the different stages of the perturbation experimentl Baseline Metastable New steady-state steady-state steady-state CODin (mg/l ) 8500 8500 8500 CODe (mg/l) 370 :1: 537- 4970 :1: 630 280 :1: 103 VSS (g/l) 1060 :1: 160 720 i160 1000 :t 180 Gas (ml/L-d) 430 :1: 20 230 :1: 30 440 :1: 260 CH4 (%) 50 :l: 2.0 26 :1: 6.0 49 :1: 2.0 pH 7.02 :1: 0.04 6.14 i 0.17 6.97 :1: 0.04 Acetate (mg/l) 150 :1: 30 2320 :1: 570 120 :1: 100 Propionate (mg/l) . 26 i 15 403 :1: 110 18 i 9.0 Butyrate (mg/l) 24 :1: 8.0 1030 :1: 550 8 i 4.0 Iso-butyrate (mg/1) 17 :1: 10 90 i 60 2 :1: 0.8 1: All the data shown in the columns of transient steady-state and new steady-state are averaged values for six day cycle during substrate perturbation period. 2: Standard deviation. (440 to 460 ml/L-d). Average gas production for three famine days in each cycle decreased steadily from ca. 100 ml/L-d to a minimum level of ca. 10 ml/L-d within ca 50 days, indicating a significant decrease in the number or activity of acetate-utilizing methanogens. Subsequently, gas production on famine days increased slightly and fluctuated around a level of ca. 25 ml/L-d for about 190 days. After day 240, gas production on famine days increased (Figure 5-8), and by day 270, it had increased to near pre-perturbation levels. A sudden increase in the gas production rate was then observed on famine days. On day 315, a peak gas production of ca. 448 mllL-d was observed on a famine day. This peak coincided with a sharp drop in acetate concentration of the reactor effluent. Gas production on famine days stabilized at a level of ca. 100 mllL- d thereafter. lll 8v cocoa gangsta 05 war—2v :ouusuoa mum baa .w-n 0.5me as: 2:2. own com omm com on fi 09 on (P/Ilu) 59:) 112 5.3.3 Changes in un-ionized VFA concentrations during the substrate perturbation period No special pH control means other than routine bicarbonate addition was employed. Under these conditions, accumulated VFA concentration (as acetic acid) correlated with reactor pH throughout the experiment. Accumulation of acetic or other volatile fatty acids will neutralize mineral alkalinity in the reactor, causing the pH to decrease. A pH-dependent equilibrium exists between the ionized and un-ionized volatile acids is illustrated for acetate in Equation 5-3: CH3COOH = CH3COO' + H+ (5-3) As the pH decreases, equilibrium shifts to the left and the un-dissociated acetic acid concentration increases. At 35 0C the ionization constant (Kca) has a value of 1.73 x 10‘5 (Kroeker, 1979). The un-ionized acetic acid concentration can be calculated by using the following equation: [CH3COOH] = [CH3COO‘] [H+]/Kca (5-4) Because of the passive diffusion of neutral compounds across the bacterial membrane, acetic acid and other VFAs in their un-ionized form would be expected to cross penetrate the bacteria cell membrane without resistance. Un-ionized VFA level is a function of VFA concentration and environmental pH, and is directly linked to inhibition of methanogenic activity (Attal et a1, 1988; Duarte and Anderson, 1982; Anderson et al., 1982; Kroeker et al., 1979). As illustrated in Figure 5-9, the un-ionized VFA concentration (as acetic acid) is usually less than 10 mg/L in well operated anaerobic reactors. An un-ionized VFA concentration above 30 mg/L tends to result in reactor failure. Therefore, an increase in VFA concentration or a decrease in pH may increase un-ionized VFA concentration to levels at which acetate-metabolizin g methanogenic activity is significantly inhibited or even stopped. pl! 113 Transition zone —'—) Digester failure 6.4 ' I I a I I 6.2“ i l .. I I 6.0- 1 1 I 5.8 'l i ‘ I V I I l 1 0 20 4o 60 80 100 Non-ionized VFA (mg/L as acetic acid) Figure 5-9. Relationship of un-ionized volatile fatty acids, pH and anaerobic reactor performance (after Attal et a1, 1988; Duarte and Anderson, 1982; Kroeker et al., 1979). l 14 The changes in un-ionized VFA concentrations (as acetic acid) in the present study and in the one day feast-one day famine perturbation (1/1 day perturbation) are presented in Figure 5-10, and summarized in Table 5-3. Experimental data show that although the average VFA concentrations were essentially the same in the metastable steady-state in both perturbation tests, due to a combined effect of lower pH value and larger fluctuations in VFA concentrations, much higher un-ionized acetate and un-ionized VFA concentrations were observed during the 3/3 day perturbation test. Table 5-3. Summary of un-ionized acetate and un-ionized VFA concentrations (as acetic acid) observed during the 3/3 day and 1]] day substrate perturbationsl Steady-state Metastable New steady-state steady-state M pH 7.2 6.14 6.97 un-ionized Ac (mg/L) 0.3 97.10 0.74 un—ionized VFA (mg/L) 1.04 142.90 0.86 W pH 7.2 6.34 7.02 un-ionized Ac (mg/L) 1.05 65.64 1.44 un-ionized VFA (mg/L) 1.43 93.36 1.61 1: Data used for un-ionized VFA calculation were taken from Tables 4-3 and 5-2 respectively. Adv-Uni. Unnatva I: l~\n.-IIIV IUQNMIHAVfiIIh~H 115 300 ., .1211?!.111.-,-1_ " _._ —-—- 1/1 Un-ionized VFA (mg/L as acitic acid) 0 100 200 300 Time (day) Figure 5-10. Changes in non-ionized VFA concentrations during the 1/1 day and 3/3 day substrate perturbations. 116 5.3.4 Free energy calculations for anaerobic oxidation of glucose metabolic intermediates. Free energy available for hydrogen and carbon dioxide conversion to methane during the pre-perturbation period varied between -3.3 to -6.2 KJ/rnol of H2 (Figure 5-1 la). This value increased to between -25 and -27 KJ/mol of H2 during the first stage of the perturbation period indicating that the removal of hydrogen was not thermodynamically limited. These values decreased to levels approaching pre-perturbation values during the beginning of the second stage. During the first two stages of the long-term substrate perturbation, acetate concentration rapidly increased and then stabilized at levels of around 3,000 mg/L. The free energy available for acetate catabolism concurrently increased from ca. -24 KJ/mol to -37 KJ/mol during this period (Figure 5-1 lb), indicating that acetate accumulation was the result of a kinetic and not thermodynamic limitation. Production of acetate exceeded the ability of the acetate utilizing methanogens to metabolize acetate by a considerable amount. This does not appear to be the situation for the accumulation of the higher molecular weight VFAs. The free energy available for oxidation of propionate and butyrate dramatically changed during the first two stages. Results of free energy calculation show that during the first stage of the substrate perturbation, free energy available for propionate and butyrate catabolism concurrently decreased from ca. -5.1 KJ/mol and -13.1 KJ/mol to average levels of 40 KJ/mol and 34 KJ/mol, respectively (Figure 5-12a and b). This was caused by significant hydrogen accumulation. As a result, oxidation of propionate and butyrate was thermodynamically impossible. During the second stage, when hydrogen concentration returned to pre-perturbation levels (< 50 ppm), the free energy for oxidation of propionate and butyrate again became favorable. Free energy available for propionate and butyrate oxidation fluctuated around -7.5 KJ/mol and -5.0 KJ/mol, respectively. When steady-state conditions were re-established, the free energy available for propionate and butyrate oxidation increased to ca. -14.2 KJ/mol and - 16.2 KJ/mol, respectively. 117 5 1 a. H2 + C02 0 .. 1. -5 ‘ 111:1: °°° M°“1Toii';roi':1°unfi_;°mari 0:31)] a on: 1.. 0 '11‘11“ ‘ . Gibb's free energy changes (kJ/mol) o 4H2 + C02 = 2H20 + CH4 -30 . . . , . , . 0 100 200 300 400 Time (day) -15 b. Acetate CH3COO- + H+ = CH4 + 002 Gibb's free energy change (lemol) 400 Time (day) Figure 5-11. Free energy available for hydrogen and carbon dioxide, and acetate conversion to methane during the substrate perturbation period. II...‘ nuns-:\~rvh.v Undue-weirV Hui—flu...- ri rvuvrflh ..r.. AfiAi‘HV 118 . a. Butyrate CH3CH2CH2COO- + 2H20 = 2CH3COO- + 2112 + H+ Gibb's free energy change (kJ/mol) Time (day) b. Propionate Gibb's free energy Change (kJ/mol) ' 7 . .. 9 '. 5"" .. - . 1' '1 '1. '.'~'--- . I“ if?!" ' "0‘. I . L a u 1.. CH3CH2COO- + 2H20= CHBCOO- + 3H2 + C02 ‘40 ' l ' I v I u 0 100 200 300 400 Time (day) Figure 5-12. Free energy available for the anaerobic oxidation of butyrate and propionate during the substrate perturbation period. 119 Free energy calculations for ethanol oxidation show that during the first stage of the substrate perturbation the free energy available for ethanol oxidation changed dramatically within each perturbation cycle (Figure 5-13). A cyclic pattern of free energy changes, driven by the pattern of hydrogen accumulation (Figure 5-5), was observed. Although the free energy available for ethanol oxidation was negative most of the time, the average free energy in each perturbation cycle decreased quickly from ca. -28 KJ/mol to ca. -5 KJ/mol in the end of the first stage. When the hydrogen concentration returned to normal (< 50 ppm) in the beginning of the second stage, free energy available for ethanol oxidation increased to near pre-perturbation values and stabilized thereafter. Free energy changes calculated for oxidation of glucose metabolic intermediates under steady—state pre-perturbation and perturbation conditions are summarized in Table 5—4. Table 54 Free energy values (KJlmol) observed for the anaerobic oxidation of glucose metabolic intermediates during substrate perturbation. Baseline Transient New Substrate Steady-state Steady-statel Steady-state Acetate -24.3 :1: 3.12 -37.4 :1: 0.7 -23.6 d: 2.8 Propionate - 5.1 :1: 1.7 -7.5 :1: 5.6 -l4.2 :1: 4.9 Butyrate -l3.l :1: 3.7 -5.0 :1: 2.7 -l6.2 :1: 6.2 Ethanol -29.83 :1: 3.7 -17.93 :1: 6.4 -35.63 :1: 4.0 Hydrogen -6.3 :1: 4.0 -2l.3 i 4.5 -5.6 :1: 4.1 1. Calculated based on data from day 80 to day 290. 2. Standard deviation 3. Estimated based on ethanol detection limit of 0.5 m g/l. Anew-I-\-lzv rind-ulna.“ hurled-I9 grill... arr.A-A-UIV 5.; The I" 120 10 MM V\ \1 Gibb's free energy change (KJlmol) r CH3CH20H + H20 = CHBCOO- + 11+ + 2H2 ~50 ' I V I ' I T I ' I ' I e I ' I v 0 6 12 18 24 30 36 42 48 54 Time (day) Figure 5-13. Free energy available for the anaerobic oxidation of ethanol during the ' substrate perturbation period. 121 5.3.5 Electron equivalents mass balance An electron balance that accounts for influent and effluent organics, cells and methane production can be used to demonstrate the relative importance of each reaction product. Electron equivalents that are derived from the complete oxidation of key substrates to carbon dioxide were determined from the following half reactions: C5Hle5 + 6HzO = 24e' + 24H+ + 6C02 (5-5) CH3COOH + 21120 = 8e' + 8H+ + 2C02 (5-6) CH3CH2COOH + 41120 = He” + 1411+ + 3C02 (5-7) CH3CH2CH2COOH + 6H20 = Me + 20H+ + 4C02 (5-8) CH3CH20H + 3HzO = 12e' + 12H+ + 2C02 (5-9) CH4 + 2H20 = 8e' + 8H+ + C02 (5-10) H2 = 2e' + 2H“' (5-11) Clip-[00.41802 + 1.6H20 = 4.3e' + 4.3H+ + C02 + 0.2NH3 (5-12) Based on the above half reactions, the electron equivalents of the glucose substrate, fermentation intermediates, methane and biomass are summarized in Table 5-5. Here, one mole of cells is assumed to be equal to 22.9 grams of VSS, based on the empirical cell formula of CHL7OQ4N02 obtained in Chapter 3. An electron equivalents balance was performed using data for the entire periodic substrate perturbation period (Figure 5-14). The distribution of the electron equivalents in reactor effluent during the different stages of the substrate perturbation is summarized in Table 5- 6. These data demonstrate that under steady-state condition, most of the electrons generated from glucose oxidation quickly flowed through the fermentation intermediates, such as VFAs and hydrogen, ending up in methane (75%) and biomass (19%). In contrast, under unstable perturbation conditions (stage one to stage three), a significant amount (ca. 65 to 70%) of electrons generated from glucose oxidation ended up in fermentation intermediates, mainly as acetate, propionate, butyrate and ethanol. Electron 122 Table 55. Electron equivalents (eq) of the anaerobic reactor substrate, fermentation intermediates and products Organics MW mol eq/mol mass g mass/mol eq Glucose 180 24 7.50 Acetic acid 60 8 7.50 Propionic acid 74 14 5.29 Butyric acid 88 20 4.40 Ethanol 46 12 3.83 Hydrogen 2 2 1.00 Methane 16 8 2.00 Biomass 22.9 4.3 5.33 Table 5-6 Electron equivalents distribution (mmol) in anaerobic reactor effluent during the 3/3 substrate perturbation (Total input electron equivalents = 104) Organics Steady-state Stage 1 Stage 2 New steady-state Methane 77.9 $1.8 22.6 $9.3 17.9 $4.9 79.8 $1.3 Biomass 19.4 $0.7 9.8 $1.4 13.7 $1.8 18.4 $0.3 Acetate 1.3 $0.4 32.0 $0.9 33.0 $6.7 1.3 $0.4 Propionate 0.2 $0.2 19.3 $2.4 8.4 $3.1 0.1 $0.03 Butyrate 0.7 $0.4 2.1 $1.3 25.5 $11.0 0.1 $0.04 Ethanol ND1 11.4 $4.0 ND ND Total 99.5 97.2 98.5 99.7 1: ND, not detectable equivalents in hydrogen were negligible throughout the perturbation period. The electrons in the measured products accounted for 93.6% of the input electrons in the feed glucose. 123 2 _ 50 E E. g; 100 " E > Propionate O ‘ . I r I ‘ l M 1/1 3/3 5/5 Reactor Figure 7-6. Effect of perturbation interval on maximum glucose and propionate utilization rates. 176 1500 a ' . ,5 1200'“ g d 900‘ .//.\0 as h -1 5 600- II . a 300‘ > ' Acetate O I ' I I I ' I M 1/1 3/3 5/5 Reactor 600 ’7'? 3 > 400‘ an E: E V 200‘ a E > Butyrate O I r I 1 I ' r M 1/1 3/3 5/5 Reactor A 200 "‘9 1 8’, 1607 > . an .. 3 120 E I E, 80 ‘ g 40‘ E 0 ' I I . I Hydro'gen M 1/1 3/3 5/5 Reactor Figure 7-7. Effect of perturbation interval on maximum acetate, butyrate and hydrogen utilization rates. 177 7.3.2 Response of anaerobic populations to glucose and temperature shocks Both glucose and temperature shocks were performed on the 1/1 daughter reactor and the constant feed chemostat (m-l) which was used as the representative of the control mother reactor. For both experiments, the shock was applied at the beginning of the change in the feed for the 1/1 reactor. Before initiation of the glucose and temperature shocks, samples were intensively collected from the daughter reactor 1/1 and the constantly fed anaerobic reactor m-l for three weeks to obtained baseline operational performance (Table 7-4). 7-4 Baseline operation results for reactor 1/1 and m-l. parameters l/ll m-l CODe (mg/L) 248 $782 427 $56 pH 6.90 10.07 6.97 $0.02 vss (glL) 1.15 $0.18 0.94 10.11 H2 (ppm) 19.7 $9.4 29.4 14.9 CH4 (%) 50 3:20 50 11.0 Gas (ml/L-d) 430 1:31 417 1:20 1: average value for two day cycle 2: Standard deviation. :1. Results of glucose shock loading test After pulse injection of the concentrated glucose stock solution, the effluent COD concentration in reactor m-l immediately increased to level of ca. 4,600 mg/L and then gradually decreased to near its pre-perturbation level over the next 4 days (Figure 7-8a). The glucose injected into reactor m-l was completely degraded within 11.5 hours at an 178 COD (mg/L) Time (day) Glucose (mg/L) Time (hour) Figure 7-8. Changes in a) COD and b) glucose concentrations during the glucose shock loadrng experiments. 179 average degradation rate of 360 mg glucose/L-h (Figures 7-8b). A rapid accumulation of glucose metabolic intermediates was observed immediately after the glucose was added. Acetate, ethanol and propionate concentrations reached their peak levels of ca. 1,100 mg/L, 500 mg/L and 300 mg/L, respectively, at the same time when glucose was observed to decrease to below detection limits (< 0.5 mg/L) (Figure 7-9a). Concurrently, a rapid accumulation in hydrogen up to 1,600 ppm was observed in the reactor headspace (Figure 7-10a). No significant butyrate accumulation was observed throughout the experimental period. The hydrogen and ethanol that accumulated were metabolized within ca. 0.5 and 1.5 days, respectively. For acetate and propionate, ca. 4 days was needed before their concentrations retumed to near the pre-shock levels. During the shock loading period a remarkable change in biomass concentration was also observed (Figure 7-10b). Reactor pH decreased concurrently from 6.98 to 6.62 with the significant accumulation of the glucose fermentation intermediates and returned to near its normal levels within ca. 4 days (Figure 7-11). The anaerobic population in the 1/1 daughter reactor appeared to be more resistant to the glucose shock. The pulse injection of glucose resulted a peak COD concentration of ca. 4500 mg/L. The effluent COD concentration returned to near its pre-shock levels within about 2 days (Figure 7-8a). The injected glucose was completely metabolized within ca. 5.5 hours at an average disappearance rate of ca. 760 mg glucose/L-h (Figure 7- 8b). This rapid degradation of glucose caused a significant accumulation of acetate, butyrate and propionate concentrations up to 760 mg/L, 150 mg/L and 70 mg/L, respectively (Figure 7-9b). After the pulse injection of glucose, ethanol accumulated up to 470 mg/L within ca 5 hours and then completely degraded during the following 24 hours. An accumulation of hydrogen with a peak level of ca. 350 ppm, was observed in the l/1 reactor headspace (Figure 7-10a). The changes in biomass and pH observed during the glucose shock loading test are presented in Figure 7-10b and 7-11 respectively. The perturbed daughter reactor retumed to its normal condition within ca. 1.5 days. 180 a 1200 ‘ m - 1 1000 " g ‘ —°— Acetate E 800 " —I— Propionate I ‘ —*— Butyrate {3 600 - Ethanol '8 E 5'5 3. 10 Time (day) b 1200 . 1 I 1 1000 ' —"-'—" Ethanol 800 '- 1 —-0— Acetate Propionate Intermediates (mg/L) Time (day) Figure 7-9. Changes inconcentrations of glucose fermentation intermediates for the a) ml and b) 1/1 reactors dunng the glucose chock loading experiments. 181 1800 1500 " Time (day) 2.0 b 1.6 A " 45 E9 1.2 m "9 0.8- 0.4- —0— 1/1 —0— M-1 0.0 ' I ' U ' ' O 2 4 6 8 Time (day) Figure 7-10. Changes in a) hydrogen and b) biomass concentrations during glucose shock loading experiments. 182 pH 6.5 - I ' I ' I ' ' Time (day) Figure 7-11. Changes in pH observed during the glucose shock loading experiments. 183 Table ‘7-5. Peak levels of glucose fermentation intermediates and other Operational parameters in reactor 1]] and m-l during glucose shock loading tests Parameters 1/1 m-l Hydrogen (ppm) 350 1600 Acetate (mg/L) 770 1100 Ethanol (mg/L) 560 510 Propionate (mg/L) 70 310 Butyrate (mg/L) 140 30 pH 6.58 6.60 VSS (glL) 1.94 1.87 Recovery time (day) 1.5 4.0 The peak levels of glucose fermentation intermediates and other operational parameters during the glucose shock loading test are summarized in Table 7-5. h. Results of temperature shock The temperature shock was performed by temporally shutting off the reactor heating system for 24 hours. Reactor temperatures decreased from 35 0C to ca. 25 0C within ca. 3 hours and remained at this level until heating resumed (Figure 7-12a). The response of the m-l and the periodically fed 1/1 reactors to the 24 hour temperature shock is presented in Figure 7-12b. In the reactor m-l, a slight increase in effluent COD during the 24 hour temperature shock period was followed by a continuous increase in effluent COD over the following 15 days. A peak COD level of ca. 1100 mg/L was observed. Subsequently, COD decreased to its normal level within 5 more days. A slight decrease in reactor pH was also observed in reactor m-l. Although the effluent 184 4O * Stop heating 6‘ 35 8 0 E 2 3' 30" 5 Resume heating l- 3 a 25- t 0 a: 20 v I I I ' I ' I ' I r -2 O 2 4 6 8 10 Time (day) 1200 .. . —0— 1/1 1000' 800‘ g, 600- 8 Resume heating 0 400 200 O I I I I ' I ' I I ' I -5 O 5 10 15 20 25 Time (day) Figure 7-12. Changes in temperature (a) and COD concentrations (h) during the temperature shock experiments. 185 COD concentration in reactor m-l increased to twice its normal value, no significant increase in effluent VFA (acetate, propionate and butyrate) concentrations was observed. Analytical results also demonstrated that the COD increase was not the result of accumulation of common glucose metabolic intermediates ethanol and lactic acid. The elevated COD concentration observed during the temperature shock period were not identified. In contrast, the community in the 1/1 reactor was not significantly affected by the short-term temperature shock. No significant change in COD or other operational parameters was observed either during the temporary temperature decrease or thereafter (Figure 7_-12b ). No changes in headspace hydrogen partial pressure was observed for either system. l 86 7.4 Discussion Experimental results presented in previous chapters illustrate that the anaerobic populations could gradually recover from the initial impacts of the periodic substrate perturbations. Additional experimental results obtained from this study demonstrate that the re-established anaerobic populations could maintain stable and effective Operation under long-term periodic substrate feed conditions. The average COD removal efficiency and gas production, as well as other operational parameters, obtained in the re-established daughter reactors were similar to those obtained in the control mother reactor which operated under constant substrate feed conditions. The only major difference in daily operational performance between the mother and the periodically fed daughter reactors, was the cyclic changes in the operational parameters of the daughter reactors. This study established that periodic substrate feeding mode could be used as an operational strategy to enhance reactor stability. In addition to high organic removal efficiency, the re-established daughter reactor population appeared to be more resistant to sudden environmental changes. Short recovery time and less serious impacts were observed for the daughter reactor community. The characteristics possessed of the daughter reactor community are most likely due to changes in microbial community. As indicated by the maximum specific substrate utilization rates, more active or "robust" species became predominant in the perturbed daughter reactor populations presumably because of a long-term selection/adaptation processes (Table 7-3). To achieve effective organic degradation, the entire anaerobic community must work syntrophically. The enhanced substrate utilization capacity in the daughter reactor population does not only benefit individual trophic groups, but is also advantageous to the entire system. Remarkable differences in the distribution of glucose fermentation intermediates were observed in reactor 1/1 and m-l after glucose shocks (Table 7-5). Much less propionate was produced in the daughter reactor. This was most likely due to reduced hydrogen accumulation in the II] system which, as discussed in Chapter 4 and 5, could dramatically change glucose fermentation routes. As usually 187 observed in perturbed anaerobic reactors (Mosey and Fernandez, 1989), a large amount of propionate was produced in the m-l reactor, indicating a significant shift in the fermentation route. Due to poor propionate utilization capacity possessed by the constantly fed anaerobic community, propionate contributed a large portion of the accumulated COD during the following recovery period in the m-l reactor. Usually, operators of anaerobic reactor systems for wastewater treatment strive to achieve a constant feed. Under such conditions, the number of electrons channeled through propionic and butyric acids is greatly decreased (Palns, 1987; Mosey 1981). This causes a decrease in the relative population of the associated acetogens. Consequently, the resulting system is susceptible to acid accumulation from influent substrate variations (Duarte and Anderson, 1982; Fongastitkul et a1, 1994). Although this drawback has long been recognized (Harper and Pohland, 1985), no suitable resolution has been reached to date. This appears to be due to the paradox that to maintain high organic removal efficiency constant feed is desirable, but in order to change substrate distribution (i.e. let more substrate flow through butyrate-propionate fermentation) some kind of perturbation is necessary. Experimental results presented in this thesis suggest a reasonable resolution -- introducing properly controlled periodic substrate perturbations. This procedure could effectively change substrate distribution so as to maintain sufficient acetogens in the community, while at the same time sustaining a high efficiency of organic removal. Maximum substrate utilization rates in the perturbed daughter reactor cultures changed with perturbation intervals. This observation shows that the maximum substrate conversion rates may be used as a group criteria to compare or estimate stability of anaerobic communities. This observation also implies that there is the possibility of adjusting the perturbation interval to achieve an "optimum" microbial community structure. The above experimental results and discussion strongly suggest that although usually considered "harmful", substrate perturbations, if properly controlled, can be used as a tool to improve operational performance of anaerobic systems. 188 7.5 Summary The re-established daughter reactor populations could maintain stable and effective operation under long-term periodic substrate feed conditions. There reactors possessed higher substrate utilization capacities, and appeared to be more resistant to environmental shocks. These experimental results suggest the potential use of the substrate perturbation as a tool to optimize anaerobic microbial community and improve system operational performance. Chapter 8 CHARACTERIZATION OF MICROBIAL COMMUNITY STRUCTURE CHANGE IN RESPONSE TO LONG-TERM PERIODIC SUBSTRATE PERTURBATIONS BY FATTY ACID METHYL ESTERS 8.1 Introduction Methane fermentation results in the conversion of organic materials to methane and carbon dioxide in the absence of molecular oxygen. Conversion of carbohydrate, fat and protein to methane requires the combined activity of fermentative, acetogenic, and methanogenic bacteria (McCarty, 1981; Novaes, 1986; Pavlostathis, 1991). These different functional groups are composed of a large number of bacteria, of which relatively few have been isolated (Zeikus, 1979; Iannotti, 1982). Although much effort has been directed towards culturing the bacteria present in anaerobic digesters, only a small portion of the microorganisms present in anaerobic digesters have been cultured (T oerien and Hattingh, 1969; Varel, 1984; Britz et al., 1994). The above studies highlight the difficulties in culturing bacteria from anaerobic environments where many bacteria are syntrophically associated via complex physiological, kinetic and thermodynamic interactions. Recognizing the limitations of conventional isolation techniques, White (1979, 1983) advocated new methods based on the analysis of lipids to obtain information about the community structure and nutritional status of complex microbial ecosystems, without the need for selective removal or growth of isolates. Fatty acid methyl esters (FAME) analysis can be used to obtain a characterization of the microbial community (Bobble and White, 1980). In bacterial cells, fatty acids occur mainly in the cell membranes as the acyl constituents of phospholipids 189 190 (Kandeda, 1991). Some fatty acids are unique to specific bacteria or to groups of bacteria and can serve as signatures for these bacteria (Ikemoto S. etal., 1978; Boe and Gjerde, 1980). Gillan and Hogg (1984) have used this concept to divide the bacterial community of mangrove sediments into several subgroups. The FAME method is frequently used to assess the structure of soil and aquifer communities (Vestal and White, 1989; Rajendran N., et al., 1992; Zelles, et al., 1992). In addition, some researchers have pioneered use of FAME to estimate metabolic status and community structure of anaerobic digesters ( Henson et al., 1985; Mikel] et al., 1987; Schropp et al., 1988; Hedrick et al., 1991). However, little is known about the FAME characteristics for specific anaerobic systems, such as the glucose fed anaerobic chemostats evaluated in this study, or the feasibility of using FAME profiles to monitor changes in microbial community. Confirmed information is still lacking for the relationship of the FAME results with major operational parameters such as COD and VFA concentrations. This chapter presents the results of using FAME analysis techniques to monitor changes in microbial community structure for the daughter reactors described in Chapters 4 to 6, during initial steady-state perturbation conditions and after recovery. 191 8.2 MATERIALS AND METHODS 8.2.1 Anaerobic reactors A 15 L working volume anaerobic chemostat (the mother reactor), operated at 10 day HRT and 35 0C with glucose as the sole carbon and energy source, was used as a source of organisms for the inoculation of daughter reactors and as a stable control. Three 1.5 L working volume anaerobic chemostats (the daughter reactors) were inoculated with cultures taken from the mother reactor and were initially operated until a steady state was achieved under the same conditions as the control mother reactor (10 day HRT and 8 gIL glucose feed). After collection of sufficient steady state baseline information, the three daughter reactors were subjected to periodic loading patterns in which glucose feed concentration was varied from 16 g/L to 0 g/L (mineral media only) in 2 day, 6 day and 10 day cycles respectively. System set up, nutrient supply and inoculation procedures are described in detail in Chapters 3 - 6, respectively. 8.2.2 Sample collection and storage Samples of reactor liquid phase were transferred into clean, dry screw cap culture tubes ( 13 mm x 100 mm), then centrifuged at 2000 g for 10 minutes; the resulting cell pellets were stored at -35 0C for subsequent extraction and analysis. 8.2.3 FAME extraction A complete description of the FAME extraction procedure is provided in Appendix G. 8.2.4 FAME analysis FAME analysis was performed by using a HP 5890A gas chromatograph (Hewlett Packard, Palo Alto, CA), equipped with a flame ionization detector (FID) and a 25-m HP Ultra 2 capillary column using UHP (99.999%) hydrogen gas as canier (50 ml/min). The oven temperature was programmed from 170 0C to 270 0C, heating for 20 min at a rate of 192 5 oC/min, followed by a 3.4-min isothermal period at 270 0C. The injector and detector temperatures were 250 0C and 300 0C, respectively. The analysis time for each sample was ca. 23.4 minutes. 8.2.5 Numerical analysis methods All FAME data consisted of peak areas expressed as percentages of the total peak area of a trace. 'Ihree numerical methods were used for the analysis of the FAME data: (1) Calculation of the overlap coefficient, So (Bousfield er al., 1983). So is a measure of the degree of overlap of two superimposed traces, both scaled to the same total area of 100: So (i, j) = 100 - 0.5 2 ”inc - xjkl (8-1) where: Xik, Xjk = the percentage areas of the kth peak for the ith and jth culture respectively. The overlap coefficient attempts to mimic the way in which traces might be compared visually. In an intuitive sense, two traces which could be superimposed exactly would be considered completely similar, whereas two which showed no overlap would be completely dissimilar. Thus the value of the overlap coefficient 80 can be defined as: 0 5 S0 (i,j) S 1 (8-2) (2) Principal-component analysis. Principal-component analysis was performed using the SYSTAT 6.0 statistical package (SYST AT. Intelligent software, Evanston, IL). (3) Cluster analysis. Cluster analysis was completed using the median linkage method and using the SYSTAT 6.0 statistical package (SYSTAT. Intelligent software, Evanston, IL) 193 8.2.6 Experimental design In this study, three daughter reactors were subjected to long-term periodic substrate perturbations. During the perturbation period, the influent glucose concentrations of the daughter reactors were alternately changed in amplitude from 16 g/L to 0 g/L (mineral media only) on 2 day, 6 day and 10 day square wave cycles respectively. A 10 day hydraulic retention time (dilution rate of 0.1 d'l) was maintained throughout the experiment in all the perturbed daughter reactors. The daughter reactors were continuously operated for a period of 450 days including ca. 50 days under steady-state baseline conditions and ca. 400 days under perturbation and recovery conditions. The microbial community changes in the perturbed daughter reactors and in the control mother reactor were monitored using fatty acids profiles by periodically sampling these reactors. Fatty acids profiles obtained before the initiation of the substrate perturbations served as baseline reference values. FAME samples were more frequently collected during the operational stages when rapid environmental changes occurred (as indicated by other Operational parameters such as COD and VFA concentrations) to trace possible microbial community changes that occurred during these relatively short time periods. This sampling schedule allowed us to explore the relationship between FAME profiles and operational parameters. 8.2.7 Nomenclature for fatty acids configuration The fatty acid designation sequence used in this Chapter is as follows: number of carbon atoms, number of double bonds, and the location of double bonds in relation to the omega (methyl) end of the fatty acid. The number of double bond are indicated after a colon symbol. When the location of the double bond is indicated as "W6", the double bond is in the 6 carbon, counting from the omega end. Special structures are designated: "a” for anteiso; "i" for iso; "cyc" for cyclopropane, and "OH" for hydroxyl group. A more detailed description of fatty acid nomenclature is presented in Appendix H. 194 8.3 RESULTS 8.3.1 Determination of minimum sample size for FAME analysis In order to determine the minimum sample size necessary for reliable FAME analysis, 5 different sample volumes were analyzed. For each level, triplicate samples were analyzed to minimize random error. The sample volumes and the equivalent biomass (as VSS) used for determination of the minimum sampling size for the mother reactor culture are presented in Table 8-1. FAME profiles for samples with a liquid volume from 3 ml to 10 ml were essentially the same (Table 8-2). When the liquid sample volume was reduced to 1 ml (equivalent to a biomass dry weight as VSS of ca. 1.12 mg), significant differences among the FAME profiles were observed. Some of the cell membrane fatty acids, which were present in relatively small abundance, disappeared from the FAME profile. Similar results were obtained from other parallel experiments (results not shown). Based on these results, a minimum biomass (as VSS) amount of ca. 3.5 mg was determined necessary for reliable FAME analysis. Because of the relatively large fluctuations in biomass concentrations in the perturbed daughter reactors, a liquid sample volume of 6 ml was chosen as the routine sampling size for FAME analysis. For samples which had extremely low biomass concentrations, a larger volume was used. The accuracy and the reproducibility of the GC-FAME analysis were evaluated with triplicate samples from the different anaerobic reactors. The coefficient of variation for most fatty acids was generally below 5% for triplicate analysis. Standard deviations higher than 5% were only observed for the fatty acids with low abundance (< 4 percent). 195 Table 8-1. Sample volumes and equivalent biomass (as VSS) of the mother reactor culture used for minimum sampling size determination Sample volume (ml) Equivalent Biomass (mg) 0.5 0.56 l 1.12 3 3.36 5 5.60 10 1 1.20 Table 8-2. Effect of sample volume on relative abundance of fatty acids for FAME analysis. Fatty acid 10 ml 5 ml 3 ml 1 ml 0.5 ml a- 13:0 02610.03 0.181002 0.221002 -a - 13:1 2.091013 2.041014 2.06101 1 - - 13:0 2. 1610.08 1.661007 1.711009 - - i- 14:0 4.141022 3.671017 3.731014 - - 14:0 12.491027 12.151027 11.981026 13.341028 15.861038 i-15:0 13.041033 13.111029 12.871038 15.671047 18.991065 a-15:0 25.031055 24.851057 24.771071 29.911072 38.6411 . 16 15:0 18.261046 18.491051 18.341038 22.951067 26.511066 i-14:0 3OH 3.391021 3.421017 3.581019 - - 14:0 20H 0.351002 0.311010 03410.09 - - i-16:0 4.621019 4.771017 4.831018 5.761023 - a— 16:0 05110.10 0.541005 05610.08 - - 16:0 4.931022 5.151027 5.191023 6.531030 - i-17:0 1.361008 1.561010 1.521012 - - a-17:0 4.581031 5.021025 5.181027 5.831031 - 17:1 w6c 0.781010 08510.09 0.861011 - - 17:0 1.191007 1.281010 1.291007 - - i-17:0 30H 0.391009 0.44101 1 04510.07 - - 17:0 20H 0.421006 0.491010 0.521006 - - a: Not detected 196 8.3.2 Characterization of microbial community structure in anaerobic reactor populations by FAME analysis Typical FAME profiles obtained from the mother and the daughter reactor populations are presented in Table 8-3. FAME profiles designated daughter 1/ 1-1, daughter 3/3-1 and daughter 5/5-1 were obtained from 1/ 1, 3/3 and 5/5 day daughter reactors under the steady-state condition that re-established after long-term perturbation; FAME profiles named daughter 3/3-2 and daughter 5/5-2 were obtained from 3/3 and 5/5 day daughter reactors during a metastable period prior to re-establishmentof a stable steady-state. At least 41 fatty acids were identified among the different bioreactor microbial populations. Under steady-state conditions, fatty acids i-15:0, a-15:0 and 14:0 were prevalent. Significant amounts of these fatty acids were observed in all steady-state anaerobic reactor populations. Fatty acids i-14:0 3OH, 15:0 and 16:0 were also present in relatively significant amounts. During the metastable operation periods, significantly different FAME profiles were observed. The prevalent fatty acids for the stressed metastable steady-state reactor populations were straight chain fatty acids 16:0 and 14:0; 18:0 and 16:1 w7c were observed in particular populations. As shown in Table 8-3, some fatty acids were detectable for only a particular reactor community. For example, fatty acids 10:0 and i- ll:0 were found only in the 1/ 1 day daughter reactor, and fatty acids a-13:0 and a-l9:0 were found only in the mother reactor community. Differences in microbial communities among the anaerobic reactor populations can be recognized by comparing the abundance of the major cell membrane fatty acids and the unique fatty acids which are present only in particular reactor communities. After a stable steady-state was re-established, the mother reactor and 1/1, 3/3 and 5/5 day daughter reactor communities appeared to have similar populations. In contrast, the structures of the stressed metastable 3/3 and 5/5 day daughter reactor communities were significantly different. 197 Table 8-3 FAME profiles for the mother and the daughter reactor microbial populations Fatty acid mother daughter daughter daughter daughter daughter 1/1-1 3/3-1 5/5-1 3/3-2 5/5-2 10:0 -a 0.1 1 - - - - i-l 1:0 - 0.28 - - - - 1 1:0 - 0.66 0.12 - - - 12:00 0.10 5.11 0.12 0.99 3.96 1.31 i- 13:00 0.16 1.48 2.24 0.26 - 0.85 a-l3:0 0.08 0.00 - - - - 13: 1 2.28 1.51 - - 0.97 4.05 13:0 0.89 0.77 0.56 - - 0.97 i-14:0 2.58 4.92 3.06 13.58 1.10 - 14:0 10.35 10.78 10.09 12.81 10.94 14.07 i-13:0 30H - 0.12 - - - - i-15:l - 0.17 - - - - i-15:0 16.36 27.06 25.34 2.46 1.25 2.96 a-15:0 26.54 10.78 1 1.50 37.22 4.54 3.70 15:1 w8c 0.41 - 0.22 - - - 15:0 9.47 2.82 9.02 3.49 1.81 8.10 i-14:0 30H 8.45 9.86 12.71 0.29 - - 14:0 20H 0.45 0.17 0.49 - - - i-16:0 3.85 5.78 5.20 4.64 - - a- 16:0 0.50 - 0.83 - - - 16:1 w9c 0.21 0.15 0.22 - 1.38 2.04 16:1 w7c 0.30 0.17 0.32 2.21 2.03 1 1.60 16:0 4.88 3.98 5.42 12.36 48.19 26.98 i-17:0 w9c 041 0.72 0.21 - - - i-17:0 2.67 4.50 2.30 - - - a-17:0 4.45 1.83 6.83 2.14 - 0.39 17:1 w8c 0.52 - 0.32 - - 4.28 17:0 cyc 0.41 0.21 - 2.21 2.92 3.52 17:0 1.77 - 1.38 0.14 2.14 2.58 16:0 3OH - - - 4.52 - 6.24 16:0 20H - - - - - - 18:1 w9c - 0.27 0.31 - 0.94 0.99 18:0 0.21 0.65 0.31 0.18 1 1.04 1.07 i-17:0 3OH 0.60 0.89 0.38 - - 0.66 17:0 20H 0.31 0.00 0.28 0.24 1.43 - i-19:1 - 4.14 - 0.26 3.42 0.79 17:0 30H 0.27 - - - - 0.50 a-191) 0.14 - - - - - 19:0 cyc, w8c 0.13 - - - 1.94 0.70 20:3 w6,9,12c 0.27 0.13 0.23 - - 0.60 20:2 w6,9c - - - - - 0.54 a: not detected 198 The percentages of saturated, branched, unsaturated and cyclopropyl fatty acids in the total cell membrance lipids have been found to be useful criteria for microbial community analysis (Guckert et al., 1986; Kaneda, 1991). In Table 8-4, these percentages are illustrasted for the different reactor populations. Distinctive differences in microbial communities existed between the steady-state and the stressed anaerobic reactor populations. Under steady-state conditions the majority (ca. 60 to 70%) of the cell wall fatty acids were branched fatty acids. Unsaturated and cyclOpropyl fatty acids were both present at low levels. In contrast, under metastable steady-state conditions, the majority Table 8-4 the abundance (%) of the saturated, branched, unsaturated and cyclopropyl fatty acids in the different reactor populations Fatty acid M Dl/l-l D3/3-1 D5/5-1 D3/3-2 D5/5-2 Saturated 27.7 24.9 27.0 30.0 77.8 61.4 Branched 66.8 68.2 70.6 60.8 6.9 9.4 Unsaturated 5.0 6.7 2.4 7.0 10.4 24.9 Cyclopropyl 0.5 0.2 0.0 2.2 4.9 4.3 Table 8-5. Similarity matrix (So, %) of FAME data for reactor communities M D l/l-l D3/3-1 DS/S-l D3/3-2 D1/1-l 69 D3/3-l 79 81 D5/5-1 60 46 45 D3/3-2 29 35 27 39 D5/5-2 38 32 36 46 60 199 of the cell membrane fatty acids were saturated fatty acids (ca. 60 to 80%). Unsaturated and cyclopropyl fatty acids were both present at relatively high levels. By grouping fatty acids, a similar conclusion can be achieved for the community population similarities of the various microbial populations as determined by the direct comparison of the FAME profiles. Quantitative taxonomic techniques were used to evaluate changes in FAME profiles. Overlap correlation coefficients, principal-components and median linkage clusters are provided in Table 8-5, Figures 8-1 and Figure 8-2, respectively. The matrix of similarity coefficients calculated for each pair of FAME profiles (Table 8-3) is presented in Table 8-5. By definition, a high percentage for So means high similarity between the two microbial communities that are being compared. Choosing the mother reactor FAME profile as the reference community (column 1 in Table 8-5), the daughter reactor communities can be ranked from most to least similar as: 3/3-1, 1/1-1, 5/5-1, 5/5-2 and 3/3-2. Based on the values of the similarity coefficients (So), the FAME profiles can be sorted into two groups: the FAME profiles obtained from steady-state reactor populations, and the FAME profiles obtained from the stressed reactor populations. From the results presented in Table 8-5 the most similar pair was 3/3-1 and 1/1-1, and the next most similar pair was the mother and daughter 3/3-1. The FAME profiles for the 3/3-2 and 5/5-2 were significantly different from the steady-state reactor FAME profiles, the similarity between these two FAME profiles was relatively high. Significant differences in the populations were evident when the FAME profiles of the different stable steady-state and metastable reactors were compared using principal- component analysis (Figure 8-1). The stable steady-state reactor communities separated from the metastable communities by both the first and the second axes (P1 and P2), which accounted for 50.1% and 31.0% of the variation in the FAME profiles, respectively. If the mother reactor FAME profile is selected as a reference, the relatedness of FAME profiles as evaluated in P1-P2 space (Figure 8-1) has exactly the same order as previously P2 200 LO SIS-2 ‘ 313-2 0.5 " SIS-1 0.0- f 0.2 0.4 0.6 0.8 1.0 Pl Figure 8-1. Principal-component analysis of the FAME profiles listed in Table 8-3. 201 DISTANCES 0.000 1 .000 D1/1 0 3/3-1 __l _l D 5/5-1 D 3/3-2 0 5/5-2 ‘ | Figure 8-2. Clustering of the anaerobic reactor populations using the l-Pearson correlation coefficient, median linkage method with FAME profiles listed in Table 8-3. 202 determined using the similarity coefficient. Principal-component analysis shows that the community structure in the steady-state 5/5 day daughter reactor community was considerably different from those found in the other three steady-state anaerobic reactor communities. This trend was also revealed by previous FAME results. Using Join-cluster analysis method, the FAME profiles presented in Table 8-3 were sorted into two groups: FAME profiles from the stable steady-state reactor communities and FAME profiles from the metastable reactor communities (Figure 8-2). The communities in the first group were further divided into three sub-groups based on similarities in structures. This classification of FAME profiles was in good accordance with other FAME analyses presented previously. 8.3.3 Feasibily of Monitoring microbial community structure changes by FAME analysis The feasibility of using FAME analysis to monitor community structure changes in anaerobic reactor populations was evaluated during the 3/3 and 5/5 day substrate perturbation experiments. These results are presented in the following two sections. 1. Experimental results for the 3/3 day perturbation conditions As described in Chapter 5, significant microbial community changes occurred as evidenced by the physical-chemical, kinetic and genetic analyses, as well as by microscopic morphological observations. The FAME profiles obtained during this perturbation period also reflected changes in the microbial communities. As indicated by the FAME profile taken at day 28, a dramatic change in the microbial population occurred compared to the initial community (Figure 8-3). The predominant cell membrane fatty acids observed at day 28 changed from a-15:0 (28%) and i-15:0 (16%) to 16:0 (48%). Some other cell membrane fatty acids which had a relatively high abundance in the original steady-state FAME profile, such as 15:0, i-l4.0 30H and a-17:0, were absent or decreased to relatively low levels. This dramatic change in the microbial population, as A g v OUflhanvPI-unn < A QR.» v hire-.hnnuv: In: (s i n 5 fit 3 203 60 50" s 40' " . 8 30- % . g 20‘ —0— 12:0 ‘52 . —D— 16:0 —0— i-19:0 10 I I 0 0 400 Time (day) 35 30 - s 3 3 1820 g i-15:0 3 a-15:0 15:0 Time (day) Figure 8-3. Changes in major cell membrane fatty acids abundance during perturbation period in 3/3 day reactor populations. 204 indicated by the FAME analysis, coincided with the rapid accumulation in COD and VFAs (stage 1a, see Chapter 5). Subsequent FAME profiles obtained from day 88 to day 287 (the metastable steady- state) show that the structure of the perturbed daughter reactor community was not stable. The predominant cell membrane fatty acids shifted from 16:0 to 12:0, 18:0 and i-19:1. This change corresponded to changes in other parameters such as VFAs (see Chapter 5). A rapid change in FAME composition occurred between day 303 to 323. During this relatively short time period, the predominant cell membrane fatty acids shifted from 12:0 and i-19:1 to i-15:0, a-15:0, 15:0 and i-14:0 3OH, respectively, similar to what was observed in the original steady-state FAME profiles before the perturbation was applied. This change corresponded to the third stage of the 3/3 day perturbation during which time rapid degradation of the accumulated VFAs took place. The FAME profiles during the 3/3 day perturbation were also analyzed by sorting FAME constituents into saturated, branched, unsaturated and cyclopropyl fatty acids (Figure 8-4). This grouping procedure provided additional information about changes in the microbial community structure. After initiating the substrate perturbation, the abundance of the saturated cell membrane fatty acids increased from ca. 27% to 70%, while the branched cell membrane fatty acids concurrently decreased from ca. 70% to a level of ca. 6.4%. The abundance of saturated and branched fatty acids returned to near pre-perturbation levels of ca. 30% and 63%, respectively, at which point reactor performance again stabilized with high COD removal efficiency. When high COD removal efficiency was observed, an abundance ratio of saturated fatty acids to branched fatty acids of ca 2/1 was observed. When COD removal was poor, this ratio shifted dramatically to less than 1/10. The FAME profiles obtained from the 3/3 day reactor populations were evaluated using numerical methods. The similarity coefficients, calculated based on the initial steady-state community FAME profile, suggest changes in populations presented in the perturbed anaerobic reactor (Figure 8-5). Dramatic changes in the microbial community, _ Census Cm wow—=2.— V .VI”. Dun—Wu Cl 53 «5:. m $350.30 88528:: emcee—Em emfiaawm lei ._ V m. u m. 9. w a I I \om I ( . cc 11% 206 So (%) o . . . . - ' 300 400 Time (day) Figure 8-5. Changes in similarity coefficient So during perturbation period in 3/3 day reactor populations. 207 as indicated by the values of the similarity coefficient, So, were observed during the first and the third stages of the substrate perturbation experiment. FAME profiles with low So values were obtained during periods of metastable operation. This was when glucose fermentation intermediates (ie VFAs) accumulated. The correspondence between the So values and reactor effluent COD concentrations suggests that efficient COD removal was achieved by anaerobic communities which have a structure that is similar to that of the initial steady-state microbial community. Principal-component analyses illustrate the entire sequence of changes in community structure that occurred during the 3/3 day substrate perturbation period (Figure 8-6). The FAME profiles corresponding to stable steady-state and metastable steady-state are well separated in Pl-P2 space. Some of the FAME profiles belonging to the transient state lay between the steady-state and stressed FAME profiles in the Pl-P2 space. A zone appears to exist in P1-P2 space corresponding to "healthy" operation and a relatively high COD removal efficiency. By contrast, FAME profiles located far from the "steady-state zone" appear to reflect stressed conditions. This trend is illustrated by the FAME profiles for the 3/3 day perturbation experiment. Using cluster analysis, FAME profiles obtained from the 3/3 day perturbation experiment can be sorted into two groups (Figure 8-7). The first group includes the original FAME profile obtained before the perturbation initiated and the FAME profiles obtained when a high COD removal steady-state was re-established. The second group includes the FAME profiles obtained during the transient state and stressed metastable steady-state period. This group was further divided into two subgroups in which the first subgroup contains FAME profiles obtained from day 28 to day 110, and the second subgroup contains FAME profiles obtained from day 120 to day 311. The fact that two distinctive subgroups of FAME profiles existed during the metastable steady-state indicates that significant changes in the microbial population took place during this period. The similarity coefficient (So) as well as the relative abundance of the saturated and branched fatty acids correlated well with COD removal efficiency (Figure 8-8). Similar 208 1.0 0.8 ‘ 0.6 " 0.4 " 0.2 " P2 0.0 ‘ -0.2 " -04 ‘ 16 15 1 E 0.0 Figure 8-6. Principal-component analyses of the FAME profiles obtained from 3/3 day 0.2 0.4 0.6 0.8 P1 reactor populations. * 209 accusing c988 hue Qm So...“ confine .8an m2 , ii 0.2“ 0.0 . . . - . . 1 . . 0 5 10 15 20 25 Time (day) Figure 8-15. Changes in similarity coefficient So during the starvation period for the mother and the 1/1 day daughter reactor cultures. 221 0.15 1 M at h e r 0.10 ~ 0 - 2 0.05 [E . 4 5 3 g 0.00 " -0.05 ‘ . 6 E] -010 ‘ - n -0.15 I 0.95 1.00 1.05 Pl 0.15 Da u g hte r 0.10 ‘ ‘ 1 0.05 ‘ 2 . 43' [gr 3 0.00 - 5 -0.05 ' 6 ' c1 -O.10 ‘ D 7 -0.15 I 0.95 1.00 1.05 P1 Figure 8-16. Principal component analysis results for the FAME profiles obtained during the starvation test. 222 during the starvation test a slightly larger change occurred in the mother reactor population. 8.3.5 Effect of shock loading on FAME profiles The effect of shock loading conditions on FAME profiles were investigated using a constantly fed anaerobic reactor (m-l) and periodically fed daughter reactor 1/ 1. During this experiment, 6 grams of glucose was pulse injected into each of the two, 1.5 liter working volume anaerobic reactors while maintaining normal organic and hydraulic loading rates. Significant changes were observed in the FAME profiles during the shock loading. The most remarkable changes were observed in the abundance of the individual cell membrane fatty acids 12:0, 15:0 and 17:1 in the constantly fed anaerobic reactor m-l. The most noticeable changes in the abundance of the individual cell membrane fatty acids in the 1/1 day daughter reactor were 13:1, a-15:0 and i-14:0 30H (Figure 8—17). The abundance of the total saturated, unsaturated and branched cell membrane fatty acids changed rapidly just after the initiation of the shock loading test (Figure 8-18). The largest changes in the abundance of the grouped cell membrane fatty acids occurred during the first few days, when dramatic changes in reactor glucose fermentation intermediates took place. The similarity coefficients also reflected the changes in the microbial populations (Figure 8-19). The largest changes, as indicated by similarity coefficient (So), occurred in the first few days of the shock loading test for both anaerobic reactor populations. Based on the similarity coefficients obtained during the shock loading experiment more significant changes apparently occuned in the constantly fed anaerobic reactor population. 223 2%: 8 S “e r: = .e < O I I I I I I I fl 0 4 8 12 16 Time (day) —0— 13:1 + i-15:0 —El— a-15:O A -I-— i-14:030H e: 8 5 ‘3 :I E 40 “Cl SR1 12 16 Time (day) Figure 8-17. Changes in major cell membrane fatty acids abundance during shock loading experiment. 224 100 m-1 -—0— Saturated . + Branched 80 —D— Unsaturated F: 8 E a '6 I 5 A < Time (day) 100 1/1 —0— Saturated + Branched 80 d —D— Unsaturated s 3 a E a '6 I 5 .G < _A E 6 9 12 15 Time (day) Figure 8-18. Changes in grouped cell membrane fatty acids abundance during the glucose shock loading test. ' 225 .68 9.68— 082? 882w 05 wet—6 moweaeo “come—.080 55:85 67m BewE 3.3 as; 2 S w e c p p p n . n i o 7.: I om S IOII .. 1 0% S O . m C 1 00 cm 01 C 00" 226 8.3.6 FAME analysis results for mother and 1]] day daughter reactor MPN enrichments In order to obtain FAME characteristics for the predominant species in each trophic group in the mother and the 1/1 daughter reactor (after steady-state was re-established) populations, MPN enrichments were analyzed. The MPN enrichments used for FAME analysis included glucose fermenting bacteria, butyrate- and propionate-utilizing acetogens, and hydrogen- and acetate-utilizing methanogens. Due to the extremely low biomass concentrations in the MPN tubes, reliable FAME profiles were only obtained from the glucose fed MPN enrichments (Figure 8-20a). Both the mother and the daughter reactor FAME profiles, presented in Figure 8-20, came from the highest (10'9) dilution glucose MPN tubes. In both these samples, the cultures appeared to be morphologically homogeneous. Although these two cultures in the glucose MPN enrichments cannot be morphologically distinguished, two different predominant fermentative bacteria did exist, as indicated by the results of FAME profiles. The major cell membrane fatty acids for the fermentative bacteria in the mother reactor MPN enrichment were 16:0 (51.4%), 16:1 w7c (14.7%), 14:0 (12.5%) respectively, while the major cell membrane fatty acids for the fermentative bacteria in the daughter reactor enrichment were 18:0 (33.3%), i19:0 (16.9%), 12:0 (16.6%), respectively. A low similarity coefficient value of 0.27 also indicated the significant difference between the two glucose fermentative populations. The population difference indicated by FAME analysis was confirmed by the DNA analysis results. Cultures taken from the same glucose feed MPN tubes of the mother and the daughter reactors were directly used for polymerase chain reaction (PCR) amplification. The results of the H93, 11 restriction endonuclease digest of 16S rRNA polymerase chain reaction (PCR) amplified from the glucose fed MPN tube cultures are shown in Figure 8-21. Population differences can be observed from the distinctive restriction fragment patterns for the two reactor MPN cultures. 227 i J A 119:1 19:0 cyc w8c L Peak name 11 O Q 0 16:1 7 w c C] Daughter 14:0 4 I Mother 12:0 p fi I I’ I T I t I v . I 0 10 20 30 40 50 60 GC area (%) 4 19:0 cyc w8c 18.0 17:0 cyc ' 16:0 16:1w7c W Peak name “.0 'I/I/I’IA * Mother-2 120 g I Mother-1 0 10 20 30 40 50 60 CC area (%) Figure 8-20. FAME profiles obtained from mother and 1/1 daughter reactor glucose MPN enrichment cultures. 228 Figure 8-21. DNA analysis results for mother and 1/1 daughter reactor glucose MPN enrichment cultures. Lane 1, DNA standard (123 bp); lane 2 blank; lane 3 and 4, 10‘7 dilution MPN tubes of the mother reactor; lane 5 and 6, 10'8 dilution MPN tubes of the mother reactor; lane 7 and 8, 10‘7 dilution MPN tubes of the daughter reactor. 229 In contrast to the population difference observed between the mother and daughter reactor MPN enrichments, FAME profiles for the mother reactor glucose MPN enrichments from two independent MPN enumeration tests performed 8 months apart were quite similar (Figure 8-21b). A relatively high similarity coefficient value of 0.86 was obtained between these two samples. This result indicates that the predominant glucose fermentative population in the mother reactor was quite stable during the experimental period. The above observations further confirmed the feasibility of using FAME profiles as indicator to identify or distinguish individual microbial species in the anaerobic systems. 230 8.4 DISCUSSION Most of substrate perturbation experiments reported in the literature focus on anaerobic system response in terms of kinetics and operational performance, (i.e. in substrate degradation rate, accumulation of metabolic intermediates and changes in biomass concentration). Little is known about the effects of substrate perturbations on anaerobic microbial community structure. This is, in part, due to the lack of a reliable technique for monitoring of changes in microbial communities. Changes in anaerobic populations can not be effectively monitored or identified using ordinary analytical methods, such as microorganism isolation and incubation, microscopic observation or kinetic parameter evaluation. This work establishes that the effects of environmental perturbations on microbial communities can be effectively investigated using FAME analysis. FAME analysis revealed community structure differences between the periodic feed daughter reactor and the control mother reactor communities and made it possible to monitor population changes during the entire perturbation period. In response to the periodic substrate perturbation, significant changes in microbial populations occurred in all the perturbed daughter reactor communities. This observation was in good agreement with other observations such as changes in COD and VFA concentrations, substrate degradation rates, microscopic examination and DNA analysis. For example, a solid correlation was observed between the similarity coefficient (So) or the abundance of the grouped cell membrane fatty acids and conventional reactor operational parameters, such as COD and total VFA concentrations. The abundance of the saturated, branched and unsaturated cell membrane fatty acids were previously found to be a useful tool for the analysis of microbial community structure in aquifer and soil ecological systems (Guckert et al., 1986; Kaneda, 1991). Guckert et al. (1985) reported that the abundance of the grouped cell membrane fatty acids 23 l were quite different among the marine sediments incubated under aerobic, facultative and anaerobic conditions. During the present study, the abundance of the total saturated, unsaturated and branched cell membrane fatty acids was found to sensitively reflect the community structure changes in anaerobic reactor communities. Similarity in the abundance of the grouped cell membrane fatty acids was found in all the stable reactor communities. The abundance of the saturated, branched and unsaturated cell membrane fatty acids accounted to ca. 30%, 65% and 5% of the total cell membrane lipids under normal operational conditions. This ratio significantly changed during the unbalanced transient and metastable steady-state conditions indicating dramatic changes in the microbial communities. FAME profile is known as a multith description of a viable microbial commrmity structure (White, 1983). Each trophic group of bacteria usually has its own special composition and predominant cell membrane fatty acids (Ikemoto et al., 1978; Kaneda, 1991). The similar abundance of the grouped cell membrane fatty acids observed for the steady-state anaerobic populations illustrates that under given operating conditions, a certain microbial community structure is required to achieve effective glucose conversion to methane. To effectively cany out methane fermentation, the fermentative, acetogenic and methanogenic bacteria must work syntrophically, and sufficient organisms from each group (or guild) of bacteria must be present in the total community. Any significant changes in microbial composition will directly change the abundance of the grouped cell membrane fatty acids. Theoretically, two different kind of population changes may occur in anaerobic systems. The first is the community structure changes caused by imbalance among different major trophic groups of anaerobic bacteria. The second is community structure changes resulting from a shift of predominant species within a major trophic group or guild of anaerobic bacteria. These two kinds of community structure changes may occur 232 individually or simultaneously. The first kind of changes can be easily observed using ordinary analysis methods because it is often accompanied by large changes in conventional operational parameters. The second type of change in community structure may not be easily observed by ordinary analysis. This is because a shift in the predominant species within the same trophic group may result in only slight changes in kinetics of substrate utilization. Both of the above changes in community structure may be identified using FAME analysis. Considering the fact that each major group of anaerobic bacteria has its own unique abundance in grouped cell membrane fatty acids (Kaneda, 1991), large changes in abundance of the grouped cell wall fatty acids may indicate imbalance among the major trophic groups of anaerobic populations. On the other hand, changes in the abundance of individual cell membrane fatty acids, which do not largely change the abundance of the grouped cell membrane fatty acids, may indicate a shift in predominant species within the same trophic groups of the anaerobic communities. However, this speculation will need additional experimental verification. The abundance of cyclopropyl fatty acids is a stress marker in many FAME related biosystem studies ( Knivett and Cullen, 1965; Lepage et al., 1987; Hedrick, 1991). Hedrick (1991) reported that the abundance of the cell membrane fatty acid cyc 17:0 was much more variable in disturbed samples and was significantly greater in an overfed reactor population than in a healthy population. During an acetone-butanol fermentation study, Lepage et al., (1987) found that the abundance of cyclopropyl fatty acids considerably increased when the studied population was exposed to high concentration of solvents or cultivated under decreasedtemperature conditions. In the present study, a higher abundance of the cyclopropyl fatty acids was often observed during the transient stage of the substrate perturbations. At that time, the most significant change in environmental conditions and microbial population imbalance often took place. These 233 results indicate that the increased abundance of cyclopropyl fatty acids may also have some use as a stress marker for anaerobic reactor communities. Results of FAME analysis obtained from the mother and the daughter reactor MPN enrichments show that FAME profiles can be used to distinguish changes within a trophic group in anaerobic systems. Considering the fact that the anaerobic community is composed of several specific subgroups of fermentative, acetogenic and methanogenic bacteria and that each group of bacteria may possess its own unique composition of cell membrane fatty acids, then mathematical analysis of anaerobic microbial communities should be possible if the FAME composition (the fingerprint of anaerobes) can be obtained for the predominant anaerobic microorganisms present in specific metabolic groups. Procedures for numerical FAME profile analysis have been developed and verified in this experiment for effective study of microbial population changes in the anaerobic reactors. One of the major advantages of using numerical methods to treat raw FAME data is that quantitative comparisons of the individual microbial communities can be performed. By this method, the information contained in the raw FAME profiles can be more accurately interpreted and represented for further analysis. This essentially avoids subjective error and uncertainty resulting from using "vassal" or other semi-empirical methods for FAME data interpretation. This was verified by using the similarity coefficient (So), clustering and principal component analysis methods to treat FAME profiles. Highly consistent conclusions were reached using these three different numerical treatments of the same FAME profiles. Clustering analysis results indicated that perturbed anaerobic communities underwent several different structural changes. Clear evidence has been visually presented, using principal-component analysis, that the perturbed microbial communities tend to return to their original steady-state composition after adaptation to the environmental perturbation. 234 This is reasonable because only under conditions where the substrate supplied to the anaerobic community can be completely metabolized does the maximum free energy become available for the entire anaerobic community. This is likely the major driving force for the disturbed anaerobic system to return to its steady-state (or attractor) condition. Through principal component analysis, it has also been observed that a zone of "good health" apparently existed in the Pl-P2 space which was valid for all the anaerobic reactor populations examined in this study. This zone will likely change to some degree depending upon the substrate fed to the system. This was not examined herein. Although FAME analysis cannot answer all experimental questions regarding anaerobic reactor microbial community structure changes and operational performance, it does represent an additional set of tools for studying the anaerobic systems. Based on the findings in this work, FAME analysis can be effectively used to identify and monitor anaerobic reactor population community structure changes. If enough steady-state baseline information has been obtained to develop comparison standards (the "heath zone" in Pl-P2 space, the "heath ratio" of the abundance of grouped cell wall fatty acids, ect.), it may be possible to use FAME profiles to assess anaerobic reactor operational performance. Analysis techniques for FAME profiles can obviously be extended for use in other biological treatment systems. . .1 . . . . 1r zl‘vf. - t _» v , r.. {I . ,‘7_lllrr‘ " . .1',’ 7" ' . I .3 ’ll'“ J Am ‘rfl '11.. ' 1' ‘ " 1..» 71.063“ . ~ .- ‘-‘\- - ~r<>illi...w.~ . - ’ . . r. l a 14.171... .1 I . .' H r. . 1”,, N.,, _ . . . ,., , IINW. .. . "1111111“ . .1 ( ..l ..- . .-"mutu 1,5,”! .. r. :‘--,..l‘_ "'r.« .. -‘.dl 'L:l‘1.l"- _Ill.'l.:‘.; : ., 21;. ..». ,r ' . 1'. a width“ . l‘ '.‘rv.r..-1Im‘l:.ur luzr :r-mxn. .. 1:1 ' ., «m... l‘HIffiJI.M . . .- 91113111115111“ -‘lll‘ri.‘{ ,Z l». " ‘. ' . ' .1 . I. ‘ “n‘ 19M .‘ I’- I 109mm” lune-(10mm .~ ~. , . . . ,1, 31151.1”. , ' I 4 Slut-V1150“ (lgmmb 51 . x I . r' . . "In .mmn; mural-M- “mu W' 5d” 413151.01} m -.- 1 1 - .-' . ;,- mum. and“ “dunk” '(le llaw 11'. .~ .I'l .. .' whim 5:11 7.. "gun m, rLI 4, WWI jot-II“ I'vlrli'uulr. -' .- E I'lll l '1"! ’IN’\‘I w . “mlma'i(lruutwiurtn. Hr'i l mauupindw' (‘113‘3’ILE n 235 8.5 SUMMARY Cell membrane fatty acids are widely studied group of lipids with sufficient taxonomic diversity that can be used in defining microbial community structure. Experimental results presented in this chapter illustrate that FAME analysis can be successfully used as a tool to identify and monitor microbial community structure changes in anaerobic reactors. The abundance of the grouped cell membrane fatty acids was an important indicator of changes in anaerobic reactor community structure and reactor performance. A relatively stable ”health" ratio among these grouped cell membrane fatty acids exist under specific operational conditions. Large changes in these ratios indicated community structure changes which were related to unbalanced operational performance. Using numerical analysis methods, the information contained in the raw FAME profiles can be more accurately and quantitatively interpreted and used for identifying and monitoring microbial community structure changes. By using principal-component analysis it is possible to "visually" trace microbial community structure changes in the P1- P2 space. Based on comparison with other operation parameters, a 'health' zone can be established in the P1-P2 space. In this manner, FAME profiles can be used to graphically ascertain whether the reactor is operating well or not. The microbial community structure changes illustrated by FAME analysis were in good accordance with other experimental observations such as chemical-physical, kinetic, genetic and morphological analysis results. Chapter 9 CONCLUSIONS AND ENGINEERING SIGNIFICANCE A. Conclusion 1. The microbial communities from steady-state anaerobic chemostats were sensitive to an applied periodic substrate feeding pattern. The rapid accumulation of glucose fermentation intermediates indicates an apparent decrease in the number or activity of methanogens during the transient and metastable steady-state periods. 2. The acetate-utilizing methanogens were seriously affected by the periodic substrate feeding pattern. Under the 1/1 and 3/3 day perturbation conditions, extremely long recovery times were observed for the acetate-utilizing methanogens. Experimental results suggest that in addition to the inherent slow growth rate of the acetate utilizing methanogens, pH related un-ionized VFA inhibition may be partially responsible for the protracted period required for the anaerobic communities to adapt to the 1/1 and 3/3 day perturbation conditions. 3. Through a long-term adaptation, the acetate-utilizing methanogens apparently acquired the ability to effectively metabolize acetate at pH as low as 5.9 to 6.2 and VFA concentration up to ca. 4,000 mg/L (as acetate), a level that was highly inhibitory for the original acetate-utilizing methanogens. 4. Hz-utilizing methanogens were impacted by the periodic substrate perturbations, especially under 1/1 and 3/3 day perturbation conditions. However, the adaptation time for the hydrogen-utilizing methanogens was relatively quick compared to that of the acetate- utilizing methanogens. 236 237 5. The fermentative bacteria were affected by the periodic substrate perturbation with dramatic changes in the glucose fermentation pathway. Under elevated hydrogen partial pressure, the electrons flowing through the propionate fermentation route increased from less than 5 percent to ca. 40 percent, indicating that the fermentative bacteria responded to the environmental changes by quickly shifting the pathway of fermentation to maximize free energy under the altered environmental conditions. 6. A cyclic and alternating accumulation of acetate and butyrate was observed during the 3/3 day perturbation test. Results of thermodynamic calculations indicate that accumulated acetate reduced the available free energy to levels at which anaerobic oxidation of butyrate could not proceed. Once acetate concentrations decreased and butyrate concentrations increased, the cycle reversed, and acetate accumulated. 7. The anaerobic systems did gradually adapt to the periodic substrate perturbations through changes in the microbial communities as evidenced by changes in kinetics of substrate utilization, FAME and DNA analysis and morphological observations. 8. The populations that became established in the 1/1 and 3/3 reactors could maintain long-term stable and effective operation. One of the reactors (III) was evaluated to assess its resistance to environmental impact. This reactor community appeared to be more resistant to environment shocks, such as pulse organic loading and a short term temperature decrease, compared to the control mother reactor community. 9. The perturbation interval is an important control parameter for anaerobic chemostat systems operated under periodic feeding conditions. It can directly affect recovery and operational stability of the perturbed anaerobic systems. Based on results of this study, stable communities can likely only be established and maintained for a certain range of perturbation intervals. 10 FAME analysis can be effectively used as a tool to identify and monitor changes in anaerobic microbial communities. Using numerical analysis methods, the information contained in raw FAME profiles can be accurately and quantitatively interpreted for comparison of microbial communities. 238 11. Well controlled periodic substrate perturbations can be used as a tool to improve operational performance and robustness of anaerobic systems. b. Engineering importance of this study. Many perturbation experiments have been performed to investigate the response of anaerobic systems to increased organic loading rate and other environmental shocks. These experiments have mainly focused on inhibitory effects, recovery processes and thermodynamic changes that resulted from the perturbations. In this sense, environmental perturbations, on engineered biological systems, are usually considered as having only negative effects. The results of this study, however, suggest potential benefits of perturbations on engineered systems. For example, others have reported a tendency toward gradually increased maximum substrate utilization rates when an anaerobic system was subjected to successive shock loadings (Cohen, et al., 1982). Although perturbation results often indicated microbial community changes (Harper and Poland, 1985; Fongastitkul, 1994), little effort was directed at evaluating the effects of environmental perturbations on microbial communities. This study demonstrates that anaerobic communities can be selected or "optimized" by well controlled periodic substrate perturbations. Controlled substrate perturbations can therefore be used as a tool to improve operational performance and robustness. Substrate perturbations can create "hostile" or harsh environmental conditions. This can break down the existing dynamic balance among the specific trophic groups (guilds) of the anaerobic community. Only the "robust" species which possessed or can develop the requisite ability for survive are retained under such conditions. less competitive species, which may dominate the original community will eventually be eliminated or their population size reduced. Long-term periodic substrate perturbations selected for these populations possessing higher substrate utilization capacities. Periodic substrate perturbations also changed fermentation intermediate distribution, resulting in more stable microbial 239 communities. The adaptive communities are better able to cope with unfavorable environmental changes. The above observation has ecological and engineering implications. Unlike microbial communities in nature which are always exposed to unstable environments (and are subjected to natural selection thereby), microbial communities in engineered biological systems are for most part developed and maintained under carefully controlled, steady- state conditions. Operational conditions in engineered biological systems are controlled so that the microbial communities can achieve "optimum performance", but, paradoxically, this practice may result in the selection of population with reduced ability to cope with unfavorable environmental conditions, such as temperature drop, pH change, pulse loading and other impacts resulting from operational errors. The situation is worse for anaerobic populations due to the inherent slow growth rates possessed by the key microorganisms in these systems. Results presented in this thesis suggest that properly controlled perturbations might be used to select for robust populations to replace "idle" or less competitive ones and to keep the population "strong" and "vigorous". This method might be viewed as an ecological method or technique to distinguish itself from other physical, chemical or biological methods which used to improve biological system operational performance. This new operational strategy is, obviously, not only applicable for anaerobic chemostats. Through more experiment and practice, its application can be expected to benefit other biological treatment systems. Chapter 10 RECOMMENDATIONS FOR FUTURE RESEARCH . Perform long-term periodic substrate perturbations under pH controlled conditions to investigate the extent Of the effect Of the perturbation with the role Of pH eliminated, and tO observe adaptation and the resultant population changes. . Isolate and identify predominant microorganisms in each individual subgroups Of the mother and the daughter reactor populations during different stages Of the perturbation. This will allow better understanding Of the microbial community changes during the perturbation periods. . Determine if a unique microbial community is established under given perturbation conditions by simultaneously initiating perturbations on replicate anaerobic reactors under identical conditions. . Examine the response Of periodic substrate perturbations on other anaerobic systems such as UASB or anaerobic FBR which are often employed for full-scale anaerobic treatment. . Conduct pilot or full scale substrate periodic feeding experiment. Before use Of periodic substrate feeding as a practical Operational strategy, a pilot or full scale experiment should be performed. 240 APPENDIX A APPENDIX A GENERAL ANALYTICAL METHODS 1. Sample collection and storage Gas samples were taken from the reactor gas sampling port using 1,000 r11 Unimetrics gas-tight valve syringes (Unimetrics Corp., Shorewood, IL) equipped with 25 gauge needles, and were analyzed immediately after taken from the reactor. Liquid samples were withdrawn from the reactor liquid phase using 10 ml syringes equipped with 22 gauge needles, centrifuged using a Hermle compact centrifuge (Model Z203, National Labnet Co., Woodbridge, NJ) at 3,500 x g for 10 minutes, and either analyzed immediately or frozen at -35 0C and stored for further analysis. 2. Analytical methods 1. COD, pH, SS, VSS and alkalinity Chemical oxygen demand (COD), pH, suspended solids (SS), volatile suspended solids (V SS) and alkalinity were determined in accordance with procedures from Standard Methods for the Examination Of Water and Wastewater (APHA, 1989). COD was analyzed using Hach COD glass vials and a Hach COD reactor (Model 45600 Hach Company, Loveland, CO). The COD values of the digested samples were determined using the ferrous ammonium sulfate titration method. pH was measured using an Orion Model 720 pH meter equipped with an Orion glass Ag/AgCl combination pH electrode (Orion Research Inc., Boston, Ma). SS were analyzed using 0.45 um glass fiber filtration disc (Whatrnan International Ltd, Maidstone, England). VSS were determined by ignition in a muffle furnace at 550 0C. 241 242 ii. Volatile fatty acids Volatile fatty acids (acetate, propionate, butyrate and iso-butyrate) and ethanol were measured using a Perkin-Elmer 8700 GC (Perkin-Elmer Co., PalO Alto, CA) equipped with a flame ionization detector (FID). VFAs and ethanol were separated at 125 0C using GP 10% SP-1200/ 1% H3PO4 on 80/ 100 Chromosorb in a glass column (6' x 2 mm ID) using nitrogen (50 mein.) as a carrier. The temperature Of injector and detector were 180 and 250 0C respectively. After centrifuged at 3500 g for 10 minutes, the supematant was acidified using concentrated H3PO4 (1/20, vol. acid/sample). Injection volumes were 0.1 to 0.4 ul depending on sample concentration. iii. Methane and carbon dioxide Methane and carbon dioxide were determined using a Varian model 3700 GC (Varian Associates Inc., Norwalk, CT) equipped with a thermal conductivity detector (TCD). CH4 and C02 were separated at 130 0C on 80/ 100 carbonsphere in a stainless steel column (6' x 1/8" ID ) using helium (40 mein.) as a carrier. The temperatures Of the injector and detector were 180 and 250 0C, respectively. An injection volume Of 0.2 ml was used for all analyses. iv. Gas phase hydrogen partial pressure Gas phase hydrogen partial pressure was measured using a RGA3 Reduction Gas Analyzer (Trace Analytical, MenlO Park, CA). The reduction gas detector was Operated at temperature Of 265 0C with 30 ml/min Of helium as carrier gas. The reaction bed temperature was 90 0C. An injection volume Of 1 ml was used for all analyses. v. Optical density The Optical density Of anaerobic reactor culture was determined by using a Perkin Elmer LAMBDA 6 UV/V IS spectrophotometer (Perkin-Elmer Co., Palo Alto, CA) at waveleght Of 650 nm against DI water. 243 vi. C, H, N elemental contents The elemental contents Of biomass carbon, hydrogen and nitrogen were determined by using a PE 2400 CHN elemental analyzer equipped with a P-E AD-4 Ultramicrobalance (Perkin-Elmer Co., Norwalk, CO). APPENDIX B APPENDIX B MPN ENUMERATION METHOD The five tube per dilution MPN method (APHA, 1989) was used to estimate the population distribution in the anaerobic reactors. Glucose-fermenting bacteria, propionate- and butyrate-utilizing acetogenes, and acetate- and H2-utilizing methanogens were examined. 1. Medium and substrate preparation a. Phosphate buffered basal media (PBBM) composition NaCl 2.7 g MgClz. 61120 0.6 g CaC12.2H20 0.3 g NH4C1 3.0 g Resazurin solution 3 m1 trace mineral II 30 ml Double distilled water 3000 ml After dissolving the above chemicals in D. I. water, the media solution pH was adjusted to 7.2 - 7.4 by adding 4 N NaOH or 4 N HCl. b. Amount Of substrate solutions 10% Glucose solution 100 ml 1 M Propionate solution 100 m1 1 M Butyrate solution 100 ml 1 M Acetate solution 100 ml 244 245 c. Stock chemical solutions: (1) 10% sodium bicarbonate solution 10 g NaHCO3 dissolved in 100 ml DI water. (2) Vitamin solution The vitamin solution used in this study was adopted from WOlin et al. (1963). It contained (mg [1000 ml DI water): Biotin 2 Forlic acid 2 Pyridoxine-HCl (B6) 10 Thiamine-HCI (Bl) 5 Riboflavin (B2) 5 Nicotinic acid 5 Pantothenic acid 5 Cyanocobalamin (B 12) 0.1 P-aminobenzoic acid (PABA) 5 The above compounds were added in the listed order to ensure complete dissolution. The solution was then filter sterilized (0.25 um filter) and stored in vacuum and N2 gas flushed, and autoclaved serum bottles. (3) 2.76 M phosphate buffer solution: Dissolve 15 g KHzPO4 and 29 g K2HPO4 in 100 ml DI water. (4) 2.5% Na28-9HzO solution: Sulfide solution was prepared under an oxygen free N2 atmosphere. Double distilled water was boiled for 15 min in a flask under N2 gas atmosphere and cooled for 5 min. Subsequently, 25 g Of Nags-9H20 crystal were transferred into the flask containing 1,000 ml DI water and dissolved by stirring. After the solution had cooled to room temperature (25 0C), concentrated HCl was used to adjust pH to 9.5 under a N2 atmosphere. 246 (5) Trace metal solution, it contained (g/1000 ml): FeC12-4H20 1.5 ZnC12 0.07 MnC12-4H20 0.1 H3BO3 0.06 COC12~6H20 0.19 CuC12-2H20 0.002 NiC12-6H20 0.024 N32M004°2H20 0.36 HCl (12N) 6.4 ml The FeC12I4H20 was first dissolved in the HCl, then this solution was mixed with 950 ml DI water. Subsequently all the other salts were dissolved into this solution. pH was adjust to 6.0 with l N NaOH. Final solution volume was adjusted to 1 L with deionized water. Substrate and chemical stock solutions were transferred into serum bottles(1/3 full), sealed with rubber stoppers and aluminum crimpers, there alternately evaluated under vacuum and flashed with N2 gas at 14 psig three times. Serum bottles were autoclaved at 121 0C for 25 min. 2. Dilution used for each substrate Glucose fermenters 10-5 - 10-11 H2-lltilizing methanogens 10'4 - 1040 Acetate-utilizing methanogens 10‘4 - 10—10 propionate-utilizing acetogens 10'4 - 10‘10 Butyrate-utilizing acetogens 10'4 - 10'10 For each dilution 5 tubes were used. 3. MPN tube preparation protocol (1) Boil and dispense media solution under a N2 gas purge. (2) Dispense 9 ml solution into each pressure tube and seal the tube with rubber stopper and aluminum ring. 247 (3) After autoclaving at 121 0C for 25 min, add the following amounts of autoclaved chemical agents to each tube: 10% HCO3’ solution 0.1 ml Vitamin solution 0.1 ml (after filtering through 0.2 pm glass fiber filter) Phosphate buffer 0.1 ml 2.5% Nags-91120 0.1 ml (4) Pressurize the tubes by 5% C02 + 95% H2 mixture at 14 psig except the tubes for H2 degraders (5) Add proper substrate to pressure tube as follows: 10% glucose 0.2 ml 1 M acetate 0.2 ml 1 M propionate 0.2 ml 1 M butyrate 0.2 ml 80% H2 + 20% C02 mixture 20 psig 4. Incubation After preparation, MPN tubes were incubated at 35 0C. The incubation time used for glucose fermenters was 14 days. The incubation time for all other microorganisms was 60 days. 5. Test result determination Test results reported here were based on CH4 detection for acetogens and methanogens. Tubes that had a CH4 concentration above 0.5 mM were evaluated as positive. The results for glucose-utilizing fermentative bacteria were recorded on the basis Of increase in Optical density compared to non-inoculated control tubes. Tubes with an Optical density > 0.2 at 650 nm were scored as positive. APPENDIX C APPENDIX C. EPIFLUORESCENCE MICROSCOPY DIRECT COUNTING METHODS Fluorescent dyes can improve the visualization Of individual microorganisms by binding to specific cell components ( Bratbak, 1985; Bitton, et al., 1993). Acridine Orange has affinity for DNA, RNA, acidic polysaccharides; while DTAF has affinity for protein. 1. Acridine orange direct count (DNA stain) Materials (1). 0.2 um pore size 25 mm diameter polycarbonate membrane filters (Fisher NO. 11021); (2). 1% Formaldehyde solution; (3). 0.1% Acridine Orange (Sigma C. I. 46005) solution; (4). 0.1% Polyoxyethylene sorbitan mono-OleateCI‘ ween 80, Sigma NO. P-1754) solution; (5). 80% ethanol solution; (6). 80% isopropanol solution; (7). Sterilized distilled water; (8). TWA-30 Microscope. Procedure 1. Take 1 ml sample from the reactor and immediately mix with 1 ml of 1% formaldehyde solution and 8 ml water in a 14 ml sterilized falcon tube and store at 4 0C for further analysis. 2. Serially dilute sample so that about 30 - 60 cells per grids can be Observed. 3. Add 1 ml Of 0.1% A0 solution into the sample (1 ml sample 8 ml water), mix well and store in the dark for 30 min. 248 249 4. Wash filtration glass column with water first and then and 80% ethanol. 5. Put filtration membrane on the filter. 6. Wash the membrane with 1 ml of the 0.1% Tween 80 solution. 7. Add 2 ml sample and 8 m1 sterilized DI water into the filtration column for filtration. 8. Add 2-3 ml 25% isoproponol solution to the frlter column to wash the membrane. 9. Remove the membrane from the fitter and place on a microscope slide (add a drop of oil on the slide first). 10. Count the orange spots under UV light on the Observation grids. In order to get reliable results, repeated counting 5-10 times should be performed for each sample. Calculation: Effective filter area is 231 mmz. Using the 63 X objective the graticule field is 156 X 156 um, the total area is 0024336 mm2. Therefore the effective filter area contains: 231 I 0.024336 = 9492.1 graticule field The total bacteria number in the sample can be calculated: Bacteria NOJml = 9492.1 x Nf x Nd Where: Nf. Average bacteria number Observed on graticule field Nd: sample dilution. 2. DTAF direct count (protein stain) Reagents: l. Phosphate Buffered Saline 0.05M NazHPO4 (7.8 g/L) 0.85% NaCl (8.5 M.) The NazHPO4 is diluted in 085‘? NaCl and adjusted to pH 9.0 and filtered through 0.2 rtm membrane prior to use. 250 2. DTAF fluorescent stain: 5-(4, 6-dichlorotriazin-2-yl) aminofluorescein dissolved 2 mg in 10 ml of PBS immediately prior to use. Stain is stable for several hours. 3. Distilled water (from CME lab) 4. 25% isopropanol-250 m1 Filter through a 0.2 tun membrane prior to use. 5. Immersion Oil - Type FF. 6. 0.1% formaldehyde solution. Materials 1. Falcon tubes - 17 x 100 mm; 2. DispO tube - 12 x 75 mm; 3. Plain microscope slides; 4. Cover glass - 22 mm sq.; 5. Filter - Black - 25 mm, 0.2 um; 6. Warring Blender; 7. Stop Watch; 8. Adjustable pipettes with tips to deliver 0.5, 1.0 and 5 ml; 9. Filter Apparatus; 10. Forceps; 11. Fluorescent Microscope; 12. 400/270 Kodak slide film (MSU Biochemical Store 36-roll 1442355, 24-roll 1442351). Procedure 1. Dilute the sample properly so that in the microscOpc grid about 100 cells can be observed. 2. In the dilution tube add 1 ml of 0.1 % fonnaldchydc solution plus 8 ml tilt-crud water. 3. Add 0.5 ml of the diluted culture in 12 "' 75 mm 31.1,“ ml“ 1\\ m duplicate. 25 l 4. Add 0.5 ml of DTAF stain, mix well by vortexing. 5. Place sample in the dark for 30 min. 6. After 30 min. mix all tubes on vortex again. 7. Place black filter on filter apparatus and Attach filter cylinder. 8. Add 5 ml of PBS in the cylinder. 9. Add all the stained sample to the PBS (5 ml) 10 Rinse the tube with a little PBS and pour into the cylinder. 11. Turn on the vacuum pump, open the valve to the cylinder, until filter is dry. 12. Add 5 ml of PBS-twice, applying vacuums until filter is dry between each 5 ml. 13. Add 5 ml Of distilled water — applying vacuum until dry. 14. Add 5 ml 25% isopropanol - applying vacuum until dry. 15. Close valve and turn Off vacuum pump. 16. Remove filter and place on microscope slide, add one drop Of immersion oil to filter, cover with cover glass. Cover glass may be sealed with fingernail polish. 17. Examine under the fluorescent microscope using the grid to facilitate counting. Calculation: Effective filter area is 231 mmz. Using the 63x objective the graticule field is 156 X 156 run, the total area is 0.024336 mmz. Therefore the effective filter area contains: 231 / 0.024336 = 9492.1 graticule field The total bacteria number in the sample can be calculated: Bacteria NOJml = 9492.1 x Nf x Nd Where: Nf. Average bacteria number observed on graticule field, Nd: sample dilution. APPENDIX D APPENDIX D PROTEIN MEASUREMENT PROCEDURE Protein analysis was performed according to the MBS method (Bradford, 1976; Sedmak and Grossberg, 1977). In this method the diluted samples were first digested with 10 N NaOH for two hours. Then 5 N HCl was added to neutralize the solution. Coomassie protein assay reagent was added to the neutralized samples and reacted for 20 min. The mixed liquor was transferred into 1.5 ml 1 cm path light polystyrene curettes and the absorbence Obtained using on spectrophotometer at a wavelength Of 595 nm compared to a blank. Materials: 1. 10 N NaOH solution (Cat NO 88255-1, Fisher scientific, Pittsburgh, PA); 2. 5 N HCl solution (Cat NO LC15360-2, Fisher scientific, Pittsburgh, PA); 3. Coomassie protein assay reagent (Cat NO 23200, Pierce, Rockford, IL); 4. 10 ml glass tube with screw cap; 5. 1.5 ml, 1 cm path polystyrene curette (Cat No LS-2410-100, Life Science Productes, Inc., Denver, CO); 6. Perkin-Elmer Lambda 6 UV/V IS spectrometer (Perkin-Elmer Corporation, Norwalk, CT). Procedure: 1. Add 1 ml diluted sample (1/ 10 of the original concentration) into 10 ml glass tube. 2. Add 0.5 ml 10 N NaOH solution into the tube mixing well with the diluted sample and digest for 2 hours. 3. Add 1 ml 5 N HCl solution into the tube to neutralize the solution. 4. Add 2 ml Coomassie protein assay reagent into the tube and react for 20 min. 252 253 5. Read absorption at wavelength 595 nm against Coomassie protein assay reagent blank (use 1 m1 DI water to replace diluted sample). 6. Calculate protein concentration by using standard absorption curve. 7. In this experiment the standard calibration curve was prepared by using standard protein solution with the following protein concentrations: 10, 20, 40, 60, 80 and 100 ug/ml. APPENDIX E APPENDIX E EFFECT OF pH AND VFAS ON MAXIMUM HYDROGEN UTILIZATION RATES The effect of pH and VFAs on the maximum hydrogen utilization rates Of the mother reactor culture was investigated using batch experiments. The effect Of VFA concentration on conversion of Hz-COZ was tested in 158 mL serum bottles containing 60 ml Of the mother reactor culture. The bottles were pressurized by addition of 20 psig of H2-C02 gas (80:20) into headspace. VFA concentrations examined using duplicate samples for each test are listed in Table E-l. Table E-l. VFA concentrations examined in the Hz-COz degradation experiment Acids Concentrations (mg/L) Acetate 200 1000 1500 2000 3000 4500 Propionate 100 500 1000 1600 3000 4000 Butyrate 100 1600 3200 4600 6200 The effect Of pH on conversion rates of H2-C02 by the mother reactor culture was evaluated using 158 ml serum bottles. Culture conditions were similar to the VFA studies. The pH of the culture was adjusted to 5.0, 5.5, 6.3, 6.8, 7.6 and 8.2 by adding different amount of 5N NaOH or 5N HCl solutions into liquid phase. 254 255 Bottles were incubated in a shaking bed at 35 0C. H2-C02 conversion rates were calculated based on methane production. Results reported were averages Of duplicate samples. The effect of pH on the conversion rates of H2-C02 is presented in Figure E-l. The Optimum pH for maximum hydrogen conversion by the mother reactor culture was ca. 5.5 to 6.8. When the pH was increased or decreased beyond this range, a significant decrease in hydrogen utilization rate was Observed. The effects of VFA concentrations on the maximum hydrogen utilization rate are presented in Figure E-2 (a-c). When acetate, propionate and butyrate concentrations were less than 2,000, 1,500 and 2,000 mg/L, respectively, the H2-C02 utilization rates were not effected. When the VFAs concentrations were increased to ca. 4,500, 3,500 and 6,000 mg/L, respectively, the Hz-COz utilization rates were decreased to ca. 87%, 89% and 87%, respectively, Of the maximum utilization rates measured under normal operation conditions (VFAs < 200 mg/L). V/VO 256 1.0 " 0.8 " 0.6 " 0.4 ‘ 0.2‘ 0.0 4. (l 5.0 6.0 7.0 8.0 Figure E-l. pH effect on hydrogen conversion rates. 257 1.0 1M 0.8 " ° 0.6 " Z d > d 0-4 y = 1.0284 - 3.774lc-5x R"2 = 0800 0.2 ‘ ‘ a 00 u I u I v I u I u (l l()()() 2()()() 3000 4000 5000 Acetate (mg/L) 1.0 ~ "‘ H D . U U 8 Ci 0.8 " o d 2 0.6 > ‘ y = 1.0096 - 2.869413-5x 1292 = 0.914 0.4 ‘ 0.2 ‘ ' b 0.1) . 1 I I I I . I () 1000 2000 3001) 4000 Propionate (mg/L) 1.1)L-\.I\.\‘\ ‘ O 0.8 ‘ ° 0.6 ‘ K . > 0 4 - y = 1.0100 - 2.539le-5x R"2 = 0.786 0.2 ‘ - c (1.0 I l . r . 1 . 1 . 1 . 1 () l()()() 2000 3000 4000 5000 6000 7000 Butyrate (mg/L) Figure E—2. VFA effect on hydrogen conversion rates at pH 7.4. (a) acetate; (b) propionate; (c) butyrate. APPENDIX F APPENDIX F DEVELOPMENT OF FREE ENERGY EQUATION USED FOR ACTUAL OPERATION CONDITIONS Based on Thauer er al (1977), the free energy change under physiological conditions (pH 7 and 25 0C) can be calculated as: AG' = 150" + RT 2 Vi 1n Ai (F-l) where: R = gas constant, 8.3143 J/K mol, T = Temperature, K, Vi = The stoichiometric coefficient for component A1 in a biological reaction ( i = 1,2,3,... ) and is negative for reactants and positive for products, Ai = the physiological concentration Of the component i in the reaction, mol, AGO' = the increment of free energy at standard physiological conditions (25 0C, 1 atrn and pH 7), KJ/mol. 150" was calculated from AG0 according to the following equation: [1100' = AGO + m Af'(H+) (F-2) Where: AG0 = free energy change under standard condition (25 0C, 1 atrn and unit solutes concentrations), KJ/mol, m = net number of protons in the reaction ( m is negative when more protons are consumed than formed), AF (11*) = the free energy of formation of a proton at pH 7. Af'(H+) in equation (F-2) can be written as: Al’(H+) = RT ln 10-7 (F-3) 258 259 Substitute Af'(H+) from equation (F-3) to equation (F-2), get: AGO. = AGo + m RT In 10‘7 (E4) TO use the same principle, the ”standard free energy change” under operation pH other than pH 7 can be calculated as: A001,, = AGO + m Ar'(H+) = AGo + m RT ln(H+) (F-5) Where: AGO'op = Standard free energy change at operating pH, KJ/mol, [H‘l'] = Proton concentration in the reactor solution, mol/L. Resolve AGo from equation (F-4) and substitute it to equation (F-S), get: Aco'op = A09 + m RT 1n([H+l/10-7) (F-6) Under Operation conditions, the free energy change can be calculated using following equation: AG'op = AGo'op ‘1' RT 2 VI 1“ AI (F‘7) Substitute equation (F-6) into equation (E7) and rearrange it, get: Ac'op = AGO' + m RT ln([H+]/10‘7) + RT 2 Vi 1n A1 = 1160' + 2.303 RT [rog([H+1/1o-7)m + 2 v1 Log Ai] (F-8) where the free energy units are K], the substrate concentrations are in mole and the gas pressure is atm. APPENDIX G APPENDIX G FAME EXTRACTION PROCEDURE 1. Saponification Samples were treated by strong methanolic base combined with heat to lyse the cells. Fatty acids were cleaved from the cell lipids and were converted to their sodium salts. (1). First add 1.0 ml methanolic base into each Of the culture tubes. (2). Tightly seal each tube with a clean Teflon-lined screw cap. (3). Vortex the tubes for 5- 10 seconds. (4). Place a rack of the sample tubes into a boiling water bath at 100 1 2 0C. (5). After 5 minutes, remove the tubes from the boiling water and cool them slowly. (6). Vortex the tubes for 5-10 seconds then return the tubes to the water bath and continue heating the tubes for additional 25 minutes. (7). After a total of 30 minutes of saponification in the water bath, remove and set the rack of tubes in a pan Of room temperature water. 2. Methylation Methylation converts the fatty acids (as sodium salts) to fatty acids methyl esters which increases the volatility Of the fatty acids for the GC analysis. (1). Add 2 ml methylation reagent tO each tube. (2). Tightly cap the tube and vortex the solution for 5-10 seconds. (3). Heat the tube at 80 0C water bath for 10 minutes. (4). Cool the tubes to room temperature by place the tubes in a tray Of room temperature water. 260 261 3. Extraction Fatty acid methyl esters are removed from the acidic aqueous phase and transferred to an organic phase with a liquid-liquid extraction procedure. (1). Add 1.25 ml hexane/MTBE extraction solvent to each tube. (2). Place batch of tubes in a laboratory rotator and mix end-over-end for 10 minutes. (3). Use clean Pasteur pipette to remove and discard the lower aqueous phase. (4). Add 3 m1 base wash solution to each tube and rotate the tubes end-to-end for 5 minutes. (5). Centrifuge at 2000 rpm for 3 minutes to clarify the interface between phases. 4. Base wash (1). Add 3.0 ml Of reagent 4, the base wash, into each tube. (2). Tightly cap and rotate the tubes end-over-end for five minutes. (3). Brief centrifugation (three minutes at 2000 rpm) is recommended to clarify the interface between the phases when an emulsion is presented. 5. Tramfer extract Transfer of the extract to sample bottle. After that remove the upper solvent phase to CC autosampler bottle for run GC analysis. APPENDIX H APPENDIX H NOMENCLATURE OF FATTY ACIDS A. Straight Chain H 11.11 11 11 H H H 11 11 11 11 11 11 H o l l | l l l l I I l I I | I I ll H- c-c--c-c-c-c-c-c-c-c-c-c-c-c-c-c-011 l l | l I l l l | l l l l l l H 11 11 H 11 H 11 11 ll 11 11 11 H H 11 16151413121110987654321 The above represents the straight chain fatty acid palmitic acid, written as 16:0, where the " l 6" represents the number of carbons in the compound, and the number after the colon indicates the number of double bonds in the carbon chain (in this case, none). Note the carboxyl group(COOH) at the right. ‘ Note that these compounds may also be written with the letter C in front of the number, sue}! a C1620. The letter "C" stands for "carbons" in the compound, and a compound Wfitten C16:0 is the same as one written 16:0. B- Unsaturated l - C18 conformation 1: 1 ::-—r1—-:: I ::-—ri—-:: I :r-—r1——:: I ::-—ri——:: I ::——re—-:: I ::-—r5—-:: I ::-—r) 16 15 14 13 12 11 10 262 263 'I'he designation 16:1 indicates that the compound has 16 carbons and 1 double bond. The above represents the unsaturated fatty acid 16:1 cis 9. Note that both hydrogens at the double bond are on the same side. 2. trans conformation 61514131211109 The above represents the unsaturated fatty acid 16:1 trans 9. Note that the hydrogens at the double bond are on Opposite sides of the compound. C. Iso T HH-C-HHHHHHHHHHHHHHO lllllllllllllllll "‘if“???WW"???"f‘°'°"'c'c'°" I I H H H H H H H H H H H H H Jill‘l 16151413121110987654321 The above represents the fatty acid 17:0 ISO. A methyl group occurs at the second to the last carbon in the chain. D. Anteiso 1‘ TT“‘?‘”?TTTTH"“"""° H-?-¢':-?-¢l:-f-c-cuf-tlz-c-c|:-cl:-r|:-tl:-é-g-0H HHHHHHHHHlIllIllLEIIlliJl 161514131211109 8 7 6 s 4 3 21 264 The above represents the fatty acid 17:0 AN'I'E. A methyl group occurs at the third to the last carbon in the chain. E. Cyclopropane H H \/ H H H H H H c H H H H H H H o I I I I I I /\ I l I I I I I II 14 .. c - c- c- c- c- c - c - c - c - c - c - c- c- c- c- c - OH I I I I I I I I I I I I I I I H H H H H H H H H H H H H H H 16 15 14 13 12 11 1o 9 a 7 6 s 4 3 2 1 “The above represents the fatty acid 17:0 CYCLO 9-10. In this case, this compound is made from 16:1 cis 9 with the addition of the carbon group at the double bond position. F. Hydroxy l . T'he 2-hydroxy H H H H H H H H H H H H H H 0H 0 IIIIIIIIIIIIIIIII 11 - c - c - c - c - c - c - c - c - c - c - c- c- c- c- c - c - 0H II I I | | IIII I II I I H H H H H H H H H H H H H H H 16 1s 14 13 12 11 1o 9_ s 7 6 s 4 3 -2 1 The above represents the fatty acid 16:0 20H, where a hydroxyl group is added at the 2 (alpha) position. 2- The 3-hydroxy H H H H H H H H H H H H H OH H o I I I I l I I I I I I I I I I II H - c - c c c c - c- c- c- c c c c - c - c - c - c - OH I I I I I I I I I I I I I I I H H H H H H H H H H H H H H H 16 1s 14 13 12 11 1o 9 a 7 6 s 4 3 2 1 The above represents the fatty acid 16:0 30H, where a hydroxyl group is added at the 3Cbem) position. ‘ 265 3. Other hydroxys Hydroxyl functional groups may occur at other positions besides the second and third carbon. These are not common across the numerous species of bacteria which have been looked at to-date. G. Mixed functional groups I HH-c-HH H H H H H H H H H H H m o IIIIIIIIIIIIIIIII H c- c c c-c-c c c c- c-c-c- c- c- c- c m I'I I IIIIIIIIIWIII H H H H H H H H H H H H H H H 16 16 14 13 12 11 1o 9 a 7 6 s ‘4 3 2 1 Combinations of the various functional groups also occur. The above represents the fatty acid 17:0 ISO 201-1 H. Fatty acid methyl'ester H H H H H H H H H H H H H H H o H I I I I I I I I l I I I I | I H I H -c- c c c- c-c- c- c- c- c-c-c-c-c-c C-O-C-H I I I I I | I I I I I I I I I I H H H H H H H H H H H H H H H H 16 1s 14 13 12 11 1o 9 a 7 6 s 4 3 2 1 The above represents the fatty acid methyl ester 16:0, written as 16:0 FAME on the MIS printed reports. A methyl group is added to the carboxyl group to increase volatility f0? GC analysis. I. Dimethyl acetal T If Iil III If T Iii H H H H H H H H O-CH3 "~1111-1-1-1-1-1-1-1-1-1-1-1-1 H H H H H H H H H H H H H H H O-CH3 161514131211109 8 7 6 5 4 3 21 266 The above represents dimethyl aldehyde 16:0, written as 16:0 DMA on the MIS printed reports. Dimethyl acetals occur as analogs of the fatty acids present in anaerobic bacteria, and can contain any of the above functional groups. They result form the ether- linked lipids in plasmologens. J. Normal hydrocarbon H H H H H H H H H H H H H H H H | I | | I | I I I I I I I I I | H -c- c- c-c c c c-c-c-c-c-c-c-c-c-c-H I I I I I I I I I I | I I I I I H H H H H H H H H H H H H H H H 16 15 14 13 12 11 1o 9 a 7 6 5 4‘ 3 2 1 The above represents normal hydrocarbon 16:0, also written as 11 16:0. K. Aldehyde H H H H H H H H H H H H H H H 0 I I I I I I | I | I I I I I I II H c- c- c- c-c-c- c- c c- c-c-c-c-c-c-c-H I I I I I I I I I I I I I I I H H H H' H H H H H H H H H H H 161514131211109 8 7 6 5 4 3 2 1 The above represents aldehyde 16:0. L- Alcohols 1111111111111111 I I I l " 'I'I'I'I'I'I'I'I'I’I"°"c I I I" I’H I I I H H H H H H H H H H H H H H H H 161514131211109 8 7 6 5 4 3 21 The above represents the alcohol 2-octadecanol. This compound, as well as 2- elc=03anol (20 carbons long), occur in some species of Mycobacterium. LIST OF REFERENCES LIST OF REFERENCES Agargy, F. J., Cole, R. D. and E. A. Pearson (1963) Kinetic and activity parameters of anaerobic fermentation systems, SEAL Report 63-2, Sanitary Engineering Research Laboratory, University of California, Berkeley. Ahring, B. K. and‘P. Westermann ( 1987) Kinetics of butyrate, acetate, and hydrogen metabolism in a thermophilic, anaerobic, butyrate-degrading triculture. Appl and Environ Microbiol 53: 434-439. Ahring, B. K. and P. Westermann (1988) Product inhibition of butyrate metabolism by acetate and hydrogen in a thermophilic coculture. Appl Environ Microbiol 54: 2393-2397. Aivasidis, A., K.-H. Bastin and C. Wandrey (1988) Optimization of selection stress in a chemostat, Proceedings of the Fifth International Symposium on Anaerobic Digestion, held on Bologna, Italy, 22-26 May, Pergamon Press, Oxford, P35-46. Alphenaar, P. A. (1993) Phosphorus requirement in high-rate anaerobic wastewater treatment, Was. Res. 27:749-756. American Public Health Association (1989) Standard methods for the examination of water and wastewater. 17th ed. (A. D. Eaton, L. S. Clesceri and A. E. Greenberg, ed.) American Public Health Association, Washington, DC. Anderson, G. K., T. Donnelly and K. J. McKeown (1982) Identification and control of inhibition in the anaerobic treatment of industrial wastewaters, Proc. Biochem. 13:28-32. Anderson, G. K. and G. Yang G (1992) Determination of bicarbonate and total volatile acid concentration in anaerobic digesters using a simple titration, Water Environ. Res. 64: 53-64. Anderson, G. K., B. Kasapgil and O. Ince (1994) Microbiological study of two-stage anaerobic digestion during start-up Was. Res. 28:2383-2392. Archer, P. B. and G. E. Powell (1985) Dependence of the specific growth rate of methanogenic mutualistic cocultures on the methanogens. Arch. Microbiol., 141:133-137. Atlas, R. M. and R. Bartha, Microbial Ecology, The Benjamin/Cummings Publishing Company, Inc. Menlo Park, CA 94025, 1987. Attal, A., F. Ehlinger, J. M. Audic and G. M. Faup (1988) pH inhibition mechanisms of acetogenic, actoclastic and hydrogenophilic populations. Proceedings of the Fifth Intemational Symposium on Anaerobic Digestion, Bologna, Italy, 22-26 May, Pergamon Press, Oxford, 71-78. 267 268 Bae J. and P. L. McCarty (1993) Variation of carbon monoxide production during methane fermentation of glucose, Wat. Environ. Res. 65:890-898. Barford, J. P. (1989) Start-up, Dynamics, and Control of Anaerobic Digesters, Handbook on Anaerobic fermentations, Larry E. Erickson and Daniel, Yee-Chak F. (ed), Marcel Dekker, Inc., New Work. Barnes T., Influence of organic shock loadings on the performance of an anaerobic fluidized bed system, paper presented at the 37th Purdue Industrial Waste Conference, Purdue, IN, April 1983. Benefield L. D. and C. W. Randall, Biological Process Design for Wastewater Treatment, Prentice-Hall, Inc., Englewood, NJ. 1980. Bitton, G., B. Koopman, K. Jung, G. Voiland and M. Kotob (1993) Modification of the standard epifluorescence microscopic method for total bacterial counts in environmental samples, Wat. Res. 27:1109-1112. Bobble, R. J. and D. C. White (1980) Characterization of benthic microbial community structure by high-resolution gas chromatography of fatty acid methyl esters, Appl. Environ. Microbiol. 39:1212-1222. Boe, B. and J. Gjerde (1980) Fatty acid pattern in the classification of some representatives of the families Enterobacteriaceae and Vibrionaceae , J. of Gen Microbiol 116:41-49. Boone, D. R. (1987) Effects of hydrogen pressure during growth and effects of pregrowth with hydrogen on acetate degradation by Methanosarcina species, Appl. Environ. Microbiol. 53: 83-87. Bousfield, I. J. (1983) Numerical analysis of total fatty acid profiles in the identification of coryneform, noncoryneform and some other bacteria, J. Gen Microbiol 129: 375-394. Bradford, M (1976) A rapid and sensitive method for the quantitation of microgram quantities of protein utilizing the principle of protein dye-binding. Anal. Biochem. 72:248- 253. Bratbak, G. (1985) Bacterial biovolume and biomass estimations, Appl Environ Microbiol 49: 1488-1493. Britz, T. J. (1994) Acidogenic microbial species diversity in anaerobic digesters treating different substrates, Wat. Sci. Tech. 30:55-61. Bryant, M. P. (1959) Bacterial species of the rumen. Bacteriol. Rev. 23: 39-66. Bryant, M. P., (1977) Microbiology of the rumen, in M. J. Swenson (ed) Dukes physiology of domestic animals, 9th revised edition. Comell University Press, Ithaca, N. Y. P287-304. Callander, I. J., and J. P. Barford (1983) Recent advances in anaerobic digestion technology. Proc. Biochem. 18:24-30. Chang I. T., T. Noike and J. Matsumoto (1982) Effect of retention time and feed substrate concentration of methanogenisis in anaerobic digestion. Proc. of JSCE No. 320, 67-76. 269 Chynoweth, D. P. and R. A. Mah (1977) Bacterial populations and end products during anaerobic sludge fermentation of glucose, J. WPCF 49: 405-412. Cohen, A., R. J. Zoetemeyer, A. ven Deursen and J. G. ven Andel (1979) Anaerobic digestion of glucose with separated acid production and methane formation. Wat. Res. 13: 371-379. Cohen, A., B. Distel, A. van Deursen and J. G. van Andel (1981) Role of anaerobic spore forming bacteria in the acidogenesis of glucose: changes induced by discontinuous or low-rate feed supply. Wat. Res. 15:59-68. Cohen, A., A. M. Breure, J. G. van Andel and A. van Deursen (1982) Influence of phase separation on the anaerobic digestion of glucose-II stability and kinetic responses to shock loadings. Wat. Res. 16:254-263. Cohen, T., Optimization of anaerobic digestion of soluble carbohydrate containing wastewater by phase separation. Ph. D dissertation, University van Amsterdam, Nieuwe Achtergracht 127, 1018 WS Amsterdam, The Netherlands, 1983. Cord R. (1988) The capacity of hydrogenotrophic anaerobic bacteria to compete for traces of hydrogen depends on the redox potential of the terminal electron acceptor, Arch. Microbiol. 149:350-357. Costello, D. J ., P. F. Greenfield and P. L. Lee (1991) Dynamic modeling of a single-stage high rate anaerobic reactor-I. model development, Wat. Res. 25:847-858. Dinopoulou, G., R. M. Sterritt and J. N. Lester (1988) Anaerobic acidogenesis of a complex wastewater: II. kinetics of growth, inhibition and product formation, Biotech Bioeng 31:969-978. Dolfing, J. and R. A. Prins (1996) Methanogenic ”Food Chains", American Society for Microbiology News, 62:117-118. Duarte, A. C. and G. K. Anderson (1982) Inhibition modeling in anaerobic digestion. Wat Sci Tech 14:749-763. Dwyer, D. F., E. Weeg-Aressens, D. R. Shelton and J. M. Tiedje (1988) Bioenergetic conditions of butyrate metabolism by a syntrophic, anaerobic bacterium in coculture with hydrogen-oxidizing methanogenic and sulfidogenic bacteria. Appl Environ Microbiol 54:1354-1359. Ehrlich, T., Bacteria in Nature, ed. J. S. Poindexter and E. R. Leadbetter, Plenum Press, New York, 1985. Fenchel, T., Ecology of Protozoa, Science Tech Publishers, Madison, Wisconsin, 1987. Fongastitkul, P., D. S. Mavinic and K. V. Lo (1994) A Two-phased anaerobic digestion process: concept, process failure and maximum system loading rate. Wat Environ Res J 66:243-254. Fox, P., and M. T. Suidan (1996) Shock and transient loading on anaerobic reactor coupled with adsorber, J. of Environ. Eng. 68:18-24. Fukuzaki, S. (1990) Inhibition of the fermentation of propionate to methane by hydrogen, acetate, and propionate, Appl. Environ. Microbiol. 48:719-723. 270 Ghosh, S. and F. G. Pohland (1974) Kinetics of substrate assimilation and product formation in anaerobic digestion, J. WPCF. 46:748-754. Gillan, F. T. and R. W. Hogg (1984) A method for the estimation of bacterial biomass and community structure in mangrove-associated sediments. J. of Microbiological Methods 2:275-293. Guckert, J. B. (1985) Phospholipid easter-linked fatty acid profiles as reproducible assays for changes in prokaryotic community structure of estuarine sediments. FEMS Microbiol. Ecol. 31:147-158. Guckert, J. B., M. A. Hood and D. C. White (1986) Phospholipid caster-linked fatty acid profile changes during nutrient deprivation of Vibrio cholerae: increases in the trans lcis ratio and proportions of cyclopropyl fatty acids, Appl Environ Microbiol 52:794-801. Gujer, W. and A. J. B. Zehnder (1983) Conversion processes in anaerobic digestion, Wat. Sci. Tech. 15:127-167. Gupta, A. J. R. V. Flora, M. Gupta, G. D. Sayles and M. T. Suidan (1994) Methanogenesis and sulfate reduction in chemostats-l. kinetic studies and experiments. Wat. Res. 28:781-793. Harper S. R. and F. G. Poland (1985) Recent developments in hydrogen management during anaerobic biological wastewate treatment, Biotechnol. Bioeng. 28:585-602. Hattingh, H. J. W., J. P. Kotze, P. G. Thiel, D. F. Toerien and M. L. Siebert (1967) Biological changes during the adaptation of an anaerobic digester to a synthetic substrate, Wat. Res. 1:255-277. Hedrick, D. B. (1991) Disturbance, starvation, and overfeeding stresses detected by microbial lipid biomarkers in high-solids high-yield methanogenic reactors. J. of Ind. Microbiol. 8:91-98. Heijnen, J. J ., M. C. M. van Loosdrecht and L. Tijhuis (1991) A black box mathematical model to calculate auto- and heterotrophic biomass yield based on Gibbs energy dissipation, Biotechnol. Bioeng. 40:1139-1154. Henson, J. M., P. H. Smith and D. C. Which (1985) Examination of thermophilic methane-producing digesters by analysis of bacteria lipids, Appl. Environ. Microbiol. 50:1428-1433. Herbert, D. (1975) Stoichiometric aspects of microbial growth: Continuous culture 6. Applications and New Fields (Edited by Dean A. C. R., Ellwood P. G, Evans C. G. T. & Meling J .). Ellis Horwood Ltd, Chichester, UK. Heyes, R. H. and R. J. Hall (1981) Anaerobic digestion modeling-the role of H2, Biotechnology Letters, 3:431-436. Heyes, R. H. and R. J. Hall (1983) Kinetics of two subgroups of propionate-using organisms in anaerobic digestion, Appl. Environ. Microbiol., 46: 710-715. Hickey, R., W. Wu and R. Jones (1991a) The start-up, operation and monitoring of high- rate anaerobic treatment systems, Water Sci. Tech. 24:207-231. 271 Hickey, R. F. and M. S. Switzenbaum (1991b) The response and utility of hydrogen and carbon monoxide as process indicators of anaerobic digesters subject to organic and hydraulic overloads. Res J WPCF 63:129-140. Hickey, R. F. and M. S. Switzenbaum (1991c) Thermodynamics of volatile fatty acid accumulation in anaerobic digesters subjected to increases in hydraulic and organic loading. Res J WPCF 63:141-144. Hill D. T. and C. L. Barth (1974) A dynamic model for simulation of animal waste digestion. J. WPCF 49:2129-2143. Hobson, P. N. and B. G. Shaw (1978) Inhibition of methane production by Methanobacteriumfarmicicum . Wat. Res. 10:849-852. Hungate, R. E. (1975) The rumen microbial ecosystem, Arch. Microbiol. 70: 39-66. Iannotti, E. L., (1982) Characterization of bacteria from a swine manure digester. Appl. Environ. Micriobiol. 43: 136-143. Ikemoto, S. (1978) Cellular fatty acid composition in Pseudomonas species, Gen. Appl. Microbiol. 24:199-213. Jain M. K., J. G. Zeikus and L. Bhatnagar, Methanogens, Anaerobic Microbiology (P. N. Levett, ed.) Oxfoed University Press, New York, 1991, pp223-245. Jeris, J. S. and P. L. McCarty (1962) The biochemistry of methane fermentation using C14 traces, Proceedings of 17th Industrial Waste Conference, Puedue University, P 181- 197. Kaneda, T. (1991) Iso- and Anteiso-fatty acids in bacteria: Biosynthesis, function, and taxonomic significance, Microbiol. Rev. 55:288-302. Kaspar, H. F. and K. Wuhrmann (1978) Kinetic parameters and relative turnovers of some important catabolic reactions in digesting sludge. Appl Environ Microbiol 36:1-7. Kisaalita, W. S., K. V. Lo and K. L. Pinder ( 1989) Influence of dilution rate on the acidogenic phase products distribution during two-phase lactose anaerobiosis, Biotech. Bioeng. 34:1235-1250. Knivett, V. A. and J. Cullen (1965) Some factors affecting cyclopropyl acid formation in Escherichia Cali. Biochem. J. 96:771-776. Kroeker, E. J. (1979) Anaerobic treatment process stability, J. WPCF. 51:718-727. Krumholz, L. R. and C. W. Forsberg (1983) Association of methanogenic bacteria with rumen protozoa, Can. J. Microbiol. 29:676-680. Labib, F., J. F. Ferguson, M. M. Benjamin, M. Merigh and N. L. Ricker (1993) Mathematical-modeling of an anaerobic butyrate degrading consortia-predicting their response to organic overloads, Environ Sci Tech 27:2673-2684. Lawrence, A. W. and P. L. McCarty (1969) Kinetics of methane fermentation in anaerobic treatment. J. WPCF 41: Res. Suppl. R1-Rl7. 272 Lawrence, A. W. and P. L. McCarty (1970) United basis for biological treatment design and operation, J. Sanitary Engineering Division, Proceedings of the American Society of Civil Engineers, 34:757-778. Lepage, C. (1987) Changes in membrane lipid composition of Clostridium acetobutylicum during acetone-butanol fermentation: effects of solvents, growth temperature, and pH. J. Gen. Microbiol. 133: 103-110. Iettinga, G., S. W. Hobma, L. W. Hulshoff Pol, de Zeeuw, W., de Jong, P., Grin, p., and R. Roersma (1983) Design operation and economy of anaerobic treatment. Wat. Sci. Tech. 15: 177-195. Levenspiel, 0. Chemical reaction engineering, John Wiley & Sons, New Tork, 1972. Li, Y. Y., H. H. Fang and T. Noike (1994) Recent developments in microbiology and biochemistry of homoacetogens and sytrophic acetogens. Proc. Posters of the 7th International Symposium on Anaerobic Digestion. 23-27 Jauary, 1994, Cape Town, South Africa, pp. 54-57. Linden, C. James, Industrially important strains and pathways, in Handbook on Anaerobic Fermentations, ed. Erickson, E. Larry and Daniel Y-C Fung, Marcel Dekker, Inc., New York, 1989. ‘ Lovely, D. R. (1985) Minimum threshold for hydrogenmetabolism inmethanogenic bacteria, Appl. Environ. Microbiol. 56: 1598-1607. Maestrojuan, G. M. and D. R. Boone (1991) Characterization of Methanosarcina barkeri MST and 227, Methanosarcina mazei S-6T, and Methanosarcina vacuolalata Z-761T. Int. J, Syst. Bacteriol. 41:267-274. Massol-Deya, A., D. Odelson, R. Hickey and J. Tiedje (1995) Bacterial community fingerprinting of amplified 168 and 16-238 ribosomal DNA gene sequences and restriction endonuclease analysis (ARDRA). In Molecular Microbial Ecology Manual (A. D. L. Akkerrnans, J. D. van Elsas, and F. J. deBruijn, eds) . Kluwer Academic Publishers, Netherlands. pp. 3.3.2:1-8. Mawson, A. J ., R. L. Earle and V. F. Larsen (1991) Degradation of acetic and propionic acids in the methane fermentation, Wat. Res. 25: 1549-1554. McCarty, P. L (1964) Anaerobic waste treatment fundamentals-part one, Public Works. 95: 107-112. McCarty, P. L (1964) Anaerobic waste treatment fundamentals-part two. Public Works. 95: 123-126. McCarty, P. L (1965) Thermodynamics of biological synthesis and growth, Second International Conference on Water Pollution Research, Permagon Press, New York, P169- 199. McCarty, P. L (1966), Kinetics of waste assimilation in anaerobic treatment, In Developments in Industrial Microbiology, Vol. 7, American Institute of Biological Sciences, Washington DC. 273 McCarty, P. L (1981) One hundred years of anaerobic treatment. Proceedings of the Second International Symposium on Anaerobic Digestion, held on Travemunde, Federal Republic of Germany, Elsevier Biomedical Press, Amsterdam, P3-22. McCarty, P. L., One hundred years of anaerobic treatment, in Anaerobic digestion 1981, ed. D. E. Hughes et al., Elsevier Biomedical Press, New York, 3, 1983. McCarty P. L. and D. Smith, Effect of hydrogen concentration on population distribution and kinetics in methanogenesis of propionate, Paper presented at the First Symposium on Biotechnological Advances in Processing Municipal Wastes for Fuels and Chemicals, August 15-17, 1984, Minneapolis, Minnesota. McCarty, P. L. and F. E. Mosey (1991) Modeling of anaerobic digestion process (a discussion of concepts). Wat Sci Tech 24:17-33. McInemey, M. J., M. P. Bryant, and D. A. Stafford (1979) Anaerobic bacterium that degrade fatty acids in syntrophic association with methanogens. Arch. Microbiol. 122:129-135. Meakin, S. A., S. A., J. H. E. Hash, W. D. Murry, K. J. Kennedy and G. D. Sportt (1991) A generally applicable technique for the extraction of restrictable DNA from methanodenic bacteria, J Microbiol Meth 14: 119-126. Moletta, R., D. Verrier and G. Albagnac (1986) Dynamic modeling of anaerobic digestion, Wat. Res. 20:427-434. Mosey, F. E (1981) Anaerobic biological treatment of food industry waste water. Wat Pollut Control 80: 273-289. Mosey, F. E (1983a) Mathematical modeling of the anaerobic digestion process: regulatory mechanisms for the formation of short-chain volatile acids from glucose. Wat Sci Tech 15:209-232. Mosey, F. E (1983b) Kinetic descriptions of anaerobic digestion, Proceedings of the third international symposium on anaerobic digestion, Boston, MA. pp37-44. Mosey, F. E. and X. A. Fernandez (1989) Patterns of hydrogen in biogas from the anaerobic digestion of milk-sugars, Wat. Sci. Tech. 21:187-196. Novaes, R. F. V (1986) Microbiology of anaerobic digestion. Wat Sci Tech 18:1-14. Novak, J. T. and D. A. Carlson (1970) The kinetics of anaerobic long chain fatty acids degradation, J. WPCF 42: 1932-1943. Odelson, D. A. and J. A. Breznak (1985) Nutrition and growth characteristics of Trichomitopsis termopsis, a cellulolytic protozoa from termites, Appl. Environ. Microbiol. 49:614-621. O'Rourke, J. T. (1968) Kinetics of anaerobic treatment at reduced temperatures, Ph. D. Dissertation, Stanford University. Palns, Sam-soon (1987) Hypothesis for pelletization in the upflow Anaerobic sludge bed reactor. Water SA, 13:69-80. 274 Pavlostathis, S. G. and E. Giraldo-Gomez (1991) Kinetics of anaerobic treatment. Wat Sci Tech 24:35-59. Peck, M. W., J. M. Skilton, F. R. Hawkes and D. L. Hawkes (1986) Effects of temperature shock treatments on the stability of anaerobic digesters operated on separated cattle slurry, Was. Res. 20:453-462. Rajendran, N. (1992) Variation in Microbiol biomass and community structure in sediments of eutrophic bays as determined by phospholipid ester-linked fatty acids. Appl. Environ. Microbiol. 58:562-571. Ross, R. W. and M. Smollen (1981) Engineering problems with the anaerobic digestion of soluble organic wastes. Proceedings of the Second International Symposium on Anaerobic Digestion, held on Travemunde, Federal Republic of Germany, Elsevier Biomedical Press, Amsterdam, P89-106. Rozzi, A., S. Merlinio and R. Passino (1985) Development of a four p0pulation model of the anaerobic degradation of carbohydrates, Environ. Technol. Letters 62610-619. Schirndt, E. L. and E. A. Paul, Microscopic methods for soil microorganisms. P. 803-814. A. Page (ed.), Methods of soil analysis, Part 2. Chemical and microbiological properties. ASA-SSA Inc., Madison, WI. 1982. Sedmak, J. J. and S. E. Grossberg (1977) A rapid sensitive, and versatile assay for protein using Coomassie brilliant blue G250. Anal. Biochem. 79:544-552. Shea, T. G., W. A. Pretorius, R. D. Cole and E. A. Pearson (1968) Kinetics of hydrogen assimilation in the methane fermentation, Wat. Res. 2:833-848. Smith, M. R. and R. A. Mah. (1978) Growth and methanogenesis by methanosarcina strain 227 on acetate and methanol, Appl. Environ. Microbiol. 44: 870-879. Smith, P. Daniel, Hydrogenotrophic control in methanogenic processes, Ph D dissertation, Stanford University, 1987. Smith, P. D. and P. L. McCarty (1989) Energetics and rate effects on methanogenesis of ethanol and propionate in perturbed CSTR's, Biotechnol. Bioeng. 34:39-54. Smith, P. D. and P. L. McCarty (1990) Factors governing methane fluctuations following shock loading of digesters, J. WPCF. 62:58-64. Speece, R. E. & P. L. McCarty (1964) Nutrient requirements and biological solids accumulation in anaerobic digestion, In "Advances in Water Pollution Research", Pergamon Press, Oxford, 11, 305-322. Steven J. 8., J. P. Tommy, T. M. Al and D. C. White (1986) The relationship of eubacterial and methanogic community structure in anaerobic digesters, Procedings of 10th annual conference on energy from biomass and waste, Washington D. C. April 7-10, 1986. Steward, M. J. (1958) Reaction kinetics and operation parameters of continuous-flow anaerobic-fermentation processes, I. E. R. Series 90, Report No. 4, Sanitary Engineering Research Laboratory, University of Califomia, Berkeley. 275 Thauer, R. K., K. Jungerrnann and K. Decker (1977) Energy conservation in chemotrophic anaerobic bacteria. Bacteriol. Rev. 41: 100-180. Thiele, J. H. and J. G. Zeikus (1988a) Control of interspecies electron flow during anaerobic digestion: the role of formate versus hydrogen transfers during syntrophic methanogenesis in flocs. Appl. Environ. Microbiol. 54:20-29. Thiele, J. H. and J. G. Zeikus (1988b) Interactions between hydrogen- and forrnate- producing anaerobic bacteria and methanogens during anaerobic digestion. In Handbook on Anaerobic Fermentations, ed. L. E. Erickson and D. Y.-C. Fung. Dekker, Amstterdam, pp 537-595. Tiedje, J. S. Sirnkins and S. Groffman (1995) Redefining the concept and scope of detrital food chains, unpublished paper. Tijhuis, L., M. C. M. van Loosdrecht and J. J. Heijnen (1991) A thermodynamically based correlation for maintenance Gibbs energy requirements in aerobic and anaerobic chemotrophic growth, Biotechnol. Bioeng. 42:509-519. Toerien, D. F. (1966) Enrichment culture studies on aerobic and facultative anaerobic bacteria found in anaerobic digesters, Wat. Res. 1: 147-155. Toerien, D. F. and W. H. J. Hattingh (1969) Anaerobic digestion I. the microbiology of anaerobic digestion, Was. Res. 3:385-416. Tsai, Y.-1. and B. H. Olson (1991) Rapid method for direct extraction of DNA from soil and sediments. Appl Environ Microbi0157: 1017-1024. Valcke, D and W. Verstraete (1983) A practical method to estimate the acetoclastic methanogenic biomass in anaerobic sludges. J WPCF 55:1191-1195. van Bruggen, J. J. (1984) Methanobacterium forrnicicum, an endosymbiont of the anaerobic ciliate Metopus striatus McMurrich. Arch Microbiol. 139:1-7. van den Berg, L., G. B. Patel, D. S. Clark and C. P. Lentz (1976) Factors affecting rate of methane formation from acetic acid by enriched methanogenic culture. Can. J Microbiol. 22: 1312-1319. Varel, V. H. (1984) Characteristics of some fermentative bacteria from a thermophilic methane-producing fermenter. Microb. Ecol. 10:15-24. Vestal, J. R. and D. C. White (1989) Lipid analysis in microbial ecology. BioScience 39:535-541. Visser, A. Y. Gao and G. Lettiga (1993) Effects of short-term temperature increases on the mesophilic anaerobic breakdown of sulfate containing synthetic wastewater, Was. Res. 27:541-550. Vogels, G. D. (1980) Association of methanogenic bacteria with rumen Ciliates. Appl. and Envir. Micriol. 40:608-612. Weast, R. C. (ed.) CRC Handbook of Chemstry and Physics, 51th ed. CRC Press, Inc. Boca Raton, FL, 1979. 276 White, D. C., Analysis of microorganisms in terms of quantity and activity in natural environments, in Slater J. H. et al., (ed.), Microbes in their natural environments. p. 37-66. Cambridge University Press, New York, 1979. White, D. C. (1983) Lipid analysis of sediments for microbial biomass and community structure. ASTM Spec. Tech. Pub]. No. 673. P. 87-103. Williams, A. G. and C. G. Harfoot (1976) Factors affecting the uptake and metabolism of soluble carbohydrates by the rumen Ciliate Dasytricha ruminantium isolated from ovine rumen contents by filtration. J. Gen. Microbiol. 96:125-136. Wu, W., M. K. Jian, J. H. Thiele and J. G. Zeikus (1995) Effect of storage on performance of methanogenic granules, Wat. Res. 29:1445-1452. Yong S, T. and M. R. Okos (1987) Kinetic study and mathematical modeling of methanogenesis of acetate using pure cultures of methanogens, Biotechnol. Bioeng. 30:661-667. Zehnder, J.T., C. Grotenhuis and L. W. Hulshoff Pol (ed), Granular anaerobic sludge; microbiology and Technology. Prodoc, Wagenegen, the Netherlands. Zehnder, A. J. B., K. Ingvorsen and T. Marti (1981) Microbiology of methane bacteria. In Hughes, D. E., D. A. Stanford, B. I. Wheatley, W. Baader, G. Lettinga, E. J. Nyns, W. Versraete, and R. L. Wentworth (ed). Anaerobic digestion 1981. Elsevier Biolmedical Press, Amsterdam. P45-68. Zeikus, J. G. (1977) The biology of methanogenic bacteria, Bacteriol. Rev. 41: 514-541. Zeikus, J. G., Microbial population in digesters, in Anaerobic Digestion, Stafford D. A., B. I. Wheatley and D. E. Hughes ed, Applied Science Publishers LTD, Essex, England, 1979. Zelles, L (1992) Signature fatty acids in phospholipids and lipopolysaccharides as indicators of microbial biomass and community structure in agricultural soils. Soil. Biol. Biochm. 24:317-323. Zhang, T. C. and T. Noike (1994) Influence of retention time on reactor performance and bacterial trophic populations in anaerobic digestion processes, Wat. Res. 28:27-36. Zinder, S. H. (1990) Conversion of acetate acid to methane by therrnophiles, FEMS Microbiology Reviews 75:125-138. Zoetemeyer, R. J., J. C. ven den Heuvel and A. Cohen (1982a) pH Influence on acidogenic dissirnilation of glucose in anaerobic digester. Wat. Res. 16:303-311. Zoetemeyer, R. J., A. J. C. M. Matthusen, A. Cohen and C. Boelhouwer (1982b) Production inhibition in the acid forming stage of the anaerobic digestion process. Wat. Res. 16:633-639. Zhou J.-Z., Fries, M. R., Chee-Sanford, J. and J. M. Tiegje (1995) Phylogenetic analysis of a new group of denitrifiers capable of anaerobic growth on toluene: description of Azoarcus toluyricus sp. nov., Int. J. Syst. Bacteriol. 45:500-506.