A N A E R O B I C D I G E S T I O N O F C H E E S E W H E Y I N A N U P F L O W A N A E R O B I C S L U D G E B L A N K E T R E A C T O R By Jing-Qing Yan B. Eng. South-Central University of Technology, China M. Sc. Dalian Institute of Physical Chemistry, Chinese Academy of Sciences A T H E S I S S U B M I T T E D I N P A R T I A L F U L F I L L M E N T O F T H E R E Q U I R E M E N T S F O R T H E D E G R E E O F D O C T O R O F P H I L O S O P H Y in T H E F A C U L T Y O F G R A D U A T E S T U D I E S C H E M I C A L E N G I N E E R I N G We accept this thesis as conforming to the required standard T H E U N I V E R S I T Y O F B R I T I S H C O L U M B I A March 1991 © Jing-Qing Yan, 1991 In presenting this thesis in partial fulfilment of the requirements for an advanced degree at the University of British Columbia, I agree that the Library shall make it freely available for reference and study. I further agree that permission for extensive copying of this thesis for scholarly purposes may be granted by the head of my department or by his or her representatives. It is understood that copying or publication of this thesis for financial gain shall not be allowed without my written permission. Chemical Engineering The University of British Columbia 2075 Wesbrook Place Vancouver, Canada V6T 1W5 Date: Dedication To my parents for their love, encouragement and sacrifice. 11 A B S T R A C T The anaerobic digestion of cheese whey was studied in an upfiow anaerobic sludge blanket reactor for its start-up characteristics, the effects of various process parameters, the effect of sulfate addition and the determination of optimal operating conditions. Start-up of an U A S B reactor treating cheese whey was extremely difficult due to its tendency to acidify. Various start-up strategies were tested to facilitate start-up and to ensure stable operation. Among the operating parameters, sludge loading rate was the most critical for proper start-up of the U A S B reactor. The initial sludge loading rate during start-up period should not exceed 0.25 g C O D / g VSS . The response of whey digestion to several process parameters was investigated. W i t h -out pH-control, over 97% C O D removal was obtained for influent concentrations from 5 to 28.8 g C O D / 1 and H R T of 5 days. However, instability was observed when the influent concentration was increased to 38.1 g C O D / 1 . Gas production from whey is affected by organic loading rate ( O L R ) . A t an O L R less than 4 g C O D / l - d , higher influent strength resulted i n a higher methane production rate. When the O L R was greater than 6, higher strength feed or shorter hydraulic retention time ( H R T ) produced less methane. From the profiles of substrate concentration measured at various levels above the bottom of the reactor, two reaction stages, acidogenesis and methanogenesis were dis-tinguished. It was experimentally illustrated that the rate of acidogenesis is much faster than the rate of methanogenesis in a whey anaerobic digestion system. The accumu-lation of V F A s in the first stage being faster than its assimilation in the second stage creates a distinct acidogenic phase in the bottom of the reactor. The instability caused i i i by high influent concentration could be attributed to the accumulation of VFAs beyond the assimilative capacity of the methanogenic stage. A set of empirical models for accumulation and degradation of VFAs was developed using linear regression analysis. The requirement for maintaining this system in a dy-namic balance was that the degradation capacity for VFA in the second stage be greater than the accumulation of VFA in the first stage. Based on this idea, the optimal influent concentration was given as between 25 to 30 g COD/1 for system stability. A hypothesis was proposed in this study that a proper amount of sulfate may be applied to moderate the detrimental influence of excess hydrogen on a stressed anaerobic reactor. The effect of sulfate was tested to study the biochemical mechanism. The permissible influent COD concentration was increased from 30 g COD/1 to 50 g COD/1 by using sulfate addition. The pH in the reactor was on the average 0.8 units higher and the concentration of butyric acid in the acidogenic phase much lower with added sulfate than without sulfate addition. The significant improvement of process stability and treatment efficiency made by the addition of sulfate clearly illustrated that sulfate acted like a stimulator which helped to maintain conditions favorable to methanogenesis. The mechanism of this stimulation is explained according to thermodynamics and hydrogen regulation which suggested that sulfate is able to promote the /3-oxidation of VFAs by consuming hydrogen. A two-stage inhibition mechanism was proposed to explain the inhibition of high VFA concentrations and the stimulation of sulfate. Higher hydrogen pressure is the cause of preliminary inhibition, resulting in the accumulation of VFAs, which subsequently inhibit the activity and growth of methanogens in the second inhibition stage. The mechanism of inhibition of methanogens from VFAs was interpreted as being caused by the acidification of the internal cytoplasm and destruction of the pH gradient by non-ionized acids based on the theory of bacterial membrane transport. A new control strategy for stabilization iv of an anaerobic system is recommended. Under the optimal operating conditions based on the results in the first three steps, over 97% reduction of C O D was achieved when the influent C O D was 30 g / l using an H R T of 2 days, an O L R of 16.61 g C O D / l - d and sulfate concentration of 0.2 g/1. v Table of Contents Dedication ii A B S T R A C T iii List of Tables xi List of Figures x v List of Abbreviations xix Acknowledgement x x 1 I N T R O D U C T I O N 1 1.1 DISPOSAL/UTILIZATION OF CHEESE WHEY 1 1.2 ANAEROBIC METHANE FERMENTATION OF WHEY 2 1.3 PROBLEMS ASSOCIATED WITH THE ANAEROBIC DIGESTION OF CHEESE WHEY 5 1.3.1 Inadequate Buffering Capacity 5 1.3.2 Micronutrient Deficiency 6 2 L I T E R A T U R E R E V I E W 8 2.1 DEVELOPMENT OF REACTOR DESIGN TECHNOLOGY/UASB RE-ACTOR 8 2.2 ANAEROBIC TREATMENT OF CHEESE WHEY IN DIFFERENT RE-ACTOR CONFIGURATIONS 13 vi 2.3 E N V I R O N M E N T A L F A C T O R S IN A N A E R O B I C D I G E S T I O N 14 2.3.1 p H Control in the Anaerobic Process 14 2.3.2 Nutrient Requirement 16 2.3.3 Sulfate Effect 18 2.4 M I C R O B I O L O G Y O F A N A E R O B I C D I G E S T I O N 25 2.4.1 Anaerobic Microorganisms 25 2.4.2 Microbial Interaction in Anaerobic Digestion 30 2.4.3 Hydrogen Regulations 36 3 R E S E A R C H O B J E C T I V E S 40 4 E X P E R I M E N T A L M E T H O D 43 4.1 R E A C T O R S E T - U P 43 4.2 F E E D S U B S T R A T E 45 4.3 S E E D S L U D G E 47 4.4 E X P E R I M E N T D E S I G N 48 4.5 R E A C T O R O P E R A T I O N 50 4.6 A N A L Y S I S 51 5 S T A R T - U P A N D E F F E C T S OF PROCESS P A R A M E T E R S 52 5.1 I N T R O D U C T I O N 52 5.2 S T U D Y IN S T A R T - U P O F T H E R E A C T O R 53 5.2.1 Difficulties in Start-up 53 5.2.2 Start-up Performance 55 5.2.3 V F A Behavior in Reactor Start-up 66 5.3 S T E A D Y S T A T E O P E R A T I O N 67 5.4 U N S T E A D Y - S T A T E P E R F O R M A N C E 70 vii 5.5 G A S P R O D U C T I O N 70 5.5.1 The Effect of Influent Concentration 70 5.5.2 The Effect of Hydraulic Retention Time ( H R T ) 76 5.5.3 The Effect of Organic Loading Rate ( O L R ) 81 5.6 T R E A T M E N T E F F I C I E N C Y 85 5.7 S U M M A R Y 88 6 DISTRIBUTIONS OF S L U D G E A N D S U B S T R A T E S 90 6.1 I N T R O D U C T I O N 90 6.2 R E S U L T S 91 6.2.1 Distribution and Behaviour of the Sludge 91 6.2.2 Growth of Sludge 100 6.2.3 Profiles of C O D , V F A and p H 101 6.3 S U M M A R Y 112 7 E F F E C T S OF S U L F A T E O N A N A E R O B I C D I G E S T I O N O F W H E Y 116 7.1 G E N E R A L R E M A R K S 116 7.2 H Y P O T H E S I S 117 7.3 R E S U L T S IN B A T C H E X P E R I M E N T S 121 7.4 R E S U L T S IN C O N T I N U O U S E X P E R I M E N T S 122 7.4.1 Effect of Sulfate on Reactor Operation: T ime Profile 122 7.4.2 Effect of Sulfate on Steady State Performance 134 7.4.3 The Improvement of Reactor Stability 139 7.4.4 . The Effect of Sulfate on Buffer Capacity 141 7.4.5 The Improvement of Treatment Efficiency 141 7.4.6 The Effect of Sulfate on Profiles of p H , Sludge and V F A s 141 viii 7.5 M E C H A N I S M O F I N H I B I T I O N O F H I G H C O N C E N T R A T I O N O F V F A s A N D L O W P H 154 7.6 M E C H A N I S M O F S T I M U L A T I O N B Y S U L F A T E 158 7.7 S U M M A R Y 168 8 T R E A T M E N T E F F I C I E N C Y IN O P T I M A L O P E R A T I O N 170 9 C O N C L U S I O N S , C O N T R I B U T I O N S A N D R E C O M M E N D A T I O N S 174 9.1 C O N T R I B U T I O N S A N D C O N C L U S I O N S 174 9.2 R E C O M M E N D A T I O N S F O R F U T U R E R E S E A R C H 178 Bibliography 180 Appendices 200 A E M P I R I C A L M O D E L 200 A . l T H E P R O G R A M O F M I N I T A B 200 A.2 A N A L Y S I S O F T H E A C C U M U L A T I O N O F A C E T I C A C I D 201 A.2.1 Determination of the Equation 201 A.2.2 Analysis of the Adequacy of the Equation 203 A.2.3 The Comparison between the Experimental Data and the Model . 210 A . 3 S U M M A R Y O F E M P I R I C A L M O D E L S 210 A.3.1 In the Absence of Sulfate 210 A.3.2 In the Presence of Sulfate 219 B E F F E C T S OF S U L F A T E IN B A T C H E X P E R I M E N T S 221 B . l I N T R O D U C T I O N 221 B.2 P R E L I M I N A R Y E X P E R I M E N T A L D E S I G N 222 B.2.1 Variables 222 B.2.2 Experimental Design 224 B.3 EXECUTION OF T H E EXPERIMENT 225 B.3.1 Substrate 225 B.3.2 Seed Sludge (Bacteria) 225 B.3.3 Reactor Operation 225 B.3.4 Chemical Analyses 226 B.4 RESULTS AND DISCUSSION 226 B.4.1 Effects of Sulfate (S), Feed Strength (Co) and Seed Sludge Amount (A) on the Methane Production (Cif 4%) 2 2 6 B.4.2 Effects of S, Co and A on pH of the System 234 B.5 CONCLUSIONS 240 B. 6 CRITIQUE OF T H E EXPERIMENT AND F U T U R E P L A N 240 C T H E C A L C U L A T I O N OF NON-IONIZED ACID 243 C l T H E O R Y 243 C. 2 CALCULATION OF NON-IONIZED ACID 244 D STATISTICAL ANALYSIS OF pH 245 E D A T A TABLES 250 x List of Tables 2.1 Kinetic Characteristics of Sulfate Reducers and Methanogens 23 2.2 Stoichiometry and Change in Free Energy of Reaction 28 2.3 Methanogenic Substrates and Reactions 31 2.4 Partial Reaction for a Hypothetical Glucose Fermentation W i t h -out and W i t h a Methanogen 34 4.1 Characteristics of Cheddar Cheese Whey 46 4.2 Experimental Design 49 5.1 Sludge Loading in Start-up 54 5.2 Experimental Results in Steady State Performance 68 5.3 C O D Balance 71 5.4 Gas Production Rate and Composition for Different Influent Concentrations at 5 days H R T 73 5.5 Methane Production Rate at Influent of 20.5 g C O D / 1 78 5.6 Methane Production Rate at Influent of 41.1 g C O D / 1 79 5.7 V F A in the Reactor and Effluent at Different H R T s 82 5.8 Effects of Influent Concentration and H R T on Gas Production Rate for Similar O L R 84 6.1 Distribution of Sludge and Substrates 93 6.2 Relation between Sludge Distribution and Biogas 99 6.3 Sludge in the U A S B Reactor 102 6.4 A c e t i c A c i d ( A A ) 114 7.1 Free E n e r g y C h a n g e s of R e a c t i o n s I n v o l v e d i n M e t a b o l i s m of S o m e O r g a n i c M a t t e r 120 7.2 E x p e r i m e n t a l Resul ts i n S t e a d y State 135 7.3 Sludge in the R e a c t o r 140 7.4 Profiles of the U A S B R e a c t o r w i t h Sulfate A d d i t i o n 150 7.5 C a l c u l a t e d C o n c e n t r a t i o n of N o n - i o n i z e d V F A s 155 8.1 Results in the O p t i m a l O p e r a t i o n 172 8.2 C o m p a r i s o n of A n a e r o b i c T r e a t m e n t Processes for Cheese W h e y 173 A . l A c c u m u l a t i o n of A c e t i c A c i d 202 A.2 Stepwise A n a l y s i s for A c c u m u l a t i o n of A c e t i c A c i d 203 A.3 Stepwise A n a l y s i s for A c c u m u l a t i o n of A c e t i c A c i d 204 A.4 Stepwise A n a l y s i s for A c c u m u l a t i o n of A c e t i c A c i d 205 A.5 Stepwise A n a l y s i s for A c c u m u l a t i o n of A c e t i c A c i d 206 A.6 Stepwise A n a l y s i s for A c c u m u l a t i o n of A c e t i c A c i d 207 A.7 Stepwise A n a l y s i s for A c c u m u l a t i o n of A c e t i c A c i d 208 A.8 Stepwise A n a l y s i s for A c c u m u l a t i o n of A c e t i c A c i d 209 A. 9 Regression A n a l y s i s for D e g r a d a t i o n of A c e t i c A c i d 217 B. l Independent V a r i a b l e s i n B a t c h E x p e r i m e n t 222 B.2 E x p e r i m e n t a l D e s i g n of B a t c h R u n 223 B.3 E x p e r i m e n t a l Resul ts 227 B.4 Yates A n a l y s i s 229 B.5 A N O V A T a b l e . 233 B.6 95% C o n f i d e n c e Interval 235 xii B.7 E x p e r i m e n t a l D e s i g n B a s e d on Steepest A s c e n t 241 D.l C a l c u l a t i o n of S t a n d a r d Deviat ions 247 D.2 t-test 248 D. 3 t-test 249 E. l S l u d g e G r o w t h 251 E.2 T h e P r o p i o n i c A c i d ( P A ) 252 E.3 T h e T o t a l Volat i le Fa t ty A c i d ( T V F A ) 253 E.4 A n a l y s i s for A c c u m u l a t i o n and D e g r a d a t i o n of V F A 254 E.5 A n a l y s i s for A c c u m u l a t i o n and D e g r a d a t i o n of V F A 255 E.6 A n a l y s i s for A c c u m u l a t i o n and D e g r a d a t i o n of V F A 256 E.7 A n a l y s i s for A c c u m u l a t i o n and D e g r a d a t i o n of V F A 257 E.8 A n a l y s i s for A c c u m u l a t i o n a n d D e g r a d a t i o n of V F A 258 E.9 A n a l y s i s for A c c u m u l a t i o n and D e g r a d a t i o n of V F A 259 E.10 A n a l y s i s for A c c u m u l a t i o n and D e g r a d a t i o n of V F A 260 E . l l A n a l y s i s for A c c u m u l a t i o n and D e g r a d a t i o n of V F A 261 E.12 A n a l y s i s for A c c u m u l a t i o n and D e g r a d a t i o n of V F A 262 E.13 A n a l y s i s for A c c u m u l a t i o n a n d D e g r a d a t i o n of V F A 263 E.1'4 A n a l y s i s for A c c u m u l a t i o n and D e g r a d a t i o n of V F A 264 E.15 A n a l y s i s for A c c u m u l a t i o n a n d D e g r a d a t i o n of V F A 265 E.16 A n a l y s i s for A c c u m u l a t i o n and D e g r a d a t i o n of V F A 266 E.17 A n a l y s i s for A c c u m u l a t i o n and D e g r a d a t i o n of V F A 267 E.18 A n a l y s i s for A c c u m u l a t i o n and D e g r a d a t i o n of V F A 268 E.19 A n a l y s i s for A c c u m u l a t i o n and D e g r a d a t i o n of V F A 269 E.20 A n a l y s i s for A c c u m u l a t i o n and D e g r a d a t i o n of V F A 270 E.21 A n a l y s i s for A c c u m u l a t i o n and D e g r a d a t i o n of V F A 271 xiii E.22 A n a l y s i s for A c c u m u l a t i o n a n d D e g r a d a t i o n of V F A 272 E.23 A n a l y s i s for A c c u m u l a t i o n a n d D e g r a d a t i o n of V F A 273 E.24 A n a l y s i s for A c c u m u l a t i o n a n d D e g r a d a t i o n of V F A 274 xiv List of F igures 2.1 M e t a b o l i c D i s t i n c t i o n of the M i c r o b i a l P o p u l a t i o n i n A n a e r o b i c Diges t ion 27 2.2 M e t a b o l i c P a t h w a y Inside A c i d - f o r m i n g B a c t e r i a 33 2.3 G r a p h i c a l Representat ion of the H y d r o g e n - d e p e n d e n t T h e r m o -d y n a m i c Favorabi l i ty of Acetogenic O x i d a t i o n s a n d Inorgani c Respirat ions Associa ted w i t h the A n a e r o b i c D e g r a d a t i o n of W a s t e O r ganics 39 4.1 Schematic D i a g r a m of U A S B S y s t e m 44 5.1 O r g a n i c L o a d i n g Rate versus T i m e 57 5.2 Influent C O D versus T i m e 58 5.3 Eff luent C O D versus T i m e 59 5.4 C O D R e m o v a l versus T i m e 60 5.5 M e t h a n e P r o d u c t i o n R a t e versus T i m e 61 5.6 Biogas C o m p o s i t i o n versus T i m e 62 5.7 Eff luent Volat i le Fa t ty A c i d s versus T i m e 63 5.8 Vola t i le S u s p e n d e d Sol id ( V S S ) i n the Eff luent 64 5.9 C o m p a r i s o n of C O D , V F A a n d Biogas P r o d u c t i o n 65 5.10 Q u a l i t y of Effluent as a F u n c t i o n of Influent C o n c e n t r a t i o n (at H R T = 5 days) 69 5.11 Effect of Influent C o n c e n t r a t i o n on M e t h a n e P r o d u c t i o n 74 5.12 Effect of Influent C o n c e n t r a t i o n on Biogas C o m p o s i t i o n 75 xv 5 . 1 3 Effect of H R T on M e t h a n e P r o d u c t i o n 8 0 5 . 1 4 Effect of H R T on M e t h a n e C o m p o s i t i o n 8 3 5 . 1 5 Effect of O r g a n i c L o a d i n g R a t e o n M e t h a n e P r o d u c t i o n 8 6 5 . 1 6 R e l a t i o n Between M e t h a n e P r o d u c t i o n and O L R 8 7 6 . 1 Profiles of Sludge at Different Influent Concentra t ions 9 4 6 . 2 S ludge D i s t r i b u t i o n in the U A S B Reac tor 9 6 6 . 3 S ludge C o n c e n t r a t i o n in the U A S B R e a c t o r 9 7 6 . 4 S ludge D i s t r i b i t i o n versus G a s P r o d u c t i o n 9 8 6 . 5 S ludge G r o w t h 1 0 3 6 . 6 p H Profi le 1 0 4 6 . 7 Prof i le of A c e t i c A c i d 1 0 5 6 . 8 Prof i le of P r o p i o n i c A c i d 1 0 6 6 . 9 Prof i le of C O D 1 0 7 6 . 1 0 Prof i le of C O D R e d u c t i o n 1 0 8 6 . 1 1 Profi les of p H a n d V F A s 1 1 3 6 . 1 2 A c c u m u l a t i o n a n d D e g r a d a t i o n of A c e t i c A c i d 1 1 5 7 . 1 M i c r o b i a l E c o l o g y for the A n a e r o b i c Diges t ion Process 1 1 9 7 . 2 C O D C o n c e n t r a t i o n versus T i m e 1 2 3 7 . 3 O r g a n i c L o a d i n g R a t e versus T i m e 1 2 4 7 . 4 Eff luent C O D versus T i m e 1 2 5 7 . 5 C O D R e m o v a l versus T i m e 1 2 6 7 . 6 Effluent V o l a t i l e F a t t y A c i d V e r s u s T i m e 1 2 7 7 . 7 Eff luent p H versus T i m e 1 2 8 7 . 8 M e t h a n e P r o d u c t i o n R a t e versus T i m e 1 2 9 7 . 9 Biogas C o m p o s i t i o n versus T i m e 1 3 0 xvi 7.10 V S S in the Effluent versus T i m e 131 7.11 Effect of Influent C o n c e n t r a t i o n on M e t h a n e P r o d u c t i o n a n d C o m p o s i t i o n 136 7.12 Effect of O L R on M e t h a n e P r o d u c t i o n 137 7.13 Effect of Sulfate on p H Profiles 143 7.14 C o m p a r i s o n of p H without Sulfate a n d w i t h Sulfate 144 7.15 Profi le of Sludge 145 7.16 Profi le of p H 146 7.17 Profi le of Acet i c A c i d 147 7.18 Profi le of P r o p i o n i c A c i d 148 7.19 Profi le of C O D . . 149 7.20 Effect of Sulfate on the Prof i le of A c e t i c A c i d : ( A ) W i t h o u t S, (B) W i t h S 151 7.21 Effect of Sulfate on the Prof i le of P r o p i o n i c A c i d : ( A ) W i t h o u t S, (B) W i t h S 152 7.22 Effect of Sulfate on B u t y r i c A c i d C o n c e n t r a t i o n i n the A c i d o -genic Phase 153 7.23 Effect of Sulfate on the A c c u m u l a t i o n ( A ) a n d the D e g r a d a t i o n (B) of P r o p i o n i c A c i d 163 7.24 Effect of Sulfate on the Permiss ible C o n c e n t r a t i o n D e t e r m i n e d b y A c c u m u l a t i o n and the D e g r a d a t i o n of P r o p i o n i c A c i d . ( A ) i n the A b s e n c e of Sulfate; (B) i n the Presence of Sulfate 16 4 7.25 A c c u m u l a t i o n and D e g r a d a t i o n of A c e t i c A c i d , P r o p i o n i c A c i d a n d T o t a l V F A s in the A b s e n c e of Sulfate 165 7.26 A c c u m u l a t i o n and D e g r a d a t i o n of A c e t i c A c i d , P r o p i o n i c A c i d a n d T o t a l V F A s in the Presence of Sulfate 166 xvii 7.27 T w o - S t a g e I n h i b i t i o n M e c h a n i s m 169 A . l F i t t e d V a l u e s versus the A c t u a l Values 211 A.2 A n a l y s i s of residuals 212 A.3 A n a l y s i s of residuals 213 A.4 A n a l y s i s of residuals 214 A.5 A n a l y s i s of residuals 215 A . 6 C o m p a r i s o n of E x p e r i m e n t a l D a t a w i t h M o d e l 216 B. l Effect of sulfate o n B i o g a s C o m p o s i t i o n 228 B.2 Effect of Sulfate o n the Biogas C o m p o s i t i o n 231 B.3 Interact ion Effects of Sulfate (S), Influent C O D ( C ) , and Sludge ( A ) : (a) S a n d C , (b) S a n d A , (c) C a n d A 232 B.4 Effect of Sulfate on p H 236 B.5 Interact ion Effects of Sulfate (S), Influent C O D ( C ) , and Sludge ( A ) : (a) S a n d A , (b) S a n d C , (c) C a n d A 237 B.6 F i t t e d p H versus A c t u a l V a l u e 238 B.7 A n a l y s i s of R e s i d u a l s 239 xvm Lis t of A b b r e v i a t i o n s ATP adenosine triphosphate ADP adenosine diphosphate BOD biochemical oxygen demand COD chemical oxygen demand HOA hydrogen oxidizing acetotrophs HOM hydrogen oxidizing methanogens HRT hydraulic retention time MPB methane producing bacteria NAD nictinamide adenine dinucleoticide NHOA non-hydrogen oxidizing acetotrophs NRB nitrogen reducing bacteria OHPA obligate hydrogen-producing acetogens OLR organic loading rate Pi phosphate pmf proton motive force SLR sludge loading rate SRB sulfate reducing bacteria SRT solids retention time TSS total suspended solid UASB upflow anaerobic sludge blanket reactor VFAs volatile fatty acids VSS volatile suspended solid xix A c k n o w l e d g e m e n t The author would like to express her sincere gratitude and appreciation to her thesis com-mittee members, D r . K . V . L o , Dr.K.L.Pinder and Dr.R.Branion, for their strong support, special guidance and valuable advice. Thanks should be given to the Universitj' of British Columbia and Canadian people for providing the award of University Graduate Fellowship in the second and third year of P h . D . program, and for their kindness that have made my study and life here possible and easier. The Author is also very grateful to many individuals in Department of Chemical and Department of Bio-Resource Engineering for their invaluable assistance and great deal of spiritual support. The author would like to thank Dr.P.H.Liao for his technical and laboratory advice, and to Adeline Lau for her assistance in laboratory analyses. Special thanks should be given to my dear husband and loving son, Tingwei Zhou and Xing Zhou. Without their sincere undersdanding, loving, moral and spiritual support, the thesis would never been finished. Finally, the author would like to thank her friends, Ms. Kathelly Gunn and M r . Richard Barrow for their caring and sincere friendship. xx C h a p t e r 1 I N T R O D U C T I O N 1.1 D I S P O S A L / U T I L I Z A T I O N O F C H E E S E W H E Y Cheese whey is a by-product of cheese production. Each pound of cheese produced results in five to ten pounds of fluid whey. In the U . S . A . , approximately 2 7 x l 0 9 tonnes of whey are produced each year, while 2 million tonnes of whey are generated annually in Canada. W i t h the increasing cheese demand in North America, f luid whey production is tending to increase. Cheese whey contains about 5% lactose, 1% protein, 0.3% fat and 0.6% ash (Loehr,1977). The C O D of cheese whey ranges from 60,000 to 70,000 mg C O D / l i t e r , depending on the type of cheese process. Every 1000 gallons per day of raw whey discharged into a sewage treatment plant can impose a load equivalent to that from 1800 people. Every 1000 gallons of raw whey discharged into a stream requires for its oxidation the dissolved oxy-gen in over 4,500,000 gallons of unpolluted water. The high organic content of whey, the trend towards increasing production of cheese whey and stricter pollution control standards have led to an expensive and difficult waste disposal problem for the cheese manufacturers. Cheese whey represents about 90% of the milk used in the cheese manufacturing process. A number of solutions have been proposed to recover nutrients and reduce the pollution level resulting from whey. These rely on converting the whey or various components of it to marketable products. Considering that whey has a high content of 1 Chapter 1. INTRODUCTION 2 lactose and protein, several investigators have developed new schemes for whey treatment with an emphasis on product recovery and new product development (Castillo et al. 1982; Friend et al. 1982). These efforts include fermentation of whey to ethanol for beverage production (Palmer 1978, 1979) or gasohol production; drying of whey into powder which may be used as animal feed or as a supplement in human foods (Muller, 1979; Modler, 1980); separation of whey components by membrane technology (Teixeira, 1982) or fermentation of whey for protein production. Many of these schemes, however, are limited to larger producers due to economic constraints. But even so, they don't completely solve the final waste disposal problem of whey. Further treatment is needed to meet the requirement for waste discharge. Biological waste treatment systems, either aerobic, anaerobic or combinations of them can be used to treat a wide variety of waste streams and are capable of reducing the levels of pollutants to meet even the most stringent requirements. However, the activated sludge process, which is one of the most commonly applied methods in waste treatment, is unsuitable for the treatment of very high strength wastes such as whey due to its large energy consumption for aeration, which leads to high operating costs. In contrast, anaerobic systems have much lower operating due to the low yield of cells and low energy consumption (no aeration) and produce methane which can be used as an energy source. 1.2 A N A E R O B I C M E T H A N E F E R M E N T A T I O N O F W H E Y As a process to convert the organic materials contained in biomass into useful energy (methane) and to reduce the emission of pollutants from industrial and agricultural wastes, anaerobic fermentation has already exhibited its potential. The growing real-ization of the potential of anaerobic treatment is evident from the large number of the Chapter 1. INTRODUCTION 3 research reports published on this process each year. Significant advances have been made in extending this process for the treatment of a variety of wastes . W i t h regard to cheese whey treatment, these advantages are pronounced because cheese whey constitutes a high strength organic waste. Several studies have shown that anaerobic treatment can achieve an adequate removal of chemical and biological oxygen demand from cheese whey and that the methane production is close to the theoretical yield. It has been estimated that 1 liter of whey can generate about 45 liters of gas with a C i / 4 content of around 55%, given that an expected C O D removal efficiency of 80% is achievable. For every liter of whey, 20 liters of C i f 4 can be generated, which is equivalent to an energy production of 700 Btu's . In Canada, 2 x 106 tonnes of whey are produced every year. This means 1.4 x 10 1 2 Btu's of energy could be obtained by anaerobic methanogenesis of whey. According to the results of a survey among cheese producers in New York state, up to 46% of the oil and gas needs of a cheese plant could be supplied by methane generated from whey (Switzenbaun, 1982). In spite of its present significance and its optimistic future potential, the anaerobic process has not had a favorable reputation because the conventional process usually suf-fers from a long start-up period, a slow digestion rate and unstable process performance, all largely due to the low growth rate of the anaerobic bacteria. These drawbacks have prevented the anaerobic process from having wide application. In the last decade, ad-vances in the microbiology and biochemistry of the anaerobic process, along with the advances in high-rate digestor technology have led to a great improvement in treatment efficiency and process control. One of the major advantages of anaerobic digestion is the relatively low yield of microbial cells (sludge) in comparison with the aerobic process. However, this becomes a disadvantage with respect to its longer period of start-up and its poor capability for tolerating load shock and other drastic changes in environmental conditions. It is now Chapter 1. INTRODUCTION 4 well understood that the efficiency of the anaerobic process is strongly dependent on the solid retention time (SRT). SRT requires considerably more attention in anaerobic digestion than in the aerobic process. Most of the new anaerobic reactor designs attempt to maximize SRT, and in turn the concentration of biomass in the reactor to permit a reduction in the required hydraulic retention time (HRT) . The anaerobic filter was the first design which allowed the SRT to be independent from H R T . Since then, a series of high-rate digestors haveeen developed. The development of novel high SRT designs for biological reactors has resulted in many new anaerobic systems which are suitable for the treatment of concentrated organic wastewaters. These new anaerobic digestors include the upflow and downflow anaerobic filter, the fluidized bed reactor,the expanded bed reactor and the sludge blanket reactor. The Upflow Anaerobic Sludge Blanket ( U A S B ) reactor is one of these innovative reactor designs (Lettinga et al.1980) . Two distinguishing features of the U A S B reactor are the installation of the 3-phase separator in the upper part of the reactor and the ability to cultivate sludge in granules. These two features permit the maintenance of a high concentration of biomass in the digestor. The U A S B reactor has been widely used to successfully treat a variety of wastewaters (Wang et al. 1985; Lettinga et al . 1985). However, little research has been done on the use of this reactor for treating cheese whey. Few studies on treating acidic substrates, such as monosodium glutamate wastewater and cheese whey, with an U A S B reactor are available so far (Wu and Zhang 1983; Samson et al. 1984). It has been suggested that the development of an anaerobic sludge with high activity and settleability is associated with the proper start-up procedures (Lettinga et al. 1979). Therefore, an assessment pf the technical feasibility of treating cheese whey in an U A S B reactor, concentrating on the start-up procedures, is necessary. Chapter 1. INTRODUCTION 5 1.3 P R O B L E M S A S S O C I A T E D W I T H T H E A N A E R O B I C D I G E S T I O N O F C H E E S E W H E Y The anaerobic digestion of cheese whey has been investigated by using different di-gestor configurations (Hakansson, 1977; Clanton et al., 1980; Boening and Larsen, 1982; Switzenbaun, 1979; Hickey and Owens, 1981; Sutton and Li , 1981). Treatment efficiency was affected by reactor type, experimental method, pH-control, nutritional supplement and waste strength. Unlike the other substrates which are usually used as feed for anaer-obic digestion, the treatment of cheese whey is much more difficult . Unsuccessful exper-iments have been reported, apparently due to its high organic strength and its tendency towards rapid acidification. A primary obstacle is a lack of fundamental understanding of the process. The problems encountered in this process can be attributed to inadequate buffering capacity and micronutrient deficiency. It was reported that stable operation of an anaerobic fermentation reactor with cheese whey could not be maintained unless pH control or additional nutrients (or mixture with manure) were applied. 1.3.1 Inadequate B u f f e r i n g C a p a c i t y Examining the effects of medium and inoculum on the treatment of whey and cellulose in an anaerobic fixed-bed reactor, Nordstedt and Thomas (1984) found that the reactor could not achieve stable operation within 30 days without p H control. Treating full strength whey in a fixed-film reactor, Marshall and Timbers (1982) reported that addition of N a O H was needed to maintain stability. Dehaast et al. (1983) used dilute whey as feed substrate and neutralized the acids to avoid a rapid p H drop in their experiment. In the study conducted by Follmann and Mark! (1979), a pH-static process was used in which the p H value was the control signal for whey feed. When the p H increased beyond 7.0, a Chapter 1. INTRODUCTION 6 pump was automatically triggered to add substrate until the p H fell to 6.95. Wildenauer and Winter also used a pH-titrate control unit in their fixed-film system to treat high strength acidic whey. Using an expanded bed reactor Switzenbaum and Danskin (1982) found that when the influent strength was increased from 5 to 20 g COD/l i ter , the C O D removal decreased from 83% to 58%. It was suggested that the physiological balance between the methane-producing organism and the hydrogen- and acid-producing organisms was more easily upset because of the great difference in the rates of acidogenesis and methanogenesis for an easily biodegradable carbon source, such as whey. It seems that a certain level of influent concentration represents a kind of inhibition to the system. 1.3.2 M i c r o n u t r i e n t Defic iency In their study of the effects of temperature on an anaerobic film expanded bed reactor ( A F E B ) treating cheese whey, Kelly and Switzenbaun (1984) noticed that in the absence of nutrient supplement, the C O D removal efficiency was much lower as compared with Switzenbaun and Danskin's results (1982) for the same loading rate and influent concen-tration. It was assumed that the poor removal efficiency was due to nutritional limita-tions. Some necessary nutrients, that were present in the tap water used by Switzenbaun and Dankin, were absent from the tap water used for dilution by Kelly and Switzenbaun. The results of experiments on nutrient requirements indicated that trace nutrients had a significant influence on reactor performance. Operation of the A F E B reactor improved after the addition of the nutrients. Gas production increased within 24 hrs.. C O D removal efficiencies increased from 60.3% for the nutrient-limited experiment to 80% for the nutrient-supplied experiment. The volatile organic acid ( V F A ) concentration in the effluent of the nutrient-limited experiment was at least three times higher than the V F A Chapter 1. INTRODUCTION 7 concentration of the nutrient-supplied experiment. Based on the theoretical value of the C / N ratio, DeHaast et al.(1983) performed a study in a downflow-fixed film reactor treating deproteinated cheese whey with different C/N-ratios ranging from 7.5 to 73. The results indicated that no decrease in efficiency and stability occurred even at the highest C / N ratio. It is known from the literature (Henze 1982) that the nutrient requirement is a function of organic loading rate (OLR) . Since the O L R used by Haast and his co-workers was only 2.6 g COD/l i ter day, their conclusions could not be regarded as being universal. Treating both whole and diluted cheese whey in an anaerobic rotating biological con-tact reactor, Lo and Liao(1986) noticed that the system could not be maintained at steady-state when the H R T was decreased below 5 days. However, a significant im-provement in efficiency and stability was obtained by using a mixture of cheese whey and screened dairy manure instead of cheese whey alone as the feed substrate (Lo and Liao 1987). In this case, the reactor could be operated successfully at a H R T of only 2 days without the addition of buffering and nutritional reagents. This study showed that something which existed in the manure was necessary for maintaining the stability of the system. The question remains as to what nutrients contained in manure help to restore the system and maintain it's stability. C h a p t e r 2 L I T E R A T U R E R E V I E W 2.1 D E V E L O P M E N T O F R E A C T O R D E S I G N T E C H N O L O G Y / U A S B R E -A C T O R In traditional anaerobic sludge treatment the solid retention time (SRT) and the hydraulic retention time (HRT) were almost identical. More recent studies have denned SRT as the crucial design and operational parameter because of the very low growth rate of anaerobic microbes. Proper SRT control provides sufficient acclimation time to allow efficient treatment. Young and McCarty (1969) were among the first to recognize this " S R T effect" when they introduced the concept of the anaerobic filter, in which the ratio of S R T and H R T could be increased to 100 (Henze 1982). The independence of S R T from H R T has proven to be a turning point in the study of the anaerobic process and has made anaerobic wastewater treatment economically interesting as compared to the aerobic process. Since then, the development of the anaerobic process has focused on the maintenance and accumulation of a high concentration of biomass in the reactor. A variety of high-rate reactors has been developed (Callander 1983). These high-rate, anaerobic reactors differ in the way in which biomass is retained. They can be broadly divided into two systems: attached growth system and suspended growth systems (Stronach 1986). The former includes upflow and downflow fixed-film reactors, expanded bed reactors and fluidized bed reactors, in which a high SRT is achieved by retaining the biomass as a film on the inert support media, packing material in a fixed-film reactor or particle material in 8 Chapter 2. LITERATURE REVIEW 9 fiuidized bed reactor. The latter is called an Upflow Anaerobic Sludge Blanket (UASB) reactor in which high biomass levels are accumulated and retained in the digestor by means of an interior gas-liquid-solid separator that relies for its effectiveness on the design of the separator as well as on sludge settleability. The U A S B reactor, which was initially developed by Lettinga and his co-workers, has already been recognized as one of the most promising methods for anaerobic treatment of organic waste (Lettinga et al 1984). Instead of using a packing material to support and concentrate biological growth, the upflow sludge blanket operates as a suspended growth system. Two distinguishing features of the U A S B reactor permit it to maintain a high concentration of biomass. These two features are a 3-phase separator and the quality of the cultivated granular sludge. The gas-liquid-solid separator (Lettinga et al 1980; Van der Meer 1982) is mounted in the upper part of the reactor. This arrangement has the following advantages: • The bottom plates of the settler can serve as gas separator. • No additional space is required. • The sludge separated in the settler can flow directly back into the reactor without any mechanical means, such as a pump or scrapers, being required. • The sludge is not exposed to "strange" conditions, it remains within the system. In order to enhance the return of the sludge from the settler, the following conditions have to be fulfilled: • Gas bubbles must be separated before the mixture of water and sludge enters the settling compartment. Chapter 2. LITERATURE REVIEW 10 • To avoid gas production in the settling compartment, the retention time of the sludge in the settler must be short. • The inclined wall of the settler should be at an angle of approximately 50° . • The surface load on the settler should be kept below about 2 to 2.5 m/hr. • The sludge present at the liquid-gas interface in the gas collector should be kept well immersed. The key to successful operation of the U A S B system is that the hydraulic loading rate and upflow velocities of fluid must not exceed the sludge particle settling rate. The second principal feature of the U A S B reactor is its ability to cultivate granular sludge with good settleability and high activity. The development of this granular sludge, in terms of both specific activity and settleability in the U A S B reactor, is associated with the composition of the wastewater, initial seeding and environmental conditions, including a temperature of 30 to 3 5 ° C and a p H of 6.5 to 7.2 (Wu 1985). Calcium ions play a positive effect on the flocculating ability of sludge, presumably due to the improvement of the mechanical strength of the floes (Lettinga 1980a; Hulshoff 1982 ). It has been found, using yeast, sugar beet or potato waste as substrates, that gran-ulation of sludge proceeded satisfactorily, whereas problems arose with distillery, corn starch and rendering wastes (Hulsoft pol et.al,1983). Granule characteristics may be detrimentally affected by significant amounts of suspended solid in the influent. The properties of the initial seed have an effect on sludge aggregation. Lettinga et al. (1985) suggested that thicker types of digested sewage sludge (approximately 60 kg T S / m 3 ) are preferred because the thicker sludge has better settleability, although it is Chapter 2. LITERATURE REVIEW 11 generally lower in specific methanogenic activity than thinner types of digested sewage sludge (i.e. 40 kg TS/m 3). The microorganisms themselves could function as a filtering medium (Lettinga 1981; Frostell 1981). To facilitate the filtration, the feed inlet system should introduce the influent wastewater homogeneously by using a distributor at the base of the reactor column. One of the most serious limitations of the UASB process is the considerable time (4-8 weeks or longer) for start-up. The washout of sludge during the initial phases of operation is significant. It has been noticed that the development of a sludge with high specific activity and settleability could be highly dependent on the start-up procedure. Both sludge loading rate and hydraulic loading rate are the most important factors which affect the properties of anaerobic sludge (Wu 1985). Considerable attention has already been paid to the start up procedure (De Zeeuw 1985; Hulshoff et al 1983, 1984; De Zeeuw and Lettinga 1980; Lettinga 1979, 1985; Dubourguier et al 1983; Ross 1983). It was indicated that the initial sludge loading should not exceed 0.2-0.4 kg COD/kg VSS day until the volatile fatty acids were well removed. Sludge granules formed rapidly only at loading rates in excess of 0.6 kg COD/kg VSS day (Lettinga et al. 1979). In the UASB system, agitation is provided by both hydraulic upflow and rising gas bubbles. A proper upflow velocity and the agitation due to rising gas bubbles are nec-essary for sludge aggregation. Higher hydraulic loading rates, ranging between 0.25-0.4m3/m2 hr, are favorable to the granulation of sludge because at these higher loading rates light floe sludge will move upward and be discharged from the reactor, whereas heavy granular sludge will move downward and grow quickly. The ability of the 3-phase separator to maintain a high concentration of biomass in the reactor makes the UASB reactor capable of attaining high rates of methane production, high rates of conversion for dilute and mostly soluble wastes, and of accepting organic and Chapter 2. LITERATURE REVIEW 12 hydraulic shock loads and temperature fluctuations. Since there is no need for a recycle pump, mechanical mixer, nor support media, the U A S B reactor is simple in construction and operation, has low energy consumption and is cheap to run and to maintain. In order to mathematically describe and optimize the U A S B process, various kinetic and dynamic models have been proposed , which include the fluid-flow pattern , the kinetics of substrate conversion and bacterial growth and the sludge distribution and behavior (Bolle et al 1986a, 1986b; Buijs et al 1980, 1982; Heertjes and Ven der Meer 1978, 1982,). A model based on a mass balance for the sludge in the blanket has been developed and experimentally checked for the physical behavior of the sludge in the blanket. Also, using stimulus response experiments with a L i + tracer, the dynamics of the fluid flow in a U A S B reactor were studied (Heertjes 1978). From the model, three distinct parts of the sludge could be distinguished. Both sludge bed and blanket can be described as perfectly mixed tank reactors with short-circuiting flows; the settler volume acts like a plug-flow region. Bolle and his co-workers proposed a dynamic model of a continuous working U A S B reactor which included the integration of the fluid flow pattern in the reactor, the kinetic behavior of the bacteria and the mass transport phenomena between different compartments and different phases. The mathematic equations are able to predict the various observable, nonobservable or difficult to observe state variables and were prepared for computation and simulation The wastewaters that have been treated include: sugar beet, potato starch, maize starch, alcohol, yeast, brewery, slaughter house, dairy, paper mill, distiller's grain super-natant, synthetic fatty acid, sewage and bean production wastes Although the application of the U A S B reactor to the treatment of industrial waste-water has been widely reported (Lettinga et al 1985; Wang et al 1985), few studies on treating acidic substrates so far are available. Using an U A S B reactor in the anaerobic treatment of monosodium glutamate wastewater with a p H of 2.0 and a chemical oxygen Chapter 2. LITERATURE REVIEW 13 demand (COD) concentration of 3 to 4.5 g/liter, Wu and Zhang (1983) reported a rapid decrease in pH in the reactor resulting in instability and a complete loss of bacterial activity. Lettinga(1979) suggested that improper procedures used in a UASB start-up period could lead to the development of sludge with low specific activity and poor settleability. Treating cheese factory effluent with an average strength of 2 g COD/liter, Samson et al (1984) found that the granular sludge in the UASB reactor tended to float, and a large amount of sludge was washed out. It might have occurred in the acidic environment because a massive growth of filamentous bacteria created a bulky sludge with poor settleability. 2.2 A N A E R O B I C T R E A T M E N T O F C H E E S E W H E Y I N D I F F E R E N T R E A C T O R C O N F I G U R A T I O N S The anaerobic digestion of cheese whey using different reactor configurations has been reported. These reports show that anaerobic treatment can achieve significant removal of chemical oxygen demand (COD) from whey. It is evident that treatment efficiency is affected by reactor type, waste strength and the experimental method used. Without the addition of nutrients and pH-control, a fixed-film reactor was able to treat higher strength substrates and had higher COD removal efficiency than a fluidized bed. However, a longer HRT was needed in the fixed-film reactor than in the fluidized bed. It was reported by several authors that at least 5 days of HRT were needed in the fixed-film reactor system (Hakansson 1977; Claton et al 1980; Boening and Larsen 1982; Switzenbaun 1979; Hickey and Owens 1981; Sutton and Li 1981); whereas the fluidized bed reactor could be operated at much shorter HRTs. In a study of the anaerobic digestion of high strength, acidic whey using a pH-controlled, up-fiow fixed-film loop reactor, Wildenauer and Winter Chapter 2. LITERATURE REVIEW 14 (1985) noticed that re-circulation of the reactor liquid helped to significantly improve the process efficiency. They suggested that the circulation of liquid was essential for fast gas expulsion. In addition, according to Monteith's studies (1981), liquid circulation could improve the mixing behavior of the reactor contents, in turn, reducing any dead volume and facilitating the contact between substrate and biomass . Furthermore, liquid circulation could minimize pH gradients along the reactor column and maintain a good pH environment for methane producing bacteria. These considerations help to explain why fluidized bed and expanded bed reactors are able to be operated at much lower HRT. Therefore, liquid circulation would be recommended to improve process operation in a stationary system. 2.3 E N V I R O N M E N T A L F A C T O R S I N A N A E R O B I C D I G E S T I O N 2.3.1 p H C o n t r o l i n the A n a e r o b i c Process The successful control of the anaerobic treatment process depends upon a knowledge of the various environmental factors which affect the microorganisms responsible for waste degradation. Of the various control factors, pH is one of the most important. Under anaerobic conditions, pH is controlled by the interaction of the carbonic acid system and a net strong base, the latter being the net result of the activities of the VFAs, ammonia and any other strong acids and bases present. This control depends on the maintenance of an adequate bicarbonate buffer system to counteract any acidity due to the carbon dioxide and organic acids produced during the anaerobic treatment. Alkalinity is a measure of the buffering capacity of the digestor contents. It consists of bicarbonate, carbonate, ammonia and hydroxide components. VFA, hydrosulphuric and orthophosphoric acids should also be considered. McCarty (1964) indicated that a Chapter 2. LITERATURE REVIEW 15 bicarbonate alkalinity in the range of 2.5 to 5.0 g CaCOa/liter (25 to 50 m M ) provided a safe buffering capacity for anaerobic treatment of waste. Hydrosulphuric and orthophos-phoric acid systems under anaerobic conditions were reported to provide very limited buffering capacity as they existed in extremely low concentrations. The dissociation con-stants of acetic and propionic acids allow them to be considered as weak acids (Capri and Marnis,1975; Stronach et al. 1986). Between p H values of 6.0-7.7, the buffering function of ammonia and the V F A s could be negligible. In an anaerobic system with a substrate which has a rapid acidification rate, the accumulation of intermediate fatty acid products might reach a concentration that ex-ceeds the system's buffering capacity (Mah 1969; Chynoweth and Mah 1977). Cheese whey is a substrate in which a majority of its contents are easily acidified. A n instability in the p H of an anaerobic system of cheese whey was often observed. Although high V F A concentrations with low p H values are particularly detrimental to methanogenic activity through the toxic action of the unionized V F A (Andrews et al 1971; Kroeker 1979), a high cation concentration caused by neutralizing agents, added to restore the p H , may also inhibit methanogenic activity (McCarty 1964; Kugelman 1971). Therefore, neutralization by base addition is not the best way to control p H . Recent studies, based on thermodynamic considerations and the hydrogen regulation function, have suggested that the accumulation of intermediate organic acids, during pe-riods of overloading or other process stress is, due to hydrogen sensitivities moderated by syntrophic associations between hydrogen-producing bacteria (acidogens) and hydrogen consuming bacteria (methanogens, sulfate reducing bacteria or S R B and nitrate reducing bacteria or N R B ) . Propionate, lactate and ethanol are produced when the accumulation of hydrogen is beyond the collective assimilative capacity of these hydrogenotrophs in-cluding methanogens, SRB and N R B . Since methanogens cannot use those end products directly as substrate, their accumulation leads to a depression in p H . Chapter 2. LITERATURE REVIEW 16 Considering that the group of acid and methane forming bacteria each display differ-ent requirements with respect to both environmental factors and nutrients for optimal growth, the concept of phase separation was proposed (Babbitt and Baumann 1958; An-drews and Pearson 1965; Pohland and Ghosh 1971; Fan et al. 1973). However, it should be realized that an one-phase anaerobic process can not be regarded simply as the sum of the acidogenic and the methanogenic phases of the two-phase process (Cohen 1980) because of the role of interspecies hydrogen transfer which is an important interaction that occurs between non-methanogens and methanogens in an anaerobic environment (Wolin 1974; Ianotti et al 1973; Hungate 1967; Scheifinger et al 1975; Mah et al 1977; Zeikus 1977; Patel and Roth 1978; Winfrey et al 1977). It might be argued that the phys-ical separation of methanogens and acidogens (two-phase system) can not enhance the anaerobic byconversion rate for some wastes since the necessary interspecies hydrogen transfer function is disturbed 2.3.2 N u t r i e n t R e q u i r e m e n t The anaerobic process is dependent on bacteria which require certain nutrients for growth. In addition to the fundamental requirements for macronutrients such as carbon and ni-trogen, the inability of a great number of anaerobes to synthesise some essential vitamins or amino acids necessitates their supplementation with such micronutrients for growth and activity. Since municipal waste sludge normally contains a variety of these nutrients, it usually can provide an ideal environment for growth. However, industrial wastes are frequently more specific in composition, micronutrients must be added to optimize the process. Chapter 2. LITERATURE REVIEW 17 Our lack of knowledge of the nutritional requirements of the methanogens has hin-dered the development of the anaerobic process. One of the earliest studies of the nu-tritional needs of the anaerobic process was conducted by Speece and McCarty (1964). Since then, attention to nutritional requirements has been paid by only few researchers. Such research has indicated that the microbial regeneration time is a function of the nutrients present. The rate of substrate metabolism may also be limited by nutrient limitations. Specific substrate utilization rates can be increased several fold when all the required nutrients are in excess (Speece 1985). Generally, nitrogen is the major nutrient, other than an energy source, for microbial systems. Speece and McCarty (1964) determined that the nitrogen requirements for the anaerobic digestion of fatty acids, carbohydrates and protein obeyed the following growth equations: for Glucose and Starch A = 0.46F - 0.088M (2.1) for Amino and Fatty Acids A = 0.054F - 0.03M (2.2) for Nutrient Broth A = 0.076F - 0.014M (2.3) Where A=substrate synthesized into biomass F=substrate metabolized by biomass Chapter 2. LITERATURE REVIEW 18 M=mass of biomass in system The nitrogen requirements for all substrates were equal to A/9.4. It was noted that microbial synthesis and, thus the nitrogen requirement, for carbohy-drate digestion is about six times greater than for proteins and fatty acids (Speece 1985). This has a significant impact on some nitrogen-deficient biomass feedstocks. Therefore, anaerobic conversion of high carbohydrate content feedstocks to methane gas deserves special consideration because of the relatively high ratio of microbial synthesis to sub-strate consumption. Speece (1964) also found that two stages exist in the anaerobic digestion of complex substrates, one in which the BOD remains constant and another in which the BOD is reduced due to production of methane. In the case of carbohydrates the major nutrient requirements and major biological solids accumulation result from the first or constant BOD stage of digestion. 2.3.3 Sulfate Effect Among the nutrient studies, the function of sulfate has received much less attention than other nutritional compounds. Therefore, the sulfur requirement for methanogens has not been extensively documented. However, many industrial wastes contain sulfate, and the role of sulfate and the sulfate-reducing bacteria (SRB) in anaerobic digestion cannot be underestimated. Species of the genera Desulfovibrio and Desulfotomaculum are routinely isolated from digestors. The reason why this family of organisms must be considered in anaerobic digestion processes is shown by the examination of the similarity and complementary and com-petitive relationship between them and the other native groups of microorganisms, i.e., Chapter 2. LITERATURE REVIEW 19 mainly acetogens and methanogens. It is known that methanogenesis and sulfate reduction are the two main terminal processes in the complete anaerobic mineralization of organic matter. As all of the methane-producing bacteria (MPB) are known to be able to couple the oxidation of molecular H2 with the reduction of its electron acceptor CO2 to yield the electron sink product CH4, all of the SBR are capable of coupling the oxidation of molecular H2 with the reduction of its electron acceptor SO^ to yield the electron sink product H2S. The electron donor in either of these cases may be supplied as a dissolved gas from an exogenous source or through a syntrophic association with an obligate proton reducing or hydrogen-producing acetogen (OHPA). Certain VFAs are metabolized by the SRB to acetate, CO2, and H2 in the presence of a methanogen, thus switching roles in mutualistic pairing. This type of intimate association or mutualism, is an example of interspecies hydrogen transport and plays an important role in this fermentation (Postgate 1984). In anaerobic digestion, sulfate reduction is usually considered undesirable not only because of the very toxic and corrosive gas, H2S, produced from the reduction of sulfate, (therefore creating a problem with regarded its removal from the biogas) but also because of the inhibition of methanogenesis. This inhibition has been explained in terms of the level of sulfide produced and the resultant cytotoxicity. Very recently this inhibition has been subjected to reaction rate kinetics and thermodynamic analysis. Kristjanson (1982) found that methane formation is generally absent in marine sed-iments when the sulfate concentration is high. He.attributed the apparent inhibition of methane formation by sulfate to a variety of factors including hydrogen sulfide inhibi-tion (Cappenberg, 1974), kinetic competition for substrate (Abram and Nedwell 1978 a,b; Bryant et al.1977; Mountfort 1980; Winfry and Zeikus 1977), and thermodynamic considerations (Zehnder 1978). Chapter 2. LITERATURE REVIEW 20 O n theoretical grounds, the higher free energy change associated with sulfate re-duction to H2S as compared to CO2 reduction to CH4 cannot in principle explain the inhibition of sulfide to methanogenesis (McCarty 1972; Thauer et al. 1977). The free energy change of sulfate reduction to H2S with H2 is 151 K J / m o l sulfate whereas that of CO2 reduction to CH^ with i f 2 is only 135 K J / m o l carbon dioxide. Therefore, the most likely mechanism is probably competition for substrates. Inhibition caused by the activity of sulfate reducers has been attributed to competition for the common substrates, since in the presence of excess substrates both processes can take place (Mountfort et al. 1980; Oremland and Taylor 1978; Winfrey and Zeikus 1977). The competition between sulfate reducers and methanogens in sediments or mixed cultures for H2 and acetate has been investigated by several workers. Qualitatively, their data indicate that sulfate reducers effectively compete with methanogens for both substrates. The fact that the addition of sulfate inhibits methanogenesis in anaerobic digestion can be attributed to the finding that the coupled sulfate reducing reactions are thermodynamically more favorable than are the methanogenic reactions. Sulfate reducers have demonstrated a higher affinity for hydrogen than methanogens in marine sediments (Ormaland et al. 1978 and 1982, Martens et al.1974, Abram et al. 1978a, 1978b, Sorensen et al. 1981), freshwater sediments (Winfrey et al.1977, Strayer et al. 1978, Cappenberg 1974 and 1975) and in pure or mixed culture with methanogens (Kristjiansson et al. 1982). Schonheit et al.(1982) has shown that in the case of H2 the Ks (affinity con-stant) value of a typical sulfate reducer was found to be about 5-fold lower than that of a methanogen isolated from a similar habitat. They also reported that the appar-ent Ks value of D.postgatei for acetate is lower than that of M . barkeri by a factor of 15. The explanation for the sulfate-reducing bacteria in general having a lower Ks for acetate than the methanogens might be the different mechanisms of acetate utilization. D.postgatei oxidizes acetate via the citric acid cycle (Thauer 1982) and Methanosarcina Chapter 2. LITERATURE REVIEW 21 cleaves acetate by a still unknown mechanism (Zehnder and Brock 1979). The most straightforward explanation as to why SRB has a lower Ks for H2 might be the differ-ent topology of the common hydrogenase enzyme system. The hydrogenase reaction is a specific control mechanism in strict and facultative anaerobes which governs the flow of electrons. In the M P B , hydrogenase is free in the cytoplasm, while in the SRB it is located within the periplasmic space (Kristjiansson et. al 1983, Stronach et al 1986). Such a location presents less of an osmotic barrier. Another possibility is that thermody-namics indirectly puts a constraint on the kinetic parameters of the biological reactions involved. As mentioned previously, the free energy change of sulfate reduction is higher than that of C02 reduction. The Haldane equation (Fresht 1977) predicts that the Ks of an enzyme is partly determined by the free energy change associated with the particular reaction. Their experimental results show that CHA production and sulfate reduction are not mutually exclusive and in the presence of excess H2 they have no effect on each other. When the H2 supply becomes rate limiting, however, competition does take place. The methanogenic bacteria are not inhibited by the activity of the sulfate reducing bacteria but have a lower affinity for the common substrate which results in suppression. In other words, the difference in substrate affinities accounts for the inhibition of methanogenesis from H2 and C02 in sulfate rich environments where the H2 concentration is well below 5 M . Complete conversion of organic waste is possible by SRB even with total methanogenic inhibition (Stephen, 1986). A comparison of the kinetics of hydrogen and acetate uptake by M P B and SRB (Table 2-1) suggests that higher organic waste conversion rates may be available through sulfate reduction than through methanogenesis. Moreover, S R B are not limited to one or two-carbon substrates, as are methanogens. This approach therefore may hold possibilities for reducing propionic acid and hydrogen in a stressed reactor, in order to reestablish equilibrium with the hydrogen removal system. However, Chapter 2. LITERATURE REVIEW 22 such an approach has major disadvantages, including the loss of energy available from methane and the production of sulfide. Limited references in recent literature do reveal that the sulfur requirement is part of a complex picture. Observations have shown that a special relationship exists between the H2 utilizing methanogens and the sulfate reducing bacteria. It has been suggested that the sulfate reducing bacteria might help to maintain the anaerobic conditions required for the growth of methanogens (Stephen 1985). In addition, the methanogens are dependent on the production of sulfide for growth. Ronnow and Gunnersson (1981) reported that a thermophilic methanogenic bac-terium has a specific sulfide requirement for methane production and growth. Scherer and Sahm (1981) stated that optimal growth of M. barkeri occurred on a defined medium containing methanol when 2.5 to 4.0 m M sodium sulfide was added. They also reported that iron sulfide, zinc sulfide or L-methionine could also act as sulfur sources. How-ever, the addition of sodium sulfide to a sulfide depleted media failed to restore growth. Mountfort and Asher (1979) found that a part of the sulfide requirement for growth is used as a precursor of H S - C o M which is 40% sulfide by weight. According to Ronnow and Gunnarsson's studies, 2.6% of the cell mass of M . thermoautotrophicum is sulfur. At sulfide levels below 0.1 m M , growth of M . thermoautotrophicum was poor and the methane production rate decreased. After injection of sulfide, a large increase in methane production was noted within 30 minutes. They noticed a linear relation between growth, methane production and sulfide concentration. They also noted that sulfate or thiosulfate could not replace the sulfide requirement. In subsequent work Ronnow and Gunnarsson (1982) noted that when sulfide was added in increments of 20 mg/1, the methane production rate increased until sulfide was completely depleted in the medium (as determined by sulfide analysis) and that the methane production rate then decreased. This observation provides further indication of a Chapter 2. LITERATURE REVIEW 23 Table 2.1: K i n e t i c Characterist ics of Sulfate R e d u c e r s a n d M e t h a n o g e n s Reference Culture Substrate J max Ks Y ( d a y 1 ) (mg C O D / l ) ( g V S S / g C O D ) Ghosh methane phase acetate 0.49 4200 Massey methane phase acidified glucose 0.43 395 Smith Methanosarcina barken acetate 0.60 320 0.04 Zhender Huser Methanothrix soehngenii Methanothrix soehngenii acetate acetate 0.11 0.16 30 45 0.03 0.02 Hungate Rumen bacteria H 2 / C 0 2 5.4 2*10-4 Shea digesting sludge H 2 / C 0 2 1.1 0.75atm Badziong Desulfovibrio vulgaris H 2 / S 0 4 = 3.6 0.09 Badziong Desulfovibrio vulgaris H 2 / S 0 4 = 5.0 0.15 Lovley Lake sediment H 2 / S 0 4 = O.OOlatm Middleton digester sludge acetate /S0 4 = 8.3 0.05 L i u Desulfovibrio vulgaris l a c t a t e / S 0 4 = 0.12 Schonheit Desulfovibrio postgatei acetate /SO^ 13 Liu Desulfovibrio oreintis l a c t a t e / S 0 4 = 0.04 Chapter 2. LITERATURE REVIEW 24 sulfur compound requirement for high-rate methane formation rates. They further stated that although methanogenic bacteria contain large amounts of coenzyme M , between 0.3 to 16 fimole/g dry weight (Balch and Wolfe 1979), coenzyme M cannot account for a large part of the sulfur absorbed. A large pool of sulfur compounds, possibly low molecular weight compounds, still remains to be discovered. Methionine and cysteine obviously account for a large part of the sulfur. Zehnder (1980) noted the effects of various sulfide concentrations on the growth and specific methane production rate of M . arboriphilus at p H 7.0. Optimal growth and specific rate of methane production required the presence of between 10~6 and 1 0 " 3 M sulfide. A sulfur source of about 0.85 m M was found to be essential for degradation of cellulose to methane (Khan and Trottier 1978). At 9 m M , all inorganic sulfur compounds other than sulfate inhibited both cellulose degradation and methane formation. The inhibition increased in the following order: thiosulfate C H 3 C O O ' + HCO3- + H + +3H 2 +18.2 C H 3 C H 2 C H 2 C O O - +2H 20 —-> 2CH3COO- + H + +2H 2 +11.5 2CH3CHOHCOO' + 4H 2 0 —> 2CH3COO- +2HCO3- +2H++4H2 -1.9 C H 3 C H 2 O H + H 2 0 — > C H 3 C O O " + H + + 2 H 2 +2.3 B. H2-utilizing methanogens and desulfovibrios : 4 H 2 +HCO3- + H + -~-> 2 C H 4 + 3H 2 0 -32.4 C . Syntrophic association of coculture (A+B): 4 C H 3 C H 2 C O O - + 3 H 2 0 — > 4 C H 3 C 0 0 - + H C 0 3 - + H + + 3 C H 4 -24.4 2 C H 3 C H 2 C H 2 C O O - + H C 0 3 + H 2 0 — > 4 C H 3 C 0 0 - + C H 4 + H + -9.4 2 C H 3 C H O H C O O - + H 2 0 — > 2 C H 3 C 0 0 - + H C 0 3 - + H + + C H 4 -34.3 2 C H 3 C H 2 O H + H C 0 3 - — > 2 C H 3 C O O - + C H 4 + H 2 0 + H + -27.3 Chapter 2. LITERATURE REVIEW 29 There is another group of bacteria, the homoacetogenic bacteria, which can ferment a very wide spectrum of multi or one carbon compounds (e.g. sugars, acids, CO2, CO, H2, etc) to acetic acid (Yang, 1984). As a consequence of consuming hydrogen rather than producing it, homoacetogens may lower the hydrogen partial pressure during anaer-obic digestion. However, in a normal anaerobic digestion system, methanogens appear to successfully out-compete homoacetogenic bacteria for hydrogen since H2 is mainly used for reducing C02 to CH4. So far, little is known about the functional importance of homoacetogenic metabolism in anaerobic digestion or the metabolic interaction of homoacetogens and methanogens (Zeikus 1981). Acetate, C02 and H2 are catabolized to the terminal products by methanogens (Fig-ure 2.1). The methanogens are a unique but diverse group of organisms. Methanogenic bacteria isolated to date are limited to catabolism of one- and two carbon compounds (Table 2.3). Most of them utilize hydrogen and one-carbon compounds such as carbon monoxide, carbon dioxide, and formate as substrates for methane production. There are two known genera of methanogens, Methanosarcina and Methanothrix, which can utilize the two-carbon compound, acetate. Methanosarcina, which can utilize H2/C02 as well as acetate and other one-carbon compounds, are classified as hydrogen-oxidizing-acetotrophs ( H O A ) . Since methanothrix are unable to use hydrogen.in combination with carbon dioxide, they are classified as non-hydrogen-oxidizing acetotrophs ( N H O A ) . Those which can utilize H2/C02 and other one- carbon compounds, but do not cleave acetate are named hydrogen-oxidizing methanogens (HOM). Acetoclastic methanogens convert acetic acid into end productions: carbon dioxide and methane according to the reaction CH3COOH = CHA + C02 (2.9) Chapter 2. LITERATURE REVIEW 30 The conversion of acetate accounts for 60% to 75% of the methane formed and is of critical importance in maintaining a relatively stable fermentation and in determining the maximum volatile solids loading rate at a given hydraulic retention time. However, up to date, only few microbes have been found which can produce C i f 4 from acetic acid so far. Therefore they most likely do not play a major role in acetate utilization in methanogenic fermentations. Perhaps, since they are difficult to isolate due to a close symbiotic relationship with other bacteria, they grow slowly (minimum doubling time of 2-3 days), and they are extremely sensitive to oxygen, the acetate utilizating methanogenic bacteria of major important have not yet been isolated. The hydrogen-utilizing bacteria (hydrogen-oxidation methanogens) are responsible for removing almost all of the hydrogen from the system. They grow quite quickly with minimum doubling times of around 6 hours and control the redox potential of the digestion process. The traces of hydrogen that they leave behind are believed to regulate both the total rates of acid production and acid oxidation. They are the autopilot of the anaerobic digestion process. 2.4.2 M i c r o b i a l Interaction in A n a e r o b i c D i g e s t i o n A microbial ecological population distribution is determined when a fragile compromise is forged among the various species present and their environment; and it is this com-promise, involving synergistic and antagonistic interrelationships, that makes anaerobic digestion both possible and limited. It has been accepted that sugar is fermented mainly via the Embden-Meyerhof-Parnas pathway to pyruvate, which is then further catabolized to fatty acids, alcohols, hydrogen and CO2- N A D H generated in glycosis must be reoxidized to N A D before the degradation Chapter 2. LITERATURE REVIEW 31 Table 2.3: M e t h a n o g e n i c Substrates a n d React ions Substrate Reaction A G ° ( K c a l ) 1. H 2 - C 0 2 : H 2 +HCO3- + H + —> C H 4 + 3 H 2 0 - 32.4 2. Formate: H C O O " + H 2 0 + H + —> C H 4 + 3 H C 0 3 - - 31.2 3. Acetate: CH3COO- + H 2 0 —> C H 4 +HCO3- -7.4 4. Methanol: 4 C H 3 O H —-> 3 C H 4 + HCO3- + H + + H 2 0 -75.2 5. Methyl amines: 4 C H 3 N H 3 + + 3H 2 0 —> 3 C H 4 + HCO3- + 4NH 4 ++H+ -53.8 2 ( C H 3 ) 2 N H 2 + +3H 2 0 —> 3 C H 4 + HCO3- + 2 N H 4 + + H + -52.5 4 ( C H 3 ) 3 N H + + 9H 2 0 —> 9CH 4 + 3HCO3- + 4NH4++3H+ -159.8 6. C O : 4CO + 2 H 2 0 — > C H 4 + 3 C 0 2 -43.1 Chapter 2. LITERATURE REVIEW 32 of organic matter can continue (Figure 2.2). In single-culture system, electrons from pyruvate are used for the production of ethanol, and H2 is not a significant product. Regeneration of N A D for glycolysis is dependent on ethanol formation. This is due to the fact that the evolution of H2 from reduced nicotinamide adenine dinucleoticide ( N A D H ) is thermodynamically unfavorable under standard conditions: NADH + H+ = H2 + NAD+ (2.10) A G ° = + 4 . 3 Kcals When the H2-utilizing methanogens are cocultured with fermentative bacteria, the metabolic activities of the latter are shifted to generate more oxidized end products, e.g. acetate. Wolin (1976) illustrated this phenomenon by a hypothetical, but not unrealistic, fermentation pathway for the formation of ethanol, acetate, C 0 2 and H2 from glucose (Table 2.4). When the partial pressure of H2 is reduced, the free energy change becomes increasingly negative. Therefore, with methanogens, N A D is regenerated by formation of H2, which is removed by methanogens. Thus electrons from N A D H are used for H2 formation and methanogenesis in the mixed system rather than for ethanol formation. Through the methanogenic H2 removal, the cocultures not only alter the fermentation, but also provide an additional mole of A T P per mole of glucose fermented (Table 2.4). The effect of the interspecies i / 2 - transport reaction on fermentation has been demon-strated by several experiments (Zeikus 1982; Wol in 1974; Chung 1976; Weimer 1977) The # 2 -producing acetogenic bacteria catabolize the products of the hydrolysis stage, mainly propionate, butyrate, long-chain fatty acids, alcohols, and probably aromatic and other organic acids to acetate, H2 and C02. However, these organisms cannot catabolize Chapter 2. LITERATURE R E V I E W c a r b o h y d r a t e s i g l u c o s e glucose 6 phosphate — A T P -*-ADP fructose 1:6 diphosphate AOP ATP glucose 2 [glyceraldehyde phosphate] 2(phcsphati 2NAD 2NA0H*2H « 1 2[aeetyl CoAJ-211:3 diphosphoglycerafe] 2C02 l — 1.A0P 2HSGJA H 0 - 4 — A A T P •\— 2 [pyruvic acid] L 2NA0H*2H 2NA0-2H. ® NAO 2HSCoA ® | 2 N A I 2 NAOH 1 2ri • UHAOH* 4H • • 2 Glyceraldehyde-3-p +2 ADP Glyceraldehyde-3-p + 2 NAD+ +4 ADP + 2 pi — > 2 Pyruvate + 2 NADH + 4 ATP + 2 H+ 2 Pyruvate + 2 CoASH — > 2 Acetyl-SCoA + 2 H 2 + 2 C 0 2 A. Glucose-fermenting organism alone: Acetyl-SCoA + pi + ADP —> Acetate +CoASH + ATP Acetyl-SCoA + NADH + H+ —> Acetyldehyde + NAD+ + CoASH Acetyldehyde + NADH + H+ —> Ethanol + NAD+ Sum: Glucose + 3 ADP + 3 pi —> Ethanol + Acetate + 2 H 2 + 2 C 0 2 + 3 ATP B. Glucose-fermenting organic plus methanogen : 2 Acetyl-SCoA + 2 pi + 2 ADP —> 2 Acetate + 2 CoASH + 3 ATP 2 NADH + 2 H+ —> 2 NAD+ + 2 H 2 4 H 2 + C 0 2 —> C H 4 + 2 H 2 0 sum: Glucose + 4 ADP + 4 pi —> 2 Acetate + C H 4 + C 0 2 + 4 ATP Chapter 2. LITERATURE REVIEW 35 these substrates to acetate when H2 in the environment is not at an extremely low level. As is illustrated in Table 2.2 , only when acetogenic bacteria are placed in syntrophic association with i/2-utilizing organisms, such as methanogens and desulfovibrio, do the reactions become energetically favorable to proceed. (Mclnerney 1979). Although microbial interactions among diverse trophic groups of microbes in methane fermentation are not clear, methanogens have been considered as effective bio-regulators of anaerobic digestion through proton and electron exchange. By removing the metabo-lite of the acid-formers, acetate, methanogens provide a constantly favorable environ-ment for the fermentative bacteria. Otherwise most of them would be inhibited by their own metabolite. Also, electron transfer (or H2 transfer) between methanogens and iJ2-producers creates favorable conditions for metabolism of certain metabolites. The fermentation pathway of fermentative bacteria may be altered to gain more energy. On the other hand, the accumulation of intermediate organics, which are metabolites of acid-formers and have a detrimental effect on methanogenesis during periods of over-loading or other process stress, is due to hydrogen sensitivity moderated by syntrophic associations between hydrogen-producing bacteria (acidogens) and hydrogen-consuming bacteria (methanogens, SRB, NRB) . The physical separation of methanogenesis and aci-dogenesis (two-phase) in an attempt to optimize growth conditions for each group of bacteria theoretically cannot enhance anaerobic bioconversion rates since the necessary interspecies hydrogen transfer function is disturbed. Chapter 2. LITERATURE REVIEW 36 2.4.3 H y d r o g e n Regulat ions G e n e r a l Perspect ive Hydrogen in relatively low concentration appears to regulate the overall conversion pro-cess by throttling the acidogenic reactions at several points in the glucolytic pathway. Accumulation of hydrogen obviously needs an alternate method of electron disposal for N A D regeneration. This need drives the fermentation of pyruvate to propionate, lactate and ethanol and/or the fermentation of acetyl-CoA to butyrate. Since the methanogens cannot use these end products as substrates directly, their accumulation leads to a prob-lematic depression in p H . The subsequent degradation of these acids to acetate is also hydrogen dependent and is mediated by obligate hydrogen producing acetogens (OHPA) linked to the various hydrogen oxidizers. In addition to their hydrogen sensitivities, the number of O H P A and associated H O M (hydrogen oxidation methanogens) bacteria existing in an anaerobic process may vary, depending on its stability history. Under a process upset, the O H P A -H O M must increase to accommodate the accumulated hydrogen and propionate. If a long period of stable operation ensues, the relative proportion of propionate may decrease, and the number of O H P A will decrease accordingly. This dynamic balance is considered central to the overall stability and efficiency of many anaerobic processes and is regarded as the key to stabilizing and improving anaerobic treatments. In single-phase processes, the proportion of O H P A and methanogenic bacteria re-sponds to the relatively uncontrolled production of acidogenic products, which may alter-nately reach inhibitory and substrate-limiting proportions. A number of researchers have also demonstrated that the environmental requirements of acidogenic and methanogenic enrichments are quite different from each other as reflected in terms of p H and O R P (Hammer et al. 1969, Dirasian 1963, Niktin 1968, Blanc 1973 ). The two-phase system Chapter 2. LITERATURE REVIEW 37 was originally proposed to address these issues. However, the role of hydrogen was not emphasized as the primary control factor in earlier investigations. Although the two-phase process seems best suited for the treatment of soluble-type wastewaters, which present a high potential for volatile acids accumulation,(Ghosh et al. 1981, 1983a) su-perior performance has also been demonstrated for particulate-type substrates such as sewage sludge and agricultural wastes. (Ghosh 1983b, Keenan 1976, Normann et al. 1977). However, these studies failed to show definitive reasons for process improvements. T h e r m o d y n a m i c Quant i f ica t ion of H y d r o g e n Effects In view of the reported differences in hydrogen effects, and the traditional difficulties in adequately defining the redox nature of the wastewater substrate, at hydrogen levels of 10~ 6 to 1 0 - 1 0 atm, the regulatory effects of hydrogen on O H P A and H O M have been illustrated using thermodynamic models and equilibrium assumptions by several authors (McCarty 1971,1981, Mclnerney 1981, Gujer 1983). Thermodynamic calculations associ-ated with those half reactions during anaerobic stabilization of organic wastes to methane indicated that propionic acid oxidation to acetate becomes favorable only at a hydrogen partial pressure below 1 0 - 4 atm, while butyric acid oxidation becomes favorable at or below 1 0 - 3 atm (Figure 2.3). Figure 2.3 provides insight into the product formation pat-terns to be expected in a two-phase process. Evident from Figure 2.3 is the preference of sulfate reduction over bicarbonate respiration at all hydrogen partial pressures and the preference of acetate cleavage by S R B over methanogens. Chapter 2. LITERATURE REVIEW 38 T h e R o l e of H y d r o g e n i n R e g u l a t i o n of M e t h a n o g e n e s i s Standard free energy levels associated with the methanation reaction, listed in Table 2.3, indicate that the favorability of acetate cleavage is an order of magnitude lower than H2/CO2 and methanol conversions (at standard state of 1 atm of H2 in gaseous phase). However, this high level of hydrogen is uncommon in well-operating anaerobic wastewater treatment systems. At more realistic hydrogen concentrations, acetate cleavage competes more favorably with the methanogenic respiration of bicarbonate, as shown in Figure 2.3 In addition, hydrogen regulation in an anaerobic reactor is a result of the integrated effects of the specific capabilities of O H P A , H O A , N O H A , and H O M , and may be de-pendent on the cultures selected by seeding, substrates and operational procedures. For example, species of Methanosarcina ( H O A ) which utilize both acetate and H2/CO2 as substrates may be subject to catabolic repression of acetate cleavage by low levels of hydrogen. However, other investigators have reported a complete and rapid inhibition of acetate cleavage in the presence of H2/CO2 (Baresi et al.). H2/CO2 is preferentially utilized over acetate in mixtures of these substrates by Methanosarcona species to the extent that acetate cleavage is inhibited unt i l H2 is exhausted (Mclnerney 1981, Ferguson 1983, M a h 1978) Chapter 2. LITERATURE REVIEW 39 Gibb's Free Energy Change (KJ/reactlon) Figure 2.3: Graphical Representation of the Hydrogen-dependent Thermody-namic Favorability of Acetogenic Oxidations and Inorganic Respirations As-sociated 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 respiration of bicarbonate ( C 0 2 ) , 6 Methanogenic respiration of bicarbonate, 7 Respiration of sulfate to sulfite, 8 Respiration of sulfite to sulfide, 9 Methanogenic cleavage of acetic acid, 10 SRB-mediated cleavage of acetic acid (from Harper, 1986) C h a p t e r 3 R E S E A R C H O B J E C T I V E S A considerable amount of research has been directed over the years at using different reactor configurations to establish the development of an industrial anaerobic digestor for cheese whey. However, most of the previous studies were feasibility assessments of the process and were generally preliminary in nature and incomplete. Unsuccessful fermentation experiments have often been reported apparently due to the high organic concentration of cheese whey and its tendency toward rapid acidification. p H quickly droped and gas production and C O D reduction decreased. However, surprisingly little work has been done so far to explain the reason for this system instability. In addition, although the high organic strength of cheese whey can supply a useful source of energy for cheese manufacturers by means of the anaerobic fermentation, industrial applications have been very limited due to the unreliability of the process. It is therefore necessary to carry out a study to increase the fundamental understand-ing of the process, to find an explanation for its instability, to efficiently control such instability and to optimize the process. The special objectives include: (1) To assess the technical feasibility of treating cheese whey in a U A S B reactor, and to determine the effects of start-up procedures on reactor performance and treatment efficiency. (2) To determine the effects of influent concentration, H R T and O L R on the treatment efficiency, and to determine the inhibitory effects of influent concentration and organic 40 Chapter 3. RESEARCH OBJECTIVES 41 loading rate. Previous studies have shown that, when influent concentration was increased to a cer-tain level, the system became unstable. This was attributed to an upset in the physiolog-ical balance between the methane-producers and hydrogen-and acid-producers because of the great difference in the rates of acidogenesis and methanogenesis. This research wil l explore the difference in the rates between these two stages, the threshold level of the influent concentration and wil l try to obtain the optimal influent concentration for the particular system. (3)To assess the effect of sulfate addition on the buffering capacity and stability of the anaerobic system, and to investigate the mechanisms of inhibition. As has been discussed in earlier sections, knowing the importance of the nutritional requirements of methanogens is important to the overall knowledge of an anaerobic pro-cesses. Among these required nutrients, sulfur should be given more attention for special substrates such as whey. A limited number of literature references, which were cited in section 2 of the literature review, revealed that a sulfur requirement is part of a complex picture. On one hand, sulfate reducing bacteria help to maintain the anaerobic conditions required for the growth of methanogens, while on the other hand the addition of sulfate inhibits methanogenesis because sulfate reducers have demonstrated a higher affinity for hydrogen than methanogens. Methanogens and sulfate reducers are not mutually exclu-sive. In the presence of excess hydrogen, they have no effect on each other. W h e n the hydrogen supply becomes rate l imiting, competition does take place. From the point of view of thermodynamics and hydrogen regulation, the hydrogenotrophic association would help to maintain the reaction condition favorable for methanogenesis. It is expected that the results of this research wi l l yield the proper sulfate concentration needed to moderate the detrimental influences of excess hydrogen on a stressed anaerobic reactor. Chapter 3. RESEARCH OBJECTIVES 42 Obviously, the practice of sulfate addition holds numerous potential challenges, in-cluding the control of the inhibitory and corrosive hydrogen sulfide gas. However, such a study might yield valuable information on the utility of manipulating hydrogen to control acidogenesis and methanogenesis. Moreover, a number of industrial wastewaters contain significant levels of sulfates. The successful anaerobic treatment of these wastewaters requires an understanding of the competition for methanogenic precursors by SRB as well as of the associated microbiology and substrate conversion kinetics. (4)To explore maximum treatment efficiency by using optimal control parameters based on all the above studies . C h a p t e r 4 E X P E R I M E N T A L M E T H O D 4.1 R E A C T O R S E T - U P The flow sheet of the U A S B system is presented in Figure 4.1. The reactor was made of acrylic pipe with an inner diameter of 11.5 cm (4.5 in.) and a height of 168 cm (60 in.). The total volume and working volume of the reactor were 17.5 and 14.3 liters, respectively. A series of sampling ports were fitted at intervals on one side of the reactor to permit sampling for the analysis of the reactor contents. The reactor was operated in an upflow, continuous mode. A distinguishing feature of this U A S B reactor design is a conical three-phase separator located at the top of the reactor. The three phases refer to gas, liquid and solid (sludge). The basic principle of separation by the separator is explained as follows. Gas bubbles produced by the bacteria rise to the top of the reactor where they are first separated from the liquid and collected in the gas chamber between the column and the top cone. The biogas passes through a water seal column outside the reactor, which is used to regulate the liquid level in the separator, then goes to a wet gasmeter for measurement. The lower baffle, a small cone suspended below the larger,top cone is used to minimize the possibility of gas entering the separator by blocking the uprising gas bubbles from the open tip of the larger cone. The separation of sludge is accomplished by fluid dynamics. The floes of sludge with poorer settleability rise through buoyancy provided by rising gas bubbles, and travel with 43 Chapter 4. EXPERIMENTAL METHOD 44 BIOGAS THREE PHAS: SEPARATOR TEMPERATURE CONTROLLER BIOGAS =X*=10 9 s p L I N =x» G A P S 1 M HEATING TAPE FEED PUMP WATER SEAL WET GAS METER I- V-FEED T A N K Figure 4.1: Schematic Diagram of U A S B System Chapter 4. EXPERIMENTAL METHOD 45 the liquid into the upper cone. The upflow velocity of the fluid decreases as the cross sectional area increases. When the upflow velocity of fluid decreases to a value equal to downflow velocity of the sludge, the rising particles slow down, drop and return back to the reactor. This design was efficient in retaining a large percentage of the bacteria in the reactor. The whey, stored in a refrigerator at 4 ° C , was introduced continuously into the bottom of the reactor by a peristaltic pump through a copper-alloy coil immersed in a water bath to bring the feed temperature to about 34° C . The effluent left the reactor from the top of the settlling chamber and was collected in a plastic container. Fermentation temperature was maintained at 33 ± 1° C using a feedback controller and a number of external electric heater pads wrapped around the reactor. In this set-up, there was no p H control system. 4.2 F E E D S U B S T R A T E Cheddar cheese whey used in this study was obtained from the Fraser Valley Milk Pro-ducers Association's cheese producing plant at Abbotsford, British Columbia, Canada. Typical whey composition is presented in Table 4-1. The manufacture of cheese from either whole or skim milk produces, in addition to cheese itself, a greenish-yellow fluid known as whey. Whole milk is used to produce natural or processed cheeses such as cheddar, and the resulting whey has a p H in the range of 4.5 to 6. The lower p H is the result of the acid developed during coagulation. As indicated in Table 4.1, although the basic nutritional requirement of nitrogen in whey for bacterial growth seems to be adequate according to the criteria of C O D : N : P being 100:5:1, the extremely low ammonium nitrogen is perhaps the main cause of the apter 4. EXPERIMENTAL METHOD 46 Table 4.1: Characteristics of Chedder Cheese Whey Total solid (TS) 5.66-5.88 % Volatile solid (VS) 4.52-4.70 % Total Chemical Oxygen Demand (TCOD) 64-67 g/1 Soluble Chemical Oxygen Demand (SCOD) 62-65.5 g/1 SCOD as percent of TCOD 96-97 % Biological Oxygen Demand (BOD) 60-62 g/1 BOD as percent of TCOD 95-96 % Total Kjeldahl nitrogen (TKN) 2.75-3.05 g/1 Ammonium Nitrogen 2.8-3.0 mg/1 Nitrate 0.45-0.70 mg/1 Total Phosphorus 0.34-0.37 g/1 P H 4-5.5 Chapter 4. EXPERIMENTAL METHOD 47 lower buffering capacity of whey. In addition, a shortage of phosphorus is apparent. Two characteristics of raw whey are its unusually high S C O D to T C O D ratio and extremely low p H . These can create difficulties during anaerobic processing. Fresh cheese whey was contained in 5 50-liter tanks and transported from the cheese plant to the large freezer of the bio-resource lab at U B C once every 3 months, and stored there at -30° C . About one week before it was needed, a portion of the frozen whey was moved into a cold room at 4° C for defrosting. Then it was prepared to the desired C O D concentration by mixing with cool tap water, adjusted to a p H of about 7 by using 5% N a O H , and transferred to the small scale, feed tank, where it was kept refrigerated at 4° C and ready for use. 4.3 S E E D S L U D G E The seed sludge was originally obtained from the effluent of an anaerobic rotating bi-ological reactor ( A n R B C ) in the laboratory and had been stored in a plastic container in a cool room at 4° C for several months. The bacterial concentration in the effluent from the A n R B C was very low, even in the dense portion in the bottom. To obtain the necessary concentration and activity of seed sludge to ensure a successful start-up of a U A S B reactor, it had to be incubated. First, the dense portion of the effluent in the container bottom was moved to a glass jar at a room temperature of 22° C . The sludge was acclimated using 200 ml of raw cheese whey daily for more than 30 days before it was put into the reactor. At the time of seeding, the sludge concentration was 3% T S and 2.1% V S , respectively. Four liters of this effluent were used as seed . Chapter 4. EXPERIMENTAL METHOD 48 4.4 E X P E R I M E N T DESIGN It was the author's intention to attempt to achieve a high treatment efficiency initially by using a U A S B reactor without pH-control or nutritional addition, and to optimize the process through experimental investigations. This research was carried out in 18 months in 4 steps, in which a variety of influent concentrations and HRTs were used as shown in Table 4.2. For each step, the reactor was started using 4 liters of the seed sludge and 6 liters of diluted cheese whey of 5 g C O D J\. The remaining 7 liters (the digestor capacity ) were filled by continuously adding whey at the rate of 2.5 liters/day. The experiment was started with a very dilute influent concentration of about 5 g C O D / 1 in order to observe the effect of a wide range of substrate strengths on treatment efficiency and gas production. After 40 days of operation, the reactor was stabilized at a constant gas production and effluent properties, including C O D , V F A , T S S and p H . Then the influent concentration was increased stepwise from approximately 10 to 40 g C O D / 1 at a constant H R T of 5 days during the preliminary assessment of technical feasibility. The stepwise increase of feed strength not only allowed the microbes to gradually acclimate themselves to the new environment with a minimum negative impact, but also permitted a full view of the performance profile over the entire range of the operational settings with the objective of locating the optimum performance conditions. The tests on the effects of sulfate ions were first carried out in batch experiments to get a basic sense of sulfate behavior in the anaerobic process. A 3-factor experimental design was used in the batch experiment. The details are given in Appendix A . O n the basis of the results of the batch experiments, continuous flow experiments were designed to gain more accurate information about the sulfate effects. Two reactors were operated simultaneously at the same temperature with the same seed, but with different operating Chapter 4. EXPERIMENTAL METHOD 49 Table 4.2: Experimental Design Experiment H R T (days) Parameters Inf. Cone, (g COD/1) Sulfate (g/1) Start-up 5 5-9.9 0 Effect of Inf. Cone. 5 5-38 0 Effect, of H R T 24-5 20 40 oo Effect of Sulfate 5 15-50 0.2 0.3 0.5 Optimal Operation 7.5 3.8 2.0 26-32 0.2 0.2 0.2 Chapter 4. EXPERIMENTAL METHOD 50 conditions. Finally, an experiment was designed using the optimal operating conditions found from the previous experimental results to achieve the optimum treatment efficiency. 4.5 R E A C T O R O P E R A T I O N The reactors were continuously fed and operated at 33° C . The feed rate was monitored and recorded every 3 hours to ensure a constant daily feed. The production of biogas was measured daily. Samples were taken daily of the influent and effluent for analysis of C O D , V F A , TSS, VSS and p H . The analysis of the gas composition was also performed every day. During the time of changing loading rate, the biogas was sampled up to three or four times daily. When the gas production rate and the effluent C O D were stationary again (3 to 5 days for gas production and 6 to 10 days for C O D ) the reactor was at steady state. Samples of the mixture of liquid and solid in amount of 200 to 300 ml (just enough for chemical analyses) were then taken from the 10 sampling ports mounted in the reactor wall for measurement of the sludge, C O D , butyric, propionic, acetic acids and pH for each operating condition. The amount of sample taken from each sample port was recorded for the calculation of sludge growth in the reactor. To obtain reliable results of reactor profiles, only one port was sampled every hour from sample port 10 (top) to port 1 (bottom). For each subsequent increment of influent concentration, an operating period of 2 to 3 H R T s was maintained to ensure stable operation. Chapter 4. EXPERIMENTAL METHOD 51 4.6 ANALYSIS Analyses conducted on the influent and effluent were as follows: total solids (TS), total suspended solids (TSS), volatile suspended solid (VSS) and ash according to the "stan-dard methods" ( A P H A , 1975). Chemical oxygen demand ( C O D ) was determined by a colorimetric method using an optical fiber instrument (Knechtal 1978). Gas production was measured by a wet gas meter and then corrected to the standard temperature and pressure (STP) . Both gas composition and volatile fatty acids (VFAs) were analyzed on a Hewlett Packard 5890a gas chromatograph, using an external standard. The G C was equipped with both a flame ionization detector and a thermal conductivity detector with separate columns; a Porapak column was used for gas analysis and a carbopack C column for V F A s . Total Kjeldahl nitrogen ( T K N ) , and ammonia nitrogen (NH3 — N) were determined using a block digestor and a Technicon Auto Analyzer II (Schulmann et al 1973). C h a p t e r 5 S T A R T - U P A N D E F F E C T S O F P R O C E S S P A R A M E T E R S 5.1 I N T R O D U C T I O N Cheese whey acidifies easily and frequently causes problems when treated biologically. Some frustrating, unsuccessful experiments have led to the suggestion that the system was hard to maintain stable without p H control. The necessity of some micronutrients was also seriously considered. In view of the obvious difficulties in start-up of a cheese whey anaerobic system and maintenance of its stability, this part of the dissertation centers on the feasability of anaerobic digestion of cheese whey by using a new reactor configuration, a U A S B reactor, in an effort to increase VSS concentration in the reactor. The other major differences of this experiment from previous research are that no p H control or nutriential addition were applied in order to find the mechanism of the reported inhibition. The U A S B process has been described in the literature as one of the innovative reactor designs which permits efficient and economical treatment because of its ability to retain higher VSS concentrations in the reactor. However, the application of a U A S B reactor to treat high strength acidic wastes has not been established. This could be due to improper start-up procedures and/or a poor understanding of the importance of the control parameters during the start-up period. This study included the following investigations: 1. Proper start-up procedures for a cheese whey U A S B system. 52 Chapter 5. START-UP AND EFFECTS OF PROCESS PARAMETERS 53 2. Determination of the quality of effluent and treatment efficiency when the reactor was operated over a wide range of influent concentrations. Various operating parameters were to be used to assess the effectiveness of the process. 3. Effect of influent concentration on gas production and system performance. 5.2 S T U D Y I N S T A R T - U P O F T H E R E A C T O R 5.2.1 Difficult ies i n S t a r t - u p Use of the U A S B process is dependent on good sludge floe formation. The procedure of start-up is important for the development of active sludge with both high specific activity and settleability. Considerable attention has been directed to the start-up of the U A S B reactor because of the very slow growth of anaerobic microbes. It was found in this experiment, that for cheese whey, the start-up of a U A S B reactor was even more arduous. This might be attributed to the inherent characteristic of the substrate of being extremely easily acidified. Various start-up strategies were tried. The parameters applied in the period of start-up are presented in Table 5.1. The results have shown that sludge loading rate was the most critical parameter and must be carefully controlled. The reactor 1 which was fed with the lowest strength whey of 5 g C O D / 1 operated very well. The other two (reactor 2 and reactor 3), which were started with higher strength whey, experienced difficulties. Over the first 3 days, reactor 2 was fed at a rate of 2.7, 2.4 and 2.6 liters daily with an influent concentration of 30 g C O D / 1 . The methane content in the biogas was extremely low, only 19.8% on the third day. The reactor finally failed to start-up because the sludge loading rate (SLR) reached 0.38 to 0.43 g C O D / g VSS.d . Then, 5 liters of seed sludge with a VSS of 24.2 g / l was added. Chapter 5. START-UP AND EFFECTS OF PROCESS PARAMETERS 54 Table 5.1: S ludge L o a d i n g i n S t a r t - u p R e a c t o r Influent (g C O D / 1 ) F e e d (1/d) S L R (g C O D / g VS!|>) P e r f o r m a n c e R i Rr R2 R/ i 5 30 a 28b 20 63 20 2.8 2.6 2.6 1.0 1.0 0.5 0.162 0.300 0.490 0.190 0.214 0.256 stable failure failure stable stable stable N o t e : In a & b , the reactor was seeded i n different a m o u n t of sludge Chapter 5. START-UP AND EFFECTS OF PROCESS PARAMETERS 55 The total amount of VSS in the reactor was 121 g VSS. The daily feed of 2.1 liters of influent with a concentration of 28 g C O D / 1 caused the methane content to drop from 51.9% to 44.3%, and further to 25.2%, since the S L R was as high as 0.49 g C O D / g V S S . d at that point. A n attempt to recover the reactor was made by reducing the feed to 0.9 liters daily. Immediately, biogas composition rose to 44.1% methane by the next day. 2.2 1/d feed was tried again. However, the reactor was entirely unstable which was indicated by drastic drop in methane content of the biogas from 44% to 34%. A number of attempts to start-up the cheese whey U A S B reactor indicated that a successful start-up was strongly related to the SLR. Table 5.1 clearly shows this relation-ship between the performance of the reactor and SLR. No matter how strong the influent concentration, up to 63 g C O D / 1 for reactor 3, or how large the feed rate, up to 2.8 1/d for reactor 1, if the S L R was lower than 0.26 g C O D / g VSS~d, the reactors were able to start without any problem. The reactor start-up was destined to fail, on the other hand, if the S L R employed was beyond 0.26 to 0.3 g C O D / g VSS-d . From this study it can be concluded that S L R is the most important parameter to be controlled during the start-up of an U A S B reactor. The permissible S L R for a cheese whey anaerobic system was 0.25 g C O D / g V S S - d . 5.2.2 S t a r t - u p P e r f o r m a n c e This part of the study presents the behavior of the reactor which was initially fed the low strength of 5 g C O D / 1 at H R T of 5 days. The first 48 days were considered as the start-up period during which two influent concentrations (5 and 9.93 g COD/1) were used. The U A S B reactor performance during the start-up period is shown in Figures 5.1 to 5.8. Figure 5.3 shows the effluent C O D concentration vs time of reactor operation. The Chapter 5. START-UP AND EFFECTS OF PROCESS PARAMETERS 56 C O D concentration decreased significantly with the length of time of operation. On day 40, in spite of high influent concentration of 17.7 g C O D / 1 , effluent C O D was reduced from 1500 to 110 mg/1 (Figure 5.3). The same trend of increasing efficiency in terms of C O D removal and gas production was also observed, as shown in Figures 5.2 and 5.3. Within the first 40 days, the C O D removal efficiency increased from 70 to 97%. In the meantime, biogas composition increased from 48 to 57% methane and the gas production rate reached a value of 2.5 liters CHu per liter feed per day within 15 days and remained at approximately the same level under the same loading rate. Comparing C O D and V F A in the effluent and gas production, it could be noticed that the lowest effluent V F A , the highest methane content in the produced biogas and the highest C O D removal were reached after 40 days of operation. However, the highest gas production rate was reached much earlier at day 15 (Figure 5.9). The effluent VSS concentrations are shown in Figure 5.8. Due to the poor settling quality of the seed sludge, a large amount of sludge left the reactor at the beginning of the operation. After 15 days of operation, the amount of sludge remaining in the reactor was reduced from 86 to 60 g V S S . At day 15, 67.2 g VSS of seed sludge were added to the reactor to maintain the specific activity of the sludge. As a result of natural selection and sludge particle growth, the sludge settleability improved gradually. At day 30, the TSS content in the effluent was as low as 0.1 g/1. Many factors have been shown to be associated with the start-up of the U A S B reactor and its later performance. Attention has also been directed to the use of different seed sludge in order to shorten the start-up period. Starting with a poor quality digested sewage sludge, De Zeeuw (1985) was able to cultivate a highly active biomass (specific activity of 0.75 g C O D / g VSS d) within a period of 6 weeks (organic loading rate of 7.5 g C O D / 1 d). Compared to the sewage sludge used by De Zeeuw , the seed used in this study had a Chapter 5. START-UP AND EFFECTS OF PROCESS PARAMETERS 57 TJ 8 ~ Q O O o 0 1 A TJ <0 O _J o c « a 6 a -1 .vj •> , , 1 1 ' 1 • 1 ' — —1 ' 1 ' 1 ' 10 20 30 40 SO Time (days) eo 70 80 Figure 5.1: O r g a n i c L o a d i n g R a t e versus T i m e Chapter 5. START-UP AND EFFECTS OF PROCESS PARAMETERS 58 32 28 H 24 H S 20 Q O c C H ^ O - O O - C Cv-<><>0 O-CD c o i 10 20 30 40 50 Time (days) eo 70 80 Figure 5.2: Influent C O D versus Time Chapter 5. START-UP AND EFFECTS OF PROCESS PARAMETERS 59 Time (days) Figure 5.3: Effluent C O D versus T i m e Chapter 5. START-UP AND EFFECTS OF PROCESS PARAMETERS Time (days) Figure 5.4: C O D R e m o v a l versus T i m e Chapter 5. START-UP AND EFFECTS OF PROCESS PARAMETERS 61 12 TJ © « / « CC u T J o o c « o 2 • ) -r~ 10 - 1 — 20 -1— SO 40 —I— 60 eo - r -70 eo Tim« (day8) Figure 5.5: M e t h a n e P r o d u c t i o n R a t e versus T i m e Chapter 5. START-UP AND EFFECTS OF PROCESS PARAMETERS 62 60 ^ 55 H C O co 50 O a E o O -4-* c o J D H — UJ Time (days) Figure 5.8: Volatile Suspended Solid (VSS) in the Effluent Chapters. START-UP AND EFFECTS OF PROCESS PARAMETERS 65 Z O 90-O D Q £ 80 Q O ° 70 CQ0FB3UDCN A / • / 56 U I 50 48 says was** A / • • V XI 25 P I O 15 1 icnwcmouacN —1 1 T 1 1 0 V 20 30 40 50 600 V 4001 I £ 200 \ « \ fas 1 1 1 1 1 1 0 O 20 30 40 50 M M Figure 5.9: C o m p a r i s o n of C O D , V F A a n d Biogas P r o d u c t i o n Chapter 5. START-UP AND EFFECTS OF PROCESS PARAMETERS 66 lower VSS content and poorer settleability. This was demonstrated by the loss of a large amount of sludge (Figure 5.8), high V F A concentrations and low gas production at the beginning of the start-up. However, the seed was able to degrade acetate and propionate resulting in very low effluent V F A concentrations within a period of 40 days, reaching 0.71 g C O D / g V S S - d of specific activity after an operation period of 70 days. 5.2.3 V F A B e h a v i o r in Reac tor Start -up Figure 5.7 shows the daily V F A s concentrations in the effluent of the reactor. It was found that V F A s were sensitive parameters which responded to loading change and reactor performance. At the very beginning of the reactor operation, V F A s in the effluent were fairly high due to the lower activity and concentration of the sludge at that time. Then they were gradually decreased. A n increase in loading rate increased the V F A concentration. At day 30, the peak in acetic acid concentration was a response to the increase of influent concentration from 4.56 to 9.9 g C O D / 1 . As long as the sludge was acclimated and had built up enough concentration and achieved a high specific activity, the V F A in the effluent would decrease to a very low level. In this case, after 45 days from start, the total V F A in the effluent was as low as about 40. Although at days 45 and 60 when the influent concentration was increased from 10 to 18 g C O D / 1 and from 18 to 28 g C O D / 1 respectively, there was no peak observed. It is very important during the start-up of a U A S B reactor that the O L R be increased only after the majority of V F A is removed. At day 12, when an attempt was made to decrease the H R T from 5 days to 4, a sudden increase in V F A was detected. By using V F A concentration as a sensitive indicator while varying the start-up parameters it has been shown that 0.2-0.25 g C O D / g VSS would be recommended for start-up of the reactor. Chapter 5. START-UP AND EFFECTS OF PROCESS PARAMETERS 67 5.3 S T E A D Y S T A T E O P E R A T I O N The steady-state performance of whey digestion as a function of influent concentration is summarized in Table 5.2. A C O D removal efficiency of over 97% was maintained after 45 days of reactor operation. A 98% C O D reduction with a gas production rate of 9.57 1 CH4 / l feed-d was achieved at a loading rate of 5.96 g C O D / 1 d and an influent concentration of 28.8 g C O D / 1 . Figure 5-10 illustrates the quality of effluent as a function of influent concentration. In spite of the high influent C O D concentration of 28.8 g C O D / 1 , the effluent C O D concentrations were between 400 and 500 mg/1. Effluent p H increased with an increase in influent concentration. V F A contents of the effluent were in the range of 76-16 mg/1. A comparison of the behavior of a U A S B reactor with a fixed-film reactor for cheese whey treatment showed that the growth rate and bacterial activity were somewhat dif-ferent for the suspended-growth system ( U A S B system) and the attached-growth system (fixed-film) (Lettinga et al 1979, Kelly and Switzenbaum 1984, Nordstedt and Thomas 1984, Wildenauer and Winter 1985, L O and Lioa 1986). Lower effluent V F A concentra-tions were reported for the U A S B reactor than for the fixed-film system. The effluent V F A concentration has proven to be a good indicator for the condition of an anaerobic reactor. Such a comparison revealed that a higher activity of biomass was developed in the U A S B reactor than in the attached growth system with respect to the degradation of V F A . C O D balances were determined for the four experimental conditions. A mass balance of the system enables the sludge formed in the reactor to be calculated (Table 5.3). The observed yields calculated from the C O D balance over the whole experimental period were in the range of 0.098 to 0.166 g T S S formed/g C O D removed. This appears to be within a reasonable range for this type of waste which is primarily carbohydrate Chapter 5. START-UP AND EFFECTS OF PROCESS PARAMETERS 68 Table 5.2: E x p e r i m e n t a l Resul ts in Steady State P e r f o r m a n c e Influent g C O D / 1 O L R g C O D / 1 d Effluent C O D p H mg/1 V F A mg/1 C O D removal % 1 C H 4 / 1 reactor d G A S l C f W g C O D d C H 4 % 4.56 0.91 120 6.75 35 98 1.37 0.30 56.5 7.30 1.36 215 6.92 56 97 2.35 0.32 51.5 9.93 1.97 191 7.08 16 98 3.25 0.33 48.2 17.70 3.54 286 7.05 18 99 5.80 0.33 47.0 21.50 4.23 348 7.10 20 99 7.14 0.34 46.5 28.80 5.96 457 7.18 18 99 9.57 0.33 46.1 38.10 7.77 505 7.26 20 99 11.20 0.30 42.5 Figure 5.10: T h e Quality of Effluent as a Function of Influent Concentration (at H R T = 5 days) Chapter 5. START-UP AND EFFECTS OF PROCESS PARAMETERS 70 (Switzenbaum and Danskin 1982, Henze and Harremoes 1982). 5.4 U N S T E A D Y - S T A T E P E R F O R M A N C E The process became unstable 14 days after the reactor was fed influent at a concentration of 38.1 g C O D / 1 . This was first noticed by a decrease in gas production from 70 to 61 1/d and an increase in effluent acetic acid and propionic acid concentrations to 80 and 64 mg/1, respectively, following an increase of effluent C O D from 500 to 643 mg/1. The influent concentration was decreased to 28.5 g C O D / 1 until steady-state was re-established in 20 days and then increased to 41.1 g C O D / 1 . A n instability appeared again and finally the reactor was entirely upset. This apparently indicated that the system was overloaded and the influent strength of about 38 g C O D / 1 appears to be a barrier or threshold for this system. Similar findings were reported by Switzenbaum and Danskin (1982) when they treated cheese whey by using an expanded bed reactor. The C O D reduction was decreased from 83% to 58 % and the process became unstable when they increased the influent concentration from 5 to 20 g C O D / 1 (Switzenbaum et al 1982). It was suggested that in treating high strength whey the physical balance between the methanogens and the hydrogen and acid-producing bacteria was more easily upset in this system. This will be the main topic of this thesis and will receive further consideration in a later chapter. 5.5 G A S P R O D U C T I O N 5.5.1 T h e Effect of Influent C o n c e n t r a t i o n In this part of the study, the reactor was operated at a constant H R T of 5 days, and the influent C O D concentration, and therefore O L R , was varied. Biogas production rate and Table 5.3 C O D Balance (/) (2) (3) W (5) «$) (7) (8) ,(9) Influent Influent Effluent Effluent CH\ production Ctft in effluent Total COD out Sludge formed Sludge accumulation factor concentration {mgCODd'1) concentration {mgCODd-') {mgCODd'') (mgCODd'1) {mgCODd-1) (mgCODd-1) (mg VSS formed/ (mgCODr*) (mgCODl-') mg COD removed) . 4 780 13623 132 376 9885 235 10496 3 127 0166 9 934 28312 191 544 23447 235 24228 4084 0104 17725 50515 286 815 41848 235 42898 7618 0108 28 700 81795 457 1304 69049 235 70 587 17 208 0098 The above parameters were calculated as follows: (2) = 2-85 * (1), 2-85(1 .d) is the amount of feed (4) = 2-85*(3) (5) 1 g COD = 395 ml CH* at 32°C (6) assume effluent is saturated with C H 4 at 32-8ml C H . L (7) = (4)+ (5)+ (6) (8) = (2)-(7) (9) = (8)/[(2) - (4)]. Assume 10 mg solid = 1-42 mg COD Chapter 5. START-UP AND EFFECTS OF PROCESS PARAMETERS 72 methane composition are presented in Table 5.4. A plot of gas production rate in terms of liters of methane produced per gram C O D added per day as a function of the influent concentration is shown in Figure 5.11. The production rate of methane was increased with increasing influent C O D concentration up to a concentration of 28.8 g C O D / 1 , then the trend reversed. The methane production rates ranged between 0.219 and 0.313 1/g C O D d. Figure 5.12 demonstrates the effect of influent concentration on the methane com-position. It has been believed that gas composition is a function of the nature of the biodegradable portion of the feed. It would be an interesting study to relate the biodegradable portion of substrates to the methane composition. Degradation of car-bohydrates yields C02 and CH4 in equal quantities. A higher proportion of CH4 is generated from proteins and fatty substances, and this may reach a level as high as 75% methane and 25% carbon dioxide. A high C 0 2 content in the gas phase in this sys-tem (up to 45% and higher) indicates the domination of acidogenic fermentation over methanogenesis. In general, a lower influent C O D concentration yields a higher methane composition in the produced biogas. This implies that the dominance of acidogenesis over methano-genesis is enhanced by an increase in substrate strength in the cheese whey anaerobic process. Therefore, it could also be possible to develop a useful technique based on this phenomena for quickly determining stability or to monitor any stress in an anaerobic sys-tem from the spectrum of produced biogas composition since the methane composition is such a sensitive parameter which responds to any environmental change and reflects the reactor operation. The methane in the produced biogas rapidly diminished from 56% down to 48% with an increase in influent concentration up to a certain level (9.9 g COD/1) . There was little change in biogas methane composition with a further increase in influent concentration Influent Organic loading Gas Gas production concentration rate composition (g COD litre-') (g COD litre'1 (CH4%) (litres biogas (litres biogas (litres CH, day-') litre'' day'1) g-' COD day'1) g'1 COD day'1) 4-56 091 56-1 2-45 0-537 0-282 7-30 1-36 51-5 4-60 0-630 0-302 9.94 1-97 48-2 6-73 0-677 0-304 17-7 3-56 47-0 12-4 0-647 0-305 21-5 4-20 46-6 15-3 0-714 0-310 28-8 5-96 46-1 20-8 0-721 0-311 38-1 7-77 42-5 265 0-705 0-300-41-1 8-14 41-0 23-9 0-581 0-222 Or CO T a b l e : 5.4 jjj Gas Production Rate and Composition for Different Influent Concentration at 5 days HRT ^ I o O I co ? I I 8 Chapter 5. START-UP AND EFFECTS OF PROCESS PARAMETERS 74 Figure 5.11: Effect of Influent Concentration on Methane Production Chapter 5. START-UP AND EFFECTS OF PROCESS PARAMETERS Figure 5.12: Effect of Influent Concentration on Biogas Composition Chapter 5. START-UP AND EFFECTS OF PROCESS PARAMETERS 76 to 28.8 g C O D / 1 . The methane composition decreased when the influent C O D was raised to more than 28.8 g C O D / 1 . It showed that no simple linear relationship exists between the gas composition and the influent concentration. The average methane production rate was about 0.32 1/g C O D - d . The methane compositions decreased after the influent concentration was raised to more than 28.8 g C O D / 1 (loading rate great than 6 g C O D / l - d ) . At this loading rate, instability was observed. The highest methane production rate of 9.57 1 CH4 / l feed d was obtained at a loading rate of 6 g C O D / 1 d and an influent concentration of 28.8 g C O D / 1 . 5.5.2 T h e Effect of H y d r a u l i c R e t e n t i o n T i m e ( H R T ) In this part of the study, the U A S B reactor was first tested at an influent concentration of 20.5 g C O D / 1 , while the operating H R T was varied from 24.7 to 4.9 days. The experiment was then conducted over the same range of H R T s with a higher feed concentration (41.1 g COD/1) . The methane production rates are summarized in Tables 5.5 and 5.6. The methane production rate was increased with increasing H R T , which agrees with previous research (Preffer, 1974). Figure 5.13 shows the relationship between methane production rate and H R T . This study gave almost the same results as were reported by Eckenfelder and Ford that at 22.5 days H R T , the methane production rate was almost twice that at 8.8 days H R T . A more interesting finding was that there is an interaction between H R T and feed strength.. In other words, the effect of H R T was related to influent concentration. At HRTs longer than 10 days, the methane production rate was similar for two influent con-centrations as shown in Figure 5.13. At short H R T s , the effect of H R T on methane pro-duction rate was more pronounced for the higher influent concentration (41.1 g COD/1) than for the lower influent concentration (20.5 g COD/1) . The methane production rate Chapter 5. START-UP AND EFFECTS OF PROCESS PARAMETERS 77 decreased slightly from 0.356 to 0.341 1 CH^/g C O D - d at an influent concentration of 20.5 g C O D / 1 when H R T was decreased from 10.1 to 6.8 days. However, under the same range of operating H R T , a sharp decrease in methane production rate from 0.376 to 0.249 1 CHA/g C O D - d was observed for tests with the influent concentration of 41.1 g C O D / 1 . The results also indicated that for a given H R T , a higher methane production rate (1 CH4/g C O D d) was obtained at a lower influent C O D concentration at HRTs shorter than 10 days. The effect of H R T on methane production rate can be represented by the slopes of the linear equations. for Co=20.5 g C O D / 1 Gp = 0.0090 + 0.292 (5.1) for Co=41.1 g C O D / 1 , 9 > 10 days Gp = 0.0100 + 0.286 (5.2) for Co=41.1 g C O D / 1 , 9 < 10 days Gp = 0.0270 + 0.107 (5.3) where $=hydraulic retention time, days Co=influent concentration, g C O D / 1 Gp=methane production rate, 1 CH4/g C O D d Equations (5.1) and (5.2) showed similar slopes and intercepts at HRTs longer than 10 days for both influent concentrations. Equation (5.3) has a slope 2.7 times greater than Table 5.5 Methane Production Rate at an Influent Concentration of 20-5 g C O D litre"1 HRT Influent Organic loading Gas Gas production (days) concentration rate composition (gCOD litre'') (g COD litre-' day-') (CH4%) (litresgas litre-' (litres gas g~1 (litres CH4g-' feed day'') COD day-') COD day-') 24-7 18-2 0-74 50-6 190 1-021 0-492 23-1 18-2 0-79 50-6 19-2 1-054 0-497 16-5 20-2 1-23 48-5 18-4 0-909 0-411 10-1 20-0 1-98 47-4 15-9 0-807 0-356 6-8 20-5 3-01 47-0 160 0-780 0-341 5-2 21-0 4-08 46-5 154 0-731 0-317 Table 5.6 Methane Production Rate at an Influent Concentration of 41-1 g C O D litre" 1 HRT (days) Influent concentration (g,COD litre'') Organic loading rate (g COD litre'1 day'1) Gas composition (CH4%) (litresgas litre'1 feed day'') Gas production (litres gas g'1 COD day'') (litres CH4 g'' COD day'1) 23-9 41-1 1-73 49-7 • 43-3 1-053 0-490 10-8 4 M 3-82 46-5 34-5 0-846 0-367 7-1 4 M 5-84 44-7 27-1 0-658 0-274 6-0 41-1 6-91 43-2 25-2 0-618 0-249 5-0 41-1 8-21 41-0 23-9 0-581 0-222 Chapter 5. START-UP AND EFFECTS OF PROCESS PARAMETERS Hydraulic Retention Tim* (days) Figure 5.13: Effect of H R T on Methane Production Chapter 5. START-UP AND EFFECTS OF PROCESS PARAMETERS 81 that of equation (5.2), and the methane production rates decreased rapidly when the reactor was fed with an influent concentration of 41.1 g C O D / 1 . This dramatic decrease in the methane production rate indicates some inhibition of the methanogenesis, a fact supported by the V F A s data in Table 5.7. The concentration of V F A s generally increases with a decrease in H R T or an increase in organic loading rate. There was a surge of propionic acid when the H R T was decreased from 10 to 7 days. As the H R T was further decreased, propionic acid concentration increased, followed by an increase in acetic acid. The effect of H R T on methane composition is presented in Figure 5.14. As with biogas production, the methane in the biogas decreased as the H R T decreased. 5.5.3 T h e Effect of O r g a n i c L o a d i n g R a t e ( O L R ) Figure 5.15 shows that at a constant influent C O D concentration the methane production rate decreased with an increase in organic loading rate (a decrease of H R T ) . At an O L R less than 4 g C O D / 1 d, a higher influent concentration resulted in a higher methane production rate. Longer HRTs led to higher methane production rates at a similar O L R (Table 5.8). When the O L R was greater than 4 g C O D / 1 d, an increase in influent concentration or a decrease in H R T resulted in a decrease in the methane production rate. The observed effects of O L R on biogas production were in good agreement with the findings of a number of investigators. For example, Varel et al had quite similar experimental results. Studying methane production from beef cattle waste, Varel et al (1980) found that for a given VS loading rate, higher methane production rates are possible at higher influent V S concentrations and longer H R T up to an influent V S concentration of about 8%. At a constant H R T of 5 days, a change in influent concentration resulted in a small change of methane production rate, which remained relatively constant up to an influent Table 5.7 V F A in the Reactor and Effluent at Differential H R T Co § o o ^ ' o co CO 2 Influent COD HRT (g COD litre-') (days) VFA at 60 cm in the reactor (mg litre'') VFA in the effluent (mg litre'') Acetic Propionic Iso-Butyric Butyric Acetic Propionic Iso-Butyric Butyric COD Removal (%) 20-5 23-8 18 0 0 0 10 0 0 0 99 .16-4 20 0 0 0 5 0 0 0 99 101 23 0 0 0 15 0 0 0 98 6-8 22 2 0 0 26 0 0 0 98 5-8 45 14 0 0 10 0 0 0 98 41-1 23-8 29 20 4 0 10 3 0 0 99 10-4 278 762 0 0 6 0 0 0 95 6-8 266 1204 0 0 54 838 0 0 92 j 5-9 993 1636 168 809 398 1252 111 0 86 5-0 556 2231 76 158 208 1793 57 0 81 00 Chapter 5. START-UP AND EFFECTS OF PROCESS PARAMETERS 83 55 o c o 52-49-• influent 20.5 g COD I ' 1 to o a E o O co co (5 46-43- O 40-12 i 16 — r ~ 20 -r~ 24 Hydraulic Retention Time (days) Figure 5.14: Effect of H R T o n B i o g a s C o m p o s i t i o n •s Table 5.8 to ? Effects of Influent Concentration and H R T on Gas Production Rate for Similar Organic Loading Rate ^ § a o o I co CO 2 Organic loading rate (g COD litre-' day'1) Influent concentration (g COD litre'1) HRT (days) Gas composition (CH4%) Gas production (litres gasg-' COD day-') (litres CH4 g ' 1 COD day '') 0-90 4-56 5-0 56-3 0-537 0-302 0-79 18-2 23-0 50-6 1-054 0-533 1-36 7-30 5-0 51-5 0-630 0-324 1-23 20-2 16-4 48-5 0-888 0-431 1-92 9-85 5-0 48-2 0-678 0-327 1-97 200 10-1 47-4 0-806 0-382 1-73 41-1 23-9 49-7 1-053 0-524 4-08 21-0 5-1 46-5 0-731 0-340 3-82 41-1 10-8 46-6 0-846 0-394 5-96 28-1 5-0 46-1 0-723 0-333 5-84 411 7-1 44-7 0-658 0-294 2El oo Chapter 5. START-UP AND EFFECTS OF PROCESS PARAMETERS 85 concentration of 28.8 g C O D / 1 , then decreased (see Table 5.5 and Figure 5.16). The highest methane production rate of 9.57 1 CH4/\ feed-d (0.33 1 CH4/g C O D ) was obtained at an O L R of 5.96 g C O D / l - d and influent concentration of 28.8 g C O D / 1 (Table 5.2). These results imply that the reactor can be operated at short H R T s , as long as the O L R does not exceed 6 g C O D / l - d . A simple linear relationship between O L R and methane production in terms of liters CH4 per liter reactor per day was found in O L R range of 2 to 8 g C O D / l - d (Figure 5.16) 5.6 T R E A T M E N T E F F I C I E N C Y The highest C O D removal efficiency for raw whey treated in a chemostat (Clanton et al. 1985) was only 58% for a completely-mixed anaerobic reactor. Wildenauer and Winter (1985) achieved a C O D removal of 95%, close to the results of this study with the help of pH-control, which maintained the p H at about 6.7. Without p H control in the whey digestion process, lower organic loading rates were usually applied to the reactor and relatively lower C O D reductions were obtained as indicated in Table 8-2. In this study, the U A S B process accommodated fairly well to whey strength up to 28.8 g C O D / 1 and maintained a C O D reduction of over 97% . These results indicated that the U A S B reactor can tolerate a higher wastewater strength although this type reactor was supposed to be best for dilute wastes, typically under 20 g C O D / 1 and at moderate O L R . Chapter 5. START-UP AND EFFECTS OF PROCESS PARAMETERS 86 Figure 5.15: Effect of Organic Loading Rate on Methane Production Chapter 5. START-UP AND EFFECTS OF PROCESS PARAMETERS 87 Figure 5.16: Relation Between Methane Production and O L R Chapter 5. START-UP AND EFFECTS OF PROCESS PARAMETERS 88 5.7 S U M M A R Y 1. The performance of the U A S B reactor was evaluated in the range of 4.5 to 38.8 g C O D / 1 of cheese whey influent concentrations at a 5 days hydraulic retention time. The results of the present experiments have shown that anaerobic digestion using a U A S B reactor can be an efficient treatment method for diluted cheese whey. Over 97% C O D removal was achieved. Without p H control and nutritional addition, the system could treat cheese whey successfully up to a concentration of about 29 g C O D / 1 . 2. It is known that reactor performance strongly depends on the start-up procedure. Various start-up strategies were used to facilitate the start-up of an U A S B reactor and ensure stable operation. Among the operating parameters, sludge loading rate was the most important during start-up. In the very beginning, the initial sludge loading should not exceed 0.25 g C O D / g VSS d. 3. V F A s in the effluent were found to be a very useful indicator for primary adaptation of the sludge and for monitoring the system's stability. The loading rate should be increased only after the V F A concentrations are greatly reduced . Too rapid an increase of O L R may result in the loss of activity of the sludge. 4. The methane production rate increased with increasing H R T . At HRTs longer than 10 days, the methane production rate was similar for two influent concentrations. At short H R T s , the effect of H R T on methane production rate was more pronounced for the higher influent concentration (41.1 g COD/1) than for the lower influent concentration (20.5 g COD/1) . For a given H R T , a higher methane production rate (1 CH4/g C O D - d ) was obtained at a lower influent C O D concentration and at those HRTs shorter than 10 days. From the point of view of fuel gas production, the organic loading rate was a critical parameter. At an O L R less than 4 g C O D / l - d , the reactor fed with a higher influent concentration yielded a higher methane production rate. When O L R was greater Chapter 5. START-UP AND EFFECTS OF PROCESS PARAMETERS 89 than 6 g C O D / l - d , the higher influent concentration or shorter H R T produced a lower methane production rate. Therefore, the optimal O L R for this particular system with regard to methane production would be between 4-6 g C O D / l - d . For a H R T of 5 days, the optimal influent concentration should be between 20-30 g C O D / 1 . 5. When the influent concentration was increased to 38 g C O D / 1 , instability was observed. For this system the influent concentration should be maintained below 30 g C O D / 1 at H R T of 5 days. Further studies were emphasize a search for the reasons for the inhibition caused by high influent concentrations and to seek an efficient way to prevent instabilities. Chapter 6 DISTRIBUTIONS OF S L U D G E A N D S U B S T R A T E S 6.1 I N T R O D U C T I O N The U A S B process has been widely investigated since it was developed. A great number of research papers which have been published each year reveal that it is an effective process for anaerobic wastewater treatment.(Lettinga et al.1985, 1984, 1983, 1979; Wang et al. 1985; W u et al. 1985; Sayed et al.1984). The treatment of cheese whey in a U A S B reactor has further demonstrated that the easily acidified substrates, such as cheese whey, can be treated by using an U A S B reactor, if the proper start-up procedure is used. Satisfactory treatment efficiencies have been obtained (Yan et al. 1989). In order to describe and optimize the U A S B process, various models have been pro-posed to explain the kinetics of anaerobic digestion in a U A S B reactor, which include the fluid flow pattern, the kinetics of substrate conversion and bacterial growth, and the sludge distribution and behavior in the reactor (Bolle et al.1986a, 1986b; Buijs et al.1982, 1980; Heertjes et al.1978,1982; Ven der Meer et al. 1983). However, the substrate dis-tributions in the reactor have not been reported, in spite of their obvious importance for better understanding and for optimizing the process. Such a procedure is time and effort consuming and can be tedious because of the huge number of experiments and analyses that must be performed to obtain the details of the profiles. This study was motivated by the surprising shortage of data about the substrate distribution which is needed for modelling and optimization. 90 Chapter 6. DISTRIBUTIONS OF SLUDGE AND SUBSTRATES 91 In this part of the research, therefore, an extensive study of the profiles in a U A S B reactor, i.e.the distribution of the sludge, C O D , V F A s and pH under a wide spectrum of operating conditions was conducted. There were 5 different operating conditions, in which the influent C O D was increased stepwise from 5 to 40 g C O D / 1 at a H R T of 5 days. For each operating condition, samples were taken from the influent, the corresponding effluent and 10 sampling ports mounted along the column of the reactor. This study was originally designed, as was stated above, to gain detailed information about the sludge and substrate distributions in the U A S B reactor and to serve as the basis of future modelling research, and to increase the understanding of the U A S B process. It was also expected to provide information about the cause of the instability of the process which can result in a rapid upset. As these experiments progressed, it was found that the profiles of the substrates truly did provide much information about the instability mechanisms, which will be discussed in this chapter. The results which' are summarized in Table 6.1 have shown definitively that two reaction stages, i.e. acidogenesis and methanogenesis could be distinguished in different regions in the same reactor. 6.2 RESULTS 6.2.1 Distribution and Behaviour of the Sludge The profiles of sludge concentration at different influent concentrations are presented in Figure 6.1. The curves show that two sludge regions exist in the reactor. The dense sludge phase was retained in the lowest part, below 30 cm from the bottom, constituting a sludge bed with VSS of 18 to 56 g/1. In this zone, the sludge concentration varied with the location. For example, at 4 cm height, the sludge content was 35 g VSS/1 for an influent of 28.8 g C O D / 1 , while the sludge concentration was 25 g VSS/1 at 25 cm height at the same influent concentration. Above 37 cm height (sampling port 4), there was Chapter 6. DISTRIBUTIONS OF SLUDGE AND SUBSTRATES 92 a sludge blanket. The average VSS concentration in this area ranged from 2 to 10.g/1, depending on the loading rate. Unlike the sludge in the bed, the sludge concentration in the blanket was constant and quite homogeneously distributed. As indicated in Fig.6.1, the plots of sludge concentrations in the blanket (above 38 cm) vs. the reactor height were horizontal lines parallel to the x axis. The results indicated that, in the course of the experiment, the sludge in the blanket was completely mixed, but the sludge in the bed was not well mixed. The mixing, which was brought about by gas evolution, might be insufficient under these experimental conditions in the lower zone, although the biogas production was as high as 75 1/d which equals 0.30 m 3 / m 2 - h . Generally, when the organic loading was increased, the sludge bed and blanket both expanded upward, as the result of gas lift and sludge settleability deterioration. As loading was being increased, more gas was produced, creating more flotation to lift the sludge upward. At the lowest influent concentration of 4.5 g C O D / 1 , the sludge bed occupied as little as one eighth of the total working volume of the reactor (below 13 cm from the bottom). W i t h an increase in the influent concentration, mainly as the result of the increase in the gas production, the sludge bed extended to the upper part of the reactor. The sludge bed reached a height of 25 cm when the influent concentration was increased to 9.93 g C O D / 1 . When the loading rate was increased to such an extent that it exceeded the sludge digestive capacity, the sludge would not be able to maintain its vitality. In effect, its settleability deteriorated. Hereafter, poorly settled sludge tended to be suspended. It was particularly true in the case of an influent concentration of 38.1 g C O D / 1 . A large amount of floating sludge was collected in the effluent at this time. It was common when the reactor was subjected to a new higher organic loading rate, that the increase of gas production was followed by a surge of sludge wash out due to the sudden increase in floatation. If the new operating conditions were within the capacity of the reactor, a new adaptation and stabilization would be re-established, resulting in a Chapter 6. THE DISTRIBUTIONS OF SLUDGE AND SUBSTRATES 93 Table 6.1: the Distribution of Sludge and Substrates Influent Height VSS COD P H AA PA BA COD g COD/I cm g/1 mg/1 mg/1 mg/1 mg/l Removal 4.56 4.0 26.18 3390 382 92 480 0.26 12.5 19.04 800 152 60 58 0.82 25.0 5.42 570 162 100 0 0.88 37.5 1.96 410 50 56 0 0.91 50.0 1.69 400 45 52 0 0.91 62.5 1.51 370 48 56 0 0.92 87.5 1.42 370 120 60 0 0.92 112.5 1.39 330 112 44 0 0.93 9.93 4.0 35.92 7520 4.52 688 134 1140 0.24 12.5 18.18 330 6.40 96 14 14 0.97 25.0 14.38 290 6.48 60 10 0 0.97 37.5 5.23 240 6.68 194 0 0 0.97 50.0 3.96 230 6.65 96 0 0 0.98 62.5 3.16 220 6.69 338 0 0 0.98 87.5 2.76 220 6.68 82 0 0 0.98 112.5 2.65 220 6.66 30 0 0 0.98 17.70 4.0 31.12 10940 4.40 876 290 1454 0.38 12.5 20.50 560 6.50 20 3 0 0.97 25.0 18:55 590 6.54 22 3 0 0.97 37.5 6.19 350 6.72 32 0 0 0.98 50.0 4.96 380 6.72 20 0 0 0.98 62.5 4.74 400 6.70 24 0 0 0.98 87.5 4.73 370 6.71 16 0 0 0.98 112.5 4.59 340 6.68 18 0 0 0.98 28.80 4.0 35.39 15300 4.50 1166 748 0.49 12.5 32.11 970 6.90 40 36 0.97 25.0 25.07 700 7.05 18 34 0.97 37.5 7.62 520 7.10 20 0 0.98 50.0 9.35 500 7.10 18 0 0.98 62.5 7.77 480 7.10 16 0 0.98 87.5 7.60 480 7.10 22 0 0.98 112.5 7.87 440 7.15 18 0 0.98 38.10 4.0 56.56 31500 2.89 2685 231 243 0.17 12.5 38.20 35250 3.12 2895 468 255 0.07 25.0 39.77 33240 3.12 2883 537 495 0.13 37.5 31.43 2640 6.70 702 1336 108 0.93 50.0 3.38 2760 6.70 596 1192 160 0.93 62.5 4.84 2760 6.70 396 846 44 0.93 87.5 3.26 2790 6.75 376 722 2 0.91 112.5 2.69 2690 7.20 262 694 20 0.91 Chapter 6. THE DISTRIBUTIONS OF SLUDGE AND SUBSTRATES 94 70 Figure 6.1: Profiles of Sludge at Different Influent Concentrations Chapter 6. DISTRIBUTIONS OF SLUDGE AND SUBSTRATES 95 new balance between sludge and substrates. Figures 6.2 and 6.3, which show the concentration of the sludge at each location for different influent concentrations, illustrate that the sludge moved upward with an increase in the influent concentration. For example, the portion of sludge at 4 cm height decreased from 41% for the influent concentration of 4.56 g C O D / 1 to 29% for an influent C O D of 28.8 g/1, which is different from the higher location of the reactor, at 25 cm height, rhere the sludge increased from 9% to 19% for the same range of influent concentration change. The relation between the sludge distribution and the gas production is shown in Figure 6.4. and Table 6.2. The sludge concentration in the bed did not change significantly with gas production, which is indicated by the VSS concentration at sampling ports 1 and 2 in Figure 6.4a, while the sludge concentration in the blanket varied with the gas production (Figure 6.4c). The most significant effect of gas production on the sludge distribution was observed in the area of sampling port 3, which is at the junction between the bed and the blanket. Good linear relations between the sludge and gas production in the blanket, which are shown in Figure 6-4c, were fitted with the following equations: for the area between the bed and the blanket (Figure 6.4b) X = 8.710 + 5.30 (6.1) for the blanket (Figure 6.4c) X = 3.50 + 0.7 (6.2) where X is the sludge. concentration in g VSS/1 and 0 is the biogas production 1/d. The greater slope of equation 6.1 compared to equation 6.2 indicates the more significant Chapter 6. THE DISTRIBUTIONS OF SLUDGE AND SUBSTRATES 96 4«-i Reactor Height (cm) Figure 6.2: Sludge Distribution in the U A S B Reactor Chapter 6. THE DISTRIBUTIONS OF SLUDGE AND SUBSTRATES 97 40-i Reactor Height (cm) Figure 6.3: Sludge Concentration in the U A S B Reactor Chapter 6. THE DISTRIBUTIONS OF SLUDGE AND SUBSTRATES 98 iS) > o o 3 TJ A: Sludge in Bed • somple port $1 O somple port #2 •r io CO CO > c o o c c o o> 3 6H ^ o C: Sludge in the Blanket 9 « ° ™ P * * p o r t #4 ^ ) s o m p l o p o r t #5 O l o m p l * p e r l #6 7^ » o m P * * P o r t #8 ^ » o m p l * p o r t #10 10 20 30 40 50 60 70 Gas Production ( I d ) Figure 6.4:! Sludge Distribition versus the Gas Production 8 Table 6.2: Relation between Sludge Distribution and the Biogas Input g C O d / l , Biogas l/d ' 1# 2# SludgeConcentration g v s s / i 3# 4# 5# 6# 8# 10# 4.56 6.96 26.18 19.40 5.43 1.96 1.69 1.51 1.42 1.39 9.93 19.11 35.92 18.18 14.38 5.23 3.96 3.16 2.75 2.65 17.73 35.10 34.58 20.50 18.55 6.19 4.96 4.74 4.73 4.59 28.70 59.02 35.39 32.11 25.07 7.62 9.35 7.77 7.77 7.80 1 I 3 O § Co CO to Chapter 6. DISTRIBUTIONS OF SLUDGE AND SUBSTRATES 100 effect of gas flow rate on the sludge distribution in the area between the bed and the blanket than in the blanket. 6.2.2 G r o w t h of Sludge The total amount of sludge in the reactor was determined by sludge profiles over the height of the reactor for each experimental condition. There was a net increase in VSS concentration in the reactor with loading rate as is shown in Table 6.3. The small amount of biomass formation at the beginning was attributed to the low food/sludge ratio of 0.164 g C O D / g VSS d. A significant increase in VSS in the lower regions of the reactor took place only after the influent concentration was increased to 17.7 g C O D / 1 , corresponding to a sludge loading rate of 0.547 g C O D / g V S S . These results together with the process operating conditions, which are presented in Table 6.3, provided the possibility of calculating the biomass yield coefficient Y and the decay constant K d using the following model: where Y i s the yield coefficient in terms of g V S S / g C O D , X is the sludge concentration in the reactor in term of g VSS/1 and K d is the decay constant of the biomass. A plot of ( d X / d t ) / X against ( d F / d t ) / X gave a straight line with Rsq of 0.8 (Figure 6.5), from which the yield coefficient Y and the decay constant K d can be derived (see Table E . l ) . They are 0.058 g V S S / g C O D and 0.02 d'1, respectively. The observed yields calculated from the C O D balance (see chapter 5) over the whole (6.3) Chapter 6. DISTRIBUTIONS OF SLUDGE AND SUBSTRATES 101 experimental period (0.098 to 0.166 g TSS formed/g C O D removed) agree quite well with the sludge growth yield of 0.058 g V S S / g C O D as estimated from the sludge growth kinetics.(The ratio of VSS to TSS is about 0.49). 6.2.3 Profi les of C O D , V F A and p H C O D , V F A and p H were monitored at influent concentrations of 4.56, 9.93, 17.1, 28.8 and 38.1 g C O D / 1 , (pH was not measured at an influent concentration of 4.56 g COD/1) . The p H , acetic acid, propionic acid and C O D profiles are presented in Figures 6.6, 6.7, 6.8, 6.9 and 6.10, respectively. Each curve represents the effect of different influent concentration at steady state except for a concentration of 38.1 g C O D / 1 . Below- 4 cm, the pH was in the range for 4 to 5, V F A concentrations were high (up to 2895 mg/1 of acetic acid) and C O D reduction was between 17 to 49%, depending on the activity of the sludge and the influent concentration. Above 12 cm, the p H increased to 6.4 and the volatile fatty acid decreased to 100 mg/1 or less. More than 60% of total C O D reduction occurred between 4 cm and 12.5 cm above the reactor bottom except for an influent concentration of 38.1 g C O D / 1. The two completely different sets of p H C O D and V F A s values demonstrate that two separate reaction phases: acidogenesis and methanogenesis were established in the reactor. The acidogenic phase was in the bottom below 4 cm, which was indicated by lower p H , higher V F A and lower C O D reduction (Figure 6.11). Above 4 cm, methanogenesis took place. This was demonstrated by higher p H , lower V F A s and high C O D removal. So far, in anaerobic digestion studies, phase separation has been accomplished in two separate reactors by controlling the pH and dilution rate. It is believed that this is the first time that the two phases were reported in the same reactor. It was thought that this would be true for all substrates in a plug flow reactor. Interestingly enough, when using a U A S B reactor to treat baker's yeast wastewater in the same lab, two distinct Table 6.3: Sludge in the U A S B Reactor OLR Input Sludge Ratio Sludge Lost in Sludge Effluent Sample growth g COD/1 d g COD/1 g VSS g C O D / g VSS g VSS g VSS g VSS 5.00 86.5 0.164 - - -0.91 4.56 133.6* 0.099 22.36 - 3.06 1.97 9.93 107.4 0.262 20.11 13.58 7.37 3.54 17.7 91.7 0.547 4.87 17.87 7.05 5.96 28.7 114.2 0.711 2.32 13.72 38.46 7.77 38.1 164.9 0.654 2.52 13.92 67.14 * Add 67.23 g VSS of sludge to digester at day 15. Chapter 6. THE DISTRIBUTIONS OF SLUDGE AND SUBSTRATES 0.8 Figure 6.5: S ludge G r o w t h Chapter 6. THE DISTRIBUTIONS OF SLUDGE AND SUBSTRATES Figure 6.6: p H Profile Chapter 6. THE DISTRIBUTIONS OF SLUDGE AND SUBSTRATES 105 Chapter 6. THE DISTRIBUTIONS OF SLUDGE AND SUBSTRATES 106 Influent O 4.56 o coo I" • 9.93 o COO f A 17.1 a COD r X 28.8 a COO l " V 38.1 Q COO r Eaocfor H*lgM (oO Figure 6.8: Profi le of P r o p i o n i c A c i d Chapter 6. THE DISTRIBUTIONS OF SLUDGE AND SUBSTRATES Influent O 4.56 o COD i" • 8.93 g COD r' A t7.1 g COD I" X 28.8 g COO I'' V 38.1 g COD I'' —J— 20 —T— •40 — r -60 — r -80 R«actor Height (cm) — I — 100 120 Figure 6.9: • P r o f i l e of C O D Chapter 6. THE DISTRIBUTIONS OF SLUDGE AND SUBSTRATES 108 Figure 6.10: Profi le of C O D R e d u c t i o n Chapter 6. DISTRIBUTIONS OF SLUDGE AND SUBSTRATES 109 phases were not observed. The different results could be due to the inherent chemical characteristics of the substrates. Anaerobic digestion is a biological process in which a series of parallel and consecutive reactions take place. From an oversimplified point of view, it has been accepted that only two major steps, acidogenesis and methanogensis, are considered essential, and generally, the second one is extremely slow and therefore is the rate-controlling step. If the two major steps remain in balance, the intermediate products, i.e. V F A s , would not be detected. Therefore, two phases wouldn't be seen in one reactor. It is only possible to observe the two phases in cases in which the reaction rates of the two steps are very different. In other words, the observation of two phases in one reactor is only possible for some particular substrates, such as cheese whey which is easily converted into short chain acids by acidogenic bacteria. When more fatty acids are formed than can be converted, V F A s accumulation occurs and the p H drops. The accumulation of V F A s in the first step being faster than the assimilative capacity in the second step creates a distinct acidogenic phase. The backer's yeast wastewater contained high concentrations of hard-to-degrade organic material that could not be easily acidified. In contrast, whey has a tendency of rapid acidification. The observation of two phases indicates that the anaerobic system using whey was not maintained in dynamic balance even at very low organic loading rates even though the overall system, from the effluent analyses, appears to be very stable at influent concentrations below 28 g C O D / 1 . With an increase in influent concentration, the V F A s and C O D in the reactor grad-ually increased. The curves for C O D , acetic acid, propionic acid and p H distribution as a function of the height didn't change markedly until an influent concentration of 38.1 g C O D / 1 was applied . At this condition, the acidogenic as well as the methanogenic zones extended upward. Much higher C O D , acetic and propionic acid concentrations were ac-cumulated at the bottom. These high concentrations also extended to a height of 37.5 cm above the reactor bottom. For example, the p H value remained around 3 at a height Chapter 6. DISTRIBUTIONS OF SLUDGE AND SUBSTRATES 110 of 37.5 cm. In particular, propionic acid concentrations remained high throughout the reactor (Figure 6.11). Low pH values (below 3.2) were also observed in this region. Con-sequently, the overall reactor performance was affected. The process became unstable 14 days after the reactor was initially fed at this loading. This was indicated by a decrease in gas production from 67 to 61 liters/ day, and an increase in effluent C O D from 55 to 643 mg C O D / 1 and also an increase in effluent acetic and propionic acid concentrations to 80 and 64 mg/1, respectively. The upward extension of the acidogenic as well as the methanogenic zones causes the intrusion of the acidogenic phase into the region previously occupied by methanogens. The methanogens, previously highly active, in this region could be rendered inactive un-der the acidic environment. Moreover, the methanogens could not be easily replenished in a newly established methanogenic phase to counteract the accumulation of V F A con-centrations due to the very slow growth rate of the anaerobes. It can be predicted that the acidogenic region will extend into the upper portion of the reactor as the substrate loading is increased until the whole region is occupied by the acidogenic reaction and fermentation fails. This is the bottleneck of a suspended growth system. In general, the maximum influent concentration accepted in the U A S B process has been 30 g C O D / 1 to the best knowledge of author based on a literature search, even for those substrates which were not quickly acidified. With an increase in influent concentration, V F A and C O D in the reactor increased. In other words, more of the V F A s produced in the first step accumulated. This could indicate that the rate of acidogenesis increased with the increase in influent C O D . Let the acetic acid concentration of sample port # 1 represent the accumulation of V F A in the first phase. The difference in acetic acid between # 1 and # 2 represents the degradation capacity of V F A in the second phase (Table 6.4). The requirement for maintaining the anaerobic system in a dynamic balance is that the degradation capacity Chapter 6. DISTRIBUTIONS OF SLUDGE AND SUBSTRATES 111 of V F A in the second phase is greater than the accumulation of V F A in the first phase. Based on this idea, the best influent concentration can be found with regard to the system stability. Using linear regression analysis, a set of empirical models for accumulation of acetic acid and propionic acid with increase of influent concentration has been developed which is the best fit to the experimental data (see Appendix A) . The accumulation and degradation of acetic acid are shown in Figure 6.12. The degradation rate first increased until the influent concentration reached 20 g COD/1 , then declined. Between 15-28 g C O D / 1 , the degradation curve is above the accumulation curve , which means that in this region the degradation capacity exceeds the accumulation capacity. Therefore, this would be the optimal influent concentration for a cheese whey anaerobic fermentation system. This conclusion agrees with the experimental results. The experimental results indicated that the majority of the C O D was reduced below a height of 13 cm in the reactor (more than 80%). The C O D concentration decreased from 3.4 g/1 at 4 cm to 0.8 g/1 at 13 cm for an influent C O D of 4.56 g/1, and from 15.3 g/1 to 0.97 g/1 for an influent C O D of 28.1 g/1. The rest of volume of the reactor functions as a settler so that the height of the reactor could effectively be reduced from 140 cm to AA = -0.27 + 0.183Co - 0.0102Co 2 + 0.000197Co3 (6.4) PA = 0.112 - 0.00893Co + 0.00108GV (6.5) 30 cm. Chapter 6. DISTRIBUTIONS OF SLUDGE AND SUBSTRATES 112 From these results the suggested operating conditions and reactor size can be de-scribed as follows : • The optimum influent concentration would be 25 to 30 g C O D / 1 • The reactor height should be reduced to 40 cm or less for the reactor with a diameter of 12 cm. 6.3 S U M M A R Y Profiles of the sludge concentration showed that two sludge regions, a sludge bed with high density VSS and a sludge blanket, exist in the U A S B reactor. The distribution of the sludge was strongly dependent on the process conditions. With an increase of the loading rate, the sludge bed expanded so that the sludge concentration in the blanket and also in the area between the bed and blanket, varied because of gas production. The two reaction stages: acidogenesis and methanogenesis were distinguished in the U A S B reactor by the profiles of the substrates, which indicated that cheese whey is very easily converted to short chain fatty acids and that the rate of the first step is much faster than the second step. The appearance of two stages in the same reactor was associated with either the nature of the substrate or process stress and could be attributed to the fact that the rate of V F A s production exceeded the rate of their utilization. The optimum influent concentration would be between 25-30 g C O D / 1 . The reactor height could be reduced to 40 cm or less. There is an upper influent concentration threshold of cheese whey for stable operation of the U A S B system. If the feed strength is in excess of this threshold value, such as 30 g C O D / 1 at H R T of 5 days, instability occurs because the accumulation of V F A s from the first phase exceeds the degradation capacity of the second phase. Chapter 6. THE DISTRIBUTIONS OF SLUDGE AND SUBSTRATES 113 Figure 6.11: Profiles of p H and VFAs Chapter 6. THE DISTRIBUTIONS OF SLUDGE AND SUBSTRATES 114 Table 6.4: Acetic Acid (AA) Input (g COD/1) A A at 1# (g/1) A A at 2# (g/1) A A i -AAo (g/D 4.56 0.382 0.152 0.23 9.93 0.688 0.096 0.59 17.7 0.876 0.02 0.856 28.8 1.166 0.04 1.126 38.1 2.895 2.553 0.342 1# sample port 1 2# sample port 2 Chapter 6. DISTRIBUTIONS OF SLUDGE AND SUBSTRATES 115 Influent concentration (g/IJ Figure 6.12: Accumulation and Degradation of Acetic Acid C h a p t e r 7 E F F E C T S O F S U L F A T E O N A N A E R O B I C D I G E S T I O N O F W H E Y 7.1 G E N E R A L R E M A R K S High concentrations of sulfate have been thought to be inhibitory to methanogenic bac-teria. The inhibition of the highly fastidious methane producing bacteria(MPB), due to the presence of sulfate, is usually interpreted in relation to the levels of sulfide produced via the SRB (Lawrence et al 1966 and 1965, Cappenberg 1974, Kroiss et al 1983, Speece et al 1983) . Previous studies have paid more attention to the inhibition caused by sulfate on the anaerobic digestion and several mechanisms have been proposed to explain the inhibitory effect. The kinetics of competition for the available electron donors between sulfate-reducing bacteria (SRB) and methane-producing bacteria ( M P B ) have received considerable attention,and it has been concluded that the SRB apparently have a higher affinity (lower Ks) for hydrogen and acetate, which are the major methane precursors relative to the M P B (Abrum et al 1978, Kristjiansson et al 1982, Schonheit et al 1982). The reason might be the periplasmic location of the hydrogenase of sulfate reducing bac-teria (Badziong and Thauer 1980; Bell et al. 1974). Besides, the toxicity of sulfide or free H2S produced by microbial reduction of sulfate is also thought to be a factor of primary importance. On the other hand, from the point of view of thermodynamics and hydrogen regulation (Stephen et al 1986, Jack et al 1989), the presence of sulfate appears to help maintain the anaerobic conditions required for the growth of methanogens. Surprising little research 116 Chapter 7. EFFECTS OF SULFATE ON ANAEROBIC DIGESTION OF WHEY 117 has been done on this aspect, but rather, to provide information which supports this idea theoretically. Motivated by a need to explore the inhibition mechanism caused by high substrate strength and a desire to improve the stability and efficiency of the system as well, an effort was made to examine the effect of sulfate on system performance. The sulfate function in the cheese whey anaerobic fermentation will be the subject in this chapter. 7.2 H Y P O T H E S I S It is well known that high organic loading results in an inhibitory effect in the anaero-bic digestion of cheese whey. As the preliminary results of Chapter 5 indicated, when influent concentration was increased to 38.1 g C O D / 1 system instability occurred. The distribution of substrates in the reactor given in Chapter 6 further provided evidence that the instability was the result of the ease of conversion of cheese whey into short chain V F A s . The two major steps, acidogenesis and methanogenesis in anaerobic fermentation of whey, have been shown to have very different rates of metabolism. When the accu-mulation of V F A in the first step exceeds the capacity of the methanogenesis as influent C O D approaches a threshold, such as 38 g C O D / 1 , catabolism leads to the system failure. A n optimal influent concentration as being found in Chapter 6 can be chosen to avoid the problem. When higher influent concentrations are desirable to achieve the satisfaction of treatment efficiency, however, the question then arises as to how to enhance and control the process stability. Theoretically, the system stability can be enhanced by increasing the activity or concentration of the methanogens, or by inhibiting acidogenic activity, or creating a microbial association which can help to degrade V F A s . p H control was chosen as a reliable method to neutralize acids for maintaining stability Chapter 7. EFFECTS OF SULFATE ON ANAEROBIC DIGESTION OF WHEY 118 of an anaerobic reactor. However, the link between p H and system stability has not been clearly shown. In addition, other possible strategies have not been explored for the improvement of the stability of control. Recent studies have shown that molecular H2 and interspecies hydrogen transport play an important role in anaerobic digestion since anaerobic ,/3-oxidation of long chain fatty acids is considered to be the rate-limiting step for the digestion of soluble substrates (Figure 7.1, Stephen 1986). As previously described (in the literature review), the free energy changes of /3-oxidation are feasible only when the reaction can be " pulled" to the right by the continuous removal of H2 (Table 7.1, Stephen 1986). In terms of control strategies, then it is apparent that operation of the anaerobic reactor at the lowest H2 pressure will minimize the accumulation of fatty acids. Inhibition of the M P B by the SRB has been recognized in relation to the value of Ks, which suggests a higher affinity of the S R B toward H2 (Kristjiansson et al 1982, 1983). The free energy change for the oxidation of reduced pyridine dinucleotides becomes ever more favorable as the H2 partial pressure diminishes. The H2 partial pressures generated by the O H P A (obligate hydrogen-producing acetogens) can be kept at a low enough value by associating them with a i^-util izing organism, either a M P B or a S R B . Therefore, a hypothesis was proposed that the rate-limiting step (for soluble sub-strates) for anaerobic digestion, /3-oxidation, or degradation of fatty acids can be en-hanced through the presence of sulfate. In other words, a proper amount of sulfate may be applied to moderate the detrimental influence of excess H2 on a stressed anaerobic reactor. A study of the effect of sulfate ions (sodium sulfate) in anaerobic fermentation was conducted to determine whether or not they help maintain a favorable environment for methanogens and how they effect the anaerobic digestion system in terms of methane production, C O D reduction, p H , V F A and stability of the system. Chapter 7. EFFECTS OF SULFATE ON ANAEROBIC DIGESTION OF WHEY 119 ORGANIC MATTER carbohydrates proteins v I ACID- FORMING BACTERIA acetic (o) -p(B>-pyruvic t ( C H 3 C O C r + H C O 3 - +H+ +3H 2 +18.2 B . C H 3 C H 2 C H 2 C O O - + 2 H 2 0 — > 2 C H 3 C O O - + H+ + 2 H 2 +11.5 C . 2 C H 3 C H O H C O O " + 4H 2 0 —-> 2 C H 3 C O O - + 2 H C O 3 - + 2 H++4H 2 -1.9 D . C H 3 C H 2 O H + H 2 0 —> C H 3 C O O - + H+ + 2H 2 +2.3 2. H2-utilizing methanogens and desulfovibrios : E.4H 2 + H C O 3 - + H+ —> 2CH 4 + 3H 2 0 -32.4 F.4H 2 + S0 4 = + H+ —> HS; + 4H 2 0 -36.3 3. Acetate-utilizing methanogens G . 2 C H 3 C O O - + 2H 2 0 —> 2 C H 4 +2HCO3- -14.8 4. Syntrophic association of coculture : (A+E) 4CH 3CH 2COO-+3H 20 —>4CH 3COO-+HC0 3-+H ++3CH 4 -24.4 (B+E) 2CH3CH 2CH 2COO-+HC0 3+H 26—>4CH 3COO-+CH 4+H+ -9.4 (C+E) 2CH 3 CHOHCOO--fH 2 0—>2CH 3 COO-+HC0 3 -+H + +CH 4 -34.3 (D+E) 2CH3CH20H+HC03---->2CH3COO-+CH4-l-H2O-|-H+ -27.4 (A+F) 4CH3CH2COO-+3S04=—>4CH3COO-+4HC03-+H+-l-3HS- -36.1 (B+F) 2CH 3CH 2CH 2COO-+S0 4=—>4CH 3COO-+H++HS- -13.3 (C+F) 2CH 3CHOHCOO-+S0 4=—>2CH 3COO-+2HC0 3-+H 2S -38.2 (C+E+G) 2CH3CHOHCOO-+3H20—>3CH4+3HC03-+H+ -49.1 Chapter 7. EFFECTS OF SULFATE ON ANAEROBIC DIGESTION OF WHEY 121 7.3 R E S U L T S I N B A T C H E X P E R I M E N T S A preliminary experiment was first conducted in batch reactors. The details are presented in Appendix B. A n interesting finding was that the effect of sulfate on the gas composition was related to the feed strength. At lower feed strengths, the CH4% decreased when sulfate was added. However, no difference in gas composition was observed at higher feed strengths for both continuous and batch experiments. It would seem that the SRB competed with methane bacteria more effectively at lower substrate concentration. No significant inhibition was observed even when the ratio of C O D to sulfate was as low as 5 and the sodium sulfate concentration was as high as 60 mM/1. This can be attributed to the fact that this substrate had higher solubility and higher feed strength than that used in the previous studies (compared to 0.5 - 1 g COD/1) . Using higher solubility substrates as feed, such as cheese whey, the inhibition threshold concentration might be higher than that of other substrates which have lower solubility since excess hydrogen exists in such systems. Therefore, the critical inhibition value of sulfate varies from one substrate to another. T h e effect of sulfate on p H is very significant. The p H was higher when sulfate was applied. This implies that a proper concentration of sulfate might be able to increases the p H stability of an anaerobic process by competition for hydrogen and fatty acids between the sulfate reducing bacteria and the methane bacteria, helping to maintain stable operation. Batch experiments could not indicate a meaningful relationships between sulfate con-centration and other responses, such as C O D and V F A s concentration. A continuous experimental mode was therefore chosen to determine these effects. It turned out that the effect of sulfate addition was far more noticeable in continuous operation than in Chapter 7. EFFECTS OF SULFATE ON ANAEROBIC DIGESTION OF WHEY 122 batch experiments. 7.4 R E S U L T S I N C O N T I N U O U S E X P E R I M E N T S 7.4.1 Effect of Sulfate on R e a c t o r O p e r a t i o n : T i m e Profi le During the 300 days of operation, the loading rate of the reactor was increased through 7 steps from 1.2 (the sludge loading rate of 0.2 g C O D / g VSS-d).to 10 g C O D / l - d . The first 3 load changes were accomplished by the reduction of the H R T from 15 to 5 days at a constant influent concentration of 15 g C O D / 1 with an addition of 0.2 g/1 sodium sulfate. Following that, the change of load was done by changing the influent concentration at a H R T of about 5 days. For each subsequent increment of influent concentration, an operating period of 6 to 10 HRTs was maintained to ensure stable operation. The results are graphically presented in Figure 7.2 to 7.10. Several stages could be identified from the time profile of the reactor performance. The first 45 days could be considered the start-up period. The effluent concentration decreased from 1.95 to 0.35 g C O D / 1 , while C O D reduction increased from 87% to 98%. The reactor start-up was also indicated by other process parameters. Methane content in the biogas increased from 40% to 51%, then stabilized at about 48%. p H of the effluent increased from an average of 7.5 to a value of about 8.2. The content of V F A in the effluent significantly decreased during this period of time. After 45 days operation, the total V F A in the effluent was maintained below 0.3 g/1. Note that there was no difference between the results with sulfate addition and with no-addition, (in comparison with the results in Chapter 5). From day 46 to 134, in the first 4 increases of organic loading rate (OLR) of 1.2, 1.8, 2.8 and 4 g C O D / 1 d, the system was stable. The effluent C O D , V F A and gas composition Chapter 7. EFFECTS OF SULFATE ON ANAEROBIC DIGESTION OF WHEY 123 50 H Q O O 40 3 C o VL 30 CO V_ *-> c CD O o 2 0 O *•> 3 o 10 Start S=0.2 g/1 Steady S=0.2 g/1 Transient S=0.2 g/1! Unsteady S=0.3g/1 O In • Out S=0.3 g/1 %rroq&3>1 S=0.3 g/l •> / % • • Load _A Shock 1 w \ / \ shock 2 • 1 • "i — — ' 1 ' 1 — 1 < ,— • i 1 40 ao 120 160 Time (days) 200 240 260 Figure 7.9: Biogas C o m p o s i t i o n versus T i m e Chapter 7. EFFECTS OF SULFATE ON ANAEROBIC DIGESTION OF WHEY 131 Star! S=0.2g/1 Steady S=0.2 g/l Transient S=0.2 g/l 280 Time (days) Figure 7.10: VSS in the Effluent versus Time Chapter 7. EFFECTS OF SULFATE ON ANAEROBIC DIGESTION OF WHEY 132 were constant, which indicated that the reactor was in a very active and underloaded conditions without any stress up to an O L R of 4 g C O D / 1 d. The gas production, in terms of liters of methane per liter of feed per day, had not changed as the O R L increased stepwise to 4 g C O D / 1 d. From Figure7.2 we can see that the same influent C O D of 15 g/1 was applied for the first 3 steps. The O L R was changed by changing the H R T during this period of time. Independent evidence (Yan et al 1988), which showed that the effect of H R T on gas production was not significant for low influent concentrations, agrees with this result. Similar to the previous results without the addition of sulfate (Yan et al 1989), when an influent concentration of 30 g C O D / 1 ( O L R of 5.5 g C O D / 1 d) was applied on the 138th day, the system experienced a non-steady state condition. Only the sensitive parameters, such as V F A , pH and gas composition showed a respond to a "transient state". Gas composition immediately declined to 40% methane. Four days after the new load was applied, V F A , especially propionic acid increased to 500 mg/1 and p H dropped to 7.3. The transient state returned to steady state within 1.5 HRTs. The results of this experiment demonstrated once again that an influent concentration of 40 g C O D / 1 is the threshold concentration for the stability of this system, as is shown for the case without sulfate addition. The instability of the system appeared on day 15 after the load change (day 200). It was first detected by a drop in methane content in the biogas from 47% to 30%, pH fell and VSS in the effluent rose. A large amount of sludge, up to 4.3 g VSS/1, left the reactor with the effluent. The total V F A in the effluent accumulated to 2000 mg/1. The ratio of propionic to acetic acid increased from 0.5-1 at steady state to 1.5, and continued up to 2.5-5. The results,..under these conditions with the addition of sulfate, were less favorable than the author expected, based on the concept of interspecies hydrogen transfer and the functions of hydrogenotrophic association of S R B with M P B in the fermentation process. Chapter 7. EFFECTS OF SULFATE ON ANAEROBIC DIGESTION OF WHEY 133 It was assumed that 0.2 g/1 sulfate was not enough to create a favorable environment for methane fermentation of high strength cheese whey up to 40 g C O D / 1 , which represents a ratio of C O D to sulfate of 200. Further tests were then made with higher sulfate concentration in the hope that this would enhance the activity of the SRB. It was first necessary to return the reactor operation to a stable steady state. Various strategies have been tried to bring the reactor back to normal operation. Maintaining the feed concentration at 40 g C O D / 1 , the O L R was decreased from 8.5 to 3 g C O D / 1 d (daily feed of 1.1 liters), then increased step wise, until it again reached 8 g C O D / 1 d. It took 30 days (day 204-233) to re-establish the full activity of the methane bacteria. Both acetic acid and propionic acid decreased well below 200 mg/1. VSS in the effluent decreased, remaining below 1-2 g/1. After day 235, a feed of 50 g C O D / 1 with 0.3 g/1 sulfate concentration was used. The reactor remained stable until day 262, 15 days later the influent concentration was increased to 50 g C O D / 1 . The gas composition suddenly dropped to 40% CH4 and further declined to 37% the next day. To avoid an entire upset of the reactor, more sulfate was added into the feed substrate to rise the sulfate concentration to 0.5 g/1 in the feed. However, the addition of sulfate was not able to immediately reverse the decline of reactor performance. Methane content in the biogas remained about 39% for a while (day 263-267). Both acetic and propionic acids kept increasing from day 265 to 272, which showed that the high concentration of cheese whey had already upset the reactor. 5 days later, the gas composition return to about 43% CH4, and both acetic and propionic acids dropped below 0.2 g/1. A much lower butyric acid concentration than found during the previous stressed period indicated that the bacteria were less severely inhibited this time. A high concentration cheese whey of 69 g C O D / 1 with 0.5 g/1 sulfate was used during the last stage from day 282 on. A low methane content of only 31.8% in biogas on the Chapter 7. EFFECTS OF SULFATE ON ANAEROBIC DIGESTION OF WHEY 134 third day of the last stage indicated that the reactor was subjected to stress. Further addition of sulfate up to 1 g/1 no longer improved the reactor stability. 7.4.2 Effect of Sulfate on Steady State P e r f o r m a n c e Steady state was denned as the condition in which the system parameters remained con-stant within ± 1 0 % over the period of operation. For each operating condition, at least two H R T s were needed to reach a new steady-state. The results for whey digestion at steady state as a function of influent concentration are summarized in Table 7.2. In com-parison with no-sulfate addition, the results at steady state were quite similar in terms of gas production and C O D removal with the exception that with sulfate addition the reac-tor could treat a higher allowable influent concentration and organic loading rate, which shows that no inhibition occurred at these operating conditions. The results show little effect of sulfate on gas composition or methane production rate for continuous operation in the U A S B reactor (Figure 7.11 a and b). The presence of sulfate also appeared to have no influence on the relationship between methane production and organic loading (Figure 7.12). The addition of sulfate resulted in similar volumes of biogas produced in terms of 1 CH4 j\ reactor per day as those in the absence of sulfate. With the increase of influent concentration, the percentage o(CH4 in the produced biogas gradually diminished which indicated that the domination of acidogenesis in the digestion of whey increased with an increase in influent concentration. Gas production in terms of liters C / / 4 per gram C O D per day also declined when influent concentration was increased. O n the other hand, the presence of sulfate did affect the yield of methane in the batch experiments (Appendix B). The studies on the inhibitory effects of sulfate on the activity of methanogens led Isa et al (1986) to conclude that the competitive nature of S R B relative to M R B was dependent on the nature of the feed substrate. The effect of Chapter 7. EFFECTS OF SULFATE ON ANAEROBIC DIGESTION OF WHEY 135 Table 7.2: E x p e r i m e n t a l Resul ts i n Steady State Effluent. Influent OLR C O D pH COD Gas C H 4 % g COD/1 g COD/l .d g/1 Removal% 1/1 d 15.50 2,03 0.33 7.95 97 5.2 47.25 3.13 0.28 7.87 98 5.5 47.41 20.96 4.21 0.32 7.73 98 7.2 45.50 30.50 5.58 0.63 7.56 98 9.1 43.92 41.30 7.52 0.68 7.56 98 10.9 41.07 50.91 11.41 1.16 7.77 97 13.9 40.10 Chapter 7. EFFECTS OF SULFATE ON ANAEROBIC DIGESTION OF WHEY 136 Figure 7.11: Effect of Influent Concentration on Methane Production and Com-position Chapter 7. EFFECTS OF SULFATE ON ANAEROBIC DIGESTION OF WHEY 137 Figure 7.12: Effect of O L R on Methane Production Chapter 7. EFFECTS OF SULFATE ON ANAEROBIC DIGESTION OF WHEY 138 sulfate on biogas composition can be attributed to the different substrate mix used in two experiments, basically the ratio of substrate to sulfate. In the batch experiments, the concentration of whey and sulfate were compatible, since the ratio of C O D to sulfate ranged between 5 to 10. Only small amounts of sulfate were employed in the continuous experiments, a range of ratios from 100 to 200. The substrate concentration used (0.5-5 g COD/1) by Isa was also much lower than those used in the continuous experiments of this study. In addition, the different types of substrate used in the two cases should be also taken into account. Using acetate as a feed material for digestion, Isa found that both specific methane production and CH^ increased as the feed concentration was increased. This was interpreted as meaning that the S R B became more competitive at lower substrate strengths. Since acetate was the only carbon source and cleavage of acetate therefore was the only major reaction involved in the digestor, it is reasonable to believe that the methanogenic cleavage of acetate was inhibited by S R B at a very low H2 partial pressure. When whey, which mainly consists of lactose, was used as substrate for anaerobic digestion, the reactions involved in interspecies H2 transport were more complicated. H2 as an interspecies plays many roles in whey fermentation. In this case, SRB could be competing for H2 with i^ -using methanogens and could inhibit both i^ -using methanogens and acetoclastic bacteria. O n the other hand S R B might also promote and stimulate the degradation of fatty acids by removing excess hydrogen. The final result would be the overall effect of these two opposing mechanism. The amount of sludge was monitored by measuring the VS S profile in the reactor and the VSS content in the effluent. These results are presented in Table 7.3. A plot of ( d X / d t ) / X vs (dS/dt) /dX provides the means for determining Y (yield coefficient ) and Kd (decay coefficient), which were 0.053 g V S S / g C O D and 0.00047 day^. The sludge yield was the same, while the decay rate was much lower than in the process without sulfate. Chapter 7. EFFECTS OF SULFATE ON ANAEROBIC DIGESTION OF WHEY 139 Despite the fact that methane production, effluent C O D and C O D reduction at steady state for the two cases (with sulfate addition and without sulfate addition) were virtu-ally the same, several remarks can be made about the difference in the two operating conditions. 7.4.3 The Improvement of Reactor Stability The impressive improvement in the reactor stability with the addition of sulfate was a significant result. During the first 180 days, the reactor experienced a number of load shocks. The first load shock appeared on day 106 when the reactor was fed 3.4 1/d of cheese whey with a concentration of 20 g C O D / 1 daily. Two days later, 5.4 1 of the same concentration were pumped into the reactor again. Surprisingly, except for gas production and gas composition, which drastically decreased then rapidly recovered, there was no other evidence that the reactor was becoming unstable. This is seen by considering the constant VFAs and C O D concentration in the effluent. The second load shock was applied on day 167. 3.9 1/d of 31 g C O D / 1 cheese whey was used. At this time, both the gas production and effluent V F A decreased immediately, then returned to their normal values shortly afterwards. The reactor was so stable that it was expected that an influent concentration of 40 g C O D / 1 could be applied. Although 0.2 g/1 of sodium sulfate concentration in the feed did improve the bacterial resistance to shock loads, it could not help with further increases in the loading rate. As in the earlier studies, the influent concentration of 40 g C O D / 1 or O L R of about 8 g C O D / l - d caused instability. Even though the VSS profiles show that the total amount of sludge was greater than that without sulfate, it could not help to improve the O L R , nor could the 0.2 g/1 of sulfate added. It was interesting to find that no H2S was detected when only 0.2 g/1 sulfate was added to feed, while 0.3-0.6% H2S was observed for 0.3 g/1 sulfate addition. It could be explained that sulfate was first used as nutrient for M P B 139 Chapter 7. EFFECTS OF SULFATE ON ANAEROBIC DIGESTION OF WHEY 140 Table 7.3: Sludge in the reactor O L R Inf-COD Sludge F/S Eff. lost in Net Growth Sample g COD/l-dg COD/1 gVSS gVSS g VSS • gVSS 128.95 0.250 - - -2.96-3.13 14.4 142.21 0.254 -51.75 - 65.01 4.03 20.96 .206.13 0.254 64.26 14.48 142.66 • 5.50 3L24 253.34 0.308 115.83 33.59 185.53 7.56 41.36 152.52 198.24 24.1 9-11. 50.42 224.52 0.56 130.56 18.57 220.86 Chapter 7. EFFECTS OF SULFATE ON ANAEROBIC DIGESTION OF WHEY 141 and S R B rather than hydrogen utilizing reagent. 7.4.4 T h e Effect of Sulfate on Buffer C a p a c i t y Figures 7.13 and 7.14 show the effect of sulfate on pH by comparing the results with and without sulfate addition. A statistical analysis of the data using Minitab t-test is presented in Appendix D. It was evident that there was a substantial increase in p H with sulfate addition, an average of 0.6 (from 0.48 to 0.82) p H units higher than without sulfate. 7.4.5 T h e I m p r o v e m e n t of Treatment Efficiency The improvement in the treatment efficiency can be seen from the higher allowable in-fluent concentration with further sulfate addition of 0.3 - 0.5 g/1 (Table 7.2). Instead of 30-38 g C O D / 1 without sulfate (Table 5.2), the highest permissible feed concentration reached 50 g C O D / 1 in the experiments with sulfate of 0.5 g/1. Accordingly, the organic loading rate rose to 11.41 g C O D / l - d , which is higher than without sulfate, 7 g C O D / 1 d. 7.4.6 T h e Effect of Sulfate on Profiles of p H , Sludge a n d V F A s Profiles of p H , sludge and V F A s as a function of influent concentration are showed in Figure 7.15 to 7.19 and Table 7.4. In general, it can be said that sulfate addition in the amount applied here did not change the shapes of the profiles. Two different stages were distinguished again for the first 4 operating conditions, as was noticed in earlier experiments. Acetic acid, propionic acid and C O D were well reduced below a height of 12 cm, and the border between the two stages appeared clearly. Also, two sludge regions, a sludge bed and sludge blanket, Chapter 7. EFFECTS OF SULFATE ON ANAEROBIC DIGESTION OF WHEY 142 were observed as shown in Figure 7.15. However, the sludge was darker and thicker in the reactor with sulfate addition than without sulfate. The profiles of the V F A s showed a change when the influent concentration was in-creased to 50 g C O D / 1 , presumably due to the fact that the organic loading rate of the reactor approached its upper limit. Under this circumstance, concentrations of acetic acid and propionic acid drastically increased and the acidogenic stage extended upward. At a height of 50 cm, 1100 mg/1 of acetic acid and 1400 mg/1 of propionic acid were detected. Acetic acid concentration was essentially zero at a height of 80 cm, while propionic acid maintained a high concentration of 1600-2000 mg/1. In spite of the high propionic acid concentration in the reactor, the system was fairly stable. The establishment of biomass activity and the fermenter performance can largely be interpreted by its V F A values. Total V F A concentrations greater than 2000 mg/1, or acetic acid alone being greater than 800 mg/1 and a P / A ratio greater than 1.4 are the predictions of anaerobic process failure, as suggested in the literature (Hill et al,1987). The extremely high propionic acid concentration observed in the reactor with faily stable operation showed that the reactor performance was superior due to the sulfate addition. Figures 20 and 21 show the difference in acetic and propionic acids in the reactor for the absence and presence of sulfate. When the influent concentration was increased to 38.1 g C O D / 1 without the addition of sulfate, acetic acid and propionic acid were as high as 3000 and 1000 mg/1, respectively. However, acetic acid and propionic acid were only 30 and 100 mg/1 at the influent concentration of 40 g C O D / 1 / when sulfate was added. A substantial decrease in butyric acid concentration in the acidogenic stage caused by sulfate addition was observed. Figure 7.22 shows this effect. In the absence of sulfate, a considerable concentration of butyric acid was accumulated in the first stage (acidoge-nesis), but was hardly detectable in the presence of sulfate. Chapter 7. EFFECTS OF SULFATE ON ANAEROBIC DIGESTION OF WHEY 143 Without Sulfate x-- X x X -• — = • 4-- V V Influent 4.45 fl COO I ' 1 « 9.93 g COP »" A 17.1 g COD I " X 28.8 g COD I ' ' V 38.1 g COD I ' 1 With Sulfate 7-6 -5-20 Influent O 15 g COP r* • 20 g COD I'1 A 30 g COP I'* X 40 g COD l'% V SO g COD I'1 0 40 —r— 60 I 80 100 120 Reactor Height (cm) Figure 7.13: Effect of Sulfate o n p H Profiles 7. EFFECTS OF SULFATE ON ANAEROBIC DIGESTION OF WHEY 144 8 2 • Without Sulfate o With Sulfate ^ e - ° or T r 1 1 1 I ' 1 I I 1 i—'. I 1-0 10 20 30 AO Influent ConcentrationCg C O D I"1) Figure 7.14: C o m p a r i s o n of p H w i t h o u t Sulfate a n d w i t h Sulfate Chapter 7. EFFECTS OF SULFATE ON ANAEROBIC DIGESTION OF WHEY 145 Figure 7.15: Prof i le of S l u d g e Chapter 7. EFFECTS OF SULFATE ON ANAEROBIC DIGESTION OF WHEY 146 6-5-Influent O 15 fl COD I'1 • 20 g COD I" A 30 g COD I" X 40 g COD I'1 V 50 g COD I'1 —r— 40 —r— 60 - r -80 20 Reactor Height (cm) 100 120 Figure 7.16: Prof i le of p H Chapter 7. EFFECTS OF SULFATE ON ANAEROBIC DIGESTION OF WHEY 147 Influent O 15 o C O D r • 20 fl C O D I Reactor Hlght (cm) Figure 7.17: P r o f i l e of A c e t i c A c i d Chapter 7. EFFECTS OF SULFATE ON ANAEROBIC DIGESTION OF WHEY 148 Figure 7.18: Prof i le of P r o p i o n i c A c i d Chapter 7. EFFECTS OF SULFATE ON ANAEROBIC DIGESTION OF WHEY 149 Figure 7.19: Prof i le of C O D Chapter 7. EFFECTS OF SULFATE ON ANAEROBIC DIGESTION OF WHEY 150 Table 7.4: Profiles of the U A S B Reac tor w i t h Sulfate A d d i t i o n Influent " Height. VSS C O D pH A A P A B A IBA C O D g COD/1 cm g/1 mg/1 mg/1 mg/1 mg/1 mg/1 Removal • 14.40 4.0 39.42 10478 5.05 1038 297 36 941 0.31 15.11 12.5 31.72 1071 6.64 103 64 20 0 0.93 25.0 22.63 696 6.60 44 34 0 0 0.95 37.5 21.42 456 6.79 45 34 0 0 0.97 50.0 10.84 437 6.82 45 40 0 0 0.97 62.5 4.11 430 6.81 54 12 0 0 0.97 87.5 3.85 363 6.81 43 8 0 0 0.98 112.5 2.98 388 6.78 51 26 0 0 0.97 18.51 4.0 47.25 12835 5.56 1163 399 63 687 0.36 20.96 12.5 37.05 1339 6.87 43 35 4 14 0.94 25.0 36.26 739 6.89 38 22 1 0 0.96 37.5 28.16 308 6.89 44 8 1 0 0.99 50.0 8.52 398 6.91 18 9 2 0 0.98 62.5 9.82 289 6.92 32 0 3 0 0.99 87.5 8.28 399 7.07 28 0 0 0 0.98 112.5 7.77 407 6.93 27 3 . 0 0 0.98 30.08 4.0 38.47 15985 6.05 1413 901 75 364 0.47 31.24 12.5 40.77 1232 6.96 73 21 1 0 0.96 28.73 25.0 34.92 842 6.92 67 17 1 0 0.97 37.5 33.01 654 6.90 62 14 2 0 0.98 50.0 34.36 556 6.90 57 14 1 0 0.98 62.5 9.77 590 6.96 61 8 1 0 0.98 87.5 8.86 479 7.02 63 13 2 0 0.98 112.5 11.70 455 6.93 62 8 2 0 0.98 40.48 4.0 32.82 18422 6.64 1600 1050 0.55 12.5 31.50 1848 6.89 57 40 0.95 25.0 12.94 1466 7.00 20 105 0.96 37.5 12.17 880 7.01 20 96 0.98 50.0 13.24 919 6.99 21 71 0.98 62.5 11.10 910 6.98 21 56 0.98 87.5 9.21 929 6.96 20 30 0.98 112.5 8.02 980 6.89 20 51 0.98 50.91 4.0 43.98 21449 6.85 2127 2377 163 421 0.57 47.05 12.5 39.38 8067 7.08 1731 2268 349 173 0.84 25.0 29.64 4836 7.05 1154 1685 123 145 0.91 37.5 18.40 4869 7.16 1109 1815 119 73 0.90 50.0 14.94 4834 7.28 935 1586 193 50 0.91 62.5 12.74 4821 7.31 941 1268 99 46 0.91 87.5 12.70 4690 7.38 81 1902 131 41 0.91 112.5 12.82 4692 7.40 95 1700 125 41 0.91 Chapter 7. EFFECTS OF SULFATE ON ANAEROBIC DIGESTION OF WHEY 151 Figure 7.20: Effect of Sulfate on the Profile of Acetic Acid: (B) With S (A) Without S, Chapter 7. EFFECTS OF SULFATE ON ANAEROBIC DIGESTION OF WHEY 152 10000 o> \ooo4 -TO u < o c O Q . O i_ CL 100^ Influent O 4.4S n C O D I" • 9.93 q C O D I" A 17.1 g C O D I" X 28.8 g C O D I" V 38.1 g C O D I" 10000 O) 1000 E "o < o c o o i_ 100 In fluent O 15 q C O D 1" • 20 q C O D 1"' A 30 fl C O D . I " X 40 fl C O D 1" V 50 q C O D 1*' Reactor Height (cm) Figure 7.21: Effect of Sulfate on the Profile of Propionic Acid: (A) Without S, (B) W i t h S Chapter 7. EFFECTS OF SULFATE ON ANAEROBIC DIGESTION OF WHEY 153 Figure 7.22: Effect of Sulfate on Butyric Acid Concentration in the Acidogenic Phase Chapter 7. EFFECTS OF SULFATE ON ANAEROBIC DIGESTION OF WHEY 154 7.5 M E C H A N I S M O F I N H I B I T I O N O F H I G H C O N C E N T R A T I O N O F V F A s A N D L O W p H Inhibition and toxicity in anaerobic digestion are subjects which have received consider-able attention due to the important role they play in digester failure and also because they must be allowed for in the correct design and operation of reactors. Much informa-tion can be found about the effects of various chemicals and environmental factors on anaerobic digestion, but the data tend to be dispersed and without any basic unifying theory. With highly soluble carbohydrate substrates, anaerobic digestion systems are always found to be easily upset. Under loading stress or impending failure conditions, low pHs and high V F A s concentrations are commonly observed, particularly for whey anaerobic digestion systems. p H control has been demonstrated to be crucial for maintaining sta-bility. Inhibition modeling of digestors has been mainly restricted to the apparent effects of high concentrations of V F A s on methanogenic bacteria (Andrews 1968). However, the inhibition mechanism is not fully understood. It has been accepted that non-ionized V F A s are inhibitors for methanogens and sug-gested that the methanogenic bacteria are inhibited either by hydrogen ions or. by their substrate, the volatile acid (Andrews 1965, 1968)). The relationship between non-ionized V F A s and p H is that when pH decreases, the non-ionized V F A s increase. Calculated concentrations of non-ionized V F A s (Appendix D) for cases with and without sulfate are presented in Table 7.5, which shows how p H affects the concentration of non-ionized acid. A dramatic decrease in the concentration of non-ionized acids was obtained due to the increase in pH when sulfate was added. A theoretical explanation for why non-ionized acid acts as inhibitors can be given, combining bacterial membrane transport and the Mitchell chemiosmotic hypothesis. The Chapter 7. EFFECTS OF SULFATE ON ANAEROBIC DIGESTION OF WHEY 155 Table 7.5: C a l c u l a t e d C o n c e n t r a t i o n of N o n - i o n i z e d V F A s Non-ionized Acids C H 4 % p H A A ( m g / l ) PA(mg/ l ) Without Sulfate 56 4.0 324 78 51 4.2 510 105 48 4.4 612 202 45 4.5 753 483 39 3.9 2355 1336 With Sulfate 47 5.05 263 100 46 5.56 155 55 44 6.05 69 44 41 6.64 21 14 40 6.85 17 19 Chapter 7. EFFECTS OF SULFATE ON ANAEROBIC DIGESTION OF WHEY 156 accumulation of non-ionized acids leads to the acidification of the internal cytoplasmic pH and destruction of the pH gradient which is necessary for A T P synthesis. Bacteria have a cytoplasmic membrane that acts as a permeability barrier for hy-drophilic and charged molecules. A peptidoglycan layer, that surrounds the cytoplasmic membrane, confers rigidity and shape on the bacteria. In gram-negative bacteria, an additional outer membrane serves as a barrier to large hydrophilic and to hydrophobic molecules (Hancock 1984). Three different kinds of bacterial membrane transport: pas-sive diffusion, facilitated diffusion and active transport, were elucidated by Harold and Brock et al (Harold 1977, Brock 1984). Passive diffusion is a transport mechanism by which neutral molecules tend to equilibrate across the membrane. The driving force for transport is a concentration gradient. Water, oxygen and carbon dioxide are transported by passive diffusion across the cytoplasmic membrane. In the case of facilitated diffusion, the permeating molecule combines with a membrane carrier and is transported inside the cell along its concentration gradient. For active transport, a specific carrier is generally required for each solute. Three categories of active transport: ATP-dependent, group translocation and transport coupled to the pmf (proton motive force), are recognized. The pmf is a chemiosmotic gradient across the bacterial cytoplasmic membrane that can be considered to have two distinct components: an electrical potential (interior negative) and a p H gradient (interior alkaline). Translocation of protons outside the cell membrane thus increases both components of the pmf. Major roles of the pmf are in the produc-tion of A T P by the membrane-bound ATPase enzyme complex, and for the transport of substrates. In ATP-dependent transport, the hydrolysis of A T P drives the internal accumulation of solutes such as negatively charged amino acids. In group translocation, the solute is modified during its transport (e.g. sugars by phosphoenol pyruvate). In transport coupled to the pmf, cations, anions or neutral molecules can be co-transported with protons or other cations such that the molecule is neutral or carries a net positive Chapter 7. EFFECTS OF SULFATE ON ANAEROBIC DIGESTION OF WHEY 157 charge when it crosses the membrane. For neutral molecules, such as sugars or amino acids, the carrier proteins effectively transfer a positively charged molecule where protons are bound to the carrier for its activation. According to the Mitchell chemiosmotic hypothesis (Mitchell 1966), a trans-membrane p H gradient is generated between electron transport and phosphorylation of A D P . Elec-tron transport causes H+ ions to be pumped outward across the bacterial membrane. This proton expulsion leads to an increase in the pH gradient and the membrane poten-tial, and concomitantly an increase in the proton motive force. This p H gradient (interior more alkaline) is the high-energy intermediate required for A T P synthesis. When the p H of the medium decreases, the non-ionized V F A s consequently accumu-lated. The acetic, propionic and butyric acids in their non-ionized form will penetrat the bacterial membrane without any resistance based on the principle of the passive diffusion of neutral compounds across the bacterial membrane. A dissociation of the acids inside the cytoplasm is then provoked by the higher cytoplasmic internal p H . Thereby, protons are released and the cytoplasm is acidified, leading to the p H gradient dissipation. As a consequence, less energy will be available for the synthesis of bacteria and the attainable growth rate will be lowered. The acetic and propionic acids act as uncoupling agents for p H gradient destruction and potential membrane modification. Hence, the undissociated forms of V F A s are the inhibitors. Different bacteria have different systems and capacity to maintain their p H gradient. Aerobic neutrophilic bacteria, like E.Coli , are capable of maintaining an internal p H near 7.6 for an external p H range of 5.5 to 8.5 (Padan et al., 1981). Some anaerobic bacteria, such as Clostridium pasteurian and Closdium thermoaceticum also show a similar system of p H gradient. However, methanogenic bacteria have only a limited capacity to maintain a constant internal p H . It has been explained that the i / + - A T P a s e system, combined with an antiport cation, could be responsible for the maintenance of a high pH gradient at a Chapter 7. EFFECTS OF SULFATE ON ANAEROBIC DIGESTION OF WHEY 158 low external pH (Kobayashi et al 1982). Not all researchers share in the same belief that high V F A concentration are the inhibitors which cause failure. A n argument has existed for years that accumulation of V F A s are the result of unfavorable conditions for anaerobic process rather than the cause of inhibition. A further discussion on this topic will be given in next section. 7.6 M E C H A N I S M O F S T I M U L A T I O N B Y S U L F A T E . Microbial sulfate reduction is a process in which certain bacteria use sulfate as the elec-tron acceptor during metabolism of organic matter. The kinetics of competition for available electron donors (acetate and hydrogen) by M P B and S R B have received atten-tion in the literature. Although acetate oxidizing S R B have been isolated and identified (Desulfobacter postgatei), they do not seem to be found among digester bioflora (Hocks et al 1984, Mulder 1984). As previously mentioned, all S R B are able to perform the lithotrophic reduction pf SO^2: 4tf 2 + H+ + SO~2 .= HS- + 4H20 (7.6) Therefore, this common modality for H2 utilization sets up a situation of competition between the M P B and the SRB. Molecular H2 is generated in two distinct steps of the anaerobic digestion process, i.e. fermentation and /3 oxidation. The fermentation of the hydrolytic products - amino acid and glucose - is performed by a number of acidogenic Clostridial species native to anaerobic reactors. Hydrogen gas is evolved when pyruvate, the end product of glycoly-sis, is decarboxylated and dehydrogenated to acetate via the phosphoroclastic reaction. H2 is also generated in the anaerobic oxidation of volatile and long chain acids. This is performed by a number of native obligate syntrophic bacteria (usually referred to as Chapter 7. EFFECTS OF SULFATE ON ANAEROBIC DIGESTION OF WHEY 159 obligate proton reducing or hydrogen-producing acetogen [OHPA]). In this process ac-etate units are split off endwise from the chain with molecular H2 being the main sink for electrons. The stoichiometry is as follows (Gujer and Zehnder 1983): (~CH2CH2-) + 2H20 = CH3C02H + 2H2 (7.7) This reaction, mediated by pyridine dinucleotides, is believed to be inhibited by in-creasing partial pressure of H2, as its ORP, -0.32 V , is higher than the standard hydrogen electrode. Oxidation of fatty acids is feasible only when the reaction can be "pulled" to the right by the continuous removal of H2. The argument is, therefore, whether or not the fermentation of pyruvate to acetate is i?2-sensitive. It was thought that this reaction seemed not be inhibited by increasing par-tial pressures of H2 (Gottschalk 1986, Cohen 1979), considering its oxidation/reduction potential (ORP) of -0.68 V , which is lower than the standard hydrogen electrode po-tential of -0.42 V . On the other hand, it has been long believed that fermentation of acetic, propionic and butyric acids from sugar via the E M P pathway is regulated by the availability of H2. Figure 2.2 (from Mosey) provides the main route - the Embden Meyerholf pathway for conversion of sugar to organic acid. The acid-forming bacteria use this pathway to obtain energy from the oxidation of glucose to acetate. During the course of this oxidation, hydrogen atoms removed from the glucose are transferred first to the carrier molecule NAZ) + , converting it to N A D H + H+ and then released into so-lution as dissolved hydrogen gas. In order for catabolism to proceed continuously, given a constant pool of N A D , the N A D H produced during substrate-level phosphorylation of glyceraldehyde-3-phosphate and the oxidative decarboxylation of pyruvic acid to acetyl-C o A must be regenerated. This function is accomplished by the reduction of protons to form hydrogen gas, which is subsequently removed by the hydrogenotrophs such as Chapter 7. EFFECTS OF SULFATE ON ANAEROBIC DIGESTION OF WHEY 160 methanogens, SRB and N R B through interspecies hydrogen transfer. Accumulations of hydrogen need an alternate method of electron disposal for N A D H regeneration. This need is fulfilled by the fermentation of pyruvic to propionate, lactate, and ethanol and/or by the fermentation of acetyl-CoA to butyric acid. Accumulation of H2 during the formation of acetic acid from pyruvate pushes the reaction in the directions which will release the stress from a high pressure of H2 (see line D in Figure 2.3) by oxidation of acetic acid to butyric acid. Further accumulation of H2 would shift the fermentation of pyruvate from producing acetic acid to butyric and propionic acids. The evidence that the concentrations of butyric acid as well as propionic acid decreased due to the presence of sulfate led to a new interpretation of hydrogen regulation of the overall conversion process by throttling the acidogenic reaction at several points in the glycolytic pathway. First, S R B , which have the capacity to utilize H2 might be able to maintain the H2 pressure low enough (say well below 10~4 atm), allowing the fermentation of pyruvate to continue going to acetate. Secondly, as can be seen from Figure 7-1, the removal of H2 can move the metabolism to that of an O H P A . The oxidation of butyric acid can be motivated when H2 pressure drops to below 10~3 atm according to thermodynamic calculations, so can be the oxidation of propionic acid at H2 pressure below 10 _ 4 (Figure 2.3). No matter what the fate of butyric and propionic acids might be, the results indicated that fermentation is H2 sensitive and is regulated by H2 pressure. A question remains whether S R B are able to change the pathway or promote (3-oxidation by providing a H2 sink under the conditions of the present study. Taking a closer look at Figure 7.21 (also Tables 6.1 and 7.4) it was easy to notice that the concentrations of propionic acid in the acidogenic stage ( sample port 1) were the same for two experiments (in the presence of sulfate and in the absence of sulfate), while in the methanogenic stage (sample port 2) the propionic concentrations were much lower Chapter 7. EFFECTS OF SULFATE ON ANAEROBIC DIGESTION OF WHEY 161 with the addition of sulfate than without sulfate addition. A possible explanation for this observation is that pyruvic originally was fermented to propionic and butyric acids, perhaps acetic acid as well. Then propionic acid was further broken down to acetic acid. The lower the H2 pressure, the more propionic acid was oxidized, thus the lower the concentration of propionic acid would be. The results from these experiment suggest that S R B are more likely to promote /5-oxidation by providing a H2 sink, due to the level of H2 pressure to which the sulfate could reduce. A reason for the decrease of butyric acid being far more noticeable than that of propionic acid after the addition of sulfate can be explained by the effect of the different H2 partial pressure. The regulatory effects of hydrogen on acetogenic and methanogenic bacteria have been conveniently illustrated using thermodynamic models and equilibrium assumptions by several authors (Zhender et al.1980, Gujer 1983)). Thermodynamic calculations (see Figure 2.3) indicate that a partial pressure of hydrogen of 1 0 - 3 atm would favor the oxidation of butyric acid, while propionic acid oxidation to acetate becomes favorable only at H2 partial pressure below 1 0 - 4 atm. If H2 is maintained at sufficiently low levels (10~ 4atm), the production of propionic acid needs never occur. The high concentration of butyric acid in the experiment without sulfate addition indicated that the hydrogen pressure must have been greater than 1 0 - 3 atm, which inhibited the oxidation of butyric acid. The decline of H2 pressure to below 1 0 - 3 by S R B due to the addition of sulfate then promoted the further oxidation of butyric acid to acetic acid. However, the level of H2 pressure must be higher than 10~4 even with the addition of sulfate since the concentration of propionic acid remained very high in acidogenic stage. More evidence were found to support the stimulation function of sulfate from Figures 7.23 to 7.26. They were made based on the equations developed in Appendix A , which Chapter 7. EFFECTS OF SULFATE ON ANAEROBIC DIGESTION OF WHEY 162 were statistically fitted to the experimental data. Figure 7.23B shows that the degradation of propionic acid in the absence of sulfate was much lower than that with sulfate, although the accumulation of propionic acid was similar for both experiments (Figure 7.23A). These results indicate that oxidation of propionic acid in the absence of sulfate was inhibited. Supporting evidence was also found that the permissible concentration determined by utilization of propionic acid was much lower in the absence of sulfate (Figure 7.24). Figure 7.25 shows the accumulation and degradation of acetic acid, propionic acid and total V F A s in the absence of sulfate. Figure 7.26 is for the case when sulfate addition was applied. It is interesting to note that in the absence of sulfate, the permissible influent concentration determined by propionic acid concentration was 22 g C O D / 1 , which was much lower than the value given by acetic acid and total V F A s . These were 28 and 30 g C O D / 1 respectively. The permissible concentration was defined as the highest influent concentration at which the degradation of V F A s was greater than their accumulation (as described in Chapter 6). W i t h sulfate added, the permissible influent concentration calculated from propionic acid was 39 g C O D / 1 , higher than the one determined by V F A s , 35 g C O D / 1 (Figure 7.26). These results show that in the absence of sulfate, degradation of propionic acid is the rate-limiting step. When sulfate was applied to the feed stream, oxidation of propionic acid was no longer the controlling step for the overall process due to the stimulation function of the SRB's consumption of hydrogen. No single parameters could be identified as being responsible for the control of the overall process in the presence of sulfate. It is postulated that the' rate-limiting step of the complicated reactions involved in anaerobic fermentation can be changed and that hydrogen pressure plays a central role in causing the shift between those different controlling steps. Acetoclastic activity could be the limiting steps at hydrogen pressures smaller than 10~4 atm. Once the hydrogen Chapter 7. EFFECTS OF SULFATE ON ANAEROBIC DIGESTION OF WHEY 163 1.6 1 -C O <—' 1.2-Aci 0 . 8 -*c o Prop! 0 . 4 -B: D«g of PA With and Without Sulfate Without Sulfate ^ With Sulfate 0 10 20 30 40 50 Influent Concentration (cj C O D I ') Figure 7.23: Effect of Sulfate on the Accumulation(A) and the Degradation (B) of Propionic Acid Chapter 7. EFFECTS OF SULFATE ON ANAEROBIC DIGESTION OF WHEY 164 Figure 7.24: Effect of Sulfate on the Permissible Concentration Determined by Accumulation and the Degradation of Propionic Acid. (A) in the Absence of Sulfate; (B) in the Presence of Sulfate Chapter 7. EFFECTS OF SULFATE ON ANAEROBIC DIGESTION OF WHEY 165 0 5 10 15 20 25 30 35 AO 45 50 Influent Concentration (g COD I ') Figure 7.25: Accumulation and Degradation of Acetic Acid, Propionic Acid and total VFAs in the Absence of Sulfate - - - •-Chapter 7. EFFECTS OF SULFATE ON ANAEROBIC DIGESTION OF WHEY 166 2-1.5 CD < S o-5 < A: Acc and Deg of AA Acc of AA Deg of AA i—'—i—•—i—•—i—•—r 1.5-"o < 1 o "c o o l_ Q . 0 B: Acc and Deg of PA Acc of PA Deg_of_j^ A - T 1 1 1 1 1 1 1 1 J 1 1 1 J 1 J I f -5-4-3H CO < > 2-.o H C: Acc and Deg of Total VFAs Acc of VFAs i—1—r i • I • i • r 0 5 10 15 20 25 30 35 40 45 50 Influent Concentration (g COD I ') Figure 7.26: Accumulation and Degradation of Acetic A c i d , Propionic A c i d and total V F A s in the Presence of Sulfate Chapter 7. EFFECTS OF SULFATE ON ANAEROBIC DIGESTION OF WHEY 167 pressure is greater than 10 - 4 , the overall reaction rate might be determined by the rate of oxidation of propionic acid as well as butyric acid. The similarity in the shape of the acetic acid accumulation curves for the two exper-iments (with sulfate and without sulfate) implies that S R B and M P B do not appear to compete for acetic acid. As a result of the above discussion, a new inhibition scheme, two-stage inhibition in anaerobic fermentation can be suggested, which could answer the question of whether V F A s are the inhibitor or the results of inhibition (Figure 7.27). Inhibition of an anaerobic system occurs in the following two stages. First, high hydrogen pressure drives the pyruvate fermentation to produce propionic and butyric acids rather than acetate, which is called preliminary inhibition. In this preliminary inhibition, higher hydrogen pressure is the cause and the accumulation of V F A s is the result. The accumulation of V F A s in the system subsequently predominates and the consequence of this accumulation causes the direct inhibition of the activity of methane bacteria in the second inhibition stage. High V F A concentrations are the result of unfavorable conditions for the anaerobic process, and also the cause of failure due to the second inhibition. Once a high concentration of V F A s or a low p H is detected, the system has already suffered the second inhibition. Obviously, there is a more effective way to control the process prior to the observation of an apparent accumulation of V F A s , which is to maintain the hydrogen pressure low enough by using hydrogenotrophic association. The advantage of the new control strategy over the old pH control system is that it prevents the system from the first inhibition at the very beginning. Moreover, hydrogenotrophic association is able to promote j3-oxidation of butyric and propionic acids, in turn, to increase the conversion of acetate. Theoretically, it would increase the production of methane. Chapter 7. EFFECTS OF SULFATE ON ANAEROBIC DIGESTION OF WHEY 168 7.7 S U M M A R Y The significant improvement in process stability and treatment efficiency made by adding sulfate has clearly illustrated that sulfate acts like a stimulator which helps to maintain a favorable condition for methanogenesis. The mechanism of the stimulation can be explained according to thermodynamics and hydrogen regulation. It is that sulfate is able to promote the /3-oxidation of V F A s by consuming hydrogen. The results showing the profiles of p H and substrate concentrations elucidated the fact that the rate-limiting step of the complicated reactions involved in anaerobic fermentation is changed and interspecies hydrogen transfer plays a central role. The significant decrease of butyric acid in the first stage indicated that sulfate serves as a hydrogen sink. The conversion of pyruvate to acetic acid offers a solution for the removal of excess hydrogen, improving the overall stability. A two-stage inhibition mechanism in anaerobic fermentation was proposed. Higher hydrogen pressure is the cause of preliminary inhibition, resulting in the accumulation of V F A s which subsequently inhibited the activity and growth of methanogens in the second inhibition stage. The mechanism of inhibition of methanogens from V F A s was interpreted as being caused by the acidification of the internal cytoplasmic p H and the destruction of p H gradient by non-ionized acids based on the theory of bacteria membrane transport. A new control strategy for anaerobic system stability was recommended. That is to maintain the hydrogen presure at low level through hydrogenotrophic association. Chapter 7. EFFECTS OF SULFATE ON ANAEROBIC DIGESTION OF WHEY 169 H 2 He ACETIC <4r BUTYRIC H 2 PYRUVIC H* PROPIONIC FIRST ffigh H 2 pressure INHIBITOR VFAs RESULTS SECOND VFAs INHIBITOR Figure 7.27: Two-Stage Inhibition Mechanism Chapter 8 T R E A T M E N T E F F I C I E N C Y IN O P T I M A L O P E R A T I O N A n experiment was conducted in a 3.04 1 U A S B reactor under optimal operating condi-tions. The reactor configuration was the same as before except that the reactor height was reduced from 168 cm to 30 cm. The optimal operating conditions were selected from the results found in the first three sets of experiments and gave highest treatment efficiency and reliable system stability. These conditions included optimal influent concentration of about 30 g C O D / 1 , reactor height of 30 to 40 cm (for the same diameter) and amount of sulfate addition of 0.2 g/1. The reactor was seeded with 1.5 1 of sludge with a VSS concentration of 35.48 g/1. The total amount of VSS in the reactor was 53.12 g/1. The start-up procedure strictly followed the recommendations from the start-up ex-periment, being certain that the sludge loading was less than 0.25 g C O D / g V S S . In the first 2 weeks the reactor was fed 0.4 1 of whey daily, which corresponds to a specific sludge loading of 0.22 g C O D / g VSS. The daily feed was increased to 0.8 1 after the 3rd week. The gradual increase in feed rate was continued until it finally reached 1.5 1/d after a month. The influent concentrations were in the range from 26 to 32 g C O D / 1 and sulfate was added at a concentration of 0.2 g/1. The experimental results are presented in Table 8-1. A very high treatment efficiency was obtained in this study. Over 97% C O D reduction was achieved with an influent concentration of 32.6 g C O D / 1 , H R T of 2 days and an organic loading rate of 16.61 g C O D / 1 d. 170 Chapter 8. TREATMENT EFFICIENCY IN OPTIMAL OPERATION 171 By comparing the results from this work to other treatment systems given in Table 8-2, it can be seen that the highest C O D reduction for high values of organic loading were obtained using the U A S B reactor. Chapter 8. TREATMENT EFFICIENCY IN OPTIMAL OPERATION 172 Table 8.1: Results i n the O p t i m a l O p e r a t i o n Input g C O D / 1 Feed 1 O L R g C O D / 1 d p H Output g C O D / l Gas 1/d C H 4 % Reduction C O D % 29.46 0.84 8.05 8.09 0.70 14.07 42.51 97 1.16 11.39 7.79 0.69 18.12 46.09 97 1.39 13.65 7.96 0.63 20.59 43.67 98 32.59 1.47 15.96 7.74 0.80 20.36 42.38 97 1.53 16.61 7.69 0.80 21.42 42.44 97 Influent Concentration: 29-32 g C O D / 1 Volume Of Reactor: 3 littres Sulfate Concentration: 0.2 g/1 Table 8.2: Comparison of Anaerobic Treatment Process for Cheese Whey Reactor Waste HRT Temp (°C) P H control Raw Waste g COD/1 Loading g COD/1 d Removal % Reference CMRj raw 14-70 38 7.5 69 - 18-58 Wildcnauer (1985) Fixed film sour whey 5 days 35 6.7 79 14.0 95 Boening (1982) AAFEB 2 whey powder 8.9-27hrs 14-16hrs 28-31 35 No 10 5-15 8.9-27 8.2-22 77-93 61-92 Switzenbaum (1982) AnRBC 3 raw whey 5 days 6-11 days • 35 35 .No 64 61-70 10.2 6.3-10 76 82-93 Lo & Liao (1986) UASB 4 ; raw 5 days -: 33 No 5-28.7 0.91-6 97-99 this whey 5 days 2 days 33 No 41 32.6 7.9-8.2 16.61 81-86 96-98 study 1 OMR completely mixed reactor. 2 AAFEB Anaerobic attached film expanded reactor. 3. AnRBC Anaerobic rotating biological reactor. 4. UASB Upflow anaerobic sludge blanket reactor. C h a p t e r 9 C O N C L U S I O N S , C O N T R I B U T I O N S A N D R E C O M M E N D A T I O N S 9.1 C O N T R I B U T I O N S A N D C O N C L U S I O N S Cheese manufacture, one of the biggest food industries in North Americans facing the challenge of a difficult waste disposal problem. Despite the fact that a number of studies on the anaerobic treatment of cheese whey during last decade have proved that anaerobic fermentation of cheese whey could be an alternative solution for waste disposal, the system was suspect due to many unsuccessful experiments and the difficulties frequently encountered in maintaining a stable operation. This study attempted to overcome these difficulties by solving some of the major problems which existed in previous studies by increasing the basic knowledge of this process; to in turn, facilitate the operation of anaerobic reactors and to lead to its better application by industries. Concentration profiles of various parameters in a U A S B reactor are necessary for the development of a dynamic model for the process which can be used for its optimization. From the measurement of these profiles, several interesting conclusions have been drawn. It is believed that this study is the first to show in anaerobic fermentation studies that the two stages, acidogenesis and methanogenesis, occur separately in the same reactor. The significance of this observation is the realization that two major steps in the whey anaerobic fermentation process are not necessarily in dynamic balance since the two reaction rates are very different. This explains why this system is easily upset. This observation also gave a better understanding of how to efficiently control the system's 174 Chapter 9. CONCLUSIONS, CONTRIBUTIONS AND RECOMMENDATIONS 175 stability by preventing the accumulation of the products in the first stage, e.g. V F A s . This in turn has led to the experimental study of sulfate effects. Sulfate ions have long been believed to be toxic to anaerobic bacteria. The previ-ous studies on the effect of sulfate have been limited to its inhibition effects. These experimental tests first illustrated the stimulation function of sulfate ions, which greatly supported the fundamental concept of a hydrogen regulation theory. As a result of the stimulation, a significant improvement in process stability and treatment efficiency was achieved. The knowledge gained from the study on sulfate stimulation provided a better understanding of the inhibition mechanism of anaerobic digestion and had significant implications for the handling of other high strength soluble carbohydrate wastes. Based on this experimental work, the major conclusions are summarized as follows. 1. The results from the preliminary feasibility assessment experiments have shown that the anaerobic digestion of cheese whey using a U A S B reactor can be an efficient treatment method for diluted cheese whey. Without p H control and nutrient addition, the system could successfully treat cheese whey up to a concentration of about 29 g C O D / 1 . Over 97% C O D removal was achieved. 2. The start-up procedure is very important for successful reactor performance. Var-ious start-up strategies were tested in the present studies to facilitate start-up of the U A S B reactor and to ensure stable operation. Among the operating parameters, sludge loading rate was the most critical for proper start-up of the U A S B reactor. The initial sludge loading during the start-up period should not exceed 0.25 g C O D / g VSS-d . 3. V F A s were found to be a very useful indicator for primary adaptation of sludge and for monitoring the system stability. The loading rate could be increased only after the V F A concentrations in the reactor were at a low value. A rapid increase of O L R may result in the loss of activity of the sludge. 4. The methane production rate was a function of O L R . At an O L R less than 4 g Chapter 9. CONCLUSIONS, CONTRIBUTIONS AND RECOMMENDATIONS 176 C O D / l - d , the reactor fed with a higher influent concentration yielded a higher methane production rate. When the O L R was greater than 6 g C O D / l - d , a higher influent con-centration or shorter H R T led to a lower methane production rate. The optimal O L R for this particular system was between 4-6 g C O D / l - d . If we choose a H R T of 5 days, the optimal influent concentration should be between 20-30 g C O D / 1 . 5. The profile of the sludge showed that two sludge regions, sludge bed with high density VSS and sludge blanket, existed in the U A S B reactor. The distribution of the sludge was dependent on the process parameters. With an increase in the loading rate the sludge bed expanded, mainly due to the gas production, so that the sludge concentration in the blanket, especially in the area between the bed and blanket, varied with gas flow rate. 6. For highly soluble and easily acidified substrates, such as cheese whey, difficulties existed in maintaining stable operation. The results from this study provided a better understanding of the cause of the instability. The observation that two reaction stages, acidogenesis and methanogenesis, were distinguished in the U A S B reactor clearly indi-cated that cheese whey is very easily converted to short chain fatty acids and that the rate of the first step is much higher than that of the second step. The appearance of two stages in the same reactor was caused by the excessive accumulation of V F A s from the acidogenesis stage which was the result of either the nature of the substrates or process stress. 7. There is a threshold of influent concentration of cheese whey at a given H R T for stable operation of a U A S B system. If the feed strength exceeds this threshold, instability occurs. For this system the influent concentration should be maintained below 30 g C O D / 1 at a H R T of 5 days. The optimum influent concentration would be 25 to 30 g C O D / 1 based on the finding that system stability can be maintained if the degradation capacity for V F A s is greater than their accumulation. Chapter 9. CONCLUSIONS, CONTRIBUTIONS AND RECOMMENDATIONS 177 8. A l l measured substrate profiles at different operating conditions showed that more than 80% of the C O D reduction took place below a height of 12 cm in the reactor and the sludge concentration was as low as 5 to 10 g VSS/1 above a height of 30 cm. It is therefore recommended that the reactor height should be reduced to 40 cm or less for this reactor with a diameter of 12 cm. 9. The significant improvement of process stability and treatment efficiency made by the addition of sulfate clearly illustrated that sulfate acted like a stimulator which helps to maintain favorable conditions for methanogenesis. The mechanism of this stimulation is explained according to thermodynamics and hydrogen regulation which suggested that sulfate is able to promote the /3-oxidation of V F A s by consuming hydrogen. 10. A two-stage inhibition mechanism in anaerobic fermentation was proposed. Higher hydrogen pressure is the cause of preliminary inhibition, resulting in the ac-cumulation of V F A s which subsequently inhibit the activity and growth of methanogens in the second inhibition stage. The mechanism of inhibition of methanogens from V F A s was interpreted as being caused by the acidification of the internal cytoplasmic p H and destruction of the p H gradient by non-ionized acids based on the theory of bacterial membrane transport. A new control strategy for anaerobic system stability by using hydrogenotrophic association was recommended. 11. Under the optimal operating conditions, over 97% C O D reduction has been achieved at an influent concentration of 32.6 g C O D / 1 , an H R T of 2 days and an organic loading of 16.61 g C O D / 1 d. This is much higher than what has previously been reported in the literature. Chapter 9. CONCLUSIONS, CONTRIBUTIONS AND RECOMMENDATIONS 178 9.2 R E C O M M E N D A T I O N S F O R F U T U R E R E S E A R C H The following areas are recommended for future investigation based on the findings made in this research: 1. It would be instructive to conduct research on the quantitative relationship be-tween hydrogen pressure and the concentration or accumulation rate of each volatile fatty acids using different substrates as well as different amounts of sulfate or other hydrogen utilizating reagents. It is known that the conversion of pyruvate to butyric, propionic and acetic acids is regulated by hydrogen pressure. The significant reduction of butyric acid shown by profiles of the V F A s indicated that sulfate serves as a hydrogen sink. The conversion of pyruvate to acetic acid offers a solution for the removal of excess hydrogen, improving the overall stability and the process efficiency. A quantitative relationship is definitely needed for the development of a mathematical model to describe and predict the possible products and their concentrations in the acidogenic step and their effect on system stability. These models would be very useful for further development of a more efficient control strategy for the anaerobic fermentation processes. They would be based on hydrogen partial pressure control as it is especially important according to the two-stage inhibition mechanism. 2. It is recommended that an extensive study on inhibition kinetics in anaerobic processes could be made following the two-stage inhibition mechanism proposed in this dissertation. The discovery of new inhibition kinetics could be made on the concept that unexpected high hydrogen pressure is the inhibitor for preliminary inhibition and V F A s are both the result of preliminary inhibition and the inhibitor for the secondary inhibition. 3. This research provided the information on the concentration profiles in the U A S B reactor, which has extended our understanding of the anaerobic process and the causes Chapter 9. CONCLUSIONS, CONTRIBUTIONS AND RECOMMENDATIONS 179 of instability. However, information about the substrate concentrations between sam-ple port 1 and 2 is lacking. Interesting information, related to inhibition and instability, might be presented if there were more sample ports located between port 1 and 2. A n ex-tensive spectrum of substrate concentrations along the reactor column could be attained by constructing the column with a smaller diameter and a larger number of sampling ports. Improved technology such as on-line analytic equipment should also be considered due to the large number of samples and the necessity for simultaneous monitoring at all levels. 4. The positive effects of sulfate on the anaerobic process as a stimulator have been experimentally demonstrated in the present research. However, the cause of the higher concentration of iso-butyric acid in the presence of sulfate remains unknown. Further-more, an explanation of the different behaviors shown by butyric and propionic acids could be made according to thermodynamics and hydrogen regulation. This needs to be verified before a complete conclusion can be drawn. Finally, determination of the optimum sulfate addition and the critical value with respect to system stability are rec-ommended. 5. It was found that no H2S was detected when only 0.2 g/1 sulfate was added to feed, while 0.3-0.6% H2S was observed for 0.3 g/1 sulfate addition. This could be explained that sulfate was first used as nutrient for M P B and SRB rather than hydrogen utilizing reagent. Only when extra sulfate was provided, such as in the case of 0.3 g/1 sulfate added, might a portion of sulfate act as hydrogen utilizing reagent. A further study is needed to determine the quantitative relationship between the amount of hydrogen reduced and the amount sulfate applied. Bibliography A . P . H . A . American Public Health Association (1975) Standard methods for the exami-nation of water and wastewater. 14 edition. Washington D . C . Abram,J.W. and Nedwell, D . B . (1978a). Inhibition of methanogenesis by sulfate-reducig bacteria competing for transferred hydrogen. Arch. Microbiol. 117: 89-92. Abram, J .W. and Nedwell, D . B . (1978b). Hydrogen as a substrate for mathyanogensis and sulfate reduction in anaerobic salmarsh sediments. Arch. Microbiol. 117: 93-97. Aibas, Shoda, M . and Nagentam, M . (1968) Kinetics of production in alcohol fermenta-tion. Biotechnol. Bioeng. 10 845-864. Andrews, J .F . and Graef, S.P. (1971) Dynamic modeling and simulation of the anaerobic digestion process. Adv.Chem.Ser. 105: 126-162. Andrews, J .F . (1968) A mathematical model for the continuous culture of microorganisms utilizing inhibitory substrates. Biotechnol. Bioeng. 10 (6): 707-723. Andrews, J .F . and Pearson, E . A . (1965) Kinetics and characteristics of volatile acid production in anaerobic fermentation processes. Int. J . Air Wat. Poll. 9: 439-461. Anonymous, May 1, 1968, U S D A studies reverse Osmosis as Whey Disposal Method, New Release No. U S D A 1396-68, Washington, D . C . Anonymous, Volume III, Industrial Waste Profiles, No.9 Dairies, September, 1967, The Cost of Clean Water, F W P C A , Washington, D . C . Babbit, H . E . and Baumann (1958) Sewerage and sewage treatment.John Wiley and Sons, New York. Badziong, W . and Thauer, R . K . (1980). Vectorial election transport in desulfovibrio 180 Bibliography 181 vulgaris (marburg) growing on hydrogen plus sulfate as sole energy source. Arch. Microbiol. 125: 167-174. Balch, W . E . and Wolfe, R.S. (1979). Specificity and biological distribution of coenzyme M (2-mercaptoethanesulfonic acid. J . Bacteriol. 137: 264-273. Baresi, L . et al. (1978). Methanogenesis from acetate: Enrichmant studies. Appl . Environ. Mictobiol. 36:186-Bernstein, S. and Tzeng, C . H . (1977). Commercial production of alcohol by the fer-mentation of acid and/or sweet whey, E P A 600/2-77/133. Industrial Environment Protection Laboratory, Office of Research and Development, Environment Protec-tion Agency, Cincinnati, Ohio. Blanc, F . F . and Molof, A . H . (1973). Electrode potential monitoring and electrolytic control in anaerobic digestion. J . Water Pollu. Contr. Fed. 45:655. Boening, P . H . and Larsen, V . F . (1982). Anaerobic fluidized bed whey treatment. Biotech-nol. Bioeng. 24: 2539-2556. Boening, P . H . and Larsen, V . F . (1982). Anaerobic fluidized bed whey treatment, Biotech-nol. Bioeng., 24,2539-56 Bolle, W . L . , J.van Breugel, G.C.van Eybergen, N . W . F . Kossen and W.van Gils. (1986). Kinetics of anaerobic purification of industrial wastewater. Biotech. Bioeng. 28, 542-548. Bolle, W . L . , J . van Breugel, G . C . van Eybergen, N . W . F . Kossen and R.J . Zoetemeyer. (1986). Modeling the liquid flow in up- flow anaerobic sludge blanket reactor. Biotechnology and Bioengineering 28,1615-1620. Bolle, W . L . , J . van Breugel, G.C.van Eybergen, N . W . F . Kossen and W.van Gils. (1986). A n integral dynamic model for the U A S B reactor. Biotechnology and Bioengineer-ing 28, 1621-1636 Brock, T . D . , D . W . Smith and M . T . Madigan. (1984). Biology of microorganisms, 4th Bibliography 182 ed. Prentice-Hall,Englewood Cliffs, N.J . Bryant, M.P . , Campbell, L . L . , Reddy, C . A . and Crabill, M . R . (1977). Growth of desul-fovibrio in lactate or ethanol media low in sulfate in association with /7 2-utilizing methanogenic bacteria. Appl . Envron. Microbiol. 33: 1162-1169. Buijs, C , P . M . Heertjes, and R.R. van der Meer. (1982). Distribution and behavior of sludge in upflow reactor for anaerobic treatment of wastewater. Biotech. Bioeng. 24, 1975 -1989. Buijs, W. , P . M . Heertjes and R.R. van der Meer (1980) Modelling and scale-up of anaer-obic conversion of dissolved fatty acids. Advances in Biotech. 1, 529-535. Buswell, A . M . and Mueller, H.F.,(1952). Mechanism of methane fermentation. Industrial and Engineering Chemistry, 44:550- 552 Callander, I.J. and Barford, J.P. (1983). Recent advances in anaerobic digestion technol-ogy. Process Biochemistry 18(4): 24 -37. Callander, I.J. and Barford, J.P. (1983). Cheese whey anaerobic digestion - effect of chemical floculant addition. Biotechnology letters, 5:153-158 Canviro Consultants Ltd.(1985). Final report on the demonstration of anaerobic diges-tion and energy recovery at Millbank Cheese and Butter Ltd. , Ontario Ministry of Industry and Trade, Toronto. Cappenberg, T . E . (1974). Intrerelationship between sulfate -reducing and methane-producing bacteria inbottom deposits of a fresh-water lake. I. Field observations; II. Inhibition experiments; III. experiments with C 1 4 labelled substrates Antonie van leeuwenhoek J . Microbiol Serol 40:285, 297, 457. Cappenberg, T . E . (1975). A Study of mixed continuous cultures of sulfate-reducing and methane-reducing bacteria. Microbial Ecol. 2:61-Capri, M . G . and Marais, C R . (1975) p H adjustment in anaerobic digestion. Water Res. 9: 307-313. Bibliography 183 Castillo, F.J . , Izaguirre, M . E . , Michelena V and Moreno B (1982). Optimization of fermentation condition for ethanol production from whey. Biotech. Lett. 4(9): 567-572. Chartrain, M . and Zeikus, J . G . (1986a). Microbial ecophysiology of whey biomethana-tion: Intermediary metabolism of lactose degradation in continuous culture. Ap-plied and Environmental Microbiology, 51(1):180-187 Chartrain, M . and Zeikus, J . G . (1986b). Microbial ecophysiology of whey biometha-nation: Characterization of bacterial trophic population and prevalant species in continuous culture. Applied and Environmental Microbiology, 51(1):188-196 Chian, E . S . K . and DeWalle, F .B. (1976). Treatment of high strength acidic wastewater with a completely mixed anaerobic filter. Water Res. 11:295-302. Clanton, C . J . et al. (1980). Anaerobic digestion of cheese whey-a pilot project. Paper No. 80-152, presented at the 1980 North central Region American Society of Engineers. October 23-24 Fargo, N D . Cohen, A . , Zoetemeyer, R.J. , Van Deusen A . and Van Andel J . G . (1979). Anaerobic digestion of glucose with separated acid production and methane formation. Water Res. 13:571-580. Cohen, A . , Breure, A . M . , Van Anoel, J . G . and Van Deursen, A . (1980). Influence of phase separation on the anaerobic digestion of glucose - 1. Maximum COD-turnover rate during continuous operation. Water Res. 14:1439-1448. Cork, D . J . (1985). Microbial conversion of sulfate to sulfur:An alternative to gypsumdis-posal. In mizrahi a,ven Wezel A.L.(ed.): Advances in Biotechnological Processes. Vol.4, pp 183-209. New York: Alan R.Lis,Inc. Cork, D . J . and Mah, S.C. (1982). Acid gas bioconversion favors sulfur production. In Scott C D (ed.): "Fourth Symposium on Biotechnologyin Energy Production and Bibliography 184 Conservation. Biotechnol. Bioeng. Symp., Vol.12, pp 285-300. Gatlinberg. Ten-nessee. Counotte, G . H . M . (1981). Regulation of lactate metabolism in the rumen. Ph.D. Thesis, Utrecht State University, Utrecht. Currie, Coopers and Lybrand Ltd. , November (1981). The Market Potential for Modified Whey and Lactose in Canada. De Zeeuw, W. (1985). Acclimatization of anaerobic sludge for U A S B reactor start-up. Netherlands Jounal of Agricultural Science 33: 81-84. De Zeeuw, W . and Lettinga, G . (1980) Acclimatization of digested sewage sludge during start-up of an up-flow anaerobic sludge blanket (UASB) reactor. Proceeding of the 35th Industrial Waste Conference. Purdue University 39-47. DeHaast, J. , Britz, T . J . , Novello, J .C.and Lategan, P . M . (1983). Anaerobic digestion of cheese whey using a stationary fixed-bed reactor. New Zealand Journal of Dairy science and Technology, 18:261-271. DeHaast, J . , Britz, T . J . , Novello, J .C .and Verwey, E . W . , (1985) Anaerobic digestion of deproteinated cheese whey. Journal of Dairy Research, 52:457-467. Delany, R . A . M . (1981). Recent developments of the utilization of whey. Cultured Daily Products Journal, 16(2):ll-22. Dirasian, H . A . (1963). Electrode potentials developed during sludge digestion. J.Water Pollut.Contr. Fed., 35:424-428. Dubourguier, H . C . , Preesier, G . , Samain, E . and Albagnac, G . (1983). Granular methanogenic sludge; microbial and structural analysis. Poster Presented at Second Conference "Energy from, Biomass" Venice, March 1983. Eastman, J .A.and Ferguson, J .F . (1981). Solubilization of particulate organic carbon during the acid phase of anaerobic digestion. J . Water Pollut. Contr. Fed. 53:523-Edwards, V . H . (1970). The influnce of high substrate concentration on microbiol kinetics. Bi bliography 185 Biotechnol. Bioeng. 12: 679-712. Everson, T . C . (1979). Whey derived gasohol - a reality . Proceeding of Whey Products Conference (5th), Minneapolis, Minnesota, October 9-10. Fan, L . T . , Erickson, L . E . , Baltes, J . C . and Shah, P.S. (1973). Analysis and Optimization of two digestors. J . Water Pollut. Control Fed. 45: 591-610. Ferguson, T . J . and Mah, R . A . (1983). Effect of H2- C02 on methanogenesis from acetate or methanol in methanosarcina Spp. Appl . Environ. Microbiol. 46(2): 348-355. Fersht, A . (1977). Enzyme structure and mechanism. Freeman and Co, Reading, San Francisco. Follmann, V . H . and Markl, H. (1979). pH-static for the continuous anaerobic fermenta-tion of highly polluted effluents to methane. Milchwissenschaft 34(6):341-350 Friend, B . A . , Cunnigham, M . L . and Shahani, K . M . (1982). Industrial alcohol production via whey and grain fermentation. Agric. Wastes 4: 55-62. Friend, B . A . and Shahani, K . M . (1979). Whey fermentation. New Zealand Journal of Dairy Science and Technology, 14(2):143-156. Frostell, B. (1981). Anaerobic treatment in a sludge bed systemcomparedwith a filter system, Journal W P C F , 53(2) 216-222. Ghaly, A . E . and Singh, R . K . (1985). Land application of cheese whey, Proceeding of the Fifth International Symposium on Agricultural Wastes, St.Joseph, Michigan Gholsh, S. et al. (1983b). Two-stage upflow anaerobic digestion of concentrated sludge, in Proceedings of the 5th Symposium on Biotechnology for fuels and Chemicals, Gatlinberg, T N , May. Ghosh, S. and Henry, M.P. (1981). Stabilization and gasification of soft drink manufac-turing waste by conventional and two- phase anaerobic digestion, in Procedings of the 36th purdue Industrial Wastes Conference, Lafaytte, IN, pp.292. Ghosh, S. et al (1975). Anaerobic acidogenesis of sewage sludge J . Water Pollu. Contr. Bibliography 186 Fed. 47:1-14. Ghosh, S. et al. (1983a). Methane production from industrial wastes by two-phase anaerobic digestion, in Proceedings of the 6th Symposium on Energy from Biomass and Wastes. Orlando, F L . Jan. Gottschalk, G . (1986). Bacteria metabolism. Belin: Springer- Verlag. Grady, C .P .L . and Lim, H . C . (1980) Biological waste treatment. Marcel Dekker, Inc. New York. G u , X.S. (1982). Mathematic Models of Biological WasteWater Treatment. QingHua University, China. Gujer, W . and Zhender, J .B . (1977). Conversion process in anaerobic digestion. Progr. Wat. Technol. 15(8/9):127-167 Hakansson, H. (1977) Anaerobic treatment of whey and whey permeate in submerged filter. Inst. Vatten-luftewardsoal Pupl . B , 406. Hammer, M . J . et al. (1969). Dialysis separation of sewage sludge digestion. J . A m . Soc. Civil Eng. SA5, 907 Harold, F . M . (1966). Inorganic polyphosphates in biology: structure, metabolism and function. Bacteriol. Rev. 30:772-794. Harper, S.R. and Pohland, F . G . (1986) Biotechnology report: recent development in hydrogen management during anaerobic biological wastewater treatment. Biotech-nology and Bioengineering, 28: 585-602. Heertjes, P . M . , L . J . Kuijvenhoven, and R.R. van der Meer. (1982). Fluid flow pattern in upflow reactors for anaerobic treatment of beet sugar factory wastewater. Biotech. Bioeng. 24:443- 459. Heertjes, P . M . and R.R. van der Meer. (1978). Dynamic of liquid flow in an up-flow reactor used for anaerobic treatment of wastewater. Biotech. Bioeng. 20:1577-1594. Bibliography 187 Henze, M.R. and Harremoes, P. (1982). Anaerobic treatment of wastewater in fixed-film reactor: a literature review. Water Science and Technology 15: 1-101. Hichey, R .F . and Owens, R .W. (1981). Methane generation from high -strength indus-trial waste with the anaerobic biological fluidized bed. Biotechnology and Bio-engineering Symp. No. 11: 399-413, Third Symposium, Biotechnology in Energy Production and Conservation, Gatlinburg, Tennessee. Hill, D . J . (1979). Effects of carbon: nitrogen ratio on anaerobic digestion of dairy manure. Agricultural Wastes, 1:267-278. Hill, D . T . , Cobb,S.A. and Bolte, J.P. (1987). Using volatile fatty acid relationship to predict anaerobic digester failure, Transations of A S A E , 30(2):496-501 Hoeks, F . W . J . M . M . , Ten Hoopen H . J . G . , Roels, J . A . and Kuenen, J . G . (1984). Anaero-bic treatment of "acid water" (methane production in a sulfate-rich environment). In "Innovation in Biotechnology" Amsterdam: Elsevir Science. Honcock, R . E . W . (1984). Alteration in outer membrane permeability. Ann. Rev. M i -crobiol. 38:237-264 Hulshoff Pol, L .W. , Dolnng, J., Straaten, K . V . , De Zeeuw, W . J . and Lettinga, G . (1984). Pelletization of anaerobic sludge in U A S B reactor on sucrose containing substrates. Proceeding 3rd int. Symp. on Microbial Ecology. Hulshoff Pol, L .W. , De Zeeuw, W . , Dolnng, J . and Lettinga, G . (1983) Start-up and sludge granulation in UASB-reactor. Paper at Agricultural University, Department of Water Pollution Control, Wagening, Netherlands. Hulshoff Pol, L .W. , Dolfing, J. , de Zeeuw, W . and Lettinga, G . (1982). Cultivation of a well adapted pelletized methanogenic sludge, Biotechnology Letters., 4(5):329-332. Hungate, R . E . (1967). Hydrogen as an intermediate in rumen fermentation. Arch. M i -crobiol. 59: 158-164. Ianotti, F .L . , Kafkewitz, D. , Wolin, M . J . and Bryant, M . P . (1973). Glucose fermentation Bibliography 188 products of ruminococcus albus grown in continuous culture with vibrio succino-gens; changes caused by interspecies transfer of H2- J- Bact. 114: 1231-1240. Ierusalimsky, N . D . (1967). Bottlenecks in metabolism as a growth rate controlling factor. In: Powell E O et al. (ed) Microbial Physicology and Continuous Culture. Her Majesty's Stationary Office, U K . Inmarint Ltd. , 1979 Economic Evaluation of Feeding Liquid Whey to livestock, Part 2 Project, Food Research Branch, Agriculture Canada, Ottawa. Isa, Z., Grusemeyer, S.and Verstraete, W . 1986. Sulfate reduction related to methane production in high-rate anaerobic digestion: Microbiological Aspects. Appl . Envi-ron. Microbiol. 51:580- 587. Keenan, J . D . (1976). Multiple staged methane recovery from solid wastes. J. Environ. Sci. Health, All(8/9):525-531. Kelly, C R . and Switzenbaum, M.S . (1984). Anaerobic treatment: temperature and nutrient effects. Agric. Waste 10:135-54 Kelly, C R . and Switzenbount, M.S.(1984). Anaerobic treatment: temperature and nu-trient effects. Agricultural Waste 10: 135-154. Khan, A . W . and Trottier, T . M . (1978). Effect of sulfur- containing compounds on anaer-obic degradation of cellulose to methane by mixed cultures obtained from sewage sludge. Appl . Env. Microbiol. 35: 1027-1034. Kisaalita, W.S . (1987). Anaerobic fermentation of whey: Acidogenesis Ph.D thesis. The University of British Columbia. Knechtel, J .R. (1978). A more economical method in the determination of chemical oxygen demand, Water Pollution Control, 116:25-7 Kobayashi, H . , Murakami, N . and Unemoto, T . (1982). Regulation of the cyptoplasmic p H in Streptococcus faeclis. J .Biol . Chem., 257:13246-252. Bibliography 189 Kristjansson, T . K . , Schonheit, P. and Thuaer,R.K. (1982). Differental Ks value for hy-drogen of methanogenic bacteria and sulfate reducing bacteria: an explaination for the apparant inhibition of methanogenesis by sulfate. Arch. Microbiol. 131: 278-282. Kristjansson, T . K . and Schonheit, P. (1983). Why do sulfate reducing bacteria outcom-pete methanogenic bacteria for substrate. Oecologia 60:264-271 Kroeker, E.J . , Schulte, D . D . , Sparling, A . B . and Lapp, H . M . (1979) Anaerobic treatment process stability. J. Wat. Pollut. Control Fed. 51: 718-727. Kroiss, H . and Wabnegg, F.P.(1983). Sulfide toxicity with anaerobic wastewater treat-ment. Proceeding of the European Symposium on Anaerobic Wastewater Treat-ment ( A W W T ) p.72-85. The Hague, Netherlands. Kugelman, I.J. and Chin, K . K . (1971) Toxicity, synergism and antagonism in anaerobic waste treatment processes. Adv. Chem. Ser. 105: 55-90. Laidler, K . J . (1968). The chemical kinetics of enzyme action. Oxford University Press Glassgow. Lawrence, A . W . , McCarty, P.L.,Guerin, F . J .A . (1966). The effectc of sulfide on anaerobic treatment. Air Water Pollut. Int.J. 10:207-221. Lawrence, A . W . and McCarty, P.L. (1965). The role of sulfide in preventingheavy metal toxicity inanaerobic treatment. W P C F 37:392-399 Lawrence, A . W . (1971). Application of process kinetics to design of anaerobic processes. Advances in Chemistry Series 105.163 -189. Lettinga, G . and Vinken, J . N . (1980a). Feasibility of the upflow anaerobic sludge blanket (UASB) process for the treament of low-strength wastes. Proceedings of the 35th Industrial waste conference at Purdue University, Lafayette, Indiana, 625-634. Lettinga, G . , De Zeeuw W., Holshoff Pol, L A V . , Wiegant, W and Rinzema, A . (1985). Anaerobic Wastewater treatment based on biomass retention with emphasis on Bibliography 190 UASB-process. Proceedings of the 4th international Sympsium on Anaerobic di-gestion, Guangzhou, China, 279-252. Lettinga, G . , Hobma, S.W., De Zeeuw, W. , Van Velsen F . M . , Hulshoff Pol, L . W . and Klapwijk, A . (1980). Use of the upflow sludge blanket (USB) reactor concept for biological wastewater treatment, especially for anaerobic treatment. Biotech, and Bioeng. 22: 699-734. Lettinga, G . , Van Velsen, A . F . M . , De Zeeuw, W. and Homba, S.W. (1979) The appli-cation of anaerobic digestion to industrial pollution treatment. Proceeding of the 1st International Symposium on Anaerobic Digestion, University College, Card i f f , Wales. 167-187. Lettinga, G . , Roersma, R . E . , Grin, P. (1983). Anaerobic treatment of raw domes-tic sewage at ambient temperatures using a granular sludge bed UASB-reactors. Biotech. Bioeng. 25:1701-1710 Lettinga, G . 1981. Anaerobic digestion for energy saving and production, in: Energy from Biomass, 1st E . C . Conference, Palz W. Chartier,P. and Hall, D . O . (editor). Applied Sceince Publishers Ltd. , London, 264-278. Lettinga, G . , De Zeeuw, W . , Holshoff pol L . W . , Wiegant, W . and Rinzema, A . (1985). Anaerobic wastewater treatment based on biomass retention with emphasis on UASB-process. Proceedings of the 4th International Symposium on Anaerobic Digestion, Guangzhou, China. 279-302. Lettinga, G . , Huishoff Pol, L . W . , Koster, I.W., Wiegant, W . W . , De Zeeuw, W. , Roersma, R . E . and Homba,S.W. (1984). High-rate anaerobic wastewater treatment using the U A S B reactor under a wide range of temperature conditions. Biotechnology and Genetic Engineering Reveiw. Lettinga, G . , Hulshoff Pol, L . W . , Koster, I.W., Wiegant, W . M . , de Zeeruw, W . J. , Bibliography 191 Roersma, R . E . and Hobma, S.W. (1984). High-rate anaerobic waste water treat-ment using the U A S B -reactor under a wide range of temperature conditions. Biotechnology and Genetic Engineering Reviews. Vol.2, 253 Lettinga, G . , Van Velsen, A . F . M . , De Zeeuw, VV. and Hobma, S.W. (1979). The appli-cation of anaerobic digestion to industrial pollution treatment. Proceedings of the 1st International symposium on Anaerobic Digestion, University College, Cardiff, Wales. 167-187. Lettinga,G., van Velsen, A . F . M . , Hobma, S.W., de Zeeuw, W . and Klapwijk, A. (1980).Use of the upflow sludge blanket (USB) reactor concept for biological wastewater treat-ment, especially for anaerobic treatment. Biotech. Bioeng. 22: 699-734. Lo, K . V . and Liao, P.H.(1987). Mesophilic anaerobic digestion of a mixture of cheese whey and dairy manure. Biomass, In press. Lo, K . V . and Liao, P .H. (1986). Digestion of cheese whey with anaerobic rotating biological contact reactor, Biomass 10: 243-52 Lo, K . V . and Liao, P .H. (1986). Digestion of Cheese whey with anaerobic rotating biological contact reactor. Biomass 10: 243-252. Loehr, C L . (1977). Pollution Control for Agriculture. Academic Press, New York, N Y . Mah, R . A . et al. (1978). Studies on an acetate-fermentating strain of methanosarcina. Appl . Environ. Microbiol. 35:1174-1188 Mah, R . A . , Ward, D . M . , Barsi, L . and Glass, T . L . (1977). Biogenesis of methane. Ann. Rev. Microbiol. 31: 309-341 Marshall, D . and Timbers, G . E . (1982). Development and testing of a prototype fixed-film anaerobic digestor. paper No. 82 -6519, Winter Meeting, A S A E St. Joseph, Michigan. Marshall, D and Timbers, G . E . (1982). Development and testing of a prototype fixed-film anaerobic digester. Paper No. 82-6519, Winter meeting of A S A E , St Joseph, Bibliography 192 Michigan Marshall, P . G . , Dunkley, W . L . and Lower, E . , (1968). Fractionation and Concentration of Whey by Reverse Osmosis, Food Technology, Vol.22, No.8, pp.37-44, August. Martens, C.S. and Berner, R . A . (1974). Methane production in the interstitial water of sulfate-depleted marine sediments. See. 185:1167-75. McCarty, P.L. (1972). Energetics of organic matter degradation. Water Pollution M i -crobiology, Wiley -Interscience, New york 91-113. McCarty, P .L . (1981). One hundrad years of anaerobic treatment, paper presented at the 2nd International Symposium on anaerobic digestion. Travemunde, Germany, August. McCarty, P .L . (1964). Anaerobic waste treatment fundamentals, part 2environmental requirements and control. Public Works 95: 123-126 McDonough, F . E . (1968) Whey Concentration by Reverse Osmosis, Food Engineering, Vol . 40, No.3, March. Mclnerney, M . J . and Bryant, M.P. (1981). Review of Methane fermentation fundamen-tals, in Fuel Gas Production from Biomass, C R C Press, Inc. Boca Raton, F L Mclnerney, M . J . and Bryant, M.P . (1979). Anaerobic bacteria that degrade fatty acids in syntrophic association with methanogens. Arch. Microbiol. 122:129-135. Modler, H . W . , Muller, P .G. , Elliot, J . T . and Emmons, D . B . (1980). Economic and technical aspects of feeding whey to livestock. Journal of Dairy Science, 63:838-855. Monod, J . (1942) Rechereches sur la croissance des cultures bacteriennes, Harman, Paris. Monteith, H . D . and Stephonson, J.P. (1981) Mixing efficiences in full-scale anaerobic digestion by tracer methods. J . of Wat. Pollut. Control Fed. 54(l):78-84. Mosey, F . E . (1983). Mathematical modeling of the anaerobic digestion process: regula-tory mechanisms for the formation of short-chain volatile acids from glucose. Water Bibliography 193 Science technology 15: 209-232. Mounfort, D . O . , Asher, R .A. ,Mays , E .L. ,and Tiedje, J . E . (1980). Carbon and elec-tron flow in mud and sandflat intertidal sediments of Delaware inlet, Nelson, New Zealand. Appl . Environ. Microbiol. 39:686-694. Mounfort, D . O . and Asher, R . A . (1979) Effect of inorganic sulfide on the growth and metabolism of Methanosarcina bacteri strain D M . A p p l . Environ. Micobiol. 37: 670-675. Mulder, A . (1984). The effects of high sulfate concentrations on the methane fermentation of wastewater. In "Innovation in Biotechnology". Amsterdam: Elsevir Science. Muller, P . G . (1979). Economic evaluation of feeding liquid whey to livestock, Technical Report, Food Research Institute, Research Branch, Agriculture Canada, Ottawa Niktin, G . A . (1968). Change in oxidation-reduction potential during fermentation in different media by batch cultures of methane-forming bacteria. Microbiol. 4(5):510. Nordstedt, R . A . and Thomas, M . V . (1984) Innocular requirements for start-up of anaer-obic fixed bed reactors. Paper No. 84 -4090, Summer Meeting of A S A E . University of Tennessee, Knoxville. Normann, J . and Frostell, B. (1977). Anaerobic wastewater treatment in a two-stage reactor of a new design, in Proceedings of the 32rd Purdue Industrial Waste Con-ference pp.387 Oremland, R.S. (1982). Methane production and simultanous sulfate reduction in anoxic salt marsh sediments. Nature (London) 296:143-150. Oremland, R.S. and Taylor, B .F . (1978) Sulfate reduction and methanogenesis in marine sediments. Geochim Cosmochim Acta 42: 209-214. Padan, E . , Zilberstein, D. and Schuldiner, S. (1981). p H homeostasis in bacteria. Biochim. Biophys. Acta., 650:151 -166. Bibliography 194 Palmer, G . M . (1978). The utilization of cheese whey for wine production. No.8038 63-01-0, United States Environmental Protection Agency. Industrial Environmental Research Laboratory, Cincinnatti, Ohio. Palmer, G . M . (1979). The conversion of cheese whey into an alcoholic beverage. Proceed-ings of Whey Products Conference(5th) held in Minneapolis, Minnesota, October 9 -10. Agricultural Research Service, United States Department of Agriculture. Patel, G . B . and Roth, L . A . (1978). Acetic acid and Hydrogen metabolism during cocul-ture of an acetic acid producing bacterium with methanogenic bacteria. Can. J . Microbiol. 24: 1009-1010. Pohland, F . G . and Ghosh, S. (1971). Anaerobic stabilization of organic wastes: two-phase concept. Envir. Letters 1: 255-263. Postgate, J .R. (1984). The sulfate reducing bacteria. Cambridge, Cambridge University Press. Reynolds, P.J. and Colleran, E . (1986) Comparison of start -up and operation of anaero-bic fixed-bed and hybrid sludge -bed/fixed-bed reactors treating whey wastewater, P r o c . E W P C A Conf.Anaerobic Treatment, Amsterdam, September, pp.515 Ronnow, P . H . and Gunnarsson, L . A . H . (1982) Response of growth and methane produc-tion to limiting amounts of sulfide and ammonia in two thermophilic methanogenic bacteria. F E M S Microbiol. Lett. 14: 311-315. Ronnow, P .H. and Gunnarsson, L . A . H . (1981), Sulfide dependent methane production and growth of thermophilic methanogenic bacterium.Appl. Envir. Microbiol 42: 580-584. Ross, W . R . (1983) The phenomenon of sludge pelletization in the anaerobic treatment of a maize processing waste. Poster Paper, 3rd Int.Symp. on Anaerobic Digestion, Boston, U S A . Samons, R., Van den Berg, B., Peter, R. and Hade, C . (1984) Dairy Waste treatment Bibliography 195 using industrial scale fixed-film and upflow sludge bed anaerobic digestors: design and start-up experience. Proceedings of the 39th Industrial Waste Conference, Purdue University, pp. 235-241. Samson, R., Van den Berg, B., Peters, R. and Hade, C . (1984). Dairy waste treatment using industrial scale fixed-film and up -flow sludge bed digestors: design and start-up experience. Proceedings of the 39th Industrial Waste Conference, Purdue University, 235-241. Sayed, Sameh, de Zeeuw, W.and Lettinga, G . (1984). Anaerobic treatment of slaughter-house waste using a flocculant sludge UASB-reactor. Agricultural Waste. 11:197-209. Scheifinger, C . C . , Linehan, B. and Wolin, M . J . (1975). H2 production by Slenomonas ru-minantium in the absence and presence of methanogenic bacteria. Appl . Microbiol. 29:480-483. Scherer, P. and Sahm (1981). Influent of sulfur containing compounds on the growth of Methanosarcina bacter in a defined medium.European J . Appl . Microbiol. Biotech-nol. 12: 28-35. Schonheit, P., Kristjansson, J . K . and Thuaer, R . K . (1982). Kinetic mechanism for the ability of sulfate reducers to out-compete methanogens for acetate. Arch. Micro-biol. 132: 285-288. Schulmann, G . E . , Stanley, M . A . and Knudsen, D . (1973). Automated total nitrogen analysis of soil and plant sample, Proc. Soil Sci. Soc. America, 37: 480-481 Singh, R . K . and Ghaly, A . E . (1984). Single cell protein production from cheese whey, A S A E PAper No. 84 -6528, St. Joseph, Michigan. Singh, R .K. and Ghaly, A . E : (1985). Feasibility of cheese whey processing for production of food and feed supplement C S A E paper No. 85-504, Ottawa, Ontario. Sorensen, J . (1981). Volatile fatty acids and hydrogen as substrates for sulfate-reducing Bibliography 196 bacteria in anaerobic marine sediment. Appl . Enviro. Microbiol., 42: 5-16. Speece, R . E . and McCarty, P.L. (1964). Nutrient requirements and biological solid ac-cumulation in anaerobic digestion. Adv. in Wat.Poll. Res.2: 305-322. Speece, R . E . (1985). Review-microbiol conversion of biomass to methane, unpublished paper in Environmental Studies Institute, Drexel University. Speece, R . E . and Parkin, G .F . (1983). The response of methane bacteria to toxicity. Proceedings of the 3rd International Symposium on Anaerobic Digestion, p.23-35. Evans and Faulkner. Inc., Watertown, Mass. Stephen, R . H . and Frederick, G.P. (1986). Biotechnology Report: Recent develop-ments in hydrogen management during anaerobic biological wastewater treatemnt. Biotech. Bioeng. 28: 585- 602. Strayer, R .F . and Tiedjie, J . M . (1978). Kinetic parameters of the conversion of methane precursors to methane in a hypereutrophic lake sediment. Appl . Environ. Micro-biol. 36: 330-339 Stronach, S .M. , Rudd, T . and Lester, J . N . (1986). Anaerobic digestion processes in industrial wastewater treatment. Betin Heidelberg , New York, Tokyo Sutton, P . M . and Li , A . (1981). Anitron and oxitron systems: high-rate anaerobic bio-logical treatment systems for industry. Proceedings of 36th Industrial Waste Con-ference, Purdue University. Switzenbaum, M.S. (1979). Methane generation from whey for energy production and pollution control. Energy Optimization of Water and Wastewater Management for Municipal and Industrial Applications. Department of Energy conference, New Orleans, L A . Switzenbaum, M.S. and Danskin, S.C. (1982). Anaerobic expanded bed treatment of whey. Agricultural Waste, 4: 411-426. Teixeira, A . A . , Johnson, D . E . and Zall, R.R. (1982). Outlook for the utilization in the Bibliography 197 US whey processing industry. Presented at the Winter meeting of the American Society of Agricultural Engineers. December 15-17, Chicago, Illinois. Thuaer, R . K . , Jungermann, K. and Decker, K. (1977) Energy conservation in chemotrophic anaerobic bacteria. Bacterial Rev. 41: 100-180. Thuaer, R . K . (1982). Dissimilatory sulfate reduction with acetate as electron donor. In: Philosophical Transactions of the Royal Society, London. Thusman, J .F . and Douglas, J . C . (1989). Influent of sulfate and sulfate-reducing bacteria on anaerobic digestion technology. Biological Waste Treatment. Alan R. Liss, Inc. Van der Meer, R.R. and Heertjes, P . M . (1983). Mathematical description of anaerobic treatment of wastewater in upflow reactors. Biotech. Bioeng. 25: 2531-56. Ven den Meer, R.R. and Vletter, R.de (1982). Anaerobic treatment of wastewater: the gas-liquid-sludge separator. Journal W P C F 54(11): 1482-92. Wang, Z., Chen, Z. and Qian, Z. (1985). Status quo and prospects on the study of anaerobic disposal for industrial wastewater. Proceedings of the 4th International Symposium on Anaerobic Digestion, Guangzhou, China 259-277. Weimer, P.J. and Zeikus, J . G . (1980). One carbon metabolism in methanogensis bacteria, Cellular characterization and growth of methanosarcina barkeri. Arch. Microbiol.,119:49-Wildenauer, F . X . and Winter, J . (1985). Anaerobic digestion of high-strength acidic whey in a pH-controlled up-flow fixed-film loop reactor. Appl . Microbiol. Biotechnol., 22: 367-72 Williams, D . W . (1984). Anaerobic digestion of cheese whey, paper No. 84-4095. A S E A Summer meeting, University of Tennessee, Knoxvill. Winfrey, M . R . and Zeikus, J . G . (1977). Effect of sulfate on carbon and electron flow dur-ing microbial methanogenesis in freshwater sediments. Appl . Environ. Microbiol. 33: 275-283. Bibliography 198 Winfrey, M.R. , Nelson, D . G . R . , Klevickis, S.C. and Zeikus, J . G . (1977). Association of hydrogen metabolism with methanogenesis in Lake Mendota sediments. A p p l . Envir. Microbiol. 33(2): 312-318. Wolin, M . J . (1974). Metabolic interactions among intestinal microorganisms. A m . J .Clin. Nutr. 27: 1320-1328. Wolin, M . J . (1976). Interaction between # 2 -P r o ducing and methane-producing species, in H.G.Schlegel, G.Gottschalk and N.Pfennig (Eds.), Sympsium on Microbial Pro-duction and Utilization of Gases (H2, CO2, C O ) . E . Goltze K . G . , Gottingen, pp 141-150. Wu, W. , Hu, J. and G u , X. (1985). Properties of granular sludge in up-flow anaerobic sludge blanket (UASB) reactor and its formation. Proceedings of the 4th Interna-tional Symposium on Anaerobic Digestion, Guangzhou, China. 339-351. Wu, Y . and Zhang, W . (1983). Report on pilot-scale experiment disposing monosodium glutamate wastewater with up-flow anaerobic digestion. Annual report at Guangzhou Institute of Energy Conversion, Chinese Academy of Sciences. Yan, J .Q. , Lo, K . V . and Liao, P .H. (1988). Methane production from cheese whey. Biomass (17) 185-202 Yan, J . Q . , Lo, K . V . and Liao P .H. (1989). Anaerobic digestion of cheese whey using up-flow anaerobic sludge blanket reactor. Biological Waste 27: 289-305 Yan, J . Q . , Liao, P.H. and Lo, K . V . (1990). Anaerobic digestion of cheese whey using an up-flow anaerobic sludge blanket reactor: 3.sludge and substrate profiles. Biomass 21: 257 -271 Yang, S.T.(1984). Kinetics study of defined methane fermentation of lactose, P h . D . Thesis, Purdue University. Yano, T . , Nakahara, T . , Kamiyama,S. and Yamada, K. (1966). Kinetic studies on mi-crobiol activities in concentrated solution. Agr. Biol. Chem. 30(1): 42-48. Bibliography 199 Young, J . C . and McCarty, P.L. (1969). The anaerobic filter for waste treatment J . Water Pollut. Control Fed. 41: R160 -R163. Zall, R . H . (1979). Whey treatment and utilization, Proceeding of the 1989 Food Pro-cessing Waste Conference, Atlanta, Georgia. Zehnder, A . J . B . (1978). Ecology of methane formation. In:R Mitchell (ed) Water pollu-tion microbiology, Vol.2 John Wiley and Sons Inc. New York, pp.349-376. Zehnder, A . J . B . and Brock, T . D . (1979). Biological energy production in the apparent absence of electron transport and substrate level phosphorylation. F E B S Lett. 107: 1-3. Zehnder, A . J . B . , Huser, B . A . , Brock, T . D . and Wuhrman, K. (1980). Characterization of an acetate-decarboxylating, non -hydrogen-oxidizing methane bacterium. Arch. Microbiol. I l l : 199-205. Zeikus, J . G . (1977). The biology of methanogenic bacteria. Bact. Rev. 41:514-541. Zeikus, J . G . (1982). Microbial intermediary metabolism in anaerobic digestion, in D . E . Hughes: Anaerobic digestion'1981. Elsevier Biomedical Press, N.Y . pp.23-35 Appendix A E M P I R I C A L M O D E L In such circumstances when the mechanism underlying a process is not understood suffi-ciently well, or is too complicated to allow an exact model to be postulated from theory, an empirical model may be useful. Particularly if it is desired to approximate the response only over limited ranges of the variables. A . l T H E P R O G R A M OF MINITAB Minitab's regression is used to investigate the relationship among variables, particularly those relationships that enable the experimenter to predict a variable from one or more others. Minitab's regression has three features: 1. It applies simple linear regression to investigate the form of a linear relationship between two variables 2. Multiple regression is for the investigation of the form of the linear relationship between one variable and several other variables 3. Stepwise regression allows one to explore the relative importance of various pre-dictor variables. To invoke the regression command, the necessary information, including the depen-dent variable and its location (column number where it is stored) and independent vari-ables as well as their locations, must be provided. After performing a simple linear regression by R E G R E S S , M I N I T A B first gives the default output. The first fine is the estimated regression equation. The second fine is 200 Appendix A. EMPIRICAL MODEL 201 the estimate for each of the regression coefficients, their estimated standard deviations, the t-statistics and p-values for testing whether the coefficient is different from zero. The third line is the the estimated standard deviation of the regression, the R -squared value, or coefficient of determination and the adjusted R-squared. The final part of the output is the significance test for the regression in an A N O V A form, with a F-statistic and p-value as before. A n additional output can be got by issuing the " B R I E F L E V E L 3" Command in conjunction with the " R E G R E S S " C O M M A N D as show in Table A-6. Additional infor-mation is the independent value, the dependent (response) value, the fitted or predicted value (the estimated mean value of the dependent variable from the independent value and the regression equation), the standard error of the fitted value, the residual and the standardized residual. Use of S T E P W I S E command allows one to select from a large set of candidates the independent variables that best predict the dependent variable. A.2 ANALYSIS OF T H E A C C U M U L A T I O N OF A C E T I C ACID A.2.1 Determination of the Equation Using the data in Table A - l and running the program of Minitab's regression, we have following results: Due to the shortage of data from the original experimental design, we can not com-pletely go through the Stepwise procedure, as shows in Table A.2. But, through step by step of addition of each parameter, it was found that the inclusion of the squared terms increased the value of R-squared and decreased the S value. Here S is called "the standard deviation of y about the regression fine" or "the standard error of estimate". The value S can be thought of as a measure of how much the observed y values differ Appendix A. EMPIRICAL MODEL 202 Table A . l : A c c u m u l a t i o n of A c e t i c A c i d In the Absence of Sulfate Influent (g COD/1) AA at 1# (g/1) AA at 2# • (g/1) AAi-AAo (g/1) 4.56 0.382 0.152 0.23 9.93 0.688 0.096 0.59 17.7 0.876 0.02 0.856 28.8 1.166 0.04 1.126 38.1 2.895 2.553 0.342 1# sample port 1 2# sample port 2 In the presence of Sulfate Influent (g GOD/1) AA at 1# (g/1) AA at 2# (g/1) AAi-AAo (g/D 15 1.038 0.103 0.935 20 1.163 0.043 1.120 30 1.403 0.073 1.330 40 1.600 0.057 1.543 50 2.127 1.714 0.413 Appendix A. EMPIRICAL MODEL 203 Table A.2: Stepwise Analys is for A c c u m u l a t i o n of A c e t i c A c i d MTB > MTB > s t e p w i s e y i n c 2 , x i n c l c 3 - c 6 S T E P W I S E R E G R E S S I O N OF A A ON 5 P R E D I C T O R S , WITH N = 5 S T E P 1 C O N S T A N T 0 . 5 1 4 3 C o * * 3 0 . 0 0 0 0 4 T - R A T I O 8 . 5 0 S 0 . 2 0 7 R - S Q 9 6 . 0 1 M O R E ? ( Y E S , N O , SUBCOMMAND, OR H E L P ) S U B O S U B O j;uu<:> from the corresponding average y value If a cubic term was included, the R-squared value went further up to 96 and the S value went down. It was obvious that the best equation was one which included influent concentration, its square and cubic terms as follow: AA = -0 .27 + 0.183Co - 0.0102Co2 + 0.000197Co3 ( A . l ) A . 2 . 2 A n a l y s i s of the A d e q u a c y of the E q u a t i o n The evaluation of the residuals and fitted values can be used to examine the model ade-quacy. Residual is the difference between the observation and the fitted value determined by regression equation. Thus residuals tell us how the model missed in fitting the data. Appendix A. EMPIRICAL MODEL 204 Table A.3: Stepwise A n a l y s i s for A c c u m u l a t i o n of A c e t i c A c i d The regression equation i s AA = - 0.032 + 0.0601 < Co Predictor Coef Stdev t - r a t i o P Constant -0.0317 0.3560 -0.09 0 .935 Co 0.06010 0.01528 3.93 0.029 s = 0.4185 R-sq = 83.8% R-sq(adj) = = 78.3% Analysis of 1 Variance SOURCE DF SS MS F P Regression 1 2 .7089 2.7089 15. 47 0.029 Error 3 0 .5255 0.1752 T o t a l 4 3 .2344 Obs. Co AA F i t Stdev.Fit Residual St .Resic 1 4.6 0.382 0.242 0.299 0 .140 0. 48 2 9.9 0.688 0.565 0.241 0.123 0. 36 3 17.7 0.876 1.032 0.190 -0.156 -0. 42 4 28.8 1.166 1.699 0.232 -0.533 -1. 53 5 38.1 2.685 2.258 0.336 0.427 1. 71 Appendix A. EMPIRICAL MODEL 205 Table A.4: Stepwise Analysis for A c c u m u l a t i o n of A c e t i c A c i d MTB > MTB > regr c l 2 c2 c6 The regression equation i s AA = - 0.058 + 0.0597 Co + 0.171 Co**1.5 Predictor Constant Co Co**1.5 Coef -0.0577 0.05969 0.1711 Stdev 0.3745 0.01603 0.2004 t - r a t i o -0.15 3.72 0.85 P 0.892 0.065 0. 483 s = 0.4388 R-sq = 88.1% Analysis of Variance R-sq(adj) = 76.2% SOURCE DF SS MS Regression 2 2.8494 1.4247 Er r o r 2 0.3850 0.1925 T o t a l 4 3.2344 F 7.40 0. P 119 SOURCE Co Co**1.5 DF 1 1 SEQ SS 2.7089 0.1405 CONTINUE? Obs. Co AA F i t Stdev.Fit Residual St.Resid 1 4.6 0.382 0.386 0.355 -0.004 -0.01 2 9.9 0.688 0.364 0.345 0.324 1.20 3 17.7 0.876 1.170 0.256 -0.294 -0.83 4 28.8 1.166 1.490 0.345 -0.324 -1.20 5 38.1 2.685 2.387 0.384 0.298 1.40 Appendix A. EMPIRICAL MODEL 206 Table A.5: Stepwise Analysis for Accumulation of Acetic Acid MTB > regr c l 2 c2 c5 The regression equation i s AA = 1.87 + 0.180 Co - 1.02 Co**0.5 Predictor Coef Stdev t - r a t i o P Constant 1 .870 1.739 1.08 0.395 Co 0. 1799 0.1084 1.66 0.239 Co**0.5 -1. 0164 0.9113 -1.12 0.381 s = 0.4025 R-sq = 90.0% R-sq(adj) = = 80.0% Analysis of Variance SOURCE DF SS MS F P Regression 2 2. 9104 1.4552 8.98 0.100 Erro r 2 0. 3240 0.1620 T o t a l 4 3. 2344 SOURCE DF SEQ SS Co 1 2. 7089 Co**0.5 1 0. 2015 CONTINUE? Obs. Co AA F i t Stdev.Fit Residual St.Resid 1 4.( 0.382 0.520 0.380 -0.138 -1.04 2 9.9 0.688 0.453 0.252 0.235 0.75 3 17.7 0.876 0.778 0.292 0.098 0.35' 4 28.8 1.166 1.596 0.242 -0.430 -1.34 5 38.1 2.685 2.450 0.366 0.235 1.41 Appendix A. EMPIRICAL MODEL 207 Table A.6: Stepwise Analys is for A c c u m u l a t i o n of A c e t i c A c i d MTB > b r i e f output l e v e l 3 MTB > regr c l 2 c2 c3 The regression equation i s AA = 0.666 - 0.0381 Co + 0.00230 Co**2 Predictor Coef Stdev t - r a t i o p Constant 0.6665 0.4642 1.44 0.288 Co -0.03808 0.05469 -0.70 0.558 Co**2 0.002299 0.001252 1.84 0.208 s = 0.3128 R-sq = 93.9% R-sq(adj) = 87.9% Analysis of Variance SOURCE DF SS MS F p Regression 2 3.0387 1.5193 15.53 0.061 Erro r 2 0.1957 0.0979 T o t a l 4 3.2344 SOURCE DF SEQ SS Co 1 2.7089 Co**2 1 0.3298 CONTINUE? Obs. Co AA F i t Stdev.Fit Residual 1 4.6 0.382 0.541 0.276 -0.159 2 9.9 0.688 0.515 0.182 0.173 .3 17.7 0.876 0.713 0.225 0.163 4 28.8 1.166 1.476 0.212 -0.310 5 38.1 2.685 2.552 0.298 0.133 St.Resid -1.08 0.68 0.75 -1.35 1.40 Appendix A. EMPIRICAL MODEL 208 Table A.7: Stepwise A n a l y s i s for A c c u m u l a t i o n of A c e t i c A c i d MTB > b r i e f output l e v e l 3 MTB > r e g r c l 2 c2 c4 The r e g r e s s i o n e q u a t i o n i s AA = 0.540 - 0.0026 Co +0.000039 Co**3 P r e d i c t o r Coef Stdev C o n s t a n t 0.5400 0.3161 Co -0.00260 0.02695 Co**3 0.00003910 0.00001578 s = 0.2541 R-sq = 96.0% A n a l y s i s of V a r i a n c e SOURCE DF SS R e g r e s s i o n 2 3.1053 E r r o r 2 0.1291 T o t a l 4 3.2344 SOURCE DF SEQ SS Co 1 2.7089 Co**3 1 0.3964 CONTINUE? t - r a t i o 1.71 -0.10 2. 4 8 P 0.230 0.932 0.132 R-s q ( a d j ) = 92.0% MS 1.5526 0.0646 F 24.05 P 0.040 CONTINUE? Obs. Co AA 1 4.6 0.382 2 9.9 0.688 3 17.7 0.876 4 28.8 1-166 5 38.1 2.685 F i t S t d e v . F i t R e s i d u a l 0.532 0.552 0.711 1.399 2.603 0.216 0.146 0.174 0.186 0.247 -0.150 0.136 0.165 -0.233 0.082 S t . R e s i d -1.12 0.65 0.89 -1.34 1.39 Appendix A . EMPIRICAL MODEL 209 Table A.8: Stepwise A n a l y s i s for A c c u m u l a t i o n of A c e t i c A c i d MTB > b r i e f output l e v e l 3 MTB > regr c l 3 c2-c4 * NOTE * Co i s h i g h l y c o r r e l a t e d with other p r e d i c t o r v a r i a b l e s * NOTE * Co**2 Is hi g h l y c o r r e l a t e d with other p r e d i c t o r v a r i a b l e s * NOTE * Co**3 i s h i g h l y c o r r e l a t e d with other p r e d i c t o r v a r i a b l e s The r e g r e s s i o n equation is AA = — 0 . 2 7 0 + 0 .183 Co - 0 . 0 1 0 2 Co**2 +0.000197 Co**3 P r e d i c t o r Coef Stdev t - r a t i o P Constant - 0 . 2 7 0 3 0 .1449 -1 . 8 7 0.313 Co 0 .18258 0 .02967 6 . 1 5 0 .103 C o * * 2 -0 .010247 0 .001609 -6.37 0.099 Co**3 0 .00019653 0 .00002496 7.88 0.080 s = 0 . 0 5 5 7 3 R-sq = 99.9% R-sq(adj) = 99.6% A n a l y s i s of Variance SOURCE DF SS MS F P Regression 3 3 . 2313 1 . 0 7 7 1 346.81 0 . 039 E r r o r 1 0 . 0031 0 . 0 0 3 1 T o t a l 4 3 . 2344 CONTINUE? SOURCE ' DF SEQ SS ' Co i 2 . 7089 Co* * 2 1 0 . 3298 Co * * 3 1 0 . 1926 Obs. Co AA P i t S t d e v . F i t Residual S t . R e s i d 1 4 .6 0 . 3 8 2 0 0 . 3 6 7 8 0 . 0 5 3 9 0 . 0 1 4 2 1 . 00 2 9 .9 0 . 6 8 8 0 0 . 7 2 4 7 0 . 0 4 1 9 - 0 . 0 3 6 7 - 1 . 00 3 1 7 . 7 0 . 8 7 6 0 0 . 8 4 0 8 0 . 0 4 3 2 0 . 0 3 5 2 1 . 00 4 28 .8 1 . 1 6 6 0 . 1 . 1 8 3 2 0 . 0 5 3 0 - 0 . 0 1 7 2 - 1 . 00 5 3 8 . 1 2 . 6 8 5 0 2 . 6 8 0 5 0 . 0 5 5 5 0 . 0 0 4 5 1 . 00 X X denotes an obs. whose X value gives i t large i n f l u e n c e . Appendix A. EMPIRICAL MODEL 210 Figure A - l is a plot of the fitted values vs the actual values. This plot shows a ver3r good fit between the model and the data. In addition, plots of the residuals vs fit, actual values and independent variables were all made (Figures A-2 to A-5). Strong pattern or tendency in the residual plot indicates that we probably have a poor model. If the model is serious wrong, the residuals would tend to more positive for some parts of x values, and more negtive for others. However, none of them showed up any problem regarding the model adequacy. A.2.3 The Comparison between the Experimental Data and the Model Figure A - 6 presents the experimental data (points) and the model (curve). It shows a good fit. A.3 S U M M A R Y OF EMPIRICAL M O D E L S Following exactly the same procedure as described in the above section, the models for degradation of acetic acid and models for accumulation and degradation of propionic acid and total V F A s in both the absence and presence of sulfate were developed as fo A.3.1 In the Absence of Sulfate Accumulation of AA (acetic acid) AA = -0.27 + 0.183Co + -0.0102Co2 + 0.000197Co3 ' (A.2) Degradation of A A Ra (Table A.9) Appendix A. EMPIRICAL MODEL 211 MTB > MTB > plot c31 c l _ * 2.40 + f i t 1.60 + _ * 0.80+ * * _ * + + — - - - + + + - : A A 0.50 1.00 1.50 2.00 2.50 Figure A . l : Fitted Values versus the Actual Values Appendix A. EMPIRICAL MODEL 212 MTB > p r i n t c l c31 c22 ROW AA f i t r e s i d u a l 1 0. 382 0. 3678 0 .0141891 2 0. 688 0. 7247 -0 .0367000 3 0. 876 0. 8408 0 .0352088 4 1. 166 1. 1832 -0 .0172389 5 2 . 685 2. 6805 0 .0045412 MTB > MTB > •II > MTB > p l o t c22 C31 - * 0.025+ r e s i d u a l - * - * 0.000+ - * -0.025+ - * + + + + + f i t 0.50 1.00 1.50 2.00 2.50 MTB > MTB > Figure A.2: A n a l y s i s of residuals Appendix A. EMPIRICAL MODEL 213 MTB > MTB > p l o t c22 c l 0.025+ r e s i d u a l - * 0.000+ -0.025+ 1.50 2.00 + -2.50 -AA 0.50 1.00 MTB > Figure A.3: A n a l y s i s of residuals Appendix A. EMPIRICAL MODEL 214 MTB > p l o t cc22 c2 0.025+ r e s i d u a l - ' * 0.000+ -0.025+ + + + + Co 7 - ° 14-0 21.0 28.0 35.0 MTB > MTB > Figure A.4: Analysis of residuals Appendix A. EMPIRICAL MODEL 215 MTB > MTB > p l o t c22 c3 0.025+ r e s i d u a l - * 0.000+ — * -0.025+ +-0 300 +-600 + + + Co**2 900 1200 1500 MTB > Figure A.5: Analysis of residuals Appendix A. EMPIRICAL MODEL 216 0 x 10 20 30 40 50 hfijent Concentjation (g CCD/ft Figure A.6: Comparison of Experimental Data with Model Appendix A. EMPIRICAL MODEL 217 Table A.9: Regression Analysis for Degradation of Acetic Acid MTB > print cl2-cl5 ROW l / r C C**2 C**(-2) 1 4 .34783 4 .56 20.79 0 .0480917 2 1 .68919 9 .93 98.60 0 .0101415 3 1 .16822 17 .70 313.29 0 .0031919 4 0 .88810 28 .80 829.44 0 .0012056 5 2 .92398 38 .10 1451.61 0 .0006889 MTB > The regression equation is l / r = 3.68 - 0.293 C + 0.00707 C**2 + 38.1 C**(-2) Predictor Constant C C**2 C**(-2) Coef 3.682 -0.2927 0.007075 38.14 Stdev 1.993 0.1703 .003305 30.45 t-ratio 1.85 -1.72 2.14 1.25 P 0.316 0.335 0.278 0.429 s = 0.4739 R-sq = 97.3% Analysis of Variance SOURCE DF SS Regression 3 7.9592 Error 1 0.2246 Total 4 8.1838 R-sq(adj) = 89.0% MS 2.6531 0.2246 F 11.81 P 0.210 SOURCE C C**2 C**(-2) CONTINUE? DF 1 1 1 SEQ SS 0.7734 6.8336 0.3523 Obs. C l / r F i t Stdev.Fit Residual St.Resid 1 4.6 4.348 4.328 0.474 0.020 1.00 X 2 9.9 1.689 1.859 0.442 -0.170 -1.00 3 17.7 1.168 0.839 0.341 0.329 1.00 4 28.8 0.888 1.166 0.384 -0.277 -1.00 5 38.1 2.924 2.825 0.464 0.099 1.00 X denotes an obs. whose X value gives i t large influence. Appendix A. EMPIRICAL MODEL 218 = 3.68 - 0.293Co + 0.00707GV + 38.lCo~ 2 (A.3) Ra Accumulation of propionic acid (PA) PA = 0.112 - 0.00893Co + 0.00108GV (A.4) Degradation of PA Rp -J- = -18.3 + 0.488Co + 216CO-1 (A.5) Rp Accumulation of VFA VFA = 2.12 - 0.0325Co - 5.76Co _ 1 (A.6) Degradation of VFA Rv — = -0.648 + 0.0288Co + 9.02CO" 1 Rv (A.7) Appendix A. EMPIRICAL MODEL A . 3 . 2 In the Presence of Sulfate Accumulation of A A AAs = 0.504 + 0.0590Co - 0.00172CV + 0.000024Co = 15.6 - 0.63Co + 0 .00815CV - 104Co _ 1 Rs Accumulation of P A PAs = 0.603 - 0.039lCo + 0.001416V 4- = 11-0 - 0.54Co -f 0.00715GV Rs Accumulation of V F A VFA = 1.77 - 0.0552Co + 0.00215CV Degradation of V F A Appendix A. EMPIRICAL MODEL 220 - j - = 3.36 - 0.213Co + 0.00367GV (A.13) Rs Appendix B E F F E C T S OF S U L F A T E IN B A T C H E X P E R I M E N T S B . l I N T R O D U C T I O N When wastewater is subjected to anaerobic digestion, a high concentration of sulfate has been thought to be toxic for methanogenic bacteria (see literature review). Recent research has greatly increase the knowledge about the function of sulfate in the anaerobic process. It is known that the sulfate requirement of methanogens in anaerobic digestion is a complex function. On one hand, sulfate reducing bacteria help to maintain -the anaerobic condition required for the growth of methanogens. Sulfur, which is the reduced product of sulfate in the anaerobic process, is also a nutrient necessary for methanogens. O n the other hand, the addition of sulfate inhibits methanogenesis. The inhibition from sulfate is attributed to that methanogens and sulfate reducing bacteria competing for the common substrates, acetic acid and hydrogen. In the presence of excess hydrogen, they have no effects on each other. When the substrates supply becomes rate hmiting, however, competition does take place. As described in earlier chapters, unsuccessful experiments on anaerobic digestion of cheese whey have been reported by a number of researchers. Little research, however, has been done so far to solve or to explain these failures. The problems encountered in previous studies has been generally attributed to inadequate buffering capacity and micronutrient deficiency of cheese whey. According to the mechanisms considered in Chapter 7, it can be assumed that sulfate, in the proper concentration, may be applied 221 Appendix B. EFFECTS OF SULFATE IN BATCH EXPERIMENTS Table B . l : Independent Variables in Batch Experiment 222 Variables Levels + . -Sulfate (g/1) 0.6 0 Cheese W h e y C o n c e n t r a t i o n (g C O D / 1 ) 6 3 Sludge Concentra t ion (g/1) 1.5 0.75 to moderate the detrimental influences of excess hydrogen on a stressed anaerobic reactor. The practical question is what sulfate concentration in an industrial wastewater of a given composition, such as cheese whey, makes anaerobic digestion possible and success-ful. This preliminary experiment, therefore, was carried out to find how sulfate affects the anaerobic digestion system B.2 P R E L I M I N A R Y E X P E R I M E N T A L D E S I G N B.2.1 Variables The independent variables are presented in Table B - l . The concentration of cheese whey was chosen based on consideration of the possible organic loading rate during the start-up period for successful treatment. According to previous studies for a cheese whey anaerobic process, the organic loading rate cannot Appendix B. EFFECTS OF SULFATE IN BATCH EXPERIMENTS 223 Table B.2: Experimental Design of Batch R u n Table B-2 Experimental Design of Batch R u n R u n S C o A 1 _ 2 •+ - -3 - + -4 + + -5 - - + 6 + - . + 7 - + + 8 + + Appendix B. EFFECTS OF SULFATE IN BATCH EXPERIMENTS 224 exceed 0.3 g C O D / g VSS at the start-up. The maximum concentration of sludge available in the lab is about 20 g VSS/1. Therefore, the upper limit of cheese whey concentration was 6 g C O D / 1 . A 3 g C O D / 1 difference would make the response variables, such as gas composition and treatment efficiencj', different. So 3 g C O D / 1 was the lower limit. The possible lower limit of sludge concentration might be 10 g/1 based on the same consideration of organic loading rate. The concentration of sulfate was chosen based on the literature review. The toxic levels of sulfate for methanogens varied from one study to another. Generally, the optimal concentration of sulfate for methanogens growth in the anaerobic process was reported in the range of 0.1-1 m M , which is equal to 0.014-0.14 g Na2S04/\. Based on the mechanism of competition of both methanogens and sulfate reducing bacteria for common substrates, it would be reasonable to consider the ratio of C O D concentration and sulfate in the experiment. The only information that could be had from literature indicated that when the ratio of C O D to sulfate decreased below 10, sulfate reduction would occur and methanogenesis would be inhibited. The upper limit concentration of sulfate was set in such a way that the lowest ratio of C O D / N a 2 S 0 4 was about 5. Therefore, the sulfate concentration should be 0.6 g/1 at its upper limit. The lower Hmit was set at zero to maximize the difference. B.2.2 E x p e r i m e n t a l D e s i g n A full 2 3 factorial design with duplicates was used in this experiment. The layout is shown in Table B-2 Appendix B. EFFECTS OF SULFATE IN BATCH EXPERIMENTS 225 B . 3 E X E C U T I O N O F T H E E X P E R I M E N T B.3.1 Substrate Cheddar cheese whey was used in this study. The C O D concentration of the original substrate was as high as 65 g C O D / 1 and pH as low as 4. The cheese whey was dilute to the desired concentrations, 6 gCOD/1 and 3 g C O D / 1 . Then potassium hydroxide was used to adjust the pH to 7 B.3 .2 Seed Sludge (Bacteria) The seed sludge was obtained from the effluent of a laboratory-scale, upflow anaerobic sludge blanket (UASB) reactor and had a concentration of about 20 g VSS/1. B.3 .3 Reac tor O p e r a t i o n 16 flasks with volume of 250 ml were used to as the batch reactors in this experiment. Each flask was seeded 75 ml sludge with two different VSS concentrations as desired in experimental design. They were sampled and fed once a week. In the first week each of flasks was fed 100 ml substrate liquid containing different concentrations of sulfate and cheese whey according to the experimental design. In the subsequent week, from each of them was drawn a 60 ml sample for analysis and fed 60 ml of the substrates which ware exactly the same composition as the first time. A l l the flasks were randomly located in the incubator under a constant temperature of 3 5 ° C . Appendix B. EFFECTS OF SULFATE IN BATCH EXPERIMENTS 226 B.3.4 C h e m i c a l Analyses Gas production was collected daily for each reactor. Gas composition was analyzed ev-ery other day. The liquid in the flasks was withdrawn once a week for analysis of C O D (Chemical Oxygen Demand), V F A (Volatile Fatty Acids), p H and residual sulfate. Both gas composition and V F A were analyzed on a Hewlett Packard 5890A gas chromatog-raph}'. C O D was determined by a colorimetric method. Sulfate analysis was conducted according to the "standard methods". B .4 R E S U L T S A N D D I S C U S S I O N B.4 .1 Effects of Sulfate (S), Feed Strength ( C o ) a n d Seed S l u d g e A m o u n t ( A ) on the M e t h a n e P r o d u c t i o n (CH4%) A n anaerobic process is considered to be a step-wise process which proceeds in different successive stages. Anaerobic conversion of the acid products into methane and carbon dioxide, also called "methanogenesis", is the terminal and rate-limiting step of the whole sequence. Methane formation is accomplished by the methane bacteria. Methane content in the produced gas is one of the indicators of activity of the methane bacteria as well as of the system's performance. The experimental results are summarized in Table B-3. Subscripts a, b and c stand for sampling time, the first, the third and the fourth week from the start of the experiment respectively. Using the J A S S program, the main effects of S, Co and A and their interactions were estimated, which are presented in Table B-4. During the first week of the experiment, the effect of Co was the most significant one. As we can see from the cube plots (Figure B- l ) and the data analysis, CH4% increased an average of about 7.4 units when the influent concentration decreased from 6 g C O D / 1 to 3 g C O D / 1 . In comparison with the effect of Co , the effects of S and A were less Appendix B. EFFECTS OF SULFATE IN BATCH EXPERIMENTS Table B.3: E x p e r i m e n t a l Results 227 R o w S u l f a t e C O D S l u d g e C H 4 % - a C H 4 % - b C H 4 % - c p H 1 -1 1 -1 60.15 47.65 48.48 6.500 2 1 -1 -1 59.29 45.26 48.41 6.725 3 -1 -1 -1 52.99 49.30 45.47 6.445 4 1 1 -1 52.09 49.25 46.81 6.615 5 -1 1 1 60.30 48.06 49.88 6.770 6 1 -1 1 57.70 47.08 46.65 6.805 7 -1 -1 1 52.25 46.61 54.02 6.500 8 1 1 1 50.11 46.92 48.12 6.750 9 -1 1 -1 61.13 48.12 49.00 6.500 10 1 -1 -1 59.66 44.74 46.48 6.715 11 -1 . -1 -1 54.82 50.06 44.84 6.430 12 1 1 -1 51.75 49.70 47.12 6.605 13 -1 1 1 60.10 53.36 51.48 6.785 14 1 -1 1 58.62 49.05 47.56 6.830 15 -1 -1 1 51.92 47.00 47.28 6.490 16 1 1 1 51.55 47.19 46.87 6.740 significant during the first week. The less significant effect of S might be attributed to two reasons. O n one hand, sulfate reducing bacteria (SRB) might not have built up and less competitive in the first week. It could also be due to the fact that growth of S R B was limited by the low concentration of sulfate in the culture. A higher concentration of sulfate addition might have led to a statistically significant effect of sulfate on the gas composition. As time passed, it became clear that the effect of S on the gas composition was somehow related to the feed strength. Look at the upper surface (higher strength) and the lower surface (lower strength) of the cube plot (Figure B-2, C r 7 4 % in the third week). For lower feed strength of 3 g C O D / 1 , the Cr7 4 % in the generated gas changed from 47.9% in the absence of sulfate to 45% in the present of sulfate, or from 50.7% to 48.1%. While Appendix B. EFFECTS OF SULFATE IN BATCH EXPERIMENTS 228 CH4I-a 52.1 / / : / : / : - 50.8 / / / 1 53.9 51.9 60.2 / / / - 58.2 1 / e / 9 / d / / « -1 £0.6 59.5 1 -1 s self -1 . 1 CH4I-b 4S.8 / / : / : / ! - 47.1 / / / 1 49.7 49.5 50.7 / / / -1 47.9 sulf - 48.1 1 / e / g / d / o 45.0 1 -1 s 1 Figure B . l : Effect of sulfate on Biogas C o m p o s i t i o n Appendix B. EFFECTS OF SULFATE IN BATCH EXPERIMENTS 229 Table B.4: Yates A n a l y s i s C H 4 % - a C H 4%-b C H 4 % - c p H Average 55.90 48.08 48.03 6.64 -1.61 -1.37 -1.55 0.17 Co -7.43 0.33 -0.92 -0.13 A -1.12 0.15 1.91 0.14 SCo -0.01 1.39 0.88 0.04 SA -0.04 0.17 -1.18 -0.03 CoA -0.29 -2.80 1.16 -0.05 GoAS 0.40 0.05 -0.67 0.06 Appendix B. EFFECTS OF SULFATE IN BATCH EXPERIMENTS 230 CH4% was almost the same for both cases of absence and presence of sulfate when higher C O D concentration was applied. This result demonstrated that the S R B competed more effectively at lower substrate level. In contrast to sulfate, the effects of Co became less significant after 3 weeks running. This is because the microbes in the reactor had been acclimatized and their concentration increased. The interaction effects of S and Co, Co and A were not negligible compared to the main effect of S although they didn't cause inhibition. The interaction plot of CH4% (Figure B-3) shows that the interactions between sulfate and feed strength (the ratio of C O D / N a S 0 4 ) , and feed strength and sludge concentration (sludge loading) are more significant than the interaction of sulfate and sludge concentration. It has been reported that the inhibition of sulfate on the activity of methanogens is related to the ratio of feed strength and sulfate concentration in terms of GOT)/Na2S04. When the ratio is smaller than 10, inhibition could occur. In this study, the ratio was even as low as 5, but no inhibition was noticed. We would suggest that the critical value of sulfate causing inhibition to an anaerobic process may vary with substrates. For a highly soluble feed, such as cheese whey, a higher sulfate concentration can be tolerated since excess hydrogen and fatty acid exist in this system. A n unusual observation was made that the effect of sludge on the gas composition was negative in the first week. That meant that the more sludge in the reactor, the less methane that was produced in the experiment. Generally, a higher concentration of sludge should result, in higher methane production since the sludge loading is lower. The only explanation for this observation was that the lower concentration of sludge was made in such a way that 35 ml sludge was mixed with 35 ml of effluent from an anaerobic digester, which contained bacteria and a higher p H (about 8.0), which might have increased the buffer capacity of the system. Therefore, in spite of a low concentration of sludge, it could have higher activity and buffer capacity than that of the original sludge. Appendix B. EFFECTS OF SULFATE IN BATCH EXPERIMENTS 231 CH4%-b 1 50.7 48.1 s 1 I I u j I d I I g i i e 1 I -1 47.9 45.0 s u l f . -1 1 Co= -1 CH4%-b 46.8 47.1 I I 4 9 . 7 4 9 . 5 s u l f -1 1 co= 1 Figure B.2: Effect of Sulfate on the Biogas C o m p o s i t i o n Appendix B. EFFECTS OF SULFATE IN BATCH EXPERIMENTS 232 OM-b 52 52 * - C - „ 4S * C+ +C f T > 111 I t ! _P 4 8 + 1 _ + f » H « » I I I 111 f f i : f . f ] 44 t 44 * -1 sulf sulf t«ol> t«ol> tvol> twol> twol> twol> intplot Cl C H 4 I-b 52 + 48 + » " in t u r n n u n t i i n t 44 • Co Figure B.3: Interaction Effects of Sulfate (S), Influent C O D (C), and Sludge (A): (a) S and C , (b) S and A , (c) C and A Appendix B. EFFECTS OF SULFATE IN BATCH EXPERIMENTS 233 Table B.5: A N O V A Table Source DF SS MS F F c r i t Block 2 6 5 6 . 4 9 3 2 8 . 2 5 4 2 . 0 7 2 . 8 0 Treat 15 1 4 9 . 5 4 9 . 9 7 1 . 2 8 1 .88 Exp.E 30 2 3 4 . 0 9 7 . 8 0 Total 47 1 0 4 0 . 1 2 Where DF=degreee of freedom, SS=sum of square, MS=raean square. The unusual phenomenon disappeared and the effect of sludge on the methane production could be ignored in subsequent weeks experiments because the sludge in both cases was acclimatized. It was noticed in this experiment that there was a time effect in batch culture. Con-sidering time effect as the block effect (the first week and the third week), we constructed the ANOVA table (Table B-5) to find how significant the block effect was. From the F values in Table B-5 we can see that the block effect was very significant. This indicated that the effects of feed strength and sludge concentration changed with time. None of the main effects was clearly statistically significant from this experiment ex-cept the feed strength in the first week based on the calculation of a 95 % confidence Appendix B. EFFECTS OF SULFATE IN BATCH EXPERIMENTS 234 interval. Therefore, the scales of the factors should be made larger to increase the dis-crimination, especially the sulfate concentration. The upper limit of sulfate concentration was based on the previous studies, but it still didn't cause any significant effects, which also revealed that the critical inhibition value of sulfate on cheese whey is higher than that of other substrates which have lower solubility. B.4.2 Effects of S, C o a n d A o n p H of the S y s t e m p H is one of the important parameters for an anaerobic system. The optimum p H for methanogenesis is between 7.0.to 8.0 units. Below 6.0, methanogenesis will be inhibited. Therefore, p H value could be an indicator of the system stability and its performance. The higher the p H , the more stable the system is. From the results of yates analysis we can see that the effect of sulfate on p H was very significant. By using feed containing sulfate, pHs averaged 0.2 units higher than using feed without sulfate. This indicated that, at these experimental conditions, the addition of sulfate helped to maintain the stability of the system since it can increase the buffer capacity of a cheese whey anaerobic digestion system. The effects of Co and A were also significant. p H decreased with the increase of feed strength and increased with the increase of the amount of the seed sludge. There was no unusual observation as in the case of methane production since the p H data were taken after 3 weeks operation. The effects on p H had an estimated standard deviation of 0.0096 with 8 degree of freedom. To determine the confidence interval from the N=16-run ts N 4 2.306 x 0.0096 2 + 0.011 ( B . l ) Appendix B. EFFECTS OF SULFATE IN BATCH EXPERIMENTS 235 Table B.6: 95% C o n f i d e n c e Interval s 0 . 1 7 + 0 . 0 1 o r 0 . 1 8 t o 0 . 1 6 Co - 0 . 1 3 + 0 . 0 1 o r - 0 . 1 2 t o - 0 . 1 4 A 0 . 1 4 + 0 . 0 1 o r 0 . 1 5 t o 0 . 1 3 SCO 0 . 0 4 + 0 . 0 1 d'r 0 . 0 5 t o 0 . 0 3 SA - 0 . 0 3 + 0 . 0 1 o r - 0 . 0 2 t o - 0 04 COA - 0 . 0 5 + 0 . 0 1 o r - 0 . 0 4 t o - 0 . 0 6 SCO A 0 . 0 6 + 0 . 0 1 o r 0 . 0 7 t o 0 . 0 5 The 95% confidence intervals are shown in Table B-6. Since none of these effects include zero in the confidence interval, it can be stated that all of these effects are significant at the 95% confidence level. However, the intplot of Fig.B-5 shows that the interactions are not important in. these operating conditions. The predominant effects are basically the first order terms. Figure B-6 is the plot of fitted value vs the actual p H value with a linear model. Note the fit is fairly good. So the interaction effects on p H can be negligible. This allows one to use the method of steepest ascent in the next experimental step. T h e different results between the two responses, % C i ? 4 and p H , show that p H is more sensitive to the process parameters than % C i ? 4 . As a general procedure, the residuals were plotted. None of those plots appeared anything unusual. Figure B-7 demonstrates that residuals basically center at zero. Appendix B. EFFECTS OF SULFATE IN BATCH EXPERIMENTS two!) twol) effects Effect Std Error t-Ratio Effects Effect ! . S Average £.63781 .00240 -27653E+4 i 1 salf (S) .17063 .00430 35.54180 < • * I Co ' t«ol> cube ' PH .495 .745 / / / : / : / : / : PH: ' / ! / ! fi i s subtracted 1 .438 .610 ! Read .438 as 6.438 I t I < C ! .778 !-- .818 1 o : / ': " / e i i . ! / g : / : / d J I o -1 .500 .720 1 -1 s suit -I 1 Figure B.4: Effect of Sulfate o n p H Appendix B. EFFECTS OF SULFATE IN BATCH EXPERIMENTS 237 pH 6.8 + 6.8 + " L ,,-C a i l . D M " I I • I I . i n i i II I ! |~1 " | | " , *C 6.6 + 1+" , , " 6.6 + C-" ,," I I . I I i i it . i i i i I I I I . II " I I - I-1 - C+' 6.4 + 6.4 + tvo!> t«ol> intplot Cl pH £.8 + - 1 * tl c r i 11 11 i n i i n 6.6 + l - , f f f "+! M i l l l i f t III I.J 6.4 + -1 1 -1 ;alf suit Co Figure B.5: Interaction Effects of Sulfate (S), Influent C O D (C), and Sludge (A): (a) S and A , (b) S and C , (c) C and A Appendix B. EFFECTS OF SULFATE IN BATCH EXPERIMENTS tvol> plot fitted pH 5.75 + fitted 6.50 + «3 / + 1 +- +-^ -6.4 6.5 6.6 6.7 6.8 pH Figure B.6: F i t ted p H versus A c t u a l V a l u e Appendix B. EFFECTS OF SULFATE IN BATCH EXPERIMENTS 239 .10 + .05 + r e s i d u a l t * t .00 + t .05 + * i i + 1 1 0 5 10 15 OBSERVATION Figure B .7: A n a l y s i s of Residuals Appendix B. EFFECTS OF SULFATE IN BATCH EXPERIMENTS 240 B.5 CONCLUSIONS 1. A n interesting finding is that the effects of sulfate on the gas composition was related to the feed strength. At lower feed strength, V0CH4 decreased when sulfate was used. However, no difference in gas composition between sulfate used and the control case was observed for higher feed strength. This illustrated that SRB competed with methane bacteria more effectively at lower substrate levels. 2. Contrary to reports in the literature, no inhibition was observed even when the ratio of C O D and sulfate was as low as 5 and the sulfate concentration was as high as 60 m.M/1. This can be attributed contributed to cheese whey characteristics of high solubility , and high feed strength (compared to 0.5 -1 g C O D / 1 in the previous studies). Therefore, the critical inhibition value of sulfate varies from one substrate to another. Using higher solubility substrates as feed, such as cheese whey, the inhibition value of sulfate is higher than that of other substrates which have lower solubility since excess hydrogen exists in such a system. 3. The effect of sulfate on p H is significant. p H is higher when sulfate was applied. This indicates that a proper concentration of sulfate can help to maintain stability since sulfate in a proper concentration increases the buffering capacity of an anaerobic process by competing for excess hydrogen and fatty acids with methane bacteria. 4. p H significantly decreased with the increase of feed strength and increased with the increase of sludge amount. B.6 CRITIQUE OF T H E E X P E R I M E N T A N D F U T U R E P L A N The upper limitation of sulfate was set based on a literature review. This sulfate concen-tration could cause inhibition in some anaerobic treatment of wastewaters, but it seemed too low to produce a significant effect on the gas composition for cheese whey. Much Appendix B. EFFECTS OF SULFATE IN BATCH EXPERIMENTS 241 Table B.7: E x p e r i m e n t a l Design B a s e d o n Steepest A s c e n t S g / i Co g COD/1 A g VSS C o d i n g i n c r e m e n t 0 .3 1.5 0.5 C o e f f i c i e n t s o f coded u n i t 0 ,085 -0.066 0.071 C o e f f i c i e n t s o f o r i g i n a l 0 .0255 -0.099 0.036 A d j u s t i n g c o e f f i c i e n t s 0 .25 -1 0.36 D e s i g n c e n t r e p o i n t 0 .3 4.5 1.13 P o s s i b l e s t e e p e s t p a t h 0 .55 3.5 1.5 0 .8 2.5 " 1.86 - 1 .05 1.5 2.22 1 .3 0.5 2.56 more sulfate should be used in order to obtain a possibly significant effect. So far, we don't know what the possible inhibition value of sulfate for a cheese whey anaerobic process is because we believe that the value varies with the substrates. Since a linear relationship was observed between p H and the operation variables (no interactions), it would be useful to use a steepest ascent strategy to search optimum operating conditions. Table B - 7 below shows values for the three operating variables along the direction of steepest ascent. It was originally proposed to test more response behaviors, such as C O D , volatile fatty acids and so on. However, these responses could not give meaningful information, perhaps due to the disadvantages of batch experimentation and the short period used for a biological process. A continuous experimental mode will be set up to gain more Appendix B. EFFECTS OF SULFATE IN BATCH EXPERIMENTS 242 accurate determination of these effects. One may choose 6 different levels of sulfate ranging between 0.5 to 10 g/1 for two levels of 5 and 20 g C O D / 1 cheese whey feed. For a long run, the initial seed amount is not a very important parameter. So, a 6 x 2 experimental design will be carried out. Hence the interaction effects may have some effects even though they didn't play significant roles in the effects on the CHA %. Future studies would be necessary to interpret the interaction effects Appendix C T H E C A L C U L A T I O N OF NON-IONIZED ACID C l T H E O R Y A weak acid or base is a weak electrolyte, which is only partially ionized in an aqueous solution, and as a result equilibrium is established between the ionized and unionized chemical forms. According to the Bronsted theory, an acid is a proton donor, whereas a base is a proton acceptor. A chemical equation representing the ionization of a weak monoprotic acid may written as HA—>H++A~ (Cl) the equilibrium constant expression for equation C - l has the form KA ~ ~\HAY • (C-2) Where K a represents the thermodynamic equilibrium constant for the acid reaction. By introducing the definition of p H and p K a , this equation becomes 243 Appendix C. THE CALCULATION OF NON-IONIZED ACID 244 PH =PKa + Log^-t (C .3) This equation is known as the Henderson-Hasselbalch equation and is useful in cal-culation of pH of a solution. C.2 C A L C U L A T I O N O F N O N - I O N I Z E D A C I D Let X represent the fraction of the initial acid concentration which is ionized, from equation (C-l) HA = H+ + A~ (C.4) C o - X X X where Co is termed the initial concentration of the weak acid, Co-X represents the non-ionized acid. Substituting x for A~ and Co-X fo HA, we have the results in Table Appendix D STATISTICAL ANALYSIS OF pH Comparison of p H in the absence and presence of sulfate was carried out by using a statistical program Minitab. Minitab is a general purpose statistical computing system which can be used to solve various statistical problems. Before carrying out any statistical test, some assumptions should be made, ie the error is independent, identically and normally distributed ( U N D (0.0)) and the samples are representive. To compare two sets of independent data , the standard deviations and other statistics are calculated. Independent data means the members of the two sets are not related, thus two groups of p H measurements are not linked to the same influent concentration, even though the substrate and operation temperature are the same. The results are given in Table D - l . T h e important difference between the two sets of p H data can be seen clearly from the dotplot. The p H in the absence of sulfate is much lower than that observed in the presence of sulfate. The p H in the absence of sulfate ranges from 6.75 to 7.18, giving a average of 7.037. The p H in the presence of sulfate has a range of 7.55 to 7.87, resulting in a mean value of 7.687. Next we need to know how much difference there is between these two means. That can be done by the t-test using the T W O S A M P L E command as shown in Table D-2. The 95% confidence interval means that if these 7 p H values were a random sample from a large population of p H sampling, then a 95% confidence interval for the average change 245 Appendix D. STATISTICAL ANALYSIS OF pH 246 due to the addition of sulfate in that population would be 0.485 to 0.815. Thus, we would be fairly certain that p H is increased, on the average, by at least 0.48 and maybe by as much as 0.82 p H units. As shows in Table D-2, the confidence interval is from 0.485 to 0.815, which does not include zero, thus it can be stated that the p H without sulfate is significantly different from the pH with sulfate. It should be mentioned that if the two populations have the same standard deviation, a subcommand of P O O L E D can follow T W O S A M P L E , allowing one to get smaller confi-dence interval and slightly more likelihood to reject a true null hypothesis. The P O O L E D means the standard deviation from the two samples are pooled to get an estimate of the common standard deviation. This difference essentially disappears with moderately large sample size. It was recommended that it's better not to use the P O O L E D subcommand. If the standard deviations are equal, you have lost little as shows in Table D-3. The use of P O O L E D did not make very much difference. If they are unequal, you may have gained a lot. If you use the pooled procedure when it is not appropriate, such as, when the standard deviation of the two population are not equal, you could be seriously misled. For example, you might falsely claim to have evidence that the two populations differ when they really do not. A t-test can be further made to test the statistic significance. T=8.68 > t(0.025, 12)= 2.179 here 0.025=a/2, a=0.05 and 1 2 = ^ -f n2-2=7+7-2 The critical T for 95% confidence and 12 degrees of freedom is 2.179 which is smaller than the calculated t=8.68. The conclusion is that the p H of the effluent was significantly increased by adding sulfate. Appendix D. STATISTICAL ANALYSIS OF PH 247 Table D . l : Calcula t ion of S t a n d a r d D e v i a t i o n s MTB > p r i n t c l c2 ROW pH pH-S 1 6.75 7.75 2 6.92 7.87 3 7.08 7.73 4 7.05 7.58 5 7.10 7.56 6 • 7.18 7.55 7 7.18 7.77 MTB > MTB > MTU > MTB > d o t p l o t c l c2; SUBO same. -+-pH 6.75 +-7.00 +-7.25 +-7.50 +-7.75 +•-pH-S 8.00 MTB > MTB > describe c l c2 pH pH-S N MEAN MEDIAN TRMEAN STDEV SEMEAN 7 7.0371 7.0800 7.0371 0.1543 0.0583 7 7.6871 7.7300 7.6871 0.1242 0.0469 PH pH-S MIN 6.7500 7.5500 MAX 7.1800 7.8700 Ql 6.9200 7.5600 Q3. 7.1800 7.7700 Appendix D. STATISTICAL ANALYSIS OF pH 248 Table D.2: t-test MTB > twosample c2 c l TWOSAMPLE T FOR pH-S VS pH N MEAN STDEV SE MEAN pH-S 7 7.687 0.124 0.047 pH 7 7.037 0.154 0.058 95 PCT CI FOR MU pH-S - MU pH: (0.485, 0.815) TTEST MU pH-S = MU pH (VS NE): T= 8.68 P=0.0000 DF= 11 MTB > Appendix D. STATISTICAL ANALYSIS OF pH 249 Table D.3: t-test MTB > MTB > tvosaraple c l c2; SUBO pooled. TWOSAMPLE T FOR pH VS pH-S N MEAN STDEV SE MEAN pH 7 7.037 0.154 0.058 pH-S 7 7.687 0.124 0.047 95 PCT CI FOR MU pH - MU pH-S: (-0.813, -0.487) TTEST MU pH = MU pH-S (VS NE) : T= -8.68 P = 0.0000 DF= 12 POOLED STDEV = 0.140 A p p e n d i x E D A T A T A B L E S 250 Appendix E. DATA TABLES 251 Table E . l : Sludge G r o w t h KINETICS SI IN. COD g/i .5...Q00 4.560 9.930 17.700 28.000 ANALYSIS LOADIG g/1 d 0.910 1.970 3.540 EFF.COD g/i 0.132 0.191 0.286 X SLUDGE g VSS £6.540 133.600 107.400 91.700 (dF/dt)/X RATIO OF F/S g/g d 5.960 0.457 114.200 :*8-.100 7.770 4.877 164.900 ,097 .263 .548 dX 3.060 7.370 7.050 dt g VSS day 17.000 30.000 10.000 (dX/dt)/X 0.0013 0.0023 0.0077 0.696 38.46.0 ' 12J)00 0.0281 0.656 67.140 16.000 0.0254 FOR (dX/dt)/X=Y*(dF/dt)/X-Kd Regression Output: Constant Kd=-0.015742 Std-Err of Y Est 0.007 R Squared 0.799 No. of Observations 4.000 Degrees of Freedom 2.000 X Coefficient(s) Y=0.0584564447 Std Err of Coef. 0.0208 Appendix E. DATA TABLES Table E.2: T h e P r o p i o n i c A c i d ( P A ) In the Absence of Sulfate Influent (g COD/1) P A at 1# (g/1) P A at 2# (g/1) P A r P A 9 (g/1) 4.56 0.092 0.060 0.032 9.93 0.134 0.014 0.120 17.7 0.290 0.330 -0.040 28.8 0.748 0.036 0.712 38.1 1.336 1.192 0.144 In the Presence of Sulfate Influent (g COD/1) P A at 1# (g/1) P A at 2# (g/1) PAi-PAo (g/1) 15 0.279 0.064 0.215 20 0.399 0.035 0.364 30 0.901 0.021 0.880 40 1.050 0.040 1.010 50 2.268 1.685 0.583 Appendix E. DATA TABLES 253 Table E.3: T h e T o t a l Vola t i le F a t t y A c i d ( T V F A ) In the Absence of Sulfate Influent (g COD/1) T V F A at 1# (g/1) T V F A at 2# (g/1) T V F A r T V F A 9 (g/D 4.56 0.958 0.274 0.684 9.93 1.962 0.124 1.838 17.7 2.624 0.023 2.601 28.8 2.174 0.076 2.09S 38.1 3.585 2.148 1.437 In the Presence of Sulfate Influent (g COD/1) A A at 1# (g/1) A A at 2# (g/1) A A 1 - A A 9 (g/D "* 15 1.335 0167 1.168 20" 1.562 0.078 . 1.484 30 2.314 0.094 2.220 40 2.650 0.097 2.553 50 4.504 3.999 0.505 Appendix E. DATA TABLES 254 Table E.4: Analys is for A c c u m u l a t i o n and Degradat ion of V F A ROW AAs Co Co**2 Co** 3 Co**0.5 1 1 .103 15 225 3375 3.87298 2 1 .136 20 400 8000 4.47214 3 1 .413 30 900 27000 5.47723 4 1 .600 40 1600 64000 6.32456 5 2 .127 50 2500 125000 7.07107 M T U > Co**0.5 i s highly c o r r e l a t e d with other X v a r i a b l e s Co**0.5 has been removed from the equation * NOTE * * NOTE * * NOTE * Co i s hig h l y c o r r e l a t e d with other Co**2 i s hig h l y c o r r e l a t e d with other Co**3 i s hig h l y correlated with other predictor v a r i a b l e s predictor v a r i a b l e s predictor v a r i a b l e s The regression equation i s AAs = 0.504 + 0.0590 Co - 0.00172 Co**2 +0.000024 Co**3 P r e d i c t o r Constant Co Co**2 Co**3 Coef 0.5036 0.05899 -0.001724 0.00002382 Stdev 0.9074 0.09825 0.003221 0.00003267 t - r a t i o 0.55 0.60 -0.54 0.73 P 0.677 0.656 0.687 0.599 s = 0.07773 R-sq = 99.1% A n a l y s i s of Variance R-sq(adj) = 96.5% SOURCE Regress ion E r r o r CONTINUE? DF .3 1 SS 0.69183 0.00604 MS 0.23061 0.00604 F 38.16 P 0.118 CONTINUE? T o t a l 0.69787 SOURCE Co Co**2 Co**3 DF 1 1 1 SEQ SS 0.65956 0.02906 0.00321 Obs. 1 2 3 4 5 Co 15.0 20.0 30.0 40.0 50.0 AAs 1.1030 1.1360 1.4130 1.6000 2.1270 F i t Stdev.Fit Residual 1.0809 1.1843 1.3647 1.6290 2.1201 0.0745 0.0609 0.0609 0.0721 0.0774 0.0221 -0.0483 0.0483 -0.0290 0.0069 St.Resid 1.00 -1.00 1.00 -1.00 1.00 X X denotes an obs. whose X value gives i t large influence. Appendix E. DATA TABLES 255 Table E.5: A n a l y s i s for A c c u m u l a t i o n a n d D e g r a d a t i o n of V F A ROW A l A2 1/Rs C C**2 C** ( - l ) C**0.5 1 1. 038 0 .103 1. 06952 15 225 0 .0666667 3.87298 2 1. 163 0 .043 0. 89286 20 400 0 .0500000 4.47214 3 1. 403 0 .073 0. 75188 30 900 0 .0333333 5.47723 4 1. 600 0 .057 0. 64809 40 1600 0 .0250000 6.32456 5 2. 127 1 .714 2. 42131 50 2500 0 .0200000 7.07107 M T U > * C**0.5 i s hig h l y c o r r e l a t e d with other X v a r i a b l e s * C**0.5 has been removed from the equation * NOTE * C i s highly c o r r e l a t e d with other p r e d i c t o r v a r i a b l e s * NOTE * C**2 i s highly c o r r e l a t e d with other p r e d i c t o r v a r i a b l e s * NOTE * C**(-l) i s highly c o r r e l a t e d with other p r e d i c t o r v a r i a b l e s The regression equation i s 1/Rs = 15.6 - 0.630 C + 0.00815 C**2 - 104 C * * ( - l ) P r e d i c t o r Constant C C**2 C * * ( - l ) Coef 15.569 -0.6297 0.008148 -104.01 Stdev 9 .801 0.3424 0.003674 84.95 t - r a t i o 1.59 -1.84 2.22 -1.22 P 0.358 0.317 0.270 0. 436 s = 0.3100 R-sq = 95.4% An a l y s i s of Variance SOURCE Regression E r r o r CONTINUE? DF 3 1 SS 2.00290 0.09611 R-sq(adj) = 81.7% MS 0.66763 0.09611 F 6.95 P 0.270 T o t a l 2.09901 SOURCE C C**2 C * * ( - l ) DF 1 1 1 SEQ SS 0.71137 1.14746 0.14408 Obs. C 1/Rs F i t S t d e v . F i t Residual St.Res i d 1 15.0 1.070 1.022 0.306 0.048 1.00 2 20.0 0.893 1.032 0.277 -0.140 -1.00 3 30.0 0.752 0.543 0.229 0.209 1. 00 4 40.0 0.648 0.816 0.261 -0.167 -1.00 5 50.0 2.421 2.371 0.306 0.050 1.00 MTB > M T U > Appendix E. DATA TABLES 256 Table E.6: A n a l y s i s for A c c u m u l a t i o n a n d D e g r a d a t i o n of V F A ROW Co PA1 PA2 1/r Co**2 Co**3 Co**(-l) 1 4 .56 0 .092 0 .060 31 .2500 20.79 94.8 -1.00000 2 9 .93 0 .134 0 .014 8 .3333 98.60 979.1 1.00000 3 17 .70 0 .290 0 .003 3 .4843 313.29 5545.2 -1.00000 4 28 .80 0 .748 0 .036 1 . 4045 829.44 23887.9 1.00002 5 38 .10 1 . 336 1 .192 6 .9444 1451.61 55306.3 -1.00014 MTB > MTII > MTB > MTB > regr c2 4 c l c5-c8 * NOTE * Co i s highly c o r r e l a t e d with other p r e d i c t o r v a r i a b l e s * NOTE * Co**2 i s highly c o r r e l a t e d with other p r e d i c t o r v a r i a b l e s * NOTE * Co**3 i s highly c o r r e l a t e d with other p r e d i c t o r v a r i a b l e s The r e g r e s s i o n equation i s PA1 = 0.108 - 0.00749 Co +0.000984 Co**2 +0.000002 Co**3 + 0.00213 Co**( P r e d i c t o r Coef Constant 0.107665 Co -0.00748717 Co**2 0.00098372 Co**3 0.00000159 Co* * ( - l ) 0.00212939 Stdev 0.000000 0.00000000 0.00000000 0.00000000 0.00000000 t - r a t i o * P * * * s = * A n a l y s i s of Variance SOURCE Regress ion E r r o r T o t a l DF 4 0 4 SS 1.102920 1.102920 MS 0.275730 * F P CONTINUE? SOURCE Co Co**2 Co**3 Co * * ( - l ) DF 1 1 1 1 SEQ SS 1.030516 0.072380 0.000007 0.000017 Obs. Co PA1 F i t S t d e v . F i t Residual St.Resid 1 4.6 0.09200 0.09200 * 0.00000 * 2 9.9 0.13400 0.13400 * 0.00000 * 3 17.7 0.29000 0.29000 * 0.00000 * 4 28.8 0.74800 0.74800 * 0.00000 * 5 38.1 1.33600 1.33600 * 0.00000 * MTB > MTB > Appendix E. DATA TABLES 257 Table E.7: A n a l y s i s for A c c u m u l a t i o n a n d D e g r a d a t i o n of V F A MTB > s t e p w i s e y i n c2 x i n c l c 5 c 6 c7 * ERROR * ARGUMENT IS A CONSTANT OR MATRIX, BUT A COLUMN WAS EXPECTED MTB > s t e p w i s e c2 4 c l c5 c6 c7 c8 * ERROR * I L L E G A L ARGUMENT MTB > s t e p w i s e c2 4 c l c 5 - c 7 STEPWISE REGRESSION OF PA1 ON 4 P R E D I C T O R S , WITH N = 5 STEP 1 2 CONSTANT 0 .04401 0 .11244 C o * * 2 0 .00088 0 .00108 T - R A T 1 0 35 .41 76 .93 Co - 0 . 0 0 8 9 3 T - R A T I O - 1 4 . 6 0 S 0 .0296 0 .00350 R-SQ 99 .76 100 .00 MORE? ( Y E S , NO, SUBCOMMAND, OR HELP) S U B O S U B O Appendix E. DATA TABLES 258 Table E.8: A n a l y s i s for A c c u m u l a t i o n a n d D e g r a d a t i o n of V F A MTB > stepwise c2 4 c l c5-c7 STEPWISE REGRESSION OF PAl STEP 1 2 CONSTANT 0.04401 0.11244 Co**2 0.00088 0.00108 T-RATIO 35.41 76.93 ON 4 PREDICTORS, WITH N = Co T-RATIO -0.00893 -14.60 S 0.0296 0.00350 R-SQ 99.76 100.00 MORE? (YES, NO, SUBCOMMAND, OR HELP) SUBO :;mt MTB > regr c2 2 c l c5 The regression equation i s PAl = 0.112 - 0.00893 Co + 0.00108 Co**2 Pr e d i c t o r Coef Stdev t - r a t i o p Constant 0.112443 0.005189 21.67 0.002 Co -0.0089276 0.0006114 -14.60 0.005 Co**2 0.00107690 0.00001400 76.93 0.000 s = 0.003497 R-sq Ana l y s i s of Variance 100.0% SOURCE Regression Err o r T o t a l DF 2 2 4 SS 1.10290 0.00002 1.10292 R-sq(adj) = 100.0% MS F 0.55145 45084.95 0.00001 P 0.000 SOURCE Co Co**2 DF 1 1 SEQ SS 1.03052 0.07238 CONTINUE? Obs . 1 2 3 4 5 Co 4.6 9.9 17.7 28.8 38.1 PAl ,09200 ,13400 .29000 .74800 F i t S t d e v . F i t Residual 1.33600 0.09413 0.12998 0.29181 0.74855 1.33554 0.00309 0.00203 0.00251 0.00237 0.00333 -0.00213 0.00402 -0.00181 -0.00055 0.00046 St.Resid -1. 30 1.41 -0.74 -0.21 0.44 Appendix E. DATA TABLES 259 Table E.9: A n a l y s i s for A c c u m u l a t i o n a n d D e g r a d a t i o n of V F A ROW Co PAl PA2 1/r Co**2 Co**3 Co**(-l) 1 4 .56 0 .092 0 .060 31 .2500 20.79 94.8 0.219298 2 9 .93 0 .134 0 .014 8 .3333 98.60 979.1 0.100705 3 17 .70 0 .290 0 .003 3 .4843 313.29 5545.2 0.056497 4 28 .80 0 .748 0 .036 1 .4045 829.44 23887.9 0.034722 5 38 .10 1 .336 1 .192 6 .9444 1451.61 55306 . 3 0.026247 MTB > stepwise c4 4 c l c5-c7 STEPWISE REGRESSION OF 1/r ON 4 PREDICTORS, WITH N STEP 1 2 CONSTANT -2.369 -9 .280 C o * * ( - l ) 145 184 T-RATIO 5.30 17.67 Co**3 0.00020 T-RATIO 5.80 S 4.32 1.25 R-SQ 90.35 99.46 MORE? (YES, NO, SUBCOMMAND, OR HELP) SUBO MTB > regr c4 4 c l c5-c7 * NOTE * Co i s highly correlated with other p r e d i c t o r v a r i a b l e s * Co**2 i s highly correlated with other X v a r i a b l e s Co**2 has been removed from the equation The r e g r e s s i o n equation i s 1/r = - 11.4 + 0.112 Co +0.000160 Co**3 + 192 Co**(-l) P r e d i c t o r Constant Co Co**3 C o * * ( - l ) Coef -11.438 0.1119 0.0001597 191.59 Stdev 8.721 0.4396 0.0001797 34.37 s = 1.717 R-sq = 99.5% A n a l y s i s of Variance SOURCE Regress ion E r r o r T o t a l CONTINUE? DF 3 1 4 SS 576.66 2.95 579.61 t - r a t i o -1.31 0.25 0.89 5.57 P 0.415 0.841 0.537 0.113 R-sq(adj) = 98.0% MS 192.22 2.95 F 65.19 P 0.091 Appendix E. DATA TABLES Table E.10: A n a l y s i s for A c c u m u l a t i o n a n d D e g r a d a t i o n of V F A SOURCE DF SEQ SS Co 1 243. 14 Co**3 1 241. 92 Co**(-l) 1 91. 61 Obs. Co 1/r F i t Stdev.Fit Res idua l St.Resid 1 4.6 31 .250 31 .101 1.711 0.149 1.00 X 2 9.9 8 . 333 9 .123 1.525 -0.789 -1.00 3 17.7 3 .484 2 .252 1.196 1.232 1.00 4 28.8 1 .404 2 .252 1. 493 -0.848 -1.00 5 38.1 6 .944 6 .688 1.698 0.256 1.00 X denotes an obs. whose X value gives i t large i n f l u e n c e . M T B > M T B > Appendix E. DATA TABLES Table E . l l : Analysis for Accumulation and Degradation of V F A MTB > MTB > regr c4 2 c l c7 The r e g r e s s i o n equation i s l / r = - 18.3 + 0.488 Co + 216 Co**(-l) P r e d i c t o r Coef Stdev t - r a t i o P Constant -18 .311 3. 819 -4.79 0.041 Co 0. 4884 0.1115 4.38 0.048 Co**(-1) 216.: L8 19 .30 11.20 0.008 s = 1.625 R-sq = 99 .1% R-sq(adj) = 98.2% Anal y s i s of Var iance SOURCE DF SS MS F P Regress ion 2 574 . 33 287.17 108.80 0 . 009 Error 2 5. 28 2.64 T o t a l 4 579 . 61 SOURCE DF SEQ SS Co 1 243. 14 Co**(-l) 1 331. 20 C O N T 1 NI1K? Obs. Co l / r F i t S tdev.Fit Residual St.Resid 1 4.6 31 .250 31. 324 1.601 -0.074 -0. 27 2 9.9 8 .333 8. 310 1.154 0.024 0 .02 3 17.7 3 .484 2. 548 1.087 0.936 0.78 4 28.8 1 .404 3. 262 0.917 -1.858 -1.39 5 38.1 6 .944 5. 972 1. 414 0.972 1.22 M T U > Appendix E. DATA TABLES Table E.12: A n a l y s i s for A c c u m u l a t i o n a n d D e g r a d a t i o n of The r e g r e s s i o n equation i s 1/r = - 2.37 +- 145 Co**(-l) P r e d i c t o r Constant C o * * ( - l ) Coef -2.369 144.61 Stdev 3.070 27.28 t - r a t i o -0.77 5.30 P 0.497 0.013 s = 4.317 R-sq = 90.4% A n a l y s i s of Variance SOURCE Regress ion E r r o r T o t a l DF 1 3 4 SS 523.70 55.91 579.61 R-sq(adj) = 87.1% MS 523.70 18.64 F 28.10 P 0.013 Obs. Co * * ( - l ) 1/r F i t Stdev.Fit Residual St.Resid 1 0.219 31.25 29.34 4.08 1.91 1.35 2 0.101 8.33 12.19 1.96 -3.86 -1.00 3 0.056 3. 48 5.80 2.11 -2.32 -0.61 4 0.035 1.40 2.65 2.41 -1.25 -0.35 5 0.026 6.94 1.43 2.55 5. 52 1.59 MTB > Appendix E. DATA TABLES 2 6 3 Table E.13: A n a l y s i s for A c c u m u l a t i o n a n d D e g r a d a t i o n of V F A The r e g r e s s i o n equation i s PAsl = 0.275 - 0.0160 Cso +• 0.00105 Cso**2 4- 0.117 1/Cos P r e d i c t o r Constant Cso Cso**2 1/Cos Coef 0.2746 -0.01595 0.0010453 0.1168 SOURCE Regress ion E r r o r T o t a l DF 3 1 4 Stdev 0.8816 0.06160 .0009478 0.1202 s = 0.2411 R-sq = 97.7% A n a l y s i s of Variance SS 2.44093 0.05811 2.49904 t - r a t i o 0.31 -0.26 1.10 0.97 P 0.808 0.839 0.469 0.509 R-sq(adj) = 90.7% MS 0.81364 0.05811 F 14.. 00 P 0.193 SOURCE Cso Cso**2 1/Cos DF 1 1 1 SEQ SS 2.23270 0.15334 0.05489 CONTINUE? CONTINUE? Obs. Cso PAsl F i t S t d e v . F i t Residual St.Resid 1 15.0 0.279 0.387 0.215 -0.108 -1.00 2 20.0 0.399 0.257 0.195 0.142 1.00 3 30.0 0.901 0.854 0.236 0.047 1.00 4 40.0 1.050 1.192 0.195 -0.142 -1.00 5 50.0 2.268 2.207 0.233 0.061 1.00 f Appendix E. DATA TABLES 264 Table E.14: A n a l y s i s for A c c u m u l a t i o n a n d D e g r a d a t i o n of V F A MTB MTB > regr c22 2 c21 c25 The r e g r e s s i o n equation i s PAsl = 0.603 - 0.0391 Cso + 0.00141 Cso**2 P r e d i c t o r Coef Stdev Constant 0.6026 0.8030 Cso -0.03910 0.05602 Cso**2 0.0014122 0.0008572 s = 0.2377 R-sq = 95.5% A n a l y s i s of Variance SOURCE Regress ion E r r o r T o t a l DF 2 2 4 SS 2.3860 0.1130 2.4990 t - r a t i o 0.75 -0 .70 1.65 P 0.531 0.557 0.241 R-sq(adj) = 91.0% MS 1.1930 0.0565 F 21.11 P 0.045 SOURCE Cso Cso**2 DF 1 1 SEQ SS 2.2327 0.1533 Obs. Cso PAsl F i t S tdev.Fit Residual St.Resid 1 15.0 0.279 0.334 0.205 -0.055 -0.46 2 20.0 0.399 0.386 0.141 0.013 0.07 3 30.0 0.901 0.701 0.174 0.200 1.24 4 40.0 1.050 1.298 0.159 -0.248 -1. 40 5 50.0 2.268 2.178 0.228 0.090 1.34 MTB > MTB > MTU > Appendix E. DATA TABLES Table E.15: A n a l y s i s for A c c u m u l a t i o n a n d D e g r a d a t i o n of V F A MTB > MTB > regr c l 2 2 e l l c l 5 The r e g r e s s i o n equation i s C12 = 1.77 - 0.0552 Cso + 0.00215 Cso**2 P r e d i c t o r Constant Cso Cso**2 Coef 1.768 -0.05523 0.002148 Stdev 1.115 0.07776 0.001190 t - r a t i o 1.59 -0.71 1.81 P 0.254 0. 551 0.213 s = 0.3299 R-sq = 96.5% A n a l y s i s of Variance R-sq(adj) = 93.1% SOURCE Regression E r r o r T o t a l DF 2 2 4 SS 6.0888 0.2177 6.3065 MS 3.0444 0.1089 F 27.97 P 0.035 SOURCE Cso Cso**2 DF 1 1 SEQ SS 5.7340 0.3549 CONTINUE? Obs. 1 2 3 4 5 Cso 15.0 20.0 30.0 40.0 50.0 C12 .335 .562 .314 .650 .504 F i t S t d e v . F i t Residual 1.423 1.523 2.045 2.996 4.378 0.285 0.195 0.241 0.221 0.317 -0.088 0.039 0.269 -0.346 0.126 St.Resid -0.53 0.15 1.20 -1.41 1.36 MTB > MTB > Appendix E. DATA TABLES 266 Table E.16: A n a l y s i s for A c c u m u l a t i o n a n d D e g r a d a t i o n of V F A MTB > p r i n t c21-c28 )W Cso PAsl PAs2 1/rs Cso**2 Cso**3 P s l -Ps2 1 15 0.279 0.064 4 .65116 225 3375 0 . 215 2 20 0.399 0.035 2 .74725 400 8000 0 .364 3 30 0.901 0.021 1 .13636 900 27000 0 .880 4 40 1.050 0.040 0 .99010 1600 64000 1 .010 5 50 2.268 1.685 1 .71527 2500 125000 0 .583 MTB > stepwise c24 4 c21 c25 c26 c28 STEPWISE REGRESSION OF 1/rs ON 4 PREDICTORS, WITH N STEP CONSTANT 1 -0 .5539 2 -3.4673 1/Cos T-RATIO 71.84 3.89 120.65 129.91 Cso**3 T-RATIO 0.00002 62.83 S 0.709 0.0196 R-SQ 83.46 99.99 MORE? (YES, NO, SUBCOMMAND, OR HELP) Appendix E. DATA TABLES 267 Table E.17: Analys is for A c c u m u l a t i o n a n d D e g r a d a t i o n of V F A MTB > regr c24 2 c21 c25 The r e g r e s s i o n equation i s 1/rs = 11.0 - 0.540 Cso + 0.00715 Cso**2 Pred i c t o r Constant Cso Cso**2 Coef 10.9582 -0.54036 0.007148 Stdev 0.9559 0.06669 0.001020 t - r a t i o 11.46 -8.10 7.00 P 0.008 0.015 0.020 s = 0.2830 R-sq = 98.2% A n a l y s i s of Variance SOURCE Regress ion E r r o r T o t a l DF 2 2 4 SS 8.9662 0.1601 9.1263 R-sq(adj) = 96.5% MS 4.4831 0.0801 F 56.00 P 0.018 SOURCE Cso Cso**2 DF 1 1 SEQ SS 5.0380 3.9282 CONTINUE? Obs . 1 2 3 4 5 Cso 15.0 20 30 40 50 1/rs 4.651 2.747 1.136 0.990 1.715 F i t S t d e v . F i t Residual 4.461 3.010 1.180 0.780 1.809 0.244 0.167 0.207 0.189 0.272 0.190 -0.263 -0.044 0.210 -0.094 St.Resid 1.33 -1.15 -0.23 1.00 -1.18 MTI Appendix E. DATA TABLES 268 Table E.18: A n a l y s i s for A c c u m u l a t i o n a n d D e g r a d a t i o n of V F A MTB > p r i n t c l - c 8 ROW Co VFA1 VFA 2 1/rr co**2 1/Co 1 4 .56 0.958 0.274 1 .46199 20.79 0. 219298 2 9 .93 1.962 0.124 0 .54407 98.60 0. 100705 3 17 .70 2.624 0.023 0 .38447 313.29 0. 056497 4 28 .80 2.174 0.076 0 .47664 829.44 0. 034722 5 38 .10 3.585 2.148 0 .69589 1451.61 0. 026247 MTB > MTB > M T U > * Co**0.5 i s h i g h l y c o r r e l a t e d with other X v a r i a b l e s * Co**0.5 has been removed from the equation * NOTE * Co i s highly c o r r e l a t e d with other p r e d i c t o r v a r i a b l e s The r e g r e s s i o n equation i s VFA1 = 5.91 - 0.236 Co + 0.00484 co**2 - 18.3 1/Co P r e d i c t o r Coef Stdev t - r a t i o p Constant 5 .908 4. 018 1.47 0.380Co -0. 2357 0.2704 -0.87 0.544 co**2 0 .004839 0.004825 1.00 0.499 1/Co -18.32 14 .32 -1.28 0. 422 s = 0.5842 R-sq = 90.7% R-sq(adj) = 62.9% A n a l y s i s of Variance SOURCE DF SS MS F P Regression 3 3. 3382 1.1127 3.26 0 . 382 E r r o r 1 0. 3413 0.3413 T o t a l 4 3. 6795 CONTINUE? SOURCE DF SEQ SS Co 1 2. 7596 co**2 1 0. 0197 1/Co 1 0. 5589 Obs. Co VFA1 F i t S t d e v . F i t Residual St. Resid 1 4 . 6 0.958 0. 916 0.583 0.042 1. 00 2 9 . 9 1.962 2. 200 0.534 -0.238 •1. 00 3 17. 7 2.624 2. 217 0.420 0. 407 1. 00 4 28. 8 2.174 2. 498 0 .486 -0.324 •1. 00 5 38. 1 3.585 3. 472 0.573 0.113 1. 00 X d f t n o t f t s an o b s . u h o s f t X v a l u e q i v « s i t l a r g f ; i n f 1 iiftncft. Appendix E. DATA TABLES 269 Table E.19: A n a l y s i s for A c c u m u l a t i o n a n d D e g r a d a t i o n of V F A MTB > regr c2 2 c l c6 The r e g r e s s i o n equation i s VFA1 = 2.12 + 0.0325 Co - 5.76 1/Co P r e d i c t o r Coef Stdev t - r a t i o P Constant 2 .121 1.375 1.54 0.263 Co 0.03246 0.04016 0.81 0.504 1/Co -5 .762 6.950 -0.83 0.494 s = 0.5851 R-sq = 81.4% R -sq(adj) = = 62.8% A n a l y s i s of Variance SOURCE DF SS MS F P Regression 2 2. 9949 1 . 4974 4.37 0.186 E r r o r 2 0. 6846 0 .3423 T o t a l 4 3. 6795 SOURCE DF SEQ SS Co 1 2 . 7596 1/Co 1 0. 2353 CONTINUE? Obs. Co VFA1 F i t S t d e v , F i t Residual St.Res i d 1 4.6 0.958 1.006 0.577 -0.048 . -0. 48 2 9.9 1.962 1.863 0.416 0.099 0. 24 3 17.7 2.624 2.370 0.391 0.254 0. 58 4 28.8 2.174 2.856 0.330 -0.682 -1. 41 5 38.1 3.585 3.207 0.509 0.378 1. 31 Appendix E. DATA TABLES 270 Table E.20: A n a l y s i s for A c c u m u l a t i o n a n d D e g r a d a t i o n of V F A * Co**0.5 i s h i g h l y c o r r e l a t e d with other X v a r i a b l e s Co**0.5 has been removed from the equation * NOTE * Co i s highly c o r r e l a t e d with other p r e d i c t o r v a r i a b l e s The r e g r e s s i o n equation i s 1/rr = - 0.579 + 0.0239 Co +0.000088 co**2 + 8.79 1/Co P r e d i c t o r Constant Co co**2 1/Co Coef -0.5789 0.02394 0.0000880 8.7929 s = 0.02197 R-sq A n a l y s i s of Variance Stdev 0.1511 0.01017 0.0001814 0.5385 99.9% t - r a t i o -3.83 2.35 0.48 16. 33 P 0.163 0.256 0.713 0.039 R-sq(adj) = 99.7% SOURCE Regression E r r o r T o t a l DF 3 1 4 SS 0.75313 0.00048 0.75361 MS 0.25104 0.00048 F 520.17 P 032 CONTINUE? SOURCE CO . CO**2 1/Co DF 1 1 1 SEQ SS 0.17626 0.44817 0.12869 Obs. Co 1/rr F i t S t d e v . F i t Res i d u a l St.Resid 1 4.6 1 .46199 1. 46040 0.02191 0.00159 1.00 X 2 9.9 0 .54407 0. 55301 0.02007 -0.00894 -1.00 3 17.7 0 .38447 0. 36918 0.01577 0.01529 1.00 4 28.8 0 .47664 0. 48883 0.01828 -0.01218 -1.00 5 38.1 0 .69589 0. 69165 0.02155 0.00424 1.00 Appendix E. DATA TABLES Table E.21: Analysis for Accumulation and Degradation of V F A MTB > MTB > regr c4 2 c l c6 The r e g r e s s i o n equation i s 1/rr = - 0.648 + 0.0288 Co + 9.02 1/Co P r e d i c t o r Constant Co 1/Co Coef -0.64769 0.028812 9.0212 Stdev 0.04058 0.001185 0.2051 t - r a t i o -15.96 24.31 43.99 P 0.004 0.002 0.001 s = 0.01726 R-sq An a l y s i s of Variance 99.9% R-sq(adj) = 99.8% SOURCE Regress ion E r r o r T o t a l DF 2 2 4 SS 0.75301 0.00060 0.75361 MS 0.37651 0.00030 F 1263.40 P 0.001 SOURCE Co 1/Co DF 1 1 SEQ SS 0.17626 0.57675 CONTINUE? Obs. 1 2 3 4 5 Co 4.6 9.9 17.7 28.8 38.1 1/rr .46199 .54407 .38447 .47664 .69589 F i t S t d e v . F i t Residual 1.46204 0.54690 0.37196 0.49534 0.68683 0.01701 0.01227 0.01155 0.00974 0.01503 -0.00005 -0.00283 0.01251 -0.01869 0.00906 St.Resid -0.02 -0. 0. -1. 1. .23 .97 .31 .07 MTII Appendix E. DATA TABLES 272 Table E.22: A n a l y s i s for A c c u m u l a t i o n a n d D e g r a d a t i o n of V F A ROW Cso C12 C13 1/Rs Cso**2 t f 1 - t f 2 1/Cos 1 15 1 . 335 0 . 167 0. 85616 225 1 .168 0 .0666667 2 20 1 .562 0 .078 0. 67385 400 1 .484 0 .0500000 3 30 2 .314 0 .094 0. 45045 900 2 .220 0 .0333333 4 40 2 .650 0 . 097 0. 39170 1600 2 .553 0 .0250000 5 50 4 .504 3 .999 1. 98020 2500 0 .505 0 .0200000 MTB > 1 > SUBO n MTB > regr c l 2 3 e l l c l 5 c l 7 * NOTE * Cso i s highly c o r r e l a t e d with other * NOTE * Cso**2 i s highly c o r r e l a t e d with other * NOTE * 1/Cos i s highly c o r r e l a t e d with other p r e d i c t o r v a r i a b l e s p r e d i c t o r v a r i a b l e s p r e d i c t o r v a r i a b l e s The regression equation i s C12 = 11.2 - 0.378 Cso + 0.00553 Cso**2 - 82.0 1/Cos P r e d i c t o r Constant Cso Cso**2 1/Cos Coef 11.17 -0.3781 0.005526 -82.00 Stdev 11.32 0.3953 0.004243 98.10 t - r a t i o 0.99 -0.96 1.30 -0.84 P 0. 504 0 . 514 0.417 0 . 557 s = 0.3580 R-sq = 98.0% A n a l y s i s of Variance R-sq(adj) = 91.9% SOURCE Regression E r r o r T o t a l DP 3 1 4 SS 6.1784 0.1282 6.3065 MS 2.0595 0.1282 F 16.07 P 0.181 CONTINUE? SOURCE Cso Cso**2 1/Cos DF 1 1 1 . SEQ SS 5.7340 0.3549 0.0895 Obs. Cso C12 F i t Stdev. F i t R e s idual St.Res id 1 15.0 1.335 1.280 0. 354 0.055 1.00 2 20.0 1.562 1.723 0. 320 -0.161 -1.00 3 30.0 2.314 2.072 0. 264 0.242 1.00 4 40.0 2.650 2.843 0 . 301 -0.193 -1.00 5 50.0 4.504 4.446 0. 353 0.058 • 1.00 Appendix E. DATA TABLES 273 Table E.23: Analysis for A c c u m u l a t i o n a n d D e g r a d a t i o n of V F A MTB > MTB > regr c l 4 3 e l l c l 5 cl7 * NOTE * Cso i s highly c o r r e l a t e d with other * NOTE * Cso**2 i s highly c o r r e l a t e d with other * NOTE * 1/Cos i s highly c o r r e l a t e d with other p r e d i c t o r v a r i a b l e s p r e d i c t o r v a r i a b l e s p r e d i c t o r v a r i a b l e s The r e g r e s s i o n equation i s 1/Rs = 14.0 - 0.577 Cso + 0.00748 Cso**2 - 92.4 1/Cos P r e d i c t o r Constant Cso Cso**2 1/Cos Coef 13.958 -0.5773 0.007480 -92.42 Stdev • 7.578 0.2647 0.002841 65.69 t - r a t io 1.84 -2.18 2.63 -1.41 P 0.317 0.274 0.231 0 . 393 s = 0.2397 R-sq = 96.6% A n a l y s i s of Variance R-sq(adj) = 86.3% SOURCE Regression E r r o r T o t a l DF 3 1 4 SS 1.61853 0.05747 1.67600 MS 0.53951 0.05747 F 9.39 P 0.234 CONTINUE? SOURCE Cso Cso**2 1/Cos DF 1 1 1 SEQ SS 0.46789 1.03689 0.11375 Obs. 1 2 3 4 5 Cso 15.0 20.0 30.0 40.0 50.0 1/Rs 0.856 0.674 0.450 0.392 1.980 F i t S t d e v . F i t Residual 0.819 0.782 0.289 0.521 1.942 0.237 0.214 0.177 0.202 0.237 0.037 -0.108 0.162 -0.129 0.039 St.Resid 1.00 -1.00 1.00 -1.00 1.00 MTB > MTU > Appendix E. DATA TABLES 274 Table E.24: A n a l y s i s for A c c u m u l a t i o n a n d D e g r a d a t i o n of V F A MTB > regr c l 4 2 e l l c l 5 The r e g r e s s i o n equation i s 1/Rs = 3.3G - 0.213 Cso + 0.00367 Cso**2 Pred i c t o r Constant Cso Cso**2 Coef 3.3566 -0.21347 0.003672 Stdev 0.9884 0.06896 0.001055 t - r a t io 3.40 -3.10 3.48 P 0. 077 0.090 0.074 s = 0.2926 R-sq = 89.8% A n a l y s i s of Variance SOURCE Regression E r r o r T o t a l DF 2 2 4 SS 1.50479 0.17121 1.67600 R-sq(adj) = 79.6% MS 0.75239 0.08561 F 8.79 P 0.102 SOURCE Cso Cso**2 DF 1 1 SEQ SS 0.46789 1.03689 CONTINUE? Obs. 1 2 3 4 5 Cso 15.0 20 30 40 50 1/Rs 0.856 0.674 0.450 0.392 1.980 F i t Stdev.Fit Residual 0.981 0.556 0.258 0.694 1.864 0.253 0.173 0.214 0.196 0.281 -0.125 0.118 0.193 -0.302 0.116 St.Resid -0.8 5 0.50 0.97 -1.39 1.41 MTB > A p p e n d i x 1 PUBLICATIONS 1990 Yan J.Q., Lioa P.H. & Lo K.V., Anaerobic Digestion of Cheese Whey Using an Upflow Anaerobic Sludge Blanket Reactor 3.Sludge and Substrate Profiles , Biomass (21) 257-271. 1989 Yan J.Q.,Lo K.V.& Pinder K.L., Anaerobic Digestion of Whey in UASB Reactor (Treatment Efficiency), Paper No.89-6570, Winter Meeting of ASAE, New Orleans. 1989 Yan J.Q.,Lioa P.H. & Lo K.V., Methane Production from Cheese Whey, Biomass (17) 185-202 . 1989 Yan J.Q.,Liao P.H. & Lo K.V., Anaerobic Digestion of Cheese Whey Using Up-flow Anaerobic Sludge Blanket reactor (Start-up), Biological Waste,(21) 289-305. 1987 Yan J.Q., Advances in Nitrification and Denitrification Processes. Environmental Science and Technology, No.l.(Chinese) 1986 Yan J.Q., Denitrification of Photosensitive Emulphor Wastewater, Environmental Science and Technology, No.4.(Chinese) 1986 Zhou T.W. & Yan J.Q., The Bifurcation Behavior of the Steady States of Continuous Stirred Tank reactor. Journal of Engineering Mathematics. Vol.3, No.2. (Chinese) 1985 Yan J.Q. The Studies of Anaerobic Expanded Bed, paper in Guangzhou Institute of Energy Conversion .Chinese Academic of Sciences. 1983 Yan J.Q>. Studies in Solar Energy Storage Using Adsorption and Desorption. Energy Conversion No.l .(Chinese) 1981 Yan J;Q.. Studies in the Pulse reaction Kinetics, Master's thesis, Dalian Institute of Chemical Physics, Chinese Academy of Sciences. 1980 Yan J.Q., Discrimination of The Kinetic Models of Catalytic Pulse Reaction. The Paper presented in the 1st Conference of Catalysis and Kinetics, Chinese Chemical Society, Chengdou. China, July. 1980 Yan J.Q. & Jiang B.N., The Determination of adsorption Parameters under the Pulse reaction Conditions, The paper presented in the 1st Conference of Catalysis and Kinetics, Chinese Chemical Society, Chengdou. China, July .