:^; :-;--:-: ■■■: ■,- wmmt>> Hkfl • WHS** u** 1 1IBHAHY OF THE U N IVLR.SITY Of ILL1 NOIS 628 IJL65c no. 33-35 ENGINEERING HWf. ROOM ENSI&''. ftUtf s T oon^1 S T Ch3rging this mate ™l is re- sponsible for it. return on or before the Latest Date stamped below Thef,, mutilation, and underlining of books . ...c university. Digitized by the Internet Archive in 2013 http://archive.org/details/flocculationclar35pers ;|VIL ENGINEERING STUDIES SANITARY ENGINEERING SERIES NO. 35 FLOCCULATION AND CLARIFICATION IN ACTIVATED SLUDGE SYSTEMS By EDWARD RICHARD PERSHE Supported By FEDERAL WATER POLLUTION CONTROL ADMINISTRATION RESEARCH FELLOWSHIP FI-WP-21,618 RESEARCH PROJECT WP-00640 DEPARTMENT OF CIVIL ENGINEERING UNIVERSITY OF ILLINOIS URBANA, ILLINOIS JUNE, 1966 FLOCCULATXON AND CLARIFICATION IN ACTIVATED SLUDGE SYSTEMS by EDWARD RICHARD PERSHE THESIS Submitted in Partial Fulfillment of the Requirements for the Degree of Doctor of Philosophy in Sanitary Engineering in the Graduate College of the University of Illinois, 1966 Urbana, Illinois FLOCCULATION AND CLARIFICATION IN ACTIVATED SLUDGE SYSTEMS Edward Richard Pershe, Ph.D. Department of Civil Engineering University of Illinois, 1966 The factors affecting flocculation and clarifi- cation in activated sludge systems were investigated. This was done by varying the substrate loading rates of both batch and continuously fed activated sludge systems and observing how such loading rates influenced the ability of the activated sludges to flocculate and produce clear supernatants after settling. It was found that increased loading rates and decreased stabilization times were primarily responsible for the disruption of good activated sludge systems. It was shown that the optical density of the supernatant liquor from settled activated sludges ex- hibited progressive deterioration through an increase in turbidity as the substrate loading was increased and as the activated sludges went from a good flocculating con- dition to one which was nonf locculating . The optical density was also found to be a better control parameter than the sludge volume index which remained almost constant until the activated sludge system failed under high substrate loading . 2 Oxygen uptake rates of the activated sludge organisms were measured so as to determine whether there were any significant variations in the uptake rates in the various activated sludge systems. It was found that poorly flocculating sludges had high oxygen uptake rates. Attempts were made to flocculate nonf locculating activated sludges by using various flocculating agents. Inorganic salts were used to determine whether these ions were particularly bene- ficial in promoting f locculation and could be used to explain the mechanism of f locculation. It was found that conventional flocculating agents such as alum and lime could promote flocculation, however, ions of neutral salts as well as pH adjustment had no beneficial effect. Examination of good flocculating and nonf locculating activated sludges by Infrared spectra analysis showed that no significant changes in chemical composition could be detected. Zeta potential was not found to play any significant role. Treatment of a bulking sludge showed that several types of treatments could be successfully utilized for control of a bulking condition. Ill AC KNOWLEDGMENTS The author wishes to express his sincere appreciation to all those who helped hira in the investi- gation and in the preparation of this thesis. He wishes especially to thank the following s Dr. R. S. Engelbrecht for his advice and en- couragement during the conduct of the research and preparation of the thesis „ Dr. R. E. Speece for his special interest in the investigation and for his stimulating and helpful sug- gestions during the conduct of the research. The author is also deeply grateful to him for the time he devoted to reading and discussing the thesis. All personnel in the Sanitary Engineering Laboratory who have contributed to the successful completion of the research «, The National Science Foundation for providing financial support through a Science Faculty Fellowship during the initial year of study . Support for the final period of study was provided through a Research Fellowship from the Division of Water Supply and Pollution Control, Uo S Public Health Service, Department of Health, Education, Welfare, and by financial aid from the Ford Foundation. IV TABLE OF CONTENTS ACKNOWLEDGMENTS LIST OF TABLES Page iii viii LIST OF FIGURES ± I . INTRODUCTION 1 II . LITERATURE REVIEW 3 IIIc SCOPE OF INVESTIGATION 23 IV. MATERIALS AND METHODS 27 A. Experimental Design 27 1. Studies of variable loading rates 27 2. Studies of continuous flow activated sludge systems 29 3. Effect of limiting oxygen supply 31 4. Studies of flocculation without chemicals 32 5. Chemical flocculation studies 33 6. Bulking sludge control studies 34 Bo Analytical Techniques 35 1. Mixed liquor suspended solids 35 2 . Chemical oxygen demand 36 3. Oxygen uptake 36 4. Sludge volume index 36 5. Deoxyribonucleic acid 37 Page 6. Supernatant turbidity by optical density 38 7. Infrared spectrum analysis 40 8. Mobility 42 C. Experimental Technique and Equipment 42 1. Source of sludge 42 2. Variable loading rate studies using batch units 45 3. Continuous flow studies 50 4. Limiting oxygen supply studies 53 5. Studies of flocculation without chemicals 56 6. Chemical flocculation studies 57 7. Bulking sludge studies 59 V. RESULTS AND DISCUSSION 63 A. Variable Loading Rate Studies Using Batch Units 63 1„ Effect of F/M/day loading on substrate removal 54 2„ Effect of P/M/day loading on sludge settleability and effluent quality 77 3. Effect of F/M/day loading on oxygen uptake 81 4. Effect of stabilization time on substrate removal 83 5 Effect of stabilization time on sludge settleability and effluent quality 89 60 Effect of stabilization time on oxygen uptake 93 vi Page 7, Effect of F/M/day loading on oxygen uptake gg 8. Relationship of oxygen uptake to substrate removal gg 9o Relationship of oxygen uptake to settleability and effluent quality 103 10 e Effect of F/M/day loading on cell replication in relation to sludge settleability 105 B. Continuous Feed Studies 108 1. Infrared absorption spectra analysis 110 2. Oxygen uptake characteristics HI C. Limiting Oxygen Supply Studies 117 1. Limited oxygenation using spargers 117 2. Limited oxygenation using stirring 124 D. Electrophoretic Studies 130 Ee Flocculation Studies Without the Use of Chemicals 132 1. Effect of starvation 132 2 Effect of centrifugation and resuspension 134 3 Effect of mixing a flocculating activated sludge with a nonf locculating sludge i36 F Chemical Flocculation Studies 138 lo Effect of alum on nonf locculating activated sludge 138 2o Studies with calcium hydroxide 140 Vll Page 3. Studies with calcium salts 141 4„ Studies with sodium hydroxide 142 G. Bulking Sludge Control Studies 143 1. Effect of using bentonite 143 2. Effect of using alum with bentonite 150 3. Effect of using chlorine to control bulking ^.51 VI. SUMMARY AND CONCLUSIONS 152 VII. SUGGESTIONS FOR FUTURE WOR* 1 59 APPENDIX A APPENDIX B APPENDIX C APPENDIX D REFERENCES VITA 160 162 164 165 166 172 viii LIST OF TABLES Page Loading Factors and BOD Removal Efficiencies for the Complete Spectrum of Activated Sludge Process Variations c Concentrations of Substrates and Inorganic Salts in Fill and Draw Batch Units After Feeding Batch Unit Treatments for Bulking Sludge Control Study Summary of F/M/Feeding Ratios and F/M/Day Loading Ranges for Batch Units in Variable Loading Rate Studies Sludge Volume Index and Optical Density Values for Variable Loading Rate Studies 75 Average Solids Concentrations During Contact Periods in Variable F/M/Day Loading Studies 78 Batch Unit Loading Values for Cell Replication Studies 8 Summary of Cell Replication Studies 44 61 65 107 109 Characteristics of Continuously Fed Units at Various F/M/Day Ratios 2.13 Effluent Clarification Characteristics of Nonflocculating Activated Sludge Organisms 135 Results of Mixing a Flocculating Sludge With a Nonflocculating Sludge 2.37 IX LIST OF FIGURES Page Schematic Growth Curve of Activated Sludge Processes (18) 2 Typical Activated Sludge Process Flow Diagram Showing Effect of Sludge Volume Index on Mass Balance (17) 9 3 Effect of Sludge Volume Index and Return Sludge Ratio on Mixed Liquor Suspended Solids Level in the Aeration Tank 10 4 Effect of Loading Factor on Sludge Volume Index 12 5 Effect of BOD Loading on Sludge Volume Index 13 6 Effect of Substrate Loading on Effluent Quality 14 Standard Curve for Deoxyribonucleic Acid Determination, Burton's Method 39 8 Relationship Between Optical Density and Suspended Solids 41 9 Batch Unit for Variable F/M/Day Studies with Automatic Feeding Device 47 Continuous Flow Activated Sludge Unit 51 Apparatus Used for Limiting Oxygen Supply Study-Mixing by Recirculating Gas Pump 54 Changes in System Parameters During 8 Hour Feeding Cycle 66 Changes in System Parameters During 6 Hour Feeding Cycle 67 Changes in System Parameters During 4 Hour Feeding Cycle 68 Changes in System Parameters During 3 Hour Feeding Cycle 69 Figure Page 16 Changes in System Parameters During 2 Hour and 1 Hour Feeding Cycles 70 17 Oxygen Uptake by Washed Suspensions of Activated Sludge — 8 Hour Feeding Cycle 71 18 Oxygen Uptake by Washed Suspensions of Activated Sludge — 6 Hour Feeding Cycle 72 19 Oxygen Uptake by Washed Suspensions of Activated Sludge — 4 Hour Feeding Cycle 73 20 Oxygen Uptake by Washed Suspensions of Activated Sludge--3 Hour, 2 Hour, and 1 Hour Feeding Cycles 74 21 Effect of Contact Time and Biological Solids on Total COD Removal at Various Loading Rates 76 22 Effect of F/M/Day Loading on SVI and Effluent Quality of Laboratory Activated Sludge Units 79 23 Relationship Between F/M/Day Loading and Oxygen Uptake for Contact Time and for Feed Cycle Time in Terns of Applied COD 82 24 Relationship Between Stabilization Time and Oxygen Uptake for Contact Period and for Feed Cycle 84 25 Effect of Stabilization Time and Loading Rate on 30-Minute COD Removal Capacity 85 26 Effect of Loading Rate on 30-Minute COD Removal 88 Effect of Stabilization Time on Total COD Removal at Different Loading Rates 90 28 Effect of Stabilization Time on SVI and Effluent Quality of Laboratory Activated Sludge Units 91 29 Effect of Stabilization Time and Loading Rate on Oxygen Consumption Rates of Activated Sludge Organisms 94 xi Figure Page 30 Relationship of Stabilization Time to Net Oxygen Uptake Rate 95 31 Relationship Between Oxygen Uptake During Last 30 Minutes of Feed Cycle and Stabilization Time at Different F/M/Feeding Ratios 97 32 Effect of Loading Rate on Net Rate of Oxygen Uptake 93 33 Effect of Loading Rate on Oxygen Uptake During Last 30 Minutes of the Feed Cycle Time 100 34 Relationship Between Initial 30-Minute Oxygen Uptake and Terminal 30-Minute Oxygen Uptake ^0"L 35 Relationship of COD Removal and Loading Rate to Average Net Oxygen Consumption Rate of Activated Sludge Organisms 102 36 Relationship of Oxygen Consumption Rates of Activated Sludge Organisms to SVI of Activated Sludge and Effluent Quality 104 Deoxyribonucleic Acid in Activated Sludge — 4 Hour, 3 Hour, and 2.25 Hour Feeding Cycles 106 38 Oxygen Uptake by Unwashed Suspensions of Activated Sludge from Continuous Feed Unit 112 39 Limiting Oxygen Supply Study. Unit 1, Control, 100 Per Cent of Oxygen Demand Allowed, Mixing by Sparger 118 40 Limiting Oxygen Supply Study. Unit 2, 40 Per Cent of Chemical Oxygen Demand Allowed, Mixing by Sparger 120 *1 Limiting Oxygen Supply Study. Unit 3, 20 Per Cent of Chemical Oxygen Demand Allowed, Mixing by Sparger 121 Figure xn Page 42 Limiting Oxygen Supply Study „ Unit 4, 10 Per Cent of Chemical Oxygen Demand Allowed, Mixing by Sparger 122 43 Limiting Oxygen Supply Study. Unit 1, Control, 100 Per Cent of Chemical Oxygen Demand Allowed, Mixing by Stirring 125 44 Limiting Oxygen Supply Study. Unit 2, 40 Per Cent of Chemical Oxygen Demand Allowed, Mixing by Stirring 12 6 45 Limiting Oxygen Supply Study. Unit 3, 20 Per Cent of Chemical Oxygen Demand Allowed , Mixing by Stirring 127 46 Limiting Oxygen Supply Study. Unit 4, 10 Per Cent of Chemical Oxygen Demand Allowed, Mixing by Stirring 12 8 47 Effect of Settling Time on Supernatant Quality of Activated Sludge Cultures 133 48 Effect of Chemical Treatment on the Flocculation of Nonf iocculating Activated Sludge Cultures ^9 23-Hr Parameters for Unit 1 Fed 2.0 gm COD/ Liter/Day and 50 mg Bentonite/Liter/Day— Per Cent Wasting 144 23-Hr Parameters for Unit 2 Fed 2.0 gm COD/ Liter/Day, 50 mg Bentonite/Liter/Day, 100 mg Alum/Liter/Day — Per Cent Wasting 145 51 23-Hr Parameters for Unit 3 Fed 2.0 gm COD/ Liter/Day and 50 mg Bentonite/Liter/Day — 50 Per Cent Daily Wasting of MLSS 146 23-Hr Parameters for Unit 4 Fed 2 gm COD/ Liter/Day and 100 mg Alum/Liter/Day — 50 Per Cent Daily Wasting of MLSS 147 Figure xm Page 53 23-Hr Parameters for Unit 5 Fed 2 gm COD/ Liter/Day, 50 mg Bentonite/Liter/Day, and 100 mg Alum/Liter/Day — 50 Per Cent Daily Wasting of MLSS 148 54 23-Hr Parameters for Unit 6 Fed 2 gm COD/ Liter/Day with Chlorine Added to Control Bulking by Tapleshay Method — 50 Per Cent Daily Wasting of MLSS 2.49 I . INTRODUCTION An integral part of the activated sludge process is solids separation and disruptions that come about in the process are usually related to this phase. Present loading rates are limited by the inability to separate the biological solids from the effluent. This indicates that much study of the solids separation phase is required for further develop- ment of the process o Much attention has already been devoted to sub- strate and biochemical oxygen demand (BOD) removal in the activated sludge process and advantage has been taken of its property to rapidly remove waste and substrate within a short contact period (1,2). Further attention is needed now in the development of the liquid-solid separation phase in order to more economically design sedimentation and sludge reaeration units, and to improve the efficiency of such units in existing plants (3) . Such developments will also permit further economy through the use of higher food to micro- organism (F/M) loadings as well as reduce the carry-over of solids in small prefabricated activated sludge plants. Since the failure of an activated sludge process as a result of the loss of sludge settleabiiity is very often difficult to remedy, it is important that all factors 2 which can cause such failures be studied fully in order to safeguard the treatment process and to prevent damage which can arise from the discharge of biological solids to receiving streams. Many factors have been cited as being responsible for the flocculation of activated sludge systems. Most authorities believe that F/M loading rates and stabilization time are of primary importance in flocculation (4) although other factors such as type of substrate, oxygen supply, mixing, changes in pH, nutrient deficiency, and temperature changes may be of equal if not greater importance in certain circumstances, it appears, therefore, that flocculation in an activated sludge process is dependent upon a combination of several factors. The research reported in the following presentation was undertaken to investigate some of the principal factors which affect flocculation and clarification in activated sludge systems and to thereby contribute to the improvement of the activated sludge process. II ■ LITERATURE REVIEW Shortly following the experiments of Ardern and Lockett (5) which demonstrated the feasibility of treating sewage with flocculant biological solids, the first con- tinuous flow activated sludge plant was put into service at Houston, Texas in 1917 (6) . In the years that followed, much was learned about the activated sludge process that was un- known at the time of its discovery. For quite some time following the discovery of the activated sludge process, there was a controversy as to whether the process was biological or physical in nature. Buswell and Long (7) are among those credited with proving that the process is biological rather than physical. This important finding guided investigators to more effective re- search in the fields of biochemistry and microbiology. In 1930 new knowledge concerning oxygen require- ments in relation to activated sludge concentrations , temper- ature, substrate, nutrition, and bacterial growth was obtained (8,9). These studies showed that the oxygen re- quirements of the conventional activated sludge process diminish as the treatment of waste progresses and that tapereo aeration could be used for greater economy (10) . Other modifications which were developed to obtain economy for a given degree of treatment were step aeration (11) and modified aeration (12,13). In 1951 Ulrich and Smith (1,2) showed how the great adsorptive properties of well stabilized activated sludge could be exploited in a process called biosorption. In 1954 Wuhrman (14) demon- strated that BOD loadings of 190 lb BOD/day/1000 cu ft could be treated with removals on the order of 80 per cent with a high rate activated sludge process. Other processes which were developed included the dispersed growth system (15} and extended aeration (16) . A list of activated sludge process variations with their respective loading factors and BOD re- moval efficiencies is shown in Table 1. Figure 1 presents a mass-time curve on which are shown the operating ranges of the conventional activated sludge process and three major modifications (18) . The curve is shown divided into three major phases designated as logarithmic growth, declining growth, and the endogenous phase. The log growth phase is characterized by high F/M ratios. In the declining growth phase, decreased net synthe- sis results. Finally in the endogenous phase, F/M ratios become quite low and "starvation" commences. Under steady flow operating conditions of the activated sludge process,, log growth cannot be practically utilized because of the resulting dispersed growth. When TABLE 1 LOADING FACTORS AND BOD REMOVAL EFFICIENCIES FOR THE COMPLETE SPECTRUM OF ACTIVATED SLUDGE PROCESS VARIATIONS* M ' " ~ Name of Process Loading Factor lb BOD/day per 1000 cu ft BOD Removal Efficiency per cent Extended Aeration 20 75-85 ConVc Activated Sis jdge 35 95 Tapered Aeration 35 95 Step Aeration 50 90-95 Activated Aeration 50 80-85 Contact Stabilization 70 85-90 Hatfield Process 70 85-90 Krauss Process 100 85-90 High Rate 100 60-75 Modified Aeration 100 60-75 Rapid Bloc 150 90-95 Supra Activation 400 55-65 ♦After Stewart (17) Loq Declininq Endogenous Growth Growth Phase Time Figure 1. Schematic Growth Curve of Activated Sludge Processes (18) pure substrates such as glucose are brought in contact with an acclimated activated sludge,, zero order kinetics prevail and both substrate removal and excess sludge growth follow a linear pattern (19). However, it has been found in many biological oxidation systems that there is a very high initial rate of BOD removal immediately on contact of waste with sludge which is followed by a slower rate of removal. The removal reaction becomes especially complex when a heterogeneous waste mixture is oxidized. However, the rate usually follows first order kinetics. The various waste constituents are often subject to varying reaction rates re- sulting in a decreasing overall rate as the various com- ponents are oxidized. In some cases, a small fraction of BOD does not appear to be removed even after long aeration periods (20,21) . Experience in the use of the activated sludge pro- cess shows that the process can be operated continuously only in a few F/M ranges. If the process deviates too much from these ranges, it becomes difficult to handle and control, or i.t becomes uneconomical from the standpoint of design because 3f the high recirculation return ratios which must be used to naintain the proper solids concentrations in the aeration rank (18) . Trial and error experience has shown that there »re four ranges wherein it is possible to operate the 8 activated sludge process continuously in a satisfactory manner. These ranges are shown in Figure 1 and are desig- nated as dispersed growth, high rate or modified activated sludge, conventional activated sludge, and extended aeration (18). In order to have an activated sludge process operate at steady flow continuously, it is important to maintain suspended solids of good quality. The activated sludge roust flocculate well and settle rapidly. The ability to remove and concentrate solids in the system is vital to the successful operation of the activated sludge process. This is illustrated in Figures 2 and 3. Figure 2 shows that for steady flow conditions, the product of the mixed liquor volatile suspended solids (MLVSS or h Q ) concentration in the aeration tank and the sludge volume index (SVI) is a constant. If the value of the SVI increases as a result of bulking of the sludge, then the value for the MLVSS concentration must decrease . Figure 3 shows that as the SVI value increases, it becomes increasingly difficult and costly to maintain the proper concentration of mixed liquor suspended solids in the aeration tank. At very high SVI values, it may be impossible to recirculate enough return sludge to maintain adequate solids concentration in the aeration tank. As Figure 3 shows, "hen an activated sludge has a SVI of 400, the maximum solids fluent low, Q (Q + R) @ A, Return Sludge Flow = R Solids Cone. ■ B Effluent Flow, Q Excess Sludge Flow = 1.5% Q SVI (Q +- R)A_ - R-B - R 'l° S SVI Or A s • SVI = 10 R/Q - Constant (R/Q)*l Figure 2. Typical Activated Sludge Process Flow Diagram Showing Effect of Sludge Volume Index on Mass Balance (17) 10 u •H \ 0* E « H H ■o 4) '0 c 11 a 10 3 CO 9 tr 6000 bOOO — 4000 3000 2000 1000 Figure 3 12 3 4 5 Return Sludge Ratio, R/Q Effect of Sludge Volume Index and Return Sludge Ratio on Mixed Liquor Suspended Solids Level in the Aeration Tank 11 concentration that can be maintained in the aeration tanks is about 2100 mg/1 regardless of the amount of recirculation. As stated above, in order to have a properly functioning activated sludge process, the quality of the sludge must be such that it flocculates well and settles quickly leaving a clear effluent. Most biological waste treatment studies have emphasized substrate or BOD removal and although settleability of the sludge is an important factor in the process, it has more or less been a secondary objective. Because of this, field trial and error ex- perience has been used to delineate the so-called safe F/M loading ranges in which the activated sludge process can be operated. A particular range is selected for operational purposes because it produces the desired degree of treatment and also because the settleability of the sludge as indica- ted by the SVI can be controlled. Figures 4 and 5 present similar if not slightly conflicting relationships regarding the effect of F/M loading on the SVI. The most important feature of the two curves is the range in which serious bulk- ing of the activated sludge can be expected. This range extends between the F/M/day loading rates of 0.5 and 2.0. Ironically, the curves suggest that extremes of loading pro- duce the best results. The effect of F/M loading on sub- strate removal is shown in Figure 6 and indicates that 1 2 400 — 300 — 200 _ 100 - 12 3 4 5 Lb BOD Applied/Lb Vol. Sludge/Day (F/M) Figure 4. Effect of Loading Factor on Sludge Volume Index 13 x V c M V E 3 .-I > V CP *0 3 —i 500 — 400 300 200 100 Data from Lesperance (18) J 25 2.0 1.0 Lb BOD/Lb MLS S /Day Figure 5. Effect of BOD Loading on Sludge Volume Index 14 LOO Figure 6. 12 3 4 Lb BOD Applied/Lb Volatile Sludge/Day Effect of Substrate Loading on Effluent Quality 15 substrate removal may vary from 96 to 58 per cent as the F/M/day loading increases from very low loadings to a load- ing of 5 lb BOD/day/lb of volatile solids (17). Pilot plant studies by Logan and Budtf (22) showed that low F/M ratios were necessary to keep the activated sludge in a good state. They obtained a minimum sludge volume index at a loading rate of about 0.3 lb BOD/day/lb activated sludge and showed that loadings above or below this value caused higher SVI values, They also observed that the SVI value lagged behind the applied load about 4 to 6 days. Settled sewage from a primary settling tank was used as the substrate, Stewart (17) obtained results which were contrary to those reported by Logan and Budd. He found a maximum SVI occurred at an F/M of 0.6 lb BOD/day/lb MLSS and that lower SVI values were obtained for both higher and lower F/M ratios. The F/M ratio of 0.6 is within the 0.5 to 1.0 F/M range considered to cause bulking (23). Siddiqi (24) showed that F/M ratios exceeding 9.6 lb COD/day/lb MLSS could be handled in a batch system without incurring a bulking sludge using a substrate which consisted of a mixture of glucose and yeast extract in a ratio of 5:1. Orford, et al.„ (25) attempted to show that a log relationship existed between the SVI and the loading rate, 16 however, the wide variation of values obtained indicates that the relationship is not justified. The previous mentioned studies illustrate that the trial and error development of the activated sludge process does not provide sufficient fundamental knowledge which can help explain why a given process works well at one loading rate while a similar loading may not work well under other conditions. Undoubtedly the trial and error method will continue to be used in the future until the flocculation and clarification process is clearly understood. However, since the biological oxidation process is presently being taxed to its utmost ability, solids separation must be improved in order to exploit the activated sludge process, in this re- spect, research and study into the factors which can affect flocculation and clarification show much promise. Although they have not materially influenced sewage treatment plant design and operation to any great extent, lab studies delving into the causes of flocculation and clarification have been carried on for some time con- currently with field trial and error methods for activated sludge development . Theriault (26) stated that the reactivation of the activated sludge and the stabilization process were closely allied. He hypothesized that the gelatinous matrix of the 17 activated sludge behaves like a base exchanging substance in performing clarification and that the matrix was rejuvenated by the bacteria residing on it. Ruchhoft, et al., (27) showed that clarification and reactivation could be carried out using pure cultures of microorganisms that were isolated from good activated sludge. The fundamental characteristics of the organisms that were isolated and used to demonstrate clarification and re- activation were related to the zooglea-forroing group. These studies were among the first which were undertaken in an effort to determine what caused activated sludge to flocculate. McKinney and Horwood (28), and McKinney (29), began studies to ascertain the causes of floeeulation along the same lines as Theriault. They isolated and identified 11 organisms capable of forming floe similar to that in an activated sludge process when the organisms were aerated in a nutrient solution. However, they found that floe forma- tion in the bacterial cultures was not as good as that ob- tained when sewage seeded sludge was used. Most of the BOD removal by the organisms was accomplished within the first two hours of aeration. Similar results were reported by ran Gils (30) . Some aspects of the work by McKinney and Horwood 18 (28) are of interest. Through microscopic observations it was found that bacterial cells would clump together after several hours aeration for no apparent reason. They be- lieved that the bacterial cells were joined by their capsules The chemical composition of the slime layer or capsule was hypothesized to contain electric charges which resulted from certain ionizing groups and it was believed that the inter- action of oppositely charged ions with the bacterial cells caused the flocculation phenomena. However, McKinney re- ported later (31) that all bacteria examined for electric charge on their surfaces had surface charges below that con- sidered critical for bacterial flocculation (32) . From his flocculation studies with pure cultures and the F/M ratios used in various activated sludge pro- cesses, . McKinney (31) concluded that biological flocculation as it occurred in biological waste treatment processes was not the result of any special zooglea-producing bacteria but rather it was the result of the energy content of the system or the F/M ratio. He stated that high F/M ratios produced a dispersed or nonf locculent growth and that lowering the F/M ratio slightly would result in partial flocculation. If the F/M ratio was lowered still further, better flocculation would take place. It was also found that excess aeration caused a loss of microorganism activity without a; 19 flocculation. Studies by van Gils (30) confirmed the pure culture studies of McKinney and Korwood (28), however, very few factors were found which could be correlated with the floccu- lation of activated sludge. Crabtree, et al. ( (33) studied pure cultures of Zooglea ramigera and found that production of an intra- cellular storage granule, identified as poly-beta-hydroxy- butyrate, could be correlated with flocculation of the organisms. When the organisms were grown in growth limiting organic substrates, the polymer did not accumulate and the culture did not flocculate. When the carbon to nitrogen ratio exceeded 20:1, the accumulation of granules was en- hanced as was flocculation. It is known that many types of protozoa feed on bacteria or ingest suspended organic debris and that some secrete mucoid substances which cause bacteria to floccu- late (3) . The significance of these activites in relation to the efficiency of treatment plants is, however, still not known. Certain species of ciliated protozoa are usually dominant in activated sludge plants producing effluent of low suspended solids and their presence is often taken as indicative of a sludge in healthy condition. The numbers and types of protozoa can fluctuate quite widely even in plants 20 operating under fairly constant conditions. It ,as been es- timated that the total concentration or sludge consumed per day by protozoa must be of the order of 190 parts per million or about 6 per cent of the total concentration (3). Sludge bulking may be of two types. One type re- sults from high carbohydrate and causes Sj^haerotilys growths. The other type is known as zoogleal bulking and results from high F/M ratios. Heukelekian and Weisberg (34) have shown that the zoogleal type of bulking is accompanied by a high bound water content whereas the bound water content of a Sphaerotilus bulked sludge remains rather constant. Ac- tivated sludges having good settleability and a low svi were found to have a low bound water content. Sludges which had a high SVI also had a high bound water content. It was found that addition of chlorine has an im- mediate effect in reducing the SVI and the bound water (34) . It was also shown that the bound water content of a bacterial culture decreases with the length of the aeration period. Solvation is considered to be an important stability factor * colloidal suspensions and hydration of lyophillic colloids >»ch as bacterial cells is assumed to prevent the aggregation 'f particles (35) . Gould (36) reports that if the activated sludge is eparated from the sewage immediately after BOD removal, it 21 is found to retain a high fraction of its purification power and may be immediately recirculated. The purification amounts to 60 or 70 per cent and the aeration process time required is only one or two hours, if instead of removing the sludge after this initial period, aeration is continued, the sludge becomes bulky. The added purification is of little practical value because the sludge may not be efficiently separated from the sewage. Continued aeration for a total of perhaps six hours brings the sludge back to its original settling properties. Tenney and Stumm (37) have shown that dispersed bacterial cultures can be readily flocculated with chemical coagulants. From their studies they postulated that- bacterial self-fiocculation can be interpreted in terms of an interaction of naturally produced polyelectrolytes which form bridges between the individual microbial particles. They also proposed that bacterial waste products such as de- gradation intermediates and polysaccharides play an important role as polymers in biof locculation. Smallwood (38) attempted to show that agglutination of the activated sludge was related to the ion-exchange Properties of the sludge. It was found that activated s had a very small ion-exchange capacity of the order of one milliequivalent per 100 grams of dried sludge for many in- 22 organic cations. It was further found that the exchange capacity varied during the progress of the purification reaching a maximum during the first hour. High concentra- tions of inorganic salts were found to have only a slight effect on f locculation. Smallwood concluded that no single mechanism can account for all the phenomena observed in the activated sludge process. 23 III. SCOPE OF INVESTIGATION The basic objective of this study was to investi- gate those factors affecting f loeculation and clarification in activated sludge systems, in the conventional activated sludge process, sewage is mixed under aeration with a sufficient amount of biologically active sludge for a suitable period of time. The finely divided suspended and dissolved solids in the sewage are transferred to or are accumulated by the sludge floe particles in which large numbers of living organisms are able to exist. The organ- isms oxidize the organic matter and the floe particles increase in size. If the activated sludge suspension is allowed to stand under quiescent conditions following the oxidation process, the floe particles settle out in a rela- tively short period of time leaving a clear, relatively stable supernatant. The liquid-solid separation phase is normally per- formed in the final clarifier. The efficiency of floecu- lation in the settling process can be evaluated by observing the degree of clarification in the effluent as indicated by the suspended solids concentration. The quality of the sludge floe can be judged by its ability to settle within a relatively short period and by its density or degree of 24 thickening. Although much study has been devoted to de- velopment of biological systems which can satisfactorily treat various types of wastes, little attention has been given to the factors which affect the all-important liquid- solid separation phase as influenced by flocculation of such systems. In this investigation of the factors affecting flocculation and clarification in activated sludge systems, care was taken not to allow any variation of conditions to take place which might prevent a clear evaluation of the factor being studied. Substrate composition, oxygen supply, pH, nutrient concentrations, and temperature were all controlled as much as possible. F/M/feeding ratios of 0.2, 0.4, and 0.8 were employed in batch fed activated sludge systems in order to study their effect on the SVI of the activated sludge and the clarity of the effluent. The F/M/day loadings of the batch units were varied by increasing the number of feedings per day. The feeding frequency was increased from 3 to 24 feedings per day for each F/M/feeding ratio studied. A wide spectrum of F/M/day loading rates was covered extending from approximately 0.5 to 20. The biological solids growth and substrate removal were studied during each growth cycle and Warburg studies were made to observe oxygen uptake character- 25 istics of the sludge organisms, Separate batch studies were also conducted in order to study the effects on an activated sludge system when the amount of oxygen the system was allowed to have during the growth cycle was limited. Efforts were made to correlate the results obtained from the batch studies with those of a continuous flow system. A continuous flow unit was operated at different F/M/day ratios to determine how the SVI of the activated sludge in a continuous flow system compared with the results obtained from batch units which were operated at about the same loadings. Oxygen uptake studies were also made of the continuous flow unit sludge organisms. Samples of activated sludge from the continuous flow unit were taken during periods when the sludge settled well and when it settled poorly. The samples were subjected to infrared spectrum analysis to determine whether changes in the chemical composition of the culture took place. Samples of activated sludge were also taken during the growth cycles of batch operated units to see whether the organisms showed any difference in charge density as measured by their electro- phoretic mobility. Flocculation studies of nonf locculating activated sludge cultures were performed both without and with the use of chemicals in order to determine whether such methods had 26 any effect on f lobulation, in the studies without chemicals, samples of nonf locculating activated sludge were centrifuged and resuspended under various conditions, m the chemical flocculation studies the effect of trivalent aluminum and various calcium compounds in promoting chemical flocculation was observed. Studies were also made of treatments to correct a bulking activated sludge condition in order to determine if some relationship to flocculation could be obtained. The treatments consisted of dosing the bulking sludge with various substances at the time of feeding. The substances used for treating the bulking sludge were bentonite, alum, combinations of bentonite and alum, and chlorine. The effect of wasting during treatment was also studied. 27 IV o MATERIALS AND METHODS A. Experimental Design 1. Studies of variable loading rates In this study the F/M/day loading factor was studied under carefully controlled conditions. Batch units were used in order that observations could be made of the changes in various parameters as they occur during the waste treatment period. The units were fed at desired inter- vals in a manner similar to a conventional activated sludge process. This method is described later. The substrate chosen for this study was a mixture of glucose and yeast extract in a ratio of 5:1 as COD. Most activated sludge bacteria require either preformed vitamins or amino acids, or both (39). Zoogloea ramigera . one of the organisms reportedly responsible for the flocculation of activated sludge, has been shown to require certain growth factors (33,39). Since the use of a pure carbohydrate sub- strate would have restricted the establishment of a well- balanced heterogeneous microbial population, yeast extract was used jointly with glucose. Yeast extract contains most of the amino acids and growth factors ordinarily required by microorganisms. The pH of the activated sludge cultures was 28 buffered in a range which excluded selection for filamentous organisms. Periodic microscopic observations were also made to insure that such microorganisms did not gain predominance. Sufficient oxygen was made available to the organisms in the reactor units at all times and nutrients including nitrogen were supplied in slight excess to known metabolic require- ments (40,41) . Thus all factors known to cause interference were controlled so that a variation in the F/M/day ratio would indicate whether this factor was of importance in the floccu- lation of activated sludge systems. Furthermore, the loading rate at which deterioration of the settling process took place was determined and could be used as a basis for es- tablishing design criteria. The parameters studied to observe the performance of the activated sludge systems were the growth of biological solids, soluble chemical oxygen demand (COD) remaining, the sludge volume index (SVI), and the clarity of the supernatant liquor after settling as measured by the optical density (0D) . From the solids growth and the soluble COD remaining data, contact and stabilization periods were closely esti- mated. These parameters in turn were used to further define the effects of varying F/M/day ratios on the flocculation phase . 29 In addition, the oxygen uptake characteristics of the activated sludge cultures at each specific F/M/day load- ing ratio were studied in a Warburg respiroraeter to see how they correlated with other performance parameters. Since oxygen uptake rates are closely allied with the metabolic activity of the sludge organisms as well as the age of the cultures, flocculation and ability of the sludge to settle could be described in terms of the oxygen uptake behavior of the organisms. Although previous investigators had not been successful in showing that electric charge on the bacterial cell had a role in the flocculation of microorganisms (31), electrophoretic measurements were made of activated sludge particles during various periods of the growth cycle under varying F/M/day loading rates. 2. Studies of continuous flow activated sludge systems Since the results obtained from batch type studies may differ from those obtained from continuous flow units, it was considered of fundamental importance that the results obtained from batch studies dealing with loading rates be correlated by continuous flow studies. More specifically, continuous flow (chemostat) studies were undertaken to de- termine whether the same results would be obtained for compa- rable F/M/day loading ratios. 30 The continuous flow activated sludge unit was fed substrate, nutrients, and buffer of the same type and rela- tive concentration as in the batch units. On the basis of the results obtained in the batch studies, F/M/day loading ratios were selected which produced three types of activated sludge quality. These were: (i) a low F/M/day ratio which would be expected to produce a good settling sludge with a high quality effluent, fii) a high F/M/day ratio which would be expected to produce a nonf locculating culture and (iii) an intermediate F/M/day ration which might produce an ac- tivated sludge having partial qualities of (i) and (ii) . Satisfactory agreement between the batch fed and continuous feed units would permit an explanation of the continuous flow process from the results of the batch studies. The parameters studied included the F/M/day ratio, the SVI, supernatant clarity as measured by the optical density, and oxygen uptake characteristics. Samples of the sludge cultures were also taken during the continuous flow studies for the purpose of infra- red analysis. Some investigators (42) have shown that pure cultures of certain organisms will produce characteristic ab- sorption bands in the infrared spectrum during certain periods in the age of the cultures. The appearance of absorption bands in bacterial spectra has been correlated with the use 31 of a particular substrate (43), and the presence of inter- mediate or end products of the bacterial metabolism (44) . Walters (45) has shown that a sharp absorption band at 5.7 microns is obtained for activated sludge cultures. The substance responsible for the 5.7— micron band has been isolated and identified as polymerized-hydroxybutyric acid. If the activated sludge cultures underwent any significant changes when the F/M/day ratio was changed from a low value to a high one,, the infrared analysis would be expected to show this. 3. Effect of limiting oxygen supply As a further check on the oxygen uptake character- istics of activated sludge cultures as studied in the Warburg respirometer, studies were undertaken to see what effect limiting the oxygen supply to the culture would have on the growth and substrate removal capacity of activated sludge systems as well as the SVI of the sludge and the turbidity. Accordingly, batch fed units were started at the same F/M/day loading rate in a manner similar to other batch studies except that the amount of oxygen that the unit was allowed to have during the growth cycle was limited. The units were mixed in two different ways. One method used gas recirculating pumps to diffuse the available oxygen supply in the mixed liquor. The other mixing method used magnetic 32 stirrers. These experiments were designed to observe the effects on growth and other performance characteristics when only a portion of the chemical oxygen demand was supplied. 4. Studies of flocculation without chemicals These studies were undertaken to determine how an activated sludge culture in a nonf locculating state would respond to efforts to flocculate it without the use of chemicals . The first experiment of the series consisted of investigating the effect of starvation on a nonf locculating activated sludge. The sludge was "starved" by subjecting it to a long endogenous period of approximately two weeks. Since good settling activated sludges obtain their settling characteristics during the endogenous phase of the growth cycle, the nonf locculating activated sludge culture would be expected to improve its ability to flocculate and produce a clear effluent. After the starvation period, the nonf loccu- lating activated sludge was observed for improvements in flocculation character. The object of the second study was to determine whether a nonf locculating activated sludge could be made to improve its ability to flocculate by forcing the nonfloccu- Lating cells to make contact with each other. Since non- 33 flocculating activated sludge cultures are greatly hydrated, the barrier to flocculation might be removed if the cells could be forced together to push the layer of water mole- cules aside. If this could be done, it would be expected that the culture would flocculate easily if it was re- suspended. Accordingly, samples of nonf locculating ac- tivated sludges were centrifuged and resuspended under various conditions. The cultures were then observed to see if improvement in flocculation was obtained. The third study consisted of investigating what effect a good settling activated sludge has on one which is nonf locculating. Good settling activated sludges are said to possess polymers which cause a cross-linking pattern and thereby cause flocculation (37). If a satisfactory floccu- lating activated sludge is blended with one which is a non- flocculating system, it may be expected to promote an improvement in the latter. A mixture of the two sludges was made in a ratio of 1:1 and after a 20-minute aeration period, the mixture was examined for improvement in sludge settleability and effluent clarity. 5. Chemical flocculation studies Chemical flocculation studies were employed to determine whether the bacterial surface characteristics in nonflocculating activated sludge cultures would respond to 34 chemical treatment and thereby improve their flocculating properties. Experiments were designed to show the effects of the electrovalency of the cation, the merit of adding salts containing the same cation in tooth soluble and in- soluble form, and the effect of pH. Alum was selected to demonstrate the effect of the trivalent aluminum cation while calcium hydroxide was employed to show the effect of the divalent calcium cation. Since very little calcium hydroxide is soluble in aqueous solution, calcium hydroxide was also used in soluble form only to find what effect the colloidal material represented by insoluble calcium hydroxide would have in f locculation. Sodium hydroxide was used in order to explain the effect that hydroxide ion has on nonf locculating cultures. After the addition of the chemicals, the culture samples were slowly flocculated for a brief period and then allowed to settle. The supernatant from the samples was examined for clarity as denoted by the optical density. 6. Bulking sludge control studies These studies were undertaken to determine whether a bulking sludge could be controlled by using additions of bentonite, alum, combinations of bentonite and alum, or chlorine. Since activated sludge has a low ion-exchange 35 capacity (38), additions of bentonite would increase the ion exchange capacity and provide a more reactive environment for a strong cation like trivalent aluminum or its hydroxide. Under such conditions, good coagulation could take place and the bulked activated sludge would be expected to settle and produce a clear supernatant. Bentonite additions would also be expected to promote perikinetic coagulation. Additions of chlorine have been successfully em- ployed to correct a bulking activated sludge (46,47,48). Therefore, this method of controlling a bulking sludge con- dition was also selected for study to compare its effective- ness with other methods. Accordingly, the additives used to control bulking were added daily with the substrate to an activated sludge which had been previously induced to bulk* The activated sludges in the reactor units were tested periodically for solids concentration, SVI, and supernatant clarity after settling. B. Analytical Techniques 1. Mixed liquor suspended solids Mixed liquor suspended solids were determined by the membrane filter technique described by Winneberger et al. (49) e Standard filter disks having a pore size of 0.45 micron and a diameter of 47 mm were used for all determina- 36 tions. 2. Chemical oxygen demand Details of the Chemical Oxygen Demand test, COD, are described in Std. Methods (50). Silver sulfate was used as a catalyst in all determinations,, A correction for chlorides was eliminated by treating the sample with mercuric sulfate before refluxing. 3. Oxygen uptake Oxygen uptake was measured with a Warburg respi- rometer. Calibration of flasks and manometers was ac- complished by using the ferricyanide-hydrazine method (51). All sludge organisms, with the exception of those from the continuous flow unit, were washed once in 0„02 M phosphate buffer, pH 7.2, then resuspended in the same buffer solution to prevent variation in the pH„ The substrate and nutrients used to feed the sludge suspensions were placed in the flask sidearm. Flask contents were allowed to equilibrate in the bath for 20 minutes at 20° C before the Warburg studies were begun by tipping the sidearm contents into the main compart- ment. A shaking rate of 200 strokes per minute was used. Brodie fluid was used in the manometers. 4. Sludge volume index Details of the sludge volume index determination, SVI, are described in Standard Methods (50) . The SVI is the 37 volume in ml occupied by one gram of activated sludge after settling the aerated liquor for 30 minutes. Because of the limitation of the sample volumes, a 100-ml graduate was used instead of the prescribed lOOO-ml graduate. The following relationship was used to calculate the SVI: ml of settled sludge percent suspended solids 5„ Deoxyribonucleic acid Deoxyribonucleic acid, DNA, was isolated and de- termined coloriraetrically by means of a diphenylamine re- action after the method described by Burton (52). Because it was necessary to limit the size of sample taken for the test, a modification of the method was used„ All samples were frozen as quickly as possible after they were taken and were kept in the frozen state until all samples were ready for analysis. The following procedure was employed for the iso- lation and colorimetric determination of DNA. Samples were thawed in cold water, placed in plastic centrifuge tubes and centrifuged at 12,000x g for 10 minutes. The supernatant was discarded and 5 ml cold saline solution was added to each centrifuged pellet. The pellet was resuspended in the saline solution and 4 ml of the suspension were transferred to a small pyrex centrifuge tube. Then 0.2 ml concentrated per- 38 chloric acid was added to each tube and gently mixed. The tubes were immersed in a hot bath for 30 minutes at 70° C. Following removal from the hot bath, the tubes were allowed to cool slightly before they were centrifuged at 12,000x g for 10 minutes. A 2 ml aliquot of the super- natant liquid was placed in a glass tube following which 4 ml of freshly prepared diphenylamine reagent were added and gently mixed. The tubes were allowed to stand at room temperature for 20 hours before optical density readings were made using a wavelength of 600 millimicrons. To obtain concentrations of DNA, optical densities were compared with those usea to prepare a standard absorb- ance curve for DNA. The standard curve was prepared using highly polymerized calf-thymus DNA and is shown in Figure 7. 6. Supernatant turbidity by optical density The turbidity of the supernatant liquor of all ac- tivated sludge suspensions which had been allowed to settle 30 minutes was determined by optical density measurement using a wavelength of 530 millimicrons. This determination was selected to evaluate supernatant quality because of its relative simplicity and the small amount of time required to make the test. Optical density has been ysed by others (24,41,4s, 61) to measure growth in bacterial suspensions. In order to 39 1.0 Figure / 100 200 300 400 DNA, micrograms in tube Standard Curve for Deoxyribonucleic Acid Determination, Burton's Method 500 40 demonstrate that light absorption by bacterial suspensions is proportional to the concentration of the suspension as measured by the optical density, a bacterial suspension having a concentration of approximately 2000 rag/liter after settling 30 minutes was diluted to concentrations of approximately 1500, 1000, 750, 500, 250, and 125 rog/liter. The optical density of each of the seven suspensions as well as their suspended solids concentration was determined. The respective data are presented in Figure 8 and are desig- nated by numerals 1 through 7 for successive lower con- centrations. The optical densities of other bacterial concentrations made at random intervals during the course of the investigation are also shown . Although there is some variation in the optical densities at high mixed liquor suspended solids concentrations, there is a good correlation between solids concentration and optical density at con- centrations less than 1000 mg/liter. The instrument used to make the optical density readings was a Lumetron Colorimeter Model S-48K, manufactured by Photovolt Corporation, New York, N. Y. 7. Infrared spectrum analysis Infrared spectrum analysis of sludge samples was performed in the Infrared Analytical Laboratory, Department of Chemistry and Chemical Engineering, University of 41 1.0 b — 0.6 — 0.4 - 0.2 — Figure 8 500 1000 1500 2000 Mixed Liquor Suspended Solids, mg/liter Relationship Between Optical Density and Suspended Solids 2500 42 Illinois. A Perkin-Elmer Model 521 Grating Infrared Spectrophotometer was used. 8. Mobility The procedure for determining the mobility of colloidal particles is described by Black and Smith (53) . Equipment modifications included the following. A Zeiss binocular microscope was used to make the observations. The lens system consisted of a 32x dry objective and 16x eyepieces. Darkfield illumination was employed. Direct current required to maintain the electric field was furnished by a converter which allowed the use of 110-volt line current instead of a system of batteries. The depth of the microelectrophoresis cell used to make the measurements was approximately 570 microns. C. Experimental Technique and Equipment 1. Source of sludge Fill and draw batch units were used to supply all sludge used in the various phases of the investigation. They were carefully maintained so as to provide a uniform quality of good settling activated sludge at all times, a. Description of fill and draw batch units These units were 2-liter pyrex cylinders having an inside diameter of 3.5 inches and a depth of 20.5 inches. The top of the cylinder was fitted with a removable plastic 43 cap to prevent loss of liquid from spraying. The mixed liquor was aerated by means of a carborundum diffuser stone suspended from a flexible air tube. When filled to the 2-liter operating level, the units had a freeboard of approximately 3 inches. b. Substrate and nutrients A mixture of glucose and yeast extract in a ratio of 5 si as COD was used as the substrate. Nitrogen in the form of ammonium ion was supplied to maintain a COD:N ratio of 15 (40) . Nutrients were supplied to meet minimal medium requirements of most heterotrophic organisms as suggested by Davis and Mingioli (41) . The pH of the mixed liquor was kept between 7.2 and 7.5 by the addition of a phosphate buffer. Table 2 lists the substrate and nutrients used to maintain the batch units. Co Feeding and maintenance The fill and draw batch units were fed once a day at a regular time. The procedure used was as follows, After 23 hours aeration, the volume of each batch unit was brought up to 2 liters by adding distilled water to replace the water lost through evaporation., Solid matter adhering to the inner surfaces of the unit was resuspended by scraping., Then while the reactor contents were being thoroughly mixed, 50 per cent of the mixed liquor (one liter) was drawn out of 44 TABLE 2 CONCENTRATIONS OF SUBSTRATES AND INORGANIC SALTS IN FILL AND DRAW BATCH UNITS AFTER FEEDING Glucose Compound Concentration rag/1 1667 (as COD) Yeast extract 333 (as CQD) NH 4 C1 430 K 2 HP0 4 2070 KH 2 P0 4 MgS0 4 .7 H 2 FeS0 4 . 7H 2 ZnS0 4 . 7H 2 MnS0 4 . 3H 2 CaCl 2 . 2H 2 640 100 2.5 2.5 2.5 13 45 each unit and wasted. The remaining mixed liquor was allowed to settle one hour and the supernatant was then drawn off to a depth of approximately one inch over the settled sludge. Substrate and other nutrients were added and the total volume of the unit was made up to 2 liters with tap water. Aeration of the units was then resumed. Using this procedure, the mixed liquor suspended solids con- centration achieved an equilibrium of 2000 mg/1 at the end of the growth period. Since glucose comprised 83 per cent of the COD supplied to the units, it was necessary to carefully observe the settling character of the activated sludge cultures in each unit and to make periodic microscopic examinations to see if unusual members of filamentous organism group were present. Such organisms, if allowed to predominate, are capable of causing an activated sludge to bulk (54) . When this occurred, the entire contents of the batch unit were discarded and a new culture was started using as seed material, either activated sludge procured from the Urbana- Champaign municipal treatment plant, or waste activated sludge from one of the laboratory batch units. 2. Variable loading rate studies using batch units In these studies, variable loading ratios were obtained by varying the frequency of feeding of the ac- 46 tivated sludge units. The frequencies used in the study were 3, 4, 6, 8, 12, and 24 feedings per day. Since the higher frequencies would have required an abnormal amount of attention throughout the day, the batch units were fed using an automatic feeding device. The principles of complete mixing (55) were used to obtain the proper F/M ratio at the time of the feeding. The mathematical relationships involved are presented in Appendix A. a. Description of batch unit and feeding device Figure 9 shows a typical batch unit with its auto- matic feeding device which was used in this study. Three such units were employed. Each unit was fed at a particular F/M/feeding. The F/M/feeding for Unit Mo. 1 was 0.2, for Unit No. 2, 0.4, and for Unit No, 3, 0.8. Thus a wide spectrum of F/M/day loading ratios was covered in the study ranging from a low of 0„6 to a high of 19,2. The reactor vessel consisted of a plexiglass cylinder which had a 4-inch inside diameter and a 13-inch depth. It had a removable top and the freeboard in the unit when filled with 2 liters of liquid was 3,5 inches. Air was supplied to the unit through a carborundum diffuser after it had been saturated with water vapor, b. Substrate and nutrients A mixture of glucose and yeast extract in a ratio 47 Rotameter Water Vapor \ Saturator Vent Tube Liquid Level Control Tube Pinch Valve "B" So C ■— -T lenoid / ontrol — ' teactor Clock Timer jn Substrate Reservoir Unit Feeder Pinch Valve M A" I Q - 110V Power Source iqure 9. Batch Unit for Variable F/M/Day Studies with Automatic Feeding Device 48 of 5:1 was used as substrate. The same proportions of nutrients and buffering salts were used as in the case of the fill and draw batch units (Table 2). Since the mixed liquor suspended solids concentration was designed to be 1000 rag/1 immediately after feeding, the substrate con- centrations in the three batch units after feeding were approximately 200 mg/1 COD in Unit No. 1, 400 mg/1 COD in Unit No. 2, and 800 rag/1 COD in Unit No. 3, Because some nondegradable COD remained in the unit at the time of feed- ing, a portion of this COD continued to remain after the feeding, hence, the actual 0-hour COD concentration was always a little higher than the valine used for experimental design. c. Feeding procedure A suitable volume of substrate was prepared and placed in the appropriate substrate reservoir. Computation of the substrate concentrations for the three units is pre- sented in Appendix B. The clock timer was set to actuate the solenoid controlling the valves on the automatic feeding device at specified intervals conforming to the feeding frequency o During intervals between feedings the solenoid was not energized and pinch valve "A" was kept open while pinch valve "B" was kept closed. During this period sub- strate was free to flow from the substrate reservoir to the 49 unit feeder. The volume in the unit feeder could be ad- justed to contain any volume up to 540 ml by means of a liquid level control tube. This tube was a small diameter capillary tube which acted as a liquid level control as well as a vent for the unit feeder. At the indicated time to feed the reactor, the timer device energized the solenoid valve which opened pinch valve "B" while simultaneously closing valve "A". The solenoid remained energized for a minimum period of 15 minutes. During this period the volume of substrate con- tained in the unit feeder proceeded to flow into the reactor and an equal volume of mixed reactor liquor was discharged through the overflow. The discharge of substrate into the reactor required approximately one minute and upon completion, the reactor unit contained theoretical con- centrations of substrate and biological solids in a ratio which corresponded to the F/M ratio for the unit. Studies by Siddiqi (24) on the hydraulic performance of similar units showed that experimental results were in very good agreement with those theoretically expected. During operation, the inner surfaces of the reactor units were kept free of bacterial growths by periodic scarping. Small amounts of water were added to offset losses due to evaporation, however, as the frequency of the feedings 50 increased, such losses became negligible . Before the batch units could be used for making experimental determinations, they were acclimated for a sufficiently long period to insure that equilibrium had been established . The unit was assumed to have reached a state of equilibrium when the mixed liquor suspended solids and filtrate COD concentrations were the same on two successive days In all cases, a minimum of 21 feedings was used as a criterion of thorough acclimation. Banerji (56) has shown that activated sludges are normally acclimated within a very short period, usually 6 feed cycles, even when there is a change in substrate. 3. Continuous flow studies Continuous flow studies were undertaken to see if correlation with the batch type results could be obtained in a continuous flow unit or chemostat. In such a unit, one particular loading rate ratio can be maintained for an in- definite period of time provided that the rate of substrate flow into the unit, the concentration of substrate, the temperature, and the metabolic characteristics of the culture remain constant (57) . a. Description of continuous flow unit Figure 10 shows the continuous flow unit employed in this study. Substrate was pumped at a constant rate from 51 Electrolytic Pump Waste Substrate Reservoir Pigure 10. Continuous Flow Activated Sludge unit 52 the substrate reservoir to the chemostat by means of the electrolytic pump (58). The chemostat had an inside diameter of 2.5 inches and a depth of 9 inches. The overflow was adjusted so that the proper volume in the reactor was con- tained. Air was supplied to the activated sludge culture after it was saturated with water vapor„ b. Substrate and nutrients A mixture of glucose and yeast extract in a ratio of 5:1 was used as the substrate. Nutrients and buffer salts were supplied in the same relative proportions to the :0D as shown in Table 2. c. Chemostat operation After choosing the desired mixed liquor suspended solids concentration at equilibrium in the chemostat, the :oncentration of COD in the substrate feed was determined >ased on 50 per cent synthesis. By setting the electrolytic lump to deliver a constant rate of flow, the desired P/M/day ation was obtained by selecting the proper volume in the hemostat. Computations illustrating this are presented in ppendix C. Prior to making experimental determinations, the nit was acclimated at a particular F/M/day loading for a inimum of 48 hours. 53 4. Limiting oxygen supply studies These studies were made using two different methods of mixing. In one set of experiments, an enclosed volume of air was used to provide a certain amount of oxygen to satisfy the chemical oxygen demand. The enclosed air was recirculated throughout the mixed liquor during a 12 hour period by means of a gas recirculating pump. In the other set of experiments, the batch units were stirred with teflon coated magnetic stirring bars rotated by external magnetic stirring devices . The oxygen in the air volume was able to enter the mixed liquor only through the liquid-air interface during a 12-hour period except that the liquid surface was not broken as in the case when the gas recirculating pump was used. a. Description of units mixed by recirculating pumps Four one-liter bottles, each containing 750 ml of mixed liquor were used as the batch reactors . One unit was used as the control and was not enclosed. Air was supplied to this unit from the air supply line. The other three units were arranged so that they contained the proper air volumes and air could be continuously passed through the mixed liquor by the recirculating pump. Figure 11 shows a typical arrangement. During the study, the control, Unit No. 1, was 54 Reactor CO Absorption Vial Reserve Oxygen Supply Gas Recirculating Purop Pressure Equalizer Figure 11. Apparatus Used for Limiting Oxygen Supply Study --Mixing by Recirculating Gas Pump 55 furnished all the oxygen it required to satisfy the oxygen demand of the mixed liquor. Unit No. 2 was furnished 40 per cent of its oxygen requirement, Unit No. 3 was furnished 20 per cent, and Unit No, 4 was furnished 10 per cent. Compu- tations for this study are presented in Appendix D. b. Description of units mixed by stirring bars As in the experiment described above, 4 batch units of the sarae size and having the same enclosed volumes of air to supply given amounts of oxygen to satisfy the oxygen demand were used Q The apparatus looked much the same as that shown in Figure 11 except that a gas circu- lation pump was not employed. Mixing was done by magnetic stirrers. The same computations in Appendix D for the previous experiment apply here also. Co Substrate and nutrients A mixture of glucose and yeast extract in a ratio of 5?1 was used as the substrate. Nutrients and buffer in the same relative proportion to the substrate were provided as shown in Table 2, d„ Feeding procedure After 11 hours of oxygenation by one of the mixing forms described in "a" or "b" above, one-third of the contents in each reactor or 250 ml was drawn off and wasted, rhe units were then allowed to settle for one hour after 56 which 250 ml of supernatant was drawn off. The required amount of substrate and nutrients was then added and the volume of each unit was brought up to 750 ml with tap water. Fresh potassium hydroxide, 12 M, was placed in the carbon dioxide absorption vial and oxygenation was resumed by the method used for the experiment. 5. Studies of flocculation without chemicals a. Source of flocculating sludge The source of the flocculating activated sludge required in certain portions of this study is described in section C.l.a. b. Source of nonf locculating sludge Nonflocculating activated sludge was obtained using a fill and draw unit similar to that described for the growth of flocculating activated sludge. The substrate and nutrients were also similar, however, the feeding procedure was slightly different. The nonflocculating activated sludge fill and draw unit was fed once a day. After a 23-hour aeration period, the unit was allowed to settle for one hour without any wasting of the mixed liquor. Following the settling period, 200 ml of the turbid super- natant liquor was drawn off to be used as seed for a new batch of culture. The remainder of the mixed liquor and any settled solids was wasted. Substrate, nutrients, and 57 buffering salts were added to the seed organisms and the unit was made up to 2 liters before aeration was resumed. After approximately 5 days, a nonf locculating culture was developed which showed poor flocculation and clarification properties at the end of the 23-hour growth cycle. 6. Chemical flocculation studies a. Source of nonf locculating sludge Since the volume of nonf locculating activated sludge required in these studies was quite large, a large well aerated plexiglass tank having a capacity of 36 liters was used to grow the required amount of culture. The seed organisms were obtained from the fill and draw unit described for the growth of nonf locculating activated sludge, The same concentrations of substrate and nutrients were also used,, The progress of the culture's growth was followed by taking optical density readings of the culture solution. When maximum growth was reached as indicated by the optical density, the culture was used for flocculation studies „ b. Chemicals used The chemicals used for the flocculation studies and other experiments weres alum (as ,Al2(SC>4) 3 • I8H2O) , calcium hydroxide, calcium carbonate, calcium chloride, and sodium hydroxide. Calcium hydroxide was used as a slurry and also in soluble form. The soluble calcium hydroxide 58 was prepared by centrifuging a solution containing an excess of the chemical. The strength of the soluble calcium hy- droxide solution was determined by titration with a standard sulfuric acid solution. c. General procedure used for chemical flocculation After determining the range of concentrations for the chemical to be used in the flocculation study, a standard solution of the chemical was prepared and the volumes of chemical solution and distilled water which was added to make all chemical additions of equal volume, were calculated. All chemical additions were based on 500 ml volumes which in- cluded the volume of the culture and the volume of the chemical addition „ All flocculation experiments were per- formed in 800 ml pyrex beakers . The flocculation experiment was initiated by pour- ing the chemical dosage previously prepared ancj placed in a test tube into its respective beaker. Each beaker contained the same volume of nonf locculating activated sludge suspension. The chemical was quickly stirred in the culture for about 10 seconds. The beaker contents were then flocculated slowly for 20 minutes by means of a stirring machine used for jar tests. After the flocculation period, the treated samples were allowed to settle for one hour and the supernatant was tested for clarity and pH. 59 7» Bulking sludge studies a. Description of batch units These studies were conducted in fill and draw batch units which were fed once a day. The units were the same as those described in section C,l„a. b. Substrate and nutrients The substrate used in these studies was Metrecal, a liquid dietary manufactured by the Mead Johnson Company. According to Banerji (56), Metrecal has a COD/N ratio of 24 which would be suitable for bacterial metabolic requirements (40). However, since nearly all of the nitrogen in Metrecal is probably organic, additional nitrogen was supplied in the form of ammonium ion to satisfy 75 per cent of a COD/N ratio of 15 o No other nutrients or buffer compound was used . Co Source of bulking sludge The bulking sludge for the study was supplied from fill and draw batch units which were fed Metrecal equivalent to 2 grams COD/liter/day. The units were wasted daily of 50 per cent of the mixed liquor . The initial seed organisms were obtained from batch units which had good settling ac- tivated sludge. After 5 days of feeding with the Metrecal substrate, the SVI increased from 106 to 640. At this point the sludge from all the bulking units was mixed and the experiments for the control of bulking sludge were begun. 60 The SVI for the batch unit used as the control continued to increase and the culture in this unit was almost entirely nonflocculating after 4 days. Operation of the control unit was therefore discontinued because of its permanent non- flocculating character o d. Feeding and study procedure Table 3 shows a list of the fill and draw batch units used in this study with their respective treatments for controlling the bulking activated sludge. The per cent wasting of mixed liquor for each unit is also indicated. The procedure used in feeding and treating the units was as follows. Following a 23-hour aeration period, the reactor unit volume was brought up to 2 liters by adding distilled water to replace any water lost by evapo- ration. The inner surfaces of each unit were carefully scraped to resuspend any adhering solids and wasting was performed as required. Aliquots of the wasted liquor were used to determine the mixed liquor suspended solids con- centration, filtrate COD, pH, and SVI for each unit. Peri- odic microscopic examination of sludge in each unit was also made. After a one-hour settling period, supernatant liquor was removed to a depth of one inch above the sludge- supernatant interface. All units, except Unit No. 6, were then fed Metrecal equivalent to 2 g COD/liter and Nitrogen 61 TABLE 3 BATCH UNIT TREATMENTS FOR BULKING SLUDGE CONTROL STUDY Unit No. Description of Treatment Mixed Liquor Wasting per cent 1 50 mg/1 bentonite* 2 50 rag/1 bentonite, 100 mg/1 alum** 3 50 mg/1 bentonite 50 4 100 mg/1 alum 50 5 50 mg/1 bentonite, 100 mg/1 alum 50 6 Chlorine*** 50 *No. 90 Volclay Zoo mesh. **100 rag/1 CaC0 3 alkalinity in the form of NaHC03 was provided with all alum treatments to provide for aluminum hydroxide formation. ***Chlorine dosage was determined from modification of Tapleshay formula (47). 62 equivalent to 125 mg/liter plus their respective treatment for controlling bulking sludge. The sludge in Unit No. 6 was treated with chlorine for bulking control. Chlorine additions were calculated using a modification of the formula given by Tapleshay (47). The modified formula used in this study was the following expressions -5 SVI x w x 10 = C where SVI ■ sludge volume index W = mg of solids being treated with chlorine C = mg Chlorine to be added per feeding Normally one or two mg of chlorine were required per day to control sludge bulking . The chlorine in solution form* was added just prior to the feeding of the units to allow for a 5-minute contact time Following the contact period at pH 7*0, Unit No, 6 was fed the Metrecal substrate and nitrogen supplement like the other units. All 6 batch units were then made up to 2 liters with tap water and aeration was resumed . ♦Prepared from a commercial bleach (Chlorox) 63 V. RESULTS AND DISCUSSION The results of the investigation are presented in graphical or tabular form wherever possible. In order to facilitate the presentation and explanation of the experi- mental data, a discussion of the results is included with each phase of study. A. Variable Loading Rate Studies Using Batch Units These studies were conducted on well acclimated activated sludges „ The substrate used was a mixture of glucose and yeast extract in a ratio of 5:1 as COD. Food to microorganism ratios were obtained by adding the required amount of substrate, in terms of COD, to a given weight of microorganisms as determined by the membrane filter test (49) . Inorganic salts were added with the substrate to satisfy mineral requirements and to provide suitable buffer- ing capacity. The relative concentrations of the inorganic salts with respect to the substrate concentration in the reactor unit after feeding are shown in Table 2. The loading rates of the batch units, as indi- cated by the F/M/day ratios, were varied by varying the number of feedings of substrate per day. Since three batch units were used in these studies, each having a different F/M/feeding ratio, a wide spectrum of loading rates was ob- 64 tained by increasing the number of feedings per day from 3 to 24. A summary of the F/M/feeding ratios and the F/M/day loading ranges is shown in Table 4. Six F/M/day loading rates were studied for each batch unit. These loading rates were obtained for feedings made at 8-, 6- , 4-, 3-, 2-, and 1-hour intervals. The basic experimental data from these studies is presented in Figures 12 through 20. Figures 12 through 16 show the changes in the biological solids and in the substrate remaining during the respective growth cycle times. Figures 17 through 20 show the oxygen uptake characteristics for the activated sludges during the growth cycle periods indicated in Figures 12 through 16. In addition to the parameters mentioned above, data were secured regarding the settleability of the sludge and the quality of the settled supernatant at the end of the feed cycle. This information is presented in Table 5. 1. Effect of F/M/day loading on substrate removal Figure 21 shows that substrate or COD removal during contact time, at a given F/M/feeding, is independent of a time-mass product (S a -t). Mass, S a , is the average con- centration of biological solids in contact with the substrate during the contact time (t). The contact time is obtained from the COD remaining parameter shown in Figures 12 through 65 TABLE 4 SUMMARY OF F/M/FEEDING RATIOS AND F/M/DAY LOADING RANGES FOR BATCH UNITS IN VARIABLE LOADING RATE STUDIES F/M/day Batch F/M/feeding Range Unit gm COD per g ra COD/ No « gm solids gm solids/ day 1 0.2 0.6 to 4.8* 2 0.4 1.2 to 9.6 3 0.8 2.4 to 19.2 ♦Values are for increase in feedings from 3 to 24 per day. a 1400 - 1200 1000 800 600 400 200 Time, hours Figure 12. Changes in System Parameters During 8 Hour Feeding Cycle 67 u *-> •H en g (V 4J (C u •H C c •H •P c E U ■P Q) Q 1400 1200 1000 800 - 600 400 _ 200 Time, hours 13. Changes in System Parameters During 6 Hour Feeding Cycle 68 u ■p •H H \ en E 73 •H H \ E V QJ fl) •H "0 C H C •H ■P id c E V ■P 0) Q 1400 1200 1000 *o Unit 3 MLSS, F/M/Day 6.01 Unit 2 MLSS, F/M/Day 3.96 cr~o""^ Unit 1 MLSS. F/M/Day 2.18 800 600 400 200 Unit 3 COD Remaining Unit 2 COD Remaining Unit 1 COD Remaining — ID Time, Hours 8 igure 15. Changes in System Parameters During 3 Hour Feeding Cycle 70 u 4J E E C 0) x 200 1 2 3 4 5 6 150 100 50 Unit 1. 2, 3, 1. 2, 3, Endogenous, MLSS F/M 0.2, " 0.4, " 0.8, 990 rag/liter 1110 1110 990 1110 1110 Time, hours Figure 17. Oxygen Uptake by Washed Suspensions of Activated Sludge — 8 Hour Feeding Cycle 72 250 - 200 - 150 - 100 _ 50 - 1 1 ! 1 1. Unit 1, Endogenous, MLSS 1030 mg/liter 2. ■I 2 " " 1000 3. 3 " 910 4. " 1, F/M 0.2, " 1030 5. " 2, " 0.4, " 1000 6. " 3, " 0.8, 910 — y?6 — — _^ 5 m — ^-O^- 04 — lifc^ — ■^.~- ' = s- : ^^- , "~~"~"'~ 3 1 | 4 6 Time, hours 8 Figure 18. Oxygen Uptake by Washed Suspensions of Activated Sludge--6 Hour Feeding Cycle 73 250 200 150 100 50 1. Unit \ \ 1, Endogenous, MLSS 1 I ■ 985 mg/liter —2. 2 " 950 " - 3. 3 " 1080 4. 1. F/M 0.2, 985 5. 2, P/M 0.4, 950 6. 3, F/M 0.8, 1080 76 — — — / J3 5 — — / ^04 — utr^^i. jr^*^ ~ 3 1 1 4 6 Time, hours 8 Figure 19. Oxygen Uptake by Washed Suspensions of Activated Sludge-~4 Hour Feeding Cycle 74 300 •50 200 150 100 50 - T 3 4 5 6 7 8 9 10, 11, 2, 13, 14. 15. 16. 17. -18. Unit 1. Endogenous, 2. M 3 3 e . 1. F/M 0.2, 2 " 0.4, 3. " 0.8, L, Endogenous, 2, t 3, M 1, F/M 0.2, 2, " 0.4, 3, " 0.8, 1, Endogenous, 2, # 3, 1, F/M 0.2, 2, " 0,4. 3, 0.8, 945 mg/liter 910 1260 945 910 1260 940 1000 1095 940 1000 1095 1040 980 1050 1040 980 1050 3 12 Time, hours Figure 20. Oxygen Uptake by Washed Suspensions of Activated Sludge--3 Hour, 2 Hour, and 1 Hour Feeding Cycles 75 TABLE 5 SLUDGE VOLUME INDEX AND OPTICAL DENSITY VALUES FOR VARIABLE LOADING RATE STUDIES Feed Cycle Time Hours Batch Unit No. Actual F/M/day Loading SVI Optical Density 8 1 2 3 0.81 1.57 2.54 33 29 29 0.035 0,03 0.042 6 1 2 3 1.13 1.84 3.51 30 33 31 0.02 0.04 0.045 4 1 2 3 1.61 2.69 4.85 28 40 0.016 0.022 0.26 3 1 2 3 2.18 3.96 6.01 31 37 0.015 0.095 0.54 2 1 2 3 3.10 5.31 9.15 27 0.09 0.28 0.57 1 1 2 6.35 21.7 27 0.128 0.205 3 """ *"* 0.375 800 - 600 400 200 76 s Unit 3, F/M/Feeding 0.8 6 «— 3 4 - 8 Unit 2, F/M/Feeding 0.4 V&8 * 2 "D-D" □1 Unit L, F/M/Feeding 0.2 8 4 2 J~ 3 1 Q £p o_Q_ Numeral denotes length of feeding cycle in hours I I J 0.5 1.0 1.5 S a t x 10 a 2.0 2.5 Figure 21. Effect of Contact Time and Biological Solids Concentration on Total COD Removal at Various Loading Rates 77 16. The contact time extends from 0-hour, when the unit is fed, to the time when the COD remaining becomes relatively constant. The remainder of the feed cycle time may be con- sidered to be the stabilization time. The average mass of solids under aeration during the contact time was obtained from the biological solids parameter. Table 6 shows the average solids concentrations under aeration during contact periods for the three batch units and all of the feed cycle periods studied. 2. Effect of F/M/day loading on sludge settleability and effluent quality Figure 22 shows the effect of F/M/day loading on SVI and effluent quality. The sludge volume index remains almost constant until an F/M/day loading of 3.5 to 4.0 gra COD/gm solids/day is reached. At this point, further em- ployment of the SVI as a useful parameter was abandoned be- cause of the increased amounts of suspended solids in the supernatant liquor. Although a blanket of sludge was dis- cernible in the cylinder after a 100-ral volume of mixed liquor had been allowed to settle for 30 minutes, the in- creased turbidity, reflected by the high optical density readings, indicated that dispersed growth was partially present. When the optical density of the settled super- natant exceeded 0.15, the SVI was not computed. As Figure TABLE 6 AVERAGE SOLIDS CONCENTRATIONS DURING CONTACT PERIODS IN VARIABLE F/M/DAY LOADING STUDIES 78 Field Cycle Time, Hours No. of Feedings per Day Solids Unit 2 Concentration Unit Unit 2 3 8 3 860 935 1190 6 4 950 1100 1210 4 6 1000 1155 1265 3 8 930 1020 1355 2 12 1155 1240 1400 1 24 1055 935 T ™ 79 300 200 100 200 100 V V Optical Density Effluent COD- 4-2 O ci O 02- r-1-^ 6 * 2 - oi-i 0.4 0.3 0.2 •H W C a 10 u •H a o 0.1 2-3 6-10 V^ -^ Do8-3* — Denotes 8 Hour Feed Cycle for ff-2 L 3-l Un it 3 °3-l Figure 22 2 4 6 8 F/M/Day Loading, gm COD/gro solids/day 10 Effect of F/M/Day Loading on SVI and Effluent Quality of Laboratory Activated Sludge Units 80 8 shows, an optical density of 0„15 indicates that the sus- pended solids concentration of the liquid being examined is approximately 450 mg/liter. According to Figure 22, the maximum substrate loading which will assure a desirable SVI is 4 gm COD/gm solids/day. The optical density, unlike the SVI, gradually increases with the loading, even at quite low F/M/day load- ings, and surpasses a value of 0.1 at a loading of 4 gm COD/gm solids/day. Figure 8 shows that an optical density of 0.1 indicates that the suspended solids concentration is about 300 mg/1. This is a rather high suspended solids concentration and could not be tolerated in a conventional activated sludge plant . Laboratory activated sludge systems often do not produce effluents of good quality because of the lightness of the floe which results from using pure compounds like glucose, and also because of the tendency of the system to produce pin-point floe when the F/M/day load- ings are low (17). Excessive aeration or high turbulence in laboratory activated sludge units can also produce pin- point floe and increased suspended solids in the effluent (59). The limit for F/M/day loading is much better re- lated to the optical density of the settled effluent liquor than to the SVI because the latter remains almost constant 81 and shows no significant change until the critical loading is reached. The optical density, on the other hand, gradu- ally increases with the loading and therefore is capable of indicating that the point of failure is being approached. The effluent COD increases at a uniform rate as the F/M/day loading is increased, however, there does not appear to be any unusual increase in soluble COD in the effluent. As Figure 21 shows, the COD removal in all units was good regardless of the F/M/day loading. Approximately 95 per cent of the COD was removed. 3. Effect of F/M/day loading on oxygen uptake Figure 23 shows the relationship between oxygen uptake, expressed as per cent of the applied COD, during both the contact period and the total feed cycle period to the F/M/day loading. The graph shows that as the F/M/day loading is increased, the oxygen consumption, in per cent of the applied COD, during the contact period increases from 6 per cent at an F/M/day of 0.6 to about 17 per cent when the F/M/day loading is 5.0. The oxygen uptake over the whole feed cycle period, on the other hand, decreases from a value of 25 per cent of the applied COD at an F/M/day loading of 1.0, to about 17 per cent at a loading of 6.0. From the point of junction of the two curves shown in Figure 23, both curves continue downward as the F/M/day loading increases. 82 30 25 v u & u ■p c u u ^•§20 u U 0) fiki V ■HH H au i:) o a; 0) (0 o a c C 10 cr >i X ■p c i •H U rH a tj a q) < U o « 'D v c a a D c 0) o 4J c a> o n 04 25 20 Oxygen Uptake for Feed Cycle Time 15 01-2 0\o2-l 2-2 10 1-1 .0 4-3 3-1 ^-3 Oxygen Uptake for Contact Period 8-3 \-2 ^_JL£i 8-2 4- •1 6-3 - 1 821 /. Denotes 4 Hour Feed Cycle Time — Unit 1 2 4 6 Stabilization Time, hours 8 Figure 24. Relationship Between Stabilization Time and Oxygen Uptake for Contact Period and for Feed Cycle 85 (0 •D •H fH w E \ Q O u E •H 04 o « •p c u •p c •H I o •0 > E & § 1 I -1 J78 600 5*6 500 2 //— Unit 3, ''F/M/Feeding 0.8 — 400 %3 yv4 6 v/ — 300 2 / 4 ^-Unit 2, F/M/Feeding 0.4 — 200 fl 2o_2i^ 4 -^ — ^06 - — Unit 1, F/M/Feeding 0.2 ~— 1 f\ ^"\ ^ — Numeral denot :es length of feed 100 cycle pe riod 1 1 1 2 4 6 Stabilization Time, hours 8 Figure 25. Effect of Stabilization Time and Loading Rate on 30-Minute COD Removal Capacity 86 on the capacity of the activated sludge to remove COD in a short contact period of 30 minutes at given F/M/feeding ratios. The 30-minute contact period was chosen for com- parative purposes because it is within the full contact period of all the feed cycle times studied. Also, samples taken for study at this time do not show the wide fluctu- ations in analytical results that samples taken sooner often show because of incomplete mixing in the reactor unit . Figure 25 shows that when activated sludges have the same stabilization time, the system acclimated to the highest F/M/feeding will have the greatest capacity to take up substrate. Furthermore, the COD uptake capacity in- creases at a greater rate for the systems acclimated to a higher F/M/feeding as stabilization time is increased. For example, when the stabilization time of 2 hours is employed, a system loaded with an F/M/feeding of 0.2 will take up 180 mg COD/gm solids in a 30-minute contact period, and a system Loaded at an F/M/feeding of 0.8 will take up about 340 mg/gm. However, if the stabilization time is increased to 6 hours, bhe system loaded at an F/M/feeding of 0.2 will increase its -0D removal capacity to about 215 mg/gm which represents a let increase of 35 mg/gm or 16.3 per cent. The system loaded *ith an F/M/feeding of 0.8 will increase its COD removal :apacity to 625 mg/gm which represents a net increase of 285 87 mg/gm or about 84 per cent. For a given stabilization time the MLSS are maintained in a more active state when the loading rate is higher. Figure 26 shows the effect of loading rate on the capacity of the activated sludge to remove COD in a 30- minute contact period . This relationship is presented in order to better explain Figure 25. Figure 26 shows that as the F/M/day loading is increased in a unit operated at a given F/M/feeding, the 30-minute COD removal capacity of the activated sludge decreases. The amount of the decrease de- pends upon the F/M/feeding of the unit; proportionately greater decreases appear to be demonstrated at the higher F/N/feeding ratios. More importantly, Figure 26 shows that the 30-rctinute removal capacity of the activated sludges is linear for a given feed cycle time as the F/M/feeding changes. Since the contact periods during the study were of the order of one hour, the feed cycle times may be con- sidered to be closely related to the stabilization times „ Figure 26 therefore shows that a more active sludge is main- tained when the food is added in larger increments, because of increasing synthesis at high F/M ratios, than when the food is added in smaller increments. In the latter case the organisms are maintained at a more or less stationary point on the growth curve. Thus completely mixed or extended 88 600 — 1 1 1 1 / Unit 3, ^«/^— F/M/Feeding 0.8 500 — 8 J 6/ 4 / 3 / >v / — 400 /s — 2 Hour Feed A* Cycle Time 300 ^ — Unit 2, N. F/M/Feeding 0.4 200 100 O fnit 1. F/M/Feeding 0.2 2 4 6 8 F/M/Day Loading, gm COD/gm solids/day 10 Figure 26. Effect of Loading Rate on 30-Minute COD Removal 89 aeration plants probably have a much lower proportion of active MLSS than contact stabilization plants. While the stabilization time or the feed cycle time was shown to be of significance in determining the 30- minute contact removal capacity of the solids as shown in Figures 25 and 26, Figure 27 shows that stabilization time is not necessarily a critical factor since essentially all of the soluble COD is removed at each loading rate and feed- ing frequency o For example , when the activated sludge system is operated at an F/M/feeding of 0.2, the COD re- moval capacity of the sludge remains constant at 200 rag/gm solids even though stabilization times become progressively shorter as the frequency of feeding, and hence the F/M/day loading, is increased. This indicates that the micro- organism population of the system adjusts itself to remove the substrate during the feed cycle time allotted. Lower COD removals/gm of solids were obtained during the contact time as the F/M/feeding was increased because the average concentration of solids under aeration during the contact period in the higher F/M/feeding system was greater. 5. Effect of stabilization time on sludge settle- ability and effluent quality The effect of stabilization time on SVI, the optical density, and the effluent COD is shown in Figure 28. 90 E \ § u E U •P IQ ■P C u n •P c •H u 3 Q •D > E a Q O u 600 500 400 01 300 2 D 200 - 01 100 V6 Unit 3, F/M/Feeding 0.8 8 Unit 2, F/M/Feeding 0.4 8 D4 D6 Unit 1, F/M/Feeding 0.2 -o- -Q. % 02 ^ — Numeral denotes length of feed cycle period 2 4 6 Stabilization Time, hours Figure 27. Effect of Stabilization Time on Total COD Removal at Different Loading Rates 91 300 - 200 - 100 - 100 - u 200 0) 4* •H H \ E Q O U c <1) 3 H «M u 2 4 6 Stabilization Time, hours Figure 28. Effect of Stabilization Time on SVI and Effluent Quality of Laboratory Activated Sludge Units 92 As the stabilization time is decreased, the SVI does not vary significantly until a stabilization time of about one hour is reached. When the stabilization time is less than one hour, the effluent turbidity becomes quite high indi- cating that dispersed growth is taking place. Because ex- cessive turbidity in the settled supernatant shows that an appreciable concentration of solids is still unsettled, the SVI value was not considered to be a satisfactory parameter at high turbidities. The effluent COD, as shown by Figure 28, corre- lates quite well with the results shown in Figure 22. Figure 28 shows that the effluent COD does not show any un- usual increase until a stabilization time of less than one hour is used. In Figure 22, the increase in the F/M/day loading is comparable to a decrease in the stabilization time, and the graph shows that no unusual increases in effluent COD were observed. There appears to be a more useful correlation be- tween optical density and F/M/day loading than between optical density and stabilization time as comparison of Figures 22 and 28 shows. In Figure 28 the optical density of the supernatant, like the SVI, remains quite constant until the critical stabilization time of about one hour is reached and gives no warning of impending upset. On the 93 other hand, the optical density appears to gradually in- crease as the F/M/day loading increases as shown in Figure 22 and therefore, the optical density appears to be a more sensitive parameter for indicating effluent quality as well as serving warning of oncoming upset conditions. In general, Figure 28 shows that proper per- formance of an activated sludge system seems to require a minimum of one hour for sludge stabilization and preferably two hours within the limits studied. 6. Effect of stabilization time on oxygen uptake Figure 29 shows that as the stabilization time decreases for a given F/M/feeding, the average net oxygen uptake/gm solids/hr increases. The average net oxygen up- take is obtained by deducting the endogenous respiration from the total oxygen uptake for the growth cycle time. Figure 29 also shows that for a given period of stabili- zation, the oxygen uptake rate increases for increasing F/M/feeding values. Figure 30, which has been plotted from the data in Figure 29, shows that the increase in oxygen uptake for a given stabilization time at increasing F/M/ feeding values is linear. This indicates that as activated sludge systems are loaded at increasingly higher F/M/feed- ing values, the microorganism populations shift from slower to faster rates of metabolism and from low to high oxygen 94 E E ro ■p a D C a) o> >* x o a) m M > < Numeral denotes length of feed cycle Unit 3, F/M/Peeding 0.8 — Unit 2, F/M/Feeding 0.4 F/M/Feeding 0.2 V6 2 4 6 Stabilization Time, hours 8 Figure 29. Effect of Stabilization Time and Loading Rate on Oxygen Consumption Rates of Activated Sludge Organisms 95 100 u w ■H .-I V) e CN E 0) nj ■p a D C 0) D> >, X o -p 0) z V CP ra n > < 80 _ 60 40 20 Stabi lization Time, Hours 0.2 0.4 F/M/Feedinq Ratio 0.8 Figure 30. Relationship of Stabilization Time to Net Oxygen Uptake Rate 96 requirements. The relationship between the oxygen uptake in the last 30 minutes of the feed cycle time and the stabilization time is shown in Figure 31. For a given F/M/feeding, the oxygen uptake gradually decreases as the stabilization time in the system increases. This illustrates that even though the uptake rate may appear to be relatively high in the closing phases of the feed cycle, the organisms are probably undergoing stabilization . For example, Figure 15 shows that for a 3-hour feeding cycle time all 3 batch units removed the applied substrate and that the sludges in the 3 units had at least one hour of stabilization time. The oxygen up- take curves for the sludges in the units are shown in Figure 20. These curves show that there is not much reduction in the oxygen uptake during the last hour of the feed cycle time except in the case of Unit No. 1. Thus, oxygen uptake curves do not necessarily show a break in the slope of the curve when the system is undergoing endogenous respiration. 7. Effect of F/M/day loading on oxygen uptake Figure 32 shows the effect of F/M/day loading on the net oxygen uptake rate. As the graph shows, a linear relationship exists between F/M/day loading and oxygen up- take. The oxygen uptake rates vary from a low of about 5 ag/gm/hr at an F/M/day of 1.0, to a rate of 50 mg/gm/hr at 9 7 0) K U •0 V V h Hi CO 4J 3 U £ to O -H ■P 10 (0 (0 E »J on \ C O ■H H CT 3 e Q a D C V en >i x o 100 Numeral denotes length of feed cycle period Unit 3, F/M/Feeding 0.8 Unit 2, F/M/Feeding 0.4 Unit 1, F/M/Feeding 0.2 ,6 Figure 31. 2 4 6 Stabilization Time, hours Relationship Between Oxygen Uptake During Last 30 Minutes of Feed Cycle and Stabilization Time at Different F/M/Feeding Ratios 100 Figure 32 2 4 6 8 F/M/Day Loading, gm COD/gra solids/day Effect of Loading Rate on Net Rate of Oxygen Uptake 10 99 an F/M/day of 9.0. The oxygen uptake rates of activated sludges from sewage treatment plants show considerable variability. Most data from conventionally loaded plants lie between 4 and 7 mg 2 /gm solids/hr at 20 to 25° C (19). These values are low when compared to high rate plants where oxygen uptake of activated sludge is normally between 15 and 25 mg/gm solids/hr at 25° C (19) . Figure 33 shows the effect of loading rate on oxygen uptake during the last 30 minutes of the feed cycle time. As in Figure 32, a linear relationship is observed between the F/M/day loading and the oxygen uptake. The curve shows that the endogenous respiration rate ranges from 6 mg/gm solids/hr at a F/M/day of 1.0, to 46 mg/gra solids/hr at a F/M/day of 9.0. Figure 34 is presented to show the relationship between the initial 30-rainute oxygen uptake and the termi- nal 30-minute oxygen uptake of the feed cycle time, in general the curve shows that the initial oxygen uptake is considerably greater than the oxygen uptake in the terminal phase of the feed cycle because the latter takes place during endogenous respiration. 8. Relationship of oxygen uptake to substrate removal Figure 35 shows the relationship of the net oxy- gen uptake rate to total COD removal over the contact period 100 100 o a) o e 0) h 2 4 6 8 F/M/I)ay Loading, gm COD/gro solids/day Figure 33. Effect of Loading Rate on Oxygen Uptake During the Last 30 Minutes of the Peed Cycle Time 101 100 60 100 20 40 Terminal 30-Minute Oxygen Uptake, mg 2 /gm solids/hr Figure 34. Relationship Between Initial 30-Minute Oxygen Uptake and Terminal 30-Minute Oxygen Uptake 102 600 500 400 300 - 200 IOC F/M/Feeding 0.8 2^ Unit 2, F/M/Feeding 0.4 — Unit 1, F/M/Feeding 0.2 Numeral denotes feed cycle time in hours 10 20 30 40 50 Average Net Oxygen Uptake Rate, mg 0_/gm solids /hr Figure 35. Relationship of COD Removal and Loading- Rate to Average Net Oxygen Consumption Rate of Activated Sludge Organisms 103 in an activated sludge unit. Basically the curves show that the metabolisms of the microorganism populations in the various systems do not change greatly even though the oxygen consumption rates may change appreciably. For example, the metabolism of the microorganisms in the unit having a F/M/ feeding o 0.2 did not affect the COD removal capacity sig- nificantly even though the oxygen uptake rate of the organ- isms increased from 5 to 16 mg/gm/hr as the F/M/day loading was increased. The transition of a population from one having a low oxygen uptake rate to one having a high rate appears to be smooth provided that the F/M/day loadings are increased gradually. This strongly suggests that a heterogeneous population gradually rids itself of those segments of its population which cannot compete for the food supply in the feed cycle time allotted. 9. Relationship of oxygen uptake to settleability and effluent quality Figure 36 shows the relationship of net oxygen uptake rates to settleability of activated sludge systems and effluent quality. As in previous cases, the SVI here also shows very little change and is a reliable indicator of sludge settling quality until the organisms attain a net oxygen uptake rate greater than 30 mg/gm/hr. At this 104 0.6 0.5 0.4 _ ■p •H W c A H ft •H -P a O 0.3 0.2 0.1 Figure 36. Optical Density 100 SVI 50 > 20 40 60 80 Average Net Oxygen Uptake, mg 2 /gm solids/hr Relationship of Oxygen Consumption Rates of Activated Sludge Organisms to SVI of Activated Sludge and Effluent Quality 105 and higher oxygen uptake rates, the turbidity of the settled mixed liquor becomes quite pronounced preventing the use of the SVI as a reliable parameter . As shown by the graph, the optical density gradually increases as the net oxygen uptake rate increases and becomes of inferior quality when the oxy- gen uptake rate of the sludge reaches 30 mg 2 /gm solids/hr. At this point the optical density of the supernatant has a value of 0.15c 10. Effect of F/M/day loading on cell replication in relation to sludge settleability The effect which F/M/day loading has on cell replication in relation to sludge settleability and optical density is shown in Figure 37. Three F/M/day loadings with an F/M/feeding of 0,5 were studied . The SVI and optical density values for the loading rates are presented in Table 7. Table 7 shows that for a F/M/day of 3.0, the activated sludge settled well and had a fairly good optical density. The values for the loading of 5.33 showed that the system had passed into an undesirable operational phase as indicated by the high optical density. The F/M/day loading of 4.0 appeared to produce an effluent of inter- mediate quality. The relationships indicated in Figure 37 show 106 E •a i X O 40 — 30 20 10 1. Endog., F/M/Day 2.0, MLSS 870 mg/liter F/M/Day 2.0, MLSS 870 mg/liter Endog., F/M/Day 3.5, MLSS 935 mg/liter F/M/Day 3.5, MLSS 935 mg/liter Endog.. F/M/Day 5.0, MLSS 945 mg/liter 2 3 4 '5 6 F/M/Day 5.0, MLSS 945 mg/liter 15 30 45 Time, minutes Figure 38. Oxygen Uptake by Unwashed Suspensions of Activated Sludge from a Continuous Feed Unit 113 TABLE 9 CHARACTERISTICS OF CONTINUOUSLY FED UNITS AT VARIOUS F/M/DAY RATIOS ■ A./* c„t Optical Total ^SS 0x yg en C0D % °2 F/M/day &vi Density Oxygen in Uptake in Uptake Uptake Flask Rate Flask is of mg/1 mg/1 mg/gra/hr mg/1 COD 2,0 38 o 06 15 870 17.3 67 22,4 3.5 75 0.17 26 935 27.8 117 22.2 5.0 ~ 0.27 34 945 36.0 167 20.4 114 was chosen to approximate the total delivery of COD to the chemostat over a one-hour period because continuous feeding within the Warburg flask could not be made. The total oxygen uptake for the one hour period was 15 mg/1 which is 22.4 per cent of the theoretically applied COD. The sludge volume index of 38 and the optical density of 0.06 denote a good settling sludge and an excellent effluent. For the P/M/day ratio of 3.5, the oxygen uptake follows an initial rate which is quite similar to that shown for an P/M/day of 2.0, however, this initial high rate con- tinues for a longer period and it is only after 30 minutes have elapsed that the slope of the uptake curve approximates that of the endogenous unit. The initial concentration of COD in the flask was 117 mg/1. This amount of COD was chosen to approximate the total delivery of COD to the chemostat over a one-hour period because continuous feeding within the Warburg flask could not be made. The total oxygen uptake for the one-hour period was 26 mg/1 which is 22.2 per cent of the applied COD. The sludge volume index of 75 and the optical density of 0.17 denote a borderline activated sludge system because the optical density is a little higher than desired to permit the use of the SVI as a parameter. For the F/M/day loading of 5.0, the oxygen uptake curve follows a high rate which is sustained for almost 30 115 minutes. Following this a slight reduction in the uptake occurs, however, the change in slope of the curve is not as well defined as it is in the cases of the two lower F/M/day loadings . The initial concentration of COD in the flask was 167 rag/1. This amount of COD was chosen to approximate the total delivery of COD to the cheraostat over a one-hour period because continuous feeding within the Warburg flask could not be made. The total oxygen uptake for the one hour period was 34 mg/1 which is 20.2 per cent of the applied COD. The SVI for the sludge was not calculated because the optical density of 0.27 was too high. These latter para- meters are typical of an activated sludge which settles poorly. The results obtained from the chemostat show them to be in agreement with those obtained from the batch fed units with respect to oxygen uptake, SVI, and optical density as shown in Figures 36 and 38. The chief difference between the good settling activated sludges and those which settle poorly or not at all is the difference in the oxygen uptake rates. At an F/M/day loading of 2.0 the oxygen up- take rate for the organisms was about 17 mg/gm/hr, The organisms from the unit fed at an F/M/day loading of 5.0, on the other hand, had an oxygen uptake rate of 36 mg/gm/hr. The batch studies showed net oxygen uptake rates of 12 and 116 28 mg/gm/hr which are in good agreement with the result from the continuous flow organisms when allowance is made for endogenous respiration to get the net uptake. The net oxygen uptake rates for the continuous flow unit organisms are 10 and 24 mg/gm/hr for F/M/day loadings of 2.0 and 5.0 respectively. Another characteristic of the systems which settle poorly appears to be the almost continuous high rate of oxygen uptake over the entire feed cycle. At low F/M/day loading rates, where good settleability of the sludge is a feature, the organisms demonstrate an initial high oxygen uptake rate which decreases abruptly to that of the endo- genous rate. The continuous feed studies also showed that oxygen uptake in terms of the applied COD was almost constant regardless of the F/M/day loading and ranged from 22.4 per cent for an F/M/day of 2.0 to 20.4 per cent for an F/M/day of 5.0. Values of 21 and 22 per cent were ob- tained for respective loadings in the batch studies as shown in Figure 24 . This indicates that the overall performance or efficiency of the unit is the same regardless of the F/M/day loading or the types of organisms in the equilibrium population of the system. 117 C. Limiting Oxygen Supply Studies In these studies, the amount of oxygen supplied to activated sludge systems was limited to 40, 20, and 10 per cent of the applied COD. These limits on oxygen supply were chosen because Warburg respirometer studies had shown that only about 20 per cent of the theoretical oxygen demand was utilized during substrate oxidation and stabilization of the biological solids. All units were fed at an F/M/day loading of 2.0; feedings were made at 12-hour intervals with an F/M/feeding of 1.0. Eleven hours of each feed cycle were devoted to mixing by spargers or magnetic stirrers and one hour was used to waste one-third of the mixed liquor, settle the mixed liquor before wasting one-third of the settled super- natant, and feeding. The studies were divided into two sets of experi- ments. In the first phase a calculated enclosed volume of oxygen, supplied as an appropriate volume of air, was con- tinuously recirculated through spargers in the mixed liquor, while in the second phase, mixing of the system contents was accomplished by means of magnetic stirring bars. 1. Limited oxygenation using spargers The results of this study are shown in Figures 39 118 2400 F/M/Day 2.0, 12 Hour Feed Cycle MLSS 0.3 2100 1800 1500 Optical Density 400 - 200 \ - 100 Figure 39. u Limiting Oxygen Supply Study. Unit 1, Control, 100 Per Cent of Oxygen Demand Allowed, Mixing by Sparger 119 through 42. The control, Unit 1, (Figure 39) which was supplied with all the oxygen that the organisms could use, maintained top performance throughout the six-day study period. The equilibrium solids level at the end of the feeding cycle was quite steady at about 2100 mg/1 . The excellent settling character of the sludge is indicated by the SVI values which ranged from a low of 18 to a high of 26. The optical density values for the settled supernatant indicated a high quality effluent as did also the filtrate COD values of the effluent liquor. The latter remained more or less constant at about 50 mg/1. Figures 40 through 42 illustrate the effects which are produced as a result of providing a limited oxy- gen supply to the activated sludge systems. Figure 40 shows the results obtained from Unit 2 which was supplied with 40 per cent of the theoretical oxygen demand. This unit performed well until Bay-4, then the pump which was used to recirculate the enclosed air volume began to draw a vacuum and necessarily terminated the study. However, the results up to the time of the pump failure are indica- tive of the performance of a unit supplied with sufficient oxygen as shown by the control, Unit 1 (Figure 39). The MLSS concentration remained at the equilibrium level. Ef- fluent filtrate COD values were similar to those obtained 120 2400 u 0) ■p •H -H N. D> E (0 c Q O U *J c 0) 3 H >w (4-1 u 2100 1800 - 1500 200 400 - Figure 40. Limiting Oxygen Supply Study. Unit 2, 40 Per Cent of Chemical Oxygen Demand Allowed, Mixing by Sparger 121 2400 -I 0.3 2100 1800 - 1500 400 200 F/M/Day 2.0, 12 Hour Feed Cycle MLSS Figure 41. Limiting Oxygen Supply Study. Unit 3, 20 Per Cent of Chemical Oxygen Demand Allowed, Mixing by Sparger 122 2400 2100 1800 - 1500 1200 900 - 600 Figure 42 Limiting Oxygen Supply Study. Unit 4. 10 Per Cent of Chemical Oxygen Demand Allowed, Mixing by Sparger 123 in the control as were also the SVI and the optical density values. In short, the unit was provided with sufficient oxygen supply to take care of substrate oxidation and stabilization of the biological solids. The results for Unit 3 which was furnished 20 per cent of the theoretical oxygen demand demonstrate that when a sufficient supply of oxygen is furnished as determined by Warburg studies, unit performance may decrease. As shown in Figure 41, the MLSS concentration of the system con- tinued at the normal equilibrium level until Day-2, after which an unexplainable abrupt drop in the solids concentra- tion took place. The loss in growth is reflected in the simultaneous increase in the effluent COD which increased suddenly from 45 mg/1 on Day-2 to 466 mg/1 on Day-4. The SVI values were consistently low ranging from about 20 to 30 and the optical density of the effluent liquor increased from 0.04 to 0.122 on Day-3. In all outward respects the unit appeared to function well and yet the fact that failure was imminent is certainly indicated by the high effluent COD and the loss in growth. It appears that the unit which was supplied with only 10 per cent of the theoretical oxygen demand (Figure 42) felt the effects of a deficient oxygen supply from the very start of the experiment. The biological solids level 124 progressively declined, decreasing from a value of about 2000 mg/1 on Day-0 to 660 rag/1 on Bay-5. The loss in growth is also indicated by the very high filtrate COD values in the effluent. As shown by the graph, Unit 4 could not maintain a system which would perform satisfacto- rily. In spite of the fact that the solids concentration kept decreasing steadily, the optical density of the super- natant showed a steady increase up to Day-3. This was probably the result of dispersed growth setting in because of the increasing F/M/day loading on the unit. 2. Limited oxygenation using stirring The results of these studies are presented in Figures 43 through 46. The studies were made to see whether activated sludge systems could be maintained on a limited supply of oxygen and to study the effects of the limitation on the performance of the systems. In general the results indicate that the method of stirring which was used to distribute the oxygen supply throughout the con- tents of the units was not satisfactory. The results for the control, Unit 1, are shown in Figure 43. This unit was supposed to receive all the oxygen it required to satisfy the oxygen demand after being fed. Unlike the ex- cellent performance of the control which was furnished air through a sparger, the control in this phase of the study 125 u 0) P •H H E (0 (0 C (0 § u p c 0) 3

Time, days Figure 45 Limiting Oxygen Supply Study. Unit 3, 20 Per Cent of Chemical Oxygen Demand Allowed, Mixing by Stirring 128 u •H H E S c 0) 9 2100 1800 1500 1200 900 600 300 F/M/Day 2.0, 12 Hour Feed Cycle Time, days igure 46. Limiting Oxygen Supply Study. Unit 4, 10 Per Cent of Chemical Oxygen Demand Allowed, Mixing by Stirring 129 showed progressive deterioration after only the first half day of operation. The results for the units which were supplied with 40, 20, and 10 per cent of the theoretical oxygen deraand are almost identical although the results obtained for Unit 4 (10 per cent of theoretical oxygen supplied) indicate that a faster rate of deterioration took place in this unit than in any of the others. These studies have shown that the conversion of a system from a flocculating phase to one which is nonfloccu- lating does not occur automatically with the introduction of a high F/M ratio. The reason for this apparent delay can be ascertained from other phases of study undertaken in the investigation. In some of the other experimental phases where high F/M ratios were employed and where the activated sludge underwent a change from a good settling sludge to one which was poor settling, sufficient oxygen was always being dispelled throughout the mixed liquor. Figure 32 shows that higher oxygen uptake rates are ob- served when the F/M/day loading is increased. However, if the oxygen concentration in the mixed liquor is very low and the method of oxygenation is so poor that a satis- factory concentration of oxygen cannot be maintained in the mixed liquor, then those organisms which have high 130 oxygen uptake rates will not be able to outgrow the organ- isms which have low uptake rates because there is not enough oxygen to provide the gradient that is necessary to transport the oxygen into the cell. Since organisms with high oxygen requirements are probably unable to thrive or multiply in a reduced oxygen environment, this may be the reason why the activated sludge cultures in the study did not immediately change from a good settling systems to nonf locculating systems. D. Electrophoretic Studies These studies were undertaken to establish whether electric charge on bacterial cells or activated sludge particles played any significant role in biof locculation phenomena. The studies were abandoned after inconclusive results were obtained. Mention is made of these studies to show that this line of endeavor was investigated. Measurement of particle mobility was made using Black's modification of the Moyer method (53). This method of determining particle mobility has been used by several investigators in studies involving colloidal particles with varying degrees of success. In general, investigations of this nature in the past have led to inconclusive if not contradictory results. The application of mobility measure- ment to the study of activated sludge particles appears to 131 be particularly limited and difficult because of the many arbitrary conditions that roust necessarily be imposed to assure standardization of the sample particles, in addi- tion, the instrumental problems which arose because of variation in particle size, gravitational effects, and variations caused by voltage changes proved too difficult to overcome. According to the hypothesis advanced by McKinney (31) lowering of the surface charge below 15 millivolts re- sulted in agglutination of some pure bacterial suspensions. McKinney 's results were obtained by utilizing bacterial suspensions directly without any standardized procedure (62). Thus the suspensions which McKinney worked with were high in salts and nutrients. Such solutions normally give very low mobility values and therefore have low zeta potentials as well. In the studies undertaken in this phase of the investigation, mobility measurements of activated sludge particles and organisms were found to produce zeta potentials exceeding 20 millivolts and in many cases were above 25 millivolts. Furthermore, no significant difference in particle mobility could be detected between sludge particles taken from good settling activated sludges operating at low P/M ratios and those taken from poor settling systems oper- 132 ating at high F/M ratios. E. Flocculation Studies Without the Use of Chemicals 1. Effect of starvation Nonflocculating activated sludge suspensions are very stable and similar to colloidal solutions in may re- spects . Figure 47 shows the effect of settling time on both flocculating and nonflocculating cultures. The floccu- lating culture showed a decrease in optical density from 0.27 to 0.015 after a 30-minute settling period. This amounted to a decrease in turbidity of 94.5 per cent (see Figure 8). The nonflocculating activated sludge, on the other hand, showed very little change in the optical density which dropped from 0.515 to 0.500 in the same 30- rainute interval. This amounted to a decrease in turbidity of only 3 per cent . Nonflocculating bacterial cultures will remain stable frequently for many hours, When they are starved, however, there is a reduction in optical density due to the loss of solids and the resulting residual mass shows some tendency to settle. Figure 47 shows the effect of settling time on the nonflocculating activated sludge mentioned above after it had been starved under constant aeration for a period of 2 weeks. The optical density of the starved culture declined from 0.165 to 0.07 in a 30-minute 133 0.6 Nonflocculating Culture a. ~Cr 0.4 0.3 - 0.2 0.1 Flocculating Culture Nonflocculating Culture After 2 Weeks Starvation 0.5 1.0 Settling Time, hours 1.5 igure 47. Effect of Settling Time on Supernatant Quality of Activated Sludge Cultures 134 settling period. This represents a decrease in settleable suspended matter of 57.5 per cent. 2. Effect of centrifugation and resuspension In this study the microbial cells of a nonfloccu- lating activated sludge culture were brought together by centrifugation to see if forced clumping or aggregation of cellular matter would promote f locculation. The results of the study are presented in Table 10. Four samples of a nonfloeculating activated sludge culture were centrlfuged at 12,000x g in a centrifuge for 15 minutes. A fifth sample which was not centrifuged was used as the control. This sample was designated as Sample io„ 1. After centrifugation, one sample designated as Sample No. 2 was resuspended immediately in its mother liquor. The supernatant of a third sample designated Sample No* 3 was discarded and an equal volume of 0.02 M phosphate buffer, pH 7.2, was used to resuspend the cells. Samples designated as Nos. 4 and 5 were allowed to stand in pellet form after removing mother liquor for 3 hours after which they were treated in a manner similar to Samples 2 and 3 respectively. Optical density readings of all samples were taken immediately following their re- suspension and also after they had been allowed to stand for a period of 2 hours. 135 TABLE 10 EFFLUENT CLARIFICATION CHARACTERISTICS OF NONFLOCCULATING ACTIVATED SLUDGE ORGANISMS Sa "J ple Treatment Optical Density No ° 0-hour 2-hour 1 Control, not centrifuged 0.515 0.485 2 Centrifuged and resuspended in mother liquor 0.510 0.485 3 Centrifuged and resuspended in o 02 M phosphate buffer 0.495 0.475 4 Centrifuged, allowed to stand 3 hours, then re- suspended in mother liquor 0.500 0.480 5 Centrifuged, allowed to stand 3 hours, then re- suspended in 0.02 M phosphate buffer 0.575 0.550 136 The results of the study show that the optical density in all of the samples underwent very little change after the 2-hour settling period. This indicates that in- duced clumping or aggregation of bacterial cells does not promote the permanent type of floe found in conventional activated sludge systems nor does it increase the settling ability of the organisms. It was noticed during washing that when good settling activated sludge is centrifuged and resuspended, the particles settle very quickly. 3. Effect of mixing a flocculating activated sludge with a nonf locculating sludge In this study, a good settling activated sludge was thoroughly mixed with a nonf locculating activated sludge by aeration for 20 minutes. The sludges had been aerated for 24 hours before they were used in the study. The concentration of each sludge was approximately 1000 mg/1 and the ratio of the mixture of the two sludges was 1:1. Following aeration, the mixture was allowed to settle for 30 minutes at which time the optical density of the supernatant liquor was observed. The results presented in Table 11 show that the supernatant of the mixed sludges failed to show any im- provement over that of the nonf locculating activated sludge itself. For some reason the nonf locculating activated 137 TABLE 11 RESULTS OF MIXING A FLOCCULATING SLUDGE WITH A NONFLOCCULATING SLUDGE Sample Condition Optical Density Supernatant of untreated non- flocculating sludge before and after 30 minutes settling 0.26 Sample 1 diluted in lsl ratio with distilled water before and after 30 minutes settling 0.13 Supernatant of good settling sludge after 30 minutes settling 0.02 Supernatant of lsl mixture of good settling sludge and non- flocculating sludge after 20 minutes aeration and 30 minutes settling 0.15 138 sludge suspension was so stable it effectively resisted any attempts to flocculate it. A possible explanation may be that the cells of the nonf locculating activated sludge are so highly hydrated with tightly bound water molecules that they act like lubricated spheres and slip among other cells and particles without adhering to them. F. Chemical Flocculation Studies The results of all the chemical flocculation studies are shown in Figure 48. Additions of chemical coagulants or salts are presented as milliequivalents of compound per liter (meq) rather than as milligrams per liter in order to compare the effects of the different ions. 1. Effect of alum on nonf locculating activated sludge As shown in Figure 48, aluminum sulfate is very effective as a coagulating agent. Excellent clarification was obtained applying a chemical dose of approximately one meq/1 or about 120 rag/1 alum. The turbidity of the super- natant for this dosage decreased approximately 90 per cent as the optical density dropped from 0.24 to 0.025. Very little additional clarification was obtained by increasing the alum dosage above one tr.ieq/1 . Tenney and Stumm (37) have observed that when bacterial cultures are in the acid pH-range (pH less than 0.30 0.25 0.20 O Ca(OH) 2 in slurry form • Ca(OH) 2 in soluble form □ CaCl. v NaOH A A1 2 (S0 4 ) 3 .18 H 2 Figure 48. 4 8 12 16 Concentration, meq/lxter Effect of Chemical Treatment on the F] occulation of Nonf locculating Activated Sludge Cultures 140 4) where the charge of the microbial surface is less nega- tive than in near neutral solutions and thus supposedly more amenable to coagulation, "free" trivalent aluminum ions at substantial concentrations far above the critical concentration necessary for the coagulation of lyophobic colloids, are not capable of coagulating microbial cells. They suggested that the aggregation produced by the aluminum salts is caused by the interaction of linear polymers, resulting from hydrolysis of the salt, with the dispersed cells. 2. Studies with calcium hydroxide The results obtained using calcium hydroxide in both the slurry and soluble forms to flocculate a non- flocculating activated sludge indicate that the saturated solution is far more effective in producing a clarified supernatant as shown in Figure 48. The presence of colloi- dal particles in the calcium hydroxide slurry does not appear to benefit the flocculation process by providing nuclei for floe particles. The high degree of clarifica- tion produced by the soluble calcium hydroxide would suggest that calcium ions and a high pll were responsible for the excellent results obtained. Application of 4 meq/1 of soluble calcium hydroxide reduced the optical density of the nonflocculating sludge from 0.24 to 0.02. This amounted 141 to a decrease in turbidity of almost 92 per cent. The same dosage using lime in slurry form reduced the optical density to 0.18 which amounted to a decrease in turbidity of only 25 per cent „ Comparison of the soluble calcium hydroxide re- sults with those obtained using alum show that alum is the more effective coagulant in conformance with the Schulze- Hardy rule. The general similarity of the two curves also suggests that the reaction mechanisms of the two coagulants are probably alike. Approximately 4 meq/1 of the soluble calcium hydroxide or about 150 mg/l were required to obtain the same degree of clarification as 1 meq/1 alum or about 120 mg/l alum. 3o Studies with calcium salts In attempting to ascertain what effects calcium carbonate and calcium chloride would have on the floccu- lation of nonf locculating activated sludge, the results shown in Figure 48 were obtained „ Since calcium carbonate is highly insoluble, any effect it might have on the flocculation process would be attributed to the development of floe particles using the colloidal calcium carbonate particles as nuclei. As the results show, application of calcium carbonate increased the turbidity of the suspension . This suggests that the 142 colloids in the calcium hydroxide slurry are of no bene- ficial effect either in the floeculation process and are probably detrimental. The results obtained using calcium chloride show that little improvement in clarification is obtained from this calcium salt also. Since calcium chloride is highly soluble, the results indicate that calcium ion by itself is not capable of causing floeculation. 4. Studies with sodium hydroxide The results obtained by treating nonf locculating activated sludge with sodium hydroxide (Figure 48) indi- cate that the presence of the hydroxide ion actually has a deleterious effect. A comparison of the three treatments employing hydroxide ion as the anion is quite interesting. At a concentration of 4 roeq/1 hydroxide ion, the soluble calcium hydroxide reduced the turbidity of the suspension almost 92 per cent, the calcium hydroxide slurry reduced the turbidity about 25 per cent, and the sodium hydroxide dosage increased the optical density about 17 per cent. Although these studies fail to provide a clear explanation underlying the cause of flocculation in a non- flocculating activated sludge culture by additions of calcium hydroxide, it appears as though both the calcium ion and the hydroxide ion are of importance. Excess 143 quantities of lime, however, introduce particulate matter which affects the turbidity of the supernatant. Go Bulking Sludge Control Studies ilfcs of these s are shown in Figures 49 through 54, 1 Effect of using bentonite Figure 49 shows the effect of adding 50 mg/1 bentonite daily to control a bulking sludge which had an SVI of 640 on Day-0. During an 11-day period, the mixed liquor suspended solids in a non-wasting system attained a maximum concentration of 6200 ng/1. The SVI of the system dropped from a value of 640 on Day-0 to 380 on Bay-1. Thereafter, the SVI declined steadily and reached a value of 150 on the last day of the study. As the graph indicates the MLSS concentration gradually increased thus causing the F/M/dlay loading to steadily decrease as well. The lowered F/M/day loading plus the bentonite weighting appears to be the reason why the SVI came under control gradually. Figure 51 shows the results of feeding a batch unit 50 mg bentonite/i/day along with the substrate „ How- ever in this case, the mixed liquor suspended solids was wasted each day. As in 1, the SVI came from a value of 640 under almost immediate control on Bay-0 to 115 on Day-1. Thereafter the SVI declined at 144 7000 — 6000 — u V •H r-i \ E 5000 — 4000 — 3000 — 2000 — 1000 Figure 49. 23-Hr Parameters for Unit 1 Fed 2.0 gm COD/ Liter/Day and 50 mg Bentonite/Liter/Day--0 Per Cent Wasting 145 7000 6000 5000 — u 0) v H E a 4000 — 3000 - 2000 - 1000 Figure 50. 23-Hr Parameters for Unit 2 Fed 2.0 gm COD/ Liter/Day, 50 mg Bentonite/Liter/Day. 100 mg Alum/Liter/Day--0 Per Cent Wasting 146 7000 Effluent COD 6000 200 £ cr E 100 5000 u to Figure 51. 2 3-Hr Parameters for Unit 3 Fed 2.0 gm COD/ Liter/Day and 50 rog Bentoni te/Liter/Day--50 Per Cent D^ily Wasting of MLSS J 47 7000 — 6000 — .000 u 48 7000 — 6000 5000 — u QJ ■H ■H \ CP E w 4000 3000 - 2000 1000 - 200 100 — 300 200 — 100 u \ E Q O O G 3 CO > Figure 53. 23-Kr Parameters for Unit 5 Fed 2 gm COD/Liter/ Day, 50 mg Bentonite/Li ter/Day. and 100 mg Alum/ Liter/Day- -50 Per Cent Daily Wasting of MLSS 149 7000 _ - 200 6000 — 5000 — u 100 Figure 54. 23-Hr Parameters for Unit 6 Fed 2.0 gm COD/ Liter/Day with Chlorine Added to Control Bulking by Tapleshay Method — 50 Per Cent Daily Wasting of MLSS 150 a low rate reaching an average value of 80 on the last day of the study. The curves show that the unit was kept under excellent control during the entire 11-day period. Thus bentonite additions alone were capable of reducing the SVI. 2. Effect of using alum with bentonite Figures 50 and 53 show the results obtained when 50 mg bentonite/1/day and 100 mg alum/1/day are added to the activated sludges in Units 2 and 5 respectively. The mixed liquor suspended solids in Unit 2 were not wasted during the study period and the solids concentration reached a value of approximately 6800 mg/1 . The higher ac- cumulation of solids over that shown in Figure 49 for Unit 1 is due to the accumulation of the alum in the floe particles. Figures 49 and 50 are similar in that the SVI values for both units were somewhat erratic over the study period. The effect of 50 per cent daily wasting is shown in Figure 53 . The MLSS concentration in Unit 5 was main- tained at a rather uniform level, however, a gradual in- crease in the SVI took place following the Day-6 de- termination as the SVI increased from 115 to 285 on Day-11. There does not appear to be any satisfactory explanation for the gradual increase in the SVI. Figure 52 shows the results obtained for Unit 4 which was fed 100 mg alum/1/day and was wasted at a rate of 151 50 per cent of the MLSS each day. The erratic behavior pattern of the SVI values suggests that alum is not as good as bentonite in keeping the SVI under control. A comparison of the results obtained from using bentonite and alum treat- ments to control bulking sludge shows that bentonite alone gave better results than alum alone or in combination with alum. 3. Effect of using chlorine to control bulking Figure 54 shows the results of using additions of chlorine by the Tapleshay method (47) to control sludge bulking. The results show that the SVI was kept under good control during the ll-day study period. On the average, one to two mg chlorine per gra of sludge solids being returned was used daily to control bulking. 152 VI. SUMMARY AND CONCLUSIONS The studies in this investigation have shown that F/M/day loading rates are associated with the flocculation of activated sludge systems as well as the clarification of the effluent. Maximum loadings on the order of 4.5 gm COD/ gm sludge solids/day in batch studies (approximately 10 times conventional activated sludge plant loadings) were used with- out causing a bulking sludge, however, increased F/M/day loadings were found to deteriorate the effluent quality of the system. A better correlation was found between F/M/day loading and supernatant clarity as measured by the optical density than between F/M/day loading and SVI. The SVI did not change in value as the substrate loading was increased until there was failure of the system. On the other hand, the supernatant clarity showed gradual deterioration as evidenced by increased optical density values as the loading increased. This indicates that conventional plant operation can be more reliably guided with respect to maintaining a good flocculating activated sludge system by using the optical density of the settled effluent as a control parameter rather than the SVI. The effect of stabilization time on the SVI of the sludge and clarification of the effluent was not found to be a significant factor as long as the stabilization time did 153 not decrease below 1 to 2 hours. When lower stabilization times were employed, the presence of dispersed growth was found almost immediately. Accordingly, longer stabilization times may be used to restore flocculating properties of an activated sludge. The onset of process deterioration was masked by the SVI parameter which remained almost constant throughout the range of good f lobulation, however, optical density readings of the settled supernatant liquor demon- strated greater sensitivity and a gradual increase in the optical density was observed as the F/M/day loading was in- creased . Oxygen uptake rates of the sludge organisms were also correlated with the breakdown of system performance. Uptake rates were found to increase almost linearly with the P/M/day loading, however, no unusual or sudden changes in the uptake rate with respect to loading could be observed during the transition of the activated sludge system from one which had good flocculation to one which was nonfloccu- lating. Increases in oxygen uptake rates were found to correlate better with increased F/M/feeding ratios and also with decreases in the stabilization time rather than with increases in F/M/day loadings. Thus, an activated sludge having a high rate of oxygen consumption could serve to warn that a good flocculating sludge was being transformed 154 into one which was nonf locculating., Continuous flow studies were found to yield approximately the same results as the batch units wi^h respect to F/M/day loading on the flocculation and clari- fication of activated sludge systems. Differences in the chemical structure of good flocculating and poor floccu- lating sludges could not be determined by infrared spectra analysis. The studies of limiting oxygen supply to activated sludge systems showed that such systems are capable of per- forming quite well when the amount of oxygen supplied is sufficient to meet approximately 20 to 40 per cent of the theoretical oxygen demand as COD. These results confirmed the oxygen uptake observed in Warburg studies. The failure of the system which occurred when an insufficient quantity of oxygen was furnished was gradual and the transition of the systems from good flocculating to nonf locculation was retarded, very probably because the oxygen tension in the mixed liquor was insufficient to provide the proper gradient for the transport of oxygen into the bacterial cells. Electrophoresis studies to establish the associ- ation of electric charge on bacterial cells with floccu- lation of the system were not successful. The results showed that there was little difference in the mobility 155 values of good flocculating and poor flocculating activated sludges. Thus, zeta potential does not appear to play any significant role in the flocculation of activated sludge systems. Attempts to promote the flocculation of non- flocculating activated sludges by centrif ugation and re- suspension showed that although such methods were not able to promote flocculation, they did indicate that the probable reason for the nonf Peculating character of the sludge was due to some special surface characteristic such as the presence of bound water . No evidence was obtained during these studies to support the theory that external bacterial polymers were present or contributed to flocculation. Chemical flocculation studies showed that trivalent aluminum ion is very effective in coagulating non- flocculating activated sludge . Divalent soluble calcium ion and pH values on the order of 8.3 were also found to be effective although 4 times as much lime was necessary to achieve the same degree of clarification as alum on an equivalent weight basis. Attempts to associate flocculation of the organisms with the calcium ion in a neutral salt form or with high pH by the addition of sodium hydroxide showed that the effect of flocculation could not be attributed to either alone. The studies also showed that additions of 156 colloidal particles, such as are present in calcium hy- droxide and calcium carbonate suspensions, to promote flocculation by providing nuclei for the floe particles were not effective. The effects of bentonite, alum, solids wasting, and chlorine on controlling the bulking of an activated sludge were studied . it was observed that all treatments gave good results. However, in the series of experiments in which bentonite and alum were used, bentonite alone was found to be more effective than alum alone or in combination with alum. Although bentonite probably has some weighting effect on floe particles, the reason for its effectiveness in controlling bulking is mora likely due to the increase in exchange capacity which it provides. The addition of chlorine to control bulking was shown to be both effective and flexible. These observations and findings indicate that several factors, i.e. exchange capacity, stabilization time, F/M/day loading, and oxygen supply, may be responsible for flocculation and clarification in conventional activated sludge treatment plants. However, for the most part, it has been substantiated that careful control of the F/M/day loading on activated sludge systems coupled with providing a satisfactory stabilization time of the settled sludge 157 solids will often be sufficient to assure good flocculation and clarification as well as economical operation. In view of the findings of this investigation, the following conclusions may be drawn s (i) The primary cause of nonf lobulation or sludge bulking involves a combination of high F/Vday loading and low stabilization time, (ii) Optical density measurement is a better parameter than SVI for the purpose of monitoring effluent quality and giving warning of impending upset of the process. It is more sensitive than the SVI and is a more rapid de- termination. Ciii) Long stabilization times are not of im- portance in the flocculation of activated sludge systems, however, a minimum of 2 hours stabilization time should be provided to insure good performance in batch fed systems. Civ) The differences in the flocculating and clarifying properties of good and poor flocculating acti- vated sludges is not related to a difference in chemical structure of the organisms as measured by infrared spectrum analysis or to electric charge on the sludge particles. (v) Successful treatment of a bulking sludge can be accomplished by using additions of bentonite or chlorine. 158 The latter provides a very flexible means of control. (vi) Particulate matter, such as is present in a lime slurry, does not promote flocculation of nonfloccu- latirag sludges „ (vli) The results of continuous flo« studies can be correlated with those from batch units. (viii) For optimum flocculation and clarification the maximum F/M/day ratio for substrate loading of batch fed systems is approximately 4.0 gm COD/g solids/day. (ix) The upper net oxygen uptake rate of good flocculating activated sludge organisms is approximately 30 mg 2 /gm solids/hour. 159 VII. SUGGESTIONS FOR FUTURE WORK Suggestions for future work which stem from this investigation and appear most promising are: (i) Studies of chlorine additions to highly loaded or nonflocculating activated sludge systems in order to improve flocculation and BOD removal. (ii) Continuous flow studies using bentonite and other forms of particulate matter to control activated sludge bulking at high loading rates. (iii) Flocculation studies employing other sub- trates and sewage to observe whether similar relationships are obtained. (iv) Flocculation studies employing various coagulants, salts, and polyelectrolytes to improve activated sludge flocculation and clarification . 160 APPENDIX A MATHEMATICAL RELATIONSHIPS OF COMPLETE MIXING AS APPLIED TO FEEDING OF BATCH UNITS (24) For any completely mixed system in which there is no concentration difference from one part of the system to the other, the material balance can be shown as: Input - Output = Accumulation (1) If the flow rate into the reactor is Q from a reservoir having a substrate concentration of C R . Equation (1) can be written as: <=r • o) - cc t . o) = d(Ct • v) (2) r dt l ' where V is the tank volume and C t is the concentration of the substrate in the tank at any time t. If the flow to the tank ceases after a period T, and initially at ft « Q, C t = 0, Equation (2) on integration and substitution of these limits yields: 9. . T c T -V l 5; ' l - e < 3 > where C T is the concentration of substrate in the tank at t = To Similarly for the displacement of sludge from the tank by the inflowing feed, Equation (1) can be rewritten as: <3(S_ . V) -S«.Q = dt (4) 161 where S t is the concentration of sludge in the tank at any time t. If the sludge concentration in the reactor when the feeding is initiated (t = 0) is equal to S Q and S T when the feeding ceases (t = T) , Equation (4) on integration and substitution of these limits yields, -Q . T Sip §f = * (5) Using Equations (3) and (5) the concentration of feed in the reservoir and the rate and period of inflow of the feed can be calculated for a given set of operating parameters . 162 APPENDIX B COMPUTATIONS FOR BATCH STUDIES Unit No. 1: Design F/M/feeding =0.2 F C T m = §j = °- 2 S T selected to be 1000 mg/1 in all units, Then C T = 1000 x 0.2 - 200 mg/1 COD s o = S T / 0.5 C T = 1000 / 0.5(200) - 1100 mg/1 Maximum volume of reactor is 2 liters "55 QT S. V ___ -2 SI ] T = e or ilOO s e S T 1000 QT = 0.191 liters C T ~% -0.0955 ~ =l-e V or 200 =l- e R C^ C R = 2220 mg/1 COD Concentration of Feed Solution = 2220 mg COD/1 v °l ume = 191 ml Unit No. 2% Design F/M/feeding =0.4 Concentration of Feed Solution = 2400 mg COD/1 Volume " » •■ = 365 ml Reactor volume used = 2 liters 163 Unit No. 3: Design F/M/feeding = 0.8 Concentration of Feed Solution = 2810 mg COD/1 Volume " " M ene . 505 ml Reactor volume used =1.5 liters 164 APPENDIX C P/M/DAY FOR CHEMOSTAT Design F/M = COD/M =2.0 Daily flow delivered by pump = 1800 ml/day (I = 0.08 amps) For concentration of organisms in chemostat of 1000 mg/1, COD of substrate should be 2000 mg COD/1 allow- ing for 50 per cent synthesis. Total amount of substrate COD fed to cheraostat/day, 1800 x 2 = 3600 mg COD Since CPJ* . 2.0, then M required is 3600 or 1800 mg and the n 2 required volume of the chemostat ■ 1800 = 1800 ml 1 165 APPENDIX D SAMPLE COMPUTATIONS FOR OXYGEN LIMITING STUDIES Reactor Volume of Unit 750 ml MLSS @ equilibrium 2000 mg/1 Mass of organisms in reactor = 2000 x .75 = 1500 mg If one-third of MLSS is wasted, then have left 2/3 x 1500 = 1000 mg For F/M or COD/M/feeding of 1.0, must add 1000 mg COD Oxygen volume to satisfy 40 per cent of COD added 1000 mg x 0.4 = 400 mg 2 If room temperature is approximately 23° C, then 1 mole 2 occupies 296 273 x 22.4 = 24.3 lit ers Since 1 mole of 2 = 32 gm, each ml of 2 contains 32,000 _ 24,300 = 1.32 mg 0^ Therefore require IPO = 75.8 ml 0, at 23° C to 1 . 32 ^ satisfy 40 per cent of theoretical COD Since atmosphere contains 20 per cent 2 then use ^~ - 380 ml of air 166 REFERENCES 1. Ullrich, A. H., and Smith, M. W., "The Biosorption Process of Sewage and Waste Treatment." Sewage and Industrial Wastes , 23, 1248 (1951). 2 * ^iv 1 ?:/" H I' and Smith ' M ' W " "Operation Experience with Activated Sludge Biosorption at Austin, Texas " Sewage and Industrial Wastes, 29, 400 (1957) . 3. Water Pollution Research 1964 , Ministry of Technology Her Majesty's Stationery Office, London (1965). 4. Haseltine, T. R., "Design of Activated Sludge Plants " Water and Sewage Works, 103, R-271 (1956) . 5. Ardern, E., and Lockett, W. T., "Experiments on the Oxidation of Sewage Without the Aid of Filters," Journal Society of Chemical Industry , 33, 523 (1914) . 6. Sawyer, C. N., "Milestones in the Development of the Activated Sludge Process." Journal Water Pollution Control Federation , 37, 151 (1965). 7. Buswell, A. M , and Long, H. L., "Microbiology and Theory of Activated Sludge." Journal American Water Works Association , 10, 309 (1929). 8. Grant, S., Hurwitz, E., and Mohlman, F. W., "The Oxygen Requirements of the Activated Sludge Process." Sewage Works Journal , 2, 228 (1930) . 9. Theriault, E. J e , and McNamee, P. D., "Sludge Aeration Experiments — Rate of Disappearance of Oxygen." Industrial and Engineering Chemistry , 22 , 1330 (1930) . 10. Kessler, L. H., Rohlich, G. A., and Smart, J., "Tapered Aeration of Activated Sludges." Municipal Sanitation , 7, 268 (1936). 11. Gould, R. H., "Operating Experiences in New York City." Sewage Works Journal , 14 , 70 (1942) . 12. Setter, L. R., "Modified Sewage Aeration — Part I." Sewage Works Journal . 15_, 629 (1943). 167 13. Setter, L. R., and Edwards, G. P., "Modified Sewage Aeration-Part II." Sewage Works Journal, 16, 278 (1944). 14. Wuhrmann, K. f "High-Rate Activate Sludge Treatment and Its Relation to Stream Sanitation." Sewage and Industrial Wastes , 26, 1 (1954). a_ gng xnaustnal 15. Heukelekian. H "Aeration of Soluble Organic Wastes with Nonflocculent Growths." Industrial and Engineering Chemistry, 41, 1412 (1949). £±Sa 16. Mccarty, P. L ., and Brodersen, c. P., "Theory of Extended Aeration Activated Sludge." Journal Water Pollution Control Federation , 34 , 1095 (1962). ~ 17. Stewart, M. J., "Activated Sludge Process Variations The Complete Spectrins Part III. Effluent Quality* Process Loading Relationships." Water and Sewage Works 111, R-260 (1964) . a —' L8. Lesperance, To w. , "A Generalized Approach to Activated Sludge. Part II, Developing the Process." Water Work s and Wastes Engineering . 2, 52 (1965). L9. Wuhrmann, K., "Factors Affecting Efficiency and Solids Production in the Activated Sludge Process." In W. W Eckenfelder and Brother Joseph McCabe, (ed.) f Biological Treatment of Sewage and Industrial Wastes , Volume 1/ Reinhold Publication Corporation, New York (1956) . !0. Eckenfelder, W. W , "Kinetic Relationships in the Biooxidation of Sewage and Industrial Wastes." Proceed - iUSS of the. 14th Industrial Waste Confe rence, Purdue UniVo, Engineering Extension Series , 104 , 495 (1959). 1. Eckenfelder, W. W., "Applications of Kinetics of Activated Sludge to Process Design." In W. w. Eckenfelder and Brother Joseph McCabe , ~ed .) , Advances in Biological Waste Treatment , Pergamon Press, New York (1963) . R. P., and Budd, W. E., "Effect of BOD Loadings on Activated Sludge Plant Operation." In W. W. Eckenfelder and Brother Joseph McCabe, (ed.), Biological Treatment of Sewage and Industrial Wastes , Volume I, Reinhold Publishing Corporation, New York (1956) . 168 23. Joint Committee of the Water Pollution Control Federation and the American Society of Civil Engineers, Sewage Treatment Plant .Design, Water Pollution Control Federation. Washington, D. c, (1959). 24 ° Prn™S R \ T H r " R ° le ° f EnZymSS ±n Contact Stabilization Process . Universi ty, of Illinois Sanitary Engineerin g series No _30, Urbana, Illinois (1965) . 25 . Orford, H. E., Heukelekian, H. f and Isenberg, E., "Effect ■e Loading and Dissolved Oxygen on the Performance of the Activated Sludge Process ." in W w. Eckenf elder and Brother Joseph McCabe, (ed.), Advances in Biological Waste Treatment . Pergamon Press, New York (1963K — 26. Theriault, E. J., "Studies of Sewage Treatment III The Clarification of Sewage. A Review." Sewage Works Journal , 7, 377 (1935). 27. Ruchhoft, C. C, Butterfield, C. T. # and McNamee, P d and Wattle, E., "Studies of Sewage Purification." Sewage Works Journal , 11 , 195 (1939) . 28. McKinney, R. E a , and Horwood, M. P., "Fundamental Approach to the Activated Sludge Process. I. Fioc- Producing Bacteria." Sewage and Industrial Wastes 24 117 (1952). ' — 29. McKinney, R. E., "Fundamental Approach to the Activated Sludge Process. II. a Proposed Theory of Floe Formation." Sewage and Industrial Wastes. 24, 280 (1952). 30. 31 van Gils, I B H. W., "Bacteriology of Activated Sludge " Research Institute for Public Health Engineering Report No. 32, The Hague, Hoi" ', Ro E„, "Biological Flocculation. " In w. W. Eckenfelder and Brother Joseph McCabe, (ed.), Biological Treatment of Sewage and Industrial Wastes ,. Volume I, Reinhold Publishing Corporation, New York (1956). 32. Northrop, J. H., and De Kruif, P. H., "The Stability of Bacterial Suspensions. II. The Agglutination of the Bacillus of Rabbit Septicemia and of Bacillus typhosus by Electrolytes." Journal of General Physiology 4 639 (1922). ~~ ' -' 169 33, Crabtree, K., McCoy, E., Boyle, w„ C. and Rohlich G A "Isolation, Identification, and Metabolic Role of the Sudanophilic Granules of Zoogloea rarciigera ." Applied Microbiology , 13, 218 (1965) . 34, Heukelekian, H , and Weisberg, E„, "Bound Water and Activated Sludge Bulking," Sewage and Industrial Wastes 28, 558 (1956). ~ — ' . _S_£2*. 35o Jirgensons, B., and Straumanis, M D E., A Short Textbook of Colloid Chemistry,, John Wiley and Sons, New York 36. Gould, R. Ho, "Sewage Aeration Practice in Mew York." American Society of Civil Engineers, Procee dings Separate No. 107 (1953) o ~ ~ 37. Tenney, M„ W„, and Stumra, W a , "Chemical Flocculation of Microorganisms in Biological Waste Treatment." Journal Water Pollution Control Federation , 37 , 1370 (1965) . 38 o SG»liwood, c„ "Adsorption and Assimilation in Activated Sludge " Proceedings American Society of Civil [ineers, 83, SA4, 1334 (1957). 39. Bias, F. F., and Bhat, J. V., "Microbial Ecology of Activated Sludge. 1. Dominant Bacteria." Applied Microbiology ,, 12 , 412 (1964). 40. Sawyer, C. N , "Bacterial Nutrition and Synthesis." In W. W. Eckenf elder and Brother Joseph McCabe, (ed.), Biological Treatment of Sewage and Industrial Wastes , Volume I, Reinhood Publication Corporation, New York (1956) . 41. Cruickshank, R., (ed.), Hand Book of Bacteriology , E. and S. Livingstone Ltd., Edinburgh and London (1960). 42. Haynes, W. C, Melvin, E, H., Locke, J. M., Glass, C. A., and Senti, F. R., "Certain Factors Affecting the Infrared Spectra of Selected Microorganisms." Applied Micro - biology , 6, 298 (1958). 43. Levine, S., Stevenson, H. J. R., Chambers, L. A., and Kenner, B„ A., ""Infrared Spectrophotometry of Enteric :eria." Journal of Bacteriology, 65, 10 (1953b). 170 44. Levine, S., Stevenson, H. J„ r., and Bordner, R. H "Identification of Glycogen in Whole Bacterial Cells by Infrared Spectrophotometry." Science , 118 , 141 (1953a). 45. Walters, C. P., "Biochemical Storage Phenomena in Activated Sludge," Unpublished Ph.D. Thesis, Universitv of Illinois (1966) . 46. Smith, R. S., and Purdy, W. c„, "The Use of Chlorine for the Correction of Bulking in the Activated Sludge Process." Sewage Works Journal , 8, 223 (1936). 47. Tapleshay, j„ A., "Control of Sludge Index by Chlorination of Return Sludge." Sewage Works Journal , 17, 1210 (1945). 48. Grune, W. N., "Sewage Chlorination in Review." Water and Sewage Works. 103, R-283 (1956) „ 49. Winneberger, J a H., Austin, J. H„, and Carol, A. K., "Membrane Filter Weight Determinations ." Journal Water Pollution Control Federation . 35 , 807 (1963) . 50. Standard Methods for the Exmination of Water and Sewage , 12th Edition, American Public Health Association (1965)'. 51. Urabreit, W. W., Burris, R. H., and Stauffer, J. F., Manometric Techniques , Burgess Publishing Company, Minneapolis (1964) . 52. Burton, K., "A Study of the Conditions and Mechanism of the Biphenyiamine Reaction for the Colorimetric Estimation of DNA." Journal of Bioch e mistry , 62, 315 (1956). 53. Black, A. P., and Smith, A. L , "Determination of the Mobility of Colloidal Particles by Microelectrophoresis." Journal American Water Works Association , 54 , 926 (1962) . 54. Ruchhoft, C. C, and Kachmar, J. P., "Studies of Sewage Purification." Sewage Works Journal , 13 , 3 (1941). 55. Rich, L. G., Unit Processes of Sanit ary Engineering , John Wiley and Sons, New York (1963). 56. Banerji, S. K., "Biological Removal of Colloidal Waste in Activated Sludge Process." Universitv of Illinois Sanitary Engineering Series No. 29, Urbana, Illinois (1965) . 171 57 . Herbert, D., Elsworth, R , and Telling, R. c., "The Continuous Culture of Bacteria; A Theoretical and Experimental Study." Journal of General Microbioloqv 14. 601 (1956). aX ' 58. Symons, J. M e , "Simple, Continuous Flow, Low and Variable Rate Pump." Journal Water Pollution Control Federation , 35, 1480 (1963) . 59. Ludzack, F„ J., "Laboratory Model Activated Sludge Unit." Journal Water Pollution Control Federation 32 605 (1960) . ~~ ' — ' 60. McKinney, R. E., "Bacterial Flocculation in Relationship to Aerobic Waste Treatment Processes." Progress Report on Project E-269 (C) , Southwest Foundation for Research and Education, San Antonio, Texas (1953) . 61. Aukamp, D. R., "Changes in Biomass of Activated Sludge," Unpublished Master of Science Thesis, University of Illinois (1966). 172 VITA Edward Richard Pershe was born at Omaha, Nebraska on July 30, 1924 . After graduating from South High School in Omaha in 1942, he attended one year at Creighton University before entering the U. s» Army in December 1943. After serving overseas in New Guinea, the Philippines, and Japan, he was honorably discharged in February 1946. He resumed his education and completed the civil engineering curriculum at the University of Illinois, receiving the degree of Bachelor of Science in June 1949. In October 1950 he received the degree of Master of Science in Sanitary Engineering from the Massachusetts Institute of Technology. He began his professional career as a sanitary engineer in the Medical Service Corps of the U. S. Air Force in 1950. After his honorable discharge in 1953, he spent five years in consulting engineering offices gaining valuable experience in design, supervision of construction, and report writing. In September 1958 he began his teaching career at the University of Nebraska. After receiving a National Science Foundation science faculty fellowship, he took leave of absence to study sanitary engineering at the University of Illinois. He is a member of the American Society of Civil 173 Engineers, the American Water Works Association, and the Nebraska Water Pollution Control Association. As a student he was a fraternal member of Chi Epsilon and Sigma Tau. He is the co-author of a paper with C. N. Sawyer and F. S. Howard entitled, "Scientific Basis for the Liming of Digesters," which was published in Sewage and Industrial Wastes Journal in August 1954, and is the author of a paper entitled, "The Development of the Sanitary Engineer," which was published in the Journal Water Pollution Control Federation in December 1963 . He is a registered professional engineer in the states of Ohio and Nebraska and is a reserve officer in the U S. Public Health Service with the rank of Lieutenant Commander. At the present time he is an Associate Professor of Civil Engineering at Northeastern University, Boston, Massachusetts. jRbANA 3 0112 027806527