1 15 1 https://www.nal.usda.gov/exhibits/speccoll/files/original/3212323b839b196dbdfe478477abbf1a.pdf 92c1536bb9111654f7aecf37f8743b79 PDF Text Text Item ID Number: Author Corporate Author 00105 Cobb H D ' Laboratory of Cellular Physiology, Trinity University, San Antonio, TX Roport/Artltio TitlO Biodegradation of Phenolic Paint-Stripping Waste: Laboratory Evaluation of a Fixed Film Batch Reactor Journal/Book Title Year 1979 Month/Day May Color Number of Images us DOSCrlptOll NOtOS F09635-77-C-0156; JON 21037W77; Program Element 63723F Friday, December 08, 2000 Page 105 of 106 �Cobb, I L L ) . , et al 1979 l ~ KTTR-79-11 BiodegradaLion of 1'honolic paint-stripping waste: Laboratory evaluation of a fixed Ellin batch rec.ctor biuutUKADATIQN OF PHENOLIC PAINT-STRIPPING WASTE: LABORATORY EVALUATION OF A FIXED FILM BATCH REACTOR HOWELL D COBB, JR JOHN W EGAN WILLIAM E OLIVE, JR DANIEL J HANSEN LABORATORY OF CELLULAR PHYSIOLOGY TRINITY UNIVERSITY SAN ANTONIO TX 78242 MAY 1979 FINAL REPORT OCTOBER 1977-JANUARY 1979 APPROVED FOR PUBLIC RELEASE; DISTRIBUTION UNLIMITED ENGINEERING AND SERVICES LABORATORY AIR FORCE ENGINEERING AND SERVICES CENTER TYNDALL AIR FORCE BASE, FLORIDA 32403 �UNCLAfiSIF'TF.n SECURITY CLASSIFICATION OF THIS PAGE (Whtn f>mt»'Entered) READ INSTRUCTIONS BEFORE COMPLETING FORM REPORT DOCUMENTATION PAGE I. REPORT NUMBER 2. GOVT ACCESSION NO 3. RECIPIENT'S CATALOG NUMBER ESL-TR-79-11 S. TYPE OF REPORT * PERIOD COVERED 4. TITLE (end Subtitle) BIODEGRADATION OF PHENOLIC PAINT-STRIPPING WASTE; Laboratory Evaluation of a Fixed Film Batch Reactor Final Report Oct 1977 - Jan 1979 6. PERFORMING ORG. REPORT NUMBER 8. CONTRACT OR GRANT NUMBERO) 7. AUTHORO) Howe11 D. Cobb, Jr John W. Egan William E. Olive, Jr Daniel J. Hansen 9. PERFORMING ORGANIZATION NAME AND ADDRESS F08635-77-C-0156 10. PROGRAM ELEMENT, PROJECT, TASK AREA A WORK UNIT NUMBERS Laboratory of Cellular Physiology Trinity University San Antonio TX 78242 JON 21037W77 Program Element 63723F 12. REPORT DATE II. CONTROLLING OFFICE NAME AND ADDRESS May 1979 HQ AFESC/RDVW Tyndall AF3 FL 32403 13. N U M B E R OF PAGES 110 TT~MONITORING AGENCY NAME A ADORESSf/f dlllerent from Controlling Office) 15. SECURITY CLASS, (ol Ihti report) UNCLASSIFIED I5«. DECLASSIFICATION/DOWNGRADING SCHEDULE (6. DISTRIBUTION STATEMENT (ot thi* Report) Approved for public release; distribution unlimited. 17. DISTRIBUTION STATEMENT (ol the mbtltmet entered In Block 30, II different from Report) 18. SUPPLEMENTARY NOTES Available in DDC 9. KEY WORDS (Continue on revetee elde II neceeeury *"d Identify by block number) Trickling filter Phenol Paint stripper Biodegradation Reactor genetics Chromium Toxic waste Fixed-film 0. ABSTRACT (Continue on reverse i/de // neee*e»ry mnd Identity by block number) USAF aircraft and ground support equipment require the protection of durable epoxy-polyurethane surface coatings. Maintenance of such painted surfaces using phenol - and chromium-containing strippers has created a waste disposal problem that is aggravated by the centralization of large aircraft depainting operations. Phe present investigation studied performance of a selectively-seeded, dedicated-function, trickling filter-type biodegradation unit. The specific waste target was the concentrated phenolic waste water produced at the Kelly (continued) • DD , JANRM7, 1473 UNCLASSIFIED SECURITY CLASSIFICATION OF THIS PAGE (When Dmlm Entered) �tTNCT-ASSTFIEn SECURITY-CLASSIFICATION OF THIS PACEfWisn Data Enterod) Block 20 continued. AFB-ALC depaint facility. Three filter units utilizing a top-down fluid-air flow batch operation were built at Trinity University, San Antonio, TX. Experiments were run examining solid, support media choice, bed length and volume, ventilation requirements, hydraulic surface loading, phenol concentration and loading, rate kinetics, chromium tolerance, starvation response, and temperature v effects. Results indicate that a small plastic bio-ring medium in this type filter operation, coupled with a genetically selected ecoysytem of column microorganisms, is an effective means of dealing with this waste stream. Requiring only coarse ( . 8 mesh) pretreatment filtration, the three 3 ft columns showed #1 essentially zero-order kinetics* handling concentrations of 1000 ppm at a , hydraulic loading of 6.1 gpm/ft with an average removal rate of 3 GM/Hr per ft of reactor at 20. C ( 8 F and 5.7 GM/Hr/ft at 30°C ( 6 F . Effluent phenol^con6°) 8°) centrations of less that 1 ppm were achieved. Compressed air at 0.16 cfm/ft supplied adequate1 filter ventilation. Phenol:nitrogen:phosphorus ratios of 1:10:48 were adequate for maximal rates; TRIS buffer at 1 g/L maintained optimum pH at 7-8. (A senior-controlled pH-stat system might replace the buffer, with inexpensive .alkaliL_pH adjustment.) Concentrations up to 3500 ppm phenol, and starvation for periods up to six weeks, were tolerated with recovery to normal within 7 ' hours/; overnight and weekend shutdowns were routinely handled with no 2 effect on; rate;. Routine process effluent contained only low amounts of fine suspended solids- Bed depths up to 11.5 ft yielded similar rates-per-unitvolume. It was- theorized that the batch process with, its alternating starvation/loading cycles selects for a microbial community better able to cope with occasional wider swings in this cycle. A thin-film reactor conserves the genes of its adapted! community more efficiently than other, reactor types. The combination of a batch operated thin-ffilm reactor could, on theoretical grounds, be expected to show stability, under conditions which would normally upset the> function of a continuous-flow- system. This study showed that such a system could also offer competitive phenol removal rates. The data, summarized in this report suggests that a batch fixed film process may have, advantages over other biological unit processes for some phenolic wasterstreams. Further, the literature surveyed indicates that various recovery schemes and physical/chemical treatment processes such as incineration, adsorption, and chemical oxidation may be more appropriate technologies, for the treatment of certain phenolic wastes. The selection of a process for a particular wastestream should, then, involve an evaluation of all existing technologies, including, the batch process described here, from a treatment efficiency, capital cost, and- operation and jnaintenance standpoint. �PREFACE This report was prepared by the Laboratory of Cellular Physiology, Trinity University, San Antonio, Texas under Contract No. FO8635-77-C0156 with the Civil and Environmental Engineering Development Office (CEEDO). Detachment 1, Armament Development and Test Center (ADTC)/ECW,. Tyndall AFB FL (26 Oct 77 - 30 Sep 78). Effective 1 March 1979 CEEDO was inactivated and became the Engineering and Services Laboratory (ESL), a. directorate of the Air Force Engineering and Services Center located on Tyndall AFB Florida 32403. Capt Jay A. Farmwald was program manager for this work. This report has been reviewed by the Information Office ( 1 and is 0) releasable to the National Technical Information Service (NTIS). At KTIS it will be available to the general public, including foreign nations. This technical report has been reviewed and is approved for publication. A. FARMWALD, Capt, USAF, BSC >ject Officer CROWLEY, MajjTUSAF Chief, Envlronics DivisM JOSEPH S. PIZZUTO,_ Col, USAF,_BSC Director, Engineering and Services Lab (The reverse of this page is blank) �TABLE OF CONTENTS Section I. Page 1 A. BACKGROUND 1 B. II. INTRODUCTION BIOPROCESSING ALTERNATIVES 3 PHASE ONE: AND MEDIA A. DEVELOPMENT OF SEED CULTURES 7 7 B. MEDIA FORMULATIONS 9 C. III. ISOLATION AND MAINTENANCE OF BACTERIAL STRAINS SEED CULTURE GROWTH STUDIES PHASE II: A. 11 PILOT PLANT STUDIES 16 PLANT DESIGN 16 1. Rationale for Batch Approach 16 2. Pilot Plant Design 18 B. COLUMN SEEDING 22 C. SUPPORT MEDIA SELECTION 23 1. 23 Support Media Design Requirements... 2. Methods and Materials "... 24 3. Media Performance D. 27 ACTUAL WASTE WATER EXPERIMENTS 34 1. Methods and Materials 34 2. Bound Volume Absorption Effect 38 3. Aeration Effects 43 4. Loading Effects 51 a. 51 Parameters ill �TABLE OF CONTENTS (cont'd) Section Page b. Reactor Flow Regime 53 c. Hydraulic Surface Loading Effects 53 d.. Organic Loading and Substrate Concentration Effects 60 5. 68 6. Starvation Recovery 71 7. Bed Volume and Depth Effects (Tandem Mode) 75 8. Temperature Studies 78 9. IV, Chromium Tolerance Biomass Evolution 80 DISCUSSION A. 33 83 PROCESS PROTOCOLS 90 1. General 90 2. Batch 91 3. B. GENETIC ASPECTS OF DEDICATED FUNCTION REACTORS Continuous-flow 92 4. Design Equations 93 a. Batch 94 c. VI. 93 b. V-. Kinetics Plug-flow 96 CONCLUSIONS 98 RECOMMENDATIONS 101 LITERATURE CITED 103 BIBLIOGRAPHY 108 iv �LIST 01? TAULES Table 1 Page Analysis of Paint Stripper and Paint Stripping Waste Water 5 2 Media Composition 8 3 Tap Water Analysis, San Antonio, Texas 8 4 Standard Run Conditions 20 5 Rate Data for Column Support Media Experiment 33 6 Aeration Effects 50 7 Countercurrent Filter Ventilation Effects.. 58 & Chromium Reduction 69 9 Rate Data for Temperature Experiments 77 LIST OF FIGURES Figure 1 Growth and phenol utilizaton on two selected mineral salts formulations!, 13 Growth and phenol utilization on five selected mineral salts formulations 14 3 Trickling filter pilot plant schematic.... 19 4 Triple column pilot plant 21 5 Three types of support media 25 6a Growth and utilization rates for three types of support media 28 2 b 29 c 30 d 31 �Figure Page 7 Mixed liquor turbidity 37 8 Column equilibration "sponge effect" 37 9 Graphic representation of column equilibration „ 41 10 Air flow patterns in trickling filter unit. 44 11 Conceptual model for bound volume V^ 46 12 Oxygen saturation for ventilated filter run 48 Oxygen saturation for non-ventilated filter run 48 14 Dye tracer pulse showing plug flow regime.. 54 15 Hydraulic loading effects 56 16a Substrate loading effects (raw data) 61 16b Substrate loading effects (rate vs. concentration plot) 61 1.7 Phenol inhibition model 64 18 Ortho- and meta-ring fission pathways 65 19 Six day starvation test 70 20 Oxygen consumption for slime and pure 13 , culture organisms 74 21 Tandem mode schematic 76 22 Phenol degradation rates for single column and tandem mode 76 23 Laboratory selective protocol vs. facility non-selective application 85 24 Batch and continuous flow schematics 89 25 Simplified on-site pilot plant arrangement. 102 vi �I. INTRODUCTION A. BACKGROUND The Air Force maintains five Air Logistics Centers (ALC's) which support large-scale paint stripping operations. These centers are responsible for the maintenance, depainting, and repainting of operational aircraft and ground support equipment. Current Air Force paint systems employ polyurethane topcoats with epoxy primers. Depainting these materials requires the use of paint removers which contain significant concentrations of phenol and chromium (see Table 1) ( ) The resultant 1. waste water represents the single major source of phenolic waste within the Air Force. Although 1983 EPA guidelines do not include prescribed technology for phenolic paint stripping wastes per se, the ability of industrial and municipal processing installations to meet the extremely low phenol concentration discharge requirements (0.02 ppm for Best Available Control Technology Economically Achievable--BACTEA) (2) will be affected by loadings from paint stripping facilities. The problem is aggra- vated by the centralization of large aircraft depainting operations. Process alternatives for the treatment of phenolic waste waters have been heavily investigated by both the Government and private industry. Systems examined have included various recovery schemes, incineration, adsorption, chemical oxidation, and biological degradation. For point-source streams �totaling less than 50 gpm at 2000 ppm, recovery of phenols does not appear economically feasible (3). Incineration is energy-intensive and practical only where generated heat can be used for additional purposes. Activated charcoal adsorption may be practical, depending strongly on .waste component •;impact on carbon renewal capacity, on costs associated with stringent pretreatment filtration, and on long-term ©Derating .costs. Chemical oxidation schemes can be dependable .and ieffective on high-strength point-source waste; but if more ..than one target toxicant is present, the number of treatment •grteps '.may -increase. Its normal use is as an adjunct--either _a pre treatment or polishing step--to .some other treatment: rprocess. Commercial analysts have found installation costs for b.ipprocessing systems to be slightly higher 'than activated .C.arbon, but operating costs to be significantly .loWer . 4 . () For an excellent summary and further references the reader -is referred to an article by K. H. Lanouette ( ) 5. The Air Force is continuing to evaluate alternative ..processes in their efforts to meet current interim .and future 'final discharge standards in a cost-effective manner,. Control technologies have been demonstrated for permanganate and ozone oxidation (6) and carbon adsorption ( ) The use of existing 1. military biological treatment facilities has al-so been .evaluated -(7). However, since the date of that Air Force investigation point-source treatment technology has advanced. �.B. BIOPROCESSING ALTERNATIVES Biological treatment of phenolic wastes has proven economical and reliable for certain industrial wastes at large and intermediate scales (2,4,8,9). Generally, biological process design, system cost, and full-scale reaction rates vary, depending on the exact nature of the waste water to be treated. Due to the complex kinetics that can occur when other toxic compounds or alternate carbon sources are present, the • common engineering practice is to evaluate the efficiency of teleprocessing on the specific waste stream (5,10). Most applications of biological phenol degradation have been in the petrochemical and plastics industries (4,IT,12,13), where activated sludge systems operating in the continuous-flow mode were early introduced (14,15). This treatment scheme evolved from the standard municipal facility experience, not entirely from recognized advantages of the method. Augmentation of municipal and industrial facilities sludge populations with adapted or mutated organisms has helped to improve process reliability (16,17); longer cell retention times can conserve these adapted organisms and lead to improved efficiency (5,18,19). Trickling filter operations have been investigated and proven feasible, especially for low-flow waste streams (20). High recycle rate filters are the rule. In larger systems, trickling filters have been used for "roughing" high-strength phenolic industrial waste in conjunction with oxidation lagoons (21,22,23,24). Commercial toxic waste water treatment engineering companies offer a variety of bioprocessing �approaches: The normal procedure includes development of strains suited to the particular waste stream, in a manner similar to that involved in this study, followed by seeding of an -appropriate type reactor. A unique aspect of ALC depainting waste streams is the high chromium content of the strippers currently in use. Several commercial concerns market mixtures of "adapted, .mutated",I;JBludge organisms designed for seeding municipal and industrial waste facilities. These products are based primar- ily on,soil pseudomonad organisms such as Pseudomonas putida, and Pseudomonas aeruginosa (17), and have proven useful in other systems (25). However their kinetics in this system .are not documented, and,of -the mutant strains on the market, none are recommended by .their maker/s to .operate at the chromium concentration..-of ,the KAFB facility,-raw waste-stream (17.5-59.5 ppm) (1, 7). ,;.(See .Table .1. and .Section III. .D. 5.) Dilution .to the required 2 ppm tolerance.level would entail an increase in waste stream volume of X20, on the average. Novel experimental systems such as fluidized bed bioreactors (25) have reported high degradation rates, but influent .streams are required to be very dilute, and such reactors have,hypercritical flow regimes; that combination of conditions makes them unsuitable for applications where constant ^attention is .not desirable. In the current study, the application of a genetically-adapted ecoflyfltcm to a narrowly defined waste stream �TABLE 1 ANALYSIS OF PAINT STRIPPER AND PAINT STRIPPING WASTEWATER NOTE: All values in mg/1 COMPONENT PAINT STRIPPER PHENOL 200,000 METHYLENE CHLORIDE 600,000 100,000 50,000 40,000 10,000 2,400 2,400 _ _ _ SURFACTANTS PARAFFIN WAX METHYL CELLULOSE WATER CHROMIUM TOTAL CHROMIUM +6 TOTAL PHOSPHATE (AS P) SUSPENDED SOLIDS VOLATILE SOLIDS _ _ _ TOTAL SOLIDS COD COD FILTERED (0,45u) TOC TOC FILTERED (0,45u) _ _ _ _ _ _ OIL AND GREASE * As reported by Perrotti C 3 ALC (KAFB) wastestrearn. PAINT STRIPPING* WASTEWATER 1010 - 4060 75 - 2000 120 _ _ _ 4000 _ - _ _ _ _ _ _ _ _ _ _ _ 17,5 _ _ _59,5 _ - _ 10 107 458 800 9200 7250 2710 2520 8,4 - 28 303 2700 3830 36400 35100 14400 13600 66,3 for the San Antonio �(a dedicated function system) is observed at the pilot scale. Research and reaction data are obtained using a batch trickling filter process on point-source paint stripping waste water generated at the San Antonio ALC depaint facility. The results demonstrate that the specific high phenol/chromium w$a.te s.tream qan be economically and efficiently detoxified to low levels on site using a biological system. The utiliza- tion of a, batch process protocol to maintain selectivepressures on the ecosystem is partly responsible for the unusual system kinetics, and suggests that full benefit pf genetic engineering ts not being obtained using standard tre.atment; techniques. �II. PHASE ONE: A. DEVELOPMENT OF SEED .CULTURES ;AND MEDIA ISOLATION AND MAINTENANCE OF BACTERIAL STRAINS Initial sampling for phenol degraders for use in these experiments was done in the paint stripping room (Bldg 375 KAFB), at locations expected to harbor microorganisms already exposed to concentrations of stripping waste. Bushnell and Haas (26 ) reported that hydrocarbon degraders would most likely be found in areas where such selective forces might already have acted. Samples on sterile cotton swabs were used to inoculate phenol agar slants (for composition, see Table 2 ) at the stripping site; the slants were then transported to our laboratory where they were incubated for four days at , 37°C., then used to inoculate flasks containing 50 ml of phenol broth (see Table 2 ) . After incubating for two days at 37°C., broth cultures were streaked for isolation onto | phenol agar plates. The isolated colonies appearing on | these plates were separated, yielding 24 possible strains, \ based on colony morphology. Tests for the ability to efficiently metabolize phenol in a broth medium were run on each isolate. Strain? i J10, J13, J20 and J24 (see Section II.C.) were selected for further study. These stocks were maintained on phenol agar slants and plates, transferred semi-monthly, and were never exposed to other carbon sources. i �T A B L E 2. MEDIA COMPOSITIONS NOTE: All •mound given In grama/liter unleaa otherwise apecifled. BflS (BASAL MINERAL SALTS) SWW McSW CACiJZ NH4N03 FeCL3 NACL K2HP04 KH2PW 0.1 0.01 1,0 0.0005 0.1 2.0 1.0 (SYNTHETIC WASTE MATER) FLASK COUMI PHENOL 0.50 - 1,0 SIGMA 7-9 TRIS 6,00 - 1,0 NH4N03 0.24 - 0,5 KH2P04 0.09 - 0.2 TAP MATER TO MAKE 1 L - 1 L PH 8.0 - 8,0 H20 TO HAKE 1 L pH 8.0 PHENOL BROTH PHENOL 0.5 , BUS TO MAKE 1 L pH 8.0 PHENOL PLATES I SLANTS PHENOL AGAR BUS TO HAKE PH AHM (ACTUAL WASTE HATER, FORTIFIED) INITIAL FEEDING FORMULA ( • h experiment) •e INTRA-EXPERIMENT FEEDINGS KAFB WASTE WATER(') KAFB MM 125 ML/L PAINT STRIPPER (*) PAINT STRIPPER (*) SIGMA 7-9 TRIS 1.0 SIOHA 7-9 TRIS 1,0 NH4N03 0,5 NH4N03 0.5 KH2PM 0,2 KH2P04 0.2 TAP MATER <•) RECYCLED SPENT pH 8.0 WASTE TO HAKE 1 L .PH 0,5 20.0 1L 8.0 * The anowta of cap water ( foe dilution ), and of paint atrlpper required to fortify AWW to deelred run concentration varied vfth actual KAFB waate phenol concentration. T A B L E 3, TAP WATER ANALYSIS, SAN ANTONIO, TEXAS SPECIES PPM SPECIES PPM CA++ MG++ NA++ FE++ MN++ 71.3 14,8 C03-S04-CLN03-- 259,3 25,4 15,1 1,65 8,5 0,02 0,002 8,0 �B. MEDIA FORMULATIONS For any microorganism to operate efficiently, certain mineral requirements must be met. Table 2 lists the various.media that were developed for the study's two phases. Discussed fully here are those which apply to the growth parameter "flask" experiments; the others are more thoroughly handled in their appropriate Sections. The Basal Mineral Salts (BMS) medium met all the basic mineral requirements for these organisms and was well-buffered within the physiological pH range (pH 7 - 8 ) . Its formulation was derived on the basis of prior experience with phenolic degraders (.27 ) . This became the standard medium against which performance by others was measured in the flask growth studies. However, supplementation of this magnitude in an operational degradative facility was recognized as undesirable, and very likely unnecessary. An analysis of San Antonio tap water (Table 3 ) revealed that, with the exception of nitrogen and phosphorus, most of the required minerals were present in concentrations sufficient to sustain cell densities of 1 mg/ml ( 28 ). Based on the assumption that bacterial cells contain approximately 8% nitrogen and 1 - 2% phosphorus, it was estimated that supplementation of tap water base by 0.242 g/L NH4N03 and 0.091 g/L KH2P04 would be required to maintain this density, or a 0:P ratio of 1:48.2, and a 0:N ratio of 1:10. �A concentration of 6 g/L TRIS* was used to buffer at the desired pH for flask runs, but this proved unnecessarily high; a much lower amount was used in the Pilot Plant Synthetic Waste Water (SWW) medium. The SWW formulation Chus represented an economical basic recipe, which was varied experimentally to de.termine the stringency of mineral and carbon: source requirements. It was recognized that, under field conditions, tap water mineral contributions would be significantly augmented, by additions from aircraft, parts, metal pallets, drain trough gratings, the paint and stripper themselves, the concrete, floor, etc.; and that other carbon sxxurcfes would be present.. These parameters were left until the £inal. phase, for study (see Section III. D. ). * THIS (hydroxyMthyl) tminoaetlune 99.0-99.51 Buffer Grade Chemical Co., St. Loula, Missouri. 10 �C. SEED CULTURE GROWTH STUDIES Flask phenol broth culture growth studies were conducted on selected bacterial strains. The primary criteria established to select these strains from the 24 isolates were rapidity of growth in phenol broth, colony morphology and vigor on phenol agar plates. Strains J10 and J13, both pseudomonads, and strains J20 and J24, almost identical subspecies in the Coryneform group, were chosen for the growth studies. Organisms from stock subcultures were used to inoculate spinner flasks containing one liter of phenol broth. After cultures achieved useful turbidimetric density in log phase,, the cells were harvested by centrifugation (5000 G) and resuspended in a small volume of BMS medium. Nephelometer flasks containing either BMS-based phenol broth, or SWW test media were inoculated with sufficient cell suspension to yield an initial cell concentration of 0.1 mg/ml (dry weight-tovolume) , and placed on an Eberbach reciprocating shaker for incubation at room temperature (25° C). Throughout all phases of this study, cell concentrations were determined turbidimetrically with a Klett-Sommerson colorimeter. One ml samples were drawn at predetermined intervals and quickly frozen, pending phenol and pH determinations. Phenol concentration was measured with a Barber Coleman Series 5000 gas chromatograph, equipped with a flame ionization detector (FID), under the following conditions: 11 column packing �10% SP-2100 on 80/100 Supelcoport; column temp. 180° C., injector temp. 190° C., detector temp. 225° C., nitrogen flow 63 ml/min., air at 40 psig., hydrogen at 20 psig; and sample size\3' ul. The preliminary growth and phenol metabolism experiments indicated that growth rates of J20 and J24 were about three times those of J10 and J13, and that phenol metabolism rates of J20 and J24 were ebout one-half those of J10 and J13. Typical performances are presented in Figure 1. The disparities between growth versus phenol netabol'ism exhibited by the pseudomonads were disconcerting.. It was clear that these organisms were degrading the phenol incompletely. Pigment buildup in the broth was substantial, a phenomenon reported by other investigators (29, 30). Since tfapid column seeding was regarded as a prime quality in the prospective inoculants, Coryneform strains J20 and J24 were selected for use. An added incentive was their heavy mucoid colony appearance, promising the quick development of a substantial slime layer on the Pilot Plant's support media, which would in turn contribute a foothold for other incidentally introduced organisms as the column ecosystem evolved. (See Section III.D.10.) Figure 2 illustrates growth curves and phenol utilization rates typical of J20 and J24. The figure shows response to a variety of mineral salts compositions. When raised on SWW, J20 achieved phenol utilization 12 �MINERRL SflLTS J13 J20 4 5 TIME(HOURS) MINERRL SFILTS J13 J20 ID J13 B£ O J13SW RRTE B.3 IG/H 7.1 O J20SVW 3.8 A J20 oe 2 3 4 4.1 S TIME(HOURS) FIGURE 1. Growth and p h e n o l u t i l i z a t i o n by strains J13 and J20 in two s e l e c t e d f o r m u l a t i o n s of m i n e r a l salts. 13 �MINERRL 5RLT5 J20 LJ G 9 12 TIME(HOURS) 1200 MINERRL 5RLTS J20 D O A <> X G B£ RRTE 3Hi 2X5W SVWHRLFN SNW HflLF P 9 12 5.0 IC/H 4.2 +.1 1.1 0.7 15 TIME(HOURS) FIGURE 2. Growth and phenol utilization by strain J20 in five selected formulations of mineral salts media. 2XSWWdouble SWW mineral salts; SWW J$N and ^P« half the NH4N03 and KH2?04 concentrations of SWW. Rates are in milligrams per hour (mg/hr). l/i �rates of 4.2 mg/hr, comparing favorably to its performance on BMS (5.0 mg/hr). A doubling of the nitrogen and phosphorus component concentrations in the SWW failed to increase phenol degradation rates, although the additional minerals did produce slightly higher cell densities. However, reduction by one-half of either nitrogen or phosphorus produced drastic reductions in the rates of phenol utilization and growth. From these studies it was concluded that San Antonio tap water is capable of providing a significant proportion of total mineral needs for the inoculant strains selected; that additional mineral input from paint, paint stripper, and other sources would, under field conditions, provide a still larger share. This confirms that it will only be necessary to provide a cheap nitrogen and phosphorus source to a pilot plant, together with a means to maintain a near-optimum pH, in order to assure satisfactorily high rates of microbial phenol metabolism on the Kelly waste. 15 �III. PHASE II: A. PILOT PLANT STUDIES PLANT DESIGN 1. Rationale for Batc:h Approach Rose ( 31 ), Grady (32 ) and others have argued that it is possible to extend batch findings to continuous flow applications in trickling filter operations. From the testing standpoint, where inhibition may be involved, it is easier to obtain kinetic data from batch runs than to attempt such data extraction from transient loadings of continuous-flow systems ( 33 ). But more than just experimental convenience recommends a batch approach to the Kelly AFB paint stripping waste. As is discussed in Section IV, trickling filter batch processes may conserve engineered bacterial gene constellations better than the more common means of waste bioprocessing. Space and configuration requirements for trickling filters are more appropriate to retrofitting an operation than activated sludge systems. Trickling filter operating costs are primarily controlled by energy used in recycle pumping and forced aeration. These costs compare favorably to activated sludge facilities which, due to higher biomass densities in the mixed liquor, require a more expensive form of forced aeration and heavierduty sludge pumps. Trickling filter batch operations are more suited to intermittent waste production and variable phenol concentration. Furthermore, their output tends to require less secondary treatment, carrying less suspended solids. And effluent quality may be controlled simply by altering batch time. 16 Much �work on sequencing batch reactors has been done by Irvine's group ( 34 , 35 , 36 , 37 ); periodic operations in other fields demonstrate advantages that apply equally to unsteady-state bioprocessing techniques ( 38, 39 ). When applied to the Kelly problem, the combination of advantages and untried possibilities made study of the batch mode seem a worthwhile alternative to more conventional continuousflow treatment. 17 �2 . Pilot Plant Design While the Phase I experiments were in progress, the design of the Pilot Plant system was finalized and its construction implemented. Three identical units were built, capable of independent operation for simultaneous studies, and also able to operate in a closed-loop, or "tandem" mode for the study of factors related to filter volume and depth. Figure 3 is a detailed schematic of an individual trickling filter unit. The two major structural parts are the column (one foot in diameter, four feet six inches high), and the 160-liter (42.3-gallon) base reservoir. Each column can be loaded with an appropriate support medium; in this case, either ceramic or plastic bio-saddles, or plastic bio-rings (see Section III.B below). A Flotec R2P1 variable-flow pump allows cycling of phenolic waste (synthetic or actual) over the column support media,with surface loadings 2 from 3.3 to 6.7 gpm/ft .Compressed air is provided to each unit via a metered inlet at the top of the column, and through a l"-diameter sparging stone low in the reservoir. Liquid flow rates are monitored with a Cole-Parmer magnetic flow meter; air flow is monitored and controlled with a ball-type metering valve at each entry site. Isolated column operation, or closed-loop, tandem mode operation is effected by the settings of valves in the liquid circuits. Liquid line access is provided below the flow meter site; access to the reservoir and column effluent 18 �AIR FLOW METER .jjTOP AIR INLET 1 RECYCLE LI If=M=n ] 1 Tiyp-B u>-/.^ p^^g . ' * MAGNETIC JL FLOW METER LJ 1 I.* . * 4 1 1 BOTTOM AIR INLET ^*=H= TANDEM MODE WW INLET . * )| * > RECYCLE PUMP/ I S.":-^w £/ i^i! !^i ! H' i" AIR FLOW METER ERFORATED DISC . . TANDEM MODE ^ H WW OUTPUT Qj— ACCESS PORT fc VENT -ff • W/ GATE VALVE '!'i |i' :; SPARGING STONE ^ 3. COLUMN ,\ ' ••* r '. ' (~\—-*"—^^*" m i* '. ••*.'- •". / ; FIGURE SUPPORT MEDIA SAMPLE PORT •* .' * ^ B ^ M Denotes valve. SPLASH DISTRIBUTOR U SUPPORT .: ^MEDIA .; DRAIN & =HN= SAMPLE LINE " V1fr 3 1U '| ^ ^1 L MIXED LIQUOR RESERVOIR ]' 6" 1 Trickling filter pilot plant schematic. 19 �stream is through a large-orifice gate valve in an upper corner; access to the column support media is through a. frontside plugged port at the three-foot line of the column. Throughout the study, the battery of filter units was housed inside a controlled environment chamber allowing '' maintenance of ambient temperature within ±0.5° C. of desired value (Figure 4). Except where noted, all studies were conducted at 20° C. (68° F.)(See Table 4). r . T A B L E 1, STANDARD RUN CONDITIONS PARAMETER REACTOR VOLUME D b FILTER BED • DEPTH (ONE COLUMN) • •* FILTER HYDRAULIC FLOW RATE Lfl 2.5 L/min- dcm TEMPERATURE ORGANIC LOADING (INITIAL) INITIAL PHENOL CONCENTRATION n dcm 68±2° F 8.0 8.0 o 2.82 g/L-day 177 lb/1000 ft3 -day C0]0 500-600 ppm pH INITIAL PH 0.53 ft3/min 2041° C ML T see § IV. C. 15 L/min f 2.6-31.7 GAL see § IV. C. ML Ly 42.3 GAL 10-120 L V 0.68 ft3/min.ft2 160 L T MIXED LIQUOR AERATION RATE 6.1 gpm/ ft2 2.1 L/min • dcm2 V SYSTEM HYDRAULIC LOADING 4.8 gpm 2 V/AR MIXED LIQUOR BATCH VOLUME 3.82 ft 18 L/min f MIXED LIQUOR TANK VOLUME -0.785 ft2 11.6 dcm IR HYDRAULIC, SURFACE LOADING 3 ft3 7.3 dcm2 R AR ENGLISH 85 L V REACTOR CROSS-SECT I'ONAL AREA FILTER VENTILATION LOADING METRIC SYMBOL L ty « 100 cur 20 . 500-600 ppm �FIGURE 4. Triple columns of the nhenolic waste blodegration pilot plant located in an environmental chamber at Trinity University. Note the three exnerimcntal. packing materials. 21 �B. COLUMN SEEDING To obtain a large inoculum for commencing column operations, a Virtis Model 43-100 Fermenter containing 16 liters of phenol broth was inoculated with strains J20 and J24. Phenol concentrations and pH were monitored and main- tained at nominal levels until the cell density had increased to approximately 1.2 mg/ml. During microscopic examination it was determined that approximately 10% of the population consisted of a pseudomonad invader closely resembling strain J13. Inasmuch as sterile technique was not to be imposed on column handling, it was decided to allow the mixture to become the primary inoculum, and to follow subsequent expected population changes microscopically. * Column seeding was accomplished using 4.5L of the primary seeding mix diluted to 40L with Synthetic Waste Water (SWW). This liquid phase was then recycled over the columns •j r in tandem mode for 30• min. to insure equal seeding inocula; 1 .-'. /'.• ' ' ' ' . . .• then valving was switched to isolated column mode. At a ,* ; •• . * • i • . '•, • 2 surface loading of 6.7 gpm/ft with 15 L/min forced aeration through the bottom inlet, the columns were run for 20 days. During this attachment phase, phenol concentration and.pH were monitored. When phenol concentration would near depletion, five liters of spent liquor (l/8th of Vj^) would be removed a.nd replaced with fresh SWW media containing Tris buffer, NH4N03, KH2P04, and supplemented with enough reagent phenol to bring the waste liquor up to 500 ppm. 22 �C. SUPPORT MEDIA SELECTION 1. Support Media Design Requirements One objective of this study was to select a support medium material for a phenolic waste degradation trickling filter unit which would yield a satisfactory compromise between engineering considerations and chemical/ 4 biological demands. It is known that the degradation potential of trickling filters is heavily dependent upon filter biomass. Biomass growth, in turn, can be maximized with support media having large specific area (surface area-to-volume ratios), and by low liquid flow rates. Unfortunately these are precisely the conditions which lead to flow blockage by particulates and sloughing biomass. Processes using once-through bottom- up fluid-air flow design have encountered this problem, which increases maintenance costs and down time, offsetting the high efficiencies of such units. As in the classic municipal trickling filter, such processes have found filamentous growths particularly annoying. Their solutions have included extremely fine (25-micron) pretreatment filtration and periodic support media replacement, both highly undesirable procedures for industrial operations which routinely produce large amounts of particulates in their waste. It was felt that, with a small sacrifice in potential rate, a system design incorporating much higher hydraulic surface loadings, coupled with coarser support media, might avoid these difficulties. 23 �2. Methods and Materials It was hypothesized that there would exist an optimum specific area,a (surface area/unit volume), for the packing medium which would support an active, stable biomass, tolerant of particulate loading. To test this parameter, two medias were chosen exhibiting a wide range of a.: a plastic saddle with a=20.67 dcm2/L (63 ft2/ft3); and a plastic bioring with a=34.12 dcm2/L (104 ft2/ft3)* Further, the surface characteristics of the support media in a high rate filter might prove critical for biomass anchorage, channeling patterns, composition of the • 11.>final organism population, and thus for phenol degradation performance. The most important changes would be those in the nature of the biomass population. Large shear forces i*. '• would demand a far more tenacious slime layer. To test for ••:••..•'•. . * surface effects, a ceramic saddle, comparable in specific ;>'-•: J ;';\; ' " • .-' • area to the plastic saddle was selected, with a=25.59 dcm2/L ' " ''' " •• 9 ' ' ( 7 ft2/ft3) . These solid media are illustrated in .Figure 5. .8 From the engineering point of view, plastic substrates present a much greater ease in handling than the ceramic saddles. There is a seven-fold savings in weight per unit volume using plastic saddles. (This advantage is only negligibly reduced for the bio-rings which, with a higher packing density, weigh slightly more per unit volume than the plastic saddles.) Both of the substrate materials are satisfactorily inert and durable. * In fact, the principal Plutlc »«ddl«: Actlfll 01-0100; plMtlc blorlnf: Act If 11 10-0160; c«raalc saddle: Actlfll 20-0160; Morton Co.. Akron, Ohio. 24 �CO 4-1 M-l co 4J CO co CSJ 4J <N U U-l CO iJ 14-1 (0 OJ en ft 4J CM oo on vO PLASTIC BIORING CERAMIC BIOSADDLE PLASTIC BIOSADDLE FIGURE 5. Three types of support media studied, shown from two perspectives at actual size, with a graphic depiction of their relative weight and surface area/unit volume. 25 �points remaining in doubt were related to packing density effects on hydraulic flow, and the suitability of the smooth plastic surface as an anchor site for our developing ecosystem of phenol degrading organisms. To insure similar startup conditions in media evaluation runs, fresh SWW was added according to the previously discussed protocol and circulated through all three columns in series for 30 min. to insure an equal distribution of microbes in the liquid phase. Experiments were conducted at batch volumes of 20, 40, 80, and 120 liters at hydraulic surface loadings of 2.47L/min-dcm^ (6 gpm/ft^) for approximately 50 hrs. each. When phenol concentrations approached low levels, one-quarter of the waste water volume was replaced with an equal volume of fresh standard SWW, plus additional reagent phenol sufficient to bring the entire liquid phase volume to the desired concentration. (This .procedure was more.economical •'•'' '•'•'&"' ••"''' '' ': • ' • : •' '•• '. of water and chemicals than discarding and replenishing the , • ' < . ' ' • ' ' . . • » • entire liquid'/phase.) During the run, 5 .ml samples were ! ••> ' "'" taken from the mixed liquor reservoir at appropriate intervals and frozen until GC analysis could be performed. Turbidi- metric and pH measurements were obtained at the time of sampling. At the end of each volume run the entire liquid ohase was discarded. 26 �3. Media Performance The phenol degradation rate curves in Figure 6 show that the bio-rings have a slight but persistent advantage over the other two support media. The relatively low overall rates are indicative of the immaturity of all three columns. (These are overall rates, expressed in"Table 5 as grams/hr-ft^ for each condition.). Observations of surface attachment and wall growth confirmed that all three support materials presented no hazard to colonization. However, Column 3, packed with bio-rings, exhibited more rapid biomass development and much less wall growth than the coarser materials. The close packing of the bio-rings against the column walls and each other prevented the gross channeling of waste flow that tended to develop with saddles. This effectively reduced superficial..flow velocities-, favoring attachment. Later experience with the bio-ring material unequivocally demonstrated that growth densities attained in these early support media experiments were far from the maximum possible, and that these initial rates of phenol degradation were low compared to those attained months later as the column ecosystems matured. The cell growth patterns seen in the upper graphs of Figure 6 are variations on the classic curve. The two saddle materials yield similar trends; the bio-ring data is more erratic. Since all columns were begun initially with 27 �SUPPORT MEDIfl 20LC5.3GHLJ Q CERflMIC STOCLES O PLflSTIC SflDDLES A PLflSTIC BIOR1NGS 20 30 TIME(HOURS) 1600 SUPPORT MEDIR 20L1.5.3GFL) Q CERHHIC SflDGLES O PLflSTIC STOLES A PLRSTIC BIOR1NGS TIMECHDURS) FIGURE 6a. Cell growth and phenol utilization rates for the FIGURE 6 a-d. three types of support media at batch volumes of a) 20L; b) 40L; c) 80L; and d) 120L. Interruptions in rate curves occur at intrarun feedings. See Table 5 for rate data. 28 �0.6 SUPPORT MEDIfl 40LU0.6GFLS) (3 CHROMIC STOLES CD PLRSTIC STOLES A PLflSTIC BIQRINDS 0.0 20 30 TIME(HOURS) 1600 SUPPORT MED in 40L .-ClO.GGflL) GERflMIC OTOLES O PLfETIC SRHJES PLfETtC BIOR1NQS 20 30 TIMECHOURS) FIGURE 6b. 29 �0.5 SUPPORT MEDIR 80L (21GflLJ m CERRMIC STOES O PLASTIC SflDOLES A PLflSTIC BIORJNGS 0.4 LJ a .0.2 u LU 0.1 LJ 0.0 10 20 .30 10 50 TIME(HOURS) .1600 SUPPORT MEDIfl BOL C21GRLD CD CERRMIC STOLES O PLffinC STOLES PLflSTIC BIORINQS 20 30 TIME(HOURS) FIGURE 6c. 30 10 GO �SUPPORT MED in 120L C32GRLJ O PLflSTIC SfCDLES A, PLflSTIC BIOR1NQS 20 30 10 TIME(HOURS) 1100 SUPPORT MEDin 120L (32GRU ID CERRUIC STOLES O FUSTIC SfCdES A PLflSTIC BIORINGS 20 30 TIMECHOURS) FIGURE 6d, 31 10 �the same liquid phase populations and identical chemical environments at each TQt the variation seen at 20L and 40L for the bio-rings is difficult to explain. However in both these cases the rate data in Table 5 puts bio-ring performance first. Evidentally microbes attached to the support medium were already appreciably enhancing Column 3's efficiency. In later experiments utilizing the bio-rings, this trend toward very low liquid phase cell densities coupled with high phenol removal efficiency became quite pronounced. There are clear effluent processing advantages to low particulate concentrations in the filter effluent. In calculating rates for this initial set of experiments, a quantity we term the "bound liquid volume", Vfc (fully discussed in the following section), was not taken into account. We were unaware of the effect at the time. However subsequent reevaluation of data for these early experiments has unequivocally shown that V^ is very small for immature columns, as would be expected from the analogy of a "sponge effect": very few absorption sites are available on the stationary phase of a filter at low biomass densities-there is not much of a column "phenol sponge". Additionally, at this early stage the filter units were not being ventilated, but were sparged in the liquid phase, a condition affecting absolute but not relative performance. In electing the bio-rings an medium of choice 32 �for all subsequent experiments, we were persuaded by its advantages in both major areas of concern: First, the extreme light weight of the basic material promised to greatly ease structural requirements for large facilities. Second, biomass attachment on the vast surface area offered much greater potential for rapid attainment of high zoogleal film densities in a compact reactor. The higher surface porosity of the ceramic material, said to promote rapid microbial attachment, did not enable that massive medium to approach che performance of the bio-rings. i T A B L E 5, RATE DATA FOR COLUMN SUPPORT MEDIA (PACKING) EXPERIMENTS NOTE: All races in grama/hour • ft3 of support medium SUPPORT MEDIA 20L V O L U M E 4 0L 8 0L PLASTIC BIORINGS 1,4 1,5 1,9 2,2 1,7 PLASTIC SADDLES 1,1 0,7 1,3 1,7 1,2 CERAMIC SADDLES 1,1 1,0 1,5 1,8 1,3 33 AVERAGES 1 2 0L �D. ACTUAL WASTE WATER EXPERIMENTS 1. Methods and Materials Stringent testing of the pilot plant and its ecosystem was performed using waste water obtained from the Kelly Air Force Base paint stripping facility at Bldg. 375. (This same waste water has been characterized previously by Perrotti (1) for bench and pilot scale carbon adsorption studies. A summary of his findings appears in Table 1.) The waste was pumped directly from the facility discharge sump •jt through a #18 USA Standard Testing Sieve into 15 gal. drums. Care was taken to obtain waste only during or shortly after active paint stripping operations, allowing observation of maximum concentration conditions. (Since the sump is regular- ly drained by pumping into sanitary sewer lines, the phenol concentration of waste water fluctuates widely, as alternating strip and wash operations are performed.) The drums were then transported back to the laboratory at Trinity University, o where they were tested and stored at 20 C. until used. The phenol concentration of raw actual waste water (AWW) was found to vary between 750 and 3500 ppm during the five month collection period, which correlates well with Perrotti's findings. At the close of the column packing selection phase, Columns 1 and 2 were emptied, washed, and repacked with the * Provided by p»»om«l of D«t 1, HQ ADTC/ECV. USAF Syttna Commni, Tynd.ll API, Florida. 34 �Actifil 10-0160 bio-ring material. Synthetic waste feeds were then provided and the assembly was run in the tandem (closedloop) mode for two weeks. At that time, individual column runs indicated comparable rates for all three columns; units 1 and 2 were then considered seeded and ready for experimental use. The Actual Waste Water experiments followed an SOP similar to that used during the media performance studies with the following exception: after running the initial batch of mineral salts-fortified AWW, subsequent batches within the experiment were brought to standard initial phenol concentration by draining one-eighth of the spent liquid volume, and replacing that volume with fortified AWW, per the formulation detailed in Table 2. Although turbidimetric measurements were regularly taken on samples from the mixed liquor during runs, their utility was compromised by an unexpected phenomenon. Where in earlier \ runs cell densities had followed a standard growth curve with substrate disappearance (Figure 6), virtually all experiments in the AWW phase demonstrated a reverse trend. Immediately on run startup the mixed liquor would give relatively high Klett readings--due to the innate turbidity of the AWW, washdown from the filter, and sediment mobilized by the fill procedure. These readings would rapidly decline to a constant level around 35 Klett units--equivalent to approximately .2 mg/ml suspended 35 �solids (Figure 7)--and stay there throughout even a long run. Consequently the usual estimates of biomass production based on mixed liquor cell densities have not been made. It is possible to say, however, that mixed liquor cell densities were never high during the AWW studies, and were significantly lower than those 'normal1 growth curves recorded for the early experiments, done when the filter biomass was immature. 36 �MIXED LIQUOR TURBIDITY CR1WI) TIME(HOURS) FIGURE 7. Mixed liquor turbidity from mature column runs at a batch volume of SOL. 1200 COLUMN EQUILIBRRTION D RECYCLE LIKE 0 COLUMN EFFLUENT 0.4 O.G TIME(HOUR) FIGURE 8. "Column sponge" effect during equilibration period. Samples were taken simultaneously from the recycle line and column effluent. 37 �2. Bound Volume Absorption Effect Biomass established on the column was observed to have a strong absorptive affinity for phenol at the onset of a run. At startup, samples of column effluent typically showed a phenol concentration 5070 lower than influent samples of the same plug, with a rapid decline in difference leading to a state of equilibrium over a 30 to 60 minute period (see Figure 8). This "column sponge" effect presented early difficulties in interpreting system dynamics. There are two possible mechanisms which may account for this effect. On the one hand, there is evidence of phenol- specific adsorption in the biomass, which could be due to microbially-produced compounds with a binding affinity for phenol, such as certain extracellular enzymes. A second mechanism would be absorption, either specific or nonspecific. Specific absorption would describe phenol diffusion into slimes with high phenol diffusion constants. Nonspecific absorption would apply to slimes showing no distinct diffusion "preference" for phenol over other soluble organics. To facilitate mathematical handling of the effect, we elected to adopt a diffusion model for the purposes of this study. A counterpart to the "column sponge" exists in activated sludge systems (19, 40) and has been put to work in the contact stabilization process. But for other activated sludge systems, initial apparent high rates of substrate uptake must be accounted for only during unsteady-state situations, 38 �such as with hydraulic or organic loading transients (41, 42, 43, 44, 45). Inasmuch as transient states are normally abated by system design, the effect has drawn little attention. Batch systems, on the other hand, typically exhibit extremes of substrate concentration oscillation. The "sponge" effect can become paramount in determining absolute rates (mass/time) in cases where reactor volume (and the sponge volume) and mixed liquor batch volume do not differ greatly—as in pilot scale projects. Conceptually, the sponge effect is due to diffusionequilibration of phenol in the newly recharged mixed liquor with absorptive sites in the filter biomass (see Figure 11) , consisting of water, polymeric secretions, bacterial and fungal cells and cellular debris. Theoretical diffusion properties of such bacterial films have been discussed (46, 47), but model equations have limited pragmatic value. There are indications that these materials can show surprising adsorptive specificities, unique to both the particular biomass and substrate of interest. A high rate trickling filter exposed to limited numbers of carbon sources appears to evolve a characteristic substrate absorptive capacity range. For operational purposes, we have termed this capacity the "bound volume", V^, defined as that additional water volume which, when added to the initial mixed liquor volume V, will result in dilution of the initial absolute mass of substrate added, 0. , to an effective initial o 39 �equilibrium concentration of [0e] V , or (1) b + VML Rearranging, - VML () 2 C0e]t Except for inflections following equilibration, and at the low end of [ ] the kinetics of substrate removal for this system 0, are essentially linear. If a zero order rate is assumed during the equilibration period, then [0e]t may be determined as o indicated in Figure 9, or: [0]t = [0c]t (3) o •i-1 Rearranging: [ L - C0 0. [0e], t o = [ .t 01 () 4 No attempt was made in this study to thoroughly examine V,, nor to model its variation. It was treated as an operational parameter, empirically derived. Vb appeared to increase to a steady-state plateau of 35L as the filter biomass �FIGURE 9 . Representation of the equilibration process and its relationship to rate calculation using [0 ]t . [0Mj] is the mixed liquor phenol concentration just prior to commencement of batch recycle. [ p o is the filter effluent phenol concentration of 0] the first slug through after startup. Sampling of recycle line and column effluent shows a rapid convergence to a steady-state difference. P represents the point at which mass transfer of mixed liquor-borne phenol into the column slime (and Vjj) no longer dominates metabolic removal. Continued steep slope of recycle line curve to E is due to lag in liquid mixing of low-phenol column effluent with V-^, in batch reservoir. Exact determinants of P have not been defined; in our system, turnover occurred within five minutes of startup, and showed variation with batch size. After equilibration A[0]jj was found to represent the instantaneous metabolic rate of the filter bed. If metabolic rate is assumed linear during equilibration, as after, then extension of recycle and column effluent lines to Y-intercept yields a useful imaginary quantity, [0e]o, conceptualized as the phenol concentration of the combined volumes, V^L+V^. In view of the massive amount of phenol originally entering Vb, one can visualize V^ as constantly at equilibrium with the recycling waste stream, minus an amount due to metabolism. Therefore the calculation of rate from recycle line (i.e., mixed liquor) phenol concentration data alone must consider the waste volume acted upon by the biomass to include V._+V, ( 8 . 4) ML b 41 �matured, from which value it deviated only when the columns were stressed by starvation or phenol overload. The fact that V, was essentially independent of 0 , giving similar values o over the wide range of [0e], derived, was supporting evidence t o that the actual microbial metabolism of phenol during the equilibration period was indeed linear. As a point of interest, a "bound" volume of 35L, if evenly distributed over the specific surface area of the column packing media, would form a layer 0.12 cm thick, a not unlikely value for the observed film depth. The impact of the bound volume effect was greater on this study than. would be the case in field type units, where the recycling mixed liquor volume would normally far exceed the range of bound volume on the filter bed. Pilot scale data herein reports [0] as measured in the mixed liquor in ppm. However, rates in ppm/hr can be compared only within a q run. Between runs, the absolute rate in grams/hour-ft must be used. The absolute rate between any two points can be L" determined using (VVJT+ V, ) as the volume term. In all cases where absolute rates are reported, it has been calculated using the formula: R D [] 0 _[0] ( t ft ^ ~ b (5) V R to yield the rate in g/hour-ft^ of reactor volume, where both [ ] . and [ ] . 0^ 0. are mixed liquor phenol concentrations measured t C i i-l after equilibration, at times t. and t._-,. 42 �3. Aeration Effects For this study, it was realized that ventilation served a dual purpose, the first being provision of oxygen to a dense biomass requiring a theoretical (^-to-phenol mass ratio of 2.4:1 ( 5 ). The second purpose stemmed from the presence of highly volatile methylene chloride (CI^CH Cl) , a major paint stripper component, with the potential for interfering with microbial metabolism in the filter. As shown in Figure 10, in order to rule out oxygen limitation, forced rather than diffusion ventilation was elected; this method also served to prevent buildup of toxic volatiles. In the liquid phase, where we expected some microbial growth and phenol metabolism to take place, sparging was introduced to supply any additional oxygen demand. Vortex currents generated by the rising bubbles also aided in mixing the returned column effluent. When aerating the mixed liquor phase, compressed air was' admitted through a 1"-diameter sparging stone located near the bottom of the liquid reservoir, then vented out the reservoir access port. When ventilating the filter, compressed air was admitted through the inlet located at the top of the column, passed down through the support medium and expelled through the outlet (access) port. This separation allowed a semi-quantitative analysis of the roles played by ventilation versus mixed liquor aeration. Mueller ( 49 ) has pointed out that oxygen may become the rate limiting nutrient in high organically loaded 43 �AIR b{Xj= INLET ~ FILTER VENTILATION AIR INLET MIXED LIQUOR AERATION FIGURE 10. Air flow patterns for filter ventilation and mixed liquor aeration. �filters. Metcalf and Eddy, Inc. ( 19 ) discuss this, and recommend that air be supplied in the range of 0.03-0.30 L/mindcm^ ( 0 . 1 - 1 . 0 cfm/ft^), either by natural convection or forced ventilation, in an effort to keep that layer of bulk liquid closest to the ambient air at saturation. Figure 11 illustrates 'model1 and 'actual1 conditions existing in the Pilot Plant filter bed. The model provides for a layer of stagnant fluid on the biomass, called the "bound volume", V^, which represents the effect of all non-metabolic absorption sites for phenol within the biomass (2. above). layer. Flowing down over this is the mobile waste liquor Conventionally all these "layers" are treated as differentiated and of defined thickness. In reality, as indicated in the inset drawing, the exceedingly complex surface of the slimes—its actual surface area far greater than that provided by the media specific surface area, a--is well-exposed to ventilating gas flow, with the depth of the overlying fluid layers, both flowing and stagnant, varying locally over a wide range. Oxygen from the waste stream as well as from'ventilating air is available for biological oxidation by diffusion into the biomass. By-products and C02 diffuse out of the slime into the flowing liquid phase, eventually coming to equilibrium with the ventilating gas stream. Most models depict this as occurring through the differentiated liquid layers. But scanning electron microscopic observation of the 45 �B I0 NAS S ANAEROBIC FIGURE 11. A conceptual model illustrating bound volume, Vfo, and its relationship to conditions existing in a thinfilm reactor. 46 �slimes makes this simple picture seem incomplete; the biomass surface is highly textured on both a gross and fine scale. Virtually direct access to ventilating gases is a requirement for optimal oxygen transfer, expecially where a thick zoogleal film is involved. On the other hand, mass transfer of phenol would seem to require greater exposure of the film to the waste liquor stream. Calculations based on the known volume of mobile fluid on the column (V ), and the probable volume of biomass, indicate that the transiting waste liquor volume would theoretically be distributed in a layer of extreme thinness over the biomass surface. Under actual operat- ing conditions such uniformity is never the case. It is there- fore likely that oxygen and phenol availablity exist simultaneously as limiting factors at different sites at the microecological level. An optimal supply of. both materials would occur under somewhat turbulent filter flow conditions. A combination of random feed distribution (locally irregular hydrodynamic flew) and ventilating gas flow would therefore contribute to removal efficiencies in the filter. Figures 12 and 13 illustrate column influent and effluent oxygen concentrations under ventilation and nonventilation conditions. olism It is clear from the effect on metab- that ventilation is a far more significant source of oxygen than the dissolved gases in the influent stream. decreasing rates in Figure 13 (The reflect oxygen limitation setting in as trapped ventilating gas was depleted after ventilation 47 �VENTILflTED FILTER BOO 100 m RECYCLE © 0; RECYCLE LINE A 08 FILTER EFFUBIT 700 600 20 10 GO BO 100 TIME(MINUTES) FIGURE 12. Recycle line and column effluent % oxygen saturation levels during a normal run with both mixed liquor sparging and filter ventilation. The initial equilibration period is included in the graph to show that assumption of linear removal during this period is supported by 02 data. 1000 NON VENTILflTED FILTER GJ RECYCLE LIIC O 0, RECYCLE LINE A 02 FILTER EFFLUENT TIMECHOURS) FIGURE 13. The importance of gas-phase 02 in the filter bed can be inferred from the rapid depletion of column effluent dissolved 02 following ventilation shutdown at T+30m, even though mixed liquor sparging continued. Again equilibration period data is included. 48 �shutdown at equilibrium, T+30 min.) Table 6 gives rates obtained for several protocols of aeration. A series of unreported experiments deter- mined that batch size did not significantly alter rates; however, runs were made at 40 and 120 L (10.6 and 31.7 gals.) as spot checks of variation. No reasons were found to change the assessment. The tabulated results reconfirm filter shutdown when not ventilated, leaving a mixed liquor phase which is inactive. The low degradation rates are due to oxygen-limited metabolism on the column since, under the reverse condition, wich the column ventilated and no liquid sparging, rates are normal. There is essentially no anaerobic phenol metabolism in this system. Aeration of both phases together at 15 L/min did not produce any significant change in performance, indicating no oxygen transfer advantage to sparging the mixed liquor in either a scaled-up batch or continuous-flow facility. For standard conditions throughout this study, air was supplied at 2.05 L/min-dcm^ (.67 cfpm/ft^). Oxygen was not limiting for ventilation rates within the range of values tested. There was a modest trend toward slightly higher per- formance at the highest rates; however, it is likely that the extra energy required to supply roughly twice the air volume to the column in order to obtain a mere 7 percent increase in removal rate would not be cost effective. A third aeration approach attempted briefly in this study was counter-current ventilation. 49 In this mode, �sparged air from the mixed liquor was forced to flow up through the filter and allowed to exit at the top. Waste liquor holdup on the column was appreciable, and under the control of the filter hydraulic loading. Possibilities for use of this mode, and certain other ramifications, are further discussed in 4.C., below. TABLE AERATION FILTER VENTILATION 2 L/min-dcmf 4,11 2,05 2,05 1,37 0,69 0,00 MIXED LIQUOR AERATION 6, EFFECTS PHENOL UTILIZATION RATE 4 0L L/min | 12 0 L grams/hr-ft 15 15 0 10 5 15 — 3,1 3,1 2,9 2,7 0,9 50 3,5 3,2 — — — 0,5 �4. Loading Effects a. Parameters Primary design parameters for fixed film systems include organic loading L (mass of substrate/reactor volume time); hydraulic surface loading, Ls (reactor influent flow volume/reactor cross-sectional area time). Process efficiencies in fixed film systems are a function of organic loading because of the upper limit on total biomass related to maximum film thickness, support media specific surface area, a, and reactor volume, VR. One aspect of organic loading — the substrate concentration-becomes load-limiting in systems where substrate removal \ follows anything other than zero-order kinetics (50), and especially where a significant carbon source exhibits concentration-dependent toxic or inhibitory effects. Such a target substrate affects its own degradation rate, as well as that of other waste components, placing design constraints on substrate concentration limits and control provisions. This situation applies to phenol-loaded systems, and is of major importance when phenol constitutes the primary carbon source in the waste stream, as it does for the Kelly AFB paint stripping facility. Findings related in d., below, explore factors bearing on substrate loading and concentration effects, 51 �Hydraulic surface loading, qR/AR--where AR is the reactor cross-sectional area—is an operational variable with complex effects. Fluid motion through a trickling filter, as a source of shear forces acting on the biomass, exerts powerful selective effects on filter organisms and establishes upper limits on film thickness. Its interplay with hydrodynamic characteristics of reactor, media, and biomass establishes the mean hydraulic retention —— 9,,, and ~^^^~~~~~~~ volume, VC , of time, M transit —— the filter. Combined with the characteristic porosity of the column media, it also controls filter sensitivity to influent particulates. In this study, the base drain arrangement was nonrestrictive of fluid flow; media effects were characteristic for the packing material and constant; biomass effects stablized after maturity (V, plateau); and the.value of qR, the reactor influent flow rate, was not varied widely enough to alter QT, significantly. For a thorough discussion of the effects of liquid residence time, in low and high rate filters, see the article by Klemetson (51). In cases where concentration and hydraulic loading considerations establish organic loading limits, in both continuous-flow and batch systems, the desired organic loading (capacity) is achievable through manipulation of reactor volume. (Effects of changes in VR are handled in D.7. below) Within a given reactor volume, one may alter structural retention characteristics to increase V. in order to obtain phenol removal from a larger number of particles per pass 52 �through the filter, thereby increasing the effective column rate. Although an indirect aspect of hydraulic effects, this was not included in the variables examined in this study. Reference is made to incidental findings related to V in c., below. b. Reactor Flow Regime Dye tracer slugs were used to verify flow characteristics through the reactor and to measure the mean hydraulic retention time, 0H ( 19 ). Ten ml of erythrosin red dye in aqueous solution were injected into the waste stream entering the filter dispersion head. Samples were collected from the column effluent at 5 second intervals, and optical absorbance determined at 525 nm on an Hitachi-Perkin Elmer Model 139 UV-VIS spectrophotometer. absorbance versus time in Figure 14 The plot of represents a 5th order computer approximation combining multiple trials under identical conditions. Technically, the small degree of longitudinal dispersion exhibited qualifies this filter configuration as an "arbitrary flow" reactor whose flow may be approximated by ideal plug flow relationships ( 52f 53 ). c. Hydraulic Surface Loading .Effects One of the design goals for this dedicated function trickling filter was high immunity to blockage, coupled with relaxed requirements for waste stream pretreatment 53 �PLUGFL01W 0.6 0.5 • cn 0.3 - CE LJ 0.2 4 Q_ a o.i 4 0.0 o.o 0.5 1.0 1.5 2.0 TIME CMINUTE5) FIGURE 14. Dye tracer pulse revealing approximate plugflow regime for normal filter conditions. �clarification. A second goal was highest rate compatible with compact reactor structure. The latter dictated that high specific area packing media be used, in order to obtain a large active biomass (Section III. C., above). The smaller diameter packing and denser biomass, in turn, meant higher waste flow rates through the filter would be needed. The effect of surface loading, Lg, on phenol removal rate is illustrated in Figure 15. These experiments were conducted in batch mode at initial phenol concentrations between 500 and 600 ppm in order to minimize any concentration effects on rate. (Growth studies had earlier indicated linear rates for initial concentrations in this range). The span in Ls , from 1.37L/min-dcm2 to 2.74L/ 2 9 mn-dcm 2 (3.3gpm/ft to 6.7 gpm/ft ), plus the zero loading condition, represented the delivery range available from the Flotec R2P1 pumps yielding satisfactory dispersion of waste at the column head. Modest effects on rates were noted, increasing with flow, indicating that an optimum was reached 2 between 5-6.7 gpm/ft . The rate that was selected for standard operating conditions throughout the rest of the study 2 was 6 gpm/ft , thus the standard operating procedure (SOP) encompassed an optimum value insofar as metabolic rates were concerned. The zero-flow condition represented, of course, only mixed liquor metabolism since there was no waste .water contact with the column. The low value of 0.2 g/hr reemphasized 55 �GOO HYDRRULIC LORDING EFFECTS 40LC10.GGRL) 0 Gft_/MlN-FT2 0.« G/H-FT3 3.+ 2.50 5.1 2.73 6.1 2.73 2 3 1 TIMECHOURS) HYDRRULIC LORDING EFFECTS 120L C32GRL) 0 GFL/MIN-FT2 0.20 G^H-FT1 2.57 5.1 3.13 G. 1 3.10 6.7 3.23 TIMECHOURS) FIGURE 15. P h e n o l d e g r a d a t i o n r a t e s o b s e r v e d at s e l e c t e d h y d r a u l i c l o a d i n g s a t A O L a n d 120L b a t c h v o l u m e s . 56 �the small contribution of mixed liquor metabolism to overall performance in this sytem. Column transit volume, Vt, was observed to increase from 4L at the low flow to 7L at the high flow rate, measured as the immediate decrease in V^ when new batch recycling was begun, minus the plumbing volume. Effluent oxygen concentration data demonstrated only relatively slight reductions from the near saturated state existing at the filter head with standard condition air flow (see D.3. above). There- fore it was considered unlikely that the modest rate increases observed were due to better aeration at the dispersion head. An increase in dispersion efficiency and in Vt was probably responsible, with more of the biomass in contact with phenolbearing waste liquor at any given time. The implication is that a reactor design yielding a larger Vt might improve system efficiency. A brief series of experiments early in the study sought to evaluate a countercurrent mode of ventilation, as was briefly mentioned in D.3. above. Air was bubbled through the mixed liquor, then forced to flow upward through the filter against the descending waste liquor flow, exiting at the top. A large fluid buildup was observed on the column, and foaming at the top required entrapment. Observed rates were significantly higher than the most recent base rates. 57 �Degradation rates under several countercurrent air and hydraulic flow values are presented in Table 7. Even at an air up-flow of 0.69 L/min-dcm2 (0.18 cfpm/ft2)--equal to the lowest top-down flow ever applied in the study--phenol degradation occurred at a rate virtually equal to top-down nominal. Phenol removal moderately increased at the highest ventilation. However, hydraulic loading changes resulted in the strongest effects on unit removal efficiency. rate in Table The high of 5.7 g/hr-ft3 is equivalent to a phenol 7 loading of 4.83g/L-day (302.7 lb/1000ft3-day) while operating in a batch mode. TABLE 7. COUNTERCURRENT FILTER VENTILATION EFFECTS MOTE: All phenol utilization ratal In g/hr-ft3 FILTER VENTILATIONmod*) LOADING (Count«rcurr«nt HYDRAULIC SURFACE LOADING (Countercurraat MM) L.- 5.05 gpm-ft 2 f^- 0.53 eft. VV °-(B e*"-"2 'ML-'V 0.23 cfm.ft2 0.45 cfm.ft2 0.67 cfm-ft2 3.03 gpm-ft2 2.99 3.13 3,42 2,17 TOP-DOWN NOMINAL BASE RATE 2.73 5.05 n>m-ft2 3,76 ' TOP-DOWN NOMINAL BASE RATE 58 7.07 gpn-ft2 5.70 2.73 �Observation confirmed that the major impact of waste liquor surface loading was on the transit fluid volume, Vt. Countercurrent airflow served to obstruct the base drain configuration, and to increase turbulence in the filter bed flow. The net effects were: Longer hydraulic retention times, greater biomass contact with transiting liquid, longer ventilating air contact time, and improved reactor efficiency. Unfortunately, the mixed liquor tank was unable to contain the resulting pressure, dampening our spirits considerably. Aeration procedures were returned to low pressure, top-down normal. The episode underscored the potential of transit volume as a design variable affecting trickling filter efficiency, and countercurrent ventilation as a means of controlling it. Section IV.C.3. offers a discussion of applications. 59 �d. Organic Loading and Substrate Concentration Effects To determine the kinetics and loading tolerance of the system, batches were run at a volume of 20L under initial phenol concentrations varying from 700 Figure 16a. illustrates raw data obtained. linearity is noteworthy. to 3300 ppm. The trend toward In Figure 16b. the rate-vs-concentra- tion relationships are more apparent, including a clearcut, mounting depression of overall rate with increased initial loading. (The curves in Figure 16b. were obtained by treating raw data points with a least-squares-fit, 2nd order approximation, then calculating rates between generated data points. This procedure would smooth out between-point variability and might be criticized as disguising Monod trends; however, as the raw data shows, this does not appear to be the case, down to very low phenol concentrations. Additionally this procedure is useful in clarifying the role of [ ] in determining overall 0Q rate.) The known toxicity of phenol was expected to alter the conventional Monod rate-vs-substrate concentration relationship, depicted by the dashed curve in Figure (54 ). 17 But the uniqueness of the filter's biomass made it impossible to predict what form this alteration might take. In earlier work we had reported on inhibitory effects of high cresol concentrations (>1400 ppm) on pure culture growth systems demonstrating apparent non-Monodian kinetics ( 27 ). Yang 60 �SUBSTRflTE LORDING 20L C5.3Gni) 3500 0 10 5 15 20 25 TIME(HOURS) FIGURE 16a. Raw d a t a for s u b s t r a t e l o a d i n g r u n s . SUBSTRRTE LORDING 3 20L C5.3GRL) A-A-A-A-A-A m 2 UD v^< Ld I— CC i 0 350 700 1050 1400 1750 2100 2150 2800 3150 3500 PHENOL CONC.CPPM) FIGURE 16b. R a t e vs. c o n c e n t r a t i o n p l o t , u s i n g 2nd o r d e r a p p r o x i m a t i o n s o f r a w d a t a , showing t h e o v e r a l l i n h i b i t i o n of batch rates at the test volume with.increasing [0]Q (hence higher organic loading). • 61 �and Humphrey (55) found classical Monod kinetics in their work with Trichosporon cutaneum (a yeast) and Pseudomonas putida pure and mixed batch cultures, for phenol concentrations up to 100 ppm, and severe inhibition of phenol uptake rate beyond that. Activated sludge studies of transient loading surges for continuous -flow nonphenolic waste treatment systems have reported an inhibitory phenomenon ( 0 , but inhibition due to 4) nontoxic organics probably involves different mechanisms. Rate data for the Pilot Plant system plotted in Figure 16b portray an unexpected dimension of inhibition. Within a given run, degradation can be approximated by nearzero-order kinetics to a low substrate concentration. But the envelope rate curve (ABCDE) resembles a modified noncompetitive enzyme inhibition equation of the form max 1 + (Ks/ where R = batch reaction rate corresponding to some [ ] 0o (R assumed constant for a given [0]o> i- e zero-order kinetics ); R.UlciJC = highest initial rate measured over the noninhibiting range of [ ] 0; K = Monod' s half-velocity constant for BLaxs [0] = initial batch phenol concentration; K.^ = the inhibition constant for phenol, which for this unusual type of inhibition interaction can be 62 �written as K. = where phenol may be doing double duty as substrate and inhibitor, I; i.e., [I] = [ ]. 00 In these experiments, KI =-2000 ppm, R max * 3'3 8/hr-ft3, Ks * 1 ppm. The curve envelope ABCDE suggests a homotropic inhibitory action on a K3-type regulatory enzyme (56), but the family of curves representing actual runs at specific [ ] ' 0os does not support such a simple explanation (see Figure 16b). Some possibilities include: a) one or more of the primary enzymes in the degradation scheme--perhaps the membrane-associated hydroxylase (Figure 18)--are exhibiting a complex, time-dependent regulatory effect; b) the long term rate inhibition is due to a mechanism involving more than mere inhibition of ring-fissioning enzymes, possibly extending to suppression of RNA synthesis and/or translation throughout the biomass; c) the control may be exerted by some other component(s), either of the waste water or of phenol degradation; or d) the effect may be some combination of these or other unrecognized mechanisms . Applied to the test ecosystem, the utility of equation (6) lies in its ability to predict the system initial rate R, once Rnicix has been established at the non- 63 �PHENOL INHIBITION MODEL m Rmax t i B IE i; i; I! li I: i: KB » I «2200 «1000 «3200 [0], FIGURE 17. A graphic representation of standard Monod model .m (heavy dashed curve), together with inhibition model for phenol in this system. Batches with initial concentrations [ ] falling along the line from A 00 to D show near zero-order rates to very low ppm. For 0 < [ ] ^ 1000 ppm 0Q (C to E), Monod-type kinetics apply, and all batch rates follow the dashed curve to c. KB for this system is low, perhaps below 1 ppm. max is approximately 3.3 g/hr-ft3 at 20°C. 64 �PHENOL NADU Hydroxylase (membrane bound) V A H)O' NAD ORTHO-RING FISSION PATHWAY META-RING FISSION PATHWAY meia o.xvgcnase Dchydrogcnase NAD. NADH a OH 2-HYDROXYMUCONIC SEMIALDEHYDE )OH COOH Hydrolusc COOH "COOH 4-OXALOCROTONATE oVis-MUCONATE 1 1 1 0-KETOADIPATE i 1 SUCCINATE + ACETVL-CoA • HCOOH CH., COOH 2-KETOPENT-4ENOIC ACID I 1 ACETALDfHYDE + PYKUVATE FIGURE 18. O r t h o - and m e t a - r i n g f i s s i o n p a t h w a y s known to e x i s t i n t h e T r i n i t y e c o s y s t e m . T h e h y d r o x y l a s e e n z y m e Is b e l i e v e d to be m e m b r a n e - b o u n d , and may be the i n i t i a l target of inhibiting e f f e c t s of phenol. 65 �inhibitory phenol concentrations. Further, since near zero order (linear) rates can be expected to very low [0], R has been assumed representative of the overall rate for the batch begun at that concentration, with the following proviso: For these loading runs, the next batch run at "normal" loadings ( 0 o - 800 ppm) gave standard baseline [] rates, indicating that the phenomenon was transitory rather than lethal. of two hours. Between-batch processing time was on the order It may be that the overall rate inhibition would lift within a run, provided that the batch time were long enough. Batch time lengthens for larger V^ to Vp ratios. In such cases, R would not necessarily represent a close approximation of the run maximum, nor would it be zero-order. Without further experimentation it is difficult to assess the importance of the inhibition effect reported here. Studies reporting overall Monod kinetics have not described this phenomenon ( 37 , 46 , 57 ). Those demon- strating zero-order kinetics have generally not tested the upper limits of their ecosystems to the same extent, nor have their routine treatments of their test units been as consistently demanding as those utilized in this study ( 10 , 19 ). In any case, taking the volume of the Pilot Plant reactor to be 85L (3 ft3), removal rates of 2.7 g/1 day (169 lb/1000 ft3-day) have been routinely handled by this system under stated standard conditions at 20° C., with mixed liquor phenol concentrations of 1000 ppm for hydraulic loado ings between 3.3 and 6.7 gpm/ft . Within a range of 66 �~l£[0]_<1000 ppm the system kinetics are approximately zero-order, (Under counter-current aeration conditions, preliminary rates of 4.83 g/L-day--303 lb/1000 ft3-day—were achieved.) These results are comparable to those of pilot-scale activated sludge systems, which found that 120 to 250 lb/1000 ft3-day loadings were the highest practical without excessive foaming (5,14,15 ). 67 �5. Chromium Tolerance Perhaps the most serious aspect of the Kelly waste water is the high concentration of chromium in the toxic hexavalent state encountered even after washdown dilution. In a technical report ( 7 ), the Air Force Environmental Health Laboratory found that a maximum concentration of 0.14 mg'/L » (0.14 ppm) chromium in the waste influent to a biological treatment facility could be tolerated with a reduction efficiency of about 507o. Higher concentrations led to rapid poisoning of the . sludge mass. Currently, one of the better commercial pre-adapted bacterial preparations on the market (PHENOBAC), may be used only on waste water where hexavalent chromium is present at less than 2 ppm ( 17). KAFB stripping facility chromium concentration is routinely 20 times that (see Table 1 ). .Heavy metals toxicity tests on the J-series of organisms had earlier shown sensitivity only to silver ion (Ag**) ( 58 ). Inasmuch as the presence of chromium in the waste water never appeared to influence phenol utilization, and the Contract Statement of Work did not mention a concern for this parameter, i chromium tests were done only periodically during this study. The performance of the Coryneforms and the remainder of the • column ecosystem is dramatically in evidence in Table 8. These represent the only four runs on which chromium data was taken. Reductions of hexavalent chromium as high as 99.4% were achieved in 3.5 hours, with initial concentrations of 16.5 ppm. Performance varied, with a low of 36.4% reduction from 16.5 ppm; 68 �TABLE CHROMIUM 8, REDUCTION NOTE: All values in mg/1.* BATCH NUMBER CHROMIUM BATCH . START 1 TOTAL 2 +6 16,5 16,5 3 TOTAL +6 12,3 9,0 4 TOTAL +6 27,7 16,5 11,8 11,8 TOTAL +6 BATCH FINISH 7,2 0,1 4,3 0,1 22,0 111,5 1,8 1,8 BATCH TIME (HR) 3,5 3,5 3,5 ' 3 , 5 3,5 3,5 2,5 2,5 56,4 99,4 65,0 98,9 PERCENT REDUCTION 20,6 36,4 84,7 84,7 *Chromium assays performed by Chemical Processes Lab (MANCB) KAFB. Total chromium reported as tri + hexavalent. 69 �Che higher total chromium in Batch 3 may have influenced +6 reduction. A curious aspect of the test results is the apparent removal of chromium from the waste water. The columns developed a yellow coloration during the AWW runs; it is possible some chromium becomes locked up in the biomass. One would expect removal of chromium to decline with time to some equilibrium, balanced by sloughing of chromium-rich biomass into the effluent. Yet Batch 4, accomplished some 4 months after, the first three batches, demonstrates the highest total chromium removal. The metal sequestering mechanism.of the, Trinity filter ecosystem is an. intriguing puzzle which remains to be adequately investigated. STRRVflTION 6 DRY 20L C5:3GRU m BflSERHTE 2.73G/H-FT1 O 1ST FEEDING 1.33 -- • A 3d FEEDING 1.80 2 3 TIME (HOURS) FIGURE 19. Six day s t a r v a t i o n test 70 �6. Starvation Recovery Paint stripping at any facility is an intermittent process, and circumstances are conceivable during which phenol availability to the columns might be restricted for extended periods of time. Waste processing facilities exposed to intermittent phenolics input have historically presented severe lossof-acclimation problems to industry ( 5, 15). In keeping with the "ecosystem training" treatment philosophy, our standard procedure included routine overnight and weekend starvation periods. Except when runs happened to occupy these time periods, the filters were allowed to run out of phenol; recycle of spent waste was continued. Prior to experiments, test batches were run to check for base rate changes. During nine months of such experiments, no significant variations were found in base rate for even the first after-weekend runs. But to study the capacity of our ecosystem to withstand longer periodic substrate shortage, a Six Day Starvation Test was performed. The protocol for this set called for running a batch (20L) of AWW under standard operating procedures, to establish a column base rate (see Figure 19). After GC analysis confirmed less than 1 ppm residual phenol, this same spent waste water was recycled for an additional six days with no further replenishments of any kind. At the end of this time, two successive runs were . made in which nominal substrate loadings were applied in a mixed liquor volume of 20L. The column exhibited no lag time in phenol uptake, and was operating at approximately 50% of base 71 �rate on the first run. Within 48 hours column performance had returned to base rates, and computed V^ had reached 33L from a low of approximately 18L, indicating that the biomass had reached its normal plateau. A much more severe test was inadvertently imposed on Column 1. For more than six weeks, mechanical difficulties forced the complete shutdown of this unit. During this long period the system merely sat, liquid phase drained out, and slowly dehydrated. When started up once again, without reseeding, utilizing AWW and standard initial run conditions, a great deal of suspended solids was observed flushed into the liquid phase tank. However the phenol degradation rate was 40% of nominal (1.33 g/hr.ft^ overall), and three subsequent runs showed rapidly increasing overall rates with a return to baseline within 72 hours. Clearly the ecosystem established on the columns possesses a useful resilience, even in the face of unusually traumatic conditions, The linear kinetics of these runs were again at odds with data from standard waste treatment theory. To exist during the starvation period, column organisms would be forced to switch over to metabolism of other carbon sources. Two major pools of stored carbon were recognized: a) intracellular storage granules, such as the plentiful metaphosphate inclusions microscopically observed in the Coryneforms; and b) the extracellular polymeric secretions and dead cell lysates of the slime itself, 72 �Tests of pure culture Coryneforms recovered from the columns had shown a complete loss of phenol pathway enzymes after even short 2-generation exposures to rich carbon sources such as tryptic soy and/or glucose media (Egan, Olive and Burkholz, unpublished observations). However, the presence of high phenol-metabolizing enzyme levels was repeatedly verified in column slime material subjected to up to 8 weeks of continuous starvation. Figure 20 is a graphic interpretation of preliminary data from an allied investigation into the genetics of J-series organisms ( 59 ). Using a YSI Model 53 Biological Oxygen Monitor, suspensions of freshly obtained slime material, and of pure culture Coryneform bacteria recently isolated from the filters, were compared for their ability to use phenol as a. metabolic substrate. Each of the first two steps in the breakdown of phenol involves consumption of a molecule of oxygen (Figure 18); an increasing rate of disappearance of dissolved 02 in the test cell upon addition of phenol indicates the presence of the ring-fission enzymes. The figure shows the dramatic difference between phenol-induced and non-induced (tryptic-soy-ralsed) column J20s. Phenol-induced column J20s allowed to enter cryptic growth in a phenol-depleted SWW medium were non-induced for phenol. But column slime suspensions, even when starved for phenol for more than eight weeks, showed strong and immediate uptake of 02These results support the interpretation that one or more slime 73 �organisms possess "constitutive induction" ring-fission enzyme control. A discussion of the genetic as well as practical importance of these findings appears in Section IV.A. COLUMN SLIME .100 < STARVATION (iHR) STARVATION (6 WEEKS) 50 & 0 TIME PURE CULTURE, COLUMN J20/24 PHENOL-GROWN (IPHR) TRYPTIC SOY-GROWN (2HR) .100 £ 1 0 TIME ,100 > 0-GROWN,CRYPTIC GROWTH Q2HR> m (ER) ~ endogenous respiration I 50 V) (SM) - substrate metabolism CM O (0) - injection of phenol TIME FIGURE 20. O x y g e n c o n s u m p t i o n due to phenol m e t a b o l i s m for f i l t e r slime and p u r e c u l t u r e o r g a n i s m s grown on e i t h e r phenol or t r y p t l c soy. C u r v e s I l l u s t r a t e that phenolstarved filter s l i m e m a i n t a i n s high phenol enzyme levels over l o n g p e r i o d s o f t i m e . 7/4 �7. bed Volume and Depth EFfecLs (Tandem Mode) Effects of changes in reactor volume, V R , arid flow path length were studied by utilizing the "tandem run" mode of the triple tower Pilot Plant. In .the tandem mode, a single mixed liquor volume was cycled in a closed loop through all three towers (Figure 21). Comparison of the rate-per-unit-volume obtained at 3 ft o of .the packed bed volume (one-column rate) to that o obtained at 9 ft (three-column rate), shows a modest increase traced to slight differences in V^ between columns (see Figure 22). Here, the fluid streams were redistributed at each 3.83-ft interval by the splash heads; in an actual scale-up, the increased bed depth may interact with hydraulic factors in a way not completely predictable from the pilot scale. The pooling represented by the mixed liquor tank volumes probably had little effect on the observed rates, as we had demonstrated in prior runs. Our protocol of batch treatment, which conserved significant portions of the spent waste water as part of the, diluent for new concentrated waste input, may have been expected to lead to slight toxic by-product buildup in the mixed liquor. This might have accounted for the clear dichotomy in importance of the mixed liquor versus filter removal activities, except that, immediately following the periodic total replacement of V « . no significant differences 75 �r o ••-o r O r D FIGURE 21. Tandem mode operation. Arrows indicate flow direction of waste water.. 120 L was the total batch volume for both the tandem and isolated column run. RERCTOR VOLUME EFFECT Q 1 CO.. ROTE 2.90 G/H-FT3 O I OX. 3.50 A 3CO5. 3.EG 2 4 6 TIMECHOURS) FIGURE 22. Phenol degradation rates for 120 L runs in an Isolated column, versus rates in tandem mode. V^=35 L for each column in circuit. '/ft �in either rate or liquid phase growth were detected. During very long batch times, in cases where the mixed liquor volume Vj_ is rather large compared to VR, this circumstance might alter. One would expect more growth in Vj^ if it is left to progress long enough. But, if batch times are short, efforts to improve liquid phase phenol removal contributions by aeration may prove uneconomical and unnecessary. Implications for scaled-up batch and continuousflow systems are discussed in Section IV. C. T A B L E 9, RATE DATA FOR TEMPERATURE EXPERIMENTS . NOTE: All rates in grams /hour- ft COLUMN TEMP C 20 L 20 3,9 3,8 30 6,0 6,2 V0LUME 40 L 60 L 80 L 3,9 5,6 77 3 AVERAGE .RATES 3,9 5,9 �8. Temperature Studies Temperature conditions for the majority of experiments in this study were established by the technical tasks assigned under the Contract Statement of Work. The temperature of interest was 20° ±1° C. Previous studies by this laboratory ( 28 ) had established that test organisms originally isolated from the Kelly AFB facility were capable of growth and utilization of phenols at temperatures of 15° C., 25° C., and 35° C. the primary test temperature (20° C), reflects the mean annual temperature in the San Antonio area, and was chosen to evaluate pilot plant operation for that reason. To check expectations of performance under other ambient temperature conditions, this lab ran an additional series of studies with the current pilot plant at a temperature of 30° C. Comparison of results with those obtained at the. nominal annual mean are given in Table 9. Performance at 30° C. was approximately 1.8 times the average rate for 20° C., a result expected due to the well-known tendency for enzyme reaction velocities to double for each 10° C. rise ( 19 ). Implications of this data are clear. At.higher than mean temperatures, a trickling filter unit of this type would have an effectively increased efficiency of approximately RT - R™ i ion x (1.8) 20 78 \ 10 �where T is the new temperature in °C., R RT and is the rate at the new temperature, is the average rate experienced at 20° C., /T-20\ exp( iA ) is an exponential term indicating the fraction of a ten-degree-Centigrade interval represented by the temperature increase or decrease. At operating temperatures that are ten degrees below annual mean, the degradation rate for biological systems, might, therefore be expected to be about 44.570 of its value at the mean. There is much to be gained from maintaining warmer temperatures, and reactor design should incorporate suitable features. Temperature stabilization may be obtained by several means: insulation; locating compact columns within environmentally controlled areas; controlling liquid phase storage temperature; providing a cold-weather heating system for the units (e.g., a passive solar-type heat exchange system); or any combination thereof. 79 �9. Biomass Evolution Ordinary multiple-target waste processing reactors exhibit several stages of biomass development from startup through maturity (60 ). Beginning with representatives of relatively few microorganism species, colonization by fortuitous arrivals will commonly bring in bacteria, algae, fungi (including yeasts), protozoans, and finally many higher life forms including worms, arthropod larvae, snails, etc. It .might be expected that toxic waste dedicated-function reactors would.... show a much more restricted clientele. Initially, the filter biomass consisted entirely of bacteria. Upon dilution onto nutrient agar plates, only the original J20/24 mixture of Corynebacterium sp_. and invading Pseudomonas sp. were identified. Over sixty days, during f the synthetic waste •water experiments, . a distinct ; r; •'} , 4 ' :' ' : » •. segregation took place. The pseudomonads came to dominate the i.-'l'V ."• • • • mixed liquor phase, with the Coryneforms found principally on • ? t ,-• . y -• • .• the column and in the light foam atop the mixed liquor.. *. Shortly after the introduction of KAFB actual waste water, a colonizing species of Penicillium was identified. Additionally, the pseudomonad population in the liquid phase dropped dramatically. The cause of this decrease has not been discovered; it does not appear to be due to waste toxicity, since these pseudomonads (as well as other strains from laboratory stock) grow well on AWW agar plates in pure culture. It is possible that an antibiotic substance, perhaps produced 80 �by the fungus population on the columns, is limiting microbial growth in the liquid phase. After ninety days of operation on AWW, colonization by nematodes and mites was observed. The microscopic, transparent mites resembled Platyseius tenuipes, commonly found in activated sludge sewage processing plants. These mites survived in mixed slime cultures grown on 1000 ppm phenol agar plates. They appeared to thrive on bacterial polysacchari.de secretions adhering to the whiskerlike conidiophore processes of fungal growths. The nematodes, too, appeared immune to the •. phenolic environment, both in the filter beds and on agar plates. Microorganisms isolated from the very mature columns demonstrated curious interdependencies. The Coryneforms retained their overall genetic integrity, based on morphology, growth substrate preferences, and pigmentation. However, the only identifiable pseudomonad species recovered showed surprising fastidiousness, growing on phenol agar only in the presence of the Coryneform. By contrast, at least two varieties of Penicillium were isolated, one able to grow on all phenolic substrates tested (phenol and the cresols), the other specific for para-cresol. Fruiting bodies were in evidence on the fungi after 5 days' growth on phenol agar. These organisms have been maintained for further study. The original appearance of the biomass as it began to build in small catchments was a clumped olive drab floe on the young filter bed packing. 81 When changeover to AWW �was accomplished, it had already assumed a more filamentous texture. Within a month of exposure to the Kelly waste, its color had altered to a red-orange mottled yellow. Conceivably this change was due to the accumulation of chromium compounds in the slimes. Examination of bio-rings withdrawn from the columns has shown that the biomass is irregularly clumped on available surfaces, of a crumbly flocced texture, and is by no means "slimy". Electron microscopic observation reveals a surprisingly uniform ground forming a highly invaginated, almost honeycombed structure; a picture of a virtual column "sponge". Though the filters attained a predictably low species diversity, those organisms successfully colonizing the filters presented a strange assortment. In nontoxic multiple-target installations, the biomass ecology can be very critical to proper filter functioning ( 60 ). The exact contribution to stability by these dedicated function filter inhabitants, however, remains to be evaluated. �IV. DISCUSSION A. GENETIC ASPECTS OF DEDICATED FUNCTION REACTORS Originally,biological waste processing systems handled low-toxicity, multiple-target wastes. The fortuitous populations of such systems evolved genetic control mechanisms appropriate :o a rich environment where metabolic versatility was encouraged. Recently,when toxic wastes were introduced as additional targets ::or such '.systems, it was found that the nutritional hyperversatility of bacteria such as the soil pseudomonads made them especially useful as "buffer organisms" in a complex biomass, i able to switch to consuming the toxic substrate when its concentration rose, limiting its toxic effects on the whole population. Deliberately adapted and mutated, they became the first biotechnological tools of the waste water engineer, augmenting the fortuitous populations in common use. Still--as with nonaugmented reactor populations--problems of shock load shutdown and variable efficiency have persisted, apparently due to dilution or loss of critical genes in the biomass. Regular reseeding, coupled with designed increases in cell retention time—up to 50 days—have been necessary to counteract these problems ( 5 ). Point-source toxic wastes, however, present an unusual opportunity for bioprocessing not encountered in multiple-target facilities not even in the relatively spe- cialized industrial park waste processing plants now in existence. These limited-target streams are unique to the 83 �process generating them; the combination of toxic components, concentration ranges, and production schedule presents a specific, narrowly-defined genetic challenge to a biological system. Genetic response to that challenge is equally specific: Even uncontrolled toxic waste reactor populations show limited variety, reflecting the selection pressure of a hostile but constant environment. To seed a limited-target facility with artificially, genetically streamlined organisms is one obvious way to govern biomass development; but this does not carry the idea of controlled reactor populations to its most practical, logical conclusion. As indicated in Figure 23 , common toxic waste process protocols can actually select against those expensive lab-reared genes, simply by removing some of the selection pressures that spawned them. To make the most of the metabolic power than can be constructed in microorganisms today, the waste processing protocols must maintain those gene constellations by applying appropriate selective pressures. The more narrowly defined the waste, the more advantage may be taken of genetic streamlining. This is the utility of point-source treatment using a dedicated function reactor. Reactor conditions in continuous-flow systems can be described as environmentally stable. If microorganisms can take best advantage of a fixed environment, as suggested above, then unsteady-state batch treatment systems would seem to be at: odds with an effort to streamline and specialize the biomass �MASS BREEDING — SEEDING WASTE FACILITY PROTOCOL LABORATORY PROTOCOL DESELECTION ft WASTAGE SELECTION CONSERVATION WILD-TYPE WILD-TYPE > ISOLATION FIGURE 23. Inefficient coupling of laboratory selective protocol with non-selective, treatment facility application in standard systems. 85 �population, This might be true if "fixed environment" meant "steady-state". But there are good arguments that this is not the case. The object of dedicated function treatment is to metabolize specific toxic wastes and attain a desired discharge concentration. This means that, at some point in the process, the biomass will be consuming this waste at very low concentrations. In a typical continuous-flow system, process design maintains a low substrate concentration throughout the biomass Ai continuously. Influent lines may carry high substrate loads at .'•** • ; high concentrations, but : the actual levels delivered to the biomass are deliberately diluted down, using recycled effluent and/or some other diluent. In many waste facilities this protocol is an absolute necessity, because only a small portion of the biomass is made up of organisms actually able to consume .v:..-.,'."y>x;.v the substrate; and many of th'ese are not "switched on" to it • • • '. <••.<?.ft:f :"'''' both because it is dilute, -and because it is only one of many other target carbon sources'r This means that, if the enzyme systems responsible for oxidizing phenol follow Monod kinetics, then reactor efficiency at low concentrations may be significantly /-i . less than that possible att higher concentrations. In an activated sludge continuous-flow facility, the reactor may be operating near the half-velocity concentration most of the time; its removal efficiency per unit volume could be much less than maximal, By contrast, a batch treatment system--particularly 86 �where an active biomass is always available, as in a trickling filter bed--spends only a fraction of its operating time at the low concentration condition. Throughout most of the batch cycle, the biomass is exposed to high enough substrate levels to sustain maximal reaction rates. If Monod kinetics apply, rates will drop to levels similar to that encountered in continuous-flow only after most of the substrate load has been degraded. This implies that a smaller batch reactor and system volume could yield the same organic loading capacity as that provided by a larger activated sludge continuous-flow system. The above argument would not hold if rates were totally zero-order and of the same magnitude for both a batch and continuous-flow trickling filter system. Our study did in fact demonstrate near-zero order rates to very low phenol concentrations for the batch-operated Trinity trickling filters. This does not at all imply that the same kinetics would exist if the same system were operated in a continuous-flow mode (see Figure 24 ) : Since the effluent is maintained at some steady- state concentration, and the influent feed line is presumably equalized upstream of the filter, the filter is always presented with a constant, diluted input concentration while running. No depth of the bed would be operating under zero substrate concentration (in a high rate filter, the concentration strati- fication is not so marked as in a low rate filter), so about the same removal efficiency could be expected from all bed depths. However, the very constancy of the environment may 87 �lead to undesirable genetic alterations in the biomass. As mentioned earlier, the laboratory selection process usually involves both selection against organisms unable to tolerate high toxicant loadings, and selection for those which grow best on or use the most target substrate. In addition to selecting our column organisms in this manner, our protocol for the pilot plant also included regular starvation periods of 12 to 48 hours (overnight and weekends), coupled with the regular phenol shock loading that a 1000 ppm batch feeding entails. The resistance of the biomass to long periods of dehydration and complete starvation was directly influenced by this treatment. Possibly even the linear kinetics observed were also a result of such protocol-generated selective pressures. As was demonstrated in Section III.D.3., the phenol oxidizing enzyme activity of the filter slime did not decline even after many weeks of phenol starvation. This would be characteristic of mutant, "constitutive enzyme" genes--enzyme production genes whose induction, or "switching on," was no longer tied to the presence of phenol or a phenol metabolite. In other constitutive mutations it has been shown that enzyme production control has been shifted to some common cellular metabolic intermediate such as succinate, or an amino acid ( 61 ). in a dedicated function unit, this would mean that one could not shock the biomass with sudden nominal loadings, as may occur after even brief periods of target absence in typical sludge systems. Also, HS substrate (phenol) MM �c'.oncenLraLLon dropped during processing, Lhls decline would lead to the usual "switching off" of crucial enzyme production genes; and this would tend to erase the familiar "Monod drop" in rate at low substrate concentration. It is difficult to estimate the degree of change that might occur if our normal treatment protocol were to be permanently discontinued, but certainly the tendency would be for both shock load and starvation resistance to decline as critical population genetic configurations were allowed to mutate away from those originally selected under the laboratory protocol. This would be the situation in any facility where organisms were originated under one set of conditions, then left to operate under another. BATCH CONTINUOUS FLOW IP cl RECYClf EFFLUENT LINE INFLUENT (A) RECYCLE LINE INFLUENT (B) EFFLUENT TREATMENT RESERVOIR T REA E Nt RESERVOIR (DRAIN AND FILL) (PLUG FLOW KITH RECYCLE) FIGURE 24. B a t c h and c o n t i n u o u s f l o w s c h e m a t i c s f o r trickling f i l t e r s . Continuous flow applications may use e i t h e r i n f l u e n t i n j e c t i o n s i t e A or B. 89 �B. PROCESS PROTOCOLS 1. General The overall objective of this study was to evaluate a bioprocessing approach to handling the phenolic paint stripping waste generated at the Kelly AFB facility. To obtain this information, a batch approach was utilized for a number of reasons. a. We felt that: The application of genetic engineering to waste processing demanded efficient control of reactor environment in order to conserve desirable mutant traits. Resistance to starvation could best be maintained by continuing selection pressures for this trait in the operating filter biomass; shock load resistance would follow from regular exposure to sudden high loadings, in a manner similar to the isolation/selection protocols. > b. Batch pilot scale rate equations would be reasonably predictive of performance for large-scale reactors, provided good distribution was planned and the relationship between hydraulic loading and bed depth (mass velocity) on reaction rate was evaluated. As Rose (31 ) maintained, because of the complex fluid dynamics, direct scale-up should be attempted only if such a relationship could be estimated, or if the pilot plant was of the same bed depth as the proposed large •unit. This study evaluated rates at bed depths of 3.83 ft (11.67 dcm), 7.66 ft (23.34 dcm), and 11.49 ft (35.02 dcm) for constant hydraulic loadings, and found rates to remain constant �on a per-unit-of-volume basis (Section III.D.7.). c. ( 37 As has since been stated by Irvine and Richter ), the application possibilities of batch processing have been overlong neglected. 2. Batch At a time when discharge standards are tightening, a batch operated reactor has the advantage of controlled discharge quality. Provided production rate and reactor capacity are matched, effluent phenol concentrations are governed by batch processing time. When waste flows are irregular, as is often the case when control technologies are applied at the pointsource, batch operations are to be preferred. With a properly selected, conditioned biomass, as this study has shown, intermittent short-term shutdown can be accomplished with little effect on reactor efficiency, and with only mild transient effects even for long-term shutdowns. This indicates a much greater degree of resilience is possible than has generally been attributed to batch operations in the past. We believe this is a function of "teaching" the biomass how to cope with such challenges, by designing them into the protocol. The higher labor and handling costs associated with batch facilities can be lessened if modern process control technology is considered. A batch system built around reliable fluid switching, under electronic control, properly fail-safed and provided with back-up critical parts,will demand little 91 �more technical expertise to operate — and less human attention— than well-designed continuous-flow facilities. 3. Continuous-flow The kinetic data summarized in this report were derived using a batch reactor which operated in a low dispersion plug-flow regime. Grady (62 ), Smith ( 63 ), and others have stated that continuous-flow trickling filter installations could be designed based on batch data using plugflow relationships. Smith warns, however, that parameters affecting rate must be given appropriate values in the scaleup, and that it is frequently difficult to insure that conditions such as temperature and flow turbulence remain equivalent. For constant flow waste streams, or regularly periodic streams of predictable phenolic content, a continuousflow system has certain advantages. Traditionally this mode requires a low level of skilled supervision, and fluid switching is kept to a minimum. Parameters are chosen so that the average output of the reactor falls within guidelines. This means, of course, that ungoverned changes in input concentration or other factors can cause effluent to exceed regulation limits. If this is to be prevented, it is necessary to provide toxicant concentration test capabilities, and a means to adjust appropriate parameter values to bring effluent back within guidelines. A disadvantage of the continuous-flow process is, again, that the filter biomass is not prepared to deal with I. ramjl.pntn, and mny not bo oh IP to bundle- Forced abutdown periods 92 �as handily as the oscillation-acclimated batch system. 4. Design Equations a. Kinetics As discussed in Section III.D.4.C., the overall inhibiting effect of phenol on microbial metabolism in this system results in a zero-order batch rate that is based on the initial phenol concentration. Or, from equation () 6: R max 1 + (Ka/[0]0) + (C0]0/Ki)2 where Rtnax * 3.3 g/hr-ft3 Kg = 1 ppm K. - 2000 ppm. Temperature effects are related to rate by equation ( ) : 8 RT = RT x (1.8) A generalized rate expression for phenol, based on Kornegay's derivation ( 64 ) for Monod kinetics has been proposed below. Due to our inability to directly measure biomass produced in our system, we have been unable to verify the applicability of this model. However, its basic precepts do not violate what is known of system performance: R VR = -rA Dbdz, 93 (-9) �™ /Y KQ + L + [0] / a-x-d n 0 () ( n 1 R VR = rate of decay of 0 within the reactor; Dfc = depth of reactor bed; [0] = phenol concentration ( Os[0]slOOO ppm); K, Kg = Monod coefficients; L = diffusion coefficient of phenol; Y = growth-yield coefficient; x = mass of organisms per unit volume of active film; a = specific surface area of bed media; d = depth of the active zoogleal film. For a given medium, organism population, temperature and pH, waste water makeup, hydraulic loading rate and aeration sufficiency, all the terms become constant except for [ ] 0. We are currently working on a theoretical treatment relating the measurable parameter, V^, to biomass on the filter, hoping to find a useful replacement for the conceptual terms a - x - d , and possibly L, in Kornagey's expression. Design equations for batch, plug-flow, and plug-flow with recycle may be derived through a mass balance on phenol through the system: b. Batch For a batch system, one obtains nc l VM, l> ' -V"-'-- ( \ ~i- dC 01) �»b V ML VR R - hntch = mixed Liquor volume; = reactor volume; = rate of phenol degradation per unit volxime of reactor; Cj^ = initial concentration; Co = final concentration. I For zero-order kinetics, this reduces to V MT t. = —Sf_ A[0] RVR where (2 1) A[0] = desired change in effluent [ ] 0. In a system such as that depicted in Figure 24, when operated in a batch mode, the batch volume stored in the treatment reservoir, Vfi, may be defined as V B - Qp ^tb+tf+td+t8)' where ( 3 1 ) V = V ML: B Qp - raw waste production rate; t, = batch process time; tf = batch fill time; t, = batch drain time; ts = a safety allowance; R, V,,, and [0]O have their usual connotations. K The value of VJL,, the bound volume , may also have to be taken into consideration in a suitable algorithm if a microprocessor unit is monitoring both raw waste and batch Liquor [ ] in order to account for the initial rapid equilibration 0, 95 �effect described in Section III.D.2. If Vb is linearly related to biomass volume, it may be expected to be approximately equal to 4070 of the reactor volume, VK. Depending upon batch volume D in a significant parameter even in a scaledup batch reactor. c. Plug- flow Although untried for this system, a mass balance equation for a simple plug- flow reactor operating in continuous -flow mode is X - - dC (15) and for plug- flow with recycle, C VR " Cq R + q0) \ o 1 dC (16) q = raw waste influent flow rate; qR = recycle rate of effluent through filter; and all other variables have their usual meanings. This equation is sometimes expressed qR in terms of the "recycle ratio", . . . Part of the importance 96 �of this term derives from the loading limits of the reactor. For instance, in our system, 1000 ppm (Cmax) is the highest combined raw waste-plus-recycled effluent phenol concentration compatible with maximal filter rate. recycle ratio, This sets a minimal depending upon the maximal expected raw waste phenol concentration. In the case of the Kelly AFB facility, this would be about 4500 ppm (C^) ; and generally q R . (c i " Cmax) - C0) But, in addition, the recycle ratio required hydraulic loading. is also controlled by the Inasmuch as the primary effect of hydraulic loading in the trinity filters was to modulate the *i • transit volume, Vt, on the column—thus affecting mixed liquor retention time, and through that, the rate- -it is clear that some trade-offs are possible between designed retention volume • <••-•..". and the hydraulic regime. These manipulations were beyond the scope of the study. It would be safe to say that a much lower hydraulic loading than was actually used could be permitted, providing Vt was increased by structural design or counter- current ventilation scheme, and provision was made for the increased sensitivity to blockage that a lower flow implies. The preceding discussion applies to a continuous -flow system. For batch mode, equation (15) holds if C. = initial batch phenol concentration, and C = final batch phenol concentration. In an on-site pilot plant, provision for operation in both a batch and continuous -flow mode might be made for further evaluation of these parameters, 97 �V. CONCLUSIONS A. The Kelly AFB-ALC paint stripping facility waste stream can be efficiently handled by a biodegradative process at the point-source. B. Batch operations conducted over long periods of time appear to condition trickling filter biomass to both starvation and shock loading, eliminating the need for continuous reseeding by lab-reared stock. C. A batch-operated top-down air-fluid flow trickling filter, seeded with the Trinity University genetically tailored ecosystem of microorganisms, can routinely process initial phenol concentrations of up to 1000 ppm with an average removal rate of 3.3 g/hr-ft3 of reactor at 20°C., in the presence of total chromium concentrations ranging from 11.8 to 59.9 ppm. Effluent phenol concentrations of less than 1 ppm are achievable. D. Plastic bio-ring type support medium with a specific surface area a = 104 ft^/ft^ (diameter = 0.43 in.) provides anchorage for a dense biomass, imposing only coarse pretreatment filtration (#18 mesh). E. Filter bed depth scale-up to 11.5 ft should be possible if adequate waste stream dispersion and ventilation (0.22 cfm/ft^) are maintained. Hydraulic surface loadings, L g , between 3.4 - r\ 6.7 gpm/ft^ yielded comparable rates of phenol degradation; one rate-controlling factor influenced by Ls was the transit volume, Vt, on the column. Indications are that Vt-increasing configuration or flow-scheme changes, such as counter-current 98 �air-fluid flow, could do much to increase the already high performance of this type unit, c.ven at lower LS . F. Combined starvation and dehydration interludes supply- ing no inputs of any kind for more than six weeks were survived with rapid (72-hr) recovery to normal performance levels with neither special treatment nor reseeding. Routine 12-hr and 72-hr overnight and weekend 'down' periods were tolerated with no alterations in base rates. G. Operation at 20°C. (68°F.) gave average rates of 3.3 g/h-ft 3 with 99.9% phenol removal; operation at 30°C. (86°F) increased rates to 5.7 g/h-ft3 with 99.970 removal. H. All trace minerals"and some bulk requirements are satisfied by the combination of tap water and waste water components. A 0:P ratio of 1:48 and 0:N ration of 1:10 were adopted as sufficient supplementation for this pilot operation. KH2P04 and NH4N03 were used in this study; other .possible phosphorus or nitrogen sources include urea or diammonium phosphate (NH4)2P04. Analyses indicate that the Kelly waste water itself may provide some of the phosphorus requirement. I. In-system conditions should be maintained at pH 6-9, with an optimum range between 7-8. In an on-site system, buffer may be used, or pH governed by sensor-controlled addition of inexpensive NaOH or NH^OH. J. Batch operation of this system results in low, fine, suspended solids concentrations in the effluent (0.2 mg/ml = 200 ppm), indicating that a field system could dispense with 99 �even a simple sand clarlfier if discharge was to another waste facility. K. Reaction kinetics are approximately zero-order over the range l£[0]£l.OOO ppm; phenol inhibition results in a complex but near-zero-order, initial concentration-dependent rate for batch startup concentrations between 1000 and 3500 ppm. L. Standard steady-state continuous-flow operation of this process, if kinetics remained approximately zero-order,, would require a reactor of similar volume to the batch reactor. How- ever the kinetics of this study may be peculiar to a filter biomass utilized in a batch mode. Detailed design equations have been developed and are presented in Section IV.B.4. M. Both automated batch and continuous-flow systems are currently in use. Application of automated process control to the Kelly problem would mean a more efficient system with little change in personnel or skill levels required of the maintaining organization. 100 �VI. RECOMMENDATIONS A. It is recommended that; the Air Force consider implementation of a pilot scale on-site trickling filter bioprocessing system at the KAFB depaint facility. B. Economics of architecture and scale suggests structuring control design and pilot plant capacity to allow continued usefulness of pilot-scale facilities in a scale-up. C. Two moderate-size trickling filter units would permit flexibility of experimental design for evaluating batch, sequenced batch, and continuous-flow operation, as well as the testing of other parameters. Continuity of treatment during maintenance would also be possible. D. Figure 25 is a sketch of a possible on-site pilot- scale facility. For maximum benefits from the process, the final clarifier should be dispensed with, and effluent ducted via the existing channel to the Kelly AFB industrial waste facility. The influx of adapted organisms would provide the facility with useful genetic E. augmentation. Facility size and flow handling is a function of discharge requirements. Using a batch mode, effluent quality (phenol and chromium concentration) discharged to the industrial facility could be regulated at any convenient concentration, from less than 1 ppm upward. F. In view of the rapidly developing capabilities of automated process control technology, it may be worthwhile from the standpoint of both economics and system efficiency i to plan for its use even in such a low-flow ^application. 101 �FACILITY SUMP INDUSTRIAL HASTE FACILITY VIADUCT EQUALIZATION t STORAGE TANK f . 2. ..TREATMENT RESERVOIR FIGURE 25. Simplified on-site pilot plant arrangement for KAFB-ALC d e p a i n t f a c i l i t y . L i q u i d c i r c u i t s would a l l o w p a r a l l e l o r series o p e r a t i o n i n b a t c h , s e q u e n c i n g b a t c h , o r continuous-fl£»w mode. 102 �LITERATURE CITED 1. Perrotti, A.E. 1975. Activated carbon treatment of phenolic paint stripping wastewater. Facet Enterprises Industries, Inc., Final Report to Air Force Civil Engineering Center AFSC, AFCEC-TR-75-14. 2. Quality criteria for water. Protection Agency, 1977. 3. Wurm, H. J. 1968. The treatment of phenolic wastes. Purdue Industrial Waste Conference, Part 2, p. 1054,West Lafayette, Indiana. 4. Nijst, S. J. 1978. Treating aqueous effluents of the petrochemical industry. Shell International Chemi Maatschappij B. V. The Hague, Nederlands. 5. Lanouette, K. H. 1977. Treatment of phenolic wastes. Chemical Engineering Deskbook, October, pp.99-106. 6. Pettit, G. A. 1952. Chemical oxidation of phenolic wastes with chlorine—Discussion. Sewage and Industrial Wastes, 24:(6) 761. 7. Fishburn, G.A., R.A. Callahan, A.M. Elliott, and W.M. Melvin, Jr. 1970. Biotreatability and toxicity of selected phenolic paint strippers. USAF Environmental Health Laboratory (AFLC) Technical Report, EHL(K) 70-22. 8. Harrison, J.L. 1972. Biooxidation of oil refinery wastewater does work! Part 1., Winnipeg Refinery, 27th Annual Purdue Industrial Waste Conference. 9. Adams, C.E., Jr., R.M. Stein, W.W. Eckenfelder. Jr. 1974. Treatment of two coke plant wastewaters. to meet guideline criteria. Part 2. 29th Annual Purdue Industrial Waste Conference. U.S. Environmental 10. Sybron Corp. 1977. The BCA chemical waste treatment program. Biochemical Product Bulletin, Biochemical Corporation of America, Salem, Virginia. 11. Development document for effluent limitations guidelines for petroleum refining. U.S.Environmental Protection Agency, EPA 440/1-74-014-a, April 1977. 103 �12. Huber, L. 1967. Disposal of effluents from petroleum refineries and petrochemical plants. 22nd Annual Purdue Industrial Waste Conference. 13. Singleton, K.G. 1974. Developments in the biological treatment of waste water produced in synthetic resin manufacture at CIBA-Gcigy (U.K.) Ltd. Duxford, Cambridge. 14. Kostenbader, P.p., and J.W. Flecksteiner. 1969. Biological oxidation of coke plant weak ammonia liquor. J. Water Poll. Control Fed. 41:(2). 15. Nakashio, Makio. 1969. Phenolic wastes treatment by activated sludge process. J. Ferment. Techno1. 47: 389-393. 16. Radhakrishnana, I., andA.K.S. Ray. 1971. Activated sludge<-studies with phenol bacteria. J. Water Poll. Control. Fed. 43: (11). 17. Phenols. 1377A. 18. Sherrard, J.H. 1977. Kinetics and stoichiometry of completely mixed activated sludge. J. Water Poll. Control Fed. 49:19.68. 19. Me.tcalf and; Eddy, Inc. 1972. Wastewater Engineering: Collection, Treatment, Disposal. McGraw-Hill series in water resources'and environmental engineering, p. 377. McGraw-Hill. New York. 20. Rao, I.P.S. 1969. Pilot plant studies on treatment of phenolic wastes by high-rate deep trickling filture. Environmental Health, 11:366. 21. Stenquist, R.J., e_t al. 1977. Long-term performance of a coupled tri^klTng filter-activated sludge, plant.. J. Water Poll. Control Fed. 49:2265. 22. Lohmann, J.H. 1977. Possibilities of the one and two stage biological purification. Prog. Water Techno1. (G.B.) 8:229. 23. Wilhelmi, A.R., and R.B. Ely. 1977. The treatment of1 toxic industrial'wastes by a two step process. 30th Annual Purdue Industrial Waste Conference, Ext. Ser. 30:288. 1977. PHENOBAC. Corp. Technical Data Sheet 104 �24. KckonCelder, W.W. L978. Lecture noto.s on MoJ.o^l c;i I waste treatment. Summer Institute in Water Pollution Control, Manhattan College, N.Y.C., N.Y. 25. Holladay, D.W., C.W. Hancher, D.D. Chilcote, and C.D. Scott. 1976. Biodegradation of phenolic waste liquors in stirred-tank, columnar, and fluidized bed bioreactors. AIChE 69th Annual Meeting, Chicago. x 26. Bushnell, L.D., and H.F. Haas. 1941. The utilization of certain hydrocarbons by microorganisms. J. Bacteriol. 41:653-675. 27. Olive, W.E., Jr., H.D. Cobb, Jr., and R.M. Atherton. 1976. Biological treatment of cresylic acid laden waste water, pp. 381-388. In J.M. Sharpley and A.M. Kaplan (eds.), Proceedings of the Third International Biodegradation Symposium. Applied Science Publishers, Inc. London. 28. Olive, W.E., Jr. 1975. An ecological approach to the biological treatment of cresylic acid laden waste water. Masters' Thesis, Trinity University, San Antonio, Texas. 29. Dagley, S., and M.D. Patel. 1957. Oxidation of p-cresol and related compounds by a Pseudonpnas. Biochem. J. 66:227-233. 30. Ribbons, D.W. 1966. Metabolism of o-cresol by Pseudo- monas aeruginosa strain TL. J.Gen.Microbiol. 44: 221-231. 31. Rose, H. plants. 1978. Chemical reactor design for process McGraw-Hill, Inc. N.Y.C., N.Y. 32. Grady, C.P.L. 1977. Notes from Civil Engineering 452: Biochemical operations in environmental engineering. Purdue University. 33. Yano, T., and S. Koga. 1969. Biotechnological Bioengineering, 11:139. 34. Dennis, R.W., and R.L. Irvine. 1978. Effect of fill to react ratio on sequencing batch biological reactors. J. Water Poll. Control Fed. 105: 35. Irvine, R.L., and W.B. Davis. 1971. Use of sequencing batch reactors for waste treatment--CPC International, Corpus Christi, Texas. 26th Annual Purdue Industrial Waste Conference. 105 �36. Irvine, R.L., T.P. Fox, and R.O. Richter. 1977. Investigation of fill and batch periods of sequencing batch biological reactors. Water Research, 11:713-717 37. Irvine, R.L., and R.O. Richter. 1978. Comparative evaluation of sequencing batch reactors. J. Environ. Engineering Div., ASCE 104:503-514. 38. Horn, F.J., and R.C. Lin. 1967. Periodic processes: a variational approach. Industrial and Engineering Chemistry: Process Design and Development, 6:21-30. 39. Douglas, J.M. 1967. Periodic reactor operation. Industrial and Engineering Chemistry Process Design and Development, 6:43-48. 40. Selna, M.W., and E.D. Schroeder. 1978. Response of activated sludge processes to organic transients-Kinetics. J. Water Poll. Control Fed. 104:944-956. 41. Eckhoff, D.W., and D.I. Jenkins. 1966. Transient loading effects in the activated sludge process. Third International Conference on Water Poll. Res. Munich. 42. McLellan, J.C., and A.W. Busch. 1969. Hydraulic and process aspects of reactor design. II: Response to variations. 24th Annual Purdue Industrial Waste Conference. 43. Storer, F.F., and A.F. Gaudy. 1969. Computational analysis of transient response to quantitative shock loadings of heterogeneous populations in continuous cultures. Environ. Sci. and Tech. 3:143. 44. Adams, C.E., and W.W. Eckenfelder. 1970. Response of activated sludge to organic transient loadings. J. San. Eng. Div. ASCE 96:(SA2) 333. 45. Sherrard, J.H., and A.W. Lawrence. 1975. Response of activated sludge to step increases in loading. J. Water Poll. Control Fed. 47:1848. 46. Atkinson, B., and A.J. Knights. 1975. Microbial film fermenters: their present and future applications. Bioengineering Report, Biotechnol. Bioeng. 17:12451267. 47. Riltmann, B.E., and.P.L.McCarty. 1978. Variable-order model of bacLcrLal-fL1m kinetics. J. Environ. Enp,. Div. ASCE 104:889-900. 106 �48. Egan, J.W., H.D. Cobb, Jr., W.E. Olive, Jr., and D. Hansen. Unconventional phenolics degradation kinetics in a dedicated-function biodegradation facility. (In preparation.) To be presented to 1979 Annual International Conference of the American Society for Microbiology, Los Angeles. 49. Mueller, J. 1978. Lecture notes on biological waste treatment. Summer Institute in Water Pollution Control, Manhattan College. N.Y.C., N.Y. 50. Harremoes, P. 1976. The significance of pore diffusion to filter denitrification. J. Water Poll. Control Fed. 48:377. 51. Klemetson, S.L. 1978. Biological filters. Lit. Review, pp. 1075-1076. J. Water Poll. Control Fed., June. 52. Wehner, J.F., and R.H. Wilhelm. 1958. Boundary conditions of flow reactor. Chem. Eng. Sci. 6:89. 53. Levenspiel, 0. 1962. Chemical Reaction Engineering. Wiley and Sons, N.Y. 54. Monod, J. 1949. The growth of bacterial cultures. Annu. Rev. Microbiol. 3:371-394. 55. Yang, R.D., and A.E. Humphrey. 1975. Dynamic and steady state studies of phenol biodegradation in pure and mixed cultures. Biotechnol. and Bioeng. 17:1211-1235. 56. Wood, W.B., J.H. Wilson, R.M. Benbow, and L.E. Hood. 1974. Biochemistry: A Problems Approach. W.A. Benjamin, Inc., Menlo Park, California. 57. Williamson, K., and P.L. McCarty. 1976. Verification studies of the biofilm model for bacterial substrate utilization. J. Water Poll. Control Fed. 48:231. 58. Mitchell, C.D. 1974. An ecological approach to the biological oxidation of cresylic acid, Phase I: isolation and identification of the bacteria involved in the degradation of cresol. Masters' Thesis, Trinity University, San Antonio, Texas. 59. Cobb, H.D., Jr., W.E. Olive, Jr., and J.W. Egan. Investigations into the genetic basis of phenolic waste biodegradation. Final Report on AFOSR Grant No. 76-2875D. In preparation. 107 �60. Hawkes, H.A. 1963. The Ecology of Waste Water Treatment. Permagon Press, The MacMillan Company, N.Y. 61. Novick, A., and M. Weiner. 1957. Proceedings of the National Academy of Sciences U.S.A. 43:553. 62. Meagher, R.B., G.M. McCorkle, M.K. Ornston, and L.N. Ornston. 1972. J. Bacteriol. 111:465. 63. Grady, C.P.L., Jr. Simplified optimization of activated sludge-process. J. Environ. Eng. Div., Proc. AMCE, 103:413. 64. Kornegay, B.H. 1969. Characteristics and kinetics of fixed film biological reactors. Final Report FWPCA Research Contract WPO 1181, Clemson University, South Carolina. BIBLIOGRAPHY A literature search and critical analysis of biological trickling filter studies--Vol. 1, Dow Chemical Company, Water Poll. Control Res. Series, U.S.E.P.A., Dec. 1971, Balakrishnan> S., W^W. Eckenfelder, and' C. Brown. 1969. Organic removal by a selected trickling filter media. Water and Wastes Engineering, 6:A-22 to A-25. Barker, A.N. 1949. Some microbiological aspects of sewage purification. J. Inst. Serv. Purif. 1:7-22. Behn, V.C., and P. Monadjemi. 1968. Developments in biological filtration. Advances in Water Quality Improvement, Vol. I. E.F. Gloyna and W.W. Eckerifelder (eds.).University of Texas Press, Austin: 204-214. Brewer, W.E., D. Bird, M.A. Struck, Jr., E.W. Poth, Jr., and W.W. Melvin, Jr. 1968. Toxicity of untreated Kelly AF-B sewage to fathead minnows (Pimephales promelas) . REHL (K) 68-20. Buhr, H.O., e_t al. 1977. Research needs for automation of wastewater treatment systems. U.S. Nat'l. Tech. Inform. Serv. PB-262816/2GA; Gov.- Rep. Announce. 77:7. 108 �'Chakrabarty, A.M. (ed.) 1978. Genetic Engineering. CRC Press. » Charaklis, W.G. 1978. Microbial reaction rate expressions. J. Environ. Eng. Div., Proc. ASCE 104:531-534. Cook, E.E., and L.P. Herning. 1978. Shock load attenuation trickling filter. J. Environ. Eng. Div. ASCE 104:461-469. Daigger, G.T. , and C.P.L. Grady, Jr. 1977. Knctor.s aFrect.i.np. effluent quality from fill-and-draw activated sludge reactors. J. Water Poll. Control Fed. 49:2390. Eckenfelder, W.W. 1961. Trickling filtration design and performance. J. Sanitary Eng. Div. ASCE 87:(SA4) 33-45. Flanagan, M.J., e_t al. 1977. Automatic dissolved oxygen control. J. Environ. Eng. Div., Proc. ASCE 103:707. Haller, H.D. 1978. Degradation of mono-substituted benzoates and phenols by wastewater. J. Water Poll. Control Fed. 2771-2777. Hawkes, H.A. 1961. bed problems. An ecological approach to some bacteria J. Inst. Serv. Purif. 2:105-132. Heller, A.W. 1957. Some factors in the selection of a phenol recovery process. Proc. 12th Annual Purdue Industrial Waste Conference, pp. 103-122. Helmers, E.N., J.D. Frame, A.E. Greenburg, & C.N. Sawyer. 1952. Nutritional requirements in the biological stabilization of industrial wastes; III--Treatment with supplementary nutrients. Sewage Industr. Wastes, 24:496-507. Hsu, K.H., L.E. Erickson, and L.T. Fan. 1977. Pressure drop, gas hold-up, and oxygen transfer in tower systems. Biotechnol. Bioeng. 19:247-265. Ingold, R.S. 1940. Oxidation-reduction enzymes in activated sludge. Sewage Wks. J. 12:862-874. Interaction of heavy metals & biological sewage treatment processes. 1969. Public Health Service publication No. 999-WP-22, Division of Water Supply & Pollution Control, U.S. Dept. of Health, Education & Welfare. 109 �Kroop, R.H. 1973. Treatment of phenolic aircraft paint stripping wastewater. 28th Annual Purdue Industrial Waste Conference, Part 2. Lloyd, LI. 1944. The sewage bacteria bed fauna in its natural setting. Nature (London) 154:397. Molvar, A.E. 1977. Selected applications of instrumentation & automation in wastewater treatment facilities. U.S. Nat'l. Tech. Inform. Serv. PB-263777; Gov. Rep. Announce. 77:10. Nebel, C. f R.D«. Gottschling, J.L. Holmes, &P.C. Urangst. 1976. Ozone oxidation of phenolic effluents. Welsbach Ozone Systems, Corp. Neufeld, R.D., et al. 1977. A kinetic model & equilibrium relationsHTp~For heavy metal accumulation. J. Water Poll. Control Fed. 49:489. Petru, I.A. 1958. Temperature & air flow in filters, significance in trickling filter efficiency. Contact. Rec. 69:15-21. Fitter, P. 1976. Determination of biological degradability of organic substances. Water Research, 10:231. Schroeder, E.D. 1975. The relationship between process configuration and process performance. J. Water Poll. Control Fed. 47:1005-1011. Schulze, K.L. 1960. Load and efficiency of trickling filters. J. Water Poll. Control Fed. 32:(3) 245-261. Show-Jong Yeh and C.R. Jenkins. 1978. Pure oxygen fixedfilm reactor. J. of Environ. Eng. Div. ASCE 104:611-623. Smith, R.E., et al. 1977. Phosphorus conservation in a contact stabilization system. Proc. 30th Annual Purdue Industrial Waste Conference, Ext. Ser. 30:666. Tischler, L.F., and W.W. Eckenfelder. 1969. Linear substrate removal in the activated sludge process. Advances in Water Poll. Res. 361. Permagon Press, Oxford, U.K. Winter, T.H. 1973. Economic evaluation of phenolic waste treatment systems. International Water Conference Institute of Engineers. 110 �INITIAL DISTRIBUTION LIST DDC/DDA HQ AFSC/DL HQ AFSC/SD HQ USAF/RDPS HQ USAF/LEEV HQ USAF/SGPA SAF/MIQ SAF/OI AFIT/library AFIT/DE EPA/ORD EPA Industrial Environmental 2 1 1 1 1 1 1 1 1 1 1 1 Research Laboratory USAMB-RDL 1 US Army/CERL . USA Chief, R&D/EQ USN Chief, R&D/EQ NCEL, Code 25111 HQ AUL/LSE 71-249 HQ USAFA/Library OEHL/CC USAF Hospital, Wiesbaden OL-AD; USAF OEHL OUSDR&E HQ AFESC/TST HQ AFESC/DEV HQ AFESC/RDVW AFRCE/WR AFRCE/CR AFRCE/ER AFOSR/CC AFLC/DE AFLC/MA WR-ALC/DE OC-ALC/DE SM-ALC/DE O-ALC/DE 1 1 1. 1 1 1 1 1 1 1 1 1 10 1 1 1 1 1 1 1 1 1 1 Dept of Biology Trinity University 7 USAF Hospital BEALE/SGPM USAF OEHL/EC HQ AFTEC/SGB 1 1 1 111 (The reverse of this page is blank) � Dublin Core The Dublin Core metadata element set is common to all Omeka records, including items, files, and collections. For more information see, http://dublincore.org/documents/dces/. Title A name given to the resource Alvin L. Young Collection on Agent Orange Description An account of the resource <p style="margin-top: -1em; line-height: 1.2em;">The Alvin L. Young Collection on Agent Orange comprises 120 linear feet and spans the late 1800s to 2005; however, the bulk of the coverage is from the 1960s to the 1980s and there are many undated items. The collection was donated to Special Collections of the National Agricultural Library in 1985 by Dr. Alvin L. Young (1942- ). Dr. Young developed the collection as he conducted extensive research on the military defoliant Agent Orange. The collection is in good condition and includes letters, memoranda, books, reports, press releases, journal and newspaper clippings, field logs and notebooks, newsletters, maps, booklets and pamphlets, photographs, memorabilia, and audiotapes of an interview with Dr. Young.</p> <p>For more about this collection, <a href="/exhibits/speccoll/exhibits/show/alvin-l--young-collection-on-a">view the Agent Orange Exhibit.</a></p> Text A resource consisting primarily of words for reading. Examples include books, letters, dissertations, poems, newspapers, articles, archives of mailing lists. Note that facsimiles or images of texts are still of the genre Text. Box The box containing the original item. 012 Folder The folder containing the original item. 0105 Series The series number of the original item. Series II Dublin Core The Dublin Core metadata element set is common to all Omeka records, including items, files, and collections. For more information see, http://dublincore.org/documents/dces/. Creator An entity primarily responsible for making the resource Cobb, H.D. John W. Egan William E. Olive, Jr. Daniel J. Hansen Description An account of the resource <strong>Corporate Author: </strong>Laboratory of Cellular Physiology, Trinity University, San Antonio, TX Date A point or period of time associated with an event in the lifecycle of the resource 1979-05-01 Title A name given to the resource Biodegradation of Phenolic Paint-Stripping Waste: Laboratory Evaluation of a Fixed Film Batch Reactor Subject The topic of the resource toxic wastes